diff --git a/.env.example b/.env.example
deleted file mode 100644
index e66a4d53..00000000
--- a/.env.example
+++ /dev/null
@@ -1 +0,0 @@
-VITE_API_URL=http://127.0.0.1:5001/api
\ No newline at end of file
diff --git a/README.md b/README.md
index 9b8f6404..1e10674e 100644
--- a/README.md
+++ b/README.md
@@ -56,6 +56,71 @@ npm run dev
 
 This command will run the React app in development with Vite.js script
 
+## Rodando  o backend Flask
+
+No powerShell mude para o diretório backend
+```
+cd backend
+```
+Crie arquivos __init__.py em todos diretórios dentro de backend, inclusive nele.
+
+backend/
+   __init__.py
+   main.py
+   testes/
+      __init__.py
+      teste_exemplo.py
+
+MacOS
+Dentro da pasta backend no terminal, execute os seguintes comandos:
+
+```
+export PYTHONPATH=$(pwd)/backend
+export FLASK_APP=main
+flask run
+```
+
+Dentro da pasta backend/testes, execute:
+
+```
+export PYTHONPATH=$(pwd)/backend
+pytest
+Windows (Powershell)
+```
+Dentro da pasta backend no PowerShell, execute os seguintes comandos:
+
+```
+$env:PYTHONPATH = "$(pwd)/backend"
+$env:FLASK_APP = "main"
+flask run
+```
+
+
+Crie a venv
+```
+python -m venv venv
+```
+Ative a venv
+```
+venv\Scripts\activate
+```
+e agora você deve instalar todas as dependencias do projeto 
+
+```
+pip install -r requirements.txt
+```
+
+obs: caso sejam necessárias novas bibliotecas e deseja atualizar o requirements.txt por favor recirar arquivo.
+```
+pip freeze > requirements.txt
+```
+
+Rodando servidor Flask localmente na rota 5000.
+```
+python main.py
+```
+
+
 ## Running the tests
 
 There are two types of tests configured in the base project: unit tests using Vitest with React Testing Library and E2E acceptance tests using Cypress with Cucumber. It's interesting to create the **.env.testing** at the root of the project the same way it was created to run in development, changing the necessary values.
diff --git a/backend/__init__.py b/backend/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/__pycache__/__init__.cpython-312.pyc b/backend/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..75c33608
Binary files /dev/null and b/backend/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/__pycache__/blueprints.cpython-312.pyc b/backend/__pycache__/blueprints.cpython-312.pyc
new file mode 100644
index 00000000..5e0c7657
Binary files /dev/null and b/backend/__pycache__/blueprints.cpython-312.pyc differ
diff --git a/backend/__pycache__/blueprints.cpython-313.pyc b/backend/__pycache__/blueprints.cpython-313.pyc
new file mode 100644
index 00000000..124f8b20
Binary files /dev/null and b/backend/__pycache__/blueprints.cpython-313.pyc differ
diff --git a/backend/__pycache__/config.cpython-312.pyc b/backend/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..9d69a1d8
Binary files /dev/null and b/backend/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/__pycache__/config.cpython-313.pyc b/backend/__pycache__/config.cpython-313.pyc
new file mode 100644
index 00000000..c48abe4e
Binary files /dev/null and b/backend/__pycache__/config.cpython-313.pyc differ
diff --git a/backend/__pycache__/main.cpython-312.pyc b/backend/__pycache__/main.cpython-312.pyc
new file mode 100644
index 00000000..d509dd82
Binary files /dev/null and b/backend/__pycache__/main.cpython-312.pyc differ
diff --git a/backend/__pycache__/main.cpython-313.pyc b/backend/__pycache__/main.cpython-313.pyc
new file mode 100644
index 00000000..aa3ab5e9
Binary files /dev/null and b/backend/__pycache__/main.cpython-313.pyc differ
diff --git a/backend/blueprints.py b/backend/blueprints.py
new file mode 100644
index 00000000..5430f27f
--- /dev/null
+++ b/backend/blueprints.py
@@ -0,0 +1,20 @@
+from backend.rotas.login import login_bp
+from backend.rotas.cadastro import cadastro_bp
+from backend.rotas.logout import logout_bp
+from backend.rotas.criar_solicitacao_manutencao import criar_manutencao_bp
+from backend.rotas.criar_solicitacao_recursos import criar_recursos_bp
+from backend.rotas.excluir_solicitacao_manutencao import excluir_manutencao_bp
+from backend.rotas.excluir_solicitacao_recursos import excluir_recursos_bp
+from backend.rotas.editar_solicitacao_manutencao import editar_manutencao_bp
+from backend.rotas.editar_solicitacao_recursos import editar_recursos_bp
+from backend.rotas.reservas import reservas_bp
+from backend.rotas.salas import salas_bp
+from backend.rotas.usuario import usuarios_bp
+
+# Registra os Blueprints
+def registrarBlueprints(app):
+    blueprints = [
+       login_bp, cadastro_bp, logout_bp, criar_manutencao_bp, criar_recursos_bp, excluir_manutencao_bp, excluir_recursos_bp, editar_recursos_bp, editar_manutencao_bp
+    ]
+    for blueprint in blueprints:
+        app.register_blueprint(blueprint)
\ No newline at end of file
diff --git a/backend/config.py b/backend/config.py
new file mode 100644
index 00000000..f7e562c4
--- /dev/null
+++ b/backend/config.py
@@ -0,0 +1,7 @@
+import os
+
+class Config:
+    SECRET_KEY = os.urandom(24)
+    SQLALCHEMY_DATABASE_URI = "sqlite:///users.db"
+    SQLALCHEMY_TRACK_MODIFICATIONS = False
+
diff --git a/backend/features/cadastroServico.feature b/backend/features/cadastroServico.feature
new file mode 100644
index 00000000..b3ac6a2c
--- /dev/null
+++ b/backend/features/cadastroServico.feature
@@ -0,0 +1,100 @@
+Feature: Serviço de Cadastro de Usuários
+
+  Scenario: Sucesso no cadastro de usuário
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Demosténes"
+    And ele informa o CPF "126.455.789-00"
+    And ele informa o email "demostenessouza@example.com"
+    And ele informa se é professor "N"
+    And ele informa a senha "SecurePassword123"
+    And ele informa a confirmação da senha "SecurePassword123"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Cadastro criado com sucesso!"
+    And o status code deve ser "201"
+
+  Scenario: Sucesso no cadastro de professor
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Tatiana"
+    And ele informa o CPF "331.879.789-33"
+    And ele informa o email "vanessasilva@example.com"
+    And ele informa se é professor "S"
+    And ele informa o SIAPE "101110"
+    And ele informa a senha "12345678"
+    And ele informa a confirmação da senha "12345678"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Cadastro criado com sucesso!"
+    And o status code deve ser "201"
+
+  Scenario: Fracasso no cadastro por campos obrigatórios não preenchidos
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "João"
+    And ele informa o CPF "987.654.321-00"
+    And ele informa o email "joao@iat.com"
+    And ele informa se é professor "N"
+    And ele informa a senha "Password123"
+    And ele deixa o campo "Confirmar Senha" com ""
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Confirmar Senha é obrigatório."
+    And o status code deve ser "400"
+
+  Scenario: Fracasso no cadastro por duplicação de ID única
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Carlos Mendes"
+    And ele informa o CPF "126.456.789-00"
+    And ele informa o email "demostenes@example.com"
+    And ele informa se é professor "S"
+    And ele informa o SIAPE "010101"
+    And ele informa a senha "Password456"
+    And ele informa a confirmação da senha "Password456"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Erro: email/cpf já está registrado."
+    And o status code deve ser "409"
+
+  Scenario: Fracasso no cadastro por senhas que não coincidem
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Beatriz"
+    And ele informa o CPF "789.456.123-00"
+    And ele informa o email "Beatriz.oliveira@example.com"
+    And ele informa se é professor "N"
+    And ele informa a senha "MyPassword123"
+    And ele informa a confirmação da senha "DifferentPassword123"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "As senhas não coincidem."
+    And o status code deve ser "400"
+
+    Scenario: Fracasso no cadastro por formato inválido de email
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Lucas"
+    And ele informa o CPF "987.654.321-00"
+    And ele informa o email "lucas.example.com"
+    And ele informa se é professor "N"
+    And ele informa a senha "SenhaForte123"
+    And ele informa a confirmação da senha "SenhaForte123"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Formato de email inválido. Use um email válido, como exemplo@dominio.com."
+    And o status code deve ser "400"
+
+  Scenario: Fracasso no cadastro por formato inválido de CPF
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Fabricio"
+    And ele informa o CPF "123"
+    And ele informa o email "fabricio@example.com"
+    And ele informa se é professor "N"
+    And ele informa a senha "SenhaSegura456"
+    And ele informa a confirmação da senha "SenhaSegura456"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "CPF inválido. Digite um CPF válido no formato XXX.XXX.XXX-XX."
+    And o status code deve ser "400"
+  
+  Scenario: Fracasso no cadastro por siape já registrado
+    Given o usuário deseja se cadastrar
+    When ele informa o nome "Max"
+    And ele informa o CPF "987.654.321-00"
+    And ele informa o email "max@gmail.com"
+    And ele informa se é professor "S"
+    And ele informa o SIAPE "101010"
+    And ele informa a senha "Senha123"
+    And ele informa a confirmação da senha "Senha123"
+    And ele envia uma requisição POST para "/api/cadastro"
+    Then a resposta deve conter a mensagem "Erro: siape já está registrado."
+    And o status code deve ser "409"
\ No newline at end of file
diff --git a/backend/features/loginServico.feature b/backend/features/loginServico.feature
new file mode 100644
index 00000000..aa132915
--- /dev/null
+++ b/backend/features/loginServico.feature
@@ -0,0 +1,21 @@
+Feature: Autenticação de usuários
+
+  Scenario: Sucesso no login
+    Given o usuário possui o email "demostenes@example.com" e a senha "SecurePassword123" válidos
+    When ele envia uma requisição POST para "/api/login" com os dados "demostenes@example.com" e "SecurePassword123"
+    Then a resposta deve conter os dados do usuário
+    And o status code deve ser "200"
+
+  Scenario: Fracasso no login por senha incorreta
+    Given o usuário possui o email "demostenes@example.com" válido e a senha "SecureIncorreta123" inválida
+    When ele envia uma requisição POST para "/api/login" com os dados "demostenes@example.com" e "SecureIncorreta123"
+    Then a resposta deve conter a mensagem "Usuário ou senha inválidos."
+    And o status code deve ser "401"
+
+  Scenario: Fracasso no login por falta de email ou senha
+    Given o usuário envia uma requisição sem email ou senha
+    When ele envia uma requisição POST para "/api/login" com os dados " " e " "
+    Then a resposta deve conter a mensagem "Usuário e senha são obrigatórios."
+    And o status code deve ser "400"
+
+ 
\ No newline at end of file
diff --git a/backend/features/manutencao.feature b/backend/features/manutencao.feature
new file mode 100644
index 00000000..9121bc97
--- /dev/null
+++ b/backend/features/manutencao.feature
@@ -0,0 +1,28 @@
+Feature: Criar/remover solicitação de manutenção de sala
+
+    Scenario: sucesso ao criar uma solicitação de manutenção para uma reserva concluída
+    Given o professor possui uma reserva de sala reserva_id "1" que já foi encerrada
+    When ele envia uma requisição POST /solicitacoes/manutencao com os dados: reserva_id: "1", descricao: "Mesa quebrada."
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de manutenção foi criada!" e um status "201"
+    And a reserva reserva_id: "1" possui uma solicitação de manutenção com descricao: "Mesa quebrada."
+
+    Scenario: fracasso ao criar uma solicitação de manutenção sem preencher o campo descricao
+    Given o professor possui uma reserva de sala reserva_id "1" que já foi encerrada
+    When ele envia uma requisição POST /solicitacoes/manutencao com os dados: reserva_id: "1", descricao: " "
+    Then o sistema retorna "erro" "O campo 'descricao' não pode estar vazio." e um status "400"
+
+    Scenario: sucesso ao editar uma solicitação de manutenção existente
+    Given o professor já criou uma solicitação de manutenção associada a reserva_id "1"
+    When ele envia uma requisição PUT /solicitacoes/manutencao/"1" contendo o ID da solicitação de manutenção e a alteração descricao: "Mesa e cadeira quebradas."
+    Then o sistema retorna "mensagem" "Solicitação de manutenção atualizada com sucesso" e um status "200"
+    And o sistema atualiza os detalhes da solicitação com descricao: "Mesa e cadeira quebradas."
+
+    Scenario: fracasso ao editar solicitação de manutenção sem preencher o campo descricao
+    Given o professor já criou uma solicitação de manutenção associada a reserva_id "1"
+    When ele envia uma requisição PUT /solicitacoes/manutencao/"1" contendo o ID da solicitação de manutenção e a alteração descricao: " "
+    Then o sistema retorna "erro" "A descrição da manutenção não pode ser vazia" e um status "400"
+
+    Scenario: sucesso ao excluir uma solicitação de manutenção existente
+    Given o professor já criou uma solicitação de manutenção associada a reserva_id "1"
+    When ele envia uma requisição DELETE /solicitacoes/manutencao/"1",
+    Then o sistema remove a solicitação do banco de dados e retorna um status "204"
\ No newline at end of file
diff --git a/backend/features/recursos.feature b/backend/features/recursos.feature
new file mode 100644
index 00000000..3861f897
--- /dev/null
+++ b/backend/features/recursos.feature
@@ -0,0 +1,57 @@
+Feature: Criar/remover solicitação de recursos de sala
+
+    Scenario: sucesso ao criar solicitação de recursos para uma reserva ativa com todos os campos preenchidos
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: "Projetor, Teclado", itens_nao_listados: "Extensão elétrica", observacoes: "Para aula prática"
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de recursos foi criada!" e um status "201"
+    And a reserva_id "1" possui uma solicitação com recursos "Projetor, Teclado", itens_nao_listados "Extensão elétrica" e observacoes "Para aula prática"
+
+    Scenario: fracasso ao criar solicitação de recursos com campos preenchidos com espaços ou não preenchidos
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: " ", itens_nao_listados: " ", observacoes: " "
+    Then o sistema retorna "erro" "Você deve selecionar um recurso ou especificar itens não listados." e um status "400"
+
+    Scenario: fracasso ao criar solicitação de recursos com apenas o campo observacoes preenchido
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: " ", itens_nao_listados: " ", observacoes: "Para aula prática"
+    Then o sistema retorna "erro" "Você deve selecionar um recurso ou especificar itens não listados." e um status "400"
+
+    Scenario: sucesso ao criar solicitação de recursos sem preencher o campo itens_nao_listados
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: "Projetor, Teclado", itens_nao_listados: " ", observacoes: "Para aula prática"
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de recursos foi criada!" e um status "201"
+    And a reserva_id "1" possui uma solicitação com recursos "Projetor, Teclado", itens_nao_listados " " e observacoes "Para aula prática"
+
+    Scenario: sucesso ao criar solicitação de recursos sem preencher o campo itens_nao_listados e o campo observacoes
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: "Projetor, Teclado", itens_nao_listados: " ", observacoes: " "
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de recursos foi criada!" e um status "201"
+    And a reserva_id "1" possui uma solicitação com recursos "Projetor, Teclado", itens_nao_listados " " e observacoes " "
+
+    Scenario: sucesso ao criar solicitação de recursos com o campo de recursos vazio e apenas especificando os itens não listados
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: " ", itens_nao_listados: "Extensão elétrica", observacoes: " "
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de recursos foi criada!" e um status "201"
+    And a reserva_id "1" possui uma solicitação com recursos " ", itens_nao_listados "Extensão elétrica" e observacoes " "
+
+    Scenario: sucesso ao criar solicitação de recursos com o campo de recursos vazio e especificando os itens não listados e as observacoes
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: " ", itens_nao_listados: "Extensão elétrica", observacoes: "Para aula prática"
+    Then o sistema retorna "mensagem" "Parabéns, sua solicitação de recursos foi criada!" e um status "201"
+    And a reserva_id "1" possui uma solicitação com recursos " ", itens_nao_listados "Extensão elétrica" e observacoes "Para aula prática"
+
+    Scenario: sucesso ao editar uma solicitação de recursos existente
+    Given o professor possui uma solicitação de recursos associada a reserva_id "1"
+    When ele envia uma requisição PUT /solicitacoes/recursos/"1" contendo o ID da solicitação e os novos detalhes da solicitação recursos: "Projetor, Caixas de som", observacoes: "Para evento especial"
+    Then o sistema retorna "mensagem" "Solicitação de recursos atualizada com sucesso" e um status "200"
+    And o sistema atualiza os detalhes da solicitação: recursos: "Projetor, Caixas de som", observacoes: "Para evento especial"
+
+    Scenario: fracasso ao editar uma solicitação de recursos com apenas o campo observacoes preenchido
+    Given o professor possui uma reserva ativa com reserva_id "1"
+    When ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "1", recursos: " ", itens_nao_listados: " ", observacoes: "Para aula prática"
+    Then o sistema retorna "erro" "Você deve selecionar um recurso ou especificar itens não listados." e um status "400"
+
+    Scenario: sucesso ao excluir uma solicitação de recursos existente
+    Given o professor possui uma solicitação de recursos associada a reserva_id "1"
+    When ele envia uma requisição DELETE /solicitacoes/recursos/"1" contendo o ID da solicitação
+    Then o sistema remove a solicitação do banco de dados e retorna um status "204"
\ No newline at end of file
diff --git a/backend/features/reservas.feature b/backend/features/reservas.feature
new file mode 100644
index 00000000..00c45ded
--- /dev/null
+++ b/backend/features/reservas.feature
@@ -0,0 +1,41 @@
+Feature: API de Reservas
+
+    Scenario: Criar uma reserva com sucesso
+        Given a sala de id "5" está disponível no dia "2025-03-03" das "14:00" às "15:00"
+        And o professor de id "3" não tem uma reserva no dia "2025-03-03" das "14:00" às "15:00"
+        When uma requisição "POST" for enviada para "/api/reservas/3" com o corpo: "{"sala_id": 5,"data": "2025-03-03","start_time": "14:00","end_time": "15:00"}"
+        Then o status da resposta deve ser "201"
+        And o JSON da resposta deve conter "mensagem": "Reserva criada com sucesso!"
+        And o JSON da reserva deve conter "sala_id": "5"
+        And o JSON da reserva deve conter "professor_id": "3"
+
+
+    Scenario: Erro ao tentar reservar uma sala já ocupada
+        Given a sala de id "5" não está disponível no dia "2025-02-20" das "14:30" às "16:00"
+        When uma requisição "POST" for enviada para "/api/reservas/3" com o corpo: "{"sala_id": 5,"data": "2025-02-20","start_time": "14:30","end_time": "16:00"}"
+        Then o status da resposta deve ser "409"
+        And o JSON da resposta deve conter "erro": "Sala já reservada para esse horário"
+
+
+    Scenario: Erro ao tentar reservar com campos ausentes
+        Given a sala de id "5" está disponível no dia "2025-01-21" das "14:00" às "15:00"
+        And o professor de id "3" não tem uma reserva no dia "2025-01-21" das "14:00" às "15:00"
+        When uma requisição "POST" for enviada para "/api/reservas/3" com o corpo: "{"data": "2025-01-21","start_time": "14:00","end_time": "15:00"}"
+        Then o status da resposta deve ser "400"
+        And o JSON da resposta deve conter "erro": "Campos obrigatórios ausentes"
+
+
+    Scenario: Cancelar uma reserva com sucesso
+        Given o professor de id "3" tem uma reserva ativa de id "2"
+        When uma requisição "DELETE" for enviada para "/api/reservas/2" com o corpo: """"
+        Then o status da resposta deve ser "200"
+        And o JSON da resposta deve conter "mensagem": "Reserva cancelada!"
+        And o JSON da reserva deve conter "id": "2"
+
+
+    Scenario: Erro ao tentar cancelar uma reserva inexistente
+        When uma requisição "DELETE" for enviada para "/api/reservas/99" com o corpo: """"
+        Then o status da resposta deve ser "404"
+        And o JSON da resposta deve conter "erro": "Reserva não encontrada."
+
+
diff --git a/backend/features/salas.feature b/backend/features/salas.feature
new file mode 100644
index 00000000..1cbc2309
--- /dev/null
+++ b/backend/features/salas.feature
@@ -0,0 +1,43 @@
+Feature: API de Salas
+
+    Scenario: Criar sala com sucesso
+      Given não existe uma sala com nome "E901"
+      When uma requisição "POST" for enviada para "/api/salas" com o corpo: "{"nome": "E901", "tipo": "Reunião", "lugares": 20, "andar": 3, "equipamentos": ["Projetor", "Ar-condicionado"]}"
+      Then o status da resposta deve ser "201"
+      And o JSON da resposta deve conter "mensagem": "Sala criada com sucesso!"
+      And o JSON da sala deve conter "andar": "3"
+      And o JSON da sala deve conter "nome": "E901"
+      And o JSON da sala deve conter "tipo": "Reunião"
+      And o JSON da sala deve conter "lugares": "20"
+
+
+    Scenario: Erro ao tentar criar sala com nome existente
+      Given existe uma sala com nome "E001"
+      When uma requisição "POST" for enviada para "/api/salas" com o corpo: "{"nome": "E001", "tipo": "Reunião", "lugares": 20, "andar": 3, "equipamentos": ["Projetor", "Ar-condicionado"]}"
+      Then o status da resposta deve ser "409"
+      And o JSON da resposta deve conter "erro": "Já existe uma sala com esse nome"
+
+
+    Scenario: Buscar todas as salas disponíveis
+        When uma requisição "GET" for enviada para "/api/salas" com o corpo: """"
+        Then o status da resposta deve ser "200"
+        And o JSON da resposta deve conter uma lista de salas com todos os dados
+
+
+    Scenario: Erro ao buscar salas com tempo não informado
+      When uma requisição "GET" for enviada para "/api/salas?data=2025-02-25&start_time=14:00" com o corpo: """"
+      Then o status da resposta deve ser "400"
+      And o JSON da resposta deve conter "erro": "tempo não informado"
+
+
+    Scenario: Erro ao buscar salas sem data preenchida
+      When uma requisição "GET" for enviada para "/api/salas?start_time=14:00&end_time=15:00" com o corpo: """"
+      Then o status da resposta deve ser "400"
+      And o JSON da resposta deve conter "erro": "data não informada"
+
+
+    Scenario: Erro ao tentar deletar sala com reserva ativa
+      Given a sala de id "1" tem uma reserva ativa
+      When uma requisição "DELETE" for enviada para "/api/salas/1" com o corpo: """"
+      Then o status da resposta deve ser "409"
+      And o JSON da resposta deve conter "erro": "Sala possui reservas ativas e não pode ser deletada"
\ No newline at end of file
diff --git a/backend/instance/users.db b/backend/instance/users.db
new file mode 100644
index 00000000..8e170886
Binary files /dev/null and b/backend/instance/users.db differ
diff --git a/backend/main.py b/backend/main.py
new file mode 100644
index 00000000..240ca48b
--- /dev/null
+++ b/backend/main.py
@@ -0,0 +1,19 @@
+from flask import Flask
+from flask_cors import CORS
+from backend.blueprints import registrarBlueprints
+from backend.modelo.extensao import db
+from backend.config import Config
+
+app = Flask(__name__)
+app.config.from_object(Config)
+CORS(app)  # Permite conexão com o frontend
+db.init_app(app)
+registrarBlueprints(app)    
+
+with app.app_context():
+    db.drop_all()
+    db.create_all()
+    
+if __name__ == "__main__":
+    app.run(debug=True, port=5000)
+
diff --git a/backend/modelo/__init__.py b/backend/modelo/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/modelo/__pycache__/__init__.cpython-312.pyc b/backend/modelo/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..4d586061
Binary files /dev/null and b/backend/modelo/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/modelo/__pycache__/extensao.cpython-312.pyc b/backend/modelo/__pycache__/extensao.cpython-312.pyc
new file mode 100644
index 00000000..41ff179f
Binary files /dev/null and b/backend/modelo/__pycache__/extensao.cpython-312.pyc differ
diff --git a/backend/modelo/__pycache__/extensao.cpython-313.pyc b/backend/modelo/__pycache__/extensao.cpython-313.pyc
new file mode 100644
index 00000000..9dd2a510
Binary files /dev/null and b/backend/modelo/__pycache__/extensao.cpython-313.pyc differ
diff --git a/backend/modelo/__pycache__/reservas.cpython-313.pyc b/backend/modelo/__pycache__/reservas.cpython-313.pyc
new file mode 100644
index 00000000..15ad1844
Binary files /dev/null and b/backend/modelo/__pycache__/reservas.cpython-313.pyc differ
diff --git a/backend/modelo/__pycache__/salas.cpython-313.pyc b/backend/modelo/__pycache__/salas.cpython-313.pyc
new file mode 100644
index 00000000..2c565cd4
Binary files /dev/null and b/backend/modelo/__pycache__/salas.cpython-313.pyc differ
diff --git a/backend/modelo/__pycache__/solicitacaomanutencao.cpython-312.pyc b/backend/modelo/__pycache__/solicitacaomanutencao.cpython-312.pyc
new file mode 100644
index 00000000..f475de29
Binary files /dev/null and b/backend/modelo/__pycache__/solicitacaomanutencao.cpython-312.pyc differ
diff --git a/backend/modelo/__pycache__/solicitacaorecursos.cpython-312.pyc b/backend/modelo/__pycache__/solicitacaorecursos.cpython-312.pyc
new file mode 100644
index 00000000..71bbff2a
Binary files /dev/null and b/backend/modelo/__pycache__/solicitacaorecursos.cpython-312.pyc differ
diff --git a/backend/modelo/__pycache__/usuario.cpython-312.pyc b/backend/modelo/__pycache__/usuario.cpython-312.pyc
new file mode 100644
index 00000000..ca14eca7
Binary files /dev/null and b/backend/modelo/__pycache__/usuario.cpython-312.pyc differ
diff --git a/backend/modelo/__pycache__/usuario.cpython-313.pyc b/backend/modelo/__pycache__/usuario.cpython-313.pyc
new file mode 100644
index 00000000..56cdffcc
Binary files /dev/null and b/backend/modelo/__pycache__/usuario.cpython-313.pyc differ
diff --git a/backend/modelo/extensao.py b/backend/modelo/extensao.py
new file mode 100644
index 00000000..f0b13d6f
--- /dev/null
+++ b/backend/modelo/extensao.py
@@ -0,0 +1,3 @@
+from flask_sqlalchemy import SQLAlchemy
+
+db = SQLAlchemy()
diff --git a/backend/modelo/reservas.py b/backend/modelo/reservas.py
new file mode 100644
index 00000000..3e48958e
--- /dev/null
+++ b/backend/modelo/reservas.py
@@ -0,0 +1,50 @@
+# TODO: ver quais dados ainda precisa colocar aqui
+import copy
+
+mock_reservas = [
+    {
+        "id": 1,
+        "professor_id": 3,
+        "sala_id": 5,
+        "data": "2025-02-20",
+        "start_time": "14:00",
+        "end_time": "15:00",
+        "status": "ativa"
+    },
+
+    {
+        "id": 2,
+        "professor_id": 3,
+        "sala_id": 3,
+        "data": "2025-02-21",
+        "start_time": "09:00",
+        "end_time": "11:00",
+        "status": "ativa"
+    },
+
+    {
+        "id": 3,
+        "professor_id": 3,
+        "sala_id": 1,
+        "data": "2025-02-22",
+        "start_time": "08:00",
+        "end_time": "10:00",
+        "status": "ativa"
+    },
+
+    {
+        "id": 4,
+        "professor_id": 3,
+        "sala_id": 2,
+        "data": "2025-02-20",
+        "start_time": "16:00",
+        "end_time": "18:00",
+        "status": "ativa"
+    }
+]
+
+MOCK_RESERVAS_COPY = copy.copy(mock_reservas)
+
+def reservas_reset():
+    mock_reservas.clear()
+    mock_reservas.extend(copy.copy(MOCK_RESERVAS_COPY))
\ No newline at end of file
diff --git a/backend/modelo/salas.py b/backend/modelo/salas.py
new file mode 100644
index 00000000..4bff9860
--- /dev/null
+++ b/backend/modelo/salas.py
@@ -0,0 +1,97 @@
+import copy
+
+EQUIPAMENTOS = [
+    "Ar-condicionado", "Cabo P2", "Cabo HDMI", "Cabo VGA", "Microfone",
+    "Extensão", "Mesa de som", "Passador", "Televisor", "Projetor",
+    "Carregador", "Pen Drive", "Mouse", "Teclado", "Monitor", "USB-C",
+    "Cafeteira", "Gelágua"
+]
+
+mock_salas = [
+    {
+        "id": 1,
+        "nome": "E001",
+        "tipo": "Reunião",
+        "lugares": 14,
+        "andar": 15,
+        "equipamentos": ["Ar-condicionado", "Televisor"],
+        "average_rating": 4.0,
+        "review_count": 2
+    },
+    {
+        "id": 2,
+        "nome": "E002",
+        "tipo": "Auditório",
+        "lugares": 50,
+        "andar": 10,
+        "equipamentos": ["Projetor", "Microfone", "Mesa de som"],
+        "average_rating": 5.0,
+        "review_count": 5
+    },
+    {
+        "id": 3,
+        "nome": "E003",
+        "tipo": "Reunião",
+        "lugares": 8,
+        "andar": 7,
+        "equipamentos": ["Cabo HDMI", "Cabo VGA", "Extensão"],
+        "average_rating": 3.5,
+        "review_count": 3
+    },
+    {
+        "id": 4,
+        "nome": "E004",
+        "tipo": "Auditório",
+        "lugares": 80,
+        "andar": 5,
+        "equipamentos": ["Projetor", "Mesa de som", "Passador", "Teclado", "Mouse"],
+        "average_rating": 4.8,
+        "review_count": 10
+    },
+    {
+        "id": 5,
+        "nome": "E005",
+        "tipo": "Reunião",
+        "lugares": 12,
+        "andar": 9,
+        "equipamentos": ["Monitor", "USB-C", "Cabo P2"],
+        "average_rating": 4.2,
+        "review_count": 6
+    },
+    {
+        "id": 6,
+        "nome": "E006",
+        "tipo": "Reunião",
+        "lugares": 6,
+        "andar": 3,
+        "equipamentos": ["Ar-condicionado", "Gelágua", "Cafeteira"],
+        "average_rating": 3.9,
+        "review_count": 4
+    },
+    {
+        "id": 7,
+        "nome": "E007",
+        "tipo": "Auditório",
+        "lugares": 100,
+        "andar": 2,
+        "equipamentos": ["Projetor", "Microfone", "Mesa de som", "Extensão"],
+        "average_rating": 4.7,
+        "review_count": 8
+    },
+    {
+        "id": 8,
+        "nome": "E008",
+        "tipo": "Reunião",
+        "lugares": 10,
+        "andar": 4,
+        "equipamentos": ["Carregador", "Pen Drive", "USB-C"],
+        "average_rating": 4.5,
+        "review_count": 7
+    }
+]
+
+MOCK_SALAS_COPY = copy.copy(mock_salas)
+
+def salas_reset():
+    mock_salas.clear()
+    mock_salas.extend(copy.copy(MOCK_SALAS_COPY))
diff --git a/backend/modelo/solicitacaomanutencao.py b/backend/modelo/solicitacaomanutencao.py
new file mode 100644
index 00000000..ea49267e
--- /dev/null
+++ b/backend/modelo/solicitacaomanutencao.py
@@ -0,0 +1,7 @@
+from backend.modelo.extensao import db
+import datetime
+    
+class SolicitacaoManutencao(db.Model):
+    id = db.Column(db.Integer, primary_key=True)
+    reserva_id = db.Column(db.Integer, nullable=False)
+    descricao = db.Column(db.Text, nullable=False)  # Descrição da manutenção
\ No newline at end of file
diff --git a/backend/modelo/solicitacaorecursos.py b/backend/modelo/solicitacaorecursos.py
new file mode 100644
index 00000000..7dc3b497
--- /dev/null
+++ b/backend/modelo/solicitacaorecursos.py
@@ -0,0 +1,9 @@
+from backend.modelo.extensao import db
+import datetime
+
+class SolicitacaoRecursos(db.Model):
+    id = db.Column(db.Integer, primary_key=True)
+    reserva_id = db.Column(db.Integer, nullable=False)  # Associa à reserva
+    recursos = db.Column(db.Text, nullable=False)  # Itens solicitados
+    itens_nao_listados = db.Column(db.Text, nullable=True)
+    observacoes = db.Column(db.Text, nullable=True)
\ No newline at end of file
diff --git a/backend/modelo/usuario.py b/backend/modelo/usuario.py
new file mode 100644
index 00000000..2ea3611b
--- /dev/null
+++ b/backend/modelo/usuario.py
@@ -0,0 +1,11 @@
+from backend.modelo.extensao import db
+
+# Modelo da Tabela de Usuários
+class Usuario(db.Model):
+    id = db.Column(db.Integer, primary_key=True)
+    nome = db.Column(db.String(100), nullable=False)
+    cpf = db.Column(db.String(14), unique=True, nullable=False)
+    email = db.Column(db.String(100), unique=True, nullable=False)
+    professor = db.Column(db.String(1), nullable=False)  # "S" ou "N"
+    siape = db.Column(db.String(6), unique=True)
+    senha = db.Column(db.String(200), nullable=False)
\ No newline at end of file
diff --git a/backend/requirements.txt b/backend/requirements.txt
new file mode 100644
index 00000000..f86f6786
--- /dev/null
+++ b/backend/requirements.txt
@@ -0,0 +1,30 @@
+blinker==1.9.0
+certifi==2025.1.31
+charset-normalizer==3.4.1
+click==8.1.8
+colorama==0.4.6
+Flask==3.1.0
+Flask-Cors==5.0.0
+Flask-JWT-Extended==4.7.1
+Flask-SQLAlchemy==3.1.1
+gherkin-official==29.0.0
+greenlet==3.1.1
+idna==3.10
+iniconfig==2.0.0
+itsdangerous==2.2.0
+Jinja2==3.1.5
+Mako==1.3.9
+MarkupSafe==3.0.2
+packaging==24.2
+parse==1.20.2
+parse_type==0.6.4
+pluggy==1.5.0
+PyJWT==2.10.1
+pytest==8.3.4
+pytest-bdd==8.1.0
+requests==2.32.3
+six==1.17.0
+SQLAlchemy==2.0.38
+typing_extensions==4.12.2
+urllib3==2.3.0
+Werkzeug==3.1.3
diff --git a/backend/rotas/__init__.py b/backend/rotas/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/rotas/__pycache__/__init__.cpython-312.pyc b/backend/rotas/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3f93da27
Binary files /dev/null and b/backend/rotas/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/cadastro.cpython-312.pyc b/backend/rotas/__pycache__/cadastro.cpython-312.pyc
new file mode 100644
index 00000000..d57b64f0
Binary files /dev/null and b/backend/rotas/__pycache__/cadastro.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/cadastro.cpython-313.pyc b/backend/rotas/__pycache__/cadastro.cpython-313.pyc
new file mode 100644
index 00000000..4eedf8d4
Binary files /dev/null and b/backend/rotas/__pycache__/cadastro.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/criar_solicitacao_manutencao.cpython-312.pyc b/backend/rotas/__pycache__/criar_solicitacao_manutencao.cpython-312.pyc
new file mode 100644
index 00000000..e25435ad
Binary files /dev/null and b/backend/rotas/__pycache__/criar_solicitacao_manutencao.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/criar_solicitacao_recursos.cpython-312.pyc b/backend/rotas/__pycache__/criar_solicitacao_recursos.cpython-312.pyc
new file mode 100644
index 00000000..6e7a1754
Binary files /dev/null and b/backend/rotas/__pycache__/criar_solicitacao_recursos.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/editar_solicitacao_manutencao.cpython-312.pyc b/backend/rotas/__pycache__/editar_solicitacao_manutencao.cpython-312.pyc
new file mode 100644
index 00000000..e22af9e3
Binary files /dev/null and b/backend/rotas/__pycache__/editar_solicitacao_manutencao.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/editar_solicitacao_recursos.cpython-312.pyc b/backend/rotas/__pycache__/editar_solicitacao_recursos.cpython-312.pyc
new file mode 100644
index 00000000..38649e7d
Binary files /dev/null and b/backend/rotas/__pycache__/editar_solicitacao_recursos.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/excluir_solicitacao_manutencao.cpython-312.pyc b/backend/rotas/__pycache__/excluir_solicitacao_manutencao.cpython-312.pyc
new file mode 100644
index 00000000..a3e25b9f
Binary files /dev/null and b/backend/rotas/__pycache__/excluir_solicitacao_manutencao.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/excluir_solicitacao_recursos.cpython-312.pyc b/backend/rotas/__pycache__/excluir_solicitacao_recursos.cpython-312.pyc
new file mode 100644
index 00000000..22bb541b
Binary files /dev/null and b/backend/rotas/__pycache__/excluir_solicitacao_recursos.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/login.cpython-312.pyc b/backend/rotas/__pycache__/login.cpython-312.pyc
new file mode 100644
index 00000000..05b489cf
Binary files /dev/null and b/backend/rotas/__pycache__/login.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/login.cpython-313.pyc b/backend/rotas/__pycache__/login.cpython-313.pyc
new file mode 100644
index 00000000..3ae180e0
Binary files /dev/null and b/backend/rotas/__pycache__/login.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/logout.cpython-312.pyc b/backend/rotas/__pycache__/logout.cpython-312.pyc
new file mode 100644
index 00000000..6f29e6e6
Binary files /dev/null and b/backend/rotas/__pycache__/logout.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/logout.cpython-313.pyc b/backend/rotas/__pycache__/logout.cpython-313.pyc
new file mode 100644
index 00000000..9de5c3ac
Binary files /dev/null and b/backend/rotas/__pycache__/logout.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/reservas.cpython-313.pyc b/backend/rotas/__pycache__/reservas.cpython-313.pyc
new file mode 100644
index 00000000..1d93a0b2
Binary files /dev/null and b/backend/rotas/__pycache__/reservas.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/salas.cpython-313.pyc b/backend/rotas/__pycache__/salas.cpython-313.pyc
new file mode 100644
index 00000000..f457e816
Binary files /dev/null and b/backend/rotas/__pycache__/salas.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/testecookies.cpython-312.pyc b/backend/rotas/__pycache__/testecookies.cpython-312.pyc
new file mode 100644
index 00000000..693c4203
Binary files /dev/null and b/backend/rotas/__pycache__/testecookies.cpython-312.pyc differ
diff --git a/backend/rotas/__pycache__/testecookies.cpython-313.pyc b/backend/rotas/__pycache__/testecookies.cpython-313.pyc
new file mode 100644
index 00000000..356508dd
Binary files /dev/null and b/backend/rotas/__pycache__/testecookies.cpython-313.pyc differ
diff --git a/backend/rotas/__pycache__/usuario.cpython-313.pyc b/backend/rotas/__pycache__/usuario.cpython-313.pyc
new file mode 100644
index 00000000..b4a0efec
Binary files /dev/null and b/backend/rotas/__pycache__/usuario.cpython-313.pyc differ
diff --git a/backend/rotas/cadastro.py b/backend/rotas/cadastro.py
new file mode 100644
index 00000000..fd3d9737
--- /dev/null
+++ b/backend/rotas/cadastro.py
@@ -0,0 +1,79 @@
+from flask import Blueprint, request, jsonify
+from werkzeug.security import generate_password_hash
+from backend.modelo.usuario import Usuario
+from backend.modelo.extensao import db
+import re
+from sqlalchemy import exists  
+
+cadastro_bp = Blueprint("cadastro", __name__)
+
+def validarEmail(email):
+    """Valida o formato do email."""
+    padraoEmail = r'^[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+$'
+    return bool(re.match(padraoEmail, email))
+
+def validarCpf(cpf):
+    """Valida se o CPF está no formato correto XXX.XXX.XXX-XX"""
+    padraoCpf = r'^\d{3}\.\d{3}\.\d{3}-\d{2}$'
+    return bool(re.match(padraoCpf, cpf))
+
+@cadastro_bp.route("/api/cadastro", methods=["POST"])
+def cadastro():
+    data = request.get_json()
+
+    nome = data.get("nome", "").strip()
+    cpf = data.get("cpf", "").strip()
+    email = data.get("email", "").lower().strip()
+    professor = data.get("professor", "").upper().strip()
+    siape = data.get("siape", "").strip() if professor == "S" else None
+    senha = data.get("senha", "").strip()
+    confirmar_senha = data.get("confirmarSenha", "").strip()
+
+    camposObrigatorios = {
+        "nome": nome,
+        "cpf": cpf,
+        "email": email,
+        "senha": senha,
+        "Confirmar Senha": confirmar_senha
+    }
+    if professor == "S":
+        camposObrigatorios["siape"] = siape
+
+    for campo, valor in camposObrigatorios.items():
+        if not valor:
+            return jsonify({"error": f"{campo} é obrigatório."}), 400
+
+    if not validarEmail(email):
+        return jsonify({"error": "Formato de email inválido. Use um email válido, como exemplo@dominio.com."}), 400
+
+    if not validarCpf(cpf):
+        return jsonify({"error": "CPF inválido. Digite um CPF válido no formato XXX.XXX.XXX-XX."}), 400
+
+    usuarioExistente = db.session.query(
+        exists().where((Usuario.email == email) | (Usuario.cpf == cpf))
+    ).scalar()
+
+    if usuarioExistente:
+        return jsonify({"error": "Erro: email/cpf já está registrado."}), 409
+
+    if professor == "S":
+        siapeExistente = db.session.query(
+            exists().where(Usuario.siape == siape)
+        ).scalar()
+
+        if siapeExistente:
+            return jsonify({"error": "Erro: siape já está registrado."}), 409
+
+    if senha != confirmar_senha:
+        return jsonify({"error": "As senhas não coincidem."}), 400
+
+    senhaHash = generate_password_hash(senha)
+
+    novoUsuario = Usuario(
+        nome=nome, cpf=cpf, email=email, professor=professor, siape=siape, senha=senhaHash
+    )
+
+    db.session.add(novoUsuario)
+    db.session.commit()
+
+    return jsonify({"message": "Cadastro criado com sucesso!"}), 201
diff --git a/backend/rotas/criar_solicitacao_manutencao.py b/backend/rotas/criar_solicitacao_manutencao.py
new file mode 100644
index 00000000..71ca786c
--- /dev/null
+++ b/backend/rotas/criar_solicitacao_manutencao.py
@@ -0,0 +1,22 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+criar_manutencao_bp = Blueprint("criarmanutencao", __name__)
+
+@criar_manutencao_bp.route("/solicitacoes/manutencao", methods=["POST"])
+def criar_solicitacao_manutencao():
+    dados = request.json
+
+    # Verifica se 'descricao' está vazio ou None
+    if not dados.get("descricao").strip():
+        return jsonify({"erro": "O campo 'descricao' não pode estar vazio."}), 400
+
+    solicitacao = SolicitacaoManutencao(
+        reserva_id=dados.get("reserva_id"),
+        descricao=dados.get("descricao")
+    )
+    db.session.add(solicitacao)
+    db.session.commit()
+    return jsonify({"id": solicitacao.id, "mensagem": "Parabéns, sua solicitação de manutenção foi criada!"}), 201
\ No newline at end of file
diff --git a/backend/rotas/criar_solicitacao_recursos.py b/backend/rotas/criar_solicitacao_recursos.py
new file mode 100644
index 00000000..6d4e74e3
--- /dev/null
+++ b/backend/rotas/criar_solicitacao_recursos.py
@@ -0,0 +1,38 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+criar_recursos_bp = Blueprint("criarrecursos", __name__)
+
+@criar_recursos_bp.route("/solicitacoes/recursos", methods=["POST"])
+def criar_solicitacao_recursos():
+    dados = request.json
+    
+    recursos = dados.get("recursos", "").strip()
+    itens_nao_listados = dados.get("itens_nao_listados", "").strip()
+
+    # Validação: se recursos estiver vazio, itens_nao_listados deve estar preenchido
+    if not recursos and not itens_nao_listados:
+        return jsonify({"erro": "Você deve selecionar um recurso ou especificar itens não listados."}), 400
+
+    # Verifica se 'recursos' é None ou uma string vazia
+    #if not dados.get("recursos"):  
+        #return jsonify({"error": "O campo 'recursos' não pode estar vazio."}), 400
+
+    #solicitacao = SolicitacaoRecursos(
+        #reserva_id=dados.get("reserva_id"),
+        #recursos=dados.get("recursos"),
+        #itens_nao_listados=dados.get("itens_nao_listados"),
+        #observacoes=dados.get("observacoes")
+    #)
+
+    solicitacao = SolicitacaoRecursos(
+    reserva_id=dados.get("reserva_id"),
+    recursos=recursos,
+    itens_nao_listados=itens_nao_listados,
+    observacoes=dados.get("observacoes", "").strip()
+    )
+    db.session.add(solicitacao)
+    db.session.commit()
+    return jsonify({"id": solicitacao.id, "mensagem": "Parabéns, sua solicitação de recursos foi criada!"}), 201
\ No newline at end of file
diff --git a/backend/rotas/editar_solicitacao_manutencao.py b/backend/rotas/editar_solicitacao_manutencao.py
new file mode 100644
index 00000000..15965cf1
--- /dev/null
+++ b/backend/rotas/editar_solicitacao_manutencao.py
@@ -0,0 +1,25 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+editar_manutencao_bp = Blueprint("editarmanutencao", __name__)
+
+@editar_manutencao_bp.route("/solicitacoes/manutencao/<int:id>", methods=["PUT"])
+def editar_solicitacao_manutencao(id):
+    solicitacao = SolicitacaoManutencao.query.get_or_404(id)
+    dados = request.json
+
+        # Verifica se a nova descrição é vazia ou apenas espaços em branco
+    nova_descricao = dados.get("descricao", solicitacao.descricao)
+    if not nova_descricao.strip():
+        return jsonify({"erro": "A descrição da manutenção não pode ser vazia"}), 400
+
+    solicitacao.descricao = nova_descricao
+    db.session.commit()
+    return jsonify({
+    "mensagem": "Solicitação de manutenção atualizada com sucesso",
+    "id": solicitacao.id,
+    #"reserva_id": solicitacao.reserva_id,
+    "descricao": solicitacao.descricao
+}), 200
\ No newline at end of file
diff --git a/backend/rotas/editar_solicitacao_recursos.py b/backend/rotas/editar_solicitacao_recursos.py
new file mode 100644
index 00000000..7c300459
--- /dev/null
+++ b/backend/rotas/editar_solicitacao_recursos.py
@@ -0,0 +1,26 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+editar_recursos_bp = Blueprint("editarrecursos", __name__)
+
+@editar_recursos_bp.route("/solicitacoes/recursos/<int:id>", methods=["PUT"])
+def editar_solicitacao_recursos(id):
+    solicitacao = SolicitacaoRecursos.query.get_or_404(id)
+    dados = request.json
+
+     # Verifica se ambos os campos estão vazios
+    recursos = dados.get("recursos", "").strip()
+    itens_nao_listados = dados.get("itens_nao_listados", "").strip()
+
+    if not recursos and not itens_nao_listados:
+        return jsonify({"erro": "Você deve preencher pelo menos 'recursos' ou 'itens_nao_listados'."}), 400
+   
+    # Atualiza os campos apenas se foram passados na requisição
+    solicitacao.recursos = recursos if "recursos" in dados else solicitacao.recursos
+    solicitacao.itens_nao_listados = itens_nao_listados if "itens_nao_listados" in dados else solicitacao.itens_nao_listados
+    solicitacao.observacoes = dados.get("observacoes", solicitacao.observacoes)
+
+    db.session.commit()
+    return jsonify({"mensagem": "Solicitação de recursos atualizada com sucesso"}), 200
\ No newline at end of file
diff --git a/backend/rotas/excluir_solicitacao_manutencao.py b/backend/rotas/excluir_solicitacao_manutencao.py
new file mode 100644
index 00000000..085ea25f
--- /dev/null
+++ b/backend/rotas/excluir_solicitacao_manutencao.py
@@ -0,0 +1,13 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+excluir_manutencao_bp = Blueprint("excluirmanutencao", __name__)
+
+@excluir_manutencao_bp.route("/solicitacoes/manutencao/<int:id>", methods=["DELETE"])
+def excluir_solicitacao_manutencao(id):
+    solicitacao = SolicitacaoManutencao.query.get_or_404(id)
+    db.session.delete(solicitacao)
+    db.session.commit()
+    return jsonify({"message": "Solicitação excluída"}), 204
diff --git a/backend/rotas/excluir_solicitacao_recursos.py b/backend/rotas/excluir_solicitacao_recursos.py
new file mode 100644
index 00000000..c3290537
--- /dev/null
+++ b/backend/rotas/excluir_solicitacao_recursos.py
@@ -0,0 +1,13 @@
+from flask import Blueprint, request, jsonify
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+
+excluir_recursos_bp = Blueprint("excluirrecursos", __name__)
+
+@excluir_recursos_bp.route("/solicitacoes/recursos/<int:id>", methods=["DELETE"])
+def excluir_solicitacao_recursos(id):
+    solicitacao = SolicitacaoRecursos.query.get_or_404(id)
+    db.session.delete(solicitacao)
+    db.session.commit()
+    return jsonify({"message": "Solicitação excluída"}), 204
diff --git a/backend/rotas/login.py b/backend/rotas/login.py
new file mode 100644
index 00000000..70d96faf
--- /dev/null
+++ b/backend/rotas/login.py
@@ -0,0 +1,23 @@
+from flask import Blueprint, jsonify, request, make_response
+from backend.modelo.usuario import Usuario
+from werkzeug.security import check_password_hash
+
+login_bp = Blueprint("login", __name__)
+
+@login_bp.route("/api/login", methods=["POST"])
+def login():
+    data = request.get_json() or {}
+
+    # Pegando os valores com `.get()` para evitar erro caso estejam ausentes
+    email = data.get("email", "").strip()
+    senha = data.get("senha", "").strip()
+
+    if not email or not senha:
+        return jsonify({"error": "Usuário e senha são obrigatórios."}), 400
+
+    # Buscar usuário no banco
+    usuario = Usuario.query.filter_by(email=email).first()
+    if not usuario or not check_password_hash(usuario.senha, senha):
+        return jsonify({"error": "Usuário ou senha inválidos."}), 401
+
+    return jsonify({"redirect": "reserva"}), 200
diff --git a/backend/rotas/logout.py b/backend/rotas/logout.py
new file mode 100644
index 00000000..f9d69aaa
--- /dev/null
+++ b/backend/rotas/logout.py
@@ -0,0 +1,8 @@
+from flask import Blueprint, jsonify
+
+logout_bp = Blueprint("logout", __name__)
+
+@logout_bp.route("/api/logout", methods=["POST"])
+def logout():
+    response = jsonify({"message": "Logout realizado com sucesso."})
+    return response, 200
diff --git a/backend/rotas/reservas.py b/backend/rotas/reservas.py
new file mode 100644
index 00000000..d71f6ef1
--- /dev/null
+++ b/backend/rotas/reservas.py
@@ -0,0 +1,93 @@
+from flask import Blueprint, request, jsonify
+from modelo.reservas import mock_reservas
+from datetime import datetime
+
+reservas_bp = Blueprint('reservas', __name__)
+
+def parse_time(time_str):
+    return datetime.strptime(time_str, "%H:%M").time()
+
+# Get reservas
+@reservas_bp.route('/api/reservas', methods=['GET'])
+def get_reservas():
+    return jsonify(mock_reservas), 200
+
+# Get reservas para um usuário
+@reservas_bp.route('/api/reservas/<int:professor_id>', methods=['GET'])
+def get_reservas_professor(professor_id):
+    user_reservas = [
+        reserva for reserva in mock_reservas if reserva['professor_id'] == professor_id
+    ]
+
+    if not user_reservas:
+        return jsonify({'mensagem': 'Nenhuma reserva encontrada'}), 404
+
+    return jsonify(user_reservas), 200
+
+# Create reserva
+@reservas_bp.route('/api/reservas/<int:professor_id>', methods=['POST'])
+def create_reserva(professor_id):
+    """
+    Exemplo de body
+    {
+        "sala_id": 3,
+        "data": "2025-02-25",
+        "start_time": "14:00",
+        "end_time": "15:00"
+    }
+    """
+
+    dados = request.get_json()
+    sala_id = dados.get('sala_id')
+    data = dados.get('data')
+    start_time = dados.get('start_time')
+    end_time = dados.get('end_time')
+    status = "ativa"
+
+    # Checa se o body inclui o que precisa
+    if not sala_id or not data or not start_time or not end_time:
+        return jsonify({"erro": "Campos obrigatórios ausentes"}), 400
+
+    parsed_start_time = parse_time(start_time)
+    parsed_end_time = parse_time(end_time)
+
+    for reserva in mock_reservas:
+        if reserva['sala_id'] == sala_id and reserva['data'] == data and reserva['status'] == 'ativa':
+            reserva_start = parse_time(reserva['start_time'])
+            reserva_end = parse_time(reserva['end_time'])
+
+            # Sobreposição de horários
+            if not (reserva_end <= parsed_start_time or reserva_start >= parsed_end_time):
+                return jsonify({'erro': 'Sala já reservada para esse horário'}), 409
+
+    for reserva in mock_reservas:
+        if reserva['professor_id'] == professor_id and reserva['data'] == data and reserva['status'] == 'ativa':
+            professor_reserva_start = parse_time(reserva["start_time"])
+            professor_reserva_end = parse_time(reserva["end_time"])
+
+            if not (professor_reserva_end <= parsed_start_time or professor_reserva_start >= parsed_end_time):
+                return jsonify({"erro": "Professor já possui uma reserva nesse horário"}), 409
+
+    new_reserva = {
+        "id": len(mock_reservas) + 1,
+        "sala_id": sala_id,
+        "professor_id": professor_id,
+        "data": data,
+        "start_time": start_time,
+        "end_time": end_time,
+        "status": status
+    }
+    mock_reservas.append(new_reserva)
+
+    return jsonify({"mensagem": "Reserva criada com sucesso!", "reservation": new_reserva}), 201
+
+
+# Cancela reserva
+@reservas_bp.route('/api/reservas/<int:reserva_id>', methods=['DELETE'])
+def cancel_reserva(reserva_id):
+    for reserva in mock_reservas:
+        if reserva['id'] == reserva_id:
+            reserva['status'] = 'cancelada'
+            return jsonify({"mensagem": "Reserva cancelada!", "reservation": reserva}), 200
+
+    return jsonify({"erro": "Reserva não encontrada."}), 404
\ No newline at end of file
diff --git a/backend/rotas/salas.py b/backend/rotas/salas.py
new file mode 100644
index 00000000..3eff63ab
--- /dev/null
+++ b/backend/rotas/salas.py
@@ -0,0 +1,120 @@
+from flask import Blueprint, request, jsonify
+from modelo.reservas import mock_reservas
+from modelo.salas import mock_salas, EQUIPAMENTOS
+from datetime import datetime
+
+salas_bp = Blueprint('salas', __name__)
+
+# Convert "HH:MM" string to datetime.time
+def parse_time(time_str):
+    return datetime.strptime(time_str, "%H:%M").time()
+
+# Criar sala
+@salas_bp.route('/api/salas', methods=['POST'])
+def create_sala():
+    """
+    Exemplo de body:
+    {
+        "nome": "Sala E003",
+        "tipo": "Reunião",
+        "lugares": 20,
+        "andar": 3,
+        "equipamentos": ["Projetor", "Ar-condicionado"]
+    }
+    """
+
+    dados = request.get_json()
+
+    nome = dados.get("nome")
+    tipo = dados.get("tipo")
+    lugares = dados.get("lugares")
+    andar = dados.get("andar")
+    equipamentos = dados.get("equipamentos", [])
+
+
+    if not nome or not tipo or not lugares or not andar:
+        return jsonify({"erro": "Campos obrigatórios ausentes"}), 400
+
+    if tipo not in ["Reunião", "Auditório"]:
+        return jsonify({"erro": "Tipo de sala inválido. Escolha entre 'Reunião' ou 'Auditório'"}), 400
+
+    equipamentos_invalidos = [eq for eq in equipamentos if eq not in EQUIPAMENTOS]
+    if equipamentos_invalidos:
+        return jsonify({"erro": f"Equipamentos inválidos: {equipamentos_invalidos}"}), 400
+
+    if any(sala["nome"] == nome for sala in mock_salas):
+        return jsonify({"erro": "Já existe uma sala com esse nome"}), 409
+
+    nova_sala = {
+        "id": len(mock_salas) + 1,
+        "nome": nome,
+        "tipo": tipo,
+        "lugares": lugares,
+        "andar": andar,
+        "equipamentos": equipamentos
+    }
+    mock_salas.append(nova_sala)
+
+    return jsonify({"mensagem": "Sala criada com sucesso!", "sala": nova_sala}), 201
+
+
+# Get salas
+@salas_bp.route('/api/salas', methods=['GET'])
+def get_salas_disponiveis():
+    equipamentos_filtro = request.args.getlist("equipamentos")  # Get filter from query params
+    data = request.args.get("data")
+    start_time = request.args.get("start_time")
+    end_time = request.args.get("end_time")
+
+    if not equipamentos_filtro and not data and not start_time and not end_time:
+        return jsonify(mock_salas), 200
+    elif not data:
+        return jsonify({'erro': 'data não informada'}), 400
+    elif not start_time or not end_time:
+        return jsonify({'erro': 'tempo não informado'}), 400
+
+
+    salas_filtradas = [
+        sala for sala in mock_salas
+        if not equipamentos_filtro or all(eq in sala["equipamentos"] for eq in equipamentos_filtro)
+    ]
+
+    if data and start_time and end_time:
+        salas_disponiveis = []
+        for sala in salas_filtradas:
+            reservas_existentes = [
+                reserva for reserva in mock_reservas
+                if reserva["sala_id"] == sala["id"] and reserva["data"] == data and
+                   not (parse_time(reserva["end_time"]) <= parse_time(start_time) or
+                        parse_time(reserva["start_time"]) >= parse_time(end_time)) and
+                   reserva["status"] == "ativa"
+            ]
+
+            if not reservas_existentes:
+                salas_disponiveis.append(sala)
+    else:
+        salas_disponiveis = salas_filtradas
+
+    if not salas_disponiveis:
+        return jsonify({'mensagem': 'nenhuma sala encontrada'}), 404
+
+    return jsonify(salas_disponiveis), 200
+
+@salas_bp.route('/api/salas/<int:sala_id>', methods=['DELETE'])
+def delete_sala(sala_id):
+
+    sala = next((s for s in mock_salas if s["id"] == sala_id), None)
+    if not sala:
+        return jsonify({"erro": "Sala não encontrada"}), 404
+
+    reservas_ativas = [
+        reserva for reserva in mock_reservas
+        if reserva["sala_id"] == sala_id and reserva["status"] == "ativa"
+    ]
+
+    if reservas_ativas:
+        return jsonify({"erro": "Sala possui reservas ativas e não pode ser deletada"}), 409
+
+    mock_salas.remove(sala)
+
+    return jsonify({"mensagem": "Sala deletada com sucesso!"}), 200
\ No newline at end of file
diff --git a/backend/rotas/usuario.py b/backend/rotas/usuario.py
new file mode 100644
index 00000000..6923e96d
--- /dev/null
+++ b/backend/rotas/usuario.py
@@ -0,0 +1,35 @@
+from flask import Blueprint, request, jsonify
+from modelo.extensao import db
+from modelo.usuario import Usuario
+
+usuarios_bp = Blueprint('usuarios', __name__)
+
+@usuarios_bp.route('/api/usuarios', methods=['GET'])
+def get_usuarios():
+    usuarios = Usuario.query.all()
+
+    usuarios_json = [
+        {
+            "id": usuario.id,
+            "nome": usuario.nome,
+            "cpf": usuario.cpf,
+            "email": usuario.email,
+            "professor": usuario.professor
+        }
+        for usuario in usuarios
+    ]
+
+    return jsonify(usuarios_json), 200
+
+
+@usuarios_bp.route("/api/usuarios/<int:usuario_id>", methods=["DELETE"])
+def deletar_usuario(usuario_id):
+    usuario = Usuario.query.get(usuario_id)
+
+    if not usuario:
+        return jsonify({"erro": "Usuário não encontrado"}), 404
+
+    db.session.delete(usuario)
+    db.session.commit()
+
+    return jsonify({"menssagem": "Usuário deletado com sucesso"}), 200
\ No newline at end of file
diff --git a/backend/testes/__init__.py b/backend/testes/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/testes/__pycache__/__init__.cpython-312.pyc b/backend/testes/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2aaf3ca2
Binary files /dev/null and b/backend/testes/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/testes/__pycache__/common_step_definitions.cpython-313.pyc b/backend/testes/__pycache__/common_step_definitions.cpython-313.pyc
new file mode 100644
index 00000000..64c12849
Binary files /dev/null and b/backend/testes/__pycache__/common_step_definitions.cpython-313.pyc differ
diff --git a/backend/testes/__pycache__/conftest.cpython-312-pytest-8.3.4.pyc b/backend/testes/__pycache__/conftest.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..a4c60acf
Binary files /dev/null and b/backend/testes/__pycache__/conftest.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/testes/__pycache__/conftest.cpython-313-pytest-8.1.0.pyc b/backend/testes/__pycache__/conftest.cpython-313-pytest-8.1.0.pyc
new file mode 100644
index 00000000..a9b021ea
Binary files /dev/null and b/backend/testes/__pycache__/conftest.cpython-313-pytest-8.1.0.pyc differ
diff --git a/backend/testes/__pycache__/conftest.cpython-313-pytest-8.3.4.pyc b/backend/testes/__pycache__/conftest.cpython-313-pytest-8.3.4.pyc
new file mode 100644
index 00000000..8020514f
Binary files /dev/null and b/backend/testes/__pycache__/conftest.cpython-313-pytest-8.3.4.pyc differ
diff --git a/backend/testes/__pycache__/test_manutencao.cpython-312-pytest-8.3.4.pyc b/backend/testes/__pycache__/test_manutencao.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..9e430675
Binary files /dev/null and b/backend/testes/__pycache__/test_manutencao.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/testes/__pycache__/test_recursos.cpython-312-pytest-8.3.4.pyc b/backend/testes/__pycache__/test_recursos.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..e8efa680
Binary files /dev/null and b/backend/testes/__pycache__/test_recursos.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.1.0.pyc b/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.1.0.pyc
new file mode 100644
index 00000000..44f4ff64
Binary files /dev/null and b/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.1.0.pyc differ
diff --git a/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.3.4.pyc b/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.3.4.pyc
new file mode 100644
index 00000000..77705211
Binary files /dev/null and b/backend/testes/__pycache__/test_reservas.cpython-313-pytest-8.3.4.pyc differ
diff --git a/backend/testes/__pycache__/test_salas.cpython-313-pytest-8.1.0.pyc b/backend/testes/__pycache__/test_salas.cpython-313-pytest-8.1.0.pyc
new file mode 100644
index 00000000..c692097d
Binary files /dev/null and b/backend/testes/__pycache__/test_salas.cpython-313-pytest-8.1.0.pyc differ
diff --git a/backend/testes/__pycache__/teste_solicitacoes.cpython-312-pytest-8.3.4.pyc b/backend/testes/__pycache__/teste_solicitacoes.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..3cae02c9
Binary files /dev/null and b/backend/testes/__pycache__/teste_solicitacoes.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/testes/common_step_definitions.py b/backend/testes/common_step_definitions.py
new file mode 100644
index 00000000..a2ced8ee
--- /dev/null
+++ b/backend/testes/common_step_definitions.py
@@ -0,0 +1,65 @@
+import json
+import requests
+from pytest_bdd import when, then, given, parsers
+
+
+BASE_URL = 'http://127.0.0.1:5000'
+
+@when(parsers.parse('uma requisição "{metodo}" for enviada para "{url}" com o corpo: "{body}"'))
+def mandar_requisicao(metodo, url, body):
+    url = f'{BASE_URL}/{url}'
+
+    metodo = metodo.upper()
+    request_method = {
+        "POST": requests.post,
+        "GET": requests.get,
+        "DELETE": requests.delete,
+        "PUT": requests.put
+    }.get(metodo)
+
+    assert request_method is not None, f"Método HTTP inválido: {metodo}"
+
+    json_body = json.loads(body) if metodo in ["POST", "PUT"] else None
+
+    response = request_method(url, json=json_body)
+
+    mandar_requisicao.response = response
+
+
+@then(parsers.parse('o status da resposta deve ser "{status}"'))
+def checar_status(status):
+    assert mandar_requisicao.response.status_code == int(status), f'mensagem: "{mandar_requisicao.response.text}"'
+
+@then(parsers.parse('o JSON da resposta deve conter "{atributo}": "{string}"'))
+def checar_mensagem(atributo, string):
+    response_json = mandar_requisicao.response.json()
+
+    assert response_json.get(atributo) == string, (
+        f"Erro: Esperado '{atributo}': '{string}', obtido '{response_json.get(atributo)}'"
+    )
+
+
+@given(parsers.parse('a sala de id "{sala_id:d}" está disponível no dia "{data}" das "{start_time}" às "{end_time}"'))
+def sala_disponivel(sala_id, data, start_time, end_time):
+    url = f'{BASE_URL}/api/salas?data={data}&start_time={start_time}&end_time={end_time}'
+    response = requests.get(url)
+
+    assert response.status_code == 200, f"Erro ao buscar salas disponíveis: {response.text}"
+
+    sala_encontrada = any(sala['id'] == sala_id for sala in response.json())
+
+    assert sala_encontrada, (f'Sala {sala_id} não está disponível no horário solicitado ({data} das {start_time} '
+                             f'às {end_time})')
+
+
+@given(parsers.parse('a sala de id "{sala_id:d}" não está disponível no dia "{data}" das "{start_time}" às "{end_time}"'))
+def sala_nao_disponivel(sala_id, data, start_time, end_time):
+    url = f'{BASE_URL}/api/salas?data={data}&start_time={start_time}&end_time={end_time}'
+    response = requests.get(url)
+
+    assert response.status_code == 200, f"Erro ao buscar salas disponíveis: {response.text}"
+
+    sala_encontrada = any(sala['id'] == sala_id for sala in response.json())
+
+    assert not sala_encontrada, (f'Sala {sala_id} está disponível no horário solicitado ({data} das {start_time} '
+                             f'às {end_time})')
\ No newline at end of file
diff --git a/backend/testes/conftest.py b/backend/testes/conftest.py
new file mode 100644
index 00000000..ab57e669
--- /dev/null
+++ b/backend/testes/conftest.py
@@ -0,0 +1,96 @@
+import pytest
+from flask import Flask
+from backend.rotas.criar_solicitacao_manutencao import criar_manutencao_bp
+from backend.rotas.criar_solicitacao_recursos import criar_recursos_bp
+from backend.rotas.editar_solicitacao_manutencao import editar_manutencao_bp
+from backend.rotas.editar_solicitacao_recursos import editar_recursos_bp
+from backend.rotas.excluir_solicitacao_manutencao import excluir_manutencao_bp
+from backend.rotas.excluir_solicitacao_recursos import excluir_recursos_bp
+from backend.modelo.extensao import db
+from backend.modelo.solicitacaomanutencao import SolicitacaoManutencao
+from backend.modelo.solicitacaorecursos import SolicitacaoRecursos
+from ..rotas.cadastro import cadastro_bp
+from ..rotas.login import login_bp
+from ..modelo.extensao import db
+from ..modelo.usuario import Usuario
+from werkzeug.security import generate_password_hash
+
+@pytest.fixture(scope="module")
+def app():
+    aplicacao = Flask(__name__)
+    aplicacao.register_blueprint(criar_manutencao_bp)
+    aplicacao.register_blueprint(criar_recursos_bp)
+    aplicacao.register_blueprint(editar_manutencao_bp)
+    aplicacao.register_blueprint(editar_recursos_bp)
+    aplicacao.register_blueprint(excluir_manutencao_bp)
+    aplicacao.register_blueprint(excluir_recursos_bp)
+    aplicacao.register_blueprint(cadastro_bp)
+    aplicacao.register_blueprint(login_bp)
+    aplicacao.config["TESTING"] = True
+    aplicacao.config["SQLALCHEMY_DATABASE_URI"] = "sqlite:///test.db"
+    aplicacao.config["SQLALCHEMY_TRACK_MODIFICATIONS"] = False
+
+    with aplicacao.app_context():
+        db.init_app(aplicacao)
+        db.create_all()
+        yield aplicacao
+        db.session.remove()
+        db.drop_all()
+
+@pytest.fixture(scope="module")
+def client(app):
+    with app.test_client() as client:
+        yield client
+
+@pytest.fixture(scope="module", autouse=True)
+def setup_database(app):
+    """Popula o banco de dados com usuários de teste antes dos testes rodarem."""
+    with app.app_context():
+        db.create_all()
+        
+        usuario1 = Usuario(
+            nome="Demosténes",
+            cpf="126.456.789-00",
+            email="demostenes@example.com",
+            professor="N",
+            siape=None,
+            senha=generate_password_hash("SecurePassword123"),
+        )
+
+        usuario2 = Usuario(
+            nome="Vanessa",
+            cpf="321.879.789-33",
+            email="vanessa@example.com",
+            professor="S",
+            siape="101010",
+            senha=generate_password_hash("12345678"),
+        )
+        
+        manutencao1 = SolicitacaoManutencao(
+            reserva_id=1,
+            descricao="Mesa quebrada."
+        )
+
+        # Criando solicitações de recursos
+        recurso1 = SolicitacaoRecursos(
+            reserva_id=2,
+            recursos="Projetor, Teclado",
+            itens_nao_listados="Extensão elétrica",
+            observacoes="Para aula prática"
+        )
+#        recurso2 = SolicitacaoRecursos(
+#            reserva_id=3,
+#            descricao="Necessidade de cadeiras extras.",
+#            quantidade=5
+#        )
+
+        db.session.add_all([manutencao1, recurso1])
+        db.session.add_all([usuario1, usuario2])
+        db.session.commit()        
+        yield 
+        db.session.remove()
+        db.drop_all()
+
+@pytest.fixture
+def contexto():
+    return {}
diff --git a/backend/testes/test_cadastro_servicos.py b/backend/testes/test_cadastro_servicos.py
new file mode 100644
index 00000000..6eb8730c
--- /dev/null
+++ b/backend/testes/test_cadastro_servicos.py
@@ -0,0 +1,89 @@
+from pytest_bdd import scenario, given, when, then, parsers
+import pytest
+
+# Cenários de teste_
+@scenario("../features/CadastroServico.feature", "Sucesso no cadastro de usuário")
+def teste_SucessoUsuario():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Sucesso no cadastro de professor")
+def teste_SucessoProfessor():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por campos obrigatórios não preenchidos")
+def teste_ErroObrigatorio():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por siape já registrado")
+def teste_ErroSiapeRegistrado():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por duplicação de ID única")
+def teste_ErroCadastroDuplo():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por senhas que não coincidem")
+def teste_ErroCadastroSenha():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por formato inválido de email")
+def teste_ErroEmailInvalido():
+    pass
+
+@scenario("../features/CadastroServico.feature", "Fracasso no cadastro por formato inválido de CPF")
+def teste_ErroCpfInvalido():
+    pass
+
+# Given
+@given("o usuário deseja se cadastrar")
+def usuarioInicioCadastro(contexto):
+    contexto["dados_cadastro"] = {}
+
+# When
+@when(parsers.parse('ele informa o nome "{nome}"'))
+def informarNome(contexto, nome):
+    contexto["dados_cadastro"]["nome"] = nome
+
+@when(parsers.parse('ele informa o CPF "{cpf}"'))
+def informarCpf(contexto, cpf):
+    contexto["dados_cadastro"]["cpf"] = cpf
+
+@when(parsers.parse('ele informa o email "{email}"'))
+def informarEmail(contexto, email):
+    contexto["dados_cadastro"]["email"] = email
+
+@when(parsers.parse('ele informa se é professor "{professor}"'))
+def informarProfessor(contexto, professor):
+    contexto["dados_cadastro"]["professor"] = professor
+
+@when(parsers.parse('ele informa o SIAPE "{siape}"'))
+def informarSiape(contexto, siape):
+    contexto["dados_cadastro"]["siape"] = siape
+
+@when(parsers.parse('ele informa a senha "{senha}"'))
+def informarSenha(contexto, senha):
+    contexto["dados_cadastro"]["senha"] = senha
+
+@when(parsers.parse('ele informa a confirmação da senha "{confirmar_senha}"'))
+def confirmarSenha(contexto, confirmar_senha):
+    contexto["dados_cadastro"]["confirmarSenha"] = confirmar_senha
+
+@when(parsers.parse('ele envia uma requisição POST para "/api/cadastro"'))
+def enviarCadastro(client, contexto):
+    resposta = client.post("/api/cadastro", json=contexto["dados_cadastro"])
+    contexto["resposta"] = resposta
+    
+@when(parsers.parse('ele deixa o campo "Confirmar Senha" com ""'))
+def senhaVazio(contexto):
+    pass
+# Then
+@then(parsers.parse('a resposta deve conter a mensagem "{mensagem}"'))
+def verificarMensagem(contexto, mensagem):
+    resposta_json = contexto["resposta"].get_json()
+    mensagem_real = resposta_json.get("message") or resposta_json.get("error")
+    assert mensagem_real == mensagem, f"Esperado: {mensagem}, Recebido: {mensagem_real}"
+
+@then(parsers.parse('o status code deve ser "{status_code}"'))
+def verificarStatusCode(contexto, status_code):
+    status_code = int(status_code)
+    assert contexto["resposta"].status_code == status_code, f"Esperado: {status_code}, Recebido: {contexto['resposta'].status_code}"
diff --git a/backend/testes/test_login_servicos.py b/backend/testes/test_login_servicos.py
new file mode 100644
index 00000000..3a0174a0
--- /dev/null
+++ b/backend/testes/test_login_servicos.py
@@ -0,0 +1,55 @@
+from pytest_bdd import scenario, given, when, then, parsers
+import pytest
+
+# Cenários de teste_
+@scenario("../features/loginServico.feature", "Sucesso no login")
+def teste_SucessoLogin():
+    pass
+
+@scenario("../features/loginServico.feature", "Fracasso no login por senha incorreta")
+def teste_FracassoLogin():
+    pass
+
+@scenario("../features/loginServico.feature", "Fracasso no login por falta de email ou senha")
+def teste_FracassoLoginSemEmailOuSenha():
+    pass
+
+# Given
+@given(parsers.parse('o usuário possui o email "{email}" e a senha "{senha}" válidos'))
+def usuarioCredenciais(contexto, email, senha):
+    contexto["email"] = email
+    contexto["senha"] = senha
+
+@given(parsers.parse('o usuário possui o email "{email}" válido e a senha "{senha}" inválida'))
+def usuarioSenhaInvalida(contexto, email, senha):
+    contexto["email"] = email
+    contexto["senha"] = senha
+
+@given("o usuário envia uma requisição sem email ou senha")
+def usuarioSemCredenciais(contexto):
+    contexto["email"] = ""
+    contexto["senha"] = ""
+
+# When
+@when(parsers.parse('ele envia uma requisição POST para "/api/login" com os dados "{email}" e "{senha}"'))
+def enviarRequisicaoLogin(client, contexto, email, senha):
+    resposta = client.post("/api/login", json={"email": email, "senha": senha})
+    contexto["resposta"] = resposta
+
+@then("a resposta deve conter os dados do usuário")
+def verificarRespostaSucesso(contexto):
+    respostaJson = contexto["resposta"].get_json()
+    print(f"✅ [DEBUG] Resposta esperada: redirect='reserva'")
+    print(f"✅ [DEBUG] Resposta recebida: {respostaJson}")
+    
+    assert respostaJson["redirect"] == "reserva", f"Esperado: reserva, Recebido: {respostaJson}"
+# Then
+@then(parsers.parse('a resposta deve conter a mensagem "{mensagem}"'))
+def verificarRespostaFalha(contexto, mensagem):
+    respostaJson = contexto["resposta"].get_json()
+    assert respostaJson["error"] == mensagem, f"Esperado: {mensagem}, Recebido: {respostaJson}"
+
+@then(parsers.parse('o status code deve ser "{status_code}"'))
+def verificarStatusCode(contexto, status_code):
+    status_code = int(status_code)
+    assert contexto["resposta"].status_code == status_code, f"Esperado: {status_code}, Recebido: {contexto['resposta'].status_code}"
diff --git a/backend/testes/test_manutencao.py b/backend/testes/test_manutencao.py
new file mode 100644
index 00000000..b26ecb0c
--- /dev/null
+++ b/backend/testes/test_manutencao.py
@@ -0,0 +1,77 @@
+from pytest_bdd import scenario, given, when, then, parsers
+import pytest
+
+# Cenários de teste_
+@scenario("../features/manutencao.feature", "sucesso ao criar uma solicitação de manutenção para uma reserva concluída")
+def teste_SucessoCriarManutencao():
+    pass
+
+@scenario("../features/manutencao.feature", "fracasso ao criar uma solicitação de manutenção sem preencher o campo descricao")
+def teste_FracassoCriarManutencao():
+    pass
+
+@scenario("../features/manutencao.feature", "sucesso ao editar uma solicitação de manutenção existente")
+def teste_SucessoEditarManutencao():
+    pass
+
+@scenario("../features/manutencao.feature", "fracasso ao editar solicitação de manutenção sem preencher o campo descricao")
+def teste_FracassoEditarManutencao():
+    pass
+
+@scenario("../features/manutencao.feature", "sucesso ao excluir uma solicitação de manutenção existente")
+def teste_SucessoExcluirManutencao():
+    pass
+
+# Given
+@given(parsers.parse('o professor possui uma reserva de sala reserva_id "{reserva_id}" que já foi encerrada'))
+def reserva_encerrada(contexto, reserva_id):
+    contexto["reserva_id"] = int(reserva_id)
+
+@given(parsers.parse('o professor já criou uma solicitação de manutenção associada a reserva_id "{reserva_id}"'))
+def criar_solicitacao_previa(contexto, reserva_id):
+    contexto["reserva_id"] = int(reserva_id)
+    contexto["descricao"] = "Mesa quebrada."
+    contexto["manutencao_id"] = 1 # Simulando um ID gerado pelo sistema
+
+# When
+@when(parsers.parse('ele envia uma requisição POST /solicitacoes/manutencao com os dados: reserva_id: "{reserva_id}", descricao: "{descricao}"'))
+def criar_solicitacao(client, contexto, reserva_id, descricao):
+    resposta = client.post("/solicitacoes/manutencao", json={"reserva_id": reserva_id, "descricao": descricao})
+    contexto["resposta"] = resposta
+    contexto["descricao"] = descricao
+
+@when(parsers.parse('ele envia uma requisição PUT /solicitacoes/manutencao/{id} contendo o ID da solicitação de manutenção e a alteração descricao: "{descricao}"'))
+def editar_solicitacao(client, contexto, descricao):
+    import json
+    manutencao_id = contexto["manutencao_id"]
+    resposta = client.put(f"/solicitacoes/manutencao/{manutencao_id}", json={"descricao": descricao})
+    contexto["resposta"] = resposta
+    dados = json.loads(resposta.data)
+    contexto["descricao"] = dados.get("descricao")
+
+@when(parsers.parse('ele envia uma requisição DELETE /solicitacoes/manutencao/{id}'))
+def deletar_solicitacao(client, contexto):
+    manutencao_id = contexto["manutencao_id"]
+    resposta = client.delete(f"/solicitacoes/manutencao/{manutencao_id}")
+    contexto["resposta"] = resposta
+
+
+# Then
+@then(parsers.parse('o sistema retorna "{mensagem_tipo}" "{mensagem}" e um status "{status}"'))
+def verificar_resposta(contexto, mensagem_tipo, mensagem, status):
+    resposta = contexto["resposta"]
+    assert resposta.status_code == int(status), f"{resposta}"
+    assert resposta.get_json()[mensagem_tipo] == mensagem
+
+@then(parsers.parse('a reserva reserva_id: "{reserva_id}" possui uma solicitação de manutenção com descricao: "{descricao}"'))
+def verificar_solicitacao_de_manutencao(contexto, reserva_id, descricao):
+    assert contexto["reserva_id"] == int(reserva_id)
+    assert contexto["descricao"] == descricao
+
+@then(parsers.parse('o sistema atualiza os detalhes da solicitação com descricao: "{descricao}"'))
+def verificar_solicitacao_de_manutencao(contexto, descricao):
+    assert contexto["descricao"] == descricao
+
+@then(parsers.parse('o sistema remove a solicitação do banco de dados e retorna um status "{status}"'))
+def verificar_exclusao_de_manutencao(contexto, status):
+    assert contexto["resposta"].status_code == int(status)
\ No newline at end of file
diff --git a/backend/testes/test_recursos.py b/backend/testes/test_recursos.py
new file mode 100644
index 00000000..15fcb0c2
--- /dev/null
+++ b/backend/testes/test_recursos.py
@@ -0,0 +1,102 @@
+from pytest_bdd import scenario, given, when, then, parsers
+import pytest
+
+# Cenários de teste
+@scenario("../features/recursos.feature", "sucesso ao criar solicitação de recursos para uma reserva ativa com todos os campos preenchidos")
+def teste_SucessoCriarSolicitacao():
+    pass
+
+@scenario("../features/recursos.feature", "fracasso ao criar solicitação de recursos com campos preenchidos com espaços ou não preenchidos")
+def teste_FracassoCriarSolicitacaoCamposVazios():
+    pass
+
+@scenario("../features/recursos.feature", "fracasso ao criar solicitação de recursos com apenas o campo observacoes preenchido")
+def teste_FracassoCriarSolicitacaoSomenteObservacoes():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao criar solicitação de recursos sem preencher o campo itens_nao_listados")
+def teste_SucessoCriarSolicitacaoSemItensNaoListados():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao criar solicitação de recursos sem preencher o campo itens_nao_listados e o campo observacoes")
+def teste_SucessoCriarSolicitacaoBasica():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao criar solicitação de recursos com o campo de recursos vazio e apenas especificando os itens não listados")
+def teste_SucessoCriarSolicitacaoComItensNaoListados():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao criar solicitação de recursos com o campo de recursos vazio e especificando os itens não listados e as observacoes")
+def teste_SucessoCriarSolicitacaoComItensNaoListadosObs():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao editar uma solicitação de recursos existente")
+def teste_SucessoEditarSolicitacao():
+    pass
+
+@scenario("../features/recursos.feature", "fracasso ao editar uma solicitação de recursos com apenas o campo observacoes preenchido")
+def teste_FracassoEditarSolicitacao():
+    pass
+
+@scenario("../features/recursos.feature", "sucesso ao excluir uma solicitação de recursos existente")
+def teste_SucessoExcluirSolicitacao():
+    pass
+
+# Given
+@given(parsers.parse('o professor possui uma reserva ativa com reserva_id "{reserva_id}"'))
+def reserva_ativa(contexto, reserva_id):
+    contexto["reserva_id"] = int(reserva_id)
+
+@given(parsers.parse('o professor possui uma solicitação de recursos associada a reserva_id "{reserva_id}"'))
+def criar_solicitacao_previa(contexto, reserva_id):
+    contexto["reserva_id"] = int(reserva_id)
+    contexto["recursos"] = "Projetor, Teclado"
+    contexto["solicitacao_id"] = 1
+
+# When
+@when(parsers.parse('ele envia uma requisição POST /solicitacoes/recursos com os dados: reserva_id: "{reserva_id}", recursos: "{recursos}", itens_nao_listados: "{itens_nao_listados}", observacoes: "{observacoes}"'))
+def criar_solicitacao(client, contexto, reserva_id, recursos, itens_nao_listados, observacoes):
+    resposta = client.post("/solicitacoes/recursos", json={"reserva_id": reserva_id, "recursos": recursos, "itens_nao_listados": itens_nao_listados, "observacoes": observacoes})
+    contexto["resposta"] = resposta
+    contexto["recursos"] = recursos
+    contexto["itens_nao_listados"] = itens_nao_listados
+    contexto["observacoes"] = observacoes
+
+@when(parsers.parse('ele envia uma requisição PUT /solicitacoes/recursos/{id} contendo o ID da solicitação e os novos detalhes da solicitação recursos: "{recursos}", observacoes: "{observacoes}"'))
+def editar_solicitacao(client, contexto, recursos, observacoes):
+    import json
+    solicitacao_id = contexto["solicitacao_id"]
+    resposta = client.put(f"/solicitacoes/recursos/{solicitacao_id}", json={"recursos": recursos, "observacoes": observacoes})
+    contexto["resposta"] = resposta
+    dados = json.loads(resposta.data)
+    contexto["recursos"] = recursos
+    contexto["observacoes"] = observacoes
+
+@when(parsers.parse('ele envia uma requisição DELETE /solicitacoes/recursos/{id} contendo o ID da solicitação'))
+def deletar_solicitacao(client, contexto):
+    solicitacao_id = contexto["solicitacao_id"]
+    resposta = client.delete(f"/solicitacoes/recursos/{solicitacao_id}")
+    contexto["resposta"] = resposta
+
+# Then
+@then(parsers.parse('o sistema retorna "{mensagem_tipo}" "{mensagem}" e um status "{status}"'))
+def verificar_resposta(contexto, mensagem_tipo, mensagem, status):
+    resposta = contexto["resposta"]
+    assert resposta.status_code == int(status)
+    assert resposta.get_json()[mensagem_tipo] == mensagem
+
+@then(parsers.parse('a reserva_id "{reserva_id}" possui uma solicitação com recursos "{recursos}", itens_nao_listados "{itens_nao_listados}" e observacoes "{observacoes}"'))
+def verificar_solicitacao(contexto, reserva_id, recursos, itens_nao_listados, observacoes):
+    assert contexto["reserva_id"] == int(reserva_id)
+    assert contexto["recursos"] == recursos
+    assert contexto["itens_nao_listados"] == itens_nao_listados
+    assert contexto["observacoes"] == observacoes
+
+@then(parsers.parse('o sistema atualiza os detalhes da solicitação: recursos: "{recursos}", observacoes: "{observacoes}"'))
+def verificar_solicitacao_de_manutencao(contexto, recursos, observacoes):
+    assert contexto["recursos"] == recursos
+    assert contexto["observacoes"] == observacoes
+
+@then(parsers.parse('o sistema remove a solicitação do banco de dados e retorna um status "{status}"'))
+def verificar_exclusao(contexto, status):
+    assert contexto["resposta"].status_code == int(status)
diff --git a/backend/testes/test_reservas.py b/backend/testes/test_reservas.py
new file mode 100644
index 00000000..9b4df0d2
--- /dev/null
+++ b/backend/testes/test_reservas.py
@@ -0,0 +1,78 @@
+import requests
+from pytest_bdd import scenario, when, then, given, parsers
+from common_step_definitions import *
+
+BASE_URL = 'http://127.0.0.1:5000'
+
+
+@scenario('../features/reservas.feature', 'Criar uma reserva com sucesso')
+def test_criar_uma_reserva_com_sucesso():
+    pass
+
+@scenario('../features/reservas.feature', 'Erro ao tentar reservar uma sala já ocupada')
+def test_erro_ao_tentar_reservar_sala_ocupada():
+    pass
+
+@scenario('../features/reservas.feature', 'Erro ao tentar reservar com campos ausentes')
+def test_erro_ao_tentar_reservar_campos_ausentes():
+    pass
+
+@scenario('../features/reservas.feature', 'Cancelar uma reserva com sucesso')
+def test_cancelar_uma_reserva_com_sucesso():
+    pass
+
+@scenario('../features/reservas.feature', 'Erro ao tentar cancelar uma reserva inexistente')
+def test_cancelar_uma_reserva_inexistente():
+    pass
+
+
+@given(parsers.parse('o professor de id "{professor_id:d}" não tem uma reserva no dia "{data}" '
+                     'das "{start_time}" às "{end_time}"'))
+def professor_disponivel(professor_id, data, start_time, end_time):
+    url = f'{BASE_URL}/api/reservas/{professor_id}'
+    response = requests.get(url)
+    reservas = response.json()
+
+    for reserva in reservas:
+        if reserva['data'] == data and reserva['status'] == ['ativa']:
+            assert (start_time >= reserva['end_time'] or end_time <= reserva['start_time']), \
+                f'Professor {professor_id} já tem uma reserva nesse horário'
+
+
+@given(parsers.parse('o professor de id "{professor_id:d}" tem uma reserva ativa de id "{reserva_id}"'))
+def professor_disponivel(professor_id, reserva_id):
+    url = f'{BASE_URL}/api/reservas/{professor_id}'
+    response = requests.get(url)
+
+    assert response.status_code == 200, f'Erro ao buscar reservas do professor {professor_id}'
+
+    reservas = response.json()
+
+    reserva_encontrada = any(
+        reserva['id'] == int(reserva_id) and reserva["status"] for reserva in reservas
+    )
+
+    assert reserva_encontrada, f"O professor {professor_id} não tem uma reserva ativa de id {reserva_id}"
+
+
+@then(parsers.parse('o JSON da reserva deve conter "{atributo}": "{valor}"'))
+def checar_atributo(atributo, valor):
+    response_json = mandar_requisicao.response.json()
+
+    assert "reservation" in response_json, (
+        f"Erro: 'reservation' não encontrado na resposta:\n{response_json}"
+    )
+
+    assert atributo in response_json["reservation"], (
+        f"Erro: Atributo '{atributo}' não encontrado dentro de 'reservation':\n{response_json}"
+    )
+
+    if valor.isdigit():
+        valor = int(valor)
+
+    valor_real = response_json["reservation"][atributo]
+
+    assert valor_real == valor, (
+        f"Erro: Para '{atributo}', esperado '{valor}' ({type(valor).__name__}), "
+        f"obtido '{valor_real}' ({type(valor_real).__name__})"
+    )
\ No newline at end of file
diff --git a/backend/testes/test_salas.py b/backend/testes/test_salas.py
new file mode 100644
index 00000000..2d74b17f
--- /dev/null
+++ b/backend/testes/test_salas.py
@@ -0,0 +1,126 @@
+import json
+import requests
+from pytest_bdd import scenario, when, then, given, parsers
+
+from common_step_definitions import *
+
+
+BASE_URL = 'http://127.0.0.1:5000'
+
+@scenario('../features/salas.feature', 'Criar sala com sucesso')
+def test_criar_sala_com_sucesso():
+    pass
+
+@scenario('../features/salas.feature', 'Buscar todas as salas disponíveis')
+def test_buscar_salas():
+    pass
+
+@scenario('../features/salas.feature', 'Erro ao buscar salas sem data preenchida')
+def test_erro_ao_buscar_salas_data():
+    pass
+
+@scenario('../features/salas.feature', 'Erro ao buscar salas com tempo não informado')
+def test_erro_ao_buscar_salas_tempo():
+    pass
+
+@scenario('../features/salas.feature', 'Erro ao tentar deletar sala com reserva ativa')
+def test_erro_ao_tentar_deletar_sala_com_reserva():
+    pass
+
+@given(parsers.parse('a sala de id "{sala_id}" tem uma reserva ativa'))
+def sala_tem_reserva_ativa(sala_id):
+    sala_id = int(sala_id)
+
+    response = requests.get(f"{BASE_URL}/api/reservas")
+    assert response.status_code == 200, f"Erro ao buscar reservas: {response.text}"
+
+    reservas = response.json()
+
+    reserva_existente = any(
+        reserva for reserva in reservas if reserva["sala_id"] == sala_id and reserva["status"] == "ativa"
+    )
+
+    if not reserva_existente:
+        reserva_body = {
+            "sala_id": sala_id,
+            "data": "2025-02-25",
+            "start_time": "14:00",
+            "end_time": "15:00"
+        }
+
+        reserva_response = requests.post(f"{BASE_URL}/api/reservas/3", json=reserva_body)  # Aqui eu crio só para o professor de id 3
+        assert reserva_response.status_code == 201, f"Erro ao criar reserva ativa: {reserva_response.text}"
+
+
+@given(parsers.parse('existe uma sala com nome "{nome}"'))
+def existe_uma_sala_com_nome(nome):
+    response = requests.get(f"{BASE_URL}/api/salas")
+
+    salas = response.json()
+
+
+    sala_existente = next((sala for sala in salas if sala["nome"] == nome), None)
+
+    if not sala_existente:
+        sala_body = {
+            "nome": nome,
+            "tipo": "Reunião",
+            "lugares": 10,
+            "andar": 0,
+            "equipamentos": ["Projetor"]
+        }
+
+        sala_response = requests.post(f"{BASE_URL}/api/salas", json=sala_body)
+        assert sala_response.status_code == 201, f"Erro ao criar sala: {response.text}"
+
+@given(parsers.parse('não existe uma sala com nome "{nome}"'))
+def sala_nao_existe(nome):
+
+    response = requests.get(f"{BASE_URL}/api/salas")
+    assert response.status_code == 200, "Erro ao buscar salas: " + response.text
+
+    salas = response.json()
+
+    for sala in salas:
+        if sala["nome"] == nome:
+            sala_id = sala["id"]
+
+            # Deleta se necessário
+            delete_response = requests.delete(f"{BASE_URL}/api/salas/{sala_id}")
+            assert delete_response.status_code in [200, 204], "Erro ao deletar sala: " + delete_response.text
+
+
+@then(parsers.parse('o JSON da sala deve conter "{atributo}": "{valor}"'))
+def checar_atributo(atributo, valor):
+    response_json = mandar_requisicao.response.json()
+
+    assert "sala" in response_json, f"Erro; 'sala' não encontrado na resposta: \n{response_json}"
+
+    assert atributo in response_json["sala"], \
+        f"Erro: Atributo '{atributo}' não encontrado dentro de 'sala': \n{response_json}"
+
+    if valor.isdigit():
+        valor = int(valor)
+
+    valor_real = response_json["sala"][atributo]
+
+    assert valor_real == valor, (
+        f"Erro: Para '{atributo}', esperado '{valor}' ({type(valor).__name__}), "
+        f"obtido '{valor_real}' ({type(valor_real).__name__})"
+    )
+
+
+@then('o JSON da resposta deve conter uma lista de salas com todos os dados')
+def checar_lista_de_salas():
+    response_json = mandar_requisicao.response.json()
+
+    assert isinstance(response_json, list), f'Uma lista de salas não for retornada'
+    assert len(response_json) > 0, f'A lista de salas retornada está vazia'
+
+    for sala in response_json:
+        assert "id" in sala, f"'id' faltando: {sala}"
+        assert "tipo" in sala, f"'tipo' faltando: {sala}"
+        assert "lugares" in sala, f"'lugares' faltando: {sala}"
+        assert 'andar' in sala, f"'andar' faltando: {sala}"
+        assert 'equipamentos' in sala, f"'equipamentos' faltando: {sala}"
+        assert 'nome' in sala, f"'nome' faltando: {sala}"
diff --git a/backend/venv/bin/Activate.ps1 b/backend/venv/bin/Activate.ps1
new file mode 100644
index 00000000..b49d77ba
--- /dev/null
+++ b/backend/venv/bin/Activate.ps1
@@ -0,0 +1,247 @@
+<#
+.Synopsis
+Activate a Python virtual environment for the current PowerShell session.
+
+.Description
+Pushes the python executable for a virtual environment to the front of the
+$Env:PATH environment variable and sets the prompt to signify that you are
+in a Python virtual environment. Makes use of the command line switches as
+well as the `pyvenv.cfg` file values present in the virtual environment.
+
+.Parameter VenvDir
+Path to the directory that contains the virtual environment to activate. The
+default value for this is the parent of the directory that the Activate.ps1
+script is located within.
+
+.Parameter Prompt
+The prompt prefix to display when this virtual environment is activated. By
+default, this prompt is the name of the virtual environment folder (VenvDir)
+surrounded by parentheses and followed by a single space (ie. '(.venv) ').
+
+.Example
+Activate.ps1
+Activates the Python virtual environment that contains the Activate.ps1 script.
+
+.Example
+Activate.ps1 -Verbose
+Activates the Python virtual environment that contains the Activate.ps1 script,
+and shows extra information about the activation as it executes.
+
+.Example
+Activate.ps1 -VenvDir C:\Users\MyUser\Common\.venv
+Activates the Python virtual environment located in the specified location.
+
+.Example
+Activate.ps1 -Prompt "MyPython"
+Activates the Python virtual environment that contains the Activate.ps1 script,
+and prefixes the current prompt with the specified string (surrounded in
+parentheses) while the virtual environment is active.
+
+.Notes
+On Windows, it may be required to enable this Activate.ps1 script by setting the
+execution policy for the user. You can do this by issuing the following PowerShell
+command:
+
+PS C:\> Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
+
+For more information on Execution Policies: 
+https://go.microsoft.com/fwlink/?LinkID=135170
+
+#>
+Param(
+    [Parameter(Mandatory = $false)]
+    [String]
+    $VenvDir,
+    [Parameter(Mandatory = $false)]
+    [String]
+    $Prompt
+)
+
+<# Function declarations --------------------------------------------------- #>
+
+<#
+.Synopsis
+Remove all shell session elements added by the Activate script, including the
+addition of the virtual environment's Python executable from the beginning of
+the PATH variable.
+
+.Parameter NonDestructive
+If present, do not remove this function from the global namespace for the
+session.
+
+#>
+function global:deactivate ([switch]$NonDestructive) {
+    # Revert to original values
+
+    # The prior prompt:
+    if (Test-Path -Path Function:_OLD_VIRTUAL_PROMPT) {
+        Copy-Item -Path Function:_OLD_VIRTUAL_PROMPT -Destination Function:prompt
+        Remove-Item -Path Function:_OLD_VIRTUAL_PROMPT
+    }
+
+    # The prior PYTHONHOME:
+    if (Test-Path -Path Env:_OLD_VIRTUAL_PYTHONHOME) {
+        Copy-Item -Path Env:_OLD_VIRTUAL_PYTHONHOME -Destination Env:PYTHONHOME
+        Remove-Item -Path Env:_OLD_VIRTUAL_PYTHONHOME
+    }
+
+    # The prior PATH:
+    if (Test-Path -Path Env:_OLD_VIRTUAL_PATH) {
+        Copy-Item -Path Env:_OLD_VIRTUAL_PATH -Destination Env:PATH
+        Remove-Item -Path Env:_OLD_VIRTUAL_PATH
+    }
+
+    # Just remove the VIRTUAL_ENV altogether:
+    if (Test-Path -Path Env:VIRTUAL_ENV) {
+        Remove-Item -Path env:VIRTUAL_ENV
+    }
+
+    # Just remove VIRTUAL_ENV_PROMPT altogether.
+    if (Test-Path -Path Env:VIRTUAL_ENV_PROMPT) {
+        Remove-Item -Path env:VIRTUAL_ENV_PROMPT
+    }
+
+    # Just remove the _PYTHON_VENV_PROMPT_PREFIX altogether:
+    if (Get-Variable -Name "_PYTHON_VENV_PROMPT_PREFIX" -ErrorAction SilentlyContinue) {
+        Remove-Variable -Name _PYTHON_VENV_PROMPT_PREFIX -Scope Global -Force
+    }
+
+    # Leave deactivate function in the global namespace if requested:
+    if (-not $NonDestructive) {
+        Remove-Item -Path function:deactivate
+    }
+}
+
+<#
+.Description
+Get-PyVenvConfig parses the values from the pyvenv.cfg file located in the
+given folder, and returns them in a map.
+
+For each line in the pyvenv.cfg file, if that line can be parsed into exactly
+two strings separated by `=` (with any amount of whitespace surrounding the =)
+then it is considered a `key = value` line. The left hand string is the key,
+the right hand is the value.
+
+If the value starts with a `'` or a `"` then the first and last character is
+stripped from the value before being captured.
+
+.Parameter ConfigDir
+Path to the directory that contains the `pyvenv.cfg` file.
+#>
+function Get-PyVenvConfig(
+    [String]
+    $ConfigDir
+) {
+    Write-Verbose "Given ConfigDir=$ConfigDir, obtain values in pyvenv.cfg"
+
+    # Ensure the file exists, and issue a warning if it doesn't (but still allow the function to continue).
+    $pyvenvConfigPath = Join-Path -Resolve -Path $ConfigDir -ChildPath 'pyvenv.cfg' -ErrorAction Continue
+
+    # An empty map will be returned if no config file is found.
+    $pyvenvConfig = @{ }
+
+    if ($pyvenvConfigPath) {
+
+        Write-Verbose "File exists, parse `key = value` lines"
+        $pyvenvConfigContent = Get-Content -Path $pyvenvConfigPath
+
+        $pyvenvConfigContent | ForEach-Object {
+            $keyval = $PSItem -split "\s*=\s*", 2
+            if ($keyval[0] -and $keyval[1]) {
+                $val = $keyval[1]
+
+                # Remove extraneous quotations around a string value.
+                if ("'""".Contains($val.Substring(0, 1))) {
+                    $val = $val.Substring(1, $val.Length - 2)
+                }
+
+                $pyvenvConfig[$keyval[0]] = $val
+                Write-Verbose "Adding Key: '$($keyval[0])'='$val'"
+            }
+        }
+    }
+    return $pyvenvConfig
+}
+
+
+<# Begin Activate script --------------------------------------------------- #>
+
+# Determine the containing directory of this script
+$VenvExecPath = Split-Path -Parent $MyInvocation.MyCommand.Definition
+$VenvExecDir = Get-Item -Path $VenvExecPath
+
+Write-Verbose "Activation script is located in path: '$VenvExecPath'"
+Write-Verbose "VenvExecDir Fullname: '$($VenvExecDir.FullName)"
+Write-Verbose "VenvExecDir Name: '$($VenvExecDir.Name)"
+
+# Set values required in priority: CmdLine, ConfigFile, Default
+# First, get the location of the virtual environment, it might not be
+# VenvExecDir if specified on the command line.
+if ($VenvDir) {
+    Write-Verbose "VenvDir given as parameter, using '$VenvDir' to determine values"
+}
+else {
+    Write-Verbose "VenvDir not given as a parameter, using parent directory name as VenvDir."
+    $VenvDir = $VenvExecDir.Parent.FullName.TrimEnd("\\/")
+    Write-Verbose "VenvDir=$VenvDir"
+}
+
+# Next, read the `pyvenv.cfg` file to determine any required value such
+# as `prompt`.
+$pyvenvCfg = Get-PyVenvConfig -ConfigDir $VenvDir
+
+# Next, set the prompt from the command line, or the config file, or
+# just use the name of the virtual environment folder.
+if ($Prompt) {
+    Write-Verbose "Prompt specified as argument, using '$Prompt'"
+}
+else {
+    Write-Verbose "Prompt not specified as argument to script, checking pyvenv.cfg value"
+    if ($pyvenvCfg -and $pyvenvCfg['prompt']) {
+        Write-Verbose "  Setting based on value in pyvenv.cfg='$($pyvenvCfg['prompt'])'"
+        $Prompt = $pyvenvCfg['prompt'];
+    }
+    else {
+        Write-Verbose "  Setting prompt based on parent's directory's name. (Is the directory name passed to venv module when creating the virtual environment)"
+        Write-Verbose "  Got leaf-name of $VenvDir='$(Split-Path -Path $venvDir -Leaf)'"
+        $Prompt = Split-Path -Path $venvDir -Leaf
+    }
+}
+
+Write-Verbose "Prompt = '$Prompt'"
+Write-Verbose "VenvDir='$VenvDir'"
+
+# Deactivate any currently active virtual environment, but leave the
+# deactivate function in place.
+deactivate -nondestructive
+
+# Now set the environment variable VIRTUAL_ENV, used by many tools to determine
+# that there is an activated venv.
+$env:VIRTUAL_ENV = $VenvDir
+
+if (-not $Env:VIRTUAL_ENV_DISABLE_PROMPT) {
+
+    Write-Verbose "Setting prompt to '$Prompt'"
+
+    # Set the prompt to include the env name
+    # Make sure _OLD_VIRTUAL_PROMPT is global
+    function global:_OLD_VIRTUAL_PROMPT { "" }
+    Copy-Item -Path function:prompt -Destination function:_OLD_VIRTUAL_PROMPT
+    New-Variable -Name _PYTHON_VENV_PROMPT_PREFIX -Description "Python virtual environment prompt prefix" -Scope Global -Option ReadOnly -Visibility Public -Value $Prompt
+
+    function global:prompt {
+        Write-Host -NoNewline -ForegroundColor Green "($_PYTHON_VENV_PROMPT_PREFIX) "
+        _OLD_VIRTUAL_PROMPT
+    }
+    $env:VIRTUAL_ENV_PROMPT = $Prompt
+}
+
+# Clear PYTHONHOME
+if (Test-Path -Path Env:PYTHONHOME) {
+    Copy-Item -Path Env:PYTHONHOME -Destination Env:_OLD_VIRTUAL_PYTHONHOME
+    Remove-Item -Path Env:PYTHONHOME
+}
+
+# Add the venv to the PATH
+Copy-Item -Path Env:PATH -Destination Env:_OLD_VIRTUAL_PATH
+$Env:PATH = "$VenvExecDir$([System.IO.Path]::PathSeparator)$Env:PATH"
diff --git a/backend/venv/bin/activate b/backend/venv/bin/activate
new file mode 100644
index 00000000..eb59a125
--- /dev/null
+++ b/backend/venv/bin/activate
@@ -0,0 +1,70 @@
+# This file must be used with "source bin/activate" *from bash*
+# You cannot run it directly
+
+deactivate () {
+    # reset old environment variables
+    if [ -n "${_OLD_VIRTUAL_PATH:-}" ] ; then
+        PATH="${_OLD_VIRTUAL_PATH:-}"
+        export PATH
+        unset _OLD_VIRTUAL_PATH
+    fi
+    if [ -n "${_OLD_VIRTUAL_PYTHONHOME:-}" ] ; then
+        PYTHONHOME="${_OLD_VIRTUAL_PYTHONHOME:-}"
+        export PYTHONHOME
+        unset _OLD_VIRTUAL_PYTHONHOME
+    fi
+
+    # Call hash to forget past commands. Without forgetting
+    # past commands the $PATH changes we made may not be respected
+    hash -r 2> /dev/null
+
+    if [ -n "${_OLD_VIRTUAL_PS1:-}" ] ; then
+        PS1="${_OLD_VIRTUAL_PS1:-}"
+        export PS1
+        unset _OLD_VIRTUAL_PS1
+    fi
+
+    unset VIRTUAL_ENV
+    unset VIRTUAL_ENV_PROMPT
+    if [ ! "${1:-}" = "nondestructive" ] ; then
+    # Self destruct!
+        unset -f deactivate
+    fi
+}
+
+# unset irrelevant variables
+deactivate nondestructive
+
+# on Windows, a path can contain colons and backslashes and has to be converted:
+if [ "${OSTYPE:-}" = "cygwin" ] || [ "${OSTYPE:-}" = "msys" ] ; then
+    # transform D:\path\to\venv to /d/path/to/venv on MSYS
+    # and to /cygdrive/d/path/to/venv on Cygwin
+    export VIRTUAL_ENV=$(cygpath "/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv")
+else
+    # use the path as-is
+    export VIRTUAL_ENV="/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv"
+fi
+
+_OLD_VIRTUAL_PATH="$PATH"
+PATH="$VIRTUAL_ENV/bin:$PATH"
+export PATH
+
+# unset PYTHONHOME if set
+# this will fail if PYTHONHOME is set to the empty string (which is bad anyway)
+# could use `if (set -u; : $PYTHONHOME) ;` in bash
+if [ -n "${PYTHONHOME:-}" ] ; then
+    _OLD_VIRTUAL_PYTHONHOME="${PYTHONHOME:-}"
+    unset PYTHONHOME
+fi
+
+if [ -z "${VIRTUAL_ENV_DISABLE_PROMPT:-}" ] ; then
+    _OLD_VIRTUAL_PS1="${PS1:-}"
+    PS1="(venv) ${PS1:-}"
+    export PS1
+    VIRTUAL_ENV_PROMPT="(venv) "
+    export VIRTUAL_ENV_PROMPT
+fi
+
+# Call hash to forget past commands. Without forgetting
+# past commands the $PATH changes we made may not be respected
+hash -r 2> /dev/null
diff --git a/backend/venv/bin/activate.csh b/backend/venv/bin/activate.csh
new file mode 100644
index 00000000..1ae0f43f
--- /dev/null
+++ b/backend/venv/bin/activate.csh
@@ -0,0 +1,27 @@
+# This file must be used with "source bin/activate.csh" *from csh*.
+# You cannot run it directly.
+
+# Created by Davide Di Blasi <davidedb@gmail.com>.
+# Ported to Python 3.3 venv by Andrew Svetlov <andrew.svetlov@gmail.com>
+
+alias deactivate 'test $?_OLD_VIRTUAL_PATH != 0 && setenv PATH "$_OLD_VIRTUAL_PATH" && unset _OLD_VIRTUAL_PATH; rehash; test $?_OLD_VIRTUAL_PROMPT != 0 && set prompt="$_OLD_VIRTUAL_PROMPT" && unset _OLD_VIRTUAL_PROMPT; unsetenv VIRTUAL_ENV; unsetenv VIRTUAL_ENV_PROMPT; test "\!:*" != "nondestructive" && unalias deactivate'
+
+# Unset irrelevant variables.
+deactivate nondestructive
+
+setenv VIRTUAL_ENV "/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv"
+
+set _OLD_VIRTUAL_PATH="$PATH"
+setenv PATH "$VIRTUAL_ENV/bin:$PATH"
+
+
+set _OLD_VIRTUAL_PROMPT="$prompt"
+
+if (! "$?VIRTUAL_ENV_DISABLE_PROMPT") then
+    set prompt = "(venv) $prompt"
+    setenv VIRTUAL_ENV_PROMPT "(venv) "
+endif
+
+alias pydoc python -m pydoc
+
+rehash
diff --git a/backend/venv/bin/activate.fish b/backend/venv/bin/activate.fish
new file mode 100644
index 00000000..fd1e9f40
--- /dev/null
+++ b/backend/venv/bin/activate.fish
@@ -0,0 +1,69 @@
+# This file must be used with "source <venv>/bin/activate.fish" *from fish*
+# (https://fishshell.com/). You cannot run it directly.
+
+function deactivate  -d "Exit virtual environment and return to normal shell environment"
+    # reset old environment variables
+    if test -n "$_OLD_VIRTUAL_PATH"
+        set -gx PATH $_OLD_VIRTUAL_PATH
+        set -e _OLD_VIRTUAL_PATH
+    end
+    if test -n "$_OLD_VIRTUAL_PYTHONHOME"
+        set -gx PYTHONHOME $_OLD_VIRTUAL_PYTHONHOME
+        set -e _OLD_VIRTUAL_PYTHONHOME
+    end
+
+    if test -n "$_OLD_FISH_PROMPT_OVERRIDE"
+        set -e _OLD_FISH_PROMPT_OVERRIDE
+        # prevents error when using nested fish instances (Issue #93858)
+        if functions -q _old_fish_prompt
+            functions -e fish_prompt
+            functions -c _old_fish_prompt fish_prompt
+            functions -e _old_fish_prompt
+        end
+    end
+
+    set -e VIRTUAL_ENV
+    set -e VIRTUAL_ENV_PROMPT
+    if test "$argv[1]" != "nondestructive"
+        # Self-destruct!
+        functions -e deactivate
+    end
+end
+
+# Unset irrelevant variables.
+deactivate nondestructive
+
+set -gx VIRTUAL_ENV "/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv"
+
+set -gx _OLD_VIRTUAL_PATH $PATH
+set -gx PATH "$VIRTUAL_ENV/bin" $PATH
+
+# Unset PYTHONHOME if set.
+if set -q PYTHONHOME
+    set -gx _OLD_VIRTUAL_PYTHONHOME $PYTHONHOME
+    set -e PYTHONHOME
+end
+
+if test -z "$VIRTUAL_ENV_DISABLE_PROMPT"
+    # fish uses a function instead of an env var to generate the prompt.
+
+    # Save the current fish_prompt function as the function _old_fish_prompt.
+    functions -c fish_prompt _old_fish_prompt
+
+    # With the original prompt function renamed, we can override with our own.
+    function fish_prompt
+        # Save the return status of the last command.
+        set -l old_status $status
+
+        # Output the venv prompt; color taken from the blue of the Python logo.
+        printf "%s%s%s" (set_color 4B8BBE) "(venv) " (set_color normal)
+
+        # Restore the return status of the previous command.
+        echo "exit $old_status" | .
+        # Output the original/"old" prompt.
+        _old_fish_prompt
+    end
+
+    set -gx _OLD_FISH_PROMPT_OVERRIDE "$VIRTUAL_ENV"
+    set -gx VIRTUAL_ENV_PROMPT "(venv) "
+end
diff --git a/backend/venv/bin/flask b/backend/venv/bin/flask
new file mode 100755
index 00000000..38f5d8dd
--- /dev/null
+++ b/backend/venv/bin/flask
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from flask.cli import main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(main())
diff --git a/backend/venv/bin/gherkin b/backend/venv/bin/gherkin
new file mode 100755
index 00000000..b211ef27
--- /dev/null
+++ b/backend/venv/bin/gherkin
@@ -0,0 +1,20 @@
+#!/usr/bin/env sh
+# Use "make GHERKIN_PYTHON_VERSION=python2 ..." to use python2
+if [ -z "$GHERKIN_PYTHON_VERSION" ];
+then
+    if [ -x "$(command -v python)" ]
+    then
+        GHERKIN_PYTHON_VERSION=python
+    elif [ -x "$(command -v python3)" ]
+    then
+        GHERKIN_PYTHON_VERSION=python3
+    elif [ -x "$(command -v python2)" ]
+    then
+        GHERKIN_PYTHON_VERSION=python2
+    else
+        echo "Neiter python, python3 or python2 found on PATH, exiting"
+        exit 1
+    fi
+fi
+
+$GHERKIN_PYTHON_VERSION -m gherkin $@;
diff --git a/backend/venv/bin/mako-render b/backend/venv/bin/mako-render
new file mode 100755
index 00000000..167f8bf0
--- /dev/null
+++ b/backend/venv/bin/mako-render
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from mako.cmd import cmdline
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(cmdline())
diff --git a/backend/venv/bin/normalizer b/backend/venv/bin/normalizer
new file mode 100755
index 00000000..a859d073
--- /dev/null
+++ b/backend/venv/bin/normalizer
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from charset_normalizer import cli
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(cli.cli_detect())
diff --git a/backend/venv/bin/pip b/backend/venv/bin/pip
new file mode 100755
index 00000000..16fbdb39
--- /dev/null
+++ b/backend/venv/bin/pip
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pip._internal.cli.main import main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(main())
diff --git a/backend/venv/bin/pip3 b/backend/venv/bin/pip3
new file mode 100755
index 00000000..16fbdb39
--- /dev/null
+++ b/backend/venv/bin/pip3
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pip._internal.cli.main import main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(main())
diff --git a/backend/venv/bin/pip3.12 b/backend/venv/bin/pip3.12
new file mode 100755
index 00000000..16fbdb39
--- /dev/null
+++ b/backend/venv/bin/pip3.12
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pip._internal.cli.main import main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(main())
diff --git a/backend/venv/bin/py.test b/backend/venv/bin/py.test
new file mode 100755
index 00000000..be8beb47
--- /dev/null
+++ b/backend/venv/bin/py.test
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pytest import console_main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(console_main())
diff --git a/backend/venv/bin/pytest b/backend/venv/bin/pytest
new file mode 100755
index 00000000..be8beb47
--- /dev/null
+++ b/backend/venv/bin/pytest
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pytest import console_main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(console_main())
diff --git a/backend/venv/bin/pytest-bdd b/backend/venv/bin/pytest-bdd
new file mode 100755
index 00000000..6f9455f8
--- /dev/null
+++ b/backend/venv/bin/pytest-bdd
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from pytest_bdd.scripts import main
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(main())
diff --git a/backend/venv/bin/python b/backend/venv/bin/python
new file mode 120000
index 00000000..11b9d885
--- /dev/null
+++ b/backend/venv/bin/python
@@ -0,0 +1 @@
+python3.12
\ No newline at end of file
diff --git a/backend/venv/bin/python3 b/backend/venv/bin/python3
new file mode 120000
index 00000000..11b9d885
--- /dev/null
+++ b/backend/venv/bin/python3
@@ -0,0 +1 @@
+python3.12
\ No newline at end of file
diff --git a/backend/venv/bin/python3.12 b/backend/venv/bin/python3.12
new file mode 120000
index 00000000..a7a5fcca
--- /dev/null
+++ b/backend/venv/bin/python3.12
@@ -0,0 +1 @@
+/Library/Frameworks/Python.framework/Versions/3.12/bin/python3.12
\ No newline at end of file
diff --git a/backend/venv/bin/tree-cli b/backend/venv/bin/tree-cli
new file mode 100755
index 00000000..063bbff8
--- /dev/null
+++ b/backend/venv/bin/tree-cli
@@ -0,0 +1,8 @@
+#!/Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv/bin/python3.12
+# -*- coding: utf-8 -*-
+import re
+import sys
+from Tree.cli import create_tree
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(create_tree())
diff --git a/backend/venv/include/site/python3.12/greenlet/greenlet.h b/backend/venv/include/site/python3.12/greenlet/greenlet.h
new file mode 100644
index 00000000..d02a16e4
--- /dev/null
+++ b/backend/venv/include/site/python3.12/greenlet/greenlet.h
@@ -0,0 +1,164 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+
+/* Greenlet object interface */
+
+#ifndef Py_GREENLETOBJECT_H
+#define Py_GREENLETOBJECT_H
+
+
+#include <Python.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* This is deprecated and undocumented. It does not change. */
+#define GREENLET_VERSION "1.0.0"
+
+#ifndef GREENLET_MODULE
+#define implementation_ptr_t void*
+#endif
+
+typedef struct _greenlet {
+    PyObject_HEAD
+    PyObject* weakreflist;
+    PyObject* dict;
+    implementation_ptr_t pimpl;
+} PyGreenlet;
+
+#define PyGreenlet_Check(op) (op && PyObject_TypeCheck(op, &PyGreenlet_Type))
+
+
+/* C API functions */
+
+/* Total number of symbols that are exported */
+#define PyGreenlet_API_pointers 12
+
+#define PyGreenlet_Type_NUM 0
+#define PyExc_GreenletError_NUM 1
+#define PyExc_GreenletExit_NUM 2
+
+#define PyGreenlet_New_NUM 3
+#define PyGreenlet_GetCurrent_NUM 4
+#define PyGreenlet_Throw_NUM 5
+#define PyGreenlet_Switch_NUM 6
+#define PyGreenlet_SetParent_NUM 7
+
+#define PyGreenlet_MAIN_NUM 8
+#define PyGreenlet_STARTED_NUM 9
+#define PyGreenlet_ACTIVE_NUM 10
+#define PyGreenlet_GET_PARENT_NUM 11
+
+#ifndef GREENLET_MODULE
+/* This section is used by modules that uses the greenlet C API */
+static void** _PyGreenlet_API = NULL;
+
+#    define PyGreenlet_Type \
+        (*(PyTypeObject*)_PyGreenlet_API[PyGreenlet_Type_NUM])
+
+#    define PyExc_GreenletError \
+        ((PyObject*)_PyGreenlet_API[PyExc_GreenletError_NUM])
+
+#    define PyExc_GreenletExit \
+        ((PyObject*)_PyGreenlet_API[PyExc_GreenletExit_NUM])
+
+/*
+ * PyGreenlet_New(PyObject *args)
+ *
+ * greenlet.greenlet(run, parent=None)
+ */
+#    define PyGreenlet_New                                        \
+        (*(PyGreenlet * (*)(PyObject * run, PyGreenlet * parent)) \
+             _PyGreenlet_API[PyGreenlet_New_NUM])
+
+/*
+ * PyGreenlet_GetCurrent(void)
+ *
+ * greenlet.getcurrent()
+ */
+#    define PyGreenlet_GetCurrent \
+        (*(PyGreenlet * (*)(void)) _PyGreenlet_API[PyGreenlet_GetCurrent_NUM])
+
+/*
+ * PyGreenlet_Throw(
+ *         PyGreenlet *greenlet,
+ *         PyObject *typ,
+ *         PyObject *val,
+ *         PyObject *tb)
+ *
+ * g.throw(...)
+ */
+#    define PyGreenlet_Throw                 \
+        (*(PyObject * (*)(PyGreenlet * self, \
+                          PyObject * typ,    \
+                          PyObject * val,    \
+                          PyObject * tb))    \
+             _PyGreenlet_API[PyGreenlet_Throw_NUM])
+
+/*
+ * PyGreenlet_Switch(PyGreenlet *greenlet, PyObject *args)
+ *
+ * g.switch(*args, **kwargs)
+ */
+#    define PyGreenlet_Switch                                              \
+        (*(PyObject *                                                      \
+           (*)(PyGreenlet * greenlet, PyObject * args, PyObject * kwargs)) \
+             _PyGreenlet_API[PyGreenlet_Switch_NUM])
+
+/*
+ * PyGreenlet_SetParent(PyObject *greenlet, PyObject *new_parent)
+ *
+ * g.parent = new_parent
+ */
+#    define PyGreenlet_SetParent                                 \
+        (*(int (*)(PyGreenlet * greenlet, PyGreenlet * nparent)) \
+             _PyGreenlet_API[PyGreenlet_SetParent_NUM])
+
+/*
+ * PyGreenlet_GetParent(PyObject* greenlet)
+ *
+ * return greenlet.parent;
+ *
+ * This could return NULL even if there is no exception active.
+ * If it does not return NULL, you are responsible for decrementing the
+ * reference count.
+ */
+#     define PyGreenlet_GetParent                                    \
+    (*(PyGreenlet* (*)(PyGreenlet*))                                 \
+     _PyGreenlet_API[PyGreenlet_GET_PARENT_NUM])
+
+/*
+ * deprecated, undocumented alias.
+ */
+#     define PyGreenlet_GET_PARENT PyGreenlet_GetParent
+
+#     define PyGreenlet_MAIN                                         \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_MAIN_NUM])
+
+#     define PyGreenlet_STARTED                                      \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_STARTED_NUM])
+
+#     define PyGreenlet_ACTIVE                                       \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_ACTIVE_NUM])
+
+
+
+
+/* Macro that imports greenlet and initializes C API */
+/* NOTE: This has actually moved to ``greenlet._greenlet._C_API``, but we
+   keep the older definition to be sure older code that might have a copy of
+   the header still works. */
+#    define PyGreenlet_Import()                                               \
+        {                                                                     \
+            _PyGreenlet_API = (void**)PyCapsule_Import("greenlet._C_API", 0); \
+        }
+
+#endif /* GREENLET_MODULE */
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_GREENLETOBJECT_H */
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/LICENSE
new file mode 100644
index 00000000..46d932f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/LICENSE
@@ -0,0 +1,7 @@
+Copyright (C) 2016 Cory Dolphin, Olin College
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/METADATA
new file mode 100644
index 00000000..99902fe9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/METADATA
@@ -0,0 +1,148 @@
+Metadata-Version: 2.1
+Name: Flask-Cors
+Version: 5.0.0
+Summary: A Flask extension adding a decorator for CORS support
+Home-page: https://github.com/corydolphin/flask-cors
+Author: Cory Dolphin
+Author-email: corydolphin@gmail.com
+License: MIT
+Platform: any
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+License-File: LICENSE
+Requires-Dist: Flask >=0.9
+
+Flask-CORS
+==========
+
+|Build Status| |Latest Version| |Supported Python versions|
+|License|
+
+A Flask extension for handling Cross Origin Resource Sharing (CORS), making cross-origin AJAX possible.
+
+This package has a simple philosophy: when you want to enable CORS, you wish to enable it for all use cases on a domain.
+This means no mucking around with different allowed headers, methods, etc.
+
+By default, submission of cookies across domains is disabled due to the security implications.
+Please see the documentation for how to enable credential'ed requests, and please make sure you add some sort of `CSRF <http://en.wikipedia.org/wiki/Cross-site_request_forgery>`__ protection before doing so!
+
+Installation
+------------
+
+Install the extension with using pip, or easy\_install.
+
+.. code:: bash
+
+    $ pip install -U flask-cors
+
+Usage
+-----
+
+This package exposes a Flask extension which by default enables CORS support on all routes, for all origins and methods.
+It allows parameterization of all CORS headers on a per-resource level.
+The package also contains a decorator, for those who prefer this approach.
+
+Simple Usage
+~~~~~~~~~~~~
+
+In the simplest case, initialize the Flask-Cors extension with default arguments in order to allow CORS for all domains on all routes.
+See the full list of options in the `documentation <https://flask-cors.corydolphin.com/en/latest/api.html#extension>`__.
+
+.. code:: python
+
+
+    from flask import Flask
+    from flask_cors import CORS
+
+    app = Flask(__name__)
+    CORS(app)
+
+    @app.route("/")
+    def helloWorld():
+      return "Hello, cross-origin-world!"
+
+Resource specific CORS
+^^^^^^^^^^^^^^^^^^^^^^
+
+Alternatively, you can specify CORS options on a resource and origin level of granularity by passing a dictionary as the `resources` option, mapping paths to a set of options.
+See the full list of options in the `documentation <https://flask-cors.corydolphin.com/en/latest/api.html#extension>`__.
+
+.. code:: python
+
+    app = Flask(__name__)
+    cors = CORS(app, resources={r"/api/*": {"origins": "*"}})
+
+    @app.route("/api/v1/users")
+    def list_users():
+      return "user example"
+
+Route specific CORS via decorator
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+This extension also exposes a simple decorator to decorate flask routes with.
+Simply add ``@cross_origin()`` below a call to Flask's ``@app.route(..)`` to allow CORS on a given route.
+See the full list of options in the `decorator documentation <https://flask-cors.corydolphin.com/en/latest/api.html#decorator>`__.
+
+.. code:: python
+
+    @app.route("/")
+    @cross_origin()
+    def helloWorld():
+      return "Hello, cross-origin-world!"
+
+Documentation
+-------------
+
+For a full list of options, please see the full `documentation <https://flask-cors.corydolphin.com/en/latest/api.html>`__
+
+Troubleshooting
+---------------
+
+If things aren't working as you expect, enable logging to help understand what is going on under the hood, and why.
+
+.. code:: python
+
+    logging.getLogger('flask_cors').level = logging.DEBUG
+
+
+Tests
+-----
+
+A simple set of tests is included in ``test/``.
+To run, install nose, and simply invoke ``nosetests`` or ``python setup.py test`` to exercise the tests.
+
+If nosetests does not work for you, due to it no longer working with newer python versions.
+You can use pytest to run the tests instead.
+
+Contributing
+------------
+
+Questions, comments or improvements?
+Please create an issue on `Github <https://github.com/corydolphin/flask-cors>`__, tweet at `@corydolphin <https://twitter.com/corydolphin>`__ or send me an email.
+I do my best to include every contribution proposed in any way that I can.
+
+Credits
+-------
+
+This Flask extension is based upon the `Decorator for the HTTP Access Control <https://web.archive.org/web/20190128010149/http://flask.pocoo.org/snippets/56/>`__ written by Armin Ronacher.
+
+.. |Build Status| image:: https://github.com/corydolphin/flask-cors/actions/workflows/unittests.yaml/badge.svg
+   :target: https://travis-ci.org/corydolphin/flask-cors
+.. |Latest Version| image:: https://img.shields.io/pypi/v/Flask-Cors.svg
+   :target: https://pypi.python.org/pypi/Flask-Cors/
+.. |Supported Python versions| image:: https://img.shields.io/pypi/pyversions/Flask-Cors.svg
+   :target: https://img.shields.io/pypi/pyversions/Flask-Cors.svg
+.. |License| image:: http://img.shields.io/:license-mit-blue.svg
+   :target: https://pypi.python.org/pypi/Flask-Cors/
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/RECORD
new file mode 100644
index 00000000..2a59f843
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/RECORD
@@ -0,0 +1,17 @@
+Flask_Cors-5.0.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+Flask_Cors-5.0.0.dist-info/LICENSE,sha256=bhob3FSDTB4HQMvOXV9vLK4chG_Sp_SCsRZJWU-vvV0,1069
+Flask_Cors-5.0.0.dist-info/METADATA,sha256=V2L_s849dFlZXsOhcgXVqv5Slj_JKSVuiiuRgDOft5s,5474
+Flask_Cors-5.0.0.dist-info/RECORD,,
+Flask_Cors-5.0.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+Flask_Cors-5.0.0.dist-info/WHEEL,sha256=WDDPHYzpiOIm6GP1C2_8y8W6q16ICddAgOHlhTje9Qc,109
+Flask_Cors-5.0.0.dist-info/top_level.txt,sha256=aWye_0QNZPp_QtPF4ZluLHqnyVLT9CPJsfiGhwqkWuo,11
+flask_cors/__init__.py,sha256=wZDCvPTHspA2g1VV7KyKN7R-uCdBnirTlsCzgPDcQtI,792
+flask_cors/__pycache__/__init__.cpython-312.pyc,,
+flask_cors/__pycache__/core.cpython-312.pyc,,
+flask_cors/__pycache__/decorator.cpython-312.pyc,,
+flask_cors/__pycache__/extension.cpython-312.pyc,,
+flask_cors/__pycache__/version.cpython-312.pyc,,
+flask_cors/core.py,sha256=y76xxLasWTdV_3ka19IxpdJPOgROBZQZ5L8t20IjqRA,14252
+flask_cors/decorator.py,sha256=BeJsyX1wYhVKWN04FAhb6z8YqffiRr7wKqwzHPap4bw,5009
+flask_cors/extension.py,sha256=gzv6zWUwSDYlGHBWzMuTI_hoQ7gQmp9DlcAcrKTVHdw,8602
+flask_cors/version.py,sha256=JzYPYpvaglqIJRGCDrh5-hYmXI0ISrDDed0V1QQZAGU,22
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/WHEEL
new file mode 100644
index 00000000..5850bf02
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/WHEEL
@@ -0,0 +1,6 @@
+Wheel-Version: 1.0
+Generator: setuptools (74.0.0)
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/top_level.txt
new file mode 100644
index 00000000..27af9881
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_Cors-5.0.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+flask_cors
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/LICENSE
new file mode 100644
index 00000000..53090c15
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2016 Lily Acadia Gilbert
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/METADATA
new file mode 100644
index 00000000..2036b282
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/METADATA
@@ -0,0 +1,113 @@
+Metadata-Version: 2.1
+Name: Flask-JWT-Extended
+Version: 4.7.1
+Summary: Extended JWT integration with Flask
+Home-page: https://github.com/vimalloc/flask-jwt-extended
+Author: Lily Acadia Gilbert
+Author-email: lily.gilbert@hey.com
+License: MIT
+Keywords: flask,jwt,json web token
+Platform: any
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Framework :: Flask
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.9,<4
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: Werkzeug>=0.14
+Requires-Dist: Flask<4.0,>=2.0
+Requires-Dist: PyJWT<3.0,>=2.0
+Provides-Extra: asymmetric-crypto
+Requires-Dist: cryptography>=3.3.1; extra == "asymmetric-crypto"
+
+# Flask-JWT-Extended
+
+### Features
+
+Flask-JWT-Extended not only adds support for using JSON Web Tokens (JWT) to Flask for protecting routes,
+but also many helpful (and **optional**) features built in to make working with JSON Web Tokens
+easier. These include:
+
+-   Adding custom claims to JSON Web Tokens
+-   Automatic user loading (`current_user`).
+-   Custom claims validation on received tokens
+-   [Refresh tokens](https://auth0.com/blog/refresh-tokens-what-are-they-and-when-to-use-them/)
+-   First class support for fresh tokens for making sensitive changes.
+-   Token revoking/blocklisting
+-   Storing tokens in cookies and CSRF protection
+
+### Usage
+
+[View the documentation online](https://flask-jwt-extended.readthedocs.io/en/stable/)
+
+### Upgrading from 3.x.x to 4.0.0
+
+[View the changes](https://flask-jwt-extended.readthedocs.io/en/stable/v4_upgrade_guide/)
+
+### Changelog
+
+You can view the changelog [here](https://github.com/vimalloc/flask-jwt-extended/releases).
+This project follows [semantic versioning](https://semver.org/).
+
+### Chatting
+
+Come chat with the community or ask questions at https://discord.gg/EJBsbFd
+
+### Contributing
+
+Before making any changes, make sure to install the development requirements
+and setup the git hooks which will automatically lint and format your changes.
+
+```bash
+pip install -r requirements.txt
+pre-commit install
+```
+
+We require 100% code coverage in our unit tests. You can run the tests locally
+with `tox` which ensures that all tests pass, tests provide complete code coverage,
+documentation builds, and style guide are adhered to
+
+```bash
+tox
+```
+
+A subset of checks can also be ran by adding an argument to tox. The available
+arguments are:
+
+-   py37, py38, py39, py310, py311, py312, pypy3
+    -   Run unit tests on the given python version
+-   mypy
+    -   Run mypy type checking
+-   coverage
+    -   Run a code coverage check
+-   docs
+    -   Ensure documentation builds and there are no broken links
+-   style
+    -   Ensure style guide is adhered to
+
+```bash
+tox -e py38
+```
+
+We also require features to be well documented. You can generate a local copy
+of the documentation by going to the `docs` directory and running:
+
+```bash
+make clean && make html && open _build/html/index.html
+```
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/RECORD
new file mode 100644
index 00000000..d6bcde87
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/RECORD
@@ -0,0 +1,28 @@
+Flask_JWT_Extended-4.7.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+Flask_JWT_Extended-4.7.1.dist-info/LICENSE,sha256=IZ7JSdUnG3ZAFNkV8wskplpL1tTu-ImxNFJM2Zfci5g,1076
+Flask_JWT_Extended-4.7.1.dist-info/METADATA,sha256=GuhF8kzZOl9nKVkGgeRx5X7KEa6FiNpv9xjRWZgiz9I,3812
+Flask_JWT_Extended-4.7.1.dist-info/RECORD,,
+Flask_JWT_Extended-4.7.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+Flask_JWT_Extended-4.7.1.dist-info/WHEEL,sha256=pxeNX5JdtCe58PUSYP9upmc7jdRPgvT0Gm9kb1SHlVw,109
+Flask_JWT_Extended-4.7.1.dist-info/top_level.txt,sha256=ScNvc4cNmrFFSnAe_rENqiMpfhog1M2HkaUaydRBbUU,19
+flask_jwt_extended/__init__.py,sha256=q3b7uG1tp7jXv9L2WokVrh920M7J8mRGRQKaCc2KbNo,1179
+flask_jwt_extended/__pycache__/__init__.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/config.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/default_callbacks.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/exceptions.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/internal_utils.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/jwt_manager.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/tokens.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/typing.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/utils.cpython-312.pyc,,
+flask_jwt_extended/__pycache__/view_decorators.cpython-312.pyc,,
+flask_jwt_extended/config.py,sha256=8hvI2WEUPcDRFumnWACjPQSNm5Sn10RfVzagP4Tv46g,10083
+flask_jwt_extended/default_callbacks.py,sha256=S-zUiOS5nMiC4n3URpUpGf9LfNA5bT2kWHVJdp14Ngs,5273
+flask_jwt_extended/exceptions.py,sha256=kleki6CySt5u-BiAwJ-uskWb8rySnXHLC-w-YvMqyk4,2400
+flask_jwt_extended/internal_utils.py,sha256=XDnS8GLQ2oaZka5FbfiuE9KBgxAmmHssuKR3mLDP1Jk,3350
+flask_jwt_extended/jwt_manager.py,sha256=OjTJhl2xPVq3Z0FH874NNtHR4VeOl2y75dnQLdPL8GM,23989
+flask_jwt_extended/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flask_jwt_extended/tokens.py,sha256=GOreSqdqx1rAg3nV_DMn2ltqAngLpu2281zx6Y6V2M4,3131
+flask_jwt_extended/typing.py,sha256=l4nGV2_ARq2_BUK6CjnrOp5i43qXJ3yqLPy8YOs1soA,170
+flask_jwt_extended/utils.py,sha256=iA6Fk-S4DDmRruKNCwkB6KK70cuPWKOTVb8vsKUq2P4,15964
+flask_jwt_extended/view_decorators.py,sha256=gtDCZ4-3e6eldJREMukIjD9SzoYa1bScKdInUtPIhIs,13583
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/WHEEL
new file mode 100644
index 00000000..104f3874
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/WHEEL
@@ -0,0 +1,6 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/top_level.txt
new file mode 100644
index 00000000..ff6d8841
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Flask_JWT_Extended-4.7.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+flask_jwt_extended
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/LICENSE
new file mode 100644
index 00000000..4c18adee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/LICENSE
@@ -0,0 +1,19 @@
+Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+of the Software, and to permit persons to whom the Software is furnished to do
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/METADATA
new file mode 100644
index 00000000..44211198
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/METADATA
@@ -0,0 +1,87 @@
+Metadata-Version: 2.2
+Name: Mako
+Version: 1.3.9
+Summary: A super-fast templating language that borrows the best ideas from the existing templating languages.
+Home-page: https://www.makotemplates.org/
+Author: Mike Bayer
+Author-email: mike@zzzcomputing.com
+License: MIT
+Project-URL: Documentation, https://docs.makotemplates.org
+Project-URL: Issue Tracker, https://github.com/sqlalchemy/mako
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: MarkupSafe>=0.9.2
+Provides-Extra: testing
+Requires-Dist: pytest; extra == "testing"
+Provides-Extra: babel
+Requires-Dist: Babel; extra == "babel"
+Provides-Extra: lingua
+Requires-Dist: lingua; extra == "lingua"
+
+=========================
+Mako Templates for Python
+=========================
+
+Mako is a template library written in Python. It provides a familiar, non-XML 
+syntax which compiles into Python modules for maximum performance. Mako's 
+syntax and API borrows from the best ideas of many others, including Django
+templates, Cheetah, Myghty, and Genshi. Conceptually, Mako is an embedded 
+Python (i.e. Python Server Page) language, which refines the familiar ideas
+of componentized layout and inheritance to produce one of the most 
+straightforward and flexible models available, while also maintaining close 
+ties to Python calling and scoping semantics.
+
+Nutshell
+========
+
+::
+
+    <%inherit file="base.html"/>
+    <%
+        rows = [[v for v in range(0,10)] for row in range(0,10)]
+    %>
+    <table>
+        % for row in rows:
+            ${makerow(row)}
+        % endfor
+    </table>
+
+    <%def name="makerow(row)">
+        <tr>
+        % for name in row:
+            <td>${name}</td>\
+        % endfor
+        </tr>
+    </%def>
+
+Philosophy
+===========
+
+Python is a great scripting language. Don't reinvent the wheel...your templates can handle it !
+
+Documentation
+==============
+
+See documentation for Mako at https://docs.makotemplates.org/en/latest/
+
+License
+========
+
+Mako is licensed under an MIT-style license (see LICENSE).
+Other incorporated projects may be licensed under different licenses.
+All licenses allow for non-commercial and commercial use.
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/RECORD
new file mode 100644
index 00000000..faa4cc5a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/RECORD
@@ -0,0 +1,74 @@
+../../../bin/mako-render,sha256=gmpgXbPuhNNaU9wwucYMxdowRklJUVhrtBm4iSZ9CAI,270
+Mako-1.3.9.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+Mako-1.3.9.dist-info/LICENSE,sha256=aNcGTlPj_6jp9CCp5gS9LiBZ2cMwSS-m69TrPUgRFok,1098
+Mako-1.3.9.dist-info/METADATA,sha256=gImMYAE3i8pGSGz0W39Ch2SiUkYphuu5CDwgs2cviNA,2896
+Mako-1.3.9.dist-info/RECORD,,
+Mako-1.3.9.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+Mako-1.3.9.dist-info/entry_points.txt,sha256=LsKkUsOsJQYbJ2M72hZCm968wi5K8Ywb5uFxCuN8Obk,512
+Mako-1.3.9.dist-info/top_level.txt,sha256=LItdH8cDPetpUu8rUyBG3DObS6h9Gcpr9j_WLj2S-R0,5
+mako/__init__.py,sha256=0Umu9OqA4gU1dLRLTcAPieeVTbYoSlzmUeZk9FMt0N8,242
+mako/__pycache__/__init__.cpython-312.pyc,,
+mako/__pycache__/_ast_util.cpython-312.pyc,,
+mako/__pycache__/ast.cpython-312.pyc,,
+mako/__pycache__/cache.cpython-312.pyc,,
+mako/__pycache__/cmd.cpython-312.pyc,,
+mako/__pycache__/codegen.cpython-312.pyc,,
+mako/__pycache__/compat.cpython-312.pyc,,
+mako/__pycache__/exceptions.cpython-312.pyc,,
+mako/__pycache__/filters.cpython-312.pyc,,
+mako/__pycache__/lexer.cpython-312.pyc,,
+mako/__pycache__/lookup.cpython-312.pyc,,
+mako/__pycache__/parsetree.cpython-312.pyc,,
+mako/__pycache__/pygen.cpython-312.pyc,,
+mako/__pycache__/pyparser.cpython-312.pyc,,
+mako/__pycache__/runtime.cpython-312.pyc,,
+mako/__pycache__/template.cpython-312.pyc,,
+mako/__pycache__/util.cpython-312.pyc,,
+mako/_ast_util.py,sha256=hCbfnnizWEa3xRCA-uVyShC2HohSpmVvexz5as_lHc8,20247
+mako/ast.py,sha256=xYrdSiJFbf1CxJ9tU9pcPEWK0BYfwF2aDNDNLQG9PqQ,6642
+mako/cache.py,sha256=kA6FKGl5NeTBnSTcnhoPkSaeJ0JeYpF6GM8qzEZGSts,7680
+mako/cmd.py,sha256=Y2Y6VxNCYwO2Y8EXOdLTf5tpYgfdoondV6Ehlbh8a84,2813
+mako/codegen.py,sha256=N49EH57CcTUfKGzI8V7GFBrAEoFhRo9lkdXBzpFYzBY,47736
+mako/compat.py,sha256=owGbObdoF0C-6rCCs6Vnk3YGHv0bf0PpTja55Woxqb4,1820
+mako/exceptions.py,sha256=87Djuoi1IS5zyVFml1Z5zpCP1IoI7UMOH3h4ejt3x3g,12530
+mako/ext/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+mako/ext/__pycache__/__init__.cpython-312.pyc,,
+mako/ext/__pycache__/autohandler.cpython-312.pyc,,
+mako/ext/__pycache__/babelplugin.cpython-312.pyc,,
+mako/ext/__pycache__/beaker_cache.cpython-312.pyc,,
+mako/ext/__pycache__/extract.cpython-312.pyc,,
+mako/ext/__pycache__/linguaplugin.cpython-312.pyc,,
+mako/ext/__pycache__/preprocessors.cpython-312.pyc,,
+mako/ext/__pycache__/pygmentplugin.cpython-312.pyc,,
+mako/ext/__pycache__/turbogears.cpython-312.pyc,,
+mako/ext/autohandler.py,sha256=QuhNiPSF1LZ53awQ1Qfpwu1Zi9RGOmrtCDnwCLgQzeE,1885
+mako/ext/babelplugin.py,sha256=mmt5fG3pcYII1QsrLEV3wH1ltj-C7uKl8vBIpAjxsKY,2091
+mako/ext/beaker_cache.py,sha256=4quuJQuXRKKUyF6kM43LQhJT1J2z1KSglRZVDW-7c1I,2578
+mako/ext/extract.py,sha256=ZUeaRL2jWcUlrpnhXFXJB0CUJfvQVGBF9tJr5nKJAWI,4659
+mako/ext/linguaplugin.py,sha256=sUZalJSI_XeON9aRBb2hds-ilVQaxHKlfCg_nAl7EuU,1935
+mako/ext/preprocessors.py,sha256=HYG45idRJUwJkDpswEELL8lPFLisQzgDhW5EHpTDGkI,576
+mako/ext/pygmentplugin.py,sha256=TnpJDyQeWTTGHZraMDpw8FB3PNzbmXhaZcjyIBudyi0,4753
+mako/ext/turbogears.py,sha256=egv8hradAnnSJwxtmW8uXAsqPUzX8cZZCXclmO_8hMA,2141
+mako/filters.py,sha256=IBXyGOby4eFE3UGvNLhJlzbe1FfwJ2dcEr1-gKO_Ljc,4658
+mako/lexer.py,sha256=GMq8yf0dEn04-xw2EVDEzaOLXMSVVQz9HyUbfwKnZKg,16321
+mako/lookup.py,sha256=4ALORJiL0wIdDvK1okW8rbjq2jL5F_TNASekDFQSULY,12428
+mako/parsetree.py,sha256=-wmyX_mklAoKhc-7Psx0U15EKpNSd8oGRXFmvk2JQWo,19021
+mako/pygen.py,sha256=689_jR0GG_8Am62Dmlm5h59VY6eyZAU3GroodqEDnZ0,10416
+mako/pyparser.py,sha256=3XNxTFQMLJCrRLQdtT5aKLOMU2faRiKfdmadS47vNm8,7478
+mako/runtime.py,sha256=ZsUEN22nX3d3dECQujF69mBKDQS6yVv2nvz_0eTvFGg,27804
+mako/template.py,sha256=Vn5nLoBY-YzSQJKvRPFGb4fiPyZryA5Q-8mWutv0b9A,23563
+mako/testing/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+mako/testing/__pycache__/__init__.cpython-312.pyc,,
+mako/testing/__pycache__/_config.cpython-312.pyc,,
+mako/testing/__pycache__/assertions.cpython-312.pyc,,
+mako/testing/__pycache__/config.cpython-312.pyc,,
+mako/testing/__pycache__/exclusions.cpython-312.pyc,,
+mako/testing/__pycache__/fixtures.cpython-312.pyc,,
+mako/testing/__pycache__/helpers.cpython-312.pyc,,
+mako/testing/_config.py,sha256=k-qpnsnbXUoN-ykMN5BRpg84i1x0p6UsAddKQnrIytU,3566
+mako/testing/assertions.py,sha256=pfbGl84QlW7QWGg3_lo3wP8XnBAVo9AjzNp2ajmn7FA,5161
+mako/testing/config.py,sha256=wmYVZfzGvOK3mJUZpzmgO8-iIgvaCH41Woi4yDpxq6E,323
+mako/testing/exclusions.py,sha256=_t6ADKdatk3f18tOfHV_ZY6u_ZwQsKphZ2MXJVSAOcI,1553
+mako/testing/fixtures.py,sha256=nEp7wTusf7E0n3Q-BHJW2s_t1vx0KB9poadQ1BmIJzE,3044
+mako/testing/helpers.py,sha256=z4HAactwlht4ut1cbvxKt1QLb3yLPk1U7cnh5BwVUlc,1623
+mako/util.py,sha256=SNYeX2_PmajQJIR3-S1Yqxxylz8lwS65rC8YbCdTkUU,10638
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/WHEEL
new file mode 100644
index 00000000..505164bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/entry_points.txt
new file mode 100644
index 00000000..30f31b2b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/entry_points.txt
@@ -0,0 +1,18 @@
+[babel.extractors]
+mako = mako.ext.babelplugin:extract [babel]
+
+[console_scripts]
+mako-render = mako.cmd:cmdline
+
+[lingua.extractors]
+mako = mako.ext.linguaplugin:LinguaMakoExtractor [lingua]
+
+[pygments.lexers]
+css+mako = mako.ext.pygmentplugin:MakoCssLexer
+html+mako = mako.ext.pygmentplugin:MakoHtmlLexer
+js+mako = mako.ext.pygmentplugin:MakoJavascriptLexer
+mako = mako.ext.pygmentplugin:MakoLexer
+xml+mako = mako.ext.pygmentplugin:MakoXmlLexer
+
+[python.templating.engines]
+mako = mako.ext.turbogears:TGPlugin
diff --git a/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/top_level.txt
new file mode 100644
index 00000000..2951cdd4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Mako-1.3.9.dist-info/top_level.txt
@@ -0,0 +1 @@
+mako
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/LICENSE.txt
new file mode 100644
index 00000000..9d227a0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2010 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/METADATA
new file mode 100644
index 00000000..82261f2a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/METADATA
@@ -0,0 +1,92 @@
+Metadata-Version: 2.1
+Name: MarkupSafe
+Version: 3.0.2
+Summary: Safely add untrusted strings to HTML/XML markup.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+License: Copyright 2010 Pallets
+        
+        Redistribution and use in source and binary forms, with or without
+        modification, are permitted provided that the following conditions are
+        met:
+        
+        1.  Redistributions of source code must retain the above copyright
+            notice, this list of conditions and the following disclaimer.
+        
+        2.  Redistributions in binary form must reproduce the above copyright
+            notice, this list of conditions and the following disclaimer in the
+            documentation and/or other materials provided with the distribution.
+        
+        3.  Neither the name of the copyright holder nor the names of its
+            contributors may be used to endorse or promote products derived from
+            this software without specific prior written permission.
+        
+        THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+        "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+        LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+        PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+        HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+        SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+        TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+        PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+        LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+        NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+        SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+        
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Documentation, https://markupsafe.palletsprojects.com/
+Project-URL: Changes, https://markupsafe.palletsprojects.com/changes/
+Project-URL: Source, https://github.com/pallets/markupsafe/
+Project-URL: Chat, https://discord.gg/pallets
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Classifier: Topic :: Text Processing :: Markup :: HTML
+Classifier: Typing :: Typed
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE.txt
+
+# MarkupSafe
+
+MarkupSafe implements a text object that escapes characters so it is
+safe to use in HTML and XML. Characters that have special meanings are
+replaced so that they display as the actual characters. This mitigates
+injection attacks, meaning untrusted user input can safely be displayed
+on a page.
+
+
+## Examples
+
+```pycon
+>>> from markupsafe import Markup, escape
+
+>>> # escape replaces special characters and wraps in Markup
+>>> escape("<script>alert(document.cookie);</script>")
+Markup('&lt;script&gt;alert(document.cookie);&lt;/script&gt;')
+
+>>> # wrap in Markup to mark text "safe" and prevent escaping
+>>> Markup("<strong>Hello</strong>")
+Markup('<strong>hello</strong>')
+
+>>> escape(Markup("<strong>Hello</strong>"))
+Markup('<strong>hello</strong>')
+
+>>> # Markup is a str subclass
+>>> # methods and operators escape their arguments
+>>> template = Markup("Hello <em>{name}</em>")
+>>> template.format(name='"World"')
+Markup('Hello <em>&#34;World&#34;</em>')
+```
+
+## Donate
+
+The Pallets organization develops and supports MarkupSafe and other
+popular packages. In order to grow the community of contributors and
+users, and allow the maintainers to devote more time to the projects,
+[please donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/RECORD
new file mode 100644
index 00000000..8e0da81b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/RECORD
@@ -0,0 +1,15 @@
+MarkupSafe-3.0.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+MarkupSafe-3.0.2.dist-info/LICENSE.txt,sha256=SJqOEQhQntmKN7uYPhHg9-HTHwvY-Zp5yESOf_N9B-o,1475
+MarkupSafe-3.0.2.dist-info/METADATA,sha256=aAwbZhSmXdfFuMM-rEHpeiHRkBOGESyVLJIuwzHP-nw,3975
+MarkupSafe-3.0.2.dist-info/RECORD,,
+MarkupSafe-3.0.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+MarkupSafe-3.0.2.dist-info/WHEEL,sha256=T94HOVPNbYE6jyG6QmiIglWJ01nwJvHIWFgubY8hhjc,109
+MarkupSafe-3.0.2.dist-info/top_level.txt,sha256=qy0Plje5IJuvsCBjejJyhDCjEAdcDLK_2agVcex8Z6U,11
+markupsafe/__init__.py,sha256=sr-U6_27DfaSrj5jnHYxWN-pvhM27sjlDplMDPZKm7k,13214
+markupsafe/__pycache__/__init__.cpython-312.pyc,,
+markupsafe/__pycache__/_native.cpython-312.pyc,,
+markupsafe/_native.py,sha256=hSLs8Jmz5aqayuengJJ3kdT5PwNpBWpKrmQSdipndC8,210
+markupsafe/_speedups.c,sha256=O7XulmTo-epI6n2FtMVOrJXl8EAaIwD2iNYmBI5SEoQ,4149
+markupsafe/_speedups.cpython-312-darwin.so,sha256=LAX7ul6PQsfO_zZxF631eSzN6U7PoGVTR_Wwd2pGav8,50624
+markupsafe/_speedups.pyi,sha256=ENd1bYe7gbBUf2ywyYWOGUpnXOHNJ-cgTNqetlW8h5k,41
+markupsafe/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/WHEEL
new file mode 100644
index 00000000..edd13a08
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.2.0)
+Root-Is-Purelib: false
+Tag: cp312-cp312-macosx_11_0_arm64
+
diff --git a/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/top_level.txt
new file mode 100644
index 00000000..75bf7292
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/MarkupSafe-3.0.2.dist-info/top_level.txt
@@ -0,0 +1 @@
+markupsafe
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libXau.6.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libXau.6.dylib
new file mode 100755
index 00000000..e9e85c9e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libXau.6.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlicommon.1.1.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlicommon.1.1.0.dylib
new file mode 100755
index 00000000..3244552c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlicommon.1.1.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlidec.1.1.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlidec.1.1.0.dylib
new file mode 100755
index 00000000..6ad36684
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libbrotlidec.1.1.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libfreetype.6.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libfreetype.6.dylib
new file mode 100755
index 00000000..07be8747
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libfreetype.6.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libharfbuzz.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libharfbuzz.0.dylib
new file mode 100755
index 00000000..6e38ec25
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libharfbuzz.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libjpeg.62.4.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libjpeg.62.4.0.dylib
new file mode 100755
index 00000000..baa9d6b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libjpeg.62.4.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblcms2.2.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblcms2.2.dylib
new file mode 100755
index 00000000..99319729
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblcms2.2.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblzma.5.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblzma.5.dylib
new file mode 100755
index 00000000..166552cc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/liblzma.5.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libopenjp2.2.5.3.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libopenjp2.2.5.3.dylib
new file mode 100755
index 00000000..d2e758ce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libopenjp2.2.5.3.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libpng16.16.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libpng16.16.dylib
new file mode 100755
index 00000000..f88397c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libpng16.16.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libsharpyuv.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libsharpyuv.0.dylib
new file mode 100755
index 00000000..1cd14824
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libsharpyuv.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libtiff.6.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libtiff.6.dylib
new file mode 100755
index 00000000..4233a227
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libtiff.6.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebp.7.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebp.7.dylib
new file mode 100755
index 00000000..492ce9d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebp.7.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpdemux.2.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpdemux.2.dylib
new file mode 100755
index 00000000..0fc5bfdb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpdemux.2.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpmux.3.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpmux.3.dylib
new file mode 100755
index 00000000..2dda2d9e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libwebpmux.3.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libxcb.1.1.0.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libxcb.1.1.0.dylib
new file mode 100755
index 00000000..e8eb6871
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libxcb.1.1.0.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libz.1.3.1.zlib-ng.dylib b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libz.1.3.1.zlib-ng.dylib
new file mode 100755
index 00000000..cc4f7b4f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/.dylibs/libz.1.3.1.zlib-ng.dylib differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/BdfFontFile.py b/backend/venv/lib/python3.12/site-packages/PIL/BdfFontFile.py
new file mode 100644
index 00000000..bc1416c7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/BdfFontFile.py
@@ -0,0 +1,133 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# bitmap distribution font (bdf) file parser
+#
+# history:
+# 1996-05-16 fl   created (as bdf2pil)
+# 1997-08-25 fl   converted to FontFile driver
+# 2001-05-25 fl   removed bogus __init__ call
+# 2002-11-20 fl   robustification (from Kevin Cazabon, Dmitry Vasiliev)
+# 2003-04-22 fl   more robustification (from Graham Dumpleton)
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1997-2003 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+"""
+Parse X Bitmap Distribution Format (BDF)
+"""
+from __future__ import annotations
+
+from typing import BinaryIO
+
+from . import FontFile, Image
+
+bdf_slant = {
+    "R": "Roman",
+    "I": "Italic",
+    "O": "Oblique",
+    "RI": "Reverse Italic",
+    "RO": "Reverse Oblique",
+    "OT": "Other",
+}
+
+bdf_spacing = {"P": "Proportional", "M": "Monospaced", "C": "Cell"}
+
+
+def bdf_char(
+    f: BinaryIO,
+) -> (
+    tuple[
+        str,
+        int,
+        tuple[tuple[int, int], tuple[int, int, int, int], tuple[int, int, int, int]],
+        Image.Image,
+    ]
+    | None
+):
+    # skip to STARTCHAR
+    while True:
+        s = f.readline()
+        if not s:
+            return None
+        if s[:9] == b"STARTCHAR":
+            break
+    id = s[9:].strip().decode("ascii")
+
+    # load symbol properties
+    props = {}
+    while True:
+        s = f.readline()
+        if not s or s[:6] == b"BITMAP":
+            break
+        i = s.find(b" ")
+        props[s[:i].decode("ascii")] = s[i + 1 : -1].decode("ascii")
+
+    # load bitmap
+    bitmap = bytearray()
+    while True:
+        s = f.readline()
+        if not s or s[:7] == b"ENDCHAR":
+            break
+        bitmap += s[:-1]
+
+    # The word BBX
+    # followed by the width in x (BBw), height in y (BBh),
+    # and x and y displacement (BBxoff0, BByoff0)
+    # of the lower left corner from the origin of the character.
+    width, height, x_disp, y_disp = (int(p) for p in props["BBX"].split())
+
+    # The word DWIDTH
+    # followed by the width in x and y of the character in device pixels.
+    dwx, dwy = (int(p) for p in props["DWIDTH"].split())
+
+    bbox = (
+        (dwx, dwy),
+        (x_disp, -y_disp - height, width + x_disp, -y_disp),
+        (0, 0, width, height),
+    )
+
+    try:
+        im = Image.frombytes("1", (width, height), bitmap, "hex", "1")
+    except ValueError:
+        # deal with zero-width characters
+        im = Image.new("1", (width, height))
+
+    return id, int(props["ENCODING"]), bbox, im
+
+
+class BdfFontFile(FontFile.FontFile):
+    """Font file plugin for the X11 BDF format."""
+
+    def __init__(self, fp: BinaryIO) -> None:
+        super().__init__()
+
+        s = fp.readline()
+        if s[:13] != b"STARTFONT 2.1":
+            msg = "not a valid BDF file"
+            raise SyntaxError(msg)
+
+        props = {}
+        comments = []
+
+        while True:
+            s = fp.readline()
+            if not s or s[:13] == b"ENDPROPERTIES":
+                break
+            i = s.find(b" ")
+            props[s[:i].decode("ascii")] = s[i + 1 : -1].decode("ascii")
+            if s[:i] in [b"COMMENT", b"COPYRIGHT"]:
+                if s.find(b"LogicalFontDescription") < 0:
+                    comments.append(s[i + 1 : -1].decode("ascii"))
+
+        while True:
+            c = bdf_char(fp)
+            if not c:
+                break
+            id, ch, (xy, dst, src), im = c
+            if 0 <= ch < len(self.glyph):
+                self.glyph[ch] = xy, dst, src, im
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/BlpImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/BlpImagePlugin.py
new file mode 100644
index 00000000..c932b3b9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/BlpImagePlugin.py
@@ -0,0 +1,501 @@
+"""
+Blizzard Mipmap Format (.blp)
+Jerome Leclanche <jerome@leclan.ch>
+
+The contents of this file are hereby released in the public domain (CC0)
+Full text of the CC0 license:
+  https://creativecommons.org/publicdomain/zero/1.0/
+
+BLP1 files, used mostly in Warcraft III, are not fully supported.
+All types of BLP2 files used in World of Warcraft are supported.
+
+The BLP file structure consists of a header, up to 16 mipmaps of the
+texture
+
+Texture sizes must be powers of two, though the two dimensions do
+not have to be equal; 512x256 is valid, but 512x200 is not.
+The first mipmap (mipmap #0) is the full size image; each subsequent
+mipmap halves both dimensions. The final mipmap should be 1x1.
+
+BLP files come in many different flavours:
+* JPEG-compressed (type == 0) - only supported for BLP1.
+* RAW images (type == 1, encoding == 1). Each mipmap is stored as an
+  array of 8-bit values, one per pixel, left to right, top to bottom.
+  Each value is an index to the palette.
+* DXT-compressed (type == 1, encoding == 2):
+- DXT1 compression is used if alpha_encoding == 0.
+  - An additional alpha bit is used if alpha_depth == 1.
+  - DXT3 compression is used if alpha_encoding == 1.
+  - DXT5 compression is used if alpha_encoding == 7.
+"""
+
+from __future__ import annotations
+
+import abc
+import os
+import struct
+from enum import IntEnum
+from io import BytesIO
+from typing import IO
+
+from . import Image, ImageFile
+
+
+class Format(IntEnum):
+    JPEG = 0
+
+
+class Encoding(IntEnum):
+    UNCOMPRESSED = 1
+    DXT = 2
+    UNCOMPRESSED_RAW_BGRA = 3
+
+
+class AlphaEncoding(IntEnum):
+    DXT1 = 0
+    DXT3 = 1
+    DXT5 = 7
+
+
+def unpack_565(i: int) -> tuple[int, int, int]:
+    return ((i >> 11) & 0x1F) << 3, ((i >> 5) & 0x3F) << 2, (i & 0x1F) << 3
+
+
+def decode_dxt1(
+    data: bytes, alpha: bool = False
+) -> tuple[bytearray, bytearray, bytearray, bytearray]:
+    """
+    input: one "row" of data (i.e. will produce 4*width pixels)
+    """
+
+    blocks = len(data) // 8  # number of blocks in row
+    ret = (bytearray(), bytearray(), bytearray(), bytearray())
+
+    for block_index in range(blocks):
+        # Decode next 8-byte block.
+        idx = block_index * 8
+        color0, color1, bits = struct.unpack_from("<HHI", data, idx)
+
+        r0, g0, b0 = unpack_565(color0)
+        r1, g1, b1 = unpack_565(color1)
+
+        # Decode this block into 4x4 pixels
+        # Accumulate the results onto our 4 row accumulators
+        for j in range(4):
+            for i in range(4):
+                # get next control op and generate a pixel
+
+                control = bits & 3
+                bits = bits >> 2
+
+                a = 0xFF
+                if control == 0:
+                    r, g, b = r0, g0, b0
+                elif control == 1:
+                    r, g, b = r1, g1, b1
+                elif control == 2:
+                    if color0 > color1:
+                        r = (2 * r0 + r1) // 3
+                        g = (2 * g0 + g1) // 3
+                        b = (2 * b0 + b1) // 3
+                    else:
+                        r = (r0 + r1) // 2
+                        g = (g0 + g1) // 2
+                        b = (b0 + b1) // 2
+                elif control == 3:
+                    if color0 > color1:
+                        r = (2 * r1 + r0) // 3
+                        g = (2 * g1 + g0) // 3
+                        b = (2 * b1 + b0) // 3
+                    else:
+                        r, g, b, a = 0, 0, 0, 0
+
+                if alpha:
+                    ret[j].extend([r, g, b, a])
+                else:
+                    ret[j].extend([r, g, b])
+
+    return ret
+
+
+def decode_dxt3(data: bytes) -> tuple[bytearray, bytearray, bytearray, bytearray]:
+    """
+    input: one "row" of data (i.e. will produce 4*width pixels)
+    """
+
+    blocks = len(data) // 16  # number of blocks in row
+    ret = (bytearray(), bytearray(), bytearray(), bytearray())
+
+    for block_index in range(blocks):
+        idx = block_index * 16
+        block = data[idx : idx + 16]
+        # Decode next 16-byte block.
+        bits = struct.unpack_from("<8B", block)
+        color0, color1 = struct.unpack_from("<HH", block, 8)
+
+        (code,) = struct.unpack_from("<I", block, 12)
+
+        r0, g0, b0 = unpack_565(color0)
+        r1, g1, b1 = unpack_565(color1)
+
+        for j in range(4):
+            high = False  # Do we want the higher bits?
+            for i in range(4):
+                alphacode_index = (4 * j + i) // 2
+                a = bits[alphacode_index]
+                if high:
+                    high = False
+                    a >>= 4
+                else:
+                    high = True
+                    a &= 0xF
+                a *= 17  # We get a value between 0 and 15
+
+                color_code = (code >> 2 * (4 * j + i)) & 0x03
+
+                if color_code == 0:
+                    r, g, b = r0, g0, b0
+                elif color_code == 1:
+                    r, g, b = r1, g1, b1
+                elif color_code == 2:
+                    r = (2 * r0 + r1) // 3
+                    g = (2 * g0 + g1) // 3
+                    b = (2 * b0 + b1) // 3
+                elif color_code == 3:
+                    r = (2 * r1 + r0) // 3
+                    g = (2 * g1 + g0) // 3
+                    b = (2 * b1 + b0) // 3
+
+                ret[j].extend([r, g, b, a])
+
+    return ret
+
+
+def decode_dxt5(data: bytes) -> tuple[bytearray, bytearray, bytearray, bytearray]:
+    """
+    input: one "row" of data (i.e. will produce 4 * width pixels)
+    """
+
+    blocks = len(data) // 16  # number of blocks in row
+    ret = (bytearray(), bytearray(), bytearray(), bytearray())
+
+    for block_index in range(blocks):
+        idx = block_index * 16
+        block = data[idx : idx + 16]
+        # Decode next 16-byte block.
+        a0, a1 = struct.unpack_from("<BB", block)
+
+        bits = struct.unpack_from("<6B", block, 2)
+        alphacode1 = bits[2] | (bits[3] << 8) | (bits[4] << 16) | (bits[5] << 24)
+        alphacode2 = bits[0] | (bits[1] << 8)
+
+        color0, color1 = struct.unpack_from("<HH", block, 8)
+
+        (code,) = struct.unpack_from("<I", block, 12)
+
+        r0, g0, b0 = unpack_565(color0)
+        r1, g1, b1 = unpack_565(color1)
+
+        for j in range(4):
+            for i in range(4):
+                # get next control op and generate a pixel
+                alphacode_index = 3 * (4 * j + i)
+
+                if alphacode_index <= 12:
+                    alphacode = (alphacode2 >> alphacode_index) & 0x07
+                elif alphacode_index == 15:
+                    alphacode = (alphacode2 >> 15) | ((alphacode1 << 1) & 0x06)
+                else:  # alphacode_index >= 18 and alphacode_index <= 45
+                    alphacode = (alphacode1 >> (alphacode_index - 16)) & 0x07
+
+                if alphacode == 0:
+                    a = a0
+                elif alphacode == 1:
+                    a = a1
+                elif a0 > a1:
+                    a = ((8 - alphacode) * a0 + (alphacode - 1) * a1) // 7
+                elif alphacode == 6:
+                    a = 0
+                elif alphacode == 7:
+                    a = 255
+                else:
+                    a = ((6 - alphacode) * a0 + (alphacode - 1) * a1) // 5
+
+                color_code = (code >> 2 * (4 * j + i)) & 0x03
+
+                if color_code == 0:
+                    r, g, b = r0, g0, b0
+                elif color_code == 1:
+                    r, g, b = r1, g1, b1
+                elif color_code == 2:
+                    r = (2 * r0 + r1) // 3
+                    g = (2 * g0 + g1) // 3
+                    b = (2 * b0 + b1) // 3
+                elif color_code == 3:
+                    r = (2 * r1 + r0) // 3
+                    g = (2 * g1 + g0) // 3
+                    b = (2 * b1 + b0) // 3
+
+                ret[j].extend([r, g, b, a])
+
+    return ret
+
+
+class BLPFormatError(NotImplementedError):
+    pass
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] in (b"BLP1", b"BLP2")
+
+
+class BlpImageFile(ImageFile.ImageFile):
+    """
+    Blizzard Mipmap Format
+    """
+
+    format = "BLP"
+    format_description = "Blizzard Mipmap Format"
+
+    def _open(self) -> None:
+        self.magic = self.fp.read(4)
+        if not _accept(self.magic):
+            msg = f"Bad BLP magic {repr(self.magic)}"
+            raise BLPFormatError(msg)
+
+        compression = struct.unpack("<i", self.fp.read(4))[0]
+        if self.magic == b"BLP1":
+            alpha = struct.unpack("<I", self.fp.read(4))[0] != 0
+        else:
+            encoding = struct.unpack("<b", self.fp.read(1))[0]
+            alpha = struct.unpack("<b", self.fp.read(1))[0] != 0
+            alpha_encoding = struct.unpack("<b", self.fp.read(1))[0]
+            self.fp.seek(1, os.SEEK_CUR)  # mips
+
+        self._size = struct.unpack("<II", self.fp.read(8))
+
+        args: tuple[int, int, bool] | tuple[int, int, bool, int]
+        if self.magic == b"BLP1":
+            encoding = struct.unpack("<i", self.fp.read(4))[0]
+            self.fp.seek(4, os.SEEK_CUR)  # subtype
+
+            args = (compression, encoding, alpha)
+            offset = 28
+        else:
+            args = (compression, encoding, alpha, alpha_encoding)
+            offset = 20
+
+        decoder = self.magic.decode()
+
+        self._mode = "RGBA" if alpha else "RGB"
+        self.tile = [ImageFile._Tile(decoder, (0, 0) + self.size, offset, args)]
+
+
+class _BLPBaseDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        try:
+            self._read_header()
+            self._load()
+        except struct.error as e:
+            msg = "Truncated BLP file"
+            raise OSError(msg) from e
+        return -1, 0
+
+    @abc.abstractmethod
+    def _load(self) -> None:
+        pass
+
+    def _read_header(self) -> None:
+        self._offsets = struct.unpack("<16I", self._safe_read(16 * 4))
+        self._lengths = struct.unpack("<16I", self._safe_read(16 * 4))
+
+    def _safe_read(self, length: int) -> bytes:
+        assert self.fd is not None
+        return ImageFile._safe_read(self.fd, length)
+
+    def _read_palette(self) -> list[tuple[int, int, int, int]]:
+        ret = []
+        for i in range(256):
+            try:
+                b, g, r, a = struct.unpack("<4B", self._safe_read(4))
+            except struct.error:
+                break
+            ret.append((b, g, r, a))
+        return ret
+
+    def _read_bgra(
+        self, palette: list[tuple[int, int, int, int]], alpha: bool
+    ) -> bytearray:
+        data = bytearray()
+        _data = BytesIO(self._safe_read(self._lengths[0]))
+        while True:
+            try:
+                (offset,) = struct.unpack("<B", _data.read(1))
+            except struct.error:
+                break
+            b, g, r, a = palette[offset]
+            d: tuple[int, ...] = (r, g, b)
+            if alpha:
+                d += (a,)
+            data.extend(d)
+        return data
+
+
+class BLP1Decoder(_BLPBaseDecoder):
+    def _load(self) -> None:
+        self._compression, self._encoding, alpha = self.args
+
+        if self._compression == Format.JPEG:
+            self._decode_jpeg_stream()
+
+        elif self._compression == 1:
+            if self._encoding in (4, 5):
+                palette = self._read_palette()
+                data = self._read_bgra(palette, alpha)
+                self.set_as_raw(data)
+            else:
+                msg = f"Unsupported BLP encoding {repr(self._encoding)}"
+                raise BLPFormatError(msg)
+        else:
+            msg = f"Unsupported BLP compression {repr(self._encoding)}"
+            raise BLPFormatError(msg)
+
+    def _decode_jpeg_stream(self) -> None:
+        from .JpegImagePlugin import JpegImageFile
+
+        (jpeg_header_size,) = struct.unpack("<I", self._safe_read(4))
+        jpeg_header = self._safe_read(jpeg_header_size)
+        assert self.fd is not None
+        self._safe_read(self._offsets[0] - self.fd.tell())  # What IS this?
+        data = self._safe_read(self._lengths[0])
+        data = jpeg_header + data
+        image = JpegImageFile(BytesIO(data))
+        Image._decompression_bomb_check(image.size)
+        if image.mode == "CMYK":
+            decoder_name, extents, offset, args = image.tile[0]
+            assert isinstance(args, tuple)
+            image.tile = [
+                ImageFile._Tile(decoder_name, extents, offset, (args[0], "CMYK"))
+            ]
+        r, g, b = image.convert("RGB").split()
+        reversed_image = Image.merge("RGB", (b, g, r))
+        self.set_as_raw(reversed_image.tobytes())
+
+
+class BLP2Decoder(_BLPBaseDecoder):
+    def _load(self) -> None:
+        self._compression, self._encoding, alpha, self._alpha_encoding = self.args
+
+        palette = self._read_palette()
+
+        assert self.fd is not None
+        self.fd.seek(self._offsets[0])
+
+        if self._compression == 1:
+            # Uncompressed or DirectX compression
+
+            if self._encoding == Encoding.UNCOMPRESSED:
+                data = self._read_bgra(palette, alpha)
+
+            elif self._encoding == Encoding.DXT:
+                data = bytearray()
+                if self._alpha_encoding == AlphaEncoding.DXT1:
+                    linesize = (self.state.xsize + 3) // 4 * 8
+                    for yb in range((self.state.ysize + 3) // 4):
+                        for d in decode_dxt1(self._safe_read(linesize), alpha):
+                            data += d
+
+                elif self._alpha_encoding == AlphaEncoding.DXT3:
+                    linesize = (self.state.xsize + 3) // 4 * 16
+                    for yb in range((self.state.ysize + 3) // 4):
+                        for d in decode_dxt3(self._safe_read(linesize)):
+                            data += d
+
+                elif self._alpha_encoding == AlphaEncoding.DXT5:
+                    linesize = (self.state.xsize + 3) // 4 * 16
+                    for yb in range((self.state.ysize + 3) // 4):
+                        for d in decode_dxt5(self._safe_read(linesize)):
+                            data += d
+                else:
+                    msg = f"Unsupported alpha encoding {repr(self._alpha_encoding)}"
+                    raise BLPFormatError(msg)
+            else:
+                msg = f"Unknown BLP encoding {repr(self._encoding)}"
+                raise BLPFormatError(msg)
+
+        else:
+            msg = f"Unknown BLP compression {repr(self._compression)}"
+            raise BLPFormatError(msg)
+
+        self.set_as_raw(data)
+
+
+class BLPEncoder(ImageFile.PyEncoder):
+    _pushes_fd = True
+
+    def _write_palette(self) -> bytes:
+        data = b""
+        assert self.im is not None
+        palette = self.im.getpalette("RGBA", "RGBA")
+        for i in range(len(palette) // 4):
+            r, g, b, a = palette[i * 4 : (i + 1) * 4]
+            data += struct.pack("<4B", b, g, r, a)
+        while len(data) < 256 * 4:
+            data += b"\x00" * 4
+        return data
+
+    def encode(self, bufsize: int) -> tuple[int, int, bytes]:
+        palette_data = self._write_palette()
+
+        offset = 20 + 16 * 4 * 2 + len(palette_data)
+        data = struct.pack("<16I", offset, *((0,) * 15))
+
+        assert self.im is not None
+        w, h = self.im.size
+        data += struct.pack("<16I", w * h, *((0,) * 15))
+
+        data += palette_data
+
+        for y in range(h):
+            for x in range(w):
+                data += struct.pack("<B", self.im.getpixel((x, y)))
+
+        return len(data), 0, data
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode != "P":
+        msg = "Unsupported BLP image mode"
+        raise ValueError(msg)
+
+    magic = b"BLP1" if im.encoderinfo.get("blp_version") == "BLP1" else b"BLP2"
+    fp.write(magic)
+
+    assert im.palette is not None
+    fp.write(struct.pack("<i", 1))  # Uncompressed or DirectX compression
+
+    alpha_depth = 1 if im.palette.mode == "RGBA" else 0
+    if magic == b"BLP1":
+        fp.write(struct.pack("<L", alpha_depth))
+    else:
+        fp.write(struct.pack("<b", Encoding.UNCOMPRESSED))
+        fp.write(struct.pack("<b", alpha_depth))
+        fp.write(struct.pack("<b", 0))  # alpha encoding
+        fp.write(struct.pack("<b", 0))  # mips
+    fp.write(struct.pack("<II", *im.size))
+    if magic == b"BLP1":
+        fp.write(struct.pack("<i", 5))
+        fp.write(struct.pack("<i", 0))
+
+    ImageFile._save(im, fp, [ImageFile._Tile("BLP", (0, 0) + im.size, 0, im.mode)])
+
+
+Image.register_open(BlpImageFile.format, BlpImageFile, _accept)
+Image.register_extension(BlpImageFile.format, ".blp")
+Image.register_decoder("BLP1", BLP1Decoder)
+Image.register_decoder("BLP2", BLP2Decoder)
+
+Image.register_save(BlpImageFile.format, _save)
+Image.register_encoder("BLP", BLPEncoder)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/BmpImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/BmpImagePlugin.py
new file mode 100644
index 00000000..bf8f2957
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/BmpImagePlugin.py
@@ -0,0 +1,511 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# BMP file handler
+#
+# Windows (and OS/2) native bitmap storage format.
+#
+# history:
+# 1995-09-01 fl   Created
+# 1996-04-30 fl   Added save
+# 1997-08-27 fl   Fixed save of 1-bit images
+# 1998-03-06 fl   Load P images as L where possible
+# 1998-07-03 fl   Load P images as 1 where possible
+# 1998-12-29 fl   Handle small palettes
+# 2002-12-30 fl   Fixed load of 1-bit palette images
+# 2003-04-21 fl   Fixed load of 1-bit monochrome images
+# 2003-04-23 fl   Added limited support for BI_BITFIELDS compression
+#
+# Copyright (c) 1997-2003 by Secret Labs AB
+# Copyright (c) 1995-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+from typing import IO, Any
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i16le as i16
+from ._binary import i32le as i32
+from ._binary import o8
+from ._binary import o16le as o16
+from ._binary import o32le as o32
+
+#
+# --------------------------------------------------------------------
+# Read BMP file
+
+BIT2MODE = {
+    # bits => mode, rawmode
+    1: ("P", "P;1"),
+    4: ("P", "P;4"),
+    8: ("P", "P"),
+    16: ("RGB", "BGR;15"),
+    24: ("RGB", "BGR"),
+    32: ("RGB", "BGRX"),
+}
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:2] == b"BM"
+
+
+def _dib_accept(prefix: bytes) -> bool:
+    return i32(prefix) in [12, 40, 52, 56, 64, 108, 124]
+
+
+# =============================================================================
+# Image plugin for the Windows BMP format.
+# =============================================================================
+class BmpImageFile(ImageFile.ImageFile):
+    """Image plugin for the Windows Bitmap format (BMP)"""
+
+    # ------------------------------------------------------------- Description
+    format_description = "Windows Bitmap"
+    format = "BMP"
+
+    # -------------------------------------------------- BMP Compression values
+    COMPRESSIONS = {"RAW": 0, "RLE8": 1, "RLE4": 2, "BITFIELDS": 3, "JPEG": 4, "PNG": 5}
+    for k, v in COMPRESSIONS.items():
+        vars()[k] = v
+
+    def _bitmap(self, header: int = 0, offset: int = 0) -> None:
+        """Read relevant info about the BMP"""
+        read, seek = self.fp.read, self.fp.seek
+        if header:
+            seek(header)
+        # read bmp header size @offset 14 (this is part of the header size)
+        file_info: dict[str, bool | int | tuple[int, ...]] = {
+            "header_size": i32(read(4)),
+            "direction": -1,
+        }
+
+        # -------------------- If requested, read header at a specific position
+        # read the rest of the bmp header, without its size
+        assert isinstance(file_info["header_size"], int)
+        header_data = ImageFile._safe_read(self.fp, file_info["header_size"] - 4)
+
+        # ------------------------------- Windows Bitmap v2, IBM OS/2 Bitmap v1
+        # ----- This format has different offsets because of width/height types
+        # 12: BITMAPCOREHEADER/OS21XBITMAPHEADER
+        if file_info["header_size"] == 12:
+            file_info["width"] = i16(header_data, 0)
+            file_info["height"] = i16(header_data, 2)
+            file_info["planes"] = i16(header_data, 4)
+            file_info["bits"] = i16(header_data, 6)
+            file_info["compression"] = self.COMPRESSIONS["RAW"]
+            file_info["palette_padding"] = 3
+
+        # --------------------------------------------- Windows Bitmap v3 to v5
+        #  40: BITMAPINFOHEADER
+        #  52: BITMAPV2HEADER
+        #  56: BITMAPV3HEADER
+        #  64: BITMAPCOREHEADER2/OS22XBITMAPHEADER
+        # 108: BITMAPV4HEADER
+        # 124: BITMAPV5HEADER
+        elif file_info["header_size"] in (40, 52, 56, 64, 108, 124):
+            file_info["y_flip"] = header_data[7] == 0xFF
+            file_info["direction"] = 1 if file_info["y_flip"] else -1
+            file_info["width"] = i32(header_data, 0)
+            file_info["height"] = (
+                i32(header_data, 4)
+                if not file_info["y_flip"]
+                else 2**32 - i32(header_data, 4)
+            )
+            file_info["planes"] = i16(header_data, 8)
+            file_info["bits"] = i16(header_data, 10)
+            file_info["compression"] = i32(header_data, 12)
+            # byte size of pixel data
+            file_info["data_size"] = i32(header_data, 16)
+            file_info["pixels_per_meter"] = (
+                i32(header_data, 20),
+                i32(header_data, 24),
+            )
+            file_info["colors"] = i32(header_data, 28)
+            file_info["palette_padding"] = 4
+            assert isinstance(file_info["pixels_per_meter"], tuple)
+            self.info["dpi"] = tuple(x / 39.3701 for x in file_info["pixels_per_meter"])
+            if file_info["compression"] == self.COMPRESSIONS["BITFIELDS"]:
+                masks = ["r_mask", "g_mask", "b_mask"]
+                if len(header_data) >= 48:
+                    if len(header_data) >= 52:
+                        masks.append("a_mask")
+                    else:
+                        file_info["a_mask"] = 0x0
+                    for idx, mask in enumerate(masks):
+                        file_info[mask] = i32(header_data, 36 + idx * 4)
+                else:
+                    # 40 byte headers only have the three components in the
+                    # bitfields masks, ref:
+                    # https://msdn.microsoft.com/en-us/library/windows/desktop/dd183376(v=vs.85).aspx
+                    # See also
+                    # https://github.com/python-pillow/Pillow/issues/1293
+                    # There is a 4th component in the RGBQuad, in the alpha
+                    # location, but it is listed as a reserved component,
+                    # and it is not generally an alpha channel
+                    file_info["a_mask"] = 0x0
+                    for mask in masks:
+                        file_info[mask] = i32(read(4))
+                assert isinstance(file_info["r_mask"], int)
+                assert isinstance(file_info["g_mask"], int)
+                assert isinstance(file_info["b_mask"], int)
+                assert isinstance(file_info["a_mask"], int)
+                file_info["rgb_mask"] = (
+                    file_info["r_mask"],
+                    file_info["g_mask"],
+                    file_info["b_mask"],
+                )
+                file_info["rgba_mask"] = (
+                    file_info["r_mask"],
+                    file_info["g_mask"],
+                    file_info["b_mask"],
+                    file_info["a_mask"],
+                )
+        else:
+            msg = f"Unsupported BMP header type ({file_info['header_size']})"
+            raise OSError(msg)
+
+        # ------------------ Special case : header is reported 40, which
+        # ---------------------- is shorter than real size for bpp >= 16
+        assert isinstance(file_info["width"], int)
+        assert isinstance(file_info["height"], int)
+        self._size = file_info["width"], file_info["height"]
+
+        # ------- If color count was not found in the header, compute from bits
+        assert isinstance(file_info["bits"], int)
+        file_info["colors"] = (
+            file_info["colors"]
+            if file_info.get("colors", 0)
+            else (1 << file_info["bits"])
+        )
+        assert isinstance(file_info["colors"], int)
+        if offset == 14 + file_info["header_size"] and file_info["bits"] <= 8:
+            offset += 4 * file_info["colors"]
+
+        # ---------------------- Check bit depth for unusual unsupported values
+        self._mode, raw_mode = BIT2MODE.get(file_info["bits"], ("", ""))
+        if not self.mode:
+            msg = f"Unsupported BMP pixel depth ({file_info['bits']})"
+            raise OSError(msg)
+
+        # ---------------- Process BMP with Bitfields compression (not palette)
+        decoder_name = "raw"
+        if file_info["compression"] == self.COMPRESSIONS["BITFIELDS"]:
+            SUPPORTED: dict[int, list[tuple[int, ...]]] = {
+                32: [
+                    (0xFF0000, 0xFF00, 0xFF, 0x0),
+                    (0xFF000000, 0xFF0000, 0xFF00, 0x0),
+                    (0xFF000000, 0xFF00, 0xFF, 0x0),
+                    (0xFF000000, 0xFF0000, 0xFF00, 0xFF),
+                    (0xFF, 0xFF00, 0xFF0000, 0xFF000000),
+                    (0xFF0000, 0xFF00, 0xFF, 0xFF000000),
+                    (0xFF000000, 0xFF00, 0xFF, 0xFF0000),
+                    (0x0, 0x0, 0x0, 0x0),
+                ],
+                24: [(0xFF0000, 0xFF00, 0xFF)],
+                16: [(0xF800, 0x7E0, 0x1F), (0x7C00, 0x3E0, 0x1F)],
+            }
+            MASK_MODES = {
+                (32, (0xFF0000, 0xFF00, 0xFF, 0x0)): "BGRX",
+                (32, (0xFF000000, 0xFF0000, 0xFF00, 0x0)): "XBGR",
+                (32, (0xFF000000, 0xFF00, 0xFF, 0x0)): "BGXR",
+                (32, (0xFF000000, 0xFF0000, 0xFF00, 0xFF)): "ABGR",
+                (32, (0xFF, 0xFF00, 0xFF0000, 0xFF000000)): "RGBA",
+                (32, (0xFF0000, 0xFF00, 0xFF, 0xFF000000)): "BGRA",
+                (32, (0xFF000000, 0xFF00, 0xFF, 0xFF0000)): "BGAR",
+                (32, (0x0, 0x0, 0x0, 0x0)): "BGRA",
+                (24, (0xFF0000, 0xFF00, 0xFF)): "BGR",
+                (16, (0xF800, 0x7E0, 0x1F)): "BGR;16",
+                (16, (0x7C00, 0x3E0, 0x1F)): "BGR;15",
+            }
+            if file_info["bits"] in SUPPORTED:
+                if (
+                    file_info["bits"] == 32
+                    and file_info["rgba_mask"] in SUPPORTED[file_info["bits"]]
+                ):
+                    assert isinstance(file_info["rgba_mask"], tuple)
+                    raw_mode = MASK_MODES[(file_info["bits"], file_info["rgba_mask"])]
+                    self._mode = "RGBA" if "A" in raw_mode else self.mode
+                elif (
+                    file_info["bits"] in (24, 16)
+                    and file_info["rgb_mask"] in SUPPORTED[file_info["bits"]]
+                ):
+                    assert isinstance(file_info["rgb_mask"], tuple)
+                    raw_mode = MASK_MODES[(file_info["bits"], file_info["rgb_mask"])]
+                else:
+                    msg = "Unsupported BMP bitfields layout"
+                    raise OSError(msg)
+            else:
+                msg = "Unsupported BMP bitfields layout"
+                raise OSError(msg)
+        elif file_info["compression"] == self.COMPRESSIONS["RAW"]:
+            if file_info["bits"] == 32 and header == 22:  # 32-bit .cur offset
+                raw_mode, self._mode = "BGRA", "RGBA"
+        elif file_info["compression"] in (
+            self.COMPRESSIONS["RLE8"],
+            self.COMPRESSIONS["RLE4"],
+        ):
+            decoder_name = "bmp_rle"
+        else:
+            msg = f"Unsupported BMP compression ({file_info['compression']})"
+            raise OSError(msg)
+
+        # --------------- Once the header is processed, process the palette/LUT
+        if self.mode == "P":  # Paletted for 1, 4 and 8 bit images
+            # ---------------------------------------------------- 1-bit images
+            if not (0 < file_info["colors"] <= 65536):
+                msg = f"Unsupported BMP Palette size ({file_info['colors']})"
+                raise OSError(msg)
+            else:
+                assert isinstance(file_info["palette_padding"], int)
+                padding = file_info["palette_padding"]
+                palette = read(padding * file_info["colors"])
+                grayscale = True
+                indices = (
+                    (0, 255)
+                    if file_info["colors"] == 2
+                    else list(range(file_info["colors"]))
+                )
+
+                # ----------------- Check if grayscale and ignore palette if so
+                for ind, val in enumerate(indices):
+                    rgb = palette[ind * padding : ind * padding + 3]
+                    if rgb != o8(val) * 3:
+                        grayscale = False
+
+                # ------- If all colors are gray, white or black, ditch palette
+                if grayscale:
+                    self._mode = "1" if file_info["colors"] == 2 else "L"
+                    raw_mode = self.mode
+                else:
+                    self._mode = "P"
+                    self.palette = ImagePalette.raw(
+                        "BGRX" if padding == 4 else "BGR", palette
+                    )
+
+        # ---------------------------- Finally set the tile data for the plugin
+        self.info["compression"] = file_info["compression"]
+        args: list[Any] = [raw_mode]
+        if decoder_name == "bmp_rle":
+            args.append(file_info["compression"] == self.COMPRESSIONS["RLE4"])
+        else:
+            assert isinstance(file_info["width"], int)
+            args.append(((file_info["width"] * file_info["bits"] + 31) >> 3) & (~3))
+        args.append(file_info["direction"])
+        self.tile = [
+            ImageFile._Tile(
+                decoder_name,
+                (0, 0, file_info["width"], file_info["height"]),
+                offset or self.fp.tell(),
+                tuple(args),
+            )
+        ]
+
+    def _open(self) -> None:
+        """Open file, check magic number and read header"""
+        # read 14 bytes: magic number, filesize, reserved, header final offset
+        head_data = self.fp.read(14)
+        # choke if the file does not have the required magic bytes
+        if not _accept(head_data):
+            msg = "Not a BMP file"
+            raise SyntaxError(msg)
+        # read the start position of the BMP image data (u32)
+        offset = i32(head_data, 10)
+        # load bitmap information (offset=raster info)
+        self._bitmap(offset=offset)
+
+
+class BmpRleDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+        rle4 = self.args[1]
+        data = bytearray()
+        x = 0
+        dest_length = self.state.xsize * self.state.ysize
+        while len(data) < dest_length:
+            pixels = self.fd.read(1)
+            byte = self.fd.read(1)
+            if not pixels or not byte:
+                break
+            num_pixels = pixels[0]
+            if num_pixels:
+                # encoded mode
+                if x + num_pixels > self.state.xsize:
+                    # Too much data for row
+                    num_pixels = max(0, self.state.xsize - x)
+                if rle4:
+                    first_pixel = o8(byte[0] >> 4)
+                    second_pixel = o8(byte[0] & 0x0F)
+                    for index in range(num_pixels):
+                        if index % 2 == 0:
+                            data += first_pixel
+                        else:
+                            data += second_pixel
+                else:
+                    data += byte * num_pixels
+                x += num_pixels
+            else:
+                if byte[0] == 0:
+                    # end of line
+                    while len(data) % self.state.xsize != 0:
+                        data += b"\x00"
+                    x = 0
+                elif byte[0] == 1:
+                    # end of bitmap
+                    break
+                elif byte[0] == 2:
+                    # delta
+                    bytes_read = self.fd.read(2)
+                    if len(bytes_read) < 2:
+                        break
+                    right, up = self.fd.read(2)
+                    data += b"\x00" * (right + up * self.state.xsize)
+                    x = len(data) % self.state.xsize
+                else:
+                    # absolute mode
+                    if rle4:
+                        # 2 pixels per byte
+                        byte_count = byte[0] // 2
+                        bytes_read = self.fd.read(byte_count)
+                        for byte_read in bytes_read:
+                            data += o8(byte_read >> 4)
+                            data += o8(byte_read & 0x0F)
+                    else:
+                        byte_count = byte[0]
+                        bytes_read = self.fd.read(byte_count)
+                        data += bytes_read
+                    if len(bytes_read) < byte_count:
+                        break
+                    x += byte[0]
+
+                    # align to 16-bit word boundary
+                    if self.fd.tell() % 2 != 0:
+                        self.fd.seek(1, os.SEEK_CUR)
+        rawmode = "L" if self.mode == "L" else "P"
+        self.set_as_raw(bytes(data), rawmode, (0, self.args[-1]))
+        return -1, 0
+
+
+# =============================================================================
+# Image plugin for the DIB format (BMP alias)
+# =============================================================================
+class DibImageFile(BmpImageFile):
+    format = "DIB"
+    format_description = "Windows Bitmap"
+
+    def _open(self) -> None:
+        self._bitmap()
+
+
+#
+# --------------------------------------------------------------------
+# Write BMP file
+
+
+SAVE = {
+    "1": ("1", 1, 2),
+    "L": ("L", 8, 256),
+    "P": ("P", 8, 256),
+    "RGB": ("BGR", 24, 0),
+    "RGBA": ("BGRA", 32, 0),
+}
+
+
+def _dib_save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    _save(im, fp, filename, False)
+
+
+def _save(
+    im: Image.Image, fp: IO[bytes], filename: str | bytes, bitmap_header: bool = True
+) -> None:
+    try:
+        rawmode, bits, colors = SAVE[im.mode]
+    except KeyError as e:
+        msg = f"cannot write mode {im.mode} as BMP"
+        raise OSError(msg) from e
+
+    info = im.encoderinfo
+
+    dpi = info.get("dpi", (96, 96))
+
+    # 1 meter == 39.3701 inches
+    ppm = tuple(int(x * 39.3701 + 0.5) for x in dpi)
+
+    stride = ((im.size[0] * bits + 7) // 8 + 3) & (~3)
+    header = 40  # or 64 for OS/2 version 2
+    image = stride * im.size[1]
+
+    if im.mode == "1":
+        palette = b"".join(o8(i) * 4 for i in (0, 255))
+    elif im.mode == "L":
+        palette = b"".join(o8(i) * 4 for i in range(256))
+    elif im.mode == "P":
+        palette = im.im.getpalette("RGB", "BGRX")
+        colors = len(palette) // 4
+    else:
+        palette = None
+
+    # bitmap header
+    if bitmap_header:
+        offset = 14 + header + colors * 4
+        file_size = offset + image
+        if file_size > 2**32 - 1:
+            msg = "File size is too large for the BMP format"
+            raise ValueError(msg)
+        fp.write(
+            b"BM"  # file type (magic)
+            + o32(file_size)  # file size
+            + o32(0)  # reserved
+            + o32(offset)  # image data offset
+        )
+
+    # bitmap info header
+    fp.write(
+        o32(header)  # info header size
+        + o32(im.size[0])  # width
+        + o32(im.size[1])  # height
+        + o16(1)  # planes
+        + o16(bits)  # depth
+        + o32(0)  # compression (0=uncompressed)
+        + o32(image)  # size of bitmap
+        + o32(ppm[0])  # resolution
+        + o32(ppm[1])  # resolution
+        + o32(colors)  # colors used
+        + o32(colors)  # colors important
+    )
+
+    fp.write(b"\0" * (header - 40))  # padding (for OS/2 format)
+
+    if palette:
+        fp.write(palette)
+
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, (rawmode, stride, -1))]
+    )
+
+
+#
+# --------------------------------------------------------------------
+# Registry
+
+
+Image.register_open(BmpImageFile.format, BmpImageFile, _accept)
+Image.register_save(BmpImageFile.format, _save)
+
+Image.register_extension(BmpImageFile.format, ".bmp")
+
+Image.register_mime(BmpImageFile.format, "image/bmp")
+
+Image.register_decoder("bmp_rle", BmpRleDecoder)
+
+Image.register_open(DibImageFile.format, DibImageFile, _dib_accept)
+Image.register_save(DibImageFile.format, _dib_save)
+
+Image.register_extension(DibImageFile.format, ".dib")
+
+Image.register_mime(DibImageFile.format, "image/bmp")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/BufrStubImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/BufrStubImagePlugin.py
new file mode 100644
index 00000000..0ee2f653
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/BufrStubImagePlugin.py
@@ -0,0 +1,76 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# BUFR stub adapter
+#
+# Copyright (c) 1996-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+
+_handler = None
+
+
+def register_handler(handler: ImageFile.StubHandler | None) -> None:
+    """
+    Install application-specific BUFR image handler.
+
+    :param handler: Handler object.
+    """
+    global _handler
+    _handler = handler
+
+
+# --------------------------------------------------------------------
+# Image adapter
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"BUFR" or prefix[:4] == b"ZCZC"
+
+
+class BufrStubImageFile(ImageFile.StubImageFile):
+    format = "BUFR"
+    format_description = "BUFR"
+
+    def _open(self) -> None:
+        offset = self.fp.tell()
+
+        if not _accept(self.fp.read(4)):
+            msg = "Not a BUFR file"
+            raise SyntaxError(msg)
+
+        self.fp.seek(offset)
+
+        # make something up
+        self._mode = "F"
+        self._size = 1, 1
+
+        loader = self._load()
+        if loader:
+            loader.open(self)
+
+    def _load(self) -> ImageFile.StubHandler | None:
+        return _handler
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if _handler is None or not hasattr(_handler, "save"):
+        msg = "BUFR save handler not installed"
+        raise OSError(msg)
+    _handler.save(im, fp, filename)
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(BufrStubImageFile.format, BufrStubImageFile, _accept)
+Image.register_save(BufrStubImageFile.format, _save)
+
+Image.register_extension(BufrStubImageFile.format, ".bufr")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ContainerIO.py b/backend/venv/lib/python3.12/site-packages/PIL/ContainerIO.py
new file mode 100644
index 00000000..ec9e66c7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ContainerIO.py
@@ -0,0 +1,173 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# a class to read from a container file
+#
+# History:
+# 1995-06-18 fl     Created
+# 1995-09-07 fl     Added readline(), readlines()
+#
+# Copyright (c) 1997-2001 by Secret Labs AB
+# Copyright (c) 1995 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+from collections.abc import Iterable
+from typing import IO, AnyStr, NoReturn
+
+
+class ContainerIO(IO[AnyStr]):
+    """
+    A file object that provides read access to a part of an existing
+    file (for example a TAR file).
+    """
+
+    def __init__(self, file: IO[AnyStr], offset: int, length: int) -> None:
+        """
+        Create file object.
+
+        :param file: Existing file.
+        :param offset: Start of region, in bytes.
+        :param length: Size of region, in bytes.
+        """
+        self.fh: IO[AnyStr] = file
+        self.pos = 0
+        self.offset = offset
+        self.length = length
+        self.fh.seek(offset)
+
+    ##
+    # Always false.
+
+    def isatty(self) -> bool:
+        return False
+
+    def seekable(self) -> bool:
+        return True
+
+    def seek(self, offset: int, mode: int = io.SEEK_SET) -> int:
+        """
+        Move file pointer.
+
+        :param offset: Offset in bytes.
+        :param mode: Starting position. Use 0 for beginning of region, 1
+           for current offset, and 2 for end of region.  You cannot move
+           the pointer outside the defined region.
+        :returns: Offset from start of region, in bytes.
+        """
+        if mode == 1:
+            self.pos = self.pos + offset
+        elif mode == 2:
+            self.pos = self.length + offset
+        else:
+            self.pos = offset
+        # clamp
+        self.pos = max(0, min(self.pos, self.length))
+        self.fh.seek(self.offset + self.pos)
+        return self.pos
+
+    def tell(self) -> int:
+        """
+        Get current file pointer.
+
+        :returns: Offset from start of region, in bytes.
+        """
+        return self.pos
+
+    def readable(self) -> bool:
+        return True
+
+    def read(self, n: int = -1) -> AnyStr:
+        """
+        Read data.
+
+        :param n: Number of bytes to read. If omitted, zero or negative,
+            read until end of region.
+        :returns: An 8-bit string.
+        """
+        if n > 0:
+            n = min(n, self.length - self.pos)
+        else:
+            n = self.length - self.pos
+        if n <= 0:  # EOF
+            return b"" if "b" in self.fh.mode else ""  # type: ignore[return-value]
+        self.pos = self.pos + n
+        return self.fh.read(n)
+
+    def readline(self, n: int = -1) -> AnyStr:
+        """
+        Read a line of text.
+
+        :param n: Number of bytes to read. If omitted, zero or negative,
+            read until end of line.
+        :returns: An 8-bit string.
+        """
+        s: AnyStr = b"" if "b" in self.fh.mode else ""  # type: ignore[assignment]
+        newline_character = b"\n" if "b" in self.fh.mode else "\n"
+        while True:
+            c = self.read(1)
+            if not c:
+                break
+            s = s + c
+            if c == newline_character or len(s) == n:
+                break
+        return s
+
+    def readlines(self, n: int | None = -1) -> list[AnyStr]:
+        """
+        Read multiple lines of text.
+
+        :param n: Number of lines to read. If omitted, zero, negative or None,
+            read until end of region.
+        :returns: A list of 8-bit strings.
+        """
+        lines = []
+        while True:
+            s = self.readline()
+            if not s:
+                break
+            lines.append(s)
+            if len(lines) == n:
+                break
+        return lines
+
+    def writable(self) -> bool:
+        return False
+
+    def write(self, b: AnyStr) -> NoReturn:
+        raise NotImplementedError()
+
+    def writelines(self, lines: Iterable[AnyStr]) -> NoReturn:
+        raise NotImplementedError()
+
+    def truncate(self, size: int | None = None) -> int:
+        raise NotImplementedError()
+
+    def __enter__(self) -> ContainerIO[AnyStr]:
+        return self
+
+    def __exit__(self, *args: object) -> None:
+        self.close()
+
+    def __iter__(self) -> ContainerIO[AnyStr]:
+        return self
+
+    def __next__(self) -> AnyStr:
+        line = self.readline()
+        if not line:
+            msg = "end of region"
+            raise StopIteration(msg)
+        return line
+
+    def fileno(self) -> int:
+        return self.fh.fileno()
+
+    def flush(self) -> None:
+        self.fh.flush()
+
+    def close(self) -> None:
+        self.fh.close()
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/CurImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/CurImagePlugin.py
new file mode 100644
index 00000000..c4be0cec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/CurImagePlugin.py
@@ -0,0 +1,75 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Windows Cursor support for PIL
+#
+# notes:
+#       uses BmpImagePlugin.py to read the bitmap data.
+#
+# history:
+#       96-05-27 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import BmpImagePlugin, Image, ImageFile
+from ._binary import i16le as i16
+from ._binary import i32le as i32
+
+#
+# --------------------------------------------------------------------
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"\0\0\2\0"
+
+
+##
+# Image plugin for Windows Cursor files.
+
+
+class CurImageFile(BmpImagePlugin.BmpImageFile):
+    format = "CUR"
+    format_description = "Windows Cursor"
+
+    def _open(self) -> None:
+        offset = self.fp.tell()
+
+        # check magic
+        s = self.fp.read(6)
+        if not _accept(s):
+            msg = "not a CUR file"
+            raise SyntaxError(msg)
+
+        # pick the largest cursor in the file
+        m = b""
+        for i in range(i16(s, 4)):
+            s = self.fp.read(16)
+            if not m:
+                m = s
+            elif s[0] > m[0] and s[1] > m[1]:
+                m = s
+        if not m:
+            msg = "No cursors were found"
+            raise TypeError(msg)
+
+        # load as bitmap
+        self._bitmap(i32(m, 12) + offset)
+
+        # patch up the bitmap height
+        self._size = self.size[0], self.size[1] // 2
+        d, e, o, a = self.tile[0]
+        self.tile[0] = ImageFile._Tile(d, (0, 0) + self.size, o, a)
+
+
+#
+# --------------------------------------------------------------------
+
+Image.register_open(CurImageFile.format, CurImageFile, _accept)
+
+Image.register_extension(CurImageFile.format, ".cur")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/DcxImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/DcxImagePlugin.py
new file mode 100644
index 00000000..f67f27d7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/DcxImagePlugin.py
@@ -0,0 +1,80 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# DCX file handling
+#
+# DCX is a container file format defined by Intel, commonly used
+# for fax applications.  Each DCX file consists of a directory
+# (a list of file offsets) followed by a set of (usually 1-bit)
+# PCX files.
+#
+# History:
+# 1995-09-09 fl   Created
+# 1996-03-20 fl   Properly derived from PcxImageFile.
+# 1998-07-15 fl   Renamed offset attribute to avoid name clash
+# 2002-07-30 fl   Fixed file handling
+#
+# Copyright (c) 1997-98 by Secret Labs AB.
+# Copyright (c) 1995-96 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image
+from ._binary import i32le as i32
+from .PcxImagePlugin import PcxImageFile
+
+MAGIC = 0x3ADE68B1  # QUIZ: what's this value, then?
+
+
+def _accept(prefix: bytes) -> bool:
+    return len(prefix) >= 4 and i32(prefix) == MAGIC
+
+
+##
+# Image plugin for the Intel DCX format.
+
+
+class DcxImageFile(PcxImageFile):
+    format = "DCX"
+    format_description = "Intel DCX"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        # Header
+        s = self.fp.read(4)
+        if not _accept(s):
+            msg = "not a DCX file"
+            raise SyntaxError(msg)
+
+        # Component directory
+        self._offset = []
+        for i in range(1024):
+            offset = i32(self.fp.read(4))
+            if not offset:
+                break
+            self._offset.append(offset)
+
+        self._fp = self.fp
+        self.frame = -1
+        self.n_frames = len(self._offset)
+        self.is_animated = self.n_frames > 1
+        self.seek(0)
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        self.frame = frame
+        self.fp = self._fp
+        self.fp.seek(self._offset[frame])
+        PcxImageFile._open(self)
+
+    def tell(self) -> int:
+        return self.frame
+
+
+Image.register_open(DcxImageFile.format, DcxImageFile, _accept)
+
+Image.register_extension(DcxImageFile.format, ".dcx")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/DdsImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/DdsImagePlugin.py
new file mode 100644
index 00000000..9349e284
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/DdsImagePlugin.py
@@ -0,0 +1,573 @@
+"""
+A Pillow loader for .dds files (S3TC-compressed aka DXTC)
+Jerome Leclanche <jerome@leclan.ch>
+
+Documentation:
+https://web.archive.org/web/20170802060935/http://oss.sgi.com/projects/ogl-sample/registry/EXT/texture_compression_s3tc.txt
+
+The contents of this file are hereby released in the public domain (CC0)
+Full text of the CC0 license:
+https://creativecommons.org/publicdomain/zero/1.0/
+"""
+
+from __future__ import annotations
+
+import io
+import struct
+import sys
+from enum import IntEnum, IntFlag
+from typing import IO
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i32le as i32
+from ._binary import o8
+from ._binary import o32le as o32
+
+# Magic ("DDS ")
+DDS_MAGIC = 0x20534444
+
+
+# DDS flags
+class DDSD(IntFlag):
+    CAPS = 0x1
+    HEIGHT = 0x2
+    WIDTH = 0x4
+    PITCH = 0x8
+    PIXELFORMAT = 0x1000
+    MIPMAPCOUNT = 0x20000
+    LINEARSIZE = 0x80000
+    DEPTH = 0x800000
+
+
+# DDS caps
+class DDSCAPS(IntFlag):
+    COMPLEX = 0x8
+    TEXTURE = 0x1000
+    MIPMAP = 0x400000
+
+
+class DDSCAPS2(IntFlag):
+    CUBEMAP = 0x200
+    CUBEMAP_POSITIVEX = 0x400
+    CUBEMAP_NEGATIVEX = 0x800
+    CUBEMAP_POSITIVEY = 0x1000
+    CUBEMAP_NEGATIVEY = 0x2000
+    CUBEMAP_POSITIVEZ = 0x4000
+    CUBEMAP_NEGATIVEZ = 0x8000
+    VOLUME = 0x200000
+
+
+# Pixel Format
+class DDPF(IntFlag):
+    ALPHAPIXELS = 0x1
+    ALPHA = 0x2
+    FOURCC = 0x4
+    PALETTEINDEXED8 = 0x20
+    RGB = 0x40
+    LUMINANCE = 0x20000
+
+
+# dxgiformat.h
+class DXGI_FORMAT(IntEnum):
+    UNKNOWN = 0
+    R32G32B32A32_TYPELESS = 1
+    R32G32B32A32_FLOAT = 2
+    R32G32B32A32_UINT = 3
+    R32G32B32A32_SINT = 4
+    R32G32B32_TYPELESS = 5
+    R32G32B32_FLOAT = 6
+    R32G32B32_UINT = 7
+    R32G32B32_SINT = 8
+    R16G16B16A16_TYPELESS = 9
+    R16G16B16A16_FLOAT = 10
+    R16G16B16A16_UNORM = 11
+    R16G16B16A16_UINT = 12
+    R16G16B16A16_SNORM = 13
+    R16G16B16A16_SINT = 14
+    R32G32_TYPELESS = 15
+    R32G32_FLOAT = 16
+    R32G32_UINT = 17
+    R32G32_SINT = 18
+    R32G8X24_TYPELESS = 19
+    D32_FLOAT_S8X24_UINT = 20
+    R32_FLOAT_X8X24_TYPELESS = 21
+    X32_TYPELESS_G8X24_UINT = 22
+    R10G10B10A2_TYPELESS = 23
+    R10G10B10A2_UNORM = 24
+    R10G10B10A2_UINT = 25
+    R11G11B10_FLOAT = 26
+    R8G8B8A8_TYPELESS = 27
+    R8G8B8A8_UNORM = 28
+    R8G8B8A8_UNORM_SRGB = 29
+    R8G8B8A8_UINT = 30
+    R8G8B8A8_SNORM = 31
+    R8G8B8A8_SINT = 32
+    R16G16_TYPELESS = 33
+    R16G16_FLOAT = 34
+    R16G16_UNORM = 35
+    R16G16_UINT = 36
+    R16G16_SNORM = 37
+    R16G16_SINT = 38
+    R32_TYPELESS = 39
+    D32_FLOAT = 40
+    R32_FLOAT = 41
+    R32_UINT = 42
+    R32_SINT = 43
+    R24G8_TYPELESS = 44
+    D24_UNORM_S8_UINT = 45
+    R24_UNORM_X8_TYPELESS = 46
+    X24_TYPELESS_G8_UINT = 47
+    R8G8_TYPELESS = 48
+    R8G8_UNORM = 49
+    R8G8_UINT = 50
+    R8G8_SNORM = 51
+    R8G8_SINT = 52
+    R16_TYPELESS = 53
+    R16_FLOAT = 54
+    D16_UNORM = 55
+    R16_UNORM = 56
+    R16_UINT = 57
+    R16_SNORM = 58
+    R16_SINT = 59
+    R8_TYPELESS = 60
+    R8_UNORM = 61
+    R8_UINT = 62
+    R8_SNORM = 63
+    R8_SINT = 64
+    A8_UNORM = 65
+    R1_UNORM = 66
+    R9G9B9E5_SHAREDEXP = 67
+    R8G8_B8G8_UNORM = 68
+    G8R8_G8B8_UNORM = 69
+    BC1_TYPELESS = 70
+    BC1_UNORM = 71
+    BC1_UNORM_SRGB = 72
+    BC2_TYPELESS = 73
+    BC2_UNORM = 74
+    BC2_UNORM_SRGB = 75
+    BC3_TYPELESS = 76
+    BC3_UNORM = 77
+    BC3_UNORM_SRGB = 78
+    BC4_TYPELESS = 79
+    BC4_UNORM = 80
+    BC4_SNORM = 81
+    BC5_TYPELESS = 82
+    BC5_UNORM = 83
+    BC5_SNORM = 84
+    B5G6R5_UNORM = 85
+    B5G5R5A1_UNORM = 86
+    B8G8R8A8_UNORM = 87
+    B8G8R8X8_UNORM = 88
+    R10G10B10_XR_BIAS_A2_UNORM = 89
+    B8G8R8A8_TYPELESS = 90
+    B8G8R8A8_UNORM_SRGB = 91
+    B8G8R8X8_TYPELESS = 92
+    B8G8R8X8_UNORM_SRGB = 93
+    BC6H_TYPELESS = 94
+    BC6H_UF16 = 95
+    BC6H_SF16 = 96
+    BC7_TYPELESS = 97
+    BC7_UNORM = 98
+    BC7_UNORM_SRGB = 99
+    AYUV = 100
+    Y410 = 101
+    Y416 = 102
+    NV12 = 103
+    P010 = 104
+    P016 = 105
+    OPAQUE_420 = 106
+    YUY2 = 107
+    Y210 = 108
+    Y216 = 109
+    NV11 = 110
+    AI44 = 111
+    IA44 = 112
+    P8 = 113
+    A8P8 = 114
+    B4G4R4A4_UNORM = 115
+    P208 = 130
+    V208 = 131
+    V408 = 132
+    SAMPLER_FEEDBACK_MIN_MIP_OPAQUE = 189
+    SAMPLER_FEEDBACK_MIP_REGION_USED_OPAQUE = 190
+
+
+class D3DFMT(IntEnum):
+    UNKNOWN = 0
+    R8G8B8 = 20
+    A8R8G8B8 = 21
+    X8R8G8B8 = 22
+    R5G6B5 = 23
+    X1R5G5B5 = 24
+    A1R5G5B5 = 25
+    A4R4G4B4 = 26
+    R3G3B2 = 27
+    A8 = 28
+    A8R3G3B2 = 29
+    X4R4G4B4 = 30
+    A2B10G10R10 = 31
+    A8B8G8R8 = 32
+    X8B8G8R8 = 33
+    G16R16 = 34
+    A2R10G10B10 = 35
+    A16B16G16R16 = 36
+    A8P8 = 40
+    P8 = 41
+    L8 = 50
+    A8L8 = 51
+    A4L4 = 52
+    V8U8 = 60
+    L6V5U5 = 61
+    X8L8V8U8 = 62
+    Q8W8V8U8 = 63
+    V16U16 = 64
+    A2W10V10U10 = 67
+    D16_LOCKABLE = 70
+    D32 = 71
+    D15S1 = 73
+    D24S8 = 75
+    D24X8 = 77
+    D24X4S4 = 79
+    D16 = 80
+    D32F_LOCKABLE = 82
+    D24FS8 = 83
+    D32_LOCKABLE = 84
+    S8_LOCKABLE = 85
+    L16 = 81
+    VERTEXDATA = 100
+    INDEX16 = 101
+    INDEX32 = 102
+    Q16W16V16U16 = 110
+    R16F = 111
+    G16R16F = 112
+    A16B16G16R16F = 113
+    R32F = 114
+    G32R32F = 115
+    A32B32G32R32F = 116
+    CxV8U8 = 117
+    A1 = 118
+    A2B10G10R10_XR_BIAS = 119
+    BINARYBUFFER = 199
+
+    UYVY = i32(b"UYVY")
+    R8G8_B8G8 = i32(b"RGBG")
+    YUY2 = i32(b"YUY2")
+    G8R8_G8B8 = i32(b"GRGB")
+    DXT1 = i32(b"DXT1")
+    DXT2 = i32(b"DXT2")
+    DXT3 = i32(b"DXT3")
+    DXT4 = i32(b"DXT4")
+    DXT5 = i32(b"DXT5")
+    DX10 = i32(b"DX10")
+    BC4S = i32(b"BC4S")
+    BC4U = i32(b"BC4U")
+    BC5S = i32(b"BC5S")
+    BC5U = i32(b"BC5U")
+    ATI1 = i32(b"ATI1")
+    ATI2 = i32(b"ATI2")
+    MULTI2_ARGB8 = i32(b"MET1")
+
+
+# Backward compatibility layer
+module = sys.modules[__name__]
+for item in DDSD:
+    assert item.name is not None
+    setattr(module, f"DDSD_{item.name}", item.value)
+for item1 in DDSCAPS:
+    assert item1.name is not None
+    setattr(module, f"DDSCAPS_{item1.name}", item1.value)
+for item2 in DDSCAPS2:
+    assert item2.name is not None
+    setattr(module, f"DDSCAPS2_{item2.name}", item2.value)
+for item3 in DDPF:
+    assert item3.name is not None
+    setattr(module, f"DDPF_{item3.name}", item3.value)
+
+DDS_FOURCC = DDPF.FOURCC
+DDS_RGB = DDPF.RGB
+DDS_RGBA = DDPF.RGB | DDPF.ALPHAPIXELS
+DDS_LUMINANCE = DDPF.LUMINANCE
+DDS_LUMINANCEA = DDPF.LUMINANCE | DDPF.ALPHAPIXELS
+DDS_ALPHA = DDPF.ALPHA
+DDS_PAL8 = DDPF.PALETTEINDEXED8
+
+DDS_HEADER_FLAGS_TEXTURE = DDSD.CAPS | DDSD.HEIGHT | DDSD.WIDTH | DDSD.PIXELFORMAT
+DDS_HEADER_FLAGS_MIPMAP = DDSD.MIPMAPCOUNT
+DDS_HEADER_FLAGS_VOLUME = DDSD.DEPTH
+DDS_HEADER_FLAGS_PITCH = DDSD.PITCH
+DDS_HEADER_FLAGS_LINEARSIZE = DDSD.LINEARSIZE
+
+DDS_HEIGHT = DDSD.HEIGHT
+DDS_WIDTH = DDSD.WIDTH
+
+DDS_SURFACE_FLAGS_TEXTURE = DDSCAPS.TEXTURE
+DDS_SURFACE_FLAGS_MIPMAP = DDSCAPS.COMPLEX | DDSCAPS.MIPMAP
+DDS_SURFACE_FLAGS_CUBEMAP = DDSCAPS.COMPLEX
+
+DDS_CUBEMAP_POSITIVEX = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_POSITIVEX
+DDS_CUBEMAP_NEGATIVEX = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_NEGATIVEX
+DDS_CUBEMAP_POSITIVEY = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_POSITIVEY
+DDS_CUBEMAP_NEGATIVEY = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_NEGATIVEY
+DDS_CUBEMAP_POSITIVEZ = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_POSITIVEZ
+DDS_CUBEMAP_NEGATIVEZ = DDSCAPS2.CUBEMAP | DDSCAPS2.CUBEMAP_NEGATIVEZ
+
+DXT1_FOURCC = D3DFMT.DXT1
+DXT3_FOURCC = D3DFMT.DXT3
+DXT5_FOURCC = D3DFMT.DXT5
+
+DXGI_FORMAT_R8G8B8A8_TYPELESS = DXGI_FORMAT.R8G8B8A8_TYPELESS
+DXGI_FORMAT_R8G8B8A8_UNORM = DXGI_FORMAT.R8G8B8A8_UNORM
+DXGI_FORMAT_R8G8B8A8_UNORM_SRGB = DXGI_FORMAT.R8G8B8A8_UNORM_SRGB
+DXGI_FORMAT_BC5_TYPELESS = DXGI_FORMAT.BC5_TYPELESS
+DXGI_FORMAT_BC5_UNORM = DXGI_FORMAT.BC5_UNORM
+DXGI_FORMAT_BC5_SNORM = DXGI_FORMAT.BC5_SNORM
+DXGI_FORMAT_BC6H_UF16 = DXGI_FORMAT.BC6H_UF16
+DXGI_FORMAT_BC6H_SF16 = DXGI_FORMAT.BC6H_SF16
+DXGI_FORMAT_BC7_TYPELESS = DXGI_FORMAT.BC7_TYPELESS
+DXGI_FORMAT_BC7_UNORM = DXGI_FORMAT.BC7_UNORM
+DXGI_FORMAT_BC7_UNORM_SRGB = DXGI_FORMAT.BC7_UNORM_SRGB
+
+
+class DdsImageFile(ImageFile.ImageFile):
+    format = "DDS"
+    format_description = "DirectDraw Surface"
+
+    def _open(self) -> None:
+        if not _accept(self.fp.read(4)):
+            msg = "not a DDS file"
+            raise SyntaxError(msg)
+        (header_size,) = struct.unpack("<I", self.fp.read(4))
+        if header_size != 124:
+            msg = f"Unsupported header size {repr(header_size)}"
+            raise OSError(msg)
+        header_bytes = self.fp.read(header_size - 4)
+        if len(header_bytes) != 120:
+            msg = f"Incomplete header: {len(header_bytes)} bytes"
+            raise OSError(msg)
+        header = io.BytesIO(header_bytes)
+
+        flags, height, width = struct.unpack("<3I", header.read(12))
+        self._size = (width, height)
+        extents = (0, 0) + self.size
+
+        pitch, depth, mipmaps = struct.unpack("<3I", header.read(12))
+        struct.unpack("<11I", header.read(44))  # reserved
+
+        # pixel format
+        pfsize, pfflags, fourcc, bitcount = struct.unpack("<4I", header.read(16))
+        n = 0
+        rawmode = None
+        if pfflags & DDPF.RGB:
+            # Texture contains uncompressed RGB data
+            if pfflags & DDPF.ALPHAPIXELS:
+                self._mode = "RGBA"
+                mask_count = 4
+            else:
+                self._mode = "RGB"
+                mask_count = 3
+
+            masks = struct.unpack(f"<{mask_count}I", header.read(mask_count * 4))
+            self.tile = [ImageFile._Tile("dds_rgb", extents, 0, (bitcount, masks))]
+            return
+        elif pfflags & DDPF.LUMINANCE:
+            if bitcount == 8:
+                self._mode = "L"
+            elif bitcount == 16 and pfflags & DDPF.ALPHAPIXELS:
+                self._mode = "LA"
+            else:
+                msg = f"Unsupported bitcount {bitcount} for {pfflags}"
+                raise OSError(msg)
+        elif pfflags & DDPF.PALETTEINDEXED8:
+            self._mode = "P"
+            self.palette = ImagePalette.raw("RGBA", self.fp.read(1024))
+            self.palette.mode = "RGBA"
+        elif pfflags & DDPF.FOURCC:
+            offset = header_size + 4
+            if fourcc == D3DFMT.DXT1:
+                self._mode = "RGBA"
+                self.pixel_format = "DXT1"
+                n = 1
+            elif fourcc == D3DFMT.DXT3:
+                self._mode = "RGBA"
+                self.pixel_format = "DXT3"
+                n = 2
+            elif fourcc == D3DFMT.DXT5:
+                self._mode = "RGBA"
+                self.pixel_format = "DXT5"
+                n = 3
+            elif fourcc in (D3DFMT.BC4U, D3DFMT.ATI1):
+                self._mode = "L"
+                self.pixel_format = "BC4"
+                n = 4
+            elif fourcc == D3DFMT.BC5S:
+                self._mode = "RGB"
+                self.pixel_format = "BC5S"
+                n = 5
+            elif fourcc in (D3DFMT.BC5U, D3DFMT.ATI2):
+                self._mode = "RGB"
+                self.pixel_format = "BC5"
+                n = 5
+            elif fourcc == D3DFMT.DX10:
+                offset += 20
+                # ignoring flags which pertain to volume textures and cubemaps
+                (dxgi_format,) = struct.unpack("<I", self.fp.read(4))
+                self.fp.read(16)
+                if dxgi_format in (
+                    DXGI_FORMAT.BC1_UNORM,
+                    DXGI_FORMAT.BC1_TYPELESS,
+                ):
+                    self._mode = "RGBA"
+                    self.pixel_format = "BC1"
+                    n = 1
+                elif dxgi_format in (DXGI_FORMAT.BC4_TYPELESS, DXGI_FORMAT.BC4_UNORM):
+                    self._mode = "L"
+                    self.pixel_format = "BC4"
+                    n = 4
+                elif dxgi_format in (DXGI_FORMAT.BC5_TYPELESS, DXGI_FORMAT.BC5_UNORM):
+                    self._mode = "RGB"
+                    self.pixel_format = "BC5"
+                    n = 5
+                elif dxgi_format == DXGI_FORMAT.BC5_SNORM:
+                    self._mode = "RGB"
+                    self.pixel_format = "BC5S"
+                    n = 5
+                elif dxgi_format == DXGI_FORMAT.BC6H_UF16:
+                    self._mode = "RGB"
+                    self.pixel_format = "BC6H"
+                    n = 6
+                elif dxgi_format == DXGI_FORMAT.BC6H_SF16:
+                    self._mode = "RGB"
+                    self.pixel_format = "BC6HS"
+                    n = 6
+                elif dxgi_format in (
+                    DXGI_FORMAT.BC7_TYPELESS,
+                    DXGI_FORMAT.BC7_UNORM,
+                    DXGI_FORMAT.BC7_UNORM_SRGB,
+                ):
+                    self._mode = "RGBA"
+                    self.pixel_format = "BC7"
+                    n = 7
+                    if dxgi_format == DXGI_FORMAT.BC7_UNORM_SRGB:
+                        self.info["gamma"] = 1 / 2.2
+                elif dxgi_format in (
+                    DXGI_FORMAT.R8G8B8A8_TYPELESS,
+                    DXGI_FORMAT.R8G8B8A8_UNORM,
+                    DXGI_FORMAT.R8G8B8A8_UNORM_SRGB,
+                ):
+                    self._mode = "RGBA"
+                    if dxgi_format == DXGI_FORMAT.R8G8B8A8_UNORM_SRGB:
+                        self.info["gamma"] = 1 / 2.2
+                else:
+                    msg = f"Unimplemented DXGI format {dxgi_format}"
+                    raise NotImplementedError(msg)
+            else:
+                msg = f"Unimplemented pixel format {repr(fourcc)}"
+                raise NotImplementedError(msg)
+        else:
+            msg = f"Unknown pixel format flags {pfflags}"
+            raise NotImplementedError(msg)
+
+        if n:
+            self.tile = [
+                ImageFile._Tile("bcn", extents, offset, (n, self.pixel_format))
+            ]
+        else:
+            self.tile = [ImageFile._Tile("raw", extents, 0, rawmode or self.mode)]
+
+    def load_seek(self, pos: int) -> None:
+        pass
+
+
+class DdsRgbDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+        bitcount, masks = self.args
+
+        # Some masks will be padded with zeros, e.g. R 0b11 G 0b1100
+        # Calculate how many zeros each mask is padded with
+        mask_offsets = []
+        # And the maximum value of each channel without the padding
+        mask_totals = []
+        for mask in masks:
+            offset = 0
+            if mask != 0:
+                while mask >> (offset + 1) << (offset + 1) == mask:
+                    offset += 1
+            mask_offsets.append(offset)
+            mask_totals.append(mask >> offset)
+
+        data = bytearray()
+        bytecount = bitcount // 8
+        dest_length = self.state.xsize * self.state.ysize * len(masks)
+        while len(data) < dest_length:
+            value = int.from_bytes(self.fd.read(bytecount), "little")
+            for i, mask in enumerate(masks):
+                masked_value = value & mask
+                # Remove the zero padding, and scale it to 8 bits
+                data += o8(
+                    int(((masked_value >> mask_offsets[i]) / mask_totals[i]) * 255)
+                )
+        self.set_as_raw(data)
+        return -1, 0
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode not in ("RGB", "RGBA", "L", "LA"):
+        msg = f"cannot write mode {im.mode} as DDS"
+        raise OSError(msg)
+
+    alpha = im.mode[-1] == "A"
+    if im.mode[0] == "L":
+        pixel_flags = DDPF.LUMINANCE
+        rawmode = im.mode
+        if alpha:
+            rgba_mask = [0x000000FF, 0x000000FF, 0x000000FF]
+        else:
+            rgba_mask = [0xFF000000, 0xFF000000, 0xFF000000]
+    else:
+        pixel_flags = DDPF.RGB
+        rawmode = im.mode[::-1]
+        rgba_mask = [0x00FF0000, 0x0000FF00, 0x000000FF]
+
+        if alpha:
+            r, g, b, a = im.split()
+            im = Image.merge("RGBA", (a, r, g, b))
+    if alpha:
+        pixel_flags |= DDPF.ALPHAPIXELS
+    rgba_mask.append(0xFF000000 if alpha else 0)
+
+    flags = DDSD.CAPS | DDSD.HEIGHT | DDSD.WIDTH | DDSD.PITCH | DDSD.PIXELFORMAT
+    bitcount = len(im.getbands()) * 8
+    pitch = (im.width * bitcount + 7) // 8
+
+    fp.write(
+        o32(DDS_MAGIC)
+        + struct.pack(
+            "<7I",
+            124,  # header size
+            flags,  # flags
+            im.height,
+            im.width,
+            pitch,
+            0,  # depth
+            0,  # mipmaps
+        )
+        + struct.pack("11I", *((0,) * 11))  # reserved
+        # pfsize, pfflags, fourcc, bitcount
+        + struct.pack("<4I", 32, pixel_flags, 0, bitcount)
+        + struct.pack("<4I", *rgba_mask)  # dwRGBABitMask
+        + struct.pack("<5I", DDSCAPS.TEXTURE, 0, 0, 0, 0)
+    )
+    ImageFile._save(im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, rawmode)])
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"DDS "
+
+
+Image.register_open(DdsImageFile.format, DdsImageFile, _accept)
+Image.register_decoder("dds_rgb", DdsRgbDecoder)
+Image.register_save(DdsImageFile.format, _save)
+Image.register_extension(DdsImageFile.format, ".dds")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/EpsImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/EpsImagePlugin.py
new file mode 100644
index 00000000..36ba15ec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/EpsImagePlugin.py
@@ -0,0 +1,474 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# EPS file handling
+#
+# History:
+# 1995-09-01 fl   Created (0.1)
+# 1996-05-18 fl   Don't choke on "atend" fields, Ghostscript interface (0.2)
+# 1996-08-22 fl   Don't choke on floating point BoundingBox values
+# 1996-08-23 fl   Handle files from Macintosh (0.3)
+# 2001-02-17 fl   Use 're' instead of 'regex' (Python 2.1) (0.4)
+# 2003-09-07 fl   Check gs.close status (from Federico Di Gregorio) (0.5)
+# 2014-05-07 e    Handling of EPS with binary preview and fixed resolution
+#                 resizing
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1995-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import os
+import re
+import subprocess
+import sys
+import tempfile
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i32le as i32
+
+# --------------------------------------------------------------------
+
+
+split = re.compile(r"^%%([^:]*):[ \t]*(.*)[ \t]*$")
+field = re.compile(r"^%[%!\w]([^:]*)[ \t]*$")
+
+gs_binary: str | bool | None = None
+gs_windows_binary = None
+
+
+def has_ghostscript() -> bool:
+    global gs_binary, gs_windows_binary
+    if gs_binary is None:
+        if sys.platform.startswith("win"):
+            if gs_windows_binary is None:
+                import shutil
+
+                for binary in ("gswin32c", "gswin64c", "gs"):
+                    if shutil.which(binary) is not None:
+                        gs_windows_binary = binary
+                        break
+                else:
+                    gs_windows_binary = False
+            gs_binary = gs_windows_binary
+        else:
+            try:
+                subprocess.check_call(["gs", "--version"], stdout=subprocess.DEVNULL)
+                gs_binary = "gs"
+            except OSError:
+                gs_binary = False
+    return gs_binary is not False
+
+
+def Ghostscript(
+    tile: list[ImageFile._Tile],
+    size: tuple[int, int],
+    fp: IO[bytes],
+    scale: int = 1,
+    transparency: bool = False,
+) -> Image.core.ImagingCore:
+    """Render an image using Ghostscript"""
+    global gs_binary
+    if not has_ghostscript():
+        msg = "Unable to locate Ghostscript on paths"
+        raise OSError(msg)
+    assert isinstance(gs_binary, str)
+
+    # Unpack decoder tile
+    args = tile[0].args
+    assert isinstance(args, tuple)
+    length, bbox = args
+
+    # Hack to support hi-res rendering
+    scale = int(scale) or 1
+    width = size[0] * scale
+    height = size[1] * scale
+    # resolution is dependent on bbox and size
+    res_x = 72.0 * width / (bbox[2] - bbox[0])
+    res_y = 72.0 * height / (bbox[3] - bbox[1])
+
+    out_fd, outfile = tempfile.mkstemp()
+    os.close(out_fd)
+
+    infile_temp = None
+    if hasattr(fp, "name") and os.path.exists(fp.name):
+        infile = fp.name
+    else:
+        in_fd, infile_temp = tempfile.mkstemp()
+        os.close(in_fd)
+        infile = infile_temp
+
+        # Ignore length and offset!
+        # Ghostscript can read it
+        # Copy whole file to read in Ghostscript
+        with open(infile_temp, "wb") as f:
+            # fetch length of fp
+            fp.seek(0, io.SEEK_END)
+            fsize = fp.tell()
+            # ensure start position
+            # go back
+            fp.seek(0)
+            lengthfile = fsize
+            while lengthfile > 0:
+                s = fp.read(min(lengthfile, 100 * 1024))
+                if not s:
+                    break
+                lengthfile -= len(s)
+                f.write(s)
+
+    if transparency:
+        # "RGBA"
+        device = "pngalpha"
+    else:
+        # "pnmraw" automatically chooses between
+        # PBM ("1"), PGM ("L"), and PPM ("RGB").
+        device = "pnmraw"
+
+    # Build Ghostscript command
+    command = [
+        gs_binary,
+        "-q",  # quiet mode
+        f"-g{width:d}x{height:d}",  # set output geometry (pixels)
+        f"-r{res_x:f}x{res_y:f}",  # set input DPI (dots per inch)
+        "-dBATCH",  # exit after processing
+        "-dNOPAUSE",  # don't pause between pages
+        "-dSAFER",  # safe mode
+        f"-sDEVICE={device}",
+        f"-sOutputFile={outfile}",  # output file
+        # adjust for image origin
+        "-c",
+        f"{-bbox[0]} {-bbox[1]} translate",
+        "-f",
+        infile,  # input file
+        # showpage (see https://bugs.ghostscript.com/show_bug.cgi?id=698272)
+        "-c",
+        "showpage",
+    ]
+
+    # push data through Ghostscript
+    try:
+        startupinfo = None
+        if sys.platform.startswith("win"):
+            startupinfo = subprocess.STARTUPINFO()
+            startupinfo.dwFlags |= subprocess.STARTF_USESHOWWINDOW
+        subprocess.check_call(command, startupinfo=startupinfo)
+        with Image.open(outfile) as out_im:
+            out_im.load()
+            return out_im.im.copy()
+    finally:
+        try:
+            os.unlink(outfile)
+            if infile_temp:
+                os.unlink(infile_temp)
+        except OSError:
+            pass
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"%!PS" or (len(prefix) >= 4 and i32(prefix) == 0xC6D3D0C5)
+
+
+##
+# Image plugin for Encapsulated PostScript. This plugin supports only
+# a few variants of this format.
+
+
+class EpsImageFile(ImageFile.ImageFile):
+    """EPS File Parser for the Python Imaging Library"""
+
+    format = "EPS"
+    format_description = "Encapsulated Postscript"
+
+    mode_map = {1: "L", 2: "LAB", 3: "RGB", 4: "CMYK"}
+
+    def _open(self) -> None:
+        (length, offset) = self._find_offset(self.fp)
+
+        # go to offset - start of "%!PS"
+        self.fp.seek(offset)
+
+        self._mode = "RGB"
+
+        # When reading header comments, the first comment is used.
+        # When reading trailer comments, the last comment is used.
+        bounding_box: list[int] | None = None
+        imagedata_size: tuple[int, int] | None = None
+
+        byte_arr = bytearray(255)
+        bytes_mv = memoryview(byte_arr)
+        bytes_read = 0
+        reading_header_comments = True
+        reading_trailer_comments = False
+        trailer_reached = False
+
+        def check_required_header_comments() -> None:
+            """
+            The EPS specification requires that some headers exist.
+            This should be checked when the header comments formally end,
+            when image data starts, or when the file ends, whichever comes first.
+            """
+            if "PS-Adobe" not in self.info:
+                msg = 'EPS header missing "%!PS-Adobe" comment'
+                raise SyntaxError(msg)
+            if "BoundingBox" not in self.info:
+                msg = 'EPS header missing "%%BoundingBox" comment'
+                raise SyntaxError(msg)
+
+        def read_comment(s: str) -> bool:
+            nonlocal bounding_box, reading_trailer_comments
+            try:
+                m = split.match(s)
+            except re.error as e:
+                msg = "not an EPS file"
+                raise SyntaxError(msg) from e
+
+            if not m:
+                return False
+
+            k, v = m.group(1, 2)
+            self.info[k] = v
+            if k == "BoundingBox":
+                if v == "(atend)":
+                    reading_trailer_comments = True
+                elif not bounding_box or (trailer_reached and reading_trailer_comments):
+                    try:
+                        # Note: The DSC spec says that BoundingBox
+                        # fields should be integers, but some drivers
+                        # put floating point values there anyway.
+                        bounding_box = [int(float(i)) for i in v.split()]
+                    except Exception:
+                        pass
+            return True
+
+        while True:
+            byte = self.fp.read(1)
+            if byte == b"":
+                # if we didn't read a byte we must be at the end of the file
+                if bytes_read == 0:
+                    if reading_header_comments:
+                        check_required_header_comments()
+                    break
+            elif byte in b"\r\n":
+                # if we read a line ending character, ignore it and parse what
+                # we have already read. if we haven't read any other characters,
+                # continue reading
+                if bytes_read == 0:
+                    continue
+            else:
+                # ASCII/hexadecimal lines in an EPS file must not exceed
+                # 255 characters, not including line ending characters
+                if bytes_read >= 255:
+                    # only enforce this for lines starting with a "%",
+                    # otherwise assume it's binary data
+                    if byte_arr[0] == ord("%"):
+                        msg = "not an EPS file"
+                        raise SyntaxError(msg)
+                    else:
+                        if reading_header_comments:
+                            check_required_header_comments()
+                            reading_header_comments = False
+                        # reset bytes_read so we can keep reading
+                        # data until the end of the line
+                        bytes_read = 0
+                byte_arr[bytes_read] = byte[0]
+                bytes_read += 1
+                continue
+
+            if reading_header_comments:
+                # Load EPS header
+
+                # if this line doesn't start with a "%",
+                # or does start with "%%EndComments",
+                # then we've reached the end of the header/comments
+                if byte_arr[0] != ord("%") or bytes_mv[:13] == b"%%EndComments":
+                    check_required_header_comments()
+                    reading_header_comments = False
+                    continue
+
+                s = str(bytes_mv[:bytes_read], "latin-1")
+                if not read_comment(s):
+                    m = field.match(s)
+                    if m:
+                        k = m.group(1)
+                        if k[:8] == "PS-Adobe":
+                            self.info["PS-Adobe"] = k[9:]
+                        else:
+                            self.info[k] = ""
+                    elif s[0] == "%":
+                        # handle non-DSC PostScript comments that some
+                        # tools mistakenly put in the Comments section
+                        pass
+                    else:
+                        msg = "bad EPS header"
+                        raise OSError(msg)
+            elif bytes_mv[:11] == b"%ImageData:":
+                # Check for an "ImageData" descriptor
+                # https://www.adobe.com/devnet-apps/photoshop/fileformatashtml/#50577413_pgfId-1035096
+
+                # If we've already read an "ImageData" descriptor,
+                # don't read another one.
+                if imagedata_size:
+                    bytes_read = 0
+                    continue
+
+                # Values:
+                # columns
+                # rows
+                # bit depth (1 or 8)
+                # mode (1: L, 2: LAB, 3: RGB, 4: CMYK)
+                # number of padding channels
+                # block size (number of bytes per row per channel)
+                # binary/ascii (1: binary, 2: ascii)
+                # data start identifier (the image data follows after a single line
+                #   consisting only of this quoted value)
+                image_data_values = byte_arr[11:bytes_read].split(None, 7)
+                columns, rows, bit_depth, mode_id = (
+                    int(value) for value in image_data_values[:4]
+                )
+
+                if bit_depth == 1:
+                    self._mode = "1"
+                elif bit_depth == 8:
+                    try:
+                        self._mode = self.mode_map[mode_id]
+                    except ValueError:
+                        break
+                else:
+                    break
+
+                # Parse the columns and rows after checking the bit depth and mode
+                # in case the bit depth and/or mode are invalid.
+                imagedata_size = columns, rows
+            elif bytes_mv[:5] == b"%%EOF":
+                break
+            elif trailer_reached and reading_trailer_comments:
+                # Load EPS trailer
+                s = str(bytes_mv[:bytes_read], "latin-1")
+                read_comment(s)
+            elif bytes_mv[:9] == b"%%Trailer":
+                trailer_reached = True
+            bytes_read = 0
+
+        # A "BoundingBox" is always required,
+        # even if an "ImageData" descriptor size exists.
+        if not bounding_box:
+            msg = "cannot determine EPS bounding box"
+            raise OSError(msg)
+
+        # An "ImageData" size takes precedence over the "BoundingBox".
+        self._size = imagedata_size or (
+            bounding_box[2] - bounding_box[0],
+            bounding_box[3] - bounding_box[1],
+        )
+
+        self.tile = [
+            ImageFile._Tile("eps", (0, 0) + self.size, offset, (length, bounding_box))
+        ]
+
+    def _find_offset(self, fp: IO[bytes]) -> tuple[int, int]:
+        s = fp.read(4)
+
+        if s == b"%!PS":
+            # for HEAD without binary preview
+            fp.seek(0, io.SEEK_END)
+            length = fp.tell()
+            offset = 0
+        elif i32(s) == 0xC6D3D0C5:
+            # FIX for: Some EPS file not handled correctly / issue #302
+            # EPS can contain binary data
+            # or start directly with latin coding
+            # more info see:
+            # https://web.archive.org/web/20160528181353/http://partners.adobe.com/public/developer/en/ps/5002.EPSF_Spec.pdf
+            s = fp.read(8)
+            offset = i32(s)
+            length = i32(s, 4)
+        else:
+            msg = "not an EPS file"
+            raise SyntaxError(msg)
+
+        return length, offset
+
+    def load(
+        self, scale: int = 1, transparency: bool = False
+    ) -> Image.core.PixelAccess | None:
+        # Load EPS via Ghostscript
+        if self.tile:
+            self.im = Ghostscript(self.tile, self.size, self.fp, scale, transparency)
+            self._mode = self.im.mode
+            self._size = self.im.size
+            self.tile = []
+        return Image.Image.load(self)
+
+    def load_seek(self, pos: int) -> None:
+        # we can't incrementally load, so force ImageFile.parser to
+        # use our custom load method by defining this method.
+        pass
+
+
+# --------------------------------------------------------------------
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes, eps: int = 1) -> None:
+    """EPS Writer for the Python Imaging Library."""
+
+    # make sure image data is available
+    im.load()
+
+    # determine PostScript image mode
+    if im.mode == "L":
+        operator = (8, 1, b"image")
+    elif im.mode == "RGB":
+        operator = (8, 3, b"false 3 colorimage")
+    elif im.mode == "CMYK":
+        operator = (8, 4, b"false 4 colorimage")
+    else:
+        msg = "image mode is not supported"
+        raise ValueError(msg)
+
+    if eps:
+        # write EPS header
+        fp.write(b"%!PS-Adobe-3.0 EPSF-3.0\n")
+        fp.write(b"%%Creator: PIL 0.1 EpsEncode\n")
+        # fp.write("%%CreationDate: %s"...)
+        fp.write(b"%%%%BoundingBox: 0 0 %d %d\n" % im.size)
+        fp.write(b"%%Pages: 1\n")
+        fp.write(b"%%EndComments\n")
+        fp.write(b"%%Page: 1 1\n")
+        fp.write(b"%%ImageData: %d %d " % im.size)
+        fp.write(b'%d %d 0 1 1 "%s"\n' % operator)
+
+    # image header
+    fp.write(b"gsave\n")
+    fp.write(b"10 dict begin\n")
+    fp.write(b"/buf %d string def\n" % (im.size[0] * operator[1]))
+    fp.write(b"%d %d scale\n" % im.size)
+    fp.write(b"%d %d 8\n" % im.size)  # <= bits
+    fp.write(b"[%d 0 0 -%d 0 %d]\n" % (im.size[0], im.size[1], im.size[1]))
+    fp.write(b"{ currentfile buf readhexstring pop } bind\n")
+    fp.write(operator[2] + b"\n")
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+    ImageFile._save(im, fp, [ImageFile._Tile("eps", (0, 0) + im.size)])
+
+    fp.write(b"\n%%%%EndBinary\n")
+    fp.write(b"grestore end\n")
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+# --------------------------------------------------------------------
+
+
+Image.register_open(EpsImageFile.format, EpsImageFile, _accept)
+
+Image.register_save(EpsImageFile.format, _save)
+
+Image.register_extensions(EpsImageFile.format, [".ps", ".eps"])
+
+Image.register_mime(EpsImageFile.format, "application/postscript")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ExifTags.py b/backend/venv/lib/python3.12/site-packages/PIL/ExifTags.py
new file mode 100644
index 00000000..2280d5ce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ExifTags.py
@@ -0,0 +1,382 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# EXIF tags
+#
+# Copyright (c) 2003 by Secret Labs AB
+#
+# See the README file for information on usage and redistribution.
+#
+
+"""
+This module provides constants and clear-text names for various
+well-known EXIF tags.
+"""
+from __future__ import annotations
+
+from enum import IntEnum
+
+
+class Base(IntEnum):
+    # possibly incomplete
+    InteropIndex = 0x0001
+    ProcessingSoftware = 0x000B
+    NewSubfileType = 0x00FE
+    SubfileType = 0x00FF
+    ImageWidth = 0x0100
+    ImageLength = 0x0101
+    BitsPerSample = 0x0102
+    Compression = 0x0103
+    PhotometricInterpretation = 0x0106
+    Thresholding = 0x0107
+    CellWidth = 0x0108
+    CellLength = 0x0109
+    FillOrder = 0x010A
+    DocumentName = 0x010D
+    ImageDescription = 0x010E
+    Make = 0x010F
+    Model = 0x0110
+    StripOffsets = 0x0111
+    Orientation = 0x0112
+    SamplesPerPixel = 0x0115
+    RowsPerStrip = 0x0116
+    StripByteCounts = 0x0117
+    MinSampleValue = 0x0118
+    MaxSampleValue = 0x0119
+    XResolution = 0x011A
+    YResolution = 0x011B
+    PlanarConfiguration = 0x011C
+    PageName = 0x011D
+    FreeOffsets = 0x0120
+    FreeByteCounts = 0x0121
+    GrayResponseUnit = 0x0122
+    GrayResponseCurve = 0x0123
+    T4Options = 0x0124
+    T6Options = 0x0125
+    ResolutionUnit = 0x0128
+    PageNumber = 0x0129
+    TransferFunction = 0x012D
+    Software = 0x0131
+    DateTime = 0x0132
+    Artist = 0x013B
+    HostComputer = 0x013C
+    Predictor = 0x013D
+    WhitePoint = 0x013E
+    PrimaryChromaticities = 0x013F
+    ColorMap = 0x0140
+    HalftoneHints = 0x0141
+    TileWidth = 0x0142
+    TileLength = 0x0143
+    TileOffsets = 0x0144
+    TileByteCounts = 0x0145
+    SubIFDs = 0x014A
+    InkSet = 0x014C
+    InkNames = 0x014D
+    NumberOfInks = 0x014E
+    DotRange = 0x0150
+    TargetPrinter = 0x0151
+    ExtraSamples = 0x0152
+    SampleFormat = 0x0153
+    SMinSampleValue = 0x0154
+    SMaxSampleValue = 0x0155
+    TransferRange = 0x0156
+    ClipPath = 0x0157
+    XClipPathUnits = 0x0158
+    YClipPathUnits = 0x0159
+    Indexed = 0x015A
+    JPEGTables = 0x015B
+    OPIProxy = 0x015F
+    JPEGProc = 0x0200
+    JpegIFOffset = 0x0201
+    JpegIFByteCount = 0x0202
+    JpegRestartInterval = 0x0203
+    JpegLosslessPredictors = 0x0205
+    JpegPointTransforms = 0x0206
+    JpegQTables = 0x0207
+    JpegDCTables = 0x0208
+    JpegACTables = 0x0209
+    YCbCrCoefficients = 0x0211
+    YCbCrSubSampling = 0x0212
+    YCbCrPositioning = 0x0213
+    ReferenceBlackWhite = 0x0214
+    XMLPacket = 0x02BC
+    RelatedImageFileFormat = 0x1000
+    RelatedImageWidth = 0x1001
+    RelatedImageLength = 0x1002
+    Rating = 0x4746
+    RatingPercent = 0x4749
+    ImageID = 0x800D
+    CFARepeatPatternDim = 0x828D
+    BatteryLevel = 0x828F
+    Copyright = 0x8298
+    ExposureTime = 0x829A
+    FNumber = 0x829D
+    IPTCNAA = 0x83BB
+    ImageResources = 0x8649
+    ExifOffset = 0x8769
+    InterColorProfile = 0x8773
+    ExposureProgram = 0x8822
+    SpectralSensitivity = 0x8824
+    GPSInfo = 0x8825
+    ISOSpeedRatings = 0x8827
+    OECF = 0x8828
+    Interlace = 0x8829
+    TimeZoneOffset = 0x882A
+    SelfTimerMode = 0x882B
+    SensitivityType = 0x8830
+    StandardOutputSensitivity = 0x8831
+    RecommendedExposureIndex = 0x8832
+    ISOSpeed = 0x8833
+    ISOSpeedLatitudeyyy = 0x8834
+    ISOSpeedLatitudezzz = 0x8835
+    ExifVersion = 0x9000
+    DateTimeOriginal = 0x9003
+    DateTimeDigitized = 0x9004
+    OffsetTime = 0x9010
+    OffsetTimeOriginal = 0x9011
+    OffsetTimeDigitized = 0x9012
+    ComponentsConfiguration = 0x9101
+    CompressedBitsPerPixel = 0x9102
+    ShutterSpeedValue = 0x9201
+    ApertureValue = 0x9202
+    BrightnessValue = 0x9203
+    ExposureBiasValue = 0x9204
+    MaxApertureValue = 0x9205
+    SubjectDistance = 0x9206
+    MeteringMode = 0x9207
+    LightSource = 0x9208
+    Flash = 0x9209
+    FocalLength = 0x920A
+    Noise = 0x920D
+    ImageNumber = 0x9211
+    SecurityClassification = 0x9212
+    ImageHistory = 0x9213
+    TIFFEPStandardID = 0x9216
+    MakerNote = 0x927C
+    UserComment = 0x9286
+    SubsecTime = 0x9290
+    SubsecTimeOriginal = 0x9291
+    SubsecTimeDigitized = 0x9292
+    AmbientTemperature = 0x9400
+    Humidity = 0x9401
+    Pressure = 0x9402
+    WaterDepth = 0x9403
+    Acceleration = 0x9404
+    CameraElevationAngle = 0x9405
+    XPTitle = 0x9C9B
+    XPComment = 0x9C9C
+    XPAuthor = 0x9C9D
+    XPKeywords = 0x9C9E
+    XPSubject = 0x9C9F
+    FlashPixVersion = 0xA000
+    ColorSpace = 0xA001
+    ExifImageWidth = 0xA002
+    ExifImageHeight = 0xA003
+    RelatedSoundFile = 0xA004
+    ExifInteroperabilityOffset = 0xA005
+    FlashEnergy = 0xA20B
+    SpatialFrequencyResponse = 0xA20C
+    FocalPlaneXResolution = 0xA20E
+    FocalPlaneYResolution = 0xA20F
+    FocalPlaneResolutionUnit = 0xA210
+    SubjectLocation = 0xA214
+    ExposureIndex = 0xA215
+    SensingMethod = 0xA217
+    FileSource = 0xA300
+    SceneType = 0xA301
+    CFAPattern = 0xA302
+    CustomRendered = 0xA401
+    ExposureMode = 0xA402
+    WhiteBalance = 0xA403
+    DigitalZoomRatio = 0xA404
+    FocalLengthIn35mmFilm = 0xA405
+    SceneCaptureType = 0xA406
+    GainControl = 0xA407
+    Contrast = 0xA408
+    Saturation = 0xA409
+    Sharpness = 0xA40A
+    DeviceSettingDescription = 0xA40B
+    SubjectDistanceRange = 0xA40C
+    ImageUniqueID = 0xA420
+    CameraOwnerName = 0xA430
+    BodySerialNumber = 0xA431
+    LensSpecification = 0xA432
+    LensMake = 0xA433
+    LensModel = 0xA434
+    LensSerialNumber = 0xA435
+    CompositeImage = 0xA460
+    CompositeImageCount = 0xA461
+    CompositeImageExposureTimes = 0xA462
+    Gamma = 0xA500
+    PrintImageMatching = 0xC4A5
+    DNGVersion = 0xC612
+    DNGBackwardVersion = 0xC613
+    UniqueCameraModel = 0xC614
+    LocalizedCameraModel = 0xC615
+    CFAPlaneColor = 0xC616
+    CFALayout = 0xC617
+    LinearizationTable = 0xC618
+    BlackLevelRepeatDim = 0xC619
+    BlackLevel = 0xC61A
+    BlackLevelDeltaH = 0xC61B
+    BlackLevelDeltaV = 0xC61C
+    WhiteLevel = 0xC61D
+    DefaultScale = 0xC61E
+    DefaultCropOrigin = 0xC61F
+    DefaultCropSize = 0xC620
+    ColorMatrix1 = 0xC621
+    ColorMatrix2 = 0xC622
+    CameraCalibration1 = 0xC623
+    CameraCalibration2 = 0xC624
+    ReductionMatrix1 = 0xC625
+    ReductionMatrix2 = 0xC626
+    AnalogBalance = 0xC627
+    AsShotNeutral = 0xC628
+    AsShotWhiteXY = 0xC629
+    BaselineExposure = 0xC62A
+    BaselineNoise = 0xC62B
+    BaselineSharpness = 0xC62C
+    BayerGreenSplit = 0xC62D
+    LinearResponseLimit = 0xC62E
+    CameraSerialNumber = 0xC62F
+    LensInfo = 0xC630
+    ChromaBlurRadius = 0xC631
+    AntiAliasStrength = 0xC632
+    ShadowScale = 0xC633
+    DNGPrivateData = 0xC634
+    MakerNoteSafety = 0xC635
+    CalibrationIlluminant1 = 0xC65A
+    CalibrationIlluminant2 = 0xC65B
+    BestQualityScale = 0xC65C
+    RawDataUniqueID = 0xC65D
+    OriginalRawFileName = 0xC68B
+    OriginalRawFileData = 0xC68C
+    ActiveArea = 0xC68D
+    MaskedAreas = 0xC68E
+    AsShotICCProfile = 0xC68F
+    AsShotPreProfileMatrix = 0xC690
+    CurrentICCProfile = 0xC691
+    CurrentPreProfileMatrix = 0xC692
+    ColorimetricReference = 0xC6BF
+    CameraCalibrationSignature = 0xC6F3
+    ProfileCalibrationSignature = 0xC6F4
+    AsShotProfileName = 0xC6F6
+    NoiseReductionApplied = 0xC6F7
+    ProfileName = 0xC6F8
+    ProfileHueSatMapDims = 0xC6F9
+    ProfileHueSatMapData1 = 0xC6FA
+    ProfileHueSatMapData2 = 0xC6FB
+    ProfileToneCurve = 0xC6FC
+    ProfileEmbedPolicy = 0xC6FD
+    ProfileCopyright = 0xC6FE
+    ForwardMatrix1 = 0xC714
+    ForwardMatrix2 = 0xC715
+    PreviewApplicationName = 0xC716
+    PreviewApplicationVersion = 0xC717
+    PreviewSettingsName = 0xC718
+    PreviewSettingsDigest = 0xC719
+    PreviewColorSpace = 0xC71A
+    PreviewDateTime = 0xC71B
+    RawImageDigest = 0xC71C
+    OriginalRawFileDigest = 0xC71D
+    SubTileBlockSize = 0xC71E
+    RowInterleaveFactor = 0xC71F
+    ProfileLookTableDims = 0xC725
+    ProfileLookTableData = 0xC726
+    OpcodeList1 = 0xC740
+    OpcodeList2 = 0xC741
+    OpcodeList3 = 0xC74E
+    NoiseProfile = 0xC761
+
+
+"""Maps EXIF tags to tag names."""
+TAGS = {
+    **{i.value: i.name for i in Base},
+    0x920C: "SpatialFrequencyResponse",
+    0x9214: "SubjectLocation",
+    0x9215: "ExposureIndex",
+    0x828E: "CFAPattern",
+    0x920B: "FlashEnergy",
+    0x9216: "TIFF/EPStandardID",
+}
+
+
+class GPS(IntEnum):
+    GPSVersionID = 0x00
+    GPSLatitudeRef = 0x01
+    GPSLatitude = 0x02
+    GPSLongitudeRef = 0x03
+    GPSLongitude = 0x04
+    GPSAltitudeRef = 0x05
+    GPSAltitude = 0x06
+    GPSTimeStamp = 0x07
+    GPSSatellites = 0x08
+    GPSStatus = 0x09
+    GPSMeasureMode = 0x0A
+    GPSDOP = 0x0B
+    GPSSpeedRef = 0x0C
+    GPSSpeed = 0x0D
+    GPSTrackRef = 0x0E
+    GPSTrack = 0x0F
+    GPSImgDirectionRef = 0x10
+    GPSImgDirection = 0x11
+    GPSMapDatum = 0x12
+    GPSDestLatitudeRef = 0x13
+    GPSDestLatitude = 0x14
+    GPSDestLongitudeRef = 0x15
+    GPSDestLongitude = 0x16
+    GPSDestBearingRef = 0x17
+    GPSDestBearing = 0x18
+    GPSDestDistanceRef = 0x19
+    GPSDestDistance = 0x1A
+    GPSProcessingMethod = 0x1B
+    GPSAreaInformation = 0x1C
+    GPSDateStamp = 0x1D
+    GPSDifferential = 0x1E
+    GPSHPositioningError = 0x1F
+
+
+"""Maps EXIF GPS tags to tag names."""
+GPSTAGS = {i.value: i.name for i in GPS}
+
+
+class Interop(IntEnum):
+    InteropIndex = 0x0001
+    InteropVersion = 0x0002
+    RelatedImageFileFormat = 0x1000
+    RelatedImageWidth = 0x1001
+    RelatedImageHeight = 0x1002
+
+
+class IFD(IntEnum):
+    Exif = 0x8769
+    GPSInfo = 0x8825
+    MakerNote = 0x927C
+    Makernote = 0x927C  # Deprecated
+    Interop = 0xA005
+    IFD1 = -1
+
+
+class LightSource(IntEnum):
+    Unknown = 0x00
+    Daylight = 0x01
+    Fluorescent = 0x02
+    Tungsten = 0x03
+    Flash = 0x04
+    Fine = 0x09
+    Cloudy = 0x0A
+    Shade = 0x0B
+    DaylightFluorescent = 0x0C
+    DayWhiteFluorescent = 0x0D
+    CoolWhiteFluorescent = 0x0E
+    WhiteFluorescent = 0x0F
+    StandardLightA = 0x11
+    StandardLightB = 0x12
+    StandardLightC = 0x13
+    D55 = 0x14
+    D65 = 0x15
+    D75 = 0x16
+    D50 = 0x17
+    ISO = 0x18
+    Other = 0xFF
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/FitsImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/FitsImagePlugin.py
new file mode 100644
index 00000000..6bbd2641
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/FitsImagePlugin.py
@@ -0,0 +1,152 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# FITS file handling
+#
+# Copyright (c) 1998-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import gzip
+import math
+
+from . import Image, ImageFile
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:6] == b"SIMPLE"
+
+
+class FitsImageFile(ImageFile.ImageFile):
+    format = "FITS"
+    format_description = "FITS"
+
+    def _open(self) -> None:
+        assert self.fp is not None
+
+        headers: dict[bytes, bytes] = {}
+        header_in_progress = False
+        decoder_name = ""
+        while True:
+            header = self.fp.read(80)
+            if not header:
+                msg = "Truncated FITS file"
+                raise OSError(msg)
+            keyword = header[:8].strip()
+            if keyword in (b"SIMPLE", b"XTENSION"):
+                header_in_progress = True
+            elif headers and not header_in_progress:
+                # This is now a data unit
+                break
+            elif keyword == b"END":
+                # Seek to the end of the header unit
+                self.fp.seek(math.ceil(self.fp.tell() / 2880) * 2880)
+                if not decoder_name:
+                    decoder_name, offset, args = self._parse_headers(headers)
+
+                header_in_progress = False
+                continue
+
+            if decoder_name:
+                # Keep going to read past the headers
+                continue
+
+            value = header[8:].split(b"/")[0].strip()
+            if value.startswith(b"="):
+                value = value[1:].strip()
+            if not headers and (not _accept(keyword) or value != b"T"):
+                msg = "Not a FITS file"
+                raise SyntaxError(msg)
+            headers[keyword] = value
+
+        if not decoder_name:
+            msg = "No image data"
+            raise ValueError(msg)
+
+        offset += self.fp.tell() - 80
+        self.tile = [ImageFile._Tile(decoder_name, (0, 0) + self.size, offset, args)]
+
+    def _get_size(
+        self, headers: dict[bytes, bytes], prefix: bytes
+    ) -> tuple[int, int] | None:
+        naxis = int(headers[prefix + b"NAXIS"])
+        if naxis == 0:
+            return None
+
+        if naxis == 1:
+            return 1, int(headers[prefix + b"NAXIS1"])
+        else:
+            return int(headers[prefix + b"NAXIS1"]), int(headers[prefix + b"NAXIS2"])
+
+    def _parse_headers(
+        self, headers: dict[bytes, bytes]
+    ) -> tuple[str, int, tuple[str | int, ...]]:
+        prefix = b""
+        decoder_name = "raw"
+        offset = 0
+        if (
+            headers.get(b"XTENSION") == b"'BINTABLE'"
+            and headers.get(b"ZIMAGE") == b"T"
+            and headers[b"ZCMPTYPE"] == b"'GZIP_1  '"
+        ):
+            no_prefix_size = self._get_size(headers, prefix) or (0, 0)
+            number_of_bits = int(headers[b"BITPIX"])
+            offset = no_prefix_size[0] * no_prefix_size[1] * (number_of_bits // 8)
+
+            prefix = b"Z"
+            decoder_name = "fits_gzip"
+
+        size = self._get_size(headers, prefix)
+        if not size:
+            return "", 0, ()
+
+        self._size = size
+
+        number_of_bits = int(headers[prefix + b"BITPIX"])
+        if number_of_bits == 8:
+            self._mode = "L"
+        elif number_of_bits == 16:
+            self._mode = "I;16"
+        elif number_of_bits == 32:
+            self._mode = "I"
+        elif number_of_bits in (-32, -64):
+            self._mode = "F"
+
+        args: tuple[str | int, ...]
+        if decoder_name == "raw":
+            args = (self.mode, 0, -1)
+        else:
+            args = (number_of_bits,)
+        return decoder_name, offset, args
+
+
+class FitsGzipDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+        value = gzip.decompress(self.fd.read())
+
+        rows = []
+        offset = 0
+        number_of_bits = min(self.args[0] // 8, 4)
+        for y in range(self.state.ysize):
+            row = bytearray()
+            for x in range(self.state.xsize):
+                row += value[offset + (4 - number_of_bits) : offset + 4]
+                offset += 4
+            rows.append(row)
+        self.set_as_raw(bytes([pixel for row in rows[::-1] for pixel in row]))
+        return -1, 0
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(FitsImageFile.format, FitsImageFile, _accept)
+Image.register_decoder("fits_gzip", FitsGzipDecoder)
+
+Image.register_extensions(FitsImageFile.format, [".fit", ".fits"])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/FliImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/FliImagePlugin.py
new file mode 100644
index 00000000..b534b30a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/FliImagePlugin.py
@@ -0,0 +1,175 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# FLI/FLC file handling.
+#
+# History:
+#       95-09-01 fl     Created
+#       97-01-03 fl     Fixed parser, setup decoder tile
+#       98-07-15 fl     Renamed offset attribute to avoid name clash
+#
+# Copyright (c) Secret Labs AB 1997-98.
+# Copyright (c) Fredrik Lundh 1995-97.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i16le as i16
+from ._binary import i32le as i32
+from ._binary import o8
+
+#
+# decoder
+
+
+def _accept(prefix: bytes) -> bool:
+    return (
+        len(prefix) >= 6
+        and i16(prefix, 4) in [0xAF11, 0xAF12]
+        and i16(prefix, 14) in [0, 3]  # flags
+    )
+
+
+##
+# Image plugin for the FLI/FLC animation format.  Use the <b>seek</b>
+# method to load individual frames.
+
+
+class FliImageFile(ImageFile.ImageFile):
+    format = "FLI"
+    format_description = "Autodesk FLI/FLC Animation"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        # HEAD
+        s = self.fp.read(128)
+        if not (_accept(s) and s[20:22] == b"\x00\x00"):
+            msg = "not an FLI/FLC file"
+            raise SyntaxError(msg)
+
+        # frames
+        self.n_frames = i16(s, 6)
+        self.is_animated = self.n_frames > 1
+
+        # image characteristics
+        self._mode = "P"
+        self._size = i16(s, 8), i16(s, 10)
+
+        # animation speed
+        duration = i32(s, 16)
+        magic = i16(s, 4)
+        if magic == 0xAF11:
+            duration = (duration * 1000) // 70
+        self.info["duration"] = duration
+
+        # look for palette
+        palette = [(a, a, a) for a in range(256)]
+
+        s = self.fp.read(16)
+
+        self.__offset = 128
+
+        if i16(s, 4) == 0xF100:
+            # prefix chunk; ignore it
+            self.__offset = self.__offset + i32(s)
+            self.fp.seek(self.__offset)
+            s = self.fp.read(16)
+
+        if i16(s, 4) == 0xF1FA:
+            # look for palette chunk
+            number_of_subchunks = i16(s, 6)
+            chunk_size: int | None = None
+            for _ in range(number_of_subchunks):
+                if chunk_size is not None:
+                    self.fp.seek(chunk_size - 6, os.SEEK_CUR)
+                s = self.fp.read(6)
+                chunk_type = i16(s, 4)
+                if chunk_type in (4, 11):
+                    self._palette(palette, 2 if chunk_type == 11 else 0)
+                    break
+                chunk_size = i32(s)
+                if not chunk_size:
+                    break
+
+        self.palette = ImagePalette.raw(
+            "RGB", b"".join(o8(r) + o8(g) + o8(b) for (r, g, b) in palette)
+        )
+
+        # set things up to decode first frame
+        self.__frame = -1
+        self._fp = self.fp
+        self.__rewind = self.fp.tell()
+        self.seek(0)
+
+    def _palette(self, palette: list[tuple[int, int, int]], shift: int) -> None:
+        # load palette
+
+        i = 0
+        for e in range(i16(self.fp.read(2))):
+            s = self.fp.read(2)
+            i = i + s[0]
+            n = s[1]
+            if n == 0:
+                n = 256
+            s = self.fp.read(n * 3)
+            for n in range(0, len(s), 3):
+                r = s[n] << shift
+                g = s[n + 1] << shift
+                b = s[n + 2] << shift
+                palette[i] = (r, g, b)
+                i += 1
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        if frame < self.__frame:
+            self._seek(0)
+
+        for f in range(self.__frame + 1, frame + 1):
+            self._seek(f)
+
+    def _seek(self, frame: int) -> None:
+        if frame == 0:
+            self.__frame = -1
+            self._fp.seek(self.__rewind)
+            self.__offset = 128
+        else:
+            # ensure that the previous frame was loaded
+            self.load()
+
+        if frame != self.__frame + 1:
+            msg = f"cannot seek to frame {frame}"
+            raise ValueError(msg)
+        self.__frame = frame
+
+        # move to next frame
+        self.fp = self._fp
+        self.fp.seek(self.__offset)
+
+        s = self.fp.read(4)
+        if not s:
+            msg = "missing frame size"
+            raise EOFError(msg)
+
+        framesize = i32(s)
+
+        self.decodermaxblock = framesize
+        self.tile = [ImageFile._Tile("fli", (0, 0) + self.size, self.__offset)]
+
+        self.__offset += framesize
+
+    def tell(self) -> int:
+        return self.__frame
+
+
+#
+# registry
+
+Image.register_open(FliImageFile.format, FliImageFile, _accept)
+
+Image.register_extensions(FliImageFile.format, [".fli", ".flc"])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/FontFile.py b/backend/venv/lib/python3.12/site-packages/PIL/FontFile.py
new file mode 100644
index 00000000..1e0c1c16
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/FontFile.py
@@ -0,0 +1,134 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# base class for raster font file parsers
+#
+# history:
+# 1997-06-05 fl   created
+# 1997-08-19 fl   restrict image width
+#
+# Copyright (c) 1997-1998 by Secret Labs AB
+# Copyright (c) 1997-1998 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+from typing import BinaryIO
+
+from . import Image, _binary
+
+WIDTH = 800
+
+
+def puti16(
+    fp: BinaryIO, values: tuple[int, int, int, int, int, int, int, int, int, int]
+) -> None:
+    """Write network order (big-endian) 16-bit sequence"""
+    for v in values:
+        if v < 0:
+            v += 65536
+        fp.write(_binary.o16be(v))
+
+
+class FontFile:
+    """Base class for raster font file handlers."""
+
+    bitmap: Image.Image | None = None
+
+    def __init__(self) -> None:
+        self.info: dict[bytes, bytes | int] = {}
+        self.glyph: list[
+            tuple[
+                tuple[int, int],
+                tuple[int, int, int, int],
+                tuple[int, int, int, int],
+                Image.Image,
+            ]
+            | None
+        ] = [None] * 256
+
+    def __getitem__(self, ix: int) -> (
+        tuple[
+            tuple[int, int],
+            tuple[int, int, int, int],
+            tuple[int, int, int, int],
+            Image.Image,
+        ]
+        | None
+    ):
+        return self.glyph[ix]
+
+    def compile(self) -> None:
+        """Create metrics and bitmap"""
+
+        if self.bitmap:
+            return
+
+        # create bitmap large enough to hold all data
+        h = w = maxwidth = 0
+        lines = 1
+        for glyph in self.glyph:
+            if glyph:
+                d, dst, src, im = glyph
+                h = max(h, src[3] - src[1])
+                w = w + (src[2] - src[0])
+                if w > WIDTH:
+                    lines += 1
+                    w = src[2] - src[0]
+                maxwidth = max(maxwidth, w)
+
+        xsize = maxwidth
+        ysize = lines * h
+
+        if xsize == 0 and ysize == 0:
+            return
+
+        self.ysize = h
+
+        # paste glyphs into bitmap
+        self.bitmap = Image.new("1", (xsize, ysize))
+        self.metrics: list[
+            tuple[tuple[int, int], tuple[int, int, int, int], tuple[int, int, int, int]]
+            | None
+        ] = [None] * 256
+        x = y = 0
+        for i in range(256):
+            glyph = self[i]
+            if glyph:
+                d, dst, src, im = glyph
+                xx = src[2] - src[0]
+                x0, y0 = x, y
+                x = x + xx
+                if x > WIDTH:
+                    x, y = 0, y + h
+                    x0, y0 = x, y
+                    x = xx
+                s = src[0] + x0, src[1] + y0, src[2] + x0, src[3] + y0
+                self.bitmap.paste(im.crop(src), s)
+                self.metrics[i] = d, dst, s
+
+    def save(self, filename: str) -> None:
+        """Save font"""
+
+        self.compile()
+
+        # font data
+        if not self.bitmap:
+            msg = "No bitmap created"
+            raise ValueError(msg)
+        self.bitmap.save(os.path.splitext(filename)[0] + ".pbm", "PNG")
+
+        # font metrics
+        with open(os.path.splitext(filename)[0] + ".pil", "wb") as fp:
+            fp.write(b"PILfont\n")
+            fp.write(f";;;;;;{self.ysize};\n".encode("ascii"))  # HACK!!!
+            fp.write(b"DATA\n")
+            for id in range(256):
+                m = self.metrics[id]
+                if not m:
+                    puti16(fp, (0,) * 10)
+                else:
+                    puti16(fp, m[0] + m[1] + m[2])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/FpxImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/FpxImagePlugin.py
new file mode 100644
index 00000000..4cfcb067
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/FpxImagePlugin.py
@@ -0,0 +1,257 @@
+#
+# THIS IS WORK IN PROGRESS
+#
+# The Python Imaging Library.
+# $Id$
+#
+# FlashPix support for PIL
+#
+# History:
+# 97-01-25 fl   Created (reads uncompressed RGB images only)
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import olefile
+
+from . import Image, ImageFile
+from ._binary import i32le as i32
+
+# we map from colour field tuples to (mode, rawmode) descriptors
+MODES = {
+    # opacity
+    (0x00007FFE,): ("A", "L"),
+    # monochrome
+    (0x00010000,): ("L", "L"),
+    (0x00018000, 0x00017FFE): ("RGBA", "LA"),
+    # photo YCC
+    (0x00020000, 0x00020001, 0x00020002): ("RGB", "YCC;P"),
+    (0x00028000, 0x00028001, 0x00028002, 0x00027FFE): ("RGBA", "YCCA;P"),
+    # standard RGB (NIFRGB)
+    (0x00030000, 0x00030001, 0x00030002): ("RGB", "RGB"),
+    (0x00038000, 0x00038001, 0x00038002, 0x00037FFE): ("RGBA", "RGBA"),
+}
+
+
+#
+# --------------------------------------------------------------------
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:8] == olefile.MAGIC
+
+
+##
+# Image plugin for the FlashPix images.
+
+
+class FpxImageFile(ImageFile.ImageFile):
+    format = "FPX"
+    format_description = "FlashPix"
+
+    def _open(self) -> None:
+        #
+        # read the OLE directory and see if this is a likely
+        # to be a FlashPix file
+
+        try:
+            self.ole = olefile.OleFileIO(self.fp)
+        except OSError as e:
+            msg = "not an FPX file; invalid OLE file"
+            raise SyntaxError(msg) from e
+
+        root = self.ole.root
+        if not root or root.clsid != "56616700-C154-11CE-8553-00AA00A1F95B":
+            msg = "not an FPX file; bad root CLSID"
+            raise SyntaxError(msg)
+
+        self._open_index(1)
+
+    def _open_index(self, index: int = 1) -> None:
+        #
+        # get the Image Contents Property Set
+
+        prop = self.ole.getproperties(
+            [f"Data Object Store {index:06d}", "\005Image Contents"]
+        )
+
+        # size (highest resolution)
+
+        assert isinstance(prop[0x1000002], int)
+        assert isinstance(prop[0x1000003], int)
+        self._size = prop[0x1000002], prop[0x1000003]
+
+        size = max(self.size)
+        i = 1
+        while size > 64:
+            size = size // 2
+            i += 1
+        self.maxid = i - 1
+
+        # mode.  instead of using a single field for this, flashpix
+        # requires you to specify the mode for each channel in each
+        # resolution subimage, and leaves it to the decoder to make
+        # sure that they all match.  for now, we'll cheat and assume
+        # that this is always the case.
+
+        id = self.maxid << 16
+
+        s = prop[0x2000002 | id]
+
+        if not isinstance(s, bytes) or (bands := i32(s, 4)) > 4:
+            msg = "Invalid number of bands"
+            raise OSError(msg)
+
+        # note: for now, we ignore the "uncalibrated" flag
+        colors = tuple(i32(s, 8 + i * 4) & 0x7FFFFFFF for i in range(bands))
+
+        self._mode, self.rawmode = MODES[colors]
+
+        # load JPEG tables, if any
+        self.jpeg = {}
+        for i in range(256):
+            id = 0x3000001 | (i << 16)
+            if id in prop:
+                self.jpeg[i] = prop[id]
+
+        self._open_subimage(1, self.maxid)
+
+    def _open_subimage(self, index: int = 1, subimage: int = 0) -> None:
+        #
+        # setup tile descriptors for a given subimage
+
+        stream = [
+            f"Data Object Store {index:06d}",
+            f"Resolution {subimage:04d}",
+            "Subimage 0000 Header",
+        ]
+
+        fp = self.ole.openstream(stream)
+
+        # skip prefix
+        fp.read(28)
+
+        # header stream
+        s = fp.read(36)
+
+        size = i32(s, 4), i32(s, 8)
+        # tilecount = i32(s, 12)
+        tilesize = i32(s, 16), i32(s, 20)
+        # channels = i32(s, 24)
+        offset = i32(s, 28)
+        length = i32(s, 32)
+
+        if size != self.size:
+            msg = "subimage mismatch"
+            raise OSError(msg)
+
+        # get tile descriptors
+        fp.seek(28 + offset)
+        s = fp.read(i32(s, 12) * length)
+
+        x = y = 0
+        xsize, ysize = size
+        xtile, ytile = tilesize
+        self.tile = []
+
+        for i in range(0, len(s), length):
+            x1 = min(xsize, x + xtile)
+            y1 = min(ysize, y + ytile)
+
+            compression = i32(s, i + 8)
+
+            if compression == 0:
+                self.tile.append(
+                    ImageFile._Tile(
+                        "raw",
+                        (x, y, x1, y1),
+                        i32(s, i) + 28,
+                        self.rawmode,
+                    )
+                )
+
+            elif compression == 1:
+                # FIXME: the fill decoder is not implemented
+                self.tile.append(
+                    ImageFile._Tile(
+                        "fill",
+                        (x, y, x1, y1),
+                        i32(s, i) + 28,
+                        (self.rawmode, s[12:16]),
+                    )
+                )
+
+            elif compression == 2:
+                internal_color_conversion = s[14]
+                jpeg_tables = s[15]
+                rawmode = self.rawmode
+
+                if internal_color_conversion:
+                    # The image is stored as usual (usually YCbCr).
+                    if rawmode == "RGBA":
+                        # For "RGBA", data is stored as YCbCrA based on
+                        # negative RGB. The following trick works around
+                        # this problem :
+                        jpegmode, rawmode = "YCbCrK", "CMYK"
+                    else:
+                        jpegmode = None  # let the decoder decide
+
+                else:
+                    # The image is stored as defined by rawmode
+                    jpegmode = rawmode
+
+                self.tile.append(
+                    ImageFile._Tile(
+                        "jpeg",
+                        (x, y, x1, y1),
+                        i32(s, i) + 28,
+                        (rawmode, jpegmode),
+                    )
+                )
+
+                # FIXME: jpeg tables are tile dependent; the prefix
+                # data must be placed in the tile descriptor itself!
+
+                if jpeg_tables:
+                    self.tile_prefix = self.jpeg[jpeg_tables]
+
+            else:
+                msg = "unknown/invalid compression"
+                raise OSError(msg)
+
+            x = x + xtile
+            if x >= xsize:
+                x, y = 0, y + ytile
+                if y >= ysize:
+                    break  # isn't really required
+
+        self.stream = stream
+        self._fp = self.fp
+        self.fp = None
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if not self.fp:
+            self.fp = self.ole.openstream(self.stream[:2] + ["Subimage 0000 Data"])
+
+        return ImageFile.ImageFile.load(self)
+
+    def close(self) -> None:
+        self.ole.close()
+        super().close()
+
+    def __exit__(self, *args: object) -> None:
+        self.ole.close()
+        super().__exit__()
+
+
+#
+# --------------------------------------------------------------------
+
+
+Image.register_open(FpxImageFile.format, FpxImageFile, _accept)
+
+Image.register_extension(FpxImageFile.format, ".fpx")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/FtexImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/FtexImagePlugin.py
new file mode 100644
index 00000000..0516b760
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/FtexImagePlugin.py
@@ -0,0 +1,115 @@
+"""
+A Pillow loader for .ftc and .ftu files (FTEX)
+Jerome Leclanche <jerome@leclan.ch>
+
+The contents of this file are hereby released in the public domain (CC0)
+Full text of the CC0 license:
+  https://creativecommons.org/publicdomain/zero/1.0/
+
+Independence War 2: Edge Of Chaos - Texture File Format - 16 October 2001
+
+The textures used for 3D objects in Independence War 2: Edge Of Chaos are in a
+packed custom format called FTEX. This file format uses file extensions FTC
+and FTU.
+* FTC files are compressed textures (using standard texture compression).
+* FTU files are not compressed.
+Texture File Format
+The FTC and FTU texture files both use the same format. This
+has the following structure:
+{header}
+{format_directory}
+{data}
+Where:
+{header} = {
+    u32:magic,
+    u32:version,
+    u32:width,
+    u32:height,
+    u32:mipmap_count,
+    u32:format_count
+}
+
+* The "magic" number is "FTEX".
+* "width" and "height" are the dimensions of the texture.
+* "mipmap_count" is the number of mipmaps in the texture.
+* "format_count" is the number of texture formats (different versions of the
+same texture) in this file.
+
+{format_directory} = format_count * { u32:format, u32:where }
+
+The format value is 0 for DXT1 compressed textures and 1 for 24-bit RGB
+uncompressed textures.
+The texture data for a format starts at the position "where" in the file.
+
+Each set of texture data in the file has the following structure:
+{data} = format_count * { u32:mipmap_size, mipmap_size * { u8 } }
+* "mipmap_size" is the number of bytes in that mip level. For compressed
+textures this is the size of the texture data compressed with DXT1. For 24 bit
+uncompressed textures, this is 3 * width * height. Following this are the image
+bytes for that mipmap level.
+
+Note: All data is stored in little-Endian (Intel) byte order.
+"""
+
+from __future__ import annotations
+
+import struct
+from enum import IntEnum
+from io import BytesIO
+
+from . import Image, ImageFile
+
+MAGIC = b"FTEX"
+
+
+class Format(IntEnum):
+    DXT1 = 0
+    UNCOMPRESSED = 1
+
+
+class FtexImageFile(ImageFile.ImageFile):
+    format = "FTEX"
+    format_description = "Texture File Format (IW2:EOC)"
+
+    def _open(self) -> None:
+        if not _accept(self.fp.read(4)):
+            msg = "not an FTEX file"
+            raise SyntaxError(msg)
+        struct.unpack("<i", self.fp.read(4))  # version
+        self._size = struct.unpack("<2i", self.fp.read(8))
+        mipmap_count, format_count = struct.unpack("<2i", self.fp.read(8))
+
+        self._mode = "RGB"
+
+        # Only support single-format files.
+        # I don't know of any multi-format file.
+        assert format_count == 1
+
+        format, where = struct.unpack("<2i", self.fp.read(8))
+        self.fp.seek(where)
+        (mipmap_size,) = struct.unpack("<i", self.fp.read(4))
+
+        data = self.fp.read(mipmap_size)
+
+        if format == Format.DXT1:
+            self._mode = "RGBA"
+            self.tile = [ImageFile._Tile("bcn", (0, 0) + self.size, 0, (1,))]
+        elif format == Format.UNCOMPRESSED:
+            self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, 0, "RGB")]
+        else:
+            msg = f"Invalid texture compression format: {repr(format)}"
+            raise ValueError(msg)
+
+        self.fp.close()
+        self.fp = BytesIO(data)
+
+    def load_seek(self, pos: int) -> None:
+        pass
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == MAGIC
+
+
+Image.register_open(FtexImageFile.format, FtexImageFile, _accept)
+Image.register_extensions(FtexImageFile.format, [".ftc", ".ftu"])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GbrImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/GbrImagePlugin.py
new file mode 100644
index 00000000..f319d7e8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GbrImagePlugin.py
@@ -0,0 +1,103 @@
+#
+# The Python Imaging Library
+#
+# load a GIMP brush file
+#
+# History:
+#       96-03-14 fl     Created
+#       16-01-08 es     Version 2
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+# Copyright (c) Eric Soroos 2016.
+#
+# See the README file for information on usage and redistribution.
+#
+#
+# See https://github.com/GNOME/gimp/blob/mainline/devel-docs/gbr.txt for
+# format documentation.
+#
+# This code Interprets version 1 and 2 .gbr files.
+# Version 1 files are obsolete, and should not be used for new
+#   brushes.
+# Version 2 files are saved by GIMP v2.8 (at least)
+# Version 3 files have a format specifier of 18 for 16bit floats in
+#   the color depth field. This is currently unsupported by Pillow.
+from __future__ import annotations
+
+from . import Image, ImageFile
+from ._binary import i32be as i32
+
+
+def _accept(prefix: bytes) -> bool:
+    return len(prefix) >= 8 and i32(prefix, 0) >= 20 and i32(prefix, 4) in (1, 2)
+
+
+##
+# Image plugin for the GIMP brush format.
+
+
+class GbrImageFile(ImageFile.ImageFile):
+    format = "GBR"
+    format_description = "GIMP brush file"
+
+    def _open(self) -> None:
+        header_size = i32(self.fp.read(4))
+        if header_size < 20:
+            msg = "not a GIMP brush"
+            raise SyntaxError(msg)
+        version = i32(self.fp.read(4))
+        if version not in (1, 2):
+            msg = f"Unsupported GIMP brush version: {version}"
+            raise SyntaxError(msg)
+
+        width = i32(self.fp.read(4))
+        height = i32(self.fp.read(4))
+        color_depth = i32(self.fp.read(4))
+        if width <= 0 or height <= 0:
+            msg = "not a GIMP brush"
+            raise SyntaxError(msg)
+        if color_depth not in (1, 4):
+            msg = f"Unsupported GIMP brush color depth: {color_depth}"
+            raise SyntaxError(msg)
+
+        if version == 1:
+            comment_length = header_size - 20
+        else:
+            comment_length = header_size - 28
+            magic_number = self.fp.read(4)
+            if magic_number != b"GIMP":
+                msg = "not a GIMP brush, bad magic number"
+                raise SyntaxError(msg)
+            self.info["spacing"] = i32(self.fp.read(4))
+
+        comment = self.fp.read(comment_length)[:-1]
+
+        if color_depth == 1:
+            self._mode = "L"
+        else:
+            self._mode = "RGBA"
+
+        self._size = width, height
+
+        self.info["comment"] = comment
+
+        # Image might not be small
+        Image._decompression_bomb_check(self.size)
+
+        # Data is an uncompressed block of w * h * bytes/pixel
+        self._data_size = width * height * color_depth
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self._im is None:
+            self.im = Image.core.new(self.mode, self.size)
+            self.frombytes(self.fp.read(self._data_size))
+        return Image.Image.load(self)
+
+
+#
+# registry
+
+
+Image.register_open(GbrImageFile.format, GbrImageFile, _accept)
+Image.register_extension(GbrImageFile.format, ".gbr")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GdImageFile.py b/backend/venv/lib/python3.12/site-packages/PIL/GdImageFile.py
new file mode 100644
index 00000000..fc4801e9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GdImageFile.py
@@ -0,0 +1,102 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# GD file handling
+#
+# History:
+# 1996-04-12 fl   Created
+#
+# Copyright (c) 1997 by Secret Labs AB.
+# Copyright (c) 1996 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+
+"""
+.. note::
+    This format cannot be automatically recognized, so the
+    class is not registered for use with :py:func:`PIL.Image.open()`.  To open a
+    gd file, use the :py:func:`PIL.GdImageFile.open()` function instead.
+
+.. warning::
+    THE GD FORMAT IS NOT DESIGNED FOR DATA INTERCHANGE.  This
+    implementation is provided for convenience and demonstrational
+    purposes only.
+"""
+from __future__ import annotations
+
+from typing import IO
+
+from . import ImageFile, ImagePalette, UnidentifiedImageError
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._typing import StrOrBytesPath
+
+
+class GdImageFile(ImageFile.ImageFile):
+    """
+    Image plugin for the GD uncompressed format.  Note that this format
+    is not supported by the standard :py:func:`PIL.Image.open()` function.  To use
+    this plugin, you have to import the :py:mod:`PIL.GdImageFile` module and
+    use the :py:func:`PIL.GdImageFile.open()` function.
+    """
+
+    format = "GD"
+    format_description = "GD uncompressed images"
+
+    def _open(self) -> None:
+        # Header
+        assert self.fp is not None
+
+        s = self.fp.read(1037)
+
+        if i16(s) not in [65534, 65535]:
+            msg = "Not a valid GD 2.x .gd file"
+            raise SyntaxError(msg)
+
+        self._mode = "L"  # FIXME: "P"
+        self._size = i16(s, 2), i16(s, 4)
+
+        true_color = s[6]
+        true_color_offset = 2 if true_color else 0
+
+        # transparency index
+        tindex = i32(s, 7 + true_color_offset)
+        if tindex < 256:
+            self.info["transparency"] = tindex
+
+        self.palette = ImagePalette.raw(
+            "XBGR", s[7 + true_color_offset + 4 : 7 + true_color_offset + 4 + 256 * 4]
+        )
+
+        self.tile = [
+            ImageFile._Tile(
+                "raw",
+                (0, 0) + self.size,
+                7 + true_color_offset + 4 + 256 * 4,
+                "L",
+            )
+        ]
+
+
+def open(fp: StrOrBytesPath | IO[bytes], mode: str = "r") -> GdImageFile:
+    """
+    Load texture from a GD image file.
+
+    :param fp: GD file name, or an opened file handle.
+    :param mode: Optional mode.  In this version, if the mode argument
+        is given, it must be "r".
+    :returns: An image instance.
+    :raises OSError: If the image could not be read.
+    """
+    if mode != "r":
+        msg = "bad mode"
+        raise ValueError(msg)
+
+    try:
+        return GdImageFile(fp)
+    except SyntaxError as e:
+        msg = "cannot identify this image file"
+        raise UnidentifiedImageError(msg) from e
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GifImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/GifImagePlugin.py
new file mode 100644
index 00000000..47022d58
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GifImagePlugin.py
@@ -0,0 +1,1197 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# GIF file handling
+#
+# History:
+# 1995-09-01 fl   Created
+# 1996-12-14 fl   Added interlace support
+# 1996-12-30 fl   Added animation support
+# 1997-01-05 fl   Added write support, fixed local colour map bug
+# 1997-02-23 fl   Make sure to load raster data in getdata()
+# 1997-07-05 fl   Support external decoder (0.4)
+# 1998-07-09 fl   Handle all modes when saving (0.5)
+# 1998-07-15 fl   Renamed offset attribute to avoid name clash
+# 2001-04-16 fl   Added rewind support (seek to frame 0) (0.6)
+# 2001-04-17 fl   Added palette optimization (0.7)
+# 2002-06-06 fl   Added transparency support for save (0.8)
+# 2004-02-24 fl   Disable interlacing for small images
+#
+# Copyright (c) 1997-2004 by Secret Labs AB
+# Copyright (c) 1995-2004 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import itertools
+import math
+import os
+import subprocess
+from enum import IntEnum
+from functools import cached_property
+from typing import IO, TYPE_CHECKING, Any, Literal, NamedTuple, Union
+
+from . import (
+    Image,
+    ImageChops,
+    ImageFile,
+    ImageMath,
+    ImageOps,
+    ImagePalette,
+    ImageSequence,
+)
+from ._binary import i16le as i16
+from ._binary import o8
+from ._binary import o16le as o16
+
+if TYPE_CHECKING:
+    from . import _imaging
+    from ._typing import Buffer
+
+
+class LoadingStrategy(IntEnum):
+    """.. versionadded:: 9.1.0"""
+
+    RGB_AFTER_FIRST = 0
+    RGB_AFTER_DIFFERENT_PALETTE_ONLY = 1
+    RGB_ALWAYS = 2
+
+
+#: .. versionadded:: 9.1.0
+LOADING_STRATEGY = LoadingStrategy.RGB_AFTER_FIRST
+
+# --------------------------------------------------------------------
+# Identify/read GIF files
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:6] in [b"GIF87a", b"GIF89a"]
+
+
+##
+# Image plugin for GIF images.  This plugin supports both GIF87 and
+# GIF89 images.
+
+
+class GifImageFile(ImageFile.ImageFile):
+    format = "GIF"
+    format_description = "Compuserve GIF"
+    _close_exclusive_fp_after_loading = False
+
+    global_palette = None
+
+    def data(self) -> bytes | None:
+        s = self.fp.read(1)
+        if s and s[0]:
+            return self.fp.read(s[0])
+        return None
+
+    def _is_palette_needed(self, p: bytes) -> bool:
+        for i in range(0, len(p), 3):
+            if not (i // 3 == p[i] == p[i + 1] == p[i + 2]):
+                return True
+        return False
+
+    def _open(self) -> None:
+        # Screen
+        s = self.fp.read(13)
+        if not _accept(s):
+            msg = "not a GIF file"
+            raise SyntaxError(msg)
+
+        self.info["version"] = s[:6]
+        self._size = i16(s, 6), i16(s, 8)
+        flags = s[10]
+        bits = (flags & 7) + 1
+
+        if flags & 128:
+            # get global palette
+            self.info["background"] = s[11]
+            # check if palette contains colour indices
+            p = self.fp.read(3 << bits)
+            if self._is_palette_needed(p):
+                p = ImagePalette.raw("RGB", p)
+                self.global_palette = self.palette = p
+
+        self._fp = self.fp  # FIXME: hack
+        self.__rewind = self.fp.tell()
+        self._n_frames: int | None = None
+        self._seek(0)  # get ready to read first frame
+
+    @property
+    def n_frames(self) -> int:
+        if self._n_frames is None:
+            current = self.tell()
+            try:
+                while True:
+                    self._seek(self.tell() + 1, False)
+            except EOFError:
+                self._n_frames = self.tell() + 1
+            self.seek(current)
+        return self._n_frames
+
+    @cached_property
+    def is_animated(self) -> bool:
+        if self._n_frames is not None:
+            return self._n_frames != 1
+
+        current = self.tell()
+        if current:
+            return True
+
+        try:
+            self._seek(1, False)
+            is_animated = True
+        except EOFError:
+            is_animated = False
+
+        self.seek(current)
+        return is_animated
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        if frame < self.__frame:
+            self._im = None
+            self._seek(0)
+
+        last_frame = self.__frame
+        for f in range(self.__frame + 1, frame + 1):
+            try:
+                self._seek(f)
+            except EOFError as e:
+                self.seek(last_frame)
+                msg = "no more images in GIF file"
+                raise EOFError(msg) from e
+
+    def _seek(self, frame: int, update_image: bool = True) -> None:
+        if frame == 0:
+            # rewind
+            self.__offset = 0
+            self.dispose: _imaging.ImagingCore | None = None
+            self.__frame = -1
+            self._fp.seek(self.__rewind)
+            self.disposal_method = 0
+            if "comment" in self.info:
+                del self.info["comment"]
+        else:
+            # ensure that the previous frame was loaded
+            if self.tile and update_image:
+                self.load()
+
+        if frame != self.__frame + 1:
+            msg = f"cannot seek to frame {frame}"
+            raise ValueError(msg)
+
+        self.fp = self._fp
+        if self.__offset:
+            # backup to last frame
+            self.fp.seek(self.__offset)
+            while self.data():
+                pass
+            self.__offset = 0
+
+        s = self.fp.read(1)
+        if not s or s == b";":
+            msg = "no more images in GIF file"
+            raise EOFError(msg)
+
+        palette: ImagePalette.ImagePalette | Literal[False] | None = None
+
+        info: dict[str, Any] = {}
+        frame_transparency = None
+        interlace = None
+        frame_dispose_extent = None
+        while True:
+            if not s:
+                s = self.fp.read(1)
+            if not s or s == b";":
+                break
+
+            elif s == b"!":
+                #
+                # extensions
+                #
+                s = self.fp.read(1)
+                block = self.data()
+                if s[0] == 249 and block is not None:
+                    #
+                    # graphic control extension
+                    #
+                    flags = block[0]
+                    if flags & 1:
+                        frame_transparency = block[3]
+                    info["duration"] = i16(block, 1) * 10
+
+                    # disposal method - find the value of bits 4 - 6
+                    dispose_bits = 0b00011100 & flags
+                    dispose_bits = dispose_bits >> 2
+                    if dispose_bits:
+                        # only set the dispose if it is not
+                        # unspecified. I'm not sure if this is
+                        # correct, but it seems to prevent the last
+                        # frame from looking odd for some animations
+                        self.disposal_method = dispose_bits
+                elif s[0] == 254:
+                    #
+                    # comment extension
+                    #
+                    comment = b""
+
+                    # Read this comment block
+                    while block:
+                        comment += block
+                        block = self.data()
+
+                    if "comment" in info:
+                        # If multiple comment blocks in frame, separate with \n
+                        info["comment"] += b"\n" + comment
+                    else:
+                        info["comment"] = comment
+                    s = None
+                    continue
+                elif s[0] == 255 and frame == 0 and block is not None:
+                    #
+                    # application extension
+                    #
+                    info["extension"] = block, self.fp.tell()
+                    if block[:11] == b"NETSCAPE2.0":
+                        block = self.data()
+                        if block and len(block) >= 3 and block[0] == 1:
+                            self.info["loop"] = i16(block, 1)
+                while self.data():
+                    pass
+
+            elif s == b",":
+                #
+                # local image
+                #
+                s = self.fp.read(9)
+
+                # extent
+                x0, y0 = i16(s, 0), i16(s, 2)
+                x1, y1 = x0 + i16(s, 4), y0 + i16(s, 6)
+                if (x1 > self.size[0] or y1 > self.size[1]) and update_image:
+                    self._size = max(x1, self.size[0]), max(y1, self.size[1])
+                    Image._decompression_bomb_check(self._size)
+                frame_dispose_extent = x0, y0, x1, y1
+                flags = s[8]
+
+                interlace = (flags & 64) != 0
+
+                if flags & 128:
+                    bits = (flags & 7) + 1
+                    p = self.fp.read(3 << bits)
+                    if self._is_palette_needed(p):
+                        palette = ImagePalette.raw("RGB", p)
+                    else:
+                        palette = False
+
+                # image data
+                bits = self.fp.read(1)[0]
+                self.__offset = self.fp.tell()
+                break
+            s = None
+
+        if interlace is None:
+            msg = "image not found in GIF frame"
+            raise EOFError(msg)
+
+        self.__frame = frame
+        if not update_image:
+            return
+
+        self.tile = []
+
+        if self.dispose:
+            self.im.paste(self.dispose, self.dispose_extent)
+
+        self._frame_palette = palette if palette is not None else self.global_palette
+        self._frame_transparency = frame_transparency
+        if frame == 0:
+            if self._frame_palette:
+                if LOADING_STRATEGY == LoadingStrategy.RGB_ALWAYS:
+                    self._mode = "RGBA" if frame_transparency is not None else "RGB"
+                else:
+                    self._mode = "P"
+            else:
+                self._mode = "L"
+
+            if palette:
+                self.palette = palette
+            elif self.global_palette:
+                from copy import copy
+
+                self.palette = copy(self.global_palette)
+            else:
+                self.palette = None
+        else:
+            if self.mode == "P":
+                if (
+                    LOADING_STRATEGY != LoadingStrategy.RGB_AFTER_DIFFERENT_PALETTE_ONLY
+                    or palette
+                ):
+                    if "transparency" in self.info:
+                        self.im.putpalettealpha(self.info["transparency"], 0)
+                        self.im = self.im.convert("RGBA", Image.Dither.FLOYDSTEINBERG)
+                        self._mode = "RGBA"
+                        del self.info["transparency"]
+                    else:
+                        self._mode = "RGB"
+                        self.im = self.im.convert("RGB", Image.Dither.FLOYDSTEINBERG)
+
+        def _rgb(color: int) -> tuple[int, int, int]:
+            if self._frame_palette:
+                if color * 3 + 3 > len(self._frame_palette.palette):
+                    color = 0
+                return tuple(self._frame_palette.palette[color * 3 : color * 3 + 3])
+            else:
+                return (color, color, color)
+
+        self.dispose = None
+        self.dispose_extent = frame_dispose_extent
+        if self.dispose_extent and self.disposal_method >= 2:
+            try:
+                if self.disposal_method == 2:
+                    # replace with background colour
+
+                    # only dispose the extent in this frame
+                    x0, y0, x1, y1 = self.dispose_extent
+                    dispose_size = (x1 - x0, y1 - y0)
+
+                    Image._decompression_bomb_check(dispose_size)
+
+                    # by convention, attempt to use transparency first
+                    dispose_mode = "P"
+                    color = self.info.get("transparency", frame_transparency)
+                    if color is not None:
+                        if self.mode in ("RGB", "RGBA"):
+                            dispose_mode = "RGBA"
+                            color = _rgb(color) + (0,)
+                    else:
+                        color = self.info.get("background", 0)
+                        if self.mode in ("RGB", "RGBA"):
+                            dispose_mode = "RGB"
+                            color = _rgb(color)
+                    self.dispose = Image.core.fill(dispose_mode, dispose_size, color)
+                else:
+                    # replace with previous contents
+                    if self._im is not None:
+                        # only dispose the extent in this frame
+                        self.dispose = self._crop(self.im, self.dispose_extent)
+                    elif frame_transparency is not None:
+                        x0, y0, x1, y1 = self.dispose_extent
+                        dispose_size = (x1 - x0, y1 - y0)
+
+                        Image._decompression_bomb_check(dispose_size)
+                        dispose_mode = "P"
+                        color = frame_transparency
+                        if self.mode in ("RGB", "RGBA"):
+                            dispose_mode = "RGBA"
+                            color = _rgb(frame_transparency) + (0,)
+                        self.dispose = Image.core.fill(
+                            dispose_mode, dispose_size, color
+                        )
+            except AttributeError:
+                pass
+
+        if interlace is not None:
+            transparency = -1
+            if frame_transparency is not None:
+                if frame == 0:
+                    if LOADING_STRATEGY != LoadingStrategy.RGB_ALWAYS:
+                        self.info["transparency"] = frame_transparency
+                elif self.mode not in ("RGB", "RGBA"):
+                    transparency = frame_transparency
+            self.tile = [
+                ImageFile._Tile(
+                    "gif",
+                    (x0, y0, x1, y1),
+                    self.__offset,
+                    (bits, interlace, transparency),
+                )
+            ]
+
+        if info.get("comment"):
+            self.info["comment"] = info["comment"]
+        for k in ["duration", "extension"]:
+            if k in info:
+                self.info[k] = info[k]
+            elif k in self.info:
+                del self.info[k]
+
+    def load_prepare(self) -> None:
+        temp_mode = "P" if self._frame_palette else "L"
+        self._prev_im = None
+        if self.__frame == 0:
+            if self._frame_transparency is not None:
+                self.im = Image.core.fill(
+                    temp_mode, self.size, self._frame_transparency
+                )
+        elif self.mode in ("RGB", "RGBA"):
+            self._prev_im = self.im
+            if self._frame_palette:
+                self.im = Image.core.fill("P", self.size, self._frame_transparency or 0)
+                self.im.putpalette("RGB", *self._frame_palette.getdata())
+            else:
+                self._im = None
+        if not self._prev_im and self._im is not None and self.size != self.im.size:
+            expanded_im = Image.core.fill(self.im.mode, self.size)
+            if self._frame_palette:
+                expanded_im.putpalette("RGB", *self._frame_palette.getdata())
+            expanded_im.paste(self.im, (0, 0) + self.im.size)
+
+            self.im = expanded_im
+        self._mode = temp_mode
+        self._frame_palette = None
+
+        super().load_prepare()
+
+    def load_end(self) -> None:
+        if self.__frame == 0:
+            if self.mode == "P" and LOADING_STRATEGY == LoadingStrategy.RGB_ALWAYS:
+                if self._frame_transparency is not None:
+                    self.im.putpalettealpha(self._frame_transparency, 0)
+                    self._mode = "RGBA"
+                else:
+                    self._mode = "RGB"
+                self.im = self.im.convert(self.mode, Image.Dither.FLOYDSTEINBERG)
+            return
+        if not self._prev_im:
+            return
+        if self.size != self._prev_im.size:
+            if self._frame_transparency is not None:
+                expanded_im = Image.core.fill("RGBA", self.size)
+            else:
+                expanded_im = Image.core.fill("P", self.size)
+                expanded_im.putpalette("RGB", "RGB", self.im.getpalette())
+                expanded_im = expanded_im.convert("RGB")
+            expanded_im.paste(self._prev_im, (0, 0) + self._prev_im.size)
+
+            self._prev_im = expanded_im
+            assert self._prev_im is not None
+        if self._frame_transparency is not None:
+            self.im.putpalettealpha(self._frame_transparency, 0)
+            frame_im = self.im.convert("RGBA")
+        else:
+            frame_im = self.im.convert("RGB")
+
+        assert self.dispose_extent is not None
+        frame_im = self._crop(frame_im, self.dispose_extent)
+
+        self.im = self._prev_im
+        self._mode = self.im.mode
+        if frame_im.mode == "RGBA":
+            self.im.paste(frame_im, self.dispose_extent, frame_im)
+        else:
+            self.im.paste(frame_im, self.dispose_extent)
+
+    def tell(self) -> int:
+        return self.__frame
+
+
+# --------------------------------------------------------------------
+# Write GIF files
+
+
+RAWMODE = {"1": "L", "L": "L", "P": "P"}
+
+
+def _normalize_mode(im: Image.Image) -> Image.Image:
+    """
+    Takes an image (or frame), returns an image in a mode that is appropriate
+    for saving in a Gif.
+
+    It may return the original image, or it may return an image converted to
+    palette or 'L' mode.
+
+    :param im: Image object
+    :returns: Image object
+    """
+    if im.mode in RAWMODE:
+        im.load()
+        return im
+    if Image.getmodebase(im.mode) == "RGB":
+        im = im.convert("P", palette=Image.Palette.ADAPTIVE)
+        assert im.palette is not None
+        if im.palette.mode == "RGBA":
+            for rgba in im.palette.colors:
+                if rgba[3] == 0:
+                    im.info["transparency"] = im.palette.colors[rgba]
+                    break
+        return im
+    return im.convert("L")
+
+
+_Palette = Union[bytes, bytearray, list[int], ImagePalette.ImagePalette]
+
+
+def _normalize_palette(
+    im: Image.Image, palette: _Palette | None, info: dict[str, Any]
+) -> Image.Image:
+    """
+    Normalizes the palette for image.
+      - Sets the palette to the incoming palette, if provided.
+      - Ensures that there's a palette for L mode images
+      - Optimizes the palette if necessary/desired.
+
+    :param im: Image object
+    :param palette: bytes object containing the source palette, or ....
+    :param info: encoderinfo
+    :returns: Image object
+    """
+    source_palette = None
+    if palette:
+        # a bytes palette
+        if isinstance(palette, (bytes, bytearray, list)):
+            source_palette = bytearray(palette[:768])
+        if isinstance(palette, ImagePalette.ImagePalette):
+            source_palette = bytearray(palette.palette)
+
+    if im.mode == "P":
+        if not source_palette:
+            im_palette = im.getpalette(None)
+            assert im_palette is not None
+            source_palette = bytearray(im_palette)
+    else:  # L-mode
+        if not source_palette:
+            source_palette = bytearray(i // 3 for i in range(768))
+        im.palette = ImagePalette.ImagePalette("RGB", palette=source_palette)
+    assert source_palette is not None
+
+    if palette:
+        used_palette_colors: list[int | None] = []
+        assert im.palette is not None
+        for i in range(0, len(source_palette), 3):
+            source_color = tuple(source_palette[i : i + 3])
+            index = im.palette.colors.get(source_color)
+            if index in used_palette_colors:
+                index = None
+            used_palette_colors.append(index)
+        for i, index in enumerate(used_palette_colors):
+            if index is None:
+                for j in range(len(used_palette_colors)):
+                    if j not in used_palette_colors:
+                        used_palette_colors[i] = j
+                        break
+        dest_map: list[int] = []
+        for index in used_palette_colors:
+            assert index is not None
+            dest_map.append(index)
+        im = im.remap_palette(dest_map)
+    else:
+        optimized_palette_colors = _get_optimize(im, info)
+        if optimized_palette_colors is not None:
+            im = im.remap_palette(optimized_palette_colors, source_palette)
+            if "transparency" in info:
+                try:
+                    info["transparency"] = optimized_palette_colors.index(
+                        info["transparency"]
+                    )
+                except ValueError:
+                    del info["transparency"]
+            return im
+
+    assert im.palette is not None
+    im.palette.palette = source_palette
+    return im
+
+
+def _write_single_frame(
+    im: Image.Image,
+    fp: IO[bytes],
+    palette: _Palette | None,
+) -> None:
+    im_out = _normalize_mode(im)
+    for k, v in im_out.info.items():
+        if isinstance(k, str):
+            im.encoderinfo.setdefault(k, v)
+    im_out = _normalize_palette(im_out, palette, im.encoderinfo)
+
+    for s in _get_global_header(im_out, im.encoderinfo):
+        fp.write(s)
+
+    # local image header
+    flags = 0
+    if get_interlace(im):
+        flags = flags | 64
+    _write_local_header(fp, im, (0, 0), flags)
+
+    im_out.encoderconfig = (8, get_interlace(im))
+    ImageFile._save(
+        im_out, fp, [ImageFile._Tile("gif", (0, 0) + im.size, 0, RAWMODE[im_out.mode])]
+    )
+
+    fp.write(b"\0")  # end of image data
+
+
+def _getbbox(
+    base_im: Image.Image, im_frame: Image.Image
+) -> tuple[Image.Image, tuple[int, int, int, int] | None]:
+    palette_bytes = [
+        bytes(im.palette.palette) if im.palette else b"" for im in (base_im, im_frame)
+    ]
+    if palette_bytes[0] != palette_bytes[1]:
+        im_frame = im_frame.convert("RGBA")
+        base_im = base_im.convert("RGBA")
+    delta = ImageChops.subtract_modulo(im_frame, base_im)
+    return delta, delta.getbbox(alpha_only=False)
+
+
+class _Frame(NamedTuple):
+    im: Image.Image
+    bbox: tuple[int, int, int, int] | None
+    encoderinfo: dict[str, Any]
+
+
+def _write_multiple_frames(
+    im: Image.Image, fp: IO[bytes], palette: _Palette | None
+) -> bool:
+    duration = im.encoderinfo.get("duration")
+    disposal = im.encoderinfo.get("disposal", im.info.get("disposal"))
+
+    im_frames: list[_Frame] = []
+    previous_im: Image.Image | None = None
+    frame_count = 0
+    background_im = None
+    for imSequence in itertools.chain([im], im.encoderinfo.get("append_images", [])):
+        for im_frame in ImageSequence.Iterator(imSequence):
+            # a copy is required here since seek can still mutate the image
+            im_frame = _normalize_mode(im_frame.copy())
+            if frame_count == 0:
+                for k, v in im_frame.info.items():
+                    if k == "transparency":
+                        continue
+                    if isinstance(k, str):
+                        im.encoderinfo.setdefault(k, v)
+
+            encoderinfo = im.encoderinfo.copy()
+            if "transparency" in im_frame.info:
+                encoderinfo.setdefault("transparency", im_frame.info["transparency"])
+            im_frame = _normalize_palette(im_frame, palette, encoderinfo)
+            if isinstance(duration, (list, tuple)):
+                encoderinfo["duration"] = duration[frame_count]
+            elif duration is None and "duration" in im_frame.info:
+                encoderinfo["duration"] = im_frame.info["duration"]
+            if isinstance(disposal, (list, tuple)):
+                encoderinfo["disposal"] = disposal[frame_count]
+            frame_count += 1
+
+            diff_frame = None
+            if im_frames and previous_im:
+                # delta frame
+                delta, bbox = _getbbox(previous_im, im_frame)
+                if not bbox:
+                    # This frame is identical to the previous frame
+                    if encoderinfo.get("duration"):
+                        im_frames[-1].encoderinfo["duration"] += encoderinfo["duration"]
+                    continue
+                if im_frames[-1].encoderinfo.get("disposal") == 2:
+                    if background_im is None:
+                        color = im.encoderinfo.get(
+                            "transparency", im.info.get("transparency", (0, 0, 0))
+                        )
+                        background = _get_background(im_frame, color)
+                        background_im = Image.new("P", im_frame.size, background)
+                        first_palette = im_frames[0].im.palette
+                        assert first_palette is not None
+                        background_im.putpalette(first_palette, first_palette.mode)
+                    bbox = _getbbox(background_im, im_frame)[1]
+                elif encoderinfo.get("optimize") and im_frame.mode != "1":
+                    if "transparency" not in encoderinfo:
+                        assert im_frame.palette is not None
+                        try:
+                            encoderinfo["transparency"] = (
+                                im_frame.palette._new_color_index(im_frame)
+                            )
+                        except ValueError:
+                            pass
+                    if "transparency" in encoderinfo:
+                        # When the delta is zero, fill the image with transparency
+                        diff_frame = im_frame.copy()
+                        fill = Image.new("P", delta.size, encoderinfo["transparency"])
+                        if delta.mode == "RGBA":
+                            r, g, b, a = delta.split()
+                            mask = ImageMath.lambda_eval(
+                                lambda args: args["convert"](
+                                    args["max"](
+                                        args["max"](
+                                            args["max"](args["r"], args["g"]), args["b"]
+                                        ),
+                                        args["a"],
+                                    )
+                                    * 255,
+                                    "1",
+                                ),
+                                r=r,
+                                g=g,
+                                b=b,
+                                a=a,
+                            )
+                        else:
+                            if delta.mode == "P":
+                                # Convert to L without considering palette
+                                delta_l = Image.new("L", delta.size)
+                                delta_l.putdata(delta.getdata())
+                                delta = delta_l
+                            mask = ImageMath.lambda_eval(
+                                lambda args: args["convert"](args["im"] * 255, "1"),
+                                im=delta,
+                            )
+                        diff_frame.paste(fill, mask=ImageOps.invert(mask))
+            else:
+                bbox = None
+            previous_im = im_frame
+            im_frames.append(_Frame(diff_frame or im_frame, bbox, encoderinfo))
+
+    if len(im_frames) == 1:
+        if "duration" in im.encoderinfo:
+            # Since multiple frames will not be written, use the combined duration
+            im.encoderinfo["duration"] = im_frames[0].encoderinfo["duration"]
+        return False
+
+    for frame_data in im_frames:
+        im_frame = frame_data.im
+        if not frame_data.bbox:
+            # global header
+            for s in _get_global_header(im_frame, frame_data.encoderinfo):
+                fp.write(s)
+            offset = (0, 0)
+        else:
+            # compress difference
+            if not palette:
+                frame_data.encoderinfo["include_color_table"] = True
+
+            im_frame = im_frame.crop(frame_data.bbox)
+            offset = frame_data.bbox[:2]
+        _write_frame_data(fp, im_frame, offset, frame_data.encoderinfo)
+    return True
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    _save(im, fp, filename, save_all=True)
+
+
+def _save(
+    im: Image.Image, fp: IO[bytes], filename: str | bytes, save_all: bool = False
+) -> None:
+    # header
+    if "palette" in im.encoderinfo or "palette" in im.info:
+        palette = im.encoderinfo.get("palette", im.info.get("palette"))
+    else:
+        palette = None
+        im.encoderinfo.setdefault("optimize", True)
+
+    if not save_all or not _write_multiple_frames(im, fp, palette):
+        _write_single_frame(im, fp, palette)
+
+    fp.write(b";")  # end of file
+
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+def get_interlace(im: Image.Image) -> int:
+    interlace = im.encoderinfo.get("interlace", 1)
+
+    # workaround for @PIL153
+    if min(im.size) < 16:
+        interlace = 0
+
+    return interlace
+
+
+def _write_local_header(
+    fp: IO[bytes], im: Image.Image, offset: tuple[int, int], flags: int
+) -> None:
+    try:
+        transparency = im.encoderinfo["transparency"]
+    except KeyError:
+        transparency = None
+
+    if "duration" in im.encoderinfo:
+        duration = int(im.encoderinfo["duration"] / 10)
+    else:
+        duration = 0
+
+    disposal = int(im.encoderinfo.get("disposal", 0))
+
+    if transparency is not None or duration != 0 or disposal:
+        packed_flag = 1 if transparency is not None else 0
+        packed_flag |= disposal << 2
+
+        fp.write(
+            b"!"
+            + o8(249)  # extension intro
+            + o8(4)  # length
+            + o8(packed_flag)  # packed fields
+            + o16(duration)  # duration
+            + o8(transparency or 0)  # transparency index
+            + o8(0)
+        )
+
+    include_color_table = im.encoderinfo.get("include_color_table")
+    if include_color_table:
+        palette_bytes = _get_palette_bytes(im)
+        color_table_size = _get_color_table_size(palette_bytes)
+        if color_table_size:
+            flags = flags | 128  # local color table flag
+            flags = flags | color_table_size
+
+    fp.write(
+        b","
+        + o16(offset[0])  # offset
+        + o16(offset[1])
+        + o16(im.size[0])  # size
+        + o16(im.size[1])
+        + o8(flags)  # flags
+    )
+    if include_color_table and color_table_size:
+        fp.write(_get_header_palette(palette_bytes))
+    fp.write(o8(8))  # bits
+
+
+def _save_netpbm(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    # Unused by default.
+    # To use, uncomment the register_save call at the end of the file.
+    #
+    # If you need real GIF compression and/or RGB quantization, you
+    # can use the external NETPBM/PBMPLUS utilities.  See comments
+    # below for information on how to enable this.
+    tempfile = im._dump()
+
+    try:
+        with open(filename, "wb") as f:
+            if im.mode != "RGB":
+                subprocess.check_call(
+                    ["ppmtogif", tempfile], stdout=f, stderr=subprocess.DEVNULL
+                )
+            else:
+                # Pipe ppmquant output into ppmtogif
+                # "ppmquant 256 %s | ppmtogif > %s" % (tempfile, filename)
+                quant_cmd = ["ppmquant", "256", tempfile]
+                togif_cmd = ["ppmtogif"]
+                quant_proc = subprocess.Popen(
+                    quant_cmd, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL
+                )
+                togif_proc = subprocess.Popen(
+                    togif_cmd,
+                    stdin=quant_proc.stdout,
+                    stdout=f,
+                    stderr=subprocess.DEVNULL,
+                )
+
+                # Allow ppmquant to receive SIGPIPE if ppmtogif exits
+                assert quant_proc.stdout is not None
+                quant_proc.stdout.close()
+
+                retcode = quant_proc.wait()
+                if retcode:
+                    raise subprocess.CalledProcessError(retcode, quant_cmd)
+
+                retcode = togif_proc.wait()
+                if retcode:
+                    raise subprocess.CalledProcessError(retcode, togif_cmd)
+    finally:
+        try:
+            os.unlink(tempfile)
+        except OSError:
+            pass
+
+
+# Force optimization so that we can test performance against
+# cases where it took lots of memory and time previously.
+_FORCE_OPTIMIZE = False
+
+
+def _get_optimize(im: Image.Image, info: dict[str, Any]) -> list[int] | None:
+    """
+    Palette optimization is a potentially expensive operation.
+
+    This function determines if the palette should be optimized using
+    some heuristics, then returns the list of palette entries in use.
+
+    :param im: Image object
+    :param info: encoderinfo
+    :returns: list of indexes of palette entries in use, or None
+    """
+    if im.mode in ("P", "L") and info and info.get("optimize"):
+        # Potentially expensive operation.
+
+        # The palette saves 3 bytes per color not used, but palette
+        # lengths are restricted to 3*(2**N) bytes. Max saving would
+        # be 768 -> 6 bytes if we went all the way down to 2 colors.
+        # * If we're over 128 colors, we can't save any space.
+        # * If there aren't any holes, it's not worth collapsing.
+        # * If we have a 'large' image, the palette is in the noise.
+
+        # create the new palette if not every color is used
+        optimise = _FORCE_OPTIMIZE or im.mode == "L"
+        if optimise or im.width * im.height < 512 * 512:
+            # check which colors are used
+            used_palette_colors = []
+            for i, count in enumerate(im.histogram()):
+                if count:
+                    used_palette_colors.append(i)
+
+            if optimise or max(used_palette_colors) >= len(used_palette_colors):
+                return used_palette_colors
+
+            assert im.palette is not None
+            num_palette_colors = len(im.palette.palette) // Image.getmodebands(
+                im.palette.mode
+            )
+            current_palette_size = 1 << (num_palette_colors - 1).bit_length()
+            if (
+                # check that the palette would become smaller when saved
+                len(used_palette_colors) <= current_palette_size // 2
+                # check that the palette is not already the smallest possible size
+                and current_palette_size > 2
+            ):
+                return used_palette_colors
+    return None
+
+
+def _get_color_table_size(palette_bytes: bytes) -> int:
+    # calculate the palette size for the header
+    if not palette_bytes:
+        return 0
+    elif len(palette_bytes) < 9:
+        return 1
+    else:
+        return math.ceil(math.log(len(palette_bytes) // 3, 2)) - 1
+
+
+def _get_header_palette(palette_bytes: bytes) -> bytes:
+    """
+    Returns the palette, null padded to the next power of 2 (*3) bytes
+    suitable for direct inclusion in the GIF header
+
+    :param palette_bytes: Unpadded palette bytes, in RGBRGB form
+    :returns: Null padded palette
+    """
+    color_table_size = _get_color_table_size(palette_bytes)
+
+    # add the missing amount of bytes
+    # the palette has to be 2<<n in size
+    actual_target_size_diff = (2 << color_table_size) - len(palette_bytes) // 3
+    if actual_target_size_diff > 0:
+        palette_bytes += o8(0) * 3 * actual_target_size_diff
+    return palette_bytes
+
+
+def _get_palette_bytes(im: Image.Image) -> bytes:
+    """
+    Gets the palette for inclusion in the gif header
+
+    :param im: Image object
+    :returns: Bytes, len<=768 suitable for inclusion in gif header
+    """
+    if not im.palette:
+        return b""
+
+    palette = bytes(im.palette.palette)
+    if im.palette.mode == "RGBA":
+        palette = b"".join(palette[i * 4 : i * 4 + 3] for i in range(len(palette) // 3))
+    return palette
+
+
+def _get_background(
+    im: Image.Image,
+    info_background: int | tuple[int, int, int] | tuple[int, int, int, int] | None,
+) -> int:
+    background = 0
+    if info_background:
+        if isinstance(info_background, tuple):
+            # WebPImagePlugin stores an RGBA value in info["background"]
+            # So it must be converted to the same format as GifImagePlugin's
+            # info["background"] - a global color table index
+            assert im.palette is not None
+            try:
+                background = im.palette.getcolor(info_background, im)
+            except ValueError as e:
+                if str(e) not in (
+                    # If all 256 colors are in use,
+                    # then there is no need for the background color
+                    "cannot allocate more than 256 colors",
+                    # Ignore non-opaque WebP background
+                    "cannot add non-opaque RGBA color to RGB palette",
+                ):
+                    raise
+        else:
+            background = info_background
+    return background
+
+
+def _get_global_header(im: Image.Image, info: dict[str, Any]) -> list[bytes]:
+    """Return a list of strings representing a GIF header"""
+
+    # Header Block
+    # https://www.matthewflickinger.com/lab/whatsinagif/bits_and_bytes.asp
+
+    version = b"87a"
+    if im.info.get("version") == b"89a" or (
+        info
+        and (
+            "transparency" in info
+            or info.get("loop") is not None
+            or info.get("duration")
+            or info.get("comment")
+        )
+    ):
+        version = b"89a"
+
+    background = _get_background(im, info.get("background"))
+
+    palette_bytes = _get_palette_bytes(im)
+    color_table_size = _get_color_table_size(palette_bytes)
+
+    header = [
+        b"GIF"  # signature
+        + version  # version
+        + o16(im.size[0])  # canvas width
+        + o16(im.size[1]),  # canvas height
+        # Logical Screen Descriptor
+        # size of global color table + global color table flag
+        o8(color_table_size + 128),  # packed fields
+        # background + reserved/aspect
+        o8(background) + o8(0),
+        # Global Color Table
+        _get_header_palette(palette_bytes),
+    ]
+    if info.get("loop") is not None:
+        header.append(
+            b"!"
+            + o8(255)  # extension intro
+            + o8(11)
+            + b"NETSCAPE2.0"
+            + o8(3)
+            + o8(1)
+            + o16(info["loop"])  # number of loops
+            + o8(0)
+        )
+    if info.get("comment"):
+        comment_block = b"!" + o8(254)  # extension intro
+
+        comment = info["comment"]
+        if isinstance(comment, str):
+            comment = comment.encode()
+        for i in range(0, len(comment), 255):
+            subblock = comment[i : i + 255]
+            comment_block += o8(len(subblock)) + subblock
+
+        comment_block += o8(0)
+        header.append(comment_block)
+    return header
+
+
+def _write_frame_data(
+    fp: IO[bytes],
+    im_frame: Image.Image,
+    offset: tuple[int, int],
+    params: dict[str, Any],
+) -> None:
+    try:
+        im_frame.encoderinfo = params
+
+        # local image header
+        _write_local_header(fp, im_frame, offset, 0)
+
+        ImageFile._save(
+            im_frame,
+            fp,
+            [ImageFile._Tile("gif", (0, 0) + im_frame.size, 0, RAWMODE[im_frame.mode])],
+        )
+
+        fp.write(b"\0")  # end of image data
+    finally:
+        del im_frame.encoderinfo
+
+
+# --------------------------------------------------------------------
+# Legacy GIF utilities
+
+
+def getheader(
+    im: Image.Image, palette: _Palette | None = None, info: dict[str, Any] | None = None
+) -> tuple[list[bytes], list[int] | None]:
+    """
+    Legacy Method to get Gif data from image.
+
+    Warning:: May modify image data.
+
+    :param im: Image object
+    :param palette: bytes object containing the source palette, or ....
+    :param info: encoderinfo
+    :returns: tuple of(list of header items, optimized palette)
+
+    """
+    if info is None:
+        info = {}
+
+    used_palette_colors = _get_optimize(im, info)
+
+    if "background" not in info and "background" in im.info:
+        info["background"] = im.info["background"]
+
+    im_mod = _normalize_palette(im, palette, info)
+    im.palette = im_mod.palette
+    im.im = im_mod.im
+    header = _get_global_header(im, info)
+
+    return header, used_palette_colors
+
+
+def getdata(
+    im: Image.Image, offset: tuple[int, int] = (0, 0), **params: Any
+) -> list[bytes]:
+    """
+    Legacy Method
+
+    Return a list of strings representing this image.
+    The first string is a local image header, the rest contains
+    encoded image data.
+
+    To specify duration, add the time in milliseconds,
+    e.g. ``getdata(im_frame, duration=1000)``
+
+    :param im: Image object
+    :param offset: Tuple of (x, y) pixels. Defaults to (0, 0)
+    :param \\**params: e.g. duration or other encoder info parameters
+    :returns: List of bytes containing GIF encoded frame data
+
+    """
+    from io import BytesIO
+
+    class Collector(BytesIO):
+        data = []
+
+        def write(self, data: Buffer) -> int:
+            self.data.append(data)
+            return len(data)
+
+    im.load()  # make sure raster data is available
+
+    fp = Collector()
+
+    _write_frame_data(fp, im, offset, params)
+
+    return fp.data
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(GifImageFile.format, GifImageFile, _accept)
+Image.register_save(GifImageFile.format, _save)
+Image.register_save_all(GifImageFile.format, _save_all)
+Image.register_extension(GifImageFile.format, ".gif")
+Image.register_mime(GifImageFile.format, "image/gif")
+
+#
+# Uncomment the following line if you wish to use NETPBM/PBMPLUS
+# instead of the built-in "uncompressed" GIF encoder
+
+# Image.register_save(GifImageFile.format, _save_netpbm)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GimpGradientFile.py b/backend/venv/lib/python3.12/site-packages/PIL/GimpGradientFile.py
new file mode 100644
index 00000000..220eac57
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GimpGradientFile.py
@@ -0,0 +1,149 @@
+#
+# Python Imaging Library
+# $Id$
+#
+# stuff to read (and render) GIMP gradient files
+#
+# History:
+#       97-08-23 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+
+"""
+Stuff to translate curve segments to palette values (derived from
+the corresponding code in GIMP, written by Federico Mena Quintero.
+See the GIMP distribution for more information.)
+"""
+from __future__ import annotations
+
+from math import log, pi, sin, sqrt
+from typing import IO, Callable
+
+from ._binary import o8
+
+EPSILON = 1e-10
+""""""  # Enable auto-doc for data member
+
+
+def linear(middle: float, pos: float) -> float:
+    if pos <= middle:
+        if middle < EPSILON:
+            return 0.0
+        else:
+            return 0.5 * pos / middle
+    else:
+        pos = pos - middle
+        middle = 1.0 - middle
+        if middle < EPSILON:
+            return 1.0
+        else:
+            return 0.5 + 0.5 * pos / middle
+
+
+def curved(middle: float, pos: float) -> float:
+    return pos ** (log(0.5) / log(max(middle, EPSILON)))
+
+
+def sine(middle: float, pos: float) -> float:
+    return (sin((-pi / 2.0) + pi * linear(middle, pos)) + 1.0) / 2.0
+
+
+def sphere_increasing(middle: float, pos: float) -> float:
+    return sqrt(1.0 - (linear(middle, pos) - 1.0) ** 2)
+
+
+def sphere_decreasing(middle: float, pos: float) -> float:
+    return 1.0 - sqrt(1.0 - linear(middle, pos) ** 2)
+
+
+SEGMENTS = [linear, curved, sine, sphere_increasing, sphere_decreasing]
+""""""  # Enable auto-doc for data member
+
+
+class GradientFile:
+    gradient: (
+        list[
+            tuple[
+                float,
+                float,
+                float,
+                list[float],
+                list[float],
+                Callable[[float, float], float],
+            ]
+        ]
+        | None
+    ) = None
+
+    def getpalette(self, entries: int = 256) -> tuple[bytes, str]:
+        assert self.gradient is not None
+        palette = []
+
+        ix = 0
+        x0, x1, xm, rgb0, rgb1, segment = self.gradient[ix]
+
+        for i in range(entries):
+            x = i / (entries - 1)
+
+            while x1 < x:
+                ix += 1
+                x0, x1, xm, rgb0, rgb1, segment = self.gradient[ix]
+
+            w = x1 - x0
+
+            if w < EPSILON:
+                scale = segment(0.5, 0.5)
+            else:
+                scale = segment((xm - x0) / w, (x - x0) / w)
+
+            # expand to RGBA
+            r = o8(int(255 * ((rgb1[0] - rgb0[0]) * scale + rgb0[0]) + 0.5))
+            g = o8(int(255 * ((rgb1[1] - rgb0[1]) * scale + rgb0[1]) + 0.5))
+            b = o8(int(255 * ((rgb1[2] - rgb0[2]) * scale + rgb0[2]) + 0.5))
+            a = o8(int(255 * ((rgb1[3] - rgb0[3]) * scale + rgb0[3]) + 0.5))
+
+            # add to palette
+            palette.append(r + g + b + a)
+
+        return b"".join(palette), "RGBA"
+
+
+class GimpGradientFile(GradientFile):
+    """File handler for GIMP's gradient format."""
+
+    def __init__(self, fp: IO[bytes]) -> None:
+        if fp.readline()[:13] != b"GIMP Gradient":
+            msg = "not a GIMP gradient file"
+            raise SyntaxError(msg)
+
+        line = fp.readline()
+
+        # GIMP 1.2 gradient files don't contain a name, but GIMP 1.3 files do
+        if line.startswith(b"Name: "):
+            line = fp.readline().strip()
+
+        count = int(line)
+
+        self.gradient = []
+
+        for i in range(count):
+            s = fp.readline().split()
+            w = [float(x) for x in s[:11]]
+
+            x0, x1 = w[0], w[2]
+            xm = w[1]
+            rgb0 = w[3:7]
+            rgb1 = w[7:11]
+
+            segment = SEGMENTS[int(s[11])]
+            cspace = int(s[12])
+
+            if cspace != 0:
+                msg = "cannot handle HSV colour space"
+                raise OSError(msg)
+
+            self.gradient.append((x0, x1, xm, rgb0, rgb1, segment))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GimpPaletteFile.py b/backend/venv/lib/python3.12/site-packages/PIL/GimpPaletteFile.py
new file mode 100644
index 00000000..4cad0ebe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GimpPaletteFile.py
@@ -0,0 +1,58 @@
+#
+# Python Imaging Library
+# $Id$
+#
+# stuff to read GIMP palette files
+#
+# History:
+# 1997-08-23 fl     Created
+# 2004-09-07 fl     Support GIMP 2.0 palette files.
+#
+# Copyright (c) Secret Labs AB 1997-2004.  All rights reserved.
+# Copyright (c) Fredrik Lundh 1997-2004.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import re
+from typing import IO
+
+from ._binary import o8
+
+
+class GimpPaletteFile:
+    """File handler for GIMP's palette format."""
+
+    rawmode = "RGB"
+
+    def __init__(self, fp: IO[bytes]) -> None:
+        palette = [o8(i) * 3 for i in range(256)]
+
+        if fp.readline()[:12] != b"GIMP Palette":
+            msg = "not a GIMP palette file"
+            raise SyntaxError(msg)
+
+        for i in range(256):
+            s = fp.readline()
+            if not s:
+                break
+
+            # skip fields and comment lines
+            if re.match(rb"\w+:|#", s):
+                continue
+            if len(s) > 100:
+                msg = "bad palette file"
+                raise SyntaxError(msg)
+
+            v = tuple(map(int, s.split()[:3]))
+            if len(v) != 3:
+                msg = "bad palette entry"
+                raise ValueError(msg)
+
+            palette[i] = o8(v[0]) + o8(v[1]) + o8(v[2])
+
+        self.palette = b"".join(palette)
+
+    def getpalette(self) -> tuple[bytes, str]:
+        return self.palette, self.rawmode
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/GribStubImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/GribStubImagePlugin.py
new file mode 100644
index 00000000..e9aa084b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/GribStubImagePlugin.py
@@ -0,0 +1,76 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# GRIB stub adapter
+#
+# Copyright (c) 1996-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+
+_handler = None
+
+
+def register_handler(handler: ImageFile.StubHandler | None) -> None:
+    """
+    Install application-specific GRIB image handler.
+
+    :param handler: Handler object.
+    """
+    global _handler
+    _handler = handler
+
+
+# --------------------------------------------------------------------
+# Image adapter
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"GRIB" and prefix[7] == 1
+
+
+class GribStubImageFile(ImageFile.StubImageFile):
+    format = "GRIB"
+    format_description = "GRIB"
+
+    def _open(self) -> None:
+        offset = self.fp.tell()
+
+        if not _accept(self.fp.read(8)):
+            msg = "Not a GRIB file"
+            raise SyntaxError(msg)
+
+        self.fp.seek(offset)
+
+        # make something up
+        self._mode = "F"
+        self._size = 1, 1
+
+        loader = self._load()
+        if loader:
+            loader.open(self)
+
+    def _load(self) -> ImageFile.StubHandler | None:
+        return _handler
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if _handler is None or not hasattr(_handler, "save"):
+        msg = "GRIB save handler not installed"
+        raise OSError(msg)
+    _handler.save(im, fp, filename)
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(GribStubImageFile.format, GribStubImageFile, _accept)
+Image.register_save(GribStubImageFile.format, _save)
+
+Image.register_extension(GribStubImageFile.format, ".grib")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/Hdf5StubImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/Hdf5StubImagePlugin.py
new file mode 100644
index 00000000..cc9e73de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/Hdf5StubImagePlugin.py
@@ -0,0 +1,76 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# HDF5 stub adapter
+#
+# Copyright (c) 2000-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+
+_handler = None
+
+
+def register_handler(handler: ImageFile.StubHandler | None) -> None:
+    """
+    Install application-specific HDF5 image handler.
+
+    :param handler: Handler object.
+    """
+    global _handler
+    _handler = handler
+
+
+# --------------------------------------------------------------------
+# Image adapter
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:8] == b"\x89HDF\r\n\x1a\n"
+
+
+class HDF5StubImageFile(ImageFile.StubImageFile):
+    format = "HDF5"
+    format_description = "HDF5"
+
+    def _open(self) -> None:
+        offset = self.fp.tell()
+
+        if not _accept(self.fp.read(8)):
+            msg = "Not an HDF file"
+            raise SyntaxError(msg)
+
+        self.fp.seek(offset)
+
+        # make something up
+        self._mode = "F"
+        self._size = 1, 1
+
+        loader = self._load()
+        if loader:
+            loader.open(self)
+
+    def _load(self) -> ImageFile.StubHandler | None:
+        return _handler
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if _handler is None or not hasattr(_handler, "save"):
+        msg = "HDF5 save handler not installed"
+        raise OSError(msg)
+    _handler.save(im, fp, filename)
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(HDF5StubImageFile.format, HDF5StubImageFile, _accept)
+Image.register_save(HDF5StubImageFile.format, _save)
+
+Image.register_extensions(HDF5StubImageFile.format, [".h5", ".hdf"])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/IcnsImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/IcnsImagePlugin.py
new file mode 100644
index 00000000..9757b2b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/IcnsImagePlugin.py
@@ -0,0 +1,412 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# macOS icns file decoder, based on icns.py by Bob Ippolito.
+#
+# history:
+# 2004-10-09 fl   Turned into a PIL plugin; removed 2.3 dependencies.
+# 2020-04-04      Allow saving on all operating systems.
+#
+# Copyright (c) 2004 by Bob Ippolito.
+# Copyright (c) 2004 by Secret Labs.
+# Copyright (c) 2004 by Fredrik Lundh.
+# Copyright (c) 2014 by Alastair Houghton.
+# Copyright (c) 2020 by Pan Jing.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import os
+import struct
+import sys
+from typing import IO
+
+from . import Image, ImageFile, PngImagePlugin, features
+from ._deprecate import deprecate
+
+enable_jpeg2k = features.check_codec("jpg_2000")
+if enable_jpeg2k:
+    from . import Jpeg2KImagePlugin
+
+MAGIC = b"icns"
+HEADERSIZE = 8
+
+
+def nextheader(fobj: IO[bytes]) -> tuple[bytes, int]:
+    return struct.unpack(">4sI", fobj.read(HEADERSIZE))
+
+
+def read_32t(
+    fobj: IO[bytes], start_length: tuple[int, int], size: tuple[int, int, int]
+) -> dict[str, Image.Image]:
+    # The 128x128 icon seems to have an extra header for some reason.
+    (start, length) = start_length
+    fobj.seek(start)
+    sig = fobj.read(4)
+    if sig != b"\x00\x00\x00\x00":
+        msg = "Unknown signature, expecting 0x00000000"
+        raise SyntaxError(msg)
+    return read_32(fobj, (start + 4, length - 4), size)
+
+
+def read_32(
+    fobj: IO[bytes], start_length: tuple[int, int], size: tuple[int, int, int]
+) -> dict[str, Image.Image]:
+    """
+    Read a 32bit RGB icon resource.  Seems to be either uncompressed or
+    an RLE packbits-like scheme.
+    """
+    (start, length) = start_length
+    fobj.seek(start)
+    pixel_size = (size[0] * size[2], size[1] * size[2])
+    sizesq = pixel_size[0] * pixel_size[1]
+    if length == sizesq * 3:
+        # uncompressed ("RGBRGBGB")
+        indata = fobj.read(length)
+        im = Image.frombuffer("RGB", pixel_size, indata, "raw", "RGB", 0, 1)
+    else:
+        # decode image
+        im = Image.new("RGB", pixel_size, None)
+        for band_ix in range(3):
+            data = []
+            bytesleft = sizesq
+            while bytesleft > 0:
+                byte = fobj.read(1)
+                if not byte:
+                    break
+                byte_int = byte[0]
+                if byte_int & 0x80:
+                    blocksize = byte_int - 125
+                    byte = fobj.read(1)
+                    for i in range(blocksize):
+                        data.append(byte)
+                else:
+                    blocksize = byte_int + 1
+                    data.append(fobj.read(blocksize))
+                bytesleft -= blocksize
+                if bytesleft <= 0:
+                    break
+            if bytesleft != 0:
+                msg = f"Error reading channel [{repr(bytesleft)} left]"
+                raise SyntaxError(msg)
+            band = Image.frombuffer("L", pixel_size, b"".join(data), "raw", "L", 0, 1)
+            im.im.putband(band.im, band_ix)
+    return {"RGB": im}
+
+
+def read_mk(
+    fobj: IO[bytes], start_length: tuple[int, int], size: tuple[int, int, int]
+) -> dict[str, Image.Image]:
+    # Alpha masks seem to be uncompressed
+    start = start_length[0]
+    fobj.seek(start)
+    pixel_size = (size[0] * size[2], size[1] * size[2])
+    sizesq = pixel_size[0] * pixel_size[1]
+    band = Image.frombuffer("L", pixel_size, fobj.read(sizesq), "raw", "L", 0, 1)
+    return {"A": band}
+
+
+def read_png_or_jpeg2000(
+    fobj: IO[bytes], start_length: tuple[int, int], size: tuple[int, int, int]
+) -> dict[str, Image.Image]:
+    (start, length) = start_length
+    fobj.seek(start)
+    sig = fobj.read(12)
+
+    im: Image.Image
+    if sig[:8] == b"\x89PNG\x0d\x0a\x1a\x0a":
+        fobj.seek(start)
+        im = PngImagePlugin.PngImageFile(fobj)
+        Image._decompression_bomb_check(im.size)
+        return {"RGBA": im}
+    elif (
+        sig[:4] == b"\xff\x4f\xff\x51"
+        or sig[:4] == b"\x0d\x0a\x87\x0a"
+        or sig == b"\x00\x00\x00\x0cjP  \x0d\x0a\x87\x0a"
+    ):
+        if not enable_jpeg2k:
+            msg = (
+                "Unsupported icon subimage format (rebuild PIL "
+                "with JPEG 2000 support to fix this)"
+            )
+            raise ValueError(msg)
+        # j2k, jpc or j2c
+        fobj.seek(start)
+        jp2kstream = fobj.read(length)
+        f = io.BytesIO(jp2kstream)
+        im = Jpeg2KImagePlugin.Jpeg2KImageFile(f)
+        Image._decompression_bomb_check(im.size)
+        if im.mode != "RGBA":
+            im = im.convert("RGBA")
+        return {"RGBA": im}
+    else:
+        msg = "Unsupported icon subimage format"
+        raise ValueError(msg)
+
+
+class IcnsFile:
+    SIZES = {
+        (512, 512, 2): [(b"ic10", read_png_or_jpeg2000)],
+        (512, 512, 1): [(b"ic09", read_png_or_jpeg2000)],
+        (256, 256, 2): [(b"ic14", read_png_or_jpeg2000)],
+        (256, 256, 1): [(b"ic08", read_png_or_jpeg2000)],
+        (128, 128, 2): [(b"ic13", read_png_or_jpeg2000)],
+        (128, 128, 1): [
+            (b"ic07", read_png_or_jpeg2000),
+            (b"it32", read_32t),
+            (b"t8mk", read_mk),
+        ],
+        (64, 64, 1): [(b"icp6", read_png_or_jpeg2000)],
+        (32, 32, 2): [(b"ic12", read_png_or_jpeg2000)],
+        (48, 48, 1): [(b"ih32", read_32), (b"h8mk", read_mk)],
+        (32, 32, 1): [
+            (b"icp5", read_png_or_jpeg2000),
+            (b"il32", read_32),
+            (b"l8mk", read_mk),
+        ],
+        (16, 16, 2): [(b"ic11", read_png_or_jpeg2000)],
+        (16, 16, 1): [
+            (b"icp4", read_png_or_jpeg2000),
+            (b"is32", read_32),
+            (b"s8mk", read_mk),
+        ],
+    }
+
+    def __init__(self, fobj: IO[bytes]) -> None:
+        """
+        fobj is a file-like object as an icns resource
+        """
+        # signature : (start, length)
+        self.dct = {}
+        self.fobj = fobj
+        sig, filesize = nextheader(fobj)
+        if not _accept(sig):
+            msg = "not an icns file"
+            raise SyntaxError(msg)
+        i = HEADERSIZE
+        while i < filesize:
+            sig, blocksize = nextheader(fobj)
+            if blocksize <= 0:
+                msg = "invalid block header"
+                raise SyntaxError(msg)
+            i += HEADERSIZE
+            blocksize -= HEADERSIZE
+            self.dct[sig] = (i, blocksize)
+            fobj.seek(blocksize, io.SEEK_CUR)
+            i += blocksize
+
+    def itersizes(self) -> list[tuple[int, int, int]]:
+        sizes = []
+        for size, fmts in self.SIZES.items():
+            for fmt, reader in fmts:
+                if fmt in self.dct:
+                    sizes.append(size)
+                    break
+        return sizes
+
+    def bestsize(self) -> tuple[int, int, int]:
+        sizes = self.itersizes()
+        if not sizes:
+            msg = "No 32bit icon resources found"
+            raise SyntaxError(msg)
+        return max(sizes)
+
+    def dataforsize(self, size: tuple[int, int, int]) -> dict[str, Image.Image]:
+        """
+        Get an icon resource as {channel: array}.  Note that
+        the arrays are bottom-up like windows bitmaps and will likely
+        need to be flipped or transposed in some way.
+        """
+        dct = {}
+        for code, reader in self.SIZES[size]:
+            desc = self.dct.get(code)
+            if desc is not None:
+                dct.update(reader(self.fobj, desc, size))
+        return dct
+
+    def getimage(
+        self, size: tuple[int, int] | tuple[int, int, int] | None = None
+    ) -> Image.Image:
+        if size is None:
+            size = self.bestsize()
+        elif len(size) == 2:
+            size = (size[0], size[1], 1)
+        channels = self.dataforsize(size)
+
+        im = channels.get("RGBA")
+        if im:
+            return im
+
+        im = channels["RGB"].copy()
+        try:
+            im.putalpha(channels["A"])
+        except KeyError:
+            pass
+        return im
+
+
+##
+# Image plugin for Mac OS icons.
+
+
+class IcnsImageFile(ImageFile.ImageFile):
+    """
+    PIL image support for Mac OS .icns files.
+    Chooses the best resolution, but will possibly load
+    a different size image if you mutate the size attribute
+    before calling 'load'.
+
+    The info dictionary has a key 'sizes' that is a list
+    of sizes that the icns file has.
+    """
+
+    format = "ICNS"
+    format_description = "Mac OS icns resource"
+
+    def _open(self) -> None:
+        self.icns = IcnsFile(self.fp)
+        self._mode = "RGBA"
+        self.info["sizes"] = self.icns.itersizes()
+        self.best_size = self.icns.bestsize()
+        self.size = (
+            self.best_size[0] * self.best_size[2],
+            self.best_size[1] * self.best_size[2],
+        )
+
+    @property  # type: ignore[override]
+    def size(self) -> tuple[int, int] | tuple[int, int, int]:
+        return self._size
+
+    @size.setter
+    def size(self, value: tuple[int, int] | tuple[int, int, int]) -> None:
+        if len(value) == 3:
+            deprecate("Setting size to (width, height, scale)", 12, "load(scale)")
+            if value in self.info["sizes"]:
+                self._size = value  # type: ignore[assignment]
+                return
+        else:
+            # Check that a matching size exists,
+            # or that there is a scale that would create a size that matches
+            for size in self.info["sizes"]:
+                simple_size = size[0] * size[2], size[1] * size[2]
+                scale = simple_size[0] // value[0]
+                if simple_size[1] / value[1] == scale:
+                    self._size = value
+                    return
+        msg = "This is not one of the allowed sizes of this image"
+        raise ValueError(msg)
+
+    def load(self, scale: int | None = None) -> Image.core.PixelAccess | None:
+        if scale is not None or len(self.size) == 3:
+            if scale is None and len(self.size) == 3:
+                scale = self.size[2]
+            assert scale is not None
+            width, height = self.size[:2]
+            self.size = width * scale, height * scale
+            self.best_size = width, height, scale
+
+        px = Image.Image.load(self)
+        if self._im is not None and self.im.size == self.size:
+            # Already loaded
+            return px
+        self.load_prepare()
+        # This is likely NOT the best way to do it, but whatever.
+        im = self.icns.getimage(self.best_size)
+
+        # If this is a PNG or JPEG 2000, it won't be loaded yet
+        px = im.load()
+
+        self.im = im.im
+        self._mode = im.mode
+        self.size = im.size
+
+        return px
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    """
+    Saves the image as a series of PNG files,
+    that are then combined into a .icns file.
+    """
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+    sizes = {
+        b"ic07": 128,
+        b"ic08": 256,
+        b"ic09": 512,
+        b"ic10": 1024,
+        b"ic11": 32,
+        b"ic12": 64,
+        b"ic13": 256,
+        b"ic14": 512,
+    }
+    provided_images = {im.width: im for im in im.encoderinfo.get("append_images", [])}
+    size_streams = {}
+    for size in set(sizes.values()):
+        image = (
+            provided_images[size]
+            if size in provided_images
+            else im.resize((size, size))
+        )
+
+        temp = io.BytesIO()
+        image.save(temp, "png")
+        size_streams[size] = temp.getvalue()
+
+    entries = []
+    for type, size in sizes.items():
+        stream = size_streams[size]
+        entries.append((type, HEADERSIZE + len(stream), stream))
+
+    # Header
+    fp.write(MAGIC)
+    file_length = HEADERSIZE  # Header
+    file_length += HEADERSIZE + 8 * len(entries)  # TOC
+    file_length += sum(entry[1] for entry in entries)
+    fp.write(struct.pack(">i", file_length))
+
+    # TOC
+    fp.write(b"TOC ")
+    fp.write(struct.pack(">i", HEADERSIZE + len(entries) * HEADERSIZE))
+    for entry in entries:
+        fp.write(entry[0])
+        fp.write(struct.pack(">i", entry[1]))
+
+    # Data
+    for entry in entries:
+        fp.write(entry[0])
+        fp.write(struct.pack(">i", entry[1]))
+        fp.write(entry[2])
+
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == MAGIC
+
+
+Image.register_open(IcnsImageFile.format, IcnsImageFile, _accept)
+Image.register_extension(IcnsImageFile.format, ".icns")
+
+Image.register_save(IcnsImageFile.format, _save)
+Image.register_mime(IcnsImageFile.format, "image/icns")
+
+if __name__ == "__main__":
+    if len(sys.argv) < 2:
+        print("Syntax: python3 IcnsImagePlugin.py [file]")
+        sys.exit()
+
+    with open(sys.argv[1], "rb") as fp:
+        imf = IcnsImageFile(fp)
+        for size in imf.info["sizes"]:
+            width, height, scale = imf.size = size
+            imf.save(f"out-{width}-{height}-{scale}.png")
+        with Image.open(sys.argv[1]) as im:
+            im.save("out.png")
+        if sys.platform == "windows":
+            os.startfile("out.png")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/IcoImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/IcoImagePlugin.py
new file mode 100644
index 00000000..e879f180
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/IcoImagePlugin.py
@@ -0,0 +1,381 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Windows Icon support for PIL
+#
+# History:
+#       96-05-27 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+
+# This plugin is a refactored version of Win32IconImagePlugin by Bryan Davis
+# <casadebender@gmail.com>.
+# https://code.google.com/archive/p/casadebender/wikis/Win32IconImagePlugin.wiki
+#
+# Icon format references:
+#   * https://en.wikipedia.org/wiki/ICO_(file_format)
+#   * https://msdn.microsoft.com/en-us/library/ms997538.aspx
+from __future__ import annotations
+
+import warnings
+from io import BytesIO
+from math import ceil, log
+from typing import IO, NamedTuple
+
+from . import BmpImagePlugin, Image, ImageFile, PngImagePlugin
+from ._binary import i16le as i16
+from ._binary import i32le as i32
+from ._binary import o8
+from ._binary import o16le as o16
+from ._binary import o32le as o32
+
+#
+# --------------------------------------------------------------------
+
+_MAGIC = b"\0\0\1\0"
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    fp.write(_MAGIC)  # (2+2)
+    bmp = im.encoderinfo.get("bitmap_format") == "bmp"
+    sizes = im.encoderinfo.get(
+        "sizes",
+        [(16, 16), (24, 24), (32, 32), (48, 48), (64, 64), (128, 128), (256, 256)],
+    )
+    frames = []
+    provided_ims = [im] + im.encoderinfo.get("append_images", [])
+    width, height = im.size
+    for size in sorted(set(sizes)):
+        if size[0] > width or size[1] > height or size[0] > 256 or size[1] > 256:
+            continue
+
+        for provided_im in provided_ims:
+            if provided_im.size != size:
+                continue
+            frames.append(provided_im)
+            if bmp:
+                bits = BmpImagePlugin.SAVE[provided_im.mode][1]
+                bits_used = [bits]
+                for other_im in provided_ims:
+                    if other_im.size != size:
+                        continue
+                    bits = BmpImagePlugin.SAVE[other_im.mode][1]
+                    if bits not in bits_used:
+                        # Another image has been supplied for this size
+                        # with a different bit depth
+                        frames.append(other_im)
+                        bits_used.append(bits)
+            break
+        else:
+            # TODO: invent a more convenient method for proportional scalings
+            frame = provided_im.copy()
+            frame.thumbnail(size, Image.Resampling.LANCZOS, reducing_gap=None)
+            frames.append(frame)
+    fp.write(o16(len(frames)))  # idCount(2)
+    offset = fp.tell() + len(frames) * 16
+    for frame in frames:
+        width, height = frame.size
+        # 0 means 256
+        fp.write(o8(width if width < 256 else 0))  # bWidth(1)
+        fp.write(o8(height if height < 256 else 0))  # bHeight(1)
+
+        bits, colors = BmpImagePlugin.SAVE[frame.mode][1:] if bmp else (32, 0)
+        fp.write(o8(colors))  # bColorCount(1)
+        fp.write(b"\0")  # bReserved(1)
+        fp.write(b"\0\0")  # wPlanes(2)
+        fp.write(o16(bits))  # wBitCount(2)
+
+        image_io = BytesIO()
+        if bmp:
+            frame.save(image_io, "dib")
+
+            if bits != 32:
+                and_mask = Image.new("1", size)
+                ImageFile._save(
+                    and_mask,
+                    image_io,
+                    [ImageFile._Tile("raw", (0, 0) + size, 0, ("1", 0, -1))],
+                )
+        else:
+            frame.save(image_io, "png")
+        image_io.seek(0)
+        image_bytes = image_io.read()
+        if bmp:
+            image_bytes = image_bytes[:8] + o32(height * 2) + image_bytes[12:]
+        bytes_len = len(image_bytes)
+        fp.write(o32(bytes_len))  # dwBytesInRes(4)
+        fp.write(o32(offset))  # dwImageOffset(4)
+        current = fp.tell()
+        fp.seek(offset)
+        fp.write(image_bytes)
+        offset = offset + bytes_len
+        fp.seek(current)
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == _MAGIC
+
+
+class IconHeader(NamedTuple):
+    width: int
+    height: int
+    nb_color: int
+    reserved: int
+    planes: int
+    bpp: int
+    size: int
+    offset: int
+    dim: tuple[int, int]
+    square: int
+    color_depth: int
+
+
+class IcoFile:
+    def __init__(self, buf: IO[bytes]) -> None:
+        """
+        Parse image from file-like object containing ico file data
+        """
+
+        # check magic
+        s = buf.read(6)
+        if not _accept(s):
+            msg = "not an ICO file"
+            raise SyntaxError(msg)
+
+        self.buf = buf
+        self.entry = []
+
+        # Number of items in file
+        self.nb_items = i16(s, 4)
+
+        # Get headers for each item
+        for i in range(self.nb_items):
+            s = buf.read(16)
+
+            # See Wikipedia
+            width = s[0] or 256
+            height = s[1] or 256
+
+            # No. of colors in image (0 if >=8bpp)
+            nb_color = s[2]
+            bpp = i16(s, 6)
+            icon_header = IconHeader(
+                width=width,
+                height=height,
+                nb_color=nb_color,
+                reserved=s[3],
+                planes=i16(s, 4),
+                bpp=i16(s, 6),
+                size=i32(s, 8),
+                offset=i32(s, 12),
+                dim=(width, height),
+                square=width * height,
+                # See Wikipedia notes about color depth.
+                # We need this just to differ images with equal sizes
+                color_depth=bpp or (nb_color != 0 and ceil(log(nb_color, 2))) or 256,
+            )
+
+            self.entry.append(icon_header)
+
+        self.entry = sorted(self.entry, key=lambda x: x.color_depth)
+        # ICO images are usually squares
+        self.entry = sorted(self.entry, key=lambda x: x.square, reverse=True)
+
+    def sizes(self) -> set[tuple[int, int]]:
+        """
+        Get a set of all available icon sizes and color depths.
+        """
+        return {(h.width, h.height) for h in self.entry}
+
+    def getentryindex(self, size: tuple[int, int], bpp: int | bool = False) -> int:
+        for i, h in enumerate(self.entry):
+            if size == h.dim and (bpp is False or bpp == h.color_depth):
+                return i
+        return 0
+
+    def getimage(self, size: tuple[int, int], bpp: int | bool = False) -> Image.Image:
+        """
+        Get an image from the icon
+        """
+        return self.frame(self.getentryindex(size, bpp))
+
+    def frame(self, idx: int) -> Image.Image:
+        """
+        Get an image from frame idx
+        """
+
+        header = self.entry[idx]
+
+        self.buf.seek(header.offset)
+        data = self.buf.read(8)
+        self.buf.seek(header.offset)
+
+        im: Image.Image
+        if data[:8] == PngImagePlugin._MAGIC:
+            # png frame
+            im = PngImagePlugin.PngImageFile(self.buf)
+            Image._decompression_bomb_check(im.size)
+        else:
+            # XOR + AND mask bmp frame
+            im = BmpImagePlugin.DibImageFile(self.buf)
+            Image._decompression_bomb_check(im.size)
+
+            # change tile dimension to only encompass XOR image
+            im._size = (im.size[0], int(im.size[1] / 2))
+            d, e, o, a = im.tile[0]
+            im.tile[0] = ImageFile._Tile(d, (0, 0) + im.size, o, a)
+
+            # figure out where AND mask image starts
+            if header.bpp == 32:
+                # 32-bit color depth icon image allows semitransparent areas
+                # PIL's DIB format ignores transparency bits, recover them.
+                # The DIB is packed in BGRX byte order where X is the alpha
+                # channel.
+
+                # Back up to start of bmp data
+                self.buf.seek(o)
+                # extract every 4th byte (eg. 3,7,11,15,...)
+                alpha_bytes = self.buf.read(im.size[0] * im.size[1] * 4)[3::4]
+
+                # convert to an 8bpp grayscale image
+                try:
+                    mask = Image.frombuffer(
+                        "L",  # 8bpp
+                        im.size,  # (w, h)
+                        alpha_bytes,  # source chars
+                        "raw",  # raw decoder
+                        ("L", 0, -1),  # 8bpp inverted, unpadded, reversed
+                    )
+                except ValueError:
+                    if ImageFile.LOAD_TRUNCATED_IMAGES:
+                        mask = None
+                    else:
+                        raise
+            else:
+                # get AND image from end of bitmap
+                w = im.size[0]
+                if (w % 32) > 0:
+                    # bitmap row data is aligned to word boundaries
+                    w += 32 - (im.size[0] % 32)
+
+                # the total mask data is
+                # padded row size * height / bits per char
+
+                total_bytes = int((w * im.size[1]) / 8)
+                and_mask_offset = header.offset + header.size - total_bytes
+
+                self.buf.seek(and_mask_offset)
+                mask_data = self.buf.read(total_bytes)
+
+                # convert raw data to image
+                try:
+                    mask = Image.frombuffer(
+                        "1",  # 1 bpp
+                        im.size,  # (w, h)
+                        mask_data,  # source chars
+                        "raw",  # raw decoder
+                        ("1;I", int(w / 8), -1),  # 1bpp inverted, padded, reversed
+                    )
+                except ValueError:
+                    if ImageFile.LOAD_TRUNCATED_IMAGES:
+                        mask = None
+                    else:
+                        raise
+
+                # now we have two images, im is XOR image and mask is AND image
+
+            # apply mask image as alpha channel
+            if mask:
+                im = im.convert("RGBA")
+                im.putalpha(mask)
+
+        return im
+
+
+##
+# Image plugin for Windows Icon files.
+
+
+class IcoImageFile(ImageFile.ImageFile):
+    """
+    PIL read-only image support for Microsoft Windows .ico files.
+
+    By default the largest resolution image in the file will be loaded. This
+    can be changed by altering the 'size' attribute before calling 'load'.
+
+    The info dictionary has a key 'sizes' that is a list of the sizes available
+    in the icon file.
+
+    Handles classic, XP and Vista icon formats.
+
+    When saving, PNG compression is used. Support for this was only added in
+    Windows Vista. If you are unable to view the icon in Windows, convert the
+    image to "RGBA" mode before saving.
+
+    This plugin is a refactored version of Win32IconImagePlugin by Bryan Davis
+    <casadebender@gmail.com>.
+    https://code.google.com/archive/p/casadebender/wikis/Win32IconImagePlugin.wiki
+    """
+
+    format = "ICO"
+    format_description = "Windows Icon"
+
+    def _open(self) -> None:
+        self.ico = IcoFile(self.fp)
+        self.info["sizes"] = self.ico.sizes()
+        self.size = self.ico.entry[0].dim
+        self.load()
+
+    @property
+    def size(self) -> tuple[int, int]:
+        return self._size
+
+    @size.setter
+    def size(self, value: tuple[int, int]) -> None:
+        if value not in self.info["sizes"]:
+            msg = "This is not one of the allowed sizes of this image"
+            raise ValueError(msg)
+        self._size = value
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self._im is not None and self.im.size == self.size:
+            # Already loaded
+            return Image.Image.load(self)
+        im = self.ico.getimage(self.size)
+        # if tile is PNG, it won't really be loaded yet
+        im.load()
+        self.im = im.im
+        self._mode = im.mode
+        if im.palette:
+            self.palette = im.palette
+        if im.size != self.size:
+            warnings.warn("Image was not the expected size")
+
+            index = self.ico.getentryindex(self.size)
+            sizes = list(self.info["sizes"])
+            sizes[index] = im.size
+            self.info["sizes"] = set(sizes)
+
+            self.size = im.size
+        return None
+
+    def load_seek(self, pos: int) -> None:
+        # Flag the ImageFile.Parser so that it
+        # just does all the decode at the end.
+        pass
+
+
+#
+# --------------------------------------------------------------------
+
+
+Image.register_open(IcoImageFile.format, IcoImageFile, _accept)
+Image.register_save(IcoImageFile.format, _save)
+Image.register_extension(IcoImageFile.format, ".ico")
+
+Image.register_mime(IcoImageFile.format, "image/x-icon")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/ImImagePlugin.py
new file mode 100644
index 00000000..b4215a0b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImImagePlugin.py
@@ -0,0 +1,386 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# IFUNC IM file handling for PIL
+#
+# history:
+# 1995-09-01 fl   Created.
+# 1997-01-03 fl   Save palette images
+# 1997-01-08 fl   Added sequence support
+# 1997-01-23 fl   Added P and RGB save support
+# 1997-05-31 fl   Read floating point images
+# 1997-06-22 fl   Save floating point images
+# 1997-08-27 fl   Read and save 1-bit images
+# 1998-06-25 fl   Added support for RGB+LUT images
+# 1998-07-02 fl   Added support for YCC images
+# 1998-07-15 fl   Renamed offset attribute to avoid name clash
+# 1998-12-29 fl   Added I;16 support
+# 2001-02-17 fl   Use 're' instead of 'regex' (Python 2.1) (0.7)
+# 2003-09-26 fl   Added LA/PA support
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1995-2001 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+import re
+from typing import IO, Any
+
+from . import Image, ImageFile, ImagePalette
+
+# --------------------------------------------------------------------
+# Standard tags
+
+COMMENT = "Comment"
+DATE = "Date"
+EQUIPMENT = "Digitalization equipment"
+FRAMES = "File size (no of images)"
+LUT = "Lut"
+NAME = "Name"
+SCALE = "Scale (x,y)"
+SIZE = "Image size (x*y)"
+MODE = "Image type"
+
+TAGS = {
+    COMMENT: 0,
+    DATE: 0,
+    EQUIPMENT: 0,
+    FRAMES: 0,
+    LUT: 0,
+    NAME: 0,
+    SCALE: 0,
+    SIZE: 0,
+    MODE: 0,
+}
+
+OPEN = {
+    # ifunc93/p3cfunc formats
+    "0 1 image": ("1", "1"),
+    "L 1 image": ("1", "1"),
+    "Greyscale image": ("L", "L"),
+    "Grayscale image": ("L", "L"),
+    "RGB image": ("RGB", "RGB;L"),
+    "RLB image": ("RGB", "RLB"),
+    "RYB image": ("RGB", "RLB"),
+    "B1 image": ("1", "1"),
+    "B2 image": ("P", "P;2"),
+    "B4 image": ("P", "P;4"),
+    "X 24 image": ("RGB", "RGB"),
+    "L 32 S image": ("I", "I;32"),
+    "L 32 F image": ("F", "F;32"),
+    # old p3cfunc formats
+    "RGB3 image": ("RGB", "RGB;T"),
+    "RYB3 image": ("RGB", "RYB;T"),
+    # extensions
+    "LA image": ("LA", "LA;L"),
+    "PA image": ("LA", "PA;L"),
+    "RGBA image": ("RGBA", "RGBA;L"),
+    "RGBX image": ("RGB", "RGBX;L"),
+    "CMYK image": ("CMYK", "CMYK;L"),
+    "YCC image": ("YCbCr", "YCbCr;L"),
+}
+
+# ifunc95 extensions
+for i in ["8", "8S", "16", "16S", "32", "32F"]:
+    OPEN[f"L {i} image"] = ("F", f"F;{i}")
+    OPEN[f"L*{i} image"] = ("F", f"F;{i}")
+for i in ["16", "16L", "16B"]:
+    OPEN[f"L {i} image"] = (f"I;{i}", f"I;{i}")
+    OPEN[f"L*{i} image"] = (f"I;{i}", f"I;{i}")
+for i in ["32S"]:
+    OPEN[f"L {i} image"] = ("I", f"I;{i}")
+    OPEN[f"L*{i} image"] = ("I", f"I;{i}")
+for j in range(2, 33):
+    OPEN[f"L*{j} image"] = ("F", f"F;{j}")
+
+
+# --------------------------------------------------------------------
+# Read IM directory
+
+split = re.compile(rb"^([A-Za-z][^:]*):[ \t]*(.*)[ \t]*$")
+
+
+def number(s: Any) -> float:
+    try:
+        return int(s)
+    except ValueError:
+        return float(s)
+
+
+##
+# Image plugin for the IFUNC IM file format.
+
+
+class ImImageFile(ImageFile.ImageFile):
+    format = "IM"
+    format_description = "IFUNC Image Memory"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        # Quick rejection: if there's not an LF among the first
+        # 100 bytes, this is (probably) not a text header.
+
+        if b"\n" not in self.fp.read(100):
+            msg = "not an IM file"
+            raise SyntaxError(msg)
+        self.fp.seek(0)
+
+        n = 0
+
+        # Default values
+        self.info[MODE] = "L"
+        self.info[SIZE] = (512, 512)
+        self.info[FRAMES] = 1
+
+        self.rawmode = "L"
+
+        while True:
+            s = self.fp.read(1)
+
+            # Some versions of IFUNC uses \n\r instead of \r\n...
+            if s == b"\r":
+                continue
+
+            if not s or s == b"\0" or s == b"\x1A":
+                break
+
+            # FIXME: this may read whole file if not a text file
+            s = s + self.fp.readline()
+
+            if len(s) > 100:
+                msg = "not an IM file"
+                raise SyntaxError(msg)
+
+            if s[-2:] == b"\r\n":
+                s = s[:-2]
+            elif s[-1:] == b"\n":
+                s = s[:-1]
+
+            try:
+                m = split.match(s)
+            except re.error as e:
+                msg = "not an IM file"
+                raise SyntaxError(msg) from e
+
+            if m:
+                k, v = m.group(1, 2)
+
+                # Don't know if this is the correct encoding,
+                # but a decent guess (I guess)
+                k = k.decode("latin-1", "replace")
+                v = v.decode("latin-1", "replace")
+
+                # Convert value as appropriate
+                if k in [FRAMES, SCALE, SIZE]:
+                    v = v.replace("*", ",")
+                    v = tuple(map(number, v.split(",")))
+                    if len(v) == 1:
+                        v = v[0]
+                elif k == MODE and v in OPEN:
+                    v, self.rawmode = OPEN[v]
+
+                # Add to dictionary. Note that COMMENT tags are
+                # combined into a list of strings.
+                if k == COMMENT:
+                    if k in self.info:
+                        self.info[k].append(v)
+                    else:
+                        self.info[k] = [v]
+                else:
+                    self.info[k] = v
+
+                if k in TAGS:
+                    n += 1
+
+            else:
+                msg = f"Syntax error in IM header: {s.decode('ascii', 'replace')}"
+                raise SyntaxError(msg)
+
+        if not n:
+            msg = "Not an IM file"
+            raise SyntaxError(msg)
+
+        # Basic attributes
+        self._size = self.info[SIZE]
+        self._mode = self.info[MODE]
+
+        # Skip forward to start of image data
+        while s and s[:1] != b"\x1A":
+            s = self.fp.read(1)
+        if not s:
+            msg = "File truncated"
+            raise SyntaxError(msg)
+
+        if LUT in self.info:
+            # convert lookup table to palette or lut attribute
+            palette = self.fp.read(768)
+            greyscale = 1  # greyscale palette
+            linear = 1  # linear greyscale palette
+            for i in range(256):
+                if palette[i] == palette[i + 256] == palette[i + 512]:
+                    if palette[i] != i:
+                        linear = 0
+                else:
+                    greyscale = 0
+            if self.mode in ["L", "LA", "P", "PA"]:
+                if greyscale:
+                    if not linear:
+                        self.lut = list(palette[:256])
+                else:
+                    if self.mode in ["L", "P"]:
+                        self._mode = self.rawmode = "P"
+                    elif self.mode in ["LA", "PA"]:
+                        self._mode = "PA"
+                        self.rawmode = "PA;L"
+                    self.palette = ImagePalette.raw("RGB;L", palette)
+            elif self.mode == "RGB":
+                if not greyscale or not linear:
+                    self.lut = list(palette)
+
+        self.frame = 0
+
+        self.__offset = offs = self.fp.tell()
+
+        self._fp = self.fp  # FIXME: hack
+
+        if self.rawmode[:2] == "F;":
+            # ifunc95 formats
+            try:
+                # use bit decoder (if necessary)
+                bits = int(self.rawmode[2:])
+                if bits not in [8, 16, 32]:
+                    self.tile = [
+                        ImageFile._Tile(
+                            "bit", (0, 0) + self.size, offs, (bits, 8, 3, 0, -1)
+                        )
+                    ]
+                    return
+            except ValueError:
+                pass
+
+        if self.rawmode in ["RGB;T", "RYB;T"]:
+            # Old LabEye/3PC files.  Would be very surprised if anyone
+            # ever stumbled upon such a file ;-)
+            size = self.size[0] * self.size[1]
+            self.tile = [
+                ImageFile._Tile("raw", (0, 0) + self.size, offs, ("G", 0, -1)),
+                ImageFile._Tile("raw", (0, 0) + self.size, offs + size, ("R", 0, -1)),
+                ImageFile._Tile(
+                    "raw", (0, 0) + self.size, offs + 2 * size, ("B", 0, -1)
+                ),
+            ]
+        else:
+            # LabEye/IFUNC files
+            self.tile = [
+                ImageFile._Tile("raw", (0, 0) + self.size, offs, (self.rawmode, 0, -1))
+            ]
+
+    @property
+    def n_frames(self) -> int:
+        return self.info[FRAMES]
+
+    @property
+    def is_animated(self) -> bool:
+        return self.info[FRAMES] > 1
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+
+        self.frame = frame
+
+        if self.mode == "1":
+            bits = 1
+        else:
+            bits = 8 * len(self.mode)
+
+        size = ((self.size[0] * bits + 7) // 8) * self.size[1]
+        offs = self.__offset + frame * size
+
+        self.fp = self._fp
+
+        self.tile = [
+            ImageFile._Tile("raw", (0, 0) + self.size, offs, (self.rawmode, 0, -1))
+        ]
+
+    def tell(self) -> int:
+        return self.frame
+
+
+#
+# --------------------------------------------------------------------
+# Save IM files
+
+
+SAVE = {
+    # mode: (im type, raw mode)
+    "1": ("0 1", "1"),
+    "L": ("Greyscale", "L"),
+    "LA": ("LA", "LA;L"),
+    "P": ("Greyscale", "P"),
+    "PA": ("LA", "PA;L"),
+    "I": ("L 32S", "I;32S"),
+    "I;16": ("L 16", "I;16"),
+    "I;16L": ("L 16L", "I;16L"),
+    "I;16B": ("L 16B", "I;16B"),
+    "F": ("L 32F", "F;32F"),
+    "RGB": ("RGB", "RGB;L"),
+    "RGBA": ("RGBA", "RGBA;L"),
+    "RGBX": ("RGBX", "RGBX;L"),
+    "CMYK": ("CMYK", "CMYK;L"),
+    "YCbCr": ("YCC", "YCbCr;L"),
+}
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    try:
+        image_type, rawmode = SAVE[im.mode]
+    except KeyError as e:
+        msg = f"Cannot save {im.mode} images as IM"
+        raise ValueError(msg) from e
+
+    frames = im.encoderinfo.get("frames", 1)
+
+    fp.write(f"Image type: {image_type} image\r\n".encode("ascii"))
+    if filename:
+        # Each line must be 100 characters or less,
+        # or: SyntaxError("not an IM file")
+        # 8 characters are used for "Name: " and "\r\n"
+        # Keep just the filename, ditch the potentially overlong path
+        if isinstance(filename, bytes):
+            filename = filename.decode("ascii")
+        name, ext = os.path.splitext(os.path.basename(filename))
+        name = "".join([name[: 92 - len(ext)], ext])
+
+        fp.write(f"Name: {name}\r\n".encode("ascii"))
+    fp.write(f"Image size (x*y): {im.size[0]}*{im.size[1]}\r\n".encode("ascii"))
+    fp.write(f"File size (no of images): {frames}\r\n".encode("ascii"))
+    if im.mode in ["P", "PA"]:
+        fp.write(b"Lut: 1\r\n")
+    fp.write(b"\000" * (511 - fp.tell()) + b"\032")
+    if im.mode in ["P", "PA"]:
+        im_palette = im.im.getpalette("RGB", "RGB;L")
+        colors = len(im_palette) // 3
+        palette = b""
+        for i in range(3):
+            palette += im_palette[colors * i : colors * (i + 1)]
+            palette += b"\x00" * (256 - colors)
+        fp.write(palette)  # 768 bytes
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, (rawmode, 0, -1))]
+    )
+
+
+#
+# --------------------------------------------------------------------
+# Registry
+
+
+Image.register_open(ImImageFile.format, ImImageFile)
+Image.register_save(ImImageFile.format, _save)
+
+Image.register_extension(ImImageFile.format, ".im")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/Image.py b/backend/venv/lib/python3.12/site-packages/PIL/Image.py
new file mode 100644
index 00000000..dff3d063
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/Image.py
@@ -0,0 +1,4197 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# the Image class wrapper
+#
+# partial release history:
+# 1995-09-09 fl   Created
+# 1996-03-11 fl   PIL release 0.0 (proof of concept)
+# 1996-04-30 fl   PIL release 0.1b1
+# 1999-07-28 fl   PIL release 1.0 final
+# 2000-06-07 fl   PIL release 1.1
+# 2000-10-20 fl   PIL release 1.1.1
+# 2001-05-07 fl   PIL release 1.1.2
+# 2002-03-15 fl   PIL release 1.1.3
+# 2003-05-10 fl   PIL release 1.1.4
+# 2005-03-28 fl   PIL release 1.1.5
+# 2006-12-02 fl   PIL release 1.1.6
+# 2009-11-15 fl   PIL release 1.1.7
+#
+# Copyright (c) 1997-2009 by Secret Labs AB.  All rights reserved.
+# Copyright (c) 1995-2009 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+from __future__ import annotations
+
+import abc
+import atexit
+import builtins
+import io
+import logging
+import math
+import os
+import re
+import struct
+import sys
+import tempfile
+import warnings
+from collections.abc import Callable, Iterator, MutableMapping, Sequence
+from enum import IntEnum
+from types import ModuleType
+from typing import (
+    IO,
+    TYPE_CHECKING,
+    Any,
+    Literal,
+    Protocol,
+    cast,
+)
+
+# VERSION was removed in Pillow 6.0.0.
+# PILLOW_VERSION was removed in Pillow 9.0.0.
+# Use __version__ instead.
+from . import (
+    ExifTags,
+    ImageMode,
+    TiffTags,
+    UnidentifiedImageError,
+    __version__,
+    _plugins,
+)
+from ._binary import i32le, o32be, o32le
+from ._deprecate import deprecate
+from ._util import DeferredError, is_path
+
+ElementTree: ModuleType | None
+try:
+    from defusedxml import ElementTree
+except ImportError:
+    ElementTree = None
+
+logger = logging.getLogger(__name__)
+
+
+class DecompressionBombWarning(RuntimeWarning):
+    pass
+
+
+class DecompressionBombError(Exception):
+    pass
+
+
+WARN_POSSIBLE_FORMATS: bool = False
+
+# Limit to around a quarter gigabyte for a 24-bit (3 bpp) image
+MAX_IMAGE_PIXELS: int | None = int(1024 * 1024 * 1024 // 4 // 3)
+
+
+try:
+    # If the _imaging C module is not present, Pillow will not load.
+    # Note that other modules should not refer to _imaging directly;
+    # import Image and use the Image.core variable instead.
+    # Also note that Image.core is not a publicly documented interface,
+    # and should be considered private and subject to change.
+    from . import _imaging as core
+
+    if __version__ != getattr(core, "PILLOW_VERSION", None):
+        msg = (
+            "The _imaging extension was built for another version of Pillow or PIL:\n"
+            f"Core version: {getattr(core, 'PILLOW_VERSION', None)}\n"
+            f"Pillow version: {__version__}"
+        )
+        raise ImportError(msg)
+
+except ImportError as v:
+    core = DeferredError.new(ImportError("The _imaging C module is not installed."))
+    # Explanations for ways that we know we might have an import error
+    if str(v).startswith("Module use of python"):
+        # The _imaging C module is present, but not compiled for
+        # the right version (windows only).  Print a warning, if
+        # possible.
+        warnings.warn(
+            "The _imaging extension was built for another version of Python.",
+            RuntimeWarning,
+        )
+    elif str(v).startswith("The _imaging extension"):
+        warnings.warn(str(v), RuntimeWarning)
+    # Fail here anyway. Don't let people run with a mostly broken Pillow.
+    # see docs/porting.rst
+    raise
+
+
+def isImageType(t: Any) -> TypeGuard[Image]:
+    """
+    Checks if an object is an image object.
+
+    .. warning::
+
+       This function is for internal use only.
+
+    :param t: object to check if it's an image
+    :returns: True if the object is an image
+    """
+    deprecate("Image.isImageType(im)", 12, "isinstance(im, Image.Image)")
+    return hasattr(t, "im")
+
+
+#
+# Constants
+
+
+# transpose
+class Transpose(IntEnum):
+    FLIP_LEFT_RIGHT = 0
+    FLIP_TOP_BOTTOM = 1
+    ROTATE_90 = 2
+    ROTATE_180 = 3
+    ROTATE_270 = 4
+    TRANSPOSE = 5
+    TRANSVERSE = 6
+
+
+# transforms (also defined in Imaging.h)
+class Transform(IntEnum):
+    AFFINE = 0
+    EXTENT = 1
+    PERSPECTIVE = 2
+    QUAD = 3
+    MESH = 4
+
+
+# resampling filters (also defined in Imaging.h)
+class Resampling(IntEnum):
+    NEAREST = 0
+    BOX = 4
+    BILINEAR = 2
+    HAMMING = 5
+    BICUBIC = 3
+    LANCZOS = 1
+
+
+_filters_support = {
+    Resampling.BOX: 0.5,
+    Resampling.BILINEAR: 1.0,
+    Resampling.HAMMING: 1.0,
+    Resampling.BICUBIC: 2.0,
+    Resampling.LANCZOS: 3.0,
+}
+
+
+# dithers
+class Dither(IntEnum):
+    NONE = 0
+    ORDERED = 1  # Not yet implemented
+    RASTERIZE = 2  # Not yet implemented
+    FLOYDSTEINBERG = 3  # default
+
+
+# palettes/quantizers
+class Palette(IntEnum):
+    WEB = 0
+    ADAPTIVE = 1
+
+
+class Quantize(IntEnum):
+    MEDIANCUT = 0
+    MAXCOVERAGE = 1
+    FASTOCTREE = 2
+    LIBIMAGEQUANT = 3
+
+
+module = sys.modules[__name__]
+for enum in (Transpose, Transform, Resampling, Dither, Palette, Quantize):
+    for item in enum:
+        setattr(module, item.name, item.value)
+
+
+if hasattr(core, "DEFAULT_STRATEGY"):
+    DEFAULT_STRATEGY = core.DEFAULT_STRATEGY
+    FILTERED = core.FILTERED
+    HUFFMAN_ONLY = core.HUFFMAN_ONLY
+    RLE = core.RLE
+    FIXED = core.FIXED
+
+
+# --------------------------------------------------------------------
+# Registries
+
+if TYPE_CHECKING:
+    import mmap
+    from xml.etree.ElementTree import Element
+
+    from IPython.lib.pretty import PrettyPrinter
+
+    from . import ImageFile, ImageFilter, ImagePalette, ImageQt, TiffImagePlugin
+    from ._typing import CapsuleType, NumpyArray, StrOrBytesPath, TypeGuard
+ID: list[str] = []
+OPEN: dict[
+    str,
+    tuple[
+        Callable[[IO[bytes], str | bytes], ImageFile.ImageFile],
+        Callable[[bytes], bool | str] | None,
+    ],
+] = {}
+MIME: dict[str, str] = {}
+SAVE: dict[str, Callable[[Image, IO[bytes], str | bytes], None]] = {}
+SAVE_ALL: dict[str, Callable[[Image, IO[bytes], str | bytes], None]] = {}
+EXTENSION: dict[str, str] = {}
+DECODERS: dict[str, type[ImageFile.PyDecoder]] = {}
+ENCODERS: dict[str, type[ImageFile.PyEncoder]] = {}
+
+# --------------------------------------------------------------------
+# Modes
+
+_ENDIAN = "<" if sys.byteorder == "little" else ">"
+
+
+def _conv_type_shape(im: Image) -> tuple[tuple[int, ...], str]:
+    m = ImageMode.getmode(im.mode)
+    shape: tuple[int, ...] = (im.height, im.width)
+    extra = len(m.bands)
+    if extra != 1:
+        shape += (extra,)
+    return shape, m.typestr
+
+
+MODES = [
+    "1",
+    "CMYK",
+    "F",
+    "HSV",
+    "I",
+    "I;16",
+    "I;16B",
+    "I;16L",
+    "I;16N",
+    "L",
+    "LA",
+    "La",
+    "LAB",
+    "P",
+    "PA",
+    "RGB",
+    "RGBA",
+    "RGBa",
+    "RGBX",
+    "YCbCr",
+]
+
+# raw modes that may be memory mapped.  NOTE: if you change this, you
+# may have to modify the stride calculation in map.c too!
+_MAPMODES = ("L", "P", "RGBX", "RGBA", "CMYK", "I;16", "I;16L", "I;16B")
+
+
+def getmodebase(mode: str) -> str:
+    """
+    Gets the "base" mode for given mode.  This function returns "L" for
+    images that contain grayscale data, and "RGB" for images that
+    contain color data.
+
+    :param mode: Input mode.
+    :returns: "L" or "RGB".
+    :exception KeyError: If the input mode was not a standard mode.
+    """
+    return ImageMode.getmode(mode).basemode
+
+
+def getmodetype(mode: str) -> str:
+    """
+    Gets the storage type mode.  Given a mode, this function returns a
+    single-layer mode suitable for storing individual bands.
+
+    :param mode: Input mode.
+    :returns: "L", "I", or "F".
+    :exception KeyError: If the input mode was not a standard mode.
+    """
+    return ImageMode.getmode(mode).basetype
+
+
+def getmodebandnames(mode: str) -> tuple[str, ...]:
+    """
+    Gets a list of individual band names.  Given a mode, this function returns
+    a tuple containing the names of individual bands (use
+    :py:method:`~PIL.Image.getmodetype` to get the mode used to store each
+    individual band.
+
+    :param mode: Input mode.
+    :returns: A tuple containing band names.  The length of the tuple
+        gives the number of bands in an image of the given mode.
+    :exception KeyError: If the input mode was not a standard mode.
+    """
+    return ImageMode.getmode(mode).bands
+
+
+def getmodebands(mode: str) -> int:
+    """
+    Gets the number of individual bands for this mode.
+
+    :param mode: Input mode.
+    :returns: The number of bands in this mode.
+    :exception KeyError: If the input mode was not a standard mode.
+    """
+    return len(ImageMode.getmode(mode).bands)
+
+
+# --------------------------------------------------------------------
+# Helpers
+
+_initialized = 0
+
+
+def preinit() -> None:
+    """
+    Explicitly loads BMP, GIF, JPEG, PPM and PPM file format drivers.
+
+    It is called when opening or saving images.
+    """
+
+    global _initialized
+    if _initialized >= 1:
+        return
+
+    try:
+        from . import BmpImagePlugin
+
+        assert BmpImagePlugin
+    except ImportError:
+        pass
+    try:
+        from . import GifImagePlugin
+
+        assert GifImagePlugin
+    except ImportError:
+        pass
+    try:
+        from . import JpegImagePlugin
+
+        assert JpegImagePlugin
+    except ImportError:
+        pass
+    try:
+        from . import PpmImagePlugin
+
+        assert PpmImagePlugin
+    except ImportError:
+        pass
+    try:
+        from . import PngImagePlugin
+
+        assert PngImagePlugin
+    except ImportError:
+        pass
+
+    _initialized = 1
+
+
+def init() -> bool:
+    """
+    Explicitly initializes the Python Imaging Library. This function
+    loads all available file format drivers.
+
+    It is called when opening or saving images if :py:meth:`~preinit()` is
+    insufficient, and by :py:meth:`~PIL.features.pilinfo`.
+    """
+
+    global _initialized
+    if _initialized >= 2:
+        return False
+
+    parent_name = __name__.rpartition(".")[0]
+    for plugin in _plugins:
+        try:
+            logger.debug("Importing %s", plugin)
+            __import__(f"{parent_name}.{plugin}", globals(), locals(), [])
+        except ImportError as e:
+            logger.debug("Image: failed to import %s: %s", plugin, e)
+
+    if OPEN or SAVE:
+        _initialized = 2
+        return True
+    return False
+
+
+# --------------------------------------------------------------------
+# Codec factories (used by tobytes/frombytes and ImageFile.load)
+
+
+def _getdecoder(
+    mode: str, decoder_name: str, args: Any, extra: tuple[Any, ...] = ()
+) -> core.ImagingDecoder | ImageFile.PyDecoder:
+    # tweak arguments
+    if args is None:
+        args = ()
+    elif not isinstance(args, tuple):
+        args = (args,)
+
+    try:
+        decoder = DECODERS[decoder_name]
+    except KeyError:
+        pass
+    else:
+        return decoder(mode, *args + extra)
+
+    try:
+        # get decoder
+        decoder = getattr(core, f"{decoder_name}_decoder")
+    except AttributeError as e:
+        msg = f"decoder {decoder_name} not available"
+        raise OSError(msg) from e
+    return decoder(mode, *args + extra)
+
+
+def _getencoder(
+    mode: str, encoder_name: str, args: Any, extra: tuple[Any, ...] = ()
+) -> core.ImagingEncoder | ImageFile.PyEncoder:
+    # tweak arguments
+    if args is None:
+        args = ()
+    elif not isinstance(args, tuple):
+        args = (args,)
+
+    try:
+        encoder = ENCODERS[encoder_name]
+    except KeyError:
+        pass
+    else:
+        return encoder(mode, *args + extra)
+
+    try:
+        # get encoder
+        encoder = getattr(core, f"{encoder_name}_encoder")
+    except AttributeError as e:
+        msg = f"encoder {encoder_name} not available"
+        raise OSError(msg) from e
+    return encoder(mode, *args + extra)
+
+
+# --------------------------------------------------------------------
+# Simple expression analyzer
+
+
+class ImagePointTransform:
+    """
+    Used with :py:meth:`~PIL.Image.Image.point` for single band images with more than
+    8 bits, this represents an affine transformation, where the value is multiplied by
+    ``scale`` and ``offset`` is added.
+    """
+
+    def __init__(self, scale: float, offset: float) -> None:
+        self.scale = scale
+        self.offset = offset
+
+    def __neg__(self) -> ImagePointTransform:
+        return ImagePointTransform(-self.scale, -self.offset)
+
+    def __add__(self, other: ImagePointTransform | float) -> ImagePointTransform:
+        if isinstance(other, ImagePointTransform):
+            return ImagePointTransform(
+                self.scale + other.scale, self.offset + other.offset
+            )
+        return ImagePointTransform(self.scale, self.offset + other)
+
+    __radd__ = __add__
+
+    def __sub__(self, other: ImagePointTransform | float) -> ImagePointTransform:
+        return self + -other
+
+    def __rsub__(self, other: ImagePointTransform | float) -> ImagePointTransform:
+        return other + -self
+
+    def __mul__(self, other: ImagePointTransform | float) -> ImagePointTransform:
+        if isinstance(other, ImagePointTransform):
+            return NotImplemented
+        return ImagePointTransform(self.scale * other, self.offset * other)
+
+    __rmul__ = __mul__
+
+    def __truediv__(self, other: ImagePointTransform | float) -> ImagePointTransform:
+        if isinstance(other, ImagePointTransform):
+            return NotImplemented
+        return ImagePointTransform(self.scale / other, self.offset / other)
+
+
+def _getscaleoffset(
+    expr: Callable[[ImagePointTransform], ImagePointTransform | float]
+) -> tuple[float, float]:
+    a = expr(ImagePointTransform(1, 0))
+    return (a.scale, a.offset) if isinstance(a, ImagePointTransform) else (0, a)
+
+
+# --------------------------------------------------------------------
+# Implementation wrapper
+
+
+class SupportsGetData(Protocol):
+    def getdata(
+        self,
+    ) -> tuple[Transform, Sequence[int]]: ...
+
+
+class Image:
+    """
+    This class represents an image object.  To create
+    :py:class:`~PIL.Image.Image` objects, use the appropriate factory
+    functions.  There's hardly ever any reason to call the Image constructor
+    directly.
+
+    * :py:func:`~PIL.Image.open`
+    * :py:func:`~PIL.Image.new`
+    * :py:func:`~PIL.Image.frombytes`
+    """
+
+    format: str | None = None
+    format_description: str | None = None
+    _close_exclusive_fp_after_loading = True
+
+    def __init__(self) -> None:
+        # FIXME: take "new" parameters / other image?
+        # FIXME: turn mode and size into delegating properties?
+        self._im: core.ImagingCore | DeferredError | None = None
+        self._mode = ""
+        self._size = (0, 0)
+        self.palette: ImagePalette.ImagePalette | None = None
+        self.info: dict[str | tuple[int, int], Any] = {}
+        self.readonly = 0
+        self._exif: Exif | None = None
+
+    @property
+    def im(self) -> core.ImagingCore:
+        if isinstance(self._im, DeferredError):
+            raise self._im.ex
+        assert self._im is not None
+        return self._im
+
+    @im.setter
+    def im(self, im: core.ImagingCore) -> None:
+        self._im = im
+
+    @property
+    def width(self) -> int:
+        return self.size[0]
+
+    @property
+    def height(self) -> int:
+        return self.size[1]
+
+    @property
+    def size(self) -> tuple[int, int]:
+        return self._size
+
+    @property
+    def mode(self) -> str:
+        return self._mode
+
+    def _new(self, im: core.ImagingCore) -> Image:
+        new = Image()
+        new.im = im
+        new._mode = im.mode
+        new._size = im.size
+        if im.mode in ("P", "PA"):
+            if self.palette:
+                new.palette = self.palette.copy()
+            else:
+                from . import ImagePalette
+
+                new.palette = ImagePalette.ImagePalette()
+        new.info = self.info.copy()
+        return new
+
+    # Context manager support
+    def __enter__(self):
+        return self
+
+    def _close_fp(self):
+        if getattr(self, "_fp", False):
+            if self._fp != self.fp:
+                self._fp.close()
+            self._fp = DeferredError(ValueError("Operation on closed image"))
+        if self.fp:
+            self.fp.close()
+
+    def __exit__(self, *args):
+        if hasattr(self, "fp"):
+            if getattr(self, "_exclusive_fp", False):
+                self._close_fp()
+            self.fp = None
+
+    def close(self) -> None:
+        """
+        Closes the file pointer, if possible.
+
+        This operation will destroy the image core and release its memory.
+        The image data will be unusable afterward.
+
+        This function is required to close images that have multiple frames or
+        have not had their file read and closed by the
+        :py:meth:`~PIL.Image.Image.load` method. See :ref:`file-handling` for
+        more information.
+        """
+        if hasattr(self, "fp"):
+            try:
+                self._close_fp()
+                self.fp = None
+            except Exception as msg:
+                logger.debug("Error closing: %s", msg)
+
+        if getattr(self, "map", None):
+            self.map: mmap.mmap | None = None
+
+        # Instead of simply setting to None, we're setting up a
+        # deferred error that will better explain that the core image
+        # object is gone.
+        self._im = DeferredError(ValueError("Operation on closed image"))
+
+    def _copy(self) -> None:
+        self.load()
+        self.im = self.im.copy()
+        self.readonly = 0
+
+    def _ensure_mutable(self) -> None:
+        if self.readonly:
+            self._copy()
+        else:
+            self.load()
+
+    def _dump(
+        self, file: str | None = None, format: str | None = None, **options: Any
+    ) -> str:
+        suffix = ""
+        if format:
+            suffix = f".{format}"
+
+        if not file:
+            f, filename = tempfile.mkstemp(suffix)
+            os.close(f)
+        else:
+            filename = file
+            if not filename.endswith(suffix):
+                filename = filename + suffix
+
+        self.load()
+
+        if not format or format == "PPM":
+            self.im.save_ppm(filename)
+        else:
+            self.save(filename, format, **options)
+
+        return filename
+
+    def __eq__(self, other: object) -> bool:
+        if self.__class__ is not other.__class__:
+            return False
+        assert isinstance(other, Image)
+        return (
+            self.mode == other.mode
+            and self.size == other.size
+            and self.info == other.info
+            and self.getpalette() == other.getpalette()
+            and self.tobytes() == other.tobytes()
+        )
+
+    def __repr__(self) -> str:
+        return (
+            f"<{self.__class__.__module__}.{self.__class__.__name__} "
+            f"image mode={self.mode} size={self.size[0]}x{self.size[1]} "
+            f"at 0x{id(self):X}>"
+        )
+
+    def _repr_pretty_(self, p: PrettyPrinter, cycle: bool) -> None:
+        """IPython plain text display support"""
+
+        # Same as __repr__ but without unpredictable id(self),
+        # to keep Jupyter notebook `text/plain` output stable.
+        p.text(
+            f"<{self.__class__.__module__}.{self.__class__.__name__} "
+            f"image mode={self.mode} size={self.size[0]}x{self.size[1]}>"
+        )
+
+    def _repr_image(self, image_format: str, **kwargs: Any) -> bytes | None:
+        """Helper function for iPython display hook.
+
+        :param image_format: Image format.
+        :returns: image as bytes, saved into the given format.
+        """
+        b = io.BytesIO()
+        try:
+            self.save(b, image_format, **kwargs)
+        except Exception:
+            return None
+        return b.getvalue()
+
+    def _repr_png_(self) -> bytes | None:
+        """iPython display hook support for PNG format.
+
+        :returns: PNG version of the image as bytes
+        """
+        return self._repr_image("PNG", compress_level=1)
+
+    def _repr_jpeg_(self) -> bytes | None:
+        """iPython display hook support for JPEG format.
+
+        :returns: JPEG version of the image as bytes
+        """
+        return self._repr_image("JPEG")
+
+    @property
+    def __array_interface__(self) -> dict[str, str | bytes | int | tuple[int, ...]]:
+        # numpy array interface support
+        new: dict[str, str | bytes | int | tuple[int, ...]] = {"version": 3}
+        if self.mode == "1":
+            # Binary images need to be extended from bits to bytes
+            # See: https://github.com/python-pillow/Pillow/issues/350
+            new["data"] = self.tobytes("raw", "L")
+        else:
+            new["data"] = self.tobytes()
+        new["shape"], new["typestr"] = _conv_type_shape(self)
+        return new
+
+    def __getstate__(self) -> list[Any]:
+        im_data = self.tobytes()  # load image first
+        return [self.info, self.mode, self.size, self.getpalette(), im_data]
+
+    def __setstate__(self, state: list[Any]) -> None:
+        Image.__init__(self)
+        info, mode, size, palette, data = state[:5]
+        self.info = info
+        self._mode = mode
+        self._size = size
+        self.im = core.new(mode, size)
+        if mode in ("L", "LA", "P", "PA") and palette:
+            self.putpalette(palette)
+        self.frombytes(data)
+
+    def tobytes(self, encoder_name: str = "raw", *args: Any) -> bytes:
+        """
+        Return image as a bytes object.
+
+        .. warning::
+
+            This method returns the raw image data from the internal
+            storage.  For compressed image data (e.g. PNG, JPEG) use
+            :meth:`~.save`, with a BytesIO parameter for in-memory
+            data.
+
+        :param encoder_name: What encoder to use.  The default is to
+                             use the standard "raw" encoder.
+
+                             A list of C encoders can be seen under
+                             codecs section of the function array in
+                             :file:`_imaging.c`. Python encoders are
+                             registered within the relevant plugins.
+        :param args: Extra arguments to the encoder.
+        :returns: A :py:class:`bytes` object.
+        """
+
+        encoder_args: Any = args
+        if len(encoder_args) == 1 and isinstance(encoder_args[0], tuple):
+            # may pass tuple instead of argument list
+            encoder_args = encoder_args[0]
+
+        if encoder_name == "raw" and encoder_args == ():
+            encoder_args = self.mode
+
+        self.load()
+
+        if self.width == 0 or self.height == 0:
+            return b""
+
+        # unpack data
+        e = _getencoder(self.mode, encoder_name, encoder_args)
+        e.setimage(self.im)
+
+        bufsize = max(65536, self.size[0] * 4)  # see RawEncode.c
+
+        output = []
+        while True:
+            bytes_consumed, errcode, data = e.encode(bufsize)
+            output.append(data)
+            if errcode:
+                break
+        if errcode < 0:
+            msg = f"encoder error {errcode} in tobytes"
+            raise RuntimeError(msg)
+
+        return b"".join(output)
+
+    def tobitmap(self, name: str = "image") -> bytes:
+        """
+        Returns the image converted to an X11 bitmap.
+
+        .. note:: This method only works for mode "1" images.
+
+        :param name: The name prefix to use for the bitmap variables.
+        :returns: A string containing an X11 bitmap.
+        :raises ValueError: If the mode is not "1"
+        """
+
+        self.load()
+        if self.mode != "1":
+            msg = "not a bitmap"
+            raise ValueError(msg)
+        data = self.tobytes("xbm")
+        return b"".join(
+            [
+                f"#define {name}_width {self.size[0]}\n".encode("ascii"),
+                f"#define {name}_height {self.size[1]}\n".encode("ascii"),
+                f"static char {name}_bits[] = {{\n".encode("ascii"),
+                data,
+                b"};",
+            ]
+        )
+
+    def frombytes(
+        self,
+        data: bytes | bytearray | SupportsArrayInterface,
+        decoder_name: str = "raw",
+        *args: Any,
+    ) -> None:
+        """
+        Loads this image with pixel data from a bytes object.
+
+        This method is similar to the :py:func:`~PIL.Image.frombytes` function,
+        but loads data into this image instead of creating a new image object.
+        """
+
+        if self.width == 0 or self.height == 0:
+            return
+
+        decoder_args: Any = args
+        if len(decoder_args) == 1 and isinstance(decoder_args[0], tuple):
+            # may pass tuple instead of argument list
+            decoder_args = decoder_args[0]
+
+        # default format
+        if decoder_name == "raw" and decoder_args == ():
+            decoder_args = self.mode
+
+        # unpack data
+        d = _getdecoder(self.mode, decoder_name, decoder_args)
+        d.setimage(self.im)
+        s = d.decode(data)
+
+        if s[0] >= 0:
+            msg = "not enough image data"
+            raise ValueError(msg)
+        if s[1] != 0:
+            msg = "cannot decode image data"
+            raise ValueError(msg)
+
+    def load(self) -> core.PixelAccess | None:
+        """
+        Allocates storage for the image and loads the pixel data.  In
+        normal cases, you don't need to call this method, since the
+        Image class automatically loads an opened image when it is
+        accessed for the first time.
+
+        If the file associated with the image was opened by Pillow, then this
+        method will close it. The exception to this is if the image has
+        multiple frames, in which case the file will be left open for seek
+        operations. See :ref:`file-handling` for more information.
+
+        :returns: An image access object.
+        :rtype: :py:class:`.PixelAccess`
+        """
+        if self._im is not None and self.palette and self.palette.dirty:
+            # realize palette
+            mode, arr = self.palette.getdata()
+            self.im.putpalette(self.palette.mode, mode, arr)
+            self.palette.dirty = 0
+            self.palette.rawmode = None
+            if "transparency" in self.info and mode in ("LA", "PA"):
+                if isinstance(self.info["transparency"], int):
+                    self.im.putpalettealpha(self.info["transparency"], 0)
+                else:
+                    self.im.putpalettealphas(self.info["transparency"])
+                self.palette.mode = "RGBA"
+            else:
+                self.palette.palette = self.im.getpalette(
+                    self.palette.mode, self.palette.mode
+                )
+
+        if self._im is not None:
+            return self.im.pixel_access(self.readonly)
+        return None
+
+    def verify(self) -> None:
+        """
+        Verifies the contents of a file. For data read from a file, this
+        method attempts to determine if the file is broken, without
+        actually decoding the image data.  If this method finds any
+        problems, it raises suitable exceptions.  If you need to load
+        the image after using this method, you must reopen the image
+        file.
+        """
+        pass
+
+    def convert(
+        self,
+        mode: str | None = None,
+        matrix: tuple[float, ...] | None = None,
+        dither: Dither | None = None,
+        palette: Palette = Palette.WEB,
+        colors: int = 256,
+    ) -> Image:
+        """
+        Returns a converted copy of this image. For the "P" mode, this
+        method translates pixels through the palette.  If mode is
+        omitted, a mode is chosen so that all information in the image
+        and the palette can be represented without a palette.
+
+        This supports all possible conversions between "L", "RGB" and "CMYK". The
+        ``matrix`` argument only supports "L" and "RGB".
+
+        When translating a color image to grayscale (mode "L"),
+        the library uses the ITU-R 601-2 luma transform::
+
+            L = R * 299/1000 + G * 587/1000 + B * 114/1000
+
+        The default method of converting a grayscale ("L") or "RGB"
+        image into a bilevel (mode "1") image uses Floyd-Steinberg
+        dither to approximate the original image luminosity levels. If
+        dither is ``None``, all values larger than 127 are set to 255 (white),
+        all other values to 0 (black). To use other thresholds, use the
+        :py:meth:`~PIL.Image.Image.point` method.
+
+        When converting from "RGBA" to "P" without a ``matrix`` argument,
+        this passes the operation to :py:meth:`~PIL.Image.Image.quantize`,
+        and ``dither`` and ``palette`` are ignored.
+
+        When converting from "PA", if an "RGBA" palette is present, the alpha
+        channel from the image will be used instead of the values from the palette.
+
+        :param mode: The requested mode. See: :ref:`concept-modes`.
+        :param matrix: An optional conversion matrix.  If given, this
+           should be 4- or 12-tuple containing floating point values.
+        :param dither: Dithering method, used when converting from
+           mode "RGB" to "P" or from "RGB" or "L" to "1".
+           Available methods are :data:`Dither.NONE` or :data:`Dither.FLOYDSTEINBERG`
+           (default). Note that this is not used when ``matrix`` is supplied.
+        :param palette: Palette to use when converting from mode "RGB"
+           to "P".  Available palettes are :data:`Palette.WEB` or
+           :data:`Palette.ADAPTIVE`.
+        :param colors: Number of colors to use for the :data:`Palette.ADAPTIVE`
+           palette. Defaults to 256.
+        :rtype: :py:class:`~PIL.Image.Image`
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        if mode in ("BGR;15", "BGR;16", "BGR;24"):
+            deprecate(mode, 12)
+
+        self.load()
+
+        has_transparency = "transparency" in self.info
+        if not mode and self.mode == "P":
+            # determine default mode
+            if self.palette:
+                mode = self.palette.mode
+            else:
+                mode = "RGB"
+            if mode == "RGB" and has_transparency:
+                mode = "RGBA"
+        if not mode or (mode == self.mode and not matrix):
+            return self.copy()
+
+        if matrix:
+            # matrix conversion
+            if mode not in ("L", "RGB"):
+                msg = "illegal conversion"
+                raise ValueError(msg)
+            im = self.im.convert_matrix(mode, matrix)
+            new_im = self._new(im)
+            if has_transparency and self.im.bands == 3:
+                transparency = new_im.info["transparency"]
+
+                def convert_transparency(
+                    m: tuple[float, ...], v: tuple[int, int, int]
+                ) -> int:
+                    value = m[0] * v[0] + m[1] * v[1] + m[2] * v[2] + m[3] * 0.5
+                    return max(0, min(255, int(value)))
+
+                if mode == "L":
+                    transparency = convert_transparency(matrix, transparency)
+                elif len(mode) == 3:
+                    transparency = tuple(
+                        convert_transparency(matrix[i * 4 : i * 4 + 4], transparency)
+                        for i in range(0, len(transparency))
+                    )
+                new_im.info["transparency"] = transparency
+            return new_im
+
+        if mode == "P" and self.mode == "RGBA":
+            return self.quantize(colors)
+
+        trns = None
+        delete_trns = False
+        # transparency handling
+        if has_transparency:
+            if (self.mode in ("1", "L", "I", "I;16") and mode in ("LA", "RGBA")) or (
+                self.mode == "RGB" and mode in ("La", "LA", "RGBa", "RGBA")
+            ):
+                # Use transparent conversion to promote from transparent
+                # color to an alpha channel.
+                new_im = self._new(
+                    self.im.convert_transparent(mode, self.info["transparency"])
+                )
+                del new_im.info["transparency"]
+                return new_im
+            elif self.mode in ("L", "RGB", "P") and mode in ("L", "RGB", "P"):
+                t = self.info["transparency"]
+                if isinstance(t, bytes):
+                    # Dragons. This can't be represented by a single color
+                    warnings.warn(
+                        "Palette images with Transparency expressed in bytes should be "
+                        "converted to RGBA images"
+                    )
+                    delete_trns = True
+                else:
+                    # get the new transparency color.
+                    # use existing conversions
+                    trns_im = new(self.mode, (1, 1))
+                    if self.mode == "P":
+                        assert self.palette is not None
+                        trns_im.putpalette(self.palette, self.palette.mode)
+                        if isinstance(t, tuple):
+                            err = "Couldn't allocate a palette color for transparency"
+                            assert trns_im.palette is not None
+                            try:
+                                t = trns_im.palette.getcolor(t, self)
+                            except ValueError as e:
+                                if str(e) == "cannot allocate more than 256 colors":
+                                    # If all 256 colors are in use,
+                                    # then there is no need for transparency
+                                    t = None
+                                else:
+                                    raise ValueError(err) from e
+                    if t is None:
+                        trns = None
+                    else:
+                        trns_im.putpixel((0, 0), t)
+
+                        if mode in ("L", "RGB"):
+                            trns_im = trns_im.convert(mode)
+                        else:
+                            # can't just retrieve the palette number, got to do it
+                            # after quantization.
+                            trns_im = trns_im.convert("RGB")
+                        trns = trns_im.getpixel((0, 0))
+
+            elif self.mode == "P" and mode in ("LA", "PA", "RGBA"):
+                t = self.info["transparency"]
+                delete_trns = True
+
+                if isinstance(t, bytes):
+                    self.im.putpalettealphas(t)
+                elif isinstance(t, int):
+                    self.im.putpalettealpha(t, 0)
+                else:
+                    msg = "Transparency for P mode should be bytes or int"
+                    raise ValueError(msg)
+
+        if mode == "P" and palette == Palette.ADAPTIVE:
+            im = self.im.quantize(colors)
+            new_im = self._new(im)
+            from . import ImagePalette
+
+            new_im.palette = ImagePalette.ImagePalette(
+                "RGB", new_im.im.getpalette("RGB")
+            )
+            if delete_trns:
+                # This could possibly happen if we requantize to fewer colors.
+                # The transparency would be totally off in that case.
+                del new_im.info["transparency"]
+            if trns is not None:
+                try:
+                    new_im.info["transparency"] = new_im.palette.getcolor(
+                        cast(tuple[int, ...], trns),  # trns was converted to RGB
+                        new_im,
+                    )
+                except Exception:
+                    # if we can't make a transparent color, don't leave the old
+                    # transparency hanging around to mess us up.
+                    del new_im.info["transparency"]
+                    warnings.warn("Couldn't allocate palette entry for transparency")
+            return new_im
+
+        if "LAB" in (self.mode, mode):
+            im = self
+            if mode == "LAB":
+                if im.mode not in ("RGB", "RGBA", "RGBX"):
+                    im = im.convert("RGBA")
+                other_mode = im.mode
+            else:
+                other_mode = mode
+            if other_mode in ("RGB", "RGBA", "RGBX"):
+                from . import ImageCms
+
+                srgb = ImageCms.createProfile("sRGB")
+                lab = ImageCms.createProfile("LAB")
+                profiles = [lab, srgb] if im.mode == "LAB" else [srgb, lab]
+                transform = ImageCms.buildTransform(
+                    profiles[0], profiles[1], im.mode, mode
+                )
+                return transform.apply(im)
+
+        # colorspace conversion
+        if dither is None:
+            dither = Dither.FLOYDSTEINBERG
+
+        try:
+            im = self.im.convert(mode, dither)
+        except ValueError:
+            try:
+                # normalize source image and try again
+                modebase = getmodebase(self.mode)
+                if modebase == self.mode:
+                    raise
+                im = self.im.convert(modebase)
+                im = im.convert(mode, dither)
+            except KeyError as e:
+                msg = "illegal conversion"
+                raise ValueError(msg) from e
+
+        new_im = self._new(im)
+        if mode == "P" and palette != Palette.ADAPTIVE:
+            from . import ImagePalette
+
+            new_im.palette = ImagePalette.ImagePalette("RGB", im.getpalette("RGB"))
+        if delete_trns:
+            # crash fail if we leave a bytes transparency in an rgb/l mode.
+            del new_im.info["transparency"]
+        if trns is not None:
+            if new_im.mode == "P" and new_im.palette:
+                try:
+                    new_im.info["transparency"] = new_im.palette.getcolor(
+                        cast(tuple[int, ...], trns), new_im  # trns was converted to RGB
+                    )
+                except ValueError as e:
+                    del new_im.info["transparency"]
+                    if str(e) != "cannot allocate more than 256 colors":
+                        # If all 256 colors are in use,
+                        # then there is no need for transparency
+                        warnings.warn(
+                            "Couldn't allocate palette entry for transparency"
+                        )
+            else:
+                new_im.info["transparency"] = trns
+        return new_im
+
+    def quantize(
+        self,
+        colors: int = 256,
+        method: int | None = None,
+        kmeans: int = 0,
+        palette: Image | None = None,
+        dither: Dither = Dither.FLOYDSTEINBERG,
+    ) -> Image:
+        """
+        Convert the image to 'P' mode with the specified number
+        of colors.
+
+        :param colors: The desired number of colors, <= 256
+        :param method: :data:`Quantize.MEDIANCUT` (median cut),
+                       :data:`Quantize.MAXCOVERAGE` (maximum coverage),
+                       :data:`Quantize.FASTOCTREE` (fast octree),
+                       :data:`Quantize.LIBIMAGEQUANT` (libimagequant; check support
+                       using :py:func:`PIL.features.check_feature` with
+                       ``feature="libimagequant"``).
+
+                       By default, :data:`Quantize.MEDIANCUT` will be used.
+
+                       The exception to this is RGBA images. :data:`Quantize.MEDIANCUT`
+                       and :data:`Quantize.MAXCOVERAGE` do not support RGBA images, so
+                       :data:`Quantize.FASTOCTREE` is used by default instead.
+        :param kmeans: Integer greater than or equal to zero.
+        :param palette: Quantize to the palette of given
+                        :py:class:`PIL.Image.Image`.
+        :param dither: Dithering method, used when converting from
+           mode "RGB" to "P" or from "RGB" or "L" to "1".
+           Available methods are :data:`Dither.NONE` or :data:`Dither.FLOYDSTEINBERG`
+           (default).
+        :returns: A new image
+        """
+
+        self.load()
+
+        if method is None:
+            # defaults:
+            method = Quantize.MEDIANCUT
+            if self.mode == "RGBA":
+                method = Quantize.FASTOCTREE
+
+        if self.mode == "RGBA" and method not in (
+            Quantize.FASTOCTREE,
+            Quantize.LIBIMAGEQUANT,
+        ):
+            # Caller specified an invalid mode.
+            msg = (
+                "Fast Octree (method == 2) and libimagequant (method == 3) "
+                "are the only valid methods for quantizing RGBA images"
+            )
+            raise ValueError(msg)
+
+        if palette:
+            # use palette from reference image
+            palette.load()
+            if palette.mode != "P":
+                msg = "bad mode for palette image"
+                raise ValueError(msg)
+            if self.mode not in {"RGB", "L"}:
+                msg = "only RGB or L mode images can be quantized to a palette"
+                raise ValueError(msg)
+            im = self.im.convert("P", dither, palette.im)
+            new_im = self._new(im)
+            assert palette.palette is not None
+            new_im.palette = palette.palette.copy()
+            return new_im
+
+        if kmeans < 0:
+            msg = "kmeans must not be negative"
+            raise ValueError(msg)
+
+        im = self._new(self.im.quantize(colors, method, kmeans))
+
+        from . import ImagePalette
+
+        mode = im.im.getpalettemode()
+        palette_data = im.im.getpalette(mode, mode)[: colors * len(mode)]
+        im.palette = ImagePalette.ImagePalette(mode, palette_data)
+
+        return im
+
+    def copy(self) -> Image:
+        """
+        Copies this image. Use this method if you wish to paste things
+        into an image, but still retain the original.
+
+        :rtype: :py:class:`~PIL.Image.Image`
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+        self.load()
+        return self._new(self.im.copy())
+
+    __copy__ = copy
+
+    def crop(self, box: tuple[float, float, float, float] | None = None) -> Image:
+        """
+        Returns a rectangular region from this image. The box is a
+        4-tuple defining the left, upper, right, and lower pixel
+        coordinate. See :ref:`coordinate-system`.
+
+        Note: Prior to Pillow 3.4.0, this was a lazy operation.
+
+        :param box: The crop rectangle, as a (left, upper, right, lower)-tuple.
+        :rtype: :py:class:`~PIL.Image.Image`
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        if box is None:
+            return self.copy()
+
+        if box[2] < box[0]:
+            msg = "Coordinate 'right' is less than 'left'"
+            raise ValueError(msg)
+        elif box[3] < box[1]:
+            msg = "Coordinate 'lower' is less than 'upper'"
+            raise ValueError(msg)
+
+        self.load()
+        return self._new(self._crop(self.im, box))
+
+    def _crop(
+        self, im: core.ImagingCore, box: tuple[float, float, float, float]
+    ) -> core.ImagingCore:
+        """
+        Returns a rectangular region from the core image object im.
+
+        This is equivalent to calling im.crop((x0, y0, x1, y1)), but
+        includes additional sanity checks.
+
+        :param im: a core image object
+        :param box: The crop rectangle, as a (left, upper, right, lower)-tuple.
+        :returns: A core image object.
+        """
+
+        x0, y0, x1, y1 = map(int, map(round, box))
+
+        absolute_values = (abs(x1 - x0), abs(y1 - y0))
+
+        _decompression_bomb_check(absolute_values)
+
+        return im.crop((x0, y0, x1, y1))
+
+    def draft(
+        self, mode: str | None, size: tuple[int, int] | None
+    ) -> tuple[str, tuple[int, int, float, float]] | None:
+        """
+        Configures the image file loader so it returns a version of the
+        image that as closely as possible matches the given mode and
+        size. For example, you can use this method to convert a color
+        JPEG to grayscale while loading it.
+
+        If any changes are made, returns a tuple with the chosen ``mode`` and
+        ``box`` with coordinates of the original image within the altered one.
+
+        Note that this method modifies the :py:class:`~PIL.Image.Image` object
+        in place. If the image has already been loaded, this method has no
+        effect.
+
+        Note: This method is not implemented for most images. It is
+        currently implemented only for JPEG and MPO images.
+
+        :param mode: The requested mode.
+        :param size: The requested size in pixels, as a 2-tuple:
+           (width, height).
+        """
+        pass
+
+    def _expand(self, xmargin: int, ymargin: int | None = None) -> Image:
+        if ymargin is None:
+            ymargin = xmargin
+        self.load()
+        return self._new(self.im.expand(xmargin, ymargin))
+
+    def filter(self, filter: ImageFilter.Filter | type[ImageFilter.Filter]) -> Image:
+        """
+        Filters this image using the given filter.  For a list of
+        available filters, see the :py:mod:`~PIL.ImageFilter` module.
+
+        :param filter: Filter kernel.
+        :returns: An :py:class:`~PIL.Image.Image` object."""
+
+        from . import ImageFilter
+
+        self.load()
+
+        if callable(filter):
+            filter = filter()
+        if not hasattr(filter, "filter"):
+            msg = "filter argument should be ImageFilter.Filter instance or class"
+            raise TypeError(msg)
+
+        multiband = isinstance(filter, ImageFilter.MultibandFilter)
+        if self.im.bands == 1 or multiband:
+            return self._new(filter.filter(self.im))
+
+        ims = [
+            self._new(filter.filter(self.im.getband(c))) for c in range(self.im.bands)
+        ]
+        return merge(self.mode, ims)
+
+    def getbands(self) -> tuple[str, ...]:
+        """
+        Returns a tuple containing the name of each band in this image.
+        For example, ``getbands`` on an RGB image returns ("R", "G", "B").
+
+        :returns: A tuple containing band names.
+        :rtype: tuple
+        """
+        return ImageMode.getmode(self.mode).bands
+
+    def getbbox(self, *, alpha_only: bool = True) -> tuple[int, int, int, int] | None:
+        """
+        Calculates the bounding box of the non-zero regions in the
+        image.
+
+        :param alpha_only: Optional flag, defaulting to ``True``.
+           If ``True`` and the image has an alpha channel, trim transparent pixels.
+           Otherwise, trim pixels when all channels are zero.
+           Keyword-only argument.
+        :returns: The bounding box is returned as a 4-tuple defining the
+           left, upper, right, and lower pixel coordinate. See
+           :ref:`coordinate-system`. If the image is completely empty, this
+           method returns None.
+
+        """
+
+        self.load()
+        return self.im.getbbox(alpha_only)
+
+    def getcolors(
+        self, maxcolors: int = 256
+    ) -> list[tuple[int, tuple[int, ...]]] | list[tuple[int, float]] | None:
+        """
+        Returns a list of colors used in this image.
+
+        The colors will be in the image's mode. For example, an RGB image will
+        return a tuple of (red, green, blue) color values, and a P image will
+        return the index of the color in the palette.
+
+        :param maxcolors: Maximum number of colors.  If this number is
+           exceeded, this method returns None.  The default limit is
+           256 colors.
+        :returns: An unsorted list of (count, pixel) values.
+        """
+
+        self.load()
+        if self.mode in ("1", "L", "P"):
+            h = self.im.histogram()
+            out: list[tuple[int, float]] = [(h[i], i) for i in range(256) if h[i]]
+            if len(out) > maxcolors:
+                return None
+            return out
+        return self.im.getcolors(maxcolors)
+
+    def getdata(self, band: int | None = None) -> core.ImagingCore:
+        """
+        Returns the contents of this image as a sequence object
+        containing pixel values.  The sequence object is flattened, so
+        that values for line one follow directly after the values of
+        line zero, and so on.
+
+        Note that the sequence object returned by this method is an
+        internal PIL data type, which only supports certain sequence
+        operations.  To convert it to an ordinary sequence (e.g. for
+        printing), use ``list(im.getdata())``.
+
+        :param band: What band to return.  The default is to return
+           all bands.  To return a single band, pass in the index
+           value (e.g. 0 to get the "R" band from an "RGB" image).
+        :returns: A sequence-like object.
+        """
+
+        self.load()
+        if band is not None:
+            return self.im.getband(band)
+        return self.im  # could be abused
+
+    def getextrema(self) -> tuple[float, float] | tuple[tuple[int, int], ...]:
+        """
+        Gets the minimum and maximum pixel values for each band in
+        the image.
+
+        :returns: For a single-band image, a 2-tuple containing the
+           minimum and maximum pixel value.  For a multi-band image,
+           a tuple containing one 2-tuple for each band.
+        """
+
+        self.load()
+        if self.im.bands > 1:
+            return tuple(self.im.getband(i).getextrema() for i in range(self.im.bands))
+        return self.im.getextrema()
+
+    def getxmp(self) -> dict[str, Any]:
+        """
+        Returns a dictionary containing the XMP tags.
+        Requires defusedxml to be installed.
+
+        :returns: XMP tags in a dictionary.
+        """
+
+        def get_name(tag: str) -> str:
+            return re.sub("^{[^}]+}", "", tag)
+
+        def get_value(element: Element) -> str | dict[str, Any] | None:
+            value: dict[str, Any] = {get_name(k): v for k, v in element.attrib.items()}
+            children = list(element)
+            if children:
+                for child in children:
+                    name = get_name(child.tag)
+                    child_value = get_value(child)
+                    if name in value:
+                        if not isinstance(value[name], list):
+                            value[name] = [value[name]]
+                        value[name].append(child_value)
+                    else:
+                        value[name] = child_value
+            elif value:
+                if element.text:
+                    value["text"] = element.text
+            else:
+                return element.text
+            return value
+
+        if ElementTree is None:
+            warnings.warn("XMP data cannot be read without defusedxml dependency")
+            return {}
+        if "xmp" not in self.info:
+            return {}
+        root = ElementTree.fromstring(self.info["xmp"].rstrip(b"\x00"))
+        return {get_name(root.tag): get_value(root)}
+
+    def getexif(self) -> Exif:
+        """
+        Gets EXIF data from the image.
+
+        :returns: an :py:class:`~PIL.Image.Exif` object.
+        """
+        if self._exif is None:
+            self._exif = Exif()
+        elif self._exif._loaded:
+            return self._exif
+        self._exif._loaded = True
+
+        exif_info = self.info.get("exif")
+        if exif_info is None:
+            if "Raw profile type exif" in self.info:
+                exif_info = bytes.fromhex(
+                    "".join(self.info["Raw profile type exif"].split("\n")[3:])
+                )
+            elif hasattr(self, "tag_v2"):
+                self._exif.bigtiff = self.tag_v2._bigtiff
+                self._exif.endian = self.tag_v2._endian
+                self._exif.load_from_fp(self.fp, self.tag_v2._offset)
+        if exif_info is not None:
+            self._exif.load(exif_info)
+
+        # XMP tags
+        if ExifTags.Base.Orientation not in self._exif:
+            xmp_tags = self.info.get("XML:com.adobe.xmp")
+            if xmp_tags:
+                match = re.search(r'tiff:Orientation(="|>)([0-9])', xmp_tags)
+                if match:
+                    self._exif[ExifTags.Base.Orientation] = int(match[2])
+
+        return self._exif
+
+    def _reload_exif(self) -> None:
+        if self._exif is None or not self._exif._loaded:
+            return
+        self._exif._loaded = False
+        self.getexif()
+
+    def get_child_images(self) -> list[ImageFile.ImageFile]:
+        child_images = []
+        exif = self.getexif()
+        ifds = []
+        if ExifTags.Base.SubIFDs in exif:
+            subifd_offsets = exif[ExifTags.Base.SubIFDs]
+            if subifd_offsets:
+                if not isinstance(subifd_offsets, tuple):
+                    subifd_offsets = (subifd_offsets,)
+                for subifd_offset in subifd_offsets:
+                    ifds.append((exif._get_ifd_dict(subifd_offset), subifd_offset))
+        ifd1 = exif.get_ifd(ExifTags.IFD.IFD1)
+        if ifd1 and ifd1.get(ExifTags.Base.JpegIFOffset):
+            assert exif._info is not None
+            ifds.append((ifd1, exif._info.next))
+
+        offset = None
+        for ifd, ifd_offset in ifds:
+            current_offset = self.fp.tell()
+            if offset is None:
+                offset = current_offset
+
+            fp = self.fp
+            if ifd is not None:
+                thumbnail_offset = ifd.get(ExifTags.Base.JpegIFOffset)
+                if thumbnail_offset is not None:
+                    thumbnail_offset += getattr(self, "_exif_offset", 0)
+                    self.fp.seek(thumbnail_offset)
+                    data = self.fp.read(ifd.get(ExifTags.Base.JpegIFByteCount))
+                    fp = io.BytesIO(data)
+
+            with open(fp) as im:
+                from . import TiffImagePlugin
+
+                if thumbnail_offset is None and isinstance(
+                    im, TiffImagePlugin.TiffImageFile
+                ):
+                    im._frame_pos = [ifd_offset]
+                    im._seek(0)
+                im.load()
+                child_images.append(im)
+
+        if offset is not None:
+            self.fp.seek(offset)
+        return child_images
+
+    def getim(self) -> CapsuleType:
+        """
+        Returns a capsule that points to the internal image memory.
+
+        :returns: A capsule object.
+        """
+
+        self.load()
+        return self.im.ptr
+
+    def getpalette(self, rawmode: str | None = "RGB") -> list[int] | None:
+        """
+        Returns the image palette as a list.
+
+        :param rawmode: The mode in which to return the palette. ``None`` will
+           return the palette in its current mode.
+
+           .. versionadded:: 9.1.0
+
+        :returns: A list of color values [r, g, b, ...], or None if the
+           image has no palette.
+        """
+
+        self.load()
+        try:
+            mode = self.im.getpalettemode()
+        except ValueError:
+            return None  # no palette
+        if rawmode is None:
+            rawmode = mode
+        return list(self.im.getpalette(mode, rawmode))
+
+    @property
+    def has_transparency_data(self) -> bool:
+        """
+        Determine if an image has transparency data, whether in the form of an
+        alpha channel, a palette with an alpha channel, or a "transparency" key
+        in the info dictionary.
+
+        Note the image might still appear solid, if all of the values shown
+        within are opaque.
+
+        :returns: A boolean.
+        """
+        if (
+            self.mode in ("LA", "La", "PA", "RGBA", "RGBa")
+            or "transparency" in self.info
+        ):
+            return True
+        if self.mode == "P":
+            assert self.palette is not None
+            return self.palette.mode.endswith("A")
+        return False
+
+    def apply_transparency(self) -> None:
+        """
+        If a P mode image has a "transparency" key in the info dictionary,
+        remove the key and instead apply the transparency to the palette.
+        Otherwise, the image is unchanged.
+        """
+        if self.mode != "P" or "transparency" not in self.info:
+            return
+
+        from . import ImagePalette
+
+        palette = self.getpalette("RGBA")
+        assert palette is not None
+        transparency = self.info["transparency"]
+        if isinstance(transparency, bytes):
+            for i, alpha in enumerate(transparency):
+                palette[i * 4 + 3] = alpha
+        else:
+            palette[transparency * 4 + 3] = 0
+        self.palette = ImagePalette.ImagePalette("RGBA", bytes(palette))
+        self.palette.dirty = 1
+
+        del self.info["transparency"]
+
+    def getpixel(
+        self, xy: tuple[int, int] | list[int]
+    ) -> float | tuple[int, ...] | None:
+        """
+        Returns the pixel value at a given position.
+
+        :param xy: The coordinate, given as (x, y). See
+           :ref:`coordinate-system`.
+        :returns: The pixel value.  If the image is a multi-layer image,
+           this method returns a tuple.
+        """
+
+        self.load()
+        return self.im.getpixel(tuple(xy))
+
+    def getprojection(self) -> tuple[list[int], list[int]]:
+        """
+        Get projection to x and y axes
+
+        :returns: Two sequences, indicating where there are non-zero
+            pixels along the X-axis and the Y-axis, respectively.
+        """
+
+        self.load()
+        x, y = self.im.getprojection()
+        return list(x), list(y)
+
+    def histogram(
+        self, mask: Image | None = None, extrema: tuple[float, float] | None = None
+    ) -> list[int]:
+        """
+        Returns a histogram for the image. The histogram is returned as a
+        list of pixel counts, one for each pixel value in the source
+        image. Counts are grouped into 256 bins for each band, even if
+        the image has more than 8 bits per band. If the image has more
+        than one band, the histograms for all bands are concatenated (for
+        example, the histogram for an "RGB" image contains 768 values).
+
+        A bilevel image (mode "1") is treated as a grayscale ("L") image
+        by this method.
+
+        If a mask is provided, the method returns a histogram for those
+        parts of the image where the mask image is non-zero. The mask
+        image must have the same size as the image, and be either a
+        bi-level image (mode "1") or a grayscale image ("L").
+
+        :param mask: An optional mask.
+        :param extrema: An optional tuple of manually-specified extrema.
+        :returns: A list containing pixel counts.
+        """
+        self.load()
+        if mask:
+            mask.load()
+            return self.im.histogram((0, 0), mask.im)
+        if self.mode in ("I", "F"):
+            return self.im.histogram(
+                extrema if extrema is not None else self.getextrema()
+            )
+        return self.im.histogram()
+
+    def entropy(
+        self, mask: Image | None = None, extrema: tuple[float, float] | None = None
+    ) -> float:
+        """
+        Calculates and returns the entropy for the image.
+
+        A bilevel image (mode "1") is treated as a grayscale ("L")
+        image by this method.
+
+        If a mask is provided, the method employs the histogram for
+        those parts of the image where the mask image is non-zero.
+        The mask image must have the same size as the image, and be
+        either a bi-level image (mode "1") or a grayscale image ("L").
+
+        :param mask: An optional mask.
+        :param extrema: An optional tuple of manually-specified extrema.
+        :returns: A float value representing the image entropy
+        """
+        self.load()
+        if mask:
+            mask.load()
+            return self.im.entropy((0, 0), mask.im)
+        if self.mode in ("I", "F"):
+            return self.im.entropy(
+                extrema if extrema is not None else self.getextrema()
+            )
+        return self.im.entropy()
+
+    def paste(
+        self,
+        im: Image | str | float | tuple[float, ...],
+        box: Image | tuple[int, int, int, int] | tuple[int, int] | None = None,
+        mask: Image | None = None,
+    ) -> None:
+        """
+        Pastes another image into this image. The box argument is either
+        a 2-tuple giving the upper left corner, a 4-tuple defining the
+        left, upper, right, and lower pixel coordinate, or None (same as
+        (0, 0)). See :ref:`coordinate-system`. If a 4-tuple is given, the size
+        of the pasted image must match the size of the region.
+
+        If the modes don't match, the pasted image is converted to the mode of
+        this image (see the :py:meth:`~PIL.Image.Image.convert` method for
+        details).
+
+        Instead of an image, the source can be a integer or tuple
+        containing pixel values.  The method then fills the region
+        with the given color.  When creating RGB images, you can
+        also use color strings as supported by the ImageColor module.
+
+        If a mask is given, this method updates only the regions
+        indicated by the mask. You can use either "1", "L", "LA", "RGBA"
+        or "RGBa" images (if present, the alpha band is used as mask).
+        Where the mask is 255, the given image is copied as is.  Where
+        the mask is 0, the current value is preserved.  Intermediate
+        values will mix the two images together, including their alpha
+        channels if they have them.
+
+        See :py:meth:`~PIL.Image.Image.alpha_composite` if you want to
+        combine images with respect to their alpha channels.
+
+        :param im: Source image or pixel value (integer, float or tuple).
+        :param box: An optional 4-tuple giving the region to paste into.
+           If a 2-tuple is used instead, it's treated as the upper left
+           corner.  If omitted or None, the source is pasted into the
+           upper left corner.
+
+           If an image is given as the second argument and there is no
+           third, the box defaults to (0, 0), and the second argument
+           is interpreted as a mask image.
+        :param mask: An optional mask image.
+        """
+
+        if isinstance(box, Image):
+            if mask is not None:
+                msg = "If using second argument as mask, third argument must be None"
+                raise ValueError(msg)
+            # abbreviated paste(im, mask) syntax
+            mask = box
+            box = None
+
+        if box is None:
+            box = (0, 0)
+
+        if len(box) == 2:
+            # upper left corner given; get size from image or mask
+            if isinstance(im, Image):
+                size = im.size
+            elif isinstance(mask, Image):
+                size = mask.size
+            else:
+                # FIXME: use self.size here?
+                msg = "cannot determine region size; use 4-item box"
+                raise ValueError(msg)
+            box += (box[0] + size[0], box[1] + size[1])
+
+        source: core.ImagingCore | str | float | tuple[float, ...]
+        if isinstance(im, str):
+            from . import ImageColor
+
+            source = ImageColor.getcolor(im, self.mode)
+        elif isinstance(im, Image):
+            im.load()
+            if self.mode != im.mode:
+                if self.mode != "RGB" or im.mode not in ("LA", "RGBA", "RGBa"):
+                    # should use an adapter for this!
+                    im = im.convert(self.mode)
+            source = im.im
+        else:
+            source = im
+
+        self._ensure_mutable()
+
+        if mask:
+            mask.load()
+            self.im.paste(source, box, mask.im)
+        else:
+            self.im.paste(source, box)
+
+    def alpha_composite(
+        self, im: Image, dest: Sequence[int] = (0, 0), source: Sequence[int] = (0, 0)
+    ) -> None:
+        """'In-place' analog of Image.alpha_composite. Composites an image
+        onto this image.
+
+        :param im: image to composite over this one
+        :param dest: Optional 2 tuple (left, top) specifying the upper
+          left corner in this (destination) image.
+        :param source: Optional 2 (left, top) tuple for the upper left
+          corner in the overlay source image, or 4 tuple (left, top, right,
+          bottom) for the bounds of the source rectangle
+
+        Performance Note: Not currently implemented in-place in the core layer.
+        """
+
+        if not isinstance(source, (list, tuple)):
+            msg = "Source must be a list or tuple"
+            raise ValueError(msg)
+        if not isinstance(dest, (list, tuple)):
+            msg = "Destination must be a list or tuple"
+            raise ValueError(msg)
+
+        if len(source) == 4:
+            overlay_crop_box = tuple(source)
+        elif len(source) == 2:
+            overlay_crop_box = tuple(source) + im.size
+        else:
+            msg = "Source must be a sequence of length 2 or 4"
+            raise ValueError(msg)
+
+        if not len(dest) == 2:
+            msg = "Destination must be a sequence of length 2"
+            raise ValueError(msg)
+        if min(source) < 0:
+            msg = "Source must be non-negative"
+            raise ValueError(msg)
+
+        # over image, crop if it's not the whole image.
+        if overlay_crop_box == (0, 0) + im.size:
+            overlay = im
+        else:
+            overlay = im.crop(overlay_crop_box)
+
+        # target for the paste
+        box = tuple(dest) + (dest[0] + overlay.width, dest[1] + overlay.height)
+
+        # destination image. don't copy if we're using the whole image.
+        if box == (0, 0) + self.size:
+            background = self
+        else:
+            background = self.crop(box)
+
+        result = alpha_composite(background, overlay)
+        self.paste(result, box)
+
+    def point(
+        self,
+        lut: (
+            Sequence[float]
+            | NumpyArray
+            | Callable[[int], float]
+            | Callable[[ImagePointTransform], ImagePointTransform | float]
+            | ImagePointHandler
+        ),
+        mode: str | None = None,
+    ) -> Image:
+        """
+        Maps this image through a lookup table or function.
+
+        :param lut: A lookup table, containing 256 (or 65536 if
+           self.mode=="I" and mode == "L") values per band in the
+           image.  A function can be used instead, it should take a
+           single argument. The function is called once for each
+           possible pixel value, and the resulting table is applied to
+           all bands of the image.
+
+           It may also be an :py:class:`~PIL.Image.ImagePointHandler`
+           object::
+
+               class Example(Image.ImagePointHandler):
+                 def point(self, im: Image) -> Image:
+                   # Return result
+        :param mode: Output mode (default is same as input). This can only be used if
+           the source image has mode "L" or "P", and the output has mode "1" or the
+           source image mode is "I" and the output mode is "L".
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        self.load()
+
+        if isinstance(lut, ImagePointHandler):
+            return lut.point(self)
+
+        if callable(lut):
+            # if it isn't a list, it should be a function
+            if self.mode in ("I", "I;16", "F"):
+                # check if the function can be used with point_transform
+                # UNDONE wiredfool -- I think this prevents us from ever doing
+                # a gamma function point transform on > 8bit images.
+                scale, offset = _getscaleoffset(lut)  # type: ignore[arg-type]
+                return self._new(self.im.point_transform(scale, offset))
+            # for other modes, convert the function to a table
+            flatLut = [lut(i) for i in range(256)] * self.im.bands  # type: ignore[arg-type]
+        else:
+            flatLut = lut
+
+        if self.mode == "F":
+            # FIXME: _imaging returns a confusing error message for this case
+            msg = "point operation not supported for this mode"
+            raise ValueError(msg)
+
+        if mode != "F":
+            flatLut = [round(i) for i in flatLut]
+        return self._new(self.im.point(flatLut, mode))
+
+    def putalpha(self, alpha: Image | int) -> None:
+        """
+        Adds or replaces the alpha layer in this image.  If the image
+        does not have an alpha layer, it's converted to "LA" or "RGBA".
+        The new layer must be either "L" or "1".
+
+        :param alpha: The new alpha layer.  This can either be an "L" or "1"
+           image having the same size as this image, or an integer.
+        """
+
+        self._ensure_mutable()
+
+        if self.mode not in ("LA", "PA", "RGBA"):
+            # attempt to promote self to a matching alpha mode
+            try:
+                mode = getmodebase(self.mode) + "A"
+                try:
+                    self.im.setmode(mode)
+                except (AttributeError, ValueError) as e:
+                    # do things the hard way
+                    im = self.im.convert(mode)
+                    if im.mode not in ("LA", "PA", "RGBA"):
+                        msg = "alpha channel could not be added"
+                        raise ValueError(msg) from e  # sanity check
+                    self.im = im
+                self._mode = self.im.mode
+            except KeyError as e:
+                msg = "illegal image mode"
+                raise ValueError(msg) from e
+
+        if self.mode in ("LA", "PA"):
+            band = 1
+        else:
+            band = 3
+
+        if isinstance(alpha, Image):
+            # alpha layer
+            if alpha.mode not in ("1", "L"):
+                msg = "illegal image mode"
+                raise ValueError(msg)
+            alpha.load()
+            if alpha.mode == "1":
+                alpha = alpha.convert("L")
+        else:
+            # constant alpha
+            try:
+                self.im.fillband(band, alpha)
+            except (AttributeError, ValueError):
+                # do things the hard way
+                alpha = new("L", self.size, alpha)
+            else:
+                return
+
+        self.im.putband(alpha.im, band)
+
+    def putdata(
+        self,
+        data: Sequence[float] | Sequence[Sequence[int]] | core.ImagingCore | NumpyArray,
+        scale: float = 1.0,
+        offset: float = 0.0,
+    ) -> None:
+        """
+        Copies pixel data from a flattened sequence object into the image. The
+        values should start at the upper left corner (0, 0), continue to the
+        end of the line, followed directly by the first value of the second
+        line, and so on. Data will be read until either the image or the
+        sequence ends. The scale and offset values are used to adjust the
+        sequence values: **pixel = value*scale + offset**.
+
+        :param data: A flattened sequence object.
+        :param scale: An optional scale value.  The default is 1.0.
+        :param offset: An optional offset value.  The default is 0.0.
+        """
+
+        self._ensure_mutable()
+
+        self.im.putdata(data, scale, offset)
+
+    def putpalette(
+        self,
+        data: ImagePalette.ImagePalette | bytes | Sequence[int],
+        rawmode: str = "RGB",
+    ) -> None:
+        """
+        Attaches a palette to this image.  The image must be a "P", "PA", "L"
+        or "LA" image.
+
+        The palette sequence must contain at most 256 colors, made up of one
+        integer value for each channel in the raw mode.
+        For example, if the raw mode is "RGB", then it can contain at most 768
+        values, made up of red, green and blue values for the corresponding pixel
+        index in the 256 colors.
+        If the raw mode is "RGBA", then it can contain at most 1024 values,
+        containing red, green, blue and alpha values.
+
+        Alternatively, an 8-bit string may be used instead of an integer sequence.
+
+        :param data: A palette sequence (either a list or a string).
+        :param rawmode: The raw mode of the palette. Either "RGB", "RGBA", or a mode
+           that can be transformed to "RGB" or "RGBA" (e.g. "R", "BGR;15", "RGBA;L").
+        """
+        from . import ImagePalette
+
+        if self.mode not in ("L", "LA", "P", "PA"):
+            msg = "illegal image mode"
+            raise ValueError(msg)
+        if isinstance(data, ImagePalette.ImagePalette):
+            if data.rawmode is not None:
+                palette = ImagePalette.raw(data.rawmode, data.palette)
+            else:
+                palette = ImagePalette.ImagePalette(palette=data.palette)
+                palette.dirty = 1
+        else:
+            if not isinstance(data, bytes):
+                data = bytes(data)
+            palette = ImagePalette.raw(rawmode, data)
+        self._mode = "PA" if "A" in self.mode else "P"
+        self.palette = palette
+        self.palette.mode = "RGBA" if "A" in rawmode else "RGB"
+        self.load()  # install new palette
+
+    def putpixel(
+        self, xy: tuple[int, int], value: float | tuple[int, ...] | list[int]
+    ) -> None:
+        """
+        Modifies the pixel at the given position. The color is given as
+        a single numerical value for single-band images, and a tuple for
+        multi-band images. In addition to this, RGB and RGBA tuples are
+        accepted for P and PA images.
+
+        Note that this method is relatively slow.  For more extensive changes,
+        use :py:meth:`~PIL.Image.Image.paste` or the :py:mod:`~PIL.ImageDraw`
+        module instead.
+
+        See:
+
+        * :py:meth:`~PIL.Image.Image.paste`
+        * :py:meth:`~PIL.Image.Image.putdata`
+        * :py:mod:`~PIL.ImageDraw`
+
+        :param xy: The pixel coordinate, given as (x, y). See
+           :ref:`coordinate-system`.
+        :param value: The pixel value.
+        """
+
+        if self.readonly:
+            self._copy()
+        self.load()
+
+        if (
+            self.mode in ("P", "PA")
+            and isinstance(value, (list, tuple))
+            and len(value) in [3, 4]
+        ):
+            # RGB or RGBA value for a P or PA image
+            if self.mode == "PA":
+                alpha = value[3] if len(value) == 4 else 255
+                value = value[:3]
+            assert self.palette is not None
+            palette_index = self.palette.getcolor(tuple(value), self)
+            value = (palette_index, alpha) if self.mode == "PA" else palette_index
+        return self.im.putpixel(xy, value)
+
+    def remap_palette(
+        self, dest_map: list[int], source_palette: bytes | bytearray | None = None
+    ) -> Image:
+        """
+        Rewrites the image to reorder the palette.
+
+        :param dest_map: A list of indexes into the original palette.
+           e.g. ``[1,0]`` would swap a two item palette, and ``list(range(256))``
+           is the identity transform.
+        :param source_palette: Bytes or None.
+        :returns:  An :py:class:`~PIL.Image.Image` object.
+
+        """
+        from . import ImagePalette
+
+        if self.mode not in ("L", "P"):
+            msg = "illegal image mode"
+            raise ValueError(msg)
+
+        bands = 3
+        palette_mode = "RGB"
+        if source_palette is None:
+            if self.mode == "P":
+                self.load()
+                palette_mode = self.im.getpalettemode()
+                if palette_mode == "RGBA":
+                    bands = 4
+                source_palette = self.im.getpalette(palette_mode, palette_mode)
+            else:  # L-mode
+                source_palette = bytearray(i // 3 for i in range(768))
+        elif len(source_palette) > 768:
+            bands = 4
+            palette_mode = "RGBA"
+
+        palette_bytes = b""
+        new_positions = [0] * 256
+
+        # pick only the used colors from the palette
+        for i, oldPosition in enumerate(dest_map):
+            palette_bytes += source_palette[
+                oldPosition * bands : oldPosition * bands + bands
+            ]
+            new_positions[oldPosition] = i
+
+        # replace the palette color id of all pixel with the new id
+
+        # Palette images are [0..255], mapped through a 1 or 3
+        # byte/color map.  We need to remap the whole image
+        # from palette 1 to palette 2. New_positions is
+        # an array of indexes into palette 1.  Palette 2 is
+        # palette 1 with any holes removed.
+
+        # We're going to leverage the convert mechanism to use the
+        # C code to remap the image from palette 1 to palette 2,
+        # by forcing the source image into 'L' mode and adding a
+        # mapping 'L' mode palette, then converting back to 'L'
+        # sans palette thus converting the image bytes, then
+        # assigning the optimized RGB palette.
+
+        # perf reference, 9500x4000 gif, w/~135 colors
+        # 14 sec prepatch, 1 sec postpatch with optimization forced.
+
+        mapping_palette = bytearray(new_positions)
+
+        m_im = self.copy()
+        m_im._mode = "P"
+
+        m_im.palette = ImagePalette.ImagePalette(
+            palette_mode, palette=mapping_palette * bands
+        )
+        # possibly set palette dirty, then
+        # m_im.putpalette(mapping_palette, 'L')  # converts to 'P'
+        # or just force it.
+        # UNDONE -- this is part of the general issue with palettes
+        m_im.im.putpalette(palette_mode, palette_mode + ";L", m_im.palette.tobytes())
+
+        m_im = m_im.convert("L")
+
+        m_im.putpalette(palette_bytes, palette_mode)
+        m_im.palette = ImagePalette.ImagePalette(palette_mode, palette=palette_bytes)
+
+        if "transparency" in self.info:
+            try:
+                m_im.info["transparency"] = dest_map.index(self.info["transparency"])
+            except ValueError:
+                if "transparency" in m_im.info:
+                    del m_im.info["transparency"]
+
+        return m_im
+
+    def _get_safe_box(
+        self,
+        size: tuple[int, int],
+        resample: Resampling,
+        box: tuple[float, float, float, float],
+    ) -> tuple[int, int, int, int]:
+        """Expands the box so it includes adjacent pixels
+        that may be used by resampling with the given resampling filter.
+        """
+        filter_support = _filters_support[resample] - 0.5
+        scale_x = (box[2] - box[0]) / size[0]
+        scale_y = (box[3] - box[1]) / size[1]
+        support_x = filter_support * scale_x
+        support_y = filter_support * scale_y
+
+        return (
+            max(0, int(box[0] - support_x)),
+            max(0, int(box[1] - support_y)),
+            min(self.size[0], math.ceil(box[2] + support_x)),
+            min(self.size[1], math.ceil(box[3] + support_y)),
+        )
+
+    def resize(
+        self,
+        size: tuple[int, int] | list[int] | NumpyArray,
+        resample: int | None = None,
+        box: tuple[float, float, float, float] | None = None,
+        reducing_gap: float | None = None,
+    ) -> Image:
+        """
+        Returns a resized copy of this image.
+
+        :param size: The requested size in pixels, as a tuple or array:
+           (width, height).
+        :param resample: An optional resampling filter.  This can be
+           one of :py:data:`Resampling.NEAREST`, :py:data:`Resampling.BOX`,
+           :py:data:`Resampling.BILINEAR`, :py:data:`Resampling.HAMMING`,
+           :py:data:`Resampling.BICUBIC` or :py:data:`Resampling.LANCZOS`.
+           If the image has mode "1" or "P", it is always set to
+           :py:data:`Resampling.NEAREST`. If the image mode is "BGR;15",
+           "BGR;16" or "BGR;24", then the default filter is
+           :py:data:`Resampling.NEAREST`. Otherwise, the default filter is
+           :py:data:`Resampling.BICUBIC`. See: :ref:`concept-filters`.
+        :param box: An optional 4-tuple of floats providing
+           the source image region to be scaled.
+           The values must be within (0, 0, width, height) rectangle.
+           If omitted or None, the entire source is used.
+        :param reducing_gap: Apply optimization by resizing the image
+           in two steps. First, reducing the image by integer times
+           using :py:meth:`~PIL.Image.Image.reduce`.
+           Second, resizing using regular resampling. The last step
+           changes size no less than by ``reducing_gap`` times.
+           ``reducing_gap`` may be None (no first step is performed)
+           or should be greater than 1.0. The bigger ``reducing_gap``,
+           the closer the result to the fair resampling.
+           The smaller ``reducing_gap``, the faster resizing.
+           With ``reducing_gap`` greater or equal to 3.0, the result is
+           indistinguishable from fair resampling in most cases.
+           The default value is None (no optimization).
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        if resample is None:
+            bgr = self.mode.startswith("BGR;")
+            resample = Resampling.NEAREST if bgr else Resampling.BICUBIC
+        elif resample not in (
+            Resampling.NEAREST,
+            Resampling.BILINEAR,
+            Resampling.BICUBIC,
+            Resampling.LANCZOS,
+            Resampling.BOX,
+            Resampling.HAMMING,
+        ):
+            msg = f"Unknown resampling filter ({resample})."
+
+            filters = [
+                f"{filter[1]} ({filter[0]})"
+                for filter in (
+                    (Resampling.NEAREST, "Image.Resampling.NEAREST"),
+                    (Resampling.LANCZOS, "Image.Resampling.LANCZOS"),
+                    (Resampling.BILINEAR, "Image.Resampling.BILINEAR"),
+                    (Resampling.BICUBIC, "Image.Resampling.BICUBIC"),
+                    (Resampling.BOX, "Image.Resampling.BOX"),
+                    (Resampling.HAMMING, "Image.Resampling.HAMMING"),
+                )
+            ]
+            msg += f" Use {', '.join(filters[:-1])} or {filters[-1]}"
+            raise ValueError(msg)
+
+        if reducing_gap is not None and reducing_gap < 1.0:
+            msg = "reducing_gap must be 1.0 or greater"
+            raise ValueError(msg)
+
+        if box is None:
+            box = (0, 0) + self.size
+
+        size = tuple(size)
+        if self.size == size and box == (0, 0) + self.size:
+            return self.copy()
+
+        if self.mode in ("1", "P"):
+            resample = Resampling.NEAREST
+
+        if self.mode in ["LA", "RGBA"] and resample != Resampling.NEAREST:
+            im = self.convert({"LA": "La", "RGBA": "RGBa"}[self.mode])
+            im = im.resize(size, resample, box)
+            return im.convert(self.mode)
+
+        self.load()
+
+        if reducing_gap is not None and resample != Resampling.NEAREST:
+            factor_x = int((box[2] - box[0]) / size[0] / reducing_gap) or 1
+            factor_y = int((box[3] - box[1]) / size[1] / reducing_gap) or 1
+            if factor_x > 1 or factor_y > 1:
+                reduce_box = self._get_safe_box(size, cast(Resampling, resample), box)
+                factor = (factor_x, factor_y)
+                self = (
+                    self.reduce(factor, box=reduce_box)
+                    if callable(self.reduce)
+                    else Image.reduce(self, factor, box=reduce_box)
+                )
+                box = (
+                    (box[0] - reduce_box[0]) / factor_x,
+                    (box[1] - reduce_box[1]) / factor_y,
+                    (box[2] - reduce_box[0]) / factor_x,
+                    (box[3] - reduce_box[1]) / factor_y,
+                )
+
+        return self._new(self.im.resize(size, resample, box))
+
+    def reduce(
+        self,
+        factor: int | tuple[int, int],
+        box: tuple[int, int, int, int] | None = None,
+    ) -> Image:
+        """
+        Returns a copy of the image reduced ``factor`` times.
+        If the size of the image is not dividable by ``factor``,
+        the resulting size will be rounded up.
+
+        :param factor: A greater than 0 integer or tuple of two integers
+           for width and height separately.
+        :param box: An optional 4-tuple of ints providing
+           the source image region to be reduced.
+           The values must be within ``(0, 0, width, height)`` rectangle.
+           If omitted or ``None``, the entire source is used.
+        """
+        if not isinstance(factor, (list, tuple)):
+            factor = (factor, factor)
+
+        if box is None:
+            box = (0, 0) + self.size
+
+        if factor == (1, 1) and box == (0, 0) + self.size:
+            return self.copy()
+
+        if self.mode in ["LA", "RGBA"]:
+            im = self.convert({"LA": "La", "RGBA": "RGBa"}[self.mode])
+            im = im.reduce(factor, box)
+            return im.convert(self.mode)
+
+        self.load()
+
+        return self._new(self.im.reduce(factor, box))
+
+    def rotate(
+        self,
+        angle: float,
+        resample: Resampling = Resampling.NEAREST,
+        expand: int | bool = False,
+        center: tuple[float, float] | None = None,
+        translate: tuple[int, int] | None = None,
+        fillcolor: float | tuple[float, ...] | str | None = None,
+    ) -> Image:
+        """
+        Returns a rotated copy of this image.  This method returns a
+        copy of this image, rotated the given number of degrees counter
+        clockwise around its centre.
+
+        :param angle: In degrees counter clockwise.
+        :param resample: An optional resampling filter.  This can be
+           one of :py:data:`Resampling.NEAREST` (use nearest neighbour),
+           :py:data:`Resampling.BILINEAR` (linear interpolation in a 2x2
+           environment), or :py:data:`Resampling.BICUBIC` (cubic spline
+           interpolation in a 4x4 environment). If omitted, or if the image has
+           mode "1" or "P", it is set to :py:data:`Resampling.NEAREST`.
+           See :ref:`concept-filters`.
+        :param expand: Optional expansion flag.  If true, expands the output
+           image to make it large enough to hold the entire rotated image.
+           If false or omitted, make the output image the same size as the
+           input image.  Note that the expand flag assumes rotation around
+           the center and no translation.
+        :param center: Optional center of rotation (a 2-tuple).  Origin is
+           the upper left corner.  Default is the center of the image.
+        :param translate: An optional post-rotate translation (a 2-tuple).
+        :param fillcolor: An optional color for area outside the rotated image.
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        angle = angle % 360.0
+
+        # Fast paths regardless of filter, as long as we're not
+        # translating or changing the center.
+        if not (center or translate):
+            if angle == 0:
+                return self.copy()
+            if angle == 180:
+                return self.transpose(Transpose.ROTATE_180)
+            if angle in (90, 270) and (expand or self.width == self.height):
+                return self.transpose(
+                    Transpose.ROTATE_90 if angle == 90 else Transpose.ROTATE_270
+                )
+
+        # Calculate the affine matrix.  Note that this is the reverse
+        # transformation (from destination image to source) because we
+        # want to interpolate the (discrete) destination pixel from
+        # the local area around the (floating) source pixel.
+
+        # The matrix we actually want (note that it operates from the right):
+        # (1, 0, tx)   (1, 0, cx)   ( cos a, sin a, 0)   (1, 0, -cx)
+        # (0, 1, ty) * (0, 1, cy) * (-sin a, cos a, 0) * (0, 1, -cy)
+        # (0, 0,  1)   (0, 0,  1)   (     0,     0, 1)   (0, 0,   1)
+
+        # The reverse matrix is thus:
+        # (1, 0, cx)   ( cos -a, sin -a, 0)   (1, 0, -cx)   (1, 0, -tx)
+        # (0, 1, cy) * (-sin -a, cos -a, 0) * (0, 1, -cy) * (0, 1, -ty)
+        # (0, 0,  1)   (      0,      0, 1)   (0, 0,   1)   (0, 0,   1)
+
+        # In any case, the final translation may be updated at the end to
+        # compensate for the expand flag.
+
+        w, h = self.size
+
+        if translate is None:
+            post_trans = (0, 0)
+        else:
+            post_trans = translate
+        if center is None:
+            center = (w / 2, h / 2)
+
+        angle = -math.radians(angle)
+        matrix = [
+            round(math.cos(angle), 15),
+            round(math.sin(angle), 15),
+            0.0,
+            round(-math.sin(angle), 15),
+            round(math.cos(angle), 15),
+            0.0,
+        ]
+
+        def transform(x: float, y: float, matrix: list[float]) -> tuple[float, float]:
+            (a, b, c, d, e, f) = matrix
+            return a * x + b * y + c, d * x + e * y + f
+
+        matrix[2], matrix[5] = transform(
+            -center[0] - post_trans[0], -center[1] - post_trans[1], matrix
+        )
+        matrix[2] += center[0]
+        matrix[5] += center[1]
+
+        if expand:
+            # calculate output size
+            xx = []
+            yy = []
+            for x, y in ((0, 0), (w, 0), (w, h), (0, h)):
+                transformed_x, transformed_y = transform(x, y, matrix)
+                xx.append(transformed_x)
+                yy.append(transformed_y)
+            nw = math.ceil(max(xx)) - math.floor(min(xx))
+            nh = math.ceil(max(yy)) - math.floor(min(yy))
+
+            # We multiply a translation matrix from the right.  Because of its
+            # special form, this is the same as taking the image of the
+            # translation vector as new translation vector.
+            matrix[2], matrix[5] = transform(-(nw - w) / 2.0, -(nh - h) / 2.0, matrix)
+            w, h = nw, nh
+
+        return self.transform(
+            (w, h), Transform.AFFINE, matrix, resample, fillcolor=fillcolor
+        )
+
+    def save(
+        self, fp: StrOrBytesPath | IO[bytes], format: str | None = None, **params: Any
+    ) -> None:
+        """
+        Saves this image under the given filename.  If no format is
+        specified, the format to use is determined from the filename
+        extension, if possible.
+
+        Keyword options can be used to provide additional instructions
+        to the writer. If a writer doesn't recognise an option, it is
+        silently ignored. The available options are described in the
+        :doc:`image format documentation
+        <../handbook/image-file-formats>` for each writer.
+
+        You can use a file object instead of a filename. In this case,
+        you must always specify the format. The file object must
+        implement the ``seek``, ``tell``, and ``write``
+        methods, and be opened in binary mode.
+
+        :param fp: A filename (string), os.PathLike object or file object.
+        :param format: Optional format override.  If omitted, the
+           format to use is determined from the filename extension.
+           If a file object was used instead of a filename, this
+           parameter should always be used.
+        :param params: Extra parameters to the image writer.
+        :returns: None
+        :exception ValueError: If the output format could not be determined
+           from the file name.  Use the format option to solve this.
+        :exception OSError: If the file could not be written.  The file
+           may have been created, and may contain partial data.
+        """
+
+        filename: str | bytes = ""
+        open_fp = False
+        if is_path(fp):
+            filename = os.fspath(fp)
+            open_fp = True
+        elif fp == sys.stdout:
+            try:
+                fp = sys.stdout.buffer
+            except AttributeError:
+                pass
+        if not filename and hasattr(fp, "name") and is_path(fp.name):
+            # only set the name for metadata purposes
+            filename = os.fspath(fp.name)
+
+        # may mutate self!
+        self._ensure_mutable()
+
+        save_all = params.pop("save_all", False)
+        self.encoderinfo = {**getattr(self, "encoderinfo", {}), **params}
+        self.encoderconfig: tuple[Any, ...] = ()
+
+        preinit()
+
+        filename_ext = os.path.splitext(filename)[1].lower()
+        ext = filename_ext.decode() if isinstance(filename_ext, bytes) else filename_ext
+
+        if not format:
+            if ext not in EXTENSION:
+                init()
+            try:
+                format = EXTENSION[ext]
+            except KeyError as e:
+                msg = f"unknown file extension: {ext}"
+                raise ValueError(msg) from e
+
+        if format.upper() not in SAVE:
+            init()
+        if save_all:
+            save_handler = SAVE_ALL[format.upper()]
+        else:
+            save_handler = SAVE[format.upper()]
+
+        created = False
+        if open_fp:
+            created = not os.path.exists(filename)
+            if params.get("append", False):
+                # Open also for reading ("+"), because TIFF save_all
+                # writer needs to go back and edit the written data.
+                fp = builtins.open(filename, "r+b")
+            else:
+                fp = builtins.open(filename, "w+b")
+        else:
+            fp = cast(IO[bytes], fp)
+
+        try:
+            save_handler(self, fp, filename)
+        except Exception:
+            if open_fp:
+                fp.close()
+            if created:
+                try:
+                    os.remove(filename)
+                except PermissionError:
+                    pass
+            raise
+        finally:
+            try:
+                del self.encoderinfo
+            except AttributeError:
+                pass
+        if open_fp:
+            fp.close()
+
+    def seek(self, frame: int) -> None:
+        """
+        Seeks to the given frame in this sequence file. If you seek
+        beyond the end of the sequence, the method raises an
+        ``EOFError`` exception. When a sequence file is opened, the
+        library automatically seeks to frame 0.
+
+        See :py:meth:`~PIL.Image.Image.tell`.
+
+        If defined, :attr:`~PIL.Image.Image.n_frames` refers to the
+        number of available frames.
+
+        :param frame: Frame number, starting at 0.
+        :exception EOFError: If the call attempts to seek beyond the end
+            of the sequence.
+        """
+
+        # overridden by file handlers
+        if frame != 0:
+            msg = "no more images in file"
+            raise EOFError(msg)
+
+    def show(self, title: str | None = None) -> None:
+        """
+        Displays this image. This method is mainly intended for debugging purposes.
+
+        This method calls :py:func:`PIL.ImageShow.show` internally. You can use
+        :py:func:`PIL.ImageShow.register` to override its default behaviour.
+
+        The image is first saved to a temporary file. By default, it will be in
+        PNG format.
+
+        On Unix, the image is then opened using the **xdg-open**, **display**,
+        **gm**, **eog** or **xv** utility, depending on which one can be found.
+
+        On macOS, the image is opened with the native Preview application.
+
+        On Windows, the image is opened with the standard PNG display utility.
+
+        :param title: Optional title to use for the image window, where possible.
+        """
+
+        _show(self, title=title)
+
+    def split(self) -> tuple[Image, ...]:
+        """
+        Split this image into individual bands. This method returns a
+        tuple of individual image bands from an image. For example,
+        splitting an "RGB" image creates three new images each
+        containing a copy of one of the original bands (red, green,
+        blue).
+
+        If you need only one band, :py:meth:`~PIL.Image.Image.getchannel`
+        method can be more convenient and faster.
+
+        :returns: A tuple containing bands.
+        """
+
+        self.load()
+        if self.im.bands == 1:
+            return (self.copy(),)
+        return tuple(map(self._new, self.im.split()))
+
+    def getchannel(self, channel: int | str) -> Image:
+        """
+        Returns an image containing a single channel of the source image.
+
+        :param channel: What channel to return. Could be index
+          (0 for "R" channel of "RGB") or channel name
+          ("A" for alpha channel of "RGBA").
+        :returns: An image in "L" mode.
+
+        .. versionadded:: 4.3.0
+        """
+        self.load()
+
+        if isinstance(channel, str):
+            try:
+                channel = self.getbands().index(channel)
+            except ValueError as e:
+                msg = f'The image has no channel "{channel}"'
+                raise ValueError(msg) from e
+
+        return self._new(self.im.getband(channel))
+
+    def tell(self) -> int:
+        """
+        Returns the current frame number. See :py:meth:`~PIL.Image.Image.seek`.
+
+        If defined, :attr:`~PIL.Image.Image.n_frames` refers to the
+        number of available frames.
+
+        :returns: Frame number, starting with 0.
+        """
+        return 0
+
+    def thumbnail(
+        self,
+        size: tuple[float, float],
+        resample: Resampling = Resampling.BICUBIC,
+        reducing_gap: float | None = 2.0,
+    ) -> None:
+        """
+        Make this image into a thumbnail.  This method modifies the
+        image to contain a thumbnail version of itself, no larger than
+        the given size.  This method calculates an appropriate thumbnail
+        size to preserve the aspect of the image, calls the
+        :py:meth:`~PIL.Image.Image.draft` method to configure the file reader
+        (where applicable), and finally resizes the image.
+
+        Note that this function modifies the :py:class:`~PIL.Image.Image`
+        object in place.  If you need to use the full resolution image as well,
+        apply this method to a :py:meth:`~PIL.Image.Image.copy` of the original
+        image.
+
+        :param size: The requested size in pixels, as a 2-tuple:
+           (width, height).
+        :param resample: Optional resampling filter.  This can be one
+           of :py:data:`Resampling.NEAREST`, :py:data:`Resampling.BOX`,
+           :py:data:`Resampling.BILINEAR`, :py:data:`Resampling.HAMMING`,
+           :py:data:`Resampling.BICUBIC` or :py:data:`Resampling.LANCZOS`.
+           If omitted, it defaults to :py:data:`Resampling.BICUBIC`.
+           (was :py:data:`Resampling.NEAREST` prior to version 2.5.0).
+           See: :ref:`concept-filters`.
+        :param reducing_gap: Apply optimization by resizing the image
+           in two steps. First, reducing the image by integer times
+           using :py:meth:`~PIL.Image.Image.reduce` or
+           :py:meth:`~PIL.Image.Image.draft` for JPEG images.
+           Second, resizing using regular resampling. The last step
+           changes size no less than by ``reducing_gap`` times.
+           ``reducing_gap`` may be None (no first step is performed)
+           or should be greater than 1.0. The bigger ``reducing_gap``,
+           the closer the result to the fair resampling.
+           The smaller ``reducing_gap``, the faster resizing.
+           With ``reducing_gap`` greater or equal to 3.0, the result is
+           indistinguishable from fair resampling in most cases.
+           The default value is 2.0 (very close to fair resampling
+           while still being faster in many cases).
+        :returns: None
+        """
+
+        provided_size = tuple(map(math.floor, size))
+
+        def preserve_aspect_ratio() -> tuple[int, int] | None:
+            def round_aspect(number: float, key: Callable[[int], float]) -> int:
+                return max(min(math.floor(number), math.ceil(number), key=key), 1)
+
+            x, y = provided_size
+            if x >= self.width and y >= self.height:
+                return None
+
+            aspect = self.width / self.height
+            if x / y >= aspect:
+                x = round_aspect(y * aspect, key=lambda n: abs(aspect - n / y))
+            else:
+                y = round_aspect(
+                    x / aspect, key=lambda n: 0 if n == 0 else abs(aspect - x / n)
+                )
+            return x, y
+
+        preserved_size = preserve_aspect_ratio()
+        if preserved_size is None:
+            return
+        final_size = preserved_size
+
+        box = None
+        if reducing_gap is not None:
+            res = self.draft(
+                None, (int(size[0] * reducing_gap), int(size[1] * reducing_gap))
+            )
+            if res is not None:
+                box = res[1]
+
+        if self.size != final_size:
+            im = self.resize(final_size, resample, box=box, reducing_gap=reducing_gap)
+
+            self.im = im.im
+            self._size = final_size
+            self._mode = self.im.mode
+
+        self.readonly = 0
+
+    # FIXME: the different transform methods need further explanation
+    # instead of bloating the method docs, add a separate chapter.
+    def transform(
+        self,
+        size: tuple[int, int],
+        method: Transform | ImageTransformHandler | SupportsGetData,
+        data: Sequence[Any] | None = None,
+        resample: int = Resampling.NEAREST,
+        fill: int = 1,
+        fillcolor: float | tuple[float, ...] | str | None = None,
+    ) -> Image:
+        """
+        Transforms this image.  This method creates a new image with the
+        given size, and the same mode as the original, and copies data
+        to the new image using the given transform.
+
+        :param size: The output size in pixels, as a 2-tuple:
+           (width, height).
+        :param method: The transformation method.  This is one of
+          :py:data:`Transform.EXTENT` (cut out a rectangular subregion),
+          :py:data:`Transform.AFFINE` (affine transform),
+          :py:data:`Transform.PERSPECTIVE` (perspective transform),
+          :py:data:`Transform.QUAD` (map a quadrilateral to a rectangle), or
+          :py:data:`Transform.MESH` (map a number of source quadrilaterals
+          in one operation).
+
+          It may also be an :py:class:`~PIL.Image.ImageTransformHandler`
+          object::
+
+            class Example(Image.ImageTransformHandler):
+                def transform(self, size, data, resample, fill=1):
+                    # Return result
+
+          Implementations of :py:class:`~PIL.Image.ImageTransformHandler`
+          for some of the :py:class:`Transform` methods are provided
+          in :py:mod:`~PIL.ImageTransform`.
+
+          It may also be an object with a ``method.getdata`` method
+          that returns a tuple supplying new ``method`` and ``data`` values::
+
+            class Example:
+                def getdata(self):
+                    method = Image.Transform.EXTENT
+                    data = (0, 0, 100, 100)
+                    return method, data
+        :param data: Extra data to the transformation method.
+        :param resample: Optional resampling filter.  It can be one of
+           :py:data:`Resampling.NEAREST` (use nearest neighbour),
+           :py:data:`Resampling.BILINEAR` (linear interpolation in a 2x2
+           environment), or :py:data:`Resampling.BICUBIC` (cubic spline
+           interpolation in a 4x4 environment). If omitted, or if the image
+           has mode "1" or "P", it is set to :py:data:`Resampling.NEAREST`.
+           See: :ref:`concept-filters`.
+        :param fill: If ``method`` is an
+          :py:class:`~PIL.Image.ImageTransformHandler` object, this is one of
+          the arguments passed to it. Otherwise, it is unused.
+        :param fillcolor: Optional fill color for the area outside the
+           transform in the output image.
+        :returns: An :py:class:`~PIL.Image.Image` object.
+        """
+
+        if self.mode in ("LA", "RGBA") and resample != Resampling.NEAREST:
+            return (
+                self.convert({"LA": "La", "RGBA": "RGBa"}[self.mode])
+                .transform(size, method, data, resample, fill, fillcolor)
+                .convert(self.mode)
+            )
+
+        if isinstance(method, ImageTransformHandler):
+            return method.transform(size, self, resample=resample, fill=fill)
+
+        if hasattr(method, "getdata"):
+            # compatibility w. old-style transform objects
+            method, data = method.getdata()
+
+        if data is None:
+            msg = "missing method data"
+            raise ValueError(msg)
+
+        im = new(self.mode, size, fillcolor)
+        if self.mode == "P" and self.palette:
+            im.palette = self.palette.copy()
+        im.info = self.info.copy()
+        if method == Transform.MESH:
+            # list of quads
+            for box, quad in data:
+                im.__transformer(
+                    box, self, Transform.QUAD, quad, resample, fillcolor is None
+                )
+        else:
+            im.__transformer(
+                (0, 0) + size, self, method, data, resample, fillcolor is None
+            )
+
+        return im
+
+    def __transformer(
+        self,
+        box: tuple[int, int, int, int],
+        image: Image,
+        method: Transform,
+        data: Sequence[float],
+        resample: int = Resampling.NEAREST,
+        fill: bool = True,
+    ) -> None:
+        w = box[2] - box[0]
+        h = box[3] - box[1]
+
+        if method == Transform.AFFINE:
+            data = data[:6]
+
+        elif method == Transform.EXTENT:
+            # convert extent to an affine transform
+            x0, y0, x1, y1 = data
+            xs = (x1 - x0) / w
+            ys = (y1 - y0) / h
+            method = Transform.AFFINE
+            data = (xs, 0, x0, 0, ys, y0)
+
+        elif method == Transform.PERSPECTIVE:
+            data = data[:8]
+
+        elif method == Transform.QUAD:
+            # quadrilateral warp.  data specifies the four corners
+            # given as NW, SW, SE, and NE.
+            nw = data[:2]
+            sw = data[2:4]
+            se = data[4:6]
+            ne = data[6:8]
+            x0, y0 = nw
+            As = 1.0 / w
+            At = 1.0 / h
+            data = (
+                x0,
+                (ne[0] - x0) * As,
+                (sw[0] - x0) * At,
+                (se[0] - sw[0] - ne[0] + x0) * As * At,
+                y0,
+                (ne[1] - y0) * As,
+                (sw[1] - y0) * At,
+                (se[1] - sw[1] - ne[1] + y0) * As * At,
+            )
+
+        else:
+            msg = "unknown transformation method"
+            raise ValueError(msg)
+
+        if resample not in (
+            Resampling.NEAREST,
+            Resampling.BILINEAR,
+            Resampling.BICUBIC,
+        ):
+            if resample in (Resampling.BOX, Resampling.HAMMING, Resampling.LANCZOS):
+                unusable: dict[int, str] = {
+                    Resampling.BOX: "Image.Resampling.BOX",
+                    Resampling.HAMMING: "Image.Resampling.HAMMING",
+                    Resampling.LANCZOS: "Image.Resampling.LANCZOS",
+                }
+                msg = unusable[resample] + f" ({resample}) cannot be used."
+            else:
+                msg = f"Unknown resampling filter ({resample})."
+
+            filters = [
+                f"{filter[1]} ({filter[0]})"
+                for filter in (
+                    (Resampling.NEAREST, "Image.Resampling.NEAREST"),
+                    (Resampling.BILINEAR, "Image.Resampling.BILINEAR"),
+                    (Resampling.BICUBIC, "Image.Resampling.BICUBIC"),
+                )
+            ]
+            msg += f" Use {', '.join(filters[:-1])} or {filters[-1]}"
+            raise ValueError(msg)
+
+        image.load()
+
+        self.load()
+
+        if image.mode in ("1", "P"):
+            resample = Resampling.NEAREST
+
+        self.im.transform(box, image.im, method, data, resample, fill)
+
+    def transpose(self, method: Transpose) -> Image:
+        """
+        Transpose image (flip or rotate in 90 degree steps)
+
+        :param method: One of :py:data:`Transpose.FLIP_LEFT_RIGHT`,
+          :py:data:`Transpose.FLIP_TOP_BOTTOM`, :py:data:`Transpose.ROTATE_90`,
+          :py:data:`Transpose.ROTATE_180`, :py:data:`Transpose.ROTATE_270`,
+          :py:data:`Transpose.TRANSPOSE` or :py:data:`Transpose.TRANSVERSE`.
+        :returns: Returns a flipped or rotated copy of this image.
+        """
+
+        self.load()
+        return self._new(self.im.transpose(method))
+
+    def effect_spread(self, distance: int) -> Image:
+        """
+        Randomly spread pixels in an image.
+
+        :param distance: Distance to spread pixels.
+        """
+        self.load()
+        return self._new(self.im.effect_spread(distance))
+
+    def toqimage(self) -> ImageQt.ImageQt:
+        """Returns a QImage copy of this image"""
+        from . import ImageQt
+
+        if not ImageQt.qt_is_installed:
+            msg = "Qt bindings are not installed"
+            raise ImportError(msg)
+        return ImageQt.toqimage(self)
+
+    def toqpixmap(self) -> ImageQt.QPixmap:
+        """Returns a QPixmap copy of this image"""
+        from . import ImageQt
+
+        if not ImageQt.qt_is_installed:
+            msg = "Qt bindings are not installed"
+            raise ImportError(msg)
+        return ImageQt.toqpixmap(self)
+
+
+# --------------------------------------------------------------------
+# Abstract handlers.
+
+
+class ImagePointHandler:
+    """
+    Used as a mixin by point transforms
+    (for use with :py:meth:`~PIL.Image.Image.point`)
+    """
+
+    @abc.abstractmethod
+    def point(self, im: Image) -> Image:
+        pass
+
+
+class ImageTransformHandler:
+    """
+    Used as a mixin by geometry transforms
+    (for use with :py:meth:`~PIL.Image.Image.transform`)
+    """
+
+    @abc.abstractmethod
+    def transform(
+        self,
+        size: tuple[int, int],
+        image: Image,
+        **options: Any,
+    ) -> Image:
+        pass
+
+
+# --------------------------------------------------------------------
+# Factories
+
+#
+# Debugging
+
+
+def _wedge() -> Image:
+    """Create grayscale wedge (for debugging only)"""
+
+    return Image()._new(core.wedge("L"))
+
+
+def _check_size(size: Any) -> None:
+    """
+    Common check to enforce type and sanity check on size tuples
+
+    :param size: Should be a 2 tuple of (width, height)
+    :returns: None, or raises a ValueError
+    """
+
+    if not isinstance(size, (list, tuple)):
+        msg = "Size must be a list or tuple"
+        raise ValueError(msg)
+    if len(size) != 2:
+        msg = "Size must be a sequence of length 2"
+        raise ValueError(msg)
+    if size[0] < 0 or size[1] < 0:
+        msg = "Width and height must be >= 0"
+        raise ValueError(msg)
+
+
+def new(
+    mode: str,
+    size: tuple[int, int] | list[int],
+    color: float | tuple[float, ...] | str | None = 0,
+) -> Image:
+    """
+    Creates a new image with the given mode and size.
+
+    :param mode: The mode to use for the new image. See:
+       :ref:`concept-modes`.
+    :param size: A 2-tuple, containing (width, height) in pixels.
+    :param color: What color to use for the image.  Default is black.
+       If given, this should be a single integer or floating point value
+       for single-band modes, and a tuple for multi-band modes (one value
+       per band).  When creating RGB or HSV images, you can also use color
+       strings as supported by the ImageColor module.  If the color is
+       None, the image is not initialised.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    if mode in ("BGR;15", "BGR;16", "BGR;24"):
+        deprecate(mode, 12)
+
+    _check_size(size)
+
+    if color is None:
+        # don't initialize
+        return Image()._new(core.new(mode, size))
+
+    if isinstance(color, str):
+        # css3-style specifier
+
+        from . import ImageColor
+
+        color = ImageColor.getcolor(color, mode)
+
+    im = Image()
+    if (
+        mode == "P"
+        and isinstance(color, (list, tuple))
+        and all(isinstance(i, int) for i in color)
+    ):
+        color_ints: tuple[int, ...] = cast(tuple[int, ...], tuple(color))
+        if len(color_ints) == 3 or len(color_ints) == 4:
+            # RGB or RGBA value for a P image
+            from . import ImagePalette
+
+            im.palette = ImagePalette.ImagePalette()
+            color = im.palette.getcolor(color_ints)
+    return im._new(core.fill(mode, size, color))
+
+
+def frombytes(
+    mode: str,
+    size: tuple[int, int],
+    data: bytes | bytearray | SupportsArrayInterface,
+    decoder_name: str = "raw",
+    *args: Any,
+) -> Image:
+    """
+    Creates a copy of an image memory from pixel data in a buffer.
+
+    In its simplest form, this function takes three arguments
+    (mode, size, and unpacked pixel data).
+
+    You can also use any pixel decoder supported by PIL. For more
+    information on available decoders, see the section
+    :ref:`Writing Your Own File Codec <file-codecs>`.
+
+    Note that this function decodes pixel data only, not entire images.
+    If you have an entire image in a string, wrap it in a
+    :py:class:`~io.BytesIO` object, and use :py:func:`~PIL.Image.open` to load
+    it.
+
+    :param mode: The image mode. See: :ref:`concept-modes`.
+    :param size: The image size.
+    :param data: A byte buffer containing raw data for the given mode.
+    :param decoder_name: What decoder to use.
+    :param args: Additional parameters for the given decoder.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    _check_size(size)
+
+    im = new(mode, size)
+    if im.width != 0 and im.height != 0:
+        decoder_args: Any = args
+        if len(decoder_args) == 1 and isinstance(decoder_args[0], tuple):
+            # may pass tuple instead of argument list
+            decoder_args = decoder_args[0]
+
+        if decoder_name == "raw" and decoder_args == ():
+            decoder_args = mode
+
+        im.frombytes(data, decoder_name, decoder_args)
+    return im
+
+
+def frombuffer(
+    mode: str,
+    size: tuple[int, int],
+    data: bytes | SupportsArrayInterface,
+    decoder_name: str = "raw",
+    *args: Any,
+) -> Image:
+    """
+    Creates an image memory referencing pixel data in a byte buffer.
+
+    This function is similar to :py:func:`~PIL.Image.frombytes`, but uses data
+    in the byte buffer, where possible.  This means that changes to the
+    original buffer object are reflected in this image).  Not all modes can
+    share memory; supported modes include "L", "RGBX", "RGBA", and "CMYK".
+
+    Note that this function decodes pixel data only, not entire images.
+    If you have an entire image file in a string, wrap it in a
+    :py:class:`~io.BytesIO` object, and use :py:func:`~PIL.Image.open` to load it.
+
+    The default parameters used for the "raw" decoder differs from that used for
+    :py:func:`~PIL.Image.frombytes`. This is a bug, and will probably be fixed in a
+    future release. The current release issues a warning if you do this; to disable
+    the warning, you should provide the full set of parameters. See below for details.
+
+    :param mode: The image mode. See: :ref:`concept-modes`.
+    :param size: The image size.
+    :param data: A bytes or other buffer object containing raw
+        data for the given mode.
+    :param decoder_name: What decoder to use.
+    :param args: Additional parameters for the given decoder.  For the
+        default encoder ("raw"), it's recommended that you provide the
+        full set of parameters::
+
+            frombuffer(mode, size, data, "raw", mode, 0, 1)
+
+    :returns: An :py:class:`~PIL.Image.Image` object.
+
+    .. versionadded:: 1.1.4
+    """
+
+    _check_size(size)
+
+    # may pass tuple instead of argument list
+    if len(args) == 1 and isinstance(args[0], tuple):
+        args = args[0]
+
+    if decoder_name == "raw":
+        if args == ():
+            args = mode, 0, 1
+        if args[0] in _MAPMODES:
+            im = new(mode, (0, 0))
+            im = im._new(core.map_buffer(data, size, decoder_name, 0, args))
+            if mode == "P":
+                from . import ImagePalette
+
+                im.palette = ImagePalette.ImagePalette("RGB", im.im.getpalette("RGB"))
+            im.readonly = 1
+            return im
+
+    return frombytes(mode, size, data, decoder_name, args)
+
+
+class SupportsArrayInterface(Protocol):
+    """
+    An object that has an ``__array_interface__`` dictionary.
+    """
+
+    @property
+    def __array_interface__(self) -> dict[str, Any]:
+        raise NotImplementedError()
+
+
+def fromarray(obj: SupportsArrayInterface, mode: str | None = None) -> Image:
+    """
+    Creates an image memory from an object exporting the array interface
+    (using the buffer protocol)::
+
+      from PIL import Image
+      import numpy as np
+      a = np.zeros((5, 5))
+      im = Image.fromarray(a)
+
+    If ``obj`` is not contiguous, then the ``tobytes`` method is called
+    and :py:func:`~PIL.Image.frombuffer` is used.
+
+    In the case of NumPy, be aware that Pillow modes do not always correspond
+    to NumPy dtypes. Pillow modes only offer 1-bit pixels, 8-bit pixels,
+    32-bit signed integer pixels, and 32-bit floating point pixels.
+
+    Pillow images can also be converted to arrays::
+
+      from PIL import Image
+      import numpy as np
+      im = Image.open("hopper.jpg")
+      a = np.asarray(im)
+
+    When converting Pillow images to arrays however, only pixel values are
+    transferred. This means that P and PA mode images will lose their palette.
+
+    :param obj: Object with array interface
+    :param mode: Optional mode to use when reading ``obj``. Will be determined from
+      type if ``None``.
+
+      This will not be used to convert the data after reading, but will be used to
+      change how the data is read::
+
+        from PIL import Image
+        import numpy as np
+        a = np.full((1, 1), 300)
+        im = Image.fromarray(a, mode="L")
+        im.getpixel((0, 0))  # 44
+        im = Image.fromarray(a, mode="RGB")
+        im.getpixel((0, 0))  # (44, 1, 0)
+
+      See: :ref:`concept-modes` for general information about modes.
+    :returns: An image object.
+
+    .. versionadded:: 1.1.6
+    """
+    arr = obj.__array_interface__
+    shape = arr["shape"]
+    ndim = len(shape)
+    strides = arr.get("strides", None)
+    if mode is None:
+        try:
+            typekey = (1, 1) + shape[2:], arr["typestr"]
+        except KeyError as e:
+            msg = "Cannot handle this data type"
+            raise TypeError(msg) from e
+        try:
+            mode, rawmode = _fromarray_typemap[typekey]
+        except KeyError as e:
+            typekey_shape, typestr = typekey
+            msg = f"Cannot handle this data type: {typekey_shape}, {typestr}"
+            raise TypeError(msg) from e
+    else:
+        rawmode = mode
+    if mode in ["1", "L", "I", "P", "F"]:
+        ndmax = 2
+    elif mode == "RGB":
+        ndmax = 3
+    else:
+        ndmax = 4
+    if ndim > ndmax:
+        msg = f"Too many dimensions: {ndim} > {ndmax}."
+        raise ValueError(msg)
+
+    size = 1 if ndim == 1 else shape[1], shape[0]
+    if strides is not None:
+        if hasattr(obj, "tobytes"):
+            obj = obj.tobytes()
+        elif hasattr(obj, "tostring"):
+            obj = obj.tostring()
+        else:
+            msg = "'strides' requires either tobytes() or tostring()"
+            raise ValueError(msg)
+
+    return frombuffer(mode, size, obj, "raw", rawmode, 0, 1)
+
+
+def fromqimage(im: ImageQt.QImage) -> ImageFile.ImageFile:
+    """Creates an image instance from a QImage image"""
+    from . import ImageQt
+
+    if not ImageQt.qt_is_installed:
+        msg = "Qt bindings are not installed"
+        raise ImportError(msg)
+    return ImageQt.fromqimage(im)
+
+
+def fromqpixmap(im: ImageQt.QPixmap) -> ImageFile.ImageFile:
+    """Creates an image instance from a QPixmap image"""
+    from . import ImageQt
+
+    if not ImageQt.qt_is_installed:
+        msg = "Qt bindings are not installed"
+        raise ImportError(msg)
+    return ImageQt.fromqpixmap(im)
+
+
+_fromarray_typemap = {
+    # (shape, typestr) => mode, rawmode
+    # first two members of shape are set to one
+    ((1, 1), "|b1"): ("1", "1;8"),
+    ((1, 1), "|u1"): ("L", "L"),
+    ((1, 1), "|i1"): ("I", "I;8"),
+    ((1, 1), "<u2"): ("I", "I;16"),
+    ((1, 1), ">u2"): ("I", "I;16B"),
+    ((1, 1), "<i2"): ("I", "I;16S"),
+    ((1, 1), ">i2"): ("I", "I;16BS"),
+    ((1, 1), "<u4"): ("I", "I;32"),
+    ((1, 1), ">u4"): ("I", "I;32B"),
+    ((1, 1), "<i4"): ("I", "I;32S"),
+    ((1, 1), ">i4"): ("I", "I;32BS"),
+    ((1, 1), "<f4"): ("F", "F;32F"),
+    ((1, 1), ">f4"): ("F", "F;32BF"),
+    ((1, 1), "<f8"): ("F", "F;64F"),
+    ((1, 1), ">f8"): ("F", "F;64BF"),
+    ((1, 1, 2), "|u1"): ("LA", "LA"),
+    ((1, 1, 3), "|u1"): ("RGB", "RGB"),
+    ((1, 1, 4), "|u1"): ("RGBA", "RGBA"),
+    # shortcuts:
+    ((1, 1), f"{_ENDIAN}i4"): ("I", "I"),
+    ((1, 1), f"{_ENDIAN}f4"): ("F", "F"),
+}
+
+
+def _decompression_bomb_check(size: tuple[int, int]) -> None:
+    if MAX_IMAGE_PIXELS is None:
+        return
+
+    pixels = max(1, size[0]) * max(1, size[1])
+
+    if pixels > 2 * MAX_IMAGE_PIXELS:
+        msg = (
+            f"Image size ({pixels} pixels) exceeds limit of {2 * MAX_IMAGE_PIXELS} "
+            "pixels, could be decompression bomb DOS attack."
+        )
+        raise DecompressionBombError(msg)
+
+    if pixels > MAX_IMAGE_PIXELS:
+        warnings.warn(
+            f"Image size ({pixels} pixels) exceeds limit of {MAX_IMAGE_PIXELS} pixels, "
+            "could be decompression bomb DOS attack.",
+            DecompressionBombWarning,
+        )
+
+
+def open(
+    fp: StrOrBytesPath | IO[bytes],
+    mode: Literal["r"] = "r",
+    formats: list[str] | tuple[str, ...] | None = None,
+) -> ImageFile.ImageFile:
+    """
+    Opens and identifies the given image file.
+
+    This is a lazy operation; this function identifies the file, but
+    the file remains open and the actual image data is not read from
+    the file until you try to process the data (or call the
+    :py:meth:`~PIL.Image.Image.load` method).  See
+    :py:func:`~PIL.Image.new`. See :ref:`file-handling`.
+
+    :param fp: A filename (string), os.PathLike object or a file object.
+       The file object must implement ``file.read``,
+       ``file.seek``, and ``file.tell`` methods,
+       and be opened in binary mode. The file object will also seek to zero
+       before reading.
+    :param mode: The mode.  If given, this argument must be "r".
+    :param formats: A list or tuple of formats to attempt to load the file in.
+       This can be used to restrict the set of formats checked.
+       Pass ``None`` to try all supported formats. You can print the set of
+       available formats by running ``python3 -m PIL`` or using
+       the :py:func:`PIL.features.pilinfo` function.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    :exception FileNotFoundError: If the file cannot be found.
+    :exception PIL.UnidentifiedImageError: If the image cannot be opened and
+       identified.
+    :exception ValueError: If the ``mode`` is not "r", or if a ``StringIO``
+       instance is used for ``fp``.
+    :exception TypeError: If ``formats`` is not ``None``, a list or a tuple.
+    """
+
+    if mode != "r":
+        msg = f"bad mode {repr(mode)}"  # type: ignore[unreachable]
+        raise ValueError(msg)
+    elif isinstance(fp, io.StringIO):
+        msg = (  # type: ignore[unreachable]
+            "StringIO cannot be used to open an image. "
+            "Binary data must be used instead."
+        )
+        raise ValueError(msg)
+
+    if formats is None:
+        formats = ID
+    elif not isinstance(formats, (list, tuple)):
+        msg = "formats must be a list or tuple"  # type: ignore[unreachable]
+        raise TypeError(msg)
+
+    exclusive_fp = False
+    filename: str | bytes = ""
+    if is_path(fp):
+        filename = os.fspath(fp)
+
+    if filename:
+        fp = builtins.open(filename, "rb")
+        exclusive_fp = True
+    else:
+        fp = cast(IO[bytes], fp)
+
+    try:
+        fp.seek(0)
+    except (AttributeError, io.UnsupportedOperation):
+        fp = io.BytesIO(fp.read())
+        exclusive_fp = True
+
+    prefix = fp.read(16)
+
+    preinit()
+
+    warning_messages: list[str] = []
+
+    def _open_core(
+        fp: IO[bytes],
+        filename: str | bytes,
+        prefix: bytes,
+        formats: list[str] | tuple[str, ...],
+    ) -> ImageFile.ImageFile | None:
+        for i in formats:
+            i = i.upper()
+            if i not in OPEN:
+                init()
+            try:
+                factory, accept = OPEN[i]
+                result = not accept or accept(prefix)
+                if isinstance(result, str):
+                    warning_messages.append(result)
+                elif result:
+                    fp.seek(0)
+                    im = factory(fp, filename)
+                    _decompression_bomb_check(im.size)
+                    return im
+            except (SyntaxError, IndexError, TypeError, struct.error) as e:
+                if WARN_POSSIBLE_FORMATS:
+                    warning_messages.append(i + " opening failed. " + str(e))
+            except BaseException:
+                if exclusive_fp:
+                    fp.close()
+                raise
+        return None
+
+    im = _open_core(fp, filename, prefix, formats)
+
+    if im is None and formats is ID:
+        checked_formats = ID.copy()
+        if init():
+            im = _open_core(
+                fp,
+                filename,
+                prefix,
+                tuple(format for format in formats if format not in checked_formats),
+            )
+
+    if im:
+        im._exclusive_fp = exclusive_fp
+        return im
+
+    if exclusive_fp:
+        fp.close()
+    for message in warning_messages:
+        warnings.warn(message)
+    msg = "cannot identify image file %r" % (filename if filename else fp)
+    raise UnidentifiedImageError(msg)
+
+
+#
+# Image processing.
+
+
+def alpha_composite(im1: Image, im2: Image) -> Image:
+    """
+    Alpha composite im2 over im1.
+
+    :param im1: The first image. Must have mode RGBA.
+    :param im2: The second image.  Must have mode RGBA, and the same size as
+       the first image.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    im1.load()
+    im2.load()
+    return im1._new(core.alpha_composite(im1.im, im2.im))
+
+
+def blend(im1: Image, im2: Image, alpha: float) -> Image:
+    """
+    Creates a new image by interpolating between two input images, using
+    a constant alpha::
+
+        out = image1 * (1.0 - alpha) + image2 * alpha
+
+    :param im1: The first image.
+    :param im2: The second image.  Must have the same mode and size as
+       the first image.
+    :param alpha: The interpolation alpha factor.  If alpha is 0.0, a
+       copy of the first image is returned. If alpha is 1.0, a copy of
+       the second image is returned. There are no restrictions on the
+       alpha value. If necessary, the result is clipped to fit into
+       the allowed output range.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    im1.load()
+    im2.load()
+    return im1._new(core.blend(im1.im, im2.im, alpha))
+
+
+def composite(image1: Image, image2: Image, mask: Image) -> Image:
+    """
+    Create composite image by blending images using a transparency mask.
+
+    :param image1: The first image.
+    :param image2: The second image.  Must have the same mode and
+       size as the first image.
+    :param mask: A mask image.  This image can have mode
+       "1", "L", or "RGBA", and must have the same size as the
+       other two images.
+    """
+
+    image = image2.copy()
+    image.paste(image1, None, mask)
+    return image
+
+
+def eval(image: Image, *args: Callable[[int], float]) -> Image:
+    """
+    Applies the function (which should take one argument) to each pixel
+    in the given image. If the image has more than one band, the same
+    function is applied to each band. Note that the function is
+    evaluated once for each possible pixel value, so you cannot use
+    random components or other generators.
+
+    :param image: The input image.
+    :param function: A function object, taking one integer argument.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    return image.point(args[0])
+
+
+def merge(mode: str, bands: Sequence[Image]) -> Image:
+    """
+    Merge a set of single band images into a new multiband image.
+
+    :param mode: The mode to use for the output image. See:
+        :ref:`concept-modes`.
+    :param bands: A sequence containing one single-band image for
+        each band in the output image.  All bands must have the
+        same size.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+
+    if getmodebands(mode) != len(bands) or "*" in mode:
+        msg = "wrong number of bands"
+        raise ValueError(msg)
+    for band in bands[1:]:
+        if band.mode != getmodetype(mode):
+            msg = "mode mismatch"
+            raise ValueError(msg)
+        if band.size != bands[0].size:
+            msg = "size mismatch"
+            raise ValueError(msg)
+    for band in bands:
+        band.load()
+    return bands[0]._new(core.merge(mode, *[b.im for b in bands]))
+
+
+# --------------------------------------------------------------------
+# Plugin registry
+
+
+def register_open(
+    id: str,
+    factory: (
+        Callable[[IO[bytes], str | bytes], ImageFile.ImageFile]
+        | type[ImageFile.ImageFile]
+    ),
+    accept: Callable[[bytes], bool | str] | None = None,
+) -> None:
+    """
+    Register an image file plugin.  This function should not be used
+    in application code.
+
+    :param id: An image format identifier.
+    :param factory: An image file factory method.
+    :param accept: An optional function that can be used to quickly
+       reject images having another format.
+    """
+    id = id.upper()
+    if id not in ID:
+        ID.append(id)
+    OPEN[id] = factory, accept
+
+
+def register_mime(id: str, mimetype: str) -> None:
+    """
+    Registers an image MIME type by populating ``Image.MIME``. This function
+    should not be used in application code.
+
+    ``Image.MIME`` provides a mapping from image format identifiers to mime
+    formats, but :py:meth:`~PIL.ImageFile.ImageFile.get_format_mimetype` can
+    provide a different result for specific images.
+
+    :param id: An image format identifier.
+    :param mimetype: The image MIME type for this format.
+    """
+    MIME[id.upper()] = mimetype
+
+
+def register_save(
+    id: str, driver: Callable[[Image, IO[bytes], str | bytes], None]
+) -> None:
+    """
+    Registers an image save function.  This function should not be
+    used in application code.
+
+    :param id: An image format identifier.
+    :param driver: A function to save images in this format.
+    """
+    SAVE[id.upper()] = driver
+
+
+def register_save_all(
+    id: str, driver: Callable[[Image, IO[bytes], str | bytes], None]
+) -> None:
+    """
+    Registers an image function to save all the frames
+    of a multiframe format.  This function should not be
+    used in application code.
+
+    :param id: An image format identifier.
+    :param driver: A function to save images in this format.
+    """
+    SAVE_ALL[id.upper()] = driver
+
+
+def register_extension(id: str, extension: str) -> None:
+    """
+    Registers an image extension.  This function should not be
+    used in application code.
+
+    :param id: An image format identifier.
+    :param extension: An extension used for this format.
+    """
+    EXTENSION[extension.lower()] = id.upper()
+
+
+def register_extensions(id: str, extensions: list[str]) -> None:
+    """
+    Registers image extensions.  This function should not be
+    used in application code.
+
+    :param id: An image format identifier.
+    :param extensions: A list of extensions used for this format.
+    """
+    for extension in extensions:
+        register_extension(id, extension)
+
+
+def registered_extensions() -> dict[str, str]:
+    """
+    Returns a dictionary containing all file extensions belonging
+    to registered plugins
+    """
+    init()
+    return EXTENSION
+
+
+def register_decoder(name: str, decoder: type[ImageFile.PyDecoder]) -> None:
+    """
+    Registers an image decoder.  This function should not be
+    used in application code.
+
+    :param name: The name of the decoder
+    :param decoder: An ImageFile.PyDecoder object
+
+    .. versionadded:: 4.1.0
+    """
+    DECODERS[name] = decoder
+
+
+def register_encoder(name: str, encoder: type[ImageFile.PyEncoder]) -> None:
+    """
+    Registers an image encoder.  This function should not be
+    used in application code.
+
+    :param name: The name of the encoder
+    :param encoder: An ImageFile.PyEncoder object
+
+    .. versionadded:: 4.1.0
+    """
+    ENCODERS[name] = encoder
+
+
+# --------------------------------------------------------------------
+# Simple display support.
+
+
+def _show(image: Image, **options: Any) -> None:
+    from . import ImageShow
+
+    ImageShow.show(image, **options)
+
+
+# --------------------------------------------------------------------
+# Effects
+
+
+def effect_mandelbrot(
+    size: tuple[int, int], extent: tuple[float, float, float, float], quality: int
+) -> Image:
+    """
+    Generate a Mandelbrot set covering the given extent.
+
+    :param size: The requested size in pixels, as a 2-tuple:
+       (width, height).
+    :param extent: The extent to cover, as a 4-tuple:
+       (x0, y0, x1, y1).
+    :param quality: Quality.
+    """
+    return Image()._new(core.effect_mandelbrot(size, extent, quality))
+
+
+def effect_noise(size: tuple[int, int], sigma: float) -> Image:
+    """
+    Generate Gaussian noise centered around 128.
+
+    :param size: The requested size in pixels, as a 2-tuple:
+       (width, height).
+    :param sigma: Standard deviation of noise.
+    """
+    return Image()._new(core.effect_noise(size, sigma))
+
+
+def linear_gradient(mode: str) -> Image:
+    """
+    Generate 256x256 linear gradient from black to white, top to bottom.
+
+    :param mode: Input mode.
+    """
+    return Image()._new(core.linear_gradient(mode))
+
+
+def radial_gradient(mode: str) -> Image:
+    """
+    Generate 256x256 radial gradient from black to white, centre to edge.
+
+    :param mode: Input mode.
+    """
+    return Image()._new(core.radial_gradient(mode))
+
+
+# --------------------------------------------------------------------
+# Resources
+
+
+def _apply_env_variables(env: dict[str, str] | None = None) -> None:
+    env_dict = env if env is not None else os.environ
+
+    for var_name, setter in [
+        ("PILLOW_ALIGNMENT", core.set_alignment),
+        ("PILLOW_BLOCK_SIZE", core.set_block_size),
+        ("PILLOW_BLOCKS_MAX", core.set_blocks_max),
+    ]:
+        if var_name not in env_dict:
+            continue
+
+        var = env_dict[var_name].lower()
+
+        units = 1
+        for postfix, mul in [("k", 1024), ("m", 1024 * 1024)]:
+            if var.endswith(postfix):
+                units = mul
+                var = var[: -len(postfix)]
+
+        try:
+            var_int = int(var) * units
+        except ValueError:
+            warnings.warn(f"{var_name} is not int")
+            continue
+
+        try:
+            setter(var_int)
+        except ValueError as e:
+            warnings.warn(f"{var_name}: {e}")
+
+
+_apply_env_variables()
+atexit.register(core.clear_cache)
+
+
+if TYPE_CHECKING:
+    _ExifBase = MutableMapping[int, Any]
+else:
+    _ExifBase = MutableMapping
+
+
+class Exif(_ExifBase):
+    """
+    This class provides read and write access to EXIF image data::
+
+      from PIL import Image
+      im = Image.open("exif.png")
+      exif = im.getexif()  # Returns an instance of this class
+
+    Information can be read and written, iterated over or deleted::
+
+      print(exif[274])  # 1
+      exif[274] = 2
+      for k, v in exif.items():
+        print("Tag", k, "Value", v)  # Tag 274 Value 2
+      del exif[274]
+
+    To access information beyond IFD0, :py:meth:`~PIL.Image.Exif.get_ifd`
+    returns a dictionary::
+
+      from PIL import ExifTags
+      im = Image.open("exif_gps.jpg")
+      exif = im.getexif()
+      gps_ifd = exif.get_ifd(ExifTags.IFD.GPSInfo)
+      print(gps_ifd)
+
+    Other IFDs include ``ExifTags.IFD.Exif``, ``ExifTags.IFD.MakerNote``,
+    ``ExifTags.IFD.Interop`` and ``ExifTags.IFD.IFD1``.
+
+    :py:mod:`~PIL.ExifTags` also has enum classes to provide names for data::
+
+      print(exif[ExifTags.Base.Software])  # PIL
+      print(gps_ifd[ExifTags.GPS.GPSDateStamp])  # 1999:99:99 99:99:99
+    """
+
+    endian: str | None = None
+    bigtiff = False
+    _loaded = False
+
+    def __init__(self) -> None:
+        self._data: dict[int, Any] = {}
+        self._hidden_data: dict[int, Any] = {}
+        self._ifds: dict[int, dict[int, Any]] = {}
+        self._info: TiffImagePlugin.ImageFileDirectory_v2 | None = None
+        self._loaded_exif: bytes | None = None
+
+    def _fixup(self, value: Any) -> Any:
+        try:
+            if len(value) == 1 and isinstance(value, tuple):
+                return value[0]
+        except Exception:
+            pass
+        return value
+
+    def _fixup_dict(self, src_dict: dict[int, Any]) -> dict[int, Any]:
+        # Helper function
+        # returns a dict with any single item tuples/lists as individual values
+        return {k: self._fixup(v) for k, v in src_dict.items()}
+
+    def _get_ifd_dict(
+        self, offset: int, group: int | None = None
+    ) -> dict[int, Any] | None:
+        try:
+            # an offset pointer to the location of the nested embedded IFD.
+            # It should be a long, but may be corrupted.
+            self.fp.seek(offset)
+        except (KeyError, TypeError):
+            return None
+        else:
+            from . import TiffImagePlugin
+
+            info = TiffImagePlugin.ImageFileDirectory_v2(self.head, group=group)
+            info.load(self.fp)
+            return self._fixup_dict(dict(info))
+
+    def _get_head(self) -> bytes:
+        version = b"\x2B" if self.bigtiff else b"\x2A"
+        if self.endian == "<":
+            head = b"II" + version + b"\x00" + o32le(8)
+        else:
+            head = b"MM\x00" + version + o32be(8)
+        if self.bigtiff:
+            head += o32le(8) if self.endian == "<" else o32be(8)
+            head += b"\x00\x00\x00\x00"
+        return head
+
+    def load(self, data: bytes) -> None:
+        # Extract EXIF information.  This is highly experimental,
+        # and is likely to be replaced with something better in a future
+        # version.
+
+        # The EXIF record consists of a TIFF file embedded in a JPEG
+        # application marker (!).
+        if data == self._loaded_exif:
+            return
+        self._loaded_exif = data
+        self._data.clear()
+        self._hidden_data.clear()
+        self._ifds.clear()
+        while data and data.startswith(b"Exif\x00\x00"):
+            data = data[6:]
+        if not data:
+            self._info = None
+            return
+
+        self.fp: IO[bytes] = io.BytesIO(data)
+        self.head = self.fp.read(8)
+        # process dictionary
+        from . import TiffImagePlugin
+
+        self._info = TiffImagePlugin.ImageFileDirectory_v2(self.head)
+        self.endian = self._info._endian
+        self.fp.seek(self._info.next)
+        self._info.load(self.fp)
+
+    def load_from_fp(self, fp: IO[bytes], offset: int | None = None) -> None:
+        self._loaded_exif = None
+        self._data.clear()
+        self._hidden_data.clear()
+        self._ifds.clear()
+
+        # process dictionary
+        from . import TiffImagePlugin
+
+        self.fp = fp
+        if offset is not None:
+            self.head = self._get_head()
+        else:
+            self.head = self.fp.read(8)
+        self._info = TiffImagePlugin.ImageFileDirectory_v2(self.head)
+        if self.endian is None:
+            self.endian = self._info._endian
+        if offset is None:
+            offset = self._info.next
+        self.fp.tell()
+        self.fp.seek(offset)
+        self._info.load(self.fp)
+
+    def _get_merged_dict(self) -> dict[int, Any]:
+        merged_dict = dict(self)
+
+        # get EXIF extension
+        if ExifTags.IFD.Exif in self:
+            ifd = self._get_ifd_dict(self[ExifTags.IFD.Exif], ExifTags.IFD.Exif)
+            if ifd:
+                merged_dict.update(ifd)
+
+        # GPS
+        if ExifTags.IFD.GPSInfo in self:
+            merged_dict[ExifTags.IFD.GPSInfo] = self._get_ifd_dict(
+                self[ExifTags.IFD.GPSInfo], ExifTags.IFD.GPSInfo
+            )
+
+        return merged_dict
+
+    def tobytes(self, offset: int = 8) -> bytes:
+        from . import TiffImagePlugin
+
+        head = self._get_head()
+        ifd = TiffImagePlugin.ImageFileDirectory_v2(ifh=head)
+        for tag, ifd_dict in self._ifds.items():
+            if tag not in self:
+                ifd[tag] = ifd_dict
+        for tag, value in self.items():
+            if tag in [
+                ExifTags.IFD.Exif,
+                ExifTags.IFD.GPSInfo,
+            ] and not isinstance(value, dict):
+                value = self.get_ifd(tag)
+                if (
+                    tag == ExifTags.IFD.Exif
+                    and ExifTags.IFD.Interop in value
+                    and not isinstance(value[ExifTags.IFD.Interop], dict)
+                ):
+                    value = value.copy()
+                    value[ExifTags.IFD.Interop] = self.get_ifd(ExifTags.IFD.Interop)
+            ifd[tag] = value
+        return b"Exif\x00\x00" + head + ifd.tobytes(offset)
+
+    def get_ifd(self, tag: int) -> dict[int, Any]:
+        if tag not in self._ifds:
+            if tag == ExifTags.IFD.IFD1:
+                if self._info is not None and self._info.next != 0:
+                    ifd = self._get_ifd_dict(self._info.next)
+                    if ifd is not None:
+                        self._ifds[tag] = ifd
+            elif tag in [ExifTags.IFD.Exif, ExifTags.IFD.GPSInfo]:
+                offset = self._hidden_data.get(tag, self.get(tag))
+                if offset is not None:
+                    ifd = self._get_ifd_dict(offset, tag)
+                    if ifd is not None:
+                        self._ifds[tag] = ifd
+            elif tag in [ExifTags.IFD.Interop, ExifTags.IFD.MakerNote]:
+                if ExifTags.IFD.Exif not in self._ifds:
+                    self.get_ifd(ExifTags.IFD.Exif)
+                tag_data = self._ifds[ExifTags.IFD.Exif][tag]
+                if tag == ExifTags.IFD.MakerNote:
+                    from .TiffImagePlugin import ImageFileDirectory_v2
+
+                    if tag_data[:8] == b"FUJIFILM":
+                        ifd_offset = i32le(tag_data, 8)
+                        ifd_data = tag_data[ifd_offset:]
+
+                        makernote = {}
+                        for i in range(0, struct.unpack("<H", ifd_data[:2])[0]):
+                            ifd_tag, typ, count, data = struct.unpack(
+                                "<HHL4s", ifd_data[i * 12 + 2 : (i + 1) * 12 + 2]
+                            )
+                            try:
+                                (
+                                    unit_size,
+                                    handler,
+                                ) = ImageFileDirectory_v2._load_dispatch[typ]
+                            except KeyError:
+                                continue
+                            size = count * unit_size
+                            if size > 4:
+                                (offset,) = struct.unpack("<L", data)
+                                data = ifd_data[offset - 12 : offset + size - 12]
+                            else:
+                                data = data[:size]
+
+                            if len(data) != size:
+                                warnings.warn(
+                                    "Possibly corrupt EXIF MakerNote data.  "
+                                    f"Expecting to read {size} bytes but only got "
+                                    f"{len(data)}. Skipping tag {ifd_tag}"
+                                )
+                                continue
+
+                            if not data:
+                                continue
+
+                            makernote[ifd_tag] = handler(
+                                ImageFileDirectory_v2(), data, False
+                            )
+                        self._ifds[tag] = dict(self._fixup_dict(makernote))
+                    elif self.get(0x010F) == "Nintendo":
+                        makernote = {}
+                        for i in range(0, struct.unpack(">H", tag_data[:2])[0]):
+                            ifd_tag, typ, count, data = struct.unpack(
+                                ">HHL4s", tag_data[i * 12 + 2 : (i + 1) * 12 + 2]
+                            )
+                            if ifd_tag == 0x1101:
+                                # CameraInfo
+                                (offset,) = struct.unpack(">L", data)
+                                self.fp.seek(offset)
+
+                                camerainfo: dict[str, int | bytes] = {
+                                    "ModelID": self.fp.read(4)
+                                }
+
+                                self.fp.read(4)
+                                # Seconds since 2000
+                                camerainfo["TimeStamp"] = i32le(self.fp.read(12))
+
+                                self.fp.read(4)
+                                camerainfo["InternalSerialNumber"] = self.fp.read(4)
+
+                                self.fp.read(12)
+                                parallax = self.fp.read(4)
+                                handler = ImageFileDirectory_v2._load_dispatch[
+                                    TiffTags.FLOAT
+                                ][1]
+                                camerainfo["Parallax"] = handler(
+                                    ImageFileDirectory_v2(), parallax, False
+                                )[0]
+
+                                self.fp.read(4)
+                                camerainfo["Category"] = self.fp.read(2)
+
+                                makernote = {0x1101: camerainfo}
+                        self._ifds[tag] = makernote
+                else:
+                    # Interop
+                    ifd = self._get_ifd_dict(tag_data, tag)
+                    if ifd is not None:
+                        self._ifds[tag] = ifd
+        ifd = self._ifds.setdefault(tag, {})
+        if tag == ExifTags.IFD.Exif and self._hidden_data:
+            ifd = {
+                k: v
+                for (k, v) in ifd.items()
+                if k not in (ExifTags.IFD.Interop, ExifTags.IFD.MakerNote)
+            }
+        return ifd
+
+    def hide_offsets(self) -> None:
+        for tag in (ExifTags.IFD.Exif, ExifTags.IFD.GPSInfo):
+            if tag in self:
+                self._hidden_data[tag] = self[tag]
+                del self[tag]
+
+    def __str__(self) -> str:
+        if self._info is not None:
+            # Load all keys into self._data
+            for tag in self._info:
+                self[tag]
+
+        return str(self._data)
+
+    def __len__(self) -> int:
+        keys = set(self._data)
+        if self._info is not None:
+            keys.update(self._info)
+        return len(keys)
+
+    def __getitem__(self, tag: int) -> Any:
+        if self._info is not None and tag not in self._data and tag in self._info:
+            self._data[tag] = self._fixup(self._info[tag])
+            del self._info[tag]
+        return self._data[tag]
+
+    def __contains__(self, tag: object) -> bool:
+        return tag in self._data or (self._info is not None and tag in self._info)
+
+    def __setitem__(self, tag: int, value: Any) -> None:
+        if self._info is not None and tag in self._info:
+            del self._info[tag]
+        self._data[tag] = value
+
+    def __delitem__(self, tag: int) -> None:
+        if self._info is not None and tag in self._info:
+            del self._info[tag]
+        else:
+            del self._data[tag]
+
+    def __iter__(self) -> Iterator[int]:
+        keys = set(self._data)
+        if self._info is not None:
+            keys.update(self._info)
+        return iter(keys)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageChops.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageChops.py
new file mode 100644
index 00000000..29a5c995
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageChops.py
@@ -0,0 +1,311 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# standard channel operations
+#
+# History:
+# 1996-03-24 fl   Created
+# 1996-08-13 fl   Added logical operations (for "1" images)
+# 2000-10-12 fl   Added offset method (from Image.py)
+#
+# Copyright (c) 1997-2000 by Secret Labs AB
+# Copyright (c) 1996-2000 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+
+from __future__ import annotations
+
+from . import Image
+
+
+def constant(image: Image.Image, value: int) -> Image.Image:
+    """Fill a channel with a given gray level.
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    return Image.new("L", image.size, value)
+
+
+def duplicate(image: Image.Image) -> Image.Image:
+    """Copy a channel. Alias for :py:meth:`PIL.Image.Image.copy`.
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    return image.copy()
+
+
+def invert(image: Image.Image) -> Image.Image:
+    """
+    Invert an image (channel). ::
+
+        out = MAX - image
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image.load()
+    return image._new(image.im.chop_invert())
+
+
+def lighter(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Compares the two images, pixel by pixel, and returns a new image containing
+    the lighter values. ::
+
+        out = max(image1, image2)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_lighter(image2.im))
+
+
+def darker(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Compares the two images, pixel by pixel, and returns a new image containing
+    the darker values. ::
+
+        out = min(image1, image2)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_darker(image2.im))
+
+
+def difference(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Returns the absolute value of the pixel-by-pixel difference between the two
+    images. ::
+
+        out = abs(image1 - image2)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_difference(image2.im))
+
+
+def multiply(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Superimposes two images on top of each other.
+
+    If you multiply an image with a solid black image, the result is black. If
+    you multiply with a solid white image, the image is unaffected. ::
+
+        out = image1 * image2 / MAX
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_multiply(image2.im))
+
+
+def screen(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Superimposes two inverted images on top of each other. ::
+
+        out = MAX - ((MAX - image1) * (MAX - image2) / MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_screen(image2.im))
+
+
+def soft_light(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Superimposes two images on top of each other using the Soft Light algorithm
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_soft_light(image2.im))
+
+
+def hard_light(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Superimposes two images on top of each other using the Hard Light algorithm
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_hard_light(image2.im))
+
+
+def overlay(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """
+    Superimposes two images on top of each other using the Overlay algorithm
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_overlay(image2.im))
+
+
+def add(
+    image1: Image.Image, image2: Image.Image, scale: float = 1.0, offset: float = 0
+) -> Image.Image:
+    """
+    Adds two images, dividing the result by scale and adding the
+    offset. If omitted, scale defaults to 1.0, and offset to 0.0. ::
+
+        out = ((image1 + image2) / scale + offset)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_add(image2.im, scale, offset))
+
+
+def subtract(
+    image1: Image.Image, image2: Image.Image, scale: float = 1.0, offset: float = 0
+) -> Image.Image:
+    """
+    Subtracts two images, dividing the result by scale and adding the offset.
+    If omitted, scale defaults to 1.0, and offset to 0.0. ::
+
+        out = ((image1 - image2) / scale + offset)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_subtract(image2.im, scale, offset))
+
+
+def add_modulo(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """Add two images, without clipping the result. ::
+
+        out = ((image1 + image2) % MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_add_modulo(image2.im))
+
+
+def subtract_modulo(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """Subtract two images, without clipping the result. ::
+
+        out = ((image1 - image2) % MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_subtract_modulo(image2.im))
+
+
+def logical_and(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """Logical AND between two images.
+
+    Both of the images must have mode "1". If you would like to perform a
+    logical AND on an image with a mode other than "1", try
+    :py:meth:`~PIL.ImageChops.multiply` instead, using a black-and-white mask
+    as the second image. ::
+
+        out = ((image1 and image2) % MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_and(image2.im))
+
+
+def logical_or(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """Logical OR between two images.
+
+    Both of the images must have mode "1". ::
+
+        out = ((image1 or image2) % MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_or(image2.im))
+
+
+def logical_xor(image1: Image.Image, image2: Image.Image) -> Image.Image:
+    """Logical XOR between two images.
+
+    Both of the images must have mode "1". ::
+
+        out = ((bool(image1) != bool(image2)) % MAX)
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    image1.load()
+    image2.load()
+    return image1._new(image1.im.chop_xor(image2.im))
+
+
+def blend(image1: Image.Image, image2: Image.Image, alpha: float) -> Image.Image:
+    """Blend images using constant transparency weight. Alias for
+    :py:func:`PIL.Image.blend`.
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    return Image.blend(image1, image2, alpha)
+
+
+def composite(
+    image1: Image.Image, image2: Image.Image, mask: Image.Image
+) -> Image.Image:
+    """Create composite using transparency mask. Alias for
+    :py:func:`PIL.Image.composite`.
+
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    return Image.composite(image1, image2, mask)
+
+
+def offset(image: Image.Image, xoffset: int, yoffset: int | None = None) -> Image.Image:
+    """Returns a copy of the image where data has been offset by the given
+    distances. Data wraps around the edges. If ``yoffset`` is omitted, it
+    is assumed to be equal to ``xoffset``.
+
+    :param image: Input image.
+    :param xoffset: The horizontal distance.
+    :param yoffset: The vertical distance.  If omitted, both
+        distances are set to the same value.
+    :rtype: :py:class:`~PIL.Image.Image`
+    """
+
+    if yoffset is None:
+        yoffset = xoffset
+    image.load()
+    return image._new(image.im.offset(xoffset, yoffset))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageCms.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageCms.py
new file mode 100644
index 00000000..fdfbee78
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageCms.py
@@ -0,0 +1,1125 @@
+# The Python Imaging Library.
+# $Id$
+
+# Optional color management support, based on Kevin Cazabon's PyCMS
+# library.
+
+# Originally released under LGPL.  Graciously donated to PIL in
+# March 2009, for distribution under the standard PIL license
+
+# History:
+
+# 2009-03-08 fl   Added to PIL.
+
+# Copyright (C) 2002-2003 Kevin Cazabon
+# Copyright (c) 2009 by Fredrik Lundh
+# Copyright (c) 2013 by Eric Soroos
+
+# See the README file for information on usage and redistribution.  See
+# below for the original description.
+from __future__ import annotations
+
+import operator
+import sys
+from enum import IntEnum, IntFlag
+from functools import reduce
+from typing import Any, Literal, SupportsFloat, SupportsInt, Union
+
+from . import Image, __version__
+from ._deprecate import deprecate
+from ._typing import SupportsRead
+
+try:
+    from . import _imagingcms as core
+
+    _CmsProfileCompatible = Union[
+        str, SupportsRead[bytes], core.CmsProfile, "ImageCmsProfile"
+    ]
+except ImportError as ex:
+    # Allow error import for doc purposes, but error out when accessing
+    # anything in core.
+    from ._util import DeferredError
+
+    core = DeferredError.new(ex)
+
+_DESCRIPTION = """
+pyCMS
+
+    a Python / PIL interface to the littleCMS ICC Color Management System
+    Copyright (C) 2002-2003 Kevin Cazabon
+    kevin@cazabon.com
+    https://www.cazabon.com
+
+    pyCMS home page:  https://www.cazabon.com/pyCMS
+    littleCMS home page:  https://www.littlecms.com
+    (littleCMS is Copyright (C) 1998-2001 Marti Maria)
+
+    Originally released under LGPL.  Graciously donated to PIL in
+    March 2009, for distribution under the standard PIL license
+
+    The pyCMS.py module provides a "clean" interface between Python/PIL and
+    pyCMSdll, taking care of some of the more complex handling of the direct
+    pyCMSdll functions, as well as error-checking and making sure that all
+    relevant data is kept together.
+
+    While it is possible to call pyCMSdll functions directly, it's not highly
+    recommended.
+
+    Version History:
+
+        1.0.0 pil       Oct 2013 Port to LCMS 2.
+
+        0.1.0 pil mod   March 10, 2009
+
+                        Renamed display profile to proof profile. The proof
+                        profile is the profile of the device that is being
+                        simulated, not the profile of the device which is
+                        actually used to display/print the final simulation
+                        (that'd be the output profile) - also see LCMSAPI.txt
+                        input colorspace -> using 'renderingIntent' -> proof
+                        colorspace -> using 'proofRenderingIntent' -> output
+                        colorspace
+
+                        Added LCMS FLAGS support.
+                        Added FLAGS["SOFTPROOFING"] as default flag for
+                        buildProofTransform (otherwise the proof profile/intent
+                        would be ignored).
+
+        0.1.0 pil       March 2009 - added to PIL, as PIL.ImageCms
+
+        0.0.2 alpha     Jan 6, 2002
+
+                        Added try/except statements around type() checks of
+                        potential CObjects... Python won't let you use type()
+                        on them, and raises a TypeError (stupid, if you ask
+                        me!)
+
+                        Added buildProofTransformFromOpenProfiles() function.
+                        Additional fixes in DLL, see DLL code for details.
+
+        0.0.1 alpha     first public release, Dec. 26, 2002
+
+    Known to-do list with current version (of Python interface, not pyCMSdll):
+
+        none
+
+"""
+
+_VERSION = "1.0.0 pil"
+
+
+def __getattr__(name: str) -> Any:
+    if name == "DESCRIPTION":
+        deprecate("PIL.ImageCms.DESCRIPTION", 12)
+        return _DESCRIPTION
+    elif name == "VERSION":
+        deprecate("PIL.ImageCms.VERSION", 12)
+        return _VERSION
+    elif name == "FLAGS":
+        deprecate("PIL.ImageCms.FLAGS", 12, "PIL.ImageCms.Flags")
+        return _FLAGS
+    msg = f"module '{__name__}' has no attribute '{name}'"
+    raise AttributeError(msg)
+
+
+# --------------------------------------------------------------------.
+
+
+#
+# intent/direction values
+
+
+class Intent(IntEnum):
+    PERCEPTUAL = 0
+    RELATIVE_COLORIMETRIC = 1
+    SATURATION = 2
+    ABSOLUTE_COLORIMETRIC = 3
+
+
+class Direction(IntEnum):
+    INPUT = 0
+    OUTPUT = 1
+    PROOF = 2
+
+
+#
+# flags
+
+
+class Flags(IntFlag):
+    """Flags and documentation are taken from ``lcms2.h``."""
+
+    NONE = 0
+    NOCACHE = 0x0040
+    """Inhibit 1-pixel cache"""
+    NOOPTIMIZE = 0x0100
+    """Inhibit optimizations"""
+    NULLTRANSFORM = 0x0200
+    """Don't transform anyway"""
+    GAMUTCHECK = 0x1000
+    """Out of Gamut alarm"""
+    SOFTPROOFING = 0x4000
+    """Do softproofing"""
+    BLACKPOINTCOMPENSATION = 0x2000
+    NOWHITEONWHITEFIXUP = 0x0004
+    """Don't fix scum dot"""
+    HIGHRESPRECALC = 0x0400
+    """Use more memory to give better accuracy"""
+    LOWRESPRECALC = 0x0800
+    """Use less memory to minimize resources"""
+    # this should be 8BITS_DEVICELINK, but that is not a valid name in Python:
+    USE_8BITS_DEVICELINK = 0x0008
+    """Create 8 bits devicelinks"""
+    GUESSDEVICECLASS = 0x0020
+    """Guess device class (for ``transform2devicelink``)"""
+    KEEP_SEQUENCE = 0x0080
+    """Keep profile sequence for devicelink creation"""
+    FORCE_CLUT = 0x0002
+    """Force CLUT optimization"""
+    CLUT_POST_LINEARIZATION = 0x0001
+    """create postlinearization tables if possible"""
+    CLUT_PRE_LINEARIZATION = 0x0010
+    """create prelinearization tables if possible"""
+    NONEGATIVES = 0x8000
+    """Prevent negative numbers in floating point transforms"""
+    COPY_ALPHA = 0x04000000
+    """Alpha channels are copied on ``cmsDoTransform()``"""
+    NODEFAULTRESOURCEDEF = 0x01000000
+
+    _GRIDPOINTS_1 = 1 << 16
+    _GRIDPOINTS_2 = 2 << 16
+    _GRIDPOINTS_4 = 4 << 16
+    _GRIDPOINTS_8 = 8 << 16
+    _GRIDPOINTS_16 = 16 << 16
+    _GRIDPOINTS_32 = 32 << 16
+    _GRIDPOINTS_64 = 64 << 16
+    _GRIDPOINTS_128 = 128 << 16
+
+    @staticmethod
+    def GRIDPOINTS(n: int) -> Flags:
+        """
+        Fine-tune control over number of gridpoints
+
+        :param n: :py:class:`int` in range ``0 <= n <= 255``
+        """
+        return Flags.NONE | ((n & 0xFF) << 16)
+
+
+_MAX_FLAG = reduce(operator.or_, Flags)
+
+
+_FLAGS = {
+    "MATRIXINPUT": 1,
+    "MATRIXOUTPUT": 2,
+    "MATRIXONLY": (1 | 2),
+    "NOWHITEONWHITEFIXUP": 4,  # Don't hot fix scum dot
+    # Don't create prelinearization tables on precalculated transforms
+    # (internal use):
+    "NOPRELINEARIZATION": 16,
+    "GUESSDEVICECLASS": 32,  # Guess device class (for transform2devicelink)
+    "NOTCACHE": 64,  # Inhibit 1-pixel cache
+    "NOTPRECALC": 256,
+    "NULLTRANSFORM": 512,  # Don't transform anyway
+    "HIGHRESPRECALC": 1024,  # Use more memory to give better accuracy
+    "LOWRESPRECALC": 2048,  # Use less memory to minimize resources
+    "WHITEBLACKCOMPENSATION": 8192,
+    "BLACKPOINTCOMPENSATION": 8192,
+    "GAMUTCHECK": 4096,  # Out of Gamut alarm
+    "SOFTPROOFING": 16384,  # Do softproofing
+    "PRESERVEBLACK": 32768,  # Black preservation
+    "NODEFAULTRESOURCEDEF": 16777216,  # CRD special
+    "GRIDPOINTS": lambda n: (n & 0xFF) << 16,  # Gridpoints
+}
+
+
+# --------------------------------------------------------------------.
+# Experimental PIL-level API
+# --------------------------------------------------------------------.
+
+##
+# Profile.
+
+
+class ImageCmsProfile:
+    def __init__(self, profile: str | SupportsRead[bytes] | core.CmsProfile) -> None:
+        """
+        :param profile: Either a string representing a filename,
+            a file like object containing a profile or a
+            low-level profile object
+
+        """
+
+        if isinstance(profile, str):
+            if sys.platform == "win32":
+                profile_bytes_path = profile.encode()
+                try:
+                    profile_bytes_path.decode("ascii")
+                except UnicodeDecodeError:
+                    with open(profile, "rb") as f:
+                        self._set(core.profile_frombytes(f.read()))
+                    return
+            self._set(core.profile_open(profile), profile)
+        elif hasattr(profile, "read"):
+            self._set(core.profile_frombytes(profile.read()))
+        elif isinstance(profile, core.CmsProfile):
+            self._set(profile)
+        else:
+            msg = "Invalid type for Profile"  # type: ignore[unreachable]
+            raise TypeError(msg)
+
+    def _set(self, profile: core.CmsProfile, filename: str | None = None) -> None:
+        self.profile = profile
+        self.filename = filename
+        self.product_name = None  # profile.product_name
+        self.product_info = None  # profile.product_info
+
+    def tobytes(self) -> bytes:
+        """
+        Returns the profile in a format suitable for embedding in
+        saved images.
+
+        :returns: a bytes object containing the ICC profile.
+        """
+
+        return core.profile_tobytes(self.profile)
+
+
+class ImageCmsTransform(Image.ImagePointHandler):
+    """
+    Transform.  This can be used with the procedural API, or with the standard
+    :py:func:`~PIL.Image.Image.point` method.
+
+    Will return the output profile in the ``output.info['icc_profile']``.
+    """
+
+    def __init__(
+        self,
+        input: ImageCmsProfile,
+        output: ImageCmsProfile,
+        input_mode: str,
+        output_mode: str,
+        intent: Intent = Intent.PERCEPTUAL,
+        proof: ImageCmsProfile | None = None,
+        proof_intent: Intent = Intent.ABSOLUTE_COLORIMETRIC,
+        flags: Flags = Flags.NONE,
+    ):
+        supported_modes = (
+            "RGB",
+            "RGBA",
+            "RGBX",
+            "CMYK",
+            "I;16",
+            "I;16L",
+            "I;16B",
+            "YCbCr",
+            "LAB",
+            "L",
+            "1",
+        )
+        for mode in (input_mode, output_mode):
+            if mode not in supported_modes:
+                deprecate(
+                    mode,
+                    12,
+                    {
+                        "L;16": "I;16 or I;16L",
+                        "L:16B": "I;16B",
+                        "YCCA": "YCbCr",
+                        "YCC": "YCbCr",
+                    }.get(mode),
+                )
+        if proof is None:
+            self.transform = core.buildTransform(
+                input.profile, output.profile, input_mode, output_mode, intent, flags
+            )
+        else:
+            self.transform = core.buildProofTransform(
+                input.profile,
+                output.profile,
+                proof.profile,
+                input_mode,
+                output_mode,
+                intent,
+                proof_intent,
+                flags,
+            )
+        # Note: inputMode and outputMode are for pyCMS compatibility only
+        self.input_mode = self.inputMode = input_mode
+        self.output_mode = self.outputMode = output_mode
+
+        self.output_profile = output
+
+    def point(self, im: Image.Image) -> Image.Image:
+        return self.apply(im)
+
+    def apply(self, im: Image.Image, imOut: Image.Image | None = None) -> Image.Image:
+        if imOut is None:
+            imOut = Image.new(self.output_mode, im.size, None)
+        self.transform.apply(im.getim(), imOut.getim())
+        imOut.info["icc_profile"] = self.output_profile.tobytes()
+        return imOut
+
+    def apply_in_place(self, im: Image.Image) -> Image.Image:
+        if im.mode != self.output_mode:
+            msg = "mode mismatch"
+            raise ValueError(msg)  # wrong output mode
+        self.transform.apply(im.getim(), im.getim())
+        im.info["icc_profile"] = self.output_profile.tobytes()
+        return im
+
+
+def get_display_profile(handle: SupportsInt | None = None) -> ImageCmsProfile | None:
+    """
+    (experimental) Fetches the profile for the current display device.
+
+    :returns: ``None`` if the profile is not known.
+    """
+
+    if sys.platform != "win32":
+        return None
+
+    from . import ImageWin  # type: ignore[unused-ignore, unreachable]
+
+    if isinstance(handle, ImageWin.HDC):
+        profile = core.get_display_profile_win32(int(handle), 1)
+    else:
+        profile = core.get_display_profile_win32(int(handle or 0))
+    if profile is None:
+        return None
+    return ImageCmsProfile(profile)
+
+
+# --------------------------------------------------------------------.
+# pyCMS compatible layer
+# --------------------------------------------------------------------.
+
+
+class PyCMSError(Exception):
+    """(pyCMS) Exception class.
+    This is used for all errors in the pyCMS API."""
+
+    pass
+
+
+def profileToProfile(
+    im: Image.Image,
+    inputProfile: _CmsProfileCompatible,
+    outputProfile: _CmsProfileCompatible,
+    renderingIntent: Intent = Intent.PERCEPTUAL,
+    outputMode: str | None = None,
+    inPlace: bool = False,
+    flags: Flags = Flags.NONE,
+) -> Image.Image | None:
+    """
+    (pyCMS) Applies an ICC transformation to a given image, mapping from
+    ``inputProfile`` to ``outputProfile``.
+
+    If the input or output profiles specified are not valid filenames, a
+    :exc:`PyCMSError` will be raised.  If ``inPlace`` is ``True`` and
+    ``outputMode != im.mode``, a :exc:`PyCMSError` will be raised.
+    If an error occurs during application of the profiles,
+    a :exc:`PyCMSError` will be raised.
+    If ``outputMode`` is not a mode supported by the ``outputProfile`` (or by pyCMS),
+    a :exc:`PyCMSError` will be raised.
+
+    This function applies an ICC transformation to im from ``inputProfile``'s
+    color space to ``outputProfile``'s color space using the specified rendering
+    intent to decide how to handle out-of-gamut colors.
+
+    ``outputMode`` can be used to specify that a color mode conversion is to
+    be done using these profiles, but the specified profiles must be able
+    to handle that mode.  I.e., if converting im from RGB to CMYK using
+    profiles, the input profile must handle RGB data, and the output
+    profile must handle CMYK data.
+
+    :param im: An open :py:class:`~PIL.Image.Image` object (i.e. Image.new(...)
+        or Image.open(...), etc.)
+    :param inputProfile: String, as a valid filename path to the ICC input
+        profile you wish to use for this image, or a profile object
+    :param outputProfile: String, as a valid filename path to the ICC output
+        profile you wish to use for this image, or a profile object
+    :param renderingIntent: Integer (0-3) specifying the rendering intent you
+        wish to use for the transform
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+        they do.
+    :param outputMode: A valid PIL mode for the output image (i.e. "RGB",
+        "CMYK", etc.).  Note: if rendering the image "inPlace", outputMode
+        MUST be the same mode as the input, or omitted completely.  If
+        omitted, the outputMode will be the same as the mode of the input
+        image (im.mode)
+    :param inPlace: Boolean.  If ``True``, the original image is modified in-place,
+        and ``None`` is returned.  If ``False`` (default), a new
+        :py:class:`~PIL.Image.Image` object is returned with the transform applied.
+    :param flags: Integer (0-...) specifying additional flags
+    :returns: Either None or a new :py:class:`~PIL.Image.Image` object, depending on
+        the value of ``inPlace``
+    :exception PyCMSError:
+    """
+
+    if outputMode is None:
+        outputMode = im.mode
+
+    if not isinstance(renderingIntent, int) or not (0 <= renderingIntent <= 3):
+        msg = "renderingIntent must be an integer between 0 and 3"
+        raise PyCMSError(msg)
+
+    if not isinstance(flags, int) or not (0 <= flags <= _MAX_FLAG):
+        msg = f"flags must be an integer between 0 and {_MAX_FLAG}"
+        raise PyCMSError(msg)
+
+    try:
+        if not isinstance(inputProfile, ImageCmsProfile):
+            inputProfile = ImageCmsProfile(inputProfile)
+        if not isinstance(outputProfile, ImageCmsProfile):
+            outputProfile = ImageCmsProfile(outputProfile)
+        transform = ImageCmsTransform(
+            inputProfile,
+            outputProfile,
+            im.mode,
+            outputMode,
+            renderingIntent,
+            flags=flags,
+        )
+        if inPlace:
+            transform.apply_in_place(im)
+            imOut = None
+        else:
+            imOut = transform.apply(im)
+    except (OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+    return imOut
+
+
+def getOpenProfile(
+    profileFilename: str | SupportsRead[bytes] | core.CmsProfile,
+) -> ImageCmsProfile:
+    """
+    (pyCMS) Opens an ICC profile file.
+
+    The PyCMSProfile object can be passed back into pyCMS for use in creating
+    transforms and such (as in ImageCms.buildTransformFromOpenProfiles()).
+
+    If ``profileFilename`` is not a valid filename for an ICC profile,
+    a :exc:`PyCMSError` will be raised.
+
+    :param profileFilename: String, as a valid filename path to the ICC profile
+        you wish to open, or a file-like object.
+    :returns: A CmsProfile class object.
+    :exception PyCMSError:
+    """
+
+    try:
+        return ImageCmsProfile(profileFilename)
+    except (OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def buildTransform(
+    inputProfile: _CmsProfileCompatible,
+    outputProfile: _CmsProfileCompatible,
+    inMode: str,
+    outMode: str,
+    renderingIntent: Intent = Intent.PERCEPTUAL,
+    flags: Flags = Flags.NONE,
+) -> ImageCmsTransform:
+    """
+    (pyCMS) Builds an ICC transform mapping from the ``inputProfile`` to the
+    ``outputProfile``. Use applyTransform to apply the transform to a given
+    image.
+
+    If the input or output profiles specified are not valid filenames, a
+    :exc:`PyCMSError` will be raised. If an error occurs during creation
+    of the transform, a :exc:`PyCMSError` will be raised.
+
+    If ``inMode`` or ``outMode`` are not a mode supported by the ``outputProfile``
+    (or by pyCMS), a :exc:`PyCMSError` will be raised.
+
+    This function builds and returns an ICC transform from the ``inputProfile``
+    to the ``outputProfile`` using the ``renderingIntent`` to determine what to do
+    with out-of-gamut colors.  It will ONLY work for converting images that
+    are in ``inMode`` to images that are in ``outMode`` color format (PIL mode,
+    i.e. "RGB", "RGBA", "CMYK", etc.).
+
+    Building the transform is a fair part of the overhead in
+    ImageCms.profileToProfile(), so if you're planning on converting multiple
+    images using the same input/output settings, this can save you time.
+    Once you have a transform object, it can be used with
+    ImageCms.applyProfile() to convert images without the need to re-compute
+    the lookup table for the transform.
+
+    The reason pyCMS returns a class object rather than a handle directly
+    to the transform is that it needs to keep track of the PIL input/output
+    modes that the transform is meant for.  These attributes are stored in
+    the ``inMode`` and ``outMode`` attributes of the object (which can be
+    manually overridden if you really want to, but I don't know of any
+    time that would be of use, or would even work).
+
+    :param inputProfile: String, as a valid filename path to the ICC input
+        profile you wish to use for this transform, or a profile object
+    :param outputProfile: String, as a valid filename path to the ICC output
+        profile you wish to use for this transform, or a profile object
+    :param inMode: String, as a valid PIL mode that the appropriate profile
+        also supports (i.e. "RGB", "RGBA", "CMYK", etc.)
+    :param outMode: String, as a valid PIL mode that the appropriate profile
+        also supports (i.e. "RGB", "RGBA", "CMYK", etc.)
+    :param renderingIntent: Integer (0-3) specifying the rendering intent you
+        wish to use for the transform
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+        they do.
+    :param flags: Integer (0-...) specifying additional flags
+    :returns: A CmsTransform class object.
+    :exception PyCMSError:
+    """
+
+    if not isinstance(renderingIntent, int) or not (0 <= renderingIntent <= 3):
+        msg = "renderingIntent must be an integer between 0 and 3"
+        raise PyCMSError(msg)
+
+    if not isinstance(flags, int) or not (0 <= flags <= _MAX_FLAG):
+        msg = f"flags must be an integer between 0 and {_MAX_FLAG}"
+        raise PyCMSError(msg)
+
+    try:
+        if not isinstance(inputProfile, ImageCmsProfile):
+            inputProfile = ImageCmsProfile(inputProfile)
+        if not isinstance(outputProfile, ImageCmsProfile):
+            outputProfile = ImageCmsProfile(outputProfile)
+        return ImageCmsTransform(
+            inputProfile, outputProfile, inMode, outMode, renderingIntent, flags=flags
+        )
+    except (OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def buildProofTransform(
+    inputProfile: _CmsProfileCompatible,
+    outputProfile: _CmsProfileCompatible,
+    proofProfile: _CmsProfileCompatible,
+    inMode: str,
+    outMode: str,
+    renderingIntent: Intent = Intent.PERCEPTUAL,
+    proofRenderingIntent: Intent = Intent.ABSOLUTE_COLORIMETRIC,
+    flags: Flags = Flags.SOFTPROOFING,
+) -> ImageCmsTransform:
+    """
+    (pyCMS) Builds an ICC transform mapping from the ``inputProfile`` to the
+    ``outputProfile``, but tries to simulate the result that would be
+    obtained on the ``proofProfile`` device.
+
+    If the input, output, or proof profiles specified are not valid
+    filenames, a :exc:`PyCMSError` will be raised.
+
+    If an error occurs during creation of the transform,
+    a :exc:`PyCMSError` will be raised.
+
+    If ``inMode`` or ``outMode`` are not a mode supported by the ``outputProfile``
+    (or by pyCMS), a :exc:`PyCMSError` will be raised.
+
+    This function builds and returns an ICC transform from the ``inputProfile``
+    to the ``outputProfile``, but tries to simulate the result that would be
+    obtained on the ``proofProfile`` device using ``renderingIntent`` and
+    ``proofRenderingIntent`` to determine what to do with out-of-gamut
+    colors.  This is known as "soft-proofing".  It will ONLY work for
+    converting images that are in ``inMode`` to images that are in outMode
+    color format (PIL mode, i.e. "RGB", "RGBA", "CMYK", etc.).
+
+    Usage of the resulting transform object is exactly the same as with
+    ImageCms.buildTransform().
+
+    Proof profiling is generally used when using an output device to get a
+    good idea of what the final printed/displayed image would look like on
+    the ``proofProfile`` device when it's quicker and easier to use the
+    output device for judging color.  Generally, this means that the
+    output device is a monitor, or a dye-sub printer (etc.), and the simulated
+    device is something more expensive, complicated, or time consuming
+    (making it difficult to make a real print for color judgement purposes).
+
+    Soft-proofing basically functions by adjusting the colors on the
+    output device to match the colors of the device being simulated. However,
+    when the simulated device has a much wider gamut than the output
+    device, you may obtain marginal results.
+
+    :param inputProfile: String, as a valid filename path to the ICC input
+        profile you wish to use for this transform, or a profile object
+    :param outputProfile: String, as a valid filename path to the ICC output
+        (monitor, usually) profile you wish to use for this transform, or a
+        profile object
+    :param proofProfile: String, as a valid filename path to the ICC proof
+        profile you wish to use for this transform, or a profile object
+    :param inMode: String, as a valid PIL mode that the appropriate profile
+        also supports (i.e. "RGB", "RGBA", "CMYK", etc.)
+    :param outMode: String, as a valid PIL mode that the appropriate profile
+        also supports (i.e. "RGB", "RGBA", "CMYK", etc.)
+    :param renderingIntent: Integer (0-3) specifying the rendering intent you
+        wish to use for the input->proof (simulated) transform
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+        they do.
+    :param proofRenderingIntent: Integer (0-3) specifying the rendering intent
+        you wish to use for proof->output transform
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+        they do.
+    :param flags: Integer (0-...) specifying additional flags
+    :returns: A CmsTransform class object.
+    :exception PyCMSError:
+    """
+
+    if not isinstance(renderingIntent, int) or not (0 <= renderingIntent <= 3):
+        msg = "renderingIntent must be an integer between 0 and 3"
+        raise PyCMSError(msg)
+
+    if not isinstance(flags, int) or not (0 <= flags <= _MAX_FLAG):
+        msg = f"flags must be an integer between 0 and {_MAX_FLAG}"
+        raise PyCMSError(msg)
+
+    try:
+        if not isinstance(inputProfile, ImageCmsProfile):
+            inputProfile = ImageCmsProfile(inputProfile)
+        if not isinstance(outputProfile, ImageCmsProfile):
+            outputProfile = ImageCmsProfile(outputProfile)
+        if not isinstance(proofProfile, ImageCmsProfile):
+            proofProfile = ImageCmsProfile(proofProfile)
+        return ImageCmsTransform(
+            inputProfile,
+            outputProfile,
+            inMode,
+            outMode,
+            renderingIntent,
+            proofProfile,
+            proofRenderingIntent,
+            flags,
+        )
+    except (OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+buildTransformFromOpenProfiles = buildTransform
+buildProofTransformFromOpenProfiles = buildProofTransform
+
+
+def applyTransform(
+    im: Image.Image, transform: ImageCmsTransform, inPlace: bool = False
+) -> Image.Image | None:
+    """
+    (pyCMS) Applies a transform to a given image.
+
+    If ``im.mode != transform.input_mode``, a :exc:`PyCMSError` is raised.
+
+    If ``inPlace`` is ``True`` and ``transform.input_mode != transform.output_mode``, a
+    :exc:`PyCMSError` is raised.
+
+    If ``im.mode``, ``transform.input_mode`` or ``transform.output_mode`` is not
+    supported by pyCMSdll or the profiles you used for the transform, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while the transform is being applied,
+    a :exc:`PyCMSError` is raised.
+
+    This function applies a pre-calculated transform (from
+    ImageCms.buildTransform() or ImageCms.buildTransformFromOpenProfiles())
+    to an image. The transform can be used for multiple images, saving
+    considerable calculation time if doing the same conversion multiple times.
+
+    If you want to modify im in-place instead of receiving a new image as
+    the return value, set ``inPlace`` to ``True``.  This can only be done if
+    ``transform.input_mode`` and ``transform.output_mode`` are the same, because we
+    can't change the mode in-place (the buffer sizes for some modes are
+    different).  The default behavior is to return a new :py:class:`~PIL.Image.Image`
+    object of the same dimensions in mode ``transform.output_mode``.
+
+    :param im: An :py:class:`~PIL.Image.Image` object, and ``im.mode`` must be the same
+        as the ``input_mode`` supported by the transform.
+    :param transform: A valid CmsTransform class object
+    :param inPlace: Bool.  If ``True``, ``im`` is modified in place and ``None`` is
+        returned, if ``False``, a new :py:class:`~PIL.Image.Image` object with the
+        transform applied is returned (and ``im`` is not changed). The default is
+        ``False``.
+    :returns: Either ``None``, or a new :py:class:`~PIL.Image.Image` object,
+        depending on the value of ``inPlace``. The profile will be returned in
+        the image's ``info['icc_profile']``.
+    :exception PyCMSError:
+    """
+
+    try:
+        if inPlace:
+            transform.apply_in_place(im)
+            imOut = None
+        else:
+            imOut = transform.apply(im)
+    except (TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+    return imOut
+
+
+def createProfile(
+    colorSpace: Literal["LAB", "XYZ", "sRGB"], colorTemp: SupportsFloat = 0
+) -> core.CmsProfile:
+    """
+    (pyCMS) Creates a profile.
+
+    If colorSpace not in ``["LAB", "XYZ", "sRGB"]``,
+    a :exc:`PyCMSError` is raised.
+
+    If using LAB and ``colorTemp`` is not a positive integer,
+    a :exc:`PyCMSError` is raised.
+
+    If an error occurs while creating the profile,
+    a :exc:`PyCMSError` is raised.
+
+    Use this function to create common profiles on-the-fly instead of
+    having to supply a profile on disk and knowing the path to it.  It
+    returns a normal CmsProfile object that can be passed to
+    ImageCms.buildTransformFromOpenProfiles() to create a transform to apply
+    to images.
+
+    :param colorSpace: String, the color space of the profile you wish to
+        create.
+        Currently only "LAB", "XYZ", and "sRGB" are supported.
+    :param colorTemp: Positive number for the white point for the profile, in
+        degrees Kelvin (i.e. 5000, 6500, 9600, etc.).  The default is for D50
+        illuminant if omitted (5000k).  colorTemp is ONLY applied to LAB
+        profiles, and is ignored for XYZ and sRGB.
+    :returns: A CmsProfile class object
+    :exception PyCMSError:
+    """
+
+    if colorSpace not in ["LAB", "XYZ", "sRGB"]:
+        msg = (
+            f"Color space not supported for on-the-fly profile creation ({colorSpace})"
+        )
+        raise PyCMSError(msg)
+
+    if colorSpace == "LAB":
+        try:
+            colorTemp = float(colorTemp)
+        except (TypeError, ValueError) as e:
+            msg = f'Color temperature must be numeric, "{colorTemp}" not valid'
+            raise PyCMSError(msg) from e
+
+    try:
+        return core.createProfile(colorSpace, colorTemp)
+    except (TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileName(profile: _CmsProfileCompatible) -> str:
+    """
+
+    (pyCMS) Gets the internal product name for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile,
+    a :exc:`PyCMSError` is raised If an error occurs while trying
+    to obtain the name tag, a :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the INTERNAL name of the profile (stored
+    in an ICC tag in the profile itself), usually the one used when the
+    profile was originally created.  Sometimes this tag also contains
+    additional information supplied by the creator.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal name of the profile as stored
+        in an ICC tag.
+    :exception PyCMSError:
+    """
+
+    try:
+        # add an extra newline to preserve pyCMS compatibility
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        # do it in python, not c.
+        #    // name was "%s - %s" (model, manufacturer) || Description ,
+        #    // but if the Model and Manufacturer were the same or the model
+        #    // was long, Just the model,  in 1.x
+        model = profile.profile.model
+        manufacturer = profile.profile.manufacturer
+
+        if not (model or manufacturer):
+            return (profile.profile.profile_description or "") + "\n"
+        if not manufacturer or (model and len(model) > 30):
+            return f"{model}\n"
+        return f"{model} - {manufacturer}\n"
+
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileInfo(profile: _CmsProfileCompatible) -> str:
+    """
+    (pyCMS) Gets the internal product information for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile,
+    a :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the info tag,
+    a :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the information stored in the profile's
+    info tag.  This often contains details about the profile, and how it
+    was created, as supplied by the creator.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal profile information stored in
+        an ICC tag.
+    :exception PyCMSError:
+    """
+
+    try:
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        # add an extra newline to preserve pyCMS compatibility
+        # Python, not C. the white point bits weren't working well,
+        # so skipping.
+        # info was description \r\n\r\n copyright \r\n\r\n K007 tag \r\n\r\n whitepoint
+        description = profile.profile.profile_description
+        cpright = profile.profile.copyright
+        elements = [element for element in (description, cpright) if element]
+        return "\r\n\r\n".join(elements) + "\r\n\r\n"
+
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileCopyright(profile: _CmsProfileCompatible) -> str:
+    """
+    (pyCMS) Gets the copyright for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the copyright tag,
+    a :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the information stored in the profile's
+    copyright tag.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal profile information stored in
+        an ICC tag.
+    :exception PyCMSError:
+    """
+    try:
+        # add an extra newline to preserve pyCMS compatibility
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        return (profile.profile.copyright or "") + "\n"
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileManufacturer(profile: _CmsProfileCompatible) -> str:
+    """
+    (pyCMS) Gets the manufacturer for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the manufacturer tag, a
+    :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the information stored in the profile's
+    manufacturer tag.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal profile information stored in
+        an ICC tag.
+    :exception PyCMSError:
+    """
+    try:
+        # add an extra newline to preserve pyCMS compatibility
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        return (profile.profile.manufacturer or "") + "\n"
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileModel(profile: _CmsProfileCompatible) -> str:
+    """
+    (pyCMS) Gets the model for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the model tag,
+    a :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the information stored in the profile's
+    model tag.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal profile information stored in
+        an ICC tag.
+    :exception PyCMSError:
+    """
+
+    try:
+        # add an extra newline to preserve pyCMS compatibility
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        return (profile.profile.model or "") + "\n"
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getProfileDescription(profile: _CmsProfileCompatible) -> str:
+    """
+    (pyCMS) Gets the description for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the description tag,
+    a :exc:`PyCMSError` is raised.
+
+    Use this function to obtain the information stored in the profile's
+    description tag.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: A string containing the internal profile information stored in an
+        ICC tag.
+    :exception PyCMSError:
+    """
+
+    try:
+        # add an extra newline to preserve pyCMS compatibility
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        return (profile.profile.profile_description or "") + "\n"
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def getDefaultIntent(profile: _CmsProfileCompatible) -> int:
+    """
+    (pyCMS) Gets the default intent name for the given profile.
+
+    If ``profile`` isn't a valid CmsProfile object or filename to a profile, a
+    :exc:`PyCMSError` is raised.
+
+    If an error occurs while trying to obtain the default intent, a
+    :exc:`PyCMSError` is raised.
+
+    Use this function to determine the default (and usually best optimized)
+    rendering intent for this profile.  Most profiles support multiple
+    rendering intents, but are intended mostly for one type of conversion.
+    If you wish to use a different intent than returned, use
+    ImageCms.isIntentSupported() to verify it will work first.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :returns: Integer 0-3 specifying the default rendering intent for this
+        profile.
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+            they do.
+    :exception PyCMSError:
+    """
+
+    try:
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        return profile.profile.rendering_intent
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def isIntentSupported(
+    profile: _CmsProfileCompatible, intent: Intent, direction: Direction
+) -> Literal[-1, 1]:
+    """
+    (pyCMS) Checks if a given intent is supported.
+
+    Use this function to verify that you can use your desired
+    ``intent`` with ``profile``, and that ``profile`` can be used for the
+    input/output/proof profile as you desire.
+
+    Some profiles are created specifically for one "direction", can cannot
+    be used for others. Some profiles can only be used for certain
+    rendering intents, so it's best to either verify this before trying
+    to create a transform with them (using this function), or catch the
+    potential :exc:`PyCMSError` that will occur if they don't
+    support the modes you select.
+
+    :param profile: EITHER a valid CmsProfile object, OR a string of the
+        filename of an ICC profile.
+    :param intent: Integer (0-3) specifying the rendering intent you wish to
+        use with this profile
+
+            ImageCms.Intent.PERCEPTUAL            = 0 (DEFAULT)
+            ImageCms.Intent.RELATIVE_COLORIMETRIC = 1
+            ImageCms.Intent.SATURATION            = 2
+            ImageCms.Intent.ABSOLUTE_COLORIMETRIC = 3
+
+        see the pyCMS documentation for details on rendering intents and what
+            they do.
+    :param direction: Integer specifying if the profile is to be used for
+        input, output, or proof
+
+            INPUT  = 0 (or use ImageCms.Direction.INPUT)
+            OUTPUT = 1 (or use ImageCms.Direction.OUTPUT)
+            PROOF  = 2 (or use ImageCms.Direction.PROOF)
+
+    :returns: 1 if the intent/direction are supported, -1 if they are not.
+    :exception PyCMSError:
+    """
+
+    try:
+        if not isinstance(profile, ImageCmsProfile):
+            profile = ImageCmsProfile(profile)
+        # FIXME: I get different results for the same data w. different
+        # compilers.  Bug in LittleCMS or in the binding?
+        if profile.profile.is_intent_supported(intent, direction):
+            return 1
+        else:
+            return -1
+    except (AttributeError, OSError, TypeError, ValueError) as v:
+        raise PyCMSError(v) from v
+
+
+def versions() -> tuple[str, str | None, str, str]:
+    """
+    (pyCMS) Fetches versions.
+    """
+
+    deprecate(
+        "PIL.ImageCms.versions()",
+        12,
+        '(PIL.features.version("littlecms2"), sys.version, PIL.__version__)',
+    )
+    return _VERSION, core.littlecms_version, sys.version.split()[0], __version__
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageColor.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageColor.py
new file mode 100644
index 00000000..9a15a8eb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageColor.py
@@ -0,0 +1,320 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# map CSS3-style colour description strings to RGB
+#
+# History:
+# 2002-10-24 fl   Added support for CSS-style color strings
+# 2002-12-15 fl   Added RGBA support
+# 2004-03-27 fl   Fixed remaining int() problems for Python 1.5.2
+# 2004-07-19 fl   Fixed gray/grey spelling issues
+# 2009-03-05 fl   Fixed rounding error in grayscale calculation
+#
+# Copyright (c) 2002-2004 by Secret Labs AB
+# Copyright (c) 2002-2004 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import re
+from functools import lru_cache
+
+from . import Image
+
+
+@lru_cache
+def getrgb(color: str) -> tuple[int, int, int] | tuple[int, int, int, int]:
+    """
+     Convert a color string to an RGB or RGBA tuple. If the string cannot be
+     parsed, this function raises a :py:exc:`ValueError` exception.
+
+    .. versionadded:: 1.1.4
+
+    :param color: A color string
+    :return: ``(red, green, blue[, alpha])``
+    """
+    if len(color) > 100:
+        msg = "color specifier is too long"
+        raise ValueError(msg)
+    color = color.lower()
+
+    rgb = colormap.get(color, None)
+    if rgb:
+        if isinstance(rgb, tuple):
+            return rgb
+        rgb_tuple = getrgb(rgb)
+        assert len(rgb_tuple) == 3
+        colormap[color] = rgb_tuple
+        return rgb_tuple
+
+    # check for known string formats
+    if re.match("#[a-f0-9]{3}$", color):
+        return int(color[1] * 2, 16), int(color[2] * 2, 16), int(color[3] * 2, 16)
+
+    if re.match("#[a-f0-9]{4}$", color):
+        return (
+            int(color[1] * 2, 16),
+            int(color[2] * 2, 16),
+            int(color[3] * 2, 16),
+            int(color[4] * 2, 16),
+        )
+
+    if re.match("#[a-f0-9]{6}$", color):
+        return int(color[1:3], 16), int(color[3:5], 16), int(color[5:7], 16)
+
+    if re.match("#[a-f0-9]{8}$", color):
+        return (
+            int(color[1:3], 16),
+            int(color[3:5], 16),
+            int(color[5:7], 16),
+            int(color[7:9], 16),
+        )
+
+    m = re.match(r"rgb\(\s*(\d+)\s*,\s*(\d+)\s*,\s*(\d+)\s*\)$", color)
+    if m:
+        return int(m.group(1)), int(m.group(2)), int(m.group(3))
+
+    m = re.match(r"rgb\(\s*(\d+)%\s*,\s*(\d+)%\s*,\s*(\d+)%\s*\)$", color)
+    if m:
+        return (
+            int((int(m.group(1)) * 255) / 100.0 + 0.5),
+            int((int(m.group(2)) * 255) / 100.0 + 0.5),
+            int((int(m.group(3)) * 255) / 100.0 + 0.5),
+        )
+
+    m = re.match(
+        r"hsl\(\s*(\d+\.?\d*)\s*,\s*(\d+\.?\d*)%\s*,\s*(\d+\.?\d*)%\s*\)$", color
+    )
+    if m:
+        from colorsys import hls_to_rgb
+
+        rgb_floats = hls_to_rgb(
+            float(m.group(1)) / 360.0,
+            float(m.group(3)) / 100.0,
+            float(m.group(2)) / 100.0,
+        )
+        return (
+            int(rgb_floats[0] * 255 + 0.5),
+            int(rgb_floats[1] * 255 + 0.5),
+            int(rgb_floats[2] * 255 + 0.5),
+        )
+
+    m = re.match(
+        r"hs[bv]\(\s*(\d+\.?\d*)\s*,\s*(\d+\.?\d*)%\s*,\s*(\d+\.?\d*)%\s*\)$", color
+    )
+    if m:
+        from colorsys import hsv_to_rgb
+
+        rgb_floats = hsv_to_rgb(
+            float(m.group(1)) / 360.0,
+            float(m.group(2)) / 100.0,
+            float(m.group(3)) / 100.0,
+        )
+        return (
+            int(rgb_floats[0] * 255 + 0.5),
+            int(rgb_floats[1] * 255 + 0.5),
+            int(rgb_floats[2] * 255 + 0.5),
+        )
+
+    m = re.match(r"rgba\(\s*(\d+)\s*,\s*(\d+)\s*,\s*(\d+)\s*,\s*(\d+)\s*\)$", color)
+    if m:
+        return int(m.group(1)), int(m.group(2)), int(m.group(3)), int(m.group(4))
+    msg = f"unknown color specifier: {repr(color)}"
+    raise ValueError(msg)
+
+
+@lru_cache
+def getcolor(color: str, mode: str) -> int | tuple[int, ...]:
+    """
+    Same as :py:func:`~PIL.ImageColor.getrgb` for most modes. However, if
+    ``mode`` is HSV, converts the RGB value to a HSV value, or if ``mode`` is
+    not color or a palette image, converts the RGB value to a grayscale value.
+    If the string cannot be parsed, this function raises a :py:exc:`ValueError`
+    exception.
+
+    .. versionadded:: 1.1.4
+
+    :param color: A color string
+    :param mode: Convert result to this mode
+    :return: ``graylevel, (graylevel, alpha) or (red, green, blue[, alpha])``
+    """
+    # same as getrgb, but converts the result to the given mode
+    rgb, alpha = getrgb(color), 255
+    if len(rgb) == 4:
+        alpha = rgb[3]
+        rgb = rgb[:3]
+
+    if mode == "HSV":
+        from colorsys import rgb_to_hsv
+
+        r, g, b = rgb
+        h, s, v = rgb_to_hsv(r / 255, g / 255, b / 255)
+        return int(h * 255), int(s * 255), int(v * 255)
+    elif Image.getmodebase(mode) == "L":
+        r, g, b = rgb
+        # ITU-R Recommendation 601-2 for nonlinear RGB
+        # scaled to 24 bits to match the convert's implementation.
+        graylevel = (r * 19595 + g * 38470 + b * 7471 + 0x8000) >> 16
+        if mode[-1] == "A":
+            return graylevel, alpha
+        return graylevel
+    elif mode[-1] == "A":
+        return rgb + (alpha,)
+    return rgb
+
+
+colormap: dict[str, str | tuple[int, int, int]] = {
+    # X11 colour table from https://drafts.csswg.org/css-color-4/, with
+    # gray/grey spelling issues fixed.  This is a superset of HTML 4.0
+    # colour names used in CSS 1.
+    "aliceblue": "#f0f8ff",
+    "antiquewhite": "#faebd7",
+    "aqua": "#00ffff",
+    "aquamarine": "#7fffd4",
+    "azure": "#f0ffff",
+    "beige": "#f5f5dc",
+    "bisque": "#ffe4c4",
+    "black": "#000000",
+    "blanchedalmond": "#ffebcd",
+    "blue": "#0000ff",
+    "blueviolet": "#8a2be2",
+    "brown": "#a52a2a",
+    "burlywood": "#deb887",
+    "cadetblue": "#5f9ea0",
+    "chartreuse": "#7fff00",
+    "chocolate": "#d2691e",
+    "coral": "#ff7f50",
+    "cornflowerblue": "#6495ed",
+    "cornsilk": "#fff8dc",
+    "crimson": "#dc143c",
+    "cyan": "#00ffff",
+    "darkblue": "#00008b",
+    "darkcyan": "#008b8b",
+    "darkgoldenrod": "#b8860b",
+    "darkgray": "#a9a9a9",
+    "darkgrey": "#a9a9a9",
+    "darkgreen": "#006400",
+    "darkkhaki": "#bdb76b",
+    "darkmagenta": "#8b008b",
+    "darkolivegreen": "#556b2f",
+    "darkorange": "#ff8c00",
+    "darkorchid": "#9932cc",
+    "darkred": "#8b0000",
+    "darksalmon": "#e9967a",
+    "darkseagreen": "#8fbc8f",
+    "darkslateblue": "#483d8b",
+    "darkslategray": "#2f4f4f",
+    "darkslategrey": "#2f4f4f",
+    "darkturquoise": "#00ced1",
+    "darkviolet": "#9400d3",
+    "deeppink": "#ff1493",
+    "deepskyblue": "#00bfff",
+    "dimgray": "#696969",
+    "dimgrey": "#696969",
+    "dodgerblue": "#1e90ff",
+    "firebrick": "#b22222",
+    "floralwhite": "#fffaf0",
+    "forestgreen": "#228b22",
+    "fuchsia": "#ff00ff",
+    "gainsboro": "#dcdcdc",
+    "ghostwhite": "#f8f8ff",
+    "gold": "#ffd700",
+    "goldenrod": "#daa520",
+    "gray": "#808080",
+    "grey": "#808080",
+    "green": "#008000",
+    "greenyellow": "#adff2f",
+    "honeydew": "#f0fff0",
+    "hotpink": "#ff69b4",
+    "indianred": "#cd5c5c",
+    "indigo": "#4b0082",
+    "ivory": "#fffff0",
+    "khaki": "#f0e68c",
+    "lavender": "#e6e6fa",
+    "lavenderblush": "#fff0f5",
+    "lawngreen": "#7cfc00",
+    "lemonchiffon": "#fffacd",
+    "lightblue": "#add8e6",
+    "lightcoral": "#f08080",
+    "lightcyan": "#e0ffff",
+    "lightgoldenrodyellow": "#fafad2",
+    "lightgreen": "#90ee90",
+    "lightgray": "#d3d3d3",
+    "lightgrey": "#d3d3d3",
+    "lightpink": "#ffb6c1",
+    "lightsalmon": "#ffa07a",
+    "lightseagreen": "#20b2aa",
+    "lightskyblue": "#87cefa",
+    "lightslategray": "#778899",
+    "lightslategrey": "#778899",
+    "lightsteelblue": "#b0c4de",
+    "lightyellow": "#ffffe0",
+    "lime": "#00ff00",
+    "limegreen": "#32cd32",
+    "linen": "#faf0e6",
+    "magenta": "#ff00ff",
+    "maroon": "#800000",
+    "mediumaquamarine": "#66cdaa",
+    "mediumblue": "#0000cd",
+    "mediumorchid": "#ba55d3",
+    "mediumpurple": "#9370db",
+    "mediumseagreen": "#3cb371",
+    "mediumslateblue": "#7b68ee",
+    "mediumspringgreen": "#00fa9a",
+    "mediumturquoise": "#48d1cc",
+    "mediumvioletred": "#c71585",
+    "midnightblue": "#191970",
+    "mintcream": "#f5fffa",
+    "mistyrose": "#ffe4e1",
+    "moccasin": "#ffe4b5",
+    "navajowhite": "#ffdead",
+    "navy": "#000080",
+    "oldlace": "#fdf5e6",
+    "olive": "#808000",
+    "olivedrab": "#6b8e23",
+    "orange": "#ffa500",
+    "orangered": "#ff4500",
+    "orchid": "#da70d6",
+    "palegoldenrod": "#eee8aa",
+    "palegreen": "#98fb98",
+    "paleturquoise": "#afeeee",
+    "palevioletred": "#db7093",
+    "papayawhip": "#ffefd5",
+    "peachpuff": "#ffdab9",
+    "peru": "#cd853f",
+    "pink": "#ffc0cb",
+    "plum": "#dda0dd",
+    "powderblue": "#b0e0e6",
+    "purple": "#800080",
+    "rebeccapurple": "#663399",
+    "red": "#ff0000",
+    "rosybrown": "#bc8f8f",
+    "royalblue": "#4169e1",
+    "saddlebrown": "#8b4513",
+    "salmon": "#fa8072",
+    "sandybrown": "#f4a460",
+    "seagreen": "#2e8b57",
+    "seashell": "#fff5ee",
+    "sienna": "#a0522d",
+    "silver": "#c0c0c0",
+    "skyblue": "#87ceeb",
+    "slateblue": "#6a5acd",
+    "slategray": "#708090",
+    "slategrey": "#708090",
+    "snow": "#fffafa",
+    "springgreen": "#00ff7f",
+    "steelblue": "#4682b4",
+    "tan": "#d2b48c",
+    "teal": "#008080",
+    "thistle": "#d8bfd8",
+    "tomato": "#ff6347",
+    "turquoise": "#40e0d0",
+    "violet": "#ee82ee",
+    "wheat": "#f5deb3",
+    "white": "#ffffff",
+    "whitesmoke": "#f5f5f5",
+    "yellow": "#ffff00",
+    "yellowgreen": "#9acd32",
+}
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw.py
new file mode 100644
index 00000000..d8e4c0c6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw.py
@@ -0,0 +1,1218 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# drawing interface operations
+#
+# History:
+# 1996-04-13 fl   Created (experimental)
+# 1996-08-07 fl   Filled polygons, ellipses.
+# 1996-08-13 fl   Added text support
+# 1998-06-28 fl   Handle I and F images
+# 1998-12-29 fl   Added arc; use arc primitive to draw ellipses
+# 1999-01-10 fl   Added shape stuff (experimental)
+# 1999-02-06 fl   Added bitmap support
+# 1999-02-11 fl   Changed all primitives to take options
+# 1999-02-20 fl   Fixed backwards compatibility
+# 2000-10-12 fl   Copy on write, when necessary
+# 2001-02-18 fl   Use default ink for bitmap/text also in fill mode
+# 2002-10-24 fl   Added support for CSS-style color strings
+# 2002-12-10 fl   Added experimental support for RGBA-on-RGB drawing
+# 2002-12-11 fl   Refactored low-level drawing API (work in progress)
+# 2004-08-26 fl   Made Draw() a factory function, added getdraw() support
+# 2004-09-04 fl   Added width support to line primitive
+# 2004-09-10 fl   Added font mode handling
+# 2006-06-19 fl   Added font bearing support (getmask2)
+#
+# Copyright (c) 1997-2006 by Secret Labs AB
+# Copyright (c) 1996-2006 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import math
+import struct
+from collections.abc import Sequence
+from types import ModuleType
+from typing import TYPE_CHECKING, Any, AnyStr, Callable, Union, cast
+
+from . import Image, ImageColor
+from ._deprecate import deprecate
+from ._typing import Coords
+
+# experimental access to the outline API
+Outline: Callable[[], Image.core._Outline] | None
+try:
+    Outline = Image.core.outline
+except AttributeError:
+    Outline = None
+
+if TYPE_CHECKING:
+    from . import ImageDraw2, ImageFont
+
+_Ink = Union[float, tuple[int, ...], str]
+
+"""
+A simple 2D drawing interface for PIL images.
+<p>
+Application code should use the <b>Draw</b> factory, instead of
+directly.
+"""
+
+
+class ImageDraw:
+    font: (
+        ImageFont.ImageFont | ImageFont.FreeTypeFont | ImageFont.TransposedFont | None
+    ) = None
+
+    def __init__(self, im: Image.Image, mode: str | None = None) -> None:
+        """
+        Create a drawing instance.
+
+        :param im: The image to draw in.
+        :param mode: Optional mode to use for color values.  For RGB
+           images, this argument can be RGB or RGBA (to blend the
+           drawing into the image).  For all other modes, this argument
+           must be the same as the image mode.  If omitted, the mode
+           defaults to the mode of the image.
+        """
+        im.load()
+        if im.readonly:
+            im._copy()  # make it writeable
+        blend = 0
+        if mode is None:
+            mode = im.mode
+        if mode != im.mode:
+            if mode == "RGBA" and im.mode == "RGB":
+                blend = 1
+            else:
+                msg = "mode mismatch"
+                raise ValueError(msg)
+        if mode == "P":
+            self.palette = im.palette
+        else:
+            self.palette = None
+        self._image = im
+        self.im = im.im
+        self.draw = Image.core.draw(self.im, blend)
+        self.mode = mode
+        if mode in ("I", "F"):
+            self.ink = self.draw.draw_ink(1)
+        else:
+            self.ink = self.draw.draw_ink(-1)
+        if mode in ("1", "P", "I", "F"):
+            # FIXME: fix Fill2 to properly support matte for I+F images
+            self.fontmode = "1"
+        else:
+            self.fontmode = "L"  # aliasing is okay for other modes
+        self.fill = False
+
+    def getfont(
+        self,
+    ) -> ImageFont.ImageFont | ImageFont.FreeTypeFont | ImageFont.TransposedFont:
+        """
+        Get the current default font.
+
+        To set the default font for this ImageDraw instance::
+
+            from PIL import ImageDraw, ImageFont
+            draw.font = ImageFont.truetype("Tests/fonts/FreeMono.ttf")
+
+        To set the default font for all future ImageDraw instances::
+
+            from PIL import ImageDraw, ImageFont
+            ImageDraw.ImageDraw.font = ImageFont.truetype("Tests/fonts/FreeMono.ttf")
+
+        If the current default font is ``None``,
+        it is initialized with ``ImageFont.load_default()``.
+
+        :returns: An image font."""
+        if not self.font:
+            # FIXME: should add a font repository
+            from . import ImageFont
+
+            self.font = ImageFont.load_default()
+        return self.font
+
+    def _getfont(
+        self, font_size: float | None
+    ) -> ImageFont.ImageFont | ImageFont.FreeTypeFont | ImageFont.TransposedFont:
+        if font_size is not None:
+            from . import ImageFont
+
+            return ImageFont.load_default(font_size)
+        else:
+            return self.getfont()
+
+    def _getink(
+        self, ink: _Ink | None, fill: _Ink | None = None
+    ) -> tuple[int | None, int | None]:
+        result_ink = None
+        result_fill = None
+        if ink is None and fill is None:
+            if self.fill:
+                result_fill = self.ink
+            else:
+                result_ink = self.ink
+        else:
+            if ink is not None:
+                if isinstance(ink, str):
+                    ink = ImageColor.getcolor(ink, self.mode)
+                if self.palette and isinstance(ink, tuple):
+                    ink = self.palette.getcolor(ink, self._image)
+                result_ink = self.draw.draw_ink(ink)
+            if fill is not None:
+                if isinstance(fill, str):
+                    fill = ImageColor.getcolor(fill, self.mode)
+                if self.palette and isinstance(fill, tuple):
+                    fill = self.palette.getcolor(fill, self._image)
+                result_fill = self.draw.draw_ink(fill)
+        return result_ink, result_fill
+
+    def arc(
+        self,
+        xy: Coords,
+        start: float,
+        end: float,
+        fill: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw an arc."""
+        ink, fill = self._getink(fill)
+        if ink is not None:
+            self.draw.draw_arc(xy, start, end, ink, width)
+
+    def bitmap(
+        self, xy: Sequence[int], bitmap: Image.Image, fill: _Ink | None = None
+    ) -> None:
+        """Draw a bitmap."""
+        bitmap.load()
+        ink, fill = self._getink(fill)
+        if ink is None:
+            ink = fill
+        if ink is not None:
+            self.draw.draw_bitmap(xy, bitmap.im, ink)
+
+    def chord(
+        self,
+        xy: Coords,
+        start: float,
+        end: float,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a chord."""
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_chord(xy, start, end, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            self.draw.draw_chord(xy, start, end, ink, 0, width)
+
+    def ellipse(
+        self,
+        xy: Coords,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw an ellipse."""
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_ellipse(xy, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            self.draw.draw_ellipse(xy, ink, 0, width)
+
+    def circle(
+        self,
+        xy: Sequence[float],
+        radius: float,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a circle given center coordinates and a radius."""
+        ellipse_xy = (xy[0] - radius, xy[1] - radius, xy[0] + radius, xy[1] + radius)
+        self.ellipse(ellipse_xy, fill, outline, width)
+
+    def line(
+        self,
+        xy: Coords,
+        fill: _Ink | None = None,
+        width: int = 0,
+        joint: str | None = None,
+    ) -> None:
+        """Draw a line, or a connected sequence of line segments."""
+        ink = self._getink(fill)[0]
+        if ink is not None:
+            self.draw.draw_lines(xy, ink, width)
+            if joint == "curve" and width > 4:
+                points: Sequence[Sequence[float]]
+                if isinstance(xy[0], (list, tuple)):
+                    points = cast(Sequence[Sequence[float]], xy)
+                else:
+                    points = [
+                        cast(Sequence[float], tuple(xy[i : i + 2]))
+                        for i in range(0, len(xy), 2)
+                    ]
+                for i in range(1, len(points) - 1):
+                    point = points[i]
+                    angles = [
+                        math.degrees(math.atan2(end[0] - start[0], start[1] - end[1]))
+                        % 360
+                        for start, end in (
+                            (points[i - 1], point),
+                            (point, points[i + 1]),
+                        )
+                    ]
+                    if angles[0] == angles[1]:
+                        # This is a straight line, so no joint is required
+                        continue
+
+                    def coord_at_angle(
+                        coord: Sequence[float], angle: float
+                    ) -> tuple[float, ...]:
+                        x, y = coord
+                        angle -= 90
+                        distance = width / 2 - 1
+                        return tuple(
+                            p + (math.floor(p_d) if p_d > 0 else math.ceil(p_d))
+                            for p, p_d in (
+                                (x, distance * math.cos(math.radians(angle))),
+                                (y, distance * math.sin(math.radians(angle))),
+                            )
+                        )
+
+                    flipped = (
+                        angles[1] > angles[0] and angles[1] - 180 > angles[0]
+                    ) or (angles[1] < angles[0] and angles[1] + 180 > angles[0])
+                    coords = [
+                        (point[0] - width / 2 + 1, point[1] - width / 2 + 1),
+                        (point[0] + width / 2 - 1, point[1] + width / 2 - 1),
+                    ]
+                    if flipped:
+                        start, end = (angles[1] + 90, angles[0] + 90)
+                    else:
+                        start, end = (angles[0] - 90, angles[1] - 90)
+                    self.pieslice(coords, start - 90, end - 90, fill)
+
+                    if width > 8:
+                        # Cover potential gaps between the line and the joint
+                        if flipped:
+                            gap_coords = [
+                                coord_at_angle(point, angles[0] + 90),
+                                point,
+                                coord_at_angle(point, angles[1] + 90),
+                            ]
+                        else:
+                            gap_coords = [
+                                coord_at_angle(point, angles[0] - 90),
+                                point,
+                                coord_at_angle(point, angles[1] - 90),
+                            ]
+                        self.line(gap_coords, fill, width=3)
+
+    def shape(
+        self,
+        shape: Image.core._Outline,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+    ) -> None:
+        """(Experimental) Draw a shape."""
+        shape.close()
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_outline(shape, fill_ink, 1)
+        if ink is not None and ink != fill_ink:
+            self.draw.draw_outline(shape, ink, 0)
+
+    def pieslice(
+        self,
+        xy: Coords,
+        start: float,
+        end: float,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a pieslice."""
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_pieslice(xy, start, end, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            self.draw.draw_pieslice(xy, start, end, ink, 0, width)
+
+    def point(self, xy: Coords, fill: _Ink | None = None) -> None:
+        """Draw one or more individual pixels."""
+        ink, fill = self._getink(fill)
+        if ink is not None:
+            self.draw.draw_points(xy, ink)
+
+    def polygon(
+        self,
+        xy: Coords,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a polygon."""
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_polygon(xy, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            if width == 1:
+                self.draw.draw_polygon(xy, ink, 0, width)
+            elif self.im is not None:
+                # To avoid expanding the polygon outwards,
+                # use the fill as a mask
+                mask = Image.new("1", self.im.size)
+                mask_ink = self._getink(1)[0]
+
+                fill_im = mask.copy()
+                draw = Draw(fill_im)
+                draw.draw.draw_polygon(xy, mask_ink, 1)
+
+                ink_im = mask.copy()
+                draw = Draw(ink_im)
+                width = width * 2 - 1
+                draw.draw.draw_polygon(xy, mask_ink, 0, width)
+
+                mask.paste(ink_im, mask=fill_im)
+
+                im = Image.new(self.mode, self.im.size)
+                draw = Draw(im)
+                draw.draw.draw_polygon(xy, ink, 0, width)
+                self.im.paste(im.im, (0, 0) + im.size, mask.im)
+
+    def regular_polygon(
+        self,
+        bounding_circle: Sequence[Sequence[float] | float],
+        n_sides: int,
+        rotation: float = 0,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a regular polygon."""
+        xy = _compute_regular_polygon_vertices(bounding_circle, n_sides, rotation)
+        self.polygon(xy, fill, outline, width)
+
+    def rectangle(
+        self,
+        xy: Coords,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+    ) -> None:
+        """Draw a rectangle."""
+        ink, fill_ink = self._getink(outline, fill)
+        if fill_ink is not None:
+            self.draw.draw_rectangle(xy, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            self.draw.draw_rectangle(xy, ink, 0, width)
+
+    def rounded_rectangle(
+        self,
+        xy: Coords,
+        radius: float = 0,
+        fill: _Ink | None = None,
+        outline: _Ink | None = None,
+        width: int = 1,
+        *,
+        corners: tuple[bool, bool, bool, bool] | None = None,
+    ) -> None:
+        """Draw a rounded rectangle."""
+        if isinstance(xy[0], (list, tuple)):
+            (x0, y0), (x1, y1) = cast(Sequence[Sequence[float]], xy)
+        else:
+            x0, y0, x1, y1 = cast(Sequence[float], xy)
+        if x1 < x0:
+            msg = "x1 must be greater than or equal to x0"
+            raise ValueError(msg)
+        if y1 < y0:
+            msg = "y1 must be greater than or equal to y0"
+            raise ValueError(msg)
+        if corners is None:
+            corners = (True, True, True, True)
+
+        d = radius * 2
+
+        x0 = round(x0)
+        y0 = round(y0)
+        x1 = round(x1)
+        y1 = round(y1)
+        full_x, full_y = False, False
+        if all(corners):
+            full_x = d >= x1 - x0 - 1
+            if full_x:
+                # The two left and two right corners are joined
+                d = x1 - x0
+            full_y = d >= y1 - y0 - 1
+            if full_y:
+                # The two top and two bottom corners are joined
+                d = y1 - y0
+            if full_x and full_y:
+                # If all corners are joined, that is a circle
+                return self.ellipse(xy, fill, outline, width)
+
+        if d == 0 or not any(corners):
+            # If the corners have no curve,
+            # or there are no corners,
+            # that is a rectangle
+            return self.rectangle(xy, fill, outline, width)
+
+        r = int(d // 2)
+        ink, fill_ink = self._getink(outline, fill)
+
+        def draw_corners(pieslice: bool) -> None:
+            parts: tuple[tuple[tuple[float, float, float, float], int, int], ...]
+            if full_x:
+                # Draw top and bottom halves
+                parts = (
+                    ((x0, y0, x0 + d, y0 + d), 180, 360),
+                    ((x0, y1 - d, x0 + d, y1), 0, 180),
+                )
+            elif full_y:
+                # Draw left and right halves
+                parts = (
+                    ((x0, y0, x0 + d, y0 + d), 90, 270),
+                    ((x1 - d, y0, x1, y0 + d), 270, 90),
+                )
+            else:
+                # Draw four separate corners
+                parts = tuple(
+                    part
+                    for i, part in enumerate(
+                        (
+                            ((x0, y0, x0 + d, y0 + d), 180, 270),
+                            ((x1 - d, y0, x1, y0 + d), 270, 360),
+                            ((x1 - d, y1 - d, x1, y1), 0, 90),
+                            ((x0, y1 - d, x0 + d, y1), 90, 180),
+                        )
+                    )
+                    if corners[i]
+                )
+            for part in parts:
+                if pieslice:
+                    self.draw.draw_pieslice(*(part + (fill_ink, 1)))
+                else:
+                    self.draw.draw_arc(*(part + (ink, width)))
+
+        if fill_ink is not None:
+            draw_corners(True)
+
+            if full_x:
+                self.draw.draw_rectangle((x0, y0 + r + 1, x1, y1 - r - 1), fill_ink, 1)
+            elif x1 - r - 1 > x0 + r + 1:
+                self.draw.draw_rectangle((x0 + r + 1, y0, x1 - r - 1, y1), fill_ink, 1)
+            if not full_x and not full_y:
+                left = [x0, y0, x0 + r, y1]
+                if corners[0]:
+                    left[1] += r + 1
+                if corners[3]:
+                    left[3] -= r + 1
+                self.draw.draw_rectangle(left, fill_ink, 1)
+
+                right = [x1 - r, y0, x1, y1]
+                if corners[1]:
+                    right[1] += r + 1
+                if corners[2]:
+                    right[3] -= r + 1
+                self.draw.draw_rectangle(right, fill_ink, 1)
+        if ink is not None and ink != fill_ink and width != 0:
+            draw_corners(False)
+
+            if not full_x:
+                top = [x0, y0, x1, y0 + width - 1]
+                if corners[0]:
+                    top[0] += r + 1
+                if corners[1]:
+                    top[2] -= r + 1
+                self.draw.draw_rectangle(top, ink, 1)
+
+                bottom = [x0, y1 - width + 1, x1, y1]
+                if corners[3]:
+                    bottom[0] += r + 1
+                if corners[2]:
+                    bottom[2] -= r + 1
+                self.draw.draw_rectangle(bottom, ink, 1)
+            if not full_y:
+                left = [x0, y0, x0 + width - 1, y1]
+                if corners[0]:
+                    left[1] += r + 1
+                if corners[3]:
+                    left[3] -= r + 1
+                self.draw.draw_rectangle(left, ink, 1)
+
+                right = [x1 - width + 1, y0, x1, y1]
+                if corners[1]:
+                    right[1] += r + 1
+                if corners[2]:
+                    right[3] -= r + 1
+                self.draw.draw_rectangle(right, ink, 1)
+
+    def _multiline_check(self, text: AnyStr) -> bool:
+        split_character = "\n" if isinstance(text, str) else b"\n"
+
+        return split_character in text
+
+    def _multiline_split(self, text: AnyStr) -> list[AnyStr]:
+        return text.split("\n" if isinstance(text, str) else b"\n")
+
+    def _multiline_spacing(
+        self,
+        font: ImageFont.ImageFont | ImageFont.FreeTypeFont | ImageFont.TransposedFont,
+        spacing: float,
+        stroke_width: float,
+    ) -> float:
+        return (
+            self.textbbox((0, 0), "A", font, stroke_width=stroke_width)[3]
+            + stroke_width
+            + spacing
+        )
+
+    def text(
+        self,
+        xy: tuple[float, float],
+        text: AnyStr,
+        fill: _Ink | None = None,
+        font: (
+            ImageFont.ImageFont
+            | ImageFont.FreeTypeFont
+            | ImageFont.TransposedFont
+            | None
+        ) = None,
+        anchor: str | None = None,
+        spacing: float = 4,
+        align: str = "left",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        stroke_fill: _Ink | None = None,
+        embedded_color: bool = False,
+        *args: Any,
+        **kwargs: Any,
+    ) -> None:
+        """Draw text."""
+        if embedded_color and self.mode not in ("RGB", "RGBA"):
+            msg = "Embedded color supported only in RGB and RGBA modes"
+            raise ValueError(msg)
+
+        if font is None:
+            font = self._getfont(kwargs.get("font_size"))
+
+        if self._multiline_check(text):
+            return self.multiline_text(
+                xy,
+                text,
+                fill,
+                font,
+                anchor,
+                spacing,
+                align,
+                direction,
+                features,
+                language,
+                stroke_width,
+                stroke_fill,
+                embedded_color,
+            )
+
+        def getink(fill: _Ink | None) -> int:
+            ink, fill_ink = self._getink(fill)
+            if ink is None:
+                assert fill_ink is not None
+                return fill_ink
+            return ink
+
+        def draw_text(ink: int, stroke_width: float = 0) -> None:
+            mode = self.fontmode
+            if stroke_width == 0 and embedded_color:
+                mode = "RGBA"
+            coord = []
+            for i in range(2):
+                coord.append(int(xy[i]))
+            start = (math.modf(xy[0])[0], math.modf(xy[1])[0])
+            try:
+                mask, offset = font.getmask2(  # type: ignore[union-attr,misc]
+                    text,
+                    mode,
+                    direction=direction,
+                    features=features,
+                    language=language,
+                    stroke_width=stroke_width,
+                    anchor=anchor,
+                    ink=ink,
+                    start=start,
+                    *args,
+                    **kwargs,
+                )
+                coord = [coord[0] + offset[0], coord[1] + offset[1]]
+            except AttributeError:
+                try:
+                    mask = font.getmask(  # type: ignore[misc]
+                        text,
+                        mode,
+                        direction,
+                        features,
+                        language,
+                        stroke_width,
+                        anchor,
+                        ink,
+                        start=start,
+                        *args,
+                        **kwargs,
+                    )
+                except TypeError:
+                    mask = font.getmask(text)
+            if mode == "RGBA":
+                # font.getmask2(mode="RGBA") returns color in RGB bands and mask in A
+                # extract mask and set text alpha
+                color, mask = mask, mask.getband(3)
+                ink_alpha = struct.pack("i", ink)[3]
+                color.fillband(3, ink_alpha)
+                x, y = coord
+                if self.im is not None:
+                    self.im.paste(
+                        color, (x, y, x + mask.size[0], y + mask.size[1]), mask
+                    )
+            else:
+                self.draw.draw_bitmap(coord, mask, ink)
+
+        ink = getink(fill)
+        if ink is not None:
+            stroke_ink = None
+            if stroke_width:
+                stroke_ink = getink(stroke_fill) if stroke_fill is not None else ink
+
+            if stroke_ink is not None:
+                # Draw stroked text
+                draw_text(stroke_ink, stroke_width)
+
+                # Draw normal text
+                draw_text(ink, 0)
+            else:
+                # Only draw normal text
+                draw_text(ink)
+
+    def multiline_text(
+        self,
+        xy: tuple[float, float],
+        text: AnyStr,
+        fill: _Ink | None = None,
+        font: (
+            ImageFont.ImageFont
+            | ImageFont.FreeTypeFont
+            | ImageFont.TransposedFont
+            | None
+        ) = None,
+        anchor: str | None = None,
+        spacing: float = 4,
+        align: str = "left",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        stroke_fill: _Ink | None = None,
+        embedded_color: bool = False,
+        *,
+        font_size: float | None = None,
+    ) -> None:
+        if direction == "ttb":
+            msg = "ttb direction is unsupported for multiline text"
+            raise ValueError(msg)
+
+        if anchor is None:
+            anchor = "la"
+        elif len(anchor) != 2:
+            msg = "anchor must be a 2 character string"
+            raise ValueError(msg)
+        elif anchor[1] in "tb":
+            msg = "anchor not supported for multiline text"
+            raise ValueError(msg)
+
+        if font is None:
+            font = self._getfont(font_size)
+
+        widths = []
+        max_width: float = 0
+        lines = self._multiline_split(text)
+        line_spacing = self._multiline_spacing(font, spacing, stroke_width)
+        for line in lines:
+            line_width = self.textlength(
+                line, font, direction=direction, features=features, language=language
+            )
+            widths.append(line_width)
+            max_width = max(max_width, line_width)
+
+        top = xy[1]
+        if anchor[1] == "m":
+            top -= (len(lines) - 1) * line_spacing / 2.0
+        elif anchor[1] == "d":
+            top -= (len(lines) - 1) * line_spacing
+
+        for idx, line in enumerate(lines):
+            left = xy[0]
+            width_difference = max_width - widths[idx]
+
+            # first align left by anchor
+            if anchor[0] == "m":
+                left -= width_difference / 2.0
+            elif anchor[0] == "r":
+                left -= width_difference
+
+            # then align by align parameter
+            if align == "left":
+                pass
+            elif align == "center":
+                left += width_difference / 2.0
+            elif align == "right":
+                left += width_difference
+            else:
+                msg = 'align must be "left", "center" or "right"'
+                raise ValueError(msg)
+
+            self.text(
+                (left, top),
+                line,
+                fill,
+                font,
+                anchor,
+                direction=direction,
+                features=features,
+                language=language,
+                stroke_width=stroke_width,
+                stroke_fill=stroke_fill,
+                embedded_color=embedded_color,
+            )
+            top += line_spacing
+
+    def textlength(
+        self,
+        text: AnyStr,
+        font: (
+            ImageFont.ImageFont
+            | ImageFont.FreeTypeFont
+            | ImageFont.TransposedFont
+            | None
+        ) = None,
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        embedded_color: bool = False,
+        *,
+        font_size: float | None = None,
+    ) -> float:
+        """Get the length of a given string, in pixels with 1/64 precision."""
+        if self._multiline_check(text):
+            msg = "can't measure length of multiline text"
+            raise ValueError(msg)
+        if embedded_color and self.mode not in ("RGB", "RGBA"):
+            msg = "Embedded color supported only in RGB and RGBA modes"
+            raise ValueError(msg)
+
+        if font is None:
+            font = self._getfont(font_size)
+        mode = "RGBA" if embedded_color else self.fontmode
+        return font.getlength(text, mode, direction, features, language)
+
+    def textbbox(
+        self,
+        xy: tuple[float, float],
+        text: AnyStr,
+        font: (
+            ImageFont.ImageFont
+            | ImageFont.FreeTypeFont
+            | ImageFont.TransposedFont
+            | None
+        ) = None,
+        anchor: str | None = None,
+        spacing: float = 4,
+        align: str = "left",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        embedded_color: bool = False,
+        *,
+        font_size: float | None = None,
+    ) -> tuple[float, float, float, float]:
+        """Get the bounding box of a given string, in pixels."""
+        if embedded_color and self.mode not in ("RGB", "RGBA"):
+            msg = "Embedded color supported only in RGB and RGBA modes"
+            raise ValueError(msg)
+
+        if font is None:
+            font = self._getfont(font_size)
+
+        if self._multiline_check(text):
+            return self.multiline_textbbox(
+                xy,
+                text,
+                font,
+                anchor,
+                spacing,
+                align,
+                direction,
+                features,
+                language,
+                stroke_width,
+                embedded_color,
+            )
+
+        mode = "RGBA" if embedded_color else self.fontmode
+        bbox = font.getbbox(
+            text, mode, direction, features, language, stroke_width, anchor
+        )
+        return bbox[0] + xy[0], bbox[1] + xy[1], bbox[2] + xy[0], bbox[3] + xy[1]
+
+    def multiline_textbbox(
+        self,
+        xy: tuple[float, float],
+        text: AnyStr,
+        font: (
+            ImageFont.ImageFont
+            | ImageFont.FreeTypeFont
+            | ImageFont.TransposedFont
+            | None
+        ) = None,
+        anchor: str | None = None,
+        spacing: float = 4,
+        align: str = "left",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        embedded_color: bool = False,
+        *,
+        font_size: float | None = None,
+    ) -> tuple[float, float, float, float]:
+        if direction == "ttb":
+            msg = "ttb direction is unsupported for multiline text"
+            raise ValueError(msg)
+
+        if anchor is None:
+            anchor = "la"
+        elif len(anchor) != 2:
+            msg = "anchor must be a 2 character string"
+            raise ValueError(msg)
+        elif anchor[1] in "tb":
+            msg = "anchor not supported for multiline text"
+            raise ValueError(msg)
+
+        if font is None:
+            font = self._getfont(font_size)
+
+        widths = []
+        max_width: float = 0
+        lines = self._multiline_split(text)
+        line_spacing = self._multiline_spacing(font, spacing, stroke_width)
+        for line in lines:
+            line_width = self.textlength(
+                line,
+                font,
+                direction=direction,
+                features=features,
+                language=language,
+                embedded_color=embedded_color,
+            )
+            widths.append(line_width)
+            max_width = max(max_width, line_width)
+
+        top = xy[1]
+        if anchor[1] == "m":
+            top -= (len(lines) - 1) * line_spacing / 2.0
+        elif anchor[1] == "d":
+            top -= (len(lines) - 1) * line_spacing
+
+        bbox: tuple[float, float, float, float] | None = None
+
+        for idx, line in enumerate(lines):
+            left = xy[0]
+            width_difference = max_width - widths[idx]
+
+            # first align left by anchor
+            if anchor[0] == "m":
+                left -= width_difference / 2.0
+            elif anchor[0] == "r":
+                left -= width_difference
+
+            # then align by align parameter
+            if align == "left":
+                pass
+            elif align == "center":
+                left += width_difference / 2.0
+            elif align == "right":
+                left += width_difference
+            else:
+                msg = 'align must be "left", "center" or "right"'
+                raise ValueError(msg)
+
+            bbox_line = self.textbbox(
+                (left, top),
+                line,
+                font,
+                anchor,
+                direction=direction,
+                features=features,
+                language=language,
+                stroke_width=stroke_width,
+                embedded_color=embedded_color,
+            )
+            if bbox is None:
+                bbox = bbox_line
+            else:
+                bbox = (
+                    min(bbox[0], bbox_line[0]),
+                    min(bbox[1], bbox_line[1]),
+                    max(bbox[2], bbox_line[2]),
+                    max(bbox[3], bbox_line[3]),
+                )
+
+            top += line_spacing
+
+        if bbox is None:
+            return xy[0], xy[1], xy[0], xy[1]
+        return bbox
+
+
+def Draw(im: Image.Image, mode: str | None = None) -> ImageDraw:
+    """
+    A simple 2D drawing interface for PIL images.
+
+    :param im: The image to draw in.
+    :param mode: Optional mode to use for color values.  For RGB
+       images, this argument can be RGB or RGBA (to blend the
+       drawing into the image).  For all other modes, this argument
+       must be the same as the image mode.  If omitted, the mode
+       defaults to the mode of the image.
+    """
+    try:
+        return getattr(im, "getdraw")(mode)
+    except AttributeError:
+        return ImageDraw(im, mode)
+
+
+def getdraw(
+    im: Image.Image | None = None, hints: list[str] | None = None
+) -> tuple[ImageDraw2.Draw | None, ModuleType]:
+    """
+    :param im: The image to draw in.
+    :param hints: An optional list of hints. Deprecated.
+    :returns: A (drawing context, drawing resource factory) tuple.
+    """
+    if hints is not None:
+        deprecate("'hints' parameter", 12)
+    from . import ImageDraw2
+
+    draw = ImageDraw2.Draw(im) if im is not None else None
+    return draw, ImageDraw2
+
+
+def floodfill(
+    image: Image.Image,
+    xy: tuple[int, int],
+    value: float | tuple[int, ...],
+    border: float | tuple[int, ...] | None = None,
+    thresh: float = 0,
+) -> None:
+    """
+    .. warning:: This method is experimental.
+
+    Fills a bounded region with a given color.
+
+    :param image: Target image.
+    :param xy: Seed position (a 2-item coordinate tuple). See
+        :ref:`coordinate-system`.
+    :param value: Fill color.
+    :param border: Optional border value.  If given, the region consists of
+        pixels with a color different from the border color.  If not given,
+        the region consists of pixels having the same color as the seed
+        pixel.
+    :param thresh: Optional threshold value which specifies a maximum
+        tolerable difference of a pixel value from the 'background' in
+        order for it to be replaced. Useful for filling regions of
+        non-homogeneous, but similar, colors.
+    """
+    # based on an implementation by Eric S. Raymond
+    # amended by yo1995 @20180806
+    pixel = image.load()
+    assert pixel is not None
+    x, y = xy
+    try:
+        background = pixel[x, y]
+        if _color_diff(value, background) <= thresh:
+            return  # seed point already has fill color
+        pixel[x, y] = value
+    except (ValueError, IndexError):
+        return  # seed point outside image
+    edge = {(x, y)}
+    # use a set to keep record of current and previous edge pixels
+    # to reduce memory consumption
+    full_edge = set()
+    while edge:
+        new_edge = set()
+        for x, y in edge:  # 4 adjacent method
+            for s, t in ((x + 1, y), (x - 1, y), (x, y + 1), (x, y - 1)):
+                # If already processed, or if a coordinate is negative, skip
+                if (s, t) in full_edge or s < 0 or t < 0:
+                    continue
+                try:
+                    p = pixel[s, t]
+                except (ValueError, IndexError):
+                    pass
+                else:
+                    full_edge.add((s, t))
+                    if border is None:
+                        fill = _color_diff(p, background) <= thresh
+                    else:
+                        fill = p not in (value, border)
+                    if fill:
+                        pixel[s, t] = value
+                        new_edge.add((s, t))
+        full_edge = edge  # discard pixels processed
+        edge = new_edge
+
+
+def _compute_regular_polygon_vertices(
+    bounding_circle: Sequence[Sequence[float] | float], n_sides: int, rotation: float
+) -> list[tuple[float, float]]:
+    """
+    Generate a list of vertices for a 2D regular polygon.
+
+    :param bounding_circle: The bounding circle is a sequence defined
+        by a point and radius. The polygon is inscribed in this circle.
+        (e.g. ``bounding_circle=(x, y, r)`` or ``((x, y), r)``)
+    :param n_sides: Number of sides
+        (e.g. ``n_sides=3`` for a triangle, ``6`` for a hexagon)
+    :param rotation: Apply an arbitrary rotation to the polygon
+        (e.g. ``rotation=90``, applies a 90 degree rotation)
+    :return: List of regular polygon vertices
+        (e.g. ``[(25, 50), (50, 50), (50, 25), (25, 25)]``)
+
+    How are the vertices computed?
+    1. Compute the following variables
+        - theta: Angle between the apothem & the nearest polygon vertex
+        - side_length: Length of each polygon edge
+        - centroid: Center of bounding circle (1st, 2nd elements of bounding_circle)
+        - polygon_radius: Polygon radius (last element of bounding_circle)
+        - angles: Location of each polygon vertex in polar grid
+            (e.g. A square with 0 degree rotation => [225.0, 315.0, 45.0, 135.0])
+
+    2. For each angle in angles, get the polygon vertex at that angle
+        The vertex is computed using the equation below.
+            X= xcos(φ) + ysin(φ)
+            Y= −xsin(φ) + ycos(φ)
+
+        Note:
+            φ = angle in degrees
+            x = 0
+            y = polygon_radius
+
+        The formula above assumes rotation around the origin.
+        In our case, we are rotating around the centroid.
+        To account for this, we use the formula below
+            X = xcos(φ) + ysin(φ) + centroid_x
+            Y = −xsin(φ) + ycos(φ) + centroid_y
+    """
+    # 1. Error Handling
+    # 1.1 Check `n_sides` has an appropriate value
+    if not isinstance(n_sides, int):
+        msg = "n_sides should be an int"  # type: ignore[unreachable]
+        raise TypeError(msg)
+    if n_sides < 3:
+        msg = "n_sides should be an int > 2"
+        raise ValueError(msg)
+
+    # 1.2 Check `bounding_circle` has an appropriate value
+    if not isinstance(bounding_circle, (list, tuple)):
+        msg = "bounding_circle should be a sequence"
+        raise TypeError(msg)
+
+    if len(bounding_circle) == 3:
+        if not all(isinstance(i, (int, float)) for i in bounding_circle):
+            msg = "bounding_circle should only contain numeric data"
+            raise ValueError(msg)
+
+        *centroid, polygon_radius = cast(list[float], list(bounding_circle))
+    elif len(bounding_circle) == 2 and isinstance(bounding_circle[0], (list, tuple)):
+        if not all(
+            isinstance(i, (int, float)) for i in bounding_circle[0]
+        ) or not isinstance(bounding_circle[1], (int, float)):
+            msg = "bounding_circle should only contain numeric data"
+            raise ValueError(msg)
+
+        if len(bounding_circle[0]) != 2:
+            msg = "bounding_circle centre should contain 2D coordinates (e.g. (x, y))"
+            raise ValueError(msg)
+
+        centroid = cast(list[float], list(bounding_circle[0]))
+        polygon_radius = cast(float, bounding_circle[1])
+    else:
+        msg = (
+            "bounding_circle should contain 2D coordinates "
+            "and a radius (e.g. (x, y, r) or ((x, y), r) )"
+        )
+        raise ValueError(msg)
+
+    if polygon_radius <= 0:
+        msg = "bounding_circle radius should be > 0"
+        raise ValueError(msg)
+
+    # 1.3 Check `rotation` has an appropriate value
+    if not isinstance(rotation, (int, float)):
+        msg = "rotation should be an int or float"  # type: ignore[unreachable]
+        raise ValueError(msg)
+
+    # 2. Define Helper Functions
+    def _apply_rotation(point: list[float], degrees: float) -> tuple[float, float]:
+        return (
+            round(
+                point[0] * math.cos(math.radians(360 - degrees))
+                - point[1] * math.sin(math.radians(360 - degrees))
+                + centroid[0],
+                2,
+            ),
+            round(
+                point[1] * math.cos(math.radians(360 - degrees))
+                + point[0] * math.sin(math.radians(360 - degrees))
+                + centroid[1],
+                2,
+            ),
+        )
+
+    def _compute_polygon_vertex(angle: float) -> tuple[float, float]:
+        start_point = [polygon_radius, 0]
+        return _apply_rotation(start_point, angle)
+
+    def _get_angles(n_sides: int, rotation: float) -> list[float]:
+        angles = []
+        degrees = 360 / n_sides
+        # Start with the bottom left polygon vertex
+        current_angle = (270 - 0.5 * degrees) + rotation
+        for _ in range(0, n_sides):
+            angles.append(current_angle)
+            current_angle += degrees
+            if current_angle > 360:
+                current_angle -= 360
+        return angles
+
+    # 3. Variable Declarations
+    angles = _get_angles(n_sides, rotation)
+
+    # 4. Compute Vertices
+    return [_compute_polygon_vertex(angle) for angle in angles]
+
+
+def _color_diff(
+    color1: float | tuple[int, ...], color2: float | tuple[int, ...]
+) -> float:
+    """
+    Uses 1-norm distance to calculate difference between two values.
+    """
+    first = color1 if isinstance(color1, tuple) else (color1,)
+    second = color2 if isinstance(color2, tuple) else (color2,)
+
+    return sum(abs(first[i] - second[i]) for i in range(0, len(second)))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw2.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw2.py
new file mode 100644
index 00000000..3d68658e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageDraw2.py
@@ -0,0 +1,243 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# WCK-style drawing interface operations
+#
+# History:
+# 2003-12-07 fl   created
+# 2005-05-15 fl   updated; added to PIL as ImageDraw2
+# 2005-05-15 fl   added text support
+# 2005-05-20 fl   added arc/chord/pieslice support
+#
+# Copyright (c) 2003-2005 by Secret Labs AB
+# Copyright (c) 2003-2005 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+
+
+"""
+(Experimental) WCK-style drawing interface operations
+
+.. seealso:: :py:mod:`PIL.ImageDraw`
+"""
+from __future__ import annotations
+
+from typing import Any, AnyStr, BinaryIO
+
+from . import Image, ImageColor, ImageDraw, ImageFont, ImagePath
+from ._typing import Coords, StrOrBytesPath
+
+
+class Pen:
+    """Stores an outline color and width."""
+
+    def __init__(self, color: str, width: int = 1, opacity: int = 255) -> None:
+        self.color = ImageColor.getrgb(color)
+        self.width = width
+
+
+class Brush:
+    """Stores a fill color"""
+
+    def __init__(self, color: str, opacity: int = 255) -> None:
+        self.color = ImageColor.getrgb(color)
+
+
+class Font:
+    """Stores a TrueType font and color"""
+
+    def __init__(
+        self, color: str, file: StrOrBytesPath | BinaryIO, size: float = 12
+    ) -> None:
+        # FIXME: add support for bitmap fonts
+        self.color = ImageColor.getrgb(color)
+        self.font = ImageFont.truetype(file, size)
+
+
+class Draw:
+    """
+    (Experimental) WCK-style drawing interface
+    """
+
+    def __init__(
+        self,
+        image: Image.Image | str,
+        size: tuple[int, int] | list[int] | None = None,
+        color: float | tuple[float, ...] | str | None = None,
+    ) -> None:
+        if isinstance(image, str):
+            if size is None:
+                msg = "If image argument is mode string, size must be a list or tuple"
+                raise ValueError(msg)
+            image = Image.new(image, size, color)
+        self.draw = ImageDraw.Draw(image)
+        self.image = image
+        self.transform: tuple[float, float, float, float, float, float] | None = None
+
+    def flush(self) -> Image.Image:
+        return self.image
+
+    def render(
+        self,
+        op: str,
+        xy: Coords,
+        pen: Pen | Brush | None,
+        brush: Brush | Pen | None = None,
+        **kwargs: Any,
+    ) -> None:
+        # handle color arguments
+        outline = fill = None
+        width = 1
+        if isinstance(pen, Pen):
+            outline = pen.color
+            width = pen.width
+        elif isinstance(brush, Pen):
+            outline = brush.color
+            width = brush.width
+        if isinstance(brush, Brush):
+            fill = brush.color
+        elif isinstance(pen, Brush):
+            fill = pen.color
+        # handle transformation
+        if self.transform:
+            path = ImagePath.Path(xy)
+            path.transform(self.transform)
+            xy = path
+        # render the item
+        if op in ("arc", "line"):
+            kwargs.setdefault("fill", outline)
+        else:
+            kwargs.setdefault("fill", fill)
+            kwargs.setdefault("outline", outline)
+        if op == "line":
+            kwargs.setdefault("width", width)
+        getattr(self.draw, op)(xy, **kwargs)
+
+    def settransform(self, offset: tuple[float, float]) -> None:
+        """Sets a transformation offset."""
+        (xoffset, yoffset) = offset
+        self.transform = (1, 0, xoffset, 0, 1, yoffset)
+
+    def arc(
+        self,
+        xy: Coords,
+        pen: Pen | Brush | None,
+        start: float,
+        end: float,
+        *options: Any,
+    ) -> None:
+        """
+        Draws an arc (a portion of a circle outline) between the start and end
+        angles, inside the given bounding box.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.arc`
+        """
+        self.render("arc", xy, pen, *options, start=start, end=end)
+
+    def chord(
+        self,
+        xy: Coords,
+        pen: Pen | Brush | None,
+        start: float,
+        end: float,
+        *options: Any,
+    ) -> None:
+        """
+        Same as :py:meth:`~PIL.ImageDraw2.Draw.arc`, but connects the end points
+        with a straight line.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.chord`
+        """
+        self.render("chord", xy, pen, *options, start=start, end=end)
+
+    def ellipse(self, xy: Coords, pen: Pen | Brush | None, *options: Any) -> None:
+        """
+        Draws an ellipse inside the given bounding box.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.ellipse`
+        """
+        self.render("ellipse", xy, pen, *options)
+
+    def line(self, xy: Coords, pen: Pen | Brush | None, *options: Any) -> None:
+        """
+        Draws a line between the coordinates in the ``xy`` list.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.line`
+        """
+        self.render("line", xy, pen, *options)
+
+    def pieslice(
+        self,
+        xy: Coords,
+        pen: Pen | Brush | None,
+        start: float,
+        end: float,
+        *options: Any,
+    ) -> None:
+        """
+        Same as arc, but also draws straight lines between the end points and the
+        center of the bounding box.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.pieslice`
+        """
+        self.render("pieslice", xy, pen, *options, start=start, end=end)
+
+    def polygon(self, xy: Coords, pen: Pen | Brush | None, *options: Any) -> None:
+        """
+        Draws a polygon.
+
+        The polygon outline consists of straight lines between the given
+        coordinates, plus a straight line between the last and the first
+        coordinate.
+
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.polygon`
+        """
+        self.render("polygon", xy, pen, *options)
+
+    def rectangle(self, xy: Coords, pen: Pen | Brush | None, *options: Any) -> None:
+        """
+        Draws a rectangle.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.rectangle`
+        """
+        self.render("rectangle", xy, pen, *options)
+
+    def text(self, xy: tuple[float, float], text: AnyStr, font: Font) -> None:
+        """
+        Draws the string at the given position.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.text`
+        """
+        if self.transform:
+            path = ImagePath.Path(xy)
+            path.transform(self.transform)
+            xy = path
+        self.draw.text(xy, text, font=font.font, fill=font.color)
+
+    def textbbox(
+        self, xy: tuple[float, float], text: AnyStr, font: Font
+    ) -> tuple[float, float, float, float]:
+        """
+        Returns bounding box (in pixels) of given text.
+
+        :return: ``(left, top, right, bottom)`` bounding box
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.textbbox`
+        """
+        if self.transform:
+            path = ImagePath.Path(xy)
+            path.transform(self.transform)
+            xy = path
+        return self.draw.textbbox(xy, text, font=font.font)
+
+    def textlength(self, text: AnyStr, font: Font) -> float:
+        """
+        Returns length (in pixels) of given text.
+        This is the amount by which following text should be offset.
+
+        .. seealso:: :py:meth:`PIL.ImageDraw.ImageDraw.textlength`
+        """
+        return self.draw.textlength(text, font=font.font)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageEnhance.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageEnhance.py
new file mode 100644
index 00000000..0e7e6dd8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageEnhance.py
@@ -0,0 +1,113 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# image enhancement classes
+#
+# For a background, see "Image Processing By Interpolation and
+# Extrapolation", Paul Haeberli and Douglas Voorhies.  Available
+# at http://www.graficaobscura.com/interp/index.html
+#
+# History:
+# 1996-03-23 fl  Created
+# 2009-06-16 fl  Fixed mean calculation
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image, ImageFilter, ImageStat
+
+
+class _Enhance:
+    image: Image.Image
+    degenerate: Image.Image
+
+    def enhance(self, factor: float) -> Image.Image:
+        """
+        Returns an enhanced image.
+
+        :param factor: A floating point value controlling the enhancement.
+                       Factor 1.0 always returns a copy of the original image,
+                       lower factors mean less color (brightness, contrast,
+                       etc), and higher values more. There are no restrictions
+                       on this value.
+        :rtype: :py:class:`~PIL.Image.Image`
+        """
+        return Image.blend(self.degenerate, self.image, factor)
+
+
+class Color(_Enhance):
+    """Adjust image color balance.
+
+    This class can be used to adjust the colour balance of an image, in
+    a manner similar to the controls on a colour TV set. An enhancement
+    factor of 0.0 gives a black and white image. A factor of 1.0 gives
+    the original image.
+    """
+
+    def __init__(self, image: Image.Image) -> None:
+        self.image = image
+        self.intermediate_mode = "L"
+        if "A" in image.getbands():
+            self.intermediate_mode = "LA"
+
+        if self.intermediate_mode != image.mode:
+            image = image.convert(self.intermediate_mode).convert(image.mode)
+        self.degenerate = image
+
+
+class Contrast(_Enhance):
+    """Adjust image contrast.
+
+    This class can be used to control the contrast of an image, similar
+    to the contrast control on a TV set. An enhancement factor of 0.0
+    gives a solid gray image. A factor of 1.0 gives the original image.
+    """
+
+    def __init__(self, image: Image.Image) -> None:
+        self.image = image
+        if image.mode != "L":
+            image = image.convert("L")
+        mean = int(ImageStat.Stat(image).mean[0] + 0.5)
+        self.degenerate = Image.new("L", image.size, mean)
+        if self.degenerate.mode != self.image.mode:
+            self.degenerate = self.degenerate.convert(self.image.mode)
+
+        if "A" in self.image.getbands():
+            self.degenerate.putalpha(self.image.getchannel("A"))
+
+
+class Brightness(_Enhance):
+    """Adjust image brightness.
+
+    This class can be used to control the brightness of an image.  An
+    enhancement factor of 0.0 gives a black image. A factor of 1.0 gives the
+    original image.
+    """
+
+    def __init__(self, image: Image.Image) -> None:
+        self.image = image
+        self.degenerate = Image.new(image.mode, image.size, 0)
+
+        if "A" in image.getbands():
+            self.degenerate.putalpha(image.getchannel("A"))
+
+
+class Sharpness(_Enhance):
+    """Adjust image sharpness.
+
+    This class can be used to adjust the sharpness of an image. An
+    enhancement factor of 0.0 gives a blurred image, a factor of 1.0 gives the
+    original image, and a factor of 2.0 gives a sharpened image.
+    """
+
+    def __init__(self, image: Image.Image) -> None:
+        self.image = image
+        self.degenerate = image.filter(ImageFilter.SMOOTH)
+
+        if "A" in image.getbands():
+            self.degenerate.putalpha(image.getchannel("A"))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageFile.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageFile.py
new file mode 100644
index 00000000..5d0f87a9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageFile.py
@@ -0,0 +1,832 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# base class for image file handlers
+#
+# history:
+# 1995-09-09 fl   Created
+# 1996-03-11 fl   Fixed load mechanism.
+# 1996-04-15 fl   Added pcx/xbm decoders.
+# 1996-04-30 fl   Added encoders.
+# 1996-12-14 fl   Added load helpers
+# 1997-01-11 fl   Use encode_to_file where possible
+# 1997-08-27 fl   Flush output in _save
+# 1998-03-05 fl   Use memory mapping for some modes
+# 1999-02-04 fl   Use memory mapping also for "I;16" and "I;16B"
+# 1999-05-31 fl   Added image parser
+# 2000-10-12 fl   Set readonly flag on memory-mapped images
+# 2002-03-20 fl   Use better messages for common decoder errors
+# 2003-04-21 fl   Fall back on mmap/map_buffer if map is not available
+# 2003-10-30 fl   Added StubImageFile class
+# 2004-02-25 fl   Made incremental parser more robust
+#
+# Copyright (c) 1997-2004 by Secret Labs AB
+# Copyright (c) 1995-2004 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import abc
+import io
+import itertools
+import os
+import struct
+import sys
+from typing import IO, TYPE_CHECKING, Any, NamedTuple, cast
+
+from . import Image
+from ._deprecate import deprecate
+from ._util import is_path
+
+if TYPE_CHECKING:
+    from ._typing import StrOrBytesPath
+
+MAXBLOCK = 65536
+
+SAFEBLOCK = 1024 * 1024
+
+LOAD_TRUNCATED_IMAGES = False
+"""Whether or not to load truncated image files. User code may change this."""
+
+ERRORS = {
+    -1: "image buffer overrun error",
+    -2: "decoding error",
+    -3: "unknown error",
+    -8: "bad configuration",
+    -9: "out of memory error",
+}
+"""
+Dict of known error codes returned from :meth:`.PyDecoder.decode`,
+:meth:`.PyEncoder.encode` :meth:`.PyEncoder.encode_to_pyfd` and
+:meth:`.PyEncoder.encode_to_file`.
+"""
+
+
+#
+# --------------------------------------------------------------------
+# Helpers
+
+
+def _get_oserror(error: int, *, encoder: bool) -> OSError:
+    try:
+        msg = Image.core.getcodecstatus(error)
+    except AttributeError:
+        msg = ERRORS.get(error)
+    if not msg:
+        msg = f"{'encoder' if encoder else 'decoder'} error {error}"
+    msg += f" when {'writing' if encoder else 'reading'} image file"
+    return OSError(msg)
+
+
+def raise_oserror(error: int) -> OSError:
+    deprecate(
+        "raise_oserror",
+        12,
+        action="It is only useful for translating error codes returned by a codec's "
+        "decode() method, which ImageFile already does automatically.",
+    )
+    raise _get_oserror(error, encoder=False)
+
+
+def _tilesort(t: _Tile) -> int:
+    # sort on offset
+    return t[2]
+
+
+class _Tile(NamedTuple):
+    codec_name: str
+    extents: tuple[int, int, int, int] | None
+    offset: int = 0
+    args: tuple[Any, ...] | str | None = None
+
+
+#
+# --------------------------------------------------------------------
+# ImageFile base class
+
+
+class ImageFile(Image.Image):
+    """Base class for image file format handlers."""
+
+    def __init__(
+        self, fp: StrOrBytesPath | IO[bytes], filename: str | bytes | None = None
+    ) -> None:
+        super().__init__()
+
+        self._min_frame = 0
+
+        self.custom_mimetype: str | None = None
+
+        self.tile: list[_Tile] = []
+        """ A list of tile descriptors """
+
+        self.readonly = 1  # until we know better
+
+        self.decoderconfig: tuple[Any, ...] = ()
+        self.decodermaxblock = MAXBLOCK
+
+        if is_path(fp):
+            # filename
+            self.fp = open(fp, "rb")
+            self.filename = os.fspath(fp)
+            self._exclusive_fp = True
+        else:
+            # stream
+            self.fp = cast(IO[bytes], fp)
+            self.filename = filename if filename is not None else ""
+            # can be overridden
+            self._exclusive_fp = False
+
+        try:
+            try:
+                self._open()
+            except (
+                IndexError,  # end of data
+                TypeError,  # end of data (ord)
+                KeyError,  # unsupported mode
+                EOFError,  # got header but not the first frame
+                struct.error,
+            ) as v:
+                raise SyntaxError(v) from v
+
+            if not self.mode or self.size[0] <= 0 or self.size[1] <= 0:
+                msg = "not identified by this driver"
+                raise SyntaxError(msg)
+        except BaseException:
+            # close the file only if we have opened it this constructor
+            if self._exclusive_fp:
+                self.fp.close()
+            raise
+
+    def _open(self) -> None:
+        pass
+
+    def get_format_mimetype(self) -> str | None:
+        if self.custom_mimetype:
+            return self.custom_mimetype
+        if self.format is not None:
+            return Image.MIME.get(self.format.upper())
+        return None
+
+    def __setstate__(self, state: list[Any]) -> None:
+        self.tile = []
+        super().__setstate__(state)
+
+    def verify(self) -> None:
+        """Check file integrity"""
+
+        # raise exception if something's wrong.  must be called
+        # directly after open, and closes file when finished.
+        if self._exclusive_fp:
+            self.fp.close()
+        self.fp = None
+
+    def load(self) -> Image.core.PixelAccess | None:
+        """Load image data based on tile list"""
+
+        if not self.tile and self._im is None:
+            msg = "cannot load this image"
+            raise OSError(msg)
+
+        pixel = Image.Image.load(self)
+        if not self.tile:
+            return pixel
+
+        self.map: mmap.mmap | None = None
+        use_mmap = self.filename and len(self.tile) == 1
+        # As of pypy 2.1.0, memory mapping was failing here.
+        use_mmap = use_mmap and not hasattr(sys, "pypy_version_info")
+
+        readonly = 0
+
+        # look for read/seek overrides
+        if hasattr(self, "load_read"):
+            read = self.load_read
+            # don't use mmap if there are custom read/seek functions
+            use_mmap = False
+        else:
+            read = self.fp.read
+
+        if hasattr(self, "load_seek"):
+            seek = self.load_seek
+            use_mmap = False
+        else:
+            seek = self.fp.seek
+
+        if use_mmap:
+            # try memory mapping
+            decoder_name, extents, offset, args = self.tile[0]
+            if isinstance(args, str):
+                args = (args, 0, 1)
+            if (
+                decoder_name == "raw"
+                and isinstance(args, tuple)
+                and len(args) >= 3
+                and args[0] == self.mode
+                and args[0] in Image._MAPMODES
+            ):
+                try:
+                    # use mmap, if possible
+                    import mmap
+
+                    with open(self.filename) as fp:
+                        self.map = mmap.mmap(fp.fileno(), 0, access=mmap.ACCESS_READ)
+                    if offset + self.size[1] * args[1] > self.map.size():
+                        msg = "buffer is not large enough"
+                        raise OSError(msg)
+                    self.im = Image.core.map_buffer(
+                        self.map, self.size, decoder_name, offset, args
+                    )
+                    readonly = 1
+                    # After trashing self.im,
+                    # we might need to reload the palette data.
+                    if self.palette:
+                        self.palette.dirty = 1
+                except (AttributeError, OSError, ImportError):
+                    self.map = None
+
+        self.load_prepare()
+        err_code = -3  # initialize to unknown error
+        if not self.map:
+            # sort tiles in file order
+            self.tile.sort(key=_tilesort)
+
+            # FIXME: This is a hack to handle TIFF's JpegTables tag.
+            prefix = getattr(self, "tile_prefix", b"")
+
+            # Remove consecutive duplicates that only differ by their offset
+            self.tile = [
+                list(tiles)[-1]
+                for _, tiles in itertools.groupby(
+                    self.tile, lambda tile: (tile[0], tile[1], tile[3])
+                )
+            ]
+            for decoder_name, extents, offset, args in self.tile:
+                seek(offset)
+                decoder = Image._getdecoder(
+                    self.mode, decoder_name, args, self.decoderconfig
+                )
+                try:
+                    decoder.setimage(self.im, extents)
+                    if decoder.pulls_fd:
+                        decoder.setfd(self.fp)
+                        err_code = decoder.decode(b"")[1]
+                    else:
+                        b = prefix
+                        while True:
+                            try:
+                                s = read(self.decodermaxblock)
+                            except (IndexError, struct.error) as e:
+                                # truncated png/gif
+                                if LOAD_TRUNCATED_IMAGES:
+                                    break
+                                else:
+                                    msg = "image file is truncated"
+                                    raise OSError(msg) from e
+
+                            if not s:  # truncated jpeg
+                                if LOAD_TRUNCATED_IMAGES:
+                                    break
+                                else:
+                                    msg = (
+                                        "image file is truncated "
+                                        f"({len(b)} bytes not processed)"
+                                    )
+                                    raise OSError(msg)
+
+                            b = b + s
+                            n, err_code = decoder.decode(b)
+                            if n < 0:
+                                break
+                            b = b[n:]
+                finally:
+                    # Need to cleanup here to prevent leaks
+                    decoder.cleanup()
+
+        self.tile = []
+        self.readonly = readonly
+
+        self.load_end()
+
+        if self._exclusive_fp and self._close_exclusive_fp_after_loading:
+            self.fp.close()
+        self.fp = None
+
+        if not self.map and not LOAD_TRUNCATED_IMAGES and err_code < 0:
+            # still raised if decoder fails to return anything
+            raise _get_oserror(err_code, encoder=False)
+
+        return Image.Image.load(self)
+
+    def load_prepare(self) -> None:
+        # create image memory if necessary
+        if self._im is None:
+            self.im = Image.core.new(self.mode, self.size)
+        # create palette (optional)
+        if self.mode == "P":
+            Image.Image.load(self)
+
+    def load_end(self) -> None:
+        # may be overridden
+        pass
+
+    # may be defined for contained formats
+    # def load_seek(self, pos: int) -> None:
+    #     pass
+
+    # may be defined for blocked formats (e.g. PNG)
+    # def load_read(self, read_bytes: int) -> bytes:
+    #     pass
+
+    def _seek_check(self, frame: int) -> bool:
+        if (
+            frame < self._min_frame
+            # Only check upper limit on frames if additional seek operations
+            # are not required to do so
+            or (
+                not (hasattr(self, "_n_frames") and self._n_frames is None)
+                and frame >= getattr(self, "n_frames") + self._min_frame
+            )
+        ):
+            msg = "attempt to seek outside sequence"
+            raise EOFError(msg)
+
+        return self.tell() != frame
+
+
+class StubHandler:
+    def open(self, im: StubImageFile) -> None:
+        pass
+
+    @abc.abstractmethod
+    def load(self, im: StubImageFile) -> Image.Image:
+        pass
+
+
+class StubImageFile(ImageFile):
+    """
+    Base class for stub image loaders.
+
+    A stub loader is an image loader that can identify files of a
+    certain format, but relies on external code to load the file.
+    """
+
+    def _open(self) -> None:
+        msg = "StubImageFile subclass must implement _open"
+        raise NotImplementedError(msg)
+
+    def load(self) -> Image.core.PixelAccess | None:
+        loader = self._load()
+        if loader is None:
+            msg = f"cannot find loader for this {self.format} file"
+            raise OSError(msg)
+        image = loader.load(self)
+        assert image is not None
+        # become the other object (!)
+        self.__class__ = image.__class__  # type: ignore[assignment]
+        self.__dict__ = image.__dict__
+        return image.load()
+
+    def _load(self) -> StubHandler | None:
+        """(Hook) Find actual image loader."""
+        msg = "StubImageFile subclass must implement _load"
+        raise NotImplementedError(msg)
+
+
+class Parser:
+    """
+    Incremental image parser.  This class implements the standard
+    feed/close consumer interface.
+    """
+
+    incremental = None
+    image: Image.Image | None = None
+    data: bytes | None = None
+    decoder: Image.core.ImagingDecoder | PyDecoder | None = None
+    offset = 0
+    finished = 0
+
+    def reset(self) -> None:
+        """
+        (Consumer) Reset the parser.  Note that you can only call this
+        method immediately after you've created a parser; parser
+        instances cannot be reused.
+        """
+        assert self.data is None, "cannot reuse parsers"
+
+    def feed(self, data: bytes) -> None:
+        """
+        (Consumer) Feed data to the parser.
+
+        :param data: A string buffer.
+        :exception OSError: If the parser failed to parse the image file.
+        """
+        # collect data
+
+        if self.finished:
+            return
+
+        if self.data is None:
+            self.data = data
+        else:
+            self.data = self.data + data
+
+        # parse what we have
+        if self.decoder:
+            if self.offset > 0:
+                # skip header
+                skip = min(len(self.data), self.offset)
+                self.data = self.data[skip:]
+                self.offset = self.offset - skip
+                if self.offset > 0 or not self.data:
+                    return
+
+            n, e = self.decoder.decode(self.data)
+
+            if n < 0:
+                # end of stream
+                self.data = None
+                self.finished = 1
+                if e < 0:
+                    # decoding error
+                    self.image = None
+                    raise _get_oserror(e, encoder=False)
+                else:
+                    # end of image
+                    return
+            self.data = self.data[n:]
+
+        elif self.image:
+            # if we end up here with no decoder, this file cannot
+            # be incrementally parsed.  wait until we've gotten all
+            # available data
+            pass
+
+        else:
+            # attempt to open this file
+            try:
+                with io.BytesIO(self.data) as fp:
+                    im = Image.open(fp)
+            except OSError:
+                pass  # not enough data
+            else:
+                flag = hasattr(im, "load_seek") or hasattr(im, "load_read")
+                if flag or len(im.tile) != 1:
+                    # custom load code, or multiple tiles
+                    self.decode = None
+                else:
+                    # initialize decoder
+                    im.load_prepare()
+                    d, e, o, a = im.tile[0]
+                    im.tile = []
+                    self.decoder = Image._getdecoder(im.mode, d, a, im.decoderconfig)
+                    self.decoder.setimage(im.im, e)
+
+                    # calculate decoder offset
+                    self.offset = o
+                    if self.offset <= len(self.data):
+                        self.data = self.data[self.offset :]
+                        self.offset = 0
+
+                self.image = im
+
+    def __enter__(self) -> Parser:
+        return self
+
+    def __exit__(self, *args: object) -> None:
+        self.close()
+
+    def close(self) -> Image.Image:
+        """
+        (Consumer) Close the stream.
+
+        :returns: An image object.
+        :exception OSError: If the parser failed to parse the image file either
+                            because it cannot be identified or cannot be
+                            decoded.
+        """
+        # finish decoding
+        if self.decoder:
+            # get rid of what's left in the buffers
+            self.feed(b"")
+            self.data = self.decoder = None
+            if not self.finished:
+                msg = "image was incomplete"
+                raise OSError(msg)
+        if not self.image:
+            msg = "cannot parse this image"
+            raise OSError(msg)
+        if self.data:
+            # incremental parsing not possible; reopen the file
+            # not that we have all data
+            with io.BytesIO(self.data) as fp:
+                try:
+                    self.image = Image.open(fp)
+                finally:
+                    self.image.load()
+        return self.image
+
+
+# --------------------------------------------------------------------
+
+
+def _save(im: Image.Image, fp: IO[bytes], tile: list[_Tile], bufsize: int = 0) -> None:
+    """Helper to save image based on tile list
+
+    :param im: Image object.
+    :param fp: File object.
+    :param tile: Tile list.
+    :param bufsize: Optional buffer size
+    """
+
+    im.load()
+    if not hasattr(im, "encoderconfig"):
+        im.encoderconfig = ()
+    tile.sort(key=_tilesort)
+    # FIXME: make MAXBLOCK a configuration parameter
+    # It would be great if we could have the encoder specify what it needs
+    # But, it would need at least the image size in most cases. RawEncode is
+    # a tricky case.
+    bufsize = max(MAXBLOCK, bufsize, im.size[0] * 4)  # see RawEncode.c
+    try:
+        fh = fp.fileno()
+        fp.flush()
+        _encode_tile(im, fp, tile, bufsize, fh)
+    except (AttributeError, io.UnsupportedOperation) as exc:
+        _encode_tile(im, fp, tile, bufsize, None, exc)
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+def _encode_tile(
+    im: Image.Image,
+    fp: IO[bytes],
+    tile: list[_Tile],
+    bufsize: int,
+    fh: int | None,
+    exc: BaseException | None = None,
+) -> None:
+    for encoder_name, extents, offset, args in tile:
+        if offset > 0:
+            fp.seek(offset)
+        encoder = Image._getencoder(im.mode, encoder_name, args, im.encoderconfig)
+        try:
+            encoder.setimage(im.im, extents)
+            if encoder.pushes_fd:
+                encoder.setfd(fp)
+                errcode = encoder.encode_to_pyfd()[1]
+            else:
+                if exc:
+                    # compress to Python file-compatible object
+                    while True:
+                        errcode, data = encoder.encode(bufsize)[1:]
+                        fp.write(data)
+                        if errcode:
+                            break
+                else:
+                    # slight speedup: compress to real file object
+                    assert fh is not None
+                    errcode = encoder.encode_to_file(fh, bufsize)
+            if errcode < 0:
+                raise _get_oserror(errcode, encoder=True) from exc
+        finally:
+            encoder.cleanup()
+
+
+def _safe_read(fp: IO[bytes], size: int) -> bytes:
+    """
+    Reads large blocks in a safe way.  Unlike fp.read(n), this function
+    doesn't trust the user.  If the requested size is larger than
+    SAFEBLOCK, the file is read block by block.
+
+    :param fp: File handle.  Must implement a <b>read</b> method.
+    :param size: Number of bytes to read.
+    :returns: A string containing <i>size</i> bytes of data.
+
+    Raises an OSError if the file is truncated and the read cannot be completed
+
+    """
+    if size <= 0:
+        return b""
+    if size <= SAFEBLOCK:
+        data = fp.read(size)
+        if len(data) < size:
+            msg = "Truncated File Read"
+            raise OSError(msg)
+        return data
+    blocks: list[bytes] = []
+    remaining_size = size
+    while remaining_size > 0:
+        block = fp.read(min(remaining_size, SAFEBLOCK))
+        if not block:
+            break
+        blocks.append(block)
+        remaining_size -= len(block)
+    if sum(len(block) for block in blocks) < size:
+        msg = "Truncated File Read"
+        raise OSError(msg)
+    return b"".join(blocks)
+
+
+class PyCodecState:
+    def __init__(self) -> None:
+        self.xsize = 0
+        self.ysize = 0
+        self.xoff = 0
+        self.yoff = 0
+
+    def extents(self) -> tuple[int, int, int, int]:
+        return self.xoff, self.yoff, self.xoff + self.xsize, self.yoff + self.ysize
+
+
+class PyCodec:
+    fd: IO[bytes] | None
+
+    def __init__(self, mode: str, *args: Any) -> None:
+        self.im: Image.core.ImagingCore | None = None
+        self.state = PyCodecState()
+        self.fd = None
+        self.mode = mode
+        self.init(args)
+
+    def init(self, args: tuple[Any, ...]) -> None:
+        """
+        Override to perform codec specific initialization
+
+        :param args: Tuple of arg items from the tile entry
+        :returns: None
+        """
+        self.args = args
+
+    def cleanup(self) -> None:
+        """
+        Override to perform codec specific cleanup
+
+        :returns: None
+        """
+        pass
+
+    def setfd(self, fd: IO[bytes]) -> None:
+        """
+        Called from ImageFile to set the Python file-like object
+
+        :param fd: A Python file-like object
+        :returns: None
+        """
+        self.fd = fd
+
+    def setimage(
+        self,
+        im: Image.core.ImagingCore,
+        extents: tuple[int, int, int, int] | None = None,
+    ) -> None:
+        """
+        Called from ImageFile to set the core output image for the codec
+
+        :param im: A core image object
+        :param extents: a 4 tuple of (x0, y0, x1, y1) defining the rectangle
+            for this tile
+        :returns: None
+        """
+
+        # following c code
+        self.im = im
+
+        if extents:
+            (x0, y0, x1, y1) = extents
+        else:
+            (x0, y0, x1, y1) = (0, 0, 0, 0)
+
+        if x0 == 0 and x1 == 0:
+            self.state.xsize, self.state.ysize = self.im.size
+        else:
+            self.state.xoff = x0
+            self.state.yoff = y0
+            self.state.xsize = x1 - x0
+            self.state.ysize = y1 - y0
+
+        if self.state.xsize <= 0 or self.state.ysize <= 0:
+            msg = "Size cannot be negative"
+            raise ValueError(msg)
+
+        if (
+            self.state.xsize + self.state.xoff > self.im.size[0]
+            or self.state.ysize + self.state.yoff > self.im.size[1]
+        ):
+            msg = "Tile cannot extend outside image"
+            raise ValueError(msg)
+
+
+class PyDecoder(PyCodec):
+    """
+    Python implementation of a format decoder. Override this class and
+    add the decoding logic in the :meth:`decode` method.
+
+    See :ref:`Writing Your Own File Codec in Python<file-codecs-py>`
+    """
+
+    _pulls_fd = False
+
+    @property
+    def pulls_fd(self) -> bool:
+        return self._pulls_fd
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        """
+        Override to perform the decoding process.
+
+        :param buffer: A bytes object with the data to be decoded.
+        :returns: A tuple of ``(bytes consumed, errcode)``.
+            If finished with decoding return -1 for the bytes consumed.
+            Err codes are from :data:`.ImageFile.ERRORS`.
+        """
+        msg = "unavailable in base decoder"
+        raise NotImplementedError(msg)
+
+    def set_as_raw(
+        self, data: bytes, rawmode: str | None = None, extra: tuple[Any, ...] = ()
+    ) -> None:
+        """
+        Convenience method to set the internal image from a stream of raw data
+
+        :param data: Bytes to be set
+        :param rawmode: The rawmode to be used for the decoder.
+            If not specified, it will default to the mode of the image
+        :param extra: Extra arguments for the decoder.
+        :returns: None
+        """
+
+        if not rawmode:
+            rawmode = self.mode
+        d = Image._getdecoder(self.mode, "raw", rawmode, extra)
+        assert self.im is not None
+        d.setimage(self.im, self.state.extents())
+        s = d.decode(data)
+
+        if s[0] >= 0:
+            msg = "not enough image data"
+            raise ValueError(msg)
+        if s[1] != 0:
+            msg = "cannot decode image data"
+            raise ValueError(msg)
+
+
+class PyEncoder(PyCodec):
+    """
+    Python implementation of a format encoder. Override this class and
+    add the decoding logic in the :meth:`encode` method.
+
+    See :ref:`Writing Your Own File Codec in Python<file-codecs-py>`
+    """
+
+    _pushes_fd = False
+
+    @property
+    def pushes_fd(self) -> bool:
+        return self._pushes_fd
+
+    def encode(self, bufsize: int) -> tuple[int, int, bytes]:
+        """
+        Override to perform the encoding process.
+
+        :param bufsize: Buffer size.
+        :returns: A tuple of ``(bytes encoded, errcode, bytes)``.
+            If finished with encoding return 1 for the error code.
+            Err codes are from :data:`.ImageFile.ERRORS`.
+        """
+        msg = "unavailable in base encoder"
+        raise NotImplementedError(msg)
+
+    def encode_to_pyfd(self) -> tuple[int, int]:
+        """
+        If ``pushes_fd`` is ``True``, then this method will be used,
+        and ``encode()`` will only be called once.
+
+        :returns: A tuple of ``(bytes consumed, errcode)``.
+            Err codes are from :data:`.ImageFile.ERRORS`.
+        """
+        if not self.pushes_fd:
+            return 0, -8  # bad configuration
+        bytes_consumed, errcode, data = self.encode(0)
+        if data:
+            assert self.fd is not None
+            self.fd.write(data)
+        return bytes_consumed, errcode
+
+    def encode_to_file(self, fh: int, bufsize: int) -> int:
+        """
+        :param fh: File handle.
+        :param bufsize: Buffer size.
+
+        :returns: If finished successfully, return 0.
+            Otherwise, return an error code. Err codes are from
+            :data:`.ImageFile.ERRORS`.
+        """
+        errcode = 0
+        while errcode == 0:
+            status, errcode, buf = self.encode(bufsize)
+            if status > 0:
+                os.write(fh, buf[status:])
+        return errcode
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageFilter.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageFilter.py
new file mode 100644
index 00000000..b350e56f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageFilter.py
@@ -0,0 +1,605 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# standard filters
+#
+# History:
+# 1995-11-27 fl   Created
+# 2002-06-08 fl   Added rank and mode filters
+# 2003-09-15 fl   Fixed rank calculation in rank filter; added expand call
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1995-2002 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import abc
+import functools
+from collections.abc import Sequence
+from types import ModuleType
+from typing import TYPE_CHECKING, Any, Callable, cast
+
+if TYPE_CHECKING:
+    from . import _imaging
+    from ._typing import NumpyArray
+
+
+class Filter:
+    @abc.abstractmethod
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        pass
+
+
+class MultibandFilter(Filter):
+    pass
+
+
+class BuiltinFilter(MultibandFilter):
+    filterargs: tuple[Any, ...]
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        if image.mode == "P":
+            msg = "cannot filter palette images"
+            raise ValueError(msg)
+        return image.filter(*self.filterargs)
+
+
+class Kernel(BuiltinFilter):
+    """
+    Create a convolution kernel. This only supports 3x3 and 5x5 integer and floating
+    point kernels.
+
+    Kernels can only be applied to "L" and "RGB" images.
+
+    :param size: Kernel size, given as (width, height). This must be (3,3) or (5,5).
+    :param kernel: A sequence containing kernel weights. The kernel will be flipped
+                   vertically before being applied to the image.
+    :param scale: Scale factor. If given, the result for each pixel is divided by this
+                  value. The default is the sum of the kernel weights.
+    :param offset: Offset. If given, this value is added to the result, after it has
+                   been divided by the scale factor.
+    """
+
+    name = "Kernel"
+
+    def __init__(
+        self,
+        size: tuple[int, int],
+        kernel: Sequence[float],
+        scale: float | None = None,
+        offset: float = 0,
+    ) -> None:
+        if scale is None:
+            # default scale is sum of kernel
+            scale = functools.reduce(lambda a, b: a + b, kernel)
+        if size[0] * size[1] != len(kernel):
+            msg = "not enough coefficients in kernel"
+            raise ValueError(msg)
+        self.filterargs = size, scale, offset, kernel
+
+
+class RankFilter(Filter):
+    """
+    Create a rank filter.  The rank filter sorts all pixels in
+    a window of the given size, and returns the ``rank``'th value.
+
+    :param size: The kernel size, in pixels.
+    :param rank: What pixel value to pick.  Use 0 for a min filter,
+                 ``size * size / 2`` for a median filter, ``size * size - 1``
+                 for a max filter, etc.
+    """
+
+    name = "Rank"
+
+    def __init__(self, size: int, rank: int) -> None:
+        self.size = size
+        self.rank = rank
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        if image.mode == "P":
+            msg = "cannot filter palette images"
+            raise ValueError(msg)
+        image = image.expand(self.size // 2, self.size // 2)
+        return image.rankfilter(self.size, self.rank)
+
+
+class MedianFilter(RankFilter):
+    """
+    Create a median filter. Picks the median pixel value in a window with the
+    given size.
+
+    :param size: The kernel size, in pixels.
+    """
+
+    name = "Median"
+
+    def __init__(self, size: int = 3) -> None:
+        self.size = size
+        self.rank = size * size // 2
+
+
+class MinFilter(RankFilter):
+    """
+    Create a min filter.  Picks the lowest pixel value in a window with the
+    given size.
+
+    :param size: The kernel size, in pixels.
+    """
+
+    name = "Min"
+
+    def __init__(self, size: int = 3) -> None:
+        self.size = size
+        self.rank = 0
+
+
+class MaxFilter(RankFilter):
+    """
+    Create a max filter.  Picks the largest pixel value in a window with the
+    given size.
+
+    :param size: The kernel size, in pixels.
+    """
+
+    name = "Max"
+
+    def __init__(self, size: int = 3) -> None:
+        self.size = size
+        self.rank = size * size - 1
+
+
+class ModeFilter(Filter):
+    """
+    Create a mode filter. Picks the most frequent pixel value in a box with the
+    given size.  Pixel values that occur only once or twice are ignored; if no
+    pixel value occurs more than twice, the original pixel value is preserved.
+
+    :param size: The kernel size, in pixels.
+    """
+
+    name = "Mode"
+
+    def __init__(self, size: int = 3) -> None:
+        self.size = size
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        return image.modefilter(self.size)
+
+
+class GaussianBlur(MultibandFilter):
+    """Blurs the image with a sequence of extended box filters, which
+    approximates a Gaussian kernel. For details on accuracy see
+    <https://www.mia.uni-saarland.de/Publications/gwosdek-ssvm11.pdf>
+
+    :param radius: Standard deviation of the Gaussian kernel. Either a sequence of two
+                   numbers for x and y, or a single number for both.
+    """
+
+    name = "GaussianBlur"
+
+    def __init__(self, radius: float | Sequence[float] = 2) -> None:
+        self.radius = radius
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        xy = self.radius
+        if isinstance(xy, (int, float)):
+            xy = (xy, xy)
+        if xy == (0, 0):
+            return image.copy()
+        return image.gaussian_blur(xy)
+
+
+class BoxBlur(MultibandFilter):
+    """Blurs the image by setting each pixel to the average value of the pixels
+    in a square box extending radius pixels in each direction.
+    Supports float radius of arbitrary size. Uses an optimized implementation
+    which runs in linear time relative to the size of the image
+    for any radius value.
+
+    :param radius: Size of the box in a direction. Either a sequence of two numbers for
+                   x and y, or a single number for both.
+
+                   Radius 0 does not blur, returns an identical image.
+                   Radius 1 takes 1 pixel in each direction, i.e. 9 pixels in total.
+    """
+
+    name = "BoxBlur"
+
+    def __init__(self, radius: float | Sequence[float]) -> None:
+        xy = radius if isinstance(radius, (tuple, list)) else (radius, radius)
+        if xy[0] < 0 or xy[1] < 0:
+            msg = "radius must be >= 0"
+            raise ValueError(msg)
+        self.radius = radius
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        xy = self.radius
+        if isinstance(xy, (int, float)):
+            xy = (xy, xy)
+        if xy == (0, 0):
+            return image.copy()
+        return image.box_blur(xy)
+
+
+class UnsharpMask(MultibandFilter):
+    """Unsharp mask filter.
+
+    See Wikipedia's entry on `digital unsharp masking`_ for an explanation of
+    the parameters.
+
+    :param radius: Blur Radius
+    :param percent: Unsharp strength, in percent
+    :param threshold: Threshold controls the minimum brightness change that
+      will be sharpened
+
+    .. _digital unsharp masking: https://en.wikipedia.org/wiki/Unsharp_masking#Digital_unsharp_masking
+
+    """
+
+    name = "UnsharpMask"
+
+    def __init__(
+        self, radius: float = 2, percent: int = 150, threshold: int = 3
+    ) -> None:
+        self.radius = radius
+        self.percent = percent
+        self.threshold = threshold
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        return image.unsharp_mask(self.radius, self.percent, self.threshold)
+
+
+class BLUR(BuiltinFilter):
+    name = "Blur"
+    # fmt: off
+    filterargs = (5, 5), 16, 0, (
+        1, 1, 1, 1, 1,
+        1, 0, 0, 0, 1,
+        1, 0, 0, 0, 1,
+        1, 0, 0, 0, 1,
+        1, 1, 1, 1, 1,
+    )
+    # fmt: on
+
+
+class CONTOUR(BuiltinFilter):
+    name = "Contour"
+    # fmt: off
+    filterargs = (3, 3), 1, 255, (
+        -1, -1, -1,
+        -1,  8, -1,
+        -1, -1, -1,
+    )
+    # fmt: on
+
+
+class DETAIL(BuiltinFilter):
+    name = "Detail"
+    # fmt: off
+    filterargs = (3, 3), 6, 0, (
+        0,  -1,  0,
+        -1, 10, -1,
+        0,  -1,  0,
+    )
+    # fmt: on
+
+
+class EDGE_ENHANCE(BuiltinFilter):
+    name = "Edge-enhance"
+    # fmt: off
+    filterargs = (3, 3), 2, 0, (
+        -1, -1, -1,
+        -1, 10, -1,
+        -1, -1, -1,
+    )
+    # fmt: on
+
+
+class EDGE_ENHANCE_MORE(BuiltinFilter):
+    name = "Edge-enhance More"
+    # fmt: off
+    filterargs = (3, 3), 1, 0, (
+        -1, -1, -1,
+        -1,  9, -1,
+        -1, -1, -1,
+    )
+    # fmt: on
+
+
+class EMBOSS(BuiltinFilter):
+    name = "Emboss"
+    # fmt: off
+    filterargs = (3, 3), 1, 128, (
+        -1, 0, 0,
+        0,  1, 0,
+        0,  0, 0,
+    )
+    # fmt: on
+
+
+class FIND_EDGES(BuiltinFilter):
+    name = "Find Edges"
+    # fmt: off
+    filterargs = (3, 3), 1, 0, (
+        -1, -1, -1,
+        -1,  8, -1,
+        -1, -1, -1,
+    )
+    # fmt: on
+
+
+class SHARPEN(BuiltinFilter):
+    name = "Sharpen"
+    # fmt: off
+    filterargs = (3, 3), 16, 0, (
+        -2, -2, -2,
+        -2, 32, -2,
+        -2, -2, -2,
+    )
+    # fmt: on
+
+
+class SMOOTH(BuiltinFilter):
+    name = "Smooth"
+    # fmt: off
+    filterargs = (3, 3), 13, 0, (
+        1, 1, 1,
+        1, 5, 1,
+        1, 1, 1,
+    )
+    # fmt: on
+
+
+class SMOOTH_MORE(BuiltinFilter):
+    name = "Smooth More"
+    # fmt: off
+    filterargs = (5, 5), 100, 0, (
+        1, 1,  1, 1, 1,
+        1, 5,  5, 5, 1,
+        1, 5, 44, 5, 1,
+        1, 5,  5, 5, 1,
+        1, 1,  1, 1, 1,
+    )
+    # fmt: on
+
+
+class Color3DLUT(MultibandFilter):
+    """Three-dimensional color lookup table.
+
+    Transforms 3-channel pixels using the values of the channels as coordinates
+    in the 3D lookup table and interpolating the nearest elements.
+
+    This method allows you to apply almost any color transformation
+    in constant time by using pre-calculated decimated tables.
+
+    .. versionadded:: 5.2.0
+
+    :param size: Size of the table. One int or tuple of (int, int, int).
+                 Minimal size in any dimension is 2, maximum is 65.
+    :param table: Flat lookup table. A list of ``channels * size**3``
+                  float elements or a list of ``size**3`` channels-sized
+                  tuples with floats. Channels are changed first,
+                  then first dimension, then second, then third.
+                  Value 0.0 corresponds lowest value of output, 1.0 highest.
+    :param channels: Number of channels in the table. Could be 3 or 4.
+                     Default is 3.
+    :param target_mode: A mode for the result image. Should have not less
+                        than ``channels`` channels. Default is ``None``,
+                        which means that mode wouldn't be changed.
+    """
+
+    name = "Color 3D LUT"
+
+    def __init__(
+        self,
+        size: int | tuple[int, int, int],
+        table: Sequence[float] | Sequence[Sequence[int]] | NumpyArray,
+        channels: int = 3,
+        target_mode: str | None = None,
+        **kwargs: bool,
+    ) -> None:
+        if channels not in (3, 4):
+            msg = "Only 3 or 4 output channels are supported"
+            raise ValueError(msg)
+        self.size = size = self._check_size(size)
+        self.channels = channels
+        self.mode = target_mode
+
+        # Hidden flag `_copy_table=False` could be used to avoid extra copying
+        # of the table if the table is specially made for the constructor.
+        copy_table = kwargs.get("_copy_table", True)
+        items = size[0] * size[1] * size[2]
+        wrong_size = False
+
+        numpy: ModuleType | None = None
+        if hasattr(table, "shape"):
+            try:
+                import numpy
+            except ImportError:
+                pass
+
+        if numpy and isinstance(table, numpy.ndarray):
+            numpy_table: NumpyArray = table
+            if copy_table:
+                numpy_table = numpy_table.copy()
+
+            if numpy_table.shape in [
+                (items * channels,),
+                (items, channels),
+                (size[2], size[1], size[0], channels),
+            ]:
+                table = numpy_table.reshape(items * channels)
+            else:
+                wrong_size = True
+
+        else:
+            if copy_table:
+                table = list(table)
+
+            # Convert to a flat list
+            if table and isinstance(table[0], (list, tuple)):
+                raw_table = cast(Sequence[Sequence[int]], table)
+                flat_table: list[int] = []
+                for pixel in raw_table:
+                    if len(pixel) != channels:
+                        msg = (
+                            "The elements of the table should "
+                            f"have a length of {channels}."
+                        )
+                        raise ValueError(msg)
+                    flat_table.extend(pixel)
+                table = flat_table
+
+        if wrong_size or len(table) != items * channels:
+            msg = (
+                "The table should have either channels * size**3 float items "
+                "or size**3 items of channels-sized tuples with floats. "
+                f"Table should be: {channels}x{size[0]}x{size[1]}x{size[2]}. "
+                f"Actual length: {len(table)}"
+            )
+            raise ValueError(msg)
+        self.table = table
+
+    @staticmethod
+    def _check_size(size: Any) -> tuple[int, int, int]:
+        try:
+            _, _, _ = size
+        except ValueError as e:
+            msg = "Size should be either an integer or a tuple of three integers."
+            raise ValueError(msg) from e
+        except TypeError:
+            size = (size, size, size)
+        size = tuple(int(x) for x in size)
+        for size_1d in size:
+            if not 2 <= size_1d <= 65:
+                msg = "Size should be in [2, 65] range."
+                raise ValueError(msg)
+        return size
+
+    @classmethod
+    def generate(
+        cls,
+        size: int | tuple[int, int, int],
+        callback: Callable[[float, float, float], tuple[float, ...]],
+        channels: int = 3,
+        target_mode: str | None = None,
+    ) -> Color3DLUT:
+        """Generates new LUT using provided callback.
+
+        :param size: Size of the table. Passed to the constructor.
+        :param callback: Function with three parameters which correspond
+                         three color channels. Will be called ``size**3``
+                         times with values from 0.0 to 1.0 and should return
+                         a tuple with ``channels`` elements.
+        :param channels: The number of channels which should return callback.
+        :param target_mode: Passed to the constructor of the resulting
+                            lookup table.
+        """
+        size_1d, size_2d, size_3d = cls._check_size(size)
+        if channels not in (3, 4):
+            msg = "Only 3 or 4 output channels are supported"
+            raise ValueError(msg)
+
+        table: list[float] = [0] * (size_1d * size_2d * size_3d * channels)
+        idx_out = 0
+        for b in range(size_3d):
+            for g in range(size_2d):
+                for r in range(size_1d):
+                    table[idx_out : idx_out + channels] = callback(
+                        r / (size_1d - 1), g / (size_2d - 1), b / (size_3d - 1)
+                    )
+                    idx_out += channels
+
+        return cls(
+            (size_1d, size_2d, size_3d),
+            table,
+            channels=channels,
+            target_mode=target_mode,
+            _copy_table=False,
+        )
+
+    def transform(
+        self,
+        callback: Callable[..., tuple[float, ...]],
+        with_normals: bool = False,
+        channels: int | None = None,
+        target_mode: str | None = None,
+    ) -> Color3DLUT:
+        """Transforms the table values using provided callback and returns
+        a new LUT with altered values.
+
+        :param callback: A function which takes old lookup table values
+                         and returns a new set of values. The number
+                         of arguments which function should take is
+                         ``self.channels`` or ``3 + self.channels``
+                         if ``with_normals`` flag is set.
+                         Should return a tuple of ``self.channels`` or
+                         ``channels`` elements if it is set.
+        :param with_normals: If true, ``callback`` will be called with
+                             coordinates in the color cube as the first
+                             three arguments. Otherwise, ``callback``
+                             will be called only with actual color values.
+        :param channels: The number of channels in the resulting lookup table.
+        :param target_mode: Passed to the constructor of the resulting
+                            lookup table.
+        """
+        if channels not in (None, 3, 4):
+            msg = "Only 3 or 4 output channels are supported"
+            raise ValueError(msg)
+        ch_in = self.channels
+        ch_out = channels or ch_in
+        size_1d, size_2d, size_3d = self.size
+
+        table: list[float] = [0] * (size_1d * size_2d * size_3d * ch_out)
+        idx_in = 0
+        idx_out = 0
+        for b in range(size_3d):
+            for g in range(size_2d):
+                for r in range(size_1d):
+                    values = self.table[idx_in : idx_in + ch_in]
+                    if with_normals:
+                        values = callback(
+                            r / (size_1d - 1),
+                            g / (size_2d - 1),
+                            b / (size_3d - 1),
+                            *values,
+                        )
+                    else:
+                        values = callback(*values)
+                    table[idx_out : idx_out + ch_out] = values
+                    idx_in += ch_in
+                    idx_out += ch_out
+
+        return type(self)(
+            self.size,
+            table,
+            channels=ch_out,
+            target_mode=target_mode or self.mode,
+            _copy_table=False,
+        )
+
+    def __repr__(self) -> str:
+        r = [
+            f"{self.__class__.__name__} from {self.table.__class__.__name__}",
+            "size={:d}x{:d}x{:d}".format(*self.size),
+            f"channels={self.channels:d}",
+        ]
+        if self.mode:
+            r.append(f"target_mode={self.mode}")
+        return "<{}>".format(" ".join(r))
+
+    def filter(self, image: _imaging.ImagingCore) -> _imaging.ImagingCore:
+        from . import Image
+
+        return image.color_lut_3d(
+            self.mode or image.mode,
+            Image.Resampling.BILINEAR,
+            self.channels,
+            self.size[0],
+            self.size[1],
+            self.size[2],
+            self.table,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageFont.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageFont.py
new file mode 100644
index 00000000..d8c26556
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageFont.py
@@ -0,0 +1,1338 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PIL raster font management
+#
+# History:
+# 1996-08-07 fl   created (experimental)
+# 1997-08-25 fl   minor adjustments to handle fonts from pilfont 0.3
+# 1999-02-06 fl   rewrote most font management stuff in C
+# 1999-03-17 fl   take pth files into account in load_path (from Richard Jones)
+# 2001-02-17 fl   added freetype support
+# 2001-05-09 fl   added TransposedFont wrapper class
+# 2002-03-04 fl   make sure we have a "L" or "1" font
+# 2002-12-04 fl   skip non-directory entries in the system path
+# 2003-04-29 fl   add embedded default font
+# 2003-09-27 fl   added support for truetype charmap encodings
+#
+# Todo:
+# Adapt to PILFONT2 format (16-bit fonts, compressed, single file)
+#
+# Copyright (c) 1997-2003 by Secret Labs AB
+# Copyright (c) 1996-2003 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+
+from __future__ import annotations
+
+import base64
+import os
+import sys
+import warnings
+from enum import IntEnum
+from io import BytesIO
+from types import ModuleType
+from typing import IO, TYPE_CHECKING, Any, BinaryIO, TypedDict, cast
+
+from . import Image, features
+from ._typing import StrOrBytesPath
+from ._util import DeferredError, is_path
+
+if TYPE_CHECKING:
+    from . import ImageFile
+    from ._imaging import ImagingFont
+    from ._imagingft import Font
+
+
+class Axis(TypedDict):
+    minimum: int | None
+    default: int | None
+    maximum: int | None
+    name: bytes | None
+
+
+class Layout(IntEnum):
+    BASIC = 0
+    RAQM = 1
+
+
+MAX_STRING_LENGTH = 1_000_000
+
+
+core: ModuleType | DeferredError
+try:
+    from . import _imagingft as core
+except ImportError as ex:
+    core = DeferredError.new(ex)
+
+
+def _string_length_check(text: str | bytes | bytearray) -> None:
+    if MAX_STRING_LENGTH is not None and len(text) > MAX_STRING_LENGTH:
+        msg = "too many characters in string"
+        raise ValueError(msg)
+
+
+# FIXME: add support for pilfont2 format (see FontFile.py)
+
+# --------------------------------------------------------------------
+# Font metrics format:
+#       "PILfont" LF
+#       fontdescriptor LF
+#       (optional) key=value... LF
+#       "DATA" LF
+#       binary data: 256*10*2 bytes (dx, dy, dstbox, srcbox)
+#
+# To place a character, cut out srcbox and paste at dstbox,
+# relative to the character position.  Then move the character
+# position according to dx, dy.
+# --------------------------------------------------------------------
+
+
+class ImageFont:
+    """PIL font wrapper"""
+
+    font: ImagingFont
+
+    def _load_pilfont(self, filename: str) -> None:
+        with open(filename, "rb") as fp:
+            image: ImageFile.ImageFile | None = None
+            root = os.path.splitext(filename)[0]
+
+            for ext in (".png", ".gif", ".pbm"):
+                if image:
+                    image.close()
+                try:
+                    fullname = root + ext
+                    image = Image.open(fullname)
+                except Exception:
+                    pass
+                else:
+                    if image and image.mode in ("1", "L"):
+                        break
+            else:
+                if image:
+                    image.close()
+
+                msg = f"cannot find glyph data file {root}.{{gif|pbm|png}}"
+                raise OSError(msg)
+
+            self.file = fullname
+
+            self._load_pilfont_data(fp, image)
+            image.close()
+
+    def _load_pilfont_data(self, file: IO[bytes], image: Image.Image) -> None:
+        # read PILfont header
+        if file.readline() != b"PILfont\n":
+            msg = "Not a PILfont file"
+            raise SyntaxError(msg)
+        file.readline().split(b";")
+        self.info = []  # FIXME: should be a dictionary
+        while True:
+            s = file.readline()
+            if not s or s == b"DATA\n":
+                break
+            self.info.append(s)
+
+        # read PILfont metrics
+        data = file.read(256 * 20)
+
+        # check image
+        if image.mode not in ("1", "L"):
+            msg = "invalid font image mode"
+            raise TypeError(msg)
+
+        image.load()
+
+        self.font = Image.core.font(image.im, data)
+
+    def getmask(
+        self, text: str | bytes, mode: str = "", *args: Any, **kwargs: Any
+    ) -> Image.core.ImagingCore:
+        """
+        Create a bitmap for the text.
+
+        If the font uses antialiasing, the bitmap should have mode ``L`` and use a
+        maximum value of 255. Otherwise, it should have mode ``1``.
+
+        :param text: Text to render.
+        :param mode: Used by some graphics drivers to indicate what mode the
+                     driver prefers; if empty, the renderer may return either
+                     mode. Note that the mode is always a string, to simplify
+                     C-level implementations.
+
+                     .. versionadded:: 1.1.5
+
+        :return: An internal PIL storage memory instance as defined by the
+                 :py:mod:`PIL.Image.core` interface module.
+        """
+        _string_length_check(text)
+        Image._decompression_bomb_check(self.font.getsize(text))
+        return self.font.getmask(text, mode)
+
+    def getbbox(
+        self, text: str | bytes | bytearray, *args: Any, **kwargs: Any
+    ) -> tuple[int, int, int, int]:
+        """
+        Returns bounding box (in pixels) of given text.
+
+        .. versionadded:: 9.2.0
+
+        :param text: Text to render.
+
+        :return: ``(left, top, right, bottom)`` bounding box
+        """
+        _string_length_check(text)
+        width, height = self.font.getsize(text)
+        return 0, 0, width, height
+
+    def getlength(
+        self, text: str | bytes | bytearray, *args: Any, **kwargs: Any
+    ) -> int:
+        """
+        Returns length (in pixels) of given text.
+        This is the amount by which following text should be offset.
+
+        .. versionadded:: 9.2.0
+        """
+        _string_length_check(text)
+        width, height = self.font.getsize(text)
+        return width
+
+
+##
+# Wrapper for FreeType fonts.  Application code should use the
+# <b>truetype</b> factory function to create font objects.
+
+
+class FreeTypeFont:
+    """FreeType font wrapper (requires _imagingft service)"""
+
+    font: Font
+    font_bytes: bytes
+
+    def __init__(
+        self,
+        font: StrOrBytesPath | BinaryIO,
+        size: float = 10,
+        index: int = 0,
+        encoding: str = "",
+        layout_engine: Layout | None = None,
+    ) -> None:
+        # FIXME: use service provider instead
+
+        if isinstance(core, DeferredError):
+            raise core.ex
+
+        if size <= 0:
+            msg = f"font size must be greater than 0, not {size}"
+            raise ValueError(msg)
+
+        self.path = font
+        self.size = size
+        self.index = index
+        self.encoding = encoding
+
+        try:
+            from packaging.version import parse as parse_version
+        except ImportError:
+            pass
+        else:
+            if freetype_version := features.version_module("freetype2"):
+                if parse_version(freetype_version) < parse_version("2.9.1"):
+                    warnings.warn(
+                        "Support for FreeType 2.9.0 is deprecated and will be removed "
+                        "in Pillow 12 (2025-10-15). Please upgrade to FreeType 2.9.1 "
+                        "or newer, preferably FreeType 2.10.4 which fixes "
+                        "CVE-2020-15999.",
+                        DeprecationWarning,
+                    )
+
+        if layout_engine not in (Layout.BASIC, Layout.RAQM):
+            layout_engine = Layout.BASIC
+            if core.HAVE_RAQM:
+                layout_engine = Layout.RAQM
+        elif layout_engine == Layout.RAQM and not core.HAVE_RAQM:
+            warnings.warn(
+                "Raqm layout was requested, but Raqm is not available. "
+                "Falling back to basic layout."
+            )
+            layout_engine = Layout.BASIC
+
+        self.layout_engine = layout_engine
+
+        def load_from_bytes(f: IO[bytes]) -> None:
+            self.font_bytes = f.read()
+            self.font = core.getfont(
+                "", size, index, encoding, self.font_bytes, layout_engine
+            )
+
+        if is_path(font):
+            font = os.fspath(font)
+            if sys.platform == "win32":
+                font_bytes_path = font if isinstance(font, bytes) else font.encode()
+                try:
+                    font_bytes_path.decode("ascii")
+                except UnicodeDecodeError:
+                    # FreeType cannot load fonts with non-ASCII characters on Windows
+                    # So load it into memory first
+                    with open(font, "rb") as f:
+                        load_from_bytes(f)
+                    return
+            self.font = core.getfont(
+                font, size, index, encoding, layout_engine=layout_engine
+            )
+        else:
+            load_from_bytes(cast(IO[bytes], font))
+
+    def __getstate__(self) -> list[Any]:
+        return [self.path, self.size, self.index, self.encoding, self.layout_engine]
+
+    def __setstate__(self, state: list[Any]) -> None:
+        path, size, index, encoding, layout_engine = state
+        FreeTypeFont.__init__(self, path, size, index, encoding, layout_engine)
+
+    def getname(self) -> tuple[str | None, str | None]:
+        """
+        :return: A tuple of the font family (e.g. Helvetica) and the font style
+            (e.g. Bold)
+        """
+        return self.font.family, self.font.style
+
+    def getmetrics(self) -> tuple[int, int]:
+        """
+        :return: A tuple of the font ascent (the distance from the baseline to
+            the highest outline point) and descent (the distance from the
+            baseline to the lowest outline point, a negative value)
+        """
+        return self.font.ascent, self.font.descent
+
+    def getlength(
+        self,
+        text: str | bytes,
+        mode: str = "",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+    ) -> float:
+        """
+        Returns length (in pixels with 1/64 precision) of given text when rendered
+        in font with provided direction, features, and language.
+
+        This is the amount by which following text should be offset.
+        Text bounding box may extend past the length in some fonts,
+        e.g. when using italics or accents.
+
+        The result is returned as a float; it is a whole number if using basic layout.
+
+        Note that the sum of two lengths may not equal the length of a concatenated
+        string due to kerning. If you need to adjust for kerning, include the following
+        character and subtract its length.
+
+        For example, instead of ::
+
+          hello = font.getlength("Hello")
+          world = font.getlength("World")
+          hello_world = hello + world  # not adjusted for kerning
+          assert hello_world == font.getlength("HelloWorld")  # may fail
+
+        use ::
+
+          hello = font.getlength("HelloW") - font.getlength("W")  # adjusted for kerning
+          world = font.getlength("World")
+          hello_world = hello + world  # adjusted for kerning
+          assert hello_world == font.getlength("HelloWorld")  # True
+
+        or disable kerning with (requires libraqm) ::
+
+          hello = draw.textlength("Hello", font, features=["-kern"])
+          world = draw.textlength("World", font, features=["-kern"])
+          hello_world = hello + world  # kerning is disabled, no need to adjust
+          assert hello_world == draw.textlength("HelloWorld", font, features=["-kern"])
+
+        .. versionadded:: 8.0.0
+
+        :param text: Text to measure.
+        :param mode: Used by some graphics drivers to indicate what mode the
+                     driver prefers; if empty, the renderer may return either
+                     mode. Note that the mode is always a string, to simplify
+                     C-level implementations.
+
+        :param direction: Direction of the text. It can be 'rtl' (right to
+                          left), 'ltr' (left to right) or 'ttb' (top to bottom).
+                          Requires libraqm.
+
+        :param features: A list of OpenType font features to be used during text
+                         layout. This is usually used to turn on optional
+                         font features that are not enabled by default,
+                         for example 'dlig' or 'ss01', but can be also
+                         used to turn off default font features for
+                         example '-liga' to disable ligatures or '-kern'
+                         to disable kerning.  To get all supported
+                         features, see
+                         https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist
+                         Requires libraqm.
+
+        :param language: Language of the text. Different languages may use
+                         different glyph shapes or ligatures. This parameter tells
+                         the font which language the text is in, and to apply the
+                         correct substitutions as appropriate, if available.
+                         It should be a `BCP 47 language code
+                         <https://www.w3.org/International/articles/language-tags/>`_
+                         Requires libraqm.
+
+        :return: Either width for horizontal text, or height for vertical text.
+        """
+        _string_length_check(text)
+        return self.font.getlength(text, mode, direction, features, language) / 64
+
+    def getbbox(
+        self,
+        text: str | bytes,
+        mode: str = "",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        anchor: str | None = None,
+    ) -> tuple[float, float, float, float]:
+        """
+        Returns bounding box (in pixels) of given text relative to given anchor
+        when rendered in font with provided direction, features, and language.
+
+        Use :py:meth:`getlength()` to get the offset of following text with
+        1/64 pixel precision. The bounding box includes extra margins for
+        some fonts, e.g. italics or accents.
+
+        .. versionadded:: 8.0.0
+
+        :param text: Text to render.
+        :param mode: Used by some graphics drivers to indicate what mode the
+                     driver prefers; if empty, the renderer may return either
+                     mode. Note that the mode is always a string, to simplify
+                     C-level implementations.
+
+        :param direction: Direction of the text. It can be 'rtl' (right to
+                          left), 'ltr' (left to right) or 'ttb' (top to bottom).
+                          Requires libraqm.
+
+        :param features: A list of OpenType font features to be used during text
+                         layout. This is usually used to turn on optional
+                         font features that are not enabled by default,
+                         for example 'dlig' or 'ss01', but can be also
+                         used to turn off default font features for
+                         example '-liga' to disable ligatures or '-kern'
+                         to disable kerning.  To get all supported
+                         features, see
+                         https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist
+                         Requires libraqm.
+
+        :param language: Language of the text. Different languages may use
+                         different glyph shapes or ligatures. This parameter tells
+                         the font which language the text is in, and to apply the
+                         correct substitutions as appropriate, if available.
+                         It should be a `BCP 47 language code
+                         <https://www.w3.org/International/articles/language-tags/>`_
+                         Requires libraqm.
+
+        :param stroke_width: The width of the text stroke.
+
+        :param anchor:  The text anchor alignment. Determines the relative location of
+                        the anchor to the text. The default alignment is top left,
+                        specifically ``la`` for horizontal text and ``lt`` for
+                        vertical text. See :ref:`text-anchors` for details.
+
+        :return: ``(left, top, right, bottom)`` bounding box
+        """
+        _string_length_check(text)
+        size, offset = self.font.getsize(
+            text, mode, direction, features, language, anchor
+        )
+        left, top = offset[0] - stroke_width, offset[1] - stroke_width
+        width, height = size[0] + 2 * stroke_width, size[1] + 2 * stroke_width
+        return left, top, left + width, top + height
+
+    def getmask(
+        self,
+        text: str | bytes,
+        mode: str = "",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        anchor: str | None = None,
+        ink: int = 0,
+        start: tuple[float, float] | None = None,
+    ) -> Image.core.ImagingCore:
+        """
+        Create a bitmap for the text.
+
+        If the font uses antialiasing, the bitmap should have mode ``L`` and use a
+        maximum value of 255. If the font has embedded color data, the bitmap
+        should have mode ``RGBA``. Otherwise, it should have mode ``1``.
+
+        :param text: Text to render.
+        :param mode: Used by some graphics drivers to indicate what mode the
+                     driver prefers; if empty, the renderer may return either
+                     mode. Note that the mode is always a string, to simplify
+                     C-level implementations.
+
+                     .. versionadded:: 1.1.5
+
+        :param direction: Direction of the text. It can be 'rtl' (right to
+                          left), 'ltr' (left to right) or 'ttb' (top to bottom).
+                          Requires libraqm.
+
+                          .. versionadded:: 4.2.0
+
+        :param features: A list of OpenType font features to be used during text
+                         layout. This is usually used to turn on optional
+                         font features that are not enabled by default,
+                         for example 'dlig' or 'ss01', but can be also
+                         used to turn off default font features for
+                         example '-liga' to disable ligatures or '-kern'
+                         to disable kerning.  To get all supported
+                         features, see
+                         https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist
+                         Requires libraqm.
+
+                         .. versionadded:: 4.2.0
+
+        :param language: Language of the text. Different languages may use
+                         different glyph shapes or ligatures. This parameter tells
+                         the font which language the text is in, and to apply the
+                         correct substitutions as appropriate, if available.
+                         It should be a `BCP 47 language code
+                         <https://www.w3.org/International/articles/language-tags/>`_
+                         Requires libraqm.
+
+                         .. versionadded:: 6.0.0
+
+        :param stroke_width: The width of the text stroke.
+
+                         .. versionadded:: 6.2.0
+
+        :param anchor:  The text anchor alignment. Determines the relative location of
+                        the anchor to the text. The default alignment is top left,
+                        specifically ``la`` for horizontal text and ``lt`` for
+                        vertical text. See :ref:`text-anchors` for details.
+
+                         .. versionadded:: 8.0.0
+
+        :param ink: Foreground ink for rendering in RGBA mode.
+
+                         .. versionadded:: 8.0.0
+
+        :param start: Tuple of horizontal and vertical offset, as text may render
+                      differently when starting at fractional coordinates.
+
+                         .. versionadded:: 9.4.0
+
+        :return: An internal PIL storage memory instance as defined by the
+                 :py:mod:`PIL.Image.core` interface module.
+        """
+        return self.getmask2(
+            text,
+            mode,
+            direction=direction,
+            features=features,
+            language=language,
+            stroke_width=stroke_width,
+            anchor=anchor,
+            ink=ink,
+            start=start,
+        )[0]
+
+    def getmask2(
+        self,
+        text: str | bytes,
+        mode: str = "",
+        direction: str | None = None,
+        features: list[str] | None = None,
+        language: str | None = None,
+        stroke_width: float = 0,
+        anchor: str | None = None,
+        ink: int = 0,
+        start: tuple[float, float] | None = None,
+        *args: Any,
+        **kwargs: Any,
+    ) -> tuple[Image.core.ImagingCore, tuple[int, int]]:
+        """
+        Create a bitmap for the text.
+
+        If the font uses antialiasing, the bitmap should have mode ``L`` and use a
+        maximum value of 255. If the font has embedded color data, the bitmap
+        should have mode ``RGBA``. Otherwise, it should have mode ``1``.
+
+        :param text: Text to render.
+        :param mode: Used by some graphics drivers to indicate what mode the
+                     driver prefers; if empty, the renderer may return either
+                     mode. Note that the mode is always a string, to simplify
+                     C-level implementations.
+
+                     .. versionadded:: 1.1.5
+
+        :param direction: Direction of the text. It can be 'rtl' (right to
+                          left), 'ltr' (left to right) or 'ttb' (top to bottom).
+                          Requires libraqm.
+
+                          .. versionadded:: 4.2.0
+
+        :param features: A list of OpenType font features to be used during text
+                         layout. This is usually used to turn on optional
+                         font features that are not enabled by default,
+                         for example 'dlig' or 'ss01', but can be also
+                         used to turn off default font features for
+                         example '-liga' to disable ligatures or '-kern'
+                         to disable kerning.  To get all supported
+                         features, see
+                         https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist
+                         Requires libraqm.
+
+                         .. versionadded:: 4.2.0
+
+        :param language: Language of the text. Different languages may use
+                         different glyph shapes or ligatures. This parameter tells
+                         the font which language the text is in, and to apply the
+                         correct substitutions as appropriate, if available.
+                         It should be a `BCP 47 language code
+                         <https://www.w3.org/International/articles/language-tags/>`_
+                         Requires libraqm.
+
+                         .. versionadded:: 6.0.0
+
+        :param stroke_width: The width of the text stroke.
+
+                         .. versionadded:: 6.2.0
+
+        :param anchor:  The text anchor alignment. Determines the relative location of
+                        the anchor to the text. The default alignment is top left,
+                        specifically ``la`` for horizontal text and ``lt`` for
+                        vertical text. See :ref:`text-anchors` for details.
+
+                         .. versionadded:: 8.0.0
+
+        :param ink: Foreground ink for rendering in RGBA mode.
+
+                         .. versionadded:: 8.0.0
+
+        :param start: Tuple of horizontal and vertical offset, as text may render
+                      differently when starting at fractional coordinates.
+
+                         .. versionadded:: 9.4.0
+
+        :return: A tuple of an internal PIL storage memory instance as defined by the
+                 :py:mod:`PIL.Image.core` interface module, and the text offset, the
+                 gap between the starting coordinate and the first marking
+        """
+        _string_length_check(text)
+        if start is None:
+            start = (0, 0)
+
+        def fill(width: int, height: int) -> Image.core.ImagingCore:
+            size = (width, height)
+            Image._decompression_bomb_check(size)
+            return Image.core.fill("RGBA" if mode == "RGBA" else "L", size)
+
+        return self.font.render(
+            text,
+            fill,
+            mode,
+            direction,
+            features,
+            language,
+            stroke_width,
+            anchor,
+            ink,
+            start[0],
+            start[1],
+        )
+
+    def font_variant(
+        self,
+        font: StrOrBytesPath | BinaryIO | None = None,
+        size: float | None = None,
+        index: int | None = None,
+        encoding: str | None = None,
+        layout_engine: Layout | None = None,
+    ) -> FreeTypeFont:
+        """
+        Create a copy of this FreeTypeFont object,
+        using any specified arguments to override the settings.
+
+        Parameters are identical to the parameters used to initialize this
+        object.
+
+        :return: A FreeTypeFont object.
+        """
+        if font is None:
+            try:
+                font = BytesIO(self.font_bytes)
+            except AttributeError:
+                font = self.path
+        return FreeTypeFont(
+            font=font,
+            size=self.size if size is None else size,
+            index=self.index if index is None else index,
+            encoding=self.encoding if encoding is None else encoding,
+            layout_engine=layout_engine or self.layout_engine,
+        )
+
+    def get_variation_names(self) -> list[bytes]:
+        """
+        :returns: A list of the named styles in a variation font.
+        :exception OSError: If the font is not a variation font.
+        """
+        try:
+            names = self.font.getvarnames()
+        except AttributeError as e:
+            msg = "FreeType 2.9.1 or greater is required"
+            raise NotImplementedError(msg) from e
+        return [name.replace(b"\x00", b"") for name in names]
+
+    def set_variation_by_name(self, name: str | bytes) -> None:
+        """
+        :param name: The name of the style.
+        :exception OSError: If the font is not a variation font.
+        """
+        names = self.get_variation_names()
+        if not isinstance(name, bytes):
+            name = name.encode()
+        index = names.index(name) + 1
+
+        if index == getattr(self, "_last_variation_index", None):
+            # When the same name is set twice in a row,
+            # there is an 'unknown freetype error'
+            # https://savannah.nongnu.org/bugs/?56186
+            return
+        self._last_variation_index = index
+
+        self.font.setvarname(index)
+
+    def get_variation_axes(self) -> list[Axis]:
+        """
+        :returns: A list of the axes in a variation font.
+        :exception OSError: If the font is not a variation font.
+        """
+        try:
+            axes = self.font.getvaraxes()
+        except AttributeError as e:
+            msg = "FreeType 2.9.1 or greater is required"
+            raise NotImplementedError(msg) from e
+        for axis in axes:
+            if axis["name"]:
+                axis["name"] = axis["name"].replace(b"\x00", b"")
+        return axes
+
+    def set_variation_by_axes(self, axes: list[float]) -> None:
+        """
+        :param axes: A list of values for each axis.
+        :exception OSError: If the font is not a variation font.
+        """
+        try:
+            self.font.setvaraxes(axes)
+        except AttributeError as e:
+            msg = "FreeType 2.9.1 or greater is required"
+            raise NotImplementedError(msg) from e
+
+
+class TransposedFont:
+    """Wrapper for writing rotated or mirrored text"""
+
+    def __init__(
+        self, font: ImageFont | FreeTypeFont, orientation: Image.Transpose | None = None
+    ):
+        """
+        Wrapper that creates a transposed font from any existing font
+        object.
+
+        :param font: A font object.
+        :param orientation: An optional orientation.  If given, this should
+            be one of Image.Transpose.FLIP_LEFT_RIGHT, Image.Transpose.FLIP_TOP_BOTTOM,
+            Image.Transpose.ROTATE_90, Image.Transpose.ROTATE_180, or
+            Image.Transpose.ROTATE_270.
+        """
+        self.font = font
+        self.orientation = orientation  # any 'transpose' argument, or None
+
+    def getmask(
+        self, text: str | bytes, mode: str = "", *args: Any, **kwargs: Any
+    ) -> Image.core.ImagingCore:
+        im = self.font.getmask(text, mode, *args, **kwargs)
+        if self.orientation is not None:
+            return im.transpose(self.orientation)
+        return im
+
+    def getbbox(
+        self, text: str | bytes, *args: Any, **kwargs: Any
+    ) -> tuple[int, int, float, float]:
+        # TransposedFont doesn't support getmask2, move top-left point to (0, 0)
+        # this has no effect on ImageFont and simulates anchor="lt" for FreeTypeFont
+        left, top, right, bottom = self.font.getbbox(text, *args, **kwargs)
+        width = right - left
+        height = bottom - top
+        if self.orientation in (Image.Transpose.ROTATE_90, Image.Transpose.ROTATE_270):
+            return 0, 0, height, width
+        return 0, 0, width, height
+
+    def getlength(self, text: str | bytes, *args: Any, **kwargs: Any) -> float:
+        if self.orientation in (Image.Transpose.ROTATE_90, Image.Transpose.ROTATE_270):
+            msg = "text length is undefined for text rotated by 90 or 270 degrees"
+            raise ValueError(msg)
+        return self.font.getlength(text, *args, **kwargs)
+
+
+def load(filename: str) -> ImageFont:
+    """
+    Load a font file. This function loads a font object from the given
+    bitmap font file, and returns the corresponding font object. For loading TrueType
+    or OpenType fonts instead, see :py:func:`~PIL.ImageFont.truetype`.
+
+    :param filename: Name of font file.
+    :return: A font object.
+    :exception OSError: If the file could not be read.
+    """
+    f = ImageFont()
+    f._load_pilfont(filename)
+    return f
+
+
+def truetype(
+    font: StrOrBytesPath | BinaryIO,
+    size: float = 10,
+    index: int = 0,
+    encoding: str = "",
+    layout_engine: Layout | None = None,
+) -> FreeTypeFont:
+    """
+    Load a TrueType or OpenType font from a file or file-like object,
+    and create a font object. This function loads a font object from the given
+    file or file-like object, and creates a font object for a font of the given
+    size. For loading bitmap fonts instead, see :py:func:`~PIL.ImageFont.load`
+    and :py:func:`~PIL.ImageFont.load_path`.
+
+    Pillow uses FreeType to open font files. On Windows, be aware that FreeType
+    will keep the file open as long as the FreeTypeFont object exists. Windows
+    limits the number of files that can be open in C at once to 512, so if many
+    fonts are opened simultaneously and that limit is approached, an
+    ``OSError`` may be thrown, reporting that FreeType "cannot open resource".
+    A workaround would be to copy the file(s) into memory, and open that instead.
+
+    This function requires the _imagingft service.
+
+    :param font: A filename or file-like object containing a TrueType font.
+                 If the file is not found in this filename, the loader may also
+                 search in other directories, such as:
+
+                 * The :file:`fonts/` directory on Windows,
+                 * :file:`/Library/Fonts/`, :file:`/System/Library/Fonts/`
+                   and :file:`~/Library/Fonts/` on macOS.
+                 * :file:`~/.local/share/fonts`, :file:`/usr/local/share/fonts`,
+                   and :file:`/usr/share/fonts` on Linux; or those specified by
+                   the ``XDG_DATA_HOME`` and ``XDG_DATA_DIRS`` environment variables
+                   for user-installed and system-wide fonts, respectively.
+
+    :param size: The requested size, in pixels.
+    :param index: Which font face to load (default is first available face).
+    :param encoding: Which font encoding to use (default is Unicode). Possible
+                     encodings include (see the FreeType documentation for more
+                     information):
+
+                     * "unic" (Unicode)
+                     * "symb" (Microsoft Symbol)
+                     * "ADOB" (Adobe Standard)
+                     * "ADBE" (Adobe Expert)
+                     * "ADBC" (Adobe Custom)
+                     * "armn" (Apple Roman)
+                     * "sjis" (Shift JIS)
+                     * "gb  " (PRC)
+                     * "big5"
+                     * "wans" (Extended Wansung)
+                     * "joha" (Johab)
+                     * "lat1" (Latin-1)
+
+                     This specifies the character set to use. It does not alter the
+                     encoding of any text provided in subsequent operations.
+    :param layout_engine: Which layout engine to use, if available:
+                     :attr:`.ImageFont.Layout.BASIC` or :attr:`.ImageFont.Layout.RAQM`.
+                     If it is available, Raqm layout will be used by default.
+                     Otherwise, basic layout will be used.
+
+                     Raqm layout is recommended for all non-English text. If Raqm layout
+                     is not required, basic layout will have better performance.
+
+                     You can check support for Raqm layout using
+                     :py:func:`PIL.features.check_feature` with ``feature="raqm"``.
+
+                     .. versionadded:: 4.2.0
+    :return: A font object.
+    :exception OSError: If the file could not be read.
+    :exception ValueError: If the font size is not greater than zero.
+    """
+
+    def freetype(font: StrOrBytesPath | BinaryIO) -> FreeTypeFont:
+        return FreeTypeFont(font, size, index, encoding, layout_engine)
+
+    try:
+        return freetype(font)
+    except OSError:
+        if not is_path(font):
+            raise
+        ttf_filename = os.path.basename(font)
+
+        dirs = []
+        if sys.platform == "win32":
+            # check the windows font repository
+            # NOTE: must use uppercase WINDIR, to work around bugs in
+            # 1.5.2's os.environ.get()
+            windir = os.environ.get("WINDIR")
+            if windir:
+                dirs.append(os.path.join(windir, "fonts"))
+        elif sys.platform in ("linux", "linux2"):
+            data_home = os.environ.get("XDG_DATA_HOME")
+            if not data_home:
+                # The freedesktop spec defines the following default directory for
+                # when XDG_DATA_HOME is unset or empty. This user-level directory
+                # takes precedence over system-level directories.
+                data_home = os.path.expanduser("~/.local/share")
+            xdg_dirs = [data_home]
+
+            data_dirs = os.environ.get("XDG_DATA_DIRS")
+            if not data_dirs:
+                # Similarly, defaults are defined for the system-level directories
+                data_dirs = "/usr/local/share:/usr/share"
+            xdg_dirs += data_dirs.split(":")
+
+            dirs += [os.path.join(xdg_dir, "fonts") for xdg_dir in xdg_dirs]
+        elif sys.platform == "darwin":
+            dirs += [
+                "/Library/Fonts",
+                "/System/Library/Fonts",
+                os.path.expanduser("~/Library/Fonts"),
+            ]
+
+        ext = os.path.splitext(ttf_filename)[1]
+        first_font_with_a_different_extension = None
+        for directory in dirs:
+            for walkroot, walkdir, walkfilenames in os.walk(directory):
+                for walkfilename in walkfilenames:
+                    if ext and walkfilename == ttf_filename:
+                        return freetype(os.path.join(walkroot, walkfilename))
+                    elif not ext and os.path.splitext(walkfilename)[0] == ttf_filename:
+                        fontpath = os.path.join(walkroot, walkfilename)
+                        if os.path.splitext(fontpath)[1] == ".ttf":
+                            return freetype(fontpath)
+                        if not ext and first_font_with_a_different_extension is None:
+                            first_font_with_a_different_extension = fontpath
+        if first_font_with_a_different_extension:
+            return freetype(first_font_with_a_different_extension)
+        raise
+
+
+def load_path(filename: str | bytes) -> ImageFont:
+    """
+    Load font file. Same as :py:func:`~PIL.ImageFont.load`, but searches for a
+    bitmap font along the Python path.
+
+    :param filename: Name of font file.
+    :return: A font object.
+    :exception OSError: If the file could not be read.
+    """
+    if not isinstance(filename, str):
+        filename = filename.decode("utf-8")
+    for directory in sys.path:
+        try:
+            return load(os.path.join(directory, filename))
+        except OSError:
+            pass
+    msg = f'cannot find font file "{filename}" in sys.path'
+    if os.path.exists(filename):
+        msg += f', did you mean ImageFont.load("{filename}") instead?'
+
+    raise OSError(msg)
+
+
+def load_default_imagefont() -> ImageFont:
+    f = ImageFont()
+    f._load_pilfont_data(
+        # courB08
+        BytesIO(
+            base64.b64decode(
+                b"""
+UElMZm9udAo7Ozs7OzsxMDsKREFUQQoAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAYAAAAA//8AAQAAAAAAAAABAAEA
+BgAAAAH/+gADAAAAAQAAAAMABgAGAAAAAf/6AAT//QADAAAABgADAAYAAAAA//kABQABAAYAAAAL
+AAgABgAAAAD/+AAFAAEACwAAABAACQAGAAAAAP/5AAUAAAAQAAAAFQAHAAYAAP////oABQAAABUA
+AAAbAAYABgAAAAH/+QAE//wAGwAAAB4AAwAGAAAAAf/5AAQAAQAeAAAAIQAIAAYAAAAB//kABAAB
+ACEAAAAkAAgABgAAAAD/+QAE//0AJAAAACgABAAGAAAAAP/6AAX//wAoAAAALQAFAAYAAAAB//8A
+BAACAC0AAAAwAAMABgAAAAD//AAF//0AMAAAADUAAQAGAAAAAf//AAMAAAA1AAAANwABAAYAAAAB
+//kABQABADcAAAA7AAgABgAAAAD/+QAFAAAAOwAAAEAABwAGAAAAAP/5AAYAAABAAAAARgAHAAYA
+AAAA//kABQAAAEYAAABLAAcABgAAAAD/+QAFAAAASwAAAFAABwAGAAAAAP/5AAYAAABQAAAAVgAH
+AAYAAAAA//kABQAAAFYAAABbAAcABgAAAAD/+QAFAAAAWwAAAGAABwAGAAAAAP/5AAUAAABgAAAA
+ZQAHAAYAAAAA//kABQAAAGUAAABqAAcABgAAAAD/+QAFAAAAagAAAG8ABwAGAAAAAf/8AAMAAABv
+AAAAcQAEAAYAAAAA//wAAwACAHEAAAB0AAYABgAAAAD/+gAE//8AdAAAAHgABQAGAAAAAP/7AAT/
+/gB4AAAAfAADAAYAAAAB//oABf//AHwAAACAAAUABgAAAAD/+gAFAAAAgAAAAIUABgAGAAAAAP/5
+AAYAAQCFAAAAiwAIAAYAAP////oABgAAAIsAAACSAAYABgAA////+gAFAAAAkgAAAJgABgAGAAAA
+AP/6AAUAAACYAAAAnQAGAAYAAP////oABQAAAJ0AAACjAAYABgAA////+gAFAAAAowAAAKkABgAG
+AAD////6AAUAAACpAAAArwAGAAYAAAAA//oABQAAAK8AAAC0AAYABgAA////+gAGAAAAtAAAALsA
+BgAGAAAAAP/6AAQAAAC7AAAAvwAGAAYAAP////oABQAAAL8AAADFAAYABgAA////+gAGAAAAxQAA
+AMwABgAGAAD////6AAUAAADMAAAA0gAGAAYAAP////oABQAAANIAAADYAAYABgAA////+gAGAAAA
+2AAAAN8ABgAGAAAAAP/6AAUAAADfAAAA5AAGAAYAAP////oABQAAAOQAAADqAAYABgAAAAD/+gAF
+AAEA6gAAAO8ABwAGAAD////6AAYAAADvAAAA9gAGAAYAAAAA//oABQAAAPYAAAD7AAYABgAA////
++gAFAAAA+wAAAQEABgAGAAD////6AAYAAAEBAAABCAAGAAYAAP////oABgAAAQgAAAEPAAYABgAA
+////+gAGAAABDwAAARYABgAGAAAAAP/6AAYAAAEWAAABHAAGAAYAAP////oABgAAARwAAAEjAAYA
+BgAAAAD/+gAFAAABIwAAASgABgAGAAAAAf/5AAQAAQEoAAABKwAIAAYAAAAA//kABAABASsAAAEv
+AAgABgAAAAH/+QAEAAEBLwAAATIACAAGAAAAAP/5AAX//AEyAAABNwADAAYAAAAAAAEABgACATcA
+AAE9AAEABgAAAAH/+QAE//wBPQAAAUAAAwAGAAAAAP/7AAYAAAFAAAABRgAFAAYAAP////kABQAA
+AUYAAAFMAAcABgAAAAD/+wAFAAABTAAAAVEABQAGAAAAAP/5AAYAAAFRAAABVwAHAAYAAAAA//sA
+BQAAAVcAAAFcAAUABgAAAAD/+QAFAAABXAAAAWEABwAGAAAAAP/7AAYAAgFhAAABZwAHAAYAAP//
+//kABQAAAWcAAAFtAAcABgAAAAD/+QAGAAABbQAAAXMABwAGAAAAAP/5AAQAAgFzAAABdwAJAAYA
+AP////kABgAAAXcAAAF+AAcABgAAAAD/+QAGAAABfgAAAYQABwAGAAD////7AAUAAAGEAAABigAF
+AAYAAP////sABQAAAYoAAAGQAAUABgAAAAD/+wAFAAABkAAAAZUABQAGAAD////7AAUAAgGVAAAB
+mwAHAAYAAAAA//sABgACAZsAAAGhAAcABgAAAAD/+wAGAAABoQAAAacABQAGAAAAAP/7AAYAAAGn
+AAABrQAFAAYAAAAA//kABgAAAa0AAAGzAAcABgAA////+wAGAAABswAAAboABQAGAAD////7AAUA
+AAG6AAABwAAFAAYAAP////sABgAAAcAAAAHHAAUABgAAAAD/+wAGAAABxwAAAc0ABQAGAAD////7
+AAYAAgHNAAAB1AAHAAYAAAAA//sABQAAAdQAAAHZAAUABgAAAAH/+QAFAAEB2QAAAd0ACAAGAAAA
+Av/6AAMAAQHdAAAB3gAHAAYAAAAA//kABAABAd4AAAHiAAgABgAAAAD/+wAF//0B4gAAAecAAgAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAYAAAAB
+//sAAwACAecAAAHpAAcABgAAAAD/+QAFAAEB6QAAAe4ACAAGAAAAAP/5AAYAAAHuAAAB9AAHAAYA
+AAAA//oABf//AfQAAAH5AAUABgAAAAD/+QAGAAAB+QAAAf8ABwAGAAAAAv/5AAMAAgH/AAACAAAJ
+AAYAAAAA//kABQABAgAAAAIFAAgABgAAAAH/+gAE//sCBQAAAggAAQAGAAAAAP/5AAYAAAIIAAAC
+DgAHAAYAAAAB//kABf/+Ag4AAAISAAUABgAA////+wAGAAACEgAAAhkABQAGAAAAAP/7AAX//gIZ
+AAACHgADAAYAAAAA//wABf/9Ah4AAAIjAAEABgAAAAD/+QAHAAACIwAAAioABwAGAAAAAP/6AAT/
++wIqAAACLgABAAYAAAAA//kABP/8Ai4AAAIyAAMABgAAAAD/+gAFAAACMgAAAjcABgAGAAAAAf/5
+AAT//QI3AAACOgAEAAYAAAAB//kABP/9AjoAAAI9AAQABgAAAAL/+QAE//sCPQAAAj8AAgAGAAD/
+///7AAYAAgI/AAACRgAHAAYAAAAA//kABgABAkYAAAJMAAgABgAAAAH//AAD//0CTAAAAk4AAQAG
+AAAAAf//AAQAAgJOAAACUQADAAYAAAAB//kABP/9AlEAAAJUAAQABgAAAAH/+QAF//4CVAAAAlgA
+BQAGAAD////7AAYAAAJYAAACXwAFAAYAAP////kABgAAAl8AAAJmAAcABgAA////+QAGAAACZgAA
+Am0ABwAGAAD////5AAYAAAJtAAACdAAHAAYAAAAA//sABQACAnQAAAJ5AAcABgAA////9wAGAAAC
+eQAAAoAACQAGAAD////3AAYAAAKAAAAChwAJAAYAAP////cABgAAAocAAAKOAAkABgAA////9wAG
+AAACjgAAApUACQAGAAD////4AAYAAAKVAAACnAAIAAYAAP////cABgAAApwAAAKjAAkABgAA////
++gAGAAACowAAAqoABgAGAAAAAP/6AAUAAgKqAAACrwAIAAYAAP////cABQAAAq8AAAK1AAkABgAA
+////9wAFAAACtQAAArsACQAGAAD////3AAUAAAK7AAACwQAJAAYAAP////gABQAAAsEAAALHAAgA
+BgAAAAD/9wAEAAACxwAAAssACQAGAAAAAP/3AAQAAALLAAACzwAJAAYAAAAA//cABAAAAs8AAALT
+AAkABgAAAAD/+AAEAAAC0wAAAtcACAAGAAD////6AAUAAALXAAAC3QAGAAYAAP////cABgAAAt0A
+AALkAAkABgAAAAD/9wAFAAAC5AAAAukACQAGAAAAAP/3AAUAAALpAAAC7gAJAAYAAAAA//cABQAA
+Au4AAALzAAkABgAAAAD/9wAFAAAC8wAAAvgACQAGAAAAAP/4AAUAAAL4AAAC/QAIAAYAAAAA//oA
+Bf//Av0AAAMCAAUABgAA////+gAGAAADAgAAAwkABgAGAAD////3AAYAAAMJAAADEAAJAAYAAP//
+//cABgAAAxAAAAMXAAkABgAA////9wAGAAADFwAAAx4ACQAGAAD////4AAYAAAAAAAoABwASAAYA
+AP////cABgAAAAcACgAOABMABgAA////+gAFAAAADgAKABQAEAAGAAD////6AAYAAAAUAAoAGwAQ
+AAYAAAAA//gABgAAABsACgAhABIABgAAAAD/+AAGAAAAIQAKACcAEgAGAAAAAP/4AAYAAAAnAAoA
+LQASAAYAAAAA//gABgAAAC0ACgAzABIABgAAAAD/+QAGAAAAMwAKADkAEQAGAAAAAP/3AAYAAAA5
+AAoAPwATAAYAAP////sABQAAAD8ACgBFAA8ABgAAAAD/+wAFAAIARQAKAEoAEQAGAAAAAP/4AAUA
+AABKAAoATwASAAYAAAAA//gABQAAAE8ACgBUABIABgAAAAD/+AAFAAAAVAAKAFkAEgAGAAAAAP/5
+AAUAAABZAAoAXgARAAYAAAAA//gABgAAAF4ACgBkABIABgAAAAD/+AAGAAAAZAAKAGoAEgAGAAAA
+AP/4AAYAAABqAAoAcAASAAYAAAAA//kABgAAAHAACgB2ABEABgAAAAD/+AAFAAAAdgAKAHsAEgAG
+AAD////4AAYAAAB7AAoAggASAAYAAAAA//gABQAAAIIACgCHABIABgAAAAD/+AAFAAAAhwAKAIwA
+EgAGAAAAAP/4AAUAAACMAAoAkQASAAYAAAAA//gABQAAAJEACgCWABIABgAAAAD/+QAFAAAAlgAK
+AJsAEQAGAAAAAP/6AAX//wCbAAoAoAAPAAYAAAAA//oABQABAKAACgClABEABgAA////+AAGAAAA
+pQAKAKwAEgAGAAD////4AAYAAACsAAoAswASAAYAAP////gABgAAALMACgC6ABIABgAA////+QAG
+AAAAugAKAMEAEQAGAAD////4AAYAAgDBAAoAyAAUAAYAAP////kABQACAMgACgDOABMABgAA////
++QAGAAIAzgAKANUAEw==
+"""
+            )
+        ),
+        Image.open(
+            BytesIO(
+                base64.b64decode(
+                    b"""
+iVBORw0KGgoAAAANSUhEUgAAAx4AAAAUAQAAAAArMtZoAAAEwElEQVR4nABlAJr/AHVE4czCI/4u
+Mc4b7vuds/xzjz5/3/7u/n9vMe7vnfH/9++vPn/xyf5zhxzjt8GHw8+2d83u8x27199/nxuQ6Od9
+M43/5z2I+9n9ZtmDBwMQECDRQw/eQIQohJXxpBCNVE6QCCAAAAD//wBlAJr/AgALyj1t/wINwq0g
+LeNZUworuN1cjTPIzrTX6ofHWeo3v336qPzfEwRmBnHTtf95/fglZK5N0PDgfRTslpGBvz7LFc4F
+IUXBWQGjQ5MGCx34EDFPwXiY4YbYxavpnhHFrk14CDAAAAD//wBlAJr/AgKqRooH2gAgPeggvUAA
+Bu2WfgPoAwzRAABAAAAAAACQgLz/3Uv4Gv+gX7BJgDeeGP6AAAD1NMDzKHD7ANWr3loYbxsAD791
+NAADfcoIDyP44K/jv4Y63/Z+t98Ovt+ub4T48LAAAAD//wBlAJr/AuplMlADJAAAAGuAphWpqhMx
+in0A/fRvAYBABPgBwBUgABBQ/sYAyv9g0bCHgOLoGAAAAAAAREAAwI7nr0ArYpow7aX8//9LaP/9
+SjdavWA8ePHeBIKB//81/83ndznOaXx379wAAAD//wBlAJr/AqDxW+D3AABAAbUh/QMnbQag/gAY
+AYDAAACgtgD/gOqAAAB5IA/8AAAk+n9w0AAA8AAAmFRJuPo27ciC0cD5oeW4E7KA/wD3ECMAn2tt
+y8PgwH8AfAxFzC0JzeAMtratAsC/ffwAAAD//wBlAJr/BGKAyCAA4AAAAvgeYTAwHd1kmQF5chkG
+ABoMIHcL5xVpTfQbUqzlAAAErwAQBgAAEOClA5D9il08AEh/tUzdCBsXkbgACED+woQg8Si9VeqY
+lODCn7lmF6NhnAEYgAAA/NMIAAAAAAD//2JgjLZgVGBg5Pv/Tvpc8hwGBjYGJADjHDrAwPzAjv/H
+/Wf3PzCwtzcwHmBgYGcwbZz8wHaCAQMDOwMDQ8MCBgYOC3W7mp+f0w+wHOYxO3OG+e376hsMZjk3
+AAAAAP//YmCMY2A4wMAIN5e5gQETPD6AZisDAwMDgzSDAAPjByiHcQMDAwMDg1nOze1lByRu5/47
+c4859311AYNZzg0AAAAA//9iYGDBYihOIIMuwIjGL39/fwffA8b//xv/P2BPtzzHwCBjUQAAAAD/
+/yLFBrIBAAAA//9i1HhcwdhizX7u8NZNzyLbvT97bfrMf/QHI8evOwcSqGUJAAAA//9iYBB81iSw
+pEE170Qrg5MIYydHqwdDQRMrAwcVrQAAAAD//2J4x7j9AAMDn8Q/BgYLBoaiAwwMjPdvMDBYM1Tv
+oJodAAAAAP//Yqo/83+dxePWlxl3npsel9lvLfPcqlE9725C+acfVLMEAAAA//9i+s9gwCoaaGMR
+evta/58PTEWzr21hufPjA8N+qlnBwAAAAAD//2JiWLci5v1+HmFXDqcnULE/MxgYGBj+f6CaJQAA
+AAD//2Ji2FrkY3iYpYC5qDeGgeEMAwPDvwQBBoYvcTwOVLMEAAAA//9isDBgkP///0EOg9z35v//
+Gc/eeW7BwPj5+QGZhANUswMAAAD//2JgqGBgYGBgqEMXlvhMPUsAAAAA//8iYDd1AAAAAP//AwDR
+w7IkEbzhVQAAAABJRU5ErkJggg==
+"""
+                )
+            )
+        ),
+    )
+    return f
+
+
+def load_default(size: float | None = None) -> FreeTypeFont | ImageFont:
+    """If FreeType support is available, load a version of Aileron Regular,
+    https://dotcolon.net/font/aileron, with a more limited character set.
+
+    Otherwise, load a "better than nothing" font.
+
+    .. versionadded:: 1.1.4
+
+    :param size: The font size of Aileron Regular.
+
+        .. versionadded:: 10.1.0
+
+    :return: A font object.
+    """
+    if isinstance(core, ModuleType) or size is not None:
+        return truetype(
+            BytesIO(
+                base64.b64decode(
+                    b"""
+AAEAAAAPAIAAAwBwRkZUTYwDlUAAADFoAAAAHEdERUYAqADnAAAo8AAAACRHUE9ThhmITwAAKfgAA
+AduR1NVQnHxefoAACkUAAAA4k9TLzJovoHLAAABeAAAAGBjbWFw5lFQMQAAA6gAAAGqZ2FzcP//AA
+MAACjoAAAACGdseWYmRXoPAAAGQAAAHfhoZWFkE18ayQAAAPwAAAA2aGhlYQboArEAAAE0AAAAJGh
+tdHjjERZ8AAAB2AAAAdBsb2NhuOexrgAABVQAAADqbWF4cAC7AEYAAAFYAAAAIG5hbWUr+h5lAAAk
+OAAAA6Jwb3N0D3oPTQAAJ9wAAAEKAAEAAAABGhxJDqIhXw889QALA+gAAAAA0Bqf2QAAAADhCh2h/
+2r/LgOxAyAAAAAIAAIAAAAAAAAAAQAAA8r/GgAAA7j/av9qA7EAAQAAAAAAAAAAAAAAAAAAAHQAAQ
+AAAHQAQwAFAAAAAAACAAAAAQABAAAAQAAAAAAAAAADAfoBkAAFAAgCigJYAAAASwKKAlgAAAFeADI
+BPgAAAAAFAAAAAAAAAAAAAAcAAAAAAAAAAAAAAABVS1dOAEAAIPsCAwL/GgDIA8oA5iAAAJMAAAAA
+AhICsgAAACAAAwH0AAAAAAAAAU0AAADYAAAA8gA5AVMAVgJEAEYCRAA1AuQAKQKOAEAAsAArATsAZ
+AE7AB4CMABVAkQAUADc/+EBEgAgANwAJQEv//sCRAApAkQAggJEADwCRAAtAkQAIQJEADkCRAArAk
+QAMgJEACwCRAAxANwAJQDc/+ECRABnAkQAUAJEAEQB8wAjA1QANgJ/AB0CcwBkArsALwLFAGQCSwB
+kAjcAZALGAC8C2gBkAQgAZAIgADcCYQBkAj8AZANiAGQCzgBkAuEALwJWAGQC3QAvAmsAZAJJADQC
+ZAAiAqoAXgJuACADuAAaAnEAGQJFABMCTwAuATMAYgEv//sBJwAiAkQAUAH0ADIBLAApAhMAJAJjA
+EoCEQAeAmcAHgIlAB4BIgAVAmcAHgJRAEoA7gA+AOn/8wIKAEoA9wBGA1cASgJRAEoCSgAeAmMASg
+JnAB4BSgBKAcsAGAE5ABQCUABCAgIAAQMRAAEB4v/6AgEAAQHOABQBLwBAAPoAYAEvACECRABNA0Y
+AJAItAHgBKgAcAkQAUAEsAHQAygAgAi0AOQD3ADYA9wAWAaEANgGhABYCbAAlAYMAeAGDADkA6/9q
+AhsAFAIKABUB/QAVAAAAAwAAAAMAAAAcAAEAAAAAAKQAAwABAAAAHAAEAIgAAAAeABAAAwAOAH4Aq
+QCrALEAtAC3ALsgGSAdICYgOiBEISL7Av//AAAAIACpAKsAsAC0ALcAuyAYIBwgJiA5IEQhIvsB//
+//4/+5/7j/tP+y/7D/reBR4E/gR+A14CzfTwVxAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAEGAAABAAAAAAAAAAECAAAAAgAAAAAAAAAAAAAAAAAAAAEAAAMEBQYHCAkKCwwNDg8QERIT
+FBUWFxgZGhscHR4fICEiIyQlJicoKSorLC0uLzAxMjM0NTY3ODk6Ozw9Pj9AQUJDREVGR0hJSktMT
+U5PUFFSU1RVVldYWVpbXF1eX2BhAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGQAAA
+AAAAAAYnFmAAAAAABlAAAAAAAAAAAAAAAAAAAAAAAAAAAAY2htAAAAAAAAAABrbGlqAAAAAHAAbm9
+ycwBnAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAmACYAJgAmAD4AUgCCAMoBCgFO
+AVwBcgGIAaYBvAHKAdYB6AH2AgwCIAJKAogCpgLWAw4DIgNkA5wDugPUA+gD/AQQBEYEogS8BPoFJ
+gVSBWoFgAWwBcoF1gX6BhQGJAZMBmgGiga0BuIHGgdUB2YHkAeiB8AH3AfyCAoIHAgqCDoITghcCG
+oIogjSCPoJKglYCXwJwgnqCgIKKApACl4Klgq8CtwLDAs8C1YLjAuyC9oL7gwMDCYMSAxgDKAMrAz
+qDQoNTA1mDYQNoA2uDcAN2g3oDfYODA4iDkoOXA5sDnoOnA7EDvwAAAAFAAAAAAH0ArwAAwAGAAkA
+DAAPAAAxESERAxMhExcRASELARETAfT6qv6syKr+jgFUqsiqArz9RAGLAP/+1P8B/v3VAP8BLP4CA
+P8AAgA5//IAuQKyAAMACwAANyMDMwIyFhQGIiY0oE4MZk84JCQ4JLQB/v3AJDgkJDgAAgBWAeUBPA
+LfAAMABwAAEyMnMxcjJzOmRgpagkYKWgHl+vr6AAAAAAIARgAAAf4CsgAbAB8AAAEHMxUjByM3Iwc
+jNyM1MzcjNTM3MwczNzMHMxUrAQczAZgdZXEvOi9bLzovWmYdZXEvOi9bLzovWp9bHlsBn4w429vb
+2ziMONvb29s4jAAAAAMANf+mAg4DDAAfACYALAAAJRQGBxUjNS4BJzMeARcRLgE0Njc1MxUeARcjJ
+icVHgEBFBYXNQ4BExU+ATU0Ag5xWDpgcgRcBz41Xl9oVTpVYwpcC1ttXP6cLTQuM5szOrVRZwlOTQ
+ZqVzZECAEAGlukZAlOTQdrUG8O7iNlAQgxNhDlCDj+8/YGOjReAAAAAAUAKf/yArsCvAAHAAsAFQA
+dACcAABIyFhQGIiY0EyMBMwQiBhUUFjI2NTQSMhYUBiImNDYiBhUUFjI2NTR5iFBQiFCVVwHAV/5c
+OiMjOiPmiFBQiFCxOiMjOiMCvFaSVlaS/ZoCsjIzMC80NC8w/uNWklZWkhozMC80NC8wAAAAAgBA/
+/ICbgLAACIALgAAARUjEQYjIiY1NDY3LgE1NDYzMhcVJiMiBhUUFhcWOwE1MxUFFBYzMjc1IyIHDg
+ECbmBcYYOOVkg7R4hsQjY4Q0RNRD4SLDxW/pJUXzksPCkUUk0BgUb+zBVUZ0BkDw5RO1huCkULQzp
+COAMBcHDHRz0J/AIHRQAAAAEAKwHlAIUC3wADAAATIycze0YKWgHl+gAAAAABAGT/sAEXAwwACQAA
+EzMGEBcjLgE0Nt06dXU6OUBAAwzG/jDGVePs4wAAAAEAHv+wANEDDAAJAAATMx4BFAYHIzYQHjo5Q
+EA5OnUDDFXj7ONVxgHQAAAAAQBVAFIB2wHbAA4AAAE3FwcXBycHJzcnNxcnMwEtmxOfcTJjYzJxnx
+ObCj4BKD07KYolmZkliik7PbMAAQBQAFUB9AIlAAsAAAEjFSM1IzUzNTMVMwH0tTq1tTq1AR/Kyjj
+OzgAAAAAB/+H/iACMAGQABAAANwcjNzOMWlFOXVrS3AAAAQAgAP8A8gE3AAMAABMjNTPy0tIA/zgA
+AQAl//IApQByAAcAADYyFhQGIiY0STgkJDgkciQ4JCQ4AAAAAf/7/+IBNALQAAMAABcjEzM5Pvs+H
+gLuAAAAAAIAKf/yAhsCwAADAAcAABIgECA2IBAgKQHy/g5gATL+zgLA/TJEAkYAAAAAAQCCAAABlg
+KyAAgAAAERIxEHNTc2MwGWVr6SIygCsv1OAldxW1sWAAEAPAAAAg4CwAAZAAA3IRUhNRM+ATU0JiM
+iDwEjNz4BMzIWFRQGB7kBUv4x+kI2QTt+EAFWAQp8aGVtSl5GRjEA/0RVLzlLmAoKa3FsUkNxXQAA
+AAEALf/yAhYCwAAqAAABHgEVFAYjIi8BMxceATMyNjU0KwE1MzI2NTQmIyIGDwEjNz4BMzIWFRQGA
+YxBSZJo2RUBVgEHV0JBUaQREUBUQzc5TQcBVgEKfGhfcEMBbxJbQl1x0AoKRkZHPn9GSD80QUVCCg
+pfbGBPOlgAAAACACEAAAIkArIACgAPAAAlIxUjNSE1ATMRMyMRBg8BAiRXVv6qAVZWV60dHLCurq4
+rAdn+QgFLMibzAAABADn/8gIZArIAHQAAATIWFRQGIyIvATMXFjMyNjU0JiMiByMTIRUhBzc2ATNv
+d5Fl1RQBVgIad0VSTkVhL1IwAYj+vh8rMAHHgGdtgcUKCoFXTU5bYgGRRvAuHQAAAAACACv/8gITA
+sAAFwAjAAABMhYVFAYjIhE0NjMyFh8BIycmIyIDNzYTMjY1NCYjIgYVFBYBLmp7imr0l3RZdAgBXA
+IYZ5wKJzU6QVNJSz5SUAHSgWltiQFGxcNlVQoKdv7sPiz+ZF1LTmJbU0lhAAAAAQAyAAACGgKyAAY
+AAAEVASMBITUCGv6oXAFL/oECsij9dgJsRgAAAAMALP/xAhgCwAAWACAALAAAAR4BFRQGIyImNTQ2
+Ny4BNTQ2MhYVFAYmIgYVFBYyNjU0AzI2NTQmIyIGFRQWAZQ5S5BmbIpPOjA7ecp5P2F8Q0J8RIVJS
+0pLTEtOAW0TXTxpZ2ZqPF0SE1A3VWVlVTdQ/UU0N0RENzT9/ko+Ok1NOj1LAAIAMf/yAhkCwAAXAC
+MAAAEyERQGIyImLwEzFxYzMhMHBiMiJjU0NhMyNjU0JiMiBhUUFgEl9Jd0WXQIAVwCGGecCic1SWp
+7imo+UlBAQVNJAsD+usXDZVUKCnYBFD4sgWltif5kW1NJYV1LTmIAAAACACX/8gClAiAABwAPAAAS
+MhYUBiImNBIyFhQGIiY0STgkJDgkJDgkJDgkAiAkOCQkOP52JDgkJDgAAAAC/+H/iAClAiAABwAMA
+AASMhYUBiImNBMHIzczSTgkJDgkaFpSTl4CICQ4JCQ4/mba5gAAAQBnAB4B+AH0AAYAAAENARUlNS
+UB+P6qAVb+bwGRAbCmpkbJRMkAAAIAUAC7AfQBuwADAAcAAAEhNSERITUhAfT+XAGk/lwBpAGDOP8
+AOAABAEQAHgHVAfQABgAAARUFNS0BNQHV/m8BVv6qAStEyUSmpkYAAAAAAgAj//IB1ALAABgAIAAA
+ATIWFRQHDgEHIz4BNz4BNTQmIyIGByM+ARIyFhQGIiY0AQRibmktIAJWBSEqNig+NTlHBFoDezQ4J
+CQ4JALAZ1BjaS03JS1DMD5LLDQ/SUVgcv2yJDgkJDgAAAAAAgA2/5gDFgKYADYAQgAAAQMGFRQzMj
+Y1NCYjIg4CFRQWMzI2NxcGIyImNTQ+AjMyFhUUBiMiJwcGIyImNTQ2MzIfATcHNzYmIyIGFRQzMjY
+Cej8EJjJJlnBAfGQ+oHtAhjUYg5OPx0h2k06Os3xRWQsVLjY5VHtdPBwJETcJDyUoOkZEJz8B0f74
+EQ8kZl6EkTFZjVOLlyknMVm1pmCiaTq4lX6CSCknTVRmmR8wPdYnQzxuSWVGAAIAHQAAAncCsgAHA
+AoAACUjByMTMxMjATMDAcj+UVz4dO5d/sjPZPT0ArL9TgE6ATQAAAADAGQAAAJMArIAEAAbACcAAA
+EeARUUBgcGKwERMzIXFhUUJRUzMjc2NTQnJiMTPgE1NCcmKwEVMzIBvkdHZkwiNt7LOSGq/oeFHBt
+hahIlSTM+cB8Yj5UWAW8QT0VYYgwFArIEF5Fv1eMED2NfDAL93AU+N24PBP0AAAAAAQAv//ICjwLA
+ABsAAAEyFh8BIycmIyIGFRQWMzI/ATMHDgEjIiY1NDYBdX+PCwFWAiKiaHx5ZaIiAlYBCpWBk6a0A
+sCAagoKpqN/gaOmCgplhcicn8sAAAIAZAAAAp8CsgAMABkAAAEeARUUBgcGKwERMzITPgE1NCYnJi
+sBETMyAY59lJp8IzXN0jUVWmdjWRs5d3I4Aq4QqJWUug8EArL9mQ+PeHGHDgX92gAAAAABAGQAAAI
+vArIACwAAJRUhESEVIRUhFSEVAi/+NQHB/pUBTf6zRkYCskbwRvAAAAABAGQAAAIlArIACQAAExUh
+FSERIxEhFboBQ/69VgHBAmzwRv7KArJGAAAAAAEAL//yAo8CwAAfAAABMxEjNQcGIyImNTQ2MzIWH
+wEjJyYjIgYVFBYzMjY1IwGP90wfPnWTprSSf48LAVYCIqJofHllVG+hAU3+s3hARsicn8uAagoKpq
+N/gaN1XAAAAAEAZAAAAowCsgALAAABESMRIREjETMRIRECjFb+hFZWAXwCsv1OAS7+0gKy/sQBPAA
+AAAABAGQAAAC6ArIAAwAAMyMRM7pWVgKyAAABADf/8gHoArIAEwAAAREUBw4BIyImLwEzFxYzMjc2
+NREB6AIFcGpgbQIBVgIHfXQKAQKy/lYxIltob2EpKYyEFD0BpwAAAAABAGQAAAJ0ArIACwAACQEjA
+wcVIxEzEQEzATsBJ3ntQlZWAVVlAWH+nwEnR+ACsv6RAW8AAQBkAAACLwKyAAUAACUVIREzEQIv/j
+VWRkYCsv2UAAABAGQAAAMUArIAFAAAAREjETQ3BgcDIwMmJxYVESMRMxsBAxRWAiMxemx8NxsCVo7
+MywKy/U4BY7ZLco7+nAFmoFxLtP6dArL9lwJpAAAAAAEAZAAAAoACsgANAAAhIwEWFREjETMBJjUR
+MwKAhP67A1aEAUUDVAJeeov+pwKy/aJ5jAFZAAAAAgAv//ICuwLAAAkAEwAAEiAWFRQGICY1NBIyN
+jU0JiIGFRTbATSsrP7MrNrYenrYegLAxaKhxsahov47nIeIm5uIhwACAGQAAAJHArIADgAYAAABHg
+EVFAYHBisBESMRMzITNjQnJisBETMyAZRUX2VOHzuAVtY7GlxcGDWIiDUCrgtnVlVpCgT+5gKy/rU
+V1BUF/vgAAAACAC//zAK9AsAAEgAcAAAlFhcHJiMiBwYjIiY1NDYgFhUUJRQWMjY1NCYiBgI9PUMx
+UDcfKh8omqysATSs/dR62Hp62HpICTg7NgkHxqGixcWitbWHnJyHiJubAAIAZAAAAlgCsgAXACMAA
+CUWFyMmJyYnJisBESMRMzIXHgEVFAYHFiUzMjc+ATU0JyYrAQIqDCJfGQwNWhAhglbiOx9QXEY1Tv
+6bhDATMj1lGSyMtYgtOXR0BwH+1wKyBApbU0BSESRAAgVAOGoQBAABADT/8gIoAsAAJQAAATIWFyM
+uASMiBhUUFhceARUUBiMiJiczHgEzMjY1NCYnLgE1NDYBOmd2ClwGS0E6SUNRdW+HZnKKC1wPWkQ9
+Uk1cZGuEAsBwXUJHNjQ3OhIbZVZZbm5kREo+NT5DFRdYUFdrAAAAAAEAIgAAAmQCsgAHAAABIxEjE
+SM1IQJk9lb2AkICbP2UAmxGAAEAXv/yAmQCsgAXAAABERQHDgEiJicmNREzERQXHgEyNjc2NRECZA
+IIgfCBCAJWAgZYmlgGAgKy/k0qFFxzc1wUKgGz/lUrEkRQUEQSKwGrAAAAAAEAIAAAAnoCsgAGAAA
+hIwMzGwEzAYJ07l3N1FwCsv2PAnEAAAEAGgAAA7ECsgAMAAABAyMLASMDMxsBMxsBA7HAcZyicrZi
+kaB0nJkCsv1OAlP9rQKy/ZsCW/2kAmYAAAEAGQAAAm8CsgALAAAhCwEjEwMzGwEzAxMCCsrEY/bkY
+re+Y/D6AST+3AFcAVb+5gEa/q3+oQAAAQATAAACUQKyAAgAAAERIxEDMxsBMwFdVvRjwLphARD+8A
+EQAaL+sQFPAAABAC4AAAI5ArIACQAAJRUhNQEhNSEVAQI5/fUBof57Aen+YUZGQgIqRkX92QAAAAA
+BAGL/sAEFAwwABwAAARUjETMVIxEBBWlpowMMOP0UOANcAAAB//v/4gE0AtAAAwAABSMDMwE0Pvs+
+HgLuAAAAAQAi/7AAxQMMAAcAABcjNTMRIzUzxaNpaaNQOALsOAABAFAA1wH0AmgABgAAJQsBIxMzE
+wGwjY1GsESw1wFZ/qcBkf5vAAAAAQAy/6oBwv/iAAMAAAUhNSEBwv5wAZBWOAAAAAEAKQJEALYCsg
+ADAAATIycztjhVUAJEbgAAAAACACT/8gHQAiAAHQAlAAAhJwcGIyImNTQ2OwE1NCcmIyIHIz4BMzI
+XFh0BFBcnMjY9ASYVFAF6CR0wVUtgkJoiAgdgaQlaBm1Zrg4DCuQ9R+5MOSFQR1tbDiwUUXBUXowf
+J8c9SjRORzYSgVwAAAAAAgBK//ICRQLfABEAHgAAATIWFRQGIyImLwEVIxEzETc2EzI2NTQmIyIGH
+QEUFgFUcYCVbiNJEyNWVigySElcU01JXmECIJd4i5QTEDRJAt/+3jkq/hRuZV55ZWsdX14AAQAe//
+IB9wIgABgAAAEyFhcjJiMiBhUUFjMyNjczDgEjIiY1NDYBF152DFocbEJXU0A1Rw1aE3pbaoKQAiB
+oWH5qZm1tPDlaXYuLgZcAAAACAB7/8gIZAt8AEQAeAAABESM1BwYjIiY1NDYzMhYfAREDMjY9ATQm
+IyIGFRQWAhlWKDJacYCVbiNJEyOnSV5hQUlcUwLf/SFVOSqXeIuUExA0ARb9VWVrHV9ebmVeeQACA
+B7/8gH9AiAAFQAbAAABFAchHgEzMjY3Mw4BIyImNTQ2MzIWJyIGByEmAf0C/oAGUkA1SwlaD4FXbI
+WObmt45UBVBwEqDQEYFhNjWD84W16Oh3+akU9aU60AAAEAFQAAARoC8gAWAAATBh0BMxUjESMRIzU
+zNTQ3PgEzMhcVJqcDbW1WOTkDB0k8Hx5oAngVITRC/jQBzEIsJRs5PwVHEwAAAAIAHv8uAhkCIAAi
+AC8AAAERFAcOASMiLwEzFx4BMzI2NzY9AQcGIyImNTQ2MzIWHwE1AzI2PQE0JiMiBhUUFgIZAQSEd
+NwRAVcBBU5DTlUDASgyWnGAlW4jSRMjp0leYUFJXFMCEv5wSh1zeq8KCTI8VU0ZIQk5Kpd4i5QTED
+RJ/iJlax1fXm5lXnkAAQBKAAACCgLkABcAAAEWFREjETQnLgEHDgEdASMRMxE3NjMyFgIIAlYCBDs
+6RVRWViE5UVViAYUbQP7WASQxGzI7AQJyf+kC5P7TPSxUAAACAD4AAACsAsAABwALAAASMhYUBiIm
+NBMjETNeLiAgLiBiVlYCwCAuICAu/WACEgAC//P/LgCnAsAABwAVAAASMhYUBiImNBcRFAcGIyInN
+RY3NjURWS4gIC4gYgMLcRwNSgYCAsAgLiAgLo79wCUbZAJGBzMOHgJEAAAAAQBKAAACCALfAAsAAC
+EnBxUjETMREzMHEwGTwTJWVvdu9/rgN6kC3/4oAQv6/ugAAQBG//wA3gLfAA8AABMRFBceATcVBiM
+iJicmNRGcAQIcIxkkKi4CAQLf/bkhERoSBD4EJC8SNAJKAAAAAQBKAAADEAIgACQAAAEWFREjETQn
+JiMiFREjETQnJiMiFREjETMVNzYzMhYXNzYzMhYDCwVWBAxedFYEDF50VlYiJko7ThAvJkpEVAGfI
+jn+vAEcQyRZ1v76ARxDJFnW/voCEk08HzYtRB9HAAAAAAEASgAAAgoCIAAWAAABFhURIxE0JyYjIg
+YdASMRMxU3NjMyFgIIAlYCCXBEVVZWITlRVWIBhRtA/tYBJDEbbHR/6QISWz0sVAAAAAACAB7/8gI
+sAiAABwARAAASIBYUBiAmNBIyNjU0JiIGFRSlAQCHh/8Ah7ieWlqeWgIgn/Cfn/D+s3ZfYHV1YF8A
+AgBK/zwCRQIgABEAHgAAATIWFRQGIyImLwERIxEzFTc2EzI2NTQmIyIGHQEUFgFUcYCVbiNJEyNWV
+igySElcU01JXmECIJd4i5QTEDT+8wLWVTkq/hRuZV55ZWsdX14AAgAe/zwCGQIgABEAHgAAAREjEQ
+cGIyImNTQ2MzIWHwE1AzI2PQE0JiMiBhUUFgIZVigyWnGAlW4jSRMjp0leYUFJXFMCEv0qARk5Kpd
+4i5QTEDRJ/iJlax1fXm5lXnkAAQBKAAABPgIeAA0AAAEyFxUmBhURIxEzFTc2ARoWDkdXVlYwIwIe
+B0EFVlf+0gISU0cYAAEAGP/yAa0CIAAjAAATMhYXIyYjIgYVFBYXHgEVFAYjIiYnMxYzMjY1NCYnL
+gE1NDbkV2MJWhNdKy04PF1XbVhWbgxaE2ktOjlEUllkAiBaS2MrJCUoEBlPQkhOVFZoKCUmLhIWSE
+BIUwAAAAEAFP/4ARQCiQAXAAATERQXHgE3FQYjIiYnJjURIzUzNTMVMxWxAQMmMx8qMjMEAUdHVmM
+BzP7PGw4mFgY/BSwxDjQBNUJ7e0IAAAABAEL/8gICAhIAFwAAAREjNQcGIyImJyY1ETMRFBceATMy
+Nj0BAgJWITlRT2EKBVYEBkA1RFECEv3uWj4qTToiOQE+/tIlJC43c4DpAAAAAAEAAQAAAfwCEgAGA
+AABAyMDMxsBAfzJaclfop8CEv3uAhL+LQHTAAABAAEAAAMLAhIADAAAAQMjCwEjAzMbATMbAQMLqW
+Z2dmapY3t0a3Z7AhL97gG+/kICEv5AAcD+QwG9AAAB//oAAAHWAhIACwAAARMjJwcjEwMzFzczARq
+8ZIuKY763ZoWFYwEO/vLV1QEMAQbNzQAAAQAB/y4B+wISABEAAAEDDgEjIic1FjMyNj8BAzMbAQH7
+2iFZQB8NDRIpNhQH02GenQIS/cFVUAJGASozEwIt/i4B0gABABQAAAGxAg4ACQAAJRUhNQEhNSEVA
+QGx/mMBNP7iAYL+zkREQgGIREX+ewAAAAABAED/sAEOAwwALAAAASMiBhUUFxYVFAYHHgEVFAcGFR
+QWOwEVIyImNTQ3NjU0JzU2NTQnJjU0NjsBAQ4MKiMLDS4pKS4NCyMqDAtERAwLUlILDERECwLUGBk
+WTlsgKzUFBTcrIFtOFhkYOC87GFVMIkUIOAhFIkxVGDsvAAAAAAEAYP84AJoDIAADAAAXIxEzmjo6
+yAPoAAEAIf+wAO8DDAAsAAATFQYVFBcWFRQGKwE1MzI2NTQnJjU0NjcuATU0NzY1NCYrATUzMhYVF
+AcGFRTvUgsMREQLDCojCw0uKSkuDQsjKgwLREQMCwF6OAhFIkxVGDsvOBgZFk5bICs1BQU3KyBbTh
+YZGDgvOxhVTCJFAAABAE0A3wH2AWQAEwAAATMUIyImJyYjIhUjNDMyFhcWMzIBvjhuGywtQR0xOG4
+bLC1BHTEBZIURGCNMhREYIwAAAwAk/94DIgLoAAcAEQApAAAAIBYQBiAmECQgBhUUFiA2NTQlMhYX
+IyYjIgYUFjMyNjczDgEjIiY1NDYBAQFE3d3+vN0CB/7wubkBELn+xVBnD1wSWDo+QTcqOQZcEmZWX
+HN2Aujg/rbg4AFKpr+Mjb6+jYxbWEldV5ZZNShLVn5na34AAgB4AFIB9AGeAAUACwAAAQcXIyc3Mw
+cXIyc3AUqJiUmJifOJiUmJiQGepqampqampqYAAAIAHAHSAQ4CwAAHAA8AABIyFhQGIiY0NiIGFBY
+yNjRgakREakSTNCEhNCECwEJqQkJqCiM4IyM4AAAAAAIAUAAAAfQCCwALAA8AAAEzFSMVIzUjNTM1
+MxMhNSEBP7W1OrW1OrX+XAGkAVs4tLQ4sP31OAAAAQB0AkQBAQKyAAMAABMjNzOsOD1QAkRuAAAAA
+AEAIADsAKoBdgAHAAASMhYUBiImNEg6KCg6KAF2KDooKDoAAAIAOQBSAbUBngAFAAsAACUHIzcnMw
+UHIzcnMwELiUmJiUkBM4lJiYlJ+KampqampqYAAAABADYB5QDhAt8ABAAAEzczByM2Xk1OXQHv8Po
+AAQAWAeUAwQLfAAQAABMHIzczwV5NTl0C1fD6AAIANgHlAYsC3wAEAAkAABM3MwcjPwEzByM2Xk1O
+XapeTU5dAe/w+grw+gAAAgAWAeUBawLfAAQACQAAEwcjNzMXByM3M8FeTU5dql5NTl0C1fD6CvD6A
+AADACX/8gI1AHIABwAPABcAADYyFhQGIiY0NjIWFAYiJjQ2MhYUBiImNEk4JCQ4JOw4JCQ4JOw4JC
+Q4JHIkOCQkOCQkOCQkOCQkOCQkOAAAAAEAeABSAUoBngAFAAABBxcjJzcBSomJSYmJAZ6mpqamAAA
+AAAEAOQBSAQsBngAFAAAlByM3JzMBC4lJiYlJ+KampgAAAf9qAAABgQKyAAMAACsBATM/VwHAVwKy
+AAAAAAIAFAHIAdwClAAHABQAABMVIxUjNSM1BRUjNwcjJxcjNTMXN9pKMkoByDICKzQqATJLKysCl
+CmjoykBy46KiY3Lm5sAAQAVAAABvALyABgAAAERIxEjESMRIzUzNTQ3NjMyFxUmBgcGHQEBvFbCVj
+k5AxHHHx5iVgcDAg798gHM/jQBzEIOJRuWBUcIJDAVIRYAAAABABX//AHkAvIAJQAAJR4BNxUGIyI
+mJyY1ESYjIgcGHQEzFSMRIxEjNTM1NDc2MzIXERQBowIcIxkkKi4CAR4nXgwDbW1WLy8DEbNdOmYa
+EQQ/BCQvEjQCFQZWFSEWQv40AcxCDiUblhP9uSEAAAAAAAAWAQ4AAQAAAAAAAAATACgAAQAAAAAAA
+QAHAEwAAQAAAAAAAgAHAGQAAQAAAAAAAwAaAKIAAQAAAAAABAAHAM0AAQAAAAAABQA8AU8AAQAAAA
+AABgAPAawAAQAAAAAACAALAdQAAQAAAAAACQALAfgAAQAAAAAACwAXAjQAAQAAAAAADAAXAnwAAwA
+BBAkAAAAmAAAAAwABBAkAAQAOADwAAwABBAkAAgAOAFQAAwABBAkAAwA0AGwAAwABBAkABAAOAL0A
+AwABBAkABQB4ANUAAwABBAkABgAeAYwAAwABBAkACAAWAbwAAwABBAkACQAWAeAAAwABBAkACwAuA
+gQAAwABBAkADAAuAkwATgBvACAAUgBpAGcAaAB0AHMAIABSAGUAcwBlAHIAdgBlAGQALgAATm8gUm
+lnaHRzIFJlc2VydmVkLgAAQQBpAGwAZQByAG8AbgAAQWlsZXJvbgAAUgBlAGcAdQBsAGEAcgAAUmV
+ndWxhcgAAMQAuADEAMAAyADsAVQBLAFcATgA7AEEAaQBsAGUAcgBvAG4ALQBSAGUAZwB1AGwAYQBy
+AAAxLjEwMjtVS1dOO0FpbGVyb24tUmVndWxhcgAAQQBpAGwAZQByAG8AbgAAQWlsZXJvbgAAVgBlA
+HIAcwBpAG8AbgAgADEALgAxADAAMgA7AFAAUwAgADAAMAAxAC4AMQAwADIAOwBoAG8AdABjAG8Abg
+B2ACAAMQAuADAALgA3ADAAOwBtAGEAawBlAG8AdABmAC4AbABpAGIAMgAuADUALgA1ADgAMwAyADk
+AAFZlcnNpb24gMS4xMDI7UFMgMDAxLjEwMjtob3Rjb252IDEuMC43MDttYWtlb3RmLmxpYjIuNS41
+ODMyOQAAQQBpAGwAZQByAG8AbgAtAFIAZQBnAHUAbABhAHIAAEFpbGVyb24tUmVndWxhcgAAUwBvA
+HIAYQAgAFMAYQBnAGEAbgBvAABTb3JhIFNhZ2FubwAAUwBvAHIAYQAgAFMAYQBnAGEAbgBvAABTb3
+JhIFNhZ2FubwAAaAB0AHQAcAA6AC8ALwB3AHcAdwAuAGQAbwB0AGMAbwBsAG8AbgAuAG4AZQB0AAB
+odHRwOi8vd3d3LmRvdGNvbG9uLm5ldAAAaAB0AHQAcAA6AC8ALwB3AHcAdwAuAGQAbwB0AGMAbwBs
+AG8AbgAuAG4AZQB0AABodHRwOi8vd3d3LmRvdGNvbG9uLm5ldAAAAAACAAAAAAAA/4MAMgAAAAAAA
+AAAAAAAAAAAAAAAAAAAAHQAAAABAAIAAwAEAAUABgAHAAgACQAKAAsADAANAA4ADwAQABEAEgATAB
+QAFQAWABcAGAAZABoAGwAcAB0AHgAfACAAIQAiACMAJAAlACYAJwAoACkAKgArACwALQAuAC8AMAA
+xADIAMwA0ADUANgA3ADgAOQA6ADsAPAA9AD4APwBAAEEAQgBDAEQARQBGAEcASABJAEoASwBMAE0A
+TgBPAFAAUQBSAFMAVABVAFYAVwBYAFkAWgBbAFwAXQBeAF8AYABhAIsAqQCDAJMAjQDDAKoAtgC3A
+LQAtQCrAL4AvwC8AIwAwADBAAAAAAAB//8AAgABAAAADAAAABwAAAACAAIAAwBxAAEAcgBzAAIABA
+AAAAIAAAABAAAACgBMAGYAAkRGTFQADmxhdG4AGgAEAAAAAP//AAEAAAAWAANDQVQgAB5NT0wgABZ
+ST00gABYAAP//AAEAAAAA//8AAgAAAAEAAmxpZ2EADmxvY2wAFAAAAAEAAQAAAAEAAAACAAYAEAAG
+AAAAAgASADQABAAAAAEATAADAAAAAgAQABYAAQAcAAAAAQABAE8AAQABAGcAAQABAE8AAwAAAAIAE
+AAWAAEAHAAAAAEAAQAvAAEAAQBnAAEAAQAvAAEAGgABAAgAAgAGAAwAcwACAE8AcgACAEwAAQABAE
+kAAAABAAAACgBGAGAAAkRGTFQADmxhdG4AHAAEAAAAAP//AAIAAAABABYAA0NBVCAAFk1PTCAAFlJ
+PTSAAFgAA//8AAgAAAAEAAmNwc3AADmtlcm4AFAAAAAEAAAAAAAEAAQACAAYADgABAAAAAQASAAIA
+AAACAB4ANgABAAoABQAFAAoAAgABACQAPQAAAAEAEgAEAAAAAQAMAAEAOP/nAAEAAQAkAAIGigAEA
+AAFJAXKABoAGQAA//gAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAD/sv+4/+z/7v/MAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAD/xAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA/9T/6AAAAAD/8QAA
+ABD/vQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/7gAAAAAAAAAAAAAAAAAA//MAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABIAAAAAAAAAAP/5AAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAP/gAAD/4AAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA//L/9AAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAA/+gAAAAAAAkAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAP/zAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAP/mAAAAAAAAAAAAAAAAAAD
+/4gAA//AAAAAA//YAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/+AAAAAAAAP/OAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/zv/qAAAAAP/0AAAACAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAP/ZAAD/egAA/1kAAAAA/5D/rgAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAD/9AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAD/8AAA/7b/8P+wAAD/8P/E/98AAAAA/8P/+P/0//oAAAAAAAAAAAAA//gA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA/+AAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/w//C/9MAAP/SAAD/9wAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAD/yAAA/+kAAAAA//QAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/9wAAAAD//QAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAP/2AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAP/cAAAAAAAAAAAAAAAA/7YAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAP/8AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD/6AAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAQAkAFAAEAAAAAQACwAAABcA
+BgAAAAAAAAAIAA4AAAAAAAsAEgAAAAAAAAATABkAAwANAAAAAQAJAAAAAAAAAAAAAAAAAAAAGAAAA
+AAABwAAAAAAAAAAAAAAFQAFAAAAAAAYABgAAAAUAAAACgAAAAwAAgAPABEAFgAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAFAAEAEQBdAAYAAAAAAAAAAAAAAAAAAAAAAAA
+AAAAAAAAAAAAAAAAAAAAAAAAAAQAAAAcAAAAAAAAABwAAAAAACAAAAAAAAAAAAAcAAAAHAAAAEwAJ
+ABUADgAPAAAACwAQAAAAAAAAAAAAAAAAAAUAGAACAAIAAgAAAAIAGAAXAAAAGAAAABYAFgACABYAA
+gAWAAAAEQADAAoAFAAMAA0ABAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAASAAAAEgAGAAEAHgAkAC
+YAJwApACoALQAuAC8AMgAzADcAOAA5ADoAPAA9AEUASABOAE8AUgBTAFUAVwBZAFoAWwBcAF0AcwA
+AAAAAAQAAAADa3tfFAAAAANAan9kAAAAA4QodoQ==
+"""
+                )
+            ),
+            10 if size is None else size,
+            layout_engine=Layout.BASIC,
+        )
+    return load_default_imagefont()
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageGrab.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageGrab.py
new file mode 100644
index 00000000..fe27bfae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageGrab.py
@@ -0,0 +1,183 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# screen grabber
+#
+# History:
+# 2001-04-26 fl  created
+# 2001-09-17 fl  use builtin driver, if present
+# 2002-11-19 fl  added grabclipboard support
+#
+# Copyright (c) 2001-2002 by Secret Labs AB
+# Copyright (c) 2001-2002 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+
+from . import Image
+
+
+def grab(
+    bbox: tuple[int, int, int, int] | None = None,
+    include_layered_windows: bool = False,
+    all_screens: bool = False,
+    xdisplay: str | None = None,
+) -> Image.Image:
+    im: Image.Image
+    if xdisplay is None:
+        if sys.platform == "darwin":
+            fh, filepath = tempfile.mkstemp(".png")
+            os.close(fh)
+            args = ["screencapture"]
+            if bbox:
+                left, top, right, bottom = bbox
+                args += ["-R", f"{left},{top},{right-left},{bottom-top}"]
+            subprocess.call(args + ["-x", filepath])
+            im = Image.open(filepath)
+            im.load()
+            os.unlink(filepath)
+            if bbox:
+                im_resized = im.resize((right - left, bottom - top))
+                im.close()
+                return im_resized
+            return im
+        elif sys.platform == "win32":
+            offset, size, data = Image.core.grabscreen_win32(
+                include_layered_windows, all_screens
+            )
+            im = Image.frombytes(
+                "RGB",
+                size,
+                data,
+                # RGB, 32-bit line padding, origin lower left corner
+                "raw",
+                "BGR",
+                (size[0] * 3 + 3) & -4,
+                -1,
+            )
+            if bbox:
+                x0, y0 = offset
+                left, top, right, bottom = bbox
+                im = im.crop((left - x0, top - y0, right - x0, bottom - y0))
+            return im
+    # Cast to Optional[str] needed for Windows and macOS.
+    display_name: str | None = xdisplay
+    try:
+        if not Image.core.HAVE_XCB:
+            msg = "Pillow was built without XCB support"
+            raise OSError(msg)
+        size, data = Image.core.grabscreen_x11(display_name)
+    except OSError:
+        if (
+            display_name is None
+            and sys.platform not in ("darwin", "win32")
+            and shutil.which("gnome-screenshot")
+        ):
+            fh, filepath = tempfile.mkstemp(".png")
+            os.close(fh)
+            subprocess.call(["gnome-screenshot", "-f", filepath])
+            im = Image.open(filepath)
+            im.load()
+            os.unlink(filepath)
+            if bbox:
+                im_cropped = im.crop(bbox)
+                im.close()
+                return im_cropped
+            return im
+        else:
+            raise
+    else:
+        im = Image.frombytes("RGB", size, data, "raw", "BGRX", size[0] * 4, 1)
+        if bbox:
+            im = im.crop(bbox)
+        return im
+
+
+def grabclipboard() -> Image.Image | list[str] | None:
+    if sys.platform == "darwin":
+        p = subprocess.run(
+            ["osascript", "-e", "get the clipboard as «class PNGf»"],
+            capture_output=True,
+        )
+        if p.returncode != 0:
+            return None
+
+        import binascii
+
+        data = io.BytesIO(binascii.unhexlify(p.stdout[11:-3]))
+        return Image.open(data)
+    elif sys.platform == "win32":
+        fmt, data = Image.core.grabclipboard_win32()
+        if fmt == "file":  # CF_HDROP
+            import struct
+
+            o = struct.unpack_from("I", data)[0]
+            if data[16] != 0:
+                files = data[o:].decode("utf-16le").split("\0")
+            else:
+                files = data[o:].decode("mbcs").split("\0")
+            return files[: files.index("")]
+        if isinstance(data, bytes):
+            data = io.BytesIO(data)
+            if fmt == "png":
+                from . import PngImagePlugin
+
+                return PngImagePlugin.PngImageFile(data)
+            elif fmt == "DIB":
+                from . import BmpImagePlugin
+
+                return BmpImagePlugin.DibImageFile(data)
+        return None
+    else:
+        if os.getenv("WAYLAND_DISPLAY"):
+            session_type = "wayland"
+        elif os.getenv("DISPLAY"):
+            session_type = "x11"
+        else:  # Session type check failed
+            session_type = None
+
+        if shutil.which("wl-paste") and session_type in ("wayland", None):
+            args = ["wl-paste", "-t", "image"]
+        elif shutil.which("xclip") and session_type in ("x11", None):
+            args = ["xclip", "-selection", "clipboard", "-t", "image/png", "-o"]
+        else:
+            msg = "wl-paste or xclip is required for ImageGrab.grabclipboard() on Linux"
+            raise NotImplementedError(msg)
+
+        p = subprocess.run(args, capture_output=True)
+        if p.returncode != 0:
+            err = p.stderr
+            for silent_error in [
+                # wl-paste, when the clipboard is empty
+                b"Nothing is copied",
+                # Ubuntu/Debian wl-paste, when the clipboard is empty
+                b"No selection",
+                # Ubuntu/Debian wl-paste, when an image isn't available
+                b"No suitable type of content copied",
+                # wl-paste or Ubuntu/Debian xclip, when an image isn't available
+                b" not available",
+                # xclip, when an image isn't available
+                b"cannot convert ",
+                # xclip, when the clipboard isn't initialized
+                b"xclip: Error: There is no owner for the ",
+            ]:
+                if silent_error in err:
+                    return None
+            msg = f"{args[0]} error"
+            if err:
+                msg += f": {err.strip().decode()}"
+            raise ChildProcessError(msg)
+
+        data = io.BytesIO(p.stdout)
+        im = Image.open(data)
+        im.load()
+        return im
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageMath.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageMath.py
new file mode 100644
index 00000000..484797f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageMath.py
@@ -0,0 +1,368 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# a simple math add-on for the Python Imaging Library
+#
+# History:
+# 1999-02-15 fl   Original PIL Plus release
+# 2005-05-05 fl   Simplified and cleaned up for PIL 1.1.6
+# 2005-09-12 fl   Fixed int() and float() for Python 2.4.1
+#
+# Copyright (c) 1999-2005 by Secret Labs AB
+# Copyright (c) 2005 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import builtins
+from types import CodeType
+from typing import Any, Callable
+
+from . import Image, _imagingmath
+from ._deprecate import deprecate
+
+
+class _Operand:
+    """Wraps an image operand, providing standard operators"""
+
+    def __init__(self, im: Image.Image):
+        self.im = im
+
+    def __fixup(self, im1: _Operand | float) -> Image.Image:
+        # convert image to suitable mode
+        if isinstance(im1, _Operand):
+            # argument was an image.
+            if im1.im.mode in ("1", "L"):
+                return im1.im.convert("I")
+            elif im1.im.mode in ("I", "F"):
+                return im1.im
+            else:
+                msg = f"unsupported mode: {im1.im.mode}"
+                raise ValueError(msg)
+        else:
+            # argument was a constant
+            if isinstance(im1, (int, float)) and self.im.mode in ("1", "L", "I"):
+                return Image.new("I", self.im.size, im1)
+            else:
+                return Image.new("F", self.im.size, im1)
+
+    def apply(
+        self,
+        op: str,
+        im1: _Operand | float,
+        im2: _Operand | float | None = None,
+        mode: str | None = None,
+    ) -> _Operand:
+        im_1 = self.__fixup(im1)
+        if im2 is None:
+            # unary operation
+            out = Image.new(mode or im_1.mode, im_1.size, None)
+            try:
+                op = getattr(_imagingmath, f"{op}_{im_1.mode}")
+            except AttributeError as e:
+                msg = f"bad operand type for '{op}'"
+                raise TypeError(msg) from e
+            _imagingmath.unop(op, out.getim(), im_1.getim())
+        else:
+            # binary operation
+            im_2 = self.__fixup(im2)
+            if im_1.mode != im_2.mode:
+                # convert both arguments to floating point
+                if im_1.mode != "F":
+                    im_1 = im_1.convert("F")
+                if im_2.mode != "F":
+                    im_2 = im_2.convert("F")
+            if im_1.size != im_2.size:
+                # crop both arguments to a common size
+                size = (
+                    min(im_1.size[0], im_2.size[0]),
+                    min(im_1.size[1], im_2.size[1]),
+                )
+                if im_1.size != size:
+                    im_1 = im_1.crop((0, 0) + size)
+                if im_2.size != size:
+                    im_2 = im_2.crop((0, 0) + size)
+            out = Image.new(mode or im_1.mode, im_1.size, None)
+            try:
+                op = getattr(_imagingmath, f"{op}_{im_1.mode}")
+            except AttributeError as e:
+                msg = f"bad operand type for '{op}'"
+                raise TypeError(msg) from e
+            _imagingmath.binop(op, out.getim(), im_1.getim(), im_2.getim())
+        return _Operand(out)
+
+    # unary operators
+    def __bool__(self) -> bool:
+        # an image is "true" if it contains at least one non-zero pixel
+        return self.im.getbbox() is not None
+
+    def __abs__(self) -> _Operand:
+        return self.apply("abs", self)
+
+    def __pos__(self) -> _Operand:
+        return self
+
+    def __neg__(self) -> _Operand:
+        return self.apply("neg", self)
+
+    # binary operators
+    def __add__(self, other: _Operand | float) -> _Operand:
+        return self.apply("add", self, other)
+
+    def __radd__(self, other: _Operand | float) -> _Operand:
+        return self.apply("add", other, self)
+
+    def __sub__(self, other: _Operand | float) -> _Operand:
+        return self.apply("sub", self, other)
+
+    def __rsub__(self, other: _Operand | float) -> _Operand:
+        return self.apply("sub", other, self)
+
+    def __mul__(self, other: _Operand | float) -> _Operand:
+        return self.apply("mul", self, other)
+
+    def __rmul__(self, other: _Operand | float) -> _Operand:
+        return self.apply("mul", other, self)
+
+    def __truediv__(self, other: _Operand | float) -> _Operand:
+        return self.apply("div", self, other)
+
+    def __rtruediv__(self, other: _Operand | float) -> _Operand:
+        return self.apply("div", other, self)
+
+    def __mod__(self, other: _Operand | float) -> _Operand:
+        return self.apply("mod", self, other)
+
+    def __rmod__(self, other: _Operand | float) -> _Operand:
+        return self.apply("mod", other, self)
+
+    def __pow__(self, other: _Operand | float) -> _Operand:
+        return self.apply("pow", self, other)
+
+    def __rpow__(self, other: _Operand | float) -> _Operand:
+        return self.apply("pow", other, self)
+
+    # bitwise
+    def __invert__(self) -> _Operand:
+        return self.apply("invert", self)
+
+    def __and__(self, other: _Operand | float) -> _Operand:
+        return self.apply("and", self, other)
+
+    def __rand__(self, other: _Operand | float) -> _Operand:
+        return self.apply("and", other, self)
+
+    def __or__(self, other: _Operand | float) -> _Operand:
+        return self.apply("or", self, other)
+
+    def __ror__(self, other: _Operand | float) -> _Operand:
+        return self.apply("or", other, self)
+
+    def __xor__(self, other: _Operand | float) -> _Operand:
+        return self.apply("xor", self, other)
+
+    def __rxor__(self, other: _Operand | float) -> _Operand:
+        return self.apply("xor", other, self)
+
+    def __lshift__(self, other: _Operand | float) -> _Operand:
+        return self.apply("lshift", self, other)
+
+    def __rshift__(self, other: _Operand | float) -> _Operand:
+        return self.apply("rshift", self, other)
+
+    # logical
+    def __eq__(self, other: _Operand | float) -> _Operand:  # type: ignore[override]
+        return self.apply("eq", self, other)
+
+    def __ne__(self, other: _Operand | float) -> _Operand:  # type: ignore[override]
+        return self.apply("ne", self, other)
+
+    def __lt__(self, other: _Operand | float) -> _Operand:
+        return self.apply("lt", self, other)
+
+    def __le__(self, other: _Operand | float) -> _Operand:
+        return self.apply("le", self, other)
+
+    def __gt__(self, other: _Operand | float) -> _Operand:
+        return self.apply("gt", self, other)
+
+    def __ge__(self, other: _Operand | float) -> _Operand:
+        return self.apply("ge", self, other)
+
+
+# conversions
+def imagemath_int(self: _Operand) -> _Operand:
+    return _Operand(self.im.convert("I"))
+
+
+def imagemath_float(self: _Operand) -> _Operand:
+    return _Operand(self.im.convert("F"))
+
+
+# logical
+def imagemath_equal(self: _Operand, other: _Operand | float | None) -> _Operand:
+    return self.apply("eq", self, other, mode="I")
+
+
+def imagemath_notequal(self: _Operand, other: _Operand | float | None) -> _Operand:
+    return self.apply("ne", self, other, mode="I")
+
+
+def imagemath_min(self: _Operand, other: _Operand | float | None) -> _Operand:
+    return self.apply("min", self, other)
+
+
+def imagemath_max(self: _Operand, other: _Operand | float | None) -> _Operand:
+    return self.apply("max", self, other)
+
+
+def imagemath_convert(self: _Operand, mode: str) -> _Operand:
+    return _Operand(self.im.convert(mode))
+
+
+ops = {
+    "int": imagemath_int,
+    "float": imagemath_float,
+    "equal": imagemath_equal,
+    "notequal": imagemath_notequal,
+    "min": imagemath_min,
+    "max": imagemath_max,
+    "convert": imagemath_convert,
+}
+
+
+def lambda_eval(
+    expression: Callable[[dict[str, Any]], Any],
+    options: dict[str, Any] = {},
+    **kw: Any,
+) -> Any:
+    """
+    Returns the result of an image function.
+
+    :py:mod:`~PIL.ImageMath` only supports single-layer images. To process multi-band
+    images, use the :py:meth:`~PIL.Image.Image.split` method or
+    :py:func:`~PIL.Image.merge` function.
+
+    :param expression: A function that receives a dictionary.
+    :param options: Values to add to the function's dictionary. Deprecated.
+                    You can instead use one or more keyword arguments.
+    :param **kw: Values to add to the function's dictionary.
+    :return: The expression result. This is usually an image object, but can
+             also be an integer, a floating point value, or a pixel tuple,
+             depending on the expression.
+    """
+
+    if options:
+        deprecate(
+            "ImageMath.lambda_eval options",
+            12,
+            "ImageMath.lambda_eval keyword arguments",
+        )
+
+    args: dict[str, Any] = ops.copy()
+    args.update(options)
+    args.update(kw)
+    for k, v in args.items():
+        if isinstance(v, Image.Image):
+            args[k] = _Operand(v)
+
+    out = expression(args)
+    try:
+        return out.im
+    except AttributeError:
+        return out
+
+
+def unsafe_eval(
+    expression: str,
+    options: dict[str, Any] = {},
+    **kw: Any,
+) -> Any:
+    """
+    Evaluates an image expression. This uses Python's ``eval()`` function to process
+    the expression string, and carries the security risks of doing so. It is not
+    recommended to process expressions without considering this.
+    :py:meth:`~lambda_eval` is a more secure alternative.
+
+    :py:mod:`~PIL.ImageMath` only supports single-layer images. To process multi-band
+    images, use the :py:meth:`~PIL.Image.Image.split` method or
+    :py:func:`~PIL.Image.merge` function.
+
+    :param expression: A string containing a Python-style expression.
+    :param options: Values to add to the evaluation context. Deprecated.
+                    You can instead use one or more keyword arguments.
+    :param **kw: Values to add to the evaluation context.
+    :return: The evaluated expression. This is usually an image object, but can
+             also be an integer, a floating point value, or a pixel tuple,
+             depending on the expression.
+    """
+
+    if options:
+        deprecate(
+            "ImageMath.unsafe_eval options",
+            12,
+            "ImageMath.unsafe_eval keyword arguments",
+        )
+
+    # build execution namespace
+    args: dict[str, Any] = ops.copy()
+    for k in list(options.keys()) + list(kw.keys()):
+        if "__" in k or hasattr(builtins, k):
+            msg = f"'{k}' not allowed"
+            raise ValueError(msg)
+
+    args.update(options)
+    args.update(kw)
+    for k, v in args.items():
+        if isinstance(v, Image.Image):
+            args[k] = _Operand(v)
+
+    compiled_code = compile(expression, "<string>", "eval")
+
+    def scan(code: CodeType) -> None:
+        for const in code.co_consts:
+            if type(const) is type(compiled_code):
+                scan(const)
+
+        for name in code.co_names:
+            if name not in args and name != "abs":
+                msg = f"'{name}' not allowed"
+                raise ValueError(msg)
+
+    scan(compiled_code)
+    out = builtins.eval(expression, {"__builtins": {"abs": abs}}, args)
+    try:
+        return out.im
+    except AttributeError:
+        return out
+
+
+def eval(
+    expression: str,
+    _dict: dict[str, Any] = {},
+    **kw: Any,
+) -> Any:
+    """
+    Evaluates an image expression.
+
+    Deprecated. Use lambda_eval() or unsafe_eval() instead.
+
+    :param expression: A string containing a Python-style expression.
+    :param _dict: Values to add to the evaluation context.  You
+                  can either use a dictionary, or one or more keyword
+                  arguments.
+    :return: The evaluated expression. This is usually an image object, but can
+             also be an integer, a floating point value, or a pixel tuple,
+             depending on the expression.
+
+    ..  deprecated:: 10.3.0
+    """
+
+    deprecate(
+        "ImageMath.eval",
+        12,
+        "ImageMath.lambda_eval or ImageMath.unsafe_eval",
+    )
+    return unsafe_eval(expression, _dict, **kw)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageMode.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageMode.py
new file mode 100644
index 00000000..92a08d2c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageMode.py
@@ -0,0 +1,92 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# standard mode descriptors
+#
+# History:
+# 2006-03-20 fl   Added
+#
+# Copyright (c) 2006 by Secret Labs AB.
+# Copyright (c) 2006 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import sys
+from functools import lru_cache
+from typing import NamedTuple
+
+from ._deprecate import deprecate
+
+
+class ModeDescriptor(NamedTuple):
+    """Wrapper for mode strings."""
+
+    mode: str
+    bands: tuple[str, ...]
+    basemode: str
+    basetype: str
+    typestr: str
+
+    def __str__(self) -> str:
+        return self.mode
+
+
+@lru_cache
+def getmode(mode: str) -> ModeDescriptor:
+    """Gets a mode descriptor for the given mode."""
+    endian = "<" if sys.byteorder == "little" else ">"
+
+    modes = {
+        # core modes
+        # Bits need to be extended to bytes
+        "1": ("L", "L", ("1",), "|b1"),
+        "L": ("L", "L", ("L",), "|u1"),
+        "I": ("L", "I", ("I",), f"{endian}i4"),
+        "F": ("L", "F", ("F",), f"{endian}f4"),
+        "P": ("P", "L", ("P",), "|u1"),
+        "RGB": ("RGB", "L", ("R", "G", "B"), "|u1"),
+        "RGBX": ("RGB", "L", ("R", "G", "B", "X"), "|u1"),
+        "RGBA": ("RGB", "L", ("R", "G", "B", "A"), "|u1"),
+        "CMYK": ("RGB", "L", ("C", "M", "Y", "K"), "|u1"),
+        "YCbCr": ("RGB", "L", ("Y", "Cb", "Cr"), "|u1"),
+        # UNDONE - unsigned |u1i1i1
+        "LAB": ("RGB", "L", ("L", "A", "B"), "|u1"),
+        "HSV": ("RGB", "L", ("H", "S", "V"), "|u1"),
+        # extra experimental modes
+        "RGBa": ("RGB", "L", ("R", "G", "B", "a"), "|u1"),
+        "BGR;15": ("RGB", "L", ("B", "G", "R"), "|u1"),
+        "BGR;16": ("RGB", "L", ("B", "G", "R"), "|u1"),
+        "BGR;24": ("RGB", "L", ("B", "G", "R"), "|u1"),
+        "LA": ("L", "L", ("L", "A"), "|u1"),
+        "La": ("L", "L", ("L", "a"), "|u1"),
+        "PA": ("RGB", "L", ("P", "A"), "|u1"),
+    }
+    if mode in modes:
+        if mode in ("BGR;15", "BGR;16", "BGR;24"):
+            deprecate(mode, 12)
+        base_mode, base_type, bands, type_str = modes[mode]
+        return ModeDescriptor(mode, bands, base_mode, base_type, type_str)
+
+    mapping_modes = {
+        # I;16 == I;16L, and I;32 == I;32L
+        "I;16": "<u2",
+        "I;16S": "<i2",
+        "I;16L": "<u2",
+        "I;16LS": "<i2",
+        "I;16B": ">u2",
+        "I;16BS": ">i2",
+        "I;16N": f"{endian}u2",
+        "I;16NS": f"{endian}i2",
+        "I;32": "<u4",
+        "I;32B": ">u4",
+        "I;32L": "<u4",
+        "I;32S": "<i4",
+        "I;32BS": ">i4",
+        "I;32LS": "<i4",
+    }
+
+    type_str = mapping_modes[mode]
+    return ModeDescriptor(mode, ("I",), "L", "L", type_str)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageMorph.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageMorph.py
new file mode 100644
index 00000000..f0a066b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageMorph.py
@@ -0,0 +1,265 @@
+# A binary morphology add-on for the Python Imaging Library
+#
+# History:
+#   2014-06-04 Initial version.
+#
+# Copyright (c) 2014 Dov Grobgeld <dov.grobgeld@gmail.com>
+from __future__ import annotations
+
+import re
+
+from . import Image, _imagingmorph
+
+LUT_SIZE = 1 << 9
+
+# fmt: off
+ROTATION_MATRIX = [
+    6, 3, 0,
+    7, 4, 1,
+    8, 5, 2,
+]
+MIRROR_MATRIX = [
+    2, 1, 0,
+    5, 4, 3,
+    8, 7, 6,
+]
+# fmt: on
+
+
+class LutBuilder:
+    """A class for building a MorphLut from a descriptive language
+
+    The input patterns is a list of a strings sequences like these::
+
+        4:(...
+           .1.
+           111)->1
+
+    (whitespaces including linebreaks are ignored). The option 4
+    describes a series of symmetry operations (in this case a
+    4-rotation), the pattern is described by:
+
+    - . or X - Ignore
+    - 1 - Pixel is on
+    - 0 - Pixel is off
+
+    The result of the operation is described after "->" string.
+
+    The default is to return the current pixel value, which is
+    returned if no other match is found.
+
+    Operations:
+
+    - 4 - 4 way rotation
+    - N - Negate
+    - 1 - Dummy op for no other operation (an op must always be given)
+    - M - Mirroring
+
+    Example::
+
+        lb = LutBuilder(patterns = ["4:(... .1. 111)->1"])
+        lut = lb.build_lut()
+
+    """
+
+    def __init__(
+        self, patterns: list[str] | None = None, op_name: str | None = None
+    ) -> None:
+        if patterns is not None:
+            self.patterns = patterns
+        else:
+            self.patterns = []
+        self.lut: bytearray | None = None
+        if op_name is not None:
+            known_patterns = {
+                "corner": ["1:(... ... ...)->0", "4:(00. 01. ...)->1"],
+                "dilation4": ["4:(... .0. .1.)->1"],
+                "dilation8": ["4:(... .0. .1.)->1", "4:(... .0. ..1)->1"],
+                "erosion4": ["4:(... .1. .0.)->0"],
+                "erosion8": ["4:(... .1. .0.)->0", "4:(... .1. ..0)->0"],
+                "edge": [
+                    "1:(... ... ...)->0",
+                    "4:(.0. .1. ...)->1",
+                    "4:(01. .1. ...)->1",
+                ],
+            }
+            if op_name not in known_patterns:
+                msg = f"Unknown pattern {op_name}!"
+                raise Exception(msg)
+
+            self.patterns = known_patterns[op_name]
+
+    def add_patterns(self, patterns: list[str]) -> None:
+        self.patterns += patterns
+
+    def build_default_lut(self) -> None:
+        symbols = [0, 1]
+        m = 1 << 4  # pos of current pixel
+        self.lut = bytearray(symbols[(i & m) > 0] for i in range(LUT_SIZE))
+
+    def get_lut(self) -> bytearray | None:
+        return self.lut
+
+    def _string_permute(self, pattern: str, permutation: list[int]) -> str:
+        """string_permute takes a pattern and a permutation and returns the
+        string permuted according to the permutation list.
+        """
+        assert len(permutation) == 9
+        return "".join(pattern[p] for p in permutation)
+
+    def _pattern_permute(
+        self, basic_pattern: str, options: str, basic_result: int
+    ) -> list[tuple[str, int]]:
+        """pattern_permute takes a basic pattern and its result and clones
+        the pattern according to the modifications described in the $options
+        parameter. It returns a list of all cloned patterns."""
+        patterns = [(basic_pattern, basic_result)]
+
+        # rotations
+        if "4" in options:
+            res = patterns[-1][1]
+            for i in range(4):
+                patterns.append(
+                    (self._string_permute(patterns[-1][0], ROTATION_MATRIX), res)
+                )
+        # mirror
+        if "M" in options:
+            n = len(patterns)
+            for pattern, res in patterns[:n]:
+                patterns.append((self._string_permute(pattern, MIRROR_MATRIX), res))
+
+        # negate
+        if "N" in options:
+            n = len(patterns)
+            for pattern, res in patterns[:n]:
+                # Swap 0 and 1
+                pattern = pattern.replace("0", "Z").replace("1", "0").replace("Z", "1")
+                res = 1 - int(res)
+                patterns.append((pattern, res))
+
+        return patterns
+
+    def build_lut(self) -> bytearray:
+        """Compile all patterns into a morphology lut.
+
+        TBD :Build based on (file) morphlut:modify_lut
+        """
+        self.build_default_lut()
+        assert self.lut is not None
+        patterns = []
+
+        # Parse and create symmetries of the patterns strings
+        for p in self.patterns:
+            m = re.search(r"(\w*):?\s*\((.+?)\)\s*->\s*(\d)", p.replace("\n", ""))
+            if not m:
+                msg = 'Syntax error in pattern "' + p + '"'
+                raise Exception(msg)
+            options = m.group(1)
+            pattern = m.group(2)
+            result = int(m.group(3))
+
+            # Get rid of spaces
+            pattern = pattern.replace(" ", "").replace("\n", "")
+
+            patterns += self._pattern_permute(pattern, options, result)
+
+        # compile the patterns into regular expressions for speed
+        compiled_patterns = []
+        for pattern in patterns:
+            p = pattern[0].replace(".", "X").replace("X", "[01]")
+            compiled_patterns.append((re.compile(p), pattern[1]))
+
+        # Step through table and find patterns that match.
+        # Note that all the patterns are searched. The last one
+        # caught overrides
+        for i in range(LUT_SIZE):
+            # Build the bit pattern
+            bitpattern = bin(i)[2:]
+            bitpattern = ("0" * (9 - len(bitpattern)) + bitpattern)[::-1]
+
+            for pattern, r in compiled_patterns:
+                if pattern.match(bitpattern):
+                    self.lut[i] = [0, 1][r]
+
+        return self.lut
+
+
+class MorphOp:
+    """A class for binary morphological operators"""
+
+    def __init__(
+        self,
+        lut: bytearray | None = None,
+        op_name: str | None = None,
+        patterns: list[str] | None = None,
+    ) -> None:
+        """Create a binary morphological operator"""
+        self.lut = lut
+        if op_name is not None:
+            self.lut = LutBuilder(op_name=op_name).build_lut()
+        elif patterns is not None:
+            self.lut = LutBuilder(patterns=patterns).build_lut()
+
+    def apply(self, image: Image.Image) -> tuple[int, Image.Image]:
+        """Run a single morphological operation on an image
+
+        Returns a tuple of the number of changed pixels and the
+        morphed image"""
+        if self.lut is None:
+            msg = "No operator loaded"
+            raise Exception(msg)
+
+        if image.mode != "L":
+            msg = "Image mode must be L"
+            raise ValueError(msg)
+        outimage = Image.new(image.mode, image.size, None)
+        count = _imagingmorph.apply(bytes(self.lut), image.getim(), outimage.getim())
+        return count, outimage
+
+    def match(self, image: Image.Image) -> list[tuple[int, int]]:
+        """Get a list of coordinates matching the morphological operation on
+        an image.
+
+        Returns a list of tuples of (x,y) coordinates
+        of all matching pixels. See :ref:`coordinate-system`."""
+        if self.lut is None:
+            msg = "No operator loaded"
+            raise Exception(msg)
+
+        if image.mode != "L":
+            msg = "Image mode must be L"
+            raise ValueError(msg)
+        return _imagingmorph.match(bytes(self.lut), image.getim())
+
+    def get_on_pixels(self, image: Image.Image) -> list[tuple[int, int]]:
+        """Get a list of all turned on pixels in a binary image
+
+        Returns a list of tuples of (x,y) coordinates
+        of all matching pixels. See :ref:`coordinate-system`."""
+
+        if image.mode != "L":
+            msg = "Image mode must be L"
+            raise ValueError(msg)
+        return _imagingmorph.get_on_pixels(image.getim())
+
+    def load_lut(self, filename: str) -> None:
+        """Load an operator from an mrl file"""
+        with open(filename, "rb") as f:
+            self.lut = bytearray(f.read())
+
+        if len(self.lut) != LUT_SIZE:
+            self.lut = None
+            msg = "Wrong size operator file!"
+            raise Exception(msg)
+
+    def save_lut(self, filename: str) -> None:
+        """Save an operator to an mrl file"""
+        if self.lut is None:
+            msg = "No operator loaded"
+            raise Exception(msg)
+        with open(filename, "wb") as f:
+            f.write(self.lut)
+
+    def set_lut(self, lut: bytearray | None) -> None:
+        """Set the lut from an external source"""
+        self.lut = lut
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageOps.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageOps.py
new file mode 100644
index 00000000..bb29cc0d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageOps.py
@@ -0,0 +1,731 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# standard image operations
+#
+# History:
+# 2001-10-20 fl   Created
+# 2001-10-23 fl   Added autocontrast operator
+# 2001-12-18 fl   Added Kevin's fit operator
+# 2004-03-14 fl   Fixed potential division by zero in equalize
+# 2005-05-05 fl   Fixed equalize for low number of values
+#
+# Copyright (c) 2001-2004 by Secret Labs AB
+# Copyright (c) 2001-2004 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import functools
+import operator
+import re
+from collections.abc import Sequence
+from typing import Protocol, cast
+
+from . import ExifTags, Image, ImagePalette
+
+#
+# helpers
+
+
+def _border(border: int | tuple[int, ...]) -> tuple[int, int, int, int]:
+    if isinstance(border, tuple):
+        if len(border) == 2:
+            left, top = right, bottom = border
+        elif len(border) == 4:
+            left, top, right, bottom = border
+    else:
+        left = top = right = bottom = border
+    return left, top, right, bottom
+
+
+def _color(color: str | int | tuple[int, ...], mode: str) -> int | tuple[int, ...]:
+    if isinstance(color, str):
+        from . import ImageColor
+
+        color = ImageColor.getcolor(color, mode)
+    return color
+
+
+def _lut(image: Image.Image, lut: list[int]) -> Image.Image:
+    if image.mode == "P":
+        # FIXME: apply to lookup table, not image data
+        msg = "mode P support coming soon"
+        raise NotImplementedError(msg)
+    elif image.mode in ("L", "RGB"):
+        if image.mode == "RGB" and len(lut) == 256:
+            lut = lut + lut + lut
+        return image.point(lut)
+    else:
+        msg = f"not supported for mode {image.mode}"
+        raise OSError(msg)
+
+
+#
+# actions
+
+
+def autocontrast(
+    image: Image.Image,
+    cutoff: float | tuple[float, float] = 0,
+    ignore: int | Sequence[int] | None = None,
+    mask: Image.Image | None = None,
+    preserve_tone: bool = False,
+) -> Image.Image:
+    """
+    Maximize (normalize) image contrast. This function calculates a
+    histogram of the input image (or mask region), removes ``cutoff`` percent of the
+    lightest and darkest pixels from the histogram, and remaps the image
+    so that the darkest pixel becomes black (0), and the lightest
+    becomes white (255).
+
+    :param image: The image to process.
+    :param cutoff: The percent to cut off from the histogram on the low and
+                   high ends. Either a tuple of (low, high), or a single
+                   number for both.
+    :param ignore: The background pixel value (use None for no background).
+    :param mask: Histogram used in contrast operation is computed using pixels
+                 within the mask. If no mask is given the entire image is used
+                 for histogram computation.
+    :param preserve_tone: Preserve image tone in Photoshop-like style autocontrast.
+
+                          .. versionadded:: 8.2.0
+
+    :return: An image.
+    """
+    if preserve_tone:
+        histogram = image.convert("L").histogram(mask)
+    else:
+        histogram = image.histogram(mask)
+
+    lut = []
+    for layer in range(0, len(histogram), 256):
+        h = histogram[layer : layer + 256]
+        if ignore is not None:
+            # get rid of outliers
+            if isinstance(ignore, int):
+                h[ignore] = 0
+            else:
+                for ix in ignore:
+                    h[ix] = 0
+        if cutoff:
+            # cut off pixels from both ends of the histogram
+            if not isinstance(cutoff, tuple):
+                cutoff = (cutoff, cutoff)
+            # get number of pixels
+            n = 0
+            for ix in range(256):
+                n = n + h[ix]
+            # remove cutoff% pixels from the low end
+            cut = int(n * cutoff[0] // 100)
+            for lo in range(256):
+                if cut > h[lo]:
+                    cut = cut - h[lo]
+                    h[lo] = 0
+                else:
+                    h[lo] -= cut
+                    cut = 0
+                if cut <= 0:
+                    break
+            # remove cutoff% samples from the high end
+            cut = int(n * cutoff[1] // 100)
+            for hi in range(255, -1, -1):
+                if cut > h[hi]:
+                    cut = cut - h[hi]
+                    h[hi] = 0
+                else:
+                    h[hi] -= cut
+                    cut = 0
+                if cut <= 0:
+                    break
+        # find lowest/highest samples after preprocessing
+        for lo in range(256):
+            if h[lo]:
+                break
+        for hi in range(255, -1, -1):
+            if h[hi]:
+                break
+        if hi <= lo:
+            # don't bother
+            lut.extend(list(range(256)))
+        else:
+            scale = 255.0 / (hi - lo)
+            offset = -lo * scale
+            for ix in range(256):
+                ix = int(ix * scale + offset)
+                if ix < 0:
+                    ix = 0
+                elif ix > 255:
+                    ix = 255
+                lut.append(ix)
+    return _lut(image, lut)
+
+
+def colorize(
+    image: Image.Image,
+    black: str | tuple[int, ...],
+    white: str | tuple[int, ...],
+    mid: str | int | tuple[int, ...] | None = None,
+    blackpoint: int = 0,
+    whitepoint: int = 255,
+    midpoint: int = 127,
+) -> Image.Image:
+    """
+    Colorize grayscale image.
+    This function calculates a color wedge which maps all black pixels in
+    the source image to the first color and all white pixels to the
+    second color. If ``mid`` is specified, it uses three-color mapping.
+    The ``black`` and ``white`` arguments should be RGB tuples or color names;
+    optionally you can use three-color mapping by also specifying ``mid``.
+    Mapping positions for any of the colors can be specified
+    (e.g. ``blackpoint``), where these parameters are the integer
+    value corresponding to where the corresponding color should be mapped.
+    These parameters must have logical order, such that
+    ``blackpoint <= midpoint <= whitepoint`` (if ``mid`` is specified).
+
+    :param image: The image to colorize.
+    :param black: The color to use for black input pixels.
+    :param white: The color to use for white input pixels.
+    :param mid: The color to use for midtone input pixels.
+    :param blackpoint: an int value [0, 255] for the black mapping.
+    :param whitepoint: an int value [0, 255] for the white mapping.
+    :param midpoint: an int value [0, 255] for the midtone mapping.
+    :return: An image.
+    """
+
+    # Initial asserts
+    assert image.mode == "L"
+    if mid is None:
+        assert 0 <= blackpoint <= whitepoint <= 255
+    else:
+        assert 0 <= blackpoint <= midpoint <= whitepoint <= 255
+
+    # Define colors from arguments
+    rgb_black = cast(Sequence[int], _color(black, "RGB"))
+    rgb_white = cast(Sequence[int], _color(white, "RGB"))
+    rgb_mid = cast(Sequence[int], _color(mid, "RGB")) if mid is not None else None
+
+    # Empty lists for the mapping
+    red = []
+    green = []
+    blue = []
+
+    # Create the low-end values
+    for i in range(0, blackpoint):
+        red.append(rgb_black[0])
+        green.append(rgb_black[1])
+        blue.append(rgb_black[2])
+
+    # Create the mapping (2-color)
+    if rgb_mid is None:
+        range_map = range(0, whitepoint - blackpoint)
+
+        for i in range_map:
+            red.append(
+                rgb_black[0] + i * (rgb_white[0] - rgb_black[0]) // len(range_map)
+            )
+            green.append(
+                rgb_black[1] + i * (rgb_white[1] - rgb_black[1]) // len(range_map)
+            )
+            blue.append(
+                rgb_black[2] + i * (rgb_white[2] - rgb_black[2]) // len(range_map)
+            )
+
+    # Create the mapping (3-color)
+    else:
+        range_map1 = range(0, midpoint - blackpoint)
+        range_map2 = range(0, whitepoint - midpoint)
+
+        for i in range_map1:
+            red.append(
+                rgb_black[0] + i * (rgb_mid[0] - rgb_black[0]) // len(range_map1)
+            )
+            green.append(
+                rgb_black[1] + i * (rgb_mid[1] - rgb_black[1]) // len(range_map1)
+            )
+            blue.append(
+                rgb_black[2] + i * (rgb_mid[2] - rgb_black[2]) // len(range_map1)
+            )
+        for i in range_map2:
+            red.append(rgb_mid[0] + i * (rgb_white[0] - rgb_mid[0]) // len(range_map2))
+            green.append(
+                rgb_mid[1] + i * (rgb_white[1] - rgb_mid[1]) // len(range_map2)
+            )
+            blue.append(rgb_mid[2] + i * (rgb_white[2] - rgb_mid[2]) // len(range_map2))
+
+    # Create the high-end values
+    for i in range(0, 256 - whitepoint):
+        red.append(rgb_white[0])
+        green.append(rgb_white[1])
+        blue.append(rgb_white[2])
+
+    # Return converted image
+    image = image.convert("RGB")
+    return _lut(image, red + green + blue)
+
+
+def contain(
+    image: Image.Image, size: tuple[int, int], method: int = Image.Resampling.BICUBIC
+) -> Image.Image:
+    """
+    Returns a resized version of the image, set to the maximum width and height
+    within the requested size, while maintaining the original aspect ratio.
+
+    :param image: The image to resize.
+    :param size: The requested output size in pixels, given as a
+                 (width, height) tuple.
+    :param method: Resampling method to use. Default is
+                   :py:attr:`~PIL.Image.Resampling.BICUBIC`.
+                   See :ref:`concept-filters`.
+    :return: An image.
+    """
+
+    im_ratio = image.width / image.height
+    dest_ratio = size[0] / size[1]
+
+    if im_ratio != dest_ratio:
+        if im_ratio > dest_ratio:
+            new_height = round(image.height / image.width * size[0])
+            if new_height != size[1]:
+                size = (size[0], new_height)
+        else:
+            new_width = round(image.width / image.height * size[1])
+            if new_width != size[0]:
+                size = (new_width, size[1])
+    return image.resize(size, resample=method)
+
+
+def cover(
+    image: Image.Image, size: tuple[int, int], method: int = Image.Resampling.BICUBIC
+) -> Image.Image:
+    """
+    Returns a resized version of the image, so that the requested size is
+    covered, while maintaining the original aspect ratio.
+
+    :param image: The image to resize.
+    :param size: The requested output size in pixels, given as a
+                 (width, height) tuple.
+    :param method: Resampling method to use. Default is
+                   :py:attr:`~PIL.Image.Resampling.BICUBIC`.
+                   See :ref:`concept-filters`.
+    :return: An image.
+    """
+
+    im_ratio = image.width / image.height
+    dest_ratio = size[0] / size[1]
+
+    if im_ratio != dest_ratio:
+        if im_ratio < dest_ratio:
+            new_height = round(image.height / image.width * size[0])
+            if new_height != size[1]:
+                size = (size[0], new_height)
+        else:
+            new_width = round(image.width / image.height * size[1])
+            if new_width != size[0]:
+                size = (new_width, size[1])
+    return image.resize(size, resample=method)
+
+
+def pad(
+    image: Image.Image,
+    size: tuple[int, int],
+    method: int = Image.Resampling.BICUBIC,
+    color: str | int | tuple[int, ...] | None = None,
+    centering: tuple[float, float] = (0.5, 0.5),
+) -> Image.Image:
+    """
+    Returns a resized and padded version of the image, expanded to fill the
+    requested aspect ratio and size.
+
+    :param image: The image to resize and crop.
+    :param size: The requested output size in pixels, given as a
+                 (width, height) tuple.
+    :param method: Resampling method to use. Default is
+                   :py:attr:`~PIL.Image.Resampling.BICUBIC`.
+                   See :ref:`concept-filters`.
+    :param color: The background color of the padded image.
+    :param centering: Control the position of the original image within the
+                      padded version.
+
+                          (0.5, 0.5) will keep the image centered
+                          (0, 0) will keep the image aligned to the top left
+                          (1, 1) will keep the image aligned to the bottom
+                          right
+    :return: An image.
+    """
+
+    resized = contain(image, size, method)
+    if resized.size == size:
+        out = resized
+    else:
+        out = Image.new(image.mode, size, color)
+        if resized.palette:
+            palette = resized.getpalette()
+            if palette is not None:
+                out.putpalette(palette)
+        if resized.width != size[0]:
+            x = round((size[0] - resized.width) * max(0, min(centering[0], 1)))
+            out.paste(resized, (x, 0))
+        else:
+            y = round((size[1] - resized.height) * max(0, min(centering[1], 1)))
+            out.paste(resized, (0, y))
+    return out
+
+
+def crop(image: Image.Image, border: int = 0) -> Image.Image:
+    """
+    Remove border from image.  The same amount of pixels are removed
+    from all four sides.  This function works on all image modes.
+
+    .. seealso:: :py:meth:`~PIL.Image.Image.crop`
+
+    :param image: The image to crop.
+    :param border: The number of pixels to remove.
+    :return: An image.
+    """
+    left, top, right, bottom = _border(border)
+    return image.crop((left, top, image.size[0] - right, image.size[1] - bottom))
+
+
+def scale(
+    image: Image.Image, factor: float, resample: int = Image.Resampling.BICUBIC
+) -> Image.Image:
+    """
+    Returns a rescaled image by a specific factor given in parameter.
+    A factor greater than 1 expands the image, between 0 and 1 contracts the
+    image.
+
+    :param image: The image to rescale.
+    :param factor: The expansion factor, as a float.
+    :param resample: Resampling method to use. Default is
+                     :py:attr:`~PIL.Image.Resampling.BICUBIC`.
+                     See :ref:`concept-filters`.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+    if factor == 1:
+        return image.copy()
+    elif factor <= 0:
+        msg = "the factor must be greater than 0"
+        raise ValueError(msg)
+    else:
+        size = (round(factor * image.width), round(factor * image.height))
+        return image.resize(size, resample)
+
+
+class SupportsGetMesh(Protocol):
+    """
+    An object that supports the ``getmesh`` method, taking an image as an
+    argument, and returning a list of tuples. Each tuple contains two tuples,
+    the source box as a tuple of 4 integers, and a tuple of 8 integers for the
+    final quadrilateral, in order of top left, bottom left, bottom right, top
+    right.
+    """
+
+    def getmesh(
+        self, image: Image.Image
+    ) -> list[
+        tuple[tuple[int, int, int, int], tuple[int, int, int, int, int, int, int, int]]
+    ]: ...
+
+
+def deform(
+    image: Image.Image,
+    deformer: SupportsGetMesh,
+    resample: int = Image.Resampling.BILINEAR,
+) -> Image.Image:
+    """
+    Deform the image.
+
+    :param image: The image to deform.
+    :param deformer: A deformer object.  Any object that implements a
+                    ``getmesh`` method can be used.
+    :param resample: An optional resampling filter. Same values possible as
+       in the PIL.Image.transform function.
+    :return: An image.
+    """
+    return image.transform(
+        image.size, Image.Transform.MESH, deformer.getmesh(image), resample
+    )
+
+
+def equalize(image: Image.Image, mask: Image.Image | None = None) -> Image.Image:
+    """
+    Equalize the image histogram. This function applies a non-linear
+    mapping to the input image, in order to create a uniform
+    distribution of grayscale values in the output image.
+
+    :param image: The image to equalize.
+    :param mask: An optional mask.  If given, only the pixels selected by
+                 the mask are included in the analysis.
+    :return: An image.
+    """
+    if image.mode == "P":
+        image = image.convert("RGB")
+    h = image.histogram(mask)
+    lut = []
+    for b in range(0, len(h), 256):
+        histo = [_f for _f in h[b : b + 256] if _f]
+        if len(histo) <= 1:
+            lut.extend(list(range(256)))
+        else:
+            step = (functools.reduce(operator.add, histo) - histo[-1]) // 255
+            if not step:
+                lut.extend(list(range(256)))
+            else:
+                n = step // 2
+                for i in range(256):
+                    lut.append(n // step)
+                    n = n + h[i + b]
+    return _lut(image, lut)
+
+
+def expand(
+    image: Image.Image,
+    border: int | tuple[int, ...] = 0,
+    fill: str | int | tuple[int, ...] = 0,
+) -> Image.Image:
+    """
+    Add border to the image
+
+    :param image: The image to expand.
+    :param border: Border width, in pixels.
+    :param fill: Pixel fill value (a color value).  Default is 0 (black).
+    :return: An image.
+    """
+    left, top, right, bottom = _border(border)
+    width = left + image.size[0] + right
+    height = top + image.size[1] + bottom
+    color = _color(fill, image.mode)
+    if image.palette:
+        palette = ImagePalette.ImagePalette(palette=image.getpalette())
+        if isinstance(color, tuple) and (len(color) == 3 or len(color) == 4):
+            color = palette.getcolor(color)
+    else:
+        palette = None
+    out = Image.new(image.mode, (width, height), color)
+    if palette:
+        out.putpalette(palette.palette)
+    out.paste(image, (left, top))
+    return out
+
+
+def fit(
+    image: Image.Image,
+    size: tuple[int, int],
+    method: int = Image.Resampling.BICUBIC,
+    bleed: float = 0.0,
+    centering: tuple[float, float] = (0.5, 0.5),
+) -> Image.Image:
+    """
+    Returns a resized and cropped version of the image, cropped to the
+    requested aspect ratio and size.
+
+    This function was contributed by Kevin Cazabon.
+
+    :param image: The image to resize and crop.
+    :param size: The requested output size in pixels, given as a
+                 (width, height) tuple.
+    :param method: Resampling method to use. Default is
+                   :py:attr:`~PIL.Image.Resampling.BICUBIC`.
+                   See :ref:`concept-filters`.
+    :param bleed: Remove a border around the outside of the image from all
+                  four edges. The value is a decimal percentage (use 0.01 for
+                  one percent). The default value is 0 (no border).
+                  Cannot be greater than or equal to 0.5.
+    :param centering: Control the cropping position.  Use (0.5, 0.5) for
+                      center cropping (e.g. if cropping the width, take 50% off
+                      of the left side, and therefore 50% off the right side).
+                      (0.0, 0.0) will crop from the top left corner (i.e. if
+                      cropping the width, take all of the crop off of the right
+                      side, and if cropping the height, take all of it off the
+                      bottom).  (1.0, 0.0) will crop from the bottom left
+                      corner, etc. (i.e. if cropping the width, take all of the
+                      crop off the left side, and if cropping the height take
+                      none from the top, and therefore all off the bottom).
+    :return: An image.
+    """
+
+    # by Kevin Cazabon, Feb 17/2000
+    # kevin@cazabon.com
+    # https://www.cazabon.com
+
+    centering_x, centering_y = centering
+
+    if not 0.0 <= centering_x <= 1.0:
+        centering_x = 0.5
+    if not 0.0 <= centering_y <= 1.0:
+        centering_y = 0.5
+
+    if not 0.0 <= bleed < 0.5:
+        bleed = 0.0
+
+    # calculate the area to use for resizing and cropping, subtracting
+    # the 'bleed' around the edges
+
+    # number of pixels to trim off on Top and Bottom, Left and Right
+    bleed_pixels = (bleed * image.size[0], bleed * image.size[1])
+
+    live_size = (
+        image.size[0] - bleed_pixels[0] * 2,
+        image.size[1] - bleed_pixels[1] * 2,
+    )
+
+    # calculate the aspect ratio of the live_size
+    live_size_ratio = live_size[0] / live_size[1]
+
+    # calculate the aspect ratio of the output image
+    output_ratio = size[0] / size[1]
+
+    # figure out if the sides or top/bottom will be cropped off
+    if live_size_ratio == output_ratio:
+        # live_size is already the needed ratio
+        crop_width = live_size[0]
+        crop_height = live_size[1]
+    elif live_size_ratio >= output_ratio:
+        # live_size is wider than what's needed, crop the sides
+        crop_width = output_ratio * live_size[1]
+        crop_height = live_size[1]
+    else:
+        # live_size is taller than what's needed, crop the top and bottom
+        crop_width = live_size[0]
+        crop_height = live_size[0] / output_ratio
+
+    # make the crop
+    crop_left = bleed_pixels[0] + (live_size[0] - crop_width) * centering_x
+    crop_top = bleed_pixels[1] + (live_size[1] - crop_height) * centering_y
+
+    crop = (crop_left, crop_top, crop_left + crop_width, crop_top + crop_height)
+
+    # resize the image and return it
+    return image.resize(size, method, box=crop)
+
+
+def flip(image: Image.Image) -> Image.Image:
+    """
+    Flip the image vertically (top to bottom).
+
+    :param image: The image to flip.
+    :return: An image.
+    """
+    return image.transpose(Image.Transpose.FLIP_TOP_BOTTOM)
+
+
+def grayscale(image: Image.Image) -> Image.Image:
+    """
+    Convert the image to grayscale.
+
+    :param image: The image to convert.
+    :return: An image.
+    """
+    return image.convert("L")
+
+
+def invert(image: Image.Image) -> Image.Image:
+    """
+    Invert (negate) the image.
+
+    :param image: The image to invert.
+    :return: An image.
+    """
+    lut = list(range(255, -1, -1))
+    return image.point(lut) if image.mode == "1" else _lut(image, lut)
+
+
+def mirror(image: Image.Image) -> Image.Image:
+    """
+    Flip image horizontally (left to right).
+
+    :param image: The image to mirror.
+    :return: An image.
+    """
+    return image.transpose(Image.Transpose.FLIP_LEFT_RIGHT)
+
+
+def posterize(image: Image.Image, bits: int) -> Image.Image:
+    """
+    Reduce the number of bits for each color channel.
+
+    :param image: The image to posterize.
+    :param bits: The number of bits to keep for each channel (1-8).
+    :return: An image.
+    """
+    mask = ~(2 ** (8 - bits) - 1)
+    lut = [i & mask for i in range(256)]
+    return _lut(image, lut)
+
+
+def solarize(image: Image.Image, threshold: int = 128) -> Image.Image:
+    """
+    Invert all pixel values above a threshold.
+
+    :param image: The image to solarize.
+    :param threshold: All pixels above this grayscale level are inverted.
+    :return: An image.
+    """
+    lut = []
+    for i in range(256):
+        if i < threshold:
+            lut.append(i)
+        else:
+            lut.append(255 - i)
+    return _lut(image, lut)
+
+
+def exif_transpose(image: Image.Image, *, in_place: bool = False) -> Image.Image | None:
+    """
+    If an image has an EXIF Orientation tag, other than 1, transpose the image
+    accordingly, and remove the orientation data.
+
+    :param image: The image to transpose.
+    :param in_place: Boolean. Keyword-only argument.
+        If ``True``, the original image is modified in-place, and ``None`` is returned.
+        If ``False`` (default), a new :py:class:`~PIL.Image.Image` object is returned
+        with the transposition applied. If there is no transposition, a copy of the
+        image will be returned.
+    """
+    image.load()
+    image_exif = image.getexif()
+    orientation = image_exif.get(ExifTags.Base.Orientation, 1)
+    method = {
+        2: Image.Transpose.FLIP_LEFT_RIGHT,
+        3: Image.Transpose.ROTATE_180,
+        4: Image.Transpose.FLIP_TOP_BOTTOM,
+        5: Image.Transpose.TRANSPOSE,
+        6: Image.Transpose.ROTATE_270,
+        7: Image.Transpose.TRANSVERSE,
+        8: Image.Transpose.ROTATE_90,
+    }.get(orientation)
+    if method is not None:
+        if in_place:
+            image.im = image.im.transpose(method)
+            image._size = image.im.size
+        else:
+            transposed_image = image.transpose(method)
+        exif_image = image if in_place else transposed_image
+
+        exif = exif_image.getexif()
+        if ExifTags.Base.Orientation in exif:
+            del exif[ExifTags.Base.Orientation]
+            if "exif" in exif_image.info:
+                exif_image.info["exif"] = exif.tobytes()
+            elif "Raw profile type exif" in exif_image.info:
+                exif_image.info["Raw profile type exif"] = exif.tobytes().hex()
+            for key in ("XML:com.adobe.xmp", "xmp"):
+                if key in exif_image.info:
+                    for pattern in (
+                        r'tiff:Orientation="([0-9])"',
+                        r"<tiff:Orientation>([0-9])</tiff:Orientation>",
+                    ):
+                        value = exif_image.info[key]
+                        exif_image.info[key] = (
+                            re.sub(pattern, "", value)
+                            if isinstance(value, str)
+                            else re.sub(pattern.encode(), b"", value)
+                        )
+        if not in_place:
+            return transposed_image
+    elif not in_place:
+        return image.copy()
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImagePalette.py b/backend/venv/lib/python3.12/site-packages/PIL/ImagePalette.py
new file mode 100644
index 00000000..183f8552
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImagePalette.py
@@ -0,0 +1,285 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# image palette object
+#
+# History:
+# 1996-03-11 fl   Rewritten.
+# 1997-01-03 fl   Up and running.
+# 1997-08-23 fl   Added load hack
+# 2001-04-16 fl   Fixed randint shadow bug in random()
+#
+# Copyright (c) 1997-2001 by Secret Labs AB
+# Copyright (c) 1996-1997 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import array
+from collections.abc import Sequence
+from typing import IO, TYPE_CHECKING
+
+from . import GimpGradientFile, GimpPaletteFile, ImageColor, PaletteFile
+
+if TYPE_CHECKING:
+    from . import Image
+
+
+class ImagePalette:
+    """
+    Color palette for palette mapped images
+
+    :param mode: The mode to use for the palette. See:
+        :ref:`concept-modes`. Defaults to "RGB"
+    :param palette: An optional palette. If given, it must be a bytearray,
+        an array or a list of ints between 0-255. The list must consist of
+        all channels for one color followed by the next color (e.g. RGBRGBRGB).
+        Defaults to an empty palette.
+    """
+
+    def __init__(
+        self,
+        mode: str = "RGB",
+        palette: Sequence[int] | bytes | bytearray | None = None,
+    ) -> None:
+        self.mode = mode
+        self.rawmode: str | None = None  # if set, palette contains raw data
+        self.palette = palette or bytearray()
+        self.dirty: int | None = None
+
+    @property
+    def palette(self) -> Sequence[int] | bytes | bytearray:
+        return self._palette
+
+    @palette.setter
+    def palette(self, palette: Sequence[int] | bytes | bytearray) -> None:
+        self._colors: dict[tuple[int, ...], int] | None = None
+        self._palette = palette
+
+    @property
+    def colors(self) -> dict[tuple[int, ...], int]:
+        if self._colors is None:
+            mode_len = len(self.mode)
+            self._colors = {}
+            for i in range(0, len(self.palette), mode_len):
+                color = tuple(self.palette[i : i + mode_len])
+                if color in self._colors:
+                    continue
+                self._colors[color] = i // mode_len
+        return self._colors
+
+    @colors.setter
+    def colors(self, colors: dict[tuple[int, ...], int]) -> None:
+        self._colors = colors
+
+    def copy(self) -> ImagePalette:
+        new = ImagePalette()
+
+        new.mode = self.mode
+        new.rawmode = self.rawmode
+        if self.palette is not None:
+            new.palette = self.palette[:]
+        new.dirty = self.dirty
+
+        return new
+
+    def getdata(self) -> tuple[str, Sequence[int] | bytes | bytearray]:
+        """
+        Get palette contents in format suitable for the low-level
+        ``im.putpalette`` primitive.
+
+        .. warning:: This method is experimental.
+        """
+        if self.rawmode:
+            return self.rawmode, self.palette
+        return self.mode, self.tobytes()
+
+    def tobytes(self) -> bytes:
+        """Convert palette to bytes.
+
+        .. warning:: This method is experimental.
+        """
+        if self.rawmode:
+            msg = "palette contains raw palette data"
+            raise ValueError(msg)
+        if isinstance(self.palette, bytes):
+            return self.palette
+        arr = array.array("B", self.palette)
+        return arr.tobytes()
+
+    # Declare tostring as an alias for tobytes
+    tostring = tobytes
+
+    def _new_color_index(
+        self, image: Image.Image | None = None, e: Exception | None = None
+    ) -> int:
+        if not isinstance(self.palette, bytearray):
+            self._palette = bytearray(self.palette)
+        index = len(self.palette) // 3
+        special_colors: tuple[int | tuple[int, ...] | None, ...] = ()
+        if image:
+            special_colors = (
+                image.info.get("background"),
+                image.info.get("transparency"),
+            )
+            while index in special_colors:
+                index += 1
+        if index >= 256:
+            if image:
+                # Search for an unused index
+                for i, count in reversed(list(enumerate(image.histogram()))):
+                    if count == 0 and i not in special_colors:
+                        index = i
+                        break
+            if index >= 256:
+                msg = "cannot allocate more than 256 colors"
+                raise ValueError(msg) from e
+        return index
+
+    def getcolor(
+        self,
+        color: tuple[int, ...],
+        image: Image.Image | None = None,
+    ) -> int:
+        """Given an rgb tuple, allocate palette entry.
+
+        .. warning:: This method is experimental.
+        """
+        if self.rawmode:
+            msg = "palette contains raw palette data"
+            raise ValueError(msg)
+        if isinstance(color, tuple):
+            if self.mode == "RGB":
+                if len(color) == 4:
+                    if color[3] != 255:
+                        msg = "cannot add non-opaque RGBA color to RGB palette"
+                        raise ValueError(msg)
+                    color = color[:3]
+            elif self.mode == "RGBA":
+                if len(color) == 3:
+                    color += (255,)
+            try:
+                return self.colors[color]
+            except KeyError as e:
+                # allocate new color slot
+                index = self._new_color_index(image, e)
+                assert isinstance(self._palette, bytearray)
+                self.colors[color] = index
+                if index * 3 < len(self.palette):
+                    self._palette = (
+                        self._palette[: index * 3]
+                        + bytes(color)
+                        + self._palette[index * 3 + 3 :]
+                    )
+                else:
+                    self._palette += bytes(color)
+                self.dirty = 1
+                return index
+        else:
+            msg = f"unknown color specifier: {repr(color)}"  # type: ignore[unreachable]
+            raise ValueError(msg)
+
+    def save(self, fp: str | IO[str]) -> None:
+        """Save palette to text file.
+
+        .. warning:: This method is experimental.
+        """
+        if self.rawmode:
+            msg = "palette contains raw palette data"
+            raise ValueError(msg)
+        if isinstance(fp, str):
+            fp = open(fp, "w")
+        fp.write("# Palette\n")
+        fp.write(f"# Mode: {self.mode}\n")
+        for i in range(256):
+            fp.write(f"{i}")
+            for j in range(i * len(self.mode), (i + 1) * len(self.mode)):
+                try:
+                    fp.write(f" {self.palette[j]}")
+                except IndexError:
+                    fp.write(" 0")
+            fp.write("\n")
+        fp.close()
+
+
+# --------------------------------------------------------------------
+# Internal
+
+
+def raw(rawmode: str, data: Sequence[int] | bytes | bytearray) -> ImagePalette:
+    palette = ImagePalette()
+    palette.rawmode = rawmode
+    palette.palette = data
+    palette.dirty = 1
+    return palette
+
+
+# --------------------------------------------------------------------
+# Factories
+
+
+def make_linear_lut(black: int, white: float) -> list[int]:
+    if black == 0:
+        return [int(white * i // 255) for i in range(256)]
+
+    msg = "unavailable when black is non-zero"
+    raise NotImplementedError(msg)  # FIXME
+
+
+def make_gamma_lut(exp: float) -> list[int]:
+    return [int(((i / 255.0) ** exp) * 255.0 + 0.5) for i in range(256)]
+
+
+def negative(mode: str = "RGB") -> ImagePalette:
+    palette = list(range(256 * len(mode)))
+    palette.reverse()
+    return ImagePalette(mode, [i // len(mode) for i in palette])
+
+
+def random(mode: str = "RGB") -> ImagePalette:
+    from random import randint
+
+    palette = [randint(0, 255) for _ in range(256 * len(mode))]
+    return ImagePalette(mode, palette)
+
+
+def sepia(white: str = "#fff0c0") -> ImagePalette:
+    bands = [make_linear_lut(0, band) for band in ImageColor.getrgb(white)]
+    return ImagePalette("RGB", [bands[i % 3][i // 3] for i in range(256 * 3)])
+
+
+def wedge(mode: str = "RGB") -> ImagePalette:
+    palette = list(range(256 * len(mode)))
+    return ImagePalette(mode, [i // len(mode) for i in palette])
+
+
+def load(filename: str) -> tuple[bytes, str]:
+    # FIXME: supports GIMP gradients only
+
+    with open(filename, "rb") as fp:
+        paletteHandlers: list[
+            type[
+                GimpPaletteFile.GimpPaletteFile
+                | GimpGradientFile.GimpGradientFile
+                | PaletteFile.PaletteFile
+            ]
+        ] = [
+            GimpPaletteFile.GimpPaletteFile,
+            GimpGradientFile.GimpGradientFile,
+            PaletteFile.PaletteFile,
+        ]
+        for paletteHandler in paletteHandlers:
+            try:
+                fp.seek(0)
+                lut = paletteHandler(fp).getpalette()
+                if lut:
+                    break
+            except (SyntaxError, ValueError):
+                pass
+        else:
+            msg = "cannot load palette"
+            raise OSError(msg)
+
+    return lut  # data, rawmode
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImagePath.py b/backend/venv/lib/python3.12/site-packages/PIL/ImagePath.py
new file mode 100644
index 00000000..77e8a609
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImagePath.py
@@ -0,0 +1,20 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# path interface
+#
+# History:
+# 1996-11-04 fl   Created
+# 2002-04-14 fl   Added documentation stub class
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image
+
+Path = Image.core.path
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageQt.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageQt.py
new file mode 100644
index 00000000..2cc40f85
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageQt.py
@@ -0,0 +1,219 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# a simple Qt image interface.
+#
+# history:
+# 2006-06-03 fl: created
+# 2006-06-04 fl: inherit from QImage instead of wrapping it
+# 2006-06-05 fl: removed toimage helper; move string support to ImageQt
+# 2013-11-13 fl: add support for Qt5 (aurelien.ballier@cyclonit.com)
+#
+# Copyright (c) 2006 by Secret Labs AB
+# Copyright (c) 2006 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import sys
+from io import BytesIO
+from typing import TYPE_CHECKING, Any, Callable, Union
+
+from . import Image
+from ._util import is_path
+
+if TYPE_CHECKING:
+    import PyQt6
+    import PySide6
+
+    from . import ImageFile
+
+    QBuffer: type
+    QByteArray = Union[PyQt6.QtCore.QByteArray, PySide6.QtCore.QByteArray]
+    QIODevice = Union[PyQt6.QtCore.QIODevice, PySide6.QtCore.QIODevice]
+    QImage = Union[PyQt6.QtGui.QImage, PySide6.QtGui.QImage]
+    QPixmap = Union[PyQt6.QtGui.QPixmap, PySide6.QtGui.QPixmap]
+
+qt_version: str | None
+qt_versions = [
+    ["6", "PyQt6"],
+    ["side6", "PySide6"],
+]
+
+# If a version has already been imported, attempt it first
+qt_versions.sort(key=lambda version: version[1] in sys.modules, reverse=True)
+for version, qt_module in qt_versions:
+    try:
+        qRgba: Callable[[int, int, int, int], int]
+        if qt_module == "PyQt6":
+            from PyQt6.QtCore import QBuffer, QIODevice
+            from PyQt6.QtGui import QImage, QPixmap, qRgba
+        elif qt_module == "PySide6":
+            from PySide6.QtCore import QBuffer, QIODevice
+            from PySide6.QtGui import QImage, QPixmap, qRgba
+    except (ImportError, RuntimeError):
+        continue
+    qt_is_installed = True
+    qt_version = version
+    break
+else:
+    qt_is_installed = False
+    qt_version = None
+
+
+def rgb(r: int, g: int, b: int, a: int = 255) -> int:
+    """(Internal) Turns an RGB color into a Qt compatible color integer."""
+    # use qRgb to pack the colors, and then turn the resulting long
+    # into a negative integer with the same bitpattern.
+    return qRgba(r, g, b, a) & 0xFFFFFFFF
+
+
+def fromqimage(im: QImage | QPixmap) -> ImageFile.ImageFile:
+    """
+    :param im: QImage or PIL ImageQt object
+    """
+    buffer = QBuffer()
+    qt_openmode: object
+    if qt_version == "6":
+        try:
+            qt_openmode = getattr(QIODevice, "OpenModeFlag")
+        except AttributeError:
+            qt_openmode = getattr(QIODevice, "OpenMode")
+    else:
+        qt_openmode = QIODevice
+    buffer.open(getattr(qt_openmode, "ReadWrite"))
+    # preserve alpha channel with png
+    # otherwise ppm is more friendly with Image.open
+    if im.hasAlphaChannel():
+        im.save(buffer, "png")
+    else:
+        im.save(buffer, "ppm")
+
+    b = BytesIO()
+    b.write(buffer.data())
+    buffer.close()
+    b.seek(0)
+
+    return Image.open(b)
+
+
+def fromqpixmap(im: QPixmap) -> ImageFile.ImageFile:
+    return fromqimage(im)
+
+
+def align8to32(bytes: bytes, width: int, mode: str) -> bytes:
+    """
+    converts each scanline of data from 8 bit to 32 bit aligned
+    """
+
+    bits_per_pixel = {"1": 1, "L": 8, "P": 8, "I;16": 16}[mode]
+
+    # calculate bytes per line and the extra padding if needed
+    bits_per_line = bits_per_pixel * width
+    full_bytes_per_line, remaining_bits_per_line = divmod(bits_per_line, 8)
+    bytes_per_line = full_bytes_per_line + (1 if remaining_bits_per_line else 0)
+
+    extra_padding = -bytes_per_line % 4
+
+    # already 32 bit aligned by luck
+    if not extra_padding:
+        return bytes
+
+    new_data = [
+        bytes[i * bytes_per_line : (i + 1) * bytes_per_line] + b"\x00" * extra_padding
+        for i in range(len(bytes) // bytes_per_line)
+    ]
+
+    return b"".join(new_data)
+
+
+def _toqclass_helper(im: Image.Image | str | QByteArray) -> dict[str, Any]:
+    data = None
+    colortable = None
+    exclusive_fp = False
+
+    # handle filename, if given instead of image name
+    if hasattr(im, "toUtf8"):
+        # FIXME - is this really the best way to do this?
+        im = str(im.toUtf8(), "utf-8")
+    if is_path(im):
+        im = Image.open(im)
+        exclusive_fp = True
+    assert isinstance(im, Image.Image)
+
+    qt_format = getattr(QImage, "Format") if qt_version == "6" else QImage
+    if im.mode == "1":
+        format = getattr(qt_format, "Format_Mono")
+    elif im.mode == "L":
+        format = getattr(qt_format, "Format_Indexed8")
+        colortable = [rgb(i, i, i) for i in range(256)]
+    elif im.mode == "P":
+        format = getattr(qt_format, "Format_Indexed8")
+        palette = im.getpalette()
+        assert palette is not None
+        colortable = [rgb(*palette[i : i + 3]) for i in range(0, len(palette), 3)]
+    elif im.mode == "RGB":
+        # Populate the 4th channel with 255
+        im = im.convert("RGBA")
+
+        data = im.tobytes("raw", "BGRA")
+        format = getattr(qt_format, "Format_RGB32")
+    elif im.mode == "RGBA":
+        data = im.tobytes("raw", "BGRA")
+        format = getattr(qt_format, "Format_ARGB32")
+    elif im.mode == "I;16":
+        im = im.point(lambda i: i * 256)
+
+        format = getattr(qt_format, "Format_Grayscale16")
+    else:
+        if exclusive_fp:
+            im.close()
+        msg = f"unsupported image mode {repr(im.mode)}"
+        raise ValueError(msg)
+
+    size = im.size
+    __data = data or align8to32(im.tobytes(), size[0], im.mode)
+    if exclusive_fp:
+        im.close()
+    return {"data": __data, "size": size, "format": format, "colortable": colortable}
+
+
+if qt_is_installed:
+
+    class ImageQt(QImage):  # type: ignore[misc]
+        def __init__(self, im: Image.Image | str | QByteArray) -> None:
+            """
+            An PIL image wrapper for Qt.  This is a subclass of PyQt's QImage
+            class.
+
+            :param im: A PIL Image object, or a file name (given either as
+                Python string or a PyQt string object).
+            """
+            im_data = _toqclass_helper(im)
+            # must keep a reference, or Qt will crash!
+            # All QImage constructors that take data operate on an existing
+            # buffer, so this buffer has to hang on for the life of the image.
+            # Fixes https://github.com/python-pillow/Pillow/issues/1370
+            self.__data = im_data["data"]
+            super().__init__(
+                self.__data,
+                im_data["size"][0],
+                im_data["size"][1],
+                im_data["format"],
+            )
+            if im_data["colortable"]:
+                self.setColorTable(im_data["colortable"])
+
+
+def toqimage(im: Image.Image | str | QByteArray) -> ImageQt:
+    return ImageQt(im)
+
+
+def toqpixmap(im: Image.Image | str | QByteArray) -> QPixmap:
+    qimage = toqimage(im)
+    pixmap = getattr(QPixmap, "fromImage")(qimage)
+    if qt_version == "6":
+        pixmap.detach()
+    return pixmap
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageSequence.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageSequence.py
new file mode 100644
index 00000000..a6fc340d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageSequence.py
@@ -0,0 +1,86 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# sequence support classes
+#
+# history:
+# 1997-02-20 fl     Created
+#
+# Copyright (c) 1997 by Secret Labs AB.
+# Copyright (c) 1997 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+##
+from __future__ import annotations
+
+from typing import Callable
+
+from . import Image
+
+
+class Iterator:
+    """
+    This class implements an iterator object that can be used to loop
+    over an image sequence.
+
+    You can use the ``[]`` operator to access elements by index. This operator
+    will raise an :py:exc:`IndexError` if you try to access a nonexistent
+    frame.
+
+    :param im: An image object.
+    """
+
+    def __init__(self, im: Image.Image) -> None:
+        if not hasattr(im, "seek"):
+            msg = "im must have seek method"
+            raise AttributeError(msg)
+        self.im = im
+        self.position = getattr(self.im, "_min_frame", 0)
+
+    def __getitem__(self, ix: int) -> Image.Image:
+        try:
+            self.im.seek(ix)
+            return self.im
+        except EOFError as e:
+            msg = "end of sequence"
+            raise IndexError(msg) from e
+
+    def __iter__(self) -> Iterator:
+        return self
+
+    def __next__(self) -> Image.Image:
+        try:
+            self.im.seek(self.position)
+            self.position += 1
+            return self.im
+        except EOFError as e:
+            msg = "end of sequence"
+            raise StopIteration(msg) from e
+
+
+def all_frames(
+    im: Image.Image | list[Image.Image],
+    func: Callable[[Image.Image], Image.Image] | None = None,
+) -> list[Image.Image]:
+    """
+    Applies a given function to all frames in an image or a list of images.
+    The frames are returned as a list of separate images.
+
+    :param im: An image, or a list of images.
+    :param func: The function to apply to all of the image frames.
+    :returns: A list of images.
+    """
+    if not isinstance(im, list):
+        im = [im]
+
+    ims = []
+    for imSequence in im:
+        current = imSequence.tell()
+
+        ims += [im_frame.copy() for im_frame in Iterator(imSequence)]
+
+        imSequence.seek(current)
+    return [func(im) for im in ims] if func else ims
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageShow.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageShow.py
new file mode 100644
index 00000000..d62893d9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageShow.py
@@ -0,0 +1,360 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# im.show() drivers
+#
+# History:
+# 2008-04-06 fl   Created
+#
+# Copyright (c) Secret Labs AB 2008.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import abc
+import os
+import shutil
+import subprocess
+import sys
+from shlex import quote
+from typing import Any
+
+from . import Image
+
+_viewers = []
+
+
+def register(viewer: type[Viewer] | Viewer, order: int = 1) -> None:
+    """
+    The :py:func:`register` function is used to register additional viewers::
+
+        from PIL import ImageShow
+        ImageShow.register(MyViewer())  # MyViewer will be used as a last resort
+        ImageShow.register(MySecondViewer(), 0)  # MySecondViewer will be prioritised
+        ImageShow.register(ImageShow.XVViewer(), 0)  # XVViewer will be prioritised
+
+    :param viewer: The viewer to be registered.
+    :param order:
+        Zero or a negative integer to prepend this viewer to the list,
+        a positive integer to append it.
+    """
+    if isinstance(viewer, type) and issubclass(viewer, Viewer):
+        viewer = viewer()
+    if order > 0:
+        _viewers.append(viewer)
+    else:
+        _viewers.insert(0, viewer)
+
+
+def show(image: Image.Image, title: str | None = None, **options: Any) -> bool:
+    r"""
+    Display a given image.
+
+    :param image: An image object.
+    :param title: Optional title. Not all viewers can display the title.
+    :param \**options: Additional viewer options.
+    :returns: ``True`` if a suitable viewer was found, ``False`` otherwise.
+    """
+    for viewer in _viewers:
+        if viewer.show(image, title=title, **options):
+            return True
+    return False
+
+
+class Viewer:
+    """Base class for viewers."""
+
+    # main api
+
+    def show(self, image: Image.Image, **options: Any) -> int:
+        """
+        The main function for displaying an image.
+        Converts the given image to the target format and displays it.
+        """
+
+        if not (
+            image.mode in ("1", "RGBA")
+            or (self.format == "PNG" and image.mode in ("I;16", "LA"))
+        ):
+            base = Image.getmodebase(image.mode)
+            if image.mode != base:
+                image = image.convert(base)
+
+        return self.show_image(image, **options)
+
+    # hook methods
+
+    format: str | None = None
+    """The format to convert the image into."""
+    options: dict[str, Any] = {}
+    """Additional options used to convert the image."""
+
+    def get_format(self, image: Image.Image) -> str | None:
+        """Return format name, or ``None`` to save as PGM/PPM."""
+        return self.format
+
+    def get_command(self, file: str, **options: Any) -> str:
+        """
+        Returns the command used to display the file.
+        Not implemented in the base class.
+        """
+        msg = "unavailable in base viewer"
+        raise NotImplementedError(msg)
+
+    def save_image(self, image: Image.Image) -> str:
+        """Save to temporary file and return filename."""
+        return image._dump(format=self.get_format(image), **self.options)
+
+    def show_image(self, image: Image.Image, **options: Any) -> int:
+        """Display the given image."""
+        return self.show_file(self.save_image(image), **options)
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        os.system(self.get_command(path, **options))  # nosec
+        return 1
+
+
+# --------------------------------------------------------------------
+
+
+class WindowsViewer(Viewer):
+    """The default viewer on Windows is the default system application for PNG files."""
+
+    format = "PNG"
+    options = {"compress_level": 1, "save_all": True}
+
+    def get_command(self, file: str, **options: Any) -> str:
+        return (
+            f'start "Pillow" /WAIT "{file}" '
+            "&& ping -n 4 127.0.0.1 >NUL "
+            f'&& del /f "{file}"'
+        )
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        subprocess.Popen(
+            self.get_command(path, **options),
+            shell=True,
+            creationflags=getattr(subprocess, "CREATE_NO_WINDOW"),
+        )  # nosec
+        return 1
+
+
+if sys.platform == "win32":
+    register(WindowsViewer)
+
+
+class MacViewer(Viewer):
+    """The default viewer on macOS using ``Preview.app``."""
+
+    format = "PNG"
+    options = {"compress_level": 1, "save_all": True}
+
+    def get_command(self, file: str, **options: Any) -> str:
+        # on darwin open returns immediately resulting in the temp
+        # file removal while app is opening
+        command = "open -a Preview.app"
+        command = f"({command} {quote(file)}; sleep 20; rm -f {quote(file)})&"
+        return command
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        subprocess.call(["open", "-a", "Preview.app", path])
+        executable = sys.executable or shutil.which("python3")
+        if executable:
+            subprocess.Popen(
+                [
+                    executable,
+                    "-c",
+                    "import os, sys, time; time.sleep(20); os.remove(sys.argv[1])",
+                    path,
+                ]
+            )
+        return 1
+
+
+if sys.platform == "darwin":
+    register(MacViewer)
+
+
+class UnixViewer(Viewer):
+    format = "PNG"
+    options = {"compress_level": 1, "save_all": True}
+
+    @abc.abstractmethod
+    def get_command_ex(self, file: str, **options: Any) -> tuple[str, str]:
+        pass
+
+    def get_command(self, file: str, **options: Any) -> str:
+        command = self.get_command_ex(file, **options)[0]
+        return f"{command} {quote(file)}"
+
+
+class XDGViewer(UnixViewer):
+    """
+    The freedesktop.org ``xdg-open`` command.
+    """
+
+    def get_command_ex(self, file: str, **options: Any) -> tuple[str, str]:
+        command = executable = "xdg-open"
+        return command, executable
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        subprocess.Popen(["xdg-open", path])
+        return 1
+
+
+class DisplayViewer(UnixViewer):
+    """
+    The ImageMagick ``display`` command.
+    This viewer supports the ``title`` parameter.
+    """
+
+    def get_command_ex(
+        self, file: str, title: str | None = None, **options: Any
+    ) -> tuple[str, str]:
+        command = executable = "display"
+        if title:
+            command += f" -title {quote(title)}"
+        return command, executable
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        args = ["display"]
+        title = options.get("title")
+        if title:
+            args += ["-title", title]
+        args.append(path)
+
+        subprocess.Popen(args)
+        return 1
+
+
+class GmDisplayViewer(UnixViewer):
+    """The GraphicsMagick ``gm display`` command."""
+
+    def get_command_ex(self, file: str, **options: Any) -> tuple[str, str]:
+        executable = "gm"
+        command = "gm display"
+        return command, executable
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        subprocess.Popen(["gm", "display", path])
+        return 1
+
+
+class EogViewer(UnixViewer):
+    """The GNOME Image Viewer ``eog`` command."""
+
+    def get_command_ex(self, file: str, **options: Any) -> tuple[str, str]:
+        executable = "eog"
+        command = "eog -n"
+        return command, executable
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        subprocess.Popen(["eog", "-n", path])
+        return 1
+
+
+class XVViewer(UnixViewer):
+    """
+    The X Viewer ``xv`` command.
+    This viewer supports the ``title`` parameter.
+    """
+
+    def get_command_ex(
+        self, file: str, title: str | None = None, **options: Any
+    ) -> tuple[str, str]:
+        # note: xv is pretty outdated.  most modern systems have
+        # imagemagick's display command instead.
+        command = executable = "xv"
+        if title:
+            command += f" -name {quote(title)}"
+        return command, executable
+
+    def show_file(self, path: str, **options: Any) -> int:
+        """
+        Display given file.
+        """
+        if not os.path.exists(path):
+            raise FileNotFoundError
+        args = ["xv"]
+        title = options.get("title")
+        if title:
+            args += ["-name", title]
+        args.append(path)
+
+        subprocess.Popen(args)
+        return 1
+
+
+if sys.platform not in ("win32", "darwin"):  # unixoids
+    if shutil.which("xdg-open"):
+        register(XDGViewer)
+    if shutil.which("display"):
+        register(DisplayViewer)
+    if shutil.which("gm"):
+        register(GmDisplayViewer)
+    if shutil.which("eog"):
+        register(EogViewer)
+    if shutil.which("xv"):
+        register(XVViewer)
+
+
+class IPythonViewer(Viewer):
+    """The viewer for IPython frontends."""
+
+    def show_image(self, image: Image.Image, **options: Any) -> int:
+        ipython_display(image)
+        return 1
+
+
+try:
+    from IPython.display import display as ipython_display
+except ImportError:
+    pass
+else:
+    register(IPythonViewer)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) < 2:
+        print("Syntax: python3 ImageShow.py imagefile [title]")
+        sys.exit()
+
+    with Image.open(sys.argv[1]) as im:
+        print(show(im, *sys.argv[2:]))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageStat.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageStat.py
new file mode 100644
index 00000000..8bc50452
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageStat.py
@@ -0,0 +1,160 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# global image statistics
+#
+# History:
+# 1996-04-05 fl   Created
+# 1997-05-21 fl   Added mask; added rms, var, stddev attributes
+# 1997-08-05 fl   Added median
+# 1998-07-05 hk   Fixed integer overflow error
+#
+# Notes:
+# This class shows how to implement delayed evaluation of attributes.
+# To get a certain value, simply access the corresponding attribute.
+# The __getattr__ dispatcher takes care of the rest.
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996-97.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import math
+from functools import cached_property
+
+from . import Image
+
+
+class Stat:
+    def __init__(
+        self, image_or_list: Image.Image | list[int], mask: Image.Image | None = None
+    ) -> None:
+        """
+        Calculate statistics for the given image. If a mask is included,
+        only the regions covered by that mask are included in the
+        statistics. You can also pass in a previously calculated histogram.
+
+        :param image: A PIL image, or a precalculated histogram.
+
+            .. note::
+
+                For a PIL image, calculations rely on the
+                :py:meth:`~PIL.Image.Image.histogram` method. The pixel counts are
+                grouped into 256 bins, even if the image has more than 8 bits per
+                channel. So ``I`` and ``F`` mode images have a maximum ``mean``,
+                ``median`` and ``rms`` of 255, and cannot have an ``extrema`` maximum
+                of more than 255.
+
+        :param mask: An optional mask.
+        """
+        if isinstance(image_or_list, Image.Image):
+            self.h = image_or_list.histogram(mask)
+        elif isinstance(image_or_list, list):
+            self.h = image_or_list
+        else:
+            msg = "first argument must be image or list"  # type: ignore[unreachable]
+            raise TypeError(msg)
+        self.bands = list(range(len(self.h) // 256))
+
+    @cached_property
+    def extrema(self) -> list[tuple[int, int]]:
+        """
+        Min/max values for each band in the image.
+
+        .. note::
+            This relies on the :py:meth:`~PIL.Image.Image.histogram` method, and
+            simply returns the low and high bins used. This is correct for
+            images with 8 bits per channel, but fails for other modes such as
+            ``I`` or ``F``. Instead, use :py:meth:`~PIL.Image.Image.getextrema` to
+            return per-band extrema for the image. This is more correct and
+            efficient because, for non-8-bit modes, the histogram method uses
+            :py:meth:`~PIL.Image.Image.getextrema` to determine the bins used.
+        """
+
+        def minmax(histogram: list[int]) -> tuple[int, int]:
+            res_min, res_max = 255, 0
+            for i in range(256):
+                if histogram[i]:
+                    res_min = i
+                    break
+            for i in range(255, -1, -1):
+                if histogram[i]:
+                    res_max = i
+                    break
+            return res_min, res_max
+
+        return [minmax(self.h[i:]) for i in range(0, len(self.h), 256)]
+
+    @cached_property
+    def count(self) -> list[int]:
+        """Total number of pixels for each band in the image."""
+        return [sum(self.h[i : i + 256]) for i in range(0, len(self.h), 256)]
+
+    @cached_property
+    def sum(self) -> list[float]:
+        """Sum of all pixels for each band in the image."""
+
+        v = []
+        for i in range(0, len(self.h), 256):
+            layer_sum = 0.0
+            for j in range(256):
+                layer_sum += j * self.h[i + j]
+            v.append(layer_sum)
+        return v
+
+    @cached_property
+    def sum2(self) -> list[float]:
+        """Squared sum of all pixels for each band in the image."""
+
+        v = []
+        for i in range(0, len(self.h), 256):
+            sum2 = 0.0
+            for j in range(256):
+                sum2 += (j**2) * float(self.h[i + j])
+            v.append(sum2)
+        return v
+
+    @cached_property
+    def mean(self) -> list[float]:
+        """Average (arithmetic mean) pixel level for each band in the image."""
+        return [self.sum[i] / self.count[i] for i in self.bands]
+
+    @cached_property
+    def median(self) -> list[int]:
+        """Median pixel level for each band in the image."""
+
+        v = []
+        for i in self.bands:
+            s = 0
+            half = self.count[i] // 2
+            b = i * 256
+            for j in range(256):
+                s = s + self.h[b + j]
+                if s > half:
+                    break
+            v.append(j)
+        return v
+
+    @cached_property
+    def rms(self) -> list[float]:
+        """RMS (root-mean-square) for each band in the image."""
+        return [math.sqrt(self.sum2[i] / self.count[i]) for i in self.bands]
+
+    @cached_property
+    def var(self) -> list[float]:
+        """Variance for each band in the image."""
+        return [
+            (self.sum2[i] - (self.sum[i] ** 2.0) / self.count[i]) / self.count[i]
+            for i in self.bands
+        ]
+
+    @cached_property
+    def stddev(self) -> list[float]:
+        """Standard deviation for each band in the image."""
+        return [math.sqrt(self.var[i]) for i in self.bands]
+
+
+Global = Stat  # compatibility
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageTk.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageTk.py
new file mode 100644
index 00000000..bf29fdba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageTk.py
@@ -0,0 +1,290 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# a Tk display interface
+#
+# History:
+# 96-04-08 fl   Created
+# 96-09-06 fl   Added getimage method
+# 96-11-01 fl   Rewritten, removed image attribute and crop method
+# 97-05-09 fl   Use PyImagingPaste method instead of image type
+# 97-05-12 fl   Minor tweaks to match the IFUNC95 interface
+# 97-05-17 fl   Support the "pilbitmap" booster patch
+# 97-06-05 fl   Added file= and data= argument to image constructors
+# 98-03-09 fl   Added width and height methods to Image classes
+# 98-07-02 fl   Use default mode for "P" images without palette attribute
+# 98-07-02 fl   Explicitly destroy Tkinter image objects
+# 99-07-24 fl   Support multiple Tk interpreters (from Greg Couch)
+# 99-07-26 fl   Automatically hook into Tkinter (if possible)
+# 99-08-15 fl   Hook uses _imagingtk instead of _imaging
+#
+# Copyright (c) 1997-1999 by Secret Labs AB
+# Copyright (c) 1996-1997 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import tkinter
+from io import BytesIO
+from typing import TYPE_CHECKING, Any, cast
+
+from . import Image, ImageFile
+
+if TYPE_CHECKING:
+    from ._typing import CapsuleType
+
+# --------------------------------------------------------------------
+# Check for Tkinter interface hooks
+
+
+def _get_image_from_kw(kw: dict[str, Any]) -> ImageFile.ImageFile | None:
+    source = None
+    if "file" in kw:
+        source = kw.pop("file")
+    elif "data" in kw:
+        source = BytesIO(kw.pop("data"))
+    if not source:
+        return None
+    return Image.open(source)
+
+
+def _pyimagingtkcall(
+    command: str, photo: PhotoImage | tkinter.PhotoImage, ptr: CapsuleType
+) -> None:
+    tk = photo.tk
+    try:
+        tk.call(command, photo, repr(ptr))
+    except tkinter.TclError:
+        # activate Tkinter hook
+        # may raise an error if it cannot attach to Tkinter
+        from . import _imagingtk
+
+        _imagingtk.tkinit(tk.interpaddr())
+        tk.call(command, photo, repr(ptr))
+
+
+# --------------------------------------------------------------------
+# PhotoImage
+
+
+class PhotoImage:
+    """
+    A Tkinter-compatible photo image.  This can be used
+    everywhere Tkinter expects an image object.  If the image is an RGBA
+    image, pixels having alpha 0 are treated as transparent.
+
+    The constructor takes either a PIL image, or a mode and a size.
+    Alternatively, you can use the ``file`` or ``data`` options to initialize
+    the photo image object.
+
+    :param image: Either a PIL image, or a mode string.  If a mode string is
+                  used, a size must also be given.
+    :param size: If the first argument is a mode string, this defines the size
+                 of the image.
+    :keyword file: A filename to load the image from (using
+                   ``Image.open(file)``).
+    :keyword data: An 8-bit string containing image data (as loaded from an
+                   image file).
+    """
+
+    def __init__(
+        self,
+        image: Image.Image | str | None = None,
+        size: tuple[int, int] | None = None,
+        **kw: Any,
+    ) -> None:
+        # Tk compatibility: file or data
+        if image is None:
+            image = _get_image_from_kw(kw)
+
+        if image is None:
+            msg = "Image is required"
+            raise ValueError(msg)
+        elif isinstance(image, str):
+            mode = image
+            image = None
+
+            if size is None:
+                msg = "If first argument is mode, size is required"
+                raise ValueError(msg)
+        else:
+            # got an image instead of a mode
+            mode = image.mode
+            if mode == "P":
+                # palette mapped data
+                image.apply_transparency()
+                image.load()
+                mode = image.palette.mode if image.palette else "RGB"
+            size = image.size
+            kw["width"], kw["height"] = size
+
+        if mode not in ["1", "L", "RGB", "RGBA"]:
+            mode = Image.getmodebase(mode)
+
+        self.__mode = mode
+        self.__size = size
+        self.__photo = tkinter.PhotoImage(**kw)
+        self.tk = self.__photo.tk
+        if image:
+            self.paste(image)
+
+    def __del__(self) -> None:
+        try:
+            name = self.__photo.name
+        except AttributeError:
+            return
+        self.__photo.name = None
+        try:
+            self.__photo.tk.call("image", "delete", name)
+        except Exception:
+            pass  # ignore internal errors
+
+    def __str__(self) -> str:
+        """
+        Get the Tkinter photo image identifier.  This method is automatically
+        called by Tkinter whenever a PhotoImage object is passed to a Tkinter
+        method.
+
+        :return: A Tkinter photo image identifier (a string).
+        """
+        return str(self.__photo)
+
+    def width(self) -> int:
+        """
+        Get the width of the image.
+
+        :return: The width, in pixels.
+        """
+        return self.__size[0]
+
+    def height(self) -> int:
+        """
+        Get the height of the image.
+
+        :return: The height, in pixels.
+        """
+        return self.__size[1]
+
+    def paste(self, im: Image.Image) -> None:
+        """
+        Paste a PIL image into the photo image.  Note that this can
+        be very slow if the photo image is displayed.
+
+        :param im: A PIL image. The size must match the target region.  If the
+                   mode does not match, the image is converted to the mode of
+                   the bitmap image.
+        """
+        # convert to blittable
+        ptr = im.getim()
+        image = im.im
+        if not image.isblock() or im.mode != self.__mode:
+            block = Image.core.new_block(self.__mode, im.size)
+            image.convert2(block, image)  # convert directly between buffers
+            ptr = block.ptr
+
+        _pyimagingtkcall("PyImagingPhoto", self.__photo, ptr)
+
+
+# --------------------------------------------------------------------
+# BitmapImage
+
+
+class BitmapImage:
+    """
+    A Tkinter-compatible bitmap image.  This can be used everywhere Tkinter
+    expects an image object.
+
+    The given image must have mode "1".  Pixels having value 0 are treated as
+    transparent.  Options, if any, are passed on to Tkinter.  The most commonly
+    used option is ``foreground``, which is used to specify the color for the
+    non-transparent parts.  See the Tkinter documentation for information on
+    how to specify colours.
+
+    :param image: A PIL image.
+    """
+
+    def __init__(self, image: Image.Image | None = None, **kw: Any) -> None:
+        # Tk compatibility: file or data
+        if image is None:
+            image = _get_image_from_kw(kw)
+
+        if image is None:
+            msg = "Image is required"
+            raise ValueError(msg)
+        self.__mode = image.mode
+        self.__size = image.size
+
+        self.__photo = tkinter.BitmapImage(data=image.tobitmap(), **kw)
+
+    def __del__(self) -> None:
+        try:
+            name = self.__photo.name
+        except AttributeError:
+            return
+        self.__photo.name = None
+        try:
+            self.__photo.tk.call("image", "delete", name)
+        except Exception:
+            pass  # ignore internal errors
+
+    def width(self) -> int:
+        """
+        Get the width of the image.
+
+        :return: The width, in pixels.
+        """
+        return self.__size[0]
+
+    def height(self) -> int:
+        """
+        Get the height of the image.
+
+        :return: The height, in pixels.
+        """
+        return self.__size[1]
+
+    def __str__(self) -> str:
+        """
+        Get the Tkinter bitmap image identifier.  This method is automatically
+        called by Tkinter whenever a BitmapImage object is passed to a Tkinter
+        method.
+
+        :return: A Tkinter bitmap image identifier (a string).
+        """
+        return str(self.__photo)
+
+
+def getimage(photo: PhotoImage) -> Image.Image:
+    """Copies the contents of a PhotoImage to a PIL image memory."""
+    im = Image.new("RGBA", (photo.width(), photo.height()))
+
+    _pyimagingtkcall("PyImagingPhotoGet", photo, im.getim())
+
+    return im
+
+
+def _show(image: Image.Image, title: str | None) -> None:
+    """Helper for the Image.show method."""
+
+    class UI(tkinter.Label):
+        def __init__(self, master: tkinter.Toplevel, im: Image.Image) -> None:
+            self.image: BitmapImage | PhotoImage
+            if im.mode == "1":
+                self.image = BitmapImage(im, foreground="white", master=master)
+            else:
+                self.image = PhotoImage(im, master=master)
+            if TYPE_CHECKING:
+                image = cast(tkinter._Image, self.image)
+            else:
+                image = self.image
+            super().__init__(master, image=image, bg="black", bd=0)
+
+    if not getattr(tkinter, "_default_root"):
+        msg = "tkinter not initialized"
+        raise OSError(msg)
+    top = tkinter.Toplevel()
+    if title:
+        top.title(title)
+    UI(top, image).pack()
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageTransform.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageTransform.py
new file mode 100644
index 00000000..a3d8f441
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageTransform.py
@@ -0,0 +1,136 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# transform wrappers
+#
+# History:
+# 2002-04-08 fl   Created
+#
+# Copyright (c) 2002 by Secret Labs AB
+# Copyright (c) 2002 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from collections.abc import Sequence
+from typing import Any
+
+from . import Image
+
+
+class Transform(Image.ImageTransformHandler):
+    """Base class for other transforms defined in :py:mod:`~PIL.ImageTransform`."""
+
+    method: Image.Transform
+
+    def __init__(self, data: Sequence[Any]) -> None:
+        self.data = data
+
+    def getdata(self) -> tuple[Image.Transform, Sequence[int]]:
+        return self.method, self.data
+
+    def transform(
+        self,
+        size: tuple[int, int],
+        image: Image.Image,
+        **options: Any,
+    ) -> Image.Image:
+        """Perform the transform. Called from :py:meth:`.Image.transform`."""
+        # can be overridden
+        method, data = self.getdata()
+        return image.transform(size, method, data, **options)
+
+
+class AffineTransform(Transform):
+    """
+    Define an affine image transform.
+
+    This function takes a 6-tuple (a, b, c, d, e, f) which contain the first
+    two rows from an affine transform matrix. For each pixel (x, y) in the
+    output image, the new value is taken from a position (a x + b y + c,
+    d x + e y + f) in the input image, rounded to nearest pixel.
+
+    This function can be used to scale, translate, rotate, and shear the
+    original image.
+
+    See :py:meth:`.Image.transform`
+
+    :param matrix: A 6-tuple (a, b, c, d, e, f) containing the first two rows
+        from an affine transform matrix.
+    """
+
+    method = Image.Transform.AFFINE
+
+
+class PerspectiveTransform(Transform):
+    """
+    Define a perspective image transform.
+
+    This function takes an 8-tuple (a, b, c, d, e, f, g, h). For each pixel
+    (x, y) in the output image, the new value is taken from a position
+    ((a x + b y + c) / (g x + h y + 1), (d x + e y + f) / (g x + h y + 1)) in
+    the input image, rounded to nearest pixel.
+
+    This function can be used to scale, translate, rotate, and shear the
+    original image.
+
+    See :py:meth:`.Image.transform`
+
+    :param matrix: An 8-tuple (a, b, c, d, e, f, g, h).
+    """
+
+    method = Image.Transform.PERSPECTIVE
+
+
+class ExtentTransform(Transform):
+    """
+    Define a transform to extract a subregion from an image.
+
+    Maps a rectangle (defined by two corners) from the image to a rectangle of
+    the given size. The resulting image will contain data sampled from between
+    the corners, such that (x0, y0) in the input image will end up at (0,0) in
+    the output image, and (x1, y1) at size.
+
+    This method can be used to crop, stretch, shrink, or mirror an arbitrary
+    rectangle in the current image. It is slightly slower than crop, but about
+    as fast as a corresponding resize operation.
+
+    See :py:meth:`.Image.transform`
+
+    :param bbox: A 4-tuple (x0, y0, x1, y1) which specifies two points in the
+        input image's coordinate system. See :ref:`coordinate-system`.
+    """
+
+    method = Image.Transform.EXTENT
+
+
+class QuadTransform(Transform):
+    """
+    Define a quad image transform.
+
+    Maps a quadrilateral (a region defined by four corners) from the image to a
+    rectangle of the given size.
+
+    See :py:meth:`.Image.transform`
+
+    :param xy: An 8-tuple (x0, y0, x1, y1, x2, y2, x3, y3) which contain the
+        upper left, lower left, lower right, and upper right corner of the
+        source quadrilateral.
+    """
+
+    method = Image.Transform.QUAD
+
+
+class MeshTransform(Transform):
+    """
+    Define a mesh image transform.  A mesh transform consists of one or more
+    individual quad transforms.
+
+    See :py:meth:`.Image.transform`
+
+    :param data: A list of (bbox, quad) tuples.
+    """
+
+    method = Image.Transform.MESH
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImageWin.py b/backend/venv/lib/python3.12/site-packages/PIL/ImageWin.py
new file mode 100644
index 00000000..98c28f29
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImageWin.py
@@ -0,0 +1,247 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# a Windows DIB display interface
+#
+# History:
+# 1996-05-20 fl   Created
+# 1996-09-20 fl   Fixed subregion exposure
+# 1997-09-21 fl   Added draw primitive (for tzPrint)
+# 2003-05-21 fl   Added experimental Window/ImageWindow classes
+# 2003-09-05 fl   Added fromstring/tostring methods
+#
+# Copyright (c) Secret Labs AB 1997-2003.
+# Copyright (c) Fredrik Lundh 1996-2003.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image
+
+
+class HDC:
+    """
+    Wraps an HDC integer. The resulting object can be passed to the
+    :py:meth:`~PIL.ImageWin.Dib.draw` and :py:meth:`~PIL.ImageWin.Dib.expose`
+    methods.
+    """
+
+    def __init__(self, dc: int) -> None:
+        self.dc = dc
+
+    def __int__(self) -> int:
+        return self.dc
+
+
+class HWND:
+    """
+    Wraps an HWND integer. The resulting object can be passed to the
+    :py:meth:`~PIL.ImageWin.Dib.draw` and :py:meth:`~PIL.ImageWin.Dib.expose`
+    methods, instead of a DC.
+    """
+
+    def __init__(self, wnd: int) -> None:
+        self.wnd = wnd
+
+    def __int__(self) -> int:
+        return self.wnd
+
+
+class Dib:
+    """
+    A Windows bitmap with the given mode and size.  The mode can be one of "1",
+    "L", "P", or "RGB".
+
+    If the display requires a palette, this constructor creates a suitable
+    palette and associates it with the image. For an "L" image, 128 graylevels
+    are allocated. For an "RGB" image, a 6x6x6 colour cube is used, together
+    with 20 graylevels.
+
+    To make sure that palettes work properly under Windows, you must call the
+    ``palette`` method upon certain events from Windows.
+
+    :param image: Either a PIL image, or a mode string. If a mode string is
+                  used, a size must also be given.  The mode can be one of "1",
+                  "L", "P", or "RGB".
+    :param size: If the first argument is a mode string, this
+                 defines the size of the image.
+    """
+
+    def __init__(
+        self, image: Image.Image | str, size: tuple[int, int] | None = None
+    ) -> None:
+        if isinstance(image, str):
+            mode = image
+            image = ""
+            if size is None:
+                msg = "If first argument is mode, size is required"
+                raise ValueError(msg)
+        else:
+            mode = image.mode
+            size = image.size
+        if mode not in ["1", "L", "P", "RGB"]:
+            mode = Image.getmodebase(mode)
+        self.image = Image.core.display(mode, size)
+        self.mode = mode
+        self.size = size
+        if image:
+            assert not isinstance(image, str)
+            self.paste(image)
+
+    def expose(self, handle: int | HDC | HWND) -> None:
+        """
+        Copy the bitmap contents to a device context.
+
+        :param handle: Device context (HDC), cast to a Python integer, or an
+                       HDC or HWND instance.  In PythonWin, you can use
+                       ``CDC.GetHandleAttrib()`` to get a suitable handle.
+        """
+        handle_int = int(handle)
+        if isinstance(handle, HWND):
+            dc = self.image.getdc(handle_int)
+            try:
+                self.image.expose(dc)
+            finally:
+                self.image.releasedc(handle_int, dc)
+        else:
+            self.image.expose(handle_int)
+
+    def draw(
+        self,
+        handle: int | HDC | HWND,
+        dst: tuple[int, int, int, int],
+        src: tuple[int, int, int, int] | None = None,
+    ) -> None:
+        """
+        Same as expose, but allows you to specify where to draw the image, and
+        what part of it to draw.
+
+        The destination and source areas are given as 4-tuple rectangles. If
+        the source is omitted, the entire image is copied. If the source and
+        the destination have different sizes, the image is resized as
+        necessary.
+        """
+        if src is None:
+            src = (0, 0) + self.size
+        handle_int = int(handle)
+        if isinstance(handle, HWND):
+            dc = self.image.getdc(handle_int)
+            try:
+                self.image.draw(dc, dst, src)
+            finally:
+                self.image.releasedc(handle_int, dc)
+        else:
+            self.image.draw(handle_int, dst, src)
+
+    def query_palette(self, handle: int | HDC | HWND) -> int:
+        """
+        Installs the palette associated with the image in the given device
+        context.
+
+        This method should be called upon **QUERYNEWPALETTE** and
+        **PALETTECHANGED** events from Windows. If this method returns a
+        non-zero value, one or more display palette entries were changed, and
+        the image should be redrawn.
+
+        :param handle: Device context (HDC), cast to a Python integer, or an
+                       HDC or HWND instance.
+        :return: The number of entries that were changed (if one or more entries,
+                 this indicates that the image should be redrawn).
+        """
+        handle_int = int(handle)
+        if isinstance(handle, HWND):
+            handle = self.image.getdc(handle_int)
+            try:
+                result = self.image.query_palette(handle)
+            finally:
+                self.image.releasedc(handle, handle)
+        else:
+            result = self.image.query_palette(handle_int)
+        return result
+
+    def paste(
+        self, im: Image.Image, box: tuple[int, int, int, int] | None = None
+    ) -> None:
+        """
+        Paste a PIL image into the bitmap image.
+
+        :param im: A PIL image.  The size must match the target region.
+                   If the mode does not match, the image is converted to the
+                   mode of the bitmap image.
+        :param box: A 4-tuple defining the left, upper, right, and
+                    lower pixel coordinate.  See :ref:`coordinate-system`. If
+                    None is given instead of a tuple, all of the image is
+                    assumed.
+        """
+        im.load()
+        if self.mode != im.mode:
+            im = im.convert(self.mode)
+        if box:
+            self.image.paste(im.im, box)
+        else:
+            self.image.paste(im.im)
+
+    def frombytes(self, buffer: bytes) -> None:
+        """
+        Load display memory contents from byte data.
+
+        :param buffer: A buffer containing display data (usually
+                       data returned from :py:func:`~PIL.ImageWin.Dib.tobytes`)
+        """
+        self.image.frombytes(buffer)
+
+    def tobytes(self) -> bytes:
+        """
+        Copy display memory contents to bytes object.
+
+        :return: A bytes object containing display data.
+        """
+        return self.image.tobytes()
+
+
+class Window:
+    """Create a Window with the given title size."""
+
+    def __init__(
+        self, title: str = "PIL", width: int | None = None, height: int | None = None
+    ) -> None:
+        self.hwnd = Image.core.createwindow(
+            title, self.__dispatcher, width or 0, height or 0
+        )
+
+    def __dispatcher(self, action: str, *args: int) -> None:
+        getattr(self, f"ui_handle_{action}")(*args)
+
+    def ui_handle_clear(self, dc: int, x0: int, y0: int, x1: int, y1: int) -> None:
+        pass
+
+    def ui_handle_damage(self, x0: int, y0: int, x1: int, y1: int) -> None:
+        pass
+
+    def ui_handle_destroy(self) -> None:
+        pass
+
+    def ui_handle_repair(self, dc: int, x0: int, y0: int, x1: int, y1: int) -> None:
+        pass
+
+    def ui_handle_resize(self, width: int, height: int) -> None:
+        pass
+
+    def mainloop(self) -> None:
+        Image.core.eventloop()
+
+
+class ImageWindow(Window):
+    """Create an image window which displays the given image."""
+
+    def __init__(self, image: Image.Image | Dib, title: str = "PIL") -> None:
+        if not isinstance(image, Dib):
+            image = Dib(image)
+        self.image = image
+        width, height = image.size
+        super().__init__(title, width=width, height=height)
+
+    def ui_handle_repair(self, dc: int, x0: int, y0: int, x1: int, y1: int) -> None:
+        self.image.draw(dc, (x0, y0, x1, y1))
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/ImtImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/ImtImagePlugin.py
new file mode 100644
index 00000000..068cd5c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/ImtImagePlugin.py
@@ -0,0 +1,103 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# IM Tools support for PIL
+#
+# history:
+# 1996-05-27 fl   Created (read 8-bit images only)
+# 2001-02-17 fl   Use 're' instead of 'regex' (Python 2.1) (0.2)
+#
+# Copyright (c) Secret Labs AB 1997-2001.
+# Copyright (c) Fredrik Lundh 1996-2001.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import re
+
+from . import Image, ImageFile
+
+#
+# --------------------------------------------------------------------
+
+field = re.compile(rb"([a-z]*) ([^ \r\n]*)")
+
+
+##
+# Image plugin for IM Tools images.
+
+
+class ImtImageFile(ImageFile.ImageFile):
+    format = "IMT"
+    format_description = "IM Tools"
+
+    def _open(self) -> None:
+        # Quick rejection: if there's not a LF among the first
+        # 100 bytes, this is (probably) not a text header.
+
+        assert self.fp is not None
+
+        buffer = self.fp.read(100)
+        if b"\n" not in buffer:
+            msg = "not an IM file"
+            raise SyntaxError(msg)
+
+        xsize = ysize = 0
+
+        while True:
+            if buffer:
+                s = buffer[:1]
+                buffer = buffer[1:]
+            else:
+                s = self.fp.read(1)
+            if not s:
+                break
+
+            if s == b"\x0C":
+                # image data begins
+                self.tile = [
+                    ImageFile._Tile(
+                        "raw",
+                        (0, 0) + self.size,
+                        self.fp.tell() - len(buffer),
+                        self.mode,
+                    )
+                ]
+
+                break
+
+            else:
+                # read key/value pair
+                if b"\n" not in buffer:
+                    buffer += self.fp.read(100)
+                lines = buffer.split(b"\n")
+                s += lines.pop(0)
+                buffer = b"\n".join(lines)
+                if len(s) == 1 or len(s) > 100:
+                    break
+                if s[0] == ord(b"*"):
+                    continue  # comment
+
+                m = field.match(s)
+                if not m:
+                    break
+                k, v = m.group(1, 2)
+                if k == b"width":
+                    xsize = int(v)
+                    self._size = xsize, ysize
+                elif k == b"height":
+                    ysize = int(v)
+                    self._size = xsize, ysize
+                elif k == b"pixel" and v == b"n8":
+                    self._mode = "L"
+
+
+#
+# --------------------------------------------------------------------
+
+Image.register_open(ImtImageFile.format, ImtImageFile)
+
+#
+# no extension registered (".im" is simply too common)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/IptcImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/IptcImagePlugin.py
new file mode 100644
index 00000000..60ab7c83
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/IptcImagePlugin.py
@@ -0,0 +1,249 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# IPTC/NAA file handling
+#
+# history:
+# 1995-10-01 fl   Created
+# 1998-03-09 fl   Cleaned up and added to PIL
+# 2002-06-18 fl   Added getiptcinfo helper
+#
+# Copyright (c) Secret Labs AB 1997-2002.
+# Copyright (c) Fredrik Lundh 1995.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from collections.abc import Sequence
+from io import BytesIO
+from typing import cast
+
+from . import Image, ImageFile
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._deprecate import deprecate
+
+COMPRESSION = {1: "raw", 5: "jpeg"}
+
+
+def __getattr__(name: str) -> bytes:
+    if name == "PAD":
+        deprecate("IptcImagePlugin.PAD", 12)
+        return b"\0\0\0\0"
+    msg = f"module '{__name__}' has no attribute '{name}'"
+    raise AttributeError(msg)
+
+
+#
+# Helpers
+
+
+def _i(c: bytes) -> int:
+    return i32((b"\0\0\0\0" + c)[-4:])
+
+
+def _i8(c: int | bytes) -> int:
+    return c if isinstance(c, int) else c[0]
+
+
+def i(c: bytes) -> int:
+    """.. deprecated:: 10.2.0"""
+    deprecate("IptcImagePlugin.i", 12)
+    return _i(c)
+
+
+def dump(c: Sequence[int | bytes]) -> None:
+    """.. deprecated:: 10.2.0"""
+    deprecate("IptcImagePlugin.dump", 12)
+    for i in c:
+        print(f"{_i8(i):02x}", end=" ")
+    print()
+
+
+##
+# Image plugin for IPTC/NAA datastreams.  To read IPTC/NAA fields
+# from TIFF and JPEG files, use the <b>getiptcinfo</b> function.
+
+
+class IptcImageFile(ImageFile.ImageFile):
+    format = "IPTC"
+    format_description = "IPTC/NAA"
+
+    def getint(self, key: tuple[int, int]) -> int:
+        return _i(self.info[key])
+
+    def field(self) -> tuple[tuple[int, int] | None, int]:
+        #
+        # get a IPTC field header
+        s = self.fp.read(5)
+        if not s.strip(b"\x00"):
+            return None, 0
+
+        tag = s[1], s[2]
+
+        # syntax
+        if s[0] != 0x1C or tag[0] not in [1, 2, 3, 4, 5, 6, 7, 8, 9, 240]:
+            msg = "invalid IPTC/NAA file"
+            raise SyntaxError(msg)
+
+        # field size
+        size = s[3]
+        if size > 132:
+            msg = "illegal field length in IPTC/NAA file"
+            raise OSError(msg)
+        elif size == 128:
+            size = 0
+        elif size > 128:
+            size = _i(self.fp.read(size - 128))
+        else:
+            size = i16(s, 3)
+
+        return tag, size
+
+    def _open(self) -> None:
+        # load descriptive fields
+        while True:
+            offset = self.fp.tell()
+            tag, size = self.field()
+            if not tag or tag == (8, 10):
+                break
+            if size:
+                tagdata = self.fp.read(size)
+            else:
+                tagdata = None
+            if tag in self.info:
+                if isinstance(self.info[tag], list):
+                    self.info[tag].append(tagdata)
+                else:
+                    self.info[tag] = [self.info[tag], tagdata]
+            else:
+                self.info[tag] = tagdata
+
+        # mode
+        layers = self.info[(3, 60)][0]
+        component = self.info[(3, 60)][1]
+        if (3, 65) in self.info:
+            id = self.info[(3, 65)][0] - 1
+        else:
+            id = 0
+        if layers == 1 and not component:
+            self._mode = "L"
+        elif layers == 3 and component:
+            self._mode = "RGB"[id]
+        elif layers == 4 and component:
+            self._mode = "CMYK"[id]
+
+        # size
+        self._size = self.getint((3, 20)), self.getint((3, 30))
+
+        # compression
+        try:
+            compression = COMPRESSION[self.getint((3, 120))]
+        except KeyError as e:
+            msg = "Unknown IPTC image compression"
+            raise OSError(msg) from e
+
+        # tile
+        if tag == (8, 10):
+            self.tile = [
+                ImageFile._Tile("iptc", (0, 0) + self.size, offset, compression)
+            ]
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if len(self.tile) != 1 or self.tile[0][0] != "iptc":
+            return ImageFile.ImageFile.load(self)
+
+        offset, compression = self.tile[0][2:]
+
+        self.fp.seek(offset)
+
+        # Copy image data to temporary file
+        o = BytesIO()
+        if compression == "raw":
+            # To simplify access to the extracted file,
+            # prepend a PPM header
+            o.write(b"P5\n%d %d\n255\n" % self.size)
+        while True:
+            type, size = self.field()
+            if type != (8, 10):
+                break
+            while size > 0:
+                s = self.fp.read(min(size, 8192))
+                if not s:
+                    break
+                o.write(s)
+                size -= len(s)
+
+        with Image.open(o) as _im:
+            _im.load()
+            self.im = _im.im
+        return None
+
+
+Image.register_open(IptcImageFile.format, IptcImageFile)
+
+Image.register_extension(IptcImageFile.format, ".iim")
+
+
+def getiptcinfo(
+    im: ImageFile.ImageFile,
+) -> dict[tuple[int, int], bytes | list[bytes]] | None:
+    """
+    Get IPTC information from TIFF, JPEG, or IPTC file.
+
+    :param im: An image containing IPTC data.
+    :returns: A dictionary containing IPTC information, or None if
+        no IPTC information block was found.
+    """
+    from . import JpegImagePlugin, TiffImagePlugin
+
+    data = None
+
+    info: dict[tuple[int, int], bytes | list[bytes]] = {}
+    if isinstance(im, IptcImageFile):
+        # return info dictionary right away
+        for k, v in im.info.items():
+            if isinstance(k, tuple):
+                info[k] = v
+        return info
+
+    elif isinstance(im, JpegImagePlugin.JpegImageFile):
+        # extract the IPTC/NAA resource
+        photoshop = im.info.get("photoshop")
+        if photoshop:
+            data = photoshop.get(0x0404)
+
+    elif isinstance(im, TiffImagePlugin.TiffImageFile):
+        # get raw data from the IPTC/NAA tag (PhotoShop tags the data
+        # as 4-byte integers, so we cannot use the get method...)
+        try:
+            data = im.tag_v2[TiffImagePlugin.IPTC_NAA_CHUNK]
+        except KeyError:
+            pass
+
+    if data is None:
+        return None  # no properties
+
+    # create an IptcImagePlugin object without initializing it
+    class FakeImage:
+        pass
+
+    fake_im = FakeImage()
+    fake_im.__class__ = IptcImageFile  # type: ignore[assignment]
+    iptc_im = cast(IptcImageFile, fake_im)
+
+    # parse the IPTC information chunk
+    iptc_im.info = {}
+    iptc_im.fp = BytesIO(data)
+
+    try:
+        iptc_im._open()
+    except (IndexError, KeyError):
+        pass  # expected failure
+
+    for k, v in iptc_im.info.items():
+        if isinstance(k, tuple):
+            info[k] = v
+    return info
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/Jpeg2KImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/Jpeg2KImagePlugin.py
new file mode 100644
index 00000000..67828358
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/Jpeg2KImagePlugin.py
@@ -0,0 +1,443 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# JPEG2000 file handling
+#
+# History:
+# 2014-03-12 ajh  Created
+# 2021-06-30 rogermb  Extract dpi information from the 'resc' header box
+#
+# Copyright (c) 2014 Coriolis Systems Limited
+# Copyright (c) 2014 Alastair Houghton
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import os
+import struct
+from collections.abc import Callable
+from typing import IO, cast
+
+from . import Image, ImageFile, ImagePalette, _binary
+
+
+class BoxReader:
+    """
+    A small helper class to read fields stored in JPEG2000 header boxes
+    and to easily step into and read sub-boxes.
+    """
+
+    def __init__(self, fp: IO[bytes], length: int = -1) -> None:
+        self.fp = fp
+        self.has_length = length >= 0
+        self.length = length
+        self.remaining_in_box = -1
+
+    def _can_read(self, num_bytes: int) -> bool:
+        if self.has_length and self.fp.tell() + num_bytes > self.length:
+            # Outside box: ensure we don't read past the known file length
+            return False
+        if self.remaining_in_box >= 0:
+            # Inside box contents: ensure read does not go past box boundaries
+            return num_bytes <= self.remaining_in_box
+        else:
+            return True  # No length known, just read
+
+    def _read_bytes(self, num_bytes: int) -> bytes:
+        if not self._can_read(num_bytes):
+            msg = "Not enough data in header"
+            raise SyntaxError(msg)
+
+        data = self.fp.read(num_bytes)
+        if len(data) < num_bytes:
+            msg = f"Expected to read {num_bytes} bytes but only got {len(data)}."
+            raise OSError(msg)
+
+        if self.remaining_in_box > 0:
+            self.remaining_in_box -= num_bytes
+        return data
+
+    def read_fields(self, field_format: str) -> tuple[int | bytes, ...]:
+        size = struct.calcsize(field_format)
+        data = self._read_bytes(size)
+        return struct.unpack(field_format, data)
+
+    def read_boxes(self) -> BoxReader:
+        size = self.remaining_in_box
+        data = self._read_bytes(size)
+        return BoxReader(io.BytesIO(data), size)
+
+    def has_next_box(self) -> bool:
+        if self.has_length:
+            return self.fp.tell() + self.remaining_in_box < self.length
+        else:
+            return True
+
+    def next_box_type(self) -> bytes:
+        # Skip the rest of the box if it has not been read
+        if self.remaining_in_box > 0:
+            self.fp.seek(self.remaining_in_box, os.SEEK_CUR)
+        self.remaining_in_box = -1
+
+        # Read the length and type of the next box
+        lbox, tbox = cast(tuple[int, bytes], self.read_fields(">I4s"))
+        if lbox == 1:
+            lbox = cast(int, self.read_fields(">Q")[0])
+            hlen = 16
+        else:
+            hlen = 8
+
+        if lbox < hlen or not self._can_read(lbox - hlen):
+            msg = "Invalid header length"
+            raise SyntaxError(msg)
+
+        self.remaining_in_box = lbox - hlen
+        return tbox
+
+
+def _parse_codestream(fp: IO[bytes]) -> tuple[tuple[int, int], str]:
+    """Parse the JPEG 2000 codestream to extract the size and component
+    count from the SIZ marker segment, returning a PIL (size, mode) tuple."""
+
+    hdr = fp.read(2)
+    lsiz = _binary.i16be(hdr)
+    siz = hdr + fp.read(lsiz - 2)
+    lsiz, rsiz, xsiz, ysiz, xosiz, yosiz, _, _, _, _, csiz = struct.unpack_from(
+        ">HHIIIIIIIIH", siz
+    )
+
+    size = (xsiz - xosiz, ysiz - yosiz)
+    if csiz == 1:
+        ssiz = struct.unpack_from(">B", siz, 38)
+        if (ssiz[0] & 0x7F) + 1 > 8:
+            mode = "I;16"
+        else:
+            mode = "L"
+    elif csiz == 2:
+        mode = "LA"
+    elif csiz == 3:
+        mode = "RGB"
+    elif csiz == 4:
+        mode = "RGBA"
+    else:
+        msg = "unable to determine J2K image mode"
+        raise SyntaxError(msg)
+
+    return size, mode
+
+
+def _res_to_dpi(num: int, denom: int, exp: int) -> float | None:
+    """Convert JPEG2000's (numerator, denominator, exponent-base-10) resolution,
+    calculated as (num / denom) * 10^exp and stored in dots per meter,
+    to floating-point dots per inch."""
+    if denom == 0:
+        return None
+    return (254 * num * (10**exp)) / (10000 * denom)
+
+
+def _parse_jp2_header(
+    fp: IO[bytes],
+) -> tuple[
+    tuple[int, int],
+    str,
+    str | None,
+    tuple[float, float] | None,
+    ImagePalette.ImagePalette | None,
+]:
+    """Parse the JP2 header box to extract size, component count,
+    color space information, and optionally DPI information,
+    returning a (size, mode, mimetype, dpi) tuple."""
+
+    # Find the JP2 header box
+    reader = BoxReader(fp)
+    header = None
+    mimetype = None
+    while reader.has_next_box():
+        tbox = reader.next_box_type()
+
+        if tbox == b"jp2h":
+            header = reader.read_boxes()
+            break
+        elif tbox == b"ftyp":
+            if reader.read_fields(">4s")[0] == b"jpx ":
+                mimetype = "image/jpx"
+    assert header is not None
+
+    size = None
+    mode = None
+    bpc = None
+    nc = None
+    dpi = None  # 2-tuple of DPI info, or None
+    palette = None
+
+    while header.has_next_box():
+        tbox = header.next_box_type()
+
+        if tbox == b"ihdr":
+            height, width, nc, bpc = header.read_fields(">IIHB")
+            assert isinstance(height, int)
+            assert isinstance(width, int)
+            assert isinstance(bpc, int)
+            size = (width, height)
+            if nc == 1 and (bpc & 0x7F) > 8:
+                mode = "I;16"
+            elif nc == 1:
+                mode = "L"
+            elif nc == 2:
+                mode = "LA"
+            elif nc == 3:
+                mode = "RGB"
+            elif nc == 4:
+                mode = "RGBA"
+        elif tbox == b"colr" and nc == 4:
+            meth, _, _, enumcs = header.read_fields(">BBBI")
+            if meth == 1 and enumcs == 12:
+                mode = "CMYK"
+        elif tbox == b"pclr" and mode in ("L", "LA"):
+            ne, npc = header.read_fields(">HB")
+            assert isinstance(ne, int)
+            assert isinstance(npc, int)
+            max_bitdepth = 0
+            for bitdepth in header.read_fields(">" + ("B" * npc)):
+                assert isinstance(bitdepth, int)
+                if bitdepth > max_bitdepth:
+                    max_bitdepth = bitdepth
+            if max_bitdepth <= 8:
+                palette = ImagePalette.ImagePalette("RGBA" if npc == 4 else "RGB")
+                for i in range(ne):
+                    color: list[int] = []
+                    for value in header.read_fields(">" + ("B" * npc)):
+                        assert isinstance(value, int)
+                        color.append(value)
+                    palette.getcolor(tuple(color))
+                mode = "P" if mode == "L" else "PA"
+        elif tbox == b"res ":
+            res = header.read_boxes()
+            while res.has_next_box():
+                tres = res.next_box_type()
+                if tres == b"resc":
+                    vrcn, vrcd, hrcn, hrcd, vrce, hrce = res.read_fields(">HHHHBB")
+                    assert isinstance(vrcn, int)
+                    assert isinstance(vrcd, int)
+                    assert isinstance(hrcn, int)
+                    assert isinstance(hrcd, int)
+                    assert isinstance(vrce, int)
+                    assert isinstance(hrce, int)
+                    hres = _res_to_dpi(hrcn, hrcd, hrce)
+                    vres = _res_to_dpi(vrcn, vrcd, vrce)
+                    if hres is not None and vres is not None:
+                        dpi = (hres, vres)
+                    break
+
+    if size is None or mode is None:
+        msg = "Malformed JP2 header"
+        raise SyntaxError(msg)
+
+    return size, mode, mimetype, dpi, palette
+
+
+##
+# Image plugin for JPEG2000 images.
+
+
+class Jpeg2KImageFile(ImageFile.ImageFile):
+    format = "JPEG2000"
+    format_description = "JPEG 2000 (ISO 15444)"
+
+    def _open(self) -> None:
+        sig = self.fp.read(4)
+        if sig == b"\xff\x4f\xff\x51":
+            self.codec = "j2k"
+            self._size, self._mode = _parse_codestream(self.fp)
+            self._parse_comment()
+        else:
+            sig = sig + self.fp.read(8)
+
+            if sig == b"\x00\x00\x00\x0cjP  \x0d\x0a\x87\x0a":
+                self.codec = "jp2"
+                header = _parse_jp2_header(self.fp)
+                self._size, self._mode, self.custom_mimetype, dpi, self.palette = header
+                if dpi is not None:
+                    self.info["dpi"] = dpi
+                if self.fp.read(12).endswith(b"jp2c\xff\x4f\xff\x51"):
+                    hdr = self.fp.read(2)
+                    length = _binary.i16be(hdr)
+                    self.fp.seek(length - 2, os.SEEK_CUR)
+                    self._parse_comment()
+            else:
+                msg = "not a JPEG 2000 file"
+                raise SyntaxError(msg)
+
+        self._reduce = 0
+        self.layers = 0
+
+        fd = -1
+        length = -1
+
+        try:
+            fd = self.fp.fileno()
+            length = os.fstat(fd).st_size
+        except Exception:
+            fd = -1
+            try:
+                pos = self.fp.tell()
+                self.fp.seek(0, io.SEEK_END)
+                length = self.fp.tell()
+                self.fp.seek(pos)
+            except Exception:
+                length = -1
+
+        self.tile = [
+            ImageFile._Tile(
+                "jpeg2k",
+                (0, 0) + self.size,
+                0,
+                (self.codec, self._reduce, self.layers, fd, length),
+            )
+        ]
+
+    def _parse_comment(self) -> None:
+        while True:
+            marker = self.fp.read(2)
+            if not marker:
+                break
+            typ = marker[1]
+            if typ in (0x90, 0xD9):
+                # Start of tile or end of codestream
+                break
+            hdr = self.fp.read(2)
+            length = _binary.i16be(hdr)
+            if typ == 0x64:
+                # Comment
+                self.info["comment"] = self.fp.read(length - 2)[2:]
+                break
+            else:
+                self.fp.seek(length - 2, os.SEEK_CUR)
+
+    @property  # type: ignore[override]
+    def reduce(
+        self,
+    ) -> (
+        Callable[[int | tuple[int, int], tuple[int, int, int, int] | None], Image.Image]
+        | int
+    ):
+        # https://github.com/python-pillow/Pillow/issues/4343 found that the
+        # new Image 'reduce' method was shadowed by this plugin's 'reduce'
+        # property. This attempts to allow for both scenarios
+        return self._reduce or super().reduce
+
+    @reduce.setter
+    def reduce(self, value: int) -> None:
+        self._reduce = value
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self.tile and self._reduce:
+            power = 1 << self._reduce
+            adjust = power >> 1
+            self._size = (
+                int((self.size[0] + adjust) / power),
+                int((self.size[1] + adjust) / power),
+            )
+
+            # Update the reduce and layers settings
+            t = self.tile[0]
+            assert isinstance(t[3], tuple)
+            t3 = (t[3][0], self._reduce, self.layers, t[3][3], t[3][4])
+            self.tile = [ImageFile._Tile(t[0], (0, 0) + self.size, t[2], t3)]
+
+        return ImageFile.ImageFile.load(self)
+
+
+def _accept(prefix: bytes) -> bool:
+    return (
+        prefix[:4] == b"\xff\x4f\xff\x51"
+        or prefix[:12] == b"\x00\x00\x00\x0cjP  \x0d\x0a\x87\x0a"
+    )
+
+
+# ------------------------------------------------------------
+# Save support
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    # Get the keyword arguments
+    info = im.encoderinfo
+
+    if isinstance(filename, str):
+        filename = filename.encode()
+    if filename.endswith(b".j2k") or info.get("no_jp2", False):
+        kind = "j2k"
+    else:
+        kind = "jp2"
+
+    offset = info.get("offset", None)
+    tile_offset = info.get("tile_offset", None)
+    tile_size = info.get("tile_size", None)
+    quality_mode = info.get("quality_mode", "rates")
+    quality_layers = info.get("quality_layers", None)
+    if quality_layers is not None and not (
+        isinstance(quality_layers, (list, tuple))
+        and all(
+            isinstance(quality_layer, (int, float)) for quality_layer in quality_layers
+        )
+    ):
+        msg = "quality_layers must be a sequence of numbers"
+        raise ValueError(msg)
+
+    num_resolutions = info.get("num_resolutions", 0)
+    cblk_size = info.get("codeblock_size", None)
+    precinct_size = info.get("precinct_size", None)
+    irreversible = info.get("irreversible", False)
+    progression = info.get("progression", "LRCP")
+    cinema_mode = info.get("cinema_mode", "no")
+    mct = info.get("mct", 0)
+    signed = info.get("signed", False)
+    comment = info.get("comment")
+    if isinstance(comment, str):
+        comment = comment.encode()
+    plt = info.get("plt", False)
+
+    fd = -1
+    if hasattr(fp, "fileno"):
+        try:
+            fd = fp.fileno()
+        except Exception:
+            fd = -1
+
+    im.encoderconfig = (
+        offset,
+        tile_offset,
+        tile_size,
+        quality_mode,
+        quality_layers,
+        num_resolutions,
+        cblk_size,
+        precinct_size,
+        irreversible,
+        progression,
+        cinema_mode,
+        mct,
+        signed,
+        fd,
+        comment,
+        plt,
+    )
+
+    ImageFile._save(im, fp, [ImageFile._Tile("jpeg2k", (0, 0) + im.size, 0, kind)])
+
+
+# ------------------------------------------------------------
+# Registry stuff
+
+
+Image.register_open(Jpeg2KImageFile.format, Jpeg2KImageFile, _accept)
+Image.register_save(Jpeg2KImageFile.format, _save)
+
+Image.register_extensions(
+    Jpeg2KImageFile.format, [".jp2", ".j2k", ".jpc", ".jpf", ".jpx", ".j2c"]
+)
+
+Image.register_mime(Jpeg2KImageFile.format, "image/jp2")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/JpegImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/JpegImagePlugin.py
new file mode 100644
index 00000000..457690aa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/JpegImagePlugin.py
@@ -0,0 +1,905 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# JPEG (JFIF) file handling
+#
+# See "Digital Compression and Coding of Continuous-Tone Still Images,
+# Part 1, Requirements and Guidelines" (CCITT T.81 / ISO 10918-1)
+#
+# History:
+# 1995-09-09 fl   Created
+# 1995-09-13 fl   Added full parser
+# 1996-03-25 fl   Added hack to use the IJG command line utilities
+# 1996-05-05 fl   Workaround Photoshop 2.5 CMYK polarity bug
+# 1996-05-28 fl   Added draft support, JFIF version (0.1)
+# 1996-12-30 fl   Added encoder options, added progression property (0.2)
+# 1997-08-27 fl   Save mode 1 images as BW (0.3)
+# 1998-07-12 fl   Added YCbCr to draft and save methods (0.4)
+# 1998-10-19 fl   Don't hang on files using 16-bit DQT's (0.4.1)
+# 2001-04-16 fl   Extract DPI settings from JFIF files (0.4.2)
+# 2002-07-01 fl   Skip pad bytes before markers; identify Exif files (0.4.3)
+# 2003-04-25 fl   Added experimental EXIF decoder (0.5)
+# 2003-06-06 fl   Added experimental EXIF GPSinfo decoder
+# 2003-09-13 fl   Extract COM markers
+# 2009-09-06 fl   Added icc_profile support (from Florian Hoech)
+# 2009-03-06 fl   Changed CMYK handling; always use Adobe polarity (0.6)
+# 2009-03-08 fl   Added subsampling support (from Justin Huff).
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1995-1996 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import array
+import io
+import math
+import os
+import struct
+import subprocess
+import sys
+import tempfile
+import warnings
+from typing import IO, TYPE_CHECKING, Any
+
+from . import Image, ImageFile
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._binary import o8
+from ._binary import o16be as o16
+from ._deprecate import deprecate
+from .JpegPresets import presets
+
+if TYPE_CHECKING:
+    from .MpoImagePlugin import MpoImageFile
+
+#
+# Parser
+
+
+def Skip(self: JpegImageFile, marker: int) -> None:
+    n = i16(self.fp.read(2)) - 2
+    ImageFile._safe_read(self.fp, n)
+
+
+def APP(self: JpegImageFile, marker: int) -> None:
+    #
+    # Application marker.  Store these in the APP dictionary.
+    # Also look for well-known application markers.
+
+    n = i16(self.fp.read(2)) - 2
+    s = ImageFile._safe_read(self.fp, n)
+
+    app = f"APP{marker & 15}"
+
+    self.app[app] = s  # compatibility
+    self.applist.append((app, s))
+
+    if marker == 0xFFE0 and s[:4] == b"JFIF":
+        # extract JFIF information
+        self.info["jfif"] = version = i16(s, 5)  # version
+        self.info["jfif_version"] = divmod(version, 256)
+        # extract JFIF properties
+        try:
+            jfif_unit = s[7]
+            jfif_density = i16(s, 8), i16(s, 10)
+        except Exception:
+            pass
+        else:
+            if jfif_unit == 1:
+                self.info["dpi"] = jfif_density
+            elif jfif_unit == 2:  # cm
+                # 1 dpcm = 2.54 dpi
+                self.info["dpi"] = tuple(d * 2.54 for d in jfif_density)
+            self.info["jfif_unit"] = jfif_unit
+            self.info["jfif_density"] = jfif_density
+    elif marker == 0xFFE1 and s[:6] == b"Exif\0\0":
+        # extract EXIF information
+        if "exif" in self.info:
+            self.info["exif"] += s[6:]
+        else:
+            self.info["exif"] = s
+            self._exif_offset = self.fp.tell() - n + 6
+    elif marker == 0xFFE1 and s[:29] == b"http://ns.adobe.com/xap/1.0/\x00":
+        self.info["xmp"] = s.split(b"\x00", 1)[1]
+    elif marker == 0xFFE2 and s[:5] == b"FPXR\0":
+        # extract FlashPix information (incomplete)
+        self.info["flashpix"] = s  # FIXME: value will change
+    elif marker == 0xFFE2 and s[:12] == b"ICC_PROFILE\0":
+        # Since an ICC profile can be larger than the maximum size of
+        # a JPEG marker (64K), we need provisions to split it into
+        # multiple markers. The format defined by the ICC specifies
+        # one or more APP2 markers containing the following data:
+        #   Identifying string      ASCII "ICC_PROFILE\0"  (12 bytes)
+        #   Marker sequence number  1, 2, etc (1 byte)
+        #   Number of markers       Total of APP2's used (1 byte)
+        #   Profile data            (remainder of APP2 data)
+        # Decoders should use the marker sequence numbers to
+        # reassemble the profile, rather than assuming that the APP2
+        # markers appear in the correct sequence.
+        self.icclist.append(s)
+    elif marker == 0xFFED and s[:14] == b"Photoshop 3.0\x00":
+        # parse the image resource block
+        offset = 14
+        photoshop = self.info.setdefault("photoshop", {})
+        while s[offset : offset + 4] == b"8BIM":
+            try:
+                offset += 4
+                # resource code
+                code = i16(s, offset)
+                offset += 2
+                # resource name (usually empty)
+                name_len = s[offset]
+                # name = s[offset+1:offset+1+name_len]
+                offset += 1 + name_len
+                offset += offset & 1  # align
+                # resource data block
+                size = i32(s, offset)
+                offset += 4
+                data = s[offset : offset + size]
+                if code == 0x03ED:  # ResolutionInfo
+                    photoshop[code] = {
+                        "XResolution": i32(data, 0) / 65536,
+                        "DisplayedUnitsX": i16(data, 4),
+                        "YResolution": i32(data, 8) / 65536,
+                        "DisplayedUnitsY": i16(data, 12),
+                    }
+                else:
+                    photoshop[code] = data
+                offset += size
+                offset += offset & 1  # align
+            except struct.error:
+                break  # insufficient data
+
+    elif marker == 0xFFEE and s[:5] == b"Adobe":
+        self.info["adobe"] = i16(s, 5)
+        # extract Adobe custom properties
+        try:
+            adobe_transform = s[11]
+        except IndexError:
+            pass
+        else:
+            self.info["adobe_transform"] = adobe_transform
+    elif marker == 0xFFE2 and s[:4] == b"MPF\0":
+        # extract MPO information
+        self.info["mp"] = s[4:]
+        # offset is current location minus buffer size
+        # plus constant header size
+        self.info["mpoffset"] = self.fp.tell() - n + 4
+
+
+def COM(self: JpegImageFile, marker: int) -> None:
+    #
+    # Comment marker.  Store these in the APP dictionary.
+    n = i16(self.fp.read(2)) - 2
+    s = ImageFile._safe_read(self.fp, n)
+
+    self.info["comment"] = s
+    self.app["COM"] = s  # compatibility
+    self.applist.append(("COM", s))
+
+
+def SOF(self: JpegImageFile, marker: int) -> None:
+    #
+    # Start of frame marker.  Defines the size and mode of the
+    # image.  JPEG is colour blind, so we use some simple
+    # heuristics to map the number of layers to an appropriate
+    # mode.  Note that this could be made a bit brighter, by
+    # looking for JFIF and Adobe APP markers.
+
+    n = i16(self.fp.read(2)) - 2
+    s = ImageFile._safe_read(self.fp, n)
+    self._size = i16(s, 3), i16(s, 1)
+
+    self.bits = s[0]
+    if self.bits != 8:
+        msg = f"cannot handle {self.bits}-bit layers"
+        raise SyntaxError(msg)
+
+    self.layers = s[5]
+    if self.layers == 1:
+        self._mode = "L"
+    elif self.layers == 3:
+        self._mode = "RGB"
+    elif self.layers == 4:
+        self._mode = "CMYK"
+    else:
+        msg = f"cannot handle {self.layers}-layer images"
+        raise SyntaxError(msg)
+
+    if marker in [0xFFC2, 0xFFC6, 0xFFCA, 0xFFCE]:
+        self.info["progressive"] = self.info["progression"] = 1
+
+    if self.icclist:
+        # fixup icc profile
+        self.icclist.sort()  # sort by sequence number
+        if self.icclist[0][13] == len(self.icclist):
+            profile = [p[14:] for p in self.icclist]
+            icc_profile = b"".join(profile)
+        else:
+            icc_profile = None  # wrong number of fragments
+        self.info["icc_profile"] = icc_profile
+        self.icclist = []
+
+    for i in range(6, len(s), 3):
+        t = s[i : i + 3]
+        # 4-tuples: id, vsamp, hsamp, qtable
+        self.layer.append((t[0], t[1] // 16, t[1] & 15, t[2]))
+
+
+def DQT(self: JpegImageFile, marker: int) -> None:
+    #
+    # Define quantization table.  Note that there might be more
+    # than one table in each marker.
+
+    # FIXME: The quantization tables can be used to estimate the
+    # compression quality.
+
+    n = i16(self.fp.read(2)) - 2
+    s = ImageFile._safe_read(self.fp, n)
+    while len(s):
+        v = s[0]
+        precision = 1 if (v // 16 == 0) else 2  # in bytes
+        qt_length = 1 + precision * 64
+        if len(s) < qt_length:
+            msg = "bad quantization table marker"
+            raise SyntaxError(msg)
+        data = array.array("B" if precision == 1 else "H", s[1:qt_length])
+        if sys.byteorder == "little" and precision > 1:
+            data.byteswap()  # the values are always big-endian
+        self.quantization[v & 15] = [data[i] for i in zigzag_index]
+        s = s[qt_length:]
+
+
+#
+# JPEG marker table
+
+MARKER = {
+    0xFFC0: ("SOF0", "Baseline DCT", SOF),
+    0xFFC1: ("SOF1", "Extended Sequential DCT", SOF),
+    0xFFC2: ("SOF2", "Progressive DCT", SOF),
+    0xFFC3: ("SOF3", "Spatial lossless", SOF),
+    0xFFC4: ("DHT", "Define Huffman table", Skip),
+    0xFFC5: ("SOF5", "Differential sequential DCT", SOF),
+    0xFFC6: ("SOF6", "Differential progressive DCT", SOF),
+    0xFFC7: ("SOF7", "Differential spatial", SOF),
+    0xFFC8: ("JPG", "Extension", None),
+    0xFFC9: ("SOF9", "Extended sequential DCT (AC)", SOF),
+    0xFFCA: ("SOF10", "Progressive DCT (AC)", SOF),
+    0xFFCB: ("SOF11", "Spatial lossless DCT (AC)", SOF),
+    0xFFCC: ("DAC", "Define arithmetic coding conditioning", Skip),
+    0xFFCD: ("SOF13", "Differential sequential DCT (AC)", SOF),
+    0xFFCE: ("SOF14", "Differential progressive DCT (AC)", SOF),
+    0xFFCF: ("SOF15", "Differential spatial (AC)", SOF),
+    0xFFD0: ("RST0", "Restart 0", None),
+    0xFFD1: ("RST1", "Restart 1", None),
+    0xFFD2: ("RST2", "Restart 2", None),
+    0xFFD3: ("RST3", "Restart 3", None),
+    0xFFD4: ("RST4", "Restart 4", None),
+    0xFFD5: ("RST5", "Restart 5", None),
+    0xFFD6: ("RST6", "Restart 6", None),
+    0xFFD7: ("RST7", "Restart 7", None),
+    0xFFD8: ("SOI", "Start of image", None),
+    0xFFD9: ("EOI", "End of image", None),
+    0xFFDA: ("SOS", "Start of scan", Skip),
+    0xFFDB: ("DQT", "Define quantization table", DQT),
+    0xFFDC: ("DNL", "Define number of lines", Skip),
+    0xFFDD: ("DRI", "Define restart interval", Skip),
+    0xFFDE: ("DHP", "Define hierarchical progression", SOF),
+    0xFFDF: ("EXP", "Expand reference component", Skip),
+    0xFFE0: ("APP0", "Application segment 0", APP),
+    0xFFE1: ("APP1", "Application segment 1", APP),
+    0xFFE2: ("APP2", "Application segment 2", APP),
+    0xFFE3: ("APP3", "Application segment 3", APP),
+    0xFFE4: ("APP4", "Application segment 4", APP),
+    0xFFE5: ("APP5", "Application segment 5", APP),
+    0xFFE6: ("APP6", "Application segment 6", APP),
+    0xFFE7: ("APP7", "Application segment 7", APP),
+    0xFFE8: ("APP8", "Application segment 8", APP),
+    0xFFE9: ("APP9", "Application segment 9", APP),
+    0xFFEA: ("APP10", "Application segment 10", APP),
+    0xFFEB: ("APP11", "Application segment 11", APP),
+    0xFFEC: ("APP12", "Application segment 12", APP),
+    0xFFED: ("APP13", "Application segment 13", APP),
+    0xFFEE: ("APP14", "Application segment 14", APP),
+    0xFFEF: ("APP15", "Application segment 15", APP),
+    0xFFF0: ("JPG0", "Extension 0", None),
+    0xFFF1: ("JPG1", "Extension 1", None),
+    0xFFF2: ("JPG2", "Extension 2", None),
+    0xFFF3: ("JPG3", "Extension 3", None),
+    0xFFF4: ("JPG4", "Extension 4", None),
+    0xFFF5: ("JPG5", "Extension 5", None),
+    0xFFF6: ("JPG6", "Extension 6", None),
+    0xFFF7: ("JPG7", "Extension 7", None),
+    0xFFF8: ("JPG8", "Extension 8", None),
+    0xFFF9: ("JPG9", "Extension 9", None),
+    0xFFFA: ("JPG10", "Extension 10", None),
+    0xFFFB: ("JPG11", "Extension 11", None),
+    0xFFFC: ("JPG12", "Extension 12", None),
+    0xFFFD: ("JPG13", "Extension 13", None),
+    0xFFFE: ("COM", "Comment", COM),
+}
+
+
+def _accept(prefix: bytes) -> bool:
+    # Magic number was taken from https://en.wikipedia.org/wiki/JPEG
+    return prefix[:3] == b"\xFF\xD8\xFF"
+
+
+##
+# Image plugin for JPEG and JFIF images.
+
+
+class JpegImageFile(ImageFile.ImageFile):
+    format = "JPEG"
+    format_description = "JPEG (ISO 10918)"
+
+    def _open(self) -> None:
+        s = self.fp.read(3)
+
+        if not _accept(s):
+            msg = "not a JPEG file"
+            raise SyntaxError(msg)
+        s = b"\xFF"
+
+        # Create attributes
+        self.bits = self.layers = 0
+        self._exif_offset = 0
+
+        # JPEG specifics (internal)
+        self.layer: list[tuple[int, int, int, int]] = []
+        self._huffman_dc: dict[Any, Any] = {}
+        self._huffman_ac: dict[Any, Any] = {}
+        self.quantization: dict[int, list[int]] = {}
+        self.app: dict[str, bytes] = {}  # compatibility
+        self.applist: list[tuple[str, bytes]] = []
+        self.icclist: list[bytes] = []
+
+        while True:
+            i = s[0]
+            if i == 0xFF:
+                s = s + self.fp.read(1)
+                i = i16(s)
+            else:
+                # Skip non-0xFF junk
+                s = self.fp.read(1)
+                continue
+
+            if i in MARKER:
+                name, description, handler = MARKER[i]
+                if handler is not None:
+                    handler(self, i)
+                if i == 0xFFDA:  # start of scan
+                    rawmode = self.mode
+                    if self.mode == "CMYK":
+                        rawmode = "CMYK;I"  # assume adobe conventions
+                    self.tile = [
+                        ImageFile._Tile("jpeg", (0, 0) + self.size, 0, (rawmode, ""))
+                    ]
+                    # self.__offset = self.fp.tell()
+                    break
+                s = self.fp.read(1)
+            elif i in {0, 0xFFFF}:
+                # padded marker or junk; move on
+                s = b"\xff"
+            elif i == 0xFF00:  # Skip extraneous data (escaped 0xFF)
+                s = self.fp.read(1)
+            else:
+                msg = "no marker found"
+                raise SyntaxError(msg)
+
+        self._read_dpi_from_exif()
+
+    def __getattr__(self, name: str) -> Any:
+        if name in ("huffman_ac", "huffman_dc"):
+            deprecate(name, 12)
+            return getattr(self, "_" + name)
+        raise AttributeError(name)
+
+    def __getstate__(self) -> list[Any]:
+        return super().__getstate__() + [self.layers, self.layer]
+
+    def __setstate__(self, state: list[Any]) -> None:
+        super().__setstate__(state)
+        self.layers, self.layer = state[5:]
+
+    def load_read(self, read_bytes: int) -> bytes:
+        """
+        internal: read more image data
+        For premature EOF and LOAD_TRUNCATED_IMAGES adds EOI marker
+        so libjpeg can finish decoding
+        """
+        s = self.fp.read(read_bytes)
+
+        if not s and ImageFile.LOAD_TRUNCATED_IMAGES and not hasattr(self, "_ended"):
+            # Premature EOF.
+            # Pretend file is finished adding EOI marker
+            self._ended = True
+            return b"\xFF\xD9"
+
+        return s
+
+    def draft(
+        self, mode: str | None, size: tuple[int, int] | None
+    ) -> tuple[str, tuple[int, int, float, float]] | None:
+        if len(self.tile) != 1:
+            return None
+
+        # Protect from second call
+        if self.decoderconfig:
+            return None
+
+        d, e, o, a = self.tile[0]
+        scale = 1
+        original_size = self.size
+
+        assert isinstance(a, tuple)
+        if a[0] == "RGB" and mode in ["L", "YCbCr"]:
+            self._mode = mode
+            a = mode, ""
+
+        if size:
+            scale = min(self.size[0] // size[0], self.size[1] // size[1])
+            for s in [8, 4, 2, 1]:
+                if scale >= s:
+                    break
+            assert e is not None
+            e = (
+                e[0],
+                e[1],
+                (e[2] - e[0] + s - 1) // s + e[0],
+                (e[3] - e[1] + s - 1) // s + e[1],
+            )
+            self._size = ((self.size[0] + s - 1) // s, (self.size[1] + s - 1) // s)
+            scale = s
+
+        self.tile = [ImageFile._Tile(d, e, o, a)]
+        self.decoderconfig = (scale, 0)
+
+        box = (0, 0, original_size[0] / scale, original_size[1] / scale)
+        return self.mode, box
+
+    def load_djpeg(self) -> None:
+        # ALTERNATIVE: handle JPEGs via the IJG command line utilities
+
+        f, path = tempfile.mkstemp()
+        os.close(f)
+        if os.path.exists(self.filename):
+            subprocess.check_call(["djpeg", "-outfile", path, self.filename])
+        else:
+            try:
+                os.unlink(path)
+            except OSError:
+                pass
+
+            msg = "Invalid Filename"
+            raise ValueError(msg)
+
+        try:
+            with Image.open(path) as _im:
+                _im.load()
+                self.im = _im.im
+        finally:
+            try:
+                os.unlink(path)
+            except OSError:
+                pass
+
+        self._mode = self.im.mode
+        self._size = self.im.size
+
+        self.tile = []
+
+    def _getexif(self) -> dict[int, Any] | None:
+        return _getexif(self)
+
+    def _read_dpi_from_exif(self) -> None:
+        # If DPI isn't in JPEG header, fetch from EXIF
+        if "dpi" in self.info or "exif" not in self.info:
+            return
+        try:
+            exif = self.getexif()
+            resolution_unit = exif[0x0128]
+            x_resolution = exif[0x011A]
+            try:
+                dpi = float(x_resolution[0]) / x_resolution[1]
+            except TypeError:
+                dpi = x_resolution
+            if math.isnan(dpi):
+                msg = "DPI is not a number"
+                raise ValueError(msg)
+            if resolution_unit == 3:  # cm
+                # 1 dpcm = 2.54 dpi
+                dpi *= 2.54
+            self.info["dpi"] = dpi, dpi
+        except (
+            struct.error,  # truncated EXIF
+            KeyError,  # dpi not included
+            SyntaxError,  # invalid/unreadable EXIF
+            TypeError,  # dpi is an invalid float
+            ValueError,  # dpi is an invalid float
+            ZeroDivisionError,  # invalid dpi rational value
+        ):
+            self.info["dpi"] = 72, 72
+
+    def _getmp(self) -> dict[int, Any] | None:
+        return _getmp(self)
+
+
+def _getexif(self: JpegImageFile) -> dict[int, Any] | None:
+    if "exif" not in self.info:
+        return None
+    return self.getexif()._get_merged_dict()
+
+
+def _getmp(self: JpegImageFile) -> dict[int, Any] | None:
+    # Extract MP information.  This method was inspired by the "highly
+    # experimental" _getexif version that's been in use for years now,
+    # itself based on the ImageFileDirectory class in the TIFF plugin.
+
+    # The MP record essentially consists of a TIFF file embedded in a JPEG
+    # application marker.
+    try:
+        data = self.info["mp"]
+    except KeyError:
+        return None
+    file_contents = io.BytesIO(data)
+    head = file_contents.read(8)
+    endianness = ">" if head[:4] == b"\x4d\x4d\x00\x2a" else "<"
+    # process dictionary
+    from . import TiffImagePlugin
+
+    try:
+        info = TiffImagePlugin.ImageFileDirectory_v2(head)
+        file_contents.seek(info.next)
+        info.load(file_contents)
+        mp = dict(info)
+    except Exception as e:
+        msg = "malformed MP Index (unreadable directory)"
+        raise SyntaxError(msg) from e
+    # it's an error not to have a number of images
+    try:
+        quant = mp[0xB001]
+    except KeyError as e:
+        msg = "malformed MP Index (no number of images)"
+        raise SyntaxError(msg) from e
+    # get MP entries
+    mpentries = []
+    try:
+        rawmpentries = mp[0xB002]
+        for entrynum in range(0, quant):
+            unpackedentry = struct.unpack_from(
+                f"{endianness}LLLHH", rawmpentries, entrynum * 16
+            )
+            labels = ("Attribute", "Size", "DataOffset", "EntryNo1", "EntryNo2")
+            mpentry = dict(zip(labels, unpackedentry))
+            mpentryattr = {
+                "DependentParentImageFlag": bool(mpentry["Attribute"] & (1 << 31)),
+                "DependentChildImageFlag": bool(mpentry["Attribute"] & (1 << 30)),
+                "RepresentativeImageFlag": bool(mpentry["Attribute"] & (1 << 29)),
+                "Reserved": (mpentry["Attribute"] & (3 << 27)) >> 27,
+                "ImageDataFormat": (mpentry["Attribute"] & (7 << 24)) >> 24,
+                "MPType": mpentry["Attribute"] & 0x00FFFFFF,
+            }
+            if mpentryattr["ImageDataFormat"] == 0:
+                mpentryattr["ImageDataFormat"] = "JPEG"
+            else:
+                msg = "unsupported picture format in MPO"
+                raise SyntaxError(msg)
+            mptypemap = {
+                0x000000: "Undefined",
+                0x010001: "Large Thumbnail (VGA Equivalent)",
+                0x010002: "Large Thumbnail (Full HD Equivalent)",
+                0x020001: "Multi-Frame Image (Panorama)",
+                0x020002: "Multi-Frame Image: (Disparity)",
+                0x020003: "Multi-Frame Image: (Multi-Angle)",
+                0x030000: "Baseline MP Primary Image",
+            }
+            mpentryattr["MPType"] = mptypemap.get(mpentryattr["MPType"], "Unknown")
+            mpentry["Attribute"] = mpentryattr
+            mpentries.append(mpentry)
+        mp[0xB002] = mpentries
+    except KeyError as e:
+        msg = "malformed MP Index (bad MP Entry)"
+        raise SyntaxError(msg) from e
+    # Next we should try and parse the individual image unique ID list;
+    # we don't because I've never seen this actually used in a real MPO
+    # file and so can't test it.
+    return mp
+
+
+# --------------------------------------------------------------------
+# stuff to save JPEG files
+
+RAWMODE = {
+    "1": "L",
+    "L": "L",
+    "RGB": "RGB",
+    "RGBX": "RGB",
+    "CMYK": "CMYK;I",  # assume adobe conventions
+    "YCbCr": "YCbCr",
+}
+
+# fmt: off
+zigzag_index = (
+    0,  1,  5,  6, 14, 15, 27, 28,
+    2,  4,  7, 13, 16, 26, 29, 42,
+    3,  8, 12, 17, 25, 30, 41, 43,
+    9, 11, 18, 24, 31, 40, 44, 53,
+    10, 19, 23, 32, 39, 45, 52, 54,
+    20, 22, 33, 38, 46, 51, 55, 60,
+    21, 34, 37, 47, 50, 56, 59, 61,
+    35, 36, 48, 49, 57, 58, 62, 63,
+)
+
+samplings = {
+    (1, 1, 1, 1, 1, 1): 0,
+    (2, 1, 1, 1, 1, 1): 1,
+    (2, 2, 1, 1, 1, 1): 2,
+}
+# fmt: on
+
+
+def get_sampling(im: Image.Image) -> int:
+    # There's no subsampling when images have only 1 layer
+    # (grayscale images) or when they are CMYK (4 layers),
+    # so set subsampling to the default value.
+    #
+    # NOTE: currently Pillow can't encode JPEG to YCCK format.
+    # If YCCK support is added in the future, subsampling code will have
+    # to be updated (here and in JpegEncode.c) to deal with 4 layers.
+    if not isinstance(im, JpegImageFile) or im.layers in (1, 4):
+        return -1
+    sampling = im.layer[0][1:3] + im.layer[1][1:3] + im.layer[2][1:3]
+    return samplings.get(sampling, -1)
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.width == 0 or im.height == 0:
+        msg = "cannot write empty image as JPEG"
+        raise ValueError(msg)
+
+    try:
+        rawmode = RAWMODE[im.mode]
+    except KeyError as e:
+        msg = f"cannot write mode {im.mode} as JPEG"
+        raise OSError(msg) from e
+
+    info = im.encoderinfo
+
+    dpi = [round(x) for x in info.get("dpi", (0, 0))]
+
+    quality = info.get("quality", -1)
+    subsampling = info.get("subsampling", -1)
+    qtables = info.get("qtables")
+
+    if quality == "keep":
+        quality = -1
+        subsampling = "keep"
+        qtables = "keep"
+    elif quality in presets:
+        preset = presets[quality]
+        quality = -1
+        subsampling = preset.get("subsampling", -1)
+        qtables = preset.get("quantization")
+    elif not isinstance(quality, int):
+        msg = "Invalid quality setting"
+        raise ValueError(msg)
+    else:
+        if subsampling in presets:
+            subsampling = presets[subsampling].get("subsampling", -1)
+        if isinstance(qtables, str) and qtables in presets:
+            qtables = presets[qtables].get("quantization")
+
+    if subsampling == "4:4:4":
+        subsampling = 0
+    elif subsampling == "4:2:2":
+        subsampling = 1
+    elif subsampling == "4:2:0":
+        subsampling = 2
+    elif subsampling == "4:1:1":
+        # For compatibility. Before Pillow 4.3, 4:1:1 actually meant 4:2:0.
+        # Set 4:2:0 if someone is still using that value.
+        subsampling = 2
+    elif subsampling == "keep":
+        if im.format != "JPEG":
+            msg = "Cannot use 'keep' when original image is not a JPEG"
+            raise ValueError(msg)
+        subsampling = get_sampling(im)
+
+    def validate_qtables(
+        qtables: (
+            str | tuple[list[int], ...] | list[list[int]] | dict[int, list[int]] | None
+        )
+    ) -> list[list[int]] | None:
+        if qtables is None:
+            return qtables
+        if isinstance(qtables, str):
+            try:
+                lines = [
+                    int(num)
+                    for line in qtables.splitlines()
+                    for num in line.split("#", 1)[0].split()
+                ]
+            except ValueError as e:
+                msg = "Invalid quantization table"
+                raise ValueError(msg) from e
+            else:
+                qtables = [lines[s : s + 64] for s in range(0, len(lines), 64)]
+        if isinstance(qtables, (tuple, list, dict)):
+            if isinstance(qtables, dict):
+                qtables = [
+                    qtables[key] for key in range(len(qtables)) if key in qtables
+                ]
+            elif isinstance(qtables, tuple):
+                qtables = list(qtables)
+            if not (0 < len(qtables) < 5):
+                msg = "None or too many quantization tables"
+                raise ValueError(msg)
+            for idx, table in enumerate(qtables):
+                try:
+                    if len(table) != 64:
+                        msg = "Invalid quantization table"
+                        raise TypeError(msg)
+                    table_array = array.array("H", table)
+                except TypeError as e:
+                    msg = "Invalid quantization table"
+                    raise ValueError(msg) from e
+                else:
+                    qtables[idx] = list(table_array)
+            return qtables
+
+    if qtables == "keep":
+        if im.format != "JPEG":
+            msg = "Cannot use 'keep' when original image is not a JPEG"
+            raise ValueError(msg)
+        qtables = getattr(im, "quantization", None)
+    qtables = validate_qtables(qtables)
+
+    extra = info.get("extra", b"")
+
+    MAX_BYTES_IN_MARKER = 65533
+    xmp = info.get("xmp")
+    if xmp:
+        overhead_len = 29  # b"http://ns.adobe.com/xap/1.0/\x00"
+        max_data_bytes_in_marker = MAX_BYTES_IN_MARKER - overhead_len
+        if len(xmp) > max_data_bytes_in_marker:
+            msg = "XMP data is too long"
+            raise ValueError(msg)
+        size = o16(2 + overhead_len + len(xmp))
+        extra += b"\xFF\xE1" + size + b"http://ns.adobe.com/xap/1.0/\x00" + xmp
+
+    icc_profile = info.get("icc_profile")
+    if icc_profile:
+        overhead_len = 14  # b"ICC_PROFILE\0" + o8(i) + o8(len(markers))
+        max_data_bytes_in_marker = MAX_BYTES_IN_MARKER - overhead_len
+        markers = []
+        while icc_profile:
+            markers.append(icc_profile[:max_data_bytes_in_marker])
+            icc_profile = icc_profile[max_data_bytes_in_marker:]
+        i = 1
+        for marker in markers:
+            size = o16(2 + overhead_len + len(marker))
+            extra += (
+                b"\xFF\xE2"
+                + size
+                + b"ICC_PROFILE\0"
+                + o8(i)
+                + o8(len(markers))
+                + marker
+            )
+            i += 1
+
+    comment = info.get("comment", im.info.get("comment"))
+
+    # "progressive" is the official name, but older documentation
+    # says "progression"
+    # FIXME: issue a warning if the wrong form is used (post-1.1.7)
+    progressive = info.get("progressive", False) or info.get("progression", False)
+
+    optimize = info.get("optimize", False)
+
+    exif = info.get("exif", b"")
+    if isinstance(exif, Image.Exif):
+        exif = exif.tobytes()
+    if len(exif) > MAX_BYTES_IN_MARKER:
+        msg = "EXIF data is too long"
+        raise ValueError(msg)
+
+    # get keyword arguments
+    im.encoderconfig = (
+        quality,
+        progressive,
+        info.get("smooth", 0),
+        optimize,
+        info.get("keep_rgb", False),
+        info.get("streamtype", 0),
+        dpi[0],
+        dpi[1],
+        subsampling,
+        info.get("restart_marker_blocks", 0),
+        info.get("restart_marker_rows", 0),
+        qtables,
+        comment,
+        extra,
+        exif,
+    )
+
+    # if we optimize, libjpeg needs a buffer big enough to hold the whole image
+    # in a shot. Guessing on the size, at im.size bytes. (raw pixel size is
+    # channels*size, this is a value that's been used in a django patch.
+    # https://github.com/matthewwithanm/django-imagekit/issues/50
+    bufsize = 0
+    if optimize or progressive:
+        # CMYK can be bigger
+        if im.mode == "CMYK":
+            bufsize = 4 * im.size[0] * im.size[1]
+        # keep sets quality to -1, but the actual value may be high.
+        elif quality >= 95 or quality == -1:
+            bufsize = 2 * im.size[0] * im.size[1]
+        else:
+            bufsize = im.size[0] * im.size[1]
+        if exif:
+            bufsize += len(exif) + 5
+        if extra:
+            bufsize += len(extra) + 1
+    else:
+        # The EXIF info needs to be written as one block, + APP1, + one spare byte.
+        # Ensure that our buffer is big enough. Same with the icc_profile block.
+        bufsize = max(bufsize, len(exif) + 5, len(extra) + 1)
+
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("jpeg", (0, 0) + im.size, 0, rawmode)], bufsize
+    )
+
+
+def _save_cjpeg(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    # ALTERNATIVE: handle JPEGs via the IJG command line utilities.
+    tempfile = im._dump()
+    subprocess.check_call(["cjpeg", "-outfile", filename, tempfile])
+    try:
+        os.unlink(tempfile)
+    except OSError:
+        pass
+
+
+##
+# Factory for making JPEG and MPO instances
+def jpeg_factory(
+    fp: IO[bytes], filename: str | bytes | None = None
+) -> JpegImageFile | MpoImageFile:
+    im = JpegImageFile(fp, filename)
+    try:
+        mpheader = im._getmp()
+        if mpheader is not None and mpheader[45057] > 1:
+            for segment, content in im.applist:
+                if segment == "APP1" and b' hdrgm:Version="' in content:
+                    # Ultra HDR images are not yet supported
+                    return im
+            # It's actually an MPO
+            from .MpoImagePlugin import MpoImageFile
+
+            # Don't reload everything, just convert it.
+            im = MpoImageFile.adopt(im, mpheader)
+    except (TypeError, IndexError):
+        # It is really a JPEG
+        pass
+    except SyntaxError:
+        warnings.warn(
+            "Image appears to be a malformed MPO file, it will be "
+            "interpreted as a base JPEG file"
+        )
+    return im
+
+
+# ---------------------------------------------------------------------
+# Registry stuff
+
+Image.register_open(JpegImageFile.format, jpeg_factory, _accept)
+Image.register_save(JpegImageFile.format, _save)
+
+Image.register_extensions(JpegImageFile.format, [".jfif", ".jpe", ".jpg", ".jpeg"])
+
+Image.register_mime(JpegImageFile.format, "image/jpeg")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/JpegPresets.py b/backend/venv/lib/python3.12/site-packages/PIL/JpegPresets.py
new file mode 100644
index 00000000..d0e64a35
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/JpegPresets.py
@@ -0,0 +1,242 @@
+"""
+JPEG quality settings equivalent to the Photoshop settings.
+Can be used when saving JPEG files.
+
+The following presets are available by default:
+``web_low``, ``web_medium``, ``web_high``, ``web_very_high``, ``web_maximum``,
+``low``, ``medium``, ``high``, ``maximum``.
+More presets can be added to the :py:data:`presets` dict if needed.
+
+To apply the preset, specify::
+
+  quality="preset_name"
+
+To apply only the quantization table::
+
+  qtables="preset_name"
+
+To apply only the subsampling setting::
+
+  subsampling="preset_name"
+
+Example::
+
+  im.save("image_name.jpg", quality="web_high")
+
+Subsampling
+-----------
+
+Subsampling is the practice of encoding images by implementing less resolution
+for chroma information than for luma information.
+(ref.: https://en.wikipedia.org/wiki/Chroma_subsampling)
+
+Possible subsampling values are 0, 1 and 2 that correspond to 4:4:4, 4:2:2 and
+4:2:0.
+
+You can get the subsampling of a JPEG with the
+:func:`.JpegImagePlugin.get_sampling` function.
+
+In JPEG compressed data a JPEG marker is used instead of an EXIF tag.
+(ref.: https://exiv2.org/tags.html)
+
+
+Quantization tables
+-------------------
+
+They are values use by the DCT (Discrete cosine transform) to remove
+*unnecessary* information from the image (the lossy part of the compression).
+(ref.: https://en.wikipedia.org/wiki/Quantization_matrix#Quantization_matrices,
+https://en.wikipedia.org/wiki/JPEG#Quantization)
+
+You can get the quantization tables of a JPEG with::
+
+  im.quantization
+
+This will return a dict with a number of lists. You can pass this dict
+directly as the qtables argument when saving a JPEG.
+
+The quantization table format in presets is a list with sublists. These formats
+are interchangeable.
+
+Libjpeg ref.:
+https://web.archive.org/web/20120328125543/http://www.jpegcameras.com/libjpeg/libjpeg-3.html
+
+"""
+
+from __future__ import annotations
+
+# fmt: off
+presets = {
+            'web_low':      {'subsampling':  2,  # "4:2:0"
+                             'quantization': [
+                               [20, 16, 25, 39, 50, 46, 62, 68,
+                                16, 18, 23, 38, 38, 53, 65, 68,
+                                25, 23, 31, 38, 53, 65, 68, 68,
+                                39, 38, 38, 53, 65, 68, 68, 68,
+                                50, 38, 53, 65, 68, 68, 68, 68,
+                                46, 53, 65, 68, 68, 68, 68, 68,
+                                62, 65, 68, 68, 68, 68, 68, 68,
+                                68, 68, 68, 68, 68, 68, 68, 68],
+                               [21, 25, 32, 38, 54, 68, 68, 68,
+                                25, 28, 24, 38, 54, 68, 68, 68,
+                                32, 24, 32, 43, 66, 68, 68, 68,
+                                38, 38, 43, 53, 68, 68, 68, 68,
+                                54, 54, 66, 68, 68, 68, 68, 68,
+                                68, 68, 68, 68, 68, 68, 68, 68,
+                                68, 68, 68, 68, 68, 68, 68, 68,
+                                68, 68, 68, 68, 68, 68, 68, 68]
+                              ]},
+            'web_medium':   {'subsampling':  2,  # "4:2:0"
+                             'quantization': [
+                               [16, 11, 11, 16, 23, 27, 31, 30,
+                                11, 12, 12, 15, 20, 23, 23, 30,
+                                11, 12, 13, 16, 23, 26, 35, 47,
+                                16, 15, 16, 23, 26, 37, 47, 64,
+                                23, 20, 23, 26, 39, 51, 64, 64,
+                                27, 23, 26, 37, 51, 64, 64, 64,
+                                31, 23, 35, 47, 64, 64, 64, 64,
+                                30, 30, 47, 64, 64, 64, 64, 64],
+                               [17, 15, 17, 21, 20, 26, 38, 48,
+                                15, 19, 18, 17, 20, 26, 35, 43,
+                                17, 18, 20, 22, 26, 30, 46, 53,
+                                21, 17, 22, 28, 30, 39, 53, 64,
+                                20, 20, 26, 30, 39, 48, 64, 64,
+                                26, 26, 30, 39, 48, 63, 64, 64,
+                                38, 35, 46, 53, 64, 64, 64, 64,
+                                48, 43, 53, 64, 64, 64, 64, 64]
+                             ]},
+            'web_high':     {'subsampling':  0,  # "4:4:4"
+                             'quantization': [
+                               [6,   4,  4,  6,  9, 11, 12, 16,
+                                4,   5,  5,  6,  8, 10, 12, 12,
+                                4,   5,  5,  6, 10, 12, 14, 19,
+                                6,   6,  6, 11, 12, 15, 19, 28,
+                                9,   8, 10, 12, 16, 20, 27, 31,
+                                11, 10, 12, 15, 20, 27, 31, 31,
+                                12, 12, 14, 19, 27, 31, 31, 31,
+                                16, 12, 19, 28, 31, 31, 31, 31],
+                               [7,   7, 13, 24, 26, 31, 31, 31,
+                                7,  12, 16, 21, 31, 31, 31, 31,
+                                13, 16, 17, 31, 31, 31, 31, 31,
+                                24, 21, 31, 31, 31, 31, 31, 31,
+                                26, 31, 31, 31, 31, 31, 31, 31,
+                                31, 31, 31, 31, 31, 31, 31, 31,
+                                31, 31, 31, 31, 31, 31, 31, 31,
+                                31, 31, 31, 31, 31, 31, 31, 31]
+                             ]},
+            'web_very_high': {'subsampling':  0,  # "4:4:4"
+                              'quantization': [
+                               [2,   2,  2,  2,  3,  4,  5,  6,
+                                2,   2,  2,  2,  3,  4,  5,  6,
+                                2,   2,  2,  2,  4,  5,  7,  9,
+                                2,   2,  2,  4,  5,  7,  9, 12,
+                                3,   3,  4,  5,  8, 10, 12, 12,
+                                4,   4,  5,  7, 10, 12, 12, 12,
+                                5,   5,  7,  9, 12, 12, 12, 12,
+                                6,   6,  9, 12, 12, 12, 12, 12],
+                               [3,   3,  5,  9, 13, 15, 15, 15,
+                                3,   4,  6, 11, 14, 12, 12, 12,
+                                5,   6,  9, 14, 12, 12, 12, 12,
+                                9,  11, 14, 12, 12, 12, 12, 12,
+                                13, 14, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12]
+                              ]},
+            'web_maximum':  {'subsampling':  0,  # "4:4:4"
+                             'quantization': [
+                                [1,  1,  1,  1,  1,  1,  1,  1,
+                                 1,  1,  1,  1,  1,  1,  1,  1,
+                                 1,  1,  1,  1,  1,  1,  1,  2,
+                                 1,  1,  1,  1,  1,  1,  2,  2,
+                                 1,  1,  1,  1,  1,  2,  2,  3,
+                                 1,  1,  1,  1,  2,  2,  3,  3,
+                                 1,  1,  1,  2,  2,  3,  3,  3,
+                                 1,  1,  2,  2,  3,  3,  3,  3],
+                                [1,  1,  1,  2,  2,  3,  3,  3,
+                                 1,  1,  1,  2,  3,  3,  3,  3,
+                                 1,  1,  1,  3,  3,  3,  3,  3,
+                                 2,  2,  3,  3,  3,  3,  3,  3,
+                                 2,  3,  3,  3,  3,  3,  3,  3,
+                                 3,  3,  3,  3,  3,  3,  3,  3,
+                                 3,  3,  3,  3,  3,  3,  3,  3,
+                                 3,  3,  3,  3,  3,  3,  3,  3]
+                             ]},
+            'low':          {'subsampling':  2,  # "4:2:0"
+                             'quantization': [
+                               [18, 14, 14, 21, 30, 35, 34, 17,
+                                14, 16, 16, 19, 26, 23, 12, 12,
+                                14, 16, 17, 21, 23, 12, 12, 12,
+                                21, 19, 21, 23, 12, 12, 12, 12,
+                                30, 26, 23, 12, 12, 12, 12, 12,
+                                35, 23, 12, 12, 12, 12, 12, 12,
+                                34, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12],
+                               [20, 19, 22, 27, 20, 20, 17, 17,
+                                19, 25, 23, 14, 14, 12, 12, 12,
+                                22, 23, 14, 14, 12, 12, 12, 12,
+                                27, 14, 14, 12, 12, 12, 12, 12,
+                                20, 14, 12, 12, 12, 12, 12, 12,
+                                20, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12]
+                             ]},
+            'medium':       {'subsampling':  2,  # "4:2:0"
+                             'quantization': [
+                               [12,  8,  8, 12, 17, 21, 24, 17,
+                                8,   9,  9, 11, 15, 19, 12, 12,
+                                8,   9, 10, 12, 19, 12, 12, 12,
+                                12, 11, 12, 21, 12, 12, 12, 12,
+                                17, 15, 19, 12, 12, 12, 12, 12,
+                                21, 19, 12, 12, 12, 12, 12, 12,
+                                24, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12],
+                               [13, 11, 13, 16, 20, 20, 17, 17,
+                                11, 14, 14, 14, 14, 12, 12, 12,
+                                13, 14, 14, 14, 12, 12, 12, 12,
+                                16, 14, 14, 12, 12, 12, 12, 12,
+                                20, 14, 12, 12, 12, 12, 12, 12,
+                                20, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12]
+                             ]},
+            'high':         {'subsampling':  0,  # "4:4:4"
+                             'quantization': [
+                               [6,   4,  4,  6,  9, 11, 12, 16,
+                                4,   5,  5,  6,  8, 10, 12, 12,
+                                4,   5,  5,  6, 10, 12, 12, 12,
+                                6,   6,  6, 11, 12, 12, 12, 12,
+                                9,   8, 10, 12, 12, 12, 12, 12,
+                                11, 10, 12, 12, 12, 12, 12, 12,
+                                12, 12, 12, 12, 12, 12, 12, 12,
+                                16, 12, 12, 12, 12, 12, 12, 12],
+                               [7,   7, 13, 24, 20, 20, 17, 17,
+                                7,  12, 16, 14, 14, 12, 12, 12,
+                                13, 16, 14, 14, 12, 12, 12, 12,
+                                24, 14, 14, 12, 12, 12, 12, 12,
+                                20, 14, 12, 12, 12, 12, 12, 12,
+                                20, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12,
+                                17, 12, 12, 12, 12, 12, 12, 12]
+                             ]},
+            'maximum':      {'subsampling':  0,  # "4:4:4"
+                             'quantization': [
+                               [2,   2,  2,  2,  3,  4,  5,  6,
+                                2,   2,  2,  2,  3,  4,  5,  6,
+                                2,   2,  2,  2,  4,  5,  7,  9,
+                                2,   2,  2,  4,  5,  7,  9, 12,
+                                3,   3,  4,  5,  8, 10, 12, 12,
+                                4,   4,  5,  7, 10, 12, 12, 12,
+                                5,   5,  7,  9, 12, 12, 12, 12,
+                                6,   6,  9, 12, 12, 12, 12, 12],
+                               [3,   3,  5,  9, 13, 15, 15, 15,
+                                3,   4,  6, 10, 14, 12, 12, 12,
+                                5,   6,  9, 14, 12, 12, 12, 12,
+                                9,  10, 14, 12, 12, 12, 12, 12,
+                                13, 14, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12,
+                                15, 12, 12, 12, 12, 12, 12, 12]
+                             ]},
+}
+# fmt: on
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/McIdasImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/McIdasImagePlugin.py
new file mode 100644
index 00000000..5dd031be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/McIdasImagePlugin.py
@@ -0,0 +1,80 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Basic McIdas support for PIL
+#
+# History:
+# 1997-05-05 fl  Created (8-bit images only)
+# 2009-03-08 fl  Added 16/32-bit support.
+#
+# Thanks to Richard Jones and Craig Swank for specs and samples.
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import struct
+
+from . import Image, ImageFile
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:8] == b"\x00\x00\x00\x00\x00\x00\x00\x04"
+
+
+##
+# Image plugin for McIdas area images.
+
+
+class McIdasImageFile(ImageFile.ImageFile):
+    format = "MCIDAS"
+    format_description = "McIdas area file"
+
+    def _open(self) -> None:
+        # parse area file directory
+        assert self.fp is not None
+
+        s = self.fp.read(256)
+        if not _accept(s) or len(s) != 256:
+            msg = "not an McIdas area file"
+            raise SyntaxError(msg)
+
+        self.area_descriptor_raw = s
+        self.area_descriptor = w = [0] + list(struct.unpack("!64i", s))
+
+        # get mode
+        if w[11] == 1:
+            mode = rawmode = "L"
+        elif w[11] == 2:
+            # FIXME: add memory map support
+            mode = "I"
+            rawmode = "I;16B"
+        elif w[11] == 4:
+            # FIXME: add memory map support
+            mode = "I"
+            rawmode = "I;32B"
+        else:
+            msg = "unsupported McIdas format"
+            raise SyntaxError(msg)
+
+        self._mode = mode
+        self._size = w[10], w[9]
+
+        offset = w[34] + w[15]
+        stride = w[15] + w[10] * w[11] * w[14]
+
+        self.tile = [
+            ImageFile._Tile("raw", (0, 0) + self.size, offset, (rawmode, stride, 1))
+        ]
+
+
+# --------------------------------------------------------------------
+# registry
+
+Image.register_open(McIdasImageFile.format, McIdasImageFile, _accept)
+
+# no default extension
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/MicImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/MicImagePlugin.py
new file mode 100644
index 00000000..5f23a34b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/MicImagePlugin.py
@@ -0,0 +1,107 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Microsoft Image Composer support for PIL
+#
+# Notes:
+#       uses TiffImagePlugin.py to read the actual image streams
+#
+# History:
+#       97-01-20 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import olefile
+
+from . import Image, TiffImagePlugin
+
+#
+# --------------------------------------------------------------------
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:8] == olefile.MAGIC
+
+
+##
+# Image plugin for Microsoft's Image Composer file format.
+
+
+class MicImageFile(TiffImagePlugin.TiffImageFile):
+    format = "MIC"
+    format_description = "Microsoft Image Composer"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        # read the OLE directory and see if this is a likely
+        # to be a Microsoft Image Composer file
+
+        try:
+            self.ole = olefile.OleFileIO(self.fp)
+        except OSError as e:
+            msg = "not an MIC file; invalid OLE file"
+            raise SyntaxError(msg) from e
+
+        # find ACI subfiles with Image members (maybe not the
+        # best way to identify MIC files, but what the... ;-)
+
+        self.images = [
+            path
+            for path in self.ole.listdir()
+            if path[1:] and path[0][-4:] == ".ACI" and path[1] == "Image"
+        ]
+
+        # if we didn't find any images, this is probably not
+        # an MIC file.
+        if not self.images:
+            msg = "not an MIC file; no image entries"
+            raise SyntaxError(msg)
+
+        self.frame = -1
+        self._n_frames = len(self.images)
+        self.is_animated = self._n_frames > 1
+
+        self.__fp = self.fp
+        self.seek(0)
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        try:
+            filename = self.images[frame]
+        except IndexError as e:
+            msg = "no such frame"
+            raise EOFError(msg) from e
+
+        self.fp = self.ole.openstream(filename)
+
+        TiffImagePlugin.TiffImageFile._open(self)
+
+        self.frame = frame
+
+    def tell(self) -> int:
+        return self.frame
+
+    def close(self) -> None:
+        self.__fp.close()
+        self.ole.close()
+        super().close()
+
+    def __exit__(self, *args: object) -> None:
+        self.__fp.close()
+        self.ole.close()
+        super().__exit__()
+
+
+#
+# --------------------------------------------------------------------
+
+Image.register_open(MicImageFile.format, MicImageFile, _accept)
+
+Image.register_extension(MicImageFile.format, ".mic")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/MpegImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/MpegImagePlugin.py
new file mode 100644
index 00000000..ad4d3e93
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/MpegImagePlugin.py
@@ -0,0 +1,88 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# MPEG file handling
+#
+# History:
+#       95-09-09 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1995.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image, ImageFile
+from ._binary import i8
+from ._typing import SupportsRead
+
+#
+# Bitstream parser
+
+
+class BitStream:
+    def __init__(self, fp: SupportsRead[bytes]) -> None:
+        self.fp = fp
+        self.bits = 0
+        self.bitbuffer = 0
+
+    def next(self) -> int:
+        return i8(self.fp.read(1))
+
+    def peek(self, bits: int) -> int:
+        while self.bits < bits:
+            c = self.next()
+            if c < 0:
+                self.bits = 0
+                continue
+            self.bitbuffer = (self.bitbuffer << 8) + c
+            self.bits += 8
+        return self.bitbuffer >> (self.bits - bits) & (1 << bits) - 1
+
+    def skip(self, bits: int) -> None:
+        while self.bits < bits:
+            self.bitbuffer = (self.bitbuffer << 8) + i8(self.fp.read(1))
+            self.bits += 8
+        self.bits = self.bits - bits
+
+    def read(self, bits: int) -> int:
+        v = self.peek(bits)
+        self.bits = self.bits - bits
+        return v
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"\x00\x00\x01\xb3"
+
+
+##
+# Image plugin for MPEG streams.  This plugin can identify a stream,
+# but it cannot read it.
+
+
+class MpegImageFile(ImageFile.ImageFile):
+    format = "MPEG"
+    format_description = "MPEG"
+
+    def _open(self) -> None:
+        assert self.fp is not None
+
+        s = BitStream(self.fp)
+        if s.read(32) != 0x1B3:
+            msg = "not an MPEG file"
+            raise SyntaxError(msg)
+
+        self._mode = "RGB"
+        self._size = s.read(12), s.read(12)
+
+
+# --------------------------------------------------------------------
+# Registry stuff
+
+Image.register_open(MpegImageFile.format, MpegImageFile, _accept)
+
+Image.register_extensions(MpegImageFile.format, [".mpg", ".mpeg"])
+
+Image.register_mime(MpegImageFile.format, "video/mpeg")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/MpoImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/MpoImagePlugin.py
new file mode 100644
index 00000000..71f89a09
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/MpoImagePlugin.py
@@ -0,0 +1,190 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# MPO file handling
+#
+# See "Multi-Picture Format" (CIPA DC-007-Translation 2009, Standard of the
+# Camera & Imaging Products Association)
+#
+# The multi-picture object combines multiple JPEG images (with a modified EXIF
+# data format) into a single file. While it can theoretically be used much like
+# a GIF animation, it is commonly used to represent 3D photographs and is (as
+# of this writing) the most commonly used format by 3D cameras.
+#
+# History:
+# 2014-03-13 Feneric   Created
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import itertools
+import os
+import struct
+from typing import IO, Any, cast
+
+from . import (
+    Image,
+    ImageFile,
+    ImageSequence,
+    JpegImagePlugin,
+    TiffImagePlugin,
+)
+from ._binary import o32le
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    JpegImagePlugin._save(im, fp, filename)
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    append_images = im.encoderinfo.get("append_images", [])
+    if not append_images and not getattr(im, "is_animated", False):
+        _save(im, fp, filename)
+        return
+
+    mpf_offset = 28
+    offsets: list[int] = []
+    for imSequence in itertools.chain([im], append_images):
+        for im_frame in ImageSequence.Iterator(imSequence):
+            if not offsets:
+                # APP2 marker
+                im_frame.encoderinfo["extra"] = (
+                    b"\xFF\xE2" + struct.pack(">H", 6 + 82) + b"MPF\0" + b" " * 82
+                )
+                exif = im_frame.encoderinfo.get("exif")
+                if isinstance(exif, Image.Exif):
+                    exif = exif.tobytes()
+                    im_frame.encoderinfo["exif"] = exif
+                if exif:
+                    mpf_offset += 4 + len(exif)
+
+                JpegImagePlugin._save(im_frame, fp, filename)
+                offsets.append(fp.tell())
+            else:
+                im_frame.save(fp, "JPEG")
+                offsets.append(fp.tell() - offsets[-1])
+
+    ifd = TiffImagePlugin.ImageFileDirectory_v2()
+    ifd[0xB000] = b"0100"
+    ifd[0xB001] = len(offsets)
+
+    mpentries = b""
+    data_offset = 0
+    for i, size in enumerate(offsets):
+        if i == 0:
+            mptype = 0x030000  # Baseline MP Primary Image
+        else:
+            mptype = 0x000000  # Undefined
+        mpentries += struct.pack("<LLLHH", mptype, size, data_offset, 0, 0)
+        if i == 0:
+            data_offset -= mpf_offset
+        data_offset += size
+    ifd[0xB002] = mpentries
+
+    fp.seek(mpf_offset)
+    fp.write(b"II\x2A\x00" + o32le(8) + ifd.tobytes(8))
+    fp.seek(0, os.SEEK_END)
+
+
+##
+# Image plugin for MPO images.
+
+
+class MpoImageFile(JpegImagePlugin.JpegImageFile):
+    format = "MPO"
+    format_description = "MPO (CIPA DC-007)"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        self.fp.seek(0)  # prep the fp in order to pass the JPEG test
+        JpegImagePlugin.JpegImageFile._open(self)
+        self._after_jpeg_open()
+
+    def _after_jpeg_open(self, mpheader: dict[int, Any] | None = None) -> None:
+        self.mpinfo = mpheader if mpheader is not None else self._getmp()
+        if self.mpinfo is None:
+            msg = "Image appears to be a malformed MPO file"
+            raise ValueError(msg)
+        self.n_frames = self.mpinfo[0xB001]
+        self.__mpoffsets = [
+            mpent["DataOffset"] + self.info["mpoffset"] for mpent in self.mpinfo[0xB002]
+        ]
+        self.__mpoffsets[0] = 0
+        # Note that the following assertion will only be invalid if something
+        # gets broken within JpegImagePlugin.
+        assert self.n_frames == len(self.__mpoffsets)
+        del self.info["mpoffset"]  # no longer needed
+        self.is_animated = self.n_frames > 1
+        self._fp = self.fp  # FIXME: hack
+        self._fp.seek(self.__mpoffsets[0])  # get ready to read first frame
+        self.__frame = 0
+        self.offset = 0
+        # for now we can only handle reading and individual frame extraction
+        self.readonly = 1
+
+    def load_seek(self, pos: int) -> None:
+        self._fp.seek(pos)
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        self.fp = self._fp
+        self.offset = self.__mpoffsets[frame]
+
+        original_exif = self.info.get("exif")
+        if "exif" in self.info:
+            del self.info["exif"]
+
+        self.fp.seek(self.offset + 2)  # skip SOI marker
+        if not self.fp.read(2):
+            msg = "No data found for frame"
+            raise ValueError(msg)
+        self.fp.seek(self.offset)
+        JpegImagePlugin.JpegImageFile._open(self)
+        if self.info.get("exif") != original_exif:
+            self._reload_exif()
+
+        self.tile = [
+            ImageFile._Tile("jpeg", (0, 0) + self.size, self.offset, self.tile[0][-1])
+        ]
+        self.__frame = frame
+
+    def tell(self) -> int:
+        return self.__frame
+
+    @staticmethod
+    def adopt(
+        jpeg_instance: JpegImagePlugin.JpegImageFile,
+        mpheader: dict[int, Any] | None = None,
+    ) -> MpoImageFile:
+        """
+        Transform the instance of JpegImageFile into
+        an instance of MpoImageFile.
+        After the call, the JpegImageFile is extended
+        to be an MpoImageFile.
+
+        This is essentially useful when opening a JPEG
+        file that reveals itself as an MPO, to avoid
+        double call to _open.
+        """
+        jpeg_instance.__class__ = MpoImageFile
+        mpo_instance = cast(MpoImageFile, jpeg_instance)
+        mpo_instance._after_jpeg_open(mpheader)
+        return mpo_instance
+
+
+# ---------------------------------------------------------------------
+# Registry stuff
+
+# Note that since MPO shares a factory with JPEG, we do not need to do a
+# separate registration for it here.
+# Image.register_open(MpoImageFile.format,
+#                     JpegImagePlugin.jpeg_factory, _accept)
+Image.register_save(MpoImageFile.format, _save)
+Image.register_save_all(MpoImageFile.format, _save_all)
+
+Image.register_extension(MpoImageFile.format, ".mpo")
+
+Image.register_mime(MpoImageFile.format, "image/mpo")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/MspImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/MspImagePlugin.py
new file mode 100644
index 00000000..ef6ae87f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/MspImagePlugin.py
@@ -0,0 +1,200 @@
+#
+# The Python Imaging Library.
+#
+# MSP file handling
+#
+# This is the format used by the Paint program in Windows 1 and 2.
+#
+# History:
+#       95-09-05 fl     Created
+#       97-01-03 fl     Read/write MSP images
+#       17-02-21 es     Fixed RLE interpretation
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1995-97.
+# Copyright (c) Eric Soroos 2017.
+#
+# See the README file for information on usage and redistribution.
+#
+# More info on this format: https://archive.org/details/gg243631
+# Page 313:
+# Figure 205. Windows Paint Version 1: "DanM" Format
+# Figure 206. Windows Paint Version 2: "LinS" Format. Used in Windows V2.03
+#
+# See also: https://www.fileformat.info/format/mspaint/egff.htm
+from __future__ import annotations
+
+import io
+import struct
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i16le as i16
+from ._binary import o16le as o16
+
+#
+# read MSP files
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] in [b"DanM", b"LinS"]
+
+
+##
+# Image plugin for Windows MSP images.  This plugin supports both
+# uncompressed (Windows 1.0).
+
+
+class MspImageFile(ImageFile.ImageFile):
+    format = "MSP"
+    format_description = "Windows Paint"
+
+    def _open(self) -> None:
+        # Header
+        assert self.fp is not None
+
+        s = self.fp.read(32)
+        if not _accept(s):
+            msg = "not an MSP file"
+            raise SyntaxError(msg)
+
+        # Header checksum
+        checksum = 0
+        for i in range(0, 32, 2):
+            checksum = checksum ^ i16(s, i)
+        if checksum != 0:
+            msg = "bad MSP checksum"
+            raise SyntaxError(msg)
+
+        self._mode = "1"
+        self._size = i16(s, 4), i16(s, 6)
+
+        if s[:4] == b"DanM":
+            self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, 32, "1")]
+        else:
+            self.tile = [ImageFile._Tile("MSP", (0, 0) + self.size, 32)]
+
+
+class MspDecoder(ImageFile.PyDecoder):
+    # The algo for the MSP decoder is from
+    # https://www.fileformat.info/format/mspaint/egff.htm
+    # cc-by-attribution -- That page references is taken from the
+    # Encyclopedia of Graphics File Formats and is licensed by
+    # O'Reilly under the Creative Common/Attribution license
+    #
+    # For RLE encoded files, the 32byte header is followed by a scan
+    # line map, encoded as one 16bit word of encoded byte length per
+    # line.
+    #
+    # NOTE: the encoded length of the line can be 0. This was not
+    # handled in the previous version of this encoder, and there's no
+    # mention of how to handle it in the documentation. From the few
+    # examples I've seen, I've assumed that it is a fill of the
+    # background color, in this case, white.
+    #
+    #
+    # Pseudocode of the decoder:
+    # Read a BYTE value as the RunType
+    #  If the RunType value is zero
+    #   Read next byte as the RunCount
+    #   Read the next byte as the RunValue
+    #   Write the RunValue byte RunCount times
+    #  If the RunType value is non-zero
+    #   Use this value as the RunCount
+    #   Read and write the next RunCount bytes literally
+    #
+    #  e.g.:
+    #  0x00 03 ff 05 00 01 02 03 04
+    #  would yield the bytes:
+    #  0xff ff ff 00 01 02 03 04
+    #
+    # which are then interpreted as a bit packed mode '1' image
+
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+
+        img = io.BytesIO()
+        blank_line = bytearray((0xFF,) * ((self.state.xsize + 7) // 8))
+        try:
+            self.fd.seek(32)
+            rowmap = struct.unpack_from(
+                f"<{self.state.ysize}H", self.fd.read(self.state.ysize * 2)
+            )
+        except struct.error as e:
+            msg = "Truncated MSP file in row map"
+            raise OSError(msg) from e
+
+        for x, rowlen in enumerate(rowmap):
+            try:
+                if rowlen == 0:
+                    img.write(blank_line)
+                    continue
+                row = self.fd.read(rowlen)
+                if len(row) != rowlen:
+                    msg = f"Truncated MSP file, expected {rowlen} bytes on row {x}"
+                    raise OSError(msg)
+                idx = 0
+                while idx < rowlen:
+                    runtype = row[idx]
+                    idx += 1
+                    if runtype == 0:
+                        (runcount, runval) = struct.unpack_from("Bc", row, idx)
+                        img.write(runval * runcount)
+                        idx += 2
+                    else:
+                        runcount = runtype
+                        img.write(row[idx : idx + runcount])
+                        idx += runcount
+
+            except struct.error as e:
+                msg = f"Corrupted MSP file in row {x}"
+                raise OSError(msg) from e
+
+        self.set_as_raw(img.getvalue(), "1")
+
+        return -1, 0
+
+
+Image.register_decoder("MSP", MspDecoder)
+
+
+#
+# write MSP files (uncompressed only)
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode != "1":
+        msg = f"cannot write mode {im.mode} as MSP"
+        raise OSError(msg)
+
+    # create MSP header
+    header = [0] * 16
+
+    header[0], header[1] = i16(b"Da"), i16(b"nM")  # version 1
+    header[2], header[3] = im.size
+    header[4], header[5] = 1, 1
+    header[6], header[7] = 1, 1
+    header[8], header[9] = im.size
+
+    checksum = 0
+    for h in header:
+        checksum = checksum ^ h
+    header[12] = checksum  # FIXME: is this the right field?
+
+    # header
+    for h in header:
+        fp.write(o16(h))
+
+    # image body
+    ImageFile._save(im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 32, "1")])
+
+
+#
+# registry
+
+Image.register_open(MspImageFile.format, MspImageFile, _accept)
+Image.register_save(MspImageFile.format, _save)
+
+Image.register_extension(MspImageFile.format, ".msp")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PSDraw.py b/backend/venv/lib/python3.12/site-packages/PIL/PSDraw.py
new file mode 100644
index 00000000..02939d26
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PSDraw.py
@@ -0,0 +1,234 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# Simple PostScript graphics interface
+#
+# History:
+# 1996-04-20 fl   Created
+# 1999-01-10 fl   Added gsave/grestore to image method
+# 2005-05-04 fl   Fixed floating point issue in image (from Eric Etheridge)
+#
+# Copyright (c) 1997-2005 by Secret Labs AB.  All rights reserved.
+# Copyright (c) 1996 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import sys
+from typing import IO, TYPE_CHECKING
+
+from . import EpsImagePlugin
+
+##
+# Simple PostScript graphics interface.
+
+
+class PSDraw:
+    """
+    Sets up printing to the given file. If ``fp`` is omitted,
+    ``sys.stdout.buffer`` is assumed.
+    """
+
+    def __init__(self, fp: IO[bytes] | None = None) -> None:
+        if not fp:
+            fp = sys.stdout.buffer
+        self.fp = fp
+
+    def begin_document(self, id: str | None = None) -> None:
+        """Set up printing of a document. (Write PostScript DSC header.)"""
+        # FIXME: incomplete
+        self.fp.write(
+            b"%!PS-Adobe-3.0\n"
+            b"save\n"
+            b"/showpage { } def\n"
+            b"%%EndComments\n"
+            b"%%BeginDocument\n"
+        )
+        # self.fp.write(ERROR_PS)  # debugging!
+        self.fp.write(EDROFF_PS)
+        self.fp.write(VDI_PS)
+        self.fp.write(b"%%EndProlog\n")
+        self.isofont: dict[bytes, int] = {}
+
+    def end_document(self) -> None:
+        """Ends printing. (Write PostScript DSC footer.)"""
+        self.fp.write(b"%%EndDocument\nrestore showpage\n%%End\n")
+        if hasattr(self.fp, "flush"):
+            self.fp.flush()
+
+    def setfont(self, font: str, size: int) -> None:
+        """
+        Selects which font to use.
+
+        :param font: A PostScript font name
+        :param size: Size in points.
+        """
+        font_bytes = bytes(font, "UTF-8")
+        if font_bytes not in self.isofont:
+            # reencode font
+            self.fp.write(
+                b"/PSDraw-%s ISOLatin1Encoding /%s E\n" % (font_bytes, font_bytes)
+            )
+            self.isofont[font_bytes] = 1
+        # rough
+        self.fp.write(b"/F0 %d /PSDraw-%s F\n" % (size, font_bytes))
+
+    def line(self, xy0: tuple[int, int], xy1: tuple[int, int]) -> None:
+        """
+        Draws a line between the two points. Coordinates are given in
+        PostScript point coordinates (72 points per inch, (0, 0) is the lower
+        left corner of the page).
+        """
+        self.fp.write(b"%d %d %d %d Vl\n" % (*xy0, *xy1))
+
+    def rectangle(self, box: tuple[int, int, int, int]) -> None:
+        """
+        Draws a rectangle.
+
+        :param box: A tuple of four integers, specifying left, bottom, width and
+           height.
+        """
+        self.fp.write(b"%d %d M 0 %d %d Vr\n" % box)
+
+    def text(self, xy: tuple[int, int], text: str) -> None:
+        """
+        Draws text at the given position. You must use
+        :py:meth:`~PIL.PSDraw.PSDraw.setfont` before calling this method.
+        """
+        text_bytes = bytes(text, "UTF-8")
+        text_bytes = b"\\(".join(text_bytes.split(b"("))
+        text_bytes = b"\\)".join(text_bytes.split(b")"))
+        self.fp.write(b"%d %d M (%s) S\n" % (xy + (text_bytes,)))
+
+    if TYPE_CHECKING:
+        from . import Image
+
+    def image(
+        self, box: tuple[int, int, int, int], im: Image.Image, dpi: int | None = None
+    ) -> None:
+        """Draw a PIL image, centered in the given box."""
+        # default resolution depends on mode
+        if not dpi:
+            if im.mode == "1":
+                dpi = 200  # fax
+            else:
+                dpi = 100  # grayscale
+        # image size (on paper)
+        x = im.size[0] * 72 / dpi
+        y = im.size[1] * 72 / dpi
+        # max allowed size
+        xmax = float(box[2] - box[0])
+        ymax = float(box[3] - box[1])
+        if x > xmax:
+            y = y * xmax / x
+            x = xmax
+        if y > ymax:
+            x = x * ymax / y
+            y = ymax
+        dx = (xmax - x) / 2 + box[0]
+        dy = (ymax - y) / 2 + box[1]
+        self.fp.write(b"gsave\n%f %f translate\n" % (dx, dy))
+        if (x, y) != im.size:
+            # EpsImagePlugin._save prints the image at (0,0,xsize,ysize)
+            sx = x / im.size[0]
+            sy = y / im.size[1]
+            self.fp.write(b"%f %f scale\n" % (sx, sy))
+        EpsImagePlugin._save(im, self.fp, "", 0)
+        self.fp.write(b"\ngrestore\n")
+
+
+# --------------------------------------------------------------------
+# PostScript driver
+
+#
+# EDROFF.PS -- PostScript driver for Edroff 2
+#
+# History:
+# 94-01-25 fl: created (edroff 2.04)
+#
+# Copyright (c) Fredrik Lundh 1994.
+#
+
+
+EDROFF_PS = b"""\
+/S { show } bind def
+/P { moveto show } bind def
+/M { moveto } bind def
+/X { 0 rmoveto } bind def
+/Y { 0 exch rmoveto } bind def
+/E {    findfont
+        dup maxlength dict begin
+        {
+                1 index /FID ne { def } { pop pop } ifelse
+        } forall
+        /Encoding exch def
+        dup /FontName exch def
+        currentdict end definefont pop
+} bind def
+/F {    findfont exch scalefont dup setfont
+        [ exch /setfont cvx ] cvx bind def
+} bind def
+"""
+
+#
+# VDI.PS -- PostScript driver for VDI meta commands
+#
+# History:
+# 94-01-25 fl: created (edroff 2.04)
+#
+# Copyright (c) Fredrik Lundh 1994.
+#
+
+VDI_PS = b"""\
+/Vm { moveto } bind def
+/Va { newpath arcn stroke } bind def
+/Vl { moveto lineto stroke } bind def
+/Vc { newpath 0 360 arc closepath } bind def
+/Vr {   exch dup 0 rlineto
+        exch dup 0 exch rlineto
+        exch neg 0 rlineto
+        0 exch neg rlineto
+        setgray fill } bind def
+/Tm matrix def
+/Ve {   Tm currentmatrix pop
+        translate scale newpath 0 0 .5 0 360 arc closepath
+        Tm setmatrix
+} bind def
+/Vf { currentgray exch setgray fill setgray } bind def
+"""
+
+#
+# ERROR.PS -- Error handler
+#
+# History:
+# 89-11-21 fl: created (pslist 1.10)
+#
+
+ERROR_PS = b"""\
+/landscape false def
+/errorBUF 200 string def
+/errorNL { currentpoint 10 sub exch pop 72 exch moveto } def
+errordict begin /handleerror {
+    initmatrix /Courier findfont 10 scalefont setfont
+    newpath 72 720 moveto $error begin /newerror false def
+    (PostScript Error) show errorNL errorNL
+    (Error: ) show
+        /errorname load errorBUF cvs show errorNL errorNL
+    (Command: ) show
+        /command load dup type /stringtype ne { errorBUF cvs } if show
+        errorNL errorNL
+    (VMstatus: ) show
+        vmstatus errorBUF cvs show ( bytes available, ) show
+        errorBUF cvs show ( bytes used at level ) show
+        errorBUF cvs show errorNL errorNL
+    (Operand stargck: ) show errorNL /ostargck load {
+        dup type /stringtype ne { errorBUF cvs } if 72 0 rmoveto show errorNL
+    } forall errorNL
+    (Execution stargck: ) show errorNL /estargck load {
+        dup type /stringtype ne { errorBUF cvs } if 72 0 rmoveto show errorNL
+    } forall
+    end showpage
+} def end
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PaletteFile.py b/backend/venv/lib/python3.12/site-packages/PIL/PaletteFile.py
new file mode 100644
index 00000000..81652e5e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PaletteFile.py
@@ -0,0 +1,54 @@
+#
+# Python Imaging Library
+# $Id$
+#
+# stuff to read simple, teragon-style palette files
+#
+# History:
+#       97-08-23 fl     Created
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from typing import IO
+
+from ._binary import o8
+
+
+class PaletteFile:
+    """File handler for Teragon-style palette files."""
+
+    rawmode = "RGB"
+
+    def __init__(self, fp: IO[bytes]) -> None:
+        palette = [o8(i) * 3 for i in range(256)]
+
+        while True:
+            s = fp.readline()
+
+            if not s:
+                break
+            if s[:1] == b"#":
+                continue
+            if len(s) > 100:
+                msg = "bad palette file"
+                raise SyntaxError(msg)
+
+            v = [int(x) for x in s.split()]
+            try:
+                [i, r, g, b] = v
+            except ValueError:
+                [i, r] = v
+                g = b = r
+
+            if 0 <= i <= 255:
+                palette[i] = o8(r) + o8(g) + o8(b)
+
+        self.palette = b"".join(palette)
+
+    def getpalette(self) -> tuple[bytes, str]:
+        return self.palette, self.rawmode
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PalmImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PalmImagePlugin.py
new file mode 100644
index 00000000..b3324537
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PalmImagePlugin.py
@@ -0,0 +1,232 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+
+##
+# Image plugin for Palm pixmap images (output only).
+##
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import o8
+from ._binary import o16be as o16b
+
+# fmt: off
+_Palm8BitColormapValues = (
+    (255, 255, 255), (255, 204, 255), (255, 153, 255), (255, 102, 255),
+    (255,  51, 255), (255,   0, 255), (255, 255, 204), (255, 204, 204),
+    (255, 153, 204), (255, 102, 204), (255,  51, 204), (255,   0, 204),
+    (255, 255, 153), (255, 204, 153), (255, 153, 153), (255, 102, 153),
+    (255,  51, 153), (255,   0, 153), (204, 255, 255), (204, 204, 255),
+    (204, 153, 255), (204, 102, 255), (204,  51, 255), (204,   0, 255),
+    (204, 255, 204), (204, 204, 204), (204, 153, 204), (204, 102, 204),
+    (204,  51, 204), (204,   0, 204), (204, 255, 153), (204, 204, 153),
+    (204, 153, 153), (204, 102, 153), (204,  51, 153), (204,   0, 153),
+    (153, 255, 255), (153, 204, 255), (153, 153, 255), (153, 102, 255),
+    (153,  51, 255), (153,   0, 255), (153, 255, 204), (153, 204, 204),
+    (153, 153, 204), (153, 102, 204), (153,  51, 204), (153,   0, 204),
+    (153, 255, 153), (153, 204, 153), (153, 153, 153), (153, 102, 153),
+    (153,  51, 153), (153,   0, 153), (102, 255, 255), (102, 204, 255),
+    (102, 153, 255), (102, 102, 255), (102,  51, 255), (102,   0, 255),
+    (102, 255, 204), (102, 204, 204), (102, 153, 204), (102, 102, 204),
+    (102,  51, 204), (102,   0, 204), (102, 255, 153), (102, 204, 153),
+    (102, 153, 153), (102, 102, 153), (102,  51, 153), (102,   0, 153),
+    (51,  255, 255), (51,  204, 255), (51,  153, 255), (51,  102, 255),
+    (51,   51, 255), (51,    0, 255), (51,  255, 204), (51,  204, 204),
+    (51,  153, 204), (51,  102, 204), (51,   51, 204), (51,    0, 204),
+    (51,  255, 153), (51,  204, 153), (51,  153, 153), (51,  102, 153),
+    (51,   51, 153), (51,    0, 153), (0,   255, 255), (0,   204, 255),
+    (0,   153, 255), (0,   102, 255), (0,    51, 255), (0,     0, 255),
+    (0,   255, 204), (0,   204, 204), (0,   153, 204), (0,   102, 204),
+    (0,    51, 204), (0,     0, 204), (0,   255, 153), (0,   204, 153),
+    (0,   153, 153), (0,   102, 153), (0,    51, 153), (0,     0, 153),
+    (255, 255, 102), (255, 204, 102), (255, 153, 102), (255, 102, 102),
+    (255,  51, 102), (255,   0, 102), (255, 255,  51), (255, 204,  51),
+    (255, 153,  51), (255, 102,  51), (255,  51,  51), (255,   0,  51),
+    (255, 255,   0), (255, 204,   0), (255, 153,   0), (255, 102,   0),
+    (255,  51,   0), (255,   0,   0), (204, 255, 102), (204, 204, 102),
+    (204, 153, 102), (204, 102, 102), (204,  51, 102), (204,   0, 102),
+    (204, 255,  51), (204, 204,  51), (204, 153,  51), (204, 102,  51),
+    (204,  51,  51), (204,   0,  51), (204, 255,   0), (204, 204,   0),
+    (204, 153,   0), (204, 102,   0), (204,  51,   0), (204,   0,   0),
+    (153, 255, 102), (153, 204, 102), (153, 153, 102), (153, 102, 102),
+    (153,  51, 102), (153,   0, 102), (153, 255,  51), (153, 204,  51),
+    (153, 153,  51), (153, 102,  51), (153,  51,  51), (153,   0,  51),
+    (153, 255,   0), (153, 204,   0), (153, 153,   0), (153, 102,   0),
+    (153,  51,   0), (153,   0,   0), (102, 255, 102), (102, 204, 102),
+    (102, 153, 102), (102, 102, 102), (102,  51, 102), (102,   0, 102),
+    (102, 255,  51), (102, 204,  51), (102, 153,  51), (102, 102,  51),
+    (102,  51,  51), (102,   0,  51), (102, 255,   0), (102, 204,   0),
+    (102, 153,   0), (102, 102,   0), (102,  51,   0), (102,   0,   0),
+    (51,  255, 102), (51,  204, 102), (51,  153, 102), (51,  102, 102),
+    (51,   51, 102), (51,    0, 102), (51,  255,  51), (51,  204,  51),
+    (51,  153,  51), (51,  102,  51), (51,   51,  51), (51,    0,  51),
+    (51,  255,   0), (51,  204,   0), (51,  153,   0), (51,  102,   0),
+    (51,   51,   0), (51,    0,   0), (0,   255, 102), (0,   204, 102),
+    (0,   153, 102), (0,   102, 102), (0,    51, 102), (0,     0, 102),
+    (0,   255,  51), (0,   204,  51), (0,   153,  51), (0,   102,  51),
+    (0,    51,  51), (0,     0,  51), (0,   255,   0), (0,   204,   0),
+    (0,   153,   0), (0,   102,   0), (0,    51,   0), (17,   17,  17),
+    (34,   34,  34), (68,   68,  68), (85,   85,  85), (119, 119, 119),
+    (136, 136, 136), (170, 170, 170), (187, 187, 187), (221, 221, 221),
+    (238, 238, 238), (192, 192, 192), (128,   0,   0), (128,   0, 128),
+    (0,   128,   0), (0,   128, 128), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0),
+    (0,     0,   0), (0,     0,   0), (0,     0,   0), (0,     0,   0))
+# fmt: on
+
+
+# so build a prototype image to be used for palette resampling
+def build_prototype_image() -> Image.Image:
+    image = Image.new("L", (1, len(_Palm8BitColormapValues)))
+    image.putdata(list(range(len(_Palm8BitColormapValues))))
+    palettedata: tuple[int, ...] = ()
+    for colormapValue in _Palm8BitColormapValues:
+        palettedata += colormapValue
+    palettedata += (0, 0, 0) * (256 - len(_Palm8BitColormapValues))
+    image.putpalette(palettedata)
+    return image
+
+
+Palm8BitColormapImage = build_prototype_image()
+
+# OK, we now have in Palm8BitColormapImage,
+# a "P"-mode image with the right palette
+#
+# --------------------------------------------------------------------
+
+_FLAGS = {"custom-colormap": 0x4000, "is-compressed": 0x8000, "has-transparent": 0x2000}
+
+_COMPRESSION_TYPES = {"none": 0xFF, "rle": 0x01, "scanline": 0x00}
+
+
+#
+# --------------------------------------------------------------------
+
+##
+# (Internal) Image save plugin for the Palm format.
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode == "P":
+        # we assume this is a color Palm image with the standard colormap,
+        # unless the "info" dict has a "custom-colormap" field
+
+        rawmode = "P"
+        bpp = 8
+        version = 1
+
+    elif im.mode == "L":
+        if im.encoderinfo.get("bpp") in (1, 2, 4):
+            # this is 8-bit grayscale, so we shift it to get the high-order bits,
+            # and invert it because
+            # Palm does grayscale from white (0) to black (1)
+            bpp = im.encoderinfo["bpp"]
+            maxval = (1 << bpp) - 1
+            shift = 8 - bpp
+            im = im.point(lambda x: maxval - (x >> shift))
+        elif im.info.get("bpp") in (1, 2, 4):
+            # here we assume that even though the inherent mode is 8-bit grayscale,
+            # only the lower bpp bits are significant.
+            # We invert them to match the Palm.
+            bpp = im.info["bpp"]
+            maxval = (1 << bpp) - 1
+            im = im.point(lambda x: maxval - (x & maxval))
+        else:
+            msg = f"cannot write mode {im.mode} as Palm"
+            raise OSError(msg)
+
+        # we ignore the palette here
+        im._mode = "P"
+        rawmode = f"P;{bpp}"
+        version = 1
+
+    elif im.mode == "1":
+        # monochrome -- write it inverted, as is the Palm standard
+        rawmode = "1;I"
+        bpp = 1
+        version = 0
+
+    else:
+        msg = f"cannot write mode {im.mode} as Palm"
+        raise OSError(msg)
+
+    #
+    # make sure image data is available
+    im.load()
+
+    # write header
+
+    cols = im.size[0]
+    rows = im.size[1]
+
+    rowbytes = int((cols + (16 // bpp - 1)) / (16 // bpp)) * 2
+    transparent_index = 0
+    compression_type = _COMPRESSION_TYPES["none"]
+
+    flags = 0
+    if im.mode == "P" and "custom-colormap" in im.info:
+        assert im.palette is not None
+        flags = flags & _FLAGS["custom-colormap"]
+        colormapsize = 4 * 256 + 2
+        colormapmode = im.palette.mode
+        colormap = im.getdata().getpalette()
+    else:
+        colormapsize = 0
+
+    if "offset" in im.info:
+        offset = (rowbytes * rows + 16 + 3 + colormapsize) // 4
+    else:
+        offset = 0
+
+    fp.write(o16b(cols) + o16b(rows) + o16b(rowbytes) + o16b(flags))
+    fp.write(o8(bpp))
+    fp.write(o8(version))
+    fp.write(o16b(offset))
+    fp.write(o8(transparent_index))
+    fp.write(o8(compression_type))
+    fp.write(o16b(0))  # reserved by Palm
+
+    # now write colormap if necessary
+
+    if colormapsize > 0:
+        fp.write(o16b(256))
+        for i in range(256):
+            fp.write(o8(i))
+            if colormapmode == "RGB":
+                fp.write(
+                    o8(colormap[3 * i])
+                    + o8(colormap[3 * i + 1])
+                    + o8(colormap[3 * i + 2])
+                )
+            elif colormapmode == "RGBA":
+                fp.write(
+                    o8(colormap[4 * i])
+                    + o8(colormap[4 * i + 1])
+                    + o8(colormap[4 * i + 2])
+                )
+
+    # now convert data to raw form
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, (rawmode, rowbytes, 1))]
+    )
+
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+#
+# --------------------------------------------------------------------
+
+Image.register_save("Palm", _save)
+
+Image.register_extension("Palm", ".palm")
+
+Image.register_mime("Palm", "image/palm")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PcdImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PcdImagePlugin.py
new file mode 100644
index 00000000..ac40383f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PcdImagePlugin.py
@@ -0,0 +1,64 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PCD file handling
+#
+# History:
+#       96-05-10 fl     Created
+#       96-05-27 fl     Added draft mode (128x192, 256x384)
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image, ImageFile
+
+##
+# Image plugin for PhotoCD images.  This plugin only reads the 768x512
+# image from the file; higher resolutions are encoded in a proprietary
+# encoding.
+
+
+class PcdImageFile(ImageFile.ImageFile):
+    format = "PCD"
+    format_description = "Kodak PhotoCD"
+
+    def _open(self) -> None:
+        # rough
+        assert self.fp is not None
+
+        self.fp.seek(2048)
+        s = self.fp.read(2048)
+
+        if s[:4] != b"PCD_":
+            msg = "not a PCD file"
+            raise SyntaxError(msg)
+
+        orientation = s[1538] & 3
+        self.tile_post_rotate = None
+        if orientation == 1:
+            self.tile_post_rotate = 90
+        elif orientation == 3:
+            self.tile_post_rotate = -90
+
+        self._mode = "RGB"
+        self._size = 768, 512  # FIXME: not correct for rotated images!
+        self.tile = [ImageFile._Tile("pcd", (0, 0) + self.size, 96 * 2048)]
+
+    def load_end(self) -> None:
+        if self.tile_post_rotate:
+            # Handle rotated PCDs
+            self.im = self.im.rotate(self.tile_post_rotate)
+            self._size = self.im.size
+
+
+#
+# registry
+
+Image.register_open(PcdImageFile.format, PcdImageFile)
+
+Image.register_extension(PcdImageFile.format, ".pcd")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PcfFontFile.py b/backend/venv/lib/python3.12/site-packages/PIL/PcfFontFile.py
new file mode 100644
index 00000000..0d1968b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PcfFontFile.py
@@ -0,0 +1,254 @@
+#
+# THIS IS WORK IN PROGRESS
+#
+# The Python Imaging Library
+# $Id$
+#
+# portable compiled font file parser
+#
+# history:
+# 1997-08-19 fl   created
+# 2003-09-13 fl   fixed loading of unicode fonts
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1997-2003 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+from typing import BinaryIO, Callable
+
+from . import FontFile, Image
+from ._binary import i8
+from ._binary import i16be as b16
+from ._binary import i16le as l16
+from ._binary import i32be as b32
+from ._binary import i32le as l32
+
+# --------------------------------------------------------------------
+# declarations
+
+PCF_MAGIC = 0x70636601  # "\x01fcp"
+
+PCF_PROPERTIES = 1 << 0
+PCF_ACCELERATORS = 1 << 1
+PCF_METRICS = 1 << 2
+PCF_BITMAPS = 1 << 3
+PCF_INK_METRICS = 1 << 4
+PCF_BDF_ENCODINGS = 1 << 5
+PCF_SWIDTHS = 1 << 6
+PCF_GLYPH_NAMES = 1 << 7
+PCF_BDF_ACCELERATORS = 1 << 8
+
+BYTES_PER_ROW: list[Callable[[int], int]] = [
+    lambda bits: ((bits + 7) >> 3),
+    lambda bits: ((bits + 15) >> 3) & ~1,
+    lambda bits: ((bits + 31) >> 3) & ~3,
+    lambda bits: ((bits + 63) >> 3) & ~7,
+]
+
+
+def sz(s: bytes, o: int) -> bytes:
+    return s[o : s.index(b"\0", o)]
+
+
+class PcfFontFile(FontFile.FontFile):
+    """Font file plugin for the X11 PCF format."""
+
+    name = "name"
+
+    def __init__(self, fp: BinaryIO, charset_encoding: str = "iso8859-1"):
+        self.charset_encoding = charset_encoding
+
+        magic = l32(fp.read(4))
+        if magic != PCF_MAGIC:
+            msg = "not a PCF file"
+            raise SyntaxError(msg)
+
+        super().__init__()
+
+        count = l32(fp.read(4))
+        self.toc = {}
+        for i in range(count):
+            type = l32(fp.read(4))
+            self.toc[type] = l32(fp.read(4)), l32(fp.read(4)), l32(fp.read(4))
+
+        self.fp = fp
+
+        self.info = self._load_properties()
+
+        metrics = self._load_metrics()
+        bitmaps = self._load_bitmaps(metrics)
+        encoding = self._load_encoding()
+
+        #
+        # create glyph structure
+
+        for ch, ix in enumerate(encoding):
+            if ix is not None:
+                (
+                    xsize,
+                    ysize,
+                    left,
+                    right,
+                    width,
+                    ascent,
+                    descent,
+                    attributes,
+                ) = metrics[ix]
+                self.glyph[ch] = (
+                    (width, 0),
+                    (left, descent - ysize, xsize + left, descent),
+                    (0, 0, xsize, ysize),
+                    bitmaps[ix],
+                )
+
+    def _getformat(
+        self, tag: int
+    ) -> tuple[BinaryIO, int, Callable[[bytes], int], Callable[[bytes], int]]:
+        format, size, offset = self.toc[tag]
+
+        fp = self.fp
+        fp.seek(offset)
+
+        format = l32(fp.read(4))
+
+        if format & 4:
+            i16, i32 = b16, b32
+        else:
+            i16, i32 = l16, l32
+
+        return fp, format, i16, i32
+
+    def _load_properties(self) -> dict[bytes, bytes | int]:
+        #
+        # font properties
+
+        properties = {}
+
+        fp, format, i16, i32 = self._getformat(PCF_PROPERTIES)
+
+        nprops = i32(fp.read(4))
+
+        # read property description
+        p = [(i32(fp.read(4)), i8(fp.read(1)), i32(fp.read(4))) for _ in range(nprops)]
+
+        if nprops & 3:
+            fp.seek(4 - (nprops & 3), io.SEEK_CUR)  # pad
+
+        data = fp.read(i32(fp.read(4)))
+
+        for k, s, v in p:
+            property_value: bytes | int = sz(data, v) if s else v
+            properties[sz(data, k)] = property_value
+
+        return properties
+
+    def _load_metrics(self) -> list[tuple[int, int, int, int, int, int, int, int]]:
+        #
+        # font metrics
+
+        metrics: list[tuple[int, int, int, int, int, int, int, int]] = []
+
+        fp, format, i16, i32 = self._getformat(PCF_METRICS)
+
+        append = metrics.append
+
+        if (format & 0xFF00) == 0x100:
+            # "compressed" metrics
+            for i in range(i16(fp.read(2))):
+                left = i8(fp.read(1)) - 128
+                right = i8(fp.read(1)) - 128
+                width = i8(fp.read(1)) - 128
+                ascent = i8(fp.read(1)) - 128
+                descent = i8(fp.read(1)) - 128
+                xsize = right - left
+                ysize = ascent + descent
+                append((xsize, ysize, left, right, width, ascent, descent, 0))
+
+        else:
+            # "jumbo" metrics
+            for i in range(i32(fp.read(4))):
+                left = i16(fp.read(2))
+                right = i16(fp.read(2))
+                width = i16(fp.read(2))
+                ascent = i16(fp.read(2))
+                descent = i16(fp.read(2))
+                attributes = i16(fp.read(2))
+                xsize = right - left
+                ysize = ascent + descent
+                append((xsize, ysize, left, right, width, ascent, descent, attributes))
+
+        return metrics
+
+    def _load_bitmaps(
+        self, metrics: list[tuple[int, int, int, int, int, int, int, int]]
+    ) -> list[Image.Image]:
+        #
+        # bitmap data
+
+        fp, format, i16, i32 = self._getformat(PCF_BITMAPS)
+
+        nbitmaps = i32(fp.read(4))
+
+        if nbitmaps != len(metrics):
+            msg = "Wrong number of bitmaps"
+            raise OSError(msg)
+
+        offsets = [i32(fp.read(4)) for _ in range(nbitmaps)]
+
+        bitmap_sizes = [i32(fp.read(4)) for _ in range(4)]
+
+        # byteorder = format & 4  # non-zero => MSB
+        bitorder = format & 8  # non-zero => MSB
+        padindex = format & 3
+
+        bitmapsize = bitmap_sizes[padindex]
+        offsets.append(bitmapsize)
+
+        data = fp.read(bitmapsize)
+
+        pad = BYTES_PER_ROW[padindex]
+        mode = "1;R"
+        if bitorder:
+            mode = "1"
+
+        bitmaps = []
+        for i in range(nbitmaps):
+            xsize, ysize = metrics[i][:2]
+            b, e = offsets[i : i + 2]
+            bitmaps.append(
+                Image.frombytes("1", (xsize, ysize), data[b:e], "raw", mode, pad(xsize))
+            )
+
+        return bitmaps
+
+    def _load_encoding(self) -> list[int | None]:
+        fp, format, i16, i32 = self._getformat(PCF_BDF_ENCODINGS)
+
+        first_col, last_col = i16(fp.read(2)), i16(fp.read(2))
+        first_row, last_row = i16(fp.read(2)), i16(fp.read(2))
+
+        i16(fp.read(2))  # default
+
+        nencoding = (last_col - first_col + 1) * (last_row - first_row + 1)
+
+        # map character code to bitmap index
+        encoding: list[int | None] = [None] * min(256, nencoding)
+
+        encoding_offsets = [i16(fp.read(2)) for _ in range(nencoding)]
+
+        for i in range(first_col, len(encoding)):
+            try:
+                encoding_offset = encoding_offsets[
+                    ord(bytearray([i]).decode(self.charset_encoding))
+                ]
+                if encoding_offset != 0xFFFF:
+                    encoding[i] = encoding_offset
+            except UnicodeDecodeError:
+                # character is not supported in selected encoding
+                pass
+
+        return encoding
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PcxImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PcxImagePlugin.py
new file mode 100644
index 00000000..32436cea
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PcxImagePlugin.py
@@ -0,0 +1,229 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PCX file handling
+#
+# This format was originally used by ZSoft's popular PaintBrush
+# program for the IBM PC.  It is also supported by many MS-DOS and
+# Windows applications, including the Windows PaintBrush program in
+# Windows 3.
+#
+# history:
+# 1995-09-01 fl   Created
+# 1996-05-20 fl   Fixed RGB support
+# 1997-01-03 fl   Fixed 2-bit and 4-bit support
+# 1999-02-03 fl   Fixed 8-bit support (broken in 1.0b1)
+# 1999-02-07 fl   Added write support
+# 2002-06-09 fl   Made 2-bit and 4-bit support a bit more robust
+# 2002-07-30 fl   Seek from to current position, not beginning of file
+# 2003-06-03 fl   Extract DPI settings (info["dpi"])
+#
+# Copyright (c) 1997-2003 by Secret Labs AB.
+# Copyright (c) 1995-2003 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import logging
+from typing import IO
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i16le as i16
+from ._binary import o8
+from ._binary import o16le as o16
+
+logger = logging.getLogger(__name__)
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[0] == 10 and prefix[1] in [0, 2, 3, 5]
+
+
+##
+# Image plugin for Paintbrush images.
+
+
+class PcxImageFile(ImageFile.ImageFile):
+    format = "PCX"
+    format_description = "Paintbrush"
+
+    def _open(self) -> None:
+        # header
+        assert self.fp is not None
+
+        s = self.fp.read(128)
+        if not _accept(s):
+            msg = "not a PCX file"
+            raise SyntaxError(msg)
+
+        # image
+        bbox = i16(s, 4), i16(s, 6), i16(s, 8) + 1, i16(s, 10) + 1
+        if bbox[2] <= bbox[0] or bbox[3] <= bbox[1]:
+            msg = "bad PCX image size"
+            raise SyntaxError(msg)
+        logger.debug("BBox: %s %s %s %s", *bbox)
+
+        # format
+        version = s[1]
+        bits = s[3]
+        planes = s[65]
+        provided_stride = i16(s, 66)
+        logger.debug(
+            "PCX version %s, bits %s, planes %s, stride %s",
+            version,
+            bits,
+            planes,
+            provided_stride,
+        )
+
+        self.info["dpi"] = i16(s, 12), i16(s, 14)
+
+        if bits == 1 and planes == 1:
+            mode = rawmode = "1"
+
+        elif bits == 1 and planes in (2, 4):
+            mode = "P"
+            rawmode = f"P;{planes}L"
+            self.palette = ImagePalette.raw("RGB", s[16:64])
+
+        elif version == 5 and bits == 8 and planes == 1:
+            mode = rawmode = "L"
+            # FIXME: hey, this doesn't work with the incremental loader !!!
+            self.fp.seek(-769, io.SEEK_END)
+            s = self.fp.read(769)
+            if len(s) == 769 and s[0] == 12:
+                # check if the palette is linear grayscale
+                for i in range(256):
+                    if s[i * 3 + 1 : i * 3 + 4] != o8(i) * 3:
+                        mode = rawmode = "P"
+                        break
+                if mode == "P":
+                    self.palette = ImagePalette.raw("RGB", s[1:])
+            self.fp.seek(128)
+
+        elif version == 5 and bits == 8 and planes == 3:
+            mode = "RGB"
+            rawmode = "RGB;L"
+
+        else:
+            msg = "unknown PCX mode"
+            raise OSError(msg)
+
+        self._mode = mode
+        self._size = bbox[2] - bbox[0], bbox[3] - bbox[1]
+
+        # Don't trust the passed in stride.
+        # Calculate the approximate position for ourselves.
+        # CVE-2020-35653
+        stride = (self._size[0] * bits + 7) // 8
+
+        # While the specification states that this must be even,
+        # not all images follow this
+        if provided_stride != stride:
+            stride += stride % 2
+
+        bbox = (0, 0) + self.size
+        logger.debug("size: %sx%s", *self.size)
+
+        self.tile = [
+            ImageFile._Tile("pcx", bbox, self.fp.tell(), (rawmode, planes * stride))
+        ]
+
+
+# --------------------------------------------------------------------
+# save PCX files
+
+
+SAVE = {
+    # mode: (version, bits, planes, raw mode)
+    "1": (2, 1, 1, "1"),
+    "L": (5, 8, 1, "L"),
+    "P": (5, 8, 1, "P"),
+    "RGB": (5, 8, 3, "RGB;L"),
+}
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    try:
+        version, bits, planes, rawmode = SAVE[im.mode]
+    except KeyError as e:
+        msg = f"Cannot save {im.mode} images as PCX"
+        raise ValueError(msg) from e
+
+    # bytes per plane
+    stride = (im.size[0] * bits + 7) // 8
+    # stride should be even
+    stride += stride % 2
+    # Stride needs to be kept in sync with the PcxEncode.c version.
+    # Ideally it should be passed in in the state, but the bytes value
+    # gets overwritten.
+
+    logger.debug(
+        "PcxImagePlugin._save: xwidth: %d, bits: %d, stride: %d",
+        im.size[0],
+        bits,
+        stride,
+    )
+
+    # under windows, we could determine the current screen size with
+    # "Image.core.display_mode()[1]", but I think that's overkill...
+
+    screen = im.size
+
+    dpi = 100, 100
+
+    # PCX header
+    fp.write(
+        o8(10)
+        + o8(version)
+        + o8(1)
+        + o8(bits)
+        + o16(0)
+        + o16(0)
+        + o16(im.size[0] - 1)
+        + o16(im.size[1] - 1)
+        + o16(dpi[0])
+        + o16(dpi[1])
+        + b"\0" * 24
+        + b"\xFF" * 24
+        + b"\0"
+        + o8(planes)
+        + o16(stride)
+        + o16(1)
+        + o16(screen[0])
+        + o16(screen[1])
+        + b"\0" * 54
+    )
+
+    assert fp.tell() == 128
+
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("pcx", (0, 0) + im.size, 0, (rawmode, bits * planes))]
+    )
+
+    if im.mode == "P":
+        # colour palette
+        fp.write(o8(12))
+        palette = im.im.getpalette("RGB", "RGB")
+        palette += b"\x00" * (768 - len(palette))
+        fp.write(palette)  # 768 bytes
+    elif im.mode == "L":
+        # grayscale palette
+        fp.write(o8(12))
+        for i in range(256):
+            fp.write(o8(i) * 3)
+
+
+# --------------------------------------------------------------------
+# registry
+
+
+Image.register_open(PcxImageFile.format, PcxImageFile, _accept)
+Image.register_save(PcxImageFile.format, _save)
+
+Image.register_extension(PcxImageFile.format, ".pcx")
+
+Image.register_mime(PcxImageFile.format, "image/x-pcx")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PdfImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PdfImagePlugin.py
new file mode 100644
index 00000000..e9c20ddc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PdfImagePlugin.py
@@ -0,0 +1,311 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PDF (Acrobat) file handling
+#
+# History:
+# 1996-07-16 fl   Created
+# 1997-01-18 fl   Fixed header
+# 2004-02-21 fl   Fixes for 1/L/CMYK images, etc.
+# 2004-02-24 fl   Fixes for 1 and P images.
+#
+# Copyright (c) 1997-2004 by Secret Labs AB.  All rights reserved.
+# Copyright (c) 1996-1997 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+##
+# Image plugin for PDF images (output only).
+##
+from __future__ import annotations
+
+import io
+import math
+import os
+import time
+from typing import IO, Any
+
+from . import Image, ImageFile, ImageSequence, PdfParser, __version__, features
+
+#
+# --------------------------------------------------------------------
+
+# object ids:
+#  1. catalogue
+#  2. pages
+#  3. image
+#  4. page
+#  5. page contents
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    _save(im, fp, filename, save_all=True)
+
+
+##
+# (Internal) Image save plugin for the PDF format.
+
+
+def _write_image(
+    im: Image.Image,
+    filename: str | bytes,
+    existing_pdf: PdfParser.PdfParser,
+    image_refs: list[PdfParser.IndirectReference],
+) -> tuple[PdfParser.IndirectReference, str]:
+    # FIXME: Should replace ASCIIHexDecode with RunLengthDecode
+    # (packbits) or LZWDecode (tiff/lzw compression).  Note that
+    # PDF 1.2 also supports Flatedecode (zip compression).
+
+    params = None
+    decode = None
+
+    #
+    # Get image characteristics
+
+    width, height = im.size
+
+    dict_obj: dict[str, Any] = {"BitsPerComponent": 8}
+    if im.mode == "1":
+        if features.check("libtiff"):
+            decode_filter = "CCITTFaxDecode"
+            dict_obj["BitsPerComponent"] = 1
+            params = PdfParser.PdfArray(
+                [
+                    PdfParser.PdfDict(
+                        {
+                            "K": -1,
+                            "BlackIs1": True,
+                            "Columns": width,
+                            "Rows": height,
+                        }
+                    )
+                ]
+            )
+        else:
+            decode_filter = "DCTDecode"
+        dict_obj["ColorSpace"] = PdfParser.PdfName("DeviceGray")
+        procset = "ImageB"  # grayscale
+    elif im.mode == "L":
+        decode_filter = "DCTDecode"
+        # params = f"<< /Predictor 15 /Columns {width-2} >>"
+        dict_obj["ColorSpace"] = PdfParser.PdfName("DeviceGray")
+        procset = "ImageB"  # grayscale
+    elif im.mode == "LA":
+        decode_filter = "JPXDecode"
+        # params = f"<< /Predictor 15 /Columns {width-2} >>"
+        procset = "ImageB"  # grayscale
+        dict_obj["SMaskInData"] = 1
+    elif im.mode == "P":
+        decode_filter = "ASCIIHexDecode"
+        palette = im.getpalette()
+        assert palette is not None
+        dict_obj["ColorSpace"] = [
+            PdfParser.PdfName("Indexed"),
+            PdfParser.PdfName("DeviceRGB"),
+            len(palette) // 3 - 1,
+            PdfParser.PdfBinary(palette),
+        ]
+        procset = "ImageI"  # indexed color
+
+        if "transparency" in im.info:
+            smask = im.convert("LA").getchannel("A")
+            smask.encoderinfo = {}
+
+            image_ref = _write_image(smask, filename, existing_pdf, image_refs)[0]
+            dict_obj["SMask"] = image_ref
+    elif im.mode == "RGB":
+        decode_filter = "DCTDecode"
+        dict_obj["ColorSpace"] = PdfParser.PdfName("DeviceRGB")
+        procset = "ImageC"  # color images
+    elif im.mode == "RGBA":
+        decode_filter = "JPXDecode"
+        procset = "ImageC"  # color images
+        dict_obj["SMaskInData"] = 1
+    elif im.mode == "CMYK":
+        decode_filter = "DCTDecode"
+        dict_obj["ColorSpace"] = PdfParser.PdfName("DeviceCMYK")
+        procset = "ImageC"  # color images
+        decode = [1, 0, 1, 0, 1, 0, 1, 0]
+    else:
+        msg = f"cannot save mode {im.mode}"
+        raise ValueError(msg)
+
+    #
+    # image
+
+    op = io.BytesIO()
+
+    if decode_filter == "ASCIIHexDecode":
+        ImageFile._save(im, op, [ImageFile._Tile("hex", (0, 0) + im.size, 0, im.mode)])
+    elif decode_filter == "CCITTFaxDecode":
+        im.save(
+            op,
+            "TIFF",
+            compression="group4",
+            # use a single strip
+            strip_size=math.ceil(width / 8) * height,
+        )
+    elif decode_filter == "DCTDecode":
+        Image.SAVE["JPEG"](im, op, filename)
+    elif decode_filter == "JPXDecode":
+        del dict_obj["BitsPerComponent"]
+        Image.SAVE["JPEG2000"](im, op, filename)
+    else:
+        msg = f"unsupported PDF filter ({decode_filter})"
+        raise ValueError(msg)
+
+    stream = op.getvalue()
+    filter: PdfParser.PdfArray | PdfParser.PdfName
+    if decode_filter == "CCITTFaxDecode":
+        stream = stream[8:]
+        filter = PdfParser.PdfArray([PdfParser.PdfName(decode_filter)])
+    else:
+        filter = PdfParser.PdfName(decode_filter)
+
+    image_ref = image_refs.pop(0)
+    existing_pdf.write_obj(
+        image_ref,
+        stream=stream,
+        Type=PdfParser.PdfName("XObject"),
+        Subtype=PdfParser.PdfName("Image"),
+        Width=width,  # * 72.0 / x_resolution,
+        Height=height,  # * 72.0 / y_resolution,
+        Filter=filter,
+        Decode=decode,
+        DecodeParms=params,
+        **dict_obj,
+    )
+
+    return image_ref, procset
+
+
+def _save(
+    im: Image.Image, fp: IO[bytes], filename: str | bytes, save_all: bool = False
+) -> None:
+    is_appending = im.encoderinfo.get("append", False)
+    filename_str = filename.decode() if isinstance(filename, bytes) else filename
+    if is_appending:
+        existing_pdf = PdfParser.PdfParser(f=fp, filename=filename_str, mode="r+b")
+    else:
+        existing_pdf = PdfParser.PdfParser(f=fp, filename=filename_str, mode="w+b")
+
+    dpi = im.encoderinfo.get("dpi")
+    if dpi:
+        x_resolution = dpi[0]
+        y_resolution = dpi[1]
+    else:
+        x_resolution = y_resolution = im.encoderinfo.get("resolution", 72.0)
+
+    info = {
+        "title": (
+            None if is_appending else os.path.splitext(os.path.basename(filename))[0]
+        ),
+        "author": None,
+        "subject": None,
+        "keywords": None,
+        "creator": None,
+        "producer": None,
+        "creationDate": None if is_appending else time.gmtime(),
+        "modDate": None if is_appending else time.gmtime(),
+    }
+    for k, default in info.items():
+        v = im.encoderinfo.get(k) if k in im.encoderinfo else default
+        if v:
+            existing_pdf.info[k[0].upper() + k[1:]] = v
+
+    #
+    # make sure image data is available
+    im.load()
+
+    existing_pdf.start_writing()
+    existing_pdf.write_header()
+    existing_pdf.write_comment(f"created by Pillow {__version__} PDF driver")
+
+    #
+    # pages
+    ims = [im]
+    if save_all:
+        append_images = im.encoderinfo.get("append_images", [])
+        for append_im in append_images:
+            append_im.encoderinfo = im.encoderinfo.copy()
+            ims.append(append_im)
+    number_of_pages = 0
+    image_refs = []
+    page_refs = []
+    contents_refs = []
+    for im in ims:
+        im_number_of_pages = 1
+        if save_all:
+            im_number_of_pages = getattr(im, "n_frames", 1)
+        number_of_pages += im_number_of_pages
+        for i in range(im_number_of_pages):
+            image_refs.append(existing_pdf.next_object_id(0))
+            if im.mode == "P" and "transparency" in im.info:
+                image_refs.append(existing_pdf.next_object_id(0))
+
+            page_refs.append(existing_pdf.next_object_id(0))
+            contents_refs.append(existing_pdf.next_object_id(0))
+            existing_pdf.pages.append(page_refs[-1])
+
+    #
+    # catalog and list of pages
+    existing_pdf.write_catalog()
+
+    page_number = 0
+    for im_sequence in ims:
+        im_pages: ImageSequence.Iterator | list[Image.Image] = (
+            ImageSequence.Iterator(im_sequence) if save_all else [im_sequence]
+        )
+        for im in im_pages:
+            image_ref, procset = _write_image(im, filename, existing_pdf, image_refs)
+
+            #
+            # page
+
+            existing_pdf.write_page(
+                page_refs[page_number],
+                Resources=PdfParser.PdfDict(
+                    ProcSet=[PdfParser.PdfName("PDF"), PdfParser.PdfName(procset)],
+                    XObject=PdfParser.PdfDict(image=image_ref),
+                ),
+                MediaBox=[
+                    0,
+                    0,
+                    im.width * 72.0 / x_resolution,
+                    im.height * 72.0 / y_resolution,
+                ],
+                Contents=contents_refs[page_number],
+            )
+
+            #
+            # page contents
+
+            page_contents = b"q %f 0 0 %f 0 0 cm /image Do Q\n" % (
+                im.width * 72.0 / x_resolution,
+                im.height * 72.0 / y_resolution,
+            )
+
+            existing_pdf.write_obj(contents_refs[page_number], stream=page_contents)
+
+            page_number += 1
+
+    #
+    # trailer
+    existing_pdf.write_xref_and_trailer()
+    if hasattr(fp, "flush"):
+        fp.flush()
+    existing_pdf.close()
+
+
+#
+# --------------------------------------------------------------------
+
+
+Image.register_save("PDF", _save)
+Image.register_save_all("PDF", _save_all)
+
+Image.register_extension("PDF", ".pdf")
+
+Image.register_mime("PDF", "application/pdf")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PdfParser.py b/backend/venv/lib/python3.12/site-packages/PIL/PdfParser.py
new file mode 100644
index 00000000..7cb2d241
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PdfParser.py
@@ -0,0 +1,1073 @@
+from __future__ import annotations
+
+import calendar
+import codecs
+import collections
+import mmap
+import os
+import re
+import time
+import zlib
+from typing import IO, TYPE_CHECKING, Any, NamedTuple, Union
+
+
+# see 7.9.2.2 Text String Type on page 86 and D.3 PDFDocEncoding Character Set
+# on page 656
+def encode_text(s: str) -> bytes:
+    return codecs.BOM_UTF16_BE + s.encode("utf_16_be")
+
+
+PDFDocEncoding = {
+    0x16: "\u0017",
+    0x18: "\u02D8",
+    0x19: "\u02C7",
+    0x1A: "\u02C6",
+    0x1B: "\u02D9",
+    0x1C: "\u02DD",
+    0x1D: "\u02DB",
+    0x1E: "\u02DA",
+    0x1F: "\u02DC",
+    0x80: "\u2022",
+    0x81: "\u2020",
+    0x82: "\u2021",
+    0x83: "\u2026",
+    0x84: "\u2014",
+    0x85: "\u2013",
+    0x86: "\u0192",
+    0x87: "\u2044",
+    0x88: "\u2039",
+    0x89: "\u203A",
+    0x8A: "\u2212",
+    0x8B: "\u2030",
+    0x8C: "\u201E",
+    0x8D: "\u201C",
+    0x8E: "\u201D",
+    0x8F: "\u2018",
+    0x90: "\u2019",
+    0x91: "\u201A",
+    0x92: "\u2122",
+    0x93: "\uFB01",
+    0x94: "\uFB02",
+    0x95: "\u0141",
+    0x96: "\u0152",
+    0x97: "\u0160",
+    0x98: "\u0178",
+    0x99: "\u017D",
+    0x9A: "\u0131",
+    0x9B: "\u0142",
+    0x9C: "\u0153",
+    0x9D: "\u0161",
+    0x9E: "\u017E",
+    0xA0: "\u20AC",
+}
+
+
+def decode_text(b: bytes) -> str:
+    if b[: len(codecs.BOM_UTF16_BE)] == codecs.BOM_UTF16_BE:
+        return b[len(codecs.BOM_UTF16_BE) :].decode("utf_16_be")
+    else:
+        return "".join(PDFDocEncoding.get(byte, chr(byte)) for byte in b)
+
+
+class PdfFormatError(RuntimeError):
+    """An error that probably indicates a syntactic or semantic error in the
+    PDF file structure"""
+
+    pass
+
+
+def check_format_condition(condition: bool, error_message: str) -> None:
+    if not condition:
+        raise PdfFormatError(error_message)
+
+
+class IndirectReferenceTuple(NamedTuple):
+    object_id: int
+    generation: int
+
+
+class IndirectReference(IndirectReferenceTuple):
+    def __str__(self) -> str:
+        return f"{self.object_id} {self.generation} R"
+
+    def __bytes__(self) -> bytes:
+        return self.__str__().encode("us-ascii")
+
+    def __eq__(self, other: object) -> bool:
+        if self.__class__ is not other.__class__:
+            return False
+        assert isinstance(other, IndirectReference)
+        return other.object_id == self.object_id and other.generation == self.generation
+
+    def __ne__(self, other: object) -> bool:
+        return not (self == other)
+
+    def __hash__(self) -> int:
+        return hash((self.object_id, self.generation))
+
+
+class IndirectObjectDef(IndirectReference):
+    def __str__(self) -> str:
+        return f"{self.object_id} {self.generation} obj"
+
+
+class XrefTable:
+    def __init__(self) -> None:
+        self.existing_entries: dict[int, tuple[int, int]] = (
+            {}
+        )  # object ID => (offset, generation)
+        self.new_entries: dict[int, tuple[int, int]] = (
+            {}
+        )  # object ID => (offset, generation)
+        self.deleted_entries = {0: 65536}  # object ID => generation
+        self.reading_finished = False
+
+    def __setitem__(self, key: int, value: tuple[int, int]) -> None:
+        if self.reading_finished:
+            self.new_entries[key] = value
+        else:
+            self.existing_entries[key] = value
+        if key in self.deleted_entries:
+            del self.deleted_entries[key]
+
+    def __getitem__(self, key: int) -> tuple[int, int]:
+        try:
+            return self.new_entries[key]
+        except KeyError:
+            return self.existing_entries[key]
+
+    def __delitem__(self, key: int) -> None:
+        if key in self.new_entries:
+            generation = self.new_entries[key][1] + 1
+            del self.new_entries[key]
+            self.deleted_entries[key] = generation
+        elif key in self.existing_entries:
+            generation = self.existing_entries[key][1] + 1
+            self.deleted_entries[key] = generation
+        elif key in self.deleted_entries:
+            generation = self.deleted_entries[key]
+        else:
+            msg = f"object ID {key} cannot be deleted because it doesn't exist"
+            raise IndexError(msg)
+
+    def __contains__(self, key: int) -> bool:
+        return key in self.existing_entries or key in self.new_entries
+
+    def __len__(self) -> int:
+        return len(
+            set(self.existing_entries.keys())
+            | set(self.new_entries.keys())
+            | set(self.deleted_entries.keys())
+        )
+
+    def keys(self) -> set[int]:
+        return (
+            set(self.existing_entries.keys()) - set(self.deleted_entries.keys())
+        ) | set(self.new_entries.keys())
+
+    def write(self, f: IO[bytes]) -> int:
+        keys = sorted(set(self.new_entries.keys()) | set(self.deleted_entries.keys()))
+        deleted_keys = sorted(set(self.deleted_entries.keys()))
+        startxref = f.tell()
+        f.write(b"xref\n")
+        while keys:
+            # find a contiguous sequence of object IDs
+            prev: int | None = None
+            for index, key in enumerate(keys):
+                if prev is None or prev + 1 == key:
+                    prev = key
+                else:
+                    contiguous_keys = keys[:index]
+                    keys = keys[index:]
+                    break
+            else:
+                contiguous_keys = keys
+                keys = []
+            f.write(b"%d %d\n" % (contiguous_keys[0], len(contiguous_keys)))
+            for object_id in contiguous_keys:
+                if object_id in self.new_entries:
+                    f.write(b"%010d %05d n \n" % self.new_entries[object_id])
+                else:
+                    this_deleted_object_id = deleted_keys.pop(0)
+                    check_format_condition(
+                        object_id == this_deleted_object_id,
+                        f"expected the next deleted object ID to be {object_id}, "
+                        f"instead found {this_deleted_object_id}",
+                    )
+                    try:
+                        next_in_linked_list = deleted_keys[0]
+                    except IndexError:
+                        next_in_linked_list = 0
+                    f.write(
+                        b"%010d %05d f \n"
+                        % (next_in_linked_list, self.deleted_entries[object_id])
+                    )
+        return startxref
+
+
+class PdfName:
+    name: bytes
+
+    def __init__(self, name: PdfName | bytes | str) -> None:
+        if isinstance(name, PdfName):
+            self.name = name.name
+        elif isinstance(name, bytes):
+            self.name = name
+        else:
+            self.name = name.encode("us-ascii")
+
+    def name_as_str(self) -> str:
+        return self.name.decode("us-ascii")
+
+    def __eq__(self, other: object) -> bool:
+        return (
+            isinstance(other, PdfName) and other.name == self.name
+        ) or other == self.name
+
+    def __hash__(self) -> int:
+        return hash(self.name)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({repr(self.name)})"
+
+    @classmethod
+    def from_pdf_stream(cls, data: bytes) -> PdfName:
+        return cls(PdfParser.interpret_name(data))
+
+    allowed_chars = set(range(33, 127)) - {ord(c) for c in "#%/()<>[]{}"}
+
+    def __bytes__(self) -> bytes:
+        result = bytearray(b"/")
+        for b in self.name:
+            if b in self.allowed_chars:
+                result.append(b)
+            else:
+                result.extend(b"#%02X" % b)
+        return bytes(result)
+
+
+class PdfArray(list[Any]):
+    def __bytes__(self) -> bytes:
+        return b"[ " + b" ".join(pdf_repr(x) for x in self) + b" ]"
+
+
+if TYPE_CHECKING:
+    _DictBase = collections.UserDict[Union[str, bytes], Any]
+else:
+    _DictBase = collections.UserDict
+
+
+class PdfDict(_DictBase):
+    def __setattr__(self, key: str, value: Any) -> None:
+        if key == "data":
+            collections.UserDict.__setattr__(self, key, value)
+        else:
+            self[key.encode("us-ascii")] = value
+
+    def __getattr__(self, key: str) -> str | time.struct_time:
+        try:
+            value = self[key.encode("us-ascii")]
+        except KeyError as e:
+            raise AttributeError(key) from e
+        if isinstance(value, bytes):
+            value = decode_text(value)
+        if key.endswith("Date"):
+            if value.startswith("D:"):
+                value = value[2:]
+
+            relationship = "Z"
+            if len(value) > 17:
+                relationship = value[14]
+                offset = int(value[15:17]) * 60
+                if len(value) > 20:
+                    offset += int(value[18:20])
+
+            format = "%Y%m%d%H%M%S"[: len(value) - 2]
+            value = time.strptime(value[: len(format) + 2], format)
+            if relationship in ["+", "-"]:
+                offset *= 60
+                if relationship == "+":
+                    offset *= -1
+                value = time.gmtime(calendar.timegm(value) + offset)
+        return value
+
+    def __bytes__(self) -> bytes:
+        out = bytearray(b"<<")
+        for key, value in self.items():
+            if value is None:
+                continue
+            value = pdf_repr(value)
+            out.extend(b"\n")
+            out.extend(bytes(PdfName(key)))
+            out.extend(b" ")
+            out.extend(value)
+        out.extend(b"\n>>")
+        return bytes(out)
+
+
+class PdfBinary:
+    def __init__(self, data: list[int] | bytes) -> None:
+        self.data = data
+
+    def __bytes__(self) -> bytes:
+        return b"<%s>" % b"".join(b"%02X" % b for b in self.data)
+
+
+class PdfStream:
+    def __init__(self, dictionary: PdfDict, buf: bytes) -> None:
+        self.dictionary = dictionary
+        self.buf = buf
+
+    def decode(self) -> bytes:
+        try:
+            filter = self.dictionary[b"Filter"]
+        except KeyError:
+            return self.buf
+        if filter == b"FlateDecode":
+            try:
+                expected_length = self.dictionary[b"DL"]
+            except KeyError:
+                expected_length = self.dictionary[b"Length"]
+            return zlib.decompress(self.buf, bufsize=int(expected_length))
+        else:
+            msg = f"stream filter {repr(filter)} unknown/unsupported"
+            raise NotImplementedError(msg)
+
+
+def pdf_repr(x: Any) -> bytes:
+    if x is True:
+        return b"true"
+    elif x is False:
+        return b"false"
+    elif x is None:
+        return b"null"
+    elif isinstance(x, (PdfName, PdfDict, PdfArray, PdfBinary)):
+        return bytes(x)
+    elif isinstance(x, (int, float)):
+        return str(x).encode("us-ascii")
+    elif isinstance(x, time.struct_time):
+        return b"(D:" + time.strftime("%Y%m%d%H%M%SZ", x).encode("us-ascii") + b")"
+    elif isinstance(x, dict):
+        return bytes(PdfDict(x))
+    elif isinstance(x, list):
+        return bytes(PdfArray(x))
+    elif isinstance(x, str):
+        return pdf_repr(encode_text(x))
+    elif isinstance(x, bytes):
+        # XXX escape more chars? handle binary garbage
+        x = x.replace(b"\\", b"\\\\")
+        x = x.replace(b"(", b"\\(")
+        x = x.replace(b")", b"\\)")
+        return b"(" + x + b")"
+    else:
+        return bytes(x)
+
+
+class PdfParser:
+    """Based on
+    https://www.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/PDF32000_2008.pdf
+    Supports PDF up to 1.4
+    """
+
+    def __init__(
+        self,
+        filename: str | None = None,
+        f: IO[bytes] | None = None,
+        buf: bytes | bytearray | None = None,
+        start_offset: int = 0,
+        mode: str = "rb",
+    ) -> None:
+        if buf and f:
+            msg = "specify buf or f or filename, but not both buf and f"
+            raise RuntimeError(msg)
+        self.filename = filename
+        self.buf: bytes | bytearray | mmap.mmap | None = buf
+        self.f = f
+        self.start_offset = start_offset
+        self.should_close_buf = False
+        self.should_close_file = False
+        if filename is not None and f is None:
+            self.f = f = open(filename, mode)
+            self.should_close_file = True
+        if f is not None:
+            self.buf = self.get_buf_from_file(f)
+            self.should_close_buf = True
+            if not filename and hasattr(f, "name"):
+                self.filename = f.name
+        self.cached_objects: dict[IndirectReference, Any] = {}
+        self.root_ref: IndirectReference | None
+        self.info_ref: IndirectReference | None
+        self.pages_ref: IndirectReference | None
+        self.last_xref_section_offset: int | None
+        if self.buf:
+            self.read_pdf_info()
+        else:
+            self.file_size_total = self.file_size_this = 0
+            self.root = PdfDict()
+            self.root_ref = None
+            self.info = PdfDict()
+            self.info_ref = None
+            self.page_tree_root = PdfDict()
+            self.pages: list[IndirectReference] = []
+            self.orig_pages: list[IndirectReference] = []
+            self.pages_ref = None
+            self.last_xref_section_offset = None
+            self.trailer_dict: dict[bytes, Any] = {}
+            self.xref_table = XrefTable()
+        self.xref_table.reading_finished = True
+        if f:
+            self.seek_end()
+
+    def __enter__(self) -> PdfParser:
+        return self
+
+    def __exit__(self, *args: object) -> None:
+        self.close()
+
+    def start_writing(self) -> None:
+        self.close_buf()
+        self.seek_end()
+
+    def close_buf(self) -> None:
+        if isinstance(self.buf, mmap.mmap):
+            self.buf.close()
+        self.buf = None
+
+    def close(self) -> None:
+        if self.should_close_buf:
+            self.close_buf()
+        if self.f is not None and self.should_close_file:
+            self.f.close()
+            self.f = None
+
+    def seek_end(self) -> None:
+        assert self.f is not None
+        self.f.seek(0, os.SEEK_END)
+
+    def write_header(self) -> None:
+        assert self.f is not None
+        self.f.write(b"%PDF-1.4\n")
+
+    def write_comment(self, s: str) -> None:
+        assert self.f is not None
+        self.f.write(f"% {s}\n".encode())
+
+    def write_catalog(self) -> IndirectReference:
+        assert self.f is not None
+        self.del_root()
+        self.root_ref = self.next_object_id(self.f.tell())
+        self.pages_ref = self.next_object_id(0)
+        self.rewrite_pages()
+        self.write_obj(self.root_ref, Type=PdfName(b"Catalog"), Pages=self.pages_ref)
+        self.write_obj(
+            self.pages_ref,
+            Type=PdfName(b"Pages"),
+            Count=len(self.pages),
+            Kids=self.pages,
+        )
+        return self.root_ref
+
+    def rewrite_pages(self) -> None:
+        pages_tree_nodes_to_delete = []
+        for i, page_ref in enumerate(self.orig_pages):
+            page_info = self.cached_objects[page_ref]
+            del self.xref_table[page_ref.object_id]
+            pages_tree_nodes_to_delete.append(page_info[PdfName(b"Parent")])
+            if page_ref not in self.pages:
+                # the page has been deleted
+                continue
+            # make dict keys into strings for passing to write_page
+            stringified_page_info = {}
+            for key, value in page_info.items():
+                # key should be a PdfName
+                stringified_page_info[key.name_as_str()] = value
+            stringified_page_info["Parent"] = self.pages_ref
+            new_page_ref = self.write_page(None, **stringified_page_info)
+            for j, cur_page_ref in enumerate(self.pages):
+                if cur_page_ref == page_ref:
+                    # replace the page reference with the new one
+                    self.pages[j] = new_page_ref
+        # delete redundant Pages tree nodes from xref table
+        for pages_tree_node_ref in pages_tree_nodes_to_delete:
+            while pages_tree_node_ref:
+                pages_tree_node = self.cached_objects[pages_tree_node_ref]
+                if pages_tree_node_ref.object_id in self.xref_table:
+                    del self.xref_table[pages_tree_node_ref.object_id]
+                pages_tree_node_ref = pages_tree_node.get(b"Parent", None)
+        self.orig_pages = []
+
+    def write_xref_and_trailer(
+        self, new_root_ref: IndirectReference | None = None
+    ) -> None:
+        assert self.f is not None
+        if new_root_ref:
+            self.del_root()
+            self.root_ref = new_root_ref
+        if self.info:
+            self.info_ref = self.write_obj(None, self.info)
+        start_xref = self.xref_table.write(self.f)
+        num_entries = len(self.xref_table)
+        trailer_dict: dict[str | bytes, Any] = {
+            b"Root": self.root_ref,
+            b"Size": num_entries,
+        }
+        if self.last_xref_section_offset is not None:
+            trailer_dict[b"Prev"] = self.last_xref_section_offset
+        if self.info:
+            trailer_dict[b"Info"] = self.info_ref
+        self.last_xref_section_offset = start_xref
+        self.f.write(
+            b"trailer\n"
+            + bytes(PdfDict(trailer_dict))
+            + b"\nstartxref\n%d\n%%%%EOF" % start_xref
+        )
+
+    def write_page(
+        self, ref: int | IndirectReference | None, *objs: Any, **dict_obj: Any
+    ) -> IndirectReference:
+        obj_ref = self.pages[ref] if isinstance(ref, int) else ref
+        if "Type" not in dict_obj:
+            dict_obj["Type"] = PdfName(b"Page")
+        if "Parent" not in dict_obj:
+            dict_obj["Parent"] = self.pages_ref
+        return self.write_obj(obj_ref, *objs, **dict_obj)
+
+    def write_obj(
+        self, ref: IndirectReference | None, *objs: Any, **dict_obj: Any
+    ) -> IndirectReference:
+        assert self.f is not None
+        f = self.f
+        if ref is None:
+            ref = self.next_object_id(f.tell())
+        else:
+            self.xref_table[ref.object_id] = (f.tell(), ref.generation)
+        f.write(bytes(IndirectObjectDef(*ref)))
+        stream = dict_obj.pop("stream", None)
+        if stream is not None:
+            dict_obj["Length"] = len(stream)
+        if dict_obj:
+            f.write(pdf_repr(dict_obj))
+        for obj in objs:
+            f.write(pdf_repr(obj))
+        if stream is not None:
+            f.write(b"stream\n")
+            f.write(stream)
+            f.write(b"\nendstream\n")
+        f.write(b"endobj\n")
+        return ref
+
+    def del_root(self) -> None:
+        if self.root_ref is None:
+            return
+        del self.xref_table[self.root_ref.object_id]
+        del self.xref_table[self.root[b"Pages"].object_id]
+
+    @staticmethod
+    def get_buf_from_file(f: IO[bytes]) -> bytes | mmap.mmap:
+        if hasattr(f, "getbuffer"):
+            return f.getbuffer()
+        elif hasattr(f, "getvalue"):
+            return f.getvalue()
+        else:
+            try:
+                return mmap.mmap(f.fileno(), 0, access=mmap.ACCESS_READ)
+            except ValueError:  # cannot mmap an empty file
+                return b""
+
+    def read_pdf_info(self) -> None:
+        assert self.buf is not None
+        self.file_size_total = len(self.buf)
+        self.file_size_this = self.file_size_total - self.start_offset
+        self.read_trailer()
+        check_format_condition(
+            self.trailer_dict.get(b"Root") is not None, "Root is missing"
+        )
+        self.root_ref = self.trailer_dict[b"Root"]
+        assert self.root_ref is not None
+        self.info_ref = self.trailer_dict.get(b"Info", None)
+        self.root = PdfDict(self.read_indirect(self.root_ref))
+        if self.info_ref is None:
+            self.info = PdfDict()
+        else:
+            self.info = PdfDict(self.read_indirect(self.info_ref))
+        check_format_condition(b"Type" in self.root, "/Type missing in Root")
+        check_format_condition(
+            self.root[b"Type"] == b"Catalog", "/Type in Root is not /Catalog"
+        )
+        check_format_condition(
+            self.root.get(b"Pages") is not None, "/Pages missing in Root"
+        )
+        check_format_condition(
+            isinstance(self.root[b"Pages"], IndirectReference),
+            "/Pages in Root is not an indirect reference",
+        )
+        self.pages_ref = self.root[b"Pages"]
+        assert self.pages_ref is not None
+        self.page_tree_root = self.read_indirect(self.pages_ref)
+        self.pages = self.linearize_page_tree(self.page_tree_root)
+        # save the original list of page references
+        # in case the user modifies, adds or deletes some pages
+        # and we need to rewrite the pages and their list
+        self.orig_pages = self.pages[:]
+
+    def next_object_id(self, offset: int | None = None) -> IndirectReference:
+        try:
+            # TODO: support reuse of deleted objects
+            reference = IndirectReference(max(self.xref_table.keys()) + 1, 0)
+        except ValueError:
+            reference = IndirectReference(1, 0)
+        if offset is not None:
+            self.xref_table[reference.object_id] = (offset, 0)
+        return reference
+
+    delimiter = rb"[][()<>{}/%]"
+    delimiter_or_ws = rb"[][()<>{}/%\000\011\012\014\015\040]"
+    whitespace = rb"[\000\011\012\014\015\040]"
+    whitespace_or_hex = rb"[\000\011\012\014\015\0400-9a-fA-F]"
+    whitespace_optional = whitespace + b"*"
+    whitespace_mandatory = whitespace + b"+"
+    # No "\012" aka "\n" or "\015" aka "\r":
+    whitespace_optional_no_nl = rb"[\000\011\014\040]*"
+    newline_only = rb"[\r\n]+"
+    newline = whitespace_optional_no_nl + newline_only + whitespace_optional_no_nl
+    re_trailer_end = re.compile(
+        whitespace_mandatory
+        + rb"trailer"
+        + whitespace_optional
+        + rb"<<(.*>>)"
+        + newline
+        + rb"startxref"
+        + newline
+        + rb"([0-9]+)"
+        + newline
+        + rb"%%EOF"
+        + whitespace_optional
+        + rb"$",
+        re.DOTALL,
+    )
+    re_trailer_prev = re.compile(
+        whitespace_optional
+        + rb"trailer"
+        + whitespace_optional
+        + rb"<<(.*?>>)"
+        + newline
+        + rb"startxref"
+        + newline
+        + rb"([0-9]+)"
+        + newline
+        + rb"%%EOF"
+        + whitespace_optional,
+        re.DOTALL,
+    )
+
+    def read_trailer(self) -> None:
+        assert self.buf is not None
+        search_start_offset = len(self.buf) - 16384
+        if search_start_offset < self.start_offset:
+            search_start_offset = self.start_offset
+        m = self.re_trailer_end.search(self.buf, search_start_offset)
+        check_format_condition(m is not None, "trailer end not found")
+        # make sure we found the LAST trailer
+        last_match = m
+        while m:
+            last_match = m
+            m = self.re_trailer_end.search(self.buf, m.start() + 16)
+        if not m:
+            m = last_match
+        assert m is not None
+        trailer_data = m.group(1)
+        self.last_xref_section_offset = int(m.group(2))
+        self.trailer_dict = self.interpret_trailer(trailer_data)
+        self.xref_table = XrefTable()
+        self.read_xref_table(xref_section_offset=self.last_xref_section_offset)
+        if b"Prev" in self.trailer_dict:
+            self.read_prev_trailer(self.trailer_dict[b"Prev"])
+
+    def read_prev_trailer(self, xref_section_offset: int) -> None:
+        assert self.buf is not None
+        trailer_offset = self.read_xref_table(xref_section_offset=xref_section_offset)
+        m = self.re_trailer_prev.search(
+            self.buf[trailer_offset : trailer_offset + 16384]
+        )
+        check_format_condition(m is not None, "previous trailer not found")
+        assert m is not None
+        trailer_data = m.group(1)
+        check_format_condition(
+            int(m.group(2)) == xref_section_offset,
+            "xref section offset in previous trailer doesn't match what was expected",
+        )
+        trailer_dict = self.interpret_trailer(trailer_data)
+        if b"Prev" in trailer_dict:
+            self.read_prev_trailer(trailer_dict[b"Prev"])
+
+    re_whitespace_optional = re.compile(whitespace_optional)
+    re_name = re.compile(
+        whitespace_optional
+        + rb"/([!-$&'*-.0-;=?-Z\\^-z|~]+)(?="
+        + delimiter_or_ws
+        + rb")"
+    )
+    re_dict_start = re.compile(whitespace_optional + rb"<<")
+    re_dict_end = re.compile(whitespace_optional + rb">>" + whitespace_optional)
+
+    @classmethod
+    def interpret_trailer(cls, trailer_data: bytes) -> dict[bytes, Any]:
+        trailer = {}
+        offset = 0
+        while True:
+            m = cls.re_name.match(trailer_data, offset)
+            if not m:
+                m = cls.re_dict_end.match(trailer_data, offset)
+                check_format_condition(
+                    m is not None and m.end() == len(trailer_data),
+                    "name not found in trailer, remaining data: "
+                    + repr(trailer_data[offset:]),
+                )
+                break
+            key = cls.interpret_name(m.group(1))
+            assert isinstance(key, bytes)
+            value, value_offset = cls.get_value(trailer_data, m.end())
+            trailer[key] = value
+            if value_offset is None:
+                break
+            offset = value_offset
+        check_format_condition(
+            b"Size" in trailer and isinstance(trailer[b"Size"], int),
+            "/Size not in trailer or not an integer",
+        )
+        check_format_condition(
+            b"Root" in trailer and isinstance(trailer[b"Root"], IndirectReference),
+            "/Root not in trailer or not an indirect reference",
+        )
+        return trailer
+
+    re_hashes_in_name = re.compile(rb"([^#]*)(#([0-9a-fA-F]{2}))?")
+
+    @classmethod
+    def interpret_name(cls, raw: bytes, as_text: bool = False) -> str | bytes:
+        name = b""
+        for m in cls.re_hashes_in_name.finditer(raw):
+            if m.group(3):
+                name += m.group(1) + bytearray.fromhex(m.group(3).decode("us-ascii"))
+            else:
+                name += m.group(1)
+        if as_text:
+            return name.decode("utf-8")
+        else:
+            return bytes(name)
+
+    re_null = re.compile(whitespace_optional + rb"null(?=" + delimiter_or_ws + rb")")
+    re_true = re.compile(whitespace_optional + rb"true(?=" + delimiter_or_ws + rb")")
+    re_false = re.compile(whitespace_optional + rb"false(?=" + delimiter_or_ws + rb")")
+    re_int = re.compile(
+        whitespace_optional + rb"([-+]?[0-9]+)(?=" + delimiter_or_ws + rb")"
+    )
+    re_real = re.compile(
+        whitespace_optional
+        + rb"([-+]?([0-9]+\.[0-9]*|[0-9]*\.[0-9]+))(?="
+        + delimiter_or_ws
+        + rb")"
+    )
+    re_array_start = re.compile(whitespace_optional + rb"\[")
+    re_array_end = re.compile(whitespace_optional + rb"]")
+    re_string_hex = re.compile(
+        whitespace_optional + rb"<(" + whitespace_or_hex + rb"*)>"
+    )
+    re_string_lit = re.compile(whitespace_optional + rb"\(")
+    re_indirect_reference = re.compile(
+        whitespace_optional
+        + rb"([-+]?[0-9]+)"
+        + whitespace_mandatory
+        + rb"([-+]?[0-9]+)"
+        + whitespace_mandatory
+        + rb"R(?="
+        + delimiter_or_ws
+        + rb")"
+    )
+    re_indirect_def_start = re.compile(
+        whitespace_optional
+        + rb"([-+]?[0-9]+)"
+        + whitespace_mandatory
+        + rb"([-+]?[0-9]+)"
+        + whitespace_mandatory
+        + rb"obj(?="
+        + delimiter_or_ws
+        + rb")"
+    )
+    re_indirect_def_end = re.compile(
+        whitespace_optional + rb"endobj(?=" + delimiter_or_ws + rb")"
+    )
+    re_comment = re.compile(
+        rb"(" + whitespace_optional + rb"%[^\r\n]*" + newline + rb")*"
+    )
+    re_stream_start = re.compile(whitespace_optional + rb"stream\r?\n")
+    re_stream_end = re.compile(
+        whitespace_optional + rb"endstream(?=" + delimiter_or_ws + rb")"
+    )
+
+    @classmethod
+    def get_value(
+        cls,
+        data: bytes | bytearray | mmap.mmap,
+        offset: int,
+        expect_indirect: IndirectReference | None = None,
+        max_nesting: int = -1,
+    ) -> tuple[Any, int | None]:
+        if max_nesting == 0:
+            return None, None
+        m = cls.re_comment.match(data, offset)
+        if m:
+            offset = m.end()
+        m = cls.re_indirect_def_start.match(data, offset)
+        if m:
+            check_format_condition(
+                int(m.group(1)) > 0,
+                "indirect object definition: object ID must be greater than 0",
+            )
+            check_format_condition(
+                int(m.group(2)) >= 0,
+                "indirect object definition: generation must be non-negative",
+            )
+            check_format_condition(
+                expect_indirect is None
+                or expect_indirect
+                == IndirectReference(int(m.group(1)), int(m.group(2))),
+                "indirect object definition different than expected",
+            )
+            object, object_offset = cls.get_value(
+                data, m.end(), max_nesting=max_nesting - 1
+            )
+            if object_offset is None:
+                return object, None
+            m = cls.re_indirect_def_end.match(data, object_offset)
+            check_format_condition(
+                m is not None, "indirect object definition end not found"
+            )
+            assert m is not None
+            return object, m.end()
+        check_format_condition(
+            not expect_indirect, "indirect object definition not found"
+        )
+        m = cls.re_indirect_reference.match(data, offset)
+        if m:
+            check_format_condition(
+                int(m.group(1)) > 0,
+                "indirect object reference: object ID must be greater than 0",
+            )
+            check_format_condition(
+                int(m.group(2)) >= 0,
+                "indirect object reference: generation must be non-negative",
+            )
+            return IndirectReference(int(m.group(1)), int(m.group(2))), m.end()
+        m = cls.re_dict_start.match(data, offset)
+        if m:
+            offset = m.end()
+            result: dict[Any, Any] = {}
+            m = cls.re_dict_end.match(data, offset)
+            current_offset: int | None = offset
+            while not m:
+                assert current_offset is not None
+                key, current_offset = cls.get_value(
+                    data, current_offset, max_nesting=max_nesting - 1
+                )
+                if current_offset is None:
+                    return result, None
+                value, current_offset = cls.get_value(
+                    data, current_offset, max_nesting=max_nesting - 1
+                )
+                result[key] = value
+                if current_offset is None:
+                    return result, None
+                m = cls.re_dict_end.match(data, current_offset)
+            current_offset = m.end()
+            m = cls.re_stream_start.match(data, current_offset)
+            if m:
+                stream_len = result.get(b"Length")
+                if stream_len is None or not isinstance(stream_len, int):
+                    msg = f"bad or missing Length in stream dict ({stream_len})"
+                    raise PdfFormatError(msg)
+                stream_data = data[m.end() : m.end() + stream_len]
+                m = cls.re_stream_end.match(data, m.end() + stream_len)
+                check_format_condition(m is not None, "stream end not found")
+                assert m is not None
+                current_offset = m.end()
+                return PdfStream(PdfDict(result), stream_data), current_offset
+            return PdfDict(result), current_offset
+        m = cls.re_array_start.match(data, offset)
+        if m:
+            offset = m.end()
+            results = []
+            m = cls.re_array_end.match(data, offset)
+            current_offset = offset
+            while not m:
+                assert current_offset is not None
+                value, current_offset = cls.get_value(
+                    data, current_offset, max_nesting=max_nesting - 1
+                )
+                results.append(value)
+                if current_offset is None:
+                    return results, None
+                m = cls.re_array_end.match(data, current_offset)
+            return results, m.end()
+        m = cls.re_null.match(data, offset)
+        if m:
+            return None, m.end()
+        m = cls.re_true.match(data, offset)
+        if m:
+            return True, m.end()
+        m = cls.re_false.match(data, offset)
+        if m:
+            return False, m.end()
+        m = cls.re_name.match(data, offset)
+        if m:
+            return PdfName(cls.interpret_name(m.group(1))), m.end()
+        m = cls.re_int.match(data, offset)
+        if m:
+            return int(m.group(1)), m.end()
+        m = cls.re_real.match(data, offset)
+        if m:
+            # XXX Decimal instead of float???
+            return float(m.group(1)), m.end()
+        m = cls.re_string_hex.match(data, offset)
+        if m:
+            # filter out whitespace
+            hex_string = bytearray(
+                b for b in m.group(1) if b in b"0123456789abcdefABCDEF"
+            )
+            if len(hex_string) % 2 == 1:
+                # append a 0 if the length is not even - yes, at the end
+                hex_string.append(ord(b"0"))
+            return bytearray.fromhex(hex_string.decode("us-ascii")), m.end()
+        m = cls.re_string_lit.match(data, offset)
+        if m:
+            return cls.get_literal_string(data, m.end())
+        # return None, offset  # fallback (only for debugging)
+        msg = f"unrecognized object: {repr(data[offset : offset + 32])}"
+        raise PdfFormatError(msg)
+
+    re_lit_str_token = re.compile(
+        rb"(\\[nrtbf()\\])|(\\[0-9]{1,3})|(\\(\r\n|\r|\n))|(\r\n|\r|\n)|(\()|(\))"
+    )
+    escaped_chars = {
+        b"n": b"\n",
+        b"r": b"\r",
+        b"t": b"\t",
+        b"b": b"\b",
+        b"f": b"\f",
+        b"(": b"(",
+        b")": b")",
+        b"\\": b"\\",
+        ord(b"n"): b"\n",
+        ord(b"r"): b"\r",
+        ord(b"t"): b"\t",
+        ord(b"b"): b"\b",
+        ord(b"f"): b"\f",
+        ord(b"("): b"(",
+        ord(b")"): b")",
+        ord(b"\\"): b"\\",
+    }
+
+    @classmethod
+    def get_literal_string(
+        cls, data: bytes | bytearray | mmap.mmap, offset: int
+    ) -> tuple[bytes, int]:
+        nesting_depth = 0
+        result = bytearray()
+        for m in cls.re_lit_str_token.finditer(data, offset):
+            result.extend(data[offset : m.start()])
+            if m.group(1):
+                result.extend(cls.escaped_chars[m.group(1)[1]])
+            elif m.group(2):
+                result.append(int(m.group(2)[1:], 8))
+            elif m.group(3):
+                pass
+            elif m.group(5):
+                result.extend(b"\n")
+            elif m.group(6):
+                result.extend(b"(")
+                nesting_depth += 1
+            elif m.group(7):
+                if nesting_depth == 0:
+                    return bytes(result), m.end()
+                result.extend(b")")
+                nesting_depth -= 1
+            offset = m.end()
+        msg = "unfinished literal string"
+        raise PdfFormatError(msg)
+
+    re_xref_section_start = re.compile(whitespace_optional + rb"xref" + newline)
+    re_xref_subsection_start = re.compile(
+        whitespace_optional
+        + rb"([0-9]+)"
+        + whitespace_mandatory
+        + rb"([0-9]+)"
+        + whitespace_optional
+        + newline_only
+    )
+    re_xref_entry = re.compile(rb"([0-9]{10}) ([0-9]{5}) ([fn])( \r| \n|\r\n)")
+
+    def read_xref_table(self, xref_section_offset: int) -> int:
+        assert self.buf is not None
+        subsection_found = False
+        m = self.re_xref_section_start.match(
+            self.buf, xref_section_offset + self.start_offset
+        )
+        check_format_condition(m is not None, "xref section start not found")
+        assert m is not None
+        offset = m.end()
+        while True:
+            m = self.re_xref_subsection_start.match(self.buf, offset)
+            if not m:
+                check_format_condition(
+                    subsection_found, "xref subsection start not found"
+                )
+                break
+            subsection_found = True
+            offset = m.end()
+            first_object = int(m.group(1))
+            num_objects = int(m.group(2))
+            for i in range(first_object, first_object + num_objects):
+                m = self.re_xref_entry.match(self.buf, offset)
+                check_format_condition(m is not None, "xref entry not found")
+                assert m is not None
+                offset = m.end()
+                is_free = m.group(3) == b"f"
+                if not is_free:
+                    generation = int(m.group(2))
+                    new_entry = (int(m.group(1)), generation)
+                    if i not in self.xref_table:
+                        self.xref_table[i] = new_entry
+        return offset
+
+    def read_indirect(self, ref: IndirectReference, max_nesting: int = -1) -> Any:
+        offset, generation = self.xref_table[ref[0]]
+        check_format_condition(
+            generation == ref[1],
+            f"expected to find generation {ref[1]} for object ID {ref[0]} in xref "
+            f"table, instead found generation {generation} at offset {offset}",
+        )
+        assert self.buf is not None
+        value = self.get_value(
+            self.buf,
+            offset + self.start_offset,
+            expect_indirect=IndirectReference(*ref),
+            max_nesting=max_nesting,
+        )[0]
+        self.cached_objects[ref] = value
+        return value
+
+    def linearize_page_tree(
+        self, node: PdfDict | None = None
+    ) -> list[IndirectReference]:
+        page_node = node if node is not None else self.page_tree_root
+        check_format_condition(
+            page_node[b"Type"] == b"Pages", "/Type of page tree node is not /Pages"
+        )
+        pages = []
+        for kid in page_node[b"Kids"]:
+            kid_object = self.read_indirect(kid)
+            if kid_object[b"Type"] == b"Page":
+                pages.append(kid)
+            else:
+                pages.extend(self.linearize_page_tree(node=kid_object))
+        return pages
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PixarImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PixarImagePlugin.py
new file mode 100644
index 00000000..5c465bbd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PixarImagePlugin.py
@@ -0,0 +1,72 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PIXAR raster support for PIL
+#
+# history:
+#       97-01-29 fl     Created
+#
+# notes:
+#       This is incomplete; it is based on a few samples created with
+#       Photoshop 2.5 and 3.0, and a summary description provided by
+#       Greg Coats <gcoats@labiris.er.usgs.gov>.  Hopefully, "L" and
+#       "RGBA" support will be added in future versions.
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1997.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image, ImageFile
+from ._binary import i16le as i16
+
+#
+# helpers
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"\200\350\000\000"
+
+
+##
+# Image plugin for PIXAR raster images.
+
+
+class PixarImageFile(ImageFile.ImageFile):
+    format = "PIXAR"
+    format_description = "PIXAR raster image"
+
+    def _open(self) -> None:
+        # assuming a 4-byte magic label
+        assert self.fp is not None
+
+        s = self.fp.read(4)
+        if not _accept(s):
+            msg = "not a PIXAR file"
+            raise SyntaxError(msg)
+
+        # read rest of header
+        s = s + self.fp.read(508)
+
+        self._size = i16(s, 418), i16(s, 416)
+
+        # get channel/depth descriptions
+        mode = i16(s, 424), i16(s, 426)
+
+        if mode == (14, 2):
+            self._mode = "RGB"
+        # FIXME: to be continued...
+
+        # create tile descriptor (assuming "dumped")
+        self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, 1024, self.mode)]
+
+
+#
+# --------------------------------------------------------------------
+
+Image.register_open(PixarImageFile.format, PixarImageFile, _accept)
+
+Image.register_extension(PixarImageFile.format, ".pxr")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PngImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PngImagePlugin.py
new file mode 100644
index 00000000..4b97992a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PngImagePlugin.py
@@ -0,0 +1,1544 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PNG support code
+#
+# See "PNG (Portable Network Graphics) Specification, version 1.0;
+# W3C Recommendation", 1996-10-01, Thomas Boutell (ed.).
+#
+# history:
+# 1996-05-06 fl   Created (couldn't resist it)
+# 1996-12-14 fl   Upgraded, added read and verify support (0.2)
+# 1996-12-15 fl   Separate PNG stream parser
+# 1996-12-29 fl   Added write support, added getchunks
+# 1996-12-30 fl   Eliminated circular references in decoder (0.3)
+# 1998-07-12 fl   Read/write 16-bit images as mode I (0.4)
+# 2001-02-08 fl   Added transparency support (from Zircon) (0.5)
+# 2001-04-16 fl   Don't close data source in "open" method (0.6)
+# 2004-02-24 fl   Don't even pretend to support interlaced files (0.7)
+# 2004-08-31 fl   Do basic sanity check on chunk identifiers (0.8)
+# 2004-09-20 fl   Added PngInfo chunk container
+# 2004-12-18 fl   Added DPI read support (based on code by Niki Spahiev)
+# 2008-08-13 fl   Added tRNS support for RGB images
+# 2009-03-06 fl   Support for preserving ICC profiles (by Florian Hoech)
+# 2009-03-08 fl   Added zTXT support (from Lowell Alleman)
+# 2009-03-29 fl   Read interlaced PNG files (from Conrado Porto Lopes Gouvua)
+#
+# Copyright (c) 1997-2009 by Secret Labs AB
+# Copyright (c) 1996 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import itertools
+import logging
+import re
+import struct
+import warnings
+import zlib
+from collections.abc import Callable
+from enum import IntEnum
+from typing import IO, TYPE_CHECKING, Any, NamedTuple, NoReturn, cast
+
+from . import Image, ImageChops, ImageFile, ImagePalette, ImageSequence
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._binary import o8
+from ._binary import o16be as o16
+from ._binary import o32be as o32
+
+if TYPE_CHECKING:
+    from . import _imaging
+
+logger = logging.getLogger(__name__)
+
+is_cid = re.compile(rb"\w\w\w\w").match
+
+
+_MAGIC = b"\211PNG\r\n\032\n"
+
+
+_MODES = {
+    # supported bits/color combinations, and corresponding modes/rawmodes
+    # Grayscale
+    (1, 0): ("1", "1"),
+    (2, 0): ("L", "L;2"),
+    (4, 0): ("L", "L;4"),
+    (8, 0): ("L", "L"),
+    (16, 0): ("I;16", "I;16B"),
+    # Truecolour
+    (8, 2): ("RGB", "RGB"),
+    (16, 2): ("RGB", "RGB;16B"),
+    # Indexed-colour
+    (1, 3): ("P", "P;1"),
+    (2, 3): ("P", "P;2"),
+    (4, 3): ("P", "P;4"),
+    (8, 3): ("P", "P"),
+    # Grayscale with alpha
+    (8, 4): ("LA", "LA"),
+    (16, 4): ("RGBA", "LA;16B"),  # LA;16B->LA not yet available
+    # Truecolour with alpha
+    (8, 6): ("RGBA", "RGBA"),
+    (16, 6): ("RGBA", "RGBA;16B"),
+}
+
+
+_simple_palette = re.compile(b"^\xff*\x00\xff*$")
+
+MAX_TEXT_CHUNK = ImageFile.SAFEBLOCK
+"""
+Maximum decompressed size for a iTXt or zTXt chunk.
+Eliminates decompression bombs where compressed chunks can expand 1000x.
+See :ref:`Text in PNG File Format<png-text>`.
+"""
+MAX_TEXT_MEMORY = 64 * MAX_TEXT_CHUNK
+"""
+Set the maximum total text chunk size.
+See :ref:`Text in PNG File Format<png-text>`.
+"""
+
+
+# APNG frame disposal modes
+class Disposal(IntEnum):
+    OP_NONE = 0
+    """
+    No disposal is done on this frame before rendering the next frame.
+    See :ref:`Saving APNG sequences<apng-saving>`.
+    """
+    OP_BACKGROUND = 1
+    """
+    This frame’s modified region is cleared to fully transparent black before rendering
+    the next frame.
+    See :ref:`Saving APNG sequences<apng-saving>`.
+    """
+    OP_PREVIOUS = 2
+    """
+    This frame’s modified region is reverted to the previous frame’s contents before
+    rendering the next frame.
+    See :ref:`Saving APNG sequences<apng-saving>`.
+    """
+
+
+# APNG frame blend modes
+class Blend(IntEnum):
+    OP_SOURCE = 0
+    """
+    All color components of this frame, including alpha, overwrite the previous output
+    image contents.
+    See :ref:`Saving APNG sequences<apng-saving>`.
+    """
+    OP_OVER = 1
+    """
+    This frame should be alpha composited with the previous output image contents.
+    See :ref:`Saving APNG sequences<apng-saving>`.
+    """
+
+
+def _safe_zlib_decompress(s: bytes) -> bytes:
+    dobj = zlib.decompressobj()
+    plaintext = dobj.decompress(s, MAX_TEXT_CHUNK)
+    if dobj.unconsumed_tail:
+        msg = "Decompressed data too large for PngImagePlugin.MAX_TEXT_CHUNK"
+        raise ValueError(msg)
+    return plaintext
+
+
+def _crc32(data: bytes, seed: int = 0) -> int:
+    return zlib.crc32(data, seed) & 0xFFFFFFFF
+
+
+# --------------------------------------------------------------------
+# Support classes.  Suitable for PNG and related formats like MNG etc.
+
+
+class ChunkStream:
+    def __init__(self, fp: IO[bytes]) -> None:
+        self.fp: IO[bytes] | None = fp
+        self.queue: list[tuple[bytes, int, int]] | None = []
+
+    def read(self) -> tuple[bytes, int, int]:
+        """Fetch a new chunk. Returns header information."""
+        cid = None
+
+        assert self.fp is not None
+        if self.queue:
+            cid, pos, length = self.queue.pop()
+            self.fp.seek(pos)
+        else:
+            s = self.fp.read(8)
+            cid = s[4:]
+            pos = self.fp.tell()
+            length = i32(s)
+
+        if not is_cid(cid):
+            if not ImageFile.LOAD_TRUNCATED_IMAGES:
+                msg = f"broken PNG file (chunk {repr(cid)})"
+                raise SyntaxError(msg)
+
+        return cid, pos, length
+
+    def __enter__(self) -> ChunkStream:
+        return self
+
+    def __exit__(self, *args: object) -> None:
+        self.close()
+
+    def close(self) -> None:
+        self.queue = self.fp = None
+
+    def push(self, cid: bytes, pos: int, length: int) -> None:
+        assert self.queue is not None
+        self.queue.append((cid, pos, length))
+
+    def call(self, cid: bytes, pos: int, length: int) -> bytes:
+        """Call the appropriate chunk handler"""
+
+        logger.debug("STREAM %r %s %s", cid, pos, length)
+        return getattr(self, f"chunk_{cid.decode('ascii')}")(pos, length)
+
+    def crc(self, cid: bytes, data: bytes) -> None:
+        """Read and verify checksum"""
+
+        # Skip CRC checks for ancillary chunks if allowed to load truncated
+        # images
+        # 5th byte of first char is 1 [specs, section 5.4]
+        if ImageFile.LOAD_TRUNCATED_IMAGES and (cid[0] >> 5 & 1):
+            self.crc_skip(cid, data)
+            return
+
+        assert self.fp is not None
+        try:
+            crc1 = _crc32(data, _crc32(cid))
+            crc2 = i32(self.fp.read(4))
+            if crc1 != crc2:
+                msg = f"broken PNG file (bad header checksum in {repr(cid)})"
+                raise SyntaxError(msg)
+        except struct.error as e:
+            msg = f"broken PNG file (incomplete checksum in {repr(cid)})"
+            raise SyntaxError(msg) from e
+
+    def crc_skip(self, cid: bytes, data: bytes) -> None:
+        """Read checksum"""
+
+        assert self.fp is not None
+        self.fp.read(4)
+
+    def verify(self, endchunk: bytes = b"IEND") -> list[bytes]:
+        # Simple approach; just calculate checksum for all remaining
+        # blocks.  Must be called directly after open.
+
+        cids = []
+
+        assert self.fp is not None
+        while True:
+            try:
+                cid, pos, length = self.read()
+            except struct.error as e:
+                msg = "truncated PNG file"
+                raise OSError(msg) from e
+
+            if cid == endchunk:
+                break
+            self.crc(cid, ImageFile._safe_read(self.fp, length))
+            cids.append(cid)
+
+        return cids
+
+
+class iTXt(str):
+    """
+    Subclass of string to allow iTXt chunks to look like strings while
+    keeping their extra information
+
+    """
+
+    lang: str | bytes | None
+    tkey: str | bytes | None
+
+    @staticmethod
+    def __new__(
+        cls, text: str, lang: str | None = None, tkey: str | None = None
+    ) -> iTXt:
+        """
+        :param cls: the class to use when creating the instance
+        :param text: value for this key
+        :param lang: language code
+        :param tkey: UTF-8 version of the key name
+        """
+
+        self = str.__new__(cls, text)
+        self.lang = lang
+        self.tkey = tkey
+        return self
+
+
+class PngInfo:
+    """
+    PNG chunk container (for use with save(pnginfo=))
+
+    """
+
+    def __init__(self) -> None:
+        self.chunks: list[tuple[bytes, bytes, bool]] = []
+
+    def add(self, cid: bytes, data: bytes, after_idat: bool = False) -> None:
+        """Appends an arbitrary chunk. Use with caution.
+
+        :param cid: a byte string, 4 bytes long.
+        :param data: a byte string of the encoded data
+        :param after_idat: for use with private chunks. Whether the chunk
+                           should be written after IDAT
+
+        """
+
+        self.chunks.append((cid, data, after_idat))
+
+    def add_itxt(
+        self,
+        key: str | bytes,
+        value: str | bytes,
+        lang: str | bytes = "",
+        tkey: str | bytes = "",
+        zip: bool = False,
+    ) -> None:
+        """Appends an iTXt chunk.
+
+        :param key: latin-1 encodable text key name
+        :param value: value for this key
+        :param lang: language code
+        :param tkey: UTF-8 version of the key name
+        :param zip: compression flag
+
+        """
+
+        if not isinstance(key, bytes):
+            key = key.encode("latin-1", "strict")
+        if not isinstance(value, bytes):
+            value = value.encode("utf-8", "strict")
+        if not isinstance(lang, bytes):
+            lang = lang.encode("utf-8", "strict")
+        if not isinstance(tkey, bytes):
+            tkey = tkey.encode("utf-8", "strict")
+
+        if zip:
+            self.add(
+                b"iTXt",
+                key + b"\0\x01\0" + lang + b"\0" + tkey + b"\0" + zlib.compress(value),
+            )
+        else:
+            self.add(b"iTXt", key + b"\0\0\0" + lang + b"\0" + tkey + b"\0" + value)
+
+    def add_text(
+        self, key: str | bytes, value: str | bytes | iTXt, zip: bool = False
+    ) -> None:
+        """Appends a text chunk.
+
+        :param key: latin-1 encodable text key name
+        :param value: value for this key, text or an
+           :py:class:`PIL.PngImagePlugin.iTXt` instance
+        :param zip: compression flag
+
+        """
+        if isinstance(value, iTXt):
+            return self.add_itxt(
+                key,
+                value,
+                value.lang if value.lang is not None else b"",
+                value.tkey if value.tkey is not None else b"",
+                zip=zip,
+            )
+
+        # The tEXt chunk stores latin-1 text
+        if not isinstance(value, bytes):
+            try:
+                value = value.encode("latin-1", "strict")
+            except UnicodeError:
+                return self.add_itxt(key, value, zip=zip)
+
+        if not isinstance(key, bytes):
+            key = key.encode("latin-1", "strict")
+
+        if zip:
+            self.add(b"zTXt", key + b"\0\0" + zlib.compress(value))
+        else:
+            self.add(b"tEXt", key + b"\0" + value)
+
+
+# --------------------------------------------------------------------
+# PNG image stream (IHDR/IEND)
+
+
+class _RewindState(NamedTuple):
+    info: dict[str | tuple[int, int], Any]
+    tile: list[ImageFile._Tile]
+    seq_num: int | None
+
+
+class PngStream(ChunkStream):
+    def __init__(self, fp: IO[bytes]) -> None:
+        super().__init__(fp)
+
+        # local copies of Image attributes
+        self.im_info: dict[str | tuple[int, int], Any] = {}
+        self.im_text: dict[str, str | iTXt] = {}
+        self.im_size = (0, 0)
+        self.im_mode = ""
+        self.im_tile: list[ImageFile._Tile] = []
+        self.im_palette: tuple[str, bytes] | None = None
+        self.im_custom_mimetype: str | None = None
+        self.im_n_frames: int | None = None
+        self._seq_num: int | None = None
+        self.rewind_state = _RewindState({}, [], None)
+
+        self.text_memory = 0
+
+    def check_text_memory(self, chunklen: int) -> None:
+        self.text_memory += chunklen
+        if self.text_memory > MAX_TEXT_MEMORY:
+            msg = (
+                "Too much memory used in text chunks: "
+                f"{self.text_memory}>MAX_TEXT_MEMORY"
+            )
+            raise ValueError(msg)
+
+    def save_rewind(self) -> None:
+        self.rewind_state = _RewindState(
+            self.im_info.copy(),
+            self.im_tile,
+            self._seq_num,
+        )
+
+    def rewind(self) -> None:
+        self.im_info = self.rewind_state.info.copy()
+        self.im_tile = self.rewind_state.tile
+        self._seq_num = self.rewind_state.seq_num
+
+    def chunk_iCCP(self, pos: int, length: int) -> bytes:
+        # ICC profile
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        # according to PNG spec, the iCCP chunk contains:
+        # Profile name  1-79 bytes (character string)
+        # Null separator        1 byte (null character)
+        # Compression method    1 byte (0)
+        # Compressed profile    n bytes (zlib with deflate compression)
+        i = s.find(b"\0")
+        logger.debug("iCCP profile name %r", s[:i])
+        comp_method = s[i + 1]
+        logger.debug("Compression method %s", comp_method)
+        if comp_method != 0:
+            msg = f"Unknown compression method {comp_method} in iCCP chunk"
+            raise SyntaxError(msg)
+        try:
+            icc_profile = _safe_zlib_decompress(s[i + 2 :])
+        except ValueError:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                icc_profile = None
+            else:
+                raise
+        except zlib.error:
+            icc_profile = None  # FIXME
+        self.im_info["icc_profile"] = icc_profile
+        return s
+
+    def chunk_IHDR(self, pos: int, length: int) -> bytes:
+        # image header
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if length < 13:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                return s
+            msg = "Truncated IHDR chunk"
+            raise ValueError(msg)
+        self.im_size = i32(s, 0), i32(s, 4)
+        try:
+            self.im_mode, self.im_rawmode = _MODES[(s[8], s[9])]
+        except Exception:
+            pass
+        if s[12]:
+            self.im_info["interlace"] = 1
+        if s[11]:
+            msg = "unknown filter category"
+            raise SyntaxError(msg)
+        return s
+
+    def chunk_IDAT(self, pos: int, length: int) -> NoReturn:
+        # image data
+        if "bbox" in self.im_info:
+            tile = [ImageFile._Tile("zip", self.im_info["bbox"], pos, self.im_rawmode)]
+        else:
+            if self.im_n_frames is not None:
+                self.im_info["default_image"] = True
+            tile = [ImageFile._Tile("zip", (0, 0) + self.im_size, pos, self.im_rawmode)]
+        self.im_tile = tile
+        self.im_idat = length
+        msg = "image data found"
+        raise EOFError(msg)
+
+    def chunk_IEND(self, pos: int, length: int) -> NoReturn:
+        msg = "end of PNG image"
+        raise EOFError(msg)
+
+    def chunk_PLTE(self, pos: int, length: int) -> bytes:
+        # palette
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if self.im_mode == "P":
+            self.im_palette = "RGB", s
+        return s
+
+    def chunk_tRNS(self, pos: int, length: int) -> bytes:
+        # transparency
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if self.im_mode == "P":
+            if _simple_palette.match(s):
+                # tRNS contains only one full-transparent entry,
+                # other entries are full opaque
+                i = s.find(b"\0")
+                if i >= 0:
+                    self.im_info["transparency"] = i
+            else:
+                # otherwise, we have a byte string with one alpha value
+                # for each palette entry
+                self.im_info["transparency"] = s
+        elif self.im_mode in ("1", "L", "I;16"):
+            self.im_info["transparency"] = i16(s)
+        elif self.im_mode == "RGB":
+            self.im_info["transparency"] = i16(s), i16(s, 2), i16(s, 4)
+        return s
+
+    def chunk_gAMA(self, pos: int, length: int) -> bytes:
+        # gamma setting
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        self.im_info["gamma"] = i32(s) / 100000.0
+        return s
+
+    def chunk_cHRM(self, pos: int, length: int) -> bytes:
+        # chromaticity, 8 unsigned ints, actual value is scaled by 100,000
+        # WP x,y, Red x,y, Green x,y Blue x,y
+
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        raw_vals = struct.unpack(f">{len(s) // 4}I", s)
+        self.im_info["chromaticity"] = tuple(elt / 100000.0 for elt in raw_vals)
+        return s
+
+    def chunk_sRGB(self, pos: int, length: int) -> bytes:
+        # srgb rendering intent, 1 byte
+        # 0 perceptual
+        # 1 relative colorimetric
+        # 2 saturation
+        # 3 absolute colorimetric
+
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if length < 1:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                return s
+            msg = "Truncated sRGB chunk"
+            raise ValueError(msg)
+        self.im_info["srgb"] = s[0]
+        return s
+
+    def chunk_pHYs(self, pos: int, length: int) -> bytes:
+        # pixels per unit
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if length < 9:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                return s
+            msg = "Truncated pHYs chunk"
+            raise ValueError(msg)
+        px, py = i32(s, 0), i32(s, 4)
+        unit = s[8]
+        if unit == 1:  # meter
+            dpi = px * 0.0254, py * 0.0254
+            self.im_info["dpi"] = dpi
+        elif unit == 0:
+            self.im_info["aspect"] = px, py
+        return s
+
+    def chunk_tEXt(self, pos: int, length: int) -> bytes:
+        # text
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        try:
+            k, v = s.split(b"\0", 1)
+        except ValueError:
+            # fallback for broken tEXt tags
+            k = s
+            v = b""
+        if k:
+            k_str = k.decode("latin-1", "strict")
+            v_str = v.decode("latin-1", "replace")
+
+            self.im_info[k_str] = v if k == b"exif" else v_str
+            self.im_text[k_str] = v_str
+            self.check_text_memory(len(v_str))
+
+        return s
+
+    def chunk_zTXt(self, pos: int, length: int) -> bytes:
+        # compressed text
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        try:
+            k, v = s.split(b"\0", 1)
+        except ValueError:
+            k = s
+            v = b""
+        if v:
+            comp_method = v[0]
+        else:
+            comp_method = 0
+        if comp_method != 0:
+            msg = f"Unknown compression method {comp_method} in zTXt chunk"
+            raise SyntaxError(msg)
+        try:
+            v = _safe_zlib_decompress(v[1:])
+        except ValueError:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                v = b""
+            else:
+                raise
+        except zlib.error:
+            v = b""
+
+        if k:
+            k_str = k.decode("latin-1", "strict")
+            v_str = v.decode("latin-1", "replace")
+
+            self.im_info[k_str] = self.im_text[k_str] = v_str
+            self.check_text_memory(len(v_str))
+
+        return s
+
+    def chunk_iTXt(self, pos: int, length: int) -> bytes:
+        # international text
+        assert self.fp is not None
+        r = s = ImageFile._safe_read(self.fp, length)
+        try:
+            k, r = r.split(b"\0", 1)
+        except ValueError:
+            return s
+        if len(r) < 2:
+            return s
+        cf, cm, r = r[0], r[1], r[2:]
+        try:
+            lang, tk, v = r.split(b"\0", 2)
+        except ValueError:
+            return s
+        if cf != 0:
+            if cm == 0:
+                try:
+                    v = _safe_zlib_decompress(v)
+                except ValueError:
+                    if ImageFile.LOAD_TRUNCATED_IMAGES:
+                        return s
+                    else:
+                        raise
+                except zlib.error:
+                    return s
+            else:
+                return s
+        if k == b"XML:com.adobe.xmp":
+            self.im_info["xmp"] = v
+        try:
+            k_str = k.decode("latin-1", "strict")
+            lang_str = lang.decode("utf-8", "strict")
+            tk_str = tk.decode("utf-8", "strict")
+            v_str = v.decode("utf-8", "strict")
+        except UnicodeError:
+            return s
+
+        self.im_info[k_str] = self.im_text[k_str] = iTXt(v_str, lang_str, tk_str)
+        self.check_text_memory(len(v_str))
+
+        return s
+
+    def chunk_eXIf(self, pos: int, length: int) -> bytes:
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        self.im_info["exif"] = b"Exif\x00\x00" + s
+        return s
+
+    # APNG chunks
+    def chunk_acTL(self, pos: int, length: int) -> bytes:
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if length < 8:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                return s
+            msg = "APNG contains truncated acTL chunk"
+            raise ValueError(msg)
+        if self.im_n_frames is not None:
+            self.im_n_frames = None
+            warnings.warn("Invalid APNG, will use default PNG image if possible")
+            return s
+        n_frames = i32(s)
+        if n_frames == 0 or n_frames > 0x80000000:
+            warnings.warn("Invalid APNG, will use default PNG image if possible")
+            return s
+        self.im_n_frames = n_frames
+        self.im_info["loop"] = i32(s, 4)
+        self.im_custom_mimetype = "image/apng"
+        return s
+
+    def chunk_fcTL(self, pos: int, length: int) -> bytes:
+        assert self.fp is not None
+        s = ImageFile._safe_read(self.fp, length)
+        if length < 26:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                return s
+            msg = "APNG contains truncated fcTL chunk"
+            raise ValueError(msg)
+        seq = i32(s)
+        if (self._seq_num is None and seq != 0) or (
+            self._seq_num is not None and self._seq_num != seq - 1
+        ):
+            msg = "APNG contains frame sequence errors"
+            raise SyntaxError(msg)
+        self._seq_num = seq
+        width, height = i32(s, 4), i32(s, 8)
+        px, py = i32(s, 12), i32(s, 16)
+        im_w, im_h = self.im_size
+        if px + width > im_w or py + height > im_h:
+            msg = "APNG contains invalid frames"
+            raise SyntaxError(msg)
+        self.im_info["bbox"] = (px, py, px + width, py + height)
+        delay_num, delay_den = i16(s, 20), i16(s, 22)
+        if delay_den == 0:
+            delay_den = 100
+        self.im_info["duration"] = float(delay_num) / float(delay_den) * 1000
+        self.im_info["disposal"] = s[24]
+        self.im_info["blend"] = s[25]
+        return s
+
+    def chunk_fdAT(self, pos: int, length: int) -> bytes:
+        assert self.fp is not None
+        if length < 4:
+            if ImageFile.LOAD_TRUNCATED_IMAGES:
+                s = ImageFile._safe_read(self.fp, length)
+                return s
+            msg = "APNG contains truncated fDAT chunk"
+            raise ValueError(msg)
+        s = ImageFile._safe_read(self.fp, 4)
+        seq = i32(s)
+        if self._seq_num != seq - 1:
+            msg = "APNG contains frame sequence errors"
+            raise SyntaxError(msg)
+        self._seq_num = seq
+        return self.chunk_IDAT(pos + 4, length - 4)
+
+
+# --------------------------------------------------------------------
+# PNG reader
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:8] == _MAGIC
+
+
+##
+# Image plugin for PNG images.
+
+
+class PngImageFile(ImageFile.ImageFile):
+    format = "PNG"
+    format_description = "Portable network graphics"
+
+    def _open(self) -> None:
+        if not _accept(self.fp.read(8)):
+            msg = "not a PNG file"
+            raise SyntaxError(msg)
+        self._fp = self.fp
+        self.__frame = 0
+
+        #
+        # Parse headers up to the first IDAT or fDAT chunk
+
+        self.private_chunks: list[tuple[bytes, bytes] | tuple[bytes, bytes, bool]] = []
+        self.png: PngStream | None = PngStream(self.fp)
+
+        while True:
+            #
+            # get next chunk
+
+            cid, pos, length = self.png.read()
+
+            try:
+                s = self.png.call(cid, pos, length)
+            except EOFError:
+                break
+            except AttributeError:
+                logger.debug("%r %s %s (unknown)", cid, pos, length)
+                s = ImageFile._safe_read(self.fp, length)
+                if cid[1:2].islower():
+                    self.private_chunks.append((cid, s))
+
+            self.png.crc(cid, s)
+
+        #
+        # Copy relevant attributes from the PngStream.  An alternative
+        # would be to let the PngStream class modify these attributes
+        # directly, but that introduces circular references which are
+        # difficult to break if things go wrong in the decoder...
+        # (believe me, I've tried ;-)
+
+        self._mode = self.png.im_mode
+        self._size = self.png.im_size
+        self.info = self.png.im_info
+        self._text: dict[str, str | iTXt] | None = None
+        self.tile = self.png.im_tile
+        self.custom_mimetype = self.png.im_custom_mimetype
+        self.n_frames = self.png.im_n_frames or 1
+        self.default_image = self.info.get("default_image", False)
+
+        if self.png.im_palette:
+            rawmode, data = self.png.im_palette
+            self.palette = ImagePalette.raw(rawmode, data)
+
+        if cid == b"fdAT":
+            self.__prepare_idat = length - 4
+        else:
+            self.__prepare_idat = length  # used by load_prepare()
+
+        if self.png.im_n_frames is not None:
+            self._close_exclusive_fp_after_loading = False
+            self.png.save_rewind()
+            self.__rewind_idat = self.__prepare_idat
+            self.__rewind = self._fp.tell()
+            if self.default_image:
+                # IDAT chunk contains default image and not first animation frame
+                self.n_frames += 1
+            self._seek(0)
+        self.is_animated = self.n_frames > 1
+
+    @property
+    def text(self) -> dict[str, str | iTXt]:
+        # experimental
+        if self._text is None:
+            # iTxt, tEXt and zTXt chunks may appear at the end of the file
+            # So load the file to ensure that they are read
+            if self.is_animated:
+                frame = self.__frame
+                # for APNG, seek to the final frame before loading
+                self.seek(self.n_frames - 1)
+            self.load()
+            if self.is_animated:
+                self.seek(frame)
+        assert self._text is not None
+        return self._text
+
+    def verify(self) -> None:
+        """Verify PNG file"""
+
+        if self.fp is None:
+            msg = "verify must be called directly after open"
+            raise RuntimeError(msg)
+
+        # back up to beginning of IDAT block
+        self.fp.seek(self.tile[0][2] - 8)
+
+        assert self.png is not None
+        self.png.verify()
+        self.png.close()
+
+        if self._exclusive_fp:
+            self.fp.close()
+        self.fp = None
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+        if frame < self.__frame:
+            self._seek(0, True)
+
+        last_frame = self.__frame
+        for f in range(self.__frame + 1, frame + 1):
+            try:
+                self._seek(f)
+            except EOFError as e:
+                self.seek(last_frame)
+                msg = "no more images in APNG file"
+                raise EOFError(msg) from e
+
+    def _seek(self, frame: int, rewind: bool = False) -> None:
+        assert self.png is not None
+
+        self.dispose: _imaging.ImagingCore | None
+        dispose_extent = None
+        if frame == 0:
+            if rewind:
+                self._fp.seek(self.__rewind)
+                self.png.rewind()
+                self.__prepare_idat = self.__rewind_idat
+                self._im = None
+                self.info = self.png.im_info
+                self.tile = self.png.im_tile
+                self.fp = self._fp
+            self._prev_im = None
+            self.dispose = None
+            self.default_image = self.info.get("default_image", False)
+            self.dispose_op = self.info.get("disposal")
+            self.blend_op = self.info.get("blend")
+            dispose_extent = self.info.get("bbox")
+            self.__frame = 0
+        else:
+            if frame != self.__frame + 1:
+                msg = f"cannot seek to frame {frame}"
+                raise ValueError(msg)
+
+            # ensure previous frame was loaded
+            self.load()
+
+            if self.dispose:
+                self.im.paste(self.dispose, self.dispose_extent)
+            self._prev_im = self.im.copy()
+
+            self.fp = self._fp
+
+            # advance to the next frame
+            if self.__prepare_idat:
+                ImageFile._safe_read(self.fp, self.__prepare_idat)
+                self.__prepare_idat = 0
+            frame_start = False
+            while True:
+                self.fp.read(4)  # CRC
+
+                try:
+                    cid, pos, length = self.png.read()
+                except (struct.error, SyntaxError):
+                    break
+
+                if cid == b"IEND":
+                    msg = "No more images in APNG file"
+                    raise EOFError(msg)
+                if cid == b"fcTL":
+                    if frame_start:
+                        # there must be at least one fdAT chunk between fcTL chunks
+                        msg = "APNG missing frame data"
+                        raise SyntaxError(msg)
+                    frame_start = True
+
+                try:
+                    self.png.call(cid, pos, length)
+                except UnicodeDecodeError:
+                    break
+                except EOFError:
+                    if cid == b"fdAT":
+                        length -= 4
+                        if frame_start:
+                            self.__prepare_idat = length
+                            break
+                    ImageFile._safe_read(self.fp, length)
+                except AttributeError:
+                    logger.debug("%r %s %s (unknown)", cid, pos, length)
+                    ImageFile._safe_read(self.fp, length)
+
+            self.__frame = frame
+            self.tile = self.png.im_tile
+            self.dispose_op = self.info.get("disposal")
+            self.blend_op = self.info.get("blend")
+            dispose_extent = self.info.get("bbox")
+
+            if not self.tile:
+                msg = "image not found in APNG frame"
+                raise EOFError(msg)
+        if dispose_extent:
+            self.dispose_extent: tuple[float, float, float, float] = dispose_extent
+
+        # setup frame disposal (actual disposal done when needed in the next _seek())
+        if self._prev_im is None and self.dispose_op == Disposal.OP_PREVIOUS:
+            self.dispose_op = Disposal.OP_BACKGROUND
+
+        self.dispose = None
+        if self.dispose_op == Disposal.OP_PREVIOUS:
+            if self._prev_im:
+                self.dispose = self._prev_im.copy()
+                self.dispose = self._crop(self.dispose, self.dispose_extent)
+        elif self.dispose_op == Disposal.OP_BACKGROUND:
+            self.dispose = Image.core.fill(self.mode, self.size)
+            self.dispose = self._crop(self.dispose, self.dispose_extent)
+
+    def tell(self) -> int:
+        return self.__frame
+
+    def load_prepare(self) -> None:
+        """internal: prepare to read PNG file"""
+
+        if self.info.get("interlace"):
+            self.decoderconfig = self.decoderconfig + (1,)
+
+        self.__idat = self.__prepare_idat  # used by load_read()
+        ImageFile.ImageFile.load_prepare(self)
+
+    def load_read(self, read_bytes: int) -> bytes:
+        """internal: read more image data"""
+
+        assert self.png is not None
+        while self.__idat == 0:
+            # end of chunk, skip forward to next one
+
+            self.fp.read(4)  # CRC
+
+            cid, pos, length = self.png.read()
+
+            if cid not in [b"IDAT", b"DDAT", b"fdAT"]:
+                self.png.push(cid, pos, length)
+                return b""
+
+            if cid == b"fdAT":
+                try:
+                    self.png.call(cid, pos, length)
+                except EOFError:
+                    pass
+                self.__idat = length - 4  # sequence_num has already been read
+            else:
+                self.__idat = length  # empty chunks are allowed
+
+        # read more data from this chunk
+        if read_bytes <= 0:
+            read_bytes = self.__idat
+        else:
+            read_bytes = min(read_bytes, self.__idat)
+
+        self.__idat = self.__idat - read_bytes
+
+        return self.fp.read(read_bytes)
+
+    def load_end(self) -> None:
+        """internal: finished reading image data"""
+        assert self.png is not None
+        if self.__idat != 0:
+            self.fp.read(self.__idat)
+        while True:
+            self.fp.read(4)  # CRC
+
+            try:
+                cid, pos, length = self.png.read()
+            except (struct.error, SyntaxError):
+                break
+
+            if cid == b"IEND":
+                break
+            elif cid == b"fcTL" and self.is_animated:
+                # start of the next frame, stop reading
+                self.__prepare_idat = 0
+                self.png.push(cid, pos, length)
+                break
+
+            try:
+                self.png.call(cid, pos, length)
+            except UnicodeDecodeError:
+                break
+            except EOFError:
+                if cid == b"fdAT":
+                    length -= 4
+                try:
+                    ImageFile._safe_read(self.fp, length)
+                except OSError as e:
+                    if ImageFile.LOAD_TRUNCATED_IMAGES:
+                        break
+                    else:
+                        raise e
+            except AttributeError:
+                logger.debug("%r %s %s (unknown)", cid, pos, length)
+                s = ImageFile._safe_read(self.fp, length)
+                if cid[1:2].islower():
+                    self.private_chunks.append((cid, s, True))
+        self._text = self.png.im_text
+        if not self.is_animated:
+            self.png.close()
+            self.png = None
+        else:
+            if self._prev_im and self.blend_op == Blend.OP_OVER:
+                updated = self._crop(self.im, self.dispose_extent)
+                if self.im.mode == "RGB" and "transparency" in self.info:
+                    mask = updated.convert_transparent(
+                        "RGBA", self.info["transparency"]
+                    )
+                else:
+                    if self.im.mode == "P" and "transparency" in self.info:
+                        t = self.info["transparency"]
+                        if isinstance(t, bytes):
+                            updated.putpalettealphas(t)
+                        elif isinstance(t, int):
+                            updated.putpalettealpha(t)
+                    mask = updated.convert("RGBA")
+                self._prev_im.paste(updated, self.dispose_extent, mask)
+                self.im = self._prev_im
+
+    def _getexif(self) -> dict[int, Any] | None:
+        if "exif" not in self.info:
+            self.load()
+        if "exif" not in self.info and "Raw profile type exif" not in self.info:
+            return None
+        return self.getexif()._get_merged_dict()
+
+    def getexif(self) -> Image.Exif:
+        if "exif" not in self.info:
+            self.load()
+
+        return super().getexif()
+
+
+# --------------------------------------------------------------------
+# PNG writer
+
+_OUTMODES = {
+    # supported PIL modes, and corresponding rawmode, bit depth and color type
+    "1": ("1", b"\x01", b"\x00"),
+    "L;1": ("L;1", b"\x01", b"\x00"),
+    "L;2": ("L;2", b"\x02", b"\x00"),
+    "L;4": ("L;4", b"\x04", b"\x00"),
+    "L": ("L", b"\x08", b"\x00"),
+    "LA": ("LA", b"\x08", b"\x04"),
+    "I": ("I;16B", b"\x10", b"\x00"),
+    "I;16": ("I;16B", b"\x10", b"\x00"),
+    "I;16B": ("I;16B", b"\x10", b"\x00"),
+    "P;1": ("P;1", b"\x01", b"\x03"),
+    "P;2": ("P;2", b"\x02", b"\x03"),
+    "P;4": ("P;4", b"\x04", b"\x03"),
+    "P": ("P", b"\x08", b"\x03"),
+    "RGB": ("RGB", b"\x08", b"\x02"),
+    "RGBA": ("RGBA", b"\x08", b"\x06"),
+}
+
+
+def putchunk(fp: IO[bytes], cid: bytes, *data: bytes) -> None:
+    """Write a PNG chunk (including CRC field)"""
+
+    byte_data = b"".join(data)
+
+    fp.write(o32(len(byte_data)) + cid)
+    fp.write(byte_data)
+    crc = _crc32(byte_data, _crc32(cid))
+    fp.write(o32(crc))
+
+
+class _idat:
+    # wrap output from the encoder in IDAT chunks
+
+    def __init__(self, fp: IO[bytes], chunk: Callable[..., None]) -> None:
+        self.fp = fp
+        self.chunk = chunk
+
+    def write(self, data: bytes) -> None:
+        self.chunk(self.fp, b"IDAT", data)
+
+
+class _fdat:
+    # wrap encoder output in fdAT chunks
+
+    def __init__(self, fp: IO[bytes], chunk: Callable[..., None], seq_num: int) -> None:
+        self.fp = fp
+        self.chunk = chunk
+        self.seq_num = seq_num
+
+    def write(self, data: bytes) -> None:
+        self.chunk(self.fp, b"fdAT", o32(self.seq_num), data)
+        self.seq_num += 1
+
+
+class _Frame(NamedTuple):
+    im: Image.Image
+    bbox: tuple[int, int, int, int] | None
+    encoderinfo: dict[str, Any]
+
+
+def _write_multiple_frames(
+    im: Image.Image,
+    fp: IO[bytes],
+    chunk: Callable[..., None],
+    mode: str,
+    rawmode: str,
+    default_image: Image.Image | None,
+    append_images: list[Image.Image],
+) -> Image.Image | None:
+    duration = im.encoderinfo.get("duration")
+    loop = im.encoderinfo.get("loop", im.info.get("loop", 0))
+    disposal = im.encoderinfo.get("disposal", im.info.get("disposal", Disposal.OP_NONE))
+    blend = im.encoderinfo.get("blend", im.info.get("blend", Blend.OP_SOURCE))
+
+    if default_image:
+        chain = itertools.chain(append_images)
+    else:
+        chain = itertools.chain([im], append_images)
+
+    im_frames: list[_Frame] = []
+    frame_count = 0
+    for im_seq in chain:
+        for im_frame in ImageSequence.Iterator(im_seq):
+            if im_frame.mode == mode:
+                im_frame = im_frame.copy()
+            else:
+                im_frame = im_frame.convert(mode)
+            encoderinfo = im.encoderinfo.copy()
+            if isinstance(duration, (list, tuple)):
+                encoderinfo["duration"] = duration[frame_count]
+            elif duration is None and "duration" in im_frame.info:
+                encoderinfo["duration"] = im_frame.info["duration"]
+            if isinstance(disposal, (list, tuple)):
+                encoderinfo["disposal"] = disposal[frame_count]
+            if isinstance(blend, (list, tuple)):
+                encoderinfo["blend"] = blend[frame_count]
+            frame_count += 1
+
+            if im_frames:
+                previous = im_frames[-1]
+                prev_disposal = previous.encoderinfo.get("disposal")
+                prev_blend = previous.encoderinfo.get("blend")
+                if prev_disposal == Disposal.OP_PREVIOUS and len(im_frames) < 2:
+                    prev_disposal = Disposal.OP_BACKGROUND
+
+                if prev_disposal == Disposal.OP_BACKGROUND:
+                    base_im = previous.im.copy()
+                    dispose = Image.core.fill("RGBA", im.size, (0, 0, 0, 0))
+                    bbox = previous.bbox
+                    if bbox:
+                        dispose = dispose.crop(bbox)
+                    else:
+                        bbox = (0, 0) + im.size
+                    base_im.paste(dispose, bbox)
+                elif prev_disposal == Disposal.OP_PREVIOUS:
+                    base_im = im_frames[-2].im
+                else:
+                    base_im = previous.im
+                delta = ImageChops.subtract_modulo(
+                    im_frame.convert("RGBA"), base_im.convert("RGBA")
+                )
+                bbox = delta.getbbox(alpha_only=False)
+                if (
+                    not bbox
+                    and prev_disposal == encoderinfo.get("disposal")
+                    and prev_blend == encoderinfo.get("blend")
+                    and "duration" in encoderinfo
+                ):
+                    previous.encoderinfo["duration"] += encoderinfo["duration"]
+                    continue
+            else:
+                bbox = None
+            im_frames.append(_Frame(im_frame, bbox, encoderinfo))
+
+    if len(im_frames) == 1 and not default_image:
+        return im_frames[0].im
+
+    # animation control
+    chunk(
+        fp,
+        b"acTL",
+        o32(len(im_frames)),  # 0: num_frames
+        o32(loop),  # 4: num_plays
+    )
+
+    # default image IDAT (if it exists)
+    if default_image:
+        if im.mode != mode:
+            im = im.convert(mode)
+        ImageFile._save(
+            im,
+            cast(IO[bytes], _idat(fp, chunk)),
+            [ImageFile._Tile("zip", (0, 0) + im.size, 0, rawmode)],
+        )
+
+    seq_num = 0
+    for frame, frame_data in enumerate(im_frames):
+        im_frame = frame_data.im
+        if not frame_data.bbox:
+            bbox = (0, 0) + im_frame.size
+        else:
+            bbox = frame_data.bbox
+            im_frame = im_frame.crop(bbox)
+        size = im_frame.size
+        encoderinfo = frame_data.encoderinfo
+        frame_duration = int(round(encoderinfo.get("duration", 0)))
+        frame_disposal = encoderinfo.get("disposal", disposal)
+        frame_blend = encoderinfo.get("blend", blend)
+        # frame control
+        chunk(
+            fp,
+            b"fcTL",
+            o32(seq_num),  # sequence_number
+            o32(size[0]),  # width
+            o32(size[1]),  # height
+            o32(bbox[0]),  # x_offset
+            o32(bbox[1]),  # y_offset
+            o16(frame_duration),  # delay_numerator
+            o16(1000),  # delay_denominator
+            o8(frame_disposal),  # dispose_op
+            o8(frame_blend),  # blend_op
+        )
+        seq_num += 1
+        # frame data
+        if frame == 0 and not default_image:
+            # first frame must be in IDAT chunks for backwards compatibility
+            ImageFile._save(
+                im_frame,
+                cast(IO[bytes], _idat(fp, chunk)),
+                [ImageFile._Tile("zip", (0, 0) + im_frame.size, 0, rawmode)],
+            )
+        else:
+            fdat_chunks = _fdat(fp, chunk, seq_num)
+            ImageFile._save(
+                im_frame,
+                cast(IO[bytes], fdat_chunks),
+                [ImageFile._Tile("zip", (0, 0) + im_frame.size, 0, rawmode)],
+            )
+            seq_num = fdat_chunks.seq_num
+    return None
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    _save(im, fp, filename, save_all=True)
+
+
+def _save(
+    im: Image.Image,
+    fp: IO[bytes],
+    filename: str | bytes,
+    chunk: Callable[..., None] = putchunk,
+    save_all: bool = False,
+) -> None:
+    # save an image to disk (called by the save method)
+
+    if save_all:
+        default_image = im.encoderinfo.get(
+            "default_image", im.info.get("default_image")
+        )
+        modes = set()
+        sizes = set()
+        append_images = im.encoderinfo.get("append_images", [])
+        for im_seq in itertools.chain([im], append_images):
+            for im_frame in ImageSequence.Iterator(im_seq):
+                modes.add(im_frame.mode)
+                sizes.add(im_frame.size)
+        for mode in ("RGBA", "RGB", "P"):
+            if mode in modes:
+                break
+        else:
+            mode = modes.pop()
+        size = tuple(max(frame_size[i] for frame_size in sizes) for i in range(2))
+    else:
+        size = im.size
+        mode = im.mode
+
+    outmode = mode
+    if mode == "P":
+        #
+        # attempt to minimize storage requirements for palette images
+        if "bits" in im.encoderinfo:
+            # number of bits specified by user
+            colors = min(1 << im.encoderinfo["bits"], 256)
+        else:
+            # check palette contents
+            if im.palette:
+                colors = max(min(len(im.palette.getdata()[1]) // 3, 256), 1)
+            else:
+                colors = 256
+
+        if colors <= 16:
+            if colors <= 2:
+                bits = 1
+            elif colors <= 4:
+                bits = 2
+            else:
+                bits = 4
+            outmode += f";{bits}"
+
+    # encoder options
+    im.encoderconfig = (
+        im.encoderinfo.get("optimize", False),
+        im.encoderinfo.get("compress_level", -1),
+        im.encoderinfo.get("compress_type", -1),
+        im.encoderinfo.get("dictionary", b""),
+    )
+
+    # get the corresponding PNG mode
+    try:
+        rawmode, bit_depth, color_type = _OUTMODES[outmode]
+    except KeyError as e:
+        msg = f"cannot write mode {mode} as PNG"
+        raise OSError(msg) from e
+
+    #
+    # write minimal PNG file
+
+    fp.write(_MAGIC)
+
+    chunk(
+        fp,
+        b"IHDR",
+        o32(size[0]),  # 0: size
+        o32(size[1]),
+        bit_depth,
+        color_type,
+        b"\0",  # 10: compression
+        b"\0",  # 11: filter category
+        b"\0",  # 12: interlace flag
+    )
+
+    chunks = [b"cHRM", b"gAMA", b"sBIT", b"sRGB", b"tIME"]
+
+    icc = im.encoderinfo.get("icc_profile", im.info.get("icc_profile"))
+    if icc:
+        # ICC profile
+        # according to PNG spec, the iCCP chunk contains:
+        # Profile name  1-79 bytes (character string)
+        # Null separator        1 byte (null character)
+        # Compression method    1 byte (0)
+        # Compressed profile    n bytes (zlib with deflate compression)
+        name = b"ICC Profile"
+        data = name + b"\0\0" + zlib.compress(icc)
+        chunk(fp, b"iCCP", data)
+
+        # You must either have sRGB or iCCP.
+        # Disallow sRGB chunks when an iCCP-chunk has been emitted.
+        chunks.remove(b"sRGB")
+
+    info = im.encoderinfo.get("pnginfo")
+    if info:
+        chunks_multiple_allowed = [b"sPLT", b"iTXt", b"tEXt", b"zTXt"]
+        for info_chunk in info.chunks:
+            cid, data = info_chunk[:2]
+            if cid in chunks:
+                chunks.remove(cid)
+                chunk(fp, cid, data)
+            elif cid in chunks_multiple_allowed:
+                chunk(fp, cid, data)
+            elif cid[1:2].islower():
+                # Private chunk
+                after_idat = len(info_chunk) == 3 and info_chunk[2]
+                if not after_idat:
+                    chunk(fp, cid, data)
+
+    if im.mode == "P":
+        palette_byte_number = colors * 3
+        palette_bytes = im.im.getpalette("RGB")[:palette_byte_number]
+        while len(palette_bytes) < palette_byte_number:
+            palette_bytes += b"\0"
+        chunk(fp, b"PLTE", palette_bytes)
+
+    transparency = im.encoderinfo.get("transparency", im.info.get("transparency", None))
+
+    if transparency or transparency == 0:
+        if im.mode == "P":
+            # limit to actual palette size
+            alpha_bytes = colors
+            if isinstance(transparency, bytes):
+                chunk(fp, b"tRNS", transparency[:alpha_bytes])
+            else:
+                transparency = max(0, min(255, transparency))
+                alpha = b"\xFF" * transparency + b"\0"
+                chunk(fp, b"tRNS", alpha[:alpha_bytes])
+        elif im.mode in ("1", "L", "I", "I;16"):
+            transparency = max(0, min(65535, transparency))
+            chunk(fp, b"tRNS", o16(transparency))
+        elif im.mode == "RGB":
+            red, green, blue = transparency
+            chunk(fp, b"tRNS", o16(red) + o16(green) + o16(blue))
+        else:
+            if "transparency" in im.encoderinfo:
+                # don't bother with transparency if it's an RGBA
+                # and it's in the info dict. It's probably just stale.
+                msg = "cannot use transparency for this mode"
+                raise OSError(msg)
+    else:
+        if im.mode == "P" and im.im.getpalettemode() == "RGBA":
+            alpha = im.im.getpalette("RGBA", "A")
+            alpha_bytes = colors
+            chunk(fp, b"tRNS", alpha[:alpha_bytes])
+
+    dpi = im.encoderinfo.get("dpi")
+    if dpi:
+        chunk(
+            fp,
+            b"pHYs",
+            o32(int(dpi[0] / 0.0254 + 0.5)),
+            o32(int(dpi[1] / 0.0254 + 0.5)),
+            b"\x01",
+        )
+
+    if info:
+        chunks = [b"bKGD", b"hIST"]
+        for info_chunk in info.chunks:
+            cid, data = info_chunk[:2]
+            if cid in chunks:
+                chunks.remove(cid)
+                chunk(fp, cid, data)
+
+    exif = im.encoderinfo.get("exif")
+    if exif:
+        if isinstance(exif, Image.Exif):
+            exif = exif.tobytes(8)
+        if exif.startswith(b"Exif\x00\x00"):
+            exif = exif[6:]
+        chunk(fp, b"eXIf", exif)
+
+    single_im: Image.Image | None = im
+    if save_all:
+        single_im = _write_multiple_frames(
+            im, fp, chunk, mode, rawmode, default_image, append_images
+        )
+    if single_im:
+        ImageFile._save(
+            single_im,
+            cast(IO[bytes], _idat(fp, chunk)),
+            [ImageFile._Tile("zip", (0, 0) + single_im.size, 0, rawmode)],
+        )
+
+    if info:
+        for info_chunk in info.chunks:
+            cid, data = info_chunk[:2]
+            if cid[1:2].islower():
+                # Private chunk
+                after_idat = len(info_chunk) == 3 and info_chunk[2]
+                if after_idat:
+                    chunk(fp, cid, data)
+
+    chunk(fp, b"IEND", b"")
+
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+# --------------------------------------------------------------------
+# PNG chunk converter
+
+
+def getchunks(im: Image.Image, **params: Any) -> list[tuple[bytes, bytes, bytes]]:
+    """Return a list of PNG chunks representing this image."""
+    from io import BytesIO
+
+    chunks = []
+
+    def append(fp: IO[bytes], cid: bytes, *data: bytes) -> None:
+        byte_data = b"".join(data)
+        crc = o32(_crc32(byte_data, _crc32(cid)))
+        chunks.append((cid, byte_data, crc))
+
+    fp = BytesIO()
+
+    try:
+        im.encoderinfo = params
+        _save(im, fp, "", append)
+    finally:
+        del im.encoderinfo
+
+    return chunks
+
+
+# --------------------------------------------------------------------
+# Registry
+
+Image.register_open(PngImageFile.format, PngImageFile, _accept)
+Image.register_save(PngImageFile.format, _save)
+Image.register_save_all(PngImageFile.format, _save_all)
+
+Image.register_extensions(PngImageFile.format, [".png", ".apng"])
+
+Image.register_mime(PngImageFile.format, "image/png")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PpmImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PpmImagePlugin.py
new file mode 100644
index 00000000..4e779df1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PpmImagePlugin.py
@@ -0,0 +1,375 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# PPM support for PIL
+#
+# History:
+#       96-03-24 fl     Created
+#       98-03-06 fl     Write RGBA images (as RGB, that is)
+#
+# Copyright (c) Secret Labs AB 1997-98.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import math
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i16be as i16
+from ._binary import o8
+from ._binary import o32le as o32
+
+#
+# --------------------------------------------------------------------
+
+b_whitespace = b"\x20\x09\x0a\x0b\x0c\x0d"
+
+MODES = {
+    # standard
+    b"P1": "1",
+    b"P2": "L",
+    b"P3": "RGB",
+    b"P4": "1",
+    b"P5": "L",
+    b"P6": "RGB",
+    # extensions
+    b"P0CMYK": "CMYK",
+    b"Pf": "F",
+    # PIL extensions (for test purposes only)
+    b"PyP": "P",
+    b"PyRGBA": "RGBA",
+    b"PyCMYK": "CMYK",
+}
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[0:1] == b"P" and prefix[1] in b"0123456fy"
+
+
+##
+# Image plugin for PBM, PGM, and PPM images.
+
+
+class PpmImageFile(ImageFile.ImageFile):
+    format = "PPM"
+    format_description = "Pbmplus image"
+
+    def _read_magic(self) -> bytes:
+        assert self.fp is not None
+
+        magic = b""
+        # read until whitespace or longest available magic number
+        for _ in range(6):
+            c = self.fp.read(1)
+            if not c or c in b_whitespace:
+                break
+            magic += c
+        return magic
+
+    def _read_token(self) -> bytes:
+        assert self.fp is not None
+
+        token = b""
+        while len(token) <= 10:  # read until next whitespace or limit of 10 characters
+            c = self.fp.read(1)
+            if not c:
+                break
+            elif c in b_whitespace:  # token ended
+                if not token:
+                    # skip whitespace at start
+                    continue
+                break
+            elif c == b"#":
+                # ignores rest of the line; stops at CR, LF or EOF
+                while self.fp.read(1) not in b"\r\n":
+                    pass
+                continue
+            token += c
+        if not token:
+            # Token was not even 1 byte
+            msg = "Reached EOF while reading header"
+            raise ValueError(msg)
+        elif len(token) > 10:
+            msg = f"Token too long in file header: {token.decode()}"
+            raise ValueError(msg)
+        return token
+
+    def _open(self) -> None:
+        assert self.fp is not None
+
+        magic_number = self._read_magic()
+        try:
+            mode = MODES[magic_number]
+        except KeyError:
+            msg = "not a PPM file"
+            raise SyntaxError(msg)
+        self._mode = mode
+
+        if magic_number in (b"P1", b"P4"):
+            self.custom_mimetype = "image/x-portable-bitmap"
+        elif magic_number in (b"P2", b"P5"):
+            self.custom_mimetype = "image/x-portable-graymap"
+        elif magic_number in (b"P3", b"P6"):
+            self.custom_mimetype = "image/x-portable-pixmap"
+
+        self._size = int(self._read_token()), int(self._read_token())
+
+        decoder_name = "raw"
+        if magic_number in (b"P1", b"P2", b"P3"):
+            decoder_name = "ppm_plain"
+
+        args: str | tuple[str | int, ...]
+        if mode == "1":
+            args = "1;I"
+        elif mode == "F":
+            scale = float(self._read_token())
+            if scale == 0.0 or not math.isfinite(scale):
+                msg = "scale must be finite and non-zero"
+                raise ValueError(msg)
+            self.info["scale"] = abs(scale)
+
+            rawmode = "F;32F" if scale < 0 else "F;32BF"
+            args = (rawmode, 0, -1)
+        else:
+            maxval = int(self._read_token())
+            if not 0 < maxval < 65536:
+                msg = "maxval must be greater than 0 and less than 65536"
+                raise ValueError(msg)
+            if maxval > 255 and mode == "L":
+                self._mode = "I"
+
+            rawmode = mode
+            if decoder_name != "ppm_plain":
+                # If maxval matches a bit depth, use the raw decoder directly
+                if maxval == 65535 and mode == "L":
+                    rawmode = "I;16B"
+                elif maxval != 255:
+                    decoder_name = "ppm"
+
+            args = rawmode if decoder_name == "raw" else (rawmode, maxval)
+        self.tile = [
+            ImageFile._Tile(decoder_name, (0, 0) + self.size, self.fp.tell(), args)
+        ]
+
+
+#
+# --------------------------------------------------------------------
+
+
+class PpmPlainDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+    _comment_spans: bool
+
+    def _read_block(self) -> bytes:
+        assert self.fd is not None
+
+        return self.fd.read(ImageFile.SAFEBLOCK)
+
+    def _find_comment_end(self, block: bytes, start: int = 0) -> int:
+        a = block.find(b"\n", start)
+        b = block.find(b"\r", start)
+        return min(a, b) if a * b > 0 else max(a, b)  # lowest nonnegative index (or -1)
+
+    def _ignore_comments(self, block: bytes) -> bytes:
+        if self._comment_spans:
+            # Finish current comment
+            while block:
+                comment_end = self._find_comment_end(block)
+                if comment_end != -1:
+                    # Comment ends in this block
+                    # Delete tail of comment
+                    block = block[comment_end + 1 :]
+                    break
+                else:
+                    # Comment spans whole block
+                    # So read the next block, looking for the end
+                    block = self._read_block()
+
+        # Search for any further comments
+        self._comment_spans = False
+        while True:
+            comment_start = block.find(b"#")
+            if comment_start == -1:
+                # No comment found
+                break
+            comment_end = self._find_comment_end(block, comment_start)
+            if comment_end != -1:
+                # Comment ends in this block
+                # Delete comment
+                block = block[:comment_start] + block[comment_end + 1 :]
+            else:
+                # Comment continues to next block(s)
+                block = block[:comment_start]
+                self._comment_spans = True
+                break
+        return block
+
+    def _decode_bitonal(self) -> bytearray:
+        """
+        This is a separate method because in the plain PBM format, all data tokens are
+        exactly one byte, so the inter-token whitespace is optional.
+        """
+        data = bytearray()
+        total_bytes = self.state.xsize * self.state.ysize
+
+        while len(data) != total_bytes:
+            block = self._read_block()  # read next block
+            if not block:
+                # eof
+                break
+
+            block = self._ignore_comments(block)
+
+            tokens = b"".join(block.split())
+            for token in tokens:
+                if token not in (48, 49):
+                    msg = b"Invalid token for this mode: %s" % bytes([token])
+                    raise ValueError(msg)
+            data = (data + tokens)[:total_bytes]
+        invert = bytes.maketrans(b"01", b"\xFF\x00")
+        return data.translate(invert)
+
+    def _decode_blocks(self, maxval: int) -> bytearray:
+        data = bytearray()
+        max_len = 10
+        out_byte_count = 4 if self.mode == "I" else 1
+        out_max = 65535 if self.mode == "I" else 255
+        bands = Image.getmodebands(self.mode)
+        total_bytes = self.state.xsize * self.state.ysize * bands * out_byte_count
+
+        half_token = b""
+        while len(data) != total_bytes:
+            block = self._read_block()  # read next block
+            if not block:
+                if half_token:
+                    block = bytearray(b" ")  # flush half_token
+                else:
+                    # eof
+                    break
+
+            block = self._ignore_comments(block)
+
+            if half_token:
+                block = half_token + block  # stitch half_token to new block
+                half_token = b""
+
+            tokens = block.split()
+
+            if block and not block[-1:].isspace():  # block might split token
+                half_token = tokens.pop()  # save half token for later
+                if len(half_token) > max_len:  # prevent buildup of half_token
+                    msg = (
+                        b"Token too long found in data: %s" % half_token[: max_len + 1]
+                    )
+                    raise ValueError(msg)
+
+            for token in tokens:
+                if len(token) > max_len:
+                    msg = b"Token too long found in data: %s" % token[: max_len + 1]
+                    raise ValueError(msg)
+                value = int(token)
+                if value < 0:
+                    msg_str = f"Channel value is negative: {value}"
+                    raise ValueError(msg_str)
+                if value > maxval:
+                    msg_str = f"Channel value too large for this mode: {value}"
+                    raise ValueError(msg_str)
+                value = round(value / maxval * out_max)
+                data += o32(value) if self.mode == "I" else o8(value)
+                if len(data) == total_bytes:  # finished!
+                    break
+        return data
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        self._comment_spans = False
+        if self.mode == "1":
+            data = self._decode_bitonal()
+            rawmode = "1;8"
+        else:
+            maxval = self.args[-1]
+            data = self._decode_blocks(maxval)
+            rawmode = "I;32" if self.mode == "I" else self.mode
+        self.set_as_raw(bytes(data), rawmode)
+        return -1, 0
+
+
+class PpmDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+
+        data = bytearray()
+        maxval = self.args[-1]
+        in_byte_count = 1 if maxval < 256 else 2
+        out_byte_count = 4 if self.mode == "I" else 1
+        out_max = 65535 if self.mode == "I" else 255
+        bands = Image.getmodebands(self.mode)
+        dest_length = self.state.xsize * self.state.ysize * bands * out_byte_count
+        while len(data) < dest_length:
+            pixels = self.fd.read(in_byte_count * bands)
+            if len(pixels) < in_byte_count * bands:
+                # eof
+                break
+            for b in range(bands):
+                value = (
+                    pixels[b] if in_byte_count == 1 else i16(pixels, b * in_byte_count)
+                )
+                value = min(out_max, round(value / maxval * out_max))
+                data += o32(value) if self.mode == "I" else o8(value)
+        rawmode = "I;32" if self.mode == "I" else self.mode
+        self.set_as_raw(bytes(data), rawmode)
+        return -1, 0
+
+
+#
+# --------------------------------------------------------------------
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode == "1":
+        rawmode, head = "1;I", b"P4"
+    elif im.mode == "L":
+        rawmode, head = "L", b"P5"
+    elif im.mode in ("I", "I;16"):
+        rawmode, head = "I;16B", b"P5"
+    elif im.mode in ("RGB", "RGBA"):
+        rawmode, head = "RGB", b"P6"
+    elif im.mode == "F":
+        rawmode, head = "F;32F", b"Pf"
+    else:
+        msg = f"cannot write mode {im.mode} as PPM"
+        raise OSError(msg)
+    fp.write(head + b"\n%d %d\n" % im.size)
+    if head == b"P6":
+        fp.write(b"255\n")
+    elif head == b"P5":
+        if rawmode == "L":
+            fp.write(b"255\n")
+        else:
+            fp.write(b"65535\n")
+    elif head == b"Pf":
+        fp.write(b"-1.0\n")
+    row_order = -1 if im.mode == "F" else 1
+    ImageFile._save(
+        im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, (rawmode, 0, row_order))]
+    )
+
+
+#
+# --------------------------------------------------------------------
+
+
+Image.register_open(PpmImageFile.format, PpmImageFile, _accept)
+Image.register_save(PpmImageFile.format, _save)
+
+Image.register_decoder("ppm", PpmDecoder)
+Image.register_decoder("ppm_plain", PpmPlainDecoder)
+
+Image.register_extensions(PpmImageFile.format, [".pbm", ".pgm", ".ppm", ".pnm", ".pfm"])
+
+Image.register_mime(PpmImageFile.format, "image/x-portable-anymap")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/PsdImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/PsdImagePlugin.py
new file mode 100644
index 00000000..8ff5e390
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/PsdImagePlugin.py
@@ -0,0 +1,332 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# Adobe PSD 2.5/3.0 file handling
+#
+# History:
+# 1995-09-01 fl   Created
+# 1997-01-03 fl   Read most PSD images
+# 1997-01-18 fl   Fixed P and CMYK support
+# 2001-10-21 fl   Added seek/tell support (for layers)
+#
+# Copyright (c) 1997-2001 by Secret Labs AB.
+# Copyright (c) 1995-2001 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+from functools import cached_property
+from typing import IO
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i8
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._binary import si16be as si16
+from ._binary import si32be as si32
+
+MODES = {
+    # (photoshop mode, bits) -> (pil mode, required channels)
+    (0, 1): ("1", 1),
+    (0, 8): ("L", 1),
+    (1, 8): ("L", 1),
+    (2, 8): ("P", 1),
+    (3, 8): ("RGB", 3),
+    (4, 8): ("CMYK", 4),
+    (7, 8): ("L", 1),  # FIXME: multilayer
+    (8, 8): ("L", 1),  # duotone
+    (9, 8): ("LAB", 3),
+}
+
+
+# --------------------------------------------------------------------.
+# read PSD images
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"8BPS"
+
+
+##
+# Image plugin for Photoshop images.
+
+
+class PsdImageFile(ImageFile.ImageFile):
+    format = "PSD"
+    format_description = "Adobe Photoshop"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        read = self.fp.read
+
+        #
+        # header
+
+        s = read(26)
+        if not _accept(s) or i16(s, 4) != 1:
+            msg = "not a PSD file"
+            raise SyntaxError(msg)
+
+        psd_bits = i16(s, 22)
+        psd_channels = i16(s, 12)
+        psd_mode = i16(s, 24)
+
+        mode, channels = MODES[(psd_mode, psd_bits)]
+
+        if channels > psd_channels:
+            msg = "not enough channels"
+            raise OSError(msg)
+        if mode == "RGB" and psd_channels == 4:
+            mode = "RGBA"
+            channels = 4
+
+        self._mode = mode
+        self._size = i32(s, 18), i32(s, 14)
+
+        #
+        # color mode data
+
+        size = i32(read(4))
+        if size:
+            data = read(size)
+            if mode == "P" and size == 768:
+                self.palette = ImagePalette.raw("RGB;L", data)
+
+        #
+        # image resources
+
+        self.resources = []
+
+        size = i32(read(4))
+        if size:
+            # load resources
+            end = self.fp.tell() + size
+            while self.fp.tell() < end:
+                read(4)  # signature
+                id = i16(read(2))
+                name = read(i8(read(1)))
+                if not (len(name) & 1):
+                    read(1)  # padding
+                data = read(i32(read(4)))
+                if len(data) & 1:
+                    read(1)  # padding
+                self.resources.append((id, name, data))
+                if id == 1039:  # ICC profile
+                    self.info["icc_profile"] = data
+
+        #
+        # layer and mask information
+
+        self._layers_position = None
+
+        size = i32(read(4))
+        if size:
+            end = self.fp.tell() + size
+            size = i32(read(4))
+            if size:
+                self._layers_position = self.fp.tell()
+                self._layers_size = size
+            self.fp.seek(end)
+        self._n_frames: int | None = None
+
+        #
+        # image descriptor
+
+        self.tile = _maketile(self.fp, mode, (0, 0) + self.size, channels)
+
+        # keep the file open
+        self._fp = self.fp
+        self.frame = 1
+        self._min_frame = 1
+
+    @cached_property
+    def layers(
+        self,
+    ) -> list[tuple[str, str, tuple[int, int, int, int], list[ImageFile._Tile]]]:
+        layers = []
+        if self._layers_position is not None:
+            self._fp.seek(self._layers_position)
+            _layer_data = io.BytesIO(ImageFile._safe_read(self._fp, self._layers_size))
+            layers = _layerinfo(_layer_data, self._layers_size)
+        self._n_frames = len(layers)
+        return layers
+
+    @property
+    def n_frames(self) -> int:
+        if self._n_frames is None:
+            self._n_frames = len(self.layers)
+        return self._n_frames
+
+    @property
+    def is_animated(self) -> bool:
+        return len(self.layers) > 1
+
+    def seek(self, layer: int) -> None:
+        if not self._seek_check(layer):
+            return
+
+        # seek to given layer (1..max)
+        try:
+            _, mode, _, tile = self.layers[layer - 1]
+            self._mode = mode
+            self.tile = tile
+            self.frame = layer
+            self.fp = self._fp
+        except IndexError as e:
+            msg = "no such layer"
+            raise EOFError(msg) from e
+
+    def tell(self) -> int:
+        # return layer number (0=image, 1..max=layers)
+        return self.frame
+
+
+def _layerinfo(
+    fp: IO[bytes], ct_bytes: int
+) -> list[tuple[str, str, tuple[int, int, int, int], list[ImageFile._Tile]]]:
+    # read layerinfo block
+    layers = []
+
+    def read(size: int) -> bytes:
+        return ImageFile._safe_read(fp, size)
+
+    ct = si16(read(2))
+
+    # sanity check
+    if ct_bytes < (abs(ct) * 20):
+        msg = "Layer block too short for number of layers requested"
+        raise SyntaxError(msg)
+
+    for _ in range(abs(ct)):
+        # bounding box
+        y0 = si32(read(4))
+        x0 = si32(read(4))
+        y1 = si32(read(4))
+        x1 = si32(read(4))
+
+        # image info
+        bands = []
+        ct_types = i16(read(2))
+        if ct_types > 4:
+            fp.seek(ct_types * 6 + 12, io.SEEK_CUR)
+            size = i32(read(4))
+            fp.seek(size, io.SEEK_CUR)
+            continue
+
+        for _ in range(ct_types):
+            type = i16(read(2))
+
+            if type == 65535:
+                b = "A"
+            else:
+                b = "RGBA"[type]
+
+            bands.append(b)
+            read(4)  # size
+
+        # figure out the image mode
+        bands.sort()
+        if bands == ["R"]:
+            mode = "L"
+        elif bands == ["B", "G", "R"]:
+            mode = "RGB"
+        elif bands == ["A", "B", "G", "R"]:
+            mode = "RGBA"
+        else:
+            mode = ""  # unknown
+
+        # skip over blend flags and extra information
+        read(12)  # filler
+        name = ""
+        size = i32(read(4))  # length of the extra data field
+        if size:
+            data_end = fp.tell() + size
+
+            length = i32(read(4))
+            if length:
+                fp.seek(length - 16, io.SEEK_CUR)
+
+            length = i32(read(4))
+            if length:
+                fp.seek(length, io.SEEK_CUR)
+
+            length = i8(read(1))
+            if length:
+                # Don't know the proper encoding,
+                # Latin-1 should be a good guess
+                name = read(length).decode("latin-1", "replace")
+
+            fp.seek(data_end)
+        layers.append((name, mode, (x0, y0, x1, y1)))
+
+    # get tiles
+    layerinfo = []
+    for i, (name, mode, bbox) in enumerate(layers):
+        tile = []
+        for m in mode:
+            t = _maketile(fp, m, bbox, 1)
+            if t:
+                tile.extend(t)
+        layerinfo.append((name, mode, bbox, tile))
+
+    return layerinfo
+
+
+def _maketile(
+    file: IO[bytes], mode: str, bbox: tuple[int, int, int, int], channels: int
+) -> list[ImageFile._Tile]:
+    tiles = []
+    read = file.read
+
+    compression = i16(read(2))
+
+    xsize = bbox[2] - bbox[0]
+    ysize = bbox[3] - bbox[1]
+
+    offset = file.tell()
+
+    if compression == 0:
+        #
+        # raw compression
+        for channel in range(channels):
+            layer = mode[channel]
+            if mode == "CMYK":
+                layer += ";I"
+            tiles.append(ImageFile._Tile("raw", bbox, offset, layer))
+            offset = offset + xsize * ysize
+
+    elif compression == 1:
+        #
+        # packbits compression
+        i = 0
+        bytecount = read(channels * ysize * 2)
+        offset = file.tell()
+        for channel in range(channels):
+            layer = mode[channel]
+            if mode == "CMYK":
+                layer += ";I"
+            tiles.append(ImageFile._Tile("packbits", bbox, offset, layer))
+            for y in range(ysize):
+                offset = offset + i16(bytecount, i)
+                i += 2
+
+    file.seek(offset)
+
+    if offset & 1:
+        read(1)  # padding
+
+    return tiles
+
+
+# --------------------------------------------------------------------
+# registry
+
+
+Image.register_open(PsdImageFile.format, PsdImageFile, _accept)
+
+Image.register_extension(PsdImageFile.format, ".psd")
+
+Image.register_mime(PsdImageFile.format, "image/vnd.adobe.photoshop")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/QoiImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/QoiImagePlugin.py
new file mode 100644
index 00000000..01cc868b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/QoiImagePlugin.py
@@ -0,0 +1,115 @@
+#
+# The Python Imaging Library.
+#
+# QOI support for PIL
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+
+from . import Image, ImageFile
+from ._binary import i32be as i32
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] == b"qoif"
+
+
+class QoiImageFile(ImageFile.ImageFile):
+    format = "QOI"
+    format_description = "Quite OK Image"
+
+    def _open(self) -> None:
+        if not _accept(self.fp.read(4)):
+            msg = "not a QOI file"
+            raise SyntaxError(msg)
+
+        self._size = i32(self.fp.read(4)), i32(self.fp.read(4))
+
+        channels = self.fp.read(1)[0]
+        self._mode = "RGB" if channels == 3 else "RGBA"
+
+        self.fp.seek(1, os.SEEK_CUR)  # colorspace
+        self.tile = [ImageFile._Tile("qoi", (0, 0) + self._size, self.fp.tell())]
+
+
+class QoiDecoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+    _previous_pixel: bytes | bytearray | None = None
+    _previously_seen_pixels: dict[int, bytes | bytearray] = {}
+
+    def _add_to_previous_pixels(self, value: bytes | bytearray) -> None:
+        self._previous_pixel = value
+
+        r, g, b, a = value
+        hash_value = (r * 3 + g * 5 + b * 7 + a * 11) % 64
+        self._previously_seen_pixels[hash_value] = value
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+
+        self._previously_seen_pixels = {}
+        self._add_to_previous_pixels(bytearray((0, 0, 0, 255)))
+
+        data = bytearray()
+        bands = Image.getmodebands(self.mode)
+        dest_length = self.state.xsize * self.state.ysize * bands
+        while len(data) < dest_length:
+            byte = self.fd.read(1)[0]
+            value: bytes | bytearray
+            if byte == 0b11111110 and self._previous_pixel:  # QOI_OP_RGB
+                value = bytearray(self.fd.read(3)) + self._previous_pixel[3:]
+            elif byte == 0b11111111:  # QOI_OP_RGBA
+                value = self.fd.read(4)
+            else:
+                op = byte >> 6
+                if op == 0:  # QOI_OP_INDEX
+                    op_index = byte & 0b00111111
+                    value = self._previously_seen_pixels.get(
+                        op_index, bytearray((0, 0, 0, 0))
+                    )
+                elif op == 1 and self._previous_pixel:  # QOI_OP_DIFF
+                    value = bytearray(
+                        (
+                            (self._previous_pixel[0] + ((byte & 0b00110000) >> 4) - 2)
+                            % 256,
+                            (self._previous_pixel[1] + ((byte & 0b00001100) >> 2) - 2)
+                            % 256,
+                            (self._previous_pixel[2] + (byte & 0b00000011) - 2) % 256,
+                            self._previous_pixel[3],
+                        )
+                    )
+                elif op == 2 and self._previous_pixel:  # QOI_OP_LUMA
+                    second_byte = self.fd.read(1)[0]
+                    diff_green = (byte & 0b00111111) - 32
+                    diff_red = ((second_byte & 0b11110000) >> 4) - 8
+                    diff_blue = (second_byte & 0b00001111) - 8
+
+                    value = bytearray(
+                        tuple(
+                            (self._previous_pixel[i] + diff_green + diff) % 256
+                            for i, diff in enumerate((diff_red, 0, diff_blue))
+                        )
+                    )
+                    value += self._previous_pixel[3:]
+                elif op == 3 and self._previous_pixel:  # QOI_OP_RUN
+                    run_length = (byte & 0b00111111) + 1
+                    value = self._previous_pixel
+                    if bands == 3:
+                        value = value[:3]
+                    data += value * run_length
+                    continue
+            self._add_to_previous_pixels(value)
+
+            if bands == 3:
+                value = value[:3]
+            data += value
+        self.set_as_raw(data)
+        return -1, 0
+
+
+Image.register_open(QoiImageFile.format, QoiImageFile, _accept)
+Image.register_decoder("qoi", QoiDecoder)
+Image.register_extension(QoiImageFile.format, ".qoi")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/SgiImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/SgiImagePlugin.py
new file mode 100644
index 00000000..44254b7a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/SgiImagePlugin.py
@@ -0,0 +1,247 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# SGI image file handling
+#
+# See "The SGI Image File Format (Draft version 0.97)", Paul Haeberli.
+# <ftp://ftp.sgi.com/graphics/SGIIMAGESPEC>
+#
+#
+# History:
+# 2017-22-07 mb   Add RLE decompression
+# 2016-16-10 mb   Add save method without compression
+# 1995-09-10 fl   Created
+#
+# Copyright (c) 2016 by Mickael Bonfill.
+# Copyright (c) 2008 by Karsten Hiddemann.
+# Copyright (c) 1997 by Secret Labs AB.
+# Copyright (c) 1995 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import os
+import struct
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i16be as i16
+from ._binary import o8
+
+
+def _accept(prefix: bytes) -> bool:
+    return len(prefix) >= 2 and i16(prefix) == 474
+
+
+MODES = {
+    (1, 1, 1): "L",
+    (1, 2, 1): "L",
+    (2, 1, 1): "L;16B",
+    (2, 2, 1): "L;16B",
+    (1, 3, 3): "RGB",
+    (2, 3, 3): "RGB;16B",
+    (1, 3, 4): "RGBA",
+    (2, 3, 4): "RGBA;16B",
+}
+
+
+##
+# Image plugin for SGI images.
+class SgiImageFile(ImageFile.ImageFile):
+    format = "SGI"
+    format_description = "SGI Image File Format"
+
+    def _open(self) -> None:
+        # HEAD
+        assert self.fp is not None
+
+        headlen = 512
+        s = self.fp.read(headlen)
+
+        if not _accept(s):
+            msg = "Not an SGI image file"
+            raise ValueError(msg)
+
+        # compression : verbatim or RLE
+        compression = s[2]
+
+        # bpc : 1 or 2 bytes (8bits or 16bits)
+        bpc = s[3]
+
+        # dimension : 1, 2 or 3 (depending on xsize, ysize and zsize)
+        dimension = i16(s, 4)
+
+        # xsize : width
+        xsize = i16(s, 6)
+
+        # ysize : height
+        ysize = i16(s, 8)
+
+        # zsize : channels count
+        zsize = i16(s, 10)
+
+        # layout
+        layout = bpc, dimension, zsize
+
+        # determine mode from bits/zsize
+        rawmode = ""
+        try:
+            rawmode = MODES[layout]
+        except KeyError:
+            pass
+
+        if rawmode == "":
+            msg = "Unsupported SGI image mode"
+            raise ValueError(msg)
+
+        self._size = xsize, ysize
+        self._mode = rawmode.split(";")[0]
+        if self.mode == "RGB":
+            self.custom_mimetype = "image/rgb"
+
+        # orientation -1 : scanlines begins at the bottom-left corner
+        orientation = -1
+
+        # decoder info
+        if compression == 0:
+            pagesize = xsize * ysize * bpc
+            if bpc == 2:
+                self.tile = [
+                    ImageFile._Tile(
+                        "SGI16",
+                        (0, 0) + self.size,
+                        headlen,
+                        (self.mode, 0, orientation),
+                    )
+                ]
+            else:
+                self.tile = []
+                offset = headlen
+                for layer in self.mode:
+                    self.tile.append(
+                        ImageFile._Tile(
+                            "raw", (0, 0) + self.size, offset, (layer, 0, orientation)
+                        )
+                    )
+                    offset += pagesize
+        elif compression == 1:
+            self.tile = [
+                ImageFile._Tile(
+                    "sgi_rle", (0, 0) + self.size, headlen, (rawmode, orientation, bpc)
+                )
+            ]
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode not in {"RGB", "RGBA", "L"}:
+        msg = "Unsupported SGI image mode"
+        raise ValueError(msg)
+
+    # Get the keyword arguments
+    info = im.encoderinfo
+
+    # Byte-per-pixel precision, 1 = 8bits per pixel
+    bpc = info.get("bpc", 1)
+
+    if bpc not in (1, 2):
+        msg = "Unsupported number of bytes per pixel"
+        raise ValueError(msg)
+
+    # Flip the image, since the origin of SGI file is the bottom-left corner
+    orientation = -1
+    # Define the file as SGI File Format
+    magic_number = 474
+    # Run-Length Encoding Compression - Unsupported at this time
+    rle = 0
+
+    # Number of dimensions (x,y,z)
+    dim = 3
+    # X Dimension = width / Y Dimension = height
+    x, y = im.size
+    if im.mode == "L" and y == 1:
+        dim = 1
+    elif im.mode == "L":
+        dim = 2
+    # Z Dimension: Number of channels
+    z = len(im.mode)
+
+    if dim in {1, 2}:
+        z = 1
+
+    # assert we've got the right number of bands.
+    if len(im.getbands()) != z:
+        msg = f"incorrect number of bands in SGI write: {z} vs {len(im.getbands())}"
+        raise ValueError(msg)
+
+    # Minimum Byte value
+    pinmin = 0
+    # Maximum Byte value (255 = 8bits per pixel)
+    pinmax = 255
+    # Image name (79 characters max, truncated below in write)
+    img_name = os.path.splitext(os.path.basename(filename))[0]
+    if isinstance(img_name, str):
+        img_name = img_name.encode("ascii", "ignore")
+    # Standard representation of pixel in the file
+    colormap = 0
+    fp.write(struct.pack(">h", magic_number))
+    fp.write(o8(rle))
+    fp.write(o8(bpc))
+    fp.write(struct.pack(">H", dim))
+    fp.write(struct.pack(">H", x))
+    fp.write(struct.pack(">H", y))
+    fp.write(struct.pack(">H", z))
+    fp.write(struct.pack(">l", pinmin))
+    fp.write(struct.pack(">l", pinmax))
+    fp.write(struct.pack("4s", b""))  # dummy
+    fp.write(struct.pack("79s", img_name))  # truncates to 79 chars
+    fp.write(struct.pack("s", b""))  # force null byte after img_name
+    fp.write(struct.pack(">l", colormap))
+    fp.write(struct.pack("404s", b""))  # dummy
+
+    rawmode = "L"
+    if bpc == 2:
+        rawmode = "L;16B"
+
+    for channel in im.split():
+        fp.write(channel.tobytes("raw", rawmode, 0, orientation))
+
+    if hasattr(fp, "flush"):
+        fp.flush()
+
+
+class SGI16Decoder(ImageFile.PyDecoder):
+    _pulls_fd = True
+
+    def decode(self, buffer: bytes | Image.SupportsArrayInterface) -> tuple[int, int]:
+        assert self.fd is not None
+        assert self.im is not None
+
+        rawmode, stride, orientation = self.args
+        pagesize = self.state.xsize * self.state.ysize
+        zsize = len(self.mode)
+        self.fd.seek(512)
+
+        for band in range(zsize):
+            channel = Image.new("L", (self.state.xsize, self.state.ysize))
+            channel.frombytes(
+                self.fd.read(2 * pagesize), "raw", "L;16B", stride, orientation
+            )
+            self.im.putband(channel.im, band)
+
+        return -1, 0
+
+
+#
+# registry
+
+
+Image.register_decoder("SGI16", SGI16Decoder)
+Image.register_open(SgiImageFile.format, SgiImageFile, _accept)
+Image.register_save(SgiImageFile.format, _save)
+Image.register_mime(SgiImageFile.format, "image/sgi")
+
+Image.register_extensions(SgiImageFile.format, [".bw", ".rgb", ".rgba", ".sgi"])
+
+# End of file
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/SpiderImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/SpiderImagePlugin.py
new file mode 100644
index 00000000..3a87d009
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/SpiderImagePlugin.py
@@ -0,0 +1,326 @@
+#
+# The Python Imaging Library.
+#
+# SPIDER image file handling
+#
+# History:
+# 2004-08-02    Created BB
+# 2006-03-02    added save method
+# 2006-03-13    added support for stack images
+#
+# Copyright (c) 2004 by Health Research Inc. (HRI) RENSSELAER, NY 12144.
+# Copyright (c) 2004 by William Baxter.
+# Copyright (c) 2004 by Secret Labs AB.
+# Copyright (c) 2004 by Fredrik Lundh.
+#
+
+##
+# Image plugin for the Spider image format. This format is used
+# by the SPIDER software, in processing image data from electron
+# microscopy and tomography.
+##
+
+#
+# SpiderImagePlugin.py
+#
+# The Spider image format is used by SPIDER software, in processing
+# image data from electron microscopy and tomography.
+#
+# Spider home page:
+# https://spider.wadsworth.org/spider_doc/spider/docs/spider.html
+#
+# Details about the Spider image format:
+# https://spider.wadsworth.org/spider_doc/spider/docs/image_doc.html
+#
+from __future__ import annotations
+
+import os
+import struct
+import sys
+from typing import IO, TYPE_CHECKING, Any, cast
+
+from . import Image, ImageFile
+
+
+def isInt(f: Any) -> int:
+    try:
+        i = int(f)
+        if f - i == 0:
+            return 1
+        else:
+            return 0
+    except (ValueError, OverflowError):
+        return 0
+
+
+iforms = [1, 3, -11, -12, -21, -22]
+
+
+# There is no magic number to identify Spider files, so just check a
+# series of header locations to see if they have reasonable values.
+# Returns no. of bytes in the header, if it is a valid Spider header,
+# otherwise returns 0
+
+
+def isSpiderHeader(t: tuple[float, ...]) -> int:
+    h = (99,) + t  # add 1 value so can use spider header index start=1
+    # header values 1,2,5,12,13,22,23 should be integers
+    for i in [1, 2, 5, 12, 13, 22, 23]:
+        if not isInt(h[i]):
+            return 0
+    # check iform
+    iform = int(h[5])
+    if iform not in iforms:
+        return 0
+    # check other header values
+    labrec = int(h[13])  # no. records in file header
+    labbyt = int(h[22])  # total no. of bytes in header
+    lenbyt = int(h[23])  # record length in bytes
+    if labbyt != (labrec * lenbyt):
+        return 0
+    # looks like a valid header
+    return labbyt
+
+
+def isSpiderImage(filename: str) -> int:
+    with open(filename, "rb") as fp:
+        f = fp.read(92)  # read 23 * 4 bytes
+    t = struct.unpack(">23f", f)  # try big-endian first
+    hdrlen = isSpiderHeader(t)
+    if hdrlen == 0:
+        t = struct.unpack("<23f", f)  # little-endian
+        hdrlen = isSpiderHeader(t)
+    return hdrlen
+
+
+class SpiderImageFile(ImageFile.ImageFile):
+    format = "SPIDER"
+    format_description = "Spider 2D image"
+    _close_exclusive_fp_after_loading = False
+
+    def _open(self) -> None:
+        # check header
+        n = 27 * 4  # read 27 float values
+        f = self.fp.read(n)
+
+        try:
+            self.bigendian = 1
+            t = struct.unpack(">27f", f)  # try big-endian first
+            hdrlen = isSpiderHeader(t)
+            if hdrlen == 0:
+                self.bigendian = 0
+                t = struct.unpack("<27f", f)  # little-endian
+                hdrlen = isSpiderHeader(t)
+            if hdrlen == 0:
+                msg = "not a valid Spider file"
+                raise SyntaxError(msg)
+        except struct.error as e:
+            msg = "not a valid Spider file"
+            raise SyntaxError(msg) from e
+
+        h = (99,) + t  # add 1 value : spider header index starts at 1
+        iform = int(h[5])
+        if iform != 1:
+            msg = "not a Spider 2D image"
+            raise SyntaxError(msg)
+
+        self._size = int(h[12]), int(h[2])  # size in pixels (width, height)
+        self.istack = int(h[24])
+        self.imgnumber = int(h[27])
+
+        if self.istack == 0 and self.imgnumber == 0:
+            # stk=0, img=0: a regular 2D image
+            offset = hdrlen
+            self._nimages = 1
+        elif self.istack > 0 and self.imgnumber == 0:
+            # stk>0, img=0: Opening the stack for the first time
+            self.imgbytes = int(h[12]) * int(h[2]) * 4
+            self.hdrlen = hdrlen
+            self._nimages = int(h[26])
+            # Point to the first image in the stack
+            offset = hdrlen * 2
+            self.imgnumber = 1
+        elif self.istack == 0 and self.imgnumber > 0:
+            # stk=0, img>0: an image within the stack
+            offset = hdrlen + self.stkoffset
+            self.istack = 2  # So Image knows it's still a stack
+        else:
+            msg = "inconsistent stack header values"
+            raise SyntaxError(msg)
+
+        if self.bigendian:
+            self.rawmode = "F;32BF"
+        else:
+            self.rawmode = "F;32F"
+        self._mode = "F"
+
+        self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, offset, self.rawmode)]
+        self._fp = self.fp  # FIXME: hack
+
+    @property
+    def n_frames(self) -> int:
+        return self._nimages
+
+    @property
+    def is_animated(self) -> bool:
+        return self._nimages > 1
+
+    # 1st image index is zero (although SPIDER imgnumber starts at 1)
+    def tell(self) -> int:
+        if self.imgnumber < 1:
+            return 0
+        else:
+            return self.imgnumber - 1
+
+    def seek(self, frame: int) -> None:
+        if self.istack == 0:
+            msg = "attempt to seek in a non-stack file"
+            raise EOFError(msg)
+        if not self._seek_check(frame):
+            return
+        self.stkoffset = self.hdrlen + frame * (self.hdrlen + self.imgbytes)
+        self.fp = self._fp
+        self.fp.seek(self.stkoffset)
+        self._open()
+
+    # returns a byte image after rescaling to 0..255
+    def convert2byte(self, depth: int = 255) -> Image.Image:
+        extrema = self.getextrema()
+        assert isinstance(extrema[0], float)
+        minimum, maximum = cast(tuple[float, float], extrema)
+        m: float = 1
+        if maximum != minimum:
+            m = depth / (maximum - minimum)
+        b = -m * minimum
+        return self.point(lambda i: i * m + b).convert("L")
+
+    if TYPE_CHECKING:
+        from . import ImageTk
+
+    # returns a ImageTk.PhotoImage object, after rescaling to 0..255
+    def tkPhotoImage(self) -> ImageTk.PhotoImage:
+        from . import ImageTk
+
+        return ImageTk.PhotoImage(self.convert2byte(), palette=256)
+
+
+# --------------------------------------------------------------------
+# Image series
+
+
+# given a list of filenames, return a list of images
+def loadImageSeries(filelist: list[str] | None = None) -> list[Image.Image] | None:
+    """create a list of :py:class:`~PIL.Image.Image` objects for use in a montage"""
+    if filelist is None or len(filelist) < 1:
+        return None
+
+    byte_imgs = []
+    for img in filelist:
+        if not os.path.exists(img):
+            print(f"unable to find {img}")
+            continue
+        try:
+            with Image.open(img) as im:
+                assert isinstance(im, SpiderImageFile)
+                byte_im = im.convert2byte()
+        except Exception:
+            if not isSpiderImage(img):
+                print(f"{img} is not a Spider image file")
+            continue
+        byte_im.info["filename"] = img
+        byte_imgs.append(byte_im)
+    return byte_imgs
+
+
+# --------------------------------------------------------------------
+# For saving images in Spider format
+
+
+def makeSpiderHeader(im: Image.Image) -> list[bytes]:
+    nsam, nrow = im.size
+    lenbyt = nsam * 4  # There are labrec records in the header
+    labrec = int(1024 / lenbyt)
+    if 1024 % lenbyt != 0:
+        labrec += 1
+    labbyt = labrec * lenbyt
+    nvalues = int(labbyt / 4)
+    if nvalues < 23:
+        return []
+
+    hdr = [0.0] * nvalues
+
+    # NB these are Fortran indices
+    hdr[1] = 1.0  # nslice (=1 for an image)
+    hdr[2] = float(nrow)  # number of rows per slice
+    hdr[3] = float(nrow)  # number of records in the image
+    hdr[5] = 1.0  # iform for 2D image
+    hdr[12] = float(nsam)  # number of pixels per line
+    hdr[13] = float(labrec)  # number of records in file header
+    hdr[22] = float(labbyt)  # total number of bytes in header
+    hdr[23] = float(lenbyt)  # record length in bytes
+
+    # adjust for Fortran indexing
+    hdr = hdr[1:]
+    hdr.append(0.0)
+    # pack binary data into a string
+    return [struct.pack("f", v) for v in hdr]
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode[0] != "F":
+        im = im.convert("F")
+
+    hdr = makeSpiderHeader(im)
+    if len(hdr) < 256:
+        msg = "Error creating Spider header"
+        raise OSError(msg)
+
+    # write the SPIDER header
+    fp.writelines(hdr)
+
+    rawmode = "F;32NF"  # 32-bit native floating point
+    ImageFile._save(im, fp, [ImageFile._Tile("raw", (0, 0) + im.size, 0, rawmode)])
+
+
+def _save_spider(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    # get the filename extension and register it with Image
+    filename_ext = os.path.splitext(filename)[1]
+    ext = filename_ext.decode() if isinstance(filename_ext, bytes) else filename_ext
+    Image.register_extension(SpiderImageFile.format, ext)
+    _save(im, fp, filename)
+
+
+# --------------------------------------------------------------------
+
+
+Image.register_open(SpiderImageFile.format, SpiderImageFile)
+Image.register_save(SpiderImageFile.format, _save_spider)
+
+if __name__ == "__main__":
+    if len(sys.argv) < 2:
+        print("Syntax: python3 SpiderImagePlugin.py [infile] [outfile]")
+        sys.exit()
+
+    filename = sys.argv[1]
+    if not isSpiderImage(filename):
+        print("input image must be in Spider format")
+        sys.exit()
+
+    with Image.open(filename) as im:
+        print(f"image: {im}")
+        print(f"format: {im.format}")
+        print(f"size: {im.size}")
+        print(f"mode: {im.mode}")
+        print("max, min: ", end=" ")
+        print(im.getextrema())
+
+        if len(sys.argv) > 2:
+            outfile = sys.argv[2]
+
+            # perform some image operation
+            im = im.transpose(Image.Transpose.FLIP_LEFT_RIGHT)
+            print(
+                f"saving a flipped version of {os.path.basename(filename)} "
+                f"as {outfile} "
+            )
+            im.save(outfile, SpiderImageFile.format)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/SunImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/SunImagePlugin.py
new file mode 100644
index 00000000..8912379e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/SunImagePlugin.py
@@ -0,0 +1,145 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Sun image file handling
+#
+# History:
+# 1995-09-10 fl   Created
+# 1996-05-28 fl   Fixed 32-bit alignment
+# 1998-12-29 fl   Import ImagePalette module
+# 2001-12-18 fl   Fixed palette loading (from Jean-Claude Rimbault)
+#
+# Copyright (c) 1997-2001 by Secret Labs AB
+# Copyright (c) 1995-1996 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i32be as i32
+
+
+def _accept(prefix: bytes) -> bool:
+    return len(prefix) >= 4 and i32(prefix) == 0x59A66A95
+
+
+##
+# Image plugin for Sun raster files.
+
+
+class SunImageFile(ImageFile.ImageFile):
+    format = "SUN"
+    format_description = "Sun Raster File"
+
+    def _open(self) -> None:
+        # The Sun Raster file header is 32 bytes in length
+        # and has the following format:
+
+        #     typedef struct _SunRaster
+        #     {
+        #         DWORD MagicNumber;      /* Magic (identification) number */
+        #         DWORD Width;            /* Width of image in pixels */
+        #         DWORD Height;           /* Height of image in pixels */
+        #         DWORD Depth;            /* Number of bits per pixel */
+        #         DWORD Length;           /* Size of image data in bytes */
+        #         DWORD Type;             /* Type of raster file */
+        #         DWORD ColorMapType;     /* Type of color map */
+        #         DWORD ColorMapLength;   /* Size of the color map in bytes */
+        #     } SUNRASTER;
+
+        assert self.fp is not None
+
+        # HEAD
+        s = self.fp.read(32)
+        if not _accept(s):
+            msg = "not an SUN raster file"
+            raise SyntaxError(msg)
+
+        offset = 32
+
+        self._size = i32(s, 4), i32(s, 8)
+
+        depth = i32(s, 12)
+        # data_length = i32(s, 16)   # unreliable, ignore.
+        file_type = i32(s, 20)
+        palette_type = i32(s, 24)  # 0: None, 1: RGB, 2: Raw/arbitrary
+        palette_length = i32(s, 28)
+
+        if depth == 1:
+            self._mode, rawmode = "1", "1;I"
+        elif depth == 4:
+            self._mode, rawmode = "L", "L;4"
+        elif depth == 8:
+            self._mode = rawmode = "L"
+        elif depth == 24:
+            if file_type == 3:
+                self._mode, rawmode = "RGB", "RGB"
+            else:
+                self._mode, rawmode = "RGB", "BGR"
+        elif depth == 32:
+            if file_type == 3:
+                self._mode, rawmode = "RGB", "RGBX"
+            else:
+                self._mode, rawmode = "RGB", "BGRX"
+        else:
+            msg = "Unsupported Mode/Bit Depth"
+            raise SyntaxError(msg)
+
+        if palette_length:
+            if palette_length > 1024:
+                msg = "Unsupported Color Palette Length"
+                raise SyntaxError(msg)
+
+            if palette_type != 1:
+                msg = "Unsupported Palette Type"
+                raise SyntaxError(msg)
+
+            offset = offset + palette_length
+            self.palette = ImagePalette.raw("RGB;L", self.fp.read(palette_length))
+            if self.mode == "L":
+                self._mode = "P"
+                rawmode = rawmode.replace("L", "P")
+
+        # 16 bit boundaries on stride
+        stride = ((self.size[0] * depth + 15) // 16) * 2
+
+        # file type: Type is the version (or flavor) of the bitmap
+        # file. The following values are typically found in the Type
+        # field:
+        # 0000h Old
+        # 0001h Standard
+        # 0002h Byte-encoded
+        # 0003h RGB format
+        # 0004h TIFF format
+        # 0005h IFF format
+        # FFFFh Experimental
+
+        # Old and standard are the same, except for the length tag.
+        # byte-encoded is run-length-encoded
+        # RGB looks similar to standard, but RGB byte order
+        # TIFF and IFF mean that they were converted from T/IFF
+        # Experimental means that it's something else.
+        # (https://www.fileformat.info/format/sunraster/egff.htm)
+
+        if file_type in (0, 1, 3, 4, 5):
+            self.tile = [
+                ImageFile._Tile("raw", (0, 0) + self.size, offset, (rawmode, stride))
+            ]
+        elif file_type == 2:
+            self.tile = [
+                ImageFile._Tile("sun_rle", (0, 0) + self.size, offset, rawmode)
+            ]
+        else:
+            msg = "Unsupported Sun Raster file type"
+            raise SyntaxError(msg)
+
+
+#
+# registry
+
+
+Image.register_open(SunImageFile.format, SunImageFile, _accept)
+
+Image.register_extension(SunImageFile.format, ".ras")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/TarIO.py b/backend/venv/lib/python3.12/site-packages/PIL/TarIO.py
new file mode 100644
index 00000000..779288b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/TarIO.py
@@ -0,0 +1,57 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# read files from within a tar file
+#
+# History:
+# 95-06-18 fl   Created
+# 96-05-28 fl   Open files in binary mode
+#
+# Copyright (c) Secret Labs AB 1997.
+# Copyright (c) Fredrik Lundh 1995-96.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+
+from . import ContainerIO
+
+
+class TarIO(ContainerIO.ContainerIO[bytes]):
+    """A file object that provides read access to a given member of a TAR file."""
+
+    def __init__(self, tarfile: str, file: str) -> None:
+        """
+        Create file object.
+
+        :param tarfile: Name of TAR file.
+        :param file: Name of member file.
+        """
+        self.fh = open(tarfile, "rb")
+
+        while True:
+            s = self.fh.read(512)
+            if len(s) != 512:
+                msg = "unexpected end of tar file"
+                raise OSError(msg)
+
+            name = s[:100].decode("utf-8")
+            i = name.find("\0")
+            if i == 0:
+                msg = "cannot find subfile"
+                raise OSError(msg)
+            if i > 0:
+                name = name[:i]
+
+            size = int(s[124:135], 8)
+
+            if file == name:
+                break
+
+            self.fh.seek((size + 511) & (~511), io.SEEK_CUR)
+
+        # Open region
+        super().__init__(self.fh, self.fh.tell(), size)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/TgaImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/TgaImagePlugin.py
new file mode 100644
index 00000000..90d5b5cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/TgaImagePlugin.py
@@ -0,0 +1,264 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# TGA file handling
+#
+# History:
+# 95-09-01 fl   created (reads 24-bit files only)
+# 97-01-04 fl   support more TGA versions, including compressed images
+# 98-07-04 fl   fixed orientation and alpha layer bugs
+# 98-09-11 fl   fixed orientation for runlength decoder
+#
+# Copyright (c) Secret Labs AB 1997-98.
+# Copyright (c) Fredrik Lundh 1995-97.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import warnings
+from typing import IO
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import i16le as i16
+from ._binary import o8
+from ._binary import o16le as o16
+
+#
+# --------------------------------------------------------------------
+# Read RGA file
+
+
+MODES = {
+    # map imagetype/depth to rawmode
+    (1, 8): "P",
+    (3, 1): "1",
+    (3, 8): "L",
+    (3, 16): "LA",
+    (2, 16): "BGRA;15Z",
+    (2, 24): "BGR",
+    (2, 32): "BGRA",
+}
+
+
+##
+# Image plugin for Targa files.
+
+
+class TgaImageFile(ImageFile.ImageFile):
+    format = "TGA"
+    format_description = "Targa"
+
+    def _open(self) -> None:
+        # process header
+        assert self.fp is not None
+
+        s = self.fp.read(18)
+
+        id_len = s[0]
+
+        colormaptype = s[1]
+        imagetype = s[2]
+
+        depth = s[16]
+
+        flags = s[17]
+
+        self._size = i16(s, 12), i16(s, 14)
+
+        # validate header fields
+        if (
+            colormaptype not in (0, 1)
+            or self.size[0] <= 0
+            or self.size[1] <= 0
+            or depth not in (1, 8, 16, 24, 32)
+        ):
+            msg = "not a TGA file"
+            raise SyntaxError(msg)
+
+        # image mode
+        if imagetype in (3, 11):
+            self._mode = "L"
+            if depth == 1:
+                self._mode = "1"  # ???
+            elif depth == 16:
+                self._mode = "LA"
+        elif imagetype in (1, 9):
+            self._mode = "P" if colormaptype else "L"
+        elif imagetype in (2, 10):
+            self._mode = "RGB" if depth == 24 else "RGBA"
+        else:
+            msg = "unknown TGA mode"
+            raise SyntaxError(msg)
+
+        # orientation
+        orientation = flags & 0x30
+        self._flip_horizontally = orientation in [0x10, 0x30]
+        if orientation in [0x20, 0x30]:
+            orientation = 1
+        elif orientation in [0, 0x10]:
+            orientation = -1
+        else:
+            msg = "unknown TGA orientation"
+            raise SyntaxError(msg)
+
+        self.info["orientation"] = orientation
+
+        if imagetype & 8:
+            self.info["compression"] = "tga_rle"
+
+        if id_len:
+            self.info["id_section"] = self.fp.read(id_len)
+
+        if colormaptype:
+            # read palette
+            start, size, mapdepth = i16(s, 3), i16(s, 5), s[7]
+            if mapdepth == 16:
+                self.palette = ImagePalette.raw(
+                    "BGRA;15Z", bytes(2 * start) + self.fp.read(2 * size)
+                )
+                self.palette.mode = "RGBA"
+            elif mapdepth == 24:
+                self.palette = ImagePalette.raw(
+                    "BGR", bytes(3 * start) + self.fp.read(3 * size)
+                )
+            elif mapdepth == 32:
+                self.palette = ImagePalette.raw(
+                    "BGRA", bytes(4 * start) + self.fp.read(4 * size)
+                )
+            else:
+                msg = "unknown TGA map depth"
+                raise SyntaxError(msg)
+
+        # setup tile descriptor
+        try:
+            rawmode = MODES[(imagetype & 7, depth)]
+            if imagetype & 8:
+                # compressed
+                self.tile = [
+                    ImageFile._Tile(
+                        "tga_rle",
+                        (0, 0) + self.size,
+                        self.fp.tell(),
+                        (rawmode, orientation, depth),
+                    )
+                ]
+            else:
+                self.tile = [
+                    ImageFile._Tile(
+                        "raw",
+                        (0, 0) + self.size,
+                        self.fp.tell(),
+                        (rawmode, 0, orientation),
+                    )
+                ]
+        except KeyError:
+            pass  # cannot decode
+
+    def load_end(self) -> None:
+        if self._flip_horizontally:
+            self.im = self.im.transpose(Image.Transpose.FLIP_LEFT_RIGHT)
+
+
+#
+# --------------------------------------------------------------------
+# Write TGA file
+
+
+SAVE = {
+    "1": ("1", 1, 0, 3),
+    "L": ("L", 8, 0, 3),
+    "LA": ("LA", 16, 0, 3),
+    "P": ("P", 8, 1, 1),
+    "RGB": ("BGR", 24, 0, 2),
+    "RGBA": ("BGRA", 32, 0, 2),
+}
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    try:
+        rawmode, bits, colormaptype, imagetype = SAVE[im.mode]
+    except KeyError as e:
+        msg = f"cannot write mode {im.mode} as TGA"
+        raise OSError(msg) from e
+
+    if "rle" in im.encoderinfo:
+        rle = im.encoderinfo["rle"]
+    else:
+        compression = im.encoderinfo.get("compression", im.info.get("compression"))
+        rle = compression == "tga_rle"
+    if rle:
+        imagetype += 8
+
+    id_section = im.encoderinfo.get("id_section", im.info.get("id_section", ""))
+    id_len = len(id_section)
+    if id_len > 255:
+        id_len = 255
+        id_section = id_section[:255]
+        warnings.warn("id_section has been trimmed to 255 characters")
+
+    if colormaptype:
+        palette = im.im.getpalette("RGB", "BGR")
+        colormaplength, colormapentry = len(palette) // 3, 24
+    else:
+        colormaplength, colormapentry = 0, 0
+
+    if im.mode in ("LA", "RGBA"):
+        flags = 8
+    else:
+        flags = 0
+
+    orientation = im.encoderinfo.get("orientation", im.info.get("orientation", -1))
+    if orientation > 0:
+        flags = flags | 0x20
+
+    fp.write(
+        o8(id_len)
+        + o8(colormaptype)
+        + o8(imagetype)
+        + o16(0)  # colormapfirst
+        + o16(colormaplength)
+        + o8(colormapentry)
+        + o16(0)
+        + o16(0)
+        + o16(im.size[0])
+        + o16(im.size[1])
+        + o8(bits)
+        + o8(flags)
+    )
+
+    if id_section:
+        fp.write(id_section)
+
+    if colormaptype:
+        fp.write(palette)
+
+    if rle:
+        ImageFile._save(
+            im,
+            fp,
+            [ImageFile._Tile("tga_rle", (0, 0) + im.size, 0, (rawmode, orientation))],
+        )
+    else:
+        ImageFile._save(
+            im,
+            fp,
+            [ImageFile._Tile("raw", (0, 0) + im.size, 0, (rawmode, 0, orientation))],
+        )
+
+    # write targa version 2 footer
+    fp.write(b"\000" * 8 + b"TRUEVISION-XFILE." + b"\000")
+
+
+#
+# --------------------------------------------------------------------
+# Registry
+
+
+Image.register_open(TgaImageFile.format, TgaImageFile)
+Image.register_save(TgaImageFile.format, _save)
+
+Image.register_extensions(TgaImageFile.format, [".tga", ".icb", ".vda", ".vst"])
+
+Image.register_mime(TgaImageFile.format, "image/x-tga")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/TiffImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/TiffImagePlugin.py
new file mode 100644
index 00000000..61eb1524
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/TiffImagePlugin.py
@@ -0,0 +1,2297 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# TIFF file handling
+#
+# TIFF is a flexible, if somewhat aged, image file format originally
+# defined by Aldus.  Although TIFF supports a wide variety of pixel
+# layouts and compression methods, the name doesn't really stand for
+# "thousands of incompatible file formats," it just feels that way.
+#
+# To read TIFF data from a stream, the stream must be seekable.  For
+# progressive decoding, make sure to use TIFF files where the tag
+# directory is placed first in the file.
+#
+# History:
+# 1995-09-01 fl   Created
+# 1996-05-04 fl   Handle JPEGTABLES tag
+# 1996-05-18 fl   Fixed COLORMAP support
+# 1997-01-05 fl   Fixed PREDICTOR support
+# 1997-08-27 fl   Added support for rational tags (from Perry Stoll)
+# 1998-01-10 fl   Fixed seek/tell (from Jan Blom)
+# 1998-07-15 fl   Use private names for internal variables
+# 1999-06-13 fl   Rewritten for PIL 1.0 (1.0)
+# 2000-10-11 fl   Additional fixes for Python 2.0 (1.1)
+# 2001-04-17 fl   Fixed rewind support (seek to frame 0) (1.2)
+# 2001-05-12 fl   Added write support for more tags (from Greg Couch) (1.3)
+# 2001-12-18 fl   Added workaround for broken Matrox library
+# 2002-01-18 fl   Don't mess up if photometric tag is missing (D. Alan Stewart)
+# 2003-05-19 fl   Check FILLORDER tag
+# 2003-09-26 fl   Added RGBa support
+# 2004-02-24 fl   Added DPI support; fixed rational write support
+# 2005-02-07 fl   Added workaround for broken Corel Draw 10 files
+# 2006-01-09 fl   Added support for float/double tags (from Russell Nelson)
+#
+# Copyright (c) 1997-2006 by Secret Labs AB.  All rights reserved.
+# Copyright (c) 1995-1997 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import io
+import itertools
+import logging
+import math
+import os
+import struct
+import warnings
+from collections.abc import Iterator, MutableMapping
+from fractions import Fraction
+from numbers import Number, Rational
+from typing import IO, TYPE_CHECKING, Any, Callable, NoReturn, cast
+
+from . import ExifTags, Image, ImageFile, ImageOps, ImagePalette, TiffTags
+from ._binary import i16be as i16
+from ._binary import i32be as i32
+from ._binary import o8
+from ._deprecate import deprecate
+from ._typing import StrOrBytesPath
+from ._util import is_path
+from .TiffTags import TYPES
+
+if TYPE_CHECKING:
+    from ._typing import Buffer, IntegralLike
+
+logger = logging.getLogger(__name__)
+
+# Set these to true to force use of libtiff for reading or writing.
+READ_LIBTIFF = False
+WRITE_LIBTIFF = False
+STRIP_SIZE = 65536
+
+II = b"II"  # little-endian (Intel style)
+MM = b"MM"  # big-endian (Motorola style)
+
+#
+# --------------------------------------------------------------------
+# Read TIFF files
+
+# a few tag names, just to make the code below a bit more readable
+OSUBFILETYPE = 255
+IMAGEWIDTH = 256
+IMAGELENGTH = 257
+BITSPERSAMPLE = 258
+COMPRESSION = 259
+PHOTOMETRIC_INTERPRETATION = 262
+FILLORDER = 266
+IMAGEDESCRIPTION = 270
+STRIPOFFSETS = 273
+SAMPLESPERPIXEL = 277
+ROWSPERSTRIP = 278
+STRIPBYTECOUNTS = 279
+X_RESOLUTION = 282
+Y_RESOLUTION = 283
+PLANAR_CONFIGURATION = 284
+RESOLUTION_UNIT = 296
+TRANSFERFUNCTION = 301
+SOFTWARE = 305
+DATE_TIME = 306
+ARTIST = 315
+PREDICTOR = 317
+COLORMAP = 320
+TILEWIDTH = 322
+TILELENGTH = 323
+TILEOFFSETS = 324
+TILEBYTECOUNTS = 325
+SUBIFD = 330
+EXTRASAMPLES = 338
+SAMPLEFORMAT = 339
+JPEGTABLES = 347
+YCBCRSUBSAMPLING = 530
+REFERENCEBLACKWHITE = 532
+COPYRIGHT = 33432
+IPTC_NAA_CHUNK = 33723  # newsphoto properties
+PHOTOSHOP_CHUNK = 34377  # photoshop properties
+ICCPROFILE = 34675
+EXIFIFD = 34665
+XMP = 700
+JPEGQUALITY = 65537  # pseudo-tag by libtiff
+
+# https://github.com/imagej/ImageJA/blob/master/src/main/java/ij/io/TiffDecoder.java
+IMAGEJ_META_DATA_BYTE_COUNTS = 50838
+IMAGEJ_META_DATA = 50839
+
+COMPRESSION_INFO = {
+    # Compression => pil compression name
+    1: "raw",
+    2: "tiff_ccitt",
+    3: "group3",
+    4: "group4",
+    5: "tiff_lzw",
+    6: "tiff_jpeg",  # obsolete
+    7: "jpeg",
+    8: "tiff_adobe_deflate",
+    32771: "tiff_raw_16",  # 16-bit padding
+    32773: "packbits",
+    32809: "tiff_thunderscan",
+    32946: "tiff_deflate",
+    34676: "tiff_sgilog",
+    34677: "tiff_sgilog24",
+    34925: "lzma",
+    50000: "zstd",
+    50001: "webp",
+}
+
+COMPRESSION_INFO_REV = {v: k for k, v in COMPRESSION_INFO.items()}
+
+OPEN_INFO = {
+    # (ByteOrder, PhotoInterpretation, SampleFormat, FillOrder, BitsPerSample,
+    #  ExtraSamples) => mode, rawmode
+    (II, 0, (1,), 1, (1,), ()): ("1", "1;I"),
+    (MM, 0, (1,), 1, (1,), ()): ("1", "1;I"),
+    (II, 0, (1,), 2, (1,), ()): ("1", "1;IR"),
+    (MM, 0, (1,), 2, (1,), ()): ("1", "1;IR"),
+    (II, 1, (1,), 1, (1,), ()): ("1", "1"),
+    (MM, 1, (1,), 1, (1,), ()): ("1", "1"),
+    (II, 1, (1,), 2, (1,), ()): ("1", "1;R"),
+    (MM, 1, (1,), 2, (1,), ()): ("1", "1;R"),
+    (II, 0, (1,), 1, (2,), ()): ("L", "L;2I"),
+    (MM, 0, (1,), 1, (2,), ()): ("L", "L;2I"),
+    (II, 0, (1,), 2, (2,), ()): ("L", "L;2IR"),
+    (MM, 0, (1,), 2, (2,), ()): ("L", "L;2IR"),
+    (II, 1, (1,), 1, (2,), ()): ("L", "L;2"),
+    (MM, 1, (1,), 1, (2,), ()): ("L", "L;2"),
+    (II, 1, (1,), 2, (2,), ()): ("L", "L;2R"),
+    (MM, 1, (1,), 2, (2,), ()): ("L", "L;2R"),
+    (II, 0, (1,), 1, (4,), ()): ("L", "L;4I"),
+    (MM, 0, (1,), 1, (4,), ()): ("L", "L;4I"),
+    (II, 0, (1,), 2, (4,), ()): ("L", "L;4IR"),
+    (MM, 0, (1,), 2, (4,), ()): ("L", "L;4IR"),
+    (II, 1, (1,), 1, (4,), ()): ("L", "L;4"),
+    (MM, 1, (1,), 1, (4,), ()): ("L", "L;4"),
+    (II, 1, (1,), 2, (4,), ()): ("L", "L;4R"),
+    (MM, 1, (1,), 2, (4,), ()): ("L", "L;4R"),
+    (II, 0, (1,), 1, (8,), ()): ("L", "L;I"),
+    (MM, 0, (1,), 1, (8,), ()): ("L", "L;I"),
+    (II, 0, (1,), 2, (8,), ()): ("L", "L;IR"),
+    (MM, 0, (1,), 2, (8,), ()): ("L", "L;IR"),
+    (II, 1, (1,), 1, (8,), ()): ("L", "L"),
+    (MM, 1, (1,), 1, (8,), ()): ("L", "L"),
+    (II, 1, (2,), 1, (8,), ()): ("L", "L"),
+    (MM, 1, (2,), 1, (8,), ()): ("L", "L"),
+    (II, 1, (1,), 2, (8,), ()): ("L", "L;R"),
+    (MM, 1, (1,), 2, (8,), ()): ("L", "L;R"),
+    (II, 1, (1,), 1, (12,), ()): ("I;16", "I;12"),
+    (II, 0, (1,), 1, (16,), ()): ("I;16", "I;16"),
+    (II, 1, (1,), 1, (16,), ()): ("I;16", "I;16"),
+    (MM, 1, (1,), 1, (16,), ()): ("I;16B", "I;16B"),
+    (II, 1, (1,), 2, (16,), ()): ("I;16", "I;16R"),
+    (II, 1, (2,), 1, (16,), ()): ("I", "I;16S"),
+    (MM, 1, (2,), 1, (16,), ()): ("I", "I;16BS"),
+    (II, 0, (3,), 1, (32,), ()): ("F", "F;32F"),
+    (MM, 0, (3,), 1, (32,), ()): ("F", "F;32BF"),
+    (II, 1, (1,), 1, (32,), ()): ("I", "I;32N"),
+    (II, 1, (2,), 1, (32,), ()): ("I", "I;32S"),
+    (MM, 1, (2,), 1, (32,), ()): ("I", "I;32BS"),
+    (II, 1, (3,), 1, (32,), ()): ("F", "F;32F"),
+    (MM, 1, (3,), 1, (32,), ()): ("F", "F;32BF"),
+    (II, 1, (1,), 1, (8, 8), (2,)): ("LA", "LA"),
+    (MM, 1, (1,), 1, (8, 8), (2,)): ("LA", "LA"),
+    (II, 2, (1,), 1, (8, 8, 8), ()): ("RGB", "RGB"),
+    (MM, 2, (1,), 1, (8, 8, 8), ()): ("RGB", "RGB"),
+    (II, 2, (1,), 2, (8, 8, 8), ()): ("RGB", "RGB;R"),
+    (MM, 2, (1,), 2, (8, 8, 8), ()): ("RGB", "RGB;R"),
+    (II, 2, (1,), 1, (8, 8, 8, 8), ()): ("RGBA", "RGBA"),  # missing ExtraSamples
+    (MM, 2, (1,), 1, (8, 8, 8, 8), ()): ("RGBA", "RGBA"),  # missing ExtraSamples
+    (II, 2, (1,), 1, (8, 8, 8, 8), (0,)): ("RGB", "RGBX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8), (0,)): ("RGB", "RGBX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8), (0, 0)): ("RGB", "RGBXX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8), (0, 0)): ("RGB", "RGBXX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (0, 0, 0)): ("RGB", "RGBXXX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (0, 0, 0)): ("RGB", "RGBXXX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8), (1,)): ("RGBA", "RGBa"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8), (1,)): ("RGBA", "RGBa"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8), (1, 0)): ("RGBA", "RGBaX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8), (1, 0)): ("RGBA", "RGBaX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (1, 0, 0)): ("RGBA", "RGBaXX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (1, 0, 0)): ("RGBA", "RGBaXX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8), (2,)): ("RGBA", "RGBA"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8), (2,)): ("RGBA", "RGBA"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8), (2, 0)): ("RGBA", "RGBAX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8), (2, 0)): ("RGBA", "RGBAX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (2, 0, 0)): ("RGBA", "RGBAXX"),
+    (MM, 2, (1,), 1, (8, 8, 8, 8, 8, 8), (2, 0, 0)): ("RGBA", "RGBAXX"),
+    (II, 2, (1,), 1, (8, 8, 8, 8), (999,)): ("RGBA", "RGBA"),  # Corel Draw 10
+    (MM, 2, (1,), 1, (8, 8, 8, 8), (999,)): ("RGBA", "RGBA"),  # Corel Draw 10
+    (II, 2, (1,), 1, (16, 16, 16), ()): ("RGB", "RGB;16L"),
+    (MM, 2, (1,), 1, (16, 16, 16), ()): ("RGB", "RGB;16B"),
+    (II, 2, (1,), 1, (16, 16, 16, 16), ()): ("RGBA", "RGBA;16L"),
+    (MM, 2, (1,), 1, (16, 16, 16, 16), ()): ("RGBA", "RGBA;16B"),
+    (II, 2, (1,), 1, (16, 16, 16, 16), (0,)): ("RGB", "RGBX;16L"),
+    (MM, 2, (1,), 1, (16, 16, 16, 16), (0,)): ("RGB", "RGBX;16B"),
+    (II, 2, (1,), 1, (16, 16, 16, 16), (1,)): ("RGBA", "RGBa;16L"),
+    (MM, 2, (1,), 1, (16, 16, 16, 16), (1,)): ("RGBA", "RGBa;16B"),
+    (II, 2, (1,), 1, (16, 16, 16, 16), (2,)): ("RGBA", "RGBA;16L"),
+    (MM, 2, (1,), 1, (16, 16, 16, 16), (2,)): ("RGBA", "RGBA;16B"),
+    (II, 3, (1,), 1, (1,), ()): ("P", "P;1"),
+    (MM, 3, (1,), 1, (1,), ()): ("P", "P;1"),
+    (II, 3, (1,), 2, (1,), ()): ("P", "P;1R"),
+    (MM, 3, (1,), 2, (1,), ()): ("P", "P;1R"),
+    (II, 3, (1,), 1, (2,), ()): ("P", "P;2"),
+    (MM, 3, (1,), 1, (2,), ()): ("P", "P;2"),
+    (II, 3, (1,), 2, (2,), ()): ("P", "P;2R"),
+    (MM, 3, (1,), 2, (2,), ()): ("P", "P;2R"),
+    (II, 3, (1,), 1, (4,), ()): ("P", "P;4"),
+    (MM, 3, (1,), 1, (4,), ()): ("P", "P;4"),
+    (II, 3, (1,), 2, (4,), ()): ("P", "P;4R"),
+    (MM, 3, (1,), 2, (4,), ()): ("P", "P;4R"),
+    (II, 3, (1,), 1, (8,), ()): ("P", "P"),
+    (MM, 3, (1,), 1, (8,), ()): ("P", "P"),
+    (II, 3, (1,), 1, (8, 8), (0,)): ("P", "PX"),
+    (II, 3, (1,), 1, (8, 8), (2,)): ("PA", "PA"),
+    (MM, 3, (1,), 1, (8, 8), (2,)): ("PA", "PA"),
+    (II, 3, (1,), 2, (8,), ()): ("P", "P;R"),
+    (MM, 3, (1,), 2, (8,), ()): ("P", "P;R"),
+    (II, 5, (1,), 1, (8, 8, 8, 8), ()): ("CMYK", "CMYK"),
+    (MM, 5, (1,), 1, (8, 8, 8, 8), ()): ("CMYK", "CMYK"),
+    (II, 5, (1,), 1, (8, 8, 8, 8, 8), (0,)): ("CMYK", "CMYKX"),
+    (MM, 5, (1,), 1, (8, 8, 8, 8, 8), (0,)): ("CMYK", "CMYKX"),
+    (II, 5, (1,), 1, (8, 8, 8, 8, 8, 8), (0, 0)): ("CMYK", "CMYKXX"),
+    (MM, 5, (1,), 1, (8, 8, 8, 8, 8, 8), (0, 0)): ("CMYK", "CMYKXX"),
+    (II, 5, (1,), 1, (16, 16, 16, 16), ()): ("CMYK", "CMYK;16L"),
+    (MM, 5, (1,), 1, (16, 16, 16, 16), ()): ("CMYK", "CMYK;16B"),
+    (II, 6, (1,), 1, (8,), ()): ("L", "L"),
+    (MM, 6, (1,), 1, (8,), ()): ("L", "L"),
+    # JPEG compressed images handled by LibTiff and auto-converted to RGBX
+    # Minimal Baseline TIFF requires YCbCr images to have 3 SamplesPerPixel
+    (II, 6, (1,), 1, (8, 8, 8), ()): ("RGB", "RGBX"),
+    (MM, 6, (1,), 1, (8, 8, 8), ()): ("RGB", "RGBX"),
+    (II, 8, (1,), 1, (8, 8, 8), ()): ("LAB", "LAB"),
+    (MM, 8, (1,), 1, (8, 8, 8), ()): ("LAB", "LAB"),
+}
+
+MAX_SAMPLESPERPIXEL = max(len(key_tp[4]) for key_tp in OPEN_INFO)
+
+PREFIXES = [
+    b"MM\x00\x2A",  # Valid TIFF header with big-endian byte order
+    b"II\x2A\x00",  # Valid TIFF header with little-endian byte order
+    b"MM\x2A\x00",  # Invalid TIFF header, assume big-endian
+    b"II\x00\x2A",  # Invalid TIFF header, assume little-endian
+    b"MM\x00\x2B",  # BigTIFF with big-endian byte order
+    b"II\x2B\x00",  # BigTIFF with little-endian byte order
+]
+
+if not getattr(Image.core, "libtiff_support_custom_tags", True):
+    deprecate("Support for LibTIFF earlier than version 4", 12)
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:4] in PREFIXES
+
+
+def _limit_rational(
+    val: float | Fraction | IFDRational, max_val: int
+) -> tuple[IntegralLike, IntegralLike]:
+    inv = abs(val) > 1
+    n_d = IFDRational(1 / val if inv else val).limit_rational(max_val)
+    return n_d[::-1] if inv else n_d
+
+
+def _limit_signed_rational(
+    val: IFDRational, max_val: int, min_val: int
+) -> tuple[IntegralLike, IntegralLike]:
+    frac = Fraction(val)
+    n_d: tuple[IntegralLike, IntegralLike] = frac.numerator, frac.denominator
+
+    if min(float(i) for i in n_d) < min_val:
+        n_d = _limit_rational(val, abs(min_val))
+
+    n_d_float = tuple(float(i) for i in n_d)
+    if max(n_d_float) > max_val:
+        n_d = _limit_rational(n_d_float[0] / n_d_float[1], max_val)
+
+    return n_d
+
+
+##
+# Wrapper for TIFF IFDs.
+
+_load_dispatch = {}
+_write_dispatch = {}
+
+
+def _delegate(op: str) -> Any:
+    def delegate(
+        self: IFDRational, *args: tuple[float, ...]
+    ) -> bool | float | Fraction:
+        return getattr(self._val, op)(*args)
+
+    return delegate
+
+
+class IFDRational(Rational):
+    """Implements a rational class where 0/0 is a legal value to match
+    the in the wild use of exif rationals.
+
+    e.g., DigitalZoomRatio - 0.00/0.00  indicates that no digital zoom was used
+    """
+
+    """ If the denominator is 0, store this as a float('nan'), otherwise store
+    as a fractions.Fraction(). Delegate as appropriate
+
+    """
+
+    __slots__ = ("_numerator", "_denominator", "_val")
+
+    def __init__(
+        self, value: float | Fraction | IFDRational, denominator: int = 1
+    ) -> None:
+        """
+        :param value: either an integer numerator, a
+        float/rational/other number, or an IFDRational
+        :param denominator: Optional integer denominator
+        """
+        self._val: Fraction | float
+        if isinstance(value, IFDRational):
+            self._numerator = value.numerator
+            self._denominator = value.denominator
+            self._val = value._val
+            return
+
+        if isinstance(value, Fraction):
+            self._numerator = value.numerator
+            self._denominator = value.denominator
+        else:
+            if TYPE_CHECKING:
+                self._numerator = cast(IntegralLike, value)
+            else:
+                self._numerator = value
+            self._denominator = denominator
+
+        if denominator == 0:
+            self._val = float("nan")
+        elif denominator == 1:
+            self._val = Fraction(value)
+        elif int(value) == value:
+            self._val = Fraction(int(value), denominator)
+        else:
+            self._val = Fraction(value / denominator)
+
+    @property
+    def numerator(self) -> IntegralLike:
+        return self._numerator
+
+    @property
+    def denominator(self) -> int:
+        return self._denominator
+
+    def limit_rational(self, max_denominator: int) -> tuple[IntegralLike, int]:
+        """
+
+        :param max_denominator: Integer, the maximum denominator value
+        :returns: Tuple of (numerator, denominator)
+        """
+
+        if self.denominator == 0:
+            return self.numerator, self.denominator
+
+        assert isinstance(self._val, Fraction)
+        f = self._val.limit_denominator(max_denominator)
+        return f.numerator, f.denominator
+
+    def __repr__(self) -> str:
+        return str(float(self._val))
+
+    def __hash__(self) -> int:
+        return self._val.__hash__()
+
+    def __eq__(self, other: object) -> bool:
+        val = self._val
+        if isinstance(other, IFDRational):
+            other = other._val
+        if isinstance(other, float):
+            val = float(val)
+        return val == other
+
+    def __getstate__(self) -> list[float | Fraction | IntegralLike]:
+        return [self._val, self._numerator, self._denominator]
+
+    def __setstate__(self, state: list[float | Fraction | IntegralLike]) -> None:
+        IFDRational.__init__(self, 0)
+        _val, _numerator, _denominator = state
+        assert isinstance(_val, (float, Fraction))
+        self._val = _val
+        if TYPE_CHECKING:
+            self._numerator = cast(IntegralLike, _numerator)
+        else:
+            self._numerator = _numerator
+        assert isinstance(_denominator, int)
+        self._denominator = _denominator
+
+    """ a = ['add','radd', 'sub', 'rsub', 'mul', 'rmul',
+             'truediv', 'rtruediv', 'floordiv', 'rfloordiv',
+             'mod','rmod', 'pow','rpow', 'pos', 'neg',
+             'abs', 'trunc', 'lt', 'gt', 'le', 'ge', 'bool',
+             'ceil', 'floor', 'round']
+        print("\n".join("__%s__ = _delegate('__%s__')" % (s,s) for s in a))
+        """
+
+    __add__ = _delegate("__add__")
+    __radd__ = _delegate("__radd__")
+    __sub__ = _delegate("__sub__")
+    __rsub__ = _delegate("__rsub__")
+    __mul__ = _delegate("__mul__")
+    __rmul__ = _delegate("__rmul__")
+    __truediv__ = _delegate("__truediv__")
+    __rtruediv__ = _delegate("__rtruediv__")
+    __floordiv__ = _delegate("__floordiv__")
+    __rfloordiv__ = _delegate("__rfloordiv__")
+    __mod__ = _delegate("__mod__")
+    __rmod__ = _delegate("__rmod__")
+    __pow__ = _delegate("__pow__")
+    __rpow__ = _delegate("__rpow__")
+    __pos__ = _delegate("__pos__")
+    __neg__ = _delegate("__neg__")
+    __abs__ = _delegate("__abs__")
+    __trunc__ = _delegate("__trunc__")
+    __lt__ = _delegate("__lt__")
+    __gt__ = _delegate("__gt__")
+    __le__ = _delegate("__le__")
+    __ge__ = _delegate("__ge__")
+    __bool__ = _delegate("__bool__")
+    __ceil__ = _delegate("__ceil__")
+    __floor__ = _delegate("__floor__")
+    __round__ = _delegate("__round__")
+    # Python >= 3.11
+    if hasattr(Fraction, "__int__"):
+        __int__ = _delegate("__int__")
+
+
+_LoaderFunc = Callable[["ImageFileDirectory_v2", bytes, bool], Any]
+
+
+def _register_loader(idx: int, size: int) -> Callable[[_LoaderFunc], _LoaderFunc]:
+    def decorator(func: _LoaderFunc) -> _LoaderFunc:
+        from .TiffTags import TYPES
+
+        if func.__name__.startswith("load_"):
+            TYPES[idx] = func.__name__[5:].replace("_", " ")
+        _load_dispatch[idx] = size, func  # noqa: F821
+        return func
+
+    return decorator
+
+
+def _register_writer(idx: int) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
+    def decorator(func: Callable[..., Any]) -> Callable[..., Any]:
+        _write_dispatch[idx] = func  # noqa: F821
+        return func
+
+    return decorator
+
+
+def _register_basic(idx_fmt_name: tuple[int, str, str]) -> None:
+    from .TiffTags import TYPES
+
+    idx, fmt, name = idx_fmt_name
+    TYPES[idx] = name
+    size = struct.calcsize(f"={fmt}")
+
+    def basic_handler(
+        self: ImageFileDirectory_v2, data: bytes, legacy_api: bool = True
+    ) -> tuple[Any, ...]:
+        return self._unpack(f"{len(data) // size}{fmt}", data)
+
+    _load_dispatch[idx] = size, basic_handler  # noqa: F821
+    _write_dispatch[idx] = lambda self, *values: (  # noqa: F821
+        b"".join(self._pack(fmt, value) for value in values)
+    )
+
+
+if TYPE_CHECKING:
+    _IFDv2Base = MutableMapping[int, Any]
+else:
+    _IFDv2Base = MutableMapping
+
+
+class ImageFileDirectory_v2(_IFDv2Base):
+    """This class represents a TIFF tag directory.  To speed things up, we
+    don't decode tags unless they're asked for.
+
+    Exposes a dictionary interface of the tags in the directory::
+
+        ifd = ImageFileDirectory_v2()
+        ifd[key] = 'Some Data'
+        ifd.tagtype[key] = TiffTags.ASCII
+        print(ifd[key])
+        'Some Data'
+
+    Individual values are returned as the strings or numbers, sequences are
+    returned as tuples of the values.
+
+    The tiff metadata type of each item is stored in a dictionary of
+    tag types in
+    :attr:`~PIL.TiffImagePlugin.ImageFileDirectory_v2.tagtype`. The types
+    are read from a tiff file, guessed from the type added, or added
+    manually.
+
+    Data Structures:
+
+        * ``self.tagtype = {}``
+
+          * Key: numerical TIFF tag number
+          * Value: integer corresponding to the data type from
+            :py:data:`.TiffTags.TYPES`
+
+          .. versionadded:: 3.0.0
+
+    'Internal' data structures:
+
+        * ``self._tags_v2 = {}``
+
+          * Key: numerical TIFF tag number
+          * Value: decoded data, as tuple for multiple values
+
+        * ``self._tagdata = {}``
+
+          * Key: numerical TIFF tag number
+          * Value: undecoded byte string from file
+
+        * ``self._tags_v1 = {}``
+
+          * Key: numerical TIFF tag number
+          * Value: decoded data in the v1 format
+
+    Tags will be found in the private attributes ``self._tagdata``, and in
+    ``self._tags_v2`` once decoded.
+
+    ``self.legacy_api`` is a value for internal use, and shouldn't be changed
+    from outside code. In cooperation with
+    :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v1`, if ``legacy_api``
+    is true, then decoded tags will be populated into both ``_tags_v1`` and
+    ``_tags_v2``. ``_tags_v2`` will be used if this IFD is used in the TIFF
+    save routine. Tags should be read from ``_tags_v1`` if
+    ``legacy_api == true``.
+
+    """
+
+    _load_dispatch: dict[int, tuple[int, _LoaderFunc]] = {}
+    _write_dispatch: dict[int, Callable[..., Any]] = {}
+
+    def __init__(
+        self,
+        ifh: bytes = b"II\x2A\x00\x00\x00\x00\x00",
+        prefix: bytes | None = None,
+        group: int | None = None,
+    ) -> None:
+        """Initialize an ImageFileDirectory.
+
+        To construct an ImageFileDirectory from a real file, pass the 8-byte
+        magic header to the constructor.  To only set the endianness, pass it
+        as the 'prefix' keyword argument.
+
+        :param ifh: One of the accepted magic headers (cf. PREFIXES); also sets
+              endianness.
+        :param prefix: Override the endianness of the file.
+        """
+        if not _accept(ifh):
+            msg = f"not a TIFF file (header {repr(ifh)} not valid)"
+            raise SyntaxError(msg)
+        self._prefix = prefix if prefix is not None else ifh[:2]
+        if self._prefix == MM:
+            self._endian = ">"
+        elif self._prefix == II:
+            self._endian = "<"
+        else:
+            msg = "not a TIFF IFD"
+            raise SyntaxError(msg)
+        self._bigtiff = ifh[2] == 43
+        self.group = group
+        self.tagtype: dict[int, int] = {}
+        """ Dictionary of tag types """
+        self.reset()
+        self.next = (
+            self._unpack("Q", ifh[8:])[0]
+            if self._bigtiff
+            else self._unpack("L", ifh[4:])[0]
+        )
+        self._legacy_api = False
+
+    prefix = property(lambda self: self._prefix)
+    offset = property(lambda self: self._offset)
+
+    @property
+    def legacy_api(self) -> bool:
+        return self._legacy_api
+
+    @legacy_api.setter
+    def legacy_api(self, value: bool) -> NoReturn:
+        msg = "Not allowing setting of legacy api"
+        raise Exception(msg)
+
+    def reset(self) -> None:
+        self._tags_v1: dict[int, Any] = {}  # will remain empty if legacy_api is false
+        self._tags_v2: dict[int, Any] = {}  # main tag storage
+        self._tagdata: dict[int, bytes] = {}
+        self.tagtype = {}  # added 2008-06-05 by Florian Hoech
+        self._next = None
+        self._offset: int | None = None
+
+    def __str__(self) -> str:
+        return str(dict(self))
+
+    def named(self) -> dict[str, Any]:
+        """
+        :returns: dict of name|key: value
+
+        Returns the complete tag dictionary, with named tags where possible.
+        """
+        return {
+            TiffTags.lookup(code, self.group).name: value
+            for code, value in self.items()
+        }
+
+    def __len__(self) -> int:
+        return len(set(self._tagdata) | set(self._tags_v2))
+
+    def __getitem__(self, tag: int) -> Any:
+        if tag not in self._tags_v2:  # unpack on the fly
+            data = self._tagdata[tag]
+            typ = self.tagtype[tag]
+            size, handler = self._load_dispatch[typ]
+            self[tag] = handler(self, data, self.legacy_api)  # check type
+        val = self._tags_v2[tag]
+        if self.legacy_api and not isinstance(val, (tuple, bytes)):
+            val = (val,)
+        return val
+
+    def __contains__(self, tag: object) -> bool:
+        return tag in self._tags_v2 or tag in self._tagdata
+
+    def __setitem__(self, tag: int, value: Any) -> None:
+        self._setitem(tag, value, self.legacy_api)
+
+    def _setitem(self, tag: int, value: Any, legacy_api: bool) -> None:
+        basetypes = (Number, bytes, str)
+
+        info = TiffTags.lookup(tag, self.group)
+        values = [value] if isinstance(value, basetypes) else value
+
+        if tag not in self.tagtype:
+            if info.type:
+                self.tagtype[tag] = info.type
+            else:
+                self.tagtype[tag] = TiffTags.UNDEFINED
+                if all(isinstance(v, IFDRational) for v in values):
+                    for v in values:
+                        assert isinstance(v, IFDRational)
+                        if v < 0:
+                            self.tagtype[tag] = TiffTags.SIGNED_RATIONAL
+                            break
+                    else:
+                        self.tagtype[tag] = TiffTags.RATIONAL
+                elif all(isinstance(v, int) for v in values):
+                    short = True
+                    signed_short = True
+                    long = True
+                    for v in values:
+                        assert isinstance(v, int)
+                        if short and not (0 <= v < 2**16):
+                            short = False
+                        if signed_short and not (-(2**15) < v < 2**15):
+                            signed_short = False
+                        if long and v < 0:
+                            long = False
+                    if short:
+                        self.tagtype[tag] = TiffTags.SHORT
+                    elif signed_short:
+                        self.tagtype[tag] = TiffTags.SIGNED_SHORT
+                    elif long:
+                        self.tagtype[tag] = TiffTags.LONG
+                    else:
+                        self.tagtype[tag] = TiffTags.SIGNED_LONG
+                elif all(isinstance(v, float) for v in values):
+                    self.tagtype[tag] = TiffTags.DOUBLE
+                elif all(isinstance(v, str) for v in values):
+                    self.tagtype[tag] = TiffTags.ASCII
+                elif all(isinstance(v, bytes) for v in values):
+                    self.tagtype[tag] = TiffTags.BYTE
+
+        if self.tagtype[tag] == TiffTags.UNDEFINED:
+            values = [
+                v.encode("ascii", "replace") if isinstance(v, str) else v
+                for v in values
+            ]
+        elif self.tagtype[tag] == TiffTags.RATIONAL:
+            values = [float(v) if isinstance(v, int) else v for v in values]
+
+        is_ifd = self.tagtype[tag] == TiffTags.LONG and isinstance(values, dict)
+        if not is_ifd:
+            values = tuple(
+                info.cvt_enum(value) if isinstance(value, str) else value
+                for value in values
+            )
+
+        dest = self._tags_v1 if legacy_api else self._tags_v2
+
+        # Three branches:
+        # Spec'd length == 1, Actual length 1, store as element
+        # Spec'd length == 1, Actual > 1, Warn and truncate. Formerly barfed.
+        # No Spec, Actual length 1, Formerly (<4.2) returned a 1 element tuple.
+        # Don't mess with the legacy api, since it's frozen.
+        if not is_ifd and (
+            (info.length == 1)
+            or self.tagtype[tag] == TiffTags.BYTE
+            or (info.length is None and len(values) == 1 and not legacy_api)
+        ):
+            # Don't mess with the legacy api, since it's frozen.
+            if legacy_api and self.tagtype[tag] in [
+                TiffTags.RATIONAL,
+                TiffTags.SIGNED_RATIONAL,
+            ]:  # rationals
+                values = (values,)
+            try:
+                (dest[tag],) = values
+            except ValueError:
+                # We've got a builtin tag with 1 expected entry
+                warnings.warn(
+                    f"Metadata Warning, tag {tag} had too many entries: "
+                    f"{len(values)}, expected 1"
+                )
+                dest[tag] = values[0]
+
+        else:
+            # Spec'd length > 1 or undefined
+            # Unspec'd, and length > 1
+            dest[tag] = values
+
+    def __delitem__(self, tag: int) -> None:
+        self._tags_v2.pop(tag, None)
+        self._tags_v1.pop(tag, None)
+        self._tagdata.pop(tag, None)
+
+    def __iter__(self) -> Iterator[int]:
+        return iter(set(self._tagdata) | set(self._tags_v2))
+
+    def _unpack(self, fmt: str, data: bytes) -> tuple[Any, ...]:
+        return struct.unpack(self._endian + fmt, data)
+
+    def _pack(self, fmt: str, *values: Any) -> bytes:
+        return struct.pack(self._endian + fmt, *values)
+
+    list(
+        map(
+            _register_basic,
+            [
+                (TiffTags.SHORT, "H", "short"),
+                (TiffTags.LONG, "L", "long"),
+                (TiffTags.SIGNED_BYTE, "b", "signed byte"),
+                (TiffTags.SIGNED_SHORT, "h", "signed short"),
+                (TiffTags.SIGNED_LONG, "l", "signed long"),
+                (TiffTags.FLOAT, "f", "float"),
+                (TiffTags.DOUBLE, "d", "double"),
+                (TiffTags.IFD, "L", "long"),
+                (TiffTags.LONG8, "Q", "long8"),
+            ],
+        )
+    )
+
+    @_register_loader(1, 1)  # Basic type, except for the legacy API.
+    def load_byte(self, data: bytes, legacy_api: bool = True) -> bytes:
+        return data
+
+    @_register_writer(1)  # Basic type, except for the legacy API.
+    def write_byte(self, data: bytes | int | IFDRational) -> bytes:
+        if isinstance(data, IFDRational):
+            data = int(data)
+        if isinstance(data, int):
+            data = bytes((data,))
+        return data
+
+    @_register_loader(2, 1)
+    def load_string(self, data: bytes, legacy_api: bool = True) -> str:
+        if data.endswith(b"\0"):
+            data = data[:-1]
+        return data.decode("latin-1", "replace")
+
+    @_register_writer(2)
+    def write_string(self, value: str | bytes | int) -> bytes:
+        # remerge of https://github.com/python-pillow/Pillow/pull/1416
+        if isinstance(value, int):
+            value = str(value)
+        if not isinstance(value, bytes):
+            value = value.encode("ascii", "replace")
+        return value + b"\0"
+
+    @_register_loader(5, 8)
+    def load_rational(
+        self, data: bytes, legacy_api: bool = True
+    ) -> tuple[tuple[int, int] | IFDRational, ...]:
+        vals = self._unpack(f"{len(data) // 4}L", data)
+
+        def combine(a: int, b: int) -> tuple[int, int] | IFDRational:
+            return (a, b) if legacy_api else IFDRational(a, b)
+
+        return tuple(combine(num, denom) for num, denom in zip(vals[::2], vals[1::2]))
+
+    @_register_writer(5)
+    def write_rational(self, *values: IFDRational) -> bytes:
+        return b"".join(
+            self._pack("2L", *_limit_rational(frac, 2**32 - 1)) for frac in values
+        )
+
+    @_register_loader(7, 1)
+    def load_undefined(self, data: bytes, legacy_api: bool = True) -> bytes:
+        return data
+
+    @_register_writer(7)
+    def write_undefined(self, value: bytes | int | IFDRational) -> bytes:
+        if isinstance(value, IFDRational):
+            value = int(value)
+        if isinstance(value, int):
+            value = str(value).encode("ascii", "replace")
+        return value
+
+    @_register_loader(10, 8)
+    def load_signed_rational(
+        self, data: bytes, legacy_api: bool = True
+    ) -> tuple[tuple[int, int] | IFDRational, ...]:
+        vals = self._unpack(f"{len(data) // 4}l", data)
+
+        def combine(a: int, b: int) -> tuple[int, int] | IFDRational:
+            return (a, b) if legacy_api else IFDRational(a, b)
+
+        return tuple(combine(num, denom) for num, denom in zip(vals[::2], vals[1::2]))
+
+    @_register_writer(10)
+    def write_signed_rational(self, *values: IFDRational) -> bytes:
+        return b"".join(
+            self._pack("2l", *_limit_signed_rational(frac, 2**31 - 1, -(2**31)))
+            for frac in values
+        )
+
+    def _ensure_read(self, fp: IO[bytes], size: int) -> bytes:
+        ret = fp.read(size)
+        if len(ret) != size:
+            msg = (
+                "Corrupt EXIF data.  "
+                f"Expecting to read {size} bytes but only got {len(ret)}. "
+            )
+            raise OSError(msg)
+        return ret
+
+    def load(self, fp: IO[bytes]) -> None:
+        self.reset()
+        self._offset = fp.tell()
+
+        try:
+            tag_count = (
+                self._unpack("Q", self._ensure_read(fp, 8))
+                if self._bigtiff
+                else self._unpack("H", self._ensure_read(fp, 2))
+            )[0]
+            for i in range(tag_count):
+                tag, typ, count, data = (
+                    self._unpack("HHQ8s", self._ensure_read(fp, 20))
+                    if self._bigtiff
+                    else self._unpack("HHL4s", self._ensure_read(fp, 12))
+                )
+
+                tagname = TiffTags.lookup(tag, self.group).name
+                typname = TYPES.get(typ, "unknown")
+                msg = f"tag: {tagname} ({tag}) - type: {typname} ({typ})"
+
+                try:
+                    unit_size, handler = self._load_dispatch[typ]
+                except KeyError:
+                    logger.debug("%s - unsupported type %s", msg, typ)
+                    continue  # ignore unsupported type
+                size = count * unit_size
+                if size > (8 if self._bigtiff else 4):
+                    here = fp.tell()
+                    (offset,) = self._unpack("Q" if self._bigtiff else "L", data)
+                    msg += f" Tag Location: {here} - Data Location: {offset}"
+                    fp.seek(offset)
+                    data = ImageFile._safe_read(fp, size)
+                    fp.seek(here)
+                else:
+                    data = data[:size]
+
+                if len(data) != size:
+                    warnings.warn(
+                        "Possibly corrupt EXIF data.  "
+                        f"Expecting to read {size} bytes but only got {len(data)}."
+                        f" Skipping tag {tag}"
+                    )
+                    logger.debug(msg)
+                    continue
+
+                if not data:
+                    logger.debug(msg)
+                    continue
+
+                self._tagdata[tag] = data
+                self.tagtype[tag] = typ
+
+                msg += " - value: "
+                msg += f"<table: {size} bytes>" if size > 32 else repr(data)
+
+                logger.debug(msg)
+
+            (self.next,) = (
+                self._unpack("Q", self._ensure_read(fp, 8))
+                if self._bigtiff
+                else self._unpack("L", self._ensure_read(fp, 4))
+            )
+        except OSError as msg:
+            warnings.warn(str(msg))
+            return
+
+    def _get_ifh(self):
+        ifh = self._prefix + self._pack("H", 43 if self._bigtiff else 42)
+        if self._bigtiff:
+            ifh += self._pack("HH", 8, 0)
+        ifh += self._pack("Q", 16) if self._bigtiff else self._pack("L", 8)
+
+        return ifh
+
+    def tobytes(self, offset: int = 0) -> bytes:
+        # FIXME What about tagdata?
+        result = self._pack("Q" if self._bigtiff else "H", len(self._tags_v2))
+
+        entries: list[tuple[int, int, int, bytes, bytes]] = []
+        offset += len(result) + len(self._tags_v2) * (20 if self._bigtiff else 12) + 4
+        stripoffsets = None
+
+        # pass 1: convert tags to binary format
+        # always write tags in ascending order
+        fmt = "Q" if self._bigtiff else "L"
+        fmt_size = 8 if self._bigtiff else 4
+        for tag, value in sorted(self._tags_v2.items()):
+            if tag == STRIPOFFSETS:
+                stripoffsets = len(entries)
+            typ = self.tagtype[tag]
+            logger.debug("Tag %s, Type: %s, Value: %s", tag, typ, repr(value))
+            is_ifd = typ == TiffTags.LONG and isinstance(value, dict)
+            if is_ifd:
+                ifd = ImageFileDirectory_v2(self._get_ifh(), group=tag)
+                values = self._tags_v2[tag]
+                for ifd_tag, ifd_value in values.items():
+                    ifd[ifd_tag] = ifd_value
+                data = ifd.tobytes(offset)
+            else:
+                values = value if isinstance(value, tuple) else (value,)
+                data = self._write_dispatch[typ](self, *values)
+
+            tagname = TiffTags.lookup(tag, self.group).name
+            typname = "ifd" if is_ifd else TYPES.get(typ, "unknown")
+            msg = f"save: {tagname} ({tag}) - type: {typname} ({typ}) - value: "
+            msg += f"<table: {len(data)} bytes>" if len(data) >= 16 else str(values)
+            logger.debug(msg)
+
+            # count is sum of lengths for string and arbitrary data
+            if is_ifd:
+                count = 1
+            elif typ in [TiffTags.BYTE, TiffTags.ASCII, TiffTags.UNDEFINED]:
+                count = len(data)
+            else:
+                count = len(values)
+            # figure out if data fits into the entry
+            if len(data) <= fmt_size:
+                entries.append((tag, typ, count, data.ljust(fmt_size, b"\0"), b""))
+            else:
+                entries.append((tag, typ, count, self._pack(fmt, offset), data))
+                offset += (len(data) + 1) // 2 * 2  # pad to word
+
+        # update strip offset data to point beyond auxiliary data
+        if stripoffsets is not None:
+            tag, typ, count, value, data = entries[stripoffsets]
+            if data:
+                size, handler = self._load_dispatch[typ]
+                values = [val + offset for val in handler(self, data, self.legacy_api)]
+                data = self._write_dispatch[typ](self, *values)
+            else:
+                value = self._pack(fmt, self._unpack(fmt, value)[0] + offset)
+            entries[stripoffsets] = tag, typ, count, value, data
+
+        # pass 2: write entries to file
+        for tag, typ, count, value, data in entries:
+            logger.debug("%s %s %s %s %s", tag, typ, count, repr(value), repr(data))
+            result += self._pack(
+                "HHQ8s" if self._bigtiff else "HHL4s", tag, typ, count, value
+            )
+
+        # -- overwrite here for multi-page --
+        result += b"\0\0\0\0"  # end of entries
+
+        # pass 3: write auxiliary data to file
+        for tag, typ, count, value, data in entries:
+            result += data
+            if len(data) & 1:
+                result += b"\0"
+
+        return result
+
+    def save(self, fp: IO[bytes]) -> int:
+        if fp.tell() == 0:  # skip TIFF header on subsequent pages
+            fp.write(self._get_ifh())
+
+        offset = fp.tell()
+        result = self.tobytes(offset)
+        fp.write(result)
+        return offset + len(result)
+
+
+ImageFileDirectory_v2._load_dispatch = _load_dispatch
+ImageFileDirectory_v2._write_dispatch = _write_dispatch
+for idx, name in TYPES.items():
+    name = name.replace(" ", "_")
+    setattr(ImageFileDirectory_v2, f"load_{name}", _load_dispatch[idx][1])
+    setattr(ImageFileDirectory_v2, f"write_{name}", _write_dispatch[idx])
+del _load_dispatch, _write_dispatch, idx, name
+
+
+# Legacy ImageFileDirectory support.
+class ImageFileDirectory_v1(ImageFileDirectory_v2):
+    """This class represents the **legacy** interface to a TIFF tag directory.
+
+    Exposes a dictionary interface of the tags in the directory::
+
+        ifd = ImageFileDirectory_v1()
+        ifd[key] = 'Some Data'
+        ifd.tagtype[key] = TiffTags.ASCII
+        print(ifd[key])
+        ('Some Data',)
+
+    Also contains a dictionary of tag types as read from the tiff image file,
+    :attr:`~PIL.TiffImagePlugin.ImageFileDirectory_v1.tagtype`.
+
+    Values are returned as a tuple.
+
+    ..  deprecated:: 3.0.0
+    """
+
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        super().__init__(*args, **kwargs)
+        self._legacy_api = True
+
+    tags = property(lambda self: self._tags_v1)
+    tagdata = property(lambda self: self._tagdata)
+
+    # defined in ImageFileDirectory_v2
+    tagtype: dict[int, int]
+    """Dictionary of tag types"""
+
+    @classmethod
+    def from_v2(cls, original: ImageFileDirectory_v2) -> ImageFileDirectory_v1:
+        """Returns an
+        :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v1`
+        instance with the same data as is contained in the original
+        :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v2`
+        instance.
+
+        :returns: :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v1`
+
+        """
+
+        ifd = cls(prefix=original.prefix)
+        ifd._tagdata = original._tagdata
+        ifd.tagtype = original.tagtype
+        ifd.next = original.next  # an indicator for multipage tiffs
+        return ifd
+
+    def to_v2(self) -> ImageFileDirectory_v2:
+        """Returns an
+        :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v2`
+        instance with the same data as is contained in the original
+        :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v1`
+        instance.
+
+        :returns: :py:class:`~PIL.TiffImagePlugin.ImageFileDirectory_v2`
+
+        """
+
+        ifd = ImageFileDirectory_v2(prefix=self.prefix)
+        ifd._tagdata = dict(self._tagdata)
+        ifd.tagtype = dict(self.tagtype)
+        ifd._tags_v2 = dict(self._tags_v2)
+        return ifd
+
+    def __contains__(self, tag: object) -> bool:
+        return tag in self._tags_v1 or tag in self._tagdata
+
+    def __len__(self) -> int:
+        return len(set(self._tagdata) | set(self._tags_v1))
+
+    def __iter__(self) -> Iterator[int]:
+        return iter(set(self._tagdata) | set(self._tags_v1))
+
+    def __setitem__(self, tag: int, value: Any) -> None:
+        for legacy_api in (False, True):
+            self._setitem(tag, value, legacy_api)
+
+    def __getitem__(self, tag: int) -> Any:
+        if tag not in self._tags_v1:  # unpack on the fly
+            data = self._tagdata[tag]
+            typ = self.tagtype[tag]
+            size, handler = self._load_dispatch[typ]
+            for legacy in (False, True):
+                self._setitem(tag, handler(self, data, legacy), legacy)
+        val = self._tags_v1[tag]
+        if not isinstance(val, (tuple, bytes)):
+            val = (val,)
+        return val
+
+
+# undone -- switch this pointer
+ImageFileDirectory = ImageFileDirectory_v1
+
+
+##
+# Image plugin for TIFF files.
+
+
+class TiffImageFile(ImageFile.ImageFile):
+    format = "TIFF"
+    format_description = "Adobe TIFF"
+    _close_exclusive_fp_after_loading = False
+
+    def __init__(
+        self,
+        fp: StrOrBytesPath | IO[bytes],
+        filename: str | bytes | None = None,
+    ) -> None:
+        self.tag_v2: ImageFileDirectory_v2
+        """ Image file directory (tag dictionary) """
+
+        self.tag: ImageFileDirectory_v1
+        """ Legacy tag entries """
+
+        super().__init__(fp, filename)
+
+    def _open(self) -> None:
+        """Open the first image in a TIFF file"""
+
+        # Header
+        ifh = self.fp.read(8)
+        if ifh[2] == 43:
+            ifh += self.fp.read(8)
+
+        self.tag_v2 = ImageFileDirectory_v2(ifh)
+
+        # setup frame pointers
+        self.__first = self.__next = self.tag_v2.next
+        self.__frame = -1
+        self._fp = self.fp
+        self._frame_pos: list[int] = []
+        self._n_frames: int | None = None
+
+        logger.debug("*** TiffImageFile._open ***")
+        logger.debug("- __first: %s", self.__first)
+        logger.debug("- ifh: %s", repr(ifh))  # Use repr to avoid str(bytes)
+
+        # and load the first frame
+        self._seek(0)
+
+    @property
+    def n_frames(self) -> int:
+        current_n_frames = self._n_frames
+        if current_n_frames is None:
+            current = self.tell()
+            self._seek(len(self._frame_pos))
+            while self._n_frames is None:
+                self._seek(self.tell() + 1)
+            self.seek(current)
+        assert self._n_frames is not None
+        return self._n_frames
+
+    def seek(self, frame: int) -> None:
+        """Select a given frame as current image"""
+        if not self._seek_check(frame):
+            return
+        self._seek(frame)
+        if self._im is not None and (
+            self.im.size != self._tile_size or self.im.mode != self.mode
+        ):
+            # The core image will no longer be used
+            self._im = None
+
+    def _seek(self, frame: int) -> None:
+        self.fp = self._fp
+
+        while len(self._frame_pos) <= frame:
+            if not self.__next:
+                msg = "no more images in TIFF file"
+                raise EOFError(msg)
+            logger.debug(
+                "Seeking to frame %s, on frame %s, __next %s, location: %s",
+                frame,
+                self.__frame,
+                self.__next,
+                self.fp.tell(),
+            )
+            if self.__next >= 2**63:
+                msg = "Unable to seek to frame"
+                raise ValueError(msg)
+            self.fp.seek(self.__next)
+            self._frame_pos.append(self.__next)
+            logger.debug("Loading tags, location: %s", self.fp.tell())
+            self.tag_v2.load(self.fp)
+            if self.tag_v2.next in self._frame_pos:
+                # This IFD has already been processed
+                # Declare this to be the end of the image
+                self.__next = 0
+            else:
+                self.__next = self.tag_v2.next
+            if self.__next == 0:
+                self._n_frames = frame + 1
+            if len(self._frame_pos) == 1:
+                self.is_animated = self.__next != 0
+            self.__frame += 1
+        self.fp.seek(self._frame_pos[frame])
+        self.tag_v2.load(self.fp)
+        if XMP in self.tag_v2:
+            self.info["xmp"] = self.tag_v2[XMP]
+        elif "xmp" in self.info:
+            del self.info["xmp"]
+        self._reload_exif()
+        # fill the legacy tag/ifd entries
+        self.tag = self.ifd = ImageFileDirectory_v1.from_v2(self.tag_v2)
+        self.__frame = frame
+        self._setup()
+
+    def tell(self) -> int:
+        """Return the current frame number"""
+        return self.__frame
+
+    def get_photoshop_blocks(self) -> dict[int, dict[str, bytes]]:
+        """
+        Returns a dictionary of Photoshop "Image Resource Blocks".
+        The keys are the image resource ID. For more information, see
+        https://www.adobe.com/devnet-apps/photoshop/fileformatashtml/#50577409_pgfId-1037727
+
+        :returns: Photoshop "Image Resource Blocks" in a dictionary.
+        """
+        blocks = {}
+        val = self.tag_v2.get(ExifTags.Base.ImageResources)
+        if val:
+            while val[:4] == b"8BIM":
+                id = i16(val[4:6])
+                n = math.ceil((val[6] + 1) / 2) * 2
+                size = i32(val[6 + n : 10 + n])
+                data = val[10 + n : 10 + n + size]
+                blocks[id] = {"data": data}
+
+                val = val[math.ceil((10 + n + size) / 2) * 2 :]
+        return blocks
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self.tile and self.use_load_libtiff:
+            return self._load_libtiff()
+        return super().load()
+
+    def load_prepare(self) -> None:
+        if self._im is None:
+            Image._decompression_bomb_check(self._tile_size)
+            self.im = Image.core.new(self.mode, self._tile_size)
+        ImageFile.ImageFile.load_prepare(self)
+
+    def load_end(self) -> None:
+        # allow closing if we're on the first frame, there's no next
+        # This is the ImageFile.load path only, libtiff specific below.
+        if not self.is_animated:
+            self._close_exclusive_fp_after_loading = True
+
+            # load IFD data from fp before it is closed
+            exif = self.getexif()
+            for key in TiffTags.TAGS_V2_GROUPS:
+                if key not in exif:
+                    continue
+                exif.get_ifd(key)
+
+        ImageOps.exif_transpose(self, in_place=True)
+        if ExifTags.Base.Orientation in self.tag_v2:
+            del self.tag_v2[ExifTags.Base.Orientation]
+
+    def _load_libtiff(self) -> Image.core.PixelAccess | None:
+        """Overload method triggered when we detect a compressed tiff
+        Calls out to libtiff"""
+
+        Image.Image.load(self)
+
+        self.load_prepare()
+
+        if not len(self.tile) == 1:
+            msg = "Not exactly one tile"
+            raise OSError(msg)
+
+        # (self._compression, (extents tuple),
+        #   0, (rawmode, self._compression, fp))
+        extents = self.tile[0][1]
+        args = self.tile[0][3]
+
+        # To be nice on memory footprint, if there's a
+        # file descriptor, use that instead of reading
+        # into a string in python.
+        try:
+            fp = hasattr(self.fp, "fileno") and self.fp.fileno()
+            # flush the file descriptor, prevents error on pypy 2.4+
+            # should also eliminate the need for fp.tell
+            # in _seek
+            if hasattr(self.fp, "flush"):
+                self.fp.flush()
+        except OSError:
+            # io.BytesIO have a fileno, but returns an OSError if
+            # it doesn't use a file descriptor.
+            fp = False
+
+        if fp:
+            assert isinstance(args, tuple)
+            args_list = list(args)
+            args_list[2] = fp
+            args = tuple(args_list)
+
+        decoder = Image._getdecoder(self.mode, "libtiff", args, self.decoderconfig)
+        try:
+            decoder.setimage(self.im, extents)
+        except ValueError as e:
+            msg = "Couldn't set the image"
+            raise OSError(msg) from e
+
+        close_self_fp = self._exclusive_fp and not self.is_animated
+        if hasattr(self.fp, "getvalue"):
+            # We've got a stringio like thing passed in. Yay for all in memory.
+            # The decoder needs the entire file in one shot, so there's not
+            # a lot we can do here other than give it the entire file.
+            # unless we could do something like get the address of the
+            # underlying string for stringio.
+            #
+            # Rearranging for supporting byteio items, since they have a fileno
+            # that returns an OSError if there's no underlying fp. Easier to
+            # deal with here by reordering.
+            logger.debug("have getvalue. just sending in a string from getvalue")
+            n, err = decoder.decode(self.fp.getvalue())
+        elif fp:
+            # we've got a actual file on disk, pass in the fp.
+            logger.debug("have fileno, calling fileno version of the decoder.")
+            if not close_self_fp:
+                self.fp.seek(0)
+            # Save and restore the file position, because libtiff will move it
+            # outside of the Python runtime, and that will confuse
+            # io.BufferedReader and possible others.
+            # NOTE: This must use os.lseek(), and not fp.tell()/fp.seek(),
+            # because the buffer read head already may not equal the actual
+            # file position, and fp.seek() may just adjust it's internal
+            # pointer and not actually seek the OS file handle.
+            pos = os.lseek(fp, 0, os.SEEK_CUR)
+            # 4 bytes, otherwise the trace might error out
+            n, err = decoder.decode(b"fpfp")
+            os.lseek(fp, pos, os.SEEK_SET)
+        else:
+            # we have something else.
+            logger.debug("don't have fileno or getvalue. just reading")
+            self.fp.seek(0)
+            # UNDONE -- so much for that buffer size thing.
+            n, err = decoder.decode(self.fp.read())
+
+        self.tile = []
+        self.readonly = 0
+
+        self.load_end()
+
+        if close_self_fp:
+            self.fp.close()
+            self.fp = None  # might be shared
+
+        if err < 0:
+            raise OSError(err)
+
+        return Image.Image.load(self)
+
+    def _setup(self) -> None:
+        """Setup this image object based on current tags"""
+
+        if 0xBC01 in self.tag_v2:
+            msg = "Windows Media Photo files not yet supported"
+            raise OSError(msg)
+
+        # extract relevant tags
+        self._compression = COMPRESSION_INFO[self.tag_v2.get(COMPRESSION, 1)]
+        self._planar_configuration = self.tag_v2.get(PLANAR_CONFIGURATION, 1)
+
+        # photometric is a required tag, but not everyone is reading
+        # the specification
+        photo = self.tag_v2.get(PHOTOMETRIC_INTERPRETATION, 0)
+
+        # old style jpeg compression images most certainly are YCbCr
+        if self._compression == "tiff_jpeg":
+            photo = 6
+
+        fillorder = self.tag_v2.get(FILLORDER, 1)
+
+        logger.debug("*** Summary ***")
+        logger.debug("- compression: %s", self._compression)
+        logger.debug("- photometric_interpretation: %s", photo)
+        logger.debug("- planar_configuration: %s", self._planar_configuration)
+        logger.debug("- fill_order: %s", fillorder)
+        logger.debug("- YCbCr subsampling: %s", self.tag_v2.get(YCBCRSUBSAMPLING))
+
+        # size
+        try:
+            xsize = self.tag_v2[IMAGEWIDTH]
+            ysize = self.tag_v2[IMAGELENGTH]
+        except KeyError as e:
+            msg = "Missing dimensions"
+            raise TypeError(msg) from e
+        if not isinstance(xsize, int) or not isinstance(ysize, int):
+            msg = "Invalid dimensions"
+            raise ValueError(msg)
+        self._tile_size = xsize, ysize
+        orientation = self.tag_v2.get(ExifTags.Base.Orientation)
+        if orientation in (5, 6, 7, 8):
+            self._size = ysize, xsize
+        else:
+            self._size = xsize, ysize
+
+        logger.debug("- size: %s", self.size)
+
+        sample_format = self.tag_v2.get(SAMPLEFORMAT, (1,))
+        if len(sample_format) > 1 and max(sample_format) == min(sample_format) == 1:
+            # SAMPLEFORMAT is properly per band, so an RGB image will
+            # be (1,1,1).  But, we don't support per band pixel types,
+            # and anything more than one band is a uint8. So, just
+            # take the first element. Revisit this if adding support
+            # for more exotic images.
+            sample_format = (1,)
+
+        bps_tuple = self.tag_v2.get(BITSPERSAMPLE, (1,))
+        extra_tuple = self.tag_v2.get(EXTRASAMPLES, ())
+        if photo in (2, 6, 8):  # RGB, YCbCr, LAB
+            bps_count = 3
+        elif photo == 5:  # CMYK
+            bps_count = 4
+        else:
+            bps_count = 1
+        bps_count += len(extra_tuple)
+        bps_actual_count = len(bps_tuple)
+        samples_per_pixel = self.tag_v2.get(
+            SAMPLESPERPIXEL,
+            3 if self._compression == "tiff_jpeg" and photo in (2, 6) else 1,
+        )
+
+        if samples_per_pixel > MAX_SAMPLESPERPIXEL:
+            # DOS check, samples_per_pixel can be a Long, and we extend the tuple below
+            logger.error(
+                "More samples per pixel than can be decoded: %s", samples_per_pixel
+            )
+            msg = "Invalid value for samples per pixel"
+            raise SyntaxError(msg)
+
+        if samples_per_pixel < bps_actual_count:
+            # If a file has more values in bps_tuple than expected,
+            # remove the excess.
+            bps_tuple = bps_tuple[:samples_per_pixel]
+        elif samples_per_pixel > bps_actual_count and bps_actual_count == 1:
+            # If a file has only one value in bps_tuple, when it should have more,
+            # presume it is the same number of bits for all of the samples.
+            bps_tuple = bps_tuple * samples_per_pixel
+
+        if len(bps_tuple) != samples_per_pixel:
+            msg = "unknown data organization"
+            raise SyntaxError(msg)
+
+        # mode: check photometric interpretation and bits per pixel
+        key = (
+            self.tag_v2.prefix,
+            photo,
+            sample_format,
+            fillorder,
+            bps_tuple,
+            extra_tuple,
+        )
+        logger.debug("format key: %s", key)
+        try:
+            self._mode, rawmode = OPEN_INFO[key]
+        except KeyError as e:
+            logger.debug("- unsupported format")
+            msg = "unknown pixel mode"
+            raise SyntaxError(msg) from e
+
+        logger.debug("- raw mode: %s", rawmode)
+        logger.debug("- pil mode: %s", self.mode)
+
+        self.info["compression"] = self._compression
+
+        xres = self.tag_v2.get(X_RESOLUTION, 1)
+        yres = self.tag_v2.get(Y_RESOLUTION, 1)
+
+        if xres and yres:
+            resunit = self.tag_v2.get(RESOLUTION_UNIT)
+            if resunit == 2:  # dots per inch
+                self.info["dpi"] = (xres, yres)
+            elif resunit == 3:  # dots per centimeter. convert to dpi
+                self.info["dpi"] = (xres * 2.54, yres * 2.54)
+            elif resunit is None:  # used to default to 1, but now 2)
+                self.info["dpi"] = (xres, yres)
+                # For backward compatibility,
+                # we also preserve the old behavior
+                self.info["resolution"] = xres, yres
+            else:  # No absolute unit of measurement
+                self.info["resolution"] = xres, yres
+
+        # build tile descriptors
+        x = y = layer = 0
+        self.tile = []
+        self.use_load_libtiff = READ_LIBTIFF or self._compression != "raw"
+        if self.use_load_libtiff:
+            # Decoder expects entire file as one tile.
+            # There's a buffer size limit in load (64k)
+            # so large g4 images will fail if we use that
+            # function.
+            #
+            # Setup the one tile for the whole image, then
+            # use the _load_libtiff function.
+
+            # libtiff handles the fillmode for us, so 1;IR should
+            # actually be 1;I. Including the R double reverses the
+            # bits, so stripes of the image are reversed.  See
+            # https://github.com/python-pillow/Pillow/issues/279
+            if fillorder == 2:
+                # Replace fillorder with fillorder=1
+                key = key[:3] + (1,) + key[4:]
+                logger.debug("format key: %s", key)
+                # this should always work, since all the
+                # fillorder==2 modes have a corresponding
+                # fillorder=1 mode
+                self._mode, rawmode = OPEN_INFO[key]
+            # YCbCr images with new jpeg compression with pixels in one plane
+            # unpacked straight into RGB values
+            if (
+                photo == 6
+                and self._compression == "jpeg"
+                and self._planar_configuration == 1
+            ):
+                rawmode = "RGB"
+            # libtiff always returns the bytes in native order.
+            # we're expecting image byte order. So, if the rawmode
+            # contains I;16, we need to convert from native to image
+            # byte order.
+            elif rawmode == "I;16":
+                rawmode = "I;16N"
+            elif rawmode.endswith(";16B") or rawmode.endswith(";16L"):
+                rawmode = rawmode[:-1] + "N"
+
+            # Offset in the tile tuple is 0, we go from 0,0 to
+            # w,h, and we only do this once -- eds
+            a = (rawmode, self._compression, False, self.tag_v2.offset)
+            self.tile.append(ImageFile._Tile("libtiff", (0, 0, xsize, ysize), 0, a))
+
+        elif STRIPOFFSETS in self.tag_v2 or TILEOFFSETS in self.tag_v2:
+            # striped image
+            if STRIPOFFSETS in self.tag_v2:
+                offsets = self.tag_v2[STRIPOFFSETS]
+                h = self.tag_v2.get(ROWSPERSTRIP, ysize)
+                w = xsize
+            else:
+                # tiled image
+                offsets = self.tag_v2[TILEOFFSETS]
+                tilewidth = self.tag_v2.get(TILEWIDTH)
+                h = self.tag_v2.get(TILELENGTH)
+                if not isinstance(tilewidth, int) or not isinstance(h, int):
+                    msg = "Invalid tile dimensions"
+                    raise ValueError(msg)
+                w = tilewidth
+
+            for offset in offsets:
+                if x + w > xsize:
+                    stride = w * sum(bps_tuple) / 8  # bytes per line
+                else:
+                    stride = 0
+
+                tile_rawmode = rawmode
+                if self._planar_configuration == 2:
+                    # each band on it's own layer
+                    tile_rawmode = rawmode[layer]
+                    # adjust stride width accordingly
+                    stride /= bps_count
+
+                args = (tile_rawmode, int(stride), 1)
+                self.tile.append(
+                    ImageFile._Tile(
+                        self._compression,
+                        (x, y, min(x + w, xsize), min(y + h, ysize)),
+                        offset,
+                        args,
+                    )
+                )
+                x = x + w
+                if x >= xsize:
+                    x, y = 0, y + h
+                    if y >= ysize:
+                        x = y = 0
+                        layer += 1
+        else:
+            logger.debug("- unsupported data organization")
+            msg = "unknown data organization"
+            raise SyntaxError(msg)
+
+        # Fix up info.
+        if ICCPROFILE in self.tag_v2:
+            self.info["icc_profile"] = self.tag_v2[ICCPROFILE]
+
+        # fixup palette descriptor
+
+        if self.mode in ["P", "PA"]:
+            palette = [o8(b // 256) for b in self.tag_v2[COLORMAP]]
+            self.palette = ImagePalette.raw("RGB;L", b"".join(palette))
+
+
+#
+# --------------------------------------------------------------------
+# Write TIFF files
+
+# little endian is default except for image modes with
+# explicit big endian byte-order
+
+SAVE_INFO = {
+    # mode => rawmode, byteorder, photometrics,
+    #           sampleformat, bitspersample, extra
+    "1": ("1", II, 1, 1, (1,), None),
+    "L": ("L", II, 1, 1, (8,), None),
+    "LA": ("LA", II, 1, 1, (8, 8), 2),
+    "P": ("P", II, 3, 1, (8,), None),
+    "PA": ("PA", II, 3, 1, (8, 8), 2),
+    "I": ("I;32S", II, 1, 2, (32,), None),
+    "I;16": ("I;16", II, 1, 1, (16,), None),
+    "I;16S": ("I;16S", II, 1, 2, (16,), None),
+    "F": ("F;32F", II, 1, 3, (32,), None),
+    "RGB": ("RGB", II, 2, 1, (8, 8, 8), None),
+    "RGBX": ("RGBX", II, 2, 1, (8, 8, 8, 8), 0),
+    "RGBA": ("RGBA", II, 2, 1, (8, 8, 8, 8), 2),
+    "CMYK": ("CMYK", II, 5, 1, (8, 8, 8, 8), None),
+    "YCbCr": ("YCbCr", II, 6, 1, (8, 8, 8), None),
+    "LAB": ("LAB", II, 8, 1, (8, 8, 8), None),
+    "I;32BS": ("I;32BS", MM, 1, 2, (32,), None),
+    "I;16B": ("I;16B", MM, 1, 1, (16,), None),
+    "I;16BS": ("I;16BS", MM, 1, 2, (16,), None),
+    "F;32BF": ("F;32BF", MM, 1, 3, (32,), None),
+}
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    try:
+        rawmode, prefix, photo, format, bits, extra = SAVE_INFO[im.mode]
+    except KeyError as e:
+        msg = f"cannot write mode {im.mode} as TIFF"
+        raise OSError(msg) from e
+
+    encoderinfo = im.encoderinfo
+    encoderconfig = im.encoderconfig
+
+    ifd = ImageFileDirectory_v2(prefix=prefix)
+    if encoderinfo.get("big_tiff"):
+        ifd._bigtiff = True
+
+    try:
+        compression = encoderinfo["compression"]
+    except KeyError:
+        compression = im.info.get("compression")
+        if isinstance(compression, int):
+            # compression value may be from BMP. Ignore it
+            compression = None
+    if compression is None:
+        compression = "raw"
+    elif compression == "tiff_jpeg":
+        # OJPEG is obsolete, so use new-style JPEG compression instead
+        compression = "jpeg"
+    elif compression == "tiff_deflate":
+        compression = "tiff_adobe_deflate"
+
+    libtiff = WRITE_LIBTIFF or compression != "raw"
+
+    # required for color libtiff images
+    ifd[PLANAR_CONFIGURATION] = 1
+
+    ifd[IMAGEWIDTH] = im.size[0]
+    ifd[IMAGELENGTH] = im.size[1]
+
+    # write any arbitrary tags passed in as an ImageFileDirectory
+    if "tiffinfo" in encoderinfo:
+        info = encoderinfo["tiffinfo"]
+    elif "exif" in encoderinfo:
+        info = encoderinfo["exif"]
+        if isinstance(info, bytes):
+            exif = Image.Exif()
+            exif.load(info)
+            info = exif
+    else:
+        info = {}
+    logger.debug("Tiffinfo Keys: %s", list(info))
+    if isinstance(info, ImageFileDirectory_v1):
+        info = info.to_v2()
+    for key in info:
+        if isinstance(info, Image.Exif) and key in TiffTags.TAGS_V2_GROUPS:
+            ifd[key] = info.get_ifd(key)
+        else:
+            ifd[key] = info.get(key)
+        try:
+            ifd.tagtype[key] = info.tagtype[key]
+        except Exception:
+            pass  # might not be an IFD. Might not have populated type
+
+    legacy_ifd = {}
+    if hasattr(im, "tag"):
+        legacy_ifd = im.tag.to_v2()
+
+    supplied_tags = {**legacy_ifd, **getattr(im, "tag_v2", {})}
+    for tag in (
+        # IFD offset that may not be correct in the saved image
+        EXIFIFD,
+        # Determined by the image format and should not be copied from legacy_ifd.
+        SAMPLEFORMAT,
+    ):
+        if tag in supplied_tags:
+            del supplied_tags[tag]
+
+    # additions written by Greg Couch, gregc@cgl.ucsf.edu
+    # inspired by image-sig posting from Kevin Cazabon, kcazabon@home.com
+    if hasattr(im, "tag_v2"):
+        # preserve tags from original TIFF image file
+        for key in (
+            RESOLUTION_UNIT,
+            X_RESOLUTION,
+            Y_RESOLUTION,
+            IPTC_NAA_CHUNK,
+            PHOTOSHOP_CHUNK,
+            XMP,
+        ):
+            if key in im.tag_v2:
+                if key == IPTC_NAA_CHUNK and im.tag_v2.tagtype[key] not in (
+                    TiffTags.BYTE,
+                    TiffTags.UNDEFINED,
+                ):
+                    del supplied_tags[key]
+                else:
+                    ifd[key] = im.tag_v2[key]
+                    ifd.tagtype[key] = im.tag_v2.tagtype[key]
+
+    # preserve ICC profile (should also work when saving other formats
+    # which support profiles as TIFF) -- 2008-06-06 Florian Hoech
+    icc = encoderinfo.get("icc_profile", im.info.get("icc_profile"))
+    if icc:
+        ifd[ICCPROFILE] = icc
+
+    for key, name in [
+        (IMAGEDESCRIPTION, "description"),
+        (X_RESOLUTION, "resolution"),
+        (Y_RESOLUTION, "resolution"),
+        (X_RESOLUTION, "x_resolution"),
+        (Y_RESOLUTION, "y_resolution"),
+        (RESOLUTION_UNIT, "resolution_unit"),
+        (SOFTWARE, "software"),
+        (DATE_TIME, "date_time"),
+        (ARTIST, "artist"),
+        (COPYRIGHT, "copyright"),
+    ]:
+        if name in encoderinfo:
+            ifd[key] = encoderinfo[name]
+
+    dpi = encoderinfo.get("dpi")
+    if dpi:
+        ifd[RESOLUTION_UNIT] = 2
+        ifd[X_RESOLUTION] = dpi[0]
+        ifd[Y_RESOLUTION] = dpi[1]
+
+    if bits != (1,):
+        ifd[BITSPERSAMPLE] = bits
+        if len(bits) != 1:
+            ifd[SAMPLESPERPIXEL] = len(bits)
+    if extra is not None:
+        ifd[EXTRASAMPLES] = extra
+    if format != 1:
+        ifd[SAMPLEFORMAT] = format
+
+    if PHOTOMETRIC_INTERPRETATION not in ifd:
+        ifd[PHOTOMETRIC_INTERPRETATION] = photo
+    elif im.mode in ("1", "L") and ifd[PHOTOMETRIC_INTERPRETATION] == 0:
+        if im.mode == "1":
+            inverted_im = im.copy()
+            px = inverted_im.load()
+            if px is not None:
+                for y in range(inverted_im.height):
+                    for x in range(inverted_im.width):
+                        px[x, y] = 0 if px[x, y] == 255 else 255
+                im = inverted_im
+        else:
+            im = ImageOps.invert(im)
+
+    if im.mode in ["P", "PA"]:
+        lut = im.im.getpalette("RGB", "RGB;L")
+        colormap = []
+        colors = len(lut) // 3
+        for i in range(3):
+            colormap += [v * 256 for v in lut[colors * i : colors * (i + 1)]]
+            colormap += [0] * (256 - colors)
+        ifd[COLORMAP] = colormap
+    # data orientation
+    w, h = ifd[IMAGEWIDTH], ifd[IMAGELENGTH]
+    stride = len(bits) * ((w * bits[0] + 7) // 8)
+    if ROWSPERSTRIP not in ifd:
+        # aim for given strip size (64 KB by default) when using libtiff writer
+        if libtiff:
+            im_strip_size = encoderinfo.get("strip_size", STRIP_SIZE)
+            rows_per_strip = 1 if stride == 0 else min(im_strip_size // stride, h)
+            # JPEG encoder expects multiple of 8 rows
+            if compression == "jpeg":
+                rows_per_strip = min(((rows_per_strip + 7) // 8) * 8, h)
+        else:
+            rows_per_strip = h
+        if rows_per_strip == 0:
+            rows_per_strip = 1
+        ifd[ROWSPERSTRIP] = rows_per_strip
+    strip_byte_counts = 1 if stride == 0 else stride * ifd[ROWSPERSTRIP]
+    strips_per_image = (h + ifd[ROWSPERSTRIP] - 1) // ifd[ROWSPERSTRIP]
+    if strip_byte_counts >= 2**16:
+        ifd.tagtype[STRIPBYTECOUNTS] = TiffTags.LONG
+    ifd[STRIPBYTECOUNTS] = (strip_byte_counts,) * (strips_per_image - 1) + (
+        stride * h - strip_byte_counts * (strips_per_image - 1),
+    )
+    ifd[STRIPOFFSETS] = tuple(
+        range(0, strip_byte_counts * strips_per_image, strip_byte_counts)
+    )  # this is adjusted by IFD writer
+    # no compression by default:
+    ifd[COMPRESSION] = COMPRESSION_INFO_REV.get(compression, 1)
+
+    if im.mode == "YCbCr":
+        for tag, default_value in {
+            YCBCRSUBSAMPLING: (1, 1),
+            REFERENCEBLACKWHITE: (0, 255, 128, 255, 128, 255),
+        }.items():
+            ifd.setdefault(tag, default_value)
+
+    blocklist = [TILEWIDTH, TILELENGTH, TILEOFFSETS, TILEBYTECOUNTS]
+    if libtiff:
+        if "quality" in encoderinfo:
+            quality = encoderinfo["quality"]
+            if not isinstance(quality, int) or quality < 0 or quality > 100:
+                msg = "Invalid quality setting"
+                raise ValueError(msg)
+            if compression != "jpeg":
+                msg = "quality setting only supported for 'jpeg' compression"
+                raise ValueError(msg)
+            ifd[JPEGQUALITY] = quality
+
+        logger.debug("Saving using libtiff encoder")
+        logger.debug("Items: %s", sorted(ifd.items()))
+        _fp = 0
+        if hasattr(fp, "fileno"):
+            try:
+                fp.seek(0)
+                _fp = fp.fileno()
+            except io.UnsupportedOperation:
+                pass
+
+        # optional types for non core tags
+        types = {}
+        # STRIPOFFSETS and STRIPBYTECOUNTS are added by the library
+        # based on the data in the strip.
+        # OSUBFILETYPE is deprecated.
+        # The other tags expect arrays with a certain length (fixed or depending on
+        # BITSPERSAMPLE, etc), passing arrays with a different length will result in
+        # segfaults. Block these tags until we add extra validation.
+        # SUBIFD may also cause a segfault.
+        blocklist += [
+            OSUBFILETYPE,
+            REFERENCEBLACKWHITE,
+            STRIPBYTECOUNTS,
+            STRIPOFFSETS,
+            TRANSFERFUNCTION,
+            SUBIFD,
+        ]
+
+        # bits per sample is a single short in the tiff directory, not a list.
+        atts: dict[int, Any] = {BITSPERSAMPLE: bits[0]}
+        # Merge the ones that we have with (optional) more bits from
+        # the original file, e.g x,y resolution so that we can
+        # save(load('')) == original file.
+        for tag, value in itertools.chain(ifd.items(), supplied_tags.items()):
+            # Libtiff can only process certain core items without adding
+            # them to the custom dictionary.
+            # Custom items are supported for int, float, unicode, string and byte
+            # values. Other types and tuples require a tagtype.
+            if tag not in TiffTags.LIBTIFF_CORE:
+                if not getattr(Image.core, "libtiff_support_custom_tags", False):
+                    continue
+
+                if tag in TiffTags.TAGS_V2_GROUPS:
+                    types[tag] = TiffTags.LONG8
+                elif tag in ifd.tagtype:
+                    types[tag] = ifd.tagtype[tag]
+                elif not (isinstance(value, (int, float, str, bytes))):
+                    continue
+                else:
+                    type = TiffTags.lookup(tag).type
+                    if type:
+                        types[tag] = type
+            if tag not in atts and tag not in blocklist:
+                if isinstance(value, str):
+                    atts[tag] = value.encode("ascii", "replace") + b"\0"
+                elif isinstance(value, IFDRational):
+                    atts[tag] = float(value)
+                else:
+                    atts[tag] = value
+
+        if SAMPLEFORMAT in atts and len(atts[SAMPLEFORMAT]) == 1:
+            atts[SAMPLEFORMAT] = atts[SAMPLEFORMAT][0]
+
+        logger.debug("Converted items: %s", sorted(atts.items()))
+
+        # libtiff always expects the bytes in native order.
+        # we're storing image byte order. So, if the rawmode
+        # contains I;16, we need to convert from native to image
+        # byte order.
+        if im.mode in ("I;16B", "I;16"):
+            rawmode = "I;16N"
+
+        # Pass tags as sorted list so that the tags are set in a fixed order.
+        # This is required by libtiff for some tags. For example, the JPEGQUALITY
+        # pseudo tag requires that the COMPRESS tag was already set.
+        tags = list(atts.items())
+        tags.sort()
+        a = (rawmode, compression, _fp, filename, tags, types)
+        encoder = Image._getencoder(im.mode, "libtiff", a, encoderconfig)
+        encoder.setimage(im.im, (0, 0) + im.size)
+        while True:
+            errcode, data = encoder.encode(ImageFile.MAXBLOCK)[1:]
+            if not _fp:
+                fp.write(data)
+            if errcode:
+                break
+        if errcode < 0:
+            msg = f"encoder error {errcode} when writing image file"
+            raise OSError(msg)
+
+    else:
+        for tag in blocklist:
+            del ifd[tag]
+        offset = ifd.save(fp)
+
+        ImageFile._save(
+            im,
+            fp,
+            [ImageFile._Tile("raw", (0, 0) + im.size, offset, (rawmode, stride, 1))],
+        )
+
+    # -- helper for multi-page save --
+    if "_debug_multipage" in encoderinfo:
+        # just to access o32 and o16 (using correct byte order)
+        setattr(im, "_debug_multipage", ifd)
+
+
+class AppendingTiffWriter(io.BytesIO):
+    fieldSizes = [
+        0,  # None
+        1,  # byte
+        1,  # ascii
+        2,  # short
+        4,  # long
+        8,  # rational
+        1,  # sbyte
+        1,  # undefined
+        2,  # sshort
+        4,  # slong
+        8,  # srational
+        4,  # float
+        8,  # double
+        4,  # ifd
+        2,  # unicode
+        4,  # complex
+        8,  # long8
+    ]
+
+    Tags = {
+        273,  # StripOffsets
+        288,  # FreeOffsets
+        324,  # TileOffsets
+        519,  # JPEGQTables
+        520,  # JPEGDCTables
+        521,  # JPEGACTables
+    }
+
+    def __init__(self, fn: StrOrBytesPath | IO[bytes], new: bool = False) -> None:
+        self.f: IO[bytes]
+        if is_path(fn):
+            self.name = fn
+            self.close_fp = True
+            try:
+                self.f = open(fn, "w+b" if new else "r+b")
+            except OSError:
+                self.f = open(fn, "w+b")
+        else:
+            self.f = cast(IO[bytes], fn)
+            self.close_fp = False
+        self.beginning = self.f.tell()
+        self.setup()
+
+    def setup(self) -> None:
+        # Reset everything.
+        self.f.seek(self.beginning, os.SEEK_SET)
+
+        self.whereToWriteNewIFDOffset: int | None = None
+        self.offsetOfNewPage = 0
+
+        self.IIMM = iimm = self.f.read(4)
+        if not iimm:
+            # empty file - first page
+            self.isFirst = True
+            return
+
+        self.isFirst = False
+        if iimm == b"II\x2a\x00":
+            self.setEndian("<")
+        elif iimm == b"MM\x00\x2a":
+            self.setEndian(">")
+        else:
+            msg = "Invalid TIFF file header"
+            raise RuntimeError(msg)
+
+        self.skipIFDs()
+        self.goToEnd()
+
+    def finalize(self) -> None:
+        if self.isFirst:
+            return
+
+        # fix offsets
+        self.f.seek(self.offsetOfNewPage)
+
+        iimm = self.f.read(4)
+        if not iimm:
+            # Make it easy to finish a frame without committing to a new one.
+            return
+
+        if iimm != self.IIMM:
+            msg = "IIMM of new page doesn't match IIMM of first page"
+            raise RuntimeError(msg)
+
+        ifd_offset = self.readLong()
+        ifd_offset += self.offsetOfNewPage
+        assert self.whereToWriteNewIFDOffset is not None
+        self.f.seek(self.whereToWriteNewIFDOffset)
+        self.writeLong(ifd_offset)
+        self.f.seek(ifd_offset)
+        self.fixIFD()
+
+    def newFrame(self) -> None:
+        # Call this to finish a frame.
+        self.finalize()
+        self.setup()
+
+    def __enter__(self) -> AppendingTiffWriter:
+        return self
+
+    def __exit__(self, *args: object) -> None:
+        if self.close_fp:
+            self.close()
+
+    def tell(self) -> int:
+        return self.f.tell() - self.offsetOfNewPage
+
+    def seek(self, offset: int, whence: int = io.SEEK_SET) -> int:
+        """
+        :param offset: Distance to seek.
+        :param whence: Whether the distance is relative to the start,
+                       end or current position.
+        :returns: The resulting position, relative to the start.
+        """
+        if whence == os.SEEK_SET:
+            offset += self.offsetOfNewPage
+
+        self.f.seek(offset, whence)
+        return self.tell()
+
+    def goToEnd(self) -> None:
+        self.f.seek(0, os.SEEK_END)
+        pos = self.f.tell()
+
+        # pad to 16 byte boundary
+        pad_bytes = 16 - pos % 16
+        if 0 < pad_bytes < 16:
+            self.f.write(bytes(pad_bytes))
+        self.offsetOfNewPage = self.f.tell()
+
+    def setEndian(self, endian: str) -> None:
+        self.endian = endian
+        self.longFmt = f"{self.endian}L"
+        self.shortFmt = f"{self.endian}H"
+        self.tagFormat = f"{self.endian}HHL"
+
+    def skipIFDs(self) -> None:
+        while True:
+            ifd_offset = self.readLong()
+            if ifd_offset == 0:
+                self.whereToWriteNewIFDOffset = self.f.tell() - 4
+                break
+
+            self.f.seek(ifd_offset)
+            num_tags = self.readShort()
+            self.f.seek(num_tags * 12, os.SEEK_CUR)
+
+    def write(self, data: Buffer, /) -> int:
+        return self.f.write(data)
+
+    def _fmt(self, field_size: int) -> str:
+        try:
+            return {2: "H", 4: "L", 8: "Q"}[field_size]
+        except KeyError:
+            msg = "offset is not supported"
+            raise RuntimeError(msg)
+
+    def _read(self, field_size: int) -> int:
+        (value,) = struct.unpack(
+            self.endian + self._fmt(field_size), self.f.read(field_size)
+        )
+        return value
+
+    def readShort(self) -> int:
+        return self._read(2)
+
+    def readLong(self) -> int:
+        return self._read(4)
+
+    @staticmethod
+    def _verify_bytes_written(bytes_written: int | None, expected: int) -> None:
+        if bytes_written is not None and bytes_written != expected:
+            msg = f"wrote only {bytes_written} bytes but wanted {expected}"
+            raise RuntimeError(msg)
+
+    def rewriteLastShortToLong(self, value: int) -> None:
+        self.f.seek(-2, os.SEEK_CUR)
+        bytes_written = self.f.write(struct.pack(self.longFmt, value))
+        self._verify_bytes_written(bytes_written, 4)
+
+    def _rewriteLast(self, value: int, field_size: int) -> None:
+        self.f.seek(-field_size, os.SEEK_CUR)
+        bytes_written = self.f.write(
+            struct.pack(self.endian + self._fmt(field_size), value)
+        )
+        self._verify_bytes_written(bytes_written, field_size)
+
+    def rewriteLastShort(self, value: int) -> None:
+        return self._rewriteLast(value, 2)
+
+    def rewriteLastLong(self, value: int) -> None:
+        return self._rewriteLast(value, 4)
+
+    def writeShort(self, value: int) -> None:
+        bytes_written = self.f.write(struct.pack(self.shortFmt, value))
+        self._verify_bytes_written(bytes_written, 2)
+
+    def writeLong(self, value: int) -> None:
+        bytes_written = self.f.write(struct.pack(self.longFmt, value))
+        self._verify_bytes_written(bytes_written, 4)
+
+    def close(self) -> None:
+        self.finalize()
+        if self.close_fp:
+            self.f.close()
+
+    def fixIFD(self) -> None:
+        num_tags = self.readShort()
+
+        for i in range(num_tags):
+            tag, field_type, count = struct.unpack(self.tagFormat, self.f.read(8))
+
+            field_size = self.fieldSizes[field_type]
+            total_size = field_size * count
+            is_local = total_size <= 4
+            if not is_local:
+                offset = self.readLong() + self.offsetOfNewPage
+                self.rewriteLastLong(offset)
+
+            if tag in self.Tags:
+                cur_pos = self.f.tell()
+
+                if is_local:
+                    self._fixOffsets(count, field_size)
+                    self.f.seek(cur_pos + 4)
+                else:
+                    self.f.seek(offset)
+                    self._fixOffsets(count, field_size)
+                    self.f.seek(cur_pos)
+
+            elif is_local:
+                # skip the locally stored value that is not an offset
+                self.f.seek(4, os.SEEK_CUR)
+
+    def _fixOffsets(self, count: int, field_size: int) -> None:
+        for i in range(count):
+            offset = self._read(field_size)
+            offset += self.offsetOfNewPage
+            if field_size == 2 and offset >= 65536:
+                # offset is now too large - we must convert shorts to longs
+                if count != 1:
+                    msg = "not implemented"
+                    raise RuntimeError(msg)  # XXX TODO
+
+                # simple case - the offset is just one and therefore it is
+                # local (not referenced with another offset)
+                self.rewriteLastShortToLong(offset)
+                self.f.seek(-10, os.SEEK_CUR)
+                self.writeShort(TiffTags.LONG)  # rewrite the type to LONG
+                self.f.seek(8, os.SEEK_CUR)
+            else:
+                self._rewriteLast(offset, field_size)
+
+    def fixOffsets(
+        self, count: int, isShort: bool = False, isLong: bool = False
+    ) -> None:
+        if isShort:
+            field_size = 2
+        elif isLong:
+            field_size = 4
+        else:
+            field_size = 0
+        return self._fixOffsets(count, field_size)
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    encoderinfo = im.encoderinfo.copy()
+    encoderconfig = im.encoderconfig
+    append_images = list(encoderinfo.get("append_images", []))
+    if not hasattr(im, "n_frames") and not append_images:
+        return _save(im, fp, filename)
+
+    cur_idx = im.tell()
+    try:
+        with AppendingTiffWriter(fp) as tf:
+            for ims in [im] + append_images:
+                ims.encoderinfo = encoderinfo
+                ims.encoderconfig = encoderconfig
+                if not hasattr(ims, "n_frames"):
+                    nfr = 1
+                else:
+                    nfr = ims.n_frames
+
+                for idx in range(nfr):
+                    ims.seek(idx)
+                    ims.load()
+                    _save(ims, tf, filename)
+                    tf.newFrame()
+    finally:
+        im.seek(cur_idx)
+
+
+#
+# --------------------------------------------------------------------
+# Register
+
+Image.register_open(TiffImageFile.format, TiffImageFile, _accept)
+Image.register_save(TiffImageFile.format, _save)
+Image.register_save_all(TiffImageFile.format, _save_all)
+
+Image.register_extensions(TiffImageFile.format, [".tif", ".tiff"])
+
+Image.register_mime(TiffImageFile.format, "image/tiff")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/TiffTags.py b/backend/venv/lib/python3.12/site-packages/PIL/TiffTags.py
new file mode 100644
index 00000000..86adaa45
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/TiffTags.py
@@ -0,0 +1,562 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# TIFF tags
+#
+# This module provides clear-text names for various well-known
+# TIFF tags.  the TIFF codec works just fine without it.
+#
+# Copyright (c) Secret Labs AB 1999.
+#
+# See the README file for information on usage and redistribution.
+#
+
+##
+# This module provides constants and clear-text names for various
+# well-known TIFF tags.
+##
+from __future__ import annotations
+
+from typing import NamedTuple
+
+
+class _TagInfo(NamedTuple):
+    value: int | None
+    name: str
+    type: int | None
+    length: int | None
+    enum: dict[str, int]
+
+
+class TagInfo(_TagInfo):
+    __slots__: list[str] = []
+
+    def __new__(
+        cls,
+        value: int | None = None,
+        name: str = "unknown",
+        type: int | None = None,
+        length: int | None = None,
+        enum: dict[str, int] | None = None,
+    ) -> TagInfo:
+        return super().__new__(cls, value, name, type, length, enum or {})
+
+    def cvt_enum(self, value: str) -> int | str:
+        # Using get will call hash(value), which can be expensive
+        # for some types (e.g. Fraction). Since self.enum is rarely
+        # used, it's usually better to test it first.
+        return self.enum.get(value, value) if self.enum else value
+
+
+def lookup(tag: int, group: int | None = None) -> TagInfo:
+    """
+    :param tag: Integer tag number
+    :param group: Which :py:data:`~PIL.TiffTags.TAGS_V2_GROUPS` to look in
+
+    .. versionadded:: 8.3.0
+
+    :returns: Taginfo namedtuple, From the ``TAGS_V2`` info if possible,
+        otherwise just populating the value and name from ``TAGS``.
+        If the tag is not recognized, "unknown" is returned for the name
+
+    """
+
+    if group is not None:
+        info = TAGS_V2_GROUPS[group].get(tag) if group in TAGS_V2_GROUPS else None
+    else:
+        info = TAGS_V2.get(tag)
+    return info or TagInfo(tag, TAGS.get(tag, "unknown"))
+
+
+##
+# Map tag numbers to tag info.
+#
+#  id: (Name, Type, Length[, enum_values])
+#
+# The length here differs from the length in the tiff spec.  For
+# numbers, the tiff spec is for the number of fields returned. We
+# agree here.  For string-like types, the tiff spec uses the length of
+# field in bytes.  In Pillow, we are using the number of expected
+# fields, in general 1 for string-like types.
+
+
+BYTE = 1
+ASCII = 2
+SHORT = 3
+LONG = 4
+RATIONAL = 5
+SIGNED_BYTE = 6
+UNDEFINED = 7
+SIGNED_SHORT = 8
+SIGNED_LONG = 9
+SIGNED_RATIONAL = 10
+FLOAT = 11
+DOUBLE = 12
+IFD = 13
+LONG8 = 16
+
+_tags_v2: dict[int, tuple[str, int, int] | tuple[str, int, int, dict[str, int]]] = {
+    254: ("NewSubfileType", LONG, 1),
+    255: ("SubfileType", SHORT, 1),
+    256: ("ImageWidth", LONG, 1),
+    257: ("ImageLength", LONG, 1),
+    258: ("BitsPerSample", SHORT, 0),
+    259: (
+        "Compression",
+        SHORT,
+        1,
+        {
+            "Uncompressed": 1,
+            "CCITT 1d": 2,
+            "Group 3 Fax": 3,
+            "Group 4 Fax": 4,
+            "LZW": 5,
+            "JPEG": 6,
+            "PackBits": 32773,
+        },
+    ),
+    262: (
+        "PhotometricInterpretation",
+        SHORT,
+        1,
+        {
+            "WhiteIsZero": 0,
+            "BlackIsZero": 1,
+            "RGB": 2,
+            "RGB Palette": 3,
+            "Transparency Mask": 4,
+            "CMYK": 5,
+            "YCbCr": 6,
+            "CieLAB": 8,
+            "CFA": 32803,  # TIFF/EP, Adobe DNG
+            "LinearRaw": 32892,  # Adobe DNG
+        },
+    ),
+    263: ("Threshholding", SHORT, 1),
+    264: ("CellWidth", SHORT, 1),
+    265: ("CellLength", SHORT, 1),
+    266: ("FillOrder", SHORT, 1),
+    269: ("DocumentName", ASCII, 1),
+    270: ("ImageDescription", ASCII, 1),
+    271: ("Make", ASCII, 1),
+    272: ("Model", ASCII, 1),
+    273: ("StripOffsets", LONG, 0),
+    274: ("Orientation", SHORT, 1),
+    277: ("SamplesPerPixel", SHORT, 1),
+    278: ("RowsPerStrip", LONG, 1),
+    279: ("StripByteCounts", LONG, 0),
+    280: ("MinSampleValue", SHORT, 0),
+    281: ("MaxSampleValue", SHORT, 0),
+    282: ("XResolution", RATIONAL, 1),
+    283: ("YResolution", RATIONAL, 1),
+    284: ("PlanarConfiguration", SHORT, 1, {"Contiguous": 1, "Separate": 2}),
+    285: ("PageName", ASCII, 1),
+    286: ("XPosition", RATIONAL, 1),
+    287: ("YPosition", RATIONAL, 1),
+    288: ("FreeOffsets", LONG, 1),
+    289: ("FreeByteCounts", LONG, 1),
+    290: ("GrayResponseUnit", SHORT, 1),
+    291: ("GrayResponseCurve", SHORT, 0),
+    292: ("T4Options", LONG, 1),
+    293: ("T6Options", LONG, 1),
+    296: ("ResolutionUnit", SHORT, 1, {"none": 1, "inch": 2, "cm": 3}),
+    297: ("PageNumber", SHORT, 2),
+    301: ("TransferFunction", SHORT, 0),
+    305: ("Software", ASCII, 1),
+    306: ("DateTime", ASCII, 1),
+    315: ("Artist", ASCII, 1),
+    316: ("HostComputer", ASCII, 1),
+    317: ("Predictor", SHORT, 1, {"none": 1, "Horizontal Differencing": 2}),
+    318: ("WhitePoint", RATIONAL, 2),
+    319: ("PrimaryChromaticities", RATIONAL, 6),
+    320: ("ColorMap", SHORT, 0),
+    321: ("HalftoneHints", SHORT, 2),
+    322: ("TileWidth", LONG, 1),
+    323: ("TileLength", LONG, 1),
+    324: ("TileOffsets", LONG, 0),
+    325: ("TileByteCounts", LONG, 0),
+    330: ("SubIFDs", LONG, 0),
+    332: ("InkSet", SHORT, 1),
+    333: ("InkNames", ASCII, 1),
+    334: ("NumberOfInks", SHORT, 1),
+    336: ("DotRange", SHORT, 0),
+    337: ("TargetPrinter", ASCII, 1),
+    338: ("ExtraSamples", SHORT, 0),
+    339: ("SampleFormat", SHORT, 0),
+    340: ("SMinSampleValue", DOUBLE, 0),
+    341: ("SMaxSampleValue", DOUBLE, 0),
+    342: ("TransferRange", SHORT, 6),
+    347: ("JPEGTables", UNDEFINED, 1),
+    # obsolete JPEG tags
+    512: ("JPEGProc", SHORT, 1),
+    513: ("JPEGInterchangeFormat", LONG, 1),
+    514: ("JPEGInterchangeFormatLength", LONG, 1),
+    515: ("JPEGRestartInterval", SHORT, 1),
+    517: ("JPEGLosslessPredictors", SHORT, 0),
+    518: ("JPEGPointTransforms", SHORT, 0),
+    519: ("JPEGQTables", LONG, 0),
+    520: ("JPEGDCTables", LONG, 0),
+    521: ("JPEGACTables", LONG, 0),
+    529: ("YCbCrCoefficients", RATIONAL, 3),
+    530: ("YCbCrSubSampling", SHORT, 2),
+    531: ("YCbCrPositioning", SHORT, 1),
+    532: ("ReferenceBlackWhite", RATIONAL, 6),
+    700: ("XMP", BYTE, 0),
+    33432: ("Copyright", ASCII, 1),
+    33723: ("IptcNaaInfo", UNDEFINED, 1),
+    34377: ("PhotoshopInfo", BYTE, 0),
+    # FIXME add more tags here
+    34665: ("ExifIFD", LONG, 1),
+    34675: ("ICCProfile", UNDEFINED, 1),
+    34853: ("GPSInfoIFD", LONG, 1),
+    36864: ("ExifVersion", UNDEFINED, 1),
+    37724: ("ImageSourceData", UNDEFINED, 1),
+    40965: ("InteroperabilityIFD", LONG, 1),
+    41730: ("CFAPattern", UNDEFINED, 1),
+    # MPInfo
+    45056: ("MPFVersion", UNDEFINED, 1),
+    45057: ("NumberOfImages", LONG, 1),
+    45058: ("MPEntry", UNDEFINED, 1),
+    45059: ("ImageUIDList", UNDEFINED, 0),  # UNDONE, check
+    45060: ("TotalFrames", LONG, 1),
+    45313: ("MPIndividualNum", LONG, 1),
+    45569: ("PanOrientation", LONG, 1),
+    45570: ("PanOverlap_H", RATIONAL, 1),
+    45571: ("PanOverlap_V", RATIONAL, 1),
+    45572: ("BaseViewpointNum", LONG, 1),
+    45573: ("ConvergenceAngle", SIGNED_RATIONAL, 1),
+    45574: ("BaselineLength", RATIONAL, 1),
+    45575: ("VerticalDivergence", SIGNED_RATIONAL, 1),
+    45576: ("AxisDistance_X", SIGNED_RATIONAL, 1),
+    45577: ("AxisDistance_Y", SIGNED_RATIONAL, 1),
+    45578: ("AxisDistance_Z", SIGNED_RATIONAL, 1),
+    45579: ("YawAngle", SIGNED_RATIONAL, 1),
+    45580: ("PitchAngle", SIGNED_RATIONAL, 1),
+    45581: ("RollAngle", SIGNED_RATIONAL, 1),
+    40960: ("FlashPixVersion", UNDEFINED, 1),
+    50741: ("MakerNoteSafety", SHORT, 1, {"Unsafe": 0, "Safe": 1}),
+    50780: ("BestQualityScale", RATIONAL, 1),
+    50838: ("ImageJMetaDataByteCounts", LONG, 0),  # Can be more than one
+    50839: ("ImageJMetaData", UNDEFINED, 1),  # see Issue #2006
+}
+_tags_v2_groups = {
+    # ExifIFD
+    34665: {
+        36864: ("ExifVersion", UNDEFINED, 1),
+        40960: ("FlashPixVersion", UNDEFINED, 1),
+        40965: ("InteroperabilityIFD", LONG, 1),
+        41730: ("CFAPattern", UNDEFINED, 1),
+    },
+    # GPSInfoIFD
+    34853: {
+        0: ("GPSVersionID", BYTE, 4),
+        1: ("GPSLatitudeRef", ASCII, 2),
+        2: ("GPSLatitude", RATIONAL, 3),
+        3: ("GPSLongitudeRef", ASCII, 2),
+        4: ("GPSLongitude", RATIONAL, 3),
+        5: ("GPSAltitudeRef", BYTE, 1),
+        6: ("GPSAltitude", RATIONAL, 1),
+        7: ("GPSTimeStamp", RATIONAL, 3),
+        8: ("GPSSatellites", ASCII, 0),
+        9: ("GPSStatus", ASCII, 2),
+        10: ("GPSMeasureMode", ASCII, 2),
+        11: ("GPSDOP", RATIONAL, 1),
+        12: ("GPSSpeedRef", ASCII, 2),
+        13: ("GPSSpeed", RATIONAL, 1),
+        14: ("GPSTrackRef", ASCII, 2),
+        15: ("GPSTrack", RATIONAL, 1),
+        16: ("GPSImgDirectionRef", ASCII, 2),
+        17: ("GPSImgDirection", RATIONAL, 1),
+        18: ("GPSMapDatum", ASCII, 0),
+        19: ("GPSDestLatitudeRef", ASCII, 2),
+        20: ("GPSDestLatitude", RATIONAL, 3),
+        21: ("GPSDestLongitudeRef", ASCII, 2),
+        22: ("GPSDestLongitude", RATIONAL, 3),
+        23: ("GPSDestBearingRef", ASCII, 2),
+        24: ("GPSDestBearing", RATIONAL, 1),
+        25: ("GPSDestDistanceRef", ASCII, 2),
+        26: ("GPSDestDistance", RATIONAL, 1),
+        27: ("GPSProcessingMethod", UNDEFINED, 0),
+        28: ("GPSAreaInformation", UNDEFINED, 0),
+        29: ("GPSDateStamp", ASCII, 11),
+        30: ("GPSDifferential", SHORT, 1),
+    },
+    # InteroperabilityIFD
+    40965: {1: ("InteropIndex", ASCII, 1), 2: ("InteropVersion", UNDEFINED, 1)},
+}
+
+# Legacy Tags structure
+# these tags aren't included above, but were in the previous versions
+TAGS: dict[int | tuple[int, int], str] = {
+    347: "JPEGTables",
+    700: "XMP",
+    # Additional Exif Info
+    32932: "Wang Annotation",
+    33434: "ExposureTime",
+    33437: "FNumber",
+    33445: "MD FileTag",
+    33446: "MD ScalePixel",
+    33447: "MD ColorTable",
+    33448: "MD LabName",
+    33449: "MD SampleInfo",
+    33450: "MD PrepDate",
+    33451: "MD PrepTime",
+    33452: "MD FileUnits",
+    33550: "ModelPixelScaleTag",
+    33723: "IptcNaaInfo",
+    33918: "INGR Packet Data Tag",
+    33919: "INGR Flag Registers",
+    33920: "IrasB Transformation Matrix",
+    33922: "ModelTiepointTag",
+    34264: "ModelTransformationTag",
+    34377: "PhotoshopInfo",
+    34735: "GeoKeyDirectoryTag",
+    34736: "GeoDoubleParamsTag",
+    34737: "GeoAsciiParamsTag",
+    34850: "ExposureProgram",
+    34852: "SpectralSensitivity",
+    34855: "ISOSpeedRatings",
+    34856: "OECF",
+    34864: "SensitivityType",
+    34865: "StandardOutputSensitivity",
+    34866: "RecommendedExposureIndex",
+    34867: "ISOSpeed",
+    34868: "ISOSpeedLatitudeyyy",
+    34869: "ISOSpeedLatitudezzz",
+    34908: "HylaFAX FaxRecvParams",
+    34909: "HylaFAX FaxSubAddress",
+    34910: "HylaFAX FaxRecvTime",
+    36864: "ExifVersion",
+    36867: "DateTimeOriginal",
+    36868: "DateTimeDigitized",
+    37121: "ComponentsConfiguration",
+    37122: "CompressedBitsPerPixel",
+    37724: "ImageSourceData",
+    37377: "ShutterSpeedValue",
+    37378: "ApertureValue",
+    37379: "BrightnessValue",
+    37380: "ExposureBiasValue",
+    37381: "MaxApertureValue",
+    37382: "SubjectDistance",
+    37383: "MeteringMode",
+    37384: "LightSource",
+    37385: "Flash",
+    37386: "FocalLength",
+    37396: "SubjectArea",
+    37500: "MakerNote",
+    37510: "UserComment",
+    37520: "SubSec",
+    37521: "SubSecTimeOriginal",
+    37522: "SubsecTimeDigitized",
+    40960: "FlashPixVersion",
+    40961: "ColorSpace",
+    40962: "PixelXDimension",
+    40963: "PixelYDimension",
+    40964: "RelatedSoundFile",
+    40965: "InteroperabilityIFD",
+    41483: "FlashEnergy",
+    41484: "SpatialFrequencyResponse",
+    41486: "FocalPlaneXResolution",
+    41487: "FocalPlaneYResolution",
+    41488: "FocalPlaneResolutionUnit",
+    41492: "SubjectLocation",
+    41493: "ExposureIndex",
+    41495: "SensingMethod",
+    41728: "FileSource",
+    41729: "SceneType",
+    41730: "CFAPattern",
+    41985: "CustomRendered",
+    41986: "ExposureMode",
+    41987: "WhiteBalance",
+    41988: "DigitalZoomRatio",
+    41989: "FocalLengthIn35mmFilm",
+    41990: "SceneCaptureType",
+    41991: "GainControl",
+    41992: "Contrast",
+    41993: "Saturation",
+    41994: "Sharpness",
+    41995: "DeviceSettingDescription",
+    41996: "SubjectDistanceRange",
+    42016: "ImageUniqueID",
+    42032: "CameraOwnerName",
+    42033: "BodySerialNumber",
+    42034: "LensSpecification",
+    42035: "LensMake",
+    42036: "LensModel",
+    42037: "LensSerialNumber",
+    42112: "GDAL_METADATA",
+    42113: "GDAL_NODATA",
+    42240: "Gamma",
+    50215: "Oce Scanjob Description",
+    50216: "Oce Application Selector",
+    50217: "Oce Identification Number",
+    50218: "Oce ImageLogic Characteristics",
+    # Adobe DNG
+    50706: "DNGVersion",
+    50707: "DNGBackwardVersion",
+    50708: "UniqueCameraModel",
+    50709: "LocalizedCameraModel",
+    50710: "CFAPlaneColor",
+    50711: "CFALayout",
+    50712: "LinearizationTable",
+    50713: "BlackLevelRepeatDim",
+    50714: "BlackLevel",
+    50715: "BlackLevelDeltaH",
+    50716: "BlackLevelDeltaV",
+    50717: "WhiteLevel",
+    50718: "DefaultScale",
+    50719: "DefaultCropOrigin",
+    50720: "DefaultCropSize",
+    50721: "ColorMatrix1",
+    50722: "ColorMatrix2",
+    50723: "CameraCalibration1",
+    50724: "CameraCalibration2",
+    50725: "ReductionMatrix1",
+    50726: "ReductionMatrix2",
+    50727: "AnalogBalance",
+    50728: "AsShotNeutral",
+    50729: "AsShotWhiteXY",
+    50730: "BaselineExposure",
+    50731: "BaselineNoise",
+    50732: "BaselineSharpness",
+    50733: "BayerGreenSplit",
+    50734: "LinearResponseLimit",
+    50735: "CameraSerialNumber",
+    50736: "LensInfo",
+    50737: "ChromaBlurRadius",
+    50738: "AntiAliasStrength",
+    50740: "DNGPrivateData",
+    50778: "CalibrationIlluminant1",
+    50779: "CalibrationIlluminant2",
+    50784: "Alias Layer Metadata",
+}
+
+TAGS_V2: dict[int, TagInfo] = {}
+TAGS_V2_GROUPS: dict[int, dict[int, TagInfo]] = {}
+
+
+def _populate() -> None:
+    for k, v in _tags_v2.items():
+        # Populate legacy structure.
+        TAGS[k] = v[0]
+        if len(v) == 4:
+            for sk, sv in v[3].items():
+                TAGS[(k, sv)] = sk
+
+        TAGS_V2[k] = TagInfo(k, *v)
+
+    for group, tags in _tags_v2_groups.items():
+        TAGS_V2_GROUPS[group] = {k: TagInfo(k, *v) for k, v in tags.items()}
+
+
+_populate()
+##
+# Map type numbers to type names -- defined in ImageFileDirectory.
+
+TYPES: dict[int, str] = {}
+
+#
+# These tags are handled by default in libtiff, without
+# adding to the custom dictionary. From tif_dir.c, searching for
+# case TIFFTAG in the _TIFFVSetField function:
+# Line: item.
+# 148: case TIFFTAG_SUBFILETYPE:
+# 151: case TIFFTAG_IMAGEWIDTH:
+# 154: case TIFFTAG_IMAGELENGTH:
+# 157: case TIFFTAG_BITSPERSAMPLE:
+# 181: case TIFFTAG_COMPRESSION:
+# 202: case TIFFTAG_PHOTOMETRIC:
+# 205: case TIFFTAG_THRESHHOLDING:
+# 208: case TIFFTAG_FILLORDER:
+# 214: case TIFFTAG_ORIENTATION:
+# 221: case TIFFTAG_SAMPLESPERPIXEL:
+# 228: case TIFFTAG_ROWSPERSTRIP:
+# 238: case TIFFTAG_MINSAMPLEVALUE:
+# 241: case TIFFTAG_MAXSAMPLEVALUE:
+# 244: case TIFFTAG_SMINSAMPLEVALUE:
+# 247: case TIFFTAG_SMAXSAMPLEVALUE:
+# 250: case TIFFTAG_XRESOLUTION:
+# 256: case TIFFTAG_YRESOLUTION:
+# 262: case TIFFTAG_PLANARCONFIG:
+# 268: case TIFFTAG_XPOSITION:
+# 271: case TIFFTAG_YPOSITION:
+# 274: case TIFFTAG_RESOLUTIONUNIT:
+# 280: case TIFFTAG_PAGENUMBER:
+# 284: case TIFFTAG_HALFTONEHINTS:
+# 288: case TIFFTAG_COLORMAP:
+# 294: case TIFFTAG_EXTRASAMPLES:
+# 298: case TIFFTAG_MATTEING:
+# 305: case TIFFTAG_TILEWIDTH:
+# 316: case TIFFTAG_TILELENGTH:
+# 327: case TIFFTAG_TILEDEPTH:
+# 333: case TIFFTAG_DATATYPE:
+# 344: case TIFFTAG_SAMPLEFORMAT:
+# 361: case TIFFTAG_IMAGEDEPTH:
+# 364: case TIFFTAG_SUBIFD:
+# 376: case TIFFTAG_YCBCRPOSITIONING:
+# 379: case TIFFTAG_YCBCRSUBSAMPLING:
+# 383: case TIFFTAG_TRANSFERFUNCTION:
+# 389: case TIFFTAG_REFERENCEBLACKWHITE:
+# 393: case TIFFTAG_INKNAMES:
+
+# Following pseudo-tags are also handled by default in libtiff:
+# TIFFTAG_JPEGQUALITY 65537
+
+# some of these are not in our TAGS_V2 dict and were included from tiff.h
+
+# This list also exists in encode.c
+LIBTIFF_CORE = {
+    255,
+    256,
+    257,
+    258,
+    259,
+    262,
+    263,
+    266,
+    274,
+    277,
+    278,
+    280,
+    281,
+    340,
+    341,
+    282,
+    283,
+    284,
+    286,
+    287,
+    296,
+    297,
+    321,
+    320,
+    338,
+    32995,
+    322,
+    323,
+    32998,
+    32996,
+    339,
+    32997,
+    330,
+    531,
+    530,
+    301,
+    532,
+    333,
+    # as above
+    269,  # this has been in our tests forever, and works
+    65537,
+}
+
+LIBTIFF_CORE.remove(255)  # We don't have support for subfiletypes
+LIBTIFF_CORE.remove(322)  # We don't have support for writing tiled images with libtiff
+LIBTIFF_CORE.remove(323)  # Tiled images
+LIBTIFF_CORE.remove(333)  # Ink Names either
+
+# Note to advanced users: There may be combinations of these
+# parameters and values that when added properly, will work and
+# produce valid tiff images that may work in your application.
+# It is safe to add and remove tags from this set from Pillow's point
+# of view so long as you test against libtiff.
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/WalImageFile.py b/backend/venv/lib/python3.12/site-packages/PIL/WalImageFile.py
new file mode 100644
index 00000000..87e32878
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/WalImageFile.py
@@ -0,0 +1,127 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# WAL file handling
+#
+# History:
+# 2003-04-23 fl   created
+#
+# Copyright (c) 2003 by Fredrik Lundh.
+#
+# See the README file for information on usage and redistribution.
+#
+
+"""
+This reader is based on the specification available from:
+https://www.flipcode.com/archives/Quake_2_BSP_File_Format.shtml
+and has been tested with a few sample files found using google.
+
+.. note::
+    This format cannot be automatically recognized, so the reader
+    is not registered for use with :py:func:`PIL.Image.open()`.
+    To open a WAL file, use the :py:func:`PIL.WalImageFile.open()` function instead.
+"""
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i32le as i32
+from ._typing import StrOrBytesPath
+
+
+class WalImageFile(ImageFile.ImageFile):
+    format = "WAL"
+    format_description = "Quake2 Texture"
+
+    def _open(self) -> None:
+        self._mode = "P"
+
+        # read header fields
+        header = self.fp.read(32 + 24 + 32 + 12)
+        self._size = i32(header, 32), i32(header, 36)
+        Image._decompression_bomb_check(self.size)
+
+        # load pixel data
+        offset = i32(header, 40)
+        self.fp.seek(offset)
+
+        # strings are null-terminated
+        self.info["name"] = header[:32].split(b"\0", 1)[0]
+        next_name = header[56 : 56 + 32].split(b"\0", 1)[0]
+        if next_name:
+            self.info["next_name"] = next_name
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self._im is None:
+            self.im = Image.core.new(self.mode, self.size)
+            self.frombytes(self.fp.read(self.size[0] * self.size[1]))
+            self.putpalette(quake2palette)
+        return Image.Image.load(self)
+
+
+def open(filename: StrOrBytesPath | IO[bytes]) -> WalImageFile:
+    """
+    Load texture from a Quake2 WAL texture file.
+
+    By default, a Quake2 standard palette is attached to the texture.
+    To override the palette, use the :py:func:`PIL.Image.Image.putpalette()` method.
+
+    :param filename: WAL file name, or an opened file handle.
+    :returns: An image instance.
+    """
+    return WalImageFile(filename)
+
+
+quake2palette = (
+    # default palette taken from piffo 0.93 by Hans Häggström
+    b"\x01\x01\x01\x0b\x0b\x0b\x12\x12\x12\x17\x17\x17\x1b\x1b\x1b\x1e"
+    b"\x1e\x1e\x22\x22\x22\x26\x26\x26\x29\x29\x29\x2c\x2c\x2c\x2f\x2f"
+    b"\x2f\x32\x32\x32\x35\x35\x35\x37\x37\x37\x3a\x3a\x3a\x3c\x3c\x3c"
+    b"\x24\x1e\x13\x22\x1c\x12\x20\x1b\x12\x1f\x1a\x10\x1d\x19\x10\x1b"
+    b"\x17\x0f\x1a\x16\x0f\x18\x14\x0d\x17\x13\x0d\x16\x12\x0d\x14\x10"
+    b"\x0b\x13\x0f\x0b\x10\x0d\x0a\x0f\x0b\x0a\x0d\x0b\x07\x0b\x0a\x07"
+    b"\x23\x23\x26\x22\x22\x25\x22\x20\x23\x21\x1f\x22\x20\x1e\x20\x1f"
+    b"\x1d\x1e\x1d\x1b\x1c\x1b\x1a\x1a\x1a\x19\x19\x18\x17\x17\x17\x16"
+    b"\x16\x14\x14\x14\x13\x13\x13\x10\x10\x10\x0f\x0f\x0f\x0d\x0d\x0d"
+    b"\x2d\x28\x20\x29\x24\x1c\x27\x22\x1a\x25\x1f\x17\x38\x2e\x1e\x31"
+    b"\x29\x1a\x2c\x25\x17\x26\x20\x14\x3c\x30\x14\x37\x2c\x13\x33\x28"
+    b"\x12\x2d\x24\x10\x28\x1f\x0f\x22\x1a\x0b\x1b\x14\x0a\x13\x0f\x07"
+    b"\x31\x1a\x16\x30\x17\x13\x2e\x16\x10\x2c\x14\x0d\x2a\x12\x0b\x27"
+    b"\x0f\x0a\x25\x0f\x07\x21\x0d\x01\x1e\x0b\x01\x1c\x0b\x01\x1a\x0b"
+    b"\x01\x18\x0a\x01\x16\x0a\x01\x13\x0a\x01\x10\x07\x01\x0d\x07\x01"
+    b"\x29\x23\x1e\x27\x21\x1c\x26\x20\x1b\x25\x1f\x1a\x23\x1d\x19\x21"
+    b"\x1c\x18\x20\x1b\x17\x1e\x19\x16\x1c\x18\x14\x1b\x17\x13\x19\x14"
+    b"\x10\x17\x13\x0f\x14\x10\x0d\x12\x0f\x0b\x0f\x0b\x0a\x0b\x0a\x07"
+    b"\x26\x1a\x0f\x23\x19\x0f\x20\x17\x0f\x1c\x16\x0f\x19\x13\x0d\x14"
+    b"\x10\x0b\x10\x0d\x0a\x0b\x0a\x07\x33\x22\x1f\x35\x29\x26\x37\x2f"
+    b"\x2d\x39\x35\x34\x37\x39\x3a\x33\x37\x39\x30\x34\x36\x2b\x31\x34"
+    b"\x27\x2e\x31\x22\x2b\x2f\x1d\x28\x2c\x17\x25\x2a\x0f\x20\x26\x0d"
+    b"\x1e\x25\x0b\x1c\x22\x0a\x1b\x20\x07\x19\x1e\x07\x17\x1b\x07\x14"
+    b"\x18\x01\x12\x16\x01\x0f\x12\x01\x0b\x0d\x01\x07\x0a\x01\x01\x01"
+    b"\x2c\x21\x21\x2a\x1f\x1f\x29\x1d\x1d\x27\x1c\x1c\x26\x1a\x1a\x24"
+    b"\x18\x18\x22\x17\x17\x21\x16\x16\x1e\x13\x13\x1b\x12\x12\x18\x10"
+    b"\x10\x16\x0d\x0d\x12\x0b\x0b\x0d\x0a\x0a\x0a\x07\x07\x01\x01\x01"
+    b"\x2e\x30\x29\x2d\x2e\x27\x2b\x2c\x26\x2a\x2a\x24\x28\x29\x23\x27"
+    b"\x27\x21\x26\x26\x1f\x24\x24\x1d\x22\x22\x1c\x1f\x1f\x1a\x1c\x1c"
+    b"\x18\x19\x19\x16\x17\x17\x13\x13\x13\x10\x0f\x0f\x0d\x0b\x0b\x0a"
+    b"\x30\x1e\x1b\x2d\x1c\x19\x2c\x1a\x17\x2a\x19\x14\x28\x17\x13\x26"
+    b"\x16\x10\x24\x13\x0f\x21\x12\x0d\x1f\x10\x0b\x1c\x0f\x0a\x19\x0d"
+    b"\x0a\x16\x0b\x07\x12\x0a\x07\x0f\x07\x01\x0a\x01\x01\x01\x01\x01"
+    b"\x28\x29\x38\x26\x27\x36\x25\x26\x34\x24\x24\x31\x22\x22\x2f\x20"
+    b"\x21\x2d\x1e\x1f\x2a\x1d\x1d\x27\x1b\x1b\x25\x19\x19\x21\x17\x17"
+    b"\x1e\x14\x14\x1b\x13\x12\x17\x10\x0f\x13\x0d\x0b\x0f\x0a\x07\x07"
+    b"\x2f\x32\x29\x2d\x30\x26\x2b\x2e\x24\x29\x2c\x21\x27\x2a\x1e\x25"
+    b"\x28\x1c\x23\x26\x1a\x21\x25\x18\x1e\x22\x14\x1b\x1f\x10\x19\x1c"
+    b"\x0d\x17\x1a\x0a\x13\x17\x07\x10\x13\x01\x0d\x0f\x01\x0a\x0b\x01"
+    b"\x01\x3f\x01\x13\x3c\x0b\x1b\x39\x10\x20\x35\x14\x23\x31\x17\x23"
+    b"\x2d\x18\x23\x29\x18\x3f\x3f\x3f\x3f\x3f\x39\x3f\x3f\x31\x3f\x3f"
+    b"\x2a\x3f\x3f\x20\x3f\x3f\x14\x3f\x3c\x12\x3f\x39\x0f\x3f\x35\x0b"
+    b"\x3f\x32\x07\x3f\x2d\x01\x3d\x2a\x01\x3b\x26\x01\x39\x21\x01\x37"
+    b"\x1d\x01\x34\x1a\x01\x32\x16\x01\x2f\x12\x01\x2d\x0f\x01\x2a\x0b"
+    b"\x01\x27\x07\x01\x23\x01\x01\x1d\x01\x01\x17\x01\x01\x10\x01\x01"
+    b"\x3d\x01\x01\x19\x19\x3f\x3f\x01\x01\x01\x01\x3f\x16\x16\x13\x10"
+    b"\x10\x0f\x0d\x0d\x0b\x3c\x2e\x2a\x36\x27\x20\x30\x21\x18\x29\x1b"
+    b"\x10\x3c\x39\x37\x37\x32\x2f\x31\x2c\x28\x2b\x26\x21\x30\x22\x20"
+)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/WebPImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/WebPImagePlugin.py
new file mode 100644
index 00000000..c7f85552
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/WebPImagePlugin.py
@@ -0,0 +1,322 @@
+from __future__ import annotations
+
+from io import BytesIO
+from typing import IO, Any
+
+from . import Image, ImageFile
+
+try:
+    from . import _webp
+
+    SUPPORTED = True
+except ImportError:
+    SUPPORTED = False
+
+
+_VP8_MODES_BY_IDENTIFIER = {
+    b"VP8 ": "RGB",
+    b"VP8X": "RGBA",
+    b"VP8L": "RGBA",  # lossless
+}
+
+
+def _accept(prefix: bytes) -> bool | str:
+    is_riff_file_format = prefix[:4] == b"RIFF"
+    is_webp_file = prefix[8:12] == b"WEBP"
+    is_valid_vp8_mode = prefix[12:16] in _VP8_MODES_BY_IDENTIFIER
+
+    if is_riff_file_format and is_webp_file and is_valid_vp8_mode:
+        if not SUPPORTED:
+            return (
+                "image file could not be identified because WEBP support not installed"
+            )
+        return True
+    return False
+
+
+class WebPImageFile(ImageFile.ImageFile):
+    format = "WEBP"
+    format_description = "WebP image"
+    __loaded = 0
+    __logical_frame = 0
+
+    def _open(self) -> None:
+        # Use the newer AnimDecoder API to parse the (possibly) animated file,
+        # and access muxed chunks like ICC/EXIF/XMP.
+        self._decoder = _webp.WebPAnimDecoder(self.fp.read())
+
+        # Get info from decoder
+        width, height, loop_count, bgcolor, frame_count, mode = self._decoder.get_info()
+        self._size = width, height
+        self.info["loop"] = loop_count
+        bg_a, bg_r, bg_g, bg_b = (
+            (bgcolor >> 24) & 0xFF,
+            (bgcolor >> 16) & 0xFF,
+            (bgcolor >> 8) & 0xFF,
+            bgcolor & 0xFF,
+        )
+        self.info["background"] = (bg_r, bg_g, bg_b, bg_a)
+        self.n_frames = frame_count
+        self.is_animated = self.n_frames > 1
+        self._mode = "RGB" if mode == "RGBX" else mode
+        self.rawmode = mode
+
+        # Attempt to read ICC / EXIF / XMP chunks from file
+        icc_profile = self._decoder.get_chunk("ICCP")
+        exif = self._decoder.get_chunk("EXIF")
+        xmp = self._decoder.get_chunk("XMP ")
+        if icc_profile:
+            self.info["icc_profile"] = icc_profile
+        if exif:
+            self.info["exif"] = exif
+        if xmp:
+            self.info["xmp"] = xmp
+
+        # Initialize seek state
+        self._reset(reset=False)
+
+    def _getexif(self) -> dict[int, Any] | None:
+        if "exif" not in self.info:
+            return None
+        return self.getexif()._get_merged_dict()
+
+    def seek(self, frame: int) -> None:
+        if not self._seek_check(frame):
+            return
+
+        # Set logical frame to requested position
+        self.__logical_frame = frame
+
+    def _reset(self, reset: bool = True) -> None:
+        if reset:
+            self._decoder.reset()
+        self.__physical_frame = 0
+        self.__loaded = -1
+        self.__timestamp = 0
+
+    def _get_next(self) -> tuple[bytes, int, int]:
+        # Get next frame
+        ret = self._decoder.get_next()
+        self.__physical_frame += 1
+
+        # Check if an error occurred
+        if ret is None:
+            self._reset()  # Reset just to be safe
+            self.seek(0)
+            msg = "failed to decode next frame in WebP file"
+            raise EOFError(msg)
+
+        # Compute duration
+        data, timestamp = ret
+        duration = timestamp - self.__timestamp
+        self.__timestamp = timestamp
+
+        # libwebp gives frame end, adjust to start of frame
+        timestamp -= duration
+        return data, timestamp, duration
+
+    def _seek(self, frame: int) -> None:
+        if self.__physical_frame == frame:
+            return  # Nothing to do
+        if frame < self.__physical_frame:
+            self._reset()  # Rewind to beginning
+        while self.__physical_frame < frame:
+            self._get_next()  # Advance to the requested frame
+
+    def load(self) -> Image.core.PixelAccess | None:
+        if self.__loaded != self.__logical_frame:
+            self._seek(self.__logical_frame)
+
+            # We need to load the image data for this frame
+            data, timestamp, duration = self._get_next()
+            self.info["timestamp"] = timestamp
+            self.info["duration"] = duration
+            self.__loaded = self.__logical_frame
+
+            # Set tile
+            if self.fp and self._exclusive_fp:
+                self.fp.close()
+            self.fp = BytesIO(data)
+            self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, 0, self.rawmode)]
+
+        return super().load()
+
+    def load_seek(self, pos: int) -> None:
+        pass
+
+    def tell(self) -> int:
+        return self.__logical_frame
+
+
+def _convert_frame(im: Image.Image) -> Image.Image:
+    # Make sure image mode is supported
+    if im.mode not in ("RGBX", "RGBA", "RGB"):
+        im = im.convert("RGBA" if im.has_transparency_data else "RGB")
+    return im
+
+
+def _save_all(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    encoderinfo = im.encoderinfo.copy()
+    append_images = list(encoderinfo.get("append_images", []))
+
+    # If total frame count is 1, then save using the legacy API, which
+    # will preserve non-alpha modes
+    total = 0
+    for ims in [im] + append_images:
+        total += getattr(ims, "n_frames", 1)
+    if total == 1:
+        _save(im, fp, filename)
+        return
+
+    background: int | tuple[int, ...] = (0, 0, 0, 0)
+    if "background" in encoderinfo:
+        background = encoderinfo["background"]
+    elif "background" in im.info:
+        background = im.info["background"]
+        if isinstance(background, int):
+            # GifImagePlugin stores a global color table index in
+            # info["background"]. So it must be converted to an RGBA value
+            palette = im.getpalette()
+            if palette:
+                r, g, b = palette[background * 3 : (background + 1) * 3]
+                background = (r, g, b, 255)
+            else:
+                background = (background, background, background, 255)
+
+    duration = im.encoderinfo.get("duration", im.info.get("duration", 0))
+    loop = im.encoderinfo.get("loop", 0)
+    minimize_size = im.encoderinfo.get("minimize_size", False)
+    kmin = im.encoderinfo.get("kmin", None)
+    kmax = im.encoderinfo.get("kmax", None)
+    allow_mixed = im.encoderinfo.get("allow_mixed", False)
+    verbose = False
+    lossless = im.encoderinfo.get("lossless", False)
+    quality = im.encoderinfo.get("quality", 80)
+    alpha_quality = im.encoderinfo.get("alpha_quality", 100)
+    method = im.encoderinfo.get("method", 0)
+    icc_profile = im.encoderinfo.get("icc_profile") or ""
+    exif = im.encoderinfo.get("exif", "")
+    if isinstance(exif, Image.Exif):
+        exif = exif.tobytes()
+    xmp = im.encoderinfo.get("xmp", "")
+    if allow_mixed:
+        lossless = False
+
+    # Sensible keyframe defaults are from gif2webp.c script
+    if kmin is None:
+        kmin = 9 if lossless else 3
+    if kmax is None:
+        kmax = 17 if lossless else 5
+
+    # Validate background color
+    if (
+        not isinstance(background, (list, tuple))
+        or len(background) != 4
+        or not all(0 <= v < 256 for v in background)
+    ):
+        msg = f"Background color is not an RGBA tuple clamped to (0-255): {background}"
+        raise OSError(msg)
+
+    # Convert to packed uint
+    bg_r, bg_g, bg_b, bg_a = background
+    background = (bg_a << 24) | (bg_r << 16) | (bg_g << 8) | (bg_b << 0)
+
+    # Setup the WebP animation encoder
+    enc = _webp.WebPAnimEncoder(
+        im.size[0],
+        im.size[1],
+        background,
+        loop,
+        minimize_size,
+        kmin,
+        kmax,
+        allow_mixed,
+        verbose,
+    )
+
+    # Add each frame
+    frame_idx = 0
+    timestamp = 0
+    cur_idx = im.tell()
+    try:
+        for ims in [im] + append_images:
+            # Get # of frames in this image
+            nfr = getattr(ims, "n_frames", 1)
+
+            for idx in range(nfr):
+                ims.seek(idx)
+
+                frame = _convert_frame(ims)
+
+                # Append the frame to the animation encoder
+                enc.add(
+                    frame.getim(),
+                    round(timestamp),
+                    lossless,
+                    quality,
+                    alpha_quality,
+                    method,
+                )
+
+                # Update timestamp and frame index
+                if isinstance(duration, (list, tuple)):
+                    timestamp += duration[frame_idx]
+                else:
+                    timestamp += duration
+                frame_idx += 1
+
+    finally:
+        im.seek(cur_idx)
+
+    # Force encoder to flush frames
+    enc.add(None, round(timestamp), lossless, quality, alpha_quality, 0)
+
+    # Get the final output from the encoder
+    data = enc.assemble(icc_profile, exif, xmp)
+    if data is None:
+        msg = "cannot write file as WebP (encoder returned None)"
+        raise OSError(msg)
+
+    fp.write(data)
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    lossless = im.encoderinfo.get("lossless", False)
+    quality = im.encoderinfo.get("quality", 80)
+    alpha_quality = im.encoderinfo.get("alpha_quality", 100)
+    icc_profile = im.encoderinfo.get("icc_profile") or ""
+    exif = im.encoderinfo.get("exif", b"")
+    if isinstance(exif, Image.Exif):
+        exif = exif.tobytes()
+    if exif.startswith(b"Exif\x00\x00"):
+        exif = exif[6:]
+    xmp = im.encoderinfo.get("xmp", "")
+    method = im.encoderinfo.get("method", 4)
+    exact = 1 if im.encoderinfo.get("exact") else 0
+
+    im = _convert_frame(im)
+
+    data = _webp.WebPEncode(
+        im.getim(),
+        lossless,
+        float(quality),
+        float(alpha_quality),
+        icc_profile,
+        method,
+        exact,
+        exif,
+        xmp,
+    )
+    if data is None:
+        msg = "cannot write file as WebP (encoder returned None)"
+        raise OSError(msg)
+
+    fp.write(data)
+
+
+Image.register_open(WebPImageFile.format, WebPImageFile, _accept)
+if SUPPORTED:
+    Image.register_save(WebPImageFile.format, _save)
+    Image.register_save_all(WebPImageFile.format, _save_all)
+    Image.register_extension(WebPImageFile.format, ".webp")
+    Image.register_mime(WebPImageFile.format, "image/webp")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/WmfImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/WmfImagePlugin.py
new file mode 100644
index 00000000..48e9823e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/WmfImagePlugin.py
@@ -0,0 +1,184 @@
+#
+# The Python Imaging Library
+# $Id$
+#
+# WMF stub codec
+#
+# history:
+# 1996-12-14 fl   Created
+# 2004-02-22 fl   Turned into a stub driver
+# 2004-02-23 fl   Added EMF support
+#
+# Copyright (c) Secret Labs AB 1997-2004.  All rights reserved.
+# Copyright (c) Fredrik Lundh 1996.
+#
+# See the README file for information on usage and redistribution.
+#
+# WMF/EMF reference documentation:
+# https://winprotocoldoc.blob.core.windows.net/productionwindowsarchives/MS-WMF/[MS-WMF].pdf
+# http://wvware.sourceforge.net/caolan/index.html
+# http://wvware.sourceforge.net/caolan/ora-wmf.html
+from __future__ import annotations
+
+from typing import IO
+
+from . import Image, ImageFile
+from ._binary import i16le as word
+from ._binary import si16le as short
+from ._binary import si32le as _long
+
+_handler = None
+
+
+def register_handler(handler: ImageFile.StubHandler | None) -> None:
+    """
+    Install application-specific WMF image handler.
+
+    :param handler: Handler object.
+    """
+    global _handler
+    _handler = handler
+
+
+if hasattr(Image.core, "drawwmf"):
+    # install default handler (windows only)
+
+    class WmfHandler(ImageFile.StubHandler):
+        def open(self, im: ImageFile.StubImageFile) -> None:
+            im._mode = "RGB"
+            self.bbox = im.info["wmf_bbox"]
+
+        def load(self, im: ImageFile.StubImageFile) -> Image.Image:
+            im.fp.seek(0)  # rewind
+            return Image.frombytes(
+                "RGB",
+                im.size,
+                Image.core.drawwmf(im.fp.read(), im.size, self.bbox),
+                "raw",
+                "BGR",
+                (im.size[0] * 3 + 3) & -4,
+                -1,
+            )
+
+    register_handler(WmfHandler())
+
+#
+# --------------------------------------------------------------------
+# Read WMF file
+
+
+def _accept(prefix: bytes) -> bool:
+    return (
+        prefix[:6] == b"\xd7\xcd\xc6\x9a\x00\x00" or prefix[:4] == b"\x01\x00\x00\x00"
+    )
+
+
+##
+# Image plugin for Windows metafiles.
+
+
+class WmfStubImageFile(ImageFile.StubImageFile):
+    format = "WMF"
+    format_description = "Windows Metafile"
+
+    def _open(self) -> None:
+        self._inch = None
+
+        # check placable header
+        s = self.fp.read(80)
+
+        if s[:6] == b"\xd7\xcd\xc6\x9a\x00\x00":
+            # placeable windows metafile
+
+            # get units per inch
+            self._inch = word(s, 14)
+            if self._inch == 0:
+                msg = "Invalid inch"
+                raise ValueError(msg)
+
+            # get bounding box
+            x0 = short(s, 6)
+            y0 = short(s, 8)
+            x1 = short(s, 10)
+            y1 = short(s, 12)
+
+            # normalize size to 72 dots per inch
+            self.info["dpi"] = 72
+            size = (
+                (x1 - x0) * self.info["dpi"] // self._inch,
+                (y1 - y0) * self.info["dpi"] // self._inch,
+            )
+
+            self.info["wmf_bbox"] = x0, y0, x1, y1
+
+            # sanity check (standard metafile header)
+            if s[22:26] != b"\x01\x00\t\x00":
+                msg = "Unsupported WMF file format"
+                raise SyntaxError(msg)
+
+        elif s[:4] == b"\x01\x00\x00\x00" and s[40:44] == b" EMF":
+            # enhanced metafile
+
+            # get bounding box
+            x0 = _long(s, 8)
+            y0 = _long(s, 12)
+            x1 = _long(s, 16)
+            y1 = _long(s, 20)
+
+            # get frame (in 0.01 millimeter units)
+            frame = _long(s, 24), _long(s, 28), _long(s, 32), _long(s, 36)
+
+            size = x1 - x0, y1 - y0
+
+            # calculate dots per inch from bbox and frame
+            xdpi = 2540.0 * (x1 - x0) / (frame[2] - frame[0])
+            ydpi = 2540.0 * (y1 - y0) / (frame[3] - frame[1])
+
+            self.info["wmf_bbox"] = x0, y0, x1, y1
+
+            if xdpi == ydpi:
+                self.info["dpi"] = xdpi
+            else:
+                self.info["dpi"] = xdpi, ydpi
+
+        else:
+            msg = "Unsupported file format"
+            raise SyntaxError(msg)
+
+        self._mode = "RGB"
+        self._size = size
+
+        loader = self._load()
+        if loader:
+            loader.open(self)
+
+    def _load(self) -> ImageFile.StubHandler | None:
+        return _handler
+
+    def load(self, dpi: int | None = None) -> Image.core.PixelAccess | None:
+        if dpi is not None and self._inch is not None:
+            self.info["dpi"] = dpi
+            x0, y0, x1, y1 = self.info["wmf_bbox"]
+            self._size = (
+                (x1 - x0) * self.info["dpi"] // self._inch,
+                (y1 - y0) * self.info["dpi"] // self._inch,
+            )
+        return super().load()
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if _handler is None or not hasattr(_handler, "save"):
+        msg = "WMF save handler not installed"
+        raise OSError(msg)
+    _handler.save(im, fp, filename)
+
+
+#
+# --------------------------------------------------------------------
+# Registry stuff
+
+
+Image.register_open(WmfStubImageFile.format, WmfStubImageFile, _accept)
+Image.register_save(WmfStubImageFile.format, _save)
+
+Image.register_extensions(WmfStubImageFile.format, [".wmf", ".emf"])
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/XVThumbImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/XVThumbImagePlugin.py
new file mode 100644
index 00000000..75333354
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/XVThumbImagePlugin.py
@@ -0,0 +1,83 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# XV Thumbnail file handler by Charles E. "Gene" Cash
+# (gcash@magicnet.net)
+#
+# see xvcolor.c and xvbrowse.c in the sources to John Bradley's XV,
+# available from ftp://ftp.cis.upenn.edu/pub/xv/
+#
+# history:
+# 98-08-15 cec  created (b/w only)
+# 98-12-09 cec  added color palette
+# 98-12-28 fl   added to PIL (with only a few very minor modifications)
+#
+# To do:
+# FIXME: make save work (this requires quantization support)
+#
+from __future__ import annotations
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import o8
+
+_MAGIC = b"P7 332"
+
+# standard color palette for thumbnails (RGB332)
+PALETTE = b""
+for r in range(8):
+    for g in range(8):
+        for b in range(4):
+            PALETTE = PALETTE + (
+                o8((r * 255) // 7) + o8((g * 255) // 7) + o8((b * 255) // 3)
+            )
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:6] == _MAGIC
+
+
+##
+# Image plugin for XV thumbnail images.
+
+
+class XVThumbImageFile(ImageFile.ImageFile):
+    format = "XVThumb"
+    format_description = "XV thumbnail image"
+
+    def _open(self) -> None:
+        # check magic
+        assert self.fp is not None
+
+        if not _accept(self.fp.read(6)):
+            msg = "not an XV thumbnail file"
+            raise SyntaxError(msg)
+
+        # Skip to beginning of next line
+        self.fp.readline()
+
+        # skip info comments
+        while True:
+            s = self.fp.readline()
+            if not s:
+                msg = "Unexpected EOF reading XV thumbnail file"
+                raise SyntaxError(msg)
+            if s[0] != 35:  # ie. when not a comment: '#'
+                break
+
+        # parse header line (already read)
+        s = s.strip().split()
+
+        self._mode = "P"
+        self._size = int(s[0]), int(s[1])
+
+        self.palette = ImagePalette.raw("RGB", PALETTE)
+
+        self.tile = [
+            ImageFile._Tile("raw", (0, 0) + self.size, self.fp.tell(), self.mode)
+        ]
+
+
+# --------------------------------------------------------------------
+
+Image.register_open(XVThumbImageFile.format, XVThumbImageFile, _accept)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/XbmImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/XbmImagePlugin.py
new file mode 100644
index 00000000..943a0447
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/XbmImagePlugin.py
@@ -0,0 +1,98 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# XBM File handling
+#
+# History:
+# 1995-09-08 fl   Created
+# 1996-11-01 fl   Added save support
+# 1997-07-07 fl   Made header parser more tolerant
+# 1997-07-22 fl   Fixed yet another parser bug
+# 2001-02-17 fl   Use 're' instead of 'regex' (Python 2.1) (0.4)
+# 2001-05-13 fl   Added hotspot handling (based on code from Bernhard Herzog)
+# 2004-02-24 fl   Allow some whitespace before first #define
+#
+# Copyright (c) 1997-2004 by Secret Labs AB
+# Copyright (c) 1996-1997 by Fredrik Lundh
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import re
+from typing import IO
+
+from . import Image, ImageFile
+
+# XBM header
+xbm_head = re.compile(
+    rb"\s*#define[ \t]+.*_width[ \t]+(?P<width>[0-9]+)[\r\n]+"
+    b"#define[ \t]+.*_height[ \t]+(?P<height>[0-9]+)[\r\n]+"
+    b"(?P<hotspot>"
+    b"#define[ \t]+[^_]*_x_hot[ \t]+(?P<xhot>[0-9]+)[\r\n]+"
+    b"#define[ \t]+[^_]*_y_hot[ \t]+(?P<yhot>[0-9]+)[\r\n]+"
+    b")?"
+    rb"[\000-\377]*_bits\[]"
+)
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix.lstrip()[:7] == b"#define"
+
+
+##
+# Image plugin for X11 bitmaps.
+
+
+class XbmImageFile(ImageFile.ImageFile):
+    format = "XBM"
+    format_description = "X11 Bitmap"
+
+    def _open(self) -> None:
+        assert self.fp is not None
+
+        m = xbm_head.match(self.fp.read(512))
+
+        if not m:
+            msg = "not a XBM file"
+            raise SyntaxError(msg)
+
+        xsize = int(m.group("width"))
+        ysize = int(m.group("height"))
+
+        if m.group("hotspot"):
+            self.info["hotspot"] = (int(m.group("xhot")), int(m.group("yhot")))
+
+        self._mode = "1"
+        self._size = xsize, ysize
+
+        self.tile = [ImageFile._Tile("xbm", (0, 0) + self.size, m.end())]
+
+
+def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
+    if im.mode != "1":
+        msg = f"cannot write mode {im.mode} as XBM"
+        raise OSError(msg)
+
+    fp.write(f"#define im_width {im.size[0]}\n".encode("ascii"))
+    fp.write(f"#define im_height {im.size[1]}\n".encode("ascii"))
+
+    hotspot = im.encoderinfo.get("hotspot")
+    if hotspot:
+        fp.write(f"#define im_x_hot {hotspot[0]}\n".encode("ascii"))
+        fp.write(f"#define im_y_hot {hotspot[1]}\n".encode("ascii"))
+
+    fp.write(b"static char im_bits[] = {\n")
+
+    ImageFile._save(im, fp, [ImageFile._Tile("xbm", (0, 0) + im.size)])
+
+    fp.write(b"};\n")
+
+
+Image.register_open(XbmImageFile.format, XbmImageFile, _accept)
+Image.register_save(XbmImageFile.format, _save)
+
+Image.register_extension(XbmImageFile.format, ".xbm")
+
+Image.register_mime(XbmImageFile.format, "image/xbm")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/XpmImagePlugin.py b/backend/venv/lib/python3.12/site-packages/PIL/XpmImagePlugin.py
new file mode 100644
index 00000000..b985aa5d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/XpmImagePlugin.py
@@ -0,0 +1,125 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# XPM File handling
+#
+# History:
+# 1996-12-29 fl   Created
+# 2001-02-17 fl   Use 're' instead of 'regex' (Python 2.1) (0.7)
+#
+# Copyright (c) Secret Labs AB 1997-2001.
+# Copyright (c) Fredrik Lundh 1996-2001.
+#
+# See the README file for information on usage and redistribution.
+#
+from __future__ import annotations
+
+import re
+
+from . import Image, ImageFile, ImagePalette
+from ._binary import o8
+
+# XPM header
+xpm_head = re.compile(b'"([0-9]*) ([0-9]*) ([0-9]*) ([0-9]*)')
+
+
+def _accept(prefix: bytes) -> bool:
+    return prefix[:9] == b"/* XPM */"
+
+
+##
+# Image plugin for X11 pixel maps.
+
+
+class XpmImageFile(ImageFile.ImageFile):
+    format = "XPM"
+    format_description = "X11 Pixel Map"
+
+    def _open(self) -> None:
+        if not _accept(self.fp.read(9)):
+            msg = "not an XPM file"
+            raise SyntaxError(msg)
+
+        # skip forward to next string
+        while True:
+            s = self.fp.readline()
+            if not s:
+                msg = "broken XPM file"
+                raise SyntaxError(msg)
+            m = xpm_head.match(s)
+            if m:
+                break
+
+        self._size = int(m.group(1)), int(m.group(2))
+
+        pal = int(m.group(3))
+        bpp = int(m.group(4))
+
+        if pal > 256 or bpp != 1:
+            msg = "cannot read this XPM file"
+            raise ValueError(msg)
+
+        #
+        # load palette description
+
+        palette = [b"\0\0\0"] * 256
+
+        for _ in range(pal):
+            s = self.fp.readline()
+            if s[-2:] == b"\r\n":
+                s = s[:-2]
+            elif s[-1:] in b"\r\n":
+                s = s[:-1]
+
+            c = s[1]
+            s = s[2:-2].split()
+
+            for i in range(0, len(s), 2):
+                if s[i] == b"c":
+                    # process colour key
+                    rgb = s[i + 1]
+                    if rgb == b"None":
+                        self.info["transparency"] = c
+                    elif rgb[:1] == b"#":
+                        # FIXME: handle colour names (see ImagePalette.py)
+                        rgb = int(rgb[1:], 16)
+                        palette[c] = (
+                            o8((rgb >> 16) & 255) + o8((rgb >> 8) & 255) + o8(rgb & 255)
+                        )
+                    else:
+                        # unknown colour
+                        msg = "cannot read this XPM file"
+                        raise ValueError(msg)
+                    break
+
+            else:
+                # missing colour key
+                msg = "cannot read this XPM file"
+                raise ValueError(msg)
+
+        self._mode = "P"
+        self.palette = ImagePalette.raw("RGB", b"".join(palette))
+
+        self.tile = [ImageFile._Tile("raw", (0, 0) + self.size, self.fp.tell(), "P")]
+
+    def load_read(self, read_bytes: int) -> bytes:
+        #
+        # load all image data in one chunk
+
+        xsize, ysize = self.size
+
+        s = [self.fp.readline()[1 : xsize + 1].ljust(xsize) for i in range(ysize)]
+
+        return b"".join(s)
+
+
+#
+# Registry
+
+
+Image.register_open(XpmImageFile.format, XpmImageFile, _accept)
+
+Image.register_extension(XpmImageFile.format, ".xpm")
+
+Image.register_mime(XpmImageFile.format, "image/xpm")
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__init__.py b/backend/venv/lib/python3.12/site-packages/PIL/__init__.py
new file mode 100644
index 00000000..09546fe6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/__init__.py
@@ -0,0 +1,86 @@
+"""Pillow (Fork of the Python Imaging Library)
+
+Pillow is the friendly PIL fork by Jeffrey A. Clark and contributors.
+    https://github.com/python-pillow/Pillow/
+
+Pillow is forked from PIL 1.1.7.
+
+PIL is the Python Imaging Library by Fredrik Lundh and contributors.
+Copyright (c) 1999 by Secret Labs AB.
+
+Use PIL.__version__ for this Pillow version.
+
+;-)
+"""
+
+from __future__ import annotations
+
+from . import _version
+
+# VERSION was removed in Pillow 6.0.0.
+# PILLOW_VERSION was removed in Pillow 9.0.0.
+# Use __version__ instead.
+__version__ = _version.__version__
+del _version
+
+
+_plugins = [
+    "BlpImagePlugin",
+    "BmpImagePlugin",
+    "BufrStubImagePlugin",
+    "CurImagePlugin",
+    "DcxImagePlugin",
+    "DdsImagePlugin",
+    "EpsImagePlugin",
+    "FitsImagePlugin",
+    "FliImagePlugin",
+    "FpxImagePlugin",
+    "FtexImagePlugin",
+    "GbrImagePlugin",
+    "GifImagePlugin",
+    "GribStubImagePlugin",
+    "Hdf5StubImagePlugin",
+    "IcnsImagePlugin",
+    "IcoImagePlugin",
+    "ImImagePlugin",
+    "ImtImagePlugin",
+    "IptcImagePlugin",
+    "JpegImagePlugin",
+    "Jpeg2KImagePlugin",
+    "McIdasImagePlugin",
+    "MicImagePlugin",
+    "MpegImagePlugin",
+    "MpoImagePlugin",
+    "MspImagePlugin",
+    "PalmImagePlugin",
+    "PcdImagePlugin",
+    "PcxImagePlugin",
+    "PdfImagePlugin",
+    "PixarImagePlugin",
+    "PngImagePlugin",
+    "PpmImagePlugin",
+    "PsdImagePlugin",
+    "QoiImagePlugin",
+    "SgiImagePlugin",
+    "SpiderImagePlugin",
+    "SunImagePlugin",
+    "TgaImagePlugin",
+    "TiffImagePlugin",
+    "WebPImagePlugin",
+    "WmfImagePlugin",
+    "XbmImagePlugin",
+    "XpmImagePlugin",
+    "XVThumbImagePlugin",
+]
+
+
+class UnidentifiedImageError(OSError):
+    """
+    Raised in :py:meth:`PIL.Image.open` if an image cannot be opened and identified.
+
+    If a PNG image raises this error, setting :data:`.ImageFile.LOAD_TRUNCATED_IMAGES`
+    to true may allow the image to be opened after all. The setting will ignore missing
+    data and checksum failures.
+    """
+
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__main__.py b/backend/venv/lib/python3.12/site-packages/PIL/__main__.py
new file mode 100644
index 00000000..043156e8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/__main__.py
@@ -0,0 +1,7 @@
+from __future__ import annotations
+
+import sys
+
+from .features import pilinfo
+
+pilinfo(supported_formats="--report" not in sys.argv)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BdfFontFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BdfFontFile.cpython-312.pyc
new file mode 100644
index 00000000..4ea7d00e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BdfFontFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BlpImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BlpImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..c398b28d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BlpImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BmpImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BmpImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..5ced33fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BmpImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BufrStubImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BufrStubImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..15131ef0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/BufrStubImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ContainerIO.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ContainerIO.cpython-312.pyc
new file mode 100644
index 00000000..570aef5f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ContainerIO.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/CurImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/CurImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..d8d2a2c5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/CurImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DcxImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DcxImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..60055dfe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DcxImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DdsImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DdsImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..ba9b8349
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/DdsImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/EpsImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/EpsImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..fcc93cca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/EpsImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ExifTags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ExifTags.cpython-312.pyc
new file mode 100644
index 00000000..fef43dcf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ExifTags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FitsImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FitsImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..e0454fc7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FitsImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FliImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FliImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..27195c24
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FliImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FontFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FontFile.cpython-312.pyc
new file mode 100644
index 00000000..6e8ce70f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FontFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FpxImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FpxImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..77adeee8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FpxImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FtexImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FtexImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..c9e00141
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/FtexImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GbrImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GbrImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..e795020a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GbrImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GdImageFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GdImageFile.cpython-312.pyc
new file mode 100644
index 00000000..a61d4514
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GdImageFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GifImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GifImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..619389c9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GifImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpGradientFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpGradientFile.cpython-312.pyc
new file mode 100644
index 00000000..7582294f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpGradientFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpPaletteFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpPaletteFile.cpython-312.pyc
new file mode 100644
index 00000000..d6488111
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GimpPaletteFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GribStubImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GribStubImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..0bdff6fd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/GribStubImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Hdf5StubImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Hdf5StubImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..f9d59727
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Hdf5StubImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcnsImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcnsImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..ffca85bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcnsImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcoImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcoImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..2d77642c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IcoImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..da69525a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Image.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Image.cpython-312.pyc
new file mode 100644
index 00000000..b6726b26
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Image.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageChops.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageChops.cpython-312.pyc
new file mode 100644
index 00000000..45f19b7b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageChops.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageCms.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageCms.cpython-312.pyc
new file mode 100644
index 00000000..7eb5fcb6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageCms.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageColor.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageColor.cpython-312.pyc
new file mode 100644
index 00000000..ca37694d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageColor.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw.cpython-312.pyc
new file mode 100644
index 00000000..449b432a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw2.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw2.cpython-312.pyc
new file mode 100644
index 00000000..f5b3ad4a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageDraw2.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageEnhance.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageEnhance.cpython-312.pyc
new file mode 100644
index 00000000..35ec318b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageEnhance.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFile.cpython-312.pyc
new file mode 100644
index 00000000..0a84f239
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFilter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFilter.cpython-312.pyc
new file mode 100644
index 00000000..a56bb1a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFilter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFont.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFont.cpython-312.pyc
new file mode 100644
index 00000000..a2fe98c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageFont.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageGrab.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageGrab.cpython-312.pyc
new file mode 100644
index 00000000..c6ee01f6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageGrab.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMath.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMath.cpython-312.pyc
new file mode 100644
index 00000000..179d68e0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMath.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMode.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMode.cpython-312.pyc
new file mode 100644
index 00000000..f8286f1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMode.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMorph.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMorph.cpython-312.pyc
new file mode 100644
index 00000000..a5543ad4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageMorph.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageOps.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageOps.cpython-312.pyc
new file mode 100644
index 00000000..7fab246c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageOps.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePalette.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePalette.cpython-312.pyc
new file mode 100644
index 00000000..65706c91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePalette.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePath.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePath.cpython-312.pyc
new file mode 100644
index 00000000..4467ee98
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImagePath.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageQt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageQt.cpython-312.pyc
new file mode 100644
index 00000000..fd05d17e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageQt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageSequence.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageSequence.cpython-312.pyc
new file mode 100644
index 00000000..ec02e9f9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageSequence.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageShow.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageShow.cpython-312.pyc
new file mode 100644
index 00000000..bff9b687
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageShow.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageStat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageStat.cpython-312.pyc
new file mode 100644
index 00000000..363643a4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageStat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTk.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTk.cpython-312.pyc
new file mode 100644
index 00000000..905cdab3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTk.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTransform.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTransform.cpython-312.pyc
new file mode 100644
index 00000000..79653909
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageTransform.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageWin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageWin.cpython-312.pyc
new file mode 100644
index 00000000..f3774089
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImageWin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImtImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImtImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..deb98c99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/ImtImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IptcImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IptcImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..9cd0e807
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/IptcImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Jpeg2KImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Jpeg2KImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..264f14a2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/Jpeg2KImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..fc593424
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegPresets.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegPresets.cpython-312.pyc
new file mode 100644
index 00000000..05a6006e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/JpegPresets.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/McIdasImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/McIdasImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..e6e61ce0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/McIdasImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MicImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MicImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..cc947137
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MicImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpegImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpegImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..8d4df346
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpegImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpoImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpoImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..d6e518ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MpoImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MspImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MspImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..849ebf57
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/MspImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PSDraw.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PSDraw.cpython-312.pyc
new file mode 100644
index 00000000..9052e6e2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PSDraw.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PaletteFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PaletteFile.cpython-312.pyc
new file mode 100644
index 00000000..2c91c24e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PaletteFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PalmImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PalmImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..89e051c4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PalmImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcdImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcdImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..2fb261a4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcdImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcfFontFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcfFontFile.cpython-312.pyc
new file mode 100644
index 00000000..da500e70
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcfFontFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcxImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcxImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..9c46fc08
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PcxImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..c60ed07d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfParser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfParser.cpython-312.pyc
new file mode 100644
index 00000000..2f517818
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PdfParser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PixarImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PixarImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..871cb83b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PixarImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PngImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PngImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..54efdb94
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PngImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PpmImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PpmImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..a397852a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PpmImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PsdImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PsdImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..d29af328
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/PsdImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/QoiImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/QoiImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..b747c824
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/QoiImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SgiImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SgiImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..60f63816
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SgiImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SpiderImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SpiderImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..55a1bf68
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SpiderImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SunImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SunImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..c6b6b971
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/SunImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TarIO.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TarIO.cpython-312.pyc
new file mode 100644
index 00000000..eb2c3ec4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TarIO.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TgaImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TgaImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..aa221fe1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TgaImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..8f84d8a8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffTags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffTags.cpython-312.pyc
new file mode 100644
index 00000000..fcb985c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/TiffTags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WalImageFile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WalImageFile.cpython-312.pyc
new file mode 100644
index 00000000..961ec785
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WalImageFile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WebPImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WebPImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..d38dae4c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WebPImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WmfImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WmfImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..4bb9461f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/WmfImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XVThumbImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XVThumbImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..4a8decca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XVThumbImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XbmImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XbmImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..b928021a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XbmImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XpmImagePlugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XpmImagePlugin.cpython-312.pyc
new file mode 100644
index 00000000..12addd5b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/XpmImagePlugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..130368e7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..8264d2a9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_binary.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_binary.cpython-312.pyc
new file mode 100644
index 00000000..d11db440
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_binary.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_deprecate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_deprecate.cpython-312.pyc
new file mode 100644
index 00000000..0882d08d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_deprecate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_tkinter_finder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_tkinter_finder.cpython-312.pyc
new file mode 100644
index 00000000..c287ad65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_tkinter_finder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_typing.cpython-312.pyc
new file mode 100644
index 00000000..93d0769f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_util.cpython-312.pyc
new file mode 100644
index 00000000..c65b101d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..4c8dce6a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/features.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/features.cpython-312.pyc
new file mode 100644
index 00000000..ac51c650
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/features.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/report.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/report.cpython-312.pyc
new file mode 100644
index 00000000..fa16e003
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/__pycache__/report.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_binary.py b/backend/venv/lib/python3.12/site-packages/PIL/_binary.py
new file mode 100644
index 00000000..4594ccce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_binary.py
@@ -0,0 +1,112 @@
+#
+# The Python Imaging Library.
+# $Id$
+#
+# Binary input/output support routines.
+#
+# Copyright (c) 1997-2003 by Secret Labs AB
+# Copyright (c) 1995-2003 by Fredrik Lundh
+# Copyright (c) 2012 by Brian Crowell
+#
+# See the README file for information on usage and redistribution.
+#
+
+
+"""Binary input/output support routines."""
+from __future__ import annotations
+
+from struct import pack, unpack_from
+
+
+def i8(c: bytes) -> int:
+    return c[0]
+
+
+def o8(i: int) -> bytes:
+    return bytes((i & 255,))
+
+
+# Input, le = little endian, be = big endian
+def i16le(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 2-bytes (16 bits) string to an unsigned integer.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from("<H", c, o)[0]
+
+
+def si16le(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 2-bytes (16 bits) string to a signed integer.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from("<h", c, o)[0]
+
+
+def si16be(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 2-bytes (16 bits) string to a signed integer, big endian.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from(">h", c, o)[0]
+
+
+def i32le(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 4-bytes (32 bits) string to an unsigned integer.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from("<I", c, o)[0]
+
+
+def si32le(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 4-bytes (32 bits) string to a signed integer.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from("<i", c, o)[0]
+
+
+def si32be(c: bytes, o: int = 0) -> int:
+    """
+    Converts a 4-bytes (32 bits) string to a signed integer, big endian.
+
+    :param c: string containing bytes to convert
+    :param o: offset of bytes to convert in string
+    """
+    return unpack_from(">i", c, o)[0]
+
+
+def i16be(c: bytes, o: int = 0) -> int:
+    return unpack_from(">H", c, o)[0]
+
+
+def i32be(c: bytes, o: int = 0) -> int:
+    return unpack_from(">I", c, o)[0]
+
+
+# Output, le = little endian, be = big endian
+def o16le(i: int) -> bytes:
+    return pack("<H", i)
+
+
+def o32le(i: int) -> bytes:
+    return pack("<I", i)
+
+
+def o16be(i: int) -> bytes:
+    return pack(">H", i)
+
+
+def o32be(i: int) -> bytes:
+    return pack(">I", i)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_deprecate.py b/backend/venv/lib/python3.12/site-packages/PIL/_deprecate.py
new file mode 100644
index 00000000..83952b39
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_deprecate.py
@@ -0,0 +1,69 @@
+from __future__ import annotations
+
+import warnings
+
+from . import __version__
+
+
+def deprecate(
+    deprecated: str,
+    when: int | None,
+    replacement: str | None = None,
+    *,
+    action: str | None = None,
+    plural: bool = False,
+) -> None:
+    """
+    Deprecations helper.
+
+    :param deprecated: Name of thing to be deprecated.
+    :param when: Pillow major version to be removed in.
+    :param replacement: Name of replacement.
+    :param action: Instead of "replacement", give a custom call to action
+        e.g. "Upgrade to new thing".
+    :param plural: if the deprecated thing is plural, needing "are" instead of "is".
+
+    Usually of the form:
+
+        "[deprecated] is deprecated and will be removed in Pillow [when] (yyyy-mm-dd).
+        Use [replacement] instead."
+
+    You can leave out the replacement sentence:
+
+        "[deprecated] is deprecated and will be removed in Pillow [when] (yyyy-mm-dd)"
+
+    Or with another call to action:
+
+        "[deprecated] is deprecated and will be removed in Pillow [when] (yyyy-mm-dd).
+        [action]."
+    """
+
+    is_ = "are" if plural else "is"
+
+    if when is None:
+        removed = "a future version"
+    elif when <= int(__version__.split(".")[0]):
+        msg = f"{deprecated} {is_} deprecated and should be removed."
+        raise RuntimeError(msg)
+    elif when == 12:
+        removed = "Pillow 12 (2025-10-15)"
+    else:
+        msg = f"Unknown removal version: {when}. Update {__name__}?"
+        raise ValueError(msg)
+
+    if replacement and action:
+        msg = "Use only one of 'replacement' and 'action'"
+        raise ValueError(msg)
+
+    if replacement:
+        action = f". Use {replacement} instead."
+    elif action:
+        action = f". {action.rstrip('.')}."
+    else:
+        action = ""
+
+    warnings.warn(
+        f"{deprecated} {is_} deprecated and will be removed in {removed}{action}",
+        DeprecationWarning,
+        stacklevel=3,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imaging.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imaging.cpython-312-darwin.so
new file mode 100755
index 00000000..65b74f0b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imaging.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imaging.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imaging.pyi
new file mode 100644
index 00000000..998bc52e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imaging.pyi
@@ -0,0 +1,31 @@
+from typing import Any
+
+class ImagingCore:
+    def __getitem__(self, index: int) -> float: ...
+    def __getattr__(self, name: str) -> Any: ...
+
+class ImagingFont:
+    def __getattr__(self, name: str) -> Any: ...
+
+class ImagingDraw:
+    def __getattr__(self, name: str) -> Any: ...
+
+class PixelAccess:
+    def __getitem__(self, xy: tuple[int, int]) -> float | tuple[int, ...]: ...
+    def __setitem__(
+        self, xy: tuple[int, int], color: float | tuple[int, ...]
+    ) -> None: ...
+
+class ImagingDecoder:
+    def __getattr__(self, name: str) -> Any: ...
+
+class ImagingEncoder:
+    def __getattr__(self, name: str) -> Any: ...
+
+class _Outline:
+    def close(self) -> None: ...
+    def __getattr__(self, name: str) -> Any: ...
+
+def font(image: ImagingCore, glyphdata: bytes) -> ImagingFont: ...
+def outline() -> _Outline: ...
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.cpython-312-darwin.so
new file mode 100755
index 00000000..677397f3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.pyi
new file mode 100644
index 00000000..ddcf93ab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imagingcms.pyi
@@ -0,0 +1,143 @@
+import datetime
+import sys
+from typing import Literal, SupportsFloat, TypedDict
+
+from ._typing import CapsuleType
+
+littlecms_version: str | None
+
+_Tuple3f = tuple[float, float, float]
+_Tuple2x3f = tuple[_Tuple3f, _Tuple3f]
+_Tuple3x3f = tuple[_Tuple3f, _Tuple3f, _Tuple3f]
+
+class _IccMeasurementCondition(TypedDict):
+    observer: int
+    backing: _Tuple3f
+    geo: str
+    flare: float
+    illuminant_type: str
+
+class _IccViewingCondition(TypedDict):
+    illuminant: _Tuple3f
+    surround: _Tuple3f
+    illuminant_type: str
+
+class CmsProfile:
+    @property
+    def rendering_intent(self) -> int: ...
+    @property
+    def creation_date(self) -> datetime.datetime | None: ...
+    @property
+    def copyright(self) -> str | None: ...
+    @property
+    def target(self) -> str | None: ...
+    @property
+    def manufacturer(self) -> str | None: ...
+    @property
+    def model(self) -> str | None: ...
+    @property
+    def profile_description(self) -> str | None: ...
+    @property
+    def screening_description(self) -> str | None: ...
+    @property
+    def viewing_condition(self) -> str | None: ...
+    @property
+    def version(self) -> float: ...
+    @property
+    def icc_version(self) -> int: ...
+    @property
+    def attributes(self) -> int: ...
+    @property
+    def header_flags(self) -> int: ...
+    @property
+    def header_manufacturer(self) -> str: ...
+    @property
+    def header_model(self) -> str: ...
+    @property
+    def device_class(self) -> str: ...
+    @property
+    def connection_space(self) -> str: ...
+    @property
+    def xcolor_space(self) -> str: ...
+    @property
+    def profile_id(self) -> bytes: ...
+    @property
+    def is_matrix_shaper(self) -> bool: ...
+    @property
+    def technology(self) -> str | None: ...
+    @property
+    def colorimetric_intent(self) -> str | None: ...
+    @property
+    def perceptual_rendering_intent_gamut(self) -> str | None: ...
+    @property
+    def saturation_rendering_intent_gamut(self) -> str | None: ...
+    @property
+    def red_colorant(self) -> _Tuple2x3f | None: ...
+    @property
+    def green_colorant(self) -> _Tuple2x3f | None: ...
+    @property
+    def blue_colorant(self) -> _Tuple2x3f | None: ...
+    @property
+    def red_primary(self) -> _Tuple2x3f | None: ...
+    @property
+    def green_primary(self) -> _Tuple2x3f | None: ...
+    @property
+    def blue_primary(self) -> _Tuple2x3f | None: ...
+    @property
+    def media_white_point_temperature(self) -> float | None: ...
+    @property
+    def media_white_point(self) -> _Tuple2x3f | None: ...
+    @property
+    def media_black_point(self) -> _Tuple2x3f | None: ...
+    @property
+    def luminance(self) -> _Tuple2x3f | None: ...
+    @property
+    def chromatic_adaptation(self) -> tuple[_Tuple3x3f, _Tuple3x3f] | None: ...
+    @property
+    def chromaticity(self) -> _Tuple3x3f | None: ...
+    @property
+    def colorant_table(self) -> list[str] | None: ...
+    @property
+    def colorant_table_out(self) -> list[str] | None: ...
+    @property
+    def intent_supported(self) -> dict[int, tuple[bool, bool, bool]] | None: ...
+    @property
+    def clut(self) -> dict[int, tuple[bool, bool, bool]] | None: ...
+    @property
+    def icc_measurement_condition(self) -> _IccMeasurementCondition | None: ...
+    @property
+    def icc_viewing_condition(self) -> _IccViewingCondition | None: ...
+    def is_intent_supported(self, intent: int, direction: int, /) -> int: ...
+
+class CmsTransform:
+    def apply(self, id_in: CapsuleType, id_out: CapsuleType) -> int: ...
+
+def profile_open(profile: str, /) -> CmsProfile: ...
+def profile_frombytes(profile: bytes, /) -> CmsProfile: ...
+def profile_tobytes(profile: CmsProfile, /) -> bytes: ...
+def buildTransform(
+    input_profile: CmsProfile,
+    output_profile: CmsProfile,
+    in_mode: str,
+    out_mode: str,
+    rendering_intent: int = 0,
+    cms_flags: int = 0,
+    /,
+) -> CmsTransform: ...
+def buildProofTransform(
+    input_profile: CmsProfile,
+    output_profile: CmsProfile,
+    proof_profile: CmsProfile,
+    in_mode: str,
+    out_mode: str,
+    rendering_intent: int = 0,
+    proof_intent: int = 0,
+    cms_flags: int = 0,
+    /,
+) -> CmsTransform: ...
+def createProfile(
+    color_space: Literal["LAB", "XYZ", "sRGB"], color_temp: SupportsFloat = 0.0, /
+) -> CmsProfile: ...
+
+if sys.platform == "win32":
+    def get_display_profile_win32(handle: int = 0, is_dc: int = 0, /) -> str | None: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.cpython-312-darwin.so
new file mode 100755
index 00000000..27178b0d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.pyi
new file mode 100644
index 00000000..9cc9822f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imagingft.pyi
@@ -0,0 +1,69 @@
+from typing import Any, Callable
+
+from . import ImageFont, _imaging
+
+class Font:
+    @property
+    def family(self) -> str | None: ...
+    @property
+    def style(self) -> str | None: ...
+    @property
+    def ascent(self) -> int: ...
+    @property
+    def descent(self) -> int: ...
+    @property
+    def height(self) -> int: ...
+    @property
+    def x_ppem(self) -> int: ...
+    @property
+    def y_ppem(self) -> int: ...
+    @property
+    def glyphs(self) -> int: ...
+    def render(
+        self,
+        string: str | bytes,
+        fill: Callable[[int, int], _imaging.ImagingCore],
+        mode: str,
+        dir: str | None,
+        features: list[str] | None,
+        lang: str | None,
+        stroke_width: float,
+        anchor: str | None,
+        foreground_ink_long: int,
+        x_start: float,
+        y_start: float,
+        /,
+    ) -> tuple[_imaging.ImagingCore, tuple[int, int]]: ...
+    def getsize(
+        self,
+        string: str | bytes | bytearray,
+        mode: str,
+        dir: str | None,
+        features: list[str] | None,
+        lang: str | None,
+        anchor: str | None,
+        /,
+    ) -> tuple[tuple[int, int], tuple[int, int]]: ...
+    def getlength(
+        self,
+        string: str | bytes,
+        mode: str,
+        dir: str | None,
+        features: list[str] | None,
+        lang: str | None,
+        /,
+    ) -> float: ...
+    def getvarnames(self) -> list[bytes]: ...
+    def getvaraxes(self) -> list[ImageFont.Axis]: ...
+    def setvarname(self, instance_index: int, /) -> None: ...
+    def setvaraxes(self, axes: list[float], /) -> None: ...
+
+def getfont(
+    filename: str | bytes,
+    size: float,
+    index: int,
+    encoding: str,
+    font_bytes: bytes,
+    layout_engine: int,
+) -> Font: ...
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.cpython-312-darwin.so
new file mode 100755
index 00000000..9172f32f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.pyi
new file mode 100644
index 00000000..e27843e5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmath.pyi
@@ -0,0 +1,3 @@
+from typing import Any
+
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.cpython-312-darwin.so
new file mode 100755
index 00000000..2bf9502f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.pyi
new file mode 100644
index 00000000..e27843e5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imagingmorph.pyi
@@ -0,0 +1,3 @@
+from typing import Any
+
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.cpython-312-darwin.so
new file mode 100755
index 00000000..e85a78e0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.pyi
new file mode 100644
index 00000000..e27843e5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_imagingtk.pyi
@@ -0,0 +1,3 @@
+from typing import Any
+
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_tkinter_finder.py b/backend/venv/lib/python3.12/site-packages/PIL/_tkinter_finder.py
new file mode 100644
index 00000000..beddfb06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_tkinter_finder.py
@@ -0,0 +1,21 @@
+""" Find compiled module linking to Tcl / Tk libraries
+"""
+
+from __future__ import annotations
+
+import sys
+import tkinter
+
+tk = getattr(tkinter, "_tkinter")
+
+try:
+    if hasattr(sys, "pypy_find_executable"):
+        TKINTER_LIB = tk.tklib_cffi.__file__
+    else:
+        TKINTER_LIB = tk.__file__
+except AttributeError:
+    # _tkinter may be compiled directly into Python, in which case __file__ is
+    # not available. load_tkinter_funcs will check the binary first in any case.
+    TKINTER_LIB = None
+
+tk_version = str(tkinter.TkVersion)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_typing.py b/backend/venv/lib/python3.12/site-packages/PIL/_typing.py
new file mode 100644
index 00000000..34a9a81e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_typing.py
@@ -0,0 +1,53 @@
+from __future__ import annotations
+
+import os
+import sys
+from collections.abc import Sequence
+from typing import TYPE_CHECKING, Any, Protocol, TypeVar, Union
+
+if TYPE_CHECKING:
+    from numbers import _IntegralLike as IntegralLike
+
+    try:
+        import numpy.typing as npt
+
+        NumpyArray = npt.NDArray[Any]  # requires numpy>=1.21
+    except (ImportError, AttributeError):
+        pass
+
+if sys.version_info >= (3, 13):
+    from types import CapsuleType
+else:
+    CapsuleType = object
+
+if sys.version_info >= (3, 12):
+    from collections.abc import Buffer
+else:
+    Buffer = Any
+
+if sys.version_info >= (3, 10):
+    from typing import TypeGuard
+else:
+    try:
+        from typing_extensions import TypeGuard
+    except ImportError:
+
+        class TypeGuard:  # type: ignore[no-redef]
+            def __class_getitem__(cls, item: Any) -> type[bool]:
+                return bool
+
+
+Coords = Union[Sequence[float], Sequence[Sequence[float]]]
+
+
+_T_co = TypeVar("_T_co", covariant=True)
+
+
+class SupportsRead(Protocol[_T_co]):
+    def read(self, length: int = ..., /) -> _T_co: ...
+
+
+StrOrBytesPath = Union[str, bytes, os.PathLike[str], os.PathLike[bytes]]
+
+
+__all__ = ["Buffer", "IntegralLike", "StrOrBytesPath", "SupportsRead", "TypeGuard"]
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_util.py b/backend/venv/lib/python3.12/site-packages/PIL/_util.py
new file mode 100644
index 00000000..8ef0d36f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_util.py
@@ -0,0 +1,26 @@
+from __future__ import annotations
+
+import os
+from typing import Any, NoReturn
+
+from ._typing import StrOrBytesPath, TypeGuard
+
+
+def is_path(f: Any) -> TypeGuard[StrOrBytesPath]:
+    return isinstance(f, (bytes, str, os.PathLike))
+
+
+class DeferredError:
+    def __init__(self, ex: BaseException):
+        self.ex = ex
+
+    def __getattr__(self, elt: str) -> NoReturn:
+        raise self.ex
+
+    @staticmethod
+    def new(ex: BaseException) -> Any:
+        """
+        Creates an object that raises the wrapped exception ``ex`` when used,
+        and casts it to :py:obj:`~typing.Any` type.
+        """
+        return DeferredError(ex)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_version.py b/backend/venv/lib/python3.12/site-packages/PIL/_version.py
new file mode 100644
index 00000000..9938a0af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_version.py
@@ -0,0 +1,4 @@
+# Master version for Pillow
+from __future__ import annotations
+
+__version__ = "11.1.0"
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_webp.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/PIL/_webp.cpython-312-darwin.so
new file mode 100755
index 00000000..3a1aca80
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/PIL/_webp.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/_webp.pyi b/backend/venv/lib/python3.12/site-packages/PIL/_webp.pyi
new file mode 100644
index 00000000..e27843e5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/_webp.pyi
@@ -0,0 +1,3 @@
+from typing import Any
+
+def __getattr__(name: str) -> Any: ...
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/features.py b/backend/venv/lib/python3.12/site-packages/PIL/features.py
new file mode 100644
index 00000000..3645e3de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/features.py
@@ -0,0 +1,357 @@
+from __future__ import annotations
+
+import collections
+import os
+import sys
+import warnings
+from typing import IO
+
+import PIL
+
+from . import Image
+from ._deprecate import deprecate
+
+modules = {
+    "pil": ("PIL._imaging", "PILLOW_VERSION"),
+    "tkinter": ("PIL._tkinter_finder", "tk_version"),
+    "freetype2": ("PIL._imagingft", "freetype2_version"),
+    "littlecms2": ("PIL._imagingcms", "littlecms_version"),
+    "webp": ("PIL._webp", "webpdecoder_version"),
+}
+
+
+def check_module(feature: str) -> bool:
+    """
+    Checks if a module is available.
+
+    :param feature: The module to check for.
+    :returns: ``True`` if available, ``False`` otherwise.
+    :raises ValueError: If the module is not defined in this version of Pillow.
+    """
+    if feature not in modules:
+        msg = f"Unknown module {feature}"
+        raise ValueError(msg)
+
+    module, ver = modules[feature]
+
+    try:
+        __import__(module)
+        return True
+    except ModuleNotFoundError:
+        return False
+    except ImportError as ex:
+        warnings.warn(str(ex))
+        return False
+
+
+def version_module(feature: str) -> str | None:
+    """
+    :param feature: The module to check for.
+    :returns:
+        The loaded version number as a string, or ``None`` if unknown or not available.
+    :raises ValueError: If the module is not defined in this version of Pillow.
+    """
+    if not check_module(feature):
+        return None
+
+    module, ver = modules[feature]
+
+    return getattr(__import__(module, fromlist=[ver]), ver)
+
+
+def get_supported_modules() -> list[str]:
+    """
+    :returns: A list of all supported modules.
+    """
+    return [f for f in modules if check_module(f)]
+
+
+codecs = {
+    "jpg": ("jpeg", "jpeglib"),
+    "jpg_2000": ("jpeg2k", "jp2klib"),
+    "zlib": ("zip", "zlib"),
+    "libtiff": ("libtiff", "libtiff"),
+}
+
+
+def check_codec(feature: str) -> bool:
+    """
+    Checks if a codec is available.
+
+    :param feature: The codec to check for.
+    :returns: ``True`` if available, ``False`` otherwise.
+    :raises ValueError: If the codec is not defined in this version of Pillow.
+    """
+    if feature not in codecs:
+        msg = f"Unknown codec {feature}"
+        raise ValueError(msg)
+
+    codec, lib = codecs[feature]
+
+    return f"{codec}_encoder" in dir(Image.core)
+
+
+def version_codec(feature: str) -> str | None:
+    """
+    :param feature: The codec to check for.
+    :returns:
+        The version number as a string, or ``None`` if not available.
+        Checked at compile time for ``jpg``, run-time otherwise.
+    :raises ValueError: If the codec is not defined in this version of Pillow.
+    """
+    if not check_codec(feature):
+        return None
+
+    codec, lib = codecs[feature]
+
+    version = getattr(Image.core, f"{lib}_version")
+
+    if feature == "libtiff":
+        return version.split("\n")[0].split("Version ")[1]
+
+    return version
+
+
+def get_supported_codecs() -> list[str]:
+    """
+    :returns: A list of all supported codecs.
+    """
+    return [f for f in codecs if check_codec(f)]
+
+
+features: dict[str, tuple[str, str | bool, str | None]] = {
+    "webp_anim": ("PIL._webp", True, None),
+    "webp_mux": ("PIL._webp", True, None),
+    "transp_webp": ("PIL._webp", True, None),
+    "raqm": ("PIL._imagingft", "HAVE_RAQM", "raqm_version"),
+    "fribidi": ("PIL._imagingft", "HAVE_FRIBIDI", "fribidi_version"),
+    "harfbuzz": ("PIL._imagingft", "HAVE_HARFBUZZ", "harfbuzz_version"),
+    "libjpeg_turbo": ("PIL._imaging", "HAVE_LIBJPEGTURBO", "libjpeg_turbo_version"),
+    "zlib_ng": ("PIL._imaging", "HAVE_ZLIBNG", "zlib_ng_version"),
+    "libimagequant": ("PIL._imaging", "HAVE_LIBIMAGEQUANT", "imagequant_version"),
+    "xcb": ("PIL._imaging", "HAVE_XCB", None),
+}
+
+
+def check_feature(feature: str) -> bool | None:
+    """
+    Checks if a feature is available.
+
+    :param feature: The feature to check for.
+    :returns: ``True`` if available, ``False`` if unavailable, ``None`` if unknown.
+    :raises ValueError: If the feature is not defined in this version of Pillow.
+    """
+    if feature not in features:
+        msg = f"Unknown feature {feature}"
+        raise ValueError(msg)
+
+    module, flag, ver = features[feature]
+
+    if isinstance(flag, bool):
+        deprecate(f'check_feature("{feature}")', 12)
+    try:
+        imported_module = __import__(module, fromlist=["PIL"])
+        if isinstance(flag, bool):
+            return flag
+        return getattr(imported_module, flag)
+    except ModuleNotFoundError:
+        return None
+    except ImportError as ex:
+        warnings.warn(str(ex))
+        return None
+
+
+def version_feature(feature: str) -> str | None:
+    """
+    :param feature: The feature to check for.
+    :returns: The version number as a string, or ``None`` if not available.
+    :raises ValueError: If the feature is not defined in this version of Pillow.
+    """
+    if not check_feature(feature):
+        return None
+
+    module, flag, ver = features[feature]
+
+    if ver is None:
+        return None
+
+    return getattr(__import__(module, fromlist=[ver]), ver)
+
+
+def get_supported_features() -> list[str]:
+    """
+    :returns: A list of all supported features.
+    """
+    supported_features = []
+    for f, (module, flag, _) in features.items():
+        if flag is True:
+            for feature, (feature_module, _) in modules.items():
+                if feature_module == module:
+                    if check_module(feature):
+                        supported_features.append(f)
+                    break
+        elif check_feature(f):
+            supported_features.append(f)
+    return supported_features
+
+
+def check(feature: str) -> bool | None:
+    """
+    :param feature: A module, codec, or feature name.
+    :returns:
+        ``True`` if the module, codec, or feature is available,
+        ``False`` or ``None`` otherwise.
+    """
+
+    if feature in modules:
+        return check_module(feature)
+    if feature in codecs:
+        return check_codec(feature)
+    if feature in features:
+        return check_feature(feature)
+    warnings.warn(f"Unknown feature '{feature}'.", stacklevel=2)
+    return False
+
+
+def version(feature: str) -> str | None:
+    """
+    :param feature:
+        The module, codec, or feature to check for.
+    :returns:
+        The version number as a string, or ``None`` if unknown or not available.
+    """
+    if feature in modules:
+        return version_module(feature)
+    if feature in codecs:
+        return version_codec(feature)
+    if feature in features:
+        return version_feature(feature)
+    return None
+
+
+def get_supported() -> list[str]:
+    """
+    :returns: A list of all supported modules, features, and codecs.
+    """
+
+    ret = get_supported_modules()
+    ret.extend(get_supported_features())
+    ret.extend(get_supported_codecs())
+    return ret
+
+
+def pilinfo(out: IO[str] | None = None, supported_formats: bool = True) -> None:
+    """
+    Prints information about this installation of Pillow.
+    This function can be called with ``python3 -m PIL``.
+    It can also be called with ``python3 -m PIL.report`` or ``python3 -m PIL --report``
+    to have "supported_formats" set to ``False``, omitting the list of all supported
+    image file formats.
+
+    :param out:
+        The output stream to print to. Defaults to ``sys.stdout`` if ``None``.
+    :param supported_formats:
+        If ``True``, a list of all supported image file formats will be printed.
+    """
+
+    if out is None:
+        out = sys.stdout
+
+    Image.init()
+
+    print("-" * 68, file=out)
+    print(f"Pillow {PIL.__version__}", file=out)
+    py_version_lines = sys.version.splitlines()
+    print(f"Python {py_version_lines[0].strip()}", file=out)
+    for py_version in py_version_lines[1:]:
+        print(f"       {py_version.strip()}", file=out)
+    print("-" * 68, file=out)
+    print(f"Python executable is {sys.executable or 'unknown'}", file=out)
+    if sys.prefix != sys.base_prefix:
+        print(f"Environment Python files loaded from {sys.prefix}", file=out)
+    print(f"System Python files loaded from {sys.base_prefix}", file=out)
+    print("-" * 68, file=out)
+    print(
+        f"Python Pillow modules loaded from {os.path.dirname(Image.__file__)}",
+        file=out,
+    )
+    print(
+        f"Binary Pillow modules loaded from {os.path.dirname(Image.core.__file__)}",
+        file=out,
+    )
+    print("-" * 68, file=out)
+
+    for name, feature in [
+        ("pil", "PIL CORE"),
+        ("tkinter", "TKINTER"),
+        ("freetype2", "FREETYPE2"),
+        ("littlecms2", "LITTLECMS2"),
+        ("webp", "WEBP"),
+        ("jpg", "JPEG"),
+        ("jpg_2000", "OPENJPEG (JPEG2000)"),
+        ("zlib", "ZLIB (PNG/ZIP)"),
+        ("libtiff", "LIBTIFF"),
+        ("raqm", "RAQM (Bidirectional Text)"),
+        ("libimagequant", "LIBIMAGEQUANT (Quantization method)"),
+        ("xcb", "XCB (X protocol)"),
+    ]:
+        if check(name):
+            v: str | None = None
+            if name == "jpg":
+                libjpeg_turbo_version = version_feature("libjpeg_turbo")
+                if libjpeg_turbo_version is not None:
+                    v = "libjpeg-turbo " + libjpeg_turbo_version
+            if v is None:
+                v = version(name)
+            if v is not None:
+                version_static = name in ("pil", "jpg")
+                if name == "littlecms2":
+                    # this check is also in src/_imagingcms.c:setup_module()
+                    version_static = tuple(int(x) for x in v.split(".")) < (2, 7)
+                t = "compiled for" if version_static else "loaded"
+                if name == "zlib":
+                    zlib_ng_version = version_feature("zlib_ng")
+                    if zlib_ng_version is not None:
+                        v += ", compiled for zlib-ng " + zlib_ng_version
+                elif name == "raqm":
+                    for f in ("fribidi", "harfbuzz"):
+                        v2 = version_feature(f)
+                        if v2 is not None:
+                            v += f", {f} {v2}"
+                print("---", feature, "support ok,", t, v, file=out)
+            else:
+                print("---", feature, "support ok", file=out)
+        else:
+            print("***", feature, "support not installed", file=out)
+    print("-" * 68, file=out)
+
+    if supported_formats:
+        extensions = collections.defaultdict(list)
+        for ext, i in Image.EXTENSION.items():
+            extensions[i].append(ext)
+
+        for i in sorted(Image.ID):
+            line = f"{i}"
+            if i in Image.MIME:
+                line = f"{line} {Image.MIME[i]}"
+            print(line, file=out)
+
+            if i in extensions:
+                print(
+                    "Extensions: {}".format(", ".join(sorted(extensions[i]))), file=out
+                )
+
+            features = []
+            if i in Image.OPEN:
+                features.append("open")
+            if i in Image.SAVE:
+                features.append("save")
+            if i in Image.SAVE_ALL:
+                features.append("save_all")
+            if i in Image.DECODERS:
+                features.append("decode")
+            if i in Image.ENCODERS:
+                features.append("encode")
+
+            print("Features: {}".format(", ".join(features)), file=out)
+            print("-" * 68, file=out)
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/py.typed b/backend/venv/lib/python3.12/site-packages/PIL/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/PIL/report.py b/backend/venv/lib/python3.12/site-packages/PIL/report.py
new file mode 100644
index 00000000..d2815e84
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PIL/report.py
@@ -0,0 +1,5 @@
+from __future__ import annotations
+
+from .features import pilinfo
+
+pilinfo(supported_formats=False)
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/AUTHORS.rst b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/AUTHORS.rst
new file mode 100644
index 00000000..88e2b6ad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/AUTHORS.rst
@@ -0,0 +1,7 @@
+Authors
+=======
+
+``pyjwt`` is currently written and maintained by `Jose Padilla <https://github.com/jpadilla>`_.
+Originally written and maintained by `Jeff Lindsay <https://github.com/progrium>`_.
+
+A full list of contributors can be found on GitHub’s `overview <https://github.com/jpadilla/pyjwt/graphs/contributors>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/LICENSE
new file mode 100644
index 00000000..fd0ecbc8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/LICENSE
@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2015-2022 José Padilla
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/METADATA
new file mode 100644
index 00000000..f31b700e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/METADATA
@@ -0,0 +1,106 @@
+Metadata-Version: 2.1
+Name: PyJWT
+Version: 2.10.1
+Summary: JSON Web Token implementation in Python
+Author-email: Jose Padilla <hello@jpadilla.com>
+License: MIT
+Project-URL: Homepage, https://github.com/jpadilla/pyjwt
+Keywords: json,jwt,security,signing,token,web
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Natural Language :: English
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Utilities
+Requires-Python: >=3.9
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+License-File: AUTHORS.rst
+Provides-Extra: crypto
+Requires-Dist: cryptography>=3.4.0; extra == "crypto"
+Provides-Extra: dev
+Requires-Dist: coverage[toml]==5.0.4; extra == "dev"
+Requires-Dist: cryptography>=3.4.0; extra == "dev"
+Requires-Dist: pre-commit; extra == "dev"
+Requires-Dist: pytest<7.0.0,>=6.0.0; extra == "dev"
+Requires-Dist: sphinx; extra == "dev"
+Requires-Dist: sphinx-rtd-theme; extra == "dev"
+Requires-Dist: zope.interface; extra == "dev"
+Provides-Extra: docs
+Requires-Dist: sphinx; extra == "docs"
+Requires-Dist: sphinx-rtd-theme; extra == "docs"
+Requires-Dist: zope.interface; extra == "docs"
+Provides-Extra: tests
+Requires-Dist: coverage[toml]==5.0.4; extra == "tests"
+Requires-Dist: pytest<7.0.0,>=6.0.0; extra == "tests"
+
+PyJWT
+=====
+
+.. image:: https://github.com/jpadilla/pyjwt/workflows/CI/badge.svg
+   :target: https://github.com/jpadilla/pyjwt/actions?query=workflow%3ACI
+
+.. image:: https://img.shields.io/pypi/v/pyjwt.svg
+   :target: https://pypi.python.org/pypi/pyjwt
+
+.. image:: https://codecov.io/gh/jpadilla/pyjwt/branch/master/graph/badge.svg
+   :target: https://codecov.io/gh/jpadilla/pyjwt
+
+.. image:: https://readthedocs.org/projects/pyjwt/badge/?version=stable
+   :target: https://pyjwt.readthedocs.io/en/stable/
+
+A Python implementation of `RFC 7519 <https://tools.ietf.org/html/rfc7519>`_. Original implementation was written by `@progrium <https://github.com/progrium>`_.
+
+Sponsor
+-------
+
+.. |auth0-logo| image:: https://github.com/user-attachments/assets/ee98379e-ee76-4bcb-943a-e25c4ea6d174
+   :width: 160px
+
++--------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| |auth0-logo| | If you want to quickly add secure token-based authentication to Python projects, feel free to check Auth0's Python SDK and free plan at `auth0.com/signup <https://auth0.com/signup?utm_source=external_sites&utm_medium=pyjwt&utm_campaign=devn_signup>`_. |
++--------------+-----------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+
+Installing
+----------
+
+Install with **pip**:
+
+.. code-block:: console
+
+    $ pip install PyJWT
+
+
+Usage
+-----
+
+.. code-block:: pycon
+
+    >>> import jwt
+    >>> encoded = jwt.encode({"some": "payload"}, "secret", algorithm="HS256")
+    >>> print(encoded)
+    eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzb21lIjoicGF5bG9hZCJ9.4twFt5NiznN84AWoo1d7KO1T_yoc0Z6XOpOVswacPZg
+    >>> jwt.decode(encoded, "secret", algorithms=["HS256"])
+    {'some': 'payload'}
+
+Documentation
+-------------
+
+View the full docs online at https://pyjwt.readthedocs.io/en/stable/
+
+
+Tests
+-----
+
+You can run tests from the project root after cloning with:
+
+.. code-block:: console
+
+    $ tox
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/RECORD
new file mode 100644
index 00000000..457a3fb8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/RECORD
@@ -0,0 +1,33 @@
+PyJWT-2.10.1.dist-info/AUTHORS.rst,sha256=klzkNGECnu2_VY7At89_xLBF3vUSDruXk3xwgUBxzwc,322
+PyJWT-2.10.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+PyJWT-2.10.1.dist-info/LICENSE,sha256=eXp6ICMdTEM-nxkR2xcx0GtYKLmPSZgZoDT3wPVvXOU,1085
+PyJWT-2.10.1.dist-info/METADATA,sha256=EkewF6D6KU8SGaaQzVYfxUUU1P_gs_dp1pYTkoYvAx8,3990
+PyJWT-2.10.1.dist-info/RECORD,,
+PyJWT-2.10.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+PyJWT-2.10.1.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
+PyJWT-2.10.1.dist-info/top_level.txt,sha256=RP5DHNyJbMq2ka0FmfTgoSaQzh7e3r5XuCWCO8a00k8,4
+jwt/__init__.py,sha256=VB2vFKuboTjcDGeZ8r-UqK_dz3NsQSQEqySSICby8Xg,1711
+jwt/__pycache__/__init__.cpython-312.pyc,,
+jwt/__pycache__/algorithms.cpython-312.pyc,,
+jwt/__pycache__/api_jwk.cpython-312.pyc,,
+jwt/__pycache__/api_jws.cpython-312.pyc,,
+jwt/__pycache__/api_jwt.cpython-312.pyc,,
+jwt/__pycache__/exceptions.cpython-312.pyc,,
+jwt/__pycache__/help.cpython-312.pyc,,
+jwt/__pycache__/jwk_set_cache.cpython-312.pyc,,
+jwt/__pycache__/jwks_client.cpython-312.pyc,,
+jwt/__pycache__/types.cpython-312.pyc,,
+jwt/__pycache__/utils.cpython-312.pyc,,
+jwt/__pycache__/warnings.cpython-312.pyc,,
+jwt/algorithms.py,sha256=cKr-XEioe0mBtqJMCaHEswqVOA1Z8Purt5Sb3Bi-5BE,30409
+jwt/api_jwk.py,sha256=6F1r7rmm8V5qEnBKA_xMjS9R7VoANe1_BL1oD2FrAjE,4451
+jwt/api_jws.py,sha256=aM8vzqQf6mRrAw7bRy-Moj_pjWsKSVQyYK896AfMjJU,11762
+jwt/api_jwt.py,sha256=OGT4hok1l5A6FH_KdcrU5g6u6EQ8B7em0r9kGM9SYgA,14512
+jwt/exceptions.py,sha256=bUIOJ-v9tjopTLS-FYOTc3kFx5WP5IZt7ksN_HE1G9Q,1211
+jwt/help.py,sha256=vFdNzjQoAch04XCMYpCkyB2blaqHAGAqQrtf9nSPkdk,1808
+jwt/jwk_set_cache.py,sha256=hBKmN-giU7-G37L_XKgc_OZu2ah4wdbj1ZNG_GkoSE8,959
+jwt/jwks_client.py,sha256=p9b-IbQqo2tEge9Zit3oSPBFNePqwho96VLbnUrHUWs,4259
+jwt/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+jwt/types.py,sha256=VnhGv_VFu5a7_mrPoSCB7HaNLrJdhM8Sq1sSfEg0gLU,99
+jwt/utils.py,sha256=hxOjvDBheBYhz-RIPiEz7Q88dSUSTMzEdKE_Ww2VdJw,3640
+jwt/warnings.py,sha256=50XWOnyNsIaqzUJTk6XHNiIDykiL763GYA92MjTKmok,59
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/WHEEL
new file mode 100644
index 00000000..ae527e7d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/top_level.txt
new file mode 100644
index 00000000..27ccc9bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/PyJWT-2.10.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+jwt
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/LICENSE
new file mode 100644
index 00000000..dfe1a4d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/LICENSE
@@ -0,0 +1,19 @@
+Copyright 2005-2025 SQLAlchemy authors and contributors <see AUTHORS file>.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+of the Software, and to permit persons to whom the Software is furnished to do
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/METADATA
new file mode 100644
index 00000000..7e70fc5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/METADATA
@@ -0,0 +1,243 @@
+Metadata-Version: 2.2
+Name: SQLAlchemy
+Version: 2.0.38
+Summary: Database Abstraction Library
+Home-page: https://www.sqlalchemy.org
+Author: Mike Bayer
+Author-email: mike_mp@zzzcomputing.com
+License: MIT
+Project-URL: Documentation, https://docs.sqlalchemy.org
+Project-URL: Issue Tracker, https://github.com/sqlalchemy/sqlalchemy/
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Database :: Front-Ends
+Requires-Python: >=3.7
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: importlib-metadata; python_version < "3.8"
+Requires-Dist: greenlet!=0.4.17; python_version < "3.14" and (platform_machine == "aarch64" or (platform_machine == "ppc64le" or (platform_machine == "x86_64" or (platform_machine == "amd64" or (platform_machine == "AMD64" or (platform_machine == "win32" or platform_machine == "WIN32"))))))
+Requires-Dist: typing-extensions>=4.6.0
+Provides-Extra: asyncio
+Requires-Dist: greenlet!=0.4.17; extra == "asyncio"
+Provides-Extra: mypy
+Requires-Dist: mypy>=0.910; extra == "mypy"
+Provides-Extra: mssql
+Requires-Dist: pyodbc; extra == "mssql"
+Provides-Extra: mssql-pymssql
+Requires-Dist: pymssql; extra == "mssql-pymssql"
+Provides-Extra: mssql-pyodbc
+Requires-Dist: pyodbc; extra == "mssql-pyodbc"
+Provides-Extra: mysql
+Requires-Dist: mysqlclient>=1.4.0; extra == "mysql"
+Provides-Extra: mysql-connector
+Requires-Dist: mysql-connector-python; extra == "mysql-connector"
+Provides-Extra: mariadb-connector
+Requires-Dist: mariadb!=1.1.10,!=1.1.2,!=1.1.5,>=1.0.1; extra == "mariadb-connector"
+Provides-Extra: oracle
+Requires-Dist: cx_oracle>=8; extra == "oracle"
+Provides-Extra: oracle-oracledb
+Requires-Dist: oracledb>=1.0.1; extra == "oracle-oracledb"
+Provides-Extra: postgresql
+Requires-Dist: psycopg2>=2.7; extra == "postgresql"
+Provides-Extra: postgresql-pg8000
+Requires-Dist: pg8000>=1.29.1; extra == "postgresql-pg8000"
+Provides-Extra: postgresql-asyncpg
+Requires-Dist: greenlet!=0.4.17; extra == "postgresql-asyncpg"
+Requires-Dist: asyncpg; extra == "postgresql-asyncpg"
+Provides-Extra: postgresql-psycopg2binary
+Requires-Dist: psycopg2-binary; extra == "postgresql-psycopg2binary"
+Provides-Extra: postgresql-psycopg2cffi
+Requires-Dist: psycopg2cffi; extra == "postgresql-psycopg2cffi"
+Provides-Extra: postgresql-psycopg
+Requires-Dist: psycopg>=3.0.7; extra == "postgresql-psycopg"
+Provides-Extra: postgresql-psycopgbinary
+Requires-Dist: psycopg[binary]>=3.0.7; extra == "postgresql-psycopgbinary"
+Provides-Extra: pymysql
+Requires-Dist: pymysql; extra == "pymysql"
+Provides-Extra: aiomysql
+Requires-Dist: greenlet!=0.4.17; extra == "aiomysql"
+Requires-Dist: aiomysql>=0.2.0; extra == "aiomysql"
+Provides-Extra: aioodbc
+Requires-Dist: greenlet!=0.4.17; extra == "aioodbc"
+Requires-Dist: aioodbc; extra == "aioodbc"
+Provides-Extra: asyncmy
+Requires-Dist: greenlet!=0.4.17; extra == "asyncmy"
+Requires-Dist: asyncmy!=0.2.4,!=0.2.6,>=0.2.3; extra == "asyncmy"
+Provides-Extra: aiosqlite
+Requires-Dist: greenlet!=0.4.17; extra == "aiosqlite"
+Requires-Dist: aiosqlite; extra == "aiosqlite"
+Requires-Dist: typing_extensions!=3.10.0.1; extra == "aiosqlite"
+Provides-Extra: sqlcipher
+Requires-Dist: sqlcipher3_binary; extra == "sqlcipher"
+
+SQLAlchemy
+==========
+
+|PyPI| |Python| |Downloads|
+
+.. |PyPI| image:: https://img.shields.io/pypi/v/sqlalchemy
+    :target: https://pypi.org/project/sqlalchemy
+    :alt: PyPI
+
+.. |Python| image:: https://img.shields.io/pypi/pyversions/sqlalchemy
+    :target: https://pypi.org/project/sqlalchemy
+    :alt: PyPI - Python Version
+
+.. |Downloads| image:: https://static.pepy.tech/badge/sqlalchemy/month
+    :target: https://pepy.tech/project/sqlalchemy
+    :alt: PyPI - Downloads
+
+
+The Python SQL Toolkit and Object Relational Mapper
+
+Introduction
+-------------
+
+SQLAlchemy is the Python SQL toolkit and Object Relational Mapper
+that gives application developers the full power and
+flexibility of SQL. SQLAlchemy provides a full suite
+of well known enterprise-level persistence patterns,
+designed for efficient and high-performing database
+access, adapted into a simple and Pythonic domain
+language.
+
+Major SQLAlchemy features include:
+
+* An industrial strength ORM, built
+  from the core on the identity map, unit of work,
+  and data mapper patterns.   These patterns
+  allow transparent persistence of objects
+  using a declarative configuration system.
+  Domain models
+  can be constructed and manipulated naturally,
+  and changes are synchronized with the
+  current transaction automatically.
+* A relationally-oriented query system, exposing
+  the full range of SQL's capabilities
+  explicitly, including joins, subqueries,
+  correlation, and most everything else,
+  in terms of the object model.
+  Writing queries with the ORM uses the same
+  techniques of relational composition you use
+  when writing SQL.  While you can drop into
+  literal SQL at any time, it's virtually never
+  needed.
+* A comprehensive and flexible system
+  of eager loading for related collections and objects.
+  Collections are cached within a session,
+  and can be loaded on individual access, all
+  at once using joins, or by query per collection
+  across the full result set.
+* A Core SQL construction system and DBAPI
+  interaction layer.  The SQLAlchemy Core is
+  separate from the ORM and is a full database
+  abstraction layer in its own right, and includes
+  an extensible Python-based SQL expression
+  language, schema metadata, connection pooling,
+  type coercion, and custom types.
+* All primary and foreign key constraints are
+  assumed to be composite and natural.  Surrogate
+  integer primary keys are of course still the
+  norm, but SQLAlchemy never assumes or hardcodes
+  to this model.
+* Database introspection and generation.  Database
+  schemas can be "reflected" in one step into
+  Python structures representing database metadata;
+  those same structures can then generate
+  CREATE statements right back out - all within
+  the Core, independent of the ORM.
+
+SQLAlchemy's philosophy:
+
+* SQL databases behave less and less like object
+  collections the more size and performance start to
+  matter; object collections behave less and less like
+  tables and rows the more abstraction starts to matter.
+  SQLAlchemy aims to accommodate both of these
+  principles.
+* An ORM doesn't need to hide the "R".   A relational
+  database provides rich, set-based functionality
+  that should be fully exposed.   SQLAlchemy's
+  ORM provides an open-ended set of patterns
+  that allow a developer to construct a custom
+  mediation layer between a domain model and
+  a relational schema, turning the so-called
+  "object relational impedance" issue into
+  a distant memory.
+* The developer, in all cases, makes all decisions
+  regarding the design, structure, and naming conventions
+  of both the object model as well as the relational
+  schema.   SQLAlchemy only provides the means
+  to automate the execution of these decisions.
+* With SQLAlchemy, there's no such thing as
+  "the ORM generated a bad query" - you
+  retain full control over the structure of
+  queries, including how joins are organized,
+  how subqueries and correlation is used, what
+  columns are requested.  Everything SQLAlchemy
+  does is ultimately the result of a developer-initiated 
+  decision.
+* Don't use an ORM if the problem doesn't need one.
+  SQLAlchemy consists of a Core and separate ORM
+  component.   The Core offers a full SQL expression
+  language that allows Pythonic construction
+  of SQL constructs that render directly to SQL
+  strings for a target database, returning
+  result sets that are essentially enhanced DBAPI
+  cursors.
+* Transactions should be the norm.  With SQLAlchemy's
+  ORM, nothing goes to permanent storage until
+  commit() is called.  SQLAlchemy encourages applications
+  to create a consistent means of delineating
+  the start and end of a series of operations.
+* Never render a literal value in a SQL statement.
+  Bound parameters are used to the greatest degree
+  possible, allowing query optimizers to cache
+  query plans effectively and making SQL injection
+  attacks a non-issue.
+
+Documentation
+-------------
+
+Latest documentation is at:
+
+https://www.sqlalchemy.org/docs/
+
+Installation / Requirements
+---------------------------
+
+Full documentation for installation is at
+`Installation <https://www.sqlalchemy.org/docs/intro.html#installation>`_.
+
+Getting Help / Development / Bug reporting
+------------------------------------------
+
+Please refer to the `SQLAlchemy Community Guide <https://www.sqlalchemy.org/support.html>`_.
+
+Code of Conduct
+---------------
+
+Above all, SQLAlchemy places great emphasis on polite, thoughtful, and
+constructive communication between users and developers.
+Please see our current Code of Conduct at
+`Code of Conduct <https://www.sqlalchemy.org/codeofconduct.html>`_.
+
+License
+-------
+
+SQLAlchemy is distributed under the `MIT license
+<https://www.opensource.org/licenses/mit-license.php>`_.
+
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/RECORD
new file mode 100644
index 00000000..8783513e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/RECORD
@@ -0,0 +1,530 @@
+SQLAlchemy-2.0.38.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+SQLAlchemy-2.0.38.dist-info/LICENSE,sha256=mCFyC1jUpWW2EyEAeorUOraZGjlZ5mzV203Z6uacffw,1100
+SQLAlchemy-2.0.38.dist-info/METADATA,sha256=qsQnUJ7IfOD1DMmH6MGG3g_L2WqZuKshzK1_PGbNnCQ,9641
+SQLAlchemy-2.0.38.dist-info/RECORD,,
+SQLAlchemy-2.0.38.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+SQLAlchemy-2.0.38.dist-info/WHEEL,sha256=VujM3ypTCyUW6hcTDdK2ej0ARVMxlU1Djlh_zWnDgqk,109
+SQLAlchemy-2.0.38.dist-info/top_level.txt,sha256=rp-ZgB7D8G11ivXON5VGPjupT1voYmWqkciDt5Uaw_Q,11
+sqlalchemy/__init__.py,sha256=vYerm6uGcyLqASO-K-scqU2qswyLASrifWJ6sRYRjds,13033
+sqlalchemy/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/__pycache__/events.cpython-312.pyc,,
+sqlalchemy/__pycache__/exc.cpython-312.pyc,,
+sqlalchemy/__pycache__/inspection.cpython-312.pyc,,
+sqlalchemy/__pycache__/log.cpython-312.pyc,,
+sqlalchemy/__pycache__/schema.cpython-312.pyc,,
+sqlalchemy/__pycache__/types.cpython-312.pyc,,
+sqlalchemy/connectors/__init__.py,sha256=YeSHsOB0YhdM6jZUvHFQFwKqNXO02MlklmGW0yCywjI,476
+sqlalchemy/connectors/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/connectors/__pycache__/aioodbc.cpython-312.pyc,,
+sqlalchemy/connectors/__pycache__/asyncio.cpython-312.pyc,,
+sqlalchemy/connectors/__pycache__/pyodbc.cpython-312.pyc,,
+sqlalchemy/connectors/aioodbc.py,sha256=KT9xi2xQ4AJgDiGPTV5h_5qi9dummmenKAvWelwza3w,5288
+sqlalchemy/connectors/asyncio.py,sha256=00claZADdFUh2iQmlpqoLhLTBxK0i79Mwd9WZqUtleM,6138
+sqlalchemy/connectors/pyodbc.py,sha256=GsW9bD0H30OMTbGDx9SdaTT_ujgpxP7TM4rfhIzD4mo,8501
+sqlalchemy/cyextension/__init__.py,sha256=4npVIjitKfUs0NQ6f3UdQBDq4ipJ0_ZNB2mpKqtc5ik,244
+sqlalchemy/cyextension/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/cyextension/collections.cpython-312-darwin.so,sha256=vGs47smcJY7BrRCdmIYpydbilOwlPZuQa7cg5z6xSlQ,247360
+sqlalchemy/cyextension/collections.pyx,sha256=L7DZ3DGKpgw2MT2ZZRRxCnrcyE5pU1NAFowWgAzQPEc,12571
+sqlalchemy/cyextension/immutabledict.cpython-312-darwin.so,sha256=4APbgDX6bwFZ6mA2ie64IkTslfXvlP64Z9Vz0xqRRWM,122336
+sqlalchemy/cyextension/immutabledict.pxd,sha256=3x3-rXG5eRQ7bBnktZ-OJ9-6ft8zToPmTDOd92iXpB0,291
+sqlalchemy/cyextension/immutabledict.pyx,sha256=KfDTYbTfebstE8xuqAtuXsHNAK0_b5q_ymUiinUe_xs,3535
+sqlalchemy/cyextension/processors.cpython-312-darwin.so,sha256=tT5thytZcM6BiexdvRqi9hmnHWe9CvVzXBdiFarrpEo,102928
+sqlalchemy/cyextension/processors.pyx,sha256=R1rHsGLEaGeBq5VeCydjClzYlivERIJ9B-XLOJlf2MQ,1792
+sqlalchemy/cyextension/resultproxy.cpython-312-darwin.so,sha256=Zc4S7XwIIvgxOV5qwqPZtkuLncz5b-wiQPmFT5oifE4,104864
+sqlalchemy/cyextension/resultproxy.pyx,sha256=eWLdyBXiBy_CLQrF5ScfWJm7X0NeelscSXedtj1zv9Q,2725
+sqlalchemy/cyextension/util.cpython-312-darwin.so,sha256=75nS344FWeCbsXQsP8JowiXCwTrAQGdxto0xzfFctec,122296
+sqlalchemy/cyextension/util.pyx,sha256=B85orxa9LddLuQEaDoVSq1XmAXIbLKxrxpvuB8ogV_o,2530
+sqlalchemy/dialects/__init__.py,sha256=4jxiSgI_fVCNXcz42gQYKEp0k07RAHyQN4ZpjaNsFUI,1770
+sqlalchemy/dialects/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/__pycache__/_typing.cpython-312.pyc,,
+sqlalchemy/dialects/_typing.py,sha256=8YwrkOa8IvmBojwwegbL5mL_0UAuzdqYiKHKANpvHMw,971
+sqlalchemy/dialects/mssql/__init__.py,sha256=6t_aNpgbMLdPE9gpHYTf9o6QfVavncztRLbr21l2NaY,1880
+sqlalchemy/dialects/mssql/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/aioodbc.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/information_schema.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/json.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/pymssql.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/__pycache__/pyodbc.cpython-312.pyc,,
+sqlalchemy/dialects/mssql/aioodbc.py,sha256=4CmhwIkZrabpG-r7_ogRVajD-nhRZSFJ0Swz2d0jIHM,2021
+sqlalchemy/dialects/mssql/base.py,sha256=2UCotpN3WBPgMddhXVP6Epc-srvNrYHCnK4kcEbjW6w,132713
+sqlalchemy/dialects/mssql/information_schema.py,sha256=v5MZz1FN72THEwF_u3Eh_2vnWdFE13RYydOioMMcuvU,8084
+sqlalchemy/dialects/mssql/json.py,sha256=F53pibuOVRzgDtjoclOI7LnkKXNVsaVfJyBH1XAhyDo,4756
+sqlalchemy/dialects/mssql/provision.py,sha256=P1tqxZ4f6Oeqn2gNi7dXl82LRLCg1-OB4eWiZc6CHek,5593
+sqlalchemy/dialects/mssql/pymssql.py,sha256=C7yAs3Pw81W1KTVNc6_0sHQuYlJ5iH82vKByY4TkB1g,4097
+sqlalchemy/dialects/mssql/pyodbc.py,sha256=CnO7KDWxbxb7AoZhp_PMDBvVSMuzwq1h4Cav2IWFWDo,27173
+sqlalchemy/dialects/mysql/__init__.py,sha256=ropOMUWrAcL-Q7h-9jQ_tb3ISAFIsNRQ8YVXvn0URl0,2206
+sqlalchemy/dialects/mysql/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/aiomysql.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/asyncmy.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/cymysql.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/dml.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/enumerated.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/expression.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/json.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/mariadb.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/mariadbconnector.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/mysqlconnector.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/mysqldb.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/pymysql.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/pyodbc.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/reflection.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/reserved_words.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/__pycache__/types.cpython-312.pyc,,
+sqlalchemy/dialects/mysql/aiomysql.py,sha256=yrujoFtAG0QvtVlgbGBUMg3kXeXlIH62tvyYTCMUfnE,10013
+sqlalchemy/dialects/mysql/asyncmy.py,sha256=rmVSf86VYxgAUROIKfVtvS-grG9aPBiLY_Gu0KJMjuo,10081
+sqlalchemy/dialects/mysql/base.py,sha256=5ih5fhH0HOGnJBhjGKdF8OeJWkh8YeiUc8qf_xgGtqw,123602
+sqlalchemy/dialects/mysql/cymysql.py,sha256=KwxSsF4a6uUd6yblhSns8uj4hgmhv4hFInTZNdmRixA,2300
+sqlalchemy/dialects/mysql/dml.py,sha256=VjnTobe_SBNF2RN6tvqa5LOn-9x4teVUyzUedZkOmdc,7768
+sqlalchemy/dialects/mysql/enumerated.py,sha256=qI5gnBYhxk9dhPeUfGiijp0qT2Puazdp27-ba_38uWQ,8447
+sqlalchemy/dialects/mysql/expression.py,sha256=3PEKPwYIZ8mVXkjUgHaj_efPBYuBNWZSnfUcJuoZddA,4121
+sqlalchemy/dialects/mysql/json.py,sha256=W31DojiRypifXKVh3PJSWP7IHqFoeKwzLl-0CJH6QRI,2269
+sqlalchemy/dialects/mysql/mariadb.py,sha256=g4v4WQuXHn556Nn6k-RgvPrmfCql1R46fIEk6UEx0U8,1450
+sqlalchemy/dialects/mysql/mariadbconnector.py,sha256=t4m6kfYBoURjNXRxlEsRajjvArNDc4lmaFGxHQh7VTo,8623
+sqlalchemy/dialects/mysql/mysqlconnector.py,sha256=gdNOGdRqvnCbLZpKjpubu_0tGRQ5Tn_2TZvbp3v9rX0,5729
+sqlalchemy/dialects/mysql/mysqldb.py,sha256=5ME7B0WI9G8tw5482YBejDg38uVMXR2oUasNDOCsAqQ,9526
+sqlalchemy/dialects/mysql/provision.py,sha256=5LCeInPvyEbGuzxSs9rnnLYkMsFpW3IJ8lC-sjTfKnk,3575
+sqlalchemy/dialects/mysql/pymysql.py,sha256=osp0em1s3Cip5Vpcj-PeaH7btHEInorO-qs351muw3Q,4082
+sqlalchemy/dialects/mysql/pyodbc.py,sha256=ZiFNJQq2qiOTzTZLmNJQ938EnS1ItVsNDa3fvNEDqnI,4298
+sqlalchemy/dialects/mysql/reflection.py,sha256=eGV9taua0nZS_HsHyAy6zjcHEHFPXmFdux-bUmtOeWs,22834
+sqlalchemy/dialects/mysql/reserved_words.py,sha256=C9npWSuhsxoVCqETxCQ1zE_UEgy4gfiHw9zI5dPkjWI,9258
+sqlalchemy/dialects/mysql/types.py,sha256=w68OASMw04xkyAc0_GtXkuEhhVqlR6LTwaOch4KaAFQ,24343
+sqlalchemy/dialects/oracle/__init__.py,sha256=rp9qPRNQAk1Yq_Zhe7SsUH8EvFgNOAh8XOF17Lkxpyo,1493
+sqlalchemy/dialects/oracle/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/cx_oracle.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/dictionary.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/oracledb.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/__pycache__/types.cpython-312.pyc,,
+sqlalchemy/dialects/oracle/base.py,sha256=27y-G7FMrTiC1u9C4acBGa_V7a62neKRuCTdCbGqY2o,128476
+sqlalchemy/dialects/oracle/cx_oracle.py,sha256=ohENTgLxGUfobRH3K8KdeZgBRPG1rX3vY-ph9blj-2g,56612
+sqlalchemy/dialects/oracle/dictionary.py,sha256=J7tGVE0KyUPZKpPLOary3HdDq1DWd29arF5udLgv8_o,19519
+sqlalchemy/dialects/oracle/oracledb.py,sha256=veqto1AUIbSxRmpUQin0ysMV8Y6sWAkzXt7W8IIl118,33771
+sqlalchemy/dialects/oracle/provision.py,sha256=ga1gNQZlXZKk7DYuYegllUejJxZXRKDGa7dbi_S_poc,8313
+sqlalchemy/dialects/oracle/types.py,sha256=axN6Yidx9tGRIUAbDpBrhMWXE-C8jSllFpTghpGOOzU,9058
+sqlalchemy/dialects/postgresql/__init__.py,sha256=kD8W-SV5e2CesvWg2MQAtncXuZFwGPfR_UODvmRXE08,3892
+sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/array.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/json.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/__pycache__/types.cpython-312.pyc,,
+sqlalchemy/dialects/postgresql/_psycopg_common.py,sha256=szME-lCjVwqnW9-USA6e8ke8N_bN3IbqnIm_oZruvqc,5696
+sqlalchemy/dialects/postgresql/array.py,sha256=28kndSQwgvNWlO4z6MUh5WYAtNSgkgBa6qSEQCIflks,13856
+sqlalchemy/dialects/postgresql/asyncpg.py,sha256=OG35rGetY-GVtyUhOQFJ9w-blbCsPUDY6ETkJ6AIOyM,41287
+sqlalchemy/dialects/postgresql/base.py,sha256=otAswEHqeRhbN9_AGMxnwDo6r872ECkiJ5FMetXfS0k,179452
+sqlalchemy/dialects/postgresql/dml.py,sha256=2SmyMeYveAgm7OnT_CJvwad2nh8BP37yT6gFs8dBYN8,12126
+sqlalchemy/dialects/postgresql/ext.py,sha256=MtN4IU5sRYvoY-E8PTltJ1CuIGb-aCwY2pHMPJcTboA,16318
+sqlalchemy/dialects/postgresql/hstore.py,sha256=wR4gmvfQWPssHwYTXEsPJTb4LkBS6x4e4XXE6smtDH4,11934
+sqlalchemy/dialects/postgresql/json.py,sha256=9sHFGTRFyNbLsANrVYookw9NOJwIPTsEBRNIOUOzOGw,11612
+sqlalchemy/dialects/postgresql/named_types.py,sha256=TEWaBCjuHM2WJoQNrQErQ6f_bUkWypGJfW71wzVJXWc,17572
+sqlalchemy/dialects/postgresql/operators.py,sha256=ay3ckNsWtqDjxDseTdKMGGqYVzST6lmfhbbYHG_bxCw,2808
+sqlalchemy/dialects/postgresql/pg8000.py,sha256=RAykzZuO3Anr6AsyK2JYr7CPb2pru6WtkrX2phCyCGU,18638
+sqlalchemy/dialects/postgresql/pg_catalog.py,sha256=lgJMn7aDuJI2XeHddLkge5NFy6oB2-aDSn8A47QpwAU,9254
+sqlalchemy/dialects/postgresql/provision.py,sha256=7pg9-nOnaK5XBzqByXNPuvi3rxtnRa3dJxdSPVq4eeA,5770
+sqlalchemy/dialects/postgresql/psycopg.py,sha256=k7zXsJj35aOXCrhsbMxwTQX5JWegrqirFJ1Hgbq-GjQ,23326
+sqlalchemy/dialects/postgresql/psycopg2.py,sha256=1KXw9RzsQEAXJazCBywdP5CwLu-HsCSDAD_Khc_rPTM,32032
+sqlalchemy/dialects/postgresql/psycopg2cffi.py,sha256=nKilJfvO9mJwk5NRw5iZDekKY5vi379tvdUJ2vn5eyQ,1756
+sqlalchemy/dialects/postgresql/ranges.py,sha256=fnaj4YgCQGO-G_S4k5ea8bYMH7SzggKJdUX5qfaNp4Y,32978
+sqlalchemy/dialects/postgresql/types.py,sha256=sjb-m-h49lbLBFh0P30G8BWgf_aKNiNyVwWEktugwRw,7286
+sqlalchemy/dialects/sqlite/__init__.py,sha256=6Xcz3nPsl8lqCcZ4-VzPRmkMrkKgAp2buKsClZelU7c,1182
+sqlalchemy/dialects/sqlite/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/aiosqlite.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/dml.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/json.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/pysqlcipher.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/__pycache__/pysqlite.cpython-312.pyc,,
+sqlalchemy/dialects/sqlite/aiosqlite.py,sha256=Uvx23OGcp7_cpRzdm46Zkbf04Yg-jVImv4ELc4WrODo,13194
+sqlalchemy/dialects/sqlite/base.py,sha256=PvwPzukomHAkufUzSqgfJcbKC2ZJAkJbVnW2BQB2T58,98271
+sqlalchemy/dialects/sqlite/dml.py,sha256=4N8qh06RuMphLoQgWw7wv5nXIrka57jIFvK2x9xTZqg,9138
+sqlalchemy/dialects/sqlite/json.py,sha256=A62xPyLRZxl2hvgTMM92jd_7jlw9UE_4Y6Udqt-8g04,2777
+sqlalchemy/dialects/sqlite/provision.py,sha256=iLJyeQSy8pfr9lwEu4_d4O_CI4OavAtkNeRi3qqys1U,5632
+sqlalchemy/dialects/sqlite/pysqlcipher.py,sha256=di8rYryfL0KAn3pRGepmunHyIRGy-4Hhr-2q_ehPzss,5371
+sqlalchemy/dialects/sqlite/pysqlite.py,sha256=rg7F1S2UOhUu6Y1xNVaqF8VbA-FsRY_Y_XpGTpkKpGs,28087
+sqlalchemy/dialects/type_migration_guidelines.txt,sha256=-uHNdmYFGB7bzUNT6i8M5nb4j6j9YUKAtW4lcBZqsMg,8239
+sqlalchemy/engine/__init__.py,sha256=EF4haWCPu95WtWx1GzcHRJ_bBmtJMznno3I2TQ-ZIHE,2818
+sqlalchemy/engine/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/_py_processors.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/_py_row.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/_py_util.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/characteristics.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/create.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/cursor.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/default.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/events.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/interfaces.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/mock.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/processors.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/reflection.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/result.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/row.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/strategies.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/url.cpython-312.pyc,,
+sqlalchemy/engine/__pycache__/util.cpython-312.pyc,,
+sqlalchemy/engine/_py_processors.py,sha256=7QxgkVOd5h1Qd22qFh-pPZdM7RBRzNjj8lWAMWrilcI,3744
+sqlalchemy/engine/_py_row.py,sha256=yNdrZe36yw6mO7x0OEbG0dGojH7CQkNReIwn9LMUPUs,3787
+sqlalchemy/engine/_py_util.py,sha256=LdpbNRQIrJo3EkmiwNkM5bxGUf4uWuL5uS_u-zHadWc,2484
+sqlalchemy/engine/base.py,sha256=9kCWrDp3ECOlQ7BHK_efYAILo3-emcPSk4F8AFRgN7E,122901
+sqlalchemy/engine/characteristics.py,sha256=PepmGApo1sL01dS1qtSbmHplu9ZCdtuSegiGI7L7NZY,4765
+sqlalchemy/engine/create.py,sha256=4gFkqV7fgJbI1906DC4zDgFFX1-xJQ96GIHIrQuc-w4,33217
+sqlalchemy/engine/cursor.py,sha256=6KIZqlwWMUMv02w_el4uNYFMYcfc7eWbkAxW27UyDLE,76305
+sqlalchemy/engine/default.py,sha256=SHM6boxcDNk7MW_Eyd0zCb557Eqf8KTdX1iTUbS0DLw,84705
+sqlalchemy/engine/events.py,sha256=4_e6Ip32ar2Eb27R4ipamiKC-7Tpg4lVz3txabhT5Rc,37400
+sqlalchemy/engine/interfaces.py,sha256=fGmcrBt8yT78ty0R3e3XUvsPh7XYDU_b1JW3QhK_MwY,113029
+sqlalchemy/engine/mock.py,sha256=_aXG1xzj_TO5UWdz8IthPj1ZJ8IlhsKw6D9mmFN_frQ,4181
+sqlalchemy/engine/processors.py,sha256=XK32bULBkuVVRa703u4-SrTCDi_a18Dxq1M09QFBEPw,2379
+sqlalchemy/engine/reflection.py,sha256=_v9zCy3h28hN4KKIUTc5_7KJv7argSgi8A011b_iCdc,75383
+sqlalchemy/engine/result.py,sha256=rgny4qFLmpj80GSdFK35Dpgc3Qk2tc3eJPpahGWVR-M,77622
+sqlalchemy/engine/row.py,sha256=BPtAwsceiRxB9ANpDNM24uQ1M_Zs0xFkSXoKR_I8xyY,12031
+sqlalchemy/engine/strategies.py,sha256=-0rieXY-iXgV83OrJZr-wozFFQn3amKKHchQ6kL-r7A,442
+sqlalchemy/engine/url.py,sha256=gaEeSEJCD0nVEb8J02rIMASrd5L2wYdq5ZXJaj7szVI,31069
+sqlalchemy/engine/util.py,sha256=4OmXwFlmnq6_vBlfUBHnz5LrI_8bT3TwgynX4wcJfnw,5682
+sqlalchemy/event/__init__.py,sha256=ZjVxFGbt9neH5AC4GFiUN5IG2O4j6Z9v2LdmyagJi9w,997
+sqlalchemy/event/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/event/__pycache__/api.cpython-312.pyc,,
+sqlalchemy/event/__pycache__/attr.cpython-312.pyc,,
+sqlalchemy/event/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/event/__pycache__/legacy.cpython-312.pyc,,
+sqlalchemy/event/__pycache__/registry.cpython-312.pyc,,
+sqlalchemy/event/api.py,sha256=NetgcQfbURaZzoxus7_801YDG_LJ7PYqaC3T1lws114,8111
+sqlalchemy/event/attr.py,sha256=YhPXVBPj63Cfyn0nS6h8Ljq0SEbD3mtAZn9HYlzGbtw,20751
+sqlalchemy/event/base.py,sha256=g5eRGX4e949srBK2gUxLYM0RrDUdtUEPS2FT_9IKZeI,15254
+sqlalchemy/event/legacy.py,sha256=lGafKAOF6PY8Bz0AqhN9Q6n-lpXqFLwdv-0T6-UBpow,8227
+sqlalchemy/event/registry.py,sha256=MNEMyR8HZhzQFgxk4Jk_Em6nXTihmGXiSIwPdUnalPM,11144
+sqlalchemy/events.py,sha256=VBRvtckn9JS3tfUfi6UstqUrvQ15J2xamcDByFysIrI,525
+sqlalchemy/exc.py,sha256=AjFBCrOl_V4vQdGegn72Y951RSRMPL6T5qjxnFTGFbM,23978
+sqlalchemy/ext/__init__.py,sha256=BkTNuOg454MpCY9QA3FLK8td7KQhD1W74fOEXxnWibE,322
+sqlalchemy/ext/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/associationproxy.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/automap.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/baked.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/compiler.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/horizontal_shard.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/hybrid.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/indexable.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/instrumentation.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/mutable.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/orderinglist.cpython-312.pyc,,
+sqlalchemy/ext/__pycache__/serializer.cpython-312.pyc,,
+sqlalchemy/ext/associationproxy.py,sha256=VhOFB1vB8hmDYQP90_VdpPI9IFzP3NENkG_eDKziVoI,66062
+sqlalchemy/ext/asyncio/__init__.py,sha256=kTIfpwsHWhqZ-VMOBZFBq66kt1XeF0hNuwOToEDe4_Y,1317
+sqlalchemy/ext/asyncio/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/engine.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/exc.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/result.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/scoping.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/__pycache__/session.cpython-312.pyc,,
+sqlalchemy/ext/asyncio/base.py,sha256=2YQ-nKaHbAm--7q6vbxbznzdwT8oPwetwAarKyu2O8E,8930
+sqlalchemy/ext/asyncio/engine.py,sha256=fe_RZrO-5DiiEgMZ3g-Lti-fdaR7z_Q8gDfPUf-30EY,48198
+sqlalchemy/ext/asyncio/exc.py,sha256=npijuILDXH2p4Q5RzhHzutKwZ5CjtqTcP-U0h9TZUmk,639
+sqlalchemy/ext/asyncio/result.py,sha256=zhhXe13vMT7OfdfGXapgtn4crtiqqctRLb3ka4mmGXY,30477
+sqlalchemy/ext/asyncio/scoping.py,sha256=4f7MX3zUd-4rA8A5wd7j0_GlqCSUxdOPfYd7BBIxkJI,52587
+sqlalchemy/ext/asyncio/session.py,sha256=2wxu06UtJGyf-be2edMFkcK4eLMh8xuGmsAlGRj0YPM,63166
+sqlalchemy/ext/automap.py,sha256=n88mktqvExwjqfsDu3yLIA4wbOIWUpQ1S35Uw3X6ffQ,61675
+sqlalchemy/ext/baked.py,sha256=w3SeRoqnPkIhPL2nRAxfVhyir2ypsiW4kmtmUGKs8qo,17753
+sqlalchemy/ext/compiler.py,sha256=f7o4qhUUldpsx4F1sQoUvdVaT2BhiemqNBCF4r_uQUo,20889
+sqlalchemy/ext/declarative/__init__.py,sha256=SuVflXOGDxx2sB2QSTqNEvqS0fyhOkh3-sy2lRsSOLA,1818
+sqlalchemy/ext/declarative/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/ext/declarative/__pycache__/extensions.cpython-312.pyc,,
+sqlalchemy/ext/declarative/extensions.py,sha256=yHUPcztU-5E1JrNyELDFWKchAnaYK6Y9-dLcqyc1nUI,19531
+sqlalchemy/ext/horizontal_shard.py,sha256=vouIehpQAuwT0HXyWyynTL3m_gcBuLcB-X8lDB0uQ8U,16691
+sqlalchemy/ext/hybrid.py,sha256=DkvNGtiQYzlEBvs1rYEDXhM8vJEXXh_6DMigsHH9w4k,52531
+sqlalchemy/ext/indexable.py,sha256=_dTOgCS96jURcQd9L-hnUMIJDe9KUMyd9gfH57vs078,11065
+sqlalchemy/ext/instrumentation.py,sha256=iCp89rvfK7buW0jJyzKTBDKyMsd06oTRJDItOk4OVSw,15707
+sqlalchemy/ext/mutable.py,sha256=7Zyh2kQq2gm3J_JwsddinIXk7qUuKWbPzRZOmTultEk,37560
+sqlalchemy/ext/mypy/__init__.py,sha256=yVNtoBDNeTl1sqRoA_fSY3o1g6M8NxqUVvAHPRLmFTw,241
+sqlalchemy/ext/mypy/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/apply.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/decl_class.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/infer.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/names.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/plugin.cpython-312.pyc,,
+sqlalchemy/ext/mypy/__pycache__/util.cpython-312.pyc,,
+sqlalchemy/ext/mypy/apply.py,sha256=v_Svc1WiBz9yBXqBVBKoCuPGN286TfVmuuCVZPlbyzo,10591
+sqlalchemy/ext/mypy/decl_class.py,sha256=Nuca4ofHkASAkdqEQlULYB7iLm_KID7Mp384seDhVGg,17384
+sqlalchemy/ext/mypy/infer.py,sha256=29vgn22Hi8E8oIZL6UJCBl6oipiPSAQjxccCEkVb410,19367
+sqlalchemy/ext/mypy/names.py,sha256=hn889DD1nlF0f3drsKi5KSGTG-JefJ2UJrrIQ4L8QWA,10479
+sqlalchemy/ext/mypy/plugin.py,sha256=9YHBp0Bwo92DbDZIUWwIr0hwXPcE4XvHs0-xshvSwUw,9750
+sqlalchemy/ext/mypy/util.py,sha256=CuW2fJ-g9YtkjcypzmrPRaFc-rAvQTzW5A2-w5VTANg,9960
+sqlalchemy/ext/orderinglist.py,sha256=MROa19cm4RZkWXuUuqc1029r7g4HrAJRc17fTHeThvI,14431
+sqlalchemy/ext/serializer.py,sha256=_z95wZMTn3G3sCGN52gwzD4CuKjrhGMr5Eu8g9MxQNg,6169
+sqlalchemy/future/__init__.py,sha256=R1h8VBwMiIUdP3QHv_tFNby557425FJOAGhUoXGvCmc,512
+sqlalchemy/future/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/future/__pycache__/engine.cpython-312.pyc,,
+sqlalchemy/future/engine.py,sha256=2nJFBQAXAE8pqe1cs-D3JjC6wUX2ya2h2e_tniuaBq0,495
+sqlalchemy/inspection.py,sha256=qKEKG37N1OjxpQeVzob1q9VwWjBbjI1x0movJG7fYJ4,5063
+sqlalchemy/log.py,sha256=e_ztNUfZM08FmTWeXN9-doD5YKW44nXxgKCUxxNs6Ow,8607
+sqlalchemy/orm/__init__.py,sha256=BICvTXpLaTNe2AiUaxnZHWzjL5miT9fd_IU-ip3OFNk,8463
+sqlalchemy/orm/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/_orm_constructors.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/_typing.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/attributes.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/bulk_persistence.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/clsregistry.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/collections.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/context.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/decl_api.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/decl_base.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/dependency.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/descriptor_props.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/dynamic.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/evaluator.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/events.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/exc.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/identity.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/instrumentation.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/interfaces.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/loading.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/mapped_collection.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/mapper.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/path_registry.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/persistence.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/properties.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/query.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/relationships.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/scoping.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/session.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/state.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/state_changes.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/strategies.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/strategy_options.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/sync.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/unitofwork.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/util.cpython-312.pyc,,
+sqlalchemy/orm/__pycache__/writeonly.cpython-312.pyc,,
+sqlalchemy/orm/_orm_constructors.py,sha256=NiAagQ1060QYS9n5y_gzPvHQQz44EN1dVtamGVtde6E,103626
+sqlalchemy/orm/_typing.py,sha256=vaYRl4_K3n-sjc9u0Rb4eWWpBOoOi92--OHqaGogRvA,4973
+sqlalchemy/orm/attributes.py,sha256=e_U0A4TGWAzL3yXVvk9YVhIRjKM4RTsIE2PNRLn8LbU,92534
+sqlalchemy/orm/base.py,sha256=DWVup_5mbQ7EbntFGTZolo5vU6xvdekp2kIADrZ0x3A,27501
+sqlalchemy/orm/bulk_persistence.py,sha256=Ciea9MhJ6ZbAi-uGy5-Kj6lodO9bfRqPq8GSf2qFshE,72663
+sqlalchemy/orm/clsregistry.py,sha256=syn6bB-Ylx-juh5GDCmNrPZ58C-z6sdwRkbZFeKysQU,17974
+sqlalchemy/orm/collections.py,sha256=XxZC8d9UX9E2R-WlNH198OPWRPmpLuYt0Y26LrdbuHc,52252
+sqlalchemy/orm/context.py,sha256=eyh7xTo3SyxIHl8_NBUqJ_GpJ0kZtmnTt32Z67cfqgs,112973
+sqlalchemy/orm/decl_api.py,sha256=SJ25fQjjKyWZDQbq5S69eiybpOzns0LkRziP10iW5-E,64969
+sqlalchemy/orm/decl_base.py,sha256=ZlZmyNVOsCPA_pThMeXuWmAhlJwlvTxdGXhnARsKxhk,83288
+sqlalchemy/orm/dependency.py,sha256=4NMhoogevOiX1Wm5B1_yY2u9MHYlIjJNNoEVRE0yLwA,47631
+sqlalchemy/orm/descriptor_props.py,sha256=LgfdiO_U5uznq5ImenfbWGV5T47bH4b_ztbzB4B7FsU,37231
+sqlalchemy/orm/dynamic.py,sha256=Z4GpcVL8rM8gi0bytQOZXw-_kKi-sExbRWGjU30dK3g,9816
+sqlalchemy/orm/evaluator.py,sha256=PKrUW1zEOvmv1XEgc_hBdYqNcyk4zjWr_rJhCEQBFIc,12353
+sqlalchemy/orm/events.py,sha256=OZtTCpI-DVaE6CY16e42GUVpci1U1GjdNO76xU-Tj5Y,127781
+sqlalchemy/orm/exc.py,sha256=zJgAIofYsWKjktqO5MFxze95GlJASziEOJJx-P5_wOU,7413
+sqlalchemy/orm/identity.py,sha256=5NFtF9ZPZWAOmtOqCPyVX2-_pQq9A5XeN2ns3Wirpv8,9249
+sqlalchemy/orm/instrumentation.py,sha256=WhElvvOWOn3Fuc-Asc5HmcKDX6EzFtBleLJKPZEc5A0,24321
+sqlalchemy/orm/interfaces.py,sha256=W6ADDLOixmm4tnSnUP_I9HFLj9MCO2bODk_WTNjkZGA,48797
+sqlalchemy/orm/loading.py,sha256=6Rd1hWtBPm7SfCUpjPQrcoUg_DSCcfhO8Qhz7SScjRE,58277
+sqlalchemy/orm/mapped_collection.py,sha256=FAqaTlOUCYqdws2KR_fW0T8mMWIrLuAxJGU5f4W1aGs,19682
+sqlalchemy/orm/mapper.py,sha256=-gkJKHeAJmIFT153WFIIySduyyLGbT5plCgSfnsa0I0,171668
+sqlalchemy/orm/path_registry.py,sha256=SbVi-7DmInstRvp_sGhURtuFak0O2gCqTvnWYj21zeE,25916
+sqlalchemy/orm/persistence.py,sha256=Uz45Cwxi7FnNiSk2crbh3TzV7b9kb85vmcvOwy5NVmw,61701
+sqlalchemy/orm/properties.py,sha256=vbx_YiSjj3tI94-G-_ghbyWYcIIJQQeGG1P-0RC8Jv4,29065
+sqlalchemy/orm/query.py,sha256=GI_go9ErXYK1BteCmIh5E9iv-jfMJkRBVIlw0XmnYyk,118540
+sqlalchemy/orm/relationships.py,sha256=C40n_-oliMgJJ0FHfwsi1-dm963CrYeKJ5HEYjLdg_o,128899
+sqlalchemy/orm/scoping.py,sha256=-SNRAewfMJ4x4Um8X-yv0k1Thz8E1_kCBmbmG1l1auo,78617
+sqlalchemy/orm/session.py,sha256=1fzksIcb9DtKcwqkS1KkZngkrEYGUHmoNW_o6l8IXQ4,196114
+sqlalchemy/orm/state.py,sha256=1vtlz674sGFmwZ8Ih9TdrslA-0nhU2G52WgV-FoG2j0,37670
+sqlalchemy/orm/state_changes.py,sha256=XJLYYhTZu7nA6uD7xupbLZ9XSzqLYwrDJgW0ZAWvVGE,6815
+sqlalchemy/orm/strategies.py,sha256=qziXv4z2bJeF2qFSj6wogc9BLlxuOnT8nOcEvocVf88,119866
+sqlalchemy/orm/strategy_options.py,sha256=wMYd4E_nRb5ei8Fr3jWeSewNY2k1-AfqtYRGOLiHOFA,85043
+sqlalchemy/orm/sync.py,sha256=RdoxnhvgNjn3Lhtoq4QjvXpj8qfOz__wyibh0FMON0A,5779
+sqlalchemy/orm/unitofwork.py,sha256=hkSIcVonoSt0WWHk019bCDEw0g2o2fg4m4yqoTGyAoo,27033
+sqlalchemy/orm/util.py,sha256=rtClCjtg0eSSC8k-L30W0v6BauJaJuh9Nf-MSqofWuQ,80831
+sqlalchemy/orm/writeonly.py,sha256=R-MVxYDw0ZQ795H21yBtgGSZXWUzSovcb_SO1mv5hoI,22305
+sqlalchemy/pool/__init__.py,sha256=niqzCv2uOZT07DOiV2inlmjrW3lZyqDXGCjnOl1IqJ4,1804
+sqlalchemy/pool/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/pool/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/pool/__pycache__/events.cpython-312.pyc,,
+sqlalchemy/pool/__pycache__/impl.cpython-312.pyc,,
+sqlalchemy/pool/base.py,sha256=mT-PHTlVUGcYRVsMB9LQwNgndjhOTOorWX5-hNRi2FM,52236
+sqlalchemy/pool/events.py,sha256=wdFfvat0fSrVF84Zzsz5E3HnVY0bhL7MPsGME-b2qa8,13149
+sqlalchemy/pool/impl.py,sha256=MLSh83SGNNtZZgZvA-5tvTIT8Dz7U95Bgt8HO_oR1Ps,18944
+sqlalchemy/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sqlalchemy/schema.py,sha256=yt4dcuMAKMleUHVidsAVAsm-JPpASFZXP2xM3pmzYHY,3194
+sqlalchemy/sql/__init__.py,sha256=Y-bZ25Zf-bxqsF2zUkpRGTjFuozNNVQHxUJV3Qmaq2M,5820
+sqlalchemy/sql/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_dml_constructors.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_elements_constructors.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_orm_types.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_py_util.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_selectable_constructors.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/_typing.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/annotation.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/cache_key.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/coercions.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/compiler.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/crud.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/ddl.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/default_comparator.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/dml.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/elements.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/events.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/expression.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/functions.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/lambdas.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/naming.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/operators.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/roles.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/schema.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/selectable.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/sqltypes.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/traversals.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/type_api.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/util.cpython-312.pyc,,
+sqlalchemy/sql/__pycache__/visitors.cpython-312.pyc,,
+sqlalchemy/sql/_dml_constructors.py,sha256=JF_XucNTfAk6Vz9fYiPWOgpIGtUkDj6VPILysLcrVhk,3795
+sqlalchemy/sql/_elements_constructors.py,sha256=eoQhkoRH0qox171ZSODyxxhj_HZEhO64rSowaN-I-v4,62630
+sqlalchemy/sql/_orm_types.py,sha256=0zeMit-V4rYZe-bB9X3xugnjFnPXH0gmeqkJou9Fows,625
+sqlalchemy/sql/_py_util.py,sha256=4KFXNvBq3hhfrr-A1J1uBml3b3CGguIf1dat9gsEHqE,2173
+sqlalchemy/sql/_selectable_constructors.py,sha256=fwVBsDHHWhngodBG205nvhM-Tb3uR1srbCnN3mPgrjA,18785
+sqlalchemy/sql/_typing.py,sha256=zYKlxXnUW_KIkGuBmBnzj-vFG1QON8_F9JN1dl9KSiM,12771
+sqlalchemy/sql/annotation.py,sha256=qHUEwbdmMD3Ybr0ez-Dyiw9l9UB_RUMHWAUIeO_r3gE,18245
+sqlalchemy/sql/base.py,sha256=kfmVNRimU5z6X6OKqMLMs1bDCFQ47BeyF_MZc23nkjY,73848
+sqlalchemy/sql/cache_key.py,sha256=ET2OIQ6jZK2FSxsdnCvhLCrNJ2Fp3zipQ-gvINgAjhQ,33668
+sqlalchemy/sql/coercions.py,sha256=p4zOKXbGYkW_RAXctoD2gxmxHrmLEKKTAancM5LG1QI,40653
+sqlalchemy/sql/compiler.py,sha256=eT_zrKvApimVfycvcTdubQK8-QAzGHm5xWKdhOgnWUY,274965
+sqlalchemy/sql/crud.py,sha256=vFegNw5557ayS4kv761zh0bx0yikEKh1ovMrhErHelg,56514
+sqlalchemy/sql/ddl.py,sha256=rfb7gDvLmn_ktgH2xiXLRTczqnMOED1eakXuGuRPklg,45641
+sqlalchemy/sql/default_comparator.py,sha256=uXLr8B-X6KbybwTjLjZ2hN-WZAvqoMhZ-DDHJX7rAUw,16707
+sqlalchemy/sql/dml.py,sha256=v67qoSe0ExVrzloy-AcYBS6qwIsDlUF4pKBVVRAHiI0,65890
+sqlalchemy/sql/elements.py,sha256=vV4-rdb0Cjwku4ldoULZiFfcgIsKDJC0M04Z6rDZ6wc,176943
+sqlalchemy/sql/events.py,sha256=iWjc_nm1vClDBLg4ZhDnY75CkBdnlDPSPe0MGBSmbiM,18312
+sqlalchemy/sql/expression.py,sha256=rw5tAm8vbd5Vm4MofTZ0ZcXsphz4z9xO_exy-gem6TM,7586
+sqlalchemy/sql/functions.py,sha256=tbBxIeAqLV3kc1YDxyt68mxw0fFy6e93ctRUZSuuf3I,63858
+sqlalchemy/sql/lambdas.py,sha256=h9sPCETBgAanLtVHQsRPHeY-hTEjM5nscq3m4bDstwM,49196
+sqlalchemy/sql/naming.py,sha256=BU0ZdSzXXKHTPhoaKMWJ3gPMoeZSJJe9-3YDYflmjJw,6858
+sqlalchemy/sql/operators.py,sha256=h5bgu31gukGdsYsN_0-1C7IGAdSCFpBxuRjOUnu1Two,76792
+sqlalchemy/sql/roles.py,sha256=drAeWbevjgFAKNcMrH_EuJ-9sSvcq4aeXwAqMXXZGYw,7662
+sqlalchemy/sql/schema.py,sha256=Bg4MorNkXodwtaIh1voyClt8UChusLuK9l00VTorgsg,230022
+sqlalchemy/sql/selectable.py,sha256=aYoBNBJe8YxgJTA9dTYOT2LIK9CBUzqadOBmw0c5Lz8,240885
+sqlalchemy/sql/sqltypes.py,sha256=yXHvZXfZJmaRvMoX4_jXqazAev33pk0Ltwl5c-D5Ha4,128609
+sqlalchemy/sql/traversals.py,sha256=7GALHt5mFceUv2SMUikIdAb9SUcSbACqhwoei5rPkxc,33664
+sqlalchemy/sql/type_api.py,sha256=wdi3nmOBRdhG6L1z21V_PwQGB8CIRouMdNKoIzJA4Zo,84440
+sqlalchemy/sql/util.py,sha256=G-2ZI6rZ7XxVu5YXaVvLrApeAk5VwSG4C--lqtglgGE,48086
+sqlalchemy/sql/visitors.py,sha256=URpw-GxxUkwjEDbD2xXJGyFJavG5lN6ISoY34JlYRS8,36319
+sqlalchemy/testing/__init__.py,sha256=GgUEqxUNCxg-92_GgBDnljUHsdCxaGPMG1TWy5tjwgk,3160
+sqlalchemy/testing/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/assertions.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/assertsql.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/asyncio.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/config.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/engines.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/entities.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/exclusions.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/pickleable.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/profiling.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/provision.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/requirements.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/schema.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/util.cpython-312.pyc,,
+sqlalchemy/testing/__pycache__/warnings.cpython-312.pyc,,
+sqlalchemy/testing/assertions.py,sha256=RFTkxGq-kDvn3JSUuT_6bU1y0vtoI6pE6ryZgV2YEx4,31439
+sqlalchemy/testing/assertsql.py,sha256=cmhtZrgPBjrqIfzFz3VBWxVNvxWoRllvmoWcUCoqsio,16817
+sqlalchemy/testing/asyncio.py,sha256=QsMzDWARFRrpLoWhuYqzYQPTUZ80fymlKrqOoDkmCmQ,3830
+sqlalchemy/testing/config.py,sha256=HySdB5_FgCW1iHAJVxYo-4wq5gUAEi0N8E93IC6M86Q,12058
+sqlalchemy/testing/engines.py,sha256=c1gFXfpo5S1dvNjGIL03mbW2eVYtUD_9M_ZEfQO2ArM,13414
+sqlalchemy/testing/entities.py,sha256=KdgTVPSALhi9KkAXj2giOYl62ld-1yZziIDBSV8E3vw,3354
+sqlalchemy/testing/exclusions.py,sha256=jzVrBXqyQlyMgvfChMjJOd0ZtReKgkJ4Ik-0mkWe6KM,12460
+sqlalchemy/testing/fixtures/__init__.py,sha256=e5YtfSlkKDRuyIZhEKBCycMX5BOO4MZ-0d97l1JDhJE,1198
+sqlalchemy/testing/fixtures/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/testing/fixtures/__pycache__/base.cpython-312.pyc,,
+sqlalchemy/testing/fixtures/__pycache__/mypy.cpython-312.pyc,,
+sqlalchemy/testing/fixtures/__pycache__/orm.cpython-312.pyc,,
+sqlalchemy/testing/fixtures/__pycache__/sql.cpython-312.pyc,,
+sqlalchemy/testing/fixtures/base.py,sha256=y5iEEdUZIft06fvAOXwKU73ciIFTO5AVgDDGzYD9nOY,12256
+sqlalchemy/testing/fixtures/mypy.py,sha256=9fuJ90F9LBki26dVEVOEtRVXG2koaK803k4nukTnA8o,11973
+sqlalchemy/testing/fixtures/orm.py,sha256=3JJoYdI2tj5-LL7AN8bVa79NV3Guo4d9p6IgheHkWGc,6095
+sqlalchemy/testing/fixtures/sql.py,sha256=ht-OD6fMZ0inxucRzRZG4kEMNicqY8oJdlKbZzHhAJc,15900
+sqlalchemy/testing/pickleable.py,sha256=G3L0xL9OtbX7wThfreRjWd0GW7q0kUKcTUuCN5ETGno,2833
+sqlalchemy/testing/plugin/__init__.py,sha256=vRfF7M763cGm9tLQDWK6TyBNHc80J1nX2fmGGxN14wY,247
+sqlalchemy/testing/plugin/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/testing/plugin/__pycache__/bootstrap.cpython-312.pyc,,
+sqlalchemy/testing/plugin/__pycache__/plugin_base.cpython-312.pyc,,
+sqlalchemy/testing/plugin/__pycache__/pytestplugin.cpython-312.pyc,,
+sqlalchemy/testing/plugin/bootstrap.py,sha256=VYnVSMb-u30hGY6xGn6iG-LqiF0CubT90AJPFY_6UiY,1685
+sqlalchemy/testing/plugin/plugin_base.py,sha256=TBWdg2XgXB6QgUUFdKLv1O9-SXMitjHLm2rNNIzXZhQ,21578
+sqlalchemy/testing/plugin/pytestplugin.py,sha256=0rRCp7RlnhJBg3gJEq0t0kJ-BCTQ34bqBE_lEQk5U3U,27656
+sqlalchemy/testing/profiling.py,sha256=SWhWiZImJvDsNn0rQyNki70xdNxZL53ZI98ihxiykbQ,10148
+sqlalchemy/testing/provision.py,sha256=6r2FTnm-t7u8MMbWo7eMhAH3qkL0w0WlmE29MUSEIu4,14702
+sqlalchemy/testing/requirements.py,sha256=MVuTKtZjeTZaYlrAU8XFIB1bhJA_AedqL_q7NwVEGiw,52956
+sqlalchemy/testing/schema.py,sha256=IImFumAdpzOyoKAs0WnaGakq8D3sSU4snD9W4LVOV3s,6513
+sqlalchemy/testing/suite/__init__.py,sha256=S8TLwTiif8xX67qlZUo5I9fl9UjZAFGSzvlptp2WoWc,722
+sqlalchemy/testing/suite/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_cte.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_ddl.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_deprecations.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_dialect.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_insert.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_reflection.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_results.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_rowcount.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_select.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_sequence.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_types.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_unicode_ddl.cpython-312.pyc,,
+sqlalchemy/testing/suite/__pycache__/test_update_delete.cpython-312.pyc,,
+sqlalchemy/testing/suite/test_cte.py,sha256=d3OWDBNhnAwlyAz_QhFk-vKSWaAI3mADVnqdtTWOuwI,6451
+sqlalchemy/testing/suite/test_ddl.py,sha256=MItp-votCzvahlRqHRagte2Omyq9XUOFdFsgzCb6_-g,12031
+sqlalchemy/testing/suite/test_deprecations.py,sha256=7C6IbxRmq7wg_DLq56f1V5RCS9iVrAv3epJZQTB-dOo,5337
+sqlalchemy/testing/suite/test_dialect.py,sha256=eGJFZCwKmLrIl66ZlkLLZf5Fq6bzWI174gQsJt2bY2c,22923
+sqlalchemy/testing/suite/test_insert.py,sha256=pR0VWMQ9JJPbnANE6634PzR0VFmWMF8im6OTahc4vsQ,18824
+sqlalchemy/testing/suite/test_reflection.py,sha256=EJvTjRDimw9k90zlI5VCkmCzf7Tv5VF9y4O3D8SZMFU,109648
+sqlalchemy/testing/suite/test_results.py,sha256=9FFBNLeXcNRIC9FHfEjFKwfV6w2Bb58ulml_M8Zdokg,16914
+sqlalchemy/testing/suite/test_rowcount.py,sha256=UVyHHQsU0TxkzV_dqCOKR1aROvIq7frKYMVjwUqLWfE,7900
+sqlalchemy/testing/suite/test_select.py,sha256=S81w-Dox6W29Tjmi6LIBJ4HuB5E8dDAzmePDm0PKTYo,61732
+sqlalchemy/testing/suite/test_sequence.py,sha256=DMqyJkL1o4GClrNjzoy7GDn_jPNPTZNvk9t5e-MVXeo,9923
+sqlalchemy/testing/suite/test_types.py,sha256=gPA6t-90Icnpj2ZzITwbqka1DB-rNOoh6_xS9dC-4HU,67805
+sqlalchemy/testing/suite/test_unicode_ddl.py,sha256=0zVc2e3zbCQag_xL4b0i7F062HblHwV46JHLMweYtcE,6141
+sqlalchemy/testing/suite/test_update_delete.py,sha256=_OxH0wggHUqPImalGEPI48RiRx6mO985Om1PtRYOCzA,3994
+sqlalchemy/testing/util.py,sha256=KsUInolFBXUPIXVZKAdb_8rQrW8yW8OCtiA3GXuYRvA,14571
+sqlalchemy/testing/warnings.py,sha256=sj4vfTtjodcfoX6FPH_Zykb4fomjmgqIYj81QPpSwH8,1546
+sqlalchemy/types.py,sha256=m3I9h6xoyT7cjeUx5XCzmaE-GHT2sJVwECiuSJl75Ss,3168
+sqlalchemy/util/__init__.py,sha256=tYWkZV6PYVfEW32zt48FCLH12VyV_kaNUa3KBAOYpSM,8312
+sqlalchemy/util/__pycache__/__init__.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/_collections.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/_concurrency_py3k.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/_has_cy.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/_py_collections.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/compat.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/concurrency.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/deprecations.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/langhelpers.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/preloaded.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/queue.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/tool_support.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/topological.cpython-312.pyc,,
+sqlalchemy/util/__pycache__/typing.cpython-312.pyc,,
+sqlalchemy/util/_collections.py,sha256=RbP4UixqNtRBUrl_QqYDiadVmELSVxxXm2drhvQaIKo,20078
+sqlalchemy/util/_concurrency_py3k.py,sha256=UtPDkb67OOVWYvBqYaQgENg0k_jOA2mQOE04XmrbYq0,9170
+sqlalchemy/util/_has_cy.py,sha256=3oh7s5iQtW9qcI8zYunCfGAKG6fzo2DIpzP5p1BnE8Q,1247
+sqlalchemy/util/_py_collections.py,sha256=irOg3nkzxmtdYfIS46un2cp0JqSiACI7WGQBg-BaEXU,16714
+sqlalchemy/util/compat.py,sha256=TdDfvL21VnBEdSUnjcx-F8XhmVFg9Mvyr67a4omWZAM,8760
+sqlalchemy/util/concurrency.py,sha256=eQVS3YDH3GwB3Uw5pbzmqEBSYTK90EbnE5mQ05fHERg,3304
+sqlalchemy/util/deprecations.py,sha256=L7D4GqeIozpjO8iVybf7jL9dDlgfTbAaQH4TQAX74qE,12012
+sqlalchemy/util/langhelpers.py,sha256=G67avnsStFbslILlbCHmsyAMnShS7RYftFr9a8uFDL8,65140
+sqlalchemy/util/preloaded.py,sha256=RMarsuhtMW8ZuvqLSuR0kwbp45VRlzKpJMLUe7p__qY,5904
+sqlalchemy/util/queue.py,sha256=w1ufhuiC7lzyiZDhciRtRz1uyxU72jRI7SWhhL-p600,10185
+sqlalchemy/util/tool_support.py,sha256=e7lWu6o1QlKq4e6c9PyDsuyFyiWe79vO72UQ_YX2pUA,6135
+sqlalchemy/util/topological.py,sha256=HcJgdCeU0XFIskgIBnTaHXfRXaulaEJRYRwKv4yPNek,3458
+sqlalchemy/util/typing.py,sha256=C4jF7QTNo0w0bjvcIqSSTOvoy8FttuZtyTzjiyoIzQQ,20920
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/WHEEL
new file mode 100644
index 00000000..d37c50b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: false
+Tag: cp312-cp312-macosx_11_0_arm64
+
diff --git a/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/top_level.txt
new file mode 100644
index 00000000..39fb2bef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/SQLAlchemy-2.0.38.dist-info/top_level.txt
@@ -0,0 +1 @@
+sqlalchemy
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/METADATA
new file mode 100644
index 00000000..1523c0fb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/METADATA
@@ -0,0 +1,78 @@
+Metadata-Version: 2.2
+Name: Tree
+Version: 0.2.4
+Summary: A package for creating and drawing trees
+Home-page: https://github.com/PixelwarStudio/PyTree
+Author: Pixelwar
+License: MIT
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.5
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Requires-Dist: Pillow
+Requires-Dist: svgwrite
+Requires-Dist: setuptools
+Requires-Dist: click
+Dynamic: author
+Dynamic: classifier
+Dynamic: description
+Dynamic: home-page
+Dynamic: license
+Dynamic: requires-dist
+Dynamic: summary
+
+PyTree
+======
+.. image:: https://img.shields.io/pypi/v/tree.svg?style=flat-square
+        :target: https://pypi.python.org/pypi/Tree
+
+.. image:: https://img.shields.io/pypi/l/Tree.svg?style=flat-square
+        :target: https://github.com/PixelwarStudio/PyTree/blob/master/LICENSE
+
+Python package, which you can use to generate and drawing trees, realistic or fractal ones.
+
+Usage
+-----
+.. code-block:: bash
+
+    $ pip install Tree
+
+.. code-block:: python
+
+    from math import pi, radians as rad
+    from Tree.core import Tree
+    from PIL import Image
+
+    branches = (
+        (.5, rad(-30)),
+        (.6, rad(30)),
+        (.4, rad(60))
+    )
+
+    def main():
+        tree = Tree(
+            pos=(0, 0, 0, -500),
+            branches=branches
+        )
+        
+        # Let the tree grow
+        tree.grow(10)
+        
+        # Move the tree in the right position, so that the tree is completly in the image
+        tree.move_in_rectangle()
+
+        im = Image.new("RGB", tree.get_size(), (239, 239, 239))
+        tree.draw_on(im, (85, 25, 0, 128, 53, 21), (0, 62, 21), 10)
+        im.show()
+
+    if __name__ == '__main__':
+        main()
+    
+.. image:: https://raw.githubusercontent.com/PixelwarStudio/PyTree/master/images/example.png
+
+Documentation
+-------------
+The documentation is hosted on Readthedocs_.
+
+.. _Readthedocs: http://pytree.readthedocs.io/en/latest/ 
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/RECORD
new file mode 100644
index 00000000..55fc4b2c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/RECORD
@@ -0,0 +1,18 @@
+../../../bin/tree-cli,sha256=bq3t1rS64OIShAOhgj7dSYX6PEflaKzepoYu_T43jyY,278
+Tree-0.2.4.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+Tree-0.2.4.dist-info/METADATA,sha256=3zfHPkG3XcVs9bAa0QaZCdpAXMGgNixDwl1Je48pJjo,2031
+Tree-0.2.4.dist-info/RECORD,,
+Tree-0.2.4.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+Tree-0.2.4.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+Tree-0.2.4.dist-info/entry_points.txt,sha256=WF51H4IX4aB_l3kxVJJ1nkJvy4mE-iIkIZxCH_0Q-hc,50
+Tree-0.2.4.dist-info/top_level.txt,sha256=KttPIq5RpLmuJYiP9KUd1Z-WGafr_N-OITgHKHDj9No,5
+Tree/__init__.py,sha256=WPzF115nNj52MzCV85miYA7tQPJGUUduPz_VRceH-W8,95
+Tree/__pycache__/__init__.cpython-312.pyc,,
+Tree/__pycache__/cli.cpython-312.pyc,,
+Tree/__pycache__/core.cpython-312.pyc,,
+Tree/__pycache__/draw.cpython-312.pyc,,
+Tree/__pycache__/utils.cpython-312.pyc,,
+Tree/cli.py,sha256=MkLH98wuXcaGFjBTEbDJhiFLNLp8dFIgvC0bYw4nUwA,2234
+Tree/core.py,sha256=99xFaOCRiGqyXmCeMA_lq_GZ5IK57ghNaSKrwWwDhf4,7845
+Tree/draw.py,sha256=2v_PLERMZArrFgU9fb8j8WDO8QiCs_DffliA4lvxfiU,3870
+Tree/utils.py,sha256=1gjSbHy4S6DeGUqpOSmLlJwAmyiR_l7GpAp80C_fUVE,2137
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/WHEEL
new file mode 100644
index 00000000..505164bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/entry_points.txt
new file mode 100644
index 00000000..810be061
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/entry_points.txt
@@ -0,0 +1,2 @@
+[console_scripts]
+tree-cli = Tree.cli:create_tree
diff --git a/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/top_level.txt
new file mode 100644
index 00000000..37da77b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree-0.2.4.dist-info/top_level.txt
@@ -0,0 +1 @@
+Tree
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__init__.py b/backend/venv/lib/python3.12/site-packages/Tree/__init__.py
new file mode 100644
index 00000000..7f39485b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree/__init__.py
@@ -0,0 +1,5 @@
+"""
+Package for creating and drawing trees.
+"""
+__version__ = "v0.2.4"
+__author__ = "Pixelwar"
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..c79a74b1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/cli.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/cli.cpython-312.pyc
new file mode 100644
index 00000000..7f353efe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/cli.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..1315e1b3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/draw.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/draw.cpython-312.pyc
new file mode 100644
index 00000000..e8e4a9ff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/draw.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..6b75f292
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/Tree/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/cli.py b/backend/venv/lib/python3.12/site-packages/Tree/cli.py
new file mode 100644
index 00000000..4b1c1674
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree/cli.py
@@ -0,0 +1,62 @@
+from math import radians
+
+import click
+
+from PIL import Image
+import svgwrite
+
+from Tree.core import Tree
+
+def get_format(path):
+    pos = path.find(".")
+
+    if pos == -1:
+        return False
+    return path[pos+1:]
+
+@click.command()
+@click.option("--length", "-l", help="The start length of tree.", type=float, default=300)
+@click.option("--branches", "-b", help="Add a branch with a scale and a angle.", multiple=True, type=(float, int), default=[[.5, 45], [.5, -45]])
+@click.option("--sigma", "-s", help="Add randomness to scale and angle.", nargs=2, type=float, default=(0, 0))
+@click.option("--age", "-a", help="Indicates how many time the tree should be iterated.",type=int, default=5)
+@click.option("--path", "-p", help="The path for saving the tree. Multiple formats supported e.g. svg.", default=None)
+@click.option("--show", help="Shows a image of the tree.", is_flag=True)
+@click.option("--stem_color1", "-sc1", help="The stem start color given as r g b", nargs=3, type=int, default=(255, 0, 255))
+@click.option("--stem_color2", "-sc2", help="The stem end color given as r g b", nargs=3, type=int, default=(255, 0, 255))
+@click.option("--leaf_color", "-lc", help="The leaf color given as r g b", nargs=3, type=int, default=(255, 255, 255))
+@click.option("--thickness", "-t", help="The start width of the first branch.", type=int, default=5)
+
+def create_tree(length, branches, sigma, age, path, show, stem_color1, stem_color2, leaf_color, thickness):
+    stem_color = stem_color1+stem_color2
+    options = [
+        stem_color,
+        leaf_color,
+        thickness
+    ]
+    
+    #Convert angles to radians
+    branches = [[branch[0], radians(branch[1])] for branch in branches]
+
+    tree = Tree((0, 0, 0, -length), branches, sigma)
+    tree.grow(times=age)
+    tree.move_in_rectangle()
+
+    form = get_format(path) if path is not None else None
+
+    if show or form not in ("svg", None):
+        im = Image.new("RGB", tree.get_size())
+        tree.draw_on(im, *options)
+
+    if form == "svg":
+        svg = svgwrite.Drawing(path)
+        tree.draw_on(svg, *options)
+        svg.save()
+
+    if form not in ("svg", None):
+        im.save(path)
+
+    if show:
+        im.show()
+
+if __name__ == "__main__":
+    create_tree()
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/core.py b/backend/venv/lib/python3.12/site-packages/Tree/core.py
new file mode 100644
index 00000000..40e64178
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree/core.py
@@ -0,0 +1,240 @@
+"""
+Module for creating trees
+"""
+from math import pi, log, sqrt
+from random import gauss
+from Tree.utils import Node
+from Tree.draw import SUPPORTED_CANVAS
+
+class Tree:
+    """The standard tree."""
+    def __init__(self, pos=(0, 0, 0, -100), branches=None, sigma=(0, 0)):
+        """The contructor.
+
+        Args:
+            pos (tupel): A tupel, holding the start and end point of the tree. (x1, y1, x2, y2)
+            branches (tupel/array): Holding array/s with scale and angle for every branch.
+            sigma (tuple): Holding the branch and angle sigma. e.g.(0.1, 0.2)
+        """
+        self.pos = pos
+        self.length = sqrt((pos[2]-pos[0])**2+(pos[3]-pos[1])**2)
+        self.branches = branches
+        self.sigma = sigma
+
+        self.comp = len(self.branches)
+        self.age = 0
+
+        self.nodes = [
+            [Node(pos[2:])]
+        ]
+
+    def get_rectangle(self):
+        """Gets the coordinates of the rectangle, in which the tree can be put.
+
+        Returns:
+            tupel: (x1, y1, x2, y2)
+        """
+        rec = [self.pos[0], self.pos[1]]*2
+        for age in self.nodes:
+            for node in age:
+                # Check max/min for x/y coords
+                for i in range(2):
+                    if rec[0+i] > node.pos[i]:
+                        rec[0+i] = node.pos[i]
+                    elif rec[2+i] < node.pos[i]:
+                        rec[2+i] = node.pos[i]
+        return tuple(rec)
+
+    def get_size(self):
+        """Get the size of the tree.
+
+        Returns:
+            tupel: (width, height)
+        """
+        rec = self.get_rectangle()
+        return (int(rec[2]-rec[0]), int(rec[3]-rec[1]))
+
+    def get_branch_length(self, age=None, pos=0):
+        """Get the length of a branch.
+
+        This method calculates the length of a branch in specific age.
+        The used formula: length * scale^age.
+
+        Args:
+            age (int): The age, for which you want to know the branch length.
+        Returns:
+            float: The length of the branch
+        """
+        if age is None:
+            age = self.age
+
+        return self.length * pow(self.branches[pos][0], age)
+
+    def get_steps_branch_len(self, length):
+        """Get, how much steps will needed for a given branch length.
+
+        Returns:
+            float: The age the tree must achieve to reach the given branch length.
+        """
+        return log(length/self.length, min(self.branches[0][0]))
+
+    def get_node_sum(self, age=None):
+        """Get sum of all branches in the tree.
+
+        Returns:
+            int: The sum of all nodes grown until the age.
+        """
+        if age is None:
+            age = self.age
+
+        return age if self.comp == 1 else int((pow(self.comp, age+1) - 1) / (self.comp - 1))
+
+    def get_node_age_sum(self, age=None):
+        """Get the sum of branches grown in an specific age.
+
+        Returns:
+            int: The sum of all nodes grown in an age.
+        """
+        if age is None:
+            age = self.age
+
+        return pow(self.comp, age)
+
+    def get_nodes(self):
+        """Get the tree nodes as list.
+
+        Returns:
+            list: A 2d-list holding the grown nodes coordinates as tupel for every age.
+                Example:
+                [
+                [(10, 40)],
+                [(20, 80), (100, 30)],
+                [(100, 90), (120, 40), ...],
+                ...
+                ]
+        """
+        nodes = []
+        for age, level in enumerate(self.nodes):
+            nodes.append([])
+            for node in level:
+                nodes[age].append(node.get_tuple())
+        return nodes
+
+    def get_branches(self):
+        """Get the tree branches as list.
+
+        Returns:
+            list: A 2d-list holding the grown branches coordinates as tupel for every age.
+                Example:
+                [
+                [(10, 40, 90, 30)],
+                [(90, 30, 100, 40), (90, 30, 300, 60)],
+                [(100, 40, 120, 70), (100, 40, 150, 90), ...],
+                ...
+                ]
+        """
+        branches = []
+        for age, level in enumerate(self.nodes):
+            branches.append([])
+            for n, node in enumerate(level):
+                if age == 0:
+                    p_node = Node(self.pos[:2])
+                else:
+                    p_node = self._get_node_parent(age-1, n)
+                branches[age].append(p_node.get_tuple() + node.get_tuple())
+
+        return branches
+
+    def move(self, delta):
+        """Move the tree.
+
+        Args:
+            delta (tupel): The adjustment of the position.
+        """
+        pos = self.pos
+        self.pos = (pos[0]+delta[0], pos[1]+delta[1], pos[2]+delta[0], pos[3]+delta[1])
+
+        # Move all nodes
+        for age in self.nodes:
+            for node in age:
+                node.move(delta)
+
+    def move_in_rectangle(self):
+        """Move the tree so that the tree fits in the rectangle."""
+        rec = self.get_rectangle()
+        self.move((-rec[0], -rec[1]))
+
+    def grow(self, times=1):
+        """Let the tree grow.
+
+        Args:
+            times (integer): Indicate how many times the tree will grow.
+        """
+        self.nodes.append([])
+
+        for n, node in enumerate(self.nodes[self.age]):
+            if self.age == 0:
+                p_node = Node(self.pos[:2])
+            else:
+                p_node = self._get_node_parent(self.age-1, n)
+            angle = node.get_node_angle(p_node)
+            for i in range(self.comp):
+                tot_angle = self.__get_total_angle(angle, i)
+                length = self.__get_total_length(self.age+1, i)
+                self.nodes[self.age+1].append(node.make_new_node(length, tot_angle))
+
+        self.age += 1
+
+        if times > 1:
+            self.grow(times-1)
+
+    def draw_on(self, canvas, stem_color, leaf_color, thickness, ages=None):
+        """Draw the tree on a canvas.
+
+        Args:
+            canvas (object): The canvas, you want to draw the tree on. Supported canvases: svgwrite.Drawing and PIL.Image (You can also add your custom libraries.)
+            stem_color (tupel): Color or gradient for the stem of the tree.
+            leaf_color (tupel): Color for the leaf (= the color for last iteration).
+            thickness (int): The start thickness of the tree.
+        """
+        if canvas.__module__ in SUPPORTED_CANVAS:
+            drawer = SUPPORTED_CANVAS[canvas.__module__]
+            drawer(self, canvas, stem_color, leaf_color, thickness, ages).draw()
+
+    def __get_total_angle(self, angle, pos):
+        """Get the total angle."""
+        tot_angle = angle - self.branches[pos][1]
+        if self.sigma[1] != 0:
+            tot_angle += gauss(0, self.sigma[1]) * pi
+        return tot_angle
+
+    def __get_total_length(self, age, pos):
+        length = self.get_branch_length(age, pos)
+        if self.sigma[0] != 0:
+            length *= (1+gauss(0, self.sigma[0]))
+        return length
+
+    def _get_node_parent(self, age, pos):
+        """Get the parent node of node, whch is located in tree's node list.
+
+        Returns:
+            object: The parent node.
+        """
+        return self.nodes[age][int(pos / self.comp)]
+
+def generate_branches(scales=None, angles=None, shift_angle=0):
+    """Generates branches with alternative system.
+
+    Args:
+        scales (tuple/array): Indicating how the branch/es length/es develop/s from age to age.
+        angles (tuple/array): Holding the branch and shift angle in radians.
+        shift_angle (float): Holding the rotation angle for all branches.
+
+    Returns:
+        branches (2d-array): A array constits of arrays holding scale and angle for every branch.
+    """
+    branches = []
+    for pos, scale in enumerate(scales):
+        angle = -sum(angles)/2 + sum(angles[:pos]) + shift_angle
+        branches.append([scale, angle])
+    return branches
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/draw.py b/backend/venv/lib/python3.12/site-packages/Tree/draw.py
new file mode 100644
index 00000000..f7c072ec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree/draw.py
@@ -0,0 +1,123 @@
+"""
+Module for drawing trees.
+"""
+from PIL import ImageDraw
+import svgwrite
+from Tree.utils import convert_color
+
+class Drawer(object):
+    """A generic class for drawing tree on acanvas."""
+    def __init__(self, tree, canvas, stem_color=(255, 255, 255), leaf_color=(230, 120, 34), thickness=1, ages=None):
+        """Constructor of drawer.
+
+        Args:
+            tree (object): The tree, which should drawn on canvas.
+            canvas (object): The canvas for drawing the tree.
+            stem_color (tupel): Color or gradient for the steam of the tree.
+            leaf_color (tupel): Color for the leaf (= the color for last iteration).
+            thickness (int): The start thickness of the tree.
+            ages (array): Contains the ages you want to draw.
+
+        Returns:
+            int: The thickness of the branch/es
+        """
+        self.canvas = canvas
+        self.tree = tree
+        self.stem_color = stem_color
+        self.leaf_color = leaf_color
+        self.thickness = thickness
+        self.ages = range(tree.age+1) if ages is None else ages
+
+    def _get_thickness(self, age):
+        """Get the thickness depending on age.
+
+        Args:
+            age (int): The age of the branch/es
+
+        Returns:
+            int: The thickness of the branch/es
+        """
+        return int((self.thickness*5)/(age+5))
+
+    def _get_color(self, age):
+        """Get the fill color depending on age.
+
+        Args:
+            age (int): The age of the branch/es
+
+        Returns:
+            tuple: (r, g, b)
+        """
+        if age == self.tree.age:
+            return self.leaf_color
+        color = self.stem_color
+        tree = self.tree
+
+        if len(color) == 3:
+            return color
+
+        diff = [color[i+3]-color[i] for i in range(3)]
+        per_age = [diff[i]/(tree.age-1) for i in range(3)]
+
+        return tuple([int(color[i]+per_age[i]*age) for i in range(3)])
+
+    def _draw_branch(self, branch, color, thickness, age):
+        """Placeholder for specific draw methods for a branch.
+
+        Args:
+            branch (tupel): The coordinates of the branch.
+            color (tupel): The color of the branch.
+            thickness (int): The thickness of the branch.
+            age (int): The age of the tree the branch is drawn.
+        """
+        pass
+
+    def draw(self):
+        """Draws the tree.
+
+        Args:
+            ages (array): Contains the ages you want to draw.
+        """
+        for age, level in enumerate(self.tree.get_branches()):
+            if age in self.ages:
+                thickness = self._get_thickness(age)
+                color = self._get_color(age)
+                for branch in level:
+                    self._draw_branch(branch, color, thickness, age)
+
+class PilDrawer(Drawer):
+    """A drawer class for drawing on PIL/Pillow images."""
+    def _draw_branch(self, branch, color, thickness, age):
+        ImageDraw.Draw(self.canvas).line(branch, color, thickness)
+
+class SvgDrawer(Drawer):
+    """A drawer class for drawing on svg documents.
+
+    Attributes:
+        group (list): Saves the groups created for every age.
+    """
+    def __init__(self, tree, canvas, color=(255, 255, 255), thickness=1):
+        super(SvgDrawer, self).__init__(tree, canvas, color, thickness)
+        self.group = []
+
+    def _draw_branch(self, branch, color, thickness, age):
+        color = convert_color(color)
+        self.group[age].add(
+            self.canvas.line(
+                start=branch[:2],
+                end=branch[2:],
+                stroke=color,
+                stroke_width=thickness
+            )
+        )
+
+    def draw(self):
+        self.group = []
+        for _ in self.ages:
+            self.group.append(self.canvas.add(svgwrite.container.Group()))
+        Drawer.draw(self)
+
+SUPPORTED_CANVAS = {
+    "PIL.Image": PilDrawer,
+    "svgwrite.drawing": SvgDrawer
+}
diff --git a/backend/venv/lib/python3.12/site-packages/Tree/utils.py b/backend/venv/lib/python3.12/site-packages/Tree/utils.py
new file mode 100644
index 00000000..a56bf22f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/Tree/utils.py
@@ -0,0 +1,78 @@
+"""
+Helper module.
+"""
+from math import atan2, cos, sin, pi, sqrt
+
+def convert_color(color):
+    """Convert color tupel(r, g, b) to string("rgb({r}, {g}, {b}").
+
+    Args:
+        color (tupel): RGB color. e.g. (134, 8, 45)
+
+    Returns:
+        string: "rgb({r}, {g}, {b}"
+    """
+    return "rgb({}, {}, {})".format(color[0], color[1], color[2])
+
+class Node(object):
+    """A node.
+
+    Attributes:
+        pos (tupel): The position of the node. (x, y)
+    """
+    def __init__(self, pos):
+        self.pos = pos
+
+    def make_new_node(self, distance, angle):
+        """Make a new node from an existing one.
+
+        This method creates a new node with a distance and angle given.
+        The position of the new node is calculated with:
+        x2 = cos(-angle)*distance+x1
+        y2 = sin(-angle)*distance+y1
+
+        Args:
+            distance (float): The distance of the original node to the new node.
+            angle (rad): The angle between the old and new node, relative to the horizont.
+
+        Returns:
+            object: The node with calculated poistion.
+        """
+        return Node((cos(-angle)*distance+self.pos[0],
+                     sin(-angle)*distance+self.pos[1]))
+
+    def get_node_angle(self, node):
+        """Get the angle beetween 2 nodes relative to the horizont.
+
+        Args:
+            node (object): The other node.
+
+        Returns:
+            rad: The angle
+        """
+        return atan2(self.pos[0]-node.pos[0], self.pos[1]-node.pos[1]) - pi / 2
+
+    def get_distance(self, node):
+        """Get the distance beetween 2 nodes
+
+        Args:
+            node (object): The other node.
+        """
+        delta = (node.pos[0]-self.pos[0], node.pos[1]-self.pos[1])
+        return sqrt(delta[0]**2+delta[1]**2)
+
+    def get_tuple(self):
+        """Get the position of the node as tuple.
+
+        Returns:
+            tupel: (x, y)
+        """
+        return self.pos
+
+    def move(self, delta):
+        """Move the node.
+
+        Args:
+            delta (tupel): A tupel, holding the adjustment of the position.
+        """
+        self.pos = (self.pos[0]+delta[0], self.pos[1]+delta[1])
diff --git a/backend/venv/lib/python3.12/site-packages/__pycache__/parse.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/__pycache__/parse.cpython-312.pyc
new file mode 100644
index 00000000..417ae9dc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/__pycache__/parse.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/__pycache__/py.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/__pycache__/py.cpython-312.pyc
new file mode 100644
index 00000000..80690508
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/__pycache__/py.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/__pycache__/six.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/__pycache__/six.cpython-312.pyc
new file mode 100644
index 00000000..255cff27
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/__pycache__/six.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/__pycache__/typing_extensions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/__pycache__/typing_extensions.cpython-312.pyc
new file mode 100644
index 00000000..aa0b8efa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/__pycache__/typing_extensions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_distutils_hack/__init__.py b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__init__.py
new file mode 100644
index 00000000..94f71b99
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__init__.py
@@ -0,0 +1,239 @@
+# don't import any costly modules
+import os
+import sys
+
+report_url = (
+    "https://github.com/pypa/setuptools/issues/new?template=distutils-deprecation.yml"
+)
+
+
+def warn_distutils_present():
+    if 'distutils' not in sys.modules:
+        return
+    import warnings
+
+    warnings.warn(
+        "Distutils was imported before Setuptools, but importing Setuptools "
+        "also replaces the `distutils` module in `sys.modules`. This may lead "
+        "to undesirable behaviors or errors. To avoid these issues, avoid "
+        "using distutils directly, ensure that setuptools is installed in the "
+        "traditional way (e.g. not an editable install), and/or make sure "
+        "that setuptools is always imported before distutils."
+    )
+
+
+def clear_distutils():
+    if 'distutils' not in sys.modules:
+        return
+    import warnings
+
+    warnings.warn(
+        "Setuptools is replacing distutils. Support for replacing "
+        "an already imported distutils is deprecated. In the future, "
+        "this condition will fail. "
+        f"Register concerns at {report_url}"
+    )
+    mods = [
+        name
+        for name in sys.modules
+        if name == "distutils" or name.startswith("distutils.")
+    ]
+    for name in mods:
+        del sys.modules[name]
+
+
+def enabled():
+    """
+    Allow selection of distutils by environment variable.
+    """
+    which = os.environ.get('SETUPTOOLS_USE_DISTUTILS', 'local')
+    if which == 'stdlib':
+        import warnings
+
+        warnings.warn(
+            "Reliance on distutils from stdlib is deprecated. Users "
+            "must rely on setuptools to provide the distutils module. "
+            "Avoid importing distutils or import setuptools first, "
+            "and avoid setting SETUPTOOLS_USE_DISTUTILS=stdlib. "
+            f"Register concerns at {report_url}"
+        )
+    return which == 'local'
+
+
+def ensure_local_distutils():
+    import importlib
+
+    clear_distutils()
+
+    # With the DistutilsMetaFinder in place,
+    # perform an import to cause distutils to be
+    # loaded from setuptools._distutils. Ref #2906.
+    with shim():
+        importlib.import_module('distutils')
+
+    # check that submodules load as expected
+    core = importlib.import_module('distutils.core')
+    assert '_distutils' in core.__file__, core.__file__
+    assert 'setuptools._distutils.log' not in sys.modules
+
+
+def do_override():
+    """
+    Ensure that the local copy of distutils is preferred over stdlib.
+
+    See https://github.com/pypa/setuptools/issues/417#issuecomment-392298401
+    for more motivation.
+    """
+    if enabled():
+        warn_distutils_present()
+        ensure_local_distutils()
+
+
+class _TrivialRe:
+    def __init__(self, *patterns) -> None:
+        self._patterns = patterns
+
+    def match(self, string):
+        return all(pat in string for pat in self._patterns)
+
+
+class DistutilsMetaFinder:
+    def find_spec(self, fullname, path, target=None):
+        # optimization: only consider top level modules and those
+        # found in the CPython test suite.
+        if path is not None and not fullname.startswith('test.'):
+            return None
+
+        method_name = 'spec_for_{fullname}'.format(**locals())
+        method = getattr(self, method_name, lambda: None)
+        return method()
+
+    def spec_for_distutils(self):
+        if self.is_cpython():
+            return None
+
+        import importlib
+        import importlib.abc
+        import importlib.util
+
+        try:
+            mod = importlib.import_module('setuptools._distutils')
+        except Exception:
+            # There are a couple of cases where setuptools._distutils
+            # may not be present:
+            # - An older Setuptools without a local distutils is
+            #   taking precedence. Ref #2957.
+            # - Path manipulation during sitecustomize removes
+            #   setuptools from the path but only after the hook
+            #   has been loaded. Ref #2980.
+            # In either case, fall back to stdlib behavior.
+            return None
+
+        class DistutilsLoader(importlib.abc.Loader):
+            def create_module(self, spec):
+                mod.__name__ = 'distutils'
+                return mod
+
+            def exec_module(self, module):
+                pass
+
+        return importlib.util.spec_from_loader(
+            'distutils', DistutilsLoader(), origin=mod.__file__
+        )
+
+    @staticmethod
+    def is_cpython():
+        """
+        Suppress supplying distutils for CPython (build and tests).
+        Ref #2965 and #3007.
+        """
+        return os.path.isfile('pybuilddir.txt')
+
+    def spec_for_pip(self):
+        """
+        Ensure stdlib distutils when running under pip.
+        See pypa/pip#8761 for rationale.
+        """
+        if sys.version_info >= (3, 12) or self.pip_imported_during_build():
+            return
+        clear_distutils()
+        self.spec_for_distutils = lambda: None
+
+    @classmethod
+    def pip_imported_during_build(cls):
+        """
+        Detect if pip is being imported in a build script. Ref #2355.
+        """
+        import traceback
+
+        return any(
+            cls.frame_file_is_setup(frame) for frame, line in traceback.walk_stack(None)
+        )
+
+    @staticmethod
+    def frame_file_is_setup(frame):
+        """
+        Return True if the indicated frame suggests a setup.py file.
+        """
+        # some frames may not have __file__ (#2940)
+        return frame.f_globals.get('__file__', '').endswith('setup.py')
+
+    def spec_for_sensitive_tests(self):
+        """
+        Ensure stdlib distutils when running select tests under CPython.
+
+        python/cpython#91169
+        """
+        clear_distutils()
+        self.spec_for_distutils = lambda: None
+
+    sensitive_tests = (
+        [
+            'test.test_distutils',
+            'test.test_peg_generator',
+            'test.test_importlib',
+        ]
+        if sys.version_info < (3, 10)
+        else [
+            'test.test_distutils',
+        ]
+    )
+
+
+for name in DistutilsMetaFinder.sensitive_tests:
+    setattr(
+        DistutilsMetaFinder,
+        f'spec_for_{name}',
+        DistutilsMetaFinder.spec_for_sensitive_tests,
+    )
+
+
+DISTUTILS_FINDER = DistutilsMetaFinder()
+
+
+def add_shim():
+    DISTUTILS_FINDER in sys.meta_path or insert_shim()
+
+
+class shim:
+    def __enter__(self) -> None:
+        insert_shim()
+
+    def __exit__(self, exc: object, value: object, tb: object) -> None:
+        _remove_shim()
+
+
+def insert_shim():
+    sys.meta_path.insert(0, DISTUTILS_FINDER)
+
+
+def _remove_shim():
+    try:
+        sys.meta_path.remove(DISTUTILS_FINDER)
+    except ValueError:
+        pass
+
+
+if sys.version_info < (3, 12):
+    # DistutilsMetaFinder can only be disabled in Python < 3.12 (PEP 632)
+    remove_shim = _remove_shim
diff --git a/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..716bb07e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/override.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/override.cpython-312.pyc
new file mode 100644
index 00000000..6225b809
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_distutils_hack/__pycache__/override.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_distutils_hack/override.py b/backend/venv/lib/python3.12/site-packages/_distutils_hack/override.py
new file mode 100644
index 00000000..2cc433a4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_distutils_hack/override.py
@@ -0,0 +1 @@
+__import__('_distutils_hack').do_override()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/__init__.py
new file mode 100644
index 00000000..8eb8ec96
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/__init__.py
@@ -0,0 +1,13 @@
+from __future__ import annotations
+
+
+__all__ = ["__version__", "version_tuple"]
+
+try:
+    from ._version import version as __version__
+    from ._version import version_tuple
+except ImportError:  # pragma: no cover
+    # broken installation, we don't even try
+    # unknown only works because we do poor mans version compare
+    __version__ = "unknown"
+    version_tuple = (0, 0, "unknown")
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..84df311c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_argcomplete.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_argcomplete.cpython-312.pyc
new file mode 100644
index 00000000..31bfa3a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_argcomplete.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..81bbd5aa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/cacheprovider.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/cacheprovider.cpython-312.pyc
new file mode 100644
index 00000000..f0153c9c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/cacheprovider.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/capture.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/capture.cpython-312.pyc
new file mode 100644
index 00000000..1dbf6b49
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/capture.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..f803eaa9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/debugging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/debugging.cpython-312.pyc
new file mode 100644
index 00000000..e742c02d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/debugging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/deprecated.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/deprecated.cpython-312.pyc
new file mode 100644
index 00000000..1597671e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/deprecated.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/doctest.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/doctest.cpython-312.pyc
new file mode 100644
index 00000000..f9aef959
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/doctest.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/faulthandler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/faulthandler.cpython-312.pyc
new file mode 100644
index 00000000..d76092ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/faulthandler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/fixtures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/fixtures.cpython-312.pyc
new file mode 100644
index 00000000..d7fd041c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/fixtures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/freeze_support.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/freeze_support.cpython-312.pyc
new file mode 100644
index 00000000..50867df0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/freeze_support.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/helpconfig.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/helpconfig.cpython-312.pyc
new file mode 100644
index 00000000..b208b32c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/helpconfig.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/hookspec.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/hookspec.cpython-312.pyc
new file mode 100644
index 00000000..513cb424
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/hookspec.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/junitxml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/junitxml.cpython-312.pyc
new file mode 100644
index 00000000..f16c4e16
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/junitxml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/legacypath.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/legacypath.cpython-312.pyc
new file mode 100644
index 00000000..62e2e994
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/legacypath.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/logging.cpython-312.pyc
new file mode 100644
index 00000000..8ed8882c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/main.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/main.cpython-312.pyc
new file mode 100644
index 00000000..2dfbb62f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/main.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/monkeypatch.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/monkeypatch.cpython-312.pyc
new file mode 100644
index 00000000..ced42a9d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/monkeypatch.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/nodes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/nodes.cpython-312.pyc
new file mode 100644
index 00000000..a86588ce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/nodes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/outcomes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/outcomes.cpython-312.pyc
new file mode 100644
index 00000000..f152c263
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/outcomes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pastebin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pastebin.cpython-312.pyc
new file mode 100644
index 00000000..7891f749
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pastebin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pathlib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pathlib.cpython-312.pyc
new file mode 100644
index 00000000..b839873e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pathlib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester.cpython-312.pyc
new file mode 100644
index 00000000..e978bd2b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester_assertions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester_assertions.cpython-312.pyc
new file mode 100644
index 00000000..162240cb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/pytester_assertions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python.cpython-312.pyc
new file mode 100644
index 00000000..f86988f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_api.cpython-312.pyc
new file mode 100644
index 00000000..a062e1de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_path.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_path.cpython-312.pyc
new file mode 100644
index 00000000..0965cc70
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/python_path.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/recwarn.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/recwarn.cpython-312.pyc
new file mode 100644
index 00000000..1a326a09
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/recwarn.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/reports.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/reports.cpython-312.pyc
new file mode 100644
index 00000000..dea3af0e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/reports.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/runner.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/runner.cpython-312.pyc
new file mode 100644
index 00000000..b3538dbe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/runner.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/scope.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/scope.cpython-312.pyc
new file mode 100644
index 00000000..26ddf88c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/scope.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setuponly.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setuponly.cpython-312.pyc
new file mode 100644
index 00000000..265e8cd0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setuponly.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setupplan.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setupplan.cpython-312.pyc
new file mode 100644
index 00000000..77374436
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/setupplan.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/skipping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/skipping.cpython-312.pyc
new file mode 100644
index 00000000..bc19aa4f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/skipping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stash.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stash.cpython-312.pyc
new file mode 100644
index 00000000..69c89d47
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stash.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stepwise.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stepwise.cpython-312.pyc
new file mode 100644
index 00000000..69bd2d62
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/stepwise.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/terminal.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/terminal.cpython-312.pyc
new file mode 100644
index 00000000..47b46c3e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/terminal.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/threadexception.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/threadexception.cpython-312.pyc
new file mode 100644
index 00000000..dedc3b73
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/threadexception.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/timing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/timing.cpython-312.pyc
new file mode 100644
index 00000000..ab7425da
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/timing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/tmpdir.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/tmpdir.cpython-312.pyc
new file mode 100644
index 00000000..4b9466f6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/tmpdir.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unittest.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unittest.cpython-312.pyc
new file mode 100644
index 00000000..b06183c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unittest.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unraisableexception.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unraisableexception.cpython-312.pyc
new file mode 100644
index 00000000..25374ada
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/unraisableexception.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warning_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warning_types.cpython-312.pyc
new file mode 100644
index 00000000..6e2e1b9e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warning_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warnings.cpython-312.pyc
new file mode 100644
index 00000000..16484930
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/__pycache__/warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_argcomplete.py b/backend/venv/lib/python3.12/site-packages/_pytest/_argcomplete.py
new file mode 100644
index 00000000..59426ef9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_argcomplete.py
@@ -0,0 +1,117 @@
+"""Allow bash-completion for argparse with argcomplete if installed.
+
+Needs argcomplete>=0.5.6 for python 3.2/3.3 (older versions fail
+to find the magic string, so _ARGCOMPLETE env. var is never set, and
+this does not need special code).
+
+Function try_argcomplete(parser) should be called directly before
+the call to ArgumentParser.parse_args().
+
+The filescompleter is what you normally would use on the positional
+arguments specification, in order to get "dirname/" after "dirn<TAB>"
+instead of the default "dirname ":
+
+   optparser.add_argument(Config._file_or_dir, nargs='*').completer=filescompleter
+
+Other, application specific, completers should go in the file
+doing the add_argument calls as they need to be specified as .completer
+attributes as well. (If argcomplete is not installed, the function the
+attribute points to will not be used).
+
+SPEEDUP
+=======
+
+The generic argcomplete script for bash-completion
+(/etc/bash_completion.d/python-argcomplete.sh)
+uses a python program to determine startup script generated by pip.
+You can speed up completion somewhat by changing this script to include
+  # PYTHON_ARGCOMPLETE_OK
+so the python-argcomplete-check-easy-install-script does not
+need to be called to find the entry point of the code and see if that is
+marked  with PYTHON_ARGCOMPLETE_OK.
+
+INSTALL/DEBUGGING
+=================
+
+To include this support in another application that has setup.py generated
+scripts:
+
+- Add the line:
+    # PYTHON_ARGCOMPLETE_OK
+  near the top of the main python entry point.
+
+- Include in the file calling parse_args():
+    from _argcomplete import try_argcomplete, filescompleter
+  Call try_argcomplete just before parse_args(), and optionally add
+  filescompleter to the positional arguments' add_argument().
+
+If things do not work right away:
+
+- Switch on argcomplete debugging with (also helpful when doing custom
+  completers):
+    export _ARC_DEBUG=1
+
+- Run:
+    python-argcomplete-check-easy-install-script $(which appname)
+    echo $?
+  will echo 0 if the magic line has been found, 1 if not.
+
+- Sometimes it helps to find early on errors using:
+    _ARGCOMPLETE=1 _ARC_DEBUG=1 appname
+  which should throw a KeyError: 'COMPLINE' (which is properly set by the
+  global argcomplete script).
+"""
+
+from __future__ import annotations
+
+import argparse
+from glob import glob
+import os
+import sys
+from typing import Any
+
+
+class FastFilesCompleter:
+    """Fast file completer class."""
+
+    def __init__(self, directories: bool = True) -> None:
+        self.directories = directories
+
+    def __call__(self, prefix: str, **kwargs: Any) -> list[str]:
+        # Only called on non option completions.
+        if os.sep in prefix[1:]:
+            prefix_dir = len(os.path.dirname(prefix) + os.sep)
+        else:
+            prefix_dir = 0
+        completion = []
+        globbed = []
+        if "*" not in prefix and "?" not in prefix:
+            # We are on unix, otherwise no bash.
+            if not prefix or prefix[-1] == os.sep:
+                globbed.extend(glob(prefix + ".*"))
+            prefix += "*"
+        globbed.extend(glob(prefix))
+        for x in sorted(globbed):
+            if os.path.isdir(x):
+                x += "/"
+            # Append stripping the prefix (like bash, not like compgen).
+            completion.append(x[prefix_dir:])
+        return completion
+
+
+if os.environ.get("_ARGCOMPLETE"):
+    try:
+        import argcomplete.completers
+    except ImportError:
+        sys.exit(-1)
+    filescompleter: FastFilesCompleter | None = FastFilesCompleter()
+
+    def try_argcomplete(parser: argparse.ArgumentParser) -> None:
+        argcomplete.autocomplete(parser, always_complete_options=False)
+
+else:
+
+    def try_argcomplete(parser: argparse.ArgumentParser) -> None:
+        pass
+
+    filescompleter = None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__init__.py
new file mode 100644
index 00000000..0bfde426
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__init__.py
@@ -0,0 +1,26 @@
+"""Python inspection/code generation API."""
+
+from __future__ import annotations
+
+from .code import Code
+from .code import ExceptionInfo
+from .code import filter_traceback
+from .code import Frame
+from .code import getfslineno
+from .code import Traceback
+from .code import TracebackEntry
+from .source import getrawcode
+from .source import Source
+
+
+__all__ = [
+    "Code",
+    "ExceptionInfo",
+    "filter_traceback",
+    "Frame",
+    "getfslineno",
+    "getrawcode",
+    "Traceback",
+    "TracebackEntry",
+    "Source",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..77cf54c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/code.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/code.cpython-312.pyc
new file mode 100644
index 00000000..907f9fc1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/code.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/source.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/source.cpython-312.pyc
new file mode 100644
index 00000000..5caef0ce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_code/__pycache__/source.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/code.py b/backend/venv/lib/python3.12/site-packages/_pytest/_code/code.py
new file mode 100644
index 00000000..fec627b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_code/code.py
@@ -0,0 +1,1409 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import ast
+import dataclasses
+import inspect
+from inspect import CO_VARARGS
+from inspect import CO_VARKEYWORDS
+from io import StringIO
+import os
+from pathlib import Path
+import re
+import sys
+import traceback
+from traceback import format_exception_only
+from types import CodeType
+from types import FrameType
+from types import TracebackType
+from typing import Any
+from typing import Callable
+from typing import ClassVar
+from typing import Final
+from typing import final
+from typing import Generic
+from typing import Iterable
+from typing import List
+from typing import Literal
+from typing import Mapping
+from typing import overload
+from typing import Pattern
+from typing import Sequence
+from typing import SupportsIndex
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+import pluggy
+
+import _pytest
+from _pytest._code.source import findsource
+from _pytest._code.source import getrawcode
+from _pytest._code.source import getstatementrange_ast
+from _pytest._code.source import Source
+from _pytest._io import TerminalWriter
+from _pytest._io.saferepr import safeformat
+from _pytest._io.saferepr import saferepr
+from _pytest.compat import get_real_func
+from _pytest.deprecated import check_ispytest
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import bestrelpath
+
+
+if sys.version_info < (3, 11):
+    from exceptiongroup import BaseExceptionGroup
+
+TracebackStyle = Literal["long", "short", "line", "no", "native", "value", "auto"]
+
+EXCEPTION_OR_MORE = Union[Type[BaseException], Tuple[Type[BaseException], ...]]
+
+
+class Code:
+    """Wrapper around Python code objects."""
+
+    __slots__ = ("raw",)
+
+    def __init__(self, obj: CodeType) -> None:
+        self.raw = obj
+
+    @classmethod
+    def from_function(cls, obj: object) -> Code:
+        return cls(getrawcode(obj))
+
+    def __eq__(self, other):
+        return self.raw == other.raw
+
+    # Ignore type because of https://github.com/python/mypy/issues/4266.
+    __hash__ = None  # type: ignore
+
+    @property
+    def firstlineno(self) -> int:
+        return self.raw.co_firstlineno - 1
+
+    @property
+    def name(self) -> str:
+        return self.raw.co_name
+
+    @property
+    def path(self) -> Path | str:
+        """Return a path object pointing to source code, or an ``str`` in
+        case of ``OSError`` / non-existing file."""
+        if not self.raw.co_filename:
+            return ""
+        try:
+            p = absolutepath(self.raw.co_filename)
+            # maybe don't try this checking
+            if not p.exists():
+                raise OSError("path check failed.")
+            return p
+        except OSError:
+            # XXX maybe try harder like the weird logic
+            # in the standard lib [linecache.updatecache] does?
+            return self.raw.co_filename
+
+    @property
+    def fullsource(self) -> Source | None:
+        """Return a _pytest._code.Source object for the full source file of the code."""
+        full, _ = findsource(self.raw)
+        return full
+
+    def source(self) -> Source:
+        """Return a _pytest._code.Source object for the code object's source only."""
+        # return source only for that part of code
+        return Source(self.raw)
+
+    def getargs(self, var: bool = False) -> tuple[str, ...]:
+        """Return a tuple with the argument names for the code object.
+
+        If 'var' is set True also return the names of the variable and
+        keyword arguments when present.
+        """
+        # Handy shortcut for getting args.
+        raw = self.raw
+        argcount = raw.co_argcount
+        if var:
+            argcount += raw.co_flags & CO_VARARGS
+            argcount += raw.co_flags & CO_VARKEYWORDS
+        return raw.co_varnames[:argcount]
+
+
+class Frame:
+    """Wrapper around a Python frame holding f_locals and f_globals
+    in which expressions can be evaluated."""
+
+    __slots__ = ("raw",)
+
+    def __init__(self, frame: FrameType) -> None:
+        self.raw = frame
+
+    @property
+    def lineno(self) -> int:
+        return self.raw.f_lineno - 1
+
+    @property
+    def f_globals(self) -> dict[str, Any]:
+        return self.raw.f_globals
+
+    @property
+    def f_locals(self) -> dict[str, Any]:
+        return self.raw.f_locals
+
+    @property
+    def code(self) -> Code:
+        return Code(self.raw.f_code)
+
+    @property
+    def statement(self) -> Source:
+        """Statement this frame is at."""
+        if self.code.fullsource is None:
+            return Source("")
+        return self.code.fullsource.getstatement(self.lineno)
+
+    def eval(self, code, **vars):
+        """Evaluate 'code' in the frame.
+
+        'vars' are optional additional local variables.
+
+        Returns the result of the evaluation.
+        """
+        f_locals = self.f_locals.copy()
+        f_locals.update(vars)
+        return eval(code, self.f_globals, f_locals)
+
+    def repr(self, object: object) -> str:
+        """Return a 'safe' (non-recursive, one-line) string repr for 'object'."""
+        return saferepr(object)
+
+    def getargs(self, var: bool = False):
+        """Return a list of tuples (name, value) for all arguments.
+
+        If 'var' is set True, also include the variable and keyword arguments
+        when present.
+        """
+        retval = []
+        for arg in self.code.getargs(var):
+            try:
+                retval.append((arg, self.f_locals[arg]))
+            except KeyError:
+                pass  # this can occur when using Psyco
+        return retval
+
+
+class TracebackEntry:
+    """A single entry in a Traceback."""
+
+    __slots__ = ("_rawentry", "_repr_style")
+
+    def __init__(
+        self,
+        rawentry: TracebackType,
+        repr_style: Literal["short", "long"] | None = None,
+    ) -> None:
+        self._rawentry: Final = rawentry
+        self._repr_style: Final = repr_style
+
+    def with_repr_style(
+        self, repr_style: Literal["short", "long"] | None
+    ) -> TracebackEntry:
+        return TracebackEntry(self._rawentry, repr_style)
+
+    @property
+    def lineno(self) -> int:
+        return self._rawentry.tb_lineno - 1
+
+    @property
+    def frame(self) -> Frame:
+        return Frame(self._rawentry.tb_frame)
+
+    @property
+    def relline(self) -> int:
+        return self.lineno - self.frame.code.firstlineno
+
+    def __repr__(self) -> str:
+        return "<TracebackEntry %s:%d>" % (self.frame.code.path, self.lineno + 1)
+
+    @property
+    def statement(self) -> Source:
+        """_pytest._code.Source object for the current statement."""
+        source = self.frame.code.fullsource
+        assert source is not None
+        return source.getstatement(self.lineno)
+
+    @property
+    def path(self) -> Path | str:
+        """Path to the source code."""
+        return self.frame.code.path
+
+    @property
+    def locals(self) -> dict[str, Any]:
+        """Locals of underlying frame."""
+        return self.frame.f_locals
+
+    def getfirstlinesource(self) -> int:
+        return self.frame.code.firstlineno
+
+    def getsource(
+        self, astcache: dict[str | Path, ast.AST] | None = None
+    ) -> Source | None:
+        """Return failing source code."""
+        # we use the passed in astcache to not reparse asttrees
+        # within exception info printing
+        source = self.frame.code.fullsource
+        if source is None:
+            return None
+        key = astnode = None
+        if astcache is not None:
+            key = self.frame.code.path
+            if key is not None:
+                astnode = astcache.get(key, None)
+        start = self.getfirstlinesource()
+        try:
+            astnode, _, end = getstatementrange_ast(
+                self.lineno, source, astnode=astnode
+            )
+        except SyntaxError:
+            end = self.lineno + 1
+        else:
+            if key is not None and astcache is not None:
+                astcache[key] = astnode
+        return source[start:end]
+
+    source = property(getsource)
+
+    def ishidden(self, excinfo: ExceptionInfo[BaseException] | None) -> bool:
+        """Return True if the current frame has a var __tracebackhide__
+        resolving to True.
+
+        If __tracebackhide__ is a callable, it gets called with the
+        ExceptionInfo instance and can decide whether to hide the traceback.
+
+        Mostly for internal use.
+        """
+        tbh: bool | Callable[[ExceptionInfo[BaseException] | None], bool] = False
+        for maybe_ns_dct in (self.frame.f_locals, self.frame.f_globals):
+            # in normal cases, f_locals and f_globals are dictionaries
+            # however via `exec(...)` / `eval(...)` they can be other types
+            # (even incorrect types!).
+            # as such, we suppress all exceptions while accessing __tracebackhide__
+            try:
+                tbh = maybe_ns_dct["__tracebackhide__"]
+            except Exception:
+                pass
+            else:
+                break
+        if tbh and callable(tbh):
+            return tbh(excinfo)
+        return tbh
+
+    def __str__(self) -> str:
+        name = self.frame.code.name
+        try:
+            line = str(self.statement).lstrip()
+        except KeyboardInterrupt:
+            raise
+        except BaseException:
+            line = "???"
+        # This output does not quite match Python's repr for traceback entries,
+        # but changing it to do so would break certain plugins.  See
+        # https://github.com/pytest-dev/pytest/pull/7535/ for details.
+        return "  File %r:%d in %s\n  %s\n" % (
+            str(self.path),
+            self.lineno + 1,
+            name,
+            line,
+        )
+
+    @property
+    def name(self) -> str:
+        """co_name of underlying code."""
+        return self.frame.code.raw.co_name
+
+
+class Traceback(List[TracebackEntry]):
+    """Traceback objects encapsulate and offer higher level access to Traceback entries."""
+
+    def __init__(
+        self,
+        tb: TracebackType | Iterable[TracebackEntry],
+    ) -> None:
+        """Initialize from given python traceback object and ExceptionInfo."""
+        if isinstance(tb, TracebackType):
+
+            def f(cur: TracebackType) -> Iterable[TracebackEntry]:
+                cur_: TracebackType | None = cur
+                while cur_ is not None:
+                    yield TracebackEntry(cur_)
+                    cur_ = cur_.tb_next
+
+            super().__init__(f(tb))
+        else:
+            super().__init__(tb)
+
+    def cut(
+        self,
+        path: os.PathLike[str] | str | None = None,
+        lineno: int | None = None,
+        firstlineno: int | None = None,
+        excludepath: os.PathLike[str] | None = None,
+    ) -> Traceback:
+        """Return a Traceback instance wrapping part of this Traceback.
+
+        By providing any combination of path, lineno and firstlineno, the
+        first frame to start the to-be-returned traceback is determined.
+
+        This allows cutting the first part of a Traceback instance e.g.
+        for formatting reasons (removing some uninteresting bits that deal
+        with handling of the exception/traceback).
+        """
+        path_ = None if path is None else os.fspath(path)
+        excludepath_ = None if excludepath is None else os.fspath(excludepath)
+        for x in self:
+            code = x.frame.code
+            codepath = code.path
+            if path is not None and str(codepath) != path_:
+                continue
+            if (
+                excludepath is not None
+                and isinstance(codepath, Path)
+                and excludepath_ in (str(p) for p in codepath.parents)  # type: ignore[operator]
+            ):
+                continue
+            if lineno is not None and x.lineno != lineno:
+                continue
+            if firstlineno is not None and x.frame.code.firstlineno != firstlineno:
+                continue
+            return Traceback(x._rawentry)
+        return self
+
+    @overload
+    def __getitem__(self, key: SupportsIndex) -> TracebackEntry: ...
+
+    @overload
+    def __getitem__(self, key: slice) -> Traceback: ...
+
+    def __getitem__(self, key: SupportsIndex | slice) -> TracebackEntry | Traceback:
+        if isinstance(key, slice):
+            return self.__class__(super().__getitem__(key))
+        else:
+            return super().__getitem__(key)
+
+    def filter(
+        self,
+        excinfo_or_fn: ExceptionInfo[BaseException] | Callable[[TracebackEntry], bool],
+        /,
+    ) -> Traceback:
+        """Return a Traceback instance with certain items removed.
+
+        If the filter is an `ExceptionInfo`, removes all the ``TracebackEntry``s
+        which are hidden (see ishidden() above).
+
+        Otherwise, the filter is a function that gets a single argument, a
+        ``TracebackEntry`` instance, and should return True when the item should
+        be added to the ``Traceback``, False when not.
+        """
+        if isinstance(excinfo_or_fn, ExceptionInfo):
+            fn = lambda x: not x.ishidden(excinfo_or_fn)  # noqa: E731
+        else:
+            fn = excinfo_or_fn
+        return Traceback(filter(fn, self))
+
+    def recursionindex(self) -> int | None:
+        """Return the index of the frame/TracebackEntry where recursion originates if
+        appropriate, None if no recursion occurred."""
+        cache: dict[tuple[Any, int, int], list[dict[str, Any]]] = {}
+        for i, entry in enumerate(self):
+            # id for the code.raw is needed to work around
+            # the strange metaprogramming in the decorator lib from pypi
+            # which generates code objects that have hash/value equality
+            # XXX needs a test
+            key = entry.frame.code.path, id(entry.frame.code.raw), entry.lineno
+            values = cache.setdefault(key, [])
+            # Since Python 3.13 f_locals is a proxy, freeze it.
+            loc = dict(entry.frame.f_locals)
+            if values:
+                for otherloc in values:
+                    if otherloc == loc:
+                        return i
+            values.append(loc)
+        return None
+
+
+E = TypeVar("E", bound=BaseException, covariant=True)
+
+
+@final
+@dataclasses.dataclass
+class ExceptionInfo(Generic[E]):
+    """Wraps sys.exc_info() objects and offers help for navigating the traceback."""
+
+    _assert_start_repr: ClassVar = "AssertionError('assert "
+
+    _excinfo: tuple[type[E], E, TracebackType] | None
+    _striptext: str
+    _traceback: Traceback | None
+
+    def __init__(
+        self,
+        excinfo: tuple[type[E], E, TracebackType] | None,
+        striptext: str = "",
+        traceback: Traceback | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._excinfo = excinfo
+        self._striptext = striptext
+        self._traceback = traceback
+
+    @classmethod
+    def from_exception(
+        cls,
+        # Ignoring error: "Cannot use a covariant type variable as a parameter".
+        # This is OK to ignore because this class is (conceptually) readonly.
+        # See https://github.com/python/mypy/issues/7049.
+        exception: E,  # type: ignore[misc]
+        exprinfo: str | None = None,
+    ) -> ExceptionInfo[E]:
+        """Return an ExceptionInfo for an existing exception.
+
+        The exception must have a non-``None`` ``__traceback__`` attribute,
+        otherwise this function fails with an assertion error. This means that
+        the exception must have been raised, or added a traceback with the
+        :py:meth:`~BaseException.with_traceback()` method.
+
+        :param exprinfo:
+            A text string helping to determine if we should strip
+            ``AssertionError`` from the output. Defaults to the exception
+            message/``__str__()``.
+
+        .. versionadded:: 7.4
+        """
+        assert exception.__traceback__, (
+            "Exceptions passed to ExcInfo.from_exception(...)"
+            " must have a non-None __traceback__."
+        )
+        exc_info = (type(exception), exception, exception.__traceback__)
+        return cls.from_exc_info(exc_info, exprinfo)
+
+    @classmethod
+    def from_exc_info(
+        cls,
+        exc_info: tuple[type[E], E, TracebackType],
+        exprinfo: str | None = None,
+    ) -> ExceptionInfo[E]:
+        """Like :func:`from_exception`, but using old-style exc_info tuple."""
+        _striptext = ""
+        if exprinfo is None and isinstance(exc_info[1], AssertionError):
+            exprinfo = getattr(exc_info[1], "msg", None)
+            if exprinfo is None:
+                exprinfo = saferepr(exc_info[1])
+            if exprinfo and exprinfo.startswith(cls._assert_start_repr):
+                _striptext = "AssertionError: "
+
+        return cls(exc_info, _striptext, _ispytest=True)
+
+    @classmethod
+    def from_current(cls, exprinfo: str | None = None) -> ExceptionInfo[BaseException]:
+        """Return an ExceptionInfo matching the current traceback.
+
+        .. warning::
+
+            Experimental API
+
+        :param exprinfo:
+            A text string helping to determine if we should strip
+            ``AssertionError`` from the output. Defaults to the exception
+            message/``__str__()``.
+        """
+        tup = sys.exc_info()
+        assert tup[0] is not None, "no current exception"
+        assert tup[1] is not None, "no current exception"
+        assert tup[2] is not None, "no current exception"
+        exc_info = (tup[0], tup[1], tup[2])
+        return ExceptionInfo.from_exc_info(exc_info, exprinfo)
+
+    @classmethod
+    def for_later(cls) -> ExceptionInfo[E]:
+        """Return an unfilled ExceptionInfo."""
+        return cls(None, _ispytest=True)
+
+    def fill_unfilled(self, exc_info: tuple[type[E], E, TracebackType]) -> None:
+        """Fill an unfilled ExceptionInfo created with ``for_later()``."""
+        assert self._excinfo is None, "ExceptionInfo was already filled"
+        self._excinfo = exc_info
+
+    @property
+    def type(self) -> type[E]:
+        """The exception class."""
+        assert (
+            self._excinfo is not None
+        ), ".type can only be used after the context manager exits"
+        return self._excinfo[0]
+
+    @property
+    def value(self) -> E:
+        """The exception value."""
+        assert (
+            self._excinfo is not None
+        ), ".value can only be used after the context manager exits"
+        return self._excinfo[1]
+
+    @property
+    def tb(self) -> TracebackType:
+        """The exception raw traceback."""
+        assert (
+            self._excinfo is not None
+        ), ".tb can only be used after the context manager exits"
+        return self._excinfo[2]
+
+    @property
+    def typename(self) -> str:
+        """The type name of the exception."""
+        assert (
+            self._excinfo is not None
+        ), ".typename can only be used after the context manager exits"
+        return self.type.__name__
+
+    @property
+    def traceback(self) -> Traceback:
+        """The traceback."""
+        if self._traceback is None:
+            self._traceback = Traceback(self.tb)
+        return self._traceback
+
+    @traceback.setter
+    def traceback(self, value: Traceback) -> None:
+        self._traceback = value
+
+    def __repr__(self) -> str:
+        if self._excinfo is None:
+            return "<ExceptionInfo for raises contextmanager>"
+        return f"<{self.__class__.__name__} {saferepr(self._excinfo[1])} tblen={len(self.traceback)}>"
+
+    def exconly(self, tryshort: bool = False) -> str:
+        """Return the exception as a string.
+
+        When 'tryshort' resolves to True, and the exception is an
+        AssertionError, only the actual exception part of the exception
+        representation is returned (so 'AssertionError: ' is removed from
+        the beginning).
+        """
+        lines = format_exception_only(self.type, self.value)
+        text = "".join(lines)
+        text = text.rstrip()
+        if tryshort:
+            if text.startswith(self._striptext):
+                text = text[len(self._striptext) :]
+        return text
+
+    def errisinstance(self, exc: EXCEPTION_OR_MORE) -> bool:
+        """Return True if the exception is an instance of exc.
+
+        Consider using ``isinstance(excinfo.value, exc)`` instead.
+        """
+        return isinstance(self.value, exc)
+
+    def _getreprcrash(self) -> ReprFileLocation | None:
+        # Find last non-hidden traceback entry that led to the exception of the
+        # traceback, or None if all hidden.
+        for i in range(-1, -len(self.traceback) - 1, -1):
+            entry = self.traceback[i]
+            if not entry.ishidden(self):
+                path, lineno = entry.frame.code.raw.co_filename, entry.lineno
+                exconly = self.exconly(tryshort=True)
+                return ReprFileLocation(path, lineno + 1, exconly)
+        return None
+
+    def getrepr(
+        self,
+        showlocals: bool = False,
+        style: TracebackStyle = "long",
+        abspath: bool = False,
+        tbfilter: bool
+        | Callable[[ExceptionInfo[BaseException]], _pytest._code.code.Traceback] = True,
+        funcargs: bool = False,
+        truncate_locals: bool = True,
+        truncate_args: bool = True,
+        chain: bool = True,
+    ) -> ReprExceptionInfo | ExceptionChainRepr:
+        """Return str()able representation of this exception info.
+
+        :param bool showlocals:
+            Show locals per traceback entry.
+            Ignored if ``style=="native"``.
+
+        :param str style:
+            long|short|line|no|native|value traceback style.
+
+        :param bool abspath:
+            If paths should be changed to absolute or left unchanged.
+
+        :param tbfilter:
+            A filter for traceback entries.
+
+            * If false, don't hide any entries.
+            * If true, hide internal entries and entries that contain a local
+              variable ``__tracebackhide__ = True``.
+            * If a callable, delegates the filtering to the callable.
+
+            Ignored if ``style`` is ``"native"``.
+
+        :param bool funcargs:
+            Show fixtures ("funcargs" for legacy purposes) per traceback entry.
+
+        :param bool truncate_locals:
+            With ``showlocals==True``, make sure locals can be safely represented as strings.
+
+        :param bool truncate_args:
+            With ``showargs==True``, make sure args can be safely represented as strings.
+
+        :param bool chain:
+            If chained exceptions in Python 3 should be shown.
+
+        .. versionchanged:: 3.9
+
+            Added the ``chain`` parameter.
+        """
+        if style == "native":
+            return ReprExceptionInfo(
+                reprtraceback=ReprTracebackNative(
+                    traceback.format_exception(
+                        self.type,
+                        self.value,
+                        self.traceback[0]._rawentry if self.traceback else None,
+                    )
+                ),
+                reprcrash=self._getreprcrash(),
+            )
+
+        fmt = FormattedExcinfo(
+            showlocals=showlocals,
+            style=style,
+            abspath=abspath,
+            tbfilter=tbfilter,
+            funcargs=funcargs,
+            truncate_locals=truncate_locals,
+            truncate_args=truncate_args,
+            chain=chain,
+        )
+        return fmt.repr_excinfo(self)
+
+    def _stringify_exception(self, exc: BaseException) -> str:
+        try:
+            notes = getattr(exc, "__notes__", [])
+        except KeyError:
+            # Workaround for https://github.com/python/cpython/issues/98778 on
+            # Python <= 3.9, and some 3.10 and 3.11 patch versions.
+            HTTPError = getattr(sys.modules.get("urllib.error", None), "HTTPError", ())
+            if sys.version_info < (3, 12) and isinstance(exc, HTTPError):
+                notes = []
+            else:
+                raise
+
+        return "\n".join(
+            [
+                str(exc),
+                *notes,
+            ]
+        )
+
+    def match(self, regexp: str | Pattern[str]) -> Literal[True]:
+        """Check whether the regular expression `regexp` matches the string
+        representation of the exception using :func:`python:re.search`.
+
+        If it matches `True` is returned, otherwise an `AssertionError` is raised.
+        """
+        __tracebackhide__ = True
+        value = self._stringify_exception(self.value)
+        msg = f"Regex pattern did not match.\n Regex: {regexp!r}\n Input: {value!r}"
+        if regexp == value:
+            msg += "\n Did you mean to `re.escape()` the regex?"
+        assert re.search(regexp, value), msg
+        # Return True to allow for "assert excinfo.match()".
+        return True
+
+    def _group_contains(
+        self,
+        exc_group: BaseExceptionGroup[BaseException],
+        expected_exception: EXCEPTION_OR_MORE,
+        match: str | Pattern[str] | None,
+        target_depth: int | None = None,
+        current_depth: int = 1,
+    ) -> bool:
+        """Return `True` if a `BaseExceptionGroup` contains a matching exception."""
+        if (target_depth is not None) and (current_depth > target_depth):
+            # already descended past the target depth
+            return False
+        for exc in exc_group.exceptions:
+            if isinstance(exc, BaseExceptionGroup):
+                if self._group_contains(
+                    exc, expected_exception, match, target_depth, current_depth + 1
+                ):
+                    return True
+            if (target_depth is not None) and (current_depth != target_depth):
+                # not at the target depth, no match
+                continue
+            if not isinstance(exc, expected_exception):
+                continue
+            if match is not None:
+                value = self._stringify_exception(exc)
+                if not re.search(match, value):
+                    continue
+            return True
+        return False
+
+    def group_contains(
+        self,
+        expected_exception: EXCEPTION_OR_MORE,
+        *,
+        match: str | Pattern[str] | None = None,
+        depth: int | None = None,
+    ) -> bool:
+        """Check whether a captured exception group contains a matching exception.
+
+        :param Type[BaseException] | Tuple[Type[BaseException]] expected_exception:
+            The expected exception type, or a tuple if one of multiple possible
+            exception types are expected.
+
+        :param str | Pattern[str] | None match:
+            If specified, a string containing a regular expression,
+            or a regular expression object, that is tested against the string
+            representation of the exception and its `PEP-678 <https://peps.python.org/pep-0678/>` `__notes__`
+            using :func:`re.search`.
+
+            To match a literal string that may contain :ref:`special characters
+            <re-syntax>`, the pattern can first be escaped with :func:`re.escape`.
+
+        :param Optional[int] depth:
+            If `None`, will search for a matching exception at any nesting depth.
+            If >= 1, will only match an exception if it's at the specified depth (depth = 1 being
+            the exceptions contained within the topmost exception group).
+
+        .. versionadded:: 8.0
+        """
+        msg = "Captured exception is not an instance of `BaseExceptionGroup`"
+        assert isinstance(self.value, BaseExceptionGroup), msg
+        msg = "`depth` must be >= 1 if specified"
+        assert (depth is None) or (depth >= 1), msg
+        return self._group_contains(self.value, expected_exception, match, depth)
+
+
+@dataclasses.dataclass
+class FormattedExcinfo:
+    """Presenting information about failing Functions and Generators."""
+
+    # for traceback entries
+    flow_marker: ClassVar = ">"
+    fail_marker: ClassVar = "E"
+
+    showlocals: bool = False
+    style: TracebackStyle = "long"
+    abspath: bool = True
+    tbfilter: bool | Callable[[ExceptionInfo[BaseException]], Traceback] = True
+    funcargs: bool = False
+    truncate_locals: bool = True
+    truncate_args: bool = True
+    chain: bool = True
+    astcache: dict[str | Path, ast.AST] = dataclasses.field(
+        default_factory=dict, init=False, repr=False
+    )
+
+    def _getindent(self, source: Source) -> int:
+        # Figure out indent for the given source.
+        try:
+            s = str(source.getstatement(len(source) - 1))
+        except KeyboardInterrupt:
+            raise
+        except BaseException:
+            try:
+                s = str(source[-1])
+            except KeyboardInterrupt:
+                raise
+            except BaseException:
+                return 0
+        return 4 + (len(s) - len(s.lstrip()))
+
+    def _getentrysource(self, entry: TracebackEntry) -> Source | None:
+        source = entry.getsource(self.astcache)
+        if source is not None:
+            source = source.deindent()
+        return source
+
+    def repr_args(self, entry: TracebackEntry) -> ReprFuncArgs | None:
+        if self.funcargs:
+            args = []
+            for argname, argvalue in entry.frame.getargs(var=True):
+                if self.truncate_args:
+                    str_repr = saferepr(argvalue)
+                else:
+                    str_repr = saferepr(argvalue, maxsize=None)
+                args.append((argname, str_repr))
+            return ReprFuncArgs(args)
+        return None
+
+    def get_source(
+        self,
+        source: Source | None,
+        line_index: int = -1,
+        excinfo: ExceptionInfo[BaseException] | None = None,
+        short: bool = False,
+    ) -> list[str]:
+        """Return formatted and marked up source lines."""
+        lines = []
+        if source is not None and line_index < 0:
+            line_index += len(source)
+        if source is None or line_index >= len(source.lines) or line_index < 0:
+            # `line_index` could still be outside `range(len(source.lines))` if
+            # we're processing AST with pathological position attributes.
+            source = Source("???")
+            line_index = 0
+        space_prefix = "    "
+        if short:
+            lines.append(space_prefix + source.lines[line_index].strip())
+        else:
+            for line in source.lines[:line_index]:
+                lines.append(space_prefix + line)
+            lines.append(self.flow_marker + "   " + source.lines[line_index])
+            for line in source.lines[line_index + 1 :]:
+                lines.append(space_prefix + line)
+        if excinfo is not None:
+            indent = 4 if short else self._getindent(source)
+            lines.extend(self.get_exconly(excinfo, indent=indent, markall=True))
+        return lines
+
+    def get_exconly(
+        self,
+        excinfo: ExceptionInfo[BaseException],
+        indent: int = 4,
+        markall: bool = False,
+    ) -> list[str]:
+        lines = []
+        indentstr = " " * indent
+        # Get the real exception information out.
+        exlines = excinfo.exconly(tryshort=True).split("\n")
+        failindent = self.fail_marker + indentstr[1:]
+        for line in exlines:
+            lines.append(failindent + line)
+            if not markall:
+                failindent = indentstr
+        return lines
+
+    def repr_locals(self, locals: Mapping[str, object]) -> ReprLocals | None:
+        if self.showlocals:
+            lines = []
+            keys = [loc for loc in locals if loc[0] != "@"]
+            keys.sort()
+            for name in keys:
+                value = locals[name]
+                if name == "__builtins__":
+                    lines.append("__builtins__ = <builtins>")
+                else:
+                    # This formatting could all be handled by the
+                    # _repr() function, which is only reprlib.Repr in
+                    # disguise, so is very configurable.
+                    if self.truncate_locals:
+                        str_repr = saferepr(value)
+                    else:
+                        str_repr = safeformat(value)
+                    # if len(str_repr) < 70 or not isinstance(value, (list, tuple, dict)):
+                    lines.append(f"{name:<10} = {str_repr}")
+                    # else:
+                    #    self._line("%-10s =\\" % (name,))
+                    #    # XXX
+                    #    pprint.pprint(value, stream=self.excinfowriter)
+            return ReprLocals(lines)
+        return None
+
+    def repr_traceback_entry(
+        self,
+        entry: TracebackEntry | None,
+        excinfo: ExceptionInfo[BaseException] | None = None,
+    ) -> ReprEntry:
+        lines: list[str] = []
+        style = (
+            entry._repr_style
+            if entry is not None and entry._repr_style is not None
+            else self.style
+        )
+        if style in ("short", "long") and entry is not None:
+            source = self._getentrysource(entry)
+            if source is None:
+                source = Source("???")
+                line_index = 0
+            else:
+                line_index = entry.lineno - entry.getfirstlinesource()
+            short = style == "short"
+            reprargs = self.repr_args(entry) if not short else None
+            s = self.get_source(source, line_index, excinfo, short=short)
+            lines.extend(s)
+            if short:
+                message = f"in {entry.name}"
+            else:
+                message = excinfo and excinfo.typename or ""
+            entry_path = entry.path
+            path = self._makepath(entry_path)
+            reprfileloc = ReprFileLocation(path, entry.lineno + 1, message)
+            localsrepr = self.repr_locals(entry.locals)
+            return ReprEntry(lines, reprargs, localsrepr, reprfileloc, style)
+        elif style == "value":
+            if excinfo:
+                lines.extend(str(excinfo.value).split("\n"))
+            return ReprEntry(lines, None, None, None, style)
+        else:
+            if excinfo:
+                lines.extend(self.get_exconly(excinfo, indent=4))
+            return ReprEntry(lines, None, None, None, style)
+
+    def _makepath(self, path: Path | str) -> str:
+        if not self.abspath and isinstance(path, Path):
+            try:
+                np = bestrelpath(Path.cwd(), path)
+            except OSError:
+                return str(path)
+            if len(np) < len(str(path)):
+                return np
+        return str(path)
+
+    def repr_traceback(self, excinfo: ExceptionInfo[BaseException]) -> ReprTraceback:
+        traceback = excinfo.traceback
+        if callable(self.tbfilter):
+            traceback = self.tbfilter(excinfo)
+        elif self.tbfilter:
+            traceback = traceback.filter(excinfo)
+
+        if isinstance(excinfo.value, RecursionError):
+            traceback, extraline = self._truncate_recursive_traceback(traceback)
+        else:
+            extraline = None
+
+        if not traceback:
+            if extraline is None:
+                extraline = "All traceback entries are hidden. Pass `--full-trace` to see hidden and internal frames."
+            entries = [self.repr_traceback_entry(None, excinfo)]
+            return ReprTraceback(entries, extraline, style=self.style)
+
+        last = traceback[-1]
+        if self.style == "value":
+            entries = [self.repr_traceback_entry(last, excinfo)]
+            return ReprTraceback(entries, None, style=self.style)
+
+        entries = [
+            self.repr_traceback_entry(entry, excinfo if last == entry else None)
+            for entry in traceback
+        ]
+        return ReprTraceback(entries, extraline, style=self.style)
+
+    def _truncate_recursive_traceback(
+        self, traceback: Traceback
+    ) -> tuple[Traceback, str | None]:
+        """Truncate the given recursive traceback trying to find the starting
+        point of the recursion.
+
+        The detection is done by going through each traceback entry and
+        finding the point in which the locals of the frame are equal to the
+        locals of a previous frame (see ``recursionindex()``).
+
+        Handle the situation where the recursion process might raise an
+        exception (for example comparing numpy arrays using equality raises a
+        TypeError), in which case we do our best to warn the user of the
+        error and show a limited traceback.
+        """
+        try:
+            recursionindex = traceback.recursionindex()
+        except Exception as e:
+            max_frames = 10
+            extraline: str | None = (
+                "!!! Recursion error detected, but an error occurred locating the origin of recursion.\n"
+                "  The following exception happened when comparing locals in the stack frame:\n"
+                f"    {type(e).__name__}: {e!s}\n"
+                f"  Displaying first and last {max_frames} stack frames out of {len(traceback)}."
+            )
+            # Type ignored because adding two instances of a List subtype
+            # currently incorrectly has type List instead of the subtype.
+            traceback = traceback[:max_frames] + traceback[-max_frames:]  # type: ignore
+        else:
+            if recursionindex is not None:
+                extraline = "!!! Recursion detected (same locals & position)"
+                traceback = traceback[: recursionindex + 1]
+            else:
+                extraline = None
+
+        return traceback, extraline
+
+    def repr_excinfo(self, excinfo: ExceptionInfo[BaseException]) -> ExceptionChainRepr:
+        repr_chain: list[tuple[ReprTraceback, ReprFileLocation | None, str | None]] = []
+        e: BaseException | None = excinfo.value
+        excinfo_: ExceptionInfo[BaseException] | None = excinfo
+        descr = None
+        seen: set[int] = set()
+        while e is not None and id(e) not in seen:
+            seen.add(id(e))
+
+            if excinfo_:
+                # Fall back to native traceback as a temporary workaround until
+                # full support for exception groups added to ExceptionInfo.
+                # See https://github.com/pytest-dev/pytest/issues/9159
+                if isinstance(e, BaseExceptionGroup):
+                    reprtraceback: ReprTracebackNative | ReprTraceback = (
+                        ReprTracebackNative(
+                            traceback.format_exception(
+                                type(excinfo_.value),
+                                excinfo_.value,
+                                excinfo_.traceback[0]._rawentry,
+                            )
+                        )
+                    )
+                else:
+                    reprtraceback = self.repr_traceback(excinfo_)
+                reprcrash = excinfo_._getreprcrash()
+            else:
+                # Fallback to native repr if the exception doesn't have a traceback:
+                # ExceptionInfo objects require a full traceback to work.
+                reprtraceback = ReprTracebackNative(
+                    traceback.format_exception(type(e), e, None)
+                )
+                reprcrash = None
+            repr_chain += [(reprtraceback, reprcrash, descr)]
+
+            if e.__cause__ is not None and self.chain:
+                e = e.__cause__
+                excinfo_ = ExceptionInfo.from_exception(e) if e.__traceback__ else None
+                descr = "The above exception was the direct cause of the following exception:"
+            elif (
+                e.__context__ is not None and not e.__suppress_context__ and self.chain
+            ):
+                e = e.__context__
+                excinfo_ = ExceptionInfo.from_exception(e) if e.__traceback__ else None
+                descr = "During handling of the above exception, another exception occurred:"
+            else:
+                e = None
+        repr_chain.reverse()
+        return ExceptionChainRepr(repr_chain)
+
+
+@dataclasses.dataclass(eq=False)
+class TerminalRepr:
+    def __str__(self) -> str:
+        # FYI this is called from pytest-xdist's serialization of exception
+        # information.
+        io = StringIO()
+        tw = TerminalWriter(file=io)
+        self.toterminal(tw)
+        return io.getvalue().strip()
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__} instance at {id(self):0x}>"
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        raise NotImplementedError()
+
+
+# This class is abstract -- only subclasses are instantiated.
+@dataclasses.dataclass(eq=False)
+class ExceptionRepr(TerminalRepr):
+    # Provided by subclasses.
+    reprtraceback: ReprTraceback
+    reprcrash: ReprFileLocation | None
+    sections: list[tuple[str, str, str]] = dataclasses.field(
+        init=False, default_factory=list
+    )
+
+    def addsection(self, name: str, content: str, sep: str = "-") -> None:
+        self.sections.append((name, content, sep))
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        for name, content, sep in self.sections:
+            tw.sep(sep, name)
+            tw.line(content)
+
+
+@dataclasses.dataclass(eq=False)
+class ExceptionChainRepr(ExceptionRepr):
+    chain: Sequence[tuple[ReprTraceback, ReprFileLocation | None, str | None]]
+
+    def __init__(
+        self,
+        chain: Sequence[tuple[ReprTraceback, ReprFileLocation | None, str | None]],
+    ) -> None:
+        # reprcrash and reprtraceback of the outermost (the newest) exception
+        # in the chain.
+        super().__init__(
+            reprtraceback=chain[-1][0],
+            reprcrash=chain[-1][1],
+        )
+        self.chain = chain
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        for element in self.chain:
+            element[0].toterminal(tw)
+            if element[2] is not None:
+                tw.line("")
+                tw.line(element[2], yellow=True)
+        super().toterminal(tw)
+
+
+@dataclasses.dataclass(eq=False)
+class ReprExceptionInfo(ExceptionRepr):
+    reprtraceback: ReprTraceback
+    reprcrash: ReprFileLocation | None
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        self.reprtraceback.toterminal(tw)
+        super().toterminal(tw)
+
+
+@dataclasses.dataclass(eq=False)
+class ReprTraceback(TerminalRepr):
+    reprentries: Sequence[ReprEntry | ReprEntryNative]
+    extraline: str | None
+    style: TracebackStyle
+
+    entrysep: ClassVar = "_ "
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        # The entries might have different styles.
+        for i, entry in enumerate(self.reprentries):
+            if entry.style == "long":
+                tw.line("")
+            entry.toterminal(tw)
+            if i < len(self.reprentries) - 1:
+                next_entry = self.reprentries[i + 1]
+                if (
+                    entry.style == "long"
+                    or entry.style == "short"
+                    and next_entry.style == "long"
+                ):
+                    tw.sep(self.entrysep)
+
+        if self.extraline:
+            tw.line(self.extraline)
+
+
+class ReprTracebackNative(ReprTraceback):
+    def __init__(self, tblines: Sequence[str]) -> None:
+        self.reprentries = [ReprEntryNative(tblines)]
+        self.extraline = None
+        self.style = "native"
+
+
+@dataclasses.dataclass(eq=False)
+class ReprEntryNative(TerminalRepr):
+    lines: Sequence[str]
+
+    style: ClassVar[TracebackStyle] = "native"
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        tw.write("".join(self.lines))
+
+
+@dataclasses.dataclass(eq=False)
+class ReprEntry(TerminalRepr):
+    lines: Sequence[str]
+    reprfuncargs: ReprFuncArgs | None
+    reprlocals: ReprLocals | None
+    reprfileloc: ReprFileLocation | None
+    style: TracebackStyle
+
+    def _write_entry_lines(self, tw: TerminalWriter) -> None:
+        """Write the source code portions of a list of traceback entries with syntax highlighting.
+
+        Usually entries are lines like these:
+
+            "     x = 1"
+            ">    assert x == 2"
+            "E    assert 1 == 2"
+
+        This function takes care of rendering the "source" portions of it (the lines without
+        the "E" prefix) using syntax highlighting, taking care to not highlighting the ">"
+        character, as doing so might break line continuations.
+        """
+        if not self.lines:
+            return
+
+        if self.style == "value":
+            # Using tw.write instead of tw.line for testing purposes due to TWMock implementation;
+            # lines written with TWMock.line and TWMock._write_source cannot be distinguished
+            # from each other, whereas lines written with TWMock.write are marked with TWMock.WRITE
+            for line in self.lines:
+                tw.write(line)
+                tw.write("\n")
+            return
+
+        # separate indents and source lines that are not failures: we want to
+        # highlight the code but not the indentation, which may contain markers
+        # such as ">   assert 0"
+        fail_marker = f"{FormattedExcinfo.fail_marker}   "
+        indent_size = len(fail_marker)
+        indents: list[str] = []
+        source_lines: list[str] = []
+        failure_lines: list[str] = []
+        for index, line in enumerate(self.lines):
+            is_failure_line = line.startswith(fail_marker)
+            if is_failure_line:
+                # from this point on all lines are considered part of the failure
+                failure_lines.extend(self.lines[index:])
+                break
+            else:
+                indents.append(line[:indent_size])
+                source_lines.append(line[indent_size:])
+
+        tw._write_source(source_lines, indents)
+
+        # failure lines are always completely red and bold
+        for line in failure_lines:
+            tw.line(line, bold=True, red=True)
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        if self.style == "short":
+            if self.reprfileloc:
+                self.reprfileloc.toterminal(tw)
+            self._write_entry_lines(tw)
+            if self.reprlocals:
+                self.reprlocals.toterminal(tw, indent=" " * 8)
+            return
+
+        if self.reprfuncargs:
+            self.reprfuncargs.toterminal(tw)
+
+        self._write_entry_lines(tw)
+
+        if self.reprlocals:
+            tw.line("")
+            self.reprlocals.toterminal(tw)
+        if self.reprfileloc:
+            if self.lines:
+                tw.line("")
+            self.reprfileloc.toterminal(tw)
+
+    def __str__(self) -> str:
+        return "{}\n{}\n{}".format(
+            "\n".join(self.lines), self.reprlocals, self.reprfileloc
+        )
+
+
+@dataclasses.dataclass(eq=False)
+class ReprFileLocation(TerminalRepr):
+    path: str
+    lineno: int
+    message: str
+
+    def __post_init__(self) -> None:
+        self.path = str(self.path)
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        # Filename and lineno output for each entry, using an output format
+        # that most editors understand.
+        msg = self.message
+        i = msg.find("\n")
+        if i != -1:
+            msg = msg[:i]
+        tw.write(self.path, bold=True, red=True)
+        tw.line(f":{self.lineno}: {msg}")
+
+
+@dataclasses.dataclass(eq=False)
+class ReprLocals(TerminalRepr):
+    lines: Sequence[str]
+
+    def toterminal(self, tw: TerminalWriter, indent="") -> None:
+        for line in self.lines:
+            tw.line(indent + line)
+
+
+@dataclasses.dataclass(eq=False)
+class ReprFuncArgs(TerminalRepr):
+    args: Sequence[tuple[str, object]]
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        if self.args:
+            linesofar = ""
+            for name, value in self.args:
+                ns = f"{name} = {value}"
+                if len(ns) + len(linesofar) + 2 > tw.fullwidth:
+                    if linesofar:
+                        tw.line(linesofar)
+                    linesofar = ns
+                else:
+                    if linesofar:
+                        linesofar += ", " + ns
+                    else:
+                        linesofar = ns
+            if linesofar:
+                tw.line(linesofar)
+            tw.line("")
+
+
+def getfslineno(obj: object) -> tuple[str | Path, int]:
+    """Return source location (path, lineno) for the given object.
+
+    If the source cannot be determined return ("", -1).
+
+    The line number is 0-based.
+    """
+    # xxx let decorators etc specify a sane ordering
+    # NOTE: this used to be done in _pytest.compat.getfslineno, initially added
+    #       in 6ec13a2b9.  It ("place_as") appears to be something very custom.
+    obj = get_real_func(obj)
+    if hasattr(obj, "place_as"):
+        obj = obj.place_as
+
+    try:
+        code = Code.from_function(obj)
+    except TypeError:
+        try:
+            fn = inspect.getsourcefile(obj) or inspect.getfile(obj)  # type: ignore[arg-type]
+        except TypeError:
+            return "", -1
+
+        fspath = fn and absolutepath(fn) or ""
+        lineno = -1
+        if fspath:
+            try:
+                _, lineno = findsource(obj)
+            except OSError:
+                pass
+        return fspath, lineno
+
+    return code.path, code.firstlineno
+
+
+# Relative paths that we use to filter traceback entries from appearing to the user;
+# see filter_traceback.
+# note: if we need to add more paths than what we have now we should probably use a list
+# for better maintenance.
+
+_PLUGGY_DIR = Path(pluggy.__file__.rstrip("oc"))
+# pluggy is either a package or a single module depending on the version
+if _PLUGGY_DIR.name == "__init__.py":
+    _PLUGGY_DIR = _PLUGGY_DIR.parent
+_PYTEST_DIR = Path(_pytest.__file__).parent
+
+
+def filter_traceback(entry: TracebackEntry) -> bool:
+    """Return True if a TracebackEntry instance should be included in tracebacks.
+
+    We hide traceback entries of:
+
+    * dynamically generated code (no code to show up for it);
+    * internal traceback from pytest or its internal libraries, py and pluggy.
+    """
+    # entry.path might sometimes return a str object when the entry
+    # points to dynamically generated code.
+    # See https://bitbucket.org/pytest-dev/py/issues/71.
+    raw_filename = entry.frame.code.raw.co_filename
+    is_generated = "<" in raw_filename and ">" in raw_filename
+    if is_generated:
+        return False
+
+    # entry.path might point to a non-existing file, in which case it will
+    # also return a str object. See #1133.
+    p = Path(entry.path)
+
+    parents = p.parents
+    if _PLUGGY_DIR in parents:
+        return False
+    if _PYTEST_DIR in parents:
+        return False
+
+    return True
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_code/source.py b/backend/venv/lib/python3.12/site-packages/_pytest/_code/source.py
new file mode 100644
index 00000000..604aff8b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_code/source.py
@@ -0,0 +1,215 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import ast
+from bisect import bisect_right
+import inspect
+import textwrap
+import tokenize
+import types
+from typing import Iterable
+from typing import Iterator
+from typing import overload
+import warnings
+
+
+class Source:
+    """An immutable object holding a source code fragment.
+
+    When using Source(...), the source lines are deindented.
+    """
+
+    def __init__(self, obj: object = None) -> None:
+        if not obj:
+            self.lines: list[str] = []
+        elif isinstance(obj, Source):
+            self.lines = obj.lines
+        elif isinstance(obj, (tuple, list)):
+            self.lines = deindent(x.rstrip("\n") for x in obj)
+        elif isinstance(obj, str):
+            self.lines = deindent(obj.split("\n"))
+        else:
+            try:
+                rawcode = getrawcode(obj)
+                src = inspect.getsource(rawcode)
+            except TypeError:
+                src = inspect.getsource(obj)  # type: ignore[arg-type]
+            self.lines = deindent(src.split("\n"))
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Source):
+            return NotImplemented
+        return self.lines == other.lines
+
+    # Ignore type because of https://github.com/python/mypy/issues/4266.
+    __hash__ = None  # type: ignore
+
+    @overload
+    def __getitem__(self, key: int) -> str: ...
+
+    @overload
+    def __getitem__(self, key: slice) -> Source: ...
+
+    def __getitem__(self, key: int | slice) -> str | Source:
+        if isinstance(key, int):
+            return self.lines[key]
+        else:
+            if key.step not in (None, 1):
+                raise IndexError("cannot slice a Source with a step")
+            newsource = Source()
+            newsource.lines = self.lines[key.start : key.stop]
+            return newsource
+
+    def __iter__(self) -> Iterator[str]:
+        return iter(self.lines)
+
+    def __len__(self) -> int:
+        return len(self.lines)
+
+    def strip(self) -> Source:
+        """Return new Source object with trailing and leading blank lines removed."""
+        start, end = 0, len(self)
+        while start < end and not self.lines[start].strip():
+            start += 1
+        while end > start and not self.lines[end - 1].strip():
+            end -= 1
+        source = Source()
+        source.lines[:] = self.lines[start:end]
+        return source
+
+    def indent(self, indent: str = " " * 4) -> Source:
+        """Return a copy of the source object with all lines indented by the
+        given indent-string."""
+        newsource = Source()
+        newsource.lines = [(indent + line) for line in self.lines]
+        return newsource
+
+    def getstatement(self, lineno: int) -> Source:
+        """Return Source statement which contains the given linenumber
+        (counted from 0)."""
+        start, end = self.getstatementrange(lineno)
+        return self[start:end]
+
+    def getstatementrange(self, lineno: int) -> tuple[int, int]:
+        """Return (start, end) tuple which spans the minimal statement region
+        which containing the given lineno."""
+        if not (0 <= lineno < len(self)):
+            raise IndexError("lineno out of range")
+        ast, start, end = getstatementrange_ast(lineno, self)
+        return start, end
+
+    def deindent(self) -> Source:
+        """Return a new Source object deindented."""
+        newsource = Source()
+        newsource.lines[:] = deindent(self.lines)
+        return newsource
+
+    def __str__(self) -> str:
+        return "\n".join(self.lines)
+
+
+#
+# helper functions
+#
+
+
+def findsource(obj) -> tuple[Source | None, int]:
+    try:
+        sourcelines, lineno = inspect.findsource(obj)
+    except Exception:
+        return None, -1
+    source = Source()
+    source.lines = [line.rstrip() for line in sourcelines]
+    return source, lineno
+
+
+def getrawcode(obj: object, trycall: bool = True) -> types.CodeType:
+    """Return code object for given function."""
+    try:
+        return obj.__code__  # type: ignore[attr-defined,no-any-return]
+    except AttributeError:
+        pass
+    if trycall:
+        call = getattr(obj, "__call__", None)
+        if call and not isinstance(obj, type):
+            return getrawcode(call, trycall=False)
+    raise TypeError(f"could not get code object for {obj!r}")
+
+
+def deindent(lines: Iterable[str]) -> list[str]:
+    return textwrap.dedent("\n".join(lines)).splitlines()
+
+
+def get_statement_startend2(lineno: int, node: ast.AST) -> tuple[int, int | None]:
+    # Flatten all statements and except handlers into one lineno-list.
+    # AST's line numbers start indexing at 1.
+    values: list[int] = []
+    for x in ast.walk(node):
+        if isinstance(x, (ast.stmt, ast.ExceptHandler)):
+            # The lineno points to the class/def, so need to include the decorators.
+            if isinstance(x, (ast.ClassDef, ast.FunctionDef, ast.AsyncFunctionDef)):
+                for d in x.decorator_list:
+                    values.append(d.lineno - 1)
+            values.append(x.lineno - 1)
+            for name in ("finalbody", "orelse"):
+                val: list[ast.stmt] | None = getattr(x, name, None)
+                if val:
+                    # Treat the finally/orelse part as its own statement.
+                    values.append(val[0].lineno - 1 - 1)
+    values.sort()
+    insert_index = bisect_right(values, lineno)
+    start = values[insert_index - 1]
+    if insert_index >= len(values):
+        end = None
+    else:
+        end = values[insert_index]
+    return start, end
+
+
+def getstatementrange_ast(
+    lineno: int,
+    source: Source,
+    assertion: bool = False,
+    astnode: ast.AST | None = None,
+) -> tuple[ast.AST, int, int]:
+    if astnode is None:
+        content = str(source)
+        # See #4260:
+        # Don't produce duplicate warnings when compiling source to find AST.
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            astnode = ast.parse(content, "source", "exec")
+
+    start, end = get_statement_startend2(lineno, astnode)
+    # We need to correct the end:
+    # - ast-parsing strips comments
+    # - there might be empty lines
+    # - we might have lesser indented code blocks at the end
+    if end is None:
+        end = len(source.lines)
+
+    if end > start + 1:
+        # Make sure we don't span differently indented code blocks
+        # by using the BlockFinder helper used which inspect.getsource() uses itself.
+        block_finder = inspect.BlockFinder()
+        # If we start with an indented line, put blockfinder to "started" mode.
+        block_finder.started = (
+            bool(source.lines[start]) and source.lines[start][0].isspace()
+        )
+        it = ((x + "\n") for x in source.lines[start:end])
+        try:
+            for tok in tokenize.generate_tokens(lambda: next(it)):
+                block_finder.tokeneater(*tok)
+        except (inspect.EndOfBlock, IndentationError):
+            end = block_finder.last + start
+        except Exception:
+            pass
+
+    # The end might still point to a comment or empty line, correct it.
+    while end:
+        line = source.lines[end - 1].lstrip()
+        if line.startswith("#") or not line:
+            end -= 1
+        else:
+            break
+    return astnode, start, end
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__init__.py
new file mode 100644
index 00000000..b0155b18
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__init__.py
@@ -0,0 +1,10 @@
+from __future__ import annotations
+
+from .terminalwriter import get_terminal_width
+from .terminalwriter import TerminalWriter
+
+
+__all__ = [
+    "TerminalWriter",
+    "get_terminal_width",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..5045ba87
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/pprint.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/pprint.cpython-312.pyc
new file mode 100644
index 00000000..8c35aff2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/pprint.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/saferepr.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/saferepr.cpython-312.pyc
new file mode 100644
index 00000000..bc9ccc8b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/saferepr.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/terminalwriter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/terminalwriter.cpython-312.pyc
new file mode 100644
index 00000000..1316cab3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/terminalwriter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/wcwidth.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/wcwidth.cpython-312.pyc
new file mode 100644
index 00000000..9e6be31a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_io/__pycache__/wcwidth.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/pprint.py b/backend/venv/lib/python3.12/site-packages/_pytest/_io/pprint.py
new file mode 100644
index 00000000..fc29989b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_io/pprint.py
@@ -0,0 +1,673 @@
+# mypy: allow-untyped-defs
+# This module was imported from the cpython standard library
+# (https://github.com/python/cpython/) at commit
+# c5140945c723ae6c4b7ee81ff720ac8ea4b52cfd (python3.12).
+#
+#
+#  Original Author:      Fred L. Drake, Jr.
+#                        fdrake@acm.org
+#
+#  This is a simple little module I wrote to make life easier.  I didn't
+#  see anything quite like it in the library, though I may have overlooked
+#  something.  I wrote this when I was trying to read some heavily nested
+#  tuples with fairly non-descriptive content.  This is modeled very much
+#  after Lisp/Scheme - style pretty-printing of lists.  If you find it
+#  useful, thank small children who sleep at night.
+from __future__ import annotations
+
+import collections as _collections
+import dataclasses as _dataclasses
+from io import StringIO as _StringIO
+import re
+import types as _types
+from typing import Any
+from typing import Callable
+from typing import IO
+from typing import Iterator
+
+
+class _safe_key:
+    """Helper function for key functions when sorting unorderable objects.
+
+    The wrapped-object will fallback to a Py2.x style comparison for
+    unorderable types (sorting first comparing the type name and then by
+    the obj ids).  Does not work recursively, so dict.items() must have
+    _safe_key applied to both the key and the value.
+
+    """
+
+    __slots__ = ["obj"]
+
+    def __init__(self, obj):
+        self.obj = obj
+
+    def __lt__(self, other):
+        try:
+            return self.obj < other.obj
+        except TypeError:
+            return (str(type(self.obj)), id(self.obj)) < (
+                str(type(other.obj)),
+                id(other.obj),
+            )
+
+
+def _safe_tuple(t):
+    """Helper function for comparing 2-tuples"""
+    return _safe_key(t[0]), _safe_key(t[1])
+
+
+class PrettyPrinter:
+    def __init__(
+        self,
+        indent: int = 4,
+        width: int = 80,
+        depth: int | None = None,
+    ) -> None:
+        """Handle pretty printing operations onto a stream using a set of
+        configured parameters.
+
+        indent
+            Number of spaces to indent for each level of nesting.
+
+        width
+            Attempted maximum number of columns in the output.
+
+        depth
+            The maximum depth to print out nested structures.
+
+        """
+        if indent < 0:
+            raise ValueError("indent must be >= 0")
+        if depth is not None and depth <= 0:
+            raise ValueError("depth must be > 0")
+        if not width:
+            raise ValueError("width must be != 0")
+        self._depth = depth
+        self._indent_per_level = indent
+        self._width = width
+
+    def pformat(self, object: Any) -> str:
+        sio = _StringIO()
+        self._format(object, sio, 0, 0, set(), 0)
+        return sio.getvalue()
+
+    def _format(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        objid = id(object)
+        if objid in context:
+            stream.write(_recursion(object))
+            return
+
+        p = self._dispatch.get(type(object).__repr__, None)
+        if p is not None:
+            context.add(objid)
+            p(self, object, stream, indent, allowance, context, level + 1)
+            context.remove(objid)
+        elif (
+            _dataclasses.is_dataclass(object)  # type:ignore[unreachable]
+            and not isinstance(object, type)
+            and object.__dataclass_params__.repr
+            and
+            # Check dataclass has generated repr method.
+            hasattr(object.__repr__, "__wrapped__")
+            and "__create_fn__" in object.__repr__.__wrapped__.__qualname__
+        ):
+            context.add(objid)  # type:ignore[unreachable]
+            self._pprint_dataclass(
+                object, stream, indent, allowance, context, level + 1
+            )
+            context.remove(objid)
+        else:
+            stream.write(self._repr(object, context, level))
+
+    def _pprint_dataclass(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        cls_name = object.__class__.__name__
+        items = [
+            (f.name, getattr(object, f.name))
+            for f in _dataclasses.fields(object)
+            if f.repr
+        ]
+        stream.write(cls_name + "(")
+        self._format_namespace_items(items, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch: dict[
+        Callable[..., str],
+        Callable[[PrettyPrinter, Any, IO[str], int, int, set[int], int], None],
+    ] = {}
+
+    def _pprint_dict(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        write = stream.write
+        write("{")
+        items = sorted(object.items(), key=_safe_tuple)
+        self._format_dict_items(items, stream, indent, allowance, context, level)
+        write("}")
+
+    _dispatch[dict.__repr__] = _pprint_dict
+
+    def _pprint_ordered_dict(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not len(object):
+            stream.write(repr(object))
+            return
+        cls = object.__class__
+        stream.write(cls.__name__ + "(")
+        self._pprint_dict(object, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[_collections.OrderedDict.__repr__] = _pprint_ordered_dict
+
+    def _pprint_list(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        stream.write("[")
+        self._format_items(object, stream, indent, allowance, context, level)
+        stream.write("]")
+
+    _dispatch[list.__repr__] = _pprint_list
+
+    def _pprint_tuple(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        stream.write("(")
+        self._format_items(object, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[tuple.__repr__] = _pprint_tuple
+
+    def _pprint_set(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not len(object):
+            stream.write(repr(object))
+            return
+        typ = object.__class__
+        if typ is set:
+            stream.write("{")
+            endchar = "}"
+        else:
+            stream.write(typ.__name__ + "({")
+            endchar = "})"
+        object = sorted(object, key=_safe_key)
+        self._format_items(object, stream, indent, allowance, context, level)
+        stream.write(endchar)
+
+    _dispatch[set.__repr__] = _pprint_set
+    _dispatch[frozenset.__repr__] = _pprint_set
+
+    def _pprint_str(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        write = stream.write
+        if not len(object):
+            write(repr(object))
+            return
+        chunks = []
+        lines = object.splitlines(True)
+        if level == 1:
+            indent += 1
+            allowance += 1
+        max_width1 = max_width = self._width - indent
+        for i, line in enumerate(lines):
+            rep = repr(line)
+            if i == len(lines) - 1:
+                max_width1 -= allowance
+            if len(rep) <= max_width1:
+                chunks.append(rep)
+            else:
+                # A list of alternating (non-space, space) strings
+                parts = re.findall(r"\S*\s*", line)
+                assert parts
+                assert not parts[-1]
+                parts.pop()  # drop empty last part
+                max_width2 = max_width
+                current = ""
+                for j, part in enumerate(parts):
+                    candidate = current + part
+                    if j == len(parts) - 1 and i == len(lines) - 1:
+                        max_width2 -= allowance
+                    if len(repr(candidate)) > max_width2:
+                        if current:
+                            chunks.append(repr(current))
+                        current = part
+                    else:
+                        current = candidate
+                if current:
+                    chunks.append(repr(current))
+        if len(chunks) == 1:
+            write(rep)
+            return
+        if level == 1:
+            write("(")
+        for i, rep in enumerate(chunks):
+            if i > 0:
+                write("\n" + " " * indent)
+            write(rep)
+        if level == 1:
+            write(")")
+
+    _dispatch[str.__repr__] = _pprint_str
+
+    def _pprint_bytes(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        write = stream.write
+        if len(object) <= 4:
+            write(repr(object))
+            return
+        parens = level == 1
+        if parens:
+            indent += 1
+            allowance += 1
+            write("(")
+        delim = ""
+        for rep in _wrap_bytes_repr(object, self._width - indent, allowance):
+            write(delim)
+            write(rep)
+            if not delim:
+                delim = "\n" + " " * indent
+        if parens:
+            write(")")
+
+    _dispatch[bytes.__repr__] = _pprint_bytes
+
+    def _pprint_bytearray(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        write = stream.write
+        write("bytearray(")
+        self._pprint_bytes(
+            bytes(object), stream, indent + 10, allowance + 1, context, level + 1
+        )
+        write(")")
+
+    _dispatch[bytearray.__repr__] = _pprint_bytearray
+
+    def _pprint_mappingproxy(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        stream.write("mappingproxy(")
+        self._format(object.copy(), stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[_types.MappingProxyType.__repr__] = _pprint_mappingproxy
+
+    def _pprint_simplenamespace(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if type(object) is _types.SimpleNamespace:
+            # The SimpleNamespace repr is "namespace" instead of the class
+            # name, so we do the same here. For subclasses; use the class name.
+            cls_name = "namespace"
+        else:
+            cls_name = object.__class__.__name__
+        items = object.__dict__.items()
+        stream.write(cls_name + "(")
+        self._format_namespace_items(items, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[_types.SimpleNamespace.__repr__] = _pprint_simplenamespace
+
+    def _format_dict_items(
+        self,
+        items: list[tuple[Any, Any]],
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not items:
+            return
+
+        write = stream.write
+        item_indent = indent + self._indent_per_level
+        delimnl = "\n" + " " * item_indent
+        for key, ent in items:
+            write(delimnl)
+            write(self._repr(key, context, level))
+            write(": ")
+            self._format(ent, stream, item_indent, 1, context, level)
+            write(",")
+
+        write("\n" + " " * indent)
+
+    def _format_namespace_items(
+        self,
+        items: list[tuple[Any, Any]],
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not items:
+            return
+
+        write = stream.write
+        item_indent = indent + self._indent_per_level
+        delimnl = "\n" + " " * item_indent
+        for key, ent in items:
+            write(delimnl)
+            write(key)
+            write("=")
+            if id(ent) in context:
+                # Special-case representation of recursion to match standard
+                # recursive dataclass repr.
+                write("...")
+            else:
+                self._format(
+                    ent,
+                    stream,
+                    item_indent + len(key) + 1,
+                    1,
+                    context,
+                    level,
+                )
+
+            write(",")
+
+        write("\n" + " " * indent)
+
+    def _format_items(
+        self,
+        items: list[Any],
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not items:
+            return
+
+        write = stream.write
+        item_indent = indent + self._indent_per_level
+        delimnl = "\n" + " " * item_indent
+
+        for item in items:
+            write(delimnl)
+            self._format(item, stream, item_indent, 1, context, level)
+            write(",")
+
+        write("\n" + " " * indent)
+
+    def _repr(self, object: Any, context: set[int], level: int) -> str:
+        return self._safe_repr(object, context.copy(), self._depth, level)
+
+    def _pprint_default_dict(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        rdf = self._repr(object.default_factory, context, level)
+        stream.write(f"{object.__class__.__name__}({rdf}, ")
+        self._pprint_dict(object, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[_collections.defaultdict.__repr__] = _pprint_default_dict
+
+    def _pprint_counter(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        stream.write(object.__class__.__name__ + "(")
+
+        if object:
+            stream.write("{")
+            items = object.most_common()
+            self._format_dict_items(items, stream, indent, allowance, context, level)
+            stream.write("}")
+
+        stream.write(")")
+
+    _dispatch[_collections.Counter.__repr__] = _pprint_counter
+
+    def _pprint_chain_map(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        if not len(object.maps) or (len(object.maps) == 1 and not len(object.maps[0])):
+            stream.write(repr(object))
+            return
+
+        stream.write(object.__class__.__name__ + "(")
+        self._format_items(object.maps, stream, indent, allowance, context, level)
+        stream.write(")")
+
+    _dispatch[_collections.ChainMap.__repr__] = _pprint_chain_map
+
+    def _pprint_deque(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        stream.write(object.__class__.__name__ + "(")
+        if object.maxlen is not None:
+            stream.write("maxlen=%d, " % object.maxlen)
+        stream.write("[")
+
+        self._format_items(object, stream, indent, allowance + 1, context, level)
+        stream.write("])")
+
+    _dispatch[_collections.deque.__repr__] = _pprint_deque
+
+    def _pprint_user_dict(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        self._format(object.data, stream, indent, allowance, context, level - 1)
+
+    _dispatch[_collections.UserDict.__repr__] = _pprint_user_dict
+
+    def _pprint_user_list(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        self._format(object.data, stream, indent, allowance, context, level - 1)
+
+    _dispatch[_collections.UserList.__repr__] = _pprint_user_list
+
+    def _pprint_user_string(
+        self,
+        object: Any,
+        stream: IO[str],
+        indent: int,
+        allowance: int,
+        context: set[int],
+        level: int,
+    ) -> None:
+        self._format(object.data, stream, indent, allowance, context, level - 1)
+
+    _dispatch[_collections.UserString.__repr__] = _pprint_user_string
+
+    def _safe_repr(
+        self, object: Any, context: set[int], maxlevels: int | None, level: int
+    ) -> str:
+        typ = type(object)
+        if typ in _builtin_scalars:
+            return repr(object)
+
+        r = getattr(typ, "__repr__", None)
+
+        if issubclass(typ, dict) and r is dict.__repr__:
+            if not object:
+                return "{}"
+            objid = id(object)
+            if maxlevels and level >= maxlevels:
+                return "{...}"
+            if objid in context:
+                return _recursion(object)
+            context.add(objid)
+            components: list[str] = []
+            append = components.append
+            level += 1
+            for k, v in sorted(object.items(), key=_safe_tuple):
+                krepr = self._safe_repr(k, context, maxlevels, level)
+                vrepr = self._safe_repr(v, context, maxlevels, level)
+                append(f"{krepr}: {vrepr}")
+            context.remove(objid)
+            return "{{{}}}".format(", ".join(components))
+
+        if (issubclass(typ, list) and r is list.__repr__) or (
+            issubclass(typ, tuple) and r is tuple.__repr__
+        ):
+            if issubclass(typ, list):
+                if not object:
+                    return "[]"
+                format = "[%s]"
+            elif len(object) == 1:
+                format = "(%s,)"
+            else:
+                if not object:
+                    return "()"
+                format = "(%s)"
+            objid = id(object)
+            if maxlevels and level >= maxlevels:
+                return format % "..."
+            if objid in context:
+                return _recursion(object)
+            context.add(objid)
+            components = []
+            append = components.append
+            level += 1
+            for o in object:
+                orepr = self._safe_repr(o, context, maxlevels, level)
+                append(orepr)
+            context.remove(objid)
+            return format % ", ".join(components)
+
+        return repr(object)
+
+
+_builtin_scalars = frozenset(
+    {str, bytes, bytearray, float, complex, bool, type(None), int}
+)
+
+
+def _recursion(object: Any) -> str:
+    return f"<Recursion on {type(object).__name__} with id={id(object)}>"
+
+
+def _wrap_bytes_repr(object: Any, width: int, allowance: int) -> Iterator[str]:
+    current = b""
+    last = len(object) // 4 * 4
+    for i in range(0, len(object), 4):
+        part = object[i : i + 4]
+        candidate = current + part
+        if i == last:
+            width -= allowance
+        if len(repr(candidate)) > width:
+            if current:
+                yield repr(current)
+            current = part
+        else:
+            current = candidate
+    if current:
+        yield repr(current)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/saferepr.py b/backend/venv/lib/python3.12/site-packages/_pytest/_io/saferepr.py
new file mode 100644
index 00000000..cee70e33
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_io/saferepr.py
@@ -0,0 +1,130 @@
+from __future__ import annotations
+
+import pprint
+import reprlib
+
+
+def _try_repr_or_str(obj: object) -> str:
+    try:
+        return repr(obj)
+    except (KeyboardInterrupt, SystemExit):
+        raise
+    except BaseException:
+        return f'{type(obj).__name__}("{obj}")'
+
+
+def _format_repr_exception(exc: BaseException, obj: object) -> str:
+    try:
+        exc_info = _try_repr_or_str(exc)
+    except (KeyboardInterrupt, SystemExit):
+        raise
+    except BaseException as inner_exc:
+        exc_info = f"unpresentable exception ({_try_repr_or_str(inner_exc)})"
+    return (
+        f"<[{exc_info} raised in repr()] {type(obj).__name__} object at 0x{id(obj):x}>"
+    )
+
+
+def _ellipsize(s: str, maxsize: int) -> str:
+    if len(s) > maxsize:
+        i = max(0, (maxsize - 3) // 2)
+        j = max(0, maxsize - 3 - i)
+        return s[:i] + "..." + s[len(s) - j :]
+    return s
+
+
+class SafeRepr(reprlib.Repr):
+    """
+    repr.Repr that limits the resulting size of repr() and includes
+    information on exceptions raised during the call.
+    """
+
+    def __init__(self, maxsize: int | None, use_ascii: bool = False) -> None:
+        """
+        :param maxsize:
+            If not None, will truncate the resulting repr to that specific size, using ellipsis
+            somewhere in the middle to hide the extra text.
+            If None, will not impose any size limits on the returning repr.
+        """
+        super().__init__()
+        # ``maxstring`` is used by the superclass, and needs to be an int; using a
+        # very large number in case maxsize is None, meaning we want to disable
+        # truncation.
+        self.maxstring = maxsize if maxsize is not None else 1_000_000_000
+        self.maxsize = maxsize
+        self.use_ascii = use_ascii
+
+    def repr(self, x: object) -> str:
+        try:
+            if self.use_ascii:
+                s = ascii(x)
+            else:
+                s = super().repr(x)
+        except (KeyboardInterrupt, SystemExit):
+            raise
+        except BaseException as exc:
+            s = _format_repr_exception(exc, x)
+        if self.maxsize is not None:
+            s = _ellipsize(s, self.maxsize)
+        return s
+
+    def repr_instance(self, x: object, level: int) -> str:
+        try:
+            s = repr(x)
+        except (KeyboardInterrupt, SystemExit):
+            raise
+        except BaseException as exc:
+            s = _format_repr_exception(exc, x)
+        if self.maxsize is not None:
+            s = _ellipsize(s, self.maxsize)
+        return s
+
+
+def safeformat(obj: object) -> str:
+    """Return a pretty printed string for the given object.
+
+    Failing __repr__ functions of user instances will be represented
+    with a short exception info.
+    """
+    try:
+        return pprint.pformat(obj)
+    except Exception as exc:
+        return _format_repr_exception(exc, obj)
+
+
+# Maximum size of overall repr of objects to display during assertion errors.
+DEFAULT_REPR_MAX_SIZE = 240
+
+
+def saferepr(
+    obj: object, maxsize: int | None = DEFAULT_REPR_MAX_SIZE, use_ascii: bool = False
+) -> str:
+    """Return a size-limited safe repr-string for the given object.
+
+    Failing __repr__ functions of user instances will be represented
+    with a short exception info and 'saferepr' generally takes
+    care to never raise exceptions itself.
+
+    This function is a wrapper around the Repr/reprlib functionality of the
+    stdlib.
+    """
+    return SafeRepr(maxsize, use_ascii).repr(obj)
+
+
+def saferepr_unlimited(obj: object, use_ascii: bool = True) -> str:
+    """Return an unlimited-size safe repr-string for the given object.
+
+    As with saferepr, failing __repr__ functions of user instances
+    will be represented with a short exception info.
+
+    This function is a wrapper around simple repr.
+
+    Note: a cleaner solution would be to alter ``saferepr``this way
+    when maxsize=None, but that might affect some other code.
+    """
+    try:
+        if use_ascii:
+            return ascii(obj)
+        return repr(obj)
+    except Exception as exc:
+        return _format_repr_exception(exc, obj)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/terminalwriter.py b/backend/venv/lib/python3.12/site-packages/_pytest/_io/terminalwriter.py
new file mode 100644
index 00000000..70ebd3d0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_io/terminalwriter.py
@@ -0,0 +1,274 @@
+"""Helper functions for writing to terminals and files."""
+
+from __future__ import annotations
+
+import os
+import shutil
+import sys
+from typing import final
+from typing import Literal
+from typing import Sequence
+from typing import TextIO
+from typing import TYPE_CHECKING
+
+from ..compat import assert_never
+from .wcwidth import wcswidth
+
+
+if TYPE_CHECKING:
+    from pygments.formatter import Formatter
+    from pygments.lexer import Lexer
+
+
+# This code was initially copied from py 1.8.1, file _io/terminalwriter.py.
+
+
+def get_terminal_width() -> int:
+    width, _ = shutil.get_terminal_size(fallback=(80, 24))
+
+    # The Windows get_terminal_size may be bogus, let's sanify a bit.
+    if width < 40:
+        width = 80
+
+    return width
+
+
+def should_do_markup(file: TextIO) -> bool:
+    if os.environ.get("PY_COLORS") == "1":
+        return True
+    if os.environ.get("PY_COLORS") == "0":
+        return False
+    if os.environ.get("NO_COLOR"):
+        return False
+    if os.environ.get("FORCE_COLOR"):
+        return True
+    return (
+        hasattr(file, "isatty") and file.isatty() and os.environ.get("TERM") != "dumb"
+    )
+
+
+@final
+class TerminalWriter:
+    _esctable = dict(
+        black=30,
+        red=31,
+        green=32,
+        yellow=33,
+        blue=34,
+        purple=35,
+        cyan=36,
+        white=37,
+        Black=40,
+        Red=41,
+        Green=42,
+        Yellow=43,
+        Blue=44,
+        Purple=45,
+        Cyan=46,
+        White=47,
+        bold=1,
+        light=2,
+        blink=5,
+        invert=7,
+    )
+
+    def __init__(self, file: TextIO | None = None) -> None:
+        if file is None:
+            file = sys.stdout
+        if hasattr(file, "isatty") and file.isatty() and sys.platform == "win32":
+            try:
+                import colorama
+            except ImportError:
+                pass
+            else:
+                file = colorama.AnsiToWin32(file).stream
+                assert file is not None
+        self._file = file
+        self.hasmarkup = should_do_markup(file)
+        self._current_line = ""
+        self._terminal_width: int | None = None
+        self.code_highlight = True
+
+    @property
+    def fullwidth(self) -> int:
+        if self._terminal_width is not None:
+            return self._terminal_width
+        return get_terminal_width()
+
+    @fullwidth.setter
+    def fullwidth(self, value: int) -> None:
+        self._terminal_width = value
+
+    @property
+    def width_of_current_line(self) -> int:
+        """Return an estimate of the width so far in the current line."""
+        return wcswidth(self._current_line)
+
+    def markup(self, text: str, **markup: bool) -> str:
+        for name in markup:
+            if name not in self._esctable:
+                raise ValueError(f"unknown markup: {name!r}")
+        if self.hasmarkup:
+            esc = [self._esctable[name] for name, on in markup.items() if on]
+            if esc:
+                text = "".join(f"\x1b[{cod}m" for cod in esc) + text + "\x1b[0m"
+        return text
+
+    def sep(
+        self,
+        sepchar: str,
+        title: str | None = None,
+        fullwidth: int | None = None,
+        **markup: bool,
+    ) -> None:
+        if fullwidth is None:
+            fullwidth = self.fullwidth
+        # The goal is to have the line be as long as possible
+        # under the condition that len(line) <= fullwidth.
+        if sys.platform == "win32":
+            # If we print in the last column on windows we are on a
+            # new line but there is no way to verify/neutralize this
+            # (we may not know the exact line width).
+            # So let's be defensive to avoid empty lines in the output.
+            fullwidth -= 1
+        if title is not None:
+            # we want 2 + 2*len(fill) + len(title) <= fullwidth
+            # i.e.    2 + 2*len(sepchar)*N + len(title) <= fullwidth
+            #         2*len(sepchar)*N <= fullwidth - len(title) - 2
+            #         N <= (fullwidth - len(title) - 2) // (2*len(sepchar))
+            N = max((fullwidth - len(title) - 2) // (2 * len(sepchar)), 1)
+            fill = sepchar * N
+            line = f"{fill} {title} {fill}"
+        else:
+            # we want len(sepchar)*N <= fullwidth
+            # i.e.    N <= fullwidth // len(sepchar)
+            line = sepchar * (fullwidth // len(sepchar))
+        # In some situations there is room for an extra sepchar at the right,
+        # in particular if we consider that with a sepchar like "_ " the
+        # trailing space is not important at the end of the line.
+        if len(line) + len(sepchar.rstrip()) <= fullwidth:
+            line += sepchar.rstrip()
+
+        self.line(line, **markup)
+
+    def write(self, msg: str, *, flush: bool = False, **markup: bool) -> None:
+        if msg:
+            current_line = msg.rsplit("\n", 1)[-1]
+            if "\n" in msg:
+                self._current_line = current_line
+            else:
+                self._current_line += current_line
+
+            msg = self.markup(msg, **markup)
+
+            try:
+                self._file.write(msg)
+            except UnicodeEncodeError:
+                # Some environments don't support printing general Unicode
+                # strings, due to misconfiguration or otherwise; in that case,
+                # print the string escaped to ASCII.
+                # When the Unicode situation improves we should consider
+                # letting the error propagate instead of masking it (see #7475
+                # for one brief attempt).
+                msg = msg.encode("unicode-escape").decode("ascii")
+                self._file.write(msg)
+
+            if flush:
+                self.flush()
+
+    def line(self, s: str = "", **markup: bool) -> None:
+        self.write(s, **markup)
+        self.write("\n")
+
+    def flush(self) -> None:
+        self._file.flush()
+
+    def _write_source(self, lines: Sequence[str], indents: Sequence[str] = ()) -> None:
+        """Write lines of source code possibly highlighted.
+
+        Keeping this private for now because the API is clunky. We should discuss how
+        to evolve the terminal writer so we can have more precise color support, for example
+        being able to write part of a line in one color and the rest in another, and so on.
+        """
+        if indents and len(indents) != len(lines):
+            raise ValueError(
+                f"indents size ({len(indents)}) should have same size as lines ({len(lines)})"
+            )
+        if not indents:
+            indents = [""] * len(lines)
+        source = "\n".join(lines)
+        new_lines = self._highlight(source).splitlines()
+        for indent, new_line in zip(indents, new_lines):
+            self.line(indent + new_line)
+
+    def _get_pygments_lexer(self, lexer: Literal["python", "diff"]) -> Lexer | None:
+        try:
+            if lexer == "python":
+                from pygments.lexers.python import PythonLexer
+
+                return PythonLexer()
+            elif lexer == "diff":
+                from pygments.lexers.diff import DiffLexer
+
+                return DiffLexer()
+            else:
+                assert_never(lexer)
+        except ModuleNotFoundError:
+            return None
+
+    def _get_pygments_formatter(self) -> Formatter | None:
+        try:
+            import pygments.util
+        except ModuleNotFoundError:
+            return None
+
+        from _pytest.config.exceptions import UsageError
+
+        theme = os.getenv("PYTEST_THEME")
+        theme_mode = os.getenv("PYTEST_THEME_MODE", "dark")
+
+        try:
+            from pygments.formatters.terminal import TerminalFormatter
+
+            return TerminalFormatter(bg=theme_mode, style=theme)
+
+        except pygments.util.ClassNotFound as e:
+            raise UsageError(
+                f"PYTEST_THEME environment variable has an invalid value: '{theme}'. "
+                "Hint: See available pygments styles with `pygmentize -L styles`."
+            ) from e
+        except pygments.util.OptionError as e:
+            raise UsageError(
+                f"PYTEST_THEME_MODE environment variable has an invalid value: '{theme_mode}'. "
+                "The allowed values are 'dark' (default) and 'light'."
+            ) from e
+
+    def _highlight(
+        self, source: str, lexer: Literal["diff", "python"] = "python"
+    ) -> str:
+        """Highlight the given source if we have markup support."""
+        if not source or not self.hasmarkup or not self.code_highlight:
+            return source
+
+        pygments_lexer = self._get_pygments_lexer(lexer)
+        if pygments_lexer is None:
+            return source
+
+        pygments_formatter = self._get_pygments_formatter()
+        if pygments_formatter is None:
+            return source
+
+        from pygments import highlight
+
+        highlighted: str = highlight(source, pygments_lexer, pygments_formatter)
+        # pygments terminal formatter may add a newline when there wasn't one.
+        # We don't want this, remove.
+        if highlighted[-1] == "\n" and source[-1] != "\n":
+            highlighted = highlighted[:-1]
+
+        # Some lexers will not set the initial color explicitly
+        # which may lead to the previous color being propagated to the
+        # start of the expression, so reset first.
+        highlighted = "\x1b[0m" + highlighted
+
+        return highlighted
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_io/wcwidth.py b/backend/venv/lib/python3.12/site-packages/_pytest/_io/wcwidth.py
new file mode 100644
index 00000000..23886ff1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_io/wcwidth.py
@@ -0,0 +1,57 @@
+from __future__ import annotations
+
+from functools import lru_cache
+import unicodedata
+
+
+@lru_cache(100)
+def wcwidth(c: str) -> int:
+    """Determine how many columns are needed to display a character in a terminal.
+
+    Returns -1 if the character is not printable.
+    Returns 0, 1 or 2 for other characters.
+    """
+    o = ord(c)
+
+    # ASCII fast path.
+    if 0x20 <= o < 0x07F:
+        return 1
+
+    # Some Cf/Zp/Zl characters which should be zero-width.
+    if (
+        o == 0x0000
+        or 0x200B <= o <= 0x200F
+        or 0x2028 <= o <= 0x202E
+        or 0x2060 <= o <= 0x2063
+    ):
+        return 0
+
+    category = unicodedata.category(c)
+
+    # Control characters.
+    if category == "Cc":
+        return -1
+
+    # Combining characters with zero width.
+    if category in ("Me", "Mn"):
+        return 0
+
+    # Full/Wide east asian characters.
+    if unicodedata.east_asian_width(c) in ("F", "W"):
+        return 2
+
+    return 1
+
+
+def wcswidth(s: str) -> int:
+    """Determine how many columns are needed to display a string in a terminal.
+
+    Returns -1 if the string contains non-printable characters.
+    """
+    width = 0
+    for c in unicodedata.normalize("NFC", s):
+        wc = wcwidth(c)
+        if wc < 0:
+            return -1
+        width += wc
+    return width
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..00b87401
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/error.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/error.cpython-312.pyc
new file mode 100644
index 00000000..c3ab5225
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/error.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/path.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/path.cpython-312.pyc
new file mode 100644
index 00000000..301639a6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/_py/__pycache__/path.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/error.py b/backend/venv/lib/python3.12/site-packages/_pytest/_py/error.py
new file mode 100644
index 00000000..ab3a4ed3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_py/error.py
@@ -0,0 +1,111 @@
+"""create errno-specific classes for IO or os calls."""
+
+from __future__ import annotations
+
+import errno
+import os
+import sys
+from typing import Callable
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+
+if TYPE_CHECKING:
+    from typing_extensions import ParamSpec
+
+    P = ParamSpec("P")
+
+R = TypeVar("R")
+
+
+class Error(EnvironmentError):
+    def __repr__(self) -> str:
+        return "{}.{} {!r}: {} ".format(
+            self.__class__.__module__,
+            self.__class__.__name__,
+            self.__class__.__doc__,
+            " ".join(map(str, self.args)),
+            # repr(self.args)
+        )
+
+    def __str__(self) -> str:
+        s = "[{}]: {}".format(
+            self.__class__.__doc__,
+            " ".join(map(str, self.args)),
+        )
+        return s
+
+
+_winerrnomap = {
+    2: errno.ENOENT,
+    3: errno.ENOENT,
+    17: errno.EEXIST,
+    18: errno.EXDEV,
+    13: errno.EBUSY,  # empty cd drive, but ENOMEDIUM seems unavailable
+    22: errno.ENOTDIR,
+    20: errno.ENOTDIR,
+    267: errno.ENOTDIR,
+    5: errno.EACCES,  # anything better?
+}
+
+
+class ErrorMaker:
+    """lazily provides Exception classes for each possible POSIX errno
+    (as defined per the 'errno' module).  All such instances
+    subclass EnvironmentError.
+    """
+
+    _errno2class: dict[int, type[Error]] = {}
+
+    def __getattr__(self, name: str) -> type[Error]:
+        if name[0] == "_":
+            raise AttributeError(name)
+        eno = getattr(errno, name)
+        cls = self._geterrnoclass(eno)
+        setattr(self, name, cls)
+        return cls
+
+    def _geterrnoclass(self, eno: int) -> type[Error]:
+        try:
+            return self._errno2class[eno]
+        except KeyError:
+            clsname = errno.errorcode.get(eno, "UnknownErrno%d" % (eno,))
+            errorcls = type(
+                clsname,
+                (Error,),
+                {"__module__": "py.error", "__doc__": os.strerror(eno)},
+            )
+            self._errno2class[eno] = errorcls
+            return errorcls
+
+    def checked_call(
+        self, func: Callable[P, R], *args: P.args, **kwargs: P.kwargs
+    ) -> R:
+        """Call a function and raise an errno-exception if applicable."""
+        __tracebackhide__ = True
+        try:
+            return func(*args, **kwargs)
+        except Error:
+            raise
+        except OSError as value:
+            if not hasattr(value, "errno"):
+                raise
+            errno = value.errno
+            if sys.platform == "win32":
+                try:
+                    cls = self._geterrnoclass(_winerrnomap[errno])
+                except KeyError:
+                    raise value
+            else:
+                # we are not on Windows, or we got a proper OSError
+                cls = self._geterrnoclass(errno)
+
+            raise cls(f"{func.__name__}{args!r}")
+
+
+_error_maker = ErrorMaker()
+checked_call = _error_maker.checked_call
+
+
+def __getattr__(attr: str) -> type[Error]:
+    return getattr(_error_maker, attr)  # type: ignore[no-any-return]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_py/path.py b/backend/venv/lib/python3.12/site-packages/_pytest/_py/path.py
new file mode 100644
index 00000000..c7ab1182
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_py/path.py
@@ -0,0 +1,1475 @@
+# mypy: allow-untyped-defs
+"""local path implementation."""
+
+from __future__ import annotations
+
+import atexit
+from contextlib import contextmanager
+import fnmatch
+import importlib.util
+import io
+import os
+from os.path import abspath
+from os.path import dirname
+from os.path import exists
+from os.path import isabs
+from os.path import isdir
+from os.path import isfile
+from os.path import islink
+from os.path import normpath
+import posixpath
+from stat import S_ISDIR
+from stat import S_ISLNK
+from stat import S_ISREG
+import sys
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Literal
+from typing import overload
+from typing import TYPE_CHECKING
+import uuid
+import warnings
+
+from . import error
+
+
+# Moved from local.py.
+iswin32 = sys.platform == "win32" or (getattr(os, "_name", False) == "nt")
+
+
+class Checkers:
+    _depend_on_existence = "exists", "link", "dir", "file"
+
+    def __init__(self, path):
+        self.path = path
+
+    def dotfile(self):
+        return self.path.basename.startswith(".")
+
+    def ext(self, arg):
+        if not arg.startswith("."):
+            arg = "." + arg
+        return self.path.ext == arg
+
+    def basename(self, arg):
+        return self.path.basename == arg
+
+    def basestarts(self, arg):
+        return self.path.basename.startswith(arg)
+
+    def relto(self, arg):
+        return self.path.relto(arg)
+
+    def fnmatch(self, arg):
+        return self.path.fnmatch(arg)
+
+    def endswith(self, arg):
+        return str(self.path).endswith(arg)
+
+    def _evaluate(self, kw):
+        from .._code.source import getrawcode
+
+        for name, value in kw.items():
+            invert = False
+            meth = None
+            try:
+                meth = getattr(self, name)
+            except AttributeError:
+                if name[:3] == "not":
+                    invert = True
+                    try:
+                        meth = getattr(self, name[3:])
+                    except AttributeError:
+                        pass
+            if meth is None:
+                raise TypeError(f"no {name!r} checker available for {self.path!r}")
+            try:
+                if getrawcode(meth).co_argcount > 1:
+                    if (not meth(value)) ^ invert:
+                        return False
+                else:
+                    if bool(value) ^ bool(meth()) ^ invert:
+                        return False
+            except (error.ENOENT, error.ENOTDIR, error.EBUSY):
+                # EBUSY feels not entirely correct,
+                # but its kind of necessary since ENOMEDIUM
+                # is not accessible in python
+                for name in self._depend_on_existence:
+                    if name in kw:
+                        if kw.get(name):
+                            return False
+                    name = "not" + name
+                    if name in kw:
+                        if not kw.get(name):
+                            return False
+        return True
+
+    _statcache: Stat
+
+    def _stat(self) -> Stat:
+        try:
+            return self._statcache
+        except AttributeError:
+            try:
+                self._statcache = self.path.stat()
+            except error.ELOOP:
+                self._statcache = self.path.lstat()
+            return self._statcache
+
+    def dir(self):
+        return S_ISDIR(self._stat().mode)
+
+    def file(self):
+        return S_ISREG(self._stat().mode)
+
+    def exists(self):
+        return self._stat()
+
+    def link(self):
+        st = self.path.lstat()
+        return S_ISLNK(st.mode)
+
+
+class NeverRaised(Exception):
+    pass
+
+
+class Visitor:
+    def __init__(self, fil, rec, ignore, bf, sort):
+        if isinstance(fil, str):
+            fil = FNMatcher(fil)
+        if isinstance(rec, str):
+            self.rec: Callable[[LocalPath], bool] = FNMatcher(rec)
+        elif not hasattr(rec, "__call__") and rec:
+            self.rec = lambda path: True
+        else:
+            self.rec = rec
+        self.fil = fil
+        self.ignore = ignore
+        self.breadthfirst = bf
+        self.optsort = cast(Callable[[Any], Any], sorted) if sort else (lambda x: x)
+
+    def gen(self, path):
+        try:
+            entries = path.listdir()
+        except self.ignore:
+            return
+        rec = self.rec
+        dirs = self.optsort(
+            [p for p in entries if p.check(dir=1) and (rec is None or rec(p))]
+        )
+        if not self.breadthfirst:
+            for subdir in dirs:
+                yield from self.gen(subdir)
+        for p in self.optsort(entries):
+            if self.fil is None or self.fil(p):
+                yield p
+        if self.breadthfirst:
+            for subdir in dirs:
+                yield from self.gen(subdir)
+
+
+class FNMatcher:
+    def __init__(self, pattern):
+        self.pattern = pattern
+
+    def __call__(self, path):
+        pattern = self.pattern
+
+        if (
+            pattern.find(path.sep) == -1
+            and iswin32
+            and pattern.find(posixpath.sep) != -1
+        ):
+            # Running on Windows, the pattern has no Windows path separators,
+            # and the pattern has one or more Posix path separators. Replace
+            # the Posix path separators with the Windows path separator.
+            pattern = pattern.replace(posixpath.sep, path.sep)
+
+        if pattern.find(path.sep) == -1:
+            name = path.basename
+        else:
+            name = str(path)  # path.strpath # XXX svn?
+            if not os.path.isabs(pattern):
+                pattern = "*" + path.sep + pattern
+        return fnmatch.fnmatch(name, pattern)
+
+
+def map_as_list(func, iter):
+    return list(map(func, iter))
+
+
+class Stat:
+    if TYPE_CHECKING:
+
+        @property
+        def size(self) -> int: ...
+
+        @property
+        def mtime(self) -> float: ...
+
+    def __getattr__(self, name: str) -> Any:
+        return getattr(self._osstatresult, "st_" + name)
+
+    def __init__(self, path, osstatresult):
+        self.path = path
+        self._osstatresult = osstatresult
+
+    @property
+    def owner(self):
+        if iswin32:
+            raise NotImplementedError("XXX win32")
+        import pwd
+
+        entry = error.checked_call(pwd.getpwuid, self.uid)  # type:ignore[attr-defined,unused-ignore]
+        return entry[0]
+
+    @property
+    def group(self):
+        """Return group name of file."""
+        if iswin32:
+            raise NotImplementedError("XXX win32")
+        import grp
+
+        entry = error.checked_call(grp.getgrgid, self.gid)  # type:ignore[attr-defined,unused-ignore]
+        return entry[0]
+
+    def isdir(self):
+        return S_ISDIR(self._osstatresult.st_mode)
+
+    def isfile(self):
+        return S_ISREG(self._osstatresult.st_mode)
+
+    def islink(self):
+        self.path.lstat()
+        return S_ISLNK(self._osstatresult.st_mode)
+
+
+def getuserid(user):
+    import pwd
+
+    if not isinstance(user, int):
+        user = pwd.getpwnam(user)[2]  # type:ignore[attr-defined,unused-ignore]
+    return user
+
+
+def getgroupid(group):
+    import grp
+
+    if not isinstance(group, int):
+        group = grp.getgrnam(group)[2]  # type:ignore[attr-defined,unused-ignore]
+    return group
+
+
+class LocalPath:
+    """Object oriented interface to os.path and other local filesystem
+    related information.
+    """
+
+    class ImportMismatchError(ImportError):
+        """raised on pyimport() if there is a mismatch of __file__'s"""
+
+    sep = os.sep
+
+    def __init__(self, path=None, expanduser=False):
+        """Initialize and return a local Path instance.
+
+        Path can be relative to the current directory.
+        If path is None it defaults to the current working directory.
+        If expanduser is True, tilde-expansion is performed.
+        Note that Path instances always carry an absolute path.
+        Note also that passing in a local path object will simply return
+        the exact same path object. Use new() to get a new copy.
+        """
+        if path is None:
+            self.strpath = error.checked_call(os.getcwd)
+        else:
+            try:
+                path = os.fspath(path)
+            except TypeError:
+                raise ValueError(
+                    "can only pass None, Path instances "
+                    "or non-empty strings to LocalPath"
+                )
+            if expanduser:
+                path = os.path.expanduser(path)
+            self.strpath = abspath(path)
+
+    if sys.platform != "win32":
+
+        def chown(self, user, group, rec=0):
+            """Change ownership to the given user and group.
+            user and group may be specified by a number or
+            by a name.  if rec is True change ownership
+            recursively.
+            """
+            uid = getuserid(user)
+            gid = getgroupid(group)
+            if rec:
+                for x in self.visit(rec=lambda x: x.check(link=0)):
+                    if x.check(link=0):
+                        error.checked_call(os.chown, str(x), uid, gid)
+            error.checked_call(os.chown, str(self), uid, gid)
+
+        def readlink(self) -> str:
+            """Return value of a symbolic link."""
+            # https://github.com/python/mypy/issues/12278
+            return error.checked_call(os.readlink, self.strpath)  # type: ignore[arg-type,return-value,unused-ignore]
+
+        def mklinkto(self, oldname):
+            """Posix style hard link to another name."""
+            error.checked_call(os.link, str(oldname), str(self))
+
+        def mksymlinkto(self, value, absolute=1):
+            """Create a symbolic link with the given value (pointing to another name)."""
+            if absolute:
+                error.checked_call(os.symlink, str(value), self.strpath)
+            else:
+                base = self.common(value)
+                # with posix local paths '/' is always a common base
+                relsource = self.__class__(value).relto(base)
+                reldest = self.relto(base)
+                n = reldest.count(self.sep)
+                target = self.sep.join(("..",) * n + (relsource,))
+                error.checked_call(os.symlink, target, self.strpath)
+
+    def __div__(self, other):
+        return self.join(os.fspath(other))
+
+    __truediv__ = __div__  # py3k
+
+    @property
+    def basename(self):
+        """Basename part of path."""
+        return self._getbyspec("basename")[0]
+
+    @property
+    def dirname(self):
+        """Dirname part of path."""
+        return self._getbyspec("dirname")[0]
+
+    @property
+    def purebasename(self):
+        """Pure base name of the path."""
+        return self._getbyspec("purebasename")[0]
+
+    @property
+    def ext(self):
+        """Extension of the path (including the '.')."""
+        return self._getbyspec("ext")[0]
+
+    def read_binary(self):
+        """Read and return a bytestring from reading the path."""
+        with self.open("rb") as f:
+            return f.read()
+
+    def read_text(self, encoding):
+        """Read and return a Unicode string from reading the path."""
+        with self.open("r", encoding=encoding) as f:
+            return f.read()
+
+    def read(self, mode="r"):
+        """Read and return a bytestring from reading the path."""
+        with self.open(mode) as f:
+            return f.read()
+
+    def readlines(self, cr=1):
+        """Read and return a list of lines from the path. if cr is False, the
+        newline will be removed from the end of each line."""
+        mode = "r"
+
+        if not cr:
+            content = self.read(mode)
+            return content.split("\n")
+        else:
+            f = self.open(mode)
+            try:
+                return f.readlines()
+            finally:
+                f.close()
+
+    def load(self):
+        """(deprecated) return object unpickled from self.read()"""
+        f = self.open("rb")
+        try:
+            import pickle
+
+            return error.checked_call(pickle.load, f)
+        finally:
+            f.close()
+
+    def move(self, target):
+        """Move this path to target."""
+        if target.relto(self):
+            raise error.EINVAL(target, "cannot move path into a subdirectory of itself")
+        try:
+            self.rename(target)
+        except error.EXDEV:  # invalid cross-device link
+            self.copy(target)
+            self.remove()
+
+    def fnmatch(self, pattern):
+        """Return true if the basename/fullname matches the glob-'pattern'.
+
+        valid pattern characters::
+
+            *       matches everything
+            ?       matches any single character
+            [seq]   matches any character in seq
+            [!seq]  matches any char not in seq
+
+        If the pattern contains a path-separator then the full path
+        is used for pattern matching and a '*' is prepended to the
+        pattern.
+
+        if the pattern doesn't contain a path-separator the pattern
+        is only matched against the basename.
+        """
+        return FNMatcher(pattern)(self)
+
+    def relto(self, relpath):
+        """Return a string which is the relative part of the path
+        to the given 'relpath'.
+        """
+        if not isinstance(relpath, (str, LocalPath)):
+            raise TypeError(f"{relpath!r}: not a string or path object")
+        strrelpath = str(relpath)
+        if strrelpath and strrelpath[-1] != self.sep:
+            strrelpath += self.sep
+        # assert strrelpath[-1] == self.sep
+        # assert strrelpath[-2] != self.sep
+        strself = self.strpath
+        if sys.platform == "win32" or getattr(os, "_name", None) == "nt":
+            if os.path.normcase(strself).startswith(os.path.normcase(strrelpath)):
+                return strself[len(strrelpath) :]
+        elif strself.startswith(strrelpath):
+            return strself[len(strrelpath) :]
+        return ""
+
+    def ensure_dir(self, *args):
+        """Ensure the path joined with args is a directory."""
+        return self.ensure(*args, dir=True)
+
+    def bestrelpath(self, dest):
+        """Return a string which is a relative path from self
+        (assumed to be a directory) to dest such that
+        self.join(bestrelpath) == dest and if not such
+        path can be determined return dest.
+        """
+        try:
+            if self == dest:
+                return os.curdir
+            base = self.common(dest)
+            if not base:  # can be the case on windows
+                return str(dest)
+            self2base = self.relto(base)
+            reldest = dest.relto(base)
+            if self2base:
+                n = self2base.count(self.sep) + 1
+            else:
+                n = 0
+            lst = [os.pardir] * n
+            if reldest:
+                lst.append(reldest)
+            target = dest.sep.join(lst)
+            return target
+        except AttributeError:
+            return str(dest)
+
+    def exists(self):
+        return self.check()
+
+    def isdir(self):
+        return self.check(dir=1)
+
+    def isfile(self):
+        return self.check(file=1)
+
+    def parts(self, reverse=False):
+        """Return a root-first list of all ancestor directories
+        plus the path itself.
+        """
+        current = self
+        lst = [self]
+        while 1:
+            last = current
+            current = current.dirpath()
+            if last == current:
+                break
+            lst.append(current)
+        if not reverse:
+            lst.reverse()
+        return lst
+
+    def common(self, other):
+        """Return the common part shared with the other path
+        or None if there is no common part.
+        """
+        last = None
+        for x, y in zip(self.parts(), other.parts()):
+            if x != y:
+                return last
+            last = x
+        return last
+
+    def __add__(self, other):
+        """Return new path object with 'other' added to the basename"""
+        return self.new(basename=self.basename + str(other))
+
+    def visit(self, fil=None, rec=None, ignore=NeverRaised, bf=False, sort=False):
+        """Yields all paths below the current one
+
+        fil is a filter (glob pattern or callable), if not matching the
+        path will not be yielded, defaulting to None (everything is
+        returned)
+
+        rec is a filter (glob pattern or callable) that controls whether
+        a node is descended, defaulting to None
+
+        ignore is an Exception class that is ignoredwhen calling dirlist()
+        on any of the paths (by default, all exceptions are reported)
+
+        bf if True will cause a breadthfirst search instead of the
+        default depthfirst. Default: False
+
+        sort if True will sort entries within each directory level.
+        """
+        yield from Visitor(fil, rec, ignore, bf, sort).gen(self)
+
+    def _sortlist(self, res, sort):
+        if sort:
+            if hasattr(sort, "__call__"):
+                warnings.warn(
+                    DeprecationWarning(
+                        "listdir(sort=callable) is deprecated and breaks on python3"
+                    ),
+                    stacklevel=3,
+                )
+                res.sort(sort)
+            else:
+                res.sort()
+
+    def __fspath__(self):
+        return self.strpath
+
+    def __hash__(self):
+        s = self.strpath
+        if iswin32:
+            s = s.lower()
+        return hash(s)
+
+    def __eq__(self, other):
+        s1 = os.fspath(self)
+        try:
+            s2 = os.fspath(other)
+        except TypeError:
+            return False
+        if iswin32:
+            s1 = s1.lower()
+            try:
+                s2 = s2.lower()
+            except AttributeError:
+                return False
+        return s1 == s2
+
+    def __ne__(self, other):
+        return not (self == other)
+
+    def __lt__(self, other):
+        return os.fspath(self) < os.fspath(other)
+
+    def __gt__(self, other):
+        return os.fspath(self) > os.fspath(other)
+
+    def samefile(self, other):
+        """Return True if 'other' references the same file as 'self'."""
+        other = os.fspath(other)
+        if not isabs(other):
+            other = abspath(other)
+        if self == other:
+            return True
+        if not hasattr(os.path, "samefile"):
+            return False
+        return error.checked_call(os.path.samefile, self.strpath, other)
+
+    def remove(self, rec=1, ignore_errors=False):
+        """Remove a file or directory (or a directory tree if rec=1).
+        if ignore_errors is True, errors while removing directories will
+        be ignored.
+        """
+        if self.check(dir=1, link=0):
+            if rec:
+                # force remove of readonly files on windows
+                if iswin32:
+                    self.chmod(0o700, rec=1)
+                import shutil
+
+                error.checked_call(
+                    shutil.rmtree, self.strpath, ignore_errors=ignore_errors
+                )
+            else:
+                error.checked_call(os.rmdir, self.strpath)
+        else:
+            if iswin32:
+                self.chmod(0o700)
+            error.checked_call(os.remove, self.strpath)
+
+    def computehash(self, hashtype="md5", chunksize=524288):
+        """Return hexdigest of hashvalue for this file."""
+        try:
+            try:
+                import hashlib as mod
+            except ImportError:
+                if hashtype == "sha1":
+                    hashtype = "sha"
+                mod = __import__(hashtype)
+            hash = getattr(mod, hashtype)()
+        except (AttributeError, ImportError):
+            raise ValueError(f"Don't know how to compute {hashtype!r} hash")
+        f = self.open("rb")
+        try:
+            while 1:
+                buf = f.read(chunksize)
+                if not buf:
+                    return hash.hexdigest()
+                hash.update(buf)
+        finally:
+            f.close()
+
+    def new(self, **kw):
+        """Create a modified version of this path.
+        the following keyword arguments modify various path parts::
+
+          a:/some/path/to/a/file.ext
+          xx                           drive
+          xxxxxxxxxxxxxxxxx            dirname
+                            xxxxxxxx   basename
+                            xxxx       purebasename
+                                 xxx   ext
+        """
+        obj = object.__new__(self.__class__)
+        if not kw:
+            obj.strpath = self.strpath
+            return obj
+        drive, dirname, basename, purebasename, ext = self._getbyspec(
+            "drive,dirname,basename,purebasename,ext"
+        )
+        if "basename" in kw:
+            if "purebasename" in kw or "ext" in kw:
+                raise ValueError(f"invalid specification {kw!r}")
+        else:
+            pb = kw.setdefault("purebasename", purebasename)
+            try:
+                ext = kw["ext"]
+            except KeyError:
+                pass
+            else:
+                if ext and not ext.startswith("."):
+                    ext = "." + ext
+            kw["basename"] = pb + ext
+
+        if "dirname" in kw and not kw["dirname"]:
+            kw["dirname"] = drive
+        else:
+            kw.setdefault("dirname", dirname)
+        kw.setdefault("sep", self.sep)
+        obj.strpath = normpath("{dirname}{sep}{basename}".format(**kw))
+        return obj
+
+    def _getbyspec(self, spec: str) -> list[str]:
+        """See new for what 'spec' can be."""
+        res = []
+        parts = self.strpath.split(self.sep)
+
+        args = filter(None, spec.split(","))
+        for name in args:
+            if name == "drive":
+                res.append(parts[0])
+            elif name == "dirname":
+                res.append(self.sep.join(parts[:-1]))
+            else:
+                basename = parts[-1]
+                if name == "basename":
+                    res.append(basename)
+                else:
+                    i = basename.rfind(".")
+                    if i == -1:
+                        purebasename, ext = basename, ""
+                    else:
+                        purebasename, ext = basename[:i], basename[i:]
+                    if name == "purebasename":
+                        res.append(purebasename)
+                    elif name == "ext":
+                        res.append(ext)
+                    else:
+                        raise ValueError(f"invalid part specification {name!r}")
+        return res
+
+    def dirpath(self, *args, **kwargs):
+        """Return the directory path joined with any given path arguments."""
+        if not kwargs:
+            path = object.__new__(self.__class__)
+            path.strpath = dirname(self.strpath)
+            if args:
+                path = path.join(*args)
+            return path
+        return self.new(basename="").join(*args, **kwargs)
+
+    def join(self, *args: os.PathLike[str], abs: bool = False) -> LocalPath:
+        """Return a new path by appending all 'args' as path
+        components.  if abs=1 is used restart from root if any
+        of the args is an absolute path.
+        """
+        sep = self.sep
+        strargs = [os.fspath(arg) for arg in args]
+        strpath = self.strpath
+        if abs:
+            newargs: list[str] = []
+            for arg in reversed(strargs):
+                if isabs(arg):
+                    strpath = arg
+                    strargs = newargs
+                    break
+                newargs.insert(0, arg)
+        # special case for when we have e.g. strpath == "/"
+        actual_sep = "" if strpath.endswith(sep) else sep
+        for arg in strargs:
+            arg = arg.strip(sep)
+            if iswin32:
+                # allow unix style paths even on windows.
+                arg = arg.strip("/")
+                arg = arg.replace("/", sep)
+            strpath = strpath + actual_sep + arg
+            actual_sep = sep
+        obj = object.__new__(self.__class__)
+        obj.strpath = normpath(strpath)
+        return obj
+
+    def open(self, mode="r", ensure=False, encoding=None):
+        """Return an opened file with the given mode.
+
+        If ensure is True, create parent directories if needed.
+        """
+        if ensure:
+            self.dirpath().ensure(dir=1)
+        if encoding:
+            return error.checked_call(
+                io.open,
+                self.strpath,
+                mode,
+                encoding=encoding,
+            )
+        return error.checked_call(open, self.strpath, mode)
+
+    def _fastjoin(self, name):
+        child = object.__new__(self.__class__)
+        child.strpath = self.strpath + self.sep + name
+        return child
+
+    def islink(self):
+        return islink(self.strpath)
+
+    def check(self, **kw):
+        """Check a path for existence and properties.
+
+        Without arguments, return True if the path exists, otherwise False.
+
+        valid checkers::
+
+            file = 1  # is a file
+            file = 0  # is not a file (may not even exist)
+            dir = 1  # is a dir
+            link = 1  # is a link
+            exists = 1  # exists
+
+        You can specify multiple checker definitions, for example::
+
+            path.check(file=1, link=1)  # a link pointing to a file
+        """
+        if not kw:
+            return exists(self.strpath)
+        if len(kw) == 1:
+            if "dir" in kw:
+                return not kw["dir"] ^ isdir(self.strpath)
+            if "file" in kw:
+                return not kw["file"] ^ isfile(self.strpath)
+        if not kw:
+            kw = {"exists": 1}
+        return Checkers(self)._evaluate(kw)
+
+    _patternchars = set("*?[" + os.sep)
+
+    def listdir(self, fil=None, sort=None):
+        """List directory contents, possibly filter by the given fil func
+        and possibly sorted.
+        """
+        if fil is None and sort is None:
+            names = error.checked_call(os.listdir, self.strpath)
+            return map_as_list(self._fastjoin, names)
+        if isinstance(fil, str):
+            if not self._patternchars.intersection(fil):
+                child = self._fastjoin(fil)
+                if exists(child.strpath):
+                    return [child]
+                return []
+            fil = FNMatcher(fil)
+        names = error.checked_call(os.listdir, self.strpath)
+        res = []
+        for name in names:
+            child = self._fastjoin(name)
+            if fil is None or fil(child):
+                res.append(child)
+        self._sortlist(res, sort)
+        return res
+
+    def size(self) -> int:
+        """Return size of the underlying file object"""
+        return self.stat().size
+
+    def mtime(self) -> float:
+        """Return last modification time of the path."""
+        return self.stat().mtime
+
+    def copy(self, target, mode=False, stat=False):
+        """Copy path to target.
+
+        If mode is True, will copy permission from path to target.
+        If stat is True, copy permission, last modification
+        time, last access time, and flags from path to target.
+        """
+        if self.check(file=1):
+            if target.check(dir=1):
+                target = target.join(self.basename)
+            assert self != target
+            copychunked(self, target)
+            if mode:
+                copymode(self.strpath, target.strpath)
+            if stat:
+                copystat(self, target)
+        else:
+
+            def rec(p):
+                return p.check(link=0)
+
+            for x in self.visit(rec=rec):
+                relpath = x.relto(self)
+                newx = target.join(relpath)
+                newx.dirpath().ensure(dir=1)
+                if x.check(link=1):
+                    newx.mksymlinkto(x.readlink())
+                    continue
+                elif x.check(file=1):
+                    copychunked(x, newx)
+                elif x.check(dir=1):
+                    newx.ensure(dir=1)
+                if mode:
+                    copymode(x.strpath, newx.strpath)
+                if stat:
+                    copystat(x, newx)
+
+    def rename(self, target):
+        """Rename this path to target."""
+        target = os.fspath(target)
+        return error.checked_call(os.rename, self.strpath, target)
+
+    def dump(self, obj, bin=1):
+        """Pickle object into path location"""
+        f = self.open("wb")
+        import pickle
+
+        try:
+            error.checked_call(pickle.dump, obj, f, bin)
+        finally:
+            f.close()
+
+    def mkdir(self, *args):
+        """Create & return the directory joined with args."""
+        p = self.join(*args)
+        error.checked_call(os.mkdir, os.fspath(p))
+        return p
+
+    def write_binary(self, data, ensure=False):
+        """Write binary data into path.   If ensure is True create
+        missing parent directories.
+        """
+        if ensure:
+            self.dirpath().ensure(dir=1)
+        with self.open("wb") as f:
+            f.write(data)
+
+    def write_text(self, data, encoding, ensure=False):
+        """Write text data into path using the specified encoding.
+        If ensure is True create missing parent directories.
+        """
+        if ensure:
+            self.dirpath().ensure(dir=1)
+        with self.open("w", encoding=encoding) as f:
+            f.write(data)
+
+    def write(self, data, mode="w", ensure=False):
+        """Write data into path.   If ensure is True create
+        missing parent directories.
+        """
+        if ensure:
+            self.dirpath().ensure(dir=1)
+        if "b" in mode:
+            if not isinstance(data, bytes):
+                raise ValueError("can only process bytes")
+        else:
+            if not isinstance(data, str):
+                if not isinstance(data, bytes):
+                    data = str(data)
+                else:
+                    data = data.decode(sys.getdefaultencoding())
+        f = self.open(mode)
+        try:
+            f.write(data)
+        finally:
+            f.close()
+
+    def _ensuredirs(self):
+        parent = self.dirpath()
+        if parent == self:
+            return self
+        if parent.check(dir=0):
+            parent._ensuredirs()
+        if self.check(dir=0):
+            try:
+                self.mkdir()
+            except error.EEXIST:
+                # race condition: file/dir created by another thread/process.
+                # complain if it is not a dir
+                if self.check(dir=0):
+                    raise
+        return self
+
+    def ensure(self, *args, **kwargs):
+        """Ensure that an args-joined path exists (by default as
+        a file). if you specify a keyword argument 'dir=True'
+        then the path is forced to be a directory path.
+        """
+        p = self.join(*args)
+        if kwargs.get("dir", 0):
+            return p._ensuredirs()
+        else:
+            p.dirpath()._ensuredirs()
+            if not p.check(file=1):
+                p.open("wb").close()
+            return p
+
+    @overload
+    def stat(self, raising: Literal[True] = ...) -> Stat: ...
+
+    @overload
+    def stat(self, raising: Literal[False]) -> Stat | None: ...
+
+    def stat(self, raising: bool = True) -> Stat | None:
+        """Return an os.stat() tuple."""
+        if raising:
+            return Stat(self, error.checked_call(os.stat, self.strpath))
+        try:
+            return Stat(self, os.stat(self.strpath))
+        except KeyboardInterrupt:
+            raise
+        except Exception:
+            return None
+
+    def lstat(self) -> Stat:
+        """Return an os.lstat() tuple."""
+        return Stat(self, error.checked_call(os.lstat, self.strpath))
+
+    def setmtime(self, mtime=None):
+        """Set modification time for the given path.  if 'mtime' is None
+        (the default) then the file's mtime is set to current time.
+
+        Note that the resolution for 'mtime' is platform dependent.
+        """
+        if mtime is None:
+            return error.checked_call(os.utime, self.strpath, mtime)
+        try:
+            return error.checked_call(os.utime, self.strpath, (-1, mtime))
+        except error.EINVAL:
+            return error.checked_call(os.utime, self.strpath, (self.atime(), mtime))
+
+    def chdir(self):
+        """Change directory to self and return old current directory"""
+        try:
+            old = self.__class__()
+        except error.ENOENT:
+            old = None
+        error.checked_call(os.chdir, self.strpath)
+        return old
+
+    @contextmanager
+    def as_cwd(self):
+        """
+        Return a context manager, which changes to the path's dir during the
+        managed "with" context.
+        On __enter__ it returns the old dir, which might be ``None``.
+        """
+        old = self.chdir()
+        try:
+            yield old
+        finally:
+            if old is not None:
+                old.chdir()
+
+    def realpath(self):
+        """Return a new path which contains no symbolic links."""
+        return self.__class__(os.path.realpath(self.strpath))
+
+    def atime(self):
+        """Return last access time of the path."""
+        return self.stat().atime
+
+    def __repr__(self):
+        return f"local({self.strpath!r})"
+
+    def __str__(self):
+        """Return string representation of the Path."""
+        return self.strpath
+
+    def chmod(self, mode, rec=0):
+        """Change permissions to the given mode. If mode is an
+        integer it directly encodes the os-specific modes.
+        if rec is True perform recursively.
+        """
+        if not isinstance(mode, int):
+            raise TypeError(f"mode {mode!r} must be an integer")
+        if rec:
+            for x in self.visit(rec=rec):
+                error.checked_call(os.chmod, str(x), mode)
+        error.checked_call(os.chmod, self.strpath, mode)
+
+    def pypkgpath(self):
+        """Return the Python package path by looking for the last
+        directory upwards which still contains an __init__.py.
+        Return None if a pkgpath cannot be determined.
+        """
+        pkgpath = None
+        for parent in self.parts(reverse=True):
+            if parent.isdir():
+                if not parent.join("__init__.py").exists():
+                    break
+                if not isimportable(parent.basename):
+                    break
+                pkgpath = parent
+        return pkgpath
+
+    def _ensuresyspath(self, ensuremode, path):
+        if ensuremode:
+            s = str(path)
+            if ensuremode == "append":
+                if s not in sys.path:
+                    sys.path.append(s)
+            else:
+                if s != sys.path[0]:
+                    sys.path.insert(0, s)
+
+    def pyimport(self, modname=None, ensuresyspath=True):
+        """Return path as an imported python module.
+
+        If modname is None, look for the containing package
+        and construct an according module name.
+        The module will be put/looked up in sys.modules.
+        if ensuresyspath is True then the root dir for importing
+        the file (taking __init__.py files into account) will
+        be prepended to sys.path if it isn't there already.
+        If ensuresyspath=="append" the root dir will be appended
+        if it isn't already contained in sys.path.
+        if ensuresyspath is False no modification of syspath happens.
+
+        Special value of ensuresyspath=="importlib" is intended
+        purely for using in pytest, it is capable only of importing
+        separate .py files outside packages, e.g. for test suite
+        without any __init__.py file. It effectively allows having
+        same-named test modules in different places and offers
+        mild opt-in via this option. Note that it works only in
+        recent versions of python.
+        """
+        if not self.check():
+            raise error.ENOENT(self)
+
+        if ensuresyspath == "importlib":
+            if modname is None:
+                modname = self.purebasename
+            spec = importlib.util.spec_from_file_location(modname, str(self))
+            if spec is None or spec.loader is None:
+                raise ImportError(f"Can't find module {modname} at location {self!s}")
+            mod = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(mod)
+            return mod
+
+        pkgpath = None
+        if modname is None:
+            pkgpath = self.pypkgpath()
+            if pkgpath is not None:
+                pkgroot = pkgpath.dirpath()
+                names = self.new(ext="").relto(pkgroot).split(self.sep)
+                if names[-1] == "__init__":
+                    names.pop()
+                modname = ".".join(names)
+            else:
+                pkgroot = self.dirpath()
+                modname = self.purebasename
+
+            self._ensuresyspath(ensuresyspath, pkgroot)
+            __import__(modname)
+            mod = sys.modules[modname]
+            if self.basename == "__init__.py":
+                return mod  # we don't check anything as we might
+                # be in a namespace package ... too icky to check
+            modfile = mod.__file__
+            assert modfile is not None
+            if modfile[-4:] in (".pyc", ".pyo"):
+                modfile = modfile[:-1]
+            elif modfile.endswith("$py.class"):
+                modfile = modfile[:-9] + ".py"
+            if modfile.endswith(os.sep + "__init__.py"):
+                if self.basename != "__init__.py":
+                    modfile = modfile[:-12]
+            try:
+                issame = self.samefile(modfile)
+            except error.ENOENT:
+                issame = False
+            if not issame:
+                ignore = os.getenv("PY_IGNORE_IMPORTMISMATCH")
+                if ignore != "1":
+                    raise self.ImportMismatchError(modname, modfile, self)
+            return mod
+        else:
+            try:
+                return sys.modules[modname]
+            except KeyError:
+                # we have a custom modname, do a pseudo-import
+                import types
+
+                mod = types.ModuleType(modname)
+                mod.__file__ = str(self)
+                sys.modules[modname] = mod
+                try:
+                    with open(str(self), "rb") as f:
+                        exec(f.read(), mod.__dict__)
+                except BaseException:
+                    del sys.modules[modname]
+                    raise
+                return mod
+
+    def sysexec(self, *argv: os.PathLike[str], **popen_opts: Any) -> str:
+        """Return stdout text from executing a system child process,
+        where the 'self' path points to executable.
+        The process is directly invoked and not through a system shell.
+        """
+        from subprocess import PIPE
+        from subprocess import Popen
+
+        popen_opts.pop("stdout", None)
+        popen_opts.pop("stderr", None)
+        proc = Popen(
+            [str(self)] + [str(arg) for arg in argv],
+            **popen_opts,
+            stdout=PIPE,
+            stderr=PIPE,
+        )
+        stdout: str | bytes
+        stdout, stderr = proc.communicate()
+        ret = proc.wait()
+        if isinstance(stdout, bytes):
+            stdout = stdout.decode(sys.getdefaultencoding())
+        if ret != 0:
+            if isinstance(stderr, bytes):
+                stderr = stderr.decode(sys.getdefaultencoding())
+            raise RuntimeError(
+                ret,
+                ret,
+                str(self),
+                stdout,
+                stderr,
+            )
+        return stdout
+
+    @classmethod
+    def sysfind(cls, name, checker=None, paths=None):
+        """Return a path object found by looking at the systems
+        underlying PATH specification. If the checker is not None
+        it will be invoked to filter matching paths.  If a binary
+        cannot be found, None is returned
+        Note: This is probably not working on plain win32 systems
+        but may work on cygwin.
+        """
+        if isabs(name):
+            p = local(name)
+            if p.check(file=1):
+                return p
+        else:
+            if paths is None:
+                if iswin32:
+                    paths = os.environ["Path"].split(";")
+                    if "" not in paths and "." not in paths:
+                        paths.append(".")
+                    try:
+                        systemroot = os.environ["SYSTEMROOT"]
+                    except KeyError:
+                        pass
+                    else:
+                        paths = [
+                            path.replace("%SystemRoot%", systemroot) for path in paths
+                        ]
+                else:
+                    paths = os.environ["PATH"].split(":")
+            tryadd = []
+            if iswin32:
+                tryadd += os.environ["PATHEXT"].split(os.pathsep)
+            tryadd.append("")
+
+            for x in paths:
+                for addext in tryadd:
+                    p = local(x).join(name, abs=True) + addext
+                    try:
+                        if p.check(file=1):
+                            if checker:
+                                if not checker(p):
+                                    continue
+                            return p
+                    except error.EACCES:
+                        pass
+        return None
+
+    @classmethod
+    def _gethomedir(cls):
+        try:
+            x = os.environ["HOME"]
+        except KeyError:
+            try:
+                x = os.environ["HOMEDRIVE"] + os.environ["HOMEPATH"]
+            except KeyError:
+                return None
+        return cls(x)
+
+    # """
+    # special class constructors for local filesystem paths
+    # """
+    @classmethod
+    def get_temproot(cls):
+        """Return the system's temporary directory
+        (where tempfiles are usually created in)
+        """
+        import tempfile
+
+        return local(tempfile.gettempdir())
+
+    @classmethod
+    def mkdtemp(cls, rootdir=None):
+        """Return a Path object pointing to a fresh new temporary directory
+        (which we created ourselves).
+        """
+        import tempfile
+
+        if rootdir is None:
+            rootdir = cls.get_temproot()
+        path = error.checked_call(tempfile.mkdtemp, dir=str(rootdir))
+        return cls(path)
+
+    @classmethod
+    def make_numbered_dir(
+        cls, prefix="session-", rootdir=None, keep=3, lock_timeout=172800
+    ):  # two days
+        """Return unique directory with a number greater than the current
+        maximum one.  The number is assumed to start directly after prefix.
+        if keep is true directories with a number less than (maxnum-keep)
+        will be removed. If .lock files are used (lock_timeout non-zero),
+        algorithm is multi-process safe.
+        """
+        if rootdir is None:
+            rootdir = cls.get_temproot()
+
+        nprefix = prefix.lower()
+
+        def parse_num(path):
+            """Parse the number out of a path (if it matches the prefix)"""
+            nbasename = path.basename.lower()
+            if nbasename.startswith(nprefix):
+                try:
+                    return int(nbasename[len(nprefix) :])
+                except ValueError:
+                    pass
+
+        def create_lockfile(path):
+            """Exclusively create lockfile. Throws when failed"""
+            mypid = os.getpid()
+            lockfile = path.join(".lock")
+            if hasattr(lockfile, "mksymlinkto"):
+                lockfile.mksymlinkto(str(mypid))
+            else:
+                fd = error.checked_call(
+                    os.open, str(lockfile), os.O_WRONLY | os.O_CREAT | os.O_EXCL, 0o644
+                )
+                with os.fdopen(fd, "w") as f:
+                    f.write(str(mypid))
+            return lockfile
+
+        def atexit_remove_lockfile(lockfile):
+            """Ensure lockfile is removed at process exit"""
+            mypid = os.getpid()
+
+            def try_remove_lockfile():
+                # in a fork() situation, only the last process should
+                # remove the .lock, otherwise the other processes run the
+                # risk of seeing their temporary dir disappear.  For now
+                # we remove the .lock in the parent only (i.e. we assume
+                # that the children finish before the parent).
+                if os.getpid() != mypid:
+                    return
+                try:
+                    lockfile.remove()
+                except error.Error:
+                    pass
+
+            atexit.register(try_remove_lockfile)
+
+        # compute the maximum number currently in use with the prefix
+        lastmax = None
+        while True:
+            maxnum = -1
+            for path in rootdir.listdir():
+                num = parse_num(path)
+                if num is not None:
+                    maxnum = max(maxnum, num)
+
+            # make the new directory
+            try:
+                udir = rootdir.mkdir(prefix + str(maxnum + 1))
+                if lock_timeout:
+                    lockfile = create_lockfile(udir)
+                    atexit_remove_lockfile(lockfile)
+            except (error.EEXIST, error.ENOENT, error.EBUSY):
+                # race condition (1): another thread/process created the dir
+                #                     in the meantime - try again
+                # race condition (2): another thread/process spuriously acquired
+                #                     lock treating empty directory as candidate
+                #                     for removal - try again
+                # race condition (3): another thread/process tried to create the lock at
+                #                     the same time (happened in Python 3.3 on Windows)
+                # https://ci.appveyor.com/project/pytestbot/py/build/1.0.21/job/ffi85j4c0lqwsfwa
+                if lastmax == maxnum:
+                    raise
+                lastmax = maxnum
+                continue
+            break
+
+        def get_mtime(path):
+            """Read file modification time"""
+            try:
+                return path.lstat().mtime
+            except error.Error:
+                pass
+
+        garbage_prefix = prefix + "garbage-"
+
+        def is_garbage(path):
+            """Check if path denotes directory scheduled for removal"""
+            bn = path.basename
+            return bn.startswith(garbage_prefix)
+
+        # prune old directories
+        udir_time = get_mtime(udir)
+        if keep and udir_time:
+            for path in rootdir.listdir():
+                num = parse_num(path)
+                if num is not None and num <= (maxnum - keep):
+                    try:
+                        # try acquiring lock to remove directory as exclusive user
+                        if lock_timeout:
+                            create_lockfile(path)
+                    except (error.EEXIST, error.ENOENT, error.EBUSY):
+                        path_time = get_mtime(path)
+                        if not path_time:
+                            # assume directory doesn't exist now
+                            continue
+                        if abs(udir_time - path_time) < lock_timeout:
+                            # assume directory with lockfile exists
+                            # and lock timeout hasn't expired yet
+                            continue
+
+                    # path dir locked for exclusive use
+                    # and scheduled for removal to avoid another thread/process
+                    # treating it as a new directory or removal candidate
+                    garbage_path = rootdir.join(garbage_prefix + str(uuid.uuid4()))
+                    try:
+                        path.rename(garbage_path)
+                        garbage_path.remove(rec=1)
+                    except KeyboardInterrupt:
+                        raise
+                    except Exception:  # this might be error.Error, WindowsError ...
+                        pass
+                if is_garbage(path):
+                    try:
+                        path.remove(rec=1)
+                    except KeyboardInterrupt:
+                        raise
+                    except Exception:  # this might be error.Error, WindowsError ...
+                        pass
+
+        # make link...
+        try:
+            username = os.environ["USER"]  # linux, et al
+        except KeyError:
+            try:
+                username = os.environ["USERNAME"]  # windows
+            except KeyError:
+                username = "current"
+
+        src = str(udir)
+        dest = src[: src.rfind("-")] + "-" + username
+        try:
+            os.unlink(dest)
+        except OSError:
+            pass
+        try:
+            os.symlink(src, dest)
+        except (OSError, AttributeError, NotImplementedError):
+            pass
+
+        return udir
+
+
+def copymode(src, dest):
+    """Copy permission from src to dst."""
+    import shutil
+
+    shutil.copymode(src, dest)
+
+
+def copystat(src, dest):
+    """Copy permission,  last modification time,
+    last access time, and flags from src to dst."""
+    import shutil
+
+    shutil.copystat(str(src), str(dest))
+
+
+def copychunked(src, dest):
+    chunksize = 524288  # half a meg of bytes
+    fsrc = src.open("rb")
+    try:
+        fdest = dest.open("wb")
+        try:
+            while 1:
+                buf = fsrc.read(chunksize)
+                if not buf:
+                    break
+                fdest.write(buf)
+        finally:
+            fdest.close()
+    finally:
+        fsrc.close()
+
+
+def isimportable(name):
+    if name and (name[0].isalpha() or name[0] == "_"):
+        name = name.replace("_", "")
+        return not name or name.isalnum()
+
+
+local = LocalPath
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/_version.py b/backend/venv/lib/python3.12/site-packages/_pytest/_version.py
new file mode 100644
index 00000000..e88f96c7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/_version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '8.3.4'
+__version_tuple__ = version_tuple = (8, 3, 4)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__init__.py
new file mode 100644
index 00000000..f2f1d029
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__init__.py
@@ -0,0 +1,192 @@
+# mypy: allow-untyped-defs
+"""Support for presenting detailed information in failing assertions."""
+
+from __future__ import annotations
+
+import sys
+from typing import Any
+from typing import Generator
+from typing import TYPE_CHECKING
+
+from _pytest.assertion import rewrite
+from _pytest.assertion import truncate
+from _pytest.assertion import util
+from _pytest.assertion.rewrite import assertstate_key
+from _pytest.config import Config
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.nodes import Item
+
+
+if TYPE_CHECKING:
+    from _pytest.main import Session
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("debugconfig")
+    group.addoption(
+        "--assert",
+        action="store",
+        dest="assertmode",
+        choices=("rewrite", "plain"),
+        default="rewrite",
+        metavar="MODE",
+        help=(
+            "Control assertion debugging tools.\n"
+            "'plain' performs no assertion debugging.\n"
+            "'rewrite' (the default) rewrites assert statements in test modules"
+            " on import to provide assert expression information."
+        ),
+    )
+    parser.addini(
+        "enable_assertion_pass_hook",
+        type="bool",
+        default=False,
+        help="Enables the pytest_assertion_pass hook. "
+        "Make sure to delete any previously generated pyc cache files.",
+    )
+    Config._add_verbosity_ini(
+        parser,
+        Config.VERBOSITY_ASSERTIONS,
+        help=(
+            "Specify a verbosity level for assertions, overriding the main level. "
+            "Higher levels will provide more detailed explanation when an assertion fails."
+        ),
+    )
+
+
+def register_assert_rewrite(*names: str) -> None:
+    """Register one or more module names to be rewritten on import.
+
+    This function will make sure that this module or all modules inside
+    the package will get their assert statements rewritten.
+    Thus you should make sure to call this before the module is
+    actually imported, usually in your __init__.py if you are a plugin
+    using a package.
+
+    :param names: The module names to register.
+    """
+    for name in names:
+        if not isinstance(name, str):
+            msg = "expected module names as *args, got {0} instead"  # type: ignore[unreachable]
+            raise TypeError(msg.format(repr(names)))
+    for hook in sys.meta_path:
+        if isinstance(hook, rewrite.AssertionRewritingHook):
+            importhook = hook
+            break
+    else:
+        # TODO(typing): Add a protocol for mark_rewrite() and use it
+        # for importhook and for PytestPluginManager.rewrite_hook.
+        importhook = DummyRewriteHook()  # type: ignore
+    importhook.mark_rewrite(*names)
+
+
+class DummyRewriteHook:
+    """A no-op import hook for when rewriting is disabled."""
+
+    def mark_rewrite(self, *names: str) -> None:
+        pass
+
+
+class AssertionState:
+    """State for the assertion plugin."""
+
+    def __init__(self, config: Config, mode) -> None:
+        self.mode = mode
+        self.trace = config.trace.root.get("assertion")
+        self.hook: rewrite.AssertionRewritingHook | None = None
+
+
+def install_importhook(config: Config) -> rewrite.AssertionRewritingHook:
+    """Try to install the rewrite hook, raise SystemError if it fails."""
+    config.stash[assertstate_key] = AssertionState(config, "rewrite")
+    config.stash[assertstate_key].hook = hook = rewrite.AssertionRewritingHook(config)
+    sys.meta_path.insert(0, hook)
+    config.stash[assertstate_key].trace("installed rewrite import hook")
+
+    def undo() -> None:
+        hook = config.stash[assertstate_key].hook
+        if hook is not None and hook in sys.meta_path:
+            sys.meta_path.remove(hook)
+
+    config.add_cleanup(undo)
+    return hook
+
+
+def pytest_collection(session: Session) -> None:
+    # This hook is only called when test modules are collected
+    # so for example not in the managing process of pytest-xdist
+    # (which does not collect test modules).
+    assertstate = session.config.stash.get(assertstate_key, None)
+    if assertstate:
+        if assertstate.hook is not None:
+            assertstate.hook.set_session(session)
+
+
+@hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_protocol(item: Item) -> Generator[None, object, object]:
+    """Setup the pytest_assertrepr_compare and pytest_assertion_pass hooks.
+
+    The rewrite module will use util._reprcompare if it exists to use custom
+    reporting via the pytest_assertrepr_compare hook.  This sets up this custom
+    comparison for the test.
+    """
+    ihook = item.ihook
+
+    def callbinrepr(op, left: object, right: object) -> str | None:
+        """Call the pytest_assertrepr_compare hook and prepare the result.
+
+        This uses the first result from the hook and then ensures the
+        following:
+        * Overly verbose explanations are truncated unless configured otherwise
+          (eg. if running in verbose mode).
+        * Embedded newlines are escaped to help util.format_explanation()
+          later.
+        * If the rewrite mode is used embedded %-characters are replaced
+          to protect later % formatting.
+
+        The result can be formatted by util.format_explanation() for
+        pretty printing.
+        """
+        hook_result = ihook.pytest_assertrepr_compare(
+            config=item.config, op=op, left=left, right=right
+        )
+        for new_expl in hook_result:
+            if new_expl:
+                new_expl = truncate.truncate_if_required(new_expl, item)
+                new_expl = [line.replace("\n", "\\n") for line in new_expl]
+                res = "\n~".join(new_expl)
+                if item.config.getvalue("assertmode") == "rewrite":
+                    res = res.replace("%", "%%")
+                return res
+        return None
+
+    saved_assert_hooks = util._reprcompare, util._assertion_pass
+    util._reprcompare = callbinrepr
+    util._config = item.config
+
+    if ihook.pytest_assertion_pass.get_hookimpls():
+
+        def call_assertion_pass_hook(lineno: int, orig: str, expl: str) -> None:
+            ihook.pytest_assertion_pass(item=item, lineno=lineno, orig=orig, expl=expl)
+
+        util._assertion_pass = call_assertion_pass_hook
+
+    try:
+        return (yield)
+    finally:
+        util._reprcompare, util._assertion_pass = saved_assert_hooks
+        util._config = None
+
+
+def pytest_sessionfinish(session: Session) -> None:
+    assertstate = session.config.stash.get(assertstate_key, None)
+    if assertstate:
+        if assertstate.hook is not None:
+            assertstate.hook.set_session(None)
+
+
+def pytest_assertrepr_compare(
+    config: Config, op: str, left: Any, right: Any
+) -> list[str] | None:
+    return util.assertrepr_compare(config=config, op=op, left=left, right=right)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9565593d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/rewrite.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/rewrite.cpython-312.pyc
new file mode 100644
index 00000000..62afbad6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/rewrite.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/truncate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/truncate.cpython-312.pyc
new file mode 100644
index 00000000..9c51ad15
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/truncate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..9ddc5113
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/rewrite.py b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/rewrite.py
new file mode 100644
index 00000000..37c09b03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/rewrite.py
@@ -0,0 +1,1211 @@
+"""Rewrite assertion AST to produce nice error messages."""
+
+from __future__ import annotations
+
+import ast
+from collections import defaultdict
+import errno
+import functools
+import importlib.abc
+import importlib.machinery
+import importlib.util
+import io
+import itertools
+import marshal
+import os
+from pathlib import Path
+from pathlib import PurePath
+import struct
+import sys
+import tokenize
+import types
+from typing import Callable
+from typing import IO
+from typing import Iterable
+from typing import Iterator
+from typing import Sequence
+from typing import TYPE_CHECKING
+
+from _pytest._io.saferepr import DEFAULT_REPR_MAX_SIZE
+from _pytest._io.saferepr import saferepr
+from _pytest._version import version
+from _pytest.assertion import util
+from _pytest.config import Config
+from _pytest.main import Session
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import fnmatch_ex
+from _pytest.stash import StashKey
+
+
+# fmt: off
+from _pytest.assertion.util import format_explanation as _format_explanation  # noqa:F401, isort:skip
+# fmt:on
+
+if TYPE_CHECKING:
+    from _pytest.assertion import AssertionState
+
+
+class Sentinel:
+    pass
+
+
+assertstate_key = StashKey["AssertionState"]()
+
+# pytest caches rewritten pycs in pycache dirs
+PYTEST_TAG = f"{sys.implementation.cache_tag}-pytest-{version}"
+PYC_EXT = ".py" + (__debug__ and "c" or "o")
+PYC_TAIL = "." + PYTEST_TAG + PYC_EXT
+
+# Special marker that denotes we have just left a scope definition
+_SCOPE_END_MARKER = Sentinel()
+
+
+class AssertionRewritingHook(importlib.abc.MetaPathFinder, importlib.abc.Loader):
+    """PEP302/PEP451 import hook which rewrites asserts."""
+
+    def __init__(self, config: Config) -> None:
+        self.config = config
+        try:
+            self.fnpats = config.getini("python_files")
+        except ValueError:
+            self.fnpats = ["test_*.py", "*_test.py"]
+        self.session: Session | None = None
+        self._rewritten_names: dict[str, Path] = {}
+        self._must_rewrite: set[str] = set()
+        # flag to guard against trying to rewrite a pyc file while we are already writing another pyc file,
+        # which might result in infinite recursion (#3506)
+        self._writing_pyc = False
+        self._basenames_to_check_rewrite = {"conftest"}
+        self._marked_for_rewrite_cache: dict[str, bool] = {}
+        self._session_paths_checked = False
+
+    def set_session(self, session: Session | None) -> None:
+        self.session = session
+        self._session_paths_checked = False
+
+    # Indirection so we can mock calls to find_spec originated from the hook during testing
+    _find_spec = importlib.machinery.PathFinder.find_spec
+
+    def find_spec(
+        self,
+        name: str,
+        path: Sequence[str | bytes] | None = None,
+        target: types.ModuleType | None = None,
+    ) -> importlib.machinery.ModuleSpec | None:
+        if self._writing_pyc:
+            return None
+        state = self.config.stash[assertstate_key]
+        if self._early_rewrite_bailout(name, state):
+            return None
+        state.trace(f"find_module called for: {name}")
+
+        # Type ignored because mypy is confused about the `self` binding here.
+        spec = self._find_spec(name, path)  # type: ignore
+
+        if spec is None and path is not None:
+            # With --import-mode=importlib, PathFinder cannot find spec without modifying `sys.path`,
+            # causing inability to assert rewriting (#12659).
+            # At this point, try using the file path to find the module spec.
+            for _path_str in path:
+                spec = importlib.util.spec_from_file_location(name, _path_str)
+                if spec is not None:
+                    break
+
+        if (
+            # the import machinery could not find a file to import
+            spec is None
+            # this is a namespace package (without `__init__.py`)
+            # there's nothing to rewrite there
+            or spec.origin is None
+            # we can only rewrite source files
+            or not isinstance(spec.loader, importlib.machinery.SourceFileLoader)
+            # if the file doesn't exist, we can't rewrite it
+            or not os.path.exists(spec.origin)
+        ):
+            return None
+        else:
+            fn = spec.origin
+
+        if not self._should_rewrite(name, fn, state):
+            return None
+
+        return importlib.util.spec_from_file_location(
+            name,
+            fn,
+            loader=self,
+            submodule_search_locations=spec.submodule_search_locations,
+        )
+
+    def create_module(
+        self, spec: importlib.machinery.ModuleSpec
+    ) -> types.ModuleType | None:
+        return None  # default behaviour is fine
+
+    def exec_module(self, module: types.ModuleType) -> None:
+        assert module.__spec__ is not None
+        assert module.__spec__.origin is not None
+        fn = Path(module.__spec__.origin)
+        state = self.config.stash[assertstate_key]
+
+        self._rewritten_names[module.__name__] = fn
+
+        # The requested module looks like a test file, so rewrite it. This is
+        # the most magical part of the process: load the source, rewrite the
+        # asserts, and load the rewritten source. We also cache the rewritten
+        # module code in a special pyc. We must be aware of the possibility of
+        # concurrent pytest processes rewriting and loading pycs. To avoid
+        # tricky race conditions, we maintain the following invariant: The
+        # cached pyc is always a complete, valid pyc. Operations on it must be
+        # atomic. POSIX's atomic rename comes in handy.
+        write = not sys.dont_write_bytecode
+        cache_dir = get_cache_dir(fn)
+        if write:
+            ok = try_makedirs(cache_dir)
+            if not ok:
+                write = False
+                state.trace(f"read only directory: {cache_dir}")
+
+        cache_name = fn.name[:-3] + PYC_TAIL
+        pyc = cache_dir / cache_name
+        # Notice that even if we're in a read-only directory, I'm going
+        # to check for a cached pyc. This may not be optimal...
+        co = _read_pyc(fn, pyc, state.trace)
+        if co is None:
+            state.trace(f"rewriting {fn!r}")
+            source_stat, co = _rewrite_test(fn, self.config)
+            if write:
+                self._writing_pyc = True
+                try:
+                    _write_pyc(state, co, source_stat, pyc)
+                finally:
+                    self._writing_pyc = False
+        else:
+            state.trace(f"found cached rewritten pyc for {fn}")
+        exec(co, module.__dict__)
+
+    def _early_rewrite_bailout(self, name: str, state: AssertionState) -> bool:
+        """A fast way to get out of rewriting modules.
+
+        Profiling has shown that the call to PathFinder.find_spec (inside of
+        the find_spec from this class) is a major slowdown, so, this method
+        tries to filter what we're sure won't be rewritten before getting to
+        it.
+        """
+        if self.session is not None and not self._session_paths_checked:
+            self._session_paths_checked = True
+            for initial_path in self.session._initialpaths:
+                # Make something as c:/projects/my_project/path.py ->
+                #     ['c:', 'projects', 'my_project', 'path.py']
+                parts = str(initial_path).split(os.sep)
+                # add 'path' to basenames to be checked.
+                self._basenames_to_check_rewrite.add(os.path.splitext(parts[-1])[0])
+
+        # Note: conftest already by default in _basenames_to_check_rewrite.
+        parts = name.split(".")
+        if parts[-1] in self._basenames_to_check_rewrite:
+            return False
+
+        # For matching the name it must be as if it was a filename.
+        path = PurePath(*parts).with_suffix(".py")
+
+        for pat in self.fnpats:
+            # if the pattern contains subdirectories ("tests/**.py" for example) we can't bail out based
+            # on the name alone because we need to match against the full path
+            if os.path.dirname(pat):
+                return False
+            if fnmatch_ex(pat, path):
+                return False
+
+        if self._is_marked_for_rewrite(name, state):
+            return False
+
+        state.trace(f"early skip of rewriting module: {name}")
+        return True
+
+    def _should_rewrite(self, name: str, fn: str, state: AssertionState) -> bool:
+        # always rewrite conftest files
+        if os.path.basename(fn) == "conftest.py":
+            state.trace(f"rewriting conftest file: {fn!r}")
+            return True
+
+        if self.session is not None:
+            if self.session.isinitpath(absolutepath(fn)):
+                state.trace(f"matched test file (was specified on cmdline): {fn!r}")
+                return True
+
+        # modules not passed explicitly on the command line are only
+        # rewritten if they match the naming convention for test files
+        fn_path = PurePath(fn)
+        for pat in self.fnpats:
+            if fnmatch_ex(pat, fn_path):
+                state.trace(f"matched test file {fn!r}")
+                return True
+
+        return self._is_marked_for_rewrite(name, state)
+
+    def _is_marked_for_rewrite(self, name: str, state: AssertionState) -> bool:
+        try:
+            return self._marked_for_rewrite_cache[name]
+        except KeyError:
+            for marked in self._must_rewrite:
+                if name == marked or name.startswith(marked + "."):
+                    state.trace(f"matched marked file {name!r} (from {marked!r})")
+                    self._marked_for_rewrite_cache[name] = True
+                    return True
+
+            self._marked_for_rewrite_cache[name] = False
+            return False
+
+    def mark_rewrite(self, *names: str) -> None:
+        """Mark import names as needing to be rewritten.
+
+        The named module or package as well as any nested modules will
+        be rewritten on import.
+        """
+        already_imported = (
+            set(names).intersection(sys.modules).difference(self._rewritten_names)
+        )
+        for name in already_imported:
+            mod = sys.modules[name]
+            if not AssertionRewriter.is_rewrite_disabled(
+                mod.__doc__ or ""
+            ) and not isinstance(mod.__loader__, type(self)):
+                self._warn_already_imported(name)
+        self._must_rewrite.update(names)
+        self._marked_for_rewrite_cache.clear()
+
+    def _warn_already_imported(self, name: str) -> None:
+        from _pytest.warning_types import PytestAssertRewriteWarning
+
+        self.config.issue_config_time_warning(
+            PytestAssertRewriteWarning(
+                f"Module already imported so cannot be rewritten: {name}"
+            ),
+            stacklevel=5,
+        )
+
+    def get_data(self, pathname: str | bytes) -> bytes:
+        """Optional PEP302 get_data API."""
+        with open(pathname, "rb") as f:
+            return f.read()
+
+    if sys.version_info >= (3, 10):
+        if sys.version_info >= (3, 12):
+            from importlib.resources.abc import TraversableResources
+        else:
+            from importlib.abc import TraversableResources
+
+        def get_resource_reader(self, name: str) -> TraversableResources:
+            if sys.version_info < (3, 11):
+                from importlib.readers import FileReader
+            else:
+                from importlib.resources.readers import FileReader
+
+            return FileReader(types.SimpleNamespace(path=self._rewritten_names[name]))
+
+
+def _write_pyc_fp(
+    fp: IO[bytes], source_stat: os.stat_result, co: types.CodeType
+) -> None:
+    # Technically, we don't have to have the same pyc format as
+    # (C)Python, since these "pycs" should never be seen by builtin
+    # import. However, there's little reason to deviate.
+    fp.write(importlib.util.MAGIC_NUMBER)
+    # https://www.python.org/dev/peps/pep-0552/
+    flags = b"\x00\x00\x00\x00"
+    fp.write(flags)
+    # as of now, bytecode header expects 32-bit numbers for size and mtime (#4903)
+    mtime = int(source_stat.st_mtime) & 0xFFFFFFFF
+    size = source_stat.st_size & 0xFFFFFFFF
+    # "<LL" stands for 2 unsigned longs, little-endian.
+    fp.write(struct.pack("<LL", mtime, size))
+    fp.write(marshal.dumps(co))
+
+
+def _write_pyc(
+    state: AssertionState,
+    co: types.CodeType,
+    source_stat: os.stat_result,
+    pyc: Path,
+) -> bool:
+    proc_pyc = f"{pyc}.{os.getpid()}"
+    try:
+        with open(proc_pyc, "wb") as fp:
+            _write_pyc_fp(fp, source_stat, co)
+    except OSError as e:
+        state.trace(f"error writing pyc file at {proc_pyc}: errno={e.errno}")
+        return False
+
+    try:
+        os.replace(proc_pyc, pyc)
+    except OSError as e:
+        state.trace(f"error writing pyc file at {pyc}: {e}")
+        # we ignore any failure to write the cache file
+        # there are many reasons, permission-denied, pycache dir being a
+        # file etc.
+        return False
+    return True
+
+
+def _rewrite_test(fn: Path, config: Config) -> tuple[os.stat_result, types.CodeType]:
+    """Read and rewrite *fn* and return the code object."""
+    stat = os.stat(fn)
+    source = fn.read_bytes()
+    strfn = str(fn)
+    tree = ast.parse(source, filename=strfn)
+    rewrite_asserts(tree, source, strfn, config)
+    co = compile(tree, strfn, "exec", dont_inherit=True)
+    return stat, co
+
+
+def _read_pyc(
+    source: Path, pyc: Path, trace: Callable[[str], None] = lambda x: None
+) -> types.CodeType | None:
+    """Possibly read a pytest pyc containing rewritten code.
+
+    Return rewritten code if successful or None if not.
+    """
+    try:
+        fp = open(pyc, "rb")
+    except OSError:
+        return None
+    with fp:
+        try:
+            stat_result = os.stat(source)
+            mtime = int(stat_result.st_mtime)
+            size = stat_result.st_size
+            data = fp.read(16)
+        except OSError as e:
+            trace(f"_read_pyc({source}): OSError {e}")
+            return None
+        # Check for invalid or out of date pyc file.
+        if len(data) != (16):
+            trace(f"_read_pyc({source}): invalid pyc (too short)")
+            return None
+        if data[:4] != importlib.util.MAGIC_NUMBER:
+            trace(f"_read_pyc({source}): invalid pyc (bad magic number)")
+            return None
+        if data[4:8] != b"\x00\x00\x00\x00":
+            trace(f"_read_pyc({source}): invalid pyc (unsupported flags)")
+            return None
+        mtime_data = data[8:12]
+        if int.from_bytes(mtime_data, "little") != mtime & 0xFFFFFFFF:
+            trace(f"_read_pyc({source}): out of date")
+            return None
+        size_data = data[12:16]
+        if int.from_bytes(size_data, "little") != size & 0xFFFFFFFF:
+            trace(f"_read_pyc({source}): invalid pyc (incorrect size)")
+            return None
+        try:
+            co = marshal.load(fp)
+        except Exception as e:
+            trace(f"_read_pyc({source}): marshal.load error {e}")
+            return None
+        if not isinstance(co, types.CodeType):
+            trace(f"_read_pyc({source}): not a code object")
+            return None
+        return co
+
+
+def rewrite_asserts(
+    mod: ast.Module,
+    source: bytes,
+    module_path: str | None = None,
+    config: Config | None = None,
+) -> None:
+    """Rewrite the assert statements in mod."""
+    AssertionRewriter(module_path, config, source).run(mod)
+
+
+def _saferepr(obj: object) -> str:
+    r"""Get a safe repr of an object for assertion error messages.
+
+    The assertion formatting (util.format_explanation()) requires
+    newlines to be escaped since they are a special character for it.
+    Normally assertion.util.format_explanation() does this but for a
+    custom repr it is possible to contain one of the special escape
+    sequences, especially '\n{' and '\n}' are likely to be present in
+    JSON reprs.
+    """
+    if isinstance(obj, types.MethodType):
+        # for bound methods, skip redundant <bound method ...> information
+        return obj.__name__
+
+    maxsize = _get_maxsize_for_saferepr(util._config)
+    return saferepr(obj, maxsize=maxsize).replace("\n", "\\n")
+
+
+def _get_maxsize_for_saferepr(config: Config | None) -> int | None:
+    """Get `maxsize` configuration for saferepr based on the given config object."""
+    if config is None:
+        verbosity = 0
+    else:
+        verbosity = config.get_verbosity(Config.VERBOSITY_ASSERTIONS)
+    if verbosity >= 2:
+        return None
+    if verbosity >= 1:
+        return DEFAULT_REPR_MAX_SIZE * 10
+    return DEFAULT_REPR_MAX_SIZE
+
+
+def _format_assertmsg(obj: object) -> str:
+    r"""Format the custom assertion message given.
+
+    For strings this simply replaces newlines with '\n~' so that
+    util.format_explanation() will preserve them instead of escaping
+    newlines.  For other objects saferepr() is used first.
+    """
+    # reprlib appears to have a bug which means that if a string
+    # contains a newline it gets escaped, however if an object has a
+    # .__repr__() which contains newlines it does not get escaped.
+    # However in either case we want to preserve the newline.
+    replaces = [("\n", "\n~"), ("%", "%%")]
+    if not isinstance(obj, str):
+        obj = saferepr(obj, _get_maxsize_for_saferepr(util._config))
+        replaces.append(("\\n", "\n~"))
+
+    for r1, r2 in replaces:
+        obj = obj.replace(r1, r2)
+
+    return obj
+
+
+def _should_repr_global_name(obj: object) -> bool:
+    if callable(obj):
+        return False
+
+    try:
+        return not hasattr(obj, "__name__")
+    except Exception:
+        return True
+
+
+def _format_boolop(explanations: Iterable[str], is_or: bool) -> str:
+    explanation = "(" + (is_or and " or " or " and ").join(explanations) + ")"
+    return explanation.replace("%", "%%")
+
+
+def _call_reprcompare(
+    ops: Sequence[str],
+    results: Sequence[bool],
+    expls: Sequence[str],
+    each_obj: Sequence[object],
+) -> str:
+    for i, res, expl in zip(range(len(ops)), results, expls):
+        try:
+            done = not res
+        except Exception:
+            done = True
+        if done:
+            break
+    if util._reprcompare is not None:
+        custom = util._reprcompare(ops[i], each_obj[i], each_obj[i + 1])
+        if custom is not None:
+            return custom
+    return expl
+
+
+def _call_assertion_pass(lineno: int, orig: str, expl: str) -> None:
+    if util._assertion_pass is not None:
+        util._assertion_pass(lineno, orig, expl)
+
+
+def _check_if_assertion_pass_impl() -> bool:
+    """Check if any plugins implement the pytest_assertion_pass hook
+    in order not to generate explanation unnecessarily (might be expensive)."""
+    return True if util._assertion_pass else False
+
+
+UNARY_MAP = {ast.Not: "not %s", ast.Invert: "~%s", ast.USub: "-%s", ast.UAdd: "+%s"}
+
+BINOP_MAP = {
+    ast.BitOr: "|",
+    ast.BitXor: "^",
+    ast.BitAnd: "&",
+    ast.LShift: "<<",
+    ast.RShift: ">>",
+    ast.Add: "+",
+    ast.Sub: "-",
+    ast.Mult: "*",
+    ast.Div: "/",
+    ast.FloorDiv: "//",
+    ast.Mod: "%%",  # escaped for string formatting
+    ast.Eq: "==",
+    ast.NotEq: "!=",
+    ast.Lt: "<",
+    ast.LtE: "<=",
+    ast.Gt: ">",
+    ast.GtE: ">=",
+    ast.Pow: "**",
+    ast.Is: "is",
+    ast.IsNot: "is not",
+    ast.In: "in",
+    ast.NotIn: "not in",
+    ast.MatMult: "@",
+}
+
+
+def traverse_node(node: ast.AST) -> Iterator[ast.AST]:
+    """Recursively yield node and all its children in depth-first order."""
+    yield node
+    for child in ast.iter_child_nodes(node):
+        yield from traverse_node(child)
+
+
+@functools.lru_cache(maxsize=1)
+def _get_assertion_exprs(src: bytes) -> dict[int, str]:
+    """Return a mapping from {lineno: "assertion test expression"}."""
+    ret: dict[int, str] = {}
+
+    depth = 0
+    lines: list[str] = []
+    assert_lineno: int | None = None
+    seen_lines: set[int] = set()
+
+    def _write_and_reset() -> None:
+        nonlocal depth, lines, assert_lineno, seen_lines
+        assert assert_lineno is not None
+        ret[assert_lineno] = "".join(lines).rstrip().rstrip("\\")
+        depth = 0
+        lines = []
+        assert_lineno = None
+        seen_lines = set()
+
+    tokens = tokenize.tokenize(io.BytesIO(src).readline)
+    for tp, source, (lineno, offset), _, line in tokens:
+        if tp == tokenize.NAME and source == "assert":
+            assert_lineno = lineno
+        elif assert_lineno is not None:
+            # keep track of depth for the assert-message `,` lookup
+            if tp == tokenize.OP and source in "([{":
+                depth += 1
+            elif tp == tokenize.OP and source in ")]}":
+                depth -= 1
+
+            if not lines:
+                lines.append(line[offset:])
+                seen_lines.add(lineno)
+            # a non-nested comma separates the expression from the message
+            elif depth == 0 and tp == tokenize.OP and source == ",":
+                # one line assert with message
+                if lineno in seen_lines and len(lines) == 1:
+                    offset_in_trimmed = offset + len(lines[-1]) - len(line)
+                    lines[-1] = lines[-1][:offset_in_trimmed]
+                # multi-line assert with message
+                elif lineno in seen_lines:
+                    lines[-1] = lines[-1][:offset]
+                # multi line assert with escaped newline before message
+                else:
+                    lines.append(line[:offset])
+                _write_and_reset()
+            elif tp in {tokenize.NEWLINE, tokenize.ENDMARKER}:
+                _write_and_reset()
+            elif lines and lineno not in seen_lines:
+                lines.append(line)
+                seen_lines.add(lineno)
+
+    return ret
+
+
+class AssertionRewriter(ast.NodeVisitor):
+    """Assertion rewriting implementation.
+
+    The main entrypoint is to call .run() with an ast.Module instance,
+    this will then find all the assert statements and rewrite them to
+    provide intermediate values and a detailed assertion error.  See
+    http://pybites.blogspot.be/2011/07/behind-scenes-of-pytests-new-assertion.html
+    for an overview of how this works.
+
+    The entry point here is .run() which will iterate over all the
+    statements in an ast.Module and for each ast.Assert statement it
+    finds call .visit() with it.  Then .visit_Assert() takes over and
+    is responsible for creating new ast statements to replace the
+    original assert statement: it rewrites the test of an assertion
+    to provide intermediate values and replace it with an if statement
+    which raises an assertion error with a detailed explanation in
+    case the expression is false and calls pytest_assertion_pass hook
+    if expression is true.
+
+    For this .visit_Assert() uses the visitor pattern to visit all the
+    AST nodes of the ast.Assert.test field, each visit call returning
+    an AST node and the corresponding explanation string.  During this
+    state is kept in several instance attributes:
+
+    :statements: All the AST statements which will replace the assert
+       statement.
+
+    :variables: This is populated by .variable() with each variable
+       used by the statements so that they can all be set to None at
+       the end of the statements.
+
+    :variable_counter: Counter to create new unique variables needed
+       by statements.  Variables are created using .variable() and
+       have the form of "@py_assert0".
+
+    :expl_stmts: The AST statements which will be executed to get
+       data from the assertion.  This is the code which will construct
+       the detailed assertion message that is used in the AssertionError
+       or for the pytest_assertion_pass hook.
+
+    :explanation_specifiers: A dict filled by .explanation_param()
+       with %-formatting placeholders and their corresponding
+       expressions to use in the building of an assertion message.
+       This is used by .pop_format_context() to build a message.
+
+    :stack: A stack of the explanation_specifiers dicts maintained by
+       .push_format_context() and .pop_format_context() which allows
+       to build another %-formatted string while already building one.
+
+    :scope: A tuple containing the current scope used for variables_overwrite.
+
+    :variables_overwrite: A dict filled with references to variables
+       that change value within an assert. This happens when a variable is
+       reassigned with the walrus operator
+
+    This state, except the variables_overwrite,  is reset on every new assert
+    statement visited and used by the other visitors.
+    """
+
+    def __init__(
+        self, module_path: str | None, config: Config | None, source: bytes
+    ) -> None:
+        super().__init__()
+        self.module_path = module_path
+        self.config = config
+        if config is not None:
+            self.enable_assertion_pass_hook = config.getini(
+                "enable_assertion_pass_hook"
+            )
+        else:
+            self.enable_assertion_pass_hook = False
+        self.source = source
+        self.scope: tuple[ast.AST, ...] = ()
+        self.variables_overwrite: defaultdict[tuple[ast.AST, ...], dict[str, str]] = (
+            defaultdict(dict)
+        )
+
+    def run(self, mod: ast.Module) -> None:
+        """Find all assert statements in *mod* and rewrite them."""
+        if not mod.body:
+            # Nothing to do.
+            return
+
+        # We'll insert some special imports at the top of the module, but after any
+        # docstrings and __future__ imports, so first figure out where that is.
+        doc = getattr(mod, "docstring", None)
+        expect_docstring = doc is None
+        if doc is not None and self.is_rewrite_disabled(doc):
+            return
+        pos = 0
+        item = None
+        for item in mod.body:
+            if (
+                expect_docstring
+                and isinstance(item, ast.Expr)
+                and isinstance(item.value, ast.Constant)
+                and isinstance(item.value.value, str)
+            ):
+                doc = item.value.value
+                if self.is_rewrite_disabled(doc):
+                    return
+                expect_docstring = False
+            elif (
+                isinstance(item, ast.ImportFrom)
+                and item.level == 0
+                and item.module == "__future__"
+            ):
+                pass
+            else:
+                break
+            pos += 1
+        # Special case: for a decorated function, set the lineno to that of the
+        # first decorator, not the `def`. Issue #4984.
+        if isinstance(item, ast.FunctionDef) and item.decorator_list:
+            lineno = item.decorator_list[0].lineno
+        else:
+            lineno = item.lineno
+        # Now actually insert the special imports.
+        if sys.version_info >= (3, 10):
+            aliases = [
+                ast.alias("builtins", "@py_builtins", lineno=lineno, col_offset=0),
+                ast.alias(
+                    "_pytest.assertion.rewrite",
+                    "@pytest_ar",
+                    lineno=lineno,
+                    col_offset=0,
+                ),
+            ]
+        else:
+            aliases = [
+                ast.alias("builtins", "@py_builtins"),
+                ast.alias("_pytest.assertion.rewrite", "@pytest_ar"),
+            ]
+        imports = [
+            ast.Import([alias], lineno=lineno, col_offset=0) for alias in aliases
+        ]
+        mod.body[pos:pos] = imports
+
+        # Collect asserts.
+        self.scope = (mod,)
+        nodes: list[ast.AST | Sentinel] = [mod]
+        while nodes:
+            node = nodes.pop()
+            if isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef, ast.ClassDef)):
+                self.scope = tuple((*self.scope, node))
+                nodes.append(_SCOPE_END_MARKER)
+            if node == _SCOPE_END_MARKER:
+                self.scope = self.scope[:-1]
+                continue
+            assert isinstance(node, ast.AST)
+            for name, field in ast.iter_fields(node):
+                if isinstance(field, list):
+                    new: list[ast.AST] = []
+                    for i, child in enumerate(field):
+                        if isinstance(child, ast.Assert):
+                            # Transform assert.
+                            new.extend(self.visit(child))
+                        else:
+                            new.append(child)
+                            if isinstance(child, ast.AST):
+                                nodes.append(child)
+                    setattr(node, name, new)
+                elif (
+                    isinstance(field, ast.AST)
+                    # Don't recurse into expressions as they can't contain
+                    # asserts.
+                    and not isinstance(field, ast.expr)
+                ):
+                    nodes.append(field)
+
+    @staticmethod
+    def is_rewrite_disabled(docstring: str) -> bool:
+        return "PYTEST_DONT_REWRITE" in docstring
+
+    def variable(self) -> str:
+        """Get a new variable."""
+        # Use a character invalid in python identifiers to avoid clashing.
+        name = "@py_assert" + str(next(self.variable_counter))
+        self.variables.append(name)
+        return name
+
+    def assign(self, expr: ast.expr) -> ast.Name:
+        """Give *expr* a name."""
+        name = self.variable()
+        self.statements.append(ast.Assign([ast.Name(name, ast.Store())], expr))
+        return ast.copy_location(ast.Name(name, ast.Load()), expr)
+
+    def display(self, expr: ast.expr) -> ast.expr:
+        """Call saferepr on the expression."""
+        return self.helper("_saferepr", expr)
+
+    def helper(self, name: str, *args: ast.expr) -> ast.expr:
+        """Call a helper in this module."""
+        py_name = ast.Name("@pytest_ar", ast.Load())
+        attr = ast.Attribute(py_name, name, ast.Load())
+        return ast.Call(attr, list(args), [])
+
+    def builtin(self, name: str) -> ast.Attribute:
+        """Return the builtin called *name*."""
+        builtin_name = ast.Name("@py_builtins", ast.Load())
+        return ast.Attribute(builtin_name, name, ast.Load())
+
+    def explanation_param(self, expr: ast.expr) -> str:
+        """Return a new named %-formatting placeholder for expr.
+
+        This creates a %-formatting placeholder for expr in the
+        current formatting context, e.g. ``%(py0)s``.  The placeholder
+        and expr are placed in the current format context so that it
+        can be used on the next call to .pop_format_context().
+        """
+        specifier = "py" + str(next(self.variable_counter))
+        self.explanation_specifiers[specifier] = expr
+        return "%(" + specifier + ")s"
+
+    def push_format_context(self) -> None:
+        """Create a new formatting context.
+
+        The format context is used for when an explanation wants to
+        have a variable value formatted in the assertion message.  In
+        this case the value required can be added using
+        .explanation_param().  Finally .pop_format_context() is used
+        to format a string of %-formatted values as added by
+        .explanation_param().
+        """
+        self.explanation_specifiers: dict[str, ast.expr] = {}
+        self.stack.append(self.explanation_specifiers)
+
+    def pop_format_context(self, expl_expr: ast.expr) -> ast.Name:
+        """Format the %-formatted string with current format context.
+
+        The expl_expr should be an str ast.expr instance constructed from
+        the %-placeholders created by .explanation_param().  This will
+        add the required code to format said string to .expl_stmts and
+        return the ast.Name instance of the formatted string.
+        """
+        current = self.stack.pop()
+        if self.stack:
+            self.explanation_specifiers = self.stack[-1]
+        keys: list[ast.expr | None] = [ast.Constant(key) for key in current.keys()]
+        format_dict = ast.Dict(keys, list(current.values()))
+        form = ast.BinOp(expl_expr, ast.Mod(), format_dict)
+        name = "@py_format" + str(next(self.variable_counter))
+        if self.enable_assertion_pass_hook:
+            self.format_variables.append(name)
+        self.expl_stmts.append(ast.Assign([ast.Name(name, ast.Store())], form))
+        return ast.Name(name, ast.Load())
+
+    def generic_visit(self, node: ast.AST) -> tuple[ast.Name, str]:
+        """Handle expressions we don't have custom code for."""
+        assert isinstance(node, ast.expr)
+        res = self.assign(node)
+        return res, self.explanation_param(self.display(res))
+
+    def visit_Assert(self, assert_: ast.Assert) -> list[ast.stmt]:
+        """Return the AST statements to replace the ast.Assert instance.
+
+        This rewrites the test of an assertion to provide
+        intermediate values and replace it with an if statement which
+        raises an assertion error with a detailed explanation in case
+        the expression is false.
+        """
+        if isinstance(assert_.test, ast.Tuple) and len(assert_.test.elts) >= 1:
+            import warnings
+
+            from _pytest.warning_types import PytestAssertRewriteWarning
+
+            # TODO: This assert should not be needed.
+            assert self.module_path is not None
+            warnings.warn_explicit(
+                PytestAssertRewriteWarning(
+                    "assertion is always true, perhaps remove parentheses?"
+                ),
+                category=None,
+                filename=self.module_path,
+                lineno=assert_.lineno,
+            )
+
+        self.statements: list[ast.stmt] = []
+        self.variables: list[str] = []
+        self.variable_counter = itertools.count()
+
+        if self.enable_assertion_pass_hook:
+            self.format_variables: list[str] = []
+
+        self.stack: list[dict[str, ast.expr]] = []
+        self.expl_stmts: list[ast.stmt] = []
+        self.push_format_context()
+        # Rewrite assert into a bunch of statements.
+        top_condition, explanation = self.visit(assert_.test)
+
+        negation = ast.UnaryOp(ast.Not(), top_condition)
+
+        if self.enable_assertion_pass_hook:  # Experimental pytest_assertion_pass hook
+            msg = self.pop_format_context(ast.Constant(explanation))
+
+            # Failed
+            if assert_.msg:
+                assertmsg = self.helper("_format_assertmsg", assert_.msg)
+                gluestr = "\n>assert "
+            else:
+                assertmsg = ast.Constant("")
+                gluestr = "assert "
+            err_explanation = ast.BinOp(ast.Constant(gluestr), ast.Add(), msg)
+            err_msg = ast.BinOp(assertmsg, ast.Add(), err_explanation)
+            err_name = ast.Name("AssertionError", ast.Load())
+            fmt = self.helper("_format_explanation", err_msg)
+            exc = ast.Call(err_name, [fmt], [])
+            raise_ = ast.Raise(exc, None)
+            statements_fail = []
+            statements_fail.extend(self.expl_stmts)
+            statements_fail.append(raise_)
+
+            # Passed
+            fmt_pass = self.helper("_format_explanation", msg)
+            orig = _get_assertion_exprs(self.source)[assert_.lineno]
+            hook_call_pass = ast.Expr(
+                self.helper(
+                    "_call_assertion_pass",
+                    ast.Constant(assert_.lineno),
+                    ast.Constant(orig),
+                    fmt_pass,
+                )
+            )
+            # If any hooks implement assert_pass hook
+            hook_impl_test = ast.If(
+                self.helper("_check_if_assertion_pass_impl"),
+                [*self.expl_stmts, hook_call_pass],
+                [],
+            )
+            statements_pass: list[ast.stmt] = [hook_impl_test]
+
+            # Test for assertion condition
+            main_test = ast.If(negation, statements_fail, statements_pass)
+            self.statements.append(main_test)
+            if self.format_variables:
+                variables: list[ast.expr] = [
+                    ast.Name(name, ast.Store()) for name in self.format_variables
+                ]
+                clear_format = ast.Assign(variables, ast.Constant(None))
+                self.statements.append(clear_format)
+
+        else:  # Original assertion rewriting
+            # Create failure message.
+            body = self.expl_stmts
+            self.statements.append(ast.If(negation, body, []))
+            if assert_.msg:
+                assertmsg = self.helper("_format_assertmsg", assert_.msg)
+                explanation = "\n>assert " + explanation
+            else:
+                assertmsg = ast.Constant("")
+                explanation = "assert " + explanation
+            template = ast.BinOp(assertmsg, ast.Add(), ast.Constant(explanation))
+            msg = self.pop_format_context(template)
+            fmt = self.helper("_format_explanation", msg)
+            err_name = ast.Name("AssertionError", ast.Load())
+            exc = ast.Call(err_name, [fmt], [])
+            raise_ = ast.Raise(exc, None)
+
+            body.append(raise_)
+
+        # Clear temporary variables by setting them to None.
+        if self.variables:
+            variables = [ast.Name(name, ast.Store()) for name in self.variables]
+            clear = ast.Assign(variables, ast.Constant(None))
+            self.statements.append(clear)
+        # Fix locations (line numbers/column offsets).
+        for stmt in self.statements:
+            for node in traverse_node(stmt):
+                if getattr(node, "lineno", None) is None:
+                    # apply the assertion location to all generated ast nodes without source location
+                    # and preserve the location of existing nodes or generated nodes with an correct location.
+                    ast.copy_location(node, assert_)
+        return self.statements
+
+    def visit_NamedExpr(self, name: ast.NamedExpr) -> tuple[ast.NamedExpr, str]:
+        # This method handles the 'walrus operator' repr of the target
+        # name if it's a local variable or _should_repr_global_name()
+        # thinks it's acceptable.
+        locs = ast.Call(self.builtin("locals"), [], [])
+        target_id = name.target.id
+        inlocs = ast.Compare(ast.Constant(target_id), [ast.In()], [locs])
+        dorepr = self.helper("_should_repr_global_name", name)
+        test = ast.BoolOp(ast.Or(), [inlocs, dorepr])
+        expr = ast.IfExp(test, self.display(name), ast.Constant(target_id))
+        return name, self.explanation_param(expr)
+
+    def visit_Name(self, name: ast.Name) -> tuple[ast.Name, str]:
+        # Display the repr of the name if it's a local variable or
+        # _should_repr_global_name() thinks it's acceptable.
+        locs = ast.Call(self.builtin("locals"), [], [])
+        inlocs = ast.Compare(ast.Constant(name.id), [ast.In()], [locs])
+        dorepr = self.helper("_should_repr_global_name", name)
+        test = ast.BoolOp(ast.Or(), [inlocs, dorepr])
+        expr = ast.IfExp(test, self.display(name), ast.Constant(name.id))
+        return name, self.explanation_param(expr)
+
+    def visit_BoolOp(self, boolop: ast.BoolOp) -> tuple[ast.Name, str]:
+        res_var = self.variable()
+        expl_list = self.assign(ast.List([], ast.Load()))
+        app = ast.Attribute(expl_list, "append", ast.Load())
+        is_or = int(isinstance(boolop.op, ast.Or))
+        body = save = self.statements
+        fail_save = self.expl_stmts
+        levels = len(boolop.values) - 1
+        self.push_format_context()
+        # Process each operand, short-circuiting if needed.
+        for i, v in enumerate(boolop.values):
+            if i:
+                fail_inner: list[ast.stmt] = []
+                # cond is set in a prior loop iteration below
+                self.expl_stmts.append(ast.If(cond, fail_inner, []))  # noqa: F821
+                self.expl_stmts = fail_inner
+                # Check if the left operand is a ast.NamedExpr and the value has already been visited
+                if (
+                    isinstance(v, ast.Compare)
+                    and isinstance(v.left, ast.NamedExpr)
+                    and v.left.target.id
+                    in [
+                        ast_expr.id
+                        for ast_expr in boolop.values[:i]
+                        if hasattr(ast_expr, "id")
+                    ]
+                ):
+                    pytest_temp = self.variable()
+                    self.variables_overwrite[self.scope][v.left.target.id] = v.left  # type:ignore[assignment]
+                    v.left.target.id = pytest_temp
+            self.push_format_context()
+            res, expl = self.visit(v)
+            body.append(ast.Assign([ast.Name(res_var, ast.Store())], res))
+            expl_format = self.pop_format_context(ast.Constant(expl))
+            call = ast.Call(app, [expl_format], [])
+            self.expl_stmts.append(ast.Expr(call))
+            if i < levels:
+                cond: ast.expr = res
+                if is_or:
+                    cond = ast.UnaryOp(ast.Not(), cond)
+                inner: list[ast.stmt] = []
+                self.statements.append(ast.If(cond, inner, []))
+                self.statements = body = inner
+        self.statements = save
+        self.expl_stmts = fail_save
+        expl_template = self.helper("_format_boolop", expl_list, ast.Constant(is_or))
+        expl = self.pop_format_context(expl_template)
+        return ast.Name(res_var, ast.Load()), self.explanation_param(expl)
+
+    def visit_UnaryOp(self, unary: ast.UnaryOp) -> tuple[ast.Name, str]:
+        pattern = UNARY_MAP[unary.op.__class__]
+        operand_res, operand_expl = self.visit(unary.operand)
+        res = self.assign(ast.copy_location(ast.UnaryOp(unary.op, operand_res), unary))
+        return res, pattern % (operand_expl,)
+
+    def visit_BinOp(self, binop: ast.BinOp) -> tuple[ast.Name, str]:
+        symbol = BINOP_MAP[binop.op.__class__]
+        left_expr, left_expl = self.visit(binop.left)
+        right_expr, right_expl = self.visit(binop.right)
+        explanation = f"({left_expl} {symbol} {right_expl})"
+        res = self.assign(
+            ast.copy_location(ast.BinOp(left_expr, binop.op, right_expr), binop)
+        )
+        return res, explanation
+
+    def visit_Call(self, call: ast.Call) -> tuple[ast.Name, str]:
+        new_func, func_expl = self.visit(call.func)
+        arg_expls = []
+        new_args = []
+        new_kwargs = []
+        for arg in call.args:
+            if isinstance(arg, ast.Name) and arg.id in self.variables_overwrite.get(
+                self.scope, {}
+            ):
+                arg = self.variables_overwrite[self.scope][arg.id]  # type:ignore[assignment]
+            res, expl = self.visit(arg)
+            arg_expls.append(expl)
+            new_args.append(res)
+        for keyword in call.keywords:
+            if isinstance(
+                keyword.value, ast.Name
+            ) and keyword.value.id in self.variables_overwrite.get(self.scope, {}):
+                keyword.value = self.variables_overwrite[self.scope][keyword.value.id]  # type:ignore[assignment]
+            res, expl = self.visit(keyword.value)
+            new_kwargs.append(ast.keyword(keyword.arg, res))
+            if keyword.arg:
+                arg_expls.append(keyword.arg + "=" + expl)
+            else:  # **args have `arg` keywords with an .arg of None
+                arg_expls.append("**" + expl)
+
+        expl = "{}({})".format(func_expl, ", ".join(arg_expls))
+        new_call = ast.copy_location(ast.Call(new_func, new_args, new_kwargs), call)
+        res = self.assign(new_call)
+        res_expl = self.explanation_param(self.display(res))
+        outer_expl = f"{res_expl}\n{{{res_expl} = {expl}\n}}"
+        return res, outer_expl
+
+    def visit_Starred(self, starred: ast.Starred) -> tuple[ast.Starred, str]:
+        # A Starred node can appear in a function call.
+        res, expl = self.visit(starred.value)
+        new_starred = ast.Starred(res, starred.ctx)
+        return new_starred, "*" + expl
+
+    def visit_Attribute(self, attr: ast.Attribute) -> tuple[ast.Name, str]:
+        if not isinstance(attr.ctx, ast.Load):
+            return self.generic_visit(attr)
+        value, value_expl = self.visit(attr.value)
+        res = self.assign(
+            ast.copy_location(ast.Attribute(value, attr.attr, ast.Load()), attr)
+        )
+        res_expl = self.explanation_param(self.display(res))
+        pat = "%s\n{%s = %s.%s\n}"
+        expl = pat % (res_expl, res_expl, value_expl, attr.attr)
+        return res, expl
+
+    def visit_Compare(self, comp: ast.Compare) -> tuple[ast.expr, str]:
+        self.push_format_context()
+        # We first check if we have overwritten a variable in the previous assert
+        if isinstance(
+            comp.left, ast.Name
+        ) and comp.left.id in self.variables_overwrite.get(self.scope, {}):
+            comp.left = self.variables_overwrite[self.scope][comp.left.id]  # type:ignore[assignment]
+        if isinstance(comp.left, ast.NamedExpr):
+            self.variables_overwrite[self.scope][comp.left.target.id] = comp.left  # type:ignore[assignment]
+        left_res, left_expl = self.visit(comp.left)
+        if isinstance(comp.left, (ast.Compare, ast.BoolOp)):
+            left_expl = f"({left_expl})"
+        res_variables = [self.variable() for i in range(len(comp.ops))]
+        load_names: list[ast.expr] = [ast.Name(v, ast.Load()) for v in res_variables]
+        store_names = [ast.Name(v, ast.Store()) for v in res_variables]
+        it = zip(range(len(comp.ops)), comp.ops, comp.comparators)
+        expls: list[ast.expr] = []
+        syms: list[ast.expr] = []
+        results = [left_res]
+        for i, op, next_operand in it:
+            if (
+                isinstance(next_operand, ast.NamedExpr)
+                and isinstance(left_res, ast.Name)
+                and next_operand.target.id == left_res.id
+            ):
+                next_operand.target.id = self.variable()
+                self.variables_overwrite[self.scope][left_res.id] = next_operand  # type:ignore[assignment]
+            next_res, next_expl = self.visit(next_operand)
+            if isinstance(next_operand, (ast.Compare, ast.BoolOp)):
+                next_expl = f"({next_expl})"
+            results.append(next_res)
+            sym = BINOP_MAP[op.__class__]
+            syms.append(ast.Constant(sym))
+            expl = f"{left_expl} {sym} {next_expl}"
+            expls.append(ast.Constant(expl))
+            res_expr = ast.copy_location(ast.Compare(left_res, [op], [next_res]), comp)
+            self.statements.append(ast.Assign([store_names[i]], res_expr))
+            left_res, left_expl = next_res, next_expl
+        # Use pytest.assertion.util._reprcompare if that's available.
+        expl_call = self.helper(
+            "_call_reprcompare",
+            ast.Tuple(syms, ast.Load()),
+            ast.Tuple(load_names, ast.Load()),
+            ast.Tuple(expls, ast.Load()),
+            ast.Tuple(results, ast.Load()),
+        )
+        if len(comp.ops) > 1:
+            res: ast.expr = ast.BoolOp(ast.And(), load_names)
+        else:
+            res = load_names[0]
+
+        return res, self.explanation_param(self.pop_format_context(expl_call))
+
+
+def try_makedirs(cache_dir: Path) -> bool:
+    """Attempt to create the given directory and sub-directories exist.
+
+    Returns True if successful or if it already exists.
+    """
+    try:
+        os.makedirs(cache_dir, exist_ok=True)
+    except (FileNotFoundError, NotADirectoryError, FileExistsError):
+        # One of the path components was not a directory:
+        # - we're in a zip file
+        # - it is a file
+        return False
+    except PermissionError:
+        return False
+    except OSError as e:
+        # as of now, EROFS doesn't have an equivalent OSError-subclass
+        #
+        # squashfuse_ll returns ENOSYS "OSError: [Errno 38] Function not
+        # implemented" for a read-only error
+        if e.errno in {errno.EROFS, errno.ENOSYS}:
+            return False
+        raise
+    return True
+
+
+def get_cache_dir(file_path: Path) -> Path:
+    """Return the cache directory to write .pyc files for the given .py file path."""
+    if sys.pycache_prefix:
+        # given:
+        #   prefix = '/tmp/pycs'
+        #   path = '/home/user/proj/test_app.py'
+        # we want:
+        #   '/tmp/pycs/home/user/proj'
+        return Path(sys.pycache_prefix) / Path(*file_path.parts[1:-1])
+    else:
+        # classic pycache directory
+        return file_path.parent / "__pycache__"
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/truncate.py b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/truncate.py
new file mode 100644
index 00000000..b67f02cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/truncate.py
@@ -0,0 +1,117 @@
+"""Utilities for truncating assertion output.
+
+Current default behaviour is to truncate assertion explanations at
+terminal lines, unless running with an assertions verbosity level of at least 2 or running on CI.
+"""
+
+from __future__ import annotations
+
+from _pytest.assertion import util
+from _pytest.config import Config
+from _pytest.nodes import Item
+
+
+DEFAULT_MAX_LINES = 8
+DEFAULT_MAX_CHARS = 8 * 80
+USAGE_MSG = "use '-vv' to show"
+
+
+def truncate_if_required(
+    explanation: list[str], item: Item, max_length: int | None = None
+) -> list[str]:
+    """Truncate this assertion explanation if the given test item is eligible."""
+    if _should_truncate_item(item):
+        return _truncate_explanation(explanation)
+    return explanation
+
+
+def _should_truncate_item(item: Item) -> bool:
+    """Whether or not this test item is eligible for truncation."""
+    verbose = item.config.get_verbosity(Config.VERBOSITY_ASSERTIONS)
+    return verbose < 2 and not util.running_on_ci()
+
+
+def _truncate_explanation(
+    input_lines: list[str],
+    max_lines: int | None = None,
+    max_chars: int | None = None,
+) -> list[str]:
+    """Truncate given list of strings that makes up the assertion explanation.
+
+    Truncates to either 8 lines, or 640 characters - whichever the input reaches
+    first, taking the truncation explanation into account. The remaining lines
+    will be replaced by a usage message.
+    """
+    if max_lines is None:
+        max_lines = DEFAULT_MAX_LINES
+    if max_chars is None:
+        max_chars = DEFAULT_MAX_CHARS
+
+    # Check if truncation required
+    input_char_count = len("".join(input_lines))
+    # The length of the truncation explanation depends on the number of lines
+    # removed but is at least 68 characters:
+    # The real value is
+    # 64 (for the base message:
+    # '...\n...Full output truncated (1 line hidden), use '-vv' to show")'
+    # )
+    # + 1 (for plural)
+    # + int(math.log10(len(input_lines) - max_lines)) (number of hidden line, at least 1)
+    # + 3 for the '...' added to the truncated line
+    # But if there's more than 100 lines it's very likely that we're going to
+    # truncate, so we don't need the exact value using log10.
+    tolerable_max_chars = (
+        max_chars + 70  # 64 + 1 (for plural) + 2 (for '99') + 3 for '...'
+    )
+    # The truncation explanation add two lines to the output
+    tolerable_max_lines = max_lines + 2
+    if (
+        len(input_lines) <= tolerable_max_lines
+        and input_char_count <= tolerable_max_chars
+    ):
+        return input_lines
+    # Truncate first to max_lines, and then truncate to max_chars if necessary
+    truncated_explanation = input_lines[:max_lines]
+    truncated_char = True
+    # We reevaluate the need to truncate chars following removal of some lines
+    if len("".join(truncated_explanation)) > tolerable_max_chars:
+        truncated_explanation = _truncate_by_char_count(
+            truncated_explanation, max_chars
+        )
+    else:
+        truncated_char = False
+
+    truncated_line_count = len(input_lines) - len(truncated_explanation)
+    if truncated_explanation[-1]:
+        # Add ellipsis and take into account part-truncated final line
+        truncated_explanation[-1] = truncated_explanation[-1] + "..."
+        if truncated_char:
+            # It's possible that we did not remove any char from this line
+            truncated_line_count += 1
+    else:
+        # Add proper ellipsis when we were able to fit a full line exactly
+        truncated_explanation[-1] = "..."
+    return [
+        *truncated_explanation,
+        "",
+        f"...Full output truncated ({truncated_line_count} line"
+        f"{'' if truncated_line_count == 1 else 's'} hidden), {USAGE_MSG}",
+    ]
+
+
+def _truncate_by_char_count(input_lines: list[str], max_chars: int) -> list[str]:
+    # Find point at which input length exceeds total allowed length
+    iterated_char_count = 0
+    for iterated_index, input_line in enumerate(input_lines):
+        if iterated_char_count + len(input_line) > max_chars:
+            break
+        iterated_char_count += len(input_line)
+
+    # Create truncated explanation with modified final line
+    truncated_result = input_lines[:iterated_index]
+    final_line = input_lines[iterated_index]
+    if final_line:
+        final_line_truncate_point = max_chars - iterated_char_count
+        final_line = final_line[:final_line_truncate_point]
+    truncated_result.append(final_line)
+    return truncated_result
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/assertion/util.py b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/util.py
new file mode 100644
index 00000000..4dc1af4a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/assertion/util.py
@@ -0,0 +1,609 @@
+# mypy: allow-untyped-defs
+"""Utilities for assertion debugging."""
+
+from __future__ import annotations
+
+import collections.abc
+import os
+import pprint
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import Iterable
+from typing import Literal
+from typing import Mapping
+from typing import Protocol
+from typing import Sequence
+from unicodedata import normalize
+
+from _pytest import outcomes
+import _pytest._code
+from _pytest._io.pprint import PrettyPrinter
+from _pytest._io.saferepr import saferepr
+from _pytest._io.saferepr import saferepr_unlimited
+from _pytest.config import Config
+
+
+# The _reprcompare attribute on the util module is used by the new assertion
+# interpretation code and assertion rewriter to detect this plugin was
+# loaded and in turn call the hooks defined here as part of the
+# DebugInterpreter.
+_reprcompare: Callable[[str, object, object], str | None] | None = None
+
+# Works similarly as _reprcompare attribute. Is populated with the hook call
+# when pytest_runtest_setup is called.
+_assertion_pass: Callable[[int, str, str], None] | None = None
+
+# Config object which is assigned during pytest_runtest_protocol.
+_config: Config | None = None
+
+
+class _HighlightFunc(Protocol):
+    def __call__(self, source: str, lexer: Literal["diff", "python"] = "python") -> str:
+        """Apply highlighting to the given source."""
+
+
+def format_explanation(explanation: str) -> str:
+    r"""Format an explanation.
+
+    Normally all embedded newlines are escaped, however there are
+    three exceptions: \n{, \n} and \n~.  The first two are intended
+    cover nested explanations, see function and attribute explanations
+    for examples (.visit_Call(), visit_Attribute()).  The last one is
+    for when one explanation needs to span multiple lines, e.g. when
+    displaying diffs.
+    """
+    lines = _split_explanation(explanation)
+    result = _format_lines(lines)
+    return "\n".join(result)
+
+
+def _split_explanation(explanation: str) -> list[str]:
+    r"""Return a list of individual lines in the explanation.
+
+    This will return a list of lines split on '\n{', '\n}' and '\n~'.
+    Any other newlines will be escaped and appear in the line as the
+    literal '\n' characters.
+    """
+    raw_lines = (explanation or "").split("\n")
+    lines = [raw_lines[0]]
+    for values in raw_lines[1:]:
+        if values and values[0] in ["{", "}", "~", ">"]:
+            lines.append(values)
+        else:
+            lines[-1] += "\\n" + values
+    return lines
+
+
+def _format_lines(lines: Sequence[str]) -> list[str]:
+    """Format the individual lines.
+
+    This will replace the '{', '}' and '~' characters of our mini formatting
+    language with the proper 'where ...', 'and ...' and ' + ...' text, taking
+    care of indentation along the way.
+
+    Return a list of formatted lines.
+    """
+    result = list(lines[:1])
+    stack = [0]
+    stackcnt = [0]
+    for line in lines[1:]:
+        if line.startswith("{"):
+            if stackcnt[-1]:
+                s = "and   "
+            else:
+                s = "where "
+            stack.append(len(result))
+            stackcnt[-1] += 1
+            stackcnt.append(0)
+            result.append(" +" + "  " * (len(stack) - 1) + s + line[1:])
+        elif line.startswith("}"):
+            stack.pop()
+            stackcnt.pop()
+            result[stack[-1]] += line[1:]
+        else:
+            assert line[0] in ["~", ">"]
+            stack[-1] += 1
+            indent = len(stack) if line.startswith("~") else len(stack) - 1
+            result.append("  " * indent + line[1:])
+    assert len(stack) == 1
+    return result
+
+
+def issequence(x: Any) -> bool:
+    return isinstance(x, collections.abc.Sequence) and not isinstance(x, str)
+
+
+def istext(x: Any) -> bool:
+    return isinstance(x, str)
+
+
+def isdict(x: Any) -> bool:
+    return isinstance(x, dict)
+
+
+def isset(x: Any) -> bool:
+    return isinstance(x, (set, frozenset))
+
+
+def isnamedtuple(obj: Any) -> bool:
+    return isinstance(obj, tuple) and getattr(obj, "_fields", None) is not None
+
+
+def isdatacls(obj: Any) -> bool:
+    return getattr(obj, "__dataclass_fields__", None) is not None
+
+
+def isattrs(obj: Any) -> bool:
+    return getattr(obj, "__attrs_attrs__", None) is not None
+
+
+def isiterable(obj: Any) -> bool:
+    try:
+        iter(obj)
+        return not istext(obj)
+    except Exception:
+        return False
+
+
+def has_default_eq(
+    obj: object,
+) -> bool:
+    """Check if an instance of an object contains the default eq
+
+    First, we check if the object's __eq__ attribute has __code__,
+    if so, we check the equally of the method code filename (__code__.co_filename)
+    to the default one generated by the dataclass and attr module
+    for dataclasses the default co_filename is <string>, for attrs class, the __eq__ should contain "attrs eq generated"
+    """
+    # inspired from https://github.com/willmcgugan/rich/blob/07d51ffc1aee6f16bd2e5a25b4e82850fb9ed778/rich/pretty.py#L68
+    if hasattr(obj.__eq__, "__code__") and hasattr(obj.__eq__.__code__, "co_filename"):
+        code_filename = obj.__eq__.__code__.co_filename
+
+        if isattrs(obj):
+            return "attrs generated eq" in code_filename
+
+        return code_filename == "<string>"  # data class
+    return True
+
+
+def assertrepr_compare(
+    config, op: str, left: Any, right: Any, use_ascii: bool = False
+) -> list[str] | None:
+    """Return specialised explanations for some operators/operands."""
+    verbose = config.get_verbosity(Config.VERBOSITY_ASSERTIONS)
+
+    # Strings which normalize equal are often hard to distinguish when printed; use ascii() to make this easier.
+    # See issue #3246.
+    use_ascii = (
+        isinstance(left, str)
+        and isinstance(right, str)
+        and normalize("NFD", left) == normalize("NFD", right)
+    )
+
+    if verbose > 1:
+        left_repr = saferepr_unlimited(left, use_ascii=use_ascii)
+        right_repr = saferepr_unlimited(right, use_ascii=use_ascii)
+    else:
+        # XXX: "15 chars indentation" is wrong
+        #      ("E       AssertionError: assert "); should use term width.
+        maxsize = (
+            80 - 15 - len(op) - 2
+        ) // 2  # 15 chars indentation, 1 space around op
+
+        left_repr = saferepr(left, maxsize=maxsize, use_ascii=use_ascii)
+        right_repr = saferepr(right, maxsize=maxsize, use_ascii=use_ascii)
+
+    summary = f"{left_repr} {op} {right_repr}"
+    highlighter = config.get_terminal_writer()._highlight
+
+    explanation = None
+    try:
+        if op == "==":
+            explanation = _compare_eq_any(left, right, highlighter, verbose)
+        elif op == "not in":
+            if istext(left) and istext(right):
+                explanation = _notin_text(left, right, verbose)
+        elif op == "!=":
+            if isset(left) and isset(right):
+                explanation = ["Both sets are equal"]
+        elif op == ">=":
+            if isset(left) and isset(right):
+                explanation = _compare_gte_set(left, right, highlighter, verbose)
+        elif op == "<=":
+            if isset(left) and isset(right):
+                explanation = _compare_lte_set(left, right, highlighter, verbose)
+        elif op == ">":
+            if isset(left) and isset(right):
+                explanation = _compare_gt_set(left, right, highlighter, verbose)
+        elif op == "<":
+            if isset(left) and isset(right):
+                explanation = _compare_lt_set(left, right, highlighter, verbose)
+
+    except outcomes.Exit:
+        raise
+    except Exception:
+        repr_crash = _pytest._code.ExceptionInfo.from_current()._getreprcrash()
+        explanation = [
+            f"(pytest_assertion plugin: representation of details failed: {repr_crash}.",
+            " Probably an object has a faulty __repr__.)",
+        ]
+
+    if not explanation:
+        return None
+
+    if explanation[0] != "":
+        explanation = ["", *explanation]
+    return [summary, *explanation]
+
+
+def _compare_eq_any(
+    left: Any, right: Any, highlighter: _HighlightFunc, verbose: int = 0
+) -> list[str]:
+    explanation = []
+    if istext(left) and istext(right):
+        explanation = _diff_text(left, right, verbose)
+    else:
+        from _pytest.python_api import ApproxBase
+
+        if isinstance(left, ApproxBase) or isinstance(right, ApproxBase):
+            # Although the common order should be obtained == expected, this ensures both ways
+            approx_side = left if isinstance(left, ApproxBase) else right
+            other_side = right if isinstance(left, ApproxBase) else left
+
+            explanation = approx_side._repr_compare(other_side)
+        elif type(left) is type(right) and (
+            isdatacls(left) or isattrs(left) or isnamedtuple(left)
+        ):
+            # Note: unlike dataclasses/attrs, namedtuples compare only the
+            # field values, not the type or field names. But this branch
+            # intentionally only handles the same-type case, which was often
+            # used in older code bases before dataclasses/attrs were available.
+            explanation = _compare_eq_cls(left, right, highlighter, verbose)
+        elif issequence(left) and issequence(right):
+            explanation = _compare_eq_sequence(left, right, highlighter, verbose)
+        elif isset(left) and isset(right):
+            explanation = _compare_eq_set(left, right, highlighter, verbose)
+        elif isdict(left) and isdict(right):
+            explanation = _compare_eq_dict(left, right, highlighter, verbose)
+
+        if isiterable(left) and isiterable(right):
+            expl = _compare_eq_iterable(left, right, highlighter, verbose)
+            explanation.extend(expl)
+
+    return explanation
+
+
+def _diff_text(left: str, right: str, verbose: int = 0) -> list[str]:
+    """Return the explanation for the diff between text.
+
+    Unless --verbose is used this will skip leading and trailing
+    characters which are identical to keep the diff minimal.
+    """
+    from difflib import ndiff
+
+    explanation: list[str] = []
+
+    if verbose < 1:
+        i = 0  # just in case left or right has zero length
+        for i in range(min(len(left), len(right))):
+            if left[i] != right[i]:
+                break
+        if i > 42:
+            i -= 10  # Provide some context
+            explanation = [
+                f"Skipping {i} identical leading characters in diff, use -v to show"
+            ]
+            left = left[i:]
+            right = right[i:]
+        if len(left) == len(right):
+            for i in range(len(left)):
+                if left[-i] != right[-i]:
+                    break
+            if i > 42:
+                i -= 10  # Provide some context
+                explanation += [
+                    f"Skipping {i} identical trailing "
+                    "characters in diff, use -v to show"
+                ]
+                left = left[:-i]
+                right = right[:-i]
+    keepends = True
+    if left.isspace() or right.isspace():
+        left = repr(str(left))
+        right = repr(str(right))
+        explanation += ["Strings contain only whitespace, escaping them using repr()"]
+    # "right" is the expected base against which we compare "left",
+    # see https://github.com/pytest-dev/pytest/issues/3333
+    explanation += [
+        line.strip("\n")
+        for line in ndiff(right.splitlines(keepends), left.splitlines(keepends))
+    ]
+    return explanation
+
+
+def _compare_eq_iterable(
+    left: Iterable[Any],
+    right: Iterable[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    if verbose <= 0 and not running_on_ci():
+        return ["Use -v to get more diff"]
+    # dynamic import to speedup pytest
+    import difflib
+
+    left_formatting = PrettyPrinter().pformat(left).splitlines()
+    right_formatting = PrettyPrinter().pformat(right).splitlines()
+
+    explanation = ["", "Full diff:"]
+    # "right" is the expected base against which we compare "left",
+    # see https://github.com/pytest-dev/pytest/issues/3333
+    explanation.extend(
+        highlighter(
+            "\n".join(
+                line.rstrip()
+                for line in difflib.ndiff(right_formatting, left_formatting)
+            ),
+            lexer="diff",
+        ).splitlines()
+    )
+    return explanation
+
+
+def _compare_eq_sequence(
+    left: Sequence[Any],
+    right: Sequence[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    comparing_bytes = isinstance(left, bytes) and isinstance(right, bytes)
+    explanation: list[str] = []
+    len_left = len(left)
+    len_right = len(right)
+    for i in range(min(len_left, len_right)):
+        if left[i] != right[i]:
+            if comparing_bytes:
+                # when comparing bytes, we want to see their ascii representation
+                # instead of their numeric values (#5260)
+                # using a slice gives us the ascii representation:
+                # >>> s = b'foo'
+                # >>> s[0]
+                # 102
+                # >>> s[0:1]
+                # b'f'
+                left_value = left[i : i + 1]
+                right_value = right[i : i + 1]
+            else:
+                left_value = left[i]
+                right_value = right[i]
+
+            explanation.append(
+                f"At index {i} diff:"
+                f" {highlighter(repr(left_value))} != {highlighter(repr(right_value))}"
+            )
+            break
+
+    if comparing_bytes:
+        # when comparing bytes, it doesn't help to show the "sides contain one or more
+        # items" longer explanation, so skip it
+
+        return explanation
+
+    len_diff = len_left - len_right
+    if len_diff:
+        if len_diff > 0:
+            dir_with_more = "Left"
+            extra = saferepr(left[len_right])
+        else:
+            len_diff = 0 - len_diff
+            dir_with_more = "Right"
+            extra = saferepr(right[len_left])
+
+        if len_diff == 1:
+            explanation += [
+                f"{dir_with_more} contains one more item: {highlighter(extra)}"
+            ]
+        else:
+            explanation += [
+                "%s contains %d more items, first extra item: %s"
+                % (dir_with_more, len_diff, highlighter(extra))
+            ]
+    return explanation
+
+
+def _compare_eq_set(
+    left: AbstractSet[Any],
+    right: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    explanation = []
+    explanation.extend(_set_one_sided_diff("left", left, right, highlighter))
+    explanation.extend(_set_one_sided_diff("right", right, left, highlighter))
+    return explanation
+
+
+def _compare_gt_set(
+    left: AbstractSet[Any],
+    right: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    explanation = _compare_gte_set(left, right, highlighter)
+    if not explanation:
+        return ["Both sets are equal"]
+    return explanation
+
+
+def _compare_lt_set(
+    left: AbstractSet[Any],
+    right: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    explanation = _compare_lte_set(left, right, highlighter)
+    if not explanation:
+        return ["Both sets are equal"]
+    return explanation
+
+
+def _compare_gte_set(
+    left: AbstractSet[Any],
+    right: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    return _set_one_sided_diff("right", right, left, highlighter)
+
+
+def _compare_lte_set(
+    left: AbstractSet[Any],
+    right: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    return _set_one_sided_diff("left", left, right, highlighter)
+
+
+def _set_one_sided_diff(
+    posn: str,
+    set1: AbstractSet[Any],
+    set2: AbstractSet[Any],
+    highlighter: _HighlightFunc,
+) -> list[str]:
+    explanation = []
+    diff = set1 - set2
+    if diff:
+        explanation.append(f"Extra items in the {posn} set:")
+        for item in diff:
+            explanation.append(highlighter(saferepr(item)))
+    return explanation
+
+
+def _compare_eq_dict(
+    left: Mapping[Any, Any],
+    right: Mapping[Any, Any],
+    highlighter: _HighlightFunc,
+    verbose: int = 0,
+) -> list[str]:
+    explanation: list[str] = []
+    set_left = set(left)
+    set_right = set(right)
+    common = set_left.intersection(set_right)
+    same = {k: left[k] for k in common if left[k] == right[k]}
+    if same and verbose < 2:
+        explanation += [f"Omitting {len(same)} identical items, use -vv to show"]
+    elif same:
+        explanation += ["Common items:"]
+        explanation += highlighter(pprint.pformat(same)).splitlines()
+    diff = {k for k in common if left[k] != right[k]}
+    if diff:
+        explanation += ["Differing items:"]
+        for k in diff:
+            explanation += [
+                highlighter(saferepr({k: left[k]}))
+                + " != "
+                + highlighter(saferepr({k: right[k]}))
+            ]
+    extra_left = set_left - set_right
+    len_extra_left = len(extra_left)
+    if len_extra_left:
+        explanation.append(
+            "Left contains %d more item%s:"
+            % (len_extra_left, "" if len_extra_left == 1 else "s")
+        )
+        explanation.extend(
+            highlighter(pprint.pformat({k: left[k] for k in extra_left})).splitlines()
+        )
+    extra_right = set_right - set_left
+    len_extra_right = len(extra_right)
+    if len_extra_right:
+        explanation.append(
+            "Right contains %d more item%s:"
+            % (len_extra_right, "" if len_extra_right == 1 else "s")
+        )
+        explanation.extend(
+            highlighter(pprint.pformat({k: right[k] for k in extra_right})).splitlines()
+        )
+    return explanation
+
+
+def _compare_eq_cls(
+    left: Any, right: Any, highlighter: _HighlightFunc, verbose: int
+) -> list[str]:
+    if not has_default_eq(left):
+        return []
+    if isdatacls(left):
+        import dataclasses
+
+        all_fields = dataclasses.fields(left)
+        fields_to_check = [info.name for info in all_fields if info.compare]
+    elif isattrs(left):
+        all_fields = left.__attrs_attrs__
+        fields_to_check = [field.name for field in all_fields if getattr(field, "eq")]
+    elif isnamedtuple(left):
+        fields_to_check = left._fields
+    else:
+        assert False
+
+    indent = "  "
+    same = []
+    diff = []
+    for field in fields_to_check:
+        if getattr(left, field) == getattr(right, field):
+            same.append(field)
+        else:
+            diff.append(field)
+
+    explanation = []
+    if same or diff:
+        explanation += [""]
+    if same and verbose < 2:
+        explanation.append(f"Omitting {len(same)} identical items, use -vv to show")
+    elif same:
+        explanation += ["Matching attributes:"]
+        explanation += highlighter(pprint.pformat(same)).splitlines()
+    if diff:
+        explanation += ["Differing attributes:"]
+        explanation += highlighter(pprint.pformat(diff)).splitlines()
+        for field in diff:
+            field_left = getattr(left, field)
+            field_right = getattr(right, field)
+            explanation += [
+                "",
+                f"Drill down into differing attribute {field}:",
+                f"{indent}{field}: {highlighter(repr(field_left))} != {highlighter(repr(field_right))}",
+            ]
+            explanation += [
+                indent + line
+                for line in _compare_eq_any(
+                    field_left, field_right, highlighter, verbose
+                )
+            ]
+    return explanation
+
+
+def _notin_text(term: str, text: str, verbose: int = 0) -> list[str]:
+    index = text.find(term)
+    head = text[:index]
+    tail = text[index + len(term) :]
+    correct_text = head + tail
+    diff = _diff_text(text, correct_text, verbose)
+    newdiff = [f"{saferepr(term, maxsize=42)} is contained here:"]
+    for line in diff:
+        if line.startswith("Skipping"):
+            continue
+        if line.startswith("- "):
+            continue
+        if line.startswith("+ "):
+            newdiff.append("  " + line[2:])
+        else:
+            newdiff.append(line)
+    return newdiff
+
+
+def running_on_ci() -> bool:
+    """Check if we're currently running on a CI system."""
+    env_vars = ["CI", "BUILD_NUMBER"]
+    return any(var in os.environ for var in env_vars)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/cacheprovider.py b/backend/venv/lib/python3.12/site-packages/_pytest/cacheprovider.py
new file mode 100644
index 00000000..1b236efd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/cacheprovider.py
@@ -0,0 +1,626 @@
+# mypy: allow-untyped-defs
+"""Implementation of the cache provider."""
+
+# This plugin was not named "cache" to avoid conflicts with the external
+# pytest-cache version.
+from __future__ import annotations
+
+import dataclasses
+import errno
+import json
+import os
+from pathlib import Path
+import tempfile
+from typing import final
+from typing import Generator
+from typing import Iterable
+
+from .pathlib import resolve_from_str
+from .pathlib import rm_rf
+from .reports import CollectReport
+from _pytest import nodes
+from _pytest._io import TerminalWriter
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureRequest
+from _pytest.main import Session
+from _pytest.nodes import Directory
+from _pytest.nodes import File
+from _pytest.reports import TestReport
+
+
+README_CONTENT = """\
+# pytest cache directory #
+
+This directory contains data from the pytest's cache plugin,
+which provides the `--lf` and `--ff` options, as well as the `cache` fixture.
+
+**Do not** commit this to version control.
+
+See [the docs](https://docs.pytest.org/en/stable/how-to/cache.html) for more information.
+"""
+
+CACHEDIR_TAG_CONTENT = b"""\
+Signature: 8a477f597d28d172789f06886806bc55
+# This file is a cache directory tag created by pytest.
+# For information about cache directory tags, see:
+#	https://bford.info/cachedir/spec.html
+"""
+
+
+@final
+@dataclasses.dataclass
+class Cache:
+    """Instance of the `cache` fixture."""
+
+    _cachedir: Path = dataclasses.field(repr=False)
+    _config: Config = dataclasses.field(repr=False)
+
+    # Sub-directory under cache-dir for directories created by `mkdir()`.
+    _CACHE_PREFIX_DIRS = "d"
+
+    # Sub-directory under cache-dir for values created by `set()`.
+    _CACHE_PREFIX_VALUES = "v"
+
+    def __init__(
+        self, cachedir: Path, config: Config, *, _ispytest: bool = False
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._cachedir = cachedir
+        self._config = config
+
+    @classmethod
+    def for_config(cls, config: Config, *, _ispytest: bool = False) -> Cache:
+        """Create the Cache instance for a Config.
+
+        :meta private:
+        """
+        check_ispytest(_ispytest)
+        cachedir = cls.cache_dir_from_config(config, _ispytest=True)
+        if config.getoption("cacheclear") and cachedir.is_dir():
+            cls.clear_cache(cachedir, _ispytest=True)
+        return cls(cachedir, config, _ispytest=True)
+
+    @classmethod
+    def clear_cache(cls, cachedir: Path, _ispytest: bool = False) -> None:
+        """Clear the sub-directories used to hold cached directories and values.
+
+        :meta private:
+        """
+        check_ispytest(_ispytest)
+        for prefix in (cls._CACHE_PREFIX_DIRS, cls._CACHE_PREFIX_VALUES):
+            d = cachedir / prefix
+            if d.is_dir():
+                rm_rf(d)
+
+    @staticmethod
+    def cache_dir_from_config(config: Config, *, _ispytest: bool = False) -> Path:
+        """Get the path to the cache directory for a Config.
+
+        :meta private:
+        """
+        check_ispytest(_ispytest)
+        return resolve_from_str(config.getini("cache_dir"), config.rootpath)
+
+    def warn(self, fmt: str, *, _ispytest: bool = False, **args: object) -> None:
+        """Issue a cache warning.
+
+        :meta private:
+        """
+        check_ispytest(_ispytest)
+        import warnings
+
+        from _pytest.warning_types import PytestCacheWarning
+
+        warnings.warn(
+            PytestCacheWarning(fmt.format(**args) if args else fmt),
+            self._config.hook,
+            stacklevel=3,
+        )
+
+    def _mkdir(self, path: Path) -> None:
+        self._ensure_cache_dir_and_supporting_files()
+        path.mkdir(exist_ok=True, parents=True)
+
+    def mkdir(self, name: str) -> Path:
+        """Return a directory path object with the given name.
+
+        If the directory does not yet exist, it will be created. You can use
+        it to manage files to e.g. store/retrieve database dumps across test
+        sessions.
+
+        .. versionadded:: 7.0
+
+        :param name:
+            Must be a string not containing a ``/`` separator.
+            Make sure the name contains your plugin or application
+            identifiers to prevent clashes with other cache users.
+        """
+        path = Path(name)
+        if len(path.parts) > 1:
+            raise ValueError("name is not allowed to contain path separators")
+        res = self._cachedir.joinpath(self._CACHE_PREFIX_DIRS, path)
+        self._mkdir(res)
+        return res
+
+    def _getvaluepath(self, key: str) -> Path:
+        return self._cachedir.joinpath(self._CACHE_PREFIX_VALUES, Path(key))
+
+    def get(self, key: str, default):
+        """Return the cached value for the given key.
+
+        If no value was yet cached or the value cannot be read, the specified
+        default is returned.
+
+        :param key:
+            Must be a ``/`` separated value. Usually the first
+            name is the name of your plugin or your application.
+        :param default:
+            The value to return in case of a cache-miss or invalid cache value.
+        """
+        path = self._getvaluepath(key)
+        try:
+            with path.open("r", encoding="UTF-8") as f:
+                return json.load(f)
+        except (ValueError, OSError):
+            return default
+
+    def set(self, key: str, value: object) -> None:
+        """Save value for the given key.
+
+        :param key:
+            Must be a ``/`` separated value. Usually the first
+            name is the name of your plugin or your application.
+        :param value:
+            Must be of any combination of basic python types,
+            including nested types like lists of dictionaries.
+        """
+        path = self._getvaluepath(key)
+        try:
+            self._mkdir(path.parent)
+        except OSError as exc:
+            self.warn(
+                f"could not create cache path {path}: {exc}",
+                _ispytest=True,
+            )
+            return
+        data = json.dumps(value, ensure_ascii=False, indent=2)
+        try:
+            f = path.open("w", encoding="UTF-8")
+        except OSError as exc:
+            self.warn(
+                f"cache could not write path {path}: {exc}",
+                _ispytest=True,
+            )
+        else:
+            with f:
+                f.write(data)
+
+    def _ensure_cache_dir_and_supporting_files(self) -> None:
+        """Create the cache dir and its supporting files."""
+        if self._cachedir.is_dir():
+            return
+
+        self._cachedir.parent.mkdir(parents=True, exist_ok=True)
+        with tempfile.TemporaryDirectory(
+            prefix="pytest-cache-files-",
+            dir=self._cachedir.parent,
+        ) as newpath:
+            path = Path(newpath)
+
+            # Reset permissions to the default, see #12308.
+            # Note: there's no way to get the current umask atomically, eek.
+            umask = os.umask(0o022)
+            os.umask(umask)
+            path.chmod(0o777 - umask)
+
+            with open(path.joinpath("README.md"), "x", encoding="UTF-8") as f:
+                f.write(README_CONTENT)
+            with open(path.joinpath(".gitignore"), "x", encoding="UTF-8") as f:
+                f.write("# Created by pytest automatically.\n*\n")
+            with open(path.joinpath("CACHEDIR.TAG"), "xb") as f:
+                f.write(CACHEDIR_TAG_CONTENT)
+
+            try:
+                path.rename(self._cachedir)
+            except OSError as e:
+                # If 2 concurrent pytests both race to the rename, the loser
+                # gets "Directory not empty" from the rename. In this case,
+                # everything is handled so just continue (while letting the
+                # temporary directory be cleaned up).
+                # On Windows, the error is a FileExistsError which translates to EEXIST.
+                if e.errno not in (errno.ENOTEMPTY, errno.EEXIST):
+                    raise
+            else:
+                # Create a directory in place of the one we just moved so that
+                # `TemporaryDirectory`'s cleanup doesn't complain.
+                #
+                # TODO: pass ignore_cleanup_errors=True when we no longer support python < 3.10.
+                # See https://github.com/python/cpython/issues/74168. Note that passing
+                # delete=False would do the wrong thing in case of errors and isn't supported
+                # until python 3.12.
+                path.mkdir()
+
+
+class LFPluginCollWrapper:
+    def __init__(self, lfplugin: LFPlugin) -> None:
+        self.lfplugin = lfplugin
+        self._collected_at_least_one_failure = False
+
+    @hookimpl(wrapper=True)
+    def pytest_make_collect_report(
+        self, collector: nodes.Collector
+    ) -> Generator[None, CollectReport, CollectReport]:
+        res = yield
+        if isinstance(collector, (Session, Directory)):
+            # Sort any lf-paths to the beginning.
+            lf_paths = self.lfplugin._last_failed_paths
+
+            # Use stable sort to prioritize last failed.
+            def sort_key(node: nodes.Item | nodes.Collector) -> bool:
+                return node.path in lf_paths
+
+            res.result = sorted(
+                res.result,
+                key=sort_key,
+                reverse=True,
+            )
+
+        elif isinstance(collector, File):
+            if collector.path in self.lfplugin._last_failed_paths:
+                result = res.result
+                lastfailed = self.lfplugin.lastfailed
+
+                # Only filter with known failures.
+                if not self._collected_at_least_one_failure:
+                    if not any(x.nodeid in lastfailed for x in result):
+                        return res
+                    self.lfplugin.config.pluginmanager.register(
+                        LFPluginCollSkipfiles(self.lfplugin), "lfplugin-collskip"
+                    )
+                    self._collected_at_least_one_failure = True
+
+                session = collector.session
+                result[:] = [
+                    x
+                    for x in result
+                    if x.nodeid in lastfailed
+                    # Include any passed arguments (not trivial to filter).
+                    or session.isinitpath(x.path)
+                    # Keep all sub-collectors.
+                    or isinstance(x, nodes.Collector)
+                ]
+
+        return res
+
+
+class LFPluginCollSkipfiles:
+    def __init__(self, lfplugin: LFPlugin) -> None:
+        self.lfplugin = lfplugin
+
+    @hookimpl
+    def pytest_make_collect_report(
+        self, collector: nodes.Collector
+    ) -> CollectReport | None:
+        if isinstance(collector, File):
+            if collector.path not in self.lfplugin._last_failed_paths:
+                self.lfplugin._skipped_files += 1
+
+                return CollectReport(
+                    collector.nodeid, "passed", longrepr=None, result=[]
+                )
+        return None
+
+
+class LFPlugin:
+    """Plugin which implements the --lf (run last-failing) option."""
+
+    def __init__(self, config: Config) -> None:
+        self.config = config
+        active_keys = "lf", "failedfirst"
+        self.active = any(config.getoption(key) for key in active_keys)
+        assert config.cache
+        self.lastfailed: dict[str, bool] = config.cache.get("cache/lastfailed", {})
+        self._previously_failed_count: int | None = None
+        self._report_status: str | None = None
+        self._skipped_files = 0  # count skipped files during collection due to --lf
+
+        if config.getoption("lf"):
+            self._last_failed_paths = self.get_last_failed_paths()
+            config.pluginmanager.register(
+                LFPluginCollWrapper(self), "lfplugin-collwrapper"
+            )
+
+    def get_last_failed_paths(self) -> set[Path]:
+        """Return a set with all Paths of the previously failed nodeids and
+        their parents."""
+        rootpath = self.config.rootpath
+        result = set()
+        for nodeid in self.lastfailed:
+            path = rootpath / nodeid.split("::")[0]
+            result.add(path)
+            result.update(path.parents)
+        return {x for x in result if x.exists()}
+
+    def pytest_report_collectionfinish(self) -> str | None:
+        if self.active and self.config.get_verbosity() >= 0:
+            return f"run-last-failure: {self._report_status}"
+        return None
+
+    def pytest_runtest_logreport(self, report: TestReport) -> None:
+        if (report.when == "call" and report.passed) or report.skipped:
+            self.lastfailed.pop(report.nodeid, None)
+        elif report.failed:
+            self.lastfailed[report.nodeid] = True
+
+    def pytest_collectreport(self, report: CollectReport) -> None:
+        passed = report.outcome in ("passed", "skipped")
+        if passed:
+            if report.nodeid in self.lastfailed:
+                self.lastfailed.pop(report.nodeid)
+                self.lastfailed.update((item.nodeid, True) for item in report.result)
+        else:
+            self.lastfailed[report.nodeid] = True
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_collection_modifyitems(
+        self, config: Config, items: list[nodes.Item]
+    ) -> Generator[None]:
+        res = yield
+
+        if not self.active:
+            return res
+
+        if self.lastfailed:
+            previously_failed = []
+            previously_passed = []
+            for item in items:
+                if item.nodeid in self.lastfailed:
+                    previously_failed.append(item)
+                else:
+                    previously_passed.append(item)
+            self._previously_failed_count = len(previously_failed)
+
+            if not previously_failed:
+                # Running a subset of all tests with recorded failures
+                # only outside of it.
+                self._report_status = "%d known failures not in selected tests" % (
+                    len(self.lastfailed),
+                )
+            else:
+                if self.config.getoption("lf"):
+                    items[:] = previously_failed
+                    config.hook.pytest_deselected(items=previously_passed)
+                else:  # --failedfirst
+                    items[:] = previously_failed + previously_passed
+
+                noun = "failure" if self._previously_failed_count == 1 else "failures"
+                suffix = " first" if self.config.getoption("failedfirst") else ""
+                self._report_status = (
+                    f"rerun previous {self._previously_failed_count} {noun}{suffix}"
+                )
+
+            if self._skipped_files > 0:
+                files_noun = "file" if self._skipped_files == 1 else "files"
+                self._report_status += f" (skipped {self._skipped_files} {files_noun})"
+        else:
+            self._report_status = "no previously failed tests, "
+            if self.config.getoption("last_failed_no_failures") == "none":
+                self._report_status += "deselecting all items."
+                config.hook.pytest_deselected(items=items[:])
+                items[:] = []
+            else:
+                self._report_status += "not deselecting items."
+
+        return res
+
+    def pytest_sessionfinish(self, session: Session) -> None:
+        config = self.config
+        if config.getoption("cacheshow") or hasattr(config, "workerinput"):
+            return
+
+        assert config.cache is not None
+        saved_lastfailed = config.cache.get("cache/lastfailed", {})
+        if saved_lastfailed != self.lastfailed:
+            config.cache.set("cache/lastfailed", self.lastfailed)
+
+
+class NFPlugin:
+    """Plugin which implements the --nf (run new-first) option."""
+
+    def __init__(self, config: Config) -> None:
+        self.config = config
+        self.active = config.option.newfirst
+        assert config.cache is not None
+        self.cached_nodeids = set(config.cache.get("cache/nodeids", []))
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_collection_modifyitems(self, items: list[nodes.Item]) -> Generator[None]:
+        res = yield
+
+        if self.active:
+            new_items: dict[str, nodes.Item] = {}
+            other_items: dict[str, nodes.Item] = {}
+            for item in items:
+                if item.nodeid not in self.cached_nodeids:
+                    new_items[item.nodeid] = item
+                else:
+                    other_items[item.nodeid] = item
+
+            items[:] = self._get_increasing_order(
+                new_items.values()
+            ) + self._get_increasing_order(other_items.values())
+            self.cached_nodeids.update(new_items)
+        else:
+            self.cached_nodeids.update(item.nodeid for item in items)
+
+        return res
+
+    def _get_increasing_order(self, items: Iterable[nodes.Item]) -> list[nodes.Item]:
+        return sorted(items, key=lambda item: item.path.stat().st_mtime, reverse=True)
+
+    def pytest_sessionfinish(self) -> None:
+        config = self.config
+        if config.getoption("cacheshow") or hasattr(config, "workerinput"):
+            return
+
+        if config.getoption("collectonly"):
+            return
+
+        assert config.cache is not None
+        config.cache.set("cache/nodeids", sorted(self.cached_nodeids))
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group.addoption(
+        "--lf",
+        "--last-failed",
+        action="store_true",
+        dest="lf",
+        help="Rerun only the tests that failed "
+        "at the last run (or all if none failed)",
+    )
+    group.addoption(
+        "--ff",
+        "--failed-first",
+        action="store_true",
+        dest="failedfirst",
+        help="Run all tests, but run the last failures first. "
+        "This may re-order tests and thus lead to "
+        "repeated fixture setup/teardown.",
+    )
+    group.addoption(
+        "--nf",
+        "--new-first",
+        action="store_true",
+        dest="newfirst",
+        help="Run tests from new files first, then the rest of the tests "
+        "sorted by file mtime",
+    )
+    group.addoption(
+        "--cache-show",
+        action="append",
+        nargs="?",
+        dest="cacheshow",
+        help=(
+            "Show cache contents, don't perform collection or tests. "
+            "Optional argument: glob (default: '*')."
+        ),
+    )
+    group.addoption(
+        "--cache-clear",
+        action="store_true",
+        dest="cacheclear",
+        help="Remove all cache contents at start of test run",
+    )
+    cache_dir_default = ".pytest_cache"
+    if "TOX_ENV_DIR" in os.environ:
+        cache_dir_default = os.path.join(os.environ["TOX_ENV_DIR"], cache_dir_default)
+    parser.addini("cache_dir", default=cache_dir_default, help="Cache directory path")
+    group.addoption(
+        "--lfnf",
+        "--last-failed-no-failures",
+        action="store",
+        dest="last_failed_no_failures",
+        choices=("all", "none"),
+        default="all",
+        help="With ``--lf``, determines whether to execute tests when there "
+        "are no previously (known) failures or when no "
+        "cached ``lastfailed`` data was found. "
+        "``all`` (the default) runs the full test suite again. "
+        "``none`` just emits a message about no known failures and exits successfully.",
+    )
+
+
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    if config.option.cacheshow and not config.option.help:
+        from _pytest.main import wrap_session
+
+        return wrap_session(config, cacheshow)
+    return None
+
+
+@hookimpl(tryfirst=True)
+def pytest_configure(config: Config) -> None:
+    config.cache = Cache.for_config(config, _ispytest=True)
+    config.pluginmanager.register(LFPlugin(config), "lfplugin")
+    config.pluginmanager.register(NFPlugin(config), "nfplugin")
+
+
+@fixture
+def cache(request: FixtureRequest) -> Cache:
+    """Return a cache object that can persist state between testing sessions.
+
+    cache.get(key, default)
+    cache.set(key, value)
+
+    Keys must be ``/`` separated strings, where the first part is usually the
+    name of your plugin or application to avoid clashes with other cache users.
+
+    Values can be any object handled by the json stdlib module.
+    """
+    assert request.config.cache is not None
+    return request.config.cache
+
+
+def pytest_report_header(config: Config) -> str | None:
+    """Display cachedir with --cache-show and if non-default."""
+    if config.option.verbose > 0 or config.getini("cache_dir") != ".pytest_cache":
+        assert config.cache is not None
+        cachedir = config.cache._cachedir
+        # TODO: evaluate generating upward relative paths
+        # starting with .., ../.. if sensible
+
+        try:
+            displaypath = cachedir.relative_to(config.rootpath)
+        except ValueError:
+            displaypath = cachedir
+        return f"cachedir: {displaypath}"
+    return None
+
+
+def cacheshow(config: Config, session: Session) -> int:
+    from pprint import pformat
+
+    assert config.cache is not None
+
+    tw = TerminalWriter()
+    tw.line("cachedir: " + str(config.cache._cachedir))
+    if not config.cache._cachedir.is_dir():
+        tw.line("cache is empty")
+        return 0
+
+    glob = config.option.cacheshow[0]
+    if glob is None:
+        glob = "*"
+
+    dummy = object()
+    basedir = config.cache._cachedir
+    vdir = basedir / Cache._CACHE_PREFIX_VALUES
+    tw.sep("-", f"cache values for {glob!r}")
+    for valpath in sorted(x for x in vdir.rglob(glob) if x.is_file()):
+        key = str(valpath.relative_to(vdir))
+        val = config.cache.get(key, dummy)
+        if val is dummy:
+            tw.line(f"{key} contains unreadable content, will be ignored")
+        else:
+            tw.line(f"{key} contains:")
+            for line in pformat(val).splitlines():
+                tw.line("  " + line)
+
+    ddir = basedir / Cache._CACHE_PREFIX_DIRS
+    if ddir.is_dir():
+        contents = sorted(ddir.rglob(glob))
+        tw.sep("-", f"cache directories for {glob!r}")
+        for p in contents:
+            # if p.is_dir():
+            #    print("%s/" % p.relative_to(basedir))
+            if p.is_file():
+                key = str(p.relative_to(basedir))
+                tw.line(f"{key} is a file of length {p.stat().st_size:d}")
+    return 0
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/capture.py b/backend/venv/lib/python3.12/site-packages/_pytest/capture.py
new file mode 100644
index 00000000..506c0b3d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/capture.py
@@ -0,0 +1,1087 @@
+# mypy: allow-untyped-defs
+"""Per-test stdout/stderr capturing mechanism."""
+
+from __future__ import annotations
+
+import abc
+import collections
+import contextlib
+import io
+from io import UnsupportedOperation
+import os
+import sys
+from tempfile import TemporaryFile
+from types import TracebackType
+from typing import Any
+from typing import AnyStr
+from typing import BinaryIO
+from typing import Final
+from typing import final
+from typing import Generator
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import Literal
+from typing import NamedTuple
+from typing import TextIO
+from typing import TYPE_CHECKING
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+from _pytest.config import Config
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import SubRequest
+from _pytest.nodes import Collector
+from _pytest.nodes import File
+from _pytest.nodes import Item
+from _pytest.reports import CollectReport
+
+
+_CaptureMethod = Literal["fd", "sys", "no", "tee-sys"]
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group._addoption(
+        "--capture",
+        action="store",
+        default="fd",
+        metavar="method",
+        choices=["fd", "sys", "no", "tee-sys"],
+        help="Per-test capturing method: one of fd|sys|no|tee-sys",
+    )
+    group._addoption(
+        "-s",
+        action="store_const",
+        const="no",
+        dest="capture",
+        help="Shortcut for --capture=no",
+    )
+
+
+def _colorama_workaround() -> None:
+    """Ensure colorama is imported so that it attaches to the correct stdio
+    handles on Windows.
+
+    colorama uses the terminal on import time. So if something does the
+    first import of colorama while I/O capture is active, colorama will
+    fail in various ways.
+    """
+    if sys.platform.startswith("win32"):
+        try:
+            import colorama  # noqa: F401
+        except ImportError:
+            pass
+
+
+def _windowsconsoleio_workaround(stream: TextIO) -> None:
+    """Workaround for Windows Unicode console handling.
+
+    Python 3.6 implemented Unicode console handling for Windows. This works
+    by reading/writing to the raw console handle using
+    ``{Read,Write}ConsoleW``.
+
+    The problem is that we are going to ``dup2`` over the stdio file
+    descriptors when doing ``FDCapture`` and this will ``CloseHandle`` the
+    handles used by Python to write to the console. Though there is still some
+    weirdness and the console handle seems to only be closed randomly and not
+    on the first call to ``CloseHandle``, or maybe it gets reopened with the
+    same handle value when we suspend capturing.
+
+    The workaround in this case will reopen stdio with a different fd which
+    also means a different handle by replicating the logic in
+    "Py_lifecycle.c:initstdio/create_stdio".
+
+    :param stream:
+        In practice ``sys.stdout`` or ``sys.stderr``, but given
+        here as parameter for unittesting purposes.
+
+    See https://github.com/pytest-dev/py/issues/103.
+    """
+    if not sys.platform.startswith("win32") or hasattr(sys, "pypy_version_info"):
+        return
+
+    # Bail out if ``stream`` doesn't seem like a proper ``io`` stream (#2666).
+    if not hasattr(stream, "buffer"):  # type: ignore[unreachable,unused-ignore]
+        return
+
+    raw_stdout = stream.buffer.raw if hasattr(stream.buffer, "raw") else stream.buffer
+
+    if not isinstance(raw_stdout, io._WindowsConsoleIO):  # type: ignore[attr-defined,unused-ignore]
+        return
+
+    def _reopen_stdio(f, mode):
+        if not hasattr(stream.buffer, "raw") and mode[0] == "w":
+            buffering = 0
+        else:
+            buffering = -1
+
+        return io.TextIOWrapper(
+            open(os.dup(f.fileno()), mode, buffering),
+            f.encoding,
+            f.errors,
+            f.newlines,
+            f.line_buffering,
+        )
+
+    sys.stdin = _reopen_stdio(sys.stdin, "rb")
+    sys.stdout = _reopen_stdio(sys.stdout, "wb")
+    sys.stderr = _reopen_stdio(sys.stderr, "wb")
+
+
+@hookimpl(wrapper=True)
+def pytest_load_initial_conftests(early_config: Config) -> Generator[None]:
+    ns = early_config.known_args_namespace
+    if ns.capture == "fd":
+        _windowsconsoleio_workaround(sys.stdout)
+    _colorama_workaround()
+    pluginmanager = early_config.pluginmanager
+    capman = CaptureManager(ns.capture)
+    pluginmanager.register(capman, "capturemanager")
+
+    # Make sure that capturemanager is properly reset at final shutdown.
+    early_config.add_cleanup(capman.stop_global_capturing)
+
+    # Finally trigger conftest loading but while capturing (issue #93).
+    capman.start_global_capturing()
+    try:
+        try:
+            yield
+        finally:
+            capman.suspend_global_capture()
+    except BaseException:
+        out, err = capman.read_global_capture()
+        sys.stdout.write(out)
+        sys.stderr.write(err)
+        raise
+
+
+# IO Helpers.
+
+
+class EncodedFile(io.TextIOWrapper):
+    __slots__ = ()
+
+    @property
+    def name(self) -> str:
+        # Ensure that file.name is a string. Workaround for a Python bug
+        # fixed in >=3.7.4: https://bugs.python.org/issue36015
+        return repr(self.buffer)
+
+    @property
+    def mode(self) -> str:
+        # TextIOWrapper doesn't expose a mode, but at least some of our
+        # tests check it.
+        return self.buffer.mode.replace("b", "")
+
+
+class CaptureIO(io.TextIOWrapper):
+    def __init__(self) -> None:
+        super().__init__(io.BytesIO(), encoding="UTF-8", newline="", write_through=True)
+
+    def getvalue(self) -> str:
+        assert isinstance(self.buffer, io.BytesIO)
+        return self.buffer.getvalue().decode("UTF-8")
+
+
+class TeeCaptureIO(CaptureIO):
+    def __init__(self, other: TextIO) -> None:
+        self._other = other
+        super().__init__()
+
+    def write(self, s: str) -> int:
+        super().write(s)
+        return self._other.write(s)
+
+
+class DontReadFromInput(TextIO):
+    @property
+    def encoding(self) -> str:
+        assert sys.__stdin__ is not None
+        return sys.__stdin__.encoding
+
+    def read(self, size: int = -1) -> str:
+        raise OSError(
+            "pytest: reading from stdin while output is captured!  Consider using `-s`."
+        )
+
+    readline = read
+
+    def __next__(self) -> str:
+        return self.readline()
+
+    def readlines(self, hint: int | None = -1) -> list[str]:
+        raise OSError(
+            "pytest: reading from stdin while output is captured!  Consider using `-s`."
+        )
+
+    def __iter__(self) -> Iterator[str]:
+        return self
+
+    def fileno(self) -> int:
+        raise UnsupportedOperation("redirected stdin is pseudofile, has no fileno()")
+
+    def flush(self) -> None:
+        raise UnsupportedOperation("redirected stdin is pseudofile, has no flush()")
+
+    def isatty(self) -> bool:
+        return False
+
+    def close(self) -> None:
+        pass
+
+    def readable(self) -> bool:
+        return False
+
+    def seek(self, offset: int, whence: int = 0) -> int:
+        raise UnsupportedOperation("redirected stdin is pseudofile, has no seek(int)")
+
+    def seekable(self) -> bool:
+        return False
+
+    def tell(self) -> int:
+        raise UnsupportedOperation("redirected stdin is pseudofile, has no tell()")
+
+    def truncate(self, size: int | None = None) -> int:
+        raise UnsupportedOperation("cannot truncate stdin")
+
+    def write(self, data: str) -> int:
+        raise UnsupportedOperation("cannot write to stdin")
+
+    def writelines(self, lines: Iterable[str]) -> None:
+        raise UnsupportedOperation("Cannot write to stdin")
+
+    def writable(self) -> bool:
+        return False
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        type: type[BaseException] | None,
+        value: BaseException | None,
+        traceback: TracebackType | None,
+    ) -> None:
+        pass
+
+    @property
+    def buffer(self) -> BinaryIO:
+        # The str/bytes doesn't actually matter in this type, so OK to fake.
+        return self  # type: ignore[return-value]
+
+
+# Capture classes.
+
+
+class CaptureBase(abc.ABC, Generic[AnyStr]):
+    EMPTY_BUFFER: AnyStr
+
+    @abc.abstractmethod
+    def __init__(self, fd: int) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def start(self) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def done(self) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def suspend(self) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def resume(self) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def writeorg(self, data: AnyStr) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def snap(self) -> AnyStr:
+        raise NotImplementedError()
+
+
+patchsysdict = {0: "stdin", 1: "stdout", 2: "stderr"}
+
+
+class NoCapture(CaptureBase[str]):
+    EMPTY_BUFFER = ""
+
+    def __init__(self, fd: int) -> None:
+        pass
+
+    def start(self) -> None:
+        pass
+
+    def done(self) -> None:
+        pass
+
+    def suspend(self) -> None:
+        pass
+
+    def resume(self) -> None:
+        pass
+
+    def snap(self) -> str:
+        return ""
+
+    def writeorg(self, data: str) -> None:
+        pass
+
+
+class SysCaptureBase(CaptureBase[AnyStr]):
+    def __init__(
+        self, fd: int, tmpfile: TextIO | None = None, *, tee: bool = False
+    ) -> None:
+        name = patchsysdict[fd]
+        self._old: TextIO = getattr(sys, name)
+        self.name = name
+        if tmpfile is None:
+            if name == "stdin":
+                tmpfile = DontReadFromInput()
+            else:
+                tmpfile = CaptureIO() if not tee else TeeCaptureIO(self._old)
+        self.tmpfile = tmpfile
+        self._state = "initialized"
+
+    def repr(self, class_name: str) -> str:
+        return "<{} {} _old={} _state={!r} tmpfile={!r}>".format(
+            class_name,
+            self.name,
+            hasattr(self, "_old") and repr(self._old) or "<UNSET>",
+            self._state,
+            self.tmpfile,
+        )
+
+    def __repr__(self) -> str:
+        return "<{} {} _old={} _state={!r} tmpfile={!r}>".format(
+            self.__class__.__name__,
+            self.name,
+            hasattr(self, "_old") and repr(self._old) or "<UNSET>",
+            self._state,
+            self.tmpfile,
+        )
+
+    def _assert_state(self, op: str, states: tuple[str, ...]) -> None:
+        assert (
+            self._state in states
+        ), "cannot {} in state {!r}: expected one of {}".format(
+            op, self._state, ", ".join(states)
+        )
+
+    def start(self) -> None:
+        self._assert_state("start", ("initialized",))
+        setattr(sys, self.name, self.tmpfile)
+        self._state = "started"
+
+    def done(self) -> None:
+        self._assert_state("done", ("initialized", "started", "suspended", "done"))
+        if self._state == "done":
+            return
+        setattr(sys, self.name, self._old)
+        del self._old
+        self.tmpfile.close()
+        self._state = "done"
+
+    def suspend(self) -> None:
+        self._assert_state("suspend", ("started", "suspended"))
+        setattr(sys, self.name, self._old)
+        self._state = "suspended"
+
+    def resume(self) -> None:
+        self._assert_state("resume", ("started", "suspended"))
+        if self._state == "started":
+            return
+        setattr(sys, self.name, self.tmpfile)
+        self._state = "started"
+
+
+class SysCaptureBinary(SysCaptureBase[bytes]):
+    EMPTY_BUFFER = b""
+
+    def snap(self) -> bytes:
+        self._assert_state("snap", ("started", "suspended"))
+        self.tmpfile.seek(0)
+        res = self.tmpfile.buffer.read()
+        self.tmpfile.seek(0)
+        self.tmpfile.truncate()
+        return res
+
+    def writeorg(self, data: bytes) -> None:
+        self._assert_state("writeorg", ("started", "suspended"))
+        self._old.flush()
+        self._old.buffer.write(data)
+        self._old.buffer.flush()
+
+
+class SysCapture(SysCaptureBase[str]):
+    EMPTY_BUFFER = ""
+
+    def snap(self) -> str:
+        self._assert_state("snap", ("started", "suspended"))
+        assert isinstance(self.tmpfile, CaptureIO)
+        res = self.tmpfile.getvalue()
+        self.tmpfile.seek(0)
+        self.tmpfile.truncate()
+        return res
+
+    def writeorg(self, data: str) -> None:
+        self._assert_state("writeorg", ("started", "suspended"))
+        self._old.write(data)
+        self._old.flush()
+
+
+class FDCaptureBase(CaptureBase[AnyStr]):
+    def __init__(self, targetfd: int) -> None:
+        self.targetfd = targetfd
+
+        try:
+            os.fstat(targetfd)
+        except OSError:
+            # FD capturing is conceptually simple -- create a temporary file,
+            # redirect the FD to it, redirect back when done. But when the
+            # target FD is invalid it throws a wrench into this lovely scheme.
+            #
+            # Tests themselves shouldn't care if the FD is valid, FD capturing
+            # should work regardless of external circumstances. So falling back
+            # to just sys capturing is not a good option.
+            #
+            # Further complications are the need to support suspend() and the
+            # possibility of FD reuse (e.g. the tmpfile getting the very same
+            # target FD). The following approach is robust, I believe.
+            self.targetfd_invalid: int | None = os.open(os.devnull, os.O_RDWR)
+            os.dup2(self.targetfd_invalid, targetfd)
+        else:
+            self.targetfd_invalid = None
+        self.targetfd_save = os.dup(targetfd)
+
+        if targetfd == 0:
+            self.tmpfile = open(os.devnull, encoding="utf-8")
+            self.syscapture: CaptureBase[str] = SysCapture(targetfd)
+        else:
+            self.tmpfile = EncodedFile(
+                TemporaryFile(buffering=0),
+                encoding="utf-8",
+                errors="replace",
+                newline="",
+                write_through=True,
+            )
+            if targetfd in patchsysdict:
+                self.syscapture = SysCapture(targetfd, self.tmpfile)
+            else:
+                self.syscapture = NoCapture(targetfd)
+
+        self._state = "initialized"
+
+    def __repr__(self) -> str:
+        return (
+            f"<{self.__class__.__name__} {self.targetfd} oldfd={self.targetfd_save} "
+            f"_state={self._state!r} tmpfile={self.tmpfile!r}>"
+        )
+
+    def _assert_state(self, op: str, states: tuple[str, ...]) -> None:
+        assert (
+            self._state in states
+        ), "cannot {} in state {!r}: expected one of {}".format(
+            op, self._state, ", ".join(states)
+        )
+
+    def start(self) -> None:
+        """Start capturing on targetfd using memorized tmpfile."""
+        self._assert_state("start", ("initialized",))
+        os.dup2(self.tmpfile.fileno(), self.targetfd)
+        self.syscapture.start()
+        self._state = "started"
+
+    def done(self) -> None:
+        """Stop capturing, restore streams, return original capture file,
+        seeked to position zero."""
+        self._assert_state("done", ("initialized", "started", "suspended", "done"))
+        if self._state == "done":
+            return
+        os.dup2(self.targetfd_save, self.targetfd)
+        os.close(self.targetfd_save)
+        if self.targetfd_invalid is not None:
+            if self.targetfd_invalid != self.targetfd:
+                os.close(self.targetfd)
+            os.close(self.targetfd_invalid)
+        self.syscapture.done()
+        self.tmpfile.close()
+        self._state = "done"
+
+    def suspend(self) -> None:
+        self._assert_state("suspend", ("started", "suspended"))
+        if self._state == "suspended":
+            return
+        self.syscapture.suspend()
+        os.dup2(self.targetfd_save, self.targetfd)
+        self._state = "suspended"
+
+    def resume(self) -> None:
+        self._assert_state("resume", ("started", "suspended"))
+        if self._state == "started":
+            return
+        self.syscapture.resume()
+        os.dup2(self.tmpfile.fileno(), self.targetfd)
+        self._state = "started"
+
+
+class FDCaptureBinary(FDCaptureBase[bytes]):
+    """Capture IO to/from a given OS-level file descriptor.
+
+    snap() produces `bytes`.
+    """
+
+    EMPTY_BUFFER = b""
+
+    def snap(self) -> bytes:
+        self._assert_state("snap", ("started", "suspended"))
+        self.tmpfile.seek(0)
+        res = self.tmpfile.buffer.read()
+        self.tmpfile.seek(0)
+        self.tmpfile.truncate()
+        return res
+
+    def writeorg(self, data: bytes) -> None:
+        """Write to original file descriptor."""
+        self._assert_state("writeorg", ("started", "suspended"))
+        os.write(self.targetfd_save, data)
+
+
+class FDCapture(FDCaptureBase[str]):
+    """Capture IO to/from a given OS-level file descriptor.
+
+    snap() produces text.
+    """
+
+    EMPTY_BUFFER = ""
+
+    def snap(self) -> str:
+        self._assert_state("snap", ("started", "suspended"))
+        self.tmpfile.seek(0)
+        res = self.tmpfile.read()
+        self.tmpfile.seek(0)
+        self.tmpfile.truncate()
+        return res
+
+    def writeorg(self, data: str) -> None:
+        """Write to original file descriptor."""
+        self._assert_state("writeorg", ("started", "suspended"))
+        # XXX use encoding of original stream
+        os.write(self.targetfd_save, data.encode("utf-8"))
+
+
+# MultiCapture
+
+
+# Generic NamedTuple only supported since Python 3.11.
+if sys.version_info >= (3, 11) or TYPE_CHECKING:
+
+    @final
+    class CaptureResult(NamedTuple, Generic[AnyStr]):
+        """The result of :method:`caplog.readouterr() <pytest.CaptureFixture.readouterr>`."""
+
+        out: AnyStr
+        err: AnyStr
+
+else:
+
+    class CaptureResult(
+        collections.namedtuple("CaptureResult", ["out", "err"]),  # noqa: PYI024
+        Generic[AnyStr],
+    ):
+        """The result of :method:`caplog.readouterr() <pytest.CaptureFixture.readouterr>`."""
+
+        __slots__ = ()
+
+
+class MultiCapture(Generic[AnyStr]):
+    _state = None
+    _in_suspended = False
+
+    def __init__(
+        self,
+        in_: CaptureBase[AnyStr] | None,
+        out: CaptureBase[AnyStr] | None,
+        err: CaptureBase[AnyStr] | None,
+    ) -> None:
+        self.in_: CaptureBase[AnyStr] | None = in_
+        self.out: CaptureBase[AnyStr] | None = out
+        self.err: CaptureBase[AnyStr] | None = err
+
+    def __repr__(self) -> str:
+        return (
+            f"<MultiCapture out={self.out!r} err={self.err!r} in_={self.in_!r} "
+            f"_state={self._state!r} _in_suspended={self._in_suspended!r}>"
+        )
+
+    def start_capturing(self) -> None:
+        self._state = "started"
+        if self.in_:
+            self.in_.start()
+        if self.out:
+            self.out.start()
+        if self.err:
+            self.err.start()
+
+    def pop_outerr_to_orig(self) -> tuple[AnyStr, AnyStr]:
+        """Pop current snapshot out/err capture and flush to orig streams."""
+        out, err = self.readouterr()
+        if out:
+            assert self.out is not None
+            self.out.writeorg(out)
+        if err:
+            assert self.err is not None
+            self.err.writeorg(err)
+        return out, err
+
+    def suspend_capturing(self, in_: bool = False) -> None:
+        self._state = "suspended"
+        if self.out:
+            self.out.suspend()
+        if self.err:
+            self.err.suspend()
+        if in_ and self.in_:
+            self.in_.suspend()
+            self._in_suspended = True
+
+    def resume_capturing(self) -> None:
+        self._state = "started"
+        if self.out:
+            self.out.resume()
+        if self.err:
+            self.err.resume()
+        if self._in_suspended:
+            assert self.in_ is not None
+            self.in_.resume()
+            self._in_suspended = False
+
+    def stop_capturing(self) -> None:
+        """Stop capturing and reset capturing streams."""
+        if self._state == "stopped":
+            raise ValueError("was already stopped")
+        self._state = "stopped"
+        if self.out:
+            self.out.done()
+        if self.err:
+            self.err.done()
+        if self.in_:
+            self.in_.done()
+
+    def is_started(self) -> bool:
+        """Whether actively capturing -- not suspended or stopped."""
+        return self._state == "started"
+
+    def readouterr(self) -> CaptureResult[AnyStr]:
+        out = self.out.snap() if self.out else ""
+        err = self.err.snap() if self.err else ""
+        # TODO: This type error is real, need to fix.
+        return CaptureResult(out, err)  # type: ignore[arg-type]
+
+
+def _get_multicapture(method: _CaptureMethod) -> MultiCapture[str]:
+    if method == "fd":
+        return MultiCapture(in_=FDCapture(0), out=FDCapture(1), err=FDCapture(2))
+    elif method == "sys":
+        return MultiCapture(in_=SysCapture(0), out=SysCapture(1), err=SysCapture(2))
+    elif method == "no":
+        return MultiCapture(in_=None, out=None, err=None)
+    elif method == "tee-sys":
+        return MultiCapture(
+            in_=None, out=SysCapture(1, tee=True), err=SysCapture(2, tee=True)
+        )
+    raise ValueError(f"unknown capturing method: {method!r}")
+
+
+# CaptureManager and CaptureFixture
+
+
+class CaptureManager:
+    """The capture plugin.
+
+    Manages that the appropriate capture method is enabled/disabled during
+    collection and each test phase (setup, call, teardown). After each of
+    those points, the captured output is obtained and attached to the
+    collection/runtest report.
+
+    There are two levels of capture:
+
+    * global: enabled by default and can be suppressed by the ``-s``
+      option. This is always enabled/disabled during collection and each test
+      phase.
+
+    * fixture: when a test function or one of its fixture depend on the
+      ``capsys`` or ``capfd`` fixtures. In this case special handling is
+      needed to ensure the fixtures take precedence over the global capture.
+    """
+
+    def __init__(self, method: _CaptureMethod) -> None:
+        self._method: Final = method
+        self._global_capturing: MultiCapture[str] | None = None
+        self._capture_fixture: CaptureFixture[Any] | None = None
+
+    def __repr__(self) -> str:
+        return (
+            f"<CaptureManager _method={self._method!r} _global_capturing={self._global_capturing!r} "
+            f"_capture_fixture={self._capture_fixture!r}>"
+        )
+
+    def is_capturing(self) -> str | bool:
+        if self.is_globally_capturing():
+            return "global"
+        if self._capture_fixture:
+            return f"fixture {self._capture_fixture.request.fixturename}"
+        return False
+
+    # Global capturing control
+
+    def is_globally_capturing(self) -> bool:
+        return self._method != "no"
+
+    def start_global_capturing(self) -> None:
+        assert self._global_capturing is None
+        self._global_capturing = _get_multicapture(self._method)
+        self._global_capturing.start_capturing()
+
+    def stop_global_capturing(self) -> None:
+        if self._global_capturing is not None:
+            self._global_capturing.pop_outerr_to_orig()
+            self._global_capturing.stop_capturing()
+            self._global_capturing = None
+
+    def resume_global_capture(self) -> None:
+        # During teardown of the python process, and on rare occasions, capture
+        # attributes can be `None` while trying to resume global capture.
+        if self._global_capturing is not None:
+            self._global_capturing.resume_capturing()
+
+    def suspend_global_capture(self, in_: bool = False) -> None:
+        if self._global_capturing is not None:
+            self._global_capturing.suspend_capturing(in_=in_)
+
+    def suspend(self, in_: bool = False) -> None:
+        # Need to undo local capsys-et-al if it exists before disabling global capture.
+        self.suspend_fixture()
+        self.suspend_global_capture(in_)
+
+    def resume(self) -> None:
+        self.resume_global_capture()
+        self.resume_fixture()
+
+    def read_global_capture(self) -> CaptureResult[str]:
+        assert self._global_capturing is not None
+        return self._global_capturing.readouterr()
+
+    # Fixture Control
+
+    def set_fixture(self, capture_fixture: CaptureFixture[Any]) -> None:
+        if self._capture_fixture:
+            current_fixture = self._capture_fixture.request.fixturename
+            requested_fixture = capture_fixture.request.fixturename
+            capture_fixture.request.raiseerror(
+                f"cannot use {requested_fixture} and {current_fixture} at the same time"
+            )
+        self._capture_fixture = capture_fixture
+
+    def unset_fixture(self) -> None:
+        self._capture_fixture = None
+
+    def activate_fixture(self) -> None:
+        """If the current item is using ``capsys`` or ``capfd``, activate
+        them so they take precedence over the global capture."""
+        if self._capture_fixture:
+            self._capture_fixture._start()
+
+    def deactivate_fixture(self) -> None:
+        """Deactivate the ``capsys`` or ``capfd`` fixture of this item, if any."""
+        if self._capture_fixture:
+            self._capture_fixture.close()
+
+    def suspend_fixture(self) -> None:
+        if self._capture_fixture:
+            self._capture_fixture._suspend()
+
+    def resume_fixture(self) -> None:
+        if self._capture_fixture:
+            self._capture_fixture._resume()
+
+    # Helper context managers
+
+    @contextlib.contextmanager
+    def global_and_fixture_disabled(self) -> Generator[None]:
+        """Context manager to temporarily disable global and current fixture capturing."""
+        do_fixture = self._capture_fixture and self._capture_fixture._is_started()
+        if do_fixture:
+            self.suspend_fixture()
+        do_global = self._global_capturing and self._global_capturing.is_started()
+        if do_global:
+            self.suspend_global_capture()
+        try:
+            yield
+        finally:
+            if do_global:
+                self.resume_global_capture()
+            if do_fixture:
+                self.resume_fixture()
+
+    @contextlib.contextmanager
+    def item_capture(self, when: str, item: Item) -> Generator[None]:
+        self.resume_global_capture()
+        self.activate_fixture()
+        try:
+            yield
+        finally:
+            self.deactivate_fixture()
+            self.suspend_global_capture(in_=False)
+
+            out, err = self.read_global_capture()
+            item.add_report_section(when, "stdout", out)
+            item.add_report_section(when, "stderr", err)
+
+    # Hooks
+
+    @hookimpl(wrapper=True)
+    def pytest_make_collect_report(
+        self, collector: Collector
+    ) -> Generator[None, CollectReport, CollectReport]:
+        if isinstance(collector, File):
+            self.resume_global_capture()
+            try:
+                rep = yield
+            finally:
+                self.suspend_global_capture()
+            out, err = self.read_global_capture()
+            if out:
+                rep.sections.append(("Captured stdout", out))
+            if err:
+                rep.sections.append(("Captured stderr", err))
+        else:
+            rep = yield
+        return rep
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_setup(self, item: Item) -> Generator[None]:
+        with self.item_capture("setup", item):
+            return (yield)
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_call(self, item: Item) -> Generator[None]:
+        with self.item_capture("call", item):
+            return (yield)
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_teardown(self, item: Item) -> Generator[None]:
+        with self.item_capture("teardown", item):
+            return (yield)
+
+    @hookimpl(tryfirst=True)
+    def pytest_keyboard_interrupt(self) -> None:
+        self.stop_global_capturing()
+
+    @hookimpl(tryfirst=True)
+    def pytest_internalerror(self) -> None:
+        self.stop_global_capturing()
+
+
+class CaptureFixture(Generic[AnyStr]):
+    """Object returned by the :fixture:`capsys`, :fixture:`capsysbinary`,
+    :fixture:`capfd` and :fixture:`capfdbinary` fixtures."""
+
+    def __init__(
+        self,
+        captureclass: type[CaptureBase[AnyStr]],
+        request: SubRequest,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self.captureclass: type[CaptureBase[AnyStr]] = captureclass
+        self.request = request
+        self._capture: MultiCapture[AnyStr] | None = None
+        self._captured_out: AnyStr = self.captureclass.EMPTY_BUFFER
+        self._captured_err: AnyStr = self.captureclass.EMPTY_BUFFER
+
+    def _start(self) -> None:
+        if self._capture is None:
+            self._capture = MultiCapture(
+                in_=None,
+                out=self.captureclass(1),
+                err=self.captureclass(2),
+            )
+            self._capture.start_capturing()
+
+    def close(self) -> None:
+        if self._capture is not None:
+            out, err = self._capture.pop_outerr_to_orig()
+            self._captured_out += out
+            self._captured_err += err
+            self._capture.stop_capturing()
+            self._capture = None
+
+    def readouterr(self) -> CaptureResult[AnyStr]:
+        """Read and return the captured output so far, resetting the internal
+        buffer.
+
+        :returns:
+            The captured content as a namedtuple with ``out`` and ``err``
+            string attributes.
+        """
+        captured_out, captured_err = self._captured_out, self._captured_err
+        if self._capture is not None:
+            out, err = self._capture.readouterr()
+            captured_out += out
+            captured_err += err
+        self._captured_out = self.captureclass.EMPTY_BUFFER
+        self._captured_err = self.captureclass.EMPTY_BUFFER
+        return CaptureResult(captured_out, captured_err)
+
+    def _suspend(self) -> None:
+        """Suspend this fixture's own capturing temporarily."""
+        if self._capture is not None:
+            self._capture.suspend_capturing()
+
+    def _resume(self) -> None:
+        """Resume this fixture's own capturing temporarily."""
+        if self._capture is not None:
+            self._capture.resume_capturing()
+
+    def _is_started(self) -> bool:
+        """Whether actively capturing -- not disabled or closed."""
+        if self._capture is not None:
+            return self._capture.is_started()
+        return False
+
+    @contextlib.contextmanager
+    def disabled(self) -> Generator[None]:
+        """Temporarily disable capturing while inside the ``with`` block."""
+        capmanager: CaptureManager = self.request.config.pluginmanager.getplugin(
+            "capturemanager"
+        )
+        with capmanager.global_and_fixture_disabled():
+            yield
+
+
+# The fixtures.
+
+
+@fixture
+def capsys(request: SubRequest) -> Generator[CaptureFixture[str]]:
+    r"""Enable text capturing of writes to ``sys.stdout`` and ``sys.stderr``.
+
+    The captured output is made available via ``capsys.readouterr()`` method
+    calls, which return a ``(out, err)`` namedtuple.
+    ``out`` and ``err`` will be ``text`` objects.
+
+    Returns an instance of :class:`CaptureFixture[str] <pytest.CaptureFixture>`.
+
+    Example:
+
+    .. code-block:: python
+
+        def test_output(capsys):
+            print("hello")
+            captured = capsys.readouterr()
+            assert captured.out == "hello\n"
+    """
+    capman: CaptureManager = request.config.pluginmanager.getplugin("capturemanager")
+    capture_fixture = CaptureFixture(SysCapture, request, _ispytest=True)
+    capman.set_fixture(capture_fixture)
+    capture_fixture._start()
+    yield capture_fixture
+    capture_fixture.close()
+    capman.unset_fixture()
+
+
+@fixture
+def capsysbinary(request: SubRequest) -> Generator[CaptureFixture[bytes]]:
+    r"""Enable bytes capturing of writes to ``sys.stdout`` and ``sys.stderr``.
+
+    The captured output is made available via ``capsysbinary.readouterr()``
+    method calls, which return a ``(out, err)`` namedtuple.
+    ``out`` and ``err`` will be ``bytes`` objects.
+
+    Returns an instance of :class:`CaptureFixture[bytes] <pytest.CaptureFixture>`.
+
+    Example:
+
+    .. code-block:: python
+
+        def test_output(capsysbinary):
+            print("hello")
+            captured = capsysbinary.readouterr()
+            assert captured.out == b"hello\n"
+    """
+    capman: CaptureManager = request.config.pluginmanager.getplugin("capturemanager")
+    capture_fixture = CaptureFixture(SysCaptureBinary, request, _ispytest=True)
+    capman.set_fixture(capture_fixture)
+    capture_fixture._start()
+    yield capture_fixture
+    capture_fixture.close()
+    capman.unset_fixture()
+
+
+@fixture
+def capfd(request: SubRequest) -> Generator[CaptureFixture[str]]:
+    r"""Enable text capturing of writes to file descriptors ``1`` and ``2``.
+
+    The captured output is made available via ``capfd.readouterr()`` method
+    calls, which return a ``(out, err)`` namedtuple.
+    ``out`` and ``err`` will be ``text`` objects.
+
+    Returns an instance of :class:`CaptureFixture[str] <pytest.CaptureFixture>`.
+
+    Example:
+
+    .. code-block:: python
+
+        def test_system_echo(capfd):
+            os.system('echo "hello"')
+            captured = capfd.readouterr()
+            assert captured.out == "hello\n"
+    """
+    capman: CaptureManager = request.config.pluginmanager.getplugin("capturemanager")
+    capture_fixture = CaptureFixture(FDCapture, request, _ispytest=True)
+    capman.set_fixture(capture_fixture)
+    capture_fixture._start()
+    yield capture_fixture
+    capture_fixture.close()
+    capman.unset_fixture()
+
+
+@fixture
+def capfdbinary(request: SubRequest) -> Generator[CaptureFixture[bytes]]:
+    r"""Enable bytes capturing of writes to file descriptors ``1`` and ``2``.
+
+    The captured output is made available via ``capfd.readouterr()`` method
+    calls, which return a ``(out, err)`` namedtuple.
+    ``out`` and ``err`` will be ``byte`` objects.
+
+    Returns an instance of :class:`CaptureFixture[bytes] <pytest.CaptureFixture>`.
+
+    Example:
+
+    .. code-block:: python
+
+        def test_system_echo(capfdbinary):
+            os.system('echo "hello"')
+            captured = capfdbinary.readouterr()
+            assert captured.out == b"hello\n"
+
+    """
+    capman: CaptureManager = request.config.pluginmanager.getplugin("capturemanager")
+    capture_fixture = CaptureFixture(FDCaptureBinary, request, _ispytest=True)
+    capman.set_fixture(capture_fixture)
+    capture_fixture._start()
+    yield capture_fixture
+    capture_fixture.close()
+    capman.unset_fixture()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/compat.py b/backend/venv/lib/python3.12/site-packages/_pytest/compat.py
new file mode 100644
index 00000000..614848e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/compat.py
@@ -0,0 +1,351 @@
+# mypy: allow-untyped-defs
+"""Python version compatibility code."""
+
+from __future__ import annotations
+
+import dataclasses
+import enum
+import functools
+import inspect
+from inspect import Parameter
+from inspect import signature
+import os
+from pathlib import Path
+import sys
+from typing import Any
+from typing import Callable
+from typing import Final
+from typing import NoReturn
+
+import py
+
+
+#: constant to prepare valuing pylib path replacements/lazy proxies later on
+#  intended for removal in pytest 8.0 or 9.0
+
+# fmt: off
+# intentional space to create a fake difference for the verification
+LEGACY_PATH = py.path. local
+# fmt: on
+
+
+def legacy_path(path: str | os.PathLike[str]) -> LEGACY_PATH:
+    """Internal wrapper to prepare lazy proxies for legacy_path instances"""
+    return LEGACY_PATH(path)
+
+
+# fmt: off
+# Singleton type for NOTSET, as described in:
+# https://www.python.org/dev/peps/pep-0484/#support-for-singleton-types-in-unions
+class NotSetType(enum.Enum):
+    token = 0
+NOTSET: Final = NotSetType.token
+# fmt: on
+
+
+def is_generator(func: object) -> bool:
+    genfunc = inspect.isgeneratorfunction(func)
+    return genfunc and not iscoroutinefunction(func)
+
+
+def iscoroutinefunction(func: object) -> bool:
+    """Return True if func is a coroutine function (a function defined with async
+    def syntax, and doesn't contain yield), or a function decorated with
+    @asyncio.coroutine.
+
+    Note: copied and modified from Python 3.5's builtin coroutines.py to avoid
+    importing asyncio directly, which in turns also initializes the "logging"
+    module as a side-effect (see issue #8).
+    """
+    return inspect.iscoroutinefunction(func) or getattr(func, "_is_coroutine", False)
+
+
+def is_async_function(func: object) -> bool:
+    """Return True if the given function seems to be an async function or
+    an async generator."""
+    return iscoroutinefunction(func) or inspect.isasyncgenfunction(func)
+
+
+def getlocation(function, curdir: str | os.PathLike[str] | None = None) -> str:
+    function = get_real_func(function)
+    fn = Path(inspect.getfile(function))
+    lineno = function.__code__.co_firstlineno
+    if curdir is not None:
+        try:
+            relfn = fn.relative_to(curdir)
+        except ValueError:
+            pass
+        else:
+            return "%s:%d" % (relfn, lineno + 1)
+    return "%s:%d" % (fn, lineno + 1)
+
+
+def num_mock_patch_args(function) -> int:
+    """Return number of arguments used up by mock arguments (if any)."""
+    patchings = getattr(function, "patchings", None)
+    if not patchings:
+        return 0
+
+    mock_sentinel = getattr(sys.modules.get("mock"), "DEFAULT", object())
+    ut_mock_sentinel = getattr(sys.modules.get("unittest.mock"), "DEFAULT", object())
+
+    return len(
+        [
+            p
+            for p in patchings
+            if not p.attribute_name
+            and (p.new is mock_sentinel or p.new is ut_mock_sentinel)
+        ]
+    )
+
+
+def getfuncargnames(
+    function: Callable[..., object],
+    *,
+    name: str = "",
+    cls: type | None = None,
+) -> tuple[str, ...]:
+    """Return the names of a function's mandatory arguments.
+
+    Should return the names of all function arguments that:
+    * Aren't bound to an instance or type as in instance or class methods.
+    * Don't have default values.
+    * Aren't bound with functools.partial.
+    * Aren't replaced with mocks.
+
+    The cls arguments indicate that the function should be treated as a bound
+    method even though it's not unless the function is a static method.
+
+    The name parameter should be the original name in which the function was collected.
+    """
+    # TODO(RonnyPfannschmidt): This function should be refactored when we
+    # revisit fixtures. The fixture mechanism should ask the node for
+    # the fixture names, and not try to obtain directly from the
+    # function object well after collection has occurred.
+
+    # The parameters attribute of a Signature object contains an
+    # ordered mapping of parameter names to Parameter instances.  This
+    # creates a tuple of the names of the parameters that don't have
+    # defaults.
+    try:
+        parameters = signature(function).parameters
+    except (ValueError, TypeError) as e:
+        from _pytest.outcomes import fail
+
+        fail(
+            f"Could not determine arguments of {function!r}: {e}",
+            pytrace=False,
+        )
+
+    arg_names = tuple(
+        p.name
+        for p in parameters.values()
+        if (
+            p.kind is Parameter.POSITIONAL_OR_KEYWORD
+            or p.kind is Parameter.KEYWORD_ONLY
+        )
+        and p.default is Parameter.empty
+    )
+    if not name:
+        name = function.__name__
+
+    # If this function should be treated as a bound method even though
+    # it's passed as an unbound method or function, remove the first
+    # parameter name.
+    if (
+        # Not using `getattr` because we don't want to resolve the staticmethod.
+        # Not using `cls.__dict__` because we want to check the entire MRO.
+        cls
+        and not isinstance(
+            inspect.getattr_static(cls, name, default=None), staticmethod
+        )
+    ):
+        arg_names = arg_names[1:]
+    # Remove any names that will be replaced with mocks.
+    if hasattr(function, "__wrapped__"):
+        arg_names = arg_names[num_mock_patch_args(function) :]
+    return arg_names
+
+
+def get_default_arg_names(function: Callable[..., Any]) -> tuple[str, ...]:
+    # Note: this code intentionally mirrors the code at the beginning of
+    # getfuncargnames, to get the arguments which were excluded from its result
+    # because they had default values.
+    return tuple(
+        p.name
+        for p in signature(function).parameters.values()
+        if p.kind in (Parameter.POSITIONAL_OR_KEYWORD, Parameter.KEYWORD_ONLY)
+        and p.default is not Parameter.empty
+    )
+
+
+_non_printable_ascii_translate_table = {
+    i: f"\\x{i:02x}" for i in range(128) if i not in range(32, 127)
+}
+_non_printable_ascii_translate_table.update(
+    {ord("\t"): "\\t", ord("\r"): "\\r", ord("\n"): "\\n"}
+)
+
+
+def ascii_escaped(val: bytes | str) -> str:
+    r"""If val is pure ASCII, return it as an str, otherwise, escape
+    bytes objects into a sequence of escaped bytes:
+
+    b'\xc3\xb4\xc5\xd6' -> r'\xc3\xb4\xc5\xd6'
+
+    and escapes strings into a sequence of escaped unicode ids, e.g.:
+
+    r'4\nV\U00043efa\x0eMXWB\x1e\u3028\u15fd\xcd\U0007d944'
+
+    Note:
+       The obvious "v.decode('unicode-escape')" will return
+       valid UTF-8 unicode if it finds them in bytes, but we
+       want to return escaped bytes for any byte, even if they match
+       a UTF-8 string.
+    """
+    if isinstance(val, bytes):
+        ret = val.decode("ascii", "backslashreplace")
+    else:
+        ret = val.encode("unicode_escape").decode("ascii")
+    return ret.translate(_non_printable_ascii_translate_table)
+
+
+@dataclasses.dataclass
+class _PytestWrapper:
+    """Dummy wrapper around a function object for internal use only.
+
+    Used to correctly unwrap the underlying function object when we are
+    creating fixtures, because we wrap the function object ourselves with a
+    decorator to issue warnings when the fixture function is called directly.
+    """
+
+    obj: Any
+
+
+def get_real_func(obj):
+    """Get the real function object of the (possibly) wrapped object by
+    functools.wraps or functools.partial."""
+    start_obj = obj
+    for i in range(100):
+        # __pytest_wrapped__ is set by @pytest.fixture when wrapping the fixture function
+        # to trigger a warning if it gets called directly instead of by pytest: we don't
+        # want to unwrap further than this otherwise we lose useful wrappings like @mock.patch (#3774)
+        new_obj = getattr(obj, "__pytest_wrapped__", None)
+        if isinstance(new_obj, _PytestWrapper):
+            obj = new_obj.obj
+            break
+        new_obj = getattr(obj, "__wrapped__", None)
+        if new_obj is None:
+            break
+        obj = new_obj
+    else:
+        from _pytest._io.saferepr import saferepr
+
+        raise ValueError(
+            f"could not find real function of {saferepr(start_obj)}\nstopped at {saferepr(obj)}"
+        )
+    if isinstance(obj, functools.partial):
+        obj = obj.func
+    return obj
+
+
+def get_real_method(obj, holder):
+    """Attempt to obtain the real function object that might be wrapping
+    ``obj``, while at the same time returning a bound method to ``holder`` if
+    the original object was a bound method."""
+    try:
+        is_method = hasattr(obj, "__func__")
+        obj = get_real_func(obj)
+    except Exception:  # pragma: no cover
+        return obj
+    if is_method and hasattr(obj, "__get__") and callable(obj.__get__):
+        obj = obj.__get__(holder)
+    return obj
+
+
+def getimfunc(func):
+    try:
+        return func.__func__
+    except AttributeError:
+        return func
+
+
+def safe_getattr(object: Any, name: str, default: Any) -> Any:
+    """Like getattr but return default upon any Exception or any OutcomeException.
+
+    Attribute access can potentially fail for 'evil' Python objects.
+    See issue #214.
+    It catches OutcomeException because of #2490 (issue #580), new outcomes
+    are derived from BaseException instead of Exception (for more details
+    check #2707).
+    """
+    from _pytest.outcomes import TEST_OUTCOME
+
+    try:
+        return getattr(object, name, default)
+    except TEST_OUTCOME:
+        return default
+
+
+def safe_isclass(obj: object) -> bool:
+    """Ignore any exception via isinstance on Python 3."""
+    try:
+        return inspect.isclass(obj)
+    except Exception:
+        return False
+
+
+def get_user_id() -> int | None:
+    """Return the current process's real user id or None if it could not be
+    determined.
+
+    :return: The user id or None if it could not be determined.
+    """
+    # mypy follows the version and platform checking expectation of PEP 484:
+    # https://mypy.readthedocs.io/en/stable/common_issues.html?highlight=platform#python-version-and-system-platform-checks
+    # Containment checks are too complex for mypy v1.5.0 and cause failure.
+    if sys.platform == "win32" or sys.platform == "emscripten":
+        # win32 does not have a getuid() function.
+        # Emscripten has a return 0 stub.
+        return None
+    else:
+        # On other platforms, a return value of -1 is assumed to indicate that
+        # the current process's real user id could not be determined.
+        ERROR = -1
+        uid = os.getuid()
+        return uid if uid != ERROR else None
+
+
+# Perform exhaustiveness checking.
+#
+# Consider this example:
+#
+#     MyUnion = Union[int, str]
+#
+#     def handle(x: MyUnion) -> int {
+#         if isinstance(x, int):
+#             return 1
+#         elif isinstance(x, str):
+#             return 2
+#         else:
+#             raise Exception('unreachable')
+#
+# Now suppose we add a new variant:
+#
+#     MyUnion = Union[int, str, bytes]
+#
+# After doing this, we must remember ourselves to go and update the handle
+# function to handle the new variant.
+#
+# With `assert_never` we can do better:
+#
+#     // raise Exception('unreachable')
+#     return assert_never(x)
+#
+# Now, if we forget to handle the new variant, the type-checker will emit a
+# compile-time error, instead of the runtime error we would have gotten
+# previously.
+#
+# This also work for Enums (if you use `is` to compare) and Literals.
+def assert_never(value: NoReturn) -> NoReturn:
+    assert False, f"Unhandled value: {value} ({type(value).__name__})"
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/config/__init__.py
new file mode 100644
index 00000000..710e03e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/config/__init__.py
@@ -0,0 +1,1973 @@
+# mypy: allow-untyped-defs
+"""Command line options, ini-file and conftest.py processing."""
+
+from __future__ import annotations
+
+import argparse
+import collections.abc
+import copy
+import dataclasses
+import enum
+from functools import lru_cache
+import glob
+import importlib.metadata
+import inspect
+import os
+import pathlib
+import re
+import shlex
+import sys
+from textwrap import dedent
+import types
+from types import FunctionType
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Final
+from typing import final
+from typing import Generator
+from typing import IO
+from typing import Iterable
+from typing import Iterator
+from typing import Sequence
+from typing import TextIO
+from typing import Type
+from typing import TYPE_CHECKING
+import warnings
+
+import pluggy
+from pluggy import HookimplMarker
+from pluggy import HookimplOpts
+from pluggy import HookspecMarker
+from pluggy import HookspecOpts
+from pluggy import PluginManager
+
+from .compat import PathAwareHookProxy
+from .exceptions import PrintHelp as PrintHelp
+from .exceptions import UsageError as UsageError
+from .findpaths import determine_setup
+from _pytest import __version__
+import _pytest._code
+from _pytest._code import ExceptionInfo
+from _pytest._code import filter_traceback
+from _pytest._code.code import TracebackStyle
+from _pytest._io import TerminalWriter
+from _pytest.config.argparsing import Argument
+from _pytest.config.argparsing import Parser
+import _pytest.deprecated
+import _pytest.hookspec
+from _pytest.outcomes import fail
+from _pytest.outcomes import Skipped
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import bestrelpath
+from _pytest.pathlib import import_path
+from _pytest.pathlib import ImportMode
+from _pytest.pathlib import resolve_package_path
+from _pytest.pathlib import safe_exists
+from _pytest.stash import Stash
+from _pytest.warning_types import PytestConfigWarning
+from _pytest.warning_types import warn_explicit_for
+
+
+if TYPE_CHECKING:
+    from _pytest.cacheprovider import Cache
+    from _pytest.terminal import TerminalReporter
+
+
+_PluggyPlugin = object
+"""A type to represent plugin objects.
+
+Plugins can be any namespace, so we can't narrow it down much, but we use an
+alias to make the intent clear.
+
+Ideally this type would be provided by pluggy itself.
+"""
+
+
+hookimpl = HookimplMarker("pytest")
+hookspec = HookspecMarker("pytest")
+
+
+@final
+class ExitCode(enum.IntEnum):
+    """Encodes the valid exit codes by pytest.
+
+    Currently users and plugins may supply other exit codes as well.
+
+    .. versionadded:: 5.0
+    """
+
+    #: Tests passed.
+    OK = 0
+    #: Tests failed.
+    TESTS_FAILED = 1
+    #: pytest was interrupted.
+    INTERRUPTED = 2
+    #: An internal error got in the way.
+    INTERNAL_ERROR = 3
+    #: pytest was misused.
+    USAGE_ERROR = 4
+    #: pytest couldn't find tests.
+    NO_TESTS_COLLECTED = 5
+
+
+class ConftestImportFailure(Exception):
+    def __init__(
+        self,
+        path: pathlib.Path,
+        *,
+        cause: Exception,
+    ) -> None:
+        self.path = path
+        self.cause = cause
+
+    def __str__(self) -> str:
+        return f"{type(self.cause).__name__}: {self.cause} (from {self.path})"
+
+
+def filter_traceback_for_conftest_import_failure(
+    entry: _pytest._code.TracebackEntry,
+) -> bool:
+    """Filter tracebacks entries which point to pytest internals or importlib.
+
+    Make a special case for importlib because we use it to import test modules and conftest files
+    in _pytest.pathlib.import_path.
+    """
+    return filter_traceback(entry) and "importlib" not in str(entry.path).split(os.sep)
+
+
+def main(
+    args: list[str] | os.PathLike[str] | None = None,
+    plugins: Sequence[str | _PluggyPlugin] | None = None,
+) -> int | ExitCode:
+    """Perform an in-process test run.
+
+    :param args:
+        List of command line arguments. If `None` or not given, defaults to reading
+        arguments directly from the process command line (:data:`sys.argv`).
+    :param plugins: List of plugin objects to be auto-registered during initialization.
+
+    :returns: An exit code.
+    """
+    old_pytest_version = os.environ.get("PYTEST_VERSION")
+    try:
+        os.environ["PYTEST_VERSION"] = __version__
+        try:
+            config = _prepareconfig(args, plugins)
+        except ConftestImportFailure as e:
+            exc_info = ExceptionInfo.from_exception(e.cause)
+            tw = TerminalWriter(sys.stderr)
+            tw.line(f"ImportError while loading conftest '{e.path}'.", red=True)
+            exc_info.traceback = exc_info.traceback.filter(
+                filter_traceback_for_conftest_import_failure
+            )
+            exc_repr = (
+                exc_info.getrepr(style="short", chain=False)
+                if exc_info.traceback
+                else exc_info.exconly()
+            )
+            formatted_tb = str(exc_repr)
+            for line in formatted_tb.splitlines():
+                tw.line(line.rstrip(), red=True)
+            return ExitCode.USAGE_ERROR
+        else:
+            try:
+                ret: ExitCode | int = config.hook.pytest_cmdline_main(config=config)
+                try:
+                    return ExitCode(ret)
+                except ValueError:
+                    return ret
+            finally:
+                config._ensure_unconfigure()
+    except UsageError as e:
+        tw = TerminalWriter(sys.stderr)
+        for msg in e.args:
+            tw.line(f"ERROR: {msg}\n", red=True)
+        return ExitCode.USAGE_ERROR
+    finally:
+        if old_pytest_version is None:
+            os.environ.pop("PYTEST_VERSION", None)
+        else:
+            os.environ["PYTEST_VERSION"] = old_pytest_version
+
+
+def console_main() -> int:
+    """The CLI entry point of pytest.
+
+    This function is not meant for programmable use; use `main()` instead.
+    """
+    # https://docs.python.org/3/library/signal.html#note-on-sigpipe
+    try:
+        code = main()
+        sys.stdout.flush()
+        return code
+    except BrokenPipeError:
+        # Python flushes standard streams on exit; redirect remaining output
+        # to devnull to avoid another BrokenPipeError at shutdown
+        devnull = os.open(os.devnull, os.O_WRONLY)
+        os.dup2(devnull, sys.stdout.fileno())
+        return 1  # Python exits with error code 1 on EPIPE
+
+
+class cmdline:  # compatibility namespace
+    main = staticmethod(main)
+
+
+def filename_arg(path: str, optname: str) -> str:
+    """Argparse type validator for filename arguments.
+
+    :path: Path of filename.
+    :optname: Name of the option.
+    """
+    if os.path.isdir(path):
+        raise UsageError(f"{optname} must be a filename, given: {path}")
+    return path
+
+
+def directory_arg(path: str, optname: str) -> str:
+    """Argparse type validator for directory arguments.
+
+    :path: Path of directory.
+    :optname: Name of the option.
+    """
+    if not os.path.isdir(path):
+        raise UsageError(f"{optname} must be a directory, given: {path}")
+    return path
+
+
+# Plugins that cannot be disabled via "-p no:X" currently.
+essential_plugins = (
+    "mark",
+    "main",
+    "runner",
+    "fixtures",
+    "helpconfig",  # Provides -p.
+)
+
+default_plugins = (
+    *essential_plugins,
+    "python",
+    "terminal",
+    "debugging",
+    "unittest",
+    "capture",
+    "skipping",
+    "legacypath",
+    "tmpdir",
+    "monkeypatch",
+    "recwarn",
+    "pastebin",
+    "assertion",
+    "junitxml",
+    "doctest",
+    "cacheprovider",
+    "freeze_support",
+    "setuponly",
+    "setupplan",
+    "stepwise",
+    "warnings",
+    "logging",
+    "reports",
+    "python_path",
+    "unraisableexception",
+    "threadexception",
+    "faulthandler",
+)
+
+builtin_plugins = set(default_plugins)
+builtin_plugins.add("pytester")
+builtin_plugins.add("pytester_assertions")
+
+
+def get_config(
+    args: list[str] | None = None,
+    plugins: Sequence[str | _PluggyPlugin] | None = None,
+) -> Config:
+    # subsequent calls to main will create a fresh instance
+    pluginmanager = PytestPluginManager()
+    config = Config(
+        pluginmanager,
+        invocation_params=Config.InvocationParams(
+            args=args or (),
+            plugins=plugins,
+            dir=pathlib.Path.cwd(),
+        ),
+    )
+
+    if args is not None:
+        # Handle any "-p no:plugin" args.
+        pluginmanager.consider_preparse(args, exclude_only=True)
+
+    for spec in default_plugins:
+        pluginmanager.import_plugin(spec)
+
+    return config
+
+
+def get_plugin_manager() -> PytestPluginManager:
+    """Obtain a new instance of the
+    :py:class:`pytest.PytestPluginManager`, with default plugins
+    already loaded.
+
+    This function can be used by integration with other tools, like hooking
+    into pytest to run tests into an IDE.
+    """
+    return get_config().pluginmanager
+
+
+def _prepareconfig(
+    args: list[str] | os.PathLike[str] | None = None,
+    plugins: Sequence[str | _PluggyPlugin] | None = None,
+) -> Config:
+    if args is None:
+        args = sys.argv[1:]
+    elif isinstance(args, os.PathLike):
+        args = [os.fspath(args)]
+    elif not isinstance(args, list):
+        msg = (  # type:ignore[unreachable]
+            "`args` parameter expected to be a list of strings, got: {!r} (type: {})"
+        )
+        raise TypeError(msg.format(args, type(args)))
+
+    config = get_config(args, plugins)
+    pluginmanager = config.pluginmanager
+    try:
+        if plugins:
+            for plugin in plugins:
+                if isinstance(plugin, str):
+                    pluginmanager.consider_pluginarg(plugin)
+                else:
+                    pluginmanager.register(plugin)
+        config = pluginmanager.hook.pytest_cmdline_parse(
+            pluginmanager=pluginmanager, args=args
+        )
+        return config
+    except BaseException:
+        config._ensure_unconfigure()
+        raise
+
+
+def _get_directory(path: pathlib.Path) -> pathlib.Path:
+    """Get the directory of a path - itself if already a directory."""
+    if path.is_file():
+        return path.parent
+    else:
+        return path
+
+
+def _get_legacy_hook_marks(
+    method: Any,
+    hook_type: str,
+    opt_names: tuple[str, ...],
+) -> dict[str, bool]:
+    if TYPE_CHECKING:
+        # abuse typeguard from importlib to avoid massive method type union that's lacking an alias
+        assert inspect.isroutine(method)
+    known_marks: set[str] = {m.name for m in getattr(method, "pytestmark", [])}
+    must_warn: list[str] = []
+    opts: dict[str, bool] = {}
+    for opt_name in opt_names:
+        opt_attr = getattr(method, opt_name, AttributeError)
+        if opt_attr is not AttributeError:
+            must_warn.append(f"{opt_name}={opt_attr}")
+            opts[opt_name] = True
+        elif opt_name in known_marks:
+            must_warn.append(f"{opt_name}=True")
+            opts[opt_name] = True
+        else:
+            opts[opt_name] = False
+    if must_warn:
+        hook_opts = ", ".join(must_warn)
+        message = _pytest.deprecated.HOOK_LEGACY_MARKING.format(
+            type=hook_type,
+            fullname=method.__qualname__,
+            hook_opts=hook_opts,
+        )
+        warn_explicit_for(cast(FunctionType, method), message)
+    return opts
+
+
+@final
+class PytestPluginManager(PluginManager):
+    """A :py:class:`pluggy.PluginManager <pluggy.PluginManager>` with
+    additional pytest-specific functionality:
+
+    * Loading plugins from the command line, ``PYTEST_PLUGINS`` env variable and
+      ``pytest_plugins`` global variables found in plugins being loaded.
+    * ``conftest.py`` loading during start-up.
+    """
+
+    def __init__(self) -> None:
+        import _pytest.assertion
+
+        super().__init__("pytest")
+
+        # -- State related to local conftest plugins.
+        # All loaded conftest modules.
+        self._conftest_plugins: set[types.ModuleType] = set()
+        # All conftest modules applicable for a directory.
+        # This includes the directory's own conftest modules as well
+        # as those of its parent directories.
+        self._dirpath2confmods: dict[pathlib.Path, list[types.ModuleType]] = {}
+        # Cutoff directory above which conftests are no longer discovered.
+        self._confcutdir: pathlib.Path | None = None
+        # If set, conftest loading is skipped.
+        self._noconftest = False
+
+        # _getconftestmodules()'s call to _get_directory() causes a stat
+        # storm when it's called potentially thousands of times in a test
+        # session (#9478), often with the same path, so cache it.
+        self._get_directory = lru_cache(256)(_get_directory)
+
+        # plugins that were explicitly skipped with pytest.skip
+        # list of (module name, skip reason)
+        # previously we would issue a warning when a plugin was skipped, but
+        # since we refactored warnings as first citizens of Config, they are
+        # just stored here to be used later.
+        self.skipped_plugins: list[tuple[str, str]] = []
+
+        self.add_hookspecs(_pytest.hookspec)
+        self.register(self)
+        if os.environ.get("PYTEST_DEBUG"):
+            err: IO[str] = sys.stderr
+            encoding: str = getattr(err, "encoding", "utf8")
+            try:
+                err = open(
+                    os.dup(err.fileno()),
+                    mode=err.mode,
+                    buffering=1,
+                    encoding=encoding,
+                )
+            except Exception:
+                pass
+            self.trace.root.setwriter(err.write)
+            self.enable_tracing()
+
+        # Config._consider_importhook will set a real object if required.
+        self.rewrite_hook = _pytest.assertion.DummyRewriteHook()
+        # Used to know when we are importing conftests after the pytest_configure stage.
+        self._configured = False
+
+    def parse_hookimpl_opts(
+        self, plugin: _PluggyPlugin, name: str
+    ) -> HookimplOpts | None:
+        """:meta private:"""
+        # pytest hooks are always prefixed with "pytest_",
+        # so we avoid accessing possibly non-readable attributes
+        # (see issue #1073).
+        if not name.startswith("pytest_"):
+            return None
+        # Ignore names which cannot be hooks.
+        if name == "pytest_plugins":
+            return None
+
+        opts = super().parse_hookimpl_opts(plugin, name)
+        if opts is not None:
+            return opts
+
+        method = getattr(plugin, name)
+        # Consider only actual functions for hooks (#3775).
+        if not inspect.isroutine(method):
+            return None
+        # Collect unmarked hooks as long as they have the `pytest_' prefix.
+        return _get_legacy_hook_marks(  # type: ignore[return-value]
+            method, "impl", ("tryfirst", "trylast", "optionalhook", "hookwrapper")
+        )
+
+    def parse_hookspec_opts(self, module_or_class, name: str) -> HookspecOpts | None:
+        """:meta private:"""
+        opts = super().parse_hookspec_opts(module_or_class, name)
+        if opts is None:
+            method = getattr(module_or_class, name)
+            if name.startswith("pytest_"):
+                opts = _get_legacy_hook_marks(  # type: ignore[assignment]
+                    method,
+                    "spec",
+                    ("firstresult", "historic"),
+                )
+        return opts
+
+    def register(self, plugin: _PluggyPlugin, name: str | None = None) -> str | None:
+        if name in _pytest.deprecated.DEPRECATED_EXTERNAL_PLUGINS:
+            warnings.warn(
+                PytestConfigWarning(
+                    "{} plugin has been merged into the core, "
+                    "please remove it from your requirements.".format(
+                        name.replace("_", "-")
+                    )
+                )
+            )
+            return None
+        plugin_name = super().register(plugin, name)
+        if plugin_name is not None:
+            self.hook.pytest_plugin_registered.call_historic(
+                kwargs=dict(
+                    plugin=plugin,
+                    plugin_name=plugin_name,
+                    manager=self,
+                )
+            )
+
+            if isinstance(plugin, types.ModuleType):
+                self.consider_module(plugin)
+        return plugin_name
+
+    def getplugin(self, name: str):
+        # Support deprecated naming because plugins (xdist e.g.) use it.
+        plugin: _PluggyPlugin | None = self.get_plugin(name)
+        return plugin
+
+    def hasplugin(self, name: str) -> bool:
+        """Return whether a plugin with the given name is registered."""
+        return bool(self.get_plugin(name))
+
+    def pytest_configure(self, config: Config) -> None:
+        """:meta private:"""
+        # XXX now that the pluginmanager exposes hookimpl(tryfirst...)
+        # we should remove tryfirst/trylast as markers.
+        config.addinivalue_line(
+            "markers",
+            "tryfirst: mark a hook implementation function such that the "
+            "plugin machinery will try to call it first/as early as possible. "
+            "DEPRECATED, use @pytest.hookimpl(tryfirst=True) instead.",
+        )
+        config.addinivalue_line(
+            "markers",
+            "trylast: mark a hook implementation function such that the "
+            "plugin machinery will try to call it last/as late as possible. "
+            "DEPRECATED, use @pytest.hookimpl(trylast=True) instead.",
+        )
+        self._configured = True
+
+    #
+    # Internal API for local conftest plugin handling.
+    #
+    def _set_initial_conftests(
+        self,
+        args: Sequence[str | pathlib.Path],
+        pyargs: bool,
+        noconftest: bool,
+        rootpath: pathlib.Path,
+        confcutdir: pathlib.Path | None,
+        invocation_dir: pathlib.Path,
+        importmode: ImportMode | str,
+        *,
+        consider_namespace_packages: bool,
+    ) -> None:
+        """Load initial conftest files given a preparsed "namespace".
+
+        As conftest files may add their own command line options which have
+        arguments ('--my-opt somepath') we might get some false positives.
+        All builtin and 3rd party plugins will have been loaded, however, so
+        common options will not confuse our logic here.
+        """
+        self._confcutdir = (
+            absolutepath(invocation_dir / confcutdir) if confcutdir else None
+        )
+        self._noconftest = noconftest
+        self._using_pyargs = pyargs
+        foundanchor = False
+        for initial_path in args:
+            path = str(initial_path)
+            # remove node-id syntax
+            i = path.find("::")
+            if i != -1:
+                path = path[:i]
+            anchor = absolutepath(invocation_dir / path)
+
+            # Ensure we do not break if what appears to be an anchor
+            # is in fact a very long option (#10169, #11394).
+            if safe_exists(anchor):
+                self._try_load_conftest(
+                    anchor,
+                    importmode,
+                    rootpath,
+                    consider_namespace_packages=consider_namespace_packages,
+                )
+                foundanchor = True
+        if not foundanchor:
+            self._try_load_conftest(
+                invocation_dir,
+                importmode,
+                rootpath,
+                consider_namespace_packages=consider_namespace_packages,
+            )
+
+    def _is_in_confcutdir(self, path: pathlib.Path) -> bool:
+        """Whether to consider the given path to load conftests from."""
+        if self._confcutdir is None:
+            return True
+        # The semantics here are literally:
+        #   Do not load a conftest if it is found upwards from confcut dir.
+        # But this is *not* the same as:
+        #   Load only conftests from confcutdir or below.
+        # At first glance they might seem the same thing, however we do support use cases where
+        # we want to load conftests that are not found in confcutdir or below, but are found
+        # in completely different directory hierarchies like packages installed
+        # in out-of-source trees.
+        # (see #9767 for a regression where the logic was inverted).
+        return path not in self._confcutdir.parents
+
+    def _try_load_conftest(
+        self,
+        anchor: pathlib.Path,
+        importmode: str | ImportMode,
+        rootpath: pathlib.Path,
+        *,
+        consider_namespace_packages: bool,
+    ) -> None:
+        self._loadconftestmodules(
+            anchor,
+            importmode,
+            rootpath,
+            consider_namespace_packages=consider_namespace_packages,
+        )
+        # let's also consider test* subdirs
+        if anchor.is_dir():
+            for x in anchor.glob("test*"):
+                if x.is_dir():
+                    self._loadconftestmodules(
+                        x,
+                        importmode,
+                        rootpath,
+                        consider_namespace_packages=consider_namespace_packages,
+                    )
+
+    def _loadconftestmodules(
+        self,
+        path: pathlib.Path,
+        importmode: str | ImportMode,
+        rootpath: pathlib.Path,
+        *,
+        consider_namespace_packages: bool,
+    ) -> None:
+        if self._noconftest:
+            return
+
+        directory = self._get_directory(path)
+
+        # Optimization: avoid repeated searches in the same directory.
+        # Assumes always called with same importmode and rootpath.
+        if directory in self._dirpath2confmods:
+            return
+
+        clist = []
+        for parent in reversed((directory, *directory.parents)):
+            if self._is_in_confcutdir(parent):
+                conftestpath = parent / "conftest.py"
+                if conftestpath.is_file():
+                    mod = self._importconftest(
+                        conftestpath,
+                        importmode,
+                        rootpath,
+                        consider_namespace_packages=consider_namespace_packages,
+                    )
+                    clist.append(mod)
+        self._dirpath2confmods[directory] = clist
+
+    def _getconftestmodules(self, path: pathlib.Path) -> Sequence[types.ModuleType]:
+        directory = self._get_directory(path)
+        return self._dirpath2confmods.get(directory, ())
+
+    def _rget_with_confmod(
+        self,
+        name: str,
+        path: pathlib.Path,
+    ) -> tuple[types.ModuleType, Any]:
+        modules = self._getconftestmodules(path)
+        for mod in reversed(modules):
+            try:
+                return mod, getattr(mod, name)
+            except AttributeError:
+                continue
+        raise KeyError(name)
+
+    def _importconftest(
+        self,
+        conftestpath: pathlib.Path,
+        importmode: str | ImportMode,
+        rootpath: pathlib.Path,
+        *,
+        consider_namespace_packages: bool,
+    ) -> types.ModuleType:
+        conftestpath_plugin_name = str(conftestpath)
+        existing = self.get_plugin(conftestpath_plugin_name)
+        if existing is not None:
+            return cast(types.ModuleType, existing)
+
+        # conftest.py files there are not in a Python package all have module
+        # name "conftest", and thus conflict with each other. Clear the existing
+        # before loading the new one, otherwise the existing one will be
+        # returned from the module cache.
+        pkgpath = resolve_package_path(conftestpath)
+        if pkgpath is None:
+            try:
+                del sys.modules[conftestpath.stem]
+            except KeyError:
+                pass
+
+        try:
+            mod = import_path(
+                conftestpath,
+                mode=importmode,
+                root=rootpath,
+                consider_namespace_packages=consider_namespace_packages,
+            )
+        except Exception as e:
+            assert e.__traceback__ is not None
+            raise ConftestImportFailure(conftestpath, cause=e) from e
+
+        self._check_non_top_pytest_plugins(mod, conftestpath)
+
+        self._conftest_plugins.add(mod)
+        dirpath = conftestpath.parent
+        if dirpath in self._dirpath2confmods:
+            for path, mods in self._dirpath2confmods.items():
+                if dirpath in path.parents or path == dirpath:
+                    if mod in mods:
+                        raise AssertionError(
+                            f"While trying to load conftest path {conftestpath!s}, "
+                            f"found that the module {mod} is already loaded with path {mod.__file__}. "
+                            "This is not supposed to happen. Please report this issue to pytest."
+                        )
+                    mods.append(mod)
+        self.trace(f"loading conftestmodule {mod!r}")
+        self.consider_conftest(mod, registration_name=conftestpath_plugin_name)
+        return mod
+
+    def _check_non_top_pytest_plugins(
+        self,
+        mod: types.ModuleType,
+        conftestpath: pathlib.Path,
+    ) -> None:
+        if (
+            hasattr(mod, "pytest_plugins")
+            and self._configured
+            and not self._using_pyargs
+        ):
+            msg = (
+                "Defining 'pytest_plugins' in a non-top-level conftest is no longer supported:\n"
+                "It affects the entire test suite instead of just below the conftest as expected.\n"
+                "  {}\n"
+                "Please move it to a top level conftest file at the rootdir:\n"
+                "  {}\n"
+                "For more information, visit:\n"
+                "  https://docs.pytest.org/en/stable/deprecations.html#pytest-plugins-in-non-top-level-conftest-files"
+            )
+            fail(msg.format(conftestpath, self._confcutdir), pytrace=False)
+
+    #
+    # API for bootstrapping plugin loading
+    #
+    #
+
+    def consider_preparse(
+        self, args: Sequence[str], *, exclude_only: bool = False
+    ) -> None:
+        """:meta private:"""
+        i = 0
+        n = len(args)
+        while i < n:
+            opt = args[i]
+            i += 1
+            if isinstance(opt, str):
+                if opt == "-p":
+                    try:
+                        parg = args[i]
+                    except IndexError:
+                        return
+                    i += 1
+                elif opt.startswith("-p"):
+                    parg = opt[2:]
+                else:
+                    continue
+                parg = parg.strip()
+                if exclude_only and not parg.startswith("no:"):
+                    continue
+                self.consider_pluginarg(parg)
+
+    def consider_pluginarg(self, arg: str) -> None:
+        """:meta private:"""
+        if arg.startswith("no:"):
+            name = arg[3:]
+            if name in essential_plugins:
+                raise UsageError(f"plugin {name} cannot be disabled")
+
+            # PR #4304: remove stepwise if cacheprovider is blocked.
+            if name == "cacheprovider":
+                self.set_blocked("stepwise")
+                self.set_blocked("pytest_stepwise")
+
+            self.set_blocked(name)
+            if not name.startswith("pytest_"):
+                self.set_blocked("pytest_" + name)
+        else:
+            name = arg
+            # Unblock the plugin.
+            self.unblock(name)
+            if not name.startswith("pytest_"):
+                self.unblock("pytest_" + name)
+            self.import_plugin(arg, consider_entry_points=True)
+
+    def consider_conftest(
+        self, conftestmodule: types.ModuleType, registration_name: str
+    ) -> None:
+        """:meta private:"""
+        self.register(conftestmodule, name=registration_name)
+
+    def consider_env(self) -> None:
+        """:meta private:"""
+        self._import_plugin_specs(os.environ.get("PYTEST_PLUGINS"))
+
+    def consider_module(self, mod: types.ModuleType) -> None:
+        """:meta private:"""
+        self._import_plugin_specs(getattr(mod, "pytest_plugins", []))
+
+    def _import_plugin_specs(
+        self, spec: None | types.ModuleType | str | Sequence[str]
+    ) -> None:
+        plugins = _get_plugin_specs_as_list(spec)
+        for import_spec in plugins:
+            self.import_plugin(import_spec)
+
+    def import_plugin(self, modname: str, consider_entry_points: bool = False) -> None:
+        """Import a plugin with ``modname``.
+
+        If ``consider_entry_points`` is True, entry point names are also
+        considered to find a plugin.
+        """
+        # Most often modname refers to builtin modules, e.g. "pytester",
+        # "terminal" or "capture".  Those plugins are registered under their
+        # basename for historic purposes but must be imported with the
+        # _pytest prefix.
+        assert isinstance(
+            modname, str
+        ), f"module name as text required, got {modname!r}"
+        if self.is_blocked(modname) or self.get_plugin(modname) is not None:
+            return
+
+        importspec = "_pytest." + modname if modname in builtin_plugins else modname
+        self.rewrite_hook.mark_rewrite(importspec)
+
+        if consider_entry_points:
+            loaded = self.load_setuptools_entrypoints("pytest11", name=modname)
+            if loaded:
+                return
+
+        try:
+            __import__(importspec)
+        except ImportError as e:
+            raise ImportError(
+                f'Error importing plugin "{modname}": {e.args[0]}'
+            ).with_traceback(e.__traceback__) from e
+
+        except Skipped as e:
+            self.skipped_plugins.append((modname, e.msg or ""))
+        else:
+            mod = sys.modules[importspec]
+            self.register(mod, modname)
+
+
+def _get_plugin_specs_as_list(
+    specs: None | types.ModuleType | str | Sequence[str],
+) -> list[str]:
+    """Parse a plugins specification into a list of plugin names."""
+    # None means empty.
+    if specs is None:
+        return []
+    # Workaround for #3899 - a submodule which happens to be called "pytest_plugins".
+    if isinstance(specs, types.ModuleType):
+        return []
+    # Comma-separated list.
+    if isinstance(specs, str):
+        return specs.split(",") if specs else []
+    # Direct specification.
+    if isinstance(specs, collections.abc.Sequence):
+        return list(specs)
+    raise UsageError(
+        f"Plugins may be specified as a sequence or a ','-separated string of plugin names. Got: {specs!r}"
+    )
+
+
+class Notset:
+    def __repr__(self):
+        return "<NOTSET>"
+
+
+notset = Notset()
+
+
+def _iter_rewritable_modules(package_files: Iterable[str]) -> Iterator[str]:
+    """Given an iterable of file names in a source distribution, return the "names" that should
+    be marked for assertion rewrite.
+
+    For example the package "pytest_mock/__init__.py" should be added as "pytest_mock" in
+    the assertion rewrite mechanism.
+
+    This function has to deal with dist-info based distributions and egg based distributions
+    (which are still very much in use for "editable" installs).
+
+    Here are the file names as seen in a dist-info based distribution:
+
+        pytest_mock/__init__.py
+        pytest_mock/_version.py
+        pytest_mock/plugin.py
+        pytest_mock.egg-info/PKG-INFO
+
+    Here are the file names as seen in an egg based distribution:
+
+        src/pytest_mock/__init__.py
+        src/pytest_mock/_version.py
+        src/pytest_mock/plugin.py
+        src/pytest_mock.egg-info/PKG-INFO
+        LICENSE
+        setup.py
+
+    We have to take in account those two distribution flavors in order to determine which
+    names should be considered for assertion rewriting.
+
+    More information:
+        https://github.com/pytest-dev/pytest-mock/issues/167
+    """
+    package_files = list(package_files)
+    seen_some = False
+    for fn in package_files:
+        is_simple_module = "/" not in fn and fn.endswith(".py")
+        is_package = fn.count("/") == 1 and fn.endswith("__init__.py")
+        if is_simple_module:
+            module_name, _ = os.path.splitext(fn)
+            # we ignore "setup.py" at the root of the distribution
+            # as well as editable installation finder modules made by setuptools
+            if module_name != "setup" and not module_name.startswith("__editable__"):
+                seen_some = True
+                yield module_name
+        elif is_package:
+            package_name = os.path.dirname(fn)
+            seen_some = True
+            yield package_name
+
+    if not seen_some:
+        # At this point we did not find any packages or modules suitable for assertion
+        # rewriting, so we try again by stripping the first path component (to account for
+        # "src" based source trees for example).
+        # This approach lets us have the common case continue to be fast, as egg-distributions
+        # are rarer.
+        new_package_files = []
+        for fn in package_files:
+            parts = fn.split("/")
+            new_fn = "/".join(parts[1:])
+            if new_fn:
+                new_package_files.append(new_fn)
+        if new_package_files:
+            yield from _iter_rewritable_modules(new_package_files)
+
+
+@final
+class Config:
+    """Access to configuration values, pluginmanager and plugin hooks.
+
+    :param PytestPluginManager pluginmanager:
+        A pytest PluginManager.
+
+    :param InvocationParams invocation_params:
+        Object containing parameters regarding the :func:`pytest.main`
+        invocation.
+    """
+
+    @final
+    @dataclasses.dataclass(frozen=True)
+    class InvocationParams:
+        """Holds parameters passed during :func:`pytest.main`.
+
+        The object attributes are read-only.
+
+        .. versionadded:: 5.1
+
+        .. note::
+
+            Note that the environment variable ``PYTEST_ADDOPTS`` and the ``addopts``
+            ini option are handled by pytest, not being included in the ``args`` attribute.
+
+            Plugins accessing ``InvocationParams`` must be aware of that.
+        """
+
+        args: tuple[str, ...]
+        """The command-line arguments as passed to :func:`pytest.main`."""
+        plugins: Sequence[str | _PluggyPlugin] | None
+        """Extra plugins, might be `None`."""
+        dir: pathlib.Path
+        """The directory from which :func:`pytest.main` was invoked. :type: pathlib.Path"""
+
+        def __init__(
+            self,
+            *,
+            args: Iterable[str],
+            plugins: Sequence[str | _PluggyPlugin] | None,
+            dir: pathlib.Path,
+        ) -> None:
+            object.__setattr__(self, "args", tuple(args))
+            object.__setattr__(self, "plugins", plugins)
+            object.__setattr__(self, "dir", dir)
+
+    class ArgsSource(enum.Enum):
+        """Indicates the source of the test arguments.
+
+        .. versionadded:: 7.2
+        """
+
+        #: Command line arguments.
+        ARGS = enum.auto()
+        #: Invocation directory.
+        INVOCATION_DIR = enum.auto()
+        INCOVATION_DIR = INVOCATION_DIR  # backwards compatibility alias
+        #: 'testpaths' configuration value.
+        TESTPATHS = enum.auto()
+
+    # Set by cacheprovider plugin.
+    cache: Cache
+
+    def __init__(
+        self,
+        pluginmanager: PytestPluginManager,
+        *,
+        invocation_params: InvocationParams | None = None,
+    ) -> None:
+        from .argparsing import FILE_OR_DIR
+        from .argparsing import Parser
+
+        if invocation_params is None:
+            invocation_params = self.InvocationParams(
+                args=(), plugins=None, dir=pathlib.Path.cwd()
+            )
+
+        self.option = argparse.Namespace()
+        """Access to command line option as attributes.
+
+        :type: argparse.Namespace
+        """
+
+        self.invocation_params = invocation_params
+        """The parameters with which pytest was invoked.
+
+        :type: InvocationParams
+        """
+
+        _a = FILE_OR_DIR
+        self._parser = Parser(
+            usage=f"%(prog)s [options] [{_a}] [{_a}] [...]",
+            processopt=self._processopt,
+            _ispytest=True,
+        )
+        self.pluginmanager = pluginmanager
+        """The plugin manager handles plugin registration and hook invocation.
+
+        :type: PytestPluginManager
+        """
+
+        self.stash = Stash()
+        """A place where plugins can store information on the config for their
+        own use.
+
+        :type: Stash
+        """
+        # Deprecated alias. Was never public. Can be removed in a few releases.
+        self._store = self.stash
+
+        self.trace = self.pluginmanager.trace.root.get("config")
+        self.hook: pluggy.HookRelay = PathAwareHookProxy(self.pluginmanager.hook)  # type: ignore[assignment]
+        self._inicache: dict[str, Any] = {}
+        self._override_ini: Sequence[str] = ()
+        self._opt2dest: dict[str, str] = {}
+        self._cleanup: list[Callable[[], None]] = []
+        self.pluginmanager.register(self, "pytestconfig")
+        self._configured = False
+        self.hook.pytest_addoption.call_historic(
+            kwargs=dict(parser=self._parser, pluginmanager=self.pluginmanager)
+        )
+        self.args_source = Config.ArgsSource.ARGS
+        self.args: list[str] = []
+
+    @property
+    def rootpath(self) -> pathlib.Path:
+        """The path to the :ref:`rootdir <rootdir>`.
+
+        :type: pathlib.Path
+
+        .. versionadded:: 6.1
+        """
+        return self._rootpath
+
+    @property
+    def inipath(self) -> pathlib.Path | None:
+        """The path to the :ref:`configfile <configfiles>`.
+
+        .. versionadded:: 6.1
+        """
+        return self._inipath
+
+    def add_cleanup(self, func: Callable[[], None]) -> None:
+        """Add a function to be called when the config object gets out of
+        use (usually coinciding with pytest_unconfigure)."""
+        self._cleanup.append(func)
+
+    def _do_configure(self) -> None:
+        assert not self._configured
+        self._configured = True
+        with warnings.catch_warnings():
+            warnings.simplefilter("default")
+            self.hook.pytest_configure.call_historic(kwargs=dict(config=self))
+
+    def _ensure_unconfigure(self) -> None:
+        if self._configured:
+            self._configured = False
+            self.hook.pytest_unconfigure(config=self)
+            self.hook.pytest_configure._call_history = []
+        while self._cleanup:
+            fin = self._cleanup.pop()
+            fin()
+
+    def get_terminal_writer(self) -> TerminalWriter:
+        terminalreporter: TerminalReporter | None = self.pluginmanager.get_plugin(
+            "terminalreporter"
+        )
+        assert terminalreporter is not None
+        return terminalreporter._tw
+
+    def pytest_cmdline_parse(
+        self, pluginmanager: PytestPluginManager, args: list[str]
+    ) -> Config:
+        try:
+            self.parse(args)
+        except UsageError:
+            # Handle --version and --help here in a minimal fashion.
+            # This gets done via helpconfig normally, but its
+            # pytest_cmdline_main is not called in case of errors.
+            if getattr(self.option, "version", False) or "--version" in args:
+                from _pytest.helpconfig import showversion
+
+                showversion(self)
+            elif (
+                getattr(self.option, "help", False) or "--help" in args or "-h" in args
+            ):
+                self._parser._getparser().print_help()
+                sys.stdout.write(
+                    "\nNOTE: displaying only minimal help due to UsageError.\n\n"
+                )
+
+            raise
+
+        return self
+
+    def notify_exception(
+        self,
+        excinfo: ExceptionInfo[BaseException],
+        option: argparse.Namespace | None = None,
+    ) -> None:
+        if option and getattr(option, "fulltrace", False):
+            style: TracebackStyle = "long"
+        else:
+            style = "native"
+        excrepr = excinfo.getrepr(
+            funcargs=True, showlocals=getattr(option, "showlocals", False), style=style
+        )
+        res = self.hook.pytest_internalerror(excrepr=excrepr, excinfo=excinfo)
+        if not any(res):
+            for line in str(excrepr).split("\n"):
+                sys.stderr.write(f"INTERNALERROR> {line}\n")
+                sys.stderr.flush()
+
+    def cwd_relative_nodeid(self, nodeid: str) -> str:
+        # nodeid's are relative to the rootpath, compute relative to cwd.
+        if self.invocation_params.dir != self.rootpath:
+            base_path_part, *nodeid_part = nodeid.split("::")
+            # Only process path part
+            fullpath = self.rootpath / base_path_part
+            relative_path = bestrelpath(self.invocation_params.dir, fullpath)
+
+            nodeid = "::".join([relative_path, *nodeid_part])
+        return nodeid
+
+    @classmethod
+    def fromdictargs(cls, option_dict, args) -> Config:
+        """Constructor usable for subprocesses."""
+        config = get_config(args)
+        config.option.__dict__.update(option_dict)
+        config.parse(args, addopts=False)
+        for x in config.option.plugins:
+            config.pluginmanager.consider_pluginarg(x)
+        return config
+
+    def _processopt(self, opt: Argument) -> None:
+        for name in opt._short_opts + opt._long_opts:
+            self._opt2dest[name] = opt.dest
+
+        if hasattr(opt, "default"):
+            if not hasattr(self.option, opt.dest):
+                setattr(self.option, opt.dest, opt.default)
+
+    @hookimpl(trylast=True)
+    def pytest_load_initial_conftests(self, early_config: Config) -> None:
+        # We haven't fully parsed the command line arguments yet, so
+        # early_config.args it not set yet. But we need it for
+        # discovering the initial conftests. So "pre-run" the logic here.
+        # It will be done for real in `parse()`.
+        args, args_source = early_config._decide_args(
+            args=early_config.known_args_namespace.file_or_dir,
+            pyargs=early_config.known_args_namespace.pyargs,
+            testpaths=early_config.getini("testpaths"),
+            invocation_dir=early_config.invocation_params.dir,
+            rootpath=early_config.rootpath,
+            warn=False,
+        )
+        self.pluginmanager._set_initial_conftests(
+            args=args,
+            pyargs=early_config.known_args_namespace.pyargs,
+            noconftest=early_config.known_args_namespace.noconftest,
+            rootpath=early_config.rootpath,
+            confcutdir=early_config.known_args_namespace.confcutdir,
+            invocation_dir=early_config.invocation_params.dir,
+            importmode=early_config.known_args_namespace.importmode,
+            consider_namespace_packages=early_config.getini(
+                "consider_namespace_packages"
+            ),
+        )
+
+    def _initini(self, args: Sequence[str]) -> None:
+        ns, unknown_args = self._parser.parse_known_and_unknown_args(
+            args, namespace=copy.copy(self.option)
+        )
+        rootpath, inipath, inicfg = determine_setup(
+            inifile=ns.inifilename,
+            args=ns.file_or_dir + unknown_args,
+            rootdir_cmd_arg=ns.rootdir or None,
+            invocation_dir=self.invocation_params.dir,
+        )
+        self._rootpath = rootpath
+        self._inipath = inipath
+        self.inicfg = inicfg
+        self._parser.extra_info["rootdir"] = str(self.rootpath)
+        self._parser.extra_info["inifile"] = str(self.inipath)
+        self._parser.addini("addopts", "Extra command line options", "args")
+        self._parser.addini("minversion", "Minimally required pytest version")
+        self._parser.addini(
+            "required_plugins",
+            "Plugins that must be present for pytest to run",
+            type="args",
+            default=[],
+        )
+        self._override_ini = ns.override_ini or ()
+
+    def _consider_importhook(self, args: Sequence[str]) -> None:
+        """Install the PEP 302 import hook if using assertion rewriting.
+
+        Needs to parse the --assert=<mode> option from the commandline
+        and find all the installed plugins to mark them for rewriting
+        by the importhook.
+        """
+        ns, unknown_args = self._parser.parse_known_and_unknown_args(args)
+        mode = getattr(ns, "assertmode", "plain")
+        if mode == "rewrite":
+            import _pytest.assertion
+
+            try:
+                hook = _pytest.assertion.install_importhook(self)
+            except SystemError:
+                mode = "plain"
+            else:
+                self._mark_plugins_for_rewrite(hook)
+        self._warn_about_missing_assertion(mode)
+
+    def _mark_plugins_for_rewrite(self, hook) -> None:
+        """Given an importhook, mark for rewrite any top-level
+        modules or packages in the distribution package for
+        all pytest plugins."""
+        self.pluginmanager.rewrite_hook = hook
+
+        if os.environ.get("PYTEST_DISABLE_PLUGIN_AUTOLOAD"):
+            # We don't autoload from distribution package entry points,
+            # no need to continue.
+            return
+
+        package_files = (
+            str(file)
+            for dist in importlib.metadata.distributions()
+            if any(ep.group == "pytest11" for ep in dist.entry_points)
+            for file in dist.files or []
+        )
+
+        for name in _iter_rewritable_modules(package_files):
+            hook.mark_rewrite(name)
+
+    def _validate_args(self, args: list[str], via: str) -> list[str]:
+        """Validate known args."""
+        self._parser._config_source_hint = via  # type: ignore
+        try:
+            self._parser.parse_known_and_unknown_args(
+                args, namespace=copy.copy(self.option)
+            )
+        finally:
+            del self._parser._config_source_hint  # type: ignore
+
+        return args
+
+    def _decide_args(
+        self,
+        *,
+        args: list[str],
+        pyargs: bool,
+        testpaths: list[str],
+        invocation_dir: pathlib.Path,
+        rootpath: pathlib.Path,
+        warn: bool,
+    ) -> tuple[list[str], ArgsSource]:
+        """Decide the args (initial paths/nodeids) to use given the relevant inputs.
+
+        :param warn: Whether can issue warnings.
+
+        :returns: The args and the args source. Guaranteed to be non-empty.
+        """
+        if args:
+            source = Config.ArgsSource.ARGS
+            result = args
+        else:
+            if invocation_dir == rootpath:
+                source = Config.ArgsSource.TESTPATHS
+                if pyargs:
+                    result = testpaths
+                else:
+                    result = []
+                    for path in testpaths:
+                        result.extend(sorted(glob.iglob(path, recursive=True)))
+                    if testpaths and not result:
+                        if warn:
+                            warning_text = (
+                                "No files were found in testpaths; "
+                                "consider removing or adjusting your testpaths configuration. "
+                                "Searching recursively from the current directory instead."
+                            )
+                            self.issue_config_time_warning(
+                                PytestConfigWarning(warning_text), stacklevel=3
+                            )
+            else:
+                result = []
+            if not result:
+                source = Config.ArgsSource.INVOCATION_DIR
+                result = [str(invocation_dir)]
+        return result, source
+
+    def _preparse(self, args: list[str], addopts: bool = True) -> None:
+        if addopts:
+            env_addopts = os.environ.get("PYTEST_ADDOPTS", "")
+            if len(env_addopts):
+                args[:] = (
+                    self._validate_args(shlex.split(env_addopts), "via PYTEST_ADDOPTS")
+                    + args
+                )
+        self._initini(args)
+        if addopts:
+            args[:] = (
+                self._validate_args(self.getini("addopts"), "via addopts config") + args
+            )
+
+        self.known_args_namespace = self._parser.parse_known_args(
+            args, namespace=copy.copy(self.option)
+        )
+        self._checkversion()
+        self._consider_importhook(args)
+        self.pluginmanager.consider_preparse(args, exclude_only=False)
+        if not os.environ.get("PYTEST_DISABLE_PLUGIN_AUTOLOAD"):
+            # Don't autoload from distribution package entry point. Only
+            # explicitly specified plugins are going to be loaded.
+            self.pluginmanager.load_setuptools_entrypoints("pytest11")
+        self.pluginmanager.consider_env()
+
+        self.known_args_namespace = self._parser.parse_known_args(
+            args, namespace=copy.copy(self.known_args_namespace)
+        )
+
+        self._validate_plugins()
+        self._warn_about_skipped_plugins()
+
+        if self.known_args_namespace.confcutdir is None:
+            if self.inipath is not None:
+                confcutdir = str(self.inipath.parent)
+            else:
+                confcutdir = str(self.rootpath)
+            self.known_args_namespace.confcutdir = confcutdir
+        try:
+            self.hook.pytest_load_initial_conftests(
+                early_config=self, args=args, parser=self._parser
+            )
+        except ConftestImportFailure as e:
+            if self.known_args_namespace.help or self.known_args_namespace.version:
+                # we don't want to prevent --help/--version to work
+                # so just let is pass and print a warning at the end
+                self.issue_config_time_warning(
+                    PytestConfigWarning(f"could not load initial conftests: {e.path}"),
+                    stacklevel=2,
+                )
+            else:
+                raise
+
+    @hookimpl(wrapper=True)
+    def pytest_collection(self) -> Generator[None, object, object]:
+        # Validate invalid ini keys after collection is done so we take in account
+        # options added by late-loading conftest files.
+        try:
+            return (yield)
+        finally:
+            self._validate_config_options()
+
+    def _checkversion(self) -> None:
+        import pytest
+
+        minver = self.inicfg.get("minversion", None)
+        if minver:
+            # Imported lazily to improve start-up time.
+            from packaging.version import Version
+
+            if not isinstance(minver, str):
+                raise pytest.UsageError(
+                    f"{self.inipath}: 'minversion' must be a single value"
+                )
+
+            if Version(minver) > Version(pytest.__version__):
+                raise pytest.UsageError(
+                    f"{self.inipath}: 'minversion' requires pytest-{minver}, actual pytest-{pytest.__version__}'"
+                )
+
+    def _validate_config_options(self) -> None:
+        for key in sorted(self._get_unknown_ini_keys()):
+            self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
+
+    def _validate_plugins(self) -> None:
+        required_plugins = sorted(self.getini("required_plugins"))
+        if not required_plugins:
+            return
+
+        # Imported lazily to improve start-up time.
+        from packaging.requirements import InvalidRequirement
+        from packaging.requirements import Requirement
+        from packaging.version import Version
+
+        plugin_info = self.pluginmanager.list_plugin_distinfo()
+        plugin_dist_info = {dist.project_name: dist.version for _, dist in plugin_info}
+
+        missing_plugins = []
+        for required_plugin in required_plugins:
+            try:
+                req = Requirement(required_plugin)
+            except InvalidRequirement:
+                missing_plugins.append(required_plugin)
+                continue
+
+            if req.name not in plugin_dist_info:
+                missing_plugins.append(required_plugin)
+            elif not req.specifier.contains(
+                Version(plugin_dist_info[req.name]), prereleases=True
+            ):
+                missing_plugins.append(required_plugin)
+
+        if missing_plugins:
+            raise UsageError(
+                "Missing required plugins: {}".format(", ".join(missing_plugins)),
+            )
+
+    def _warn_or_fail_if_strict(self, message: str) -> None:
+        if self.known_args_namespace.strict_config:
+            raise UsageError(message)
+
+        self.issue_config_time_warning(PytestConfigWarning(message), stacklevel=3)
+
+    def _get_unknown_ini_keys(self) -> list[str]:
+        parser_inicfg = self._parser._inidict
+        return [name for name in self.inicfg if name not in parser_inicfg]
+
+    def parse(self, args: list[str], addopts: bool = True) -> None:
+        # Parse given cmdline arguments into this config object.
+        assert (
+            self.args == []
+        ), "can only parse cmdline args at most once per Config object"
+        self.hook.pytest_addhooks.call_historic(
+            kwargs=dict(pluginmanager=self.pluginmanager)
+        )
+        self._preparse(args, addopts=addopts)
+        self._parser.after_preparse = True  # type: ignore
+        try:
+            args = self._parser.parse_setoption(
+                args, self.option, namespace=self.option
+            )
+            self.args, self.args_source = self._decide_args(
+                args=args,
+                pyargs=self.known_args_namespace.pyargs,
+                testpaths=self.getini("testpaths"),
+                invocation_dir=self.invocation_params.dir,
+                rootpath=self.rootpath,
+                warn=True,
+            )
+        except PrintHelp:
+            pass
+
+    def issue_config_time_warning(self, warning: Warning, stacklevel: int) -> None:
+        """Issue and handle a warning during the "configure" stage.
+
+        During ``pytest_configure`` we can't capture warnings using the ``catch_warnings_for_item``
+        function because it is not possible to have hook wrappers around ``pytest_configure``.
+
+        This function is mainly intended for plugins that need to issue warnings during
+        ``pytest_configure`` (or similar stages).
+
+        :param warning: The warning instance.
+        :param stacklevel: stacklevel forwarded to warnings.warn.
+        """
+        if self.pluginmanager.is_blocked("warnings"):
+            return
+
+        cmdline_filters = self.known_args_namespace.pythonwarnings or []
+        config_filters = self.getini("filterwarnings")
+
+        with warnings.catch_warnings(record=True) as records:
+            warnings.simplefilter("always", type(warning))
+            apply_warning_filters(config_filters, cmdline_filters)
+            warnings.warn(warning, stacklevel=stacklevel)
+
+        if records:
+            frame = sys._getframe(stacklevel - 1)
+            location = frame.f_code.co_filename, frame.f_lineno, frame.f_code.co_name
+            self.hook.pytest_warning_recorded.call_historic(
+                kwargs=dict(
+                    warning_message=records[0],
+                    when="config",
+                    nodeid="",
+                    location=location,
+                )
+            )
+
+    def addinivalue_line(self, name: str, line: str) -> None:
+        """Add a line to an ini-file option. The option must have been
+        declared but might not yet be set in which case the line becomes
+        the first line in its value."""
+        x = self.getini(name)
+        assert isinstance(x, list)
+        x.append(line)  # modifies the cached list inline
+
+    def getini(self, name: str):
+        """Return configuration value from an :ref:`ini file <configfiles>`.
+
+        If a configuration value is not defined in an
+        :ref:`ini file <configfiles>`, then the ``default`` value provided while
+        registering the configuration through
+        :func:`parser.addini <pytest.Parser.addini>` will be returned.
+        Please note that you can even provide ``None`` as a valid
+        default value.
+
+        If ``default`` is not provided while registering using
+        :func:`parser.addini <pytest.Parser.addini>`, then a default value
+        based on the ``type`` parameter passed to
+        :func:`parser.addini <pytest.Parser.addini>` will be returned.
+        The default values based on ``type`` are:
+        ``paths``, ``pathlist``, ``args`` and ``linelist`` : empty list ``[]``
+        ``bool`` : ``False``
+        ``string`` : empty string ``""``
+
+        If neither the ``default`` nor the ``type`` parameter is passed
+        while registering the configuration through
+        :func:`parser.addini <pytest.Parser.addini>`, then the configuration
+        is treated as a string and a default empty string '' is returned.
+
+        If the specified name hasn't been registered through a prior
+        :func:`parser.addini <pytest.Parser.addini>` call (usually from a
+        plugin), a ValueError is raised.
+        """
+        try:
+            return self._inicache[name]
+        except KeyError:
+            self._inicache[name] = val = self._getini(name)
+            return val
+
+    # Meant for easy monkeypatching by legacypath plugin.
+    # Can be inlined back (with no cover removed) once legacypath is gone.
+    def _getini_unknown_type(self, name: str, type: str, value: str | list[str]):
+        msg = f"unknown configuration type: {type}"
+        raise ValueError(msg, value)  # pragma: no cover
+
+    def _getini(self, name: str):
+        try:
+            description, type, default = self._parser._inidict[name]
+        except KeyError as e:
+            raise ValueError(f"unknown configuration value: {name!r}") from e
+        override_value = self._get_override_ini_value(name)
+        if override_value is None:
+            try:
+                value = self.inicfg[name]
+            except KeyError:
+                return default
+        else:
+            value = override_value
+        # Coerce the values based on types.
+        #
+        # Note: some coercions are only required if we are reading from .ini files, because
+        # the file format doesn't contain type information, but when reading from toml we will
+        # get either str or list of str values (see _parse_ini_config_from_pyproject_toml).
+        # For example:
+        #
+        #   ini:
+        #     a_line_list = "tests acceptance"
+        #   in this case, we need to split the string to obtain a list of strings.
+        #
+        #   toml:
+        #     a_line_list = ["tests", "acceptance"]
+        #   in this case, we already have a list ready to use.
+        #
+        if type == "paths":
+            dp = (
+                self.inipath.parent
+                if self.inipath is not None
+                else self.invocation_params.dir
+            )
+            input_values = shlex.split(value) if isinstance(value, str) else value
+            return [dp / x for x in input_values]
+        elif type == "args":
+            return shlex.split(value) if isinstance(value, str) else value
+        elif type == "linelist":
+            if isinstance(value, str):
+                return [t for t in map(lambda x: x.strip(), value.split("\n")) if t]
+            else:
+                return value
+        elif type == "bool":
+            return _strtobool(str(value).strip())
+        elif type == "string":
+            return value
+        elif type is None:
+            return value
+        else:
+            return self._getini_unknown_type(name, type, value)
+
+    def _getconftest_pathlist(
+        self, name: str, path: pathlib.Path
+    ) -> list[pathlib.Path] | None:
+        try:
+            mod, relroots = self.pluginmanager._rget_with_confmod(name, path)
+        except KeyError:
+            return None
+        assert mod.__file__ is not None
+        modpath = pathlib.Path(mod.__file__).parent
+        values: list[pathlib.Path] = []
+        for relroot in relroots:
+            if isinstance(relroot, os.PathLike):
+                relroot = pathlib.Path(relroot)
+            else:
+                relroot = relroot.replace("/", os.sep)
+                relroot = absolutepath(modpath / relroot)
+            values.append(relroot)
+        return values
+
+    def _get_override_ini_value(self, name: str) -> str | None:
+        value = None
+        # override_ini is a list of "ini=value" options.
+        # Always use the last item if multiple values are set for same ini-name,
+        # e.g. -o foo=bar1 -o foo=bar2 will set foo to bar2.
+        for ini_config in self._override_ini:
+            try:
+                key, user_ini_value = ini_config.split("=", 1)
+            except ValueError as e:
+                raise UsageError(
+                    f"-o/--override-ini expects option=value style (got: {ini_config!r})."
+                ) from e
+            else:
+                if key == name:
+                    value = user_ini_value
+        return value
+
+    def getoption(self, name: str, default=notset, skip: bool = False):
+        """Return command line option value.
+
+        :param name: Name of the option. You may also specify
+            the literal ``--OPT`` option instead of the "dest" option name.
+        :param default: Fallback value if no option of that name is **declared** via :hook:`pytest_addoption`.
+            Note this parameter will be ignored when the option is **declared** even if the option's value is ``None``.
+        :param skip: If ``True``, raise :func:`pytest.skip` if option is undeclared or has a ``None`` value.
+            Note that even if ``True``, if a default was specified it will be returned instead of a skip.
+        """
+        name = self._opt2dest.get(name, name)
+        try:
+            val = getattr(self.option, name)
+            if val is None and skip:
+                raise AttributeError(name)
+            return val
+        except AttributeError as e:
+            if default is not notset:
+                return default
+            if skip:
+                import pytest
+
+                pytest.skip(f"no {name!r} option found")
+            raise ValueError(f"no option named {name!r}") from e
+
+    def getvalue(self, name: str, path=None):
+        """Deprecated, use getoption() instead."""
+        return self.getoption(name)
+
+    def getvalueorskip(self, name: str, path=None):
+        """Deprecated, use getoption(skip=True) instead."""
+        return self.getoption(name, skip=True)
+
+    #: Verbosity type for failed assertions (see :confval:`verbosity_assertions`).
+    VERBOSITY_ASSERTIONS: Final = "assertions"
+    #: Verbosity type for test case execution (see :confval:`verbosity_test_cases`).
+    VERBOSITY_TEST_CASES: Final = "test_cases"
+    _VERBOSITY_INI_DEFAULT: Final = "auto"
+
+    def get_verbosity(self, verbosity_type: str | None = None) -> int:
+        r"""Retrieve the verbosity level for a fine-grained verbosity type.
+
+        :param verbosity_type: Verbosity type to get level for. If a level is
+            configured for the given type, that value will be returned. If the
+            given type is not a known verbosity type, the global verbosity
+            level will be returned. If the given type is None (default), the
+            global verbosity level will be returned.
+
+        To configure a level for a fine-grained verbosity type, the
+        configuration file should have a setting for the configuration name
+        and a numeric value for the verbosity level. A special value of "auto"
+        can be used to explicitly use the global verbosity level.
+
+        Example:
+
+        .. code-block:: ini
+
+            # content of pytest.ini
+            [pytest]
+            verbosity_assertions = 2
+
+        .. code-block:: console
+
+            pytest -v
+
+        .. code-block:: python
+
+            print(config.get_verbosity())  # 1
+            print(config.get_verbosity(Config.VERBOSITY_ASSERTIONS))  # 2
+        """
+        global_level = self.getoption("verbose", default=0)
+        assert isinstance(global_level, int)
+        if verbosity_type is None:
+            return global_level
+
+        ini_name = Config._verbosity_ini_name(verbosity_type)
+        if ini_name not in self._parser._inidict:
+            return global_level
+
+        level = self.getini(ini_name)
+        if level == Config._VERBOSITY_INI_DEFAULT:
+            return global_level
+
+        return int(level)
+
+    @staticmethod
+    def _verbosity_ini_name(verbosity_type: str) -> str:
+        return f"verbosity_{verbosity_type}"
+
+    @staticmethod
+    def _add_verbosity_ini(parser: Parser, verbosity_type: str, help: str) -> None:
+        """Add a output verbosity configuration option for the given output type.
+
+        :param parser: Parser for command line arguments and ini-file values.
+        :param verbosity_type: Fine-grained verbosity category.
+        :param help: Description of the output this type controls.
+
+        The value should be retrieved via a call to
+        :py:func:`config.get_verbosity(type) <pytest.Config.get_verbosity>`.
+        """
+        parser.addini(
+            Config._verbosity_ini_name(verbosity_type),
+            help=help,
+            type="string",
+            default=Config._VERBOSITY_INI_DEFAULT,
+        )
+
+    def _warn_about_missing_assertion(self, mode: str) -> None:
+        if not _assertion_supported():
+            if mode == "plain":
+                warning_text = (
+                    "ASSERTIONS ARE NOT EXECUTED"
+                    " and FAILING TESTS WILL PASS.  Are you"
+                    " using python -O?"
+                )
+            else:
+                warning_text = (
+                    "assertions not in test modules or"
+                    " plugins will be ignored"
+                    " because assert statements are not executed "
+                    "by the underlying Python interpreter "
+                    "(are you using python -O?)\n"
+                )
+            self.issue_config_time_warning(
+                PytestConfigWarning(warning_text),
+                stacklevel=3,
+            )
+
+    def _warn_about_skipped_plugins(self) -> None:
+        for module_name, msg in self.pluginmanager.skipped_plugins:
+            self.issue_config_time_warning(
+                PytestConfigWarning(f"skipped plugin {module_name!r}: {msg}"),
+                stacklevel=2,
+            )
+
+
+def _assertion_supported() -> bool:
+    try:
+        assert False
+    except AssertionError:
+        return True
+    else:
+        return False  # type: ignore[unreachable]
+
+
+def create_terminal_writer(
+    config: Config, file: TextIO | None = None
+) -> TerminalWriter:
+    """Create a TerminalWriter instance configured according to the options
+    in the config object.
+
+    Every code which requires a TerminalWriter object and has access to a
+    config object should use this function.
+    """
+    tw = TerminalWriter(file=file)
+
+    if config.option.color == "yes":
+        tw.hasmarkup = True
+    elif config.option.color == "no":
+        tw.hasmarkup = False
+
+    if config.option.code_highlight == "yes":
+        tw.code_highlight = True
+    elif config.option.code_highlight == "no":
+        tw.code_highlight = False
+
+    return tw
+
+
+def _strtobool(val: str) -> bool:
+    """Convert a string representation of truth to True or False.
+
+    True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values
+    are 'n', 'no', 'f', 'false', 'off', and '0'.  Raises ValueError if
+    'val' is anything else.
+
+    .. note:: Copied from distutils.util.
+    """
+    val = val.lower()
+    if val in ("y", "yes", "t", "true", "on", "1"):
+        return True
+    elif val in ("n", "no", "f", "false", "off", "0"):
+        return False
+    else:
+        raise ValueError(f"invalid truth value {val!r}")
+
+
+@lru_cache(maxsize=50)
+def parse_warning_filter(
+    arg: str, *, escape: bool
+) -> tuple[warnings._ActionKind, str, type[Warning], str, int]:
+    """Parse a warnings filter string.
+
+    This is copied from warnings._setoption with the following changes:
+
+    * Does not apply the filter.
+    * Escaping is optional.
+    * Raises UsageError so we get nice error messages on failure.
+    """
+    __tracebackhide__ = True
+    error_template = dedent(
+        f"""\
+        while parsing the following warning configuration:
+
+          {arg}
+
+        This error occurred:
+
+        {{error}}
+        """
+    )
+
+    parts = arg.split(":")
+    if len(parts) > 5:
+        doc_url = (
+            "https://docs.python.org/3/library/warnings.html#describing-warning-filters"
+        )
+        error = dedent(
+            f"""\
+            Too many fields ({len(parts)}), expected at most 5 separated by colons:
+
+              action:message:category:module:line
+
+            For more information please consult: {doc_url}
+            """
+        )
+        raise UsageError(error_template.format(error=error))
+
+    while len(parts) < 5:
+        parts.append("")
+    action_, message, category_, module, lineno_ = (s.strip() for s in parts)
+    try:
+        action: warnings._ActionKind = warnings._getaction(action_)  # type: ignore[attr-defined]
+    except warnings._OptionError as e:
+        raise UsageError(error_template.format(error=str(e))) from None
+    try:
+        category: type[Warning] = _resolve_warning_category(category_)
+    except Exception:
+        exc_info = ExceptionInfo.from_current()
+        exception_text = exc_info.getrepr(style="native")
+        raise UsageError(error_template.format(error=exception_text)) from None
+    if message and escape:
+        message = re.escape(message)
+    if module and escape:
+        module = re.escape(module) + r"\Z"
+    if lineno_:
+        try:
+            lineno = int(lineno_)
+            if lineno < 0:
+                raise ValueError("number is negative")
+        except ValueError as e:
+            raise UsageError(
+                error_template.format(error=f"invalid lineno {lineno_!r}: {e}")
+            ) from None
+    else:
+        lineno = 0
+    return action, message, category, module, lineno
+
+
+def _resolve_warning_category(category: str) -> type[Warning]:
+    """
+    Copied from warnings._getcategory, but changed so it lets exceptions (specially ImportErrors)
+    propagate so we can get access to their tracebacks (#9218).
+    """
+    __tracebackhide__ = True
+    if not category:
+        return Warning
+
+    if "." not in category:
+        import builtins as m
+
+        klass = category
+    else:
+        module, _, klass = category.rpartition(".")
+        m = __import__(module, None, None, [klass])
+    cat = getattr(m, klass)
+    if not issubclass(cat, Warning):
+        raise UsageError(f"{cat} is not a Warning subclass")
+    return cast(Type[Warning], cat)
+
+
+def apply_warning_filters(
+    config_filters: Iterable[str], cmdline_filters: Iterable[str]
+) -> None:
+    """Applies pytest-configured filters to the warnings module"""
+    # Filters should have this precedence: cmdline options, config.
+    # Filters should be applied in the inverse order of precedence.
+    for arg in config_filters:
+        warnings.filterwarnings(*parse_warning_filter(arg, escape=False))
+
+    for arg in cmdline_filters:
+        warnings.filterwarnings(*parse_warning_filter(arg, escape=True))
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..87498769
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/argparsing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/argparsing.cpython-312.pyc
new file mode 100644
index 00000000..c87a4c84
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/argparsing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..13a4dbf4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..3d1aafea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/findpaths.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/findpaths.cpython-312.pyc
new file mode 100644
index 00000000..e32de350
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/config/__pycache__/findpaths.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/argparsing.py b/backend/venv/lib/python3.12/site-packages/_pytest/config/argparsing.py
new file mode 100644
index 00000000..85aa4632
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/config/argparsing.py
@@ -0,0 +1,551 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import argparse
+from gettext import gettext
+import os
+import sys
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import final
+from typing import List
+from typing import Literal
+from typing import Mapping
+from typing import NoReturn
+from typing import Sequence
+
+import _pytest._io
+from _pytest.config.exceptions import UsageError
+from _pytest.deprecated import check_ispytest
+
+
+FILE_OR_DIR = "file_or_dir"
+
+
+class NotSet:
+    def __repr__(self) -> str:
+        return "<notset>"
+
+
+NOT_SET = NotSet()
+
+
+@final
+class Parser:
+    """Parser for command line arguments and ini-file values.
+
+    :ivar extra_info: Dict of generic param -> value to display in case
+        there's an error processing the command line arguments.
+    """
+
+    prog: str | None = None
+
+    def __init__(
+        self,
+        usage: str | None = None,
+        processopt: Callable[[Argument], None] | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._anonymous = OptionGroup("Custom options", parser=self, _ispytest=True)
+        self._groups: list[OptionGroup] = []
+        self._processopt = processopt
+        self._usage = usage
+        self._inidict: dict[str, tuple[str, str | None, Any]] = {}
+        self._ininames: list[str] = []
+        self.extra_info: dict[str, Any] = {}
+
+    def processoption(self, option: Argument) -> None:
+        if self._processopt:
+            if option.dest:
+                self._processopt(option)
+
+    def getgroup(
+        self, name: str, description: str = "", after: str | None = None
+    ) -> OptionGroup:
+        """Get (or create) a named option Group.
+
+        :param name: Name of the option group.
+        :param description: Long description for --help output.
+        :param after: Name of another group, used for ordering --help output.
+        :returns: The option group.
+
+        The returned group object has an ``addoption`` method with the same
+        signature as :func:`parser.addoption <pytest.Parser.addoption>` but
+        will be shown in the respective group in the output of
+        ``pytest --help``.
+        """
+        for group in self._groups:
+            if group.name == name:
+                return group
+        group = OptionGroup(name, description, parser=self, _ispytest=True)
+        i = 0
+        for i, grp in enumerate(self._groups):
+            if grp.name == after:
+                break
+        self._groups.insert(i + 1, group)
+        return group
+
+    def addoption(self, *opts: str, **attrs: Any) -> None:
+        """Register a command line option.
+
+        :param opts:
+            Option names, can be short or long options.
+        :param attrs:
+            Same attributes as the argparse library's :meth:`add_argument()
+            <argparse.ArgumentParser.add_argument>` function accepts.
+
+        After command line parsing, options are available on the pytest config
+        object via ``config.option.NAME`` where ``NAME`` is usually set
+        by passing a ``dest`` attribute, for example
+        ``addoption("--long", dest="NAME", ...)``.
+        """
+        self._anonymous.addoption(*opts, **attrs)
+
+    def parse(
+        self,
+        args: Sequence[str | os.PathLike[str]],
+        namespace: argparse.Namespace | None = None,
+    ) -> argparse.Namespace:
+        from _pytest._argcomplete import try_argcomplete
+
+        self.optparser = self._getparser()
+        try_argcomplete(self.optparser)
+        strargs = [os.fspath(x) for x in args]
+        return self.optparser.parse_args(strargs, namespace=namespace)
+
+    def _getparser(self) -> MyOptionParser:
+        from _pytest._argcomplete import filescompleter
+
+        optparser = MyOptionParser(self, self.extra_info, prog=self.prog)
+        groups = [*self._groups, self._anonymous]
+        for group in groups:
+            if group.options:
+                desc = group.description or group.name
+                arggroup = optparser.add_argument_group(desc)
+                for option in group.options:
+                    n = option.names()
+                    a = option.attrs()
+                    arggroup.add_argument(*n, **a)
+        file_or_dir_arg = optparser.add_argument(FILE_OR_DIR, nargs="*")
+        # bash like autocompletion for dirs (appending '/')
+        # Type ignored because typeshed doesn't know about argcomplete.
+        file_or_dir_arg.completer = filescompleter  # type: ignore
+        return optparser
+
+    def parse_setoption(
+        self,
+        args: Sequence[str | os.PathLike[str]],
+        option: argparse.Namespace,
+        namespace: argparse.Namespace | None = None,
+    ) -> list[str]:
+        parsedoption = self.parse(args, namespace=namespace)
+        for name, value in parsedoption.__dict__.items():
+            setattr(option, name, value)
+        return cast(List[str], getattr(parsedoption, FILE_OR_DIR))
+
+    def parse_known_args(
+        self,
+        args: Sequence[str | os.PathLike[str]],
+        namespace: argparse.Namespace | None = None,
+    ) -> argparse.Namespace:
+        """Parse the known arguments at this point.
+
+        :returns: An argparse namespace object.
+        """
+        return self.parse_known_and_unknown_args(args, namespace=namespace)[0]
+
+    def parse_known_and_unknown_args(
+        self,
+        args: Sequence[str | os.PathLike[str]],
+        namespace: argparse.Namespace | None = None,
+    ) -> tuple[argparse.Namespace, list[str]]:
+        """Parse the known arguments at this point, and also return the
+        remaining unknown arguments.
+
+        :returns:
+            A tuple containing an argparse namespace object for the known
+            arguments, and a list of the unknown arguments.
+        """
+        optparser = self._getparser()
+        strargs = [os.fspath(x) for x in args]
+        return optparser.parse_known_args(strargs, namespace=namespace)
+
+    def addini(
+        self,
+        name: str,
+        help: str,
+        type: Literal["string", "paths", "pathlist", "args", "linelist", "bool"]
+        | None = None,
+        default: Any = NOT_SET,
+    ) -> None:
+        """Register an ini-file option.
+
+        :param name:
+            Name of the ini-variable.
+        :param type:
+            Type of the variable. Can be:
+
+                * ``string``: a string
+                * ``bool``: a boolean
+                * ``args``: a list of strings, separated as in a shell
+                * ``linelist``: a list of strings, separated by line breaks
+                * ``paths``: a list of :class:`pathlib.Path`, separated as in a shell
+                * ``pathlist``: a list of ``py.path``, separated as in a shell
+
+            For ``paths`` and ``pathlist`` types, they are considered relative to the ini-file.
+            In case the execution is happening without an ini-file defined,
+            they will be considered relative to the current working directory (for example with ``--override-ini``).
+
+            .. versionadded:: 7.0
+                The ``paths`` variable type.
+
+            .. versionadded:: 8.1
+                Use the current working directory to resolve ``paths`` and ``pathlist`` in the absence of an ini-file.
+
+            Defaults to ``string`` if ``None`` or not passed.
+        :param default:
+            Default value if no ini-file option exists but is queried.
+
+        The value of ini-variables can be retrieved via a call to
+        :py:func:`config.getini(name) <pytest.Config.getini>`.
+        """
+        assert type in (None, "string", "paths", "pathlist", "args", "linelist", "bool")
+        if default is NOT_SET:
+            default = get_ini_default_for_type(type)
+
+        self._inidict[name] = (help, type, default)
+        self._ininames.append(name)
+
+
+def get_ini_default_for_type(
+    type: Literal["string", "paths", "pathlist", "args", "linelist", "bool"] | None,
+) -> Any:
+    """
+    Used by addini to get the default value for a given ini-option type, when
+    default is not supplied.
+    """
+    if type is None:
+        return ""
+    elif type in ("paths", "pathlist", "args", "linelist"):
+        return []
+    elif type == "bool":
+        return False
+    else:
+        return ""
+
+
+class ArgumentError(Exception):
+    """Raised if an Argument instance is created with invalid or
+    inconsistent arguments."""
+
+    def __init__(self, msg: str, option: Argument | str) -> None:
+        self.msg = msg
+        self.option_id = str(option)
+
+    def __str__(self) -> str:
+        if self.option_id:
+            return f"option {self.option_id}: {self.msg}"
+        else:
+            return self.msg
+
+
+class Argument:
+    """Class that mimics the necessary behaviour of optparse.Option.
+
+    It's currently a least effort implementation and ignoring choices
+    and integer prefixes.
+
+    https://docs.python.org/3/library/optparse.html#optparse-standard-option-types
+    """
+
+    def __init__(self, *names: str, **attrs: Any) -> None:
+        """Store params in private vars for use in add_argument."""
+        self._attrs = attrs
+        self._short_opts: list[str] = []
+        self._long_opts: list[str] = []
+        try:
+            self.type = attrs["type"]
+        except KeyError:
+            pass
+        try:
+            # Attribute existence is tested in Config._processopt.
+            self.default = attrs["default"]
+        except KeyError:
+            pass
+        self._set_opt_strings(names)
+        dest: str | None = attrs.get("dest")
+        if dest:
+            self.dest = dest
+        elif self._long_opts:
+            self.dest = self._long_opts[0][2:].replace("-", "_")
+        else:
+            try:
+                self.dest = self._short_opts[0][1:]
+            except IndexError as e:
+                self.dest = "???"  # Needed for the error repr.
+                raise ArgumentError("need a long or short option", self) from e
+
+    def names(self) -> list[str]:
+        return self._short_opts + self._long_opts
+
+    def attrs(self) -> Mapping[str, Any]:
+        # Update any attributes set by processopt.
+        attrs = "default dest help".split()
+        attrs.append(self.dest)
+        for attr in attrs:
+            try:
+                self._attrs[attr] = getattr(self, attr)
+            except AttributeError:
+                pass
+        return self._attrs
+
+    def _set_opt_strings(self, opts: Sequence[str]) -> None:
+        """Directly from optparse.
+
+        Might not be necessary as this is passed to argparse later on.
+        """
+        for opt in opts:
+            if len(opt) < 2:
+                raise ArgumentError(
+                    f"invalid option string {opt!r}: "
+                    "must be at least two characters long",
+                    self,
+                )
+            elif len(opt) == 2:
+                if not (opt[0] == "-" and opt[1] != "-"):
+                    raise ArgumentError(
+                        f"invalid short option string {opt!r}: "
+                        "must be of the form -x, (x any non-dash char)",
+                        self,
+                    )
+                self._short_opts.append(opt)
+            else:
+                if not (opt[0:2] == "--" and opt[2] != "-"):
+                    raise ArgumentError(
+                        f"invalid long option string {opt!r}: "
+                        "must start with --, followed by non-dash",
+                        self,
+                    )
+                self._long_opts.append(opt)
+
+    def __repr__(self) -> str:
+        args: list[str] = []
+        if self._short_opts:
+            args += ["_short_opts: " + repr(self._short_opts)]
+        if self._long_opts:
+            args += ["_long_opts: " + repr(self._long_opts)]
+        args += ["dest: " + repr(self.dest)]
+        if hasattr(self, "type"):
+            args += ["type: " + repr(self.type)]
+        if hasattr(self, "default"):
+            args += ["default: " + repr(self.default)]
+        return "Argument({})".format(", ".join(args))
+
+
+class OptionGroup:
+    """A group of options shown in its own section."""
+
+    def __init__(
+        self,
+        name: str,
+        description: str = "",
+        parser: Parser | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self.name = name
+        self.description = description
+        self.options: list[Argument] = []
+        self.parser = parser
+
+    def addoption(self, *opts: str, **attrs: Any) -> None:
+        """Add an option to this group.
+
+        If a shortened version of a long option is specified, it will
+        be suppressed in the help. ``addoption('--twowords', '--two-words')``
+        results in help showing ``--two-words`` only, but ``--twowords`` gets
+        accepted **and** the automatic destination is in ``args.twowords``.
+
+        :param opts:
+            Option names, can be short or long options.
+        :param attrs:
+            Same attributes as the argparse library's :meth:`add_argument()
+            <argparse.ArgumentParser.add_argument>` function accepts.
+        """
+        conflict = set(opts).intersection(
+            name for opt in self.options for name in opt.names()
+        )
+        if conflict:
+            raise ValueError(f"option names {conflict} already added")
+        option = Argument(*opts, **attrs)
+        self._addoption_instance(option, shortupper=False)
+
+    def _addoption(self, *opts: str, **attrs: Any) -> None:
+        option = Argument(*opts, **attrs)
+        self._addoption_instance(option, shortupper=True)
+
+    def _addoption_instance(self, option: Argument, shortupper: bool = False) -> None:
+        if not shortupper:
+            for opt in option._short_opts:
+                if opt[0] == "-" and opt[1].islower():
+                    raise ValueError("lowercase shortoptions reserved")
+        if self.parser:
+            self.parser.processoption(option)
+        self.options.append(option)
+
+
+class MyOptionParser(argparse.ArgumentParser):
+    def __init__(
+        self,
+        parser: Parser,
+        extra_info: dict[str, Any] | None = None,
+        prog: str | None = None,
+    ) -> None:
+        self._parser = parser
+        super().__init__(
+            prog=prog,
+            usage=parser._usage,
+            add_help=False,
+            formatter_class=DropShorterLongHelpFormatter,
+            allow_abbrev=False,
+            fromfile_prefix_chars="@",
+        )
+        # extra_info is a dict of (param -> value) to display if there's
+        # an usage error to provide more contextual information to the user.
+        self.extra_info = extra_info if extra_info else {}
+
+    def error(self, message: str) -> NoReturn:
+        """Transform argparse error message into UsageError."""
+        msg = f"{self.prog}: error: {message}"
+
+        if hasattr(self._parser, "_config_source_hint"):
+            msg = f"{msg} ({self._parser._config_source_hint})"
+
+        raise UsageError(self.format_usage() + msg)
+
+    # Type ignored because typeshed has a very complex type in the superclass.
+    def parse_args(  # type: ignore
+        self,
+        args: Sequence[str] | None = None,
+        namespace: argparse.Namespace | None = None,
+    ) -> argparse.Namespace:
+        """Allow splitting of positional arguments."""
+        parsed, unrecognized = self.parse_known_args(args, namespace)
+        if unrecognized:
+            for arg in unrecognized:
+                if arg and arg[0] == "-":
+                    lines = [
+                        "unrecognized arguments: {}".format(" ".join(unrecognized))
+                    ]
+                    for k, v in sorted(self.extra_info.items()):
+                        lines.append(f"  {k}: {v}")
+                    self.error("\n".join(lines))
+            getattr(parsed, FILE_OR_DIR).extend(unrecognized)
+        return parsed
+
+    if sys.version_info < (3, 9):  # pragma: no cover
+        # Backport of https://github.com/python/cpython/pull/14316 so we can
+        # disable long --argument abbreviations without breaking short flags.
+        def _parse_optional(
+            self, arg_string: str
+        ) -> tuple[argparse.Action | None, str, str | None] | None:
+            if not arg_string:
+                return None
+            if arg_string[0] not in self.prefix_chars:
+                return None
+            if arg_string in self._option_string_actions:
+                action = self._option_string_actions[arg_string]
+                return action, arg_string, None
+            if len(arg_string) == 1:
+                return None
+            if "=" in arg_string:
+                option_string, explicit_arg = arg_string.split("=", 1)
+                if option_string in self._option_string_actions:
+                    action = self._option_string_actions[option_string]
+                    return action, option_string, explicit_arg
+            if self.allow_abbrev or not arg_string.startswith("--"):
+                option_tuples = self._get_option_tuples(arg_string)
+                if len(option_tuples) > 1:
+                    msg = gettext(
+                        "ambiguous option: %(option)s could match %(matches)s"
+                    )
+                    options = ", ".join(option for _, option, _ in option_tuples)
+                    self.error(msg % {"option": arg_string, "matches": options})
+                elif len(option_tuples) == 1:
+                    (option_tuple,) = option_tuples
+                    return option_tuple
+            if self._negative_number_matcher.match(arg_string):
+                if not self._has_negative_number_optionals:
+                    return None
+            if " " in arg_string:
+                return None
+            return None, arg_string, None
+
+
+class DropShorterLongHelpFormatter(argparse.HelpFormatter):
+    """Shorten help for long options that differ only in extra hyphens.
+
+    - Collapse **long** options that are the same except for extra hyphens.
+    - Shortcut if there are only two options and one of them is a short one.
+    - Cache result on the action object as this is called at least 2 times.
+    """
+
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        # Use more accurate terminal width.
+        if "width" not in kwargs:
+            kwargs["width"] = _pytest._io.get_terminal_width()
+        super().__init__(*args, **kwargs)
+
+    def _format_action_invocation(self, action: argparse.Action) -> str:
+        orgstr = super()._format_action_invocation(action)
+        if orgstr and orgstr[0] != "-":  # only optional arguments
+            return orgstr
+        res: str | None = getattr(action, "_formatted_action_invocation", None)
+        if res:
+            return res
+        options = orgstr.split(", ")
+        if len(options) == 2 and (len(options[0]) == 2 or len(options[1]) == 2):
+            # a shortcut for '-h, --help' or '--abc', '-a'
+            action._formatted_action_invocation = orgstr  # type: ignore
+            return orgstr
+        return_list = []
+        short_long: dict[str, str] = {}
+        for option in options:
+            if len(option) == 2 or option[2] == " ":
+                continue
+            if not option.startswith("--"):
+                raise ArgumentError(
+                    f'long optional argument without "--": [{option}]', option
+                )
+            xxoption = option[2:]
+            shortened = xxoption.replace("-", "")
+            if shortened not in short_long or len(short_long[shortened]) < len(
+                xxoption
+            ):
+                short_long[shortened] = xxoption
+        # now short_long has been filled out to the longest with dashes
+        # **and** we keep the right option ordering from add_argument
+        for option in options:
+            if len(option) == 2 or option[2] == " ":
+                return_list.append(option)
+            if option[2:] == short_long.get(option.replace("-", "")):
+                return_list.append(option.replace(" ", "=", 1))
+        formatted_action_invocation = ", ".join(return_list)
+        action._formatted_action_invocation = formatted_action_invocation  # type: ignore
+        return formatted_action_invocation
+
+    def _split_lines(self, text, width):
+        """Wrap lines after splitting on original newlines.
+
+        This allows to have explicit line breaks in the help text.
+        """
+        import textwrap
+
+        lines = []
+        for line in text.splitlines():
+            lines.extend(textwrap.wrap(line.strip(), width))
+        return lines
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/compat.py b/backend/venv/lib/python3.12/site-packages/_pytest/config/compat.py
new file mode 100644
index 00000000..2856d85d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/config/compat.py
@@ -0,0 +1,85 @@
+from __future__ import annotations
+
+import functools
+from pathlib import Path
+from typing import Any
+from typing import Mapping
+import warnings
+
+import pluggy
+
+from ..compat import LEGACY_PATH
+from ..compat import legacy_path
+from ..deprecated import HOOK_LEGACY_PATH_ARG
+
+
+# hookname: (Path, LEGACY_PATH)
+imply_paths_hooks: Mapping[str, tuple[str, str]] = {
+    "pytest_ignore_collect": ("collection_path", "path"),
+    "pytest_collect_file": ("file_path", "path"),
+    "pytest_pycollect_makemodule": ("module_path", "path"),
+    "pytest_report_header": ("start_path", "startdir"),
+    "pytest_report_collectionfinish": ("start_path", "startdir"),
+}
+
+
+def _check_path(path: Path, fspath: LEGACY_PATH) -> None:
+    if Path(fspath) != path:
+        raise ValueError(
+            f"Path({fspath!r}) != {path!r}\n"
+            "if both path and fspath are given they need to be equal"
+        )
+
+
+class PathAwareHookProxy:
+    """
+    this helper wraps around hook callers
+    until pluggy supports fixingcalls, this one will do
+
+    it currently doesn't return full hook caller proxies for fixed hooks,
+    this may have to be changed later depending on bugs
+    """
+
+    def __init__(self, hook_relay: pluggy.HookRelay) -> None:
+        self._hook_relay = hook_relay
+
+    def __dir__(self) -> list[str]:
+        return dir(self._hook_relay)
+
+    def __getattr__(self, key: str) -> pluggy.HookCaller:
+        hook: pluggy.HookCaller = getattr(self._hook_relay, key)
+        if key not in imply_paths_hooks:
+            self.__dict__[key] = hook
+            return hook
+        else:
+            path_var, fspath_var = imply_paths_hooks[key]
+
+            @functools.wraps(hook)
+            def fixed_hook(**kw: Any) -> Any:
+                path_value: Path | None = kw.pop(path_var, None)
+                fspath_value: LEGACY_PATH | None = kw.pop(fspath_var, None)
+                if fspath_value is not None:
+                    warnings.warn(
+                        HOOK_LEGACY_PATH_ARG.format(
+                            pylib_path_arg=fspath_var, pathlib_path_arg=path_var
+                        ),
+                        stacklevel=2,
+                    )
+                if path_value is not None:
+                    if fspath_value is not None:
+                        _check_path(path_value, fspath_value)
+                    else:
+                        fspath_value = legacy_path(path_value)
+                else:
+                    assert fspath_value is not None
+                    path_value = Path(fspath_value)
+
+                kw[path_var] = path_value
+                kw[fspath_var] = fspath_value
+                return hook(**kw)
+
+            fixed_hook.name = hook.name  # type: ignore[attr-defined]
+            fixed_hook.spec = hook.spec  # type: ignore[attr-defined]
+            fixed_hook.__name__ = key
+            self.__dict__[key] = fixed_hook
+            return fixed_hook  # type: ignore[return-value]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/exceptions.py b/backend/venv/lib/python3.12/site-packages/_pytest/config/exceptions.py
new file mode 100644
index 00000000..90108eca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/config/exceptions.py
@@ -0,0 +1,13 @@
+from __future__ import annotations
+
+from typing import final
+
+
+@final
+class UsageError(Exception):
+    """Error in pytest usage or invocation."""
+
+
+class PrintHelp(Exception):
+    """Raised when pytest should print its help to skip the rest of the
+    argument parsing and validation."""
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/config/findpaths.py b/backend/venv/lib/python3.12/site-packages/_pytest/config/findpaths.py
new file mode 100644
index 00000000..ce4c990b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/config/findpaths.py
@@ -0,0 +1,228 @@
+from __future__ import annotations
+
+import os
+from pathlib import Path
+import sys
+from typing import Iterable
+from typing import Sequence
+
+import iniconfig
+
+from .exceptions import UsageError
+from _pytest.outcomes import fail
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import commonpath
+from _pytest.pathlib import safe_exists
+
+
+def _parse_ini_config(path: Path) -> iniconfig.IniConfig:
+    """Parse the given generic '.ini' file using legacy IniConfig parser, returning
+    the parsed object.
+
+    Raise UsageError if the file cannot be parsed.
+    """
+    try:
+        return iniconfig.IniConfig(str(path))
+    except iniconfig.ParseError as exc:
+        raise UsageError(str(exc)) from exc
+
+
+def load_config_dict_from_file(
+    filepath: Path,
+) -> dict[str, str | list[str]] | None:
+    """Load pytest configuration from the given file path, if supported.
+
+    Return None if the file does not contain valid pytest configuration.
+    """
+    # Configuration from ini files are obtained from the [pytest] section, if present.
+    if filepath.suffix == ".ini":
+        iniconfig = _parse_ini_config(filepath)
+
+        if "pytest" in iniconfig:
+            return dict(iniconfig["pytest"].items())
+        else:
+            # "pytest.ini" files are always the source of configuration, even if empty.
+            if filepath.name == "pytest.ini":
+                return {}
+
+    # '.cfg' files are considered if they contain a "[tool:pytest]" section.
+    elif filepath.suffix == ".cfg":
+        iniconfig = _parse_ini_config(filepath)
+
+        if "tool:pytest" in iniconfig.sections:
+            return dict(iniconfig["tool:pytest"].items())
+        elif "pytest" in iniconfig.sections:
+            # If a setup.cfg contains a "[pytest]" section, we raise a failure to indicate users that
+            # plain "[pytest]" sections in setup.cfg files is no longer supported (#3086).
+            fail(CFG_PYTEST_SECTION.format(filename="setup.cfg"), pytrace=False)
+
+    # '.toml' files are considered if they contain a [tool.pytest.ini_options] table.
+    elif filepath.suffix == ".toml":
+        if sys.version_info >= (3, 11):
+            import tomllib
+        else:
+            import tomli as tomllib
+
+        toml_text = filepath.read_text(encoding="utf-8")
+        try:
+            config = tomllib.loads(toml_text)
+        except tomllib.TOMLDecodeError as exc:
+            raise UsageError(f"{filepath}: {exc}") from exc
+
+        result = config.get("tool", {}).get("pytest", {}).get("ini_options", None)
+        if result is not None:
+            # TOML supports richer data types than ini files (strings, arrays, floats, ints, etc),
+            # however we need to convert all scalar values to str for compatibility with the rest
+            # of the configuration system, which expects strings only.
+            def make_scalar(v: object) -> str | list[str]:
+                return v if isinstance(v, list) else str(v)
+
+            return {k: make_scalar(v) for k, v in result.items()}
+
+    return None
+
+
+def locate_config(
+    invocation_dir: Path,
+    args: Iterable[Path],
+) -> tuple[Path | None, Path | None, dict[str, str | list[str]]]:
+    """Search in the list of arguments for a valid ini-file for pytest,
+    and return a tuple of (rootdir, inifile, cfg-dict)."""
+    config_names = [
+        "pytest.ini",
+        ".pytest.ini",
+        "pyproject.toml",
+        "tox.ini",
+        "setup.cfg",
+    ]
+    args = [x for x in args if not str(x).startswith("-")]
+    if not args:
+        args = [invocation_dir]
+    found_pyproject_toml: Path | None = None
+    for arg in args:
+        argpath = absolutepath(arg)
+        for base in (argpath, *argpath.parents):
+            for config_name in config_names:
+                p = base / config_name
+                if p.is_file():
+                    if p.name == "pyproject.toml" and found_pyproject_toml is None:
+                        found_pyproject_toml = p
+                    ini_config = load_config_dict_from_file(p)
+                    if ini_config is not None:
+                        return base, p, ini_config
+    if found_pyproject_toml is not None:
+        return found_pyproject_toml.parent, found_pyproject_toml, {}
+    return None, None, {}
+
+
+def get_common_ancestor(
+    invocation_dir: Path,
+    paths: Iterable[Path],
+) -> Path:
+    common_ancestor: Path | None = None
+    for path in paths:
+        if not path.exists():
+            continue
+        if common_ancestor is None:
+            common_ancestor = path
+        else:
+            if common_ancestor in path.parents or path == common_ancestor:
+                continue
+            elif path in common_ancestor.parents:
+                common_ancestor = path
+            else:
+                shared = commonpath(path, common_ancestor)
+                if shared is not None:
+                    common_ancestor = shared
+    if common_ancestor is None:
+        common_ancestor = invocation_dir
+    elif common_ancestor.is_file():
+        common_ancestor = common_ancestor.parent
+    return common_ancestor
+
+
+def get_dirs_from_args(args: Iterable[str]) -> list[Path]:
+    def is_option(x: str) -> bool:
+        return x.startswith("-")
+
+    def get_file_part_from_node_id(x: str) -> str:
+        return x.split("::")[0]
+
+    def get_dir_from_path(path: Path) -> Path:
+        if path.is_dir():
+            return path
+        return path.parent
+
+    # These look like paths but may not exist
+    possible_paths = (
+        absolutepath(get_file_part_from_node_id(arg))
+        for arg in args
+        if not is_option(arg)
+    )
+
+    return [get_dir_from_path(path) for path in possible_paths if safe_exists(path)]
+
+
+CFG_PYTEST_SECTION = "[pytest] section in {filename} files is no longer supported, change to [tool:pytest] instead."
+
+
+def determine_setup(
+    *,
+    inifile: str | None,
+    args: Sequence[str],
+    rootdir_cmd_arg: str | None,
+    invocation_dir: Path,
+) -> tuple[Path, Path | None, dict[str, str | list[str]]]:
+    """Determine the rootdir, inifile and ini configuration values from the
+    command line arguments.
+
+    :param inifile:
+        The `--inifile` command line argument, if given.
+    :param args:
+        The free command line arguments.
+    :param rootdir_cmd_arg:
+        The `--rootdir` command line argument, if given.
+    :param invocation_dir:
+        The working directory when pytest was invoked.
+    """
+    rootdir = None
+    dirs = get_dirs_from_args(args)
+    if inifile:
+        inipath_ = absolutepath(inifile)
+        inipath: Path | None = inipath_
+        inicfg = load_config_dict_from_file(inipath_) or {}
+        if rootdir_cmd_arg is None:
+            rootdir = inipath_.parent
+    else:
+        ancestor = get_common_ancestor(invocation_dir, dirs)
+        rootdir, inipath, inicfg = locate_config(invocation_dir, [ancestor])
+        if rootdir is None and rootdir_cmd_arg is None:
+            for possible_rootdir in (ancestor, *ancestor.parents):
+                if (possible_rootdir / "setup.py").is_file():
+                    rootdir = possible_rootdir
+                    break
+            else:
+                if dirs != [ancestor]:
+                    rootdir, inipath, inicfg = locate_config(invocation_dir, dirs)
+                if rootdir is None:
+                    rootdir = get_common_ancestor(
+                        invocation_dir, [invocation_dir, ancestor]
+                    )
+                    if is_fs_root(rootdir):
+                        rootdir = ancestor
+    if rootdir_cmd_arg:
+        rootdir = absolutepath(os.path.expandvars(rootdir_cmd_arg))
+        if not rootdir.is_dir():
+            raise UsageError(
+                f"Directory '{rootdir}' not found. Check your '--rootdir' option."
+            )
+    assert rootdir is not None
+    return rootdir, inipath, inicfg or {}
+
+
+def is_fs_root(p: Path) -> bool:
+    r"""
+    Return True if the given path is pointing to the root of the
+    file system ("/" on Unix and "C:\\" on Windows for example).
+    """
+    return os.path.splitdrive(str(p))[1] == os.sep
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/debugging.py b/backend/venv/lib/python3.12/site-packages/_pytest/debugging.py
new file mode 100644
index 00000000..2dfe321e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/debugging.py
@@ -0,0 +1,385 @@
+# mypy: allow-untyped-defs
+# ruff: noqa: T100
+"""Interactive debugging with PDB, the Python Debugger."""
+
+from __future__ import annotations
+
+import argparse
+import functools
+import sys
+import types
+from typing import Any
+from typing import Callable
+from typing import Generator
+import unittest
+
+from _pytest import outcomes
+from _pytest._code import ExceptionInfo
+from _pytest.capture import CaptureManager
+from _pytest.config import Config
+from _pytest.config import ConftestImportFailure
+from _pytest.config import hookimpl
+from _pytest.config import PytestPluginManager
+from _pytest.config.argparsing import Parser
+from _pytest.config.exceptions import UsageError
+from _pytest.nodes import Node
+from _pytest.reports import BaseReport
+from _pytest.runner import CallInfo
+
+
+def _validate_usepdb_cls(value: str) -> tuple[str, str]:
+    """Validate syntax of --pdbcls option."""
+    try:
+        modname, classname = value.split(":")
+    except ValueError as e:
+        raise argparse.ArgumentTypeError(
+            f"{value!r} is not in the format 'modname:classname'"
+        ) from e
+    return (modname, classname)
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group._addoption(
+        "--pdb",
+        dest="usepdb",
+        action="store_true",
+        help="Start the interactive Python debugger on errors or KeyboardInterrupt",
+    )
+    group._addoption(
+        "--pdbcls",
+        dest="usepdb_cls",
+        metavar="modulename:classname",
+        type=_validate_usepdb_cls,
+        help="Specify a custom interactive Python debugger for use with --pdb."
+        "For example: --pdbcls=IPython.terminal.debugger:TerminalPdb",
+    )
+    group._addoption(
+        "--trace",
+        dest="trace",
+        action="store_true",
+        help="Immediately break when running each test",
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    import pdb
+
+    if config.getvalue("trace"):
+        config.pluginmanager.register(PdbTrace(), "pdbtrace")
+    if config.getvalue("usepdb"):
+        config.pluginmanager.register(PdbInvoke(), "pdbinvoke")
+
+    pytestPDB._saved.append(
+        (pdb.set_trace, pytestPDB._pluginmanager, pytestPDB._config)
+    )
+    pdb.set_trace = pytestPDB.set_trace
+    pytestPDB._pluginmanager = config.pluginmanager
+    pytestPDB._config = config
+
+    # NOTE: not using pytest_unconfigure, since it might get called although
+    #       pytest_configure was not (if another plugin raises UsageError).
+    def fin() -> None:
+        (
+            pdb.set_trace,
+            pytestPDB._pluginmanager,
+            pytestPDB._config,
+        ) = pytestPDB._saved.pop()
+
+    config.add_cleanup(fin)
+
+
+class pytestPDB:
+    """Pseudo PDB that defers to the real pdb."""
+
+    _pluginmanager: PytestPluginManager | None = None
+    _config: Config | None = None
+    _saved: list[
+        tuple[Callable[..., None], PytestPluginManager | None, Config | None]
+    ] = []
+    _recursive_debug = 0
+    _wrapped_pdb_cls: tuple[type[Any], type[Any]] | None = None
+
+    @classmethod
+    def _is_capturing(cls, capman: CaptureManager | None) -> str | bool:
+        if capman:
+            return capman.is_capturing()
+        return False
+
+    @classmethod
+    def _import_pdb_cls(cls, capman: CaptureManager | None):
+        if not cls._config:
+            import pdb
+
+            # Happens when using pytest.set_trace outside of a test.
+            return pdb.Pdb
+
+        usepdb_cls = cls._config.getvalue("usepdb_cls")
+
+        if cls._wrapped_pdb_cls and cls._wrapped_pdb_cls[0] == usepdb_cls:
+            return cls._wrapped_pdb_cls[1]
+
+        if usepdb_cls:
+            modname, classname = usepdb_cls
+
+            try:
+                __import__(modname)
+                mod = sys.modules[modname]
+
+                # Handle --pdbcls=pdb:pdb.Pdb (useful e.g. with pdbpp).
+                parts = classname.split(".")
+                pdb_cls = getattr(mod, parts[0])
+                for part in parts[1:]:
+                    pdb_cls = getattr(pdb_cls, part)
+            except Exception as exc:
+                value = ":".join((modname, classname))
+                raise UsageError(
+                    f"--pdbcls: could not import {value!r}: {exc}"
+                ) from exc
+        else:
+            import pdb
+
+            pdb_cls = pdb.Pdb
+
+        wrapped_cls = cls._get_pdb_wrapper_class(pdb_cls, capman)
+        cls._wrapped_pdb_cls = (usepdb_cls, wrapped_cls)
+        return wrapped_cls
+
+    @classmethod
+    def _get_pdb_wrapper_class(cls, pdb_cls, capman: CaptureManager | None):
+        import _pytest.config
+
+        class PytestPdbWrapper(pdb_cls):
+            _pytest_capman = capman
+            _continued = False
+
+            def do_debug(self, arg):
+                cls._recursive_debug += 1
+                ret = super().do_debug(arg)
+                cls._recursive_debug -= 1
+                return ret
+
+            def do_continue(self, arg):
+                ret = super().do_continue(arg)
+                if cls._recursive_debug == 0:
+                    assert cls._config is not None
+                    tw = _pytest.config.create_terminal_writer(cls._config)
+                    tw.line()
+
+                    capman = self._pytest_capman
+                    capturing = pytestPDB._is_capturing(capman)
+                    if capturing:
+                        if capturing == "global":
+                            tw.sep(">", "PDB continue (IO-capturing resumed)")
+                        else:
+                            tw.sep(
+                                ">",
+                                f"PDB continue (IO-capturing resumed for {capturing})",
+                            )
+                        assert capman is not None
+                        capman.resume()
+                    else:
+                        tw.sep(">", "PDB continue")
+                assert cls._pluginmanager is not None
+                cls._pluginmanager.hook.pytest_leave_pdb(config=cls._config, pdb=self)
+                self._continued = True
+                return ret
+
+            do_c = do_cont = do_continue
+
+            def do_quit(self, arg):
+                """Raise Exit outcome when quit command is used in pdb.
+
+                This is a bit of a hack - it would be better if BdbQuit
+                could be handled, but this would require to wrap the
+                whole pytest run, and adjust the report etc.
+                """
+                ret = super().do_quit(arg)
+
+                if cls._recursive_debug == 0:
+                    outcomes.exit("Quitting debugger")
+
+                return ret
+
+            do_q = do_quit
+            do_exit = do_quit
+
+            def setup(self, f, tb):
+                """Suspend on setup().
+
+                Needed after do_continue resumed, and entering another
+                breakpoint again.
+                """
+                ret = super().setup(f, tb)
+                if not ret and self._continued:
+                    # pdb.setup() returns True if the command wants to exit
+                    # from the interaction: do not suspend capturing then.
+                    if self._pytest_capman:
+                        self._pytest_capman.suspend_global_capture(in_=True)
+                return ret
+
+            def get_stack(self, f, t):
+                stack, i = super().get_stack(f, t)
+                if f is None:
+                    # Find last non-hidden frame.
+                    i = max(0, len(stack) - 1)
+                    while i and stack[i][0].f_locals.get("__tracebackhide__", False):
+                        i -= 1
+                return stack, i
+
+        return PytestPdbWrapper
+
+    @classmethod
+    def _init_pdb(cls, method, *args, **kwargs):
+        """Initialize PDB debugging, dropping any IO capturing."""
+        import _pytest.config
+
+        if cls._pluginmanager is None:
+            capman: CaptureManager | None = None
+        else:
+            capman = cls._pluginmanager.getplugin("capturemanager")
+        if capman:
+            capman.suspend(in_=True)
+
+        if cls._config:
+            tw = _pytest.config.create_terminal_writer(cls._config)
+            tw.line()
+
+            if cls._recursive_debug == 0:
+                # Handle header similar to pdb.set_trace in py37+.
+                header = kwargs.pop("header", None)
+                if header is not None:
+                    tw.sep(">", header)
+                else:
+                    capturing = cls._is_capturing(capman)
+                    if capturing == "global":
+                        tw.sep(">", f"PDB {method} (IO-capturing turned off)")
+                    elif capturing:
+                        tw.sep(
+                            ">",
+                            f"PDB {method} (IO-capturing turned off for {capturing})",
+                        )
+                    else:
+                        tw.sep(">", f"PDB {method}")
+
+        _pdb = cls._import_pdb_cls(capman)(**kwargs)
+
+        if cls._pluginmanager:
+            cls._pluginmanager.hook.pytest_enter_pdb(config=cls._config, pdb=_pdb)
+        return _pdb
+
+    @classmethod
+    def set_trace(cls, *args, **kwargs) -> None:
+        """Invoke debugging via ``Pdb.set_trace``, dropping any IO capturing."""
+        frame = sys._getframe().f_back
+        _pdb = cls._init_pdb("set_trace", *args, **kwargs)
+        _pdb.set_trace(frame)
+
+
+class PdbInvoke:
+    def pytest_exception_interact(
+        self, node: Node, call: CallInfo[Any], report: BaseReport
+    ) -> None:
+        capman = node.config.pluginmanager.getplugin("capturemanager")
+        if capman:
+            capman.suspend_global_capture(in_=True)
+            out, err = capman.read_global_capture()
+            sys.stdout.write(out)
+            sys.stdout.write(err)
+        assert call.excinfo is not None
+
+        if not isinstance(call.excinfo.value, unittest.SkipTest):
+            _enter_pdb(node, call.excinfo, report)
+
+    def pytest_internalerror(self, excinfo: ExceptionInfo[BaseException]) -> None:
+        tb = _postmortem_traceback(excinfo)
+        post_mortem(tb)
+
+
+class PdbTrace:
+    @hookimpl(wrapper=True)
+    def pytest_pyfunc_call(self, pyfuncitem) -> Generator[None, object, object]:
+        wrap_pytest_function_for_tracing(pyfuncitem)
+        return (yield)
+
+
+def wrap_pytest_function_for_tracing(pyfuncitem) -> None:
+    """Change the Python function object of the given Function item by a
+    wrapper which actually enters pdb before calling the python function
+    itself, effectively leaving the user in the pdb prompt in the first
+    statement of the function."""
+    _pdb = pytestPDB._init_pdb("runcall")
+    testfunction = pyfuncitem.obj
+
+    # we can't just return `partial(pdb.runcall, testfunction)` because (on
+    # python < 3.7.4) runcall's first param is `func`, which means we'd get
+    # an exception if one of the kwargs to testfunction was called `func`.
+    @functools.wraps(testfunction)
+    def wrapper(*args, **kwargs) -> None:
+        func = functools.partial(testfunction, *args, **kwargs)
+        _pdb.runcall(func)
+
+    pyfuncitem.obj = wrapper
+
+
+def maybe_wrap_pytest_function_for_tracing(pyfuncitem) -> None:
+    """Wrap the given pytestfunct item for tracing support if --trace was given in
+    the command line."""
+    if pyfuncitem.config.getvalue("trace"):
+        wrap_pytest_function_for_tracing(pyfuncitem)
+
+
+def _enter_pdb(
+    node: Node, excinfo: ExceptionInfo[BaseException], rep: BaseReport
+) -> BaseReport:
+    # XXX we reuse the TerminalReporter's terminalwriter
+    # because this seems to avoid some encoding related troubles
+    # for not completely clear reasons.
+    tw = node.config.pluginmanager.getplugin("terminalreporter")._tw
+    tw.line()
+
+    showcapture = node.config.option.showcapture
+
+    for sectionname, content in (
+        ("stdout", rep.capstdout),
+        ("stderr", rep.capstderr),
+        ("log", rep.caplog),
+    ):
+        if showcapture in (sectionname, "all") and content:
+            tw.sep(">", "captured " + sectionname)
+            if content[-1:] == "\n":
+                content = content[:-1]
+            tw.line(content)
+
+    tw.sep(">", "traceback")
+    rep.toterminal(tw)
+    tw.sep(">", "entering PDB")
+    tb = _postmortem_traceback(excinfo)
+    rep._pdbshown = True  # type: ignore[attr-defined]
+    post_mortem(tb)
+    return rep
+
+
+def _postmortem_traceback(excinfo: ExceptionInfo[BaseException]) -> types.TracebackType:
+    from doctest import UnexpectedException
+
+    if isinstance(excinfo.value, UnexpectedException):
+        # A doctest.UnexpectedException is not useful for post_mortem.
+        # Use the underlying exception instead:
+        return excinfo.value.exc_info[2]
+    elif isinstance(excinfo.value, ConftestImportFailure):
+        # A config.ConftestImportFailure is not useful for post_mortem.
+        # Use the underlying exception instead:
+        assert excinfo.value.cause.__traceback__ is not None
+        return excinfo.value.cause.__traceback__
+    else:
+        assert excinfo._excinfo is not None
+        return excinfo._excinfo[2]
+
+
+def post_mortem(t: types.TracebackType) -> None:
+    p = pytestPDB._init_pdb("post_mortem")
+    p.reset()
+    p.interaction(None, t)
+    if p.quitting:
+        outcomes.exit("Quitting debugger")
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/deprecated.py b/backend/venv/lib/python3.12/site-packages/_pytest/deprecated.py
new file mode 100644
index 00000000..a605c24e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/deprecated.py
@@ -0,0 +1,91 @@
+"""Deprecation messages and bits of code used elsewhere in the codebase that
+is planned to be removed in the next pytest release.
+
+Keeping it in a central location makes it easy to track what is deprecated and should
+be removed when the time comes.
+
+All constants defined in this module should be either instances of
+:class:`PytestWarning`, or :class:`UnformattedWarning`
+in case of warnings which need to format their messages.
+"""
+
+from __future__ import annotations
+
+from warnings import warn
+
+from _pytest.warning_types import PytestDeprecationWarning
+from _pytest.warning_types import PytestRemovedIn9Warning
+from _pytest.warning_types import UnformattedWarning
+
+
+# set of plugins which have been integrated into the core; we use this list to ignore
+# them during registration to avoid conflicts
+DEPRECATED_EXTERNAL_PLUGINS = {
+    "pytest_catchlog",
+    "pytest_capturelog",
+    "pytest_faulthandler",
+}
+
+
+# This can be* removed pytest 8, but it's harmless and common, so no rush to remove.
+# * If you're in the future: "could have been".
+YIELD_FIXTURE = PytestDeprecationWarning(
+    "@pytest.yield_fixture is deprecated.\n"
+    "Use @pytest.fixture instead; they are the same."
+)
+
+# This deprecation is never really meant to be removed.
+PRIVATE = PytestDeprecationWarning("A private pytest class or function was used.")
+
+
+HOOK_LEGACY_PATH_ARG = UnformattedWarning(
+    PytestRemovedIn9Warning,
+    "The ({pylib_path_arg}: py.path.local) argument is deprecated, please use ({pathlib_path_arg}: pathlib.Path)\n"
+    "see https://docs.pytest.org/en/latest/deprecations.html"
+    "#py-path-local-arguments-for-hooks-replaced-with-pathlib-path",
+)
+
+NODE_CTOR_FSPATH_ARG = UnformattedWarning(
+    PytestRemovedIn9Warning,
+    "The (fspath: py.path.local) argument to {node_type_name} is deprecated. "
+    "Please use the (path: pathlib.Path) argument instead.\n"
+    "See https://docs.pytest.org/en/latest/deprecations.html"
+    "#fspath-argument-for-node-constructors-replaced-with-pathlib-path",
+)
+
+HOOK_LEGACY_MARKING = UnformattedWarning(
+    PytestDeprecationWarning,
+    "The hook{type} {fullname} uses old-style configuration options (marks or attributes).\n"
+    "Please use the pytest.hook{type}({hook_opts}) decorator instead\n"
+    " to configure the hooks.\n"
+    " See https://docs.pytest.org/en/latest/deprecations.html"
+    "#configuring-hook-specs-impls-using-markers",
+)
+
+MARKED_FIXTURE = PytestRemovedIn9Warning(
+    "Marks applied to fixtures have no effect\n"
+    "See docs: https://docs.pytest.org/en/stable/deprecations.html#applying-a-mark-to-a-fixture-function"
+)
+
+# You want to make some `__init__` or function "private".
+#
+#   def my_private_function(some, args):
+#       ...
+#
+# Do this:
+#
+#   def my_private_function(some, args, *, _ispytest: bool = False):
+#       check_ispytest(_ispytest)
+#       ...
+#
+# Change all internal/allowed calls to
+#
+#   my_private_function(some, args, _ispytest=True)
+#
+# All other calls will get the default _ispytest=False and trigger
+# the warning (possibly error in the future).
+
+
+def check_ispytest(ispytest: bool) -> None:
+    if not ispytest:
+        warn(PRIVATE, stacklevel=3)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/doctest.py b/backend/venv/lib/python3.12/site-packages/_pytest/doctest.py
new file mode 100644
index 00000000..384dea97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/doctest.py
@@ -0,0 +1,755 @@
+# mypy: allow-untyped-defs
+"""Discover and run doctests in modules and test files."""
+
+from __future__ import annotations
+
+import bdb
+from contextlib import contextmanager
+import functools
+import inspect
+import os
+from pathlib import Path
+import platform
+import sys
+import traceback
+import types
+from typing import Any
+from typing import Callable
+from typing import Generator
+from typing import Iterable
+from typing import Pattern
+from typing import Sequence
+from typing import TYPE_CHECKING
+import warnings
+
+from _pytest import outcomes
+from _pytest._code.code import ExceptionInfo
+from _pytest._code.code import ReprFileLocation
+from _pytest._code.code import TerminalRepr
+from _pytest._io import TerminalWriter
+from _pytest.compat import safe_getattr
+from _pytest.config import Config
+from _pytest.config.argparsing import Parser
+from _pytest.fixtures import fixture
+from _pytest.fixtures import TopRequest
+from _pytest.nodes import Collector
+from _pytest.nodes import Item
+from _pytest.outcomes import OutcomeException
+from _pytest.outcomes import skip
+from _pytest.pathlib import fnmatch_ex
+from _pytest.python import Module
+from _pytest.python_api import approx
+from _pytest.warning_types import PytestWarning
+
+
+if TYPE_CHECKING:
+    import doctest
+
+    from typing_extensions import Self
+
+DOCTEST_REPORT_CHOICE_NONE = "none"
+DOCTEST_REPORT_CHOICE_CDIFF = "cdiff"
+DOCTEST_REPORT_CHOICE_NDIFF = "ndiff"
+DOCTEST_REPORT_CHOICE_UDIFF = "udiff"
+DOCTEST_REPORT_CHOICE_ONLY_FIRST_FAILURE = "only_first_failure"
+
+DOCTEST_REPORT_CHOICES = (
+    DOCTEST_REPORT_CHOICE_NONE,
+    DOCTEST_REPORT_CHOICE_CDIFF,
+    DOCTEST_REPORT_CHOICE_NDIFF,
+    DOCTEST_REPORT_CHOICE_UDIFF,
+    DOCTEST_REPORT_CHOICE_ONLY_FIRST_FAILURE,
+)
+
+# Lazy definition of runner class
+RUNNER_CLASS = None
+# Lazy definition of output checker class
+CHECKER_CLASS: type[doctest.OutputChecker] | None = None
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini(
+        "doctest_optionflags",
+        "Option flags for doctests",
+        type="args",
+        default=["ELLIPSIS"],
+    )
+    parser.addini(
+        "doctest_encoding", "Encoding used for doctest files", default="utf-8"
+    )
+    group = parser.getgroup("collect")
+    group.addoption(
+        "--doctest-modules",
+        action="store_true",
+        default=False,
+        help="Run doctests in all .py modules",
+        dest="doctestmodules",
+    )
+    group.addoption(
+        "--doctest-report",
+        type=str.lower,
+        default="udiff",
+        help="Choose another output format for diffs on doctest failure",
+        choices=DOCTEST_REPORT_CHOICES,
+        dest="doctestreport",
+    )
+    group.addoption(
+        "--doctest-glob",
+        action="append",
+        default=[],
+        metavar="pat",
+        help="Doctests file matching pattern, default: test*.txt",
+        dest="doctestglob",
+    )
+    group.addoption(
+        "--doctest-ignore-import-errors",
+        action="store_true",
+        default=False,
+        help="Ignore doctest collection errors",
+        dest="doctest_ignore_import_errors",
+    )
+    group.addoption(
+        "--doctest-continue-on-failure",
+        action="store_true",
+        default=False,
+        help="For a given doctest, continue to run after the first failure",
+        dest="doctest_continue_on_failure",
+    )
+
+
+def pytest_unconfigure() -> None:
+    global RUNNER_CLASS
+
+    RUNNER_CLASS = None
+
+
+def pytest_collect_file(
+    file_path: Path,
+    parent: Collector,
+) -> DoctestModule | DoctestTextfile | None:
+    config = parent.config
+    if file_path.suffix == ".py":
+        if config.option.doctestmodules and not any(
+            (_is_setup_py(file_path), _is_main_py(file_path))
+        ):
+            return DoctestModule.from_parent(parent, path=file_path)
+    elif _is_doctest(config, file_path, parent):
+        return DoctestTextfile.from_parent(parent, path=file_path)
+    return None
+
+
+def _is_setup_py(path: Path) -> bool:
+    if path.name != "setup.py":
+        return False
+    contents = path.read_bytes()
+    return b"setuptools" in contents or b"distutils" in contents
+
+
+def _is_doctest(config: Config, path: Path, parent: Collector) -> bool:
+    if path.suffix in (".txt", ".rst") and parent.session.isinitpath(path):
+        return True
+    globs = config.getoption("doctestglob") or ["test*.txt"]
+    return any(fnmatch_ex(glob, path) for glob in globs)
+
+
+def _is_main_py(path: Path) -> bool:
+    return path.name == "__main__.py"
+
+
+class ReprFailDoctest(TerminalRepr):
+    def __init__(
+        self, reprlocation_lines: Sequence[tuple[ReprFileLocation, Sequence[str]]]
+    ) -> None:
+        self.reprlocation_lines = reprlocation_lines
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        for reprlocation, lines in self.reprlocation_lines:
+            for line in lines:
+                tw.line(line)
+            reprlocation.toterminal(tw)
+
+
+class MultipleDoctestFailures(Exception):
+    def __init__(self, failures: Sequence[doctest.DocTestFailure]) -> None:
+        super().__init__()
+        self.failures = failures
+
+
+def _init_runner_class() -> type[doctest.DocTestRunner]:
+    import doctest
+
+    class PytestDoctestRunner(doctest.DebugRunner):
+        """Runner to collect failures.
+
+        Note that the out variable in this case is a list instead of a
+        stdout-like object.
+        """
+
+        def __init__(
+            self,
+            checker: doctest.OutputChecker | None = None,
+            verbose: bool | None = None,
+            optionflags: int = 0,
+            continue_on_failure: bool = True,
+        ) -> None:
+            super().__init__(checker=checker, verbose=verbose, optionflags=optionflags)
+            self.continue_on_failure = continue_on_failure
+
+        def report_failure(
+            self,
+            out,
+            test: doctest.DocTest,
+            example: doctest.Example,
+            got: str,
+        ) -> None:
+            failure = doctest.DocTestFailure(test, example, got)
+            if self.continue_on_failure:
+                out.append(failure)
+            else:
+                raise failure
+
+        def report_unexpected_exception(
+            self,
+            out,
+            test: doctest.DocTest,
+            example: doctest.Example,
+            exc_info: tuple[type[BaseException], BaseException, types.TracebackType],
+        ) -> None:
+            if isinstance(exc_info[1], OutcomeException):
+                raise exc_info[1]
+            if isinstance(exc_info[1], bdb.BdbQuit):
+                outcomes.exit("Quitting debugger")
+            failure = doctest.UnexpectedException(test, example, exc_info)
+            if self.continue_on_failure:
+                out.append(failure)
+            else:
+                raise failure
+
+    return PytestDoctestRunner
+
+
+def _get_runner(
+    checker: doctest.OutputChecker | None = None,
+    verbose: bool | None = None,
+    optionflags: int = 0,
+    continue_on_failure: bool = True,
+) -> doctest.DocTestRunner:
+    # We need this in order to do a lazy import on doctest
+    global RUNNER_CLASS
+    if RUNNER_CLASS is None:
+        RUNNER_CLASS = _init_runner_class()
+    # Type ignored because the continue_on_failure argument is only defined on
+    # PytestDoctestRunner, which is lazily defined so can't be used as a type.
+    return RUNNER_CLASS(  # type: ignore
+        checker=checker,
+        verbose=verbose,
+        optionflags=optionflags,
+        continue_on_failure=continue_on_failure,
+    )
+
+
+class DoctestItem(Item):
+    def __init__(
+        self,
+        name: str,
+        parent: DoctestTextfile | DoctestModule,
+        runner: doctest.DocTestRunner,
+        dtest: doctest.DocTest,
+    ) -> None:
+        super().__init__(name, parent)
+        self.runner = runner
+        self.dtest = dtest
+
+        # Stuff needed for fixture support.
+        self.obj = None
+        fm = self.session._fixturemanager
+        fixtureinfo = fm.getfixtureinfo(node=self, func=None, cls=None)
+        self._fixtureinfo = fixtureinfo
+        self.fixturenames = fixtureinfo.names_closure
+        self._initrequest()
+
+    @classmethod
+    def from_parent(  # type: ignore[override]
+        cls,
+        parent: DoctestTextfile | DoctestModule,
+        *,
+        name: str,
+        runner: doctest.DocTestRunner,
+        dtest: doctest.DocTest,
+    ) -> Self:
+        # incompatible signature due to imposed limits on subclass
+        """The public named constructor."""
+        return super().from_parent(name=name, parent=parent, runner=runner, dtest=dtest)
+
+    def _initrequest(self) -> None:
+        self.funcargs: dict[str, object] = {}
+        self._request = TopRequest(self, _ispytest=True)  # type: ignore[arg-type]
+
+    def setup(self) -> None:
+        self._request._fillfixtures()
+        globs = dict(getfixture=self._request.getfixturevalue)
+        for name, value in self._request.getfixturevalue("doctest_namespace").items():
+            globs[name] = value
+        self.dtest.globs.update(globs)
+
+    def runtest(self) -> None:
+        _check_all_skipped(self.dtest)
+        self._disable_output_capturing_for_darwin()
+        failures: list[doctest.DocTestFailure] = []
+        # Type ignored because we change the type of `out` from what
+        # doctest expects.
+        self.runner.run(self.dtest, out=failures)  # type: ignore[arg-type]
+        if failures:
+            raise MultipleDoctestFailures(failures)
+
+    def _disable_output_capturing_for_darwin(self) -> None:
+        """Disable output capturing. Otherwise, stdout is lost to doctest (#985)."""
+        if platform.system() != "Darwin":
+            return
+        capman = self.config.pluginmanager.getplugin("capturemanager")
+        if capman:
+            capman.suspend_global_capture(in_=True)
+            out, err = capman.read_global_capture()
+            sys.stdout.write(out)
+            sys.stderr.write(err)
+
+    # TODO: Type ignored -- breaks Liskov Substitution.
+    def repr_failure(  # type: ignore[override]
+        self,
+        excinfo: ExceptionInfo[BaseException],
+    ) -> str | TerminalRepr:
+        import doctest
+
+        failures: (
+            Sequence[doctest.DocTestFailure | doctest.UnexpectedException] | None
+        ) = None
+        if isinstance(
+            excinfo.value, (doctest.DocTestFailure, doctest.UnexpectedException)
+        ):
+            failures = [excinfo.value]
+        elif isinstance(excinfo.value, MultipleDoctestFailures):
+            failures = excinfo.value.failures
+
+        if failures is None:
+            return super().repr_failure(excinfo)
+
+        reprlocation_lines = []
+        for failure in failures:
+            example = failure.example
+            test = failure.test
+            filename = test.filename
+            if test.lineno is None:
+                lineno = None
+            else:
+                lineno = test.lineno + example.lineno + 1
+            message = type(failure).__name__
+            # TODO: ReprFileLocation doesn't expect a None lineno.
+            reprlocation = ReprFileLocation(filename, lineno, message)  # type: ignore[arg-type]
+            checker = _get_checker()
+            report_choice = _get_report_choice(self.config.getoption("doctestreport"))
+            if lineno is not None:
+                assert failure.test.docstring is not None
+                lines = failure.test.docstring.splitlines(False)
+                # add line numbers to the left of the error message
+                assert test.lineno is not None
+                lines = [
+                    "%03d %s" % (i + test.lineno + 1, x) for (i, x) in enumerate(lines)
+                ]
+                # trim docstring error lines to 10
+                lines = lines[max(example.lineno - 9, 0) : example.lineno + 1]
+            else:
+                lines = [
+                    "EXAMPLE LOCATION UNKNOWN, not showing all tests of that example"
+                ]
+                indent = ">>>"
+                for line in example.source.splitlines():
+                    lines.append(f"??? {indent} {line}")
+                    indent = "..."
+            if isinstance(failure, doctest.DocTestFailure):
+                lines += checker.output_difference(
+                    example, failure.got, report_choice
+                ).split("\n")
+            else:
+                inner_excinfo = ExceptionInfo.from_exc_info(failure.exc_info)
+                lines += [f"UNEXPECTED EXCEPTION: {inner_excinfo.value!r}"]
+                lines += [
+                    x.strip("\n") for x in traceback.format_exception(*failure.exc_info)
+                ]
+            reprlocation_lines.append((reprlocation, lines))
+        return ReprFailDoctest(reprlocation_lines)
+
+    def reportinfo(self) -> tuple[os.PathLike[str] | str, int | None, str]:
+        return self.path, self.dtest.lineno, f"[doctest] {self.name}"
+
+
+def _get_flag_lookup() -> dict[str, int]:
+    import doctest
+
+    return dict(
+        DONT_ACCEPT_TRUE_FOR_1=doctest.DONT_ACCEPT_TRUE_FOR_1,
+        DONT_ACCEPT_BLANKLINE=doctest.DONT_ACCEPT_BLANKLINE,
+        NORMALIZE_WHITESPACE=doctest.NORMALIZE_WHITESPACE,
+        ELLIPSIS=doctest.ELLIPSIS,
+        IGNORE_EXCEPTION_DETAIL=doctest.IGNORE_EXCEPTION_DETAIL,
+        COMPARISON_FLAGS=doctest.COMPARISON_FLAGS,
+        ALLOW_UNICODE=_get_allow_unicode_flag(),
+        ALLOW_BYTES=_get_allow_bytes_flag(),
+        NUMBER=_get_number_flag(),
+    )
+
+
+def get_optionflags(config: Config) -> int:
+    optionflags_str = config.getini("doctest_optionflags")
+    flag_lookup_table = _get_flag_lookup()
+    flag_acc = 0
+    for flag in optionflags_str:
+        flag_acc |= flag_lookup_table[flag]
+    return flag_acc
+
+
+def _get_continue_on_failure(config: Config) -> bool:
+    continue_on_failure: bool = config.getvalue("doctest_continue_on_failure")
+    if continue_on_failure:
+        # We need to turn off this if we use pdb since we should stop at
+        # the first failure.
+        if config.getvalue("usepdb"):
+            continue_on_failure = False
+    return continue_on_failure
+
+
+class DoctestTextfile(Module):
+    obj = None
+
+    def collect(self) -> Iterable[DoctestItem]:
+        import doctest
+
+        # Inspired by doctest.testfile; ideally we would use it directly,
+        # but it doesn't support passing a custom checker.
+        encoding = self.config.getini("doctest_encoding")
+        text = self.path.read_text(encoding)
+        filename = str(self.path)
+        name = self.path.name
+        globs = {"__name__": "__main__"}
+
+        optionflags = get_optionflags(self.config)
+
+        runner = _get_runner(
+            verbose=False,
+            optionflags=optionflags,
+            checker=_get_checker(),
+            continue_on_failure=_get_continue_on_failure(self.config),
+        )
+
+        parser = doctest.DocTestParser()
+        test = parser.get_doctest(text, globs, name, filename, 0)
+        if test.examples:
+            yield DoctestItem.from_parent(
+                self, name=test.name, runner=runner, dtest=test
+            )
+
+
+def _check_all_skipped(test: doctest.DocTest) -> None:
+    """Raise pytest.skip() if all examples in the given DocTest have the SKIP
+    option set."""
+    import doctest
+
+    all_skipped = all(x.options.get(doctest.SKIP, False) for x in test.examples)
+    if all_skipped:
+        skip("all tests skipped by +SKIP option")
+
+
+def _is_mocked(obj: object) -> bool:
+    """Return if an object is possibly a mock object by checking the
+    existence of a highly improbable attribute."""
+    return (
+        safe_getattr(obj, "pytest_mock_example_attribute_that_shouldnt_exist", None)
+        is not None
+    )
+
+
+@contextmanager
+def _patch_unwrap_mock_aware() -> Generator[None]:
+    """Context manager which replaces ``inspect.unwrap`` with a version
+    that's aware of mock objects and doesn't recurse into them."""
+    real_unwrap = inspect.unwrap
+
+    def _mock_aware_unwrap(
+        func: Callable[..., Any], *, stop: Callable[[Any], Any] | None = None
+    ) -> Any:
+        try:
+            if stop is None or stop is _is_mocked:
+                return real_unwrap(func, stop=_is_mocked)
+            _stop = stop
+            return real_unwrap(func, stop=lambda obj: _is_mocked(obj) or _stop(func))
+        except Exception as e:
+            warnings.warn(
+                f"Got {e!r} when unwrapping {func!r}.  This is usually caused "
+                "by a violation of Python's object protocol; see e.g. "
+                "https://github.com/pytest-dev/pytest/issues/5080",
+                PytestWarning,
+            )
+            raise
+
+    inspect.unwrap = _mock_aware_unwrap
+    try:
+        yield
+    finally:
+        inspect.unwrap = real_unwrap
+
+
+class DoctestModule(Module):
+    def collect(self) -> Iterable[DoctestItem]:
+        import doctest
+
+        class MockAwareDocTestFinder(doctest.DocTestFinder):
+            py_ver_info_minor = sys.version_info[:2]
+            is_find_lineno_broken = (
+                py_ver_info_minor < (3, 11)
+                or (py_ver_info_minor == (3, 11) and sys.version_info.micro < 9)
+                or (py_ver_info_minor == (3, 12) and sys.version_info.micro < 3)
+            )
+            if is_find_lineno_broken:
+
+                def _find_lineno(self, obj, source_lines):
+                    """On older Pythons, doctest code does not take into account
+                    `@property`. https://github.com/python/cpython/issues/61648
+
+                    Moreover, wrapped Doctests need to be unwrapped so the correct
+                    line number is returned. #8796
+                    """
+                    if isinstance(obj, property):
+                        obj = getattr(obj, "fget", obj)
+
+                    if hasattr(obj, "__wrapped__"):
+                        # Get the main obj in case of it being wrapped
+                        obj = inspect.unwrap(obj)
+
+                    # Type ignored because this is a private function.
+                    return super()._find_lineno(  # type:ignore[misc]
+                        obj,
+                        source_lines,
+                    )
+
+            if sys.version_info < (3, 10):
+
+                def _find(
+                    self, tests, obj, name, module, source_lines, globs, seen
+                ) -> None:
+                    """Override _find to work around issue in stdlib.
+
+                    https://github.com/pytest-dev/pytest/issues/3456
+                    https://github.com/python/cpython/issues/69718
+                    """
+                    if _is_mocked(obj):
+                        return  # pragma: no cover
+                    with _patch_unwrap_mock_aware():
+                        # Type ignored because this is a private function.
+                        super()._find(  # type:ignore[misc]
+                            tests, obj, name, module, source_lines, globs, seen
+                        )
+
+            if sys.version_info < (3, 13):
+
+                def _from_module(self, module, object):
+                    """`cached_property` objects are never considered a part
+                    of the 'current module'. As such they are skipped by doctest.
+                    Here we override `_from_module` to check the underlying
+                    function instead. https://github.com/python/cpython/issues/107995
+                    """
+                    if isinstance(object, functools.cached_property):
+                        object = object.func
+
+                    # Type ignored because this is a private function.
+                    return super()._from_module(module, object)  # type: ignore[misc]
+
+        try:
+            module = self.obj
+        except Collector.CollectError:
+            if self.config.getvalue("doctest_ignore_import_errors"):
+                skip(f"unable to import module {self.path!r}")
+            else:
+                raise
+
+        # While doctests currently don't support fixtures directly, we still
+        # need to pick up autouse fixtures.
+        self.session._fixturemanager.parsefactories(self)
+
+        # Uses internal doctest module parsing mechanism.
+        finder = MockAwareDocTestFinder()
+        optionflags = get_optionflags(self.config)
+        runner = _get_runner(
+            verbose=False,
+            optionflags=optionflags,
+            checker=_get_checker(),
+            continue_on_failure=_get_continue_on_failure(self.config),
+        )
+
+        for test in finder.find(module, module.__name__):
+            if test.examples:  # skip empty doctests
+                yield DoctestItem.from_parent(
+                    self, name=test.name, runner=runner, dtest=test
+                )
+
+
+def _init_checker_class() -> type[doctest.OutputChecker]:
+    import doctest
+    import re
+
+    class LiteralsOutputChecker(doctest.OutputChecker):
+        # Based on doctest_nose_plugin.py from the nltk project
+        # (https://github.com/nltk/nltk) and on the "numtest" doctest extension
+        # by Sebastien Boisgerault (https://github.com/boisgera/numtest).
+
+        _unicode_literal_re = re.compile(r"(\W|^)[uU]([rR]?[\'\"])", re.UNICODE)
+        _bytes_literal_re = re.compile(r"(\W|^)[bB]([rR]?[\'\"])", re.UNICODE)
+        _number_re = re.compile(
+            r"""
+            (?P<number>
+              (?P<mantissa>
+                (?P<integer1> [+-]?\d*)\.(?P<fraction>\d+)
+                |
+                (?P<integer2> [+-]?\d+)\.
+              )
+              (?:
+                [Ee]
+                (?P<exponent1> [+-]?\d+)
+              )?
+              |
+              (?P<integer3> [+-]?\d+)
+              (?:
+                [Ee]
+                (?P<exponent2> [+-]?\d+)
+              )
+            )
+            """,
+            re.VERBOSE,
+        )
+
+        def check_output(self, want: str, got: str, optionflags: int) -> bool:
+            if super().check_output(want, got, optionflags):
+                return True
+
+            allow_unicode = optionflags & _get_allow_unicode_flag()
+            allow_bytes = optionflags & _get_allow_bytes_flag()
+            allow_number = optionflags & _get_number_flag()
+
+            if not allow_unicode and not allow_bytes and not allow_number:
+                return False
+
+            def remove_prefixes(regex: Pattern[str], txt: str) -> str:
+                return re.sub(regex, r"\1\2", txt)
+
+            if allow_unicode:
+                want = remove_prefixes(self._unicode_literal_re, want)
+                got = remove_prefixes(self._unicode_literal_re, got)
+
+            if allow_bytes:
+                want = remove_prefixes(self._bytes_literal_re, want)
+                got = remove_prefixes(self._bytes_literal_re, got)
+
+            if allow_number:
+                got = self._remove_unwanted_precision(want, got)
+
+            return super().check_output(want, got, optionflags)
+
+        def _remove_unwanted_precision(self, want: str, got: str) -> str:
+            wants = list(self._number_re.finditer(want))
+            gots = list(self._number_re.finditer(got))
+            if len(wants) != len(gots):
+                return got
+            offset = 0
+            for w, g in zip(wants, gots):
+                fraction: str | None = w.group("fraction")
+                exponent: str | None = w.group("exponent1")
+                if exponent is None:
+                    exponent = w.group("exponent2")
+                precision = 0 if fraction is None else len(fraction)
+                if exponent is not None:
+                    precision -= int(exponent)
+                if float(w.group()) == approx(float(g.group()), abs=10**-precision):
+                    # They're close enough. Replace the text we actually
+                    # got with the text we want, so that it will match when we
+                    # check the string literally.
+                    got = (
+                        got[: g.start() + offset] + w.group() + got[g.end() + offset :]
+                    )
+                    offset += w.end() - w.start() - (g.end() - g.start())
+            return got
+
+    return LiteralsOutputChecker
+
+
+def _get_checker() -> doctest.OutputChecker:
+    """Return a doctest.OutputChecker subclass that supports some
+    additional options:
+
+    * ALLOW_UNICODE and ALLOW_BYTES options to ignore u'' and b''
+      prefixes (respectively) in string literals. Useful when the same
+      doctest should run in Python 2 and Python 3.
+
+    * NUMBER to ignore floating-point differences smaller than the
+      precision of the literal number in the doctest.
+
+    An inner class is used to avoid importing "doctest" at the module
+    level.
+    """
+    global CHECKER_CLASS
+    if CHECKER_CLASS is None:
+        CHECKER_CLASS = _init_checker_class()
+    return CHECKER_CLASS()
+
+
+def _get_allow_unicode_flag() -> int:
+    """Register and return the ALLOW_UNICODE flag."""
+    import doctest
+
+    return doctest.register_optionflag("ALLOW_UNICODE")
+
+
+def _get_allow_bytes_flag() -> int:
+    """Register and return the ALLOW_BYTES flag."""
+    import doctest
+
+    return doctest.register_optionflag("ALLOW_BYTES")
+
+
+def _get_number_flag() -> int:
+    """Register and return the NUMBER flag."""
+    import doctest
+
+    return doctest.register_optionflag("NUMBER")
+
+
+def _get_report_choice(key: str) -> int:
+    """Return the actual `doctest` module flag value.
+
+    We want to do it as late as possible to avoid importing `doctest` and all
+    its dependencies when parsing options, as it adds overhead and breaks tests.
+    """
+    import doctest
+
+    return {
+        DOCTEST_REPORT_CHOICE_UDIFF: doctest.REPORT_UDIFF,
+        DOCTEST_REPORT_CHOICE_CDIFF: doctest.REPORT_CDIFF,
+        DOCTEST_REPORT_CHOICE_NDIFF: doctest.REPORT_NDIFF,
+        DOCTEST_REPORT_CHOICE_ONLY_FIRST_FAILURE: doctest.REPORT_ONLY_FIRST_FAILURE,
+        DOCTEST_REPORT_CHOICE_NONE: 0,
+    }[key]
+
+
+@fixture(scope="session")
+def doctest_namespace() -> dict[str, Any]:
+    """Fixture that returns a :py:class:`dict` that will be injected into the
+    namespace of doctests.
+
+    Usually this fixture is used in conjunction with another ``autouse`` fixture:
+
+    .. code-block:: python
+
+        @pytest.fixture(autouse=True)
+        def add_np(doctest_namespace):
+            doctest_namespace["np"] = numpy
+
+    For more details: :ref:`doctest_namespace`.
+    """
+    return dict()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/faulthandler.py b/backend/venv/lib/python3.12/site-packages/_pytest/faulthandler.py
new file mode 100644
index 00000000..d16aea1e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/faulthandler.py
@@ -0,0 +1,105 @@
+from __future__ import annotations
+
+import os
+import sys
+from typing import Generator
+
+from _pytest.config import Config
+from _pytest.config.argparsing import Parser
+from _pytest.nodes import Item
+from _pytest.stash import StashKey
+import pytest
+
+
+fault_handler_original_stderr_fd_key = StashKey[int]()
+fault_handler_stderr_fd_key = StashKey[int]()
+
+
+def pytest_addoption(parser: Parser) -> None:
+    help = (
+        "Dump the traceback of all threads if a test takes "
+        "more than TIMEOUT seconds to finish"
+    )
+    parser.addini("faulthandler_timeout", help, default=0.0)
+
+
+def pytest_configure(config: Config) -> None:
+    import faulthandler
+
+    # at teardown we want to restore the original faulthandler fileno
+    # but faulthandler has no api to return the original fileno
+    # so here we stash the stderr fileno to be used at teardown
+    # sys.stderr and sys.__stderr__ may be closed or patched during the session
+    # so we can't rely on their values being good at that point (#11572).
+    stderr_fileno = get_stderr_fileno()
+    if faulthandler.is_enabled():
+        config.stash[fault_handler_original_stderr_fd_key] = stderr_fileno
+    config.stash[fault_handler_stderr_fd_key] = os.dup(stderr_fileno)
+    faulthandler.enable(file=config.stash[fault_handler_stderr_fd_key])
+
+
+def pytest_unconfigure(config: Config) -> None:
+    import faulthandler
+
+    faulthandler.disable()
+    # Close the dup file installed during pytest_configure.
+    if fault_handler_stderr_fd_key in config.stash:
+        os.close(config.stash[fault_handler_stderr_fd_key])
+        del config.stash[fault_handler_stderr_fd_key]
+    # Re-enable the faulthandler if it was originally enabled.
+    if fault_handler_original_stderr_fd_key in config.stash:
+        faulthandler.enable(config.stash[fault_handler_original_stderr_fd_key])
+        del config.stash[fault_handler_original_stderr_fd_key]
+
+
+def get_stderr_fileno() -> int:
+    try:
+        fileno = sys.stderr.fileno()
+        # The Twisted Logger will return an invalid file descriptor since it is not backed
+        # by an FD. So, let's also forward this to the same code path as with pytest-xdist.
+        if fileno == -1:
+            raise AttributeError()
+        return fileno
+    except (AttributeError, ValueError):
+        # pytest-xdist monkeypatches sys.stderr with an object that is not an actual file.
+        # https://docs.python.org/3/library/faulthandler.html#issue-with-file-descriptors
+        # This is potentially dangerous, but the best we can do.
+        assert sys.__stderr__ is not None
+        return sys.__stderr__.fileno()
+
+
+def get_timeout_config_value(config: Config) -> float:
+    return float(config.getini("faulthandler_timeout") or 0.0)
+
+
+@pytest.hookimpl(wrapper=True, trylast=True)
+def pytest_runtest_protocol(item: Item) -> Generator[None, object, object]:
+    timeout = get_timeout_config_value(item.config)
+    if timeout > 0:
+        import faulthandler
+
+        stderr = item.config.stash[fault_handler_stderr_fd_key]
+        faulthandler.dump_traceback_later(timeout, file=stderr)
+        try:
+            return (yield)
+        finally:
+            faulthandler.cancel_dump_traceback_later()
+    else:
+        return (yield)
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_enter_pdb() -> None:
+    """Cancel any traceback dumping due to timeout before entering pdb."""
+    import faulthandler
+
+    faulthandler.cancel_dump_traceback_later()
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_exception_interact() -> None:
+    """Cancel any traceback dumping due to an interactive exception being
+    raised."""
+    import faulthandler
+
+    faulthandler.cancel_dump_traceback_later()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/fixtures.py b/backend/venv/lib/python3.12/site-packages/_pytest/fixtures.py
new file mode 100644
index 00000000..6b882fa3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/fixtures.py
@@ -0,0 +1,1932 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import abc
+from collections import defaultdict
+from collections import deque
+import dataclasses
+import functools
+import inspect
+import os
+from pathlib import Path
+import sys
+import types
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Final
+from typing import final
+from typing import Generator
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import Mapping
+from typing import MutableMapping
+from typing import NoReturn
+from typing import Optional
+from typing import OrderedDict
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import warnings
+
+import _pytest
+from _pytest import nodes
+from _pytest._code import getfslineno
+from _pytest._code import Source
+from _pytest._code.code import FormattedExcinfo
+from _pytest._code.code import TerminalRepr
+from _pytest._io import TerminalWriter
+from _pytest.compat import _PytestWrapper
+from _pytest.compat import assert_never
+from _pytest.compat import get_real_func
+from _pytest.compat import get_real_method
+from _pytest.compat import getfuncargnames
+from _pytest.compat import getimfunc
+from _pytest.compat import getlocation
+from _pytest.compat import is_generator
+from _pytest.compat import NOTSET
+from _pytest.compat import NotSetType
+from _pytest.compat import safe_getattr
+from _pytest.compat import safe_isclass
+from _pytest.config import _PluggyPlugin
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.deprecated import MARKED_FIXTURE
+from _pytest.deprecated import YIELD_FIXTURE
+from _pytest.main import Session
+from _pytest.mark import Mark
+from _pytest.mark import ParameterSet
+from _pytest.mark.structures import MarkDecorator
+from _pytest.outcomes import fail
+from _pytest.outcomes import skip
+from _pytest.outcomes import TEST_OUTCOME
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import bestrelpath
+from _pytest.scope import _ScopeName
+from _pytest.scope import HIGH_SCOPES
+from _pytest.scope import Scope
+
+
+if sys.version_info < (3, 11):
+    from exceptiongroup import BaseExceptionGroup
+
+
+if TYPE_CHECKING:
+    from _pytest.python import CallSpec2
+    from _pytest.python import Function
+    from _pytest.python import Metafunc
+
+
+# The value of the fixture -- return/yield of the fixture function (type variable).
+FixtureValue = TypeVar("FixtureValue")
+# The type of the fixture function (type variable).
+FixtureFunction = TypeVar("FixtureFunction", bound=Callable[..., object])
+# The type of a fixture function (type alias generic in fixture value).
+_FixtureFunc = Union[
+    Callable[..., FixtureValue], Callable[..., Generator[FixtureValue, None, None]]
+]
+# The type of FixtureDef.cached_result (type alias generic in fixture value).
+_FixtureCachedResult = Union[
+    Tuple[
+        # The result.
+        FixtureValue,
+        # Cache key.
+        object,
+        None,
+    ],
+    Tuple[
+        None,
+        # Cache key.
+        object,
+        # The exception and the original traceback.
+        Tuple[BaseException, Optional[types.TracebackType]],
+    ],
+]
+
+
+@dataclasses.dataclass(frozen=True)
+class PseudoFixtureDef(Generic[FixtureValue]):
+    cached_result: _FixtureCachedResult[FixtureValue]
+    _scope: Scope
+
+
+def pytest_sessionstart(session: Session) -> None:
+    session._fixturemanager = FixtureManager(session)
+
+
+def get_scope_package(
+    node: nodes.Item,
+    fixturedef: FixtureDef[object],
+) -> nodes.Node | None:
+    from _pytest.python import Package
+
+    for parent in node.iter_parents():
+        if isinstance(parent, Package) and parent.nodeid == fixturedef.baseid:
+            return parent
+    return node.session
+
+
+def get_scope_node(node: nodes.Node, scope: Scope) -> nodes.Node | None:
+    import _pytest.python
+
+    if scope is Scope.Function:
+        # Type ignored because this is actually safe, see:
+        # https://github.com/python/mypy/issues/4717
+        return node.getparent(nodes.Item)  # type: ignore[type-abstract]
+    elif scope is Scope.Class:
+        return node.getparent(_pytest.python.Class)
+    elif scope is Scope.Module:
+        return node.getparent(_pytest.python.Module)
+    elif scope is Scope.Package:
+        return node.getparent(_pytest.python.Package)
+    elif scope is Scope.Session:
+        return node.getparent(_pytest.main.Session)
+    else:
+        assert_never(scope)
+
+
+def getfixturemarker(obj: object) -> FixtureFunctionMarker | None:
+    """Return fixturemarker or None if it doesn't exist or raised
+    exceptions."""
+    return cast(
+        Optional[FixtureFunctionMarker],
+        safe_getattr(obj, "_pytestfixturefunction", None),
+    )
+
+
+# Algorithm for sorting on a per-parametrized resource setup basis.
+# It is called for Session scope first and performs sorting
+# down to the lower scopes such as to minimize number of "high scope"
+# setups and teardowns.
+
+
+@dataclasses.dataclass(frozen=True)
+class FixtureArgKey:
+    argname: str
+    param_index: int
+    scoped_item_path: Path | None
+    item_cls: type | None
+
+
+_V = TypeVar("_V")
+OrderedSet = Dict[_V, None]
+
+
+def get_parametrized_fixture_argkeys(
+    item: nodes.Item, scope: Scope
+) -> Iterator[FixtureArgKey]:
+    """Return list of keys for all parametrized arguments which match
+    the specified scope."""
+    assert scope is not Scope.Function
+
+    try:
+        callspec: CallSpec2 = item.callspec  # type: ignore[attr-defined]
+    except AttributeError:
+        return
+
+    item_cls = None
+    if scope is Scope.Session:
+        scoped_item_path = None
+    elif scope is Scope.Package:
+        # Package key = module's directory.
+        scoped_item_path = item.path.parent
+    elif scope is Scope.Module:
+        scoped_item_path = item.path
+    elif scope is Scope.Class:
+        scoped_item_path = item.path
+        item_cls = item.cls  # type: ignore[attr-defined]
+    else:
+        assert_never(scope)
+
+    for argname in callspec.indices:
+        if callspec._arg2scope[argname] != scope:
+            continue
+        param_index = callspec.indices[argname]
+        yield FixtureArgKey(argname, param_index, scoped_item_path, item_cls)
+
+
+def reorder_items(items: Sequence[nodes.Item]) -> list[nodes.Item]:
+    argkeys_by_item: dict[Scope, dict[nodes.Item, OrderedSet[FixtureArgKey]]] = {}
+    items_by_argkey: dict[
+        Scope, dict[FixtureArgKey, OrderedDict[nodes.Item, None]]
+    ] = {}
+    for scope in HIGH_SCOPES:
+        scoped_argkeys_by_item = argkeys_by_item[scope] = {}
+        scoped_items_by_argkey = items_by_argkey[scope] = defaultdict(OrderedDict)
+        for item in items:
+            argkeys = dict.fromkeys(get_parametrized_fixture_argkeys(item, scope))
+            if argkeys:
+                scoped_argkeys_by_item[item] = argkeys
+                for argkey in argkeys:
+                    scoped_items_by_argkey[argkey][item] = None
+
+    items_set = dict.fromkeys(items)
+    return list(
+        reorder_items_atscope(
+            items_set, argkeys_by_item, items_by_argkey, Scope.Session
+        )
+    )
+
+
+def reorder_items_atscope(
+    items: OrderedSet[nodes.Item],
+    argkeys_by_item: Mapping[Scope, Mapping[nodes.Item, OrderedSet[FixtureArgKey]]],
+    items_by_argkey: Mapping[
+        Scope, Mapping[FixtureArgKey, OrderedDict[nodes.Item, None]]
+    ],
+    scope: Scope,
+) -> OrderedSet[nodes.Item]:
+    if scope is Scope.Function or len(items) < 3:
+        return items
+
+    scoped_items_by_argkey = items_by_argkey[scope]
+    scoped_argkeys_by_item = argkeys_by_item[scope]
+
+    ignore: set[FixtureArgKey] = set()
+    items_deque = deque(items)
+    items_done: OrderedSet[nodes.Item] = {}
+    while items_deque:
+        no_argkey_items: OrderedSet[nodes.Item] = {}
+        slicing_argkey = None
+        while items_deque:
+            item = items_deque.popleft()
+            if item in items_done or item in no_argkey_items:
+                continue
+            argkeys = dict.fromkeys(
+                k for k in scoped_argkeys_by_item.get(item, ()) if k not in ignore
+            )
+            if not argkeys:
+                no_argkey_items[item] = None
+            else:
+                slicing_argkey, _ = argkeys.popitem()
+                # We don't have to remove relevant items from later in the
+                # deque because they'll just be ignored.
+                matching_items = [
+                    i for i in scoped_items_by_argkey[slicing_argkey] if i in items
+                ]
+                for i in reversed(matching_items):
+                    items_deque.appendleft(i)
+                    # Fix items_by_argkey order.
+                    for other_scope in HIGH_SCOPES:
+                        other_scoped_items_by_argkey = items_by_argkey[other_scope]
+                        for argkey in argkeys_by_item[other_scope].get(i, ()):
+                            other_scoped_items_by_argkey[argkey][i] = None
+                            other_scoped_items_by_argkey[argkey].move_to_end(
+                                i, last=False
+                            )
+                break
+        if no_argkey_items:
+            reordered_no_argkey_items = reorder_items_atscope(
+                no_argkey_items, argkeys_by_item, items_by_argkey, scope.next_lower()
+            )
+            items_done.update(reordered_no_argkey_items)
+        if slicing_argkey is not None:
+            ignore.add(slicing_argkey)
+    return items_done
+
+
+@dataclasses.dataclass(frozen=True)
+class FuncFixtureInfo:
+    """Fixture-related information for a fixture-requesting item (e.g. test
+    function).
+
+    This is used to examine the fixtures which an item requests statically
+    (known during collection). This includes autouse fixtures, fixtures
+    requested by the `usefixtures` marker, fixtures requested in the function
+    parameters, and the transitive closure of these.
+
+    An item may also request fixtures dynamically (using `request.getfixturevalue`);
+    these are not reflected here.
+    """
+
+    __slots__ = ("argnames", "initialnames", "names_closure", "name2fixturedefs")
+
+    # Fixture names that the item requests directly by function parameters.
+    argnames: tuple[str, ...]
+    # Fixture names that the item immediately requires. These include
+    # argnames + fixture names specified via usefixtures and via autouse=True in
+    # fixture definitions.
+    initialnames: tuple[str, ...]
+    # The transitive closure of the fixture names that the item requires.
+    # Note: can't include dynamic dependencies (`request.getfixturevalue` calls).
+    names_closure: list[str]
+    # A map from a fixture name in the transitive closure to the FixtureDefs
+    # matching the name which are applicable to this function.
+    # There may be multiple overriding fixtures with the same name. The
+    # sequence is ordered from furthest to closes to the function.
+    name2fixturedefs: dict[str, Sequence[FixtureDef[Any]]]
+
+    def prune_dependency_tree(self) -> None:
+        """Recompute names_closure from initialnames and name2fixturedefs.
+
+        Can only reduce names_closure, which means that the new closure will
+        always be a subset of the old one. The order is preserved.
+
+        This method is needed because direct parametrization may shadow some
+        of the fixtures that were included in the originally built dependency
+        tree. In this way the dependency tree can get pruned, and the closure
+        of argnames may get reduced.
+        """
+        closure: set[str] = set()
+        working_set = set(self.initialnames)
+        while working_set:
+            argname = working_set.pop()
+            # Argname may be something not included in the original names_closure,
+            # in which case we ignore it. This currently happens with pseudo
+            # FixtureDefs which wrap 'get_direct_param_fixture_func(request)'.
+            # So they introduce the new dependency 'request' which might have
+            # been missing in the original tree (closure).
+            if argname not in closure and argname in self.names_closure:
+                closure.add(argname)
+                if argname in self.name2fixturedefs:
+                    working_set.update(self.name2fixturedefs[argname][-1].argnames)
+
+        self.names_closure[:] = sorted(closure, key=self.names_closure.index)
+
+
+class FixtureRequest(abc.ABC):
+    """The type of the ``request`` fixture.
+
+    A request object gives access to the requesting test context and has a
+    ``param`` attribute in case the fixture is parametrized.
+    """
+
+    def __init__(
+        self,
+        pyfuncitem: Function,
+        fixturename: str | None,
+        arg2fixturedefs: dict[str, Sequence[FixtureDef[Any]]],
+        fixture_defs: dict[str, FixtureDef[Any]],
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        #: Fixture for which this request is being performed.
+        self.fixturename: Final = fixturename
+        self._pyfuncitem: Final = pyfuncitem
+        # The FixtureDefs for each fixture name requested by this item.
+        # Starts from the statically-known fixturedefs resolved during
+        # collection. Dynamically requested fixtures (using
+        # `request.getfixturevalue("foo")`) are added dynamically.
+        self._arg2fixturedefs: Final = arg2fixturedefs
+        # The evaluated argnames so far, mapping to the FixtureDef they resolved
+        # to.
+        self._fixture_defs: Final = fixture_defs
+        # Notes on the type of `param`:
+        # -`request.param` is only defined in parametrized fixtures, and will raise
+        #   AttributeError otherwise. Python typing has no notion of "undefined", so
+        #   this cannot be reflected in the type.
+        # - Technically `param` is only (possibly) defined on SubRequest, not
+        #   FixtureRequest, but the typing of that is still in flux so this cheats.
+        # - In the future we might consider using a generic for the param type, but
+        #   for now just using Any.
+        self.param: Any
+
+    @property
+    def _fixturemanager(self) -> FixtureManager:
+        return self._pyfuncitem.session._fixturemanager
+
+    @property
+    @abc.abstractmethod
+    def _scope(self) -> Scope:
+        raise NotImplementedError()
+
+    @property
+    def scope(self) -> _ScopeName:
+        """Scope string, one of "function", "class", "module", "package", "session"."""
+        return self._scope.value
+
+    @abc.abstractmethod
+    def _check_scope(
+        self,
+        requested_fixturedef: FixtureDef[object] | PseudoFixtureDef[object],
+        requested_scope: Scope,
+    ) -> None:
+        raise NotImplementedError()
+
+    @property
+    def fixturenames(self) -> list[str]:
+        """Names of all active fixtures in this request."""
+        result = list(self._pyfuncitem.fixturenames)
+        result.extend(set(self._fixture_defs).difference(result))
+        return result
+
+    @property
+    @abc.abstractmethod
+    def node(self):
+        """Underlying collection node (depends on current request scope)."""
+        raise NotImplementedError()
+
+    @property
+    def config(self) -> Config:
+        """The pytest config object associated with this request."""
+        return self._pyfuncitem.config
+
+    @property
+    def function(self):
+        """Test function object if the request has a per-function scope."""
+        if self.scope != "function":
+            raise AttributeError(
+                f"function not available in {self.scope}-scoped context"
+            )
+        return self._pyfuncitem.obj
+
+    @property
+    def cls(self):
+        """Class (can be None) where the test function was collected."""
+        if self.scope not in ("class", "function"):
+            raise AttributeError(f"cls not available in {self.scope}-scoped context")
+        clscol = self._pyfuncitem.getparent(_pytest.python.Class)
+        if clscol:
+            return clscol.obj
+
+    @property
+    def instance(self):
+        """Instance (can be None) on which test function was collected."""
+        if self.scope != "function":
+            return None
+        return getattr(self._pyfuncitem, "instance", None)
+
+    @property
+    def module(self):
+        """Python module object where the test function was collected."""
+        if self.scope not in ("function", "class", "module"):
+            raise AttributeError(f"module not available in {self.scope}-scoped context")
+        mod = self._pyfuncitem.getparent(_pytest.python.Module)
+        assert mod is not None
+        return mod.obj
+
+    @property
+    def path(self) -> Path:
+        """Path where the test function was collected."""
+        if self.scope not in ("function", "class", "module", "package"):
+            raise AttributeError(f"path not available in {self.scope}-scoped context")
+        return self._pyfuncitem.path
+
+    @property
+    def keywords(self) -> MutableMapping[str, Any]:
+        """Keywords/markers dictionary for the underlying node."""
+        node: nodes.Node = self.node
+        return node.keywords
+
+    @property
+    def session(self) -> Session:
+        """Pytest session object."""
+        return self._pyfuncitem.session
+
+    @abc.abstractmethod
+    def addfinalizer(self, finalizer: Callable[[], object]) -> None:
+        """Add finalizer/teardown function to be called without arguments after
+        the last test within the requesting test context finished execution."""
+        raise NotImplementedError()
+
+    def applymarker(self, marker: str | MarkDecorator) -> None:
+        """Apply a marker to a single test function invocation.
+
+        This method is useful if you don't want to have a keyword/marker
+        on all function invocations.
+
+        :param marker:
+            An object created by a call to ``pytest.mark.NAME(...)``.
+        """
+        self.node.add_marker(marker)
+
+    def raiseerror(self, msg: str | None) -> NoReturn:
+        """Raise a FixtureLookupError exception.
+
+        :param msg:
+            An optional custom error message.
+        """
+        raise FixtureLookupError(None, self, msg)
+
+    def getfixturevalue(self, argname: str) -> Any:
+        """Dynamically run a named fixture function.
+
+        Declaring fixtures via function argument is recommended where possible.
+        But if you can only decide whether to use another fixture at test
+        setup time, you may use this function to retrieve it inside a fixture
+        or test function body.
+
+        This method can be used during the test setup phase or the test run
+        phase, but during the test teardown phase a fixture's value may not
+        be available.
+
+        :param argname:
+            The fixture name.
+        :raises pytest.FixtureLookupError:
+            If the given fixture could not be found.
+        """
+        # Note that in addition to the use case described in the docstring,
+        # getfixturevalue() is also called by pytest itself during item and fixture
+        # setup to evaluate the fixtures that are requested statically
+        # (using function parameters, autouse, etc).
+
+        fixturedef = self._get_active_fixturedef(argname)
+        assert fixturedef.cached_result is not None, (
+            f'The fixture value for "{argname}" is not available.  '
+            "This can happen when the fixture has already been torn down."
+        )
+        return fixturedef.cached_result[0]
+
+    def _iter_chain(self) -> Iterator[SubRequest]:
+        """Yield all SubRequests in the chain, from self up.
+
+        Note: does *not* yield the TopRequest.
+        """
+        current = self
+        while isinstance(current, SubRequest):
+            yield current
+            current = current._parent_request
+
+    def _get_active_fixturedef(
+        self, argname: str
+    ) -> FixtureDef[object] | PseudoFixtureDef[object]:
+        if argname == "request":
+            cached_result = (self, [0], None)
+            return PseudoFixtureDef(cached_result, Scope.Function)
+
+        # If we already finished computing a fixture by this name in this item,
+        # return it.
+        fixturedef = self._fixture_defs.get(argname)
+        if fixturedef is not None:
+            self._check_scope(fixturedef, fixturedef._scope)
+            return fixturedef
+
+        # Find the appropriate fixturedef.
+        fixturedefs = self._arg2fixturedefs.get(argname, None)
+        if fixturedefs is None:
+            # We arrive here because of a dynamic call to
+            # getfixturevalue(argname) which was naturally
+            # not known at parsing/collection time.
+            fixturedefs = self._fixturemanager.getfixturedefs(argname, self._pyfuncitem)
+            if fixturedefs is not None:
+                self._arg2fixturedefs[argname] = fixturedefs
+        # No fixtures defined with this name.
+        if fixturedefs is None:
+            raise FixtureLookupError(argname, self)
+        # The are no fixtures with this name applicable for the function.
+        if not fixturedefs:
+            raise FixtureLookupError(argname, self)
+        # A fixture may override another fixture with the same name, e.g. a
+        # fixture in a module can override a fixture in a conftest, a fixture in
+        # a class can override a fixture in the module, and so on.
+        # An overriding fixture can request its own name (possibly indirectly);
+        # in this case it gets the value of the fixture it overrides, one level
+        # up.
+        # Check how many `argname`s deep we are, and take the next one.
+        # `fixturedefs` is sorted from furthest to closest, so use negative
+        # indexing to go in reverse.
+        index = -1
+        for request in self._iter_chain():
+            if request.fixturename == argname:
+                index -= 1
+        # If already consumed all of the available levels, fail.
+        if -index > len(fixturedefs):
+            raise FixtureLookupError(argname, self)
+        fixturedef = fixturedefs[index]
+
+        # Prepare a SubRequest object for calling the fixture.
+        try:
+            callspec = self._pyfuncitem.callspec
+        except AttributeError:
+            callspec = None
+        if callspec is not None and argname in callspec.params:
+            param = callspec.params[argname]
+            param_index = callspec.indices[argname]
+            # The parametrize invocation scope overrides the fixture's scope.
+            scope = callspec._arg2scope[argname]
+        else:
+            param = NOTSET
+            param_index = 0
+            scope = fixturedef._scope
+            self._check_fixturedef_without_param(fixturedef)
+        self._check_scope(fixturedef, scope)
+        subrequest = SubRequest(
+            self, scope, param, param_index, fixturedef, _ispytest=True
+        )
+
+        # Make sure the fixture value is cached, running it if it isn't
+        fixturedef.execute(request=subrequest)
+
+        self._fixture_defs[argname] = fixturedef
+        return fixturedef
+
+    def _check_fixturedef_without_param(self, fixturedef: FixtureDef[object]) -> None:
+        """Check that this request is allowed to execute this fixturedef without
+        a param."""
+        funcitem = self._pyfuncitem
+        has_params = fixturedef.params is not None
+        fixtures_not_supported = getattr(funcitem, "nofuncargs", False)
+        if has_params and fixtures_not_supported:
+            msg = (
+                f"{funcitem.name} does not support fixtures, maybe unittest.TestCase subclass?\n"
+                f"Node id: {funcitem.nodeid}\n"
+                f"Function type: {type(funcitem).__name__}"
+            )
+            fail(msg, pytrace=False)
+        if has_params:
+            frame = inspect.stack()[3]
+            frameinfo = inspect.getframeinfo(frame[0])
+            source_path = absolutepath(frameinfo.filename)
+            source_lineno = frameinfo.lineno
+            try:
+                source_path_str = str(source_path.relative_to(funcitem.config.rootpath))
+            except ValueError:
+                source_path_str = str(source_path)
+            location = getlocation(fixturedef.func, funcitem.config.rootpath)
+            msg = (
+                "The requested fixture has no parameter defined for test:\n"
+                f"    {funcitem.nodeid}\n\n"
+                f"Requested fixture '{fixturedef.argname}' defined in:\n"
+                f"{location}\n\n"
+                f"Requested here:\n"
+                f"{source_path_str}:{source_lineno}"
+            )
+            fail(msg, pytrace=False)
+
+    def _get_fixturestack(self) -> list[FixtureDef[Any]]:
+        values = [request._fixturedef for request in self._iter_chain()]
+        values.reverse()
+        return values
+
+
+@final
+class TopRequest(FixtureRequest):
+    """The type of the ``request`` fixture in a test function."""
+
+    def __init__(self, pyfuncitem: Function, *, _ispytest: bool = False) -> None:
+        super().__init__(
+            fixturename=None,
+            pyfuncitem=pyfuncitem,
+            arg2fixturedefs=pyfuncitem._fixtureinfo.name2fixturedefs.copy(),
+            fixture_defs={},
+            _ispytest=_ispytest,
+        )
+
+    @property
+    def _scope(self) -> Scope:
+        return Scope.Function
+
+    def _check_scope(
+        self,
+        requested_fixturedef: FixtureDef[object] | PseudoFixtureDef[object],
+        requested_scope: Scope,
+    ) -> None:
+        # TopRequest always has function scope so always valid.
+        pass
+
+    @property
+    def node(self):
+        return self._pyfuncitem
+
+    def __repr__(self) -> str:
+        return f"<FixtureRequest for {self.node!r}>"
+
+    def _fillfixtures(self) -> None:
+        item = self._pyfuncitem
+        for argname in item.fixturenames:
+            if argname not in item.funcargs:
+                item.funcargs[argname] = self.getfixturevalue(argname)
+
+    def addfinalizer(self, finalizer: Callable[[], object]) -> None:
+        self.node.addfinalizer(finalizer)
+
+
+@final
+class SubRequest(FixtureRequest):
+    """The type of the ``request`` fixture in a fixture function requested
+    (transitively) by a test function."""
+
+    def __init__(
+        self,
+        request: FixtureRequest,
+        scope: Scope,
+        param: Any,
+        param_index: int,
+        fixturedef: FixtureDef[object],
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        super().__init__(
+            pyfuncitem=request._pyfuncitem,
+            fixturename=fixturedef.argname,
+            fixture_defs=request._fixture_defs,
+            arg2fixturedefs=request._arg2fixturedefs,
+            _ispytest=_ispytest,
+        )
+        self._parent_request: Final[FixtureRequest] = request
+        self._scope_field: Final = scope
+        self._fixturedef: Final[FixtureDef[object]] = fixturedef
+        if param is not NOTSET:
+            self.param = param
+        self.param_index: Final = param_index
+
+    def __repr__(self) -> str:
+        return f"<SubRequest {self.fixturename!r} for {self._pyfuncitem!r}>"
+
+    @property
+    def _scope(self) -> Scope:
+        return self._scope_field
+
+    @property
+    def node(self):
+        scope = self._scope
+        if scope is Scope.Function:
+            # This might also be a non-function Item despite its attribute name.
+            node: nodes.Node | None = self._pyfuncitem
+        elif scope is Scope.Package:
+            node = get_scope_package(self._pyfuncitem, self._fixturedef)
+        else:
+            node = get_scope_node(self._pyfuncitem, scope)
+        if node is None and scope is Scope.Class:
+            # Fallback to function item itself.
+            node = self._pyfuncitem
+        assert node, f'Could not obtain a node for scope "{scope}" for function {self._pyfuncitem!r}'
+        return node
+
+    def _check_scope(
+        self,
+        requested_fixturedef: FixtureDef[object] | PseudoFixtureDef[object],
+        requested_scope: Scope,
+    ) -> None:
+        if isinstance(requested_fixturedef, PseudoFixtureDef):
+            return
+        if self._scope > requested_scope:
+            # Try to report something helpful.
+            argname = requested_fixturedef.argname
+            fixture_stack = "\n".join(
+                self._format_fixturedef_line(fixturedef)
+                for fixturedef in self._get_fixturestack()
+            )
+            requested_fixture = self._format_fixturedef_line(requested_fixturedef)
+            fail(
+                f"ScopeMismatch: You tried to access the {requested_scope.value} scoped "
+                f"fixture {argname} with a {self._scope.value} scoped request object. "
+                f"Requesting fixture stack:\n{fixture_stack}\n"
+                f"Requested fixture:\n{requested_fixture}",
+                pytrace=False,
+            )
+
+    def _format_fixturedef_line(self, fixturedef: FixtureDef[object]) -> str:
+        factory = fixturedef.func
+        path, lineno = getfslineno(factory)
+        if isinstance(path, Path):
+            path = bestrelpath(self._pyfuncitem.session.path, path)
+        signature = inspect.signature(factory)
+        return f"{path}:{lineno + 1}:  def {factory.__name__}{signature}"
+
+    def addfinalizer(self, finalizer: Callable[[], object]) -> None:
+        self._fixturedef.addfinalizer(finalizer)
+
+
+@final
+class FixtureLookupError(LookupError):
+    """Could not return a requested fixture (missing or invalid)."""
+
+    def __init__(
+        self, argname: str | None, request: FixtureRequest, msg: str | None = None
+    ) -> None:
+        self.argname = argname
+        self.request = request
+        self.fixturestack = request._get_fixturestack()
+        self.msg = msg
+
+    def formatrepr(self) -> FixtureLookupErrorRepr:
+        tblines: list[str] = []
+        addline = tblines.append
+        stack = [self.request._pyfuncitem.obj]
+        stack.extend(map(lambda x: x.func, self.fixturestack))
+        msg = self.msg
+        if msg is not None:
+            # The last fixture raise an error, let's present
+            # it at the requesting side.
+            stack = stack[:-1]
+        for function in stack:
+            fspath, lineno = getfslineno(function)
+            try:
+                lines, _ = inspect.getsourcelines(get_real_func(function))
+            except (OSError, IndexError, TypeError):
+                error_msg = "file %s, line %s: source code not available"
+                addline(error_msg % (fspath, lineno + 1))
+            else:
+                addline(f"file {fspath}, line {lineno + 1}")
+                for i, line in enumerate(lines):
+                    line = line.rstrip()
+                    addline("  " + line)
+                    if line.lstrip().startswith("def"):
+                        break
+
+        if msg is None:
+            fm = self.request._fixturemanager
+            available = set()
+            parent = self.request._pyfuncitem.parent
+            assert parent is not None
+            for name, fixturedefs in fm._arg2fixturedefs.items():
+                faclist = list(fm._matchfactories(fixturedefs, parent))
+                if faclist:
+                    available.add(name)
+            if self.argname in available:
+                msg = (
+                    f" recursive dependency involving fixture '{self.argname}' detected"
+                )
+            else:
+                msg = f"fixture '{self.argname}' not found"
+            msg += "\n available fixtures: {}".format(", ".join(sorted(available)))
+            msg += "\n use 'pytest --fixtures [testpath]' for help on them."
+
+        return FixtureLookupErrorRepr(fspath, lineno, tblines, msg, self.argname)
+
+
+class FixtureLookupErrorRepr(TerminalRepr):
+    def __init__(
+        self,
+        filename: str | os.PathLike[str],
+        firstlineno: int,
+        tblines: Sequence[str],
+        errorstring: str,
+        argname: str | None,
+    ) -> None:
+        self.tblines = tblines
+        self.errorstring = errorstring
+        self.filename = filename
+        self.firstlineno = firstlineno
+        self.argname = argname
+
+    def toterminal(self, tw: TerminalWriter) -> None:
+        # tw.line("FixtureLookupError: %s" %(self.argname), red=True)
+        for tbline in self.tblines:
+            tw.line(tbline.rstrip())
+        lines = self.errorstring.split("\n")
+        if lines:
+            tw.line(
+                f"{FormattedExcinfo.fail_marker}       {lines[0].strip()}",
+                red=True,
+            )
+            for line in lines[1:]:
+                tw.line(
+                    f"{FormattedExcinfo.flow_marker}       {line.strip()}",
+                    red=True,
+                )
+        tw.line()
+        tw.line("%s:%d" % (os.fspath(self.filename), self.firstlineno + 1))
+
+
+def call_fixture_func(
+    fixturefunc: _FixtureFunc[FixtureValue], request: FixtureRequest, kwargs
+) -> FixtureValue:
+    if is_generator(fixturefunc):
+        fixturefunc = cast(
+            Callable[..., Generator[FixtureValue, None, None]], fixturefunc
+        )
+        generator = fixturefunc(**kwargs)
+        try:
+            fixture_result = next(generator)
+        except StopIteration:
+            raise ValueError(f"{request.fixturename} did not yield a value") from None
+        finalizer = functools.partial(_teardown_yield_fixture, fixturefunc, generator)
+        request.addfinalizer(finalizer)
+    else:
+        fixturefunc = cast(Callable[..., FixtureValue], fixturefunc)
+        fixture_result = fixturefunc(**kwargs)
+    return fixture_result
+
+
+def _teardown_yield_fixture(fixturefunc, it) -> None:
+    """Execute the teardown of a fixture function by advancing the iterator
+    after the yield and ensure the iteration ends (if not it means there is
+    more than one yield in the function)."""
+    try:
+        next(it)
+    except StopIteration:
+        pass
+    else:
+        fs, lineno = getfslineno(fixturefunc)
+        fail(
+            f"fixture function has more than one 'yield':\n\n"
+            f"{Source(fixturefunc).indent()}\n"
+            f"{fs}:{lineno + 1}",
+            pytrace=False,
+        )
+
+
+def _eval_scope_callable(
+    scope_callable: Callable[[str, Config], _ScopeName],
+    fixture_name: str,
+    config: Config,
+) -> _ScopeName:
+    try:
+        # Type ignored because there is no typing mechanism to specify
+        # keyword arguments, currently.
+        result = scope_callable(fixture_name=fixture_name, config=config)  # type: ignore[call-arg]
+    except Exception as e:
+        raise TypeError(
+            f"Error evaluating {scope_callable} while defining fixture '{fixture_name}'.\n"
+            "Expected a function with the signature (*, fixture_name, config)"
+        ) from e
+    if not isinstance(result, str):
+        fail(
+            f"Expected {scope_callable} to return a 'str' while defining fixture '{fixture_name}', but it returned:\n"
+            f"{result!r}",
+            pytrace=False,
+        )
+    return result
+
+
+@final
+class FixtureDef(Generic[FixtureValue]):
+    """A container for a fixture definition.
+
+    Note: At this time, only explicitly documented fields and methods are
+    considered public stable API.
+    """
+
+    def __init__(
+        self,
+        config: Config,
+        baseid: str | None,
+        argname: str,
+        func: _FixtureFunc[FixtureValue],
+        scope: Scope | _ScopeName | Callable[[str, Config], _ScopeName] | None,
+        params: Sequence[object] | None,
+        ids: tuple[object | None, ...] | Callable[[Any], object | None] | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        # The "base" node ID for the fixture.
+        #
+        # This is a node ID prefix. A fixture is only available to a node (e.g.
+        # a `Function` item) if the fixture's baseid is a nodeid of a parent of
+        # node.
+        #
+        # For a fixture found in a Collector's object (e.g. a `Module`s module,
+        # a `Class`'s class), the baseid is the Collector's nodeid.
+        #
+        # For a fixture found in a conftest plugin, the baseid is the conftest's
+        # directory path relative to the rootdir.
+        #
+        # For other plugins, the baseid is the empty string (always matches).
+        self.baseid: Final = baseid or ""
+        # Whether the fixture was found from a node or a conftest in the
+        # collection tree. Will be false for fixtures defined in non-conftest
+        # plugins.
+        self.has_location: Final = baseid is not None
+        # The fixture factory function.
+        self.func: Final = func
+        # The name by which the fixture may be requested.
+        self.argname: Final = argname
+        if scope is None:
+            scope = Scope.Function
+        elif callable(scope):
+            scope = _eval_scope_callable(scope, argname, config)
+        if isinstance(scope, str):
+            scope = Scope.from_user(
+                scope, descr=f"Fixture '{func.__name__}'", where=baseid
+            )
+        self._scope: Final = scope
+        # If the fixture is directly parametrized, the parameter values.
+        self.params: Final = params
+        # If the fixture is directly parametrized, a tuple of explicit IDs to
+        # assign to the parameter values, or a callable to generate an ID given
+        # a parameter value.
+        self.ids: Final = ids
+        # The names requested by the fixtures.
+        self.argnames: Final = getfuncargnames(func, name=argname)
+        # If the fixture was executed, the current value of the fixture.
+        # Can change if the fixture is executed with different parameters.
+        self.cached_result: _FixtureCachedResult[FixtureValue] | None = None
+        self._finalizers: Final[list[Callable[[], object]]] = []
+
+    @property
+    def scope(self) -> _ScopeName:
+        """Scope string, one of "function", "class", "module", "package", "session"."""
+        return self._scope.value
+
+    def addfinalizer(self, finalizer: Callable[[], object]) -> None:
+        self._finalizers.append(finalizer)
+
+    def finish(self, request: SubRequest) -> None:
+        exceptions: list[BaseException] = []
+        while self._finalizers:
+            fin = self._finalizers.pop()
+            try:
+                fin()
+            except BaseException as e:
+                exceptions.append(e)
+        node = request.node
+        node.ihook.pytest_fixture_post_finalizer(fixturedef=self, request=request)
+        # Even if finalization fails, we invalidate the cached fixture
+        # value and remove all finalizers because they may be bound methods
+        # which will keep instances alive.
+        self.cached_result = None
+        self._finalizers.clear()
+        if len(exceptions) == 1:
+            raise exceptions[0]
+        elif len(exceptions) > 1:
+            msg = f'errors while tearing down fixture "{self.argname}" of {node}'
+            raise BaseExceptionGroup(msg, exceptions[::-1])
+
+    def execute(self, request: SubRequest) -> FixtureValue:
+        """Return the value of this fixture, executing it if not cached."""
+        # Ensure that the dependent fixtures requested by this fixture are loaded.
+        # This needs to be done before checking if we have a cached value, since
+        # if a dependent fixture has their cache invalidated, e.g. due to
+        # parametrization, they finalize themselves and fixtures depending on it
+        # (which will likely include this fixture) setting `self.cached_result = None`.
+        # See #4871
+        requested_fixtures_that_should_finalize_us = []
+        for argname in self.argnames:
+            fixturedef = request._get_active_fixturedef(argname)
+            # Saves requested fixtures in a list so we later can add our finalizer
+            # to them, ensuring that if a requested fixture gets torn down we get torn
+            # down first. This is generally handled by SetupState, but still currently
+            # needed when this fixture is not parametrized but depends on a parametrized
+            # fixture.
+            if not isinstance(fixturedef, PseudoFixtureDef):
+                requested_fixtures_that_should_finalize_us.append(fixturedef)
+
+        # Check for (and return) cached value/exception.
+        if self.cached_result is not None:
+            request_cache_key = self.cache_key(request)
+            cache_key = self.cached_result[1]
+            try:
+                # Attempt to make a normal == check: this might fail for objects
+                # which do not implement the standard comparison (like numpy arrays -- #6497).
+                cache_hit = bool(request_cache_key == cache_key)
+            except (ValueError, RuntimeError):
+                # If the comparison raises, use 'is' as fallback.
+                cache_hit = request_cache_key is cache_key
+
+            if cache_hit:
+                if self.cached_result[2] is not None:
+                    exc, exc_tb = self.cached_result[2]
+                    raise exc.with_traceback(exc_tb)
+                else:
+                    result = self.cached_result[0]
+                    return result
+            # We have a previous but differently parametrized fixture instance
+            # so we need to tear it down before creating a new one.
+            self.finish(request)
+            assert self.cached_result is None
+
+        # Add finalizer to requested fixtures we saved previously.
+        # We make sure to do this after checking for cached value to avoid
+        # adding our finalizer multiple times. (#12135)
+        finalizer = functools.partial(self.finish, request=request)
+        for parent_fixture in requested_fixtures_that_should_finalize_us:
+            parent_fixture.addfinalizer(finalizer)
+
+        ihook = request.node.ihook
+        try:
+            # Setup the fixture, run the code in it, and cache the value
+            # in self.cached_result
+            result = ihook.pytest_fixture_setup(fixturedef=self, request=request)
+        finally:
+            # schedule our finalizer, even if the setup failed
+            request.node.addfinalizer(finalizer)
+
+        return result
+
+    def cache_key(self, request: SubRequest) -> object:
+        return getattr(request, "param", None)
+
+    def __repr__(self) -> str:
+        return f"<FixtureDef argname={self.argname!r} scope={self.scope!r} baseid={self.baseid!r}>"
+
+
+def resolve_fixture_function(
+    fixturedef: FixtureDef[FixtureValue], request: FixtureRequest
+) -> _FixtureFunc[FixtureValue]:
+    """Get the actual callable that can be called to obtain the fixture
+    value."""
+    fixturefunc = fixturedef.func
+    # The fixture function needs to be bound to the actual
+    # request.instance so that code working with "fixturedef" behaves
+    # as expected.
+    instance = request.instance
+    if instance is not None:
+        # Handle the case where fixture is defined not in a test class, but some other class
+        # (for example a plugin class with a fixture), see #2270.
+        if hasattr(fixturefunc, "__self__") and not isinstance(
+            instance,
+            fixturefunc.__self__.__class__,
+        ):
+            return fixturefunc
+        fixturefunc = getimfunc(fixturedef.func)
+        if fixturefunc != fixturedef.func:
+            fixturefunc = fixturefunc.__get__(instance)
+    return fixturefunc
+
+
+def pytest_fixture_setup(
+    fixturedef: FixtureDef[FixtureValue], request: SubRequest
+) -> FixtureValue:
+    """Execution of fixture setup."""
+    kwargs = {}
+    for argname in fixturedef.argnames:
+        kwargs[argname] = request.getfixturevalue(argname)
+
+    fixturefunc = resolve_fixture_function(fixturedef, request)
+    my_cache_key = fixturedef.cache_key(request)
+    try:
+        result = call_fixture_func(fixturefunc, request, kwargs)
+    except TEST_OUTCOME as e:
+        if isinstance(e, skip.Exception):
+            # The test requested a fixture which caused a skip.
+            # Don't show the fixture as the skip location, as then the user
+            # wouldn't know which test skipped.
+            e._use_item_location = True
+        fixturedef.cached_result = (None, my_cache_key, (e, e.__traceback__))
+        raise
+    fixturedef.cached_result = (result, my_cache_key, None)
+    return result
+
+
+def wrap_function_to_error_out_if_called_directly(
+    function: FixtureFunction,
+    fixture_marker: FixtureFunctionMarker,
+) -> FixtureFunction:
+    """Wrap the given fixture function so we can raise an error about it being called directly,
+    instead of used as an argument in a test function."""
+    name = fixture_marker.name or function.__name__
+    message = (
+        f'Fixture "{name}" called directly. Fixtures are not meant to be called directly,\n'
+        "but are created automatically when test functions request them as parameters.\n"
+        "See https://docs.pytest.org/en/stable/explanation/fixtures.html for more information about fixtures, and\n"
+        "https://docs.pytest.org/en/stable/deprecations.html#calling-fixtures-directly about how to update your code."
+    )
+
+    @functools.wraps(function)
+    def result(*args, **kwargs):
+        fail(message, pytrace=False)
+
+    # Keep reference to the original function in our own custom attribute so we don't unwrap
+    # further than this point and lose useful wrappings like @mock.patch (#3774).
+    result.__pytest_wrapped__ = _PytestWrapper(function)  # type: ignore[attr-defined]
+
+    return cast(FixtureFunction, result)
+
+
+@final
+@dataclasses.dataclass(frozen=True)
+class FixtureFunctionMarker:
+    scope: _ScopeName | Callable[[str, Config], _ScopeName]
+    params: tuple[object, ...] | None
+    autouse: bool = False
+    ids: tuple[object | None, ...] | Callable[[Any], object | None] | None = None
+    name: str | None = None
+
+    _ispytest: dataclasses.InitVar[bool] = False
+
+    def __post_init__(self, _ispytest: bool) -> None:
+        check_ispytest(_ispytest)
+
+    def __call__(self, function: FixtureFunction) -> FixtureFunction:
+        if inspect.isclass(function):
+            raise ValueError("class fixtures not supported (maybe in the future)")
+
+        if getattr(function, "_pytestfixturefunction", False):
+            raise ValueError(
+                f"@pytest.fixture is being applied more than once to the same function {function.__name__!r}"
+            )
+
+        if hasattr(function, "pytestmark"):
+            warnings.warn(MARKED_FIXTURE, stacklevel=2)
+
+        function = wrap_function_to_error_out_if_called_directly(function, self)
+
+        name = self.name or function.__name__
+        if name == "request":
+            location = getlocation(function)
+            fail(
+                f"'request' is a reserved word for fixtures, use another name:\n  {location}",
+                pytrace=False,
+            )
+
+        # Type ignored because https://github.com/python/mypy/issues/2087.
+        function._pytestfixturefunction = self  # type: ignore[attr-defined]
+        return function
+
+
+@overload
+def fixture(
+    fixture_function: FixtureFunction,
+    *,
+    scope: _ScopeName | Callable[[str, Config], _ScopeName] = ...,
+    params: Iterable[object] | None = ...,
+    autouse: bool = ...,
+    ids: Sequence[object | None] | Callable[[Any], object | None] | None = ...,
+    name: str | None = ...,
+) -> FixtureFunction: ...
+
+
+@overload
+def fixture(
+    fixture_function: None = ...,
+    *,
+    scope: _ScopeName | Callable[[str, Config], _ScopeName] = ...,
+    params: Iterable[object] | None = ...,
+    autouse: bool = ...,
+    ids: Sequence[object | None] | Callable[[Any], object | None] | None = ...,
+    name: str | None = None,
+) -> FixtureFunctionMarker: ...
+
+
+def fixture(
+    fixture_function: FixtureFunction | None = None,
+    *,
+    scope: _ScopeName | Callable[[str, Config], _ScopeName] = "function",
+    params: Iterable[object] | None = None,
+    autouse: bool = False,
+    ids: Sequence[object | None] | Callable[[Any], object | None] | None = None,
+    name: str | None = None,
+) -> FixtureFunctionMarker | FixtureFunction:
+    """Decorator to mark a fixture factory function.
+
+    This decorator can be used, with or without parameters, to define a
+    fixture function.
+
+    The name of the fixture function can later be referenced to cause its
+    invocation ahead of running tests: test modules or classes can use the
+    ``pytest.mark.usefixtures(fixturename)`` marker.
+
+    Test functions can directly use fixture names as input arguments in which
+    case the fixture instance returned from the fixture function will be
+    injected.
+
+    Fixtures can provide their values to test functions using ``return`` or
+    ``yield`` statements. When using ``yield`` the code block after the
+    ``yield`` statement is executed as teardown code regardless of the test
+    outcome, and must yield exactly once.
+
+    :param scope:
+        The scope for which this fixture is shared; one of ``"function"``
+        (default), ``"class"``, ``"module"``, ``"package"`` or ``"session"``.
+
+        This parameter may also be a callable which receives ``(fixture_name, config)``
+        as parameters, and must return a ``str`` with one of the values mentioned above.
+
+        See :ref:`dynamic scope` in the docs for more information.
+
+    :param params:
+        An optional list of parameters which will cause multiple invocations
+        of the fixture function and all of the tests using it. The current
+        parameter is available in ``request.param``.
+
+    :param autouse:
+        If True, the fixture func is activated for all tests that can see it.
+        If False (the default), an explicit reference is needed to activate
+        the fixture.
+
+    :param ids:
+        Sequence of ids each corresponding to the params so that they are
+        part of the test id. If no ids are provided they will be generated
+        automatically from the params.
+
+    :param name:
+        The name of the fixture. This defaults to the name of the decorated
+        function. If a fixture is used in the same module in which it is
+        defined, the function name of the fixture will be shadowed by the
+        function arg that requests the fixture; one way to resolve this is to
+        name the decorated function ``fixture_<fixturename>`` and then use
+        ``@pytest.fixture(name='<fixturename>')``.
+    """
+    fixture_marker = FixtureFunctionMarker(
+        scope=scope,
+        params=tuple(params) if params is not None else None,
+        autouse=autouse,
+        ids=None if ids is None else ids if callable(ids) else tuple(ids),
+        name=name,
+        _ispytest=True,
+    )
+
+    # Direct decoration.
+    if fixture_function:
+        return fixture_marker(fixture_function)
+
+    return fixture_marker
+
+
+def yield_fixture(
+    fixture_function=None,
+    *args,
+    scope="function",
+    params=None,
+    autouse=False,
+    ids=None,
+    name=None,
+):
+    """(Return a) decorator to mark a yield-fixture factory function.
+
+    .. deprecated:: 3.0
+        Use :py:func:`pytest.fixture` directly instead.
+    """
+    warnings.warn(YIELD_FIXTURE, stacklevel=2)
+    return fixture(
+        fixture_function,
+        *args,
+        scope=scope,
+        params=params,
+        autouse=autouse,
+        ids=ids,
+        name=name,
+    )
+
+
+@fixture(scope="session")
+def pytestconfig(request: FixtureRequest) -> Config:
+    """Session-scoped fixture that returns the session's :class:`pytest.Config`
+    object.
+
+    Example::
+
+        def test_foo(pytestconfig):
+            if pytestconfig.get_verbosity() > 0:
+                ...
+
+    """
+    return request.config
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini(
+        "usefixtures",
+        type="args",
+        default=[],
+        help="List of default fixtures to be used with this project",
+    )
+    group = parser.getgroup("general")
+    group.addoption(
+        "--fixtures",
+        "--funcargs",
+        action="store_true",
+        dest="showfixtures",
+        default=False,
+        help="Show available fixtures, sorted by plugin appearance "
+        "(fixtures with leading '_' are only shown with '-v')",
+    )
+    group.addoption(
+        "--fixtures-per-test",
+        action="store_true",
+        dest="show_fixtures_per_test",
+        default=False,
+        help="Show fixtures per test",
+    )
+
+
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    if config.option.showfixtures:
+        showfixtures(config)
+        return 0
+    if config.option.show_fixtures_per_test:
+        show_fixtures_per_test(config)
+        return 0
+    return None
+
+
+def _get_direct_parametrize_args(node: nodes.Node) -> set[str]:
+    """Return all direct parametrization arguments of a node, so we don't
+    mistake them for fixtures.
+
+    Check https://github.com/pytest-dev/pytest/issues/5036.
+
+    These things are done later as well when dealing with parametrization
+    so this could be improved.
+    """
+    parametrize_argnames: set[str] = set()
+    for marker in node.iter_markers(name="parametrize"):
+        if not marker.kwargs.get("indirect", False):
+            p_argnames, _ = ParameterSet._parse_parametrize_args(
+                *marker.args, **marker.kwargs
+            )
+            parametrize_argnames.update(p_argnames)
+    return parametrize_argnames
+
+
+def deduplicate_names(*seqs: Iterable[str]) -> tuple[str, ...]:
+    """De-duplicate the sequence of names while keeping the original order."""
+    # Ideally we would use a set, but it does not preserve insertion order.
+    return tuple(dict.fromkeys(name for seq in seqs for name in seq))
+
+
+class FixtureManager:
+    """pytest fixture definitions and information is stored and managed
+    from this class.
+
+    During collection fm.parsefactories() is called multiple times to parse
+    fixture function definitions into FixtureDef objects and internal
+    data structures.
+
+    During collection of test functions, metafunc-mechanics instantiate
+    a FuncFixtureInfo object which is cached per node/func-name.
+    This FuncFixtureInfo object is later retrieved by Function nodes
+    which themselves offer a fixturenames attribute.
+
+    The FuncFixtureInfo object holds information about fixtures and FixtureDefs
+    relevant for a particular function. An initial list of fixtures is
+    assembled like this:
+
+    - ini-defined usefixtures
+    - autouse-marked fixtures along the collection chain up from the function
+    - usefixtures markers at module/class/function level
+    - test function funcargs
+
+    Subsequently the funcfixtureinfo.fixturenames attribute is computed
+    as the closure of the fixtures needed to setup the initial fixtures,
+    i.e. fixtures needed by fixture functions themselves are appended
+    to the fixturenames list.
+
+    Upon the test-setup phases all fixturenames are instantiated, retrieved
+    by a lookup of their FuncFixtureInfo.
+    """
+
+    def __init__(self, session: Session) -> None:
+        self.session = session
+        self.config: Config = session.config
+        # Maps a fixture name (argname) to all of the FixtureDefs in the test
+        # suite/plugins defined with this name. Populated by parsefactories().
+        # TODO: The order of the FixtureDefs list of each arg is significant,
+        #       explain.
+        self._arg2fixturedefs: Final[dict[str, list[FixtureDef[Any]]]] = {}
+        self._holderobjseen: Final[set[object]] = set()
+        # A mapping from a nodeid to a list of autouse fixtures it defines.
+        self._nodeid_autousenames: Final[dict[str, list[str]]] = {
+            "": self.config.getini("usefixtures"),
+        }
+        session.config.pluginmanager.register(self, "funcmanage")
+
+    def getfixtureinfo(
+        self,
+        node: nodes.Item,
+        func: Callable[..., object] | None,
+        cls: type | None,
+    ) -> FuncFixtureInfo:
+        """Calculate the :class:`FuncFixtureInfo` for an item.
+
+        If ``func`` is None, or if the item sets an attribute
+        ``nofuncargs = True``, then ``func`` is not examined at all.
+
+        :param node:
+            The item requesting the fixtures.
+        :param func:
+            The item's function.
+        :param cls:
+            If the function is a method, the method's class.
+        """
+        if func is not None and not getattr(node, "nofuncargs", False):
+            argnames = getfuncargnames(func, name=node.name, cls=cls)
+        else:
+            argnames = ()
+        usefixturesnames = self._getusefixturesnames(node)
+        autousenames = self._getautousenames(node)
+        initialnames = deduplicate_names(autousenames, usefixturesnames, argnames)
+
+        direct_parametrize_args = _get_direct_parametrize_args(node)
+
+        names_closure, arg2fixturedefs = self.getfixtureclosure(
+            parentnode=node,
+            initialnames=initialnames,
+            ignore_args=direct_parametrize_args,
+        )
+
+        return FuncFixtureInfo(argnames, initialnames, names_closure, arg2fixturedefs)
+
+    def pytest_plugin_registered(self, plugin: _PluggyPlugin, plugin_name: str) -> None:
+        # Fixtures defined in conftest plugins are only visible to within the
+        # conftest's directory. This is unlike fixtures in non-conftest plugins
+        # which have global visibility. So for conftests, construct the base
+        # nodeid from the plugin name (which is the conftest path).
+        if plugin_name and plugin_name.endswith("conftest.py"):
+            # Note: we explicitly do *not* use `plugin.__file__` here -- The
+            # difference is that plugin_name has the correct capitalization on
+            # case-insensitive systems (Windows) and other normalization issues
+            # (issue #11816).
+            conftestpath = absolutepath(plugin_name)
+            try:
+                nodeid = str(conftestpath.parent.relative_to(self.config.rootpath))
+            except ValueError:
+                nodeid = ""
+            if nodeid == ".":
+                nodeid = ""
+            if os.sep != nodes.SEP:
+                nodeid = nodeid.replace(os.sep, nodes.SEP)
+        else:
+            nodeid = None
+
+        self.parsefactories(plugin, nodeid)
+
+    def _getautousenames(self, node: nodes.Node) -> Iterator[str]:
+        """Return the names of autouse fixtures applicable to node."""
+        for parentnode in node.listchain():
+            basenames = self._nodeid_autousenames.get(parentnode.nodeid)
+            if basenames:
+                yield from basenames
+
+    def _getusefixturesnames(self, node: nodes.Item) -> Iterator[str]:
+        """Return the names of usefixtures fixtures applicable to node."""
+        for mark in node.iter_markers(name="usefixtures"):
+            yield from mark.args
+
+    def getfixtureclosure(
+        self,
+        parentnode: nodes.Node,
+        initialnames: tuple[str, ...],
+        ignore_args: AbstractSet[str],
+    ) -> tuple[list[str], dict[str, Sequence[FixtureDef[Any]]]]:
+        # Collect the closure of all fixtures, starting with the given
+        # fixturenames as the initial set.  As we have to visit all
+        # factory definitions anyway, we also return an arg2fixturedefs
+        # mapping so that the caller can reuse it and does not have
+        # to re-discover fixturedefs again for each fixturename
+        # (discovering matching fixtures for a given name/node is expensive).
+
+        fixturenames_closure = list(initialnames)
+
+        arg2fixturedefs: dict[str, Sequence[FixtureDef[Any]]] = {}
+        lastlen = -1
+        while lastlen != len(fixturenames_closure):
+            lastlen = len(fixturenames_closure)
+            for argname in fixturenames_closure:
+                if argname in ignore_args:
+                    continue
+                if argname in arg2fixturedefs:
+                    continue
+                fixturedefs = self.getfixturedefs(argname, parentnode)
+                if fixturedefs:
+                    arg2fixturedefs[argname] = fixturedefs
+                    for arg in fixturedefs[-1].argnames:
+                        if arg not in fixturenames_closure:
+                            fixturenames_closure.append(arg)
+
+        def sort_by_scope(arg_name: str) -> Scope:
+            try:
+                fixturedefs = arg2fixturedefs[arg_name]
+            except KeyError:
+                return Scope.Function
+            else:
+                return fixturedefs[-1]._scope
+
+        fixturenames_closure.sort(key=sort_by_scope, reverse=True)
+        return fixturenames_closure, arg2fixturedefs
+
+    def pytest_generate_tests(self, metafunc: Metafunc) -> None:
+        """Generate new tests based on parametrized fixtures used by the given metafunc"""
+
+        def get_parametrize_mark_argnames(mark: Mark) -> Sequence[str]:
+            args, _ = ParameterSet._parse_parametrize_args(*mark.args, **mark.kwargs)
+            return args
+
+        for argname in metafunc.fixturenames:
+            # Get the FixtureDefs for the argname.
+            fixture_defs = metafunc._arg2fixturedefs.get(argname)
+            if not fixture_defs:
+                # Will raise FixtureLookupError at setup time if not parametrized somewhere
+                # else (e.g @pytest.mark.parametrize)
+                continue
+
+            # If the test itself parametrizes using this argname, give it
+            # precedence.
+            if any(
+                argname in get_parametrize_mark_argnames(mark)
+                for mark in metafunc.definition.iter_markers("parametrize")
+            ):
+                continue
+
+            # In the common case we only look at the fixture def with the
+            # closest scope (last in the list). But if the fixture overrides
+            # another fixture, while requesting the super fixture, keep going
+            # in case the super fixture is parametrized (#1953).
+            for fixturedef in reversed(fixture_defs):
+                # Fixture is parametrized, apply it and stop.
+                if fixturedef.params is not None:
+                    metafunc.parametrize(
+                        argname,
+                        fixturedef.params,
+                        indirect=True,
+                        scope=fixturedef.scope,
+                        ids=fixturedef.ids,
+                    )
+                    break
+
+                # Not requesting the overridden super fixture, stop.
+                if argname not in fixturedef.argnames:
+                    break
+
+                # Try next super fixture, if any.
+
+    def pytest_collection_modifyitems(self, items: list[nodes.Item]) -> None:
+        # Separate parametrized setups.
+        items[:] = reorder_items(items)
+
+    def _register_fixture(
+        self,
+        *,
+        name: str,
+        func: _FixtureFunc[object],
+        nodeid: str | None,
+        scope: Scope | _ScopeName | Callable[[str, Config], _ScopeName] = "function",
+        params: Sequence[object] | None = None,
+        ids: tuple[object | None, ...] | Callable[[Any], object | None] | None = None,
+        autouse: bool = False,
+    ) -> None:
+        """Register a fixture
+
+        :param name:
+            The fixture's name.
+        :param func:
+            The fixture's implementation function.
+        :param nodeid:
+            The visibility of the fixture. The fixture will be available to the
+            node with this nodeid and its children in the collection tree.
+            None means that the fixture is visible to the entire collection tree,
+            e.g. a fixture defined for general use in a plugin.
+        :param scope:
+            The fixture's scope.
+        :param params:
+            The fixture's parametrization params.
+        :param ids:
+            The fixture's IDs.
+        :param autouse:
+            Whether this is an autouse fixture.
+        """
+        fixture_def = FixtureDef(
+            config=self.config,
+            baseid=nodeid,
+            argname=name,
+            func=func,
+            scope=scope,
+            params=params,
+            ids=ids,
+            _ispytest=True,
+        )
+
+        faclist = self._arg2fixturedefs.setdefault(name, [])
+        if fixture_def.has_location:
+            faclist.append(fixture_def)
+        else:
+            # fixturedefs with no location are at the front
+            # so this inserts the current fixturedef after the
+            # existing fixturedefs from external plugins but
+            # before the fixturedefs provided in conftests.
+            i = len([f for f in faclist if not f.has_location])
+            faclist.insert(i, fixture_def)
+        if autouse:
+            self._nodeid_autousenames.setdefault(nodeid or "", []).append(name)
+
+    @overload
+    def parsefactories(
+        self,
+        node_or_obj: nodes.Node,
+    ) -> None:
+        raise NotImplementedError()
+
+    @overload
+    def parsefactories(
+        self,
+        node_or_obj: object,
+        nodeid: str | None,
+    ) -> None:
+        raise NotImplementedError()
+
+    def parsefactories(
+        self,
+        node_or_obj: nodes.Node | object,
+        nodeid: str | NotSetType | None = NOTSET,
+    ) -> None:
+        """Collect fixtures from a collection node or object.
+
+        Found fixtures are parsed into `FixtureDef`s and saved.
+
+        If `node_or_object` is a collection node (with an underlying Python
+        object), the node's object is traversed and the node's nodeid is used to
+        determine the fixtures' visibility. `nodeid` must not be specified in
+        this case.
+
+        If `node_or_object` is an object (e.g. a plugin), the object is
+        traversed and the given `nodeid` is used to determine the fixtures'
+        visibility. `nodeid` must be specified in this case; None and "" mean
+        total visibility.
+        """
+        if nodeid is not NOTSET:
+            holderobj = node_or_obj
+        else:
+            assert isinstance(node_or_obj, nodes.Node)
+            holderobj = cast(object, node_or_obj.obj)  # type: ignore[attr-defined]
+            assert isinstance(node_or_obj.nodeid, str)
+            nodeid = node_or_obj.nodeid
+        if holderobj in self._holderobjseen:
+            return
+
+        # Avoid accessing `@property` (and other descriptors) when iterating fixtures.
+        if not safe_isclass(holderobj) and not isinstance(holderobj, types.ModuleType):
+            holderobj_tp: object = type(holderobj)
+        else:
+            holderobj_tp = holderobj
+
+        self._holderobjseen.add(holderobj)
+        for name in dir(holderobj):
+            # The attribute can be an arbitrary descriptor, so the attribute
+            # access below can raise. safe_getattr() ignores such exceptions.
+            obj_ub = safe_getattr(holderobj_tp, name, None)
+            marker = getfixturemarker(obj_ub)
+            if not isinstance(marker, FixtureFunctionMarker):
+                # Magic globals  with __getattr__ might have got us a wrong
+                # fixture attribute.
+                continue
+
+            # OK we know it is a fixture -- now safe to look up on the _instance_.
+            obj = getattr(holderobj, name)
+
+            if marker.name:
+                name = marker.name
+
+            # During fixture definition we wrap the original fixture function
+            # to issue a warning if called directly, so here we unwrap it in
+            # order to not emit the warning when pytest itself calls the
+            # fixture function.
+            func = get_real_method(obj, holderobj)
+
+            self._register_fixture(
+                name=name,
+                nodeid=nodeid,
+                func=func,
+                scope=marker.scope,
+                params=marker.params,
+                ids=marker.ids,
+                autouse=marker.autouse,
+            )
+
+    def getfixturedefs(
+        self, argname: str, node: nodes.Node
+    ) -> Sequence[FixtureDef[Any]] | None:
+        """Get FixtureDefs for a fixture name which are applicable
+        to a given node.
+
+        Returns None if there are no fixtures at all defined with the given
+        name. (This is different from the case in which there are fixtures
+        with the given name, but none applicable to the node. In this case,
+        an empty result is returned).
+
+        :param argname: Name of the fixture to search for.
+        :param node: The requesting Node.
+        """
+        try:
+            fixturedefs = self._arg2fixturedefs[argname]
+        except KeyError:
+            return None
+        return tuple(self._matchfactories(fixturedefs, node))
+
+    def _matchfactories(
+        self, fixturedefs: Iterable[FixtureDef[Any]], node: nodes.Node
+    ) -> Iterator[FixtureDef[Any]]:
+        parentnodeids = {n.nodeid for n in node.iter_parents()}
+        for fixturedef in fixturedefs:
+            if fixturedef.baseid in parentnodeids:
+                yield fixturedef
+
+
+def show_fixtures_per_test(config: Config) -> int | ExitCode:
+    from _pytest.main import wrap_session
+
+    return wrap_session(config, _show_fixtures_per_test)
+
+
+_PYTEST_DIR = Path(_pytest.__file__).parent
+
+
+def _pretty_fixture_path(invocation_dir: Path, func) -> str:
+    loc = Path(getlocation(func, invocation_dir))
+    prefix = Path("...", "_pytest")
+    try:
+        return str(prefix / loc.relative_to(_PYTEST_DIR))
+    except ValueError:
+        return bestrelpath(invocation_dir, loc)
+
+
+def _show_fixtures_per_test(config: Config, session: Session) -> None:
+    import _pytest.config
+
+    session.perform_collect()
+    invocation_dir = config.invocation_params.dir
+    tw = _pytest.config.create_terminal_writer(config)
+    verbose = config.get_verbosity()
+
+    def get_best_relpath(func) -> str:
+        loc = getlocation(func, invocation_dir)
+        return bestrelpath(invocation_dir, Path(loc))
+
+    def write_fixture(fixture_def: FixtureDef[object]) -> None:
+        argname = fixture_def.argname
+        if verbose <= 0 and argname.startswith("_"):
+            return
+        prettypath = _pretty_fixture_path(invocation_dir, fixture_def.func)
+        tw.write(f"{argname}", green=True)
+        tw.write(f" -- {prettypath}", yellow=True)
+        tw.write("\n")
+        fixture_doc = inspect.getdoc(fixture_def.func)
+        if fixture_doc:
+            write_docstring(
+                tw,
+                fixture_doc.split("\n\n", maxsplit=1)[0]
+                if verbose <= 0
+                else fixture_doc,
+            )
+        else:
+            tw.line("    no docstring available", red=True)
+
+    def write_item(item: nodes.Item) -> None:
+        # Not all items have _fixtureinfo attribute.
+        info: FuncFixtureInfo | None = getattr(item, "_fixtureinfo", None)
+        if info is None or not info.name2fixturedefs:
+            # This test item does not use any fixtures.
+            return
+        tw.line()
+        tw.sep("-", f"fixtures used by {item.name}")
+        # TODO: Fix this type ignore.
+        tw.sep("-", f"({get_best_relpath(item.function)})")  # type: ignore[attr-defined]
+        # dict key not used in loop but needed for sorting.
+        for _, fixturedefs in sorted(info.name2fixturedefs.items()):
+            assert fixturedefs is not None
+            if not fixturedefs:
+                continue
+            # Last item is expected to be the one used by the test item.
+            write_fixture(fixturedefs[-1])
+
+    for session_item in session.items:
+        write_item(session_item)
+
+
+def showfixtures(config: Config) -> int | ExitCode:
+    from _pytest.main import wrap_session
+
+    return wrap_session(config, _showfixtures_main)
+
+
+def _showfixtures_main(config: Config, session: Session) -> None:
+    import _pytest.config
+
+    session.perform_collect()
+    invocation_dir = config.invocation_params.dir
+    tw = _pytest.config.create_terminal_writer(config)
+    verbose = config.get_verbosity()
+
+    fm = session._fixturemanager
+
+    available = []
+    seen: set[tuple[str, str]] = set()
+
+    for argname, fixturedefs in fm._arg2fixturedefs.items():
+        assert fixturedefs is not None
+        if not fixturedefs:
+            continue
+        for fixturedef in fixturedefs:
+            loc = getlocation(fixturedef.func, invocation_dir)
+            if (fixturedef.argname, loc) in seen:
+                continue
+            seen.add((fixturedef.argname, loc))
+            available.append(
+                (
+                    len(fixturedef.baseid),
+                    fixturedef.func.__module__,
+                    _pretty_fixture_path(invocation_dir, fixturedef.func),
+                    fixturedef.argname,
+                    fixturedef,
+                )
+            )
+
+    available.sort()
+    currentmodule = None
+    for baseid, module, prettypath, argname, fixturedef in available:
+        if currentmodule != module:
+            if not module.startswith("_pytest."):
+                tw.line()
+                tw.sep("-", f"fixtures defined from {module}")
+                currentmodule = module
+        if verbose <= 0 and argname.startswith("_"):
+            continue
+        tw.write(f"{argname}", green=True)
+        if fixturedef.scope != "function":
+            tw.write(f" [{fixturedef.scope} scope]", cyan=True)
+        tw.write(f" -- {prettypath}", yellow=True)
+        tw.write("\n")
+        doc = inspect.getdoc(fixturedef.func)
+        if doc:
+            write_docstring(
+                tw, doc.split("\n\n", maxsplit=1)[0] if verbose <= 0 else doc
+            )
+        else:
+            tw.line("    no docstring available", red=True)
+        tw.line()
+
+
+def write_docstring(tw: TerminalWriter, doc: str, indent: str = "    ") -> None:
+    for line in doc.split("\n"):
+        tw.line(indent + line)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/freeze_support.py b/backend/venv/lib/python3.12/site-packages/_pytest/freeze_support.py
new file mode 100644
index 00000000..2ba6f9b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/freeze_support.py
@@ -0,0 +1,45 @@
+"""Provides a function to report all internal modules for using freezing
+tools."""
+
+from __future__ import annotations
+
+import types
+from typing import Iterator
+
+
+def freeze_includes() -> list[str]:
+    """Return a list of module names used by pytest that should be
+    included by cx_freeze."""
+    import _pytest
+
+    result = list(_iter_all_modules(_pytest))
+    return result
+
+
+def _iter_all_modules(
+    package: str | types.ModuleType,
+    prefix: str = "",
+) -> Iterator[str]:
+    """Iterate over the names of all modules that can be found in the given
+    package, recursively.
+
+        >>> import _pytest
+        >>> list(_iter_all_modules(_pytest))
+        ['_pytest._argcomplete', '_pytest._code.code', ...]
+    """
+    import os
+    import pkgutil
+
+    if isinstance(package, str):
+        path = package
+    else:
+        # Type ignored because typeshed doesn't define ModuleType.__path__
+        # (only defined on packages).
+        package_path = package.__path__
+        path, prefix = package_path[0], package.__name__ + "."
+    for _, name, is_package in pkgutil.iter_modules([path]):
+        if is_package:
+            for m in _iter_all_modules(os.path.join(path, name), prefix=name + "."):
+                yield prefix + m
+        else:
+            yield prefix + name
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/helpconfig.py b/backend/venv/lib/python3.12/site-packages/_pytest/helpconfig.py
new file mode 100644
index 00000000..1886d5c9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/helpconfig.py
@@ -0,0 +1,276 @@
+# mypy: allow-untyped-defs
+"""Version info, help messages, tracing configuration."""
+
+from __future__ import annotations
+
+from argparse import Action
+import os
+import sys
+from typing import Generator
+
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import PrintHelp
+from _pytest.config.argparsing import Parser
+from _pytest.terminal import TerminalReporter
+import pytest
+
+
+class HelpAction(Action):
+    """An argparse Action that will raise an exception in order to skip the
+    rest of the argument parsing when --help is passed.
+
+    This prevents argparse from quitting due to missing required arguments
+    when any are defined, for example by ``pytest_addoption``.
+    This is similar to the way that the builtin argparse --help option is
+    implemented by raising SystemExit.
+    """
+
+    def __init__(self, option_strings, dest=None, default=False, help=None):
+        super().__init__(
+            option_strings=option_strings,
+            dest=dest,
+            const=True,
+            default=default,
+            nargs=0,
+            help=help,
+        )
+
+    def __call__(self, parser, namespace, values, option_string=None):
+        setattr(namespace, self.dest, self.const)
+
+        # We should only skip the rest of the parsing after preparse is done.
+        if getattr(parser._parser, "after_preparse", False):
+            raise PrintHelp
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("debugconfig")
+    group.addoption(
+        "--version",
+        "-V",
+        action="count",
+        default=0,
+        dest="version",
+        help="Display pytest version and information about plugins. "
+        "When given twice, also display information about plugins.",
+    )
+    group._addoption(
+        "-h",
+        "--help",
+        action=HelpAction,
+        dest="help",
+        help="Show help message and configuration info",
+    )
+    group._addoption(
+        "-p",
+        action="append",
+        dest="plugins",
+        default=[],
+        metavar="name",
+        help="Early-load given plugin module name or entry point (multi-allowed). "
+        "To avoid loading of plugins, use the `no:` prefix, e.g. "
+        "`no:doctest`.",
+    )
+    group.addoption(
+        "--traceconfig",
+        "--trace-config",
+        action="store_true",
+        default=False,
+        help="Trace considerations of conftest.py files",
+    )
+    group.addoption(
+        "--debug",
+        action="store",
+        nargs="?",
+        const="pytestdebug.log",
+        dest="debug",
+        metavar="DEBUG_FILE_NAME",
+        help="Store internal tracing debug information in this log file. "
+        "This file is opened with 'w' and truncated as a result, care advised. "
+        "Default: pytestdebug.log.",
+    )
+    group._addoption(
+        "-o",
+        "--override-ini",
+        dest="override_ini",
+        action="append",
+        help='Override ini option with "option=value" style, '
+        "e.g. `-o xfail_strict=True -o cache_dir=cache`.",
+    )
+
+
+@pytest.hookimpl(wrapper=True)
+def pytest_cmdline_parse() -> Generator[None, Config, Config]:
+    config = yield
+
+    if config.option.debug:
+        # --debug | --debug <file.log> was provided.
+        path = config.option.debug
+        debugfile = open(path, "w", encoding="utf-8")
+        debugfile.write(
+            "versions pytest-{}, "
+            "python-{}\ninvocation_dir={}\ncwd={}\nargs={}\n\n".format(
+                pytest.__version__,
+                ".".join(map(str, sys.version_info)),
+                config.invocation_params.dir,
+                os.getcwd(),
+                config.invocation_params.args,
+            )
+        )
+        config.trace.root.setwriter(debugfile.write)
+        undo_tracing = config.pluginmanager.enable_tracing()
+        sys.stderr.write(f"writing pytest debug information to {path}\n")
+
+        def unset_tracing() -> None:
+            debugfile.close()
+            sys.stderr.write(f"wrote pytest debug information to {debugfile.name}\n")
+            config.trace.root.setwriter(None)
+            undo_tracing()
+
+        config.add_cleanup(unset_tracing)
+
+    return config
+
+
+def showversion(config: Config) -> None:
+    if config.option.version > 1:
+        sys.stdout.write(
+            f"This is pytest version {pytest.__version__}, imported from {pytest.__file__}\n"
+        )
+        plugininfo = getpluginversioninfo(config)
+        if plugininfo:
+            for line in plugininfo:
+                sys.stdout.write(line + "\n")
+    else:
+        sys.stdout.write(f"pytest {pytest.__version__}\n")
+
+
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    if config.option.version > 0:
+        showversion(config)
+        return 0
+    elif config.option.help:
+        config._do_configure()
+        showhelp(config)
+        config._ensure_unconfigure()
+        return 0
+    return None
+
+
+def showhelp(config: Config) -> None:
+    import textwrap
+
+    reporter: TerminalReporter | None = config.pluginmanager.get_plugin(
+        "terminalreporter"
+    )
+    assert reporter is not None
+    tw = reporter._tw
+    tw.write(config._parser.optparser.format_help())
+    tw.line()
+    tw.line(
+        "[pytest] ini-options in the first "
+        "pytest.ini|tox.ini|setup.cfg|pyproject.toml file found:"
+    )
+    tw.line()
+
+    columns = tw.fullwidth  # costly call
+    indent_len = 24  # based on argparse's max_help_position=24
+    indent = " " * indent_len
+    for name in config._parser._ininames:
+        help, type, default = config._parser._inidict[name]
+        if type is None:
+            type = "string"
+        if help is None:
+            raise TypeError(f"help argument cannot be None for {name}")
+        spec = f"{name} ({type}):"
+        tw.write(f"  {spec}")
+        spec_len = len(spec)
+        if spec_len > (indent_len - 3):
+            # Display help starting at a new line.
+            tw.line()
+            helplines = textwrap.wrap(
+                help,
+                columns,
+                initial_indent=indent,
+                subsequent_indent=indent,
+                break_on_hyphens=False,
+            )
+
+            for line in helplines:
+                tw.line(line)
+        else:
+            # Display help starting after the spec, following lines indented.
+            tw.write(" " * (indent_len - spec_len - 2))
+            wrapped = textwrap.wrap(help, columns - indent_len, break_on_hyphens=False)
+
+            if wrapped:
+                tw.line(wrapped[0])
+                for line in wrapped[1:]:
+                    tw.line(indent + line)
+
+    tw.line()
+    tw.line("Environment variables:")
+    vars = [
+        (
+            "CI",
+            "When set (regardless of value), pytest knows it is running in a "
+            "CI process and does not truncate summary info",
+        ),
+        ("BUILD_NUMBER", "Equivalent to CI"),
+        ("PYTEST_ADDOPTS", "Extra command line options"),
+        ("PYTEST_PLUGINS", "Comma-separated plugins to load during startup"),
+        ("PYTEST_DISABLE_PLUGIN_AUTOLOAD", "Set to disable plugin auto-loading"),
+        ("PYTEST_DEBUG", "Set to enable debug tracing of pytest's internals"),
+    ]
+    for name, help in vars:
+        tw.line(f"  {name:<24} {help}")
+    tw.line()
+    tw.line()
+
+    tw.line("to see available markers type: pytest --markers")
+    tw.line("to see available fixtures type: pytest --fixtures")
+    tw.line(
+        "(shown according to specified file_or_dir or current dir "
+        "if not specified; fixtures with leading '_' are only shown "
+        "with the '-v' option"
+    )
+
+    for warningreport in reporter.stats.get("warnings", []):
+        tw.line("warning : " + warningreport.message, red=True)
+
+
+conftest_options = [("pytest_plugins", "list of plugin names to load")]
+
+
+def getpluginversioninfo(config: Config) -> list[str]:
+    lines = []
+    plugininfo = config.pluginmanager.list_plugin_distinfo()
+    if plugininfo:
+        lines.append("registered third-party plugins:")
+        for plugin, dist in plugininfo:
+            loc = getattr(plugin, "__file__", repr(plugin))
+            content = f"{dist.project_name}-{dist.version} at {loc}"
+            lines.append("  " + content)
+    return lines
+
+
+def pytest_report_header(config: Config) -> list[str]:
+    lines = []
+    if config.option.debug or config.option.traceconfig:
+        lines.append(f"using: pytest-{pytest.__version__}")
+
+        verinfo = getpluginversioninfo(config)
+        if verinfo:
+            lines.extend(verinfo)
+
+    if config.option.traceconfig:
+        lines.append("active plugins:")
+        items = config.pluginmanager.list_name_plugin()
+        for name, plugin in items:
+            if hasattr(plugin, "__file__"):
+                r = plugin.__file__
+            else:
+                r = repr(plugin)
+            lines.append(f"    {name:<20}: {r}")
+    return lines
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/hookspec.py b/backend/venv/lib/python3.12/site-packages/_pytest/hookspec.py
new file mode 100644
index 00000000..0a41b0ac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/hookspec.py
@@ -0,0 +1,1333 @@
+# mypy: allow-untyped-defs
+# ruff: noqa: T100
+"""Hook specifications for pytest plugins which are invoked by pytest itself
+and by builtin plugins."""
+
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Any
+from typing import Mapping
+from typing import Sequence
+from typing import TYPE_CHECKING
+
+from pluggy import HookspecMarker
+
+from .deprecated import HOOK_LEGACY_PATH_ARG
+
+
+if TYPE_CHECKING:
+    import pdb
+    from typing import Literal
+    import warnings
+
+    from _pytest._code.code import ExceptionInfo
+    from _pytest._code.code import ExceptionRepr
+    from _pytest.compat import LEGACY_PATH
+    from _pytest.config import _PluggyPlugin
+    from _pytest.config import Config
+    from _pytest.config import ExitCode
+    from _pytest.config import PytestPluginManager
+    from _pytest.config.argparsing import Parser
+    from _pytest.fixtures import FixtureDef
+    from _pytest.fixtures import SubRequest
+    from _pytest.main import Session
+    from _pytest.nodes import Collector
+    from _pytest.nodes import Item
+    from _pytest.outcomes import Exit
+    from _pytest.python import Class
+    from _pytest.python import Function
+    from _pytest.python import Metafunc
+    from _pytest.python import Module
+    from _pytest.reports import CollectReport
+    from _pytest.reports import TestReport
+    from _pytest.runner import CallInfo
+    from _pytest.terminal import TerminalReporter
+    from _pytest.terminal import TestShortLogReport
+
+
+hookspec = HookspecMarker("pytest")
+
+# -------------------------------------------------------------------------
+# Initialization hooks called for every plugin
+# -------------------------------------------------------------------------
+
+
+@hookspec(historic=True)
+def pytest_addhooks(pluginmanager: PytestPluginManager) -> None:
+    """Called at plugin registration time to allow adding new hooks via a call to
+    :func:`pluginmanager.add_hookspecs(module_or_class, prefix) <pytest.PytestPluginManager.add_hookspecs>`.
+
+    :param pluginmanager: The pytest plugin manager.
+
+    .. note::
+        This hook is incompatible with hook wrappers.
+
+    Use in conftest plugins
+    =======================
+
+    If a conftest plugin implements this hook, it will be called immediately
+    when the conftest is registered.
+    """
+
+
+@hookspec(historic=True)
+def pytest_plugin_registered(
+    plugin: _PluggyPlugin,
+    plugin_name: str,
+    manager: PytestPluginManager,
+) -> None:
+    """A new pytest plugin got registered.
+
+    :param plugin: The plugin module or instance.
+    :param plugin_name: The name by which the plugin is registered.
+    :param manager: The pytest plugin manager.
+
+    .. note::
+        This hook is incompatible with hook wrappers.
+
+    Use in conftest plugins
+    =======================
+
+    If a conftest plugin implements this hook, it will be called immediately
+    when the conftest is registered, once for each plugin registered thus far
+    (including itself!), and for all plugins thereafter when they are
+    registered.
+    """
+
+
+@hookspec(historic=True)
+def pytest_addoption(parser: Parser, pluginmanager: PytestPluginManager) -> None:
+    """Register argparse-style options and ini-style config values,
+    called once at the beginning of a test run.
+
+    :param parser:
+        To add command line options, call
+        :py:func:`parser.addoption(...) <pytest.Parser.addoption>`.
+        To add ini-file values call :py:func:`parser.addini(...)
+        <pytest.Parser.addini>`.
+
+    :param pluginmanager:
+        The pytest plugin manager, which can be used to install :py:func:`~pytest.hookspec`'s
+        or :py:func:`~pytest.hookimpl`'s and allow one plugin to call another plugin's hooks
+        to change how command line options are added.
+
+    Options can later be accessed through the
+    :py:class:`config <pytest.Config>` object, respectively:
+
+    - :py:func:`config.getoption(name) <pytest.Config.getoption>` to
+      retrieve the value of a command line option.
+
+    - :py:func:`config.getini(name) <pytest.Config.getini>` to retrieve
+      a value read from an ini-style file.
+
+    The config object is passed around on many internal objects via the ``.config``
+    attribute or can be retrieved as the ``pytestconfig`` fixture.
+
+    .. note::
+        This hook is incompatible with hook wrappers.
+
+    Use in conftest plugins
+    =======================
+
+    If a conftest plugin implements this hook, it will be called immediately
+    when the conftest is registered.
+
+    This hook is only called for :ref:`initial conftests <pluginorder>`.
+    """
+
+
+@hookspec(historic=True)
+def pytest_configure(config: Config) -> None:
+    """Allow plugins and conftest files to perform initial configuration.
+
+    .. note::
+        This hook is incompatible with hook wrappers.
+
+    :param config: The pytest config object.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is called for every :ref:`initial conftest <pluginorder>` file
+    after command line options have been parsed. After that, the hook is called
+    for other conftest files as they are registered.
+    """
+
+
+# -------------------------------------------------------------------------
+# Bootstrapping hooks called for plugins registered early enough:
+# internal and 3rd party plugins.
+# -------------------------------------------------------------------------
+
+
+@hookspec(firstresult=True)
+def pytest_cmdline_parse(
+    pluginmanager: PytestPluginManager, args: list[str]
+) -> Config | None:
+    """Return an initialized :class:`~pytest.Config`, parsing the specified args.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    .. note::
+        This hook is only called for plugin classes passed to the
+        ``plugins`` arg when using `pytest.main`_ to perform an in-process
+        test run.
+
+    :param pluginmanager: The pytest plugin manager.
+    :param args: List of arguments passed on the command line.
+    :returns: A pytest config object.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is not called for conftest files.
+    """
+
+
+def pytest_load_initial_conftests(
+    early_config: Config, parser: Parser, args: list[str]
+) -> None:
+    """Called to implement the loading of :ref:`initial conftest files
+    <pluginorder>` ahead of command line option parsing.
+
+    :param early_config: The pytest config object.
+    :param args: Arguments passed on the command line.
+    :param parser: To add command line options.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is not called for conftest files.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_cmdline_main(config: Config) -> ExitCode | int | None:
+    """Called for performing the main command line action.
+
+    The default implementation will invoke the configure hooks and
+    :hook:`pytest_runtestloop`.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param config: The pytest config object.
+    :returns: The exit code.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is only called for :ref:`initial conftests <pluginorder>`.
+    """
+
+
+# -------------------------------------------------------------------------
+# collection hooks
+# -------------------------------------------------------------------------
+
+
+@hookspec(firstresult=True)
+def pytest_collection(session: Session) -> object | None:
+    """Perform the collection phase for the given session.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+    The return value is not used, but only stops further processing.
+
+    The default collection phase is this (see individual hooks for full details):
+
+    1. Starting from ``session`` as the initial collector:
+
+      1. ``pytest_collectstart(collector)``
+      2. ``report = pytest_make_collect_report(collector)``
+      3. ``pytest_exception_interact(collector, call, report)`` if an interactive exception occurred
+      4. For each collected node:
+
+        1. If an item, ``pytest_itemcollected(item)``
+        2. If a collector, recurse into it.
+
+      5. ``pytest_collectreport(report)``
+
+    2. ``pytest_collection_modifyitems(session, config, items)``
+
+      1. ``pytest_deselected(items)`` for any deselected items (may be called multiple times)
+
+    3. ``pytest_collection_finish(session)``
+    4. Set ``session.items`` to the list of collected items
+    5. Set ``session.testscollected`` to the number of collected items
+
+    You can implement this hook to only perform some action before collection,
+    for example the terminal plugin uses it to start displaying the collection
+    counter (and returns `None`).
+
+    :param session: The pytest session object.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is only called for :ref:`initial conftests <pluginorder>`.
+    """
+
+
+def pytest_collection_modifyitems(
+    session: Session, config: Config, items: list[Item]
+) -> None:
+    """Called after collection has been performed. May filter or re-order
+    the items in-place.
+
+    When items are deselected (filtered out from ``items``),
+    the hook :hook:`pytest_deselected` must be called explicitly
+    with the deselected items to properly notify other plugins,
+    e.g. with ``config.hook.pytest_deselected(deselected_items)``.
+
+    :param session: The pytest session object.
+    :param config: The pytest config object.
+    :param items: List of item objects.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+def pytest_collection_finish(session: Session) -> None:
+    """Called after collection has been performed and modified.
+
+    :param session: The pytest session object.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+@hookspec(
+    firstresult=True,
+    warn_on_impl_args={
+        "path": HOOK_LEGACY_PATH_ARG.format(
+            pylib_path_arg="path", pathlib_path_arg="collection_path"
+        ),
+    },
+)
+def pytest_ignore_collect(
+    collection_path: Path, path: LEGACY_PATH, config: Config
+) -> bool | None:
+    """Return ``True`` to ignore this path for collection.
+
+    Return ``None`` to let other plugins ignore the path for collection.
+
+    Returning ``False`` will forcefully *not* ignore this path for collection,
+    without giving a chance for other plugins to ignore this path.
+
+    This hook is consulted for all files and directories prior to calling
+    more specific hooks.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param collection_path: The path to analyze.
+    :type collection_path: pathlib.Path
+    :param path: The path to analyze (deprecated).
+    :param config: The pytest config object.
+
+    .. versionchanged:: 7.0.0
+        The ``collection_path`` parameter was added as a :class:`pathlib.Path`
+        equivalent of the ``path`` parameter. The ``path`` parameter
+        has been deprecated.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collection path, only
+    conftest files in parent directories of the collection path are consulted
+    (if the path is a directory, its own conftest file is *not* consulted - a
+    directory cannot ignore itself!).
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_collect_directory(path: Path, parent: Collector) -> Collector | None:
+    """Create a :class:`~pytest.Collector` for the given directory, or None if
+    not relevant.
+
+    .. versionadded:: 8.0
+
+    For best results, the returned collector should be a subclass of
+    :class:`~pytest.Directory`, but this is not required.
+
+    The new node needs to have the specified ``parent`` as a parent.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param path: The path to analyze.
+    :type path: pathlib.Path
+
+    See :ref:`custom directory collectors` for a simple example of use of this
+    hook.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collection path, only
+    conftest files in parent directories of the collection path are consulted
+    (if the path is a directory, its own conftest file is *not* consulted - a
+    directory cannot collect itself!).
+    """
+
+
+@hookspec(
+    warn_on_impl_args={
+        "path": HOOK_LEGACY_PATH_ARG.format(
+            pylib_path_arg="path", pathlib_path_arg="file_path"
+        ),
+    },
+)
+def pytest_collect_file(
+    file_path: Path, path: LEGACY_PATH, parent: Collector
+) -> Collector | None:
+    """Create a :class:`~pytest.Collector` for the given path, or None if not relevant.
+
+    For best results, the returned collector should be a subclass of
+    :class:`~pytest.File`, but this is not required.
+
+    The new node needs to have the specified ``parent`` as a parent.
+
+    :param file_path: The path to analyze.
+    :type file_path: pathlib.Path
+    :param path: The path to collect (deprecated).
+
+    .. versionchanged:: 7.0.0
+        The ``file_path`` parameter was added as a :class:`pathlib.Path`
+        equivalent of the ``path`` parameter. The ``path`` parameter
+        has been deprecated.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given file path, only
+    conftest files in parent directories of the file path are consulted.
+    """
+
+
+# logging hooks for collection
+
+
+def pytest_collectstart(collector: Collector) -> None:
+    """Collector starts collecting.
+
+    :param collector:
+        The collector.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collector, only
+    conftest files in the collector's directory and its parent directories are
+    consulted.
+    """
+
+
+def pytest_itemcollected(item: Item) -> None:
+    """We just collected a test item.
+
+    :param item:
+        The item.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_collectreport(report: CollectReport) -> None:
+    """Collector finished collecting.
+
+    :param report:
+        The collect report.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collector, only
+    conftest files in the collector's directory and its parent directories are
+    consulted.
+    """
+
+
+def pytest_deselected(items: Sequence[Item]) -> None:
+    """Called for deselected test items, e.g. by keyword.
+
+    Note that this hook has two integration aspects for plugins:
+
+    - it can be *implemented* to be notified of deselected items
+    - it must be *called* from :hook:`pytest_collection_modifyitems`
+      implementations when items are deselected (to properly notify other plugins).
+
+    May be called multiple times.
+
+    :param items:
+        The items.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_make_collect_report(collector: Collector) -> CollectReport | None:
+    """Perform :func:`collector.collect() <pytest.Collector.collect>` and return
+    a :class:`~pytest.CollectReport`.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param collector:
+        The collector.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collector, only
+    conftest files in the collector's directory and its parent directories are
+    consulted.
+    """
+
+
+# -------------------------------------------------------------------------
+# Python test function related hooks
+# -------------------------------------------------------------------------
+
+
+@hookspec(
+    firstresult=True,
+    warn_on_impl_args={
+        "path": HOOK_LEGACY_PATH_ARG.format(
+            pylib_path_arg="path", pathlib_path_arg="module_path"
+        ),
+    },
+)
+def pytest_pycollect_makemodule(
+    module_path: Path, path: LEGACY_PATH, parent
+) -> Module | None:
+    """Return a :class:`pytest.Module` collector or None for the given path.
+
+    This hook will be called for each matching test module path.
+    The :hook:`pytest_collect_file` hook needs to be used if you want to
+    create test modules for files that do not match as a test module.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param module_path: The path of the module to collect.
+    :type module_path: pathlib.Path
+    :param path: The path of the module to collect (deprecated).
+
+    .. versionchanged:: 7.0.0
+        The ``module_path`` parameter was added as a :class:`pathlib.Path`
+        equivalent of the ``path`` parameter.
+
+        The ``path`` parameter has been deprecated in favor of ``fspath``.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given parent collector,
+    only conftest files in the collector's directory and its parent directories
+    are consulted.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_pycollect_makeitem(
+    collector: Module | Class, name: str, obj: object
+) -> None | Item | Collector | list[Item | Collector]:
+    """Return a custom item/collector for a Python object in a module, or None.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param collector:
+        The module/class collector.
+    :param name:
+        The name of the object in the module/class.
+    :param obj:
+        The object.
+    :returns:
+        The created items/collectors.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given collector, only
+    conftest files in the collector's directory and its parent directories
+    are consulted.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_pyfunc_call(pyfuncitem: Function) -> object | None:
+    """Call underlying test function.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param pyfuncitem:
+        The function item.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only
+    conftest files in the item's directory and its parent directories
+    are consulted.
+    """
+
+
+def pytest_generate_tests(metafunc: Metafunc) -> None:
+    """Generate (multiple) parametrized calls to a test function.
+
+    :param metafunc:
+        The :class:`~pytest.Metafunc` helper for the test function.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given function definition,
+    only conftest files in the functions's directory and its parent directories
+    are consulted.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_make_parametrize_id(config: Config, val: object, argname: str) -> str | None:
+    """Return a user-friendly string representation of the given ``val``
+    that will be used by @pytest.mark.parametrize calls, or None if the hook
+    doesn't know about ``val``.
+
+    The parameter name is available as ``argname``, if required.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    :param config: The pytest config object.
+    :param val: The parametrized value.
+    :param argname: The automatic parameter name produced by pytest.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+# -------------------------------------------------------------------------
+# runtest related hooks
+# -------------------------------------------------------------------------
+
+
+@hookspec(firstresult=True)
+def pytest_runtestloop(session: Session) -> object | None:
+    """Perform the main runtest loop (after collection finished).
+
+    The default hook implementation performs the runtest protocol for all items
+    collected in the session (``session.items``), unless the collection failed
+    or the ``collectonly`` pytest option is set.
+
+    If at any point :py:func:`pytest.exit` is called, the loop is
+    terminated immediately.
+
+    If at any point ``session.shouldfail`` or ``session.shouldstop`` are set, the
+    loop is terminated after the runtest protocol for the current item is finished.
+
+    :param session: The pytest session object.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+    The return value is not used, but only stops further processing.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_runtest_protocol(item: Item, nextitem: Item | None) -> object | None:
+    """Perform the runtest protocol for a single test item.
+
+    The default runtest protocol is this (see individual hooks for full details):
+
+    - ``pytest_runtest_logstart(nodeid, location)``
+
+    - Setup phase:
+        - ``call = pytest_runtest_setup(item)`` (wrapped in ``CallInfo(when="setup")``)
+        - ``report = pytest_runtest_makereport(item, call)``
+        - ``pytest_runtest_logreport(report)``
+        - ``pytest_exception_interact(call, report)`` if an interactive exception occurred
+
+    - Call phase, if the setup passed and the ``setuponly`` pytest option is not set:
+        - ``call = pytest_runtest_call(item)`` (wrapped in ``CallInfo(when="call")``)
+        - ``report = pytest_runtest_makereport(item, call)``
+        - ``pytest_runtest_logreport(report)``
+        - ``pytest_exception_interact(call, report)`` if an interactive exception occurred
+
+    - Teardown phase:
+        - ``call = pytest_runtest_teardown(item, nextitem)`` (wrapped in ``CallInfo(when="teardown")``)
+        - ``report = pytest_runtest_makereport(item, call)``
+        - ``pytest_runtest_logreport(report)``
+        - ``pytest_exception_interact(call, report)`` if an interactive exception occurred
+
+    - ``pytest_runtest_logfinish(nodeid, location)``
+
+    :param item: Test item for which the runtest protocol is performed.
+    :param nextitem: The scheduled-to-be-next test item (or None if this is the end my friend).
+
+    Stops at first non-None result, see :ref:`firstresult`.
+    The return value is not used, but only stops further processing.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+def pytest_runtest_logstart(nodeid: str, location: tuple[str, int | None, str]) -> None:
+    """Called at the start of running the runtest protocol for a single item.
+
+    See :hook:`pytest_runtest_protocol` for a description of the runtest protocol.
+
+    :param nodeid: Full node ID of the item.
+    :param location: A tuple of ``(filename, lineno, testname)``
+        where ``filename`` is a file path relative to ``config.rootpath``
+        and ``lineno`` is 0-based.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_runtest_logfinish(
+    nodeid: str, location: tuple[str, int | None, str]
+) -> None:
+    """Called at the end of running the runtest protocol for a single item.
+
+    See :hook:`pytest_runtest_protocol` for a description of the runtest protocol.
+
+    :param nodeid: Full node ID of the item.
+    :param location: A tuple of ``(filename, lineno, testname)``
+        where ``filename`` is a file path relative to ``config.rootpath``
+        and ``lineno`` is 0-based.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_runtest_setup(item: Item) -> None:
+    """Called to perform the setup phase for a test item.
+
+    The default implementation runs ``setup()`` on ``item`` and all of its
+    parents (which haven't been setup yet). This includes obtaining the
+    values of fixtures required by the item (which haven't been obtained
+    yet).
+
+    :param item:
+        The item.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_runtest_call(item: Item) -> None:
+    """Called to run the test for test item (the call phase).
+
+    The default implementation calls ``item.runtest()``.
+
+    :param item:
+        The item.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_runtest_teardown(item: Item, nextitem: Item | None) -> None:
+    """Called to perform the teardown phase for a test item.
+
+    The default implementation runs the finalizers and calls ``teardown()``
+    on ``item`` and all of its parents (which need to be torn down). This
+    includes running the teardown phase of fixtures required by the item (if
+    they go out of scope).
+
+    :param item:
+        The item.
+    :param nextitem:
+        The scheduled-to-be-next test item (None if no further test item is
+        scheduled). This argument is used to perform exact teardowns, i.e.
+        calling just enough finalizers so that nextitem only needs to call
+        setup functions.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_runtest_makereport(item: Item, call: CallInfo[None]) -> TestReport | None:
+    """Called to create a :class:`~pytest.TestReport` for each of
+    the setup, call and teardown runtest phases of a test item.
+
+    See :hook:`pytest_runtest_protocol` for a description of the runtest protocol.
+
+    :param item: The item.
+    :param call: The :class:`~pytest.CallInfo` for the phase.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_runtest_logreport(report: TestReport) -> None:
+    """Process the :class:`~pytest.TestReport` produced for each
+    of the setup, call and teardown runtest phases of an item.
+
+    See :hook:`pytest_runtest_protocol` for a description of the runtest protocol.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_report_to_serializable(
+    config: Config,
+    report: CollectReport | TestReport,
+) -> dict[str, Any] | None:
+    """Serialize the given report object into a data structure suitable for
+    sending over the wire, e.g. converted to JSON.
+
+    :param config: The pytest config object.
+    :param report: The report.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. The exact details may depend
+    on the plugin which calls the hook.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_report_from_serializable(
+    config: Config,
+    data: dict[str, Any],
+) -> CollectReport | TestReport | None:
+    """Restore a report object previously serialized with
+    :hook:`pytest_report_to_serializable`.
+
+    :param config: The pytest config object.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. The exact details may depend
+    on the plugin which calls the hook.
+    """
+
+
+# -------------------------------------------------------------------------
+# Fixture related hooks
+# -------------------------------------------------------------------------
+
+
+@hookspec(firstresult=True)
+def pytest_fixture_setup(
+    fixturedef: FixtureDef[Any], request: SubRequest
+) -> object | None:
+    """Perform fixture setup execution.
+
+    :param fixturedef:
+        The fixture definition object.
+    :param request:
+        The fixture request object.
+    :returns:
+        The return value of the call to the fixture function.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    .. note::
+        If the fixture function returns None, other implementations of
+        this hook function will continue to be called, according to the
+        behavior of the :ref:`firstresult` option.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given fixture, only
+    conftest files in the fixture scope's directory and its parent directories
+    are consulted.
+    """
+
+
+def pytest_fixture_post_finalizer(
+    fixturedef: FixtureDef[Any], request: SubRequest
+) -> None:
+    """Called after fixture teardown, but before the cache is cleared, so
+    the fixture result ``fixturedef.cached_result`` is still available (not
+    ``None``).
+
+    :param fixturedef:
+        The fixture definition object.
+    :param request:
+        The fixture request object.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given fixture, only
+    conftest files in the fixture scope's directory and its parent directories
+    are consulted.
+    """
+
+
+# -------------------------------------------------------------------------
+# test session related hooks
+# -------------------------------------------------------------------------
+
+
+def pytest_sessionstart(session: Session) -> None:
+    """Called after the ``Session`` object has been created and before performing collection
+    and entering the run test loop.
+
+    :param session: The pytest session object.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is only called for :ref:`initial conftests <pluginorder>`.
+    """
+
+
+def pytest_sessionfinish(
+    session: Session,
+    exitstatus: int | ExitCode,
+) -> None:
+    """Called after whole test run finished, right before returning the exit status to the system.
+
+    :param session: The pytest session object.
+    :param exitstatus: The status which pytest will return to the system.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+def pytest_unconfigure(config: Config) -> None:
+    """Called before test process is exited.
+
+    :param config: The pytest config object.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook.
+    """
+
+
+# -------------------------------------------------------------------------
+# hooks for customizing the assert methods
+# -------------------------------------------------------------------------
+
+
+def pytest_assertrepr_compare(
+    config: Config, op: str, left: object, right: object
+) -> list[str] | None:
+    """Return explanation for comparisons in failing assert expressions.
+
+    Return None for no custom explanation, otherwise return a list
+    of strings. The strings will be joined by newlines but any newlines
+    *in* a string will be escaped. Note that all but the first line will
+    be indented slightly, the intention is for the first line to be a summary.
+
+    :param config: The pytest config object.
+    :param op: The operator, e.g. `"=="`, `"!="`, `"not in"`.
+    :param left: The left operand.
+    :param right: The right operand.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+def pytest_assertion_pass(item: Item, lineno: int, orig: str, expl: str) -> None:
+    """Called whenever an assertion passes.
+
+    .. versionadded:: 5.0
+
+    Use this hook to do some processing after a passing assertion.
+    The original assertion information is available in the `orig` string
+    and the pytest introspected assertion information is available in the
+    `expl` string.
+
+    This hook must be explicitly enabled by the ``enable_assertion_pass_hook``
+    ini-file option:
+
+    .. code-block:: ini
+
+        [pytest]
+        enable_assertion_pass_hook=true
+
+    You need to **clean the .pyc** files in your project directory and interpreter libraries
+    when enabling this option, as assertions will require to be re-written.
+
+    :param item: pytest item object of current test.
+    :param lineno: Line number of the assert statement.
+    :param orig: String with the original assertion.
+    :param expl: String with the assert explanation.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in the item's directory and its parent directories are consulted.
+    """
+
+
+# -------------------------------------------------------------------------
+# Hooks for influencing reporting (invoked from _pytest_terminal).
+# -------------------------------------------------------------------------
+
+
+@hookspec(
+    warn_on_impl_args={
+        "startdir": HOOK_LEGACY_PATH_ARG.format(
+            pylib_path_arg="startdir", pathlib_path_arg="start_path"
+        ),
+    },
+)
+def pytest_report_header(  # type:ignore[empty-body]
+    config: Config, start_path: Path, startdir: LEGACY_PATH
+) -> str | list[str]:
+    """Return a string or list of strings to be displayed as header info for terminal reporting.
+
+    :param config: The pytest config object.
+    :param start_path: The starting dir.
+    :type start_path: pathlib.Path
+    :param startdir: The starting dir (deprecated).
+
+    .. note::
+
+        Lines returned by a plugin are displayed before those of plugins which
+        ran before it.
+        If you want to have your line(s) displayed first, use
+        :ref:`trylast=True <plugin-hookorder>`.
+
+    .. versionchanged:: 7.0.0
+        The ``start_path`` parameter was added as a :class:`pathlib.Path`
+        equivalent of the ``startdir`` parameter. The ``startdir`` parameter
+        has been deprecated.
+
+    Use in conftest plugins
+    =======================
+
+    This hook is only called for :ref:`initial conftests <pluginorder>`.
+    """
+
+
+@hookspec(
+    warn_on_impl_args={
+        "startdir": HOOK_LEGACY_PATH_ARG.format(
+            pylib_path_arg="startdir", pathlib_path_arg="start_path"
+        ),
+    },
+)
+def pytest_report_collectionfinish(  # type:ignore[empty-body]
+    config: Config,
+    start_path: Path,
+    startdir: LEGACY_PATH,
+    items: Sequence[Item],
+) -> str | list[str]:
+    """Return a string or list of strings to be displayed after collection
+    has finished successfully.
+
+    These strings will be displayed after the standard "collected X items" message.
+
+    .. versionadded:: 3.2
+
+    :param config: The pytest config object.
+    :param start_path: The starting dir.
+    :type start_path: pathlib.Path
+    :param startdir: The starting dir (deprecated).
+    :param items: List of pytest items that are going to be executed; this list should not be modified.
+
+    .. note::
+
+        Lines returned by a plugin are displayed before those of plugins which
+        ran before it.
+        If you want to have your line(s) displayed first, use
+        :ref:`trylast=True <plugin-hookorder>`.
+
+    .. versionchanged:: 7.0.0
+        The ``start_path`` parameter was added as a :class:`pathlib.Path`
+        equivalent of the ``startdir`` parameter. The ``startdir`` parameter
+        has been deprecated.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+@hookspec(firstresult=True)
+def pytest_report_teststatus(  # type:ignore[empty-body]
+    report: CollectReport | TestReport, config: Config
+) -> TestShortLogReport | tuple[str, str, str | tuple[str, Mapping[str, bool]]]:
+    """Return result-category, shortletter and verbose word for status
+    reporting.
+
+    The result-category is a category in which to count the result, for
+    example "passed", "skipped", "error" or the empty string.
+
+    The shortletter is shown as testing progresses, for example ".", "s",
+    "E" or the empty string.
+
+    The verbose word is shown as testing progresses in verbose mode, for
+    example "PASSED", "SKIPPED", "ERROR" or the empty string.
+
+    pytest may style these implicitly according to the report outcome.
+    To provide explicit styling, return a tuple for the verbose word,
+    for example ``"rerun", "R", ("RERUN", {"yellow": True})``.
+
+    :param report: The report object whose status is to be returned.
+    :param config: The pytest config object.
+    :returns: The test status.
+
+    Stops at first non-None result, see :ref:`firstresult`.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+def pytest_terminal_summary(
+    terminalreporter: TerminalReporter,
+    exitstatus: ExitCode,
+    config: Config,
+) -> None:
+    """Add a section to terminal summary reporting.
+
+    :param terminalreporter: The internal terminal reporter object.
+    :param exitstatus: The exit status that will be reported back to the OS.
+    :param config: The pytest config object.
+
+    .. versionadded:: 4.2
+        The ``config`` parameter.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+@hookspec(historic=True)
+def pytest_warning_recorded(
+    warning_message: warnings.WarningMessage,
+    when: Literal["config", "collect", "runtest"],
+    nodeid: str,
+    location: tuple[str, int, str] | None,
+) -> None:
+    """Process a warning captured by the internal pytest warnings plugin.
+
+    :param warning_message:
+        The captured warning. This is the same object produced by :class:`warnings.catch_warnings`,
+        and contains the same attributes as the parameters of :py:func:`warnings.showwarning`.
+
+    :param when:
+        Indicates when the warning was captured. Possible values:
+
+        * ``"config"``: during pytest configuration/initialization stage.
+        * ``"collect"``: during test collection.
+        * ``"runtest"``: during test execution.
+
+    :param nodeid:
+        Full id of the item. Empty string for warnings that are not specific to
+        a particular node.
+
+    :param location:
+        When available, holds information about the execution context of the captured
+        warning (filename, linenumber, function). ``function`` evaluates to <module>
+        when the execution context is at the module level.
+
+    .. versionadded:: 6.0
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. If the warning is specific to a
+    particular node, only conftest files in parent directories of the node are
+    consulted.
+    """
+
+
+# -------------------------------------------------------------------------
+# Hooks for influencing skipping
+# -------------------------------------------------------------------------
+
+
+def pytest_markeval_namespace(  # type:ignore[empty-body]
+    config: Config,
+) -> dict[str, Any]:
+    """Called when constructing the globals dictionary used for
+    evaluating string conditions in xfail/skipif markers.
+
+    This is useful when the condition for a marker requires
+    objects that are expensive or impossible to obtain during
+    collection time, which is required by normal boolean
+    conditions.
+
+    .. versionadded:: 6.2
+
+    :param config: The pytest config object.
+    :returns: A dictionary of additional globals to add.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given item, only conftest
+    files in parent directories of the item are consulted.
+    """
+
+
+# -------------------------------------------------------------------------
+# error handling and internal debugging hooks
+# -------------------------------------------------------------------------
+
+
+def pytest_internalerror(
+    excrepr: ExceptionRepr,
+    excinfo: ExceptionInfo[BaseException],
+) -> bool | None:
+    """Called for internal errors.
+
+    Return True to suppress the fallback handling of printing an
+    INTERNALERROR message directly to sys.stderr.
+
+    :param excrepr: The exception repr object.
+    :param excinfo: The exception info.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+def pytest_keyboard_interrupt(
+    excinfo: ExceptionInfo[KeyboardInterrupt | Exit],
+) -> None:
+    """Called for keyboard interrupt.
+
+    :param excinfo: The exception info.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+def pytest_exception_interact(
+    node: Item | Collector,
+    call: CallInfo[Any],
+    report: CollectReport | TestReport,
+) -> None:
+    """Called when an exception was raised which can potentially be
+    interactively handled.
+
+    May be called during collection (see :hook:`pytest_make_collect_report`),
+    in which case ``report`` is a :class:`~pytest.CollectReport`.
+
+    May be called during runtest of an item (see :hook:`pytest_runtest_protocol`),
+    in which case ``report`` is a :class:`~pytest.TestReport`.
+
+    This hook is not called if the exception that was raised is an internal
+    exception like ``skip.Exception``.
+
+    :param node:
+        The item or collector.
+    :param call:
+        The call information. Contains the exception.
+    :param report:
+        The collection or test report.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest file can implement this hook. For a given node, only conftest
+    files in parent directories of the node are consulted.
+    """
+
+
+def pytest_enter_pdb(config: Config, pdb: pdb.Pdb) -> None:
+    """Called upon pdb.set_trace().
+
+    Can be used by plugins to take special action just before the python
+    debugger enters interactive mode.
+
+    :param config: The pytest config object.
+    :param pdb: The Pdb instance.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
+
+
+def pytest_leave_pdb(config: Config, pdb: pdb.Pdb) -> None:
+    """Called when leaving pdb (e.g. with continue after pdb.set_trace()).
+
+    Can be used by plugins to take special action just after the python
+    debugger leaves interactive mode.
+
+    :param config: The pytest config object.
+    :param pdb: The Pdb instance.
+
+    Use in conftest plugins
+    =======================
+
+    Any conftest plugin can implement this hook.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/junitxml.py b/backend/venv/lib/python3.12/site-packages/_pytest/junitxml.py
new file mode 100644
index 00000000..3a2cb59a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/junitxml.py
@@ -0,0 +1,697 @@
+# mypy: allow-untyped-defs
+"""Report test results in JUnit-XML format, for use with Jenkins and build
+integration servers.
+
+Based on initial code from Ross Lawley.
+
+Output conforms to
+https://github.com/jenkinsci/xunit-plugin/blob/master/src/main/resources/org/jenkinsci/plugins/xunit/types/model/xsd/junit-10.xsd
+"""
+
+from __future__ import annotations
+
+from datetime import datetime
+from datetime import timezone
+import functools
+import os
+import platform
+import re
+from typing import Callable
+from typing import Match
+import xml.etree.ElementTree as ET
+
+from _pytest import nodes
+from _pytest import timing
+from _pytest._code.code import ExceptionRepr
+from _pytest._code.code import ReprFileLocation
+from _pytest.config import Config
+from _pytest.config import filename_arg
+from _pytest.config.argparsing import Parser
+from _pytest.fixtures import FixtureRequest
+from _pytest.reports import TestReport
+from _pytest.stash import StashKey
+from _pytest.terminal import TerminalReporter
+import pytest
+
+
+xml_key = StashKey["LogXML"]()
+
+
+def bin_xml_escape(arg: object) -> str:
+    r"""Visually escape invalid XML characters.
+
+    For example, transforms
+        'hello\aworld\b'
+    into
+        'hello#x07world#x08'
+    Note that the #xABs are *not* XML escapes - missing the ampersand &#xAB.
+    The idea is to escape visually for the user rather than for XML itself.
+    """
+
+    def repl(matchobj: Match[str]) -> str:
+        i = ord(matchobj.group())
+        if i <= 0xFF:
+            return f"#x{i:02X}"
+        else:
+            return f"#x{i:04X}"
+
+    # The spec range of valid chars is:
+    # Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
+    # For an unknown(?) reason, we disallow #x7F (DEL) as well.
+    illegal_xml_re = (
+        "[^\u0009\u000a\u000d\u0020-\u007e\u0080-\ud7ff\ue000-\ufffd\u10000-\u10ffff]"
+    )
+    return re.sub(illegal_xml_re, repl, str(arg))
+
+
+def merge_family(left, right) -> None:
+    result = {}
+    for kl, vl in left.items():
+        for kr, vr in right.items():
+            if not isinstance(vl, list):
+                raise TypeError(type(vl))
+            result[kl] = vl + vr
+    left.update(result)
+
+
+families = {}
+families["_base"] = {"testcase": ["classname", "name"]}
+families["_base_legacy"] = {"testcase": ["file", "line", "url"]}
+
+# xUnit 1.x inherits legacy attributes.
+families["xunit1"] = families["_base"].copy()
+merge_family(families["xunit1"], families["_base_legacy"])
+
+# xUnit 2.x uses strict base attributes.
+families["xunit2"] = families["_base"]
+
+
+class _NodeReporter:
+    def __init__(self, nodeid: str | TestReport, xml: LogXML) -> None:
+        self.id = nodeid
+        self.xml = xml
+        self.add_stats = self.xml.add_stats
+        self.family = self.xml.family
+        self.duration = 0.0
+        self.properties: list[tuple[str, str]] = []
+        self.nodes: list[ET.Element] = []
+        self.attrs: dict[str, str] = {}
+
+    def append(self, node: ET.Element) -> None:
+        self.xml.add_stats(node.tag)
+        self.nodes.append(node)
+
+    def add_property(self, name: str, value: object) -> None:
+        self.properties.append((str(name), bin_xml_escape(value)))
+
+    def add_attribute(self, name: str, value: object) -> None:
+        self.attrs[str(name)] = bin_xml_escape(value)
+
+    def make_properties_node(self) -> ET.Element | None:
+        """Return a Junit node containing custom properties, if any."""
+        if self.properties:
+            properties = ET.Element("properties")
+            for name, value in self.properties:
+                properties.append(ET.Element("property", name=name, value=value))
+            return properties
+        return None
+
+    def record_testreport(self, testreport: TestReport) -> None:
+        names = mangle_test_address(testreport.nodeid)
+        existing_attrs = self.attrs
+        classnames = names[:-1]
+        if self.xml.prefix:
+            classnames.insert(0, self.xml.prefix)
+        attrs: dict[str, str] = {
+            "classname": ".".join(classnames),
+            "name": bin_xml_escape(names[-1]),
+            "file": testreport.location[0],
+        }
+        if testreport.location[1] is not None:
+            attrs["line"] = str(testreport.location[1])
+        if hasattr(testreport, "url"):
+            attrs["url"] = testreport.url
+        self.attrs = attrs
+        self.attrs.update(existing_attrs)  # Restore any user-defined attributes.
+
+        # Preserve legacy testcase behavior.
+        if self.family == "xunit1":
+            return
+
+        # Filter out attributes not permitted by this test family.
+        # Including custom attributes because they are not valid here.
+        temp_attrs = {}
+        for key in self.attrs:
+            if key in families[self.family]["testcase"]:
+                temp_attrs[key] = self.attrs[key]
+        self.attrs = temp_attrs
+
+    def to_xml(self) -> ET.Element:
+        testcase = ET.Element("testcase", self.attrs, time=f"{self.duration:.3f}")
+        properties = self.make_properties_node()
+        if properties is not None:
+            testcase.append(properties)
+        testcase.extend(self.nodes)
+        return testcase
+
+    def _add_simple(self, tag: str, message: str, data: str | None = None) -> None:
+        node = ET.Element(tag, message=message)
+        node.text = bin_xml_escape(data)
+        self.append(node)
+
+    def write_captured_output(self, report: TestReport) -> None:
+        if not self.xml.log_passing_tests and report.passed:
+            return
+
+        content_out = report.capstdout
+        content_log = report.caplog
+        content_err = report.capstderr
+        if self.xml.logging == "no":
+            return
+        content_all = ""
+        if self.xml.logging in ["log", "all"]:
+            content_all = self._prepare_content(content_log, " Captured Log ")
+        if self.xml.logging in ["system-out", "out-err", "all"]:
+            content_all += self._prepare_content(content_out, " Captured Out ")
+            self._write_content(report, content_all, "system-out")
+            content_all = ""
+        if self.xml.logging in ["system-err", "out-err", "all"]:
+            content_all += self._prepare_content(content_err, " Captured Err ")
+            self._write_content(report, content_all, "system-err")
+            content_all = ""
+        if content_all:
+            self._write_content(report, content_all, "system-out")
+
+    def _prepare_content(self, content: str, header: str) -> str:
+        return "\n".join([header.center(80, "-"), content, ""])
+
+    def _write_content(self, report: TestReport, content: str, jheader: str) -> None:
+        tag = ET.Element(jheader)
+        tag.text = bin_xml_escape(content)
+        self.append(tag)
+
+    def append_pass(self, report: TestReport) -> None:
+        self.add_stats("passed")
+
+    def append_failure(self, report: TestReport) -> None:
+        # msg = str(report.longrepr.reprtraceback.extraline)
+        if hasattr(report, "wasxfail"):
+            self._add_simple("skipped", "xfail-marked test passes unexpectedly")
+        else:
+            assert report.longrepr is not None
+            reprcrash: ReprFileLocation | None = getattr(
+                report.longrepr, "reprcrash", None
+            )
+            if reprcrash is not None:
+                message = reprcrash.message
+            else:
+                message = str(report.longrepr)
+            message = bin_xml_escape(message)
+            self._add_simple("failure", message, str(report.longrepr))
+
+    def append_collect_error(self, report: TestReport) -> None:
+        # msg = str(report.longrepr.reprtraceback.extraline)
+        assert report.longrepr is not None
+        self._add_simple("error", "collection failure", str(report.longrepr))
+
+    def append_collect_skipped(self, report: TestReport) -> None:
+        self._add_simple("skipped", "collection skipped", str(report.longrepr))
+
+    def append_error(self, report: TestReport) -> None:
+        assert report.longrepr is not None
+        reprcrash: ReprFileLocation | None = getattr(report.longrepr, "reprcrash", None)
+        if reprcrash is not None:
+            reason = reprcrash.message
+        else:
+            reason = str(report.longrepr)
+
+        if report.when == "teardown":
+            msg = f'failed on teardown with "{reason}"'
+        else:
+            msg = f'failed on setup with "{reason}"'
+        self._add_simple("error", bin_xml_escape(msg), str(report.longrepr))
+
+    def append_skipped(self, report: TestReport) -> None:
+        if hasattr(report, "wasxfail"):
+            xfailreason = report.wasxfail
+            if xfailreason.startswith("reason: "):
+                xfailreason = xfailreason[8:]
+            xfailreason = bin_xml_escape(xfailreason)
+            skipped = ET.Element("skipped", type="pytest.xfail", message=xfailreason)
+            self.append(skipped)
+        else:
+            assert isinstance(report.longrepr, tuple)
+            filename, lineno, skipreason = report.longrepr
+            if skipreason.startswith("Skipped: "):
+                skipreason = skipreason[9:]
+            details = f"{filename}:{lineno}: {skipreason}"
+
+            skipped = ET.Element(
+                "skipped", type="pytest.skip", message=bin_xml_escape(skipreason)
+            )
+            skipped.text = bin_xml_escape(details)
+            self.append(skipped)
+            self.write_captured_output(report)
+
+    def finalize(self) -> None:
+        data = self.to_xml()
+        self.__dict__.clear()
+        # Type ignored because mypy doesn't like overriding a method.
+        # Also the return value doesn't match...
+        self.to_xml = lambda: data  # type: ignore[method-assign]
+
+
+def _warn_incompatibility_with_xunit2(
+    request: FixtureRequest, fixture_name: str
+) -> None:
+    """Emit a PytestWarning about the given fixture being incompatible with newer xunit revisions."""
+    from _pytest.warning_types import PytestWarning
+
+    xml = request.config.stash.get(xml_key, None)
+    if xml is not None and xml.family not in ("xunit1", "legacy"):
+        request.node.warn(
+            PytestWarning(
+                f"{fixture_name} is incompatible with junit_family '{xml.family}' (use 'legacy' or 'xunit1')"
+            )
+        )
+
+
+@pytest.fixture
+def record_property(request: FixtureRequest) -> Callable[[str, object], None]:
+    """Add extra properties to the calling test.
+
+    User properties become part of the test report and are available to the
+    configured reporters, like JUnit XML.
+
+    The fixture is callable with ``name, value``. The value is automatically
+    XML-encoded.
+
+    Example::
+
+        def test_function(record_property):
+            record_property("example_key", 1)
+    """
+    _warn_incompatibility_with_xunit2(request, "record_property")
+
+    def append_property(name: str, value: object) -> None:
+        request.node.user_properties.append((name, value))
+
+    return append_property
+
+
+@pytest.fixture
+def record_xml_attribute(request: FixtureRequest) -> Callable[[str, object], None]:
+    """Add extra xml attributes to the tag for the calling test.
+
+    The fixture is callable with ``name, value``. The value is
+    automatically XML-encoded.
+    """
+    from _pytest.warning_types import PytestExperimentalApiWarning
+
+    request.node.warn(
+        PytestExperimentalApiWarning("record_xml_attribute is an experimental feature")
+    )
+
+    _warn_incompatibility_with_xunit2(request, "record_xml_attribute")
+
+    # Declare noop
+    def add_attr_noop(name: str, value: object) -> None:
+        pass
+
+    attr_func = add_attr_noop
+
+    xml = request.config.stash.get(xml_key, None)
+    if xml is not None:
+        node_reporter = xml.node_reporter(request.node.nodeid)
+        attr_func = node_reporter.add_attribute
+
+    return attr_func
+
+
+def _check_record_param_type(param: str, v: str) -> None:
+    """Used by record_testsuite_property to check that the given parameter name is of the proper
+    type."""
+    __tracebackhide__ = True
+    if not isinstance(v, str):
+        msg = "{param} parameter needs to be a string, but {g} given"  # type: ignore[unreachable]
+        raise TypeError(msg.format(param=param, g=type(v).__name__))
+
+
+@pytest.fixture(scope="session")
+def record_testsuite_property(request: FixtureRequest) -> Callable[[str, object], None]:
+    """Record a new ``<property>`` tag as child of the root ``<testsuite>``.
+
+    This is suitable to writing global information regarding the entire test
+    suite, and is compatible with ``xunit2`` JUnit family.
+
+    This is a ``session``-scoped fixture which is called with ``(name, value)``. Example:
+
+    .. code-block:: python
+
+        def test_foo(record_testsuite_property):
+            record_testsuite_property("ARCH", "PPC")
+            record_testsuite_property("STORAGE_TYPE", "CEPH")
+
+    :param name:
+        The property name.
+    :param value:
+        The property value. Will be converted to a string.
+
+    .. warning::
+
+        Currently this fixture **does not work** with the
+        `pytest-xdist <https://github.com/pytest-dev/pytest-xdist>`__ plugin. See
+        :issue:`7767` for details.
+    """
+    __tracebackhide__ = True
+
+    def record_func(name: str, value: object) -> None:
+        """No-op function in case --junit-xml was not passed in the command-line."""
+        __tracebackhide__ = True
+        _check_record_param_type("name", name)
+
+    xml = request.config.stash.get(xml_key, None)
+    if xml is not None:
+        record_func = xml.add_global_property
+    return record_func
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("terminal reporting")
+    group.addoption(
+        "--junitxml",
+        "--junit-xml",
+        action="store",
+        dest="xmlpath",
+        metavar="path",
+        type=functools.partial(filename_arg, optname="--junitxml"),
+        default=None,
+        help="Create junit-xml style report file at given path",
+    )
+    group.addoption(
+        "--junitprefix",
+        "--junit-prefix",
+        action="store",
+        metavar="str",
+        default=None,
+        help="Prepend prefix to classnames in junit-xml output",
+    )
+    parser.addini(
+        "junit_suite_name", "Test suite name for JUnit report", default="pytest"
+    )
+    parser.addini(
+        "junit_logging",
+        "Write captured log messages to JUnit report: "
+        "one of no|log|system-out|system-err|out-err|all",
+        default="no",
+    )
+    parser.addini(
+        "junit_log_passing_tests",
+        "Capture log information for passing tests to JUnit report: ",
+        type="bool",
+        default=True,
+    )
+    parser.addini(
+        "junit_duration_report",
+        "Duration time to report: one of total|call",
+        default="total",
+    )  # choices=['total', 'call'])
+    parser.addini(
+        "junit_family",
+        "Emit XML for schema: one of legacy|xunit1|xunit2",
+        default="xunit2",
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    xmlpath = config.option.xmlpath
+    # Prevent opening xmllog on worker nodes (xdist).
+    if xmlpath and not hasattr(config, "workerinput"):
+        junit_family = config.getini("junit_family")
+        config.stash[xml_key] = LogXML(
+            xmlpath,
+            config.option.junitprefix,
+            config.getini("junit_suite_name"),
+            config.getini("junit_logging"),
+            config.getini("junit_duration_report"),
+            junit_family,
+            config.getini("junit_log_passing_tests"),
+        )
+        config.pluginmanager.register(config.stash[xml_key])
+
+
+def pytest_unconfigure(config: Config) -> None:
+    xml = config.stash.get(xml_key, None)
+    if xml:
+        del config.stash[xml_key]
+        config.pluginmanager.unregister(xml)
+
+
+def mangle_test_address(address: str) -> list[str]:
+    path, possible_open_bracket, params = address.partition("[")
+    names = path.split("::")
+    # Convert file path to dotted path.
+    names[0] = names[0].replace(nodes.SEP, ".")
+    names[0] = re.sub(r"\.py$", "", names[0])
+    # Put any params back.
+    names[-1] += possible_open_bracket + params
+    return names
+
+
+class LogXML:
+    def __init__(
+        self,
+        logfile,
+        prefix: str | None,
+        suite_name: str = "pytest",
+        logging: str = "no",
+        report_duration: str = "total",
+        family="xunit1",
+        log_passing_tests: bool = True,
+    ) -> None:
+        logfile = os.path.expanduser(os.path.expandvars(logfile))
+        self.logfile = os.path.normpath(os.path.abspath(logfile))
+        self.prefix = prefix
+        self.suite_name = suite_name
+        self.logging = logging
+        self.log_passing_tests = log_passing_tests
+        self.report_duration = report_duration
+        self.family = family
+        self.stats: dict[str, int] = dict.fromkeys(
+            ["error", "passed", "failure", "skipped"], 0
+        )
+        self.node_reporters: dict[tuple[str | TestReport, object], _NodeReporter] = {}
+        self.node_reporters_ordered: list[_NodeReporter] = []
+        self.global_properties: list[tuple[str, str]] = []
+
+        # List of reports that failed on call but teardown is pending.
+        self.open_reports: list[TestReport] = []
+        self.cnt_double_fail_tests = 0
+
+        # Replaces convenience family with real family.
+        if self.family == "legacy":
+            self.family = "xunit1"
+
+    def finalize(self, report: TestReport) -> None:
+        nodeid = getattr(report, "nodeid", report)
+        # Local hack to handle xdist report order.
+        workernode = getattr(report, "node", None)
+        reporter = self.node_reporters.pop((nodeid, workernode))
+
+        for propname, propvalue in report.user_properties:
+            reporter.add_property(propname, str(propvalue))
+
+        if reporter is not None:
+            reporter.finalize()
+
+    def node_reporter(self, report: TestReport | str) -> _NodeReporter:
+        nodeid: str | TestReport = getattr(report, "nodeid", report)
+        # Local hack to handle xdist report order.
+        workernode = getattr(report, "node", None)
+
+        key = nodeid, workernode
+
+        if key in self.node_reporters:
+            # TODO: breaks for --dist=each
+            return self.node_reporters[key]
+
+        reporter = _NodeReporter(nodeid, self)
+
+        self.node_reporters[key] = reporter
+        self.node_reporters_ordered.append(reporter)
+
+        return reporter
+
+    def add_stats(self, key: str) -> None:
+        if key in self.stats:
+            self.stats[key] += 1
+
+    def _opentestcase(self, report: TestReport) -> _NodeReporter:
+        reporter = self.node_reporter(report)
+        reporter.record_testreport(report)
+        return reporter
+
+    def pytest_runtest_logreport(self, report: TestReport) -> None:
+        """Handle a setup/call/teardown report, generating the appropriate
+        XML tags as necessary.
+
+        Note: due to plugins like xdist, this hook may be called in interlaced
+        order with reports from other nodes. For example:
+
+        Usual call order:
+            -> setup node1
+            -> call node1
+            -> teardown node1
+            -> setup node2
+            -> call node2
+            -> teardown node2
+
+        Possible call order in xdist:
+            -> setup node1
+            -> call node1
+            -> setup node2
+            -> call node2
+            -> teardown node2
+            -> teardown node1
+        """
+        close_report = None
+        if report.passed:
+            if report.when == "call":  # ignore setup/teardown
+                reporter = self._opentestcase(report)
+                reporter.append_pass(report)
+        elif report.failed:
+            if report.when == "teardown":
+                # The following vars are needed when xdist plugin is used.
+                report_wid = getattr(report, "worker_id", None)
+                report_ii = getattr(report, "item_index", None)
+                close_report = next(
+                    (
+                        rep
+                        for rep in self.open_reports
+                        if (
+                            rep.nodeid == report.nodeid
+                            and getattr(rep, "item_index", None) == report_ii
+                            and getattr(rep, "worker_id", None) == report_wid
+                        )
+                    ),
+                    None,
+                )
+                if close_report:
+                    # We need to open new testcase in case we have failure in
+                    # call and error in teardown in order to follow junit
+                    # schema.
+                    self.finalize(close_report)
+                    self.cnt_double_fail_tests += 1
+            reporter = self._opentestcase(report)
+            if report.when == "call":
+                reporter.append_failure(report)
+                self.open_reports.append(report)
+                if not self.log_passing_tests:
+                    reporter.write_captured_output(report)
+            else:
+                reporter.append_error(report)
+        elif report.skipped:
+            reporter = self._opentestcase(report)
+            reporter.append_skipped(report)
+        self.update_testcase_duration(report)
+        if report.when == "teardown":
+            reporter = self._opentestcase(report)
+            reporter.write_captured_output(report)
+
+            self.finalize(report)
+            report_wid = getattr(report, "worker_id", None)
+            report_ii = getattr(report, "item_index", None)
+            close_report = next(
+                (
+                    rep
+                    for rep in self.open_reports
+                    if (
+                        rep.nodeid == report.nodeid
+                        and getattr(rep, "item_index", None) == report_ii
+                        and getattr(rep, "worker_id", None) == report_wid
+                    )
+                ),
+                None,
+            )
+            if close_report:
+                self.open_reports.remove(close_report)
+
+    def update_testcase_duration(self, report: TestReport) -> None:
+        """Accumulate total duration for nodeid from given report and update
+        the Junit.testcase with the new total if already created."""
+        if self.report_duration in {"total", report.when}:
+            reporter = self.node_reporter(report)
+            reporter.duration += getattr(report, "duration", 0.0)
+
+    def pytest_collectreport(self, report: TestReport) -> None:
+        if not report.passed:
+            reporter = self._opentestcase(report)
+            if report.failed:
+                reporter.append_collect_error(report)
+            else:
+                reporter.append_collect_skipped(report)
+
+    def pytest_internalerror(self, excrepr: ExceptionRepr) -> None:
+        reporter = self.node_reporter("internal")
+        reporter.attrs.update(classname="pytest", name="internal")
+        reporter._add_simple("error", "internal error", str(excrepr))
+
+    def pytest_sessionstart(self) -> None:
+        self.suite_start_time = timing.time()
+
+    def pytest_sessionfinish(self) -> None:
+        dirname = os.path.dirname(os.path.abspath(self.logfile))
+        # exist_ok avoids filesystem race conditions between checking path existence and requesting creation
+        os.makedirs(dirname, exist_ok=True)
+
+        with open(self.logfile, "w", encoding="utf-8") as logfile:
+            suite_stop_time = timing.time()
+            suite_time_delta = suite_stop_time - self.suite_start_time
+
+            numtests = (
+                self.stats["passed"]
+                + self.stats["failure"]
+                + self.stats["skipped"]
+                + self.stats["error"]
+                - self.cnt_double_fail_tests
+            )
+            logfile.write('<?xml version="1.0" encoding="utf-8"?>')
+
+            suite_node = ET.Element(
+                "testsuite",
+                name=self.suite_name,
+                errors=str(self.stats["error"]),
+                failures=str(self.stats["failure"]),
+                skipped=str(self.stats["skipped"]),
+                tests=str(numtests),
+                time=f"{suite_time_delta:.3f}",
+                timestamp=datetime.fromtimestamp(self.suite_start_time, timezone.utc)
+                .astimezone()
+                .isoformat(),
+                hostname=platform.node(),
+            )
+            global_properties = self._get_global_properties_node()
+            if global_properties is not None:
+                suite_node.append(global_properties)
+            for node_reporter in self.node_reporters_ordered:
+                suite_node.append(node_reporter.to_xml())
+            testsuites = ET.Element("testsuites")
+            testsuites.append(suite_node)
+            logfile.write(ET.tostring(testsuites, encoding="unicode"))
+
+    def pytest_terminal_summary(self, terminalreporter: TerminalReporter) -> None:
+        terminalreporter.write_sep("-", f"generated xml file: {self.logfile}")
+
+    def add_global_property(self, name: str, value: object) -> None:
+        __tracebackhide__ = True
+        _check_record_param_type("name", name)
+        self.global_properties.append((name, bin_xml_escape(value)))
+
+    def _get_global_properties_node(self) -> ET.Element | None:
+        """Return a Junit node containing custom properties, if any."""
+        if self.global_properties:
+            properties = ET.Element("properties")
+            for name, value in self.global_properties:
+                properties.append(ET.Element("property", name=name, value=value))
+            return properties
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/legacypath.py b/backend/venv/lib/python3.12/site-packages/_pytest/legacypath.py
new file mode 100644
index 00000000..59e8ef6e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/legacypath.py
@@ -0,0 +1,468 @@
+# mypy: allow-untyped-defs
+"""Add backward compatibility support for the legacy py path type."""
+
+from __future__ import annotations
+
+import dataclasses
+from pathlib import Path
+import shlex
+import subprocess
+from typing import Final
+from typing import final
+from typing import TYPE_CHECKING
+
+from iniconfig import SectionWrapper
+
+from _pytest.cacheprovider import Cache
+from _pytest.compat import LEGACY_PATH
+from _pytest.compat import legacy_path
+from _pytest.config import Config
+from _pytest.config import hookimpl
+from _pytest.config import PytestPluginManager
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureRequest
+from _pytest.main import Session
+from _pytest.monkeypatch import MonkeyPatch
+from _pytest.nodes import Collector
+from _pytest.nodes import Item
+from _pytest.nodes import Node
+from _pytest.pytester import HookRecorder
+from _pytest.pytester import Pytester
+from _pytest.pytester import RunResult
+from _pytest.terminal import TerminalReporter
+from _pytest.tmpdir import TempPathFactory
+
+
+if TYPE_CHECKING:
+    import pexpect
+
+
+@final
+class Testdir:
+    """
+    Similar to :class:`Pytester`, but this class works with legacy legacy_path objects instead.
+
+    All methods just forward to an internal :class:`Pytester` instance, converting results
+    to `legacy_path` objects as necessary.
+    """
+
+    __test__ = False
+
+    CLOSE_STDIN: Final = Pytester.CLOSE_STDIN
+    TimeoutExpired: Final = Pytester.TimeoutExpired
+
+    def __init__(self, pytester: Pytester, *, _ispytest: bool = False) -> None:
+        check_ispytest(_ispytest)
+        self._pytester = pytester
+
+    @property
+    def tmpdir(self) -> LEGACY_PATH:
+        """Temporary directory where tests are executed."""
+        return legacy_path(self._pytester.path)
+
+    @property
+    def test_tmproot(self) -> LEGACY_PATH:
+        return legacy_path(self._pytester._test_tmproot)
+
+    @property
+    def request(self):
+        return self._pytester._request
+
+    @property
+    def plugins(self):
+        return self._pytester.plugins
+
+    @plugins.setter
+    def plugins(self, plugins):
+        self._pytester.plugins = plugins
+
+    @property
+    def monkeypatch(self) -> MonkeyPatch:
+        return self._pytester._monkeypatch
+
+    def make_hook_recorder(self, pluginmanager) -> HookRecorder:
+        """See :meth:`Pytester.make_hook_recorder`."""
+        return self._pytester.make_hook_recorder(pluginmanager)
+
+    def chdir(self) -> None:
+        """See :meth:`Pytester.chdir`."""
+        return self._pytester.chdir()
+
+    def finalize(self) -> None:
+        return self._pytester._finalize()
+
+    def makefile(self, ext, *args, **kwargs) -> LEGACY_PATH:
+        """See :meth:`Pytester.makefile`."""
+        if ext and not ext.startswith("."):
+            # pytester.makefile is going to throw a ValueError in a way that
+            # testdir.makefile did not, because
+            # pathlib.Path is stricter suffixes than py.path
+            # This ext arguments is likely user error, but since testdir has
+            # allowed this, we will prepend "." as a workaround to avoid breaking
+            # testdir usage that worked before
+            ext = "." + ext
+        return legacy_path(self._pytester.makefile(ext, *args, **kwargs))
+
+    def makeconftest(self, source) -> LEGACY_PATH:
+        """See :meth:`Pytester.makeconftest`."""
+        return legacy_path(self._pytester.makeconftest(source))
+
+    def makeini(self, source) -> LEGACY_PATH:
+        """See :meth:`Pytester.makeini`."""
+        return legacy_path(self._pytester.makeini(source))
+
+    def getinicfg(self, source: str) -> SectionWrapper:
+        """See :meth:`Pytester.getinicfg`."""
+        return self._pytester.getinicfg(source)
+
+    def makepyprojecttoml(self, source) -> LEGACY_PATH:
+        """See :meth:`Pytester.makepyprojecttoml`."""
+        return legacy_path(self._pytester.makepyprojecttoml(source))
+
+    def makepyfile(self, *args, **kwargs) -> LEGACY_PATH:
+        """See :meth:`Pytester.makepyfile`."""
+        return legacy_path(self._pytester.makepyfile(*args, **kwargs))
+
+    def maketxtfile(self, *args, **kwargs) -> LEGACY_PATH:
+        """See :meth:`Pytester.maketxtfile`."""
+        return legacy_path(self._pytester.maketxtfile(*args, **kwargs))
+
+    def syspathinsert(self, path=None) -> None:
+        """See :meth:`Pytester.syspathinsert`."""
+        return self._pytester.syspathinsert(path)
+
+    def mkdir(self, name) -> LEGACY_PATH:
+        """See :meth:`Pytester.mkdir`."""
+        return legacy_path(self._pytester.mkdir(name))
+
+    def mkpydir(self, name) -> LEGACY_PATH:
+        """See :meth:`Pytester.mkpydir`."""
+        return legacy_path(self._pytester.mkpydir(name))
+
+    def copy_example(self, name=None) -> LEGACY_PATH:
+        """See :meth:`Pytester.copy_example`."""
+        return legacy_path(self._pytester.copy_example(name))
+
+    def getnode(self, config: Config, arg) -> Item | Collector | None:
+        """See :meth:`Pytester.getnode`."""
+        return self._pytester.getnode(config, arg)
+
+    def getpathnode(self, path):
+        """See :meth:`Pytester.getpathnode`."""
+        return self._pytester.getpathnode(path)
+
+    def genitems(self, colitems: list[Item | Collector]) -> list[Item]:
+        """See :meth:`Pytester.genitems`."""
+        return self._pytester.genitems(colitems)
+
+    def runitem(self, source):
+        """See :meth:`Pytester.runitem`."""
+        return self._pytester.runitem(source)
+
+    def inline_runsource(self, source, *cmdlineargs):
+        """See :meth:`Pytester.inline_runsource`."""
+        return self._pytester.inline_runsource(source, *cmdlineargs)
+
+    def inline_genitems(self, *args):
+        """See :meth:`Pytester.inline_genitems`."""
+        return self._pytester.inline_genitems(*args)
+
+    def inline_run(self, *args, plugins=(), no_reraise_ctrlc: bool = False):
+        """See :meth:`Pytester.inline_run`."""
+        return self._pytester.inline_run(
+            *args, plugins=plugins, no_reraise_ctrlc=no_reraise_ctrlc
+        )
+
+    def runpytest_inprocess(self, *args, **kwargs) -> RunResult:
+        """See :meth:`Pytester.runpytest_inprocess`."""
+        return self._pytester.runpytest_inprocess(*args, **kwargs)
+
+    def runpytest(self, *args, **kwargs) -> RunResult:
+        """See :meth:`Pytester.runpytest`."""
+        return self._pytester.runpytest(*args, **kwargs)
+
+    def parseconfig(self, *args) -> Config:
+        """See :meth:`Pytester.parseconfig`."""
+        return self._pytester.parseconfig(*args)
+
+    def parseconfigure(self, *args) -> Config:
+        """See :meth:`Pytester.parseconfigure`."""
+        return self._pytester.parseconfigure(*args)
+
+    def getitem(self, source, funcname="test_func"):
+        """See :meth:`Pytester.getitem`."""
+        return self._pytester.getitem(source, funcname)
+
+    def getitems(self, source):
+        """See :meth:`Pytester.getitems`."""
+        return self._pytester.getitems(source)
+
+    def getmodulecol(self, source, configargs=(), withinit=False):
+        """See :meth:`Pytester.getmodulecol`."""
+        return self._pytester.getmodulecol(
+            source, configargs=configargs, withinit=withinit
+        )
+
+    def collect_by_name(self, modcol: Collector, name: str) -> Item | Collector | None:
+        """See :meth:`Pytester.collect_by_name`."""
+        return self._pytester.collect_by_name(modcol, name)
+
+    def popen(
+        self,
+        cmdargs,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.PIPE,
+        stdin=CLOSE_STDIN,
+        **kw,
+    ):
+        """See :meth:`Pytester.popen`."""
+        return self._pytester.popen(cmdargs, stdout, stderr, stdin, **kw)
+
+    def run(self, *cmdargs, timeout=None, stdin=CLOSE_STDIN) -> RunResult:
+        """See :meth:`Pytester.run`."""
+        return self._pytester.run(*cmdargs, timeout=timeout, stdin=stdin)
+
+    def runpython(self, script) -> RunResult:
+        """See :meth:`Pytester.runpython`."""
+        return self._pytester.runpython(script)
+
+    def runpython_c(self, command):
+        """See :meth:`Pytester.runpython_c`."""
+        return self._pytester.runpython_c(command)
+
+    def runpytest_subprocess(self, *args, timeout=None) -> RunResult:
+        """See :meth:`Pytester.runpytest_subprocess`."""
+        return self._pytester.runpytest_subprocess(*args, timeout=timeout)
+
+    def spawn_pytest(self, string: str, expect_timeout: float = 10.0) -> pexpect.spawn:
+        """See :meth:`Pytester.spawn_pytest`."""
+        return self._pytester.spawn_pytest(string, expect_timeout=expect_timeout)
+
+    def spawn(self, cmd: str, expect_timeout: float = 10.0) -> pexpect.spawn:
+        """See :meth:`Pytester.spawn`."""
+        return self._pytester.spawn(cmd, expect_timeout=expect_timeout)
+
+    def __repr__(self) -> str:
+        return f"<Testdir {self.tmpdir!r}>"
+
+    def __str__(self) -> str:
+        return str(self.tmpdir)
+
+
+class LegacyTestdirPlugin:
+    @staticmethod
+    @fixture
+    def testdir(pytester: Pytester) -> Testdir:
+        """
+        Identical to :fixture:`pytester`, and provides an instance whose methods return
+        legacy ``LEGACY_PATH`` objects instead when applicable.
+
+        New code should avoid using :fixture:`testdir` in favor of :fixture:`pytester`.
+        """
+        return Testdir(pytester, _ispytest=True)
+
+
+@final
+@dataclasses.dataclass
+class TempdirFactory:
+    """Backward compatibility wrapper that implements ``py.path.local``
+    for :class:`TempPathFactory`.
+
+    .. note::
+        These days, it is preferred to use ``tmp_path_factory``.
+
+        :ref:`About the tmpdir and tmpdir_factory fixtures<tmpdir and tmpdir_factory>`.
+
+    """
+
+    _tmppath_factory: TempPathFactory
+
+    def __init__(
+        self, tmppath_factory: TempPathFactory, *, _ispytest: bool = False
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._tmppath_factory = tmppath_factory
+
+    def mktemp(self, basename: str, numbered: bool = True) -> LEGACY_PATH:
+        """Same as :meth:`TempPathFactory.mktemp`, but returns a ``py.path.local`` object."""
+        return legacy_path(self._tmppath_factory.mktemp(basename, numbered).resolve())
+
+    def getbasetemp(self) -> LEGACY_PATH:
+        """Same as :meth:`TempPathFactory.getbasetemp`, but returns a ``py.path.local`` object."""
+        return legacy_path(self._tmppath_factory.getbasetemp().resolve())
+
+
+class LegacyTmpdirPlugin:
+    @staticmethod
+    @fixture(scope="session")
+    def tmpdir_factory(request: FixtureRequest) -> TempdirFactory:
+        """Return a :class:`pytest.TempdirFactory` instance for the test session."""
+        # Set dynamically by pytest_configure().
+        return request.config._tmpdirhandler  # type: ignore
+
+    @staticmethod
+    @fixture
+    def tmpdir(tmp_path: Path) -> LEGACY_PATH:
+        """Return a temporary directory (as `legacy_path`_ object)
+        which is unique to each test function invocation.
+        The temporary directory is created as a subdirectory
+        of the base temporary directory, with configurable retention,
+        as discussed in :ref:`temporary directory location and retention`.
+
+        .. note::
+            These days, it is preferred to use ``tmp_path``.
+
+            :ref:`About the tmpdir and tmpdir_factory fixtures<tmpdir and tmpdir_factory>`.
+
+        .. _legacy_path: https://py.readthedocs.io/en/latest/path.html
+        """
+        return legacy_path(tmp_path)
+
+
+def Cache_makedir(self: Cache, name: str) -> LEGACY_PATH:
+    """Return a directory path object with the given name.
+
+    Same as :func:`mkdir`, but returns a legacy py path instance.
+    """
+    return legacy_path(self.mkdir(name))
+
+
+def FixtureRequest_fspath(self: FixtureRequest) -> LEGACY_PATH:
+    """(deprecated) The file system path of the test module which collected this test."""
+    return legacy_path(self.path)
+
+
+def TerminalReporter_startdir(self: TerminalReporter) -> LEGACY_PATH:
+    """The directory from which pytest was invoked.
+
+    Prefer to use ``startpath`` which is a :class:`pathlib.Path`.
+
+    :type: LEGACY_PATH
+    """
+    return legacy_path(self.startpath)
+
+
+def Config_invocation_dir(self: Config) -> LEGACY_PATH:
+    """The directory from which pytest was invoked.
+
+    Prefer to use :attr:`invocation_params.dir <InvocationParams.dir>`,
+    which is a :class:`pathlib.Path`.
+
+    :type: LEGACY_PATH
+    """
+    return legacy_path(str(self.invocation_params.dir))
+
+
+def Config_rootdir(self: Config) -> LEGACY_PATH:
+    """The path to the :ref:`rootdir <rootdir>`.
+
+    Prefer to use :attr:`rootpath`, which is a :class:`pathlib.Path`.
+
+    :type: LEGACY_PATH
+    """
+    return legacy_path(str(self.rootpath))
+
+
+def Config_inifile(self: Config) -> LEGACY_PATH | None:
+    """The path to the :ref:`configfile <configfiles>`.
+
+    Prefer to use :attr:`inipath`, which is a :class:`pathlib.Path`.
+
+    :type: Optional[LEGACY_PATH]
+    """
+    return legacy_path(str(self.inipath)) if self.inipath else None
+
+
+def Session_startdir(self: Session) -> LEGACY_PATH:
+    """The path from which pytest was invoked.
+
+    Prefer to use ``startpath`` which is a :class:`pathlib.Path`.
+
+    :type: LEGACY_PATH
+    """
+    return legacy_path(self.startpath)
+
+
+def Config__getini_unknown_type(self, name: str, type: str, value: str | list[str]):
+    if type == "pathlist":
+        # TODO: This assert is probably not valid in all cases.
+        assert self.inipath is not None
+        dp = self.inipath.parent
+        input_values = shlex.split(value) if isinstance(value, str) else value
+        return [legacy_path(str(dp / x)) for x in input_values]
+    else:
+        raise ValueError(f"unknown configuration type: {type}", value)
+
+
+def Node_fspath(self: Node) -> LEGACY_PATH:
+    """(deprecated) returns a legacy_path copy of self.path"""
+    return legacy_path(self.path)
+
+
+def Node_fspath_set(self: Node, value: LEGACY_PATH) -> None:
+    self.path = Path(value)
+
+
+@hookimpl(tryfirst=True)
+def pytest_load_initial_conftests(early_config: Config) -> None:
+    """Monkeypatch legacy path attributes in several classes, as early as possible."""
+    mp = MonkeyPatch()
+    early_config.add_cleanup(mp.undo)
+
+    # Add Cache.makedir().
+    mp.setattr(Cache, "makedir", Cache_makedir, raising=False)
+
+    # Add FixtureRequest.fspath property.
+    mp.setattr(FixtureRequest, "fspath", property(FixtureRequest_fspath), raising=False)
+
+    # Add TerminalReporter.startdir property.
+    mp.setattr(
+        TerminalReporter, "startdir", property(TerminalReporter_startdir), raising=False
+    )
+
+    # Add Config.{invocation_dir,rootdir,inifile} properties.
+    mp.setattr(Config, "invocation_dir", property(Config_invocation_dir), raising=False)
+    mp.setattr(Config, "rootdir", property(Config_rootdir), raising=False)
+    mp.setattr(Config, "inifile", property(Config_inifile), raising=False)
+
+    # Add Session.startdir property.
+    mp.setattr(Session, "startdir", property(Session_startdir), raising=False)
+
+    # Add pathlist configuration type.
+    mp.setattr(Config, "_getini_unknown_type", Config__getini_unknown_type)
+
+    # Add Node.fspath property.
+    mp.setattr(Node, "fspath", property(Node_fspath, Node_fspath_set), raising=False)
+
+
+@hookimpl
+def pytest_configure(config: Config) -> None:
+    """Installs the LegacyTmpdirPlugin if the ``tmpdir`` plugin is also installed."""
+    if config.pluginmanager.has_plugin("tmpdir"):
+        mp = MonkeyPatch()
+        config.add_cleanup(mp.undo)
+        # Create TmpdirFactory and attach it to the config object.
+        #
+        # This is to comply with existing plugins which expect the handler to be
+        # available at pytest_configure time, but ideally should be moved entirely
+        # to the tmpdir_factory session fixture.
+        try:
+            tmp_path_factory = config._tmp_path_factory  # type: ignore[attr-defined]
+        except AttributeError:
+            # tmpdir plugin is blocked.
+            pass
+        else:
+            _tmpdirhandler = TempdirFactory(tmp_path_factory, _ispytest=True)
+            mp.setattr(config, "_tmpdirhandler", _tmpdirhandler, raising=False)
+
+        config.pluginmanager.register(LegacyTmpdirPlugin, "legacypath-tmpdir")
+
+
+@hookimpl
+def pytest_plugin_registered(plugin: object, manager: PytestPluginManager) -> None:
+    # pytester is not loaded by default and is commonly loaded from a conftest,
+    # so checking for it in `pytest_configure` is not enough.
+    is_pytester = plugin is manager.get_plugin("pytester")
+    if is_pytester and not manager.is_registered(LegacyTestdirPlugin):
+        manager.register(LegacyTestdirPlugin, "legacypath-pytester")
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/logging.py b/backend/venv/lib/python3.12/site-packages/_pytest/logging.py
new file mode 100644
index 00000000..08c826ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/logging.py
@@ -0,0 +1,955 @@
+# mypy: allow-untyped-defs
+"""Access and control log capturing."""
+
+from __future__ import annotations
+
+from contextlib import contextmanager
+from contextlib import nullcontext
+from datetime import datetime
+from datetime import timedelta
+from datetime import timezone
+import io
+from io import StringIO
+import logging
+from logging import LogRecord
+import os
+from pathlib import Path
+import re
+from types import TracebackType
+from typing import AbstractSet
+from typing import Dict
+from typing import final
+from typing import Generator
+from typing import Generic
+from typing import List
+from typing import Literal
+from typing import Mapping
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from _pytest import nodes
+from _pytest._io import TerminalWriter
+from _pytest.capture import CaptureManager
+from _pytest.config import _strtobool
+from _pytest.config import Config
+from _pytest.config import create_terminal_writer
+from _pytest.config import hookimpl
+from _pytest.config import UsageError
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureRequest
+from _pytest.main import Session
+from _pytest.stash import StashKey
+from _pytest.terminal import TerminalReporter
+
+
+if TYPE_CHECKING:
+    logging_StreamHandler = logging.StreamHandler[StringIO]
+else:
+    logging_StreamHandler = logging.StreamHandler
+
+DEFAULT_LOG_FORMAT = "%(levelname)-8s %(name)s:%(filename)s:%(lineno)d %(message)s"
+DEFAULT_LOG_DATE_FORMAT = "%H:%M:%S"
+_ANSI_ESCAPE_SEQ = re.compile(r"\x1b\[[\d;]+m")
+caplog_handler_key = StashKey["LogCaptureHandler"]()
+caplog_records_key = StashKey[Dict[str, List[logging.LogRecord]]]()
+
+
+def _remove_ansi_escape_sequences(text: str) -> str:
+    return _ANSI_ESCAPE_SEQ.sub("", text)
+
+
+class DatetimeFormatter(logging.Formatter):
+    """A logging formatter which formats record with
+    :func:`datetime.datetime.strftime` formatter instead of
+    :func:`time.strftime` in case of microseconds in format string.
+    """
+
+    def formatTime(self, record: LogRecord, datefmt: str | None = None) -> str:
+        if datefmt and "%f" in datefmt:
+            ct = self.converter(record.created)
+            tz = timezone(timedelta(seconds=ct.tm_gmtoff), ct.tm_zone)
+            # Construct `datetime.datetime` object from `struct_time`
+            # and msecs information from `record`
+            # Using int() instead of round() to avoid it exceeding 1_000_000 and causing a ValueError (#11861).
+            dt = datetime(*ct[0:6], microsecond=int(record.msecs * 1000), tzinfo=tz)
+            return dt.strftime(datefmt)
+        # Use `logging.Formatter` for non-microsecond formats
+        return super().formatTime(record, datefmt)
+
+
+class ColoredLevelFormatter(DatetimeFormatter):
+    """A logging formatter which colorizes the %(levelname)..s part of the
+    log format passed to __init__."""
+
+    LOGLEVEL_COLOROPTS: Mapping[int, AbstractSet[str]] = {
+        logging.CRITICAL: {"red"},
+        logging.ERROR: {"red", "bold"},
+        logging.WARNING: {"yellow"},
+        logging.WARN: {"yellow"},
+        logging.INFO: {"green"},
+        logging.DEBUG: {"purple"},
+        logging.NOTSET: set(),
+    }
+    LEVELNAME_FMT_REGEX = re.compile(r"%\(levelname\)([+-.]?\d*(?:\.\d+)?s)")
+
+    def __init__(self, terminalwriter: TerminalWriter, *args, **kwargs) -> None:
+        super().__init__(*args, **kwargs)
+        self._terminalwriter = terminalwriter
+        self._original_fmt = self._style._fmt
+        self._level_to_fmt_mapping: dict[int, str] = {}
+
+        for level, color_opts in self.LOGLEVEL_COLOROPTS.items():
+            self.add_color_level(level, *color_opts)
+
+    def add_color_level(self, level: int, *color_opts: str) -> None:
+        """Add or update color opts for a log level.
+
+        :param level:
+            Log level to apply a style to, e.g. ``logging.INFO``.
+        :param color_opts:
+            ANSI escape sequence color options. Capitalized colors indicates
+            background color, i.e. ``'green', 'Yellow', 'bold'`` will give bold
+            green text on yellow background.
+
+        .. warning::
+            This is an experimental API.
+        """
+        assert self._fmt is not None
+        levelname_fmt_match = self.LEVELNAME_FMT_REGEX.search(self._fmt)
+        if not levelname_fmt_match:
+            return
+        levelname_fmt = levelname_fmt_match.group()
+
+        formatted_levelname = levelname_fmt % {"levelname": logging.getLevelName(level)}
+
+        # add ANSI escape sequences around the formatted levelname
+        color_kwargs = {name: True for name in color_opts}
+        colorized_formatted_levelname = self._terminalwriter.markup(
+            formatted_levelname, **color_kwargs
+        )
+        self._level_to_fmt_mapping[level] = self.LEVELNAME_FMT_REGEX.sub(
+            colorized_formatted_levelname, self._fmt
+        )
+
+    def format(self, record: logging.LogRecord) -> str:
+        fmt = self._level_to_fmt_mapping.get(record.levelno, self._original_fmt)
+        self._style._fmt = fmt
+        return super().format(record)
+
+
+class PercentStyleMultiline(logging.PercentStyle):
+    """A logging style with special support for multiline messages.
+
+    If the message of a record consists of multiple lines, this style
+    formats the message as if each line were logged separately.
+    """
+
+    def __init__(self, fmt: str, auto_indent: int | str | bool | None) -> None:
+        super().__init__(fmt)
+        self._auto_indent = self._get_auto_indent(auto_indent)
+
+    @staticmethod
+    def _get_auto_indent(auto_indent_option: int | str | bool | None) -> int:
+        """Determine the current auto indentation setting.
+
+        Specify auto indent behavior (on/off/fixed) by passing in
+        extra={"auto_indent": [value]} to the call to logging.log() or
+        using a --log-auto-indent [value] command line or the
+        log_auto_indent [value] config option.
+
+        Default behavior is auto-indent off.
+
+        Using the string "True" or "on" or the boolean True as the value
+        turns auto indent on, using the string "False" or "off" or the
+        boolean False or the int 0 turns it off, and specifying a
+        positive integer fixes the indentation position to the value
+        specified.
+
+        Any other values for the option are invalid, and will silently be
+        converted to the default.
+
+        :param None|bool|int|str auto_indent_option:
+            User specified option for indentation from command line, config
+            or extra kwarg. Accepts int, bool or str. str option accepts the
+            same range of values as boolean config options, as well as
+            positive integers represented in str form.
+
+        :returns:
+            Indentation value, which can be
+            -1 (automatically determine indentation) or
+            0 (auto-indent turned off) or
+            >0 (explicitly set indentation position).
+        """
+        if auto_indent_option is None:
+            return 0
+        elif isinstance(auto_indent_option, bool):
+            if auto_indent_option:
+                return -1
+            else:
+                return 0
+        elif isinstance(auto_indent_option, int):
+            return int(auto_indent_option)
+        elif isinstance(auto_indent_option, str):
+            try:
+                return int(auto_indent_option)
+            except ValueError:
+                pass
+            try:
+                if _strtobool(auto_indent_option):
+                    return -1
+            except ValueError:
+                return 0
+
+        return 0
+
+    def format(self, record: logging.LogRecord) -> str:
+        if "\n" in record.message:
+            if hasattr(record, "auto_indent"):
+                # Passed in from the "extra={}" kwarg on the call to logging.log().
+                auto_indent = self._get_auto_indent(record.auto_indent)
+            else:
+                auto_indent = self._auto_indent
+
+            if auto_indent:
+                lines = record.message.splitlines()
+                formatted = self._fmt % {**record.__dict__, "message": lines[0]}
+
+                if auto_indent < 0:
+                    indentation = _remove_ansi_escape_sequences(formatted).find(
+                        lines[0]
+                    )
+                else:
+                    # Optimizes logging by allowing a fixed indentation.
+                    indentation = auto_indent
+                lines[0] = formatted
+                return ("\n" + " " * indentation).join(lines)
+        return self._fmt % record.__dict__
+
+
+def get_option_ini(config: Config, *names: str):
+    for name in names:
+        ret = config.getoption(name)  # 'default' arg won't work as expected
+        if ret is None:
+            ret = config.getini(name)
+        if ret:
+            return ret
+
+
+def pytest_addoption(parser: Parser) -> None:
+    """Add options to control log capturing."""
+    group = parser.getgroup("logging")
+
+    def add_option_ini(option, dest, default=None, type=None, **kwargs):
+        parser.addini(
+            dest, default=default, type=type, help="Default value for " + option
+        )
+        group.addoption(option, dest=dest, **kwargs)
+
+    add_option_ini(
+        "--log-level",
+        dest="log_level",
+        default=None,
+        metavar="LEVEL",
+        help=(
+            "Level of messages to catch/display."
+            " Not set by default, so it depends on the root/parent log handler's"
+            ' effective level, where it is "WARNING" by default.'
+        ),
+    )
+    add_option_ini(
+        "--log-format",
+        dest="log_format",
+        default=DEFAULT_LOG_FORMAT,
+        help="Log format used by the logging module",
+    )
+    add_option_ini(
+        "--log-date-format",
+        dest="log_date_format",
+        default=DEFAULT_LOG_DATE_FORMAT,
+        help="Log date format used by the logging module",
+    )
+    parser.addini(
+        "log_cli",
+        default=False,
+        type="bool",
+        help='Enable log display during test run (also known as "live logging")',
+    )
+    add_option_ini(
+        "--log-cli-level", dest="log_cli_level", default=None, help="CLI logging level"
+    )
+    add_option_ini(
+        "--log-cli-format",
+        dest="log_cli_format",
+        default=None,
+        help="Log format used by the logging module",
+    )
+    add_option_ini(
+        "--log-cli-date-format",
+        dest="log_cli_date_format",
+        default=None,
+        help="Log date format used by the logging module",
+    )
+    add_option_ini(
+        "--log-file",
+        dest="log_file",
+        default=None,
+        help="Path to a file when logging will be written to",
+    )
+    add_option_ini(
+        "--log-file-mode",
+        dest="log_file_mode",
+        default="w",
+        choices=["w", "a"],
+        help="Log file open mode",
+    )
+    add_option_ini(
+        "--log-file-level",
+        dest="log_file_level",
+        default=None,
+        help="Log file logging level",
+    )
+    add_option_ini(
+        "--log-file-format",
+        dest="log_file_format",
+        default=None,
+        help="Log format used by the logging module",
+    )
+    add_option_ini(
+        "--log-file-date-format",
+        dest="log_file_date_format",
+        default=None,
+        help="Log date format used by the logging module",
+    )
+    add_option_ini(
+        "--log-auto-indent",
+        dest="log_auto_indent",
+        default=None,
+        help="Auto-indent multiline messages passed to the logging module. Accepts true|on, false|off or an integer.",
+    )
+    group.addoption(
+        "--log-disable",
+        action="append",
+        default=[],
+        dest="logger_disable",
+        help="Disable a logger by name. Can be passed multiple times.",
+    )
+
+
+_HandlerType = TypeVar("_HandlerType", bound=logging.Handler)
+
+
+# Not using @contextmanager for performance reasons.
+class catching_logs(Generic[_HandlerType]):
+    """Context manager that prepares the whole logging machinery properly."""
+
+    __slots__ = ("handler", "level", "orig_level")
+
+    def __init__(self, handler: _HandlerType, level: int | None = None) -> None:
+        self.handler = handler
+        self.level = level
+
+    def __enter__(self) -> _HandlerType:
+        root_logger = logging.getLogger()
+        if self.level is not None:
+            self.handler.setLevel(self.level)
+        root_logger.addHandler(self.handler)
+        if self.level is not None:
+            self.orig_level = root_logger.level
+            root_logger.setLevel(min(self.orig_level, self.level))
+        return self.handler
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        root_logger = logging.getLogger()
+        if self.level is not None:
+            root_logger.setLevel(self.orig_level)
+        root_logger.removeHandler(self.handler)
+
+
+class LogCaptureHandler(logging_StreamHandler):
+    """A logging handler that stores log records and the log text."""
+
+    def __init__(self) -> None:
+        """Create a new log handler."""
+        super().__init__(StringIO())
+        self.records: list[logging.LogRecord] = []
+
+    def emit(self, record: logging.LogRecord) -> None:
+        """Keep the log records in a list in addition to the log text."""
+        self.records.append(record)
+        super().emit(record)
+
+    def reset(self) -> None:
+        self.records = []
+        self.stream = StringIO()
+
+    def clear(self) -> None:
+        self.records.clear()
+        self.stream = StringIO()
+
+    def handleError(self, record: logging.LogRecord) -> None:
+        if logging.raiseExceptions:
+            # Fail the test if the log message is bad (emit failed).
+            # The default behavior of logging is to print "Logging error"
+            # to stderr with the call stack and some extra details.
+            # pytest wants to make such mistakes visible during testing.
+            raise  # noqa: PLE0704
+
+
+@final
+class LogCaptureFixture:
+    """Provides access and control of log capturing."""
+
+    def __init__(self, item: nodes.Node, *, _ispytest: bool = False) -> None:
+        check_ispytest(_ispytest)
+        self._item = item
+        self._initial_handler_level: int | None = None
+        # Dict of log name -> log level.
+        self._initial_logger_levels: dict[str | None, int] = {}
+        self._initial_disabled_logging_level: int | None = None
+
+    def _finalize(self) -> None:
+        """Finalize the fixture.
+
+        This restores the log levels and the disabled logging levels changed by :meth:`set_level`.
+        """
+        # Restore log levels.
+        if self._initial_handler_level is not None:
+            self.handler.setLevel(self._initial_handler_level)
+        for logger_name, level in self._initial_logger_levels.items():
+            logger = logging.getLogger(logger_name)
+            logger.setLevel(level)
+        # Disable logging at the original disabled logging level.
+        if self._initial_disabled_logging_level is not None:
+            logging.disable(self._initial_disabled_logging_level)
+            self._initial_disabled_logging_level = None
+
+    @property
+    def handler(self) -> LogCaptureHandler:
+        """Get the logging handler used by the fixture."""
+        return self._item.stash[caplog_handler_key]
+
+    def get_records(
+        self, when: Literal["setup", "call", "teardown"]
+    ) -> list[logging.LogRecord]:
+        """Get the logging records for one of the possible test phases.
+
+        :param when:
+            Which test phase to obtain the records from.
+            Valid values are: "setup", "call" and "teardown".
+
+        :returns: The list of captured records at the given stage.
+
+        .. versionadded:: 3.4
+        """
+        return self._item.stash[caplog_records_key].get(when, [])
+
+    @property
+    def text(self) -> str:
+        """The formatted log text."""
+        return _remove_ansi_escape_sequences(self.handler.stream.getvalue())
+
+    @property
+    def records(self) -> list[logging.LogRecord]:
+        """The list of log records."""
+        return self.handler.records
+
+    @property
+    def record_tuples(self) -> list[tuple[str, int, str]]:
+        """A list of a stripped down version of log records intended
+        for use in assertion comparison.
+
+        The format of the tuple is:
+
+            (logger_name, log_level, message)
+        """
+        return [(r.name, r.levelno, r.getMessage()) for r in self.records]
+
+    @property
+    def messages(self) -> list[str]:
+        """A list of format-interpolated log messages.
+
+        Unlike 'records', which contains the format string and parameters for
+        interpolation, log messages in this list are all interpolated.
+
+        Unlike 'text', which contains the output from the handler, log
+        messages in this list are unadorned with levels, timestamps, etc,
+        making exact comparisons more reliable.
+
+        Note that traceback or stack info (from :func:`logging.exception` or
+        the `exc_info` or `stack_info` arguments to the logging functions) is
+        not included, as this is added by the formatter in the handler.
+
+        .. versionadded:: 3.7
+        """
+        return [r.getMessage() for r in self.records]
+
+    def clear(self) -> None:
+        """Reset the list of log records and the captured log text."""
+        self.handler.clear()
+
+    def _force_enable_logging(
+        self, level: int | str, logger_obj: logging.Logger
+    ) -> int:
+        """Enable the desired logging level if the global level was disabled via ``logging.disabled``.
+
+        Only enables logging levels greater than or equal to the requested ``level``.
+
+        Does nothing if the desired ``level`` wasn't disabled.
+
+        :param level:
+            The logger level caplog should capture.
+            All logging is enabled if a non-standard logging level string is supplied.
+            Valid level strings are in :data:`logging._nameToLevel`.
+        :param logger_obj: The logger object to check.
+
+        :return: The original disabled logging level.
+        """
+        original_disable_level: int = logger_obj.manager.disable
+
+        if isinstance(level, str):
+            # Try to translate the level string to an int for `logging.disable()`
+            level = logging.getLevelName(level)
+
+        if not isinstance(level, int):
+            # The level provided was not valid, so just un-disable all logging.
+            logging.disable(logging.NOTSET)
+        elif not logger_obj.isEnabledFor(level):
+            # Each level is `10` away from other levels.
+            # https://docs.python.org/3/library/logging.html#logging-levels
+            disable_level = max(level - 10, logging.NOTSET)
+            logging.disable(disable_level)
+
+        return original_disable_level
+
+    def set_level(self, level: int | str, logger: str | None = None) -> None:
+        """Set the threshold level of a logger for the duration of a test.
+
+        Logging messages which are less severe than this level will not be captured.
+
+        .. versionchanged:: 3.4
+            The levels of the loggers changed by this function will be
+            restored to their initial values at the end of the test.
+
+        Will enable the requested logging level if it was disabled via :func:`logging.disable`.
+
+        :param level: The level.
+        :param logger: The logger to update. If not given, the root logger.
+        """
+        logger_obj = logging.getLogger(logger)
+        # Save the original log-level to restore it during teardown.
+        self._initial_logger_levels.setdefault(logger, logger_obj.level)
+        logger_obj.setLevel(level)
+        if self._initial_handler_level is None:
+            self._initial_handler_level = self.handler.level
+        self.handler.setLevel(level)
+        initial_disabled_logging_level = self._force_enable_logging(level, logger_obj)
+        if self._initial_disabled_logging_level is None:
+            self._initial_disabled_logging_level = initial_disabled_logging_level
+
+    @contextmanager
+    def at_level(self, level: int | str, logger: str | None = None) -> Generator[None]:
+        """Context manager that sets the level for capturing of logs. After
+        the end of the 'with' statement the level is restored to its original
+        value.
+
+        Will enable the requested logging level if it was disabled via :func:`logging.disable`.
+
+        :param level: The level.
+        :param logger: The logger to update. If not given, the root logger.
+        """
+        logger_obj = logging.getLogger(logger)
+        orig_level = logger_obj.level
+        logger_obj.setLevel(level)
+        handler_orig_level = self.handler.level
+        self.handler.setLevel(level)
+        original_disable_level = self._force_enable_logging(level, logger_obj)
+        try:
+            yield
+        finally:
+            logger_obj.setLevel(orig_level)
+            self.handler.setLevel(handler_orig_level)
+            logging.disable(original_disable_level)
+
+    @contextmanager
+    def filtering(self, filter_: logging.Filter) -> Generator[None]:
+        """Context manager that temporarily adds the given filter to the caplog's
+        :meth:`handler` for the 'with' statement block, and removes that filter at the
+        end of the block.
+
+        :param filter_: A custom :class:`logging.Filter` object.
+
+        .. versionadded:: 7.5
+        """
+        self.handler.addFilter(filter_)
+        try:
+            yield
+        finally:
+            self.handler.removeFilter(filter_)
+
+
+@fixture
+def caplog(request: FixtureRequest) -> Generator[LogCaptureFixture]:
+    """Access and control log capturing.
+
+    Captured logs are available through the following properties/methods::
+
+    * caplog.messages        -> list of format-interpolated log messages
+    * caplog.text            -> string containing formatted log output
+    * caplog.records         -> list of logging.LogRecord instances
+    * caplog.record_tuples   -> list of (logger_name, level, message) tuples
+    * caplog.clear()         -> clear captured records and formatted log output string
+    """
+    result = LogCaptureFixture(request.node, _ispytest=True)
+    yield result
+    result._finalize()
+
+
+def get_log_level_for_setting(config: Config, *setting_names: str) -> int | None:
+    for setting_name in setting_names:
+        log_level = config.getoption(setting_name)
+        if log_level is None:
+            log_level = config.getini(setting_name)
+        if log_level:
+            break
+    else:
+        return None
+
+    if isinstance(log_level, str):
+        log_level = log_level.upper()
+    try:
+        return int(getattr(logging, log_level, log_level))
+    except ValueError as e:
+        # Python logging does not recognise this as a logging level
+        raise UsageError(
+            f"'{log_level}' is not recognized as a logging level name for "
+            f"'{setting_name}'. Please consider passing the "
+            "logging level num instead."
+        ) from e
+
+
+# run after terminalreporter/capturemanager are configured
+@hookimpl(trylast=True)
+def pytest_configure(config: Config) -> None:
+    config.pluginmanager.register(LoggingPlugin(config), "logging-plugin")
+
+
+class LoggingPlugin:
+    """Attaches to the logging module and captures log messages for each test."""
+
+    def __init__(self, config: Config) -> None:
+        """Create a new plugin to capture log messages.
+
+        The formatter can be safely shared across all handlers so
+        create a single one for the entire test session here.
+        """
+        self._config = config
+
+        # Report logging.
+        self.formatter = self._create_formatter(
+            get_option_ini(config, "log_format"),
+            get_option_ini(config, "log_date_format"),
+            get_option_ini(config, "log_auto_indent"),
+        )
+        self.log_level = get_log_level_for_setting(config, "log_level")
+        self.caplog_handler = LogCaptureHandler()
+        self.caplog_handler.setFormatter(self.formatter)
+        self.report_handler = LogCaptureHandler()
+        self.report_handler.setFormatter(self.formatter)
+
+        # File logging.
+        self.log_file_level = get_log_level_for_setting(
+            config, "log_file_level", "log_level"
+        )
+        log_file = get_option_ini(config, "log_file") or os.devnull
+        if log_file != os.devnull:
+            directory = os.path.dirname(os.path.abspath(log_file))
+            if not os.path.isdir(directory):
+                os.makedirs(directory)
+
+        self.log_file_mode = get_option_ini(config, "log_file_mode") or "w"
+        self.log_file_handler = _FileHandler(
+            log_file, mode=self.log_file_mode, encoding="UTF-8"
+        )
+        log_file_format = get_option_ini(config, "log_file_format", "log_format")
+        log_file_date_format = get_option_ini(
+            config, "log_file_date_format", "log_date_format"
+        )
+
+        log_file_formatter = DatetimeFormatter(
+            log_file_format, datefmt=log_file_date_format
+        )
+        self.log_file_handler.setFormatter(log_file_formatter)
+
+        # CLI/live logging.
+        self.log_cli_level = get_log_level_for_setting(
+            config, "log_cli_level", "log_level"
+        )
+        if self._log_cli_enabled():
+            terminal_reporter = config.pluginmanager.get_plugin("terminalreporter")
+            # Guaranteed by `_log_cli_enabled()`.
+            assert terminal_reporter is not None
+            capture_manager = config.pluginmanager.get_plugin("capturemanager")
+            # if capturemanager plugin is disabled, live logging still works.
+            self.log_cli_handler: (
+                _LiveLoggingStreamHandler | _LiveLoggingNullHandler
+            ) = _LiveLoggingStreamHandler(terminal_reporter, capture_manager)
+        else:
+            self.log_cli_handler = _LiveLoggingNullHandler()
+        log_cli_formatter = self._create_formatter(
+            get_option_ini(config, "log_cli_format", "log_format"),
+            get_option_ini(config, "log_cli_date_format", "log_date_format"),
+            get_option_ini(config, "log_auto_indent"),
+        )
+        self.log_cli_handler.setFormatter(log_cli_formatter)
+        self._disable_loggers(loggers_to_disable=config.option.logger_disable)
+
+    def _disable_loggers(self, loggers_to_disable: list[str]) -> None:
+        if not loggers_to_disable:
+            return
+
+        for name in loggers_to_disable:
+            logger = logging.getLogger(name)
+            logger.disabled = True
+
+    def _create_formatter(self, log_format, log_date_format, auto_indent):
+        # Color option doesn't exist if terminal plugin is disabled.
+        color = getattr(self._config.option, "color", "no")
+        if color != "no" and ColoredLevelFormatter.LEVELNAME_FMT_REGEX.search(
+            log_format
+        ):
+            formatter: logging.Formatter = ColoredLevelFormatter(
+                create_terminal_writer(self._config), log_format, log_date_format
+            )
+        else:
+            formatter = DatetimeFormatter(log_format, log_date_format)
+
+        formatter._style = PercentStyleMultiline(
+            formatter._style._fmt, auto_indent=auto_indent
+        )
+
+        return formatter
+
+    def set_log_path(self, fname: str) -> None:
+        """Set the filename parameter for Logging.FileHandler().
+
+        Creates parent directory if it does not exist.
+
+        .. warning::
+            This is an experimental API.
+        """
+        fpath = Path(fname)
+
+        if not fpath.is_absolute():
+            fpath = self._config.rootpath / fpath
+
+        if not fpath.parent.exists():
+            fpath.parent.mkdir(exist_ok=True, parents=True)
+
+        # https://github.com/python/mypy/issues/11193
+        stream: io.TextIOWrapper = fpath.open(mode=self.log_file_mode, encoding="UTF-8")  # type: ignore[assignment]
+        old_stream = self.log_file_handler.setStream(stream)
+        if old_stream:
+            old_stream.close()
+
+    def _log_cli_enabled(self) -> bool:
+        """Return whether live logging is enabled."""
+        enabled = self._config.getoption(
+            "--log-cli-level"
+        ) is not None or self._config.getini("log_cli")
+        if not enabled:
+            return False
+
+        terminal_reporter = self._config.pluginmanager.get_plugin("terminalreporter")
+        if terminal_reporter is None:
+            # terminal reporter is disabled e.g. by pytest-xdist.
+            return False
+
+        return True
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_sessionstart(self) -> Generator[None]:
+        self.log_cli_handler.set_when("sessionstart")
+
+        with catching_logs(self.log_cli_handler, level=self.log_cli_level):
+            with catching_logs(self.log_file_handler, level=self.log_file_level):
+                return (yield)
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_collection(self) -> Generator[None]:
+        self.log_cli_handler.set_when("collection")
+
+        with catching_logs(self.log_cli_handler, level=self.log_cli_level):
+            with catching_logs(self.log_file_handler, level=self.log_file_level):
+                return (yield)
+
+    @hookimpl(wrapper=True)
+    def pytest_runtestloop(self, session: Session) -> Generator[None, object, object]:
+        if session.config.option.collectonly:
+            return (yield)
+
+        if self._log_cli_enabled() and self._config.get_verbosity() < 1:
+            # The verbose flag is needed to avoid messy test progress output.
+            self._config.option.verbose = 1
+
+        with catching_logs(self.log_cli_handler, level=self.log_cli_level):
+            with catching_logs(self.log_file_handler, level=self.log_file_level):
+                return (yield)  # Run all the tests.
+
+    @hookimpl
+    def pytest_runtest_logstart(self) -> None:
+        self.log_cli_handler.reset()
+        self.log_cli_handler.set_when("start")
+
+    @hookimpl
+    def pytest_runtest_logreport(self) -> None:
+        self.log_cli_handler.set_when("logreport")
+
+    def _runtest_for(self, item: nodes.Item, when: str) -> Generator[None]:
+        """Implement the internals of the pytest_runtest_xxx() hooks."""
+        with catching_logs(
+            self.caplog_handler,
+            level=self.log_level,
+        ) as caplog_handler, catching_logs(
+            self.report_handler,
+            level=self.log_level,
+        ) as report_handler:
+            caplog_handler.reset()
+            report_handler.reset()
+            item.stash[caplog_records_key][when] = caplog_handler.records
+            item.stash[caplog_handler_key] = caplog_handler
+
+            try:
+                yield
+            finally:
+                log = report_handler.stream.getvalue().strip()
+                item.add_report_section(when, "log", log)
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_setup(self, item: nodes.Item) -> Generator[None]:
+        self.log_cli_handler.set_when("setup")
+
+        empty: dict[str, list[logging.LogRecord]] = {}
+        item.stash[caplog_records_key] = empty
+        yield from self._runtest_for(item, "setup")
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_call(self, item: nodes.Item) -> Generator[None]:
+        self.log_cli_handler.set_when("call")
+
+        yield from self._runtest_for(item, "call")
+
+    @hookimpl(wrapper=True)
+    def pytest_runtest_teardown(self, item: nodes.Item) -> Generator[None]:
+        self.log_cli_handler.set_when("teardown")
+
+        try:
+            yield from self._runtest_for(item, "teardown")
+        finally:
+            del item.stash[caplog_records_key]
+            del item.stash[caplog_handler_key]
+
+    @hookimpl
+    def pytest_runtest_logfinish(self) -> None:
+        self.log_cli_handler.set_when("finish")
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_sessionfinish(self) -> Generator[None]:
+        self.log_cli_handler.set_when("sessionfinish")
+
+        with catching_logs(self.log_cli_handler, level=self.log_cli_level):
+            with catching_logs(self.log_file_handler, level=self.log_file_level):
+                return (yield)
+
+    @hookimpl
+    def pytest_unconfigure(self) -> None:
+        # Close the FileHandler explicitly.
+        # (logging.shutdown might have lost the weakref?!)
+        self.log_file_handler.close()
+
+
+class _FileHandler(logging.FileHandler):
+    """A logging FileHandler with pytest tweaks."""
+
+    def handleError(self, record: logging.LogRecord) -> None:
+        # Handled by LogCaptureHandler.
+        pass
+
+
+class _LiveLoggingStreamHandler(logging_StreamHandler):
+    """A logging StreamHandler used by the live logging feature: it will
+    write a newline before the first log message in each test.
+
+    During live logging we must also explicitly disable stdout/stderr
+    capturing otherwise it will get captured and won't appear in the
+    terminal.
+    """
+
+    # Officially stream needs to be a IO[str], but TerminalReporter
+    # isn't. So force it.
+    stream: TerminalReporter = None  # type: ignore
+
+    def __init__(
+        self,
+        terminal_reporter: TerminalReporter,
+        capture_manager: CaptureManager | None,
+    ) -> None:
+        super().__init__(stream=terminal_reporter)  # type: ignore[arg-type]
+        self.capture_manager = capture_manager
+        self.reset()
+        self.set_when(None)
+        self._test_outcome_written = False
+
+    def reset(self) -> None:
+        """Reset the handler; should be called before the start of each test."""
+        self._first_record_emitted = False
+
+    def set_when(self, when: str | None) -> None:
+        """Prepare for the given test phase (setup/call/teardown)."""
+        self._when = when
+        self._section_name_shown = False
+        if when == "start":
+            self._test_outcome_written = False
+
+    def emit(self, record: logging.LogRecord) -> None:
+        ctx_manager = (
+            self.capture_manager.global_and_fixture_disabled()
+            if self.capture_manager
+            else nullcontext()
+        )
+        with ctx_manager:
+            if not self._first_record_emitted:
+                self.stream.write("\n")
+                self._first_record_emitted = True
+            elif self._when in ("teardown", "finish"):
+                if not self._test_outcome_written:
+                    self._test_outcome_written = True
+                    self.stream.write("\n")
+            if not self._section_name_shown and self._when:
+                self.stream.section("live log " + self._when, sep="-", bold=True)
+                self._section_name_shown = True
+            super().emit(record)
+
+    def handleError(self, record: logging.LogRecord) -> None:
+        # Handled by LogCaptureHandler.
+        pass
+
+
+class _LiveLoggingNullHandler(logging.NullHandler):
+    """A logging handler used when live logging is disabled."""
+
+    def reset(self) -> None:
+        pass
+
+    def set_when(self, when: str) -> None:
+        pass
+
+    def handleError(self, record: logging.LogRecord) -> None:
+        # Handled by LogCaptureHandler.
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/main.py b/backend/venv/lib/python3.12/site-packages/_pytest/main.py
new file mode 100644
index 00000000..e5534e98
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/main.py
@@ -0,0 +1,1072 @@
+"""Core implementation of the testing process: init, session, runtest loop."""
+
+from __future__ import annotations
+
+import argparse
+import dataclasses
+import fnmatch
+import functools
+import importlib
+import importlib.util
+import os
+from pathlib import Path
+import sys
+from typing import AbstractSet
+from typing import Callable
+from typing import Dict
+from typing import final
+from typing import Iterable
+from typing import Iterator
+from typing import Literal
+from typing import overload
+from typing import Sequence
+from typing import TYPE_CHECKING
+import warnings
+
+import pluggy
+
+from _pytest import nodes
+import _pytest._code
+from _pytest.config import Config
+from _pytest.config import directory_arg
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config import PytestPluginManager
+from _pytest.config import UsageError
+from _pytest.config.argparsing import Parser
+from _pytest.config.compat import PathAwareHookProxy
+from _pytest.outcomes import exit
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import bestrelpath
+from _pytest.pathlib import fnmatch_ex
+from _pytest.pathlib import safe_exists
+from _pytest.pathlib import scandir
+from _pytest.reports import CollectReport
+from _pytest.reports import TestReport
+from _pytest.runner import collect_one_node
+from _pytest.runner import SetupState
+from _pytest.warning_types import PytestWarning
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from _pytest.fixtures import FixtureManager
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini(
+        "norecursedirs",
+        "Directory patterns to avoid for recursion",
+        type="args",
+        default=[
+            "*.egg",
+            ".*",
+            "_darcs",
+            "build",
+            "CVS",
+            "dist",
+            "node_modules",
+            "venv",
+            "{arch}",
+        ],
+    )
+    parser.addini(
+        "testpaths",
+        "Directories to search for tests when no files or directories are given on the "
+        "command line",
+        type="args",
+        default=[],
+    )
+    group = parser.getgroup("general", "Running and selection options")
+    group._addoption(
+        "-x",
+        "--exitfirst",
+        action="store_const",
+        dest="maxfail",
+        const=1,
+        help="Exit instantly on first error or failed test",
+    )
+    group = parser.getgroup("pytest-warnings")
+    group.addoption(
+        "-W",
+        "--pythonwarnings",
+        action="append",
+        help="Set which warnings to report, see -W option of Python itself",
+    )
+    parser.addini(
+        "filterwarnings",
+        type="linelist",
+        help="Each line specifies a pattern for "
+        "warnings.filterwarnings. "
+        "Processed after -W/--pythonwarnings.",
+    )
+    group._addoption(
+        "--maxfail",
+        metavar="num",
+        action="store",
+        type=int,
+        dest="maxfail",
+        default=0,
+        help="Exit after first num failures or errors",
+    )
+    group._addoption(
+        "--strict-config",
+        action="store_true",
+        help="Any warnings encountered while parsing the `pytest` section of the "
+        "configuration file raise errors",
+    )
+    group._addoption(
+        "--strict-markers",
+        action="store_true",
+        help="Markers not registered in the `markers` section of the configuration "
+        "file raise errors",
+    )
+    group._addoption(
+        "--strict",
+        action="store_true",
+        help="(Deprecated) alias to --strict-markers",
+    )
+    group._addoption(
+        "-c",
+        "--config-file",
+        metavar="FILE",
+        type=str,
+        dest="inifilename",
+        help="Load configuration from `FILE` instead of trying to locate one of the "
+        "implicit configuration files.",
+    )
+    group._addoption(
+        "--continue-on-collection-errors",
+        action="store_true",
+        default=False,
+        dest="continue_on_collection_errors",
+        help="Force test execution even if collection errors occur",
+    )
+    group._addoption(
+        "--rootdir",
+        action="store",
+        dest="rootdir",
+        help="Define root directory for tests. Can be relative path: 'root_dir', './root_dir', "
+        "'root_dir/another_dir/'; absolute path: '/home/user/root_dir'; path with variables: "
+        "'$HOME/root_dir'.",
+    )
+
+    group = parser.getgroup("collect", "collection")
+    group.addoption(
+        "--collectonly",
+        "--collect-only",
+        "--co",
+        action="store_true",
+        help="Only collect tests, don't execute them",
+    )
+    group.addoption(
+        "--pyargs",
+        action="store_true",
+        help="Try to interpret all arguments as Python packages",
+    )
+    group.addoption(
+        "--ignore",
+        action="append",
+        metavar="path",
+        help="Ignore path during collection (multi-allowed)",
+    )
+    group.addoption(
+        "--ignore-glob",
+        action="append",
+        metavar="path",
+        help="Ignore path pattern during collection (multi-allowed)",
+    )
+    group.addoption(
+        "--deselect",
+        action="append",
+        metavar="nodeid_prefix",
+        help="Deselect item (via node id prefix) during collection (multi-allowed)",
+    )
+    group.addoption(
+        "--confcutdir",
+        dest="confcutdir",
+        default=None,
+        metavar="dir",
+        type=functools.partial(directory_arg, optname="--confcutdir"),
+        help="Only load conftest.py's relative to specified dir",
+    )
+    group.addoption(
+        "--noconftest",
+        action="store_true",
+        dest="noconftest",
+        default=False,
+        help="Don't load any conftest.py files",
+    )
+    group.addoption(
+        "--keepduplicates",
+        "--keep-duplicates",
+        action="store_true",
+        dest="keepduplicates",
+        default=False,
+        help="Keep duplicate tests",
+    )
+    group.addoption(
+        "--collect-in-virtualenv",
+        action="store_true",
+        dest="collect_in_virtualenv",
+        default=False,
+        help="Don't ignore tests in a local virtualenv directory",
+    )
+    group.addoption(
+        "--import-mode",
+        default="prepend",
+        choices=["prepend", "append", "importlib"],
+        dest="importmode",
+        help="Prepend/append to sys.path when importing test modules and conftest "
+        "files. Default: prepend.",
+    )
+    parser.addini(
+        "consider_namespace_packages",
+        type="bool",
+        default=False,
+        help="Consider namespace packages when resolving module names during import",
+    )
+
+    group = parser.getgroup("debugconfig", "test session debugging and configuration")
+    group.addoption(
+        "--basetemp",
+        dest="basetemp",
+        default=None,
+        type=validate_basetemp,
+        metavar="dir",
+        help=(
+            "Base temporary directory for this test run. "
+            "(Warning: this directory is removed if it exists.)"
+        ),
+    )
+
+
+def validate_basetemp(path: str) -> str:
+    # GH 7119
+    msg = "basetemp must not be empty, the current working directory or any parent directory of it"
+
+    # empty path
+    if not path:
+        raise argparse.ArgumentTypeError(msg)
+
+    def is_ancestor(base: Path, query: Path) -> bool:
+        """Return whether query is an ancestor of base."""
+        if base == query:
+            return True
+        return query in base.parents
+
+    # check if path is an ancestor of cwd
+    if is_ancestor(Path.cwd(), Path(path).absolute()):
+        raise argparse.ArgumentTypeError(msg)
+
+    # check symlinks for ancestors
+    if is_ancestor(Path.cwd().resolve(), Path(path).resolve()):
+        raise argparse.ArgumentTypeError(msg)
+
+    return path
+
+
+def wrap_session(
+    config: Config, doit: Callable[[Config, Session], int | ExitCode | None]
+) -> int | ExitCode:
+    """Skeleton command line program."""
+    session = Session.from_config(config)
+    session.exitstatus = ExitCode.OK
+    initstate = 0
+    try:
+        try:
+            config._do_configure()
+            initstate = 1
+            config.hook.pytest_sessionstart(session=session)
+            initstate = 2
+            session.exitstatus = doit(config, session) or 0
+        except UsageError:
+            session.exitstatus = ExitCode.USAGE_ERROR
+            raise
+        except Failed:
+            session.exitstatus = ExitCode.TESTS_FAILED
+        except (KeyboardInterrupt, exit.Exception):
+            excinfo = _pytest._code.ExceptionInfo.from_current()
+            exitstatus: int | ExitCode = ExitCode.INTERRUPTED
+            if isinstance(excinfo.value, exit.Exception):
+                if excinfo.value.returncode is not None:
+                    exitstatus = excinfo.value.returncode
+                if initstate < 2:
+                    sys.stderr.write(f"{excinfo.typename}: {excinfo.value.msg}\n")
+            config.hook.pytest_keyboard_interrupt(excinfo=excinfo)
+            session.exitstatus = exitstatus
+        except BaseException:
+            session.exitstatus = ExitCode.INTERNAL_ERROR
+            excinfo = _pytest._code.ExceptionInfo.from_current()
+            try:
+                config.notify_exception(excinfo, config.option)
+            except exit.Exception as exc:
+                if exc.returncode is not None:
+                    session.exitstatus = exc.returncode
+                sys.stderr.write(f"{type(exc).__name__}: {exc}\n")
+            else:
+                if isinstance(excinfo.value, SystemExit):
+                    sys.stderr.write("mainloop: caught unexpected SystemExit!\n")
+
+    finally:
+        # Explicitly break reference cycle.
+        excinfo = None  # type: ignore
+        os.chdir(session.startpath)
+        if initstate >= 2:
+            try:
+                config.hook.pytest_sessionfinish(
+                    session=session, exitstatus=session.exitstatus
+                )
+            except exit.Exception as exc:
+                if exc.returncode is not None:
+                    session.exitstatus = exc.returncode
+                sys.stderr.write(f"{type(exc).__name__}: {exc}\n")
+        config._ensure_unconfigure()
+    return session.exitstatus
+
+
+def pytest_cmdline_main(config: Config) -> int | ExitCode:
+    return wrap_session(config, _main)
+
+
+def _main(config: Config, session: Session) -> int | ExitCode | None:
+    """Default command line protocol for initialization, session,
+    running tests and reporting."""
+    config.hook.pytest_collection(session=session)
+    config.hook.pytest_runtestloop(session=session)
+
+    if session.testsfailed:
+        return ExitCode.TESTS_FAILED
+    elif session.testscollected == 0:
+        return ExitCode.NO_TESTS_COLLECTED
+    return None
+
+
+def pytest_collection(session: Session) -> None:
+    session.perform_collect()
+
+
+def pytest_runtestloop(session: Session) -> bool:
+    if session.testsfailed and not session.config.option.continue_on_collection_errors:
+        raise session.Interrupted(
+            "%d error%s during collection"
+            % (session.testsfailed, "s" if session.testsfailed != 1 else "")
+        )
+
+    if session.config.option.collectonly:
+        return True
+
+    for i, item in enumerate(session.items):
+        nextitem = session.items[i + 1] if i + 1 < len(session.items) else None
+        item.config.hook.pytest_runtest_protocol(item=item, nextitem=nextitem)
+        if session.shouldfail:
+            raise session.Failed(session.shouldfail)
+        if session.shouldstop:
+            raise session.Interrupted(session.shouldstop)
+    return True
+
+
+def _in_venv(path: Path) -> bool:
+    """Attempt to detect if ``path`` is the root of a Virtual Environment by
+    checking for the existence of the pyvenv.cfg file.
+
+    [https://peps.python.org/pep-0405/]
+
+    For regression protection we also check for conda environments that do not include pyenv.cfg yet --
+    https://github.com/conda/conda/issues/13337 is the conda issue tracking adding pyenv.cfg.
+
+    Checking for the `conda-meta/history` file per https://github.com/pytest-dev/pytest/issues/12652#issuecomment-2246336902.
+
+    """
+    try:
+        return (
+            path.joinpath("pyvenv.cfg").is_file()
+            or path.joinpath("conda-meta", "history").is_file()
+        )
+    except OSError:
+        return False
+
+
+def pytest_ignore_collect(collection_path: Path, config: Config) -> bool | None:
+    if collection_path.name == "__pycache__":
+        return True
+
+    ignore_paths = config._getconftest_pathlist(
+        "collect_ignore", path=collection_path.parent
+    )
+    ignore_paths = ignore_paths or []
+    excludeopt = config.getoption("ignore")
+    if excludeopt:
+        ignore_paths.extend(absolutepath(x) for x in excludeopt)
+
+    if collection_path in ignore_paths:
+        return True
+
+    ignore_globs = config._getconftest_pathlist(
+        "collect_ignore_glob", path=collection_path.parent
+    )
+    ignore_globs = ignore_globs or []
+    excludeglobopt = config.getoption("ignore_glob")
+    if excludeglobopt:
+        ignore_globs.extend(absolutepath(x) for x in excludeglobopt)
+
+    if any(fnmatch.fnmatch(str(collection_path), str(glob)) for glob in ignore_globs):
+        return True
+
+    allow_in_venv = config.getoption("collect_in_virtualenv")
+    if not allow_in_venv and _in_venv(collection_path):
+        return True
+
+    if collection_path.is_dir():
+        norecursepatterns = config.getini("norecursedirs")
+        if any(fnmatch_ex(pat, collection_path) for pat in norecursepatterns):
+            return True
+
+    return None
+
+
+def pytest_collect_directory(
+    path: Path, parent: nodes.Collector
+) -> nodes.Collector | None:
+    return Dir.from_parent(parent, path=path)
+
+
+def pytest_collection_modifyitems(items: list[nodes.Item], config: Config) -> None:
+    deselect_prefixes = tuple(config.getoption("deselect") or [])
+    if not deselect_prefixes:
+        return
+
+    remaining = []
+    deselected = []
+    for colitem in items:
+        if colitem.nodeid.startswith(deselect_prefixes):
+            deselected.append(colitem)
+        else:
+            remaining.append(colitem)
+
+    if deselected:
+        config.hook.pytest_deselected(items=deselected)
+        items[:] = remaining
+
+
+class FSHookProxy:
+    def __init__(
+        self,
+        pm: PytestPluginManager,
+        remove_mods: AbstractSet[object],
+    ) -> None:
+        self.pm = pm
+        self.remove_mods = remove_mods
+
+    def __getattr__(self, name: str) -> pluggy.HookCaller:
+        x = self.pm.subset_hook_caller(name, remove_plugins=self.remove_mods)
+        self.__dict__[name] = x
+        return x
+
+
+class Interrupted(KeyboardInterrupt):
+    """Signals that the test run was interrupted."""
+
+    __module__ = "builtins"  # For py3.
+
+
+class Failed(Exception):
+    """Signals a stop as failed test run."""
+
+
+@dataclasses.dataclass
+class _bestrelpath_cache(Dict[Path, str]):
+    __slots__ = ("path",)
+
+    path: Path
+
+    def __missing__(self, path: Path) -> str:
+        r = bestrelpath(self.path, path)
+        self[path] = r
+        return r
+
+
+@final
+class Dir(nodes.Directory):
+    """Collector of files in a file system directory.
+
+    .. versionadded:: 8.0
+
+    .. note::
+
+        Python directories with an `__init__.py` file are instead collected by
+        :class:`~pytest.Package` by default. Both are :class:`~pytest.Directory`
+        collectors.
+    """
+
+    @classmethod
+    def from_parent(  # type: ignore[override]
+        cls,
+        parent: nodes.Collector,
+        *,
+        path: Path,
+    ) -> Self:
+        """The public constructor.
+
+        :param parent: The parent collector of this Dir.
+        :param path: The directory's path.
+        :type path: pathlib.Path
+        """
+        return super().from_parent(parent=parent, path=path)
+
+    def collect(self) -> Iterable[nodes.Item | nodes.Collector]:
+        config = self.config
+        col: nodes.Collector | None
+        cols: Sequence[nodes.Collector]
+        ihook = self.ihook
+        for direntry in scandir(self.path):
+            if direntry.is_dir():
+                path = Path(direntry.path)
+                if not self.session.isinitpath(path, with_parents=True):
+                    if ihook.pytest_ignore_collect(collection_path=path, config=config):
+                        continue
+                col = ihook.pytest_collect_directory(path=path, parent=self)
+                if col is not None:
+                    yield col
+
+            elif direntry.is_file():
+                path = Path(direntry.path)
+                if not self.session.isinitpath(path):
+                    if ihook.pytest_ignore_collect(collection_path=path, config=config):
+                        continue
+                cols = ihook.pytest_collect_file(file_path=path, parent=self)
+                yield from cols
+
+
+@final
+class Session(nodes.Collector):
+    """The root of the collection tree.
+
+    ``Session`` collects the initial paths given as arguments to pytest.
+    """
+
+    Interrupted = Interrupted
+    Failed = Failed
+    # Set on the session by runner.pytest_sessionstart.
+    _setupstate: SetupState
+    # Set on the session by fixtures.pytest_sessionstart.
+    _fixturemanager: FixtureManager
+    exitstatus: int | ExitCode
+
+    def __init__(self, config: Config) -> None:
+        super().__init__(
+            name="",
+            path=config.rootpath,
+            fspath=None,
+            parent=None,
+            config=config,
+            session=self,
+            nodeid="",
+        )
+        self.testsfailed = 0
+        self.testscollected = 0
+        self._shouldstop: bool | str = False
+        self._shouldfail: bool | str = False
+        self.trace = config.trace.root.get("collection")
+        self._initialpaths: frozenset[Path] = frozenset()
+        self._initialpaths_with_parents: frozenset[Path] = frozenset()
+        self._notfound: list[tuple[str, Sequence[nodes.Collector]]] = []
+        self._initial_parts: list[CollectionArgument] = []
+        self._collection_cache: dict[nodes.Collector, CollectReport] = {}
+        self.items: list[nodes.Item] = []
+
+        self._bestrelpathcache: dict[Path, str] = _bestrelpath_cache(config.rootpath)
+
+        self.config.pluginmanager.register(self, name="session")
+
+    @classmethod
+    def from_config(cls, config: Config) -> Session:
+        session: Session = cls._create(config=config)
+        return session
+
+    def __repr__(self) -> str:
+        return "<%s %s exitstatus=%r testsfailed=%d testscollected=%d>" % (
+            self.__class__.__name__,
+            self.name,
+            getattr(self, "exitstatus", "<UNSET>"),
+            self.testsfailed,
+            self.testscollected,
+        )
+
+    @property
+    def shouldstop(self) -> bool | str:
+        return self._shouldstop
+
+    @shouldstop.setter
+    def shouldstop(self, value: bool | str) -> None:
+        # The runner checks shouldfail and assumes that if it is set we are
+        # definitely stopping, so prevent unsetting it.
+        if value is False and self._shouldstop:
+            warnings.warn(
+                PytestWarning(
+                    "session.shouldstop cannot be unset after it has been set; ignoring."
+                ),
+                stacklevel=2,
+            )
+            return
+        self._shouldstop = value
+
+    @property
+    def shouldfail(self) -> bool | str:
+        return self._shouldfail
+
+    @shouldfail.setter
+    def shouldfail(self, value: bool | str) -> None:
+        # The runner checks shouldfail and assumes that if it is set we are
+        # definitely stopping, so prevent unsetting it.
+        if value is False and self._shouldfail:
+            warnings.warn(
+                PytestWarning(
+                    "session.shouldfail cannot be unset after it has been set; ignoring."
+                ),
+                stacklevel=2,
+            )
+            return
+        self._shouldfail = value
+
+    @property
+    def startpath(self) -> Path:
+        """The path from which pytest was invoked.
+
+        .. versionadded:: 7.0.0
+        """
+        return self.config.invocation_params.dir
+
+    def _node_location_to_relpath(self, node_path: Path) -> str:
+        # bestrelpath is a quite slow function.
+        return self._bestrelpathcache[node_path]
+
+    @hookimpl(tryfirst=True)
+    def pytest_collectstart(self) -> None:
+        if self.shouldfail:
+            raise self.Failed(self.shouldfail)
+        if self.shouldstop:
+            raise self.Interrupted(self.shouldstop)
+
+    @hookimpl(tryfirst=True)
+    def pytest_runtest_logreport(self, report: TestReport | CollectReport) -> None:
+        if report.failed and not hasattr(report, "wasxfail"):
+            self.testsfailed += 1
+            maxfail = self.config.getvalue("maxfail")
+            if maxfail and self.testsfailed >= maxfail:
+                self.shouldfail = "stopping after %d failures" % (self.testsfailed)
+
+    pytest_collectreport = pytest_runtest_logreport
+
+    def isinitpath(
+        self,
+        path: str | os.PathLike[str],
+        *,
+        with_parents: bool = False,
+    ) -> bool:
+        """Is path an initial path?
+
+        An initial path is a path explicitly given to pytest on the command
+        line.
+
+        :param with_parents:
+            If set, also return True if the path is a parent of an initial path.
+
+        .. versionchanged:: 8.0
+            Added the ``with_parents`` parameter.
+        """
+        # Optimization: Path(Path(...)) is much slower than isinstance.
+        path_ = path if isinstance(path, Path) else Path(path)
+        if with_parents:
+            return path_ in self._initialpaths_with_parents
+        else:
+            return path_ in self._initialpaths
+
+    def gethookproxy(self, fspath: os.PathLike[str]) -> pluggy.HookRelay:
+        # Optimization: Path(Path(...)) is much slower than isinstance.
+        path = fspath if isinstance(fspath, Path) else Path(fspath)
+        pm = self.config.pluginmanager
+        # Check if we have the common case of running
+        # hooks with all conftest.py files.
+        my_conftestmodules = pm._getconftestmodules(path)
+        remove_mods = pm._conftest_plugins.difference(my_conftestmodules)
+        proxy: pluggy.HookRelay
+        if remove_mods:
+            # One or more conftests are not in use at this path.
+            proxy = PathAwareHookProxy(FSHookProxy(pm, remove_mods))  # type: ignore[arg-type,assignment]
+        else:
+            # All plugins are active for this fspath.
+            proxy = self.config.hook
+        return proxy
+
+    def _collect_path(
+        self,
+        path: Path,
+        path_cache: dict[Path, Sequence[nodes.Collector]],
+    ) -> Sequence[nodes.Collector]:
+        """Create a Collector for the given path.
+
+        `path_cache` makes it so the same Collectors are returned for the same
+        path.
+        """
+        if path in path_cache:
+            return path_cache[path]
+
+        if path.is_dir():
+            ihook = self.gethookproxy(path.parent)
+            col: nodes.Collector | None = ihook.pytest_collect_directory(
+                path=path, parent=self
+            )
+            cols: Sequence[nodes.Collector] = (col,) if col is not None else ()
+
+        elif path.is_file():
+            ihook = self.gethookproxy(path)
+            cols = ihook.pytest_collect_file(file_path=path, parent=self)
+
+        else:
+            # Broken symlink or invalid/missing file.
+            cols = ()
+
+        path_cache[path] = cols
+        return cols
+
+    @overload
+    def perform_collect(
+        self, args: Sequence[str] | None = ..., genitems: Literal[True] = ...
+    ) -> Sequence[nodes.Item]: ...
+
+    @overload
+    def perform_collect(
+        self, args: Sequence[str] | None = ..., genitems: bool = ...
+    ) -> Sequence[nodes.Item | nodes.Collector]: ...
+
+    def perform_collect(
+        self, args: Sequence[str] | None = None, genitems: bool = True
+    ) -> Sequence[nodes.Item | nodes.Collector]:
+        """Perform the collection phase for this session.
+
+        This is called by the default :hook:`pytest_collection` hook
+        implementation; see the documentation of this hook for more details.
+        For testing purposes, it may also be called directly on a fresh
+        ``Session``.
+
+        This function normally recursively expands any collectors collected
+        from the session to their items, and only items are returned. For
+        testing purposes, this may be suppressed by passing ``genitems=False``,
+        in which case the return value contains these collectors unexpanded,
+        and ``session.items`` is empty.
+        """
+        if args is None:
+            args = self.config.args
+
+        self.trace("perform_collect", self, args)
+        self.trace.root.indent += 1
+
+        hook = self.config.hook
+
+        self._notfound = []
+        self._initial_parts = []
+        self._collection_cache = {}
+        self.items = []
+        items: Sequence[nodes.Item | nodes.Collector] = self.items
+        try:
+            initialpaths: list[Path] = []
+            initialpaths_with_parents: list[Path] = []
+            for arg in args:
+                collection_argument = resolve_collection_argument(
+                    self.config.invocation_params.dir,
+                    arg,
+                    as_pypath=self.config.option.pyargs,
+                )
+                self._initial_parts.append(collection_argument)
+                initialpaths.append(collection_argument.path)
+                initialpaths_with_parents.append(collection_argument.path)
+                initialpaths_with_parents.extend(collection_argument.path.parents)
+            self._initialpaths = frozenset(initialpaths)
+            self._initialpaths_with_parents = frozenset(initialpaths_with_parents)
+
+            rep = collect_one_node(self)
+            self.ihook.pytest_collectreport(report=rep)
+            self.trace.root.indent -= 1
+            if self._notfound:
+                errors = []
+                for arg, collectors in self._notfound:
+                    if collectors:
+                        errors.append(
+                            f"not found: {arg}\n(no match in any of {collectors!r})"
+                        )
+                    else:
+                        errors.append(f"found no collectors for {arg}")
+
+                raise UsageError(*errors)
+
+            if not genitems:
+                items = rep.result
+            else:
+                if rep.passed:
+                    for node in rep.result:
+                        self.items.extend(self.genitems(node))
+
+            self.config.pluginmanager.check_pending()
+            hook.pytest_collection_modifyitems(
+                session=self, config=self.config, items=items
+            )
+        finally:
+            self._notfound = []
+            self._initial_parts = []
+            self._collection_cache = {}
+            hook.pytest_collection_finish(session=self)
+
+        if genitems:
+            self.testscollected = len(items)
+
+        return items
+
+    def _collect_one_node(
+        self,
+        node: nodes.Collector,
+        handle_dupes: bool = True,
+    ) -> tuple[CollectReport, bool]:
+        if node in self._collection_cache and handle_dupes:
+            rep = self._collection_cache[node]
+            return rep, True
+        else:
+            rep = collect_one_node(node)
+            self._collection_cache[node] = rep
+            return rep, False
+
+    def collect(self) -> Iterator[nodes.Item | nodes.Collector]:
+        # This is a cache for the root directories of the initial paths.
+        # We can't use collection_cache for Session because of its special
+        # role as the bootstrapping collector.
+        path_cache: dict[Path, Sequence[nodes.Collector]] = {}
+
+        pm = self.config.pluginmanager
+
+        for collection_argument in self._initial_parts:
+            self.trace("processing argument", collection_argument)
+            self.trace.root.indent += 1
+
+            argpath = collection_argument.path
+            names = collection_argument.parts
+            module_name = collection_argument.module_name
+
+            # resolve_collection_argument() ensures this.
+            if argpath.is_dir():
+                assert not names, f"invalid arg {(argpath, names)!r}"
+
+            paths = [argpath]
+            # Add relevant parents of the path, from the root, e.g.
+            #   /a/b/c.py -> [/, /a, /a/b, /a/b/c.py]
+            if module_name is None:
+                # Paths outside of the confcutdir should not be considered.
+                for path in argpath.parents:
+                    if not pm._is_in_confcutdir(path):
+                        break
+                    paths.insert(0, path)
+            else:
+                # For --pyargs arguments, only consider paths matching the module
+                # name. Paths beyond the package hierarchy are not included.
+                module_name_parts = module_name.split(".")
+                for i, path in enumerate(argpath.parents, 2):
+                    if i > len(module_name_parts) or path.stem != module_name_parts[-i]:
+                        break
+                    paths.insert(0, path)
+
+            # Start going over the parts from the root, collecting each level
+            # and discarding all nodes which don't match the level's part.
+            any_matched_in_initial_part = False
+            notfound_collectors = []
+            work: list[tuple[nodes.Collector | nodes.Item, list[Path | str]]] = [
+                (self, [*paths, *names])
+            ]
+            while work:
+                matchnode, matchparts = work.pop()
+
+                # Pop'd all of the parts, this is a match.
+                if not matchparts:
+                    yield matchnode
+                    any_matched_in_initial_part = True
+                    continue
+
+                # Should have been matched by now, discard.
+                if not isinstance(matchnode, nodes.Collector):
+                    continue
+
+                # Collect this level of matching.
+                # Collecting Session (self) is done directly to avoid endless
+                # recursion to this function.
+                subnodes: Sequence[nodes.Collector | nodes.Item]
+                if isinstance(matchnode, Session):
+                    assert isinstance(matchparts[0], Path)
+                    subnodes = matchnode._collect_path(matchparts[0], path_cache)
+                else:
+                    # For backward compat, files given directly multiple
+                    # times on the command line should not be deduplicated.
+                    handle_dupes = not (
+                        len(matchparts) == 1
+                        and isinstance(matchparts[0], Path)
+                        and matchparts[0].is_file()
+                    )
+                    rep, duplicate = self._collect_one_node(matchnode, handle_dupes)
+                    if not duplicate and not rep.passed:
+                        # Report collection failures here to avoid failing to
+                        # run some test specified in the command line because
+                        # the module could not be imported (#134).
+                        matchnode.ihook.pytest_collectreport(report=rep)
+                    if not rep.passed:
+                        continue
+                    subnodes = rep.result
+
+                # Prune this level.
+                any_matched_in_collector = False
+                for node in reversed(subnodes):
+                    # Path part e.g. `/a/b/` in `/a/b/test_file.py::TestIt::test_it`.
+                    if isinstance(matchparts[0], Path):
+                        is_match = node.path == matchparts[0]
+                        if sys.platform == "win32" and not is_match:
+                            # In case the file paths do not match, fallback to samefile() to
+                            # account for short-paths on Windows (#11895).
+                            same_file = os.path.samefile(node.path, matchparts[0])
+                            # We don't want to match links to the current node,
+                            # otherwise we would match the same file more than once (#12039).
+                            is_match = same_file and (
+                                os.path.islink(node.path)
+                                == os.path.islink(matchparts[0])
+                            )
+
+                    # Name part e.g. `TestIt` in `/a/b/test_file.py::TestIt::test_it`.
+                    else:
+                        # TODO: Remove parametrized workaround once collection structure contains
+                        # parametrization.
+                        is_match = (
+                            node.name == matchparts[0]
+                            or node.name.split("[")[0] == matchparts[0]
+                        )
+                    if is_match:
+                        work.append((node, matchparts[1:]))
+                        any_matched_in_collector = True
+
+                if not any_matched_in_collector:
+                    notfound_collectors.append(matchnode)
+
+            if not any_matched_in_initial_part:
+                report_arg = "::".join((str(argpath), *names))
+                self._notfound.append((report_arg, notfound_collectors))
+
+            self.trace.root.indent -= 1
+
+    def genitems(self, node: nodes.Item | nodes.Collector) -> Iterator[nodes.Item]:
+        self.trace("genitems", node)
+        if isinstance(node, nodes.Item):
+            node.ihook.pytest_itemcollected(item=node)
+            yield node
+        else:
+            assert isinstance(node, nodes.Collector)
+            keepduplicates = self.config.getoption("keepduplicates")
+            # For backward compat, dedup only applies to files.
+            handle_dupes = not (keepduplicates and isinstance(node, nodes.File))
+            rep, duplicate = self._collect_one_node(node, handle_dupes)
+            if duplicate and not keepduplicates:
+                return
+            if rep.passed:
+                for subnode in rep.result:
+                    yield from self.genitems(subnode)
+            if not duplicate:
+                node.ihook.pytest_collectreport(report=rep)
+
+
+def search_pypath(module_name: str) -> str | None:
+    """Search sys.path for the given a dotted module name, and return its file
+    system path if found."""
+    try:
+        spec = importlib.util.find_spec(module_name)
+    # AttributeError: looks like package module, but actually filename
+    # ImportError: module does not exist
+    # ValueError: not a module name
+    except (AttributeError, ImportError, ValueError):
+        return None
+    if spec is None or spec.origin is None or spec.origin == "namespace":
+        return None
+    elif spec.submodule_search_locations:
+        return os.path.dirname(spec.origin)
+    else:
+        return spec.origin
+
+
+@dataclasses.dataclass(frozen=True)
+class CollectionArgument:
+    """A resolved collection argument."""
+
+    path: Path
+    parts: Sequence[str]
+    module_name: str | None
+
+
+def resolve_collection_argument(
+    invocation_path: Path, arg: str, *, as_pypath: bool = False
+) -> CollectionArgument:
+    """Parse path arguments optionally containing selection parts and return (fspath, names).
+
+    Command-line arguments can point to files and/or directories, and optionally contain
+    parts for specific tests selection, for example:
+
+        "pkg/tests/test_foo.py::TestClass::test_foo"
+
+    This function ensures the path exists, and returns a resolved `CollectionArgument`:
+
+        CollectionArgument(
+            path=Path("/full/path/to/pkg/tests/test_foo.py"),
+            parts=["TestClass", "test_foo"],
+            module_name=None,
+        )
+
+    When as_pypath is True, expects that the command-line argument actually contains
+    module paths instead of file-system paths:
+
+        "pkg.tests.test_foo::TestClass::test_foo"
+
+    In which case we search sys.path for a matching module, and then return the *path* to the
+    found module, which may look like this:
+
+        CollectionArgument(
+            path=Path("/home/u/myvenv/lib/site-packages/pkg/tests/test_foo.py"),
+            parts=["TestClass", "test_foo"],
+            module_name="pkg.tests.test_foo",
+        )
+
+    If the path doesn't exist, raise UsageError.
+    If the path is a directory and selection parts are present, raise UsageError.
+    """
+    base, squacket, rest = str(arg).partition("[")
+    strpath, *parts = base.split("::")
+    if parts:
+        parts[-1] = f"{parts[-1]}{squacket}{rest}"
+    module_name = None
+    if as_pypath:
+        pyarg_strpath = search_pypath(strpath)
+        if pyarg_strpath is not None:
+            module_name = strpath
+            strpath = pyarg_strpath
+    fspath = invocation_path / strpath
+    fspath = absolutepath(fspath)
+    if not safe_exists(fspath):
+        msg = (
+            "module or package not found: {arg} (missing __init__.py?)"
+            if as_pypath
+            else "file or directory not found: {arg}"
+        )
+        raise UsageError(msg.format(arg=arg))
+    if parts and fspath.is_dir():
+        msg = (
+            "package argument cannot contain :: selection parts: {arg}"
+            if as_pypath
+            else "directory argument cannot contain :: selection parts: {arg}"
+        )
+        raise UsageError(msg.format(arg=arg))
+    return CollectionArgument(
+        path=fspath,
+        parts=parts,
+        module_name=module_name,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/__init__.py b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__init__.py
new file mode 100644
index 00000000..a4f942c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__init__.py
@@ -0,0 +1,292 @@
+"""Generic mechanism for marking and selecting python functions."""
+
+from __future__ import annotations
+
+import collections
+import dataclasses
+from typing import AbstractSet
+from typing import Collection
+from typing import Iterable
+from typing import Optional
+from typing import TYPE_CHECKING
+
+from .expression import Expression
+from .expression import ParseError
+from .structures import EMPTY_PARAMETERSET_OPTION
+from .structures import get_empty_parameterset_mark
+from .structures import Mark
+from .structures import MARK_GEN
+from .structures import MarkDecorator
+from .structures import MarkGenerator
+from .structures import ParameterSet
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config import UsageError
+from _pytest.config.argparsing import NOT_SET
+from _pytest.config.argparsing import Parser
+from _pytest.stash import StashKey
+
+
+if TYPE_CHECKING:
+    from _pytest.nodes import Item
+
+
+__all__ = [
+    "MARK_GEN",
+    "Mark",
+    "MarkDecorator",
+    "MarkGenerator",
+    "ParameterSet",
+    "get_empty_parameterset_mark",
+]
+
+
+old_mark_config_key = StashKey[Optional[Config]]()
+
+
+def param(
+    *values: object,
+    marks: MarkDecorator | Collection[MarkDecorator | Mark] = (),
+    id: str | None = None,
+) -> ParameterSet:
+    """Specify a parameter in `pytest.mark.parametrize`_ calls or
+    :ref:`parametrized fixtures <fixture-parametrize-marks>`.
+
+    .. code-block:: python
+
+        @pytest.mark.parametrize(
+            "test_input,expected",
+            [
+                ("3+5", 8),
+                pytest.param("6*9", 42, marks=pytest.mark.xfail),
+            ],
+        )
+        def test_eval(test_input, expected):
+            assert eval(test_input) == expected
+
+    :param values: Variable args of the values of the parameter set, in order.
+    :param marks: A single mark or a list of marks to be applied to this parameter set.
+    :param id: The id to attribute to this parameter set.
+    """
+    return ParameterSet.param(*values, marks=marks, id=id)
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group._addoption(
+        "-k",
+        action="store",
+        dest="keyword",
+        default="",
+        metavar="EXPRESSION",
+        help="Only run tests which match the given substring expression. "
+        "An expression is a Python evaluable expression "
+        "where all names are substring-matched against test names "
+        "and their parent classes. Example: -k 'test_method or test_"
+        "other' matches all test functions and classes whose name "
+        "contains 'test_method' or 'test_other', while -k 'not test_method' "
+        "matches those that don't contain 'test_method' in their names. "
+        "-k 'not test_method and not test_other' will eliminate the matches. "
+        "Additionally keywords are matched to classes and functions "
+        "containing extra names in their 'extra_keyword_matches' set, "
+        "as well as functions which have names assigned directly to them. "
+        "The matching is case-insensitive.",
+    )
+
+    group._addoption(
+        "-m",
+        action="store",
+        dest="markexpr",
+        default="",
+        metavar="MARKEXPR",
+        help="Only run tests matching given mark expression. "
+        "For example: -m 'mark1 and not mark2'.",
+    )
+
+    group.addoption(
+        "--markers",
+        action="store_true",
+        help="show markers (builtin, plugin and per-project ones).",
+    )
+
+    parser.addini("markers", "Register new markers for test functions", "linelist")
+    parser.addini(EMPTY_PARAMETERSET_OPTION, "Default marker for empty parametersets")
+
+
+@hookimpl(tryfirst=True)
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    import _pytest.config
+
+    if config.option.markers:
+        config._do_configure()
+        tw = _pytest.config.create_terminal_writer(config)
+        for line in config.getini("markers"):
+            parts = line.split(":", 1)
+            name = parts[0]
+            rest = parts[1] if len(parts) == 2 else ""
+            tw.write(f"@pytest.mark.{name}:", bold=True)
+            tw.line(rest)
+            tw.line()
+        config._ensure_unconfigure()
+        return 0
+
+    return None
+
+
+@dataclasses.dataclass
+class KeywordMatcher:
+    """A matcher for keywords.
+
+    Given a list of names, matches any substring of one of these names. The
+    string inclusion check is case-insensitive.
+
+    Will match on the name of colitem, including the names of its parents.
+    Only matches names of items which are either a :class:`Class` or a
+    :class:`Function`.
+
+    Additionally, matches on names in the 'extra_keyword_matches' set of
+    any item, as well as names directly assigned to test functions.
+    """
+
+    __slots__ = ("_names",)
+
+    _names: AbstractSet[str]
+
+    @classmethod
+    def from_item(cls, item: Item) -> KeywordMatcher:
+        mapped_names = set()
+
+        # Add the names of the current item and any parent items,
+        # except the Session and root Directory's which are not
+        # interesting for matching.
+        import pytest
+
+        for node in item.listchain():
+            if isinstance(node, pytest.Session):
+                continue
+            if isinstance(node, pytest.Directory) and isinstance(
+                node.parent, pytest.Session
+            ):
+                continue
+            mapped_names.add(node.name)
+
+        # Add the names added as extra keywords to current or parent items.
+        mapped_names.update(item.listextrakeywords())
+
+        # Add the names attached to the current function through direct assignment.
+        function_obj = getattr(item, "function", None)
+        if function_obj:
+            mapped_names.update(function_obj.__dict__)
+
+        # Add the markers to the keywords as we no longer handle them correctly.
+        mapped_names.update(mark.name for mark in item.iter_markers())
+
+        return cls(mapped_names)
+
+    def __call__(self, subname: str, /, **kwargs: str | int | bool | None) -> bool:
+        if kwargs:
+            raise UsageError("Keyword expressions do not support call parameters.")
+        subname = subname.lower()
+        names = (name.lower() for name in self._names)
+
+        for name in names:
+            if subname in name:
+                return True
+        return False
+
+
+def deselect_by_keyword(items: list[Item], config: Config) -> None:
+    keywordexpr = config.option.keyword.lstrip()
+    if not keywordexpr:
+        return
+
+    expr = _parse_expression(keywordexpr, "Wrong expression passed to '-k'")
+
+    remaining = []
+    deselected = []
+    for colitem in items:
+        if not expr.evaluate(KeywordMatcher.from_item(colitem)):
+            deselected.append(colitem)
+        else:
+            remaining.append(colitem)
+
+    if deselected:
+        config.hook.pytest_deselected(items=deselected)
+        items[:] = remaining
+
+
+@dataclasses.dataclass
+class MarkMatcher:
+    """A matcher for markers which are present.
+
+    Tries to match on any marker names, attached to the given colitem.
+    """
+
+    __slots__ = ("own_mark_name_mapping",)
+
+    own_mark_name_mapping: dict[str, list[Mark]]
+
+    @classmethod
+    def from_markers(cls, markers: Iterable[Mark]) -> MarkMatcher:
+        mark_name_mapping = collections.defaultdict(list)
+        for mark in markers:
+            mark_name_mapping[mark.name].append(mark)
+        return cls(mark_name_mapping)
+
+    def __call__(self, name: str, /, **kwargs: str | int | bool | None) -> bool:
+        if not (matches := self.own_mark_name_mapping.get(name, [])):
+            return False
+
+        for mark in matches:
+            if all(mark.kwargs.get(k, NOT_SET) == v for k, v in kwargs.items()):
+                return True
+
+        return False
+
+
+def deselect_by_mark(items: list[Item], config: Config) -> None:
+    matchexpr = config.option.markexpr
+    if not matchexpr:
+        return
+
+    expr = _parse_expression(matchexpr, "Wrong expression passed to '-m'")
+    remaining: list[Item] = []
+    deselected: list[Item] = []
+    for item in items:
+        if expr.evaluate(MarkMatcher.from_markers(item.iter_markers())):
+            remaining.append(item)
+        else:
+            deselected.append(item)
+    if deselected:
+        config.hook.pytest_deselected(items=deselected)
+        items[:] = remaining
+
+
+def _parse_expression(expr: str, exc_message: str) -> Expression:
+    try:
+        return Expression.compile(expr)
+    except ParseError as e:
+        raise UsageError(f"{exc_message}: {expr}: {e}") from None
+
+
+def pytest_collection_modifyitems(items: list[Item], config: Config) -> None:
+    deselect_by_keyword(items, config)
+    deselect_by_mark(items, config)
+
+
+def pytest_configure(config: Config) -> None:
+    config.stash[old_mark_config_key] = MARK_GEN._config
+    MARK_GEN._config = config
+
+    empty_parameterset = config.getini(EMPTY_PARAMETERSET_OPTION)
+
+    if empty_parameterset not in ("skip", "xfail", "fail_at_collect", None, ""):
+        raise UsageError(
+            f"{EMPTY_PARAMETERSET_OPTION!s} must be one of skip, xfail or fail_at_collect"
+            f" but it is {empty_parameterset!r}"
+        )
+
+
+def pytest_unconfigure(config: Config) -> None:
+    MARK_GEN._config = config.stash.get(old_mark_config_key, None)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..f3dca097
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/expression.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/expression.cpython-312.pyc
new file mode 100644
index 00000000..f2451624
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/expression.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/structures.cpython-312.pyc
new file mode 100644
index 00000000..6efef5f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/_pytest/mark/__pycache__/structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/expression.py b/backend/venv/lib/python3.12/site-packages/_pytest/mark/expression.py
new file mode 100644
index 00000000..89cc0e94
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/mark/expression.py
@@ -0,0 +1,333 @@
+r"""Evaluate match expressions, as used by `-k` and `-m`.
+
+The grammar is:
+
+expression: expr? EOF
+expr:       and_expr ('or' and_expr)*
+and_expr:   not_expr ('and' not_expr)*
+not_expr:   'not' not_expr | '(' expr ')' | ident kwargs?
+
+ident:      (\w|:|\+|-|\.|\[|\]|\\|/)+
+kwargs:     ('(' name '=' value ( ', ' name '=' value )*  ')')
+name:       a valid ident, but not a reserved keyword
+value:      (unescaped) string literal | (-)?[0-9]+ | 'False' | 'True' | 'None'
+
+The semantics are:
+
+- Empty expression evaluates to False.
+- ident evaluates to True or False according to a provided matcher function.
+- or/and/not evaluate according to the usual boolean semantics.
+- ident with parentheses and keyword arguments evaluates to True or False according to a provided matcher function.
+"""
+
+from __future__ import annotations
+
+import ast
+import dataclasses
+import enum
+import keyword
+import re
+import types
+from typing import Iterator
+from typing import Literal
+from typing import Mapping
+from typing import NoReturn
+from typing import overload
+from typing import Protocol
+from typing import Sequence
+
+
+__all__ = [
+    "Expression",
+    "ParseError",
+]
+
+
+class TokenType(enum.Enum):
+    LPAREN = "left parenthesis"
+    RPAREN = "right parenthesis"
+    OR = "or"
+    AND = "and"
+    NOT = "not"
+    IDENT = "identifier"
+    EOF = "end of input"
+    EQUAL = "="
+    STRING = "string literal"
+    COMMA = ","
+
+
+@dataclasses.dataclass(frozen=True)
+class Token:
+    __slots__ = ("type", "value", "pos")
+    type: TokenType
+    value: str
+    pos: int
+
+
+class ParseError(Exception):
+    """The expression contains invalid syntax.
+
+    :param column: The column in the line where the error occurred (1-based).
+    :param message: A description of the error.
+    """
+
+    def __init__(self, column: int, message: str) -> None:
+        self.column = column
+        self.message = message
+
+    def __str__(self) -> str:
+        return f"at column {self.column}: {self.message}"
+
+
+class Scanner:
+    __slots__ = ("tokens", "current")
+
+    def __init__(self, input: str) -> None:
+        self.tokens = self.lex(input)
+        self.current = next(self.tokens)
+
+    def lex(self, input: str) -> Iterator[Token]:
+        pos = 0
+        while pos < len(input):
+            if input[pos] in (" ", "\t"):
+                pos += 1
+            elif input[pos] == "(":
+                yield Token(TokenType.LPAREN, "(", pos)
+                pos += 1
+            elif input[pos] == ")":
+                yield Token(TokenType.RPAREN, ")", pos)
+                pos += 1
+            elif input[pos] == "=":
+                yield Token(TokenType.EQUAL, "=", pos)
+                pos += 1
+            elif input[pos] == ",":
+                yield Token(TokenType.COMMA, ",", pos)
+                pos += 1
+            elif (quote_char := input[pos]) in ("'", '"'):
+                end_quote_pos = input.find(quote_char, pos + 1)
+                if end_quote_pos == -1:
+                    raise ParseError(
+                        pos + 1,
+                        f'closing quote "{quote_char}" is missing',
+                    )
+                value = input[pos : end_quote_pos + 1]
+                if (backslash_pos := input.find("\\")) != -1:
+                    raise ParseError(
+                        backslash_pos + 1,
+                        r'escaping with "\" not supported in marker expression',
+                    )
+                yield Token(TokenType.STRING, value, pos)
+                pos += len(value)
+            else:
+                match = re.match(r"(:?\w|:|\+|-|\.|\[|\]|\\|/)+", input[pos:])
+                if match:
+                    value = match.group(0)
+                    if value == "or":
+                        yield Token(TokenType.OR, value, pos)
+                    elif value == "and":
+                        yield Token(TokenType.AND, value, pos)
+                    elif value == "not":
+                        yield Token(TokenType.NOT, value, pos)
+                    else:
+                        yield Token(TokenType.IDENT, value, pos)
+                    pos += len(value)
+                else:
+                    raise ParseError(
+                        pos + 1,
+                        f'unexpected character "{input[pos]}"',
+                    )
+        yield Token(TokenType.EOF, "", pos)
+
+    @overload
+    def accept(self, type: TokenType, *, reject: Literal[True]) -> Token: ...
+
+    @overload
+    def accept(
+        self, type: TokenType, *, reject: Literal[False] = False
+    ) -> Token | None: ...
+
+    def accept(self, type: TokenType, *, reject: bool = False) -> Token | None:
+        if self.current.type is type:
+            token = self.current
+            if token.type is not TokenType.EOF:
+                self.current = next(self.tokens)
+            return token
+        if reject:
+            self.reject((type,))
+        return None
+
+    def reject(self, expected: Sequence[TokenType]) -> NoReturn:
+        raise ParseError(
+            self.current.pos + 1,
+            "expected {}; got {}".format(
+                " OR ".join(type.value for type in expected),
+                self.current.type.value,
+            ),
+        )
+
+
+# True, False and None are legal match expression identifiers,
+# but illegal as Python identifiers. To fix this, this prefix
+# is added to identifiers in the conversion to Python AST.
+IDENT_PREFIX = "$"
+
+
+def expression(s: Scanner) -> ast.Expression:
+    if s.accept(TokenType.EOF):
+        ret: ast.expr = ast.Constant(False)
+    else:
+        ret = expr(s)
+        s.accept(TokenType.EOF, reject=True)
+    return ast.fix_missing_locations(ast.Expression(ret))
+
+
+def expr(s: Scanner) -> ast.expr:
+    ret = and_expr(s)
+    while s.accept(TokenType.OR):
+        rhs = and_expr(s)
+        ret = ast.BoolOp(ast.Or(), [ret, rhs])
+    return ret
+
+
+def and_expr(s: Scanner) -> ast.expr:
+    ret = not_expr(s)
+    while s.accept(TokenType.AND):
+        rhs = not_expr(s)
+        ret = ast.BoolOp(ast.And(), [ret, rhs])
+    return ret
+
+
+def not_expr(s: Scanner) -> ast.expr:
+    if s.accept(TokenType.NOT):
+        return ast.UnaryOp(ast.Not(), not_expr(s))
+    if s.accept(TokenType.LPAREN):
+        ret = expr(s)
+        s.accept(TokenType.RPAREN, reject=True)
+        return ret
+    ident = s.accept(TokenType.IDENT)
+    if ident:
+        name = ast.Name(IDENT_PREFIX + ident.value, ast.Load())
+        if s.accept(TokenType.LPAREN):
+            ret = ast.Call(func=name, args=[], keywords=all_kwargs(s))
+            s.accept(TokenType.RPAREN, reject=True)
+        else:
+            ret = name
+        return ret
+
+    s.reject((TokenType.NOT, TokenType.LPAREN, TokenType.IDENT))
+
+
+BUILTIN_MATCHERS = {"True": True, "False": False, "None": None}
+
+
+def single_kwarg(s: Scanner) -> ast.keyword:
+    keyword_name = s.accept(TokenType.IDENT, reject=True)
+    if not keyword_name.value.isidentifier():
+        raise ParseError(
+            keyword_name.pos + 1,
+            f"not a valid python identifier {keyword_name.value}",
+        )
+    if keyword.iskeyword(keyword_name.value):
+        raise ParseError(
+            keyword_name.pos + 1,
+            f"unexpected reserved python keyword `{keyword_name.value}`",
+        )
+    s.accept(TokenType.EQUAL, reject=True)
+
+    if value_token := s.accept(TokenType.STRING):
+        value: str | int | bool | None = value_token.value[1:-1]  # strip quotes
+    else:
+        value_token = s.accept(TokenType.IDENT, reject=True)
+        if (
+            (number := value_token.value).isdigit()
+            or number.startswith("-")
+            and number[1:].isdigit()
+        ):
+            value = int(number)
+        elif value_token.value in BUILTIN_MATCHERS:
+            value = BUILTIN_MATCHERS[value_token.value]
+        else:
+            raise ParseError(
+                value_token.pos + 1,
+                f'unexpected character/s "{value_token.value}"',
+            )
+
+    ret = ast.keyword(keyword_name.value, ast.Constant(value))
+    return ret
+
+
+def all_kwargs(s: Scanner) -> list[ast.keyword]:
+    ret = [single_kwarg(s)]
+    while s.accept(TokenType.COMMA):
+        ret.append(single_kwarg(s))
+    return ret
+
+
+class MatcherCall(Protocol):
+    def __call__(self, name: str, /, **kwargs: str | int | bool | None) -> bool: ...
+
+
+@dataclasses.dataclass
+class MatcherNameAdapter:
+    matcher: MatcherCall
+    name: str
+
+    def __bool__(self) -> bool:
+        return self.matcher(self.name)
+
+    def __call__(self, **kwargs: str | int | bool | None) -> bool:
+        return self.matcher(self.name, **kwargs)
+
+
+class MatcherAdapter(Mapping[str, MatcherNameAdapter]):
+    """Adapts a matcher function to a locals mapping as required by eval()."""
+
+    def __init__(self, matcher: MatcherCall) -> None:
+        self.matcher = matcher
+
+    def __getitem__(self, key: str) -> MatcherNameAdapter:
+        return MatcherNameAdapter(matcher=self.matcher, name=key[len(IDENT_PREFIX) :])
+
+    def __iter__(self) -> Iterator[str]:
+        raise NotImplementedError()
+
+    def __len__(self) -> int:
+        raise NotImplementedError()
+
+
+class Expression:
+    """A compiled match expression as used by -k and -m.
+
+    The expression can be evaluated against different matchers.
+    """
+
+    __slots__ = ("code",)
+
+    def __init__(self, code: types.CodeType) -> None:
+        self.code = code
+
+    @classmethod
+    def compile(self, input: str) -> Expression:
+        """Compile a match expression.
+
+        :param input: The input expression - one line.
+        """
+        astexpr = expression(Scanner(input))
+        code: types.CodeType = compile(
+            astexpr,
+            filename="<pytest match expression>",
+            mode="eval",
+        )
+        return Expression(code)
+
+    def evaluate(self, matcher: MatcherCall) -> bool:
+        """Evaluate the match expression.
+
+        :param matcher:
+            Given an identifier, should return whether it matches or not.
+            Should be prepared to handle arbitrary strings as input.
+
+        :returns: Whether the expression matches or not.
+        """
+        ret: bool = bool(eval(self.code, {"__builtins__": {}}, MatcherAdapter(matcher)))
+        return ret
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/mark/structures.py b/backend/venv/lib/python3.12/site-packages/_pytest/mark/structures.py
new file mode 100644
index 00000000..92ade55f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/mark/structures.py
@@ -0,0 +1,615 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import collections.abc
+import dataclasses
+import inspect
+from typing import Any
+from typing import Callable
+from typing import Collection
+from typing import final
+from typing import Iterable
+from typing import Iterator
+from typing import Mapping
+from typing import MutableMapping
+from typing import NamedTuple
+from typing import overload
+from typing import Sequence
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import warnings
+
+from .._code import getfslineno
+from ..compat import ascii_escaped
+from ..compat import NOTSET
+from ..compat import NotSetType
+from _pytest.config import Config
+from _pytest.deprecated import check_ispytest
+from _pytest.deprecated import MARKED_FIXTURE
+from _pytest.outcomes import fail
+from _pytest.scope import _ScopeName
+from _pytest.warning_types import PytestUnknownMarkWarning
+
+
+if TYPE_CHECKING:
+    from ..nodes import Node
+
+
+EMPTY_PARAMETERSET_OPTION = "empty_parameter_set_mark"
+
+
+def istestfunc(func) -> bool:
+    return callable(func) and getattr(func, "__name__", "<lambda>") != "<lambda>"
+
+
+def get_empty_parameterset_mark(
+    config: Config, argnames: Sequence[str], func
+) -> MarkDecorator:
+    from ..nodes import Collector
+
+    fs, lineno = getfslineno(func)
+    reason = "got empty parameter set %r, function %s at %s:%d" % (
+        argnames,
+        func.__name__,
+        fs,
+        lineno,
+    )
+
+    requested_mark = config.getini(EMPTY_PARAMETERSET_OPTION)
+    if requested_mark in ("", None, "skip"):
+        mark = MARK_GEN.skip(reason=reason)
+    elif requested_mark == "xfail":
+        mark = MARK_GEN.xfail(reason=reason, run=False)
+    elif requested_mark == "fail_at_collect":
+        f_name = func.__name__
+        _, lineno = getfslineno(func)
+        raise Collector.CollectError(
+            "Empty parameter set in '%s' at line %d" % (f_name, lineno + 1)
+        )
+    else:
+        raise LookupError(requested_mark)
+    return mark
+
+
+class ParameterSet(NamedTuple):
+    values: Sequence[object | NotSetType]
+    marks: Collection[MarkDecorator | Mark]
+    id: str | None
+
+    @classmethod
+    def param(
+        cls,
+        *values: object,
+        marks: MarkDecorator | Collection[MarkDecorator | Mark] = (),
+        id: str | None = None,
+    ) -> ParameterSet:
+        if isinstance(marks, MarkDecorator):
+            marks = (marks,)
+        else:
+            assert isinstance(marks, collections.abc.Collection)
+
+        if id is not None:
+            if not isinstance(id, str):
+                raise TypeError(f"Expected id to be a string, got {type(id)}: {id!r}")
+            id = ascii_escaped(id)
+        return cls(values, marks, id)
+
+    @classmethod
+    def extract_from(
+        cls,
+        parameterset: ParameterSet | Sequence[object] | object,
+        force_tuple: bool = False,
+    ) -> ParameterSet:
+        """Extract from an object or objects.
+
+        :param parameterset:
+            A legacy style parameterset that may or may not be a tuple,
+            and may or may not be wrapped into a mess of mark objects.
+
+        :param force_tuple:
+            Enforce tuple wrapping so single argument tuple values
+            don't get decomposed and break tests.
+        """
+        if isinstance(parameterset, cls):
+            return parameterset
+        if force_tuple:
+            return cls.param(parameterset)
+        else:
+            # TODO: Refactor to fix this type-ignore. Currently the following
+            # passes type-checking but crashes:
+            #
+            #   @pytest.mark.parametrize(('x', 'y'), [1, 2])
+            #   def test_foo(x, y): pass
+            return cls(parameterset, marks=[], id=None)  # type: ignore[arg-type]
+
+    @staticmethod
+    def _parse_parametrize_args(
+        argnames: str | Sequence[str],
+        argvalues: Iterable[ParameterSet | Sequence[object] | object],
+        *args,
+        **kwargs,
+    ) -> tuple[Sequence[str], bool]:
+        if isinstance(argnames, str):
+            argnames = [x.strip() for x in argnames.split(",") if x.strip()]
+            force_tuple = len(argnames) == 1
+        else:
+            force_tuple = False
+        return argnames, force_tuple
+
+    @staticmethod
+    def _parse_parametrize_parameters(
+        argvalues: Iterable[ParameterSet | Sequence[object] | object],
+        force_tuple: bool,
+    ) -> list[ParameterSet]:
+        return [
+            ParameterSet.extract_from(x, force_tuple=force_tuple) for x in argvalues
+        ]
+
+    @classmethod
+    def _for_parametrize(
+        cls,
+        argnames: str | Sequence[str],
+        argvalues: Iterable[ParameterSet | Sequence[object] | object],
+        func,
+        config: Config,
+        nodeid: str,
+    ) -> tuple[Sequence[str], list[ParameterSet]]:
+        argnames, force_tuple = cls._parse_parametrize_args(argnames, argvalues)
+        parameters = cls._parse_parametrize_parameters(argvalues, force_tuple)
+        del argvalues
+
+        if parameters:
+            # Check all parameter sets have the correct number of values.
+            for param in parameters:
+                if len(param.values) != len(argnames):
+                    msg = (
+                        '{nodeid}: in "parametrize" the number of names ({names_len}):\n'
+                        "  {names}\n"
+                        "must be equal to the number of values ({values_len}):\n"
+                        "  {values}"
+                    )
+                    fail(
+                        msg.format(
+                            nodeid=nodeid,
+                            values=param.values,
+                            names=argnames,
+                            names_len=len(argnames),
+                            values_len=len(param.values),
+                        ),
+                        pytrace=False,
+                    )
+        else:
+            # Empty parameter set (likely computed at runtime): create a single
+            # parameter set with NOTSET values, with the "empty parameter set" mark applied to it.
+            mark = get_empty_parameterset_mark(config, argnames, func)
+            parameters.append(
+                ParameterSet(values=(NOTSET,) * len(argnames), marks=[mark], id=None)
+            )
+        return argnames, parameters
+
+
+@final
+@dataclasses.dataclass(frozen=True)
+class Mark:
+    """A pytest mark."""
+
+    #: Name of the mark.
+    name: str
+    #: Positional arguments of the mark decorator.
+    args: tuple[Any, ...]
+    #: Keyword arguments of the mark decorator.
+    kwargs: Mapping[str, Any]
+
+    #: Source Mark for ids with parametrize Marks.
+    _param_ids_from: Mark | None = dataclasses.field(default=None, repr=False)
+    #: Resolved/generated ids with parametrize Marks.
+    _param_ids_generated: Sequence[str] | None = dataclasses.field(
+        default=None, repr=False
+    )
+
+    def __init__(
+        self,
+        name: str,
+        args: tuple[Any, ...],
+        kwargs: Mapping[str, Any],
+        param_ids_from: Mark | None = None,
+        param_ids_generated: Sequence[str] | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        """:meta private:"""
+        check_ispytest(_ispytest)
+        # Weirdness to bypass frozen=True.
+        object.__setattr__(self, "name", name)
+        object.__setattr__(self, "args", args)
+        object.__setattr__(self, "kwargs", kwargs)
+        object.__setattr__(self, "_param_ids_from", param_ids_from)
+        object.__setattr__(self, "_param_ids_generated", param_ids_generated)
+
+    def _has_param_ids(self) -> bool:
+        return "ids" in self.kwargs or len(self.args) >= 4
+
+    def combined_with(self, other: Mark) -> Mark:
+        """Return a new Mark which is a combination of this
+        Mark and another Mark.
+
+        Combines by appending args and merging kwargs.
+
+        :param Mark other: The mark to combine with.
+        :rtype: Mark
+        """
+        assert self.name == other.name
+
+        # Remember source of ids with parametrize Marks.
+        param_ids_from: Mark | None = None
+        if self.name == "parametrize":
+            if other._has_param_ids():
+                param_ids_from = other
+            elif self._has_param_ids():
+                param_ids_from = self
+
+        return Mark(
+            self.name,
+            self.args + other.args,
+            dict(self.kwargs, **other.kwargs),
+            param_ids_from=param_ids_from,
+            _ispytest=True,
+        )
+
+
+# A generic parameter designating an object to which a Mark may
+# be applied -- a test function (callable) or class.
+# Note: a lambda is not allowed, but this can't be represented.
+Markable = TypeVar("Markable", bound=Union[Callable[..., object], type])
+
+
+@dataclasses.dataclass
+class MarkDecorator:
+    """A decorator for applying a mark on test functions and classes.
+
+    ``MarkDecorators`` are created with ``pytest.mark``::
+
+        mark1 = pytest.mark.NAME  # Simple MarkDecorator
+        mark2 = pytest.mark.NAME(name1=value)  # Parametrized MarkDecorator
+
+    and can then be applied as decorators to test functions::
+
+        @mark2
+        def test_function():
+            pass
+
+    When a ``MarkDecorator`` is called, it does the following:
+
+    1. If called with a single class as its only positional argument and no
+       additional keyword arguments, it attaches the mark to the class so it
+       gets applied automatically to all test cases found in that class.
+
+    2. If called with a single function as its only positional argument and
+       no additional keyword arguments, it attaches the mark to the function,
+       containing all the arguments already stored internally in the
+       ``MarkDecorator``.
+
+    3. When called in any other case, it returns a new ``MarkDecorator``
+       instance with the original ``MarkDecorator``'s content updated with
+       the arguments passed to this call.
+
+    Note: The rules above prevent a ``MarkDecorator`` from storing only a
+    single function or class reference as its positional argument with no
+    additional keyword or positional arguments. You can work around this by
+    using `with_args()`.
+    """
+
+    mark: Mark
+
+    def __init__(self, mark: Mark, *, _ispytest: bool = False) -> None:
+        """:meta private:"""
+        check_ispytest(_ispytest)
+        self.mark = mark
+
+    @property
+    def name(self) -> str:
+        """Alias for mark.name."""
+        return self.mark.name
+
+    @property
+    def args(self) -> tuple[Any, ...]:
+        """Alias for mark.args."""
+        return self.mark.args
+
+    @property
+    def kwargs(self) -> Mapping[str, Any]:
+        """Alias for mark.kwargs."""
+        return self.mark.kwargs
+
+    @property
+    def markname(self) -> str:
+        """:meta private:"""
+        return self.name  # for backward-compat (2.4.1 had this attr)
+
+    def with_args(self, *args: object, **kwargs: object) -> MarkDecorator:
+        """Return a MarkDecorator with extra arguments added.
+
+        Unlike calling the MarkDecorator, with_args() can be used even
+        if the sole argument is a callable/class.
+        """
+        mark = Mark(self.name, args, kwargs, _ispytest=True)
+        return MarkDecorator(self.mark.combined_with(mark), _ispytest=True)
+
+    # Type ignored because the overloads overlap with an incompatible
+    # return type. Not much we can do about that. Thankfully mypy picks
+    # the first match so it works out even if we break the rules.
+    @overload
+    def __call__(self, arg: Markable) -> Markable:  # type: ignore[overload-overlap]
+        pass
+
+    @overload
+    def __call__(self, *args: object, **kwargs: object) -> MarkDecorator:
+        pass
+
+    def __call__(self, *args: object, **kwargs: object):
+        """Call the MarkDecorator."""
+        if args and not kwargs:
+            func = args[0]
+            is_class = inspect.isclass(func)
+            if len(args) == 1 and (istestfunc(func) or is_class):
+                store_mark(func, self.mark, stacklevel=3)
+                return func
+        return self.with_args(*args, **kwargs)
+
+
+def get_unpacked_marks(
+    obj: object | type,
+    *,
+    consider_mro: bool = True,
+) -> list[Mark]:
+    """Obtain the unpacked marks that are stored on an object.
+
+    If obj is a class and consider_mro is true, return marks applied to
+    this class and all of its super-classes in MRO order. If consider_mro
+    is false, only return marks applied directly to this class.
+    """
+    if isinstance(obj, type):
+        if not consider_mro:
+            mark_lists = [obj.__dict__.get("pytestmark", [])]
+        else:
+            mark_lists = [
+                x.__dict__.get("pytestmark", []) for x in reversed(obj.__mro__)
+            ]
+        mark_list = []
+        for item in mark_lists:
+            if isinstance(item, list):
+                mark_list.extend(item)
+            else:
+                mark_list.append(item)
+    else:
+        mark_attribute = getattr(obj, "pytestmark", [])
+        if isinstance(mark_attribute, list):
+            mark_list = mark_attribute
+        else:
+            mark_list = [mark_attribute]
+    return list(normalize_mark_list(mark_list))
+
+
+def normalize_mark_list(
+    mark_list: Iterable[Mark | MarkDecorator],
+) -> Iterable[Mark]:
+    """
+    Normalize an iterable of Mark or MarkDecorator objects into a list of marks
+    by retrieving the `mark` attribute on MarkDecorator instances.
+
+    :param mark_list: marks to normalize
+    :returns: A new list of the extracted Mark objects
+    """
+    for mark in mark_list:
+        mark_obj = getattr(mark, "mark", mark)
+        if not isinstance(mark_obj, Mark):
+            raise TypeError(f"got {mark_obj!r} instead of Mark")
+        yield mark_obj
+
+
+def store_mark(obj, mark: Mark, *, stacklevel: int = 2) -> None:
+    """Store a Mark on an object.
+
+    This is used to implement the Mark declarations/decorators correctly.
+    """
+    assert isinstance(mark, Mark), mark
+
+    from ..fixtures import getfixturemarker
+
+    if getfixturemarker(obj) is not None:
+        warnings.warn(MARKED_FIXTURE, stacklevel=stacklevel)
+
+    # Always reassign name to avoid updating pytestmark in a reference that
+    # was only borrowed.
+    obj.pytestmark = [*get_unpacked_marks(obj, consider_mro=False), mark]
+
+
+# Typing for builtin pytest marks. This is cheating; it gives builtin marks
+# special privilege, and breaks modularity. But practicality beats purity...
+if TYPE_CHECKING:
+
+    class _SkipMarkDecorator(MarkDecorator):
+        @overload  # type: ignore[override,no-overload-impl]
+        def __call__(self, arg: Markable) -> Markable: ...
+
+        @overload
+        def __call__(self, reason: str = ...) -> MarkDecorator: ...
+
+    class _SkipifMarkDecorator(MarkDecorator):
+        def __call__(  # type: ignore[override]
+            self,
+            condition: str | bool = ...,
+            *conditions: str | bool,
+            reason: str = ...,
+        ) -> MarkDecorator: ...
+
+    class _XfailMarkDecorator(MarkDecorator):
+        @overload  # type: ignore[override,no-overload-impl]
+        def __call__(self, arg: Markable) -> Markable: ...
+
+        @overload
+        def __call__(
+            self,
+            condition: str | bool = False,
+            *conditions: str | bool,
+            reason: str = ...,
+            run: bool = ...,
+            raises: None | type[BaseException] | tuple[type[BaseException], ...] = ...,
+            strict: bool = ...,
+        ) -> MarkDecorator: ...
+
+    class _ParametrizeMarkDecorator(MarkDecorator):
+        def __call__(  # type: ignore[override]
+            self,
+            argnames: str | Sequence[str],
+            argvalues: Iterable[ParameterSet | Sequence[object] | object],
+            *,
+            indirect: bool | Sequence[str] = ...,
+            ids: Iterable[None | str | float | int | bool]
+            | Callable[[Any], object | None]
+            | None = ...,
+            scope: _ScopeName | None = ...,
+        ) -> MarkDecorator: ...
+
+    class _UsefixturesMarkDecorator(MarkDecorator):
+        def __call__(self, *fixtures: str) -> MarkDecorator:  # type: ignore[override]
+            ...
+
+    class _FilterwarningsMarkDecorator(MarkDecorator):
+        def __call__(self, *filters: str) -> MarkDecorator:  # type: ignore[override]
+            ...
+
+
+@final
+class MarkGenerator:
+    """Factory for :class:`MarkDecorator` objects - exposed as
+    a ``pytest.mark`` singleton instance.
+
+    Example::
+
+         import pytest
+
+
+         @pytest.mark.slowtest
+         def test_function():
+             pass
+
+    applies a 'slowtest' :class:`Mark` on ``test_function``.
+    """
+
+    # See TYPE_CHECKING above.
+    if TYPE_CHECKING:
+        skip: _SkipMarkDecorator
+        skipif: _SkipifMarkDecorator
+        xfail: _XfailMarkDecorator
+        parametrize: _ParametrizeMarkDecorator
+        usefixtures: _UsefixturesMarkDecorator
+        filterwarnings: _FilterwarningsMarkDecorator
+
+    def __init__(self, *, _ispytest: bool = False) -> None:
+        check_ispytest(_ispytest)
+        self._config: Config | None = None
+        self._markers: set[str] = set()
+
+    def __getattr__(self, name: str) -> MarkDecorator:
+        """Generate a new :class:`MarkDecorator` with the given name."""
+        if name[0] == "_":
+            raise AttributeError("Marker name must NOT start with underscore")
+
+        if self._config is not None:
+            # We store a set of markers as a performance optimisation - if a mark
+            # name is in the set we definitely know it, but a mark may be known and
+            # not in the set.  We therefore start by updating the set!
+            if name not in self._markers:
+                for line in self._config.getini("markers"):
+                    # example lines: "skipif(condition): skip the given test if..."
+                    # or "hypothesis: tests which use Hypothesis", so to get the
+                    # marker name we split on both `:` and `(`.
+                    marker = line.split(":")[0].split("(")[0].strip()
+                    self._markers.add(marker)
+
+            # If the name is not in the set of known marks after updating,
+            # then it really is time to issue a warning or an error.
+            if name not in self._markers:
+                if self._config.option.strict_markers or self._config.option.strict:
+                    fail(
+                        f"{name!r} not found in `markers` configuration option",
+                        pytrace=False,
+                    )
+
+                # Raise a specific error for common misspellings of "parametrize".
+                if name in ["parameterize", "parametrise", "parameterise"]:
+                    __tracebackhide__ = True
+                    fail(f"Unknown '{name}' mark, did you mean 'parametrize'?")
+
+                warnings.warn(
+                    f"Unknown pytest.mark.{name} - is this a typo?  You can register "
+                    "custom marks to avoid this warning - for details, see "
+                    "https://docs.pytest.org/en/stable/how-to/mark.html",
+                    PytestUnknownMarkWarning,
+                    2,
+                )
+
+        return MarkDecorator(Mark(name, (), {}, _ispytest=True), _ispytest=True)
+
+
+MARK_GEN = MarkGenerator(_ispytest=True)
+
+
+@final
+class NodeKeywords(MutableMapping[str, Any]):
+    __slots__ = ("node", "parent", "_markers")
+
+    def __init__(self, node: Node) -> None:
+        self.node = node
+        self.parent = node.parent
+        self._markers = {node.name: True}
+
+    def __getitem__(self, key: str) -> Any:
+        try:
+            return self._markers[key]
+        except KeyError:
+            if self.parent is None:
+                raise
+            return self.parent.keywords[key]
+
+    def __setitem__(self, key: str, value: Any) -> None:
+        self._markers[key] = value
+
+    # Note: we could've avoided explicitly implementing some of the methods
+    # below and use the collections.abc fallback, but that would be slow.
+
+    def __contains__(self, key: object) -> bool:
+        return (
+            key in self._markers
+            or self.parent is not None
+            and key in self.parent.keywords
+        )
+
+    def update(  # type: ignore[override]
+        self,
+        other: Mapping[str, Any] | Iterable[tuple[str, Any]] = (),
+        **kwds: Any,
+    ) -> None:
+        self._markers.update(other)
+        self._markers.update(kwds)
+
+    def __delitem__(self, key: str) -> None:
+        raise ValueError("cannot delete key in keywords dict")
+
+    def __iter__(self) -> Iterator[str]:
+        # Doesn't need to be fast.
+        yield from self._markers
+        if self.parent is not None:
+            for keyword in self.parent.keywords:
+                # self._marks and self.parent.keywords can have duplicates.
+                if keyword not in self._markers:
+                    yield keyword
+
+    def __len__(self) -> int:
+        # Doesn't need to be fast.
+        return sum(1 for keyword in self)
+
+    def __repr__(self) -> str:
+        return f"<NodeKeywords for node {self.node}>"
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/monkeypatch.py b/backend/venv/lib/python3.12/site-packages/_pytest/monkeypatch.py
new file mode 100644
index 00000000..46eb1724
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/monkeypatch.py
@@ -0,0 +1,415 @@
+# mypy: allow-untyped-defs
+"""Monkeypatching and mocking functionality."""
+
+from __future__ import annotations
+
+from contextlib import contextmanager
+import os
+import re
+import sys
+from typing import Any
+from typing import final
+from typing import Generator
+from typing import Mapping
+from typing import MutableMapping
+from typing import overload
+from typing import TypeVar
+import warnings
+
+from _pytest.fixtures import fixture
+from _pytest.warning_types import PytestWarning
+
+
+RE_IMPORT_ERROR_NAME = re.compile(r"^No module named (.*)$")
+
+
+K = TypeVar("K")
+V = TypeVar("V")
+
+
+@fixture
+def monkeypatch() -> Generator[MonkeyPatch]:
+    """A convenient fixture for monkey-patching.
+
+    The fixture provides these methods to modify objects, dictionaries, or
+    :data:`os.environ`:
+
+    * :meth:`monkeypatch.setattr(obj, name, value, raising=True) <pytest.MonkeyPatch.setattr>`
+    * :meth:`monkeypatch.delattr(obj, name, raising=True) <pytest.MonkeyPatch.delattr>`
+    * :meth:`monkeypatch.setitem(mapping, name, value) <pytest.MonkeyPatch.setitem>`
+    * :meth:`monkeypatch.delitem(obj, name, raising=True) <pytest.MonkeyPatch.delitem>`
+    * :meth:`monkeypatch.setenv(name, value, prepend=None) <pytest.MonkeyPatch.setenv>`
+    * :meth:`monkeypatch.delenv(name, raising=True) <pytest.MonkeyPatch.delenv>`
+    * :meth:`monkeypatch.syspath_prepend(path) <pytest.MonkeyPatch.syspath_prepend>`
+    * :meth:`monkeypatch.chdir(path) <pytest.MonkeyPatch.chdir>`
+    * :meth:`monkeypatch.context() <pytest.MonkeyPatch.context>`
+
+    All modifications will be undone after the requesting test function or
+    fixture has finished. The ``raising`` parameter determines if a :class:`KeyError`
+    or :class:`AttributeError` will be raised if the set/deletion operation does not have the
+    specified target.
+
+    To undo modifications done by the fixture in a contained scope,
+    use :meth:`context() <pytest.MonkeyPatch.context>`.
+    """
+    mpatch = MonkeyPatch()
+    yield mpatch
+    mpatch.undo()
+
+
+def resolve(name: str) -> object:
+    # Simplified from zope.dottedname.
+    parts = name.split(".")
+
+    used = parts.pop(0)
+    found: object = __import__(used)
+    for part in parts:
+        used += "." + part
+        try:
+            found = getattr(found, part)
+        except AttributeError:
+            pass
+        else:
+            continue
+        # We use explicit un-nesting of the handling block in order
+        # to avoid nested exceptions.
+        try:
+            __import__(used)
+        except ImportError as ex:
+            expected = str(ex).split()[-1]
+            if expected == used:
+                raise
+            else:
+                raise ImportError(f"import error in {used}: {ex}") from ex
+        found = annotated_getattr(found, part, used)
+    return found
+
+
+def annotated_getattr(obj: object, name: str, ann: str) -> object:
+    try:
+        obj = getattr(obj, name)
+    except AttributeError as e:
+        raise AttributeError(
+            f"{type(obj).__name__!r} object at {ann} has no attribute {name!r}"
+        ) from e
+    return obj
+
+
+def derive_importpath(import_path: str, raising: bool) -> tuple[str, object]:
+    if not isinstance(import_path, str) or "." not in import_path:
+        raise TypeError(f"must be absolute import path string, not {import_path!r}")
+    module, attr = import_path.rsplit(".", 1)
+    target = resolve(module)
+    if raising:
+        annotated_getattr(target, attr, ann=module)
+    return attr, target
+
+
+class Notset:
+    def __repr__(self) -> str:
+        return "<notset>"
+
+
+notset = Notset()
+
+
+@final
+class MonkeyPatch:
+    """Helper to conveniently monkeypatch attributes/items/environment
+    variables/syspath.
+
+    Returned by the :fixture:`monkeypatch` fixture.
+
+    .. versionchanged:: 6.2
+        Can now also be used directly as `pytest.MonkeyPatch()`, for when
+        the fixture is not available. In this case, use
+        :meth:`with MonkeyPatch.context() as mp: <context>` or remember to call
+        :meth:`undo` explicitly.
+    """
+
+    def __init__(self) -> None:
+        self._setattr: list[tuple[object, str, object]] = []
+        self._setitem: list[tuple[Mapping[Any, Any], object, object]] = []
+        self._cwd: str | None = None
+        self._savesyspath: list[str] | None = None
+
+    @classmethod
+    @contextmanager
+    def context(cls) -> Generator[MonkeyPatch]:
+        """Context manager that returns a new :class:`MonkeyPatch` object
+        which undoes any patching done inside the ``with`` block upon exit.
+
+        Example:
+
+        .. code-block:: python
+
+            import functools
+
+
+            def test_partial(monkeypatch):
+                with monkeypatch.context() as m:
+                    m.setattr(functools, "partial", 3)
+
+        Useful in situations where it is desired to undo some patches before the test ends,
+        such as mocking ``stdlib`` functions that might break pytest itself if mocked (for examples
+        of this see :issue:`3290`).
+        """
+        m = cls()
+        try:
+            yield m
+        finally:
+            m.undo()
+
+    @overload
+    def setattr(
+        self,
+        target: str,
+        name: object,
+        value: Notset = ...,
+        raising: bool = ...,
+    ) -> None: ...
+
+    @overload
+    def setattr(
+        self,
+        target: object,
+        name: str,
+        value: object,
+        raising: bool = ...,
+    ) -> None: ...
+
+    def setattr(
+        self,
+        target: str | object,
+        name: object | str,
+        value: object = notset,
+        raising: bool = True,
+    ) -> None:
+        """
+        Set attribute value on target, memorizing the old value.
+
+        For example:
+
+        .. code-block:: python
+
+            import os
+
+            monkeypatch.setattr(os, "getcwd", lambda: "/")
+
+        The code above replaces the :func:`os.getcwd` function by a ``lambda`` which
+        always returns ``"/"``.
+
+        For convenience, you can specify a string as ``target`` which
+        will be interpreted as a dotted import path, with the last part
+        being the attribute name:
+
+        .. code-block:: python
+
+            monkeypatch.setattr("os.getcwd", lambda: "/")
+
+        Raises :class:`AttributeError` if the attribute does not exist, unless
+        ``raising`` is set to False.
+
+        **Where to patch**
+
+        ``monkeypatch.setattr`` works by (temporarily) changing the object that a name points to with another one.
+        There can be many names pointing to any individual object, so for patching to work you must ensure
+        that you patch the name used by the system under test.
+
+        See the section :ref:`Where to patch <python:where-to-patch>` in the :mod:`unittest.mock`
+        docs for a complete explanation, which is meant for :func:`unittest.mock.patch` but
+        applies to ``monkeypatch.setattr`` as well.
+        """
+        __tracebackhide__ = True
+        import inspect
+
+        if isinstance(value, Notset):
+            if not isinstance(target, str):
+                raise TypeError(
+                    "use setattr(target, name, value) or "
+                    "setattr(target, value) with target being a dotted "
+                    "import string"
+                )
+            value = name
+            name, target = derive_importpath(target, raising)
+        else:
+            if not isinstance(name, str):
+                raise TypeError(
+                    "use setattr(target, name, value) with name being a string or "
+                    "setattr(target, value) with target being a dotted "
+                    "import string"
+                )
+
+        oldval = getattr(target, name, notset)
+        if raising and oldval is notset:
+            raise AttributeError(f"{target!r} has no attribute {name!r}")
+
+        # avoid class descriptors like staticmethod/classmethod
+        if inspect.isclass(target):
+            oldval = target.__dict__.get(name, notset)
+        self._setattr.append((target, name, oldval))
+        setattr(target, name, value)
+
+    def delattr(
+        self,
+        target: object | str,
+        name: str | Notset = notset,
+        raising: bool = True,
+    ) -> None:
+        """Delete attribute ``name`` from ``target``.
+
+        If no ``name`` is specified and ``target`` is a string
+        it will be interpreted as a dotted import path with the
+        last part being the attribute name.
+
+        Raises AttributeError it the attribute does not exist, unless
+        ``raising`` is set to False.
+        """
+        __tracebackhide__ = True
+        import inspect
+
+        if isinstance(name, Notset):
+            if not isinstance(target, str):
+                raise TypeError(
+                    "use delattr(target, name) or "
+                    "delattr(target) with target being a dotted "
+                    "import string"
+                )
+            name, target = derive_importpath(target, raising)
+
+        if not hasattr(target, name):
+            if raising:
+                raise AttributeError(name)
+        else:
+            oldval = getattr(target, name, notset)
+            # Avoid class descriptors like staticmethod/classmethod.
+            if inspect.isclass(target):
+                oldval = target.__dict__.get(name, notset)
+            self._setattr.append((target, name, oldval))
+            delattr(target, name)
+
+    def setitem(self, dic: Mapping[K, V], name: K, value: V) -> None:
+        """Set dictionary entry ``name`` to value."""
+        self._setitem.append((dic, name, dic.get(name, notset)))
+        # Not all Mapping types support indexing, but MutableMapping doesn't support TypedDict
+        dic[name] = value  # type: ignore[index]
+
+    def delitem(self, dic: Mapping[K, V], name: K, raising: bool = True) -> None:
+        """Delete ``name`` from dict.
+
+        Raises ``KeyError`` if it doesn't exist, unless ``raising`` is set to
+        False.
+        """
+        if name not in dic:
+            if raising:
+                raise KeyError(name)
+        else:
+            self._setitem.append((dic, name, dic.get(name, notset)))
+            # Not all Mapping types support indexing, but MutableMapping doesn't support TypedDict
+            del dic[name]  # type: ignore[attr-defined]
+
+    def setenv(self, name: str, value: str, prepend: str | None = None) -> None:
+        """Set environment variable ``name`` to ``value``.
+
+        If ``prepend`` is a character, read the current environment variable
+        value and prepend the ``value`` adjoined with the ``prepend``
+        character.
+        """
+        if not isinstance(value, str):
+            warnings.warn(  # type: ignore[unreachable]
+                PytestWarning(
+                    f"Value of environment variable {name} type should be str, but got "
+                    f"{value!r} (type: {type(value).__name__}); converted to str implicitly"
+                ),
+                stacklevel=2,
+            )
+            value = str(value)
+        if prepend and name in os.environ:
+            value = value + prepend + os.environ[name]
+        self.setitem(os.environ, name, value)
+
+    def delenv(self, name: str, raising: bool = True) -> None:
+        """Delete ``name`` from the environment.
+
+        Raises ``KeyError`` if it does not exist, unless ``raising`` is set to
+        False.
+        """
+        environ: MutableMapping[str, str] = os.environ
+        self.delitem(environ, name, raising=raising)
+
+    def syspath_prepend(self, path) -> None:
+        """Prepend ``path`` to ``sys.path`` list of import locations."""
+        if self._savesyspath is None:
+            self._savesyspath = sys.path[:]
+        sys.path.insert(0, str(path))
+
+        # https://github.com/pypa/setuptools/blob/d8b901bc/docs/pkg_resources.txt#L162-L171
+        # this is only needed when pkg_resources was already loaded by the namespace package
+        if "pkg_resources" in sys.modules:
+            from pkg_resources import fixup_namespace_packages
+
+            fixup_namespace_packages(str(path))
+
+        # A call to syspathinsert() usually means that the caller wants to
+        # import some dynamically created files, thus with python3 we
+        # invalidate its import caches.
+        # This is especially important when any namespace package is in use,
+        # since then the mtime based FileFinder cache (that gets created in
+        # this case already) gets not invalidated when writing the new files
+        # quickly afterwards.
+        from importlib import invalidate_caches
+
+        invalidate_caches()
+
+    def chdir(self, path: str | os.PathLike[str]) -> None:
+        """Change the current working directory to the specified path.
+
+        :param path:
+            The path to change into.
+        """
+        if self._cwd is None:
+            self._cwd = os.getcwd()
+        os.chdir(path)
+
+    def undo(self) -> None:
+        """Undo previous changes.
+
+        This call consumes the undo stack. Calling it a second time has no
+        effect unless you do more monkeypatching after the undo call.
+
+        There is generally no need to call `undo()`, since it is
+        called automatically during tear-down.
+
+        .. note::
+            The same `monkeypatch` fixture is used across a
+            single test function invocation. If `monkeypatch` is used both by
+            the test function itself and one of the test fixtures,
+            calling `undo()` will undo all of the changes made in
+            both functions.
+
+            Prefer to use :meth:`context() <pytest.MonkeyPatch.context>` instead.
+        """
+        for obj, name, value in reversed(self._setattr):
+            if value is not notset:
+                setattr(obj, name, value)
+            else:
+                delattr(obj, name)
+        self._setattr[:] = []
+        for dictionary, key, value in reversed(self._setitem):
+            if value is notset:
+                try:
+                    # Not all Mapping types support indexing, but MutableMapping doesn't support TypedDict
+                    del dictionary[key]  # type: ignore[attr-defined]
+                except KeyError:
+                    pass  # Was already deleted, so we have the desired state.
+            else:
+                # Not all Mapping types support indexing, but MutableMapping doesn't support TypedDict
+                dictionary[key] = value  # type: ignore[index]
+        self._setitem[:] = []
+        if self._savesyspath is not None:
+            sys.path[:] = self._savesyspath
+            self._savesyspath = None
+
+        if self._cwd is not None:
+            os.chdir(self._cwd)
+            self._cwd = None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/nodes.py b/backend/venv/lib/python3.12/site-packages/_pytest/nodes.py
new file mode 100644
index 00000000..51bc5174
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/nodes.py
@@ -0,0 +1,766 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import abc
+from functools import cached_property
+from inspect import signature
+import os
+import pathlib
+from pathlib import Path
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Iterable
+from typing import Iterator
+from typing import MutableMapping
+from typing import NoReturn
+from typing import overload
+from typing import TYPE_CHECKING
+from typing import TypeVar
+import warnings
+
+import pluggy
+
+import _pytest._code
+from _pytest._code import getfslineno
+from _pytest._code.code import ExceptionInfo
+from _pytest._code.code import TerminalRepr
+from _pytest._code.code import Traceback
+from _pytest._code.code import TracebackStyle
+from _pytest.compat import LEGACY_PATH
+from _pytest.config import Config
+from _pytest.config import ConftestImportFailure
+from _pytest.config.compat import _check_path
+from _pytest.deprecated import NODE_CTOR_FSPATH_ARG
+from _pytest.mark.structures import Mark
+from _pytest.mark.structures import MarkDecorator
+from _pytest.mark.structures import NodeKeywords
+from _pytest.outcomes import fail
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import commonpath
+from _pytest.stash import Stash
+from _pytest.warning_types import PytestWarning
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    # Imported here due to circular import.
+    from _pytest.main import Session
+
+
+SEP = "/"
+
+tracebackcutdir = Path(_pytest.__file__).parent
+
+
+_T = TypeVar("_T")
+
+
+def _imply_path(
+    node_type: type[Node],
+    path: Path | None,
+    fspath: LEGACY_PATH | None,
+) -> Path:
+    if fspath is not None:
+        warnings.warn(
+            NODE_CTOR_FSPATH_ARG.format(
+                node_type_name=node_type.__name__,
+            ),
+            stacklevel=6,
+        )
+    if path is not None:
+        if fspath is not None:
+            _check_path(path, fspath)
+        return path
+    else:
+        assert fspath is not None
+        return Path(fspath)
+
+
+_NodeType = TypeVar("_NodeType", bound="Node")
+
+
+class NodeMeta(abc.ABCMeta):
+    """Metaclass used by :class:`Node` to enforce that direct construction raises
+    :class:`Failed`.
+
+    This behaviour supports the indirection introduced with :meth:`Node.from_parent`,
+    the named constructor to be used instead of direct construction. The design
+    decision to enforce indirection with :class:`NodeMeta` was made as a
+    temporary aid for refactoring the collection tree, which was diagnosed to
+    have :class:`Node` objects whose creational patterns were overly entangled.
+    Once the refactoring is complete, this metaclass can be removed.
+
+    See https://github.com/pytest-dev/pytest/projects/3 for an overview of the
+    progress on detangling the :class:`Node` classes.
+    """
+
+    def __call__(cls, *k, **kw) -> NoReturn:
+        msg = (
+            "Direct construction of {name} has been deprecated, please use {name}.from_parent.\n"
+            "See "
+            "https://docs.pytest.org/en/stable/deprecations.html#node-construction-changed-to-node-from-parent"
+            " for more details."
+        ).format(name=f"{cls.__module__}.{cls.__name__}")
+        fail(msg, pytrace=False)
+
+    def _create(cls: type[_T], *k, **kw) -> _T:
+        try:
+            return super().__call__(*k, **kw)  # type: ignore[no-any-return,misc]
+        except TypeError:
+            sig = signature(getattr(cls, "__init__"))
+            known_kw = {k: v for k, v in kw.items() if k in sig.parameters}
+            from .warning_types import PytestDeprecationWarning
+
+            warnings.warn(
+                PytestDeprecationWarning(
+                    f"{cls} is not using a cooperative constructor and only takes {set(known_kw)}.\n"
+                    "See https://docs.pytest.org/en/stable/deprecations.html"
+                    "#constructors-of-custom-pytest-node-subclasses-should-take-kwargs "
+                    "for more details."
+                )
+            )
+
+            return super().__call__(*k, **known_kw)  # type: ignore[no-any-return,misc]
+
+
+class Node(abc.ABC, metaclass=NodeMeta):
+    r"""Base class of :class:`Collector` and :class:`Item`, the components of
+    the test collection tree.
+
+    ``Collector``\'s are the internal nodes of the tree, and ``Item``\'s are the
+    leaf nodes.
+    """
+
+    # Implemented in the legacypath plugin.
+    #: A ``LEGACY_PATH`` copy of the :attr:`path` attribute. Intended for usage
+    #: for methods not migrated to ``pathlib.Path`` yet, such as
+    #: :meth:`Item.reportinfo <pytest.Item.reportinfo>`. Will be deprecated in
+    #: a future release, prefer using :attr:`path` instead.
+    fspath: LEGACY_PATH
+
+    # Use __slots__ to make attribute access faster.
+    # Note that __dict__ is still available.
+    __slots__ = (
+        "name",
+        "parent",
+        "config",
+        "session",
+        "path",
+        "_nodeid",
+        "_store",
+        "__dict__",
+    )
+
+    def __init__(
+        self,
+        name: str,
+        parent: Node | None = None,
+        config: Config | None = None,
+        session: Session | None = None,
+        fspath: LEGACY_PATH | None = None,
+        path: Path | None = None,
+        nodeid: str | None = None,
+    ) -> None:
+        #: A unique name within the scope of the parent node.
+        self.name: str = name
+
+        #: The parent collector node.
+        self.parent = parent
+
+        if config:
+            #: The pytest config object.
+            self.config: Config = config
+        else:
+            if not parent:
+                raise TypeError("config or parent must be provided")
+            self.config = parent.config
+
+        if session:
+            #: The pytest session this node is part of.
+            self.session: Session = session
+        else:
+            if not parent:
+                raise TypeError("session or parent must be provided")
+            self.session = parent.session
+
+        if path is None and fspath is None:
+            path = getattr(parent, "path", None)
+        #: Filesystem path where this node was collected from (can be None).
+        self.path: pathlib.Path = _imply_path(type(self), path, fspath=fspath)
+
+        # The explicit annotation is to avoid publicly exposing NodeKeywords.
+        #: Keywords/markers collected from all scopes.
+        self.keywords: MutableMapping[str, Any] = NodeKeywords(self)
+
+        #: The marker objects belonging to this node.
+        self.own_markers: list[Mark] = []
+
+        #: Allow adding of extra keywords to use for matching.
+        self.extra_keyword_matches: set[str] = set()
+
+        if nodeid is not None:
+            assert "::()" not in nodeid
+            self._nodeid = nodeid
+        else:
+            if not self.parent:
+                raise TypeError("nodeid or parent must be provided")
+            self._nodeid = self.parent.nodeid + "::" + self.name
+
+        #: A place where plugins can store information on the node for their
+        #: own use.
+        self.stash: Stash = Stash()
+        # Deprecated alias. Was never public. Can be removed in a few releases.
+        self._store = self.stash
+
+    @classmethod
+    def from_parent(cls, parent: Node, **kw) -> Self:
+        """Public constructor for Nodes.
+
+        This indirection got introduced in order to enable removing
+        the fragile logic from the node constructors.
+
+        Subclasses can use ``super().from_parent(...)`` when overriding the
+        construction.
+
+        :param parent: The parent node of this Node.
+        """
+        if "config" in kw:
+            raise TypeError("config is not a valid argument for from_parent")
+        if "session" in kw:
+            raise TypeError("session is not a valid argument for from_parent")
+        return cls._create(parent=parent, **kw)
+
+    @property
+    def ihook(self) -> pluggy.HookRelay:
+        """fspath-sensitive hook proxy used to call pytest hooks."""
+        return self.session.gethookproxy(self.path)
+
+    def __repr__(self) -> str:
+        return "<{} {}>".format(self.__class__.__name__, getattr(self, "name", None))
+
+    def warn(self, warning: Warning) -> None:
+        """Issue a warning for this Node.
+
+        Warnings will be displayed after the test session, unless explicitly suppressed.
+
+        :param Warning warning:
+            The warning instance to issue.
+
+        :raises ValueError: If ``warning`` instance is not a subclass of Warning.
+
+        Example usage:
+
+        .. code-block:: python
+
+            node.warn(PytestWarning("some message"))
+            node.warn(UserWarning("some message"))
+
+        .. versionchanged:: 6.2
+            Any subclass of :class:`Warning` is now accepted, rather than only
+            :class:`PytestWarning <pytest.PytestWarning>` subclasses.
+        """
+        # enforce type checks here to avoid getting a generic type error later otherwise.
+        if not isinstance(warning, Warning):
+            raise ValueError(
+                f"warning must be an instance of Warning or subclass, got {warning!r}"
+            )
+        path, lineno = get_fslocation_from_item(self)
+        assert lineno is not None
+        warnings.warn_explicit(
+            warning,
+            category=None,
+            filename=str(path),
+            lineno=lineno + 1,
+        )
+
+    # Methods for ordering nodes.
+
+    @property
+    def nodeid(self) -> str:
+        """A ::-separated string denoting its collection tree address."""
+        return self._nodeid
+
+    def __hash__(self) -> int:
+        return hash(self._nodeid)
+
+    def setup(self) -> None:
+        pass
+
+    def teardown(self) -> None:
+        pass
+
+    def iter_parents(self) -> Iterator[Node]:
+        """Iterate over all parent collectors starting from and including self
+        up to the root of the collection tree.
+
+        .. versionadded:: 8.1
+        """
+        parent: Node | None = self
+        while parent is not None:
+            yield parent
+            parent = parent.parent
+
+    def listchain(self) -> list[Node]:
+        """Return a list of all parent collectors starting from the root of the
+        collection tree down to and including self."""
+        chain = []
+        item: Node | None = self
+        while item is not None:
+            chain.append(item)
+            item = item.parent
+        chain.reverse()
+        return chain
+
+    def add_marker(self, marker: str | MarkDecorator, append: bool = True) -> None:
+        """Dynamically add a marker object to the node.
+
+        :param marker:
+            The marker.
+        :param append:
+            Whether to append the marker, or prepend it.
+        """
+        from _pytest.mark import MARK_GEN
+
+        if isinstance(marker, MarkDecorator):
+            marker_ = marker
+        elif isinstance(marker, str):
+            marker_ = getattr(MARK_GEN, marker)
+        else:
+            raise ValueError("is not a string or pytest.mark.* Marker")
+        self.keywords[marker_.name] = marker_
+        if append:
+            self.own_markers.append(marker_.mark)
+        else:
+            self.own_markers.insert(0, marker_.mark)
+
+    def iter_markers(self, name: str | None = None) -> Iterator[Mark]:
+        """Iterate over all markers of the node.
+
+        :param name: If given, filter the results by the name attribute.
+        :returns: An iterator of the markers of the node.
+        """
+        return (x[1] for x in self.iter_markers_with_node(name=name))
+
+    def iter_markers_with_node(
+        self, name: str | None = None
+    ) -> Iterator[tuple[Node, Mark]]:
+        """Iterate over all markers of the node.
+
+        :param name: If given, filter the results by the name attribute.
+        :returns: An iterator of (node, mark) tuples.
+        """
+        for node in self.iter_parents():
+            for mark in node.own_markers:
+                if name is None or getattr(mark, "name", None) == name:
+                    yield node, mark
+
+    @overload
+    def get_closest_marker(self, name: str) -> Mark | None: ...
+
+    @overload
+    def get_closest_marker(self, name: str, default: Mark) -> Mark: ...
+
+    def get_closest_marker(self, name: str, default: Mark | None = None) -> Mark | None:
+        """Return the first marker matching the name, from closest (for
+        example function) to farther level (for example module level).
+
+        :param default: Fallback return value if no marker was found.
+        :param name: Name to filter by.
+        """
+        return next(self.iter_markers(name=name), default)
+
+    def listextrakeywords(self) -> set[str]:
+        """Return a set of all extra keywords in self and any parents."""
+        extra_keywords: set[str] = set()
+        for item in self.listchain():
+            extra_keywords.update(item.extra_keyword_matches)
+        return extra_keywords
+
+    def listnames(self) -> list[str]:
+        return [x.name for x in self.listchain()]
+
+    def addfinalizer(self, fin: Callable[[], object]) -> None:
+        """Register a function to be called without arguments when this node is
+        finalized.
+
+        This method can only be called when this node is active
+        in a setup chain, for example during self.setup().
+        """
+        self.session._setupstate.addfinalizer(fin, self)
+
+    def getparent(self, cls: type[_NodeType]) -> _NodeType | None:
+        """Get the closest parent node (including self) which is an instance of
+        the given class.
+
+        :param cls: The node class to search for.
+        :returns: The node, if found.
+        """
+        for node in self.iter_parents():
+            if isinstance(node, cls):
+                return node
+        return None
+
+    def _traceback_filter(self, excinfo: ExceptionInfo[BaseException]) -> Traceback:
+        return excinfo.traceback
+
+    def _repr_failure_py(
+        self,
+        excinfo: ExceptionInfo[BaseException],
+        style: TracebackStyle | None = None,
+    ) -> TerminalRepr:
+        from _pytest.fixtures import FixtureLookupError
+
+        if isinstance(excinfo.value, ConftestImportFailure):
+            excinfo = ExceptionInfo.from_exception(excinfo.value.cause)
+        if isinstance(excinfo.value, fail.Exception):
+            if not excinfo.value.pytrace:
+                style = "value"
+        if isinstance(excinfo.value, FixtureLookupError):
+            return excinfo.value.formatrepr()
+
+        tbfilter: bool | Callable[[ExceptionInfo[BaseException]], Traceback]
+        if self.config.getoption("fulltrace", False):
+            style = "long"
+            tbfilter = False
+        else:
+            tbfilter = self._traceback_filter
+            if style == "auto":
+                style = "long"
+        # XXX should excinfo.getrepr record all data and toterminal() process it?
+        if style is None:
+            if self.config.getoption("tbstyle", "auto") == "short":
+                style = "short"
+            else:
+                style = "long"
+
+        if self.config.get_verbosity() > 1:
+            truncate_locals = False
+        else:
+            truncate_locals = True
+
+        truncate_args = False if self.config.get_verbosity() > 2 else True
+
+        # excinfo.getrepr() formats paths relative to the CWD if `abspath` is False.
+        # It is possible for a fixture/test to change the CWD while this code runs, which
+        # would then result in the user seeing confusing paths in the failure message.
+        # To fix this, if the CWD changed, always display the full absolute path.
+        # It will be better to just always display paths relative to invocation_dir, but
+        # this requires a lot of plumbing (#6428).
+        try:
+            abspath = Path(os.getcwd()) != self.config.invocation_params.dir
+        except OSError:
+            abspath = True
+
+        return excinfo.getrepr(
+            funcargs=True,
+            abspath=abspath,
+            showlocals=self.config.getoption("showlocals", False),
+            style=style,
+            tbfilter=tbfilter,
+            truncate_locals=truncate_locals,
+            truncate_args=truncate_args,
+        )
+
+    def repr_failure(
+        self,
+        excinfo: ExceptionInfo[BaseException],
+        style: TracebackStyle | None = None,
+    ) -> str | TerminalRepr:
+        """Return a representation of a collection or test failure.
+
+        .. seealso:: :ref:`non-python tests`
+
+        :param excinfo: Exception information for the failure.
+        """
+        return self._repr_failure_py(excinfo, style)
+
+
+def get_fslocation_from_item(node: Node) -> tuple[str | Path, int | None]:
+    """Try to extract the actual location from a node, depending on available attributes:
+
+    * "location": a pair (path, lineno)
+    * "obj": a Python object that the node wraps.
+    * "path": just a path
+
+    :rtype: A tuple of (str|Path, int) with filename and 0-based line number.
+    """
+    # See Item.location.
+    location: tuple[str, int | None, str] | None = getattr(node, "location", None)
+    if location is not None:
+        return location[:2]
+    obj = getattr(node, "obj", None)
+    if obj is not None:
+        return getfslineno(obj)
+    return getattr(node, "path", "unknown location"), -1
+
+
+class Collector(Node, abc.ABC):
+    """Base class of all collectors.
+
+    Collector create children through `collect()` and thus iteratively build
+    the collection tree.
+    """
+
+    class CollectError(Exception):
+        """An error during collection, contains a custom message."""
+
+    @abc.abstractmethod
+    def collect(self) -> Iterable[Item | Collector]:
+        """Collect children (items and collectors) for this collector."""
+        raise NotImplementedError("abstract")
+
+    # TODO: This omits the style= parameter which breaks Liskov Substitution.
+    def repr_failure(  # type: ignore[override]
+        self, excinfo: ExceptionInfo[BaseException]
+    ) -> str | TerminalRepr:
+        """Return a representation of a collection failure.
+
+        :param excinfo: Exception information for the failure.
+        """
+        if isinstance(excinfo.value, self.CollectError) and not self.config.getoption(
+            "fulltrace", False
+        ):
+            exc = excinfo.value
+            return str(exc.args[0])
+
+        # Respect explicit tbstyle option, but default to "short"
+        # (_repr_failure_py uses "long" with "fulltrace" option always).
+        tbstyle = self.config.getoption("tbstyle", "auto")
+        if tbstyle == "auto":
+            tbstyle = "short"
+
+        return self._repr_failure_py(excinfo, style=tbstyle)
+
+    def _traceback_filter(self, excinfo: ExceptionInfo[BaseException]) -> Traceback:
+        if hasattr(self, "path"):
+            traceback = excinfo.traceback
+            ntraceback = traceback.cut(path=self.path)
+            if ntraceback == traceback:
+                ntraceback = ntraceback.cut(excludepath=tracebackcutdir)
+            return ntraceback.filter(excinfo)
+        return excinfo.traceback
+
+
+def _check_initialpaths_for_relpath(session: Session, path: Path) -> str | None:
+    for initial_path in session._initialpaths:
+        if commonpath(path, initial_path) == initial_path:
+            rel = str(path.relative_to(initial_path))
+            return "" if rel == "." else rel
+    return None
+
+
+class FSCollector(Collector, abc.ABC):
+    """Base class for filesystem collectors."""
+
+    def __init__(
+        self,
+        fspath: LEGACY_PATH | None = None,
+        path_or_parent: Path | Node | None = None,
+        path: Path | None = None,
+        name: str | None = None,
+        parent: Node | None = None,
+        config: Config | None = None,
+        session: Session | None = None,
+        nodeid: str | None = None,
+    ) -> None:
+        if path_or_parent:
+            if isinstance(path_or_parent, Node):
+                assert parent is None
+                parent = cast(FSCollector, path_or_parent)
+            elif isinstance(path_or_parent, Path):
+                assert path is None
+                path = path_or_parent
+
+        path = _imply_path(type(self), path, fspath=fspath)
+        if name is None:
+            name = path.name
+            if parent is not None and parent.path != path:
+                try:
+                    rel = path.relative_to(parent.path)
+                except ValueError:
+                    pass
+                else:
+                    name = str(rel)
+                name = name.replace(os.sep, SEP)
+        self.path = path
+
+        if session is None:
+            assert parent is not None
+            session = parent.session
+
+        if nodeid is None:
+            try:
+                nodeid = str(self.path.relative_to(session.config.rootpath))
+            except ValueError:
+                nodeid = _check_initialpaths_for_relpath(session, path)
+
+            if nodeid and os.sep != SEP:
+                nodeid = nodeid.replace(os.sep, SEP)
+
+        super().__init__(
+            name=name,
+            parent=parent,
+            config=config,
+            session=session,
+            nodeid=nodeid,
+            path=path,
+        )
+
+    @classmethod
+    def from_parent(
+        cls,
+        parent,
+        *,
+        fspath: LEGACY_PATH | None = None,
+        path: Path | None = None,
+        **kw,
+    ) -> Self:
+        """The public constructor."""
+        return super().from_parent(parent=parent, fspath=fspath, path=path, **kw)
+
+
+class File(FSCollector, abc.ABC):
+    """Base class for collecting tests from a file.
+
+    :ref:`non-python tests`.
+    """
+
+
+class Directory(FSCollector, abc.ABC):
+    """Base class for collecting files from a directory.
+
+    A basic directory collector does the following: goes over the files and
+    sub-directories in the directory and creates collectors for them by calling
+    the hooks :hook:`pytest_collect_directory` and :hook:`pytest_collect_file`,
+    after checking that they are not ignored using
+    :hook:`pytest_ignore_collect`.
+
+    The default directory collectors are :class:`~pytest.Dir` and
+    :class:`~pytest.Package`.
+
+    .. versionadded:: 8.0
+
+    :ref:`custom directory collectors`.
+    """
+
+
+class Item(Node, abc.ABC):
+    """Base class of all test invocation items.
+
+    Note that for a single function there might be multiple test invocation items.
+    """
+
+    nextitem = None
+
+    def __init__(
+        self,
+        name,
+        parent=None,
+        config: Config | None = None,
+        session: Session | None = None,
+        nodeid: str | None = None,
+        **kw,
+    ) -> None:
+        # The first two arguments are intentionally passed positionally,
+        # to keep plugins who define a node type which inherits from
+        # (pytest.Item, pytest.File) working (see issue #8435).
+        # They can be made kwargs when the deprecation above is done.
+        super().__init__(
+            name,
+            parent,
+            config=config,
+            session=session,
+            nodeid=nodeid,
+            **kw,
+        )
+        self._report_sections: list[tuple[str, str, str]] = []
+
+        #: A list of tuples (name, value) that holds user defined properties
+        #: for this test.
+        self.user_properties: list[tuple[str, object]] = []
+
+        self._check_item_and_collector_diamond_inheritance()
+
+    def _check_item_and_collector_diamond_inheritance(self) -> None:
+        """
+        Check if the current type inherits from both File and Collector
+        at the same time, emitting a warning accordingly (#8447).
+        """
+        cls = type(self)
+
+        # We inject an attribute in the type to avoid issuing this warning
+        # for the same class more than once, which is not helpful.
+        # It is a hack, but was deemed acceptable in order to avoid
+        # flooding the user in the common case.
+        attr_name = "_pytest_diamond_inheritance_warning_shown"
+        if getattr(cls, attr_name, False):
+            return
+        setattr(cls, attr_name, True)
+
+        problems = ", ".join(
+            base.__name__ for base in cls.__bases__ if issubclass(base, Collector)
+        )
+        if problems:
+            warnings.warn(
+                f"{cls.__name__} is an Item subclass and should not be a collector, "
+                f"however its bases {problems} are collectors.\n"
+                "Please split the Collectors and the Item into separate node types.\n"
+                "Pytest Doc example: https://docs.pytest.org/en/latest/example/nonpython.html\n"
+                "example pull request on a plugin: https://github.com/asmeurer/pytest-flakes/pull/40/",
+                PytestWarning,
+            )
+
+    @abc.abstractmethod
+    def runtest(self) -> None:
+        """Run the test case for this item.
+
+        Must be implemented by subclasses.
+
+        .. seealso:: :ref:`non-python tests`
+        """
+        raise NotImplementedError("runtest must be implemented by Item subclass")
+
+    def add_report_section(self, when: str, key: str, content: str) -> None:
+        """Add a new report section, similar to what's done internally to add
+        stdout and stderr captured output::
+
+            item.add_report_section("call", "stdout", "report section contents")
+
+        :param str when:
+            One of the possible capture states, ``"setup"``, ``"call"``, ``"teardown"``.
+        :param str key:
+            Name of the section, can be customized at will. Pytest uses ``"stdout"`` and
+            ``"stderr"`` internally.
+        :param str content:
+            The full contents as a string.
+        """
+        if content:
+            self._report_sections.append((when, key, content))
+
+    def reportinfo(self) -> tuple[os.PathLike[str] | str, int | None, str]:
+        """Get location information for this item for test reports.
+
+        Returns a tuple with three elements:
+
+        - The path of the test (default ``self.path``)
+        - The 0-based line number of the test (default ``None``)
+        - A name of the test to be shown (default ``""``)
+
+        .. seealso:: :ref:`non-python tests`
+        """
+        return self.path, None, ""
+
+    @cached_property
+    def location(self) -> tuple[str, int | None, str]:
+        """
+        Returns a tuple of ``(relfspath, lineno, testname)`` for this item
+        where ``relfspath`` is file path relative to ``config.rootpath``
+        and lineno is a 0-based line number.
+        """
+        location = self.reportinfo()
+        path = absolutepath(location[0])
+        relfspath = self.session._node_location_to_relpath(path)
+        assert type(location[2]) is str
+        return (relfspath, location[1], location[2])
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/outcomes.py b/backend/venv/lib/python3.12/site-packages/_pytest/outcomes.py
new file mode 100644
index 00000000..5b20803e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/outcomes.py
@@ -0,0 +1,318 @@
+"""Exception classes and constants handling test outcomes as well as
+functions creating them."""
+
+from __future__ import annotations
+
+import sys
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import NoReturn
+from typing import Protocol
+from typing import Type
+from typing import TypeVar
+
+from .warning_types import PytestDeprecationWarning
+
+
+class OutcomeException(BaseException):
+    """OutcomeException and its subclass instances indicate and contain info
+    about test and collection outcomes."""
+
+    def __init__(self, msg: str | None = None, pytrace: bool = True) -> None:
+        if msg is not None and not isinstance(msg, str):
+            error_msg = (  # type: ignore[unreachable]
+                "{} expected string as 'msg' parameter, got '{}' instead.\n"
+                "Perhaps you meant to use a mark?"
+            )
+            raise TypeError(error_msg.format(type(self).__name__, type(msg).__name__))
+        super().__init__(msg)
+        self.msg = msg
+        self.pytrace = pytrace
+
+    def __repr__(self) -> str:
+        if self.msg is not None:
+            return self.msg
+        return f"<{self.__class__.__name__} instance>"
+
+    __str__ = __repr__
+
+
+TEST_OUTCOME = (OutcomeException, Exception)
+
+
+class Skipped(OutcomeException):
+    # XXX hackish: on 3k we fake to live in the builtins
+    # in order to have Skipped exception printing shorter/nicer
+    __module__ = "builtins"
+
+    def __init__(
+        self,
+        msg: str | None = None,
+        pytrace: bool = True,
+        allow_module_level: bool = False,
+        *,
+        _use_item_location: bool = False,
+    ) -> None:
+        super().__init__(msg=msg, pytrace=pytrace)
+        self.allow_module_level = allow_module_level
+        # If true, the skip location is reported as the item's location,
+        # instead of the place that raises the exception/calls skip().
+        self._use_item_location = _use_item_location
+
+
+class Failed(OutcomeException):
+    """Raised from an explicit call to pytest.fail()."""
+
+    __module__ = "builtins"
+
+
+class Exit(Exception):
+    """Raised for immediate program exits (no tracebacks/summaries)."""
+
+    def __init__(
+        self, msg: str = "unknown reason", returncode: int | None = None
+    ) -> None:
+        self.msg = msg
+        self.returncode = returncode
+        super().__init__(msg)
+
+
+# Elaborate hack to work around https://github.com/python/mypy/issues/2087.
+# Ideally would just be `exit.Exception = Exit` etc.
+
+_F = TypeVar("_F", bound=Callable[..., object])
+_ET = TypeVar("_ET", bound=Type[BaseException])
+
+
+class _WithException(Protocol[_F, _ET]):
+    Exception: _ET
+    __call__: _F
+
+
+def _with_exception(exception_type: _ET) -> Callable[[_F], _WithException[_F, _ET]]:
+    def decorate(func: _F) -> _WithException[_F, _ET]:
+        func_with_exception = cast(_WithException[_F, _ET], func)
+        func_with_exception.Exception = exception_type
+        return func_with_exception
+
+    return decorate
+
+
+# Exposed helper methods.
+
+
+@_with_exception(Exit)
+def exit(
+    reason: str = "",
+    returncode: int | None = None,
+) -> NoReturn:
+    """Exit testing process.
+
+    :param reason:
+        The message to show as the reason for exiting pytest.  reason has a default value
+        only because `msg` is deprecated.
+
+    :param returncode:
+        Return code to be used when exiting pytest. None means the same as ``0`` (no error), same as :func:`sys.exit`.
+
+    :raises pytest.exit.Exception:
+        The exception that is raised.
+    """
+    __tracebackhide__ = True
+    raise Exit(reason, returncode)
+
+
+@_with_exception(Skipped)
+def skip(
+    reason: str = "",
+    *,
+    allow_module_level: bool = False,
+) -> NoReturn:
+    """Skip an executing test with the given message.
+
+    This function should be called only during testing (setup, call or teardown) or
+    during collection by using the ``allow_module_level`` flag.  This function can
+    be called in doctests as well.
+
+    :param reason:
+        The message to show the user as reason for the skip.
+
+    :param allow_module_level:
+        Allows this function to be called at module level.
+        Raising the skip exception at module level will stop
+        the execution of the module and prevent the collection of all tests in the module,
+        even those defined before the `skip` call.
+
+        Defaults to False.
+
+    :raises pytest.skip.Exception:
+        The exception that is raised.
+
+    .. note::
+        It is better to use the :ref:`pytest.mark.skipif ref` marker when
+        possible to declare a test to be skipped under certain conditions
+        like mismatching platforms or dependencies.
+        Similarly, use the ``# doctest: +SKIP`` directive (see :py:data:`doctest.SKIP`)
+        to skip a doctest statically.
+    """
+    __tracebackhide__ = True
+    raise Skipped(msg=reason, allow_module_level=allow_module_level)
+
+
+@_with_exception(Failed)
+def fail(reason: str = "", pytrace: bool = True) -> NoReturn:
+    """Explicitly fail an executing test with the given message.
+
+    :param reason:
+        The message to show the user as reason for the failure.
+
+    :param pytrace:
+        If False, msg represents the full failure information and no
+        python traceback will be reported.
+
+    :raises pytest.fail.Exception:
+        The exception that is raised.
+    """
+    __tracebackhide__ = True
+    raise Failed(msg=reason, pytrace=pytrace)
+
+
+class XFailed(Failed):
+    """Raised from an explicit call to pytest.xfail()."""
+
+
+@_with_exception(XFailed)
+def xfail(reason: str = "") -> NoReturn:
+    """Imperatively xfail an executing test or setup function with the given reason.
+
+    This function should be called only during testing (setup, call or teardown).
+
+    No other code is executed after using ``xfail()`` (it is implemented
+    internally by raising an exception).
+
+    :param reason:
+        The message to show the user as reason for the xfail.
+
+    .. note::
+        It is better to use the :ref:`pytest.mark.xfail ref` marker when
+        possible to declare a test to be xfailed under certain conditions
+        like known bugs or missing features.
+
+    :raises pytest.xfail.Exception:
+        The exception that is raised.
+    """
+    __tracebackhide__ = True
+    raise XFailed(reason)
+
+
+def importorskip(
+    modname: str,
+    minversion: str | None = None,
+    reason: str | None = None,
+    *,
+    exc_type: type[ImportError] | None = None,
+) -> Any:
+    """Import and return the requested module ``modname``, or skip the
+    current test if the module cannot be imported.
+
+    :param modname:
+        The name of the module to import.
+    :param minversion:
+        If given, the imported module's ``__version__`` attribute must be at
+        least this minimal version, otherwise the test is still skipped.
+    :param reason:
+        If given, this reason is shown as the message when the module cannot
+        be imported.
+    :param exc_type:
+        The exception that should be captured in order to skip modules.
+        Must be :py:class:`ImportError` or a subclass.
+
+        If the module can be imported but raises :class:`ImportError`, pytest will
+        issue a warning to the user, as often users expect the module not to be
+        found (which would raise :class:`ModuleNotFoundError` instead).
+
+        This warning can be suppressed by passing ``exc_type=ImportError`` explicitly.
+
+        See :ref:`import-or-skip-import-error` for details.
+
+
+    :returns:
+        The imported module. This should be assigned to its canonical name.
+
+    :raises pytest.skip.Exception:
+        If the module cannot be imported.
+
+    Example::
+
+        docutils = pytest.importorskip("docutils")
+
+    .. versionadded:: 8.2
+
+        The ``exc_type`` parameter.
+    """
+    import warnings
+
+    __tracebackhide__ = True
+    compile(modname, "", "eval")  # to catch syntaxerrors
+
+    # Until pytest 9.1, we will warn the user if we catch ImportError (instead of ModuleNotFoundError),
+    # as this might be hiding an installation/environment problem, which is not usually what is intended
+    # when using importorskip() (#11523).
+    # In 9.1, to keep the function signature compatible, we just change the code below to:
+    # 1. Use `exc_type = ModuleNotFoundError` if `exc_type` is not given.
+    # 2. Remove `warn_on_import` and the warning handling.
+    if exc_type is None:
+        exc_type = ImportError
+        warn_on_import_error = True
+    else:
+        warn_on_import_error = False
+
+    skipped: Skipped | None = None
+    warning: Warning | None = None
+
+    with warnings.catch_warnings():
+        # Make sure to ignore ImportWarnings that might happen because
+        # of existing directories with the same name we're trying to
+        # import but without a __init__.py file.
+        warnings.simplefilter("ignore")
+
+        try:
+            __import__(modname)
+        except exc_type as exc:
+            # Do not raise or issue warnings inside the catch_warnings() block.
+            if reason is None:
+                reason = f"could not import {modname!r}: {exc}"
+            skipped = Skipped(reason, allow_module_level=True)
+
+            if warn_on_import_error and not isinstance(exc, ModuleNotFoundError):
+                lines = [
+                    "",
+                    f"Module '{modname}' was found, but when imported by pytest it raised:",
+                    f"    {exc!r}",
+                    "In pytest 9.1 this warning will become an error by default.",
+                    "You can fix the underlying problem, or alternatively overwrite this behavior and silence this "
+                    "warning by passing exc_type=ImportError explicitly.",
+                    "See https://docs.pytest.org/en/stable/deprecations.html#pytest-importorskip-default-behavior-regarding-importerror",
+                ]
+                warning = PytestDeprecationWarning("\n".join(lines))
+
+    if warning:
+        warnings.warn(warning, stacklevel=2)
+    if skipped:
+        raise skipped
+
+    mod = sys.modules[modname]
+    if minversion is None:
+        return mod
+    verattr = getattr(mod, "__version__", None)
+    if minversion is not None:
+        # Imported lazily to improve start-up time.
+        from packaging.version import Version
+
+        if verattr is None or Version(verattr) < Version(minversion):
+            raise Skipped(
+                f"module {modname!r} has __version__ {verattr!r}, required is: {minversion!r}",
+                allow_module_level=True,
+            )
+    return mod
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/pastebin.py b/backend/venv/lib/python3.12/site-packages/_pytest/pastebin.py
new file mode 100644
index 00000000..69c011ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/pastebin.py
@@ -0,0 +1,113 @@
+# mypy: allow-untyped-defs
+"""Submit failure or test session information to a pastebin service."""
+
+from __future__ import annotations
+
+from io import StringIO
+import tempfile
+from typing import IO
+
+from _pytest.config import Config
+from _pytest.config import create_terminal_writer
+from _pytest.config.argparsing import Parser
+from _pytest.stash import StashKey
+from _pytest.terminal import TerminalReporter
+import pytest
+
+
+pastebinfile_key = StashKey[IO[bytes]]()
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("terminal reporting")
+    group._addoption(
+        "--pastebin",
+        metavar="mode",
+        action="store",
+        dest="pastebin",
+        default=None,
+        choices=["failed", "all"],
+        help="Send failed|all info to bpaste.net pastebin service",
+    )
+
+
+@pytest.hookimpl(trylast=True)
+def pytest_configure(config: Config) -> None:
+    if config.option.pastebin == "all":
+        tr = config.pluginmanager.getplugin("terminalreporter")
+        # If no terminal reporter plugin is present, nothing we can do here;
+        # this can happen when this function executes in a worker node
+        # when using pytest-xdist, for example.
+        if tr is not None:
+            # pastebin file will be UTF-8 encoded binary file.
+            config.stash[pastebinfile_key] = tempfile.TemporaryFile("w+b")
+            oldwrite = tr._tw.write
+
+            def tee_write(s, **kwargs):
+                oldwrite(s, **kwargs)
+                if isinstance(s, str):
+                    s = s.encode("utf-8")
+                config.stash[pastebinfile_key].write(s)
+
+            tr._tw.write = tee_write
+
+
+def pytest_unconfigure(config: Config) -> None:
+    if pastebinfile_key in config.stash:
+        pastebinfile = config.stash[pastebinfile_key]
+        # Get terminal contents and delete file.
+        pastebinfile.seek(0)
+        sessionlog = pastebinfile.read()
+        pastebinfile.close()
+        del config.stash[pastebinfile_key]
+        # Undo our patching in the terminal reporter.
+        tr = config.pluginmanager.getplugin("terminalreporter")
+        del tr._tw.__dict__["write"]
+        # Write summary.
+        tr.write_sep("=", "Sending information to Paste Service")
+        pastebinurl = create_new_paste(sessionlog)
+        tr.write_line(f"pastebin session-log: {pastebinurl}\n")
+
+
+def create_new_paste(contents: str | bytes) -> str:
+    """Create a new paste using the bpaste.net service.
+
+    :contents: Paste contents string.
+    :returns: URL to the pasted contents, or an error message.
+    """
+    import re
+    from urllib.parse import urlencode
+    from urllib.request import urlopen
+
+    params = {"code": contents, "lexer": "text", "expiry": "1week"}
+    url = "https://bpa.st"
+    try:
+        response: str = (
+            urlopen(url, data=urlencode(params).encode("ascii")).read().decode("utf-8")
+        )
+    except OSError as exc_info:  # urllib errors
+        return f"bad response: {exc_info}"
+    m = re.search(r'href="/raw/(\w+)"', response)
+    if m:
+        return f"{url}/show/{m.group(1)}"
+    else:
+        return "bad response: invalid format ('" + response + "')"
+
+
+def pytest_terminal_summary(terminalreporter: TerminalReporter) -> None:
+    if terminalreporter.config.option.pastebin != "failed":
+        return
+    if "failed" in terminalreporter.stats:
+        terminalreporter.write_sep("=", "Sending information to Paste Service")
+        for rep in terminalreporter.stats["failed"]:
+            try:
+                msg = rep.longrepr.reprtraceback.reprentries[-1].reprfileloc
+            except AttributeError:
+                msg = terminalreporter._getfailureheadline(rep)
+            file = StringIO()
+            tw = create_terminal_writer(terminalreporter.config, file)
+            rep.toterminal(tw)
+            s = file.getvalue()
+            assert len(s)
+            pastebinurl = create_new_paste(s)
+            terminalreporter.write_line(f"{msg} --> {pastebinurl}")
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/pathlib.py b/backend/venv/lib/python3.12/site-packages/_pytest/pathlib.py
new file mode 100644
index 00000000..dd36559c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/pathlib.py
@@ -0,0 +1,1038 @@
+from __future__ import annotations
+
+import atexit
+import contextlib
+from enum import Enum
+from errno import EBADF
+from errno import ELOOP
+from errno import ENOENT
+from errno import ENOTDIR
+import fnmatch
+from functools import partial
+from importlib.machinery import ModuleSpec
+from importlib.machinery import PathFinder
+import importlib.util
+import itertools
+import os
+from os.path import expanduser
+from os.path import expandvars
+from os.path import isabs
+from os.path import sep
+from pathlib import Path
+from pathlib import PurePath
+from posixpath import sep as posix_sep
+import shutil
+import sys
+import types
+from types import ModuleType
+from typing import Any
+from typing import Callable
+from typing import Iterable
+from typing import Iterator
+from typing import TypeVar
+import uuid
+import warnings
+
+from _pytest.compat import assert_never
+from _pytest.outcomes import skip
+from _pytest.warning_types import PytestWarning
+
+
+if sys.version_info < (3, 11):
+    from importlib._bootstrap_external import _NamespaceLoader as NamespaceLoader
+else:
+    from importlib.machinery import NamespaceLoader
+
+LOCK_TIMEOUT = 60 * 60 * 24 * 3
+
+_AnyPurePath = TypeVar("_AnyPurePath", bound=PurePath)
+
+# The following function, variables and comments were
+# copied from cpython 3.9 Lib/pathlib.py file.
+
+# EBADF - guard against macOS `stat` throwing EBADF
+_IGNORED_ERRORS = (ENOENT, ENOTDIR, EBADF, ELOOP)
+
+_IGNORED_WINERRORS = (
+    21,  # ERROR_NOT_READY - drive exists but is not accessible
+    1921,  # ERROR_CANT_RESOLVE_FILENAME - fix for broken symlink pointing to itself
+)
+
+
+def _ignore_error(exception: Exception) -> bool:
+    return (
+        getattr(exception, "errno", None) in _IGNORED_ERRORS
+        or getattr(exception, "winerror", None) in _IGNORED_WINERRORS
+    )
+
+
+def get_lock_path(path: _AnyPurePath) -> _AnyPurePath:
+    return path.joinpath(".lock")
+
+
+def on_rm_rf_error(
+    func: Callable[..., Any] | None,
+    path: str,
+    excinfo: BaseException
+    | tuple[type[BaseException], BaseException, types.TracebackType | None],
+    *,
+    start_path: Path,
+) -> bool:
+    """Handle known read-only errors during rmtree.
+
+    The returned value is used only by our own tests.
+    """
+    if isinstance(excinfo, BaseException):
+        exc = excinfo
+    else:
+        exc = excinfo[1]
+
+    # Another process removed the file in the middle of the "rm_rf" (xdist for example).
+    # More context: https://github.com/pytest-dev/pytest/issues/5974#issuecomment-543799018
+    if isinstance(exc, FileNotFoundError):
+        return False
+
+    if not isinstance(exc, PermissionError):
+        warnings.warn(
+            PytestWarning(f"(rm_rf) error removing {path}\n{type(exc)}: {exc}")
+        )
+        return False
+
+    if func not in (os.rmdir, os.remove, os.unlink):
+        if func not in (os.open,):
+            warnings.warn(
+                PytestWarning(
+                    f"(rm_rf) unknown function {func} when removing {path}:\n{type(exc)}: {exc}"
+                )
+            )
+        return False
+
+    # Chmod + retry.
+    import stat
+
+    def chmod_rw(p: str) -> None:
+        mode = os.stat(p).st_mode
+        os.chmod(p, mode | stat.S_IRUSR | stat.S_IWUSR)
+
+    # For files, we need to recursively go upwards in the directories to
+    # ensure they all are also writable.
+    p = Path(path)
+    if p.is_file():
+        for parent in p.parents:
+            chmod_rw(str(parent))
+            # Stop when we reach the original path passed to rm_rf.
+            if parent == start_path:
+                break
+    chmod_rw(str(path))
+
+    func(path)
+    return True
+
+
+def ensure_extended_length_path(path: Path) -> Path:
+    """Get the extended-length version of a path (Windows).
+
+    On Windows, by default, the maximum length of a path (MAX_PATH) is 260
+    characters, and operations on paths longer than that fail. But it is possible
+    to overcome this by converting the path to "extended-length" form before
+    performing the operation:
+    https://docs.microsoft.com/en-us/windows/win32/fileio/naming-a-file#maximum-path-length-limitation
+
+    On Windows, this function returns the extended-length absolute version of path.
+    On other platforms it returns path unchanged.
+    """
+    if sys.platform.startswith("win32"):
+        path = path.resolve()
+        path = Path(get_extended_length_path_str(str(path)))
+    return path
+
+
+def get_extended_length_path_str(path: str) -> str:
+    """Convert a path to a Windows extended length path."""
+    long_path_prefix = "\\\\?\\"
+    unc_long_path_prefix = "\\\\?\\UNC\\"
+    if path.startswith((long_path_prefix, unc_long_path_prefix)):
+        return path
+    # UNC
+    if path.startswith("\\\\"):
+        return unc_long_path_prefix + path[2:]
+    return long_path_prefix + path
+
+
+def rm_rf(path: Path) -> None:
+    """Remove the path contents recursively, even if some elements
+    are read-only."""
+    path = ensure_extended_length_path(path)
+    onerror = partial(on_rm_rf_error, start_path=path)
+    if sys.version_info >= (3, 12):
+        shutil.rmtree(str(path), onexc=onerror)
+    else:
+        shutil.rmtree(str(path), onerror=onerror)
+
+
+def find_prefixed(root: Path, prefix: str) -> Iterator[os.DirEntry[str]]:
+    """Find all elements in root that begin with the prefix, case-insensitive."""
+    l_prefix = prefix.lower()
+    for x in os.scandir(root):
+        if x.name.lower().startswith(l_prefix):
+            yield x
+
+
+def extract_suffixes(iter: Iterable[os.DirEntry[str]], prefix: str) -> Iterator[str]:
+    """Return the parts of the paths following the prefix.
+
+    :param iter: Iterator over path names.
+    :param prefix: Expected prefix of the path names.
+    """
+    p_len = len(prefix)
+    for entry in iter:
+        yield entry.name[p_len:]
+
+
+def find_suffixes(root: Path, prefix: str) -> Iterator[str]:
+    """Combine find_prefixes and extract_suffixes."""
+    return extract_suffixes(find_prefixed(root, prefix), prefix)
+
+
+def parse_num(maybe_num: str) -> int:
+    """Parse number path suffixes, returns -1 on error."""
+    try:
+        return int(maybe_num)
+    except ValueError:
+        return -1
+
+
+def _force_symlink(root: Path, target: str | PurePath, link_to: str | Path) -> None:
+    """Helper to create the current symlink.
+
+    It's full of race conditions that are reasonably OK to ignore
+    for the context of best effort linking to the latest test run.
+
+    The presumption being that in case of much parallelism
+    the inaccuracy is going to be acceptable.
+    """
+    current_symlink = root.joinpath(target)
+    try:
+        current_symlink.unlink()
+    except OSError:
+        pass
+    try:
+        current_symlink.symlink_to(link_to)
+    except Exception:
+        pass
+
+
+def make_numbered_dir(root: Path, prefix: str, mode: int = 0o700) -> Path:
+    """Create a directory with an increased number as suffix for the given prefix."""
+    for i in range(10):
+        # try up to 10 times to create the folder
+        max_existing = max(map(parse_num, find_suffixes(root, prefix)), default=-1)
+        new_number = max_existing + 1
+        new_path = root.joinpath(f"{prefix}{new_number}")
+        try:
+            new_path.mkdir(mode=mode)
+        except Exception:
+            pass
+        else:
+            _force_symlink(root, prefix + "current", new_path)
+            return new_path
+    else:
+        raise OSError(
+            "could not create numbered dir with prefix "
+            f"{prefix} in {root} after 10 tries"
+        )
+
+
+def create_cleanup_lock(p: Path) -> Path:
+    """Create a lock to prevent premature folder cleanup."""
+    lock_path = get_lock_path(p)
+    try:
+        fd = os.open(str(lock_path), os.O_WRONLY | os.O_CREAT | os.O_EXCL, 0o644)
+    except FileExistsError as e:
+        raise OSError(f"cannot create lockfile in {p}") from e
+    else:
+        pid = os.getpid()
+        spid = str(pid).encode()
+        os.write(fd, spid)
+        os.close(fd)
+        if not lock_path.is_file():
+            raise OSError("lock path got renamed after successful creation")
+        return lock_path
+
+
+def register_cleanup_lock_removal(
+    lock_path: Path, register: Any = atexit.register
+) -> Any:
+    """Register a cleanup function for removing a lock, by default on atexit."""
+    pid = os.getpid()
+
+    def cleanup_on_exit(lock_path: Path = lock_path, original_pid: int = pid) -> None:
+        current_pid = os.getpid()
+        if current_pid != original_pid:
+            # fork
+            return
+        try:
+            lock_path.unlink()
+        except OSError:
+            pass
+
+    return register(cleanup_on_exit)
+
+
+def maybe_delete_a_numbered_dir(path: Path) -> None:
+    """Remove a numbered directory if its lock can be obtained and it does
+    not seem to be in use."""
+    path = ensure_extended_length_path(path)
+    lock_path = None
+    try:
+        lock_path = create_cleanup_lock(path)
+        parent = path.parent
+
+        garbage = parent.joinpath(f"garbage-{uuid.uuid4()}")
+        path.rename(garbage)
+        rm_rf(garbage)
+    except OSError:
+        #  known races:
+        #  * other process did a cleanup at the same time
+        #  * deletable folder was found
+        #  * process cwd (Windows)
+        return
+    finally:
+        # If we created the lock, ensure we remove it even if we failed
+        # to properly remove the numbered dir.
+        if lock_path is not None:
+            try:
+                lock_path.unlink()
+            except OSError:
+                pass
+
+
+def ensure_deletable(path: Path, consider_lock_dead_if_created_before: float) -> bool:
+    """Check if `path` is deletable based on whether the lock file is expired."""
+    if path.is_symlink():
+        return False
+    lock = get_lock_path(path)
+    try:
+        if not lock.is_file():
+            return True
+    except OSError:
+        # we might not have access to the lock file at all, in this case assume
+        # we don't have access to the entire directory (#7491).
+        return False
+    try:
+        lock_time = lock.stat().st_mtime
+    except Exception:
+        return False
+    else:
+        if lock_time < consider_lock_dead_if_created_before:
+            # We want to ignore any errors while trying to remove the lock such as:
+            # - PermissionDenied, like the file permissions have changed since the lock creation;
+            # - FileNotFoundError, in case another pytest process got here first;
+            # and any other cause of failure.
+            with contextlib.suppress(OSError):
+                lock.unlink()
+                return True
+        return False
+
+
+def try_cleanup(path: Path, consider_lock_dead_if_created_before: float) -> None:
+    """Try to cleanup a folder if we can ensure it's deletable."""
+    if ensure_deletable(path, consider_lock_dead_if_created_before):
+        maybe_delete_a_numbered_dir(path)
+
+
+def cleanup_candidates(root: Path, prefix: str, keep: int) -> Iterator[Path]:
+    """List candidates for numbered directories to be removed - follows py.path."""
+    max_existing = max(map(parse_num, find_suffixes(root, prefix)), default=-1)
+    max_delete = max_existing - keep
+    entries = find_prefixed(root, prefix)
+    entries, entries2 = itertools.tee(entries)
+    numbers = map(parse_num, extract_suffixes(entries2, prefix))
+    for entry, number in zip(entries, numbers):
+        if number <= max_delete:
+            yield Path(entry)
+
+
+def cleanup_dead_symlinks(root: Path) -> None:
+    for left_dir in root.iterdir():
+        if left_dir.is_symlink():
+            if not left_dir.resolve().exists():
+                left_dir.unlink()
+
+
+def cleanup_numbered_dir(
+    root: Path, prefix: str, keep: int, consider_lock_dead_if_created_before: float
+) -> None:
+    """Cleanup for lock driven numbered directories."""
+    if not root.exists():
+        return
+    for path in cleanup_candidates(root, prefix, keep):
+        try_cleanup(path, consider_lock_dead_if_created_before)
+    for path in root.glob("garbage-*"):
+        try_cleanup(path, consider_lock_dead_if_created_before)
+
+    cleanup_dead_symlinks(root)
+
+
+def make_numbered_dir_with_cleanup(
+    root: Path,
+    prefix: str,
+    keep: int,
+    lock_timeout: float,
+    mode: int,
+) -> Path:
+    """Create a numbered dir with a cleanup lock and remove old ones."""
+    e = None
+    for i in range(10):
+        try:
+            p = make_numbered_dir(root, prefix, mode)
+            # Only lock the current dir when keep is not 0
+            if keep != 0:
+                lock_path = create_cleanup_lock(p)
+                register_cleanup_lock_removal(lock_path)
+        except Exception as exc:
+            e = exc
+        else:
+            consider_lock_dead_if_created_before = p.stat().st_mtime - lock_timeout
+            # Register a cleanup for program exit
+            atexit.register(
+                cleanup_numbered_dir,
+                root,
+                prefix,
+                keep,
+                consider_lock_dead_if_created_before,
+            )
+            return p
+    assert e is not None
+    raise e
+
+
+def resolve_from_str(input: str, rootpath: Path) -> Path:
+    input = expanduser(input)
+    input = expandvars(input)
+    if isabs(input):
+        return Path(input)
+    else:
+        return rootpath.joinpath(input)
+
+
+def fnmatch_ex(pattern: str, path: str | os.PathLike[str]) -> bool:
+    """A port of FNMatcher from py.path.common which works with PurePath() instances.
+
+    The difference between this algorithm and PurePath.match() is that the
+    latter matches "**" glob expressions for each part of the path, while
+    this algorithm uses the whole path instead.
+
+    For example:
+        "tests/foo/bar/doc/test_foo.py" matches pattern "tests/**/doc/test*.py"
+        with this algorithm, but not with PurePath.match().
+
+    This algorithm was ported to keep backward-compatibility with existing
+    settings which assume paths match according this logic.
+
+    References:
+    * https://bugs.python.org/issue29249
+    * https://bugs.python.org/issue34731
+    """
+    path = PurePath(path)
+    iswin32 = sys.platform.startswith("win")
+
+    if iswin32 and sep not in pattern and posix_sep in pattern:
+        # Running on Windows, the pattern has no Windows path separators,
+        # and the pattern has one or more Posix path separators. Replace
+        # the Posix path separators with the Windows path separator.
+        pattern = pattern.replace(posix_sep, sep)
+
+    if sep not in pattern:
+        name = path.name
+    else:
+        name = str(path)
+        if path.is_absolute() and not os.path.isabs(pattern):
+            pattern = f"*{os.sep}{pattern}"
+    return fnmatch.fnmatch(name, pattern)
+
+
+def parts(s: str) -> set[str]:
+    parts = s.split(sep)
+    return {sep.join(parts[: i + 1]) or sep for i in range(len(parts))}
+
+
+def symlink_or_skip(
+    src: os.PathLike[str] | str,
+    dst: os.PathLike[str] | str,
+    **kwargs: Any,
+) -> None:
+    """Make a symlink, or skip the test in case symlinks are not supported."""
+    try:
+        os.symlink(src, dst, **kwargs)
+    except OSError as e:
+        skip(f"symlinks not supported: {e}")
+
+
+class ImportMode(Enum):
+    """Possible values for `mode` parameter of `import_path`."""
+
+    prepend = "prepend"
+    append = "append"
+    importlib = "importlib"
+
+
+class ImportPathMismatchError(ImportError):
+    """Raised on import_path() if there is a mismatch of __file__'s.
+
+    This can happen when `import_path` is called multiple times with different filenames that has
+    the same basename but reside in packages
+    (for example "/tests1/test_foo.py" and "/tests2/test_foo.py").
+    """
+
+
+def import_path(
+    path: str | os.PathLike[str],
+    *,
+    mode: str | ImportMode = ImportMode.prepend,
+    root: Path,
+    consider_namespace_packages: bool,
+) -> ModuleType:
+    """
+    Import and return a module from the given path, which can be a file (a module) or
+    a directory (a package).
+
+    :param path:
+        Path to the file to import.
+
+    :param mode:
+        Controls the underlying import mechanism that will be used:
+
+        * ImportMode.prepend: the directory containing the module (or package, taking
+          `__init__.py` files into account) will be put at the *start* of `sys.path` before
+          being imported with `importlib.import_module`.
+
+        * ImportMode.append: same as `prepend`, but the directory will be appended
+          to the end of `sys.path`, if not already in `sys.path`.
+
+        * ImportMode.importlib: uses more fine control mechanisms provided by `importlib`
+          to import the module, which avoids having to muck with `sys.path` at all. It effectively
+          allows having same-named test modules in different places.
+
+    :param root:
+        Used as an anchor when mode == ImportMode.importlib to obtain
+        a unique name for the module being imported so it can safely be stored
+        into ``sys.modules``.
+
+    :param consider_namespace_packages:
+        If True, consider namespace packages when resolving module names.
+
+    :raises ImportPathMismatchError:
+        If after importing the given `path` and the module `__file__`
+        are different. Only raised in `prepend` and `append` modes.
+    """
+    path = Path(path)
+    mode = ImportMode(mode)
+
+    if not path.exists():
+        raise ImportError(path)
+
+    if mode is ImportMode.importlib:
+        # Try to import this module using the standard import mechanisms, but
+        # without touching sys.path.
+        try:
+            pkg_root, module_name = resolve_pkg_root_and_module_name(
+                path, consider_namespace_packages=consider_namespace_packages
+            )
+        except CouldNotResolvePathError:
+            pass
+        else:
+            # If the given module name is already in sys.modules, do not import it again.
+            with contextlib.suppress(KeyError):
+                return sys.modules[module_name]
+
+            mod = _import_module_using_spec(
+                module_name, path, pkg_root, insert_modules=False
+            )
+            if mod is not None:
+                return mod
+
+        # Could not import the module with the current sys.path, so we fall back
+        # to importing the file as a single module, not being a part of a package.
+        module_name = module_name_from_path(path, root)
+        with contextlib.suppress(KeyError):
+            return sys.modules[module_name]
+
+        mod = _import_module_using_spec(
+            module_name, path, path.parent, insert_modules=True
+        )
+        if mod is None:
+            raise ImportError(f"Can't find module {module_name} at location {path}")
+        return mod
+
+    try:
+        pkg_root, module_name = resolve_pkg_root_and_module_name(
+            path, consider_namespace_packages=consider_namespace_packages
+        )
+    except CouldNotResolvePathError:
+        pkg_root, module_name = path.parent, path.stem
+
+    # Change sys.path permanently: restoring it at the end of this function would cause surprising
+    # problems because of delayed imports: for example, a conftest.py file imported by this function
+    # might have local imports, which would fail at runtime if we restored sys.path.
+    if mode is ImportMode.append:
+        if str(pkg_root) not in sys.path:
+            sys.path.append(str(pkg_root))
+    elif mode is ImportMode.prepend:
+        if str(pkg_root) != sys.path[0]:
+            sys.path.insert(0, str(pkg_root))
+    else:
+        assert_never(mode)
+
+    importlib.import_module(module_name)
+
+    mod = sys.modules[module_name]
+    if path.name == "__init__.py":
+        return mod
+
+    ignore = os.environ.get("PY_IGNORE_IMPORTMISMATCH", "")
+    if ignore != "1":
+        module_file = mod.__file__
+        if module_file is None:
+            raise ImportPathMismatchError(module_name, module_file, path)
+
+        if module_file.endswith((".pyc", ".pyo")):
+            module_file = module_file[:-1]
+        if module_file.endswith(os.sep + "__init__.py"):
+            module_file = module_file[: -(len(os.sep + "__init__.py"))]
+
+        try:
+            is_same = _is_same(str(path), module_file)
+        except FileNotFoundError:
+            is_same = False
+
+        if not is_same:
+            raise ImportPathMismatchError(module_name, module_file, path)
+
+    return mod
+
+
+def _import_module_using_spec(
+    module_name: str, module_path: Path, module_location: Path, *, insert_modules: bool
+) -> ModuleType | None:
+    """
+    Tries to import a module by its canonical name, path, and its parent location.
+
+    :param module_name:
+        The expected module name, will become the key of `sys.modules`.
+
+    :param module_path:
+        The file path of the module, for example `/foo/bar/test_demo.py`.
+        If module is a package, pass the path to the  `__init__.py` of the package.
+        If module is a namespace package, pass directory path.
+
+    :param module_location:
+        The parent location of the module.
+        If module is a package, pass the directory containing the `__init__.py` file.
+
+    :param insert_modules:
+        If True, will call `insert_missing_modules` to create empty intermediate modules
+        with made-up module names (when importing test files not reachable from `sys.path`).
+
+    Example 1 of parent_module_*:
+
+        module_name:        "a.b.c.demo"
+        module_path:        Path("a/b/c/demo.py")
+        module_location:    Path("a/b/c/")
+        if "a.b.c" is package ("a/b/c/__init__.py" exists), then
+            parent_module_name:         "a.b.c"
+            parent_module_path:         Path("a/b/c/__init__.py")
+            parent_module_location:     Path("a/b/c/")
+        else:
+            parent_module_name:         "a.b.c"
+            parent_module_path:         Path("a/b/c")
+            parent_module_location:     Path("a/b/")
+
+    Example 2 of parent_module_*:
+
+        module_name:        "a.b.c"
+        module_path:        Path("a/b/c/__init__.py")
+        module_location:    Path("a/b/c/")
+        if  "a.b" is package ("a/b/__init__.py" exists), then
+            parent_module_name:         "a.b"
+            parent_module_path:         Path("a/b/__init__.py")
+            parent_module_location:     Path("a/b/")
+        else:
+            parent_module_name:         "a.b"
+            parent_module_path:         Path("a/b/")
+            parent_module_location:     Path("a/")
+    """
+    # Attempt to import the parent module, seems is our responsibility:
+    # https://github.com/python/cpython/blob/73906d5c908c1e0b73c5436faeff7d93698fc074/Lib/importlib/_bootstrap.py#L1308-L1311
+    parent_module_name, _, name = module_name.rpartition(".")
+    parent_module: ModuleType | None = None
+    if parent_module_name:
+        parent_module = sys.modules.get(parent_module_name)
+        if parent_module is None:
+            # Get parent_location based on location, get parent_path based on path.
+            if module_path.name == "__init__.py":
+                # If the current module is in a package,
+                # need to leave the package first and then enter the parent module.
+                parent_module_path = module_path.parent.parent
+            else:
+                parent_module_path = module_path.parent
+
+            if (parent_module_path / "__init__.py").is_file():
+                # If the parent module is a package, loading by  __init__.py file.
+                parent_module_path = parent_module_path / "__init__.py"
+
+            parent_module = _import_module_using_spec(
+                parent_module_name,
+                parent_module_path,
+                parent_module_path.parent,
+                insert_modules=insert_modules,
+            )
+
+    # Checking with sys.meta_path first in case one of its hooks can import this module,
+    # such as our own assertion-rewrite hook.
+    for meta_importer in sys.meta_path:
+        spec = meta_importer.find_spec(
+            module_name, [str(module_location), str(module_path)]
+        )
+        if spec_matches_module_path(spec, module_path):
+            break
+    else:
+        loader = None
+        if module_path.is_dir():
+            # The `spec_from_file_location` matches a loader based on the file extension by default.
+            # For a namespace package, need to manually specify a loader.
+            loader = NamespaceLoader(name, module_path, PathFinder())
+
+        spec = importlib.util.spec_from_file_location(
+            module_name, str(module_path), loader=loader
+        )
+
+    if spec_matches_module_path(spec, module_path):
+        assert spec is not None
+        # Find spec and import this module.
+        mod = importlib.util.module_from_spec(spec)
+        sys.modules[module_name] = mod
+        spec.loader.exec_module(mod)  # type: ignore[union-attr]
+
+        # Set this module as an attribute of the parent module (#12194).
+        if parent_module is not None:
+            setattr(parent_module, name, mod)
+
+        if insert_modules:
+            insert_missing_modules(sys.modules, module_name)
+        return mod
+
+    return None
+
+
+def spec_matches_module_path(module_spec: ModuleSpec | None, module_path: Path) -> bool:
+    """Return true if the given ModuleSpec can be used to import the given module path."""
+    if module_spec is None:
+        return False
+
+    if module_spec.origin:
+        return Path(module_spec.origin) == module_path
+
+    # Compare the path with the `module_spec.submodule_Search_Locations` in case
+    # the module is part of a namespace package.
+    # https://docs.python.org/3/library/importlib.html#importlib.machinery.ModuleSpec.submodule_search_locations
+    if module_spec.submodule_search_locations:  # can be None.
+        for path in module_spec.submodule_search_locations:
+            if Path(path) == module_path:
+                return True
+
+    return False
+
+
+# Implement a special _is_same function on Windows which returns True if the two filenames
+# compare equal, to circumvent os.path.samefile returning False for mounts in UNC (#7678).
+if sys.platform.startswith("win"):
+
+    def _is_same(f1: str, f2: str) -> bool:
+        return Path(f1) == Path(f2) or os.path.samefile(f1, f2)
+
+else:
+
+    def _is_same(f1: str, f2: str) -> bool:
+        return os.path.samefile(f1, f2)
+
+
+def module_name_from_path(path: Path, root: Path) -> str:
+    """
+    Return a dotted module name based on the given path, anchored on root.
+
+    For example: path="projects/src/tests/test_foo.py" and root="/projects", the
+    resulting module name will be "src.tests.test_foo".
+    """
+    path = path.with_suffix("")
+    try:
+        relative_path = path.relative_to(root)
+    except ValueError:
+        # If we can't get a relative path to root, use the full path, except
+        # for the first part ("d:\\" or "/" depending on the platform, for example).
+        path_parts = path.parts[1:]
+    else:
+        # Use the parts for the relative path to the root path.
+        path_parts = relative_path.parts
+
+    # Module name for packages do not contain the __init__ file, unless
+    # the `__init__.py` file is at the root.
+    if len(path_parts) >= 2 and path_parts[-1] == "__init__":
+        path_parts = path_parts[:-1]
+
+    # Module names cannot contain ".", normalize them to "_". This prevents
+    # a directory having a "." in the name (".env.310" for example) causing extra intermediate modules.
+    # Also, important to replace "." at the start of paths, as those are considered relative imports.
+    path_parts = tuple(x.replace(".", "_") for x in path_parts)
+
+    return ".".join(path_parts)
+
+
+def insert_missing_modules(modules: dict[str, ModuleType], module_name: str) -> None:
+    """
+    Used by ``import_path`` to create intermediate modules when using mode=importlib.
+
+    When we want to import a module as "src.tests.test_foo" for example, we need
+    to create empty modules "src" and "src.tests" after inserting "src.tests.test_foo",
+    otherwise "src.tests.test_foo" is not importable by ``__import__``.
+    """
+    module_parts = module_name.split(".")
+    while module_name:
+        parent_module_name, _, child_name = module_name.rpartition(".")
+        if parent_module_name:
+            parent_module = modules.get(parent_module_name)
+            if parent_module is None:
+                try:
+                    # If sys.meta_path is empty, calling import_module will issue
+                    # a warning and raise ModuleNotFoundError. To avoid the
+                    # warning, we check sys.meta_path explicitly and raise the error
+                    # ourselves to fall back to creating a dummy module.
+                    if not sys.meta_path:
+                        raise ModuleNotFoundError
+                    parent_module = importlib.import_module(parent_module_name)
+                except ModuleNotFoundError:
+                    parent_module = ModuleType(
+                        module_name,
+                        doc="Empty module created by pytest's importmode=importlib.",
+                    )
+                modules[parent_module_name] = parent_module
+
+            # Add child attribute to the parent that can reference the child
+            # modules.
+            if not hasattr(parent_module, child_name):
+                setattr(parent_module, child_name, modules[module_name])
+
+        module_parts.pop(-1)
+        module_name = ".".join(module_parts)
+
+
+def resolve_package_path(path: Path) -> Path | None:
+    """Return the Python package path by looking for the last
+    directory upwards which still contains an __init__.py.
+
+    Returns None if it cannot be determined.
+    """
+    result = None
+    for parent in itertools.chain((path,), path.parents):
+        if parent.is_dir():
+            if not (parent / "__init__.py").is_file():
+                break
+            if not parent.name.isidentifier():
+                break
+            result = parent
+    return result
+
+
+def resolve_pkg_root_and_module_name(
+    path: Path, *, consider_namespace_packages: bool = False
+) -> tuple[Path, str]:
+    """
+    Return the path to the directory of the root package that contains the
+    given Python file, and its module name:
+
+        src/
+            app/
+                __init__.py
+                core/
+                    __init__.py
+                    models.py
+
+    Passing the full path to `models.py` will yield Path("src") and "app.core.models".
+
+    If consider_namespace_packages is True, then we additionally check upwards in the hierarchy
+    for namespace packages:
+
+    https://packaging.python.org/en/latest/guides/packaging-namespace-packages
+
+    Raises CouldNotResolvePathError if the given path does not belong to a package (missing any __init__.py files).
+    """
+    pkg_root: Path | None = None
+    pkg_path = resolve_package_path(path)
+    if pkg_path is not None:
+        pkg_root = pkg_path.parent
+    if consider_namespace_packages:
+        start = pkg_root if pkg_root is not None else path.parent
+        for candidate in (start, *start.parents):
+            module_name = compute_module_name(candidate, path)
+            if module_name and is_importable(module_name, path):
+                # Point the pkg_root to the root of the namespace package.
+                pkg_root = candidate
+                break
+
+    if pkg_root is not None:
+        module_name = compute_module_name(pkg_root, path)
+        if module_name:
+            return pkg_root, module_name
+
+    raise CouldNotResolvePathError(f"Could not resolve for {path}")
+
+
+def is_importable(module_name: str, module_path: Path) -> bool:
+    """
+    Return if the given module path could be imported normally by Python, akin to the user
+    entering the REPL and importing the corresponding module name directly, and corresponds
+    to the module_path specified.
+
+    :param module_name:
+        Full module name that we want to check if is importable.
+        For example, "app.models".
+
+    :param module_path:
+        Full path to the python module/package we want to check if is importable.
+        For example, "/projects/src/app/models.py".
+    """
+    try:
+        # Note this is different from what we do in ``_import_module_using_spec``, where we explicitly search through
+        # sys.meta_path to be able to pass the path of the module that we want to import (``meta_importer.find_spec``).
+        # Using importlib.util.find_spec() is different, it gives the same results as trying to import
+        # the module normally in the REPL.
+        spec = importlib.util.find_spec(module_name)
+    except (ImportError, ValueError, ImportWarning):
+        return False
+    else:
+        return spec_matches_module_path(spec, module_path)
+
+
+def compute_module_name(root: Path, module_path: Path) -> str | None:
+    """Compute a module name based on a path and a root anchor."""
+    try:
+        path_without_suffix = module_path.with_suffix("")
+    except ValueError:
+        # Empty paths (such as Path.cwd()) might break meta_path hooks (like our own assertion rewriter).
+        return None
+
+    try:
+        relative = path_without_suffix.relative_to(root)
+    except ValueError:  # pragma: no cover
+        return None
+    names = list(relative.parts)
+    if not names:
+        return None
+    if names[-1] == "__init__":
+        names.pop()
+    return ".".join(names)
+
+
+class CouldNotResolvePathError(Exception):
+    """Custom exception raised by resolve_pkg_root_and_module_name."""
+
+
+def scandir(
+    path: str | os.PathLike[str],
+    sort_key: Callable[[os.DirEntry[str]], object] = lambda entry: entry.name,
+) -> list[os.DirEntry[str]]:
+    """Scan a directory recursively, in breadth-first order.
+
+    The returned entries are sorted according to the given key.
+    The default is to sort by name.
+    """
+    entries = []
+    with os.scandir(path) as s:
+        # Skip entries with symlink loops and other brokenness, so the caller
+        # doesn't have to deal with it.
+        for entry in s:
+            try:
+                entry.is_file()
+            except OSError as err:
+                if _ignore_error(err):
+                    continue
+                raise
+            entries.append(entry)
+    entries.sort(key=sort_key)  # type: ignore[arg-type]
+    return entries
+
+
+def visit(
+    path: str | os.PathLike[str], recurse: Callable[[os.DirEntry[str]], bool]
+) -> Iterator[os.DirEntry[str]]:
+    """Walk a directory recursively, in breadth-first order.
+
+    The `recurse` predicate determines whether a directory is recursed.
+
+    Entries at each directory level are sorted.
+    """
+    entries = scandir(path)
+    yield from entries
+    for entry in entries:
+        if entry.is_dir() and recurse(entry):
+            yield from visit(entry.path, recurse)
+
+
+def absolutepath(path: str | os.PathLike[str]) -> Path:
+    """Convert a path to an absolute path using os.path.abspath.
+
+    Prefer this over Path.resolve() (see #6523).
+    Prefer this over Path.absolute() (not public, doesn't normalize).
+    """
+    return Path(os.path.abspath(path))
+
+
+def commonpath(path1: Path, path2: Path) -> Path | None:
+    """Return the common part shared with the other path, or None if there is
+    no common part.
+
+    If one path is relative and one is absolute, returns None.
+    """
+    try:
+        return Path(os.path.commonpath((str(path1), str(path2))))
+    except ValueError:
+        return None
+
+
+def bestrelpath(directory: Path, dest: Path) -> str:
+    """Return a string which is a relative path from directory to dest such
+    that directory/bestrelpath == dest.
+
+    The paths must be either both absolute or both relative.
+
+    If no such path can be determined, returns dest.
+    """
+    assert isinstance(directory, Path)
+    assert isinstance(dest, Path)
+    if dest == directory:
+        return os.curdir
+    # Find the longest common directory.
+    base = commonpath(directory, dest)
+    # Can be the case on Windows for two absolute paths on different drives.
+    # Can be the case for two relative paths without common prefix.
+    # Can be the case for a relative path and an absolute path.
+    if not base:
+        return str(dest)
+    reldirectory = directory.relative_to(base)
+    reldest = dest.relative_to(base)
+    return os.path.join(
+        # Back from directory to base.
+        *([os.pardir] * len(reldirectory.parts)),
+        # Forward from base to dest.
+        *reldest.parts,
+    )
+
+
+def safe_exists(p: Path) -> bool:
+    """Like Path.exists(), but account for input arguments that might be too long (#11394)."""
+    try:
+        return p.exists()
+    except (ValueError, OSError):
+        # ValueError: stat: path too long for Windows
+        # OSError: [WinError 123] The filename, directory name, or volume label syntax is incorrect
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/py.typed b/backend/venv/lib/python3.12/site-packages/_pytest/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/pytester.py b/backend/venv/lib/python3.12/site-packages/_pytest/pytester.py
new file mode 100644
index 00000000..3f7520ee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/pytester.py
@@ -0,0 +1,1766 @@
+# mypy: allow-untyped-defs
+"""(Disabled by default) support for testing pytest and pytest plugins.
+
+PYTEST_DONT_REWRITE
+"""
+
+from __future__ import annotations
+
+import collections.abc
+import contextlib
+from fnmatch import fnmatch
+import gc
+import importlib
+from io import StringIO
+import locale
+import os
+from pathlib import Path
+import platform
+import re
+import shutil
+import subprocess
+import sys
+import traceback
+from typing import Any
+from typing import Callable
+from typing import Final
+from typing import final
+from typing import Generator
+from typing import IO
+from typing import Iterable
+from typing import Literal
+from typing import overload
+from typing import Sequence
+from typing import TextIO
+from typing import TYPE_CHECKING
+from weakref import WeakKeyDictionary
+
+from iniconfig import IniConfig
+from iniconfig import SectionWrapper
+
+from _pytest import timing
+from _pytest._code import Source
+from _pytest.capture import _get_multicapture
+from _pytest.compat import NOTSET
+from _pytest.compat import NotSetType
+from _pytest.config import _PluggyPlugin
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config import main
+from _pytest.config import PytestPluginManager
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureRequest
+from _pytest.main import Session
+from _pytest.monkeypatch import MonkeyPatch
+from _pytest.nodes import Collector
+from _pytest.nodes import Item
+from _pytest.outcomes import fail
+from _pytest.outcomes import importorskip
+from _pytest.outcomes import skip
+from _pytest.pathlib import bestrelpath
+from _pytest.pathlib import make_numbered_dir
+from _pytest.reports import CollectReport
+from _pytest.reports import TestReport
+from _pytest.tmpdir import TempPathFactory
+from _pytest.warning_types import PytestWarning
+
+
+if TYPE_CHECKING:
+    import pexpect
+
+
+pytest_plugins = ["pytester_assertions"]
+
+
+IGNORE_PAM = [  # filenames added when obtaining details about the current user
+    "/var/lib/sss/mc/passwd"
+]
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addoption(
+        "--lsof",
+        action="store_true",
+        dest="lsof",
+        default=False,
+        help="Run FD checks if lsof is available",
+    )
+
+    parser.addoption(
+        "--runpytest",
+        default="inprocess",
+        dest="runpytest",
+        choices=("inprocess", "subprocess"),
+        help=(
+            "Run pytest sub runs in tests using an 'inprocess' "
+            "or 'subprocess' (python -m main) method"
+        ),
+    )
+
+    parser.addini(
+        "pytester_example_dir", help="Directory to take the pytester example files from"
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    if config.getvalue("lsof"):
+        checker = LsofFdLeakChecker()
+        if checker.matching_platform():
+            config.pluginmanager.register(checker)
+
+    config.addinivalue_line(
+        "markers",
+        "pytester_example_path(*path_segments): join the given path "
+        "segments to `pytester_example_dir` for this test.",
+    )
+
+
+class LsofFdLeakChecker:
+    def get_open_files(self) -> list[tuple[str, str]]:
+        if sys.version_info >= (3, 11):
+            # New in Python 3.11, ignores utf-8 mode
+            encoding = locale.getencoding()
+        else:
+            encoding = locale.getpreferredencoding(False)
+        out = subprocess.run(
+            ("lsof", "-Ffn0", "-p", str(os.getpid())),
+            stdout=subprocess.PIPE,
+            stderr=subprocess.DEVNULL,
+            check=True,
+            text=True,
+            encoding=encoding,
+        ).stdout
+
+        def isopen(line: str) -> bool:
+            return line.startswith("f") and (
+                "deleted" not in line
+                and "mem" not in line
+                and "txt" not in line
+                and "cwd" not in line
+            )
+
+        open_files = []
+
+        for line in out.split("\n"):
+            if isopen(line):
+                fields = line.split("\0")
+                fd = fields[0][1:]
+                filename = fields[1][1:]
+                if filename in IGNORE_PAM:
+                    continue
+                if filename.startswith("/"):
+                    open_files.append((fd, filename))
+
+        return open_files
+
+    def matching_platform(self) -> bool:
+        try:
+            subprocess.run(("lsof", "-v"), check=True)
+        except (OSError, subprocess.CalledProcessError):
+            return False
+        else:
+            return True
+
+    @hookimpl(wrapper=True, tryfirst=True)
+    def pytest_runtest_protocol(self, item: Item) -> Generator[None, object, object]:
+        lines1 = self.get_open_files()
+        try:
+            return (yield)
+        finally:
+            if hasattr(sys, "pypy_version_info"):
+                gc.collect()
+            lines2 = self.get_open_files()
+
+            new_fds = {t[0] for t in lines2} - {t[0] for t in lines1}
+            leaked_files = [t for t in lines2 if t[0] in new_fds]
+            if leaked_files:
+                error = [
+                    f"***** {len(leaked_files)} FD leakage detected",
+                    *(str(f) for f in leaked_files),
+                    "*** Before:",
+                    *(str(f) for f in lines1),
+                    "*** After:",
+                    *(str(f) for f in lines2),
+                    f"***** {len(leaked_files)} FD leakage detected",
+                    "*** function {}:{}: {} ".format(*item.location),
+                    "See issue #2366",
+                ]
+                item.warn(PytestWarning("\n".join(error)))
+
+
+# used at least by pytest-xdist plugin
+
+
+@fixture
+def _pytest(request: FixtureRequest) -> PytestArg:
+    """Return a helper which offers a gethookrecorder(hook) method which
+    returns a HookRecorder instance which helps to make assertions about called
+    hooks."""
+    return PytestArg(request)
+
+
+class PytestArg:
+    def __init__(self, request: FixtureRequest) -> None:
+        self._request = request
+
+    def gethookrecorder(self, hook) -> HookRecorder:
+        hookrecorder = HookRecorder(hook._pm)
+        self._request.addfinalizer(hookrecorder.finish_recording)
+        return hookrecorder
+
+
+def get_public_names(values: Iterable[str]) -> list[str]:
+    """Only return names from iterator values without a leading underscore."""
+    return [x for x in values if x[0] != "_"]
+
+
+@final
+class RecordedHookCall:
+    """A recorded call to a hook.
+
+    The arguments to the hook call are set as attributes.
+    For example:
+
+    .. code-block:: python
+
+        calls = hook_recorder.getcalls("pytest_runtest_setup")
+        # Suppose pytest_runtest_setup was called once with `item=an_item`.
+        assert calls[0].item is an_item
+    """
+
+    def __init__(self, name: str, kwargs) -> None:
+        self.__dict__.update(kwargs)
+        self._name = name
+
+    def __repr__(self) -> str:
+        d = self.__dict__.copy()
+        del d["_name"]
+        return f"<RecordedHookCall {self._name!r}(**{d!r})>"
+
+    if TYPE_CHECKING:
+        # The class has undetermined attributes, this tells mypy about it.
+        def __getattr__(self, key: str): ...
+
+
+@final
+class HookRecorder:
+    """Record all hooks called in a plugin manager.
+
+    Hook recorders are created by :class:`Pytester`.
+
+    This wraps all the hook calls in the plugin manager, recording each call
+    before propagating the normal calls.
+    """
+
+    def __init__(
+        self, pluginmanager: PytestPluginManager, *, _ispytest: bool = False
+    ) -> None:
+        check_ispytest(_ispytest)
+
+        self._pluginmanager = pluginmanager
+        self.calls: list[RecordedHookCall] = []
+        self.ret: int | ExitCode | None = None
+
+        def before(hook_name: str, hook_impls, kwargs) -> None:
+            self.calls.append(RecordedHookCall(hook_name, kwargs))
+
+        def after(outcome, hook_name: str, hook_impls, kwargs) -> None:
+            pass
+
+        self._undo_wrapping = pluginmanager.add_hookcall_monitoring(before, after)
+
+    def finish_recording(self) -> None:
+        self._undo_wrapping()
+
+    def getcalls(self, names: str | Iterable[str]) -> list[RecordedHookCall]:
+        """Get all recorded calls to hooks with the given names (or name)."""
+        if isinstance(names, str):
+            names = names.split()
+        return [call for call in self.calls if call._name in names]
+
+    def assert_contains(self, entries: Sequence[tuple[str, str]]) -> None:
+        __tracebackhide__ = True
+        i = 0
+        entries = list(entries)
+        # Since Python 3.13, f_locals is not a dict, but eval requires a dict.
+        backlocals = dict(sys._getframe(1).f_locals)
+        while entries:
+            name, check = entries.pop(0)
+            for ind, call in enumerate(self.calls[i:]):
+                if call._name == name:
+                    print("NAMEMATCH", name, call)
+                    if eval(check, backlocals, call.__dict__):
+                        print("CHECKERMATCH", repr(check), "->", call)
+                    else:
+                        print("NOCHECKERMATCH", repr(check), "-", call)
+                        continue
+                    i += ind + 1
+                    break
+                print("NONAMEMATCH", name, "with", call)
+            else:
+                fail(f"could not find {name!r} check {check!r}")
+
+    def popcall(self, name: str) -> RecordedHookCall:
+        __tracebackhide__ = True
+        for i, call in enumerate(self.calls):
+            if call._name == name:
+                del self.calls[i]
+                return call
+        lines = [f"could not find call {name!r}, in:"]
+        lines.extend([f"  {x}" for x in self.calls])
+        fail("\n".join(lines))
+
+    def getcall(self, name: str) -> RecordedHookCall:
+        values = self.getcalls(name)
+        assert len(values) == 1, (name, values)
+        return values[0]
+
+    # functionality for test reports
+
+    @overload
+    def getreports(
+        self,
+        names: Literal["pytest_collectreport"],
+    ) -> Sequence[CollectReport]: ...
+
+    @overload
+    def getreports(
+        self,
+        names: Literal["pytest_runtest_logreport"],
+    ) -> Sequence[TestReport]: ...
+
+    @overload
+    def getreports(
+        self,
+        names: str | Iterable[str] = (
+            "pytest_collectreport",
+            "pytest_runtest_logreport",
+        ),
+    ) -> Sequence[CollectReport | TestReport]: ...
+
+    def getreports(
+        self,
+        names: str | Iterable[str] = (
+            "pytest_collectreport",
+            "pytest_runtest_logreport",
+        ),
+    ) -> Sequence[CollectReport | TestReport]:
+        return [x.report for x in self.getcalls(names)]
+
+    def matchreport(
+        self,
+        inamepart: str = "",
+        names: str | Iterable[str] = (
+            "pytest_runtest_logreport",
+            "pytest_collectreport",
+        ),
+        when: str | None = None,
+    ) -> CollectReport | TestReport:
+        """Return a testreport whose dotted import path matches."""
+        values = []
+        for rep in self.getreports(names=names):
+            if not when and rep.when != "call" and rep.passed:
+                # setup/teardown passing reports - let's ignore those
+                continue
+            if when and rep.when != when:
+                continue
+            if not inamepart or inamepart in rep.nodeid.split("::"):
+                values.append(rep)
+        if not values:
+            raise ValueError(
+                f"could not find test report matching {inamepart!r}: "
+                "no test reports at all!"
+            )
+        if len(values) > 1:
+            raise ValueError(
+                f"found 2 or more testreports matching {inamepart!r}: {values}"
+            )
+        return values[0]
+
+    @overload
+    def getfailures(
+        self,
+        names: Literal["pytest_collectreport"],
+    ) -> Sequence[CollectReport]: ...
+
+    @overload
+    def getfailures(
+        self,
+        names: Literal["pytest_runtest_logreport"],
+    ) -> Sequence[TestReport]: ...
+
+    @overload
+    def getfailures(
+        self,
+        names: str | Iterable[str] = (
+            "pytest_collectreport",
+            "pytest_runtest_logreport",
+        ),
+    ) -> Sequence[CollectReport | TestReport]: ...
+
+    def getfailures(
+        self,
+        names: str | Iterable[str] = (
+            "pytest_collectreport",
+            "pytest_runtest_logreport",
+        ),
+    ) -> Sequence[CollectReport | TestReport]:
+        return [rep for rep in self.getreports(names) if rep.failed]
+
+    def getfailedcollections(self) -> Sequence[CollectReport]:
+        return self.getfailures("pytest_collectreport")
+
+    def listoutcomes(
+        self,
+    ) -> tuple[
+        Sequence[TestReport],
+        Sequence[CollectReport | TestReport],
+        Sequence[CollectReport | TestReport],
+    ]:
+        passed = []
+        skipped = []
+        failed = []
+        for rep in self.getreports(
+            ("pytest_collectreport", "pytest_runtest_logreport")
+        ):
+            if rep.passed:
+                if rep.when == "call":
+                    assert isinstance(rep, TestReport)
+                    passed.append(rep)
+            elif rep.skipped:
+                skipped.append(rep)
+            else:
+                assert rep.failed, f"Unexpected outcome: {rep!r}"
+                failed.append(rep)
+        return passed, skipped, failed
+
+    def countoutcomes(self) -> list[int]:
+        return [len(x) for x in self.listoutcomes()]
+
+    def assertoutcome(self, passed: int = 0, skipped: int = 0, failed: int = 0) -> None:
+        __tracebackhide__ = True
+        from _pytest.pytester_assertions import assertoutcome
+
+        outcomes = self.listoutcomes()
+        assertoutcome(
+            outcomes,
+            passed=passed,
+            skipped=skipped,
+            failed=failed,
+        )
+
+    def clear(self) -> None:
+        self.calls[:] = []
+
+
+@fixture
+def linecomp() -> LineComp:
+    """A :class: `LineComp` instance for checking that an input linearly
+    contains a sequence of strings."""
+    return LineComp()
+
+
+@fixture(name="LineMatcher")
+def LineMatcher_fixture(request: FixtureRequest) -> type[LineMatcher]:
+    """A reference to the :class: `LineMatcher`.
+
+    This is instantiable with a list of lines (without their trailing newlines).
+    This is useful for testing large texts, such as the output of commands.
+    """
+    return LineMatcher
+
+
+@fixture
+def pytester(
+    request: FixtureRequest, tmp_path_factory: TempPathFactory, monkeypatch: MonkeyPatch
+) -> Pytester:
+    """
+    Facilities to write tests/configuration files, execute pytest in isolation, and match
+    against expected output, perfect for black-box testing of pytest plugins.
+
+    It attempts to isolate the test run from external factors as much as possible, modifying
+    the current working directory to ``path`` and environment variables during initialization.
+
+    It is particularly useful for testing plugins. It is similar to the :fixture:`tmp_path`
+    fixture but provides methods which aid in testing pytest itself.
+    """
+    return Pytester(request, tmp_path_factory, monkeypatch, _ispytest=True)
+
+
+@fixture
+def _sys_snapshot() -> Generator[None]:
+    snappaths = SysPathsSnapshot()
+    snapmods = SysModulesSnapshot()
+    yield
+    snapmods.restore()
+    snappaths.restore()
+
+
+@fixture
+def _config_for_test() -> Generator[Config]:
+    from _pytest.config import get_config
+
+    config = get_config()
+    yield config
+    config._ensure_unconfigure()  # cleanup, e.g. capman closing tmpfiles.
+
+
+# Regex to match the session duration string in the summary: "74.34s".
+rex_session_duration = re.compile(r"\d+\.\d\ds")
+# Regex to match all the counts and phrases in the summary line: "34 passed, 111 skipped".
+rex_outcome = re.compile(r"(\d+) (\w+)")
+
+
+@final
+class RunResult:
+    """The result of running a command from :class:`~pytest.Pytester`."""
+
+    def __init__(
+        self,
+        ret: int | ExitCode,
+        outlines: list[str],
+        errlines: list[str],
+        duration: float,
+    ) -> None:
+        try:
+            self.ret: int | ExitCode = ExitCode(ret)
+            """The return value."""
+        except ValueError:
+            self.ret = ret
+        self.outlines = outlines
+        """List of lines captured from stdout."""
+        self.errlines = errlines
+        """List of lines captured from stderr."""
+        self.stdout = LineMatcher(outlines)
+        """:class:`~pytest.LineMatcher` of stdout.
+
+        Use e.g. :func:`str(stdout) <pytest.LineMatcher.__str__()>` to reconstruct stdout, or the commonly used
+        :func:`stdout.fnmatch_lines() <pytest.LineMatcher.fnmatch_lines()>` method.
+        """
+        self.stderr = LineMatcher(errlines)
+        """:class:`~pytest.LineMatcher` of stderr."""
+        self.duration = duration
+        """Duration in seconds."""
+
+    def __repr__(self) -> str:
+        return (
+            "<RunResult ret=%s len(stdout.lines)=%d len(stderr.lines)=%d duration=%.2fs>"
+            % (self.ret, len(self.stdout.lines), len(self.stderr.lines), self.duration)
+        )
+
+    def parseoutcomes(self) -> dict[str, int]:
+        """Return a dictionary of outcome noun -> count from parsing the terminal
+        output that the test process produced.
+
+        The returned nouns will always be in plural form::
+
+            ======= 1 failed, 1 passed, 1 warning, 1 error in 0.13s ====
+
+        Will return ``{"failed": 1, "passed": 1, "warnings": 1, "errors": 1}``.
+        """
+        return self.parse_summary_nouns(self.outlines)
+
+    @classmethod
+    def parse_summary_nouns(cls, lines) -> dict[str, int]:
+        """Extract the nouns from a pytest terminal summary line.
+
+        It always returns the plural noun for consistency::
+
+            ======= 1 failed, 1 passed, 1 warning, 1 error in 0.13s ====
+
+        Will return ``{"failed": 1, "passed": 1, "warnings": 1, "errors": 1}``.
+        """
+        for line in reversed(lines):
+            if rex_session_duration.search(line):
+                outcomes = rex_outcome.findall(line)
+                ret = {noun: int(count) for (count, noun) in outcomes}
+                break
+        else:
+            raise ValueError("Pytest terminal summary report not found")
+
+        to_plural = {
+            "warning": "warnings",
+            "error": "errors",
+        }
+        return {to_plural.get(k, k): v for k, v in ret.items()}
+
+    def assert_outcomes(
+        self,
+        passed: int = 0,
+        skipped: int = 0,
+        failed: int = 0,
+        errors: int = 0,
+        xpassed: int = 0,
+        xfailed: int = 0,
+        warnings: int | None = None,
+        deselected: int | None = None,
+    ) -> None:
+        """
+        Assert that the specified outcomes appear with the respective
+        numbers (0 means it didn't occur) in the text output from a test run.
+
+        ``warnings`` and ``deselected`` are only checked if not None.
+        """
+        __tracebackhide__ = True
+        from _pytest.pytester_assertions import assert_outcomes
+
+        outcomes = self.parseoutcomes()
+        assert_outcomes(
+            outcomes,
+            passed=passed,
+            skipped=skipped,
+            failed=failed,
+            errors=errors,
+            xpassed=xpassed,
+            xfailed=xfailed,
+            warnings=warnings,
+            deselected=deselected,
+        )
+
+
+class SysModulesSnapshot:
+    def __init__(self, preserve: Callable[[str], bool] | None = None) -> None:
+        self.__preserve = preserve
+        self.__saved = dict(sys.modules)
+
+    def restore(self) -> None:
+        if self.__preserve:
+            self.__saved.update(
+                (k, m) for k, m in sys.modules.items() if self.__preserve(k)
+            )
+        sys.modules.clear()
+        sys.modules.update(self.__saved)
+
+
+class SysPathsSnapshot:
+    def __init__(self) -> None:
+        self.__saved = list(sys.path), list(sys.meta_path)
+
+    def restore(self) -> None:
+        sys.path[:], sys.meta_path[:] = self.__saved
+
+
+@final
+class Pytester:
+    """
+    Facilities to write tests/configuration files, execute pytest in isolation, and match
+    against expected output, perfect for black-box testing of pytest plugins.
+
+    It attempts to isolate the test run from external factors as much as possible, modifying
+    the current working directory to :attr:`path` and environment variables during initialization.
+    """
+
+    __test__ = False
+
+    CLOSE_STDIN: Final = NOTSET
+
+    class TimeoutExpired(Exception):
+        pass
+
+    def __init__(
+        self,
+        request: FixtureRequest,
+        tmp_path_factory: TempPathFactory,
+        monkeypatch: MonkeyPatch,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._request = request
+        self._mod_collections: WeakKeyDictionary[Collector, list[Item | Collector]] = (
+            WeakKeyDictionary()
+        )
+        if request.function:
+            name: str = request.function.__name__
+        else:
+            name = request.node.name
+        self._name = name
+        self._path: Path = tmp_path_factory.mktemp(name, numbered=True)
+        #: A list of plugins to use with :py:meth:`parseconfig` and
+        #: :py:meth:`runpytest`.  Initially this is an empty list but plugins can
+        #: be added to the list.  The type of items to add to the list depends on
+        #: the method using them so refer to them for details.
+        self.plugins: list[str | _PluggyPlugin] = []
+        self._sys_path_snapshot = SysPathsSnapshot()
+        self._sys_modules_snapshot = self.__take_sys_modules_snapshot()
+        self._request.addfinalizer(self._finalize)
+        self._method = self._request.config.getoption("--runpytest")
+        self._test_tmproot = tmp_path_factory.mktemp(f"tmp-{name}", numbered=True)
+
+        self._monkeypatch = mp = monkeypatch
+        self.chdir()
+        mp.setenv("PYTEST_DEBUG_TEMPROOT", str(self._test_tmproot))
+        # Ensure no unexpected caching via tox.
+        mp.delenv("TOX_ENV_DIR", raising=False)
+        # Discard outer pytest options.
+        mp.delenv("PYTEST_ADDOPTS", raising=False)
+        # Ensure no user config is used.
+        tmphome = str(self.path)
+        mp.setenv("HOME", tmphome)
+        mp.setenv("USERPROFILE", tmphome)
+        # Do not use colors for inner runs by default.
+        mp.setenv("PY_COLORS", "0")
+
+    @property
+    def path(self) -> Path:
+        """Temporary directory path used to create files/run tests from, etc."""
+        return self._path
+
+    def __repr__(self) -> str:
+        return f"<Pytester {self.path!r}>"
+
+    def _finalize(self) -> None:
+        """
+        Clean up global state artifacts.
+
+        Some methods modify the global interpreter state and this tries to
+        clean this up. It does not remove the temporary directory however so
+        it can be looked at after the test run has finished.
+        """
+        self._sys_modules_snapshot.restore()
+        self._sys_path_snapshot.restore()
+
+    def __take_sys_modules_snapshot(self) -> SysModulesSnapshot:
+        # Some zope modules used by twisted-related tests keep internal state
+        # and can't be deleted; we had some trouble in the past with
+        # `zope.interface` for example.
+        #
+        # Preserve readline due to https://bugs.python.org/issue41033.
+        # pexpect issues a SIGWINCH.
+        def preserve_module(name):
+            return name.startswith(("zope", "readline"))
+
+        return SysModulesSnapshot(preserve=preserve_module)
+
+    def make_hook_recorder(self, pluginmanager: PytestPluginManager) -> HookRecorder:
+        """Create a new :class:`HookRecorder` for a :class:`PytestPluginManager`."""
+        pluginmanager.reprec = reprec = HookRecorder(pluginmanager, _ispytest=True)  # type: ignore[attr-defined]
+        self._request.addfinalizer(reprec.finish_recording)
+        return reprec
+
+    def chdir(self) -> None:
+        """Cd into the temporary directory.
+
+        This is done automatically upon instantiation.
+        """
+        self._monkeypatch.chdir(self.path)
+
+    def _makefile(
+        self,
+        ext: str,
+        lines: Sequence[Any | bytes],
+        files: dict[str, str],
+        encoding: str = "utf-8",
+    ) -> Path:
+        items = list(files.items())
+
+        if ext is None:
+            raise TypeError("ext must not be None")
+
+        if ext and not ext.startswith("."):
+            raise ValueError(
+                f"pytester.makefile expects a file extension, try .{ext} instead of {ext}"
+            )
+
+        def to_text(s: Any | bytes) -> str:
+            return s.decode(encoding) if isinstance(s, bytes) else str(s)
+
+        if lines:
+            source = "\n".join(to_text(x) for x in lines)
+            basename = self._name
+            items.insert(0, (basename, source))
+
+        ret = None
+        for basename, value in items:
+            p = self.path.joinpath(basename).with_suffix(ext)
+            p.parent.mkdir(parents=True, exist_ok=True)
+            source_ = Source(value)
+            source = "\n".join(to_text(line) for line in source_.lines)
+            p.write_text(source.strip(), encoding=encoding)
+            if ret is None:
+                ret = p
+        assert ret is not None
+        return ret
+
+    def makefile(self, ext: str, *args: str, **kwargs: str) -> Path:
+        r"""Create new text file(s) in the test directory.
+
+        :param ext:
+            The extension the file(s) should use, including the dot, e.g. `.py`.
+        :param args:
+            All args are treated as strings and joined using newlines.
+            The result is written as contents to the file.  The name of the
+            file is based on the test function requesting this fixture.
+        :param kwargs:
+            Each keyword is the name of a file, while the value of it will
+            be written as contents of the file.
+        :returns:
+            The first created file.
+
+        Examples:
+
+        .. code-block:: python
+
+            pytester.makefile(".txt", "line1", "line2")
+
+            pytester.makefile(".ini", pytest="[pytest]\naddopts=-rs\n")
+
+        To create binary files, use :meth:`pathlib.Path.write_bytes` directly:
+
+        .. code-block:: python
+
+            filename = pytester.path.joinpath("foo.bin")
+            filename.write_bytes(b"...")
+        """
+        return self._makefile(ext, args, kwargs)
+
+    def makeconftest(self, source: str) -> Path:
+        """Write a conftest.py file.
+
+        :param source: The contents.
+        :returns: The conftest.py file.
+        """
+        return self.makepyfile(conftest=source)
+
+    def makeini(self, source: str) -> Path:
+        """Write a tox.ini file.
+
+        :param source: The contents.
+        :returns: The tox.ini file.
+        """
+        return self.makefile(".ini", tox=source)
+
+    def getinicfg(self, source: str) -> SectionWrapper:
+        """Return the pytest section from the tox.ini config file."""
+        p = self.makeini(source)
+        return IniConfig(str(p))["pytest"]
+
+    def makepyprojecttoml(self, source: str) -> Path:
+        """Write a pyproject.toml file.
+
+        :param source: The contents.
+        :returns: The pyproject.ini file.
+
+        .. versionadded:: 6.0
+        """
+        return self.makefile(".toml", pyproject=source)
+
+    def makepyfile(self, *args, **kwargs) -> Path:
+        r"""Shortcut for .makefile() with a .py extension.
+
+        Defaults to the test name with a '.py' extension, e.g test_foobar.py, overwriting
+        existing files.
+
+        Examples:
+
+        .. code-block:: python
+
+            def test_something(pytester):
+                # Initial file is created test_something.py.
+                pytester.makepyfile("foobar")
+                # To create multiple files, pass kwargs accordingly.
+                pytester.makepyfile(custom="foobar")
+                # At this point, both 'test_something.py' & 'custom.py' exist in the test directory.
+
+        """
+        return self._makefile(".py", args, kwargs)
+
+    def maketxtfile(self, *args, **kwargs) -> Path:
+        r"""Shortcut for .makefile() with a .txt extension.
+
+        Defaults to the test name with a '.txt' extension, e.g test_foobar.txt, overwriting
+        existing files.
+
+        Examples:
+
+        .. code-block:: python
+
+            def test_something(pytester):
+                # Initial file is created test_something.txt.
+                pytester.maketxtfile("foobar")
+                # To create multiple files, pass kwargs accordingly.
+                pytester.maketxtfile(custom="foobar")
+                # At this point, both 'test_something.txt' & 'custom.txt' exist in the test directory.
+
+        """
+        return self._makefile(".txt", args, kwargs)
+
+    def syspathinsert(self, path: str | os.PathLike[str] | None = None) -> None:
+        """Prepend a directory to sys.path, defaults to :attr:`path`.
+
+        This is undone automatically when this object dies at the end of each
+        test.
+
+        :param path:
+            The path.
+        """
+        if path is None:
+            path = self.path
+
+        self._monkeypatch.syspath_prepend(str(path))
+
+    def mkdir(self, name: str | os.PathLike[str]) -> Path:
+        """Create a new (sub)directory.
+
+        :param name:
+            The name of the directory, relative to the pytester path.
+        :returns:
+            The created directory.
+        :rtype: pathlib.Path
+        """
+        p = self.path / name
+        p.mkdir()
+        return p
+
+    def mkpydir(self, name: str | os.PathLike[str]) -> Path:
+        """Create a new python package.
+
+        This creates a (sub)directory with an empty ``__init__.py`` file so it
+        gets recognised as a Python package.
+        """
+        p = self.path / name
+        p.mkdir()
+        p.joinpath("__init__.py").touch()
+        return p
+
+    def copy_example(self, name: str | None = None) -> Path:
+        """Copy file from project's directory into the testdir.
+
+        :param name:
+            The name of the file to copy.
+        :return:
+            Path to the copied directory (inside ``self.path``).
+        :rtype: pathlib.Path
+        """
+        example_dir_ = self._request.config.getini("pytester_example_dir")
+        if example_dir_ is None:
+            raise ValueError("pytester_example_dir is unset, can't copy examples")
+        example_dir: Path = self._request.config.rootpath / example_dir_
+
+        for extra_element in self._request.node.iter_markers("pytester_example_path"):
+            assert extra_element.args
+            example_dir = example_dir.joinpath(*extra_element.args)
+
+        if name is None:
+            func_name = self._name
+            maybe_dir = example_dir / func_name
+            maybe_file = example_dir / (func_name + ".py")
+
+            if maybe_dir.is_dir():
+                example_path = maybe_dir
+            elif maybe_file.is_file():
+                example_path = maybe_file
+            else:
+                raise LookupError(
+                    f"{func_name} can't be found as module or package in {example_dir}"
+                )
+        else:
+            example_path = example_dir.joinpath(name)
+
+        if example_path.is_dir() and not example_path.joinpath("__init__.py").is_file():
+            shutil.copytree(example_path, self.path, symlinks=True, dirs_exist_ok=True)
+            return self.path
+        elif example_path.is_file():
+            result = self.path.joinpath(example_path.name)
+            shutil.copy(example_path, result)
+            return result
+        else:
+            raise LookupError(
+                f'example "{example_path}" is not found as a file or directory'
+            )
+
+    def getnode(self, config: Config, arg: str | os.PathLike[str]) -> Collector | Item:
+        """Get the collection node of a file.
+
+        :param config:
+           A pytest config.
+           See :py:meth:`parseconfig` and :py:meth:`parseconfigure` for creating it.
+        :param arg:
+            Path to the file.
+        :returns:
+            The node.
+        """
+        session = Session.from_config(config)
+        assert "::" not in str(arg)
+        p = Path(os.path.abspath(arg))
+        config.hook.pytest_sessionstart(session=session)
+        res = session.perform_collect([str(p)], genitems=False)[0]
+        config.hook.pytest_sessionfinish(session=session, exitstatus=ExitCode.OK)
+        return res
+
+    def getpathnode(self, path: str | os.PathLike[str]) -> Collector | Item:
+        """Return the collection node of a file.
+
+        This is like :py:meth:`getnode` but uses :py:meth:`parseconfigure` to
+        create the (configured) pytest Config instance.
+
+        :param path:
+            Path to the file.
+        :returns:
+            The node.
+        """
+        path = Path(path)
+        config = self.parseconfigure(path)
+        session = Session.from_config(config)
+        x = bestrelpath(session.path, path)
+        config.hook.pytest_sessionstart(session=session)
+        res = session.perform_collect([x], genitems=False)[0]
+        config.hook.pytest_sessionfinish(session=session, exitstatus=ExitCode.OK)
+        return res
+
+    def genitems(self, colitems: Sequence[Item | Collector]) -> list[Item]:
+        """Generate all test items from a collection node.
+
+        This recurses into the collection node and returns a list of all the
+        test items contained within.
+
+        :param colitems:
+            The collection nodes.
+        :returns:
+            The collected items.
+        """
+        session = colitems[0].session
+        result: list[Item] = []
+        for colitem in colitems:
+            result.extend(session.genitems(colitem))
+        return result
+
+    def runitem(self, source: str) -> Any:
+        """Run the "test_func" Item.
+
+        The calling test instance (class containing the test method) must
+        provide a ``.getrunner()`` method which should return a runner which
+        can run the test protocol for a single item, e.g.
+        ``_pytest.runner.runtestprotocol``.
+        """
+        # used from runner functional tests
+        item = self.getitem(source)
+        # the test class where we are called from wants to provide the runner
+        testclassinstance = self._request.instance
+        runner = testclassinstance.getrunner()
+        return runner(item)
+
+    def inline_runsource(self, source: str, *cmdlineargs) -> HookRecorder:
+        """Run a test module in process using ``pytest.main()``.
+
+        This run writes "source" into a temporary file and runs
+        ``pytest.main()`` on it, returning a :py:class:`HookRecorder` instance
+        for the result.
+
+        :param source: The source code of the test module.
+        :param cmdlineargs: Any extra command line arguments to use.
+        """
+        p = self.makepyfile(source)
+        values = [*list(cmdlineargs), p]
+        return self.inline_run(*values)
+
+    def inline_genitems(self, *args) -> tuple[list[Item], HookRecorder]:
+        """Run ``pytest.main(['--collect-only'])`` in-process.
+
+        Runs the :py:func:`pytest.main` function to run all of pytest inside
+        the test process itself like :py:meth:`inline_run`, but returns a
+        tuple of the collected items and a :py:class:`HookRecorder` instance.
+        """
+        rec = self.inline_run("--collect-only", *args)
+        items = [x.item for x in rec.getcalls("pytest_itemcollected")]
+        return items, rec
+
+    def inline_run(
+        self,
+        *args: str | os.PathLike[str],
+        plugins=(),
+        no_reraise_ctrlc: bool = False,
+    ) -> HookRecorder:
+        """Run ``pytest.main()`` in-process, returning a HookRecorder.
+
+        Runs the :py:func:`pytest.main` function to run all of pytest inside
+        the test process itself.  This means it can return a
+        :py:class:`HookRecorder` instance which gives more detailed results
+        from that run than can be done by matching stdout/stderr from
+        :py:meth:`runpytest`.
+
+        :param args:
+            Command line arguments to pass to :py:func:`pytest.main`.
+        :param plugins:
+            Extra plugin instances the ``pytest.main()`` instance should use.
+        :param no_reraise_ctrlc:
+            Typically we reraise keyboard interrupts from the child run. If
+            True, the KeyboardInterrupt exception is captured.
+        """
+        # (maybe a cpython bug?) the importlib cache sometimes isn't updated
+        # properly between file creation and inline_run (especially if imports
+        # are interspersed with file creation)
+        importlib.invalidate_caches()
+
+        plugins = list(plugins)
+        finalizers = []
+        try:
+            # Any sys.module or sys.path changes done while running pytest
+            # inline should be reverted after the test run completes to avoid
+            # clashing with later inline tests run within the same pytest test,
+            # e.g. just because they use matching test module names.
+            finalizers.append(self.__take_sys_modules_snapshot().restore)
+            finalizers.append(SysPathsSnapshot().restore)
+
+            # Important note:
+            # - our tests should not leave any other references/registrations
+            #   laying around other than possibly loaded test modules
+            #   referenced from sys.modules, as nothing will clean those up
+            #   automatically
+
+            rec = []
+
+            class Collect:
+                def pytest_configure(x, config: Config) -> None:
+                    rec.append(self.make_hook_recorder(config.pluginmanager))
+
+            plugins.append(Collect())
+            ret = main([str(x) for x in args], plugins=plugins)
+            if len(rec) == 1:
+                reprec = rec.pop()
+            else:
+
+                class reprec:  # type: ignore
+                    pass
+
+            reprec.ret = ret
+
+            # Typically we reraise keyboard interrupts from the child run
+            # because it's our user requesting interruption of the testing.
+            if ret == ExitCode.INTERRUPTED and not no_reraise_ctrlc:
+                calls = reprec.getcalls("pytest_keyboard_interrupt")
+                if calls and calls[-1].excinfo.type == KeyboardInterrupt:
+                    raise KeyboardInterrupt()
+            return reprec
+        finally:
+            for finalizer in finalizers:
+                finalizer()
+
+    def runpytest_inprocess(
+        self, *args: str | os.PathLike[str], **kwargs: Any
+    ) -> RunResult:
+        """Return result of running pytest in-process, providing a similar
+        interface to what self.runpytest() provides."""
+        syspathinsert = kwargs.pop("syspathinsert", False)
+
+        if syspathinsert:
+            self.syspathinsert()
+        now = timing.time()
+        capture = _get_multicapture("sys")
+        capture.start_capturing()
+        try:
+            try:
+                reprec = self.inline_run(*args, **kwargs)
+            except SystemExit as e:
+                ret = e.args[0]
+                try:
+                    ret = ExitCode(e.args[0])
+                except ValueError:
+                    pass
+
+                class reprec:  # type: ignore
+                    ret = ret
+
+            except Exception:
+                traceback.print_exc()
+
+                class reprec:  # type: ignore
+                    ret = ExitCode(3)
+
+        finally:
+            out, err = capture.readouterr()
+            capture.stop_capturing()
+            sys.stdout.write(out)
+            sys.stderr.write(err)
+
+        assert reprec.ret is not None
+        res = RunResult(
+            reprec.ret, out.splitlines(), err.splitlines(), timing.time() - now
+        )
+        res.reprec = reprec  # type: ignore
+        return res
+
+    def runpytest(self, *args: str | os.PathLike[str], **kwargs: Any) -> RunResult:
+        """Run pytest inline or in a subprocess, depending on the command line
+        option "--runpytest" and return a :py:class:`~pytest.RunResult`."""
+        new_args = self._ensure_basetemp(args)
+        if self._method == "inprocess":
+            return self.runpytest_inprocess(*new_args, **kwargs)
+        elif self._method == "subprocess":
+            return self.runpytest_subprocess(*new_args, **kwargs)
+        raise RuntimeError(f"Unrecognized runpytest option: {self._method}")
+
+    def _ensure_basetemp(
+        self, args: Sequence[str | os.PathLike[str]]
+    ) -> list[str | os.PathLike[str]]:
+        new_args = list(args)
+        for x in new_args:
+            if str(x).startswith("--basetemp"):
+                break
+        else:
+            new_args.append(
+                "--basetemp={}".format(self.path.parent.joinpath("basetemp"))
+            )
+        return new_args
+
+    def parseconfig(self, *args: str | os.PathLike[str]) -> Config:
+        """Return a new pytest :class:`pytest.Config` instance from given
+        commandline args.
+
+        This invokes the pytest bootstrapping code in _pytest.config to create a
+        new :py:class:`pytest.PytestPluginManager` and call the
+        :hook:`pytest_cmdline_parse` hook to create a new :class:`pytest.Config`
+        instance.
+
+        If :attr:`plugins` has been populated they should be plugin modules
+        to be registered with the plugin manager.
+        """
+        import _pytest.config
+
+        new_args = self._ensure_basetemp(args)
+        new_args = [str(x) for x in new_args]
+
+        config = _pytest.config._prepareconfig(new_args, self.plugins)  # type: ignore[arg-type]
+        # we don't know what the test will do with this half-setup config
+        # object and thus we make sure it gets unconfigured properly in any
+        # case (otherwise capturing could still be active, for example)
+        self._request.addfinalizer(config._ensure_unconfigure)
+        return config
+
+    def parseconfigure(self, *args: str | os.PathLike[str]) -> Config:
+        """Return a new pytest configured Config instance.
+
+        Returns a new :py:class:`pytest.Config` instance like
+        :py:meth:`parseconfig`, but also calls the :hook:`pytest_configure`
+        hook.
+        """
+        config = self.parseconfig(*args)
+        config._do_configure()
+        return config
+
+    def getitem(
+        self, source: str | os.PathLike[str], funcname: str = "test_func"
+    ) -> Item:
+        """Return the test item for a test function.
+
+        Writes the source to a python file and runs pytest's collection on
+        the resulting module, returning the test item for the requested
+        function name.
+
+        :param source:
+            The module source.
+        :param funcname:
+            The name of the test function for which to return a test item.
+        :returns:
+            The test item.
+        """
+        items = self.getitems(source)
+        for item in items:
+            if item.name == funcname:
+                return item
+        assert 0, f"{funcname!r} item not found in module:\n{source}\nitems: {items}"
+
+    def getitems(self, source: str | os.PathLike[str]) -> list[Item]:
+        """Return all test items collected from the module.
+
+        Writes the source to a Python file and runs pytest's collection on
+        the resulting module, returning all test items contained within.
+        """
+        modcol = self.getmodulecol(source)
+        return self.genitems([modcol])
+
+    def getmodulecol(
+        self,
+        source: str | os.PathLike[str],
+        configargs=(),
+        *,
+        withinit: bool = False,
+    ):
+        """Return the module collection node for ``source``.
+
+        Writes ``source`` to a file using :py:meth:`makepyfile` and then
+        runs the pytest collection on it, returning the collection node for the
+        test module.
+
+        :param source:
+            The source code of the module to collect.
+
+        :param configargs:
+            Any extra arguments to pass to :py:meth:`parseconfigure`.
+
+        :param withinit:
+            Whether to also write an ``__init__.py`` file to the same
+            directory to ensure it is a package.
+        """
+        if isinstance(source, os.PathLike):
+            path = self.path.joinpath(source)
+            assert not withinit, "not supported for paths"
+        else:
+            kw = {self._name: str(source)}
+            path = self.makepyfile(**kw)
+        if withinit:
+            self.makepyfile(__init__="#")
+        self.config = config = self.parseconfigure(path, *configargs)
+        return self.getnode(config, path)
+
+    def collect_by_name(self, modcol: Collector, name: str) -> Item | Collector | None:
+        """Return the collection node for name from the module collection.
+
+        Searches a module collection node for a collection node matching the
+        given name.
+
+        :param modcol: A module collection node; see :py:meth:`getmodulecol`.
+        :param name: The name of the node to return.
+        """
+        if modcol not in self._mod_collections:
+            self._mod_collections[modcol] = list(modcol.collect())
+        for colitem in self._mod_collections[modcol]:
+            if colitem.name == name:
+                return colitem
+        return None
+
+    def popen(
+        self,
+        cmdargs: Sequence[str | os.PathLike[str]],
+        stdout: int | TextIO = subprocess.PIPE,
+        stderr: int | TextIO = subprocess.PIPE,
+        stdin: NotSetType | bytes | IO[Any] | int = CLOSE_STDIN,
+        **kw,
+    ):
+        """Invoke :py:class:`subprocess.Popen`.
+
+        Calls :py:class:`subprocess.Popen` making sure the current working
+        directory is in ``PYTHONPATH``.
+
+        You probably want to use :py:meth:`run` instead.
+        """
+        env = os.environ.copy()
+        env["PYTHONPATH"] = os.pathsep.join(
+            filter(None, [os.getcwd(), env.get("PYTHONPATH", "")])
+        )
+        kw["env"] = env
+
+        if stdin is self.CLOSE_STDIN:
+            kw["stdin"] = subprocess.PIPE
+        elif isinstance(stdin, bytes):
+            kw["stdin"] = subprocess.PIPE
+        else:
+            kw["stdin"] = stdin
+
+        popen = subprocess.Popen(cmdargs, stdout=stdout, stderr=stderr, **kw)
+        if stdin is self.CLOSE_STDIN:
+            assert popen.stdin is not None
+            popen.stdin.close()
+        elif isinstance(stdin, bytes):
+            assert popen.stdin is not None
+            popen.stdin.write(stdin)
+
+        return popen
+
+    def run(
+        self,
+        *cmdargs: str | os.PathLike[str],
+        timeout: float | None = None,
+        stdin: NotSetType | bytes | IO[Any] | int = CLOSE_STDIN,
+    ) -> RunResult:
+        """Run a command with arguments.
+
+        Run a process using :py:class:`subprocess.Popen` saving the stdout and
+        stderr.
+
+        :param cmdargs:
+            The sequence of arguments to pass to :py:class:`subprocess.Popen`,
+            with path-like objects being converted to :py:class:`str`
+            automatically.
+        :param timeout:
+            The period in seconds after which to timeout and raise
+            :py:class:`Pytester.TimeoutExpired`.
+        :param stdin:
+            Optional standard input.
+
+            - If it is ``CLOSE_STDIN`` (Default), then this method calls
+              :py:class:`subprocess.Popen` with ``stdin=subprocess.PIPE``, and
+              the standard input is closed immediately after the new command is
+              started.
+
+            - If it is of type :py:class:`bytes`, these bytes are sent to the
+              standard input of the command.
+
+            - Otherwise, it is passed through to :py:class:`subprocess.Popen`.
+              For further information in this case, consult the document of the
+              ``stdin`` parameter in :py:class:`subprocess.Popen`.
+        :type stdin: _pytest.compat.NotSetType | bytes | IO[Any] | int
+        :returns:
+            The result.
+
+        """
+        __tracebackhide__ = True
+
+        cmdargs = tuple(os.fspath(arg) for arg in cmdargs)
+        p1 = self.path.joinpath("stdout")
+        p2 = self.path.joinpath("stderr")
+        print("running:", *cmdargs)
+        print("     in:", Path.cwd())
+
+        with p1.open("w", encoding="utf8") as f1, p2.open("w", encoding="utf8") as f2:
+            now = timing.time()
+            popen = self.popen(
+                cmdargs,
+                stdin=stdin,
+                stdout=f1,
+                stderr=f2,
+                close_fds=(sys.platform != "win32"),
+            )
+            if popen.stdin is not None:
+                popen.stdin.close()
+
+            def handle_timeout() -> None:
+                __tracebackhide__ = True
+
+                timeout_message = f"{timeout} second timeout expired running: {cmdargs}"
+
+                popen.kill()
+                popen.wait()
+                raise self.TimeoutExpired(timeout_message)
+
+            if timeout is None:
+                ret = popen.wait()
+            else:
+                try:
+                    ret = popen.wait(timeout)
+                except subprocess.TimeoutExpired:
+                    handle_timeout()
+
+        with p1.open(encoding="utf8") as f1, p2.open(encoding="utf8") as f2:
+            out = f1.read().splitlines()
+            err = f2.read().splitlines()
+
+        self._dump_lines(out, sys.stdout)
+        self._dump_lines(err, sys.stderr)
+
+        with contextlib.suppress(ValueError):
+            ret = ExitCode(ret)
+        return RunResult(ret, out, err, timing.time() - now)
+
+    def _dump_lines(self, lines, fp):
+        try:
+            for line in lines:
+                print(line, file=fp)
+        except UnicodeEncodeError:
+            print(f"couldn't print to {fp} because of encoding")
+
+    def _getpytestargs(self) -> tuple[str, ...]:
+        return sys.executable, "-mpytest"
+
+    def runpython(self, script: os.PathLike[str]) -> RunResult:
+        """Run a python script using sys.executable as interpreter."""
+        return self.run(sys.executable, script)
+
+    def runpython_c(self, command: str) -> RunResult:
+        """Run ``python -c "command"``."""
+        return self.run(sys.executable, "-c", command)
+
+    def runpytest_subprocess(
+        self, *args: str | os.PathLike[str], timeout: float | None = None
+    ) -> RunResult:
+        """Run pytest as a subprocess with given arguments.
+
+        Any plugins added to the :py:attr:`plugins` list will be added using the
+        ``-p`` command line option.  Additionally ``--basetemp`` is used to put
+        any temporary files and directories in a numbered directory prefixed
+        with "runpytest-" to not conflict with the normal numbered pytest
+        location for temporary files and directories.
+
+        :param args:
+            The sequence of arguments to pass to the pytest subprocess.
+        :param timeout:
+            The period in seconds after which to timeout and raise
+            :py:class:`Pytester.TimeoutExpired`.
+        :returns:
+            The result.
+        """
+        __tracebackhide__ = True
+        p = make_numbered_dir(root=self.path, prefix="runpytest-", mode=0o700)
+        args = (f"--basetemp={p}", *args)
+        plugins = [x for x in self.plugins if isinstance(x, str)]
+        if plugins:
+            args = ("-p", plugins[0], *args)
+        args = self._getpytestargs() + args
+        return self.run(*args, timeout=timeout)
+
+    def spawn_pytest(self, string: str, expect_timeout: float = 10.0) -> pexpect.spawn:
+        """Run pytest using pexpect.
+
+        This makes sure to use the right pytest and sets up the temporary
+        directory locations.
+
+        The pexpect child is returned.
+        """
+        basetemp = self.path / "temp-pexpect"
+        basetemp.mkdir(mode=0o700)
+        invoke = " ".join(map(str, self._getpytestargs()))
+        cmd = f"{invoke} --basetemp={basetemp} {string}"
+        return self.spawn(cmd, expect_timeout=expect_timeout)
+
+    def spawn(self, cmd: str, expect_timeout: float = 10.0) -> pexpect.spawn:
+        """Run a command using pexpect.
+
+        The pexpect child is returned.
+        """
+        pexpect = importorskip("pexpect", "3.0")
+        if hasattr(sys, "pypy_version_info") and "64" in platform.machine():
+            skip("pypy-64 bit not supported")
+        if not hasattr(pexpect, "spawn"):
+            skip("pexpect.spawn not available")
+        logfile = self.path.joinpath("spawn.out").open("wb")
+
+        child = pexpect.spawn(cmd, logfile=logfile, timeout=expect_timeout)
+        self._request.addfinalizer(logfile.close)
+        return child
+
+
+class LineComp:
+    def __init__(self) -> None:
+        self.stringio = StringIO()
+        """:class:`python:io.StringIO()` instance used for input."""
+
+    def assert_contains_lines(self, lines2: Sequence[str]) -> None:
+        """Assert that ``lines2`` are contained (linearly) in :attr:`stringio`'s value.
+
+        Lines are matched using :func:`LineMatcher.fnmatch_lines <pytest.LineMatcher.fnmatch_lines>`.
+        """
+        __tracebackhide__ = True
+        val = self.stringio.getvalue()
+        self.stringio.truncate(0)
+        self.stringio.seek(0)
+        lines1 = val.split("\n")
+        LineMatcher(lines1).fnmatch_lines(lines2)
+
+
+class LineMatcher:
+    """Flexible matching of text.
+
+    This is a convenience class to test large texts like the output of
+    commands.
+
+    The constructor takes a list of lines without their trailing newlines, i.e.
+    ``text.splitlines()``.
+    """
+
+    def __init__(self, lines: list[str]) -> None:
+        self.lines = lines
+        self._log_output: list[str] = []
+
+    def __str__(self) -> str:
+        """Return the entire original text.
+
+        .. versionadded:: 6.2
+            You can use :meth:`str` in older versions.
+        """
+        return "\n".join(self.lines)
+
+    def _getlines(self, lines2: str | Sequence[str] | Source) -> Sequence[str]:
+        if isinstance(lines2, str):
+            lines2 = Source(lines2)
+        if isinstance(lines2, Source):
+            lines2 = lines2.strip().lines
+        return lines2
+
+    def fnmatch_lines_random(self, lines2: Sequence[str]) -> None:
+        """Check lines exist in the output in any order (using :func:`python:fnmatch.fnmatch`)."""
+        __tracebackhide__ = True
+        self._match_lines_random(lines2, fnmatch)
+
+    def re_match_lines_random(self, lines2: Sequence[str]) -> None:
+        """Check lines exist in the output in any order (using :func:`python:re.match`)."""
+        __tracebackhide__ = True
+        self._match_lines_random(lines2, lambda name, pat: bool(re.match(pat, name)))
+
+    def _match_lines_random(
+        self, lines2: Sequence[str], match_func: Callable[[str, str], bool]
+    ) -> None:
+        __tracebackhide__ = True
+        lines2 = self._getlines(lines2)
+        for line in lines2:
+            for x in self.lines:
+                if line == x or match_func(x, line):
+                    self._log("matched: ", repr(line))
+                    break
+            else:
+                msg = f"line {line!r} not found in output"
+                self._log(msg)
+                self._fail(msg)
+
+    def get_lines_after(self, fnline: str) -> Sequence[str]:
+        """Return all lines following the given line in the text.
+
+        The given line can contain glob wildcards.
+        """
+        for i, line in enumerate(self.lines):
+            if fnline == line or fnmatch(line, fnline):
+                return self.lines[i + 1 :]
+        raise ValueError(f"line {fnline!r} not found in output")
+
+    def _log(self, *args) -> None:
+        self._log_output.append(" ".join(str(x) for x in args))
+
+    @property
+    def _log_text(self) -> str:
+        return "\n".join(self._log_output)
+
+    def fnmatch_lines(
+        self, lines2: Sequence[str], *, consecutive: bool = False
+    ) -> None:
+        """Check lines exist in the output (using :func:`python:fnmatch.fnmatch`).
+
+        The argument is a list of lines which have to match and can use glob
+        wildcards.  If they do not match a pytest.fail() is called.  The
+        matches and non-matches are also shown as part of the error message.
+
+        :param lines2: String patterns to match.
+        :param consecutive: Match lines consecutively?
+        """
+        __tracebackhide__ = True
+        self._match_lines(lines2, fnmatch, "fnmatch", consecutive=consecutive)
+
+    def re_match_lines(
+        self, lines2: Sequence[str], *, consecutive: bool = False
+    ) -> None:
+        """Check lines exist in the output (using :func:`python:re.match`).
+
+        The argument is a list of lines which have to match using ``re.match``.
+        If they do not match a pytest.fail() is called.
+
+        The matches and non-matches are also shown as part of the error message.
+
+        :param lines2: string patterns to match.
+        :param consecutive: match lines consecutively?
+        """
+        __tracebackhide__ = True
+        self._match_lines(
+            lines2,
+            lambda name, pat: bool(re.match(pat, name)),
+            "re.match",
+            consecutive=consecutive,
+        )
+
+    def _match_lines(
+        self,
+        lines2: Sequence[str],
+        match_func: Callable[[str, str], bool],
+        match_nickname: str,
+        *,
+        consecutive: bool = False,
+    ) -> None:
+        """Underlying implementation of ``fnmatch_lines`` and ``re_match_lines``.
+
+        :param Sequence[str] lines2:
+            List of string patterns to match. The actual format depends on
+            ``match_func``.
+        :param match_func:
+            A callable ``match_func(line, pattern)`` where line is the
+            captured line from stdout/stderr and pattern is the matching
+            pattern.
+        :param str match_nickname:
+            The nickname for the match function that will be logged to stdout
+            when a match occurs.
+        :param consecutive:
+            Match lines consecutively?
+        """
+        if not isinstance(lines2, collections.abc.Sequence):
+            raise TypeError(f"invalid type for lines2: {type(lines2).__name__}")
+        lines2 = self._getlines(lines2)
+        lines1 = self.lines[:]
+        extralines = []
+        __tracebackhide__ = True
+        wnick = len(match_nickname) + 1
+        started = False
+        for line in lines2:
+            nomatchprinted = False
+            while lines1:
+                nextline = lines1.pop(0)
+                if line == nextline:
+                    self._log("exact match:", repr(line))
+                    started = True
+                    break
+                elif match_func(nextline, line):
+                    self._log(f"{match_nickname}:", repr(line))
+                    self._log(
+                        "{:>{width}}".format("with:", width=wnick), repr(nextline)
+                    )
+                    started = True
+                    break
+                else:
+                    if consecutive and started:
+                        msg = f"no consecutive match: {line!r}"
+                        self._log(msg)
+                        self._log(
+                            "{:>{width}}".format("with:", width=wnick), repr(nextline)
+                        )
+                        self._fail(msg)
+                    if not nomatchprinted:
+                        self._log(
+                            "{:>{width}}".format("nomatch:", width=wnick), repr(line)
+                        )
+                        nomatchprinted = True
+                    self._log("{:>{width}}".format("and:", width=wnick), repr(nextline))
+                extralines.append(nextline)
+            else:
+                msg = f"remains unmatched: {line!r}"
+                self._log(msg)
+                self._fail(msg)
+        self._log_output = []
+
+    def no_fnmatch_line(self, pat: str) -> None:
+        """Ensure captured lines do not match the given pattern, using ``fnmatch.fnmatch``.
+
+        :param str pat: The pattern to match lines.
+        """
+        __tracebackhide__ = True
+        self._no_match_line(pat, fnmatch, "fnmatch")
+
+    def no_re_match_line(self, pat: str) -> None:
+        """Ensure captured lines do not match the given pattern, using ``re.match``.
+
+        :param str pat: The regular expression to match lines.
+        """
+        __tracebackhide__ = True
+        self._no_match_line(
+            pat, lambda name, pat: bool(re.match(pat, name)), "re.match"
+        )
+
+    def _no_match_line(
+        self, pat: str, match_func: Callable[[str, str], bool], match_nickname: str
+    ) -> None:
+        """Ensure captured lines does not have a the given pattern, using ``fnmatch.fnmatch``.
+
+        :param str pat: The pattern to match lines.
+        """
+        __tracebackhide__ = True
+        nomatch_printed = False
+        wnick = len(match_nickname) + 1
+        for line in self.lines:
+            if match_func(line, pat):
+                msg = f"{match_nickname}: {pat!r}"
+                self._log(msg)
+                self._log("{:>{width}}".format("with:", width=wnick), repr(line))
+                self._fail(msg)
+            else:
+                if not nomatch_printed:
+                    self._log("{:>{width}}".format("nomatch:", width=wnick), repr(pat))
+                    nomatch_printed = True
+                self._log("{:>{width}}".format("and:", width=wnick), repr(line))
+        self._log_output = []
+
+    def _fail(self, msg: str) -> None:
+        __tracebackhide__ = True
+        log_text = self._log_text
+        self._log_output = []
+        fail(log_text)
+
+    def str(self) -> str:
+        """Return the entire original text."""
+        return str(self)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/pytester_assertions.py b/backend/venv/lib/python3.12/site-packages/_pytest/pytester_assertions.py
new file mode 100644
index 00000000..d543798f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/pytester_assertions.py
@@ -0,0 +1,74 @@
+"""Helper plugin for pytester; should not be loaded on its own."""
+
+# This plugin contains assertions used by pytester. pytester cannot
+# contain them itself, since it is imported by the `pytest` module,
+# hence cannot be subject to assertion rewriting, which requires a
+# module to not be already imported.
+from __future__ import annotations
+
+from typing import Sequence
+
+from _pytest.reports import CollectReport
+from _pytest.reports import TestReport
+
+
+def assertoutcome(
+    outcomes: tuple[
+        Sequence[TestReport],
+        Sequence[CollectReport | TestReport],
+        Sequence[CollectReport | TestReport],
+    ],
+    passed: int = 0,
+    skipped: int = 0,
+    failed: int = 0,
+) -> None:
+    __tracebackhide__ = True
+
+    realpassed, realskipped, realfailed = outcomes
+    obtained = {
+        "passed": len(realpassed),
+        "skipped": len(realskipped),
+        "failed": len(realfailed),
+    }
+    expected = {"passed": passed, "skipped": skipped, "failed": failed}
+    assert obtained == expected, outcomes
+
+
+def assert_outcomes(
+    outcomes: dict[str, int],
+    passed: int = 0,
+    skipped: int = 0,
+    failed: int = 0,
+    errors: int = 0,
+    xpassed: int = 0,
+    xfailed: int = 0,
+    warnings: int | None = None,
+    deselected: int | None = None,
+) -> None:
+    """Assert that the specified outcomes appear with the respective
+    numbers (0 means it didn't occur) in the text output from a test run."""
+    __tracebackhide__ = True
+
+    obtained = {
+        "passed": outcomes.get("passed", 0),
+        "skipped": outcomes.get("skipped", 0),
+        "failed": outcomes.get("failed", 0),
+        "errors": outcomes.get("errors", 0),
+        "xpassed": outcomes.get("xpassed", 0),
+        "xfailed": outcomes.get("xfailed", 0),
+    }
+    expected = {
+        "passed": passed,
+        "skipped": skipped,
+        "failed": failed,
+        "errors": errors,
+        "xpassed": xpassed,
+        "xfailed": xfailed,
+    }
+    if warnings is not None:
+        obtained["warnings"] = outcomes.get("warnings", 0)
+        expected["warnings"] = warnings
+    if deselected is not None:
+        obtained["deselected"] = outcomes.get("deselected", 0)
+        expected["deselected"] = deselected
+    assert obtained == expected
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/python.py b/backend/venv/lib/python3.12/site-packages/_pytest/python.py
new file mode 100644
index 00000000..3478c34c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/python.py
@@ -0,0 +1,1679 @@
+# mypy: allow-untyped-defs
+"""Python test discovery, setup and run of test functions."""
+
+from __future__ import annotations
+
+import abc
+from collections import Counter
+from collections import defaultdict
+import dataclasses
+import enum
+import fnmatch
+from functools import partial
+import inspect
+import itertools
+import os
+from pathlib import Path
+import types
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import final
+from typing import Generator
+from typing import Iterable
+from typing import Iterator
+from typing import Literal
+from typing import Mapping
+from typing import Pattern
+from typing import Sequence
+from typing import TYPE_CHECKING
+import warnings
+
+import _pytest
+from _pytest import fixtures
+from _pytest import nodes
+from _pytest._code import filter_traceback
+from _pytest._code import getfslineno
+from _pytest._code.code import ExceptionInfo
+from _pytest._code.code import TerminalRepr
+from _pytest._code.code import Traceback
+from _pytest._io.saferepr import saferepr
+from _pytest.compat import ascii_escaped
+from _pytest.compat import get_default_arg_names
+from _pytest.compat import get_real_func
+from _pytest.compat import getimfunc
+from _pytest.compat import is_async_function
+from _pytest.compat import is_generator
+from _pytest.compat import LEGACY_PATH
+from _pytest.compat import NOTSET
+from _pytest.compat import safe_getattr
+from _pytest.compat import safe_isclass
+from _pytest.config import Config
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import FixtureDef
+from _pytest.fixtures import FixtureRequest
+from _pytest.fixtures import FuncFixtureInfo
+from _pytest.fixtures import get_scope_node
+from _pytest.main import Session
+from _pytest.mark import MARK_GEN
+from _pytest.mark import ParameterSet
+from _pytest.mark.structures import get_unpacked_marks
+from _pytest.mark.structures import Mark
+from _pytest.mark.structures import MarkDecorator
+from _pytest.mark.structures import normalize_mark_list
+from _pytest.outcomes import fail
+from _pytest.outcomes import skip
+from _pytest.pathlib import fnmatch_ex
+from _pytest.pathlib import import_path
+from _pytest.pathlib import ImportPathMismatchError
+from _pytest.pathlib import scandir
+from _pytest.scope import _ScopeName
+from _pytest.scope import Scope
+from _pytest.stash import StashKey
+from _pytest.warning_types import PytestCollectionWarning
+from _pytest.warning_types import PytestReturnNotNoneWarning
+from _pytest.warning_types import PytestUnhandledCoroutineWarning
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini(
+        "python_files",
+        type="args",
+        # NOTE: default is also used in AssertionRewritingHook.
+        default=["test_*.py", "*_test.py"],
+        help="Glob-style file patterns for Python test module discovery",
+    )
+    parser.addini(
+        "python_classes",
+        type="args",
+        default=["Test"],
+        help="Prefixes or glob names for Python test class discovery",
+    )
+    parser.addini(
+        "python_functions",
+        type="args",
+        default=["test"],
+        help="Prefixes or glob names for Python test function and method discovery",
+    )
+    parser.addini(
+        "disable_test_id_escaping_and_forfeit_all_rights_to_community_support",
+        type="bool",
+        default=False,
+        help="Disable string escape non-ASCII characters, might cause unwanted "
+        "side effects(use at your own risk)",
+    )
+
+
+def pytest_generate_tests(metafunc: Metafunc) -> None:
+    for marker in metafunc.definition.iter_markers(name="parametrize"):
+        metafunc.parametrize(*marker.args, **marker.kwargs, _param_mark=marker)
+
+
+def pytest_configure(config: Config) -> None:
+    config.addinivalue_line(
+        "markers",
+        "parametrize(argnames, argvalues): call a test function multiple "
+        "times passing in different arguments in turn. argvalues generally "
+        "needs to be a list of values if argnames specifies only one name "
+        "or a list of tuples of values if argnames specifies multiple names. "
+        "Example: @parametrize('arg1', [1,2]) would lead to two calls of the "
+        "decorated test function, one with arg1=1 and another with arg1=2."
+        "see https://docs.pytest.org/en/stable/how-to/parametrize.html for more info "
+        "and examples.",
+    )
+    config.addinivalue_line(
+        "markers",
+        "usefixtures(fixturename1, fixturename2, ...): mark tests as needing "
+        "all of the specified fixtures. see "
+        "https://docs.pytest.org/en/stable/explanation/fixtures.html#usefixtures ",
+    )
+
+
+def async_warn_and_skip(nodeid: str) -> None:
+    msg = "async def functions are not natively supported and have been skipped.\n"
+    msg += (
+        "You need to install a suitable plugin for your async framework, for example:\n"
+    )
+    msg += "  - anyio\n"
+    msg += "  - pytest-asyncio\n"
+    msg += "  - pytest-tornasync\n"
+    msg += "  - pytest-trio\n"
+    msg += "  - pytest-twisted"
+    warnings.warn(PytestUnhandledCoroutineWarning(msg.format(nodeid)))
+    skip(reason="async def function and no async plugin installed (see warnings)")
+
+
+@hookimpl(trylast=True)
+def pytest_pyfunc_call(pyfuncitem: Function) -> object | None:
+    testfunction = pyfuncitem.obj
+    if is_async_function(testfunction):
+        async_warn_and_skip(pyfuncitem.nodeid)
+    funcargs = pyfuncitem.funcargs
+    testargs = {arg: funcargs[arg] for arg in pyfuncitem._fixtureinfo.argnames}
+    result = testfunction(**testargs)
+    if hasattr(result, "__await__") or hasattr(result, "__aiter__"):
+        async_warn_and_skip(pyfuncitem.nodeid)
+    elif result is not None:
+        warnings.warn(
+            PytestReturnNotNoneWarning(
+                f"Expected None, but {pyfuncitem.nodeid} returned {result!r}, which will be an error in a "
+                "future version of pytest.  Did you mean to use `assert` instead of `return`?"
+            )
+        )
+    return True
+
+
+def pytest_collect_directory(
+    path: Path, parent: nodes.Collector
+) -> nodes.Collector | None:
+    pkginit = path / "__init__.py"
+    try:
+        has_pkginit = pkginit.is_file()
+    except PermissionError:
+        # See https://github.com/pytest-dev/pytest/issues/12120#issuecomment-2106349096.
+        return None
+    if has_pkginit:
+        return Package.from_parent(parent, path=path)
+    return None
+
+
+def pytest_collect_file(file_path: Path, parent: nodes.Collector) -> Module | None:
+    if file_path.suffix == ".py":
+        if not parent.session.isinitpath(file_path):
+            if not path_matches_patterns(
+                file_path, parent.config.getini("python_files")
+            ):
+                return None
+        ihook = parent.session.gethookproxy(file_path)
+        module: Module = ihook.pytest_pycollect_makemodule(
+            module_path=file_path, parent=parent
+        )
+        return module
+    return None
+
+
+def path_matches_patterns(path: Path, patterns: Iterable[str]) -> bool:
+    """Return whether path matches any of the patterns in the list of globs given."""
+    return any(fnmatch_ex(pattern, path) for pattern in patterns)
+
+
+def pytest_pycollect_makemodule(module_path: Path, parent) -> Module:
+    return Module.from_parent(parent, path=module_path)
+
+
+@hookimpl(trylast=True)
+def pytest_pycollect_makeitem(
+    collector: Module | Class, name: str, obj: object
+) -> None | nodes.Item | nodes.Collector | list[nodes.Item | nodes.Collector]:
+    assert isinstance(collector, (Class, Module)), type(collector)
+    # Nothing was collected elsewhere, let's do it here.
+    if safe_isclass(obj):
+        if collector.istestclass(obj, name):
+            return Class.from_parent(collector, name=name, obj=obj)
+    elif collector.istestfunction(obj, name):
+        # mock seems to store unbound methods (issue473), normalize it.
+        obj = getattr(obj, "__func__", obj)
+        # We need to try and unwrap the function if it's a functools.partial
+        # or a functools.wrapped.
+        # We mustn't if it's been wrapped with mock.patch (python 2 only).
+        if not (inspect.isfunction(obj) or inspect.isfunction(get_real_func(obj))):
+            filename, lineno = getfslineno(obj)
+            warnings.warn_explicit(
+                message=PytestCollectionWarning(
+                    f"cannot collect {name!r} because it is not a function."
+                ),
+                category=None,
+                filename=str(filename),
+                lineno=lineno + 1,
+            )
+        elif getattr(obj, "__test__", True):
+            if is_generator(obj):
+                res = Function.from_parent(collector, name=name)
+                reason = (
+                    f"yield tests were removed in pytest 4.0 - {name} will be ignored"
+                )
+                res.add_marker(MARK_GEN.xfail(run=False, reason=reason))
+                res.warn(PytestCollectionWarning(reason))
+                return res
+            else:
+                return list(collector._genfunctions(name, obj))
+    return None
+
+
+class PyobjMixin(nodes.Node):
+    """this mix-in inherits from Node to carry over the typing information
+
+    as its intended to always mix in before a node
+    its position in the mro is unaffected"""
+
+    _ALLOW_MARKERS = True
+
+    @property
+    def module(self):
+        """Python module object this node was collected from (can be None)."""
+        node = self.getparent(Module)
+        return node.obj if node is not None else None
+
+    @property
+    def cls(self):
+        """Python class object this node was collected from (can be None)."""
+        node = self.getparent(Class)
+        return node.obj if node is not None else None
+
+    @property
+    def instance(self):
+        """Python instance object the function is bound to.
+
+        Returns None if not a test method, e.g. for a standalone test function,
+        a class or a module.
+        """
+        # Overridden by Function.
+        return None
+
+    @property
+    def obj(self):
+        """Underlying Python object."""
+        obj = getattr(self, "_obj", None)
+        if obj is None:
+            self._obj = obj = self._getobj()
+            # XXX evil hack
+            # used to avoid Function marker duplication
+            if self._ALLOW_MARKERS:
+                self.own_markers.extend(get_unpacked_marks(self.obj))
+                # This assumes that `obj` is called before there is a chance
+                # to add custom keys to `self.keywords`, so no fear of overriding.
+                self.keywords.update((mark.name, mark) for mark in self.own_markers)
+        return obj
+
+    @obj.setter
+    def obj(self, value):
+        self._obj = value
+
+    def _getobj(self):
+        """Get the underlying Python object. May be overwritten by subclasses."""
+        # TODO: Improve the type of `parent` such that assert/ignore aren't needed.
+        assert self.parent is not None
+        obj = self.parent.obj  # type: ignore[attr-defined]
+        return getattr(obj, self.name)
+
+    def getmodpath(self, stopatmodule: bool = True, includemodule: bool = False) -> str:
+        """Return Python path relative to the containing module."""
+        parts = []
+        for node in self.iter_parents():
+            name = node.name
+            if isinstance(node, Module):
+                name = os.path.splitext(name)[0]
+                if stopatmodule:
+                    if includemodule:
+                        parts.append(name)
+                    break
+            parts.append(name)
+        parts.reverse()
+        return ".".join(parts)
+
+    def reportinfo(self) -> tuple[os.PathLike[str] | str, int | None, str]:
+        # XXX caching?
+        path, lineno = getfslineno(self.obj)
+        modpath = self.getmodpath()
+        return path, lineno, modpath
+
+
+# As an optimization, these builtin attribute names are pre-ignored when
+# iterating over an object during collection -- the pytest_pycollect_makeitem
+# hook is not called for them.
+# fmt: off
+class _EmptyClass: pass  # noqa: E701
+IGNORED_ATTRIBUTES = frozenset.union(
+    frozenset(),
+    # Module.
+    dir(types.ModuleType("empty_module")),
+    # Some extra module attributes the above doesn't catch.
+    {"__builtins__", "__file__", "__cached__"},
+    # Class.
+    dir(_EmptyClass),
+    # Instance.
+    dir(_EmptyClass()),
+)
+del _EmptyClass
+# fmt: on
+
+
+class PyCollector(PyobjMixin, nodes.Collector, abc.ABC):
+    def funcnamefilter(self, name: str) -> bool:
+        return self._matches_prefix_or_glob_option("python_functions", name)
+
+    def isnosetest(self, obj: object) -> bool:
+        """Look for the __test__ attribute, which is applied by the
+        @nose.tools.istest decorator.
+        """
+        # We explicitly check for "is True" here to not mistakenly treat
+        # classes with a custom __getattr__ returning something truthy (like a
+        # function) as test classes.
+        return safe_getattr(obj, "__test__", False) is True
+
+    def classnamefilter(self, name: str) -> bool:
+        return self._matches_prefix_or_glob_option("python_classes", name)
+
+    def istestfunction(self, obj: object, name: str) -> bool:
+        if self.funcnamefilter(name) or self.isnosetest(obj):
+            if isinstance(obj, (staticmethod, classmethod)):
+                # staticmethods and classmethods need to be unwrapped.
+                obj = safe_getattr(obj, "__func__", False)
+            return callable(obj) and fixtures.getfixturemarker(obj) is None
+        else:
+            return False
+
+    def istestclass(self, obj: object, name: str) -> bool:
+        if not (self.classnamefilter(name) or self.isnosetest(obj)):
+            return False
+        if inspect.isabstract(obj):
+            return False
+        return True
+
+    def _matches_prefix_or_glob_option(self, option_name: str, name: str) -> bool:
+        """Check if the given name matches the prefix or glob-pattern defined
+        in ini configuration."""
+        for option in self.config.getini(option_name):
+            if name.startswith(option):
+                return True
+            # Check that name looks like a glob-string before calling fnmatch
+            # because this is called for every name in each collected module,
+            # and fnmatch is somewhat expensive to call.
+            elif ("*" in option or "?" in option or "[" in option) and fnmatch.fnmatch(
+                name, option
+            ):
+                return True
+        return False
+
+    def collect(self) -> Iterable[nodes.Item | nodes.Collector]:
+        if not getattr(self.obj, "__test__", True):
+            return []
+
+        # Avoid random getattrs and peek in the __dict__ instead.
+        dicts = [getattr(self.obj, "__dict__", {})]
+        if isinstance(self.obj, type):
+            for basecls in self.obj.__mro__:
+                dicts.append(basecls.__dict__)
+
+        # In each class, nodes should be definition ordered.
+        # __dict__ is definition ordered.
+        seen: set[str] = set()
+        dict_values: list[list[nodes.Item | nodes.Collector]] = []
+        ihook = self.ihook
+        for dic in dicts:
+            values: list[nodes.Item | nodes.Collector] = []
+            # Note: seems like the dict can change during iteration -
+            # be careful not to remove the list() without consideration.
+            for name, obj in list(dic.items()):
+                if name in IGNORED_ATTRIBUTES:
+                    continue
+                if name in seen:
+                    continue
+                seen.add(name)
+                res = ihook.pytest_pycollect_makeitem(
+                    collector=self, name=name, obj=obj
+                )
+                if res is None:
+                    continue
+                elif isinstance(res, list):
+                    values.extend(res)
+                else:
+                    values.append(res)
+            dict_values.append(values)
+
+        # Between classes in the class hierarchy, reverse-MRO order -- nodes
+        # inherited from base classes should come before subclasses.
+        result = []
+        for values in reversed(dict_values):
+            result.extend(values)
+        return result
+
+    def _genfunctions(self, name: str, funcobj) -> Iterator[Function]:
+        modulecol = self.getparent(Module)
+        assert modulecol is not None
+        module = modulecol.obj
+        clscol = self.getparent(Class)
+        cls = clscol and clscol.obj or None
+
+        definition = FunctionDefinition.from_parent(self, name=name, callobj=funcobj)
+        fixtureinfo = definition._fixtureinfo
+
+        # pytest_generate_tests impls call metafunc.parametrize() which fills
+        # metafunc._calls, the outcome of the hook.
+        metafunc = Metafunc(
+            definition=definition,
+            fixtureinfo=fixtureinfo,
+            config=self.config,
+            cls=cls,
+            module=module,
+            _ispytest=True,
+        )
+        methods = []
+        if hasattr(module, "pytest_generate_tests"):
+            methods.append(module.pytest_generate_tests)
+        if cls is not None and hasattr(cls, "pytest_generate_tests"):
+            methods.append(cls().pytest_generate_tests)
+        self.ihook.pytest_generate_tests.call_extra(methods, dict(metafunc=metafunc))
+
+        if not metafunc._calls:
+            yield Function.from_parent(self, name=name, fixtureinfo=fixtureinfo)
+        else:
+            # Direct parametrizations taking place in module/class-specific
+            # `metafunc.parametrize` calls may have shadowed some fixtures, so make sure
+            # we update what the function really needs a.k.a its fixture closure. Note that
+            # direct parametrizations using `@pytest.mark.parametrize` have already been considered
+            # into making the closure using `ignore_args` arg to `getfixtureclosure`.
+            fixtureinfo.prune_dependency_tree()
+
+            for callspec in metafunc._calls:
+                subname = f"{name}[{callspec.id}]"
+                yield Function.from_parent(
+                    self,
+                    name=subname,
+                    callspec=callspec,
+                    fixtureinfo=fixtureinfo,
+                    keywords={callspec.id: True},
+                    originalname=name,
+                )
+
+
+def importtestmodule(
+    path: Path,
+    config: Config,
+):
+    # We assume we are only called once per module.
+    importmode = config.getoption("--import-mode")
+    try:
+        mod = import_path(
+            path,
+            mode=importmode,
+            root=config.rootpath,
+            consider_namespace_packages=config.getini("consider_namespace_packages"),
+        )
+    except SyntaxError as e:
+        raise nodes.Collector.CollectError(
+            ExceptionInfo.from_current().getrepr(style="short")
+        ) from e
+    except ImportPathMismatchError as e:
+        raise nodes.Collector.CollectError(
+            "import file mismatch:\n"
+            "imported module {!r} has this __file__ attribute:\n"
+            "  {}\n"
+            "which is not the same as the test file we want to collect:\n"
+            "  {}\n"
+            "HINT: remove __pycache__ / .pyc files and/or use a "
+            "unique basename for your test file modules".format(*e.args)
+        ) from e
+    except ImportError as e:
+        exc_info = ExceptionInfo.from_current()
+        if config.get_verbosity() < 2:
+            exc_info.traceback = exc_info.traceback.filter(filter_traceback)
+        exc_repr = (
+            exc_info.getrepr(style="short")
+            if exc_info.traceback
+            else exc_info.exconly()
+        )
+        formatted_tb = str(exc_repr)
+        raise nodes.Collector.CollectError(
+            f"ImportError while importing test module '{path}'.\n"
+            "Hint: make sure your test modules/packages have valid Python names.\n"
+            "Traceback:\n"
+            f"{formatted_tb}"
+        ) from e
+    except skip.Exception as e:
+        if e.allow_module_level:
+            raise
+        raise nodes.Collector.CollectError(
+            "Using pytest.skip outside of a test will skip the entire module. "
+            "If that's your intention, pass `allow_module_level=True`. "
+            "If you want to skip a specific test or an entire class, "
+            "use the @pytest.mark.skip or @pytest.mark.skipif decorators."
+        ) from e
+    config.pluginmanager.consider_module(mod)
+    return mod
+
+
+class Module(nodes.File, PyCollector):
+    """Collector for test classes and functions in a Python module."""
+
+    def _getobj(self):
+        return importtestmodule(self.path, self.config)
+
+    def collect(self) -> Iterable[nodes.Item | nodes.Collector]:
+        self._register_setup_module_fixture()
+        self._register_setup_function_fixture()
+        self.session._fixturemanager.parsefactories(self)
+        return super().collect()
+
+    def _register_setup_module_fixture(self) -> None:
+        """Register an autouse, module-scoped fixture for the collected module object
+        that invokes setUpModule/tearDownModule if either or both are available.
+
+        Using a fixture to invoke this methods ensures we play nicely and unsurprisingly with
+        other fixtures (#517).
+        """
+        setup_module = _get_first_non_fixture_func(
+            self.obj, ("setUpModule", "setup_module")
+        )
+        teardown_module = _get_first_non_fixture_func(
+            self.obj, ("tearDownModule", "teardown_module")
+        )
+
+        if setup_module is None and teardown_module is None:
+            return
+
+        def xunit_setup_module_fixture(request) -> Generator[None]:
+            module = request.module
+            if setup_module is not None:
+                _call_with_optional_argument(setup_module, module)
+            yield
+            if teardown_module is not None:
+                _call_with_optional_argument(teardown_module, module)
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_xunit_setup_module_fixture_{self.obj.__name__}",
+            func=xunit_setup_module_fixture,
+            nodeid=self.nodeid,
+            scope="module",
+            autouse=True,
+        )
+
+    def _register_setup_function_fixture(self) -> None:
+        """Register an autouse, function-scoped fixture for the collected module object
+        that invokes setup_function/teardown_function if either or both are available.
+
+        Using a fixture to invoke this methods ensures we play nicely and unsurprisingly with
+        other fixtures (#517).
+        """
+        setup_function = _get_first_non_fixture_func(self.obj, ("setup_function",))
+        teardown_function = _get_first_non_fixture_func(
+            self.obj, ("teardown_function",)
+        )
+        if setup_function is None and teardown_function is None:
+            return
+
+        def xunit_setup_function_fixture(request) -> Generator[None]:
+            if request.instance is not None:
+                # in this case we are bound to an instance, so we need to let
+                # setup_method handle this
+                yield
+                return
+            function = request.function
+            if setup_function is not None:
+                _call_with_optional_argument(setup_function, function)
+            yield
+            if teardown_function is not None:
+                _call_with_optional_argument(teardown_function, function)
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_xunit_setup_function_fixture_{self.obj.__name__}",
+            func=xunit_setup_function_fixture,
+            nodeid=self.nodeid,
+            scope="function",
+            autouse=True,
+        )
+
+
+class Package(nodes.Directory):
+    """Collector for files and directories in a Python packages -- directories
+    with an `__init__.py` file.
+
+    .. note::
+
+        Directories without an `__init__.py` file are instead collected by
+        :class:`~pytest.Dir` by default. Both are :class:`~pytest.Directory`
+        collectors.
+
+    .. versionchanged:: 8.0
+
+        Now inherits from :class:`~pytest.Directory`.
+    """
+
+    def __init__(
+        self,
+        fspath: LEGACY_PATH | None,
+        parent: nodes.Collector,
+        # NOTE: following args are unused:
+        config=None,
+        session=None,
+        nodeid=None,
+        path: Path | None = None,
+    ) -> None:
+        # NOTE: Could be just the following, but kept as-is for compat.
+        # super().__init__(self, fspath, parent=parent)
+        session = parent.session
+        super().__init__(
+            fspath=fspath,
+            path=path,
+            parent=parent,
+            config=config,
+            session=session,
+            nodeid=nodeid,
+        )
+
+    def setup(self) -> None:
+        init_mod = importtestmodule(self.path / "__init__.py", self.config)
+
+        # Not using fixtures to call setup_module here because autouse fixtures
+        # from packages are not called automatically (#4085).
+        setup_module = _get_first_non_fixture_func(
+            init_mod, ("setUpModule", "setup_module")
+        )
+        if setup_module is not None:
+            _call_with_optional_argument(setup_module, init_mod)
+
+        teardown_module = _get_first_non_fixture_func(
+            init_mod, ("tearDownModule", "teardown_module")
+        )
+        if teardown_module is not None:
+            func = partial(_call_with_optional_argument, teardown_module, init_mod)
+            self.addfinalizer(func)
+
+    def collect(self) -> Iterable[nodes.Item | nodes.Collector]:
+        # Always collect __init__.py first.
+        def sort_key(entry: os.DirEntry[str]) -> object:
+            return (entry.name != "__init__.py", entry.name)
+
+        config = self.config
+        col: nodes.Collector | None
+        cols: Sequence[nodes.Collector]
+        ihook = self.ihook
+        for direntry in scandir(self.path, sort_key):
+            if direntry.is_dir():
+                path = Path(direntry.path)
+                if not self.session.isinitpath(path, with_parents=True):
+                    if ihook.pytest_ignore_collect(collection_path=path, config=config):
+                        continue
+                col = ihook.pytest_collect_directory(path=path, parent=self)
+                if col is not None:
+                    yield col
+
+            elif direntry.is_file():
+                path = Path(direntry.path)
+                if not self.session.isinitpath(path):
+                    if ihook.pytest_ignore_collect(collection_path=path, config=config):
+                        continue
+                cols = ihook.pytest_collect_file(file_path=path, parent=self)
+                yield from cols
+
+
+def _call_with_optional_argument(func, arg) -> None:
+    """Call the given function with the given argument if func accepts one argument, otherwise
+    calls func without arguments."""
+    arg_count = func.__code__.co_argcount
+    if inspect.ismethod(func):
+        arg_count -= 1
+    if arg_count:
+        func(arg)
+    else:
+        func()
+
+
+def _get_first_non_fixture_func(obj: object, names: Iterable[str]) -> object | None:
+    """Return the attribute from the given object to be used as a setup/teardown
+    xunit-style function, but only if not marked as a fixture to avoid calling it twice.
+    """
+    for name in names:
+        meth: object | None = getattr(obj, name, None)
+        if meth is not None and fixtures.getfixturemarker(meth) is None:
+            return meth
+    return None
+
+
+class Class(PyCollector):
+    """Collector for test methods (and nested classes) in a Python class."""
+
+    @classmethod
+    def from_parent(cls, parent, *, name, obj=None, **kw) -> Self:  # type: ignore[override]
+        """The public constructor."""
+        return super().from_parent(name=name, parent=parent, **kw)
+
+    def newinstance(self):
+        return self.obj()
+
+    def collect(self) -> Iterable[nodes.Item | nodes.Collector]:
+        if not safe_getattr(self.obj, "__test__", True):
+            return []
+        if hasinit(self.obj):
+            assert self.parent is not None
+            self.warn(
+                PytestCollectionWarning(
+                    f"cannot collect test class {self.obj.__name__!r} because it has a "
+                    f"__init__ constructor (from: {self.parent.nodeid})"
+                )
+            )
+            return []
+        elif hasnew(self.obj):
+            assert self.parent is not None
+            self.warn(
+                PytestCollectionWarning(
+                    f"cannot collect test class {self.obj.__name__!r} because it has a "
+                    f"__new__ constructor (from: {self.parent.nodeid})"
+                )
+            )
+            return []
+
+        self._register_setup_class_fixture()
+        self._register_setup_method_fixture()
+
+        self.session._fixturemanager.parsefactories(self.newinstance(), self.nodeid)
+
+        return super().collect()
+
+    def _register_setup_class_fixture(self) -> None:
+        """Register an autouse, class scoped fixture into the collected class object
+        that invokes setup_class/teardown_class if either or both are available.
+
+        Using a fixture to invoke this methods ensures we play nicely and unsurprisingly with
+        other fixtures (#517).
+        """
+        setup_class = _get_first_non_fixture_func(self.obj, ("setup_class",))
+        teardown_class = _get_first_non_fixture_func(self.obj, ("teardown_class",))
+        if setup_class is None and teardown_class is None:
+            return
+
+        def xunit_setup_class_fixture(request) -> Generator[None]:
+            cls = request.cls
+            if setup_class is not None:
+                func = getimfunc(setup_class)
+                _call_with_optional_argument(func, cls)
+            yield
+            if teardown_class is not None:
+                func = getimfunc(teardown_class)
+                _call_with_optional_argument(func, cls)
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_xunit_setup_class_fixture_{self.obj.__qualname__}",
+            func=xunit_setup_class_fixture,
+            nodeid=self.nodeid,
+            scope="class",
+            autouse=True,
+        )
+
+    def _register_setup_method_fixture(self) -> None:
+        """Register an autouse, function scoped fixture into the collected class object
+        that invokes setup_method/teardown_method if either or both are available.
+
+        Using a fixture to invoke these methods ensures we play nicely and unsurprisingly with
+        other fixtures (#517).
+        """
+        setup_name = "setup_method"
+        setup_method = _get_first_non_fixture_func(self.obj, (setup_name,))
+        teardown_name = "teardown_method"
+        teardown_method = _get_first_non_fixture_func(self.obj, (teardown_name,))
+        if setup_method is None and teardown_method is None:
+            return
+
+        def xunit_setup_method_fixture(request) -> Generator[None]:
+            instance = request.instance
+            method = request.function
+            if setup_method is not None:
+                func = getattr(instance, setup_name)
+                _call_with_optional_argument(func, method)
+            yield
+            if teardown_method is not None:
+                func = getattr(instance, teardown_name)
+                _call_with_optional_argument(func, method)
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_xunit_setup_method_fixture_{self.obj.__qualname__}",
+            func=xunit_setup_method_fixture,
+            nodeid=self.nodeid,
+            scope="function",
+            autouse=True,
+        )
+
+
+def hasinit(obj: object) -> bool:
+    init: object = getattr(obj, "__init__", None)
+    if init:
+        return init != object.__init__
+    return False
+
+
+def hasnew(obj: object) -> bool:
+    new: object = getattr(obj, "__new__", None)
+    if new:
+        return new != object.__new__
+    return False
+
+
+@final
+@dataclasses.dataclass(frozen=True)
+class IdMaker:
+    """Make IDs for a parametrization."""
+
+    __slots__ = (
+        "argnames",
+        "parametersets",
+        "idfn",
+        "ids",
+        "config",
+        "nodeid",
+        "func_name",
+    )
+
+    # The argnames of the parametrization.
+    argnames: Sequence[str]
+    # The ParameterSets of the parametrization.
+    parametersets: Sequence[ParameterSet]
+    # Optionally, a user-provided callable to make IDs for parameters in a
+    # ParameterSet.
+    idfn: Callable[[Any], object | None] | None
+    # Optionally, explicit IDs for ParameterSets by index.
+    ids: Sequence[object | None] | None
+    # Optionally, the pytest config.
+    # Used for controlling ASCII escaping, and for calling the
+    # :hook:`pytest_make_parametrize_id` hook.
+    config: Config | None
+    # Optionally, the ID of the node being parametrized.
+    # Used only for clearer error messages.
+    nodeid: str | None
+    # Optionally, the ID of the function being parametrized.
+    # Used only for clearer error messages.
+    func_name: str | None
+
+    def make_unique_parameterset_ids(self) -> list[str]:
+        """Make a unique identifier for each ParameterSet, that may be used to
+        identify the parametrization in a node ID.
+
+        Format is <prm_1_token>-...-<prm_n_token>[counter], where prm_x_token is
+        - user-provided id, if given
+        - else an id derived from the value, applicable for certain types
+        - else <argname><parameterset index>
+        The counter suffix is appended only in case a string wouldn't be unique
+        otherwise.
+        """
+        resolved_ids = list(self._resolve_ids())
+        # All IDs must be unique!
+        if len(resolved_ids) != len(set(resolved_ids)):
+            # Record the number of occurrences of each ID.
+            id_counts = Counter(resolved_ids)
+            # Map the ID to its next suffix.
+            id_suffixes: dict[str, int] = defaultdict(int)
+            # Suffix non-unique IDs to make them unique.
+            for index, id in enumerate(resolved_ids):
+                if id_counts[id] > 1:
+                    suffix = ""
+                    if id and id[-1].isdigit():
+                        suffix = "_"
+                    new_id = f"{id}{suffix}{id_suffixes[id]}"
+                    while new_id in set(resolved_ids):
+                        id_suffixes[id] += 1
+                        new_id = f"{id}{suffix}{id_suffixes[id]}"
+                    resolved_ids[index] = new_id
+                    id_suffixes[id] += 1
+        assert len(resolved_ids) == len(
+            set(resolved_ids)
+        ), f"Internal error: {resolved_ids=}"
+        return resolved_ids
+
+    def _resolve_ids(self) -> Iterable[str]:
+        """Resolve IDs for all ParameterSets (may contain duplicates)."""
+        for idx, parameterset in enumerate(self.parametersets):
+            if parameterset.id is not None:
+                # ID provided directly - pytest.param(..., id="...")
+                yield parameterset.id
+            elif self.ids and idx < len(self.ids) and self.ids[idx] is not None:
+                # ID provided in the IDs list - parametrize(..., ids=[...]).
+                yield self._idval_from_value_required(self.ids[idx], idx)
+            else:
+                # ID not provided - generate it.
+                yield "-".join(
+                    self._idval(val, argname, idx)
+                    for val, argname in zip(parameterset.values, self.argnames)
+                )
+
+    def _idval(self, val: object, argname: str, idx: int) -> str:
+        """Make an ID for a parameter in a ParameterSet."""
+        idval = self._idval_from_function(val, argname, idx)
+        if idval is not None:
+            return idval
+        idval = self._idval_from_hook(val, argname)
+        if idval is not None:
+            return idval
+        idval = self._idval_from_value(val)
+        if idval is not None:
+            return idval
+        return self._idval_from_argname(argname, idx)
+
+    def _idval_from_function(self, val: object, argname: str, idx: int) -> str | None:
+        """Try to make an ID for a parameter in a ParameterSet using the
+        user-provided id callable, if given."""
+        if self.idfn is None:
+            return None
+        try:
+            id = self.idfn(val)
+        except Exception as e:
+            prefix = f"{self.nodeid}: " if self.nodeid is not None else ""
+            msg = "error raised while trying to determine id of parameter '{}' at position {}"
+            msg = prefix + msg.format(argname, idx)
+            raise ValueError(msg) from e
+        if id is None:
+            return None
+        return self._idval_from_value(id)
+
+    def _idval_from_hook(self, val: object, argname: str) -> str | None:
+        """Try to make an ID for a parameter in a ParameterSet by calling the
+        :hook:`pytest_make_parametrize_id` hook."""
+        if self.config:
+            id: str | None = self.config.hook.pytest_make_parametrize_id(
+                config=self.config, val=val, argname=argname
+            )
+            return id
+        return None
+
+    def _idval_from_value(self, val: object) -> str | None:
+        """Try to make an ID for a parameter in a ParameterSet from its value,
+        if the value type is supported."""
+        if isinstance(val, (str, bytes)):
+            return _ascii_escaped_by_config(val, self.config)
+        elif val is None or isinstance(val, (float, int, bool, complex)):
+            return str(val)
+        elif isinstance(val, Pattern):
+            return ascii_escaped(val.pattern)
+        elif val is NOTSET:
+            # Fallback to default. Note that NOTSET is an enum.Enum.
+            pass
+        elif isinstance(val, enum.Enum):
+            return str(val)
+        elif isinstance(getattr(val, "__name__", None), str):
+            # Name of a class, function, module, etc.
+            name: str = getattr(val, "__name__")
+            return name
+        return None
+
+    def _idval_from_value_required(self, val: object, idx: int) -> str:
+        """Like _idval_from_value(), but fails if the type is not supported."""
+        id = self._idval_from_value(val)
+        if id is not None:
+            return id
+
+        # Fail.
+        if self.func_name is not None:
+            prefix = f"In {self.func_name}: "
+        elif self.nodeid is not None:
+            prefix = f"In {self.nodeid}: "
+        else:
+            prefix = ""
+        msg = (
+            f"{prefix}ids contains unsupported value {saferepr(val)} (type: {type(val)!r}) at index {idx}. "
+            "Supported types are: str, bytes, int, float, complex, bool, enum, regex or anything with a __name__."
+        )
+        fail(msg, pytrace=False)
+
+    @staticmethod
+    def _idval_from_argname(argname: str, idx: int) -> str:
+        """Make an ID for a parameter in a ParameterSet from the argument name
+        and the index of the ParameterSet."""
+        return str(argname) + str(idx)
+
+
+@final
+@dataclasses.dataclass(frozen=True)
+class CallSpec2:
+    """A planned parameterized invocation of a test function.
+
+    Calculated during collection for a given test function's Metafunc.
+    Once collection is over, each callspec is turned into a single Item
+    and stored in item.callspec.
+    """
+
+    # arg name -> arg value which will be passed to a fixture or pseudo-fixture
+    # of the same name. (indirect or direct parametrization respectively)
+    params: dict[str, object] = dataclasses.field(default_factory=dict)
+    # arg name -> arg index.
+    indices: dict[str, int] = dataclasses.field(default_factory=dict)
+    # Used for sorting parametrized resources.
+    _arg2scope: Mapping[str, Scope] = dataclasses.field(default_factory=dict)
+    # Parts which will be added to the item's name in `[..]` separated by "-".
+    _idlist: Sequence[str] = dataclasses.field(default_factory=tuple)
+    # Marks which will be applied to the item.
+    marks: list[Mark] = dataclasses.field(default_factory=list)
+
+    def setmulti(
+        self,
+        *,
+        argnames: Iterable[str],
+        valset: Iterable[object],
+        id: str,
+        marks: Iterable[Mark | MarkDecorator],
+        scope: Scope,
+        param_index: int,
+    ) -> CallSpec2:
+        params = self.params.copy()
+        indices = self.indices.copy()
+        arg2scope = dict(self._arg2scope)
+        for arg, val in zip(argnames, valset):
+            if arg in params:
+                raise ValueError(f"duplicate parametrization of {arg!r}")
+            params[arg] = val
+            indices[arg] = param_index
+            arg2scope[arg] = scope
+        return CallSpec2(
+            params=params,
+            indices=indices,
+            _arg2scope=arg2scope,
+            _idlist=[*self._idlist, id],
+            marks=[*self.marks, *normalize_mark_list(marks)],
+        )
+
+    def getparam(self, name: str) -> object:
+        try:
+            return self.params[name]
+        except KeyError as e:
+            raise ValueError(name) from e
+
+    @property
+    def id(self) -> str:
+        return "-".join(self._idlist)
+
+
+def get_direct_param_fixture_func(request: FixtureRequest) -> Any:
+    return request.param
+
+
+# Used for storing pseudo fixturedefs for direct parametrization.
+name2pseudofixturedef_key = StashKey[Dict[str, FixtureDef[Any]]]()
+
+
+@final
+class Metafunc:
+    """Objects passed to the :hook:`pytest_generate_tests` hook.
+
+    They help to inspect a test function and to generate tests according to
+    test configuration or values specified in the class or module where a
+    test function is defined.
+    """
+
+    def __init__(
+        self,
+        definition: FunctionDefinition,
+        fixtureinfo: fixtures.FuncFixtureInfo,
+        config: Config,
+        cls=None,
+        module=None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+
+        #: Access to the underlying :class:`_pytest.python.FunctionDefinition`.
+        self.definition = definition
+
+        #: Access to the :class:`pytest.Config` object for the test session.
+        self.config = config
+
+        #: The module object where the test function is defined in.
+        self.module = module
+
+        #: Underlying Python test function.
+        self.function = definition.obj
+
+        #: Set of fixture names required by the test function.
+        self.fixturenames = fixtureinfo.names_closure
+
+        #: Class object where the test function is defined in or ``None``.
+        self.cls = cls
+
+        self._arg2fixturedefs = fixtureinfo.name2fixturedefs
+
+        # Result of parametrize().
+        self._calls: list[CallSpec2] = []
+
+    def parametrize(
+        self,
+        argnames: str | Sequence[str],
+        argvalues: Iterable[ParameterSet | Sequence[object] | object],
+        indirect: bool | Sequence[str] = False,
+        ids: Iterable[object | None] | Callable[[Any], object | None] | None = None,
+        scope: _ScopeName | None = None,
+        *,
+        _param_mark: Mark | None = None,
+    ) -> None:
+        """Add new invocations to the underlying test function using the list
+        of argvalues for the given argnames. Parametrization is performed
+        during the collection phase. If you need to setup expensive resources
+        see about setting indirect to do it rather than at test setup time.
+
+        Can be called multiple times per test function (but only on different
+        argument names), in which case each call parametrizes all previous
+        parametrizations, e.g.
+
+        ::
+
+            unparametrized:         t
+            parametrize ["x", "y"]: t[x], t[y]
+            parametrize [1, 2]:     t[x-1], t[x-2], t[y-1], t[y-2]
+
+        :param argnames:
+            A comma-separated string denoting one or more argument names, or
+            a list/tuple of argument strings.
+
+        :param argvalues:
+            The list of argvalues determines how often a test is invoked with
+            different argument values.
+
+            If only one argname was specified argvalues is a list of values.
+            If N argnames were specified, argvalues must be a list of
+            N-tuples, where each tuple-element specifies a value for its
+            respective argname.
+        :type argvalues: Iterable[_pytest.mark.structures.ParameterSet | Sequence[object] | object]
+        :param indirect:
+            A list of arguments' names (subset of argnames) or a boolean.
+            If True the list contains all names from the argnames. Each
+            argvalue corresponding to an argname in this list will
+            be passed as request.param to its respective argname fixture
+            function so that it can perform more expensive setups during the
+            setup phase of a test rather than at collection time.
+
+        :param ids:
+            Sequence of (or generator for) ids for ``argvalues``,
+            or a callable to return part of the id for each argvalue.
+
+            With sequences (and generators like ``itertools.count()``) the
+            returned ids should be of type ``string``, ``int``, ``float``,
+            ``bool``, or ``None``.
+            They are mapped to the corresponding index in ``argvalues``.
+            ``None`` means to use the auto-generated id.
+
+            If it is a callable it will be called for each entry in
+            ``argvalues``, and the return value is used as part of the
+            auto-generated id for the whole set (where parts are joined with
+            dashes ("-")).
+            This is useful to provide more specific ids for certain items, e.g.
+            dates.  Returning ``None`` will use an auto-generated id.
+
+            If no ids are provided they will be generated automatically from
+            the argvalues.
+
+        :param scope:
+            If specified it denotes the scope of the parameters.
+            The scope is used for grouping tests by parameter instances.
+            It will also override any fixture-function defined scope, allowing
+            to set a dynamic scope using test context or configuration.
+        """
+        argnames, parametersets = ParameterSet._for_parametrize(
+            argnames,
+            argvalues,
+            self.function,
+            self.config,
+            nodeid=self.definition.nodeid,
+        )
+        del argvalues
+
+        if "request" in argnames:
+            fail(
+                "'request' is a reserved name and cannot be used in @pytest.mark.parametrize",
+                pytrace=False,
+            )
+
+        if scope is not None:
+            scope_ = Scope.from_user(
+                scope, descr=f"parametrize() call in {self.function.__name__}"
+            )
+        else:
+            scope_ = _find_parametrized_scope(argnames, self._arg2fixturedefs, indirect)
+
+        self._validate_if_using_arg_names(argnames, indirect)
+
+        # Use any already (possibly) generated ids with parametrize Marks.
+        if _param_mark and _param_mark._param_ids_from:
+            generated_ids = _param_mark._param_ids_from._param_ids_generated
+            if generated_ids is not None:
+                ids = generated_ids
+
+        ids = self._resolve_parameter_set_ids(
+            argnames, ids, parametersets, nodeid=self.definition.nodeid
+        )
+
+        # Store used (possibly generated) ids with parametrize Marks.
+        if _param_mark and _param_mark._param_ids_from and generated_ids is None:
+            object.__setattr__(_param_mark._param_ids_from, "_param_ids_generated", ids)
+
+        # Add funcargs as fixturedefs to fixtureinfo.arg2fixturedefs by registering
+        # artificial "pseudo" FixtureDef's so that later at test execution time we can
+        # rely on a proper FixtureDef to exist for fixture setup.
+        node = None
+        # If we have a scope that is higher than function, we need
+        # to make sure we only ever create an according fixturedef on
+        # a per-scope basis. We thus store and cache the fixturedef on the
+        # node related to the scope.
+        if scope_ is not Scope.Function:
+            collector = self.definition.parent
+            assert collector is not None
+            node = get_scope_node(collector, scope_)
+            if node is None:
+                # If used class scope and there is no class, use module-level
+                # collector (for now).
+                if scope_ is Scope.Class:
+                    assert isinstance(collector, Module)
+                    node = collector
+                # If used package scope and there is no package, use session
+                # (for now).
+                elif scope_ is Scope.Package:
+                    node = collector.session
+                else:
+                    assert False, f"Unhandled missing scope: {scope}"
+        if node is None:
+            name2pseudofixturedef = None
+        else:
+            default: dict[str, FixtureDef[Any]] = {}
+            name2pseudofixturedef = node.stash.setdefault(
+                name2pseudofixturedef_key, default
+            )
+        arg_directness = self._resolve_args_directness(argnames, indirect)
+        for argname in argnames:
+            if arg_directness[argname] == "indirect":
+                continue
+            if name2pseudofixturedef is not None and argname in name2pseudofixturedef:
+                fixturedef = name2pseudofixturedef[argname]
+            else:
+                fixturedef = FixtureDef(
+                    config=self.config,
+                    baseid="",
+                    argname=argname,
+                    func=get_direct_param_fixture_func,
+                    scope=scope_,
+                    params=None,
+                    ids=None,
+                    _ispytest=True,
+                )
+                if name2pseudofixturedef is not None:
+                    name2pseudofixturedef[argname] = fixturedef
+            self._arg2fixturedefs[argname] = [fixturedef]
+
+        # Create the new calls: if we are parametrize() multiple times (by applying the decorator
+        # more than once) then we accumulate those calls generating the cartesian product
+        # of all calls.
+        newcalls = []
+        for callspec in self._calls or [CallSpec2()]:
+            for param_index, (param_id, param_set) in enumerate(
+                zip(ids, parametersets)
+            ):
+                newcallspec = callspec.setmulti(
+                    argnames=argnames,
+                    valset=param_set.values,
+                    id=param_id,
+                    marks=param_set.marks,
+                    scope=scope_,
+                    param_index=param_index,
+                )
+                newcalls.append(newcallspec)
+        self._calls = newcalls
+
+    def _resolve_parameter_set_ids(
+        self,
+        argnames: Sequence[str],
+        ids: Iterable[object | None] | Callable[[Any], object | None] | None,
+        parametersets: Sequence[ParameterSet],
+        nodeid: str,
+    ) -> list[str]:
+        """Resolve the actual ids for the given parameter sets.
+
+        :param argnames:
+            Argument names passed to ``parametrize()``.
+        :param ids:
+            The `ids` parameter of the ``parametrize()`` call (see docs).
+        :param parametersets:
+            The parameter sets, each containing a set of values corresponding
+            to ``argnames``.
+        :param nodeid str:
+            The nodeid of the definition item that generated this
+            parametrization.
+        :returns:
+            List with ids for each parameter set given.
+        """
+        if ids is None:
+            idfn = None
+            ids_ = None
+        elif callable(ids):
+            idfn = ids
+            ids_ = None
+        else:
+            idfn = None
+            ids_ = self._validate_ids(ids, parametersets, self.function.__name__)
+        id_maker = IdMaker(
+            argnames,
+            parametersets,
+            idfn,
+            ids_,
+            self.config,
+            nodeid=nodeid,
+            func_name=self.function.__name__,
+        )
+        return id_maker.make_unique_parameterset_ids()
+
+    def _validate_ids(
+        self,
+        ids: Iterable[object | None],
+        parametersets: Sequence[ParameterSet],
+        func_name: str,
+    ) -> list[object | None]:
+        try:
+            num_ids = len(ids)  # type: ignore[arg-type]
+        except TypeError:
+            try:
+                iter(ids)
+            except TypeError as e:
+                raise TypeError("ids must be a callable or an iterable") from e
+            num_ids = len(parametersets)
+
+        # num_ids == 0 is a special case: https://github.com/pytest-dev/pytest/issues/1849
+        if num_ids != len(parametersets) and num_ids != 0:
+            msg = "In {}: {} parameter sets specified, with different number of ids: {}"
+            fail(msg.format(func_name, len(parametersets), num_ids), pytrace=False)
+
+        return list(itertools.islice(ids, num_ids))
+
+    def _resolve_args_directness(
+        self,
+        argnames: Sequence[str],
+        indirect: bool | Sequence[str],
+    ) -> dict[str, Literal["indirect", "direct"]]:
+        """Resolve if each parametrized argument must be considered an indirect
+        parameter to a fixture of the same name, or a direct parameter to the
+        parametrized function, based on the ``indirect`` parameter of the
+        parametrized() call.
+
+        :param argnames:
+            List of argument names passed to ``parametrize()``.
+        :param indirect:
+            Same as the ``indirect`` parameter of ``parametrize()``.
+        :returns
+            A dict mapping each arg name to either "indirect" or "direct".
+        """
+        arg_directness: dict[str, Literal["indirect", "direct"]]
+        if isinstance(indirect, bool):
+            arg_directness = dict.fromkeys(
+                argnames, "indirect" if indirect else "direct"
+            )
+        elif isinstance(indirect, Sequence):
+            arg_directness = dict.fromkeys(argnames, "direct")
+            for arg in indirect:
+                if arg not in argnames:
+                    fail(
+                        f"In {self.function.__name__}: indirect fixture '{arg}' doesn't exist",
+                        pytrace=False,
+                    )
+                arg_directness[arg] = "indirect"
+        else:
+            fail(
+                f"In {self.function.__name__}: expected Sequence or boolean"
+                f" for indirect, got {type(indirect).__name__}",
+                pytrace=False,
+            )
+        return arg_directness
+
+    def _validate_if_using_arg_names(
+        self,
+        argnames: Sequence[str],
+        indirect: bool | Sequence[str],
+    ) -> None:
+        """Check if all argnames are being used, by default values, or directly/indirectly.
+
+        :param List[str] argnames: List of argument names passed to ``parametrize()``.
+        :param indirect: Same as the ``indirect`` parameter of ``parametrize()``.
+        :raises ValueError: If validation fails.
+        """
+        default_arg_names = set(get_default_arg_names(self.function))
+        func_name = self.function.__name__
+        for arg in argnames:
+            if arg not in self.fixturenames:
+                if arg in default_arg_names:
+                    fail(
+                        f"In {func_name}: function already takes an argument '{arg}' with a default value",
+                        pytrace=False,
+                    )
+                else:
+                    if isinstance(indirect, Sequence):
+                        name = "fixture" if arg in indirect else "argument"
+                    else:
+                        name = "fixture" if indirect else "argument"
+                    fail(
+                        f"In {func_name}: function uses no {name} '{arg}'",
+                        pytrace=False,
+                    )
+
+
+def _find_parametrized_scope(
+    argnames: Sequence[str],
+    arg2fixturedefs: Mapping[str, Sequence[fixtures.FixtureDef[object]]],
+    indirect: bool | Sequence[str],
+) -> Scope:
+    """Find the most appropriate scope for a parametrized call based on its arguments.
+
+    When there's at least one direct argument, always use "function" scope.
+
+    When a test function is parametrized and all its arguments are indirect
+    (e.g. fixtures), return the most narrow scope based on the fixtures used.
+
+    Related to issue #1832, based on code posted by @Kingdread.
+    """
+    if isinstance(indirect, Sequence):
+        all_arguments_are_fixtures = len(indirect) == len(argnames)
+    else:
+        all_arguments_are_fixtures = bool(indirect)
+
+    if all_arguments_are_fixtures:
+        fixturedefs = arg2fixturedefs or {}
+        used_scopes = [
+            fixturedef[-1]._scope
+            for name, fixturedef in fixturedefs.items()
+            if name in argnames
+        ]
+        # Takes the most narrow scope from used fixtures.
+        return min(used_scopes, default=Scope.Function)
+
+    return Scope.Function
+
+
+def _ascii_escaped_by_config(val: str | bytes, config: Config | None) -> str:
+    if config is None:
+        escape_option = False
+    else:
+        escape_option = config.getini(
+            "disable_test_id_escaping_and_forfeit_all_rights_to_community_support"
+        )
+    # TODO: If escaping is turned off and the user passes bytes,
+    #       will return a bytes. For now we ignore this but the
+    #       code *probably* doesn't handle this case.
+    return val if escape_option else ascii_escaped(val)  # type: ignore
+
+
+class Function(PyobjMixin, nodes.Item):
+    """Item responsible for setting up and executing a Python test function.
+
+    :param name:
+        The full function name, including any decorations like those
+        added by parametrization (``my_func[my_param]``).
+    :param parent:
+        The parent Node.
+    :param config:
+        The pytest Config object.
+    :param callspec:
+        If given, this function has been parametrized and the callspec contains
+        meta information about the parametrization.
+    :param callobj:
+        If given, the object which will be called when the Function is invoked,
+        otherwise the callobj will be obtained from ``parent`` using ``originalname``.
+    :param keywords:
+        Keywords bound to the function object for "-k" matching.
+    :param session:
+        The pytest Session object.
+    :param fixtureinfo:
+        Fixture information already resolved at this fixture node..
+    :param originalname:
+        The attribute name to use for accessing the underlying function object.
+        Defaults to ``name``. Set this if name is different from the original name,
+        for example when it contains decorations like those added by parametrization
+        (``my_func[my_param]``).
+    """
+
+    # Disable since functions handle it themselves.
+    _ALLOW_MARKERS = False
+
+    def __init__(
+        self,
+        name: str,
+        parent,
+        config: Config | None = None,
+        callspec: CallSpec2 | None = None,
+        callobj=NOTSET,
+        keywords: Mapping[str, Any] | None = None,
+        session: Session | None = None,
+        fixtureinfo: FuncFixtureInfo | None = None,
+        originalname: str | None = None,
+    ) -> None:
+        super().__init__(name, parent, config=config, session=session)
+
+        if callobj is not NOTSET:
+            self._obj = callobj
+            self._instance = getattr(callobj, "__self__", None)
+
+        #: Original function name, without any decorations (for example
+        #: parametrization adds a ``"[...]"`` suffix to function names), used to access
+        #: the underlying function object from ``parent`` (in case ``callobj`` is not given
+        #: explicitly).
+        #:
+        #: .. versionadded:: 3.0
+        self.originalname = originalname or name
+
+        # Note: when FunctionDefinition is introduced, we should change ``originalname``
+        # to a readonly property that returns FunctionDefinition.name.
+
+        self.own_markers.extend(get_unpacked_marks(self.obj))
+        if callspec:
+            self.callspec = callspec
+            self.own_markers.extend(callspec.marks)
+
+        # todo: this is a hell of a hack
+        # https://github.com/pytest-dev/pytest/issues/4569
+        # Note: the order of the updates is important here; indicates what
+        # takes priority (ctor argument over function attributes over markers).
+        # Take own_markers only; NodeKeywords handles parent traversal on its own.
+        self.keywords.update((mark.name, mark) for mark in self.own_markers)
+        self.keywords.update(self.obj.__dict__)
+        if keywords:
+            self.keywords.update(keywords)
+
+        if fixtureinfo is None:
+            fm = self.session._fixturemanager
+            fixtureinfo = fm.getfixtureinfo(self, self.obj, self.cls)
+        self._fixtureinfo: FuncFixtureInfo = fixtureinfo
+        self.fixturenames = fixtureinfo.names_closure
+        self._initrequest()
+
+    # todo: determine sound type limitations
+    @classmethod
+    def from_parent(cls, parent, **kw) -> Self:
+        """The public constructor."""
+        return super().from_parent(parent=parent, **kw)
+
+    def _initrequest(self) -> None:
+        self.funcargs: dict[str, object] = {}
+        self._request = fixtures.TopRequest(self, _ispytest=True)
+
+    @property
+    def function(self):
+        """Underlying python 'function' object."""
+        return getimfunc(self.obj)
+
+    @property
+    def instance(self):
+        try:
+            return self._instance
+        except AttributeError:
+            if isinstance(self.parent, Class):
+                # Each Function gets a fresh class instance.
+                self._instance = self._getinstance()
+            else:
+                self._instance = None
+        return self._instance
+
+    def _getinstance(self):
+        if isinstance(self.parent, Class):
+            # Each Function gets a fresh class instance.
+            return self.parent.newinstance()
+        else:
+            return None
+
+    def _getobj(self):
+        instance = self.instance
+        if instance is not None:
+            parent_obj = instance
+        else:
+            assert self.parent is not None
+            parent_obj = self.parent.obj  # type: ignore[attr-defined]
+        return getattr(parent_obj, self.originalname)
+
+    @property
+    def _pyfuncitem(self):
+        """(compatonly) for code expecting pytest-2.2 style request objects."""
+        return self
+
+    def runtest(self) -> None:
+        """Execute the underlying test function."""
+        self.ihook.pytest_pyfunc_call(pyfuncitem=self)
+
+    def setup(self) -> None:
+        self._request._fillfixtures()
+
+    def _traceback_filter(self, excinfo: ExceptionInfo[BaseException]) -> Traceback:
+        if hasattr(self, "_obj") and not self.config.getoption("fulltrace", False):
+            code = _pytest._code.Code.from_function(get_real_func(self.obj))
+            path, firstlineno = code.path, code.firstlineno
+            traceback = excinfo.traceback
+            ntraceback = traceback.cut(path=path, firstlineno=firstlineno)
+            if ntraceback == traceback:
+                ntraceback = ntraceback.cut(path=path)
+                if ntraceback == traceback:
+                    ntraceback = ntraceback.filter(filter_traceback)
+                    if not ntraceback:
+                        ntraceback = traceback
+            ntraceback = ntraceback.filter(excinfo)
+
+            # issue364: mark all but first and last frames to
+            # only show a single-line message for each frame.
+            if self.config.getoption("tbstyle", "auto") == "auto":
+                if len(ntraceback) > 2:
+                    ntraceback = Traceback(
+                        (
+                            ntraceback[0],
+                            *(t.with_repr_style("short") for t in ntraceback[1:-1]),
+                            ntraceback[-1],
+                        )
+                    )
+
+            return ntraceback
+        return excinfo.traceback
+
+    # TODO: Type ignored -- breaks Liskov Substitution.
+    def repr_failure(  # type: ignore[override]
+        self,
+        excinfo: ExceptionInfo[BaseException],
+    ) -> str | TerminalRepr:
+        style = self.config.getoption("tbstyle", "auto")
+        if style == "auto":
+            style = "long"
+        return self._repr_failure_py(excinfo, style=style)
+
+
+class FunctionDefinition(Function):
+    """This class is a stop gap solution until we evolve to have actual function
+    definition nodes and manage to get rid of ``metafunc``."""
+
+    def runtest(self) -> None:
+        raise RuntimeError("function definitions are not supposed to be run as tests")
+
+    setup = runtest
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/python_api.py b/backend/venv/lib/python3.12/site-packages/_pytest/python_api.py
new file mode 100644
index 00000000..f0035f0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/python_api.py
@@ -0,0 +1,1028 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+from collections.abc import Collection
+from collections.abc import Sized
+from decimal import Decimal
+import math
+from numbers import Complex
+import pprint
+import re
+from types import TracebackType
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ContextManager
+from typing import final
+from typing import Mapping
+from typing import overload
+from typing import Pattern
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+import _pytest._code
+from _pytest.outcomes import fail
+
+
+if TYPE_CHECKING:
+    from numpy import ndarray
+
+
+def _compare_approx(
+    full_object: object,
+    message_data: Sequence[tuple[str, str, str]],
+    number_of_elements: int,
+    different_ids: Sequence[object],
+    max_abs_diff: float,
+    max_rel_diff: float,
+) -> list[str]:
+    message_list = list(message_data)
+    message_list.insert(0, ("Index", "Obtained", "Expected"))
+    max_sizes = [0, 0, 0]
+    for index, obtained, expected in message_list:
+        max_sizes[0] = max(max_sizes[0], len(index))
+        max_sizes[1] = max(max_sizes[1], len(obtained))
+        max_sizes[2] = max(max_sizes[2], len(expected))
+    explanation = [
+        f"comparison failed. Mismatched elements: {len(different_ids)} / {number_of_elements}:",
+        f"Max absolute difference: {max_abs_diff}",
+        f"Max relative difference: {max_rel_diff}",
+    ] + [
+        f"{indexes:<{max_sizes[0]}} | {obtained:<{max_sizes[1]}} | {expected:<{max_sizes[2]}}"
+        for indexes, obtained, expected in message_list
+    ]
+    return explanation
+
+
+# builtin pytest.approx helper
+
+
+class ApproxBase:
+    """Provide shared utilities for making approximate comparisons between
+    numbers or sequences of numbers."""
+
+    # Tell numpy to use our `__eq__` operator instead of its.
+    __array_ufunc__ = None
+    __array_priority__ = 100
+
+    def __init__(self, expected, rel=None, abs=None, nan_ok: bool = False) -> None:
+        __tracebackhide__ = True
+        self.expected = expected
+        self.abs = abs
+        self.rel = rel
+        self.nan_ok = nan_ok
+        self._check_type()
+
+    def __repr__(self) -> str:
+        raise NotImplementedError
+
+    def _repr_compare(self, other_side: Any) -> list[str]:
+        return [
+            "comparison failed",
+            f"Obtained: {other_side}",
+            f"Expected: {self}",
+        ]
+
+    def __eq__(self, actual) -> bool:
+        return all(
+            a == self._approx_scalar(x) for a, x in self._yield_comparisons(actual)
+        )
+
+    def __bool__(self):
+        __tracebackhide__ = True
+        raise AssertionError(
+            "approx() is not supported in a boolean context.\nDid you mean: `assert a == approx(b)`?"
+        )
+
+    # Ignore type because of https://github.com/python/mypy/issues/4266.
+    __hash__ = None  # type: ignore
+
+    def __ne__(self, actual) -> bool:
+        return not (actual == self)
+
+    def _approx_scalar(self, x) -> ApproxScalar:
+        if isinstance(x, Decimal):
+            return ApproxDecimal(x, rel=self.rel, abs=self.abs, nan_ok=self.nan_ok)
+        return ApproxScalar(x, rel=self.rel, abs=self.abs, nan_ok=self.nan_ok)
+
+    def _yield_comparisons(self, actual):
+        """Yield all the pairs of numbers to be compared.
+
+        This is used to implement the `__eq__` method.
+        """
+        raise NotImplementedError
+
+    def _check_type(self) -> None:
+        """Raise a TypeError if the expected value is not a valid type."""
+        # This is only a concern if the expected value is a sequence.  In every
+        # other case, the approx() function ensures that the expected value has
+        # a numeric type.  For this reason, the default is to do nothing.  The
+        # classes that deal with sequences should reimplement this method to
+        # raise if there are any non-numeric elements in the sequence.
+
+
+def _recursive_sequence_map(f, x):
+    """Recursively map a function over a sequence of arbitrary depth"""
+    if isinstance(x, (list, tuple)):
+        seq_type = type(x)
+        return seq_type(_recursive_sequence_map(f, xi) for xi in x)
+    elif _is_sequence_like(x):
+        return [_recursive_sequence_map(f, xi) for xi in x]
+    else:
+        return f(x)
+
+
+class ApproxNumpy(ApproxBase):
+    """Perform approximate comparisons where the expected value is numpy array."""
+
+    def __repr__(self) -> str:
+        list_scalars = _recursive_sequence_map(
+            self._approx_scalar, self.expected.tolist()
+        )
+        return f"approx({list_scalars!r})"
+
+    def _repr_compare(self, other_side: ndarray | list[Any]) -> list[str]:
+        import itertools
+        import math
+
+        def get_value_from_nested_list(
+            nested_list: list[Any], nd_index: tuple[Any, ...]
+        ) -> Any:
+            """
+            Helper function to get the value out of a nested list, given an n-dimensional index.
+            This mimics numpy's indexing, but for raw nested python lists.
+            """
+            value: Any = nested_list
+            for i in nd_index:
+                value = value[i]
+            return value
+
+        np_array_shape = self.expected.shape
+        approx_side_as_seq = _recursive_sequence_map(
+            self._approx_scalar, self.expected.tolist()
+        )
+
+        # convert other_side to numpy array to ensure shape attribute is available
+        other_side_as_array = _as_numpy_array(other_side)
+        assert other_side_as_array is not None
+
+        if np_array_shape != other_side_as_array.shape:
+            return [
+                "Impossible to compare arrays with different shapes.",
+                f"Shapes: {np_array_shape} and {other_side_as_array.shape}",
+            ]
+
+        number_of_elements = self.expected.size
+        max_abs_diff = -math.inf
+        max_rel_diff = -math.inf
+        different_ids = []
+        for index in itertools.product(*(range(i) for i in np_array_shape)):
+            approx_value = get_value_from_nested_list(approx_side_as_seq, index)
+            other_value = get_value_from_nested_list(other_side_as_array, index)
+            if approx_value != other_value:
+                abs_diff = abs(approx_value.expected - other_value)
+                max_abs_diff = max(max_abs_diff, abs_diff)
+                if other_value == 0.0:
+                    max_rel_diff = math.inf
+                else:
+                    max_rel_diff = max(max_rel_diff, abs_diff / abs(other_value))
+                different_ids.append(index)
+
+        message_data = [
+            (
+                str(index),
+                str(get_value_from_nested_list(other_side_as_array, index)),
+                str(get_value_from_nested_list(approx_side_as_seq, index)),
+            )
+            for index in different_ids
+        ]
+        return _compare_approx(
+            self.expected,
+            message_data,
+            number_of_elements,
+            different_ids,
+            max_abs_diff,
+            max_rel_diff,
+        )
+
+    def __eq__(self, actual) -> bool:
+        import numpy as np
+
+        # self.expected is supposed to always be an array here.
+
+        if not np.isscalar(actual):
+            try:
+                actual = np.asarray(actual)
+            except Exception as e:
+                raise TypeError(f"cannot compare '{actual}' to numpy.ndarray") from e
+
+        if not np.isscalar(actual) and actual.shape != self.expected.shape:
+            return False
+
+        return super().__eq__(actual)
+
+    def _yield_comparisons(self, actual):
+        import numpy as np
+
+        # `actual` can either be a numpy array or a scalar, it is treated in
+        # `__eq__` before being passed to `ApproxBase.__eq__`, which is the
+        # only method that calls this one.
+
+        if np.isscalar(actual):
+            for i in np.ndindex(self.expected.shape):
+                yield actual, self.expected[i].item()
+        else:
+            for i in np.ndindex(self.expected.shape):
+                yield actual[i].item(), self.expected[i].item()
+
+
+class ApproxMapping(ApproxBase):
+    """Perform approximate comparisons where the expected value is a mapping
+    with numeric values (the keys can be anything)."""
+
+    def __repr__(self) -> str:
+        return f"approx({({k: self._approx_scalar(v) for k, v in self.expected.items()})!r})"
+
+    def _repr_compare(self, other_side: Mapping[object, float]) -> list[str]:
+        import math
+
+        approx_side_as_map = {
+            k: self._approx_scalar(v) for k, v in self.expected.items()
+        }
+
+        number_of_elements = len(approx_side_as_map)
+        max_abs_diff = -math.inf
+        max_rel_diff = -math.inf
+        different_ids = []
+        for (approx_key, approx_value), other_value in zip(
+            approx_side_as_map.items(), other_side.values()
+        ):
+            if approx_value != other_value:
+                if approx_value.expected is not None and other_value is not None:
+                    try:
+                        max_abs_diff = max(
+                            max_abs_diff, abs(approx_value.expected - other_value)
+                        )
+                        if approx_value.expected == 0.0:
+                            max_rel_diff = math.inf
+                        else:
+                            max_rel_diff = max(
+                                max_rel_diff,
+                                abs(
+                                    (approx_value.expected - other_value)
+                                    / approx_value.expected
+                                ),
+                            )
+                    except ZeroDivisionError:
+                        pass
+                different_ids.append(approx_key)
+
+        message_data = [
+            (str(key), str(other_side[key]), str(approx_side_as_map[key]))
+            for key in different_ids
+        ]
+
+        return _compare_approx(
+            self.expected,
+            message_data,
+            number_of_elements,
+            different_ids,
+            max_abs_diff,
+            max_rel_diff,
+        )
+
+    def __eq__(self, actual) -> bool:
+        try:
+            if set(actual.keys()) != set(self.expected.keys()):
+                return False
+        except AttributeError:
+            return False
+
+        return super().__eq__(actual)
+
+    def _yield_comparisons(self, actual):
+        for k in self.expected.keys():
+            yield actual[k], self.expected[k]
+
+    def _check_type(self) -> None:
+        __tracebackhide__ = True
+        for key, value in self.expected.items():
+            if isinstance(value, type(self.expected)):
+                msg = "pytest.approx() does not support nested dictionaries: key={!r} value={!r}\n  full mapping={}"
+                raise TypeError(msg.format(key, value, pprint.pformat(self.expected)))
+
+
+class ApproxSequenceLike(ApproxBase):
+    """Perform approximate comparisons where the expected value is a sequence of numbers."""
+
+    def __repr__(self) -> str:
+        seq_type = type(self.expected)
+        if seq_type not in (tuple, list):
+            seq_type = list
+        return f"approx({seq_type(self._approx_scalar(x) for x in self.expected)!r})"
+
+    def _repr_compare(self, other_side: Sequence[float]) -> list[str]:
+        import math
+
+        if len(self.expected) != len(other_side):
+            return [
+                "Impossible to compare lists with different sizes.",
+                f"Lengths: {len(self.expected)} and {len(other_side)}",
+            ]
+
+        approx_side_as_map = _recursive_sequence_map(self._approx_scalar, self.expected)
+
+        number_of_elements = len(approx_side_as_map)
+        max_abs_diff = -math.inf
+        max_rel_diff = -math.inf
+        different_ids = []
+        for i, (approx_value, other_value) in enumerate(
+            zip(approx_side_as_map, other_side)
+        ):
+            if approx_value != other_value:
+                abs_diff = abs(approx_value.expected - other_value)
+                max_abs_diff = max(max_abs_diff, abs_diff)
+                if other_value == 0.0:
+                    max_rel_diff = math.inf
+                else:
+                    max_rel_diff = max(max_rel_diff, abs_diff / abs(other_value))
+                different_ids.append(i)
+
+        message_data = [
+            (str(i), str(other_side[i]), str(approx_side_as_map[i]))
+            for i in different_ids
+        ]
+
+        return _compare_approx(
+            self.expected,
+            message_data,
+            number_of_elements,
+            different_ids,
+            max_abs_diff,
+            max_rel_diff,
+        )
+
+    def __eq__(self, actual) -> bool:
+        try:
+            if len(actual) != len(self.expected):
+                return False
+        except TypeError:
+            return False
+        return super().__eq__(actual)
+
+    def _yield_comparisons(self, actual):
+        return zip(actual, self.expected)
+
+    def _check_type(self) -> None:
+        __tracebackhide__ = True
+        for index, x in enumerate(self.expected):
+            if isinstance(x, type(self.expected)):
+                msg = "pytest.approx() does not support nested data structures: {!r} at index {}\n  full sequence: {}"
+                raise TypeError(msg.format(x, index, pprint.pformat(self.expected)))
+
+
+class ApproxScalar(ApproxBase):
+    """Perform approximate comparisons where the expected value is a single number."""
+
+    # Using Real should be better than this Union, but not possible yet:
+    # https://github.com/python/typeshed/pull/3108
+    DEFAULT_ABSOLUTE_TOLERANCE: float | Decimal = 1e-12
+    DEFAULT_RELATIVE_TOLERANCE: float | Decimal = 1e-6
+
+    def __repr__(self) -> str:
+        """Return a string communicating both the expected value and the
+        tolerance for the comparison being made.
+
+        For example, ``1.0 ± 1e-6``, ``(3+4j) ± 5e-6 ∠ ±180°``.
+        """
+        # Don't show a tolerance for values that aren't compared using
+        # tolerances, i.e. non-numerics and infinities. Need to call abs to
+        # handle complex numbers, e.g. (inf + 1j).
+        if (
+            isinstance(self.expected, bool)
+            or (not isinstance(self.expected, (Complex, Decimal)))
+            or math.isinf(abs(self.expected) or isinstance(self.expected, bool))
+        ):
+            return str(self.expected)
+
+        # If a sensible tolerance can't be calculated, self.tolerance will
+        # raise a ValueError.  In this case, display '???'.
+        try:
+            vetted_tolerance = f"{self.tolerance:.1e}"
+            if (
+                isinstance(self.expected, Complex)
+                and self.expected.imag
+                and not math.isinf(self.tolerance)
+            ):
+                vetted_tolerance += " ∠ ±180°"
+        except ValueError:
+            vetted_tolerance = "???"
+
+        return f"{self.expected} ± {vetted_tolerance}"
+
+    def __eq__(self, actual) -> bool:
+        """Return whether the given value is equal to the expected value
+        within the pre-specified tolerance."""
+        asarray = _as_numpy_array(actual)
+        if asarray is not None:
+            # Call ``__eq__()`` manually to prevent infinite-recursion with
+            # numpy<1.13.  See #3748.
+            return all(self.__eq__(a) for a in asarray.flat)
+
+        # Short-circuit exact equality, except for bool
+        if isinstance(self.expected, bool) and not isinstance(actual, bool):
+            return False
+        elif actual == self.expected:
+            return True
+
+        # If either type is non-numeric, fall back to strict equality.
+        # NB: we need Complex, rather than just Number, to ensure that __abs__,
+        # __sub__, and __float__ are defined. Also, consider bool to be
+        # nonnumeric, even though it has the required arithmetic.
+        if isinstance(self.expected, bool) or not (
+            isinstance(self.expected, (Complex, Decimal))
+            and isinstance(actual, (Complex, Decimal))
+        ):
+            return False
+
+        # Allow the user to control whether NaNs are considered equal to each
+        # other or not.  The abs() calls are for compatibility with complex
+        # numbers.
+        if math.isnan(abs(self.expected)):
+            return self.nan_ok and math.isnan(abs(actual))
+
+        # Infinity shouldn't be approximately equal to anything but itself, but
+        # if there's a relative tolerance, it will be infinite and infinity
+        # will seem approximately equal to everything.  The equal-to-itself
+        # case would have been short circuited above, so here we can just
+        # return false if the expected value is infinite.  The abs() call is
+        # for compatibility with complex numbers.
+        if math.isinf(abs(self.expected)):
+            return False
+
+        # Return true if the two numbers are within the tolerance.
+        result: bool = abs(self.expected - actual) <= self.tolerance
+        return result
+
+    # Ignore type because of https://github.com/python/mypy/issues/4266.
+    __hash__ = None  # type: ignore
+
+    @property
+    def tolerance(self):
+        """Return the tolerance for the comparison.
+
+        This could be either an absolute tolerance or a relative tolerance,
+        depending on what the user specified or which would be larger.
+        """
+
+        def set_default(x, default):
+            return x if x is not None else default
+
+        # Figure out what the absolute tolerance should be.  ``self.abs`` is
+        # either None or a value specified by the user.
+        absolute_tolerance = set_default(self.abs, self.DEFAULT_ABSOLUTE_TOLERANCE)
+
+        if absolute_tolerance < 0:
+            raise ValueError(
+                f"absolute tolerance can't be negative: {absolute_tolerance}"
+            )
+        if math.isnan(absolute_tolerance):
+            raise ValueError("absolute tolerance can't be NaN.")
+
+        # If the user specified an absolute tolerance but not a relative one,
+        # just return the absolute tolerance.
+        if self.rel is None:
+            if self.abs is not None:
+                return absolute_tolerance
+
+        # Figure out what the relative tolerance should be.  ``self.rel`` is
+        # either None or a value specified by the user.  This is done after
+        # we've made sure the user didn't ask for an absolute tolerance only,
+        # because we don't want to raise errors about the relative tolerance if
+        # we aren't even going to use it.
+        relative_tolerance = set_default(
+            self.rel, self.DEFAULT_RELATIVE_TOLERANCE
+        ) * abs(self.expected)
+
+        if relative_tolerance < 0:
+            raise ValueError(
+                f"relative tolerance can't be negative: {relative_tolerance}"
+            )
+        if math.isnan(relative_tolerance):
+            raise ValueError("relative tolerance can't be NaN.")
+
+        # Return the larger of the relative and absolute tolerances.
+        return max(relative_tolerance, absolute_tolerance)
+
+
+class ApproxDecimal(ApproxScalar):
+    """Perform approximate comparisons where the expected value is a Decimal."""
+
+    DEFAULT_ABSOLUTE_TOLERANCE = Decimal("1e-12")
+    DEFAULT_RELATIVE_TOLERANCE = Decimal("1e-6")
+
+
+def approx(expected, rel=None, abs=None, nan_ok: bool = False) -> ApproxBase:
+    """Assert that two numbers (or two ordered sequences of numbers) are equal to each other
+    within some tolerance.
+
+    Due to the :doc:`python:tutorial/floatingpoint`, numbers that we
+    would intuitively expect to be equal are not always so::
+
+        >>> 0.1 + 0.2 == 0.3
+        False
+
+    This problem is commonly encountered when writing tests, e.g. when making
+    sure that floating-point values are what you expect them to be.  One way to
+    deal with this problem is to assert that two floating-point numbers are
+    equal to within some appropriate tolerance::
+
+        >>> abs((0.1 + 0.2) - 0.3) < 1e-6
+        True
+
+    However, comparisons like this are tedious to write and difficult to
+    understand.  Furthermore, absolute comparisons like the one above are
+    usually discouraged because there's no tolerance that works well for all
+    situations.  ``1e-6`` is good for numbers around ``1``, but too small for
+    very big numbers and too big for very small ones.  It's better to express
+    the tolerance as a fraction of the expected value, but relative comparisons
+    like that are even more difficult to write correctly and concisely.
+
+    The ``approx`` class performs floating-point comparisons using a syntax
+    that's as intuitive as possible::
+
+        >>> from pytest import approx
+        >>> 0.1 + 0.2 == approx(0.3)
+        True
+
+    The same syntax also works for ordered sequences of numbers::
+
+        >>> (0.1 + 0.2, 0.2 + 0.4) == approx((0.3, 0.6))
+        True
+
+    ``numpy`` arrays::
+
+        >>> import numpy as np                                                          # doctest: +SKIP
+        >>> np.array([0.1, 0.2]) + np.array([0.2, 0.4]) == approx(np.array([0.3, 0.6])) # doctest: +SKIP
+        True
+
+    And for a ``numpy`` array against a scalar::
+
+        >>> import numpy as np                                         # doctest: +SKIP
+        >>> np.array([0.1, 0.2]) + np.array([0.2, 0.1]) == approx(0.3) # doctest: +SKIP
+        True
+
+    Only ordered sequences are supported, because ``approx`` needs
+    to infer the relative position of the sequences without ambiguity. This means
+    ``sets`` and other unordered sequences are not supported.
+
+    Finally, dictionary *values* can also be compared::
+
+        >>> {'a': 0.1 + 0.2, 'b': 0.2 + 0.4} == approx({'a': 0.3, 'b': 0.6})
+        True
+
+    The comparison will be true if both mappings have the same keys and their
+    respective values match the expected tolerances.
+
+    **Tolerances**
+
+    By default, ``approx`` considers numbers within a relative tolerance of
+    ``1e-6`` (i.e. one part in a million) of its expected value to be equal.
+    This treatment would lead to surprising results if the expected value was
+    ``0.0``, because nothing but ``0.0`` itself is relatively close to ``0.0``.
+    To handle this case less surprisingly, ``approx`` also considers numbers
+    within an absolute tolerance of ``1e-12`` of its expected value to be
+    equal.  Infinity and NaN are special cases.  Infinity is only considered
+    equal to itself, regardless of the relative tolerance.  NaN is not
+    considered equal to anything by default, but you can make it be equal to
+    itself by setting the ``nan_ok`` argument to True.  (This is meant to
+    facilitate comparing arrays that use NaN to mean "no data".)
+
+    Both the relative and absolute tolerances can be changed by passing
+    arguments to the ``approx`` constructor::
+
+        >>> 1.0001 == approx(1)
+        False
+        >>> 1.0001 == approx(1, rel=1e-3)
+        True
+        >>> 1.0001 == approx(1, abs=1e-3)
+        True
+
+    If you specify ``abs`` but not ``rel``, the comparison will not consider
+    the relative tolerance at all.  In other words, two numbers that are within
+    the default relative tolerance of ``1e-6`` will still be considered unequal
+    if they exceed the specified absolute tolerance.  If you specify both
+    ``abs`` and ``rel``, the numbers will be considered equal if either
+    tolerance is met::
+
+        >>> 1 + 1e-8 == approx(1)
+        True
+        >>> 1 + 1e-8 == approx(1, abs=1e-12)
+        False
+        >>> 1 + 1e-8 == approx(1, rel=1e-6, abs=1e-12)
+        True
+
+    You can also use ``approx`` to compare nonnumeric types, or dicts and
+    sequences containing nonnumeric types, in which case it falls back to
+    strict equality. This can be useful for comparing dicts and sequences that
+    can contain optional values::
+
+        >>> {"required": 1.0000005, "optional": None} == approx({"required": 1, "optional": None})
+        True
+        >>> [None, 1.0000005] == approx([None,1])
+        True
+        >>> ["foo", 1.0000005] == approx([None,1])
+        False
+
+    If you're thinking about using ``approx``, then you might want to know how
+    it compares to other good ways of comparing floating-point numbers.  All of
+    these algorithms are based on relative and absolute tolerances and should
+    agree for the most part, but they do have meaningful differences:
+
+    - ``math.isclose(a, b, rel_tol=1e-9, abs_tol=0.0)``:  True if the relative
+      tolerance is met w.r.t. either ``a`` or ``b`` or if the absolute
+      tolerance is met.  Because the relative tolerance is calculated w.r.t.
+      both ``a`` and ``b``, this test is symmetric (i.e.  neither ``a`` nor
+      ``b`` is a "reference value").  You have to specify an absolute tolerance
+      if you want to compare to ``0.0`` because there is no tolerance by
+      default.  More information: :py:func:`math.isclose`.
+
+    - ``numpy.isclose(a, b, rtol=1e-5, atol=1e-8)``: True if the difference
+      between ``a`` and ``b`` is less that the sum of the relative tolerance
+      w.r.t. ``b`` and the absolute tolerance.  Because the relative tolerance
+      is only calculated w.r.t. ``b``, this test is asymmetric and you can
+      think of ``b`` as the reference value.  Support for comparing sequences
+      is provided by :py:func:`numpy.allclose`.  More information:
+      :std:doc:`numpy:reference/generated/numpy.isclose`.
+
+    - ``unittest.TestCase.assertAlmostEqual(a, b)``: True if ``a`` and ``b``
+      are within an absolute tolerance of ``1e-7``.  No relative tolerance is
+      considered , so this function is not appropriate for very large or very
+      small numbers.  Also, it's only available in subclasses of ``unittest.TestCase``
+      and it's ugly because it doesn't follow PEP8.  More information:
+      :py:meth:`unittest.TestCase.assertAlmostEqual`.
+
+    - ``a == pytest.approx(b, rel=1e-6, abs=1e-12)``: True if the relative
+      tolerance is met w.r.t. ``b`` or if the absolute tolerance is met.
+      Because the relative tolerance is only calculated w.r.t. ``b``, this test
+      is asymmetric and you can think of ``b`` as the reference value.  In the
+      special case that you explicitly specify an absolute tolerance but not a
+      relative tolerance, only the absolute tolerance is considered.
+
+    .. note::
+
+        ``approx`` can handle numpy arrays, but we recommend the
+        specialised test helpers in :std:doc:`numpy:reference/routines.testing`
+        if you need support for comparisons, NaNs, or ULP-based tolerances.
+
+        To match strings using regex, you can use
+        `Matches <https://github.com/asottile/re-assert#re_assertmatchespattern-str-args-kwargs>`_
+        from the
+        `re_assert package <https://github.com/asottile/re-assert>`_.
+
+    .. warning::
+
+       .. versionchanged:: 3.2
+
+       In order to avoid inconsistent behavior, :py:exc:`TypeError` is
+       raised for ``>``, ``>=``, ``<`` and ``<=`` comparisons.
+       The example below illustrates the problem::
+
+           assert approx(0.1) > 0.1 + 1e-10  # calls approx(0.1).__gt__(0.1 + 1e-10)
+           assert 0.1 + 1e-10 > approx(0.1)  # calls approx(0.1).__lt__(0.1 + 1e-10)
+
+       In the second example one expects ``approx(0.1).__le__(0.1 + 1e-10)``
+       to be called. But instead, ``approx(0.1).__lt__(0.1 + 1e-10)`` is used to
+       comparison. This is because the call hierarchy of rich comparisons
+       follows a fixed behavior. More information: :py:meth:`object.__ge__`
+
+    .. versionchanged:: 3.7.1
+       ``approx`` raises ``TypeError`` when it encounters a dict value or
+       sequence element of nonnumeric type.
+
+    .. versionchanged:: 6.1.0
+       ``approx`` falls back to strict equality for nonnumeric types instead
+       of raising ``TypeError``.
+    """
+    # Delegate the comparison to a class that knows how to deal with the type
+    # of the expected value (e.g. int, float, list, dict, numpy.array, etc).
+    #
+    # The primary responsibility of these classes is to implement ``__eq__()``
+    # and ``__repr__()``.  The former is used to actually check if some
+    # "actual" value is equivalent to the given expected value within the
+    # allowed tolerance.  The latter is used to show the user the expected
+    # value and tolerance, in the case that a test failed.
+    #
+    # The actual logic for making approximate comparisons can be found in
+    # ApproxScalar, which is used to compare individual numbers.  All of the
+    # other Approx classes eventually delegate to this class.  The ApproxBase
+    # class provides some convenient methods and overloads, but isn't really
+    # essential.
+
+    __tracebackhide__ = True
+
+    if isinstance(expected, Decimal):
+        cls: type[ApproxBase] = ApproxDecimal
+    elif isinstance(expected, Mapping):
+        cls = ApproxMapping
+    elif _is_numpy_array(expected):
+        expected = _as_numpy_array(expected)
+        cls = ApproxNumpy
+    elif _is_sequence_like(expected):
+        cls = ApproxSequenceLike
+    elif isinstance(expected, Collection) and not isinstance(expected, (str, bytes)):
+        msg = f"pytest.approx() only supports ordered sequences, but got: {expected!r}"
+        raise TypeError(msg)
+    else:
+        cls = ApproxScalar
+
+    return cls(expected, rel, abs, nan_ok)
+
+
+def _is_sequence_like(expected: object) -> bool:
+    return (
+        hasattr(expected, "__getitem__")
+        and isinstance(expected, Sized)
+        and not isinstance(expected, (str, bytes))
+    )
+
+
+def _is_numpy_array(obj: object) -> bool:
+    """
+    Return true if the given object is implicitly convertible to ndarray,
+    and numpy is already imported.
+    """
+    return _as_numpy_array(obj) is not None
+
+
+def _as_numpy_array(obj: object) -> ndarray | None:
+    """
+    Return an ndarray if the given object is implicitly convertible to ndarray,
+    and numpy is already imported, otherwise None.
+    """
+    import sys
+
+    np: Any = sys.modules.get("numpy")
+    if np is not None:
+        # avoid infinite recursion on numpy scalars, which have __array__
+        if np.isscalar(obj):
+            return None
+        elif isinstance(obj, np.ndarray):
+            return obj
+        elif hasattr(obj, "__array__") or hasattr("obj", "__array_interface__"):
+            return np.asarray(obj)
+    return None
+
+
+# builtin pytest.raises helper
+
+E = TypeVar("E", bound=BaseException)
+
+
+@overload
+def raises(
+    expected_exception: type[E] | tuple[type[E], ...],
+    *,
+    match: str | Pattern[str] | None = ...,
+) -> RaisesContext[E]: ...
+
+
+@overload
+def raises(
+    expected_exception: type[E] | tuple[type[E], ...],
+    func: Callable[..., Any],
+    *args: Any,
+    **kwargs: Any,
+) -> _pytest._code.ExceptionInfo[E]: ...
+
+
+def raises(
+    expected_exception: type[E] | tuple[type[E], ...], *args: Any, **kwargs: Any
+) -> RaisesContext[E] | _pytest._code.ExceptionInfo[E]:
+    r"""Assert that a code block/function call raises an exception type, or one of its subclasses.
+
+    :param expected_exception:
+        The expected exception type, or a tuple if one of multiple possible
+        exception types are expected. Note that subclasses of the passed exceptions
+        will also match.
+
+    :kwparam str | re.Pattern[str] | None match:
+        If specified, a string containing a regular expression,
+        or a regular expression object, that is tested against the string
+        representation of the exception and its :pep:`678` `__notes__`
+        using :func:`re.search`.
+
+        To match a literal string that may contain :ref:`special characters
+        <re-syntax>`, the pattern can first be escaped with :func:`re.escape`.
+
+        (This is only used when ``pytest.raises`` is used as a context manager,
+        and passed through to the function otherwise.
+        When using ``pytest.raises`` as a function, you can use:
+        ``pytest.raises(Exc, func, match="passed on").match("my pattern")``.)
+
+    Use ``pytest.raises`` as a context manager, which will capture the exception of the given
+    type, or any of its subclasses::
+
+        >>> import pytest
+        >>> with pytest.raises(ZeroDivisionError):
+        ...    1/0
+
+    If the code block does not raise the expected exception (:class:`ZeroDivisionError` in the example
+    above), or no exception at all, the check will fail instead.
+
+    You can also use the keyword argument ``match`` to assert that the
+    exception matches a text or regex::
+
+        >>> with pytest.raises(ValueError, match='must be 0 or None'):
+        ...     raise ValueError("value must be 0 or None")
+
+        >>> with pytest.raises(ValueError, match=r'must be \d+$'):
+        ...     raise ValueError("value must be 42")
+
+    The ``match`` argument searches the formatted exception string, which includes any
+    `PEP-678 <https://peps.python.org/pep-0678/>`__ ``__notes__``:
+
+        >>> with pytest.raises(ValueError, match=r"had a note added"):  # doctest: +SKIP
+        ...     e = ValueError("value must be 42")
+        ...     e.add_note("had a note added")
+        ...     raise e
+
+    The context manager produces an :class:`ExceptionInfo` object which can be used to inspect the
+    details of the captured exception::
+
+        >>> with pytest.raises(ValueError) as exc_info:
+        ...     raise ValueError("value must be 42")
+        >>> assert exc_info.type is ValueError
+        >>> assert exc_info.value.args[0] == "value must be 42"
+
+    .. warning::
+
+       Given that ``pytest.raises`` matches subclasses, be wary of using it to match :class:`Exception` like this::
+
+           with pytest.raises(Exception):  # Careful, this will catch ANY exception raised.
+               some_function()
+
+       Because :class:`Exception` is the base class of almost all exceptions, it is easy for this to hide
+       real bugs, where the user wrote this expecting a specific exception, but some other exception is being
+       raised due to a bug introduced during a refactoring.
+
+       Avoid using ``pytest.raises`` to catch :class:`Exception` unless certain that you really want to catch
+       **any** exception raised.
+
+    .. note::
+
+       When using ``pytest.raises`` as a context manager, it's worthwhile to
+       note that normal context manager rules apply and that the exception
+       raised *must* be the final line in the scope of the context manager.
+       Lines of code after that, within the scope of the context manager will
+       not be executed. For example::
+
+           >>> value = 15
+           >>> with pytest.raises(ValueError) as exc_info:
+           ...     if value > 10:
+           ...         raise ValueError("value must be <= 10")
+           ...     assert exc_info.type is ValueError  # This will not execute.
+
+       Instead, the following approach must be taken (note the difference in
+       scope)::
+
+           >>> with pytest.raises(ValueError) as exc_info:
+           ...     if value > 10:
+           ...         raise ValueError("value must be <= 10")
+           ...
+           >>> assert exc_info.type is ValueError
+
+    **Using with** ``pytest.mark.parametrize``
+
+    When using :ref:`pytest.mark.parametrize ref`
+    it is possible to parametrize tests such that
+    some runs raise an exception and others do not.
+
+    See :ref:`parametrizing_conditional_raising` for an example.
+
+    .. seealso::
+
+        :ref:`assertraises` for more examples and detailed discussion.
+
+    **Legacy form**
+
+    It is possible to specify a callable by passing a to-be-called lambda::
+
+        >>> raises(ZeroDivisionError, lambda: 1/0)
+        <ExceptionInfo ...>
+
+    or you can specify an arbitrary callable with arguments::
+
+        >>> def f(x): return 1/x
+        ...
+        >>> raises(ZeroDivisionError, f, 0)
+        <ExceptionInfo ...>
+        >>> raises(ZeroDivisionError, f, x=0)
+        <ExceptionInfo ...>
+
+    The form above is fully supported but discouraged for new code because the
+    context manager form is regarded as more readable and less error-prone.
+
+    .. note::
+        Similar to caught exception objects in Python, explicitly clearing
+        local references to returned ``ExceptionInfo`` objects can
+        help the Python interpreter speed up its garbage collection.
+
+        Clearing those references breaks a reference cycle
+        (``ExceptionInfo`` --> caught exception --> frame stack raising
+        the exception --> current frame stack --> local variables -->
+        ``ExceptionInfo``) which makes Python keep all objects referenced
+        from that cycle (including all local variables in the current
+        frame) alive until the next cyclic garbage collection run.
+        More detailed information can be found in the official Python
+        documentation for :ref:`the try statement <python:try>`.
+    """
+    __tracebackhide__ = True
+
+    if not expected_exception:
+        raise ValueError(
+            f"Expected an exception type or a tuple of exception types, but got `{expected_exception!r}`. "
+            f"Raising exceptions is already understood as failing the test, so you don't need "
+            f"any special code to say 'this should never raise an exception'."
+        )
+    if isinstance(expected_exception, type):
+        expected_exceptions: tuple[type[E], ...] = (expected_exception,)
+    else:
+        expected_exceptions = expected_exception
+    for exc in expected_exceptions:
+        if not isinstance(exc, type) or not issubclass(exc, BaseException):
+            msg = "expected exception must be a BaseException type, not {}"  # type: ignore[unreachable]
+            not_a = exc.__name__ if isinstance(exc, type) else type(exc).__name__
+            raise TypeError(msg.format(not_a))
+
+    message = f"DID NOT RAISE {expected_exception}"
+
+    if not args:
+        match: str | Pattern[str] | None = kwargs.pop("match", None)
+        if kwargs:
+            msg = "Unexpected keyword arguments passed to pytest.raises: "
+            msg += ", ".join(sorted(kwargs))
+            msg += "\nUse context-manager form instead?"
+            raise TypeError(msg)
+        return RaisesContext(expected_exception, message, match)
+    else:
+        func = args[0]
+        if not callable(func):
+            raise TypeError(f"{func!r} object (type: {type(func)}) must be callable")
+        try:
+            func(*args[1:], **kwargs)
+        except expected_exception as e:
+            return _pytest._code.ExceptionInfo.from_exception(e)
+    fail(message)
+
+
+# This doesn't work with mypy for now. Use fail.Exception instead.
+raises.Exception = fail.Exception  # type: ignore
+
+
+@final
+class RaisesContext(ContextManager[_pytest._code.ExceptionInfo[E]]):
+    def __init__(
+        self,
+        expected_exception: type[E] | tuple[type[E], ...],
+        message: str,
+        match_expr: str | Pattern[str] | None = None,
+    ) -> None:
+        self.expected_exception = expected_exception
+        self.message = message
+        self.match_expr = match_expr
+        self.excinfo: _pytest._code.ExceptionInfo[E] | None = None
+        if self.match_expr is not None:
+            re_error = None
+            try:
+                re.compile(self.match_expr)
+            except re.error as e:
+                re_error = e
+            if re_error is not None:
+                fail(f"Invalid regex pattern provided to 'match': {re_error}")
+
+    def __enter__(self) -> _pytest._code.ExceptionInfo[E]:
+        self.excinfo = _pytest._code.ExceptionInfo.for_later()
+        return self.excinfo
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> bool:
+        __tracebackhide__ = True
+        if exc_type is None:
+            fail(self.message)
+        assert self.excinfo is not None
+        if not issubclass(exc_type, self.expected_exception):
+            return False
+        # Cast to narrow the exception type now that it's verified.
+        exc_info = cast(Tuple[Type[E], E, TracebackType], (exc_type, exc_val, exc_tb))
+        self.excinfo.fill_unfilled(exc_info)
+        if self.match_expr is not None:
+            self.excinfo.match(self.match_expr)
+        return True
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/python_path.py b/backend/venv/lib/python3.12/site-packages/_pytest/python_path.py
new file mode 100644
index 00000000..6e33c8a3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/python_path.py
@@ -0,0 +1,26 @@
+from __future__ import annotations
+
+import sys
+
+import pytest
+from pytest import Config
+from pytest import Parser
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini("pythonpath", type="paths", help="Add paths to sys.path", default=[])
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_load_initial_conftests(early_config: Config) -> None:
+    # `pythonpath = a b` will set `sys.path` to `[a, b, x, y, z, ...]`
+    for path in reversed(early_config.getini("pythonpath")):
+        sys.path.insert(0, str(path))
+
+
+@pytest.hookimpl(trylast=True)
+def pytest_unconfigure(config: Config) -> None:
+    for path in config.getini("pythonpath"):
+        path_str = str(path)
+        if path_str in sys.path:
+            sys.path.remove(path_str)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/recwarn.py b/backend/venv/lib/python3.12/site-packages/_pytest/recwarn.py
new file mode 100644
index 00000000..0dc002ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/recwarn.py
@@ -0,0 +1,364 @@
+# mypy: allow-untyped-defs
+"""Record warnings during test function execution."""
+
+from __future__ import annotations
+
+from pprint import pformat
+import re
+from types import TracebackType
+from typing import Any
+from typing import Callable
+from typing import final
+from typing import Generator
+from typing import Iterator
+from typing import overload
+from typing import Pattern
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+import warnings
+
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.outcomes import Exit
+from _pytest.outcomes import fail
+
+
+T = TypeVar("T")
+
+
+@fixture
+def recwarn() -> Generator[WarningsRecorder]:
+    """Return a :class:`WarningsRecorder` instance that records all warnings emitted by test functions.
+
+    See :ref:`warnings` for information on warning categories.
+    """
+    wrec = WarningsRecorder(_ispytest=True)
+    with wrec:
+        warnings.simplefilter("default")
+        yield wrec
+
+
+@overload
+def deprecated_call(*, match: str | Pattern[str] | None = ...) -> WarningsRecorder: ...
+
+
+@overload
+def deprecated_call(func: Callable[..., T], *args: Any, **kwargs: Any) -> T: ...
+
+
+def deprecated_call(
+    func: Callable[..., Any] | None = None, *args: Any, **kwargs: Any
+) -> WarningsRecorder | Any:
+    """Assert that code produces a ``DeprecationWarning`` or ``PendingDeprecationWarning`` or ``FutureWarning``.
+
+    This function can be used as a context manager::
+
+        >>> import warnings
+        >>> def api_call_v2():
+        ...     warnings.warn('use v3 of this api', DeprecationWarning)
+        ...     return 200
+
+        >>> import pytest
+        >>> with pytest.deprecated_call():
+        ...    assert api_call_v2() == 200
+
+    It can also be used by passing a function and ``*args`` and ``**kwargs``,
+    in which case it will ensure calling ``func(*args, **kwargs)`` produces one of
+    the warnings types above. The return value is the return value of the function.
+
+    In the context manager form you may use the keyword argument ``match`` to assert
+    that the warning matches a text or regex.
+
+    The context manager produces a list of :class:`warnings.WarningMessage` objects,
+    one for each warning raised.
+    """
+    __tracebackhide__ = True
+    if func is not None:
+        args = (func, *args)
+    return warns(
+        (DeprecationWarning, PendingDeprecationWarning, FutureWarning), *args, **kwargs
+    )
+
+
+@overload
+def warns(
+    expected_warning: type[Warning] | tuple[type[Warning], ...] = ...,
+    *,
+    match: str | Pattern[str] | None = ...,
+) -> WarningsChecker: ...
+
+
+@overload
+def warns(
+    expected_warning: type[Warning] | tuple[type[Warning], ...],
+    func: Callable[..., T],
+    *args: Any,
+    **kwargs: Any,
+) -> T: ...
+
+
+def warns(
+    expected_warning: type[Warning] | tuple[type[Warning], ...] = Warning,
+    *args: Any,
+    match: str | Pattern[str] | None = None,
+    **kwargs: Any,
+) -> WarningsChecker | Any:
+    r"""Assert that code raises a particular class of warning.
+
+    Specifically, the parameter ``expected_warning`` can be a warning class or tuple
+    of warning classes, and the code inside the ``with`` block must issue at least one
+    warning of that class or classes.
+
+    This helper produces a list of :class:`warnings.WarningMessage` objects, one for
+    each warning emitted (regardless of whether it is an ``expected_warning`` or not).
+    Since pytest 8.0, unmatched warnings are also re-emitted when the context closes.
+
+    This function can be used as a context manager::
+
+        >>> import pytest
+        >>> with pytest.warns(RuntimeWarning):
+        ...    warnings.warn("my warning", RuntimeWarning)
+
+    In the context manager form you may use the keyword argument ``match`` to assert
+    that the warning matches a text or regex::
+
+        >>> with pytest.warns(UserWarning, match='must be 0 or None'):
+        ...     warnings.warn("value must be 0 or None", UserWarning)
+
+        >>> with pytest.warns(UserWarning, match=r'must be \d+$'):
+        ...     warnings.warn("value must be 42", UserWarning)
+
+        >>> with pytest.warns(UserWarning):  # catch re-emitted warning
+        ...     with pytest.warns(UserWarning, match=r'must be \d+$'):
+        ...         warnings.warn("this is not here", UserWarning)
+        Traceback (most recent call last):
+          ...
+        Failed: DID NOT WARN. No warnings of type ...UserWarning... were emitted...
+
+    **Using with** ``pytest.mark.parametrize``
+
+    When using :ref:`pytest.mark.parametrize ref` it is possible to parametrize tests
+    such that some runs raise a warning and others do not.
+
+    This could be achieved in the same way as with exceptions, see
+    :ref:`parametrizing_conditional_raising` for an example.
+
+    """
+    __tracebackhide__ = True
+    if not args:
+        if kwargs:
+            argnames = ", ".join(sorted(kwargs))
+            raise TypeError(
+                f"Unexpected keyword arguments passed to pytest.warns: {argnames}"
+                "\nUse context-manager form instead?"
+            )
+        return WarningsChecker(expected_warning, match_expr=match, _ispytest=True)
+    else:
+        func = args[0]
+        if not callable(func):
+            raise TypeError(f"{func!r} object (type: {type(func)}) must be callable")
+        with WarningsChecker(expected_warning, _ispytest=True):
+            return func(*args[1:], **kwargs)
+
+
+class WarningsRecorder(warnings.catch_warnings):  # type:ignore[type-arg]
+    """A context manager to record raised warnings.
+
+    Each recorded warning is an instance of :class:`warnings.WarningMessage`.
+
+    Adapted from `warnings.catch_warnings`.
+
+    .. note::
+        ``DeprecationWarning`` and ``PendingDeprecationWarning`` are treated
+        differently; see :ref:`ensuring_function_triggers`.
+
+    """
+
+    def __init__(self, *, _ispytest: bool = False) -> None:
+        check_ispytest(_ispytest)
+        super().__init__(record=True)
+        self._entered = False
+        self._list: list[warnings.WarningMessage] = []
+
+    @property
+    def list(self) -> list[warnings.WarningMessage]:
+        """The list of recorded warnings."""
+        return self._list
+
+    def __getitem__(self, i: int) -> warnings.WarningMessage:
+        """Get a recorded warning by index."""
+        return self._list[i]
+
+    def __iter__(self) -> Iterator[warnings.WarningMessage]:
+        """Iterate through the recorded warnings."""
+        return iter(self._list)
+
+    def __len__(self) -> int:
+        """The number of recorded warnings."""
+        return len(self._list)
+
+    def pop(self, cls: type[Warning] = Warning) -> warnings.WarningMessage:
+        """Pop the first recorded warning which is an instance of ``cls``,
+        but not an instance of a child class of any other match.
+        Raises ``AssertionError`` if there is no match.
+        """
+        best_idx: int | None = None
+        for i, w in enumerate(self._list):
+            if w.category == cls:
+                return self._list.pop(i)  # exact match, stop looking
+            if issubclass(w.category, cls) and (
+                best_idx is None
+                or not issubclass(w.category, self._list[best_idx].category)
+            ):
+                best_idx = i
+        if best_idx is not None:
+            return self._list.pop(best_idx)
+        __tracebackhide__ = True
+        raise AssertionError(f"{cls!r} not found in warning list")
+
+    def clear(self) -> None:
+        """Clear the list of recorded warnings."""
+        self._list[:] = []
+
+    def __enter__(self) -> Self:
+        if self._entered:
+            __tracebackhide__ = True
+            raise RuntimeError(f"Cannot enter {self!r} twice")
+        _list = super().__enter__()
+        # record=True means it's None.
+        assert _list is not None
+        self._list = _list
+        warnings.simplefilter("always")
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        if not self._entered:
+            __tracebackhide__ = True
+            raise RuntimeError(f"Cannot exit {self!r} without entering first")
+
+        super().__exit__(exc_type, exc_val, exc_tb)
+
+        # Built-in catch_warnings does not reset entered state so we do it
+        # manually here for this context manager to become reusable.
+        self._entered = False
+
+
+@final
+class WarningsChecker(WarningsRecorder):
+    def __init__(
+        self,
+        expected_warning: type[Warning] | tuple[type[Warning], ...] = Warning,
+        match_expr: str | Pattern[str] | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        super().__init__(_ispytest=True)
+
+        msg = "exceptions must be derived from Warning, not %s"
+        if isinstance(expected_warning, tuple):
+            for exc in expected_warning:
+                if not issubclass(exc, Warning):
+                    raise TypeError(msg % type(exc))
+            expected_warning_tup = expected_warning
+        elif isinstance(expected_warning, type) and issubclass(
+            expected_warning, Warning
+        ):
+            expected_warning_tup = (expected_warning,)
+        else:
+            raise TypeError(msg % type(expected_warning))
+
+        self.expected_warning = expected_warning_tup
+        self.match_expr = match_expr
+
+    def matches(self, warning: warnings.WarningMessage) -> bool:
+        assert self.expected_warning is not None
+        return issubclass(warning.category, self.expected_warning) and bool(
+            self.match_expr is None or re.search(self.match_expr, str(warning.message))
+        )
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        super().__exit__(exc_type, exc_val, exc_tb)
+
+        __tracebackhide__ = True
+
+        # BaseExceptions like pytest.{skip,fail,xfail,exit} or Ctrl-C within
+        # pytest.warns should *not* trigger "DID NOT WARN" and get suppressed
+        # when the warning doesn't happen. Control-flow exceptions should always
+        # propagate.
+        if exc_val is not None and (
+            not isinstance(exc_val, Exception)
+            # Exit is an Exception, not a BaseException, for some reason.
+            or isinstance(exc_val, Exit)
+        ):
+            return
+
+        def found_str() -> str:
+            return pformat([record.message for record in self], indent=2)
+
+        try:
+            if not any(issubclass(w.category, self.expected_warning) for w in self):
+                fail(
+                    f"DID NOT WARN. No warnings of type {self.expected_warning} were emitted.\n"
+                    f" Emitted warnings: {found_str()}."
+                )
+            elif not any(self.matches(w) for w in self):
+                fail(
+                    f"DID NOT WARN. No warnings of type {self.expected_warning} matching the regex were emitted.\n"
+                    f" Regex: {self.match_expr}\n"
+                    f" Emitted warnings: {found_str()}."
+                )
+        finally:
+            # Whether or not any warnings matched, we want to re-emit all unmatched warnings.
+            for w in self:
+                if not self.matches(w):
+                    warnings.warn_explicit(
+                        message=w.message,
+                        category=w.category,
+                        filename=w.filename,
+                        lineno=w.lineno,
+                        module=w.__module__,
+                        source=w.source,
+                    )
+
+            # Currently in Python it is possible to pass other types than an
+            # `str` message when creating `Warning` instances, however this
+            # causes an exception when :func:`warnings.filterwarnings` is used
+            # to filter those warnings. See
+            # https://github.com/python/cpython/issues/103577 for a discussion.
+            # While this can be considered a bug in CPython, we put guards in
+            # pytest as the error message produced without this check in place
+            # is confusing (#10865).
+            for w in self:
+                if type(w.message) is not UserWarning:
+                    # If the warning was of an incorrect type then `warnings.warn()`
+                    # creates a UserWarning. Any other warning must have been specified
+                    # explicitly.
+                    continue
+                if not w.message.args:
+                    # UserWarning() without arguments must have been specified explicitly.
+                    continue
+                msg = w.message.args[0]
+                if isinstance(msg, str):
+                    continue
+                # It's possible that UserWarning was explicitly specified, and
+                # its first argument was not a string. But that case can't be
+                # distinguished from an invalid type.
+                raise TypeError(
+                    f"Warning must be str or Warning, got {msg!r} (type {type(msg).__name__})"
+                )
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/reports.py b/backend/venv/lib/python3.12/site-packages/_pytest/reports.py
new file mode 100644
index 00000000..77cbf773
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/reports.py
@@ -0,0 +1,636 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+import dataclasses
+from io import StringIO
+import os
+from pprint import pprint
+from typing import Any
+from typing import cast
+from typing import final
+from typing import Iterable
+from typing import Iterator
+from typing import Literal
+from typing import Mapping
+from typing import NoReturn
+from typing import Sequence
+from typing import TYPE_CHECKING
+
+from _pytest._code.code import ExceptionChainRepr
+from _pytest._code.code import ExceptionInfo
+from _pytest._code.code import ExceptionRepr
+from _pytest._code.code import ReprEntry
+from _pytest._code.code import ReprEntryNative
+from _pytest._code.code import ReprExceptionInfo
+from _pytest._code.code import ReprFileLocation
+from _pytest._code.code import ReprFuncArgs
+from _pytest._code.code import ReprLocals
+from _pytest._code.code import ReprTraceback
+from _pytest._code.code import TerminalRepr
+from _pytest._io import TerminalWriter
+from _pytest.config import Config
+from _pytest.nodes import Collector
+from _pytest.nodes import Item
+from _pytest.outcomes import fail
+from _pytest.outcomes import skip
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from _pytest.runner import CallInfo
+
+
+def getworkerinfoline(node):
+    try:
+        return node._workerinfocache
+    except AttributeError:
+        d = node.workerinfo
+        ver = "{}.{}.{}".format(*d["version_info"][:3])
+        node._workerinfocache = s = "[{}] {} -- Python {} {}".format(
+            d["id"], d["sysplatform"], ver, d["executable"]
+        )
+        return s
+
+
+class BaseReport:
+    when: str | None
+    location: tuple[str, int | None, str] | None
+    longrepr: (
+        None | ExceptionInfo[BaseException] | tuple[str, int, str] | str | TerminalRepr
+    )
+    sections: list[tuple[str, str]]
+    nodeid: str
+    outcome: Literal["passed", "failed", "skipped"]
+
+    def __init__(self, **kw: Any) -> None:
+        self.__dict__.update(kw)
+
+    if TYPE_CHECKING:
+        # Can have arbitrary fields given to __init__().
+        def __getattr__(self, key: str) -> Any: ...
+
+    def toterminal(self, out: TerminalWriter) -> None:
+        if hasattr(self, "node"):
+            worker_info = getworkerinfoline(self.node)
+            if worker_info:
+                out.line(worker_info)
+
+        longrepr = self.longrepr
+        if longrepr is None:
+            return
+
+        if hasattr(longrepr, "toterminal"):
+            longrepr_terminal = cast(TerminalRepr, longrepr)
+            longrepr_terminal.toterminal(out)
+        else:
+            try:
+                s = str(longrepr)
+            except UnicodeEncodeError:
+                s = "<unprintable longrepr>"
+            out.line(s)
+
+    def get_sections(self, prefix: str) -> Iterator[tuple[str, str]]:
+        for name, content in self.sections:
+            if name.startswith(prefix):
+                yield prefix, content
+
+    @property
+    def longreprtext(self) -> str:
+        """Read-only property that returns the full string representation of
+        ``longrepr``.
+
+        .. versionadded:: 3.0
+        """
+        file = StringIO()
+        tw = TerminalWriter(file)
+        tw.hasmarkup = False
+        self.toterminal(tw)
+        exc = file.getvalue()
+        return exc.strip()
+
+    @property
+    def caplog(self) -> str:
+        """Return captured log lines, if log capturing is enabled.
+
+        .. versionadded:: 3.5
+        """
+        return "\n".join(
+            content for (prefix, content) in self.get_sections("Captured log")
+        )
+
+    @property
+    def capstdout(self) -> str:
+        """Return captured text from stdout, if capturing is enabled.
+
+        .. versionadded:: 3.0
+        """
+        return "".join(
+            content for (prefix, content) in self.get_sections("Captured stdout")
+        )
+
+    @property
+    def capstderr(self) -> str:
+        """Return captured text from stderr, if capturing is enabled.
+
+        .. versionadded:: 3.0
+        """
+        return "".join(
+            content for (prefix, content) in self.get_sections("Captured stderr")
+        )
+
+    @property
+    def passed(self) -> bool:
+        """Whether the outcome is passed."""
+        return self.outcome == "passed"
+
+    @property
+    def failed(self) -> bool:
+        """Whether the outcome is failed."""
+        return self.outcome == "failed"
+
+    @property
+    def skipped(self) -> bool:
+        """Whether the outcome is skipped."""
+        return self.outcome == "skipped"
+
+    @property
+    def fspath(self) -> str:
+        """The path portion of the reported node, as a string."""
+        return self.nodeid.split("::")[0]
+
+    @property
+    def count_towards_summary(self) -> bool:
+        """**Experimental** Whether this report should be counted towards the
+        totals shown at the end of the test session: "1 passed, 1 failure, etc".
+
+        .. note::
+
+            This function is considered **experimental**, so beware that it is subject to changes
+            even in patch releases.
+        """
+        return True
+
+    @property
+    def head_line(self) -> str | None:
+        """**Experimental** The head line shown with longrepr output for this
+        report, more commonly during traceback representation during
+        failures::
+
+            ________ Test.foo ________
+
+
+        In the example above, the head_line is "Test.foo".
+
+        .. note::
+
+            This function is considered **experimental**, so beware that it is subject to changes
+            even in patch releases.
+        """
+        if self.location is not None:
+            fspath, lineno, domain = self.location
+            return domain
+        return None
+
+    def _get_verbose_word_with_markup(
+        self, config: Config, default_markup: Mapping[str, bool]
+    ) -> tuple[str, Mapping[str, bool]]:
+        _category, _short, verbose = config.hook.pytest_report_teststatus(
+            report=self, config=config
+        )
+
+        if isinstance(verbose, str):
+            return verbose, default_markup
+
+        if isinstance(verbose, Sequence) and len(verbose) == 2:
+            word, markup = verbose
+            if isinstance(word, str) and isinstance(markup, Mapping):
+                return word, markup
+
+        fail(  # pragma: no cover
+            "pytest_report_teststatus() hook (from a plugin) returned "
+            f"an invalid verbose value: {verbose!r}.\nExpected either a string "
+            "or a tuple of (word, markup)."
+        )
+
+    def _to_json(self) -> dict[str, Any]:
+        """Return the contents of this report as a dict of builtin entries,
+        suitable for serialization.
+
+        This was originally the serialize_report() function from xdist (ca03269).
+
+        Experimental method.
+        """
+        return _report_to_json(self)
+
+    @classmethod
+    def _from_json(cls, reportdict: dict[str, object]) -> Self:
+        """Create either a TestReport or CollectReport, depending on the calling class.
+
+        It is the callers responsibility to know which class to pass here.
+
+        This was originally the serialize_report() function from xdist (ca03269).
+
+        Experimental method.
+        """
+        kwargs = _report_kwargs_from_json(reportdict)
+        return cls(**kwargs)
+
+
+def _report_unserialization_failure(
+    type_name: str, report_class: type[BaseReport], reportdict
+) -> NoReturn:
+    url = "https://github.com/pytest-dev/pytest/issues"
+    stream = StringIO()
+    pprint("-" * 100, stream=stream)
+    pprint(f"INTERNALERROR: Unknown entry type returned: {type_name}", stream=stream)
+    pprint(f"report_name: {report_class}", stream=stream)
+    pprint(reportdict, stream=stream)
+    pprint(f"Please report this bug at {url}", stream=stream)
+    pprint("-" * 100, stream=stream)
+    raise RuntimeError(stream.getvalue())
+
+
+@final
+class TestReport(BaseReport):
+    """Basic test report object (also used for setup and teardown calls if
+    they fail).
+
+    Reports can contain arbitrary extra attributes.
+    """
+
+    __test__ = False
+    # Defined by skipping plugin.
+    # xfail reason if xfailed, otherwise not defined. Use hasattr to distinguish.
+    wasxfail: str
+
+    def __init__(
+        self,
+        nodeid: str,
+        location: tuple[str, int | None, str],
+        keywords: Mapping[str, Any],
+        outcome: Literal["passed", "failed", "skipped"],
+        longrepr: None
+        | ExceptionInfo[BaseException]
+        | tuple[str, int, str]
+        | str
+        | TerminalRepr,
+        when: Literal["setup", "call", "teardown"],
+        sections: Iterable[tuple[str, str]] = (),
+        duration: float = 0,
+        start: float = 0,
+        stop: float = 0,
+        user_properties: Iterable[tuple[str, object]] | None = None,
+        **extra,
+    ) -> None:
+        #: Normalized collection nodeid.
+        self.nodeid = nodeid
+
+        #: A (filesystempath, lineno, domaininfo) tuple indicating the
+        #: actual location of a test item - it might be different from the
+        #: collected one e.g. if a method is inherited from a different module.
+        #: The filesystempath may be relative to ``config.rootdir``.
+        #: The line number is 0-based.
+        self.location: tuple[str, int | None, str] = location
+
+        #: A name -> value dictionary containing all keywords and
+        #: markers associated with a test invocation.
+        self.keywords: Mapping[str, Any] = keywords
+
+        #: Test outcome, always one of "passed", "failed", "skipped".
+        self.outcome = outcome
+
+        #: None or a failure representation.
+        self.longrepr = longrepr
+
+        #: One of 'setup', 'call', 'teardown' to indicate runtest phase.
+        self.when = when
+
+        #: User properties is a list of tuples (name, value) that holds user
+        #: defined properties of the test.
+        self.user_properties = list(user_properties or [])
+
+        #: Tuples of str ``(heading, content)`` with extra information
+        #: for the test report. Used by pytest to add text captured
+        #: from ``stdout``, ``stderr``, and intercepted logging events. May
+        #: be used by other plugins to add arbitrary information to reports.
+        self.sections = list(sections)
+
+        #: Time it took to run just the test.
+        self.duration: float = duration
+
+        #: The system time when the call started, in seconds since the epoch.
+        self.start: float = start
+        #: The system time when the call ended, in seconds since the epoch.
+        self.stop: float = stop
+
+        self.__dict__.update(extra)
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__} {self.nodeid!r} when={self.when!r} outcome={self.outcome!r}>"
+
+    @classmethod
+    def from_item_and_call(cls, item: Item, call: CallInfo[None]) -> TestReport:
+        """Create and fill a TestReport with standard item and call info.
+
+        :param item: The item.
+        :param call: The call info.
+        """
+        when = call.when
+        # Remove "collect" from the Literal type -- only for collection calls.
+        assert when != "collect"
+        duration = call.duration
+        start = call.start
+        stop = call.stop
+        keywords = {x: 1 for x in item.keywords}
+        excinfo = call.excinfo
+        sections = []
+        if not call.excinfo:
+            outcome: Literal["passed", "failed", "skipped"] = "passed"
+            longrepr: (
+                None
+                | ExceptionInfo[BaseException]
+                | tuple[str, int, str]
+                | str
+                | TerminalRepr
+            ) = None
+        else:
+            if not isinstance(excinfo, ExceptionInfo):
+                outcome = "failed"
+                longrepr = excinfo
+            elif isinstance(excinfo.value, skip.Exception):
+                outcome = "skipped"
+                r = excinfo._getreprcrash()
+                assert (
+                    r is not None
+                ), "There should always be a traceback entry for skipping a test."
+                if excinfo.value._use_item_location:
+                    path, line = item.reportinfo()[:2]
+                    assert line is not None
+                    longrepr = os.fspath(path), line + 1, r.message
+                else:
+                    longrepr = (str(r.path), r.lineno, r.message)
+            else:
+                outcome = "failed"
+                if call.when == "call":
+                    longrepr = item.repr_failure(excinfo)
+                else:  # exception in setup or teardown
+                    longrepr = item._repr_failure_py(
+                        excinfo, style=item.config.getoption("tbstyle", "auto")
+                    )
+        for rwhen, key, content in item._report_sections:
+            sections.append((f"Captured {key} {rwhen}", content))
+        return cls(
+            item.nodeid,
+            item.location,
+            keywords,
+            outcome,
+            longrepr,
+            when,
+            sections,
+            duration,
+            start,
+            stop,
+            user_properties=item.user_properties,
+        )
+
+
+@final
+class CollectReport(BaseReport):
+    """Collection report object.
+
+    Reports can contain arbitrary extra attributes.
+    """
+
+    when = "collect"
+
+    def __init__(
+        self,
+        nodeid: str,
+        outcome: Literal["passed", "failed", "skipped"],
+        longrepr: None
+        | ExceptionInfo[BaseException]
+        | tuple[str, int, str]
+        | str
+        | TerminalRepr,
+        result: list[Item | Collector] | None,
+        sections: Iterable[tuple[str, str]] = (),
+        **extra,
+    ) -> None:
+        #: Normalized collection nodeid.
+        self.nodeid = nodeid
+
+        #: Test outcome, always one of "passed", "failed", "skipped".
+        self.outcome = outcome
+
+        #: None or a failure representation.
+        self.longrepr = longrepr
+
+        #: The collected items and collection nodes.
+        self.result = result or []
+
+        #: Tuples of str ``(heading, content)`` with extra information
+        #: for the test report. Used by pytest to add text captured
+        #: from ``stdout``, ``stderr``, and intercepted logging events. May
+        #: be used by other plugins to add arbitrary information to reports.
+        self.sections = list(sections)
+
+        self.__dict__.update(extra)
+
+    @property
+    def location(  # type:ignore[override]
+        self,
+    ) -> tuple[str, int | None, str] | None:
+        return (self.fspath, None, self.fspath)
+
+    def __repr__(self) -> str:
+        return f"<CollectReport {self.nodeid!r} lenresult={len(self.result)} outcome={self.outcome!r}>"
+
+
+class CollectErrorRepr(TerminalRepr):
+    def __init__(self, msg: str) -> None:
+        self.longrepr = msg
+
+    def toterminal(self, out: TerminalWriter) -> None:
+        out.line(self.longrepr, red=True)
+
+
+def pytest_report_to_serializable(
+    report: CollectReport | TestReport,
+) -> dict[str, Any] | None:
+    if isinstance(report, (TestReport, CollectReport)):
+        data = report._to_json()
+        data["$report_type"] = report.__class__.__name__
+        return data
+    # TODO: Check if this is actually reachable.
+    return None  # type: ignore[unreachable]
+
+
+def pytest_report_from_serializable(
+    data: dict[str, Any],
+) -> CollectReport | TestReport | None:
+    if "$report_type" in data:
+        if data["$report_type"] == "TestReport":
+            return TestReport._from_json(data)
+        elif data["$report_type"] == "CollectReport":
+            return CollectReport._from_json(data)
+        assert False, "Unknown report_type unserialize data: {}".format(
+            data["$report_type"]
+        )
+    return None
+
+
+def _report_to_json(report: BaseReport) -> dict[str, Any]:
+    """Return the contents of this report as a dict of builtin entries,
+    suitable for serialization.
+
+    This was originally the serialize_report() function from xdist (ca03269).
+    """
+
+    def serialize_repr_entry(
+        entry: ReprEntry | ReprEntryNative,
+    ) -> dict[str, Any]:
+        data = dataclasses.asdict(entry)
+        for key, value in data.items():
+            if hasattr(value, "__dict__"):
+                data[key] = dataclasses.asdict(value)
+        entry_data = {"type": type(entry).__name__, "data": data}
+        return entry_data
+
+    def serialize_repr_traceback(reprtraceback: ReprTraceback) -> dict[str, Any]:
+        result = dataclasses.asdict(reprtraceback)
+        result["reprentries"] = [
+            serialize_repr_entry(x) for x in reprtraceback.reprentries
+        ]
+        return result
+
+    def serialize_repr_crash(
+        reprcrash: ReprFileLocation | None,
+    ) -> dict[str, Any] | None:
+        if reprcrash is not None:
+            return dataclasses.asdict(reprcrash)
+        else:
+            return None
+
+    def serialize_exception_longrepr(rep: BaseReport) -> dict[str, Any]:
+        assert rep.longrepr is not None
+        # TODO: Investigate whether the duck typing is really necessary here.
+        longrepr = cast(ExceptionRepr, rep.longrepr)
+        result: dict[str, Any] = {
+            "reprcrash": serialize_repr_crash(longrepr.reprcrash),
+            "reprtraceback": serialize_repr_traceback(longrepr.reprtraceback),
+            "sections": longrepr.sections,
+        }
+        if isinstance(longrepr, ExceptionChainRepr):
+            result["chain"] = []
+            for repr_traceback, repr_crash, description in longrepr.chain:
+                result["chain"].append(
+                    (
+                        serialize_repr_traceback(repr_traceback),
+                        serialize_repr_crash(repr_crash),
+                        description,
+                    )
+                )
+        else:
+            result["chain"] = None
+        return result
+
+    d = report.__dict__.copy()
+    if hasattr(report.longrepr, "toterminal"):
+        if hasattr(report.longrepr, "reprtraceback") and hasattr(
+            report.longrepr, "reprcrash"
+        ):
+            d["longrepr"] = serialize_exception_longrepr(report)
+        else:
+            d["longrepr"] = str(report.longrepr)
+    else:
+        d["longrepr"] = report.longrepr
+    for name in d:
+        if isinstance(d[name], os.PathLike):
+            d[name] = os.fspath(d[name])
+        elif name == "result":
+            d[name] = None  # for now
+    return d
+
+
+def _report_kwargs_from_json(reportdict: dict[str, Any]) -> dict[str, Any]:
+    """Return **kwargs that can be used to construct a TestReport or
+    CollectReport instance.
+
+    This was originally the serialize_report() function from xdist (ca03269).
+    """
+
+    def deserialize_repr_entry(entry_data):
+        data = entry_data["data"]
+        entry_type = entry_data["type"]
+        if entry_type == "ReprEntry":
+            reprfuncargs = None
+            reprfileloc = None
+            reprlocals = None
+            if data["reprfuncargs"]:
+                reprfuncargs = ReprFuncArgs(**data["reprfuncargs"])
+            if data["reprfileloc"]:
+                reprfileloc = ReprFileLocation(**data["reprfileloc"])
+            if data["reprlocals"]:
+                reprlocals = ReprLocals(data["reprlocals"]["lines"])
+
+            reprentry: ReprEntry | ReprEntryNative = ReprEntry(
+                lines=data["lines"],
+                reprfuncargs=reprfuncargs,
+                reprlocals=reprlocals,
+                reprfileloc=reprfileloc,
+                style=data["style"],
+            )
+        elif entry_type == "ReprEntryNative":
+            reprentry = ReprEntryNative(data["lines"])
+        else:
+            _report_unserialization_failure(entry_type, TestReport, reportdict)
+        return reprentry
+
+    def deserialize_repr_traceback(repr_traceback_dict):
+        repr_traceback_dict["reprentries"] = [
+            deserialize_repr_entry(x) for x in repr_traceback_dict["reprentries"]
+        ]
+        return ReprTraceback(**repr_traceback_dict)
+
+    def deserialize_repr_crash(repr_crash_dict: dict[str, Any] | None):
+        if repr_crash_dict is not None:
+            return ReprFileLocation(**repr_crash_dict)
+        else:
+            return None
+
+    if (
+        reportdict["longrepr"]
+        and "reprcrash" in reportdict["longrepr"]
+        and "reprtraceback" in reportdict["longrepr"]
+    ):
+        reprtraceback = deserialize_repr_traceback(
+            reportdict["longrepr"]["reprtraceback"]
+        )
+        reprcrash = deserialize_repr_crash(reportdict["longrepr"]["reprcrash"])
+        if reportdict["longrepr"]["chain"]:
+            chain = []
+            for repr_traceback_data, repr_crash_data, description in reportdict[
+                "longrepr"
+            ]["chain"]:
+                chain.append(
+                    (
+                        deserialize_repr_traceback(repr_traceback_data),
+                        deserialize_repr_crash(repr_crash_data),
+                        description,
+                    )
+                )
+            exception_info: ExceptionChainRepr | ReprExceptionInfo = ExceptionChainRepr(
+                chain
+            )
+        else:
+            exception_info = ReprExceptionInfo(
+                reprtraceback=reprtraceback,
+                reprcrash=reprcrash,
+            )
+
+        for section in reportdict["longrepr"]["sections"]:
+            exception_info.addsection(*section)
+        reportdict["longrepr"] = exception_info
+
+    return reportdict
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/runner.py b/backend/venv/lib/python3.12/site-packages/_pytest/runner.py
new file mode 100644
index 00000000..0b60301b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/runner.py
@@ -0,0 +1,571 @@
+# mypy: allow-untyped-defs
+"""Basic collect and runtest protocol implementations."""
+
+from __future__ import annotations
+
+import bdb
+import dataclasses
+import os
+import sys
+import types
+from typing import Callable
+from typing import cast
+from typing import final
+from typing import Generic
+from typing import Literal
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from .reports import BaseReport
+from .reports import CollectErrorRepr
+from .reports import CollectReport
+from .reports import TestReport
+from _pytest import timing
+from _pytest._code.code import ExceptionChainRepr
+from _pytest._code.code import ExceptionInfo
+from _pytest._code.code import TerminalRepr
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.nodes import Collector
+from _pytest.nodes import Directory
+from _pytest.nodes import Item
+from _pytest.nodes import Node
+from _pytest.outcomes import Exit
+from _pytest.outcomes import OutcomeException
+from _pytest.outcomes import Skipped
+from _pytest.outcomes import TEST_OUTCOME
+
+
+if sys.version_info < (3, 11):
+    from exceptiongroup import BaseExceptionGroup
+
+if TYPE_CHECKING:
+    from _pytest.main import Session
+    from _pytest.terminal import TerminalReporter
+
+#
+# pytest plugin hooks.
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("terminal reporting", "Reporting", after="general")
+    group.addoption(
+        "--durations",
+        action="store",
+        type=int,
+        default=None,
+        metavar="N",
+        help="Show N slowest setup/test durations (N=0 for all)",
+    )
+    group.addoption(
+        "--durations-min",
+        action="store",
+        type=float,
+        default=0.005,
+        metavar="N",
+        help="Minimal duration in seconds for inclusion in slowest list. "
+        "Default: 0.005.",
+    )
+
+
+def pytest_terminal_summary(terminalreporter: TerminalReporter) -> None:
+    durations = terminalreporter.config.option.durations
+    durations_min = terminalreporter.config.option.durations_min
+    verbose = terminalreporter.config.get_verbosity()
+    if durations is None:
+        return
+    tr = terminalreporter
+    dlist = []
+    for replist in tr.stats.values():
+        for rep in replist:
+            if hasattr(rep, "duration"):
+                dlist.append(rep)
+    if not dlist:
+        return
+    dlist.sort(key=lambda x: x.duration, reverse=True)
+    if not durations:
+        tr.write_sep("=", "slowest durations")
+    else:
+        tr.write_sep("=", f"slowest {durations} durations")
+        dlist = dlist[:durations]
+
+    for i, rep in enumerate(dlist):
+        if verbose < 2 and rep.duration < durations_min:
+            tr.write_line("")
+            tr.write_line(
+                f"({len(dlist) - i} durations < {durations_min:g}s hidden.  Use -vv to show these durations.)"
+            )
+            break
+        tr.write_line(f"{rep.duration:02.2f}s {rep.when:<8} {rep.nodeid}")
+
+
+def pytest_sessionstart(session: Session) -> None:
+    session._setupstate = SetupState()
+
+
+def pytest_sessionfinish(session: Session) -> None:
+    session._setupstate.teardown_exact(None)
+
+
+def pytest_runtest_protocol(item: Item, nextitem: Item | None) -> bool:
+    ihook = item.ihook
+    ihook.pytest_runtest_logstart(nodeid=item.nodeid, location=item.location)
+    runtestprotocol(item, nextitem=nextitem)
+    ihook.pytest_runtest_logfinish(nodeid=item.nodeid, location=item.location)
+    return True
+
+
+def runtestprotocol(
+    item: Item, log: bool = True, nextitem: Item | None = None
+) -> list[TestReport]:
+    hasrequest = hasattr(item, "_request")
+    if hasrequest and not item._request:  # type: ignore[attr-defined]
+        # This only happens if the item is re-run, as is done by
+        # pytest-rerunfailures.
+        item._initrequest()  # type: ignore[attr-defined]
+    rep = call_and_report(item, "setup", log)
+    reports = [rep]
+    if rep.passed:
+        if item.config.getoption("setupshow", False):
+            show_test_item(item)
+        if not item.config.getoption("setuponly", False):
+            reports.append(call_and_report(item, "call", log))
+    # If the session is about to fail or stop, teardown everything - this is
+    # necessary to correctly report fixture teardown errors (see #11706)
+    if item.session.shouldfail or item.session.shouldstop:
+        nextitem = None
+    reports.append(call_and_report(item, "teardown", log, nextitem=nextitem))
+    # After all teardown hooks have been called
+    # want funcargs and request info to go away.
+    if hasrequest:
+        item._request = False  # type: ignore[attr-defined]
+        item.funcargs = None  # type: ignore[attr-defined]
+    return reports
+
+
+def show_test_item(item: Item) -> None:
+    """Show test function, parameters and the fixtures of the test item."""
+    tw = item.config.get_terminal_writer()
+    tw.line()
+    tw.write(" " * 8)
+    tw.write(item.nodeid)
+    used_fixtures = sorted(getattr(item, "fixturenames", []))
+    if used_fixtures:
+        tw.write(" (fixtures used: {})".format(", ".join(used_fixtures)))
+    tw.flush()
+
+
+def pytest_runtest_setup(item: Item) -> None:
+    _update_current_test_var(item, "setup")
+    item.session._setupstate.setup(item)
+
+
+def pytest_runtest_call(item: Item) -> None:
+    _update_current_test_var(item, "call")
+    try:
+        del sys.last_type
+        del sys.last_value
+        del sys.last_traceback
+        if sys.version_info >= (3, 12, 0):
+            del sys.last_exc  # type:ignore[attr-defined]
+    except AttributeError:
+        pass
+    try:
+        item.runtest()
+    except Exception as e:
+        # Store trace info to allow postmortem debugging
+        sys.last_type = type(e)
+        sys.last_value = e
+        if sys.version_info >= (3, 12, 0):
+            sys.last_exc = e  # type:ignore[attr-defined]
+        assert e.__traceback__ is not None
+        # Skip *this* frame
+        sys.last_traceback = e.__traceback__.tb_next
+        raise
+
+
+def pytest_runtest_teardown(item: Item, nextitem: Item | None) -> None:
+    _update_current_test_var(item, "teardown")
+    item.session._setupstate.teardown_exact(nextitem)
+    _update_current_test_var(item, None)
+
+
+def _update_current_test_var(
+    item: Item, when: Literal["setup", "call", "teardown"] | None
+) -> None:
+    """Update :envvar:`PYTEST_CURRENT_TEST` to reflect the current item and stage.
+
+    If ``when`` is None, delete ``PYTEST_CURRENT_TEST`` from the environment.
+    """
+    var_name = "PYTEST_CURRENT_TEST"
+    if when:
+        value = f"{item.nodeid} ({when})"
+        # don't allow null bytes on environment variables (see #2644, #2957)
+        value = value.replace("\x00", "(null)")
+        os.environ[var_name] = value
+    else:
+        os.environ.pop(var_name)
+
+
+def pytest_report_teststatus(report: BaseReport) -> tuple[str, str, str] | None:
+    if report.when in ("setup", "teardown"):
+        if report.failed:
+            #      category, shortletter, verbose-word
+            return "error", "E", "ERROR"
+        elif report.skipped:
+            return "skipped", "s", "SKIPPED"
+        else:
+            return "", "", ""
+    return None
+
+
+#
+# Implementation
+
+
+def call_and_report(
+    item: Item, when: Literal["setup", "call", "teardown"], log: bool = True, **kwds
+) -> TestReport:
+    ihook = item.ihook
+    if when == "setup":
+        runtest_hook: Callable[..., None] = ihook.pytest_runtest_setup
+    elif when == "call":
+        runtest_hook = ihook.pytest_runtest_call
+    elif when == "teardown":
+        runtest_hook = ihook.pytest_runtest_teardown
+    else:
+        assert False, f"Unhandled runtest hook case: {when}"
+    reraise: tuple[type[BaseException], ...] = (Exit,)
+    if not item.config.getoption("usepdb", False):
+        reraise += (KeyboardInterrupt,)
+    call = CallInfo.from_call(
+        lambda: runtest_hook(item=item, **kwds), when=when, reraise=reraise
+    )
+    report: TestReport = ihook.pytest_runtest_makereport(item=item, call=call)
+    if log:
+        ihook.pytest_runtest_logreport(report=report)
+    if check_interactive_exception(call, report):
+        ihook.pytest_exception_interact(node=item, call=call, report=report)
+    return report
+
+
+def check_interactive_exception(call: CallInfo[object], report: BaseReport) -> bool:
+    """Check whether the call raised an exception that should be reported as
+    interactive."""
+    if call.excinfo is None:
+        # Didn't raise.
+        return False
+    if hasattr(report, "wasxfail"):
+        # Exception was expected.
+        return False
+    if isinstance(call.excinfo.value, (Skipped, bdb.BdbQuit)):
+        # Special control flow exception.
+        return False
+    return True
+
+
+TResult = TypeVar("TResult", covariant=True)
+
+
+@final
+@dataclasses.dataclass
+class CallInfo(Generic[TResult]):
+    """Result/Exception info of a function invocation."""
+
+    _result: TResult | None
+    #: The captured exception of the call, if it raised.
+    excinfo: ExceptionInfo[BaseException] | None
+    #: The system time when the call started, in seconds since the epoch.
+    start: float
+    #: The system time when the call ended, in seconds since the epoch.
+    stop: float
+    #: The call duration, in seconds.
+    duration: float
+    #: The context of invocation: "collect", "setup", "call" or "teardown".
+    when: Literal["collect", "setup", "call", "teardown"]
+
+    def __init__(
+        self,
+        result: TResult | None,
+        excinfo: ExceptionInfo[BaseException] | None,
+        start: float,
+        stop: float,
+        duration: float,
+        when: Literal["collect", "setup", "call", "teardown"],
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        self._result = result
+        self.excinfo = excinfo
+        self.start = start
+        self.stop = stop
+        self.duration = duration
+        self.when = when
+
+    @property
+    def result(self) -> TResult:
+        """The return value of the call, if it didn't raise.
+
+        Can only be accessed if excinfo is None.
+        """
+        if self.excinfo is not None:
+            raise AttributeError(f"{self!r} has no valid result")
+        # The cast is safe because an exception wasn't raised, hence
+        # _result has the expected function return type (which may be
+        #  None, that's why a cast and not an assert).
+        return cast(TResult, self._result)
+
+    @classmethod
+    def from_call(
+        cls,
+        func: Callable[[], TResult],
+        when: Literal["collect", "setup", "call", "teardown"],
+        reraise: type[BaseException] | tuple[type[BaseException], ...] | None = None,
+    ) -> CallInfo[TResult]:
+        """Call func, wrapping the result in a CallInfo.
+
+        :param func:
+            The function to call. Called without arguments.
+        :type func: Callable[[], _pytest.runner.TResult]
+        :param when:
+            The phase in which the function is called.
+        :param reraise:
+            Exception or exceptions that shall propagate if raised by the
+            function, instead of being wrapped in the CallInfo.
+        """
+        excinfo = None
+        start = timing.time()
+        precise_start = timing.perf_counter()
+        try:
+            result: TResult | None = func()
+        except BaseException:
+            excinfo = ExceptionInfo.from_current()
+            if reraise is not None and isinstance(excinfo.value, reraise):
+                raise
+            result = None
+        # use the perf counter
+        precise_stop = timing.perf_counter()
+        duration = precise_stop - precise_start
+        stop = timing.time()
+        return cls(
+            start=start,
+            stop=stop,
+            duration=duration,
+            when=when,
+            result=result,
+            excinfo=excinfo,
+            _ispytest=True,
+        )
+
+    def __repr__(self) -> str:
+        if self.excinfo is None:
+            return f"<CallInfo when={self.when!r} result: {self._result!r}>"
+        return f"<CallInfo when={self.when!r} excinfo={self.excinfo!r}>"
+
+
+def pytest_runtest_makereport(item: Item, call: CallInfo[None]) -> TestReport:
+    return TestReport.from_item_and_call(item, call)
+
+
+def pytest_make_collect_report(collector: Collector) -> CollectReport:
+    def collect() -> list[Item | Collector]:
+        # Before collecting, if this is a Directory, load the conftests.
+        # If a conftest import fails to load, it is considered a collection
+        # error of the Directory collector. This is why it's done inside of the
+        # CallInfo wrapper.
+        #
+        # Note: initial conftests are loaded early, not here.
+        if isinstance(collector, Directory):
+            collector.config.pluginmanager._loadconftestmodules(
+                collector.path,
+                collector.config.getoption("importmode"),
+                rootpath=collector.config.rootpath,
+                consider_namespace_packages=collector.config.getini(
+                    "consider_namespace_packages"
+                ),
+            )
+
+        return list(collector.collect())
+
+    call = CallInfo.from_call(
+        collect, "collect", reraise=(KeyboardInterrupt, SystemExit)
+    )
+    longrepr: None | tuple[str, int, str] | str | TerminalRepr = None
+    if not call.excinfo:
+        outcome: Literal["passed", "skipped", "failed"] = "passed"
+    else:
+        skip_exceptions = [Skipped]
+        unittest = sys.modules.get("unittest")
+        if unittest is not None:
+            skip_exceptions.append(unittest.SkipTest)
+        if isinstance(call.excinfo.value, tuple(skip_exceptions)):
+            outcome = "skipped"
+            r_ = collector._repr_failure_py(call.excinfo, "line")
+            assert isinstance(r_, ExceptionChainRepr), repr(r_)
+            r = r_.reprcrash
+            assert r
+            longrepr = (str(r.path), r.lineno, r.message)
+        else:
+            outcome = "failed"
+            errorinfo = collector.repr_failure(call.excinfo)
+            if not hasattr(errorinfo, "toterminal"):
+                assert isinstance(errorinfo, str)
+                errorinfo = CollectErrorRepr(errorinfo)
+            longrepr = errorinfo
+    result = call.result if not call.excinfo else None
+    rep = CollectReport(collector.nodeid, outcome, longrepr, result)
+    rep.call = call  # type: ignore # see collect_one_node
+    return rep
+
+
+class SetupState:
+    """Shared state for setting up/tearing down test items or collectors
+    in a session.
+
+    Suppose we have a collection tree as follows:
+
+    <Session session>
+        <Module mod1>
+            <Function item1>
+        <Module mod2>
+            <Function item2>
+
+    The SetupState maintains a stack. The stack starts out empty:
+
+        []
+
+    During the setup phase of item1, setup(item1) is called. What it does
+    is:
+
+        push session to stack, run session.setup()
+        push mod1 to stack, run mod1.setup()
+        push item1 to stack, run item1.setup()
+
+    The stack is:
+
+        [session, mod1, item1]
+
+    While the stack is in this shape, it is allowed to add finalizers to
+    each of session, mod1, item1 using addfinalizer().
+
+    During the teardown phase of item1, teardown_exact(item2) is called,
+    where item2 is the next item to item1. What it does is:
+
+        pop item1 from stack, run its teardowns
+        pop mod1 from stack, run its teardowns
+
+    mod1 was popped because it ended its purpose with item1. The stack is:
+
+        [session]
+
+    During the setup phase of item2, setup(item2) is called. What it does
+    is:
+
+        push mod2 to stack, run mod2.setup()
+        push item2 to stack, run item2.setup()
+
+    Stack:
+
+        [session, mod2, item2]
+
+    During the teardown phase of item2, teardown_exact(None) is called,
+    because item2 is the last item. What it does is:
+
+        pop item2 from stack, run its teardowns
+        pop mod2 from stack, run its teardowns
+        pop session from stack, run its teardowns
+
+    Stack:
+
+        []
+
+    The end!
+    """
+
+    def __init__(self) -> None:
+        # The stack is in the dict insertion order.
+        self.stack: dict[
+            Node,
+            tuple[
+                # Node's finalizers.
+                list[Callable[[], object]],
+                # Node's exception and original traceback, if its setup raised.
+                tuple[OutcomeException | Exception, types.TracebackType | None] | None,
+            ],
+        ] = {}
+
+    def setup(self, item: Item) -> None:
+        """Setup objects along the collector chain to the item."""
+        needed_collectors = item.listchain()
+
+        # If a collector fails its setup, fail its entire subtree of items.
+        # The setup is not retried for each item - the same exception is used.
+        for col, (finalizers, exc) in self.stack.items():
+            assert col in needed_collectors, "previous item was not torn down properly"
+            if exc:
+                raise exc[0].with_traceback(exc[1])
+
+        for col in needed_collectors[len(self.stack) :]:
+            assert col not in self.stack
+            # Push onto the stack.
+            self.stack[col] = ([col.teardown], None)
+            try:
+                col.setup()
+            except TEST_OUTCOME as exc:
+                self.stack[col] = (self.stack[col][0], (exc, exc.__traceback__))
+                raise
+
+    def addfinalizer(self, finalizer: Callable[[], object], node: Node) -> None:
+        """Attach a finalizer to the given node.
+
+        The node must be currently active in the stack.
+        """
+        assert node and not isinstance(node, tuple)
+        assert callable(finalizer)
+        assert node in self.stack, (node, self.stack)
+        self.stack[node][0].append(finalizer)
+
+    def teardown_exact(self, nextitem: Item | None) -> None:
+        """Teardown the current stack up until reaching nodes that nextitem
+        also descends from.
+
+        When nextitem is None (meaning we're at the last item), the entire
+        stack is torn down.
+        """
+        needed_collectors = nextitem and nextitem.listchain() or []
+        exceptions: list[BaseException] = []
+        while self.stack:
+            if list(self.stack.keys()) == needed_collectors[: len(self.stack)]:
+                break
+            node, (finalizers, _) = self.stack.popitem()
+            these_exceptions = []
+            while finalizers:
+                fin = finalizers.pop()
+                try:
+                    fin()
+                except TEST_OUTCOME as e:
+                    these_exceptions.append(e)
+
+            if len(these_exceptions) == 1:
+                exceptions.extend(these_exceptions)
+            elif these_exceptions:
+                msg = f"errors while tearing down {node!r}"
+                exceptions.append(BaseExceptionGroup(msg, these_exceptions[::-1]))
+
+        if len(exceptions) == 1:
+            raise exceptions[0]
+        elif exceptions:
+            raise BaseExceptionGroup("errors during test teardown", exceptions[::-1])
+        if nextitem is None:
+            assert not self.stack
+
+
+def collect_one_node(collector: Collector) -> CollectReport:
+    ihook = collector.ihook
+    ihook.pytest_collectstart(collector=collector)
+    rep: CollectReport = ihook.pytest_make_collect_report(collector=collector)
+    call = rep.__dict__.pop("call", None)
+    if call and check_interactive_exception(call, rep):
+        ihook.pytest_exception_interact(node=collector, call=call, report=rep)
+    return rep
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/scope.py b/backend/venv/lib/python3.12/site-packages/_pytest/scope.py
new file mode 100644
index 00000000..976a3ba2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/scope.py
@@ -0,0 +1,91 @@
+"""
+Scope definition and related utilities.
+
+Those are defined here, instead of in the 'fixtures' module because
+their use is spread across many other pytest modules, and centralizing it in 'fixtures'
+would cause circular references.
+
+Also this makes the module light to import, as it should.
+"""
+
+from __future__ import annotations
+
+from enum import Enum
+from functools import total_ordering
+from typing import Literal
+
+
+_ScopeName = Literal["session", "package", "module", "class", "function"]
+
+
+@total_ordering
+class Scope(Enum):
+    """
+    Represents one of the possible fixture scopes in pytest.
+
+    Scopes are ordered from lower to higher, that is:
+
+              ->>> higher ->>>
+
+    Function < Class < Module < Package < Session
+
+              <<<- lower  <<<-
+    """
+
+    # Scopes need to be listed from lower to higher.
+    Function: _ScopeName = "function"
+    Class: _ScopeName = "class"
+    Module: _ScopeName = "module"
+    Package: _ScopeName = "package"
+    Session: _ScopeName = "session"
+
+    def next_lower(self) -> Scope:
+        """Return the next lower scope."""
+        index = _SCOPE_INDICES[self]
+        if index == 0:
+            raise ValueError(f"{self} is the lower-most scope")
+        return _ALL_SCOPES[index - 1]
+
+    def next_higher(self) -> Scope:
+        """Return the next higher scope."""
+        index = _SCOPE_INDICES[self]
+        if index == len(_SCOPE_INDICES) - 1:
+            raise ValueError(f"{self} is the upper-most scope")
+        return _ALL_SCOPES[index + 1]
+
+    def __lt__(self, other: Scope) -> bool:
+        self_index = _SCOPE_INDICES[self]
+        other_index = _SCOPE_INDICES[other]
+        return self_index < other_index
+
+    @classmethod
+    def from_user(
+        cls, scope_name: _ScopeName, descr: str, where: str | None = None
+    ) -> Scope:
+        """
+        Given a scope name from the user, return the equivalent Scope enum. Should be used
+        whenever we want to convert a user provided scope name to its enum object.
+
+        If the scope name is invalid, construct a user friendly message and call pytest.fail.
+        """
+        from _pytest.outcomes import fail
+
+        try:
+            # Holding this reference is necessary for mypy at the moment.
+            scope = Scope(scope_name)
+        except ValueError:
+            fail(
+                "{} {}got an unexpected scope value '{}'".format(
+                    descr, f"from {where} " if where else "", scope_name
+                ),
+                pytrace=False,
+            )
+        return scope
+
+
+_ALL_SCOPES = list(Scope)
+_SCOPE_INDICES = {scope: index for index, scope in enumerate(_ALL_SCOPES)}
+
+
+# Ordered list of scopes which can contain many tests (in practice all except Function).
+HIGH_SCOPES = [x for x in Scope if x is not Scope.Function]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/setuponly.py b/backend/venv/lib/python3.12/site-packages/_pytest/setuponly.py
new file mode 100644
index 00000000..de297f40
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/setuponly.py
@@ -0,0 +1,102 @@
+from __future__ import annotations
+
+from typing import Generator
+
+from _pytest._io.saferepr import saferepr
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config.argparsing import Parser
+from _pytest.fixtures import FixtureDef
+from _pytest.fixtures import SubRequest
+from _pytest.scope import Scope
+import pytest
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("debugconfig")
+    group.addoption(
+        "--setuponly",
+        "--setup-only",
+        action="store_true",
+        help="Only setup fixtures, do not execute tests",
+    )
+    group.addoption(
+        "--setupshow",
+        "--setup-show",
+        action="store_true",
+        help="Show setup of fixtures while executing tests",
+    )
+
+
+@pytest.hookimpl(wrapper=True)
+def pytest_fixture_setup(
+    fixturedef: FixtureDef[object], request: SubRequest
+) -> Generator[None, object, object]:
+    try:
+        return (yield)
+    finally:
+        if request.config.option.setupshow:
+            if hasattr(request, "param"):
+                # Save the fixture parameter so ._show_fixture_action() can
+                # display it now and during the teardown (in .finish()).
+                if fixturedef.ids:
+                    if callable(fixturedef.ids):
+                        param = fixturedef.ids(request.param)
+                    else:
+                        param = fixturedef.ids[request.param_index]
+                else:
+                    param = request.param
+                fixturedef.cached_param = param  # type: ignore[attr-defined]
+            _show_fixture_action(fixturedef, request.config, "SETUP")
+
+
+def pytest_fixture_post_finalizer(
+    fixturedef: FixtureDef[object], request: SubRequest
+) -> None:
+    if fixturedef.cached_result is not None:
+        config = request.config
+        if config.option.setupshow:
+            _show_fixture_action(fixturedef, request.config, "TEARDOWN")
+            if hasattr(fixturedef, "cached_param"):
+                del fixturedef.cached_param
+
+
+def _show_fixture_action(
+    fixturedef: FixtureDef[object], config: Config, msg: str
+) -> None:
+    capman = config.pluginmanager.getplugin("capturemanager")
+    if capman:
+        capman.suspend_global_capture()
+
+    tw = config.get_terminal_writer()
+    tw.line()
+    # Use smaller indentation the higher the scope: Session = 0, Package = 1, etc.
+    scope_indent = list(reversed(Scope)).index(fixturedef._scope)
+    tw.write(" " * 2 * scope_indent)
+    tw.write(
+        "{step} {scope} {fixture}".format(  # noqa: UP032 (Readability)
+            step=msg.ljust(8),  # align the output to TEARDOWN
+            scope=fixturedef.scope[0].upper(),
+            fixture=fixturedef.argname,
+        )
+    )
+
+    if msg == "SETUP":
+        deps = sorted(arg for arg in fixturedef.argnames if arg != "request")
+        if deps:
+            tw.write(" (fixtures used: {})".format(", ".join(deps)))
+
+    if hasattr(fixturedef, "cached_param"):
+        tw.write(f"[{saferepr(fixturedef.cached_param, maxsize=42)}]")
+
+    tw.flush()
+
+    if capman:
+        capman.resume_global_capture()
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    if config.option.setuponly:
+        config.option.setupshow = True
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/setupplan.py b/backend/venv/lib/python3.12/site-packages/_pytest/setupplan.py
new file mode 100644
index 00000000..4e124cce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/setupplan.py
@@ -0,0 +1,39 @@
+from __future__ import annotations
+
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config.argparsing import Parser
+from _pytest.fixtures import FixtureDef
+from _pytest.fixtures import SubRequest
+import pytest
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("debugconfig")
+    group.addoption(
+        "--setupplan",
+        "--setup-plan",
+        action="store_true",
+        help="Show what fixtures and tests would be executed but "
+        "don't execute anything",
+    )
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_fixture_setup(
+    fixturedef: FixtureDef[object], request: SubRequest
+) -> object | None:
+    # Will return a dummy fixture if the setuponly option is provided.
+    if request.config.option.setupplan:
+        my_cache_key = fixturedef.cache_key(request)
+        fixturedef.cached_result = (None, my_cache_key, None)
+        return fixturedef.cached_result
+    return None
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_cmdline_main(config: Config) -> int | ExitCode | None:
+    if config.option.setupplan:
+        config.option.setuponly = True
+        config.option.setupshow = True
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/skipping.py b/backend/venv/lib/python3.12/site-packages/_pytest/skipping.py
new file mode 100644
index 00000000..9818be2a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/skipping.py
@@ -0,0 +1,301 @@
+# mypy: allow-untyped-defs
+"""Support for skip/xfail functions and markers."""
+
+from __future__ import annotations
+
+from collections.abc import Mapping
+import dataclasses
+import os
+import platform
+import sys
+import traceback
+from typing import Generator
+from typing import Optional
+
+from _pytest.config import Config
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.mark.structures import Mark
+from _pytest.nodes import Item
+from _pytest.outcomes import fail
+from _pytest.outcomes import skip
+from _pytest.outcomes import xfail
+from _pytest.reports import BaseReport
+from _pytest.reports import TestReport
+from _pytest.runner import CallInfo
+from _pytest.stash import StashKey
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group.addoption(
+        "--runxfail",
+        action="store_true",
+        dest="runxfail",
+        default=False,
+        help="Report the results of xfail tests as if they were not marked",
+    )
+
+    parser.addini(
+        "xfail_strict",
+        "Default for the strict parameter of xfail "
+        "markers when not given explicitly (default: False)",
+        default=False,
+        type="bool",
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    if config.option.runxfail:
+        # yay a hack
+        import pytest
+
+        old = pytest.xfail
+        config.add_cleanup(lambda: setattr(pytest, "xfail", old))
+
+        def nop(*args, **kwargs):
+            pass
+
+        nop.Exception = xfail.Exception  # type: ignore[attr-defined]
+        setattr(pytest, "xfail", nop)
+
+    config.addinivalue_line(
+        "markers",
+        "skip(reason=None): skip the given test function with an optional reason. "
+        'Example: skip(reason="no way of currently testing this") skips the '
+        "test.",
+    )
+    config.addinivalue_line(
+        "markers",
+        "skipif(condition, ..., *, reason=...): "
+        "skip the given test function if any of the conditions evaluate to True. "
+        "Example: skipif(sys.platform == 'win32') skips the test if we are on the win32 platform. "
+        "See https://docs.pytest.org/en/stable/reference/reference.html#pytest-mark-skipif",
+    )
+    config.addinivalue_line(
+        "markers",
+        "xfail(condition, ..., *, reason=..., run=True, raises=None, strict=xfail_strict): "
+        "mark the test function as an expected failure if any of the conditions "
+        "evaluate to True. Optionally specify a reason for better reporting "
+        "and run=False if you don't even want to execute the test function. "
+        "If only specific exception(s) are expected, you can list them in "
+        "raises, and if the test fails in other ways, it will be reported as "
+        "a true failure. See https://docs.pytest.org/en/stable/reference/reference.html#pytest-mark-xfail",
+    )
+
+
+def evaluate_condition(item: Item, mark: Mark, condition: object) -> tuple[bool, str]:
+    """Evaluate a single skipif/xfail condition.
+
+    If an old-style string condition is given, it is eval()'d, otherwise the
+    condition is bool()'d. If this fails, an appropriately formatted pytest.fail
+    is raised.
+
+    Returns (result, reason). The reason is only relevant if the result is True.
+    """
+    # String condition.
+    if isinstance(condition, str):
+        globals_ = {
+            "os": os,
+            "sys": sys,
+            "platform": platform,
+            "config": item.config,
+        }
+        for dictionary in reversed(
+            item.ihook.pytest_markeval_namespace(config=item.config)
+        ):
+            if not isinstance(dictionary, Mapping):
+                raise ValueError(
+                    f"pytest_markeval_namespace() needs to return a dict, got {dictionary!r}"
+                )
+            globals_.update(dictionary)
+        if hasattr(item, "obj"):
+            globals_.update(item.obj.__globals__)
+        try:
+            filename = f"<{mark.name} condition>"
+            condition_code = compile(condition, filename, "eval")
+            result = eval(condition_code, globals_)
+        except SyntaxError as exc:
+            msglines = [
+                f"Error evaluating {mark.name!r} condition",
+                "    " + condition,
+                "    " + " " * (exc.offset or 0) + "^",
+                "SyntaxError: invalid syntax",
+            ]
+            fail("\n".join(msglines), pytrace=False)
+        except Exception as exc:
+            msglines = [
+                f"Error evaluating {mark.name!r} condition",
+                "    " + condition,
+                *traceback.format_exception_only(type(exc), exc),
+            ]
+            fail("\n".join(msglines), pytrace=False)
+
+    # Boolean condition.
+    else:
+        try:
+            result = bool(condition)
+        except Exception as exc:
+            msglines = [
+                f"Error evaluating {mark.name!r} condition as a boolean",
+                *traceback.format_exception_only(type(exc), exc),
+            ]
+            fail("\n".join(msglines), pytrace=False)
+
+    reason = mark.kwargs.get("reason", None)
+    if reason is None:
+        if isinstance(condition, str):
+            reason = "condition: " + condition
+        else:
+            # XXX better be checked at collection time
+            msg = (
+                f"Error evaluating {mark.name!r}: "
+                + "you need to specify reason=STRING when using booleans as conditions."
+            )
+            fail(msg, pytrace=False)
+
+    return result, reason
+
+
+@dataclasses.dataclass(frozen=True)
+class Skip:
+    """The result of evaluate_skip_marks()."""
+
+    reason: str = "unconditional skip"
+
+
+def evaluate_skip_marks(item: Item) -> Skip | None:
+    """Evaluate skip and skipif marks on item, returning Skip if triggered."""
+    for mark in item.iter_markers(name="skipif"):
+        if "condition" not in mark.kwargs:
+            conditions = mark.args
+        else:
+            conditions = (mark.kwargs["condition"],)
+
+        # Unconditional.
+        if not conditions:
+            reason = mark.kwargs.get("reason", "")
+            return Skip(reason)
+
+        # If any of the conditions are true.
+        for condition in conditions:
+            result, reason = evaluate_condition(item, mark, condition)
+            if result:
+                return Skip(reason)
+
+    for mark in item.iter_markers(name="skip"):
+        try:
+            return Skip(*mark.args, **mark.kwargs)
+        except TypeError as e:
+            raise TypeError(str(e) + " - maybe you meant pytest.mark.skipif?") from None
+
+    return None
+
+
+@dataclasses.dataclass(frozen=True)
+class Xfail:
+    """The result of evaluate_xfail_marks()."""
+
+    __slots__ = ("reason", "run", "strict", "raises")
+
+    reason: str
+    run: bool
+    strict: bool
+    raises: tuple[type[BaseException], ...] | None
+
+
+def evaluate_xfail_marks(item: Item) -> Xfail | None:
+    """Evaluate xfail marks on item, returning Xfail if triggered."""
+    for mark in item.iter_markers(name="xfail"):
+        run = mark.kwargs.get("run", True)
+        strict = mark.kwargs.get("strict", item.config.getini("xfail_strict"))
+        raises = mark.kwargs.get("raises", None)
+        if "condition" not in mark.kwargs:
+            conditions = mark.args
+        else:
+            conditions = (mark.kwargs["condition"],)
+
+        # Unconditional.
+        if not conditions:
+            reason = mark.kwargs.get("reason", "")
+            return Xfail(reason, run, strict, raises)
+
+        # If any of the conditions are true.
+        for condition in conditions:
+            result, reason = evaluate_condition(item, mark, condition)
+            if result:
+                return Xfail(reason, run, strict, raises)
+
+    return None
+
+
+# Saves the xfail mark evaluation. Can be refreshed during call if None.
+xfailed_key = StashKey[Optional[Xfail]]()
+
+
+@hookimpl(tryfirst=True)
+def pytest_runtest_setup(item: Item) -> None:
+    skipped = evaluate_skip_marks(item)
+    if skipped:
+        raise skip.Exception(skipped.reason, _use_item_location=True)
+
+    item.stash[xfailed_key] = xfailed = evaluate_xfail_marks(item)
+    if xfailed and not item.config.option.runxfail and not xfailed.run:
+        xfail("[NOTRUN] " + xfailed.reason)
+
+
+@hookimpl(wrapper=True)
+def pytest_runtest_call(item: Item) -> Generator[None]:
+    xfailed = item.stash.get(xfailed_key, None)
+    if xfailed is None:
+        item.stash[xfailed_key] = xfailed = evaluate_xfail_marks(item)
+
+    if xfailed and not item.config.option.runxfail and not xfailed.run:
+        xfail("[NOTRUN] " + xfailed.reason)
+
+    try:
+        return (yield)
+    finally:
+        # The test run may have added an xfail mark dynamically.
+        xfailed = item.stash.get(xfailed_key, None)
+        if xfailed is None:
+            item.stash[xfailed_key] = xfailed = evaluate_xfail_marks(item)
+
+
+@hookimpl(wrapper=True)
+def pytest_runtest_makereport(
+    item: Item, call: CallInfo[None]
+) -> Generator[None, TestReport, TestReport]:
+    rep = yield
+    xfailed = item.stash.get(xfailed_key, None)
+    if item.config.option.runxfail:
+        pass  # don't interfere
+    elif call.excinfo and isinstance(call.excinfo.value, xfail.Exception):
+        assert call.excinfo.value.msg is not None
+        rep.wasxfail = "reason: " + call.excinfo.value.msg
+        rep.outcome = "skipped"
+    elif not rep.skipped and xfailed:
+        if call.excinfo:
+            raises = xfailed.raises
+            if raises is not None and not isinstance(call.excinfo.value, raises):
+                rep.outcome = "failed"
+            else:
+                rep.outcome = "skipped"
+                rep.wasxfail = xfailed.reason
+        elif call.when == "call":
+            if xfailed.strict:
+                rep.outcome = "failed"
+                rep.longrepr = "[XPASS(strict)] " + xfailed.reason
+            else:
+                rep.outcome = "passed"
+                rep.wasxfail = xfailed.reason
+    return rep
+
+
+def pytest_report_teststatus(report: BaseReport) -> tuple[str, str, str] | None:
+    if hasattr(report, "wasxfail"):
+        if report.skipped:
+            return "xfailed", "x", "XFAIL"
+        elif report.passed:
+            return "xpassed", "X", "XPASS"
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/stash.py b/backend/venv/lib/python3.12/site-packages/_pytest/stash.py
new file mode 100644
index 00000000..6a9ff884
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/stash.py
@@ -0,0 +1,116 @@
+from __future__ import annotations
+
+from typing import Any
+from typing import cast
+from typing import Generic
+from typing import TypeVar
+
+
+__all__ = ["Stash", "StashKey"]
+
+
+T = TypeVar("T")
+D = TypeVar("D")
+
+
+class StashKey(Generic[T]):
+    """``StashKey`` is an object used as a key to a :class:`Stash`.
+
+    A ``StashKey`` is associated with the type ``T`` of the value of the key.
+
+    A ``StashKey`` is unique and cannot conflict with another key.
+
+    .. versionadded:: 7.0
+    """
+
+    __slots__ = ()
+
+
+class Stash:
+    r"""``Stash`` is a type-safe heterogeneous mutable mapping that
+    allows keys and value types to be defined separately from
+    where it (the ``Stash``) is created.
+
+    Usually you will be given an object which has a ``Stash``, for example
+    :class:`~pytest.Config` or a :class:`~_pytest.nodes.Node`:
+
+    .. code-block:: python
+
+        stash: Stash = some_object.stash
+
+    If a module or plugin wants to store data in this ``Stash``, it creates
+    :class:`StashKey`\s for its keys (at the module level):
+
+    .. code-block:: python
+
+        # At the top-level of the module
+        some_str_key = StashKey[str]()
+        some_bool_key = StashKey[bool]()
+
+    To store information:
+
+    .. code-block:: python
+
+        # Value type must match the key.
+        stash[some_str_key] = "value"
+        stash[some_bool_key] = True
+
+    To retrieve the information:
+
+    .. code-block:: python
+
+        # The static type of some_str is str.
+        some_str = stash[some_str_key]
+        # The static type of some_bool is bool.
+        some_bool = stash[some_bool_key]
+
+    .. versionadded:: 7.0
+    """
+
+    __slots__ = ("_storage",)
+
+    def __init__(self) -> None:
+        self._storage: dict[StashKey[Any], object] = {}
+
+    def __setitem__(self, key: StashKey[T], value: T) -> None:
+        """Set a value for key."""
+        self._storage[key] = value
+
+    def __getitem__(self, key: StashKey[T]) -> T:
+        """Get the value for key.
+
+        Raises ``KeyError`` if the key wasn't set before.
+        """
+        return cast(T, self._storage[key])
+
+    def get(self, key: StashKey[T], default: D) -> T | D:
+        """Get the value for key, or return default if the key wasn't set
+        before."""
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def setdefault(self, key: StashKey[T], default: T) -> T:
+        """Return the value of key if already set, otherwise set the value
+        of key to default and return default."""
+        try:
+            return self[key]
+        except KeyError:
+            self[key] = default
+            return default
+
+    def __delitem__(self, key: StashKey[T]) -> None:
+        """Delete the value for key.
+
+        Raises ``KeyError`` if the key wasn't set before.
+        """
+        del self._storage[key]
+
+    def __contains__(self, key: StashKey[T]) -> bool:
+        """Return whether key was set."""
+        return key in self._storage
+
+    def __len__(self) -> int:
+        """Return how many items exist in the stash."""
+        return len(self._storage)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/stepwise.py b/backend/venv/lib/python3.12/site-packages/_pytest/stepwise.py
new file mode 100644
index 00000000..c7860808
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/stepwise.py
@@ -0,0 +1,125 @@
+from __future__ import annotations
+
+from _pytest import nodes
+from _pytest.cacheprovider import Cache
+from _pytest.config import Config
+from _pytest.config.argparsing import Parser
+from _pytest.main import Session
+from _pytest.reports import TestReport
+
+
+STEPWISE_CACHE_DIR = "cache/stepwise"
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("general")
+    group.addoption(
+        "--sw",
+        "--stepwise",
+        action="store_true",
+        default=False,
+        dest="stepwise",
+        help="Exit on test failure and continue from last failing test next time",
+    )
+    group.addoption(
+        "--sw-skip",
+        "--stepwise-skip",
+        action="store_true",
+        default=False,
+        dest="stepwise_skip",
+        help="Ignore the first failing test but stop on the next failing test. "
+        "Implicitly enables --stepwise.",
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    if config.option.stepwise_skip:
+        # allow --stepwise-skip to work on its own merits.
+        config.option.stepwise = True
+    if config.getoption("stepwise"):
+        config.pluginmanager.register(StepwisePlugin(config), "stepwiseplugin")
+
+
+def pytest_sessionfinish(session: Session) -> None:
+    if not session.config.getoption("stepwise"):
+        assert session.config.cache is not None
+        if hasattr(session.config, "workerinput"):
+            # Do not update cache if this process is a xdist worker to prevent
+            # race conditions (#10641).
+            return
+        # Clear the list of failing tests if the plugin is not active.
+        session.config.cache.set(STEPWISE_CACHE_DIR, [])
+
+
+class StepwisePlugin:
+    def __init__(self, config: Config) -> None:
+        self.config = config
+        self.session: Session | None = None
+        self.report_status = ""
+        assert config.cache is not None
+        self.cache: Cache = config.cache
+        self.lastfailed: str | None = self.cache.get(STEPWISE_CACHE_DIR, None)
+        self.skip: bool = config.getoption("stepwise_skip")
+
+    def pytest_sessionstart(self, session: Session) -> None:
+        self.session = session
+
+    def pytest_collection_modifyitems(
+        self, config: Config, items: list[nodes.Item]
+    ) -> None:
+        if not self.lastfailed:
+            self.report_status = "no previously failed tests, not skipping."
+            return
+
+        # check all item nodes until we find a match on last failed
+        failed_index = None
+        for index, item in enumerate(items):
+            if item.nodeid == self.lastfailed:
+                failed_index = index
+                break
+
+        # If the previously failed test was not found among the test items,
+        # do not skip any tests.
+        if failed_index is None:
+            self.report_status = "previously failed test not found, not skipping."
+        else:
+            self.report_status = f"skipping {failed_index} already passed items."
+            deselected = items[:failed_index]
+            del items[:failed_index]
+            config.hook.pytest_deselected(items=deselected)
+
+    def pytest_runtest_logreport(self, report: TestReport) -> None:
+        if report.failed:
+            if self.skip:
+                # Remove test from the failed ones (if it exists) and unset the skip option
+                # to make sure the following tests will not be skipped.
+                if report.nodeid == self.lastfailed:
+                    self.lastfailed = None
+
+                self.skip = False
+            else:
+                # Mark test as the last failing and interrupt the test session.
+                self.lastfailed = report.nodeid
+                assert self.session is not None
+                self.session.shouldstop = (
+                    "Test failed, continuing from this test next run."
+                )
+
+        else:
+            # If the test was actually run and did pass.
+            if report.when == "call":
+                # Remove test from the failed ones, if exists.
+                if report.nodeid == self.lastfailed:
+                    self.lastfailed = None
+
+    def pytest_report_collectionfinish(self) -> str | None:
+        if self.config.get_verbosity() >= 0 and self.report_status:
+            return f"stepwise: {self.report_status}"
+        return None
+
+    def pytest_sessionfinish(self) -> None:
+        if hasattr(self.config, "workerinput"):
+            # Do not update cache if this process is a xdist worker to prevent
+            # race conditions (#10641).
+            return
+        self.cache.set(STEPWISE_CACHE_DIR, self.lastfailed)
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/terminal.py b/backend/venv/lib/python3.12/site-packages/_pytest/terminal.py
new file mode 100644
index 00000000..ed267bf5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/terminal.py
@@ -0,0 +1,1577 @@
+# mypy: allow-untyped-defs
+"""Terminal reporting of the full testing process.
+
+This is a good source for looking at the various reporting hooks.
+"""
+
+from __future__ import annotations
+
+import argparse
+from collections import Counter
+import dataclasses
+import datetime
+from functools import partial
+import inspect
+from pathlib import Path
+import platform
+import sys
+import textwrap
+from typing import Any
+from typing import Callable
+from typing import ClassVar
+from typing import final
+from typing import Generator
+from typing import Literal
+from typing import Mapping
+from typing import NamedTuple
+from typing import Sequence
+from typing import TextIO
+from typing import TYPE_CHECKING
+import warnings
+
+import pluggy
+
+from _pytest import nodes
+from _pytest import timing
+from _pytest._code import ExceptionInfo
+from _pytest._code.code import ExceptionRepr
+from _pytest._io import TerminalWriter
+from _pytest._io.wcwidth import wcswidth
+import _pytest._version
+from _pytest.assertion.util import running_on_ci
+from _pytest.config import _PluggyPlugin
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.nodes import Item
+from _pytest.nodes import Node
+from _pytest.pathlib import absolutepath
+from _pytest.pathlib import bestrelpath
+from _pytest.reports import BaseReport
+from _pytest.reports import CollectReport
+from _pytest.reports import TestReport
+
+
+if TYPE_CHECKING:
+    from _pytest.main import Session
+
+
+REPORT_COLLECTING_RESOLUTION = 0.5
+
+KNOWN_TYPES = (
+    "failed",
+    "passed",
+    "skipped",
+    "deselected",
+    "xfailed",
+    "xpassed",
+    "warnings",
+    "error",
+)
+
+_REPORTCHARS_DEFAULT = "fE"
+
+
+class MoreQuietAction(argparse.Action):
+    """A modified copy of the argparse count action which counts down and updates
+    the legacy quiet attribute at the same time.
+
+    Used to unify verbosity handling.
+    """
+
+    def __init__(
+        self,
+        option_strings: Sequence[str],
+        dest: str,
+        default: object = None,
+        required: bool = False,
+        help: str | None = None,
+    ) -> None:
+        super().__init__(
+            option_strings=option_strings,
+            dest=dest,
+            nargs=0,
+            default=default,
+            required=required,
+            help=help,
+        )
+
+    def __call__(
+        self,
+        parser: argparse.ArgumentParser,
+        namespace: argparse.Namespace,
+        values: str | Sequence[object] | None,
+        option_string: str | None = None,
+    ) -> None:
+        new_count = getattr(namespace, self.dest, 0) - 1
+        setattr(namespace, self.dest, new_count)
+        # todo Deprecate config.quiet
+        namespace.quiet = getattr(namespace, "quiet", 0) + 1
+
+
+class TestShortLogReport(NamedTuple):
+    """Used to store the test status result category, shortletter and verbose word.
+    For example ``"rerun", "R", ("RERUN", {"yellow": True})``.
+
+    :ivar category:
+        The class of result, for example ``“passed”``, ``“skipped”``, ``“error”``, or the empty string.
+
+    :ivar letter:
+        The short letter shown as testing progresses, for example ``"."``, ``"s"``, ``"E"``, or the empty string.
+
+    :ivar word:
+        Verbose word is shown as testing progresses in verbose mode, for example ``"PASSED"``, ``"SKIPPED"``,
+        ``"ERROR"``, or the empty string.
+    """
+
+    category: str
+    letter: str
+    word: str | tuple[str, Mapping[str, bool]]
+
+
+def pytest_addoption(parser: Parser) -> None:
+    group = parser.getgroup("terminal reporting", "Reporting", after="general")
+    group._addoption(
+        "-v",
+        "--verbose",
+        action="count",
+        default=0,
+        dest="verbose",
+        help="Increase verbosity",
+    )
+    group._addoption(
+        "--no-header",
+        action="store_true",
+        default=False,
+        dest="no_header",
+        help="Disable header",
+    )
+    group._addoption(
+        "--no-summary",
+        action="store_true",
+        default=False,
+        dest="no_summary",
+        help="Disable summary",
+    )
+    group._addoption(
+        "--no-fold-skipped",
+        action="store_false",
+        dest="fold_skipped",
+        default=True,
+        help="Do not fold skipped tests in short summary.",
+    )
+    group._addoption(
+        "-q",
+        "--quiet",
+        action=MoreQuietAction,
+        default=0,
+        dest="verbose",
+        help="Decrease verbosity",
+    )
+    group._addoption(
+        "--verbosity",
+        dest="verbose",
+        type=int,
+        default=0,
+        help="Set verbosity. Default: 0.",
+    )
+    group._addoption(
+        "-r",
+        action="store",
+        dest="reportchars",
+        default=_REPORTCHARS_DEFAULT,
+        metavar="chars",
+        help="Show extra test summary info as specified by chars: (f)ailed, "
+        "(E)rror, (s)kipped, (x)failed, (X)passed, "
+        "(p)assed, (P)assed with output, (a)ll except passed (p/P), or (A)ll. "
+        "(w)arnings are enabled by default (see --disable-warnings), "
+        "'N' can be used to reset the list. (default: 'fE').",
+    )
+    group._addoption(
+        "--disable-warnings",
+        "--disable-pytest-warnings",
+        default=False,
+        dest="disable_warnings",
+        action="store_true",
+        help="Disable warnings summary",
+    )
+    group._addoption(
+        "-l",
+        "--showlocals",
+        action="store_true",
+        dest="showlocals",
+        default=False,
+        help="Show locals in tracebacks (disabled by default)",
+    )
+    group._addoption(
+        "--no-showlocals",
+        action="store_false",
+        dest="showlocals",
+        help="Hide locals in tracebacks (negate --showlocals passed through addopts)",
+    )
+    group._addoption(
+        "--tb",
+        metavar="style",
+        action="store",
+        dest="tbstyle",
+        default="auto",
+        choices=["auto", "long", "short", "no", "line", "native"],
+        help="Traceback print mode (auto/long/short/line/native/no)",
+    )
+    group._addoption(
+        "--xfail-tb",
+        action="store_true",
+        dest="xfail_tb",
+        default=False,
+        help="Show tracebacks for xfail (as long as --tb != no)",
+    )
+    group._addoption(
+        "--show-capture",
+        action="store",
+        dest="showcapture",
+        choices=["no", "stdout", "stderr", "log", "all"],
+        default="all",
+        help="Controls how captured stdout/stderr/log is shown on failed tests. "
+        "Default: all.",
+    )
+    group._addoption(
+        "--fulltrace",
+        "--full-trace",
+        action="store_true",
+        default=False,
+        help="Don't cut any tracebacks (default is to cut)",
+    )
+    group._addoption(
+        "--color",
+        metavar="color",
+        action="store",
+        dest="color",
+        default="auto",
+        choices=["yes", "no", "auto"],
+        help="Color terminal output (yes/no/auto)",
+    )
+    group._addoption(
+        "--code-highlight",
+        default="yes",
+        choices=["yes", "no"],
+        help="Whether code should be highlighted (only if --color is also enabled). "
+        "Default: yes.",
+    )
+
+    parser.addini(
+        "console_output_style",
+        help='Console output: "classic", or with additional progress information '
+        '("progress" (percentage) | "count" | "progress-even-when-capture-no" (forces '
+        "progress even when capture=no)",
+        default="progress",
+    )
+    Config._add_verbosity_ini(
+        parser,
+        Config.VERBOSITY_TEST_CASES,
+        help=(
+            "Specify a verbosity level for test case execution, overriding the main level. "
+            "Higher levels will provide more detailed information about each test case executed."
+        ),
+    )
+
+
+def pytest_configure(config: Config) -> None:
+    reporter = TerminalReporter(config, sys.stdout)
+    config.pluginmanager.register(reporter, "terminalreporter")
+    if config.option.debug or config.option.traceconfig:
+
+        def mywriter(tags, args):
+            msg = " ".join(map(str, args))
+            reporter.write_line("[traceconfig] " + msg)
+
+        config.trace.root.setprocessor("pytest:config", mywriter)
+
+
+def getreportopt(config: Config) -> str:
+    reportchars: str = config.option.reportchars
+
+    old_aliases = {"F", "S"}
+    reportopts = ""
+    for char in reportchars:
+        if char in old_aliases:
+            char = char.lower()
+        if char == "a":
+            reportopts = "sxXEf"
+        elif char == "A":
+            reportopts = "PpsxXEf"
+        elif char == "N":
+            reportopts = ""
+        elif char not in reportopts:
+            reportopts += char
+
+    if not config.option.disable_warnings and "w" not in reportopts:
+        reportopts = "w" + reportopts
+    elif config.option.disable_warnings and "w" in reportopts:
+        reportopts = reportopts.replace("w", "")
+
+    return reportopts
+
+
+@hookimpl(trylast=True)  # after _pytest.runner
+def pytest_report_teststatus(report: BaseReport) -> tuple[str, str, str]:
+    letter = "F"
+    if report.passed:
+        letter = "."
+    elif report.skipped:
+        letter = "s"
+
+    outcome: str = report.outcome
+    if report.when in ("collect", "setup", "teardown") and outcome == "failed":
+        outcome = "error"
+        letter = "E"
+
+    return outcome, letter, outcome.upper()
+
+
+@dataclasses.dataclass
+class WarningReport:
+    """Simple structure to hold warnings information captured by ``pytest_warning_recorded``.
+
+    :ivar str message:
+        User friendly message about the warning.
+    :ivar str|None nodeid:
+        nodeid that generated the warning (see ``get_location``).
+    :ivar tuple fslocation:
+        File system location of the source of the warning (see ``get_location``).
+    """
+
+    message: str
+    nodeid: str | None = None
+    fslocation: tuple[str, int] | None = None
+
+    count_towards_summary: ClassVar = True
+
+    def get_location(self, config: Config) -> str | None:
+        """Return the more user-friendly information about the location of a warning, or None."""
+        if self.nodeid:
+            return self.nodeid
+        if self.fslocation:
+            filename, linenum = self.fslocation
+            relpath = bestrelpath(config.invocation_params.dir, absolutepath(filename))
+            return f"{relpath}:{linenum}"
+        return None
+
+
+@final
+class TerminalReporter:
+    def __init__(self, config: Config, file: TextIO | None = None) -> None:
+        import _pytest.config
+
+        self.config = config
+        self._numcollected = 0
+        self._session: Session | None = None
+        self._showfspath: bool | None = None
+
+        self.stats: dict[str, list[Any]] = {}
+        self._main_color: str | None = None
+        self._known_types: list[str] | None = None
+        self.startpath = config.invocation_params.dir
+        if file is None:
+            file = sys.stdout
+        self._tw = _pytest.config.create_terminal_writer(config, file)
+        self._screen_width = self._tw.fullwidth
+        self.currentfspath: None | Path | str | int = None
+        self.reportchars = getreportopt(config)
+        self.foldskipped = config.option.fold_skipped
+        self.hasmarkup = self._tw.hasmarkup
+        self.isatty = file.isatty()
+        self._progress_nodeids_reported: set[str] = set()
+        self._show_progress_info = self._determine_show_progress_info()
+        self._collect_report_last_write: float | None = None
+        self._already_displayed_warnings: int | None = None
+        self._keyboardinterrupt_memo: ExceptionRepr | None = None
+
+    def _determine_show_progress_info(self) -> Literal["progress", "count", False]:
+        """Return whether we should display progress information based on the current config."""
+        # do not show progress if we are not capturing output (#3038) unless explicitly
+        # overridden by progress-even-when-capture-no
+        if (
+            self.config.getoption("capture", "no") == "no"
+            and self.config.getini("console_output_style")
+            != "progress-even-when-capture-no"
+        ):
+            return False
+        # do not show progress if we are showing fixture setup/teardown
+        if self.config.getoption("setupshow", False):
+            return False
+        cfg: str = self.config.getini("console_output_style")
+        if cfg in {"progress", "progress-even-when-capture-no"}:
+            return "progress"
+        elif cfg == "count":
+            return "count"
+        else:
+            return False
+
+    @property
+    def verbosity(self) -> int:
+        verbosity: int = self.config.option.verbose
+        return verbosity
+
+    @property
+    def showheader(self) -> bool:
+        return self.verbosity >= 0
+
+    @property
+    def no_header(self) -> bool:
+        return bool(self.config.option.no_header)
+
+    @property
+    def no_summary(self) -> bool:
+        return bool(self.config.option.no_summary)
+
+    @property
+    def showfspath(self) -> bool:
+        if self._showfspath is None:
+            return self.config.get_verbosity(Config.VERBOSITY_TEST_CASES) >= 0
+        return self._showfspath
+
+    @showfspath.setter
+    def showfspath(self, value: bool | None) -> None:
+        self._showfspath = value
+
+    @property
+    def showlongtestinfo(self) -> bool:
+        return self.config.get_verbosity(Config.VERBOSITY_TEST_CASES) > 0
+
+    def hasopt(self, char: str) -> bool:
+        char = {"xfailed": "x", "skipped": "s"}.get(char, char)
+        return char in self.reportchars
+
+    def write_fspath_result(self, nodeid: str, res: str, **markup: bool) -> None:
+        fspath = self.config.rootpath / nodeid.split("::")[0]
+        if self.currentfspath is None or fspath != self.currentfspath:
+            if self.currentfspath is not None and self._show_progress_info:
+                self._write_progress_information_filling_space()
+            self.currentfspath = fspath
+            relfspath = bestrelpath(self.startpath, fspath)
+            self._tw.line()
+            self._tw.write(relfspath + " ")
+        self._tw.write(res, flush=True, **markup)
+
+    def write_ensure_prefix(self, prefix: str, extra: str = "", **kwargs) -> None:
+        if self.currentfspath != prefix:
+            self._tw.line()
+            self.currentfspath = prefix
+            self._tw.write(prefix)
+        if extra:
+            self._tw.write(extra, **kwargs)
+            self.currentfspath = -2
+
+    def ensure_newline(self) -> None:
+        if self.currentfspath:
+            self._tw.line()
+            self.currentfspath = None
+
+    def wrap_write(
+        self,
+        content: str,
+        *,
+        flush: bool = False,
+        margin: int = 8,
+        line_sep: str = "\n",
+        **markup: bool,
+    ) -> None:
+        """Wrap message with margin for progress info."""
+        width_of_current_line = self._tw.width_of_current_line
+        wrapped = line_sep.join(
+            textwrap.wrap(
+                " " * width_of_current_line + content,
+                width=self._screen_width - margin,
+                drop_whitespace=True,
+                replace_whitespace=False,
+            ),
+        )
+        wrapped = wrapped[width_of_current_line:]
+        self._tw.write(wrapped, flush=flush, **markup)
+
+    def write(self, content: str, *, flush: bool = False, **markup: bool) -> None:
+        self._tw.write(content, flush=flush, **markup)
+
+    def flush(self) -> None:
+        self._tw.flush()
+
+    def write_line(self, line: str | bytes, **markup: bool) -> None:
+        if not isinstance(line, str):
+            line = str(line, errors="replace")
+        self.ensure_newline()
+        self._tw.line(line, **markup)
+
+    def rewrite(self, line: str, **markup: bool) -> None:
+        """Rewinds the terminal cursor to the beginning and writes the given line.
+
+        :param erase:
+            If True, will also add spaces until the full terminal width to ensure
+            previous lines are properly erased.
+
+        The rest of the keyword arguments are markup instructions.
+        """
+        erase = markup.pop("erase", False)
+        if erase:
+            fill_count = self._tw.fullwidth - len(line) - 1
+            fill = " " * fill_count
+        else:
+            fill = ""
+        line = str(line)
+        self._tw.write("\r" + line + fill, **markup)
+
+    def write_sep(
+        self,
+        sep: str,
+        title: str | None = None,
+        fullwidth: int | None = None,
+        **markup: bool,
+    ) -> None:
+        self.ensure_newline()
+        self._tw.sep(sep, title, fullwidth, **markup)
+
+    def section(self, title: str, sep: str = "=", **kw: bool) -> None:
+        self._tw.sep(sep, title, **kw)
+
+    def line(self, msg: str, **kw: bool) -> None:
+        self._tw.line(msg, **kw)
+
+    def _add_stats(self, category: str, items: Sequence[Any]) -> None:
+        set_main_color = category not in self.stats
+        self.stats.setdefault(category, []).extend(items)
+        if set_main_color:
+            self._set_main_color()
+
+    def pytest_internalerror(self, excrepr: ExceptionRepr) -> bool:
+        for line in str(excrepr).split("\n"):
+            self.write_line("INTERNALERROR> " + line)
+        return True
+
+    def pytest_warning_recorded(
+        self,
+        warning_message: warnings.WarningMessage,
+        nodeid: str,
+    ) -> None:
+        from _pytest.warnings import warning_record_to_str
+
+        fslocation = warning_message.filename, warning_message.lineno
+        message = warning_record_to_str(warning_message)
+
+        warning_report = WarningReport(
+            fslocation=fslocation, message=message, nodeid=nodeid
+        )
+        self._add_stats("warnings", [warning_report])
+
+    def pytest_plugin_registered(self, plugin: _PluggyPlugin) -> None:
+        if self.config.option.traceconfig:
+            msg = f"PLUGIN registered: {plugin}"
+            # XXX This event may happen during setup/teardown time
+            #     which unfortunately captures our output here
+            #     which garbles our output if we use self.write_line.
+            self.write_line(msg)
+
+    def pytest_deselected(self, items: Sequence[Item]) -> None:
+        self._add_stats("deselected", items)
+
+    def pytest_runtest_logstart(
+        self, nodeid: str, location: tuple[str, int | None, str]
+    ) -> None:
+        fspath, lineno, domain = location
+        # Ensure that the path is printed before the
+        # 1st test of a module starts running.
+        if self.showlongtestinfo:
+            line = self._locationline(nodeid, fspath, lineno, domain)
+            self.write_ensure_prefix(line, "")
+            self.flush()
+        elif self.showfspath:
+            self.write_fspath_result(nodeid, "")
+            self.flush()
+
+    def pytest_runtest_logreport(self, report: TestReport) -> None:
+        self._tests_ran = True
+        rep = report
+
+        res = TestShortLogReport(
+            *self.config.hook.pytest_report_teststatus(report=rep, config=self.config)
+        )
+        category, letter, word = res.category, res.letter, res.word
+        if not isinstance(word, tuple):
+            markup = None
+        else:
+            word, markup = word
+        self._add_stats(category, [rep])
+        if not letter and not word:
+            # Probably passed setup/teardown.
+            return
+        if markup is None:
+            was_xfail = hasattr(report, "wasxfail")
+            if rep.passed and not was_xfail:
+                markup = {"green": True}
+            elif rep.passed and was_xfail:
+                markup = {"yellow": True}
+            elif rep.failed:
+                markup = {"red": True}
+            elif rep.skipped:
+                markup = {"yellow": True}
+            else:
+                markup = {}
+        self._progress_nodeids_reported.add(rep.nodeid)
+        if self.config.get_verbosity(Config.VERBOSITY_TEST_CASES) <= 0:
+            self._tw.write(letter, **markup)
+            # When running in xdist, the logreport and logfinish of multiple
+            # items are interspersed, e.g. `logreport`, `logreport`,
+            # `logfinish`, `logfinish`. To avoid the "past edge" calculation
+            # from getting confused and overflowing (#7166), do the past edge
+            # printing here and not in logfinish, except for the 100% which
+            # should only be printed after all teardowns are finished.
+            if self._show_progress_info and not self._is_last_item:
+                self._write_progress_information_if_past_edge()
+        else:
+            line = self._locationline(rep.nodeid, *rep.location)
+            running_xdist = hasattr(rep, "node")
+            if not running_xdist:
+                self.write_ensure_prefix(line, word, **markup)
+                if rep.skipped or hasattr(report, "wasxfail"):
+                    reason = _get_raw_skip_reason(rep)
+                    if self.config.get_verbosity(Config.VERBOSITY_TEST_CASES) < 2:
+                        available_width = (
+                            (self._tw.fullwidth - self._tw.width_of_current_line)
+                            - len(" [100%]")
+                            - 1
+                        )
+                        formatted_reason = _format_trimmed(
+                            " ({})", reason, available_width
+                        )
+                    else:
+                        formatted_reason = f" ({reason})"
+
+                    if reason and formatted_reason is not None:
+                        self.wrap_write(formatted_reason)
+                if self._show_progress_info:
+                    self._write_progress_information_filling_space()
+            else:
+                self.ensure_newline()
+                self._tw.write(f"[{rep.node.gateway.id}]")
+                if self._show_progress_info:
+                    self._tw.write(
+                        self._get_progress_information_message() + " ", cyan=True
+                    )
+                else:
+                    self._tw.write(" ")
+                self._tw.write(word, **markup)
+                self._tw.write(" " + line)
+                self.currentfspath = -2
+        self.flush()
+
+    @property
+    def _is_last_item(self) -> bool:
+        assert self._session is not None
+        return len(self._progress_nodeids_reported) == self._session.testscollected
+
+    @hookimpl(wrapper=True)
+    def pytest_runtestloop(self) -> Generator[None, object, object]:
+        result = yield
+
+        # Write the final/100% progress -- deferred until the loop is complete.
+        if (
+            self.config.get_verbosity(Config.VERBOSITY_TEST_CASES) <= 0
+            and self._show_progress_info
+            and self._progress_nodeids_reported
+        ):
+            self._write_progress_information_filling_space()
+
+        return result
+
+    def _get_progress_information_message(self) -> str:
+        assert self._session
+        collected = self._session.testscollected
+        if self._show_progress_info == "count":
+            if collected:
+                progress = len(self._progress_nodeids_reported)
+                counter_format = f"{{:{len(str(collected))}d}}"
+                format_string = f" [{counter_format}/{{}}]"
+                return format_string.format(progress, collected)
+            return f" [ {collected} / {collected} ]"
+        else:
+            if collected:
+                return (
+                    f" [{len(self._progress_nodeids_reported) * 100 // collected:3d}%]"
+                )
+            return " [100%]"
+
+    def _write_progress_information_if_past_edge(self) -> None:
+        w = self._width_of_current_line
+        if self._show_progress_info == "count":
+            assert self._session
+            num_tests = self._session.testscollected
+            progress_length = len(f" [{num_tests}/{num_tests}]")
+        else:
+            progress_length = len(" [100%]")
+        past_edge = w + progress_length + 1 >= self._screen_width
+        if past_edge:
+            main_color, _ = self._get_main_color()
+            msg = self._get_progress_information_message()
+            self._tw.write(msg + "\n", **{main_color: True})
+
+    def _write_progress_information_filling_space(self) -> None:
+        color, _ = self._get_main_color()
+        msg = self._get_progress_information_message()
+        w = self._width_of_current_line
+        fill = self._tw.fullwidth - w - 1
+        self.write(msg.rjust(fill), flush=True, **{color: True})
+
+    @property
+    def _width_of_current_line(self) -> int:
+        """Return the width of the current line."""
+        return self._tw.width_of_current_line
+
+    def pytest_collection(self) -> None:
+        if self.isatty:
+            if self.config.option.verbose >= 0:
+                self.write("collecting ... ", flush=True, bold=True)
+                self._collect_report_last_write = timing.time()
+        elif self.config.option.verbose >= 1:
+            self.write("collecting ... ", flush=True, bold=True)
+
+    def pytest_collectreport(self, report: CollectReport) -> None:
+        if report.failed:
+            self._add_stats("error", [report])
+        elif report.skipped:
+            self._add_stats("skipped", [report])
+        items = [x for x in report.result if isinstance(x, Item)]
+        self._numcollected += len(items)
+        if self.isatty:
+            self.report_collect()
+
+    def report_collect(self, final: bool = False) -> None:
+        if self.config.option.verbose < 0:
+            return
+
+        if not final:
+            # Only write "collecting" report every 0.5s.
+            t = timing.time()
+            if (
+                self._collect_report_last_write is not None
+                and self._collect_report_last_write > t - REPORT_COLLECTING_RESOLUTION
+            ):
+                return
+            self._collect_report_last_write = t
+
+        errors = len(self.stats.get("error", []))
+        skipped = len(self.stats.get("skipped", []))
+        deselected = len(self.stats.get("deselected", []))
+        selected = self._numcollected - deselected
+        line = "collected " if final else "collecting "
+        line += (
+            str(self._numcollected) + " item" + ("" if self._numcollected == 1 else "s")
+        )
+        if errors:
+            line += " / %d error%s" % (errors, "s" if errors != 1 else "")
+        if deselected:
+            line += " / %d deselected" % deselected
+        if skipped:
+            line += " / %d skipped" % skipped
+        if self._numcollected > selected:
+            line += " / %d selected" % selected
+        if self.isatty:
+            self.rewrite(line, bold=True, erase=True)
+            if final:
+                self.write("\n")
+        else:
+            self.write_line(line)
+
+    @hookimpl(trylast=True)
+    def pytest_sessionstart(self, session: Session) -> None:
+        self._session = session
+        self._sessionstarttime = timing.time()
+        if not self.showheader:
+            return
+        self.write_sep("=", "test session starts", bold=True)
+        verinfo = platform.python_version()
+        if not self.no_header:
+            msg = f"platform {sys.platform} -- Python {verinfo}"
+            pypy_version_info = getattr(sys, "pypy_version_info", None)
+            if pypy_version_info:
+                verinfo = ".".join(map(str, pypy_version_info[:3]))
+                msg += f"[pypy-{verinfo}-{pypy_version_info[3]}]"
+            msg += f", pytest-{_pytest._version.version}, pluggy-{pluggy.__version__}"
+            if (
+                self.verbosity > 0
+                or self.config.option.debug
+                or getattr(self.config.option, "pastebin", None)
+            ):
+                msg += " -- " + str(sys.executable)
+            self.write_line(msg)
+            lines = self.config.hook.pytest_report_header(
+                config=self.config, start_path=self.startpath
+            )
+            self._write_report_lines_from_hooks(lines)
+
+    def _write_report_lines_from_hooks(
+        self, lines: Sequence[str | Sequence[str]]
+    ) -> None:
+        for line_or_lines in reversed(lines):
+            if isinstance(line_or_lines, str):
+                self.write_line(line_or_lines)
+            else:
+                for line in line_or_lines:
+                    self.write_line(line)
+
+    def pytest_report_header(self, config: Config) -> list[str]:
+        result = [f"rootdir: {config.rootpath}"]
+
+        if config.inipath:
+            result.append("configfile: " + bestrelpath(config.rootpath, config.inipath))
+
+        if config.args_source == Config.ArgsSource.TESTPATHS:
+            testpaths: list[str] = config.getini("testpaths")
+            result.append("testpaths: {}".format(", ".join(testpaths)))
+
+        plugininfo = config.pluginmanager.list_plugin_distinfo()
+        if plugininfo:
+            result.append(
+                "plugins: {}".format(", ".join(_plugin_nameversions(plugininfo)))
+            )
+        return result
+
+    def pytest_collection_finish(self, session: Session) -> None:
+        self.report_collect(True)
+
+        lines = self.config.hook.pytest_report_collectionfinish(
+            config=self.config,
+            start_path=self.startpath,
+            items=session.items,
+        )
+        self._write_report_lines_from_hooks(lines)
+
+        if self.config.getoption("collectonly"):
+            if session.items:
+                if self.config.option.verbose > -1:
+                    self._tw.line("")
+                self._printcollecteditems(session.items)
+
+            failed = self.stats.get("failed")
+            if failed:
+                self._tw.sep("!", "collection failures")
+                for rep in failed:
+                    rep.toterminal(self._tw)
+
+    def _printcollecteditems(self, items: Sequence[Item]) -> None:
+        test_cases_verbosity = self.config.get_verbosity(Config.VERBOSITY_TEST_CASES)
+        if test_cases_verbosity < 0:
+            if test_cases_verbosity < -1:
+                counts = Counter(item.nodeid.split("::", 1)[0] for item in items)
+                for name, count in sorted(counts.items()):
+                    self._tw.line("%s: %d" % (name, count))
+            else:
+                for item in items:
+                    self._tw.line(item.nodeid)
+            return
+        stack: list[Node] = []
+        indent = ""
+        for item in items:
+            needed_collectors = item.listchain()[1:]  # strip root node
+            while stack:
+                if stack == needed_collectors[: len(stack)]:
+                    break
+                stack.pop()
+            for col in needed_collectors[len(stack) :]:
+                stack.append(col)
+                indent = (len(stack) - 1) * "  "
+                self._tw.line(f"{indent}{col}")
+                if test_cases_verbosity >= 1:
+                    obj = getattr(col, "obj", None)
+                    doc = inspect.getdoc(obj) if obj else None
+                    if doc:
+                        for line in doc.splitlines():
+                            self._tw.line("{}{}".format(indent + "  ", line))
+
+    @hookimpl(wrapper=True)
+    def pytest_sessionfinish(
+        self, session: Session, exitstatus: int | ExitCode
+    ) -> Generator[None]:
+        result = yield
+        self._tw.line("")
+        summary_exit_codes = (
+            ExitCode.OK,
+            ExitCode.TESTS_FAILED,
+            ExitCode.INTERRUPTED,
+            ExitCode.USAGE_ERROR,
+            ExitCode.NO_TESTS_COLLECTED,
+        )
+        if exitstatus in summary_exit_codes and not self.no_summary:
+            self.config.hook.pytest_terminal_summary(
+                terminalreporter=self, exitstatus=exitstatus, config=self.config
+            )
+        if session.shouldfail:
+            self.write_sep("!", str(session.shouldfail), red=True)
+        if exitstatus == ExitCode.INTERRUPTED:
+            self._report_keyboardinterrupt()
+            self._keyboardinterrupt_memo = None
+        elif session.shouldstop:
+            self.write_sep("!", str(session.shouldstop), red=True)
+        self.summary_stats()
+        return result
+
+    @hookimpl(wrapper=True)
+    def pytest_terminal_summary(self) -> Generator[None]:
+        self.summary_errors()
+        self.summary_failures()
+        self.summary_xfailures()
+        self.summary_warnings()
+        self.summary_passes()
+        self.summary_xpasses()
+        try:
+            return (yield)
+        finally:
+            self.short_test_summary()
+            # Display any extra warnings from teardown here (if any).
+            self.summary_warnings()
+
+    def pytest_keyboard_interrupt(self, excinfo: ExceptionInfo[BaseException]) -> None:
+        self._keyboardinterrupt_memo = excinfo.getrepr(funcargs=True)
+
+    def pytest_unconfigure(self) -> None:
+        if self._keyboardinterrupt_memo is not None:
+            self._report_keyboardinterrupt()
+
+    def _report_keyboardinterrupt(self) -> None:
+        excrepr = self._keyboardinterrupt_memo
+        assert excrepr is not None
+        assert excrepr.reprcrash is not None
+        msg = excrepr.reprcrash.message
+        self.write_sep("!", msg)
+        if "KeyboardInterrupt" in msg:
+            if self.config.option.fulltrace:
+                excrepr.toterminal(self._tw)
+            else:
+                excrepr.reprcrash.toterminal(self._tw)
+                self._tw.line(
+                    "(to show a full traceback on KeyboardInterrupt use --full-trace)",
+                    yellow=True,
+                )
+
+    def _locationline(
+        self, nodeid: str, fspath: str, lineno: int | None, domain: str
+    ) -> str:
+        def mkrel(nodeid: str) -> str:
+            line = self.config.cwd_relative_nodeid(nodeid)
+            if domain and line.endswith(domain):
+                line = line[: -len(domain)]
+                values = domain.split("[")
+                values[0] = values[0].replace(".", "::")  # don't replace '.' in params
+                line += "[".join(values)
+            return line
+
+        # fspath comes from testid which has a "/"-normalized path.
+        if fspath:
+            res = mkrel(nodeid)
+            if self.verbosity >= 2 and nodeid.split("::")[0] != fspath.replace(
+                "\\", nodes.SEP
+            ):
+                res += " <- " + bestrelpath(self.startpath, Path(fspath))
+        else:
+            res = "[location]"
+        return res + " "
+
+    def _getfailureheadline(self, rep):
+        head_line = rep.head_line
+        if head_line:
+            return head_line
+        return "test session"  # XXX?
+
+    def _getcrashline(self, rep):
+        try:
+            return str(rep.longrepr.reprcrash)
+        except AttributeError:
+            try:
+                return str(rep.longrepr)[:50]
+            except AttributeError:
+                return ""
+
+    #
+    # Summaries for sessionfinish.
+    #
+    def getreports(self, name: str):
+        return [x for x in self.stats.get(name, ()) if not hasattr(x, "_pdbshown")]
+
+    def summary_warnings(self) -> None:
+        if self.hasopt("w"):
+            all_warnings: list[WarningReport] | None = self.stats.get("warnings")
+            if not all_warnings:
+                return
+
+            final = self._already_displayed_warnings is not None
+            if final:
+                warning_reports = all_warnings[self._already_displayed_warnings :]
+            else:
+                warning_reports = all_warnings
+            self._already_displayed_warnings = len(warning_reports)
+            if not warning_reports:
+                return
+
+            reports_grouped_by_message: dict[str, list[WarningReport]] = {}
+            for wr in warning_reports:
+                reports_grouped_by_message.setdefault(wr.message, []).append(wr)
+
+            def collapsed_location_report(reports: list[WarningReport]) -> str:
+                locations = []
+                for w in reports:
+                    location = w.get_location(self.config)
+                    if location:
+                        locations.append(location)
+
+                if len(locations) < 10:
+                    return "\n".join(map(str, locations))
+
+                counts_by_filename = Counter(
+                    str(loc).split("::", 1)[0] for loc in locations
+                )
+                return "\n".join(
+                    "{}: {} warning{}".format(k, v, "s" if v > 1 else "")
+                    for k, v in counts_by_filename.items()
+                )
+
+            title = "warnings summary (final)" if final else "warnings summary"
+            self.write_sep("=", title, yellow=True, bold=False)
+            for message, message_reports in reports_grouped_by_message.items():
+                maybe_location = collapsed_location_report(message_reports)
+                if maybe_location:
+                    self._tw.line(maybe_location)
+                    lines = message.splitlines()
+                    indented = "\n".join("  " + x for x in lines)
+                    message = indented.rstrip()
+                else:
+                    message = message.rstrip()
+                self._tw.line(message)
+                self._tw.line()
+            self._tw.line(
+                "-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html"
+            )
+
+    def summary_passes(self) -> None:
+        self.summary_passes_combined("passed", "PASSES", "P")
+
+    def summary_xpasses(self) -> None:
+        self.summary_passes_combined("xpassed", "XPASSES", "X")
+
+    def summary_passes_combined(
+        self, which_reports: str, sep_title: str, needed_opt: str
+    ) -> None:
+        if self.config.option.tbstyle != "no":
+            if self.hasopt(needed_opt):
+                reports: list[TestReport] = self.getreports(which_reports)
+                if not reports:
+                    return
+                self.write_sep("=", sep_title)
+                for rep in reports:
+                    if rep.sections:
+                        msg = self._getfailureheadline(rep)
+                        self.write_sep("_", msg, green=True, bold=True)
+                        self._outrep_summary(rep)
+                    self._handle_teardown_sections(rep.nodeid)
+
+    def _get_teardown_reports(self, nodeid: str) -> list[TestReport]:
+        reports = self.getreports("")
+        return [
+            report
+            for report in reports
+            if report.when == "teardown" and report.nodeid == nodeid
+        ]
+
+    def _handle_teardown_sections(self, nodeid: str) -> None:
+        for report in self._get_teardown_reports(nodeid):
+            self.print_teardown_sections(report)
+
+    def print_teardown_sections(self, rep: TestReport) -> None:
+        showcapture = self.config.option.showcapture
+        if showcapture == "no":
+            return
+        for secname, content in rep.sections:
+            if showcapture != "all" and showcapture not in secname:
+                continue
+            if "teardown" in secname:
+                self._tw.sep("-", secname)
+                if content[-1:] == "\n":
+                    content = content[:-1]
+                self._tw.line(content)
+
+    def summary_failures(self) -> None:
+        style = self.config.option.tbstyle
+        self.summary_failures_combined("failed", "FAILURES", style=style)
+
+    def summary_xfailures(self) -> None:
+        show_tb = self.config.option.xfail_tb
+        style = self.config.option.tbstyle if show_tb else "no"
+        self.summary_failures_combined("xfailed", "XFAILURES", style=style)
+
+    def summary_failures_combined(
+        self,
+        which_reports: str,
+        sep_title: str,
+        *,
+        style: str,
+        needed_opt: str | None = None,
+    ) -> None:
+        if style != "no":
+            if not needed_opt or self.hasopt(needed_opt):
+                reports: list[BaseReport] = self.getreports(which_reports)
+                if not reports:
+                    return
+                self.write_sep("=", sep_title)
+                if style == "line":
+                    for rep in reports:
+                        line = self._getcrashline(rep)
+                        self.write_line(line)
+                else:
+                    for rep in reports:
+                        msg = self._getfailureheadline(rep)
+                        self.write_sep("_", msg, red=True, bold=True)
+                        self._outrep_summary(rep)
+                        self._handle_teardown_sections(rep.nodeid)
+
+    def summary_errors(self) -> None:
+        if self.config.option.tbstyle != "no":
+            reports: list[BaseReport] = self.getreports("error")
+            if not reports:
+                return
+            self.write_sep("=", "ERRORS")
+            for rep in self.stats["error"]:
+                msg = self._getfailureheadline(rep)
+                if rep.when == "collect":
+                    msg = "ERROR collecting " + msg
+                else:
+                    msg = f"ERROR at {rep.when} of {msg}"
+                self.write_sep("_", msg, red=True, bold=True)
+                self._outrep_summary(rep)
+
+    def _outrep_summary(self, rep: BaseReport) -> None:
+        rep.toterminal(self._tw)
+        showcapture = self.config.option.showcapture
+        if showcapture == "no":
+            return
+        for secname, content in rep.sections:
+            if showcapture != "all" and showcapture not in secname:
+                continue
+            self._tw.sep("-", secname)
+            if content[-1:] == "\n":
+                content = content[:-1]
+            self._tw.line(content)
+
+    def summary_stats(self) -> None:
+        if self.verbosity < -1:
+            return
+
+        session_duration = timing.time() - self._sessionstarttime
+        (parts, main_color) = self.build_summary_stats_line()
+        line_parts = []
+
+        display_sep = self.verbosity >= 0
+        if display_sep:
+            fullwidth = self._tw.fullwidth
+        for text, markup in parts:
+            with_markup = self._tw.markup(text, **markup)
+            if display_sep:
+                fullwidth += len(with_markup) - len(text)
+            line_parts.append(with_markup)
+        msg = ", ".join(line_parts)
+
+        main_markup = {main_color: True}
+        duration = f" in {format_session_duration(session_duration)}"
+        duration_with_markup = self._tw.markup(duration, **main_markup)
+        if display_sep:
+            fullwidth += len(duration_with_markup) - len(duration)
+        msg += duration_with_markup
+
+        if display_sep:
+            markup_for_end_sep = self._tw.markup("", **main_markup)
+            if markup_for_end_sep.endswith("\x1b[0m"):
+                markup_for_end_sep = markup_for_end_sep[:-4]
+            fullwidth += len(markup_for_end_sep)
+            msg += markup_for_end_sep
+
+        if display_sep:
+            self.write_sep("=", msg, fullwidth=fullwidth, **main_markup)
+        else:
+            self.write_line(msg, **main_markup)
+
+    def short_test_summary(self) -> None:
+        if not self.reportchars:
+            return
+
+        def show_simple(lines: list[str], *, stat: str) -> None:
+            failed = self.stats.get(stat, [])
+            if not failed:
+                return
+            config = self.config
+            for rep in failed:
+                color = _color_for_type.get(stat, _color_for_type_default)
+                line = _get_line_with_reprcrash_message(
+                    config, rep, self._tw, {color: True}
+                )
+                lines.append(line)
+
+        def show_xfailed(lines: list[str]) -> None:
+            xfailed = self.stats.get("xfailed", [])
+            for rep in xfailed:
+                verbose_word, verbose_markup = rep._get_verbose_word_with_markup(
+                    self.config, {_color_for_type["warnings"]: True}
+                )
+                markup_word = self._tw.markup(verbose_word, **verbose_markup)
+                nodeid = _get_node_id_with_markup(self._tw, self.config, rep)
+                line = f"{markup_word} {nodeid}"
+                reason = rep.wasxfail
+                if reason:
+                    line += " - " + str(reason)
+
+                lines.append(line)
+
+        def show_xpassed(lines: list[str]) -> None:
+            xpassed = self.stats.get("xpassed", [])
+            for rep in xpassed:
+                verbose_word, verbose_markup = rep._get_verbose_word_with_markup(
+                    self.config, {_color_for_type["warnings"]: True}
+                )
+                markup_word = self._tw.markup(verbose_word, **verbose_markup)
+                nodeid = _get_node_id_with_markup(self._tw, self.config, rep)
+                line = f"{markup_word} {nodeid}"
+                reason = rep.wasxfail
+                if reason:
+                    line += " - " + str(reason)
+                lines.append(line)
+
+        def show_skipped_folded(lines: list[str]) -> None:
+            skipped: list[CollectReport] = self.stats.get("skipped", [])
+            fskips = _folded_skips(self.startpath, skipped) if skipped else []
+            if not fskips:
+                return
+            verbose_word, verbose_markup = skipped[0]._get_verbose_word_with_markup(
+                self.config, {_color_for_type["warnings"]: True}
+            )
+            markup_word = self._tw.markup(verbose_word, **verbose_markup)
+            prefix = "Skipped: "
+            for num, fspath, lineno, reason in fskips:
+                if reason.startswith(prefix):
+                    reason = reason[len(prefix) :]
+                if lineno is not None:
+                    lines.append(
+                        "%s [%d] %s:%d: %s" % (markup_word, num, fspath, lineno, reason)
+                    )
+                else:
+                    lines.append("%s [%d] %s: %s" % (markup_word, num, fspath, reason))
+
+        def show_skipped_unfolded(lines: list[str]) -> None:
+            skipped: list[CollectReport] = self.stats.get("skipped", [])
+
+            for rep in skipped:
+                assert rep.longrepr is not None
+                assert isinstance(rep.longrepr, tuple), (rep, rep.longrepr)
+                assert len(rep.longrepr) == 3, (rep, rep.longrepr)
+
+                verbose_word, verbose_markup = rep._get_verbose_word_with_markup(
+                    self.config, {_color_for_type["warnings"]: True}
+                )
+                markup_word = self._tw.markup(verbose_word, **verbose_markup)
+                nodeid = _get_node_id_with_markup(self._tw, self.config, rep)
+                line = f"{markup_word} {nodeid}"
+                reason = rep.longrepr[2]
+                if reason:
+                    line += " - " + str(reason)
+                lines.append(line)
+
+        def show_skipped(lines: list[str]) -> None:
+            if self.foldskipped:
+                show_skipped_folded(lines)
+            else:
+                show_skipped_unfolded(lines)
+
+        REPORTCHAR_ACTIONS: Mapping[str, Callable[[list[str]], None]] = {
+            "x": show_xfailed,
+            "X": show_xpassed,
+            "f": partial(show_simple, stat="failed"),
+            "s": show_skipped,
+            "p": partial(show_simple, stat="passed"),
+            "E": partial(show_simple, stat="error"),
+        }
+
+        lines: list[str] = []
+        for char in self.reportchars:
+            action = REPORTCHAR_ACTIONS.get(char)
+            if action:  # skipping e.g. "P" (passed with output) here.
+                action(lines)
+
+        if lines:
+            self.write_sep("=", "short test summary info", cyan=True, bold=True)
+            for line in lines:
+                self.write_line(line)
+
+    def _get_main_color(self) -> tuple[str, list[str]]:
+        if self._main_color is None or self._known_types is None or self._is_last_item:
+            self._set_main_color()
+            assert self._main_color
+            assert self._known_types
+        return self._main_color, self._known_types
+
+    def _determine_main_color(self, unknown_type_seen: bool) -> str:
+        stats = self.stats
+        if "failed" in stats or "error" in stats:
+            main_color = "red"
+        elif "warnings" in stats or "xpassed" in stats or unknown_type_seen:
+            main_color = "yellow"
+        elif "passed" in stats or not self._is_last_item:
+            main_color = "green"
+        else:
+            main_color = "yellow"
+        return main_color
+
+    def _set_main_color(self) -> None:
+        unknown_types: list[str] = []
+        for found_type in self.stats:
+            if found_type:  # setup/teardown reports have an empty key, ignore them
+                if found_type not in KNOWN_TYPES and found_type not in unknown_types:
+                    unknown_types.append(found_type)
+        self._known_types = list(KNOWN_TYPES) + unknown_types
+        self._main_color = self._determine_main_color(bool(unknown_types))
+
+    def build_summary_stats_line(self) -> tuple[list[tuple[str, dict[str, bool]]], str]:
+        """
+        Build the parts used in the last summary stats line.
+
+        The summary stats line is the line shown at the end, "=== 12 passed, 2 errors in Xs===".
+
+        This function builds a list of the "parts" that make up for the text in that line, in
+        the example above it would be:
+
+            [
+                ("12 passed", {"green": True}),
+                ("2 errors", {"red": True}
+            ]
+
+        That last dict for each line is a "markup dictionary", used by TerminalWriter to
+        color output.
+
+        The final color of the line is also determined by this function, and is the second
+        element of the returned tuple.
+        """
+        if self.config.getoption("collectonly"):
+            return self._build_collect_only_summary_stats_line()
+        else:
+            return self._build_normal_summary_stats_line()
+
+    def _get_reports_to_display(self, key: str) -> list[Any]:
+        """Get test/collection reports for the given status key, such as `passed` or `error`."""
+        reports = self.stats.get(key, [])
+        return [x for x in reports if getattr(x, "count_towards_summary", True)]
+
+    def _build_normal_summary_stats_line(
+        self,
+    ) -> tuple[list[tuple[str, dict[str, bool]]], str]:
+        main_color, known_types = self._get_main_color()
+        parts = []
+
+        for key in known_types:
+            reports = self._get_reports_to_display(key)
+            if reports:
+                count = len(reports)
+                color = _color_for_type.get(key, _color_for_type_default)
+                markup = {color: True, "bold": color == main_color}
+                parts.append(("%d %s" % pluralize(count, key), markup))
+
+        if not parts:
+            parts = [("no tests ran", {_color_for_type_default: True})]
+
+        return parts, main_color
+
+    def _build_collect_only_summary_stats_line(
+        self,
+    ) -> tuple[list[tuple[str, dict[str, bool]]], str]:
+        deselected = len(self._get_reports_to_display("deselected"))
+        errors = len(self._get_reports_to_display("error"))
+
+        if self._numcollected == 0:
+            parts = [("no tests collected", {"yellow": True})]
+            main_color = "yellow"
+
+        elif deselected == 0:
+            main_color = "green"
+            collected_output = "%d %s collected" % pluralize(self._numcollected, "test")
+            parts = [(collected_output, {main_color: True})]
+        else:
+            all_tests_were_deselected = self._numcollected == deselected
+            if all_tests_were_deselected:
+                main_color = "yellow"
+                collected_output = f"no tests collected ({deselected} deselected)"
+            else:
+                main_color = "green"
+                selected = self._numcollected - deselected
+                collected_output = f"{selected}/{self._numcollected} tests collected ({deselected} deselected)"
+
+            parts = [(collected_output, {main_color: True})]
+
+        if errors:
+            main_color = _color_for_type["error"]
+            parts += [("%d %s" % pluralize(errors, "error"), {main_color: True})]
+
+        return parts, main_color
+
+
+def _get_node_id_with_markup(tw: TerminalWriter, config: Config, rep: BaseReport):
+    nodeid = config.cwd_relative_nodeid(rep.nodeid)
+    path, *parts = nodeid.split("::")
+    if parts:
+        parts_markup = tw.markup("::".join(parts), bold=True)
+        return path + "::" + parts_markup
+    else:
+        return path
+
+
+def _format_trimmed(format: str, msg: str, available_width: int) -> str | None:
+    """Format msg into format, ellipsizing it if doesn't fit in available_width.
+
+    Returns None if even the ellipsis can't fit.
+    """
+    # Only use the first line.
+    i = msg.find("\n")
+    if i != -1:
+        msg = msg[:i]
+
+    ellipsis = "..."
+    format_width = wcswidth(format.format(""))
+    if format_width + len(ellipsis) > available_width:
+        return None
+
+    if format_width + wcswidth(msg) > available_width:
+        available_width -= len(ellipsis)
+        msg = msg[:available_width]
+        while format_width + wcswidth(msg) > available_width:
+            msg = msg[:-1]
+        msg += ellipsis
+
+    return format.format(msg)
+
+
+def _get_line_with_reprcrash_message(
+    config: Config, rep: BaseReport, tw: TerminalWriter, word_markup: dict[str, bool]
+) -> str:
+    """Get summary line for a report, trying to add reprcrash message."""
+    verbose_word, verbose_markup = rep._get_verbose_word_with_markup(
+        config, word_markup
+    )
+    word = tw.markup(verbose_word, **verbose_markup)
+    node = _get_node_id_with_markup(tw, config, rep)
+
+    line = f"{word} {node}"
+    line_width = wcswidth(line)
+
+    try:
+        # Type ignored intentionally -- possible AttributeError expected.
+        msg = rep.longrepr.reprcrash.message  # type: ignore[union-attr]
+    except AttributeError:
+        pass
+    else:
+        if running_on_ci() or config.option.verbose >= 2:
+            msg = f" - {msg}"
+        else:
+            available_width = tw.fullwidth - line_width
+            msg = _format_trimmed(" - {}", msg, available_width)
+        if msg is not None:
+            line += msg
+
+    return line
+
+
+def _folded_skips(
+    startpath: Path,
+    skipped: Sequence[CollectReport],
+) -> list[tuple[int, str, int | None, str]]:
+    d: dict[tuple[str, int | None, str], list[CollectReport]] = {}
+    for event in skipped:
+        assert event.longrepr is not None
+        assert isinstance(event.longrepr, tuple), (event, event.longrepr)
+        assert len(event.longrepr) == 3, (event, event.longrepr)
+        fspath, lineno, reason = event.longrepr
+        # For consistency, report all fspaths in relative form.
+        fspath = bestrelpath(startpath, Path(fspath))
+        keywords = getattr(event, "keywords", {})
+        # Folding reports with global pytestmark variable.
+        # This is a workaround, because for now we cannot identify the scope of a skip marker
+        # TODO: Revisit after marks scope would be fixed.
+        if (
+            event.when == "setup"
+            and "skip" in keywords
+            and "pytestmark" not in keywords
+        ):
+            key: tuple[str, int | None, str] = (fspath, None, reason)
+        else:
+            key = (fspath, lineno, reason)
+        d.setdefault(key, []).append(event)
+    values: list[tuple[int, str, int | None, str]] = []
+    for key, events in d.items():
+        values.append((len(events), *key))
+    return values
+
+
+_color_for_type = {
+    "failed": "red",
+    "error": "red",
+    "warnings": "yellow",
+    "passed": "green",
+}
+_color_for_type_default = "yellow"
+
+
+def pluralize(count: int, noun: str) -> tuple[int, str]:
+    # No need to pluralize words such as `failed` or `passed`.
+    if noun not in ["error", "warnings", "test"]:
+        return count, noun
+
+    # The `warnings` key is plural. To avoid API breakage, we keep it that way but
+    # set it to singular here so we can determine plurality in the same way as we do
+    # for `error`.
+    noun = noun.replace("warnings", "warning")
+
+    return count, noun + "s" if count != 1 else noun
+
+
+def _plugin_nameversions(plugininfo) -> list[str]:
+    values: list[str] = []
+    for plugin, dist in plugininfo:
+        # Gets us name and version!
+        name = f"{dist.project_name}-{dist.version}"
+        # Questionable convenience, but it keeps things short.
+        if name.startswith("pytest-"):
+            name = name[7:]
+        # We decided to print python package names they can have more than one plugin.
+        if name not in values:
+            values.append(name)
+    return values
+
+
+def format_session_duration(seconds: float) -> str:
+    """Format the given seconds in a human readable manner to show in the final summary."""
+    if seconds < 60:
+        return f"{seconds:.2f}s"
+    else:
+        dt = datetime.timedelta(seconds=int(seconds))
+        return f"{seconds:.2f}s ({dt})"
+
+
+def _get_raw_skip_reason(report: TestReport) -> str:
+    """Get the reason string of a skip/xfail/xpass test report.
+
+    The string is just the part given by the user.
+    """
+    if hasattr(report, "wasxfail"):
+        reason = report.wasxfail
+        if reason.startswith("reason: "):
+            reason = reason[len("reason: ") :]
+        return reason
+    else:
+        assert report.skipped
+        assert isinstance(report.longrepr, tuple)
+        _, _, reason = report.longrepr
+        if reason.startswith("Skipped: "):
+            reason = reason[len("Skipped: ") :]
+        elif reason == "Skipped":
+            reason = ""
+        return reason
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/threadexception.py b/backend/venv/lib/python3.12/site-packages/_pytest/threadexception.py
new file mode 100644
index 00000000..c1ed8038
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/threadexception.py
@@ -0,0 +1,97 @@
+from __future__ import annotations
+
+import threading
+import traceback
+from types import TracebackType
+from typing import Any
+from typing import Callable
+from typing import Generator
+from typing import TYPE_CHECKING
+import warnings
+
+import pytest
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+
+# Copied from cpython/Lib/test/support/threading_helper.py, with modifications.
+class catch_threading_exception:
+    """Context manager catching threading.Thread exception using
+    threading.excepthook.
+
+    Storing exc_value using a custom hook can create a reference cycle. The
+    reference cycle is broken explicitly when the context manager exits.
+
+    Storing thread using a custom hook can resurrect it if it is set to an
+    object which is being finalized. Exiting the context manager clears the
+    stored object.
+
+    Usage:
+        with threading_helper.catch_threading_exception() as cm:
+            # code spawning a thread which raises an exception
+            ...
+            # check the thread exception: use cm.args
+            ...
+        # cm.args attribute no longer exists at this point
+        # (to break a reference cycle)
+    """
+
+    def __init__(self) -> None:
+        self.args: threading.ExceptHookArgs | None = None
+        self._old_hook: Callable[[threading.ExceptHookArgs], Any] | None = None
+
+    def _hook(self, args: threading.ExceptHookArgs) -> None:
+        self.args = args
+
+    def __enter__(self) -> Self:
+        self._old_hook = threading.excepthook
+        threading.excepthook = self._hook
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        assert self._old_hook is not None
+        threading.excepthook = self._old_hook
+        self._old_hook = None
+        del self.args
+
+
+def thread_exception_runtest_hook() -> Generator[None]:
+    with catch_threading_exception() as cm:
+        try:
+            yield
+        finally:
+            if cm.args:
+                thread_name = (
+                    "<unknown>" if cm.args.thread is None else cm.args.thread.name
+                )
+                msg = f"Exception in thread {thread_name}\n\n"
+                msg += "".join(
+                    traceback.format_exception(
+                        cm.args.exc_type,
+                        cm.args.exc_value,
+                        cm.args.exc_traceback,
+                    )
+                )
+                warnings.warn(pytest.PytestUnhandledThreadExceptionWarning(msg))
+
+
+@pytest.hookimpl(wrapper=True, trylast=True)
+def pytest_runtest_setup() -> Generator[None]:
+    yield from thread_exception_runtest_hook()
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_call() -> Generator[None]:
+    yield from thread_exception_runtest_hook()
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_teardown() -> Generator[None]:
+    yield from thread_exception_runtest_hook()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/timing.py b/backend/venv/lib/python3.12/site-packages/_pytest/timing.py
new file mode 100644
index 00000000..b23c7f69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/timing.py
@@ -0,0 +1,16 @@
+"""Indirection for time functions.
+
+We intentionally grab some "time" functions internally to avoid tests mocking "time" to affect
+pytest runtime information (issue #185).
+
+Fixture "mock_timing" also interacts with this module for pytest's own tests.
+"""
+
+from __future__ import annotations
+
+from time import perf_counter
+from time import sleep
+from time import time
+
+
+__all__ = ["perf_counter", "sleep", "time"]
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/tmpdir.py b/backend/venv/lib/python3.12/site-packages/_pytest/tmpdir.py
new file mode 100644
index 00000000..1731a4b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/tmpdir.py
@@ -0,0 +1,314 @@
+# mypy: allow-untyped-defs
+"""Support for providing temporary directories to test functions."""
+
+from __future__ import annotations
+
+import dataclasses
+import os
+from pathlib import Path
+import re
+from shutil import rmtree
+import tempfile
+from typing import Any
+from typing import Dict
+from typing import final
+from typing import Generator
+from typing import Literal
+
+from .pathlib import cleanup_dead_symlinks
+from .pathlib import LOCK_TIMEOUT
+from .pathlib import make_numbered_dir
+from .pathlib import make_numbered_dir_with_cleanup
+from .pathlib import rm_rf
+from _pytest.compat import get_user_id
+from _pytest.config import Config
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config.argparsing import Parser
+from _pytest.deprecated import check_ispytest
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureRequest
+from _pytest.monkeypatch import MonkeyPatch
+from _pytest.nodes import Item
+from _pytest.reports import TestReport
+from _pytest.stash import StashKey
+
+
+tmppath_result_key = StashKey[Dict[str, bool]]()
+RetentionType = Literal["all", "failed", "none"]
+
+
+@final
+@dataclasses.dataclass
+class TempPathFactory:
+    """Factory for temporary directories under the common base temp directory,
+    as discussed at :ref:`temporary directory location and retention`.
+    """
+
+    _given_basetemp: Path | None
+    # pluggy TagTracerSub, not currently exposed, so Any.
+    _trace: Any
+    _basetemp: Path | None
+    _retention_count: int
+    _retention_policy: RetentionType
+
+    def __init__(
+        self,
+        given_basetemp: Path | None,
+        retention_count: int,
+        retention_policy: RetentionType,
+        trace,
+        basetemp: Path | None = None,
+        *,
+        _ispytest: bool = False,
+    ) -> None:
+        check_ispytest(_ispytest)
+        if given_basetemp is None:
+            self._given_basetemp = None
+        else:
+            # Use os.path.abspath() to get absolute path instead of resolve() as it
+            # does not work the same in all platforms (see #4427).
+            # Path.absolute() exists, but it is not public (see https://bugs.python.org/issue25012).
+            self._given_basetemp = Path(os.path.abspath(str(given_basetemp)))
+        self._trace = trace
+        self._retention_count = retention_count
+        self._retention_policy = retention_policy
+        self._basetemp = basetemp
+
+    @classmethod
+    def from_config(
+        cls,
+        config: Config,
+        *,
+        _ispytest: bool = False,
+    ) -> TempPathFactory:
+        """Create a factory according to pytest configuration.
+
+        :meta private:
+        """
+        check_ispytest(_ispytest)
+        count = int(config.getini("tmp_path_retention_count"))
+        if count < 0:
+            raise ValueError(
+                f"tmp_path_retention_count must be >= 0. Current input: {count}."
+            )
+
+        policy = config.getini("tmp_path_retention_policy")
+        if policy not in ("all", "failed", "none"):
+            raise ValueError(
+                f"tmp_path_retention_policy must be either all, failed, none. Current input: {policy}."
+            )
+
+        return cls(
+            given_basetemp=config.option.basetemp,
+            trace=config.trace.get("tmpdir"),
+            retention_count=count,
+            retention_policy=policy,
+            _ispytest=True,
+        )
+
+    def _ensure_relative_to_basetemp(self, basename: str) -> str:
+        basename = os.path.normpath(basename)
+        if (self.getbasetemp() / basename).resolve().parent != self.getbasetemp():
+            raise ValueError(f"{basename} is not a normalized and relative path")
+        return basename
+
+    def mktemp(self, basename: str, numbered: bool = True) -> Path:
+        """Create a new temporary directory managed by the factory.
+
+        :param basename:
+            Directory base name, must be a relative path.
+
+        :param numbered:
+            If ``True``, ensure the directory is unique by adding a numbered
+            suffix greater than any existing one: ``basename="foo-"`` and ``numbered=True``
+            means that this function will create directories named ``"foo-0"``,
+            ``"foo-1"``, ``"foo-2"`` and so on.
+
+        :returns:
+            The path to the new directory.
+        """
+        basename = self._ensure_relative_to_basetemp(basename)
+        if not numbered:
+            p = self.getbasetemp().joinpath(basename)
+            p.mkdir(mode=0o700)
+        else:
+            p = make_numbered_dir(root=self.getbasetemp(), prefix=basename, mode=0o700)
+            self._trace("mktemp", p)
+        return p
+
+    def getbasetemp(self) -> Path:
+        """Return the base temporary directory, creating it if needed.
+
+        :returns:
+            The base temporary directory.
+        """
+        if self._basetemp is not None:
+            return self._basetemp
+
+        if self._given_basetemp is not None:
+            basetemp = self._given_basetemp
+            if basetemp.exists():
+                rm_rf(basetemp)
+            basetemp.mkdir(mode=0o700)
+            basetemp = basetemp.resolve()
+        else:
+            from_env = os.environ.get("PYTEST_DEBUG_TEMPROOT")
+            temproot = Path(from_env or tempfile.gettempdir()).resolve()
+            user = get_user() or "unknown"
+            # use a sub-directory in the temproot to speed-up
+            # make_numbered_dir() call
+            rootdir = temproot.joinpath(f"pytest-of-{user}")
+            try:
+                rootdir.mkdir(mode=0o700, exist_ok=True)
+            except OSError:
+                # getuser() likely returned illegal characters for the platform, use unknown back off mechanism
+                rootdir = temproot.joinpath("pytest-of-unknown")
+                rootdir.mkdir(mode=0o700, exist_ok=True)
+            # Because we use exist_ok=True with a predictable name, make sure
+            # we are the owners, to prevent any funny business (on unix, where
+            # temproot is usually shared).
+            # Also, to keep things private, fixup any world-readable temp
+            # rootdir's permissions. Historically 0o755 was used, so we can't
+            # just error out on this, at least for a while.
+            uid = get_user_id()
+            if uid is not None:
+                rootdir_stat = rootdir.stat()
+                if rootdir_stat.st_uid != uid:
+                    raise OSError(
+                        f"The temporary directory {rootdir} is not owned by the current user. "
+                        "Fix this and try again."
+                    )
+                if (rootdir_stat.st_mode & 0o077) != 0:
+                    os.chmod(rootdir, rootdir_stat.st_mode & ~0o077)
+            keep = self._retention_count
+            if self._retention_policy == "none":
+                keep = 0
+            basetemp = make_numbered_dir_with_cleanup(
+                prefix="pytest-",
+                root=rootdir,
+                keep=keep,
+                lock_timeout=LOCK_TIMEOUT,
+                mode=0o700,
+            )
+        assert basetemp is not None, basetemp
+        self._basetemp = basetemp
+        self._trace("new basetemp", basetemp)
+        return basetemp
+
+
+def get_user() -> str | None:
+    """Return the current user name, or None if getuser() does not work
+    in the current environment (see #1010)."""
+    try:
+        # In some exotic environments, getpass may not be importable.
+        import getpass
+
+        return getpass.getuser()
+    except (ImportError, OSError, KeyError):
+        return None
+
+
+def pytest_configure(config: Config) -> None:
+    """Create a TempPathFactory and attach it to the config object.
+
+    This is to comply with existing plugins which expect the handler to be
+    available at pytest_configure time, but ideally should be moved entirely
+    to the tmp_path_factory session fixture.
+    """
+    mp = MonkeyPatch()
+    config.add_cleanup(mp.undo)
+    _tmp_path_factory = TempPathFactory.from_config(config, _ispytest=True)
+    mp.setattr(config, "_tmp_path_factory", _tmp_path_factory, raising=False)
+
+
+def pytest_addoption(parser: Parser) -> None:
+    parser.addini(
+        "tmp_path_retention_count",
+        help="How many sessions should we keep the `tmp_path` directories, according to `tmp_path_retention_policy`.",
+        default=3,
+    )
+
+    parser.addini(
+        "tmp_path_retention_policy",
+        help="Controls which directories created by the `tmp_path` fixture are kept around, based on test outcome. "
+        "(all/failed/none)",
+        default="all",
+    )
+
+
+@fixture(scope="session")
+def tmp_path_factory(request: FixtureRequest) -> TempPathFactory:
+    """Return a :class:`pytest.TempPathFactory` instance for the test session."""
+    # Set dynamically by pytest_configure() above.
+    return request.config._tmp_path_factory  # type: ignore
+
+
+def _mk_tmp(request: FixtureRequest, factory: TempPathFactory) -> Path:
+    name = request.node.name
+    name = re.sub(r"[\W]", "_", name)
+    MAXVAL = 30
+    name = name[:MAXVAL]
+    return factory.mktemp(name, numbered=True)
+
+
+@fixture
+def tmp_path(
+    request: FixtureRequest, tmp_path_factory: TempPathFactory
+) -> Generator[Path]:
+    """Return a temporary directory (as :class:`pathlib.Path` object)
+    which is unique to each test function invocation.
+    The temporary directory is created as a subdirectory
+    of the base temporary directory, with configurable retention,
+    as discussed in :ref:`temporary directory location and retention`.
+    """
+    path = _mk_tmp(request, tmp_path_factory)
+    yield path
+
+    # Remove the tmpdir if the policy is "failed" and the test passed.
+    tmp_path_factory: TempPathFactory = request.session.config._tmp_path_factory  # type: ignore
+    policy = tmp_path_factory._retention_policy
+    result_dict = request.node.stash[tmppath_result_key]
+
+    if policy == "failed" and result_dict.get("call", True):
+        # We do a "best effort" to remove files, but it might not be possible due to some leaked resource,
+        # permissions, etc, in which case we ignore it.
+        rmtree(path, ignore_errors=True)
+
+    del request.node.stash[tmppath_result_key]
+
+
+def pytest_sessionfinish(session, exitstatus: int | ExitCode):
+    """After each session, remove base directory if all the tests passed,
+    the policy is "failed", and the basetemp is not specified by a user.
+    """
+    tmp_path_factory: TempPathFactory = session.config._tmp_path_factory
+    basetemp = tmp_path_factory._basetemp
+    if basetemp is None:
+        return
+
+    policy = tmp_path_factory._retention_policy
+    if (
+        exitstatus == 0
+        and policy == "failed"
+        and tmp_path_factory._given_basetemp is None
+    ):
+        if basetemp.is_dir():
+            # We do a "best effort" to remove files, but it might not be possible due to some leaked resource,
+            # permissions, etc, in which case we ignore it.
+            rmtree(basetemp, ignore_errors=True)
+
+    # Remove dead symlinks.
+    if basetemp.is_dir():
+        cleanup_dead_symlinks(basetemp)
+
+
+@hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_makereport(
+    item: Item, call
+) -> Generator[None, TestReport, TestReport]:
+    rep = yield
+    assert rep.when is not None
+    empty: dict[str, bool] = {}
+    item.stash.setdefault(tmppath_result_key, empty)[rep.when] = rep.passed
+    return rep
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/unittest.py b/backend/venv/lib/python3.12/site-packages/_pytest/unittest.py
new file mode 100644
index 00000000..8cecd4f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/unittest.py
@@ -0,0 +1,435 @@
+# mypy: allow-untyped-defs
+"""Discover and run std-library "unittest" style tests."""
+
+from __future__ import annotations
+
+import inspect
+import sys
+import traceback
+import types
+from typing import Any
+from typing import Callable
+from typing import Generator
+from typing import Iterable
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+
+import _pytest._code
+from _pytest.compat import is_async_function
+from _pytest.config import hookimpl
+from _pytest.fixtures import FixtureRequest
+from _pytest.nodes import Collector
+from _pytest.nodes import Item
+from _pytest.outcomes import exit
+from _pytest.outcomes import fail
+from _pytest.outcomes import skip
+from _pytest.outcomes import xfail
+from _pytest.python import Class
+from _pytest.python import Function
+from _pytest.python import Module
+from _pytest.runner import CallInfo
+import pytest
+
+
+if sys.version_info[:2] < (3, 11):
+    from exceptiongroup import ExceptionGroup
+
+if TYPE_CHECKING:
+    import unittest
+
+    import twisted.trial.unittest
+
+
+_SysExcInfoType = Union[
+    Tuple[Type[BaseException], BaseException, types.TracebackType],
+    Tuple[None, None, None],
+]
+
+
+def pytest_pycollect_makeitem(
+    collector: Module | Class, name: str, obj: object
+) -> UnitTestCase | None:
+    try:
+        # Has unittest been imported?
+        ut = sys.modules["unittest"]
+        # Is obj a subclass of unittest.TestCase?
+        # Type ignored because `ut` is an opaque module.
+        if not issubclass(obj, ut.TestCase):  # type: ignore
+            return None
+    except Exception:
+        return None
+    # Is obj a concrete class?
+    # Abstract classes can't be instantiated so no point collecting them.
+    if inspect.isabstract(obj):
+        return None
+    # Yes, so let's collect it.
+    return UnitTestCase.from_parent(collector, name=name, obj=obj)
+
+
+class UnitTestCase(Class):
+    # Marker for fixturemanger.getfixtureinfo()
+    # to declare that our children do not support funcargs.
+    nofuncargs = True
+
+    def newinstance(self):
+        # TestCase __init__ takes the method (test) name. The TestCase
+        # constructor treats the name "runTest" as a special no-op, so it can be
+        # used when a dummy instance is needed. While unittest.TestCase has a
+        # default, some subclasses omit the default (#9610), so always supply
+        # it.
+        return self.obj("runTest")
+
+    def collect(self) -> Iterable[Item | Collector]:
+        from unittest import TestLoader
+
+        cls = self.obj
+        if not getattr(cls, "__test__", True):
+            return
+
+        skipped = _is_skipped(cls)
+        if not skipped:
+            self._register_unittest_setup_method_fixture(cls)
+            self._register_unittest_setup_class_fixture(cls)
+            self._register_setup_class_fixture()
+
+        self.session._fixturemanager.parsefactories(self.newinstance(), self.nodeid)
+
+        loader = TestLoader()
+        foundsomething = False
+        for name in loader.getTestCaseNames(self.obj):
+            x = getattr(self.obj, name)
+            if not getattr(x, "__test__", True):
+                continue
+            yield TestCaseFunction.from_parent(self, name=name)
+            foundsomething = True
+
+        if not foundsomething:
+            runtest = getattr(self.obj, "runTest", None)
+            if runtest is not None:
+                ut = sys.modules.get("twisted.trial.unittest", None)
+                if ut is None or runtest != ut.TestCase.runTest:
+                    yield TestCaseFunction.from_parent(self, name="runTest")
+
+    def _register_unittest_setup_class_fixture(self, cls: type) -> None:
+        """Register an auto-use fixture to invoke setUpClass and
+        tearDownClass (#517)."""
+        setup = getattr(cls, "setUpClass", None)
+        teardown = getattr(cls, "tearDownClass", None)
+        if setup is None and teardown is None:
+            return None
+        cleanup = getattr(cls, "doClassCleanups", lambda: None)
+
+        def process_teardown_exceptions() -> None:
+            # tearDown_exceptions is a list set in the class containing exc_infos for errors during
+            # teardown for the class.
+            exc_infos = getattr(cls, "tearDown_exceptions", None)
+            if not exc_infos:
+                return
+            exceptions = [exc for (_, exc, _) in exc_infos]
+            # If a single exception, raise it directly as this provides a more readable
+            # error (hopefully this will improve in #12255).
+            if len(exceptions) == 1:
+                raise exceptions[0]
+            else:
+                raise ExceptionGroup("Unittest class cleanup errors", exceptions)
+
+        def unittest_setup_class_fixture(
+            request: FixtureRequest,
+        ) -> Generator[None]:
+            cls = request.cls
+            if _is_skipped(cls):
+                reason = cls.__unittest_skip_why__
+                raise pytest.skip.Exception(reason, _use_item_location=True)
+            if setup is not None:
+                try:
+                    setup()
+                # unittest does not call the cleanup function for every BaseException, so we
+                # follow this here.
+                except Exception:
+                    cleanup()
+                    process_teardown_exceptions()
+                    raise
+            yield
+            try:
+                if teardown is not None:
+                    teardown()
+            finally:
+                cleanup()
+                process_teardown_exceptions()
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_unittest_setUpClass_fixture_{cls.__qualname__}",
+            func=unittest_setup_class_fixture,
+            nodeid=self.nodeid,
+            scope="class",
+            autouse=True,
+        )
+
+    def _register_unittest_setup_method_fixture(self, cls: type) -> None:
+        """Register an auto-use fixture to invoke setup_method and
+        teardown_method (#517)."""
+        setup = getattr(cls, "setup_method", None)
+        teardown = getattr(cls, "teardown_method", None)
+        if setup is None and teardown is None:
+            return None
+
+        def unittest_setup_method_fixture(
+            request: FixtureRequest,
+        ) -> Generator[None]:
+            self = request.instance
+            if _is_skipped(self):
+                reason = self.__unittest_skip_why__
+                raise pytest.skip.Exception(reason, _use_item_location=True)
+            if setup is not None:
+                setup(self, request.function)
+            yield
+            if teardown is not None:
+                teardown(self, request.function)
+
+        self.session._fixturemanager._register_fixture(
+            # Use a unique name to speed up lookup.
+            name=f"_unittest_setup_method_fixture_{cls.__qualname__}",
+            func=unittest_setup_method_fixture,
+            nodeid=self.nodeid,
+            scope="function",
+            autouse=True,
+        )
+
+
+class TestCaseFunction(Function):
+    nofuncargs = True
+    _excinfo: list[_pytest._code.ExceptionInfo[BaseException]] | None = None
+
+    def _getinstance(self):
+        assert isinstance(self.parent, UnitTestCase)
+        return self.parent.obj(self.name)
+
+    # Backward compat for pytest-django; can be removed after pytest-django
+    # updates + some slack.
+    @property
+    def _testcase(self):
+        return self.instance
+
+    def setup(self) -> None:
+        # A bound method to be called during teardown() if set (see 'runtest()').
+        self._explicit_tearDown: Callable[[], None] | None = None
+        super().setup()
+
+    def teardown(self) -> None:
+        if self._explicit_tearDown is not None:
+            self._explicit_tearDown()
+            self._explicit_tearDown = None
+        self._obj = None
+        del self._instance
+        super().teardown()
+
+    def startTest(self, testcase: unittest.TestCase) -> None:
+        pass
+
+    def _addexcinfo(self, rawexcinfo: _SysExcInfoType) -> None:
+        # Unwrap potential exception info (see twisted trial support below).
+        rawexcinfo = getattr(rawexcinfo, "_rawexcinfo", rawexcinfo)
+        try:
+            excinfo = _pytest._code.ExceptionInfo[BaseException].from_exc_info(
+                rawexcinfo  # type: ignore[arg-type]
+            )
+            # Invoke the attributes to trigger storing the traceback
+            # trial causes some issue there.
+            _ = excinfo.value
+            _ = excinfo.traceback
+        except TypeError:
+            try:
+                try:
+                    values = traceback.format_exception(*rawexcinfo)
+                    values.insert(
+                        0,
+                        "NOTE: Incompatible Exception Representation, "
+                        "displaying natively:\n\n",
+                    )
+                    fail("".join(values), pytrace=False)
+                except (fail.Exception, KeyboardInterrupt):
+                    raise
+                except BaseException:
+                    fail(
+                        "ERROR: Unknown Incompatible Exception "
+                        f"representation:\n{rawexcinfo!r}",
+                        pytrace=False,
+                    )
+            except KeyboardInterrupt:
+                raise
+            except fail.Exception:
+                excinfo = _pytest._code.ExceptionInfo.from_current()
+        self.__dict__.setdefault("_excinfo", []).append(excinfo)
+
+    def addError(
+        self, testcase: unittest.TestCase, rawexcinfo: _SysExcInfoType
+    ) -> None:
+        try:
+            if isinstance(rawexcinfo[1], exit.Exception):
+                exit(rawexcinfo[1].msg)
+        except TypeError:
+            pass
+        self._addexcinfo(rawexcinfo)
+
+    def addFailure(
+        self, testcase: unittest.TestCase, rawexcinfo: _SysExcInfoType
+    ) -> None:
+        self._addexcinfo(rawexcinfo)
+
+    def addSkip(self, testcase: unittest.TestCase, reason: str) -> None:
+        try:
+            raise pytest.skip.Exception(reason, _use_item_location=True)
+        except skip.Exception:
+            self._addexcinfo(sys.exc_info())
+
+    def addExpectedFailure(
+        self,
+        testcase: unittest.TestCase,
+        rawexcinfo: _SysExcInfoType,
+        reason: str = "",
+    ) -> None:
+        try:
+            xfail(str(reason))
+        except xfail.Exception:
+            self._addexcinfo(sys.exc_info())
+
+    def addUnexpectedSuccess(
+        self,
+        testcase: unittest.TestCase,
+        reason: twisted.trial.unittest.Todo | None = None,
+    ) -> None:
+        msg = "Unexpected success"
+        if reason:
+            msg += f": {reason.reason}"
+        # Preserve unittest behaviour - fail the test. Explicitly not an XPASS.
+        try:
+            fail(msg, pytrace=False)
+        except fail.Exception:
+            self._addexcinfo(sys.exc_info())
+
+    def addSuccess(self, testcase: unittest.TestCase) -> None:
+        pass
+
+    def stopTest(self, testcase: unittest.TestCase) -> None:
+        pass
+
+    def addDuration(self, testcase: unittest.TestCase, elapsed: float) -> None:
+        pass
+
+    def runtest(self) -> None:
+        from _pytest.debugging import maybe_wrap_pytest_function_for_tracing
+
+        testcase = self.instance
+        assert testcase is not None
+
+        maybe_wrap_pytest_function_for_tracing(self)
+
+        # Let the unittest framework handle async functions.
+        if is_async_function(self.obj):
+            testcase(result=self)
+        else:
+            # When --pdb is given, we want to postpone calling tearDown() otherwise
+            # when entering the pdb prompt, tearDown() would have probably cleaned up
+            # instance variables, which makes it difficult to debug.
+            # Arguably we could always postpone tearDown(), but this changes the moment where the
+            # TestCase instance interacts with the results object, so better to only do it
+            # when absolutely needed.
+            # We need to consider if the test itself is skipped, or the whole class.
+            assert isinstance(self.parent, UnitTestCase)
+            skipped = _is_skipped(self.obj) or _is_skipped(self.parent.obj)
+            if self.config.getoption("usepdb") and not skipped:
+                self._explicit_tearDown = testcase.tearDown
+                setattr(testcase, "tearDown", lambda *args: None)
+
+            # We need to update the actual bound method with self.obj, because
+            # wrap_pytest_function_for_tracing replaces self.obj by a wrapper.
+            setattr(testcase, self.name, self.obj)
+            try:
+                testcase(result=self)
+            finally:
+                delattr(testcase, self.name)
+
+    def _traceback_filter(
+        self, excinfo: _pytest._code.ExceptionInfo[BaseException]
+    ) -> _pytest._code.Traceback:
+        traceback = super()._traceback_filter(excinfo)
+        ntraceback = traceback.filter(
+            lambda x: not x.frame.f_globals.get("__unittest"),
+        )
+        if not ntraceback:
+            ntraceback = traceback
+        return ntraceback
+
+
+@hookimpl(tryfirst=True)
+def pytest_runtest_makereport(item: Item, call: CallInfo[None]) -> None:
+    if isinstance(item, TestCaseFunction):
+        if item._excinfo:
+            call.excinfo = item._excinfo.pop(0)
+            try:
+                del call.result
+            except AttributeError:
+                pass
+
+    # Convert unittest.SkipTest to pytest.skip.
+    # This is actually only needed for nose, which reuses unittest.SkipTest for
+    # its own nose.SkipTest. For unittest TestCases, SkipTest is already
+    # handled internally, and doesn't reach here.
+    unittest = sys.modules.get("unittest")
+    if unittest and call.excinfo and isinstance(call.excinfo.value, unittest.SkipTest):
+        excinfo = call.excinfo
+        call2 = CallInfo[None].from_call(
+            lambda: pytest.skip(str(excinfo.value)), call.when
+        )
+        call.excinfo = call2.excinfo
+
+
+# Twisted trial support.
+classImplements_has_run = False
+
+
+@hookimpl(wrapper=True)
+def pytest_runtest_protocol(item: Item) -> Generator[None, object, object]:
+    if isinstance(item, TestCaseFunction) and "twisted.trial.unittest" in sys.modules:
+        ut: Any = sys.modules["twisted.python.failure"]
+        global classImplements_has_run
+        Failure__init__ = ut.Failure.__init__
+        if not classImplements_has_run:
+            from twisted.trial.itrial import IReporter
+            from zope.interface import classImplements
+
+            classImplements(TestCaseFunction, IReporter)
+            classImplements_has_run = True
+
+        def excstore(
+            self, exc_value=None, exc_type=None, exc_tb=None, captureVars=None
+        ):
+            if exc_value is None:
+                self._rawexcinfo = sys.exc_info()
+            else:
+                if exc_type is None:
+                    exc_type = type(exc_value)
+                self._rawexcinfo = (exc_type, exc_value, exc_tb)
+            try:
+                Failure__init__(
+                    self, exc_value, exc_type, exc_tb, captureVars=captureVars
+                )
+            except TypeError:
+                Failure__init__(self, exc_value, exc_type, exc_tb)
+
+        ut.Failure.__init__ = excstore
+        try:
+            res = yield
+        finally:
+            ut.Failure.__init__ = Failure__init__
+    else:
+        res = yield
+    return res
+
+
+def _is_skipped(obj) -> bool:
+    """Return True if the given object has been marked with @unittest.skip."""
+    return bool(getattr(obj, "__unittest_skip__", False))
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/unraisableexception.py b/backend/venv/lib/python3.12/site-packages/_pytest/unraisableexception.py
new file mode 100644
index 00000000..77a2de20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/unraisableexception.py
@@ -0,0 +1,100 @@
+from __future__ import annotations
+
+import sys
+import traceback
+from types import TracebackType
+from typing import Any
+from typing import Callable
+from typing import Generator
+from typing import TYPE_CHECKING
+import warnings
+
+import pytest
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+
+# Copied from cpython/Lib/test/support/__init__.py, with modifications.
+class catch_unraisable_exception:
+    """Context manager catching unraisable exception using sys.unraisablehook.
+
+    Storing the exception value (cm.unraisable.exc_value) creates a reference
+    cycle. The reference cycle is broken explicitly when the context manager
+    exits.
+
+    Storing the object (cm.unraisable.object) can resurrect it if it is set to
+    an object which is being finalized. Exiting the context manager clears the
+    stored object.
+
+    Usage:
+        with catch_unraisable_exception() as cm:
+            # code creating an "unraisable exception"
+            ...
+            # check the unraisable exception: use cm.unraisable
+            ...
+        # cm.unraisable attribute no longer exists at this point
+        # (to break a reference cycle)
+    """
+
+    def __init__(self) -> None:
+        self.unraisable: sys.UnraisableHookArgs | None = None
+        self._old_hook: Callable[[sys.UnraisableHookArgs], Any] | None = None
+
+    def _hook(self, unraisable: sys.UnraisableHookArgs) -> None:
+        # Storing unraisable.object can resurrect an object which is being
+        # finalized. Storing unraisable.exc_value creates a reference cycle.
+        self.unraisable = unraisable
+
+    def __enter__(self) -> Self:
+        self._old_hook = sys.unraisablehook
+        sys.unraisablehook = self._hook
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> None:
+        assert self._old_hook is not None
+        sys.unraisablehook = self._old_hook
+        self._old_hook = None
+        del self.unraisable
+
+
+def unraisable_exception_runtest_hook() -> Generator[None]:
+    with catch_unraisable_exception() as cm:
+        try:
+            yield
+        finally:
+            if cm.unraisable:
+                if cm.unraisable.err_msg is not None:
+                    err_msg = cm.unraisable.err_msg
+                else:
+                    err_msg = "Exception ignored in"
+                msg = f"{err_msg}: {cm.unraisable.object!r}\n\n"
+                msg += "".join(
+                    traceback.format_exception(
+                        cm.unraisable.exc_type,
+                        cm.unraisable.exc_value,
+                        cm.unraisable.exc_traceback,
+                    )
+                )
+                warnings.warn(pytest.PytestUnraisableExceptionWarning(msg))
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_setup() -> Generator[None]:
+    yield from unraisable_exception_runtest_hook()
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_call() -> Generator[None]:
+    yield from unraisable_exception_runtest_hook()
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_teardown() -> Generator[None]:
+    yield from unraisable_exception_runtest_hook()
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/warning_types.py b/backend/venv/lib/python3.12/site-packages/_pytest/warning_types.py
new file mode 100644
index 00000000..4ab14e48
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/warning_types.py
@@ -0,0 +1,166 @@
+from __future__ import annotations
+
+import dataclasses
+import inspect
+from types import FunctionType
+from typing import Any
+from typing import final
+from typing import Generic
+from typing import TypeVar
+import warnings
+
+
+class PytestWarning(UserWarning):
+    """Base class for all warnings emitted by pytest."""
+
+    __module__ = "pytest"
+
+
+@final
+class PytestAssertRewriteWarning(PytestWarning):
+    """Warning emitted by the pytest assert rewrite module."""
+
+    __module__ = "pytest"
+
+
+@final
+class PytestCacheWarning(PytestWarning):
+    """Warning emitted by the cache plugin in various situations."""
+
+    __module__ = "pytest"
+
+
+@final
+class PytestConfigWarning(PytestWarning):
+    """Warning emitted for configuration issues."""
+
+    __module__ = "pytest"
+
+
+@final
+class PytestCollectionWarning(PytestWarning):
+    """Warning emitted when pytest is not able to collect a file or symbol in a module."""
+
+    __module__ = "pytest"
+
+
+class PytestDeprecationWarning(PytestWarning, DeprecationWarning):
+    """Warning class for features that will be removed in a future version."""
+
+    __module__ = "pytest"
+
+
+class PytestRemovedIn9Warning(PytestDeprecationWarning):
+    """Warning class for features that will be removed in pytest 9."""
+
+    __module__ = "pytest"
+
+
+class PytestReturnNotNoneWarning(PytestWarning):
+    """Warning emitted when a test function is returning value other than None."""
+
+    __module__ = "pytest"
+
+
+@final
+class PytestExperimentalApiWarning(PytestWarning, FutureWarning):
+    """Warning category used to denote experiments in pytest.
+
+    Use sparingly as the API might change or even be removed completely in a
+    future version.
+    """
+
+    __module__ = "pytest"
+
+    @classmethod
+    def simple(cls, apiname: str) -> PytestExperimentalApiWarning:
+        return cls(f"{apiname} is an experimental api that may change over time")
+
+
+@final
+class PytestUnhandledCoroutineWarning(PytestReturnNotNoneWarning):
+    """Warning emitted for an unhandled coroutine.
+
+    A coroutine was encountered when collecting test functions, but was not
+    handled by any async-aware plugin.
+    Coroutine test functions are not natively supported.
+    """
+
+    __module__ = "pytest"
+
+
+@final
+class PytestUnknownMarkWarning(PytestWarning):
+    """Warning emitted on use of unknown markers.
+
+    See :ref:`mark` for details.
+    """
+
+    __module__ = "pytest"
+
+
+@final
+class PytestUnraisableExceptionWarning(PytestWarning):
+    """An unraisable exception was reported.
+
+    Unraisable exceptions are exceptions raised in :meth:`__del__ <object.__del__>`
+    implementations and similar situations when the exception cannot be raised
+    as normal.
+    """
+
+    __module__ = "pytest"
+
+
+@final
+class PytestUnhandledThreadExceptionWarning(PytestWarning):
+    """An unhandled exception occurred in a :class:`~threading.Thread`.
+
+    Such exceptions don't propagate normally.
+    """
+
+    __module__ = "pytest"
+
+
+_W = TypeVar("_W", bound=PytestWarning)
+
+
+@final
+@dataclasses.dataclass
+class UnformattedWarning(Generic[_W]):
+    """A warning meant to be formatted during runtime.
+
+    This is used to hold warnings that need to format their message at runtime,
+    as opposed to a direct message.
+    """
+
+    category: type[_W]
+    template: str
+
+    def format(self, **kwargs: Any) -> _W:
+        """Return an instance of the warning category, formatted with given kwargs."""
+        return self.category(self.template.format(**kwargs))
+
+
+def warn_explicit_for(method: FunctionType, message: PytestWarning) -> None:
+    """
+    Issue the warning :param:`message` for the definition of the given :param:`method`
+
+    this helps to log warnings for functions defined prior to finding an issue with them
+    (like hook wrappers being marked in a legacy mechanism)
+    """
+    lineno = method.__code__.co_firstlineno
+    filename = inspect.getfile(method)
+    module = method.__module__
+    mod_globals = method.__globals__
+    try:
+        warnings.warn_explicit(
+            message,
+            type(message),
+            filename=filename,
+            module=module,
+            registry=mod_globals.setdefault("__warningregistry__", {}),
+            lineno=lineno,
+        )
+    except Warning as w:
+        # If warnings are errors (e.g. -Werror), location information gets lost, so we add it to the message.
+        raise type(w)(f"{w}\n at {filename}:{lineno}") from None
diff --git a/backend/venv/lib/python3.12/site-packages/_pytest/warnings.py b/backend/venv/lib/python3.12/site-packages/_pytest/warnings.py
new file mode 100644
index 00000000..eeb47726
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/_pytest/warnings.py
@@ -0,0 +1,151 @@
+# mypy: allow-untyped-defs
+from __future__ import annotations
+
+from contextlib import contextmanager
+import sys
+from typing import Generator
+from typing import Literal
+import warnings
+
+from _pytest.config import apply_warning_filters
+from _pytest.config import Config
+from _pytest.config import parse_warning_filter
+from _pytest.main import Session
+from _pytest.nodes import Item
+from _pytest.terminal import TerminalReporter
+import pytest
+
+
+def pytest_configure(config: Config) -> None:
+    config.addinivalue_line(
+        "markers",
+        "filterwarnings(warning): add a warning filter to the given test. "
+        "see https://docs.pytest.org/en/stable/how-to/capture-warnings.html#pytest-mark-filterwarnings ",
+    )
+
+
+@contextmanager
+def catch_warnings_for_item(
+    config: Config,
+    ihook,
+    when: Literal["config", "collect", "runtest"],
+    item: Item | None,
+) -> Generator[None]:
+    """Context manager that catches warnings generated in the contained execution block.
+
+    ``item`` can be None if we are not in the context of an item execution.
+
+    Each warning captured triggers the ``pytest_warning_recorded`` hook.
+    """
+    config_filters = config.getini("filterwarnings")
+    cmdline_filters = config.known_args_namespace.pythonwarnings or []
+    with warnings.catch_warnings(record=True) as log:
+        # mypy can't infer that record=True means log is not None; help it.
+        assert log is not None
+
+        if not sys.warnoptions:
+            # If user is not explicitly configuring warning filters, show deprecation warnings by default (#2908).
+            warnings.filterwarnings("always", category=DeprecationWarning)
+            warnings.filterwarnings("always", category=PendingDeprecationWarning)
+
+        # To be enabled in pytest 9.0.0.
+        # warnings.filterwarnings("error", category=pytest.PytestRemovedIn9Warning)
+
+        apply_warning_filters(config_filters, cmdline_filters)
+
+        # apply filters from "filterwarnings" marks
+        nodeid = "" if item is None else item.nodeid
+        if item is not None:
+            for mark in item.iter_markers(name="filterwarnings"):
+                for arg in mark.args:
+                    warnings.filterwarnings(*parse_warning_filter(arg, escape=False))
+
+        try:
+            yield
+        finally:
+            for warning_message in log:
+                ihook.pytest_warning_recorded.call_historic(
+                    kwargs=dict(
+                        warning_message=warning_message,
+                        nodeid=nodeid,
+                        when=when,
+                        location=None,
+                    )
+                )
+
+
+def warning_record_to_str(warning_message: warnings.WarningMessage) -> str:
+    """Convert a warnings.WarningMessage to a string."""
+    warn_msg = warning_message.message
+    msg = warnings.formatwarning(
+        str(warn_msg),
+        warning_message.category,
+        warning_message.filename,
+        warning_message.lineno,
+        warning_message.line,
+    )
+    if warning_message.source is not None:
+        try:
+            import tracemalloc
+        except ImportError:
+            pass
+        else:
+            tb = tracemalloc.get_object_traceback(warning_message.source)
+            if tb is not None:
+                formatted_tb = "\n".join(tb.format())
+                # Use a leading new line to better separate the (large) output
+                # from the traceback to the previous warning text.
+                msg += f"\nObject allocated at:\n{formatted_tb}"
+            else:
+                # No need for a leading new line.
+                url = "https://docs.pytest.org/en/stable/how-to/capture-warnings.html#resource-warnings"
+                msg += "Enable tracemalloc to get traceback where the object was allocated.\n"
+                msg += f"See {url} for more info."
+    return msg
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_runtest_protocol(item: Item) -> Generator[None, object, object]:
+    with catch_warnings_for_item(
+        config=item.config, ihook=item.ihook, when="runtest", item=item
+    ):
+        return (yield)
+
+
+@pytest.hookimpl(wrapper=True, tryfirst=True)
+def pytest_collection(session: Session) -> Generator[None, object, object]:
+    config = session.config
+    with catch_warnings_for_item(
+        config=config, ihook=config.hook, when="collect", item=None
+    ):
+        return (yield)
+
+
+@pytest.hookimpl(wrapper=True)
+def pytest_terminal_summary(
+    terminalreporter: TerminalReporter,
+) -> Generator[None]:
+    config = terminalreporter.config
+    with catch_warnings_for_item(
+        config=config, ihook=config.hook, when="config", item=None
+    ):
+        return (yield)
+
+
+@pytest.hookimpl(wrapper=True)
+def pytest_sessionfinish(session: Session) -> Generator[None]:
+    config = session.config
+    with catch_warnings_for_item(
+        config=config, ihook=config.hook, when="config", item=None
+    ):
+        return (yield)
+
+
+@pytest.hookimpl(wrapper=True)
+def pytest_load_initial_conftests(
+    early_config: Config,
+) -> Generator[None]:
+    with catch_warnings_for_item(
+        config=early_config, ihook=early_config.hook, when="config", item=None
+    ):
+        return (yield)
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/LICENSE.txt
new file mode 100644
index 00000000..79c9825a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/LICENSE.txt
@@ -0,0 +1,20 @@
+Copyright 2010 Jason Kirtland
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be included
+in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/METADATA
new file mode 100644
index 00000000..6d343f57
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/METADATA
@@ -0,0 +1,60 @@
+Metadata-Version: 2.3
+Name: blinker
+Version: 1.9.0
+Summary: Fast, simple object-to-object and broadcast signaling
+Author: Jason Kirtland
+Maintainer-email: Pallets Ecosystem <contact@palletsprojects.com>
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python
+Classifier: Typing :: Typed
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://blinker.readthedocs.io
+Project-URL: Source, https://github.com/pallets-eco/blinker/
+
+# Blinker
+
+Blinker provides a fast dispatching system that allows any number of
+interested parties to subscribe to events, or "signals".
+
+
+## Pallets Community Ecosystem
+
+> [!IMPORTANT]\
+> This project is part of the Pallets Community Ecosystem. Pallets is the open
+> source organization that maintains Flask; Pallets-Eco enables community
+> maintenance of related projects. If you are interested in helping maintain
+> this project, please reach out on [the Pallets Discord server][discord].
+>
+> [discord]: https://discord.gg/pallets
+
+
+## Example
+
+Signal receivers can subscribe to specific senders or receive signals
+sent by any sender.
+
+```pycon
+>>> from blinker import signal
+>>> started = signal('round-started')
+>>> def each(round):
+...     print(f"Round {round}")
+...
+>>> started.connect(each)
+
+>>> def round_two(round):
+...     print("This is round two.")
+...
+>>> started.connect(round_two, sender=2)
+
+>>> for round in range(1, 4):
+...     started.send(round)
+...
+Round 1!
+Round 2!
+This is round two.
+Round 3!
+```
+
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/RECORD
new file mode 100644
index 00000000..b6a9ebee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/RECORD
@@ -0,0 +1,13 @@
+blinker-1.9.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+blinker-1.9.0.dist-info/LICENSE.txt,sha256=nrc6HzhZekqhcCXSrhvjg5Ykx5XphdTw6Xac4p-spGc,1054
+blinker-1.9.0.dist-info/METADATA,sha256=uIRiM8wjjbHkCtbCyTvctU37IAZk0kEe5kxAld1dvzA,1633
+blinker-1.9.0.dist-info/RECORD,,
+blinker-1.9.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+blinker-1.9.0.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+blinker/__init__.py,sha256=I2EdZqpy4LyjX17Hn1yzJGWCjeLaVaPzsMgHkLfj_cQ,317
+blinker/__pycache__/__init__.cpython-312.pyc,,
+blinker/__pycache__/_utilities.cpython-312.pyc,,
+blinker/__pycache__/base.cpython-312.pyc,,
+blinker/_utilities.py,sha256=0J7eeXXTUx0Ivf8asfpx0ycVkp0Eqfqnj117x2mYX9E,1675
+blinker/base.py,sha256=QpDuvXXcwJF49lUBcH5BiST46Rz9wSG7VW_p7N_027M,19132
+blinker/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker-1.9.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/__init__.py b/backend/venv/lib/python3.12/site-packages/blinker/__init__.py
new file mode 100644
index 00000000..1772fa4a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker/__init__.py
@@ -0,0 +1,17 @@
+from __future__ import annotations
+
+from .base import ANY
+from .base import default_namespace
+from .base import NamedSignal
+from .base import Namespace
+from .base import Signal
+from .base import signal
+
+__all__ = [
+    "ANY",
+    "default_namespace",
+    "NamedSignal",
+    "Namespace",
+    "Signal",
+    "signal",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0a3f3233
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/_utilities.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/_utilities.cpython-312.pyc
new file mode 100644
index 00000000..c1226d76
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/_utilities.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..ed38500d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/blinker/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/_utilities.py b/backend/venv/lib/python3.12/site-packages/blinker/_utilities.py
new file mode 100644
index 00000000..000c902a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker/_utilities.py
@@ -0,0 +1,64 @@
+from __future__ import annotations
+
+import collections.abc as c
+import inspect
+import typing as t
+from weakref import ref
+from weakref import WeakMethod
+
+T = t.TypeVar("T")
+
+
+class Symbol:
+    """A constant symbol, nicer than ``object()``. Repeated calls return the
+    same instance.
+
+    >>> Symbol('foo') is Symbol('foo')
+    True
+    >>> Symbol('foo')
+    foo
+    """
+
+    symbols: t.ClassVar[dict[str, Symbol]] = {}
+
+    def __new__(cls, name: str) -> Symbol:
+        if name in cls.symbols:
+            return cls.symbols[name]
+
+        obj = super().__new__(cls)
+        cls.symbols[name] = obj
+        return obj
+
+    def __init__(self, name: str) -> None:
+        self.name = name
+
+    def __repr__(self) -> str:
+        return self.name
+
+    def __getnewargs__(self) -> tuple[t.Any, ...]:
+        return (self.name,)
+
+
+def make_id(obj: object) -> c.Hashable:
+    """Get a stable identifier for a receiver or sender, to be used as a dict
+    key or in a set.
+    """
+    if inspect.ismethod(obj):
+        # The id of a bound method is not stable, but the id of the unbound
+        # function and instance are.
+        return id(obj.__func__), id(obj.__self__)
+
+    if isinstance(obj, (str, int)):
+        # Instances with the same value always compare equal and have the same
+        # hash, even if the id may change.
+        return obj
+
+    # Assume other types are not hashable but will always be the same instance.
+    return id(obj)
+
+
+def make_ref(obj: T, callback: c.Callable[[ref[T]], None] | None = None) -> ref[T]:
+    if inspect.ismethod(obj):
+        return WeakMethod(obj, callback)  # type: ignore[arg-type, return-value]
+
+    return ref(obj, callback)
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/base.py b/backend/venv/lib/python3.12/site-packages/blinker/base.py
new file mode 100644
index 00000000..d051b94a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/blinker/base.py
@@ -0,0 +1,512 @@
+from __future__ import annotations
+
+import collections.abc as c
+import sys
+import typing as t
+import weakref
+from collections import defaultdict
+from contextlib import contextmanager
+from functools import cached_property
+from inspect import iscoroutinefunction
+
+from ._utilities import make_id
+from ._utilities import make_ref
+from ._utilities import Symbol
+
+F = t.TypeVar("F", bound=c.Callable[..., t.Any])
+
+ANY = Symbol("ANY")
+"""Symbol for "any sender"."""
+
+ANY_ID = 0
+
+
+class Signal:
+    """A notification emitter.
+
+    :param doc: The docstring for the signal.
+    """
+
+    ANY = ANY
+    """An alias for the :data:`~blinker.ANY` sender symbol."""
+
+    set_class: type[set[t.Any]] = set
+    """The set class to use for tracking connected receivers and senders.
+    Python's ``set`` is unordered. If receivers must be dispatched in the order
+    they were connected, an ordered set implementation can be used.
+
+    .. versionadded:: 1.7
+    """
+
+    @cached_property
+    def receiver_connected(self) -> Signal:
+        """Emitted at the end of each :meth:`connect` call.
+
+        The signal sender is the signal instance, and the :meth:`connect`
+        arguments are passed through: ``receiver``, ``sender``, and ``weak``.
+
+        .. versionadded:: 1.2
+        """
+        return Signal(doc="Emitted after a receiver connects.")
+
+    @cached_property
+    def receiver_disconnected(self) -> Signal:
+        """Emitted at the end of each :meth:`disconnect` call.
+
+        The sender is the signal instance, and the :meth:`disconnect` arguments
+        are passed through: ``receiver`` and ``sender``.
+
+        This signal is emitted **only** when :meth:`disconnect` is called
+        explicitly. This signal cannot be emitted by an automatic disconnect
+        when a weakly referenced receiver or sender goes out of scope, as the
+        instance is no longer be available to be used as the sender for this
+        signal.
+
+        An alternative approach is available by subscribing to
+        :attr:`receiver_connected` and setting up a custom weakref cleanup
+        callback on weak receivers and senders.
+
+        .. versionadded:: 1.2
+        """
+        return Signal(doc="Emitted after a receiver disconnects.")
+
+    def __init__(self, doc: str | None = None) -> None:
+        if doc:
+            self.__doc__ = doc
+
+        self.receivers: dict[
+            t.Any, weakref.ref[c.Callable[..., t.Any]] | c.Callable[..., t.Any]
+        ] = {}
+        """The map of connected receivers. Useful to quickly check if any
+        receivers are connected to the signal: ``if s.receivers:``. The
+        structure and data is not part of the public API, but checking its
+        boolean value is.
+        """
+
+        self.is_muted: bool = False
+        self._by_receiver: dict[t.Any, set[t.Any]] = defaultdict(self.set_class)
+        self._by_sender: dict[t.Any, set[t.Any]] = defaultdict(self.set_class)
+        self._weak_senders: dict[t.Any, weakref.ref[t.Any]] = {}
+
+    def connect(self, receiver: F, sender: t.Any = ANY, weak: bool = True) -> F:
+        """Connect ``receiver`` to be called when the signal is sent by
+        ``sender``.
+
+        :param receiver: The callable to call when :meth:`send` is called with
+            the given ``sender``, passing ``sender`` as a positional argument
+            along with any extra keyword arguments.
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender. A receiver may be connected
+            to multiple senders by calling :meth:`connect` multiple times.
+        :param weak: Track the receiver with a :mod:`weakref`. The receiver will
+            be automatically disconnected when it is garbage collected. When
+            connecting a receiver defined within a function, set to ``False``,
+            otherwise it will be disconnected when the function scope ends.
+        """
+        receiver_id = make_id(receiver)
+        sender_id = ANY_ID if sender is ANY else make_id(sender)
+
+        if weak:
+            self.receivers[receiver_id] = make_ref(
+                receiver, self._make_cleanup_receiver(receiver_id)
+            )
+        else:
+            self.receivers[receiver_id] = receiver
+
+        self._by_sender[sender_id].add(receiver_id)
+        self._by_receiver[receiver_id].add(sender_id)
+
+        if sender is not ANY and sender_id not in self._weak_senders:
+            # store a cleanup for weakref-able senders
+            try:
+                self._weak_senders[sender_id] = make_ref(
+                    sender, self._make_cleanup_sender(sender_id)
+                )
+            except TypeError:
+                pass
+
+        if "receiver_connected" in self.__dict__ and self.receiver_connected.receivers:
+            try:
+                self.receiver_connected.send(
+                    self, receiver=receiver, sender=sender, weak=weak
+                )
+            except TypeError:
+                # TODO no explanation or test for this
+                self.disconnect(receiver, sender)
+                raise
+
+        return receiver
+
+    def connect_via(self, sender: t.Any, weak: bool = False) -> c.Callable[[F], F]:
+        """Connect the decorated function to be called when the signal is sent
+        by ``sender``.
+
+        The decorated function will be called when :meth:`send` is called with
+        the given ``sender``, passing ``sender`` as a positional argument along
+        with any extra keyword arguments.
+
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender. A receiver may be connected
+            to multiple senders by calling :meth:`connect` multiple times.
+        :param weak: Track the receiver with a :mod:`weakref`. The receiver will
+            be automatically disconnected when it is garbage collected. When
+            connecting a receiver defined within a function, set to ``False``,
+            otherwise it will be disconnected when the function scope ends.=
+
+        .. versionadded:: 1.1
+        """
+
+        def decorator(fn: F) -> F:
+            self.connect(fn, sender, weak)
+            return fn
+
+        return decorator
+
+    @contextmanager
+    def connected_to(
+        self, receiver: c.Callable[..., t.Any], sender: t.Any = ANY
+    ) -> c.Generator[None, None, None]:
+        """A context manager that temporarily connects ``receiver`` to the
+        signal while a ``with`` block executes. When the block exits, the
+        receiver is disconnected. Useful for tests.
+
+        :param receiver: The callable to call when :meth:`send` is called with
+            the given ``sender``, passing ``sender`` as a positional argument
+            along with any extra keyword arguments.
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender.
+
+        .. versionadded:: 1.1
+        """
+        self.connect(receiver, sender=sender, weak=False)
+
+        try:
+            yield None
+        finally:
+            self.disconnect(receiver)
+
+    @contextmanager
+    def muted(self) -> c.Generator[None, None, None]:
+        """A context manager that temporarily disables the signal. No receivers
+        will be called if the signal is sent, until the ``with`` block exits.
+        Useful for tests.
+        """
+        self.is_muted = True
+
+        try:
+            yield None
+        finally:
+            self.is_muted = False
+
+    def send(
+        self,
+        sender: t.Any | None = None,
+        /,
+        *,
+        _async_wrapper: c.Callable[
+            [c.Callable[..., c.Coroutine[t.Any, t.Any, t.Any]]], c.Callable[..., t.Any]
+        ]
+        | None = None,
+        **kwargs: t.Any,
+    ) -> list[tuple[c.Callable[..., t.Any], t.Any]]:
+        """Call all receivers that are connected to the given ``sender``
+        or :data:`ANY`. Each receiver is called with ``sender`` as a positional
+        argument along with any extra keyword arguments. Return a list of
+        ``(receiver, return value)`` tuples.
+
+        The order receivers are called is undefined, but can be influenced by
+        setting :attr:`set_class`.
+
+        If a receiver raises an exception, that exception will propagate up.
+        This makes debugging straightforward, with an assumption that correctly
+        implemented receivers will not raise.
+
+        :param sender: Call receivers connected to this sender, in addition to
+            those connected to :data:`ANY`.
+        :param _async_wrapper: Will be called on any receivers that are async
+            coroutines to turn them into sync callables. For example, could run
+            the receiver with an event loop.
+        :param kwargs: Extra keyword arguments to pass to each receiver.
+
+        .. versionchanged:: 1.7
+            Added the ``_async_wrapper`` argument.
+        """
+        if self.is_muted:
+            return []
+
+        results = []
+
+        for receiver in self.receivers_for(sender):
+            if iscoroutinefunction(receiver):
+                if _async_wrapper is None:
+                    raise RuntimeError("Cannot send to a coroutine function.")
+
+                result = _async_wrapper(receiver)(sender, **kwargs)
+            else:
+                result = receiver(sender, **kwargs)
+
+            results.append((receiver, result))
+
+        return results
+
+    async def send_async(
+        self,
+        sender: t.Any | None = None,
+        /,
+        *,
+        _sync_wrapper: c.Callable[
+            [c.Callable[..., t.Any]], c.Callable[..., c.Coroutine[t.Any, t.Any, t.Any]]
+        ]
+        | None = None,
+        **kwargs: t.Any,
+    ) -> list[tuple[c.Callable[..., t.Any], t.Any]]:
+        """Await all receivers that are connected to the given ``sender``
+        or :data:`ANY`. Each receiver is called with ``sender`` as a positional
+        argument along with any extra keyword arguments. Return a list of
+        ``(receiver, return value)`` tuples.
+
+        The order receivers are called is undefined, but can be influenced by
+        setting :attr:`set_class`.
+
+        If a receiver raises an exception, that exception will propagate up.
+        This makes debugging straightforward, with an assumption that correctly
+        implemented receivers will not raise.
+
+        :param sender: Call receivers connected to this sender, in addition to
+            those connected to :data:`ANY`.
+        :param _sync_wrapper: Will be called on any receivers that are sync
+            callables to turn them into async coroutines. For example,
+            could call the receiver in a thread.
+        :param kwargs: Extra keyword arguments to pass to each receiver.
+
+        .. versionadded:: 1.7
+        """
+        if self.is_muted:
+            return []
+
+        results = []
+
+        for receiver in self.receivers_for(sender):
+            if not iscoroutinefunction(receiver):
+                if _sync_wrapper is None:
+                    raise RuntimeError("Cannot send to a non-coroutine function.")
+
+                result = await _sync_wrapper(receiver)(sender, **kwargs)
+            else:
+                result = await receiver(sender, **kwargs)
+
+            results.append((receiver, result))
+
+        return results
+
+    def has_receivers_for(self, sender: t.Any) -> bool:
+        """Check if there is at least one receiver that will be called with the
+        given ``sender``. A receiver connected to :data:`ANY` will always be
+        called, regardless of sender. Does not check if weakly referenced
+        receivers are still live. See :meth:`receivers_for` for a stronger
+        search.
+
+        :param sender: Check for receivers connected to this sender, in addition
+            to those connected to :data:`ANY`.
+        """
+        if not self.receivers:
+            return False
+
+        if self._by_sender[ANY_ID]:
+            return True
+
+        if sender is ANY:
+            return False
+
+        return make_id(sender) in self._by_sender
+
+    def receivers_for(
+        self, sender: t.Any
+    ) -> c.Generator[c.Callable[..., t.Any], None, None]:
+        """Yield each receiver to be called for ``sender``, in addition to those
+        to be called for :data:`ANY`. Weakly referenced receivers that are not
+        live will be disconnected and skipped.
+
+        :param sender: Yield receivers connected to this sender, in addition
+            to those connected to :data:`ANY`.
+        """
+        # TODO: test receivers_for(ANY)
+        if not self.receivers:
+            return
+
+        sender_id = make_id(sender)
+
+        if sender_id in self._by_sender:
+            ids = self._by_sender[ANY_ID] | self._by_sender[sender_id]
+        else:
+            ids = self._by_sender[ANY_ID].copy()
+
+        for receiver_id in ids:
+            receiver = self.receivers.get(receiver_id)
+
+            if receiver is None:
+                continue
+
+            if isinstance(receiver, weakref.ref):
+                strong = receiver()
+
+                if strong is None:
+                    self._disconnect(receiver_id, ANY_ID)
+                    continue
+
+                yield strong
+            else:
+                yield receiver
+
+    def disconnect(self, receiver: c.Callable[..., t.Any], sender: t.Any = ANY) -> None:
+        """Disconnect ``receiver`` from being called when the signal is sent by
+        ``sender``.
+
+        :param receiver: A connected receiver callable.
+        :param sender: Disconnect from only this sender. By default, disconnect
+            from all senders.
+        """
+        sender_id: c.Hashable
+
+        if sender is ANY:
+            sender_id = ANY_ID
+        else:
+            sender_id = make_id(sender)
+
+        receiver_id = make_id(receiver)
+        self._disconnect(receiver_id, sender_id)
+
+        if (
+            "receiver_disconnected" in self.__dict__
+            and self.receiver_disconnected.receivers
+        ):
+            self.receiver_disconnected.send(self, receiver=receiver, sender=sender)
+
+    def _disconnect(self, receiver_id: c.Hashable, sender_id: c.Hashable) -> None:
+        if sender_id == ANY_ID:
+            if self._by_receiver.pop(receiver_id, None) is not None:
+                for bucket in self._by_sender.values():
+                    bucket.discard(receiver_id)
+
+            self.receivers.pop(receiver_id, None)
+        else:
+            self._by_sender[sender_id].discard(receiver_id)
+            self._by_receiver[receiver_id].discard(sender_id)
+
+    def _make_cleanup_receiver(
+        self, receiver_id: c.Hashable
+    ) -> c.Callable[[weakref.ref[c.Callable[..., t.Any]]], None]:
+        """Create a callback function to disconnect a weakly referenced
+        receiver when it is garbage collected.
+        """
+
+        def cleanup(ref: weakref.ref[c.Callable[..., t.Any]]) -> None:
+            # If the interpreter is shutting down, disconnecting can result in a
+            # weird ignored exception. Don't call it in that case.
+            if not sys.is_finalizing():
+                self._disconnect(receiver_id, ANY_ID)
+
+        return cleanup
+
+    def _make_cleanup_sender(
+        self, sender_id: c.Hashable
+    ) -> c.Callable[[weakref.ref[t.Any]], None]:
+        """Create a callback function to disconnect all receivers for a weakly
+        referenced sender when it is garbage collected.
+        """
+        assert sender_id != ANY_ID
+
+        def cleanup(ref: weakref.ref[t.Any]) -> None:
+            self._weak_senders.pop(sender_id, None)
+
+            for receiver_id in self._by_sender.pop(sender_id, ()):
+                self._by_receiver[receiver_id].discard(sender_id)
+
+        return cleanup
+
+    def _cleanup_bookkeeping(self) -> None:
+        """Prune unused sender/receiver bookkeeping. Not threadsafe.
+
+        Connecting & disconnecting leaves behind a small amount of bookkeeping
+        data. Typical workloads using Blinker, for example in most web apps,
+        Flask, CLI scripts, etc., are not adversely affected by this
+        bookkeeping.
+
+        With a long-running process performing dynamic signal routing with high
+        volume, e.g. connecting to function closures, senders are all unique
+        object instances. Doing all of this over and over may cause memory usage
+        to grow due to extraneous bookkeeping. (An empty ``set`` for each stale
+        sender/receiver pair.)
+
+        This method will prune that bookkeeping away, with the caveat that such
+        pruning is not threadsafe. The risk is that cleanup of a fully
+        disconnected receiver/sender pair occurs while another thread is
+        connecting that same pair. If you are in the highly dynamic, unique
+        receiver/sender situation that has lead you to this method, that failure
+        mode is perhaps not a big deal for you.
+        """
+        for mapping in (self._by_sender, self._by_receiver):
+            for ident, bucket in list(mapping.items()):
+                if not bucket:
+                    mapping.pop(ident, None)
+
+    def _clear_state(self) -> None:
+        """Disconnect all receivers and senders. Useful for tests."""
+        self._weak_senders.clear()
+        self.receivers.clear()
+        self._by_sender.clear()
+        self._by_receiver.clear()
+
+
+class NamedSignal(Signal):
+    """A named generic notification emitter. The name is not used by the signal
+    itself, but matches the key in the :class:`Namespace` that it belongs to.
+
+    :param name: The name of the signal within the namespace.
+    :param doc: The docstring for the signal.
+    """
+
+    def __init__(self, name: str, doc: str | None = None) -> None:
+        super().__init__(doc)
+
+        #: The name of this signal.
+        self.name: str = name
+
+    def __repr__(self) -> str:
+        base = super().__repr__()
+        return f"{base[:-1]}; {self.name!r}>"  # noqa: E702
+
+
+class Namespace(dict[str, NamedSignal]):
+    """A dict mapping names to signals."""
+
+    def signal(self, name: str, doc: str | None = None) -> NamedSignal:
+        """Return the :class:`NamedSignal` for the given ``name``, creating it
+        if required. Repeated calls with the same name return the same signal.
+
+        :param name: The name of the signal.
+        :param doc: The docstring of the signal.
+        """
+        if name not in self:
+            self[name] = NamedSignal(name, doc)
+
+        return self[name]
+
+
+class _PNamespaceSignal(t.Protocol):
+    def __call__(self, name: str, doc: str | None = None) -> NamedSignal: ...
+
+
+default_namespace: Namespace = Namespace()
+"""A default :class:`Namespace` for creating named signals. :func:`signal`
+creates a :class:`NamedSignal` in this namespace.
+"""
+
+signal: _PNamespaceSignal = default_namespace.signal
+"""Return a :class:`NamedSignal` in :data:`default_namespace` with the given
+``name``, creating it if required. Repeated calls with the same name return the
+same signal.
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/blinker/py.typed b/backend/venv/lib/python3.12/site-packages/blinker/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/LICENSE
new file mode 100644
index 00000000..62b076cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/LICENSE
@@ -0,0 +1,20 @@
+This package contains a modified version of ca-bundle.crt:
+
+ca-bundle.crt -- Bundle of CA Root Certificates
+
+This is a bundle of X.509 certificates of public Certificate Authorities
+(CA). These were automatically extracted from Mozilla's root certificates
+file (certdata.txt).  This file can be found in the mozilla source tree:
+https://hg.mozilla.org/mozilla-central/file/tip/security/nss/lib/ckfw/builtins/certdata.txt
+It contains the certificates in PEM format and therefore
+can be directly used with curl / libcurl / php_curl, or with
+an Apache+mod_ssl webserver for SSL client authentication.
+Just configure this file as the SSLCACertificateFile.#
+
+***** BEGIN LICENSE BLOCK *****
+This Source Code Form is subject to the terms of the Mozilla Public License,
+v. 2.0. If a copy of the MPL was not distributed with this file, You can obtain
+one at http://mozilla.org/MPL/2.0/.
+
+***** END LICENSE BLOCK *****
+@(#) $RCSfile: certdata.txt,v $ $Revision: 1.80 $ $Date: 2011/11/03 15:11:58 $
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/METADATA
new file mode 100644
index 00000000..50af1444
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/METADATA
@@ -0,0 +1,77 @@
+Metadata-Version: 2.2
+Name: certifi
+Version: 2025.1.31
+Summary: Python package for providing Mozilla's CA Bundle.
+Home-page: https://github.com/certifi/python-certifi
+Author: Kenneth Reitz
+Author-email: me@kennethreitz.com
+License: MPL-2.0
+Project-URL: Source, https://github.com/certifi/python-certifi
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Mozilla Public License 2.0 (MPL 2.0)
+Classifier: Natural Language :: English
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Requires-Python: >=3.6
+License-File: LICENSE
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: home-page
+Dynamic: license
+Dynamic: project-url
+Dynamic: requires-python
+Dynamic: summary
+
+Certifi: Python SSL Certificates
+================================
+
+Certifi provides Mozilla's carefully curated collection of Root Certificates for
+validating the trustworthiness of SSL certificates while verifying the identity
+of TLS hosts. It has been extracted from the `Requests`_ project.
+
+Installation
+------------
+
+``certifi`` is available on PyPI. Simply install it with ``pip``::
+
+    $ pip install certifi
+
+Usage
+-----
+
+To reference the installed certificate authority (CA) bundle, you can use the
+built-in function::
+
+    >>> import certifi
+
+    >>> certifi.where()
+    '/usr/local/lib/python3.7/site-packages/certifi/cacert.pem'
+
+Or from the command line::
+
+    $ python -m certifi
+    /usr/local/lib/python3.7/site-packages/certifi/cacert.pem
+
+Enjoy!
+
+.. _`Requests`: https://requests.readthedocs.io/en/master/
+
+Addition/Removal of Certificates
+--------------------------------
+
+Certifi does not support any addition/removal or other modification of the
+CA trust store content. This project is intended to provide a reliable and
+highly portable root of trust to python deployments. Look to upstream projects
+for methods to use alternate trust.
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/RECORD
new file mode 100644
index 00000000..e554dc55
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/RECORD
@@ -0,0 +1,14 @@
+certifi-2025.1.31.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+certifi-2025.1.31.dist-info/LICENSE,sha256=6TcW2mucDVpKHfYP5pWzcPBpVgPSH2-D8FPkLPwQyvc,989
+certifi-2025.1.31.dist-info/METADATA,sha256=t5kcT5aGu0dQ6_psUNZYTqnC0uCRnponewm3uYjeHbg,2451
+certifi-2025.1.31.dist-info/RECORD,,
+certifi-2025.1.31.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+certifi-2025.1.31.dist-info/top_level.txt,sha256=KMu4vUCfsjLrkPbSNdgdekS-pVJzBAJFO__nI8NF6-U,8
+certifi/__init__.py,sha256=neIaAf7BM36ygmQCmy-ZsSyjnvjWghFeu13wwEAnjj0,94
+certifi/__main__.py,sha256=xBBoj905TUWBLRGANOcf7oi6e-3dMP4cEoG9OyMs11g,243
+certifi/__pycache__/__init__.cpython-312.pyc,,
+certifi/__pycache__/__main__.cpython-312.pyc,,
+certifi/__pycache__/core.cpython-312.pyc,,
+certifi/cacert.pem,sha256=xVsh-Qf3-G1IrdCTVS-1ZRdJ_1-GBQjMu0I9bB-9gMc,297255
+certifi/core.py,sha256=qRDDFyXVJwTB_EmoGppaXU_R9qCZvhl-EzxPMuV3nTA,4426
+certifi/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/WHEEL
new file mode 100644
index 00000000..505164bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/top_level.txt
new file mode 100644
index 00000000..963eac53
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi-2025.1.31.dist-info/top_level.txt
@@ -0,0 +1 @@
+certifi
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/__init__.py b/backend/venv/lib/python3.12/site-packages/certifi/__init__.py
new file mode 100644
index 00000000..177082e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi/__init__.py
@@ -0,0 +1,4 @@
+from .core import contents, where
+
+__all__ = ["contents", "where"]
+__version__ = "2025.01.31"
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/__main__.py b/backend/venv/lib/python3.12/site-packages/certifi/__main__.py
new file mode 100644
index 00000000..8945b5da
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi/__main__.py
@@ -0,0 +1,12 @@
+import argparse
+
+from certifi import contents, where
+
+parser = argparse.ArgumentParser()
+parser.add_argument("-c", "--contents", action="store_true")
+args = parser.parse_args()
+
+if args.contents:
+    print(contents())
+else:
+    print(where())
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..1175141f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..9cd3c6d1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..05b27e40
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/certifi/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/cacert.pem b/backend/venv/lib/python3.12/site-packages/certifi/cacert.pem
new file mode 100644
index 00000000..860f259b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi/cacert.pem
@@ -0,0 +1,4897 @@
+
+# Issuer: CN=GlobalSign Root CA O=GlobalSign nv-sa OU=Root CA
+# Subject: CN=GlobalSign Root CA O=GlobalSign nv-sa OU=Root CA
+# Label: "GlobalSign Root CA"
+# Serial: 4835703278459707669005204
+# MD5 Fingerprint: 3e:45:52:15:09:51:92:e1:b7:5d:37:9f:b1:87:29:8a
+# SHA1 Fingerprint: b1:bc:96:8b:d4:f4:9d:62:2a:a8:9a:81:f2:15:01:52:a4:1d:82:9c
+# SHA256 Fingerprint: eb:d4:10:40:e4:bb:3e:c7:42:c9:e3:81:d3:1e:f2:a4:1a:48:b6:68:5c:96:e7:ce:f3:c1:df:6c:d4:33:1c:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust.net Certification Authority (2048) O=Entrust.net OU=www.entrust.net/CPS_2048 incorp. by ref. (limits liab.)/(c) 1999 Entrust.net Limited
+# Subject: CN=Entrust.net Certification Authority (2048) O=Entrust.net OU=www.entrust.net/CPS_2048 incorp. by ref. (limits liab.)/(c) 1999 Entrust.net Limited
+# Label: "Entrust.net Premium 2048 Secure Server CA"
+# Serial: 946069240
+# MD5 Fingerprint: ee:29:31:bc:32:7e:9a:e6:e8:b5:f7:51:b4:34:71:90
+# SHA1 Fingerprint: 50:30:06:09:1d:97:d4:f5:ae:39:f7:cb:e7:92:7d:7d:65:2d:34:31
+# SHA256 Fingerprint: 6d:c4:71:72:e0:1c:bc:b0:bf:62:58:0d:89:5f:e2:b8:ac:9a:d4:f8:73:80:1e:0c:10:b9:c8:37:d2:1e:b1:77
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Baltimore CyberTrust Root O=Baltimore OU=CyberTrust
+# Subject: CN=Baltimore CyberTrust Root O=Baltimore OU=CyberTrust
+# Label: "Baltimore CyberTrust Root"
+# Serial: 33554617
+# MD5 Fingerprint: ac:b6:94:a5:9c:17:e0:d7:91:52:9b:b1:97:06:a6:e4
+# SHA1 Fingerprint: d4:de:20:d0:5e:66:fc:53:fe:1a:50:88:2c:78:db:28:52:ca:e4:74
+# SHA256 Fingerprint: 16:af:57:a9:f6:76:b0:ab:12:60:95:aa:5e:ba:de:f2:2a:b3:11:19:d6:44:ac:95:cd:4b:93:db:f3:f2:6a:eb
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority O=Entrust, Inc. OU=www.entrust.net/CPS is incorporated by reference/(c) 2006 Entrust, Inc.
+# Subject: CN=Entrust Root Certification Authority O=Entrust, Inc. OU=www.entrust.net/CPS is incorporated by reference/(c) 2006 Entrust, Inc.
+# Label: "Entrust Root Certification Authority"
+# Serial: 1164660820
+# MD5 Fingerprint: d6:a5:c3:ed:5d:dd:3e:00:c1:3d:87:92:1f:1d:3f:e4
+# SHA1 Fingerprint: b3:1e:b1:b7:40:e3:6c:84:02:da:dc:37:d4:4d:f5:d4:67:49:52:f9
+# SHA256 Fingerprint: 73:c1:76:43:4f:1b:c6:d5:ad:f4:5b:0e:76:e7:27:28:7c:8d:e5:76:16:c1:e6:e6:14:1a:2b:2c:bc:7d:8e:4c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AAA Certificate Services O=Comodo CA Limited
+# Subject: CN=AAA Certificate Services O=Comodo CA Limited
+# Label: "Comodo AAA Services root"
+# Serial: 1
+# MD5 Fingerprint: 49:79:04:b0:eb:87:19:ac:47:b0:bc:11:51:9b:74:d0
+# SHA1 Fingerprint: d1:eb:23:a4:6d:17:d6:8f:d9:25:64:c2:f1:f1:60:17:64:d8:e3:49
+# SHA256 Fingerprint: d7:a7:a0:fb:5d:7e:27:31:d7:71:e9:48:4e:bc:de:f7:1d:5f:0c:3e:0a:29:48:78:2b:c8:3e:e0:ea:69:9e:f4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 2 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 2 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 2"
+# Serial: 1289
+# MD5 Fingerprint: 5e:39:7b:dd:f8:ba:ec:82:e9:ac:62:ba:0c:54:00:2b
+# SHA1 Fingerprint: ca:3a:fb:cf:12:40:36:4b:44:b2:16:20:88:80:48:39:19:93:7c:f7
+# SHA256 Fingerprint: 85:a0:dd:7d:d7:20:ad:b7:ff:05:f8:3d:54:2b:20:9d:c7:ff:45:28:f7:d6:77:b1:83:89:fe:a5:e5:c4:9e:86
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 3"
+# Serial: 1478
+# MD5 Fingerprint: 31:85:3c:62:94:97:63:b9:aa:fd:89:4e:af:6f:e0:cf
+# SHA1 Fingerprint: 1f:49:14:f7:d8:74:95:1d:dd:ae:02:c0:be:fd:3a:2d:82:75:51:85
+# SHA256 Fingerprint: 18:f1:fc:7f:20:5d:f8:ad:dd:eb:7f:e0:07:dd:57:e3:af:37:5a:9c:4d:8d:73:54:6b:f4:f1:fe:d1:e1:8d:35
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=XRamp Global Certification Authority O=XRamp Security Services Inc OU=www.xrampsecurity.com
+# Subject: CN=XRamp Global Certification Authority O=XRamp Security Services Inc OU=www.xrampsecurity.com
+# Label: "XRamp Global CA Root"
+# Serial: 107108908803651509692980124233745014957
+# MD5 Fingerprint: a1:0b:44:b3:ca:10:d8:00:6e:9d:0f:d8:0f:92:0a:d1
+# SHA1 Fingerprint: b8:01:86:d1:eb:9c:86:a5:41:04:cf:30:54:f3:4c:52:b7:e5:58:c6
+# SHA256 Fingerprint: ce:cd:dc:90:50:99:d8:da:df:c5:b1:d2:09:b7:37:cb:e2:c1:8c:fb:2c:10:c0:ff:0b:cf:0d:32:86:fc:1a:a2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=The Go Daddy Group, Inc. OU=Go Daddy Class 2 Certification Authority
+# Subject: O=The Go Daddy Group, Inc. OU=Go Daddy Class 2 Certification Authority
+# Label: "Go Daddy Class 2 CA"
+# Serial: 0
+# MD5 Fingerprint: 91:de:06:25:ab:da:fd:32:17:0c:bb:25:17:2a:84:67
+# SHA1 Fingerprint: 27:96:ba:e6:3f:18:01:e2:77:26:1b:a0:d7:77:70:02:8f:20:ee:e4
+# SHA256 Fingerprint: c3:84:6b:f2:4b:9e:93:ca:64:27:4c:0e:c6:7c:1e:cc:5e:02:4f:fc:ac:d2:d7:40:19:35:0e:81:fe:54:6a:e4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=Starfield Technologies, Inc. OU=Starfield Class 2 Certification Authority
+# Subject: O=Starfield Technologies, Inc. OU=Starfield Class 2 Certification Authority
+# Label: "Starfield Class 2 CA"
+# Serial: 0
+# MD5 Fingerprint: 32:4a:4b:bb:c8:63:69:9b:be:74:9a:c6:dd:1d:46:24
+# SHA1 Fingerprint: ad:7e:1c:28:b0:64:ef:8f:60:03:40:20:14:c3:d0:e3:37:0e:b5:8a
+# SHA256 Fingerprint: 14:65:fa:20:53:97:b8:76:fa:a6:f0:a9:95:8e:55:90:e4:0f:cc:7f:aa:4f:b7:c2:c8:67:75:21:fb:5f:b6:58
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root CA"
+# Serial: 17154717934120587862167794914071425081
+# MD5 Fingerprint: 87:ce:0b:7b:2a:0e:49:00:e1:58:71:9b:37:a8:93:72
+# SHA1 Fingerprint: 05:63:b8:63:0d:62:d7:5a:bb:c8:ab:1e:4b:df:b5:a8:99:b2:4d:43
+# SHA256 Fingerprint: 3e:90:99:b5:01:5e:8f:48:6c:00:bc:ea:9d:11:1e:e7:21:fa:ba:35:5a:89:bc:f1:df:69:56:1e:3d:c6:32:5c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root CA"
+# Serial: 10944719598952040374951832963794454346
+# MD5 Fingerprint: 79:e4:a9:84:0d:7d:3a:96:d7:c0:4f:e2:43:4c:89:2e
+# SHA1 Fingerprint: a8:98:5d:3a:65:e5:e5:c4:b2:d7:d6:6d:40:c6:dd:2f:b1:9c:54:36
+# SHA256 Fingerprint: 43:48:a0:e9:44:4c:78:cb:26:5e:05:8d:5e:89:44:b4:d8:4f:96:62:bd:26:db:25:7f:89:34:a4:43:c7:01:61
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert High Assurance EV Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert High Assurance EV Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert High Assurance EV Root CA"
+# Serial: 3553400076410547919724730734378100087
+# MD5 Fingerprint: d4:74:de:57:5c:39:b2:d3:9c:85:83:c5:c0:65:49:8a
+# SHA1 Fingerprint: 5f:b7:ee:06:33:e2:59:db:ad:0c:4c:9a:e6:d3:8f:1a:61:c7:dc:25
+# SHA256 Fingerprint: 74:31:e5:f4:c3:c1:ce:46:90:77:4f:0b:61:e0:54:40:88:3b:a9:a0:1e:d0:0b:a6:ab:d7:80:6e:d3:b1:18:cf
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SwissSign Gold CA - G2 O=SwissSign AG
+# Subject: CN=SwissSign Gold CA - G2 O=SwissSign AG
+# Label: "SwissSign Gold CA - G2"
+# Serial: 13492815561806991280
+# MD5 Fingerprint: 24:77:d9:a8:91:d1:3b:fa:88:2d:c2:ff:f8:cd:33:93
+# SHA1 Fingerprint: d8:c5:38:8a:b7:30:1b:1b:6e:d4:7a:e6:45:25:3a:6f:9f:1a:27:61
+# SHA256 Fingerprint: 62:dd:0b:e9:b9:f5:0a:16:3e:a0:f8:e7:5c:05:3b:1e:ca:57:ea:55:c8:68:8f:64:7c:68:81:f2:c8:35:7b:95
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureTrust CA O=SecureTrust Corporation
+# Subject: CN=SecureTrust CA O=SecureTrust Corporation
+# Label: "SecureTrust CA"
+# Serial: 17199774589125277788362757014266862032
+# MD5 Fingerprint: dc:32:c3:a7:6d:25:57:c7:68:09:9d:ea:2d:a9:a2:d1
+# SHA1 Fingerprint: 87:82:c6:c3:04:35:3b:cf:d2:96:92:d2:59:3e:7d:44:d9:34:ff:11
+# SHA256 Fingerprint: f1:c1:b5:0a:e5:a2:0d:d8:03:0e:c9:f6:bc:24:82:3d:d3:67:b5:25:57:59:b4:e7:1b:61:fc:e9:f7:37:5d:73
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=Secure Global CA O=SecureTrust Corporation
+# Subject: CN=Secure Global CA O=SecureTrust Corporation
+# Label: "Secure Global CA"
+# Serial: 9751836167731051554232119481456978597
+# MD5 Fingerprint: cf:f4:27:0d:d4:ed:dc:65:16:49:6d:3d:da:bf:6e:de
+# SHA1 Fingerprint: 3a:44:73:5a:e5:81:90:1f:24:86:61:46:1e:3b:9c:c4:5f:f5:3a:1b
+# SHA256 Fingerprint: 42:00:f5:04:3a:c8:59:0e:bb:52:7d:20:9e:d1:50:30:29:fb:cb:d4:1c:a1:b5:06:ec:27:f1:5a:de:7d:ac:69
+-----BEGIN CERTIFICATE-----
+MIIDvDCCAqSgAwIBAgIQB1YipOjUiolN9BPI8PjqpTANBgkqhkiG9w0BAQUFADBK
+MQswCQYDVQQGEwJVUzEgMB4GA1UEChMXU2VjdXJlVHJ1c3QgQ29ycG9yYXRpb24x
+GTAXBgNVBAMTEFNlY3VyZSBHbG9iYWwgQ0EwHhcNMDYxMTA3MTk0MjI4WhcNMjkx
+MjMxMTk1MjA2WjBKMQswCQYDVQQGEwJVUzEgMB4GA1UEChMXU2VjdXJlVHJ1c3Qg
+Q29ycG9yYXRpb24xGTAXBgNVBAMTEFNlY3VyZSBHbG9iYWwgQ0EwggEiMA0GCSqG
+SIb3DQEBAQUAA4IBDwAwggEKAoIBAQCvNS7YrGxVaQZx5RNoJLNP2MwhR/jxYDiJ
+iQPpvepeRlMJ3Fz1Wuj3RSoC6zFh1ykzTM7HfAo3fg+6MpjhHZevj8fcyTiW89sa
+/FHtaMbQbqR8JNGuQsiWUGMu4P51/pinX0kuleM5M2SOHqRfkNJnPLLZ/kG5VacJ
+jnIFHovdRIWCQtBJwB1g8NEXLJXr9qXBkqPFwqcIYA1gBBCWeZ4WNOaptvolRTnI
+HmX5k/Wq8VLcmZg9pYYaDDUz+kulBAYVHDGA76oYa8J719rO+TMg1fW9ajMtgQT7
+sFzUnKPiXB3jqUJ1XnvUd+85VLrJChgbEplJL4hL/VBi0XPnj3pDAgMBAAGjgZ0w
+gZowEwYJKwYBBAGCNxQCBAYeBABDAEEwCwYDVR0PBAQDAgGGMA8GA1UdEwEB/wQF
+MAMBAf8wHQYDVR0OBBYEFK9EBMJBfkiD2045AuzshHrmzsmkMDQGA1UdHwQtMCsw
+KaAnoCWGI2h0dHA6Ly9jcmwuc2VjdXJldHJ1c3QuY29tL1NHQ0EuY3JsMBAGCSsG
+AQQBgjcVAQQDAgEAMA0GCSqGSIb3DQEBBQUAA4IBAQBjGghAfaReUw132HquHw0L
+URYD7xh8yOOvaliTFGCRsoTciE6+OYo68+aCiV0BN7OrJKQVDpI1WkpEXk5X+nXO
+H0jOZvQ8QCaSmGwb7iRGDBezUqXbpZGRzzfTb+cnCDpOGR86p1hcF895P4vkp9Mm
+I50mD1hp/Ed+stCNi5O/KU9DaXR2Z0vPB4zmAve14bRDtUstFJ/53CYNv6ZHdAbY
+iNE6KTCEztI5gGIbqMdXSbxqVVFnFUq+NQfk1XWYN3kwFNspnWzFacxHVaIw98xc
+f8LDmBxrThaA63p4ZUWiABqvDA1VZDRIuJK58bRQKfJPIx/abKwfROHdI3hRW8cW
+-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO Certification Authority O=COMODO CA Limited
+# Label: "COMODO Certification Authority"
+# Serial: 104350513648249232941998508985834464573
+# MD5 Fingerprint: 5c:48:dc:f7:42:72:ec:56:94:6d:1c:cc:71:35:80:75
+# SHA1 Fingerprint: 66:31:bf:9e:f7:4f:9e:b6:c9:d5:a6:0c:ba:6a:be:d1:f7:bd:ef:7b
+# SHA256 Fingerprint: 0c:2c:d6:3d:f7:80:6f:a3:99:ed:e8:09:11:6b:57:5b:f8:79:89:f0:65:18:f9:80:8c:86:05:03:17:8b:af:66
+-----BEGIN CERTIFICATE-----
+MIIEHTCCAwWgAwIBAgIQToEtioJl4AsC7j41AkblPTANBgkqhkiG9w0BAQUFADCB
+gTELMAkGA1UEBhMCR0IxGzAZBgNVBAgTEkdyZWF0ZXIgTWFuY2hlc3RlcjEQMA4G
+A1UEBxMHU2FsZm9yZDEaMBgGA1UEChMRQ09NT0RPIENBIExpbWl0ZWQxJzAlBgNV
+BAMTHkNPTU9ETyBDZXJ0aWZpY2F0aW9uIEF1dGhvcml0eTAeFw0wNjEyMDEwMDAw
+MDBaFw0yOTEyMzEyMzU5NTlaMIGBMQswCQYDVQQGEwJHQjEbMBkGA1UECBMSR3Jl
+YXRlciBNYW5jaGVzdGVyMRAwDgYDVQQHEwdTYWxmb3JkMRowGAYDVQQKExFDT01P
+RE8gQ0EgTGltaXRlZDEnMCUGA1UEAxMeQ09NT0RPIENlcnRpZmljYXRpb24gQXV0
+aG9yaXR5MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEA0ECLi3LjkRv3
+UcEbVASY06m/weaKXTuH+7uIzg3jLz8GlvCiKVCZrts7oVewdFFxze1CkU1B/qnI
+2GqGd0S7WWaXUF601CxwRM/aN5VCaTwwxHGzUvAhTaHYujl8HJ6jJJ3ygxaYqhZ8
+Q5sVW7euNJH+1GImGEaaP+vB+fGQV+useg2L23IwambV4EajcNxo2f8ESIl33rXp
++2dtQem8Ob0y2WIC8bGoPW43nOIv4tOiJovGuFVDiOEjPqXSJDlqR6sA1KGzqSX+
+DT+nHbrTUcELpNqsOO9VUCQFZUaTNE8tja3G1CEZ0o7KBWFxB3NH5YoZEr0ETc5O
+nKVIrLsm9wIDAQABo4GOMIGLMB0GA1UdDgQWBBQLWOWLxkwVN6RAqTCpIb5HNlpW
+/zAOBgNVHQ8BAf8EBAMCAQYwDwYDVR0TAQH/BAUwAwEB/zBJBgNVHR8EQjBAMD6g
+PKA6hjhodHRwOi8vY3JsLmNvbW9kb2NhLmNvbS9DT01PRE9DZXJ0aWZpY2F0aW9u
+QXV0aG9yaXR5LmNybDANBgkqhkiG9w0BAQUFAAOCAQEAPpiem/Yb6dc5t3iuHXIY
+SdOH5EOC6z/JqvWote9VfCFSZfnVDeFs9D6Mk3ORLgLETgdxb8CPOGEIqB6BCsAv
+IC9Bi5HcSEW88cbeunZrM8gALTFGTO3nnc+IlP8zwFboJIYmuNg4ON8qa90SzMc/
+RxdMosIGlgnW2/4/PEZB31jiVg88O8EckzXZOFKs7sjsLjBOlDW0JB9LeGna8gI4
+zJVSk/BwJVmcIGfE7vmLV2H0knZ9P4SNVbfo5azV8fUZVqZa+5Acr5Pr5RzUZ5dd
+BA6+C4OmF4O5MBKgxTMVBbkN+8cFduPYSo38NBejxiEovjBFMR7HeL5YYTisO+IB
+ZQ==
+-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO ECC Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO ECC Certification Authority O=COMODO CA Limited
+# Label: "COMODO ECC Certification Authority"
+# Serial: 41578283867086692638256921589707938090
+# MD5 Fingerprint: 7c:62:ff:74:9d:31:53:5e:68:4a:d5:78:aa:1e:bf:23
+# SHA1 Fingerprint: 9f:74:4e:9f:2b:4d:ba:ec:0f:31:2c:50:b6:56:3b:8e:2d:93:c3:11
+# SHA256 Fingerprint: 17:93:92:7a:06:14:54:97:89:ad:ce:2f:8f:34:f7:f0:b6:6d:0f:3a:e3:a3:b8:4d:21:ec:15:db:ba:4f:ad:c7
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certigna O=Dhimyotis
+# Subject: CN=Certigna O=Dhimyotis
+# Label: "Certigna"
+# Serial: 18364802974209362175
+# MD5 Fingerprint: ab:57:a6:5b:7d:42:82:19:b5:d8:58:26:28:5e:fd:ff
+# SHA1 Fingerprint: b1:2e:13:63:45:86:a4:6f:1a:b2:60:68:37:58:2d:c4:ac:fd:94:97
+# SHA256 Fingerprint: e3:b6:a2:db:2e:d7:ce:48:84:2f:7a:c5:32:41:c7:b7:1d:54:14:4b:fb:40:c1:1f:3f:1d:0b:42:f5:ee:a1:2d
+-----BEGIN CERTIFICATE-----
+MIIDqDCCApCgAwIBAgIJAP7c4wEPyUj/MA0GCSqGSIb3DQEBBQUAMDQxCzAJBgNV
+BAYTAkZSMRIwEAYDVQQKDAlEaGlteW90aXMxETAPBgNVBAMMCENlcnRpZ25hMB4X
+DTA3MDYyOTE1MTMwNVoXDTI3MDYyOTE1MTMwNVowNDELMAkGA1UEBhMCRlIxEjAQ
+BgNVBAoMCURoaW15b3RpczERMA8GA1UEAwwIQ2VydGlnbmEwggEiMA0GCSqGSIb3
+DQEBAQUAA4IBDwAwggEKAoIBAQDIaPHJ1tazNHUmgh7stL7qXOEm7RFHYeGifBZ4
+QCHkYJ5ayGPhxLGWkv8YbWkj4Sti993iNi+RB7lIzw7sebYs5zRLcAglozyHGxny
+gQcPOJAZ0xH+hrTy0V4eHpbNgGzOOzGTtvKg0KmVEn2lmsxryIRWijOp5yIVUxbw
+zBfsV1/pogqYCd7jX5xv3EjjhQsVWqa6n6xI4wmy9/Qy3l40vhx4XUJbzg4ij02Q
+130yGLMLLGq/jj8UEYkgDncUtT2UCIf3JR7VsmAA7G8qKCVuKj4YYxclPz5EIBb2
+JsglrgVKtOdjLPOMFlN+XPsRGgjBRmKfIrjxwo1p3Po6WAbfAgMBAAGjgbwwgbkw
+DwYDVR0TAQH/BAUwAwEB/zAdBgNVHQ4EFgQUGu3+QTmQtCRZvgHyUtVF9lo53BEw
+ZAYDVR0jBF0wW4AUGu3+QTmQtCRZvgHyUtVF9lo53BGhOKQ2MDQxCzAJBgNVBAYT
+AkZSMRIwEAYDVQQKDAlEaGlteW90aXMxETAPBgNVBAMMCENlcnRpZ25hggkA/tzj
+AQ/JSP8wDgYDVR0PAQH/BAQDAgEGMBEGCWCGSAGG+EIBAQQEAwIABzANBgkqhkiG
+9w0BAQUFAAOCAQEAhQMeknH2Qq/ho2Ge6/PAD/Kl1NqV5ta+aDY9fm4fTIrv0Q8h
+bV6lUmPOEvjvKtpv6zf+EwLHyzs+ImvaYS5/1HI93TDhHkxAGYwP15zRgzB7mFnc
+fca5DClMoTOi62c6ZYTTluLtdkVwj7Ur3vkj1kluPBS1xp81HlDQwY9qcEQCYsuu
+HWhBp6pX6FOqB9IG9tUUBguRA3UsbHK1YZWaDYu5Def131TN3ubY1gkIl2PlwS6w
+t0QmwCbAr1UwnjvVNioZBPRcHv/PLLf/0P2HQBHVESO7SMAhqaQoLf0V+LBOK/Qw
+WyH8EZE0vkHve52Xdf+XlcCWWC/qu0bXu+TZLg==
+-----END CERTIFICATE-----
+
+# Issuer: O=Chunghwa Telecom Co., Ltd. OU=ePKI Root Certification Authority
+# Subject: O=Chunghwa Telecom Co., Ltd. OU=ePKI Root Certification Authority
+# Label: "ePKI Root Certification Authority"
+# Serial: 28956088682735189655030529057352760477
+# MD5 Fingerprint: 1b:2e:00:ca:26:06:90:3d:ad:fe:6f:15:68:d3:6b:b3
+# SHA1 Fingerprint: 67:65:0d:f1:7e:8e:7e:5b:82:40:a4:f4:56:4b:cf:e2:3d:69:c6:f0
+# SHA256 Fingerprint: c0:a6:f4:dc:63:a2:4b:fd:cf:54:ef:2a:6a:08:2a:0a:72:de:35:80:3e:2f:f5:ff:52:7a:e5:d8:72:06:df:d5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=certSIGN OU=certSIGN ROOT CA
+# Subject: O=certSIGN OU=certSIGN ROOT CA
+# Label: "certSIGN ROOT CA"
+# Serial: 35210227249154
+# MD5 Fingerprint: 18:98:c0:d6:e9:3a:fc:f9:b0:f5:0c:f7:4b:01:44:17
+# SHA1 Fingerprint: fa:b7:ee:36:97:26:62:fb:2d:b0:2a:f6:bf:03:fd:e8:7c:4b:2f:9b
+# SHA256 Fingerprint: ea:a9:62:c4:fa:4a:6b:af:eb:e4:15:19:6d:35:1c:cd:88:8d:4f:53:f3:fa:8a:e6:d7:c4:66:a9:4e:60:42:bb
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny O=NetLock Kft. OU=Tan\xfas\xedtv\xe1nykiad\xf3k (Certification Services)
+# Subject: CN=NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny O=NetLock Kft. OU=Tan\xfas\xedtv\xe1nykiad\xf3k (Certification Services)
+# Label: "NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny"
+# Serial: 80544274841616
+# MD5 Fingerprint: c5:a1:b7:ff:73:dd:d6:d7:34:32:18:df:fc:3c:ad:88
+# SHA1 Fingerprint: 06:08:3f:59:3f:15:a1:04:a0:69:a4:6b:a9:03:d0:06:b7:97:09:91
+# SHA256 Fingerprint: 6c:61:da:c3:a2:de:f0:31:50:6b:e0:36:d2:a6:fe:40:19:94:fb:d1:3d:f9:c8:d4:66:59:92:74:c4:46:ec:98
+-----BEGIN CERTIFICATE-----
+MIIEFTCCAv2gAwIBAgIGSUEs5AAQMA0GCSqGSIb3DQEBCwUAMIGnMQswCQYDVQQG
+EwJIVTERMA8GA1UEBwwIQnVkYXBlc3QxFTATBgNVBAoMDE5ldExvY2sgS2Z0LjE3
+MDUGA1UECwwuVGFuw7pzw610dsOhbnlraWFkw7NrIChDZXJ0aWZpY2F0aW9uIFNl
+cnZpY2VzKTE1MDMGA1UEAwwsTmV0TG9jayBBcmFueSAoQ2xhc3MgR29sZCkgRsWR
+dGFuw7pzw610dsOhbnkwHhcNMDgxMjExMTUwODIxWhcNMjgxMjA2MTUwODIxWjCB
+pzELMAkGA1UEBhMCSFUxETAPBgNVBAcMCEJ1ZGFwZXN0MRUwEwYDVQQKDAxOZXRM
+b2NrIEtmdC4xNzA1BgNVBAsMLlRhbsO6c8OtdHbDoW55a2lhZMOzayAoQ2VydGlm
+aWNhdGlvbiBTZXJ2aWNlcykxNTAzBgNVBAMMLE5ldExvY2sgQXJhbnkgKENsYXNz
+IEdvbGQpIEbFkXRhbsO6c8OtdHbDoW55MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8A
+MIIBCgKCAQEAxCRec75LbRTDofTjl5Bu0jBFHjzuZ9lk4BqKf8owyoPjIMHj9DrT
+lF8afFttvzBPhCf2nx9JvMaZCpDyD/V/Q4Q3Y1GLeqVw/HpYzY6b7cNGbIRwXdrz
+AZAj/E4wqX7hJ2Pn7WQ8oLjJM2P+FpD/sLj916jAwJRDC7bVWaaeVtAkH3B5r9s5
+VA1lddkVQZQBr17s9o3x/61k/iCa11zr/qYfCGSji3ZVrR47KGAuhyXoqq8fxmRG
+ILdwfzzeSNuWU7c5d+Qa4scWhHaXWy+7GRWF+GmF9ZmnqfI0p6m2pgP8b4Y9VHx2
+BJtr+UBdADTHLpl1neWIA6pN+APSQnbAGwIDAKiLo0UwQzASBgNVHRMBAf8ECDAG
+AQH/AgEEMA4GA1UdDwEB/wQEAwIBBjAdBgNVHQ4EFgQUzPpnk/C2uNClwB7zU/2M
+U9+D15YwDQYJKoZIhvcNAQELBQADggEBAKt/7hwWqZw8UQCgwBEIBaeZ5m8BiFRh
+bvG5GK1Krf6BQCOUL/t1fC8oS2IkgYIL9WHxHG64YTjrgfpioTtaYtOUZcTh5m2C
++C8lcLIhJsFyUR+MLMOEkMNaj7rP9KdlpeuY0fsFskZ1FSNqb4VjMIDw1Z4fKRzC
+bLBQWV2QWzuoDTDPv31/zvGdg73JRm4gpvlhUbohL3u+pRVjodSVh/GeufOJ8z2F
+uLjbvrW5KfnaNwUASZQDhETnv0Mxz3WLJdH0pmT1kvarBes96aULNmLazAZfNou2
+XjG4Kvte9nHfRCaexOYNkbQudZWAUWpLMKawYqGT8ZvYzsRjdT9ZR7E=
+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsec e-Szigno Root CA 2009 O=Microsec Ltd.
+# Subject: CN=Microsec e-Szigno Root CA 2009 O=Microsec Ltd.
+# Label: "Microsec e-Szigno Root CA 2009"
+# Serial: 14014712776195784473
+# MD5 Fingerprint: f8:49:f4:03:bc:44:2d:83:be:48:69:7d:29:64:fc:b1
+# SHA1 Fingerprint: 89:df:74:fe:5c:f4:0f:4a:80:f9:e3:37:7d:54:da:91:e1:01:31:8e
+# SHA256 Fingerprint: 3c:5f:81:fe:a5:fa:b8:2c:64:bf:a2:ea:ec:af:cd:e8:e0:77:fc:86:20:a7:ca:e5:37:16:3d:f3:6e:db:f3:78
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R3
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R3
+# Label: "GlobalSign Root CA - R3"
+# Serial: 4835703278459759426209954
+# MD5 Fingerprint: c5:df:b8:49:ca:05:13:55:ee:2d:ba:1a:c3:3e:b0:28
+# SHA1 Fingerprint: d6:9b:56:11:48:f0:1c:77:c5:45:78:c1:09:26:df:5b:85:69:76:ad
+# SHA256 Fingerprint: cb:b5:22:d7:b7:f1:27:ad:6a:01:13:86:5b:df:1c:d4:10:2e:7d:07:59:af:63:5a:7c:f4:72:0d:c9:63:c5:3b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Izenpe.com O=IZENPE S.A.
+# Subject: CN=Izenpe.com O=IZENPE S.A.
+# Label: "Izenpe.com"
+# Serial: 917563065490389241595536686991402621
+# MD5 Fingerprint: a6:b0:cd:85:80:da:5c:50:34:a3:39:90:2f:55:67:73
+# SHA1 Fingerprint: 2f:78:3d:25:52:18:a7:4a:65:39:71:b5:2c:a2:9c:45:15:6f:e9:19
+# SHA256 Fingerprint: 25:30:cc:8e:98:32:15:02:ba:d9:6f:9b:1f:ba:1b:09:9e:2d:29:9e:0f:45:48:bb:91:4f:36:3b:c0:d4:53:1f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Go Daddy Root Certificate Authority - G2 O=GoDaddy.com, Inc.
+# Subject: CN=Go Daddy Root Certificate Authority - G2 O=GoDaddy.com, Inc.
+# Label: "Go Daddy Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: 80:3a:bc:22:c1:e6:fb:8d:9b:3b:27:4a:32:1b:9a:01
+# SHA1 Fingerprint: 47:be:ab:c9:22:ea:e8:0e:78:78:34:62:a7:9f:45:c2:54:fd:e6:8b
+# SHA256 Fingerprint: 45:14:0b:32:47:eb:9c:c8:c5:b4:f0:d7:b5:30:91:f7:32:92:08:9e:6e:5a:63:e2:74:9d:d3:ac:a9:19:8e:da
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Starfield Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Subject: CN=Starfield Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Label: "Starfield Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: d6:39:81:c6:52:7e:96:69:fc:fc:ca:66:ed:05:f2:96
+# SHA1 Fingerprint: b5:1c:06:7c:ee:2b:0c:3d:f8:55:ab:2d:92:f4:fe:39:d4:e7:0f:0e
+# SHA256 Fingerprint: 2c:e1:cb:0b:f9:d2:f9:e1:02:99:3f:be:21:51:52:c3:b2:dd:0c:ab:de:1c:68:e5:31:9b:83:91:54:db:b7:f5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Starfield Services Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Subject: CN=Starfield Services Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Label: "Starfield Services Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: 17:35:74:af:7b:61:1c:eb:f4:f9:3c:e2:ee:40:f9:a2
+# SHA1 Fingerprint: 92:5a:8f:8d:2c:6d:04:e0:66:5f:59:6a:ff:22:d8:63:e8:25:6f:3f
+# SHA256 Fingerprint: 56:8d:69:05:a2:c8:87:08:a4:b3:02:51:90:ed:cf:ed:b1:97:4a:60:6a:13:c6:e5:29:0f:cb:2a:e6:3e:da:b5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Commercial O=AffirmTrust
+# Subject: CN=AffirmTrust Commercial O=AffirmTrust
+# Label: "AffirmTrust Commercial"
+# Serial: 8608355977964138876
+# MD5 Fingerprint: 82:92:ba:5b:ef:cd:8a:6f:a6:3d:55:f9:84:f6:d6:b7
+# SHA1 Fingerprint: f9:b5:b6:32:45:5f:9c:be:ec:57:5f:80:dc:e9:6e:2c:c7:b2:78:b7
+# SHA256 Fingerprint: 03:76:ab:1d:54:c5:f9:80:3c:e4:b2:e2:01:a0:ee:7e:ef:7b:57:b6:36:e8:a9:3c:9b:8d:48:60:c9:6f:5f:a7
+-----BEGIN CERTIFICATE-----
+MIIDTDCCAjSgAwIBAgIId3cGJyapsXwwDQYJKoZIhvcNAQELBQAwRDELMAkGA1UE
+BhMCVVMxFDASBgNVBAoMC0FmZmlybVRydXN0MR8wHQYDVQQDDBZBZmZpcm1UcnVz
+dCBDb21tZXJjaWFsMB4XDTEwMDEyOTE0MDYwNloXDTMwMTIzMTE0MDYwNlowRDEL
+MAkGA1UEBhMCVVMxFDASBgNVBAoMC0FmZmlybVRydXN0MR8wHQYDVQQDDBZBZmZp
+cm1UcnVzdCBDb21tZXJjaWFsMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKC
+AQEA9htPZwcroRX1BiLLHwGy43NFBkRJLLtJJRTWzsO3qyxPxkEylFf6EqdbDuKP
+Hx6GGaeqtS25Xw2Kwq+FNXkyLbscYjfysVtKPcrNcV/pQr6U6Mje+SJIZMblq8Yr
+ba0F8PrVC8+a5fBQpIs7R6UjW3p6+DM/uO+Zl+MgwdYoic+U+7lF7eNAFxHUdPAL
+MeIrJmqbTFeurCA+ukV6BfO9m2kVrn1OIGPENXY6BwLJN/3HR+7o8XYdcxXyl6S1
+yHp52UKqK39c/s4mT6NmgTWvRLpUHhwwMmWd5jyTXlBOeuM61G7MGvv50jeuJCqr
+VwMiKA1JdX+3KNp1v47j3A55MQIDAQABo0IwQDAdBgNVHQ4EFgQUnZPGU4teyq8/
+nx4P5ZmVvCT2lI8wDwYDVR0TAQH/BAUwAwEB/zAOBgNVHQ8BAf8EBAMCAQYwDQYJ
+KoZIhvcNAQELBQADggEBAFis9AQOzcAN/wr91LoWXym9e2iZWEnStB03TX8nfUYG
+XUPGhi4+c7ImfU+TqbbEKpqrIZcUsd6M06uJFdhrJNTxFq7YpFzUf1GO7RgBsZNj
+vbz4YYCanrHOQnDiqX0GJX0nof5v7LMeJNrjS1UaADs1tDvZ110w/YETifLCBivt
+Z8SOyUOyXGsViQK8YvxO8rUzqrJv0wqiUOP2O+guRMLbZjipM1ZI8W0bM40NjD9g
+N53Tym1+NH4Nn3J2ixufcv1SNUFFApYvHLKac0khsUlHRUe072o0EclNmsxZt9YC
+nlpOZbWUrhvfKbAW8b8Angc6F2S1BLUjIZkKlTuXfO8=
+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Networking O=AffirmTrust
+# Subject: CN=AffirmTrust Networking O=AffirmTrust
+# Label: "AffirmTrust Networking"
+# Serial: 8957382827206547757
+# MD5 Fingerprint: 42:65:ca:be:01:9a:9a:4c:a9:8c:41:49:cd:c0:d5:7f
+# SHA1 Fingerprint: 29:36:21:02:8b:20:ed:02:f5:66:c5:32:d1:d6:ed:90:9f:45:00:2f
+# SHA256 Fingerprint: 0a:81:ec:5a:92:97:77:f1:45:90:4a:f3:8d:5d:50:9f:66:b5:e2:c5:8f:cd:b5:31:05:8b:0e:17:f3:f0:b4:1b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Premium O=AffirmTrust
+# Subject: CN=AffirmTrust Premium O=AffirmTrust
+# Label: "AffirmTrust Premium"
+# Serial: 7893706540734352110
+# MD5 Fingerprint: c4:5d:0e:48:b6:ac:28:30:4e:0a:bc:f9:38:16:87:57
+# SHA1 Fingerprint: d8:a6:33:2c:e0:03:6f:b1:85:f6:63:4f:7d:6a:06:65:26:32:28:27
+# SHA256 Fingerprint: 70:a7:3f:7f:37:6b:60:07:42:48:90:45:34:b1:14:82:d5:bf:0e:69:8e:cc:49:8d:f5:25:77:eb:f2:e9:3b:9a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Premium ECC O=AffirmTrust
+# Subject: CN=AffirmTrust Premium ECC O=AffirmTrust
+# Label: "AffirmTrust Premium ECC"
+# Serial: 8401224907861490260
+# MD5 Fingerprint: 64:b0:09:55:cf:b1:d5:99:e2:be:13:ab:a6:5d:ea:4d
+# SHA1 Fingerprint: b8:23:6b:00:2f:1d:16:86:53:01:55:6c:11:a4:37:ca:eb:ff:c3:bb
+# SHA256 Fingerprint: bd:71:fd:f6:da:97:e4:cf:62:d1:64:7a:dd:25:81:b0:7d:79:ad:f8:39:7e:b4:ec:ba:9c:5e:84:88:82:14:23
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Network CA O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Network CA O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Network CA"
+# Serial: 279744
+# MD5 Fingerprint: d5:e9:81:40:c5:18:69:fc:46:2c:89:75:62:0f:aa:78
+# SHA1 Fingerprint: 07:e0:32:e0:20:b7:2c:3f:19:2f:06:28:a2:59:3a:19:a7:0f:06:9e
+# SHA256 Fingerprint: 5c:58:46:8d:55:f5:8e:49:7e:74:39:82:d2:b5:00:10:b6:d1:65:37:4a:cf:83:a7:d4:a3:2d:b7:68:c4:40:8e
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA Root Certification Authority O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA Root Certification Authority O=TAIWAN-CA OU=Root CA
+# Label: "TWCA Root Certification Authority"
+# Serial: 1
+# MD5 Fingerprint: aa:08:8f:f6:f9:7b:b7:f2:b1:a7:1e:9b:ea:ea:bd:79
+# SHA1 Fingerprint: cf:9e:87:6d:d3:eb:fc:42:26:97:a3:b5:a3:7a:a0:76:a9:06:23:48
+# SHA256 Fingerprint: bf:d8:8f:e1:10:1c:41:ae:3e:80:1b:f8:be:56:35:0e:e9:ba:d1:a6:b9:bd:51:5e:dc:5c:6d:5b:87:11:ac:44
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=SECOM Trust Systems CO.,LTD. OU=Security Communication RootCA2
+# Subject: O=SECOM Trust Systems CO.,LTD. OU=Security Communication RootCA2
+# Label: "Security Communication RootCA2"
+# Serial: 0
+# MD5 Fingerprint: 6c:39:7d:a4:0e:55:59:b2:3f:d6:41:b1:12:50:de:43
+# SHA1 Fingerprint: 5f:3b:8c:f2:f8:10:b3:7d:78:b4:ce:ec:19:19:c3:73:34:b9:c7:74
+# SHA256 Fingerprint: 51:3b:2c:ec:b8:10:d4:cd:e5:dd:85:39:1a:df:c6:c2:dd:60:d8:7b:b7:36:d2:b5:21:48:4a:a4:7a:0e:be:f6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Actalis Authentication Root CA O=Actalis S.p.A./03358520967
+# Subject: CN=Actalis Authentication Root CA O=Actalis S.p.A./03358520967
+# Label: "Actalis Authentication Root CA"
+# Serial: 6271844772424770508
+# MD5 Fingerprint: 69:c1:0d:4f:07:a3:1b:c3:fe:56:3d:04:bc:11:f6:a6
+# SHA1 Fingerprint: f3:73:b3:87:06:5a:28:84:8a:f2:f3:4a:ce:19:2b:dd:c7:8e:9c:ac
+# SHA256 Fingerprint: 55:92:60:84:ec:96:3a:64:b9:6e:2a:be:01:ce:0b:a8:6a:64:fb:fe:bc:c7:aa:b5:af:c1:55:b3:7f:d7:60:66
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Buypass Class 2 Root CA O=Buypass AS-983163327
+# Subject: CN=Buypass Class 2 Root CA O=Buypass AS-983163327
+# Label: "Buypass Class 2 Root CA"
+# Serial: 2
+# MD5 Fingerprint: 46:a7:d2:fe:45:fb:64:5a:a8:59:90:9b:78:44:9b:29
+# SHA1 Fingerprint: 49:0a:75:74:de:87:0a:47:fe:58:ee:f6:c7:6b:eb:c6:0b:12:40:99
+# SHA256 Fingerprint: 9a:11:40:25:19:7c:5b:b9:5d:94:e6:3d:55:cd:43:79:08:47:b6:46:b2:3c:df:11:ad:a4:a0:0e:ff:15:fb:48
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Buypass Class 3 Root CA O=Buypass AS-983163327
+# Subject: CN=Buypass Class 3 Root CA O=Buypass AS-983163327
+# Label: "Buypass Class 3 Root CA"
+# Serial: 2
+# MD5 Fingerprint: 3d:3b:18:9e:2c:64:5a:e8:d5:88:ce:0e:f9:37:c2:ec
+# SHA1 Fingerprint: da:fa:f7:fa:66:84:ec:06:8f:14:50:bd:c7:c2:81:a5:bc:a9:64:57
+# SHA256 Fingerprint: ed:f7:eb:bc:a2:7a:2a:38:4d:38:7b:7d:40:10:c6:66:e2:ed:b4:84:3e:4c:29:b4:ae:1d:5b:93:32:e6:b2:4d
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=T-TeleSec GlobalRoot Class 3 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Subject: CN=T-TeleSec GlobalRoot Class 3 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Label: "T-TeleSec GlobalRoot Class 3"
+# Serial: 1
+# MD5 Fingerprint: ca:fb:40:a8:4e:39:92:8a:1d:fe:8e:2f:c4:27:ea:ef
+# SHA1 Fingerprint: 55:a6:72:3e:cb:f2:ec:cd:c3:23:74:70:19:9d:2a:be:11:e3:81:d1
+# SHA256 Fingerprint: fd:73:da:d3:1c:64:4f:f1:b4:3b:ef:0c:cd:da:96:71:0b:9c:d9:87:5e:ca:7e:31:70:7a:f3:e9:6d:52:2b:bd
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST Root Class 3 CA 2 2009 O=D-Trust GmbH
+# Subject: CN=D-TRUST Root Class 3 CA 2 2009 O=D-Trust GmbH
+# Label: "D-TRUST Root Class 3 CA 2 2009"
+# Serial: 623603
+# MD5 Fingerprint: cd:e0:25:69:8d:47:ac:9c:89:35:90:f7:fd:51:3d:2f
+# SHA1 Fingerprint: 58:e8:ab:b0:36:15:33:fb:80:f7:9b:1b:6d:29:d3:ff:8d:5f:00:f0
+# SHA256 Fingerprint: 49:e7:a4:42:ac:f0:ea:62:87:05:00:54:b5:25:64:b6:50:e4:f4:9e:42:e3:48:d6:aa:38:e0:39:e9:57:b1:c1
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST Root Class 3 CA 2 EV 2009 O=D-Trust GmbH
+# Subject: CN=D-TRUST Root Class 3 CA 2 EV 2009 O=D-Trust GmbH
+# Label: "D-TRUST Root Class 3 CA 2 EV 2009"
+# Serial: 623604
+# MD5 Fingerprint: aa:c6:43:2c:5e:2d:cd:c4:34:c0:50:4f:11:02:4f:b6
+# SHA1 Fingerprint: 96:c9:1b:0b:95:b4:10:98:42:fa:d0:d8:22:79:fe:60:fa:b9:16:83
+# SHA256 Fingerprint: ee:c5:49:6b:98:8c:e9:86:25:b9:34:09:2e:ec:29:08:be:d0:b0:f3:16:c2:d4:73:0c:84:ea:f1:f3:d3:48:81
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CA Disig Root R2 O=Disig a.s.
+# Subject: CN=CA Disig Root R2 O=Disig a.s.
+# Label: "CA Disig Root R2"
+# Serial: 10572350602393338211
+# MD5 Fingerprint: 26:01:fb:d8:27:a7:17:9a:45:54:38:1a:43:01:3b:03
+# SHA1 Fingerprint: b5:61:eb:ea:a4:de:e4:25:4b:69:1a:98:a5:57:47:c2:34:c7:d9:71
+# SHA256 Fingerprint: e2:3d:4a:03:6d:7b:70:e9:f5:95:b1:42:20:79:d2:b9:1e:df:bb:1f:b6:51:a0:63:3e:aa:8a:9d:c5:f8:07:03
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=ACCVRAIZ1 O=ACCV OU=PKIACCV
+# Subject: CN=ACCVRAIZ1 O=ACCV OU=PKIACCV
+# Label: "ACCVRAIZ1"
+# Serial: 6828503384748696800
+# MD5 Fingerprint: d0:a0:5a:ee:05:b6:09:94:21:a1:7d:f1:b2:29:82:02
+# SHA1 Fingerprint: 93:05:7a:88:15:c6:4f:ce:88:2f:fa:91:16:52:28:78:bc:53:64:17
+# SHA256 Fingerprint: 9a:6e:c0:12:e1:a7:da:9d:be:34:19:4d:47:8a:d7:c0:db:18:22:fb:07:1d:f1:29:81:49:6e:d1:04:38:41:13
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA Global Root CA O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA Global Root CA O=TAIWAN-CA OU=Root CA
+# Label: "TWCA Global Root CA"
+# Serial: 3262
+# MD5 Fingerprint: f9:03:7e:cf:e6:9e:3c:73:7a:2a:90:07:69:ff:2b:96
+# SHA1 Fingerprint: 9c:bb:48:53:f6:a4:f6:d3:52:a4:e8:32:52:55:60:13:f5:ad:af:65
+# SHA256 Fingerprint: 59:76:90:07:f7:68:5d:0f:cd:50:87:2f:9f:95:d5:75:5a:5b:2b:45:7d:81:f3:69:2b:61:0a:98:67:2f:0e:1b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TeliaSonera Root CA v1 O=TeliaSonera
+# Subject: CN=TeliaSonera Root CA v1 O=TeliaSonera
+# Label: "TeliaSonera Root CA v1"
+# Serial: 199041966741090107964904287217786801558
+# MD5 Fingerprint: 37:41:49:1b:18:56:9a:26:f5:ad:c2:66:fb:40:a5:4c
+# SHA1 Fingerprint: 43:13:bb:96:f1:d5:86:9b:c1:4e:6a:92:f6:cf:f6:34:69:87:82:37
+# SHA256 Fingerprint: dd:69:36:fe:21:f8:f0:77:c1:23:a1:a5:21:c1:22:24:f7:22:55:b7:3e:03:a7:26:06:93:e8:a2:4b:0f:a3:89
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=T-TeleSec GlobalRoot Class 2 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Subject: CN=T-TeleSec GlobalRoot Class 2 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Label: "T-TeleSec GlobalRoot Class 2"
+# Serial: 1
+# MD5 Fingerprint: 2b:9b:9e:e4:7b:6c:1f:00:72:1a:cc:c1:77:79:df:6a
+# SHA1 Fingerprint: 59:0d:2d:7d:88:4f:40:2e:61:7e:a5:62:32:17:65:cf:17:d8:94:e9
+# SHA256 Fingerprint: 91:e2:f5:78:8d:58:10:eb:a7:ba:58:73:7d:e1:54:8a:8e:ca:cd:01:45:98:bc:0b:14:3e:04:1b:17:05:25:52
+-----BEGIN CERTIFICATE-----
+MIIDwzCCAqugAwIBAgIBATANBgkqhkiG9w0BAQsFADCBgjELMAkGA1UEBhMCREUx
+KzApBgNVBAoMIlQtU3lzdGVtcyBFbnRlcnByaXNlIFNlcnZpY2VzIEdtYkgxHzAd
+BgNVBAsMFlQtU3lzdGVtcyBUcnVzdCBDZW50ZXIxJTAjBgNVBAMMHFQtVGVsZVNl
+YyBHbG9iYWxSb290IENsYXNzIDIwHhcNMDgxMDAxMTA0MDE0WhcNMzMxMDAxMjM1
+OTU5WjCBgjELMAkGA1UEBhMCREUxKzApBgNVBAoMIlQtU3lzdGVtcyBFbnRlcnBy
+aXNlIFNlcnZpY2VzIEdtYkgxHzAdBgNVBAsMFlQtU3lzdGVtcyBUcnVzdCBDZW50
+ZXIxJTAjBgNVBAMMHFQtVGVsZVNlYyBHbG9iYWxSb290IENsYXNzIDIwggEiMA0G
+CSqGSIb3DQEBAQUAA4IBDwAwggEKAoIBAQCqX9obX+hzkeXaXPSi5kfl82hVYAUd
+AqSzm1nzHoqvNK38DcLZSBnuaY/JIPwhqgcZ7bBcrGXHX+0CfHt8LRvWurmAwhiC
+FoT6ZrAIxlQjgeTNuUk/9k9uN0goOA/FvudocP05l03Sx5iRUKrERLMjfTlH6VJi
+1hKTXrcxlkIF+3anHqP1wvzpesVsqXFP6st4vGCvx9702cu+fjOlbpSD8DT6Iavq
+jnKgP6TeMFvvhk1qlVtDRKgQFRzlAVfFmPHmBiiRqiDFt1MmUUOyCxGVWOHAD3bZ
+wI18gfNycJ5v/hqO2V81xrJvNHy+SE/iWjnX2J14np+GPgNeGYtEotXHAgMBAAGj
+QjBAMA8GA1UdEwEB/wQFMAMBAf8wDgYDVR0PAQH/BAQDAgEGMB0GA1UdDgQWBBS/
+WSA2AHmgoCJrjNXyYdK4LMuCSjANBgkqhkiG9w0BAQsFAAOCAQEAMQOiYQsfdOhy
+NsZt+U2e+iKo4YFWz827n+qrkRk4r6p8FU3ztqONpfSO9kSpp+ghla0+AGIWiPAC
+uvxhI+YzmzB6azZie60EI4RYZeLbK4rnJVM3YlNfvNoBYimipidx5joifsFvHZVw
+IEoHNN/q/xWA5brXethbdXwFeilHfkCoMRN3zUA7tFFHei4R40cR3p1m0IvVVGb6
+g1XqfMIpiRvpb7PO4gWEyS8+eIVibslfwXhjdFjASBgMmTnrpMwatXlajRWc2BQN
+9noHV8cigwUtPJslJj0Ys6lDfMjIq2SPDqO/nBudMNva0Bkuqjzx+zOAduTNrRlP
+BSeOE6Fuwg==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot 2011 O=Atos
+# Subject: CN=Atos TrustedRoot 2011 O=Atos
+# Label: "Atos TrustedRoot 2011"
+# Serial: 6643877497813316402
+# MD5 Fingerprint: ae:b9:c4:32:4b:ac:7f:5d:66:cc:77:94:bb:2a:77:56
+# SHA1 Fingerprint: 2b:b1:f5:3e:55:0c:1d:c5:f1:d4:e6:b7:6a:46:4b:55:06:02:ac:21
+# SHA256 Fingerprint: f3:56:be:a2:44:b7:a9:1e:b3:5d:53:ca:9a:d7:86:4a:ce:01:8e:2d:35:d5:f8:f9:6d:df:68:a6:f4:1a:a4:74
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 1 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 1 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 1 G3"
+# Serial: 687049649626669250736271037606554624078720034195
+# MD5 Fingerprint: a4:bc:5b:3f:fe:37:9a:fa:64:f0:e2:fa:05:3d:0b:ab
+# SHA1 Fingerprint: 1b:8e:ea:57:96:29:1a:c9:39:ea:b8:0a:81:1a:73:73:c0:93:79:67
+# SHA256 Fingerprint: 8a:86:6f:d1:b2:76:b5:7e:57:8e:92:1c:65:82:8a:2b:ed:58:e9:f2:f2:88:05:41:34:b7:f1:f4:bf:c9:cc:74
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 2 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 2 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 2 G3"
+# Serial: 390156079458959257446133169266079962026824725800
+# MD5 Fingerprint: af:0c:86:6e:bf:40:2d:7f:0b:3e:12:50:ba:12:3d:06
+# SHA1 Fingerprint: 09:3c:61:f3:8b:8b:dc:7d:55:df:75:38:02:05:00:e1:25:f5:c8:36
+# SHA256 Fingerprint: 8f:e4:fb:0a:f9:3a:4d:0d:67:db:0b:eb:b2:3e:37:c7:1b:f3:25:dc:bc:dd:24:0e:a0:4d:af:58:b4:7e:18:40
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 3 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 3 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 3 G3"
+# Serial: 268090761170461462463995952157327242137089239581
+# MD5 Fingerprint: df:7d:b9:ad:54:6f:68:a1:df:89:57:03:97:43:b0:d7
+# SHA1 Fingerprint: 48:12:bd:92:3c:a8:c4:39:06:e7:30:6d:27:96:e6:a4:cf:22:2e:7d
+# SHA256 Fingerprint: 88:ef:81:de:20:2e:b0:18:45:2e:43:f8:64:72:5c:ea:5f:bd:1f:c2:d9:d2:05:73:07:09:c5:d8:b8:69:0f:46
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root G2 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root G2 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root G2"
+# Serial: 15385348160840213938643033620894905419
+# MD5 Fingerprint: 92:38:b9:f8:63:24:82:65:2c:57:33:e6:fe:81:8f:9d
+# SHA1 Fingerprint: a1:4b:48:d9:43:ee:0a:0e:40:90:4f:3c:e0:a4:c0:91:93:51:5d:3f
+# SHA256 Fingerprint: 7d:05:eb:b6:82:33:9f:8c:94:51:ee:09:4e:eb:fe:fa:79:53:a1:14:ed:b2:f4:49:49:45:2f:ab:7d:2f:c1:85
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root G3 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root G3 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root G3"
+# Serial: 15459312981008553731928384953135426796
+# MD5 Fingerprint: 7c:7f:65:31:0c:81:df:8d:ba:3e:99:e2:5c:ad:6e:fb
+# SHA1 Fingerprint: f5:17:a2:4f:9a:48:c6:c9:f8:a2:00:26:9f:dc:0f:48:2c:ab:30:89
+# SHA256 Fingerprint: 7e:37:cb:8b:4c:47:09:0c:ab:36:55:1b:a6:f4:5d:b8:40:68:0f:ba:16:6a:95:2d:b1:00:71:7f:43:05:3f:c2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root G2 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root G2 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root G2"
+# Serial: 4293743540046975378534879503202253541
+# MD5 Fingerprint: e4:a6:8a:c8:54:ac:52:42:46:0a:fd:72:48:1b:2a:44
+# SHA1 Fingerprint: df:3c:24:f9:bf:d6:66:76:1b:26:80:73:fe:06:d1:cc:8d:4f:82:a4
+# SHA256 Fingerprint: cb:3c:cb:b7:60:31:e5:e0:13:8f:8d:d3:9a:23:f9:de:47:ff:c3:5e:43:c1:14:4c:ea:27:d4:6a:5a:b1:cb:5f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root G3 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root G3 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root G3"
+# Serial: 7089244469030293291760083333884364146
+# MD5 Fingerprint: f5:5d:a4:50:a5:fb:28:7e:1e:0f:0d:cc:96:57:56:ca
+# SHA1 Fingerprint: 7e:04:de:89:6a:3e:66:6d:00:e6:87:d3:3f:fa:d9:3b:e8:3d:34:9e
+# SHA256 Fingerprint: 31:ad:66:48:f8:10:41:38:c7:38:f3:9e:a4:32:01:33:39:3e:3a:18:cc:02:29:6e:f9:7c:2a:c9:ef:67:31:d0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Trusted Root G4 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Trusted Root G4 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Trusted Root G4"
+# Serial: 7451500558977370777930084869016614236
+# MD5 Fingerprint: 78:f2:fc:aa:60:1f:2f:b4:eb:c9:37:ba:53:2e:75:49
+# SHA1 Fingerprint: dd:fb:16:cd:49:31:c9:73:a2:03:7d:3f:c8:3a:4d:7d:77:5d:05:e4
+# SHA256 Fingerprint: 55:2f:7b:dc:f1:a7:af:9e:6c:e6:72:01:7f:4f:12:ab:f7:72:40:c7:8e:76:1a:c2:03:d1:d9:d2:0a:c8:99:88
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO RSA Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO RSA Certification Authority O=COMODO CA Limited
+# Label: "COMODO RSA Certification Authority"
+# Serial: 101909084537582093308941363524873193117
+# MD5 Fingerprint: 1b:31:b0:71:40:36:cc:14:36:91:ad:c4:3e:fd:ec:18
+# SHA1 Fingerprint: af:e5:d2:44:a8:d1:19:42:30:ff:47:9f:e2:f8:97:bb:cd:7a:8c:b4
+# SHA256 Fingerprint: 52:f0:e1:c4:e5:8e:c6:29:29:1b:60:31:7f:07:46:71:b8:5d:7e:a8:0d:5b:07:27:34:63:53:4b:32:b4:02:34
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=USERTrust RSA Certification Authority O=The USERTRUST Network
+# Subject: CN=USERTrust RSA Certification Authority O=The USERTRUST Network
+# Label: "USERTrust RSA Certification Authority"
+# Serial: 2645093764781058787591871645665788717
+# MD5 Fingerprint: 1b:fe:69:d1:91:b7:19:33:a3:72:a8:0f:e1:55:e5:b5
+# SHA1 Fingerprint: 2b:8f:1b:57:33:0d:bb:a2:d0:7a:6c:51:f7:0e:e9:0d:da:b9:ad:8e
+# SHA256 Fingerprint: e7:93:c9:b0:2f:d8:aa:13:e2:1c:31:22:8a:cc:b0:81:19:64:3b:74:9c:89:89:64:b1:74:6d:46:c3:d4:cb:d2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=USERTrust ECC Certification Authority O=The USERTRUST Network
+# Subject: CN=USERTrust ECC Certification Authority O=The USERTRUST Network
+# Label: "USERTrust ECC Certification Authority"
+# Serial: 123013823720199481456569720443997572134
+# MD5 Fingerprint: fa:68:bc:d9:b5:7f:ad:fd:c9:1d:06:83:28:cc:24:c1
+# SHA1 Fingerprint: d1:cb:ca:5d:b2:d5:2a:7f:69:3b:67:4d:e5:f0:5a:1d:0c:95:7d:f0
+# SHA256 Fingerprint: 4f:f4:60:d5:4b:9c:86:da:bf:bc:fc:57:12:e0:40:0d:2b:ed:3f:bc:4d:4f:bd:aa:86:e0:6a:dc:d2:a9:ad:7a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R5
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R5
+# Label: "GlobalSign ECC Root CA - R5"
+# Serial: 32785792099990507226680698011560947931244
+# MD5 Fingerprint: 9f:ad:3b:1c:02:1e:8a:ba:17:74:38:81:0c:a2:bc:08
+# SHA1 Fingerprint: 1f:24:c6:30:cd:a4:18:ef:20:69:ff:ad:4f:dd:5f:46:3a:1b:69:aa
+# SHA256 Fingerprint: 17:9f:bc:14:8a:3d:d0:0f:d2:4e:a1:34:58:cc:43:bf:a7:f5:9c:81:82:d7:83:a5:13:f6:eb:ec:10:0c:89:24
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=IdenTrust Commercial Root CA 1 O=IdenTrust
+# Subject: CN=IdenTrust Commercial Root CA 1 O=IdenTrust
+# Label: "IdenTrust Commercial Root CA 1"
+# Serial: 13298821034946342390520003877796839426
+# MD5 Fingerprint: b3:3e:77:73:75:ee:a0:d3:e3:7e:49:63:49:59:bb:c7
+# SHA1 Fingerprint: df:71:7e:aa:4a:d9:4e:c9:55:84:99:60:2d:48:de:5f:bc:f0:3a:25
+# SHA256 Fingerprint: 5d:56:49:9b:e4:d2:e0:8b:cf:ca:d0:8a:3e:38:72:3d:50:50:3b:de:70:69:48:e4:2f:55:60:30:19:e5:28:ae
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=IdenTrust Public Sector Root CA 1 O=IdenTrust
+# Subject: CN=IdenTrust Public Sector Root CA 1 O=IdenTrust
+# Label: "IdenTrust Public Sector Root CA 1"
+# Serial: 13298821034946342390521976156843933698
+# MD5 Fingerprint: 37:06:a5:b0:fc:89:9d:ba:f4:6b:8c:1a:64:cd:d5:ba
+# SHA1 Fingerprint: ba:29:41:60:77:98:3f:f4:f3:ef:f2:31:05:3b:2e:ea:6d:4d:45:fd
+# SHA256 Fingerprint: 30:d0:89:5a:9a:44:8a:26:20:91:63:55:22:d1:f5:20:10:b5:86:7a:ca:e1:2c:78:ef:95:8f:d4:f4:38:9f:2f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority - G2 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2009 Entrust, Inc. - for authorized use only
+# Subject: CN=Entrust Root Certification Authority - G2 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2009 Entrust, Inc. - for authorized use only
+# Label: "Entrust Root Certification Authority - G2"
+# Serial: 1246989352
+# MD5 Fingerprint: 4b:e2:c9:91:96:65:0c:f4:0e:5a:93:92:a0:0a:fe:b2
+# SHA1 Fingerprint: 8c:f4:27:fd:79:0c:3a:d1:66:06:8d:e8:1e:57:ef:bb:93:22:72:d4
+# SHA256 Fingerprint: 43:df:57:74:b0:3e:7f:ef:5f:e4:0d:93:1a:7b:ed:f1:bb:2e:6b:42:73:8c:4e:6d:38:41:10:3d:3a:a7:f3:39
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority - EC1 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2012 Entrust, Inc. - for authorized use only
+# Subject: CN=Entrust Root Certification Authority - EC1 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2012 Entrust, Inc. - for authorized use only
+# Label: "Entrust Root Certification Authority - EC1"
+# Serial: 51543124481930649114116133369
+# MD5 Fingerprint: b6:7e:1d:f0:58:c5:49:6c:24:3b:3d:ed:98:18:ed:bc
+# SHA1 Fingerprint: 20:d8:06:40:df:9b:25:f5:12:25:3a:11:ea:f7:59:8a:eb:14:b5:47
+# SHA256 Fingerprint: 02:ed:0e:b2:8c:14:da:45:16:5c:56:67:91:70:0d:64:51:d7:fb:56:f0:b2:ab:1d:3b:8e:b0:70:e5:6e:df:f5
+-----BEGIN CERTIFICATE-----
+MIIC+TCCAoCgAwIBAgINAKaLeSkAAAAAUNCR+TAKBggqhkjOPQQDAzCBvzELMAkG
+A1UEBhMCVVMxFjAUBgNVBAoTDUVudHJ1c3QsIEluYy4xKDAmBgNVBAsTH1NlZSB3
+d3cuZW50cnVzdC5uZXQvbGVnYWwtdGVybXMxOTA3BgNVBAsTMChjKSAyMDEyIEVu
+dHJ1c3QsIEluYy4gLSBmb3IgYXV0aG9yaXplZCB1c2Ugb25seTEzMDEGA1UEAxMq
+RW50cnVzdCBSb290IENlcnRpZmljYXRpb24gQXV0aG9yaXR5IC0gRUMxMB4XDTEy
+MTIxODE1MjUzNloXDTM3MTIxODE1NTUzNlowgb8xCzAJBgNVBAYTAlVTMRYwFAYD
+VQQKEw1FbnRydXN0LCBJbmMuMSgwJgYDVQQLEx9TZWUgd3d3LmVudHJ1c3QubmV0
+L2xlZ2FsLXRlcm1zMTkwNwYDVQQLEzAoYykgMjAxMiBFbnRydXN0LCBJbmMuIC0g
+Zm9yIGF1dGhvcml6ZWQgdXNlIG9ubHkxMzAxBgNVBAMTKkVudHJ1c3QgUm9vdCBD
+ZXJ0aWZpY2F0aW9uIEF1dGhvcml0eSAtIEVDMTB2MBAGByqGSM49AgEGBSuBBAAi
+A2IABIQTydC6bUF74mzQ61VfZgIaJPRbiWlH47jCffHyAsWfoPZb1YsGGYZPUxBt
+ByQnoaD41UcZYUx9ypMn6nQM72+WCf5j7HBdNq1nd67JnXxVRDqiY1Ef9eNi1KlH
+Bz7MIKNCMEAwDgYDVR0PAQH/BAQDAgEGMA8GA1UdEwEB/wQFMAMBAf8wHQYDVR0O
+BBYEFLdj5xrdjekIplWDpOBqUEFlEUJJMAoGCCqGSM49BAMDA2cAMGQCMGF52OVC
+R98crlOZF7ZvHH3hvxGU0QOIdeSNiaSKd0bebWHvAvX7td/M/k7//qnmpwIwW5nX
+hTcGtXsI/esni0qU+eH6p44mCOh8kmhtc9hvJqwhAriZtyZBWyVgrtBIGu4G
+-----END CERTIFICATE-----
+
+# Issuer: CN=CFCA EV ROOT O=China Financial Certification Authority
+# Subject: CN=CFCA EV ROOT O=China Financial Certification Authority
+# Label: "CFCA EV ROOT"
+# Serial: 407555286
+# MD5 Fingerprint: 74:e1:b6:ed:26:7a:7a:44:30:33:94:ab:7b:27:81:30
+# SHA1 Fingerprint: e2:b8:29:4b:55:84:ab:6b:58:c2:90:46:6c:ac:3f:b8:39:8f:84:83
+# SHA256 Fingerprint: 5c:c3:d7:8e:4e:1d:5e:45:54:7a:04:e6:87:3e:64:f9:0c:f9:53:6d:1c:cc:2e:f8:00:f3:55:c4:c5:fd:70:fd
+-----BEGIN CERTIFICATE-----
+MIIFjTCCA3WgAwIBAgIEGErM1jANBgkqhkiG9w0BAQsFADBWMQswCQYDVQQGEwJD
+TjEwMC4GA1UECgwnQ2hpbmEgRmluYW5jaWFsIENlcnRpZmljYXRpb24gQXV0aG9y
+aXR5MRUwEwYDVQQDDAxDRkNBIEVWIFJPT1QwHhcNMTIwODA4MDMwNzAxWhcNMjkx
+MjMxMDMwNzAxWjBWMQswCQYDVQQGEwJDTjEwMC4GA1UECgwnQ2hpbmEgRmluYW5j
+aWFsIENlcnRpZmljYXRpb24gQXV0aG9yaXR5MRUwEwYDVQQDDAxDRkNBIEVWIFJP
+T1QwggIiMA0GCSqGSIb3DQEBAQUAA4ICDwAwggIKAoICAQDXXWvNED8fBVnVBU03
+sQ7smCuOFR36k0sXgiFxEFLXUWRwFsJVaU2OFW2fvwwbwuCjZ9YMrM8irq93VCpL
+TIpTUnrD7i7es3ElweldPe6hL6P3KjzJIx1qqx2hp/Hz7KDVRM8Vz3IvHWOX6Jn5
+/ZOkVIBMUtRSqy5J35DNuF++P96hyk0g1CXohClTt7GIH//62pCfCqktQT+x8Rgp
+7hZZLDRJGqgG16iI0gNyejLi6mhNbiyWZXvKWfry4t3uMCz7zEasxGPrb382KzRz
+EpR/38wmnvFyXVBlWY9ps4deMm/DGIq1lY+wejfeWkU7xzbh72fROdOXW3NiGUgt
+hxwG+3SYIElz8AXSG7Ggo7cbcNOIabla1jj0Ytwli3i/+Oh+uFzJlU9fpy25IGvP
+a931DfSCt/SyZi4QKPaXWnuWFo8BGS1sbn85WAZkgwGDg8NNkt0yxoekN+kWzqot
+aK8KgWU6cMGbrU1tVMoqLUuFG7OA5nBFDWteNfB/O7ic5ARwiRIlk9oKmSJgamNg
+TnYGmE69g60dWIolhdLHZR4tjsbftsbhf4oEIRUpdPA+nJCdDC7xij5aqgwJHsfV
+PKPtl8MeNPo4+QgO48BdK4PRVmrJtqhUUy54Mmc9gn900PvhtgVguXDbjgv5E1hv
+cWAQUhC5wUEJ73IfZzF4/5YFjQIDAQABo2MwYTAfBgNVHSMEGDAWgBTj/i39KNAL
+tbq2osS/BqoFjJP7LzAPBgNVHRMBAf8EBTADAQH/MA4GA1UdDwEB/wQEAwIBBjAd
+BgNVHQ4EFgQU4/4t/SjQC7W6tqLEvwaqBYyT+y8wDQYJKoZIhvcNAQELBQADggIB
+ACXGumvrh8vegjmWPfBEp2uEcwPenStPuiB/vHiyz5ewG5zz13ku9Ui20vsXiObT
+ej/tUxPQ4i9qecsAIyjmHjdXNYmEwnZPNDatZ8POQQaIxffu2Bq41gt/UP+TqhdL
+jOztUmCypAbqTuv0axn96/Ua4CUqmtzHQTb3yHQFhDmVOdYLO6Qn+gjYXB74BGBS
+ESgoA//vU2YApUo0FmZ8/Qmkrp5nGm9BC2sGE5uPhnEFtC+NiWYzKXZUmhH4J/qy
+P5Hgzg0b8zAarb8iXRvTvyUFTeGSGn+ZnzxEk8rUQElsgIfXBDrDMlI1Dlb4pd19
+xIsNER9Tyx6yF7Zod1rg1MvIB671Oi6ON7fQAUtDKXeMOZePglr4UeWJoBjnaH9d
+Ci77o0cOPaYjesYBx4/IXr9tgFa+iiS6M+qf4TIRnvHST4D2G0CvOJ4RUHlzEhLN
+5mydLIhyPDCBBpEi6lmt2hkuIsKNuYyH4Ga8cyNfIWRjgEj1oDwYPZTISEEdQLpe
+/v5WOaHIz16eGWRGENoXkbcFgKyLmZJ956LYBws2J+dIeWCKw9cTXPhyQN9Ky8+Z
+AAoACxGV2lZFA4gKn2fQ1XmxqI1AbQ3CekD6819kR5LLU7m7Wc5P/dAVUwHY3+vZ
+5nbv0CO7O6l5s9UCKc2Jo5YPSjXnTkLAdc0Hz+Ys63su
+-----END CERTIFICATE-----
+
+# Issuer: CN=OISTE WISeKey Global Root GB CA O=WISeKey OU=OISTE Foundation Endorsed
+# Subject: CN=OISTE WISeKey Global Root GB CA O=WISeKey OU=OISTE Foundation Endorsed
+# Label: "OISTE WISeKey Global Root GB CA"
+# Serial: 157768595616588414422159278966750757568
+# MD5 Fingerprint: a4:eb:b9:61:28:2e:b7:2f:98:b0:35:26:90:99:51:1d
+# SHA1 Fingerprint: 0f:f9:40:76:18:d3:d7:6a:4b:98:f0:a8:35:9e:0c:fd:27:ac:cc:ed
+# SHA256 Fingerprint: 6b:9c:08:e8:6e:b0:f7:67:cf:ad:65:cd:98:b6:21:49:e5:49:4a:67:f5:84:5e:7b:d1:ed:01:9f:27:b8:6b:d6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SZAFIR ROOT CA2 O=Krajowa Izba Rozliczeniowa S.A.
+# Subject: CN=SZAFIR ROOT CA2 O=Krajowa Izba Rozliczeniowa S.A.
+# Label: "SZAFIR ROOT CA2"
+# Serial: 357043034767186914217277344587386743377558296292
+# MD5 Fingerprint: 11:64:c1:89:b0:24:b1:8c:b1:07:7e:89:9e:51:9e:99
+# SHA1 Fingerprint: e2:52:fa:95:3f:ed:db:24:60:bd:6e:28:f3:9c:cc:cf:5e:b3:3f:de
+# SHA256 Fingerprint: a1:33:9d:33:28:1a:0b:56:e5:57:d3:d3:2b:1c:e7:f9:36:7e:b0:94:bd:5f:a7:2a:7e:50:04:c8:de:d7:ca:fe
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Network CA 2 O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Network CA 2 O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Network CA 2"
+# Serial: 44979900017204383099463764357512596969
+# MD5 Fingerprint: 6d:46:9e:d9:25:6d:08:23:5b:5e:74:7d:1e:27:db:f2
+# SHA1 Fingerprint: d3:dd:48:3e:2b:bf:4c:05:e8:af:10:f5:fa:76:26:cf:d3:dc:30:92
+# SHA256 Fingerprint: b6:76:f2:ed:da:e8:77:5c:d3:6c:b0:f6:3c:d1:d4:60:39:61:f4:9e:62:65:ba:01:3a:2f:03:07:b6:d0:b8:04
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hellenic Academic and Research Institutions RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Subject: CN=Hellenic Academic and Research Institutions RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Label: "Hellenic Academic and Research Institutions RootCA 2015"
+# Serial: 0
+# MD5 Fingerprint: ca:ff:e2:db:03:d9:cb:4b:e9:0f:ad:84:fd:7b:18:ce
+# SHA1 Fingerprint: 01:0c:06:95:a6:98:19:14:ff:bf:5f:c6:b0:b6:95:ea:29:e9:12:a6
+# SHA256 Fingerprint: a0:40:92:9a:02:ce:53:b4:ac:f4:f2:ff:c6:98:1c:e4:49:6f:75:5e:6d:45:fe:0b:2a:69:2b:cd:52:52:3f:36
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hellenic Academic and Research Institutions ECC RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Subject: CN=Hellenic Academic and Research Institutions ECC RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Label: "Hellenic Academic and Research Institutions ECC RootCA 2015"
+# Serial: 0
+# MD5 Fingerprint: 81:e5:b4:17:eb:c2:f5:e1:4b:0d:41:7b:49:92:fe:ef
+# SHA1 Fingerprint: 9f:f1:71:8d:92:d5:9a:f3:7d:74:97:b4:bc:6f:84:68:0b:ba:b6:66
+# SHA256 Fingerprint: 44:b5:45:aa:8a:25:e6:5a:73:ca:15:dc:27:fc:36:d2:4c:1c:b9:95:3a:06:65:39:b1:15:82:dc:48:7b:48:33
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=ISRG Root X1 O=Internet Security Research Group
+# Subject: CN=ISRG Root X1 O=Internet Security Research Group
+# Label: "ISRG Root X1"
+# Serial: 172886928669790476064670243504169061120
+# MD5 Fingerprint: 0c:d2:f9:e0:da:17:73:e9:ed:86:4d:a5:e3:70:e7:4e
+# SHA1 Fingerprint: ca:bd:2a:79:a1:07:6a:31:f2:1d:25:36:35:cb:03:9d:43:29:a5:e8
+# SHA256 Fingerprint: 96:bc:ec:06:26:49:76:f3:74:60:77:9a:cf:28:c5:a7:cf:e8:a3:c0:aa:e1:1a:8f:fc:ee:05:c0:bd:df:08:c6
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: O=FNMT-RCM OU=AC RAIZ FNMT-RCM
+# Subject: O=FNMT-RCM OU=AC RAIZ FNMT-RCM
+# Label: "AC RAIZ FNMT-RCM"
+# Serial: 485876308206448804701554682760554759
+# MD5 Fingerprint: e2:09:04:b4:d3:bd:d1:a0:14:fd:1a:d2:47:c4:57:1d
+# SHA1 Fingerprint: ec:50:35:07:b2:15:c4:95:62:19:e2:a8:9a:5b:42:99:2c:4c:2c:20
+# SHA256 Fingerprint: eb:c5:57:0c:29:01:8c:4d:67:b1:aa:12:7b:af:12:f7:03:b4:61:1e:bc:17:b7:da:b5:57:38:94:17:9b:93:fa
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 1 O=Amazon
+# Subject: CN=Amazon Root CA 1 O=Amazon
+# Label: "Amazon Root CA 1"
+# Serial: 143266978916655856878034712317230054538369994
+# MD5 Fingerprint: 43:c6:bf:ae:ec:fe:ad:2f:18:c6:88:68:30:fc:c8:e6
+# SHA1 Fingerprint: 8d:a7:f9:65:ec:5e:fc:37:91:0f:1c:6e:59:fd:c1:cc:6a:6e:de:16
+# SHA256 Fingerprint: 8e:cd:e6:88:4f:3d:87:b1:12:5b:a3:1a:c3:fc:b1:3d:70:16:de:7f:57:cc:90:4f:e1:cb:97:c6:ae:98:19:6e
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 2 O=Amazon
+# Subject: CN=Amazon Root CA 2 O=Amazon
+# Label: "Amazon Root CA 2"
+# Serial: 143266982885963551818349160658925006970653239
+# MD5 Fingerprint: c8:e5:8d:ce:a8:42:e2:7a:c0:2a:5c:7c:9e:26:bf:66
+# SHA1 Fingerprint: 5a:8c:ef:45:d7:a6:98:59:76:7a:8c:8b:44:96:b5:78:cf:47:4b:1a
+# SHA256 Fingerprint: 1b:a5:b2:aa:8c:65:40:1a:82:96:01:18:f8:0b:ec:4f:62:30:4d:83:ce:c4:71:3a:19:c3:9c:01:1e:a4:6d:b4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 3 O=Amazon
+# Subject: CN=Amazon Root CA 3 O=Amazon
+# Label: "Amazon Root CA 3"
+# Serial: 143266986699090766294700635381230934788665930
+# MD5 Fingerprint: a0:d4:ef:0b:f7:b5:d8:49:95:2a:ec:f5:c4:fc:81:87
+# SHA1 Fingerprint: 0d:44:dd:8c:3c:8c:1a:1a:58:75:64:81:e9:0f:2e:2a:ff:b3:d2:6e
+# SHA256 Fingerprint: 18:ce:6c:fe:7b:f1:4e:60:b2:e3:47:b8:df:e8:68:cb:31:d0:2e:bb:3a:da:27:15:69:f5:03:43:b4:6d:b3:a4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 4 O=Amazon
+# Subject: CN=Amazon Root CA 4 O=Amazon
+# Label: "Amazon Root CA 4"
+# Serial: 143266989758080763974105200630763877849284878
+# MD5 Fingerprint: 89:bc:27:d5:eb:17:8d:06:6a:69:d5:fd:89:47:b4:cd
+# SHA1 Fingerprint: f6:10:84:07:d6:f8:bb:67:98:0c:c2:e2:44:c2:eb:ae:1c:ef:63:be
+# SHA256 Fingerprint: e3:5d:28:41:9e:d0:20:25:cf:a6:90:38:cd:62:39:62:45:8d:a5:c6:95:fb:de:a3:c2:2b:0b:fb:25:89:70:92
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1 O=Turkiye Bilimsel ve Teknolojik Arastirma Kurumu - TUBITAK OU=Kamu Sertifikasyon Merkezi - Kamu SM
+# Subject: CN=TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1 O=Turkiye Bilimsel ve Teknolojik Arastirma Kurumu - TUBITAK OU=Kamu Sertifikasyon Merkezi - Kamu SM
+# Label: "TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1"
+# Serial: 1
+# MD5 Fingerprint: dc:00:81:dc:69:2f:3e:2f:b0:3b:f6:3d:5a:91:8e:49
+# SHA1 Fingerprint: 31:43:64:9b:ec:ce:27:ec:ed:3a:3f:0b:8f:0d:e4:e8:91:dd:ee:ca
+# SHA256 Fingerprint: 46:ed:c3:68:90:46:d5:3a:45:3f:b3:10:4a:b8:0d:ca:ec:65:8b:26:60:ea:16:29:dd:7e:86:79:90:64:87:16
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GDCA TrustAUTH R5 ROOT O=GUANG DONG CERTIFICATE AUTHORITY CO.,LTD.
+# Subject: CN=GDCA TrustAUTH R5 ROOT O=GUANG DONG CERTIFICATE AUTHORITY CO.,LTD.
+# Label: "GDCA TrustAUTH R5 ROOT"
+# Serial: 9009899650740120186
+# MD5 Fingerprint: 63:cc:d9:3d:34:35:5c:6f:53:a3:e2:08:70:48:1f:b4
+# SHA1 Fingerprint: 0f:36:38:5b:81:1a:25:c3:9b:31:4e:83:ca:e9:34:66:70:cc:74:b4
+# SHA256 Fingerprint: bf:ff:8f:d0:44:33:48:7d:6a:8a:a6:0c:1a:29:76:7a:9f:c2:bb:b0:5e:42:0f:71:3a:13:b9:92:89:1d:38:93
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com Root Certification Authority RSA O=SSL Corporation
+# Subject: CN=SSL.com Root Certification Authority RSA O=SSL Corporation
+# Label: "SSL.com Root Certification Authority RSA"
+# Serial: 8875640296558310041
+# MD5 Fingerprint: 86:69:12:c0:70:f1:ec:ac:ac:c2:d5:bc:a5:5b:a1:29
+# SHA1 Fingerprint: b7:ab:33:08:d1:ea:44:77:ba:14:80:12:5a:6f:bd:a9:36:49:0c:bb
+# SHA256 Fingerprint: 85:66:6a:56:2e:e0:be:5c:e9:25:c1:d8:89:0a:6f:76:a8:7e:c1:6d:4d:7d:5f:29:ea:74:19:cf:20:12:3b:69
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com Root Certification Authority ECC O=SSL Corporation
+# Subject: CN=SSL.com Root Certification Authority ECC O=SSL Corporation
+# Label: "SSL.com Root Certification Authority ECC"
+# Serial: 8495723813297216424
+# MD5 Fingerprint: 2e:da:e4:39:7f:9c:8f:37:d1:70:9f:26:17:51:3a:8e
+# SHA1 Fingerprint: c3:19:7c:39:24:e6:54:af:1b:c4:ab:20:95:7a:e2:c3:0e:13:02:6a
+# SHA256 Fingerprint: 34:17:bb:06:cc:60:07:da:1b:96:1c:92:0b:8a:b4:ce:3f:ad:82:0e:4a:a3:0b:9a:cb:c4:a7:4e:bd:ce:bc:65
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com EV Root Certification Authority RSA R2 O=SSL Corporation
+# Subject: CN=SSL.com EV Root Certification Authority RSA R2 O=SSL Corporation
+# Label: "SSL.com EV Root Certification Authority RSA R2"
+# Serial: 6248227494352943350
+# MD5 Fingerprint: e1:1e:31:58:1a:ae:54:53:02:f6:17:6a:11:7b:4d:95
+# SHA1 Fingerprint: 74:3a:f0:52:9b:d0:32:a0:f4:4a:83:cd:d4:ba:a9:7b:7c:2e:c4:9a
+# SHA256 Fingerprint: 2e:7b:f1:6c:c2:24:85:a7:bb:e2:aa:86:96:75:07:61:b0:ae:39:be:3b:2f:e9:d0:cc:6d:4e:f7:34:91:42:5c
+-----BEGIN CERTIFICATE-----
+MIIF6zCCA9OgAwIBAgIIVrYpzTS8ePYwDQYJKoZIhvcNAQELBQAwgYIxCzAJBgNV
+BAYTAlVTMQ4wDAYDVQQIDAVUZXhhczEQMA4GA1UEBwwHSG91c3RvbjEYMBYGA1UE
+CgwPU1NMIENvcnBvcmF0aW9uMTcwNQYDVQQDDC5TU0wuY29tIEVWIFJvb3QgQ2Vy
+dGlmaWNhdGlvbiBBdXRob3JpdHkgUlNBIFIyMB4XDTE3MDUzMTE4MTQzN1oXDTQy
+MDUzMDE4MTQzN1owgYIxCzAJBgNVBAYTAlVTMQ4wDAYDVQQIDAVUZXhhczEQMA4G
+A1UEBwwHSG91c3RvbjEYMBYGA1UECgwPU1NMIENvcnBvcmF0aW9uMTcwNQYDVQQD
+DC5TU0wuY29tIEVWIFJvb3QgQ2VydGlmaWNhdGlvbiBBdXRob3JpdHkgUlNBIFIy
+MIICIjANBgkqhkiG9w0BAQEFAAOCAg8AMIICCgKCAgEAjzZlQOHWTcDXtOlG2mvq
+M0fNTPl9fb69LT3w23jhhqXZuglXaO1XPqDQCEGD5yhBJB/jchXQARr7XnAjssuf
+OePPxU7Gkm0mxnu7s9onnQqG6YE3Bf7wcXHswxzpY6IXFJ3vG2fThVUCAtZJycxa
+4bH3bzKfydQ7iEGonL3Lq9ttewkfokxykNorCPzPPFTOZw+oz12WGQvE43LrrdF9
+HSfvkusQv1vrO6/PgN3B0pYEW3p+pKk8OHakYo6gOV7qd89dAFmPZiw+B6KjBSYR
+aZfqhbcPlgtLyEDhULouisv3D5oi53+aNxPN8k0TayHRwMwi8qFG9kRpnMphNQcA
+b9ZhCBHqurj26bNg5U257J8UZslXWNvNh2n4ioYSA0e/ZhN2rHd9NCSFg83XqpyQ
+Gp8hLH94t2S42Oim9HizVcuE0jLEeK6jj2HdzghTreyI/BXkmg3mnxp3zkyPuBQV
+PWKchjgGAGYS5Fl2WlPAApiiECtoRHuOec4zSnaqW4EWG7WK2NAAe15itAnWhmMO
+pgWVSbooi4iTsjQc2KRVbrcc0N6ZVTsj9CLg+SlmJuwgUHfbSguPvuUCYHBBXtSu
+UDkiFCbLsjtzdFVHB3mBOagwE0TlBIqulhMlQg+5U8Sb/M3kHN48+qvWBkofZ6aY
+MBzdLNvcGJVXZsb/XItW9XcCAwEAAaNjMGEwDwYDVR0TAQH/BAUwAwEB/zAfBgNV
+HSMEGDAWgBT5YLvU49U09rj1BoAlp3PbRmmonjAdBgNVHQ4EFgQU+WC71OPVNPa4
+9QaAJadz20ZpqJ4wDgYDVR0PAQH/BAQDAgGGMA0GCSqGSIb3DQEBCwUAA4ICAQBW
+s47LCp1Jjr+kxJG7ZhcFUZh1++VQLHqe8RT6q9OKPv+RKY9ji9i0qVQBDb6Thi/5
+Sm3HXvVX+cpVHBK+Rw82xd9qt9t1wkclf7nxY/hoLVUE0fKNsKTPvDxeH3jnpaAg
+cLAExbf3cqfeIg29MyVGjGSSJuM+LmOW2puMPfgYCdcDzH2GguDKBAdRUNf/ktUM
+79qGn5nX67evaOI5JpS6aLe/g9Pqemc9YmeuJeVy6OLk7K4S9ksrPJ/psEDzOFSz
+/bdoyNrGj1E8svuR3Bznm53htw1yj+KkxKl4+esUrMZDBcJlOSgYAsOCsp0FvmXt
+ll9ldDz7CTUue5wT/RsPXcdtgTpWD8w74a8CLyKsRspGPKAcTNZEtF4uXBVmCeEm
+Kf7GUmG6sXP/wwyc5WxqlD8UykAWlYTzWamsX0xhk23RO8yilQwipmdnRC652dKK
+QbNmC1r7fSOl8hqw/96bg5Qu0T/fkreRrwU7ZcegbLHNYhLDkBvjJc40vG93drEQ
+w/cFGsDWr3RiSBd3kmmQYRzelYB0VI8YHMPzA9C/pEN1hlMYegouCRw2n5H9gooi
+S9EOUCXdywMMF8mDAAhONU2Ki+3wApRmLER/y5UnlhetCTCstnEXbosX9hwJ1C07
+mKVx01QT2WDz9UtmT/rx7iASjbSsV7FFY6GsdqnC+w==
+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com EV Root Certification Authority ECC O=SSL Corporation
+# Subject: CN=SSL.com EV Root Certification Authority ECC O=SSL Corporation
+# Label: "SSL.com EV Root Certification Authority ECC"
+# Serial: 3182246526754555285
+# MD5 Fingerprint: 59:53:22:65:83:42:01:54:c0:ce:42:b9:5a:7c:f2:90
+# SHA1 Fingerprint: 4c:dd:51:a3:d1:f5:20:32:14:b0:c6:c5:32:23:03:91:c7:46:42:6d
+# SHA256 Fingerprint: 22:a2:c1:f7:bd:ed:70:4c:c1:e7:01:b5:f4:08:c3:10:88:0f:e9:56:b5:de:2a:4a:44:f9:9c:87:3a:25:a7:c8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R6
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R6
+# Label: "GlobalSign Root CA - R6"
+# Serial: 1417766617973444989252670301619537
+# MD5 Fingerprint: 4f:dd:07:e4:d4:22:64:39:1e:0c:37:42:ea:d1:c6:ae
+# SHA1 Fingerprint: 80:94:64:0e:b5:a7:a1:ca:11:9c:1f:dd:d5:9f:81:02:63:a7:fb:d1
+# SHA256 Fingerprint: 2c:ab:ea:fe:37:d0:6c:a2:2a:ba:73:91:c0:03:3d:25:98:29:52:c4:53:64:73:49:76:3a:3a:b5:ad:6c:cf:69
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=OISTE WISeKey Global Root GC CA O=WISeKey OU=OISTE Foundation Endorsed
+# Subject: CN=OISTE WISeKey Global Root GC CA O=WISeKey OU=OISTE Foundation Endorsed
+# Label: "OISTE WISeKey Global Root GC CA"
+# Serial: 44084345621038548146064804565436152554
+# MD5 Fingerprint: a9:d6:b9:2d:2f:93:64:f8:a5:69:ca:91:e9:68:07:23
+# SHA1 Fingerprint: e0:11:84:5e:34:de:be:88:81:b9:9c:f6:16:26:d1:96:1f:c3:b9:31
+# SHA256 Fingerprint: 85:60:f9:1c:36:24:da:ba:95:70:b5:fe:a0:db:e3:6f:f1:1a:83:23:be:94:86:85:4f:b3:f3:4a:55:71:19:8d
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=UCA Global G2 Root O=UniTrust
+# Subject: CN=UCA Global G2 Root O=UniTrust
+# Label: "UCA Global G2 Root"
+# Serial: 124779693093741543919145257850076631279
+# MD5 Fingerprint: 80:fe:f0:c4:4a:f0:5c:62:32:9f:1c:ba:78:a9:50:f8
+# SHA1 Fingerprint: 28:f9:78:16:19:7a:ff:18:25:18:aa:44:fe:c1:a0:ce:5c:b6:4c:8a
+# SHA256 Fingerprint: 9b:ea:11:c9:76:fe:01:47:64:c1:be:56:a6:f9:14:b5:a5:60:31:7a:bd:99:88:39:33:82:e5:16:1a:a0:49:3c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=UCA Extended Validation Root O=UniTrust
+# Subject: CN=UCA Extended Validation Root O=UniTrust
+# Label: "UCA Extended Validation Root"
+# Serial: 106100277556486529736699587978573607008
+# MD5 Fingerprint: a1:f3:5f:43:c6:34:9b:da:bf:8c:7e:05:53:ad:96:e2
+# SHA1 Fingerprint: a3:a1:b0:6f:24:61:23:4a:e3:36:a5:c2:37:fc:a6:ff:dd:f0:d7:3a
+# SHA256 Fingerprint: d4:3a:f9:b3:54:73:75:5c:96:84:fc:06:d7:d8:cb:70:ee:5c:28:e7:73:fb:29:4e:b4:1e:e7:17:22:92:4d:24
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certigna Root CA O=Dhimyotis OU=0002 48146308100036
+# Subject: CN=Certigna Root CA O=Dhimyotis OU=0002 48146308100036
+# Label: "Certigna Root CA"
+# Serial: 269714418870597844693661054334862075617
+# MD5 Fingerprint: 0e:5c:30:62:27:eb:5b:bc:d7:ae:62:ba:e9:d5:df:77
+# SHA1 Fingerprint: 2d:0d:52:14:ff:9e:ad:99:24:01:74:20:47:6e:6c:85:27:27:f5:43
+# SHA256 Fingerprint: d4:8d:3d:23:ee:db:50:a4:59:e5:51:97:60:1c:27:77:4b:9d:7b:18:c9:4d:5a:05:95:11:a1:02:50:b9:31:68
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign Root CA - G1 O=eMudhra Technologies Limited OU=emSign PKI
+# Subject: CN=emSign Root CA - G1 O=eMudhra Technologies Limited OU=emSign PKI
+# Label: "emSign Root CA - G1"
+# Serial: 235931866688319308814040
+# MD5 Fingerprint: 9c:42:84:57:dd:cb:0b:a7:2e:95:ad:b6:f3:da:bc:ac
+# SHA1 Fingerprint: 8a:c7:ad:8f:73:ac:4e:c1:b5:75:4d:a5:40:f4:fc:cf:7c:b5:8e:8c
+# SHA256 Fingerprint: 40:f6:af:03:46:a9:9a:a1:cd:1d:55:5a:4e:9c:ce:62:c7:f9:63:46:03:ee:40:66:15:83:3d:c8:c8:d0:03:67
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign ECC Root CA - G3 O=eMudhra Technologies Limited OU=emSign PKI
+# Subject: CN=emSign ECC Root CA - G3 O=eMudhra Technologies Limited OU=emSign PKI
+# Label: "emSign ECC Root CA - G3"
+# Serial: 287880440101571086945156
+# MD5 Fingerprint: ce:0b:72:d1:9f:88:8e:d0:50:03:e8:e3:b8:8b:67:40
+# SHA1 Fingerprint: 30:43:fa:4f:f2:57:dc:a0:c3:80:ee:2e:58:ea:78:b2:3f:e6:bb:c1
+# SHA256 Fingerprint: 86:a1:ec:ba:08:9c:4a:8d:3b:be:27:34:c6:12:ba:34:1d:81:3e:04:3c:f9:e8:a8:62:cd:5c:57:a3:6b:be:6b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign Root CA - C1 O=eMudhra Inc OU=emSign PKI
+# Subject: CN=emSign Root CA - C1 O=eMudhra Inc OU=emSign PKI
+# Label: "emSign Root CA - C1"
+# Serial: 825510296613316004955058
+# MD5 Fingerprint: d8:e3:5d:01:21:fa:78:5a:b0:df:ba:d2:ee:2a:5f:68
+# SHA1 Fingerprint: e7:2e:f1:df:fc:b2:09:28:cf:5d:d4:d5:67:37:b1:51:cb:86:4f:01
+# SHA256 Fingerprint: 12:56:09:aa:30:1d:a0:a2:49:b9:7a:82:39:cb:6a:34:21:6f:44:dc:ac:9f:39:54:b1:42:92:f2:e8:c8:60:8f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign ECC Root CA - C3 O=eMudhra Inc OU=emSign PKI
+# Subject: CN=emSign ECC Root CA - C3 O=eMudhra Inc OU=emSign PKI
+# Label: "emSign ECC Root CA - C3"
+# Serial: 582948710642506000014504
+# MD5 Fingerprint: 3e:53:b3:a3:81:ee:d7:10:f8:d3:b0:1d:17:92:f5:d5
+# SHA1 Fingerprint: b6:af:43:c2:9b:81:53:7d:f6:ef:6b:c3:1f:1f:60:15:0c:ee:48:66
+# SHA256 Fingerprint: bc:4d:80:9b:15:18:9d:78:db:3e:1d:8c:f4:f9:72:6a:79:5d:a1:64:3c:a5:f1:35:8e:1d:db:0e:dc:0d:7e:b3
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hongkong Post Root CA 3 O=Hongkong Post
+# Subject: CN=Hongkong Post Root CA 3 O=Hongkong Post
+# Label: "Hongkong Post Root CA 3"
+# Serial: 46170865288971385588281144162979347873371282084
+# MD5 Fingerprint: 11:fc:9f:bd:73:30:02:8a:fd:3f:f3:58:b9:cb:20:f0
+# SHA1 Fingerprint: 58:a2:d0:ec:20:52:81:5b:c1:f3:f8:64:02:24:4e:c2:8e:02:4b:02
+# SHA256 Fingerprint: 5a:2f:c0:3f:0c:83:b0:90:bb:fa:40:60:4b:09:88:44:6c:76:36:18:3d:f9:84:6e:17:10:1a:44:7f:b8:ef:d6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsoft ECC Root Certificate Authority 2017 O=Microsoft Corporation
+# Subject: CN=Microsoft ECC Root Certificate Authority 2017 O=Microsoft Corporation
+# Label: "Microsoft ECC Root Certificate Authority 2017"
+# Serial: 136839042543790627607696632466672567020
+# MD5 Fingerprint: dd:a1:03:e6:4a:93:10:d1:bf:f0:19:42:cb:fe:ed:67
+# SHA1 Fingerprint: 99:9a:64:c3:7f:f4:7d:9f:ab:95:f1:47:69:89:14:60:ee:c4:c3:c5
+# SHA256 Fingerprint: 35:8d:f3:9d:76:4a:f9:e1:b7:66:e9:c9:72:df:35:2e:e1:5c:fa:c2:27:af:6a:d1:d7:0e:8e:4a:6e:dc:ba:02
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsoft RSA Root Certificate Authority 2017 O=Microsoft Corporation
+# Subject: CN=Microsoft RSA Root Certificate Authority 2017 O=Microsoft Corporation
+# Label: "Microsoft RSA Root Certificate Authority 2017"
+# Serial: 40975477897264996090493496164228220339
+# MD5 Fingerprint: 10:ff:00:ff:cf:c9:f8:c7:7a:c0:ee:35:8e:c9:0f:47
+# SHA1 Fingerprint: 73:a5:e6:4a:3b:ff:83:16:ff:0e:dc:cc:61:8a:90:6e:4e:ae:4d:74
+# SHA256 Fingerprint: c7:41:f7:0f:4b:2a:8d:88:bf:2e:71:c1:41:22:ef:53:ef:10:eb:a0:cf:a5:e6:4c:fa:20:f4:18:85:30:73:e0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=e-Szigno Root CA 2017 O=Microsec Ltd.
+# Subject: CN=e-Szigno Root CA 2017 O=Microsec Ltd.
+# Label: "e-Szigno Root CA 2017"
+# Serial: 411379200276854331539784714
+# MD5 Fingerprint: de:1f:f6:9e:84:ae:a7:b4:21:ce:1e:58:7d:d1:84:98
+# SHA1 Fingerprint: 89:d4:83:03:4f:9e:9a:48:80:5f:72:37:d4:a9:a6:ef:cb:7c:1f:d1
+# SHA256 Fingerprint: be:b0:0b:30:83:9b:9b:c3:2c:32:e4:44:79:05:95:06:41:f2:64:21:b1:5e:d0:89:19:8b:51:8a:e2:ea:1b:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=CERTSIGN SA OU=certSIGN ROOT CA G2
+# Subject: O=CERTSIGN SA OU=certSIGN ROOT CA G2
+# Label: "certSIGN Root CA G2"
+# Serial: 313609486401300475190
+# MD5 Fingerprint: 8c:f1:75:8a:c6:19:cf:94:b7:f7:65:20:87:c3:97:c7
+# SHA1 Fingerprint: 26:f9:93:b4:ed:3d:28:27:b0:b9:4b:a7:e9:15:1d:a3:8d:92:e5:32
+# SHA256 Fingerprint: 65:7c:fe:2f:a7:3f:aa:38:46:25:71:f3:32:a2:36:3a:46:fc:e7:02:09:51:71:07:02:cd:fb:b6:ee:da:33:05
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global Certification Authority"
+# Serial: 1846098327275375458322922162
+# MD5 Fingerprint: f8:1c:18:2d:2f:ba:5f:6d:a1:6c:bc:c7:ab:91:c7:0e
+# SHA1 Fingerprint: 2f:8f:36:4f:e1:58:97:44:21:59:87:a5:2a:9a:d0:69:95:26:7f:b5
+# SHA256 Fingerprint: 97:55:20:15:f5:dd:fc:3c:87:88:c0:06:94:45:55:40:88:94:45:00:84:f1:00:86:70:86:bc:1a:2b:b5:8d:c8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global ECC P256 Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global ECC P256 Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global ECC P256 Certification Authority"
+# Serial: 4151900041497450638097112925
+# MD5 Fingerprint: 5b:44:e3:8d:5d:36:86:26:e8:0d:05:d2:59:a7:83:54
+# SHA1 Fingerprint: b4:90:82:dd:45:0c:be:8b:5b:b1:66:d3:e2:a4:08:26:cd:ed:42:cf
+# SHA256 Fingerprint: 94:5b:bc:82:5e:a5:54:f4:89:d1:fd:51:a7:3d:df:2e:a6:24:ac:70:19:a0:52:05:22:5c:22:a7:8c:cf:a8:b4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global ECC P384 Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global ECC P384 Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global ECC P384 Certification Authority"
+# Serial: 2704997926503831671788816187
+# MD5 Fingerprint: ea:cf:60:c4:3b:b9:15:29:40:a1:97:ed:78:27:93:d6
+# SHA1 Fingerprint: e7:f3:a3:c8:cf:6f:c3:04:2e:6d:0e:67:32:c5:9e:68:95:0d:5e:d2
+# SHA256 Fingerprint: 55:90:38:59:c8:c0:c3:eb:b8:75:9e:ce:4e:25:57:22:5f:f5:75:8b:bd:38:eb:d4:82:76:60:1e:1b:d5:80:97
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=NAVER Global Root Certification Authority O=NAVER BUSINESS PLATFORM Corp.
+# Subject: CN=NAVER Global Root Certification Authority O=NAVER BUSINESS PLATFORM Corp.
+# Label: "NAVER Global Root Certification Authority"
+# Serial: 9013692873798656336226253319739695165984492813
+# MD5 Fingerprint: c8:7e:41:f6:25:3b:f5:09:b3:17:e8:46:3d:bf:d0:9b
+# SHA1 Fingerprint: 8f:6b:f2:a9:27:4a:da:14:a0:c4:f4:8e:61:27:f9:c0:1e:78:5d:d1
+# SHA256 Fingerprint: 88:f4:38:dc:f8:ff:d1:fa:8f:42:91:15:ff:e5:f8:2a:e1:e0:6e:0c:70:c3:75:fa:ad:71:7b:34:a4:9e:72:65
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AC RAIZ FNMT-RCM SERVIDORES SEGUROS O=FNMT-RCM OU=Ceres
+# Subject: CN=AC RAIZ FNMT-RCM SERVIDORES SEGUROS O=FNMT-RCM OU=Ceres
+# Label: "AC RAIZ FNMT-RCM SERVIDORES SEGUROS"
+# Serial: 131542671362353147877283741781055151509
+# MD5 Fingerprint: 19:36:9c:52:03:2f:d2:d1:bb:23:cc:dd:1e:12:55:bb
+# SHA1 Fingerprint: 62:ff:d9:9e:c0:65:0d:03:ce:75:93:d2:ed:3f:2d:32:c9:e3:e5:4a
+# SHA256 Fingerprint: 55:41:53:b1:3d:2c:f9:dd:b7:53:bf:be:1a:4e:0a:e0:8d:0a:a4:18:70:58:fe:60:a2:b8:62:b2:e4:b8:7b:cb
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign Root R46 O=GlobalSign nv-sa
+# Subject: CN=GlobalSign Root R46 O=GlobalSign nv-sa
+# Label: "GlobalSign Root R46"
+# Serial: 1552617688466950547958867513931858518042577
+# MD5 Fingerprint: c4:14:30:e4:fa:66:43:94:2a:6a:1b:24:5f:19:d0:ef
+# SHA1 Fingerprint: 53:a2:b0:4b:ca:6b:d6:45:e6:39:8a:8e:c4:0d:d2:bf:77:c3:a2:90
+# SHA256 Fingerprint: 4f:a3:12:6d:8d:3a:11:d1:c4:85:5a:4f:80:7c:ba:d6:cf:91:9d:3a:5a:88:b0:3b:ea:2c:63:72:d9:3c:40:c9
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign Root E46 O=GlobalSign nv-sa
+# Subject: CN=GlobalSign Root E46 O=GlobalSign nv-sa
+# Label: "GlobalSign Root E46"
+# Serial: 1552617690338932563915843282459653771421763
+# MD5 Fingerprint: b5:b8:66:ed:de:08:83:e3:c9:e2:01:34:06:ac:51:6f
+# SHA1 Fingerprint: 39:b4:6c:d5:fe:80:06:eb:e2:2f:4a:bb:08:33:a0:af:db:b9:dd:84
+# SHA256 Fingerprint: cb:b9:c4:4d:84:b8:04:3e:10:50:ea:31:a6:9f:51:49:55:d7:bf:d2:e2:c6:b4:93:01:01:9a:d6:1d:9f:50:58
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=GLOBALTRUST 2020 O=e-commerce monitoring GmbH
+# Subject: CN=GLOBALTRUST 2020 O=e-commerce monitoring GmbH
+# Label: "GLOBALTRUST 2020"
+# Serial: 109160994242082918454945253
+# MD5 Fingerprint: 8a:c7:6f:cb:6d:e3:cc:a2:f1:7c:83:fa:0e:78:d7:e8
+# SHA1 Fingerprint: d0:67:c1:13:51:01:0c:aa:d0:c7:6a:65:37:31:16:26:4f:53:71:a2
+# SHA256 Fingerprint: 9a:29:6a:51:82:d1:d4:51:a2:e3:7f:43:9b:74:da:af:a2:67:52:33:29:f9:0f:9a:0d:20:07:c3:34:e2:3c:9a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=ANF Secure Server Root CA O=ANF Autoridad de Certificacion OU=ANF CA Raiz
+# Subject: CN=ANF Secure Server Root CA O=ANF Autoridad de Certificacion OU=ANF CA Raiz
+# Label: "ANF Secure Server Root CA"
+# Serial: 996390341000653745
+# MD5 Fingerprint: 26:a6:44:5a:d9:af:4e:2f:b2:1d:b6:65:b0:4e:e8:96
+# SHA1 Fingerprint: 5b:6e:68:d0:cc:15:b6:a0:5f:1e:c1:5f:ae:02:fc:6b:2f:5d:6f:74
+# SHA256 Fingerprint: fb:8f:ec:75:91:69:b9:10:6b:1e:51:16:44:c6:18:c5:13:04:37:3f:6c:06:43:08:8d:8b:ef:fd:1b:99:75:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum EC-384 CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Subject: CN=Certum EC-384 CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Label: "Certum EC-384 CA"
+# Serial: 160250656287871593594747141429395092468
+# MD5 Fingerprint: b6:65:b3:96:60:97:12:a1:ec:4e:e1:3d:a3:c6:c9:f1
+# SHA1 Fingerprint: f3:3e:78:3c:ac:df:f4:a2:cc:ac:67:55:69:56:d7:e5:16:3c:e1:ed
+# SHA256 Fingerprint: 6b:32:80:85:62:53:18:aa:50:d1:73:c9:8d:8b:da:09:d5:7e:27:41:3d:11:4c:f7:87:a0:f5:d0:6c:03:0c:f6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Root CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Root CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Root CA"
+# Serial: 40870380103424195783807378461123655149
+# MD5 Fingerprint: 51:e1:c2:e7:fe:4c:84:af:59:0e:2f:f4:54:6f:ea:29
+# SHA1 Fingerprint: c8:83:44:c0:18:ae:9f:cc:f1:87:b7:8f:22:d1:c5:d7:45:84:ba:e5
+# SHA256 Fingerprint: fe:76:96:57:38:55:77:3e:37:a9:5e:7a:d4:d9:cc:96:c3:01:57:c1:5d:31:76:5b:a9:b1:57:04:e1:ae:78:fd
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TunTrust Root CA O=Agence Nationale de Certification Electronique
+# Subject: CN=TunTrust Root CA O=Agence Nationale de Certification Electronique
+# Label: "TunTrust Root CA"
+# Serial: 108534058042236574382096126452369648152337120275
+# MD5 Fingerprint: 85:13:b9:90:5b:36:5c:b6:5e:b8:5a:f8:e0:31:57:b4
+# SHA1 Fingerprint: cf:e9:70:84:0f:e0:73:0f:9d:f6:0c:7f:2c:4b:ee:20:46:34:9c:bb
+# SHA256 Fingerprint: 2e:44:10:2a:b5:8c:b8:54:19:45:1c:8e:19:d9:ac:f3:66:2c:af:bc:61:4b:6a:53:96:0a:30:f7:d0:e2:eb:41
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=HARICA TLS RSA Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Subject: CN=HARICA TLS RSA Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Label: "HARICA TLS RSA Root CA 2021"
+# Serial: 76817823531813593706434026085292783742
+# MD5 Fingerprint: 65:47:9b:58:86:dd:2c:f0:fc:a2:84:1f:1e:96:c4:91
+# SHA1 Fingerprint: 02:2d:05:82:fa:88:ce:14:0c:06:79:de:7f:14:10:e9:45:d7:a5:6d
+# SHA256 Fingerprint: d9:5d:0e:8e:da:79:52:5b:f9:be:b1:1b:14:d2:10:0d:32:94:98:5f:0c:62:d9:fa:bd:9c:d9:99:ec:cb:7b:1d
+-----BEGIN CERTIFICATE-----
+MIIFpDCCA4ygAwIBAgIQOcqTHO9D88aOk8f0ZIk4fjANBgkqhkiG9w0BAQsFADBs
+MQswCQYDVQQGEwJHUjE3MDUGA1UECgwuSGVsbGVuaWMgQWNhZGVtaWMgYW5kIFJl
+c2VhcmNoIEluc3RpdHV0aW9ucyBDQTEkMCIGA1UEAwwbSEFSSUNBIFRMUyBSU0Eg
+Um9vdCBDQSAyMDIxMB4XDTIxMDIxOTEwNTUzOFoXDTQ1MDIxMzEwNTUzN1owbDEL
+MAkGA1UEBhMCR1IxNzA1BgNVBAoMLkhlbGxlbmljIEFjYWRlbWljIGFuZCBSZXNl
+YXJjaCBJbnN0aXR1dGlvbnMgQ0ExJDAiBgNVBAMMG0hBUklDQSBUTFMgUlNBIFJv
+b3QgQ0EgMjAyMTCCAiIwDQYJKoZIhvcNAQEBBQADggIPADCCAgoCggIBAIvC569l
+mwVnlskNJLnQDmT8zuIkGCyEf3dRywQRNrhe7Wlxp57kJQmXZ8FHws+RFjZiPTgE
+4VGC/6zStGndLuwRo0Xua2s7TL+MjaQenRG56Tj5eg4MmOIjHdFOY9TnuEFE+2uv
+a9of08WRiFukiZLRgeaMOVig1mlDqa2YUlhu2wr7a89o+uOkXjpFc5gH6l8Cct4M
+pbOfrqkdtx2z/IpZ525yZa31MJQjB/OCFks1mJxTuy/K5FrZx40d/JiZ+yykgmvw
+Kh+OC19xXFyuQnspiYHLA6OZyoieC0AJQTPb5lh6/a6ZcMBaD9YThnEvdmn8kN3b
+LW7R8pv1GmuebxWMevBLKKAiOIAkbDakO/IwkfN4E8/BPzWr8R0RI7VDIp4BkrcY
+AuUR0YLbFQDMYTfBKnya4dC6s1BG7oKsnTH4+yPiAwBIcKMJJnkVU2DzOFytOOqB
+AGMUuTNe3QvboEUHGjMJ+E20pwKmafTCWQWIZYVWrkvL4N48fS0ayOn7H6NhStYq
+E613TBoYm5EPWNgGVMWX+Ko/IIqmhaZ39qb8HOLubpQzKoNQhArlT4b4UEV4AIHr
+W2jjJo3Me1xR9BQsQL4aYB16cmEdH2MtiKrOokWQCPxrvrNQKlr9qEgYRtaQQJKQ
+CoReaDH46+0N0x3GfZkYVVYnZS6NRcUk7M7jAgMBAAGjQjBAMA8GA1UdEwEB/wQF
+MAMBAf8wHQYDVR0OBBYEFApII6ZgpJIKM+qTW8VX6iVNvRLuMA4GA1UdDwEB/wQE
+AwIBhjANBgkqhkiG9w0BAQsFAAOCAgEAPpBIqm5iFSVmewzVjIuJndftTgfvnNAU
+X15QvWiWkKQUEapobQk1OUAJ2vQJLDSle1mESSmXdMgHHkdt8s4cUCbjnj1AUz/3
+f5Z2EMVGpdAgS1D0NTsY9FVqQRtHBmg8uwkIYtlfVUKqrFOFrJVWNlar5AWMxaja
+H6NpvVMPxP/cyuN+8kyIhkdGGvMA9YCRotxDQpSbIPDRzbLrLFPCU3hKTwSUQZqP
+JzLB5UkZv/HywouoCjkxKLR9YjYsTewfM7Z+d21+UPCfDtcRj88YxeMn/ibvBZ3P
+zzfF0HvaO7AWhAw6k9a+F9sPPg4ZeAnHqQJyIkv3N3a6dcSFA1pj1bF1BcK5vZSt
+jBWZp5N99sXzqnTPBIWUmAD04vnKJGW/4GKvyMX6ssmeVkjaef2WdhW+o45WxLM0
+/L5H9MG0qPzVMIho7suuyWPEdr6sOBjhXlzPrjoiUevRi7PzKzMHVIf6tLITe7pT
+BGIBnfHAT+7hOtSLIBD6Alfm78ELt5BGnBkpjNxvoEppaZS3JGWg/6w/zgH7IS79
+aPib8qXPMThcFarmlwDB31qlpzmq6YR/PFGoOtmUW4y/Twhx5duoXNTSpv4Ao8YW
+xw/ogM4cKGR0GQjTQuPOAF1/sdwTsOEFy9EgqoZ0njnnkf3/W9b3raYvAwtt41dU
+63ZTGI0RmLo=
+-----END CERTIFICATE-----
+
+# Issuer: CN=HARICA TLS ECC Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Subject: CN=HARICA TLS ECC Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Label: "HARICA TLS ECC Root CA 2021"
+# Serial: 137515985548005187474074462014555733966
+# MD5 Fingerprint: ae:f7:4c:e5:66:35:d1:b7:9b:8c:22:93:74:d3:4b:b0
+# SHA1 Fingerprint: bc:b0:c1:9d:e9:98:92:70:19:38:57:e9:8d:a7:b4:5d:6e:ee:01:48
+# SHA256 Fingerprint: 3f:99:cc:47:4a:cf:ce:4d:fe:d5:87:94:66:5e:47:8d:15:47:73:9f:2e:78:0f:1b:b4:ca:9b:13:30:97:d4:01
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Autoridad de Certificacion Firmaprofesional CIF A62634068
+# Subject: CN=Autoridad de Certificacion Firmaprofesional CIF A62634068
+# Label: "Autoridad de Certificacion Firmaprofesional CIF A62634068"
+# Serial: 1977337328857672817
+# MD5 Fingerprint: 4e:6e:9b:54:4c:ca:b7:fa:48:e4:90:b1:15:4b:1c:a3
+# SHA1 Fingerprint: 0b:be:c2:27:22:49:cb:39:aa:db:35:5c:53:e3:8c:ae:78:ff:b6:fe
+# SHA256 Fingerprint: 57:de:05:83:ef:d2:b2:6e:03:61:da:99:da:9d:f4:64:8d:ef:7e:e8:44:1c:3b:72:8a:fa:9b:cd:e0:f9:b2:6a
+-----BEGIN CERTIFICATE-----
+MIIGFDCCA/ygAwIBAgIIG3Dp0v+ubHEwDQYJKoZIhvcNAQELBQAwUTELMAkGA1UE
+BhMCRVMxQjBABgNVBAMMOUF1dG9yaWRhZCBkZSBDZXJ0aWZpY2FjaW9uIEZpcm1h
+cHJvZmVzaW9uYWwgQ0lGIEE2MjYzNDA2ODAeFw0xNDA5MjMxNTIyMDdaFw0zNjA1
+MDUxNTIyMDdaMFExCzAJBgNVBAYTAkVTMUIwQAYDVQQDDDlBdXRvcmlkYWQgZGUg
+Q2VydGlmaWNhY2lvbiBGaXJtYXByb2Zlc2lvbmFsIENJRiBBNjI2MzQwNjgwggIi
+MA0GCSqGSIb3DQEBAQUAA4ICDwAwggIKAoICAQDKlmuO6vj78aI14H9M2uDDUtd9
+thDIAl6zQyrET2qyyhxdKJp4ERppWVevtSBC5IsP5t9bpgOSL/UR5GLXMnE42QQM
+cas9UX4PB99jBVzpv5RvwSmCwLTaUbDBPLutN0pcyvFLNg4kq7/DhHf9qFD0sefG
+L9ItWY16Ck6WaVICqjaY7Pz6FIMMNx/Jkjd/14Et5cS54D40/mf0PmbR0/RAz15i
+NA9wBj4gGFrO93IbJWyTdBSTo3OxDqqHECNZXyAFGUftaI6SEspd/NYrspI8IM/h
+X68gvqB2f3bl7BqGYTM+53u0P6APjqK5am+5hyZvQWyIplD9amML9ZMWGxmPsu2b
+m8mQ9QEM3xk9Dz44I8kvjwzRAv4bVdZO0I08r0+k8/6vKtMFnXkIoctXMbScyJCy
+Z/QYFpM6/EfY0XiWMR+6KwxfXZmtY4laJCB22N/9q06mIqqdXuYnin1oKaPnirja
+EbsXLZmdEyRG98Xi2J+Of8ePdG1asuhy9azuJBCtLxTa/y2aRnFHvkLfuwHb9H/T
+KI8xWVvTyQKmtFLKbpf7Q8UIJm+K9Lv9nyiqDdVF8xM6HdjAeI9BZzwelGSuewvF
+6NkBiDkal4ZkQdU7hwxu+g/GvUgUvzlN1J5Bto+WHWOWk9mVBngxaJ43BjuAiUVh
+OSPHG0SjFeUc+JIwuwIDAQABo4HvMIHsMB0GA1UdDgQWBBRlzeurNR4APn7VdMAc
+tHNHDhpkLzASBgNVHRMBAf8ECDAGAQH/AgEBMIGmBgNVHSAEgZ4wgZswgZgGBFUd
+IAAwgY8wLwYIKwYBBQUHAgEWI2h0dHA6Ly93d3cuZmlybWFwcm9mZXNpb25hbC5j
+b20vY3BzMFwGCCsGAQUFBwICMFAeTgBQAGEAcwBlAG8AIABkAGUAIABsAGEAIABC
+AG8AbgBhAG4AbwB2AGEAIAA0ADcAIABCAGEAcgBjAGUAbABvAG4AYQAgADAAOAAw
+ADEANzAOBgNVHQ8BAf8EBAMCAQYwDQYJKoZIhvcNAQELBQADggIBAHSHKAIrdx9m
+iWTtj3QuRhy7qPj4Cx2Dtjqn6EWKB7fgPiDL4QjbEwj4KKE1soCzC1HA01aajTNF
+Sa9J8OA9B3pFE1r/yJfY0xgsfZb43aJlQ3CTkBW6kN/oGbDbLIpgD7dvlAceHabJ
+hfa9NPhAeGIQcDq+fUs5gakQ1JZBu/hfHAsdCPKxsIl68veg4MSPi3i1O1ilI45P
+Vf42O+AMt8oqMEEgtIDNrvx2ZnOorm7hfNoD6JQg5iKj0B+QXSBTFCZX2lSX3xZE
+EAEeiGaPcjiT3SC3NL7X8e5jjkd5KAb881lFJWAiMxujX6i6KtoaPc1A6ozuBRWV
+1aUsIC+nmCjuRfzxuIgALI9C2lHVnOUTaHFFQ4ueCyE8S1wF3BqfmI7avSKecs2t
+CsvMo2ebKHTEm9caPARYpoKdrcd7b/+Alun4jWq9GJAd/0kakFI3ky88Al2CdgtR
+5xbHV/g4+afNmyJU72OwFW1TZQNKXkqgsqeOSQBZONXH9IBk9W6VULgRfhVwOEqw
+f9DEMnDAGf/JOC0ULGb0QkTmVXYbgBVX/8Cnp6o5qtjTcNAuuuuUavpfNIbnYrX9
+ivAwhZTJryQCL2/W3Wf+47BVTwSYT6RBVuKT0Gro1vP7ZeDOdcQxWQzugsgMYDNK
+GbqEZycPvEJdvSRUDewdcAZfpLz6IHxV
+-----END CERTIFICATE-----
+
+# Issuer: CN=vTrus ECC Root CA O=iTrusChina Co.,Ltd.
+# Subject: CN=vTrus ECC Root CA O=iTrusChina Co.,Ltd.
+# Label: "vTrus ECC Root CA"
+# Serial: 630369271402956006249506845124680065938238527194
+# MD5 Fingerprint: de:4b:c1:f5:52:8c:9b:43:e1:3e:8f:55:54:17:8d:85
+# SHA1 Fingerprint: f6:9c:db:b0:fc:f6:02:13:b6:52:32:a6:a3:91:3f:16:70:da:c3:e1
+# SHA256 Fingerprint: 30:fb:ba:2c:32:23:8e:2a:98:54:7a:f9:79:31:e5:50:42:8b:9b:3f:1c:8e:eb:66:33:dc:fa:86:c5:b2:7d:d3
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=vTrus Root CA O=iTrusChina Co.,Ltd.
+# Subject: CN=vTrus Root CA O=iTrusChina Co.,Ltd.
+# Label: "vTrus Root CA"
+# Serial: 387574501246983434957692974888460947164905180485
+# MD5 Fingerprint: b8:c9:37:df:fa:6b:31:84:64:c5:ea:11:6a:1b:75:fc
+# SHA1 Fingerprint: 84:1a:69:fb:f5:cd:1a:25:34:13:3d:e3:f8:fc:b8:99:d0:c9:14:b7
+# SHA256 Fingerprint: 8a:71:de:65:59:33:6f:42:6c:26:e5:38:80:d0:0d:88:a1:8d:a4:c6:a9:1f:0d:cb:61:94:e2:06:c5:c9:63:87
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=ISRG Root X2 O=Internet Security Research Group
+# Subject: CN=ISRG Root X2 O=Internet Security Research Group
+# Label: "ISRG Root X2"
+# Serial: 87493402998870891108772069816698636114
+# MD5 Fingerprint: d3:9e:c4:1e:23:3c:a6:df:cf:a3:7e:6d:e0:14:e6:e5
+# SHA1 Fingerprint: bd:b1:b9:3c:d5:97:8d:45:c6:26:14:55:f8:db:95:c7:5a:d1:53:af
+# SHA256 Fingerprint: 69:72:9b:8e:15:a8:6e:fc:17:7a:57:af:b7:17:1d:fc:64:ad:d2:8c:2f:ca:8c:f1:50:7e:34:45:3c:cb:14:70
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=HiPKI Root CA - G1 O=Chunghwa Telecom Co., Ltd.
+# Subject: CN=HiPKI Root CA - G1 O=Chunghwa Telecom Co., Ltd.
+# Label: "HiPKI Root CA - G1"
+# Serial: 60966262342023497858655262305426234976
+# MD5 Fingerprint: 69:45:df:16:65:4b:e8:68:9a:8f:76:5f:ff:80:9e:d3
+# SHA1 Fingerprint: 6a:92:e4:a8:ee:1b:ec:96:45:37:e3:29:57:49:cd:96:e3:e5:d2:60
+# SHA256 Fingerprint: f0:15:ce:3c:c2:39:bf:ef:06:4b:e9:f1:d2:c4:17:e1:a0:26:4a:0a:94:be:1f:0c:8d:12:18:64:eb:69:49:cc
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R4
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R4
+# Label: "GlobalSign ECC Root CA - R4"
+# Serial: 159662223612894884239637590694
+# MD5 Fingerprint: 26:29:f8:6d:e1:88:bf:a2:65:7f:aa:c4:cd:0f:7f:fc
+# SHA1 Fingerprint: 6b:a0:b0:98:e1:71:ef:5a:ad:fe:48:15:80:77:10:f4:bd:6f:0b:28
+# SHA256 Fingerprint: b0:85:d7:0b:96:4f:19:1a:73:e4:af:0d:54:ae:7a:0e:07:aa:fd:af:9b:71:dd:08:62:13:8a:b7:32:5a:24:a2
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R1 O=Google Trust Services LLC
+# Subject: CN=GTS Root R1 O=Google Trust Services LLC
+# Label: "GTS Root R1"
+# Serial: 159662320309726417404178440727
+# MD5 Fingerprint: 05:fe:d0:bf:71:a8:a3:76:63:da:01:e0:d8:52:dc:40
+# SHA1 Fingerprint: e5:8c:1c:c4:91:3b:38:63:4b:e9:10:6e:e3:ad:8e:6b:9d:d9:81:4a
+# SHA256 Fingerprint: d9:47:43:2a:bd:e7:b7:fa:90:fc:2e:6b:59:10:1b:12:80:e0:e1:c7:e4:e4:0f:a3:c6:88:7f:ff:57:a7:f4:cf
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R2 O=Google Trust Services LLC
+# Subject: CN=GTS Root R2 O=Google Trust Services LLC
+# Label: "GTS Root R2"
+# Serial: 159662449406622349769042896298
+# MD5 Fingerprint: 1e:39:c0:53:e6:1e:29:82:0b:ca:52:55:36:5d:57:dc
+# SHA1 Fingerprint: 9a:44:49:76:32:db:de:fa:d0:bc:fb:5a:7b:17:bd:9e:56:09:24:94
+# SHA256 Fingerprint: 8d:25:cd:97:22:9d:bf:70:35:6b:da:4e:b3:cc:73:40:31:e2:4c:f0:0f:af:cf:d3:2d:c7:6e:b5:84:1c:7e:a8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R3 O=Google Trust Services LLC
+# Subject: CN=GTS Root R3 O=Google Trust Services LLC
+# Label: "GTS Root R3"
+# Serial: 159662495401136852707857743206
+# MD5 Fingerprint: 3e:e7:9d:58:02:94:46:51:94:e5:e0:22:4a:8b:e7:73
+# SHA1 Fingerprint: ed:e5:71:80:2b:c8:92:b9:5b:83:3c:d2:32:68:3f:09:cd:a0:1e:46
+# SHA256 Fingerprint: 34:d8:a7:3e:e2:08:d9:bc:db:0d:95:65:20:93:4b:4e:40:e6:94:82:59:6e:8b:6f:73:c8:42:6b:01:0a:6f:48
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R4 O=Google Trust Services LLC
+# Subject: CN=GTS Root R4 O=Google Trust Services LLC
+# Label: "GTS Root R4"
+# Serial: 159662532700760215368942768210
+# MD5 Fingerprint: 43:96:83:77:19:4d:76:b3:9d:65:52:e4:1d:22:a5:e8
+# SHA1 Fingerprint: 77:d3:03:67:b5:e0:0c:15:f6:0c:38:61:df:7c:e1:3b:92:46:4d:47
+# SHA256 Fingerprint: 34:9d:fa:40:58:c5:e2:63:12:3b:39:8a:e7:95:57:3c:4e:13:13:c8:3f:e6:8f:93:55:6c:d5:e8:03:1b:3c:7d
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=Telia Root CA v2 O=Telia Finland Oyj
+# Subject: CN=Telia Root CA v2 O=Telia Finland Oyj
+# Label: "Telia Root CA v2"
+# Serial: 7288924052977061235122729490515358
+# MD5 Fingerprint: 0e:8f:ac:aa:82:df:85:b1:f4:dc:10:1c:fc:99:d9:48
+# SHA1 Fingerprint: b9:99:cd:d1:73:50:8a:c4:47:05:08:9c:8c:88:fb:be:a0:2b:40:cd
+# SHA256 Fingerprint: 24:2b:69:74:2f:cb:1e:5b:2a:bf:98:89:8b:94:57:21:87:54:4e:5b:4d:99:11:78:65:73:62:1f:6a:74:b8:2c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST BR Root CA 1 2020 O=D-Trust GmbH
+# Subject: CN=D-TRUST BR Root CA 1 2020 O=D-Trust GmbH
+# Label: "D-TRUST BR Root CA 1 2020"
+# Serial: 165870826978392376648679885835942448534
+# MD5 Fingerprint: b5:aa:4b:d5:ed:f7:e3:55:2e:8f:72:0a:f3:75:b8:ed
+# SHA1 Fingerprint: 1f:5b:98:f0:e3:b5:f7:74:3c:ed:e6:b0:36:7d:32:cd:f4:09:41:67
+# SHA256 Fingerprint: e5:9a:aa:81:60:09:c2:2b:ff:5b:25:ba:d3:7d:f3:06:f0:49:79:7c:1f:81:d8:5a:b0:89:e6:57:bd:8f:00:44
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST EV Root CA 1 2020 O=D-Trust GmbH
+# Subject: CN=D-TRUST EV Root CA 1 2020 O=D-Trust GmbH
+# Label: "D-TRUST EV Root CA 1 2020"
+# Serial: 126288379621884218666039612629459926992
+# MD5 Fingerprint: 8c:2d:9d:70:9f:48:99:11:06:11:fb:e9:cb:30:c0:6e
+# SHA1 Fingerprint: 61:db:8c:21:59:69:03:90:d8:7c:9c:12:86:54:cf:9d:3d:f4:dd:07
+# SHA256 Fingerprint: 08:17:0d:1a:a3:64:53:90:1a:2f:95:92:45:e3:47:db:0c:8d:37:ab:aa:bc:56:b8:1a:a1:00:dc:95:89:70:db
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert TLS ECC P384 Root G5 O=DigiCert, Inc.
+# Subject: CN=DigiCert TLS ECC P384 Root G5 O=DigiCert, Inc.
+# Label: "DigiCert TLS ECC P384 Root G5"
+# Serial: 13129116028163249804115411775095713523
+# MD5 Fingerprint: d3:71:04:6a:43:1c:db:a6:59:e1:a8:a3:aa:c5:71:ed
+# SHA1 Fingerprint: 17:f3:de:5e:9f:0f:19:e9:8e:f6:1f:32:26:6e:20:c4:07:ae:30:ee
+# SHA256 Fingerprint: 01:8e:13:f0:77:25:32:cf:80:9b:d1:b1:72:81:86:72:83:fc:48:c6:e1:3b:e9:c6:98:12:85:4a:49:0c:1b:05
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert TLS RSA4096 Root G5 O=DigiCert, Inc.
+# Subject: CN=DigiCert TLS RSA4096 Root G5 O=DigiCert, Inc.
+# Label: "DigiCert TLS RSA4096 Root G5"
+# Serial: 11930366277458970227240571539258396554
+# MD5 Fingerprint: ac:fe:f7:34:96:a9:f2:b3:b4:12:4b:e4:27:41:6f:e1
+# SHA1 Fingerprint: a7:88:49:dc:5d:7c:75:8c:8c:de:39:98:56:b3:aa:d0:b2:a5:71:35
+# SHA256 Fingerprint: 37:1a:00:dc:05:33:b3:72:1a:7e:eb:40:e8:41:9e:70:79:9d:2b:0a:0f:2c:1d:80:69:31:65:f7:ce:c4:ad:75
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certainly Root R1 O=Certainly
+# Subject: CN=Certainly Root R1 O=Certainly
+# Label: "Certainly Root R1"
+# Serial: 188833316161142517227353805653483829216
+# MD5 Fingerprint: 07:70:d4:3e:82:87:a0:fa:33:36:13:f4:fa:33:e7:12
+# SHA1 Fingerprint: a0:50:ee:0f:28:71:f4:27:b2:12:6d:6f:50:96:25:ba:cc:86:42:af
+# SHA256 Fingerprint: 77:b8:2c:d8:64:4c:43:05:f7:ac:c5:cb:15:6b:45:67:50:04:03:3d:51:c6:0c:62:02:a8:e0:c3:34:67:d3:a0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certainly Root E1 O=Certainly
+# Subject: CN=Certainly Root E1 O=Certainly
+# Label: "Certainly Root E1"
+# Serial: 8168531406727139161245376702891150584
+# MD5 Fingerprint: 0a:9e:ca:cd:3e:52:50:c6:36:f3:4b:a3:ed:a7:53:e9
+# SHA1 Fingerprint: f9:e1:6d:dc:01:89:cf:d5:82:45:63:3e:c5:37:7d:c2:eb:93:6f:2b
+# SHA256 Fingerprint: b4:58:5f:22:e4:ac:75:6a:4e:86:12:a1:36:1c:5d:9d:03:1a:93:fd:84:fe:bb:77:8f:a3:06:8b:0f:c4:2d:c2
+-----BEGIN CERTIFICATE-----
+MIIB9zCCAX2gAwIBAgIQBiUzsUcDMydc+Y2aub/M+DAKBggqhkjOPQQDAzA9MQsw
+CQYDVQQGEwJVUzESMBAGA1UEChMJQ2VydGFpbmx5MRowGAYDVQQDExFDZXJ0YWlu
+bHkgUm9vdCBFMTAeFw0yMTA0MDEwMDAwMDBaFw00NjA0MDEwMDAwMDBaMD0xCzAJ
+BgNVBAYTAlVTMRIwEAYDVQQKEwlDZXJ0YWlubHkxGjAYBgNVBAMTEUNlcnRhaW5s
+eSBSb290IEUxMHYwEAYHKoZIzj0CAQYFK4EEACIDYgAE3m/4fxzf7flHh4axpMCK
++IKXgOqPyEpeKn2IaKcBYhSRJHpcnqMXfYqGITQYUBsQ3tA3SybHGWCA6TS9YBk2
+QNYphwk8kXr2vBMj3VlOBF7PyAIcGFPBMdjaIOlEjeR2o0IwQDAOBgNVHQ8BAf8E
+BAMCAQYwDwYDVR0TAQH/BAUwAwEB/zAdBgNVHQ4EFgQU8ygYy2R17ikq6+2uI1g4
+hevIIgcwCgYIKoZIzj0EAwMDaAAwZQIxALGOWiDDshliTd6wT99u0nCK8Z9+aozm
+ut6Dacpps6kFtZaSF4fC0urQe87YQVt8rgIwRt7qy12a7DLCZRawTDBcMPPaTnOG
+BtjOiQRINzf43TNRnXCve1XYAS59BWQOhriR
+-----END CERTIFICATE-----
+
+# Issuer: CN=Security Communication ECC RootCA1 O=SECOM Trust Systems CO.,LTD.
+# Subject: CN=Security Communication ECC RootCA1 O=SECOM Trust Systems CO.,LTD.
+# Label: "Security Communication ECC RootCA1"
+# Serial: 15446673492073852651
+# MD5 Fingerprint: 7e:43:b0:92:68:ec:05:43:4c:98:ab:5d:35:2e:7e:86
+# SHA1 Fingerprint: b8:0e:26:a9:bf:d2:b2:3b:c0:ef:46:c9:ba:c7:bb:f6:1d:0d:41:41
+# SHA256 Fingerprint: e7:4f:bd:a5:5b:d5:64:c4:73:a3:6b:44:1a:a7:99:c8:a6:8e:07:74:40:e8:28:8b:9f:a1:e5:0e:4b:ba:ca:11
+-----BEGIN CERTIFICATE-----
+MIICODCCAb6gAwIBAgIJANZdm7N4gS7rMAoGCCqGSM49BAMDMGExCzAJBgNVBAYT
+AkpQMSUwIwYDVQQKExxTRUNPTSBUcnVzdCBTeXN0ZW1zIENPLixMVEQuMSswKQYD
+VQQDEyJTZWN1cml0eSBDb21tdW5pY2F0aW9uIEVDQyBSb290Q0ExMB4XDTE2MDYx
+NjA1MTUyOFoXDTM4MDExODA1MTUyOFowYTELMAkGA1UEBhMCSlAxJTAjBgNVBAoT
+HFNFQ09NIFRydXN0IFN5c3RlbXMgQ08uLExURC4xKzApBgNVBAMTIlNlY3VyaXR5
+IENvbW11bmljYXRpb24gRUNDIFJvb3RDQTEwdjAQBgcqhkjOPQIBBgUrgQQAIgNi
+AASkpW9gAwPDvTH00xecK4R1rOX9PVdu12O/5gSJko6BnOPpR27KkBLIE+Cnnfdl
+dB9sELLo5OnvbYUymUSxXv3MdhDYW72ixvnWQuRXdtyQwjWpS4g8EkdtXP9JTxpK
+ULGjQjBAMB0GA1UdDgQWBBSGHOf+LaVKiwj+KBH6vqNm+GBZLzAOBgNVHQ8BAf8E
+BAMCAQYwDwYDVR0TAQH/BAUwAwEB/zAKBggqhkjOPQQDAwNoADBlAjAVXUI9/Lbu
+9zuxNuie9sRGKEkz0FhDKmMpzE2xtHqiuQ04pV1IKv3LsnNdo4gIxwwCMQDAqy0O
+be0YottT6SXbVQjgUMzfRGEWgqtJsLKB7HOHeLRMsmIbEvoWTSVLY70eN9k=
+-----END CERTIFICATE-----
+
+# Issuer: CN=BJCA Global Root CA1 O=BEIJING CERTIFICATE AUTHORITY
+# Subject: CN=BJCA Global Root CA1 O=BEIJING CERTIFICATE AUTHORITY
+# Label: "BJCA Global Root CA1"
+# Serial: 113562791157148395269083148143378328608
+# MD5 Fingerprint: 42:32:99:76:43:33:36:24:35:07:82:9b:28:f9:d0:90
+# SHA1 Fingerprint: d5:ec:8d:7b:4c:ba:79:f4:e7:e8:cb:9d:6b:ae:77:83:10:03:21:6a
+# SHA256 Fingerprint: f3:89:6f:88:fe:7c:0a:88:27:66:a7:fa:6a:d2:74:9f:b5:7a:7f:3e:98:fb:76:9c:1f:a7:b0:9c:2c:44:d5:ae
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=BJCA Global Root CA2 O=BEIJING CERTIFICATE AUTHORITY
+# Subject: CN=BJCA Global Root CA2 O=BEIJING CERTIFICATE AUTHORITY
+# Label: "BJCA Global Root CA2"
+# Serial: 58605626836079930195615843123109055211
+# MD5 Fingerprint: 5e:0a:f6:47:5f:a6:14:e8:11:01:95:3f:4d:01:eb:3c
+# SHA1 Fingerprint: f4:27:86:eb:6e:b8:6d:88:31:67:02:fb:ba:66:a4:53:00:aa:7a:a6
+# SHA256 Fingerprint: 57:4d:f6:93:1e:27:80:39:66:7b:72:0a:fd:c1:60:0f:c2:7e:b6:6d:d3:09:29:79:fb:73:85:64:87:21:28:82
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Sectigo Public Server Authentication Root E46 O=Sectigo Limited
+# Subject: CN=Sectigo Public Server Authentication Root E46 O=Sectigo Limited
+# Label: "Sectigo Public Server Authentication Root E46"
+# Serial: 88989738453351742415770396670917916916
+# MD5 Fingerprint: 28:23:f8:b2:98:5c:37:16:3b:3e:46:13:4e:b0:b3:01
+# SHA1 Fingerprint: ec:8a:39:6c:40:f0:2e:bc:42:75:d4:9f:ab:1c:1a:5b:67:be:d2:9a
+# SHA256 Fingerprint: c9:0f:26:f0:fb:1b:40:18:b2:22:27:51:9b:5c:a2:b5:3e:2c:a5:b3:be:5c:f1:8e:fe:1b:ef:47:38:0c:53:83
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Sectigo Public Server Authentication Root R46 O=Sectigo Limited
+# Subject: CN=Sectigo Public Server Authentication Root R46 O=Sectigo Limited
+# Label: "Sectigo Public Server Authentication Root R46"
+# Serial: 156256931880233212765902055439220583700
+# MD5 Fingerprint: 32:10:09:52:00:d5:7e:6c:43:df:15:c0:b1:16:93:e5
+# SHA1 Fingerprint: ad:98:f9:f3:e4:7d:75:3b:65:d4:82:b3:a4:52:17:bb:6e:f5:e4:38
+# SHA256 Fingerprint: 7b:b6:47:a6:2a:ee:ac:88:bf:25:7a:a5:22:d0:1f:fe:a3:95:e0:ab:45:c7:3f:93:f6:56:54:ec:38:f2:5a:06
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com TLS RSA Root CA 2022 O=SSL Corporation
+# Subject: CN=SSL.com TLS RSA Root CA 2022 O=SSL Corporation
+# Label: "SSL.com TLS RSA Root CA 2022"
+# Serial: 148535279242832292258835760425842727825
+# MD5 Fingerprint: d8:4e:c6:59:30:d8:fe:a0:d6:7a:5a:2c:2c:69:78:da
+# SHA1 Fingerprint: ec:2c:83:40:72:af:26:95:10:ff:0e:f2:03:ee:31:70:f6:78:9d:ca
+# SHA256 Fingerprint: 8f:af:7d:2e:2c:b4:70:9b:b8:e0:b3:36:66:bf:75:a5:dd:45:b5:de:48:0f:8e:a8:d4:bf:e6:be:bc:17:f2:ed
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com TLS ECC Root CA 2022 O=SSL Corporation
+# Subject: CN=SSL.com TLS ECC Root CA 2022 O=SSL Corporation
+# Label: "SSL.com TLS ECC Root CA 2022"
+# Serial: 26605119622390491762507526719404364228
+# MD5 Fingerprint: 99:d7:5c:f1:51:36:cc:e9:ce:d9:19:2e:77:71:56:c5
+# SHA1 Fingerprint: 9f:5f:d9:1a:54:6d:f5:0c:71:f0:ee:7a:bd:17:49:98:84:73:e2:39
+# SHA256 Fingerprint: c3:2f:fd:9f:46:f9:36:d1:6c:36:73:99:09:59:43:4b:9a:d6:0a:af:bb:9e:7c:f3:36:54:f1:44:cc:1b:a1:43
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot Root CA ECC TLS 2021 O=Atos
+# Subject: CN=Atos TrustedRoot Root CA ECC TLS 2021 O=Atos
+# Label: "Atos TrustedRoot Root CA ECC TLS 2021"
+# Serial: 81873346711060652204712539181482831616
+# MD5 Fingerprint: 16:9f:ad:f1:70:ad:79:d6:ed:29:b4:d1:c5:79:70:a8
+# SHA1 Fingerprint: 9e:bc:75:10:42:b3:02:f3:81:f4:f7:30:62:d4:8f:c3:a7:51:b2:dd
+# SHA256 Fingerprint: b2:fa:e5:3e:14:cc:d7:ab:92:12:06:47:01:ae:27:9c:1d:89:88:fa:cb:77:5f:a8:a0:08:91:4e:66:39:88:a8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot Root CA RSA TLS 2021 O=Atos
+# Subject: CN=Atos TrustedRoot Root CA RSA TLS 2021 O=Atos
+# Label: "Atos TrustedRoot Root CA RSA TLS 2021"
+# Serial: 111436099570196163832749341232207667876
+# MD5 Fingerprint: d4:d3:46:b8:9a:c0:9c:76:5d:9e:3a:c3:b9:99:31:d2
+# SHA1 Fingerprint: 18:52:3b:0d:06:37:e4:d6:3a:df:23:e4:98:fb:5b:16:fb:86:74:48
+# SHA256 Fingerprint: 81:a9:08:8e:a5:9f:b3:64:c5:48:a6:f8:55:59:09:9b:6f:04:05:ef:bf:18:e5:32:4e:c9:f4:57:ba:00:11:2f
+-----BEGIN CERTIFICATE-----
+MIIFZDCCA0ygAwIBAgIQU9XP5hmTC/srBRLYwiqipDANBgkqhkiG9w0BAQwFADBM
+MS4wLAYDVQQDDCVBdG9zIFRydXN0ZWRSb290IFJvb3QgQ0EgUlNBIFRMUyAyMDIx
+MQ0wCwYDVQQKDARBdG9zMQswCQYDVQQGEwJERTAeFw0yMTA0MjIwOTIxMTBaFw00
+MTA0MTcwOTIxMDlaMEwxLjAsBgNVBAMMJUF0b3MgVHJ1c3RlZFJvb3QgUm9vdCBD
+QSBSU0EgVExTIDIwMjExDTALBgNVBAoMBEF0b3MxCzAJBgNVBAYTAkRFMIICIjAN
+BgkqhkiG9w0BAQEFAAOCAg8AMIICCgKCAgEAtoAOxHm9BYx9sKOdTSJNy/BBl01Z
+4NH+VoyX8te9j2y3I49f1cTYQcvyAh5x5en2XssIKl4w8i1mx4QbZFc4nXUtVsYv
+Ye+W/CBGvevUez8/fEc4BKkbqlLfEzfTFRVOvV98r61jx3ncCHvVoOX3W3WsgFWZ
+kmGbzSoXfduP9LVq6hdKZChmFSlsAvFr1bqjM9xaZ6cF4r9lthawEO3NUDPJcFDs
+GY6wx/J0W2tExn2WuZgIWWbeKQGb9Cpt0xU6kGpn8bRrZtkh68rZYnxGEFzedUln
+nkL5/nWpo63/dgpnQOPF943HhZpZnmKaau1Fh5hnstVKPNe0OwANwI8f4UDErmwh
+3El+fsqyjW22v5MvoVw+j8rtgI5Y4dtXz4U2OLJxpAmMkokIiEjxQGMYsluMWuPD
+0xeqqxmjLBvk1cbiZnrXghmmOxYsL3GHX0WelXOTwkKBIROW1527k2gV+p2kHYzy
+geBYBr3JtuP2iV2J+axEoctr+hbxx1A9JNr3w+SH1VbxT5Aw+kUJWdo0zuATHAR8
+ANSbhqRAvNncTFd+rrcztl524WWLZt+NyteYr842mIycg5kDcPOvdO3GDjbnvezB
+c6eUWsuSZIKmAMFwoW4sKeFYV+xafJlrJaSQOoD0IJ2azsct+bJLKZWD6TWNp0lI
+pw9MGZHQ9b8Q4HECAwEAAaNCMEAwDwYDVR0TAQH/BAUwAwEB/zAdBgNVHQ4EFgQU
+dEmZ0f+0emhFdcN+tNzMzjkz2ggwDgYDVR0PAQH/BAQDAgGGMA0GCSqGSIb3DQEB
+DAUAA4ICAQAjQ1MkYlxt/T7Cz1UAbMVWiLkO3TriJQ2VSpfKgInuKs1l+NsW4AmS
+4BjHeJi78+xCUvuppILXTdiK/ORO/auQxDh1MoSf/7OwKwIzNsAQkG8dnK/haZPs
+o0UvFJ/1TCplQ3IM98P4lYsU84UgYt1UU90s3BiVaU+DR3BAM1h3Egyi61IxHkzJ
+qM7F78PRreBrAwA0JrRUITWXAdxfG/F851X6LWh3e9NpzNMOa7pNdkTWwhWaJuyw
+xfW70Xp0wmzNxbVe9kzmWy2B27O3Opee7c9GslA9hGCZcbUztVdF5kJHdWoOsAgM
+rr3e97sPWD2PAzHoPYJQyi9eDF20l74gNAf0xBLh7tew2VktafcxBPTy+av5EzH4
+AXcOPUIjJsyacmdRIXrMPIWo6iFqO9taPKU0nprALN+AnCng33eU0aKAQv9qTFsR
+0PXNor6uzFFcw9VUewyu1rkGd4Di7wcaaMxZUa1+XGdrudviB0JbuAEFWDlN5LuY
+o7Ey7Nmj1m+UI/87tyll5gfp77YZ6ufCOB0yiJA8EytuzO+rdwY0d4RPcuSBhPm5
+dDTedk+SKlOxJTnbPP/lPqYO5Wue/9vsL3SD3460s6neFE3/MaNFcyT6lSnMEpcE
+oji2jbDwN/zIIX8/syQbPYtuzE2wFg2WHYMfRsCbvUOZ58SWLs5fyQ==
+-----END CERTIFICATE-----
+
+# Issuer: CN=TrustAsia Global Root CA G3 O=TrustAsia Technologies, Inc.
+# Subject: CN=TrustAsia Global Root CA G3 O=TrustAsia Technologies, Inc.
+# Label: "TrustAsia Global Root CA G3"
+# Serial: 576386314500428537169965010905813481816650257167
+# MD5 Fingerprint: 30:42:1b:b7:bb:81:75:35:e4:16:4f:53:d2:94:de:04
+# SHA1 Fingerprint: 63:cf:b6:c1:27:2b:56:e4:88:8e:1c:23:9a:b6:2e:81:47:24:c3:c7
+# SHA256 Fingerprint: e0:d3:22:6a:eb:11:63:c2:e4:8f:f9:be:3b:50:b4:c6:43:1b:e7:bb:1e:ac:c5:c3:6b:5d:5e:c5:09:03:9a:08
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TrustAsia Global Root CA G4 O=TrustAsia Technologies, Inc.
+# Subject: CN=TrustAsia Global Root CA G4 O=TrustAsia Technologies, Inc.
+# Label: "TrustAsia Global Root CA G4"
+# Serial: 451799571007117016466790293371524403291602933463
+# MD5 Fingerprint: 54:dd:b2:d7:5f:d8:3e:ed:7c:e0:0b:2e:cc:ed:eb:eb
+# SHA1 Fingerprint: 57:73:a5:61:5d:80:b2:e6:ac:38:82:fc:68:07:31:ac:9f:b5:92:5a
+# SHA256 Fingerprint: be:4b:56:cb:50:56:c0:13:6a:52:6d:f4:44:50:8d:aa:36:a0:b5:4f:42:e4:ac:38:f7:2a:f4:70:e4:79:65:4c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust ECC Root-01 O=CommScope
+# Subject: CN=CommScope Public Trust ECC Root-01 O=CommScope
+# Label: "CommScope Public Trust ECC Root-01"
+# Serial: 385011430473757362783587124273108818652468453534
+# MD5 Fingerprint: 3a:40:a7:fc:03:8c:9c:38:79:2f:3a:a2:6c:b6:0a:16
+# SHA1 Fingerprint: 07:86:c0:d8:dd:8e:c0:80:98:06:98:d0:58:7a:ef:de:a6:cc:a2:5d
+# SHA256 Fingerprint: 11:43:7c:da:7b:b4:5e:41:36:5f:45:b3:9a:38:98:6b:0d:e0:0d:ef:34:8e:0c:7b:b0:87:36:33:80:0b:c3:8b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust ECC Root-02 O=CommScope
+# Subject: CN=CommScope Public Trust ECC Root-02 O=CommScope
+# Label: "CommScope Public Trust ECC Root-02"
+# Serial: 234015080301808452132356021271193974922492992893
+# MD5 Fingerprint: 59:b0:44:d5:65:4d:b8:5c:55:19:92:02:b6:d1:94:b2
+# SHA1 Fingerprint: 3c:3f:ef:57:0f:fe:65:93:86:9e:a0:fe:b0:f6:ed:8e:d1:13:c7:e5
+# SHA256 Fingerprint: 2f:fb:7f:81:3b:bb:b3:c8:9a:b4:e8:16:2d:0f:16:d7:15:09:a8:30:cc:9d:73:c2:62:e5:14:08:75:d1:ad:4a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust RSA Root-01 O=CommScope
+# Subject: CN=CommScope Public Trust RSA Root-01 O=CommScope
+# Label: "CommScope Public Trust RSA Root-01"
+# Serial: 354030733275608256394402989253558293562031411421
+# MD5 Fingerprint: 0e:b4:15:bc:87:63:5d:5d:02:73:d4:26:38:68:73:d8
+# SHA1 Fingerprint: 6d:0a:5f:f7:b4:23:06:b4:85:b3:b7:97:64:fc:ac:75:f5:33:f2:93
+# SHA256 Fingerprint: 02:bd:f9:6e:2a:45:dd:9b:f1:8f:c7:e1:db:df:21:a0:37:9b:a3:c9:c2:61:03:44:cf:d8:d6:06:fe:c1:ed:81
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust RSA Root-02 O=CommScope
+# Subject: CN=CommScope Public Trust RSA Root-02 O=CommScope
+# Label: "CommScope Public Trust RSA Root-02"
+# Serial: 480062499834624527752716769107743131258796508494
+# MD5 Fingerprint: e1:29:f9:62:7b:76:e2:96:6d:f3:d4:d7:0f:ae:1f:aa
+# SHA1 Fingerprint: ea:b0:e2:52:1b:89:93:4c:11:68:f2:d8:9a:ac:22:4c:a3:8a:57:ae
+# SHA256 Fingerprint: ff:e9:43:d7:93:42:4b:4f:7c:44:0c:1c:3d:64:8d:53:63:f3:4b:82:dc:87:aa:7a:9f:11:8f:c5:de:e1:01:f1
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=Telekom Security TLS ECC Root 2020 O=Deutsche Telekom Security GmbH
+# Subject: CN=Telekom Security TLS ECC Root 2020 O=Deutsche Telekom Security GmbH
+# Label: "Telekom Security TLS ECC Root 2020"
+# Serial: 72082518505882327255703894282316633856
+# MD5 Fingerprint: c1:ab:fe:6a:10:2c:03:8d:bc:1c:22:32:c0:85:a7:fd
+# SHA1 Fingerprint: c0:f8:96:c5:a9:3b:01:06:21:07:da:18:42:48:bc:e9:9d:88:d5:ec
+# SHA256 Fingerprint: 57:8a:f4:de:d0:85:3f:4e:59:98:db:4a:ea:f9:cb:ea:8d:94:5f:60:b6:20:a3:8d:1a:3c:13:b2:bc:7b:a8:e1
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Telekom Security TLS RSA Root 2023 O=Deutsche Telekom Security GmbH
+# Subject: CN=Telekom Security TLS RSA Root 2023 O=Deutsche Telekom Security GmbH
+# Label: "Telekom Security TLS RSA Root 2023"
+# Serial: 44676229530606711399881795178081572759
+# MD5 Fingerprint: bf:5b:eb:54:40:cd:48:71:c4:20:8d:7d:de:0a:42:f2
+# SHA1 Fingerprint: 54:d3:ac:b3:bd:57:56:f6:85:9d:ce:e5:c3:21:e2:d4:ad:83:d0:93
+# SHA256 Fingerprint: ef:c6:5c:ad:bb:59:ad:b6:ef:e8:4d:a2:23:11:b3:56:24:b7:1b:3b:1e:a0:da:8b:66:55:17:4e:c8:97:86:46
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=FIRMAPROFESIONAL CA ROOT-A WEB O=Firmaprofesional SA
+# Subject: CN=FIRMAPROFESIONAL CA ROOT-A WEB O=Firmaprofesional SA
+# Label: "FIRMAPROFESIONAL CA ROOT-A WEB"
+# Serial: 65916896770016886708751106294915943533
+# MD5 Fingerprint: 82:b2:ad:45:00:82:b0:66:63:f8:5f:c3:67:4e:ce:a3
+# SHA1 Fingerprint: a8:31:11:74:a6:14:15:0d:ca:77:dd:0e:e4:0c:5d:58:fc:a0:72:a5
+# SHA256 Fingerprint: be:f2:56:da:f2:6e:9c:69:bd:ec:16:02:35:97:98:f3:ca:f7:18:21:a0:3e:01:82:57:c5:3c:65:61:7f:3d:4a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA CYBER Root CA O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA CYBER Root CA O=TAIWAN-CA OU=Root CA
+# Label: "TWCA CYBER Root CA"
+# Serial: 85076849864375384482682434040119489222
+# MD5 Fingerprint: 0b:33:a0:97:52:95:d4:a9:fd:bb:db:6e:a3:55:5b:51
+# SHA1 Fingerprint: f6:b1:1c:1a:83:38:e9:7b:db:b3:a8:c8:33:24:e0:2d:9c:7f:26:66
+# SHA256 Fingerprint: 3f:63:bb:28:14:be:17:4e:c8:b6:43:9c:f0:8d:6d:56:f0:b7:c4:05:88:3a:56:48:a3:34:42:4d:6b:3e:c5:58
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA12 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA12 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA12"
+# Serial: 587887345431707215246142177076162061960426065942
+# MD5 Fingerprint: c6:89:ca:64:42:9b:62:08:49:0b:1e:7f:e9:07:3d:e8
+# SHA1 Fingerprint: 7a:22:1e:3d:de:1b:06:ac:9e:c8:47:70:16:8e:3c:e5:f7:6b:06:f4
+# SHA256 Fingerprint: 3f:03:4b:b5:70:4d:44:b2:d0:85:45:a0:20:57:de:93:eb:f3:90:5f:ce:72:1a:cb:c7:30:c0:6d:da:ee:90:4e
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA14 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA14 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA14"
+# Serial: 575790784512929437950770173562378038616896959179
+# MD5 Fingerprint: 71:0d:72:fa:92:19:65:5e:89:04:ac:16:33:f0:bc:d5
+# SHA1 Fingerprint: dd:50:c0:f7:79:b3:64:2e:74:a2:b8:9d:9f:d3:40:dd:bb:f0:f2:4f
+# SHA256 Fingerprint: 4b:00:9c:10:34:49:4f:9a:b5:6b:ba:3b:a1:d6:27:31:fc:4d:20:d8:95:5a:dc:ec:10:a9:25:60:72:61:e3:38
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA15 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA15 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA15"
+# Serial: 126083514594751269499665114766174399806381178503
+# MD5 Fingerprint: 13:30:fc:c4:62:a6:a9:de:b5:c1:68:af:b5:d2:31:47
+# SHA1 Fingerprint: cb:ba:83:c8:c1:5a:5d:f1:f9:73:6f:ca:d7:ef:28:13:06:4a:07:7d
+# SHA256 Fingerprint: e7:78:f0:f0:95:fe:84:37:29:cd:1a:00:82:17:9e:53:14:a9:c2:91:44:28:05:e1:fb:1d:8f:b6:b8:88:6c:3a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST BR Root CA 2 2023 O=D-Trust GmbH
+# Subject: CN=D-TRUST BR Root CA 2 2023 O=D-Trust GmbH
+# Label: "D-TRUST BR Root CA 2 2023"
+# Serial: 153168538924886464690566649552453098598
+# MD5 Fingerprint: e1:09:ed:d3:60:d4:56:1b:47:1f:b7:0c:5f:1b:5f:85
+# SHA1 Fingerprint: 2d:b0:70:ee:71:94:af:69:68:17:db:79:ce:58:9f:a0:6b:96:f7:87
+# SHA256 Fingerprint: 05:52:e6:f8:3f:df:65:e8:fa:96:70:e6:66:df:28:a4:e2:13:40:b5:10:cb:e5:25:66:f9:7c:4f:b9:4b:2b:d1
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST EV Root CA 2 2023 O=D-Trust GmbH
+# Subject: CN=D-TRUST EV Root CA 2 2023 O=D-Trust GmbH
+# Label: "D-TRUST EV Root CA 2 2023"
+# Serial: 139766439402180512324132425437959641711
+# MD5 Fingerprint: 96:b4:78:09:f0:09:cb:77:eb:bb:1b:4d:6f:36:bc:b6
+# SHA1 Fingerprint: a5:5b:d8:47:6c:8f:19:f7:4c:f4:6d:6b:b6:c2:79:82:22:df:54:8b
+# SHA256 Fingerprint: 8e:82:21:b2:e7:d4:00:78:36:a1:67:2f:0d:cc:29:9c:33:bc:07:d3:16:f1:32:fa:1a:20:6d:58:71:50:f1:ce
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/core.py b/backend/venv/lib/python3.12/site-packages/certifi/core.py
new file mode 100644
index 00000000..91f538bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/certifi/core.py
@@ -0,0 +1,114 @@
+"""
+certifi.py
+~~~~~~~~~~
+
+This module returns the installation location of cacert.pem or its contents.
+"""
+import sys
+import atexit
+
+def exit_cacert_ctx() -> None:
+    _CACERT_CTX.__exit__(None, None, None)  # type: ignore[union-attr]
+
+
+if sys.version_info >= (3, 11):
+
+    from importlib.resources import as_file, files
+
+    _CACERT_CTX = None
+    _CACERT_PATH = None
+
+    def where() -> str:
+        # This is slightly terrible, but we want to delay extracting the file
+        # in cases where we're inside of a zipimport situation until someone
+        # actually calls where(), but we don't want to re-extract the file
+        # on every call of where(), so we'll do it once then store it in a
+        # global variable.
+        global _CACERT_CTX
+        global _CACERT_PATH
+        if _CACERT_PATH is None:
+            # This is slightly janky, the importlib.resources API wants you to
+            # manage the cleanup of this file, so it doesn't actually return a
+            # path, it returns a context manager that will give you the path
+            # when you enter it and will do any cleanup when you leave it. In
+            # the common case of not needing a temporary file, it will just
+            # return the file system location and the __exit__() is a no-op.
+            #
+            # We also have to hold onto the actual context manager, because
+            # it will do the cleanup whenever it gets garbage collected, so
+            # we will also store that at the global level as well.
+            _CACERT_CTX = as_file(files("certifi").joinpath("cacert.pem"))
+            _CACERT_PATH = str(_CACERT_CTX.__enter__())
+            atexit.register(exit_cacert_ctx)
+
+        return _CACERT_PATH
+
+    def contents() -> str:
+        return files("certifi").joinpath("cacert.pem").read_text(encoding="ascii")
+
+elif sys.version_info >= (3, 7):
+
+    from importlib.resources import path as get_path, read_text
+
+    _CACERT_CTX = None
+    _CACERT_PATH = None
+
+    def where() -> str:
+        # This is slightly terrible, but we want to delay extracting the
+        # file in cases where we're inside of a zipimport situation until
+        # someone actually calls where(), but we don't want to re-extract
+        # the file on every call of where(), so we'll do it once then store
+        # it in a global variable.
+        global _CACERT_CTX
+        global _CACERT_PATH
+        if _CACERT_PATH is None:
+            # This is slightly janky, the importlib.resources API wants you
+            # to manage the cleanup of this file, so it doesn't actually
+            # return a path, it returns a context manager that will give
+            # you the path when you enter it and will do any cleanup when
+            # you leave it. In the common case of not needing a temporary
+            # file, it will just return the file system location and the
+            # __exit__() is a no-op.
+            #
+            # We also have to hold onto the actual context manager, because
+            # it will do the cleanup whenever it gets garbage collected, so
+            # we will also store that at the global level as well.
+            _CACERT_CTX = get_path("certifi", "cacert.pem")
+            _CACERT_PATH = str(_CACERT_CTX.__enter__())
+            atexit.register(exit_cacert_ctx)
+
+        return _CACERT_PATH
+
+    def contents() -> str:
+        return read_text("certifi", "cacert.pem", encoding="ascii")
+
+else:
+    import os
+    import types
+    from typing import Union
+
+    Package = Union[types.ModuleType, str]
+    Resource = Union[str, "os.PathLike"]
+
+    # This fallback will work for Python versions prior to 3.7 that lack the
+    # importlib.resources module but relies on the existing `where` function
+    # so won't address issues with environments like PyOxidizer that don't set
+    # __file__ on modules.
+    def read_text(
+        package: Package,
+        resource: Resource,
+        encoding: str = 'utf-8',
+        errors: str = 'strict'
+    ) -> str:
+        with open(where(), encoding=encoding) as data:
+            return data.read()
+
+    # If we don't have importlib.resources, then we will just do the old logic
+    # of assuming we're on the filesystem and munge the path directly.
+    def where() -> str:
+        f = os.path.dirname(__file__)
+
+        return os.path.join(f, "cacert.pem")
+
+    def contents() -> str:
+        return read_text("certifi", "cacert.pem", encoding="ascii")
diff --git a/backend/venv/lib/python3.12/site-packages/certifi/py.typed b/backend/venv/lib/python3.12/site-packages/certifi/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/LICENSE
new file mode 100644
index 00000000..9725772c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 TAHRI Ahmed R.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/METADATA
new file mode 100644
index 00000000..58f3667d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/METADATA
@@ -0,0 +1,721 @@
+Metadata-Version: 2.1
+Name: charset-normalizer
+Version: 3.4.1
+Summary: The Real First Universal Charset Detector. Open, modern and actively maintained alternative to Chardet.
+Author-email: "Ahmed R. TAHRI" <tahri.ahmed@proton.me>
+Maintainer-email: "Ahmed R. TAHRI" <tahri.ahmed@proton.me>
+License: MIT
+Project-URL: Changelog, https://github.com/jawah/charset_normalizer/blob/master/CHANGELOG.md
+Project-URL: Documentation, https://charset-normalizer.readthedocs.io/
+Project-URL: Code, https://github.com/jawah/charset_normalizer
+Project-URL: Issue tracker, https://github.com/jawah/charset_normalizer/issues
+Keywords: encoding,charset,charset-detector,detector,normalization,unicode,chardet,detect
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Text Processing :: Linguistic
+Classifier: Topic :: Utilities
+Classifier: Typing :: Typed
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: unicode-backport
+
+<h1 align="center">Charset Detection, for Everyone 👋</h1>
+
+<p align="center">
+  <sup>The Real First Universal Charset Detector</sup><br>
+  <a href="https://pypi.org/project/charset-normalizer">
+    <img src="https://img.shields.io/pypi/pyversions/charset_normalizer.svg?orange=blue" />
+  </a>
+  <a href="https://pepy.tech/project/charset-normalizer/">
+    <img alt="Download Count Total" src="https://static.pepy.tech/badge/charset-normalizer/month" />
+  </a>
+  <a href="https://bestpractices.coreinfrastructure.org/projects/7297">
+    <img src="https://bestpractices.coreinfrastructure.org/projects/7297/badge">
+  </a>
+</p>
+<p align="center">
+  <sup><i>Featured Packages</i></sup><br>
+  <a href="https://github.com/jawah/niquests">
+   <img alt="Static Badge" src="https://img.shields.io/badge/Niquests-Best_HTTP_Client-cyan">
+  </a>
+  <a href="https://github.com/jawah/wassima">
+   <img alt="Static Badge" src="https://img.shields.io/badge/Wassima-Certifi_Killer-cyan">
+  </a>
+</p>
+<p align="center">
+  <sup><i>In other language (unofficial port - by the community)</i></sup><br>
+  <a href="https://github.com/nickspring/charset-normalizer-rs">
+   <img alt="Static Badge" src="https://img.shields.io/badge/Rust-red">
+  </a>
+</p>
+
+> A library that helps you read text from an unknown charset encoding.<br /> Motivated by `chardet`,
+> I'm trying to resolve the issue by taking a new approach.
+> All IANA character set names for which the Python core library provides codecs are supported.
+
+<p align="center">
+  >>>>> <a href="https://charsetnormalizerweb.ousret.now.sh" target="_blank">👉 Try Me Online Now, Then Adopt Me 👈 </a> <<<<<
+</p>
+
+This project offers you an alternative to **Universal Charset Encoding Detector**, also known as **Chardet**.
+
+| Feature                                          | [Chardet](https://github.com/chardet/chardet) |                                         Charset Normalizer                                         | [cChardet](https://github.com/PyYoshi/cChardet) |
+|--------------------------------------------------|:---------------------------------------------:|:--------------------------------------------------------------------------------------------------:|:-----------------------------------------------:|
+| `Fast`                                           |                       ❌                       |                                                 ✅                                                  |                        ✅                        |
+| `Universal**`                                    |                       ❌                       |                                                 ✅                                                  |                        ❌                        |
+| `Reliable` **without** distinguishable standards |                       ❌                       |                                                 ✅                                                  |                        ✅                        |
+| `Reliable` **with** distinguishable standards    |                       ✅                       |                                                 ✅                                                  |                        ✅                        |
+| `License`                                        |           LGPL-2.1<br>_restrictive_           |                                                MIT                                                 |            MPL-1.1<br>_restrictive_             |
+| `Native Python`                                  |                       ✅                       |                                                 ✅                                                  |                        ❌                        |
+| `Detect spoken language`                         |                       ❌                       |                                                 ✅                                                  |                       N/A                       |
+| `UnicodeDecodeError Safety`                      |                       ❌                       |                                                 ✅                                                  |                        ❌                        |
+| `Whl Size (min)`                                 |                   193.6 kB                    |                                               42 kB                                                |                     ~200 kB                     |
+| `Supported Encoding`                             |                      33                       | 🎉 [99](https://charset-normalizer.readthedocs.io/en/latest/user/support.html#supported-encodings) |                       40                        |
+
+<p align="center">
+<img src="https://i.imgflip.com/373iay.gif" alt="Reading Normalized Text" width="226"/><img src="https://media.tenor.com/images/c0180f70732a18b4965448d33adba3d0/tenor.gif" alt="Cat Reading Text" width="200"/>
+</p>
+
+*\*\* : They are clearly using specific code for a specific encoding even if covering most of used one*<br>
+
+## ⚡ Performance
+
+This package offer better performance than its counterpart Chardet. Here are some numbers.
+
+| Package                                       | Accuracy | Mean per file (ms) | File per sec (est) |
+|-----------------------------------------------|:--------:|:------------------:|:------------------:|
+| [chardet](https://github.com/chardet/chardet) |   86 %   |       63 ms        |    16 file/sec     |
+| charset-normalizer                            | **98 %** |     **10 ms**      |    100 file/sec    |
+
+| Package                                       | 99th percentile | 95th percentile | 50th percentile |
+|-----------------------------------------------|:---------------:|:---------------:|:---------------:|
+| [chardet](https://github.com/chardet/chardet) |     265 ms      |      71 ms      |      7 ms       |
+| charset-normalizer                            |     100 ms      |      50 ms      |      5 ms       |
+
+_updated as of december 2024 using CPython 3.12_
+
+Chardet's performance on larger file (1MB+) are very poor. Expect huge difference on large payload.
+
+> Stats are generated using 400+ files using default parameters. More details on used files, see GHA workflows.
+> And yes, these results might change at any time. The dataset can be updated to include more files.
+> The actual delays heavily depends on your CPU capabilities. The factors should remain the same.
+> Keep in mind that the stats are generous and that Chardet accuracy vs our is measured using Chardet initial capability
+> (e.g. Supported Encoding) Challenge-them if you want.
+
+## ✨ Installation
+
+Using pip:
+
+```sh
+pip install charset-normalizer -U
+```
+
+## 🚀 Basic Usage
+
+### CLI
+This package comes with a CLI.
+
+```
+usage: normalizer [-h] [-v] [-a] [-n] [-m] [-r] [-f] [-t THRESHOLD]
+                  file [file ...]
+
+The Real First Universal Charset Detector. Discover originating encoding used
+on text file. Normalize text to unicode.
+
+positional arguments:
+  files                 File(s) to be analysed
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -v, --verbose         Display complementary information about file if any.
+                        Stdout will contain logs about the detection process.
+  -a, --with-alternative
+                        Output complementary possibilities if any. Top-level
+                        JSON WILL be a list.
+  -n, --normalize       Permit to normalize input file. If not set, program
+                        does not write anything.
+  -m, --minimal         Only output the charset detected to STDOUT. Disabling
+                        JSON output.
+  -r, --replace         Replace file when trying to normalize it instead of
+                        creating a new one.
+  -f, --force           Replace file without asking if you are sure, use this
+                        flag with caution.
+  -t THRESHOLD, --threshold THRESHOLD
+                        Define a custom maximum amount of chaos allowed in
+                        decoded content. 0. <= chaos <= 1.
+  --version             Show version information and exit.
+```
+
+```bash
+normalizer ./data/sample.1.fr.srt
+```
+
+or
+
+```bash
+python -m charset_normalizer ./data/sample.1.fr.srt
+```
+
+🎉 Since version 1.4.0 the CLI produce easily usable stdout result in JSON format.
+
+```json
+{
+    "path": "/home/default/projects/charset_normalizer/data/sample.1.fr.srt",
+    "encoding": "cp1252",
+    "encoding_aliases": [
+        "1252",
+        "windows_1252"
+    ],
+    "alternative_encodings": [
+        "cp1254",
+        "cp1256",
+        "cp1258",
+        "iso8859_14",
+        "iso8859_15",
+        "iso8859_16",
+        "iso8859_3",
+        "iso8859_9",
+        "latin_1",
+        "mbcs"
+    ],
+    "language": "French",
+    "alphabets": [
+        "Basic Latin",
+        "Latin-1 Supplement"
+    ],
+    "has_sig_or_bom": false,
+    "chaos": 0.149,
+    "coherence": 97.152,
+    "unicode_path": null,
+    "is_preferred": true
+}
+```
+
+### Python
+*Just print out normalized text*
+```python
+from charset_normalizer import from_path
+
+results = from_path('./my_subtitle.srt')
+
+print(str(results.best()))
+```
+
+*Upgrade your code without effort*
+```python
+from charset_normalizer import detect
+```
+
+The above code will behave the same as **chardet**. We ensure that we offer the best (reasonable) BC result possible.
+
+See the docs for advanced usage : [readthedocs.io](https://charset-normalizer.readthedocs.io/en/latest/)
+
+## 😇 Why
+
+When I started using Chardet, I noticed that it was not suited to my expectations, and I wanted to propose a
+reliable alternative using a completely different method. Also! I never back down on a good challenge!
+
+I **don't care** about the **originating charset** encoding, because **two different tables** can
+produce **two identical rendered string.**
+What I want is to get readable text, the best I can.
+
+In a way, **I'm brute forcing text decoding.** How cool is that ? 😎
+
+Don't confuse package **ftfy** with charset-normalizer or chardet. ftfy goal is to repair Unicode string whereas charset-normalizer to convert raw file in unknown encoding to unicode.
+
+## 🍰 How
+
+  - Discard all charset encoding table that could not fit the binary content.
+  - Measure noise, or the mess once opened (by chunks) with a corresponding charset encoding.
+  - Extract matches with the lowest mess detected.
+  - Additionally, we measure coherence / probe for a language.
+
+**Wait a minute**, what is noise/mess and coherence according to **YOU ?**
+
+*Noise :* I opened hundred of text files, **written by humans**, with the wrong encoding table. **I observed**, then
+**I established** some ground rules about **what is obvious** when **it seems like** a mess (aka. defining noise in rendered text).
+ I know that my interpretation of what is noise is probably incomplete, feel free to contribute in order to
+ improve or rewrite it.
+
+*Coherence :* For each language there is on earth, we have computed ranked letter appearance occurrences (the best we can). So I thought
+that intel is worth something here. So I use those records against decoded text to check if I can detect intelligent design.
+
+## ⚡ Known limitations
+
+  - Language detection is unreliable when text contains two or more languages sharing identical letters. (eg. HTML (english tags) + Turkish content (Sharing Latin characters))
+  - Every charset detector heavily depends on sufficient content. In common cases, do not bother run detection on very tiny content.
+
+## ⚠️ About Python EOLs
+
+**If you are running:**
+
+- Python >=2.7,<3.5: Unsupported
+- Python 3.5: charset-normalizer < 2.1
+- Python 3.6: charset-normalizer < 3.1
+- Python 3.7: charset-normalizer < 4.0
+
+Upgrade your Python interpreter as soon as possible.
+
+## 👤 Contributing
+
+Contributions, issues and feature requests are very much welcome.<br />
+Feel free to check [issues page](https://github.com/ousret/charset_normalizer/issues) if you want to contribute.
+
+## 📝 License
+
+Copyright © [Ahmed TAHRI @Ousret](https://github.com/Ousret).<br />
+This project is [MIT](https://github.com/Ousret/charset_normalizer/blob/master/LICENSE) licensed.
+
+Characters frequencies used in this project © 2012 [Denny Vrandečić](http://simia.net/letters/)
+
+## 💼 For Enterprise
+
+Professional support for charset-normalizer is available as part of the [Tidelift
+Subscription][1]. Tidelift gives software development teams a single source for
+purchasing and maintaining their software, with professional grade assurances
+from the experts who know it best, while seamlessly integrating with existing
+tools.
+
+[1]: https://tidelift.com/subscription/pkg/pypi-charset-normalizer?utm_source=pypi-charset-normalizer&utm_medium=readme
+
+[![OpenSSF Best Practices](https://www.bestpractices.dev/projects/7297/badge)](https://www.bestpractices.dev/projects/7297)
+
+# Changelog
+All notable changes to charset-normalizer will be documented in this file. This project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
+
+## [3.4.1](https://github.com/Ousret/charset_normalizer/compare/3.4.0...3.4.1) (2024-12-24)
+
+### Changed
+- Project metadata are now stored using `pyproject.toml` instead of `setup.cfg` using setuptools as the build backend.
+- Enforce annotation delayed loading for a simpler and consistent types in the project.
+- Optional mypyc compilation upgraded to version 1.14 for Python >= 3.8
+
+### Added
+- pre-commit configuration.
+- noxfile.
+
+### Removed
+- `build-requirements.txt` as per using `pyproject.toml` native build configuration.
+- `bin/integration.py` and `bin/serve.py` in favor of downstream integration test (see noxfile).
+- `setup.cfg` in favor of `pyproject.toml` metadata configuration.
+- Unused `utils.range_scan` function.
+
+### Fixed
+- Converting content to Unicode bytes may insert `utf_8` instead of preferred `utf-8`. (#572)
+- Deprecation warning "'count' is passed as positional argument" when converting to Unicode bytes on Python 3.13+
+
+## [3.4.0](https://github.com/Ousret/charset_normalizer/compare/3.3.2...3.4.0) (2024-10-08)
+
+### Added
+- Argument `--no-preemptive` in the CLI to prevent the detector to search for hints.
+- Support for Python 3.13 (#512)
+
+### Fixed
+- Relax the TypeError exception thrown when trying to compare a CharsetMatch with anything else than a CharsetMatch.
+- Improved the general reliability of the detector based on user feedbacks. (#520) (#509) (#498) (#407) (#537)
+- Declared charset in content (preemptive detection) not changed when converting to utf-8 bytes. (#381)
+
+## [3.3.2](https://github.com/Ousret/charset_normalizer/compare/3.3.1...3.3.2) (2023-10-31)
+
+### Fixed
+- Unintentional memory usage regression when using large payload that match several encoding (#376)
+- Regression on some detection case showcased in the documentation (#371)
+
+### Added
+- Noise (md) probe that identify malformed arabic representation due to the presence of letters in isolated form (credit to my wife)
+
+## [3.3.1](https://github.com/Ousret/charset_normalizer/compare/3.3.0...3.3.1) (2023-10-22)
+
+### Changed
+- Optional mypyc compilation upgraded to version 1.6.1 for Python >= 3.8
+- Improved the general detection reliability based on reports from the community
+
+## [3.3.0](https://github.com/Ousret/charset_normalizer/compare/3.2.0...3.3.0) (2023-09-30)
+
+### Added
+- Allow to execute the CLI (e.g. normalizer) through `python -m charset_normalizer.cli` or `python -m charset_normalizer`
+- Support for 9 forgotten encoding that are supported by Python but unlisted in `encoding.aliases` as they have no alias (#323)
+
+### Removed
+- (internal) Redundant utils.is_ascii function and unused function is_private_use_only
+- (internal) charset_normalizer.assets is moved inside charset_normalizer.constant
+
+### Changed
+- (internal) Unicode code blocks in constants are updated using the latest v15.0.0 definition to improve detection
+- Optional mypyc compilation upgraded to version 1.5.1 for Python >= 3.8
+
+### Fixed
+- Unable to properly sort CharsetMatch when both chaos/noise and coherence were close due to an unreachable condition in \_\_lt\_\_ (#350)
+
+## [3.2.0](https://github.com/Ousret/charset_normalizer/compare/3.1.0...3.2.0) (2023-06-07)
+
+### Changed
+- Typehint for function `from_path` no longer enforce `PathLike` as its first argument
+- Minor improvement over the global detection reliability
+
+### Added
+- Introduce function `is_binary` that relies on main capabilities, and optimized to detect binaries
+- Propagate `enable_fallback` argument throughout `from_bytes`, `from_path`, and `from_fp` that allow a deeper control over the detection (default True)
+- Explicit support for Python 3.12
+
+### Fixed
+- Edge case detection failure where a file would contain 'very-long' camel cased word (Issue #289)
+
+## [3.1.0](https://github.com/Ousret/charset_normalizer/compare/3.0.1...3.1.0) (2023-03-06)
+
+### Added
+- Argument `should_rename_legacy` for legacy function `detect` and disregard any new arguments without errors (PR #262)
+
+### Removed
+- Support for Python 3.6 (PR #260)
+
+### Changed
+- Optional speedup provided by mypy/c 1.0.1
+
+## [3.0.1](https://github.com/Ousret/charset_normalizer/compare/3.0.0...3.0.1) (2022-11-18)
+
+### Fixed
+- Multi-bytes cutter/chunk generator did not always cut correctly (PR #233)
+
+### Changed
+- Speedup provided by mypy/c 0.990 on Python >= 3.7
+
+## [3.0.0](https://github.com/Ousret/charset_normalizer/compare/2.1.1...3.0.0) (2022-10-20)
+
+### Added
+- Extend the capability of explain=True when cp_isolation contains at most two entries (min one), will log in details of the Mess-detector results
+- Support for alternative language frequency set in charset_normalizer.assets.FREQUENCIES
+- Add parameter `language_threshold` in `from_bytes`, `from_path` and `from_fp` to adjust the minimum expected coherence ratio
+- `normalizer --version` now specify if current version provide extra speedup (meaning mypyc compilation whl)
+
+### Changed
+- Build with static metadata using 'build' frontend
+- Make the language detection stricter
+- Optional: Module `md.py` can be compiled using Mypyc to provide an extra speedup up to 4x faster than v2.1
+
+### Fixed
+- CLI with opt --normalize fail when using full path for files
+- TooManyAccentuatedPlugin induce false positive on the mess detection when too few alpha character have been fed to it
+- Sphinx warnings when generating the documentation
+
+### Removed
+- Coherence detector no longer return 'Simple English' instead return 'English'
+- Coherence detector no longer return 'Classical Chinese' instead return 'Chinese'
+- Breaking: Method `first()` and `best()` from CharsetMatch
+- UTF-7 will no longer appear as "detected" without a recognized SIG/mark (is unreliable/conflict with ASCII)
+- Breaking: Class aliases CharsetDetector, CharsetDoctor, CharsetNormalizerMatch and CharsetNormalizerMatches
+- Breaking: Top-level function `normalize`
+- Breaking: Properties `chaos_secondary_pass`, `coherence_non_latin` and `w_counter` from CharsetMatch
+- Support for the backport `unicodedata2`
+
+## [3.0.0rc1](https://github.com/Ousret/charset_normalizer/compare/3.0.0b2...3.0.0rc1) (2022-10-18)
+
+### Added
+- Extend the capability of explain=True when cp_isolation contains at most two entries (min one), will log in details of the Mess-detector results
+- Support for alternative language frequency set in charset_normalizer.assets.FREQUENCIES
+- Add parameter `language_threshold` in `from_bytes`, `from_path` and `from_fp` to adjust the minimum expected coherence ratio
+
+### Changed
+- Build with static metadata using 'build' frontend
+- Make the language detection stricter
+
+### Fixed
+- CLI with opt --normalize fail when using full path for files
+- TooManyAccentuatedPlugin induce false positive on the mess detection when too few alpha character have been fed to it
+
+### Removed
+- Coherence detector no longer return 'Simple English' instead return 'English'
+- Coherence detector no longer return 'Classical Chinese' instead return 'Chinese'
+
+## [3.0.0b2](https://github.com/Ousret/charset_normalizer/compare/3.0.0b1...3.0.0b2) (2022-08-21)
+
+### Added
+- `normalizer --version` now specify if current version provide extra speedup (meaning mypyc compilation whl)
+
+### Removed
+- Breaking: Method `first()` and `best()` from CharsetMatch
+- UTF-7 will no longer appear as "detected" without a recognized SIG/mark (is unreliable/conflict with ASCII)
+
+### Fixed
+- Sphinx warnings when generating the documentation
+
+## [3.0.0b1](https://github.com/Ousret/charset_normalizer/compare/2.1.0...3.0.0b1) (2022-08-15)
+
+### Changed
+- Optional: Module `md.py` can be compiled using Mypyc to provide an extra speedup up to 4x faster than v2.1
+
+### Removed
+- Breaking: Class aliases CharsetDetector, CharsetDoctor, CharsetNormalizerMatch and CharsetNormalizerMatches
+- Breaking: Top-level function `normalize`
+- Breaking: Properties `chaos_secondary_pass`, `coherence_non_latin` and `w_counter` from CharsetMatch
+- Support for the backport `unicodedata2`
+
+## [2.1.1](https://github.com/Ousret/charset_normalizer/compare/2.1.0...2.1.1) (2022-08-19)
+
+### Deprecated
+- Function `normalize` scheduled for removal in 3.0
+
+### Changed
+- Removed useless call to decode in fn is_unprintable (#206)
+
+### Fixed
+- Third-party library (i18n xgettext) crashing not recognizing utf_8 (PEP 263) with underscore from [@aleksandernovikov](https://github.com/aleksandernovikov) (#204)
+
+## [2.1.0](https://github.com/Ousret/charset_normalizer/compare/2.0.12...2.1.0) (2022-06-19)
+
+### Added
+- Output the Unicode table version when running the CLI with `--version` (PR #194)
+
+### Changed
+- Re-use decoded buffer for single byte character sets from [@nijel](https://github.com/nijel) (PR #175)
+- Fixing some performance bottlenecks from [@deedy5](https://github.com/deedy5) (PR #183)
+
+### Fixed
+- Workaround potential bug in cpython with Zero Width No-Break Space located in Arabic Presentation Forms-B, Unicode 1.1 not acknowledged as space (PR #175)
+- CLI default threshold aligned with the API threshold from [@oleksandr-kuzmenko](https://github.com/oleksandr-kuzmenko) (PR #181)
+
+### Removed
+- Support for Python 3.5 (PR #192)
+
+### Deprecated
+- Use of backport unicodedata from `unicodedata2` as Python is quickly catching up, scheduled for removal in 3.0 (PR #194)
+
+## [2.0.12](https://github.com/Ousret/charset_normalizer/compare/2.0.11...2.0.12) (2022-02-12)
+
+### Fixed
+- ASCII miss-detection on rare cases (PR #170)
+
+## [2.0.11](https://github.com/Ousret/charset_normalizer/compare/2.0.10...2.0.11) (2022-01-30)
+
+### Added
+- Explicit support for Python 3.11 (PR #164)
+
+### Changed
+- The logging behavior have been completely reviewed, now using only TRACE and DEBUG levels (PR #163 #165)
+
+## [2.0.10](https://github.com/Ousret/charset_normalizer/compare/2.0.9...2.0.10) (2022-01-04)
+
+### Fixed
+- Fallback match entries might lead to UnicodeDecodeError for large bytes sequence (PR #154)
+
+### Changed
+- Skipping the language-detection (CD) on ASCII (PR #155)
+
+## [2.0.9](https://github.com/Ousret/charset_normalizer/compare/2.0.8...2.0.9) (2021-12-03)
+
+### Changed
+- Moderating the logging impact (since 2.0.8) for specific environments (PR #147)
+
+### Fixed
+- Wrong logging level applied when setting kwarg `explain` to True (PR #146)
+
+## [2.0.8](https://github.com/Ousret/charset_normalizer/compare/2.0.7...2.0.8) (2021-11-24)
+### Changed
+- Improvement over Vietnamese detection (PR #126)
+- MD improvement on trailing data and long foreign (non-pure latin) data (PR #124)
+- Efficiency improvements in cd/alphabet_languages from [@adbar](https://github.com/adbar) (PR #122)
+- call sum() without an intermediary list following PEP 289 recommendations from [@adbar](https://github.com/adbar) (PR #129)
+- Code style as refactored by Sourcery-AI (PR #131)
+- Minor adjustment on the MD around european words (PR #133)
+- Remove and replace SRTs from assets / tests (PR #139)
+- Initialize the library logger with a `NullHandler` by default from [@nmaynes](https://github.com/nmaynes) (PR #135)
+- Setting kwarg `explain` to True will add provisionally (bounded to function lifespan) a specific stream handler (PR #135)
+
+### Fixed
+- Fix large (misleading) sequence giving UnicodeDecodeError (PR #137)
+- Avoid using too insignificant chunk (PR #137)
+
+### Added
+- Add and expose function `set_logging_handler` to configure a specific StreamHandler from [@nmaynes](https://github.com/nmaynes) (PR #135)
+- Add `CHANGELOG.md` entries, format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/) (PR #141)
+
+## [2.0.7](https://github.com/Ousret/charset_normalizer/compare/2.0.6...2.0.7) (2021-10-11)
+### Added
+- Add support for Kazakh (Cyrillic) language detection (PR #109)
+
+### Changed
+- Further, improve inferring the language from a given single-byte code page (PR #112)
+- Vainly trying to leverage PEP263 when PEP3120 is not supported (PR #116)
+- Refactoring for potential performance improvements in loops from [@adbar](https://github.com/adbar) (PR #113)
+- Various detection improvement (MD+CD) (PR #117)
+
+### Removed
+- Remove redundant logging entry about detected language(s) (PR #115)
+
+### Fixed
+- Fix a minor inconsistency between Python 3.5 and other versions regarding language detection (PR #117 #102)
+
+## [2.0.6](https://github.com/Ousret/charset_normalizer/compare/2.0.5...2.0.6) (2021-09-18)
+### Fixed
+- Unforeseen regression with the loss of the backward-compatibility with some older minor of Python 3.5.x (PR #100)
+- Fix CLI crash when using --minimal output in certain cases (PR #103)
+
+### Changed
+- Minor improvement to the detection efficiency (less than 1%) (PR #106 #101)
+
+## [2.0.5](https://github.com/Ousret/charset_normalizer/compare/2.0.4...2.0.5) (2021-09-14)
+### Changed
+- The project now comply with: flake8, mypy, isort and black to ensure a better overall quality (PR #81)
+- The BC-support with v1.x was improved, the old staticmethods are restored (PR #82)
+- The Unicode detection is slightly improved (PR #93)
+- Add syntax sugar \_\_bool\_\_ for results CharsetMatches list-container (PR #91)
+
+### Removed
+- The project no longer raise warning on tiny content given for detection, will be simply logged as warning instead (PR #92)
+
+### Fixed
+- In some rare case, the chunks extractor could cut in the middle of a multi-byte character and could mislead the mess detection (PR #95)
+- Some rare 'space' characters could trip up the UnprintablePlugin/Mess detection (PR #96)
+- The MANIFEST.in was not exhaustive (PR #78)
+
+## [2.0.4](https://github.com/Ousret/charset_normalizer/compare/2.0.3...2.0.4) (2021-07-30)
+### Fixed
+- The CLI no longer raise an unexpected exception when no encoding has been found (PR #70)
+- Fix accessing the 'alphabets' property when the payload contains surrogate characters (PR #68)
+- The logger could mislead (explain=True) on detected languages and the impact of one MBCS match (PR #72)
+- Submatch factoring could be wrong in rare edge cases (PR #72)
+- Multiple files given to the CLI were ignored when publishing results to STDOUT. (After the first path) (PR #72)
+- Fix line endings from CRLF to LF for certain project files (PR #67)
+
+### Changed
+- Adjust the MD to lower the sensitivity, thus improving the global detection reliability (PR #69 #76)
+- Allow fallback on specified encoding if any (PR #71)
+
+## [2.0.3](https://github.com/Ousret/charset_normalizer/compare/2.0.2...2.0.3) (2021-07-16)
+### Changed
+- Part of the detection mechanism has been improved to be less sensitive, resulting in more accurate detection results. Especially ASCII. (PR #63)
+- According to the community wishes, the detection will fall back on ASCII or UTF-8 in a last-resort case. (PR #64)
+
+## [2.0.2](https://github.com/Ousret/charset_normalizer/compare/2.0.1...2.0.2) (2021-07-15)
+### Fixed
+- Empty/Too small JSON payload miss-detection fixed. Report from [@tseaver](https://github.com/tseaver) (PR #59)
+
+### Changed
+- Don't inject unicodedata2 into sys.modules from [@akx](https://github.com/akx) (PR #57)
+
+## [2.0.1](https://github.com/Ousret/charset_normalizer/compare/2.0.0...2.0.1) (2021-07-13)
+### Fixed
+- Make it work where there isn't a filesystem available, dropping assets frequencies.json. Report from [@sethmlarson](https://github.com/sethmlarson). (PR #55)
+- Using explain=False permanently disable the verbose output in the current runtime (PR #47)
+- One log entry (language target preemptive) was not show in logs when using explain=True (PR #47)
+- Fix undesired exception (ValueError) on getitem of instance CharsetMatches (PR #52)
+
+### Changed
+- Public function normalize default args values were not aligned with from_bytes (PR #53)
+
+### Added
+- You may now use charset aliases in cp_isolation and cp_exclusion arguments (PR #47)
+
+## [2.0.0](https://github.com/Ousret/charset_normalizer/compare/1.4.1...2.0.0) (2021-07-02)
+### Changed
+- 4x to 5 times faster than the previous 1.4.0 release. At least 2x faster than Chardet.
+- Accent has been made on UTF-8 detection, should perform rather instantaneous.
+- The backward compatibility with Chardet has been greatly improved. The legacy detect function returns an identical charset name whenever possible.
+- The detection mechanism has been slightly improved, now Turkish content is detected correctly (most of the time)
+- The program has been rewritten to ease the readability and maintainability. (+Using static typing)+
+- utf_7 detection has been reinstated.
+
+### Removed
+- This package no longer require anything when used with Python 3.5 (Dropped cached_property)
+- Removed support for these languages: Catalan, Esperanto, Kazakh, Baque, Volapük, Azeri, Galician, Nynorsk, Macedonian, and Serbocroatian.
+- The exception hook on UnicodeDecodeError has been removed.
+
+### Deprecated
+- Methods coherence_non_latin, w_counter, chaos_secondary_pass of the class CharsetMatch are now deprecated and scheduled for removal in v3.0
+
+### Fixed
+- The CLI output used the relative path of the file(s). Should be absolute.
+
+## [1.4.1](https://github.com/Ousret/charset_normalizer/compare/1.4.0...1.4.1) (2021-05-28)
+### Fixed
+- Logger configuration/usage no longer conflict with others (PR #44)
+
+## [1.4.0](https://github.com/Ousret/charset_normalizer/compare/1.3.9...1.4.0) (2021-05-21)
+### Removed
+- Using standard logging instead of using the package loguru.
+- Dropping nose test framework in favor of the maintained pytest.
+- Choose to not use dragonmapper package to help with gibberish Chinese/CJK text.
+- Require cached_property only for Python 3.5 due to constraint. Dropping for every other interpreter version.
+- Stop support for UTF-7 that does not contain a SIG.
+- Dropping PrettyTable, replaced with pure JSON output in CLI.
+
+### Fixed
+- BOM marker in a CharsetNormalizerMatch instance could be False in rare cases even if obviously present. Due to the sub-match factoring process.
+- Not searching properly for the BOM when trying utf32/16 parent codec.
+
+### Changed
+- Improving the package final size by compressing frequencies.json.
+- Huge improvement over the larges payload.
+
+### Added
+- CLI now produces JSON consumable output.
+- Return ASCII if given sequences fit. Given reasonable confidence.
+
+## [1.3.9](https://github.com/Ousret/charset_normalizer/compare/1.3.8...1.3.9) (2021-05-13)
+
+### Fixed
+- In some very rare cases, you may end up getting encode/decode errors due to a bad bytes payload (PR #40)
+
+## [1.3.8](https://github.com/Ousret/charset_normalizer/compare/1.3.7...1.3.8) (2021-05-12)
+
+### Fixed
+- Empty given payload for detection may cause an exception if trying to access the `alphabets` property. (PR #39)
+
+## [1.3.7](https://github.com/Ousret/charset_normalizer/compare/1.3.6...1.3.7) (2021-05-12)
+
+### Fixed
+- The legacy detect function should return UTF-8-SIG if sig is present in the payload. (PR #38)
+
+## [1.3.6](https://github.com/Ousret/charset_normalizer/compare/1.3.5...1.3.6) (2021-02-09)
+
+### Changed
+- Amend the previous release to allow prettytable 2.0 (PR #35)
+
+## [1.3.5](https://github.com/Ousret/charset_normalizer/compare/1.3.4...1.3.5) (2021-02-08)
+
+### Fixed
+- Fix error while using the package with a python pre-release interpreter (PR #33)
+
+### Changed
+- Dependencies refactoring, constraints revised.
+
+### Added
+- Add python 3.9 and 3.10 to the supported interpreters
+
+MIT License
+
+Copyright (c) 2025 TAHRI Ahmed R.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/RECORD
new file mode 100644
index 00000000..67606321
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/RECORD
@@ -0,0 +1,35 @@
+../../../bin/normalizer,sha256=HE_Bau9Tpn8JzlJESHpdWa_BoHp56xVlp9eKtNH7ppA,283
+charset_normalizer-3.4.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+charset_normalizer-3.4.1.dist-info/LICENSE,sha256=bQ1Bv-FwrGx9wkjJpj4lTQ-0WmDVCoJX0K-SxuJJuIc,1071
+charset_normalizer-3.4.1.dist-info/METADATA,sha256=JbyHzhmqZh_ugEn1Y7TY7CDYZA9FoU6BP25hrCNDf50,35313
+charset_normalizer-3.4.1.dist-info/RECORD,,
+charset_normalizer-3.4.1.dist-info/WHEEL,sha256=w9Wib2Cggr6a0jsmQwU18YkLmVsyOqV25G3M7UiDLV0,115
+charset_normalizer-3.4.1.dist-info/entry_points.txt,sha256=8C-Y3iXIfyXQ83Tpir2B8t-XLJYpxF5xbb38d_js-h4,65
+charset_normalizer-3.4.1.dist-info/top_level.txt,sha256=7ASyzePr8_xuZWJsnqJjIBtyV8vhEo0wBCv1MPRRi3Q,19
+charset_normalizer/__init__.py,sha256=OKRxRv2Zhnqk00tqkN0c1BtJjm165fWXLydE52IKuHc,1590
+charset_normalizer/__main__.py,sha256=yzYxMR-IhKRHYwcSlavEv8oGdwxsR89mr2X09qXGdps,109
+charset_normalizer/__pycache__/__init__.cpython-312.pyc,,
+charset_normalizer/__pycache__/__main__.cpython-312.pyc,,
+charset_normalizer/__pycache__/api.cpython-312.pyc,,
+charset_normalizer/__pycache__/cd.cpython-312.pyc,,
+charset_normalizer/__pycache__/constant.cpython-312.pyc,,
+charset_normalizer/__pycache__/legacy.cpython-312.pyc,,
+charset_normalizer/__pycache__/md.cpython-312.pyc,,
+charset_normalizer/__pycache__/models.cpython-312.pyc,,
+charset_normalizer/__pycache__/utils.cpython-312.pyc,,
+charset_normalizer/__pycache__/version.cpython-312.pyc,,
+charset_normalizer/api.py,sha256=qBRz8mJ_R5E713R6TOyqHEdnmyxbEDnCSHvx32ubDGg,22617
+charset_normalizer/cd.py,sha256=WKTo1HDb-H9HfCDc3Bfwq5jzS25Ziy9SE2a74SgTq88,12522
+charset_normalizer/cli/__init__.py,sha256=D8I86lFk2-py45JvqxniTirSj_sFyE6sjaY_0-G1shc,136
+charset_normalizer/cli/__main__.py,sha256=VGC9klOoi6_R2z8rmyrc936kv7u2A1udjjHtlmNPDTM,10410
+charset_normalizer/cli/__pycache__/__init__.cpython-312.pyc,,
+charset_normalizer/cli/__pycache__/__main__.cpython-312.pyc,,
+charset_normalizer/constant.py,sha256=4VuTcZNLew1j_8ixA-Rt_VVqNWD4pwgHOHMCMlr0964,40477
+charset_normalizer/legacy.py,sha256=yhNXsPHkBfqPXKRb-sPXNj3Bscp9-mFGcYOkJ62tg9c,2328
+charset_normalizer/md.cpython-312-darwin.so,sha256=mUvSZMr6ty3_yLX68YZ5BJQr91RHfsfhiJCWQCFBMnE,115664
+charset_normalizer/md.py,sha256=iyXXQGWl54nnLQLueMWTmUtlivO0-rTBgVkmJxIIAGU,20036
+charset_normalizer/md__mypyc.cpython-312-darwin.so,sha256=YjKTRj4XrLjhCIZIecqc47jwxEK9cYiYZVqT1A59L0w,482088
+charset_normalizer/models.py,sha256=lKXhOnIPtiakbK3i__J9wpOfzx3JDTKj7Dn3Rg0VaRI,12394
+charset_normalizer/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+charset_normalizer/utils.py,sha256=T5UHo8AS7NVMmgruWoZyqEf0WrZVcQpgUNetRoborSk,12002
+charset_normalizer/version.py,sha256=Ambcj3O8FfvdLfDLc8dkaxZx97O1IM_R4_aKGD_TDdE,115
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/WHEEL
new file mode 100644
index 00000000..028a7fd6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: false
+Tag: cp312-cp312-macosx_10_13_universal2
+
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/entry_points.txt
new file mode 100644
index 00000000..ec920125
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/entry_points.txt
@@ -0,0 +1,2 @@
+[console_scripts]
+normalizer = charset_normalizer:cli.cli_detect
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/top_level.txt
new file mode 100644
index 00000000..66958f0a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer-3.4.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+charset_normalizer
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__init__.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__init__.py
new file mode 100644
index 00000000..0d3a3799
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__init__.py
@@ -0,0 +1,48 @@
+"""
+Charset-Normalizer
+~~~~~~~~~~~~~~
+The Real First Universal Charset Detector.
+A library that helps you read text from an unknown charset encoding.
+Motivated by chardet, This package is trying to resolve the issue by taking a new approach.
+All IANA character set names for which the Python core library provides codecs are supported.
+
+Basic usage:
+   >>> from charset_normalizer import from_bytes
+   >>> results = from_bytes('Bсеки човек има право на образование. Oбразованието!'.encode('utf_8'))
+   >>> best_guess = results.best()
+   >>> str(best_guess)
+   'Bсеки човек има право на образование. Oбразованието!'
+
+Others methods and usages are available - see the full documentation
+at <https://github.com/Ousret/charset_normalizer>.
+:copyright: (c) 2021 by Ahmed TAHRI
+:license: MIT, see LICENSE for more details.
+"""
+
+from __future__ import annotations
+
+import logging
+
+from .api import from_bytes, from_fp, from_path, is_binary
+from .legacy import detect
+from .models import CharsetMatch, CharsetMatches
+from .utils import set_logging_handler
+from .version import VERSION, __version__
+
+__all__ = (
+    "from_fp",
+    "from_path",
+    "from_bytes",
+    "is_binary",
+    "detect",
+    "CharsetMatch",
+    "CharsetMatches",
+    "__version__",
+    "VERSION",
+    "set_logging_handler",
+)
+
+# Attach a NullHandler to the top level logger by default
+# https://docs.python.org/3.3/howto/logging.html#configuring-logging-for-a-library
+
+logging.getLogger("charset_normalizer").addHandler(logging.NullHandler())
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__main__.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__main__.py
new file mode 100644
index 00000000..e0e76f7b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__main__.py
@@ -0,0 +1,6 @@
+from __future__ import annotations
+
+from .cli import cli_detect
+
+if __name__ == "__main__":
+    cli_detect()
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..70efbf23
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..b8b6a296
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..be2f8051
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/cd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/cd.cpython-312.pyc
new file mode 100644
index 00000000..b3f704f0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/cd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/constant.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/constant.cpython-312.pyc
new file mode 100644
index 00000000..e5db936f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/constant.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/legacy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/legacy.cpython-312.pyc
new file mode 100644
index 00000000..37c2c998
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/legacy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/md.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/md.cpython-312.pyc
new file mode 100644
index 00000000..1f5ff041
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/md.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/models.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/models.cpython-312.pyc
new file mode 100644
index 00000000..556a6dc0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/models.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..1cd3a636
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..86b95b61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/api.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/api.py
new file mode 100644
index 00000000..2c8c0618
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/api.py
@@ -0,0 +1,668 @@
+from __future__ import annotations
+
+import logging
+from os import PathLike
+from typing import BinaryIO
+
+from .cd import (
+    coherence_ratio,
+    encoding_languages,
+    mb_encoding_languages,
+    merge_coherence_ratios,
+)
+from .constant import IANA_SUPPORTED, TOO_BIG_SEQUENCE, TOO_SMALL_SEQUENCE, TRACE
+from .md import mess_ratio
+from .models import CharsetMatch, CharsetMatches
+from .utils import (
+    any_specified_encoding,
+    cut_sequence_chunks,
+    iana_name,
+    identify_sig_or_bom,
+    is_cp_similar,
+    is_multi_byte_encoding,
+    should_strip_sig_or_bom,
+)
+
+logger = logging.getLogger("charset_normalizer")
+explain_handler = logging.StreamHandler()
+explain_handler.setFormatter(
+    logging.Formatter("%(asctime)s | %(levelname)s | %(message)s")
+)
+
+
+def from_bytes(
+    sequences: bytes | bytearray,
+    steps: int = 5,
+    chunk_size: int = 512,
+    threshold: float = 0.2,
+    cp_isolation: list[str] | None = None,
+    cp_exclusion: list[str] | None = None,
+    preemptive_behaviour: bool = True,
+    explain: bool = False,
+    language_threshold: float = 0.1,
+    enable_fallback: bool = True,
+) -> CharsetMatches:
+    """
+    Given a raw bytes sequence, return the best possibles charset usable to render str objects.
+    If there is no results, it is a strong indicator that the source is binary/not text.
+    By default, the process will extract 5 blocks of 512o each to assess the mess and coherence of a given sequence.
+    And will give up a particular code page after 20% of measured mess. Those criteria are customizable at will.
+
+    The preemptive behavior DOES NOT replace the traditional detection workflow, it prioritize a particular code page
+    but never take it for granted. Can improve the performance.
+
+    You may want to focus your attention to some code page or/and not others, use cp_isolation and cp_exclusion for that
+    purpose.
+
+    This function will strip the SIG in the payload/sequence every time except on UTF-16, UTF-32.
+    By default the library does not setup any handler other than the NullHandler, if you choose to set the 'explain'
+    toggle to True it will alter the logger configuration to add a StreamHandler that is suitable for debugging.
+    Custom logging format and handler can be set manually.
+    """
+
+    if not isinstance(sequences, (bytearray, bytes)):
+        raise TypeError(
+            "Expected object of type bytes or bytearray, got: {}".format(
+                type(sequences)
+            )
+        )
+
+    if explain:
+        previous_logger_level: int = logger.level
+        logger.addHandler(explain_handler)
+        logger.setLevel(TRACE)
+
+    length: int = len(sequences)
+
+    if length == 0:
+        logger.debug("Encoding detection on empty bytes, assuming utf_8 intention.")
+        if explain:  # Defensive: ensure exit path clean handler
+            logger.removeHandler(explain_handler)
+            logger.setLevel(previous_logger_level or logging.WARNING)
+        return CharsetMatches([CharsetMatch(sequences, "utf_8", 0.0, False, [], "")])
+
+    if cp_isolation is not None:
+        logger.log(
+            TRACE,
+            "cp_isolation is set. use this flag for debugging purpose. "
+            "limited list of encoding allowed : %s.",
+            ", ".join(cp_isolation),
+        )
+        cp_isolation = [iana_name(cp, False) for cp in cp_isolation]
+    else:
+        cp_isolation = []
+
+    if cp_exclusion is not None:
+        logger.log(
+            TRACE,
+            "cp_exclusion is set. use this flag for debugging purpose. "
+            "limited list of encoding excluded : %s.",
+            ", ".join(cp_exclusion),
+        )
+        cp_exclusion = [iana_name(cp, False) for cp in cp_exclusion]
+    else:
+        cp_exclusion = []
+
+    if length <= (chunk_size * steps):
+        logger.log(
+            TRACE,
+            "override steps (%i) and chunk_size (%i) as content does not fit (%i byte(s) given) parameters.",
+            steps,
+            chunk_size,
+            length,
+        )
+        steps = 1
+        chunk_size = length
+
+    if steps > 1 and length / steps < chunk_size:
+        chunk_size = int(length / steps)
+
+    is_too_small_sequence: bool = len(sequences) < TOO_SMALL_SEQUENCE
+    is_too_large_sequence: bool = len(sequences) >= TOO_BIG_SEQUENCE
+
+    if is_too_small_sequence:
+        logger.log(
+            TRACE,
+            "Trying to detect encoding from a tiny portion of ({}) byte(s).".format(
+                length
+            ),
+        )
+    elif is_too_large_sequence:
+        logger.log(
+            TRACE,
+            "Using lazy str decoding because the payload is quite large, ({}) byte(s).".format(
+                length
+            ),
+        )
+
+    prioritized_encodings: list[str] = []
+
+    specified_encoding: str | None = (
+        any_specified_encoding(sequences) if preemptive_behaviour else None
+    )
+
+    if specified_encoding is not None:
+        prioritized_encodings.append(specified_encoding)
+        logger.log(
+            TRACE,
+            "Detected declarative mark in sequence. Priority +1 given for %s.",
+            specified_encoding,
+        )
+
+    tested: set[str] = set()
+    tested_but_hard_failure: list[str] = []
+    tested_but_soft_failure: list[str] = []
+
+    fallback_ascii: CharsetMatch | None = None
+    fallback_u8: CharsetMatch | None = None
+    fallback_specified: CharsetMatch | None = None
+
+    results: CharsetMatches = CharsetMatches()
+
+    early_stop_results: CharsetMatches = CharsetMatches()
+
+    sig_encoding, sig_payload = identify_sig_or_bom(sequences)
+
+    if sig_encoding is not None:
+        prioritized_encodings.append(sig_encoding)
+        logger.log(
+            TRACE,
+            "Detected a SIG or BOM mark on first %i byte(s). Priority +1 given for %s.",
+            len(sig_payload),
+            sig_encoding,
+        )
+
+    prioritized_encodings.append("ascii")
+
+    if "utf_8" not in prioritized_encodings:
+        prioritized_encodings.append("utf_8")
+
+    for encoding_iana in prioritized_encodings + IANA_SUPPORTED:
+        if cp_isolation and encoding_iana not in cp_isolation:
+            continue
+
+        if cp_exclusion and encoding_iana in cp_exclusion:
+            continue
+
+        if encoding_iana in tested:
+            continue
+
+        tested.add(encoding_iana)
+
+        decoded_payload: str | None = None
+        bom_or_sig_available: bool = sig_encoding == encoding_iana
+        strip_sig_or_bom: bool = bom_or_sig_available and should_strip_sig_or_bom(
+            encoding_iana
+        )
+
+        if encoding_iana in {"utf_16", "utf_32"} and not bom_or_sig_available:
+            logger.log(
+                TRACE,
+                "Encoding %s won't be tested as-is because it require a BOM. Will try some sub-encoder LE/BE.",
+                encoding_iana,
+            )
+            continue
+        if encoding_iana in {"utf_7"} and not bom_or_sig_available:
+            logger.log(
+                TRACE,
+                "Encoding %s won't be tested as-is because detection is unreliable without BOM/SIG.",
+                encoding_iana,
+            )
+            continue
+
+        try:
+            is_multi_byte_decoder: bool = is_multi_byte_encoding(encoding_iana)
+        except (ModuleNotFoundError, ImportError):
+            logger.log(
+                TRACE,
+                "Encoding %s does not provide an IncrementalDecoder",
+                encoding_iana,
+            )
+            continue
+
+        try:
+            if is_too_large_sequence and is_multi_byte_decoder is False:
+                str(
+                    (
+                        sequences[: int(50e4)]
+                        if strip_sig_or_bom is False
+                        else sequences[len(sig_payload) : int(50e4)]
+                    ),
+                    encoding=encoding_iana,
+                )
+            else:
+                decoded_payload = str(
+                    (
+                        sequences
+                        if strip_sig_or_bom is False
+                        else sequences[len(sig_payload) :]
+                    ),
+                    encoding=encoding_iana,
+                )
+        except (UnicodeDecodeError, LookupError) as e:
+            if not isinstance(e, LookupError):
+                logger.log(
+                    TRACE,
+                    "Code page %s does not fit given bytes sequence at ALL. %s",
+                    encoding_iana,
+                    str(e),
+                )
+            tested_but_hard_failure.append(encoding_iana)
+            continue
+
+        similar_soft_failure_test: bool = False
+
+        for encoding_soft_failed in tested_but_soft_failure:
+            if is_cp_similar(encoding_iana, encoding_soft_failed):
+                similar_soft_failure_test = True
+                break
+
+        if similar_soft_failure_test:
+            logger.log(
+                TRACE,
+                "%s is deemed too similar to code page %s and was consider unsuited already. Continuing!",
+                encoding_iana,
+                encoding_soft_failed,
+            )
+            continue
+
+        r_ = range(
+            0 if not bom_or_sig_available else len(sig_payload),
+            length,
+            int(length / steps),
+        )
+
+        multi_byte_bonus: bool = (
+            is_multi_byte_decoder
+            and decoded_payload is not None
+            and len(decoded_payload) < length
+        )
+
+        if multi_byte_bonus:
+            logger.log(
+                TRACE,
+                "Code page %s is a multi byte encoding table and it appear that at least one character "
+                "was encoded using n-bytes.",
+                encoding_iana,
+            )
+
+        max_chunk_gave_up: int = int(len(r_) / 4)
+
+        max_chunk_gave_up = max(max_chunk_gave_up, 2)
+        early_stop_count: int = 0
+        lazy_str_hard_failure = False
+
+        md_chunks: list[str] = []
+        md_ratios = []
+
+        try:
+            for chunk in cut_sequence_chunks(
+                sequences,
+                encoding_iana,
+                r_,
+                chunk_size,
+                bom_or_sig_available,
+                strip_sig_or_bom,
+                sig_payload,
+                is_multi_byte_decoder,
+                decoded_payload,
+            ):
+                md_chunks.append(chunk)
+
+                md_ratios.append(
+                    mess_ratio(
+                        chunk,
+                        threshold,
+                        explain is True and 1 <= len(cp_isolation) <= 2,
+                    )
+                )
+
+                if md_ratios[-1] >= threshold:
+                    early_stop_count += 1
+
+                if (early_stop_count >= max_chunk_gave_up) or (
+                    bom_or_sig_available and strip_sig_or_bom is False
+                ):
+                    break
+        except (
+            UnicodeDecodeError
+        ) as e:  # Lazy str loading may have missed something there
+            logger.log(
+                TRACE,
+                "LazyStr Loading: After MD chunk decode, code page %s does not fit given bytes sequence at ALL. %s",
+                encoding_iana,
+                str(e),
+            )
+            early_stop_count = max_chunk_gave_up
+            lazy_str_hard_failure = True
+
+        # We might want to check the sequence again with the whole content
+        # Only if initial MD tests passes
+        if (
+            not lazy_str_hard_failure
+            and is_too_large_sequence
+            and not is_multi_byte_decoder
+        ):
+            try:
+                sequences[int(50e3) :].decode(encoding_iana, errors="strict")
+            except UnicodeDecodeError as e:
+                logger.log(
+                    TRACE,
+                    "LazyStr Loading: After final lookup, code page %s does not fit given bytes sequence at ALL. %s",
+                    encoding_iana,
+                    str(e),
+                )
+                tested_but_hard_failure.append(encoding_iana)
+                continue
+
+        mean_mess_ratio: float = sum(md_ratios) / len(md_ratios) if md_ratios else 0.0
+        if mean_mess_ratio >= threshold or early_stop_count >= max_chunk_gave_up:
+            tested_but_soft_failure.append(encoding_iana)
+            logger.log(
+                TRACE,
+                "%s was excluded because of initial chaos probing. Gave up %i time(s). "
+                "Computed mean chaos is %f %%.",
+                encoding_iana,
+                early_stop_count,
+                round(mean_mess_ratio * 100, ndigits=3),
+            )
+            # Preparing those fallbacks in case we got nothing.
+            if (
+                enable_fallback
+                and encoding_iana in ["ascii", "utf_8", specified_encoding]
+                and not lazy_str_hard_failure
+            ):
+                fallback_entry = CharsetMatch(
+                    sequences,
+                    encoding_iana,
+                    threshold,
+                    False,
+                    [],
+                    decoded_payload,
+                    preemptive_declaration=specified_encoding,
+                )
+                if encoding_iana == specified_encoding:
+                    fallback_specified = fallback_entry
+                elif encoding_iana == "ascii":
+                    fallback_ascii = fallback_entry
+                else:
+                    fallback_u8 = fallback_entry
+            continue
+
+        logger.log(
+            TRACE,
+            "%s passed initial chaos probing. Mean measured chaos is %f %%",
+            encoding_iana,
+            round(mean_mess_ratio * 100, ndigits=3),
+        )
+
+        if not is_multi_byte_decoder:
+            target_languages: list[str] = encoding_languages(encoding_iana)
+        else:
+            target_languages = mb_encoding_languages(encoding_iana)
+
+        if target_languages:
+            logger.log(
+                TRACE,
+                "{} should target any language(s) of {}".format(
+                    encoding_iana, str(target_languages)
+                ),
+            )
+
+        cd_ratios = []
+
+        # We shall skip the CD when its about ASCII
+        # Most of the time its not relevant to run "language-detection" on it.
+        if encoding_iana != "ascii":
+            for chunk in md_chunks:
+                chunk_languages = coherence_ratio(
+                    chunk,
+                    language_threshold,
+                    ",".join(target_languages) if target_languages else None,
+                )
+
+                cd_ratios.append(chunk_languages)
+
+        cd_ratios_merged = merge_coherence_ratios(cd_ratios)
+
+        if cd_ratios_merged:
+            logger.log(
+                TRACE,
+                "We detected language {} using {}".format(
+                    cd_ratios_merged, encoding_iana
+                ),
+            )
+
+        current_match = CharsetMatch(
+            sequences,
+            encoding_iana,
+            mean_mess_ratio,
+            bom_or_sig_available,
+            cd_ratios_merged,
+            (
+                decoded_payload
+                if (
+                    is_too_large_sequence is False
+                    or encoding_iana in [specified_encoding, "ascii", "utf_8"]
+                )
+                else None
+            ),
+            preemptive_declaration=specified_encoding,
+        )
+
+        results.append(current_match)
+
+        if (
+            encoding_iana in [specified_encoding, "ascii", "utf_8"]
+            and mean_mess_ratio < 0.1
+        ):
+            # If md says nothing to worry about, then... stop immediately!
+            if mean_mess_ratio == 0.0:
+                logger.debug(
+                    "Encoding detection: %s is most likely the one.",
+                    current_match.encoding,
+                )
+                if explain:  # Defensive: ensure exit path clean handler
+                    logger.removeHandler(explain_handler)
+                    logger.setLevel(previous_logger_level)
+                return CharsetMatches([current_match])
+
+            early_stop_results.append(current_match)
+
+        if (
+            len(early_stop_results)
+            and (specified_encoding is None or specified_encoding in tested)
+            and "ascii" in tested
+            and "utf_8" in tested
+        ):
+            probable_result: CharsetMatch = early_stop_results.best()  # type: ignore[assignment]
+            logger.debug(
+                "Encoding detection: %s is most likely the one.",
+                probable_result.encoding,
+            )
+            if explain:  # Defensive: ensure exit path clean handler
+                logger.removeHandler(explain_handler)
+                logger.setLevel(previous_logger_level)
+
+            return CharsetMatches([probable_result])
+
+        if encoding_iana == sig_encoding:
+            logger.debug(
+                "Encoding detection: %s is most likely the one as we detected a BOM or SIG within "
+                "the beginning of the sequence.",
+                encoding_iana,
+            )
+            if explain:  # Defensive: ensure exit path clean handler
+                logger.removeHandler(explain_handler)
+                logger.setLevel(previous_logger_level)
+            return CharsetMatches([results[encoding_iana]])
+
+    if len(results) == 0:
+        if fallback_u8 or fallback_ascii or fallback_specified:
+            logger.log(
+                TRACE,
+                "Nothing got out of the detection process. Using ASCII/UTF-8/Specified fallback.",
+            )
+
+        if fallback_specified:
+            logger.debug(
+                "Encoding detection: %s will be used as a fallback match",
+                fallback_specified.encoding,
+            )
+            results.append(fallback_specified)
+        elif (
+            (fallback_u8 and fallback_ascii is None)
+            or (
+                fallback_u8
+                and fallback_ascii
+                and fallback_u8.fingerprint != fallback_ascii.fingerprint
+            )
+            or (fallback_u8 is not None)
+        ):
+            logger.debug("Encoding detection: utf_8 will be used as a fallback match")
+            results.append(fallback_u8)
+        elif fallback_ascii:
+            logger.debug("Encoding detection: ascii will be used as a fallback match")
+            results.append(fallback_ascii)
+
+    if results:
+        logger.debug(
+            "Encoding detection: Found %s as plausible (best-candidate) for content. With %i alternatives.",
+            results.best().encoding,  # type: ignore
+            len(results) - 1,
+        )
+    else:
+        logger.debug("Encoding detection: Unable to determine any suitable charset.")
+
+    if explain:
+        logger.removeHandler(explain_handler)
+        logger.setLevel(previous_logger_level)
+
+    return results
+
+
+def from_fp(
+    fp: BinaryIO,
+    steps: int = 5,
+    chunk_size: int = 512,
+    threshold: float = 0.20,
+    cp_isolation: list[str] | None = None,
+    cp_exclusion: list[str] | None = None,
+    preemptive_behaviour: bool = True,
+    explain: bool = False,
+    language_threshold: float = 0.1,
+    enable_fallback: bool = True,
+) -> CharsetMatches:
+    """
+    Same thing than the function from_bytes but using a file pointer that is already ready.
+    Will not close the file pointer.
+    """
+    return from_bytes(
+        fp.read(),
+        steps,
+        chunk_size,
+        threshold,
+        cp_isolation,
+        cp_exclusion,
+        preemptive_behaviour,
+        explain,
+        language_threshold,
+        enable_fallback,
+    )
+
+
+def from_path(
+    path: str | bytes | PathLike,  # type: ignore[type-arg]
+    steps: int = 5,
+    chunk_size: int = 512,
+    threshold: float = 0.20,
+    cp_isolation: list[str] | None = None,
+    cp_exclusion: list[str] | None = None,
+    preemptive_behaviour: bool = True,
+    explain: bool = False,
+    language_threshold: float = 0.1,
+    enable_fallback: bool = True,
+) -> CharsetMatches:
+    """
+    Same thing than the function from_bytes but with one extra step. Opening and reading given file path in binary mode.
+    Can raise IOError.
+    """
+    with open(path, "rb") as fp:
+        return from_fp(
+            fp,
+            steps,
+            chunk_size,
+            threshold,
+            cp_isolation,
+            cp_exclusion,
+            preemptive_behaviour,
+            explain,
+            language_threshold,
+            enable_fallback,
+        )
+
+
+def is_binary(
+    fp_or_path_or_payload: PathLike | str | BinaryIO | bytes,  # type: ignore[type-arg]
+    steps: int = 5,
+    chunk_size: int = 512,
+    threshold: float = 0.20,
+    cp_isolation: list[str] | None = None,
+    cp_exclusion: list[str] | None = None,
+    preemptive_behaviour: bool = True,
+    explain: bool = False,
+    language_threshold: float = 0.1,
+    enable_fallback: bool = False,
+) -> bool:
+    """
+    Detect if the given input (file, bytes, or path) points to a binary file. aka. not a string.
+    Based on the same main heuristic algorithms and default kwargs at the sole exception that fallbacks match
+    are disabled to be stricter around ASCII-compatible but unlikely to be a string.
+    """
+    if isinstance(fp_or_path_or_payload, (str, PathLike)):
+        guesses = from_path(
+            fp_or_path_or_payload,
+            steps=steps,
+            chunk_size=chunk_size,
+            threshold=threshold,
+            cp_isolation=cp_isolation,
+            cp_exclusion=cp_exclusion,
+            preemptive_behaviour=preemptive_behaviour,
+            explain=explain,
+            language_threshold=language_threshold,
+            enable_fallback=enable_fallback,
+        )
+    elif isinstance(
+        fp_or_path_or_payload,
+        (
+            bytes,
+            bytearray,
+        ),
+    ):
+        guesses = from_bytes(
+            fp_or_path_or_payload,
+            steps=steps,
+            chunk_size=chunk_size,
+            threshold=threshold,
+            cp_isolation=cp_isolation,
+            cp_exclusion=cp_exclusion,
+            preemptive_behaviour=preemptive_behaviour,
+            explain=explain,
+            language_threshold=language_threshold,
+            enable_fallback=enable_fallback,
+        )
+    else:
+        guesses = from_fp(
+            fp_or_path_or_payload,
+            steps=steps,
+            chunk_size=chunk_size,
+            threshold=threshold,
+            cp_isolation=cp_isolation,
+            cp_exclusion=cp_exclusion,
+            preemptive_behaviour=preemptive_behaviour,
+            explain=explain,
+            language_threshold=language_threshold,
+            enable_fallback=enable_fallback,
+        )
+
+    return not guesses
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/cd.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cd.py
new file mode 100644
index 00000000..71a3ed51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cd.py
@@ -0,0 +1,395 @@
+from __future__ import annotations
+
+import importlib
+from codecs import IncrementalDecoder
+from collections import Counter
+from functools import lru_cache
+from typing import Counter as TypeCounter
+
+from .constant import (
+    FREQUENCIES,
+    KO_NAMES,
+    LANGUAGE_SUPPORTED_COUNT,
+    TOO_SMALL_SEQUENCE,
+    ZH_NAMES,
+)
+from .md import is_suspiciously_successive_range
+from .models import CoherenceMatches
+from .utils import (
+    is_accentuated,
+    is_latin,
+    is_multi_byte_encoding,
+    is_unicode_range_secondary,
+    unicode_range,
+)
+
+
+def encoding_unicode_range(iana_name: str) -> list[str]:
+    """
+    Return associated unicode ranges in a single byte code page.
+    """
+    if is_multi_byte_encoding(iana_name):
+        raise OSError("Function not supported on multi-byte code page")
+
+    decoder = importlib.import_module(f"encodings.{iana_name}").IncrementalDecoder
+
+    p: IncrementalDecoder = decoder(errors="ignore")
+    seen_ranges: dict[str, int] = {}
+    character_count: int = 0
+
+    for i in range(0x40, 0xFF):
+        chunk: str = p.decode(bytes([i]))
+
+        if chunk:
+            character_range: str | None = unicode_range(chunk)
+
+            if character_range is None:
+                continue
+
+            if is_unicode_range_secondary(character_range) is False:
+                if character_range not in seen_ranges:
+                    seen_ranges[character_range] = 0
+                seen_ranges[character_range] += 1
+                character_count += 1
+
+    return sorted(
+        [
+            character_range
+            for character_range in seen_ranges
+            if seen_ranges[character_range] / character_count >= 0.15
+        ]
+    )
+
+
+def unicode_range_languages(primary_range: str) -> list[str]:
+    """
+    Return inferred languages used with a unicode range.
+    """
+    languages: list[str] = []
+
+    for language, characters in FREQUENCIES.items():
+        for character in characters:
+            if unicode_range(character) == primary_range:
+                languages.append(language)
+                break
+
+    return languages
+
+
+@lru_cache()
+def encoding_languages(iana_name: str) -> list[str]:
+    """
+    Single-byte encoding language association. Some code page are heavily linked to particular language(s).
+    This function does the correspondence.
+    """
+    unicode_ranges: list[str] = encoding_unicode_range(iana_name)
+    primary_range: str | None = None
+
+    for specified_range in unicode_ranges:
+        if "Latin" not in specified_range:
+            primary_range = specified_range
+            break
+
+    if primary_range is None:
+        return ["Latin Based"]
+
+    return unicode_range_languages(primary_range)
+
+
+@lru_cache()
+def mb_encoding_languages(iana_name: str) -> list[str]:
+    """
+    Multi-byte encoding language association. Some code page are heavily linked to particular language(s).
+    This function does the correspondence.
+    """
+    if (
+        iana_name.startswith("shift_")
+        or iana_name.startswith("iso2022_jp")
+        or iana_name.startswith("euc_j")
+        or iana_name == "cp932"
+    ):
+        return ["Japanese"]
+    if iana_name.startswith("gb") or iana_name in ZH_NAMES:
+        return ["Chinese"]
+    if iana_name.startswith("iso2022_kr") or iana_name in KO_NAMES:
+        return ["Korean"]
+
+    return []
+
+
+@lru_cache(maxsize=LANGUAGE_SUPPORTED_COUNT)
+def get_target_features(language: str) -> tuple[bool, bool]:
+    """
+    Determine main aspects from a supported language if it contains accents and if is pure Latin.
+    """
+    target_have_accents: bool = False
+    target_pure_latin: bool = True
+
+    for character in FREQUENCIES[language]:
+        if not target_have_accents and is_accentuated(character):
+            target_have_accents = True
+        if target_pure_latin and is_latin(character) is False:
+            target_pure_latin = False
+
+    return target_have_accents, target_pure_latin
+
+
+def alphabet_languages(
+    characters: list[str], ignore_non_latin: bool = False
+) -> list[str]:
+    """
+    Return associated languages associated to given characters.
+    """
+    languages: list[tuple[str, float]] = []
+
+    source_have_accents = any(is_accentuated(character) for character in characters)
+
+    for language, language_characters in FREQUENCIES.items():
+        target_have_accents, target_pure_latin = get_target_features(language)
+
+        if ignore_non_latin and target_pure_latin is False:
+            continue
+
+        if target_have_accents is False and source_have_accents:
+            continue
+
+        character_count: int = len(language_characters)
+
+        character_match_count: int = len(
+            [c for c in language_characters if c in characters]
+        )
+
+        ratio: float = character_match_count / character_count
+
+        if ratio >= 0.2:
+            languages.append((language, ratio))
+
+    languages = sorted(languages, key=lambda x: x[1], reverse=True)
+
+    return [compatible_language[0] for compatible_language in languages]
+
+
+def characters_popularity_compare(
+    language: str, ordered_characters: list[str]
+) -> float:
+    """
+    Determine if a ordered characters list (by occurrence from most appearance to rarest) match a particular language.
+    The result is a ratio between 0. (absolutely no correspondence) and 1. (near perfect fit).
+    Beware that is function is not strict on the match in order to ease the detection. (Meaning close match is 1.)
+    """
+    if language not in FREQUENCIES:
+        raise ValueError(f"{language} not available")
+
+    character_approved_count: int = 0
+    FREQUENCIES_language_set = set(FREQUENCIES[language])
+
+    ordered_characters_count: int = len(ordered_characters)
+    target_language_characters_count: int = len(FREQUENCIES[language])
+
+    large_alphabet: bool = target_language_characters_count > 26
+
+    for character, character_rank in zip(
+        ordered_characters, range(0, ordered_characters_count)
+    ):
+        if character not in FREQUENCIES_language_set:
+            continue
+
+        character_rank_in_language: int = FREQUENCIES[language].index(character)
+        expected_projection_ratio: float = (
+            target_language_characters_count / ordered_characters_count
+        )
+        character_rank_projection: int = int(character_rank * expected_projection_ratio)
+
+        if (
+            large_alphabet is False
+            and abs(character_rank_projection - character_rank_in_language) > 4
+        ):
+            continue
+
+        if (
+            large_alphabet is True
+            and abs(character_rank_projection - character_rank_in_language)
+            < target_language_characters_count / 3
+        ):
+            character_approved_count += 1
+            continue
+
+        characters_before_source: list[str] = FREQUENCIES[language][
+            0:character_rank_in_language
+        ]
+        characters_after_source: list[str] = FREQUENCIES[language][
+            character_rank_in_language:
+        ]
+        characters_before: list[str] = ordered_characters[0:character_rank]
+        characters_after: list[str] = ordered_characters[character_rank:]
+
+        before_match_count: int = len(
+            set(characters_before) & set(characters_before_source)
+        )
+
+        after_match_count: int = len(
+            set(characters_after) & set(characters_after_source)
+        )
+
+        if len(characters_before_source) == 0 and before_match_count <= 4:
+            character_approved_count += 1
+            continue
+
+        if len(characters_after_source) == 0 and after_match_count <= 4:
+            character_approved_count += 1
+            continue
+
+        if (
+            before_match_count / len(characters_before_source) >= 0.4
+            or after_match_count / len(characters_after_source) >= 0.4
+        ):
+            character_approved_count += 1
+            continue
+
+    return character_approved_count / len(ordered_characters)
+
+
+def alpha_unicode_split(decoded_sequence: str) -> list[str]:
+    """
+    Given a decoded text sequence, return a list of str. Unicode range / alphabet separation.
+    Ex. a text containing English/Latin with a bit a Hebrew will return two items in the resulting list;
+    One containing the latin letters and the other hebrew.
+    """
+    layers: dict[str, str] = {}
+
+    for character in decoded_sequence:
+        if character.isalpha() is False:
+            continue
+
+        character_range: str | None = unicode_range(character)
+
+        if character_range is None:
+            continue
+
+        layer_target_range: str | None = None
+
+        for discovered_range in layers:
+            if (
+                is_suspiciously_successive_range(discovered_range, character_range)
+                is False
+            ):
+                layer_target_range = discovered_range
+                break
+
+        if layer_target_range is None:
+            layer_target_range = character_range
+
+        if layer_target_range not in layers:
+            layers[layer_target_range] = character.lower()
+            continue
+
+        layers[layer_target_range] += character.lower()
+
+    return list(layers.values())
+
+
+def merge_coherence_ratios(results: list[CoherenceMatches]) -> CoherenceMatches:
+    """
+    This function merge results previously given by the function coherence_ratio.
+    The return type is the same as coherence_ratio.
+    """
+    per_language_ratios: dict[str, list[float]] = {}
+    for result in results:
+        for sub_result in result:
+            language, ratio = sub_result
+            if language not in per_language_ratios:
+                per_language_ratios[language] = [ratio]
+                continue
+            per_language_ratios[language].append(ratio)
+
+    merge = [
+        (
+            language,
+            round(
+                sum(per_language_ratios[language]) / len(per_language_ratios[language]),
+                4,
+            ),
+        )
+        for language in per_language_ratios
+    ]
+
+    return sorted(merge, key=lambda x: x[1], reverse=True)
+
+
+def filter_alt_coherence_matches(results: CoherenceMatches) -> CoherenceMatches:
+    """
+    We shall NOT return "English—" in CoherenceMatches because it is an alternative
+    of "English". This function only keeps the best match and remove the em-dash in it.
+    """
+    index_results: dict[str, list[float]] = dict()
+
+    for result in results:
+        language, ratio = result
+        no_em_name: str = language.replace("—", "")
+
+        if no_em_name not in index_results:
+            index_results[no_em_name] = []
+
+        index_results[no_em_name].append(ratio)
+
+    if any(len(index_results[e]) > 1 for e in index_results):
+        filtered_results: CoherenceMatches = []
+
+        for language in index_results:
+            filtered_results.append((language, max(index_results[language])))
+
+        return filtered_results
+
+    return results
+
+
+@lru_cache(maxsize=2048)
+def coherence_ratio(
+    decoded_sequence: str, threshold: float = 0.1, lg_inclusion: str | None = None
+) -> CoherenceMatches:
+    """
+    Detect ANY language that can be identified in given sequence. The sequence will be analysed by layers.
+    A layer = Character extraction by alphabets/ranges.
+    """
+
+    results: list[tuple[str, float]] = []
+    ignore_non_latin: bool = False
+
+    sufficient_match_count: int = 0
+
+    lg_inclusion_list = lg_inclusion.split(",") if lg_inclusion is not None else []
+    if "Latin Based" in lg_inclusion_list:
+        ignore_non_latin = True
+        lg_inclusion_list.remove("Latin Based")
+
+    for layer in alpha_unicode_split(decoded_sequence):
+        sequence_frequencies: TypeCounter[str] = Counter(layer)
+        most_common = sequence_frequencies.most_common()
+
+        character_count: int = sum(o for c, o in most_common)
+
+        if character_count <= TOO_SMALL_SEQUENCE:
+            continue
+
+        popular_character_ordered: list[str] = [c for c, o in most_common]
+
+        for language in lg_inclusion_list or alphabet_languages(
+            popular_character_ordered, ignore_non_latin
+        ):
+            ratio: float = characters_popularity_compare(
+                language, popular_character_ordered
+            )
+
+            if ratio < threshold:
+                continue
+            elif ratio >= 0.8:
+                sufficient_match_count += 1
+
+            results.append((language, round(ratio, 4)))
+
+            if sufficient_match_count >= 3:
+                break
+
+    return sorted(
+        filter_alt_coherence_matches(results), key=lambda x: x[1], reverse=True
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__init__.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__init__.py
new file mode 100644
index 00000000..543a5a4d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__init__.py
@@ -0,0 +1,8 @@
+from __future__ import annotations
+
+from .__main__ import cli_detect, query_yes_no
+
+__all__ = (
+    "cli_detect",
+    "query_yes_no",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__main__.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__main__.py
new file mode 100644
index 00000000..64a290f2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__main__.py
@@ -0,0 +1,321 @@
+from __future__ import annotations
+
+import argparse
+import sys
+from json import dumps
+from os.path import abspath, basename, dirname, join, realpath
+from platform import python_version
+from unicodedata import unidata_version
+
+import charset_normalizer.md as md_module
+from charset_normalizer import from_fp
+from charset_normalizer.models import CliDetectionResult
+from charset_normalizer.version import __version__
+
+
+def query_yes_no(question: str, default: str = "yes") -> bool:
+    """Ask a yes/no question via input() and return their answer.
+
+    "question" is a string that is presented to the user.
+    "default" is the presumed answer if the user just hits <Enter>.
+        It must be "yes" (the default), "no" or None (meaning
+        an answer is required of the user).
+
+    The "answer" return value is True for "yes" or False for "no".
+
+    Credit goes to (c) https://stackoverflow.com/questions/3041986/apt-command-line-interface-like-yes-no-input
+    """
+    valid = {"yes": True, "y": True, "ye": True, "no": False, "n": False}
+    if default is None:
+        prompt = " [y/n] "
+    elif default == "yes":
+        prompt = " [Y/n] "
+    elif default == "no":
+        prompt = " [y/N] "
+    else:
+        raise ValueError("invalid default answer: '%s'" % default)
+
+    while True:
+        sys.stdout.write(question + prompt)
+        choice = input().lower()
+        if default is not None and choice == "":
+            return valid[default]
+        elif choice in valid:
+            return valid[choice]
+        else:
+            sys.stdout.write("Please respond with 'yes' or 'no' " "(or 'y' or 'n').\n")
+
+
+def cli_detect(argv: list[str] | None = None) -> int:
+    """
+    CLI assistant using ARGV and ArgumentParser
+    :param argv:
+    :return: 0 if everything is fine, anything else equal trouble
+    """
+    parser = argparse.ArgumentParser(
+        description="The Real First Universal Charset Detector. "
+        "Discover originating encoding used on text file. "
+        "Normalize text to unicode."
+    )
+
+    parser.add_argument(
+        "files", type=argparse.FileType("rb"), nargs="+", help="File(s) to be analysed"
+    )
+    parser.add_argument(
+        "-v",
+        "--verbose",
+        action="store_true",
+        default=False,
+        dest="verbose",
+        help="Display complementary information about file if any. "
+        "Stdout will contain logs about the detection process.",
+    )
+    parser.add_argument(
+        "-a",
+        "--with-alternative",
+        action="store_true",
+        default=False,
+        dest="alternatives",
+        help="Output complementary possibilities if any. Top-level JSON WILL be a list.",
+    )
+    parser.add_argument(
+        "-n",
+        "--normalize",
+        action="store_true",
+        default=False,
+        dest="normalize",
+        help="Permit to normalize input file. If not set, program does not write anything.",
+    )
+    parser.add_argument(
+        "-m",
+        "--minimal",
+        action="store_true",
+        default=False,
+        dest="minimal",
+        help="Only output the charset detected to STDOUT. Disabling JSON output.",
+    )
+    parser.add_argument(
+        "-r",
+        "--replace",
+        action="store_true",
+        default=False,
+        dest="replace",
+        help="Replace file when trying to normalize it instead of creating a new one.",
+    )
+    parser.add_argument(
+        "-f",
+        "--force",
+        action="store_true",
+        default=False,
+        dest="force",
+        help="Replace file without asking if you are sure, use this flag with caution.",
+    )
+    parser.add_argument(
+        "-i",
+        "--no-preemptive",
+        action="store_true",
+        default=False,
+        dest="no_preemptive",
+        help="Disable looking at a charset declaration to hint the detector.",
+    )
+    parser.add_argument(
+        "-t",
+        "--threshold",
+        action="store",
+        default=0.2,
+        type=float,
+        dest="threshold",
+        help="Define a custom maximum amount of noise allowed in decoded content. 0. <= noise <= 1.",
+    )
+    parser.add_argument(
+        "--version",
+        action="version",
+        version="Charset-Normalizer {} - Python {} - Unicode {} - SpeedUp {}".format(
+            __version__,
+            python_version(),
+            unidata_version,
+            "OFF" if md_module.__file__.lower().endswith(".py") else "ON",
+        ),
+        help="Show version information and exit.",
+    )
+
+    args = parser.parse_args(argv)
+
+    if args.replace is True and args.normalize is False:
+        if args.files:
+            for my_file in args.files:
+                my_file.close()
+        print("Use --replace in addition of --normalize only.", file=sys.stderr)
+        return 1
+
+    if args.force is True and args.replace is False:
+        if args.files:
+            for my_file in args.files:
+                my_file.close()
+        print("Use --force in addition of --replace only.", file=sys.stderr)
+        return 1
+
+    if args.threshold < 0.0 or args.threshold > 1.0:
+        if args.files:
+            for my_file in args.files:
+                my_file.close()
+        print("--threshold VALUE should be between 0. AND 1.", file=sys.stderr)
+        return 1
+
+    x_ = []
+
+    for my_file in args.files:
+        matches = from_fp(
+            my_file,
+            threshold=args.threshold,
+            explain=args.verbose,
+            preemptive_behaviour=args.no_preemptive is False,
+        )
+
+        best_guess = matches.best()
+
+        if best_guess is None:
+            print(
+                'Unable to identify originating encoding for "{}". {}'.format(
+                    my_file.name,
+                    (
+                        "Maybe try increasing maximum amount of chaos."
+                        if args.threshold < 1.0
+                        else ""
+                    ),
+                ),
+                file=sys.stderr,
+            )
+            x_.append(
+                CliDetectionResult(
+                    abspath(my_file.name),
+                    None,
+                    [],
+                    [],
+                    "Unknown",
+                    [],
+                    False,
+                    1.0,
+                    0.0,
+                    None,
+                    True,
+                )
+            )
+        else:
+            x_.append(
+                CliDetectionResult(
+                    abspath(my_file.name),
+                    best_guess.encoding,
+                    best_guess.encoding_aliases,
+                    [
+                        cp
+                        for cp in best_guess.could_be_from_charset
+                        if cp != best_guess.encoding
+                    ],
+                    best_guess.language,
+                    best_guess.alphabets,
+                    best_guess.bom,
+                    best_guess.percent_chaos,
+                    best_guess.percent_coherence,
+                    None,
+                    True,
+                )
+            )
+
+            if len(matches) > 1 and args.alternatives:
+                for el in matches:
+                    if el != best_guess:
+                        x_.append(
+                            CliDetectionResult(
+                                abspath(my_file.name),
+                                el.encoding,
+                                el.encoding_aliases,
+                                [
+                                    cp
+                                    for cp in el.could_be_from_charset
+                                    if cp != el.encoding
+                                ],
+                                el.language,
+                                el.alphabets,
+                                el.bom,
+                                el.percent_chaos,
+                                el.percent_coherence,
+                                None,
+                                False,
+                            )
+                        )
+
+            if args.normalize is True:
+                if best_guess.encoding.startswith("utf") is True:
+                    print(
+                        '"{}" file does not need to be normalized, as it already came from unicode.'.format(
+                            my_file.name
+                        ),
+                        file=sys.stderr,
+                    )
+                    if my_file.closed is False:
+                        my_file.close()
+                    continue
+
+                dir_path = dirname(realpath(my_file.name))
+                file_name = basename(realpath(my_file.name))
+
+                o_: list[str] = file_name.split(".")
+
+                if args.replace is False:
+                    o_.insert(-1, best_guess.encoding)
+                    if my_file.closed is False:
+                        my_file.close()
+                elif (
+                    args.force is False
+                    and query_yes_no(
+                        'Are you sure to normalize "{}" by replacing it ?'.format(
+                            my_file.name
+                        ),
+                        "no",
+                    )
+                    is False
+                ):
+                    if my_file.closed is False:
+                        my_file.close()
+                    continue
+
+                try:
+                    x_[0].unicode_path = join(dir_path, ".".join(o_))
+
+                    with open(x_[0].unicode_path, "wb") as fp:
+                        fp.write(best_guess.output())
+                except OSError as e:
+                    print(str(e), file=sys.stderr)
+                    if my_file.closed is False:
+                        my_file.close()
+                    return 2
+
+        if my_file.closed is False:
+            my_file.close()
+
+    if args.minimal is False:
+        print(
+            dumps(
+                [el.__dict__ for el in x_] if len(x_) > 1 else x_[0].__dict__,
+                ensure_ascii=True,
+                indent=4,
+            )
+        )
+    else:
+        for my_file in args.files:
+            print(
+                ", ".join(
+                    [
+                        el.encoding or "undefined"
+                        for el in x_
+                        if el.path == abspath(my_file.name)
+                    ]
+                )
+            )
+
+    return 0
+
+
+if __name__ == "__main__":
+    cli_detect()
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..091c27ca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..48302226
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/cli/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/constant.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/constant.py
new file mode 100644
index 00000000..1fb9508d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/constant.py
@@ -0,0 +1,1998 @@
+from __future__ import annotations
+
+from codecs import BOM_UTF8, BOM_UTF16_BE, BOM_UTF16_LE, BOM_UTF32_BE, BOM_UTF32_LE
+from encodings.aliases import aliases
+from re import IGNORECASE
+from re import compile as re_compile
+
+# Contain for each eligible encoding a list of/item bytes SIG/BOM
+ENCODING_MARKS: dict[str, bytes | list[bytes]] = {
+    "utf_8": BOM_UTF8,
+    "utf_7": [
+        b"\x2b\x2f\x76\x38",
+        b"\x2b\x2f\x76\x39",
+        b"\x2b\x2f\x76\x2b",
+        b"\x2b\x2f\x76\x2f",
+        b"\x2b\x2f\x76\x38\x2d",
+    ],
+    "gb18030": b"\x84\x31\x95\x33",
+    "utf_32": [BOM_UTF32_BE, BOM_UTF32_LE],
+    "utf_16": [BOM_UTF16_BE, BOM_UTF16_LE],
+}
+
+TOO_SMALL_SEQUENCE: int = 32
+TOO_BIG_SEQUENCE: int = int(10e6)
+
+UTF8_MAXIMAL_ALLOCATION: int = 1_112_064
+
+# Up-to-date Unicode ucd/15.0.0
+UNICODE_RANGES_COMBINED: dict[str, range] = {
+    "Control character": range(32),
+    "Basic Latin": range(32, 128),
+    "Latin-1 Supplement": range(128, 256),
+    "Latin Extended-A": range(256, 384),
+    "Latin Extended-B": range(384, 592),
+    "IPA Extensions": range(592, 688),
+    "Spacing Modifier Letters": range(688, 768),
+    "Combining Diacritical Marks": range(768, 880),
+    "Greek and Coptic": range(880, 1024),
+    "Cyrillic": range(1024, 1280),
+    "Cyrillic Supplement": range(1280, 1328),
+    "Armenian": range(1328, 1424),
+    "Hebrew": range(1424, 1536),
+    "Arabic": range(1536, 1792),
+    "Syriac": range(1792, 1872),
+    "Arabic Supplement": range(1872, 1920),
+    "Thaana": range(1920, 1984),
+    "NKo": range(1984, 2048),
+    "Samaritan": range(2048, 2112),
+    "Mandaic": range(2112, 2144),
+    "Syriac Supplement": range(2144, 2160),
+    "Arabic Extended-B": range(2160, 2208),
+    "Arabic Extended-A": range(2208, 2304),
+    "Devanagari": range(2304, 2432),
+    "Bengali": range(2432, 2560),
+    "Gurmukhi": range(2560, 2688),
+    "Gujarati": range(2688, 2816),
+    "Oriya": range(2816, 2944),
+    "Tamil": range(2944, 3072),
+    "Telugu": range(3072, 3200),
+    "Kannada": range(3200, 3328),
+    "Malayalam": range(3328, 3456),
+    "Sinhala": range(3456, 3584),
+    "Thai": range(3584, 3712),
+    "Lao": range(3712, 3840),
+    "Tibetan": range(3840, 4096),
+    "Myanmar": range(4096, 4256),
+    "Georgian": range(4256, 4352),
+    "Hangul Jamo": range(4352, 4608),
+    "Ethiopic": range(4608, 4992),
+    "Ethiopic Supplement": range(4992, 5024),
+    "Cherokee": range(5024, 5120),
+    "Unified Canadian Aboriginal Syllabics": range(5120, 5760),
+    "Ogham": range(5760, 5792),
+    "Runic": range(5792, 5888),
+    "Tagalog": range(5888, 5920),
+    "Hanunoo": range(5920, 5952),
+    "Buhid": range(5952, 5984),
+    "Tagbanwa": range(5984, 6016),
+    "Khmer": range(6016, 6144),
+    "Mongolian": range(6144, 6320),
+    "Unified Canadian Aboriginal Syllabics Extended": range(6320, 6400),
+    "Limbu": range(6400, 6480),
+    "Tai Le": range(6480, 6528),
+    "New Tai Lue": range(6528, 6624),
+    "Khmer Symbols": range(6624, 6656),
+    "Buginese": range(6656, 6688),
+    "Tai Tham": range(6688, 6832),
+    "Combining Diacritical Marks Extended": range(6832, 6912),
+    "Balinese": range(6912, 7040),
+    "Sundanese": range(7040, 7104),
+    "Batak": range(7104, 7168),
+    "Lepcha": range(7168, 7248),
+    "Ol Chiki": range(7248, 7296),
+    "Cyrillic Extended-C": range(7296, 7312),
+    "Georgian Extended": range(7312, 7360),
+    "Sundanese Supplement": range(7360, 7376),
+    "Vedic Extensions": range(7376, 7424),
+    "Phonetic Extensions": range(7424, 7552),
+    "Phonetic Extensions Supplement": range(7552, 7616),
+    "Combining Diacritical Marks Supplement": range(7616, 7680),
+    "Latin Extended Additional": range(7680, 7936),
+    "Greek Extended": range(7936, 8192),
+    "General Punctuation": range(8192, 8304),
+    "Superscripts and Subscripts": range(8304, 8352),
+    "Currency Symbols": range(8352, 8400),
+    "Combining Diacritical Marks for Symbols": range(8400, 8448),
+    "Letterlike Symbols": range(8448, 8528),
+    "Number Forms": range(8528, 8592),
+    "Arrows": range(8592, 8704),
+    "Mathematical Operators": range(8704, 8960),
+    "Miscellaneous Technical": range(8960, 9216),
+    "Control Pictures": range(9216, 9280),
+    "Optical Character Recognition": range(9280, 9312),
+    "Enclosed Alphanumerics": range(9312, 9472),
+    "Box Drawing": range(9472, 9600),
+    "Block Elements": range(9600, 9632),
+    "Geometric Shapes": range(9632, 9728),
+    "Miscellaneous Symbols": range(9728, 9984),
+    "Dingbats": range(9984, 10176),
+    "Miscellaneous Mathematical Symbols-A": range(10176, 10224),
+    "Supplemental Arrows-A": range(10224, 10240),
+    "Braille Patterns": range(10240, 10496),
+    "Supplemental Arrows-B": range(10496, 10624),
+    "Miscellaneous Mathematical Symbols-B": range(10624, 10752),
+    "Supplemental Mathematical Operators": range(10752, 11008),
+    "Miscellaneous Symbols and Arrows": range(11008, 11264),
+    "Glagolitic": range(11264, 11360),
+    "Latin Extended-C": range(11360, 11392),
+    "Coptic": range(11392, 11520),
+    "Georgian Supplement": range(11520, 11568),
+    "Tifinagh": range(11568, 11648),
+    "Ethiopic Extended": range(11648, 11744),
+    "Cyrillic Extended-A": range(11744, 11776),
+    "Supplemental Punctuation": range(11776, 11904),
+    "CJK Radicals Supplement": range(11904, 12032),
+    "Kangxi Radicals": range(12032, 12256),
+    "Ideographic Description Characters": range(12272, 12288),
+    "CJK Symbols and Punctuation": range(12288, 12352),
+    "Hiragana": range(12352, 12448),
+    "Katakana": range(12448, 12544),
+    "Bopomofo": range(12544, 12592),
+    "Hangul Compatibility Jamo": range(12592, 12688),
+    "Kanbun": range(12688, 12704),
+    "Bopomofo Extended": range(12704, 12736),
+    "CJK Strokes": range(12736, 12784),
+    "Katakana Phonetic Extensions": range(12784, 12800),
+    "Enclosed CJK Letters and Months": range(12800, 13056),
+    "CJK Compatibility": range(13056, 13312),
+    "CJK Unified Ideographs Extension A": range(13312, 19904),
+    "Yijing Hexagram Symbols": range(19904, 19968),
+    "CJK Unified Ideographs": range(19968, 40960),
+    "Yi Syllables": range(40960, 42128),
+    "Yi Radicals": range(42128, 42192),
+    "Lisu": range(42192, 42240),
+    "Vai": range(42240, 42560),
+    "Cyrillic Extended-B": range(42560, 42656),
+    "Bamum": range(42656, 42752),
+    "Modifier Tone Letters": range(42752, 42784),
+    "Latin Extended-D": range(42784, 43008),
+    "Syloti Nagri": range(43008, 43056),
+    "Common Indic Number Forms": range(43056, 43072),
+    "Phags-pa": range(43072, 43136),
+    "Saurashtra": range(43136, 43232),
+    "Devanagari Extended": range(43232, 43264),
+    "Kayah Li": range(43264, 43312),
+    "Rejang": range(43312, 43360),
+    "Hangul Jamo Extended-A": range(43360, 43392),
+    "Javanese": range(43392, 43488),
+    "Myanmar Extended-B": range(43488, 43520),
+    "Cham": range(43520, 43616),
+    "Myanmar Extended-A": range(43616, 43648),
+    "Tai Viet": range(43648, 43744),
+    "Meetei Mayek Extensions": range(43744, 43776),
+    "Ethiopic Extended-A": range(43776, 43824),
+    "Latin Extended-E": range(43824, 43888),
+    "Cherokee Supplement": range(43888, 43968),
+    "Meetei Mayek": range(43968, 44032),
+    "Hangul Syllables": range(44032, 55216),
+    "Hangul Jamo Extended-B": range(55216, 55296),
+    "High Surrogates": range(55296, 56192),
+    "High Private Use Surrogates": range(56192, 56320),
+    "Low Surrogates": range(56320, 57344),
+    "Private Use Area": range(57344, 63744),
+    "CJK Compatibility Ideographs": range(63744, 64256),
+    "Alphabetic Presentation Forms": range(64256, 64336),
+    "Arabic Presentation Forms-A": range(64336, 65024),
+    "Variation Selectors": range(65024, 65040),
+    "Vertical Forms": range(65040, 65056),
+    "Combining Half Marks": range(65056, 65072),
+    "CJK Compatibility Forms": range(65072, 65104),
+    "Small Form Variants": range(65104, 65136),
+    "Arabic Presentation Forms-B": range(65136, 65280),
+    "Halfwidth and Fullwidth Forms": range(65280, 65520),
+    "Specials": range(65520, 65536),
+    "Linear B Syllabary": range(65536, 65664),
+    "Linear B Ideograms": range(65664, 65792),
+    "Aegean Numbers": range(65792, 65856),
+    "Ancient Greek Numbers": range(65856, 65936),
+    "Ancient Symbols": range(65936, 66000),
+    "Phaistos Disc": range(66000, 66048),
+    "Lycian": range(66176, 66208),
+    "Carian": range(66208, 66272),
+    "Coptic Epact Numbers": range(66272, 66304),
+    "Old Italic": range(66304, 66352),
+    "Gothic": range(66352, 66384),
+    "Old Permic": range(66384, 66432),
+    "Ugaritic": range(66432, 66464),
+    "Old Persian": range(66464, 66528),
+    "Deseret": range(66560, 66640),
+    "Shavian": range(66640, 66688),
+    "Osmanya": range(66688, 66736),
+    "Osage": range(66736, 66816),
+    "Elbasan": range(66816, 66864),
+    "Caucasian Albanian": range(66864, 66928),
+    "Vithkuqi": range(66928, 67008),
+    "Linear A": range(67072, 67456),
+    "Latin Extended-F": range(67456, 67520),
+    "Cypriot Syllabary": range(67584, 67648),
+    "Imperial Aramaic": range(67648, 67680),
+    "Palmyrene": range(67680, 67712),
+    "Nabataean": range(67712, 67760),
+    "Hatran": range(67808, 67840),
+    "Phoenician": range(67840, 67872),
+    "Lydian": range(67872, 67904),
+    "Meroitic Hieroglyphs": range(67968, 68000),
+    "Meroitic Cursive": range(68000, 68096),
+    "Kharoshthi": range(68096, 68192),
+    "Old South Arabian": range(68192, 68224),
+    "Old North Arabian": range(68224, 68256),
+    "Manichaean": range(68288, 68352),
+    "Avestan": range(68352, 68416),
+    "Inscriptional Parthian": range(68416, 68448),
+    "Inscriptional Pahlavi": range(68448, 68480),
+    "Psalter Pahlavi": range(68480, 68528),
+    "Old Turkic": range(68608, 68688),
+    "Old Hungarian": range(68736, 68864),
+    "Hanifi Rohingya": range(68864, 68928),
+    "Rumi Numeral Symbols": range(69216, 69248),
+    "Yezidi": range(69248, 69312),
+    "Arabic Extended-C": range(69312, 69376),
+    "Old Sogdian": range(69376, 69424),
+    "Sogdian": range(69424, 69488),
+    "Old Uyghur": range(69488, 69552),
+    "Chorasmian": range(69552, 69600),
+    "Elymaic": range(69600, 69632),
+    "Brahmi": range(69632, 69760),
+    "Kaithi": range(69760, 69840),
+    "Sora Sompeng": range(69840, 69888),
+    "Chakma": range(69888, 69968),
+    "Mahajani": range(69968, 70016),
+    "Sharada": range(70016, 70112),
+    "Sinhala Archaic Numbers": range(70112, 70144),
+    "Khojki": range(70144, 70224),
+    "Multani": range(70272, 70320),
+    "Khudawadi": range(70320, 70400),
+    "Grantha": range(70400, 70528),
+    "Newa": range(70656, 70784),
+    "Tirhuta": range(70784, 70880),
+    "Siddham": range(71040, 71168),
+    "Modi": range(71168, 71264),
+    "Mongolian Supplement": range(71264, 71296),
+    "Takri": range(71296, 71376),
+    "Ahom": range(71424, 71504),
+    "Dogra": range(71680, 71760),
+    "Warang Citi": range(71840, 71936),
+    "Dives Akuru": range(71936, 72032),
+    "Nandinagari": range(72096, 72192),
+    "Zanabazar Square": range(72192, 72272),
+    "Soyombo": range(72272, 72368),
+    "Unified Canadian Aboriginal Syllabics Extended-A": range(72368, 72384),
+    "Pau Cin Hau": range(72384, 72448),
+    "Devanagari Extended-A": range(72448, 72544),
+    "Bhaiksuki": range(72704, 72816),
+    "Marchen": range(72816, 72896),
+    "Masaram Gondi": range(72960, 73056),
+    "Gunjala Gondi": range(73056, 73136),
+    "Makasar": range(73440, 73472),
+    "Kawi": range(73472, 73568),
+    "Lisu Supplement": range(73648, 73664),
+    "Tamil Supplement": range(73664, 73728),
+    "Cuneiform": range(73728, 74752),
+    "Cuneiform Numbers and Punctuation": range(74752, 74880),
+    "Early Dynastic Cuneiform": range(74880, 75088),
+    "Cypro-Minoan": range(77712, 77824),
+    "Egyptian Hieroglyphs": range(77824, 78896),
+    "Egyptian Hieroglyph Format Controls": range(78896, 78944),
+    "Anatolian Hieroglyphs": range(82944, 83584),
+    "Bamum Supplement": range(92160, 92736),
+    "Mro": range(92736, 92784),
+    "Tangsa": range(92784, 92880),
+    "Bassa Vah": range(92880, 92928),
+    "Pahawh Hmong": range(92928, 93072),
+    "Medefaidrin": range(93760, 93856),
+    "Miao": range(93952, 94112),
+    "Ideographic Symbols and Punctuation": range(94176, 94208),
+    "Tangut": range(94208, 100352),
+    "Tangut Components": range(100352, 101120),
+    "Khitan Small Script": range(101120, 101632),
+    "Tangut Supplement": range(101632, 101760),
+    "Kana Extended-B": range(110576, 110592),
+    "Kana Supplement": range(110592, 110848),
+    "Kana Extended-A": range(110848, 110896),
+    "Small Kana Extension": range(110896, 110960),
+    "Nushu": range(110960, 111360),
+    "Duployan": range(113664, 113824),
+    "Shorthand Format Controls": range(113824, 113840),
+    "Znamenny Musical Notation": range(118528, 118736),
+    "Byzantine Musical Symbols": range(118784, 119040),
+    "Musical Symbols": range(119040, 119296),
+    "Ancient Greek Musical Notation": range(119296, 119376),
+    "Kaktovik Numerals": range(119488, 119520),
+    "Mayan Numerals": range(119520, 119552),
+    "Tai Xuan Jing Symbols": range(119552, 119648),
+    "Counting Rod Numerals": range(119648, 119680),
+    "Mathematical Alphanumeric Symbols": range(119808, 120832),
+    "Sutton SignWriting": range(120832, 121520),
+    "Latin Extended-G": range(122624, 122880),
+    "Glagolitic Supplement": range(122880, 122928),
+    "Cyrillic Extended-D": range(122928, 123024),
+    "Nyiakeng Puachue Hmong": range(123136, 123216),
+    "Toto": range(123536, 123584),
+    "Wancho": range(123584, 123648),
+    "Nag Mundari": range(124112, 124160),
+    "Ethiopic Extended-B": range(124896, 124928),
+    "Mende Kikakui": range(124928, 125152),
+    "Adlam": range(125184, 125280),
+    "Indic Siyaq Numbers": range(126064, 126144),
+    "Ottoman Siyaq Numbers": range(126208, 126288),
+    "Arabic Mathematical Alphabetic Symbols": range(126464, 126720),
+    "Mahjong Tiles": range(126976, 127024),
+    "Domino Tiles": range(127024, 127136),
+    "Playing Cards": range(127136, 127232),
+    "Enclosed Alphanumeric Supplement": range(127232, 127488),
+    "Enclosed Ideographic Supplement": range(127488, 127744),
+    "Miscellaneous Symbols and Pictographs": range(127744, 128512),
+    "Emoticons range(Emoji)": range(128512, 128592),
+    "Ornamental Dingbats": range(128592, 128640),
+    "Transport and Map Symbols": range(128640, 128768),
+    "Alchemical Symbols": range(128768, 128896),
+    "Geometric Shapes Extended": range(128896, 129024),
+    "Supplemental Arrows-C": range(129024, 129280),
+    "Supplemental Symbols and Pictographs": range(129280, 129536),
+    "Chess Symbols": range(129536, 129648),
+    "Symbols and Pictographs Extended-A": range(129648, 129792),
+    "Symbols for Legacy Computing": range(129792, 130048),
+    "CJK Unified Ideographs Extension B": range(131072, 173792),
+    "CJK Unified Ideographs Extension C": range(173824, 177984),
+    "CJK Unified Ideographs Extension D": range(177984, 178208),
+    "CJK Unified Ideographs Extension E": range(178208, 183984),
+    "CJK Unified Ideographs Extension F": range(183984, 191472),
+    "CJK Compatibility Ideographs Supplement": range(194560, 195104),
+    "CJK Unified Ideographs Extension G": range(196608, 201552),
+    "CJK Unified Ideographs Extension H": range(201552, 205744),
+    "Tags": range(917504, 917632),
+    "Variation Selectors Supplement": range(917760, 918000),
+    "Supplementary Private Use Area-A": range(983040, 1048576),
+    "Supplementary Private Use Area-B": range(1048576, 1114112),
+}
+
+
+UNICODE_SECONDARY_RANGE_KEYWORD: list[str] = [
+    "Supplement",
+    "Extended",
+    "Extensions",
+    "Modifier",
+    "Marks",
+    "Punctuation",
+    "Symbols",
+    "Forms",
+    "Operators",
+    "Miscellaneous",
+    "Drawing",
+    "Block",
+    "Shapes",
+    "Supplemental",
+    "Tags",
+]
+
+RE_POSSIBLE_ENCODING_INDICATION = re_compile(
+    r"(?:(?:encoding)|(?:charset)|(?:coding))(?:[\:= ]{1,10})(?:[\"\']?)([a-zA-Z0-9\-_]+)(?:[\"\']?)",
+    IGNORECASE,
+)
+
+IANA_NO_ALIASES = [
+    "cp720",
+    "cp737",
+    "cp856",
+    "cp874",
+    "cp875",
+    "cp1006",
+    "koi8_r",
+    "koi8_t",
+    "koi8_u",
+]
+
+IANA_SUPPORTED: list[str] = sorted(
+    filter(
+        lambda x: x.endswith("_codec") is False
+        and x not in {"rot_13", "tactis", "mbcs"},
+        list(set(aliases.values())) + IANA_NO_ALIASES,
+    )
+)
+
+IANA_SUPPORTED_COUNT: int = len(IANA_SUPPORTED)
+
+# pre-computed code page that are similar using the function cp_similarity.
+IANA_SUPPORTED_SIMILAR: dict[str, list[str]] = {
+    "cp037": ["cp1026", "cp1140", "cp273", "cp500"],
+    "cp1026": ["cp037", "cp1140", "cp273", "cp500"],
+    "cp1125": ["cp866"],
+    "cp1140": ["cp037", "cp1026", "cp273", "cp500"],
+    "cp1250": ["iso8859_2"],
+    "cp1251": ["kz1048", "ptcp154"],
+    "cp1252": ["iso8859_15", "iso8859_9", "latin_1"],
+    "cp1253": ["iso8859_7"],
+    "cp1254": ["iso8859_15", "iso8859_9", "latin_1"],
+    "cp1257": ["iso8859_13"],
+    "cp273": ["cp037", "cp1026", "cp1140", "cp500"],
+    "cp437": ["cp850", "cp858", "cp860", "cp861", "cp862", "cp863", "cp865"],
+    "cp500": ["cp037", "cp1026", "cp1140", "cp273"],
+    "cp850": ["cp437", "cp857", "cp858", "cp865"],
+    "cp857": ["cp850", "cp858", "cp865"],
+    "cp858": ["cp437", "cp850", "cp857", "cp865"],
+    "cp860": ["cp437", "cp861", "cp862", "cp863", "cp865"],
+    "cp861": ["cp437", "cp860", "cp862", "cp863", "cp865"],
+    "cp862": ["cp437", "cp860", "cp861", "cp863", "cp865"],
+    "cp863": ["cp437", "cp860", "cp861", "cp862", "cp865"],
+    "cp865": ["cp437", "cp850", "cp857", "cp858", "cp860", "cp861", "cp862", "cp863"],
+    "cp866": ["cp1125"],
+    "iso8859_10": ["iso8859_14", "iso8859_15", "iso8859_4", "iso8859_9", "latin_1"],
+    "iso8859_11": ["tis_620"],
+    "iso8859_13": ["cp1257"],
+    "iso8859_14": [
+        "iso8859_10",
+        "iso8859_15",
+        "iso8859_16",
+        "iso8859_3",
+        "iso8859_9",
+        "latin_1",
+    ],
+    "iso8859_15": [
+        "cp1252",
+        "cp1254",
+        "iso8859_10",
+        "iso8859_14",
+        "iso8859_16",
+        "iso8859_3",
+        "iso8859_9",
+        "latin_1",
+    ],
+    "iso8859_16": [
+        "iso8859_14",
+        "iso8859_15",
+        "iso8859_2",
+        "iso8859_3",
+        "iso8859_9",
+        "latin_1",
+    ],
+    "iso8859_2": ["cp1250", "iso8859_16", "iso8859_4"],
+    "iso8859_3": ["iso8859_14", "iso8859_15", "iso8859_16", "iso8859_9", "latin_1"],
+    "iso8859_4": ["iso8859_10", "iso8859_2", "iso8859_9", "latin_1"],
+    "iso8859_7": ["cp1253"],
+    "iso8859_9": [
+        "cp1252",
+        "cp1254",
+        "cp1258",
+        "iso8859_10",
+        "iso8859_14",
+        "iso8859_15",
+        "iso8859_16",
+        "iso8859_3",
+        "iso8859_4",
+        "latin_1",
+    ],
+    "kz1048": ["cp1251", "ptcp154"],
+    "latin_1": [
+        "cp1252",
+        "cp1254",
+        "cp1258",
+        "iso8859_10",
+        "iso8859_14",
+        "iso8859_15",
+        "iso8859_16",
+        "iso8859_3",
+        "iso8859_4",
+        "iso8859_9",
+    ],
+    "mac_iceland": ["mac_roman", "mac_turkish"],
+    "mac_roman": ["mac_iceland", "mac_turkish"],
+    "mac_turkish": ["mac_iceland", "mac_roman"],
+    "ptcp154": ["cp1251", "kz1048"],
+    "tis_620": ["iso8859_11"],
+}
+
+
+CHARDET_CORRESPONDENCE: dict[str, str] = {
+    "iso2022_kr": "ISO-2022-KR",
+    "iso2022_jp": "ISO-2022-JP",
+    "euc_kr": "EUC-KR",
+    "tis_620": "TIS-620",
+    "utf_32": "UTF-32",
+    "euc_jp": "EUC-JP",
+    "koi8_r": "KOI8-R",
+    "iso8859_1": "ISO-8859-1",
+    "iso8859_2": "ISO-8859-2",
+    "iso8859_5": "ISO-8859-5",
+    "iso8859_6": "ISO-8859-6",
+    "iso8859_7": "ISO-8859-7",
+    "iso8859_8": "ISO-8859-8",
+    "utf_16": "UTF-16",
+    "cp855": "IBM855",
+    "mac_cyrillic": "MacCyrillic",
+    "gb2312": "GB2312",
+    "gb18030": "GB18030",
+    "cp932": "CP932",
+    "cp866": "IBM866",
+    "utf_8": "utf-8",
+    "utf_8_sig": "UTF-8-SIG",
+    "shift_jis": "SHIFT_JIS",
+    "big5": "Big5",
+    "cp1250": "windows-1250",
+    "cp1251": "windows-1251",
+    "cp1252": "Windows-1252",
+    "cp1253": "windows-1253",
+    "cp1255": "windows-1255",
+    "cp1256": "windows-1256",
+    "cp1254": "Windows-1254",
+    "cp949": "CP949",
+}
+
+
+COMMON_SAFE_ASCII_CHARACTERS: set[str] = {
+    "<",
+    ">",
+    "=",
+    ":",
+    "/",
+    "&",
+    ";",
+    "{",
+    "}",
+    "[",
+    "]",
+    ",",
+    "|",
+    '"',
+    "-",
+    "(",
+    ")",
+}
+
+
+KO_NAMES: set[str] = {"johab", "cp949", "euc_kr"}
+ZH_NAMES: set[str] = {"big5", "cp950", "big5hkscs", "hz"}
+
+# Logging LEVEL below DEBUG
+TRACE: int = 5
+
+
+# Language label that contain the em dash "—"
+# character are to be considered alternative seq to origin
+FREQUENCIES: dict[str, list[str]] = {
+    "English": [
+        "e",
+        "a",
+        "t",
+        "i",
+        "o",
+        "n",
+        "s",
+        "r",
+        "h",
+        "l",
+        "d",
+        "c",
+        "u",
+        "m",
+        "f",
+        "p",
+        "g",
+        "w",
+        "y",
+        "b",
+        "v",
+        "k",
+        "x",
+        "j",
+        "z",
+        "q",
+    ],
+    "English—": [
+        "e",
+        "a",
+        "t",
+        "i",
+        "o",
+        "n",
+        "s",
+        "r",
+        "h",
+        "l",
+        "d",
+        "c",
+        "m",
+        "u",
+        "f",
+        "p",
+        "g",
+        "w",
+        "b",
+        "y",
+        "v",
+        "k",
+        "j",
+        "x",
+        "z",
+        "q",
+    ],
+    "German": [
+        "e",
+        "n",
+        "i",
+        "r",
+        "s",
+        "t",
+        "a",
+        "d",
+        "h",
+        "u",
+        "l",
+        "g",
+        "o",
+        "c",
+        "m",
+        "b",
+        "f",
+        "k",
+        "w",
+        "z",
+        "p",
+        "v",
+        "ü",
+        "ä",
+        "ö",
+        "j",
+    ],
+    "French": [
+        "e",
+        "a",
+        "s",
+        "n",
+        "i",
+        "t",
+        "r",
+        "l",
+        "u",
+        "o",
+        "d",
+        "c",
+        "p",
+        "m",
+        "é",
+        "v",
+        "g",
+        "f",
+        "b",
+        "h",
+        "q",
+        "à",
+        "x",
+        "è",
+        "y",
+        "j",
+    ],
+    "Dutch": [
+        "e",
+        "n",
+        "a",
+        "i",
+        "r",
+        "t",
+        "o",
+        "d",
+        "s",
+        "l",
+        "g",
+        "h",
+        "v",
+        "m",
+        "u",
+        "k",
+        "c",
+        "p",
+        "b",
+        "w",
+        "j",
+        "z",
+        "f",
+        "y",
+        "x",
+        "ë",
+    ],
+    "Italian": [
+        "e",
+        "i",
+        "a",
+        "o",
+        "n",
+        "l",
+        "t",
+        "r",
+        "s",
+        "c",
+        "d",
+        "u",
+        "p",
+        "m",
+        "g",
+        "v",
+        "f",
+        "b",
+        "z",
+        "h",
+        "q",
+        "è",
+        "à",
+        "k",
+        "y",
+        "ò",
+    ],
+    "Polish": [
+        "a",
+        "i",
+        "o",
+        "e",
+        "n",
+        "r",
+        "z",
+        "w",
+        "s",
+        "c",
+        "t",
+        "k",
+        "y",
+        "d",
+        "p",
+        "m",
+        "u",
+        "l",
+        "j",
+        "ł",
+        "g",
+        "b",
+        "h",
+        "ą",
+        "ę",
+        "ó",
+    ],
+    "Spanish": [
+        "e",
+        "a",
+        "o",
+        "n",
+        "s",
+        "r",
+        "i",
+        "l",
+        "d",
+        "t",
+        "c",
+        "u",
+        "m",
+        "p",
+        "b",
+        "g",
+        "v",
+        "f",
+        "y",
+        "ó",
+        "h",
+        "q",
+        "í",
+        "j",
+        "z",
+        "á",
+    ],
+    "Russian": [
+        "о",
+        "а",
+        "е",
+        "и",
+        "н",
+        "с",
+        "т",
+        "р",
+        "в",
+        "л",
+        "к",
+        "м",
+        "д",
+        "п",
+        "у",
+        "г",
+        "я",
+        "ы",
+        "з",
+        "б",
+        "й",
+        "ь",
+        "ч",
+        "х",
+        "ж",
+        "ц",
+    ],
+    # Jap-Kanji
+    "Japanese": [
+        "人",
+        "一",
+        "大",
+        "亅",
+        "丁",
+        "丨",
+        "竹",
+        "笑",
+        "口",
+        "日",
+        "今",
+        "二",
+        "彳",
+        "行",
+        "十",
+        "土",
+        "丶",
+        "寸",
+        "寺",
+        "時",
+        "乙",
+        "丿",
+        "乂",
+        "气",
+        "気",
+        "冂",
+        "巾",
+        "亠",
+        "市",
+        "目",
+        "儿",
+        "見",
+        "八",
+        "小",
+        "凵",
+        "県",
+        "月",
+        "彐",
+        "門",
+        "間",
+        "木",
+        "東",
+        "山",
+        "出",
+        "本",
+        "中",
+        "刀",
+        "分",
+        "耳",
+        "又",
+        "取",
+        "最",
+        "言",
+        "田",
+        "心",
+        "思",
+        "刂",
+        "前",
+        "京",
+        "尹",
+        "事",
+        "生",
+        "厶",
+        "云",
+        "会",
+        "未",
+        "来",
+        "白",
+        "冫",
+        "楽",
+        "灬",
+        "馬",
+        "尸",
+        "尺",
+        "駅",
+        "明",
+        "耂",
+        "者",
+        "了",
+        "阝",
+        "都",
+        "高",
+        "卜",
+        "占",
+        "厂",
+        "广",
+        "店",
+        "子",
+        "申",
+        "奄",
+        "亻",
+        "俺",
+        "上",
+        "方",
+        "冖",
+        "学",
+        "衣",
+        "艮",
+        "食",
+        "自",
+    ],
+    # Jap-Katakana
+    "Japanese—": [
+        "ー",
+        "ン",
+        "ス",
+        "・",
+        "ル",
+        "ト",
+        "リ",
+        "イ",
+        "ア",
+        "ラ",
+        "ッ",
+        "ク",
+        "ド",
+        "シ",
+        "レ",
+        "ジ",
+        "タ",
+        "フ",
+        "ロ",
+        "カ",
+        "テ",
+        "マ",
+        "ィ",
+        "グ",
+        "バ",
+        "ム",
+        "プ",
+        "オ",
+        "コ",
+        "デ",
+        "ニ",
+        "ウ",
+        "メ",
+        "サ",
+        "ビ",
+        "ナ",
+        "ブ",
+        "ャ",
+        "エ",
+        "ュ",
+        "チ",
+        "キ",
+        "ズ",
+        "ダ",
+        "パ",
+        "ミ",
+        "ェ",
+        "ョ",
+        "ハ",
+        "セ",
+        "ベ",
+        "ガ",
+        "モ",
+        "ツ",
+        "ネ",
+        "ボ",
+        "ソ",
+        "ノ",
+        "ァ",
+        "ヴ",
+        "ワ",
+        "ポ",
+        "ペ",
+        "ピ",
+        "ケ",
+        "ゴ",
+        "ギ",
+        "ザ",
+        "ホ",
+        "ゲ",
+        "ォ",
+        "ヤ",
+        "ヒ",
+        "ユ",
+        "ヨ",
+        "ヘ",
+        "ゼ",
+        "ヌ",
+        "ゥ",
+        "ゾ",
+        "ヶ",
+        "ヂ",
+        "ヲ",
+        "ヅ",
+        "ヵ",
+        "ヱ",
+        "ヰ",
+        "ヮ",
+        "ヽ",
+        "゠",
+        "ヾ",
+        "ヷ",
+        "ヿ",
+        "ヸ",
+        "ヹ",
+        "ヺ",
+    ],
+    # Jap-Hiragana
+    "Japanese——": [
+        "の",
+        "に",
+        "る",
+        "た",
+        "と",
+        "は",
+        "し",
+        "い",
+        "を",
+        "で",
+        "て",
+        "が",
+        "な",
+        "れ",
+        "か",
+        "ら",
+        "さ",
+        "っ",
+        "り",
+        "す",
+        "あ",
+        "も",
+        "こ",
+        "ま",
+        "う",
+        "く",
+        "よ",
+        "き",
+        "ん",
+        "め",
+        "お",
+        "け",
+        "そ",
+        "つ",
+        "だ",
+        "や",
+        "え",
+        "ど",
+        "わ",
+        "ち",
+        "み",
+        "せ",
+        "じ",
+        "ば",
+        "へ",
+        "び",
+        "ず",
+        "ろ",
+        "ほ",
+        "げ",
+        "む",
+        "べ",
+        "ひ",
+        "ょ",
+        "ゆ",
+        "ぶ",
+        "ご",
+        "ゃ",
+        "ね",
+        "ふ",
+        "ぐ",
+        "ぎ",
+        "ぼ",
+        "ゅ",
+        "づ",
+        "ざ",
+        "ぞ",
+        "ぬ",
+        "ぜ",
+        "ぱ",
+        "ぽ",
+        "ぷ",
+        "ぴ",
+        "ぃ",
+        "ぁ",
+        "ぇ",
+        "ぺ",
+        "ゞ",
+        "ぢ",
+        "ぉ",
+        "ぅ",
+        "ゐ",
+        "ゝ",
+        "ゑ",
+        "゛",
+        "゜",
+        "ゎ",
+        "ゔ",
+        "゚",
+        "ゟ",
+        "゙",
+        "ゕ",
+        "ゖ",
+    ],
+    "Portuguese": [
+        "a",
+        "e",
+        "o",
+        "s",
+        "i",
+        "r",
+        "d",
+        "n",
+        "t",
+        "m",
+        "u",
+        "c",
+        "l",
+        "p",
+        "g",
+        "v",
+        "b",
+        "f",
+        "h",
+        "ã",
+        "q",
+        "é",
+        "ç",
+        "á",
+        "z",
+        "í",
+    ],
+    "Swedish": [
+        "e",
+        "a",
+        "n",
+        "r",
+        "t",
+        "s",
+        "i",
+        "l",
+        "d",
+        "o",
+        "m",
+        "k",
+        "g",
+        "v",
+        "h",
+        "f",
+        "u",
+        "p",
+        "ä",
+        "c",
+        "b",
+        "ö",
+        "å",
+        "y",
+        "j",
+        "x",
+    ],
+    "Chinese": [
+        "的",
+        "一",
+        "是",
+        "不",
+        "了",
+        "在",
+        "人",
+        "有",
+        "我",
+        "他",
+        "这",
+        "个",
+        "们",
+        "中",
+        "来",
+        "上",
+        "大",
+        "为",
+        "和",
+        "国",
+        "地",
+        "到",
+        "以",
+        "说",
+        "时",
+        "要",
+        "就",
+        "出",
+        "会",
+        "可",
+        "也",
+        "你",
+        "对",
+        "生",
+        "能",
+        "而",
+        "子",
+        "那",
+        "得",
+        "于",
+        "着",
+        "下",
+        "自",
+        "之",
+        "年",
+        "过",
+        "发",
+        "后",
+        "作",
+        "里",
+        "用",
+        "道",
+        "行",
+        "所",
+        "然",
+        "家",
+        "种",
+        "事",
+        "成",
+        "方",
+        "多",
+        "经",
+        "么",
+        "去",
+        "法",
+        "学",
+        "如",
+        "都",
+        "同",
+        "现",
+        "当",
+        "没",
+        "动",
+        "面",
+        "起",
+        "看",
+        "定",
+        "天",
+        "分",
+        "还",
+        "进",
+        "好",
+        "小",
+        "部",
+        "其",
+        "些",
+        "主",
+        "样",
+        "理",
+        "心",
+        "她",
+        "本",
+        "前",
+        "开",
+        "但",
+        "因",
+        "只",
+        "从",
+        "想",
+        "实",
+    ],
+    "Ukrainian": [
+        "о",
+        "а",
+        "н",
+        "і",
+        "и",
+        "р",
+        "в",
+        "т",
+        "е",
+        "с",
+        "к",
+        "л",
+        "у",
+        "д",
+        "м",
+        "п",
+        "з",
+        "я",
+        "ь",
+        "б",
+        "г",
+        "й",
+        "ч",
+        "х",
+        "ц",
+        "ї",
+    ],
+    "Norwegian": [
+        "e",
+        "r",
+        "n",
+        "t",
+        "a",
+        "s",
+        "i",
+        "o",
+        "l",
+        "d",
+        "g",
+        "k",
+        "m",
+        "v",
+        "f",
+        "p",
+        "u",
+        "b",
+        "h",
+        "å",
+        "y",
+        "j",
+        "ø",
+        "c",
+        "æ",
+        "w",
+    ],
+    "Finnish": [
+        "a",
+        "i",
+        "n",
+        "t",
+        "e",
+        "s",
+        "l",
+        "o",
+        "u",
+        "k",
+        "ä",
+        "m",
+        "r",
+        "v",
+        "j",
+        "h",
+        "p",
+        "y",
+        "d",
+        "ö",
+        "g",
+        "c",
+        "b",
+        "f",
+        "w",
+        "z",
+    ],
+    "Vietnamese": [
+        "n",
+        "h",
+        "t",
+        "i",
+        "c",
+        "g",
+        "a",
+        "o",
+        "u",
+        "m",
+        "l",
+        "r",
+        "à",
+        "đ",
+        "s",
+        "e",
+        "v",
+        "p",
+        "b",
+        "y",
+        "ư",
+        "d",
+        "á",
+        "k",
+        "ộ",
+        "ế",
+    ],
+    "Czech": [
+        "o",
+        "e",
+        "a",
+        "n",
+        "t",
+        "s",
+        "i",
+        "l",
+        "v",
+        "r",
+        "k",
+        "d",
+        "u",
+        "m",
+        "p",
+        "í",
+        "c",
+        "h",
+        "z",
+        "á",
+        "y",
+        "j",
+        "b",
+        "ě",
+        "é",
+        "ř",
+    ],
+    "Hungarian": [
+        "e",
+        "a",
+        "t",
+        "l",
+        "s",
+        "n",
+        "k",
+        "r",
+        "i",
+        "o",
+        "z",
+        "á",
+        "é",
+        "g",
+        "m",
+        "b",
+        "y",
+        "v",
+        "d",
+        "h",
+        "u",
+        "p",
+        "j",
+        "ö",
+        "f",
+        "c",
+    ],
+    "Korean": [
+        "이",
+        "다",
+        "에",
+        "의",
+        "는",
+        "로",
+        "하",
+        "을",
+        "가",
+        "고",
+        "지",
+        "서",
+        "한",
+        "은",
+        "기",
+        "으",
+        "년",
+        "대",
+        "사",
+        "시",
+        "를",
+        "리",
+        "도",
+        "인",
+        "스",
+        "일",
+    ],
+    "Indonesian": [
+        "a",
+        "n",
+        "e",
+        "i",
+        "r",
+        "t",
+        "u",
+        "s",
+        "d",
+        "k",
+        "m",
+        "l",
+        "g",
+        "p",
+        "b",
+        "o",
+        "h",
+        "y",
+        "j",
+        "c",
+        "w",
+        "f",
+        "v",
+        "z",
+        "x",
+        "q",
+    ],
+    "Turkish": [
+        "a",
+        "e",
+        "i",
+        "n",
+        "r",
+        "l",
+        "ı",
+        "k",
+        "d",
+        "t",
+        "s",
+        "m",
+        "y",
+        "u",
+        "o",
+        "b",
+        "ü",
+        "ş",
+        "v",
+        "g",
+        "z",
+        "h",
+        "c",
+        "p",
+        "ç",
+        "ğ",
+    ],
+    "Romanian": [
+        "e",
+        "i",
+        "a",
+        "r",
+        "n",
+        "t",
+        "u",
+        "l",
+        "o",
+        "c",
+        "s",
+        "d",
+        "p",
+        "m",
+        "ă",
+        "f",
+        "v",
+        "î",
+        "g",
+        "b",
+        "ș",
+        "ț",
+        "z",
+        "h",
+        "â",
+        "j",
+    ],
+    "Farsi": [
+        "ا",
+        "ی",
+        "ر",
+        "د",
+        "ن",
+        "ه",
+        "و",
+        "م",
+        "ت",
+        "ب",
+        "س",
+        "ل",
+        "ک",
+        "ش",
+        "ز",
+        "ف",
+        "گ",
+        "ع",
+        "خ",
+        "ق",
+        "ج",
+        "آ",
+        "پ",
+        "ح",
+        "ط",
+        "ص",
+    ],
+    "Arabic": [
+        "ا",
+        "ل",
+        "ي",
+        "م",
+        "و",
+        "ن",
+        "ر",
+        "ت",
+        "ب",
+        "ة",
+        "ع",
+        "د",
+        "س",
+        "ف",
+        "ه",
+        "ك",
+        "ق",
+        "أ",
+        "ح",
+        "ج",
+        "ش",
+        "ط",
+        "ص",
+        "ى",
+        "خ",
+        "إ",
+    ],
+    "Danish": [
+        "e",
+        "r",
+        "n",
+        "t",
+        "a",
+        "i",
+        "s",
+        "d",
+        "l",
+        "o",
+        "g",
+        "m",
+        "k",
+        "f",
+        "v",
+        "u",
+        "b",
+        "h",
+        "p",
+        "å",
+        "y",
+        "ø",
+        "æ",
+        "c",
+        "j",
+        "w",
+    ],
+    "Serbian": [
+        "а",
+        "и",
+        "о",
+        "е",
+        "н",
+        "р",
+        "с",
+        "у",
+        "т",
+        "к",
+        "ј",
+        "в",
+        "д",
+        "м",
+        "п",
+        "л",
+        "г",
+        "з",
+        "б",
+        "a",
+        "i",
+        "e",
+        "o",
+        "n",
+        "ц",
+        "ш",
+    ],
+    "Lithuanian": [
+        "i",
+        "a",
+        "s",
+        "o",
+        "r",
+        "e",
+        "t",
+        "n",
+        "u",
+        "k",
+        "m",
+        "l",
+        "p",
+        "v",
+        "d",
+        "j",
+        "g",
+        "ė",
+        "b",
+        "y",
+        "ų",
+        "š",
+        "ž",
+        "c",
+        "ą",
+        "į",
+    ],
+    "Slovene": [
+        "e",
+        "a",
+        "i",
+        "o",
+        "n",
+        "r",
+        "s",
+        "l",
+        "t",
+        "j",
+        "v",
+        "k",
+        "d",
+        "p",
+        "m",
+        "u",
+        "z",
+        "b",
+        "g",
+        "h",
+        "č",
+        "c",
+        "š",
+        "ž",
+        "f",
+        "y",
+    ],
+    "Slovak": [
+        "o",
+        "a",
+        "e",
+        "n",
+        "i",
+        "r",
+        "v",
+        "t",
+        "s",
+        "l",
+        "k",
+        "d",
+        "m",
+        "p",
+        "u",
+        "c",
+        "h",
+        "j",
+        "b",
+        "z",
+        "á",
+        "y",
+        "ý",
+        "í",
+        "č",
+        "é",
+    ],
+    "Hebrew": [
+        "י",
+        "ו",
+        "ה",
+        "ל",
+        "ר",
+        "ב",
+        "ת",
+        "מ",
+        "א",
+        "ש",
+        "נ",
+        "ע",
+        "ם",
+        "ד",
+        "ק",
+        "ח",
+        "פ",
+        "ס",
+        "כ",
+        "ג",
+        "ט",
+        "צ",
+        "ן",
+        "ז",
+        "ך",
+    ],
+    "Bulgarian": [
+        "а",
+        "и",
+        "о",
+        "е",
+        "н",
+        "т",
+        "р",
+        "с",
+        "в",
+        "л",
+        "к",
+        "д",
+        "п",
+        "м",
+        "з",
+        "г",
+        "я",
+        "ъ",
+        "у",
+        "б",
+        "ч",
+        "ц",
+        "й",
+        "ж",
+        "щ",
+        "х",
+    ],
+    "Croatian": [
+        "a",
+        "i",
+        "o",
+        "e",
+        "n",
+        "r",
+        "j",
+        "s",
+        "t",
+        "u",
+        "k",
+        "l",
+        "v",
+        "d",
+        "m",
+        "p",
+        "g",
+        "z",
+        "b",
+        "c",
+        "č",
+        "h",
+        "š",
+        "ž",
+        "ć",
+        "f",
+    ],
+    "Hindi": [
+        "क",
+        "र",
+        "स",
+        "न",
+        "त",
+        "म",
+        "ह",
+        "प",
+        "य",
+        "ल",
+        "व",
+        "ज",
+        "द",
+        "ग",
+        "ब",
+        "श",
+        "ट",
+        "अ",
+        "ए",
+        "थ",
+        "भ",
+        "ड",
+        "च",
+        "ध",
+        "ष",
+        "इ",
+    ],
+    "Estonian": [
+        "a",
+        "i",
+        "e",
+        "s",
+        "t",
+        "l",
+        "u",
+        "n",
+        "o",
+        "k",
+        "r",
+        "d",
+        "m",
+        "v",
+        "g",
+        "p",
+        "j",
+        "h",
+        "ä",
+        "b",
+        "õ",
+        "ü",
+        "f",
+        "c",
+        "ö",
+        "y",
+    ],
+    "Thai": [
+        "า",
+        "น",
+        "ร",
+        "อ",
+        "ก",
+        "เ",
+        "ง",
+        "ม",
+        "ย",
+        "ล",
+        "ว",
+        "ด",
+        "ท",
+        "ส",
+        "ต",
+        "ะ",
+        "ป",
+        "บ",
+        "ค",
+        "ห",
+        "แ",
+        "จ",
+        "พ",
+        "ช",
+        "ข",
+        "ใ",
+    ],
+    "Greek": [
+        "α",
+        "τ",
+        "ο",
+        "ι",
+        "ε",
+        "ν",
+        "ρ",
+        "σ",
+        "κ",
+        "η",
+        "π",
+        "ς",
+        "υ",
+        "μ",
+        "λ",
+        "ί",
+        "ό",
+        "ά",
+        "γ",
+        "έ",
+        "δ",
+        "ή",
+        "ω",
+        "χ",
+        "θ",
+        "ύ",
+    ],
+    "Tamil": [
+        "க",
+        "த",
+        "ப",
+        "ட",
+        "ர",
+        "ம",
+        "ல",
+        "ன",
+        "வ",
+        "ற",
+        "ய",
+        "ள",
+        "ச",
+        "ந",
+        "இ",
+        "ண",
+        "அ",
+        "ஆ",
+        "ழ",
+        "ங",
+        "எ",
+        "உ",
+        "ஒ",
+        "ஸ",
+    ],
+    "Kazakh": [
+        "а",
+        "ы",
+        "е",
+        "н",
+        "т",
+        "р",
+        "л",
+        "і",
+        "д",
+        "с",
+        "м",
+        "қ",
+        "к",
+        "о",
+        "б",
+        "и",
+        "у",
+        "ғ",
+        "ж",
+        "ң",
+        "з",
+        "ш",
+        "й",
+        "п",
+        "г",
+        "ө",
+    ],
+}
+
+LANGUAGE_SUPPORTED_COUNT: int = len(FREQUENCIES)
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/legacy.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/legacy.py
new file mode 100644
index 00000000..a2f53451
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/legacy.py
@@ -0,0 +1,66 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any
+from warnings import warn
+
+from .api import from_bytes
+from .constant import CHARDET_CORRESPONDENCE
+
+# TODO: remove this check when dropping Python 3.7 support
+if TYPE_CHECKING:
+    from typing_extensions import TypedDict
+
+    class ResultDict(TypedDict):
+        encoding: str | None
+        language: str
+        confidence: float | None
+
+
+def detect(
+    byte_str: bytes, should_rename_legacy: bool = False, **kwargs: Any
+) -> ResultDict:
+    """
+    chardet legacy method
+    Detect the encoding of the given byte string. It should be mostly backward-compatible.
+    Encoding name will match Chardet own writing whenever possible. (Not on encoding name unsupported by it)
+    This function is deprecated and should be used to migrate your project easily, consult the documentation for
+    further information. Not planned for removal.
+
+    :param byte_str:     The byte sequence to examine.
+    :param should_rename_legacy:  Should we rename legacy encodings
+                                  to their more modern equivalents?
+    """
+    if len(kwargs):
+        warn(
+            f"charset-normalizer disregard arguments '{','.join(list(kwargs.keys()))}' in legacy function detect()"
+        )
+
+    if not isinstance(byte_str, (bytearray, bytes)):
+        raise TypeError(  # pragma: nocover
+            "Expected object of type bytes or bytearray, got: " "{}".format(
+                type(byte_str)
+            )
+        )
+
+    if isinstance(byte_str, bytearray):
+        byte_str = bytes(byte_str)
+
+    r = from_bytes(byte_str).best()
+
+    encoding = r.encoding if r is not None else None
+    language = r.language if r is not None and r.language != "Unknown" else ""
+    confidence = 1.0 - r.chaos if r is not None else None
+
+    # Note: CharsetNormalizer does not return 'UTF-8-SIG' as the sig get stripped in the detection/normalization process
+    # but chardet does return 'utf-8-sig' and it is a valid codec name.
+    if r is not None and encoding == "utf_8" and r.bom:
+        encoding += "_sig"
+
+    if should_rename_legacy is False and encoding in CHARDET_CORRESPONDENCE:
+        encoding = CHARDET_CORRESPONDENCE[encoding]
+
+    return {
+        "encoding": encoding,
+        "language": language,
+        "confidence": confidence,
+    }
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.cpython-312-darwin.so
new file mode 100755
index 00000000..b83755a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.py
new file mode 100644
index 00000000..9ed59a86
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md.py
@@ -0,0 +1,630 @@
+from __future__ import annotations
+
+from functools import lru_cache
+from logging import getLogger
+
+from .constant import (
+    COMMON_SAFE_ASCII_CHARACTERS,
+    TRACE,
+    UNICODE_SECONDARY_RANGE_KEYWORD,
+)
+from .utils import (
+    is_accentuated,
+    is_arabic,
+    is_arabic_isolated_form,
+    is_case_variable,
+    is_cjk,
+    is_emoticon,
+    is_hangul,
+    is_hiragana,
+    is_katakana,
+    is_latin,
+    is_punctuation,
+    is_separator,
+    is_symbol,
+    is_thai,
+    is_unprintable,
+    remove_accent,
+    unicode_range,
+)
+
+
+class MessDetectorPlugin:
+    """
+    Base abstract class used for mess detection plugins.
+    All detectors MUST extend and implement given methods.
+    """
+
+    def eligible(self, character: str) -> bool:
+        """
+        Determine if given character should be fed in.
+        """
+        raise NotImplementedError  # pragma: nocover
+
+    def feed(self, character: str) -> None:
+        """
+        The main routine to be executed upon character.
+        Insert the logic in witch the text would be considered chaotic.
+        """
+        raise NotImplementedError  # pragma: nocover
+
+    def reset(self) -> None:  # pragma: no cover
+        """
+        Permit to reset the plugin to the initial state.
+        """
+        raise NotImplementedError
+
+    @property
+    def ratio(self) -> float:
+        """
+        Compute the chaos ratio based on what your feed() has seen.
+        Must NOT be lower than 0.; No restriction gt 0.
+        """
+        raise NotImplementedError  # pragma: nocover
+
+
+class TooManySymbolOrPunctuationPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._punctuation_count: int = 0
+        self._symbol_count: int = 0
+        self._character_count: int = 0
+
+        self._last_printable_char: str | None = None
+        self._frenzy_symbol_in_word: bool = False
+
+    def eligible(self, character: str) -> bool:
+        return character.isprintable()
+
+    def feed(self, character: str) -> None:
+        self._character_count += 1
+
+        if (
+            character != self._last_printable_char
+            and character not in COMMON_SAFE_ASCII_CHARACTERS
+        ):
+            if is_punctuation(character):
+                self._punctuation_count += 1
+            elif (
+                character.isdigit() is False
+                and is_symbol(character)
+                and is_emoticon(character) is False
+            ):
+                self._symbol_count += 2
+
+        self._last_printable_char = character
+
+    def reset(self) -> None:  # Abstract
+        self._punctuation_count = 0
+        self._character_count = 0
+        self._symbol_count = 0
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count == 0:
+            return 0.0
+
+        ratio_of_punctuation: float = (
+            self._punctuation_count + self._symbol_count
+        ) / self._character_count
+
+        return ratio_of_punctuation if ratio_of_punctuation >= 0.3 else 0.0
+
+
+class TooManyAccentuatedPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._character_count: int = 0
+        self._accentuated_count: int = 0
+
+    def eligible(self, character: str) -> bool:
+        return character.isalpha()
+
+    def feed(self, character: str) -> None:
+        self._character_count += 1
+
+        if is_accentuated(character):
+            self._accentuated_count += 1
+
+    def reset(self) -> None:  # Abstract
+        self._character_count = 0
+        self._accentuated_count = 0
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count < 8:
+            return 0.0
+
+        ratio_of_accentuation: float = self._accentuated_count / self._character_count
+        return ratio_of_accentuation if ratio_of_accentuation >= 0.35 else 0.0
+
+
+class UnprintablePlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._unprintable_count: int = 0
+        self._character_count: int = 0
+
+    def eligible(self, character: str) -> bool:
+        return True
+
+    def feed(self, character: str) -> None:
+        if is_unprintable(character):
+            self._unprintable_count += 1
+        self._character_count += 1
+
+    def reset(self) -> None:  # Abstract
+        self._unprintable_count = 0
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count == 0:
+            return 0.0
+
+        return (self._unprintable_count * 8) / self._character_count
+
+
+class SuspiciousDuplicateAccentPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._successive_count: int = 0
+        self._character_count: int = 0
+
+        self._last_latin_character: str | None = None
+
+    def eligible(self, character: str) -> bool:
+        return character.isalpha() and is_latin(character)
+
+    def feed(self, character: str) -> None:
+        self._character_count += 1
+        if (
+            self._last_latin_character is not None
+            and is_accentuated(character)
+            and is_accentuated(self._last_latin_character)
+        ):
+            if character.isupper() and self._last_latin_character.isupper():
+                self._successive_count += 1
+            # Worse if its the same char duplicated with different accent.
+            if remove_accent(character) == remove_accent(self._last_latin_character):
+                self._successive_count += 1
+        self._last_latin_character = character
+
+    def reset(self) -> None:  # Abstract
+        self._successive_count = 0
+        self._character_count = 0
+        self._last_latin_character = None
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count == 0:
+            return 0.0
+
+        return (self._successive_count * 2) / self._character_count
+
+
+class SuspiciousRange(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._suspicious_successive_range_count: int = 0
+        self._character_count: int = 0
+        self._last_printable_seen: str | None = None
+
+    def eligible(self, character: str) -> bool:
+        return character.isprintable()
+
+    def feed(self, character: str) -> None:
+        self._character_count += 1
+
+        if (
+            character.isspace()
+            or is_punctuation(character)
+            or character in COMMON_SAFE_ASCII_CHARACTERS
+        ):
+            self._last_printable_seen = None
+            return
+
+        if self._last_printable_seen is None:
+            self._last_printable_seen = character
+            return
+
+        unicode_range_a: str | None = unicode_range(self._last_printable_seen)
+        unicode_range_b: str | None = unicode_range(character)
+
+        if is_suspiciously_successive_range(unicode_range_a, unicode_range_b):
+            self._suspicious_successive_range_count += 1
+
+        self._last_printable_seen = character
+
+    def reset(self) -> None:  # Abstract
+        self._character_count = 0
+        self._suspicious_successive_range_count = 0
+        self._last_printable_seen = None
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count <= 13:
+            return 0.0
+
+        ratio_of_suspicious_range_usage: float = (
+            self._suspicious_successive_range_count * 2
+        ) / self._character_count
+
+        return ratio_of_suspicious_range_usage
+
+
+class SuperWeirdWordPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._word_count: int = 0
+        self._bad_word_count: int = 0
+        self._foreign_long_count: int = 0
+
+        self._is_current_word_bad: bool = False
+        self._foreign_long_watch: bool = False
+
+        self._character_count: int = 0
+        self._bad_character_count: int = 0
+
+        self._buffer: str = ""
+        self._buffer_accent_count: int = 0
+        self._buffer_glyph_count: int = 0
+
+    def eligible(self, character: str) -> bool:
+        return True
+
+    def feed(self, character: str) -> None:
+        if character.isalpha():
+            self._buffer += character
+            if is_accentuated(character):
+                self._buffer_accent_count += 1
+            if (
+                self._foreign_long_watch is False
+                and (is_latin(character) is False or is_accentuated(character))
+                and is_cjk(character) is False
+                and is_hangul(character) is False
+                and is_katakana(character) is False
+                and is_hiragana(character) is False
+                and is_thai(character) is False
+            ):
+                self._foreign_long_watch = True
+            if (
+                is_cjk(character)
+                or is_hangul(character)
+                or is_katakana(character)
+                or is_hiragana(character)
+                or is_thai(character)
+            ):
+                self._buffer_glyph_count += 1
+            return
+        if not self._buffer:
+            return
+        if (
+            character.isspace() or is_punctuation(character) or is_separator(character)
+        ) and self._buffer:
+            self._word_count += 1
+            buffer_length: int = len(self._buffer)
+
+            self._character_count += buffer_length
+
+            if buffer_length >= 4:
+                if self._buffer_accent_count / buffer_length >= 0.5:
+                    self._is_current_word_bad = True
+                # Word/Buffer ending with an upper case accentuated letter are so rare,
+                # that we will consider them all as suspicious. Same weight as foreign_long suspicious.
+                elif (
+                    is_accentuated(self._buffer[-1])
+                    and self._buffer[-1].isupper()
+                    and all(_.isupper() for _ in self._buffer) is False
+                ):
+                    self._foreign_long_count += 1
+                    self._is_current_word_bad = True
+                elif self._buffer_glyph_count == 1:
+                    self._is_current_word_bad = True
+                    self._foreign_long_count += 1
+            if buffer_length >= 24 and self._foreign_long_watch:
+                camel_case_dst = [
+                    i
+                    for c, i in zip(self._buffer, range(0, buffer_length))
+                    if c.isupper()
+                ]
+                probable_camel_cased: bool = False
+
+                if camel_case_dst and (len(camel_case_dst) / buffer_length <= 0.3):
+                    probable_camel_cased = True
+
+                if not probable_camel_cased:
+                    self._foreign_long_count += 1
+                    self._is_current_word_bad = True
+
+            if self._is_current_word_bad:
+                self._bad_word_count += 1
+                self._bad_character_count += len(self._buffer)
+                self._is_current_word_bad = False
+
+            self._foreign_long_watch = False
+            self._buffer = ""
+            self._buffer_accent_count = 0
+            self._buffer_glyph_count = 0
+        elif (
+            character not in {"<", ">", "-", "=", "~", "|", "_"}
+            and character.isdigit() is False
+            and is_symbol(character)
+        ):
+            self._is_current_word_bad = True
+            self._buffer += character
+
+    def reset(self) -> None:  # Abstract
+        self._buffer = ""
+        self._is_current_word_bad = False
+        self._foreign_long_watch = False
+        self._bad_word_count = 0
+        self._word_count = 0
+        self._character_count = 0
+        self._bad_character_count = 0
+        self._foreign_long_count = 0
+
+    @property
+    def ratio(self) -> float:
+        if self._word_count <= 10 and self._foreign_long_count == 0:
+            return 0.0
+
+        return self._bad_character_count / self._character_count
+
+
+class CjkInvalidStopPlugin(MessDetectorPlugin):
+    """
+    GB(Chinese) based encoding often render the stop incorrectly when the content does not fit and
+    can be easily detected. Searching for the overuse of '丅' and '丄'.
+    """
+
+    def __init__(self) -> None:
+        self._wrong_stop_count: int = 0
+        self._cjk_character_count: int = 0
+
+    def eligible(self, character: str) -> bool:
+        return True
+
+    def feed(self, character: str) -> None:
+        if character in {"丅", "丄"}:
+            self._wrong_stop_count += 1
+            return
+        if is_cjk(character):
+            self._cjk_character_count += 1
+
+    def reset(self) -> None:  # Abstract
+        self._wrong_stop_count = 0
+        self._cjk_character_count = 0
+
+    @property
+    def ratio(self) -> float:
+        if self._cjk_character_count < 16:
+            return 0.0
+        return self._wrong_stop_count / self._cjk_character_count
+
+
+class ArchaicUpperLowerPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._buf: bool = False
+
+        self._character_count_since_last_sep: int = 0
+
+        self._successive_upper_lower_count: int = 0
+        self._successive_upper_lower_count_final: int = 0
+
+        self._character_count: int = 0
+
+        self._last_alpha_seen: str | None = None
+        self._current_ascii_only: bool = True
+
+    def eligible(self, character: str) -> bool:
+        return True
+
+    def feed(self, character: str) -> None:
+        is_concerned = character.isalpha() and is_case_variable(character)
+        chunk_sep = is_concerned is False
+
+        if chunk_sep and self._character_count_since_last_sep > 0:
+            if (
+                self._character_count_since_last_sep <= 64
+                and character.isdigit() is False
+                and self._current_ascii_only is False
+            ):
+                self._successive_upper_lower_count_final += (
+                    self._successive_upper_lower_count
+                )
+
+            self._successive_upper_lower_count = 0
+            self._character_count_since_last_sep = 0
+            self._last_alpha_seen = None
+            self._buf = False
+            self._character_count += 1
+            self._current_ascii_only = True
+
+            return
+
+        if self._current_ascii_only is True and character.isascii() is False:
+            self._current_ascii_only = False
+
+        if self._last_alpha_seen is not None:
+            if (character.isupper() and self._last_alpha_seen.islower()) or (
+                character.islower() and self._last_alpha_seen.isupper()
+            ):
+                if self._buf is True:
+                    self._successive_upper_lower_count += 2
+                    self._buf = False
+                else:
+                    self._buf = True
+            else:
+                self._buf = False
+
+        self._character_count += 1
+        self._character_count_since_last_sep += 1
+        self._last_alpha_seen = character
+
+    def reset(self) -> None:  # Abstract
+        self._character_count = 0
+        self._character_count_since_last_sep = 0
+        self._successive_upper_lower_count = 0
+        self._successive_upper_lower_count_final = 0
+        self._last_alpha_seen = None
+        self._buf = False
+        self._current_ascii_only = True
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count == 0:
+            return 0.0
+
+        return self._successive_upper_lower_count_final / self._character_count
+
+
+class ArabicIsolatedFormPlugin(MessDetectorPlugin):
+    def __init__(self) -> None:
+        self._character_count: int = 0
+        self._isolated_form_count: int = 0
+
+    def reset(self) -> None:  # Abstract
+        self._character_count = 0
+        self._isolated_form_count = 0
+
+    def eligible(self, character: str) -> bool:
+        return is_arabic(character)
+
+    def feed(self, character: str) -> None:
+        self._character_count += 1
+
+        if is_arabic_isolated_form(character):
+            self._isolated_form_count += 1
+
+    @property
+    def ratio(self) -> float:
+        if self._character_count < 8:
+            return 0.0
+
+        isolated_form_usage: float = self._isolated_form_count / self._character_count
+
+        return isolated_form_usage
+
+
+@lru_cache(maxsize=1024)
+def is_suspiciously_successive_range(
+    unicode_range_a: str | None, unicode_range_b: str | None
+) -> bool:
+    """
+    Determine if two Unicode range seen next to each other can be considered as suspicious.
+    """
+    if unicode_range_a is None or unicode_range_b is None:
+        return True
+
+    if unicode_range_a == unicode_range_b:
+        return False
+
+    if "Latin" in unicode_range_a and "Latin" in unicode_range_b:
+        return False
+
+    if "Emoticons" in unicode_range_a or "Emoticons" in unicode_range_b:
+        return False
+
+    # Latin characters can be accompanied with a combining diacritical mark
+    # eg. Vietnamese.
+    if ("Latin" in unicode_range_a or "Latin" in unicode_range_b) and (
+        "Combining" in unicode_range_a or "Combining" in unicode_range_b
+    ):
+        return False
+
+    keywords_range_a, keywords_range_b = (
+        unicode_range_a.split(" "),
+        unicode_range_b.split(" "),
+    )
+
+    for el in keywords_range_a:
+        if el in UNICODE_SECONDARY_RANGE_KEYWORD:
+            continue
+        if el in keywords_range_b:
+            return False
+
+    # Japanese Exception
+    range_a_jp_chars, range_b_jp_chars = (
+        unicode_range_a
+        in (
+            "Hiragana",
+            "Katakana",
+        ),
+        unicode_range_b in ("Hiragana", "Katakana"),
+    )
+    if (range_a_jp_chars or range_b_jp_chars) and (
+        "CJK" in unicode_range_a or "CJK" in unicode_range_b
+    ):
+        return False
+    if range_a_jp_chars and range_b_jp_chars:
+        return False
+
+    if "Hangul" in unicode_range_a or "Hangul" in unicode_range_b:
+        if "CJK" in unicode_range_a or "CJK" in unicode_range_b:
+            return False
+        if unicode_range_a == "Basic Latin" or unicode_range_b == "Basic Latin":
+            return False
+
+    # Chinese/Japanese use dedicated range for punctuation and/or separators.
+    if ("CJK" in unicode_range_a or "CJK" in unicode_range_b) or (
+        unicode_range_a in ["Katakana", "Hiragana"]
+        and unicode_range_b in ["Katakana", "Hiragana"]
+    ):
+        if "Punctuation" in unicode_range_a or "Punctuation" in unicode_range_b:
+            return False
+        if "Forms" in unicode_range_a or "Forms" in unicode_range_b:
+            return False
+        if unicode_range_a == "Basic Latin" or unicode_range_b == "Basic Latin":
+            return False
+
+    return True
+
+
+@lru_cache(maxsize=2048)
+def mess_ratio(
+    decoded_sequence: str, maximum_threshold: float = 0.2, debug: bool = False
+) -> float:
+    """
+    Compute a mess ratio given a decoded bytes sequence. The maximum threshold does stop the computation earlier.
+    """
+
+    detectors: list[MessDetectorPlugin] = [
+        md_class() for md_class in MessDetectorPlugin.__subclasses__()
+    ]
+
+    length: int = len(decoded_sequence) + 1
+
+    mean_mess_ratio: float = 0.0
+
+    if length < 512:
+        intermediary_mean_mess_ratio_calc: int = 32
+    elif length <= 1024:
+        intermediary_mean_mess_ratio_calc = 64
+    else:
+        intermediary_mean_mess_ratio_calc = 128
+
+    for character, index in zip(decoded_sequence + "\n", range(length)):
+        for detector in detectors:
+            if detector.eligible(character):
+                detector.feed(character)
+
+        if (
+            index > 0 and index % intermediary_mean_mess_ratio_calc == 0
+        ) or index == length - 1:
+            mean_mess_ratio = sum(dt.ratio for dt in detectors)
+
+            if mean_mess_ratio >= maximum_threshold:
+                break
+
+    if debug:
+        logger = getLogger("charset_normalizer")
+
+        logger.log(
+            TRACE,
+            "Mess-detector extended-analysis start. "
+            f"intermediary_mean_mess_ratio_calc={intermediary_mean_mess_ratio_calc} mean_mess_ratio={mean_mess_ratio} "
+            f"maximum_threshold={maximum_threshold}",
+        )
+
+        if len(decoded_sequence) > 16:
+            logger.log(TRACE, f"Starting with: {decoded_sequence[:16]}")
+            logger.log(TRACE, f"Ending with: {decoded_sequence[-16::]}")
+
+        for dt in detectors:
+            logger.log(TRACE, f"{dt.__class__}: {dt.ratio}")
+
+    return round(mean_mess_ratio, 3)
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/md__mypyc.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md__mypyc.cpython-312-darwin.so
new file mode 100755
index 00000000..523afb3c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/charset_normalizer/md__mypyc.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/models.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/models.py
new file mode 100644
index 00000000..1042758f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/models.py
@@ -0,0 +1,360 @@
+from __future__ import annotations
+
+from encodings.aliases import aliases
+from hashlib import sha256
+from json import dumps
+from re import sub
+from typing import Any, Iterator, List, Tuple
+
+from .constant import RE_POSSIBLE_ENCODING_INDICATION, TOO_BIG_SEQUENCE
+from .utils import iana_name, is_multi_byte_encoding, unicode_range
+
+
+class CharsetMatch:
+    def __init__(
+        self,
+        payload: bytes,
+        guessed_encoding: str,
+        mean_mess_ratio: float,
+        has_sig_or_bom: bool,
+        languages: CoherenceMatches,
+        decoded_payload: str | None = None,
+        preemptive_declaration: str | None = None,
+    ):
+        self._payload: bytes = payload
+
+        self._encoding: str = guessed_encoding
+        self._mean_mess_ratio: float = mean_mess_ratio
+        self._languages: CoherenceMatches = languages
+        self._has_sig_or_bom: bool = has_sig_or_bom
+        self._unicode_ranges: list[str] | None = None
+
+        self._leaves: list[CharsetMatch] = []
+        self._mean_coherence_ratio: float = 0.0
+
+        self._output_payload: bytes | None = None
+        self._output_encoding: str | None = None
+
+        self._string: str | None = decoded_payload
+
+        self._preemptive_declaration: str | None = preemptive_declaration
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, CharsetMatch):
+            if isinstance(other, str):
+                return iana_name(other) == self.encoding
+            return False
+        return self.encoding == other.encoding and self.fingerprint == other.fingerprint
+
+    def __lt__(self, other: object) -> bool:
+        """
+        Implemented to make sorted available upon CharsetMatches items.
+        """
+        if not isinstance(other, CharsetMatch):
+            raise ValueError
+
+        chaos_difference: float = abs(self.chaos - other.chaos)
+        coherence_difference: float = abs(self.coherence - other.coherence)
+
+        # Below 1% difference --> Use Coherence
+        if chaos_difference < 0.01 and coherence_difference > 0.02:
+            return self.coherence > other.coherence
+        elif chaos_difference < 0.01 and coherence_difference <= 0.02:
+            # When having a difficult decision, use the result that decoded as many multi-byte as possible.
+            # preserve RAM usage!
+            if len(self._payload) >= TOO_BIG_SEQUENCE:
+                return self.chaos < other.chaos
+            return self.multi_byte_usage > other.multi_byte_usage
+
+        return self.chaos < other.chaos
+
+    @property
+    def multi_byte_usage(self) -> float:
+        return 1.0 - (len(str(self)) / len(self.raw))
+
+    def __str__(self) -> str:
+        # Lazy Str Loading
+        if self._string is None:
+            self._string = str(self._payload, self._encoding, "strict")
+        return self._string
+
+    def __repr__(self) -> str:
+        return f"<CharsetMatch '{self.encoding}' bytes({self.fingerprint})>"
+
+    def add_submatch(self, other: CharsetMatch) -> None:
+        if not isinstance(other, CharsetMatch) or other == self:
+            raise ValueError(
+                "Unable to add instance <{}> as a submatch of a CharsetMatch".format(
+                    other.__class__
+                )
+            )
+
+        other._string = None  # Unload RAM usage; dirty trick.
+        self._leaves.append(other)
+
+    @property
+    def encoding(self) -> str:
+        return self._encoding
+
+    @property
+    def encoding_aliases(self) -> list[str]:
+        """
+        Encoding name are known by many name, using this could help when searching for IBM855 when it's listed as CP855.
+        """
+        also_known_as: list[str] = []
+        for u, p in aliases.items():
+            if self.encoding == u:
+                also_known_as.append(p)
+            elif self.encoding == p:
+                also_known_as.append(u)
+        return also_known_as
+
+    @property
+    def bom(self) -> bool:
+        return self._has_sig_or_bom
+
+    @property
+    def byte_order_mark(self) -> bool:
+        return self._has_sig_or_bom
+
+    @property
+    def languages(self) -> list[str]:
+        """
+        Return the complete list of possible languages found in decoded sequence.
+        Usually not really useful. Returned list may be empty even if 'language' property return something != 'Unknown'.
+        """
+        return [e[0] for e in self._languages]
+
+    @property
+    def language(self) -> str:
+        """
+        Most probable language found in decoded sequence. If none were detected or inferred, the property will return
+        "Unknown".
+        """
+        if not self._languages:
+            # Trying to infer the language based on the given encoding
+            # Its either English or we should not pronounce ourselves in certain cases.
+            if "ascii" in self.could_be_from_charset:
+                return "English"
+
+            # doing it there to avoid circular import
+            from charset_normalizer.cd import encoding_languages, mb_encoding_languages
+
+            languages = (
+                mb_encoding_languages(self.encoding)
+                if is_multi_byte_encoding(self.encoding)
+                else encoding_languages(self.encoding)
+            )
+
+            if len(languages) == 0 or "Latin Based" in languages:
+                return "Unknown"
+
+            return languages[0]
+
+        return self._languages[0][0]
+
+    @property
+    def chaos(self) -> float:
+        return self._mean_mess_ratio
+
+    @property
+    def coherence(self) -> float:
+        if not self._languages:
+            return 0.0
+        return self._languages[0][1]
+
+    @property
+    def percent_chaos(self) -> float:
+        return round(self.chaos * 100, ndigits=3)
+
+    @property
+    def percent_coherence(self) -> float:
+        return round(self.coherence * 100, ndigits=3)
+
+    @property
+    def raw(self) -> bytes:
+        """
+        Original untouched bytes.
+        """
+        return self._payload
+
+    @property
+    def submatch(self) -> list[CharsetMatch]:
+        return self._leaves
+
+    @property
+    def has_submatch(self) -> bool:
+        return len(self._leaves) > 0
+
+    @property
+    def alphabets(self) -> list[str]:
+        if self._unicode_ranges is not None:
+            return self._unicode_ranges
+        # list detected ranges
+        detected_ranges: list[str | None] = [unicode_range(char) for char in str(self)]
+        # filter and sort
+        self._unicode_ranges = sorted(list({r for r in detected_ranges if r}))
+        return self._unicode_ranges
+
+    @property
+    def could_be_from_charset(self) -> list[str]:
+        """
+        The complete list of encoding that output the exact SAME str result and therefore could be the originating
+        encoding.
+        This list does include the encoding available in property 'encoding'.
+        """
+        return [self._encoding] + [m.encoding for m in self._leaves]
+
+    def output(self, encoding: str = "utf_8") -> bytes:
+        """
+        Method to get re-encoded bytes payload using given target encoding. Default to UTF-8.
+        Any errors will be simply ignored by the encoder NOT replaced.
+        """
+        if self._output_encoding is None or self._output_encoding != encoding:
+            self._output_encoding = encoding
+            decoded_string = str(self)
+            if (
+                self._preemptive_declaration is not None
+                and self._preemptive_declaration.lower()
+                not in ["utf-8", "utf8", "utf_8"]
+            ):
+                patched_header = sub(
+                    RE_POSSIBLE_ENCODING_INDICATION,
+                    lambda m: m.string[m.span()[0] : m.span()[1]].replace(
+                        m.groups()[0],
+                        iana_name(self._output_encoding).replace("_", "-"),  # type: ignore[arg-type]
+                    ),
+                    decoded_string[:8192],
+                    count=1,
+                )
+
+                decoded_string = patched_header + decoded_string[8192:]
+
+            self._output_payload = decoded_string.encode(encoding, "replace")
+
+        return self._output_payload  # type: ignore
+
+    @property
+    def fingerprint(self) -> str:
+        """
+        Retrieve the unique SHA256 computed using the transformed (re-encoded) payload. Not the original one.
+        """
+        return sha256(self.output()).hexdigest()
+
+
+class CharsetMatches:
+    """
+    Container with every CharsetMatch items ordered by default from most probable to the less one.
+    Act like a list(iterable) but does not implements all related methods.
+    """
+
+    def __init__(self, results: list[CharsetMatch] | None = None):
+        self._results: list[CharsetMatch] = sorted(results) if results else []
+
+    def __iter__(self) -> Iterator[CharsetMatch]:
+        yield from self._results
+
+    def __getitem__(self, item: int | str) -> CharsetMatch:
+        """
+        Retrieve a single item either by its position or encoding name (alias may be used here).
+        Raise KeyError upon invalid index or encoding not present in results.
+        """
+        if isinstance(item, int):
+            return self._results[item]
+        if isinstance(item, str):
+            item = iana_name(item, False)
+            for result in self._results:
+                if item in result.could_be_from_charset:
+                    return result
+        raise KeyError
+
+    def __len__(self) -> int:
+        return len(self._results)
+
+    def __bool__(self) -> bool:
+        return len(self._results) > 0
+
+    def append(self, item: CharsetMatch) -> None:
+        """
+        Insert a single match. Will be inserted accordingly to preserve sort.
+        Can be inserted as a submatch.
+        """
+        if not isinstance(item, CharsetMatch):
+            raise ValueError(
+                "Cannot append instance '{}' to CharsetMatches".format(
+                    str(item.__class__)
+                )
+            )
+        # We should disable the submatch factoring when the input file is too heavy (conserve RAM usage)
+        if len(item.raw) < TOO_BIG_SEQUENCE:
+            for match in self._results:
+                if match.fingerprint == item.fingerprint and match.chaos == item.chaos:
+                    match.add_submatch(item)
+                    return
+        self._results.append(item)
+        self._results = sorted(self._results)
+
+    def best(self) -> CharsetMatch | None:
+        """
+        Simply return the first match. Strict equivalent to matches[0].
+        """
+        if not self._results:
+            return None
+        return self._results[0]
+
+    def first(self) -> CharsetMatch | None:
+        """
+        Redundant method, call the method best(). Kept for BC reasons.
+        """
+        return self.best()
+
+
+CoherenceMatch = Tuple[str, float]
+CoherenceMatches = List[CoherenceMatch]
+
+
+class CliDetectionResult:
+    def __init__(
+        self,
+        path: str,
+        encoding: str | None,
+        encoding_aliases: list[str],
+        alternative_encodings: list[str],
+        language: str,
+        alphabets: list[str],
+        has_sig_or_bom: bool,
+        chaos: float,
+        coherence: float,
+        unicode_path: str | None,
+        is_preferred: bool,
+    ):
+        self.path: str = path
+        self.unicode_path: str | None = unicode_path
+        self.encoding: str | None = encoding
+        self.encoding_aliases: list[str] = encoding_aliases
+        self.alternative_encodings: list[str] = alternative_encodings
+        self.language: str = language
+        self.alphabets: list[str] = alphabets
+        self.has_sig_or_bom: bool = has_sig_or_bom
+        self.chaos: float = chaos
+        self.coherence: float = coherence
+        self.is_preferred: bool = is_preferred
+
+    @property
+    def __dict__(self) -> dict[str, Any]:  # type: ignore
+        return {
+            "path": self.path,
+            "encoding": self.encoding,
+            "encoding_aliases": self.encoding_aliases,
+            "alternative_encodings": self.alternative_encodings,
+            "language": self.language,
+            "alphabets": self.alphabets,
+            "has_sig_or_bom": self.has_sig_or_bom,
+            "chaos": self.chaos,
+            "coherence": self.coherence,
+            "unicode_path": self.unicode_path,
+            "is_preferred": self.is_preferred,
+        }
+
+    def to_json(self) -> str:
+        return dumps(self.__dict__, ensure_ascii=True, indent=4)
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/py.typed b/backend/venv/lib/python3.12/site-packages/charset_normalizer/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/utils.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/utils.py
new file mode 100644
index 00000000..0175e0a9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/utils.py
@@ -0,0 +1,408 @@
+from __future__ import annotations
+
+import importlib
+import logging
+import unicodedata
+from codecs import IncrementalDecoder
+from encodings.aliases import aliases
+from functools import lru_cache
+from re import findall
+from typing import Generator
+
+from _multibytecodec import (  # type: ignore[import-not-found,import]
+    MultibyteIncrementalDecoder,
+)
+
+from .constant import (
+    ENCODING_MARKS,
+    IANA_SUPPORTED_SIMILAR,
+    RE_POSSIBLE_ENCODING_INDICATION,
+    UNICODE_RANGES_COMBINED,
+    UNICODE_SECONDARY_RANGE_KEYWORD,
+    UTF8_MAXIMAL_ALLOCATION,
+)
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_accentuated(character: str) -> bool:
+    try:
+        description: str = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+    return (
+        "WITH GRAVE" in description
+        or "WITH ACUTE" in description
+        or "WITH CEDILLA" in description
+        or "WITH DIAERESIS" in description
+        or "WITH CIRCUMFLEX" in description
+        or "WITH TILDE" in description
+        or "WITH MACRON" in description
+        or "WITH RING ABOVE" in description
+    )
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def remove_accent(character: str) -> str:
+    decomposed: str = unicodedata.decomposition(character)
+    if not decomposed:
+        return character
+
+    codes: list[str] = decomposed.split(" ")
+
+    return chr(int(codes[0], 16))
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def unicode_range(character: str) -> str | None:
+    """
+    Retrieve the Unicode range official name from a single character.
+    """
+    character_ord: int = ord(character)
+
+    for range_name, ord_range in UNICODE_RANGES_COMBINED.items():
+        if character_ord in ord_range:
+            return range_name
+
+    return None
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_latin(character: str) -> bool:
+    try:
+        description: str = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+    return "LATIN" in description
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_punctuation(character: str) -> bool:
+    character_category: str = unicodedata.category(character)
+
+    if "P" in character_category:
+        return True
+
+    character_range: str | None = unicode_range(character)
+
+    if character_range is None:
+        return False
+
+    return "Punctuation" in character_range
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_symbol(character: str) -> bool:
+    character_category: str = unicodedata.category(character)
+
+    if "S" in character_category or "N" in character_category:
+        return True
+
+    character_range: str | None = unicode_range(character)
+
+    if character_range is None:
+        return False
+
+    return "Forms" in character_range and character_category != "Lo"
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_emoticon(character: str) -> bool:
+    character_range: str | None = unicode_range(character)
+
+    if character_range is None:
+        return False
+
+    return "Emoticons" in character_range or "Pictographs" in character_range
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_separator(character: str) -> bool:
+    if character.isspace() or character in {"｜", "+", "<", ">"}:
+        return True
+
+    character_category: str = unicodedata.category(character)
+
+    return "Z" in character_category or character_category in {"Po", "Pd", "Pc"}
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_case_variable(character: str) -> bool:
+    return character.islower() != character.isupper()
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_cjk(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "CJK" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_hiragana(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "HIRAGANA" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_katakana(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "KATAKANA" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_hangul(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "HANGUL" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_thai(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "THAI" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_arabic(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "ARABIC" in character_name
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_arabic_isolated_form(character: str) -> bool:
+    try:
+        character_name = unicodedata.name(character)
+    except ValueError:  # Defensive: unicode database outdated?
+        return False
+
+    return "ARABIC" in character_name and "ISOLATED FORM" in character_name
+
+
+@lru_cache(maxsize=len(UNICODE_RANGES_COMBINED))
+def is_unicode_range_secondary(range_name: str) -> bool:
+    return any(keyword in range_name for keyword in UNICODE_SECONDARY_RANGE_KEYWORD)
+
+
+@lru_cache(maxsize=UTF8_MAXIMAL_ALLOCATION)
+def is_unprintable(character: str) -> bool:
+    return (
+        character.isspace() is False  # includes \n \t \r \v
+        and character.isprintable() is False
+        and character != "\x1a"  # Why? Its the ASCII substitute character.
+        and character != "\ufeff"  # bug discovered in Python,
+        # Zero Width No-Break Space located in 	Arabic Presentation Forms-B, Unicode 1.1 not acknowledged as space.
+    )
+
+
+def any_specified_encoding(sequence: bytes, search_zone: int = 8192) -> str | None:
+    """
+    Extract using ASCII-only decoder any specified encoding in the first n-bytes.
+    """
+    if not isinstance(sequence, bytes):
+        raise TypeError
+
+    seq_len: int = len(sequence)
+
+    results: list[str] = findall(
+        RE_POSSIBLE_ENCODING_INDICATION,
+        sequence[: min(seq_len, search_zone)].decode("ascii", errors="ignore"),
+    )
+
+    if len(results) == 0:
+        return None
+
+    for specified_encoding in results:
+        specified_encoding = specified_encoding.lower().replace("-", "_")
+
+        encoding_alias: str
+        encoding_iana: str
+
+        for encoding_alias, encoding_iana in aliases.items():
+            if encoding_alias == specified_encoding:
+                return encoding_iana
+            if encoding_iana == specified_encoding:
+                return encoding_iana
+
+    return None
+
+
+@lru_cache(maxsize=128)
+def is_multi_byte_encoding(name: str) -> bool:
+    """
+    Verify is a specific encoding is a multi byte one based on it IANA name
+    """
+    return name in {
+        "utf_8",
+        "utf_8_sig",
+        "utf_16",
+        "utf_16_be",
+        "utf_16_le",
+        "utf_32",
+        "utf_32_le",
+        "utf_32_be",
+        "utf_7",
+    } or issubclass(
+        importlib.import_module(f"encodings.{name}").IncrementalDecoder,
+        MultibyteIncrementalDecoder,
+    )
+
+
+def identify_sig_or_bom(sequence: bytes) -> tuple[str | None, bytes]:
+    """
+    Identify and extract SIG/BOM in given sequence.
+    """
+
+    for iana_encoding in ENCODING_MARKS:
+        marks: bytes | list[bytes] = ENCODING_MARKS[iana_encoding]
+
+        if isinstance(marks, bytes):
+            marks = [marks]
+
+        for mark in marks:
+            if sequence.startswith(mark):
+                return iana_encoding, mark
+
+    return None, b""
+
+
+def should_strip_sig_or_bom(iana_encoding: str) -> bool:
+    return iana_encoding not in {"utf_16", "utf_32"}
+
+
+def iana_name(cp_name: str, strict: bool = True) -> str:
+    """Returns the Python normalized encoding name (Not the IANA official name)."""
+    cp_name = cp_name.lower().replace("-", "_")
+
+    encoding_alias: str
+    encoding_iana: str
+
+    for encoding_alias, encoding_iana in aliases.items():
+        if cp_name in [encoding_alias, encoding_iana]:
+            return encoding_iana
+
+    if strict:
+        raise ValueError(f"Unable to retrieve IANA for '{cp_name}'")
+
+    return cp_name
+
+
+def cp_similarity(iana_name_a: str, iana_name_b: str) -> float:
+    if is_multi_byte_encoding(iana_name_a) or is_multi_byte_encoding(iana_name_b):
+        return 0.0
+
+    decoder_a = importlib.import_module(f"encodings.{iana_name_a}").IncrementalDecoder
+    decoder_b = importlib.import_module(f"encodings.{iana_name_b}").IncrementalDecoder
+
+    id_a: IncrementalDecoder = decoder_a(errors="ignore")
+    id_b: IncrementalDecoder = decoder_b(errors="ignore")
+
+    character_match_count: int = 0
+
+    for i in range(255):
+        to_be_decoded: bytes = bytes([i])
+        if id_a.decode(to_be_decoded) == id_b.decode(to_be_decoded):
+            character_match_count += 1
+
+    return character_match_count / 254
+
+
+def is_cp_similar(iana_name_a: str, iana_name_b: str) -> bool:
+    """
+    Determine if two code page are at least 80% similar. IANA_SUPPORTED_SIMILAR dict was generated using
+    the function cp_similarity.
+    """
+    return (
+        iana_name_a in IANA_SUPPORTED_SIMILAR
+        and iana_name_b in IANA_SUPPORTED_SIMILAR[iana_name_a]
+    )
+
+
+def set_logging_handler(
+    name: str = "charset_normalizer",
+    level: int = logging.INFO,
+    format_string: str = "%(asctime)s | %(levelname)s | %(message)s",
+) -> None:
+    logger = logging.getLogger(name)
+    logger.setLevel(level)
+
+    handler = logging.StreamHandler()
+    handler.setFormatter(logging.Formatter(format_string))
+    logger.addHandler(handler)
+
+
+def cut_sequence_chunks(
+    sequences: bytes,
+    encoding_iana: str,
+    offsets: range,
+    chunk_size: int,
+    bom_or_sig_available: bool,
+    strip_sig_or_bom: bool,
+    sig_payload: bytes,
+    is_multi_byte_decoder: bool,
+    decoded_payload: str | None = None,
+) -> Generator[str, None, None]:
+    if decoded_payload and is_multi_byte_decoder is False:
+        for i in offsets:
+            chunk = decoded_payload[i : i + chunk_size]
+            if not chunk:
+                break
+            yield chunk
+    else:
+        for i in offsets:
+            chunk_end = i + chunk_size
+            if chunk_end > len(sequences) + 8:
+                continue
+
+            cut_sequence = sequences[i : i + chunk_size]
+
+            if bom_or_sig_available and strip_sig_or_bom is False:
+                cut_sequence = sig_payload + cut_sequence
+
+            chunk = cut_sequence.decode(
+                encoding_iana,
+                errors="ignore" if is_multi_byte_decoder else "strict",
+            )
+
+            # multi-byte bad cutting detector and adjustment
+            # not the cleanest way to perform that fix but clever enough for now.
+            if is_multi_byte_decoder and i > 0:
+                chunk_partial_size_chk: int = min(chunk_size, 16)
+
+                if (
+                    decoded_payload
+                    and chunk[:chunk_partial_size_chk] not in decoded_payload
+                ):
+                    for j in range(i, i - 4, -1):
+                        cut_sequence = sequences[j:chunk_end]
+
+                        if bom_or_sig_available and strip_sig_or_bom is False:
+                            cut_sequence = sig_payload + cut_sequence
+
+                        chunk = cut_sequence.decode(encoding_iana, errors="ignore")
+
+                        if chunk[:chunk_partial_size_chk] in decoded_payload:
+                            break
+
+            yield chunk
diff --git a/backend/venv/lib/python3.12/site-packages/charset_normalizer/version.py b/backend/venv/lib/python3.12/site-packages/charset_normalizer/version.py
new file mode 100644
index 00000000..f85e8929
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/charset_normalizer/version.py
@@ -0,0 +1,8 @@
+"""
+Expose version
+"""
+
+from __future__ import annotations
+
+__version__ = "3.4.1"
+VERSION = __version__.split(".")
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/LICENSE.txt
new file mode 100644
index 00000000..d12a8491
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2014 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/METADATA
new file mode 100644
index 00000000..366d1a7e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/METADATA
@@ -0,0 +1,74 @@
+Metadata-Version: 2.3
+Name: click
+Version: 8.1.8
+Summary: Composable command line interface toolkit
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Typing :: Typed
+Requires-Dist: colorama; platform_system == 'Windows'
+Requires-Dist: importlib-metadata; python_version < '3.8'
+Project-URL: Changes, https://click.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://click.palletsprojects.com/
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Source, https://github.com/pallets/click/
+
+# $ click_
+
+Click is a Python package for creating beautiful command line interfaces
+in a composable way with as little code as necessary. It's the "Command
+Line Interface Creation Kit". It's highly configurable but comes with
+sensible defaults out of the box.
+
+It aims to make the process of writing command line tools quick and fun
+while also preventing any frustration caused by the inability to
+implement an intended CLI API.
+
+Click in three points:
+
+-   Arbitrary nesting of commands
+-   Automatic help page generation
+-   Supports lazy loading of subcommands at runtime
+
+
+## A Simple Example
+
+```python
+import click
+
+@click.command()
+@click.option("--count", default=1, help="Number of greetings.")
+@click.option("--name", prompt="Your name", help="The person to greet.")
+def hello(count, name):
+    """Simple program that greets NAME for a total of COUNT times."""
+    for _ in range(count):
+        click.echo(f"Hello, {name}!")
+
+if __name__ == '__main__':
+    hello()
+```
+
+```
+$ python hello.py --count=3
+Your name: Click
+Hello, Click!
+Hello, Click!
+Hello, Click!
+```
+
+
+## Donate
+
+The Pallets organization develops and supports Click and other popular
+packages. In order to grow the community of contributors and users, and
+allow the maintainers to devote more time to the projects, [please
+donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
+
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/RECORD
new file mode 100644
index 00000000..2227ac11
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/RECORD
@@ -0,0 +1,39 @@
+click-8.1.8.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+click-8.1.8.dist-info/LICENSE.txt,sha256=morRBqOU6FO_4h9C9OctWSgZoigF2ZG18ydQKSkrZY0,1475
+click-8.1.8.dist-info/METADATA,sha256=WJtQ6uGS2ybLfvUE4vC0XIhIBr4yFGwjrMBR2fiCQ-Q,2263
+click-8.1.8.dist-info/RECORD,,
+click-8.1.8.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+click-8.1.8.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+click/__init__.py,sha256=j1DJeCbga4ribkv5uyvIAzI0oFN13fW9mevDKShFelo,3188
+click/__pycache__/__init__.cpython-312.pyc,,
+click/__pycache__/_compat.cpython-312.pyc,,
+click/__pycache__/_termui_impl.cpython-312.pyc,,
+click/__pycache__/_textwrap.cpython-312.pyc,,
+click/__pycache__/_winconsole.cpython-312.pyc,,
+click/__pycache__/core.cpython-312.pyc,,
+click/__pycache__/decorators.cpython-312.pyc,,
+click/__pycache__/exceptions.cpython-312.pyc,,
+click/__pycache__/formatting.cpython-312.pyc,,
+click/__pycache__/globals.cpython-312.pyc,,
+click/__pycache__/parser.cpython-312.pyc,,
+click/__pycache__/shell_completion.cpython-312.pyc,,
+click/__pycache__/termui.cpython-312.pyc,,
+click/__pycache__/testing.cpython-312.pyc,,
+click/__pycache__/types.cpython-312.pyc,,
+click/__pycache__/utils.cpython-312.pyc,,
+click/_compat.py,sha256=IGKh_J5QdfKELitnRfTGHneejWxoCw_NX9tfMbdcg3w,18730
+click/_termui_impl.py,sha256=a5z7I9gOFeMmu7Gb6_RPyQ8GPuVP1EeblixcWSPSQPk,24783
+click/_textwrap.py,sha256=10fQ64OcBUMuK7mFvh8363_uoOxPlRItZBmKzRJDgoY,1353
+click/_winconsole.py,sha256=5ju3jQkcZD0W27WEMGqmEP4y_crUVzPCqsX_FYb7BO0,7860
+click/core.py,sha256=Q1nEVdctZwvIPOlt4vfHko0TYnHCeE40UEEul8Wpyvs,114748
+click/decorators.py,sha256=7t6F-QWowtLh6F_6l-4YV4Y4yNTcqFQEu9i37zIz68s,18925
+click/exceptions.py,sha256=V7zDT6emqJ8iNl0kF1P5kpFmLMWQ1T1L7aNNKM4YR0w,9600
+click/formatting.py,sha256=Frf0-5W33-loyY_i9qrwXR8-STnW3m5gvyxLVUdyxyk,9706
+click/globals.py,sha256=cuJ6Bbo073lgEEmhjr394PeM-QFmXM-Ci-wmfsd7H5g,1954
+click/parser.py,sha256=h4sndcpF5OHrZQN8vD8IWb5OByvW7ABbhRToxovrqS8,19067
+click/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+click/shell_completion.py,sha256=TR0dXEGcvWb9Eo3aaQEXGhnvNS3FF4H4QcuLnvAvYo4,18636
+click/termui.py,sha256=dLxiS70UOvIYBda_nEEZaPAFOVDVmRs1sEPMuLDowQo,28310
+click/testing.py,sha256=3RA8anCf7TZ8-5RAF5it2Te-aWXBAL5VLasQnMiC2ZQ,16282
+click/types.py,sha256=BD5Qqq4h-8kawBmOIzJlmq4xzThAf4wCvaOLZSBDNx0,36422
+click/utils.py,sha256=ce-IrO9ilII76LGkU354pOdHbepM8UftfNH7SfMU_28,20330
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click-8.1.8.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/click/__init__.py b/backend/venv/lib/python3.12/site-packages/click/__init__.py
new file mode 100644
index 00000000..2610d0e1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/__init__.py
@@ -0,0 +1,75 @@
+"""
+Click is a simple Python module inspired by the stdlib optparse to make
+writing command line scripts fun. Unlike other modules, it's based
+around a simple API that does not come with too much magic and is
+composable.
+"""
+
+from .core import Argument as Argument
+from .core import BaseCommand as BaseCommand
+from .core import Command as Command
+from .core import CommandCollection as CommandCollection
+from .core import Context as Context
+from .core import Group as Group
+from .core import MultiCommand as MultiCommand
+from .core import Option as Option
+from .core import Parameter as Parameter
+from .decorators import argument as argument
+from .decorators import command as command
+from .decorators import confirmation_option as confirmation_option
+from .decorators import group as group
+from .decorators import help_option as help_option
+from .decorators import HelpOption as HelpOption
+from .decorators import make_pass_decorator as make_pass_decorator
+from .decorators import option as option
+from .decorators import pass_context as pass_context
+from .decorators import pass_obj as pass_obj
+from .decorators import password_option as password_option
+from .decorators import version_option as version_option
+from .exceptions import Abort as Abort
+from .exceptions import BadArgumentUsage as BadArgumentUsage
+from .exceptions import BadOptionUsage as BadOptionUsage
+from .exceptions import BadParameter as BadParameter
+from .exceptions import ClickException as ClickException
+from .exceptions import FileError as FileError
+from .exceptions import MissingParameter as MissingParameter
+from .exceptions import NoSuchOption as NoSuchOption
+from .exceptions import UsageError as UsageError
+from .formatting import HelpFormatter as HelpFormatter
+from .formatting import wrap_text as wrap_text
+from .globals import get_current_context as get_current_context
+from .parser import OptionParser as OptionParser
+from .termui import clear as clear
+from .termui import confirm as confirm
+from .termui import echo_via_pager as echo_via_pager
+from .termui import edit as edit
+from .termui import getchar as getchar
+from .termui import launch as launch
+from .termui import pause as pause
+from .termui import progressbar as progressbar
+from .termui import prompt as prompt
+from .termui import secho as secho
+from .termui import style as style
+from .termui import unstyle as unstyle
+from .types import BOOL as BOOL
+from .types import Choice as Choice
+from .types import DateTime as DateTime
+from .types import File as File
+from .types import FLOAT as FLOAT
+from .types import FloatRange as FloatRange
+from .types import INT as INT
+from .types import IntRange as IntRange
+from .types import ParamType as ParamType
+from .types import Path as Path
+from .types import STRING as STRING
+from .types import Tuple as Tuple
+from .types import UNPROCESSED as UNPROCESSED
+from .types import UUID as UUID
+from .utils import echo as echo
+from .utils import format_filename as format_filename
+from .utils import get_app_dir as get_app_dir
+from .utils import get_binary_stream as get_binary_stream
+from .utils import get_text_stream as get_text_stream
+from .utils import open_file as open_file
+
+__version__ = "8.1.8"
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7e83ce72
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_compat.cpython-312.pyc
new file mode 100644
index 00000000..3c2a4f8b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/_termui_impl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_termui_impl.cpython-312.pyc
new file mode 100644
index 00000000..a1c82858
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_termui_impl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/_textwrap.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_textwrap.cpython-312.pyc
new file mode 100644
index 00000000..30a42ab8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_textwrap.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/_winconsole.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_winconsole.cpython-312.pyc
new file mode 100644
index 00000000..8a6a168d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/_winconsole.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..c08fa7cd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/decorators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/decorators.cpython-312.pyc
new file mode 100644
index 00000000..a043db21
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/decorators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..a113e919
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/formatting.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/formatting.cpython-312.pyc
new file mode 100644
index 00000000..8a67d025
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/formatting.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/globals.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/globals.cpython-312.pyc
new file mode 100644
index 00000000..c7d72bfb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/globals.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/parser.cpython-312.pyc
new file mode 100644
index 00000000..44dced3e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/shell_completion.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/shell_completion.cpython-312.pyc
new file mode 100644
index 00000000..6f89f37d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/shell_completion.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/termui.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/termui.cpython-312.pyc
new file mode 100644
index 00000000..36f5a619
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/termui.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/testing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/testing.cpython-312.pyc
new file mode 100644
index 00000000..0e638c49
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/testing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..5bf8abb3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/click/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..bad2a1d8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/click/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/click/_compat.py b/backend/venv/lib/python3.12/site-packages/click/_compat.py
new file mode 100644
index 00000000..9153d150
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/_compat.py
@@ -0,0 +1,623 @@
+import codecs
+import io
+import os
+import re
+import sys
+import typing as t
+from weakref import WeakKeyDictionary
+
+CYGWIN = sys.platform.startswith("cygwin")
+WIN = sys.platform.startswith("win")
+auto_wrap_for_ansi: t.Optional[t.Callable[[t.TextIO], t.TextIO]] = None
+_ansi_re = re.compile(r"\033\[[;?0-9]*[a-zA-Z]")
+
+
+def _make_text_stream(
+    stream: t.BinaryIO,
+    encoding: t.Optional[str],
+    errors: t.Optional[str],
+    force_readable: bool = False,
+    force_writable: bool = False,
+) -> t.TextIO:
+    if encoding is None:
+        encoding = get_best_encoding(stream)
+    if errors is None:
+        errors = "replace"
+    return _NonClosingTextIOWrapper(
+        stream,
+        encoding,
+        errors,
+        line_buffering=True,
+        force_readable=force_readable,
+        force_writable=force_writable,
+    )
+
+
+def is_ascii_encoding(encoding: str) -> bool:
+    """Checks if a given encoding is ascii."""
+    try:
+        return codecs.lookup(encoding).name == "ascii"
+    except LookupError:
+        return False
+
+
+def get_best_encoding(stream: t.IO[t.Any]) -> str:
+    """Returns the default stream encoding if not found."""
+    rv = getattr(stream, "encoding", None) or sys.getdefaultencoding()
+    if is_ascii_encoding(rv):
+        return "utf-8"
+    return rv
+
+
+class _NonClosingTextIOWrapper(io.TextIOWrapper):
+    def __init__(
+        self,
+        stream: t.BinaryIO,
+        encoding: t.Optional[str],
+        errors: t.Optional[str],
+        force_readable: bool = False,
+        force_writable: bool = False,
+        **extra: t.Any,
+    ) -> None:
+        self._stream = stream = t.cast(
+            t.BinaryIO, _FixupStream(stream, force_readable, force_writable)
+        )
+        super().__init__(stream, encoding, errors, **extra)
+
+    def __del__(self) -> None:
+        try:
+            self.detach()
+        except Exception:
+            pass
+
+    def isatty(self) -> bool:
+        # https://bitbucket.org/pypy/pypy/issue/1803
+        return self._stream.isatty()
+
+
+class _FixupStream:
+    """The new io interface needs more from streams than streams
+    traditionally implement.  As such, this fix-up code is necessary in
+    some circumstances.
+
+    The forcing of readable and writable flags are there because some tools
+    put badly patched objects on sys (one such offender are certain version
+    of jupyter notebook).
+    """
+
+    def __init__(
+        self,
+        stream: t.BinaryIO,
+        force_readable: bool = False,
+        force_writable: bool = False,
+    ):
+        self._stream = stream
+        self._force_readable = force_readable
+        self._force_writable = force_writable
+
+    def __getattr__(self, name: str) -> t.Any:
+        return getattr(self._stream, name)
+
+    def read1(self, size: int) -> bytes:
+        f = getattr(self._stream, "read1", None)
+
+        if f is not None:
+            return t.cast(bytes, f(size))
+
+        return self._stream.read(size)
+
+    def readable(self) -> bool:
+        if self._force_readable:
+            return True
+        x = getattr(self._stream, "readable", None)
+        if x is not None:
+            return t.cast(bool, x())
+        try:
+            self._stream.read(0)
+        except Exception:
+            return False
+        return True
+
+    def writable(self) -> bool:
+        if self._force_writable:
+            return True
+        x = getattr(self._stream, "writable", None)
+        if x is not None:
+            return t.cast(bool, x())
+        try:
+            self._stream.write("")  # type: ignore
+        except Exception:
+            try:
+                self._stream.write(b"")
+            except Exception:
+                return False
+        return True
+
+    def seekable(self) -> bool:
+        x = getattr(self._stream, "seekable", None)
+        if x is not None:
+            return t.cast(bool, x())
+        try:
+            self._stream.seek(self._stream.tell())
+        except Exception:
+            return False
+        return True
+
+
+def _is_binary_reader(stream: t.IO[t.Any], default: bool = False) -> bool:
+    try:
+        return isinstance(stream.read(0), bytes)
+    except Exception:
+        return default
+        # This happens in some cases where the stream was already
+        # closed.  In this case, we assume the default.
+
+
+def _is_binary_writer(stream: t.IO[t.Any], default: bool = False) -> bool:
+    try:
+        stream.write(b"")
+    except Exception:
+        try:
+            stream.write("")
+            return False
+        except Exception:
+            pass
+        return default
+    return True
+
+
+def _find_binary_reader(stream: t.IO[t.Any]) -> t.Optional[t.BinaryIO]:
+    # We need to figure out if the given stream is already binary.
+    # This can happen because the official docs recommend detaching
+    # the streams to get binary streams.  Some code might do this, so
+    # we need to deal with this case explicitly.
+    if _is_binary_reader(stream, False):
+        return t.cast(t.BinaryIO, stream)
+
+    buf = getattr(stream, "buffer", None)
+
+    # Same situation here; this time we assume that the buffer is
+    # actually binary in case it's closed.
+    if buf is not None and _is_binary_reader(buf, True):
+        return t.cast(t.BinaryIO, buf)
+
+    return None
+
+
+def _find_binary_writer(stream: t.IO[t.Any]) -> t.Optional[t.BinaryIO]:
+    # We need to figure out if the given stream is already binary.
+    # This can happen because the official docs recommend detaching
+    # the streams to get binary streams.  Some code might do this, so
+    # we need to deal with this case explicitly.
+    if _is_binary_writer(stream, False):
+        return t.cast(t.BinaryIO, stream)
+
+    buf = getattr(stream, "buffer", None)
+
+    # Same situation here; this time we assume that the buffer is
+    # actually binary in case it's closed.
+    if buf is not None and _is_binary_writer(buf, True):
+        return t.cast(t.BinaryIO, buf)
+
+    return None
+
+
+def _stream_is_misconfigured(stream: t.TextIO) -> bool:
+    """A stream is misconfigured if its encoding is ASCII."""
+    # If the stream does not have an encoding set, we assume it's set
+    # to ASCII.  This appears to happen in certain unittest
+    # environments.  It's not quite clear what the correct behavior is
+    # but this at least will force Click to recover somehow.
+    return is_ascii_encoding(getattr(stream, "encoding", None) or "ascii")
+
+
+def _is_compat_stream_attr(stream: t.TextIO, attr: str, value: t.Optional[str]) -> bool:
+    """A stream attribute is compatible if it is equal to the
+    desired value or the desired value is unset and the attribute
+    has a value.
+    """
+    stream_value = getattr(stream, attr, None)
+    return stream_value == value or (value is None and stream_value is not None)
+
+
+def _is_compatible_text_stream(
+    stream: t.TextIO, encoding: t.Optional[str], errors: t.Optional[str]
+) -> bool:
+    """Check if a stream's encoding and errors attributes are
+    compatible with the desired values.
+    """
+    return _is_compat_stream_attr(
+        stream, "encoding", encoding
+    ) and _is_compat_stream_attr(stream, "errors", errors)
+
+
+def _force_correct_text_stream(
+    text_stream: t.IO[t.Any],
+    encoding: t.Optional[str],
+    errors: t.Optional[str],
+    is_binary: t.Callable[[t.IO[t.Any], bool], bool],
+    find_binary: t.Callable[[t.IO[t.Any]], t.Optional[t.BinaryIO]],
+    force_readable: bool = False,
+    force_writable: bool = False,
+) -> t.TextIO:
+    if is_binary(text_stream, False):
+        binary_reader = t.cast(t.BinaryIO, text_stream)
+    else:
+        text_stream = t.cast(t.TextIO, text_stream)
+        # If the stream looks compatible, and won't default to a
+        # misconfigured ascii encoding, return it as-is.
+        if _is_compatible_text_stream(text_stream, encoding, errors) and not (
+            encoding is None and _stream_is_misconfigured(text_stream)
+        ):
+            return text_stream
+
+        # Otherwise, get the underlying binary reader.
+        possible_binary_reader = find_binary(text_stream)
+
+        # If that's not possible, silently use the original reader
+        # and get mojibake instead of exceptions.
+        if possible_binary_reader is None:
+            return text_stream
+
+        binary_reader = possible_binary_reader
+
+    # Default errors to replace instead of strict in order to get
+    # something that works.
+    if errors is None:
+        errors = "replace"
+
+    # Wrap the binary stream in a text stream with the correct
+    # encoding parameters.
+    return _make_text_stream(
+        binary_reader,
+        encoding,
+        errors,
+        force_readable=force_readable,
+        force_writable=force_writable,
+    )
+
+
+def _force_correct_text_reader(
+    text_reader: t.IO[t.Any],
+    encoding: t.Optional[str],
+    errors: t.Optional[str],
+    force_readable: bool = False,
+) -> t.TextIO:
+    return _force_correct_text_stream(
+        text_reader,
+        encoding,
+        errors,
+        _is_binary_reader,
+        _find_binary_reader,
+        force_readable=force_readable,
+    )
+
+
+def _force_correct_text_writer(
+    text_writer: t.IO[t.Any],
+    encoding: t.Optional[str],
+    errors: t.Optional[str],
+    force_writable: bool = False,
+) -> t.TextIO:
+    return _force_correct_text_stream(
+        text_writer,
+        encoding,
+        errors,
+        _is_binary_writer,
+        _find_binary_writer,
+        force_writable=force_writable,
+    )
+
+
+def get_binary_stdin() -> t.BinaryIO:
+    reader = _find_binary_reader(sys.stdin)
+    if reader is None:
+        raise RuntimeError("Was not able to determine binary stream for sys.stdin.")
+    return reader
+
+
+def get_binary_stdout() -> t.BinaryIO:
+    writer = _find_binary_writer(sys.stdout)
+    if writer is None:
+        raise RuntimeError("Was not able to determine binary stream for sys.stdout.")
+    return writer
+
+
+def get_binary_stderr() -> t.BinaryIO:
+    writer = _find_binary_writer(sys.stderr)
+    if writer is None:
+        raise RuntimeError("Was not able to determine binary stream for sys.stderr.")
+    return writer
+
+
+def get_text_stdin(
+    encoding: t.Optional[str] = None, errors: t.Optional[str] = None
+) -> t.TextIO:
+    rv = _get_windows_console_stream(sys.stdin, encoding, errors)
+    if rv is not None:
+        return rv
+    return _force_correct_text_reader(sys.stdin, encoding, errors, force_readable=True)
+
+
+def get_text_stdout(
+    encoding: t.Optional[str] = None, errors: t.Optional[str] = None
+) -> t.TextIO:
+    rv = _get_windows_console_stream(sys.stdout, encoding, errors)
+    if rv is not None:
+        return rv
+    return _force_correct_text_writer(sys.stdout, encoding, errors, force_writable=True)
+
+
+def get_text_stderr(
+    encoding: t.Optional[str] = None, errors: t.Optional[str] = None
+) -> t.TextIO:
+    rv = _get_windows_console_stream(sys.stderr, encoding, errors)
+    if rv is not None:
+        return rv
+    return _force_correct_text_writer(sys.stderr, encoding, errors, force_writable=True)
+
+
+def _wrap_io_open(
+    file: t.Union[str, "os.PathLike[str]", int],
+    mode: str,
+    encoding: t.Optional[str],
+    errors: t.Optional[str],
+) -> t.IO[t.Any]:
+    """Handles not passing ``encoding`` and ``errors`` in binary mode."""
+    if "b" in mode:
+        return open(file, mode)
+
+    return open(file, mode, encoding=encoding, errors=errors)
+
+
+def open_stream(
+    filename: "t.Union[str, os.PathLike[str]]",
+    mode: str = "r",
+    encoding: t.Optional[str] = None,
+    errors: t.Optional[str] = "strict",
+    atomic: bool = False,
+) -> t.Tuple[t.IO[t.Any], bool]:
+    binary = "b" in mode
+    filename = os.fspath(filename)
+
+    # Standard streams first. These are simple because they ignore the
+    # atomic flag. Use fsdecode to handle Path("-").
+    if os.fsdecode(filename) == "-":
+        if any(m in mode for m in ["w", "a", "x"]):
+            if binary:
+                return get_binary_stdout(), False
+            return get_text_stdout(encoding=encoding, errors=errors), False
+        if binary:
+            return get_binary_stdin(), False
+        return get_text_stdin(encoding=encoding, errors=errors), False
+
+    # Non-atomic writes directly go out through the regular open functions.
+    if not atomic:
+        return _wrap_io_open(filename, mode, encoding, errors), True
+
+    # Some usability stuff for atomic writes
+    if "a" in mode:
+        raise ValueError(
+            "Appending to an existing file is not supported, because that"
+            " would involve an expensive `copy`-operation to a temporary"
+            " file. Open the file in normal `w`-mode and copy explicitly"
+            " if that's what you're after."
+        )
+    if "x" in mode:
+        raise ValueError("Use the `overwrite`-parameter instead.")
+    if "w" not in mode:
+        raise ValueError("Atomic writes only make sense with `w`-mode.")
+
+    # Atomic writes are more complicated.  They work by opening a file
+    # as a proxy in the same folder and then using the fdopen
+    # functionality to wrap it in a Python file.  Then we wrap it in an
+    # atomic file that moves the file over on close.
+    import errno
+    import random
+
+    try:
+        perm: t.Optional[int] = os.stat(filename).st_mode
+    except OSError:
+        perm = None
+
+    flags = os.O_RDWR | os.O_CREAT | os.O_EXCL
+
+    if binary:
+        flags |= getattr(os, "O_BINARY", 0)
+
+    while True:
+        tmp_filename = os.path.join(
+            os.path.dirname(filename),
+            f".__atomic-write{random.randrange(1 << 32):08x}",
+        )
+        try:
+            fd = os.open(tmp_filename, flags, 0o666 if perm is None else perm)
+            break
+        except OSError as e:
+            if e.errno == errno.EEXIST or (
+                os.name == "nt"
+                and e.errno == errno.EACCES
+                and os.path.isdir(e.filename)
+                and os.access(e.filename, os.W_OK)
+            ):
+                continue
+            raise
+
+    if perm is not None:
+        os.chmod(tmp_filename, perm)  # in case perm includes bits in umask
+
+    f = _wrap_io_open(fd, mode, encoding, errors)
+    af = _AtomicFile(f, tmp_filename, os.path.realpath(filename))
+    return t.cast(t.IO[t.Any], af), True
+
+
+class _AtomicFile:
+    def __init__(self, f: t.IO[t.Any], tmp_filename: str, real_filename: str) -> None:
+        self._f = f
+        self._tmp_filename = tmp_filename
+        self._real_filename = real_filename
+        self.closed = False
+
+    @property
+    def name(self) -> str:
+        return self._real_filename
+
+    def close(self, delete: bool = False) -> None:
+        if self.closed:
+            return
+        self._f.close()
+        os.replace(self._tmp_filename, self._real_filename)
+        self.closed = True
+
+    def __getattr__(self, name: str) -> t.Any:
+        return getattr(self._f, name)
+
+    def __enter__(self) -> "_AtomicFile":
+        return self
+
+    def __exit__(self, exc_type: t.Optional[t.Type[BaseException]], *_: t.Any) -> None:
+        self.close(delete=exc_type is not None)
+
+    def __repr__(self) -> str:
+        return repr(self._f)
+
+
+def strip_ansi(value: str) -> str:
+    return _ansi_re.sub("", value)
+
+
+def _is_jupyter_kernel_output(stream: t.IO[t.Any]) -> bool:
+    while isinstance(stream, (_FixupStream, _NonClosingTextIOWrapper)):
+        stream = stream._stream
+
+    return stream.__class__.__module__.startswith("ipykernel.")
+
+
+def should_strip_ansi(
+    stream: t.Optional[t.IO[t.Any]] = None, color: t.Optional[bool] = None
+) -> bool:
+    if color is None:
+        if stream is None:
+            stream = sys.stdin
+        return not isatty(stream) and not _is_jupyter_kernel_output(stream)
+    return not color
+
+
+# On Windows, wrap the output streams with colorama to support ANSI
+# color codes.
+# NOTE: double check is needed so mypy does not analyze this on Linux
+if sys.platform.startswith("win") and WIN:
+    from ._winconsole import _get_windows_console_stream
+
+    def _get_argv_encoding() -> str:
+        import locale
+
+        return locale.getpreferredencoding()
+
+    _ansi_stream_wrappers: t.MutableMapping[t.TextIO, t.TextIO] = WeakKeyDictionary()
+
+    def auto_wrap_for_ansi(
+        stream: t.TextIO, color: t.Optional[bool] = None
+    ) -> t.TextIO:
+        """Support ANSI color and style codes on Windows by wrapping a
+        stream with colorama.
+        """
+        try:
+            cached = _ansi_stream_wrappers.get(stream)
+        except Exception:
+            cached = None
+
+        if cached is not None:
+            return cached
+
+        import colorama
+
+        strip = should_strip_ansi(stream, color)
+        ansi_wrapper = colorama.AnsiToWin32(stream, strip=strip)
+        rv = t.cast(t.TextIO, ansi_wrapper.stream)
+        _write = rv.write
+
+        def _safe_write(s):
+            try:
+                return _write(s)
+            except BaseException:
+                ansi_wrapper.reset_all()
+                raise
+
+        rv.write = _safe_write
+
+        try:
+            _ansi_stream_wrappers[stream] = rv
+        except Exception:
+            pass
+
+        return rv
+
+else:
+
+    def _get_argv_encoding() -> str:
+        return getattr(sys.stdin, "encoding", None) or sys.getfilesystemencoding()
+
+    def _get_windows_console_stream(
+        f: t.TextIO, encoding: t.Optional[str], errors: t.Optional[str]
+    ) -> t.Optional[t.TextIO]:
+        return None
+
+
+def term_len(x: str) -> int:
+    return len(strip_ansi(x))
+
+
+def isatty(stream: t.IO[t.Any]) -> bool:
+    try:
+        return stream.isatty()
+    except Exception:
+        return False
+
+
+def _make_cached_stream_func(
+    src_func: t.Callable[[], t.Optional[t.TextIO]],
+    wrapper_func: t.Callable[[], t.TextIO],
+) -> t.Callable[[], t.Optional[t.TextIO]]:
+    cache: t.MutableMapping[t.TextIO, t.TextIO] = WeakKeyDictionary()
+
+    def func() -> t.Optional[t.TextIO]:
+        stream = src_func()
+
+        if stream is None:
+            return None
+
+        try:
+            rv = cache.get(stream)
+        except Exception:
+            rv = None
+        if rv is not None:
+            return rv
+        rv = wrapper_func()
+        try:
+            cache[stream] = rv
+        except Exception:
+            pass
+        return rv
+
+    return func
+
+
+_default_text_stdin = _make_cached_stream_func(lambda: sys.stdin, get_text_stdin)
+_default_text_stdout = _make_cached_stream_func(lambda: sys.stdout, get_text_stdout)
+_default_text_stderr = _make_cached_stream_func(lambda: sys.stderr, get_text_stderr)
+
+
+binary_streams: t.Mapping[str, t.Callable[[], t.BinaryIO]] = {
+    "stdin": get_binary_stdin,
+    "stdout": get_binary_stdout,
+    "stderr": get_binary_stderr,
+}
+
+text_streams: t.Mapping[
+    str, t.Callable[[t.Optional[str], t.Optional[str]], t.TextIO]
+] = {
+    "stdin": get_text_stdin,
+    "stdout": get_text_stdout,
+    "stderr": get_text_stderr,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/click/_termui_impl.py b/backend/venv/lib/python3.12/site-packages/click/_termui_impl.py
new file mode 100644
index 00000000..ad9f8f6c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/_termui_impl.py
@@ -0,0 +1,788 @@
+"""
+This module contains implementations for the termui module. To keep the
+import time of Click down, some infrequently used functionality is
+placed in this module and only imported as needed.
+"""
+
+import contextlib
+import math
+import os
+import sys
+import time
+import typing as t
+from gettext import gettext as _
+from io import StringIO
+from shutil import which
+from types import TracebackType
+
+from ._compat import _default_text_stdout
+from ._compat import CYGWIN
+from ._compat import get_best_encoding
+from ._compat import isatty
+from ._compat import open_stream
+from ._compat import strip_ansi
+from ._compat import term_len
+from ._compat import WIN
+from .exceptions import ClickException
+from .utils import echo
+
+V = t.TypeVar("V")
+
+if os.name == "nt":
+    BEFORE_BAR = "\r"
+    AFTER_BAR = "\n"
+else:
+    BEFORE_BAR = "\r\033[?25l"
+    AFTER_BAR = "\033[?25h\n"
+
+
+class ProgressBar(t.Generic[V]):
+    def __init__(
+        self,
+        iterable: t.Optional[t.Iterable[V]],
+        length: t.Optional[int] = None,
+        fill_char: str = "#",
+        empty_char: str = " ",
+        bar_template: str = "%(bar)s",
+        info_sep: str = "  ",
+        show_eta: bool = True,
+        show_percent: t.Optional[bool] = None,
+        show_pos: bool = False,
+        item_show_func: t.Optional[t.Callable[[t.Optional[V]], t.Optional[str]]] = None,
+        label: t.Optional[str] = None,
+        file: t.Optional[t.TextIO] = None,
+        color: t.Optional[bool] = None,
+        update_min_steps: int = 1,
+        width: int = 30,
+    ) -> None:
+        self.fill_char = fill_char
+        self.empty_char = empty_char
+        self.bar_template = bar_template
+        self.info_sep = info_sep
+        self.show_eta = show_eta
+        self.show_percent = show_percent
+        self.show_pos = show_pos
+        self.item_show_func = item_show_func
+        self.label: str = label or ""
+
+        if file is None:
+            file = _default_text_stdout()
+
+            # There are no standard streams attached to write to. For example,
+            # pythonw on Windows.
+            if file is None:
+                file = StringIO()
+
+        self.file = file
+        self.color = color
+        self.update_min_steps = update_min_steps
+        self._completed_intervals = 0
+        self.width: int = width
+        self.autowidth: bool = width == 0
+
+        if length is None:
+            from operator import length_hint
+
+            length = length_hint(iterable, -1)
+
+            if length == -1:
+                length = None
+        if iterable is None:
+            if length is None:
+                raise TypeError("iterable or length is required")
+            iterable = t.cast(t.Iterable[V], range(length))
+        self.iter: t.Iterable[V] = iter(iterable)
+        self.length = length
+        self.pos = 0
+        self.avg: t.List[float] = []
+        self.last_eta: float
+        self.start: float
+        self.start = self.last_eta = time.time()
+        self.eta_known: bool = False
+        self.finished: bool = False
+        self.max_width: t.Optional[int] = None
+        self.entered: bool = False
+        self.current_item: t.Optional[V] = None
+        self.is_hidden: bool = not isatty(self.file)
+        self._last_line: t.Optional[str] = None
+
+    def __enter__(self) -> "ProgressBar[V]":
+        self.entered = True
+        self.render_progress()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: t.Optional[t.Type[BaseException]],
+        exc_value: t.Optional[BaseException],
+        tb: t.Optional[TracebackType],
+    ) -> None:
+        self.render_finish()
+
+    def __iter__(self) -> t.Iterator[V]:
+        if not self.entered:
+            raise RuntimeError("You need to use progress bars in a with block.")
+        self.render_progress()
+        return self.generator()
+
+    def __next__(self) -> V:
+        # Iteration is defined in terms of a generator function,
+        # returned by iter(self); use that to define next(). This works
+        # because `self.iter` is an iterable consumed by that generator,
+        # so it is re-entry safe. Calling `next(self.generator())`
+        # twice works and does "what you want".
+        return next(iter(self))
+
+    def render_finish(self) -> None:
+        if self.is_hidden:
+            return
+        self.file.write(AFTER_BAR)
+        self.file.flush()
+
+    @property
+    def pct(self) -> float:
+        if self.finished:
+            return 1.0
+        return min(self.pos / (float(self.length or 1) or 1), 1.0)
+
+    @property
+    def time_per_iteration(self) -> float:
+        if not self.avg:
+            return 0.0
+        return sum(self.avg) / float(len(self.avg))
+
+    @property
+    def eta(self) -> float:
+        if self.length is not None and not self.finished:
+            return self.time_per_iteration * (self.length - self.pos)
+        return 0.0
+
+    def format_eta(self) -> str:
+        if self.eta_known:
+            t = int(self.eta)
+            seconds = t % 60
+            t //= 60
+            minutes = t % 60
+            t //= 60
+            hours = t % 24
+            t //= 24
+            if t > 0:
+                return f"{t}d {hours:02}:{minutes:02}:{seconds:02}"
+            else:
+                return f"{hours:02}:{minutes:02}:{seconds:02}"
+        return ""
+
+    def format_pos(self) -> str:
+        pos = str(self.pos)
+        if self.length is not None:
+            pos += f"/{self.length}"
+        return pos
+
+    def format_pct(self) -> str:
+        return f"{int(self.pct * 100): 4}%"[1:]
+
+    def format_bar(self) -> str:
+        if self.length is not None:
+            bar_length = int(self.pct * self.width)
+            bar = self.fill_char * bar_length
+            bar += self.empty_char * (self.width - bar_length)
+        elif self.finished:
+            bar = self.fill_char * self.width
+        else:
+            chars = list(self.empty_char * (self.width or 1))
+            if self.time_per_iteration != 0:
+                chars[
+                    int(
+                        (math.cos(self.pos * self.time_per_iteration) / 2.0 + 0.5)
+                        * self.width
+                    )
+                ] = self.fill_char
+            bar = "".join(chars)
+        return bar
+
+    def format_progress_line(self) -> str:
+        show_percent = self.show_percent
+
+        info_bits = []
+        if self.length is not None and show_percent is None:
+            show_percent = not self.show_pos
+
+        if self.show_pos:
+            info_bits.append(self.format_pos())
+        if show_percent:
+            info_bits.append(self.format_pct())
+        if self.show_eta and self.eta_known and not self.finished:
+            info_bits.append(self.format_eta())
+        if self.item_show_func is not None:
+            item_info = self.item_show_func(self.current_item)
+            if item_info is not None:
+                info_bits.append(item_info)
+
+        return (
+            self.bar_template
+            % {
+                "label": self.label,
+                "bar": self.format_bar(),
+                "info": self.info_sep.join(info_bits),
+            }
+        ).rstrip()
+
+    def render_progress(self) -> None:
+        import shutil
+
+        if self.is_hidden:
+            # Only output the label as it changes if the output is not a
+            # TTY. Use file=stderr if you expect to be piping stdout.
+            if self._last_line != self.label:
+                self._last_line = self.label
+                echo(self.label, file=self.file, color=self.color)
+
+            return
+
+        buf = []
+        # Update width in case the terminal has been resized
+        if self.autowidth:
+            old_width = self.width
+            self.width = 0
+            clutter_length = term_len(self.format_progress_line())
+            new_width = max(0, shutil.get_terminal_size().columns - clutter_length)
+            if new_width < old_width:
+                buf.append(BEFORE_BAR)
+                buf.append(" " * self.max_width)  # type: ignore
+                self.max_width = new_width
+            self.width = new_width
+
+        clear_width = self.width
+        if self.max_width is not None:
+            clear_width = self.max_width
+
+        buf.append(BEFORE_BAR)
+        line = self.format_progress_line()
+        line_len = term_len(line)
+        if self.max_width is None or self.max_width < line_len:
+            self.max_width = line_len
+
+        buf.append(line)
+        buf.append(" " * (clear_width - line_len))
+        line = "".join(buf)
+        # Render the line only if it changed.
+
+        if line != self._last_line:
+            self._last_line = line
+            echo(line, file=self.file, color=self.color, nl=False)
+            self.file.flush()
+
+    def make_step(self, n_steps: int) -> None:
+        self.pos += n_steps
+        if self.length is not None and self.pos >= self.length:
+            self.finished = True
+
+        if (time.time() - self.last_eta) < 1.0:
+            return
+
+        self.last_eta = time.time()
+
+        # self.avg is a rolling list of length <= 7 of steps where steps are
+        # defined as time elapsed divided by the total progress through
+        # self.length.
+        if self.pos:
+            step = (time.time() - self.start) / self.pos
+        else:
+            step = time.time() - self.start
+
+        self.avg = self.avg[-6:] + [step]
+
+        self.eta_known = self.length is not None
+
+    def update(self, n_steps: int, current_item: t.Optional[V] = None) -> None:
+        """Update the progress bar by advancing a specified number of
+        steps, and optionally set the ``current_item`` for this new
+        position.
+
+        :param n_steps: Number of steps to advance.
+        :param current_item: Optional item to set as ``current_item``
+            for the updated position.
+
+        .. versionchanged:: 8.0
+            Added the ``current_item`` optional parameter.
+
+        .. versionchanged:: 8.0
+            Only render when the number of steps meets the
+            ``update_min_steps`` threshold.
+        """
+        if current_item is not None:
+            self.current_item = current_item
+
+        self._completed_intervals += n_steps
+
+        if self._completed_intervals >= self.update_min_steps:
+            self.make_step(self._completed_intervals)
+            self.render_progress()
+            self._completed_intervals = 0
+
+    def finish(self) -> None:
+        self.eta_known = False
+        self.current_item = None
+        self.finished = True
+
+    def generator(self) -> t.Iterator[V]:
+        """Return a generator which yields the items added to the bar
+        during construction, and updates the progress bar *after* the
+        yielded block returns.
+        """
+        # WARNING: the iterator interface for `ProgressBar` relies on
+        # this and only works because this is a simple generator which
+        # doesn't create or manage additional state. If this function
+        # changes, the impact should be evaluated both against
+        # `iter(bar)` and `next(bar)`. `next()` in particular may call
+        # `self.generator()` repeatedly, and this must remain safe in
+        # order for that interface to work.
+        if not self.entered:
+            raise RuntimeError("You need to use progress bars in a with block.")
+
+        if self.is_hidden:
+            yield from self.iter
+        else:
+            for rv in self.iter:
+                self.current_item = rv
+
+                # This allows show_item_func to be updated before the
+                # item is processed. Only trigger at the beginning of
+                # the update interval.
+                if self._completed_intervals == 0:
+                    self.render_progress()
+
+                yield rv
+                self.update(1)
+
+            self.finish()
+            self.render_progress()
+
+
+def pager(generator: t.Iterable[str], color: t.Optional[bool] = None) -> None:
+    """Decide what method to use for paging through text."""
+    stdout = _default_text_stdout()
+
+    # There are no standard streams attached to write to. For example,
+    # pythonw on Windows.
+    if stdout is None:
+        stdout = StringIO()
+
+    if not isatty(sys.stdin) or not isatty(stdout):
+        return _nullpager(stdout, generator, color)
+    pager_cmd = (os.environ.get("PAGER", None) or "").strip()
+    if pager_cmd:
+        if WIN:
+            if _tempfilepager(generator, pager_cmd, color):
+                return
+        elif _pipepager(generator, pager_cmd, color):
+            return
+    if os.environ.get("TERM") in ("dumb", "emacs"):
+        return _nullpager(stdout, generator, color)
+    if (WIN or sys.platform.startswith("os2")) and _tempfilepager(
+        generator, "more", color
+    ):
+        return
+    if _pipepager(generator, "less", color):
+        return
+
+    import tempfile
+
+    fd, filename = tempfile.mkstemp()
+    os.close(fd)
+    try:
+        if _pipepager(generator, "more", color):
+            return
+        return _nullpager(stdout, generator, color)
+    finally:
+        os.unlink(filename)
+
+
+def _pipepager(generator: t.Iterable[str], cmd: str, color: t.Optional[bool]) -> bool:
+    """Page through text by feeding it to another program.  Invoking a
+    pager through this might support colors.
+
+    Returns True if the command was found, False otherwise and thus another
+    pager should be attempted.
+    """
+    cmd_absolute = which(cmd)
+    if cmd_absolute is None:
+        return False
+
+    import subprocess
+
+    env = dict(os.environ)
+
+    # If we're piping to less we might support colors under the
+    # condition that
+    cmd_detail = cmd.rsplit("/", 1)[-1].split()
+    if color is None and cmd_detail[0] == "less":
+        less_flags = f"{os.environ.get('LESS', '')}{' '.join(cmd_detail[1:])}"
+        if not less_flags:
+            env["LESS"] = "-R"
+            color = True
+        elif "r" in less_flags or "R" in less_flags:
+            color = True
+
+    c = subprocess.Popen(
+        [cmd_absolute],
+        shell=True,
+        stdin=subprocess.PIPE,
+        env=env,
+        errors="replace",
+        text=True,
+    )
+    assert c.stdin is not None
+    try:
+        for text in generator:
+            if not color:
+                text = strip_ansi(text)
+
+            c.stdin.write(text)
+    except (OSError, KeyboardInterrupt):
+        pass
+    else:
+        c.stdin.close()
+
+    # Less doesn't respect ^C, but catches it for its own UI purposes (aborting
+    # search or other commands inside less).
+    #
+    # That means when the user hits ^C, the parent process (click) terminates,
+    # but less is still alive, paging the output and messing up the terminal.
+    #
+    # If the user wants to make the pager exit on ^C, they should set
+    # `LESS='-K'`. It's not our decision to make.
+    while True:
+        try:
+            c.wait()
+        except KeyboardInterrupt:
+            pass
+        else:
+            break
+
+    return True
+
+
+def _tempfilepager(
+    generator: t.Iterable[str],
+    cmd: str,
+    color: t.Optional[bool],
+) -> bool:
+    """Page through text by invoking a program on a temporary file.
+
+    Returns True if the command was found, False otherwise and thus another
+    pager should be attempted.
+    """
+    # Which is necessary for Windows, it is also recommended in the Popen docs.
+    cmd_absolute = which(cmd)
+    if cmd_absolute is None:
+        return False
+
+    import subprocess
+    import tempfile
+
+    fd, filename = tempfile.mkstemp()
+    # TODO: This never terminates if the passed generator never terminates.
+    text = "".join(generator)
+    if not color:
+        text = strip_ansi(text)
+    encoding = get_best_encoding(sys.stdout)
+    with open_stream(filename, "wb")[0] as f:
+        f.write(text.encode(encoding))
+    try:
+        subprocess.call([cmd_absolute, filename])
+    except OSError:
+        # Command not found
+        pass
+    finally:
+        os.close(fd)
+        os.unlink(filename)
+
+    return True
+
+
+def _nullpager(
+    stream: t.TextIO, generator: t.Iterable[str], color: t.Optional[bool]
+) -> None:
+    """Simply print unformatted text.  This is the ultimate fallback."""
+    for text in generator:
+        if not color:
+            text = strip_ansi(text)
+        stream.write(text)
+
+
+class Editor:
+    def __init__(
+        self,
+        editor: t.Optional[str] = None,
+        env: t.Optional[t.Mapping[str, str]] = None,
+        require_save: bool = True,
+        extension: str = ".txt",
+    ) -> None:
+        self.editor = editor
+        self.env = env
+        self.require_save = require_save
+        self.extension = extension
+
+    def get_editor(self) -> str:
+        if self.editor is not None:
+            return self.editor
+        for key in "VISUAL", "EDITOR":
+            rv = os.environ.get(key)
+            if rv:
+                return rv
+        if WIN:
+            return "notepad"
+        for editor in "sensible-editor", "vim", "nano":
+            if which(editor) is not None:
+                return editor
+        return "vi"
+
+    def edit_file(self, filename: str) -> None:
+        import subprocess
+
+        editor = self.get_editor()
+        environ: t.Optional[t.Dict[str, str]] = None
+
+        if self.env:
+            environ = os.environ.copy()
+            environ.update(self.env)
+
+        try:
+            c = subprocess.Popen(f'{editor} "{filename}"', env=environ, shell=True)
+            exit_code = c.wait()
+            if exit_code != 0:
+                raise ClickException(
+                    _("{editor}: Editing failed").format(editor=editor)
+                )
+        except OSError as e:
+            raise ClickException(
+                _("{editor}: Editing failed: {e}").format(editor=editor, e=e)
+            ) from e
+
+    def edit(self, text: t.Optional[t.AnyStr]) -> t.Optional[t.AnyStr]:
+        import tempfile
+
+        if not text:
+            data = b""
+        elif isinstance(text, (bytes, bytearray)):
+            data = text
+        else:
+            if text and not text.endswith("\n"):
+                text += "\n"
+
+            if WIN:
+                data = text.replace("\n", "\r\n").encode("utf-8-sig")
+            else:
+                data = text.encode("utf-8")
+
+        fd, name = tempfile.mkstemp(prefix="editor-", suffix=self.extension)
+        f: t.BinaryIO
+
+        try:
+            with os.fdopen(fd, "wb") as f:
+                f.write(data)
+
+            # If the filesystem resolution is 1 second, like Mac OS
+            # 10.12 Extended, or 2 seconds, like FAT32, and the editor
+            # closes very fast, require_save can fail. Set the modified
+            # time to be 2 seconds in the past to work around this.
+            os.utime(name, (os.path.getatime(name), os.path.getmtime(name) - 2))
+            # Depending on the resolution, the exact value might not be
+            # recorded, so get the new recorded value.
+            timestamp = os.path.getmtime(name)
+
+            self.edit_file(name)
+
+            if self.require_save and os.path.getmtime(name) == timestamp:
+                return None
+
+            with open(name, "rb") as f:
+                rv = f.read()
+
+            if isinstance(text, (bytes, bytearray)):
+                return rv
+
+            return rv.decode("utf-8-sig").replace("\r\n", "\n")  # type: ignore
+        finally:
+            os.unlink(name)
+
+
+def open_url(url: str, wait: bool = False, locate: bool = False) -> int:
+    import subprocess
+
+    def _unquote_file(url: str) -> str:
+        from urllib.parse import unquote
+
+        if url.startswith("file://"):
+            url = unquote(url[7:])
+
+        return url
+
+    if sys.platform == "darwin":
+        args = ["open"]
+        if wait:
+            args.append("-W")
+        if locate:
+            args.append("-R")
+        args.append(_unquote_file(url))
+        null = open("/dev/null", "w")
+        try:
+            return subprocess.Popen(args, stderr=null).wait()
+        finally:
+            null.close()
+    elif WIN:
+        if locate:
+            url = _unquote_file(url)
+            args = ["explorer", f"/select,{url}"]
+        else:
+            args = ["start"]
+            if wait:
+                args.append("/WAIT")
+            args.append("")
+            args.append(url)
+        try:
+            return subprocess.call(args)
+        except OSError:
+            # Command not found
+            return 127
+    elif CYGWIN:
+        if locate:
+            url = _unquote_file(url)
+            args = ["cygstart", os.path.dirname(url)]
+        else:
+            args = ["cygstart"]
+            if wait:
+                args.append("-w")
+            args.append(url)
+        try:
+            return subprocess.call(args)
+        except OSError:
+            # Command not found
+            return 127
+
+    try:
+        if locate:
+            url = os.path.dirname(_unquote_file(url)) or "."
+        else:
+            url = _unquote_file(url)
+        c = subprocess.Popen(["xdg-open", url])
+        if wait:
+            return c.wait()
+        return 0
+    except OSError:
+        if url.startswith(("http://", "https://")) and not locate and not wait:
+            import webbrowser
+
+            webbrowser.open(url)
+            return 0
+        return 1
+
+
+def _translate_ch_to_exc(ch: str) -> t.Optional[BaseException]:
+    if ch == "\x03":
+        raise KeyboardInterrupt()
+
+    if ch == "\x04" and not WIN:  # Unix-like, Ctrl+D
+        raise EOFError()
+
+    if ch == "\x1a" and WIN:  # Windows, Ctrl+Z
+        raise EOFError()
+
+    return None
+
+
+if WIN:
+    import msvcrt
+
+    @contextlib.contextmanager
+    def raw_terminal() -> t.Iterator[int]:
+        yield -1
+
+    def getchar(echo: bool) -> str:
+        # The function `getch` will return a bytes object corresponding to
+        # the pressed character. Since Windows 10 build 1803, it will also
+        # return \x00 when called a second time after pressing a regular key.
+        #
+        # `getwch` does not share this probably-bugged behavior. Moreover, it
+        # returns a Unicode object by default, which is what we want.
+        #
+        # Either of these functions will return \x00 or \xe0 to indicate
+        # a special key, and you need to call the same function again to get
+        # the "rest" of the code. The fun part is that \u00e0 is
+        # "latin small letter a with grave", so if you type that on a French
+        # keyboard, you _also_ get a \xe0.
+        # E.g., consider the Up arrow. This returns \xe0 and then \x48. The
+        # resulting Unicode string reads as "a with grave" + "capital H".
+        # This is indistinguishable from when the user actually types
+        # "a with grave" and then "capital H".
+        #
+        # When \xe0 is returned, we assume it's part of a special-key sequence
+        # and call `getwch` again, but that means that when the user types
+        # the \u00e0 character, `getchar` doesn't return until a second
+        # character is typed.
+        # The alternative is returning immediately, but that would mess up
+        # cross-platform handling of arrow keys and others that start with
+        # \xe0. Another option is using `getch`, but then we can't reliably
+        # read non-ASCII characters, because return values of `getch` are
+        # limited to the current 8-bit codepage.
+        #
+        # Anyway, Click doesn't claim to do this Right(tm), and using `getwch`
+        # is doing the right thing in more situations than with `getch`.
+        func: t.Callable[[], str]
+
+        if echo:
+            func = msvcrt.getwche  # type: ignore
+        else:
+            func = msvcrt.getwch  # type: ignore
+
+        rv = func()
+
+        if rv in ("\x00", "\xe0"):
+            # \x00 and \xe0 are control characters that indicate special key,
+            # see above.
+            rv += func()
+
+        _translate_ch_to_exc(rv)
+        return rv
+
+else:
+    import termios
+    import tty
+
+    @contextlib.contextmanager
+    def raw_terminal() -> t.Iterator[int]:
+        f: t.Optional[t.TextIO]
+        fd: int
+
+        if not isatty(sys.stdin):
+            f = open("/dev/tty")
+            fd = f.fileno()
+        else:
+            fd = sys.stdin.fileno()
+            f = None
+
+        try:
+            old_settings = termios.tcgetattr(fd)
+
+            try:
+                tty.setraw(fd)
+                yield fd
+            finally:
+                termios.tcsetattr(fd, termios.TCSADRAIN, old_settings)
+                sys.stdout.flush()
+
+                if f is not None:
+                    f.close()
+        except termios.error:
+            pass
+
+    def getchar(echo: bool) -> str:
+        with raw_terminal() as fd:
+            ch = os.read(fd, 32).decode(get_best_encoding(sys.stdin), "replace")
+
+            if echo and isatty(sys.stdout):
+                sys.stdout.write(ch)
+
+            _translate_ch_to_exc(ch)
+            return ch
diff --git a/backend/venv/lib/python3.12/site-packages/click/_textwrap.py b/backend/venv/lib/python3.12/site-packages/click/_textwrap.py
new file mode 100644
index 00000000..b47dcbd4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/_textwrap.py
@@ -0,0 +1,49 @@
+import textwrap
+import typing as t
+from contextlib import contextmanager
+
+
+class TextWrapper(textwrap.TextWrapper):
+    def _handle_long_word(
+        self,
+        reversed_chunks: t.List[str],
+        cur_line: t.List[str],
+        cur_len: int,
+        width: int,
+    ) -> None:
+        space_left = max(width - cur_len, 1)
+
+        if self.break_long_words:
+            last = reversed_chunks[-1]
+            cut = last[:space_left]
+            res = last[space_left:]
+            cur_line.append(cut)
+            reversed_chunks[-1] = res
+        elif not cur_line:
+            cur_line.append(reversed_chunks.pop())
+
+    @contextmanager
+    def extra_indent(self, indent: str) -> t.Iterator[None]:
+        old_initial_indent = self.initial_indent
+        old_subsequent_indent = self.subsequent_indent
+        self.initial_indent += indent
+        self.subsequent_indent += indent
+
+        try:
+            yield
+        finally:
+            self.initial_indent = old_initial_indent
+            self.subsequent_indent = old_subsequent_indent
+
+    def indent_only(self, text: str) -> str:
+        rv = []
+
+        for idx, line in enumerate(text.splitlines()):
+            indent = self.initial_indent
+
+            if idx > 0:
+                indent = self.subsequent_indent
+
+            rv.append(f"{indent}{line}")
+
+        return "\n".join(rv)
diff --git a/backend/venv/lib/python3.12/site-packages/click/_winconsole.py b/backend/venv/lib/python3.12/site-packages/click/_winconsole.py
new file mode 100644
index 00000000..6b20df31
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/_winconsole.py
@@ -0,0 +1,279 @@
+# This module is based on the excellent work by Adam Bartoš who
+# provided a lot of what went into the implementation here in
+# the discussion to issue1602 in the Python bug tracker.
+#
+# There are some general differences in regards to how this works
+# compared to the original patches as we do not need to patch
+# the entire interpreter but just work in our little world of
+# echo and prompt.
+import io
+import sys
+import time
+import typing as t
+from ctypes import byref
+from ctypes import c_char
+from ctypes import c_char_p
+from ctypes import c_int
+from ctypes import c_ssize_t
+from ctypes import c_ulong
+from ctypes import c_void_p
+from ctypes import POINTER
+from ctypes import py_object
+from ctypes import Structure
+from ctypes.wintypes import DWORD
+from ctypes.wintypes import HANDLE
+from ctypes.wintypes import LPCWSTR
+from ctypes.wintypes import LPWSTR
+
+from ._compat import _NonClosingTextIOWrapper
+
+assert sys.platform == "win32"
+import msvcrt  # noqa: E402
+from ctypes import windll  # noqa: E402
+from ctypes import WINFUNCTYPE  # noqa: E402
+
+c_ssize_p = POINTER(c_ssize_t)
+
+kernel32 = windll.kernel32
+GetStdHandle = kernel32.GetStdHandle
+ReadConsoleW = kernel32.ReadConsoleW
+WriteConsoleW = kernel32.WriteConsoleW
+GetConsoleMode = kernel32.GetConsoleMode
+GetLastError = kernel32.GetLastError
+GetCommandLineW = WINFUNCTYPE(LPWSTR)(("GetCommandLineW", windll.kernel32))
+CommandLineToArgvW = WINFUNCTYPE(POINTER(LPWSTR), LPCWSTR, POINTER(c_int))(
+    ("CommandLineToArgvW", windll.shell32)
+)
+LocalFree = WINFUNCTYPE(c_void_p, c_void_p)(("LocalFree", windll.kernel32))
+
+STDIN_HANDLE = GetStdHandle(-10)
+STDOUT_HANDLE = GetStdHandle(-11)
+STDERR_HANDLE = GetStdHandle(-12)
+
+PyBUF_SIMPLE = 0
+PyBUF_WRITABLE = 1
+
+ERROR_SUCCESS = 0
+ERROR_NOT_ENOUGH_MEMORY = 8
+ERROR_OPERATION_ABORTED = 995
+
+STDIN_FILENO = 0
+STDOUT_FILENO = 1
+STDERR_FILENO = 2
+
+EOF = b"\x1a"
+MAX_BYTES_WRITTEN = 32767
+
+try:
+    from ctypes import pythonapi
+except ImportError:
+    # On PyPy we cannot get buffers so our ability to operate here is
+    # severely limited.
+    get_buffer = None
+else:
+
+    class Py_buffer(Structure):
+        _fields_ = [
+            ("buf", c_void_p),
+            ("obj", py_object),
+            ("len", c_ssize_t),
+            ("itemsize", c_ssize_t),
+            ("readonly", c_int),
+            ("ndim", c_int),
+            ("format", c_char_p),
+            ("shape", c_ssize_p),
+            ("strides", c_ssize_p),
+            ("suboffsets", c_ssize_p),
+            ("internal", c_void_p),
+        ]
+
+    PyObject_GetBuffer = pythonapi.PyObject_GetBuffer
+    PyBuffer_Release = pythonapi.PyBuffer_Release
+
+    def get_buffer(obj, writable=False):
+        buf = Py_buffer()
+        flags = PyBUF_WRITABLE if writable else PyBUF_SIMPLE
+        PyObject_GetBuffer(py_object(obj), byref(buf), flags)
+
+        try:
+            buffer_type = c_char * buf.len
+            return buffer_type.from_address(buf.buf)
+        finally:
+            PyBuffer_Release(byref(buf))
+
+
+class _WindowsConsoleRawIOBase(io.RawIOBase):
+    def __init__(self, handle):
+        self.handle = handle
+
+    def isatty(self):
+        super().isatty()
+        return True
+
+
+class _WindowsConsoleReader(_WindowsConsoleRawIOBase):
+    def readable(self):
+        return True
+
+    def readinto(self, b):
+        bytes_to_be_read = len(b)
+        if not bytes_to_be_read:
+            return 0
+        elif bytes_to_be_read % 2:
+            raise ValueError(
+                "cannot read odd number of bytes from UTF-16-LE encoded console"
+            )
+
+        buffer = get_buffer(b, writable=True)
+        code_units_to_be_read = bytes_to_be_read // 2
+        code_units_read = c_ulong()
+
+        rv = ReadConsoleW(
+            HANDLE(self.handle),
+            buffer,
+            code_units_to_be_read,
+            byref(code_units_read),
+            None,
+        )
+        if GetLastError() == ERROR_OPERATION_ABORTED:
+            # wait for KeyboardInterrupt
+            time.sleep(0.1)
+        if not rv:
+            raise OSError(f"Windows error: {GetLastError()}")
+
+        if buffer[0] == EOF:
+            return 0
+        return 2 * code_units_read.value
+
+
+class _WindowsConsoleWriter(_WindowsConsoleRawIOBase):
+    def writable(self):
+        return True
+
+    @staticmethod
+    def _get_error_message(errno):
+        if errno == ERROR_SUCCESS:
+            return "ERROR_SUCCESS"
+        elif errno == ERROR_NOT_ENOUGH_MEMORY:
+            return "ERROR_NOT_ENOUGH_MEMORY"
+        return f"Windows error {errno}"
+
+    def write(self, b):
+        bytes_to_be_written = len(b)
+        buf = get_buffer(b)
+        code_units_to_be_written = min(bytes_to_be_written, MAX_BYTES_WRITTEN) // 2
+        code_units_written = c_ulong()
+
+        WriteConsoleW(
+            HANDLE(self.handle),
+            buf,
+            code_units_to_be_written,
+            byref(code_units_written),
+            None,
+        )
+        bytes_written = 2 * code_units_written.value
+
+        if bytes_written == 0 and bytes_to_be_written > 0:
+            raise OSError(self._get_error_message(GetLastError()))
+        return bytes_written
+
+
+class ConsoleStream:
+    def __init__(self, text_stream: t.TextIO, byte_stream: t.BinaryIO) -> None:
+        self._text_stream = text_stream
+        self.buffer = byte_stream
+
+    @property
+    def name(self) -> str:
+        return self.buffer.name
+
+    def write(self, x: t.AnyStr) -> int:
+        if isinstance(x, str):
+            return self._text_stream.write(x)
+        try:
+            self.flush()
+        except Exception:
+            pass
+        return self.buffer.write(x)
+
+    def writelines(self, lines: t.Iterable[t.AnyStr]) -> None:
+        for line in lines:
+            self.write(line)
+
+    def __getattr__(self, name: str) -> t.Any:
+        return getattr(self._text_stream, name)
+
+    def isatty(self) -> bool:
+        return self.buffer.isatty()
+
+    def __repr__(self):
+        return f"<ConsoleStream name={self.name!r} encoding={self.encoding!r}>"
+
+
+def _get_text_stdin(buffer_stream: t.BinaryIO) -> t.TextIO:
+    text_stream = _NonClosingTextIOWrapper(
+        io.BufferedReader(_WindowsConsoleReader(STDIN_HANDLE)),
+        "utf-16-le",
+        "strict",
+        line_buffering=True,
+    )
+    return t.cast(t.TextIO, ConsoleStream(text_stream, buffer_stream))
+
+
+def _get_text_stdout(buffer_stream: t.BinaryIO) -> t.TextIO:
+    text_stream = _NonClosingTextIOWrapper(
+        io.BufferedWriter(_WindowsConsoleWriter(STDOUT_HANDLE)),
+        "utf-16-le",
+        "strict",
+        line_buffering=True,
+    )
+    return t.cast(t.TextIO, ConsoleStream(text_stream, buffer_stream))
+
+
+def _get_text_stderr(buffer_stream: t.BinaryIO) -> t.TextIO:
+    text_stream = _NonClosingTextIOWrapper(
+        io.BufferedWriter(_WindowsConsoleWriter(STDERR_HANDLE)),
+        "utf-16-le",
+        "strict",
+        line_buffering=True,
+    )
+    return t.cast(t.TextIO, ConsoleStream(text_stream, buffer_stream))
+
+
+_stream_factories: t.Mapping[int, t.Callable[[t.BinaryIO], t.TextIO]] = {
+    0: _get_text_stdin,
+    1: _get_text_stdout,
+    2: _get_text_stderr,
+}
+
+
+def _is_console(f: t.TextIO) -> bool:
+    if not hasattr(f, "fileno"):
+        return False
+
+    try:
+        fileno = f.fileno()
+    except (OSError, io.UnsupportedOperation):
+        return False
+
+    handle = msvcrt.get_osfhandle(fileno)
+    return bool(GetConsoleMode(handle, byref(DWORD())))
+
+
+def _get_windows_console_stream(
+    f: t.TextIO, encoding: t.Optional[str], errors: t.Optional[str]
+) -> t.Optional[t.TextIO]:
+    if (
+        get_buffer is not None
+        and encoding in {"utf-16-le", None}
+        and errors in {"strict", None}
+        and _is_console(f)
+    ):
+        func = _stream_factories.get(f.fileno())
+        if func is not None:
+            b = getattr(f, "buffer", None)
+
+            if b is None:
+                return None
+
+            return func(b)
diff --git a/backend/venv/lib/python3.12/site-packages/click/core.py b/backend/venv/lib/python3.12/site-packages/click/core.py
new file mode 100644
index 00000000..e6305011
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/core.py
@@ -0,0 +1,3047 @@
+import enum
+import errno
+import inspect
+import os
+import sys
+import typing as t
+from collections import abc
+from contextlib import contextmanager
+from contextlib import ExitStack
+from functools import update_wrapper
+from gettext import gettext as _
+from gettext import ngettext
+from itertools import repeat
+from types import TracebackType
+
+from . import types
+from .exceptions import Abort
+from .exceptions import BadParameter
+from .exceptions import ClickException
+from .exceptions import Exit
+from .exceptions import MissingParameter
+from .exceptions import UsageError
+from .formatting import HelpFormatter
+from .formatting import join_options
+from .globals import pop_context
+from .globals import push_context
+from .parser import _flag_needs_value
+from .parser import OptionParser
+from .parser import split_opt
+from .termui import confirm
+from .termui import prompt
+from .termui import style
+from .utils import _detect_program_name
+from .utils import _expand_args
+from .utils import echo
+from .utils import make_default_short_help
+from .utils import make_str
+from .utils import PacifyFlushWrapper
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .decorators import HelpOption
+    from .shell_completion import CompletionItem
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+V = t.TypeVar("V")
+
+
+def _complete_visible_commands(
+    ctx: "Context", incomplete: str
+) -> t.Iterator[t.Tuple[str, "Command"]]:
+    """List all the subcommands of a group that start with the
+    incomplete value and aren't hidden.
+
+    :param ctx: Invocation context for the group.
+    :param incomplete: Value being completed. May be empty.
+    """
+    multi = t.cast(MultiCommand, ctx.command)
+
+    for name in multi.list_commands(ctx):
+        if name.startswith(incomplete):
+            command = multi.get_command(ctx, name)
+
+            if command is not None and not command.hidden:
+                yield name, command
+
+
+def _check_multicommand(
+    base_command: "MultiCommand", cmd_name: str, cmd: "Command", register: bool = False
+) -> None:
+    if not base_command.chain or not isinstance(cmd, MultiCommand):
+        return
+    if register:
+        hint = (
+            "It is not possible to add multi commands as children to"
+            " another multi command that is in chain mode."
+        )
+    else:
+        hint = (
+            "Found a multi command as subcommand to a multi command"
+            " that is in chain mode. This is not supported."
+        )
+    raise RuntimeError(
+        f"{hint}. Command {base_command.name!r} is set to chain and"
+        f" {cmd_name!r} was added as a subcommand but it in itself is a"
+        f" multi command. ({cmd_name!r} is a {type(cmd).__name__}"
+        f" within a chained {type(base_command).__name__} named"
+        f" {base_command.name!r})."
+    )
+
+
+def batch(iterable: t.Iterable[V], batch_size: int) -> t.List[t.Tuple[V, ...]]:
+    return list(zip(*repeat(iter(iterable), batch_size)))
+
+
+@contextmanager
+def augment_usage_errors(
+    ctx: "Context", param: t.Optional["Parameter"] = None
+) -> t.Iterator[None]:
+    """Context manager that attaches extra information to exceptions."""
+    try:
+        yield
+    except BadParameter as e:
+        if e.ctx is None:
+            e.ctx = ctx
+        if param is not None and e.param is None:
+            e.param = param
+        raise
+    except UsageError as e:
+        if e.ctx is None:
+            e.ctx = ctx
+        raise
+
+
+def iter_params_for_processing(
+    invocation_order: t.Sequence["Parameter"],
+    declaration_order: t.Sequence["Parameter"],
+) -> t.List["Parameter"]:
+    """Returns all declared parameters in the order they should be processed.
+
+    The declared parameters are re-shuffled depending on the order in which
+    they were invoked, as well as the eagerness of each parameters.
+
+    The invocation order takes precedence over the declaration order. I.e. the
+    order in which the user provided them to the CLI is respected.
+
+    This behavior and its effect on callback evaluation is detailed at:
+    https://click.palletsprojects.com/en/stable/advanced/#callback-evaluation-order
+    """
+
+    def sort_key(item: "Parameter") -> t.Tuple[bool, float]:
+        try:
+            idx: float = invocation_order.index(item)
+        except ValueError:
+            idx = float("inf")
+
+        return not item.is_eager, idx
+
+    return sorted(declaration_order, key=sort_key)
+
+
+class ParameterSource(enum.Enum):
+    """This is an :class:`~enum.Enum` that indicates the source of a
+    parameter's value.
+
+    Use :meth:`click.Context.get_parameter_source` to get the
+    source for a parameter by name.
+
+    .. versionchanged:: 8.0
+        Use :class:`~enum.Enum` and drop the ``validate`` method.
+
+    .. versionchanged:: 8.0
+        Added the ``PROMPT`` value.
+    """
+
+    COMMANDLINE = enum.auto()
+    """The value was provided by the command line args."""
+    ENVIRONMENT = enum.auto()
+    """The value was provided with an environment variable."""
+    DEFAULT = enum.auto()
+    """Used the default specified by the parameter."""
+    DEFAULT_MAP = enum.auto()
+    """Used a default provided by :attr:`Context.default_map`."""
+    PROMPT = enum.auto()
+    """Used a prompt to confirm a default or provide a value."""
+
+
+class Context:
+    """The context is a special internal object that holds state relevant
+    for the script execution at every single level.  It's normally invisible
+    to commands unless they opt-in to getting access to it.
+
+    The context is useful as it can pass internal objects around and can
+    control special execution features such as reading data from
+    environment variables.
+
+    A context can be used as context manager in which case it will call
+    :meth:`close` on teardown.
+
+    :param command: the command class for this context.
+    :param parent: the parent context.
+    :param info_name: the info name for this invocation.  Generally this
+                      is the most descriptive name for the script or
+                      command.  For the toplevel script it is usually
+                      the name of the script, for commands below it it's
+                      the name of the script.
+    :param obj: an arbitrary object of user data.
+    :param auto_envvar_prefix: the prefix to use for automatic environment
+                               variables.  If this is `None` then reading
+                               from environment variables is disabled.  This
+                               does not affect manually set environment
+                               variables which are always read.
+    :param default_map: a dictionary (like object) with default values
+                        for parameters.
+    :param terminal_width: the width of the terminal.  The default is
+                           inherit from parent context.  If no context
+                           defines the terminal width then auto
+                           detection will be applied.
+    :param max_content_width: the maximum width for content rendered by
+                              Click (this currently only affects help
+                              pages).  This defaults to 80 characters if
+                              not overridden.  In other words: even if the
+                              terminal is larger than that, Click will not
+                              format things wider than 80 characters by
+                              default.  In addition to that, formatters might
+                              add some safety mapping on the right.
+    :param resilient_parsing: if this flag is enabled then Click will
+                              parse without any interactivity or callback
+                              invocation.  Default values will also be
+                              ignored.  This is useful for implementing
+                              things such as completion support.
+    :param allow_extra_args: if this is set to `True` then extra arguments
+                             at the end will not raise an error and will be
+                             kept on the context.  The default is to inherit
+                             from the command.
+    :param allow_interspersed_args: if this is set to `False` then options
+                                    and arguments cannot be mixed.  The
+                                    default is to inherit from the command.
+    :param ignore_unknown_options: instructs click to ignore options it does
+                                   not know and keeps them for later
+                                   processing.
+    :param help_option_names: optionally a list of strings that define how
+                              the default help parameter is named.  The
+                              default is ``['--help']``.
+    :param token_normalize_func: an optional function that is used to
+                                 normalize tokens (options, choices,
+                                 etc.).  This for instance can be used to
+                                 implement case insensitive behavior.
+    :param color: controls if the terminal supports ANSI colors or not.  The
+                  default is autodetection.  This is only needed if ANSI
+                  codes are used in texts that Click prints which is by
+                  default not the case.  This for instance would affect
+                  help output.
+    :param show_default: Show the default value for commands. If this
+        value is not set, it defaults to the value from the parent
+        context. ``Command.show_default`` overrides this default for the
+        specific command.
+
+    .. versionchanged:: 8.1
+        The ``show_default`` parameter is overridden by
+        ``Command.show_default``, instead of the other way around.
+
+    .. versionchanged:: 8.0
+        The ``show_default`` parameter defaults to the value from the
+        parent context.
+
+    .. versionchanged:: 7.1
+       Added the ``show_default`` parameter.
+
+    .. versionchanged:: 4.0
+        Added the ``color``, ``ignore_unknown_options``, and
+        ``max_content_width`` parameters.
+
+    .. versionchanged:: 3.0
+        Added the ``allow_extra_args`` and ``allow_interspersed_args``
+        parameters.
+
+    .. versionchanged:: 2.0
+        Added the ``resilient_parsing``, ``help_option_names``, and
+        ``token_normalize_func`` parameters.
+    """
+
+    #: The formatter class to create with :meth:`make_formatter`.
+    #:
+    #: .. versionadded:: 8.0
+    formatter_class: t.Type["HelpFormatter"] = HelpFormatter
+
+    def __init__(
+        self,
+        command: "Command",
+        parent: t.Optional["Context"] = None,
+        info_name: t.Optional[str] = None,
+        obj: t.Optional[t.Any] = None,
+        auto_envvar_prefix: t.Optional[str] = None,
+        default_map: t.Optional[t.MutableMapping[str, t.Any]] = None,
+        terminal_width: t.Optional[int] = None,
+        max_content_width: t.Optional[int] = None,
+        resilient_parsing: bool = False,
+        allow_extra_args: t.Optional[bool] = None,
+        allow_interspersed_args: t.Optional[bool] = None,
+        ignore_unknown_options: t.Optional[bool] = None,
+        help_option_names: t.Optional[t.List[str]] = None,
+        token_normalize_func: t.Optional[t.Callable[[str], str]] = None,
+        color: t.Optional[bool] = None,
+        show_default: t.Optional[bool] = None,
+    ) -> None:
+        #: the parent context or `None` if none exists.
+        self.parent = parent
+        #: the :class:`Command` for this context.
+        self.command = command
+        #: the descriptive information name
+        self.info_name = info_name
+        #: Map of parameter names to their parsed values. Parameters
+        #: with ``expose_value=False`` are not stored.
+        self.params: t.Dict[str, t.Any] = {}
+        #: the leftover arguments.
+        self.args: t.List[str] = []
+        #: protected arguments.  These are arguments that are prepended
+        #: to `args` when certain parsing scenarios are encountered but
+        #: must be never propagated to another arguments.  This is used
+        #: to implement nested parsing.
+        self.protected_args: t.List[str] = []
+        #: the collected prefixes of the command's options.
+        self._opt_prefixes: t.Set[str] = set(parent._opt_prefixes) if parent else set()
+
+        if obj is None and parent is not None:
+            obj = parent.obj
+
+        #: the user object stored.
+        self.obj: t.Any = obj
+        self._meta: t.Dict[str, t.Any] = getattr(parent, "meta", {})
+
+        #: A dictionary (-like object) with defaults for parameters.
+        if (
+            default_map is None
+            and info_name is not None
+            and parent is not None
+            and parent.default_map is not None
+        ):
+            default_map = parent.default_map.get(info_name)
+
+        self.default_map: t.Optional[t.MutableMapping[str, t.Any]] = default_map
+
+        #: This flag indicates if a subcommand is going to be executed. A
+        #: group callback can use this information to figure out if it's
+        #: being executed directly or because the execution flow passes
+        #: onwards to a subcommand. By default it's None, but it can be
+        #: the name of the subcommand to execute.
+        #:
+        #: If chaining is enabled this will be set to ``'*'`` in case
+        #: any commands are executed.  It is however not possible to
+        #: figure out which ones.  If you require this knowledge you
+        #: should use a :func:`result_callback`.
+        self.invoked_subcommand: t.Optional[str] = None
+
+        if terminal_width is None and parent is not None:
+            terminal_width = parent.terminal_width
+
+        #: The width of the terminal (None is autodetection).
+        self.terminal_width: t.Optional[int] = terminal_width
+
+        if max_content_width is None and parent is not None:
+            max_content_width = parent.max_content_width
+
+        #: The maximum width of formatted content (None implies a sensible
+        #: default which is 80 for most things).
+        self.max_content_width: t.Optional[int] = max_content_width
+
+        if allow_extra_args is None:
+            allow_extra_args = command.allow_extra_args
+
+        #: Indicates if the context allows extra args or if it should
+        #: fail on parsing.
+        #:
+        #: .. versionadded:: 3.0
+        self.allow_extra_args = allow_extra_args
+
+        if allow_interspersed_args is None:
+            allow_interspersed_args = command.allow_interspersed_args
+
+        #: Indicates if the context allows mixing of arguments and
+        #: options or not.
+        #:
+        #: .. versionadded:: 3.0
+        self.allow_interspersed_args: bool = allow_interspersed_args
+
+        if ignore_unknown_options is None:
+            ignore_unknown_options = command.ignore_unknown_options
+
+        #: Instructs click to ignore options that a command does not
+        #: understand and will store it on the context for later
+        #: processing.  This is primarily useful for situations where you
+        #: want to call into external programs.  Generally this pattern is
+        #: strongly discouraged because it's not possibly to losslessly
+        #: forward all arguments.
+        #:
+        #: .. versionadded:: 4.0
+        self.ignore_unknown_options: bool = ignore_unknown_options
+
+        if help_option_names is None:
+            if parent is not None:
+                help_option_names = parent.help_option_names
+            else:
+                help_option_names = ["--help"]
+
+        #: The names for the help options.
+        self.help_option_names: t.List[str] = help_option_names
+
+        if token_normalize_func is None and parent is not None:
+            token_normalize_func = parent.token_normalize_func
+
+        #: An optional normalization function for tokens.  This is
+        #: options, choices, commands etc.
+        self.token_normalize_func: t.Optional[t.Callable[[str], str]] = (
+            token_normalize_func
+        )
+
+        #: Indicates if resilient parsing is enabled.  In that case Click
+        #: will do its best to not cause any failures and default values
+        #: will be ignored. Useful for completion.
+        self.resilient_parsing: bool = resilient_parsing
+
+        # If there is no envvar prefix yet, but the parent has one and
+        # the command on this level has a name, we can expand the envvar
+        # prefix automatically.
+        if auto_envvar_prefix is None:
+            if (
+                parent is not None
+                and parent.auto_envvar_prefix is not None
+                and self.info_name is not None
+            ):
+                auto_envvar_prefix = (
+                    f"{parent.auto_envvar_prefix}_{self.info_name.upper()}"
+                )
+        else:
+            auto_envvar_prefix = auto_envvar_prefix.upper()
+
+        if auto_envvar_prefix is not None:
+            auto_envvar_prefix = auto_envvar_prefix.replace("-", "_")
+
+        self.auto_envvar_prefix: t.Optional[str] = auto_envvar_prefix
+
+        if color is None and parent is not None:
+            color = parent.color
+
+        #: Controls if styling output is wanted or not.
+        self.color: t.Optional[bool] = color
+
+        if show_default is None and parent is not None:
+            show_default = parent.show_default
+
+        #: Show option default values when formatting help text.
+        self.show_default: t.Optional[bool] = show_default
+
+        self._close_callbacks: t.List[t.Callable[[], t.Any]] = []
+        self._depth = 0
+        self._parameter_source: t.Dict[str, ParameterSource] = {}
+        self._exit_stack = ExitStack()
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        """Gather information that could be useful for a tool generating
+        user-facing documentation. This traverses the entire CLI
+        structure.
+
+        .. code-block:: python
+
+            with Context(cli) as ctx:
+                info = ctx.to_info_dict()
+
+        .. versionadded:: 8.0
+        """
+        return {
+            "command": self.command.to_info_dict(self),
+            "info_name": self.info_name,
+            "allow_extra_args": self.allow_extra_args,
+            "allow_interspersed_args": self.allow_interspersed_args,
+            "ignore_unknown_options": self.ignore_unknown_options,
+            "auto_envvar_prefix": self.auto_envvar_prefix,
+        }
+
+    def __enter__(self) -> "Context":
+        self._depth += 1
+        push_context(self)
+        return self
+
+    def __exit__(
+        self,
+        exc_type: t.Optional[t.Type[BaseException]],
+        exc_value: t.Optional[BaseException],
+        tb: t.Optional[TracebackType],
+    ) -> None:
+        self._depth -= 1
+        if self._depth == 0:
+            self.close()
+        pop_context()
+
+    @contextmanager
+    def scope(self, cleanup: bool = True) -> t.Iterator["Context"]:
+        """This helper method can be used with the context object to promote
+        it to the current thread local (see :func:`get_current_context`).
+        The default behavior of this is to invoke the cleanup functions which
+        can be disabled by setting `cleanup` to `False`.  The cleanup
+        functions are typically used for things such as closing file handles.
+
+        If the cleanup is intended the context object can also be directly
+        used as a context manager.
+
+        Example usage::
+
+            with ctx.scope():
+                assert get_current_context() is ctx
+
+        This is equivalent::
+
+            with ctx:
+                assert get_current_context() is ctx
+
+        .. versionadded:: 5.0
+
+        :param cleanup: controls if the cleanup functions should be run or
+                        not.  The default is to run these functions.  In
+                        some situations the context only wants to be
+                        temporarily pushed in which case this can be disabled.
+                        Nested pushes automatically defer the cleanup.
+        """
+        if not cleanup:
+            self._depth += 1
+        try:
+            with self as rv:
+                yield rv
+        finally:
+            if not cleanup:
+                self._depth -= 1
+
+    @property
+    def meta(self) -> t.Dict[str, t.Any]:
+        """This is a dictionary which is shared with all the contexts
+        that are nested.  It exists so that click utilities can store some
+        state here if they need to.  It is however the responsibility of
+        that code to manage this dictionary well.
+
+        The keys are supposed to be unique dotted strings.  For instance
+        module paths are a good choice for it.  What is stored in there is
+        irrelevant for the operation of click.  However what is important is
+        that code that places data here adheres to the general semantics of
+        the system.
+
+        Example usage::
+
+            LANG_KEY = f'{__name__}.lang'
+
+            def set_language(value):
+                ctx = get_current_context()
+                ctx.meta[LANG_KEY] = value
+
+            def get_language():
+                return get_current_context().meta.get(LANG_KEY, 'en_US')
+
+        .. versionadded:: 5.0
+        """
+        return self._meta
+
+    def make_formatter(self) -> HelpFormatter:
+        """Creates the :class:`~click.HelpFormatter` for the help and
+        usage output.
+
+        To quickly customize the formatter class used without overriding
+        this method, set the :attr:`formatter_class` attribute.
+
+        .. versionchanged:: 8.0
+            Added the :attr:`formatter_class` attribute.
+        """
+        return self.formatter_class(
+            width=self.terminal_width, max_width=self.max_content_width
+        )
+
+    def with_resource(self, context_manager: t.ContextManager[V]) -> V:
+        """Register a resource as if it were used in a ``with``
+        statement. The resource will be cleaned up when the context is
+        popped.
+
+        Uses :meth:`contextlib.ExitStack.enter_context`. It calls the
+        resource's ``__enter__()`` method and returns the result. When
+        the context is popped, it closes the stack, which calls the
+        resource's ``__exit__()`` method.
+
+        To register a cleanup function for something that isn't a
+        context manager, use :meth:`call_on_close`. Or use something
+        from :mod:`contextlib` to turn it into a context manager first.
+
+        .. code-block:: python
+
+            @click.group()
+            @click.option("--name")
+            @click.pass_context
+            def cli(ctx):
+                ctx.obj = ctx.with_resource(connect_db(name))
+
+        :param context_manager: The context manager to enter.
+        :return: Whatever ``context_manager.__enter__()`` returns.
+
+        .. versionadded:: 8.0
+        """
+        return self._exit_stack.enter_context(context_manager)
+
+    def call_on_close(self, f: t.Callable[..., t.Any]) -> t.Callable[..., t.Any]:
+        """Register a function to be called when the context tears down.
+
+        This can be used to close resources opened during the script
+        execution. Resources that support Python's context manager
+        protocol which would be used in a ``with`` statement should be
+        registered with :meth:`with_resource` instead.
+
+        :param f: The function to execute on teardown.
+        """
+        return self._exit_stack.callback(f)
+
+    def close(self) -> None:
+        """Invoke all close callbacks registered with
+        :meth:`call_on_close`, and exit all context managers entered
+        with :meth:`with_resource`.
+        """
+        self._exit_stack.close()
+        # In case the context is reused, create a new exit stack.
+        self._exit_stack = ExitStack()
+
+    @property
+    def command_path(self) -> str:
+        """The computed command path.  This is used for the ``usage``
+        information on the help page.  It's automatically created by
+        combining the info names of the chain of contexts to the root.
+        """
+        rv = ""
+        if self.info_name is not None:
+            rv = self.info_name
+        if self.parent is not None:
+            parent_command_path = [self.parent.command_path]
+
+            if isinstance(self.parent.command, Command):
+                for param in self.parent.command.get_params(self):
+                    parent_command_path.extend(param.get_usage_pieces(self))
+
+            rv = f"{' '.join(parent_command_path)} {rv}"
+        return rv.lstrip()
+
+    def find_root(self) -> "Context":
+        """Finds the outermost context."""
+        node = self
+        while node.parent is not None:
+            node = node.parent
+        return node
+
+    def find_object(self, object_type: t.Type[V]) -> t.Optional[V]:
+        """Finds the closest object of a given type."""
+        node: t.Optional[Context] = self
+
+        while node is not None:
+            if isinstance(node.obj, object_type):
+                return node.obj
+
+            node = node.parent
+
+        return None
+
+    def ensure_object(self, object_type: t.Type[V]) -> V:
+        """Like :meth:`find_object` but sets the innermost object to a
+        new instance of `object_type` if it does not exist.
+        """
+        rv = self.find_object(object_type)
+        if rv is None:
+            self.obj = rv = object_type()
+        return rv
+
+    @t.overload
+    def lookup_default(
+        self, name: str, call: "te.Literal[True]" = True
+    ) -> t.Optional[t.Any]: ...
+
+    @t.overload
+    def lookup_default(
+        self, name: str, call: "te.Literal[False]" = ...
+    ) -> t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]]: ...
+
+    def lookup_default(self, name: str, call: bool = True) -> t.Optional[t.Any]:
+        """Get the default for a parameter from :attr:`default_map`.
+
+        :param name: Name of the parameter.
+        :param call: If the default is a callable, call it. Disable to
+            return the callable instead.
+
+        .. versionchanged:: 8.0
+            Added the ``call`` parameter.
+        """
+        if self.default_map is not None:
+            value = self.default_map.get(name)
+
+            if call and callable(value):
+                return value()
+
+            return value
+
+        return None
+
+    def fail(self, message: str) -> "te.NoReturn":
+        """Aborts the execution of the program with a specific error
+        message.
+
+        :param message: the error message to fail with.
+        """
+        raise UsageError(message, self)
+
+    def abort(self) -> "te.NoReturn":
+        """Aborts the script."""
+        raise Abort()
+
+    def exit(self, code: int = 0) -> "te.NoReturn":
+        """Exits the application with a given exit code."""
+        raise Exit(code)
+
+    def get_usage(self) -> str:
+        """Helper method to get formatted usage string for the current
+        context and command.
+        """
+        return self.command.get_usage(self)
+
+    def get_help(self) -> str:
+        """Helper method to get formatted help page for the current
+        context and command.
+        """
+        return self.command.get_help(self)
+
+    def _make_sub_context(self, command: "Command") -> "Context":
+        """Create a new context of the same type as this context, but
+        for a new command.
+
+        :meta private:
+        """
+        return type(self)(command, info_name=command.name, parent=self)
+
+    @t.overload
+    def invoke(
+        __self,
+        __callback: "t.Callable[..., V]",
+        *args: t.Any,
+        **kwargs: t.Any,
+    ) -> V: ...
+
+    @t.overload
+    def invoke(
+        __self,
+        __callback: "Command",
+        *args: t.Any,
+        **kwargs: t.Any,
+    ) -> t.Any: ...
+
+    def invoke(
+        __self,
+        __callback: t.Union["Command", "t.Callable[..., V]"],
+        *args: t.Any,
+        **kwargs: t.Any,
+    ) -> t.Union[t.Any, V]:
+        """Invokes a command callback in exactly the way it expects.  There
+        are two ways to invoke this method:
+
+        1.  the first argument can be a callback and all other arguments and
+            keyword arguments are forwarded directly to the function.
+        2.  the first argument is a click command object.  In that case all
+            arguments are forwarded as well but proper click parameters
+            (options and click arguments) must be keyword arguments and Click
+            will fill in defaults.
+
+        Note that before Click 3.2 keyword arguments were not properly filled
+        in against the intention of this code and no context was created.  For
+        more information about this change and why it was done in a bugfix
+        release see :ref:`upgrade-to-3.2`.
+
+        .. versionchanged:: 8.0
+            All ``kwargs`` are tracked in :attr:`params` so they will be
+            passed if :meth:`forward` is called at multiple levels.
+        """
+        if isinstance(__callback, Command):
+            other_cmd = __callback
+
+            if other_cmd.callback is None:
+                raise TypeError(
+                    "The given command does not have a callback that can be invoked."
+                )
+            else:
+                __callback = t.cast("t.Callable[..., V]", other_cmd.callback)
+
+            ctx = __self._make_sub_context(other_cmd)
+
+            for param in other_cmd.params:
+                if param.name not in kwargs and param.expose_value:
+                    kwargs[param.name] = param.type_cast_value(  # type: ignore
+                        ctx, param.get_default(ctx)
+                    )
+
+            # Track all kwargs as params, so that forward() will pass
+            # them on in subsequent calls.
+            ctx.params.update(kwargs)
+        else:
+            ctx = __self
+
+        with augment_usage_errors(__self):
+            with ctx:
+                return __callback(*args, **kwargs)
+
+    def forward(__self, __cmd: "Command", *args: t.Any, **kwargs: t.Any) -> t.Any:
+        """Similar to :meth:`invoke` but fills in default keyword
+        arguments from the current context if the other command expects
+        it.  This cannot invoke callbacks directly, only other commands.
+
+        .. versionchanged:: 8.0
+            All ``kwargs`` are tracked in :attr:`params` so they will be
+            passed if ``forward`` is called at multiple levels.
+        """
+        # Can only forward to other commands, not direct callbacks.
+        if not isinstance(__cmd, Command):
+            raise TypeError("Callback is not a command.")
+
+        for param in __self.params:
+            if param not in kwargs:
+                kwargs[param] = __self.params[param]
+
+        return __self.invoke(__cmd, *args, **kwargs)
+
+    def set_parameter_source(self, name: str, source: ParameterSource) -> None:
+        """Set the source of a parameter. This indicates the location
+        from which the value of the parameter was obtained.
+
+        :param name: The name of the parameter.
+        :param source: A member of :class:`~click.core.ParameterSource`.
+        """
+        self._parameter_source[name] = source
+
+    def get_parameter_source(self, name: str) -> t.Optional[ParameterSource]:
+        """Get the source of a parameter. This indicates the location
+        from which the value of the parameter was obtained.
+
+        This can be useful for determining when a user specified a value
+        on the command line that is the same as the default value. It
+        will be :attr:`~click.core.ParameterSource.DEFAULT` only if the
+        value was actually taken from the default.
+
+        :param name: The name of the parameter.
+        :rtype: ParameterSource
+
+        .. versionchanged:: 8.0
+            Returns ``None`` if the parameter was not provided from any
+            source.
+        """
+        return self._parameter_source.get(name)
+
+
+class BaseCommand:
+    """The base command implements the minimal API contract of commands.
+    Most code will never use this as it does not implement a lot of useful
+    functionality but it can act as the direct subclass of alternative
+    parsing methods that do not depend on the Click parser.
+
+    For instance, this can be used to bridge Click and other systems like
+    argparse or docopt.
+
+    Because base commands do not implement a lot of the API that other
+    parts of Click take for granted, they are not supported for all
+    operations.  For instance, they cannot be used with the decorators
+    usually and they have no built-in callback system.
+
+    .. versionchanged:: 2.0
+       Added the `context_settings` parameter.
+
+    :param name: the name of the command to use unless a group overrides it.
+    :param context_settings: an optional dictionary with defaults that are
+                             passed to the context object.
+    """
+
+    #: The context class to create with :meth:`make_context`.
+    #:
+    #: .. versionadded:: 8.0
+    context_class: t.Type[Context] = Context
+    #: the default for the :attr:`Context.allow_extra_args` flag.
+    allow_extra_args = False
+    #: the default for the :attr:`Context.allow_interspersed_args` flag.
+    allow_interspersed_args = True
+    #: the default for the :attr:`Context.ignore_unknown_options` flag.
+    ignore_unknown_options = False
+
+    def __init__(
+        self,
+        name: t.Optional[str],
+        context_settings: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> None:
+        #: the name the command thinks it has.  Upon registering a command
+        #: on a :class:`Group` the group will default the command name
+        #: with this information.  You should instead use the
+        #: :class:`Context`\'s :attr:`~Context.info_name` attribute.
+        self.name = name
+
+        if context_settings is None:
+            context_settings = {}
+
+        #: an optional dictionary with defaults passed to the context.
+        self.context_settings: t.MutableMapping[str, t.Any] = context_settings
+
+    def to_info_dict(self, ctx: Context) -> t.Dict[str, t.Any]:
+        """Gather information that could be useful for a tool generating
+        user-facing documentation. This traverses the entire structure
+        below this command.
+
+        Use :meth:`click.Context.to_info_dict` to traverse the entire
+        CLI structure.
+
+        :param ctx: A :class:`Context` representing this command.
+
+        .. versionadded:: 8.0
+        """
+        return {"name": self.name}
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__} {self.name}>"
+
+    def get_usage(self, ctx: Context) -> str:
+        raise NotImplementedError("Base commands cannot get usage")
+
+    def get_help(self, ctx: Context) -> str:
+        raise NotImplementedError("Base commands cannot get help")
+
+    def make_context(
+        self,
+        info_name: t.Optional[str],
+        args: t.List[str],
+        parent: t.Optional[Context] = None,
+        **extra: t.Any,
+    ) -> Context:
+        """This function when given an info name and arguments will kick
+        off the parsing and create a new :class:`Context`.  It does not
+        invoke the actual command callback though.
+
+        To quickly customize the context class used without overriding
+        this method, set the :attr:`context_class` attribute.
+
+        :param info_name: the info name for this invocation.  Generally this
+                          is the most descriptive name for the script or
+                          command.  For the toplevel script it's usually
+                          the name of the script, for commands below it's
+                          the name of the command.
+        :param args: the arguments to parse as list of strings.
+        :param parent: the parent context if available.
+        :param extra: extra keyword arguments forwarded to the context
+                      constructor.
+
+        .. versionchanged:: 8.0
+            Added the :attr:`context_class` attribute.
+        """
+        for key, value in self.context_settings.items():
+            if key not in extra:
+                extra[key] = value
+
+        ctx = self.context_class(
+            self,  # type: ignore[arg-type]
+            info_name=info_name,
+            parent=parent,
+            **extra,
+        )
+
+        with ctx.scope(cleanup=False):
+            self.parse_args(ctx, args)
+        return ctx
+
+    def parse_args(self, ctx: Context, args: t.List[str]) -> t.List[str]:
+        """Given a context and a list of arguments this creates the parser
+        and parses the arguments, then modifies the context as necessary.
+        This is automatically invoked by :meth:`make_context`.
+        """
+        raise NotImplementedError("Base commands do not know how to parse arguments.")
+
+    def invoke(self, ctx: Context) -> t.Any:
+        """Given a context, this invokes the command.  The default
+        implementation is raising a not implemented error.
+        """
+        raise NotImplementedError("Base commands are not invocable by default")
+
+    def shell_complete(self, ctx: Context, incomplete: str) -> t.List["CompletionItem"]:
+        """Return a list of completions for the incomplete value. Looks
+        at the names of chained multi-commands.
+
+        Any command could be part of a chained multi-command, so sibling
+        commands are valid at any point during command completion. Other
+        command classes will return more completions.
+
+        :param ctx: Invocation context for this command.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        results: t.List[CompletionItem] = []
+
+        while ctx.parent is not None:
+            ctx = ctx.parent
+
+            if isinstance(ctx.command, MultiCommand) and ctx.command.chain:
+                results.extend(
+                    CompletionItem(name, help=command.get_short_help_str())
+                    for name, command in _complete_visible_commands(ctx, incomplete)
+                    if name not in ctx.protected_args
+                )
+
+        return results
+
+    @t.overload
+    def main(
+        self,
+        args: t.Optional[t.Sequence[str]] = None,
+        prog_name: t.Optional[str] = None,
+        complete_var: t.Optional[str] = None,
+        standalone_mode: "te.Literal[True]" = True,
+        **extra: t.Any,
+    ) -> "te.NoReturn": ...
+
+    @t.overload
+    def main(
+        self,
+        args: t.Optional[t.Sequence[str]] = None,
+        prog_name: t.Optional[str] = None,
+        complete_var: t.Optional[str] = None,
+        standalone_mode: bool = ...,
+        **extra: t.Any,
+    ) -> t.Any: ...
+
+    def main(
+        self,
+        args: t.Optional[t.Sequence[str]] = None,
+        prog_name: t.Optional[str] = None,
+        complete_var: t.Optional[str] = None,
+        standalone_mode: bool = True,
+        windows_expand_args: bool = True,
+        **extra: t.Any,
+    ) -> t.Any:
+        """This is the way to invoke a script with all the bells and
+        whistles as a command line application.  This will always terminate
+        the application after a call.  If this is not wanted, ``SystemExit``
+        needs to be caught.
+
+        This method is also available by directly calling the instance of
+        a :class:`Command`.
+
+        :param args: the arguments that should be used for parsing.  If not
+                     provided, ``sys.argv[1:]`` is used.
+        :param prog_name: the program name that should be used.  By default
+                          the program name is constructed by taking the file
+                          name from ``sys.argv[0]``.
+        :param complete_var: the environment variable that controls the
+                             bash completion support.  The default is
+                             ``"_<prog_name>_COMPLETE"`` with prog_name in
+                             uppercase.
+        :param standalone_mode: the default behavior is to invoke the script
+                                in standalone mode.  Click will then
+                                handle exceptions and convert them into
+                                error messages and the function will never
+                                return but shut down the interpreter.  If
+                                this is set to `False` they will be
+                                propagated to the caller and the return
+                                value of this function is the return value
+                                of :meth:`invoke`.
+        :param windows_expand_args: Expand glob patterns, user dir, and
+            env vars in command line args on Windows.
+        :param extra: extra keyword arguments are forwarded to the context
+                      constructor.  See :class:`Context` for more information.
+
+        .. versionchanged:: 8.0.1
+            Added the ``windows_expand_args`` parameter to allow
+            disabling command line arg expansion on Windows.
+
+        .. versionchanged:: 8.0
+            When taking arguments from ``sys.argv`` on Windows, glob
+            patterns, user dir, and env vars are expanded.
+
+        .. versionchanged:: 3.0
+           Added the ``standalone_mode`` parameter.
+        """
+        if args is None:
+            args = sys.argv[1:]
+
+            if os.name == "nt" and windows_expand_args:
+                args = _expand_args(args)
+        else:
+            args = list(args)
+
+        if prog_name is None:
+            prog_name = _detect_program_name()
+
+        # Process shell completion requests and exit early.
+        self._main_shell_completion(extra, prog_name, complete_var)
+
+        try:
+            try:
+                with self.make_context(prog_name, args, **extra) as ctx:
+                    rv = self.invoke(ctx)
+                    if not standalone_mode:
+                        return rv
+                    # it's not safe to `ctx.exit(rv)` here!
+                    # note that `rv` may actually contain data like "1" which
+                    # has obvious effects
+                    # more subtle case: `rv=[None, None]` can come out of
+                    # chained commands which all returned `None` -- so it's not
+                    # even always obvious that `rv` indicates success/failure
+                    # by its truthiness/falsiness
+                    ctx.exit()
+            except (EOFError, KeyboardInterrupt) as e:
+                echo(file=sys.stderr)
+                raise Abort() from e
+            except ClickException as e:
+                if not standalone_mode:
+                    raise
+                e.show()
+                sys.exit(e.exit_code)
+            except OSError as e:
+                if e.errno == errno.EPIPE:
+                    sys.stdout = t.cast(t.TextIO, PacifyFlushWrapper(sys.stdout))
+                    sys.stderr = t.cast(t.TextIO, PacifyFlushWrapper(sys.stderr))
+                    sys.exit(1)
+                else:
+                    raise
+        except Exit as e:
+            if standalone_mode:
+                sys.exit(e.exit_code)
+            else:
+                # in non-standalone mode, return the exit code
+                # note that this is only reached if `self.invoke` above raises
+                # an Exit explicitly -- thus bypassing the check there which
+                # would return its result
+                # the results of non-standalone execution may therefore be
+                # somewhat ambiguous: if there are codepaths which lead to
+                # `ctx.exit(1)` and to `return 1`, the caller won't be able to
+                # tell the difference between the two
+                return e.exit_code
+        except Abort:
+            if not standalone_mode:
+                raise
+            echo(_("Aborted!"), file=sys.stderr)
+            sys.exit(1)
+
+    def _main_shell_completion(
+        self,
+        ctx_args: t.MutableMapping[str, t.Any],
+        prog_name: str,
+        complete_var: t.Optional[str] = None,
+    ) -> None:
+        """Check if the shell is asking for tab completion, process
+        that, then exit early. Called from :meth:`main` before the
+        program is invoked.
+
+        :param prog_name: Name of the executable in the shell.
+        :param complete_var: Name of the environment variable that holds
+            the completion instruction. Defaults to
+            ``_{PROG_NAME}_COMPLETE``.
+
+        .. versionchanged:: 8.2.0
+            Dots (``.``) in ``prog_name`` are replaced with underscores (``_``).
+        """
+        if complete_var is None:
+            complete_name = prog_name.replace("-", "_").replace(".", "_")
+            complete_var = f"_{complete_name}_COMPLETE".upper()
+
+        instruction = os.environ.get(complete_var)
+
+        if not instruction:
+            return
+
+        from .shell_completion import shell_complete
+
+        rv = shell_complete(self, ctx_args, prog_name, complete_var, instruction)
+        sys.exit(rv)
+
+    def __call__(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        """Alias for :meth:`main`."""
+        return self.main(*args, **kwargs)
+
+
+class Command(BaseCommand):
+    """Commands are the basic building block of command line interfaces in
+    Click.  A basic command handles command line parsing and might dispatch
+    more parsing to commands nested below it.
+
+    :param name: the name of the command to use unless a group overrides it.
+    :param context_settings: an optional dictionary with defaults that are
+                             passed to the context object.
+    :param callback: the callback to invoke.  This is optional.
+    :param params: the parameters to register with this command.  This can
+                   be either :class:`Option` or :class:`Argument` objects.
+    :param help: the help string to use for this command.
+    :param epilog: like the help string but it's printed at the end of the
+                   help page after everything else.
+    :param short_help: the short help to use for this command.  This is
+                       shown on the command listing of the parent command.
+    :param add_help_option: by default each command registers a ``--help``
+                            option.  This can be disabled by this parameter.
+    :param no_args_is_help: this controls what happens if no arguments are
+                            provided.  This option is disabled by default.
+                            If enabled this will add ``--help`` as argument
+                            if no arguments are passed
+    :param hidden: hide this command from help outputs.
+
+    :param deprecated: issues a message indicating that
+                             the command is deprecated.
+
+    .. versionchanged:: 8.1
+        ``help``, ``epilog``, and ``short_help`` are stored unprocessed,
+        all formatting is done when outputting help text, not at init,
+        and is done even if not using the ``@command`` decorator.
+
+    .. versionchanged:: 8.0
+        Added a ``repr`` showing the command name.
+
+    .. versionchanged:: 7.1
+        Added the ``no_args_is_help`` parameter.
+
+    .. versionchanged:: 2.0
+        Added the ``context_settings`` parameter.
+    """
+
+    def __init__(
+        self,
+        name: t.Optional[str],
+        context_settings: t.Optional[t.MutableMapping[str, t.Any]] = None,
+        callback: t.Optional[t.Callable[..., t.Any]] = None,
+        params: t.Optional[t.List["Parameter"]] = None,
+        help: t.Optional[str] = None,
+        epilog: t.Optional[str] = None,
+        short_help: t.Optional[str] = None,
+        options_metavar: t.Optional[str] = "[OPTIONS]",
+        add_help_option: bool = True,
+        no_args_is_help: bool = False,
+        hidden: bool = False,
+        deprecated: bool = False,
+    ) -> None:
+        super().__init__(name, context_settings)
+        #: the callback to execute when the command fires.  This might be
+        #: `None` in which case nothing happens.
+        self.callback = callback
+        #: the list of parameters for this command in the order they
+        #: should show up in the help page and execute.  Eager parameters
+        #: will automatically be handled before non eager ones.
+        self.params: t.List[Parameter] = params or []
+        self.help = help
+        self.epilog = epilog
+        self.options_metavar = options_metavar
+        self.short_help = short_help
+        self.add_help_option = add_help_option
+        self._help_option: t.Optional[HelpOption] = None
+        self.no_args_is_help = no_args_is_help
+        self.hidden = hidden
+        self.deprecated = deprecated
+
+    def to_info_dict(self, ctx: Context) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict(ctx)
+        info_dict.update(
+            params=[param.to_info_dict() for param in self.get_params(ctx)],
+            help=self.help,
+            epilog=self.epilog,
+            short_help=self.short_help,
+            hidden=self.hidden,
+            deprecated=self.deprecated,
+        )
+        return info_dict
+
+    def get_usage(self, ctx: Context) -> str:
+        """Formats the usage line into a string and returns it.
+
+        Calls :meth:`format_usage` internally.
+        """
+        formatter = ctx.make_formatter()
+        self.format_usage(ctx, formatter)
+        return formatter.getvalue().rstrip("\n")
+
+    def get_params(self, ctx: Context) -> t.List["Parameter"]:
+        rv = self.params
+        help_option = self.get_help_option(ctx)
+
+        if help_option is not None:
+            rv = [*rv, help_option]
+
+        return rv
+
+    def format_usage(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Writes the usage line into the formatter.
+
+        This is a low-level method called by :meth:`get_usage`.
+        """
+        pieces = self.collect_usage_pieces(ctx)
+        formatter.write_usage(ctx.command_path, " ".join(pieces))
+
+    def collect_usage_pieces(self, ctx: Context) -> t.List[str]:
+        """Returns all the pieces that go into the usage line and returns
+        it as a list of strings.
+        """
+        rv = [self.options_metavar] if self.options_metavar else []
+
+        for param in self.get_params(ctx):
+            rv.extend(param.get_usage_pieces(ctx))
+
+        return rv
+
+    def get_help_option_names(self, ctx: Context) -> t.List[str]:
+        """Returns the names for the help option."""
+        all_names = set(ctx.help_option_names)
+        for param in self.params:
+            all_names.difference_update(param.opts)
+            all_names.difference_update(param.secondary_opts)
+        return list(all_names)
+
+    def get_help_option(self, ctx: Context) -> t.Optional["Option"]:
+        """Returns the help option object.
+
+        Unless ``add_help_option`` is ``False``.
+
+        .. versionchanged:: 8.1.8
+            The help option is now cached to avoid creating it multiple times.
+        """
+        help_options = self.get_help_option_names(ctx)
+
+        if not help_options or not self.add_help_option:
+            return None
+
+        # Cache the help option object in private _help_option attribute to
+        # avoid creating it multiple times. Not doing this will break the
+        # callback odering by iter_params_for_processing(), which relies on
+        # object comparison.
+        if self._help_option is None:
+            # Avoid circular import.
+            from .decorators import HelpOption
+
+            self._help_option = HelpOption(help_options)
+
+        return self._help_option
+
+    def make_parser(self, ctx: Context) -> OptionParser:
+        """Creates the underlying option parser for this command."""
+        parser = OptionParser(ctx)
+        for param in self.get_params(ctx):
+            param.add_to_parser(parser, ctx)
+        return parser
+
+    def get_help(self, ctx: Context) -> str:
+        """Formats the help into a string and returns it.
+
+        Calls :meth:`format_help` internally.
+        """
+        formatter = ctx.make_formatter()
+        self.format_help(ctx, formatter)
+        return formatter.getvalue().rstrip("\n")
+
+    def get_short_help_str(self, limit: int = 45) -> str:
+        """Gets short help for the command or makes it by shortening the
+        long help string.
+        """
+        if self.short_help:
+            text = inspect.cleandoc(self.short_help)
+        elif self.help:
+            text = make_default_short_help(self.help, limit)
+        else:
+            text = ""
+
+        if self.deprecated:
+            text = _("(Deprecated) {text}").format(text=text)
+
+        return text.strip()
+
+    def format_help(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Writes the help into the formatter if it exists.
+
+        This is a low-level method called by :meth:`get_help`.
+
+        This calls the following methods:
+
+        -   :meth:`format_usage`
+        -   :meth:`format_help_text`
+        -   :meth:`format_options`
+        -   :meth:`format_epilog`
+        """
+        self.format_usage(ctx, formatter)
+        self.format_help_text(ctx, formatter)
+        self.format_options(ctx, formatter)
+        self.format_epilog(ctx, formatter)
+
+    def format_help_text(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Writes the help text to the formatter if it exists."""
+        if self.help is not None:
+            # truncate the help text to the first form feed
+            text = inspect.cleandoc(self.help).partition("\f")[0]
+        else:
+            text = ""
+
+        if self.deprecated:
+            text = _("(Deprecated) {text}").format(text=text)
+
+        if text:
+            formatter.write_paragraph()
+
+            with formatter.indentation():
+                formatter.write_text(text)
+
+    def format_options(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Writes all the options into the formatter if they exist."""
+        opts = []
+        for param in self.get_params(ctx):
+            rv = param.get_help_record(ctx)
+            if rv is not None:
+                opts.append(rv)
+
+        if opts:
+            with formatter.section(_("Options")):
+                formatter.write_dl(opts)
+
+    def format_epilog(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Writes the epilog into the formatter if it exists."""
+        if self.epilog:
+            epilog = inspect.cleandoc(self.epilog)
+            formatter.write_paragraph()
+
+            with formatter.indentation():
+                formatter.write_text(epilog)
+
+    def parse_args(self, ctx: Context, args: t.List[str]) -> t.List[str]:
+        if not args and self.no_args_is_help and not ctx.resilient_parsing:
+            echo(ctx.get_help(), color=ctx.color)
+            ctx.exit()
+
+        parser = self.make_parser(ctx)
+        opts, args, param_order = parser.parse_args(args=args)
+
+        for param in iter_params_for_processing(param_order, self.get_params(ctx)):
+            value, args = param.handle_parse_result(ctx, opts, args)
+
+        if args and not ctx.allow_extra_args and not ctx.resilient_parsing:
+            ctx.fail(
+                ngettext(
+                    "Got unexpected extra argument ({args})",
+                    "Got unexpected extra arguments ({args})",
+                    len(args),
+                ).format(args=" ".join(map(str, args)))
+            )
+
+        ctx.args = args
+        ctx._opt_prefixes.update(parser._opt_prefixes)
+        return args
+
+    def invoke(self, ctx: Context) -> t.Any:
+        """Given a context, this invokes the attached callback (if it exists)
+        in the right way.
+        """
+        if self.deprecated:
+            message = _(
+                "DeprecationWarning: The command {name!r} is deprecated."
+            ).format(name=self.name)
+            echo(style(message, fg="red"), err=True)
+
+        if self.callback is not None:
+            return ctx.invoke(self.callback, **ctx.params)
+
+    def shell_complete(self, ctx: Context, incomplete: str) -> t.List["CompletionItem"]:
+        """Return a list of completions for the incomplete value. Looks
+        at the names of options and chained multi-commands.
+
+        :param ctx: Invocation context for this command.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        results: t.List[CompletionItem] = []
+
+        if incomplete and not incomplete[0].isalnum():
+            for param in self.get_params(ctx):
+                if (
+                    not isinstance(param, Option)
+                    or param.hidden
+                    or (
+                        not param.multiple
+                        and ctx.get_parameter_source(param.name)  # type: ignore
+                        is ParameterSource.COMMANDLINE
+                    )
+                ):
+                    continue
+
+                results.extend(
+                    CompletionItem(name, help=param.help)
+                    for name in [*param.opts, *param.secondary_opts]
+                    if name.startswith(incomplete)
+                )
+
+        results.extend(super().shell_complete(ctx, incomplete))
+        return results
+
+
+class MultiCommand(Command):
+    """A multi command is the basic implementation of a command that
+    dispatches to subcommands.  The most common version is the
+    :class:`Group`.
+
+    :param invoke_without_command: this controls how the multi command itself
+                                   is invoked.  By default it's only invoked
+                                   if a subcommand is provided.
+    :param no_args_is_help: this controls what happens if no arguments are
+                            provided.  This option is enabled by default if
+                            `invoke_without_command` is disabled or disabled
+                            if it's enabled.  If enabled this will add
+                            ``--help`` as argument if no arguments are
+                            passed.
+    :param subcommand_metavar: the string that is used in the documentation
+                               to indicate the subcommand place.
+    :param chain: if this is set to `True` chaining of multiple subcommands
+                  is enabled.  This restricts the form of commands in that
+                  they cannot have optional arguments but it allows
+                  multiple commands to be chained together.
+    :param result_callback: The result callback to attach to this multi
+        command. This can be set or changed later with the
+        :meth:`result_callback` decorator.
+    :param attrs: Other command arguments described in :class:`Command`.
+    """
+
+    allow_extra_args = True
+    allow_interspersed_args = False
+
+    def __init__(
+        self,
+        name: t.Optional[str] = None,
+        invoke_without_command: bool = False,
+        no_args_is_help: t.Optional[bool] = None,
+        subcommand_metavar: t.Optional[str] = None,
+        chain: bool = False,
+        result_callback: t.Optional[t.Callable[..., t.Any]] = None,
+        **attrs: t.Any,
+    ) -> None:
+        super().__init__(name, **attrs)
+
+        if no_args_is_help is None:
+            no_args_is_help = not invoke_without_command
+
+        self.no_args_is_help = no_args_is_help
+        self.invoke_without_command = invoke_without_command
+
+        if subcommand_metavar is None:
+            if chain:
+                subcommand_metavar = "COMMAND1 [ARGS]... [COMMAND2 [ARGS]...]..."
+            else:
+                subcommand_metavar = "COMMAND [ARGS]..."
+
+        self.subcommand_metavar = subcommand_metavar
+        self.chain = chain
+        # The result callback that is stored. This can be set or
+        # overridden with the :func:`result_callback` decorator.
+        self._result_callback = result_callback
+
+        if self.chain:
+            for param in self.params:
+                if isinstance(param, Argument) and not param.required:
+                    raise RuntimeError(
+                        "Multi commands in chain mode cannot have"
+                        " optional arguments."
+                    )
+
+    def to_info_dict(self, ctx: Context) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict(ctx)
+        commands = {}
+
+        for name in self.list_commands(ctx):
+            command = self.get_command(ctx, name)
+
+            if command is None:
+                continue
+
+            sub_ctx = ctx._make_sub_context(command)
+
+            with sub_ctx.scope(cleanup=False):
+                commands[name] = command.to_info_dict(sub_ctx)
+
+        info_dict.update(commands=commands, chain=self.chain)
+        return info_dict
+
+    def collect_usage_pieces(self, ctx: Context) -> t.List[str]:
+        rv = super().collect_usage_pieces(ctx)
+        rv.append(self.subcommand_metavar)
+        return rv
+
+    def format_options(self, ctx: Context, formatter: HelpFormatter) -> None:
+        super().format_options(ctx, formatter)
+        self.format_commands(ctx, formatter)
+
+    def result_callback(self, replace: bool = False) -> t.Callable[[F], F]:
+        """Adds a result callback to the command.  By default if a
+        result callback is already registered this will chain them but
+        this can be disabled with the `replace` parameter.  The result
+        callback is invoked with the return value of the subcommand
+        (or the list of return values from all subcommands if chaining
+        is enabled) as well as the parameters as they would be passed
+        to the main callback.
+
+        Example::
+
+            @click.group()
+            @click.option('-i', '--input', default=23)
+            def cli(input):
+                return 42
+
+            @cli.result_callback()
+            def process_result(result, input):
+                return result + input
+
+        :param replace: if set to `True` an already existing result
+                        callback will be removed.
+
+        .. versionchanged:: 8.0
+            Renamed from ``resultcallback``.
+
+        .. versionadded:: 3.0
+        """
+
+        def decorator(f: F) -> F:
+            old_callback = self._result_callback
+
+            if old_callback is None or replace:
+                self._result_callback = f
+                return f
+
+            def function(__value, *args, **kwargs):  # type: ignore
+                inner = old_callback(__value, *args, **kwargs)
+                return f(inner, *args, **kwargs)
+
+            self._result_callback = rv = update_wrapper(t.cast(F, function), f)
+            return rv  # type: ignore[return-value]
+
+        return decorator
+
+    def format_commands(self, ctx: Context, formatter: HelpFormatter) -> None:
+        """Extra format methods for multi methods that adds all the commands
+        after the options.
+        """
+        commands = []
+        for subcommand in self.list_commands(ctx):
+            cmd = self.get_command(ctx, subcommand)
+            # What is this, the tool lied about a command.  Ignore it
+            if cmd is None:
+                continue
+            if cmd.hidden:
+                continue
+
+            commands.append((subcommand, cmd))
+
+        # allow for 3 times the default spacing
+        if len(commands):
+            limit = formatter.width - 6 - max(len(cmd[0]) for cmd in commands)
+
+            rows = []
+            for subcommand, cmd in commands:
+                help = cmd.get_short_help_str(limit)
+                rows.append((subcommand, help))
+
+            if rows:
+                with formatter.section(_("Commands")):
+                    formatter.write_dl(rows)
+
+    def parse_args(self, ctx: Context, args: t.List[str]) -> t.List[str]:
+        if not args and self.no_args_is_help and not ctx.resilient_parsing:
+            echo(ctx.get_help(), color=ctx.color)
+            ctx.exit()
+
+        rest = super().parse_args(ctx, args)
+
+        if self.chain:
+            ctx.protected_args = rest
+            ctx.args = []
+        elif rest:
+            ctx.protected_args, ctx.args = rest[:1], rest[1:]
+
+        return ctx.args
+
+    def invoke(self, ctx: Context) -> t.Any:
+        def _process_result(value: t.Any) -> t.Any:
+            if self._result_callback is not None:
+                value = ctx.invoke(self._result_callback, value, **ctx.params)
+            return value
+
+        if not ctx.protected_args:
+            if self.invoke_without_command:
+                # No subcommand was invoked, so the result callback is
+                # invoked with the group return value for regular
+                # groups, or an empty list for chained groups.
+                with ctx:
+                    rv = super().invoke(ctx)
+                    return _process_result([] if self.chain else rv)
+            ctx.fail(_("Missing command."))
+
+        # Fetch args back out
+        args = [*ctx.protected_args, *ctx.args]
+        ctx.args = []
+        ctx.protected_args = []
+
+        # If we're not in chain mode, we only allow the invocation of a
+        # single command but we also inform the current context about the
+        # name of the command to invoke.
+        if not self.chain:
+            # Make sure the context is entered so we do not clean up
+            # resources until the result processor has worked.
+            with ctx:
+                cmd_name, cmd, args = self.resolve_command(ctx, args)
+                assert cmd is not None
+                ctx.invoked_subcommand = cmd_name
+                super().invoke(ctx)
+                sub_ctx = cmd.make_context(cmd_name, args, parent=ctx)
+                with sub_ctx:
+                    return _process_result(sub_ctx.command.invoke(sub_ctx))
+
+        # In chain mode we create the contexts step by step, but after the
+        # base command has been invoked.  Because at that point we do not
+        # know the subcommands yet, the invoked subcommand attribute is
+        # set to ``*`` to inform the command that subcommands are executed
+        # but nothing else.
+        with ctx:
+            ctx.invoked_subcommand = "*" if args else None
+            super().invoke(ctx)
+
+            # Otherwise we make every single context and invoke them in a
+            # chain.  In that case the return value to the result processor
+            # is the list of all invoked subcommand's results.
+            contexts = []
+            while args:
+                cmd_name, cmd, args = self.resolve_command(ctx, args)
+                assert cmd is not None
+                sub_ctx = cmd.make_context(
+                    cmd_name,
+                    args,
+                    parent=ctx,
+                    allow_extra_args=True,
+                    allow_interspersed_args=False,
+                )
+                contexts.append(sub_ctx)
+                args, sub_ctx.args = sub_ctx.args, []
+
+            rv = []
+            for sub_ctx in contexts:
+                with sub_ctx:
+                    rv.append(sub_ctx.command.invoke(sub_ctx))
+            return _process_result(rv)
+
+    def resolve_command(
+        self, ctx: Context, args: t.List[str]
+    ) -> t.Tuple[t.Optional[str], t.Optional[Command], t.List[str]]:
+        cmd_name = make_str(args[0])
+        original_cmd_name = cmd_name
+
+        # Get the command
+        cmd = self.get_command(ctx, cmd_name)
+
+        # If we can't find the command but there is a normalization
+        # function available, we try with that one.
+        if cmd is None and ctx.token_normalize_func is not None:
+            cmd_name = ctx.token_normalize_func(cmd_name)
+            cmd = self.get_command(ctx, cmd_name)
+
+        # If we don't find the command we want to show an error message
+        # to the user that it was not provided.  However, there is
+        # something else we should do: if the first argument looks like
+        # an option we want to kick off parsing again for arguments to
+        # resolve things like --help which now should go to the main
+        # place.
+        if cmd is None and not ctx.resilient_parsing:
+            if split_opt(cmd_name)[0]:
+                self.parse_args(ctx, ctx.args)
+            ctx.fail(_("No such command {name!r}.").format(name=original_cmd_name))
+        return cmd_name if cmd else None, cmd, args[1:]
+
+    def get_command(self, ctx: Context, cmd_name: str) -> t.Optional[Command]:
+        """Given a context and a command name, this returns a
+        :class:`Command` object if it exists or returns `None`.
+        """
+        raise NotImplementedError
+
+    def list_commands(self, ctx: Context) -> t.List[str]:
+        """Returns a list of subcommand names in the order they should
+        appear.
+        """
+        return []
+
+    def shell_complete(self, ctx: Context, incomplete: str) -> t.List["CompletionItem"]:
+        """Return a list of completions for the incomplete value. Looks
+        at the names of options, subcommands, and chained
+        multi-commands.
+
+        :param ctx: Invocation context for this command.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        results = [
+            CompletionItem(name, help=command.get_short_help_str())
+            for name, command in _complete_visible_commands(ctx, incomplete)
+        ]
+        results.extend(super().shell_complete(ctx, incomplete))
+        return results
+
+
+class Group(MultiCommand):
+    """A group allows a command to have subcommands attached. This is
+    the most common way to implement nesting in Click.
+
+    :param name: The name of the group command.
+    :param commands: A dict mapping names to :class:`Command` objects.
+        Can also be a list of :class:`Command`, which will use
+        :attr:`Command.name` to create the dict.
+    :param attrs: Other command arguments described in
+        :class:`MultiCommand`, :class:`Command`, and
+        :class:`BaseCommand`.
+
+    .. versionchanged:: 8.0
+        The ``commands`` argument can be a list of command objects.
+    """
+
+    #: If set, this is used by the group's :meth:`command` decorator
+    #: as the default :class:`Command` class. This is useful to make all
+    #: subcommands use a custom command class.
+    #:
+    #: .. versionadded:: 8.0
+    command_class: t.Optional[t.Type[Command]] = None
+
+    #: If set, this is used by the group's :meth:`group` decorator
+    #: as the default :class:`Group` class. This is useful to make all
+    #: subgroups use a custom group class.
+    #:
+    #: If set to the special value :class:`type` (literally
+    #: ``group_class = type``), this group's class will be used as the
+    #: default class. This makes a custom group class continue to make
+    #: custom groups.
+    #:
+    #: .. versionadded:: 8.0
+    group_class: t.Optional[t.Union[t.Type["Group"], t.Type[type]]] = None
+    # Literal[type] isn't valid, so use Type[type]
+
+    def __init__(
+        self,
+        name: t.Optional[str] = None,
+        commands: t.Optional[
+            t.Union[t.MutableMapping[str, Command], t.Sequence[Command]]
+        ] = None,
+        **attrs: t.Any,
+    ) -> None:
+        super().__init__(name, **attrs)
+
+        if commands is None:
+            commands = {}
+        elif isinstance(commands, abc.Sequence):
+            commands = {c.name: c for c in commands if c.name is not None}
+
+        #: The registered subcommands by their exported names.
+        self.commands: t.MutableMapping[str, Command] = commands
+
+    def add_command(self, cmd: Command, name: t.Optional[str] = None) -> None:
+        """Registers another :class:`Command` with this group.  If the name
+        is not provided, the name of the command is used.
+        """
+        name = name or cmd.name
+        if name is None:
+            raise TypeError("Command has no name.")
+        _check_multicommand(self, name, cmd, register=True)
+        self.commands[name] = cmd
+
+    @t.overload
+    def command(self, __func: t.Callable[..., t.Any]) -> Command: ...
+
+    @t.overload
+    def command(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Callable[[t.Callable[..., t.Any]], Command]: ...
+
+    def command(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Union[t.Callable[[t.Callable[..., t.Any]], Command], Command]:
+        """A shortcut decorator for declaring and attaching a command to
+        the group. This takes the same arguments as :func:`command` and
+        immediately registers the created command with this group by
+        calling :meth:`add_command`.
+
+        To customize the command class used, set the
+        :attr:`command_class` attribute.
+
+        .. versionchanged:: 8.1
+            This decorator can be applied without parentheses.
+
+        .. versionchanged:: 8.0
+            Added the :attr:`command_class` attribute.
+        """
+        from .decorators import command
+
+        func: t.Optional[t.Callable[..., t.Any]] = None
+
+        if args and callable(args[0]):
+            assert (
+                len(args) == 1 and not kwargs
+            ), "Use 'command(**kwargs)(callable)' to provide arguments."
+            (func,) = args
+            args = ()
+
+        if self.command_class and kwargs.get("cls") is None:
+            kwargs["cls"] = self.command_class
+
+        def decorator(f: t.Callable[..., t.Any]) -> Command:
+            cmd: Command = command(*args, **kwargs)(f)
+            self.add_command(cmd)
+            return cmd
+
+        if func is not None:
+            return decorator(func)
+
+        return decorator
+
+    @t.overload
+    def group(self, __func: t.Callable[..., t.Any]) -> "Group": ...
+
+    @t.overload
+    def group(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Callable[[t.Callable[..., t.Any]], "Group"]: ...
+
+    def group(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Union[t.Callable[[t.Callable[..., t.Any]], "Group"], "Group"]:
+        """A shortcut decorator for declaring and attaching a group to
+        the group. This takes the same arguments as :func:`group` and
+        immediately registers the created group with this group by
+        calling :meth:`add_command`.
+
+        To customize the group class used, set the :attr:`group_class`
+        attribute.
+
+        .. versionchanged:: 8.1
+            This decorator can be applied without parentheses.
+
+        .. versionchanged:: 8.0
+            Added the :attr:`group_class` attribute.
+        """
+        from .decorators import group
+
+        func: t.Optional[t.Callable[..., t.Any]] = None
+
+        if args and callable(args[0]):
+            assert (
+                len(args) == 1 and not kwargs
+            ), "Use 'group(**kwargs)(callable)' to provide arguments."
+            (func,) = args
+            args = ()
+
+        if self.group_class is not None and kwargs.get("cls") is None:
+            if self.group_class is type:
+                kwargs["cls"] = type(self)
+            else:
+                kwargs["cls"] = self.group_class
+
+        def decorator(f: t.Callable[..., t.Any]) -> "Group":
+            cmd: Group = group(*args, **kwargs)(f)
+            self.add_command(cmd)
+            return cmd
+
+        if func is not None:
+            return decorator(func)
+
+        return decorator
+
+    def get_command(self, ctx: Context, cmd_name: str) -> t.Optional[Command]:
+        return self.commands.get(cmd_name)
+
+    def list_commands(self, ctx: Context) -> t.List[str]:
+        return sorted(self.commands)
+
+
+class CommandCollection(MultiCommand):
+    """A command collection is a multi command that merges multiple multi
+    commands together into one.  This is a straightforward implementation
+    that accepts a list of different multi commands as sources and
+    provides all the commands for each of them.
+
+    See :class:`MultiCommand` and :class:`Command` for the description of
+    ``name`` and ``attrs``.
+    """
+
+    def __init__(
+        self,
+        name: t.Optional[str] = None,
+        sources: t.Optional[t.List[MultiCommand]] = None,
+        **attrs: t.Any,
+    ) -> None:
+        super().__init__(name, **attrs)
+        #: The list of registered multi commands.
+        self.sources: t.List[MultiCommand] = sources or []
+
+    def add_source(self, multi_cmd: MultiCommand) -> None:
+        """Adds a new multi command to the chain dispatcher."""
+        self.sources.append(multi_cmd)
+
+    def get_command(self, ctx: Context, cmd_name: str) -> t.Optional[Command]:
+        for source in self.sources:
+            rv = source.get_command(ctx, cmd_name)
+
+            if rv is not None:
+                if self.chain:
+                    _check_multicommand(self, cmd_name, rv)
+
+                return rv
+
+        return None
+
+    def list_commands(self, ctx: Context) -> t.List[str]:
+        rv: t.Set[str] = set()
+
+        for source in self.sources:
+            rv.update(source.list_commands(ctx))
+
+        return sorted(rv)
+
+
+def _check_iter(value: t.Any) -> t.Iterator[t.Any]:
+    """Check if the value is iterable but not a string. Raises a type
+    error, or return an iterator over the value.
+    """
+    if isinstance(value, str):
+        raise TypeError
+
+    return iter(value)
+
+
+class Parameter:
+    r"""A parameter to a command comes in two versions: they are either
+    :class:`Option`\s or :class:`Argument`\s.  Other subclasses are currently
+    not supported by design as some of the internals for parsing are
+    intentionally not finalized.
+
+    Some settings are supported by both options and arguments.
+
+    :param param_decls: the parameter declarations for this option or
+                        argument.  This is a list of flags or argument
+                        names.
+    :param type: the type that should be used.  Either a :class:`ParamType`
+                 or a Python type.  The latter is converted into the former
+                 automatically if supported.
+    :param required: controls if this is optional or not.
+    :param default: the default value if omitted.  This can also be a callable,
+                    in which case it's invoked when the default is needed
+                    without any arguments.
+    :param callback: A function to further process or validate the value
+        after type conversion. It is called as ``f(ctx, param, value)``
+        and must return the value. It is called for all sources,
+        including prompts.
+    :param nargs: the number of arguments to match.  If not ``1`` the return
+                  value is a tuple instead of single value.  The default for
+                  nargs is ``1`` (except if the type is a tuple, then it's
+                  the arity of the tuple). If ``nargs=-1``, all remaining
+                  parameters are collected.
+    :param metavar: how the value is represented in the help page.
+    :param expose_value: if this is `True` then the value is passed onwards
+                         to the command callback and stored on the context,
+                         otherwise it's skipped.
+    :param is_eager: eager values are processed before non eager ones.  This
+                     should not be set for arguments or it will inverse the
+                     order of processing.
+    :param envvar: a string or list of strings that are environment variables
+                   that should be checked.
+    :param shell_complete: A function that returns custom shell
+        completions. Used instead of the param's type completion if
+        given. Takes ``ctx, param, incomplete`` and must return a list
+        of :class:`~click.shell_completion.CompletionItem` or a list of
+        strings.
+
+    .. versionchanged:: 8.0
+        ``process_value`` validates required parameters and bounded
+        ``nargs``, and invokes the parameter callback before returning
+        the value. This allows the callback to validate prompts.
+        ``full_process_value`` is removed.
+
+    .. versionchanged:: 8.0
+        ``autocompletion`` is renamed to ``shell_complete`` and has new
+        semantics described above. The old name is deprecated and will
+        be removed in 8.1, until then it will be wrapped to match the
+        new requirements.
+
+    .. versionchanged:: 8.0
+        For ``multiple=True, nargs>1``, the default must be a list of
+        tuples.
+
+    .. versionchanged:: 8.0
+        Setting a default is no longer required for ``nargs>1``, it will
+        default to ``None``. ``multiple=True`` or ``nargs=-1`` will
+        default to ``()``.
+
+    .. versionchanged:: 7.1
+        Empty environment variables are ignored rather than taking the
+        empty string value. This makes it possible for scripts to clear
+        variables if they can't unset them.
+
+    .. versionchanged:: 2.0
+        Changed signature for parameter callback to also be passed the
+        parameter. The old callback format will still work, but it will
+        raise a warning to give you a chance to migrate the code easier.
+    """
+
+    param_type_name = "parameter"
+
+    def __init__(
+        self,
+        param_decls: t.Optional[t.Sequence[str]] = None,
+        type: t.Optional[t.Union[types.ParamType, t.Any]] = None,
+        required: bool = False,
+        default: t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]] = None,
+        callback: t.Optional[t.Callable[[Context, "Parameter", t.Any], t.Any]] = None,
+        nargs: t.Optional[int] = None,
+        multiple: bool = False,
+        metavar: t.Optional[str] = None,
+        expose_value: bool = True,
+        is_eager: bool = False,
+        envvar: t.Optional[t.Union[str, t.Sequence[str]]] = None,
+        shell_complete: t.Optional[
+            t.Callable[
+                [Context, "Parameter", str],
+                t.Union[t.List["CompletionItem"], t.List[str]],
+            ]
+        ] = None,
+    ) -> None:
+        self.name: t.Optional[str]
+        self.opts: t.List[str]
+        self.secondary_opts: t.List[str]
+        self.name, self.opts, self.secondary_opts = self._parse_decls(
+            param_decls or (), expose_value
+        )
+        self.type: types.ParamType = types.convert_type(type, default)
+
+        # Default nargs to what the type tells us if we have that
+        # information available.
+        if nargs is None:
+            if self.type.is_composite:
+                nargs = self.type.arity
+            else:
+                nargs = 1
+
+        self.required = required
+        self.callback = callback
+        self.nargs = nargs
+        self.multiple = multiple
+        self.expose_value = expose_value
+        self.default = default
+        self.is_eager = is_eager
+        self.metavar = metavar
+        self.envvar = envvar
+        self._custom_shell_complete = shell_complete
+
+        if __debug__:
+            if self.type.is_composite and nargs != self.type.arity:
+                raise ValueError(
+                    f"'nargs' must be {self.type.arity} (or None) for"
+                    f" type {self.type!r}, but it was {nargs}."
+                )
+
+            # Skip no default or callable default.
+            check_default = default if not callable(default) else None
+
+            if check_default is not None:
+                if multiple:
+                    try:
+                        # Only check the first value against nargs.
+                        check_default = next(_check_iter(check_default), None)
+                    except TypeError:
+                        raise ValueError(
+                            "'default' must be a list when 'multiple' is true."
+                        ) from None
+
+                # Can be None for multiple with empty default.
+                if nargs != 1 and check_default is not None:
+                    try:
+                        _check_iter(check_default)
+                    except TypeError:
+                        if multiple:
+                            message = (
+                                "'default' must be a list of lists when 'multiple' is"
+                                " true and 'nargs' != 1."
+                            )
+                        else:
+                            message = "'default' must be a list when 'nargs' != 1."
+
+                        raise ValueError(message) from None
+
+                    if nargs > 1 and len(check_default) != nargs:
+                        subject = "item length" if multiple else "length"
+                        raise ValueError(
+                            f"'default' {subject} must match nargs={nargs}."
+                        )
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        """Gather information that could be useful for a tool generating
+        user-facing documentation.
+
+        Use :meth:`click.Context.to_info_dict` to traverse the entire
+        CLI structure.
+
+        .. versionadded:: 8.0
+        """
+        return {
+            "name": self.name,
+            "param_type_name": self.param_type_name,
+            "opts": self.opts,
+            "secondary_opts": self.secondary_opts,
+            "type": self.type.to_info_dict(),
+            "required": self.required,
+            "nargs": self.nargs,
+            "multiple": self.multiple,
+            "default": self.default,
+            "envvar": self.envvar,
+        }
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__} {self.name}>"
+
+    def _parse_decls(
+        self, decls: t.Sequence[str], expose_value: bool
+    ) -> t.Tuple[t.Optional[str], t.List[str], t.List[str]]:
+        raise NotImplementedError()
+
+    @property
+    def human_readable_name(self) -> str:
+        """Returns the human readable name of this parameter.  This is the
+        same as the name for options, but the metavar for arguments.
+        """
+        return self.name  # type: ignore
+
+    def make_metavar(self) -> str:
+        if self.metavar is not None:
+            return self.metavar
+
+        metavar = self.type.get_metavar(self)
+
+        if metavar is None:
+            metavar = self.type.name.upper()
+
+        if self.nargs != 1:
+            metavar += "..."
+
+        return metavar
+
+    @t.overload
+    def get_default(
+        self, ctx: Context, call: "te.Literal[True]" = True
+    ) -> t.Optional[t.Any]: ...
+
+    @t.overload
+    def get_default(
+        self, ctx: Context, call: bool = ...
+    ) -> t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]]: ...
+
+    def get_default(
+        self, ctx: Context, call: bool = True
+    ) -> t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]]:
+        """Get the default for the parameter. Tries
+        :meth:`Context.lookup_default` first, then the local default.
+
+        :param ctx: Current context.
+        :param call: If the default is a callable, call it. Disable to
+            return the callable instead.
+
+        .. versionchanged:: 8.0.2
+            Type casting is no longer performed when getting a default.
+
+        .. versionchanged:: 8.0.1
+            Type casting can fail in resilient parsing mode. Invalid
+            defaults will not prevent showing help text.
+
+        .. versionchanged:: 8.0
+            Looks at ``ctx.default_map`` first.
+
+        .. versionchanged:: 8.0
+            Added the ``call`` parameter.
+        """
+        value = ctx.lookup_default(self.name, call=False)  # type: ignore
+
+        if value is None:
+            value = self.default
+
+        if call and callable(value):
+            value = value()
+
+        return value
+
+    def add_to_parser(self, parser: OptionParser, ctx: Context) -> None:
+        raise NotImplementedError()
+
+    def consume_value(
+        self, ctx: Context, opts: t.Mapping[str, t.Any]
+    ) -> t.Tuple[t.Any, ParameterSource]:
+        value = opts.get(self.name)  # type: ignore
+        source = ParameterSource.COMMANDLINE
+
+        if value is None:
+            value = self.value_from_envvar(ctx)
+            source = ParameterSource.ENVIRONMENT
+
+        if value is None:
+            value = ctx.lookup_default(self.name)  # type: ignore
+            source = ParameterSource.DEFAULT_MAP
+
+        if value is None:
+            value = self.get_default(ctx)
+            source = ParameterSource.DEFAULT
+
+        return value, source
+
+    def type_cast_value(self, ctx: Context, value: t.Any) -> t.Any:
+        """Convert and validate a value against the option's
+        :attr:`type`, :attr:`multiple`, and :attr:`nargs`.
+        """
+        if value is None:
+            return () if self.multiple or self.nargs == -1 else None
+
+        def check_iter(value: t.Any) -> t.Iterator[t.Any]:
+            try:
+                return _check_iter(value)
+            except TypeError:
+                # This should only happen when passing in args manually,
+                # the parser should construct an iterable when parsing
+                # the command line.
+                raise BadParameter(
+                    _("Value must be an iterable."), ctx=ctx, param=self
+                ) from None
+
+        if self.nargs == 1 or self.type.is_composite:
+
+            def convert(value: t.Any) -> t.Any:
+                return self.type(value, param=self, ctx=ctx)
+
+        elif self.nargs == -1:
+
+            def convert(value: t.Any) -> t.Any:  # t.Tuple[t.Any, ...]
+                return tuple(self.type(x, self, ctx) for x in check_iter(value))
+
+        else:  # nargs > 1
+
+            def convert(value: t.Any) -> t.Any:  # t.Tuple[t.Any, ...]
+                value = tuple(check_iter(value))
+
+                if len(value) != self.nargs:
+                    raise BadParameter(
+                        ngettext(
+                            "Takes {nargs} values but 1 was given.",
+                            "Takes {nargs} values but {len} were given.",
+                            len(value),
+                        ).format(nargs=self.nargs, len=len(value)),
+                        ctx=ctx,
+                        param=self,
+                    )
+
+                return tuple(self.type(x, self, ctx) for x in value)
+
+        if self.multiple:
+            return tuple(convert(x) for x in check_iter(value))
+
+        return convert(value)
+
+    def value_is_missing(self, value: t.Any) -> bool:
+        if value is None:
+            return True
+
+        if (self.nargs != 1 or self.multiple) and value == ():
+            return True
+
+        return False
+
+    def process_value(self, ctx: Context, value: t.Any) -> t.Any:
+        value = self.type_cast_value(ctx, value)
+
+        if self.required and self.value_is_missing(value):
+            raise MissingParameter(ctx=ctx, param=self)
+
+        if self.callback is not None:
+            value = self.callback(ctx, self, value)
+
+        return value
+
+    def resolve_envvar_value(self, ctx: Context) -> t.Optional[str]:
+        if self.envvar is None:
+            return None
+
+        if isinstance(self.envvar, str):
+            rv = os.environ.get(self.envvar)
+
+            if rv:
+                return rv
+        else:
+            for envvar in self.envvar:
+                rv = os.environ.get(envvar)
+
+                if rv:
+                    return rv
+
+        return None
+
+    def value_from_envvar(self, ctx: Context) -> t.Optional[t.Any]:
+        rv: t.Optional[t.Any] = self.resolve_envvar_value(ctx)
+
+        if rv is not None and self.nargs != 1:
+            rv = self.type.split_envvar_value(rv)
+
+        return rv
+
+    def handle_parse_result(
+        self, ctx: Context, opts: t.Mapping[str, t.Any], args: t.List[str]
+    ) -> t.Tuple[t.Any, t.List[str]]:
+        with augment_usage_errors(ctx, param=self):
+            value, source = self.consume_value(ctx, opts)
+            ctx.set_parameter_source(self.name, source)  # type: ignore
+
+            try:
+                value = self.process_value(ctx, value)
+            except Exception:
+                if not ctx.resilient_parsing:
+                    raise
+
+                value = None
+
+        if self.expose_value:
+            ctx.params[self.name] = value  # type: ignore
+
+        return value, args
+
+    def get_help_record(self, ctx: Context) -> t.Optional[t.Tuple[str, str]]:
+        pass
+
+    def get_usage_pieces(self, ctx: Context) -> t.List[str]:
+        return []
+
+    def get_error_hint(self, ctx: Context) -> str:
+        """Get a stringified version of the param for use in error messages to
+        indicate which param caused the error.
+        """
+        hint_list = self.opts or [self.human_readable_name]
+        return " / ".join(f"'{x}'" for x in hint_list)
+
+    def shell_complete(self, ctx: Context, incomplete: str) -> t.List["CompletionItem"]:
+        """Return a list of completions for the incomplete value. If a
+        ``shell_complete`` function was given during init, it is used.
+        Otherwise, the :attr:`type`
+        :meth:`~click.types.ParamType.shell_complete` function is used.
+
+        :param ctx: Invocation context for this command.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        if self._custom_shell_complete is not None:
+            results = self._custom_shell_complete(ctx, self, incomplete)
+
+            if results and isinstance(results[0], str):
+                from click.shell_completion import CompletionItem
+
+                results = [CompletionItem(c) for c in results]
+
+            return t.cast(t.List["CompletionItem"], results)
+
+        return self.type.shell_complete(ctx, self, incomplete)
+
+
+class Option(Parameter):
+    """Options are usually optional values on the command line and
+    have some extra features that arguments don't have.
+
+    All other parameters are passed onwards to the parameter constructor.
+
+    :param show_default: Show the default value for this option in its
+        help text. Values are not shown by default, unless
+        :attr:`Context.show_default` is ``True``. If this value is a
+        string, it shows that string in parentheses instead of the
+        actual value. This is particularly useful for dynamic options.
+        For single option boolean flags, the default remains hidden if
+        its value is ``False``.
+    :param show_envvar: Controls if an environment variable should be
+        shown on the help page. Normally, environment variables are not
+        shown.
+    :param prompt: If set to ``True`` or a non empty string then the
+        user will be prompted for input. If set to ``True`` the prompt
+        will be the option name capitalized.
+    :param confirmation_prompt: Prompt a second time to confirm the
+        value if it was prompted for. Can be set to a string instead of
+        ``True`` to customize the message.
+    :param prompt_required: If set to ``False``, the user will be
+        prompted for input only when the option was specified as a flag
+        without a value.
+    :param hide_input: If this is ``True`` then the input on the prompt
+        will be hidden from the user. This is useful for password input.
+    :param is_flag: forces this option to act as a flag.  The default is
+                    auto detection.
+    :param flag_value: which value should be used for this flag if it's
+                       enabled.  This is set to a boolean automatically if
+                       the option string contains a slash to mark two options.
+    :param multiple: if this is set to `True` then the argument is accepted
+                     multiple times and recorded.  This is similar to ``nargs``
+                     in how it works but supports arbitrary number of
+                     arguments.
+    :param count: this flag makes an option increment an integer.
+    :param allow_from_autoenv: if this is enabled then the value of this
+                               parameter will be pulled from an environment
+                               variable in case a prefix is defined on the
+                               context.
+    :param help: the help string.
+    :param hidden: hide this option from help outputs.
+    :param attrs: Other command arguments described in :class:`Parameter`.
+
+    .. versionchanged:: 8.1.0
+        Help text indentation is cleaned here instead of only in the
+        ``@option`` decorator.
+
+    .. versionchanged:: 8.1.0
+        The ``show_default`` parameter overrides
+        ``Context.show_default``.
+
+    .. versionchanged:: 8.1.0
+        The default of a single option boolean flag is not shown if the
+        default value is ``False``.
+
+    .. versionchanged:: 8.0.1
+        ``type`` is detected from ``flag_value`` if given.
+    """
+
+    param_type_name = "option"
+
+    def __init__(
+        self,
+        param_decls: t.Optional[t.Sequence[str]] = None,
+        show_default: t.Union[bool, str, None] = None,
+        prompt: t.Union[bool, str] = False,
+        confirmation_prompt: t.Union[bool, str] = False,
+        prompt_required: bool = True,
+        hide_input: bool = False,
+        is_flag: t.Optional[bool] = None,
+        flag_value: t.Optional[t.Any] = None,
+        multiple: bool = False,
+        count: bool = False,
+        allow_from_autoenv: bool = True,
+        type: t.Optional[t.Union[types.ParamType, t.Any]] = None,
+        help: t.Optional[str] = None,
+        hidden: bool = False,
+        show_choices: bool = True,
+        show_envvar: bool = False,
+        **attrs: t.Any,
+    ) -> None:
+        if help:
+            help = inspect.cleandoc(help)
+
+        default_is_missing = "default" not in attrs
+        super().__init__(param_decls, type=type, multiple=multiple, **attrs)
+
+        if prompt is True:
+            if self.name is None:
+                raise TypeError("'name' is required with 'prompt=True'.")
+
+            prompt_text: t.Optional[str] = self.name.replace("_", " ").capitalize()
+        elif prompt is False:
+            prompt_text = None
+        else:
+            prompt_text = prompt
+
+        self.prompt = prompt_text
+        self.confirmation_prompt = confirmation_prompt
+        self.prompt_required = prompt_required
+        self.hide_input = hide_input
+        self.hidden = hidden
+
+        # If prompt is enabled but not required, then the option can be
+        # used as a flag to indicate using prompt or flag_value.
+        self._flag_needs_value = self.prompt is not None and not self.prompt_required
+
+        if is_flag is None:
+            if flag_value is not None:
+                # Implicitly a flag because flag_value was set.
+                is_flag = True
+            elif self._flag_needs_value:
+                # Not a flag, but when used as a flag it shows a prompt.
+                is_flag = False
+            else:
+                # Implicitly a flag because flag options were given.
+                is_flag = bool(self.secondary_opts)
+        elif is_flag is False and not self._flag_needs_value:
+            # Not a flag, and prompt is not enabled, can be used as a
+            # flag if flag_value is set.
+            self._flag_needs_value = flag_value is not None
+
+        self.default: t.Union[t.Any, t.Callable[[], t.Any]]
+
+        if is_flag and default_is_missing and not self.required:
+            if multiple:
+                self.default = ()
+            else:
+                self.default = False
+
+        if flag_value is None:
+            flag_value = not self.default
+
+        self.type: types.ParamType
+        if is_flag and type is None:
+            # Re-guess the type from the flag value instead of the
+            # default.
+            self.type = types.convert_type(None, flag_value)
+
+        self.is_flag: bool = is_flag
+        self.is_bool_flag: bool = is_flag and isinstance(self.type, types.BoolParamType)
+        self.flag_value: t.Any = flag_value
+
+        # Counting
+        self.count = count
+        if count:
+            if type is None:
+                self.type = types.IntRange(min=0)
+            if default_is_missing:
+                self.default = 0
+
+        self.allow_from_autoenv = allow_from_autoenv
+        self.help = help
+        self.show_default = show_default
+        self.show_choices = show_choices
+        self.show_envvar = show_envvar
+
+        if __debug__:
+            if self.nargs == -1:
+                raise TypeError("nargs=-1 is not supported for options.")
+
+            if self.prompt and self.is_flag and not self.is_bool_flag:
+                raise TypeError("'prompt' is not valid for non-boolean flag.")
+
+            if not self.is_bool_flag and self.secondary_opts:
+                raise TypeError("Secondary flag is not valid for non-boolean flag.")
+
+            if self.is_bool_flag and self.hide_input and self.prompt is not None:
+                raise TypeError(
+                    "'prompt' with 'hide_input' is not valid for boolean flag."
+                )
+
+            if self.count:
+                if self.multiple:
+                    raise TypeError("'count' is not valid with 'multiple'.")
+
+                if self.is_flag:
+                    raise TypeError("'count' is not valid with 'is_flag'.")
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict.update(
+            help=self.help,
+            prompt=self.prompt,
+            is_flag=self.is_flag,
+            flag_value=self.flag_value,
+            count=self.count,
+            hidden=self.hidden,
+        )
+        return info_dict
+
+    def _parse_decls(
+        self, decls: t.Sequence[str], expose_value: bool
+    ) -> t.Tuple[t.Optional[str], t.List[str], t.List[str]]:
+        opts = []
+        secondary_opts = []
+        name = None
+        possible_names = []
+
+        for decl in decls:
+            if decl.isidentifier():
+                if name is not None:
+                    raise TypeError(f"Name '{name}' defined twice")
+                name = decl
+            else:
+                split_char = ";" if decl[:1] == "/" else "/"
+                if split_char in decl:
+                    first, second = decl.split(split_char, 1)
+                    first = first.rstrip()
+                    if first:
+                        possible_names.append(split_opt(first))
+                        opts.append(first)
+                    second = second.lstrip()
+                    if second:
+                        secondary_opts.append(second.lstrip())
+                    if first == second:
+                        raise ValueError(
+                            f"Boolean option {decl!r} cannot use the"
+                            " same flag for true/false."
+                        )
+                else:
+                    possible_names.append(split_opt(decl))
+                    opts.append(decl)
+
+        if name is None and possible_names:
+            possible_names.sort(key=lambda x: -len(x[0]))  # group long options first
+            name = possible_names[0][1].replace("-", "_").lower()
+            if not name.isidentifier():
+                name = None
+
+        if name is None:
+            if not expose_value:
+                return None, opts, secondary_opts
+            raise TypeError(
+                f"Could not determine name for option with declarations {decls!r}"
+            )
+
+        if not opts and not secondary_opts:
+            raise TypeError(
+                f"No options defined but a name was passed ({name})."
+                " Did you mean to declare an argument instead? Did"
+                f" you mean to pass '--{name}'?"
+            )
+
+        return name, opts, secondary_opts
+
+    def add_to_parser(self, parser: OptionParser, ctx: Context) -> None:
+        if self.multiple:
+            action = "append"
+        elif self.count:
+            action = "count"
+        else:
+            action = "store"
+
+        if self.is_flag:
+            action = f"{action}_const"
+
+            if self.is_bool_flag and self.secondary_opts:
+                parser.add_option(
+                    obj=self, opts=self.opts, dest=self.name, action=action, const=True
+                )
+                parser.add_option(
+                    obj=self,
+                    opts=self.secondary_opts,
+                    dest=self.name,
+                    action=action,
+                    const=False,
+                )
+            else:
+                parser.add_option(
+                    obj=self,
+                    opts=self.opts,
+                    dest=self.name,
+                    action=action,
+                    const=self.flag_value,
+                )
+        else:
+            parser.add_option(
+                obj=self,
+                opts=self.opts,
+                dest=self.name,
+                action=action,
+                nargs=self.nargs,
+            )
+
+    def get_help_record(self, ctx: Context) -> t.Optional[t.Tuple[str, str]]:
+        if self.hidden:
+            return None
+
+        any_prefix_is_slash = False
+
+        def _write_opts(opts: t.Sequence[str]) -> str:
+            nonlocal any_prefix_is_slash
+
+            rv, any_slashes = join_options(opts)
+
+            if any_slashes:
+                any_prefix_is_slash = True
+
+            if not self.is_flag and not self.count:
+                rv += f" {self.make_metavar()}"
+
+            return rv
+
+        rv = [_write_opts(self.opts)]
+
+        if self.secondary_opts:
+            rv.append(_write_opts(self.secondary_opts))
+
+        help = self.help or ""
+        extra = []
+
+        if self.show_envvar:
+            envvar = self.envvar
+
+            if envvar is None:
+                if (
+                    self.allow_from_autoenv
+                    and ctx.auto_envvar_prefix is not None
+                    and self.name is not None
+                ):
+                    envvar = f"{ctx.auto_envvar_prefix}_{self.name.upper()}"
+
+            if envvar is not None:
+                var_str = (
+                    envvar
+                    if isinstance(envvar, str)
+                    else ", ".join(str(d) for d in envvar)
+                )
+                extra.append(_("env var: {var}").format(var=var_str))
+
+        # Temporarily enable resilient parsing to avoid type casting
+        # failing for the default. Might be possible to extend this to
+        # help formatting in general.
+        resilient = ctx.resilient_parsing
+        ctx.resilient_parsing = True
+
+        try:
+            default_value = self.get_default(ctx, call=False)
+        finally:
+            ctx.resilient_parsing = resilient
+
+        show_default = False
+        show_default_is_str = False
+
+        if self.show_default is not None:
+            if isinstance(self.show_default, str):
+                show_default_is_str = show_default = True
+            else:
+                show_default = self.show_default
+        elif ctx.show_default is not None:
+            show_default = ctx.show_default
+
+        if show_default_is_str or (show_default and (default_value is not None)):
+            if show_default_is_str:
+                default_string = f"({self.show_default})"
+            elif isinstance(default_value, (list, tuple)):
+                default_string = ", ".join(str(d) for d in default_value)
+            elif inspect.isfunction(default_value):
+                default_string = _("(dynamic)")
+            elif self.is_bool_flag and self.secondary_opts:
+                # For boolean flags that have distinct True/False opts,
+                # use the opt without prefix instead of the value.
+                default_string = split_opt(
+                    (self.opts if default_value else self.secondary_opts)[0]
+                )[1]
+            elif self.is_bool_flag and not self.secondary_opts and not default_value:
+                default_string = ""
+            elif default_value == "":
+                default_string = '""'
+            else:
+                default_string = str(default_value)
+
+            if default_string:
+                extra.append(_("default: {default}").format(default=default_string))
+
+        if (
+            isinstance(self.type, types._NumberRangeBase)
+            # skip count with default range type
+            and not (self.count and self.type.min == 0 and self.type.max is None)
+        ):
+            range_str = self.type._describe_range()
+
+            if range_str:
+                extra.append(range_str)
+
+        if self.required:
+            extra.append(_("required"))
+
+        if extra:
+            extra_str = "; ".join(extra)
+            help = f"{help}  [{extra_str}]" if help else f"[{extra_str}]"
+
+        return ("; " if any_prefix_is_slash else " / ").join(rv), help
+
+    @t.overload
+    def get_default(
+        self, ctx: Context, call: "te.Literal[True]" = True
+    ) -> t.Optional[t.Any]: ...
+
+    @t.overload
+    def get_default(
+        self, ctx: Context, call: bool = ...
+    ) -> t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]]: ...
+
+    def get_default(
+        self, ctx: Context, call: bool = True
+    ) -> t.Optional[t.Union[t.Any, t.Callable[[], t.Any]]]:
+        # If we're a non boolean flag our default is more complex because
+        # we need to look at all flags in the same group to figure out
+        # if we're the default one in which case we return the flag
+        # value as default.
+        if self.is_flag and not self.is_bool_flag:
+            for param in ctx.command.params:
+                if param.name == self.name and param.default:
+                    return t.cast(Option, param).flag_value
+
+            return None
+
+        return super().get_default(ctx, call=call)
+
+    def prompt_for_value(self, ctx: Context) -> t.Any:
+        """This is an alternative flow that can be activated in the full
+        value processing if a value does not exist.  It will prompt the
+        user until a valid value exists and then returns the processed
+        value as result.
+        """
+        assert self.prompt is not None
+
+        # Calculate the default before prompting anything to be stable.
+        default = self.get_default(ctx)
+
+        # If this is a prompt for a flag we need to handle this
+        # differently.
+        if self.is_bool_flag:
+            return confirm(self.prompt, default)
+
+        return prompt(
+            self.prompt,
+            default=default,
+            type=self.type,
+            hide_input=self.hide_input,
+            show_choices=self.show_choices,
+            confirmation_prompt=self.confirmation_prompt,
+            value_proc=lambda x: self.process_value(ctx, x),
+        )
+
+    def resolve_envvar_value(self, ctx: Context) -> t.Optional[str]:
+        rv = super().resolve_envvar_value(ctx)
+
+        if rv is not None:
+            return rv
+
+        if (
+            self.allow_from_autoenv
+            and ctx.auto_envvar_prefix is not None
+            and self.name is not None
+        ):
+            envvar = f"{ctx.auto_envvar_prefix}_{self.name.upper()}"
+            rv = os.environ.get(envvar)
+
+            if rv:
+                return rv
+
+        return None
+
+    def value_from_envvar(self, ctx: Context) -> t.Optional[t.Any]:
+        rv: t.Optional[t.Any] = self.resolve_envvar_value(ctx)
+
+        if rv is None:
+            return None
+
+        value_depth = (self.nargs != 1) + bool(self.multiple)
+
+        if value_depth > 0:
+            rv = self.type.split_envvar_value(rv)
+
+            if self.multiple and self.nargs != 1:
+                rv = batch(rv, self.nargs)
+
+        return rv
+
+    def consume_value(
+        self, ctx: Context, opts: t.Mapping[str, "Parameter"]
+    ) -> t.Tuple[t.Any, ParameterSource]:
+        value, source = super().consume_value(ctx, opts)
+
+        # The parser will emit a sentinel value if the option can be
+        # given as a flag without a value. This is different from None
+        # to distinguish from the flag not being given at all.
+        if value is _flag_needs_value:
+            if self.prompt is not None and not ctx.resilient_parsing:
+                value = self.prompt_for_value(ctx)
+                source = ParameterSource.PROMPT
+            else:
+                value = self.flag_value
+                source = ParameterSource.COMMANDLINE
+
+        elif (
+            self.multiple
+            and value is not None
+            and any(v is _flag_needs_value for v in value)
+        ):
+            value = [self.flag_value if v is _flag_needs_value else v for v in value]
+            source = ParameterSource.COMMANDLINE
+
+        # The value wasn't set, or used the param's default, prompt if
+        # prompting is enabled.
+        elif (
+            source in {None, ParameterSource.DEFAULT}
+            and self.prompt is not None
+            and (self.required or self.prompt_required)
+            and not ctx.resilient_parsing
+        ):
+            value = self.prompt_for_value(ctx)
+            source = ParameterSource.PROMPT
+
+        return value, source
+
+
+class Argument(Parameter):
+    """Arguments are positional parameters to a command.  They generally
+    provide fewer features than options but can have infinite ``nargs``
+    and are required by default.
+
+    All parameters are passed onwards to the constructor of :class:`Parameter`.
+    """
+
+    param_type_name = "argument"
+
+    def __init__(
+        self,
+        param_decls: t.Sequence[str],
+        required: t.Optional[bool] = None,
+        **attrs: t.Any,
+    ) -> None:
+        if required is None:
+            if attrs.get("default") is not None:
+                required = False
+            else:
+                required = attrs.get("nargs", 1) > 0
+
+        if "multiple" in attrs:
+            raise TypeError("__init__() got an unexpected keyword argument 'multiple'.")
+
+        super().__init__(param_decls, required=required, **attrs)
+
+        if __debug__:
+            if self.default is not None and self.nargs == -1:
+                raise TypeError("'default' is not supported for nargs=-1.")
+
+    @property
+    def human_readable_name(self) -> str:
+        if self.metavar is not None:
+            return self.metavar
+        return self.name.upper()  # type: ignore
+
+    def make_metavar(self) -> str:
+        if self.metavar is not None:
+            return self.metavar
+        var = self.type.get_metavar(self)
+        if not var:
+            var = self.name.upper()  # type: ignore
+        if not self.required:
+            var = f"[{var}]"
+        if self.nargs != 1:
+            var += "..."
+        return var
+
+    def _parse_decls(
+        self, decls: t.Sequence[str], expose_value: bool
+    ) -> t.Tuple[t.Optional[str], t.List[str], t.List[str]]:
+        if not decls:
+            if not expose_value:
+                return None, [], []
+            raise TypeError("Argument is marked as exposed, but does not have a name.")
+        if len(decls) == 1:
+            name = arg = decls[0]
+            name = name.replace("-", "_").lower()
+        else:
+            raise TypeError(
+                "Arguments take exactly one parameter declaration, got"
+                f" {len(decls)}."
+            )
+        return name, [arg], []
+
+    def get_usage_pieces(self, ctx: Context) -> t.List[str]:
+        return [self.make_metavar()]
+
+    def get_error_hint(self, ctx: Context) -> str:
+        return f"'{self.make_metavar()}'"
+
+    def add_to_parser(self, parser: OptionParser, ctx: Context) -> None:
+        parser.add_argument(dest=self.name, nargs=self.nargs, obj=self)
diff --git a/backend/venv/lib/python3.12/site-packages/click/decorators.py b/backend/venv/lib/python3.12/site-packages/click/decorators.py
new file mode 100644
index 00000000..bcf8906e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/decorators.py
@@ -0,0 +1,562 @@
+import inspect
+import types
+import typing as t
+from functools import update_wrapper
+from gettext import gettext as _
+
+from .core import Argument
+from .core import Command
+from .core import Context
+from .core import Group
+from .core import Option
+from .core import Parameter
+from .globals import get_current_context
+from .utils import echo
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    P = te.ParamSpec("P")
+
+R = t.TypeVar("R")
+T = t.TypeVar("T")
+_AnyCallable = t.Callable[..., t.Any]
+FC = t.TypeVar("FC", bound=t.Union[_AnyCallable, Command])
+
+
+def pass_context(f: "t.Callable[te.Concatenate[Context, P], R]") -> "t.Callable[P, R]":
+    """Marks a callback as wanting to receive the current context
+    object as first argument.
+    """
+
+    def new_func(*args: "P.args", **kwargs: "P.kwargs") -> "R":
+        return f(get_current_context(), *args, **kwargs)
+
+    return update_wrapper(new_func, f)
+
+
+def pass_obj(f: "t.Callable[te.Concatenate[t.Any, P], R]") -> "t.Callable[P, R]":
+    """Similar to :func:`pass_context`, but only pass the object on the
+    context onwards (:attr:`Context.obj`).  This is useful if that object
+    represents the state of a nested system.
+    """
+
+    def new_func(*args: "P.args", **kwargs: "P.kwargs") -> "R":
+        return f(get_current_context().obj, *args, **kwargs)
+
+    return update_wrapper(new_func, f)
+
+
+def make_pass_decorator(
+    object_type: t.Type[T], ensure: bool = False
+) -> t.Callable[["t.Callable[te.Concatenate[T, P], R]"], "t.Callable[P, R]"]:
+    """Given an object type this creates a decorator that will work
+    similar to :func:`pass_obj` but instead of passing the object of the
+    current context, it will find the innermost context of type
+    :func:`object_type`.
+
+    This generates a decorator that works roughly like this::
+
+        from functools import update_wrapper
+
+        def decorator(f):
+            @pass_context
+            def new_func(ctx, *args, **kwargs):
+                obj = ctx.find_object(object_type)
+                return ctx.invoke(f, obj, *args, **kwargs)
+            return update_wrapper(new_func, f)
+        return decorator
+
+    :param object_type: the type of the object to pass.
+    :param ensure: if set to `True`, a new object will be created and
+                   remembered on the context if it's not there yet.
+    """
+
+    def decorator(f: "t.Callable[te.Concatenate[T, P], R]") -> "t.Callable[P, R]":
+        def new_func(*args: "P.args", **kwargs: "P.kwargs") -> "R":
+            ctx = get_current_context()
+
+            obj: t.Optional[T]
+            if ensure:
+                obj = ctx.ensure_object(object_type)
+            else:
+                obj = ctx.find_object(object_type)
+
+            if obj is None:
+                raise RuntimeError(
+                    "Managed to invoke callback without a context"
+                    f" object of type {object_type.__name__!r}"
+                    " existing."
+                )
+
+            return ctx.invoke(f, obj, *args, **kwargs)
+
+        return update_wrapper(new_func, f)
+
+    return decorator
+
+
+def pass_meta_key(
+    key: str, *, doc_description: t.Optional[str] = None
+) -> "t.Callable[[t.Callable[te.Concatenate[t.Any, P], R]], t.Callable[P, R]]":
+    """Create a decorator that passes a key from
+    :attr:`click.Context.meta` as the first argument to the decorated
+    function.
+
+    :param key: Key in ``Context.meta`` to pass.
+    :param doc_description: Description of the object being passed,
+        inserted into the decorator's docstring. Defaults to "the 'key'
+        key from Context.meta".
+
+    .. versionadded:: 8.0
+    """
+
+    def decorator(f: "t.Callable[te.Concatenate[t.Any, P], R]") -> "t.Callable[P, R]":
+        def new_func(*args: "P.args", **kwargs: "P.kwargs") -> R:
+            ctx = get_current_context()
+            obj = ctx.meta[key]
+            return ctx.invoke(f, obj, *args, **kwargs)
+
+        return update_wrapper(new_func, f)
+
+    if doc_description is None:
+        doc_description = f"the {key!r} key from :attr:`click.Context.meta`"
+
+    decorator.__doc__ = (
+        f"Decorator that passes {doc_description} as the first argument"
+        " to the decorated function."
+    )
+    return decorator
+
+
+CmdType = t.TypeVar("CmdType", bound=Command)
+
+
+# variant: no call, directly as decorator for a function.
+@t.overload
+def command(name: _AnyCallable) -> Command: ...
+
+
+# variant: with positional name and with positional or keyword cls argument:
+# @command(namearg, CommandCls, ...) or @command(namearg, cls=CommandCls, ...)
+@t.overload
+def command(
+    name: t.Optional[str],
+    cls: t.Type[CmdType],
+    **attrs: t.Any,
+) -> t.Callable[[_AnyCallable], CmdType]: ...
+
+
+# variant: name omitted, cls _must_ be a keyword argument, @command(cls=CommandCls, ...)
+@t.overload
+def command(
+    name: None = None,
+    *,
+    cls: t.Type[CmdType],
+    **attrs: t.Any,
+) -> t.Callable[[_AnyCallable], CmdType]: ...
+
+
+# variant: with optional string name, no cls argument provided.
+@t.overload
+def command(
+    name: t.Optional[str] = ..., cls: None = None, **attrs: t.Any
+) -> t.Callable[[_AnyCallable], Command]: ...
+
+
+def command(
+    name: t.Union[t.Optional[str], _AnyCallable] = None,
+    cls: t.Optional[t.Type[CmdType]] = None,
+    **attrs: t.Any,
+) -> t.Union[Command, t.Callable[[_AnyCallable], t.Union[Command, CmdType]]]:
+    r"""Creates a new :class:`Command` and uses the decorated function as
+    callback.  This will also automatically attach all decorated
+    :func:`option`\s and :func:`argument`\s as parameters to the command.
+
+    The name of the command defaults to the name of the function with
+    underscores replaced by dashes.  If you want to change that, you can
+    pass the intended name as the first argument.
+
+    All keyword arguments are forwarded to the underlying command class.
+    For the ``params`` argument, any decorated params are appended to
+    the end of the list.
+
+    Once decorated the function turns into a :class:`Command` instance
+    that can be invoked as a command line utility or be attached to a
+    command :class:`Group`.
+
+    :param name: the name of the command.  This defaults to the function
+                 name with underscores replaced by dashes.
+    :param cls: the command class to instantiate.  This defaults to
+                :class:`Command`.
+
+    .. versionchanged:: 8.1
+        This decorator can be applied without parentheses.
+
+    .. versionchanged:: 8.1
+        The ``params`` argument can be used. Decorated params are
+        appended to the end of the list.
+    """
+
+    func: t.Optional[t.Callable[[_AnyCallable], t.Any]] = None
+
+    if callable(name):
+        func = name
+        name = None
+        assert cls is None, "Use 'command(cls=cls)(callable)' to specify a class."
+        assert not attrs, "Use 'command(**kwargs)(callable)' to provide arguments."
+
+    if cls is None:
+        cls = t.cast(t.Type[CmdType], Command)
+
+    def decorator(f: _AnyCallable) -> CmdType:
+        if isinstance(f, Command):
+            raise TypeError("Attempted to convert a callback into a command twice.")
+
+        attr_params = attrs.pop("params", None)
+        params = attr_params if attr_params is not None else []
+
+        try:
+            decorator_params = f.__click_params__  # type: ignore
+        except AttributeError:
+            pass
+        else:
+            del f.__click_params__  # type: ignore
+            params.extend(reversed(decorator_params))
+
+        if attrs.get("help") is None:
+            attrs["help"] = f.__doc__
+
+        if t.TYPE_CHECKING:
+            assert cls is not None
+            assert not callable(name)
+
+        cmd = cls(
+            name=name or f.__name__.lower().replace("_", "-"),
+            callback=f,
+            params=params,
+            **attrs,
+        )
+        cmd.__doc__ = f.__doc__
+        return cmd
+
+    if func is not None:
+        return decorator(func)
+
+    return decorator
+
+
+GrpType = t.TypeVar("GrpType", bound=Group)
+
+
+# variant: no call, directly as decorator for a function.
+@t.overload
+def group(name: _AnyCallable) -> Group: ...
+
+
+# variant: with positional name and with positional or keyword cls argument:
+# @group(namearg, GroupCls, ...) or @group(namearg, cls=GroupCls, ...)
+@t.overload
+def group(
+    name: t.Optional[str],
+    cls: t.Type[GrpType],
+    **attrs: t.Any,
+) -> t.Callable[[_AnyCallable], GrpType]: ...
+
+
+# variant: name omitted, cls _must_ be a keyword argument, @group(cmd=GroupCls, ...)
+@t.overload
+def group(
+    name: None = None,
+    *,
+    cls: t.Type[GrpType],
+    **attrs: t.Any,
+) -> t.Callable[[_AnyCallable], GrpType]: ...
+
+
+# variant: with optional string name, no cls argument provided.
+@t.overload
+def group(
+    name: t.Optional[str] = ..., cls: None = None, **attrs: t.Any
+) -> t.Callable[[_AnyCallable], Group]: ...
+
+
+def group(
+    name: t.Union[str, _AnyCallable, None] = None,
+    cls: t.Optional[t.Type[GrpType]] = None,
+    **attrs: t.Any,
+) -> t.Union[Group, t.Callable[[_AnyCallable], t.Union[Group, GrpType]]]:
+    """Creates a new :class:`Group` with a function as callback.  This
+    works otherwise the same as :func:`command` just that the `cls`
+    parameter is set to :class:`Group`.
+
+    .. versionchanged:: 8.1
+        This decorator can be applied without parentheses.
+    """
+    if cls is None:
+        cls = t.cast(t.Type[GrpType], Group)
+
+    if callable(name):
+        return command(cls=cls, **attrs)(name)
+
+    return command(name, cls, **attrs)
+
+
+def _param_memo(f: t.Callable[..., t.Any], param: Parameter) -> None:
+    if isinstance(f, Command):
+        f.params.append(param)
+    else:
+        if not hasattr(f, "__click_params__"):
+            f.__click_params__ = []  # type: ignore
+
+        f.__click_params__.append(param)  # type: ignore
+
+
+def argument(
+    *param_decls: str, cls: t.Optional[t.Type[Argument]] = None, **attrs: t.Any
+) -> t.Callable[[FC], FC]:
+    """Attaches an argument to the command.  All positional arguments are
+    passed as parameter declarations to :class:`Argument`; all keyword
+    arguments are forwarded unchanged (except ``cls``).
+    This is equivalent to creating an :class:`Argument` instance manually
+    and attaching it to the :attr:`Command.params` list.
+
+    For the default argument class, refer to :class:`Argument` and
+    :class:`Parameter` for descriptions of parameters.
+
+    :param cls: the argument class to instantiate.  This defaults to
+                :class:`Argument`.
+    :param param_decls: Passed as positional arguments to the constructor of
+        ``cls``.
+    :param attrs: Passed as keyword arguments to the constructor of ``cls``.
+    """
+    if cls is None:
+        cls = Argument
+
+    def decorator(f: FC) -> FC:
+        _param_memo(f, cls(param_decls, **attrs))
+        return f
+
+    return decorator
+
+
+def option(
+    *param_decls: str, cls: t.Optional[t.Type[Option]] = None, **attrs: t.Any
+) -> t.Callable[[FC], FC]:
+    """Attaches an option to the command.  All positional arguments are
+    passed as parameter declarations to :class:`Option`; all keyword
+    arguments are forwarded unchanged (except ``cls``).
+    This is equivalent to creating an :class:`Option` instance manually
+    and attaching it to the :attr:`Command.params` list.
+
+    For the default option class, refer to :class:`Option` and
+    :class:`Parameter` for descriptions of parameters.
+
+    :param cls: the option class to instantiate.  This defaults to
+                :class:`Option`.
+    :param param_decls: Passed as positional arguments to the constructor of
+        ``cls``.
+    :param attrs: Passed as keyword arguments to the constructor of ``cls``.
+    """
+    if cls is None:
+        cls = Option
+
+    def decorator(f: FC) -> FC:
+        _param_memo(f, cls(param_decls, **attrs))
+        return f
+
+    return decorator
+
+
+def confirmation_option(*param_decls: str, **kwargs: t.Any) -> t.Callable[[FC], FC]:
+    """Add a ``--yes`` option which shows a prompt before continuing if
+    not passed. If the prompt is declined, the program will exit.
+
+    :param param_decls: One or more option names. Defaults to the single
+        value ``"--yes"``.
+    :param kwargs: Extra arguments are passed to :func:`option`.
+    """
+
+    def callback(ctx: Context, param: Parameter, value: bool) -> None:
+        if not value:
+            ctx.abort()
+
+    if not param_decls:
+        param_decls = ("--yes",)
+
+    kwargs.setdefault("is_flag", True)
+    kwargs.setdefault("callback", callback)
+    kwargs.setdefault("expose_value", False)
+    kwargs.setdefault("prompt", "Do you want to continue?")
+    kwargs.setdefault("help", "Confirm the action without prompting.")
+    return option(*param_decls, **kwargs)
+
+
+def password_option(*param_decls: str, **kwargs: t.Any) -> t.Callable[[FC], FC]:
+    """Add a ``--password`` option which prompts for a password, hiding
+    input and asking to enter the value again for confirmation.
+
+    :param param_decls: One or more option names. Defaults to the single
+        value ``"--password"``.
+    :param kwargs: Extra arguments are passed to :func:`option`.
+    """
+    if not param_decls:
+        param_decls = ("--password",)
+
+    kwargs.setdefault("prompt", True)
+    kwargs.setdefault("confirmation_prompt", True)
+    kwargs.setdefault("hide_input", True)
+    return option(*param_decls, **kwargs)
+
+
+def version_option(
+    version: t.Optional[str] = None,
+    *param_decls: str,
+    package_name: t.Optional[str] = None,
+    prog_name: t.Optional[str] = None,
+    message: t.Optional[str] = None,
+    **kwargs: t.Any,
+) -> t.Callable[[FC], FC]:
+    """Add a ``--version`` option which immediately prints the version
+    number and exits the program.
+
+    If ``version`` is not provided, Click will try to detect it using
+    :func:`importlib.metadata.version` to get the version for the
+    ``package_name``. On Python < 3.8, the ``importlib_metadata``
+    backport must be installed.
+
+    If ``package_name`` is not provided, Click will try to detect it by
+    inspecting the stack frames. This will be used to detect the
+    version, so it must match the name of the installed package.
+
+    :param version: The version number to show. If not provided, Click
+        will try to detect it.
+    :param param_decls: One or more option names. Defaults to the single
+        value ``"--version"``.
+    :param package_name: The package name to detect the version from. If
+        not provided, Click will try to detect it.
+    :param prog_name: The name of the CLI to show in the message. If not
+        provided, it will be detected from the command.
+    :param message: The message to show. The values ``%(prog)s``,
+        ``%(package)s``, and ``%(version)s`` are available. Defaults to
+        ``"%(prog)s, version %(version)s"``.
+    :param kwargs: Extra arguments are passed to :func:`option`.
+    :raise RuntimeError: ``version`` could not be detected.
+
+    .. versionchanged:: 8.0
+        Add the ``package_name`` parameter, and the ``%(package)s``
+        value for messages.
+
+    .. versionchanged:: 8.0
+        Use :mod:`importlib.metadata` instead of ``pkg_resources``. The
+        version is detected based on the package name, not the entry
+        point name. The Python package name must match the installed
+        package name, or be passed with ``package_name=``.
+    """
+    if message is None:
+        message = _("%(prog)s, version %(version)s")
+
+    if version is None and package_name is None:
+        frame = inspect.currentframe()
+        f_back = frame.f_back if frame is not None else None
+        f_globals = f_back.f_globals if f_back is not None else None
+        # break reference cycle
+        # https://docs.python.org/3/library/inspect.html#the-interpreter-stack
+        del frame
+
+        if f_globals is not None:
+            package_name = f_globals.get("__name__")
+
+            if package_name == "__main__":
+                package_name = f_globals.get("__package__")
+
+            if package_name:
+                package_name = package_name.partition(".")[0]
+
+    def callback(ctx: Context, param: Parameter, value: bool) -> None:
+        if not value or ctx.resilient_parsing:
+            return
+
+        nonlocal prog_name
+        nonlocal version
+
+        if prog_name is None:
+            prog_name = ctx.find_root().info_name
+
+        if version is None and package_name is not None:
+            metadata: t.Optional[types.ModuleType]
+
+            try:
+                from importlib import metadata
+            except ImportError:
+                # Python < 3.8
+                import importlib_metadata as metadata  # type: ignore
+
+            try:
+                version = metadata.version(package_name)  # type: ignore
+            except metadata.PackageNotFoundError:  # type: ignore
+                raise RuntimeError(
+                    f"{package_name!r} is not installed. Try passing"
+                    " 'package_name' instead."
+                ) from None
+
+        if version is None:
+            raise RuntimeError(
+                f"Could not determine the version for {package_name!r} automatically."
+            )
+
+        echo(
+            message % {"prog": prog_name, "package": package_name, "version": version},
+            color=ctx.color,
+        )
+        ctx.exit()
+
+    if not param_decls:
+        param_decls = ("--version",)
+
+    kwargs.setdefault("is_flag", True)
+    kwargs.setdefault("expose_value", False)
+    kwargs.setdefault("is_eager", True)
+    kwargs.setdefault("help", _("Show the version and exit."))
+    kwargs["callback"] = callback
+    return option(*param_decls, **kwargs)
+
+
+class HelpOption(Option):
+    """Pre-configured ``--help`` option which immediately prints the help page
+    and exits the program.
+    """
+
+    def __init__(
+        self,
+        param_decls: t.Optional[t.Sequence[str]] = None,
+        **kwargs: t.Any,
+    ) -> None:
+        if not param_decls:
+            param_decls = ("--help",)
+
+        kwargs.setdefault("is_flag", True)
+        kwargs.setdefault("expose_value", False)
+        kwargs.setdefault("is_eager", True)
+        kwargs.setdefault("help", _("Show this message and exit."))
+        kwargs.setdefault("callback", self.show_help)
+
+        super().__init__(param_decls, **kwargs)
+
+    @staticmethod
+    def show_help(ctx: Context, param: Parameter, value: bool) -> None:
+        """Callback that print the help page on ``<stdout>`` and exits."""
+        if value and not ctx.resilient_parsing:
+            echo(ctx.get_help(), color=ctx.color)
+            ctx.exit()
+
+
+def help_option(*param_decls: str, **kwargs: t.Any) -> t.Callable[[FC], FC]:
+    """Decorator for the pre-configured ``--help`` option defined above.
+
+    :param param_decls: One or more option names. Defaults to the single
+        value ``"--help"``.
+    :param kwargs: Extra arguments are passed to :func:`option`.
+    """
+    kwargs.setdefault("cls", HelpOption)
+    return option(*param_decls, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/click/exceptions.py b/backend/venv/lib/python3.12/site-packages/click/exceptions.py
new file mode 100644
index 00000000..0b831516
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/exceptions.py
@@ -0,0 +1,296 @@
+import typing as t
+from gettext import gettext as _
+from gettext import ngettext
+
+from ._compat import get_text_stderr
+from .globals import resolve_color_default
+from .utils import echo
+from .utils import format_filename
+
+if t.TYPE_CHECKING:
+    from .core import Command
+    from .core import Context
+    from .core import Parameter
+
+
+def _join_param_hints(
+    param_hint: t.Optional[t.Union[t.Sequence[str], str]],
+) -> t.Optional[str]:
+    if param_hint is not None and not isinstance(param_hint, str):
+        return " / ".join(repr(x) for x in param_hint)
+
+    return param_hint
+
+
+class ClickException(Exception):
+    """An exception that Click can handle and show to the user."""
+
+    #: The exit code for this exception.
+    exit_code = 1
+
+    def __init__(self, message: str) -> None:
+        super().__init__(message)
+        # The context will be removed by the time we print the message, so cache
+        # the color settings here to be used later on (in `show`)
+        self.show_color: t.Optional[bool] = resolve_color_default()
+        self.message = message
+
+    def format_message(self) -> str:
+        return self.message
+
+    def __str__(self) -> str:
+        return self.message
+
+    def show(self, file: t.Optional[t.IO[t.Any]] = None) -> None:
+        if file is None:
+            file = get_text_stderr()
+
+        echo(
+            _("Error: {message}").format(message=self.format_message()),
+            file=file,
+            color=self.show_color,
+        )
+
+
+class UsageError(ClickException):
+    """An internal exception that signals a usage error.  This typically
+    aborts any further handling.
+
+    :param message: the error message to display.
+    :param ctx: optionally the context that caused this error.  Click will
+                fill in the context automatically in some situations.
+    """
+
+    exit_code = 2
+
+    def __init__(self, message: str, ctx: t.Optional["Context"] = None) -> None:
+        super().__init__(message)
+        self.ctx = ctx
+        self.cmd: t.Optional[Command] = self.ctx.command if self.ctx else None
+
+    def show(self, file: t.Optional[t.IO[t.Any]] = None) -> None:
+        if file is None:
+            file = get_text_stderr()
+        color = None
+        hint = ""
+        if (
+            self.ctx is not None
+            and self.ctx.command.get_help_option(self.ctx) is not None
+        ):
+            hint = _("Try '{command} {option}' for help.").format(
+                command=self.ctx.command_path, option=self.ctx.help_option_names[0]
+            )
+            hint = f"{hint}\n"
+        if self.ctx is not None:
+            color = self.ctx.color
+            echo(f"{self.ctx.get_usage()}\n{hint}", file=file, color=color)
+        echo(
+            _("Error: {message}").format(message=self.format_message()),
+            file=file,
+            color=color,
+        )
+
+
+class BadParameter(UsageError):
+    """An exception that formats out a standardized error message for a
+    bad parameter.  This is useful when thrown from a callback or type as
+    Click will attach contextual information to it (for instance, which
+    parameter it is).
+
+    .. versionadded:: 2.0
+
+    :param param: the parameter object that caused this error.  This can
+                  be left out, and Click will attach this info itself
+                  if possible.
+    :param param_hint: a string that shows up as parameter name.  This
+                       can be used as alternative to `param` in cases
+                       where custom validation should happen.  If it is
+                       a string it's used as such, if it's a list then
+                       each item is quoted and separated.
+    """
+
+    def __init__(
+        self,
+        message: str,
+        ctx: t.Optional["Context"] = None,
+        param: t.Optional["Parameter"] = None,
+        param_hint: t.Optional[str] = None,
+    ) -> None:
+        super().__init__(message, ctx)
+        self.param = param
+        self.param_hint = param_hint
+
+    def format_message(self) -> str:
+        if self.param_hint is not None:
+            param_hint = self.param_hint
+        elif self.param is not None:
+            param_hint = self.param.get_error_hint(self.ctx)  # type: ignore
+        else:
+            return _("Invalid value: {message}").format(message=self.message)
+
+        return _("Invalid value for {param_hint}: {message}").format(
+            param_hint=_join_param_hints(param_hint), message=self.message
+        )
+
+
+class MissingParameter(BadParameter):
+    """Raised if click required an option or argument but it was not
+    provided when invoking the script.
+
+    .. versionadded:: 4.0
+
+    :param param_type: a string that indicates the type of the parameter.
+                       The default is to inherit the parameter type from
+                       the given `param`.  Valid values are ``'parameter'``,
+                       ``'option'`` or ``'argument'``.
+    """
+
+    def __init__(
+        self,
+        message: t.Optional[str] = None,
+        ctx: t.Optional["Context"] = None,
+        param: t.Optional["Parameter"] = None,
+        param_hint: t.Optional[str] = None,
+        param_type: t.Optional[str] = None,
+    ) -> None:
+        super().__init__(message or "", ctx, param, param_hint)
+        self.param_type = param_type
+
+    def format_message(self) -> str:
+        if self.param_hint is not None:
+            param_hint: t.Optional[str] = self.param_hint
+        elif self.param is not None:
+            param_hint = self.param.get_error_hint(self.ctx)  # type: ignore
+        else:
+            param_hint = None
+
+        param_hint = _join_param_hints(param_hint)
+        param_hint = f" {param_hint}" if param_hint else ""
+
+        param_type = self.param_type
+        if param_type is None and self.param is not None:
+            param_type = self.param.param_type_name
+
+        msg = self.message
+        if self.param is not None:
+            msg_extra = self.param.type.get_missing_message(self.param)
+            if msg_extra:
+                if msg:
+                    msg += f". {msg_extra}"
+                else:
+                    msg = msg_extra
+
+        msg = f" {msg}" if msg else ""
+
+        # Translate param_type for known types.
+        if param_type == "argument":
+            missing = _("Missing argument")
+        elif param_type == "option":
+            missing = _("Missing option")
+        elif param_type == "parameter":
+            missing = _("Missing parameter")
+        else:
+            missing = _("Missing {param_type}").format(param_type=param_type)
+
+        return f"{missing}{param_hint}.{msg}"
+
+    def __str__(self) -> str:
+        if not self.message:
+            param_name = self.param.name if self.param else None
+            return _("Missing parameter: {param_name}").format(param_name=param_name)
+        else:
+            return self.message
+
+
+class NoSuchOption(UsageError):
+    """Raised if click attempted to handle an option that does not
+    exist.
+
+    .. versionadded:: 4.0
+    """
+
+    def __init__(
+        self,
+        option_name: str,
+        message: t.Optional[str] = None,
+        possibilities: t.Optional[t.Sequence[str]] = None,
+        ctx: t.Optional["Context"] = None,
+    ) -> None:
+        if message is None:
+            message = _("No such option: {name}").format(name=option_name)
+
+        super().__init__(message, ctx)
+        self.option_name = option_name
+        self.possibilities = possibilities
+
+    def format_message(self) -> str:
+        if not self.possibilities:
+            return self.message
+
+        possibility_str = ", ".join(sorted(self.possibilities))
+        suggest = ngettext(
+            "Did you mean {possibility}?",
+            "(Possible options: {possibilities})",
+            len(self.possibilities),
+        ).format(possibility=possibility_str, possibilities=possibility_str)
+        return f"{self.message} {suggest}"
+
+
+class BadOptionUsage(UsageError):
+    """Raised if an option is generally supplied but the use of the option
+    was incorrect.  This is for instance raised if the number of arguments
+    for an option is not correct.
+
+    .. versionadded:: 4.0
+
+    :param option_name: the name of the option being used incorrectly.
+    """
+
+    def __init__(
+        self, option_name: str, message: str, ctx: t.Optional["Context"] = None
+    ) -> None:
+        super().__init__(message, ctx)
+        self.option_name = option_name
+
+
+class BadArgumentUsage(UsageError):
+    """Raised if an argument is generally supplied but the use of the argument
+    was incorrect.  This is for instance raised if the number of values
+    for an argument is not correct.
+
+    .. versionadded:: 6.0
+    """
+
+
+class FileError(ClickException):
+    """Raised if a file cannot be opened."""
+
+    def __init__(self, filename: str, hint: t.Optional[str] = None) -> None:
+        if hint is None:
+            hint = _("unknown error")
+
+        super().__init__(hint)
+        self.ui_filename: str = format_filename(filename)
+        self.filename = filename
+
+    def format_message(self) -> str:
+        return _("Could not open file {filename!r}: {message}").format(
+            filename=self.ui_filename, message=self.message
+        )
+
+
+class Abort(RuntimeError):
+    """An internal signalling exception that signals Click to abort."""
+
+
+class Exit(RuntimeError):
+    """An exception that indicates that the application should exit with some
+    status code.
+
+    :param code: the status code to exit with.
+    """
+
+    __slots__ = ("exit_code",)
+
+    def __init__(self, code: int = 0) -> None:
+        self.exit_code: int = code
diff --git a/backend/venv/lib/python3.12/site-packages/click/formatting.py b/backend/venv/lib/python3.12/site-packages/click/formatting.py
new file mode 100644
index 00000000..ddd2a2f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/formatting.py
@@ -0,0 +1,301 @@
+import typing as t
+from contextlib import contextmanager
+from gettext import gettext as _
+
+from ._compat import term_len
+from .parser import split_opt
+
+# Can force a width.  This is used by the test system
+FORCED_WIDTH: t.Optional[int] = None
+
+
+def measure_table(rows: t.Iterable[t.Tuple[str, str]]) -> t.Tuple[int, ...]:
+    widths: t.Dict[int, int] = {}
+
+    for row in rows:
+        for idx, col in enumerate(row):
+            widths[idx] = max(widths.get(idx, 0), term_len(col))
+
+    return tuple(y for x, y in sorted(widths.items()))
+
+
+def iter_rows(
+    rows: t.Iterable[t.Tuple[str, str]], col_count: int
+) -> t.Iterator[t.Tuple[str, ...]]:
+    for row in rows:
+        yield row + ("",) * (col_count - len(row))
+
+
+def wrap_text(
+    text: str,
+    width: int = 78,
+    initial_indent: str = "",
+    subsequent_indent: str = "",
+    preserve_paragraphs: bool = False,
+) -> str:
+    """A helper function that intelligently wraps text.  By default, it
+    assumes that it operates on a single paragraph of text but if the
+    `preserve_paragraphs` parameter is provided it will intelligently
+    handle paragraphs (defined by two empty lines).
+
+    If paragraphs are handled, a paragraph can be prefixed with an empty
+    line containing the ``\\b`` character (``\\x08``) to indicate that
+    no rewrapping should happen in that block.
+
+    :param text: the text that should be rewrapped.
+    :param width: the maximum width for the text.
+    :param initial_indent: the initial indent that should be placed on the
+                           first line as a string.
+    :param subsequent_indent: the indent string that should be placed on
+                              each consecutive line.
+    :param preserve_paragraphs: if this flag is set then the wrapping will
+                                intelligently handle paragraphs.
+    """
+    from ._textwrap import TextWrapper
+
+    text = text.expandtabs()
+    wrapper = TextWrapper(
+        width,
+        initial_indent=initial_indent,
+        subsequent_indent=subsequent_indent,
+        replace_whitespace=False,
+    )
+    if not preserve_paragraphs:
+        return wrapper.fill(text)
+
+    p: t.List[t.Tuple[int, bool, str]] = []
+    buf: t.List[str] = []
+    indent = None
+
+    def _flush_par() -> None:
+        if not buf:
+            return
+        if buf[0].strip() == "\b":
+            p.append((indent or 0, True, "\n".join(buf[1:])))
+        else:
+            p.append((indent or 0, False, " ".join(buf)))
+        del buf[:]
+
+    for line in text.splitlines():
+        if not line:
+            _flush_par()
+            indent = None
+        else:
+            if indent is None:
+                orig_len = term_len(line)
+                line = line.lstrip()
+                indent = orig_len - term_len(line)
+            buf.append(line)
+    _flush_par()
+
+    rv = []
+    for indent, raw, text in p:
+        with wrapper.extra_indent(" " * indent):
+            if raw:
+                rv.append(wrapper.indent_only(text))
+            else:
+                rv.append(wrapper.fill(text))
+
+    return "\n\n".join(rv)
+
+
+class HelpFormatter:
+    """This class helps with formatting text-based help pages.  It's
+    usually just needed for very special internal cases, but it's also
+    exposed so that developers can write their own fancy outputs.
+
+    At present, it always writes into memory.
+
+    :param indent_increment: the additional increment for each level.
+    :param width: the width for the text.  This defaults to the terminal
+                  width clamped to a maximum of 78.
+    """
+
+    def __init__(
+        self,
+        indent_increment: int = 2,
+        width: t.Optional[int] = None,
+        max_width: t.Optional[int] = None,
+    ) -> None:
+        import shutil
+
+        self.indent_increment = indent_increment
+        if max_width is None:
+            max_width = 80
+        if width is None:
+            width = FORCED_WIDTH
+            if width is None:
+                width = max(min(shutil.get_terminal_size().columns, max_width) - 2, 50)
+        self.width = width
+        self.current_indent = 0
+        self.buffer: t.List[str] = []
+
+    def write(self, string: str) -> None:
+        """Writes a unicode string into the internal buffer."""
+        self.buffer.append(string)
+
+    def indent(self) -> None:
+        """Increases the indentation."""
+        self.current_indent += self.indent_increment
+
+    def dedent(self) -> None:
+        """Decreases the indentation."""
+        self.current_indent -= self.indent_increment
+
+    def write_usage(
+        self, prog: str, args: str = "", prefix: t.Optional[str] = None
+    ) -> None:
+        """Writes a usage line into the buffer.
+
+        :param prog: the program name.
+        :param args: whitespace separated list of arguments.
+        :param prefix: The prefix for the first line. Defaults to
+            ``"Usage: "``.
+        """
+        if prefix is None:
+            prefix = f"{_('Usage:')} "
+
+        usage_prefix = f"{prefix:>{self.current_indent}}{prog} "
+        text_width = self.width - self.current_indent
+
+        if text_width >= (term_len(usage_prefix) + 20):
+            # The arguments will fit to the right of the prefix.
+            indent = " " * term_len(usage_prefix)
+            self.write(
+                wrap_text(
+                    args,
+                    text_width,
+                    initial_indent=usage_prefix,
+                    subsequent_indent=indent,
+                )
+            )
+        else:
+            # The prefix is too long, put the arguments on the next line.
+            self.write(usage_prefix)
+            self.write("\n")
+            indent = " " * (max(self.current_indent, term_len(prefix)) + 4)
+            self.write(
+                wrap_text(
+                    args, text_width, initial_indent=indent, subsequent_indent=indent
+                )
+            )
+
+        self.write("\n")
+
+    def write_heading(self, heading: str) -> None:
+        """Writes a heading into the buffer."""
+        self.write(f"{'':>{self.current_indent}}{heading}:\n")
+
+    def write_paragraph(self) -> None:
+        """Writes a paragraph into the buffer."""
+        if self.buffer:
+            self.write("\n")
+
+    def write_text(self, text: str) -> None:
+        """Writes re-indented text into the buffer.  This rewraps and
+        preserves paragraphs.
+        """
+        indent = " " * self.current_indent
+        self.write(
+            wrap_text(
+                text,
+                self.width,
+                initial_indent=indent,
+                subsequent_indent=indent,
+                preserve_paragraphs=True,
+            )
+        )
+        self.write("\n")
+
+    def write_dl(
+        self,
+        rows: t.Sequence[t.Tuple[str, str]],
+        col_max: int = 30,
+        col_spacing: int = 2,
+    ) -> None:
+        """Writes a definition list into the buffer.  This is how options
+        and commands are usually formatted.
+
+        :param rows: a list of two item tuples for the terms and values.
+        :param col_max: the maximum width of the first column.
+        :param col_spacing: the number of spaces between the first and
+                            second column.
+        """
+        rows = list(rows)
+        widths = measure_table(rows)
+        if len(widths) != 2:
+            raise TypeError("Expected two columns for definition list")
+
+        first_col = min(widths[0], col_max) + col_spacing
+
+        for first, second in iter_rows(rows, len(widths)):
+            self.write(f"{'':>{self.current_indent}}{first}")
+            if not second:
+                self.write("\n")
+                continue
+            if term_len(first) <= first_col - col_spacing:
+                self.write(" " * (first_col - term_len(first)))
+            else:
+                self.write("\n")
+                self.write(" " * (first_col + self.current_indent))
+
+            text_width = max(self.width - first_col - 2, 10)
+            wrapped_text = wrap_text(second, text_width, preserve_paragraphs=True)
+            lines = wrapped_text.splitlines()
+
+            if lines:
+                self.write(f"{lines[0]}\n")
+
+                for line in lines[1:]:
+                    self.write(f"{'':>{first_col + self.current_indent}}{line}\n")
+            else:
+                self.write("\n")
+
+    @contextmanager
+    def section(self, name: str) -> t.Iterator[None]:
+        """Helpful context manager that writes a paragraph, a heading,
+        and the indents.
+
+        :param name: the section name that is written as heading.
+        """
+        self.write_paragraph()
+        self.write_heading(name)
+        self.indent()
+        try:
+            yield
+        finally:
+            self.dedent()
+
+    @contextmanager
+    def indentation(self) -> t.Iterator[None]:
+        """A context manager that increases the indentation."""
+        self.indent()
+        try:
+            yield
+        finally:
+            self.dedent()
+
+    def getvalue(self) -> str:
+        """Returns the buffer contents."""
+        return "".join(self.buffer)
+
+
+def join_options(options: t.Sequence[str]) -> t.Tuple[str, bool]:
+    """Given a list of option strings this joins them in the most appropriate
+    way and returns them in the form ``(formatted_string,
+    any_prefix_is_slash)`` where the second item in the tuple is a flag that
+    indicates if any of the option prefixes was a slash.
+    """
+    rv = []
+    any_prefix_is_slash = False
+
+    for opt in options:
+        prefix = split_opt(opt)[0]
+
+        if prefix == "/":
+            any_prefix_is_slash = True
+
+        rv.append((len(prefix), opt))
+
+    rv.sort(key=lambda x: x[0])
+    return ", ".join(x[1] for x in rv), any_prefix_is_slash
diff --git a/backend/venv/lib/python3.12/site-packages/click/globals.py b/backend/venv/lib/python3.12/site-packages/click/globals.py
new file mode 100644
index 00000000..191e712d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/globals.py
@@ -0,0 +1,67 @@
+import typing as t
+from threading import local
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .core import Context
+
+_local = local()
+
+
+@t.overload
+def get_current_context(silent: "te.Literal[False]" = False) -> "Context": ...
+
+
+@t.overload
+def get_current_context(silent: bool = ...) -> t.Optional["Context"]: ...
+
+
+def get_current_context(silent: bool = False) -> t.Optional["Context"]:
+    """Returns the current click context.  This can be used as a way to
+    access the current context object from anywhere.  This is a more implicit
+    alternative to the :func:`pass_context` decorator.  This function is
+    primarily useful for helpers such as :func:`echo` which might be
+    interested in changing its behavior based on the current context.
+
+    To push the current context, :meth:`Context.scope` can be used.
+
+    .. versionadded:: 5.0
+
+    :param silent: if set to `True` the return value is `None` if no context
+                   is available.  The default behavior is to raise a
+                   :exc:`RuntimeError`.
+    """
+    try:
+        return t.cast("Context", _local.stack[-1])
+    except (AttributeError, IndexError) as e:
+        if not silent:
+            raise RuntimeError("There is no active click context.") from e
+
+    return None
+
+
+def push_context(ctx: "Context") -> None:
+    """Pushes a new context to the current stack."""
+    _local.__dict__.setdefault("stack", []).append(ctx)
+
+
+def pop_context() -> None:
+    """Removes the top level from the stack."""
+    _local.stack.pop()
+
+
+def resolve_color_default(color: t.Optional[bool] = None) -> t.Optional[bool]:
+    """Internal helper to get the default value of the color flag.  If a
+    value is passed it's returned unchanged, otherwise it's looked up from
+    the current context.
+    """
+    if color is not None:
+        return color
+
+    ctx = get_current_context(silent=True)
+
+    if ctx is not None:
+        return ctx.color
+
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/click/parser.py b/backend/venv/lib/python3.12/site-packages/click/parser.py
new file mode 100644
index 00000000..600b8436
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/parser.py
@@ -0,0 +1,531 @@
+"""
+This module started out as largely a copy paste from the stdlib's
+optparse module with the features removed that we do not need from
+optparse because we implement them in Click on a higher level (for
+instance type handling, help formatting and a lot more).
+
+The plan is to remove more and more from here over time.
+
+The reason this is a different module and not optparse from the stdlib
+is that there are differences in 2.x and 3.x about the error messages
+generated and optparse in the stdlib uses gettext for no good reason
+and might cause us issues.
+
+Click uses parts of optparse written by Gregory P. Ward and maintained
+by the Python Software Foundation. This is limited to code in parser.py.
+
+Copyright 2001-2006 Gregory P. Ward. All rights reserved.
+Copyright 2002-2006 Python Software Foundation. All rights reserved.
+"""
+
+# This code uses parts of optparse written by Gregory P. Ward and
+# maintained by the Python Software Foundation.
+# Copyright 2001-2006 Gregory P. Ward
+# Copyright 2002-2006 Python Software Foundation
+import typing as t
+from collections import deque
+from gettext import gettext as _
+from gettext import ngettext
+
+from .exceptions import BadArgumentUsage
+from .exceptions import BadOptionUsage
+from .exceptions import NoSuchOption
+from .exceptions import UsageError
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .core import Argument as CoreArgument
+    from .core import Context
+    from .core import Option as CoreOption
+    from .core import Parameter as CoreParameter
+
+V = t.TypeVar("V")
+
+# Sentinel value that indicates an option was passed as a flag without a
+# value but is not a flag option. Option.consume_value uses this to
+# prompt or use the flag_value.
+_flag_needs_value = object()
+
+
+def _unpack_args(
+    args: t.Sequence[str], nargs_spec: t.Sequence[int]
+) -> t.Tuple[t.Sequence[t.Union[str, t.Sequence[t.Optional[str]], None]], t.List[str]]:
+    """Given an iterable of arguments and an iterable of nargs specifications,
+    it returns a tuple with all the unpacked arguments at the first index
+    and all remaining arguments as the second.
+
+    The nargs specification is the number of arguments that should be consumed
+    or `-1` to indicate that this position should eat up all the remainders.
+
+    Missing items are filled with `None`.
+    """
+    args = deque(args)
+    nargs_spec = deque(nargs_spec)
+    rv: t.List[t.Union[str, t.Tuple[t.Optional[str], ...], None]] = []
+    spos: t.Optional[int] = None
+
+    def _fetch(c: "te.Deque[V]") -> t.Optional[V]:
+        try:
+            if spos is None:
+                return c.popleft()
+            else:
+                return c.pop()
+        except IndexError:
+            return None
+
+    while nargs_spec:
+        nargs = _fetch(nargs_spec)
+
+        if nargs is None:
+            continue
+
+        if nargs == 1:
+            rv.append(_fetch(args))
+        elif nargs > 1:
+            x = [_fetch(args) for _ in range(nargs)]
+
+            # If we're reversed, we're pulling in the arguments in reverse,
+            # so we need to turn them around.
+            if spos is not None:
+                x.reverse()
+
+            rv.append(tuple(x))
+        elif nargs < 0:
+            if spos is not None:
+                raise TypeError("Cannot have two nargs < 0")
+
+            spos = len(rv)
+            rv.append(None)
+
+    # spos is the position of the wildcard (star).  If it's not `None`,
+    # we fill it with the remainder.
+    if spos is not None:
+        rv[spos] = tuple(args)
+        args = []
+        rv[spos + 1 :] = reversed(rv[spos + 1 :])
+
+    return tuple(rv), list(args)
+
+
+def split_opt(opt: str) -> t.Tuple[str, str]:
+    first = opt[:1]
+    if first.isalnum():
+        return "", opt
+    if opt[1:2] == first:
+        return opt[:2], opt[2:]
+    return first, opt[1:]
+
+
+def normalize_opt(opt: str, ctx: t.Optional["Context"]) -> str:
+    if ctx is None or ctx.token_normalize_func is None:
+        return opt
+    prefix, opt = split_opt(opt)
+    return f"{prefix}{ctx.token_normalize_func(opt)}"
+
+
+def split_arg_string(string: str) -> t.List[str]:
+    """Split an argument string as with :func:`shlex.split`, but don't
+    fail if the string is incomplete. Ignores a missing closing quote or
+    incomplete escape sequence and uses the partial token as-is.
+
+    .. code-block:: python
+
+        split_arg_string("example 'my file")
+        ["example", "my file"]
+
+        split_arg_string("example my\\")
+        ["example", "my"]
+
+    :param string: String to split.
+    """
+    import shlex
+
+    lex = shlex.shlex(string, posix=True)
+    lex.whitespace_split = True
+    lex.commenters = ""
+    out = []
+
+    try:
+        for token in lex:
+            out.append(token)
+    except ValueError:
+        # Raised when end-of-string is reached in an invalid state. Use
+        # the partial token as-is. The quote or escape character is in
+        # lex.state, not lex.token.
+        out.append(lex.token)
+
+    return out
+
+
+class Option:
+    def __init__(
+        self,
+        obj: "CoreOption",
+        opts: t.Sequence[str],
+        dest: t.Optional[str],
+        action: t.Optional[str] = None,
+        nargs: int = 1,
+        const: t.Optional[t.Any] = None,
+    ):
+        self._short_opts = []
+        self._long_opts = []
+        self.prefixes: t.Set[str] = set()
+
+        for opt in opts:
+            prefix, value = split_opt(opt)
+            if not prefix:
+                raise ValueError(f"Invalid start character for option ({opt})")
+            self.prefixes.add(prefix[0])
+            if len(prefix) == 1 and len(value) == 1:
+                self._short_opts.append(opt)
+            else:
+                self._long_opts.append(opt)
+                self.prefixes.add(prefix)
+
+        if action is None:
+            action = "store"
+
+        self.dest = dest
+        self.action = action
+        self.nargs = nargs
+        self.const = const
+        self.obj = obj
+
+    @property
+    def takes_value(self) -> bool:
+        return self.action in ("store", "append")
+
+    def process(self, value: t.Any, state: "ParsingState") -> None:
+        if self.action == "store":
+            state.opts[self.dest] = value  # type: ignore
+        elif self.action == "store_const":
+            state.opts[self.dest] = self.const  # type: ignore
+        elif self.action == "append":
+            state.opts.setdefault(self.dest, []).append(value)  # type: ignore
+        elif self.action == "append_const":
+            state.opts.setdefault(self.dest, []).append(self.const)  # type: ignore
+        elif self.action == "count":
+            state.opts[self.dest] = state.opts.get(self.dest, 0) + 1  # type: ignore
+        else:
+            raise ValueError(f"unknown action '{self.action}'")
+        state.order.append(self.obj)
+
+
+class Argument:
+    def __init__(self, obj: "CoreArgument", dest: t.Optional[str], nargs: int = 1):
+        self.dest = dest
+        self.nargs = nargs
+        self.obj = obj
+
+    def process(
+        self,
+        value: t.Union[t.Optional[str], t.Sequence[t.Optional[str]]],
+        state: "ParsingState",
+    ) -> None:
+        if self.nargs > 1:
+            assert value is not None
+            holes = sum(1 for x in value if x is None)
+            if holes == len(value):
+                value = None
+            elif holes != 0:
+                raise BadArgumentUsage(
+                    _("Argument {name!r} takes {nargs} values.").format(
+                        name=self.dest, nargs=self.nargs
+                    )
+                )
+
+        if self.nargs == -1 and self.obj.envvar is not None and value == ():
+            # Replace empty tuple with None so that a value from the
+            # environment may be tried.
+            value = None
+
+        state.opts[self.dest] = value  # type: ignore
+        state.order.append(self.obj)
+
+
+class ParsingState:
+    def __init__(self, rargs: t.List[str]) -> None:
+        self.opts: t.Dict[str, t.Any] = {}
+        self.largs: t.List[str] = []
+        self.rargs = rargs
+        self.order: t.List[CoreParameter] = []
+
+
+class OptionParser:
+    """The option parser is an internal class that is ultimately used to
+    parse options and arguments.  It's modelled after optparse and brings
+    a similar but vastly simplified API.  It should generally not be used
+    directly as the high level Click classes wrap it for you.
+
+    It's not nearly as extensible as optparse or argparse as it does not
+    implement features that are implemented on a higher level (such as
+    types or defaults).
+
+    :param ctx: optionally the :class:`~click.Context` where this parser
+                should go with.
+    """
+
+    def __init__(self, ctx: t.Optional["Context"] = None) -> None:
+        #: The :class:`~click.Context` for this parser.  This might be
+        #: `None` for some advanced use cases.
+        self.ctx = ctx
+        #: This controls how the parser deals with interspersed arguments.
+        #: If this is set to `False`, the parser will stop on the first
+        #: non-option.  Click uses this to implement nested subcommands
+        #: safely.
+        self.allow_interspersed_args: bool = True
+        #: This tells the parser how to deal with unknown options.  By
+        #: default it will error out (which is sensible), but there is a
+        #: second mode where it will ignore it and continue processing
+        #: after shifting all the unknown options into the resulting args.
+        self.ignore_unknown_options: bool = False
+
+        if ctx is not None:
+            self.allow_interspersed_args = ctx.allow_interspersed_args
+            self.ignore_unknown_options = ctx.ignore_unknown_options
+
+        self._short_opt: t.Dict[str, Option] = {}
+        self._long_opt: t.Dict[str, Option] = {}
+        self._opt_prefixes = {"-", "--"}
+        self._args: t.List[Argument] = []
+
+    def add_option(
+        self,
+        obj: "CoreOption",
+        opts: t.Sequence[str],
+        dest: t.Optional[str],
+        action: t.Optional[str] = None,
+        nargs: int = 1,
+        const: t.Optional[t.Any] = None,
+    ) -> None:
+        """Adds a new option named `dest` to the parser.  The destination
+        is not inferred (unlike with optparse) and needs to be explicitly
+        provided.  Action can be any of ``store``, ``store_const``,
+        ``append``, ``append_const`` or ``count``.
+
+        The `obj` can be used to identify the option in the order list
+        that is returned from the parser.
+        """
+        opts = [normalize_opt(opt, self.ctx) for opt in opts]
+        option = Option(obj, opts, dest, action=action, nargs=nargs, const=const)
+        self._opt_prefixes.update(option.prefixes)
+        for opt in option._short_opts:
+            self._short_opt[opt] = option
+        for opt in option._long_opts:
+            self._long_opt[opt] = option
+
+    def add_argument(
+        self, obj: "CoreArgument", dest: t.Optional[str], nargs: int = 1
+    ) -> None:
+        """Adds a positional argument named `dest` to the parser.
+
+        The `obj` can be used to identify the option in the order list
+        that is returned from the parser.
+        """
+        self._args.append(Argument(obj, dest=dest, nargs=nargs))
+
+    def parse_args(
+        self, args: t.List[str]
+    ) -> t.Tuple[t.Dict[str, t.Any], t.List[str], t.List["CoreParameter"]]:
+        """Parses positional arguments and returns ``(values, args, order)``
+        for the parsed options and arguments as well as the leftover
+        arguments if there are any.  The order is a list of objects as they
+        appear on the command line.  If arguments appear multiple times they
+        will be memorized multiple times as well.
+        """
+        state = ParsingState(args)
+        try:
+            self._process_args_for_options(state)
+            self._process_args_for_args(state)
+        except UsageError:
+            if self.ctx is None or not self.ctx.resilient_parsing:
+                raise
+        return state.opts, state.largs, state.order
+
+    def _process_args_for_args(self, state: ParsingState) -> None:
+        pargs, args = _unpack_args(
+            state.largs + state.rargs, [x.nargs for x in self._args]
+        )
+
+        for idx, arg in enumerate(self._args):
+            arg.process(pargs[idx], state)
+
+        state.largs = args
+        state.rargs = []
+
+    def _process_args_for_options(self, state: ParsingState) -> None:
+        while state.rargs:
+            arg = state.rargs.pop(0)
+            arglen = len(arg)
+            # Double dashes always handled explicitly regardless of what
+            # prefixes are valid.
+            if arg == "--":
+                return
+            elif arg[:1] in self._opt_prefixes and arglen > 1:
+                self._process_opts(arg, state)
+            elif self.allow_interspersed_args:
+                state.largs.append(arg)
+            else:
+                state.rargs.insert(0, arg)
+                return
+
+        # Say this is the original argument list:
+        # [arg0, arg1, ..., arg(i-1), arg(i), arg(i+1), ..., arg(N-1)]
+        #                            ^
+        # (we are about to process arg(i)).
+        #
+        # Then rargs is [arg(i), ..., arg(N-1)] and largs is a *subset* of
+        # [arg0, ..., arg(i-1)] (any options and their arguments will have
+        # been removed from largs).
+        #
+        # The while loop will usually consume 1 or more arguments per pass.
+        # If it consumes 1 (eg. arg is an option that takes no arguments),
+        # then after _process_arg() is done the situation is:
+        #
+        #   largs = subset of [arg0, ..., arg(i)]
+        #   rargs = [arg(i+1), ..., arg(N-1)]
+        #
+        # If allow_interspersed_args is false, largs will always be
+        # *empty* -- still a subset of [arg0, ..., arg(i-1)], but
+        # not a very interesting subset!
+
+    def _match_long_opt(
+        self, opt: str, explicit_value: t.Optional[str], state: ParsingState
+    ) -> None:
+        if opt not in self._long_opt:
+            from difflib import get_close_matches
+
+            possibilities = get_close_matches(opt, self._long_opt)
+            raise NoSuchOption(opt, possibilities=possibilities, ctx=self.ctx)
+
+        option = self._long_opt[opt]
+        if option.takes_value:
+            # At this point it's safe to modify rargs by injecting the
+            # explicit value, because no exception is raised in this
+            # branch.  This means that the inserted value will be fully
+            # consumed.
+            if explicit_value is not None:
+                state.rargs.insert(0, explicit_value)
+
+            value = self._get_value_from_state(opt, option, state)
+
+        elif explicit_value is not None:
+            raise BadOptionUsage(
+                opt, _("Option {name!r} does not take a value.").format(name=opt)
+            )
+
+        else:
+            value = None
+
+        option.process(value, state)
+
+    def _match_short_opt(self, arg: str, state: ParsingState) -> None:
+        stop = False
+        i = 1
+        prefix = arg[0]
+        unknown_options = []
+
+        for ch in arg[1:]:
+            opt = normalize_opt(f"{prefix}{ch}", self.ctx)
+            option = self._short_opt.get(opt)
+            i += 1
+
+            if not option:
+                if self.ignore_unknown_options:
+                    unknown_options.append(ch)
+                    continue
+                raise NoSuchOption(opt, ctx=self.ctx)
+            if option.takes_value:
+                # Any characters left in arg?  Pretend they're the
+                # next arg, and stop consuming characters of arg.
+                if i < len(arg):
+                    state.rargs.insert(0, arg[i:])
+                    stop = True
+
+                value = self._get_value_from_state(opt, option, state)
+
+            else:
+                value = None
+
+            option.process(value, state)
+
+            if stop:
+                break
+
+        # If we got any unknown options we recombine the string of the
+        # remaining options and re-attach the prefix, then report that
+        # to the state as new larg.  This way there is basic combinatorics
+        # that can be achieved while still ignoring unknown arguments.
+        if self.ignore_unknown_options and unknown_options:
+            state.largs.append(f"{prefix}{''.join(unknown_options)}")
+
+    def _get_value_from_state(
+        self, option_name: str, option: Option, state: ParsingState
+    ) -> t.Any:
+        nargs = option.nargs
+
+        if len(state.rargs) < nargs:
+            if option.obj._flag_needs_value:
+                # Option allows omitting the value.
+                value = _flag_needs_value
+            else:
+                raise BadOptionUsage(
+                    option_name,
+                    ngettext(
+                        "Option {name!r} requires an argument.",
+                        "Option {name!r} requires {nargs} arguments.",
+                        nargs,
+                    ).format(name=option_name, nargs=nargs),
+                )
+        elif nargs == 1:
+            next_rarg = state.rargs[0]
+
+            if (
+                option.obj._flag_needs_value
+                and isinstance(next_rarg, str)
+                and next_rarg[:1] in self._opt_prefixes
+                and len(next_rarg) > 1
+            ):
+                # The next arg looks like the start of an option, don't
+                # use it as the value if omitting the value is allowed.
+                value = _flag_needs_value
+            else:
+                value = state.rargs.pop(0)
+        else:
+            value = tuple(state.rargs[:nargs])
+            del state.rargs[:nargs]
+
+        return value
+
+    def _process_opts(self, arg: str, state: ParsingState) -> None:
+        explicit_value = None
+        # Long option handling happens in two parts.  The first part is
+        # supporting explicitly attached values.  In any case, we will try
+        # to long match the option first.
+        if "=" in arg:
+            long_opt, explicit_value = arg.split("=", 1)
+        else:
+            long_opt = arg
+        norm_long_opt = normalize_opt(long_opt, self.ctx)
+
+        # At this point we will match the (assumed) long option through
+        # the long option matching code.  Note that this allows options
+        # like "-foo" to be matched as long options.
+        try:
+            self._match_long_opt(norm_long_opt, explicit_value, state)
+        except NoSuchOption:
+            # At this point the long option matching failed, and we need
+            # to try with short options.  However there is a special rule
+            # which says, that if we have a two character options prefix
+            # (applies to "--foo" for instance), we do not dispatch to the
+            # short option code and will instead raise the no option
+            # error.
+            if arg[:2] not in self._opt_prefixes:
+                self._match_short_opt(arg, state)
+                return
+
+            if not self.ignore_unknown_options:
+                raise
+
+            state.largs.append(arg)
diff --git a/backend/venv/lib/python3.12/site-packages/click/py.typed b/backend/venv/lib/python3.12/site-packages/click/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/click/shell_completion.py b/backend/venv/lib/python3.12/site-packages/click/shell_completion.py
new file mode 100644
index 00000000..07d0f09b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/shell_completion.py
@@ -0,0 +1,603 @@
+import os
+import re
+import typing as t
+from gettext import gettext as _
+
+from .core import Argument
+from .core import BaseCommand
+from .core import Context
+from .core import MultiCommand
+from .core import Option
+from .core import Parameter
+from .core import ParameterSource
+from .parser import split_arg_string
+from .utils import echo
+
+
+def shell_complete(
+    cli: BaseCommand,
+    ctx_args: t.MutableMapping[str, t.Any],
+    prog_name: str,
+    complete_var: str,
+    instruction: str,
+) -> int:
+    """Perform shell completion for the given CLI program.
+
+    :param cli: Command being called.
+    :param ctx_args: Extra arguments to pass to
+        ``cli.make_context``.
+    :param prog_name: Name of the executable in the shell.
+    :param complete_var: Name of the environment variable that holds
+        the completion instruction.
+    :param instruction: Value of ``complete_var`` with the completion
+        instruction and shell, in the form ``instruction_shell``.
+    :return: Status code to exit with.
+    """
+    shell, _, instruction = instruction.partition("_")
+    comp_cls = get_completion_class(shell)
+
+    if comp_cls is None:
+        return 1
+
+    comp = comp_cls(cli, ctx_args, prog_name, complete_var)
+
+    if instruction == "source":
+        echo(comp.source())
+        return 0
+
+    if instruction == "complete":
+        echo(comp.complete())
+        return 0
+
+    return 1
+
+
+class CompletionItem:
+    """Represents a completion value and metadata about the value. The
+    default metadata is ``type`` to indicate special shell handling,
+    and ``help`` if a shell supports showing a help string next to the
+    value.
+
+    Arbitrary parameters can be passed when creating the object, and
+    accessed using ``item.attr``. If an attribute wasn't passed,
+    accessing it returns ``None``.
+
+    :param value: The completion suggestion.
+    :param type: Tells the shell script to provide special completion
+        support for the type. Click uses ``"dir"`` and ``"file"``.
+    :param help: String shown next to the value if supported.
+    :param kwargs: Arbitrary metadata. The built-in implementations
+        don't use this, but custom type completions paired with custom
+        shell support could use it.
+    """
+
+    __slots__ = ("value", "type", "help", "_info")
+
+    def __init__(
+        self,
+        value: t.Any,
+        type: str = "plain",
+        help: t.Optional[str] = None,
+        **kwargs: t.Any,
+    ) -> None:
+        self.value: t.Any = value
+        self.type: str = type
+        self.help: t.Optional[str] = help
+        self._info = kwargs
+
+    def __getattr__(self, name: str) -> t.Any:
+        return self._info.get(name)
+
+
+# Only Bash >= 4.4 has the nosort option.
+_SOURCE_BASH = """\
+%(complete_func)s() {
+    local IFS=$'\\n'
+    local response
+
+    response=$(env COMP_WORDS="${COMP_WORDS[*]}" COMP_CWORD=$COMP_CWORD \
+%(complete_var)s=bash_complete $1)
+
+    for completion in $response; do
+        IFS=',' read type value <<< "$completion"
+
+        if [[ $type == 'dir' ]]; then
+            COMPREPLY=()
+            compopt -o dirnames
+        elif [[ $type == 'file' ]]; then
+            COMPREPLY=()
+            compopt -o default
+        elif [[ $type == 'plain' ]]; then
+            COMPREPLY+=($value)
+        fi
+    done
+
+    return 0
+}
+
+%(complete_func)s_setup() {
+    complete -o nosort -F %(complete_func)s %(prog_name)s
+}
+
+%(complete_func)s_setup;
+"""
+
+_SOURCE_ZSH = """\
+#compdef %(prog_name)s
+
+%(complete_func)s() {
+    local -a completions
+    local -a completions_with_descriptions
+    local -a response
+    (( ! $+commands[%(prog_name)s] )) && return 1
+
+    response=("${(@f)$(env COMP_WORDS="${words[*]}" COMP_CWORD=$((CURRENT-1)) \
+%(complete_var)s=zsh_complete %(prog_name)s)}")
+
+    for type key descr in ${response}; do
+        if [[ "$type" == "plain" ]]; then
+            if [[ "$descr" == "_" ]]; then
+                completions+=("$key")
+            else
+                completions_with_descriptions+=("$key":"$descr")
+            fi
+        elif [[ "$type" == "dir" ]]; then
+            _path_files -/
+        elif [[ "$type" == "file" ]]; then
+            _path_files -f
+        fi
+    done
+
+    if [ -n "$completions_with_descriptions" ]; then
+        _describe -V unsorted completions_with_descriptions -U
+    fi
+
+    if [ -n "$completions" ]; then
+        compadd -U -V unsorted -a completions
+    fi
+}
+
+if [[ $zsh_eval_context[-1] == loadautofunc ]]; then
+    # autoload from fpath, call function directly
+    %(complete_func)s "$@"
+else
+    # eval/source/. command, register function for later
+    compdef %(complete_func)s %(prog_name)s
+fi
+"""
+
+_SOURCE_FISH = """\
+function %(complete_func)s;
+    set -l response (env %(complete_var)s=fish_complete COMP_WORDS=(commandline -cp) \
+COMP_CWORD=(commandline -t) %(prog_name)s);
+
+    for completion in $response;
+        set -l metadata (string split "," $completion);
+
+        if test $metadata[1] = "dir";
+            __fish_complete_directories $metadata[2];
+        else if test $metadata[1] = "file";
+            __fish_complete_path $metadata[2];
+        else if test $metadata[1] = "plain";
+            echo $metadata[2];
+        end;
+    end;
+end;
+
+complete --no-files --command %(prog_name)s --arguments \
+"(%(complete_func)s)";
+"""
+
+
+class ShellComplete:
+    """Base class for providing shell completion support. A subclass for
+    a given shell will override attributes and methods to implement the
+    completion instructions (``source`` and ``complete``).
+
+    :param cli: Command being called.
+    :param prog_name: Name of the executable in the shell.
+    :param complete_var: Name of the environment variable that holds
+        the completion instruction.
+
+    .. versionadded:: 8.0
+    """
+
+    name: t.ClassVar[str]
+    """Name to register the shell as with :func:`add_completion_class`.
+    This is used in completion instructions (``{name}_source`` and
+    ``{name}_complete``).
+    """
+
+    source_template: t.ClassVar[str]
+    """Completion script template formatted by :meth:`source`. This must
+    be provided by subclasses.
+    """
+
+    def __init__(
+        self,
+        cli: BaseCommand,
+        ctx_args: t.MutableMapping[str, t.Any],
+        prog_name: str,
+        complete_var: str,
+    ) -> None:
+        self.cli = cli
+        self.ctx_args = ctx_args
+        self.prog_name = prog_name
+        self.complete_var = complete_var
+
+    @property
+    def func_name(self) -> str:
+        """The name of the shell function defined by the completion
+        script.
+        """
+        safe_name = re.sub(r"\W*", "", self.prog_name.replace("-", "_"), flags=re.ASCII)
+        return f"_{safe_name}_completion"
+
+    def source_vars(self) -> t.Dict[str, t.Any]:
+        """Vars for formatting :attr:`source_template`.
+
+        By default this provides ``complete_func``, ``complete_var``,
+        and ``prog_name``.
+        """
+        return {
+            "complete_func": self.func_name,
+            "complete_var": self.complete_var,
+            "prog_name": self.prog_name,
+        }
+
+    def source(self) -> str:
+        """Produce the shell script that defines the completion
+        function. By default this ``%``-style formats
+        :attr:`source_template` with the dict returned by
+        :meth:`source_vars`.
+        """
+        return self.source_template % self.source_vars()
+
+    def get_completion_args(self) -> t.Tuple[t.List[str], str]:
+        """Use the env vars defined by the shell script to return a
+        tuple of ``args, incomplete``. This must be implemented by
+        subclasses.
+        """
+        raise NotImplementedError
+
+    def get_completions(
+        self, args: t.List[str], incomplete: str
+    ) -> t.List[CompletionItem]:
+        """Determine the context and last complete command or parameter
+        from the complete args. Call that object's ``shell_complete``
+        method to get the completions for the incomplete value.
+
+        :param args: List of complete args before the incomplete value.
+        :param incomplete: Value being completed. May be empty.
+        """
+        ctx = _resolve_context(self.cli, self.ctx_args, self.prog_name, args)
+        obj, incomplete = _resolve_incomplete(ctx, args, incomplete)
+        return obj.shell_complete(ctx, incomplete)
+
+    def format_completion(self, item: CompletionItem) -> str:
+        """Format a completion item into the form recognized by the
+        shell script. This must be implemented by subclasses.
+
+        :param item: Completion item to format.
+        """
+        raise NotImplementedError
+
+    def complete(self) -> str:
+        """Produce the completion data to send back to the shell.
+
+        By default this calls :meth:`get_completion_args`, gets the
+        completions, then calls :meth:`format_completion` for each
+        completion.
+        """
+        args, incomplete = self.get_completion_args()
+        completions = self.get_completions(args, incomplete)
+        out = [self.format_completion(item) for item in completions]
+        return "\n".join(out)
+
+
+class BashComplete(ShellComplete):
+    """Shell completion for Bash."""
+
+    name = "bash"
+    source_template = _SOURCE_BASH
+
+    @staticmethod
+    def _check_version() -> None:
+        import shutil
+        import subprocess
+
+        bash_exe = shutil.which("bash")
+
+        if bash_exe is None:
+            match = None
+        else:
+            output = subprocess.run(
+                [bash_exe, "--norc", "-c", 'echo "${BASH_VERSION}"'],
+                stdout=subprocess.PIPE,
+            )
+            match = re.search(r"^(\d+)\.(\d+)\.\d+", output.stdout.decode())
+
+        if match is not None:
+            major, minor = match.groups()
+
+            if major < "4" or major == "4" and minor < "4":
+                echo(
+                    _(
+                        "Shell completion is not supported for Bash"
+                        " versions older than 4.4."
+                    ),
+                    err=True,
+                )
+        else:
+            echo(
+                _("Couldn't detect Bash version, shell completion is not supported."),
+                err=True,
+            )
+
+    def source(self) -> str:
+        self._check_version()
+        return super().source()
+
+    def get_completion_args(self) -> t.Tuple[t.List[str], str]:
+        cwords = split_arg_string(os.environ["COMP_WORDS"])
+        cword = int(os.environ["COMP_CWORD"])
+        args = cwords[1:cword]
+
+        try:
+            incomplete = cwords[cword]
+        except IndexError:
+            incomplete = ""
+
+        return args, incomplete
+
+    def format_completion(self, item: CompletionItem) -> str:
+        return f"{item.type},{item.value}"
+
+
+class ZshComplete(ShellComplete):
+    """Shell completion for Zsh."""
+
+    name = "zsh"
+    source_template = _SOURCE_ZSH
+
+    def get_completion_args(self) -> t.Tuple[t.List[str], str]:
+        cwords = split_arg_string(os.environ["COMP_WORDS"])
+        cword = int(os.environ["COMP_CWORD"])
+        args = cwords[1:cword]
+
+        try:
+            incomplete = cwords[cword]
+        except IndexError:
+            incomplete = ""
+
+        return args, incomplete
+
+    def format_completion(self, item: CompletionItem) -> str:
+        return f"{item.type}\n{item.value}\n{item.help if item.help else '_'}"
+
+
+class FishComplete(ShellComplete):
+    """Shell completion for Fish."""
+
+    name = "fish"
+    source_template = _SOURCE_FISH
+
+    def get_completion_args(self) -> t.Tuple[t.List[str], str]:
+        cwords = split_arg_string(os.environ["COMP_WORDS"])
+        incomplete = os.environ["COMP_CWORD"]
+        args = cwords[1:]
+
+        # Fish stores the partial word in both COMP_WORDS and
+        # COMP_CWORD, remove it from complete args.
+        if incomplete and args and args[-1] == incomplete:
+            args.pop()
+
+        return args, incomplete
+
+    def format_completion(self, item: CompletionItem) -> str:
+        if item.help:
+            return f"{item.type},{item.value}\t{item.help}"
+
+        return f"{item.type},{item.value}"
+
+
+ShellCompleteType = t.TypeVar("ShellCompleteType", bound=t.Type[ShellComplete])
+
+
+_available_shells: t.Dict[str, t.Type[ShellComplete]] = {
+    "bash": BashComplete,
+    "fish": FishComplete,
+    "zsh": ZshComplete,
+}
+
+
+def add_completion_class(
+    cls: ShellCompleteType, name: t.Optional[str] = None
+) -> ShellCompleteType:
+    """Register a :class:`ShellComplete` subclass under the given name.
+    The name will be provided by the completion instruction environment
+    variable during completion.
+
+    :param cls: The completion class that will handle completion for the
+        shell.
+    :param name: Name to register the class under. Defaults to the
+        class's ``name`` attribute.
+    """
+    if name is None:
+        name = cls.name
+
+    _available_shells[name] = cls
+
+    return cls
+
+
+def get_completion_class(shell: str) -> t.Optional[t.Type[ShellComplete]]:
+    """Look up a registered :class:`ShellComplete` subclass by the name
+    provided by the completion instruction environment variable. If the
+    name isn't registered, returns ``None``.
+
+    :param shell: Name the class is registered under.
+    """
+    return _available_shells.get(shell)
+
+
+def _is_incomplete_argument(ctx: Context, param: Parameter) -> bool:
+    """Determine if the given parameter is an argument that can still
+    accept values.
+
+    :param ctx: Invocation context for the command represented by the
+        parsed complete args.
+    :param param: Argument object being checked.
+    """
+    if not isinstance(param, Argument):
+        return False
+
+    assert param.name is not None
+    # Will be None if expose_value is False.
+    value = ctx.params.get(param.name)
+    return (
+        param.nargs == -1
+        or ctx.get_parameter_source(param.name) is not ParameterSource.COMMANDLINE
+        or (
+            param.nargs > 1
+            and isinstance(value, (tuple, list))
+            and len(value) < param.nargs
+        )
+    )
+
+
+def _start_of_option(ctx: Context, value: str) -> bool:
+    """Check if the value looks like the start of an option."""
+    if not value:
+        return False
+
+    c = value[0]
+    return c in ctx._opt_prefixes
+
+
+def _is_incomplete_option(ctx: Context, args: t.List[str], param: Parameter) -> bool:
+    """Determine if the given parameter is an option that needs a value.
+
+    :param args: List of complete args before the incomplete value.
+    :param param: Option object being checked.
+    """
+    if not isinstance(param, Option):
+        return False
+
+    if param.is_flag or param.count:
+        return False
+
+    last_option = None
+
+    for index, arg in enumerate(reversed(args)):
+        if index + 1 > param.nargs:
+            break
+
+        if _start_of_option(ctx, arg):
+            last_option = arg
+
+    return last_option is not None and last_option in param.opts
+
+
+def _resolve_context(
+    cli: BaseCommand,
+    ctx_args: t.MutableMapping[str, t.Any],
+    prog_name: str,
+    args: t.List[str],
+) -> Context:
+    """Produce the context hierarchy starting with the command and
+    traversing the complete arguments. This only follows the commands,
+    it doesn't trigger input prompts or callbacks.
+
+    :param cli: Command being called.
+    :param prog_name: Name of the executable in the shell.
+    :param args: List of complete args before the incomplete value.
+    """
+    ctx_args["resilient_parsing"] = True
+    ctx = cli.make_context(prog_name, args.copy(), **ctx_args)
+    args = ctx.protected_args + ctx.args
+
+    while args:
+        command = ctx.command
+
+        if isinstance(command, MultiCommand):
+            if not command.chain:
+                name, cmd, args = command.resolve_command(ctx, args)
+
+                if cmd is None:
+                    return ctx
+
+                ctx = cmd.make_context(name, args, parent=ctx, resilient_parsing=True)
+                args = ctx.protected_args + ctx.args
+            else:
+                sub_ctx = ctx
+
+                while args:
+                    name, cmd, args = command.resolve_command(ctx, args)
+
+                    if cmd is None:
+                        return ctx
+
+                    sub_ctx = cmd.make_context(
+                        name,
+                        args,
+                        parent=ctx,
+                        allow_extra_args=True,
+                        allow_interspersed_args=False,
+                        resilient_parsing=True,
+                    )
+                    args = sub_ctx.args
+
+                ctx = sub_ctx
+                args = [*sub_ctx.protected_args, *sub_ctx.args]
+        else:
+            break
+
+    return ctx
+
+
+def _resolve_incomplete(
+    ctx: Context, args: t.List[str], incomplete: str
+) -> t.Tuple[t.Union[BaseCommand, Parameter], str]:
+    """Find the Click object that will handle the completion of the
+    incomplete value. Return the object and the incomplete value.
+
+    :param ctx: Invocation context for the command represented by
+        the parsed complete args.
+    :param args: List of complete args before the incomplete value.
+    :param incomplete: Value being completed. May be empty.
+    """
+    # Different shells treat an "=" between a long option name and
+    # value differently. Might keep the value joined, return the "="
+    # as a separate item, or return the split name and value. Always
+    # split and discard the "=" to make completion easier.
+    if incomplete == "=":
+        incomplete = ""
+    elif "=" in incomplete and _start_of_option(ctx, incomplete):
+        name, _, incomplete = incomplete.partition("=")
+        args.append(name)
+
+    # The "--" marker tells Click to stop treating values as options
+    # even if they start with the option character. If it hasn't been
+    # given and the incomplete arg looks like an option, the current
+    # command will provide option name completions.
+    if "--" not in args and _start_of_option(ctx, incomplete):
+        return ctx.command, incomplete
+
+    params = ctx.command.get_params(ctx)
+
+    # If the last complete arg is an option name with an incomplete
+    # value, the option will provide value completions.
+    for param in params:
+        if _is_incomplete_option(ctx, args, param):
+            return param, incomplete
+
+    # It's not an option name or value. The first argument without a
+    # parsed value will provide value completions.
+    for param in params:
+        if _is_incomplete_argument(ctx, param):
+            return param, incomplete
+
+    # There were no unparsed arguments, the command may be a group that
+    # will provide command name completions.
+    return ctx.command, incomplete
diff --git a/backend/venv/lib/python3.12/site-packages/click/termui.py b/backend/venv/lib/python3.12/site-packages/click/termui.py
new file mode 100644
index 00000000..c084f196
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/termui.py
@@ -0,0 +1,784 @@
+import inspect
+import io
+import itertools
+import sys
+import typing as t
+from gettext import gettext as _
+
+from ._compat import isatty
+from ._compat import strip_ansi
+from .exceptions import Abort
+from .exceptions import UsageError
+from .globals import resolve_color_default
+from .types import Choice
+from .types import convert_type
+from .types import ParamType
+from .utils import echo
+from .utils import LazyFile
+
+if t.TYPE_CHECKING:
+    from ._termui_impl import ProgressBar
+
+V = t.TypeVar("V")
+
+# The prompt functions to use.  The doc tools currently override these
+# functions to customize how they work.
+visible_prompt_func: t.Callable[[str], str] = input
+
+_ansi_colors = {
+    "black": 30,
+    "red": 31,
+    "green": 32,
+    "yellow": 33,
+    "blue": 34,
+    "magenta": 35,
+    "cyan": 36,
+    "white": 37,
+    "reset": 39,
+    "bright_black": 90,
+    "bright_red": 91,
+    "bright_green": 92,
+    "bright_yellow": 93,
+    "bright_blue": 94,
+    "bright_magenta": 95,
+    "bright_cyan": 96,
+    "bright_white": 97,
+}
+_ansi_reset_all = "\033[0m"
+
+
+def hidden_prompt_func(prompt: str) -> str:
+    import getpass
+
+    return getpass.getpass(prompt)
+
+
+def _build_prompt(
+    text: str,
+    suffix: str,
+    show_default: bool = False,
+    default: t.Optional[t.Any] = None,
+    show_choices: bool = True,
+    type: t.Optional[ParamType] = None,
+) -> str:
+    prompt = text
+    if type is not None and show_choices and isinstance(type, Choice):
+        prompt += f" ({', '.join(map(str, type.choices))})"
+    if default is not None and show_default:
+        prompt = f"{prompt} [{_format_default(default)}]"
+    return f"{prompt}{suffix}"
+
+
+def _format_default(default: t.Any) -> t.Any:
+    if isinstance(default, (io.IOBase, LazyFile)) and hasattr(default, "name"):
+        return default.name
+
+    return default
+
+
+def prompt(
+    text: str,
+    default: t.Optional[t.Any] = None,
+    hide_input: bool = False,
+    confirmation_prompt: t.Union[bool, str] = False,
+    type: t.Optional[t.Union[ParamType, t.Any]] = None,
+    value_proc: t.Optional[t.Callable[[str], t.Any]] = None,
+    prompt_suffix: str = ": ",
+    show_default: bool = True,
+    err: bool = False,
+    show_choices: bool = True,
+) -> t.Any:
+    """Prompts a user for input.  This is a convenience function that can
+    be used to prompt a user for input later.
+
+    If the user aborts the input by sending an interrupt signal, this
+    function will catch it and raise a :exc:`Abort` exception.
+
+    :param text: the text to show for the prompt.
+    :param default: the default value to use if no input happens.  If this
+                    is not given it will prompt until it's aborted.
+    :param hide_input: if this is set to true then the input value will
+                       be hidden.
+    :param confirmation_prompt: Prompt a second time to confirm the
+        value. Can be set to a string instead of ``True`` to customize
+        the message.
+    :param type: the type to use to check the value against.
+    :param value_proc: if this parameter is provided it's a function that
+                       is invoked instead of the type conversion to
+                       convert a value.
+    :param prompt_suffix: a suffix that should be added to the prompt.
+    :param show_default: shows or hides the default value in the prompt.
+    :param err: if set to true the file defaults to ``stderr`` instead of
+                ``stdout``, the same as with echo.
+    :param show_choices: Show or hide choices if the passed type is a Choice.
+                         For example if type is a Choice of either day or week,
+                         show_choices is true and text is "Group by" then the
+                         prompt will be "Group by (day, week): ".
+
+    .. versionadded:: 8.0
+        ``confirmation_prompt`` can be a custom string.
+
+    .. versionadded:: 7.0
+        Added the ``show_choices`` parameter.
+
+    .. versionadded:: 6.0
+        Added unicode support for cmd.exe on Windows.
+
+    .. versionadded:: 4.0
+        Added the `err` parameter.
+
+    """
+
+    def prompt_func(text: str) -> str:
+        f = hidden_prompt_func if hide_input else visible_prompt_func
+        try:
+            # Write the prompt separately so that we get nice
+            # coloring through colorama on Windows
+            echo(text.rstrip(" "), nl=False, err=err)
+            # Echo a space to stdout to work around an issue where
+            # readline causes backspace to clear the whole line.
+            return f(" ")
+        except (KeyboardInterrupt, EOFError):
+            # getpass doesn't print a newline if the user aborts input with ^C.
+            # Allegedly this behavior is inherited from getpass(3).
+            # A doc bug has been filed at https://bugs.python.org/issue24711
+            if hide_input:
+                echo(None, err=err)
+            raise Abort() from None
+
+    if value_proc is None:
+        value_proc = convert_type(type, default)
+
+    prompt = _build_prompt(
+        text, prompt_suffix, show_default, default, show_choices, type
+    )
+
+    if confirmation_prompt:
+        if confirmation_prompt is True:
+            confirmation_prompt = _("Repeat for confirmation")
+
+        confirmation_prompt = _build_prompt(confirmation_prompt, prompt_suffix)
+
+    while True:
+        while True:
+            value = prompt_func(prompt)
+            if value:
+                break
+            elif default is not None:
+                value = default
+                break
+        try:
+            result = value_proc(value)
+        except UsageError as e:
+            if hide_input:
+                echo(_("Error: The value you entered was invalid."), err=err)
+            else:
+                echo(_("Error: {e.message}").format(e=e), err=err)
+            continue
+        if not confirmation_prompt:
+            return result
+        while True:
+            value2 = prompt_func(confirmation_prompt)
+            is_empty = not value and not value2
+            if value2 or is_empty:
+                break
+        if value == value2:
+            return result
+        echo(_("Error: The two entered values do not match."), err=err)
+
+
+def confirm(
+    text: str,
+    default: t.Optional[bool] = False,
+    abort: bool = False,
+    prompt_suffix: str = ": ",
+    show_default: bool = True,
+    err: bool = False,
+) -> bool:
+    """Prompts for confirmation (yes/no question).
+
+    If the user aborts the input by sending a interrupt signal this
+    function will catch it and raise a :exc:`Abort` exception.
+
+    :param text: the question to ask.
+    :param default: The default value to use when no input is given. If
+        ``None``, repeat until input is given.
+    :param abort: if this is set to `True` a negative answer aborts the
+                  exception by raising :exc:`Abort`.
+    :param prompt_suffix: a suffix that should be added to the prompt.
+    :param show_default: shows or hides the default value in the prompt.
+    :param err: if set to true the file defaults to ``stderr`` instead of
+                ``stdout``, the same as with echo.
+
+    .. versionchanged:: 8.0
+        Repeat until input is given if ``default`` is ``None``.
+
+    .. versionadded:: 4.0
+        Added the ``err`` parameter.
+    """
+    prompt = _build_prompt(
+        text,
+        prompt_suffix,
+        show_default,
+        "y/n" if default is None else ("Y/n" if default else "y/N"),
+    )
+
+    while True:
+        try:
+            # Write the prompt separately so that we get nice
+            # coloring through colorama on Windows
+            echo(prompt.rstrip(" "), nl=False, err=err)
+            # Echo a space to stdout to work around an issue where
+            # readline causes backspace to clear the whole line.
+            value = visible_prompt_func(" ").lower().strip()
+        except (KeyboardInterrupt, EOFError):
+            raise Abort() from None
+        if value in ("y", "yes"):
+            rv = True
+        elif value in ("n", "no"):
+            rv = False
+        elif default is not None and value == "":
+            rv = default
+        else:
+            echo(_("Error: invalid input"), err=err)
+            continue
+        break
+    if abort and not rv:
+        raise Abort()
+    return rv
+
+
+def echo_via_pager(
+    text_or_generator: t.Union[t.Iterable[str], t.Callable[[], t.Iterable[str]], str],
+    color: t.Optional[bool] = None,
+) -> None:
+    """This function takes a text and shows it via an environment specific
+    pager on stdout.
+
+    .. versionchanged:: 3.0
+       Added the `color` flag.
+
+    :param text_or_generator: the text to page, or alternatively, a
+                              generator emitting the text to page.
+    :param color: controls if the pager supports ANSI colors or not.  The
+                  default is autodetection.
+    """
+    color = resolve_color_default(color)
+
+    if inspect.isgeneratorfunction(text_or_generator):
+        i = t.cast(t.Callable[[], t.Iterable[str]], text_or_generator)()
+    elif isinstance(text_or_generator, str):
+        i = [text_or_generator]
+    else:
+        i = iter(t.cast(t.Iterable[str], text_or_generator))
+
+    # convert every element of i to a text type if necessary
+    text_generator = (el if isinstance(el, str) else str(el) for el in i)
+
+    from ._termui_impl import pager
+
+    return pager(itertools.chain(text_generator, "\n"), color)
+
+
+def progressbar(
+    iterable: t.Optional[t.Iterable[V]] = None,
+    length: t.Optional[int] = None,
+    label: t.Optional[str] = None,
+    show_eta: bool = True,
+    show_percent: t.Optional[bool] = None,
+    show_pos: bool = False,
+    item_show_func: t.Optional[t.Callable[[t.Optional[V]], t.Optional[str]]] = None,
+    fill_char: str = "#",
+    empty_char: str = "-",
+    bar_template: str = "%(label)s  [%(bar)s]  %(info)s",
+    info_sep: str = "  ",
+    width: int = 36,
+    file: t.Optional[t.TextIO] = None,
+    color: t.Optional[bool] = None,
+    update_min_steps: int = 1,
+) -> "ProgressBar[V]":
+    """This function creates an iterable context manager that can be used
+    to iterate over something while showing a progress bar.  It will
+    either iterate over the `iterable` or `length` items (that are counted
+    up).  While iteration happens, this function will print a rendered
+    progress bar to the given `file` (defaults to stdout) and will attempt
+    to calculate remaining time and more.  By default, this progress bar
+    will not be rendered if the file is not a terminal.
+
+    The context manager creates the progress bar.  When the context
+    manager is entered the progress bar is already created.  With every
+    iteration over the progress bar, the iterable passed to the bar is
+    advanced and the bar is updated.  When the context manager exits,
+    a newline is printed and the progress bar is finalized on screen.
+
+    Note: The progress bar is currently designed for use cases where the
+    total progress can be expected to take at least several seconds.
+    Because of this, the ProgressBar class object won't display
+    progress that is considered too fast, and progress where the time
+    between steps is less than a second.
+
+    No printing must happen or the progress bar will be unintentionally
+    destroyed.
+
+    Example usage::
+
+        with progressbar(items) as bar:
+            for item in bar:
+                do_something_with(item)
+
+    Alternatively, if no iterable is specified, one can manually update the
+    progress bar through the `update()` method instead of directly
+    iterating over the progress bar.  The update method accepts the number
+    of steps to increment the bar with::
+
+        with progressbar(length=chunks.total_bytes) as bar:
+            for chunk in chunks:
+                process_chunk(chunk)
+                bar.update(chunks.bytes)
+
+    The ``update()`` method also takes an optional value specifying the
+    ``current_item`` at the new position. This is useful when used
+    together with ``item_show_func`` to customize the output for each
+    manual step::
+
+        with click.progressbar(
+            length=total_size,
+            label='Unzipping archive',
+            item_show_func=lambda a: a.filename
+        ) as bar:
+            for archive in zip_file:
+                archive.extract()
+                bar.update(archive.size, archive)
+
+    :param iterable: an iterable to iterate over.  If not provided the length
+                     is required.
+    :param length: the number of items to iterate over.  By default the
+                   progressbar will attempt to ask the iterator about its
+                   length, which might or might not work.  If an iterable is
+                   also provided this parameter can be used to override the
+                   length.  If an iterable is not provided the progress bar
+                   will iterate over a range of that length.
+    :param label: the label to show next to the progress bar.
+    :param show_eta: enables or disables the estimated time display.  This is
+                     automatically disabled if the length cannot be
+                     determined.
+    :param show_percent: enables or disables the percentage display.  The
+                         default is `True` if the iterable has a length or
+                         `False` if not.
+    :param show_pos: enables or disables the absolute position display.  The
+                     default is `False`.
+    :param item_show_func: A function called with the current item which
+        can return a string to show next to the progress bar. If the
+        function returns ``None`` nothing is shown. The current item can
+        be ``None``, such as when entering and exiting the bar.
+    :param fill_char: the character to use to show the filled part of the
+                      progress bar.
+    :param empty_char: the character to use to show the non-filled part of
+                       the progress bar.
+    :param bar_template: the format string to use as template for the bar.
+                         The parameters in it are ``label`` for the label,
+                         ``bar`` for the progress bar and ``info`` for the
+                         info section.
+    :param info_sep: the separator between multiple info items (eta etc.)
+    :param width: the width of the progress bar in characters, 0 means full
+                  terminal width
+    :param file: The file to write to. If this is not a terminal then
+        only the label is printed.
+    :param color: controls if the terminal supports ANSI colors or not.  The
+                  default is autodetection.  This is only needed if ANSI
+                  codes are included anywhere in the progress bar output
+                  which is not the case by default.
+    :param update_min_steps: Render only when this many updates have
+        completed. This allows tuning for very fast iterators.
+
+    .. versionchanged:: 8.0
+        Output is shown even if execution time is less than 0.5 seconds.
+
+    .. versionchanged:: 8.0
+        ``item_show_func`` shows the current item, not the previous one.
+
+    .. versionchanged:: 8.0
+        Labels are echoed if the output is not a TTY. Reverts a change
+        in 7.0 that removed all output.
+
+    .. versionadded:: 8.0
+       Added the ``update_min_steps`` parameter.
+
+    .. versionchanged:: 4.0
+        Added the ``color`` parameter. Added the ``update`` method to
+        the object.
+
+    .. versionadded:: 2.0
+    """
+    from ._termui_impl import ProgressBar
+
+    color = resolve_color_default(color)
+    return ProgressBar(
+        iterable=iterable,
+        length=length,
+        show_eta=show_eta,
+        show_percent=show_percent,
+        show_pos=show_pos,
+        item_show_func=item_show_func,
+        fill_char=fill_char,
+        empty_char=empty_char,
+        bar_template=bar_template,
+        info_sep=info_sep,
+        file=file,
+        label=label,
+        width=width,
+        color=color,
+        update_min_steps=update_min_steps,
+    )
+
+
+def clear() -> None:
+    """Clears the terminal screen.  This will have the effect of clearing
+    the whole visible space of the terminal and moving the cursor to the
+    top left.  This does not do anything if not connected to a terminal.
+
+    .. versionadded:: 2.0
+    """
+    if not isatty(sys.stdout):
+        return
+
+    # ANSI escape \033[2J clears the screen, \033[1;1H moves the cursor
+    echo("\033[2J\033[1;1H", nl=False)
+
+
+def _interpret_color(
+    color: t.Union[int, t.Tuple[int, int, int], str], offset: int = 0
+) -> str:
+    if isinstance(color, int):
+        return f"{38 + offset};5;{color:d}"
+
+    if isinstance(color, (tuple, list)):
+        r, g, b = color
+        return f"{38 + offset};2;{r:d};{g:d};{b:d}"
+
+    return str(_ansi_colors[color] + offset)
+
+
+def style(
+    text: t.Any,
+    fg: t.Optional[t.Union[int, t.Tuple[int, int, int], str]] = None,
+    bg: t.Optional[t.Union[int, t.Tuple[int, int, int], str]] = None,
+    bold: t.Optional[bool] = None,
+    dim: t.Optional[bool] = None,
+    underline: t.Optional[bool] = None,
+    overline: t.Optional[bool] = None,
+    italic: t.Optional[bool] = None,
+    blink: t.Optional[bool] = None,
+    reverse: t.Optional[bool] = None,
+    strikethrough: t.Optional[bool] = None,
+    reset: bool = True,
+) -> str:
+    """Styles a text with ANSI styles and returns the new string.  By
+    default the styling is self contained which means that at the end
+    of the string a reset code is issued.  This can be prevented by
+    passing ``reset=False``.
+
+    Examples::
+
+        click.echo(click.style('Hello World!', fg='green'))
+        click.echo(click.style('ATTENTION!', blink=True))
+        click.echo(click.style('Some things', reverse=True, fg='cyan'))
+        click.echo(click.style('More colors', fg=(255, 12, 128), bg=117))
+
+    Supported color names:
+
+    * ``black`` (might be a gray)
+    * ``red``
+    * ``green``
+    * ``yellow`` (might be an orange)
+    * ``blue``
+    * ``magenta``
+    * ``cyan``
+    * ``white`` (might be light gray)
+    * ``bright_black``
+    * ``bright_red``
+    * ``bright_green``
+    * ``bright_yellow``
+    * ``bright_blue``
+    * ``bright_magenta``
+    * ``bright_cyan``
+    * ``bright_white``
+    * ``reset`` (reset the color code only)
+
+    If the terminal supports it, color may also be specified as:
+
+    -   An integer in the interval [0, 255]. The terminal must support
+        8-bit/256-color mode.
+    -   An RGB tuple of three integers in [0, 255]. The terminal must
+        support 24-bit/true-color mode.
+
+    See https://en.wikipedia.org/wiki/ANSI_color and
+    https://gist.github.com/XVilka/8346728 for more information.
+
+    :param text: the string to style with ansi codes.
+    :param fg: if provided this will become the foreground color.
+    :param bg: if provided this will become the background color.
+    :param bold: if provided this will enable or disable bold mode.
+    :param dim: if provided this will enable or disable dim mode.  This is
+                badly supported.
+    :param underline: if provided this will enable or disable underline.
+    :param overline: if provided this will enable or disable overline.
+    :param italic: if provided this will enable or disable italic.
+    :param blink: if provided this will enable or disable blinking.
+    :param reverse: if provided this will enable or disable inverse
+                    rendering (foreground becomes background and the
+                    other way round).
+    :param strikethrough: if provided this will enable or disable
+        striking through text.
+    :param reset: by default a reset-all code is added at the end of the
+                  string which means that styles do not carry over.  This
+                  can be disabled to compose styles.
+
+    .. versionchanged:: 8.0
+        A non-string ``message`` is converted to a string.
+
+    .. versionchanged:: 8.0
+       Added support for 256 and RGB color codes.
+
+    .. versionchanged:: 8.0
+        Added the ``strikethrough``, ``italic``, and ``overline``
+        parameters.
+
+    .. versionchanged:: 7.0
+        Added support for bright colors.
+
+    .. versionadded:: 2.0
+    """
+    if not isinstance(text, str):
+        text = str(text)
+
+    bits = []
+
+    if fg:
+        try:
+            bits.append(f"\033[{_interpret_color(fg)}m")
+        except KeyError:
+            raise TypeError(f"Unknown color {fg!r}") from None
+
+    if bg:
+        try:
+            bits.append(f"\033[{_interpret_color(bg, 10)}m")
+        except KeyError:
+            raise TypeError(f"Unknown color {bg!r}") from None
+
+    if bold is not None:
+        bits.append(f"\033[{1 if bold else 22}m")
+    if dim is not None:
+        bits.append(f"\033[{2 if dim else 22}m")
+    if underline is not None:
+        bits.append(f"\033[{4 if underline else 24}m")
+    if overline is not None:
+        bits.append(f"\033[{53 if overline else 55}m")
+    if italic is not None:
+        bits.append(f"\033[{3 if italic else 23}m")
+    if blink is not None:
+        bits.append(f"\033[{5 if blink else 25}m")
+    if reverse is not None:
+        bits.append(f"\033[{7 if reverse else 27}m")
+    if strikethrough is not None:
+        bits.append(f"\033[{9 if strikethrough else 29}m")
+    bits.append(text)
+    if reset:
+        bits.append(_ansi_reset_all)
+    return "".join(bits)
+
+
+def unstyle(text: str) -> str:
+    """Removes ANSI styling information from a string.  Usually it's not
+    necessary to use this function as Click's echo function will
+    automatically remove styling if necessary.
+
+    .. versionadded:: 2.0
+
+    :param text: the text to remove style information from.
+    """
+    return strip_ansi(text)
+
+
+def secho(
+    message: t.Optional[t.Any] = None,
+    file: t.Optional[t.IO[t.AnyStr]] = None,
+    nl: bool = True,
+    err: bool = False,
+    color: t.Optional[bool] = None,
+    **styles: t.Any,
+) -> None:
+    """This function combines :func:`echo` and :func:`style` into one
+    call.  As such the following two calls are the same::
+
+        click.secho('Hello World!', fg='green')
+        click.echo(click.style('Hello World!', fg='green'))
+
+    All keyword arguments are forwarded to the underlying functions
+    depending on which one they go with.
+
+    Non-string types will be converted to :class:`str`. However,
+    :class:`bytes` are passed directly to :meth:`echo` without applying
+    style. If you want to style bytes that represent text, call
+    :meth:`bytes.decode` first.
+
+    .. versionchanged:: 8.0
+        A non-string ``message`` is converted to a string. Bytes are
+        passed through without style applied.
+
+    .. versionadded:: 2.0
+    """
+    if message is not None and not isinstance(message, (bytes, bytearray)):
+        message = style(message, **styles)
+
+    return echo(message, file=file, nl=nl, err=err, color=color)
+
+
+def edit(
+    text: t.Optional[t.AnyStr] = None,
+    editor: t.Optional[str] = None,
+    env: t.Optional[t.Mapping[str, str]] = None,
+    require_save: bool = True,
+    extension: str = ".txt",
+    filename: t.Optional[str] = None,
+) -> t.Optional[t.AnyStr]:
+    r"""Edits the given text in the defined editor.  If an editor is given
+    (should be the full path to the executable but the regular operating
+    system search path is used for finding the executable) it overrides
+    the detected editor.  Optionally, some environment variables can be
+    used.  If the editor is closed without changes, `None` is returned.  In
+    case a file is edited directly the return value is always `None` and
+    `require_save` and `extension` are ignored.
+
+    If the editor cannot be opened a :exc:`UsageError` is raised.
+
+    Note for Windows: to simplify cross-platform usage, the newlines are
+    automatically converted from POSIX to Windows and vice versa.  As such,
+    the message here will have ``\n`` as newline markers.
+
+    :param text: the text to edit.
+    :param editor: optionally the editor to use.  Defaults to automatic
+                   detection.
+    :param env: environment variables to forward to the editor.
+    :param require_save: if this is true, then not saving in the editor
+                         will make the return value become `None`.
+    :param extension: the extension to tell the editor about.  This defaults
+                      to `.txt` but changing this might change syntax
+                      highlighting.
+    :param filename: if provided it will edit this file instead of the
+                     provided text contents.  It will not use a temporary
+                     file as an indirection in that case.
+    """
+    from ._termui_impl import Editor
+
+    ed = Editor(editor=editor, env=env, require_save=require_save, extension=extension)
+
+    if filename is None:
+        return ed.edit(text)
+
+    ed.edit_file(filename)
+    return None
+
+
+def launch(url: str, wait: bool = False, locate: bool = False) -> int:
+    """This function launches the given URL (or filename) in the default
+    viewer application for this file type.  If this is an executable, it
+    might launch the executable in a new session.  The return value is
+    the exit code of the launched application.  Usually, ``0`` indicates
+    success.
+
+    Examples::
+
+        click.launch('https://click.palletsprojects.com/')
+        click.launch('/my/downloaded/file', locate=True)
+
+    .. versionadded:: 2.0
+
+    :param url: URL or filename of the thing to launch.
+    :param wait: Wait for the program to exit before returning. This
+        only works if the launched program blocks. In particular,
+        ``xdg-open`` on Linux does not block.
+    :param locate: if this is set to `True` then instead of launching the
+                   application associated with the URL it will attempt to
+                   launch a file manager with the file located.  This
+                   might have weird effects if the URL does not point to
+                   the filesystem.
+    """
+    from ._termui_impl import open_url
+
+    return open_url(url, wait=wait, locate=locate)
+
+
+# If this is provided, getchar() calls into this instead.  This is used
+# for unittesting purposes.
+_getchar: t.Optional[t.Callable[[bool], str]] = None
+
+
+def getchar(echo: bool = False) -> str:
+    """Fetches a single character from the terminal and returns it.  This
+    will always return a unicode character and under certain rare
+    circumstances this might return more than one character.  The
+    situations which more than one character is returned is when for
+    whatever reason multiple characters end up in the terminal buffer or
+    standard input was not actually a terminal.
+
+    Note that this will always read from the terminal, even if something
+    is piped into the standard input.
+
+    Note for Windows: in rare cases when typing non-ASCII characters, this
+    function might wait for a second character and then return both at once.
+    This is because certain Unicode characters look like special-key markers.
+
+    .. versionadded:: 2.0
+
+    :param echo: if set to `True`, the character read will also show up on
+                 the terminal.  The default is to not show it.
+    """
+    global _getchar
+
+    if _getchar is None:
+        from ._termui_impl import getchar as f
+
+        _getchar = f
+
+    return _getchar(echo)
+
+
+def raw_terminal() -> t.ContextManager[int]:
+    from ._termui_impl import raw_terminal as f
+
+    return f()
+
+
+def pause(info: t.Optional[str] = None, err: bool = False) -> None:
+    """This command stops execution and waits for the user to press any
+    key to continue.  This is similar to the Windows batch "pause"
+    command.  If the program is not run through a terminal, this command
+    will instead do nothing.
+
+    .. versionadded:: 2.0
+
+    .. versionadded:: 4.0
+       Added the `err` parameter.
+
+    :param info: The message to print before pausing. Defaults to
+        ``"Press any key to continue..."``.
+    :param err: if set to message goes to ``stderr`` instead of
+                ``stdout``, the same as with echo.
+    """
+    if not isatty(sys.stdin) or not isatty(sys.stdout):
+        return
+
+    if info is None:
+        info = _("Press any key to continue...")
+
+    try:
+        if info:
+            echo(info, nl=False, err=err)
+        try:
+            getchar()
+        except (KeyboardInterrupt, EOFError):
+            pass
+    finally:
+        if info:
+            echo(err=err)
diff --git a/backend/venv/lib/python3.12/site-packages/click/testing.py b/backend/venv/lib/python3.12/site-packages/click/testing.py
new file mode 100644
index 00000000..772b2159
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/testing.py
@@ -0,0 +1,483 @@
+import contextlib
+import io
+import os
+import shlex
+import shutil
+import sys
+import tempfile
+import typing as t
+from types import TracebackType
+
+from . import _compat
+from . import formatting
+from . import termui
+from . import utils
+from ._compat import _find_binary_reader
+
+if t.TYPE_CHECKING:
+    from .core import BaseCommand
+
+
+class EchoingStdin:
+    def __init__(self, input: t.BinaryIO, output: t.BinaryIO) -> None:
+        self._input = input
+        self._output = output
+        self._paused = False
+
+    def __getattr__(self, x: str) -> t.Any:
+        return getattr(self._input, x)
+
+    def _echo(self, rv: bytes) -> bytes:
+        if not self._paused:
+            self._output.write(rv)
+
+        return rv
+
+    def read(self, n: int = -1) -> bytes:
+        return self._echo(self._input.read(n))
+
+    def read1(self, n: int = -1) -> bytes:
+        return self._echo(self._input.read1(n))  # type: ignore
+
+    def readline(self, n: int = -1) -> bytes:
+        return self._echo(self._input.readline(n))
+
+    def readlines(self) -> t.List[bytes]:
+        return [self._echo(x) for x in self._input.readlines()]
+
+    def __iter__(self) -> t.Iterator[bytes]:
+        return iter(self._echo(x) for x in self._input)
+
+    def __repr__(self) -> str:
+        return repr(self._input)
+
+
+@contextlib.contextmanager
+def _pause_echo(stream: t.Optional[EchoingStdin]) -> t.Iterator[None]:
+    if stream is None:
+        yield
+    else:
+        stream._paused = True
+        yield
+        stream._paused = False
+
+
+class _NamedTextIOWrapper(io.TextIOWrapper):
+    def __init__(
+        self, buffer: t.BinaryIO, name: str, mode: str, **kwargs: t.Any
+    ) -> None:
+        super().__init__(buffer, **kwargs)
+        self._name = name
+        self._mode = mode
+
+    @property
+    def name(self) -> str:
+        return self._name
+
+    @property
+    def mode(self) -> str:
+        return self._mode
+
+
+def make_input_stream(
+    input: t.Optional[t.Union[str, bytes, t.IO[t.Any]]], charset: str
+) -> t.BinaryIO:
+    # Is already an input stream.
+    if hasattr(input, "read"):
+        rv = _find_binary_reader(t.cast(t.IO[t.Any], input))
+
+        if rv is not None:
+            return rv
+
+        raise TypeError("Could not find binary reader for input stream.")
+
+    if input is None:
+        input = b""
+    elif isinstance(input, str):
+        input = input.encode(charset)
+
+    return io.BytesIO(input)
+
+
+class Result:
+    """Holds the captured result of an invoked CLI script."""
+
+    def __init__(
+        self,
+        runner: "CliRunner",
+        stdout_bytes: bytes,
+        stderr_bytes: t.Optional[bytes],
+        return_value: t.Any,
+        exit_code: int,
+        exception: t.Optional[BaseException],
+        exc_info: t.Optional[
+            t.Tuple[t.Type[BaseException], BaseException, TracebackType]
+        ] = None,
+    ):
+        #: The runner that created the result
+        self.runner = runner
+        #: The standard output as bytes.
+        self.stdout_bytes = stdout_bytes
+        #: The standard error as bytes, or None if not available
+        self.stderr_bytes = stderr_bytes
+        #: The value returned from the invoked command.
+        #:
+        #: .. versionadded:: 8.0
+        self.return_value = return_value
+        #: The exit code as integer.
+        self.exit_code = exit_code
+        #: The exception that happened if one did.
+        self.exception = exception
+        #: The traceback
+        self.exc_info = exc_info
+
+    @property
+    def output(self) -> str:
+        """The (standard) output as unicode string."""
+        return self.stdout
+
+    @property
+    def stdout(self) -> str:
+        """The standard output as unicode string."""
+        return self.stdout_bytes.decode(self.runner.charset, "replace").replace(
+            "\r\n", "\n"
+        )
+
+    @property
+    def stderr(self) -> str:
+        """The standard error as unicode string."""
+        if self.stderr_bytes is None:
+            raise ValueError("stderr not separately captured")
+        return self.stderr_bytes.decode(self.runner.charset, "replace").replace(
+            "\r\n", "\n"
+        )
+
+    def __repr__(self) -> str:
+        exc_str = repr(self.exception) if self.exception else "okay"
+        return f"<{type(self).__name__} {exc_str}>"
+
+
+class CliRunner:
+    """The CLI runner provides functionality to invoke a Click command line
+    script for unittesting purposes in a isolated environment.  This only
+    works in single-threaded systems without any concurrency as it changes the
+    global interpreter state.
+
+    :param charset: the character set for the input and output data.
+    :param env: a dictionary with environment variables for overriding.
+    :param echo_stdin: if this is set to `True`, then reading from stdin writes
+                       to stdout.  This is useful for showing examples in
+                       some circumstances.  Note that regular prompts
+                       will automatically echo the input.
+    :param mix_stderr: if this is set to `False`, then stdout and stderr are
+                       preserved as independent streams.  This is useful for
+                       Unix-philosophy apps that have predictable stdout and
+                       noisy stderr, such that each may be measured
+                       independently
+    """
+
+    def __init__(
+        self,
+        charset: str = "utf-8",
+        env: t.Optional[t.Mapping[str, t.Optional[str]]] = None,
+        echo_stdin: bool = False,
+        mix_stderr: bool = True,
+    ) -> None:
+        self.charset = charset
+        self.env: t.Mapping[str, t.Optional[str]] = env or {}
+        self.echo_stdin = echo_stdin
+        self.mix_stderr = mix_stderr
+
+    def get_default_prog_name(self, cli: "BaseCommand") -> str:
+        """Given a command object it will return the default program name
+        for it.  The default is the `name` attribute or ``"root"`` if not
+        set.
+        """
+        return cli.name or "root"
+
+    def make_env(
+        self, overrides: t.Optional[t.Mapping[str, t.Optional[str]]] = None
+    ) -> t.Mapping[str, t.Optional[str]]:
+        """Returns the environment overrides for invoking a script."""
+        rv = dict(self.env)
+        if overrides:
+            rv.update(overrides)
+        return rv
+
+    @contextlib.contextmanager
+    def isolation(
+        self,
+        input: t.Optional[t.Union[str, bytes, t.IO[t.Any]]] = None,
+        env: t.Optional[t.Mapping[str, t.Optional[str]]] = None,
+        color: bool = False,
+    ) -> t.Iterator[t.Tuple[io.BytesIO, t.Optional[io.BytesIO]]]:
+        """A context manager that sets up the isolation for invoking of a
+        command line tool.  This sets up stdin with the given input data
+        and `os.environ` with the overrides from the given dictionary.
+        This also rebinds some internals in Click to be mocked (like the
+        prompt functionality).
+
+        This is automatically done in the :meth:`invoke` method.
+
+        :param input: the input stream to put into sys.stdin.
+        :param env: the environment overrides as dictionary.
+        :param color: whether the output should contain color codes. The
+                      application can still override this explicitly.
+
+        .. versionchanged:: 8.0
+            ``stderr`` is opened with ``errors="backslashreplace"``
+            instead of the default ``"strict"``.
+
+        .. versionchanged:: 4.0
+            Added the ``color`` parameter.
+        """
+        bytes_input = make_input_stream(input, self.charset)
+        echo_input = None
+
+        old_stdin = sys.stdin
+        old_stdout = sys.stdout
+        old_stderr = sys.stderr
+        old_forced_width = formatting.FORCED_WIDTH
+        formatting.FORCED_WIDTH = 80
+
+        env = self.make_env(env)
+
+        bytes_output = io.BytesIO()
+
+        if self.echo_stdin:
+            bytes_input = echo_input = t.cast(
+                t.BinaryIO, EchoingStdin(bytes_input, bytes_output)
+            )
+
+        sys.stdin = text_input = _NamedTextIOWrapper(
+            bytes_input, encoding=self.charset, name="<stdin>", mode="r"
+        )
+
+        if self.echo_stdin:
+            # Force unbuffered reads, otherwise TextIOWrapper reads a
+            # large chunk which is echoed early.
+            text_input._CHUNK_SIZE = 1  # type: ignore
+
+        sys.stdout = _NamedTextIOWrapper(
+            bytes_output, encoding=self.charset, name="<stdout>", mode="w"
+        )
+
+        bytes_error = None
+        if self.mix_stderr:
+            sys.stderr = sys.stdout
+        else:
+            bytes_error = io.BytesIO()
+            sys.stderr = _NamedTextIOWrapper(
+                bytes_error,
+                encoding=self.charset,
+                name="<stderr>",
+                mode="w",
+                errors="backslashreplace",
+            )
+
+        @_pause_echo(echo_input)  # type: ignore
+        def visible_input(prompt: t.Optional[str] = None) -> str:
+            sys.stdout.write(prompt or "")
+            val = text_input.readline().rstrip("\r\n")
+            sys.stdout.write(f"{val}\n")
+            sys.stdout.flush()
+            return val
+
+        @_pause_echo(echo_input)  # type: ignore
+        def hidden_input(prompt: t.Optional[str] = None) -> str:
+            sys.stdout.write(f"{prompt or ''}\n")
+            sys.stdout.flush()
+            return text_input.readline().rstrip("\r\n")
+
+        @_pause_echo(echo_input)  # type: ignore
+        def _getchar(echo: bool) -> str:
+            char = sys.stdin.read(1)
+
+            if echo:
+                sys.stdout.write(char)
+
+            sys.stdout.flush()
+            return char
+
+        default_color = color
+
+        def should_strip_ansi(
+            stream: t.Optional[t.IO[t.Any]] = None, color: t.Optional[bool] = None
+        ) -> bool:
+            if color is None:
+                return not default_color
+            return not color
+
+        old_visible_prompt_func = termui.visible_prompt_func
+        old_hidden_prompt_func = termui.hidden_prompt_func
+        old__getchar_func = termui._getchar
+        old_should_strip_ansi = utils.should_strip_ansi  # type: ignore
+        old__compat_should_strip_ansi = _compat.should_strip_ansi
+        termui.visible_prompt_func = visible_input
+        termui.hidden_prompt_func = hidden_input
+        termui._getchar = _getchar
+        utils.should_strip_ansi = should_strip_ansi  # type: ignore
+        _compat.should_strip_ansi = should_strip_ansi
+
+        old_env = {}
+        try:
+            for key, value in env.items():
+                old_env[key] = os.environ.get(key)
+                if value is None:
+                    try:
+                        del os.environ[key]
+                    except Exception:
+                        pass
+                else:
+                    os.environ[key] = value
+            yield (bytes_output, bytes_error)
+        finally:
+            for key, value in old_env.items():
+                if value is None:
+                    try:
+                        del os.environ[key]
+                    except Exception:
+                        pass
+                else:
+                    os.environ[key] = value
+            sys.stdout = old_stdout
+            sys.stderr = old_stderr
+            sys.stdin = old_stdin
+            termui.visible_prompt_func = old_visible_prompt_func
+            termui.hidden_prompt_func = old_hidden_prompt_func
+            termui._getchar = old__getchar_func
+            utils.should_strip_ansi = old_should_strip_ansi  # type: ignore
+            _compat.should_strip_ansi = old__compat_should_strip_ansi
+            formatting.FORCED_WIDTH = old_forced_width
+
+    def invoke(
+        self,
+        cli: "BaseCommand",
+        args: t.Optional[t.Union[str, t.Sequence[str]]] = None,
+        input: t.Optional[t.Union[str, bytes, t.IO[t.Any]]] = None,
+        env: t.Optional[t.Mapping[str, t.Optional[str]]] = None,
+        catch_exceptions: bool = True,
+        color: bool = False,
+        **extra: t.Any,
+    ) -> Result:
+        """Invokes a command in an isolated environment.  The arguments are
+        forwarded directly to the command line script, the `extra` keyword
+        arguments are passed to the :meth:`~clickpkg.Command.main` function of
+        the command.
+
+        This returns a :class:`Result` object.
+
+        :param cli: the command to invoke
+        :param args: the arguments to invoke. It may be given as an iterable
+                     or a string. When given as string it will be interpreted
+                     as a Unix shell command. More details at
+                     :func:`shlex.split`.
+        :param input: the input data for `sys.stdin`.
+        :param env: the environment overrides.
+        :param catch_exceptions: Whether to catch any other exceptions than
+                                 ``SystemExit``.
+        :param extra: the keyword arguments to pass to :meth:`main`.
+        :param color: whether the output should contain color codes. The
+                      application can still override this explicitly.
+
+        .. versionchanged:: 8.0
+            The result object has the ``return_value`` attribute with
+            the value returned from the invoked command.
+
+        .. versionchanged:: 4.0
+            Added the ``color`` parameter.
+
+        .. versionchanged:: 3.0
+            Added the ``catch_exceptions`` parameter.
+
+        .. versionchanged:: 3.0
+            The result object has the ``exc_info`` attribute with the
+            traceback if available.
+        """
+        exc_info = None
+        with self.isolation(input=input, env=env, color=color) as outstreams:
+            return_value = None
+            exception: t.Optional[BaseException] = None
+            exit_code = 0
+
+            if isinstance(args, str):
+                args = shlex.split(args)
+
+            try:
+                prog_name = extra.pop("prog_name")
+            except KeyError:
+                prog_name = self.get_default_prog_name(cli)
+
+            try:
+                return_value = cli.main(args=args or (), prog_name=prog_name, **extra)
+            except SystemExit as e:
+                exc_info = sys.exc_info()
+                e_code = t.cast(t.Optional[t.Union[int, t.Any]], e.code)
+
+                if e_code is None:
+                    e_code = 0
+
+                if e_code != 0:
+                    exception = e
+
+                if not isinstance(e_code, int):
+                    sys.stdout.write(str(e_code))
+                    sys.stdout.write("\n")
+                    e_code = 1
+
+                exit_code = e_code
+
+            except Exception as e:
+                if not catch_exceptions:
+                    raise
+                exception = e
+                exit_code = 1
+                exc_info = sys.exc_info()
+            finally:
+                sys.stdout.flush()
+                stdout = outstreams[0].getvalue()
+                if self.mix_stderr:
+                    stderr = None
+                else:
+                    stderr = outstreams[1].getvalue()  # type: ignore
+
+        return Result(
+            runner=self,
+            stdout_bytes=stdout,
+            stderr_bytes=stderr,
+            return_value=return_value,
+            exit_code=exit_code,
+            exception=exception,
+            exc_info=exc_info,  # type: ignore
+        )
+
+    @contextlib.contextmanager
+    def isolated_filesystem(
+        self, temp_dir: t.Optional[t.Union[str, "os.PathLike[str]"]] = None
+    ) -> t.Iterator[str]:
+        """A context manager that creates a temporary directory and
+        changes the current working directory to it. This isolates tests
+        that affect the contents of the CWD to prevent them from
+        interfering with each other.
+
+        :param temp_dir: Create the temporary directory under this
+            directory. If given, the created directory is not removed
+            when exiting.
+
+        .. versionchanged:: 8.0
+            Added the ``temp_dir`` parameter.
+        """
+        cwd = os.getcwd()
+        dt = tempfile.mkdtemp(dir=temp_dir)
+        os.chdir(dt)
+
+        try:
+            yield dt
+        finally:
+            os.chdir(cwd)
+
+            if temp_dir is None:
+                try:
+                    shutil.rmtree(dt)
+                except OSError:
+                    pass
diff --git a/backend/venv/lib/python3.12/site-packages/click/types.py b/backend/venv/lib/python3.12/site-packages/click/types.py
new file mode 100644
index 00000000..a70fd58c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/types.py
@@ -0,0 +1,1093 @@
+import os
+import stat
+import sys
+import typing as t
+from datetime import datetime
+from gettext import gettext as _
+from gettext import ngettext
+
+from ._compat import _get_argv_encoding
+from ._compat import open_stream
+from .exceptions import BadParameter
+from .utils import format_filename
+from .utils import LazyFile
+from .utils import safecall
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .core import Context
+    from .core import Parameter
+    from .shell_completion import CompletionItem
+
+
+class ParamType:
+    """Represents the type of a parameter. Validates and converts values
+    from the command line or Python into the correct type.
+
+    To implement a custom type, subclass and implement at least the
+    following:
+
+    -   The :attr:`name` class attribute must be set.
+    -   Calling an instance of the type with ``None`` must return
+        ``None``. This is already implemented by default.
+    -   :meth:`convert` must convert string values to the correct type.
+    -   :meth:`convert` must accept values that are already the correct
+        type.
+    -   It must be able to convert a value if the ``ctx`` and ``param``
+        arguments are ``None``. This can occur when converting prompt
+        input.
+    """
+
+    is_composite: t.ClassVar[bool] = False
+    arity: t.ClassVar[int] = 1
+
+    #: the descriptive name of this type
+    name: str
+
+    #: if a list of this type is expected and the value is pulled from a
+    #: string environment variable, this is what splits it up.  `None`
+    #: means any whitespace.  For all parameters the general rule is that
+    #: whitespace splits them up.  The exception are paths and files which
+    #: are split by ``os.path.pathsep`` by default (":" on Unix and ";" on
+    #: Windows).
+    envvar_list_splitter: t.ClassVar[t.Optional[str]] = None
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        """Gather information that could be useful for a tool generating
+        user-facing documentation.
+
+        Use :meth:`click.Context.to_info_dict` to traverse the entire
+        CLI structure.
+
+        .. versionadded:: 8.0
+        """
+        # The class name without the "ParamType" suffix.
+        param_type = type(self).__name__.partition("ParamType")[0]
+        param_type = param_type.partition("ParameterType")[0]
+
+        # Custom subclasses might not remember to set a name.
+        if hasattr(self, "name"):
+            name = self.name
+        else:
+            name = param_type
+
+        return {"param_type": param_type, "name": name}
+
+    def __call__(
+        self,
+        value: t.Any,
+        param: t.Optional["Parameter"] = None,
+        ctx: t.Optional["Context"] = None,
+    ) -> t.Any:
+        if value is not None:
+            return self.convert(value, param, ctx)
+
+    def get_metavar(self, param: "Parameter") -> t.Optional[str]:
+        """Returns the metavar default for this param if it provides one."""
+
+    def get_missing_message(self, param: "Parameter") -> t.Optional[str]:
+        """Optionally might return extra information about a missing
+        parameter.
+
+        .. versionadded:: 2.0
+        """
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        """Convert the value to the correct type. This is not called if
+        the value is ``None`` (the missing value).
+
+        This must accept string values from the command line, as well as
+        values that are already the correct type. It may also convert
+        other compatible types.
+
+        The ``param`` and ``ctx`` arguments may be ``None`` in certain
+        situations, such as when converting prompt input.
+
+        If the value cannot be converted, call :meth:`fail` with a
+        descriptive message.
+
+        :param value: The value to convert.
+        :param param: The parameter that is using this type to convert
+            its value. May be ``None``.
+        :param ctx: The current context that arrived at this value. May
+            be ``None``.
+        """
+        return value
+
+    def split_envvar_value(self, rv: str) -> t.Sequence[str]:
+        """Given a value from an environment variable this splits it up
+        into small chunks depending on the defined envvar list splitter.
+
+        If the splitter is set to `None`, which means that whitespace splits,
+        then leading and trailing whitespace is ignored.  Otherwise, leading
+        and trailing splitters usually lead to empty items being included.
+        """
+        return (rv or "").split(self.envvar_list_splitter)
+
+    def fail(
+        self,
+        message: str,
+        param: t.Optional["Parameter"] = None,
+        ctx: t.Optional["Context"] = None,
+    ) -> "t.NoReturn":
+        """Helper method to fail with an invalid value message."""
+        raise BadParameter(message, ctx=ctx, param=param)
+
+    def shell_complete(
+        self, ctx: "Context", param: "Parameter", incomplete: str
+    ) -> t.List["CompletionItem"]:
+        """Return a list of
+        :class:`~click.shell_completion.CompletionItem` objects for the
+        incomplete value. Most types do not provide completions, but
+        some do, and this allows custom types to provide custom
+        completions as well.
+
+        :param ctx: Invocation context for this command.
+        :param param: The parameter that is requesting completion.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        return []
+
+
+class CompositeParamType(ParamType):
+    is_composite = True
+
+    @property
+    def arity(self) -> int:  # type: ignore
+        raise NotImplementedError()
+
+
+class FuncParamType(ParamType):
+    def __init__(self, func: t.Callable[[t.Any], t.Any]) -> None:
+        self.name: str = func.__name__
+        self.func = func
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict["func"] = self.func
+        return info_dict
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        try:
+            return self.func(value)
+        except ValueError:
+            try:
+                value = str(value)
+            except UnicodeError:
+                value = value.decode("utf-8", "replace")
+
+            self.fail(value, param, ctx)
+
+
+class UnprocessedParamType(ParamType):
+    name = "text"
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        return value
+
+    def __repr__(self) -> str:
+        return "UNPROCESSED"
+
+
+class StringParamType(ParamType):
+    name = "text"
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        if isinstance(value, bytes):
+            enc = _get_argv_encoding()
+            try:
+                value = value.decode(enc)
+            except UnicodeError:
+                fs_enc = sys.getfilesystemencoding()
+                if fs_enc != enc:
+                    try:
+                        value = value.decode(fs_enc)
+                    except UnicodeError:
+                        value = value.decode("utf-8", "replace")
+                else:
+                    value = value.decode("utf-8", "replace")
+            return value
+        return str(value)
+
+    def __repr__(self) -> str:
+        return "STRING"
+
+
+class Choice(ParamType):
+    """The choice type allows a value to be checked against a fixed set
+    of supported values. All of these values have to be strings.
+
+    You should only pass a list or tuple of choices. Other iterables
+    (like generators) may lead to surprising results.
+
+    The resulting value will always be one of the originally passed choices
+    regardless of ``case_sensitive`` or any ``ctx.token_normalize_func``
+    being specified.
+
+    See :ref:`choice-opts` for an example.
+
+    :param case_sensitive: Set to false to make choices case
+        insensitive. Defaults to true.
+    """
+
+    name = "choice"
+
+    def __init__(self, choices: t.Sequence[str], case_sensitive: bool = True) -> None:
+        self.choices = choices
+        self.case_sensitive = case_sensitive
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict["choices"] = self.choices
+        info_dict["case_sensitive"] = self.case_sensitive
+        return info_dict
+
+    def get_metavar(self, param: "Parameter") -> str:
+        choices_str = "|".join(self.choices)
+
+        # Use curly braces to indicate a required argument.
+        if param.required and param.param_type_name == "argument":
+            return f"{{{choices_str}}}"
+
+        # Use square braces to indicate an option or optional argument.
+        return f"[{choices_str}]"
+
+    def get_missing_message(self, param: "Parameter") -> str:
+        return _("Choose from:\n\t{choices}").format(choices=",\n\t".join(self.choices))
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        # Match through normalization and case sensitivity
+        # first do token_normalize_func, then lowercase
+        # preserve original `value` to produce an accurate message in
+        # `self.fail`
+        normed_value = value
+        normed_choices = {choice: choice for choice in self.choices}
+
+        if ctx is not None and ctx.token_normalize_func is not None:
+            normed_value = ctx.token_normalize_func(value)
+            normed_choices = {
+                ctx.token_normalize_func(normed_choice): original
+                for normed_choice, original in normed_choices.items()
+            }
+
+        if not self.case_sensitive:
+            normed_value = normed_value.casefold()
+            normed_choices = {
+                normed_choice.casefold(): original
+                for normed_choice, original in normed_choices.items()
+            }
+
+        if normed_value in normed_choices:
+            return normed_choices[normed_value]
+
+        choices_str = ", ".join(map(repr, self.choices))
+        self.fail(
+            ngettext(
+                "{value!r} is not {choice}.",
+                "{value!r} is not one of {choices}.",
+                len(self.choices),
+            ).format(value=value, choice=choices_str, choices=choices_str),
+            param,
+            ctx,
+        )
+
+    def __repr__(self) -> str:
+        return f"Choice({list(self.choices)})"
+
+    def shell_complete(
+        self, ctx: "Context", param: "Parameter", incomplete: str
+    ) -> t.List["CompletionItem"]:
+        """Complete choices that start with the incomplete value.
+
+        :param ctx: Invocation context for this command.
+        :param param: The parameter that is requesting completion.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        str_choices = map(str, self.choices)
+
+        if self.case_sensitive:
+            matched = (c for c in str_choices if c.startswith(incomplete))
+        else:
+            incomplete = incomplete.lower()
+            matched = (c for c in str_choices if c.lower().startswith(incomplete))
+
+        return [CompletionItem(c) for c in matched]
+
+
+class DateTime(ParamType):
+    """The DateTime type converts date strings into `datetime` objects.
+
+    The format strings which are checked are configurable, but default to some
+    common (non-timezone aware) ISO 8601 formats.
+
+    When specifying *DateTime* formats, you should only pass a list or a tuple.
+    Other iterables, like generators, may lead to surprising results.
+
+    The format strings are processed using ``datetime.strptime``, and this
+    consequently defines the format strings which are allowed.
+
+    Parsing is tried using each format, in order, and the first format which
+    parses successfully is used.
+
+    :param formats: A list or tuple of date format strings, in the order in
+                    which they should be tried. Defaults to
+                    ``'%Y-%m-%d'``, ``'%Y-%m-%dT%H:%M:%S'``,
+                    ``'%Y-%m-%d %H:%M:%S'``.
+    """
+
+    name = "datetime"
+
+    def __init__(self, formats: t.Optional[t.Sequence[str]] = None):
+        self.formats: t.Sequence[str] = formats or [
+            "%Y-%m-%d",
+            "%Y-%m-%dT%H:%M:%S",
+            "%Y-%m-%d %H:%M:%S",
+        ]
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict["formats"] = self.formats
+        return info_dict
+
+    def get_metavar(self, param: "Parameter") -> str:
+        return f"[{'|'.join(self.formats)}]"
+
+    def _try_to_convert_date(self, value: t.Any, format: str) -> t.Optional[datetime]:
+        try:
+            return datetime.strptime(value, format)
+        except ValueError:
+            return None
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        if isinstance(value, datetime):
+            return value
+
+        for format in self.formats:
+            converted = self._try_to_convert_date(value, format)
+
+            if converted is not None:
+                return converted
+
+        formats_str = ", ".join(map(repr, self.formats))
+        self.fail(
+            ngettext(
+                "{value!r} does not match the format {format}.",
+                "{value!r} does not match the formats {formats}.",
+                len(self.formats),
+            ).format(value=value, format=formats_str, formats=formats_str),
+            param,
+            ctx,
+        )
+
+    def __repr__(self) -> str:
+        return "DateTime"
+
+
+class _NumberParamTypeBase(ParamType):
+    _number_class: t.ClassVar[t.Type[t.Any]]
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        try:
+            return self._number_class(value)
+        except ValueError:
+            self.fail(
+                _("{value!r} is not a valid {number_type}.").format(
+                    value=value, number_type=self.name
+                ),
+                param,
+                ctx,
+            )
+
+
+class _NumberRangeBase(_NumberParamTypeBase):
+    def __init__(
+        self,
+        min: t.Optional[float] = None,
+        max: t.Optional[float] = None,
+        min_open: bool = False,
+        max_open: bool = False,
+        clamp: bool = False,
+    ) -> None:
+        self.min = min
+        self.max = max
+        self.min_open = min_open
+        self.max_open = max_open
+        self.clamp = clamp
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict.update(
+            min=self.min,
+            max=self.max,
+            min_open=self.min_open,
+            max_open=self.max_open,
+            clamp=self.clamp,
+        )
+        return info_dict
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        import operator
+
+        rv = super().convert(value, param, ctx)
+        lt_min: bool = self.min is not None and (
+            operator.le if self.min_open else operator.lt
+        )(rv, self.min)
+        gt_max: bool = self.max is not None and (
+            operator.ge if self.max_open else operator.gt
+        )(rv, self.max)
+
+        if self.clamp:
+            if lt_min:
+                return self._clamp(self.min, 1, self.min_open)  # type: ignore
+
+            if gt_max:
+                return self._clamp(self.max, -1, self.max_open)  # type: ignore
+
+        if lt_min or gt_max:
+            self.fail(
+                _("{value} is not in the range {range}.").format(
+                    value=rv, range=self._describe_range()
+                ),
+                param,
+                ctx,
+            )
+
+        return rv
+
+    def _clamp(self, bound: float, dir: "te.Literal[1, -1]", open: bool) -> float:
+        """Find the valid value to clamp to bound in the given
+        direction.
+
+        :param bound: The boundary value.
+        :param dir: 1 or -1 indicating the direction to move.
+        :param open: If true, the range does not include the bound.
+        """
+        raise NotImplementedError
+
+    def _describe_range(self) -> str:
+        """Describe the range for use in help text."""
+        if self.min is None:
+            op = "<" if self.max_open else "<="
+            return f"x{op}{self.max}"
+
+        if self.max is None:
+            op = ">" if self.min_open else ">="
+            return f"x{op}{self.min}"
+
+        lop = "<" if self.min_open else "<="
+        rop = "<" if self.max_open else "<="
+        return f"{self.min}{lop}x{rop}{self.max}"
+
+    def __repr__(self) -> str:
+        clamp = " clamped" if self.clamp else ""
+        return f"<{type(self).__name__} {self._describe_range()}{clamp}>"
+
+
+class IntParamType(_NumberParamTypeBase):
+    name = "integer"
+    _number_class = int
+
+    def __repr__(self) -> str:
+        return "INT"
+
+
+class IntRange(_NumberRangeBase, IntParamType):
+    """Restrict an :data:`click.INT` value to a range of accepted
+    values. See :ref:`ranges`.
+
+    If ``min`` or ``max`` are not passed, any value is accepted in that
+    direction. If ``min_open`` or ``max_open`` are enabled, the
+    corresponding boundary is not included in the range.
+
+    If ``clamp`` is enabled, a value outside the range is clamped to the
+    boundary instead of failing.
+
+    .. versionchanged:: 8.0
+        Added the ``min_open`` and ``max_open`` parameters.
+    """
+
+    name = "integer range"
+
+    def _clamp(  # type: ignore
+        self, bound: int, dir: "te.Literal[1, -1]", open: bool
+    ) -> int:
+        if not open:
+            return bound
+
+        return bound + dir
+
+
+class FloatParamType(_NumberParamTypeBase):
+    name = "float"
+    _number_class = float
+
+    def __repr__(self) -> str:
+        return "FLOAT"
+
+
+class FloatRange(_NumberRangeBase, FloatParamType):
+    """Restrict a :data:`click.FLOAT` value to a range of accepted
+    values. See :ref:`ranges`.
+
+    If ``min`` or ``max`` are not passed, any value is accepted in that
+    direction. If ``min_open`` or ``max_open`` are enabled, the
+    corresponding boundary is not included in the range.
+
+    If ``clamp`` is enabled, a value outside the range is clamped to the
+    boundary instead of failing. This is not supported if either
+    boundary is marked ``open``.
+
+    .. versionchanged:: 8.0
+        Added the ``min_open`` and ``max_open`` parameters.
+    """
+
+    name = "float range"
+
+    def __init__(
+        self,
+        min: t.Optional[float] = None,
+        max: t.Optional[float] = None,
+        min_open: bool = False,
+        max_open: bool = False,
+        clamp: bool = False,
+    ) -> None:
+        super().__init__(
+            min=min, max=max, min_open=min_open, max_open=max_open, clamp=clamp
+        )
+
+        if (min_open or max_open) and clamp:
+            raise TypeError("Clamping is not supported for open bounds.")
+
+    def _clamp(self, bound: float, dir: "te.Literal[1, -1]", open: bool) -> float:
+        if not open:
+            return bound
+
+        # Could use Python 3.9's math.nextafter here, but clamping an
+        # open float range doesn't seem to be particularly useful. It's
+        # left up to the user to write a callback to do it if needed.
+        raise RuntimeError("Clamping is not supported for open bounds.")
+
+
+class BoolParamType(ParamType):
+    name = "boolean"
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        if value in {False, True}:
+            return bool(value)
+
+        norm = value.strip().lower()
+
+        if norm in {"1", "true", "t", "yes", "y", "on"}:
+            return True
+
+        if norm in {"0", "false", "f", "no", "n", "off"}:
+            return False
+
+        self.fail(
+            _("{value!r} is not a valid boolean.").format(value=value), param, ctx
+        )
+
+    def __repr__(self) -> str:
+        return "BOOL"
+
+
+class UUIDParameterType(ParamType):
+    name = "uuid"
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        import uuid
+
+        if isinstance(value, uuid.UUID):
+            return value
+
+        value = value.strip()
+
+        try:
+            return uuid.UUID(value)
+        except ValueError:
+            self.fail(
+                _("{value!r} is not a valid UUID.").format(value=value), param, ctx
+            )
+
+    def __repr__(self) -> str:
+        return "UUID"
+
+
+class File(ParamType):
+    """Declares a parameter to be a file for reading or writing.  The file
+    is automatically closed once the context tears down (after the command
+    finished working).
+
+    Files can be opened for reading or writing.  The special value ``-``
+    indicates stdin or stdout depending on the mode.
+
+    By default, the file is opened for reading text data, but it can also be
+    opened in binary mode or for writing.  The encoding parameter can be used
+    to force a specific encoding.
+
+    The `lazy` flag controls if the file should be opened immediately or upon
+    first IO. The default is to be non-lazy for standard input and output
+    streams as well as files opened for reading, `lazy` otherwise. When opening a
+    file lazily for reading, it is still opened temporarily for validation, but
+    will not be held open until first IO. lazy is mainly useful when opening
+    for writing to avoid creating the file until it is needed.
+
+    Files can also be opened atomically in which case all writes go into a
+    separate file in the same folder and upon completion the file will
+    be moved over to the original location.  This is useful if a file
+    regularly read by other users is modified.
+
+    See :ref:`file-args` for more information.
+
+    .. versionchanged:: 2.0
+        Added the ``atomic`` parameter.
+    """
+
+    name = "filename"
+    envvar_list_splitter: t.ClassVar[str] = os.path.pathsep
+
+    def __init__(
+        self,
+        mode: str = "r",
+        encoding: t.Optional[str] = None,
+        errors: t.Optional[str] = "strict",
+        lazy: t.Optional[bool] = None,
+        atomic: bool = False,
+    ) -> None:
+        self.mode = mode
+        self.encoding = encoding
+        self.errors = errors
+        self.lazy = lazy
+        self.atomic = atomic
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict.update(mode=self.mode, encoding=self.encoding)
+        return info_dict
+
+    def resolve_lazy_flag(self, value: "t.Union[str, os.PathLike[str]]") -> bool:
+        if self.lazy is not None:
+            return self.lazy
+        if os.fspath(value) == "-":
+            return False
+        elif "w" in self.mode:
+            return True
+        return False
+
+    def convert(
+        self,
+        value: t.Union[str, "os.PathLike[str]", t.IO[t.Any]],
+        param: t.Optional["Parameter"],
+        ctx: t.Optional["Context"],
+    ) -> t.IO[t.Any]:
+        if _is_file_like(value):
+            return value
+
+        value = t.cast("t.Union[str, os.PathLike[str]]", value)
+
+        try:
+            lazy = self.resolve_lazy_flag(value)
+
+            if lazy:
+                lf = LazyFile(
+                    value, self.mode, self.encoding, self.errors, atomic=self.atomic
+                )
+
+                if ctx is not None:
+                    ctx.call_on_close(lf.close_intelligently)
+
+                return t.cast(t.IO[t.Any], lf)
+
+            f, should_close = open_stream(
+                value, self.mode, self.encoding, self.errors, atomic=self.atomic
+            )
+
+            # If a context is provided, we automatically close the file
+            # at the end of the context execution (or flush out).  If a
+            # context does not exist, it's the caller's responsibility to
+            # properly close the file.  This for instance happens when the
+            # type is used with prompts.
+            if ctx is not None:
+                if should_close:
+                    ctx.call_on_close(safecall(f.close))
+                else:
+                    ctx.call_on_close(safecall(f.flush))
+
+            return f
+        except OSError as e:
+            self.fail(f"'{format_filename(value)}': {e.strerror}", param, ctx)
+
+    def shell_complete(
+        self, ctx: "Context", param: "Parameter", incomplete: str
+    ) -> t.List["CompletionItem"]:
+        """Return a special completion marker that tells the completion
+        system to use the shell to provide file path completions.
+
+        :param ctx: Invocation context for this command.
+        :param param: The parameter that is requesting completion.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        return [CompletionItem(incomplete, type="file")]
+
+
+def _is_file_like(value: t.Any) -> "te.TypeGuard[t.IO[t.Any]]":
+    return hasattr(value, "read") or hasattr(value, "write")
+
+
+class Path(ParamType):
+    """The ``Path`` type is similar to the :class:`File` type, but
+    returns the filename instead of an open file. Various checks can be
+    enabled to validate the type of file and permissions.
+
+    :param exists: The file or directory needs to exist for the value to
+        be valid. If this is not set to ``True``, and the file does not
+        exist, then all further checks are silently skipped.
+    :param file_okay: Allow a file as a value.
+    :param dir_okay: Allow a directory as a value.
+    :param readable: if true, a readable check is performed.
+    :param writable: if true, a writable check is performed.
+    :param executable: if true, an executable check is performed.
+    :param resolve_path: Make the value absolute and resolve any
+        symlinks. A ``~`` is not expanded, as this is supposed to be
+        done by the shell only.
+    :param allow_dash: Allow a single dash as a value, which indicates
+        a standard stream (but does not open it). Use
+        :func:`~click.open_file` to handle opening this value.
+    :param path_type: Convert the incoming path value to this type. If
+        ``None``, keep Python's default, which is ``str``. Useful to
+        convert to :class:`pathlib.Path`.
+
+    .. versionchanged:: 8.1
+        Added the ``executable`` parameter.
+
+    .. versionchanged:: 8.0
+        Allow passing ``path_type=pathlib.Path``.
+
+    .. versionchanged:: 6.0
+        Added the ``allow_dash`` parameter.
+    """
+
+    envvar_list_splitter: t.ClassVar[str] = os.path.pathsep
+
+    def __init__(
+        self,
+        exists: bool = False,
+        file_okay: bool = True,
+        dir_okay: bool = True,
+        writable: bool = False,
+        readable: bool = True,
+        resolve_path: bool = False,
+        allow_dash: bool = False,
+        path_type: t.Optional[t.Type[t.Any]] = None,
+        executable: bool = False,
+    ):
+        self.exists = exists
+        self.file_okay = file_okay
+        self.dir_okay = dir_okay
+        self.readable = readable
+        self.writable = writable
+        self.executable = executable
+        self.resolve_path = resolve_path
+        self.allow_dash = allow_dash
+        self.type = path_type
+
+        if self.file_okay and not self.dir_okay:
+            self.name: str = _("file")
+        elif self.dir_okay and not self.file_okay:
+            self.name = _("directory")
+        else:
+            self.name = _("path")
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict.update(
+            exists=self.exists,
+            file_okay=self.file_okay,
+            dir_okay=self.dir_okay,
+            writable=self.writable,
+            readable=self.readable,
+            allow_dash=self.allow_dash,
+        )
+        return info_dict
+
+    def coerce_path_result(
+        self, value: "t.Union[str, os.PathLike[str]]"
+    ) -> "t.Union[str, bytes, os.PathLike[str]]":
+        if self.type is not None and not isinstance(value, self.type):
+            if self.type is str:
+                return os.fsdecode(value)
+            elif self.type is bytes:
+                return os.fsencode(value)
+            else:
+                return t.cast("os.PathLike[str]", self.type(value))
+
+        return value
+
+    def convert(
+        self,
+        value: "t.Union[str, os.PathLike[str]]",
+        param: t.Optional["Parameter"],
+        ctx: t.Optional["Context"],
+    ) -> "t.Union[str, bytes, os.PathLike[str]]":
+        rv = value
+
+        is_dash = self.file_okay and self.allow_dash and rv in (b"-", "-")
+
+        if not is_dash:
+            if self.resolve_path:
+                # os.path.realpath doesn't resolve symlinks on Windows
+                # until Python 3.8. Use pathlib for now.
+                import pathlib
+
+                rv = os.fsdecode(pathlib.Path(rv).resolve())
+
+            try:
+                st = os.stat(rv)
+            except OSError:
+                if not self.exists:
+                    return self.coerce_path_result(rv)
+                self.fail(
+                    _("{name} {filename!r} does not exist.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+
+            if not self.file_okay and stat.S_ISREG(st.st_mode):
+                self.fail(
+                    _("{name} {filename!r} is a file.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+            if not self.dir_okay and stat.S_ISDIR(st.st_mode):
+                self.fail(
+                    _("{name} {filename!r} is a directory.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+
+            if self.readable and not os.access(rv, os.R_OK):
+                self.fail(
+                    _("{name} {filename!r} is not readable.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+
+            if self.writable and not os.access(rv, os.W_OK):
+                self.fail(
+                    _("{name} {filename!r} is not writable.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+
+            if self.executable and not os.access(value, os.X_OK):
+                self.fail(
+                    _("{name} {filename!r} is not executable.").format(
+                        name=self.name.title(), filename=format_filename(value)
+                    ),
+                    param,
+                    ctx,
+                )
+
+        return self.coerce_path_result(rv)
+
+    def shell_complete(
+        self, ctx: "Context", param: "Parameter", incomplete: str
+    ) -> t.List["CompletionItem"]:
+        """Return a special completion marker that tells the completion
+        system to use the shell to provide path completions for only
+        directories or any paths.
+
+        :param ctx: Invocation context for this command.
+        :param param: The parameter that is requesting completion.
+        :param incomplete: Value being completed. May be empty.
+
+        .. versionadded:: 8.0
+        """
+        from click.shell_completion import CompletionItem
+
+        type = "dir" if self.dir_okay and not self.file_okay else "file"
+        return [CompletionItem(incomplete, type=type)]
+
+
+class Tuple(CompositeParamType):
+    """The default behavior of Click is to apply a type on a value directly.
+    This works well in most cases, except for when `nargs` is set to a fixed
+    count and different types should be used for different items.  In this
+    case the :class:`Tuple` type can be used.  This type can only be used
+    if `nargs` is set to a fixed number.
+
+    For more information see :ref:`tuple-type`.
+
+    This can be selected by using a Python tuple literal as a type.
+
+    :param types: a list of types that should be used for the tuple items.
+    """
+
+    def __init__(self, types: t.Sequence[t.Union[t.Type[t.Any], ParamType]]) -> None:
+        self.types: t.Sequence[ParamType] = [convert_type(ty) for ty in types]
+
+    def to_info_dict(self) -> t.Dict[str, t.Any]:
+        info_dict = super().to_info_dict()
+        info_dict["types"] = [t.to_info_dict() for t in self.types]
+        return info_dict
+
+    @property
+    def name(self) -> str:  # type: ignore
+        return f"<{' '.join(ty.name for ty in self.types)}>"
+
+    @property
+    def arity(self) -> int:  # type: ignore
+        return len(self.types)
+
+    def convert(
+        self, value: t.Any, param: t.Optional["Parameter"], ctx: t.Optional["Context"]
+    ) -> t.Any:
+        len_type = len(self.types)
+        len_value = len(value)
+
+        if len_value != len_type:
+            self.fail(
+                ngettext(
+                    "{len_type} values are required, but {len_value} was given.",
+                    "{len_type} values are required, but {len_value} were given.",
+                    len_value,
+                ).format(len_type=len_type, len_value=len_value),
+                param=param,
+                ctx=ctx,
+            )
+
+        return tuple(ty(x, param, ctx) for ty, x in zip(self.types, value))
+
+
+def convert_type(ty: t.Optional[t.Any], default: t.Optional[t.Any] = None) -> ParamType:
+    """Find the most appropriate :class:`ParamType` for the given Python
+    type. If the type isn't provided, it can be inferred from a default
+    value.
+    """
+    guessed_type = False
+
+    if ty is None and default is not None:
+        if isinstance(default, (tuple, list)):
+            # If the default is empty, ty will remain None and will
+            # return STRING.
+            if default:
+                item = default[0]
+
+                # A tuple of tuples needs to detect the inner types.
+                # Can't call convert recursively because that would
+                # incorrectly unwind the tuple to a single type.
+                if isinstance(item, (tuple, list)):
+                    ty = tuple(map(type, item))
+                else:
+                    ty = type(item)
+        else:
+            ty = type(default)
+
+        guessed_type = True
+
+    if isinstance(ty, tuple):
+        return Tuple(ty)
+
+    if isinstance(ty, ParamType):
+        return ty
+
+    if ty is str or ty is None:
+        return STRING
+
+    if ty is int:
+        return INT
+
+    if ty is float:
+        return FLOAT
+
+    if ty is bool:
+        return BOOL
+
+    if guessed_type:
+        return STRING
+
+    if __debug__:
+        try:
+            if issubclass(ty, ParamType):
+                raise AssertionError(
+                    f"Attempted to use an uninstantiated parameter type ({ty})."
+                )
+        except TypeError:
+            # ty is an instance (correct), so issubclass fails.
+            pass
+
+    return FuncParamType(ty)
+
+
+#: A dummy parameter type that just does nothing.  From a user's
+#: perspective this appears to just be the same as `STRING` but
+#: internally no string conversion takes place if the input was bytes.
+#: This is usually useful when working with file paths as they can
+#: appear in bytes and unicode.
+#:
+#: For path related uses the :class:`Path` type is a better choice but
+#: there are situations where an unprocessed type is useful which is why
+#: it is is provided.
+#:
+#: .. versionadded:: 4.0
+UNPROCESSED = UnprocessedParamType()
+
+#: A unicode string parameter type which is the implicit default.  This
+#: can also be selected by using ``str`` as type.
+STRING = StringParamType()
+
+#: An integer parameter.  This can also be selected by using ``int`` as
+#: type.
+INT = IntParamType()
+
+#: A floating point value parameter.  This can also be selected by using
+#: ``float`` as type.
+FLOAT = FloatParamType()
+
+#: A boolean parameter.  This is the default for boolean flags.  This can
+#: also be selected by using ``bool`` as a type.
+BOOL = BoolParamType()
+
+#: A UUID parameter.
+UUID = UUIDParameterType()
diff --git a/backend/venv/lib/python3.12/site-packages/click/utils.py b/backend/venv/lib/python3.12/site-packages/click/utils.py
new file mode 100644
index 00000000..836c6f21
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/click/utils.py
@@ -0,0 +1,624 @@
+import os
+import re
+import sys
+import typing as t
+from functools import update_wrapper
+from types import ModuleType
+from types import TracebackType
+
+from ._compat import _default_text_stderr
+from ._compat import _default_text_stdout
+from ._compat import _find_binary_writer
+from ._compat import auto_wrap_for_ansi
+from ._compat import binary_streams
+from ._compat import open_stream
+from ._compat import should_strip_ansi
+from ._compat import strip_ansi
+from ._compat import text_streams
+from ._compat import WIN
+from .globals import resolve_color_default
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    P = te.ParamSpec("P")
+
+R = t.TypeVar("R")
+
+
+def _posixify(name: str) -> str:
+    return "-".join(name.split()).lower()
+
+
+def safecall(func: "t.Callable[P, R]") -> "t.Callable[P, t.Optional[R]]":
+    """Wraps a function so that it swallows exceptions."""
+
+    def wrapper(*args: "P.args", **kwargs: "P.kwargs") -> t.Optional[R]:
+        try:
+            return func(*args, **kwargs)
+        except Exception:
+            pass
+        return None
+
+    return update_wrapper(wrapper, func)
+
+
+def make_str(value: t.Any) -> str:
+    """Converts a value into a valid string."""
+    if isinstance(value, bytes):
+        try:
+            return value.decode(sys.getfilesystemencoding())
+        except UnicodeError:
+            return value.decode("utf-8", "replace")
+    return str(value)
+
+
+def make_default_short_help(help: str, max_length: int = 45) -> str:
+    """Returns a condensed version of help string."""
+    # Consider only the first paragraph.
+    paragraph_end = help.find("\n\n")
+
+    if paragraph_end != -1:
+        help = help[:paragraph_end]
+
+    # Collapse newlines, tabs, and spaces.
+    words = help.split()
+
+    if not words:
+        return ""
+
+    # The first paragraph started with a "no rewrap" marker, ignore it.
+    if words[0] == "\b":
+        words = words[1:]
+
+    total_length = 0
+    last_index = len(words) - 1
+
+    for i, word in enumerate(words):
+        total_length += len(word) + (i > 0)
+
+        if total_length > max_length:  # too long, truncate
+            break
+
+        if word[-1] == ".":  # sentence end, truncate without "..."
+            return " ".join(words[: i + 1])
+
+        if total_length == max_length and i != last_index:
+            break  # not at sentence end, truncate with "..."
+    else:
+        return " ".join(words)  # no truncation needed
+
+    # Account for the length of the suffix.
+    total_length += len("...")
+
+    # remove words until the length is short enough
+    while i > 0:
+        total_length -= len(words[i]) + (i > 0)
+
+        if total_length <= max_length:
+            break
+
+        i -= 1
+
+    return " ".join(words[:i]) + "..."
+
+
+class LazyFile:
+    """A lazy file works like a regular file but it does not fully open
+    the file but it does perform some basic checks early to see if the
+    filename parameter does make sense.  This is useful for safely opening
+    files for writing.
+    """
+
+    def __init__(
+        self,
+        filename: t.Union[str, "os.PathLike[str]"],
+        mode: str = "r",
+        encoding: t.Optional[str] = None,
+        errors: t.Optional[str] = "strict",
+        atomic: bool = False,
+    ):
+        self.name: str = os.fspath(filename)
+        self.mode = mode
+        self.encoding = encoding
+        self.errors = errors
+        self.atomic = atomic
+        self._f: t.Optional[t.IO[t.Any]]
+        self.should_close: bool
+
+        if self.name == "-":
+            self._f, self.should_close = open_stream(filename, mode, encoding, errors)
+        else:
+            if "r" in mode:
+                # Open and close the file in case we're opening it for
+                # reading so that we can catch at least some errors in
+                # some cases early.
+                open(filename, mode).close()
+            self._f = None
+            self.should_close = True
+
+    def __getattr__(self, name: str) -> t.Any:
+        return getattr(self.open(), name)
+
+    def __repr__(self) -> str:
+        if self._f is not None:
+            return repr(self._f)
+        return f"<unopened file '{format_filename(self.name)}' {self.mode}>"
+
+    def open(self) -> t.IO[t.Any]:
+        """Opens the file if it's not yet open.  This call might fail with
+        a :exc:`FileError`.  Not handling this error will produce an error
+        that Click shows.
+        """
+        if self._f is not None:
+            return self._f
+        try:
+            rv, self.should_close = open_stream(
+                self.name, self.mode, self.encoding, self.errors, atomic=self.atomic
+            )
+        except OSError as e:
+            from .exceptions import FileError
+
+            raise FileError(self.name, hint=e.strerror) from e
+        self._f = rv
+        return rv
+
+    def close(self) -> None:
+        """Closes the underlying file, no matter what."""
+        if self._f is not None:
+            self._f.close()
+
+    def close_intelligently(self) -> None:
+        """This function only closes the file if it was opened by the lazy
+        file wrapper.  For instance this will never close stdin.
+        """
+        if self.should_close:
+            self.close()
+
+    def __enter__(self) -> "LazyFile":
+        return self
+
+    def __exit__(
+        self,
+        exc_type: t.Optional[t.Type[BaseException]],
+        exc_value: t.Optional[BaseException],
+        tb: t.Optional[TracebackType],
+    ) -> None:
+        self.close_intelligently()
+
+    def __iter__(self) -> t.Iterator[t.AnyStr]:
+        self.open()
+        return iter(self._f)  # type: ignore
+
+
+class KeepOpenFile:
+    def __init__(self, file: t.IO[t.Any]) -> None:
+        self._file: t.IO[t.Any] = file
+
+    def __getattr__(self, name: str) -> t.Any:
+        return getattr(self._file, name)
+
+    def __enter__(self) -> "KeepOpenFile":
+        return self
+
+    def __exit__(
+        self,
+        exc_type: t.Optional[t.Type[BaseException]],
+        exc_value: t.Optional[BaseException],
+        tb: t.Optional[TracebackType],
+    ) -> None:
+        pass
+
+    def __repr__(self) -> str:
+        return repr(self._file)
+
+    def __iter__(self) -> t.Iterator[t.AnyStr]:
+        return iter(self._file)
+
+
+def echo(
+    message: t.Optional[t.Any] = None,
+    file: t.Optional[t.IO[t.Any]] = None,
+    nl: bool = True,
+    err: bool = False,
+    color: t.Optional[bool] = None,
+) -> None:
+    """Print a message and newline to stdout or a file. This should be
+    used instead of :func:`print` because it provides better support
+    for different data, files, and environments.
+
+    Compared to :func:`print`, this does the following:
+
+    -   Ensures that the output encoding is not misconfigured on Linux.
+    -   Supports Unicode in the Windows console.
+    -   Supports writing to binary outputs, and supports writing bytes
+        to text outputs.
+    -   Supports colors and styles on Windows.
+    -   Removes ANSI color and style codes if the output does not look
+        like an interactive terminal.
+    -   Always flushes the output.
+
+    :param message: The string or bytes to output. Other objects are
+        converted to strings.
+    :param file: The file to write to. Defaults to ``stdout``.
+    :param err: Write to ``stderr`` instead of ``stdout``.
+    :param nl: Print a newline after the message. Enabled by default.
+    :param color: Force showing or hiding colors and other styles. By
+        default Click will remove color if the output does not look like
+        an interactive terminal.
+
+    .. versionchanged:: 6.0
+        Support Unicode output on the Windows console. Click does not
+        modify ``sys.stdout``, so ``sys.stdout.write()`` and ``print()``
+        will still not support Unicode.
+
+    .. versionchanged:: 4.0
+        Added the ``color`` parameter.
+
+    .. versionadded:: 3.0
+        Added the ``err`` parameter.
+
+    .. versionchanged:: 2.0
+        Support colors on Windows if colorama is installed.
+    """
+    if file is None:
+        if err:
+            file = _default_text_stderr()
+        else:
+            file = _default_text_stdout()
+
+        # There are no standard streams attached to write to. For example,
+        # pythonw on Windows.
+        if file is None:
+            return
+
+    # Convert non bytes/text into the native string type.
+    if message is not None and not isinstance(message, (str, bytes, bytearray)):
+        out: t.Optional[t.Union[str, bytes]] = str(message)
+    else:
+        out = message
+
+    if nl:
+        out = out or ""
+        if isinstance(out, str):
+            out += "\n"
+        else:
+            out += b"\n"
+
+    if not out:
+        file.flush()
+        return
+
+    # If there is a message and the value looks like bytes, we manually
+    # need to find the binary stream and write the message in there.
+    # This is done separately so that most stream types will work as you
+    # would expect. Eg: you can write to StringIO for other cases.
+    if isinstance(out, (bytes, bytearray)):
+        binary_file = _find_binary_writer(file)
+
+        if binary_file is not None:
+            file.flush()
+            binary_file.write(out)
+            binary_file.flush()
+            return
+
+    # ANSI style code support. For no message or bytes, nothing happens.
+    # When outputting to a file instead of a terminal, strip codes.
+    else:
+        color = resolve_color_default(color)
+
+        if should_strip_ansi(file, color):
+            out = strip_ansi(out)
+        elif WIN:
+            if auto_wrap_for_ansi is not None:
+                file = auto_wrap_for_ansi(file, color)  # type: ignore
+            elif not color:
+                out = strip_ansi(out)
+
+    file.write(out)  # type: ignore
+    file.flush()
+
+
+def get_binary_stream(name: "te.Literal['stdin', 'stdout', 'stderr']") -> t.BinaryIO:
+    """Returns a system stream for byte processing.
+
+    :param name: the name of the stream to open.  Valid names are ``'stdin'``,
+                 ``'stdout'`` and ``'stderr'``
+    """
+    opener = binary_streams.get(name)
+    if opener is None:
+        raise TypeError(f"Unknown standard stream '{name}'")
+    return opener()
+
+
+def get_text_stream(
+    name: "te.Literal['stdin', 'stdout', 'stderr']",
+    encoding: t.Optional[str] = None,
+    errors: t.Optional[str] = "strict",
+) -> t.TextIO:
+    """Returns a system stream for text processing.  This usually returns
+    a wrapped stream around a binary stream returned from
+    :func:`get_binary_stream` but it also can take shortcuts for already
+    correctly configured streams.
+
+    :param name: the name of the stream to open.  Valid names are ``'stdin'``,
+                 ``'stdout'`` and ``'stderr'``
+    :param encoding: overrides the detected default encoding.
+    :param errors: overrides the default error mode.
+    """
+    opener = text_streams.get(name)
+    if opener is None:
+        raise TypeError(f"Unknown standard stream '{name}'")
+    return opener(encoding, errors)
+
+
+def open_file(
+    filename: t.Union[str, "os.PathLike[str]"],
+    mode: str = "r",
+    encoding: t.Optional[str] = None,
+    errors: t.Optional[str] = "strict",
+    lazy: bool = False,
+    atomic: bool = False,
+) -> t.IO[t.Any]:
+    """Open a file, with extra behavior to handle ``'-'`` to indicate
+    a standard stream, lazy open on write, and atomic write. Similar to
+    the behavior of the :class:`~click.File` param type.
+
+    If ``'-'`` is given to open ``stdout`` or ``stdin``, the stream is
+    wrapped so that using it in a context manager will not close it.
+    This makes it possible to use the function without accidentally
+    closing a standard stream:
+
+    .. code-block:: python
+
+        with open_file(filename) as f:
+            ...
+
+    :param filename: The name or Path of the file to open, or ``'-'`` for
+        ``stdin``/``stdout``.
+    :param mode: The mode in which to open the file.
+    :param encoding: The encoding to decode or encode a file opened in
+        text mode.
+    :param errors: The error handling mode.
+    :param lazy: Wait to open the file until it is accessed. For read
+        mode, the file is temporarily opened to raise access errors
+        early, then closed until it is read again.
+    :param atomic: Write to a temporary file and replace the given file
+        on close.
+
+    .. versionadded:: 3.0
+    """
+    if lazy:
+        return t.cast(
+            t.IO[t.Any], LazyFile(filename, mode, encoding, errors, atomic=atomic)
+        )
+
+    f, should_close = open_stream(filename, mode, encoding, errors, atomic=atomic)
+
+    if not should_close:
+        f = t.cast(t.IO[t.Any], KeepOpenFile(f))
+
+    return f
+
+
+def format_filename(
+    filename: "t.Union[str, bytes, os.PathLike[str], os.PathLike[bytes]]",
+    shorten: bool = False,
+) -> str:
+    """Format a filename as a string for display. Ensures the filename can be
+    displayed by replacing any invalid bytes or surrogate escapes in the name
+    with the replacement character ``�``.
+
+    Invalid bytes or surrogate escapes will raise an error when written to a
+    stream with ``errors="strict"``. This will typically happen with ``stdout``
+    when the locale is something like ``en_GB.UTF-8``.
+
+    Many scenarios *are* safe to write surrogates though, due to PEP 538 and
+    PEP 540, including:
+
+    -   Writing to ``stderr``, which uses ``errors="backslashreplace"``.
+    -   The system has ``LANG=C.UTF-8``, ``C``, or ``POSIX``. Python opens
+        stdout and stderr with ``errors="surrogateescape"``.
+    -   None of ``LANG/LC_*`` are set. Python assumes ``LANG=C.UTF-8``.
+    -   Python is started in UTF-8 mode  with  ``PYTHONUTF8=1`` or ``-X utf8``.
+        Python opens stdout and stderr with ``errors="surrogateescape"``.
+
+    :param filename: formats a filename for UI display.  This will also convert
+                     the filename into unicode without failing.
+    :param shorten: this optionally shortens the filename to strip of the
+                    path that leads up to it.
+    """
+    if shorten:
+        filename = os.path.basename(filename)
+    else:
+        filename = os.fspath(filename)
+
+    if isinstance(filename, bytes):
+        filename = filename.decode(sys.getfilesystemencoding(), "replace")
+    else:
+        filename = filename.encode("utf-8", "surrogateescape").decode(
+            "utf-8", "replace"
+        )
+
+    return filename
+
+
+def get_app_dir(app_name: str, roaming: bool = True, force_posix: bool = False) -> str:
+    r"""Returns the config folder for the application.  The default behavior
+    is to return whatever is most appropriate for the operating system.
+
+    To give you an idea, for an app called ``"Foo Bar"``, something like
+    the following folders could be returned:
+
+    Mac OS X:
+      ``~/Library/Application Support/Foo Bar``
+    Mac OS X (POSIX):
+      ``~/.foo-bar``
+    Unix:
+      ``~/.config/foo-bar``
+    Unix (POSIX):
+      ``~/.foo-bar``
+    Windows (roaming):
+      ``C:\Users\<user>\AppData\Roaming\Foo Bar``
+    Windows (not roaming):
+      ``C:\Users\<user>\AppData\Local\Foo Bar``
+
+    .. versionadded:: 2.0
+
+    :param app_name: the application name.  This should be properly capitalized
+                     and can contain whitespace.
+    :param roaming: controls if the folder should be roaming or not on Windows.
+                    Has no effect otherwise.
+    :param force_posix: if this is set to `True` then on any POSIX system the
+                        folder will be stored in the home folder with a leading
+                        dot instead of the XDG config home or darwin's
+                        application support folder.
+    """
+    if WIN:
+        key = "APPDATA" if roaming else "LOCALAPPDATA"
+        folder = os.environ.get(key)
+        if folder is None:
+            folder = os.path.expanduser("~")
+        return os.path.join(folder, app_name)
+    if force_posix:
+        return os.path.join(os.path.expanduser(f"~/.{_posixify(app_name)}"))
+    if sys.platform == "darwin":
+        return os.path.join(
+            os.path.expanduser("~/Library/Application Support"), app_name
+        )
+    return os.path.join(
+        os.environ.get("XDG_CONFIG_HOME", os.path.expanduser("~/.config")),
+        _posixify(app_name),
+    )
+
+
+class PacifyFlushWrapper:
+    """This wrapper is used to catch and suppress BrokenPipeErrors resulting
+    from ``.flush()`` being called on broken pipe during the shutdown/final-GC
+    of the Python interpreter. Notably ``.flush()`` is always called on
+    ``sys.stdout`` and ``sys.stderr``. So as to have minimal impact on any
+    other cleanup code, and the case where the underlying file is not a broken
+    pipe, all calls and attributes are proxied.
+    """
+
+    def __init__(self, wrapped: t.IO[t.Any]) -> None:
+        self.wrapped = wrapped
+
+    def flush(self) -> None:
+        try:
+            self.wrapped.flush()
+        except OSError as e:
+            import errno
+
+            if e.errno != errno.EPIPE:
+                raise
+
+    def __getattr__(self, attr: str) -> t.Any:
+        return getattr(self.wrapped, attr)
+
+
+def _detect_program_name(
+    path: t.Optional[str] = None, _main: t.Optional[ModuleType] = None
+) -> str:
+    """Determine the command used to run the program, for use in help
+    text. If a file or entry point was executed, the file name is
+    returned. If ``python -m`` was used to execute a module or package,
+    ``python -m name`` is returned.
+
+    This doesn't try to be too precise, the goal is to give a concise
+    name for help text. Files are only shown as their name without the
+    path. ``python`` is only shown for modules, and the full path to
+    ``sys.executable`` is not shown.
+
+    :param path: The Python file being executed. Python puts this in
+        ``sys.argv[0]``, which is used by default.
+    :param _main: The ``__main__`` module. This should only be passed
+        during internal testing.
+
+    .. versionadded:: 8.0
+        Based on command args detection in the Werkzeug reloader.
+
+    :meta private:
+    """
+    if _main is None:
+        _main = sys.modules["__main__"]
+
+    if not path:
+        path = sys.argv[0]
+
+    # The value of __package__ indicates how Python was called. It may
+    # not exist if a setuptools script is installed as an egg. It may be
+    # set incorrectly for entry points created with pip on Windows.
+    # It is set to "" inside a Shiv or PEX zipapp.
+    if getattr(_main, "__package__", None) in {None, ""} or (
+        os.name == "nt"
+        and _main.__package__ == ""
+        and not os.path.exists(path)
+        and os.path.exists(f"{path}.exe")
+    ):
+        # Executed a file, like "python app.py".
+        return os.path.basename(path)
+
+    # Executed a module, like "python -m example".
+    # Rewritten by Python from "-m script" to "/path/to/script.py".
+    # Need to look at main module to determine how it was executed.
+    py_module = t.cast(str, _main.__package__)
+    name = os.path.splitext(os.path.basename(path))[0]
+
+    # A submodule like "example.cli".
+    if name != "__main__":
+        py_module = f"{py_module}.{name}"
+
+    return f"python -m {py_module.lstrip('.')}"
+
+
+def _expand_args(
+    args: t.Iterable[str],
+    *,
+    user: bool = True,
+    env: bool = True,
+    glob_recursive: bool = True,
+) -> t.List[str]:
+    """Simulate Unix shell expansion with Python functions.
+
+    See :func:`glob.glob`, :func:`os.path.expanduser`, and
+    :func:`os.path.expandvars`.
+
+    This is intended for use on Windows, where the shell does not do any
+    expansion. It may not exactly match what a Unix shell would do.
+
+    :param args: List of command line arguments to expand.
+    :param user: Expand user home directory.
+    :param env: Expand environment variables.
+    :param glob_recursive: ``**`` matches directories recursively.
+
+    .. versionchanged:: 8.1
+        Invalid glob patterns are treated as empty expansions rather
+        than raising an error.
+
+    .. versionadded:: 8.0
+
+    :meta private:
+    """
+    from glob import glob
+
+    out = []
+
+    for arg in args:
+        if user:
+            arg = os.path.expanduser(arg)
+
+        if env:
+            arg = os.path.expandvars(arg)
+
+        try:
+            matches = glob(arg, recursive=glob_recursive)
+        except re.error:
+            matches = []
+
+        if not matches:
+            out.append(arg)
+        else:
+            out.extend(matches)
+
+    return out
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/METADATA
new file mode 100644
index 00000000..a1b5c575
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/METADATA
@@ -0,0 +1,441 @@
+Metadata-Version: 2.1
+Name: colorama
+Version: 0.4.6
+Summary: Cross-platform colored terminal text.
+Project-URL: Homepage, https://github.com/tartley/colorama
+Author-email: Jonathan Hartley <tartley@tartley.com>
+License-File: LICENSE.txt
+Keywords: ansi,color,colour,crossplatform,terminal,text,windows,xplatform
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 2
+Classifier: Programming Language :: Python :: 2.7
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Terminals
+Requires-Python: !=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,!=3.4.*,!=3.5.*,!=3.6.*,>=2.7
+Description-Content-Type: text/x-rst
+
+.. image:: https://img.shields.io/pypi/v/colorama.svg
+    :target: https://pypi.org/project/colorama/
+    :alt: Latest Version
+
+.. image:: https://img.shields.io/pypi/pyversions/colorama.svg
+    :target: https://pypi.org/project/colorama/
+    :alt: Supported Python versions
+
+.. image:: https://github.com/tartley/colorama/actions/workflows/test.yml/badge.svg
+    :target: https://github.com/tartley/colorama/actions/workflows/test.yml
+    :alt: Build Status
+
+Colorama
+========
+
+Makes ANSI escape character sequences (for producing colored terminal text and
+cursor positioning) work under MS Windows.
+
+.. |donate| image:: https://www.paypalobjects.com/en_US/i/btn/btn_donate_SM.gif
+  :target: https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=2MZ9D2GMLYCUJ&item_name=Colorama&currency_code=USD
+  :alt: Donate with Paypal
+
+`PyPI for releases <https://pypi.org/project/colorama/>`_ |
+`Github for source <https://github.com/tartley/colorama>`_ |
+`Colorama for enterprise on Tidelift <https://github.com/tartley/colorama/blob/master/ENTERPRISE.md>`_
+
+If you find Colorama useful, please |donate| to the authors. Thank you!
+
+Installation
+------------
+
+Tested on CPython 2.7, 3.7, 3.8, 3.9 and 3.10 and Pypy 2.7 and 3.8.
+
+No requirements other than the standard library.
+
+.. code-block:: bash
+
+    pip install colorama
+    # or
+    conda install -c anaconda colorama
+
+Description
+-----------
+
+ANSI escape character sequences have long been used to produce colored terminal
+text and cursor positioning on Unix and Macs. Colorama makes this work on
+Windows, too, by wrapping ``stdout``, stripping ANSI sequences it finds (which
+would appear as gobbledygook in the output), and converting them into the
+appropriate win32 calls to modify the state of the terminal. On other platforms,
+Colorama does nothing.
+
+This has the upshot of providing a simple cross-platform API for printing
+colored terminal text from Python, and has the happy side-effect that existing
+applications or libraries which use ANSI sequences to produce colored output on
+Linux or Macs can now also work on Windows, simply by calling
+``colorama.just_fix_windows_console()`` (since v0.4.6) or ``colorama.init()``
+(all versions, but may have other side-effects – see below).
+
+An alternative approach is to install ``ansi.sys`` on Windows machines, which
+provides the same behaviour for all applications running in terminals. Colorama
+is intended for situations where that isn't easy (e.g., maybe your app doesn't
+have an installer.)
+
+Demo scripts in the source code repository print some colored text using
+ANSI sequences. Compare their output under Gnome-terminal's built in ANSI
+handling, versus on Windows Command-Prompt using Colorama:
+
+.. image:: https://github.com/tartley/colorama/raw/master/screenshots/ubuntu-demo.png
+    :width: 661
+    :height: 357
+    :alt: ANSI sequences on Ubuntu under gnome-terminal.
+
+.. image:: https://github.com/tartley/colorama/raw/master/screenshots/windows-demo.png
+    :width: 668
+    :height: 325
+    :alt: Same ANSI sequences on Windows, using Colorama.
+
+These screenshots show that, on Windows, Colorama does not support ANSI 'dim
+text'; it looks the same as 'normal text'.
+
+Usage
+-----
+
+Initialisation
+..............
+
+If the only thing you want from Colorama is to get ANSI escapes to work on
+Windows, then run:
+
+.. code-block:: python
+
+    from colorama import just_fix_windows_console
+    just_fix_windows_console()
+
+If you're on a recent version of Windows 10 or better, and your stdout/stderr
+are pointing to a Windows console, then this will flip the magic configuration
+switch to enable Windows' built-in ANSI support.
+
+If you're on an older version of Windows, and your stdout/stderr are pointing to
+a Windows console, then this will wrap ``sys.stdout`` and/or ``sys.stderr`` in a
+magic file object that intercepts ANSI escape sequences and issues the
+appropriate Win32 calls to emulate them.
+
+In all other circumstances, it does nothing whatsoever. Basically the idea is
+that this makes Windows act like Unix with respect to ANSI escape handling.
+
+It's safe to call this function multiple times. It's safe to call this function
+on non-Windows platforms, but it won't do anything. It's safe to call this
+function when one or both of your stdout/stderr are redirected to a file – it
+won't do anything to those streams.
+
+Alternatively, you can use the older interface with more features (but also more
+potential footguns):
+
+.. code-block:: python
+
+    from colorama import init
+    init()
+
+This does the same thing as ``just_fix_windows_console``, except for the
+following differences:
+
+- It's not safe to call ``init`` multiple times; you can end up with multiple
+  layers of wrapping and broken ANSI support.
+
+- Colorama will apply a heuristic to guess whether stdout/stderr support ANSI,
+  and if it thinks they don't, then it will wrap ``sys.stdout`` and
+  ``sys.stderr`` in a magic file object that strips out ANSI escape sequences
+  before printing them. This happens on all platforms, and can be convenient if
+  you want to write your code to emit ANSI escape sequences unconditionally, and
+  let Colorama decide whether they should actually be output. But note that
+  Colorama's heuristic is not particularly clever.
+
+- ``init`` also accepts explicit keyword args to enable/disable various
+  functionality – see below.
+
+To stop using Colorama before your program exits, simply call ``deinit()``.
+This will restore ``stdout`` and ``stderr`` to their original values, so that
+Colorama is disabled. To resume using Colorama again, call ``reinit()``; it is
+cheaper than calling ``init()`` again (but does the same thing).
+
+Most users should depend on ``colorama >= 0.4.6``, and use
+``just_fix_windows_console``. The old ``init`` interface will be supported
+indefinitely for backwards compatibility, but we don't plan to fix any issues
+with it, also for backwards compatibility.
+
+Colored Output
+..............
+
+Cross-platform printing of colored text can then be done using Colorama's
+constant shorthand for ANSI escape sequences. These are deliberately
+rudimentary, see below.
+
+.. code-block:: python
+
+    from colorama import Fore, Back, Style
+    print(Fore.RED + 'some red text')
+    print(Back.GREEN + 'and with a green background')
+    print(Style.DIM + 'and in dim text')
+    print(Style.RESET_ALL)
+    print('back to normal now')
+
+...or simply by manually printing ANSI sequences from your own code:
+
+.. code-block:: python
+
+    print('\033[31m' + 'some red text')
+    print('\033[39m') # and reset to default color
+
+...or, Colorama can be used in conjunction with existing ANSI libraries
+such as the venerable `Termcolor <https://pypi.org/project/termcolor/>`_
+the fabulous `Blessings <https://pypi.org/project/blessings/>`_,
+or the incredible `_Rich <https://pypi.org/project/rich/>`_.
+
+If you wish Colorama's Fore, Back and Style constants were more capable,
+then consider using one of the above highly capable libraries to generate
+colors, etc, and use Colorama just for its primary purpose: to convert
+those ANSI sequences to also work on Windows:
+
+SIMILARLY, do not send PRs adding the generation of new ANSI types to Colorama.
+We are only interested in converting ANSI codes to win32 API calls, not
+shortcuts like the above to generate ANSI characters.
+
+.. code-block:: python
+
+    from colorama import just_fix_windows_console
+    from termcolor import colored
+
+    # use Colorama to make Termcolor work on Windows too
+    just_fix_windows_console()
+
+    # then use Termcolor for all colored text output
+    print(colored('Hello, World!', 'green', 'on_red'))
+
+Available formatting constants are::
+
+    Fore: BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE, RESET.
+    Back: BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE, RESET.
+    Style: DIM, NORMAL, BRIGHT, RESET_ALL
+
+``Style.RESET_ALL`` resets foreground, background, and brightness. Colorama will
+perform this reset automatically on program exit.
+
+These are fairly well supported, but not part of the standard::
+
+    Fore: LIGHTBLACK_EX, LIGHTRED_EX, LIGHTGREEN_EX, LIGHTYELLOW_EX, LIGHTBLUE_EX, LIGHTMAGENTA_EX, LIGHTCYAN_EX, LIGHTWHITE_EX
+    Back: LIGHTBLACK_EX, LIGHTRED_EX, LIGHTGREEN_EX, LIGHTYELLOW_EX, LIGHTBLUE_EX, LIGHTMAGENTA_EX, LIGHTCYAN_EX, LIGHTWHITE_EX
+
+Cursor Positioning
+..................
+
+ANSI codes to reposition the cursor are supported. See ``demos/demo06.py`` for
+an example of how to generate them.
+
+Init Keyword Args
+.................
+
+``init()`` accepts some ``**kwargs`` to override default behaviour.
+
+init(autoreset=False):
+    If you find yourself repeatedly sending reset sequences to turn off color
+    changes at the end of every print, then ``init(autoreset=True)`` will
+    automate that:
+
+    .. code-block:: python
+
+        from colorama import init
+        init(autoreset=True)
+        print(Fore.RED + 'some red text')
+        print('automatically back to default color again')
+
+init(strip=None):
+    Pass ``True`` or ``False`` to override whether ANSI codes should be
+    stripped from the output. The default behaviour is to strip if on Windows
+    or if output is redirected (not a tty).
+
+init(convert=None):
+    Pass ``True`` or ``False`` to override whether to convert ANSI codes in the
+    output into win32 calls. The default behaviour is to convert if on Windows
+    and output is to a tty (terminal).
+
+init(wrap=True):
+    On Windows, Colorama works by replacing ``sys.stdout`` and ``sys.stderr``
+    with proxy objects, which override the ``.write()`` method to do their work.
+    If this wrapping causes you problems, then this can be disabled by passing
+    ``init(wrap=False)``. The default behaviour is to wrap if ``autoreset`` or
+    ``strip`` or ``convert`` are True.
+
+    When wrapping is disabled, colored printing on non-Windows platforms will
+    continue to work as normal. To do cross-platform colored output, you can
+    use Colorama's ``AnsiToWin32`` proxy directly:
+
+    .. code-block:: python
+
+        import sys
+        from colorama import init, AnsiToWin32
+        init(wrap=False)
+        stream = AnsiToWin32(sys.stderr).stream
+
+        # Python 2
+        print >>stream, Fore.BLUE + 'blue text on stderr'
+
+        # Python 3
+        print(Fore.BLUE + 'blue text on stderr', file=stream)
+
+Recognised ANSI Sequences
+.........................
+
+ANSI sequences generally take the form::
+
+    ESC [ <param> ; <param> ... <command>
+
+Where ``<param>`` is an integer, and ``<command>`` is a single letter. Zero or
+more params are passed to a ``<command>``. If no params are passed, it is
+generally synonymous with passing a single zero. No spaces exist in the
+sequence; they have been inserted here simply to read more easily.
+
+The only ANSI sequences that Colorama converts into win32 calls are::
+
+    ESC [ 0 m       # reset all (colors and brightness)
+    ESC [ 1 m       # bright
+    ESC [ 2 m       # dim (looks same as normal brightness)
+    ESC [ 22 m      # normal brightness
+
+    # FOREGROUND:
+    ESC [ 30 m      # black
+    ESC [ 31 m      # red
+    ESC [ 32 m      # green
+    ESC [ 33 m      # yellow
+    ESC [ 34 m      # blue
+    ESC [ 35 m      # magenta
+    ESC [ 36 m      # cyan
+    ESC [ 37 m      # white
+    ESC [ 39 m      # reset
+
+    # BACKGROUND
+    ESC [ 40 m      # black
+    ESC [ 41 m      # red
+    ESC [ 42 m      # green
+    ESC [ 43 m      # yellow
+    ESC [ 44 m      # blue
+    ESC [ 45 m      # magenta
+    ESC [ 46 m      # cyan
+    ESC [ 47 m      # white
+    ESC [ 49 m      # reset
+
+    # cursor positioning
+    ESC [ y;x H     # position cursor at x across, y down
+    ESC [ y;x f     # position cursor at x across, y down
+    ESC [ n A       # move cursor n lines up
+    ESC [ n B       # move cursor n lines down
+    ESC [ n C       # move cursor n characters forward
+    ESC [ n D       # move cursor n characters backward
+
+    # clear the screen
+    ESC [ mode J    # clear the screen
+
+    # clear the line
+    ESC [ mode K    # clear the line
+
+Multiple numeric params to the ``'m'`` command can be combined into a single
+sequence::
+
+    ESC [ 36 ; 45 ; 1 m     # bright cyan text on magenta background
+
+All other ANSI sequences of the form ``ESC [ <param> ; <param> ... <command>``
+are silently stripped from the output on Windows.
+
+Any other form of ANSI sequence, such as single-character codes or alternative
+initial characters, are not recognised or stripped. It would be cool to add
+them though. Let me know if it would be useful for you, via the Issues on
+GitHub.
+
+Status & Known Problems
+-----------------------
+
+I've personally only tested it on Windows XP (CMD, Console2), Ubuntu
+(gnome-terminal, xterm), and OS X.
+
+Some valid ANSI sequences aren't recognised.
+
+If you're hacking on the code, see `README-hacking.md`_. ESPECIALLY, see the
+explanation there of why we do not want PRs that allow Colorama to generate new
+types of ANSI codes.
+
+See outstanding issues and wish-list:
+https://github.com/tartley/colorama/issues
+
+If anything doesn't work for you, or doesn't do what you expected or hoped for,
+I'd love to hear about it on that issues list, would be delighted by patches,
+and would be happy to grant commit access to anyone who submits a working patch
+or two.
+
+.. _README-hacking.md: README-hacking.md
+
+License
+-------
+
+Copyright Jonathan Hartley & Arnon Yaari, 2013-2020. BSD 3-Clause license; see
+LICENSE file.
+
+Professional support
+--------------------
+
+.. |tideliftlogo| image:: https://cdn2.hubspot.net/hubfs/4008838/website/logos/logos_for_download/Tidelift_primary-shorthand-logo.png
+   :alt: Tidelift
+   :target: https://tidelift.com/subscription/pkg/pypi-colorama?utm_source=pypi-colorama&utm_medium=referral&utm_campaign=readme
+
+.. list-table::
+   :widths: 10 100
+
+   * - |tideliftlogo|
+     - Professional support for colorama is available as part of the
+       `Tidelift Subscription`_.
+       Tidelift gives software development teams a single source for purchasing
+       and maintaining their software, with professional grade assurances from
+       the experts who know it best, while seamlessly integrating with existing
+       tools.
+
+.. _Tidelift Subscription: https://tidelift.com/subscription/pkg/pypi-colorama?utm_source=pypi-colorama&utm_medium=referral&utm_campaign=readme
+
+Thanks
+------
+
+See the CHANGELOG for more thanks!
+
+* Marc Schlaich (schlamar) for a ``setup.py`` fix for Python2.5.
+* Marc Abramowitz, reported & fixed a crash on exit with closed ``stdout``,
+  providing a solution to issue #7's setuptools/distutils debate,
+  and other fixes.
+* User 'eryksun', for guidance on correctly instantiating ``ctypes.windll``.
+* Matthew McCormick for politely pointing out a longstanding crash on non-Win.
+* Ben Hoyt, for a magnificent fix under 64-bit Windows.
+* Jesse at Empty Square for submitting a fix for examples in the README.
+* User 'jamessp', an observant documentation fix for cursor positioning.
+* User 'vaal1239', Dave Mckee & Lackner Kristof for a tiny but much-needed Win7
+  fix.
+* Julien Stuyck, for wisely suggesting Python3 compatible updates to README.
+* Daniel Griffith for multiple fabulous patches.
+* Oscar Lesta for a valuable fix to stop ANSI chars being sent to non-tty
+  output.
+* Roger Binns, for many suggestions, valuable feedback, & bug reports.
+* Tim Golden for thought and much appreciated feedback on the initial idea.
+* User 'Zearin' for updates to the README file.
+* John Szakmeister for adding support for light colors
+* Charles Merriam for adding documentation to demos
+* Jurko for a fix on 64-bit Windows CPython2.5 w/o ctypes
+* Florian Bruhin for a fix when stdout or stderr are None
+* Thomas Weininger for fixing ValueError on Windows
+* Remi Rampin for better Github integration and fixes to the README file
+* Simeon Visser for closing a file handle using 'with' and updating classifiers
+  to include Python 3.3 and 3.4
+* Andy Neff for fixing RESET of LIGHT_EX colors.
+* Jonathan Hartley for the initial idea and implementation.
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/RECORD
new file mode 100644
index 00000000..f71babf1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/RECORD
@@ -0,0 +1,32 @@
+colorama-0.4.6.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+colorama-0.4.6.dist-info/METADATA,sha256=e67SnrUMOym9sz_4TjF3vxvAV4T3aF7NyqRHHH3YEMw,17158
+colorama-0.4.6.dist-info/RECORD,,
+colorama-0.4.6.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+colorama-0.4.6.dist-info/WHEEL,sha256=cdcF4Fbd0FPtw2EMIOwH-3rSOTUdTCeOSXRMD1iLUb8,105
+colorama-0.4.6.dist-info/licenses/LICENSE.txt,sha256=ysNcAmhuXQSlpxQL-zs25zrtSWZW6JEQLkKIhteTAxg,1491
+colorama/__init__.py,sha256=wePQA4U20tKgYARySLEC047ucNX-g8pRLpYBuiHlLb8,266
+colorama/__pycache__/__init__.cpython-312.pyc,,
+colorama/__pycache__/ansi.cpython-312.pyc,,
+colorama/__pycache__/ansitowin32.cpython-312.pyc,,
+colorama/__pycache__/initialise.cpython-312.pyc,,
+colorama/__pycache__/win32.cpython-312.pyc,,
+colorama/__pycache__/winterm.cpython-312.pyc,,
+colorama/ansi.py,sha256=Top4EeEuaQdBWdteKMEcGOTeKeF19Q-Wo_6_Cj5kOzQ,2522
+colorama/ansitowin32.py,sha256=vPNYa3OZbxjbuFyaVo0Tmhmy1FZ1lKMWCnT7odXpItk,11128
+colorama/initialise.py,sha256=-hIny86ClXo39ixh5iSCfUIa2f_h_bgKRDW7gqs-KLU,3325
+colorama/tests/__init__.py,sha256=MkgPAEzGQd-Rq0w0PZXSX2LadRWhUECcisJY8lSrm4Q,75
+colorama/tests/__pycache__/__init__.cpython-312.pyc,,
+colorama/tests/__pycache__/ansi_test.cpython-312.pyc,,
+colorama/tests/__pycache__/ansitowin32_test.cpython-312.pyc,,
+colorama/tests/__pycache__/initialise_test.cpython-312.pyc,,
+colorama/tests/__pycache__/isatty_test.cpython-312.pyc,,
+colorama/tests/__pycache__/utils.cpython-312.pyc,,
+colorama/tests/__pycache__/winterm_test.cpython-312.pyc,,
+colorama/tests/ansi_test.py,sha256=FeViDrUINIZcr505PAxvU4AjXz1asEiALs9GXMhwRaE,2839
+colorama/tests/ansitowin32_test.py,sha256=RN7AIhMJ5EqDsYaCjVo-o4u8JzDD4ukJbmevWKS70rY,10678
+colorama/tests/initialise_test.py,sha256=BbPy-XfyHwJ6zKozuQOvNvQZzsx9vdb_0bYXn7hsBTc,6741
+colorama/tests/isatty_test.py,sha256=Pg26LRpv0yQDB5Ac-sxgVXG7hsA1NYvapFgApZfYzZg,1866
+colorama/tests/utils.py,sha256=1IIRylG39z5-dzq09R_ngufxyPZxgldNbrxKxUGwGKE,1079
+colorama/tests/winterm_test.py,sha256=qoWFPEjym5gm2RuMwpf3pOis3a5r_PJZFCzK254JL8A,3709
+colorama/win32.py,sha256=YQOKwMTwtGBbsY4dL5HYTvwTeP9wIQra5MvPNddpxZs,6181
+colorama/winterm.py,sha256=XCQFDHjPi6AHYNdZwy0tA02H-Jh48Jp-HvCjeLeLp3U,7134
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/WHEEL
new file mode 100644
index 00000000..d79189fd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.11.1
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/licenses/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/licenses/LICENSE.txt
new file mode 100644
index 00000000..3105888e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama-0.4.6.dist-info/licenses/LICENSE.txt
@@ -0,0 +1,27 @@
+Copyright (c) 2010 Jonathan Hartley
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+* Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+
+* Redistributions in binary form must reproduce the above copyright notice,
+  this list of conditions and the following disclaimer in the documentation
+  and/or other materials provided with the distribution.
+
+* Neither the name of the copyright holders, nor those of its contributors
+  may be used to endorse or promote products derived from this software without
+  specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__init__.py b/backend/venv/lib/python3.12/site-packages/colorama/__init__.py
new file mode 100644
index 00000000..383101cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/__init__.py
@@ -0,0 +1,7 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+from .initialise import init, deinit, reinit, colorama_text, just_fix_windows_console
+from .ansi import Fore, Back, Style, Cursor
+from .ansitowin32 import AnsiToWin32
+
+__version__ = '0.4.6'
+
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..01a017ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansi.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansi.cpython-312.pyc
new file mode 100644
index 00000000..aac904bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansi.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansitowin32.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansitowin32.cpython-312.pyc
new file mode 100644
index 00000000..8a5cda2c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/ansitowin32.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/initialise.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/initialise.cpython-312.pyc
new file mode 100644
index 00000000..78aacf29
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/initialise.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/win32.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/win32.cpython-312.pyc
new file mode 100644
index 00000000..e321f089
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/win32.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/winterm.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/winterm.cpython-312.pyc
new file mode 100644
index 00000000..86915428
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/__pycache__/winterm.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/ansi.py b/backend/venv/lib/python3.12/site-packages/colorama/ansi.py
new file mode 100644
index 00000000..11ec695f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/ansi.py
@@ -0,0 +1,102 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+'''
+This module generates ANSI character codes to printing colors to terminals.
+See: http://en.wikipedia.org/wiki/ANSI_escape_code
+'''
+
+CSI = '\033['
+OSC = '\033]'
+BEL = '\a'
+
+
+def code_to_chars(code):
+    return CSI + str(code) + 'm'
+
+def set_title(title):
+    return OSC + '2;' + title + BEL
+
+def clear_screen(mode=2):
+    return CSI + str(mode) + 'J'
+
+def clear_line(mode=2):
+    return CSI + str(mode) + 'K'
+
+
+class AnsiCodes(object):
+    def __init__(self):
+        # the subclasses declare class attributes which are numbers.
+        # Upon instantiation we define instance attributes, which are the same
+        # as the class attributes but wrapped with the ANSI escape sequence
+        for name in dir(self):
+            if not name.startswith('_'):
+                value = getattr(self, name)
+                setattr(self, name, code_to_chars(value))
+
+
+class AnsiCursor(object):
+    def UP(self, n=1):
+        return CSI + str(n) + 'A'
+    def DOWN(self, n=1):
+        return CSI + str(n) + 'B'
+    def FORWARD(self, n=1):
+        return CSI + str(n) + 'C'
+    def BACK(self, n=1):
+        return CSI + str(n) + 'D'
+    def POS(self, x=1, y=1):
+        return CSI + str(y) + ';' + str(x) + 'H'
+
+
+class AnsiFore(AnsiCodes):
+    BLACK           = 30
+    RED             = 31
+    GREEN           = 32
+    YELLOW          = 33
+    BLUE            = 34
+    MAGENTA         = 35
+    CYAN            = 36
+    WHITE           = 37
+    RESET           = 39
+
+    # These are fairly well supported, but not part of the standard.
+    LIGHTBLACK_EX   = 90
+    LIGHTRED_EX     = 91
+    LIGHTGREEN_EX   = 92
+    LIGHTYELLOW_EX  = 93
+    LIGHTBLUE_EX    = 94
+    LIGHTMAGENTA_EX = 95
+    LIGHTCYAN_EX    = 96
+    LIGHTWHITE_EX   = 97
+
+
+class AnsiBack(AnsiCodes):
+    BLACK           = 40
+    RED             = 41
+    GREEN           = 42
+    YELLOW          = 43
+    BLUE            = 44
+    MAGENTA         = 45
+    CYAN            = 46
+    WHITE           = 47
+    RESET           = 49
+
+    # These are fairly well supported, but not part of the standard.
+    LIGHTBLACK_EX   = 100
+    LIGHTRED_EX     = 101
+    LIGHTGREEN_EX   = 102
+    LIGHTYELLOW_EX  = 103
+    LIGHTBLUE_EX    = 104
+    LIGHTMAGENTA_EX = 105
+    LIGHTCYAN_EX    = 106
+    LIGHTWHITE_EX   = 107
+
+
+class AnsiStyle(AnsiCodes):
+    BRIGHT    = 1
+    DIM       = 2
+    NORMAL    = 22
+    RESET_ALL = 0
+
+Fore   = AnsiFore()
+Back   = AnsiBack()
+Style  = AnsiStyle()
+Cursor = AnsiCursor()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/ansitowin32.py b/backend/venv/lib/python3.12/site-packages/colorama/ansitowin32.py
new file mode 100644
index 00000000..abf209e6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/ansitowin32.py
@@ -0,0 +1,277 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import re
+import sys
+import os
+
+from .ansi import AnsiFore, AnsiBack, AnsiStyle, Style, BEL
+from .winterm import enable_vt_processing, WinTerm, WinColor, WinStyle
+from .win32 import windll, winapi_test
+
+
+winterm = None
+if windll is not None:
+    winterm = WinTerm()
+
+
+class StreamWrapper(object):
+    '''
+    Wraps a stream (such as stdout), acting as a transparent proxy for all
+    attribute access apart from method 'write()', which is delegated to our
+    Converter instance.
+    '''
+    def __init__(self, wrapped, converter):
+        # double-underscore everything to prevent clashes with names of
+        # attributes on the wrapped stream object.
+        self.__wrapped = wrapped
+        self.__convertor = converter
+
+    def __getattr__(self, name):
+        return getattr(self.__wrapped, name)
+
+    def __enter__(self, *args, **kwargs):
+        # special method lookup bypasses __getattr__/__getattribute__, see
+        # https://stackoverflow.com/questions/12632894/why-doesnt-getattr-work-with-exit
+        # thus, contextlib magic methods are not proxied via __getattr__
+        return self.__wrapped.__enter__(*args, **kwargs)
+
+    def __exit__(self, *args, **kwargs):
+        return self.__wrapped.__exit__(*args, **kwargs)
+
+    def __setstate__(self, state):
+        self.__dict__ = state
+
+    def __getstate__(self):
+        return self.__dict__
+
+    def write(self, text):
+        self.__convertor.write(text)
+
+    def isatty(self):
+        stream = self.__wrapped
+        if 'PYCHARM_HOSTED' in os.environ:
+            if stream is not None and (stream is sys.__stdout__ or stream is sys.__stderr__):
+                return True
+        try:
+            stream_isatty = stream.isatty
+        except AttributeError:
+            return False
+        else:
+            return stream_isatty()
+
+    @property
+    def closed(self):
+        stream = self.__wrapped
+        try:
+            return stream.closed
+        # AttributeError in the case that the stream doesn't support being closed
+        # ValueError for the case that the stream has already been detached when atexit runs
+        except (AttributeError, ValueError):
+            return True
+
+
+class AnsiToWin32(object):
+    '''
+    Implements a 'write()' method which, on Windows, will strip ANSI character
+    sequences from the text, and if outputting to a tty, will convert them into
+    win32 function calls.
+    '''
+    ANSI_CSI_RE = re.compile('\001?\033\\[((?:\\d|;)*)([a-zA-Z])\002?')   # Control Sequence Introducer
+    ANSI_OSC_RE = re.compile('\001?\033\\]([^\a]*)(\a)\002?')             # Operating System Command
+
+    def __init__(self, wrapped, convert=None, strip=None, autoreset=False):
+        # The wrapped stream (normally sys.stdout or sys.stderr)
+        self.wrapped = wrapped
+
+        # should we reset colors to defaults after every .write()
+        self.autoreset = autoreset
+
+        # create the proxy wrapping our output stream
+        self.stream = StreamWrapper(wrapped, self)
+
+        on_windows = os.name == 'nt'
+        # We test if the WinAPI works, because even if we are on Windows
+        # we may be using a terminal that doesn't support the WinAPI
+        # (e.g. Cygwin Terminal). In this case it's up to the terminal
+        # to support the ANSI codes.
+        conversion_supported = on_windows and winapi_test()
+        try:
+            fd = wrapped.fileno()
+        except Exception:
+            fd = -1
+        system_has_native_ansi = not on_windows or enable_vt_processing(fd)
+        have_tty = not self.stream.closed and self.stream.isatty()
+        need_conversion = conversion_supported and not system_has_native_ansi
+
+        # should we strip ANSI sequences from our output?
+        if strip is None:
+            strip = need_conversion or not have_tty
+        self.strip = strip
+
+        # should we should convert ANSI sequences into win32 calls?
+        if convert is None:
+            convert = need_conversion and have_tty
+        self.convert = convert
+
+        # dict of ansi codes to win32 functions and parameters
+        self.win32_calls = self.get_win32_calls()
+
+        # are we wrapping stderr?
+        self.on_stderr = self.wrapped is sys.stderr
+
+    def should_wrap(self):
+        '''
+        True if this class is actually needed. If false, then the output
+        stream will not be affected, nor will win32 calls be issued, so
+        wrapping stdout is not actually required. This will generally be
+        False on non-Windows platforms, unless optional functionality like
+        autoreset has been requested using kwargs to init()
+        '''
+        return self.convert or self.strip or self.autoreset
+
+    def get_win32_calls(self):
+        if self.convert and winterm:
+            return {
+                AnsiStyle.RESET_ALL: (winterm.reset_all, ),
+                AnsiStyle.BRIGHT: (winterm.style, WinStyle.BRIGHT),
+                AnsiStyle.DIM: (winterm.style, WinStyle.NORMAL),
+                AnsiStyle.NORMAL: (winterm.style, WinStyle.NORMAL),
+                AnsiFore.BLACK: (winterm.fore, WinColor.BLACK),
+                AnsiFore.RED: (winterm.fore, WinColor.RED),
+                AnsiFore.GREEN: (winterm.fore, WinColor.GREEN),
+                AnsiFore.YELLOW: (winterm.fore, WinColor.YELLOW),
+                AnsiFore.BLUE: (winterm.fore, WinColor.BLUE),
+                AnsiFore.MAGENTA: (winterm.fore, WinColor.MAGENTA),
+                AnsiFore.CYAN: (winterm.fore, WinColor.CYAN),
+                AnsiFore.WHITE: (winterm.fore, WinColor.GREY),
+                AnsiFore.RESET: (winterm.fore, ),
+                AnsiFore.LIGHTBLACK_EX: (winterm.fore, WinColor.BLACK, True),
+                AnsiFore.LIGHTRED_EX: (winterm.fore, WinColor.RED, True),
+                AnsiFore.LIGHTGREEN_EX: (winterm.fore, WinColor.GREEN, True),
+                AnsiFore.LIGHTYELLOW_EX: (winterm.fore, WinColor.YELLOW, True),
+                AnsiFore.LIGHTBLUE_EX: (winterm.fore, WinColor.BLUE, True),
+                AnsiFore.LIGHTMAGENTA_EX: (winterm.fore, WinColor.MAGENTA, True),
+                AnsiFore.LIGHTCYAN_EX: (winterm.fore, WinColor.CYAN, True),
+                AnsiFore.LIGHTWHITE_EX: (winterm.fore, WinColor.GREY, True),
+                AnsiBack.BLACK: (winterm.back, WinColor.BLACK),
+                AnsiBack.RED: (winterm.back, WinColor.RED),
+                AnsiBack.GREEN: (winterm.back, WinColor.GREEN),
+                AnsiBack.YELLOW: (winterm.back, WinColor.YELLOW),
+                AnsiBack.BLUE: (winterm.back, WinColor.BLUE),
+                AnsiBack.MAGENTA: (winterm.back, WinColor.MAGENTA),
+                AnsiBack.CYAN: (winterm.back, WinColor.CYAN),
+                AnsiBack.WHITE: (winterm.back, WinColor.GREY),
+                AnsiBack.RESET: (winterm.back, ),
+                AnsiBack.LIGHTBLACK_EX: (winterm.back, WinColor.BLACK, True),
+                AnsiBack.LIGHTRED_EX: (winterm.back, WinColor.RED, True),
+                AnsiBack.LIGHTGREEN_EX: (winterm.back, WinColor.GREEN, True),
+                AnsiBack.LIGHTYELLOW_EX: (winterm.back, WinColor.YELLOW, True),
+                AnsiBack.LIGHTBLUE_EX: (winterm.back, WinColor.BLUE, True),
+                AnsiBack.LIGHTMAGENTA_EX: (winterm.back, WinColor.MAGENTA, True),
+                AnsiBack.LIGHTCYAN_EX: (winterm.back, WinColor.CYAN, True),
+                AnsiBack.LIGHTWHITE_EX: (winterm.back, WinColor.GREY, True),
+            }
+        return dict()
+
+    def write(self, text):
+        if self.strip or self.convert:
+            self.write_and_convert(text)
+        else:
+            self.wrapped.write(text)
+            self.wrapped.flush()
+        if self.autoreset:
+            self.reset_all()
+
+
+    def reset_all(self):
+        if self.convert:
+            self.call_win32('m', (0,))
+        elif not self.strip and not self.stream.closed:
+            self.wrapped.write(Style.RESET_ALL)
+
+
+    def write_and_convert(self, text):
+        '''
+        Write the given text to our wrapped stream, stripping any ANSI
+        sequences from the text, and optionally converting them into win32
+        calls.
+        '''
+        cursor = 0
+        text = self.convert_osc(text)
+        for match in self.ANSI_CSI_RE.finditer(text):
+            start, end = match.span()
+            self.write_plain_text(text, cursor, start)
+            self.convert_ansi(*match.groups())
+            cursor = end
+        self.write_plain_text(text, cursor, len(text))
+
+
+    def write_plain_text(self, text, start, end):
+        if start < end:
+            self.wrapped.write(text[start:end])
+            self.wrapped.flush()
+
+
+    def convert_ansi(self, paramstring, command):
+        if self.convert:
+            params = self.extract_params(command, paramstring)
+            self.call_win32(command, params)
+
+
+    def extract_params(self, command, paramstring):
+        if command in 'Hf':
+            params = tuple(int(p) if len(p) != 0 else 1 for p in paramstring.split(';'))
+            while len(params) < 2:
+                # defaults:
+                params = params + (1,)
+        else:
+            params = tuple(int(p) for p in paramstring.split(';') if len(p) != 0)
+            if len(params) == 0:
+                # defaults:
+                if command in 'JKm':
+                    params = (0,)
+                elif command in 'ABCD':
+                    params = (1,)
+
+        return params
+
+
+    def call_win32(self, command, params):
+        if command == 'm':
+            for param in params:
+                if param in self.win32_calls:
+                    func_args = self.win32_calls[param]
+                    func = func_args[0]
+                    args = func_args[1:]
+                    kwargs = dict(on_stderr=self.on_stderr)
+                    func(*args, **kwargs)
+        elif command in 'J':
+            winterm.erase_screen(params[0], on_stderr=self.on_stderr)
+        elif command in 'K':
+            winterm.erase_line(params[0], on_stderr=self.on_stderr)
+        elif command in 'Hf':     # cursor position - absolute
+            winterm.set_cursor_position(params, on_stderr=self.on_stderr)
+        elif command in 'ABCD':   # cursor position - relative
+            n = params[0]
+            # A - up, B - down, C - forward, D - back
+            x, y = {'A': (0, -n), 'B': (0, n), 'C': (n, 0), 'D': (-n, 0)}[command]
+            winterm.cursor_adjust(x, y, on_stderr=self.on_stderr)
+
+
+    def convert_osc(self, text):
+        for match in self.ANSI_OSC_RE.finditer(text):
+            start, end = match.span()
+            text = text[:start] + text[end:]
+            paramstring, command = match.groups()
+            if command == BEL:
+                if paramstring.count(";") == 1:
+                    params = paramstring.split(";")
+                    # 0 - change title and icon (we will only change title)
+                    # 1 - change icon (we don't support this)
+                    # 2 - change title
+                    if params[0] in '02':
+                        winterm.set_title(params[1])
+        return text
+
+
+    def flush(self):
+        self.wrapped.flush()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/initialise.py b/backend/venv/lib/python3.12/site-packages/colorama/initialise.py
new file mode 100644
index 00000000..d5fd4b71
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/initialise.py
@@ -0,0 +1,121 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import atexit
+import contextlib
+import sys
+
+from .ansitowin32 import AnsiToWin32
+
+
+def _wipe_internal_state_for_tests():
+    global orig_stdout, orig_stderr
+    orig_stdout = None
+    orig_stderr = None
+
+    global wrapped_stdout, wrapped_stderr
+    wrapped_stdout = None
+    wrapped_stderr = None
+
+    global atexit_done
+    atexit_done = False
+
+    global fixed_windows_console
+    fixed_windows_console = False
+
+    try:
+        # no-op if it wasn't registered
+        atexit.unregister(reset_all)
+    except AttributeError:
+        # python 2: no atexit.unregister. Oh well, we did our best.
+        pass
+
+
+def reset_all():
+    if AnsiToWin32 is not None:    # Issue #74: objects might become None at exit
+        AnsiToWin32(orig_stdout).reset_all()
+
+
+def init(autoreset=False, convert=None, strip=None, wrap=True):
+
+    if not wrap and any([autoreset, convert, strip]):
+        raise ValueError('wrap=False conflicts with any other arg=True')
+
+    global wrapped_stdout, wrapped_stderr
+    global orig_stdout, orig_stderr
+
+    orig_stdout = sys.stdout
+    orig_stderr = sys.stderr
+
+    if sys.stdout is None:
+        wrapped_stdout = None
+    else:
+        sys.stdout = wrapped_stdout = \
+            wrap_stream(orig_stdout, convert, strip, autoreset, wrap)
+    if sys.stderr is None:
+        wrapped_stderr = None
+    else:
+        sys.stderr = wrapped_stderr = \
+            wrap_stream(orig_stderr, convert, strip, autoreset, wrap)
+
+    global atexit_done
+    if not atexit_done:
+        atexit.register(reset_all)
+        atexit_done = True
+
+
+def deinit():
+    if orig_stdout is not None:
+        sys.stdout = orig_stdout
+    if orig_stderr is not None:
+        sys.stderr = orig_stderr
+
+
+def just_fix_windows_console():
+    global fixed_windows_console
+
+    if sys.platform != "win32":
+        return
+    if fixed_windows_console:
+        return
+    if wrapped_stdout is not None or wrapped_stderr is not None:
+        # Someone already ran init() and it did stuff, so we won't second-guess them
+        return
+
+    # On newer versions of Windows, AnsiToWin32.__init__ will implicitly enable the
+    # native ANSI support in the console as a side-effect. We only need to actually
+    # replace sys.stdout/stderr if we're in the old-style conversion mode.
+    new_stdout = AnsiToWin32(sys.stdout, convert=None, strip=None, autoreset=False)
+    if new_stdout.convert:
+        sys.stdout = new_stdout
+    new_stderr = AnsiToWin32(sys.stderr, convert=None, strip=None, autoreset=False)
+    if new_stderr.convert:
+        sys.stderr = new_stderr
+
+    fixed_windows_console = True
+
+@contextlib.contextmanager
+def colorama_text(*args, **kwargs):
+    init(*args, **kwargs)
+    try:
+        yield
+    finally:
+        deinit()
+
+
+def reinit():
+    if wrapped_stdout is not None:
+        sys.stdout = wrapped_stdout
+    if wrapped_stderr is not None:
+        sys.stderr = wrapped_stderr
+
+
+def wrap_stream(stream, convert, strip, autoreset, wrap):
+    if wrap:
+        wrapper = AnsiToWin32(stream,
+            convert=convert, strip=strip, autoreset=autoreset)
+        if wrapper.should_wrap():
+            stream = wrapper.stream
+    return stream
+
+
+# Use this for initial setup as well, to reduce code duplication
+_wipe_internal_state_for_tests()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__init__.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/__init__.py
new file mode 100644
index 00000000..8c5661e9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/__init__.py
@@ -0,0 +1 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6bbc56fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansi_test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansi_test.cpython-312.pyc
new file mode 100644
index 00000000..795ce5d0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansi_test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansitowin32_test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansitowin32_test.cpython-312.pyc
new file mode 100644
index 00000000..4c0aa89c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/ansitowin32_test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/initialise_test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/initialise_test.cpython-312.pyc
new file mode 100644
index 00000000..9bd6e7e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/initialise_test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/isatty_test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/isatty_test.cpython-312.pyc
new file mode 100644
index 00000000..84e8a4aa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/isatty_test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..6fc0b8c1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/winterm_test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/winterm_test.cpython-312.pyc
new file mode 100644
index 00000000..c58ba246
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/colorama/tests/__pycache__/winterm_test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/ansi_test.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/ansi_test.py
new file mode 100644
index 00000000..0a20c80f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/ansi_test.py
@@ -0,0 +1,76 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import sys
+from unittest import TestCase, main
+
+from ..ansi import Back, Fore, Style
+from ..ansitowin32 import AnsiToWin32
+
+stdout_orig = sys.stdout
+stderr_orig = sys.stderr
+
+
+class AnsiTest(TestCase):
+
+    def setUp(self):
+        # sanity check: stdout should be a file or StringIO object.
+        # It will only be AnsiToWin32 if init() has previously wrapped it
+        self.assertNotEqual(type(sys.stdout), AnsiToWin32)
+        self.assertNotEqual(type(sys.stderr), AnsiToWin32)
+
+    def tearDown(self):
+        sys.stdout = stdout_orig
+        sys.stderr = stderr_orig
+
+
+    def testForeAttributes(self):
+        self.assertEqual(Fore.BLACK, '\033[30m')
+        self.assertEqual(Fore.RED, '\033[31m')
+        self.assertEqual(Fore.GREEN, '\033[32m')
+        self.assertEqual(Fore.YELLOW, '\033[33m')
+        self.assertEqual(Fore.BLUE, '\033[34m')
+        self.assertEqual(Fore.MAGENTA, '\033[35m')
+        self.assertEqual(Fore.CYAN, '\033[36m')
+        self.assertEqual(Fore.WHITE, '\033[37m')
+        self.assertEqual(Fore.RESET, '\033[39m')
+
+        # Check the light, extended versions.
+        self.assertEqual(Fore.LIGHTBLACK_EX, '\033[90m')
+        self.assertEqual(Fore.LIGHTRED_EX, '\033[91m')
+        self.assertEqual(Fore.LIGHTGREEN_EX, '\033[92m')
+        self.assertEqual(Fore.LIGHTYELLOW_EX, '\033[93m')
+        self.assertEqual(Fore.LIGHTBLUE_EX, '\033[94m')
+        self.assertEqual(Fore.LIGHTMAGENTA_EX, '\033[95m')
+        self.assertEqual(Fore.LIGHTCYAN_EX, '\033[96m')
+        self.assertEqual(Fore.LIGHTWHITE_EX, '\033[97m')
+
+
+    def testBackAttributes(self):
+        self.assertEqual(Back.BLACK, '\033[40m')
+        self.assertEqual(Back.RED, '\033[41m')
+        self.assertEqual(Back.GREEN, '\033[42m')
+        self.assertEqual(Back.YELLOW, '\033[43m')
+        self.assertEqual(Back.BLUE, '\033[44m')
+        self.assertEqual(Back.MAGENTA, '\033[45m')
+        self.assertEqual(Back.CYAN, '\033[46m')
+        self.assertEqual(Back.WHITE, '\033[47m')
+        self.assertEqual(Back.RESET, '\033[49m')
+
+        # Check the light, extended versions.
+        self.assertEqual(Back.LIGHTBLACK_EX, '\033[100m')
+        self.assertEqual(Back.LIGHTRED_EX, '\033[101m')
+        self.assertEqual(Back.LIGHTGREEN_EX, '\033[102m')
+        self.assertEqual(Back.LIGHTYELLOW_EX, '\033[103m')
+        self.assertEqual(Back.LIGHTBLUE_EX, '\033[104m')
+        self.assertEqual(Back.LIGHTMAGENTA_EX, '\033[105m')
+        self.assertEqual(Back.LIGHTCYAN_EX, '\033[106m')
+        self.assertEqual(Back.LIGHTWHITE_EX, '\033[107m')
+
+
+    def testStyleAttributes(self):
+        self.assertEqual(Style.DIM, '\033[2m')
+        self.assertEqual(Style.NORMAL, '\033[22m')
+        self.assertEqual(Style.BRIGHT, '\033[1m')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/ansitowin32_test.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/ansitowin32_test.py
new file mode 100644
index 00000000..91ca551f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/ansitowin32_test.py
@@ -0,0 +1,294 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+from io import StringIO, TextIOWrapper
+from unittest import TestCase, main
+try:
+    from contextlib import ExitStack
+except ImportError:
+    # python 2
+    from contextlib2 import ExitStack
+
+try:
+    from unittest.mock import MagicMock, Mock, patch
+except ImportError:
+    from mock import MagicMock, Mock, patch
+
+from ..ansitowin32 import AnsiToWin32, StreamWrapper
+from ..win32 import ENABLE_VIRTUAL_TERMINAL_PROCESSING
+from .utils import osname
+
+
+class StreamWrapperTest(TestCase):
+
+    def testIsAProxy(self):
+        mockStream = Mock()
+        wrapper = StreamWrapper(mockStream, None)
+        self.assertTrue( wrapper.random_attr is mockStream.random_attr )
+
+    def testDelegatesWrite(self):
+        mockStream = Mock()
+        mockConverter = Mock()
+        wrapper = StreamWrapper(mockStream, mockConverter)
+        wrapper.write('hello')
+        self.assertTrue(mockConverter.write.call_args, (('hello',), {}))
+
+    def testDelegatesContext(self):
+        mockConverter = Mock()
+        s = StringIO()
+        with StreamWrapper(s, mockConverter) as fp:
+            fp.write(u'hello')
+        self.assertTrue(s.closed)
+
+    def testProxyNoContextManager(self):
+        mockStream = MagicMock()
+        mockStream.__enter__.side_effect = AttributeError()
+        mockConverter = Mock()
+        with self.assertRaises(AttributeError) as excinfo:
+            with StreamWrapper(mockStream, mockConverter) as wrapper:
+                wrapper.write('hello')
+
+    def test_closed_shouldnt_raise_on_closed_stream(self):
+        stream = StringIO()
+        stream.close()
+        wrapper = StreamWrapper(stream, None)
+        self.assertEqual(wrapper.closed, True)
+
+    def test_closed_shouldnt_raise_on_detached_stream(self):
+        stream = TextIOWrapper(StringIO())
+        stream.detach()
+        wrapper = StreamWrapper(stream, None)
+        self.assertEqual(wrapper.closed, True)
+
+class AnsiToWin32Test(TestCase):
+
+    def testInit(self):
+        mockStdout = Mock()
+        auto = Mock()
+        stream = AnsiToWin32(mockStdout, autoreset=auto)
+        self.assertEqual(stream.wrapped, mockStdout)
+        self.assertEqual(stream.autoreset, auto)
+
+    @patch('colorama.ansitowin32.winterm', None)
+    @patch('colorama.ansitowin32.winapi_test', lambda *_: True)
+    def testStripIsTrueOnWindows(self):
+        with osname('nt'):
+            mockStdout = Mock()
+            stream = AnsiToWin32(mockStdout)
+            self.assertTrue(stream.strip)
+
+    def testStripIsFalseOffWindows(self):
+        with osname('posix'):
+            mockStdout = Mock(closed=False)
+            stream = AnsiToWin32(mockStdout)
+            self.assertFalse(stream.strip)
+
+    def testWriteStripsAnsi(self):
+        mockStdout = Mock()
+        stream = AnsiToWin32(mockStdout)
+        stream.wrapped = Mock()
+        stream.write_and_convert = Mock()
+        stream.strip = True
+
+        stream.write('abc')
+
+        self.assertFalse(stream.wrapped.write.called)
+        self.assertEqual(stream.write_and_convert.call_args, (('abc',), {}))
+
+    def testWriteDoesNotStripAnsi(self):
+        mockStdout = Mock()
+        stream = AnsiToWin32(mockStdout)
+        stream.wrapped = Mock()
+        stream.write_and_convert = Mock()
+        stream.strip = False
+        stream.convert = False
+
+        stream.write('abc')
+
+        self.assertFalse(stream.write_and_convert.called)
+        self.assertEqual(stream.wrapped.write.call_args, (('abc',), {}))
+
+    def assert_autoresets(self, convert, autoreset=True):
+        stream = AnsiToWin32(Mock())
+        stream.convert = convert
+        stream.reset_all = Mock()
+        stream.autoreset = autoreset
+        stream.winterm = Mock()
+
+        stream.write('abc')
+
+        self.assertEqual(stream.reset_all.called, autoreset)
+
+    def testWriteAutoresets(self):
+        self.assert_autoresets(convert=True)
+        self.assert_autoresets(convert=False)
+        self.assert_autoresets(convert=True, autoreset=False)
+        self.assert_autoresets(convert=False, autoreset=False)
+
+    def testWriteAndConvertWritesPlainText(self):
+        stream = AnsiToWin32(Mock())
+        stream.write_and_convert( 'abc' )
+        self.assertEqual( stream.wrapped.write.call_args, (('abc',), {}) )
+
+    def testWriteAndConvertStripsAllValidAnsi(self):
+        stream = AnsiToWin32(Mock())
+        stream.call_win32 = Mock()
+        data = [
+            'abc\033[mdef',
+            'abc\033[0mdef',
+            'abc\033[2mdef',
+            'abc\033[02mdef',
+            'abc\033[002mdef',
+            'abc\033[40mdef',
+            'abc\033[040mdef',
+            'abc\033[0;1mdef',
+            'abc\033[40;50mdef',
+            'abc\033[50;30;40mdef',
+            'abc\033[Adef',
+            'abc\033[0Gdef',
+            'abc\033[1;20;128Hdef',
+        ]
+        for datum in data:
+            stream.wrapped.write.reset_mock()
+            stream.write_and_convert( datum )
+            self.assertEqual(
+               [args[0] for args in stream.wrapped.write.call_args_list],
+               [ ('abc',), ('def',) ]
+            )
+
+    def testWriteAndConvertSkipsEmptySnippets(self):
+        stream = AnsiToWin32(Mock())
+        stream.call_win32 = Mock()
+        stream.write_and_convert( '\033[40m\033[41m' )
+        self.assertFalse( stream.wrapped.write.called )
+
+    def testWriteAndConvertCallsWin32WithParamsAndCommand(self):
+        stream = AnsiToWin32(Mock())
+        stream.convert = True
+        stream.call_win32 = Mock()
+        stream.extract_params = Mock(return_value='params')
+        data = {
+            'abc\033[adef':         ('a', 'params'),
+            'abc\033[;;bdef':       ('b', 'params'),
+            'abc\033[0cdef':        ('c', 'params'),
+            'abc\033[;;0;;Gdef':    ('G', 'params'),
+            'abc\033[1;20;128Hdef': ('H', 'params'),
+        }
+        for datum, expected in data.items():
+            stream.call_win32.reset_mock()
+            stream.write_and_convert( datum )
+            self.assertEqual( stream.call_win32.call_args[0], expected )
+
+    def test_reset_all_shouldnt_raise_on_closed_orig_stdout(self):
+        stream = StringIO()
+        converter = AnsiToWin32(stream)
+        stream.close()
+
+        converter.reset_all()
+
+    def test_wrap_shouldnt_raise_on_closed_orig_stdout(self):
+        stream = StringIO()
+        stream.close()
+        with \
+            patch("colorama.ansitowin32.os.name", "nt"), \
+            patch("colorama.ansitowin32.winapi_test", lambda: True):
+                converter = AnsiToWin32(stream)
+        self.assertTrue(converter.strip)
+        self.assertFalse(converter.convert)
+
+    def test_wrap_shouldnt_raise_on_missing_closed_attr(self):
+        with \
+            patch("colorama.ansitowin32.os.name", "nt"), \
+            patch("colorama.ansitowin32.winapi_test", lambda: True):
+                converter = AnsiToWin32(object())
+        self.assertTrue(converter.strip)
+        self.assertFalse(converter.convert)
+
+    def testExtractParams(self):
+        stream = AnsiToWin32(Mock())
+        data = {
+            '':               (0,),
+            ';;':             (0,),
+            '2':              (2,),
+            ';;002;;':        (2,),
+            '0;1':            (0, 1),
+            ';;003;;456;;':   (3, 456),
+            '11;22;33;44;55': (11, 22, 33, 44, 55),
+        }
+        for datum, expected in data.items():
+            self.assertEqual(stream.extract_params('m', datum), expected)
+
+    def testCallWin32UsesLookup(self):
+        listener = Mock()
+        stream = AnsiToWin32(listener)
+        stream.win32_calls = {
+            1: (lambda *_, **__: listener(11),),
+            2: (lambda *_, **__: listener(22),),
+            3: (lambda *_, **__: listener(33),),
+        }
+        stream.call_win32('m', (3, 1, 99, 2))
+        self.assertEqual(
+            [a[0][0] for a in listener.call_args_list],
+            [33, 11, 22] )
+
+    def test_osc_codes(self):
+        mockStdout = Mock()
+        stream = AnsiToWin32(mockStdout, convert=True)
+        with patch('colorama.ansitowin32.winterm') as winterm:
+            data = [
+                '\033]0\x07',                      # missing arguments
+                '\033]0;foo\x08',                  # wrong OSC command
+                '\033]0;colorama_test_title\x07',  # should work
+                '\033]1;colorama_test_title\x07',  # wrong set command
+                '\033]2;colorama_test_title\x07',  # should work
+                '\033]' + ';' * 64 + '\x08',       # see issue #247
+            ]
+            for code in data:
+                stream.write(code)
+            self.assertEqual(winterm.set_title.call_count, 2)
+
+    def test_native_windows_ansi(self):
+        with ExitStack() as stack:
+            def p(a, b):
+                stack.enter_context(patch(a, b, create=True))
+            # Pretend to be on Windows
+            p("colorama.ansitowin32.os.name", "nt")
+            p("colorama.ansitowin32.winapi_test", lambda: True)
+            p("colorama.win32.winapi_test", lambda: True)
+            p("colorama.winterm.win32.windll", "non-None")
+            p("colorama.winterm.get_osfhandle", lambda _: 1234)
+
+            # Pretend that our mock stream has native ANSI support
+            p(
+                "colorama.winterm.win32.GetConsoleMode",
+                lambda _: ENABLE_VIRTUAL_TERMINAL_PROCESSING,
+            )
+            SetConsoleMode = Mock()
+            p("colorama.winterm.win32.SetConsoleMode", SetConsoleMode)
+
+            stdout = Mock()
+            stdout.closed = False
+            stdout.isatty.return_value = True
+            stdout.fileno.return_value = 1
+
+            # Our fake console says it has native vt support, so AnsiToWin32 should
+            # enable that support and do nothing else.
+            stream = AnsiToWin32(stdout)
+            SetConsoleMode.assert_called_with(1234, ENABLE_VIRTUAL_TERMINAL_PROCESSING)
+            self.assertFalse(stream.strip)
+            self.assertFalse(stream.convert)
+            self.assertFalse(stream.should_wrap())
+
+            # Now let's pretend we're on an old Windows console, that doesn't have
+            # native ANSI support.
+            p("colorama.winterm.win32.GetConsoleMode", lambda _: 0)
+            SetConsoleMode = Mock()
+            p("colorama.winterm.win32.SetConsoleMode", SetConsoleMode)
+
+            stream = AnsiToWin32(stdout)
+            SetConsoleMode.assert_called_with(1234, ENABLE_VIRTUAL_TERMINAL_PROCESSING)
+            self.assertTrue(stream.strip)
+            self.assertTrue(stream.convert)
+            self.assertTrue(stream.should_wrap())
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/initialise_test.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/initialise_test.py
new file mode 100644
index 00000000..89f9b075
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/initialise_test.py
@@ -0,0 +1,189 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import sys
+from unittest import TestCase, main, skipUnless
+
+try:
+    from unittest.mock import patch, Mock
+except ImportError:
+    from mock import patch, Mock
+
+from ..ansitowin32 import StreamWrapper
+from ..initialise import init, just_fix_windows_console, _wipe_internal_state_for_tests
+from .utils import osname, replace_by
+
+orig_stdout = sys.stdout
+orig_stderr = sys.stderr
+
+
+class InitTest(TestCase):
+
+    @skipUnless(sys.stdout.isatty(), "sys.stdout is not a tty")
+    def setUp(self):
+        # sanity check
+        self.assertNotWrapped()
+
+    def tearDown(self):
+        _wipe_internal_state_for_tests()
+        sys.stdout = orig_stdout
+        sys.stderr = orig_stderr
+
+    def assertWrapped(self):
+        self.assertIsNot(sys.stdout, orig_stdout, 'stdout should be wrapped')
+        self.assertIsNot(sys.stderr, orig_stderr, 'stderr should be wrapped')
+        self.assertTrue(isinstance(sys.stdout, StreamWrapper),
+            'bad stdout wrapper')
+        self.assertTrue(isinstance(sys.stderr, StreamWrapper),
+            'bad stderr wrapper')
+
+    def assertNotWrapped(self):
+        self.assertIs(sys.stdout, orig_stdout, 'stdout should not be wrapped')
+        self.assertIs(sys.stderr, orig_stderr, 'stderr should not be wrapped')
+
+    @patch('colorama.initialise.reset_all')
+    @patch('colorama.ansitowin32.winapi_test', lambda *_: True)
+    @patch('colorama.ansitowin32.enable_vt_processing', lambda *_: False)
+    def testInitWrapsOnWindows(self, _):
+        with osname("nt"):
+            init()
+            self.assertWrapped()
+
+    @patch('colorama.initialise.reset_all')
+    @patch('colorama.ansitowin32.winapi_test', lambda *_: False)
+    def testInitDoesntWrapOnEmulatedWindows(self, _):
+        with osname("nt"):
+            init()
+            self.assertNotWrapped()
+
+    def testInitDoesntWrapOnNonWindows(self):
+        with osname("posix"):
+            init()
+            self.assertNotWrapped()
+
+    def testInitDoesntWrapIfNone(self):
+        with replace_by(None):
+            init()
+            # We can't use assertNotWrapped here because replace_by(None)
+            # changes stdout/stderr already.
+            self.assertIsNone(sys.stdout)
+            self.assertIsNone(sys.stderr)
+
+    def testInitAutoresetOnWrapsOnAllPlatforms(self):
+        with osname("posix"):
+            init(autoreset=True)
+            self.assertWrapped()
+
+    def testInitWrapOffDoesntWrapOnWindows(self):
+        with osname("nt"):
+            init(wrap=False)
+            self.assertNotWrapped()
+
+    def testInitWrapOffIncompatibleWithAutoresetOn(self):
+        self.assertRaises(ValueError, lambda: init(autoreset=True, wrap=False))
+
+    @patch('colorama.win32.SetConsoleTextAttribute')
+    @patch('colorama.initialise.AnsiToWin32')
+    def testAutoResetPassedOn(self, mockATW32, _):
+        with osname("nt"):
+            init(autoreset=True)
+            self.assertEqual(len(mockATW32.call_args_list), 2)
+            self.assertEqual(mockATW32.call_args_list[1][1]['autoreset'], True)
+            self.assertEqual(mockATW32.call_args_list[0][1]['autoreset'], True)
+
+    @patch('colorama.initialise.AnsiToWin32')
+    def testAutoResetChangeable(self, mockATW32):
+        with osname("nt"):
+            init()
+
+            init(autoreset=True)
+            self.assertEqual(len(mockATW32.call_args_list), 4)
+            self.assertEqual(mockATW32.call_args_list[2][1]['autoreset'], True)
+            self.assertEqual(mockATW32.call_args_list[3][1]['autoreset'], True)
+
+            init()
+            self.assertEqual(len(mockATW32.call_args_list), 6)
+            self.assertEqual(
+                mockATW32.call_args_list[4][1]['autoreset'], False)
+            self.assertEqual(
+                mockATW32.call_args_list[5][1]['autoreset'], False)
+
+
+    @patch('colorama.initialise.atexit.register')
+    def testAtexitRegisteredOnlyOnce(self, mockRegister):
+        init()
+        self.assertTrue(mockRegister.called)
+        mockRegister.reset_mock()
+        init()
+        self.assertFalse(mockRegister.called)
+
+
+class JustFixWindowsConsoleTest(TestCase):
+    def _reset(self):
+        _wipe_internal_state_for_tests()
+        sys.stdout = orig_stdout
+        sys.stderr = orig_stderr
+
+    def tearDown(self):
+        self._reset()
+
+    @patch("colorama.ansitowin32.winapi_test", lambda: True)
+    def testJustFixWindowsConsole(self):
+        if sys.platform != "win32":
+            # just_fix_windows_console should be a no-op
+            just_fix_windows_console()
+            self.assertIs(sys.stdout, orig_stdout)
+            self.assertIs(sys.stderr, orig_stderr)
+        else:
+            def fake_std():
+                # Emulate stdout=not a tty, stderr=tty
+                # to check that we handle both cases correctly
+                stdout = Mock()
+                stdout.closed = False
+                stdout.isatty.return_value = False
+                stdout.fileno.return_value = 1
+                sys.stdout = stdout
+
+                stderr = Mock()
+                stderr.closed = False
+                stderr.isatty.return_value = True
+                stderr.fileno.return_value = 2
+                sys.stderr = stderr
+
+            for native_ansi in [False, True]:
+                with patch(
+                    'colorama.ansitowin32.enable_vt_processing',
+                    lambda *_: native_ansi
+                ):
+                    self._reset()
+                    fake_std()
+
+                    # Regular single-call test
+                    prev_stdout = sys.stdout
+                    prev_stderr = sys.stderr
+                    just_fix_windows_console()
+                    self.assertIs(sys.stdout, prev_stdout)
+                    if native_ansi:
+                        self.assertIs(sys.stderr, prev_stderr)
+                    else:
+                        self.assertIsNot(sys.stderr, prev_stderr)
+
+                    # second call without resetting is always a no-op
+                    prev_stdout = sys.stdout
+                    prev_stderr = sys.stderr
+                    just_fix_windows_console()
+                    self.assertIs(sys.stdout, prev_stdout)
+                    self.assertIs(sys.stderr, prev_stderr)
+
+                    self._reset()
+                    fake_std()
+
+                    # If init() runs first, just_fix_windows_console should be a no-op
+                    init()
+                    prev_stdout = sys.stdout
+                    prev_stderr = sys.stderr
+                    just_fix_windows_console()
+                    self.assertIs(prev_stdout, sys.stdout)
+                    self.assertIs(prev_stderr, sys.stderr)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/isatty_test.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/isatty_test.py
new file mode 100644
index 00000000..0f84e4be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/isatty_test.py
@@ -0,0 +1,57 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import sys
+from unittest import TestCase, main
+
+from ..ansitowin32 import StreamWrapper, AnsiToWin32
+from .utils import pycharm, replace_by, replace_original_by, StreamTTY, StreamNonTTY
+
+
+def is_a_tty(stream):
+    return StreamWrapper(stream, None).isatty()
+
+class IsattyTest(TestCase):
+
+    def test_TTY(self):
+        tty = StreamTTY()
+        self.assertTrue(is_a_tty(tty))
+        with pycharm():
+            self.assertTrue(is_a_tty(tty))
+
+    def test_nonTTY(self):
+        non_tty = StreamNonTTY()
+        self.assertFalse(is_a_tty(non_tty))
+        with pycharm():
+            self.assertFalse(is_a_tty(non_tty))
+
+    def test_withPycharm(self):
+        with pycharm():
+            self.assertTrue(is_a_tty(sys.stderr))
+            self.assertTrue(is_a_tty(sys.stdout))
+
+    def test_withPycharmTTYOverride(self):
+        tty = StreamTTY()
+        with pycharm(), replace_by(tty):
+            self.assertTrue(is_a_tty(tty))
+
+    def test_withPycharmNonTTYOverride(self):
+        non_tty = StreamNonTTY()
+        with pycharm(), replace_by(non_tty):
+            self.assertFalse(is_a_tty(non_tty))
+
+    def test_withPycharmNoneOverride(self):
+        with pycharm():
+            with replace_by(None), replace_original_by(None):
+                self.assertFalse(is_a_tty(None))
+                self.assertFalse(is_a_tty(StreamNonTTY()))
+                self.assertTrue(is_a_tty(StreamTTY()))
+
+    def test_withPycharmStreamWrapped(self):
+        with pycharm():
+            self.assertTrue(AnsiToWin32(StreamTTY()).stream.isatty())
+            self.assertFalse(AnsiToWin32(StreamNonTTY()).stream.isatty())
+            self.assertTrue(AnsiToWin32(sys.stdout).stream.isatty())
+            self.assertTrue(AnsiToWin32(sys.stderr).stream.isatty())
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/utils.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/utils.py
new file mode 100644
index 00000000..472fafb4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/utils.py
@@ -0,0 +1,49 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+from contextlib import contextmanager
+from io import StringIO
+import sys
+import os
+
+
+class StreamTTY(StringIO):
+    def isatty(self):
+        return True
+
+class StreamNonTTY(StringIO):
+    def isatty(self):
+        return False
+
+@contextmanager
+def osname(name):
+    orig = os.name
+    os.name = name
+    yield
+    os.name = orig
+
+@contextmanager
+def replace_by(stream):
+    orig_stdout = sys.stdout
+    orig_stderr = sys.stderr
+    sys.stdout = stream
+    sys.stderr = stream
+    yield
+    sys.stdout = orig_stdout
+    sys.stderr = orig_stderr
+
+@contextmanager
+def replace_original_by(stream):
+    orig_stdout = sys.__stdout__
+    orig_stderr = sys.__stderr__
+    sys.__stdout__ = stream
+    sys.__stderr__ = stream
+    yield
+    sys.__stdout__ = orig_stdout
+    sys.__stderr__ = orig_stderr
+
+@contextmanager
+def pycharm():
+    os.environ["PYCHARM_HOSTED"] = "1"
+    non_tty = StreamNonTTY()
+    with replace_by(non_tty), replace_original_by(non_tty):
+        yield
+    del os.environ["PYCHARM_HOSTED"]
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/tests/winterm_test.py b/backend/venv/lib/python3.12/site-packages/colorama/tests/winterm_test.py
new file mode 100644
index 00000000..d0955f9e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/tests/winterm_test.py
@@ -0,0 +1,131 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+import sys
+from unittest import TestCase, main, skipUnless
+
+try:
+    from unittest.mock import Mock, patch
+except ImportError:
+    from mock import Mock, patch
+
+from ..winterm import WinColor, WinStyle, WinTerm
+
+
+class WinTermTest(TestCase):
+
+    @patch('colorama.winterm.win32')
+    def testInit(self, mockWin32):
+        mockAttr = Mock()
+        mockAttr.wAttributes = 7 + 6 * 16 + 8
+        mockWin32.GetConsoleScreenBufferInfo.return_value = mockAttr
+        term = WinTerm()
+        self.assertEqual(term._fore, 7)
+        self.assertEqual(term._back, 6)
+        self.assertEqual(term._style, 8)
+
+    @skipUnless(sys.platform.startswith("win"), "requires Windows")
+    def testGetAttrs(self):
+        term = WinTerm()
+
+        term._fore = 0
+        term._back = 0
+        term._style = 0
+        self.assertEqual(term.get_attrs(), 0)
+
+        term._fore = WinColor.YELLOW
+        self.assertEqual(term.get_attrs(), WinColor.YELLOW)
+
+        term._back = WinColor.MAGENTA
+        self.assertEqual(
+            term.get_attrs(),
+            WinColor.YELLOW + WinColor.MAGENTA * 16)
+
+        term._style = WinStyle.BRIGHT
+        self.assertEqual(
+            term.get_attrs(),
+            WinColor.YELLOW + WinColor.MAGENTA * 16 + WinStyle.BRIGHT)
+
+    @patch('colorama.winterm.win32')
+    def testResetAll(self, mockWin32):
+        mockAttr = Mock()
+        mockAttr.wAttributes = 1 + 2 * 16 + 8
+        mockWin32.GetConsoleScreenBufferInfo.return_value = mockAttr
+        term = WinTerm()
+
+        term.set_console = Mock()
+        term._fore = -1
+        term._back = -1
+        term._style = -1
+
+        term.reset_all()
+
+        self.assertEqual(term._fore, 1)
+        self.assertEqual(term._back, 2)
+        self.assertEqual(term._style, 8)
+        self.assertEqual(term.set_console.called, True)
+
+    @skipUnless(sys.platform.startswith("win"), "requires Windows")
+    def testFore(self):
+        term = WinTerm()
+        term.set_console = Mock()
+        term._fore = 0
+
+        term.fore(5)
+
+        self.assertEqual(term._fore, 5)
+        self.assertEqual(term.set_console.called, True)
+
+    @skipUnless(sys.platform.startswith("win"), "requires Windows")
+    def testBack(self):
+        term = WinTerm()
+        term.set_console = Mock()
+        term._back = 0
+
+        term.back(5)
+
+        self.assertEqual(term._back, 5)
+        self.assertEqual(term.set_console.called, True)
+
+    @skipUnless(sys.platform.startswith("win"), "requires Windows")
+    def testStyle(self):
+        term = WinTerm()
+        term.set_console = Mock()
+        term._style = 0
+
+        term.style(22)
+
+        self.assertEqual(term._style, 22)
+        self.assertEqual(term.set_console.called, True)
+
+    @patch('colorama.winterm.win32')
+    def testSetConsole(self, mockWin32):
+        mockAttr = Mock()
+        mockAttr.wAttributes = 0
+        mockWin32.GetConsoleScreenBufferInfo.return_value = mockAttr
+        term = WinTerm()
+        term.windll = Mock()
+
+        term.set_console()
+
+        self.assertEqual(
+            mockWin32.SetConsoleTextAttribute.call_args,
+            ((mockWin32.STDOUT, term.get_attrs()), {})
+        )
+
+    @patch('colorama.winterm.win32')
+    def testSetConsoleOnStderr(self, mockWin32):
+        mockAttr = Mock()
+        mockAttr.wAttributes = 0
+        mockWin32.GetConsoleScreenBufferInfo.return_value = mockAttr
+        term = WinTerm()
+        term.windll = Mock()
+
+        term.set_console(on_stderr=True)
+
+        self.assertEqual(
+            mockWin32.SetConsoleTextAttribute.call_args,
+            ((mockWin32.STDERR, term.get_attrs()), {})
+        )
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/win32.py b/backend/venv/lib/python3.12/site-packages/colorama/win32.py
new file mode 100644
index 00000000..841b0e27
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/win32.py
@@ -0,0 +1,180 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+
+# from winbase.h
+STDOUT = -11
+STDERR = -12
+
+ENABLE_VIRTUAL_TERMINAL_PROCESSING = 0x0004
+
+try:
+    import ctypes
+    from ctypes import LibraryLoader
+    windll = LibraryLoader(ctypes.WinDLL)
+    from ctypes import wintypes
+except (AttributeError, ImportError):
+    windll = None
+    SetConsoleTextAttribute = lambda *_: None
+    winapi_test = lambda *_: None
+else:
+    from ctypes import byref, Structure, c_char, POINTER
+
+    COORD = wintypes._COORD
+
+    class CONSOLE_SCREEN_BUFFER_INFO(Structure):
+        """struct in wincon.h."""
+        _fields_ = [
+            ("dwSize", COORD),
+            ("dwCursorPosition", COORD),
+            ("wAttributes", wintypes.WORD),
+            ("srWindow", wintypes.SMALL_RECT),
+            ("dwMaximumWindowSize", COORD),
+        ]
+        def __str__(self):
+            return '(%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)' % (
+                self.dwSize.Y, self.dwSize.X
+                , self.dwCursorPosition.Y, self.dwCursorPosition.X
+                , self.wAttributes
+                , self.srWindow.Top, self.srWindow.Left, self.srWindow.Bottom, self.srWindow.Right
+                , self.dwMaximumWindowSize.Y, self.dwMaximumWindowSize.X
+            )
+
+    _GetStdHandle = windll.kernel32.GetStdHandle
+    _GetStdHandle.argtypes = [
+        wintypes.DWORD,
+    ]
+    _GetStdHandle.restype = wintypes.HANDLE
+
+    _GetConsoleScreenBufferInfo = windll.kernel32.GetConsoleScreenBufferInfo
+    _GetConsoleScreenBufferInfo.argtypes = [
+        wintypes.HANDLE,
+        POINTER(CONSOLE_SCREEN_BUFFER_INFO),
+    ]
+    _GetConsoleScreenBufferInfo.restype = wintypes.BOOL
+
+    _SetConsoleTextAttribute = windll.kernel32.SetConsoleTextAttribute
+    _SetConsoleTextAttribute.argtypes = [
+        wintypes.HANDLE,
+        wintypes.WORD,
+    ]
+    _SetConsoleTextAttribute.restype = wintypes.BOOL
+
+    _SetConsoleCursorPosition = windll.kernel32.SetConsoleCursorPosition
+    _SetConsoleCursorPosition.argtypes = [
+        wintypes.HANDLE,
+        COORD,
+    ]
+    _SetConsoleCursorPosition.restype = wintypes.BOOL
+
+    _FillConsoleOutputCharacterA = windll.kernel32.FillConsoleOutputCharacterA
+    _FillConsoleOutputCharacterA.argtypes = [
+        wintypes.HANDLE,
+        c_char,
+        wintypes.DWORD,
+        COORD,
+        POINTER(wintypes.DWORD),
+    ]
+    _FillConsoleOutputCharacterA.restype = wintypes.BOOL
+
+    _FillConsoleOutputAttribute = windll.kernel32.FillConsoleOutputAttribute
+    _FillConsoleOutputAttribute.argtypes = [
+        wintypes.HANDLE,
+        wintypes.WORD,
+        wintypes.DWORD,
+        COORD,
+        POINTER(wintypes.DWORD),
+    ]
+    _FillConsoleOutputAttribute.restype = wintypes.BOOL
+
+    _SetConsoleTitleW = windll.kernel32.SetConsoleTitleW
+    _SetConsoleTitleW.argtypes = [
+        wintypes.LPCWSTR
+    ]
+    _SetConsoleTitleW.restype = wintypes.BOOL
+
+    _GetConsoleMode = windll.kernel32.GetConsoleMode
+    _GetConsoleMode.argtypes = [
+        wintypes.HANDLE,
+        POINTER(wintypes.DWORD)
+    ]
+    _GetConsoleMode.restype = wintypes.BOOL
+
+    _SetConsoleMode = windll.kernel32.SetConsoleMode
+    _SetConsoleMode.argtypes = [
+        wintypes.HANDLE,
+        wintypes.DWORD
+    ]
+    _SetConsoleMode.restype = wintypes.BOOL
+
+    def _winapi_test(handle):
+        csbi = CONSOLE_SCREEN_BUFFER_INFO()
+        success = _GetConsoleScreenBufferInfo(
+            handle, byref(csbi))
+        return bool(success)
+
+    def winapi_test():
+        return any(_winapi_test(h) for h in
+                   (_GetStdHandle(STDOUT), _GetStdHandle(STDERR)))
+
+    def GetConsoleScreenBufferInfo(stream_id=STDOUT):
+        handle = _GetStdHandle(stream_id)
+        csbi = CONSOLE_SCREEN_BUFFER_INFO()
+        success = _GetConsoleScreenBufferInfo(
+            handle, byref(csbi))
+        return csbi
+
+    def SetConsoleTextAttribute(stream_id, attrs):
+        handle = _GetStdHandle(stream_id)
+        return _SetConsoleTextAttribute(handle, attrs)
+
+    def SetConsoleCursorPosition(stream_id, position, adjust=True):
+        position = COORD(*position)
+        # If the position is out of range, do nothing.
+        if position.Y <= 0 or position.X <= 0:
+            return
+        # Adjust for Windows' SetConsoleCursorPosition:
+        #    1. being 0-based, while ANSI is 1-based.
+        #    2. expecting (x,y), while ANSI uses (y,x).
+        adjusted_position = COORD(position.Y - 1, position.X - 1)
+        if adjust:
+            # Adjust for viewport's scroll position
+            sr = GetConsoleScreenBufferInfo(STDOUT).srWindow
+            adjusted_position.Y += sr.Top
+            adjusted_position.X += sr.Left
+        # Resume normal processing
+        handle = _GetStdHandle(stream_id)
+        return _SetConsoleCursorPosition(handle, adjusted_position)
+
+    def FillConsoleOutputCharacter(stream_id, char, length, start):
+        handle = _GetStdHandle(stream_id)
+        char = c_char(char.encode())
+        length = wintypes.DWORD(length)
+        num_written = wintypes.DWORD(0)
+        # Note that this is hard-coded for ANSI (vs wide) bytes.
+        success = _FillConsoleOutputCharacterA(
+            handle, char, length, start, byref(num_written))
+        return num_written.value
+
+    def FillConsoleOutputAttribute(stream_id, attr, length, start):
+        ''' FillConsoleOutputAttribute( hConsole, csbi.wAttributes, dwConSize, coordScreen, &cCharsWritten )'''
+        handle = _GetStdHandle(stream_id)
+        attribute = wintypes.WORD(attr)
+        length = wintypes.DWORD(length)
+        num_written = wintypes.DWORD(0)
+        # Note that this is hard-coded for ANSI (vs wide) bytes.
+        return _FillConsoleOutputAttribute(
+            handle, attribute, length, start, byref(num_written))
+
+    def SetConsoleTitle(title):
+        return _SetConsoleTitleW(title)
+
+    def GetConsoleMode(handle):
+        mode = wintypes.DWORD()
+        success = _GetConsoleMode(handle, byref(mode))
+        if not success:
+            raise ctypes.WinError()
+        return mode.value
+
+    def SetConsoleMode(handle, mode):
+        success = _SetConsoleMode(handle, mode)
+        if not success:
+            raise ctypes.WinError()
diff --git a/backend/venv/lib/python3.12/site-packages/colorama/winterm.py b/backend/venv/lib/python3.12/site-packages/colorama/winterm.py
new file mode 100644
index 00000000..aad867e8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/colorama/winterm.py
@@ -0,0 +1,195 @@
+# Copyright Jonathan Hartley 2013. BSD 3-Clause license, see LICENSE file.
+try:
+    from msvcrt import get_osfhandle
+except ImportError:
+    def get_osfhandle(_):
+        raise OSError("This isn't windows!")
+
+
+from . import win32
+
+# from wincon.h
+class WinColor(object):
+    BLACK   = 0
+    BLUE    = 1
+    GREEN   = 2
+    CYAN    = 3
+    RED     = 4
+    MAGENTA = 5
+    YELLOW  = 6
+    GREY    = 7
+
+# from wincon.h
+class WinStyle(object):
+    NORMAL              = 0x00 # dim text, dim background
+    BRIGHT              = 0x08 # bright text, dim background
+    BRIGHT_BACKGROUND   = 0x80 # dim text, bright background
+
+class WinTerm(object):
+
+    def __init__(self):
+        self._default = win32.GetConsoleScreenBufferInfo(win32.STDOUT).wAttributes
+        self.set_attrs(self._default)
+        self._default_fore = self._fore
+        self._default_back = self._back
+        self._default_style = self._style
+        # In order to emulate LIGHT_EX in windows, we borrow the BRIGHT style.
+        # So that LIGHT_EX colors and BRIGHT style do not clobber each other,
+        # we track them separately, since LIGHT_EX is overwritten by Fore/Back
+        # and BRIGHT is overwritten by Style codes.
+        self._light = 0
+
+    def get_attrs(self):
+        return self._fore + self._back * 16 + (self._style | self._light)
+
+    def set_attrs(self, value):
+        self._fore = value & 7
+        self._back = (value >> 4) & 7
+        self._style = value & (WinStyle.BRIGHT | WinStyle.BRIGHT_BACKGROUND)
+
+    def reset_all(self, on_stderr=None):
+        self.set_attrs(self._default)
+        self.set_console(attrs=self._default)
+        self._light = 0
+
+    def fore(self, fore=None, light=False, on_stderr=False):
+        if fore is None:
+            fore = self._default_fore
+        self._fore = fore
+        # Emulate LIGHT_EX with BRIGHT Style
+        if light:
+            self._light |= WinStyle.BRIGHT
+        else:
+            self._light &= ~WinStyle.BRIGHT
+        self.set_console(on_stderr=on_stderr)
+
+    def back(self, back=None, light=False, on_stderr=False):
+        if back is None:
+            back = self._default_back
+        self._back = back
+        # Emulate LIGHT_EX with BRIGHT_BACKGROUND Style
+        if light:
+            self._light |= WinStyle.BRIGHT_BACKGROUND
+        else:
+            self._light &= ~WinStyle.BRIGHT_BACKGROUND
+        self.set_console(on_stderr=on_stderr)
+
+    def style(self, style=None, on_stderr=False):
+        if style is None:
+            style = self._default_style
+        self._style = style
+        self.set_console(on_stderr=on_stderr)
+
+    def set_console(self, attrs=None, on_stderr=False):
+        if attrs is None:
+            attrs = self.get_attrs()
+        handle = win32.STDOUT
+        if on_stderr:
+            handle = win32.STDERR
+        win32.SetConsoleTextAttribute(handle, attrs)
+
+    def get_position(self, handle):
+        position = win32.GetConsoleScreenBufferInfo(handle).dwCursorPosition
+        # Because Windows coordinates are 0-based,
+        # and win32.SetConsoleCursorPosition expects 1-based.
+        position.X += 1
+        position.Y += 1
+        return position
+
+    def set_cursor_position(self, position=None, on_stderr=False):
+        if position is None:
+            # I'm not currently tracking the position, so there is no default.
+            # position = self.get_position()
+            return
+        handle = win32.STDOUT
+        if on_stderr:
+            handle = win32.STDERR
+        win32.SetConsoleCursorPosition(handle, position)
+
+    def cursor_adjust(self, x, y, on_stderr=False):
+        handle = win32.STDOUT
+        if on_stderr:
+            handle = win32.STDERR
+        position = self.get_position(handle)
+        adjusted_position = (position.Y + y, position.X + x)
+        win32.SetConsoleCursorPosition(handle, adjusted_position, adjust=False)
+
+    def erase_screen(self, mode=0, on_stderr=False):
+        # 0 should clear from the cursor to the end of the screen.
+        # 1 should clear from the cursor to the beginning of the screen.
+        # 2 should clear the entire screen, and move cursor to (1,1)
+        handle = win32.STDOUT
+        if on_stderr:
+            handle = win32.STDERR
+        csbi = win32.GetConsoleScreenBufferInfo(handle)
+        # get the number of character cells in the current buffer
+        cells_in_screen = csbi.dwSize.X * csbi.dwSize.Y
+        # get number of character cells before current cursor position
+        cells_before_cursor = csbi.dwSize.X * csbi.dwCursorPosition.Y + csbi.dwCursorPosition.X
+        if mode == 0:
+            from_coord = csbi.dwCursorPosition
+            cells_to_erase = cells_in_screen - cells_before_cursor
+        elif mode == 1:
+            from_coord = win32.COORD(0, 0)
+            cells_to_erase = cells_before_cursor
+        elif mode == 2:
+            from_coord = win32.COORD(0, 0)
+            cells_to_erase = cells_in_screen
+        else:
+            # invalid mode
+            return
+        # fill the entire screen with blanks
+        win32.FillConsoleOutputCharacter(handle, ' ', cells_to_erase, from_coord)
+        # now set the buffer's attributes accordingly
+        win32.FillConsoleOutputAttribute(handle, self.get_attrs(), cells_to_erase, from_coord)
+        if mode == 2:
+            # put the cursor where needed
+            win32.SetConsoleCursorPosition(handle, (1, 1))
+
+    def erase_line(self, mode=0, on_stderr=False):
+        # 0 should clear from the cursor to the end of the line.
+        # 1 should clear from the cursor to the beginning of the line.
+        # 2 should clear the entire line.
+        handle = win32.STDOUT
+        if on_stderr:
+            handle = win32.STDERR
+        csbi = win32.GetConsoleScreenBufferInfo(handle)
+        if mode == 0:
+            from_coord = csbi.dwCursorPosition
+            cells_to_erase = csbi.dwSize.X - csbi.dwCursorPosition.X
+        elif mode == 1:
+            from_coord = win32.COORD(0, csbi.dwCursorPosition.Y)
+            cells_to_erase = csbi.dwCursorPosition.X
+        elif mode == 2:
+            from_coord = win32.COORD(0, csbi.dwCursorPosition.Y)
+            cells_to_erase = csbi.dwSize.X
+        else:
+            # invalid mode
+            return
+        # fill the entire screen with blanks
+        win32.FillConsoleOutputCharacter(handle, ' ', cells_to_erase, from_coord)
+        # now set the buffer's attributes accordingly
+        win32.FillConsoleOutputAttribute(handle, self.get_attrs(), cells_to_erase, from_coord)
+
+    def set_title(self, title):
+        win32.SetConsoleTitle(title)
+
+
+def enable_vt_processing(fd):
+    if win32.windll is None or not win32.winapi_test():
+        return False
+
+    try:
+        handle = get_osfhandle(fd)
+        mode = win32.GetConsoleMode(handle)
+        win32.SetConsoleMode(
+            handle,
+            mode | win32.ENABLE_VIRTUAL_TERMINAL_PROCESSING,
+        )
+
+        mode = win32.GetConsoleMode(handle)
+        if mode & win32.ENABLE_VIRTUAL_TERMINAL_PROCESSING:
+            return True
+    # Can get TypeError in testsuite where 'fd' is a Mock()
+    except (OSError, TypeError):
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/distutils-precedence.pth b/backend/venv/lib/python3.12/site-packages/distutils-precedence.pth
new file mode 100644
index 00000000..7f009fe9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/distutils-precedence.pth
@@ -0,0 +1 @@
+import os; var = 'SETUPTOOLS_USE_DISTUTILS'; enabled = os.environ.get(var, 'local') == 'local'; enabled and __import__('_distutils_hack').add_shim(); 
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/LICENSE.txt
new file mode 100644
index 00000000..9d227a0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2010 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/METADATA
new file mode 100644
index 00000000..c49ceb9a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/METADATA
@@ -0,0 +1,81 @@
+Metadata-Version: 2.3
+Name: Flask
+Version: 3.1.0
+Summary: A simple framework for building complex web applications.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Framework :: Flask
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Classifier: Topic :: Internet :: WWW/HTTP :: WSGI
+Classifier: Topic :: Internet :: WWW/HTTP :: WSGI :: Application
+Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
+Classifier: Typing :: Typed
+Requires-Dist: Werkzeug>=3.1
+Requires-Dist: Jinja2>=3.1.2
+Requires-Dist: itsdangerous>=2.2
+Requires-Dist: click>=8.1.3
+Requires-Dist: blinker>=1.9
+Requires-Dist: importlib-metadata>=3.6; python_version < '3.10'
+Requires-Dist: asgiref>=3.2 ; extra == "async"
+Requires-Dist: python-dotenv ; extra == "dotenv"
+Project-URL: Changes, https://flask.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://flask.palletsprojects.com/
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Source, https://github.com/pallets/flask/
+Provides-Extra: async
+Provides-Extra: dotenv
+
+# Flask
+
+Flask is a lightweight [WSGI][] web application framework. It is designed
+to make getting started quick and easy, with the ability to scale up to
+complex applications. It began as a simple wrapper around [Werkzeug][]
+and [Jinja][], and has become one of the most popular Python web
+application frameworks.
+
+Flask offers suggestions, but doesn't enforce any dependencies or
+project layout. It is up to the developer to choose the tools and
+libraries they want to use. There are many extensions provided by the
+community that make adding new functionality easy.
+
+[WSGI]: https://wsgi.readthedocs.io/
+[Werkzeug]: https://werkzeug.palletsprojects.com/
+[Jinja]: https://jinja.palletsprojects.com/
+
+
+## A Simple Example
+
+```python
+# save this as app.py
+from flask import Flask
+
+app = Flask(__name__)
+
+@app.route("/")
+def hello():
+    return "Hello, World!"
+```
+
+```
+$ flask run
+  * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
+```
+
+
+## Donate
+
+The Pallets organization develops and supports Flask and the libraries
+it uses. In order to grow the community of contributors and users, and
+allow the maintainers to devote more time to the projects, [please
+donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
+
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/RECORD
new file mode 100644
index 00000000..1cb5f154
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/RECORD
@@ -0,0 +1,58 @@
+../../../bin/flask,sha256=BadfjHFgD0HrMegH3WEXoAunEthNGWeowWNVEtOrhV8,265
+flask-3.1.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+flask-3.1.0.dist-info/LICENSE.txt,sha256=SJqOEQhQntmKN7uYPhHg9-HTHwvY-Zp5yESOf_N9B-o,1475
+flask-3.1.0.dist-info/METADATA,sha256=Wvf66xwUclGRu89cNcvErpaMf1rMgcxeqIkTc4EmD90,2718
+flask-3.1.0.dist-info/RECORD,,
+flask-3.1.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flask-3.1.0.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+flask-3.1.0.dist-info/entry_points.txt,sha256=bBP7hTOS5fz9zLtC7sPofBZAlMkEvBxu7KqS6l5lvc4,40
+flask/__init__.py,sha256=6xMqdVA0FIQ2U1KVaGX3lzNCdXPzoHPaa0hvQCNcfSk,2625
+flask/__main__.py,sha256=bYt9eEaoRQWdejEHFD8REx9jxVEdZptECFsV7F49Ink,30
+flask/__pycache__/__init__.cpython-312.pyc,,
+flask/__pycache__/__main__.cpython-312.pyc,,
+flask/__pycache__/app.cpython-312.pyc,,
+flask/__pycache__/blueprints.cpython-312.pyc,,
+flask/__pycache__/cli.cpython-312.pyc,,
+flask/__pycache__/config.cpython-312.pyc,,
+flask/__pycache__/ctx.cpython-312.pyc,,
+flask/__pycache__/debughelpers.cpython-312.pyc,,
+flask/__pycache__/globals.cpython-312.pyc,,
+flask/__pycache__/helpers.cpython-312.pyc,,
+flask/__pycache__/logging.cpython-312.pyc,,
+flask/__pycache__/sessions.cpython-312.pyc,,
+flask/__pycache__/signals.cpython-312.pyc,,
+flask/__pycache__/templating.cpython-312.pyc,,
+flask/__pycache__/testing.cpython-312.pyc,,
+flask/__pycache__/typing.cpython-312.pyc,,
+flask/__pycache__/views.cpython-312.pyc,,
+flask/__pycache__/wrappers.cpython-312.pyc,,
+flask/app.py,sha256=GE7QOE_N9THDjuzKx0gUI9aF4hLh0xwBDa7hLVsjy-o,61725
+flask/blueprints.py,sha256=p5QE2lY18GItbdr_RKRpZ8Do17g0PvQGIgZkSUDhX2k,4541
+flask/cli.py,sha256=XdmkBD74SnT0jrt2gqyHrE9oXGdWlcdTrJQR-i5aApY,37093
+flask/config.py,sha256=PiqF0DPam6HW0FH4CH1hpXTBe30NSzjPEOwrz1b6kt0,13219
+flask/ctx.py,sha256=4atDhJJ_cpV1VMq4qsfU4E_61M1oN93jlS2H9gjrl58,15120
+flask/debughelpers.py,sha256=PGIDhStW_efRjpaa3zHIpo-htStJOR41Ip3OJWPYBwo,6080
+flask/globals.py,sha256=XdQZmStBmPIs8t93tjx6pO7Bm3gobAaONWkFcUHaGas,1713
+flask/helpers.py,sha256=7njmzkFJvrPSQudsgONsgQzaGrGppeBINevKgWescPk,23521
+flask/json/__init__.py,sha256=hLNR898paqoefdeAhraa5wyJy-bmRB2k2dV4EgVy2Z8,5602
+flask/json/__pycache__/__init__.cpython-312.pyc,,
+flask/json/__pycache__/provider.cpython-312.pyc,,
+flask/json/__pycache__/tag.cpython-312.pyc,,
+flask/json/provider.py,sha256=5imEzY5HjV2HoUVrQbJLqXCzMNpZXfD0Y1XqdLV2XBA,7672
+flask/json/tag.py,sha256=DhaNwuIOhdt2R74oOC9Y4Z8ZprxFYiRb5dUP5byyINw,9281
+flask/logging.py,sha256=8sM3WMTubi1cBb2c_lPkWpN0J8dMAqrgKRYLLi1dCVI,2377
+flask/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flask/sansio/README.md,sha256=-0X1tECnilmz1cogx-YhNw5d7guK7GKrq_DEV2OzlU0,228
+flask/sansio/__pycache__/app.cpython-312.pyc,,
+flask/sansio/__pycache__/blueprints.cpython-312.pyc,,
+flask/sansio/__pycache__/scaffold.cpython-312.pyc,,
+flask/sansio/app.py,sha256=Wj9NVGtiR1jvkZ9gSFd91usUlM8H0g06aPVz2sMh4bw,38116
+flask/sansio/blueprints.py,sha256=Tqe-7EkZ-tbWchm8iDoCfD848f0_3nLv6NNjeIPvHwM,24637
+flask/sansio/scaffold.py,sha256=q6wM4Y4aYMGGN_Litsj3PYKpBS3Zvut0xhDmpBEHFdo,30387
+flask/sessions.py,sha256=pImSFQIDCPtV-XSI8ttAyTTbvtRMkhDeqJ8VPZZUaf0,15430
+flask/signals.py,sha256=V7lMUww7CqgJ2ThUBn1PiatZtQanOyt7OZpu2GZI-34,750
+flask/templating.py,sha256=2TcXLT85Asflm2W9WOSFxKCmYn5e49w_Jkg9-NaaJWo,7537
+flask/testing.py,sha256=5Dxg6VZ0ZPhjwG9ReUl4TrhvkjBYvgIzV949jkY0jIU,10100
+flask/typing.py,sha256=b7mMBIeAoOcAI_vFzzhfOm7KeZ_n868SIMw6xpX5KYQ,3166
+flask/views.py,sha256=xzJx6oJqGElThtEghZN7ZQGMw5TDFyuRxUkecwRuAoA,6962
+flask/wrappers.py,sha256=jUkv4mVek2Iq4hwxd4RvqrIMb69Bv0PElDgWLmd5ORo,9406
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/entry_points.txt
new file mode 100644
index 00000000..eec6733e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask-3.1.0.dist-info/entry_points.txt
@@ -0,0 +1,3 @@
+[console_scripts]
+flask=flask.cli:main
+
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__init__.py b/backend/venv/lib/python3.12/site-packages/flask/__init__.py
new file mode 100644
index 00000000..e86eb43e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/__init__.py
@@ -0,0 +1,60 @@
+from __future__ import annotations
+
+import typing as t
+
+from . import json as json
+from .app import Flask as Flask
+from .blueprints import Blueprint as Blueprint
+from .config import Config as Config
+from .ctx import after_this_request as after_this_request
+from .ctx import copy_current_request_context as copy_current_request_context
+from .ctx import has_app_context as has_app_context
+from .ctx import has_request_context as has_request_context
+from .globals import current_app as current_app
+from .globals import g as g
+from .globals import request as request
+from .globals import session as session
+from .helpers import abort as abort
+from .helpers import flash as flash
+from .helpers import get_flashed_messages as get_flashed_messages
+from .helpers import get_template_attribute as get_template_attribute
+from .helpers import make_response as make_response
+from .helpers import redirect as redirect
+from .helpers import send_file as send_file
+from .helpers import send_from_directory as send_from_directory
+from .helpers import stream_with_context as stream_with_context
+from .helpers import url_for as url_for
+from .json import jsonify as jsonify
+from .signals import appcontext_popped as appcontext_popped
+from .signals import appcontext_pushed as appcontext_pushed
+from .signals import appcontext_tearing_down as appcontext_tearing_down
+from .signals import before_render_template as before_render_template
+from .signals import got_request_exception as got_request_exception
+from .signals import message_flashed as message_flashed
+from .signals import request_finished as request_finished
+from .signals import request_started as request_started
+from .signals import request_tearing_down as request_tearing_down
+from .signals import template_rendered as template_rendered
+from .templating import render_template as render_template
+from .templating import render_template_string as render_template_string
+from .templating import stream_template as stream_template
+from .templating import stream_template_string as stream_template_string
+from .wrappers import Request as Request
+from .wrappers import Response as Response
+
+
+def __getattr__(name: str) -> t.Any:
+    if name == "__version__":
+        import importlib.metadata
+        import warnings
+
+        warnings.warn(
+            "The '__version__' attribute is deprecated and will be removed in"
+            " Flask 3.1. Use feature detection or"
+            " 'importlib.metadata.version(\"flask\")' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return importlib.metadata.version("flask")
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__main__.py b/backend/venv/lib/python3.12/site-packages/flask/__main__.py
new file mode 100644
index 00000000..4e28416e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/__main__.py
@@ -0,0 +1,3 @@
+from .cli import main
+
+main()
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d2f4c0d3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..555f91d0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/app.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/app.cpython-312.pyc
new file mode 100644
index 00000000..33c60bf7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/app.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/blueprints.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/blueprints.cpython-312.pyc
new file mode 100644
index 00000000..8cd4be73
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/blueprints.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/cli.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/cli.cpython-312.pyc
new file mode 100644
index 00000000..2025b558
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/cli.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..dc44f78c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/ctx.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/ctx.cpython-312.pyc
new file mode 100644
index 00000000..158229de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/ctx.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/debughelpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/debughelpers.cpython-312.pyc
new file mode 100644
index 00000000..24e66423
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/debughelpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/globals.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/globals.cpython-312.pyc
new file mode 100644
index 00000000..a8c3dd33
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/globals.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/helpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/helpers.cpython-312.pyc
new file mode 100644
index 00000000..dd7277df
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/helpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/logging.cpython-312.pyc
new file mode 100644
index 00000000..a715ca8e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/sessions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/sessions.cpython-312.pyc
new file mode 100644
index 00000000..81eda563
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/sessions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/signals.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/signals.cpython-312.pyc
new file mode 100644
index 00000000..18abdacf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/signals.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/templating.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/templating.cpython-312.pyc
new file mode 100644
index 00000000..cb20f0a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/templating.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/testing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/testing.cpython-312.pyc
new file mode 100644
index 00000000..9edbe5b9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/testing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/typing.cpython-312.pyc
new file mode 100644
index 00000000..59b26ffc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/views.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/views.cpython-312.pyc
new file mode 100644
index 00000000..3a372043
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/views.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/__pycache__/wrappers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/wrappers.cpython-312.pyc
new file mode 100644
index 00000000..c104b499
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/__pycache__/wrappers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/app.py b/backend/venv/lib/python3.12/site-packages/flask/app.py
new file mode 100644
index 00000000..905b2477
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/app.py
@@ -0,0 +1,1536 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import os
+import sys
+import typing as t
+import weakref
+from datetime import timedelta
+from inspect import iscoroutinefunction
+from itertools import chain
+from types import TracebackType
+from urllib.parse import quote as _url_quote
+
+import click
+from werkzeug.datastructures import Headers
+from werkzeug.datastructures import ImmutableDict
+from werkzeug.exceptions import BadRequestKeyError
+from werkzeug.exceptions import HTTPException
+from werkzeug.exceptions import InternalServerError
+from werkzeug.routing import BuildError
+from werkzeug.routing import MapAdapter
+from werkzeug.routing import RequestRedirect
+from werkzeug.routing import RoutingException
+from werkzeug.routing import Rule
+from werkzeug.serving import is_running_from_reloader
+from werkzeug.wrappers import Response as BaseResponse
+from werkzeug.wsgi import get_host
+
+from . import cli
+from . import typing as ft
+from .ctx import AppContext
+from .ctx import RequestContext
+from .globals import _cv_app
+from .globals import _cv_request
+from .globals import current_app
+from .globals import g
+from .globals import request
+from .globals import request_ctx
+from .globals import session
+from .helpers import get_debug_flag
+from .helpers import get_flashed_messages
+from .helpers import get_load_dotenv
+from .helpers import send_from_directory
+from .sansio.app import App
+from .sansio.scaffold import _sentinel
+from .sessions import SecureCookieSessionInterface
+from .sessions import SessionInterface
+from .signals import appcontext_tearing_down
+from .signals import got_request_exception
+from .signals import request_finished
+from .signals import request_started
+from .signals import request_tearing_down
+from .templating import Environment
+from .wrappers import Request
+from .wrappers import Response
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .testing import FlaskClient
+    from .testing import FlaskCliRunner
+    from .typing import HeadersValue
+
+T_shell_context_processor = t.TypeVar(
+    "T_shell_context_processor", bound=ft.ShellContextProcessorCallable
+)
+T_teardown = t.TypeVar("T_teardown", bound=ft.TeardownCallable)
+T_template_filter = t.TypeVar("T_template_filter", bound=ft.TemplateFilterCallable)
+T_template_global = t.TypeVar("T_template_global", bound=ft.TemplateGlobalCallable)
+T_template_test = t.TypeVar("T_template_test", bound=ft.TemplateTestCallable)
+
+
+def _make_timedelta(value: timedelta | int | None) -> timedelta | None:
+    if value is None or isinstance(value, timedelta):
+        return value
+
+    return timedelta(seconds=value)
+
+
+class Flask(App):
+    """The flask object implements a WSGI application and acts as the central
+    object.  It is passed the name of the module or package of the
+    application.  Once it is created it will act as a central registry for
+    the view functions, the URL rules, template configuration and much more.
+
+    The name of the package is used to resolve resources from inside the
+    package or the folder the module is contained in depending on if the
+    package parameter resolves to an actual python package (a folder with
+    an :file:`__init__.py` file inside) or a standard module (just a ``.py`` file).
+
+    For more information about resource loading, see :func:`open_resource`.
+
+    Usually you create a :class:`Flask` instance in your main module or
+    in the :file:`__init__.py` file of your package like this::
+
+        from flask import Flask
+        app = Flask(__name__)
+
+    .. admonition:: About the First Parameter
+
+        The idea of the first parameter is to give Flask an idea of what
+        belongs to your application.  This name is used to find resources
+        on the filesystem, can be used by extensions to improve debugging
+        information and a lot more.
+
+        So it's important what you provide there.  If you are using a single
+        module, `__name__` is always the correct value.  If you however are
+        using a package, it's usually recommended to hardcode the name of
+        your package there.
+
+        For example if your application is defined in :file:`yourapplication/app.py`
+        you should create it with one of the two versions below::
+
+            app = Flask('yourapplication')
+            app = Flask(__name__.split('.')[0])
+
+        Why is that?  The application will work even with `__name__`, thanks
+        to how resources are looked up.  However it will make debugging more
+        painful.  Certain extensions can make assumptions based on the
+        import name of your application.  For example the Flask-SQLAlchemy
+        extension will look for the code in your application that triggered
+        an SQL query in debug mode.  If the import name is not properly set
+        up, that debugging information is lost.  (For example it would only
+        pick up SQL queries in `yourapplication.app` and not
+        `yourapplication.views.frontend`)
+
+    .. versionadded:: 0.7
+       The `static_url_path`, `static_folder`, and `template_folder`
+       parameters were added.
+
+    .. versionadded:: 0.8
+       The `instance_path` and `instance_relative_config` parameters were
+       added.
+
+    .. versionadded:: 0.11
+       The `root_path` parameter was added.
+
+    .. versionadded:: 1.0
+       The ``host_matching`` and ``static_host`` parameters were added.
+
+    .. versionadded:: 1.0
+       The ``subdomain_matching`` parameter was added. Subdomain
+       matching needs to be enabled manually now. Setting
+       :data:`SERVER_NAME` does not implicitly enable it.
+
+    :param import_name: the name of the application package
+    :param static_url_path: can be used to specify a different path for the
+                            static files on the web.  Defaults to the name
+                            of the `static_folder` folder.
+    :param static_folder: The folder with static files that is served at
+        ``static_url_path``. Relative to the application ``root_path``
+        or an absolute path. Defaults to ``'static'``.
+    :param static_host: the host to use when adding the static route.
+        Defaults to None. Required when using ``host_matching=True``
+        with a ``static_folder`` configured.
+    :param host_matching: set ``url_map.host_matching`` attribute.
+        Defaults to False.
+    :param subdomain_matching: consider the subdomain relative to
+        :data:`SERVER_NAME` when matching routes. Defaults to False.
+    :param template_folder: the folder that contains the templates that should
+                            be used by the application.  Defaults to
+                            ``'templates'`` folder in the root path of the
+                            application.
+    :param instance_path: An alternative instance path for the application.
+                          By default the folder ``'instance'`` next to the
+                          package or module is assumed to be the instance
+                          path.
+    :param instance_relative_config: if set to ``True`` relative filenames
+                                     for loading the config are assumed to
+                                     be relative to the instance path instead
+                                     of the application root.
+    :param root_path: The path to the root of the application files.
+        This should only be set manually when it can't be detected
+        automatically, such as for namespace packages.
+    """
+
+    default_config = ImmutableDict(
+        {
+            "DEBUG": None,
+            "TESTING": False,
+            "PROPAGATE_EXCEPTIONS": None,
+            "SECRET_KEY": None,
+            "SECRET_KEY_FALLBACKS": None,
+            "PERMANENT_SESSION_LIFETIME": timedelta(days=31),
+            "USE_X_SENDFILE": False,
+            "TRUSTED_HOSTS": None,
+            "SERVER_NAME": None,
+            "APPLICATION_ROOT": "/",
+            "SESSION_COOKIE_NAME": "session",
+            "SESSION_COOKIE_DOMAIN": None,
+            "SESSION_COOKIE_PATH": None,
+            "SESSION_COOKIE_HTTPONLY": True,
+            "SESSION_COOKIE_SECURE": False,
+            "SESSION_COOKIE_PARTITIONED": False,
+            "SESSION_COOKIE_SAMESITE": None,
+            "SESSION_REFRESH_EACH_REQUEST": True,
+            "MAX_CONTENT_LENGTH": None,
+            "MAX_FORM_MEMORY_SIZE": 500_000,
+            "MAX_FORM_PARTS": 1_000,
+            "SEND_FILE_MAX_AGE_DEFAULT": None,
+            "TRAP_BAD_REQUEST_ERRORS": None,
+            "TRAP_HTTP_EXCEPTIONS": False,
+            "EXPLAIN_TEMPLATE_LOADING": False,
+            "PREFERRED_URL_SCHEME": "http",
+            "TEMPLATES_AUTO_RELOAD": None,
+            "MAX_COOKIE_SIZE": 4093,
+            "PROVIDE_AUTOMATIC_OPTIONS": True,
+        }
+    )
+
+    #: The class that is used for request objects.  See :class:`~flask.Request`
+    #: for more information.
+    request_class: type[Request] = Request
+
+    #: The class that is used for response objects.  See
+    #: :class:`~flask.Response` for more information.
+    response_class: type[Response] = Response
+
+    #: the session interface to use.  By default an instance of
+    #: :class:`~flask.sessions.SecureCookieSessionInterface` is used here.
+    #:
+    #: .. versionadded:: 0.8
+    session_interface: SessionInterface = SecureCookieSessionInterface()
+
+    def __init__(
+        self,
+        import_name: str,
+        static_url_path: str | None = None,
+        static_folder: str | os.PathLike[str] | None = "static",
+        static_host: str | None = None,
+        host_matching: bool = False,
+        subdomain_matching: bool = False,
+        template_folder: str | os.PathLike[str] | None = "templates",
+        instance_path: str | None = None,
+        instance_relative_config: bool = False,
+        root_path: str | None = None,
+    ):
+        super().__init__(
+            import_name=import_name,
+            static_url_path=static_url_path,
+            static_folder=static_folder,
+            static_host=static_host,
+            host_matching=host_matching,
+            subdomain_matching=subdomain_matching,
+            template_folder=template_folder,
+            instance_path=instance_path,
+            instance_relative_config=instance_relative_config,
+            root_path=root_path,
+        )
+
+        #: The Click command group for registering CLI commands for this
+        #: object. The commands are available from the ``flask`` command
+        #: once the application has been discovered and blueprints have
+        #: been registered.
+        self.cli = cli.AppGroup()
+
+        # Set the name of the Click group in case someone wants to add
+        # the app's commands to another CLI tool.
+        self.cli.name = self.name
+
+        # Add a static route using the provided static_url_path, static_host,
+        # and static_folder if there is a configured static_folder.
+        # Note we do this without checking if static_folder exists.
+        # For one, it might be created while the server is running (e.g. during
+        # development). Also, Google App Engine stores static files somewhere
+        if self.has_static_folder:
+            assert (
+                bool(static_host) == host_matching
+            ), "Invalid static_host/host_matching combination"
+            # Use a weakref to avoid creating a reference cycle between the app
+            # and the view function (see #3761).
+            self_ref = weakref.ref(self)
+            self.add_url_rule(
+                f"{self.static_url_path}/<path:filename>",
+                endpoint="static",
+                host=static_host,
+                view_func=lambda **kw: self_ref().send_static_file(**kw),  # type: ignore # noqa: B950
+            )
+
+    def get_send_file_max_age(self, filename: str | None) -> int | None:
+        """Used by :func:`send_file` to determine the ``max_age`` cache
+        value for a given file path if it wasn't passed.
+
+        By default, this returns :data:`SEND_FILE_MAX_AGE_DEFAULT` from
+        the configuration of :data:`~flask.current_app`. This defaults
+        to ``None``, which tells the browser to use conditional requests
+        instead of a timed cache, which is usually preferable.
+
+        Note this is a duplicate of the same method in the Flask
+        class.
+
+        .. versionchanged:: 2.0
+            The default configuration is ``None`` instead of 12 hours.
+
+        .. versionadded:: 0.9
+        """
+        value = current_app.config["SEND_FILE_MAX_AGE_DEFAULT"]
+
+        if value is None:
+            return None
+
+        if isinstance(value, timedelta):
+            return int(value.total_seconds())
+
+        return value  # type: ignore[no-any-return]
+
+    def send_static_file(self, filename: str) -> Response:
+        """The view function used to serve files from
+        :attr:`static_folder`. A route is automatically registered for
+        this view at :attr:`static_url_path` if :attr:`static_folder` is
+        set.
+
+        Note this is a duplicate of the same method in the Flask
+        class.
+
+        .. versionadded:: 0.5
+
+        """
+        if not self.has_static_folder:
+            raise RuntimeError("'static_folder' must be set to serve static_files.")
+
+        # send_file only knows to call get_send_file_max_age on the app,
+        # call it here so it works for blueprints too.
+        max_age = self.get_send_file_max_age(filename)
+        return send_from_directory(
+            t.cast(str, self.static_folder), filename, max_age=max_age
+        )
+
+    def open_resource(
+        self, resource: str, mode: str = "rb", encoding: str | None = None
+    ) -> t.IO[t.AnyStr]:
+        """Open a resource file relative to :attr:`root_path` for reading.
+
+        For example, if the file ``schema.sql`` is next to the file
+        ``app.py`` where the ``Flask`` app is defined, it can be opened
+        with:
+
+        .. code-block:: python
+
+            with app.open_resource("schema.sql") as f:
+                conn.executescript(f.read())
+
+        :param resource: Path to the resource relative to :attr:`root_path`.
+        :param mode: Open the file in this mode. Only reading is supported,
+            valid values are ``"r"`` (or ``"rt"``) and ``"rb"``.
+        :param encoding: Open the file with this encoding when opening in text
+            mode. This is ignored when opening in binary mode.
+
+        .. versionchanged:: 3.1
+            Added the ``encoding`` parameter.
+        """
+        if mode not in {"r", "rt", "rb"}:
+            raise ValueError("Resources can only be opened for reading.")
+
+        path = os.path.join(self.root_path, resource)
+
+        if mode == "rb":
+            return open(path, mode)  # pyright: ignore
+
+        return open(path, mode, encoding=encoding)
+
+    def open_instance_resource(
+        self, resource: str, mode: str = "rb", encoding: str | None = "utf-8"
+    ) -> t.IO[t.AnyStr]:
+        """Open a resource file relative to the application's instance folder
+        :attr:`instance_path`. Unlike :meth:`open_resource`, files in the
+        instance folder can be opened for writing.
+
+        :param resource: Path to the resource relative to :attr:`instance_path`.
+        :param mode: Open the file in this mode.
+        :param encoding: Open the file with this encoding when opening in text
+            mode. This is ignored when opening in binary mode.
+
+        .. versionchanged:: 3.1
+            Added the ``encoding`` parameter.
+        """
+        path = os.path.join(self.instance_path, resource)
+
+        if "b" in mode:
+            return open(path, mode)
+
+        return open(path, mode, encoding=encoding)
+
+    def create_jinja_environment(self) -> Environment:
+        """Create the Jinja environment based on :attr:`jinja_options`
+        and the various Jinja-related methods of the app. Changing
+        :attr:`jinja_options` after this will have no effect. Also adds
+        Flask-related globals and filters to the environment.
+
+        .. versionchanged:: 0.11
+           ``Environment.auto_reload`` set in accordance with
+           ``TEMPLATES_AUTO_RELOAD`` configuration option.
+
+        .. versionadded:: 0.5
+        """
+        options = dict(self.jinja_options)
+
+        if "autoescape" not in options:
+            options["autoescape"] = self.select_jinja_autoescape
+
+        if "auto_reload" not in options:
+            auto_reload = self.config["TEMPLATES_AUTO_RELOAD"]
+
+            if auto_reload is None:
+                auto_reload = self.debug
+
+            options["auto_reload"] = auto_reload
+
+        rv = self.jinja_environment(self, **options)
+        rv.globals.update(
+            url_for=self.url_for,
+            get_flashed_messages=get_flashed_messages,
+            config=self.config,
+            # request, session and g are normally added with the
+            # context processor for efficiency reasons but for imported
+            # templates we also want the proxies in there.
+            request=request,
+            session=session,
+            g=g,
+        )
+        rv.policies["json.dumps_function"] = self.json.dumps
+        return rv
+
+    def create_url_adapter(self, request: Request | None) -> MapAdapter | None:
+        """Creates a URL adapter for the given request. The URL adapter
+        is created at a point where the request context is not yet set
+        up so the request is passed explicitly.
+
+        .. versionchanged:: 3.1
+            If :data:`SERVER_NAME` is set, it does not restrict requests to
+            only that domain, for both ``subdomain_matching`` and
+            ``host_matching``.
+
+        .. versionchanged:: 1.0
+            :data:`SERVER_NAME` no longer implicitly enables subdomain
+            matching. Use :attr:`subdomain_matching` instead.
+
+        .. versionchanged:: 0.9
+           This can be called outside a request when the URL adapter is created
+           for an application context.
+
+        .. versionadded:: 0.6
+        """
+        if request is not None:
+            if (trusted_hosts := self.config["TRUSTED_HOSTS"]) is not None:
+                request.trusted_hosts = trusted_hosts
+
+            # Check trusted_hosts here until bind_to_environ does.
+            request.host = get_host(request.environ, request.trusted_hosts)  # pyright: ignore
+            subdomain = None
+            server_name = self.config["SERVER_NAME"]
+
+            if self.url_map.host_matching:
+                # Don't pass SERVER_NAME, otherwise it's used and the actual
+                # host is ignored, which breaks host matching.
+                server_name = None
+            elif not self.subdomain_matching:
+                # Werkzeug doesn't implement subdomain matching yet. Until then,
+                # disable it by forcing the current subdomain to the default, or
+                # the empty string.
+                subdomain = self.url_map.default_subdomain or ""
+
+            return self.url_map.bind_to_environ(
+                request.environ, server_name=server_name, subdomain=subdomain
+            )
+
+        # Need at least SERVER_NAME to match/build outside a request.
+        if self.config["SERVER_NAME"] is not None:
+            return self.url_map.bind(
+                self.config["SERVER_NAME"],
+                script_name=self.config["APPLICATION_ROOT"],
+                url_scheme=self.config["PREFERRED_URL_SCHEME"],
+            )
+
+        return None
+
+    def raise_routing_exception(self, request: Request) -> t.NoReturn:
+        """Intercept routing exceptions and possibly do something else.
+
+        In debug mode, intercept a routing redirect and replace it with
+        an error if the body will be discarded.
+
+        With modern Werkzeug this shouldn't occur, since it now uses a
+        308 status which tells the browser to resend the method and
+        body.
+
+        .. versionchanged:: 2.1
+            Don't intercept 307 and 308 redirects.
+
+        :meta private:
+        :internal:
+        """
+        if (
+            not self.debug
+            or not isinstance(request.routing_exception, RequestRedirect)
+            or request.routing_exception.code in {307, 308}
+            or request.method in {"GET", "HEAD", "OPTIONS"}
+        ):
+            raise request.routing_exception  # type: ignore[misc]
+
+        from .debughelpers import FormDataRoutingRedirect
+
+        raise FormDataRoutingRedirect(request)
+
+    def update_template_context(self, context: dict[str, t.Any]) -> None:
+        """Update the template context with some commonly used variables.
+        This injects request, session, config and g into the template
+        context as well as everything template context processors want
+        to inject.  Note that the as of Flask 0.6, the original values
+        in the context will not be overridden if a context processor
+        decides to return a value with the same key.
+
+        :param context: the context as a dictionary that is updated in place
+                        to add extra variables.
+        """
+        names: t.Iterable[str | None] = (None,)
+
+        # A template may be rendered outside a request context.
+        if request:
+            names = chain(names, reversed(request.blueprints))
+
+        # The values passed to render_template take precedence. Keep a
+        # copy to re-apply after all context functions.
+        orig_ctx = context.copy()
+
+        for name in names:
+            if name in self.template_context_processors:
+                for func in self.template_context_processors[name]:
+                    context.update(self.ensure_sync(func)())
+
+        context.update(orig_ctx)
+
+    def make_shell_context(self) -> dict[str, t.Any]:
+        """Returns the shell context for an interactive shell for this
+        application.  This runs all the registered shell context
+        processors.
+
+        .. versionadded:: 0.11
+        """
+        rv = {"app": self, "g": g}
+        for processor in self.shell_context_processors:
+            rv.update(processor())
+        return rv
+
+    def run(
+        self,
+        host: str | None = None,
+        port: int | None = None,
+        debug: bool | None = None,
+        load_dotenv: bool = True,
+        **options: t.Any,
+    ) -> None:
+        """Runs the application on a local development server.
+
+        Do not use ``run()`` in a production setting. It is not intended to
+        meet security and performance requirements for a production server.
+        Instead, see :doc:`/deploying/index` for WSGI server recommendations.
+
+        If the :attr:`debug` flag is set the server will automatically reload
+        for code changes and show a debugger in case an exception happened.
+
+        If you want to run the application in debug mode, but disable the
+        code execution on the interactive debugger, you can pass
+        ``use_evalex=False`` as parameter.  This will keep the debugger's
+        traceback screen active, but disable code execution.
+
+        It is not recommended to use this function for development with
+        automatic reloading as this is badly supported.  Instead you should
+        be using the :command:`flask` command line script's ``run`` support.
+
+        .. admonition:: Keep in Mind
+
+           Flask will suppress any server error with a generic error page
+           unless it is in debug mode.  As such to enable just the
+           interactive debugger without the code reloading, you have to
+           invoke :meth:`run` with ``debug=True`` and ``use_reloader=False``.
+           Setting ``use_debugger`` to ``True`` without being in debug mode
+           won't catch any exceptions because there won't be any to
+           catch.
+
+        :param host: the hostname to listen on. Set this to ``'0.0.0.0'`` to
+            have the server available externally as well. Defaults to
+            ``'127.0.0.1'`` or the host in the ``SERVER_NAME`` config variable
+            if present.
+        :param port: the port of the webserver. Defaults to ``5000`` or the
+            port defined in the ``SERVER_NAME`` config variable if present.
+        :param debug: if given, enable or disable debug mode. See
+            :attr:`debug`.
+        :param load_dotenv: Load the nearest :file:`.env` and :file:`.flaskenv`
+            files to set environment variables. Will also change the working
+            directory to the directory containing the first file found.
+        :param options: the options to be forwarded to the underlying Werkzeug
+            server. See :func:`werkzeug.serving.run_simple` for more
+            information.
+
+        .. versionchanged:: 1.0
+            If installed, python-dotenv will be used to load environment
+            variables from :file:`.env` and :file:`.flaskenv` files.
+
+            The :envvar:`FLASK_DEBUG` environment variable will override :attr:`debug`.
+
+            Threaded mode is enabled by default.
+
+        .. versionchanged:: 0.10
+            The default port is now picked from the ``SERVER_NAME``
+            variable.
+        """
+        # Ignore this call so that it doesn't start another server if
+        # the 'flask run' command is used.
+        if os.environ.get("FLASK_RUN_FROM_CLI") == "true":
+            if not is_running_from_reloader():
+                click.secho(
+                    " * Ignoring a call to 'app.run()' that would block"
+                    " the current 'flask' CLI command.\n"
+                    "   Only call 'app.run()' in an 'if __name__ =="
+                    ' "__main__"\' guard.',
+                    fg="red",
+                )
+
+            return
+
+        if get_load_dotenv(load_dotenv):
+            cli.load_dotenv()
+
+            # if set, env var overrides existing value
+            if "FLASK_DEBUG" in os.environ:
+                self.debug = get_debug_flag()
+
+        # debug passed to method overrides all other sources
+        if debug is not None:
+            self.debug = bool(debug)
+
+        server_name = self.config.get("SERVER_NAME")
+        sn_host = sn_port = None
+
+        if server_name:
+            sn_host, _, sn_port = server_name.partition(":")
+
+        if not host:
+            if sn_host:
+                host = sn_host
+            else:
+                host = "127.0.0.1"
+
+        if port or port == 0:
+            port = int(port)
+        elif sn_port:
+            port = int(sn_port)
+        else:
+            port = 5000
+
+        options.setdefault("use_reloader", self.debug)
+        options.setdefault("use_debugger", self.debug)
+        options.setdefault("threaded", True)
+
+        cli.show_server_banner(self.debug, self.name)
+
+        from werkzeug.serving import run_simple
+
+        try:
+            run_simple(t.cast(str, host), port, self, **options)
+        finally:
+            # reset the first request information if the development server
+            # reset normally.  This makes it possible to restart the server
+            # without reloader and that stuff from an interactive shell.
+            self._got_first_request = False
+
+    def test_client(self, use_cookies: bool = True, **kwargs: t.Any) -> FlaskClient:
+        """Creates a test client for this application.  For information
+        about unit testing head over to :doc:`/testing`.
+
+        Note that if you are testing for assertions or exceptions in your
+        application code, you must set ``app.testing = True`` in order for the
+        exceptions to propagate to the test client.  Otherwise, the exception
+        will be handled by the application (not visible to the test client) and
+        the only indication of an AssertionError or other exception will be a
+        500 status code response to the test client.  See the :attr:`testing`
+        attribute.  For example::
+
+            app.testing = True
+            client = app.test_client()
+
+        The test client can be used in a ``with`` block to defer the closing down
+        of the context until the end of the ``with`` block.  This is useful if
+        you want to access the context locals for testing::
+
+            with app.test_client() as c:
+                rv = c.get('/?vodka=42')
+                assert request.args['vodka'] == '42'
+
+        Additionally, you may pass optional keyword arguments that will then
+        be passed to the application's :attr:`test_client_class` constructor.
+        For example::
+
+            from flask.testing import FlaskClient
+
+            class CustomClient(FlaskClient):
+                def __init__(self, *args, **kwargs):
+                    self._authentication = kwargs.pop("authentication")
+                    super(CustomClient,self).__init__( *args, **kwargs)
+
+            app.test_client_class = CustomClient
+            client = app.test_client(authentication='Basic ....')
+
+        See :class:`~flask.testing.FlaskClient` for more information.
+
+        .. versionchanged:: 0.4
+           added support for ``with`` block usage for the client.
+
+        .. versionadded:: 0.7
+           The `use_cookies` parameter was added as well as the ability
+           to override the client to be used by setting the
+           :attr:`test_client_class` attribute.
+
+        .. versionchanged:: 0.11
+           Added `**kwargs` to support passing additional keyword arguments to
+           the constructor of :attr:`test_client_class`.
+        """
+        cls = self.test_client_class
+        if cls is None:
+            from .testing import FlaskClient as cls
+        return cls(  # type: ignore
+            self, self.response_class, use_cookies=use_cookies, **kwargs
+        )
+
+    def test_cli_runner(self, **kwargs: t.Any) -> FlaskCliRunner:
+        """Create a CLI runner for testing CLI commands.
+        See :ref:`testing-cli`.
+
+        Returns an instance of :attr:`test_cli_runner_class`, by default
+        :class:`~flask.testing.FlaskCliRunner`. The Flask app object is
+        passed as the first argument.
+
+        .. versionadded:: 1.0
+        """
+        cls = self.test_cli_runner_class
+
+        if cls is None:
+            from .testing import FlaskCliRunner as cls
+
+        return cls(self, **kwargs)  # type: ignore
+
+    def handle_http_exception(
+        self, e: HTTPException
+    ) -> HTTPException | ft.ResponseReturnValue:
+        """Handles an HTTP exception.  By default this will invoke the
+        registered error handlers and fall back to returning the
+        exception as response.
+
+        .. versionchanged:: 1.0.3
+            ``RoutingException``, used internally for actions such as
+             slash redirects during routing, is not passed to error
+             handlers.
+
+        .. versionchanged:: 1.0
+            Exceptions are looked up by code *and* by MRO, so
+            ``HTTPException`` subclasses can be handled with a catch-all
+            handler for the base ``HTTPException``.
+
+        .. versionadded:: 0.3
+        """
+        # Proxy exceptions don't have error codes.  We want to always return
+        # those unchanged as errors
+        if e.code is None:
+            return e
+
+        # RoutingExceptions are used internally to trigger routing
+        # actions, such as slash redirects raising RequestRedirect. They
+        # are not raised or handled in user code.
+        if isinstance(e, RoutingException):
+            return e
+
+        handler = self._find_error_handler(e, request.blueprints)
+        if handler is None:
+            return e
+        return self.ensure_sync(handler)(e)  # type: ignore[no-any-return]
+
+    def handle_user_exception(
+        self, e: Exception
+    ) -> HTTPException | ft.ResponseReturnValue:
+        """This method is called whenever an exception occurs that
+        should be handled. A special case is :class:`~werkzeug
+        .exceptions.HTTPException` which is forwarded to the
+        :meth:`handle_http_exception` method. This function will either
+        return a response value or reraise the exception with the same
+        traceback.
+
+        .. versionchanged:: 1.0
+            Key errors raised from request data like ``form`` show the
+            bad key in debug mode rather than a generic bad request
+            message.
+
+        .. versionadded:: 0.7
+        """
+        if isinstance(e, BadRequestKeyError) and (
+            self.debug or self.config["TRAP_BAD_REQUEST_ERRORS"]
+        ):
+            e.show_exception = True
+
+        if isinstance(e, HTTPException) and not self.trap_http_exception(e):
+            return self.handle_http_exception(e)
+
+        handler = self._find_error_handler(e, request.blueprints)
+
+        if handler is None:
+            raise
+
+        return self.ensure_sync(handler)(e)  # type: ignore[no-any-return]
+
+    def handle_exception(self, e: Exception) -> Response:
+        """Handle an exception that did not have an error handler
+        associated with it, or that was raised from an error handler.
+        This always causes a 500 ``InternalServerError``.
+
+        Always sends the :data:`got_request_exception` signal.
+
+        If :data:`PROPAGATE_EXCEPTIONS` is ``True``, such as in debug
+        mode, the error will be re-raised so that the debugger can
+        display it. Otherwise, the original exception is logged, and
+        an :exc:`~werkzeug.exceptions.InternalServerError` is returned.
+
+        If an error handler is registered for ``InternalServerError`` or
+        ``500``, it will be used. For consistency, the handler will
+        always receive the ``InternalServerError``. The original
+        unhandled exception is available as ``e.original_exception``.
+
+        .. versionchanged:: 1.1.0
+            Always passes the ``InternalServerError`` instance to the
+            handler, setting ``original_exception`` to the unhandled
+            error.
+
+        .. versionchanged:: 1.1.0
+            ``after_request`` functions and other finalization is done
+            even for the default 500 response when there is no handler.
+
+        .. versionadded:: 0.3
+        """
+        exc_info = sys.exc_info()
+        got_request_exception.send(self, _async_wrapper=self.ensure_sync, exception=e)
+        propagate = self.config["PROPAGATE_EXCEPTIONS"]
+
+        if propagate is None:
+            propagate = self.testing or self.debug
+
+        if propagate:
+            # Re-raise if called with an active exception, otherwise
+            # raise the passed in exception.
+            if exc_info[1] is e:
+                raise
+
+            raise e
+
+        self.log_exception(exc_info)
+        server_error: InternalServerError | ft.ResponseReturnValue
+        server_error = InternalServerError(original_exception=e)
+        handler = self._find_error_handler(server_error, request.blueprints)
+
+        if handler is not None:
+            server_error = self.ensure_sync(handler)(server_error)
+
+        return self.finalize_request(server_error, from_error_handler=True)
+
+    def log_exception(
+        self,
+        exc_info: (tuple[type, BaseException, TracebackType] | tuple[None, None, None]),
+    ) -> None:
+        """Logs an exception.  This is called by :meth:`handle_exception`
+        if debugging is disabled and right before the handler is called.
+        The default implementation logs the exception as error on the
+        :attr:`logger`.
+
+        .. versionadded:: 0.8
+        """
+        self.logger.error(
+            f"Exception on {request.path} [{request.method}]", exc_info=exc_info
+        )
+
+    def dispatch_request(self) -> ft.ResponseReturnValue:
+        """Does the request dispatching.  Matches the URL and returns the
+        return value of the view or error handler.  This does not have to
+        be a response object.  In order to convert the return value to a
+        proper response object, call :func:`make_response`.
+
+        .. versionchanged:: 0.7
+           This no longer does the exception handling, this code was
+           moved to the new :meth:`full_dispatch_request`.
+        """
+        req = request_ctx.request
+        if req.routing_exception is not None:
+            self.raise_routing_exception(req)
+        rule: Rule = req.url_rule  # type: ignore[assignment]
+        # if we provide automatic options for this URL and the
+        # request came with the OPTIONS method, reply automatically
+        if (
+            getattr(rule, "provide_automatic_options", False)
+            and req.method == "OPTIONS"
+        ):
+            return self.make_default_options_response()
+        # otherwise dispatch to the handler for that endpoint
+        view_args: dict[str, t.Any] = req.view_args  # type: ignore[assignment]
+        return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)  # type: ignore[no-any-return]
+
+    def full_dispatch_request(self) -> Response:
+        """Dispatches the request and on top of that performs request
+        pre and postprocessing as well as HTTP exception catching and
+        error handling.
+
+        .. versionadded:: 0.7
+        """
+        self._got_first_request = True
+
+        try:
+            request_started.send(self, _async_wrapper=self.ensure_sync)
+            rv = self.preprocess_request()
+            if rv is None:
+                rv = self.dispatch_request()
+        except Exception as e:
+            rv = self.handle_user_exception(e)
+        return self.finalize_request(rv)
+
+    def finalize_request(
+        self,
+        rv: ft.ResponseReturnValue | HTTPException,
+        from_error_handler: bool = False,
+    ) -> Response:
+        """Given the return value from a view function this finalizes
+        the request by converting it into a response and invoking the
+        postprocessing functions.  This is invoked for both normal
+        request dispatching as well as error handlers.
+
+        Because this means that it might be called as a result of a
+        failure a special safe mode is available which can be enabled
+        with the `from_error_handler` flag.  If enabled, failures in
+        response processing will be logged and otherwise ignored.
+
+        :internal:
+        """
+        response = self.make_response(rv)
+        try:
+            response = self.process_response(response)
+            request_finished.send(
+                self, _async_wrapper=self.ensure_sync, response=response
+            )
+        except Exception:
+            if not from_error_handler:
+                raise
+            self.logger.exception(
+                "Request finalizing failed with an error while handling an error"
+            )
+        return response
+
+    def make_default_options_response(self) -> Response:
+        """This method is called to create the default ``OPTIONS`` response.
+        This can be changed through subclassing to change the default
+        behavior of ``OPTIONS`` responses.
+
+        .. versionadded:: 0.7
+        """
+        adapter = request_ctx.url_adapter
+        methods = adapter.allowed_methods()  # type: ignore[union-attr]
+        rv = self.response_class()
+        rv.allow.update(methods)
+        return rv
+
+    def ensure_sync(self, func: t.Callable[..., t.Any]) -> t.Callable[..., t.Any]:
+        """Ensure that the function is synchronous for WSGI workers.
+        Plain ``def`` functions are returned as-is. ``async def``
+        functions are wrapped to run and wait for the response.
+
+        Override this method to change how the app runs async views.
+
+        .. versionadded:: 2.0
+        """
+        if iscoroutinefunction(func):
+            return self.async_to_sync(func)
+
+        return func
+
+    def async_to_sync(
+        self, func: t.Callable[..., t.Coroutine[t.Any, t.Any, t.Any]]
+    ) -> t.Callable[..., t.Any]:
+        """Return a sync function that will run the coroutine function.
+
+        .. code-block:: python
+
+            result = app.async_to_sync(func)(*args, **kwargs)
+
+        Override this method to change how the app converts async code
+        to be synchronously callable.
+
+        .. versionadded:: 2.0
+        """
+        try:
+            from asgiref.sync import async_to_sync as asgiref_async_to_sync
+        except ImportError:
+            raise RuntimeError(
+                "Install Flask with the 'async' extra in order to use async views."
+            ) from None
+
+        return asgiref_async_to_sync(func)
+
+    def url_for(
+        self,
+        /,
+        endpoint: str,
+        *,
+        _anchor: str | None = None,
+        _method: str | None = None,
+        _scheme: str | None = None,
+        _external: bool | None = None,
+        **values: t.Any,
+    ) -> str:
+        """Generate a URL to the given endpoint with the given values.
+
+        This is called by :func:`flask.url_for`, and can be called
+        directly as well.
+
+        An *endpoint* is the name of a URL rule, usually added with
+        :meth:`@app.route() <route>`, and usually the same name as the
+        view function. A route defined in a :class:`~flask.Blueprint`
+        will prepend the blueprint's name separated by a ``.`` to the
+        endpoint.
+
+        In some cases, such as email messages, you want URLs to include
+        the scheme and domain, like ``https://example.com/hello``. When
+        not in an active request, URLs will be external by default, but
+        this requires setting :data:`SERVER_NAME` so Flask knows what
+        domain to use. :data:`APPLICATION_ROOT` and
+        :data:`PREFERRED_URL_SCHEME` should also be configured as
+        needed. This config is only used when not in an active request.
+
+        Functions can be decorated with :meth:`url_defaults` to modify
+        keyword arguments before the URL is built.
+
+        If building fails for some reason, such as an unknown endpoint
+        or incorrect values, the app's :meth:`handle_url_build_error`
+        method is called. If that returns a string, that is returned,
+        otherwise a :exc:`~werkzeug.routing.BuildError` is raised.
+
+        :param endpoint: The endpoint name associated with the URL to
+            generate. If this starts with a ``.``, the current blueprint
+            name (if any) will be used.
+        :param _anchor: If given, append this as ``#anchor`` to the URL.
+        :param _method: If given, generate the URL associated with this
+            method for the endpoint.
+        :param _scheme: If given, the URL will have this scheme if it
+            is external.
+        :param _external: If given, prefer the URL to be internal
+            (False) or require it to be external (True). External URLs
+            include the scheme and domain. When not in an active
+            request, URLs are external by default.
+        :param values: Values to use for the variable parts of the URL
+            rule. Unknown keys are appended as query string arguments,
+            like ``?a=b&c=d``.
+
+        .. versionadded:: 2.2
+            Moved from ``flask.url_for``, which calls this method.
+        """
+        req_ctx = _cv_request.get(None)
+
+        if req_ctx is not None:
+            url_adapter = req_ctx.url_adapter
+            blueprint_name = req_ctx.request.blueprint
+
+            # If the endpoint starts with "." and the request matches a
+            # blueprint, the endpoint is relative to the blueprint.
+            if endpoint[:1] == ".":
+                if blueprint_name is not None:
+                    endpoint = f"{blueprint_name}{endpoint}"
+                else:
+                    endpoint = endpoint[1:]
+
+            # When in a request, generate a URL without scheme and
+            # domain by default, unless a scheme is given.
+            if _external is None:
+                _external = _scheme is not None
+        else:
+            app_ctx = _cv_app.get(None)
+
+            # If called by helpers.url_for, an app context is active,
+            # use its url_adapter. Otherwise, app.url_for was called
+            # directly, build an adapter.
+            if app_ctx is not None:
+                url_adapter = app_ctx.url_adapter
+            else:
+                url_adapter = self.create_url_adapter(None)
+
+            if url_adapter is None:
+                raise RuntimeError(
+                    "Unable to build URLs outside an active request"
+                    " without 'SERVER_NAME' configured. Also configure"
+                    " 'APPLICATION_ROOT' and 'PREFERRED_URL_SCHEME' as"
+                    " needed."
+                )
+
+            # When outside a request, generate a URL with scheme and
+            # domain by default.
+            if _external is None:
+                _external = True
+
+        # It is an error to set _scheme when _external=False, in order
+        # to avoid accidental insecure URLs.
+        if _scheme is not None and not _external:
+            raise ValueError("When specifying '_scheme', '_external' must be True.")
+
+        self.inject_url_defaults(endpoint, values)
+
+        try:
+            rv = url_adapter.build(  # type: ignore[union-attr]
+                endpoint,
+                values,
+                method=_method,
+                url_scheme=_scheme,
+                force_external=_external,
+            )
+        except BuildError as error:
+            values.update(
+                _anchor=_anchor, _method=_method, _scheme=_scheme, _external=_external
+            )
+            return self.handle_url_build_error(error, endpoint, values)
+
+        if _anchor is not None:
+            _anchor = _url_quote(_anchor, safe="%!#$&'()*+,/:;=?@")
+            rv = f"{rv}#{_anchor}"
+
+        return rv
+
+    def make_response(self, rv: ft.ResponseReturnValue) -> Response:
+        """Convert the return value from a view function to an instance of
+        :attr:`response_class`.
+
+        :param rv: the return value from the view function. The view function
+            must return a response. Returning ``None``, or the view ending
+            without returning, is not allowed. The following types are allowed
+            for ``view_rv``:
+
+            ``str``
+                A response object is created with the string encoded to UTF-8
+                as the body.
+
+            ``bytes``
+                A response object is created with the bytes as the body.
+
+            ``dict``
+                A dictionary that will be jsonify'd before being returned.
+
+            ``list``
+                A list that will be jsonify'd before being returned.
+
+            ``generator`` or ``iterator``
+                A generator that returns ``str`` or ``bytes`` to be
+                streamed as the response.
+
+            ``tuple``
+                Either ``(body, status, headers)``, ``(body, status)``, or
+                ``(body, headers)``, where ``body`` is any of the other types
+                allowed here, ``status`` is a string or an integer, and
+                ``headers`` is a dictionary or a list of ``(key, value)``
+                tuples. If ``body`` is a :attr:`response_class` instance,
+                ``status`` overwrites the exiting value and ``headers`` are
+                extended.
+
+            :attr:`response_class`
+                The object is returned unchanged.
+
+            other :class:`~werkzeug.wrappers.Response` class
+                The object is coerced to :attr:`response_class`.
+
+            :func:`callable`
+                The function is called as a WSGI application. The result is
+                used to create a response object.
+
+        .. versionchanged:: 2.2
+            A generator will be converted to a streaming response.
+            A list will be converted to a JSON response.
+
+        .. versionchanged:: 1.1
+            A dict will be converted to a JSON response.
+
+        .. versionchanged:: 0.9
+           Previously a tuple was interpreted as the arguments for the
+           response object.
+        """
+
+        status: int | None = None
+        headers: HeadersValue | None = None
+
+        # unpack tuple returns
+        if isinstance(rv, tuple):
+            len_rv = len(rv)
+
+            # a 3-tuple is unpacked directly
+            if len_rv == 3:
+                rv, status, headers = rv  # type: ignore[misc]
+            # decide if a 2-tuple has status or headers
+            elif len_rv == 2:
+                if isinstance(rv[1], (Headers, dict, tuple, list)):
+                    rv, headers = rv  # pyright: ignore
+                else:
+                    rv, status = rv  # type: ignore[assignment,misc]
+            # other sized tuples are not allowed
+            else:
+                raise TypeError(
+                    "The view function did not return a valid response tuple."
+                    " The tuple must have the form (body, status, headers),"
+                    " (body, status), or (body, headers)."
+                )
+
+        # the body must not be None
+        if rv is None:
+            raise TypeError(
+                f"The view function for {request.endpoint!r} did not"
+                " return a valid response. The function either returned"
+                " None or ended without a return statement."
+            )
+
+        # make sure the body is an instance of the response class
+        if not isinstance(rv, self.response_class):
+            if isinstance(rv, (str, bytes, bytearray)) or isinstance(rv, cabc.Iterator):
+                # let the response class set the status and headers instead of
+                # waiting to do it manually, so that the class can handle any
+                # special logic
+                rv = self.response_class(
+                    rv,
+                    status=status,
+                    headers=headers,  # type: ignore[arg-type]
+                )
+                status = headers = None
+            elif isinstance(rv, (dict, list)):
+                rv = self.json.response(rv)
+            elif isinstance(rv, BaseResponse) or callable(rv):
+                # evaluate a WSGI callable, or coerce a different response
+                # class to the correct type
+                try:
+                    rv = self.response_class.force_type(
+                        rv,  # type: ignore[arg-type]
+                        request.environ,
+                    )
+                except TypeError as e:
+                    raise TypeError(
+                        f"{e}\nThe view function did not return a valid"
+                        " response. The return type must be a string,"
+                        " dict, list, tuple with headers or status,"
+                        " Response instance, or WSGI callable, but it"
+                        f" was a {type(rv).__name__}."
+                    ).with_traceback(sys.exc_info()[2]) from None
+            else:
+                raise TypeError(
+                    "The view function did not return a valid"
+                    " response. The return type must be a string,"
+                    " dict, list, tuple with headers or status,"
+                    " Response instance, or WSGI callable, but it was a"
+                    f" {type(rv).__name__}."
+                )
+
+        rv = t.cast(Response, rv)
+        # prefer the status if it was provided
+        if status is not None:
+            if isinstance(status, (str, bytes, bytearray)):
+                rv.status = status
+            else:
+                rv.status_code = status
+
+        # extend existing headers with provided headers
+        if headers:
+            rv.headers.update(headers)
+
+        return rv
+
+    def preprocess_request(self) -> ft.ResponseReturnValue | None:
+        """Called before the request is dispatched. Calls
+        :attr:`url_value_preprocessors` registered with the app and the
+        current blueprint (if any). Then calls :attr:`before_request_funcs`
+        registered with the app and the blueprint.
+
+        If any :meth:`before_request` handler returns a non-None value, the
+        value is handled as if it was the return value from the view, and
+        further request handling is stopped.
+        """
+        names = (None, *reversed(request.blueprints))
+
+        for name in names:
+            if name in self.url_value_preprocessors:
+                for url_func in self.url_value_preprocessors[name]:
+                    url_func(request.endpoint, request.view_args)
+
+        for name in names:
+            if name in self.before_request_funcs:
+                for before_func in self.before_request_funcs[name]:
+                    rv = self.ensure_sync(before_func)()
+
+                    if rv is not None:
+                        return rv  # type: ignore[no-any-return]
+
+        return None
+
+    def process_response(self, response: Response) -> Response:
+        """Can be overridden in order to modify the response object
+        before it's sent to the WSGI server.  By default this will
+        call all the :meth:`after_request` decorated functions.
+
+        .. versionchanged:: 0.5
+           As of Flask 0.5 the functions registered for after request
+           execution are called in reverse order of registration.
+
+        :param response: a :attr:`response_class` object.
+        :return: a new response object or the same, has to be an
+                 instance of :attr:`response_class`.
+        """
+        ctx = request_ctx._get_current_object()  # type: ignore[attr-defined]
+
+        for func in ctx._after_request_functions:
+            response = self.ensure_sync(func)(response)
+
+        for name in chain(request.blueprints, (None,)):
+            if name in self.after_request_funcs:
+                for func in reversed(self.after_request_funcs[name]):
+                    response = self.ensure_sync(func)(response)
+
+        if not self.session_interface.is_null_session(ctx.session):
+            self.session_interface.save_session(self, ctx.session, response)
+
+        return response
+
+    def do_teardown_request(
+        self,
+        exc: BaseException | None = _sentinel,  # type: ignore[assignment]
+    ) -> None:
+        """Called after the request is dispatched and the response is
+        returned, right before the request context is popped.
+
+        This calls all functions decorated with
+        :meth:`teardown_request`, and :meth:`Blueprint.teardown_request`
+        if a blueprint handled the request. Finally, the
+        :data:`request_tearing_down` signal is sent.
+
+        This is called by
+        :meth:`RequestContext.pop() <flask.ctx.RequestContext.pop>`,
+        which may be delayed during testing to maintain access to
+        resources.
+
+        :param exc: An unhandled exception raised while dispatching the
+            request. Detected from the current exception information if
+            not passed. Passed to each teardown function.
+
+        .. versionchanged:: 0.9
+            Added the ``exc`` argument.
+        """
+        if exc is _sentinel:
+            exc = sys.exc_info()[1]
+
+        for name in chain(request.blueprints, (None,)):
+            if name in self.teardown_request_funcs:
+                for func in reversed(self.teardown_request_funcs[name]):
+                    self.ensure_sync(func)(exc)
+
+        request_tearing_down.send(self, _async_wrapper=self.ensure_sync, exc=exc)
+
+    def do_teardown_appcontext(
+        self,
+        exc: BaseException | None = _sentinel,  # type: ignore[assignment]
+    ) -> None:
+        """Called right before the application context is popped.
+
+        When handling a request, the application context is popped
+        after the request context. See :meth:`do_teardown_request`.
+
+        This calls all functions decorated with
+        :meth:`teardown_appcontext`. Then the
+        :data:`appcontext_tearing_down` signal is sent.
+
+        This is called by
+        :meth:`AppContext.pop() <flask.ctx.AppContext.pop>`.
+
+        .. versionadded:: 0.9
+        """
+        if exc is _sentinel:
+            exc = sys.exc_info()[1]
+
+        for func in reversed(self.teardown_appcontext_funcs):
+            self.ensure_sync(func)(exc)
+
+        appcontext_tearing_down.send(self, _async_wrapper=self.ensure_sync, exc=exc)
+
+    def app_context(self) -> AppContext:
+        """Create an :class:`~flask.ctx.AppContext`. Use as a ``with``
+        block to push the context, which will make :data:`current_app`
+        point at this application.
+
+        An application context is automatically pushed by
+        :meth:`RequestContext.push() <flask.ctx.RequestContext.push>`
+        when handling a request, and when running a CLI command. Use
+        this to manually create a context outside of these situations.
+
+        ::
+
+            with app.app_context():
+                init_db()
+
+        See :doc:`/appcontext`.
+
+        .. versionadded:: 0.9
+        """
+        return AppContext(self)
+
+    def request_context(self, environ: WSGIEnvironment) -> RequestContext:
+        """Create a :class:`~flask.ctx.RequestContext` representing a
+        WSGI environment. Use a ``with`` block to push the context,
+        which will make :data:`request` point at this request.
+
+        See :doc:`/reqcontext`.
+
+        Typically you should not call this from your own code. A request
+        context is automatically pushed by the :meth:`wsgi_app` when
+        handling a request. Use :meth:`test_request_context` to create
+        an environment and context instead of this method.
+
+        :param environ: a WSGI environment
+        """
+        return RequestContext(self, environ)
+
+    def test_request_context(self, *args: t.Any, **kwargs: t.Any) -> RequestContext:
+        """Create a :class:`~flask.ctx.RequestContext` for a WSGI
+        environment created from the given values. This is mostly useful
+        during testing, where you may want to run a function that uses
+        request data without dispatching a full request.
+
+        See :doc:`/reqcontext`.
+
+        Use a ``with`` block to push the context, which will make
+        :data:`request` point at the request for the created
+        environment. ::
+
+            with app.test_request_context(...):
+                generate_report()
+
+        When using the shell, it may be easier to push and pop the
+        context manually to avoid indentation. ::
+
+            ctx = app.test_request_context(...)
+            ctx.push()
+            ...
+            ctx.pop()
+
+        Takes the same arguments as Werkzeug's
+        :class:`~werkzeug.test.EnvironBuilder`, with some defaults from
+        the application. See the linked Werkzeug docs for most of the
+        available arguments. Flask-specific behavior is listed here.
+
+        :param path: URL path being requested.
+        :param base_url: Base URL where the app is being served, which
+            ``path`` is relative to. If not given, built from
+            :data:`PREFERRED_URL_SCHEME`, ``subdomain``,
+            :data:`SERVER_NAME`, and :data:`APPLICATION_ROOT`.
+        :param subdomain: Subdomain name to append to
+            :data:`SERVER_NAME`.
+        :param url_scheme: Scheme to use instead of
+            :data:`PREFERRED_URL_SCHEME`.
+        :param data: The request body, either as a string or a dict of
+            form keys and values.
+        :param json: If given, this is serialized as JSON and passed as
+            ``data``. Also defaults ``content_type`` to
+            ``application/json``.
+        :param args: other positional arguments passed to
+            :class:`~werkzeug.test.EnvironBuilder`.
+        :param kwargs: other keyword arguments passed to
+            :class:`~werkzeug.test.EnvironBuilder`.
+        """
+        from .testing import EnvironBuilder
+
+        builder = EnvironBuilder(self, *args, **kwargs)
+
+        try:
+            return self.request_context(builder.get_environ())
+        finally:
+            builder.close()
+
+    def wsgi_app(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> cabc.Iterable[bytes]:
+        """The actual WSGI application. This is not implemented in
+        :meth:`__call__` so that middlewares can be applied without
+        losing a reference to the app object. Instead of doing this::
+
+            app = MyMiddleware(app)
+
+        It's a better idea to do this instead::
+
+            app.wsgi_app = MyMiddleware(app.wsgi_app)
+
+        Then you still have the original application object around and
+        can continue to call methods on it.
+
+        .. versionchanged:: 0.7
+            Teardown events for the request and app contexts are called
+            even if an unhandled error occurs. Other events may not be
+            called depending on when an error occurs during dispatch.
+            See :ref:`callbacks-and-errors`.
+
+        :param environ: A WSGI environment.
+        :param start_response: A callable accepting a status code,
+            a list of headers, and an optional exception context to
+            start the response.
+        """
+        ctx = self.request_context(environ)
+        error: BaseException | None = None
+        try:
+            try:
+                ctx.push()
+                response = self.full_dispatch_request()
+            except Exception as e:
+                error = e
+                response = self.handle_exception(e)
+            except:  # noqa: B001
+                error = sys.exc_info()[1]
+                raise
+            return response(environ, start_response)
+        finally:
+            if "werkzeug.debug.preserve_context" in environ:
+                environ["werkzeug.debug.preserve_context"](_cv_app.get())
+                environ["werkzeug.debug.preserve_context"](_cv_request.get())
+
+            if error is not None and self.should_ignore_error(error):
+                error = None
+
+            ctx.pop(error)
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> cabc.Iterable[bytes]:
+        """The WSGI server calls the Flask application object as the
+        WSGI application. This calls :meth:`wsgi_app`, which can be
+        wrapped to apply middleware.
+        """
+        return self.wsgi_app(environ, start_response)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/blueprints.py b/backend/venv/lib/python3.12/site-packages/flask/blueprints.py
new file mode 100644
index 00000000..b6d4e433
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/blueprints.py
@@ -0,0 +1,128 @@
+from __future__ import annotations
+
+import os
+import typing as t
+from datetime import timedelta
+
+from .cli import AppGroup
+from .globals import current_app
+from .helpers import send_from_directory
+from .sansio.blueprints import Blueprint as SansioBlueprint
+from .sansio.blueprints import BlueprintSetupState as BlueprintSetupState  # noqa
+from .sansio.scaffold import _sentinel
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .wrappers import Response
+
+
+class Blueprint(SansioBlueprint):
+    def __init__(
+        self,
+        name: str,
+        import_name: str,
+        static_folder: str | os.PathLike[str] | None = None,
+        static_url_path: str | None = None,
+        template_folder: str | os.PathLike[str] | None = None,
+        url_prefix: str | None = None,
+        subdomain: str | None = None,
+        url_defaults: dict[str, t.Any] | None = None,
+        root_path: str | None = None,
+        cli_group: str | None = _sentinel,  # type: ignore
+    ) -> None:
+        super().__init__(
+            name,
+            import_name,
+            static_folder,
+            static_url_path,
+            template_folder,
+            url_prefix,
+            subdomain,
+            url_defaults,
+            root_path,
+            cli_group,
+        )
+
+        #: The Click command group for registering CLI commands for this
+        #: object. The commands are available from the ``flask`` command
+        #: once the application has been discovered and blueprints have
+        #: been registered.
+        self.cli = AppGroup()
+
+        # Set the name of the Click group in case someone wants to add
+        # the app's commands to another CLI tool.
+        self.cli.name = self.name
+
+    def get_send_file_max_age(self, filename: str | None) -> int | None:
+        """Used by :func:`send_file` to determine the ``max_age`` cache
+        value for a given file path if it wasn't passed.
+
+        By default, this returns :data:`SEND_FILE_MAX_AGE_DEFAULT` from
+        the configuration of :data:`~flask.current_app`. This defaults
+        to ``None``, which tells the browser to use conditional requests
+        instead of a timed cache, which is usually preferable.
+
+        Note this is a duplicate of the same method in the Flask
+        class.
+
+        .. versionchanged:: 2.0
+            The default configuration is ``None`` instead of 12 hours.
+
+        .. versionadded:: 0.9
+        """
+        value = current_app.config["SEND_FILE_MAX_AGE_DEFAULT"]
+
+        if value is None:
+            return None
+
+        if isinstance(value, timedelta):
+            return int(value.total_seconds())
+
+        return value  # type: ignore[no-any-return]
+
+    def send_static_file(self, filename: str) -> Response:
+        """The view function used to serve files from
+        :attr:`static_folder`. A route is automatically registered for
+        this view at :attr:`static_url_path` if :attr:`static_folder` is
+        set.
+
+        Note this is a duplicate of the same method in the Flask
+        class.
+
+        .. versionadded:: 0.5
+
+        """
+        if not self.has_static_folder:
+            raise RuntimeError("'static_folder' must be set to serve static_files.")
+
+        # send_file only knows to call get_send_file_max_age on the app,
+        # call it here so it works for blueprints too.
+        max_age = self.get_send_file_max_age(filename)
+        return send_from_directory(
+            t.cast(str, self.static_folder), filename, max_age=max_age
+        )
+
+    def open_resource(
+        self, resource: str, mode: str = "rb", encoding: str | None = "utf-8"
+    ) -> t.IO[t.AnyStr]:
+        """Open a resource file relative to :attr:`root_path` for reading. The
+        blueprint-relative equivalent of the app's :meth:`~.Flask.open_resource`
+        method.
+
+        :param resource: Path to the resource relative to :attr:`root_path`.
+        :param mode: Open the file in this mode. Only reading is supported,
+            valid values are ``"r"`` (or ``"rt"``) and ``"rb"``.
+        :param encoding: Open the file with this encoding when opening in text
+            mode. This is ignored when opening in binary mode.
+
+        .. versionchanged:: 3.1
+            Added the ``encoding`` parameter.
+        """
+        if mode not in {"r", "rt", "rb"}:
+            raise ValueError("Resources can only be opened for reading.")
+
+        path = os.path.join(self.root_path, resource)
+
+        if mode == "rb":
+            return open(path, mode)  # pyright: ignore
+
+        return open(path, mode, encoding=encoding)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/cli.py b/backend/venv/lib/python3.12/site-packages/flask/cli.py
new file mode 100644
index 00000000..dd03f3c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/cli.py
@@ -0,0 +1,1133 @@
+from __future__ import annotations
+
+import ast
+import collections.abc as cabc
+import importlib.metadata
+import inspect
+import os
+import platform
+import re
+import sys
+import traceback
+import typing as t
+from functools import update_wrapper
+from operator import itemgetter
+from types import ModuleType
+
+import click
+from click.core import ParameterSource
+from werkzeug import run_simple
+from werkzeug.serving import is_running_from_reloader
+from werkzeug.utils import import_string
+
+from .globals import current_app
+from .helpers import get_debug_flag
+from .helpers import get_load_dotenv
+
+if t.TYPE_CHECKING:
+    import ssl
+
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .app import Flask
+
+
+class NoAppException(click.UsageError):
+    """Raised if an application cannot be found or loaded."""
+
+
+def find_best_app(module: ModuleType) -> Flask:
+    """Given a module instance this tries to find the best possible
+    application in the module or raises an exception.
+    """
+    from . import Flask
+
+    # Search for the most common names first.
+    for attr_name in ("app", "application"):
+        app = getattr(module, attr_name, None)
+
+        if isinstance(app, Flask):
+            return app
+
+    # Otherwise find the only object that is a Flask instance.
+    matches = [v for v in module.__dict__.values() if isinstance(v, Flask)]
+
+    if len(matches) == 1:
+        return matches[0]
+    elif len(matches) > 1:
+        raise NoAppException(
+            "Detected multiple Flask applications in module"
+            f" '{module.__name__}'. Use '{module.__name__}:name'"
+            " to specify the correct one."
+        )
+
+    # Search for app factory functions.
+    for attr_name in ("create_app", "make_app"):
+        app_factory = getattr(module, attr_name, None)
+
+        if inspect.isfunction(app_factory):
+            try:
+                app = app_factory()
+
+                if isinstance(app, Flask):
+                    return app
+            except TypeError as e:
+                if not _called_with_wrong_args(app_factory):
+                    raise
+
+                raise NoAppException(
+                    f"Detected factory '{attr_name}' in module '{module.__name__}',"
+                    " but could not call it without arguments. Use"
+                    f" '{module.__name__}:{attr_name}(args)'"
+                    " to specify arguments."
+                ) from e
+
+    raise NoAppException(
+        "Failed to find Flask application or factory in module"
+        f" '{module.__name__}'. Use '{module.__name__}:name'"
+        " to specify one."
+    )
+
+
+def _called_with_wrong_args(f: t.Callable[..., Flask]) -> bool:
+    """Check whether calling a function raised a ``TypeError`` because
+    the call failed or because something in the factory raised the
+    error.
+
+    :param f: The function that was called.
+    :return: ``True`` if the call failed.
+    """
+    tb = sys.exc_info()[2]
+
+    try:
+        while tb is not None:
+            if tb.tb_frame.f_code is f.__code__:
+                # In the function, it was called successfully.
+                return False
+
+            tb = tb.tb_next
+
+        # Didn't reach the function.
+        return True
+    finally:
+        # Delete tb to break a circular reference.
+        # https://docs.python.org/2/library/sys.html#sys.exc_info
+        del tb
+
+
+def find_app_by_string(module: ModuleType, app_name: str) -> Flask:
+    """Check if the given string is a variable name or a function. Call
+    a function to get the app instance, or return the variable directly.
+    """
+    from . import Flask
+
+    # Parse app_name as a single expression to determine if it's a valid
+    # attribute name or function call.
+    try:
+        expr = ast.parse(app_name.strip(), mode="eval").body
+    except SyntaxError:
+        raise NoAppException(
+            f"Failed to parse {app_name!r} as an attribute name or function call."
+        ) from None
+
+    if isinstance(expr, ast.Name):
+        name = expr.id
+        args = []
+        kwargs = {}
+    elif isinstance(expr, ast.Call):
+        # Ensure the function name is an attribute name only.
+        if not isinstance(expr.func, ast.Name):
+            raise NoAppException(
+                f"Function reference must be a simple name: {app_name!r}."
+            )
+
+        name = expr.func.id
+
+        # Parse the positional and keyword arguments as literals.
+        try:
+            args = [ast.literal_eval(arg) for arg in expr.args]
+            kwargs = {
+                kw.arg: ast.literal_eval(kw.value)
+                for kw in expr.keywords
+                if kw.arg is not None
+            }
+        except ValueError:
+            # literal_eval gives cryptic error messages, show a generic
+            # message with the full expression instead.
+            raise NoAppException(
+                f"Failed to parse arguments as literal values: {app_name!r}."
+            ) from None
+    else:
+        raise NoAppException(
+            f"Failed to parse {app_name!r} as an attribute name or function call."
+        )
+
+    try:
+        attr = getattr(module, name)
+    except AttributeError as e:
+        raise NoAppException(
+            f"Failed to find attribute {name!r} in {module.__name__!r}."
+        ) from e
+
+    # If the attribute is a function, call it with any args and kwargs
+    # to get the real application.
+    if inspect.isfunction(attr):
+        try:
+            app = attr(*args, **kwargs)
+        except TypeError as e:
+            if not _called_with_wrong_args(attr):
+                raise
+
+            raise NoAppException(
+                f"The factory {app_name!r} in module"
+                f" {module.__name__!r} could not be called with the"
+                " specified arguments."
+            ) from e
+    else:
+        app = attr
+
+    if isinstance(app, Flask):
+        return app
+
+    raise NoAppException(
+        "A valid Flask application was not obtained from"
+        f" '{module.__name__}:{app_name}'."
+    )
+
+
+def prepare_import(path: str) -> str:
+    """Given a filename this will try to calculate the python path, add it
+    to the search path and return the actual module name that is expected.
+    """
+    path = os.path.realpath(path)
+
+    fname, ext = os.path.splitext(path)
+    if ext == ".py":
+        path = fname
+
+    if os.path.basename(path) == "__init__":
+        path = os.path.dirname(path)
+
+    module_name = []
+
+    # move up until outside package structure (no __init__.py)
+    while True:
+        path, name = os.path.split(path)
+        module_name.append(name)
+
+        if not os.path.exists(os.path.join(path, "__init__.py")):
+            break
+
+    if sys.path[0] != path:
+        sys.path.insert(0, path)
+
+    return ".".join(module_name[::-1])
+
+
+@t.overload
+def locate_app(
+    module_name: str, app_name: str | None, raise_if_not_found: t.Literal[True] = True
+) -> Flask: ...
+
+
+@t.overload
+def locate_app(
+    module_name: str, app_name: str | None, raise_if_not_found: t.Literal[False] = ...
+) -> Flask | None: ...
+
+
+def locate_app(
+    module_name: str, app_name: str | None, raise_if_not_found: bool = True
+) -> Flask | None:
+    try:
+        __import__(module_name)
+    except ImportError:
+        # Reraise the ImportError if it occurred within the imported module.
+        # Determine this by checking whether the trace has a depth > 1.
+        if sys.exc_info()[2].tb_next:  # type: ignore[union-attr]
+            raise NoAppException(
+                f"While importing {module_name!r}, an ImportError was"
+                f" raised:\n\n{traceback.format_exc()}"
+            ) from None
+        elif raise_if_not_found:
+            raise NoAppException(f"Could not import {module_name!r}.") from None
+        else:
+            return None
+
+    module = sys.modules[module_name]
+
+    if app_name is None:
+        return find_best_app(module)
+    else:
+        return find_app_by_string(module, app_name)
+
+
+def get_version(ctx: click.Context, param: click.Parameter, value: t.Any) -> None:
+    if not value or ctx.resilient_parsing:
+        return
+
+    flask_version = importlib.metadata.version("flask")
+    werkzeug_version = importlib.metadata.version("werkzeug")
+
+    click.echo(
+        f"Python {platform.python_version()}\n"
+        f"Flask {flask_version}\n"
+        f"Werkzeug {werkzeug_version}",
+        color=ctx.color,
+    )
+    ctx.exit()
+
+
+version_option = click.Option(
+    ["--version"],
+    help="Show the Flask version.",
+    expose_value=False,
+    callback=get_version,
+    is_flag=True,
+    is_eager=True,
+)
+
+
+class ScriptInfo:
+    """Helper object to deal with Flask applications.  This is usually not
+    necessary to interface with as it's used internally in the dispatching
+    to click.  In future versions of Flask this object will most likely play
+    a bigger role.  Typically it's created automatically by the
+    :class:`FlaskGroup` but you can also manually create it and pass it
+    onwards as click object.
+
+    .. versionchanged:: 3.1
+        Added the ``load_dotenv_defaults`` parameter and attribute.
+    """
+
+    def __init__(
+        self,
+        app_import_path: str | None = None,
+        create_app: t.Callable[..., Flask] | None = None,
+        set_debug_flag: bool = True,
+        load_dotenv_defaults: bool = True,
+    ) -> None:
+        #: Optionally the import path for the Flask application.
+        self.app_import_path = app_import_path
+        #: Optionally a function that is passed the script info to create
+        #: the instance of the application.
+        self.create_app = create_app
+        #: A dictionary with arbitrary data that can be associated with
+        #: this script info.
+        self.data: dict[t.Any, t.Any] = {}
+        self.set_debug_flag = set_debug_flag
+
+        self.load_dotenv_defaults = get_load_dotenv(load_dotenv_defaults)
+        """Whether default ``.flaskenv`` and ``.env`` files should be loaded.
+
+        ``ScriptInfo`` doesn't load anything, this is for reference when doing
+        the load elsewhere during processing.
+
+        .. versionadded:: 3.1
+        """
+
+        self._loaded_app: Flask | None = None
+
+    def load_app(self) -> Flask:
+        """Loads the Flask app (if not yet loaded) and returns it.  Calling
+        this multiple times will just result in the already loaded app to
+        be returned.
+        """
+        if self._loaded_app is not None:
+            return self._loaded_app
+        app: Flask | None = None
+        if self.create_app is not None:
+            app = self.create_app()
+        else:
+            if self.app_import_path:
+                path, name = (
+                    re.split(r":(?![\\/])", self.app_import_path, maxsplit=1) + [None]
+                )[:2]
+                import_name = prepare_import(path)
+                app = locate_app(import_name, name)
+            else:
+                for path in ("wsgi.py", "app.py"):
+                    import_name = prepare_import(path)
+                    app = locate_app(import_name, None, raise_if_not_found=False)
+
+                    if app is not None:
+                        break
+
+        if app is None:
+            raise NoAppException(
+                "Could not locate a Flask application. Use the"
+                " 'flask --app' option, 'FLASK_APP' environment"
+                " variable, or a 'wsgi.py' or 'app.py' file in the"
+                " current directory."
+            )
+
+        if self.set_debug_flag:
+            # Update the app's debug flag through the descriptor so that
+            # other values repopulate as well.
+            app.debug = get_debug_flag()
+
+        self._loaded_app = app
+        return app
+
+
+pass_script_info = click.make_pass_decorator(ScriptInfo, ensure=True)
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+
+def with_appcontext(f: F) -> F:
+    """Wraps a callback so that it's guaranteed to be executed with the
+    script's application context.
+
+    Custom commands (and their options) registered under ``app.cli`` or
+    ``blueprint.cli`` will always have an app context available, this
+    decorator is not required in that case.
+
+    .. versionchanged:: 2.2
+        The app context is active for subcommands as well as the
+        decorated callback. The app context is always available to
+        ``app.cli`` command and parameter callbacks.
+    """
+
+    @click.pass_context
+    def decorator(ctx: click.Context, /, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        if not current_app:
+            app = ctx.ensure_object(ScriptInfo).load_app()
+            ctx.with_resource(app.app_context())
+
+        return ctx.invoke(f, *args, **kwargs)
+
+    return update_wrapper(decorator, f)  # type: ignore[return-value]
+
+
+class AppGroup(click.Group):
+    """This works similar to a regular click :class:`~click.Group` but it
+    changes the behavior of the :meth:`command` decorator so that it
+    automatically wraps the functions in :func:`with_appcontext`.
+
+    Not to be confused with :class:`FlaskGroup`.
+    """
+
+    def command(  # type: ignore[override]
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Callable[[t.Callable[..., t.Any]], click.Command]:
+        """This works exactly like the method of the same name on a regular
+        :class:`click.Group` but it wraps callbacks in :func:`with_appcontext`
+        unless it's disabled by passing ``with_appcontext=False``.
+        """
+        wrap_for_ctx = kwargs.pop("with_appcontext", True)
+
+        def decorator(f: t.Callable[..., t.Any]) -> click.Command:
+            if wrap_for_ctx:
+                f = with_appcontext(f)
+            return super(AppGroup, self).command(*args, **kwargs)(f)  # type: ignore[no-any-return]
+
+        return decorator
+
+    def group(  # type: ignore[override]
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Callable[[t.Callable[..., t.Any]], click.Group]:
+        """This works exactly like the method of the same name on a regular
+        :class:`click.Group` but it defaults the group class to
+        :class:`AppGroup`.
+        """
+        kwargs.setdefault("cls", AppGroup)
+        return super().group(*args, **kwargs)  # type: ignore[no-any-return]
+
+
+def _set_app(ctx: click.Context, param: click.Option, value: str | None) -> str | None:
+    if value is None:
+        return None
+
+    info = ctx.ensure_object(ScriptInfo)
+    info.app_import_path = value
+    return value
+
+
+# This option is eager so the app will be available if --help is given.
+# --help is also eager, so --app must be before it in the param list.
+# no_args_is_help bypasses eager processing, so this option must be
+# processed manually in that case to ensure FLASK_APP gets picked up.
+_app_option = click.Option(
+    ["-A", "--app"],
+    metavar="IMPORT",
+    help=(
+        "The Flask application or factory function to load, in the form 'module:name'."
+        " Module can be a dotted import or file path. Name is not required if it is"
+        " 'app', 'application', 'create_app', or 'make_app', and can be 'name(args)' to"
+        " pass arguments."
+    ),
+    is_eager=True,
+    expose_value=False,
+    callback=_set_app,
+)
+
+
+def _set_debug(ctx: click.Context, param: click.Option, value: bool) -> bool | None:
+    # If the flag isn't provided, it will default to False. Don't use
+    # that, let debug be set by env in that case.
+    source = ctx.get_parameter_source(param.name)  # type: ignore[arg-type]
+
+    if source is not None and source in (
+        ParameterSource.DEFAULT,
+        ParameterSource.DEFAULT_MAP,
+    ):
+        return None
+
+    # Set with env var instead of ScriptInfo.load so that it can be
+    # accessed early during a factory function.
+    os.environ["FLASK_DEBUG"] = "1" if value else "0"
+    return value
+
+
+_debug_option = click.Option(
+    ["--debug/--no-debug"],
+    help="Set debug mode.",
+    expose_value=False,
+    callback=_set_debug,
+)
+
+
+def _env_file_callback(
+    ctx: click.Context, param: click.Option, value: str | None
+) -> str | None:
+    try:
+        import dotenv  # noqa: F401
+    except ImportError:
+        # Only show an error if a value was passed, otherwise we still want to
+        # call load_dotenv and show a message without exiting.
+        if value is not None:
+            raise click.BadParameter(
+                "python-dotenv must be installed to load an env file.",
+                ctx=ctx,
+                param=param,
+            ) from None
+
+    # Load if a value was passed, or we want to load default files, or both.
+    if value is not None or ctx.obj.load_dotenv_defaults:
+        load_dotenv(value, load_defaults=ctx.obj.load_dotenv_defaults)
+
+    return value
+
+
+# This option is eager so env vars are loaded as early as possible to be
+# used by other options.
+_env_file_option = click.Option(
+    ["-e", "--env-file"],
+    type=click.Path(exists=True, dir_okay=False),
+    help=(
+        "Load environment variables from this file, taking precedence over"
+        " those set by '.env' and '.flaskenv'. Variables set directly in the"
+        " environment take highest precedence. python-dotenv must be installed."
+    ),
+    is_eager=True,
+    expose_value=False,
+    callback=_env_file_callback,
+)
+
+
+class FlaskGroup(AppGroup):
+    """Special subclass of the :class:`AppGroup` group that supports
+    loading more commands from the configured Flask app.  Normally a
+    developer does not have to interface with this class but there are
+    some very advanced use cases for which it makes sense to create an
+    instance of this. see :ref:`custom-scripts`.
+
+    :param add_default_commands: if this is True then the default run and
+        shell commands will be added.
+    :param add_version_option: adds the ``--version`` option.
+    :param create_app: an optional callback that is passed the script info and
+        returns the loaded app.
+    :param load_dotenv: Load the nearest :file:`.env` and :file:`.flaskenv`
+        files to set environment variables. Will also change the working
+        directory to the directory containing the first file found.
+    :param set_debug_flag: Set the app's debug flag.
+
+    .. versionchanged:: 3.1
+        ``-e path`` takes precedence over default ``.env`` and ``.flaskenv`` files.
+
+    .. versionchanged:: 2.2
+        Added the ``-A/--app``, ``--debug/--no-debug``, ``-e/--env-file`` options.
+
+    .. versionchanged:: 2.2
+        An app context is pushed when running ``app.cli`` commands, so
+        ``@with_appcontext`` is no longer required for those commands.
+
+    .. versionchanged:: 1.0
+        If installed, python-dotenv will be used to load environment variables
+        from :file:`.env` and :file:`.flaskenv` files.
+    """
+
+    def __init__(
+        self,
+        add_default_commands: bool = True,
+        create_app: t.Callable[..., Flask] | None = None,
+        add_version_option: bool = True,
+        load_dotenv: bool = True,
+        set_debug_flag: bool = True,
+        **extra: t.Any,
+    ) -> None:
+        params: list[click.Parameter] = list(extra.pop("params", None) or ())
+        # Processing is done with option callbacks instead of a group
+        # callback. This allows users to make a custom group callback
+        # without losing the behavior. --env-file must come first so
+        # that it is eagerly evaluated before --app.
+        params.extend((_env_file_option, _app_option, _debug_option))
+
+        if add_version_option:
+            params.append(version_option)
+
+        if "context_settings" not in extra:
+            extra["context_settings"] = {}
+
+        extra["context_settings"].setdefault("auto_envvar_prefix", "FLASK")
+
+        super().__init__(params=params, **extra)
+
+        self.create_app = create_app
+        self.load_dotenv = load_dotenv
+        self.set_debug_flag = set_debug_flag
+
+        if add_default_commands:
+            self.add_command(run_command)
+            self.add_command(shell_command)
+            self.add_command(routes_command)
+
+        self._loaded_plugin_commands = False
+
+    def _load_plugin_commands(self) -> None:
+        if self._loaded_plugin_commands:
+            return
+
+        if sys.version_info >= (3, 10):
+            from importlib import metadata
+        else:
+            # Use a backport on Python < 3.10. We technically have
+            # importlib.metadata on 3.8+, but the API changed in 3.10,
+            # so use the backport for consistency.
+            import importlib_metadata as metadata  # pyright: ignore
+
+        for ep in metadata.entry_points(group="flask.commands"):
+            self.add_command(ep.load(), ep.name)
+
+        self._loaded_plugin_commands = True
+
+    def get_command(self, ctx: click.Context, name: str) -> click.Command | None:
+        self._load_plugin_commands()
+        # Look up built-in and plugin commands, which should be
+        # available even if the app fails to load.
+        rv = super().get_command(ctx, name)
+
+        if rv is not None:
+            return rv
+
+        info = ctx.ensure_object(ScriptInfo)
+
+        # Look up commands provided by the app, showing an error and
+        # continuing if the app couldn't be loaded.
+        try:
+            app = info.load_app()
+        except NoAppException as e:
+            click.secho(f"Error: {e.format_message()}\n", err=True, fg="red")
+            return None
+
+        # Push an app context for the loaded app unless it is already
+        # active somehow. This makes the context available to parameter
+        # and command callbacks without needing @with_appcontext.
+        if not current_app or current_app._get_current_object() is not app:  # type: ignore[attr-defined]
+            ctx.with_resource(app.app_context())
+
+        return app.cli.get_command(ctx, name)
+
+    def list_commands(self, ctx: click.Context) -> list[str]:
+        self._load_plugin_commands()
+        # Start with the built-in and plugin commands.
+        rv = set(super().list_commands(ctx))
+        info = ctx.ensure_object(ScriptInfo)
+
+        # Add commands provided by the app, showing an error and
+        # continuing if the app couldn't be loaded.
+        try:
+            rv.update(info.load_app().cli.list_commands(ctx))
+        except NoAppException as e:
+            # When an app couldn't be loaded, show the error message
+            # without the traceback.
+            click.secho(f"Error: {e.format_message()}\n", err=True, fg="red")
+        except Exception:
+            # When any other errors occurred during loading, show the
+            # full traceback.
+            click.secho(f"{traceback.format_exc()}\n", err=True, fg="red")
+
+        return sorted(rv)
+
+    def make_context(
+        self,
+        info_name: str | None,
+        args: list[str],
+        parent: click.Context | None = None,
+        **extra: t.Any,
+    ) -> click.Context:
+        # Set a flag to tell app.run to become a no-op. If app.run was
+        # not in a __name__ == __main__ guard, it would start the server
+        # when importing, blocking whatever command is being called.
+        os.environ["FLASK_RUN_FROM_CLI"] = "true"
+
+        if "obj" not in extra and "obj" not in self.context_settings:
+            extra["obj"] = ScriptInfo(
+                create_app=self.create_app,
+                set_debug_flag=self.set_debug_flag,
+                load_dotenv_defaults=self.load_dotenv,
+            )
+
+        return super().make_context(info_name, args, parent=parent, **extra)
+
+    def parse_args(self, ctx: click.Context, args: list[str]) -> list[str]:
+        if not args and self.no_args_is_help:
+            # Attempt to load --env-file and --app early in case they
+            # were given as env vars. Otherwise no_args_is_help will not
+            # see commands from app.cli.
+            _env_file_option.handle_parse_result(ctx, {}, [])
+            _app_option.handle_parse_result(ctx, {}, [])
+
+        return super().parse_args(ctx, args)
+
+
+def _path_is_ancestor(path: str, other: str) -> bool:
+    """Take ``other`` and remove the length of ``path`` from it. Then join it
+    to ``path``. If it is the original value, ``path`` is an ancestor of
+    ``other``."""
+    return os.path.join(path, other[len(path) :].lstrip(os.sep)) == other
+
+
+def load_dotenv(
+    path: str | os.PathLike[str] | None = None, load_defaults: bool = True
+) -> bool:
+    """Load "dotenv" files to set environment variables. A given path takes
+    precedence over ``.env``, which takes precedence over ``.flaskenv``. After
+    loading and combining these files, values are only set if the key is not
+    already set in ``os.environ``.
+
+    This is a no-op if `python-dotenv`_ is not installed.
+
+    .. _python-dotenv: https://github.com/theskumar/python-dotenv#readme
+
+    :param path: Load the file at this location.
+    :param load_defaults: Search for and load the default ``.flaskenv`` and
+        ``.env`` files.
+    :return: ``True`` if at least one env var was loaded.
+
+    .. versionchanged:: 3.1
+        Added the ``load_defaults`` parameter. A given path takes precedence
+        over default files.
+
+    .. versionchanged:: 2.0
+        The current directory is not changed to the location of the
+        loaded file.
+
+    .. versionchanged:: 2.0
+        When loading the env files, set the default encoding to UTF-8.
+
+    .. versionchanged:: 1.1.0
+        Returns ``False`` when python-dotenv is not installed, or when
+        the given path isn't a file.
+
+    .. versionadded:: 1.0
+    """
+    try:
+        import dotenv
+    except ImportError:
+        if path or os.path.isfile(".env") or os.path.isfile(".flaskenv"):
+            click.secho(
+                " * Tip: There are .env files present. Install python-dotenv"
+                " to use them.",
+                fg="yellow",
+                err=True,
+            )
+
+        return False
+
+    data: dict[str, str | None] = {}
+
+    if load_defaults:
+        for default_name in (".flaskenv", ".env"):
+            if not (default_path := dotenv.find_dotenv(default_name, usecwd=True)):
+                continue
+
+            data |= dotenv.dotenv_values(default_path, encoding="utf-8")
+
+    if path is not None and os.path.isfile(path):
+        data |= dotenv.dotenv_values(path, encoding="utf-8")
+
+    for key, value in data.items():
+        if key in os.environ or value is None:
+            continue
+
+        os.environ[key] = value
+
+    return bool(data)  # True if at least one env var was loaded.
+
+
+def show_server_banner(debug: bool, app_import_path: str | None) -> None:
+    """Show extra startup messages the first time the server is run,
+    ignoring the reloader.
+    """
+    if is_running_from_reloader():
+        return
+
+    if app_import_path is not None:
+        click.echo(f" * Serving Flask app '{app_import_path}'")
+
+    if debug is not None:
+        click.echo(f" * Debug mode: {'on' if debug else 'off'}")
+
+
+class CertParamType(click.ParamType):
+    """Click option type for the ``--cert`` option. Allows either an
+    existing file, the string ``'adhoc'``, or an import for a
+    :class:`~ssl.SSLContext` object.
+    """
+
+    name = "path"
+
+    def __init__(self) -> None:
+        self.path_type = click.Path(exists=True, dir_okay=False, resolve_path=True)
+
+    def convert(
+        self, value: t.Any, param: click.Parameter | None, ctx: click.Context | None
+    ) -> t.Any:
+        try:
+            import ssl
+        except ImportError:
+            raise click.BadParameter(
+                'Using "--cert" requires Python to be compiled with SSL support.',
+                ctx,
+                param,
+            ) from None
+
+        try:
+            return self.path_type(value, param, ctx)
+        except click.BadParameter:
+            value = click.STRING(value, param, ctx).lower()
+
+            if value == "adhoc":
+                try:
+                    import cryptography  # noqa: F401
+                except ImportError:
+                    raise click.BadParameter(
+                        "Using ad-hoc certificates requires the cryptography library.",
+                        ctx,
+                        param,
+                    ) from None
+
+                return value
+
+            obj = import_string(value, silent=True)
+
+            if isinstance(obj, ssl.SSLContext):
+                return obj
+
+            raise
+
+
+def _validate_key(ctx: click.Context, param: click.Parameter, value: t.Any) -> t.Any:
+    """The ``--key`` option must be specified when ``--cert`` is a file.
+    Modifies the ``cert`` param to be a ``(cert, key)`` pair if needed.
+    """
+    cert = ctx.params.get("cert")
+    is_adhoc = cert == "adhoc"
+
+    try:
+        import ssl
+    except ImportError:
+        is_context = False
+    else:
+        is_context = isinstance(cert, ssl.SSLContext)
+
+    if value is not None:
+        if is_adhoc:
+            raise click.BadParameter(
+                'When "--cert" is "adhoc", "--key" is not used.', ctx, param
+            )
+
+        if is_context:
+            raise click.BadParameter(
+                'When "--cert" is an SSLContext object, "--key" is not used.',
+                ctx,
+                param,
+            )
+
+        if not cert:
+            raise click.BadParameter('"--cert" must also be specified.', ctx, param)
+
+        ctx.params["cert"] = cert, value
+
+    else:
+        if cert and not (is_adhoc or is_context):
+            raise click.BadParameter('Required when using "--cert".', ctx, param)
+
+    return value
+
+
+class SeparatedPathType(click.Path):
+    """Click option type that accepts a list of values separated by the
+    OS's path separator (``:``, ``;`` on Windows). Each value is
+    validated as a :class:`click.Path` type.
+    """
+
+    def convert(
+        self, value: t.Any, param: click.Parameter | None, ctx: click.Context | None
+    ) -> t.Any:
+        items = self.split_envvar_value(value)
+        # can't call no-arg super() inside list comprehension until Python 3.12
+        super_convert = super().convert
+        return [super_convert(item, param, ctx) for item in items]
+
+
+@click.command("run", short_help="Run a development server.")
+@click.option("--host", "-h", default="127.0.0.1", help="The interface to bind to.")
+@click.option("--port", "-p", default=5000, help="The port to bind to.")
+@click.option(
+    "--cert",
+    type=CertParamType(),
+    help="Specify a certificate file to use HTTPS.",
+    is_eager=True,
+)
+@click.option(
+    "--key",
+    type=click.Path(exists=True, dir_okay=False, resolve_path=True),
+    callback=_validate_key,
+    expose_value=False,
+    help="The key file to use when specifying a certificate.",
+)
+@click.option(
+    "--reload/--no-reload",
+    default=None,
+    help="Enable or disable the reloader. By default the reloader "
+    "is active if debug is enabled.",
+)
+@click.option(
+    "--debugger/--no-debugger",
+    default=None,
+    help="Enable or disable the debugger. By default the debugger "
+    "is active if debug is enabled.",
+)
+@click.option(
+    "--with-threads/--without-threads",
+    default=True,
+    help="Enable or disable multithreading.",
+)
+@click.option(
+    "--extra-files",
+    default=None,
+    type=SeparatedPathType(),
+    help=(
+        "Extra files that trigger a reload on change. Multiple paths"
+        f" are separated by {os.path.pathsep!r}."
+    ),
+)
+@click.option(
+    "--exclude-patterns",
+    default=None,
+    type=SeparatedPathType(),
+    help=(
+        "Files matching these fnmatch patterns will not trigger a reload"
+        " on change. Multiple patterns are separated by"
+        f" {os.path.pathsep!r}."
+    ),
+)
+@pass_script_info
+def run_command(
+    info: ScriptInfo,
+    host: str,
+    port: int,
+    reload: bool,
+    debugger: bool,
+    with_threads: bool,
+    cert: ssl.SSLContext | tuple[str, str | None] | t.Literal["adhoc"] | None,
+    extra_files: list[str] | None,
+    exclude_patterns: list[str] | None,
+) -> None:
+    """Run a local development server.
+
+    This server is for development purposes only. It does not provide
+    the stability, security, or performance of production WSGI servers.
+
+    The reloader and debugger are enabled by default with the '--debug'
+    option.
+    """
+    try:
+        app: WSGIApplication = info.load_app()  # pyright: ignore
+    except Exception as e:
+        if is_running_from_reloader():
+            # When reloading, print out the error immediately, but raise
+            # it later so the debugger or server can handle it.
+            traceback.print_exc()
+            err = e
+
+            def app(
+                environ: WSGIEnvironment, start_response: StartResponse
+            ) -> cabc.Iterable[bytes]:
+                raise err from None
+
+        else:
+            # When not reloading, raise the error immediately so the
+            # command fails.
+            raise e from None
+
+    debug = get_debug_flag()
+
+    if reload is None:
+        reload = debug
+
+    if debugger is None:
+        debugger = debug
+
+    show_server_banner(debug, info.app_import_path)
+
+    run_simple(
+        host,
+        port,
+        app,
+        use_reloader=reload,
+        use_debugger=debugger,
+        threaded=with_threads,
+        ssl_context=cert,
+        extra_files=extra_files,
+        exclude_patterns=exclude_patterns,
+    )
+
+
+run_command.params.insert(0, _debug_option)
+
+
+@click.command("shell", short_help="Run a shell in the app context.")
+@with_appcontext
+def shell_command() -> None:
+    """Run an interactive Python shell in the context of a given
+    Flask application.  The application will populate the default
+    namespace of this shell according to its configuration.
+
+    This is useful for executing small snippets of management code
+    without having to manually configure the application.
+    """
+    import code
+
+    banner = (
+        f"Python {sys.version} on {sys.platform}\n"
+        f"App: {current_app.import_name}\n"
+        f"Instance: {current_app.instance_path}"
+    )
+    ctx: dict[str, t.Any] = {}
+
+    # Support the regular Python interpreter startup script if someone
+    # is using it.
+    startup = os.environ.get("PYTHONSTARTUP")
+    if startup and os.path.isfile(startup):
+        with open(startup) as f:
+            eval(compile(f.read(), startup, "exec"), ctx)
+
+    ctx.update(current_app.make_shell_context())
+
+    # Site, customize, or startup script can set a hook to call when
+    # entering interactive mode. The default one sets up readline with
+    # tab and history completion.
+    interactive_hook = getattr(sys, "__interactivehook__", None)
+
+    if interactive_hook is not None:
+        try:
+            import readline
+            from rlcompleter import Completer
+        except ImportError:
+            pass
+        else:
+            # rlcompleter uses __main__.__dict__ by default, which is
+            # flask.__main__. Use the shell context instead.
+            readline.set_completer(Completer(ctx).complete)
+
+        interactive_hook()
+
+    code.interact(banner=banner, local=ctx)
+
+
+@click.command("routes", short_help="Show the routes for the app.")
+@click.option(
+    "--sort",
+    "-s",
+    type=click.Choice(("endpoint", "methods", "domain", "rule", "match")),
+    default="endpoint",
+    help=(
+        "Method to sort routes by. 'match' is the order that Flask will match routes"
+        " when dispatching a request."
+    ),
+)
+@click.option("--all-methods", is_flag=True, help="Show HEAD and OPTIONS methods.")
+@with_appcontext
+def routes_command(sort: str, all_methods: bool) -> None:
+    """Show all registered routes with endpoints and methods."""
+    rules = list(current_app.url_map.iter_rules())
+
+    if not rules:
+        click.echo("No routes were registered.")
+        return
+
+    ignored_methods = set() if all_methods else {"HEAD", "OPTIONS"}
+    host_matching = current_app.url_map.host_matching
+    has_domain = any(rule.host if host_matching else rule.subdomain for rule in rules)
+    rows = []
+
+    for rule in rules:
+        row = [
+            rule.endpoint,
+            ", ".join(sorted((rule.methods or set()) - ignored_methods)),
+        ]
+
+        if has_domain:
+            row.append((rule.host if host_matching else rule.subdomain) or "")
+
+        row.append(rule.rule)
+        rows.append(row)
+
+    headers = ["Endpoint", "Methods"]
+    sorts = ["endpoint", "methods"]
+
+    if has_domain:
+        headers.append("Host" if host_matching else "Subdomain")
+        sorts.append("domain")
+
+    headers.append("Rule")
+    sorts.append("rule")
+
+    try:
+        rows.sort(key=itemgetter(sorts.index(sort)))
+    except ValueError:
+        pass
+
+    rows.insert(0, headers)
+    widths = [max(len(row[i]) for row in rows) for i in range(len(headers))]
+    rows.insert(1, ["-" * w for w in widths])
+    template = "  ".join(f"{{{i}:<{w}}}" for i, w in enumerate(widths))
+
+    for row in rows:
+        click.echo(template.format(*row))
+
+
+cli = FlaskGroup(
+    name="flask",
+    help="""\
+A general utility script for Flask applications.
+
+An application to load must be given with the '--app' option,
+'FLASK_APP' environment variable, or with a 'wsgi.py' or 'app.py' file
+in the current directory.
+""",
+)
+
+
+def main() -> None:
+    cli.main()
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/flask/config.py b/backend/venv/lib/python3.12/site-packages/flask/config.py
new file mode 100644
index 00000000..34ef1a57
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/config.py
@@ -0,0 +1,367 @@
+from __future__ import annotations
+
+import errno
+import json
+import os
+import types
+import typing as t
+
+from werkzeug.utils import import_string
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .sansio.app import App
+
+
+T = t.TypeVar("T")
+
+
+class ConfigAttribute(t.Generic[T]):
+    """Makes an attribute forward to the config"""
+
+    def __init__(
+        self, name: str, get_converter: t.Callable[[t.Any], T] | None = None
+    ) -> None:
+        self.__name__ = name
+        self.get_converter = get_converter
+
+    @t.overload
+    def __get__(self, obj: None, owner: None) -> te.Self: ...
+
+    @t.overload
+    def __get__(self, obj: App, owner: type[App]) -> T: ...
+
+    def __get__(self, obj: App | None, owner: type[App] | None = None) -> T | te.Self:
+        if obj is None:
+            return self
+
+        rv = obj.config[self.__name__]
+
+        if self.get_converter is not None:
+            rv = self.get_converter(rv)
+
+        return rv  # type: ignore[no-any-return]
+
+    def __set__(self, obj: App, value: t.Any) -> None:
+        obj.config[self.__name__] = value
+
+
+class Config(dict):  # type: ignore[type-arg]
+    """Works exactly like a dict but provides ways to fill it from files
+    or special dictionaries.  There are two common patterns to populate the
+    config.
+
+    Either you can fill the config from a config file::
+
+        app.config.from_pyfile('yourconfig.cfg')
+
+    Or alternatively you can define the configuration options in the
+    module that calls :meth:`from_object` or provide an import path to
+    a module that should be loaded.  It is also possible to tell it to
+    use the same module and with that provide the configuration values
+    just before the call::
+
+        DEBUG = True
+        SECRET_KEY = 'development key'
+        app.config.from_object(__name__)
+
+    In both cases (loading from any Python file or loading from modules),
+    only uppercase keys are added to the config.  This makes it possible to use
+    lowercase values in the config file for temporary values that are not added
+    to the config or to define the config keys in the same file that implements
+    the application.
+
+    Probably the most interesting way to load configurations is from an
+    environment variable pointing to a file::
+
+        app.config.from_envvar('YOURAPPLICATION_SETTINGS')
+
+    In this case before launching the application you have to set this
+    environment variable to the file you want to use.  On Linux and OS X
+    use the export statement::
+
+        export YOURAPPLICATION_SETTINGS='/path/to/config/file'
+
+    On windows use `set` instead.
+
+    :param root_path: path to which files are read relative from.  When the
+                      config object is created by the application, this is
+                      the application's :attr:`~flask.Flask.root_path`.
+    :param defaults: an optional dictionary of default values
+    """
+
+    def __init__(
+        self,
+        root_path: str | os.PathLike[str],
+        defaults: dict[str, t.Any] | None = None,
+    ) -> None:
+        super().__init__(defaults or {})
+        self.root_path = root_path
+
+    def from_envvar(self, variable_name: str, silent: bool = False) -> bool:
+        """Loads a configuration from an environment variable pointing to
+        a configuration file.  This is basically just a shortcut with nicer
+        error messages for this line of code::
+
+            app.config.from_pyfile(os.environ['YOURAPPLICATION_SETTINGS'])
+
+        :param variable_name: name of the environment variable
+        :param silent: set to ``True`` if you want silent failure for missing
+                       files.
+        :return: ``True`` if the file was loaded successfully.
+        """
+        rv = os.environ.get(variable_name)
+        if not rv:
+            if silent:
+                return False
+            raise RuntimeError(
+                f"The environment variable {variable_name!r} is not set"
+                " and as such configuration could not be loaded. Set"
+                " this variable and make it point to a configuration"
+                " file"
+            )
+        return self.from_pyfile(rv, silent=silent)
+
+    def from_prefixed_env(
+        self, prefix: str = "FLASK", *, loads: t.Callable[[str], t.Any] = json.loads
+    ) -> bool:
+        """Load any environment variables that start with ``FLASK_``,
+        dropping the prefix from the env key for the config key. Values
+        are passed through a loading function to attempt to convert them
+        to more specific types than strings.
+
+        Keys are loaded in :func:`sorted` order.
+
+        The default loading function attempts to parse values as any
+        valid JSON type, including dicts and lists.
+
+        Specific items in nested dicts can be set by separating the
+        keys with double underscores (``__``). If an intermediate key
+        doesn't exist, it will be initialized to an empty dict.
+
+        :param prefix: Load env vars that start with this prefix,
+            separated with an underscore (``_``).
+        :param loads: Pass each string value to this function and use
+            the returned value as the config value. If any error is
+            raised it is ignored and the value remains a string. The
+            default is :func:`json.loads`.
+
+        .. versionadded:: 2.1
+        """
+        prefix = f"{prefix}_"
+
+        for key in sorted(os.environ):
+            if not key.startswith(prefix):
+                continue
+
+            value = os.environ[key]
+            key = key.removeprefix(prefix)
+
+            try:
+                value = loads(value)
+            except Exception:
+                # Keep the value as a string if loading failed.
+                pass
+
+            if "__" not in key:
+                # A non-nested key, set directly.
+                self[key] = value
+                continue
+
+            # Traverse nested dictionaries with keys separated by "__".
+            current = self
+            *parts, tail = key.split("__")
+
+            for part in parts:
+                # If an intermediate dict does not exist, create it.
+                if part not in current:
+                    current[part] = {}
+
+                current = current[part]
+
+            current[tail] = value
+
+        return True
+
+    def from_pyfile(
+        self, filename: str | os.PathLike[str], silent: bool = False
+    ) -> bool:
+        """Updates the values in the config from a Python file.  This function
+        behaves as if the file was imported as module with the
+        :meth:`from_object` function.
+
+        :param filename: the filename of the config.  This can either be an
+                         absolute filename or a filename relative to the
+                         root path.
+        :param silent: set to ``True`` if you want silent failure for missing
+                       files.
+        :return: ``True`` if the file was loaded successfully.
+
+        .. versionadded:: 0.7
+           `silent` parameter.
+        """
+        filename = os.path.join(self.root_path, filename)
+        d = types.ModuleType("config")
+        d.__file__ = filename
+        try:
+            with open(filename, mode="rb") as config_file:
+                exec(compile(config_file.read(), filename, "exec"), d.__dict__)
+        except OSError as e:
+            if silent and e.errno in (errno.ENOENT, errno.EISDIR, errno.ENOTDIR):
+                return False
+            e.strerror = f"Unable to load configuration file ({e.strerror})"
+            raise
+        self.from_object(d)
+        return True
+
+    def from_object(self, obj: object | str) -> None:
+        """Updates the values from the given object.  An object can be of one
+        of the following two types:
+
+        -   a string: in this case the object with that name will be imported
+        -   an actual object reference: that object is used directly
+
+        Objects are usually either modules or classes. :meth:`from_object`
+        loads only the uppercase attributes of the module/class. A ``dict``
+        object will not work with :meth:`from_object` because the keys of a
+        ``dict`` are not attributes of the ``dict`` class.
+
+        Example of module-based configuration::
+
+            app.config.from_object('yourapplication.default_config')
+            from yourapplication import default_config
+            app.config.from_object(default_config)
+
+        Nothing is done to the object before loading. If the object is a
+        class and has ``@property`` attributes, it needs to be
+        instantiated before being passed to this method.
+
+        You should not use this function to load the actual configuration but
+        rather configuration defaults.  The actual config should be loaded
+        with :meth:`from_pyfile` and ideally from a location not within the
+        package because the package might be installed system wide.
+
+        See :ref:`config-dev-prod` for an example of class-based configuration
+        using :meth:`from_object`.
+
+        :param obj: an import name or object
+        """
+        if isinstance(obj, str):
+            obj = import_string(obj)
+        for key in dir(obj):
+            if key.isupper():
+                self[key] = getattr(obj, key)
+
+    def from_file(
+        self,
+        filename: str | os.PathLike[str],
+        load: t.Callable[[t.IO[t.Any]], t.Mapping[str, t.Any]],
+        silent: bool = False,
+        text: bool = True,
+    ) -> bool:
+        """Update the values in the config from a file that is loaded
+        using the ``load`` parameter. The loaded data is passed to the
+        :meth:`from_mapping` method.
+
+        .. code-block:: python
+
+            import json
+            app.config.from_file("config.json", load=json.load)
+
+            import tomllib
+            app.config.from_file("config.toml", load=tomllib.load, text=False)
+
+        :param filename: The path to the data file. This can be an
+            absolute path or relative to the config root path.
+        :param load: A callable that takes a file handle and returns a
+            mapping of loaded data from the file.
+        :type load: ``Callable[[Reader], Mapping]`` where ``Reader``
+            implements a ``read`` method.
+        :param silent: Ignore the file if it doesn't exist.
+        :param text: Open the file in text or binary mode.
+        :return: ``True`` if the file was loaded successfully.
+
+        .. versionchanged:: 2.3
+            The ``text`` parameter was added.
+
+        .. versionadded:: 2.0
+        """
+        filename = os.path.join(self.root_path, filename)
+
+        try:
+            with open(filename, "r" if text else "rb") as f:
+                obj = load(f)
+        except OSError as e:
+            if silent and e.errno in (errno.ENOENT, errno.EISDIR):
+                return False
+
+            e.strerror = f"Unable to load configuration file ({e.strerror})"
+            raise
+
+        return self.from_mapping(obj)
+
+    def from_mapping(
+        self, mapping: t.Mapping[str, t.Any] | None = None, **kwargs: t.Any
+    ) -> bool:
+        """Updates the config like :meth:`update` ignoring items with
+        non-upper keys.
+
+        :return: Always returns ``True``.
+
+        .. versionadded:: 0.11
+        """
+        mappings: dict[str, t.Any] = {}
+        if mapping is not None:
+            mappings.update(mapping)
+        mappings.update(kwargs)
+        for key, value in mappings.items():
+            if key.isupper():
+                self[key] = value
+        return True
+
+    def get_namespace(
+        self, namespace: str, lowercase: bool = True, trim_namespace: bool = True
+    ) -> dict[str, t.Any]:
+        """Returns a dictionary containing a subset of configuration options
+        that match the specified namespace/prefix. Example usage::
+
+            app.config['IMAGE_STORE_TYPE'] = 'fs'
+            app.config['IMAGE_STORE_PATH'] = '/var/app/images'
+            app.config['IMAGE_STORE_BASE_URL'] = 'http://img.website.com'
+            image_store_config = app.config.get_namespace('IMAGE_STORE_')
+
+        The resulting dictionary `image_store_config` would look like::
+
+            {
+                'type': 'fs',
+                'path': '/var/app/images',
+                'base_url': 'http://img.website.com'
+            }
+
+        This is often useful when configuration options map directly to
+        keyword arguments in functions or class constructors.
+
+        :param namespace: a configuration namespace
+        :param lowercase: a flag indicating if the keys of the resulting
+                          dictionary should be lowercase
+        :param trim_namespace: a flag indicating if the keys of the resulting
+                          dictionary should not include the namespace
+
+        .. versionadded:: 0.11
+        """
+        rv = {}
+        for k, v in self.items():
+            if not k.startswith(namespace):
+                continue
+            if trim_namespace:
+                key = k[len(namespace) :]
+            else:
+                key = k
+            if lowercase:
+                key = key.lower()
+            rv[key] = v
+        return rv
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {dict.__repr__(self)}>"
diff --git a/backend/venv/lib/python3.12/site-packages/flask/ctx.py b/backend/venv/lib/python3.12/site-packages/flask/ctx.py
new file mode 100644
index 00000000..9b164d39
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/ctx.py
@@ -0,0 +1,449 @@
+from __future__ import annotations
+
+import contextvars
+import sys
+import typing as t
+from functools import update_wrapper
+from types import TracebackType
+
+from werkzeug.exceptions import HTTPException
+
+from . import typing as ft
+from .globals import _cv_app
+from .globals import _cv_request
+from .signals import appcontext_popped
+from .signals import appcontext_pushed
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .app import Flask
+    from .sessions import SessionMixin
+    from .wrappers import Request
+
+
+# a singleton sentinel value for parameter defaults
+_sentinel = object()
+
+
+class _AppCtxGlobals:
+    """A plain object. Used as a namespace for storing data during an
+    application context.
+
+    Creating an app context automatically creates this object, which is
+    made available as the :data:`g` proxy.
+
+    .. describe:: 'key' in g
+
+        Check whether an attribute is present.
+
+        .. versionadded:: 0.10
+
+    .. describe:: iter(g)
+
+        Return an iterator over the attribute names.
+
+        .. versionadded:: 0.10
+    """
+
+    # Define attr methods to let mypy know this is a namespace object
+    # that has arbitrary attributes.
+
+    def __getattr__(self, name: str) -> t.Any:
+        try:
+            return self.__dict__[name]
+        except KeyError:
+            raise AttributeError(name) from None
+
+    def __setattr__(self, name: str, value: t.Any) -> None:
+        self.__dict__[name] = value
+
+    def __delattr__(self, name: str) -> None:
+        try:
+            del self.__dict__[name]
+        except KeyError:
+            raise AttributeError(name) from None
+
+    def get(self, name: str, default: t.Any | None = None) -> t.Any:
+        """Get an attribute by name, or a default value. Like
+        :meth:`dict.get`.
+
+        :param name: Name of attribute to get.
+        :param default: Value to return if the attribute is not present.
+
+        .. versionadded:: 0.10
+        """
+        return self.__dict__.get(name, default)
+
+    def pop(self, name: str, default: t.Any = _sentinel) -> t.Any:
+        """Get and remove an attribute by name. Like :meth:`dict.pop`.
+
+        :param name: Name of attribute to pop.
+        :param default: Value to return if the attribute is not present,
+            instead of raising a ``KeyError``.
+
+        .. versionadded:: 0.11
+        """
+        if default is _sentinel:
+            return self.__dict__.pop(name)
+        else:
+            return self.__dict__.pop(name, default)
+
+    def setdefault(self, name: str, default: t.Any = None) -> t.Any:
+        """Get the value of an attribute if it is present, otherwise
+        set and return a default value. Like :meth:`dict.setdefault`.
+
+        :param name: Name of attribute to get.
+        :param default: Value to set and return if the attribute is not
+            present.
+
+        .. versionadded:: 0.11
+        """
+        return self.__dict__.setdefault(name, default)
+
+    def __contains__(self, item: str) -> bool:
+        return item in self.__dict__
+
+    def __iter__(self) -> t.Iterator[str]:
+        return iter(self.__dict__)
+
+    def __repr__(self) -> str:
+        ctx = _cv_app.get(None)
+        if ctx is not None:
+            return f"<flask.g of '{ctx.app.name}'>"
+        return object.__repr__(self)
+
+
+def after_this_request(
+    f: ft.AfterRequestCallable[t.Any],
+) -> ft.AfterRequestCallable[t.Any]:
+    """Executes a function after this request.  This is useful to modify
+    response objects.  The function is passed the response object and has
+    to return the same or a new one.
+
+    Example::
+
+        @app.route('/')
+        def index():
+            @after_this_request
+            def add_header(response):
+                response.headers['X-Foo'] = 'Parachute'
+                return response
+            return 'Hello World!'
+
+    This is more useful if a function other than the view function wants to
+    modify a response.  For instance think of a decorator that wants to add
+    some headers without converting the return value into a response object.
+
+    .. versionadded:: 0.9
+    """
+    ctx = _cv_request.get(None)
+
+    if ctx is None:
+        raise RuntimeError(
+            "'after_this_request' can only be used when a request"
+            " context is active, such as in a view function."
+        )
+
+    ctx._after_request_functions.append(f)
+    return f
+
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+
+def copy_current_request_context(f: F) -> F:
+    """A helper function that decorates a function to retain the current
+    request context.  This is useful when working with greenlets.  The moment
+    the function is decorated a copy of the request context is created and
+    then pushed when the function is called.  The current session is also
+    included in the copied request context.
+
+    Example::
+
+        import gevent
+        from flask import copy_current_request_context
+
+        @app.route('/')
+        def index():
+            @copy_current_request_context
+            def do_some_work():
+                # do some work here, it can access flask.request or
+                # flask.session like you would otherwise in the view function.
+                ...
+            gevent.spawn(do_some_work)
+            return 'Regular response'
+
+    .. versionadded:: 0.10
+    """
+    ctx = _cv_request.get(None)
+
+    if ctx is None:
+        raise RuntimeError(
+            "'copy_current_request_context' can only be used when a"
+            " request context is active, such as in a view function."
+        )
+
+    ctx = ctx.copy()
+
+    def wrapper(*args: t.Any, **kwargs: t.Any) -> t.Any:
+        with ctx:  # type: ignore[union-attr]
+            return ctx.app.ensure_sync(f)(*args, **kwargs)  # type: ignore[union-attr]
+
+    return update_wrapper(wrapper, f)  # type: ignore[return-value]
+
+
+def has_request_context() -> bool:
+    """If you have code that wants to test if a request context is there or
+    not this function can be used.  For instance, you may want to take advantage
+    of request information if the request object is available, but fail
+    silently if it is unavailable.
+
+    ::
+
+        class User(db.Model):
+
+            def __init__(self, username, remote_addr=None):
+                self.username = username
+                if remote_addr is None and has_request_context():
+                    remote_addr = request.remote_addr
+                self.remote_addr = remote_addr
+
+    Alternatively you can also just test any of the context bound objects
+    (such as :class:`request` or :class:`g`) for truthness::
+
+        class User(db.Model):
+
+            def __init__(self, username, remote_addr=None):
+                self.username = username
+                if remote_addr is None and request:
+                    remote_addr = request.remote_addr
+                self.remote_addr = remote_addr
+
+    .. versionadded:: 0.7
+    """
+    return _cv_request.get(None) is not None
+
+
+def has_app_context() -> bool:
+    """Works like :func:`has_request_context` but for the application
+    context.  You can also just do a boolean check on the
+    :data:`current_app` object instead.
+
+    .. versionadded:: 0.9
+    """
+    return _cv_app.get(None) is not None
+
+
+class AppContext:
+    """The app context contains application-specific information. An app
+    context is created and pushed at the beginning of each request if
+    one is not already active. An app context is also pushed when
+    running CLI commands.
+    """
+
+    def __init__(self, app: Flask) -> None:
+        self.app = app
+        self.url_adapter = app.create_url_adapter(None)
+        self.g: _AppCtxGlobals = app.app_ctx_globals_class()
+        self._cv_tokens: list[contextvars.Token[AppContext]] = []
+
+    def push(self) -> None:
+        """Binds the app context to the current context."""
+        self._cv_tokens.append(_cv_app.set(self))
+        appcontext_pushed.send(self.app, _async_wrapper=self.app.ensure_sync)
+
+    def pop(self, exc: BaseException | None = _sentinel) -> None:  # type: ignore
+        """Pops the app context."""
+        try:
+            if len(self._cv_tokens) == 1:
+                if exc is _sentinel:
+                    exc = sys.exc_info()[1]
+                self.app.do_teardown_appcontext(exc)
+        finally:
+            ctx = _cv_app.get()
+            _cv_app.reset(self._cv_tokens.pop())
+
+        if ctx is not self:
+            raise AssertionError(
+                f"Popped wrong app context. ({ctx!r} instead of {self!r})"
+            )
+
+        appcontext_popped.send(self.app, _async_wrapper=self.app.ensure_sync)
+
+    def __enter__(self) -> AppContext:
+        self.push()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type | None,
+        exc_value: BaseException | None,
+        tb: TracebackType | None,
+    ) -> None:
+        self.pop(exc_value)
+
+
+class RequestContext:
+    """The request context contains per-request information. The Flask
+    app creates and pushes it at the beginning of the request, then pops
+    it at the end of the request. It will create the URL adapter and
+    request object for the WSGI environment provided.
+
+    Do not attempt to use this class directly, instead use
+    :meth:`~flask.Flask.test_request_context` and
+    :meth:`~flask.Flask.request_context` to create this object.
+
+    When the request context is popped, it will evaluate all the
+    functions registered on the application for teardown execution
+    (:meth:`~flask.Flask.teardown_request`).
+
+    The request context is automatically popped at the end of the
+    request. When using the interactive debugger, the context will be
+    restored so ``request`` is still accessible. Similarly, the test
+    client can preserve the context after the request ends. However,
+    teardown functions may already have closed some resources such as
+    database connections.
+    """
+
+    def __init__(
+        self,
+        app: Flask,
+        environ: WSGIEnvironment,
+        request: Request | None = None,
+        session: SessionMixin | None = None,
+    ) -> None:
+        self.app = app
+        if request is None:
+            request = app.request_class(environ)
+            request.json_module = app.json
+        self.request: Request = request
+        self.url_adapter = None
+        try:
+            self.url_adapter = app.create_url_adapter(self.request)
+        except HTTPException as e:
+            self.request.routing_exception = e
+        self.flashes: list[tuple[str, str]] | None = None
+        self.session: SessionMixin | None = session
+        # Functions that should be executed after the request on the response
+        # object.  These will be called before the regular "after_request"
+        # functions.
+        self._after_request_functions: list[ft.AfterRequestCallable[t.Any]] = []
+
+        self._cv_tokens: list[
+            tuple[contextvars.Token[RequestContext], AppContext | None]
+        ] = []
+
+    def copy(self) -> RequestContext:
+        """Creates a copy of this request context with the same request object.
+        This can be used to move a request context to a different greenlet.
+        Because the actual request object is the same this cannot be used to
+        move a request context to a different thread unless access to the
+        request object is locked.
+
+        .. versionadded:: 0.10
+
+        .. versionchanged:: 1.1
+           The current session object is used instead of reloading the original
+           data. This prevents `flask.session` pointing to an out-of-date object.
+        """
+        return self.__class__(
+            self.app,
+            environ=self.request.environ,
+            request=self.request,
+            session=self.session,
+        )
+
+    def match_request(self) -> None:
+        """Can be overridden by a subclass to hook into the matching
+        of the request.
+        """
+        try:
+            result = self.url_adapter.match(return_rule=True)  # type: ignore
+            self.request.url_rule, self.request.view_args = result  # type: ignore
+        except HTTPException as e:
+            self.request.routing_exception = e
+
+    def push(self) -> None:
+        # Before we push the request context we have to ensure that there
+        # is an application context.
+        app_ctx = _cv_app.get(None)
+
+        if app_ctx is None or app_ctx.app is not self.app:
+            app_ctx = self.app.app_context()
+            app_ctx.push()
+        else:
+            app_ctx = None
+
+        self._cv_tokens.append((_cv_request.set(self), app_ctx))
+
+        # Open the session at the moment that the request context is available.
+        # This allows a custom open_session method to use the request context.
+        # Only open a new session if this is the first time the request was
+        # pushed, otherwise stream_with_context loses the session.
+        if self.session is None:
+            session_interface = self.app.session_interface
+            self.session = session_interface.open_session(self.app, self.request)
+
+            if self.session is None:
+                self.session = session_interface.make_null_session(self.app)
+
+        # Match the request URL after loading the session, so that the
+        # session is available in custom URL converters.
+        if self.url_adapter is not None:
+            self.match_request()
+
+    def pop(self, exc: BaseException | None = _sentinel) -> None:  # type: ignore
+        """Pops the request context and unbinds it by doing that.  This will
+        also trigger the execution of functions registered by the
+        :meth:`~flask.Flask.teardown_request` decorator.
+
+        .. versionchanged:: 0.9
+           Added the `exc` argument.
+        """
+        clear_request = len(self._cv_tokens) == 1
+
+        try:
+            if clear_request:
+                if exc is _sentinel:
+                    exc = sys.exc_info()[1]
+                self.app.do_teardown_request(exc)
+
+                request_close = getattr(self.request, "close", None)
+                if request_close is not None:
+                    request_close()
+        finally:
+            ctx = _cv_request.get()
+            token, app_ctx = self._cv_tokens.pop()
+            _cv_request.reset(token)
+
+            # get rid of circular dependencies at the end of the request
+            # so that we don't require the GC to be active.
+            if clear_request:
+                ctx.request.environ["werkzeug.request"] = None
+
+            if app_ctx is not None:
+                app_ctx.pop(exc)
+
+            if ctx is not self:
+                raise AssertionError(
+                    f"Popped wrong request context. ({ctx!r} instead of {self!r})"
+                )
+
+    def __enter__(self) -> RequestContext:
+        self.push()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type | None,
+        exc_value: BaseException | None,
+        tb: TracebackType | None,
+    ) -> None:
+        self.pop(exc_value)
+
+    def __repr__(self) -> str:
+        return (
+            f"<{type(self).__name__} {self.request.url!r}"
+            f" [{self.request.method}] of {self.app.name}>"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/flask/debughelpers.py b/backend/venv/lib/python3.12/site-packages/flask/debughelpers.py
new file mode 100644
index 00000000..2c8c4c48
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/debughelpers.py
@@ -0,0 +1,178 @@
+from __future__ import annotations
+
+import typing as t
+
+from jinja2.loaders import BaseLoader
+from werkzeug.routing import RequestRedirect
+
+from .blueprints import Blueprint
+from .globals import request_ctx
+from .sansio.app import App
+
+if t.TYPE_CHECKING:
+    from .sansio.scaffold import Scaffold
+    from .wrappers import Request
+
+
+class UnexpectedUnicodeError(AssertionError, UnicodeError):
+    """Raised in places where we want some better error reporting for
+    unexpected unicode or binary data.
+    """
+
+
+class DebugFilesKeyError(KeyError, AssertionError):
+    """Raised from request.files during debugging.  The idea is that it can
+    provide a better error message than just a generic KeyError/BadRequest.
+    """
+
+    def __init__(self, request: Request, key: str) -> None:
+        form_matches = request.form.getlist(key)
+        buf = [
+            f"You tried to access the file {key!r} in the request.files"
+            " dictionary but it does not exist. The mimetype for the"
+            f" request is {request.mimetype!r} instead of"
+            " 'multipart/form-data' which means that no file contents"
+            " were transmitted. To fix this error you should provide"
+            ' enctype="multipart/form-data" in your form.'
+        ]
+        if form_matches:
+            names = ", ".join(repr(x) for x in form_matches)
+            buf.append(
+                "\n\nThe browser instead transmitted some file names. "
+                f"This was submitted: {names}"
+            )
+        self.msg = "".join(buf)
+
+    def __str__(self) -> str:
+        return self.msg
+
+
+class FormDataRoutingRedirect(AssertionError):
+    """This exception is raised in debug mode if a routing redirect
+    would cause the browser to drop the method or body. This happens
+    when method is not GET, HEAD or OPTIONS and the status code is not
+    307 or 308.
+    """
+
+    def __init__(self, request: Request) -> None:
+        exc = request.routing_exception
+        assert isinstance(exc, RequestRedirect)
+        buf = [
+            f"A request was sent to '{request.url}', but routing issued"
+            f" a redirect to the canonical URL '{exc.new_url}'."
+        ]
+
+        if f"{request.base_url}/" == exc.new_url.partition("?")[0]:
+            buf.append(
+                " The URL was defined with a trailing slash. Flask"
+                " will redirect to the URL with a trailing slash if it"
+                " was accessed without one."
+            )
+
+        buf.append(
+            " Send requests to the canonical URL, or use 307 or 308 for"
+            " routing redirects. Otherwise, browsers will drop form"
+            " data.\n\n"
+            "This exception is only raised in debug mode."
+        )
+        super().__init__("".join(buf))
+
+
+def attach_enctype_error_multidict(request: Request) -> None:
+    """Patch ``request.files.__getitem__`` to raise a descriptive error
+    about ``enctype=multipart/form-data``.
+
+    :param request: The request to patch.
+    :meta private:
+    """
+    oldcls = request.files.__class__
+
+    class newcls(oldcls):  # type: ignore[valid-type, misc]
+        def __getitem__(self, key: str) -> t.Any:
+            try:
+                return super().__getitem__(key)
+            except KeyError as e:
+                if key not in request.form:
+                    raise
+
+                raise DebugFilesKeyError(request, key).with_traceback(
+                    e.__traceback__
+                ) from None
+
+    newcls.__name__ = oldcls.__name__
+    newcls.__module__ = oldcls.__module__
+    request.files.__class__ = newcls
+
+
+def _dump_loader_info(loader: BaseLoader) -> t.Iterator[str]:
+    yield f"class: {type(loader).__module__}.{type(loader).__name__}"
+    for key, value in sorted(loader.__dict__.items()):
+        if key.startswith("_"):
+            continue
+        if isinstance(value, (tuple, list)):
+            if not all(isinstance(x, str) for x in value):
+                continue
+            yield f"{key}:"
+            for item in value:
+                yield f"  - {item}"
+            continue
+        elif not isinstance(value, (str, int, float, bool)):
+            continue
+        yield f"{key}: {value!r}"
+
+
+def explain_template_loading_attempts(
+    app: App,
+    template: str,
+    attempts: list[
+        tuple[
+            BaseLoader,
+            Scaffold,
+            tuple[str, str | None, t.Callable[[], bool] | None] | None,
+        ]
+    ],
+) -> None:
+    """This should help developers understand what failed"""
+    info = [f"Locating template {template!r}:"]
+    total_found = 0
+    blueprint = None
+    if request_ctx and request_ctx.request.blueprint is not None:
+        blueprint = request_ctx.request.blueprint
+
+    for idx, (loader, srcobj, triple) in enumerate(attempts):
+        if isinstance(srcobj, App):
+            src_info = f"application {srcobj.import_name!r}"
+        elif isinstance(srcobj, Blueprint):
+            src_info = f"blueprint {srcobj.name!r} ({srcobj.import_name})"
+        else:
+            src_info = repr(srcobj)
+
+        info.append(f"{idx + 1:5}: trying loader of {src_info}")
+
+        for line in _dump_loader_info(loader):
+            info.append(f"       {line}")
+
+        if triple is None:
+            detail = "no match"
+        else:
+            detail = f"found ({triple[1] or '<string>'!r})"
+            total_found += 1
+        info.append(f"       -> {detail}")
+
+    seems_fishy = False
+    if total_found == 0:
+        info.append("Error: the template could not be found.")
+        seems_fishy = True
+    elif total_found > 1:
+        info.append("Warning: multiple loaders returned a match for the template.")
+        seems_fishy = True
+
+    if blueprint is not None and seems_fishy:
+        info.append(
+            "  The template was looked up from an endpoint that belongs"
+            f" to the blueprint {blueprint!r}."
+        )
+        info.append("  Maybe you did not place a template in the right folder?")
+        info.append("  See https://flask.palletsprojects.com/blueprints/#templates")
+
+    app.logger.info("\n".join(info))
diff --git a/backend/venv/lib/python3.12/site-packages/flask/globals.py b/backend/venv/lib/python3.12/site-packages/flask/globals.py
new file mode 100644
index 00000000..e2c410cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/globals.py
@@ -0,0 +1,51 @@
+from __future__ import annotations
+
+import typing as t
+from contextvars import ContextVar
+
+from werkzeug.local import LocalProxy
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .app import Flask
+    from .ctx import _AppCtxGlobals
+    from .ctx import AppContext
+    from .ctx import RequestContext
+    from .sessions import SessionMixin
+    from .wrappers import Request
+
+
+_no_app_msg = """\
+Working outside of application context.
+
+This typically means that you attempted to use functionality that needed
+the current application. To solve this, set up an application context
+with app.app_context(). See the documentation for more information.\
+"""
+_cv_app: ContextVar[AppContext] = ContextVar("flask.app_ctx")
+app_ctx: AppContext = LocalProxy(  # type: ignore[assignment]
+    _cv_app, unbound_message=_no_app_msg
+)
+current_app: Flask = LocalProxy(  # type: ignore[assignment]
+    _cv_app, "app", unbound_message=_no_app_msg
+)
+g: _AppCtxGlobals = LocalProxy(  # type: ignore[assignment]
+    _cv_app, "g", unbound_message=_no_app_msg
+)
+
+_no_req_msg = """\
+Working outside of request context.
+
+This typically means that you attempted to use functionality that needed
+an active HTTP request. Consult the documentation on testing for
+information about how to avoid this problem.\
+"""
+_cv_request: ContextVar[RequestContext] = ContextVar("flask.request_ctx")
+request_ctx: RequestContext = LocalProxy(  # type: ignore[assignment]
+    _cv_request, unbound_message=_no_req_msg
+)
+request: Request = LocalProxy(  # type: ignore[assignment]
+    _cv_request, "request", unbound_message=_no_req_msg
+)
+session: SessionMixin = LocalProxy(  # type: ignore[assignment]
+    _cv_request, "session", unbound_message=_no_req_msg
+)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/helpers.py b/backend/venv/lib/python3.12/site-packages/flask/helpers.py
new file mode 100644
index 00000000..a6b7e150
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/helpers.py
@@ -0,0 +1,634 @@
+from __future__ import annotations
+
+import importlib.util
+import os
+import sys
+import typing as t
+from datetime import datetime
+from functools import cache
+from functools import update_wrapper
+
+import werkzeug.utils
+from werkzeug.exceptions import abort as _wz_abort
+from werkzeug.utils import redirect as _wz_redirect
+from werkzeug.wrappers import Response as BaseResponse
+
+from .globals import _cv_request
+from .globals import current_app
+from .globals import request
+from .globals import request_ctx
+from .globals import session
+from .signals import message_flashed
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .wrappers import Response
+
+
+def get_debug_flag() -> bool:
+    """Get whether debug mode should be enabled for the app, indicated by the
+    :envvar:`FLASK_DEBUG` environment variable. The default is ``False``.
+    """
+    val = os.environ.get("FLASK_DEBUG")
+    return bool(val and val.lower() not in {"0", "false", "no"})
+
+
+def get_load_dotenv(default: bool = True) -> bool:
+    """Get whether the user has disabled loading default dotenv files by
+    setting :envvar:`FLASK_SKIP_DOTENV`. The default is ``True``, load
+    the files.
+
+    :param default: What to return if the env var isn't set.
+    """
+    val = os.environ.get("FLASK_SKIP_DOTENV")
+
+    if not val:
+        return default
+
+    return val.lower() in ("0", "false", "no")
+
+
+@t.overload
+def stream_with_context(
+    generator_or_function: t.Iterator[t.AnyStr],
+) -> t.Iterator[t.AnyStr]: ...
+
+
+@t.overload
+def stream_with_context(
+    generator_or_function: t.Callable[..., t.Iterator[t.AnyStr]],
+) -> t.Callable[[t.Iterator[t.AnyStr]], t.Iterator[t.AnyStr]]: ...
+
+
+def stream_with_context(
+    generator_or_function: t.Iterator[t.AnyStr] | t.Callable[..., t.Iterator[t.AnyStr]],
+) -> t.Iterator[t.AnyStr] | t.Callable[[t.Iterator[t.AnyStr]], t.Iterator[t.AnyStr]]:
+    """Request contexts disappear when the response is started on the server.
+    This is done for efficiency reasons and to make it less likely to encounter
+    memory leaks with badly written WSGI middlewares.  The downside is that if
+    you are using streamed responses, the generator cannot access request bound
+    information any more.
+
+    This function however can help you keep the context around for longer::
+
+        from flask import stream_with_context, request, Response
+
+        @app.route('/stream')
+        def streamed_response():
+            @stream_with_context
+            def generate():
+                yield 'Hello '
+                yield request.args['name']
+                yield '!'
+            return Response(generate())
+
+    Alternatively it can also be used around a specific generator::
+
+        from flask import stream_with_context, request, Response
+
+        @app.route('/stream')
+        def streamed_response():
+            def generate():
+                yield 'Hello '
+                yield request.args['name']
+                yield '!'
+            return Response(stream_with_context(generate()))
+
+    .. versionadded:: 0.9
+    """
+    try:
+        gen = iter(generator_or_function)  # type: ignore[arg-type]
+    except TypeError:
+
+        def decorator(*args: t.Any, **kwargs: t.Any) -> t.Any:
+            gen = generator_or_function(*args, **kwargs)  # type: ignore[operator]
+            return stream_with_context(gen)
+
+        return update_wrapper(decorator, generator_or_function)  # type: ignore[arg-type]
+
+    def generator() -> t.Iterator[t.AnyStr | None]:
+        ctx = _cv_request.get(None)
+        if ctx is None:
+            raise RuntimeError(
+                "'stream_with_context' can only be used when a request"
+                " context is active, such as in a view function."
+            )
+        with ctx:
+            # Dummy sentinel.  Has to be inside the context block or we're
+            # not actually keeping the context around.
+            yield None
+
+            # The try/finally is here so that if someone passes a WSGI level
+            # iterator in we're still running the cleanup logic.  Generators
+            # don't need that because they are closed on their destruction
+            # automatically.
+            try:
+                yield from gen
+            finally:
+                if hasattr(gen, "close"):
+                    gen.close()
+
+    # The trick is to start the generator.  Then the code execution runs until
+    # the first dummy None is yielded at which point the context was already
+    # pushed.  This item is discarded.  Then when the iteration continues the
+    # real generator is executed.
+    wrapped_g = generator()
+    next(wrapped_g)
+    return wrapped_g  # type: ignore[return-value]
+
+
+def make_response(*args: t.Any) -> Response:
+    """Sometimes it is necessary to set additional headers in a view.  Because
+    views do not have to return response objects but can return a value that
+    is converted into a response object by Flask itself, it becomes tricky to
+    add headers to it.  This function can be called instead of using a return
+    and you will get a response object which you can use to attach headers.
+
+    If view looked like this and you want to add a new header::
+
+        def index():
+            return render_template('index.html', foo=42)
+
+    You can now do something like this::
+
+        def index():
+            response = make_response(render_template('index.html', foo=42))
+            response.headers['X-Parachutes'] = 'parachutes are cool'
+            return response
+
+    This function accepts the very same arguments you can return from a
+    view function.  This for example creates a response with a 404 error
+    code::
+
+        response = make_response(render_template('not_found.html'), 404)
+
+    The other use case of this function is to force the return value of a
+    view function into a response which is helpful with view
+    decorators::
+
+        response = make_response(view_function())
+        response.headers['X-Parachutes'] = 'parachutes are cool'
+
+    Internally this function does the following things:
+
+    -   if no arguments are passed, it creates a new response argument
+    -   if one argument is passed, :meth:`flask.Flask.make_response`
+        is invoked with it.
+    -   if more than one argument is passed, the arguments are passed
+        to the :meth:`flask.Flask.make_response` function as tuple.
+
+    .. versionadded:: 0.6
+    """
+    if not args:
+        return current_app.response_class()
+    if len(args) == 1:
+        args = args[0]
+    return current_app.make_response(args)
+
+
+def url_for(
+    endpoint: str,
+    *,
+    _anchor: str | None = None,
+    _method: str | None = None,
+    _scheme: str | None = None,
+    _external: bool | None = None,
+    **values: t.Any,
+) -> str:
+    """Generate a URL to the given endpoint with the given values.
+
+    This requires an active request or application context, and calls
+    :meth:`current_app.url_for() <flask.Flask.url_for>`. See that method
+    for full documentation.
+
+    :param endpoint: The endpoint name associated with the URL to
+        generate. If this starts with a ``.``, the current blueprint
+        name (if any) will be used.
+    :param _anchor: If given, append this as ``#anchor`` to the URL.
+    :param _method: If given, generate the URL associated with this
+        method for the endpoint.
+    :param _scheme: If given, the URL will have this scheme if it is
+        external.
+    :param _external: If given, prefer the URL to be internal (False) or
+        require it to be external (True). External URLs include the
+        scheme and domain. When not in an active request, URLs are
+        external by default.
+    :param values: Values to use for the variable parts of the URL rule.
+        Unknown keys are appended as query string arguments, like
+        ``?a=b&c=d``.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.url_for``, allowing an app to override the
+        behavior.
+
+    .. versionchanged:: 0.10
+       The ``_scheme`` parameter was added.
+
+    .. versionchanged:: 0.9
+       The ``_anchor`` and ``_method`` parameters were added.
+
+    .. versionchanged:: 0.9
+       Calls ``app.handle_url_build_error`` on build errors.
+    """
+    return current_app.url_for(
+        endpoint,
+        _anchor=_anchor,
+        _method=_method,
+        _scheme=_scheme,
+        _external=_external,
+        **values,
+    )
+
+
+def redirect(
+    location: str, code: int = 302, Response: type[BaseResponse] | None = None
+) -> BaseResponse:
+    """Create a redirect response object.
+
+    If :data:`~flask.current_app` is available, it will use its
+    :meth:`~flask.Flask.redirect` method, otherwise it will use
+    :func:`werkzeug.utils.redirect`.
+
+    :param location: The URL to redirect to.
+    :param code: The status code for the redirect.
+    :param Response: The response class to use. Not used when
+        ``current_app`` is active, which uses ``app.response_class``.
+
+    .. versionadded:: 2.2
+        Calls ``current_app.redirect`` if available instead of always
+        using Werkzeug's default ``redirect``.
+    """
+    if current_app:
+        return current_app.redirect(location, code=code)
+
+    return _wz_redirect(location, code=code, Response=Response)
+
+
+def abort(code: int | BaseResponse, *args: t.Any, **kwargs: t.Any) -> t.NoReturn:
+    """Raise an :exc:`~werkzeug.exceptions.HTTPException` for the given
+    status code.
+
+    If :data:`~flask.current_app` is available, it will call its
+    :attr:`~flask.Flask.aborter` object, otherwise it will use
+    :func:`werkzeug.exceptions.abort`.
+
+    :param code: The status code for the exception, which must be
+        registered in ``app.aborter``.
+    :param args: Passed to the exception.
+    :param kwargs: Passed to the exception.
+
+    .. versionadded:: 2.2
+        Calls ``current_app.aborter`` if available instead of always
+        using Werkzeug's default ``abort``.
+    """
+    if current_app:
+        current_app.aborter(code, *args, **kwargs)
+
+    _wz_abort(code, *args, **kwargs)
+
+
+def get_template_attribute(template_name: str, attribute: str) -> t.Any:
+    """Loads a macro (or variable) a template exports.  This can be used to
+    invoke a macro from within Python code.  If you for example have a
+    template named :file:`_cider.html` with the following contents:
+
+    .. sourcecode:: html+jinja
+
+       {% macro hello(name) %}Hello {{ name }}!{% endmacro %}
+
+    You can access this from Python code like this::
+
+        hello = get_template_attribute('_cider.html', 'hello')
+        return hello('World')
+
+    .. versionadded:: 0.2
+
+    :param template_name: the name of the template
+    :param attribute: the name of the variable of macro to access
+    """
+    return getattr(current_app.jinja_env.get_template(template_name).module, attribute)
+
+
+def flash(message: str, category: str = "message") -> None:
+    """Flashes a message to the next request.  In order to remove the
+    flashed message from the session and to display it to the user,
+    the template has to call :func:`get_flashed_messages`.
+
+    .. versionchanged:: 0.3
+       `category` parameter added.
+
+    :param message: the message to be flashed.
+    :param category: the category for the message.  The following values
+                     are recommended: ``'message'`` for any kind of message,
+                     ``'error'`` for errors, ``'info'`` for information
+                     messages and ``'warning'`` for warnings.  However any
+                     kind of string can be used as category.
+    """
+    # Original implementation:
+    #
+    #     session.setdefault('_flashes', []).append((category, message))
+    #
+    # This assumed that changes made to mutable structures in the session are
+    # always in sync with the session object, which is not true for session
+    # implementations that use external storage for keeping their keys/values.
+    flashes = session.get("_flashes", [])
+    flashes.append((category, message))
+    session["_flashes"] = flashes
+    app = current_app._get_current_object()  # type: ignore
+    message_flashed.send(
+        app,
+        _async_wrapper=app.ensure_sync,
+        message=message,
+        category=category,
+    )
+
+
+def get_flashed_messages(
+    with_categories: bool = False, category_filter: t.Iterable[str] = ()
+) -> list[str] | list[tuple[str, str]]:
+    """Pulls all flashed messages from the session and returns them.
+    Further calls in the same request to the function will return
+    the same messages.  By default just the messages are returned,
+    but when `with_categories` is set to ``True``, the return value will
+    be a list of tuples in the form ``(category, message)`` instead.
+
+    Filter the flashed messages to one or more categories by providing those
+    categories in `category_filter`.  This allows rendering categories in
+    separate html blocks.  The `with_categories` and `category_filter`
+    arguments are distinct:
+
+    * `with_categories` controls whether categories are returned with message
+      text (``True`` gives a tuple, where ``False`` gives just the message text).
+    * `category_filter` filters the messages down to only those matching the
+      provided categories.
+
+    See :doc:`/patterns/flashing` for examples.
+
+    .. versionchanged:: 0.3
+       `with_categories` parameter added.
+
+    .. versionchanged:: 0.9
+        `category_filter` parameter added.
+
+    :param with_categories: set to ``True`` to also receive categories.
+    :param category_filter: filter of categories to limit return values.  Only
+                            categories in the list will be returned.
+    """
+    flashes = request_ctx.flashes
+    if flashes is None:
+        flashes = session.pop("_flashes") if "_flashes" in session else []
+        request_ctx.flashes = flashes
+    if category_filter:
+        flashes = list(filter(lambda f: f[0] in category_filter, flashes))
+    if not with_categories:
+        return [x[1] for x in flashes]
+    return flashes
+
+
+def _prepare_send_file_kwargs(**kwargs: t.Any) -> dict[str, t.Any]:
+    if kwargs.get("max_age") is None:
+        kwargs["max_age"] = current_app.get_send_file_max_age
+
+    kwargs.update(
+        environ=request.environ,
+        use_x_sendfile=current_app.config["USE_X_SENDFILE"],
+        response_class=current_app.response_class,
+        _root_path=current_app.root_path,  # type: ignore
+    )
+    return kwargs
+
+
+def send_file(
+    path_or_file: os.PathLike[t.AnyStr] | str | t.BinaryIO,
+    mimetype: str | None = None,
+    as_attachment: bool = False,
+    download_name: str | None = None,
+    conditional: bool = True,
+    etag: bool | str = True,
+    last_modified: datetime | int | float | None = None,
+    max_age: None | (int | t.Callable[[str | None], int | None]) = None,
+) -> Response:
+    """Send the contents of a file to the client.
+
+    The first argument can be a file path or a file-like object. Paths
+    are preferred in most cases because Werkzeug can manage the file and
+    get extra information from the path. Passing a file-like object
+    requires that the file is opened in binary mode, and is mostly
+    useful when building a file in memory with :class:`io.BytesIO`.
+
+    Never pass file paths provided by a user. The path is assumed to be
+    trusted, so a user could craft a path to access a file you didn't
+    intend. Use :func:`send_from_directory` to safely serve
+    user-requested paths from within a directory.
+
+    If the WSGI server sets a ``file_wrapper`` in ``environ``, it is
+    used, otherwise Werkzeug's built-in wrapper is used. Alternatively,
+    if the HTTP server supports ``X-Sendfile``, configuring Flask with
+    ``USE_X_SENDFILE = True`` will tell the server to send the given
+    path, which is much more efficient than reading it in Python.
+
+    :param path_or_file: The path to the file to send, relative to the
+        current working directory if a relative path is given.
+        Alternatively, a file-like object opened in binary mode. Make
+        sure the file pointer is seeked to the start of the data.
+    :param mimetype: The MIME type to send for the file. If not
+        provided, it will try to detect it from the file name.
+    :param as_attachment: Indicate to a browser that it should offer to
+        save the file instead of displaying it.
+    :param download_name: The default name browsers will use when saving
+        the file. Defaults to the passed file name.
+    :param conditional: Enable conditional and range responses based on
+        request headers. Requires passing a file path and ``environ``.
+    :param etag: Calculate an ETag for the file, which requires passing
+        a file path. Can also be a string to use instead.
+    :param last_modified: The last modified time to send for the file,
+        in seconds. If not provided, it will try to detect it from the
+        file path.
+    :param max_age: How long the client should cache the file, in
+        seconds. If set, ``Cache-Control`` will be ``public``, otherwise
+        it will be ``no-cache`` to prefer conditional caching.
+
+    .. versionchanged:: 2.0
+        ``download_name`` replaces the ``attachment_filename``
+        parameter. If ``as_attachment=False``, it is passed with
+        ``Content-Disposition: inline`` instead.
+
+    .. versionchanged:: 2.0
+        ``max_age`` replaces the ``cache_timeout`` parameter.
+        ``conditional`` is enabled and ``max_age`` is not set by
+        default.
+
+    .. versionchanged:: 2.0
+        ``etag`` replaces the ``add_etags`` parameter. It can be a
+        string to use instead of generating one.
+
+    .. versionchanged:: 2.0
+        Passing a file-like object that inherits from
+        :class:`~io.TextIOBase` will raise a :exc:`ValueError` rather
+        than sending an empty file.
+
+    .. versionadded:: 2.0
+        Moved the implementation to Werkzeug. This is now a wrapper to
+        pass some Flask-specific arguments.
+
+    .. versionchanged:: 1.1
+        ``filename`` may be a :class:`~os.PathLike` object.
+
+    .. versionchanged:: 1.1
+        Passing a :class:`~io.BytesIO` object supports range requests.
+
+    .. versionchanged:: 1.0.3
+        Filenames are encoded with ASCII instead of Latin-1 for broader
+        compatibility with WSGI servers.
+
+    .. versionchanged:: 1.0
+        UTF-8 filenames as specified in :rfc:`2231` are supported.
+
+    .. versionchanged:: 0.12
+        The filename is no longer automatically inferred from file
+        objects. If you want to use automatic MIME and etag support,
+        pass a filename via ``filename_or_fp`` or
+        ``attachment_filename``.
+
+    .. versionchanged:: 0.12
+        ``attachment_filename`` is preferred over ``filename`` for MIME
+        detection.
+
+    .. versionchanged:: 0.9
+        ``cache_timeout`` defaults to
+        :meth:`Flask.get_send_file_max_age`.
+
+    .. versionchanged:: 0.7
+        MIME guessing and etag support for file-like objects was
+        removed because it was unreliable. Pass a filename if you are
+        able to, otherwise attach an etag yourself.
+
+    .. versionchanged:: 0.5
+        The ``add_etags``, ``cache_timeout`` and ``conditional``
+        parameters were added. The default behavior is to add etags.
+
+    .. versionadded:: 0.2
+    """
+    return werkzeug.utils.send_file(  # type: ignore[return-value]
+        **_prepare_send_file_kwargs(
+            path_or_file=path_or_file,
+            environ=request.environ,
+            mimetype=mimetype,
+            as_attachment=as_attachment,
+            download_name=download_name,
+            conditional=conditional,
+            etag=etag,
+            last_modified=last_modified,
+            max_age=max_age,
+        )
+    )
+
+
+def send_from_directory(
+    directory: os.PathLike[str] | str,
+    path: os.PathLike[str] | str,
+    **kwargs: t.Any,
+) -> Response:
+    """Send a file from within a directory using :func:`send_file`.
+
+    .. code-block:: python
+
+        @app.route("/uploads/<path:name>")
+        def download_file(name):
+            return send_from_directory(
+                app.config['UPLOAD_FOLDER'], name, as_attachment=True
+            )
+
+    This is a secure way to serve files from a folder, such as static
+    files or uploads. Uses :func:`~werkzeug.security.safe_join` to
+    ensure the path coming from the client is not maliciously crafted to
+    point outside the specified directory.
+
+    If the final path does not point to an existing regular file,
+    raises a 404 :exc:`~werkzeug.exceptions.NotFound` error.
+
+    :param directory: The directory that ``path`` must be located under,
+        relative to the current application's root path. This *must not*
+        be a value provided by the client, otherwise it becomes insecure.
+    :param path: The path to the file to send, relative to
+        ``directory``.
+    :param kwargs: Arguments to pass to :func:`send_file`.
+
+    .. versionchanged:: 2.0
+        ``path`` replaces the ``filename`` parameter.
+
+    .. versionadded:: 2.0
+        Moved the implementation to Werkzeug. This is now a wrapper to
+        pass some Flask-specific arguments.
+
+    .. versionadded:: 0.5
+    """
+    return werkzeug.utils.send_from_directory(  # type: ignore[return-value]
+        directory, path, **_prepare_send_file_kwargs(**kwargs)
+    )
+
+
+def get_root_path(import_name: str) -> str:
+    """Find the root path of a package, or the path that contains a
+    module. If it cannot be found, returns the current working
+    directory.
+
+    Not to be confused with the value returned by :func:`find_package`.
+
+    :meta private:
+    """
+    # Module already imported and has a file attribute. Use that first.
+    mod = sys.modules.get(import_name)
+
+    if mod is not None and hasattr(mod, "__file__") and mod.__file__ is not None:
+        return os.path.dirname(os.path.abspath(mod.__file__))
+
+    # Next attempt: check the loader.
+    try:
+        spec = importlib.util.find_spec(import_name)
+
+        if spec is None:
+            raise ValueError
+    except (ImportError, ValueError):
+        loader = None
+    else:
+        loader = spec.loader
+
+    # Loader does not exist or we're referring to an unloaded main
+    # module or a main module without path (interactive sessions), go
+    # with the current working directory.
+    if loader is None:
+        return os.getcwd()
+
+    if hasattr(loader, "get_filename"):
+        filepath = loader.get_filename(import_name)  # pyright: ignore
+    else:
+        # Fall back to imports.
+        __import__(import_name)
+        mod = sys.modules[import_name]
+        filepath = getattr(mod, "__file__", None)
+
+        # If we don't have a file path it might be because it is a
+        # namespace package. In this case pick the root path from the
+        # first module that is contained in the package.
+        if filepath is None:
+            raise RuntimeError(
+                "No root path can be found for the provided module"
+                f" {import_name!r}. This can happen because the module"
+                " came from an import hook that does not provide file"
+                " name information or because it's a namespace package."
+                " In this case the root path needs to be explicitly"
+                " provided."
+            )
+
+    # filepath is import_name.py for a module, or __init__.py for a package.
+    return os.path.dirname(os.path.abspath(filepath))  # type: ignore[no-any-return]
+
+
+@cache
+def _split_blueprint_path(name: str) -> list[str]:
+    out: list[str] = [name]
+
+    if "." in name:
+        out.extend(_split_blueprint_path(name.rpartition(".")[0]))
+
+    return out
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/__init__.py b/backend/venv/lib/python3.12/site-packages/flask/json/__init__.py
new file mode 100644
index 00000000..c0941d04
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/json/__init__.py
@@ -0,0 +1,170 @@
+from __future__ import annotations
+
+import json as _json
+import typing as t
+
+from ..globals import current_app
+from .provider import _default
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from ..wrappers import Response
+
+
+def dumps(obj: t.Any, **kwargs: t.Any) -> str:
+    """Serialize data as JSON.
+
+    If :data:`~flask.current_app` is available, it will use its
+    :meth:`app.json.dumps() <flask.json.provider.JSONProvider.dumps>`
+    method, otherwise it will use :func:`json.dumps`.
+
+    :param obj: The data to serialize.
+    :param kwargs: Arguments passed to the ``dumps`` implementation.
+
+    .. versionchanged:: 2.3
+        The ``app`` parameter was removed.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.json.dumps``, allowing an app to override
+        the behavior.
+
+    .. versionchanged:: 2.0.2
+        :class:`decimal.Decimal` is supported by converting to a string.
+
+    .. versionchanged:: 2.0
+        ``encoding`` will be removed in Flask 2.1.
+
+    .. versionchanged:: 1.0.3
+        ``app`` can be passed directly, rather than requiring an app
+        context for configuration.
+    """
+    if current_app:
+        return current_app.json.dumps(obj, **kwargs)
+
+    kwargs.setdefault("default", _default)
+    return _json.dumps(obj, **kwargs)
+
+
+def dump(obj: t.Any, fp: t.IO[str], **kwargs: t.Any) -> None:
+    """Serialize data as JSON and write to a file.
+
+    If :data:`~flask.current_app` is available, it will use its
+    :meth:`app.json.dump() <flask.json.provider.JSONProvider.dump>`
+    method, otherwise it will use :func:`json.dump`.
+
+    :param obj: The data to serialize.
+    :param fp: A file opened for writing text. Should use the UTF-8
+        encoding to be valid JSON.
+    :param kwargs: Arguments passed to the ``dump`` implementation.
+
+    .. versionchanged:: 2.3
+        The ``app`` parameter was removed.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.json.dump``, allowing an app to override
+        the behavior.
+
+    .. versionchanged:: 2.0
+        Writing to a binary file, and the ``encoding`` argument, will be
+        removed in Flask 2.1.
+    """
+    if current_app:
+        current_app.json.dump(obj, fp, **kwargs)
+    else:
+        kwargs.setdefault("default", _default)
+        _json.dump(obj, fp, **kwargs)
+
+
+def loads(s: str | bytes, **kwargs: t.Any) -> t.Any:
+    """Deserialize data as JSON.
+
+    If :data:`~flask.current_app` is available, it will use its
+    :meth:`app.json.loads() <flask.json.provider.JSONProvider.loads>`
+    method, otherwise it will use :func:`json.loads`.
+
+    :param s: Text or UTF-8 bytes.
+    :param kwargs: Arguments passed to the ``loads`` implementation.
+
+    .. versionchanged:: 2.3
+        The ``app`` parameter was removed.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.json.loads``, allowing an app to override
+        the behavior.
+
+    .. versionchanged:: 2.0
+        ``encoding`` will be removed in Flask 2.1. The data must be a
+        string or UTF-8 bytes.
+
+    .. versionchanged:: 1.0.3
+        ``app`` can be passed directly, rather than requiring an app
+        context for configuration.
+    """
+    if current_app:
+        return current_app.json.loads(s, **kwargs)
+
+    return _json.loads(s, **kwargs)
+
+
+def load(fp: t.IO[t.AnyStr], **kwargs: t.Any) -> t.Any:
+    """Deserialize data as JSON read from a file.
+
+    If :data:`~flask.current_app` is available, it will use its
+    :meth:`app.json.load() <flask.json.provider.JSONProvider.load>`
+    method, otherwise it will use :func:`json.load`.
+
+    :param fp: A file opened for reading text or UTF-8 bytes.
+    :param kwargs: Arguments passed to the ``load`` implementation.
+
+    .. versionchanged:: 2.3
+        The ``app`` parameter was removed.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.json.load``, allowing an app to override
+        the behavior.
+
+    .. versionchanged:: 2.2
+        The ``app`` parameter will be removed in Flask 2.3.
+
+    .. versionchanged:: 2.0
+        ``encoding`` will be removed in Flask 2.1. The file must be text
+        mode, or binary mode with UTF-8 bytes.
+    """
+    if current_app:
+        return current_app.json.load(fp, **kwargs)
+
+    return _json.load(fp, **kwargs)
+
+
+def jsonify(*args: t.Any, **kwargs: t.Any) -> Response:
+    """Serialize the given arguments as JSON, and return a
+    :class:`~flask.Response` object with the ``application/json``
+    mimetype. A dict or list returned from a view will be converted to a
+    JSON response automatically without needing to call this.
+
+    This requires an active request or application context, and calls
+    :meth:`app.json.response() <flask.json.provider.JSONProvider.response>`.
+
+    In debug mode, the output is formatted with indentation to make it
+    easier to read. This may also be controlled by the provider.
+
+    Either positional or keyword arguments can be given, not both.
+    If no arguments are given, ``None`` is serialized.
+
+    :param args: A single value to serialize, or multiple values to
+        treat as a list to serialize.
+    :param kwargs: Treat as a dict to serialize.
+
+    .. versionchanged:: 2.2
+        Calls ``current_app.json.response``, allowing an app to override
+        the behavior.
+
+    .. versionchanged:: 2.0.2
+        :class:`decimal.Decimal` is supported by converting to a string.
+
+    .. versionchanged:: 0.11
+        Added support for serializing top-level arrays. This was a
+        security risk in ancient browsers. See :ref:`security-json`.
+
+    .. versionadded:: 0.2
+    """
+    return current_app.json.response(*args, **kwargs)  # type: ignore[return-value]
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b2a21f61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/provider.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/provider.cpython-312.pyc
new file mode 100644
index 00000000..f7889b0b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/provider.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/tag.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/tag.cpython-312.pyc
new file mode 100644
index 00000000..492be068
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/json/__pycache__/tag.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/provider.py b/backend/venv/lib/python3.12/site-packages/flask/json/provider.py
new file mode 100644
index 00000000..ea7e4753
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/json/provider.py
@@ -0,0 +1,215 @@
+from __future__ import annotations
+
+import dataclasses
+import decimal
+import json
+import typing as t
+import uuid
+import weakref
+from datetime import date
+
+from werkzeug.http import http_date
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from werkzeug.sansio.response import Response
+
+    from ..sansio.app import App
+
+
+class JSONProvider:
+    """A standard set of JSON operations for an application. Subclasses
+    of this can be used to customize JSON behavior or use different
+    JSON libraries.
+
+    To implement a provider for a specific library, subclass this base
+    class and implement at least :meth:`dumps` and :meth:`loads`. All
+    other methods have default implementations.
+
+    To use a different provider, either subclass ``Flask`` and set
+    :attr:`~flask.Flask.json_provider_class` to a provider class, or set
+    :attr:`app.json <flask.Flask.json>` to an instance of the class.
+
+    :param app: An application instance. This will be stored as a
+        :class:`weakref.proxy` on the :attr:`_app` attribute.
+
+    .. versionadded:: 2.2
+    """
+
+    def __init__(self, app: App) -> None:
+        self._app: App = weakref.proxy(app)
+
+    def dumps(self, obj: t.Any, **kwargs: t.Any) -> str:
+        """Serialize data as JSON.
+
+        :param obj: The data to serialize.
+        :param kwargs: May be passed to the underlying JSON library.
+        """
+        raise NotImplementedError
+
+    def dump(self, obj: t.Any, fp: t.IO[str], **kwargs: t.Any) -> None:
+        """Serialize data as JSON and write to a file.
+
+        :param obj: The data to serialize.
+        :param fp: A file opened for writing text. Should use the UTF-8
+            encoding to be valid JSON.
+        :param kwargs: May be passed to the underlying JSON library.
+        """
+        fp.write(self.dumps(obj, **kwargs))
+
+    def loads(self, s: str | bytes, **kwargs: t.Any) -> t.Any:
+        """Deserialize data as JSON.
+
+        :param s: Text or UTF-8 bytes.
+        :param kwargs: May be passed to the underlying JSON library.
+        """
+        raise NotImplementedError
+
+    def load(self, fp: t.IO[t.AnyStr], **kwargs: t.Any) -> t.Any:
+        """Deserialize data as JSON read from a file.
+
+        :param fp: A file opened for reading text or UTF-8 bytes.
+        :param kwargs: May be passed to the underlying JSON library.
+        """
+        return self.loads(fp.read(), **kwargs)
+
+    def _prepare_response_obj(
+        self, args: tuple[t.Any, ...], kwargs: dict[str, t.Any]
+    ) -> t.Any:
+        if args and kwargs:
+            raise TypeError("app.json.response() takes either args or kwargs, not both")
+
+        if not args and not kwargs:
+            return None
+
+        if len(args) == 1:
+            return args[0]
+
+        return args or kwargs
+
+    def response(self, *args: t.Any, **kwargs: t.Any) -> Response:
+        """Serialize the given arguments as JSON, and return a
+        :class:`~flask.Response` object with the ``application/json``
+        mimetype.
+
+        The :func:`~flask.json.jsonify` function calls this method for
+        the current application.
+
+        Either positional or keyword arguments can be given, not both.
+        If no arguments are given, ``None`` is serialized.
+
+        :param args: A single value to serialize, or multiple values to
+            treat as a list to serialize.
+        :param kwargs: Treat as a dict to serialize.
+        """
+        obj = self._prepare_response_obj(args, kwargs)
+        return self._app.response_class(self.dumps(obj), mimetype="application/json")
+
+
+def _default(o: t.Any) -> t.Any:
+    if isinstance(o, date):
+        return http_date(o)
+
+    if isinstance(o, (decimal.Decimal, uuid.UUID)):
+        return str(o)
+
+    if dataclasses and dataclasses.is_dataclass(o):
+        return dataclasses.asdict(o)  # type: ignore[arg-type]
+
+    if hasattr(o, "__html__"):
+        return str(o.__html__())
+
+    raise TypeError(f"Object of type {type(o).__name__} is not JSON serializable")
+
+
+class DefaultJSONProvider(JSONProvider):
+    """Provide JSON operations using Python's built-in :mod:`json`
+    library. Serializes the following additional data types:
+
+    -   :class:`datetime.datetime` and :class:`datetime.date` are
+        serialized to :rfc:`822` strings. This is the same as the HTTP
+        date format.
+    -   :class:`uuid.UUID` is serialized to a string.
+    -   :class:`dataclasses.dataclass` is passed to
+        :func:`dataclasses.asdict`.
+    -   :class:`~markupsafe.Markup` (or any object with a ``__html__``
+        method) will call the ``__html__`` method to get a string.
+    """
+
+    default: t.Callable[[t.Any], t.Any] = staticmethod(_default)  # type: ignore[assignment]
+    """Apply this function to any object that :meth:`json.dumps` does
+    not know how to serialize. It should return a valid JSON type or
+    raise a ``TypeError``.
+    """
+
+    ensure_ascii = True
+    """Replace non-ASCII characters with escape sequences. This may be
+    more compatible with some clients, but can be disabled for better
+    performance and size.
+    """
+
+    sort_keys = True
+    """Sort the keys in any serialized dicts. This may be useful for
+    some caching situations, but can be disabled for better performance.
+    When enabled, keys must all be strings, they are not converted
+    before sorting.
+    """
+
+    compact: bool | None = None
+    """If ``True``, or ``None`` out of debug mode, the :meth:`response`
+    output will not add indentation, newlines, or spaces. If ``False``,
+    or ``None`` in debug mode, it will use a non-compact representation.
+    """
+
+    mimetype = "application/json"
+    """The mimetype set in :meth:`response`."""
+
+    def dumps(self, obj: t.Any, **kwargs: t.Any) -> str:
+        """Serialize data as JSON to a string.
+
+        Keyword arguments are passed to :func:`json.dumps`. Sets some
+        parameter defaults from the :attr:`default`,
+        :attr:`ensure_ascii`, and :attr:`sort_keys` attributes.
+
+        :param obj: The data to serialize.
+        :param kwargs: Passed to :func:`json.dumps`.
+        """
+        kwargs.setdefault("default", self.default)
+        kwargs.setdefault("ensure_ascii", self.ensure_ascii)
+        kwargs.setdefault("sort_keys", self.sort_keys)
+        return json.dumps(obj, **kwargs)
+
+    def loads(self, s: str | bytes, **kwargs: t.Any) -> t.Any:
+        """Deserialize data as JSON from a string or bytes.
+
+        :param s: Text or UTF-8 bytes.
+        :param kwargs: Passed to :func:`json.loads`.
+        """
+        return json.loads(s, **kwargs)
+
+    def response(self, *args: t.Any, **kwargs: t.Any) -> Response:
+        """Serialize the given arguments as JSON, and return a
+        :class:`~flask.Response` object with it. The response mimetype
+        will be "application/json" and can be changed with
+        :attr:`mimetype`.
+
+        If :attr:`compact` is ``False`` or debug mode is enabled, the
+        output will be formatted to be easier to read.
+
+        Either positional or keyword arguments can be given, not both.
+        If no arguments are given, ``None`` is serialized.
+
+        :param args: A single value to serialize, or multiple values to
+            treat as a list to serialize.
+        :param kwargs: Treat as a dict to serialize.
+        """
+        obj = self._prepare_response_obj(args, kwargs)
+        dump_args: dict[str, t.Any] = {}
+
+        if (self.compact is None and self._app.debug) or self.compact is False:
+            dump_args.setdefault("indent", 2)
+        else:
+            dump_args.setdefault("separators", (",", ":"))
+
+        return self._app.response_class(
+            f"{self.dumps(obj, **dump_args)}\n", mimetype=self.mimetype
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/flask/json/tag.py b/backend/venv/lib/python3.12/site-packages/flask/json/tag.py
new file mode 100644
index 00000000..8dc3629b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/json/tag.py
@@ -0,0 +1,327 @@
+"""
+Tagged JSON
+~~~~~~~~~~~
+
+A compact representation for lossless serialization of non-standard JSON
+types. :class:`~flask.sessions.SecureCookieSessionInterface` uses this
+to serialize the session data, but it may be useful in other places. It
+can be extended to support other types.
+
+.. autoclass:: TaggedJSONSerializer
+    :members:
+
+.. autoclass:: JSONTag
+    :members:
+
+Let's see an example that adds support for
+:class:`~collections.OrderedDict`. Dicts don't have an order in JSON, so
+to handle this we will dump the items as a list of ``[key, value]``
+pairs. Subclass :class:`JSONTag` and give it the new key ``' od'`` to
+identify the type. The session serializer processes dicts first, so
+insert the new tag at the front of the order since ``OrderedDict`` must
+be processed before ``dict``.
+
+.. code-block:: python
+
+    from flask.json.tag import JSONTag
+
+    class TagOrderedDict(JSONTag):
+        __slots__ = ('serializer',)
+        key = ' od'
+
+        def check(self, value):
+            return isinstance(value, OrderedDict)
+
+        def to_json(self, value):
+            return [[k, self.serializer.tag(v)] for k, v in iteritems(value)]
+
+        def to_python(self, value):
+            return OrderedDict(value)
+
+    app.session_interface.serializer.register(TagOrderedDict, index=0)
+"""
+
+from __future__ import annotations
+
+import typing as t
+from base64 import b64decode
+from base64 import b64encode
+from datetime import datetime
+from uuid import UUID
+
+from markupsafe import Markup
+from werkzeug.http import http_date
+from werkzeug.http import parse_date
+
+from ..json import dumps
+from ..json import loads
+
+
+class JSONTag:
+    """Base class for defining type tags for :class:`TaggedJSONSerializer`."""
+
+    __slots__ = ("serializer",)
+
+    #: The tag to mark the serialized object with. If empty, this tag is
+    #: only used as an intermediate step during tagging.
+    key: str = ""
+
+    def __init__(self, serializer: TaggedJSONSerializer) -> None:
+        """Create a tagger for the given serializer."""
+        self.serializer = serializer
+
+    def check(self, value: t.Any) -> bool:
+        """Check if the given value should be tagged by this tag."""
+        raise NotImplementedError
+
+    def to_json(self, value: t.Any) -> t.Any:
+        """Convert the Python object to an object that is a valid JSON type.
+        The tag will be added later."""
+        raise NotImplementedError
+
+    def to_python(self, value: t.Any) -> t.Any:
+        """Convert the JSON representation back to the correct type. The tag
+        will already be removed."""
+        raise NotImplementedError
+
+    def tag(self, value: t.Any) -> dict[str, t.Any]:
+        """Convert the value to a valid JSON type and add the tag structure
+        around it."""
+        return {self.key: self.to_json(value)}
+
+
+class TagDict(JSONTag):
+    """Tag for 1-item dicts whose only key matches a registered tag.
+
+    Internally, the dict key is suffixed with `__`, and the suffix is removed
+    when deserializing.
+    """
+
+    __slots__ = ()
+    key = " di"
+
+    def check(self, value: t.Any) -> bool:
+        return (
+            isinstance(value, dict)
+            and len(value) == 1
+            and next(iter(value)) in self.serializer.tags
+        )
+
+    def to_json(self, value: t.Any) -> t.Any:
+        key = next(iter(value))
+        return {f"{key}__": self.serializer.tag(value[key])}
+
+    def to_python(self, value: t.Any) -> t.Any:
+        key = next(iter(value))
+        return {key[:-2]: value[key]}
+
+
+class PassDict(JSONTag):
+    __slots__ = ()
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, dict)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        # JSON objects may only have string keys, so don't bother tagging the
+        # key here.
+        return {k: self.serializer.tag(v) for k, v in value.items()}
+
+    tag = to_json
+
+
+class TagTuple(JSONTag):
+    __slots__ = ()
+    key = " t"
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, tuple)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return [self.serializer.tag(item) for item in value]
+
+    def to_python(self, value: t.Any) -> t.Any:
+        return tuple(value)
+
+
+class PassList(JSONTag):
+    __slots__ = ()
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, list)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return [self.serializer.tag(item) for item in value]
+
+    tag = to_json
+
+
+class TagBytes(JSONTag):
+    __slots__ = ()
+    key = " b"
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, bytes)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return b64encode(value).decode("ascii")
+
+    def to_python(self, value: t.Any) -> t.Any:
+        return b64decode(value)
+
+
+class TagMarkup(JSONTag):
+    """Serialize anything matching the :class:`~markupsafe.Markup` API by
+    having a ``__html__`` method to the result of that method. Always
+    deserializes to an instance of :class:`~markupsafe.Markup`."""
+
+    __slots__ = ()
+    key = " m"
+
+    def check(self, value: t.Any) -> bool:
+        return callable(getattr(value, "__html__", None))
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return str(value.__html__())
+
+    def to_python(self, value: t.Any) -> t.Any:
+        return Markup(value)
+
+
+class TagUUID(JSONTag):
+    __slots__ = ()
+    key = " u"
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, UUID)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return value.hex
+
+    def to_python(self, value: t.Any) -> t.Any:
+        return UUID(value)
+
+
+class TagDateTime(JSONTag):
+    __slots__ = ()
+    key = " d"
+
+    def check(self, value: t.Any) -> bool:
+        return isinstance(value, datetime)
+
+    def to_json(self, value: t.Any) -> t.Any:
+        return http_date(value)
+
+    def to_python(self, value: t.Any) -> t.Any:
+        return parse_date(value)
+
+
+class TaggedJSONSerializer:
+    """Serializer that uses a tag system to compactly represent objects that
+    are not JSON types. Passed as the intermediate serializer to
+    :class:`itsdangerous.Serializer`.
+
+    The following extra types are supported:
+
+    * :class:`dict`
+    * :class:`tuple`
+    * :class:`bytes`
+    * :class:`~markupsafe.Markup`
+    * :class:`~uuid.UUID`
+    * :class:`~datetime.datetime`
+    """
+
+    __slots__ = ("tags", "order")
+
+    #: Tag classes to bind when creating the serializer. Other tags can be
+    #: added later using :meth:`~register`.
+    default_tags = [
+        TagDict,
+        PassDict,
+        TagTuple,
+        PassList,
+        TagBytes,
+        TagMarkup,
+        TagUUID,
+        TagDateTime,
+    ]
+
+    def __init__(self) -> None:
+        self.tags: dict[str, JSONTag] = {}
+        self.order: list[JSONTag] = []
+
+        for cls in self.default_tags:
+            self.register(cls)
+
+    def register(
+        self,
+        tag_class: type[JSONTag],
+        force: bool = False,
+        index: int | None = None,
+    ) -> None:
+        """Register a new tag with this serializer.
+
+        :param tag_class: tag class to register. Will be instantiated with this
+            serializer instance.
+        :param force: overwrite an existing tag. If false (default), a
+            :exc:`KeyError` is raised.
+        :param index: index to insert the new tag in the tag order. Useful when
+            the new tag is a special case of an existing tag. If ``None``
+            (default), the tag is appended to the end of the order.
+
+        :raise KeyError: if the tag key is already registered and ``force`` is
+            not true.
+        """
+        tag = tag_class(self)
+        key = tag.key
+
+        if key:
+            if not force and key in self.tags:
+                raise KeyError(f"Tag '{key}' is already registered.")
+
+            self.tags[key] = tag
+
+        if index is None:
+            self.order.append(tag)
+        else:
+            self.order.insert(index, tag)
+
+    def tag(self, value: t.Any) -> t.Any:
+        """Convert a value to a tagged representation if necessary."""
+        for tag in self.order:
+            if tag.check(value):
+                return tag.tag(value)
+
+        return value
+
+    def untag(self, value: dict[str, t.Any]) -> t.Any:
+        """Convert a tagged representation back to the original type."""
+        if len(value) != 1:
+            return value
+
+        key = next(iter(value))
+
+        if key not in self.tags:
+            return value
+
+        return self.tags[key].to_python(value[key])
+
+    def _untag_scan(self, value: t.Any) -> t.Any:
+        if isinstance(value, dict):
+            # untag each item recursively
+            value = {k: self._untag_scan(v) for k, v in value.items()}
+            # untag the dict itself
+            value = self.untag(value)
+        elif isinstance(value, list):
+            # untag each item recursively
+            value = [self._untag_scan(item) for item in value]
+
+        return value
+
+    def dumps(self, value: t.Any) -> str:
+        """Tag the value and dump it to a compact JSON string."""
+        return dumps(self.tag(value), separators=(",", ":"))
+
+    def loads(self, value: str) -> t.Any:
+        """Load data from a JSON string and deserialized any tagged objects."""
+        return self._untag_scan(loads(value))
diff --git a/backend/venv/lib/python3.12/site-packages/flask/logging.py b/backend/venv/lib/python3.12/site-packages/flask/logging.py
new file mode 100644
index 00000000..0cb8f437
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/logging.py
@@ -0,0 +1,79 @@
+from __future__ import annotations
+
+import logging
+import sys
+import typing as t
+
+from werkzeug.local import LocalProxy
+
+from .globals import request
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .sansio.app import App
+
+
+@LocalProxy
+def wsgi_errors_stream() -> t.TextIO:
+    """Find the most appropriate error stream for the application. If a request
+    is active, log to ``wsgi.errors``, otherwise use ``sys.stderr``.
+
+    If you configure your own :class:`logging.StreamHandler`, you may want to
+    use this for the stream. If you are using file or dict configuration and
+    can't import this directly, you can refer to it as
+    ``ext://flask.logging.wsgi_errors_stream``.
+    """
+    if request:
+        return request.environ["wsgi.errors"]  # type: ignore[no-any-return]
+
+    return sys.stderr
+
+
+def has_level_handler(logger: logging.Logger) -> bool:
+    """Check if there is a handler in the logging chain that will handle the
+    given logger's :meth:`effective level <~logging.Logger.getEffectiveLevel>`.
+    """
+    level = logger.getEffectiveLevel()
+    current = logger
+
+    while current:
+        if any(handler.level <= level for handler in current.handlers):
+            return True
+
+        if not current.propagate:
+            break
+
+        current = current.parent  # type: ignore
+
+    return False
+
+
+#: Log messages to :func:`~flask.logging.wsgi_errors_stream` with the format
+#: ``[%(asctime)s] %(levelname)s in %(module)s: %(message)s``.
+default_handler = logging.StreamHandler(wsgi_errors_stream)  # type: ignore
+default_handler.setFormatter(
+    logging.Formatter("[%(asctime)s] %(levelname)s in %(module)s: %(message)s")
+)
+
+
+def create_logger(app: App) -> logging.Logger:
+    """Get the Flask app's logger and configure it if needed.
+
+    The logger name will be the same as
+    :attr:`app.import_name <flask.Flask.name>`.
+
+    When :attr:`~flask.Flask.debug` is enabled, set the logger level to
+    :data:`logging.DEBUG` if it is not set.
+
+    If there is no handler for the logger's effective level, add a
+    :class:`~logging.StreamHandler` for
+    :func:`~flask.logging.wsgi_errors_stream` with a basic format.
+    """
+    logger = logging.getLogger(app.name)
+
+    if app.debug and not logger.level:
+        logger.setLevel(logging.DEBUG)
+
+    if not has_level_handler(logger):
+        logger.addHandler(default_handler)
+
+    return logger
diff --git a/backend/venv/lib/python3.12/site-packages/flask/py.typed b/backend/venv/lib/python3.12/site-packages/flask/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/README.md b/backend/venv/lib/python3.12/site-packages/flask/sansio/README.md
new file mode 100644
index 00000000..623ac198
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/sansio/README.md
@@ -0,0 +1,6 @@
+# Sansio
+
+This folder contains code that can be used by alternative Flask
+implementations, for example Quart. The code therefore cannot do any
+IO, nor be part of a likely IO path. Finally this code cannot use the
+Flask globals.
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/app.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/app.cpython-312.pyc
new file mode 100644
index 00000000..8ba4513b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/app.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/blueprints.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/blueprints.cpython-312.pyc
new file mode 100644
index 00000000..db7c53c5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/blueprints.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/scaffold.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/scaffold.cpython-312.pyc
new file mode 100644
index 00000000..d996803f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask/sansio/__pycache__/scaffold.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/app.py b/backend/venv/lib/python3.12/site-packages/flask/sansio/app.py
new file mode 100644
index 00000000..36201714
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/sansio/app.py
@@ -0,0 +1,964 @@
+from __future__ import annotations
+
+import logging
+import os
+import sys
+import typing as t
+from datetime import timedelta
+from itertools import chain
+
+from werkzeug.exceptions import Aborter
+from werkzeug.exceptions import BadRequest
+from werkzeug.exceptions import BadRequestKeyError
+from werkzeug.routing import BuildError
+from werkzeug.routing import Map
+from werkzeug.routing import Rule
+from werkzeug.sansio.response import Response
+from werkzeug.utils import cached_property
+from werkzeug.utils import redirect as _wz_redirect
+
+from .. import typing as ft
+from ..config import Config
+from ..config import ConfigAttribute
+from ..ctx import _AppCtxGlobals
+from ..helpers import _split_blueprint_path
+from ..helpers import get_debug_flag
+from ..json.provider import DefaultJSONProvider
+from ..json.provider import JSONProvider
+from ..logging import create_logger
+from ..templating import DispatchingJinjaLoader
+from ..templating import Environment
+from .scaffold import _endpoint_from_view_func
+from .scaffold import find_package
+from .scaffold import Scaffold
+from .scaffold import setupmethod
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from werkzeug.wrappers import Response as BaseResponse
+
+    from ..testing import FlaskClient
+    from ..testing import FlaskCliRunner
+    from .blueprints import Blueprint
+
+T_shell_context_processor = t.TypeVar(
+    "T_shell_context_processor", bound=ft.ShellContextProcessorCallable
+)
+T_teardown = t.TypeVar("T_teardown", bound=ft.TeardownCallable)
+T_template_filter = t.TypeVar("T_template_filter", bound=ft.TemplateFilterCallable)
+T_template_global = t.TypeVar("T_template_global", bound=ft.TemplateGlobalCallable)
+T_template_test = t.TypeVar("T_template_test", bound=ft.TemplateTestCallable)
+
+
+def _make_timedelta(value: timedelta | int | None) -> timedelta | None:
+    if value is None or isinstance(value, timedelta):
+        return value
+
+    return timedelta(seconds=value)
+
+
+class App(Scaffold):
+    """The flask object implements a WSGI application and acts as the central
+    object.  It is passed the name of the module or package of the
+    application.  Once it is created it will act as a central registry for
+    the view functions, the URL rules, template configuration and much more.
+
+    The name of the package is used to resolve resources from inside the
+    package or the folder the module is contained in depending on if the
+    package parameter resolves to an actual python package (a folder with
+    an :file:`__init__.py` file inside) or a standard module (just a ``.py`` file).
+
+    For more information about resource loading, see :func:`open_resource`.
+
+    Usually you create a :class:`Flask` instance in your main module or
+    in the :file:`__init__.py` file of your package like this::
+
+        from flask import Flask
+        app = Flask(__name__)
+
+    .. admonition:: About the First Parameter
+
+        The idea of the first parameter is to give Flask an idea of what
+        belongs to your application.  This name is used to find resources
+        on the filesystem, can be used by extensions to improve debugging
+        information and a lot more.
+
+        So it's important what you provide there.  If you are using a single
+        module, `__name__` is always the correct value.  If you however are
+        using a package, it's usually recommended to hardcode the name of
+        your package there.
+
+        For example if your application is defined in :file:`yourapplication/app.py`
+        you should create it with one of the two versions below::
+
+            app = Flask('yourapplication')
+            app = Flask(__name__.split('.')[0])
+
+        Why is that?  The application will work even with `__name__`, thanks
+        to how resources are looked up.  However it will make debugging more
+        painful.  Certain extensions can make assumptions based on the
+        import name of your application.  For example the Flask-SQLAlchemy
+        extension will look for the code in your application that triggered
+        an SQL query in debug mode.  If the import name is not properly set
+        up, that debugging information is lost.  (For example it would only
+        pick up SQL queries in `yourapplication.app` and not
+        `yourapplication.views.frontend`)
+
+    .. versionadded:: 0.7
+       The `static_url_path`, `static_folder`, and `template_folder`
+       parameters were added.
+
+    .. versionadded:: 0.8
+       The `instance_path` and `instance_relative_config` parameters were
+       added.
+
+    .. versionadded:: 0.11
+       The `root_path` parameter was added.
+
+    .. versionadded:: 1.0
+       The ``host_matching`` and ``static_host`` parameters were added.
+
+    .. versionadded:: 1.0
+       The ``subdomain_matching`` parameter was added. Subdomain
+       matching needs to be enabled manually now. Setting
+       :data:`SERVER_NAME` does not implicitly enable it.
+
+    :param import_name: the name of the application package
+    :param static_url_path: can be used to specify a different path for the
+                            static files on the web.  Defaults to the name
+                            of the `static_folder` folder.
+    :param static_folder: The folder with static files that is served at
+        ``static_url_path``. Relative to the application ``root_path``
+        or an absolute path. Defaults to ``'static'``.
+    :param static_host: the host to use when adding the static route.
+        Defaults to None. Required when using ``host_matching=True``
+        with a ``static_folder`` configured.
+    :param host_matching: set ``url_map.host_matching`` attribute.
+        Defaults to False.
+    :param subdomain_matching: consider the subdomain relative to
+        :data:`SERVER_NAME` when matching routes. Defaults to False.
+    :param template_folder: the folder that contains the templates that should
+                            be used by the application.  Defaults to
+                            ``'templates'`` folder in the root path of the
+                            application.
+    :param instance_path: An alternative instance path for the application.
+                          By default the folder ``'instance'`` next to the
+                          package or module is assumed to be the instance
+                          path.
+    :param instance_relative_config: if set to ``True`` relative filenames
+                                     for loading the config are assumed to
+                                     be relative to the instance path instead
+                                     of the application root.
+    :param root_path: The path to the root of the application files.
+        This should only be set manually when it can't be detected
+        automatically, such as for namespace packages.
+    """
+
+    #: The class of the object assigned to :attr:`aborter`, created by
+    #: :meth:`create_aborter`. That object is called by
+    #: :func:`flask.abort` to raise HTTP errors, and can be
+    #: called directly as well.
+    #:
+    #: Defaults to :class:`werkzeug.exceptions.Aborter`.
+    #:
+    #: .. versionadded:: 2.2
+    aborter_class = Aborter
+
+    #: The class that is used for the Jinja environment.
+    #:
+    #: .. versionadded:: 0.11
+    jinja_environment = Environment
+
+    #: The class that is used for the :data:`~flask.g` instance.
+    #:
+    #: Example use cases for a custom class:
+    #:
+    #: 1. Store arbitrary attributes on flask.g.
+    #: 2. Add a property for lazy per-request database connectors.
+    #: 3. Return None instead of AttributeError on unexpected attributes.
+    #: 4. Raise exception if an unexpected attr is set, a "controlled" flask.g.
+    #:
+    #: In Flask 0.9 this property was called `request_globals_class` but it
+    #: was changed in 0.10 to :attr:`app_ctx_globals_class` because the
+    #: flask.g object is now application context scoped.
+    #:
+    #: .. versionadded:: 0.10
+    app_ctx_globals_class = _AppCtxGlobals
+
+    #: The class that is used for the ``config`` attribute of this app.
+    #: Defaults to :class:`~flask.Config`.
+    #:
+    #: Example use cases for a custom class:
+    #:
+    #: 1. Default values for certain config options.
+    #: 2. Access to config values through attributes in addition to keys.
+    #:
+    #: .. versionadded:: 0.11
+    config_class = Config
+
+    #: The testing flag.  Set this to ``True`` to enable the test mode of
+    #: Flask extensions (and in the future probably also Flask itself).
+    #: For example this might activate test helpers that have an
+    #: additional runtime cost which should not be enabled by default.
+    #:
+    #: If this is enabled and PROPAGATE_EXCEPTIONS is not changed from the
+    #: default it's implicitly enabled.
+    #:
+    #: This attribute can also be configured from the config with the
+    #: ``TESTING`` configuration key.  Defaults to ``False``.
+    testing = ConfigAttribute[bool]("TESTING")
+
+    #: If a secret key is set, cryptographic components can use this to
+    #: sign cookies and other things. Set this to a complex random value
+    #: when you want to use the secure cookie for instance.
+    #:
+    #: This attribute can also be configured from the config with the
+    #: :data:`SECRET_KEY` configuration key. Defaults to ``None``.
+    secret_key = ConfigAttribute[t.Union[str, bytes, None]]("SECRET_KEY")
+
+    #: A :class:`~datetime.timedelta` which is used to set the expiration
+    #: date of a permanent session.  The default is 31 days which makes a
+    #: permanent session survive for roughly one month.
+    #:
+    #: This attribute can also be configured from the config with the
+    #: ``PERMANENT_SESSION_LIFETIME`` configuration key.  Defaults to
+    #: ``timedelta(days=31)``
+    permanent_session_lifetime = ConfigAttribute[timedelta](
+        "PERMANENT_SESSION_LIFETIME",
+        get_converter=_make_timedelta,  # type: ignore[arg-type]
+    )
+
+    json_provider_class: type[JSONProvider] = DefaultJSONProvider
+    """A subclass of :class:`~flask.json.provider.JSONProvider`. An
+    instance is created and assigned to :attr:`app.json` when creating
+    the app.
+
+    The default, :class:`~flask.json.provider.DefaultJSONProvider`, uses
+    Python's built-in :mod:`json` library. A different provider can use
+    a different JSON library.
+
+    .. versionadded:: 2.2
+    """
+
+    #: Options that are passed to the Jinja environment in
+    #: :meth:`create_jinja_environment`. Changing these options after
+    #: the environment is created (accessing :attr:`jinja_env`) will
+    #: have no effect.
+    #:
+    #: .. versionchanged:: 1.1.0
+    #:     This is a ``dict`` instead of an ``ImmutableDict`` to allow
+    #:     easier configuration.
+    #:
+    jinja_options: dict[str, t.Any] = {}
+
+    #: The rule object to use for URL rules created.  This is used by
+    #: :meth:`add_url_rule`.  Defaults to :class:`werkzeug.routing.Rule`.
+    #:
+    #: .. versionadded:: 0.7
+    url_rule_class = Rule
+
+    #: The map object to use for storing the URL rules and routing
+    #: configuration parameters. Defaults to :class:`werkzeug.routing.Map`.
+    #:
+    #: .. versionadded:: 1.1.0
+    url_map_class = Map
+
+    #: The :meth:`test_client` method creates an instance of this test
+    #: client class. Defaults to :class:`~flask.testing.FlaskClient`.
+    #:
+    #: .. versionadded:: 0.7
+    test_client_class: type[FlaskClient] | None = None
+
+    #: The :class:`~click.testing.CliRunner` subclass, by default
+    #: :class:`~flask.testing.FlaskCliRunner` that is used by
+    #: :meth:`test_cli_runner`. Its ``__init__`` method should take a
+    #: Flask app object as the first argument.
+    #:
+    #: .. versionadded:: 1.0
+    test_cli_runner_class: type[FlaskCliRunner] | None = None
+
+    default_config: dict[str, t.Any]
+    response_class: type[Response]
+
+    def __init__(
+        self,
+        import_name: str,
+        static_url_path: str | None = None,
+        static_folder: str | os.PathLike[str] | None = "static",
+        static_host: str | None = None,
+        host_matching: bool = False,
+        subdomain_matching: bool = False,
+        template_folder: str | os.PathLike[str] | None = "templates",
+        instance_path: str | None = None,
+        instance_relative_config: bool = False,
+        root_path: str | None = None,
+    ) -> None:
+        super().__init__(
+            import_name=import_name,
+            static_folder=static_folder,
+            static_url_path=static_url_path,
+            template_folder=template_folder,
+            root_path=root_path,
+        )
+
+        if instance_path is None:
+            instance_path = self.auto_find_instance_path()
+        elif not os.path.isabs(instance_path):
+            raise ValueError(
+                "If an instance path is provided it must be absolute."
+                " A relative path was given instead."
+            )
+
+        #: Holds the path to the instance folder.
+        #:
+        #: .. versionadded:: 0.8
+        self.instance_path = instance_path
+
+        #: The configuration dictionary as :class:`Config`.  This behaves
+        #: exactly like a regular dictionary but supports additional methods
+        #: to load a config from files.
+        self.config = self.make_config(instance_relative_config)
+
+        #: An instance of :attr:`aborter_class` created by
+        #: :meth:`make_aborter`. This is called by :func:`flask.abort`
+        #: to raise HTTP errors, and can be called directly as well.
+        #:
+        #: .. versionadded:: 2.2
+        #:     Moved from ``flask.abort``, which calls this object.
+        self.aborter = self.make_aborter()
+
+        self.json: JSONProvider = self.json_provider_class(self)
+        """Provides access to JSON methods. Functions in ``flask.json``
+        will call methods on this provider when the application context
+        is active. Used for handling JSON requests and responses.
+
+        An instance of :attr:`json_provider_class`. Can be customized by
+        changing that attribute on a subclass, or by assigning to this
+        attribute afterwards.
+
+        The default, :class:`~flask.json.provider.DefaultJSONProvider`,
+        uses Python's built-in :mod:`json` library. A different provider
+        can use a different JSON library.
+
+        .. versionadded:: 2.2
+        """
+
+        #: A list of functions that are called by
+        #: :meth:`handle_url_build_error` when :meth:`.url_for` raises a
+        #: :exc:`~werkzeug.routing.BuildError`. Each function is called
+        #: with ``error``, ``endpoint`` and ``values``. If a function
+        #: returns ``None`` or raises a ``BuildError``, it is skipped.
+        #: Otherwise, its return value is returned by ``url_for``.
+        #:
+        #: .. versionadded:: 0.9
+        self.url_build_error_handlers: list[
+            t.Callable[[Exception, str, dict[str, t.Any]], str]
+        ] = []
+
+        #: A list of functions that are called when the application context
+        #: is destroyed.  Since the application context is also torn down
+        #: if the request ends this is the place to store code that disconnects
+        #: from databases.
+        #:
+        #: .. versionadded:: 0.9
+        self.teardown_appcontext_funcs: list[ft.TeardownCallable] = []
+
+        #: A list of shell context processor functions that should be run
+        #: when a shell context is created.
+        #:
+        #: .. versionadded:: 0.11
+        self.shell_context_processors: list[ft.ShellContextProcessorCallable] = []
+
+        #: Maps registered blueprint names to blueprint objects. The
+        #: dict retains the order the blueprints were registered in.
+        #: Blueprints can be registered multiple times, this dict does
+        #: not track how often they were attached.
+        #:
+        #: .. versionadded:: 0.7
+        self.blueprints: dict[str, Blueprint] = {}
+
+        #: a place where extensions can store application specific state.  For
+        #: example this is where an extension could store database engines and
+        #: similar things.
+        #:
+        #: The key must match the name of the extension module. For example in
+        #: case of a "Flask-Foo" extension in `flask_foo`, the key would be
+        #: ``'foo'``.
+        #:
+        #: .. versionadded:: 0.7
+        self.extensions: dict[str, t.Any] = {}
+
+        #: The :class:`~werkzeug.routing.Map` for this instance.  You can use
+        #: this to change the routing converters after the class was created
+        #: but before any routes are connected.  Example::
+        #:
+        #:    from werkzeug.routing import BaseConverter
+        #:
+        #:    class ListConverter(BaseConverter):
+        #:        def to_python(self, value):
+        #:            return value.split(',')
+        #:        def to_url(self, values):
+        #:            return ','.join(super(ListConverter, self).to_url(value)
+        #:                            for value in values)
+        #:
+        #:    app = Flask(__name__)
+        #:    app.url_map.converters['list'] = ListConverter
+        self.url_map = self.url_map_class(host_matching=host_matching)
+
+        self.subdomain_matching = subdomain_matching
+
+        # tracks internally if the application already handled at least one
+        # request.
+        self._got_first_request = False
+
+    def _check_setup_finished(self, f_name: str) -> None:
+        if self._got_first_request:
+            raise AssertionError(
+                f"The setup method '{f_name}' can no longer be called"
+                " on the application. It has already handled its first"
+                " request, any changes will not be applied"
+                " consistently.\n"
+                "Make sure all imports, decorators, functions, etc."
+                " needed to set up the application are done before"
+                " running it."
+            )
+
+    @cached_property
+    def name(self) -> str:  # type: ignore
+        """The name of the application.  This is usually the import name
+        with the difference that it's guessed from the run file if the
+        import name is main.  This name is used as a display name when
+        Flask needs the name of the application.  It can be set and overridden
+        to change the value.
+
+        .. versionadded:: 0.8
+        """
+        if self.import_name == "__main__":
+            fn: str | None = getattr(sys.modules["__main__"], "__file__", None)
+            if fn is None:
+                return "__main__"
+            return os.path.splitext(os.path.basename(fn))[0]
+        return self.import_name
+
+    @cached_property
+    def logger(self) -> logging.Logger:
+        """A standard Python :class:`~logging.Logger` for the app, with
+        the same name as :attr:`name`.
+
+        In debug mode, the logger's :attr:`~logging.Logger.level` will
+        be set to :data:`~logging.DEBUG`.
+
+        If there are no handlers configured, a default handler will be
+        added. See :doc:`/logging` for more information.
+
+        .. versionchanged:: 1.1.0
+            The logger takes the same name as :attr:`name` rather than
+            hard-coding ``"flask.app"``.
+
+        .. versionchanged:: 1.0.0
+            Behavior was simplified. The logger is always named
+            ``"flask.app"``. The level is only set during configuration,
+            it doesn't check ``app.debug`` each time. Only one format is
+            used, not different ones depending on ``app.debug``. No
+            handlers are removed, and a handler is only added if no
+            handlers are already configured.
+
+        .. versionadded:: 0.3
+        """
+        return create_logger(self)
+
+    @cached_property
+    def jinja_env(self) -> Environment:
+        """The Jinja environment used to load templates.
+
+        The environment is created the first time this property is
+        accessed. Changing :attr:`jinja_options` after that will have no
+        effect.
+        """
+        return self.create_jinja_environment()
+
+    def create_jinja_environment(self) -> Environment:
+        raise NotImplementedError()
+
+    def make_config(self, instance_relative: bool = False) -> Config:
+        """Used to create the config attribute by the Flask constructor.
+        The `instance_relative` parameter is passed in from the constructor
+        of Flask (there named `instance_relative_config`) and indicates if
+        the config should be relative to the instance path or the root path
+        of the application.
+
+        .. versionadded:: 0.8
+        """
+        root_path = self.root_path
+        if instance_relative:
+            root_path = self.instance_path
+        defaults = dict(self.default_config)
+        defaults["DEBUG"] = get_debug_flag()
+        return self.config_class(root_path, defaults)
+
+    def make_aborter(self) -> Aborter:
+        """Create the object to assign to :attr:`aborter`. That object
+        is called by :func:`flask.abort` to raise HTTP errors, and can
+        be called directly as well.
+
+        By default, this creates an instance of :attr:`aborter_class`,
+        which defaults to :class:`werkzeug.exceptions.Aborter`.
+
+        .. versionadded:: 2.2
+        """
+        return self.aborter_class()
+
+    def auto_find_instance_path(self) -> str:
+        """Tries to locate the instance path if it was not provided to the
+        constructor of the application class.  It will basically calculate
+        the path to a folder named ``instance`` next to your main file or
+        the package.
+
+        .. versionadded:: 0.8
+        """
+        prefix, package_path = find_package(self.import_name)
+        if prefix is None:
+            return os.path.join(package_path, "instance")
+        return os.path.join(prefix, "var", f"{self.name}-instance")
+
+    def create_global_jinja_loader(self) -> DispatchingJinjaLoader:
+        """Creates the loader for the Jinja2 environment.  Can be used to
+        override just the loader and keeping the rest unchanged.  It's
+        discouraged to override this function.  Instead one should override
+        the :meth:`jinja_loader` function instead.
+
+        The global loader dispatches between the loaders of the application
+        and the individual blueprints.
+
+        .. versionadded:: 0.7
+        """
+        return DispatchingJinjaLoader(self)
+
+    def select_jinja_autoescape(self, filename: str) -> bool:
+        """Returns ``True`` if autoescaping should be active for the given
+        template name. If no template name is given, returns `True`.
+
+        .. versionchanged:: 2.2
+            Autoescaping is now enabled by default for ``.svg`` files.
+
+        .. versionadded:: 0.5
+        """
+        if filename is None:
+            return True
+        return filename.endswith((".html", ".htm", ".xml", ".xhtml", ".svg"))
+
+    @property
+    def debug(self) -> bool:
+        """Whether debug mode is enabled. When using ``flask run`` to start the
+        development server, an interactive debugger will be shown for unhandled
+        exceptions, and the server will be reloaded when code changes. This maps to the
+        :data:`DEBUG` config key. It may not behave as expected if set late.
+
+        **Do not enable debug mode when deploying in production.**
+
+        Default: ``False``
+        """
+        return self.config["DEBUG"]  # type: ignore[no-any-return]
+
+    @debug.setter
+    def debug(self, value: bool) -> None:
+        self.config["DEBUG"] = value
+
+        if self.config["TEMPLATES_AUTO_RELOAD"] is None:
+            self.jinja_env.auto_reload = value
+
+    @setupmethod
+    def register_blueprint(self, blueprint: Blueprint, **options: t.Any) -> None:
+        """Register a :class:`~flask.Blueprint` on the application. Keyword
+        arguments passed to this method will override the defaults set on the
+        blueprint.
+
+        Calls the blueprint's :meth:`~flask.Blueprint.register` method after
+        recording the blueprint in the application's :attr:`blueprints`.
+
+        :param blueprint: The blueprint to register.
+        :param url_prefix: Blueprint routes will be prefixed with this.
+        :param subdomain: Blueprint routes will match on this subdomain.
+        :param url_defaults: Blueprint routes will use these default values for
+            view arguments.
+        :param options: Additional keyword arguments are passed to
+            :class:`~flask.blueprints.BlueprintSetupState`. They can be
+            accessed in :meth:`~flask.Blueprint.record` callbacks.
+
+        .. versionchanged:: 2.0.1
+            The ``name`` option can be used to change the (pre-dotted)
+            name the blueprint is registered with. This allows the same
+            blueprint to be registered multiple times with unique names
+            for ``url_for``.
+
+        .. versionadded:: 0.7
+        """
+        blueprint.register(self, options)
+
+    def iter_blueprints(self) -> t.ValuesView[Blueprint]:
+        """Iterates over all blueprints by the order they were registered.
+
+        .. versionadded:: 0.11
+        """
+        return self.blueprints.values()
+
+    @setupmethod
+    def add_url_rule(
+        self,
+        rule: str,
+        endpoint: str | None = None,
+        view_func: ft.RouteCallable | None = None,
+        provide_automatic_options: bool | None = None,
+        **options: t.Any,
+    ) -> None:
+        if endpoint is None:
+            endpoint = _endpoint_from_view_func(view_func)  # type: ignore
+        options["endpoint"] = endpoint
+        methods = options.pop("methods", None)
+
+        # if the methods are not given and the view_func object knows its
+        # methods we can use that instead.  If neither exists, we go with
+        # a tuple of only ``GET`` as default.
+        if methods is None:
+            methods = getattr(view_func, "methods", None) or ("GET",)
+        if isinstance(methods, str):
+            raise TypeError(
+                "Allowed methods must be a list of strings, for"
+                ' example: @app.route(..., methods=["POST"])'
+            )
+        methods = {item.upper() for item in methods}
+
+        # Methods that should always be added
+        required_methods: set[str] = set(getattr(view_func, "required_methods", ()))
+
+        # starting with Flask 0.8 the view_func object can disable and
+        # force-enable the automatic options handling.
+        if provide_automatic_options is None:
+            provide_automatic_options = getattr(
+                view_func, "provide_automatic_options", None
+            )
+
+        if provide_automatic_options is None:
+            if "OPTIONS" not in methods and self.config["PROVIDE_AUTOMATIC_OPTIONS"]:
+                provide_automatic_options = True
+                required_methods.add("OPTIONS")
+            else:
+                provide_automatic_options = False
+
+        # Add the required methods now.
+        methods |= required_methods
+
+        rule_obj = self.url_rule_class(rule, methods=methods, **options)
+        rule_obj.provide_automatic_options = provide_automatic_options  # type: ignore[attr-defined]
+
+        self.url_map.add(rule_obj)
+        if view_func is not None:
+            old_func = self.view_functions.get(endpoint)
+            if old_func is not None and old_func != view_func:
+                raise AssertionError(
+                    "View function mapping is overwriting an existing"
+                    f" endpoint function: {endpoint}"
+                )
+            self.view_functions[endpoint] = view_func
+
+    @setupmethod
+    def template_filter(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_filter], T_template_filter]:
+        """A decorator that is used to register custom template filter.
+        You can specify a name for the filter, otherwise the function
+        name will be used. Example::
+
+          @app.template_filter()
+          def reverse(s):
+              return s[::-1]
+
+        :param name: the optional name of the filter, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_filter) -> T_template_filter:
+            self.add_template_filter(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_template_filter(
+        self, f: ft.TemplateFilterCallable, name: str | None = None
+    ) -> None:
+        """Register a custom template filter.  Works exactly like the
+        :meth:`template_filter` decorator.
+
+        :param name: the optional name of the filter, otherwise the
+                     function name will be used.
+        """
+        self.jinja_env.filters[name or f.__name__] = f
+
+    @setupmethod
+    def template_test(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_test], T_template_test]:
+        """A decorator that is used to register custom template test.
+        You can specify a name for the test, otherwise the function
+        name will be used. Example::
+
+          @app.template_test()
+          def is_prime(n):
+              if n == 2:
+                  return True
+              for i in range(2, int(math.ceil(math.sqrt(n))) + 1):
+                  if n % i == 0:
+                      return False
+              return True
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the test, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_test) -> T_template_test:
+            self.add_template_test(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_template_test(
+        self, f: ft.TemplateTestCallable, name: str | None = None
+    ) -> None:
+        """Register a custom template test.  Works exactly like the
+        :meth:`template_test` decorator.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the test, otherwise the
+                     function name will be used.
+        """
+        self.jinja_env.tests[name or f.__name__] = f
+
+    @setupmethod
+    def template_global(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_global], T_template_global]:
+        """A decorator that is used to register a custom template global function.
+        You can specify a name for the global function, otherwise the function
+        name will be used. Example::
+
+            @app.template_global()
+            def double(n):
+                return 2 * n
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the global function, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_global) -> T_template_global:
+            self.add_template_global(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_template_global(
+        self, f: ft.TemplateGlobalCallable, name: str | None = None
+    ) -> None:
+        """Register a custom template global function. Works exactly like the
+        :meth:`template_global` decorator.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the global function, otherwise the
+                     function name will be used.
+        """
+        self.jinja_env.globals[name or f.__name__] = f
+
+    @setupmethod
+    def teardown_appcontext(self, f: T_teardown) -> T_teardown:
+        """Registers a function to be called when the application
+        context is popped. The application context is typically popped
+        after the request context for each request, at the end of CLI
+        commands, or after a manually pushed context ends.
+
+        .. code-block:: python
+
+            with app.app_context():
+                ...
+
+        When the ``with`` block exits (or ``ctx.pop()`` is called), the
+        teardown functions are called just before the app context is
+        made inactive. Since a request context typically also manages an
+        application context it would also be called when you pop a
+        request context.
+
+        When a teardown function was called because of an unhandled
+        exception it will be passed an error object. If an
+        :meth:`errorhandler` is registered, it will handle the exception
+        and the teardown will not receive it.
+
+        Teardown functions must avoid raising exceptions. If they
+        execute code that might fail they must surround that code with a
+        ``try``/``except`` block and log any errors.
+
+        The return values of teardown functions are ignored.
+
+        .. versionadded:: 0.9
+        """
+        self.teardown_appcontext_funcs.append(f)
+        return f
+
+    @setupmethod
+    def shell_context_processor(
+        self, f: T_shell_context_processor
+    ) -> T_shell_context_processor:
+        """Registers a shell context processor function.
+
+        .. versionadded:: 0.11
+        """
+        self.shell_context_processors.append(f)
+        return f
+
+    def _find_error_handler(
+        self, e: Exception, blueprints: list[str]
+    ) -> ft.ErrorHandlerCallable | None:
+        """Return a registered error handler for an exception in this order:
+        blueprint handler for a specific code, app handler for a specific code,
+        blueprint handler for an exception class, app handler for an exception
+        class, or ``None`` if a suitable handler is not found.
+        """
+        exc_class, code = self._get_exc_class_and_code(type(e))
+        names = (*blueprints, None)
+
+        for c in (code, None) if code is not None else (None,):
+            for name in names:
+                handler_map = self.error_handler_spec[name][c]
+
+                if not handler_map:
+                    continue
+
+                for cls in exc_class.__mro__:
+                    handler = handler_map.get(cls)
+
+                    if handler is not None:
+                        return handler
+        return None
+
+    def trap_http_exception(self, e: Exception) -> bool:
+        """Checks if an HTTP exception should be trapped or not.  By default
+        this will return ``False`` for all exceptions except for a bad request
+        key error if ``TRAP_BAD_REQUEST_ERRORS`` is set to ``True``.  It
+        also returns ``True`` if ``TRAP_HTTP_EXCEPTIONS`` is set to ``True``.
+
+        This is called for all HTTP exceptions raised by a view function.
+        If it returns ``True`` for any exception the error handler for this
+        exception is not called and it shows up as regular exception in the
+        traceback.  This is helpful for debugging implicitly raised HTTP
+        exceptions.
+
+        .. versionchanged:: 1.0
+            Bad request errors are not trapped by default in debug mode.
+
+        .. versionadded:: 0.8
+        """
+        if self.config["TRAP_HTTP_EXCEPTIONS"]:
+            return True
+
+        trap_bad_request = self.config["TRAP_BAD_REQUEST_ERRORS"]
+
+        # if unset, trap key errors in debug mode
+        if (
+            trap_bad_request is None
+            and self.debug
+            and isinstance(e, BadRequestKeyError)
+        ):
+            return True
+
+        if trap_bad_request:
+            return isinstance(e, BadRequest)
+
+        return False
+
+    def should_ignore_error(self, error: BaseException | None) -> bool:
+        """This is called to figure out if an error should be ignored
+        or not as far as the teardown system is concerned.  If this
+        function returns ``True`` then the teardown handlers will not be
+        passed the error.
+
+        .. versionadded:: 0.10
+        """
+        return False
+
+    def redirect(self, location: str, code: int = 302) -> BaseResponse:
+        """Create a redirect response object.
+
+        This is called by :func:`flask.redirect`, and can be called
+        directly as well.
+
+        :param location: The URL to redirect to.
+        :param code: The status code for the redirect.
+
+        .. versionadded:: 2.2
+            Moved from ``flask.redirect``, which calls this method.
+        """
+        return _wz_redirect(
+            location,
+            code=code,
+            Response=self.response_class,  # type: ignore[arg-type]
+        )
+
+    def inject_url_defaults(self, endpoint: str, values: dict[str, t.Any]) -> None:
+        """Injects the URL defaults for the given endpoint directly into
+        the values dictionary passed.  This is used internally and
+        automatically called on URL building.
+
+        .. versionadded:: 0.7
+        """
+        names: t.Iterable[str | None] = (None,)
+
+        # url_for may be called outside a request context, parse the
+        # passed endpoint instead of using request.blueprints.
+        if "." in endpoint:
+            names = chain(
+                names, reversed(_split_blueprint_path(endpoint.rpartition(".")[0]))
+            )
+
+        for name in names:
+            if name in self.url_default_functions:
+                for func in self.url_default_functions[name]:
+                    func(endpoint, values)
+
+    def handle_url_build_error(
+        self, error: BuildError, endpoint: str, values: dict[str, t.Any]
+    ) -> str:
+        """Called by :meth:`.url_for` if a
+        :exc:`~werkzeug.routing.BuildError` was raised. If this returns
+        a value, it will be returned by ``url_for``, otherwise the error
+        will be re-raised.
+
+        Each function in :attr:`url_build_error_handlers` is called with
+        ``error``, ``endpoint`` and ``values``. If a function returns
+        ``None`` or raises a ``BuildError``, it is skipped. Otherwise,
+        its return value is returned by ``url_for``.
+
+        :param error: The active ``BuildError`` being handled.
+        :param endpoint: The endpoint being built.
+        :param values: The keyword arguments passed to ``url_for``.
+        """
+        for handler in self.url_build_error_handlers:
+            try:
+                rv = handler(error, endpoint, values)
+            except BuildError as e:
+                # make error available outside except block
+                error = e
+            else:
+                if rv is not None:
+                    return rv
+
+        # Re-raise if called with an active exception, otherwise raise
+        # the passed in exception.
+        if error is sys.exc_info()[1]:
+            raise
+
+        raise error
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/blueprints.py b/backend/venv/lib/python3.12/site-packages/flask/sansio/blueprints.py
new file mode 100644
index 00000000..4f912cca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/sansio/blueprints.py
@@ -0,0 +1,632 @@
+from __future__ import annotations
+
+import os
+import typing as t
+from collections import defaultdict
+from functools import update_wrapper
+
+from .. import typing as ft
+from .scaffold import _endpoint_from_view_func
+from .scaffold import _sentinel
+from .scaffold import Scaffold
+from .scaffold import setupmethod
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .app import App
+
+DeferredSetupFunction = t.Callable[["BlueprintSetupState"], None]
+T_after_request = t.TypeVar("T_after_request", bound=ft.AfterRequestCallable[t.Any])
+T_before_request = t.TypeVar("T_before_request", bound=ft.BeforeRequestCallable)
+T_error_handler = t.TypeVar("T_error_handler", bound=ft.ErrorHandlerCallable)
+T_teardown = t.TypeVar("T_teardown", bound=ft.TeardownCallable)
+T_template_context_processor = t.TypeVar(
+    "T_template_context_processor", bound=ft.TemplateContextProcessorCallable
+)
+T_template_filter = t.TypeVar("T_template_filter", bound=ft.TemplateFilterCallable)
+T_template_global = t.TypeVar("T_template_global", bound=ft.TemplateGlobalCallable)
+T_template_test = t.TypeVar("T_template_test", bound=ft.TemplateTestCallable)
+T_url_defaults = t.TypeVar("T_url_defaults", bound=ft.URLDefaultCallable)
+T_url_value_preprocessor = t.TypeVar(
+    "T_url_value_preprocessor", bound=ft.URLValuePreprocessorCallable
+)
+
+
+class BlueprintSetupState:
+    """Temporary holder object for registering a blueprint with the
+    application.  An instance of this class is created by the
+    :meth:`~flask.Blueprint.make_setup_state` method and later passed
+    to all register callback functions.
+    """
+
+    def __init__(
+        self,
+        blueprint: Blueprint,
+        app: App,
+        options: t.Any,
+        first_registration: bool,
+    ) -> None:
+        #: a reference to the current application
+        self.app = app
+
+        #: a reference to the blueprint that created this setup state.
+        self.blueprint = blueprint
+
+        #: a dictionary with all options that were passed to the
+        #: :meth:`~flask.Flask.register_blueprint` method.
+        self.options = options
+
+        #: as blueprints can be registered multiple times with the
+        #: application and not everything wants to be registered
+        #: multiple times on it, this attribute can be used to figure
+        #: out if the blueprint was registered in the past already.
+        self.first_registration = first_registration
+
+        subdomain = self.options.get("subdomain")
+        if subdomain is None:
+            subdomain = self.blueprint.subdomain
+
+        #: The subdomain that the blueprint should be active for, ``None``
+        #: otherwise.
+        self.subdomain = subdomain
+
+        url_prefix = self.options.get("url_prefix")
+        if url_prefix is None:
+            url_prefix = self.blueprint.url_prefix
+        #: The prefix that should be used for all URLs defined on the
+        #: blueprint.
+        self.url_prefix = url_prefix
+
+        self.name = self.options.get("name", blueprint.name)
+        self.name_prefix = self.options.get("name_prefix", "")
+
+        #: A dictionary with URL defaults that is added to each and every
+        #: URL that was defined with the blueprint.
+        self.url_defaults = dict(self.blueprint.url_values_defaults)
+        self.url_defaults.update(self.options.get("url_defaults", ()))
+
+    def add_url_rule(
+        self,
+        rule: str,
+        endpoint: str | None = None,
+        view_func: ft.RouteCallable | None = None,
+        **options: t.Any,
+    ) -> None:
+        """A helper method to register a rule (and optionally a view function)
+        to the application.  The endpoint is automatically prefixed with the
+        blueprint's name.
+        """
+        if self.url_prefix is not None:
+            if rule:
+                rule = "/".join((self.url_prefix.rstrip("/"), rule.lstrip("/")))
+            else:
+                rule = self.url_prefix
+        options.setdefault("subdomain", self.subdomain)
+        if endpoint is None:
+            endpoint = _endpoint_from_view_func(view_func)  # type: ignore
+        defaults = self.url_defaults
+        if "defaults" in options:
+            defaults = dict(defaults, **options.pop("defaults"))
+
+        self.app.add_url_rule(
+            rule,
+            f"{self.name_prefix}.{self.name}.{endpoint}".lstrip("."),
+            view_func,
+            defaults=defaults,
+            **options,
+        )
+
+
+class Blueprint(Scaffold):
+    """Represents a blueprint, a collection of routes and other
+    app-related functions that can be registered on a real application
+    later.
+
+    A blueprint is an object that allows defining application functions
+    without requiring an application object ahead of time. It uses the
+    same decorators as :class:`~flask.Flask`, but defers the need for an
+    application by recording them for later registration.
+
+    Decorating a function with a blueprint creates a deferred function
+    that is called with :class:`~flask.blueprints.BlueprintSetupState`
+    when the blueprint is registered on an application.
+
+    See :doc:`/blueprints` for more information.
+
+    :param name: The name of the blueprint. Will be prepended to each
+        endpoint name.
+    :param import_name: The name of the blueprint package, usually
+        ``__name__``. This helps locate the ``root_path`` for the
+        blueprint.
+    :param static_folder: A folder with static files that should be
+        served by the blueprint's static route. The path is relative to
+        the blueprint's root path. Blueprint static files are disabled
+        by default.
+    :param static_url_path: The url to serve static files from.
+        Defaults to ``static_folder``. If the blueprint does not have
+        a ``url_prefix``, the app's static route will take precedence,
+        and the blueprint's static files won't be accessible.
+    :param template_folder: A folder with templates that should be added
+        to the app's template search path. The path is relative to the
+        blueprint's root path. Blueprint templates are disabled by
+        default. Blueprint templates have a lower precedence than those
+        in the app's templates folder.
+    :param url_prefix: A path to prepend to all of the blueprint's URLs,
+        to make them distinct from the rest of the app's routes.
+    :param subdomain: A subdomain that blueprint routes will match on by
+        default.
+    :param url_defaults: A dict of default values that blueprint routes
+        will receive by default.
+    :param root_path: By default, the blueprint will automatically set
+        this based on ``import_name``. In certain situations this
+        automatic detection can fail, so the path can be specified
+        manually instead.
+
+    .. versionchanged:: 1.1.0
+        Blueprints have a ``cli`` group to register nested CLI commands.
+        The ``cli_group`` parameter controls the name of the group under
+        the ``flask`` command.
+
+    .. versionadded:: 0.7
+    """
+
+    _got_registered_once = False
+
+    def __init__(
+        self,
+        name: str,
+        import_name: str,
+        static_folder: str | os.PathLike[str] | None = None,
+        static_url_path: str | None = None,
+        template_folder: str | os.PathLike[str] | None = None,
+        url_prefix: str | None = None,
+        subdomain: str | None = None,
+        url_defaults: dict[str, t.Any] | None = None,
+        root_path: str | None = None,
+        cli_group: str | None = _sentinel,  # type: ignore[assignment]
+    ):
+        super().__init__(
+            import_name=import_name,
+            static_folder=static_folder,
+            static_url_path=static_url_path,
+            template_folder=template_folder,
+            root_path=root_path,
+        )
+
+        if not name:
+            raise ValueError("'name' may not be empty.")
+
+        if "." in name:
+            raise ValueError("'name' may not contain a dot '.' character.")
+
+        self.name = name
+        self.url_prefix = url_prefix
+        self.subdomain = subdomain
+        self.deferred_functions: list[DeferredSetupFunction] = []
+
+        if url_defaults is None:
+            url_defaults = {}
+
+        self.url_values_defaults = url_defaults
+        self.cli_group = cli_group
+        self._blueprints: list[tuple[Blueprint, dict[str, t.Any]]] = []
+
+    def _check_setup_finished(self, f_name: str) -> None:
+        if self._got_registered_once:
+            raise AssertionError(
+                f"The setup method '{f_name}' can no longer be called on the blueprint"
+                f" '{self.name}'. It has already been registered at least once, any"
+                " changes will not be applied consistently.\n"
+                "Make sure all imports, decorators, functions, etc. needed to set up"
+                " the blueprint are done before registering it."
+            )
+
+    @setupmethod
+    def record(self, func: DeferredSetupFunction) -> None:
+        """Registers a function that is called when the blueprint is
+        registered on the application.  This function is called with the
+        state as argument as returned by the :meth:`make_setup_state`
+        method.
+        """
+        self.deferred_functions.append(func)
+
+    @setupmethod
+    def record_once(self, func: DeferredSetupFunction) -> None:
+        """Works like :meth:`record` but wraps the function in another
+        function that will ensure the function is only called once.  If the
+        blueprint is registered a second time on the application, the
+        function passed is not called.
+        """
+
+        def wrapper(state: BlueprintSetupState) -> None:
+            if state.first_registration:
+                func(state)
+
+        self.record(update_wrapper(wrapper, func))
+
+    def make_setup_state(
+        self, app: App, options: dict[str, t.Any], first_registration: bool = False
+    ) -> BlueprintSetupState:
+        """Creates an instance of :meth:`~flask.blueprints.BlueprintSetupState`
+        object that is later passed to the register callback functions.
+        Subclasses can override this to return a subclass of the setup state.
+        """
+        return BlueprintSetupState(self, app, options, first_registration)
+
+    @setupmethod
+    def register_blueprint(self, blueprint: Blueprint, **options: t.Any) -> None:
+        """Register a :class:`~flask.Blueprint` on this blueprint. Keyword
+        arguments passed to this method will override the defaults set
+        on the blueprint.
+
+        .. versionchanged:: 2.0.1
+            The ``name`` option can be used to change the (pre-dotted)
+            name the blueprint is registered with. This allows the same
+            blueprint to be registered multiple times with unique names
+            for ``url_for``.
+
+        .. versionadded:: 2.0
+        """
+        if blueprint is self:
+            raise ValueError("Cannot register a blueprint on itself")
+        self._blueprints.append((blueprint, options))
+
+    def register(self, app: App, options: dict[str, t.Any]) -> None:
+        """Called by :meth:`Flask.register_blueprint` to register all
+        views and callbacks registered on the blueprint with the
+        application. Creates a :class:`.BlueprintSetupState` and calls
+        each :meth:`record` callback with it.
+
+        :param app: The application this blueprint is being registered
+            with.
+        :param options: Keyword arguments forwarded from
+            :meth:`~Flask.register_blueprint`.
+
+        .. versionchanged:: 2.3
+            Nested blueprints now correctly apply subdomains.
+
+        .. versionchanged:: 2.1
+            Registering the same blueprint with the same name multiple
+            times is an error.
+
+        .. versionchanged:: 2.0.1
+            Nested blueprints are registered with their dotted name.
+            This allows different blueprints with the same name to be
+            nested at different locations.
+
+        .. versionchanged:: 2.0.1
+            The ``name`` option can be used to change the (pre-dotted)
+            name the blueprint is registered with. This allows the same
+            blueprint to be registered multiple times with unique names
+            for ``url_for``.
+        """
+        name_prefix = options.get("name_prefix", "")
+        self_name = options.get("name", self.name)
+        name = f"{name_prefix}.{self_name}".lstrip(".")
+
+        if name in app.blueprints:
+            bp_desc = "this" if app.blueprints[name] is self else "a different"
+            existing_at = f" '{name}'" if self_name != name else ""
+
+            raise ValueError(
+                f"The name '{self_name}' is already registered for"
+                f" {bp_desc} blueprint{existing_at}. Use 'name=' to"
+                f" provide a unique name."
+            )
+
+        first_bp_registration = not any(bp is self for bp in app.blueprints.values())
+        first_name_registration = name not in app.blueprints
+
+        app.blueprints[name] = self
+        self._got_registered_once = True
+        state = self.make_setup_state(app, options, first_bp_registration)
+
+        if self.has_static_folder:
+            state.add_url_rule(
+                f"{self.static_url_path}/<path:filename>",
+                view_func=self.send_static_file,  # type: ignore[attr-defined]
+                endpoint="static",
+            )
+
+        # Merge blueprint data into parent.
+        if first_bp_registration or first_name_registration:
+            self._merge_blueprint_funcs(app, name)
+
+        for deferred in self.deferred_functions:
+            deferred(state)
+
+        cli_resolved_group = options.get("cli_group", self.cli_group)
+
+        if self.cli.commands:
+            if cli_resolved_group is None:
+                app.cli.commands.update(self.cli.commands)
+            elif cli_resolved_group is _sentinel:
+                self.cli.name = name
+                app.cli.add_command(self.cli)
+            else:
+                self.cli.name = cli_resolved_group
+                app.cli.add_command(self.cli)
+
+        for blueprint, bp_options in self._blueprints:
+            bp_options = bp_options.copy()
+            bp_url_prefix = bp_options.get("url_prefix")
+            bp_subdomain = bp_options.get("subdomain")
+
+            if bp_subdomain is None:
+                bp_subdomain = blueprint.subdomain
+
+            if state.subdomain is not None and bp_subdomain is not None:
+                bp_options["subdomain"] = bp_subdomain + "." + state.subdomain
+            elif bp_subdomain is not None:
+                bp_options["subdomain"] = bp_subdomain
+            elif state.subdomain is not None:
+                bp_options["subdomain"] = state.subdomain
+
+            if bp_url_prefix is None:
+                bp_url_prefix = blueprint.url_prefix
+
+            if state.url_prefix is not None and bp_url_prefix is not None:
+                bp_options["url_prefix"] = (
+                    state.url_prefix.rstrip("/") + "/" + bp_url_prefix.lstrip("/")
+                )
+            elif bp_url_prefix is not None:
+                bp_options["url_prefix"] = bp_url_prefix
+            elif state.url_prefix is not None:
+                bp_options["url_prefix"] = state.url_prefix
+
+            bp_options["name_prefix"] = name
+            blueprint.register(app, bp_options)
+
+    def _merge_blueprint_funcs(self, app: App, name: str) -> None:
+        def extend(
+            bp_dict: dict[ft.AppOrBlueprintKey, list[t.Any]],
+            parent_dict: dict[ft.AppOrBlueprintKey, list[t.Any]],
+        ) -> None:
+            for key, values in bp_dict.items():
+                key = name if key is None else f"{name}.{key}"
+                parent_dict[key].extend(values)
+
+        for key, value in self.error_handler_spec.items():
+            key = name if key is None else f"{name}.{key}"
+            value = defaultdict(
+                dict,
+                {
+                    code: {exc_class: func for exc_class, func in code_values.items()}
+                    for code, code_values in value.items()
+                },
+            )
+            app.error_handler_spec[key] = value
+
+        for endpoint, func in self.view_functions.items():
+            app.view_functions[endpoint] = func
+
+        extend(self.before_request_funcs, app.before_request_funcs)
+        extend(self.after_request_funcs, app.after_request_funcs)
+        extend(
+            self.teardown_request_funcs,
+            app.teardown_request_funcs,
+        )
+        extend(self.url_default_functions, app.url_default_functions)
+        extend(self.url_value_preprocessors, app.url_value_preprocessors)
+        extend(self.template_context_processors, app.template_context_processors)
+
+    @setupmethod
+    def add_url_rule(
+        self,
+        rule: str,
+        endpoint: str | None = None,
+        view_func: ft.RouteCallable | None = None,
+        provide_automatic_options: bool | None = None,
+        **options: t.Any,
+    ) -> None:
+        """Register a URL rule with the blueprint. See :meth:`.Flask.add_url_rule` for
+        full documentation.
+
+        The URL rule is prefixed with the blueprint's URL prefix. The endpoint name,
+        used with :func:`url_for`, is prefixed with the blueprint's name.
+        """
+        if endpoint and "." in endpoint:
+            raise ValueError("'endpoint' may not contain a dot '.' character.")
+
+        if view_func and hasattr(view_func, "__name__") and "." in view_func.__name__:
+            raise ValueError("'view_func' name may not contain a dot '.' character.")
+
+        self.record(
+            lambda s: s.add_url_rule(
+                rule,
+                endpoint,
+                view_func,
+                provide_automatic_options=provide_automatic_options,
+                **options,
+            )
+        )
+
+    @setupmethod
+    def app_template_filter(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_filter], T_template_filter]:
+        """Register a template filter, available in any template rendered by the
+        application. Equivalent to :meth:`.Flask.template_filter`.
+
+        :param name: the optional name of the filter, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_filter) -> T_template_filter:
+            self.add_app_template_filter(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_app_template_filter(
+        self, f: ft.TemplateFilterCallable, name: str | None = None
+    ) -> None:
+        """Register a template filter, available in any template rendered by the
+        application. Works like the :meth:`app_template_filter` decorator. Equivalent to
+        :meth:`.Flask.add_template_filter`.
+
+        :param name: the optional name of the filter, otherwise the
+                     function name will be used.
+        """
+
+        def register_template(state: BlueprintSetupState) -> None:
+            state.app.jinja_env.filters[name or f.__name__] = f
+
+        self.record_once(register_template)
+
+    @setupmethod
+    def app_template_test(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_test], T_template_test]:
+        """Register a template test, available in any template rendered by the
+        application. Equivalent to :meth:`.Flask.template_test`.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the test, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_test) -> T_template_test:
+            self.add_app_template_test(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_app_template_test(
+        self, f: ft.TemplateTestCallable, name: str | None = None
+    ) -> None:
+        """Register a template test, available in any template rendered by the
+        application. Works like the :meth:`app_template_test` decorator. Equivalent to
+        :meth:`.Flask.add_template_test`.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the test, otherwise the
+                     function name will be used.
+        """
+
+        def register_template(state: BlueprintSetupState) -> None:
+            state.app.jinja_env.tests[name or f.__name__] = f
+
+        self.record_once(register_template)
+
+    @setupmethod
+    def app_template_global(
+        self, name: str | None = None
+    ) -> t.Callable[[T_template_global], T_template_global]:
+        """Register a template global, available in any template rendered by the
+        application. Equivalent to :meth:`.Flask.template_global`.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the global, otherwise the
+                     function name will be used.
+        """
+
+        def decorator(f: T_template_global) -> T_template_global:
+            self.add_app_template_global(f, name=name)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_app_template_global(
+        self, f: ft.TemplateGlobalCallable, name: str | None = None
+    ) -> None:
+        """Register a template global, available in any template rendered by the
+        application. Works like the :meth:`app_template_global` decorator. Equivalent to
+        :meth:`.Flask.add_template_global`.
+
+        .. versionadded:: 0.10
+
+        :param name: the optional name of the global, otherwise the
+                     function name will be used.
+        """
+
+        def register_template(state: BlueprintSetupState) -> None:
+            state.app.jinja_env.globals[name or f.__name__] = f
+
+        self.record_once(register_template)
+
+    @setupmethod
+    def before_app_request(self, f: T_before_request) -> T_before_request:
+        """Like :meth:`before_request`, but before every request, not only those handled
+        by the blueprint. Equivalent to :meth:`.Flask.before_request`.
+        """
+        self.record_once(
+            lambda s: s.app.before_request_funcs.setdefault(None, []).append(f)
+        )
+        return f
+
+    @setupmethod
+    def after_app_request(self, f: T_after_request) -> T_after_request:
+        """Like :meth:`after_request`, but after every request, not only those handled
+        by the blueprint. Equivalent to :meth:`.Flask.after_request`.
+        """
+        self.record_once(
+            lambda s: s.app.after_request_funcs.setdefault(None, []).append(f)
+        )
+        return f
+
+    @setupmethod
+    def teardown_app_request(self, f: T_teardown) -> T_teardown:
+        """Like :meth:`teardown_request`, but after every request, not only those
+        handled by the blueprint. Equivalent to :meth:`.Flask.teardown_request`.
+        """
+        self.record_once(
+            lambda s: s.app.teardown_request_funcs.setdefault(None, []).append(f)
+        )
+        return f
+
+    @setupmethod
+    def app_context_processor(
+        self, f: T_template_context_processor
+    ) -> T_template_context_processor:
+        """Like :meth:`context_processor`, but for templates rendered by every view, not
+        only by the blueprint. Equivalent to :meth:`.Flask.context_processor`.
+        """
+        self.record_once(
+            lambda s: s.app.template_context_processors.setdefault(None, []).append(f)
+        )
+        return f
+
+    @setupmethod
+    def app_errorhandler(
+        self, code: type[Exception] | int
+    ) -> t.Callable[[T_error_handler], T_error_handler]:
+        """Like :meth:`errorhandler`, but for every request, not only those handled by
+        the blueprint. Equivalent to :meth:`.Flask.errorhandler`.
+        """
+
+        def decorator(f: T_error_handler) -> T_error_handler:
+            def from_blueprint(state: BlueprintSetupState) -> None:
+                state.app.errorhandler(code)(f)
+
+            self.record_once(from_blueprint)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def app_url_value_preprocessor(
+        self, f: T_url_value_preprocessor
+    ) -> T_url_value_preprocessor:
+        """Like :meth:`url_value_preprocessor`, but for every request, not only those
+        handled by the blueprint. Equivalent to :meth:`.Flask.url_value_preprocessor`.
+        """
+        self.record_once(
+            lambda s: s.app.url_value_preprocessors.setdefault(None, []).append(f)
+        )
+        return f
+
+    @setupmethod
+    def app_url_defaults(self, f: T_url_defaults) -> T_url_defaults:
+        """Like :meth:`url_defaults`, but for every request, not only those handled by
+        the blueprint. Equivalent to :meth:`.Flask.url_defaults`.
+        """
+        self.record_once(
+            lambda s: s.app.url_default_functions.setdefault(None, []).append(f)
+        )
+        return f
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sansio/scaffold.py b/backend/venv/lib/python3.12/site-packages/flask/sansio/scaffold.py
new file mode 100644
index 00000000..3a839f5a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/sansio/scaffold.py
@@ -0,0 +1,792 @@
+from __future__ import annotations
+
+import importlib.util
+import os
+import pathlib
+import sys
+import typing as t
+from collections import defaultdict
+from functools import update_wrapper
+
+from jinja2 import BaseLoader
+from jinja2 import FileSystemLoader
+from werkzeug.exceptions import default_exceptions
+from werkzeug.exceptions import HTTPException
+from werkzeug.utils import cached_property
+
+from .. import typing as ft
+from ..helpers import get_root_path
+from ..templating import _default_template_ctx_processor
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from click import Group
+
+# a singleton sentinel value for parameter defaults
+_sentinel = object()
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+T_after_request = t.TypeVar("T_after_request", bound=ft.AfterRequestCallable[t.Any])
+T_before_request = t.TypeVar("T_before_request", bound=ft.BeforeRequestCallable)
+T_error_handler = t.TypeVar("T_error_handler", bound=ft.ErrorHandlerCallable)
+T_teardown = t.TypeVar("T_teardown", bound=ft.TeardownCallable)
+T_template_context_processor = t.TypeVar(
+    "T_template_context_processor", bound=ft.TemplateContextProcessorCallable
+)
+T_url_defaults = t.TypeVar("T_url_defaults", bound=ft.URLDefaultCallable)
+T_url_value_preprocessor = t.TypeVar(
+    "T_url_value_preprocessor", bound=ft.URLValuePreprocessorCallable
+)
+T_route = t.TypeVar("T_route", bound=ft.RouteCallable)
+
+
+def setupmethod(f: F) -> F:
+    f_name = f.__name__
+
+    def wrapper_func(self: Scaffold, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        self._check_setup_finished(f_name)
+        return f(self, *args, **kwargs)
+
+    return t.cast(F, update_wrapper(wrapper_func, f))
+
+
+class Scaffold:
+    """Common behavior shared between :class:`~flask.Flask` and
+    :class:`~flask.blueprints.Blueprint`.
+
+    :param import_name: The import name of the module where this object
+        is defined. Usually :attr:`__name__` should be used.
+    :param static_folder: Path to a folder of static files to serve.
+        If this is set, a static route will be added.
+    :param static_url_path: URL prefix for the static route.
+    :param template_folder: Path to a folder containing template files.
+        for rendering. If this is set, a Jinja loader will be added.
+    :param root_path: The path that static, template, and resource files
+        are relative to. Typically not set, it is discovered based on
+        the ``import_name``.
+
+    .. versionadded:: 2.0
+    """
+
+    cli: Group
+    name: str
+    _static_folder: str | None = None
+    _static_url_path: str | None = None
+
+    def __init__(
+        self,
+        import_name: str,
+        static_folder: str | os.PathLike[str] | None = None,
+        static_url_path: str | None = None,
+        template_folder: str | os.PathLike[str] | None = None,
+        root_path: str | None = None,
+    ):
+        #: The name of the package or module that this object belongs
+        #: to. Do not change this once it is set by the constructor.
+        self.import_name = import_name
+
+        self.static_folder = static_folder  # type: ignore
+        self.static_url_path = static_url_path
+
+        #: The path to the templates folder, relative to
+        #: :attr:`root_path`, to add to the template loader. ``None`` if
+        #: templates should not be added.
+        self.template_folder = template_folder
+
+        if root_path is None:
+            root_path = get_root_path(self.import_name)
+
+        #: Absolute path to the package on the filesystem. Used to look
+        #: up resources contained in the package.
+        self.root_path = root_path
+
+        #: A dictionary mapping endpoint names to view functions.
+        #:
+        #: To register a view function, use the :meth:`route` decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.view_functions: dict[str, ft.RouteCallable] = {}
+
+        #: A data structure of registered error handlers, in the format
+        #: ``{scope: {code: {class: handler}}}``. The ``scope`` key is
+        #: the name of a blueprint the handlers are active for, or
+        #: ``None`` for all requests. The ``code`` key is the HTTP
+        #: status code for ``HTTPException``, or ``None`` for
+        #: other exceptions. The innermost dictionary maps exception
+        #: classes to handler functions.
+        #:
+        #: To register an error handler, use the :meth:`errorhandler`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.error_handler_spec: dict[
+            ft.AppOrBlueprintKey,
+            dict[int | None, dict[type[Exception], ft.ErrorHandlerCallable]],
+        ] = defaultdict(lambda: defaultdict(dict))
+
+        #: A data structure of functions to call at the beginning of
+        #: each request, in the format ``{scope: [functions]}``. The
+        #: ``scope`` key is the name of a blueprint the functions are
+        #: active for, or ``None`` for all requests.
+        #:
+        #: To register a function, use the :meth:`before_request`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.before_request_funcs: dict[
+            ft.AppOrBlueprintKey, list[ft.BeforeRequestCallable]
+        ] = defaultdict(list)
+
+        #: A data structure of functions to call at the end of each
+        #: request, in the format ``{scope: [functions]}``. The
+        #: ``scope`` key is the name of a blueprint the functions are
+        #: active for, or ``None`` for all requests.
+        #:
+        #: To register a function, use the :meth:`after_request`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.after_request_funcs: dict[
+            ft.AppOrBlueprintKey, list[ft.AfterRequestCallable[t.Any]]
+        ] = defaultdict(list)
+
+        #: A data structure of functions to call at the end of each
+        #: request even if an exception is raised, in the format
+        #: ``{scope: [functions]}``. The ``scope`` key is the name of a
+        #: blueprint the functions are active for, or ``None`` for all
+        #: requests.
+        #:
+        #: To register a function, use the :meth:`teardown_request`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.teardown_request_funcs: dict[
+            ft.AppOrBlueprintKey, list[ft.TeardownCallable]
+        ] = defaultdict(list)
+
+        #: A data structure of functions to call to pass extra context
+        #: values when rendering templates, in the format
+        #: ``{scope: [functions]}``. The ``scope`` key is the name of a
+        #: blueprint the functions are active for, or ``None`` for all
+        #: requests.
+        #:
+        #: To register a function, use the :meth:`context_processor`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.template_context_processors: dict[
+            ft.AppOrBlueprintKey, list[ft.TemplateContextProcessorCallable]
+        ] = defaultdict(list, {None: [_default_template_ctx_processor]})
+
+        #: A data structure of functions to call to modify the keyword
+        #: arguments passed to the view function, in the format
+        #: ``{scope: [functions]}``. The ``scope`` key is the name of a
+        #: blueprint the functions are active for, or ``None`` for all
+        #: requests.
+        #:
+        #: To register a function, use the
+        #: :meth:`url_value_preprocessor` decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.url_value_preprocessors: dict[
+            ft.AppOrBlueprintKey,
+            list[ft.URLValuePreprocessorCallable],
+        ] = defaultdict(list)
+
+        #: A data structure of functions to call to modify the keyword
+        #: arguments when generating URLs, in the format
+        #: ``{scope: [functions]}``. The ``scope`` key is the name of a
+        #: blueprint the functions are active for, or ``None`` for all
+        #: requests.
+        #:
+        #: To register a function, use the :meth:`url_defaults`
+        #: decorator.
+        #:
+        #: This data structure is internal. It should not be modified
+        #: directly and its format may change at any time.
+        self.url_default_functions: dict[
+            ft.AppOrBlueprintKey, list[ft.URLDefaultCallable]
+        ] = defaultdict(list)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.name!r}>"
+
+    def _check_setup_finished(self, f_name: str) -> None:
+        raise NotImplementedError
+
+    @property
+    def static_folder(self) -> str | None:
+        """The absolute path to the configured static folder. ``None``
+        if no static folder is set.
+        """
+        if self._static_folder is not None:
+            return os.path.join(self.root_path, self._static_folder)
+        else:
+            return None
+
+    @static_folder.setter
+    def static_folder(self, value: str | os.PathLike[str] | None) -> None:
+        if value is not None:
+            value = os.fspath(value).rstrip(r"\/")
+
+        self._static_folder = value
+
+    @property
+    def has_static_folder(self) -> bool:
+        """``True`` if :attr:`static_folder` is set.
+
+        .. versionadded:: 0.5
+        """
+        return self.static_folder is not None
+
+    @property
+    def static_url_path(self) -> str | None:
+        """The URL prefix that the static route will be accessible from.
+
+        If it was not configured during init, it is derived from
+        :attr:`static_folder`.
+        """
+        if self._static_url_path is not None:
+            return self._static_url_path
+
+        if self.static_folder is not None:
+            basename = os.path.basename(self.static_folder)
+            return f"/{basename}".rstrip("/")
+
+        return None
+
+    @static_url_path.setter
+    def static_url_path(self, value: str | None) -> None:
+        if value is not None:
+            value = value.rstrip("/")
+
+        self._static_url_path = value
+
+    @cached_property
+    def jinja_loader(self) -> BaseLoader | None:
+        """The Jinja loader for this object's templates. By default this
+        is a class :class:`jinja2.loaders.FileSystemLoader` to
+        :attr:`template_folder` if it is set.
+
+        .. versionadded:: 0.5
+        """
+        if self.template_folder is not None:
+            return FileSystemLoader(os.path.join(self.root_path, self.template_folder))
+        else:
+            return None
+
+    def _method_route(
+        self,
+        method: str,
+        rule: str,
+        options: dict[str, t.Any],
+    ) -> t.Callable[[T_route], T_route]:
+        if "methods" in options:
+            raise TypeError("Use the 'route' decorator to use the 'methods' argument.")
+
+        return self.route(rule, methods=[method], **options)
+
+    @setupmethod
+    def get(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Shortcut for :meth:`route` with ``methods=["GET"]``.
+
+        .. versionadded:: 2.0
+        """
+        return self._method_route("GET", rule, options)
+
+    @setupmethod
+    def post(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Shortcut for :meth:`route` with ``methods=["POST"]``.
+
+        .. versionadded:: 2.0
+        """
+        return self._method_route("POST", rule, options)
+
+    @setupmethod
+    def put(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Shortcut for :meth:`route` with ``methods=["PUT"]``.
+
+        .. versionadded:: 2.0
+        """
+        return self._method_route("PUT", rule, options)
+
+    @setupmethod
+    def delete(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Shortcut for :meth:`route` with ``methods=["DELETE"]``.
+
+        .. versionadded:: 2.0
+        """
+        return self._method_route("DELETE", rule, options)
+
+    @setupmethod
+    def patch(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Shortcut for :meth:`route` with ``methods=["PATCH"]``.
+
+        .. versionadded:: 2.0
+        """
+        return self._method_route("PATCH", rule, options)
+
+    @setupmethod
+    def route(self, rule: str, **options: t.Any) -> t.Callable[[T_route], T_route]:
+        """Decorate a view function to register it with the given URL
+        rule and options. Calls :meth:`add_url_rule`, which has more
+        details about the implementation.
+
+        .. code-block:: python
+
+            @app.route("/")
+            def index():
+                return "Hello, World!"
+
+        See :ref:`url-route-registrations`.
+
+        The endpoint name for the route defaults to the name of the view
+        function if the ``endpoint`` parameter isn't passed.
+
+        The ``methods`` parameter defaults to ``["GET"]``. ``HEAD`` and
+        ``OPTIONS`` are added automatically.
+
+        :param rule: The URL rule string.
+        :param options: Extra options passed to the
+            :class:`~werkzeug.routing.Rule` object.
+        """
+
+        def decorator(f: T_route) -> T_route:
+            endpoint = options.pop("endpoint", None)
+            self.add_url_rule(rule, endpoint, f, **options)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def add_url_rule(
+        self,
+        rule: str,
+        endpoint: str | None = None,
+        view_func: ft.RouteCallable | None = None,
+        provide_automatic_options: bool | None = None,
+        **options: t.Any,
+    ) -> None:
+        """Register a rule for routing incoming requests and building
+        URLs. The :meth:`route` decorator is a shortcut to call this
+        with the ``view_func`` argument. These are equivalent:
+
+        .. code-block:: python
+
+            @app.route("/")
+            def index():
+                ...
+
+        .. code-block:: python
+
+            def index():
+                ...
+
+            app.add_url_rule("/", view_func=index)
+
+        See :ref:`url-route-registrations`.
+
+        The endpoint name for the route defaults to the name of the view
+        function if the ``endpoint`` parameter isn't passed. An error
+        will be raised if a function has already been registered for the
+        endpoint.
+
+        The ``methods`` parameter defaults to ``["GET"]``. ``HEAD`` is
+        always added automatically, and ``OPTIONS`` is added
+        automatically by default.
+
+        ``view_func`` does not necessarily need to be passed, but if the
+        rule should participate in routing an endpoint name must be
+        associated with a view function at some point with the
+        :meth:`endpoint` decorator.
+
+        .. code-block:: python
+
+            app.add_url_rule("/", endpoint="index")
+
+            @app.endpoint("index")
+            def index():
+                ...
+
+        If ``view_func`` has a ``required_methods`` attribute, those
+        methods are added to the passed and automatic methods. If it
+        has a ``provide_automatic_methods`` attribute, it is used as the
+        default if the parameter is not passed.
+
+        :param rule: The URL rule string.
+        :param endpoint: The endpoint name to associate with the rule
+            and view function. Used when routing and building URLs.
+            Defaults to ``view_func.__name__``.
+        :param view_func: The view function to associate with the
+            endpoint name.
+        :param provide_automatic_options: Add the ``OPTIONS`` method and
+            respond to ``OPTIONS`` requests automatically.
+        :param options: Extra options passed to the
+            :class:`~werkzeug.routing.Rule` object.
+        """
+        raise NotImplementedError
+
+    @setupmethod
+    def endpoint(self, endpoint: str) -> t.Callable[[F], F]:
+        """Decorate a view function to register it for the given
+        endpoint. Used if a rule is added without a ``view_func`` with
+        :meth:`add_url_rule`.
+
+        .. code-block:: python
+
+            app.add_url_rule("/ex", endpoint="example")
+
+            @app.endpoint("example")
+            def example():
+                ...
+
+        :param endpoint: The endpoint name to associate with the view
+            function.
+        """
+
+        def decorator(f: F) -> F:
+            self.view_functions[endpoint] = f
+            return f
+
+        return decorator
+
+    @setupmethod
+    def before_request(self, f: T_before_request) -> T_before_request:
+        """Register a function to run before each request.
+
+        For example, this can be used to open a database connection, or
+        to load the logged in user from the session.
+
+        .. code-block:: python
+
+            @app.before_request
+            def load_user():
+                if "user_id" in session:
+                    g.user = db.session.get(session["user_id"])
+
+        The function will be called without any arguments. If it returns
+        a non-``None`` value, the value is handled as if it was the
+        return value from the view, and further request handling is
+        stopped.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        executes before every request. When used on a blueprint, this executes before
+        every request that the blueprint handles. To register with a blueprint and
+        execute before every request, use :meth:`.Blueprint.before_app_request`.
+        """
+        self.before_request_funcs.setdefault(None, []).append(f)
+        return f
+
+    @setupmethod
+    def after_request(self, f: T_after_request) -> T_after_request:
+        """Register a function to run after each request to this object.
+
+        The function is called with the response object, and must return
+        a response object. This allows the functions to modify or
+        replace the response before it is sent.
+
+        If a function raises an exception, any remaining
+        ``after_request`` functions will not be called. Therefore, this
+        should not be used for actions that must execute, such as to
+        close resources. Use :meth:`teardown_request` for that.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        executes after every request. When used on a blueprint, this executes after
+        every request that the blueprint handles. To register with a blueprint and
+        execute after every request, use :meth:`.Blueprint.after_app_request`.
+        """
+        self.after_request_funcs.setdefault(None, []).append(f)
+        return f
+
+    @setupmethod
+    def teardown_request(self, f: T_teardown) -> T_teardown:
+        """Register a function to be called when the request context is
+        popped. Typically this happens at the end of each request, but
+        contexts may be pushed manually as well during testing.
+
+        .. code-block:: python
+
+            with app.test_request_context():
+                ...
+
+        When the ``with`` block exits (or ``ctx.pop()`` is called), the
+        teardown functions are called just before the request context is
+        made inactive.
+
+        When a teardown function was called because of an unhandled
+        exception it will be passed an error object. If an
+        :meth:`errorhandler` is registered, it will handle the exception
+        and the teardown will not receive it.
+
+        Teardown functions must avoid raising exceptions. If they
+        execute code that might fail they must surround that code with a
+        ``try``/``except`` block and log any errors.
+
+        The return values of teardown functions are ignored.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        executes after every request. When used on a blueprint, this executes after
+        every request that the blueprint handles. To register with a blueprint and
+        execute after every request, use :meth:`.Blueprint.teardown_app_request`.
+        """
+        self.teardown_request_funcs.setdefault(None, []).append(f)
+        return f
+
+    @setupmethod
+    def context_processor(
+        self,
+        f: T_template_context_processor,
+    ) -> T_template_context_processor:
+        """Registers a template context processor function. These functions run before
+        rendering a template. The keys of the returned dict are added as variables
+        available in the template.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        is called for every rendered template. When used on a blueprint, this is called
+        for templates rendered from the blueprint's views. To register with a blueprint
+        and affect every template, use :meth:`.Blueprint.app_context_processor`.
+        """
+        self.template_context_processors[None].append(f)
+        return f
+
+    @setupmethod
+    def url_value_preprocessor(
+        self,
+        f: T_url_value_preprocessor,
+    ) -> T_url_value_preprocessor:
+        """Register a URL value preprocessor function for all view
+        functions in the application. These functions will be called before the
+        :meth:`before_request` functions.
+
+        The function can modify the values captured from the matched url before
+        they are passed to the view. For example, this can be used to pop a
+        common language code value and place it in ``g`` rather than pass it to
+        every view.
+
+        The function is passed the endpoint name and values dict. The return
+        value is ignored.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        is called for every request. When used on a blueprint, this is called for
+        requests that the blueprint handles. To register with a blueprint and affect
+        every request, use :meth:`.Blueprint.app_url_value_preprocessor`.
+        """
+        self.url_value_preprocessors[None].append(f)
+        return f
+
+    @setupmethod
+    def url_defaults(self, f: T_url_defaults) -> T_url_defaults:
+        """Callback function for URL defaults for all view functions of the
+        application.  It's called with the endpoint and values and should
+        update the values passed in place.
+
+        This is available on both app and blueprint objects. When used on an app, this
+        is called for every request. When used on a blueprint, this is called for
+        requests that the blueprint handles. To register with a blueprint and affect
+        every request, use :meth:`.Blueprint.app_url_defaults`.
+        """
+        self.url_default_functions[None].append(f)
+        return f
+
+    @setupmethod
+    def errorhandler(
+        self, code_or_exception: type[Exception] | int
+    ) -> t.Callable[[T_error_handler], T_error_handler]:
+        """Register a function to handle errors by code or exception class.
+
+        A decorator that is used to register a function given an
+        error code.  Example::
+
+            @app.errorhandler(404)
+            def page_not_found(error):
+                return 'This page does not exist', 404
+
+        You can also register handlers for arbitrary exceptions::
+
+            @app.errorhandler(DatabaseError)
+            def special_exception_handler(error):
+                return 'Database connection failed', 500
+
+        This is available on both app and blueprint objects. When used on an app, this
+        can handle errors from every request. When used on a blueprint, this can handle
+        errors from requests that the blueprint handles. To register with a blueprint
+        and affect every request, use :meth:`.Blueprint.app_errorhandler`.
+
+        .. versionadded:: 0.7
+            Use :meth:`register_error_handler` instead of modifying
+            :attr:`error_handler_spec` directly, for application wide error
+            handlers.
+
+        .. versionadded:: 0.7
+           One can now additionally also register custom exception types
+           that do not necessarily have to be a subclass of the
+           :class:`~werkzeug.exceptions.HTTPException` class.
+
+        :param code_or_exception: the code as integer for the handler, or
+                                  an arbitrary exception
+        """
+
+        def decorator(f: T_error_handler) -> T_error_handler:
+            self.register_error_handler(code_or_exception, f)
+            return f
+
+        return decorator
+
+    @setupmethod
+    def register_error_handler(
+        self,
+        code_or_exception: type[Exception] | int,
+        f: ft.ErrorHandlerCallable,
+    ) -> None:
+        """Alternative error attach function to the :meth:`errorhandler`
+        decorator that is more straightforward to use for non decorator
+        usage.
+
+        .. versionadded:: 0.7
+        """
+        exc_class, code = self._get_exc_class_and_code(code_or_exception)
+        self.error_handler_spec[None][code][exc_class] = f
+
+    @staticmethod
+    def _get_exc_class_and_code(
+        exc_class_or_code: type[Exception] | int,
+    ) -> tuple[type[Exception], int | None]:
+        """Get the exception class being handled. For HTTP status codes
+        or ``HTTPException`` subclasses, return both the exception and
+        status code.
+
+        :param exc_class_or_code: Any exception class, or an HTTP status
+            code as an integer.
+        """
+        exc_class: type[Exception]
+
+        if isinstance(exc_class_or_code, int):
+            try:
+                exc_class = default_exceptions[exc_class_or_code]
+            except KeyError:
+                raise ValueError(
+                    f"'{exc_class_or_code}' is not a recognized HTTP"
+                    " error code. Use a subclass of HTTPException with"
+                    " that code instead."
+                ) from None
+        else:
+            exc_class = exc_class_or_code
+
+        if isinstance(exc_class, Exception):
+            raise TypeError(
+                f"{exc_class!r} is an instance, not a class. Handlers"
+                " can only be registered for Exception classes or HTTP"
+                " error codes."
+            )
+
+        if not issubclass(exc_class, Exception):
+            raise ValueError(
+                f"'{exc_class.__name__}' is not a subclass of Exception."
+                " Handlers can only be registered for Exception classes"
+                " or HTTP error codes."
+            )
+
+        if issubclass(exc_class, HTTPException):
+            return exc_class, exc_class.code
+        else:
+            return exc_class, None
+
+
+def _endpoint_from_view_func(view_func: ft.RouteCallable) -> str:
+    """Internal helper that returns the default endpoint for a given
+    function.  This always is the function name.
+    """
+    assert view_func is not None, "expected view func if endpoint is not provided."
+    return view_func.__name__
+
+
+def _find_package_path(import_name: str) -> str:
+    """Find the path that contains the package or module."""
+    root_mod_name, _, _ = import_name.partition(".")
+
+    try:
+        root_spec = importlib.util.find_spec(root_mod_name)
+
+        if root_spec is None:
+            raise ValueError("not found")
+    except (ImportError, ValueError):
+        # ImportError: the machinery told us it does not exist
+        # ValueError:
+        #    - the module name was invalid
+        #    - the module name is __main__
+        #    - we raised `ValueError` due to `root_spec` being `None`
+        return os.getcwd()
+
+    if root_spec.submodule_search_locations:
+        if root_spec.origin is None or root_spec.origin == "namespace":
+            # namespace package
+            package_spec = importlib.util.find_spec(import_name)
+
+            if package_spec is not None and package_spec.submodule_search_locations:
+                # Pick the path in the namespace that contains the submodule.
+                package_path = pathlib.Path(
+                    os.path.commonpath(package_spec.submodule_search_locations)
+                )
+                search_location = next(
+                    location
+                    for location in root_spec.submodule_search_locations
+                    if package_path.is_relative_to(location)
+                )
+            else:
+                # Pick the first path.
+                search_location = root_spec.submodule_search_locations[0]
+
+            return os.path.dirname(search_location)
+        else:
+            # package with __init__.py
+            return os.path.dirname(os.path.dirname(root_spec.origin))
+    else:
+        # module
+        return os.path.dirname(root_spec.origin)  # type: ignore[type-var, return-value]
+
+
+def find_package(import_name: str) -> tuple[str | None, str]:
+    """Find the prefix that a package is installed under, and the path
+    that it would be imported from.
+
+    The prefix is the directory containing the standard directory
+    hierarchy (lib, bin, etc.). If the package is not installed to the
+    system (:attr:`sys.prefix`) or a virtualenv (``site-packages``),
+    ``None`` is returned.
+
+    The path is the entry in :attr:`sys.path` that contains the package
+    for import. If the package is not installed, it's assumed that the
+    package was imported from the current working directory.
+    """
+    package_path = _find_package_path(import_name)
+    py_prefix = os.path.abspath(sys.prefix)
+
+    # installed to the system
+    if pathlib.PurePath(package_path).is_relative_to(py_prefix):
+        return py_prefix, package_path
+
+    site_parent, site_folder = os.path.split(package_path)
+
+    # installed to a virtualenv
+    if site_folder.lower() == "site-packages":
+        parent, folder = os.path.split(site_parent)
+
+        # Windows (prefix/lib/site-packages)
+        if folder.lower() == "lib":
+            return parent, package_path
+
+        # Unix (prefix/lib/pythonX.Y/site-packages)
+        if os.path.basename(parent).lower() == "lib":
+            return os.path.dirname(parent), package_path
+
+        # something else (prefix/site-packages)
+        return site_parent, package_path
+
+    # not installed
+    return None, package_path
diff --git a/backend/venv/lib/python3.12/site-packages/flask/sessions.py b/backend/venv/lib/python3.12/site-packages/flask/sessions.py
new file mode 100644
index 00000000..375de065
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/sessions.py
@@ -0,0 +1,398 @@
+from __future__ import annotations
+
+import collections.abc as c
+import hashlib
+import typing as t
+from collections.abc import MutableMapping
+from datetime import datetime
+from datetime import timezone
+
+from itsdangerous import BadSignature
+from itsdangerous import URLSafeTimedSerializer
+from werkzeug.datastructures import CallbackDict
+
+from .json.tag import TaggedJSONSerializer
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    import typing_extensions as te
+
+    from .app import Flask
+    from .wrappers import Request
+    from .wrappers import Response
+
+
+class SessionMixin(MutableMapping[str, t.Any]):
+    """Expands a basic dictionary with session attributes."""
+
+    @property
+    def permanent(self) -> bool:
+        """This reflects the ``'_permanent'`` key in the dict."""
+        return self.get("_permanent", False)
+
+    @permanent.setter
+    def permanent(self, value: bool) -> None:
+        self["_permanent"] = bool(value)
+
+    #: Some implementations can detect whether a session is newly
+    #: created, but that is not guaranteed. Use with caution. The mixin
+    # default is hard-coded ``False``.
+    new = False
+
+    #: Some implementations can detect changes to the session and set
+    #: this when that happens. The mixin default is hard coded to
+    #: ``True``.
+    modified = True
+
+    #: Some implementations can detect when session data is read or
+    #: written and set this when that happens. The mixin default is hard
+    #: coded to ``True``.
+    accessed = True
+
+
+class SecureCookieSession(CallbackDict[str, t.Any], SessionMixin):
+    """Base class for sessions based on signed cookies.
+
+    This session backend will set the :attr:`modified` and
+    :attr:`accessed` attributes. It cannot reliably track whether a
+    session is new (vs. empty), so :attr:`new` remains hard coded to
+    ``False``.
+    """
+
+    #: When data is changed, this is set to ``True``. Only the session
+    #: dictionary itself is tracked; if the session contains mutable
+    #: data (for example a nested dict) then this must be set to
+    #: ``True`` manually when modifying that data. The session cookie
+    #: will only be written to the response if this is ``True``.
+    modified = False
+
+    #: When data is read or written, this is set to ``True``. Used by
+    # :class:`.SecureCookieSessionInterface` to add a ``Vary: Cookie``
+    #: header, which allows caching proxies to cache different pages for
+    #: different users.
+    accessed = False
+
+    def __init__(
+        self,
+        initial: c.Mapping[str, t.Any] | c.Iterable[tuple[str, t.Any]] | None = None,
+    ) -> None:
+        def on_update(self: te.Self) -> None:
+            self.modified = True
+            self.accessed = True
+
+        super().__init__(initial, on_update)
+
+    def __getitem__(self, key: str) -> t.Any:
+        self.accessed = True
+        return super().__getitem__(key)
+
+    def get(self, key: str, default: t.Any = None) -> t.Any:
+        self.accessed = True
+        return super().get(key, default)
+
+    def setdefault(self, key: str, default: t.Any = None) -> t.Any:
+        self.accessed = True
+        return super().setdefault(key, default)
+
+
+class NullSession(SecureCookieSession):
+    """Class used to generate nicer error messages if sessions are not
+    available.  Will still allow read-only access to the empty session
+    but fail on setting.
+    """
+
+    def _fail(self, *args: t.Any, **kwargs: t.Any) -> t.NoReturn:
+        raise RuntimeError(
+            "The session is unavailable because no secret "
+            "key was set.  Set the secret_key on the "
+            "application to something unique and secret."
+        )
+
+    __setitem__ = __delitem__ = clear = pop = popitem = update = setdefault = _fail  # type: ignore # noqa: B950
+    del _fail
+
+
+class SessionInterface:
+    """The basic interface you have to implement in order to replace the
+    default session interface which uses werkzeug's securecookie
+    implementation.  The only methods you have to implement are
+    :meth:`open_session` and :meth:`save_session`, the others have
+    useful defaults which you don't need to change.
+
+    The session object returned by the :meth:`open_session` method has to
+    provide a dictionary like interface plus the properties and methods
+    from the :class:`SessionMixin`.  We recommend just subclassing a dict
+    and adding that mixin::
+
+        class Session(dict, SessionMixin):
+            pass
+
+    If :meth:`open_session` returns ``None`` Flask will call into
+    :meth:`make_null_session` to create a session that acts as replacement
+    if the session support cannot work because some requirement is not
+    fulfilled.  The default :class:`NullSession` class that is created
+    will complain that the secret key was not set.
+
+    To replace the session interface on an application all you have to do
+    is to assign :attr:`flask.Flask.session_interface`::
+
+        app = Flask(__name__)
+        app.session_interface = MySessionInterface()
+
+    Multiple requests with the same session may be sent and handled
+    concurrently. When implementing a new session interface, consider
+    whether reads or writes to the backing store must be synchronized.
+    There is no guarantee on the order in which the session for each
+    request is opened or saved, it will occur in the order that requests
+    begin and end processing.
+
+    .. versionadded:: 0.8
+    """
+
+    #: :meth:`make_null_session` will look here for the class that should
+    #: be created when a null session is requested.  Likewise the
+    #: :meth:`is_null_session` method will perform a typecheck against
+    #: this type.
+    null_session_class = NullSession
+
+    #: A flag that indicates if the session interface is pickle based.
+    #: This can be used by Flask extensions to make a decision in regards
+    #: to how to deal with the session object.
+    #:
+    #: .. versionadded:: 0.10
+    pickle_based = False
+
+    def make_null_session(self, app: Flask) -> NullSession:
+        """Creates a null session which acts as a replacement object if the
+        real session support could not be loaded due to a configuration
+        error.  This mainly aids the user experience because the job of the
+        null session is to still support lookup without complaining but
+        modifications are answered with a helpful error message of what
+        failed.
+
+        This creates an instance of :attr:`null_session_class` by default.
+        """
+        return self.null_session_class()
+
+    def is_null_session(self, obj: object) -> bool:
+        """Checks if a given object is a null session.  Null sessions are
+        not asked to be saved.
+
+        This checks if the object is an instance of :attr:`null_session_class`
+        by default.
+        """
+        return isinstance(obj, self.null_session_class)
+
+    def get_cookie_name(self, app: Flask) -> str:
+        """The name of the session cookie. Uses``app.config["SESSION_COOKIE_NAME"]``."""
+        return app.config["SESSION_COOKIE_NAME"]  # type: ignore[no-any-return]
+
+    def get_cookie_domain(self, app: Flask) -> str | None:
+        """The value of the ``Domain`` parameter on the session cookie. If not set,
+        browsers will only send the cookie to the exact domain it was set from.
+        Otherwise, they will send it to any subdomain of the given value as well.
+
+        Uses the :data:`SESSION_COOKIE_DOMAIN` config.
+
+        .. versionchanged:: 2.3
+            Not set by default, does not fall back to ``SERVER_NAME``.
+        """
+        return app.config["SESSION_COOKIE_DOMAIN"]  # type: ignore[no-any-return]
+
+    def get_cookie_path(self, app: Flask) -> str:
+        """Returns the path for which the cookie should be valid.  The
+        default implementation uses the value from the ``SESSION_COOKIE_PATH``
+        config var if it's set, and falls back to ``APPLICATION_ROOT`` or
+        uses ``/`` if it's ``None``.
+        """
+        return app.config["SESSION_COOKIE_PATH"] or app.config["APPLICATION_ROOT"]  # type: ignore[no-any-return]
+
+    def get_cookie_httponly(self, app: Flask) -> bool:
+        """Returns True if the session cookie should be httponly.  This
+        currently just returns the value of the ``SESSION_COOKIE_HTTPONLY``
+        config var.
+        """
+        return app.config["SESSION_COOKIE_HTTPONLY"]  # type: ignore[no-any-return]
+
+    def get_cookie_secure(self, app: Flask) -> bool:
+        """Returns True if the cookie should be secure.  This currently
+        just returns the value of the ``SESSION_COOKIE_SECURE`` setting.
+        """
+        return app.config["SESSION_COOKIE_SECURE"]  # type: ignore[no-any-return]
+
+    def get_cookie_samesite(self, app: Flask) -> str | None:
+        """Return ``'Strict'`` or ``'Lax'`` if the cookie should use the
+        ``SameSite`` attribute. This currently just returns the value of
+        the :data:`SESSION_COOKIE_SAMESITE` setting.
+        """
+        return app.config["SESSION_COOKIE_SAMESITE"]  # type: ignore[no-any-return]
+
+    def get_cookie_partitioned(self, app: Flask) -> bool:
+        """Returns True if the cookie should be partitioned. By default, uses
+        the value of :data:`SESSION_COOKIE_PARTITIONED`.
+
+        .. versionadded:: 3.1
+        """
+        return app.config["SESSION_COOKIE_PARTITIONED"]  # type: ignore[no-any-return]
+
+    def get_expiration_time(self, app: Flask, session: SessionMixin) -> datetime | None:
+        """A helper method that returns an expiration date for the session
+        or ``None`` if the session is linked to the browser session.  The
+        default implementation returns now + the permanent session
+        lifetime configured on the application.
+        """
+        if session.permanent:
+            return datetime.now(timezone.utc) + app.permanent_session_lifetime
+        return None
+
+    def should_set_cookie(self, app: Flask, session: SessionMixin) -> bool:
+        """Used by session backends to determine if a ``Set-Cookie`` header
+        should be set for this session cookie for this response. If the session
+        has been modified, the cookie is set. If the session is permanent and
+        the ``SESSION_REFRESH_EACH_REQUEST`` config is true, the cookie is
+        always set.
+
+        This check is usually skipped if the session was deleted.
+
+        .. versionadded:: 0.11
+        """
+
+        return session.modified or (
+            session.permanent and app.config["SESSION_REFRESH_EACH_REQUEST"]
+        )
+
+    def open_session(self, app: Flask, request: Request) -> SessionMixin | None:
+        """This is called at the beginning of each request, after
+        pushing the request context, before matching the URL.
+
+        This must return an object which implements a dictionary-like
+        interface as well as the :class:`SessionMixin` interface.
+
+        This will return ``None`` to indicate that loading failed in
+        some way that is not immediately an error. The request
+        context will fall back to using :meth:`make_null_session`
+        in this case.
+        """
+        raise NotImplementedError()
+
+    def save_session(
+        self, app: Flask, session: SessionMixin, response: Response
+    ) -> None:
+        """This is called at the end of each request, after generating
+        a response, before removing the request context. It is skipped
+        if :meth:`is_null_session` returns ``True``.
+        """
+        raise NotImplementedError()
+
+
+session_json_serializer = TaggedJSONSerializer()
+
+
+def _lazy_sha1(string: bytes = b"") -> t.Any:
+    """Don't access ``hashlib.sha1`` until runtime. FIPS builds may not include
+    SHA-1, in which case the import and use as a default would fail before the
+    developer can configure something else.
+    """
+    return hashlib.sha1(string)
+
+
+class SecureCookieSessionInterface(SessionInterface):
+    """The default session interface that stores sessions in signed cookies
+    through the :mod:`itsdangerous` module.
+    """
+
+    #: the salt that should be applied on top of the secret key for the
+    #: signing of cookie based sessions.
+    salt = "cookie-session"
+    #: the hash function to use for the signature.  The default is sha1
+    digest_method = staticmethod(_lazy_sha1)
+    #: the name of the itsdangerous supported key derivation.  The default
+    #: is hmac.
+    key_derivation = "hmac"
+    #: A python serializer for the payload.  The default is a compact
+    #: JSON derived serializer with support for some extra Python types
+    #: such as datetime objects or tuples.
+    serializer = session_json_serializer
+    session_class = SecureCookieSession
+
+    def get_signing_serializer(self, app: Flask) -> URLSafeTimedSerializer | None:
+        if not app.secret_key:
+            return None
+
+        keys: list[str | bytes] = [app.secret_key]
+
+        if fallbacks := app.config["SECRET_KEY_FALLBACKS"]:
+            keys.extend(fallbacks)
+
+        return URLSafeTimedSerializer(
+            keys,  # type: ignore[arg-type]
+            salt=self.salt,
+            serializer=self.serializer,
+            signer_kwargs={
+                "key_derivation": self.key_derivation,
+                "digest_method": self.digest_method,
+            },
+        )
+
+    def open_session(self, app: Flask, request: Request) -> SecureCookieSession | None:
+        s = self.get_signing_serializer(app)
+        if s is None:
+            return None
+        val = request.cookies.get(self.get_cookie_name(app))
+        if not val:
+            return self.session_class()
+        max_age = int(app.permanent_session_lifetime.total_seconds())
+        try:
+            data = s.loads(val, max_age=max_age)
+            return self.session_class(data)
+        except BadSignature:
+            return self.session_class()
+
+    def save_session(
+        self, app: Flask, session: SessionMixin, response: Response
+    ) -> None:
+        name = self.get_cookie_name(app)
+        domain = self.get_cookie_domain(app)
+        path = self.get_cookie_path(app)
+        secure = self.get_cookie_secure(app)
+        partitioned = self.get_cookie_partitioned(app)
+        samesite = self.get_cookie_samesite(app)
+        httponly = self.get_cookie_httponly(app)
+
+        # Add a "Vary: Cookie" header if the session was accessed at all.
+        if session.accessed:
+            response.vary.add("Cookie")
+
+        # If the session is modified to be empty, remove the cookie.
+        # If the session is empty, return without setting the cookie.
+        if not session:
+            if session.modified:
+                response.delete_cookie(
+                    name,
+                    domain=domain,
+                    path=path,
+                    secure=secure,
+                    partitioned=partitioned,
+                    samesite=samesite,
+                    httponly=httponly,
+                )
+                response.vary.add("Cookie")
+
+            return
+
+        if not self.should_set_cookie(app, session):
+            return
+
+        expires = self.get_expiration_time(app, session)
+        val = self.get_signing_serializer(app).dumps(dict(session))  # type: ignore[union-attr]
+        response.set_cookie(
+            name,
+            val,
+            expires=expires,
+            httponly=httponly,
+            domain=domain,
+            path=path,
+            secure=secure,
+            partitioned=partitioned,
+            samesite=samesite,
+        )
+        response.vary.add("Cookie")
diff --git a/backend/venv/lib/python3.12/site-packages/flask/signals.py b/backend/venv/lib/python3.12/site-packages/flask/signals.py
new file mode 100644
index 00000000..444fda99
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/signals.py
@@ -0,0 +1,17 @@
+from __future__ import annotations
+
+from blinker import Namespace
+
+# This namespace is only for signals provided by Flask itself.
+_signals = Namespace()
+
+template_rendered = _signals.signal("template-rendered")
+before_render_template = _signals.signal("before-render-template")
+request_started = _signals.signal("request-started")
+request_finished = _signals.signal("request-finished")
+request_tearing_down = _signals.signal("request-tearing-down")
+got_request_exception = _signals.signal("got-request-exception")
+appcontext_tearing_down = _signals.signal("appcontext-tearing-down")
+appcontext_pushed = _signals.signal("appcontext-pushed")
+appcontext_popped = _signals.signal("appcontext-popped")
+message_flashed = _signals.signal("message-flashed")
diff --git a/backend/venv/lib/python3.12/site-packages/flask/templating.py b/backend/venv/lib/python3.12/site-packages/flask/templating.py
new file mode 100644
index 00000000..618a3b35
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/templating.py
@@ -0,0 +1,219 @@
+from __future__ import annotations
+
+import typing as t
+
+from jinja2 import BaseLoader
+from jinja2 import Environment as BaseEnvironment
+from jinja2 import Template
+from jinja2 import TemplateNotFound
+
+from .globals import _cv_app
+from .globals import _cv_request
+from .globals import current_app
+from .globals import request
+from .helpers import stream_with_context
+from .signals import before_render_template
+from .signals import template_rendered
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from .app import Flask
+    from .sansio.app import App
+    from .sansio.scaffold import Scaffold
+
+
+def _default_template_ctx_processor() -> dict[str, t.Any]:
+    """Default template context processor.  Injects `request`,
+    `session` and `g`.
+    """
+    appctx = _cv_app.get(None)
+    reqctx = _cv_request.get(None)
+    rv: dict[str, t.Any] = {}
+    if appctx is not None:
+        rv["g"] = appctx.g
+    if reqctx is not None:
+        rv["request"] = reqctx.request
+        rv["session"] = reqctx.session
+    return rv
+
+
+class Environment(BaseEnvironment):
+    """Works like a regular Jinja2 environment but has some additional
+    knowledge of how Flask's blueprint works so that it can prepend the
+    name of the blueprint to referenced templates if necessary.
+    """
+
+    def __init__(self, app: App, **options: t.Any) -> None:
+        if "loader" not in options:
+            options["loader"] = app.create_global_jinja_loader()
+        BaseEnvironment.__init__(self, **options)
+        self.app = app
+
+
+class DispatchingJinjaLoader(BaseLoader):
+    """A loader that looks for templates in the application and all
+    the blueprint folders.
+    """
+
+    def __init__(self, app: App) -> None:
+        self.app = app
+
+    def get_source(
+        self, environment: BaseEnvironment, template: str
+    ) -> tuple[str, str | None, t.Callable[[], bool] | None]:
+        if self.app.config["EXPLAIN_TEMPLATE_LOADING"]:
+            return self._get_source_explained(environment, template)
+        return self._get_source_fast(environment, template)
+
+    def _get_source_explained(
+        self, environment: BaseEnvironment, template: str
+    ) -> tuple[str, str | None, t.Callable[[], bool] | None]:
+        attempts = []
+        rv: tuple[str, str | None, t.Callable[[], bool] | None] | None
+        trv: None | (tuple[str, str | None, t.Callable[[], bool] | None]) = None
+
+        for srcobj, loader in self._iter_loaders(template):
+            try:
+                rv = loader.get_source(environment, template)
+                if trv is None:
+                    trv = rv
+            except TemplateNotFound:
+                rv = None
+            attempts.append((loader, srcobj, rv))
+
+        from .debughelpers import explain_template_loading_attempts
+
+        explain_template_loading_attempts(self.app, template, attempts)
+
+        if trv is not None:
+            return trv
+        raise TemplateNotFound(template)
+
+    def _get_source_fast(
+        self, environment: BaseEnvironment, template: str
+    ) -> tuple[str, str | None, t.Callable[[], bool] | None]:
+        for _srcobj, loader in self._iter_loaders(template):
+            try:
+                return loader.get_source(environment, template)
+            except TemplateNotFound:
+                continue
+        raise TemplateNotFound(template)
+
+    def _iter_loaders(self, template: str) -> t.Iterator[tuple[Scaffold, BaseLoader]]:
+        loader = self.app.jinja_loader
+        if loader is not None:
+            yield self.app, loader
+
+        for blueprint in self.app.iter_blueprints():
+            loader = blueprint.jinja_loader
+            if loader is not None:
+                yield blueprint, loader
+
+    def list_templates(self) -> list[str]:
+        result = set()
+        loader = self.app.jinja_loader
+        if loader is not None:
+            result.update(loader.list_templates())
+
+        for blueprint in self.app.iter_blueprints():
+            loader = blueprint.jinja_loader
+            if loader is not None:
+                for template in loader.list_templates():
+                    result.add(template)
+
+        return list(result)
+
+
+def _render(app: Flask, template: Template, context: dict[str, t.Any]) -> str:
+    app.update_template_context(context)
+    before_render_template.send(
+        app, _async_wrapper=app.ensure_sync, template=template, context=context
+    )
+    rv = template.render(context)
+    template_rendered.send(
+        app, _async_wrapper=app.ensure_sync, template=template, context=context
+    )
+    return rv
+
+
+def render_template(
+    template_name_or_list: str | Template | list[str | Template],
+    **context: t.Any,
+) -> str:
+    """Render a template by name with the given context.
+
+    :param template_name_or_list: The name of the template to render. If
+        a list is given, the first name to exist will be rendered.
+    :param context: The variables to make available in the template.
+    """
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+    template = app.jinja_env.get_or_select_template(template_name_or_list)
+    return _render(app, template, context)
+
+
+def render_template_string(source: str, **context: t.Any) -> str:
+    """Render a template from the given source string with the given
+    context.
+
+    :param source: The source code of the template to render.
+    :param context: The variables to make available in the template.
+    """
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+    template = app.jinja_env.from_string(source)
+    return _render(app, template, context)
+
+
+def _stream(
+    app: Flask, template: Template, context: dict[str, t.Any]
+) -> t.Iterator[str]:
+    app.update_template_context(context)
+    before_render_template.send(
+        app, _async_wrapper=app.ensure_sync, template=template, context=context
+    )
+
+    def generate() -> t.Iterator[str]:
+        yield from template.generate(context)
+        template_rendered.send(
+            app, _async_wrapper=app.ensure_sync, template=template, context=context
+        )
+
+    rv = generate()
+
+    # If a request context is active, keep it while generating.
+    if request:
+        rv = stream_with_context(rv)
+
+    return rv
+
+
+def stream_template(
+    template_name_or_list: str | Template | list[str | Template],
+    **context: t.Any,
+) -> t.Iterator[str]:
+    """Render a template by name with the given context as a stream.
+    This returns an iterator of strings, which can be used as a
+    streaming response from a view.
+
+    :param template_name_or_list: The name of the template to render. If
+        a list is given, the first name to exist will be rendered.
+    :param context: The variables to make available in the template.
+
+    .. versionadded:: 2.2
+    """
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+    template = app.jinja_env.get_or_select_template(template_name_or_list)
+    return _stream(app, template, context)
+
+
+def stream_template_string(source: str, **context: t.Any) -> t.Iterator[str]:
+    """Render a template from the given source string with the given
+    context as a stream. This returns an iterator of strings, which can
+    be used as a streaming response from a view.
+
+    :param source: The source code of the template to render.
+    :param context: The variables to make available in the template.
+
+    .. versionadded:: 2.2
+    """
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+    template = app.jinja_env.from_string(source)
+    return _stream(app, template, context)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/testing.py b/backend/venv/lib/python3.12/site-packages/flask/testing.py
new file mode 100644
index 00000000..602b773b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/testing.py
@@ -0,0 +1,297 @@
+from __future__ import annotations
+
+import importlib.metadata
+import typing as t
+from contextlib import contextmanager
+from contextlib import ExitStack
+from copy import copy
+from types import TracebackType
+from urllib.parse import urlsplit
+
+import werkzeug.test
+from click.testing import CliRunner
+from werkzeug.test import Client
+from werkzeug.wrappers import Request as BaseRequest
+
+from .cli import ScriptInfo
+from .sessions import SessionMixin
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from _typeshed.wsgi import WSGIEnvironment
+    from werkzeug.test import TestResponse
+
+    from .app import Flask
+
+
+class EnvironBuilder(werkzeug.test.EnvironBuilder):
+    """An :class:`~werkzeug.test.EnvironBuilder`, that takes defaults from the
+    application.
+
+    :param app: The Flask application to configure the environment from.
+    :param path: URL path being requested.
+    :param base_url: Base URL where the app is being served, which
+        ``path`` is relative to. If not given, built from
+        :data:`PREFERRED_URL_SCHEME`, ``subdomain``,
+        :data:`SERVER_NAME`, and :data:`APPLICATION_ROOT`.
+    :param subdomain: Subdomain name to append to :data:`SERVER_NAME`.
+    :param url_scheme: Scheme to use instead of
+        :data:`PREFERRED_URL_SCHEME`.
+    :param json: If given, this is serialized as JSON and passed as
+        ``data``. Also defaults ``content_type`` to
+        ``application/json``.
+    :param args: other positional arguments passed to
+        :class:`~werkzeug.test.EnvironBuilder`.
+    :param kwargs: other keyword arguments passed to
+        :class:`~werkzeug.test.EnvironBuilder`.
+    """
+
+    def __init__(
+        self,
+        app: Flask,
+        path: str = "/",
+        base_url: str | None = None,
+        subdomain: str | None = None,
+        url_scheme: str | None = None,
+        *args: t.Any,
+        **kwargs: t.Any,
+    ) -> None:
+        assert not (base_url or subdomain or url_scheme) or (
+            base_url is not None
+        ) != bool(
+            subdomain or url_scheme
+        ), 'Cannot pass "subdomain" or "url_scheme" with "base_url".'
+
+        if base_url is None:
+            http_host = app.config.get("SERVER_NAME") or "localhost"
+            app_root = app.config["APPLICATION_ROOT"]
+
+            if subdomain:
+                http_host = f"{subdomain}.{http_host}"
+
+            if url_scheme is None:
+                url_scheme = app.config["PREFERRED_URL_SCHEME"]
+
+            url = urlsplit(path)
+            base_url = (
+                f"{url.scheme or url_scheme}://{url.netloc or http_host}"
+                f"/{app_root.lstrip('/')}"
+            )
+            path = url.path
+
+            if url.query:
+                path = f"{path}?{url.query}"
+
+        self.app = app
+        super().__init__(path, base_url, *args, **kwargs)
+
+    def json_dumps(self, obj: t.Any, **kwargs: t.Any) -> str:  # type: ignore
+        """Serialize ``obj`` to a JSON-formatted string.
+
+        The serialization will be configured according to the config associated
+        with this EnvironBuilder's ``app``.
+        """
+        return self.app.json.dumps(obj, **kwargs)
+
+
+_werkzeug_version = ""
+
+
+def _get_werkzeug_version() -> str:
+    global _werkzeug_version
+
+    if not _werkzeug_version:
+        _werkzeug_version = importlib.metadata.version("werkzeug")
+
+    return _werkzeug_version
+
+
+class FlaskClient(Client):
+    """Works like a regular Werkzeug test client but has knowledge about
+    Flask's contexts to defer the cleanup of the request context until
+    the end of a ``with`` block. For general information about how to
+    use this class refer to :class:`werkzeug.test.Client`.
+
+    .. versionchanged:: 0.12
+       `app.test_client()` includes preset default environment, which can be
+       set after instantiation of the `app.test_client()` object in
+       `client.environ_base`.
+
+    Basic usage is outlined in the :doc:`/testing` chapter.
+    """
+
+    application: Flask
+
+    def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:
+        super().__init__(*args, **kwargs)
+        self.preserve_context = False
+        self._new_contexts: list[t.ContextManager[t.Any]] = []
+        self._context_stack = ExitStack()
+        self.environ_base = {
+            "REMOTE_ADDR": "127.0.0.1",
+            "HTTP_USER_AGENT": f"Werkzeug/{_get_werkzeug_version()}",
+        }
+
+    @contextmanager
+    def session_transaction(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.Iterator[SessionMixin]:
+        """When used in combination with a ``with`` statement this opens a
+        session transaction.  This can be used to modify the session that
+        the test client uses.  Once the ``with`` block is left the session is
+        stored back.
+
+        ::
+
+            with client.session_transaction() as session:
+                session['value'] = 42
+
+        Internally this is implemented by going through a temporary test
+        request context and since session handling could depend on
+        request variables this function accepts the same arguments as
+        :meth:`~flask.Flask.test_request_context` which are directly
+        passed through.
+        """
+        if self._cookies is None:
+            raise TypeError(
+                "Cookies are disabled. Create a client with 'use_cookies=True'."
+            )
+
+        app = self.application
+        ctx = app.test_request_context(*args, **kwargs)
+        self._add_cookies_to_wsgi(ctx.request.environ)
+
+        with ctx:
+            sess = app.session_interface.open_session(app, ctx.request)
+
+        if sess is None:
+            raise RuntimeError("Session backend did not open a session.")
+
+        yield sess
+        resp = app.response_class()
+
+        if app.session_interface.is_null_session(sess):
+            return
+
+        with ctx:
+            app.session_interface.save_session(app, sess, resp)
+
+        self._update_cookies_from_response(
+            ctx.request.host.partition(":")[0],
+            ctx.request.path,
+            resp.headers.getlist("Set-Cookie"),
+        )
+
+    def _copy_environ(self, other: WSGIEnvironment) -> WSGIEnvironment:
+        out = {**self.environ_base, **other}
+
+        if self.preserve_context:
+            out["werkzeug.debug.preserve_context"] = self._new_contexts.append
+
+        return out
+
+    def _request_from_builder_args(
+        self, args: tuple[t.Any, ...], kwargs: dict[str, t.Any]
+    ) -> BaseRequest:
+        kwargs["environ_base"] = self._copy_environ(kwargs.get("environ_base", {}))
+        builder = EnvironBuilder(self.application, *args, **kwargs)
+
+        try:
+            return builder.get_request()
+        finally:
+            builder.close()
+
+    def open(
+        self,
+        *args: t.Any,
+        buffered: bool = False,
+        follow_redirects: bool = False,
+        **kwargs: t.Any,
+    ) -> TestResponse:
+        if args and isinstance(
+            args[0], (werkzeug.test.EnvironBuilder, dict, BaseRequest)
+        ):
+            if isinstance(args[0], werkzeug.test.EnvironBuilder):
+                builder = copy(args[0])
+                builder.environ_base = self._copy_environ(builder.environ_base or {})  # type: ignore[arg-type]
+                request = builder.get_request()
+            elif isinstance(args[0], dict):
+                request = EnvironBuilder.from_environ(
+                    args[0], app=self.application, environ_base=self._copy_environ({})
+                ).get_request()
+            else:
+                # isinstance(args[0], BaseRequest)
+                request = copy(args[0])
+                request.environ = self._copy_environ(request.environ)
+        else:
+            # request is None
+            request = self._request_from_builder_args(args, kwargs)
+
+        # Pop any previously preserved contexts. This prevents contexts
+        # from being preserved across redirects or multiple requests
+        # within a single block.
+        self._context_stack.close()
+
+        response = super().open(
+            request,
+            buffered=buffered,
+            follow_redirects=follow_redirects,
+        )
+        response.json_module = self.application.json  # type: ignore[assignment]
+
+        # Re-push contexts that were preserved during the request.
+        while self._new_contexts:
+            cm = self._new_contexts.pop()
+            self._context_stack.enter_context(cm)
+
+        return response
+
+    def __enter__(self) -> FlaskClient:
+        if self.preserve_context:
+            raise RuntimeError("Cannot nest client invocations")
+        self.preserve_context = True
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type | None,
+        exc_value: BaseException | None,
+        tb: TracebackType | None,
+    ) -> None:
+        self.preserve_context = False
+        self._context_stack.close()
+
+
+class FlaskCliRunner(CliRunner):
+    """A :class:`~click.testing.CliRunner` for testing a Flask app's
+    CLI commands. Typically created using
+    :meth:`~flask.Flask.test_cli_runner`. See :ref:`testing-cli`.
+    """
+
+    def __init__(self, app: Flask, **kwargs: t.Any) -> None:
+        self.app = app
+        super().__init__(**kwargs)
+
+    def invoke(  # type: ignore
+        self, cli: t.Any = None, args: t.Any = None, **kwargs: t.Any
+    ) -> t.Any:
+        """Invokes a CLI command in an isolated environment. See
+        :meth:`CliRunner.invoke <click.testing.CliRunner.invoke>` for
+        full method documentation. See :ref:`testing-cli` for examples.
+
+        If the ``obj`` argument is not given, passes an instance of
+        :class:`~flask.cli.ScriptInfo` that knows how to load the Flask
+        app being tested.
+
+        :param cli: Command object to invoke. Default is the app's
+            :attr:`~flask.app.Flask.cli` group.
+        :param args: List of strings to invoke the command with.
+
+        :return: a :class:`~click.testing.Result` object.
+        """
+        if cli is None:
+            cli = self.app.cli
+
+        if "obj" not in kwargs:
+            kwargs["obj"] = ScriptInfo(create_app=lambda: self.app)
+
+        return super().invoke(cli, args, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/flask/typing.py b/backend/venv/lib/python3.12/site-packages/flask/typing.py
new file mode 100644
index 00000000..e7234e96
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/typing.py
@@ -0,0 +1,90 @@
+from __future__ import annotations
+
+import typing as t
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from _typeshed.wsgi import WSGIApplication  # noqa: F401
+    from werkzeug.datastructures import Headers  # noqa: F401
+    from werkzeug.sansio.response import Response  # noqa: F401
+
+# The possible types that are directly convertible or are a Response object.
+ResponseValue = t.Union[
+    "Response",
+    str,
+    bytes,
+    list[t.Any],
+    # Only dict is actually accepted, but Mapping allows for TypedDict.
+    t.Mapping[str, t.Any],
+    t.Iterator[str],
+    t.Iterator[bytes],
+]
+
+# the possible types for an individual HTTP header
+# This should be a Union, but mypy doesn't pass unless it's a TypeVar.
+HeaderValue = t.Union[str, list[str], tuple[str, ...]]
+
+# the possible types for HTTP headers
+HeadersValue = t.Union[
+    "Headers",
+    t.Mapping[str, HeaderValue],
+    t.Sequence[tuple[str, HeaderValue]],
+]
+
+# The possible types returned by a route function.
+ResponseReturnValue = t.Union[
+    ResponseValue,
+    tuple[ResponseValue, HeadersValue],
+    tuple[ResponseValue, int],
+    tuple[ResponseValue, int, HeadersValue],
+    "WSGIApplication",
+]
+
+# Allow any subclass of werkzeug.Response, such as the one from Flask,
+# as a callback argument. Using werkzeug.Response directly makes a
+# callback annotated with flask.Response fail type checking.
+ResponseClass = t.TypeVar("ResponseClass", bound="Response")
+
+AppOrBlueprintKey = t.Optional[str]  # The App key is None, whereas blueprints are named
+AfterRequestCallable = t.Union[
+    t.Callable[[ResponseClass], ResponseClass],
+    t.Callable[[ResponseClass], t.Awaitable[ResponseClass]],
+]
+BeforeFirstRequestCallable = t.Union[
+    t.Callable[[], None], t.Callable[[], t.Awaitable[None]]
+]
+BeforeRequestCallable = t.Union[
+    t.Callable[[], t.Optional[ResponseReturnValue]],
+    t.Callable[[], t.Awaitable[t.Optional[ResponseReturnValue]]],
+]
+ShellContextProcessorCallable = t.Callable[[], dict[str, t.Any]]
+TeardownCallable = t.Union[
+    t.Callable[[t.Optional[BaseException]], None],
+    t.Callable[[t.Optional[BaseException]], t.Awaitable[None]],
+]
+TemplateContextProcessorCallable = t.Union[
+    t.Callable[[], dict[str, t.Any]],
+    t.Callable[[], t.Awaitable[dict[str, t.Any]]],
+]
+TemplateFilterCallable = t.Callable[..., t.Any]
+TemplateGlobalCallable = t.Callable[..., t.Any]
+TemplateTestCallable = t.Callable[..., bool]
+URLDefaultCallable = t.Callable[[str, dict[str, t.Any]], None]
+URLValuePreprocessorCallable = t.Callable[
+    [t.Optional[str], t.Optional[dict[str, t.Any]]], None
+]
+
+# This should take Exception, but that either breaks typing the argument
+# with a specific exception, or decorating multiple times with different
+# exceptions (and using a union type on the argument).
+# https://github.com/pallets/flask/issues/4095
+# https://github.com/pallets/flask/issues/4295
+# https://github.com/pallets/flask/issues/4297
+ErrorHandlerCallable = t.Union[
+    t.Callable[[t.Any], ResponseReturnValue],
+    t.Callable[[t.Any], t.Awaitable[ResponseReturnValue]],
+]
+
+RouteCallable = t.Union[
+    t.Callable[..., ResponseReturnValue],
+    t.Callable[..., t.Awaitable[ResponseReturnValue]],
+]
diff --git a/backend/venv/lib/python3.12/site-packages/flask/views.py b/backend/venv/lib/python3.12/site-packages/flask/views.py
new file mode 100644
index 00000000..53fe976d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/views.py
@@ -0,0 +1,191 @@
+from __future__ import annotations
+
+import typing as t
+
+from . import typing as ft
+from .globals import current_app
+from .globals import request
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+http_method_funcs = frozenset(
+    ["get", "post", "head", "options", "delete", "put", "trace", "patch"]
+)
+
+
+class View:
+    """Subclass this class and override :meth:`dispatch_request` to
+    create a generic class-based view. Call :meth:`as_view` to create a
+    view function that creates an instance of the class with the given
+    arguments and calls its ``dispatch_request`` method with any URL
+    variables.
+
+    See :doc:`views` for a detailed guide.
+
+    .. code-block:: python
+
+        class Hello(View):
+            init_every_request = False
+
+            def dispatch_request(self, name):
+                return f"Hello, {name}!"
+
+        app.add_url_rule(
+            "/hello/<name>", view_func=Hello.as_view("hello")
+        )
+
+    Set :attr:`methods` on the class to change what methods the view
+    accepts.
+
+    Set :attr:`decorators` on the class to apply a list of decorators to
+    the generated view function. Decorators applied to the class itself
+    will not be applied to the generated view function!
+
+    Set :attr:`init_every_request` to ``False`` for efficiency, unless
+    you need to store request-global data on ``self``.
+    """
+
+    #: The methods this view is registered for. Uses the same default
+    #: (``["GET", "HEAD", "OPTIONS"]``) as ``route`` and
+    #: ``add_url_rule`` by default.
+    methods: t.ClassVar[t.Collection[str] | None] = None
+
+    #: Control whether the ``OPTIONS`` method is handled automatically.
+    #: Uses the same default (``True``) as ``route`` and
+    #: ``add_url_rule`` by default.
+    provide_automatic_options: t.ClassVar[bool | None] = None
+
+    #: A list of decorators to apply, in order, to the generated view
+    #: function. Remember that ``@decorator`` syntax is applied bottom
+    #: to top, so the first decorator in the list would be the bottom
+    #: decorator.
+    #:
+    #: .. versionadded:: 0.8
+    decorators: t.ClassVar[list[t.Callable[..., t.Any]]] = []
+
+    #: Create a new instance of this view class for every request by
+    #: default. If a view subclass sets this to ``False``, the same
+    #: instance is used for every request.
+    #:
+    #: A single instance is more efficient, especially if complex setup
+    #: is done during init. However, storing data on ``self`` is no
+    #: longer safe across requests, and :data:`~flask.g` should be used
+    #: instead.
+    #:
+    #: .. versionadded:: 2.2
+    init_every_request: t.ClassVar[bool] = True
+
+    def dispatch_request(self) -> ft.ResponseReturnValue:
+        """The actual view function behavior. Subclasses must override
+        this and return a valid response. Any variables from the URL
+        rule are passed as keyword arguments.
+        """
+        raise NotImplementedError()
+
+    @classmethod
+    def as_view(
+        cls, name: str, *class_args: t.Any, **class_kwargs: t.Any
+    ) -> ft.RouteCallable:
+        """Convert the class into a view function that can be registered
+        for a route.
+
+        By default, the generated view will create a new instance of the
+        view class for every request and call its
+        :meth:`dispatch_request` method. If the view class sets
+        :attr:`init_every_request` to ``False``, the same instance will
+        be used for every request.
+
+        Except for ``name``, all other arguments passed to this method
+        are forwarded to the view class ``__init__`` method.
+
+        .. versionchanged:: 2.2
+            Added the ``init_every_request`` class attribute.
+        """
+        if cls.init_every_request:
+
+            def view(**kwargs: t.Any) -> ft.ResponseReturnValue:
+                self = view.view_class(  # type: ignore[attr-defined]
+                    *class_args, **class_kwargs
+                )
+                return current_app.ensure_sync(self.dispatch_request)(**kwargs)  # type: ignore[no-any-return]
+
+        else:
+            self = cls(*class_args, **class_kwargs)  # pyright: ignore
+
+            def view(**kwargs: t.Any) -> ft.ResponseReturnValue:
+                return current_app.ensure_sync(self.dispatch_request)(**kwargs)  # type: ignore[no-any-return]
+
+        if cls.decorators:
+            view.__name__ = name
+            view.__module__ = cls.__module__
+            for decorator in cls.decorators:
+                view = decorator(view)
+
+        # We attach the view class to the view function for two reasons:
+        # first of all it allows us to easily figure out what class-based
+        # view this thing came from, secondly it's also used for instantiating
+        # the view class so you can actually replace it with something else
+        # for testing purposes and debugging.
+        view.view_class = cls  # type: ignore
+        view.__name__ = name
+        view.__doc__ = cls.__doc__
+        view.__module__ = cls.__module__
+        view.methods = cls.methods  # type: ignore
+        view.provide_automatic_options = cls.provide_automatic_options  # type: ignore
+        return view
+
+
+class MethodView(View):
+    """Dispatches request methods to the corresponding instance methods.
+    For example, if you implement a ``get`` method, it will be used to
+    handle ``GET`` requests.
+
+    This can be useful for defining a REST API.
+
+    :attr:`methods` is automatically set based on the methods defined on
+    the class.
+
+    See :doc:`views` for a detailed guide.
+
+    .. code-block:: python
+
+        class CounterAPI(MethodView):
+            def get(self):
+                return str(session.get("counter", 0))
+
+            def post(self):
+                session["counter"] = session.get("counter", 0) + 1
+                return redirect(url_for("counter"))
+
+        app.add_url_rule(
+            "/counter", view_func=CounterAPI.as_view("counter")
+        )
+    """
+
+    def __init_subclass__(cls, **kwargs: t.Any) -> None:
+        super().__init_subclass__(**kwargs)
+
+        if "methods" not in cls.__dict__:
+            methods = set()
+
+            for base in cls.__bases__:
+                if getattr(base, "methods", None):
+                    methods.update(base.methods)  # type: ignore[attr-defined]
+
+            for key in http_method_funcs:
+                if hasattr(cls, key):
+                    methods.add(key.upper())
+
+            if methods:
+                cls.methods = methods
+
+    def dispatch_request(self, **kwargs: t.Any) -> ft.ResponseReturnValue:
+        meth = getattr(self, request.method.lower(), None)
+
+        # If the request method is HEAD and we don't have a handler for it
+        # retry with GET.
+        if meth is None and request.method == "HEAD":
+            meth = getattr(self, "get", None)
+
+        assert meth is not None, f"Unimplemented method {request.method!r}"
+        return current_app.ensure_sync(meth)(**kwargs)  # type: ignore[no-any-return]
diff --git a/backend/venv/lib/python3.12/site-packages/flask/wrappers.py b/backend/venv/lib/python3.12/site-packages/flask/wrappers.py
new file mode 100644
index 00000000..bab61029
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask/wrappers.py
@@ -0,0 +1,257 @@
+from __future__ import annotations
+
+import typing as t
+
+from werkzeug.exceptions import BadRequest
+from werkzeug.exceptions import HTTPException
+from werkzeug.wrappers import Request as RequestBase
+from werkzeug.wrappers import Response as ResponseBase
+
+from . import json
+from .globals import current_app
+from .helpers import _split_blueprint_path
+
+if t.TYPE_CHECKING:  # pragma: no cover
+    from werkzeug.routing import Rule
+
+
+class Request(RequestBase):
+    """The request object used by default in Flask.  Remembers the
+    matched endpoint and view arguments.
+
+    It is what ends up as :class:`~flask.request`.  If you want to replace
+    the request object used you can subclass this and set
+    :attr:`~flask.Flask.request_class` to your subclass.
+
+    The request object is a :class:`~werkzeug.wrappers.Request` subclass and
+    provides all of the attributes Werkzeug defines plus a few Flask
+    specific ones.
+    """
+
+    json_module: t.Any = json
+
+    #: The internal URL rule that matched the request.  This can be
+    #: useful to inspect which methods are allowed for the URL from
+    #: a before/after handler (``request.url_rule.methods``) etc.
+    #: Though if the request's method was invalid for the URL rule,
+    #: the valid list is available in ``routing_exception.valid_methods``
+    #: instead (an attribute of the Werkzeug exception
+    #: :exc:`~werkzeug.exceptions.MethodNotAllowed`)
+    #: because the request was never internally bound.
+    #:
+    #: .. versionadded:: 0.6
+    url_rule: Rule | None = None
+
+    #: A dict of view arguments that matched the request.  If an exception
+    #: happened when matching, this will be ``None``.
+    view_args: dict[str, t.Any] | None = None
+
+    #: If matching the URL failed, this is the exception that will be
+    #: raised / was raised as part of the request handling.  This is
+    #: usually a :exc:`~werkzeug.exceptions.NotFound` exception or
+    #: something similar.
+    routing_exception: HTTPException | None = None
+
+    _max_content_length: int | None = None
+    _max_form_memory_size: int | None = None
+    _max_form_parts: int | None = None
+
+    @property
+    def max_content_length(self) -> int | None:
+        """The maximum number of bytes that will be read during this request. If
+        this limit is exceeded, a 413 :exc:`~werkzeug.exceptions.RequestEntityTooLarge`
+        error is raised. If it is set to ``None``, no limit is enforced at the
+        Flask application level. However, if it is ``None`` and the request has
+        no ``Content-Length`` header and the WSGI server does not indicate that
+        it terminates the stream, then no data is read to avoid an infinite
+        stream.
+
+        Each request defaults to the :data:`MAX_CONTENT_LENGTH` config, which
+        defaults to ``None``. It can be set on a specific ``request`` to apply
+        the limit to that specific view. This should be set appropriately based
+        on an application's or view's specific needs.
+
+        .. versionchanged:: 3.1
+            This can be set per-request.
+
+        .. versionchanged:: 0.6
+            This is configurable through Flask config.
+        """
+        if self._max_content_length is not None:
+            return self._max_content_length
+
+        if not current_app:
+            return super().max_content_length
+
+        return current_app.config["MAX_CONTENT_LENGTH"]  # type: ignore[no-any-return]
+
+    @max_content_length.setter
+    def max_content_length(self, value: int | None) -> None:
+        self._max_content_length = value
+
+    @property
+    def max_form_memory_size(self) -> int | None:
+        """The maximum size in bytes any non-file form field may be in a
+        ``multipart/form-data`` body. If this limit is exceeded, a 413
+        :exc:`~werkzeug.exceptions.RequestEntityTooLarge` error is raised. If it
+        is set to ``None``, no limit is enforced at the Flask application level.
+
+        Each request defaults to the :data:`MAX_FORM_MEMORY_SIZE` config, which
+        defaults to ``500_000``. It can be set on a specific ``request`` to
+        apply the limit to that specific view. This should be set appropriately
+        based on an application's or view's specific needs.
+
+        .. versionchanged:: 3.1
+            This is configurable through Flask config.
+        """
+        if self._max_form_memory_size is not None:
+            return self._max_form_memory_size
+
+        if not current_app:
+            return super().max_form_memory_size
+
+        return current_app.config["MAX_FORM_MEMORY_SIZE"]  # type: ignore[no-any-return]
+
+    @max_form_memory_size.setter
+    def max_form_memory_size(self, value: int | None) -> None:
+        self._max_form_memory_size = value
+
+    @property  # type: ignore[override]
+    def max_form_parts(self) -> int | None:
+        """The maximum number of fields that may be present in a
+        ``multipart/form-data`` body. If this limit is exceeded, a 413
+        :exc:`~werkzeug.exceptions.RequestEntityTooLarge` error is raised. If it
+        is set to ``None``, no limit is enforced at the Flask application level.
+
+        Each request defaults to the :data:`MAX_FORM_PARTS` config, which
+        defaults to ``1_000``. It can be set on a specific ``request`` to apply
+        the limit to that specific view. This should be set appropriately based
+        on an application's or view's specific needs.
+
+        .. versionchanged:: 3.1
+            This is configurable through Flask config.
+        """
+        if self._max_form_parts is not None:
+            return self._max_form_parts
+
+        if not current_app:
+            return super().max_form_parts
+
+        return current_app.config["MAX_FORM_PARTS"]  # type: ignore[no-any-return]
+
+    @max_form_parts.setter
+    def max_form_parts(self, value: int | None) -> None:
+        self._max_form_parts = value
+
+    @property
+    def endpoint(self) -> str | None:
+        """The endpoint that matched the request URL.
+
+        This will be ``None`` if matching failed or has not been
+        performed yet.
+
+        This in combination with :attr:`view_args` can be used to
+        reconstruct the same URL or a modified URL.
+        """
+        if self.url_rule is not None:
+            return self.url_rule.endpoint  # type: ignore[no-any-return]
+
+        return None
+
+    @property
+    def blueprint(self) -> str | None:
+        """The registered name of the current blueprint.
+
+        This will be ``None`` if the endpoint is not part of a
+        blueprint, or if URL matching failed or has not been performed
+        yet.
+
+        This does not necessarily match the name the blueprint was
+        created with. It may have been nested, or registered with a
+        different name.
+        """
+        endpoint = self.endpoint
+
+        if endpoint is not None and "." in endpoint:
+            return endpoint.rpartition(".")[0]
+
+        return None
+
+    @property
+    def blueprints(self) -> list[str]:
+        """The registered names of the current blueprint upwards through
+        parent blueprints.
+
+        This will be an empty list if there is no current blueprint, or
+        if URL matching failed.
+
+        .. versionadded:: 2.0.1
+        """
+        name = self.blueprint
+
+        if name is None:
+            return []
+
+        return _split_blueprint_path(name)
+
+    def _load_form_data(self) -> None:
+        super()._load_form_data()
+
+        # In debug mode we're replacing the files multidict with an ad-hoc
+        # subclass that raises a different error for key errors.
+        if (
+            current_app
+            and current_app.debug
+            and self.mimetype != "multipart/form-data"
+            and not self.files
+        ):
+            from .debughelpers import attach_enctype_error_multidict
+
+            attach_enctype_error_multidict(self)
+
+    def on_json_loading_failed(self, e: ValueError | None) -> t.Any:
+        try:
+            return super().on_json_loading_failed(e)
+        except BadRequest as ebr:
+            if current_app and current_app.debug:
+                raise
+
+            raise BadRequest() from ebr
+
+
+class Response(ResponseBase):
+    """The response object that is used by default in Flask.  Works like the
+    response object from Werkzeug but is set to have an HTML mimetype by
+    default.  Quite often you don't have to create this object yourself because
+    :meth:`~flask.Flask.make_response` will take care of that for you.
+
+    If you want to replace the response object used you can subclass this and
+    set :attr:`~flask.Flask.response_class` to your subclass.
+
+    .. versionchanged:: 1.0
+        JSON support is added to the response, like the request. This is useful
+        when testing to get the test client response data as JSON.
+
+    .. versionchanged:: 1.0
+
+        Added :attr:`max_cookie_size`.
+    """
+
+    default_mimetype: str | None = "text/html"
+
+    json_module = json
+
+    autocorrect_location_header = False
+
+    @property
+    def max_cookie_size(self) -> int:  # type: ignore
+        """Read-only view of the :data:`MAX_COOKIE_SIZE` config key.
+
+        See :attr:`~werkzeug.wrappers.Response.max_cookie_size` in
+        Werkzeug's docs.
+        """
+        if current_app:
+            return current_app.config["MAX_COOKIE_SIZE"]  # type: ignore[no-any-return]
+
+        # return Werkzeug's default when not in an app context
+        return super().max_cookie_size
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__init__.py b/backend/venv/lib/python3.12/site-packages/flask_cors/__init__.py
new file mode 100644
index 00000000..458150ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_cors/__init__.py
@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+"""
+    flask_cors
+    ~~~~
+    Flask-CORS is a simple extension to Flask allowing you to support cross
+    origin resource sharing (CORS) using a simple decorator.
+
+    :copyright: (c) 2016 by Cory Dolphin.
+    :license: MIT, see LICENSE for more details.
+"""
+from .decorator import cross_origin
+from .extension import CORS
+from .version import __version__
+
+__all__ = ['CORS', 'cross_origin']
+
+# Set default logging handler to avoid "No handler found" warnings.
+import logging
+from logging import NullHandler
+
+# Set initial level to WARN. Users must manually enable logging for
+# flask_cors to see our logging.
+rootlogger = logging.getLogger(__name__)
+rootlogger.addHandler(NullHandler())
+
+if rootlogger.level == logging.NOTSET:
+    rootlogger.setLevel(logging.WARN)
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..785e3f13
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..7a70b724
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/decorator.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/decorator.cpython-312.pyc
new file mode 100644
index 00000000..a8126b22
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/decorator.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/extension.cpython-312.pyc
new file mode 100644
index 00000000..c0c511ac
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..b7fb41f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_cors/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/core.py b/backend/venv/lib/python3.12/site-packages/flask_cors/core.py
new file mode 100644
index 00000000..3fcc4d34
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_cors/core.py
@@ -0,0 +1,384 @@
+# -*- coding: utf-8 -*-
+"""
+    core
+    ~~~~
+    Core functionality shared between the extension and the decorator.
+
+    :copyright: (c) 2016 by Cory Dolphin.
+    :license: MIT, see LICENSE for more details.
+"""
+import re
+import logging
+from collections.abc import Iterable
+from datetime import timedelta
+from flask import request, current_app
+from werkzeug.datastructures import Headers, MultiDict
+
+LOG = logging.getLogger(__name__)
+
+# Response Headers
+ACL_ORIGIN = 'Access-Control-Allow-Origin'
+ACL_METHODS = 'Access-Control-Allow-Methods'
+ACL_ALLOW_HEADERS = 'Access-Control-Allow-Headers'
+ACL_EXPOSE_HEADERS = 'Access-Control-Expose-Headers'
+ACL_CREDENTIALS = 'Access-Control-Allow-Credentials'
+ACL_MAX_AGE = 'Access-Control-Max-Age'
+ACL_RESPONSE_PRIVATE_NETWORK = 'Access-Control-Allow-Private-Network'
+
+# Request Header
+ACL_REQUEST_METHOD = 'Access-Control-Request-Method'
+ACL_REQUEST_HEADERS = 'Access-Control-Request-Headers'
+ACL_REQUEST_HEADER_PRIVATE_NETWORK = 'Access-Control-Request-Private-Network'
+
+ALL_METHODS = ['GET', 'HEAD', 'POST', 'OPTIONS', 'PUT', 'PATCH', 'DELETE']
+CONFIG_OPTIONS = ['CORS_ORIGINS', 'CORS_METHODS', 'CORS_ALLOW_HEADERS',
+                  'CORS_EXPOSE_HEADERS', 'CORS_SUPPORTS_CREDENTIALS',
+                  'CORS_MAX_AGE', 'CORS_SEND_WILDCARD',
+                  'CORS_AUTOMATIC_OPTIONS', 'CORS_VARY_HEADER',
+                  'CORS_RESOURCES', 'CORS_INTERCEPT_EXCEPTIONS',
+                  'CORS_ALWAYS_SEND', 'CORS_ALLOW_PRIVATE_NETWORK']
+# Attribute added to request object by decorator to indicate that CORS
+# was evaluated, in case the decorator and extension are both applied
+# to a view.
+FLASK_CORS_EVALUATED = '_FLASK_CORS_EVALUATED'
+
+# Strange, but this gets the type of a compiled regex, which is otherwise not
+# exposed in a public API.
+RegexObject = type(re.compile(''))
+DEFAULT_OPTIONS = dict(origins='*',
+                       methods=ALL_METHODS,
+                       allow_headers='*',
+                       expose_headers=None,
+                       supports_credentials=False,
+                       max_age=None,
+                       send_wildcard=False,
+                       automatic_options=True,
+                       vary_header=True,
+                       resources=r'/*',
+                       intercept_exceptions=True,
+                       always_send=True,
+                       allow_private_network=False)
+
+
+def parse_resources(resources):
+    if isinstance(resources, dict):
+        # To make the API more consistent with the decorator, allow a
+        # resource of '*', which is not actually a valid regexp.
+        resources = [(re_fix(k), v) for k, v in resources.items()]
+
+        # Sort by regex length to provide consistency of matching and
+        # to provide a proxy for specificity of match. E.G. longer
+        # regular expressions are tried first.
+        def pattern_length(pair):
+            maybe_regex, _ = pair
+            return len(get_regexp_pattern(maybe_regex))
+
+        return sorted(resources,
+                      key=pattern_length,
+                      reverse=True)
+
+    elif isinstance(resources, str):
+        return [(re_fix(resources), {})]
+
+    elif isinstance(resources, Iterable):
+        return [(re_fix(r), {}) for r in resources]
+
+    # Type of compiled regex is not part of the public API. Test for this
+    # at runtime.
+    elif isinstance(resources,  RegexObject):
+        return [(re_fix(resources), {})]
+
+    else:
+        raise ValueError("Unexpected value for resources argument.")
+
+
+def get_regexp_pattern(regexp):
+    """
+    Helper that returns regexp pattern from given value.
+
+    :param regexp: regular expression to stringify
+    :type regexp: _sre.SRE_Pattern or str
+    :returns: string representation of given regexp pattern
+    :rtype: str
+    """
+    try:
+        return regexp.pattern
+    except AttributeError:
+        return str(regexp)
+
+
+def get_cors_origins(options, request_origin):
+    origins = options.get('origins')
+    wildcard = r'.*' in origins
+
+    # If the Origin header is not present terminate this set of steps.
+    # The request is outside the scope of this specification.-- W3Spec
+    if request_origin:
+        LOG.debug("CORS request received with 'Origin' %s", request_origin)
+
+        # If the allowed origins is an asterisk or 'wildcard', always match
+        if wildcard and options.get('send_wildcard'):
+            LOG.debug("Allowed origins are set to '*'. Sending wildcard CORS header.")
+            return ['*']
+        # If the value of the Origin header is a case-sensitive match
+        # for any of the values in list of origins
+        elif try_match_any(request_origin, origins):
+            LOG.debug("The request's Origin header matches. Sending CORS headers.", )
+            # Add a single Access-Control-Allow-Origin header, with either
+            # the value of the Origin header or the string "*" as value.
+            # -- W3Spec
+            return [request_origin]
+        else:
+            LOG.debug("The request's Origin header does not match any of allowed origins.")
+            return None
+
+
+    elif options.get('always_send'):
+        if wildcard:
+            # If wildcard is in the origins, even if 'send_wildcard' is False,
+            # simply send the wildcard. Unless supports_credentials is True,
+            # since that is forbidded by the spec..
+            # It is the most-likely to be correct thing to do (the only other
+            # option is to return nothing, which  almost certainly not what
+            # the developer wants if the '*' origin was specified.
+            if options.get('supports_credentials'):
+                return None
+            else:
+                return ['*']
+        else:
+            # Return all origins that are not regexes.
+            return sorted([o for o in origins if not probably_regex(o)])
+
+    # Terminate these steps, return the original request untouched.
+    else:
+        LOG.debug("The request did not contain an 'Origin' header. This means the browser or client did not request CORS, ensure the Origin Header is set.")
+        return None
+
+
+def get_allow_headers(options, acl_request_headers):
+    if acl_request_headers:
+        request_headers = [h.strip() for h in acl_request_headers.split(',')]
+
+        # any header that matches in the allow_headers
+        matching_headers = filter(
+            lambda h: try_match_any(h, options.get('allow_headers')),
+            request_headers
+        )
+
+        return ', '.join(sorted(matching_headers))
+
+    return None
+
+
+def get_cors_headers(options, request_headers, request_method):
+    origins_to_set = get_cors_origins(options, request_headers.get('Origin'))
+    headers = MultiDict()
+
+    if not origins_to_set:  # CORS is not enabled for this route
+        return headers
+
+    for origin in origins_to_set:
+        headers.add(ACL_ORIGIN, origin)
+
+    headers[ACL_EXPOSE_HEADERS] = options.get('expose_headers')
+
+    if options.get('supports_credentials'):
+        headers[ACL_CREDENTIALS] = 'true'  # case sensitive
+
+    if ACL_REQUEST_HEADER_PRIVATE_NETWORK in request_headers \
+            and request_headers.get(ACL_REQUEST_HEADER_PRIVATE_NETWORK) == 'true':
+        allow_private_network = 'true' if options.get('allow_private_network') else 'false'
+        headers[ACL_RESPONSE_PRIVATE_NETWORK] = allow_private_network
+
+    # This is a preflight request
+    # http://www.w3.org/TR/cors/#resource-preflight-requests
+    if request_method == 'OPTIONS':
+        acl_request_method = request_headers.get(ACL_REQUEST_METHOD, '').upper()
+
+        # If there is no Access-Control-Request-Method header or if parsing
+        # failed, do not set any additional headers
+        if acl_request_method and acl_request_method in options.get('methods'):
+
+            # If method is not a case-sensitive match for any of the values in
+            # list of methods do not set any additional headers and terminate
+            # this set of steps.
+            headers[ACL_ALLOW_HEADERS] = get_allow_headers(options, request_headers.get(ACL_REQUEST_HEADERS))
+            headers[ACL_MAX_AGE] = options.get('max_age')
+            headers[ACL_METHODS] = options.get('methods')
+        else:
+            LOG.info("The request's Access-Control-Request-Method header does not match allowed methods. CORS headers will not be applied.")
+
+    # http://www.w3.org/TR/cors/#resource-implementation
+    if options.get('vary_header'):
+        # Only set header if the origin returned will vary dynamically,
+        # i.e. if we are not returning an asterisk, and there are multiple
+        # origins that can be matched.
+        if headers[ACL_ORIGIN] == '*':
+            pass
+        elif (len(options.get('origins')) > 1 or
+              len(origins_to_set) > 1 or
+              any(map(probably_regex, options.get('origins')))):
+            headers.add('Vary', 'Origin')
+
+    return MultiDict((k, v) for k, v in headers.items() if v)
+
+
+def set_cors_headers(resp, options):
+    """
+    Performs the actual evaluation of Flask-CORS options and actually
+    modifies the response object.
+
+    This function is used both in the decorator and the after_request
+    callback
+    """
+
+    # If CORS has already been evaluated via the decorator, skip
+    if hasattr(resp, FLASK_CORS_EVALUATED):
+        LOG.debug('CORS have been already evaluated, skipping')
+        return resp
+
+    # Some libraries, like OAuthlib, set resp.headers to non Multidict
+    # objects (Werkzeug Headers work as well). This is a problem because
+    # headers allow repeated values.
+    if (not isinstance(resp.headers, Headers)
+           and not isinstance(resp.headers, MultiDict)):
+        resp.headers = MultiDict(resp.headers)
+
+    headers_to_set = get_cors_headers(options, request.headers, request.method)
+
+    LOG.debug('Settings CORS headers: %s', str(headers_to_set))
+
+    for k, v in headers_to_set.items():
+        resp.headers.add(k, v)
+
+    return resp
+
+def probably_regex(maybe_regex):
+    if isinstance(maybe_regex, RegexObject):
+        return True
+    else:
+        common_regex_chars = ['*', '\\', ']', '?', '$', '^', '[', ']', '(', ')']
+        # Use common characters used in regular expressions as a proxy
+        # for if this string is in fact a regex.
+        return any((c in maybe_regex for c in common_regex_chars))
+
+def re_fix(reg):
+    """
+        Replace the invalid regex r'*' with the valid, wildcard regex r'/.*' to
+        enable the CORS app extension to have a more user friendly api.
+    """
+    return r'.*' if reg == r'*' else reg
+
+
+def try_match_any(inst, patterns):
+    return any(try_match(inst, pattern) for pattern in patterns)
+
+
+def try_match(request_origin, maybe_regex):
+    """Safely attempts to match a pattern or string to a request origin."""
+    if isinstance(maybe_regex, RegexObject):
+        return re.match(maybe_regex, request_origin)
+    elif probably_regex(maybe_regex):
+        return re.match(maybe_regex, request_origin, flags=re.IGNORECASE)
+    else:
+        try:
+            return request_origin.lower() == maybe_regex.lower()
+        except AttributeError:
+            return request_origin == maybe_regex
+
+
+def get_cors_options(appInstance, *dicts):
+    """
+    Compute CORS options for an application by combining the DEFAULT_OPTIONS,
+    the app's configuration-specified options and any dictionaries passed. The
+    last specified option wins.
+    """
+    options = DEFAULT_OPTIONS.copy()
+    options.update(get_app_kwarg_dict(appInstance))
+    if dicts:
+        for d in dicts:
+            options.update(d)
+
+    return serialize_options(options)
+
+
+def get_app_kwarg_dict(appInstance=None):
+    """Returns the dictionary of CORS specific app configurations."""
+    app = (appInstance or current_app)
+
+    # In order to support blueprints which do not have a config attribute
+    app_config = getattr(app, 'config', {})
+
+    return {
+        k.lower().replace('cors_', ''): app_config.get(k)
+        for k in CONFIG_OPTIONS
+        if app_config.get(k) is not None
+    }
+
+
+def flexible_str(obj):
+    """
+    A more flexible str function which intelligently handles stringifying
+    strings, lists and other iterables. The results are lexographically sorted
+    to ensure generated responses are consistent when iterables such as Set
+    are used.
+    """
+    if obj is None:
+        return None
+    elif not isinstance(obj, str) and isinstance(obj, Iterable):
+        return ", ".join(str(item) for item in sorted(obj))
+    else:
+        return str(obj)
+
+
+def serialize_option(options_dict, key, upper=False):
+    if key in options_dict:
+        value = flexible_str(options_dict[key])
+        options_dict[key] = value.upper() if upper else value
+
+
+def ensure_iterable(inst):
+    """
+    Wraps scalars or string types as a list, or returns the iterable instance.
+    """
+    if isinstance(inst, str):
+        return [inst]
+    elif not isinstance(inst, Iterable):
+        return [inst]
+    else:
+        return inst
+
+def sanitize_regex_param(param):
+    return [re_fix(x) for x in ensure_iterable(param)]
+
+
+def serialize_options(opts):
+    """
+    A helper method to serialize and processes the options dictionary.
+    """
+    options = (opts or {}).copy()
+
+    for key in opts.keys():
+        if key not in DEFAULT_OPTIONS:
+            LOG.warning("Unknown option passed to Flask-CORS: %s", key)
+
+    # Ensure origins is a list of allowed origins with at least one entry.
+    options['origins'] = sanitize_regex_param(options.get('origins'))
+    options['allow_headers'] = sanitize_regex_param(options.get('allow_headers'))
+
+    # This is expressly forbidden by the spec. Raise a value error so people
+    # don't get burned in production.
+    if r'.*' in options['origins'] and options['supports_credentials'] and options['send_wildcard']:
+        raise ValueError("Cannot use supports_credentials in conjunction with"
+                         "an origin string of '*'. See: "
+                         "http://www.w3.org/TR/cors/#resource-requests")
+
+
+
+    serialize_option(options, 'expose_headers')
+    serialize_option(options, 'methods', upper=True)
+
+    if isinstance(options.get('max_age'), timedelta):
+        options['max_age'] = str(int(options['max_age'].total_seconds()))
+
+    return options
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/decorator.py b/backend/venv/lib/python3.12/site-packages/flask_cors/decorator.py
new file mode 100644
index 00000000..b61d22f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_cors/decorator.py
@@ -0,0 +1,137 @@
+# -*- coding: utf-8 -*-
+"""
+    decorator
+    ~~~~
+    This unit exposes a single decorator which should be used to wrap a
+    Flask route with. It accepts all parameters and options as
+    the CORS extension.
+
+    :copyright: (c) 2016 by Cory Dolphin.
+    :license: MIT, see LICENSE for more details.
+"""
+import logging
+from functools import update_wrapper
+from flask import make_response, request, current_app
+
+from .core import get_cors_options, set_cors_headers, FLASK_CORS_EVALUATED
+
+LOG = logging.getLogger(__name__)
+
+def cross_origin(*args, **kwargs):
+    """
+    This function is the decorator which is used to wrap a Flask route with.
+    In the simplest case, simply use the default parameters to allow all
+    origins in what is the most permissive configuration. If this method
+    modifies state or performs authentication which may be brute-forced, you
+    should add some degree of protection, such as Cross Site Request Forgery
+    protection.
+
+    :param origins:
+        The origin, or list of origins to allow requests from.
+        The origin(s) may be regular expressions, case-sensitive strings,
+        or else an asterisk
+
+        Default : '*'
+    :type origins: list, string or regex
+
+    :param methods:
+        The method or list of methods which the allowed origins are allowed to
+        access for non-simple requests.
+
+        Default : [GET, HEAD, POST, OPTIONS, PUT, PATCH, DELETE]
+    :type methods: list or string
+
+    :param expose_headers:
+        The header or list which are safe to expose to the API of a CORS API
+        specification.
+
+        Default : None
+    :type expose_headers: list or string
+
+    :param allow_headers:
+        The header or list of header field names which can be used when this
+        resource is accessed by allowed origins. The header(s) may be regular
+        expressions, case-sensitive strings, or else an asterisk.
+
+        Default : '*', allow all headers
+    :type allow_headers: list, string or regex
+
+    :param supports_credentials:
+        Allows users to make authenticated requests. If true, injects the
+        `Access-Control-Allow-Credentials` header in responses. This allows
+        cookies and credentials to be submitted across domains.
+
+        :note: This option cannot be used in conjunction with a '*' origin
+
+        Default : False
+    :type supports_credentials: bool
+
+    :param max_age:
+        The maximum time for which this CORS request maybe cached. This value
+        is set as the `Access-Control-Max-Age` header.
+
+        Default : None
+    :type max_age: timedelta, integer, string or None
+
+    :param send_wildcard: If True, and the origins parameter is `*`, a wildcard
+        `Access-Control-Allow-Origin` header is sent, rather than the
+        request's `Origin` header.
+
+        Default : False
+    :type send_wildcard: bool
+
+    :param vary_header:
+        If True, the header Vary: Origin will be returned as per the W3
+        implementation guidelines.
+
+        Setting this header when the `Access-Control-Allow-Origin` is
+        dynamically generated (e.g. when there is more than one allowed
+        origin, and an Origin than '*' is returned) informs CDNs and other
+        caches that the CORS headers are dynamic, and cannot be cached.
+
+        If False, the Vary header will never be injected or altered.
+
+        Default : True
+    :type vary_header: bool
+
+    :param automatic_options:
+        Only applies to the `cross_origin` decorator. If True, Flask-CORS will
+        override Flask's default OPTIONS handling to return CORS headers for
+        OPTIONS requests.
+
+        Default : True
+    :type automatic_options: bool
+
+    """
+    _options = kwargs
+
+    def decorator(f):
+        LOG.debug("Enabling %s for cross_origin using options:%s", f, _options)
+
+        # If True, intercept OPTIONS requests by modifying the view function,
+        # replicating Flask's default behavior, and wrapping the response with
+        # CORS headers.
+        #
+        # If f.provide_automatic_options is unset or True, Flask's route
+        # decorator (which is actually wraps the function object we return)
+        # intercepts OPTIONS handling, and requests will not have CORS headers
+        if _options.get('automatic_options', True):
+            f.required_methods = getattr(f, 'required_methods', set())
+            f.required_methods.add('OPTIONS')
+            f.provide_automatic_options = False
+
+        def wrapped_function(*args, **kwargs):
+            # Handle setting of Flask-Cors parameters
+            options = get_cors_options(current_app, _options)
+
+            if options.get('automatic_options') and request.method == 'OPTIONS':
+                resp = current_app.make_default_options_response()
+            else:
+                resp = make_response(f(*args, **kwargs))
+
+            set_cors_headers(resp, options)
+            setattr(resp, FLASK_CORS_EVALUATED, True)
+            return resp
+
+        return update_wrapper(wrapped_function, f)
+    return decorator
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/extension.py b/backend/venv/lib/python3.12/site-packages/flask_cors/extension.py
new file mode 100644
index 00000000..abf5cafe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_cors/extension.py
@@ -0,0 +1,219 @@
+# -*- coding: utf-8 -*-
+"""
+    extension
+    ~~~~
+    Flask-CORS is a simple extension to Flask allowing you to support cross
+    origin resource sharing (CORS) using a simple decorator.
+
+    :copyright: (c) 2016 by Cory Dolphin.
+    :license: MIT, see LICENSE for more details.
+"""
+import logging
+from urllib.parse import unquote_plus
+from flask import request
+
+from .core import (
+    parse_resources,
+    get_cors_options,
+    get_regexp_pattern,
+    ACL_ORIGIN,
+    try_match,
+    set_cors_headers
+)
+
+
+LOG = logging.getLogger(__name__)
+
+class CORS(object):
+    """
+    Initializes Cross Origin Resource sharing for the application. The
+    arguments are identical to :py:func:`cross_origin`, with the addition of a
+    `resources` parameter. The resources parameter defines a series of regular
+    expressions for resource paths to match and optionally, the associated
+    options to be applied to the particular resource. These options are
+    identical to the arguments to :py:func:`cross_origin`.
+
+    The settings for CORS are determined in the following order
+
+    1. Resource level settings (e.g when passed as a dictionary)
+    2. Keyword argument settings
+    3. App level configuration settings (e.g. CORS_*)
+    4. Default settings
+
+    Note: as it is possible for multiple regular expressions to match a
+    resource path, the regular expressions are first sorted by length,
+    from longest to shortest, in order to attempt to match the most
+    specific regular expression. This allows the definition of a
+    number of specific resource options, with a wildcard fallback
+    for all other resources.
+
+    :param resources:
+        The series of regular expression and (optionally) associated CORS
+        options to be applied to the given resource path.
+
+        If the argument is a dictionary, it's keys must be regular expressions,
+        and the values must be a dictionary of kwargs, identical to the kwargs
+        of this function.
+
+        If the argument is a list, it is expected to be a list of regular
+        expressions, for which the app-wide configured options are applied.
+
+        If the argument is a string, it is expected to be a regular expression
+        for which the app-wide configured options are applied.
+
+        Default : Match all and apply app-level configuration
+
+    :type resources: dict, iterable or string
+
+    :param origins:
+        The origin, or list of origins to allow requests from.
+        The origin(s) may be regular expressions, case-sensitive strings,
+        or else an asterisk.
+
+        ..  note::
+
+            origins must include the schema and the port (if not port 80),
+            e.g.,
+            `CORS(app, origins=["http://localhost:8000", "https://example.com"])`.
+
+        Default : '*'
+    :type origins: list, string or regex
+
+    :param methods:
+        The method or list of methods which the allowed origins are allowed to
+        access for non-simple requests.
+
+        Default : [GET, HEAD, POST, OPTIONS, PUT, PATCH, DELETE]
+    :type methods: list or string
+
+    :param expose_headers:
+        The header or list which are safe to expose to the API of a CORS API
+        specification.
+
+        Default : None
+    :type expose_headers: list or string
+
+    :param allow_headers:
+        The header or list of header field names which can be used when this
+        resource is accessed by allowed origins. The header(s) may be regular
+        expressions, case-sensitive strings, or else an asterisk.
+
+        Default : '*', allow all headers
+    :type allow_headers: list, string or regex
+
+    :param supports_credentials:
+        Allows users to make authenticated requests. If true, injects the
+        `Access-Control-Allow-Credentials` header in responses. This allows
+        cookies and credentials to be submitted across domains.
+
+        :note: This option cannot be used in conjunction with a '*' origin
+
+        Default : False
+    :type supports_credentials: bool
+
+    :param max_age:
+        The maximum time for which this CORS request maybe cached. This value
+        is set as the `Access-Control-Max-Age` header.
+
+        Default : None
+    :type max_age: timedelta, integer, string or None
+
+    :param send_wildcard: If True, and the origins parameter is `*`, a wildcard
+        `Access-Control-Allow-Origin` header is sent, rather than the
+        request's `Origin` header.
+
+        Default : False
+    :type send_wildcard: bool
+
+    :param vary_header:
+        If True, the header Vary: Origin will be returned as per the W3
+        implementation guidelines.
+
+        Setting this header when the `Access-Control-Allow-Origin` is
+        dynamically generated (e.g. when there is more than one allowed
+        origin, and an Origin than '*' is returned) informs CDNs and other
+        caches that the CORS headers are dynamic, and cannot be cached.
+
+        If False, the Vary header will never be injected or altered.
+
+        Default : True
+    :type vary_header: bool
+
+    :param allow_private_network:
+        If True, the response header `Access-Control-Allow-Private-Network`
+        will be set with the value 'true' whenever the request header
+        `Access-Control-Request-Private-Network` has a value 'true'.
+
+        If False, the reponse header `Access-Control-Allow-Private-Network`
+        will be set with the value 'false' whenever the request header
+        `Access-Control-Request-Private-Network` has a value of 'true'.
+
+        If the request header `Access-Control-Request-Private-Network` is
+        not present or has a value other than 'true', the response header
+        `Access-Control-Allow-Private-Network` will not be set.
+
+        Default : True
+    :type allow_private_network: bool
+    """
+
+    def __init__(self, app=None, **kwargs):
+        self._options = kwargs
+        if app is not None:
+            self.init_app(app, **kwargs)
+
+    def init_app(self, app, **kwargs):
+        # The resources and options may be specified in the App Config, the CORS constructor
+        # or the kwargs to the call to init_app.
+        options = get_cors_options(app, self._options, kwargs)
+
+        # Flatten our resources into a list of the form
+        # (pattern_or_regexp, dictionary_of_options)
+        resources = parse_resources(options.get('resources'))
+
+        # Compute the options for each resource by combining the options from
+        # the app's configuration, the constructor, the kwargs to init_app, and
+        # finally the options specified in the resources dictionary.
+        resources = [
+                     (pattern, get_cors_options(app, options, opts))
+                     for (pattern, opts) in resources
+                    ]
+
+        # Create a human-readable form of these resources by converting the compiled
+        # regular expressions into strings.
+        resources_human = {get_regexp_pattern(pattern): opts for (pattern,opts) in resources}
+        LOG.debug("Configuring CORS with resources: %s", resources_human)
+
+        cors_after_request = make_after_request_function(resources)
+        app.after_request(cors_after_request)
+
+        # Wrap exception handlers with cross_origin
+        # These error handlers will still respect the behavior of the route
+        if options.get('intercept_exceptions', True):
+            def _after_request_decorator(f):
+                def wrapped_function(*args, **kwargs):
+                    return cors_after_request(app.make_response(f(*args, **kwargs)))
+                return wrapped_function
+
+            if hasattr(app, 'handle_exception'):
+                app.handle_exception = _after_request_decorator(
+                    app.handle_exception)
+                app.handle_user_exception = _after_request_decorator(
+                    app.handle_user_exception)
+
+def make_after_request_function(resources):
+    def cors_after_request(resp):
+        # If CORS headers are set in a view decorator, pass
+        if resp.headers is not None and resp.headers.get(ACL_ORIGIN):
+            LOG.debug('CORS have been already evaluated, skipping')
+            return resp
+        normalized_path = unquote_plus(request.path)
+        for res_regex, res_options in resources:
+            if try_match(normalized_path, res_regex):
+                LOG.debug("Request to '%r' matches CORS resource '%s'. Using options: %s",
+                      request.path, get_regexp_pattern(res_regex), res_options)
+                set_cors_headers(resp, res_options)
+                break
+        else:
+            LOG.debug('No CORS rule matches')
+        return resp
+    return cors_after_request
diff --git a/backend/venv/lib/python3.12/site-packages/flask_cors/version.py b/backend/venv/lib/python3.12/site-packages/flask_cors/version.py
new file mode 100644
index 00000000..a0f66580
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_cors/version.py
@@ -0,0 +1 @@
+__version__ = '5.0.0'
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__init__.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__init__.py
new file mode 100644
index 00000000..e7fa1256
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__init__.py
@@ -0,0 +1,22 @@
+from .jwt_manager import JWTManager as JWTManager
+from .utils import create_access_token as create_access_token
+from .utils import create_refresh_token as create_refresh_token
+from .utils import current_user as current_user
+from .utils import decode_token as decode_token
+from .utils import get_csrf_token as get_csrf_token
+from .utils import get_current_user as get_current_user
+from .utils import get_jti as get_jti
+from .utils import get_jwt as get_jwt
+from .utils import get_jwt_header as get_jwt_header
+from .utils import get_jwt_identity as get_jwt_identity
+from .utils import get_jwt_request_location as get_jwt_request_location
+from .utils import get_unverified_jwt_headers as get_unverified_jwt_headers
+from .utils import set_access_cookies as set_access_cookies
+from .utils import set_refresh_cookies as set_refresh_cookies
+from .utils import unset_access_cookies as unset_access_cookies
+from .utils import unset_jwt_cookies as unset_jwt_cookies
+from .utils import unset_refresh_cookies as unset_refresh_cookies
+from .view_decorators import jwt_required as jwt_required
+from .view_decorators import verify_jwt_in_request as verify_jwt_in_request
+
+__version__ = "4.7.1"
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..4f7ba082
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..8048e7e2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/default_callbacks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/default_callbacks.cpython-312.pyc
new file mode 100644
index 00000000..77f0c01f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/default_callbacks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..727e2753
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/internal_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/internal_utils.cpython-312.pyc
new file mode 100644
index 00000000..6e56be21
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/internal_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/jwt_manager.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/jwt_manager.cpython-312.pyc
new file mode 100644
index 00000000..0a43a1b9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/jwt_manager.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/tokens.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/tokens.cpython-312.pyc
new file mode 100644
index 00000000..4652b219
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/tokens.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/typing.cpython-312.pyc
new file mode 100644
index 00000000..4600c22d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..94f2ede0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/view_decorators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/view_decorators.cpython-312.pyc
new file mode 100644
index 00000000..8fc79a65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/__pycache__/view_decorators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/config.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/config.py
new file mode 100644
index 00000000..7d256516
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/config.py
@@ -0,0 +1,319 @@
+from datetime import datetime
+from datetime import timedelta
+from datetime import timezone
+from json import JSONEncoder
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Type
+from typing import Union
+
+from flask import current_app
+from jwt.algorithms import requires_cryptography
+
+from flask_jwt_extended.internal_utils import get_json_encoder
+from flask_jwt_extended.typing import ExpiresDelta
+
+
+class _Config(object):
+    """
+    Helper object for accessing and verifying options in this extension. This
+    is meant for internal use of the application; modifying config options
+    should be done with flasks ```app.config```.
+
+    Default values for the configuration options are set in the jwt_manager
+    object. All of these values are read only. This is simply a loose wrapper
+    with some helper functionality for flasks `app.config`.
+    """
+
+    @property
+    def is_asymmetric(self) -> bool:
+        return self.algorithm in requires_cryptography
+
+    @property
+    def encode_key(self) -> str:
+        return self._private_key if self.is_asymmetric else self._secret_key
+
+    @property
+    def decode_key(self) -> str:
+        return self._public_key if self.is_asymmetric else self._secret_key
+
+    @property
+    def token_location(self) -> Sequence[str]:
+        locations = current_app.config["JWT_TOKEN_LOCATION"]
+        if isinstance(locations, str):
+            locations = (locations,)
+        elif not isinstance(locations, Iterable):
+            raise RuntimeError("JWT_TOKEN_LOCATION must be a sequence or a set")
+        elif not locations:
+            raise RuntimeError(
+                "JWT_TOKEN_LOCATION must contain at least one "
+                'of "headers", "cookies", "query_string", or "json"'
+            )
+        for location in locations:
+            if location not in ("headers", "cookies", "query_string", "json"):
+                raise RuntimeError(
+                    "JWT_TOKEN_LOCATION can only contain "
+                    '"headers", "cookies", "query_string", or "json"'
+                )
+        return locations
+
+    @property
+    def jwt_in_cookies(self) -> bool:
+        return "cookies" in self.token_location
+
+    @property
+    def jwt_in_headers(self) -> bool:
+        return "headers" in self.token_location
+
+    @property
+    def jwt_in_query_string(self) -> bool:
+        return "query_string" in self.token_location
+
+    @property
+    def jwt_in_json(self) -> bool:
+        return "json" in self.token_location
+
+    @property
+    def header_name(self) -> str:
+        name = current_app.config["JWT_HEADER_NAME"]
+        if not name:
+            raise RuntimeError("JWT_ACCESS_HEADER_NAME cannot be empty")
+        return name
+
+    @property
+    def header_type(self) -> str:
+        return current_app.config["JWT_HEADER_TYPE"]
+
+    @property
+    def query_string_name(self) -> str:
+        return current_app.config["JWT_QUERY_STRING_NAME"]
+
+    @property
+    def query_string_value_prefix(self) -> str:
+        return current_app.config["JWT_QUERY_STRING_VALUE_PREFIX"]
+
+    @property
+    def access_cookie_name(self) -> str:
+        return current_app.config["JWT_ACCESS_COOKIE_NAME"]
+
+    @property
+    def refresh_cookie_name(self) -> str:
+        return current_app.config["JWT_REFRESH_COOKIE_NAME"]
+
+    @property
+    def access_cookie_path(self) -> str:
+        return current_app.config["JWT_ACCESS_COOKIE_PATH"]
+
+    @property
+    def refresh_cookie_path(self) -> str:
+        return current_app.config["JWT_REFRESH_COOKIE_PATH"]
+
+    @property
+    def cookie_secure(self) -> bool:
+        return current_app.config["JWT_COOKIE_SECURE"]
+
+    @property
+    def cookie_domain(self) -> str:
+        return current_app.config["JWT_COOKIE_DOMAIN"]
+
+    @property
+    def session_cookie(self) -> bool:
+        return current_app.config["JWT_SESSION_COOKIE"]
+
+    @property
+    def cookie_samesite(self) -> str:
+        return current_app.config["JWT_COOKIE_SAMESITE"]
+
+    @property
+    def json_key(self) -> str:
+        return current_app.config["JWT_JSON_KEY"]
+
+    @property
+    def refresh_json_key(self) -> str:
+        return current_app.config["JWT_REFRESH_JSON_KEY"]
+
+    @property
+    def cookie_csrf_protect(self) -> bool:
+        return current_app.config["JWT_COOKIE_CSRF_PROTECT"]
+
+    @property
+    def csrf_request_methods(self) -> Iterable[str]:
+        return current_app.config["JWT_CSRF_METHODS"]
+
+    @property
+    def csrf_in_cookies(self) -> bool:
+        return current_app.config["JWT_CSRF_IN_COOKIES"]
+
+    @property
+    def access_csrf_cookie_name(self) -> str:
+        return current_app.config["JWT_ACCESS_CSRF_COOKIE_NAME"]
+
+    @property
+    def refresh_csrf_cookie_name(self) -> str:
+        return current_app.config["JWT_REFRESH_CSRF_COOKIE_NAME"]
+
+    @property
+    def access_csrf_cookie_path(self) -> str:
+        return current_app.config["JWT_ACCESS_CSRF_COOKIE_PATH"]
+
+    @property
+    def refresh_csrf_cookie_path(self) -> str:
+        return current_app.config["JWT_REFRESH_CSRF_COOKIE_PATH"]
+
+    @property
+    def access_csrf_header_name(self) -> str:
+        return current_app.config["JWT_ACCESS_CSRF_HEADER_NAME"]
+
+    @property
+    def refresh_csrf_header_name(self) -> str:
+        return current_app.config["JWT_REFRESH_CSRF_HEADER_NAME"]
+
+    @property
+    def csrf_check_form(self) -> bool:
+        return current_app.config["JWT_CSRF_CHECK_FORM"]
+
+    @property
+    def access_csrf_field_name(self) -> str:
+        return current_app.config["JWT_ACCESS_CSRF_FIELD_NAME"]
+
+    @property
+    def refresh_csrf_field_name(self) -> str:
+        return current_app.config["JWT_REFRESH_CSRF_FIELD_NAME"]
+
+    @property
+    def access_expires(self) -> ExpiresDelta:
+        delta = current_app.config["JWT_ACCESS_TOKEN_EXPIRES"]
+        if type(delta) is int:
+            delta = timedelta(seconds=delta)
+        if delta is not False:
+            try:
+                # Basically runtime typechecking. Probably a better way to do
+                # this with proper type checking
+                delta + datetime.now(timezone.utc)
+            except TypeError as e:
+                err = (
+                    "must be able to add JWT_ACCESS_TOKEN_EXPIRES to datetime.datetime"
+                )
+                raise RuntimeError(err) from e
+        return delta
+
+    @property
+    def refresh_expires(self) -> ExpiresDelta:
+        delta = current_app.config["JWT_REFRESH_TOKEN_EXPIRES"]
+        if type(delta) is int:
+            delta = timedelta(seconds=delta)
+        if delta is not False:
+            # Basically runtime typechecking. Probably a better way to do
+            # this with proper type checking
+            try:
+                delta + datetime.now(timezone.utc)
+            except TypeError as e:
+                err = (
+                    "must be able to add JWT_REFRESH_TOKEN_EXPIRES to datetime.datetime"
+                )
+                raise RuntimeError(err) from e
+        return delta
+
+    @property
+    def algorithm(self) -> str:
+        return current_app.config["JWT_ALGORITHM"]
+
+    @property
+    def decode_algorithms(self) -> List[str]:
+        algorithms = current_app.config["JWT_DECODE_ALGORITHMS"]
+        if not algorithms:
+            return [self.algorithm]
+        if self.algorithm not in algorithms:
+            algorithms.append(self.algorithm)
+        return algorithms
+
+    @property
+    def _secret_key(self) -> str:
+        key = current_app.config["JWT_SECRET_KEY"]
+        if not key:
+            key = current_app.config.get("SECRET_KEY", None)
+            if not key:
+                raise RuntimeError(
+                    "JWT_SECRET_KEY or flask SECRET_KEY "
+                    "must be set when using symmetric "
+                    'algorithm "{}"'.format(self.algorithm)
+                )
+        return key
+
+    @property
+    def _public_key(self) -> str:
+        key = current_app.config["JWT_PUBLIC_KEY"]
+        if not key:
+            raise RuntimeError(
+                "JWT_PUBLIC_KEY must be set to use "
+                "asymmetric cryptography algorithm "
+                '"{}"'.format(self.algorithm)
+            )
+        return key
+
+    @property
+    def _private_key(self) -> str:
+        key = current_app.config["JWT_PRIVATE_KEY"]
+        if not key:
+            raise RuntimeError(
+                "JWT_PRIVATE_KEY must be set to use "
+                "asymmetric cryptography algorithm "
+                '"{}"'.format(self.algorithm)
+            )
+        return key
+
+    @property
+    def cookie_max_age(self) -> Optional[int]:
+        # Returns the appropiate value for max_age for flask set_cookies. If
+        # session cookie is true, return None, otherwise return a number of
+        # seconds 1 year in the future
+        return None if self.session_cookie else 31540000  # 1 year
+
+    @property
+    def identity_claim_key(self) -> str:
+        return current_app.config["JWT_IDENTITY_CLAIM"]
+
+    @property
+    def exempt_methods(self) -> Iterable[str]:
+        return {"OPTIONS"}
+
+    @property
+    def error_msg_key(self) -> str:
+        return current_app.config["JWT_ERROR_MESSAGE_KEY"]
+
+    @property
+    def json_encoder(self) -> Type[JSONEncoder]:
+        return get_json_encoder(current_app)
+
+    @property
+    def decode_audience(self) -> Union[str, Iterable[str]]:
+        return current_app.config["JWT_DECODE_AUDIENCE"]
+
+    @property
+    def encode_audience(self) -> Union[str, Iterable[str]]:
+        return current_app.config["JWT_ENCODE_AUDIENCE"]
+
+    @property
+    def encode_issuer(self) -> str:
+        return current_app.config["JWT_ENCODE_ISSUER"]
+
+    @property
+    def decode_issuer(self) -> str:
+        return current_app.config["JWT_DECODE_ISSUER"]
+
+    @property
+    def leeway(self) -> int:
+        return current_app.config["JWT_DECODE_LEEWAY"]
+
+    @property
+    def verify_sub(self) -> bool:
+        return current_app.config["JWT_VERIFY_SUB"]
+
+    @property
+    def encode_nbf(self) -> bool:
+        return current_app.config["JWT_ENCODE_NBF"]
+
+
+config = _Config()
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/default_callbacks.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/default_callbacks.py
new file mode 100644
index 00000000..6f10e5da
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/default_callbacks.py
@@ -0,0 +1,161 @@
+"""
+These are the default methods implementations that are used in this extension.
+All of these can be updated on an app by app basis using the JWTManager
+loader decorators. For further information, check out the following links:
+
+http://flask-jwt-extended.readthedocs.io/en/latest/changing_default_behavior.html
+http://flask-jwt-extended.readthedocs.io/en/latest/tokens_from_complex_object.html
+"""
+from http import HTTPStatus
+from typing import Any
+
+from flask import jsonify
+from flask.typing import ResponseReturnValue
+
+from flask_jwt_extended.config import config
+
+
+def default_additional_claims_callback(userdata: Any) -> dict:
+    """
+    By default, we add no additional claims to the access tokens.
+
+    :param userdata: data passed in as the ```identity``` argument to the
+                     ```create_access_token``` and ```create_refresh_token```
+                     functions
+    """
+    return {}
+
+
+def default_blocklist_callback(jwt_headers: dict, jwt_data: dict) -> bool:
+    return False
+
+
+def default_jwt_headers_callback(default_headers) -> dict:
+    """
+    By default header typically consists of two parts: the type of the token,
+    which is JWT, and the signing algorithm being used, such as HMAC SHA256
+    or RSA. But we don't set the default header here we set it as empty which
+    further by default set while encoding the token
+    :return: default we set None here
+    """
+    return {}
+
+
+def default_user_identity_callback(userdata: Any) -> str:
+    """
+    By default, we use the passed in object directly as the jwt identity.
+    See this for additional info:
+
+    :param userdata: data passed in as the ```identity``` argument to the
+                     ```create_access_token``` and ```create_refresh_token```
+                     functions
+    """
+    return userdata
+
+
+def default_expired_token_callback(
+    _expired_jwt_header: dict, _expired_jwt_data: dict
+) -> ResponseReturnValue:
+    """
+    By default, if an expired token attempts to access a protected endpoint,
+    we return a generic error message with a 401 status
+    """
+    return jsonify({config.error_msg_key: "Token has expired"}), HTTPStatus.UNAUTHORIZED
+
+
+def default_invalid_token_callback(error_string: str) -> ResponseReturnValue:
+    """
+    By default, if an invalid token attempts to access a protected endpoint, we
+    return the error string for why it is not valid with a 422 status code
+
+    :param error_string: String indicating why the token is invalid
+    """
+    return (
+        jsonify({config.error_msg_key: error_string}),
+        HTTPStatus.UNPROCESSABLE_ENTITY,
+    )
+
+
+def default_unauthorized_callback(error_string: str) -> ResponseReturnValue:
+    """
+    By default, if a protected endpoint is accessed without a JWT, we return
+    the error string indicating why this is unauthorized, with a 401 status code
+
+    :param error_string: String indicating why this request is unauthorized
+    """
+    return jsonify({config.error_msg_key: error_string}), HTTPStatus.UNAUTHORIZED
+
+
+def default_needs_fresh_token_callback(
+    jwt_header: dict, jwt_data: dict
+) -> ResponseReturnValue:
+    """
+    By default, if a non-fresh jwt is used to access a ```fresh_jwt_required```
+    endpoint, we return a general error message with a 401 status code
+    """
+    return (
+        jsonify({config.error_msg_key: "Fresh token required"}),
+        HTTPStatus.UNAUTHORIZED,
+    )
+
+
+def default_revoked_token_callback(
+    jwt_header: dict, jwt_data: dict
+) -> ResponseReturnValue:
+    """
+    By default, if a revoked token is used to access a protected endpoint, we
+    return a general error message with a 401 status code
+    """
+    return (
+        jsonify({config.error_msg_key: "Token has been revoked"}),
+        HTTPStatus.UNAUTHORIZED,
+    )
+
+
+def default_user_lookup_error_callback(
+    _jwt_header: dict, jwt_data: dict
+) -> ResponseReturnValue:
+    """
+    By default, if a user_lookup callback is defined and the callback
+    function returns None, we return a general error message with a 401
+    status code
+    """
+    identity = jwt_data[config.identity_claim_key]
+    result = {config.error_msg_key: f"Error loading the user {identity}"}
+    return jsonify(result), HTTPStatus.UNAUTHORIZED
+
+
+def default_token_verification_callback(_jwt_header: dict, _jwt_data: dict) -> bool:
+    """
+    By default, we do not do any verification of the user claims.
+    """
+    return True
+
+
+def default_token_verification_failed_callback(
+    _jwt_header: dict, _jwt_data: dict
+) -> ResponseReturnValue:
+    """
+    By default, if the user claims verification failed, we return a generic
+    error message with a 400 status code
+    """
+    return (
+        jsonify({config.error_msg_key: "User claims verification failed"}),
+        HTTPStatus.BAD_REQUEST,
+    )
+
+
+def default_decode_key_callback(jwt_header: dict, jwt_data: dict) -> str:
+    """
+    By default, the decode key specified via the JWT_SECRET_KEY or
+    JWT_PUBLIC_KEY settings will be used to decode all tokens
+    """
+    return config.decode_key
+
+
+def default_encode_key_callback(identity: Any) -> str:
+    """
+    By default, the encode key specified via the JWT_SECRET_KEY or
+    JWT_PRIVATE_KEY settings will be used to encode all tokens
+    """
+    return config.encode_key
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/exceptions.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/exceptions.py
new file mode 100644
index 00000000..2ad414a8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/exceptions.py
@@ -0,0 +1,102 @@
+class JWTExtendedException(Exception):
+    """
+    Base except which all flask_jwt_extended errors extend
+    """
+
+    pass
+
+
+class JWTDecodeError(JWTExtendedException):
+    """
+    An error decoding a JWT
+    """
+
+    pass
+
+
+class InvalidHeaderError(JWTExtendedException):
+    """
+    An error getting header information from a request
+    """
+
+    pass
+
+
+class InvalidQueryParamError(JWTExtendedException):
+    """
+    An error when a query string param is not in the correct format
+    """
+
+    pass
+
+
+class NoAuthorizationError(JWTExtendedException):
+    """
+    An error raised when no authorization token was found in a protected endpoint
+    """
+
+    pass
+
+
+class CSRFError(JWTExtendedException):
+    """
+    An error with CSRF protection
+    """
+
+    pass
+
+
+class WrongTokenError(JWTExtendedException):
+    """
+    Error raised when attempting to use a refresh token to access an endpoint
+    or vice versa
+    """
+
+    pass
+
+
+class RevokedTokenError(JWTExtendedException):
+    """
+    Error raised when a revoked token attempt to access a protected endpoint
+    """
+
+    def __init__(self, jwt_header: dict, jwt_data: dict) -> None:
+        super().__init__("Token has been revoked")
+        self.jwt_header = jwt_header
+        self.jwt_data = jwt_data
+
+
+class FreshTokenRequired(JWTExtendedException):
+    """
+    Error raised when a valid, non-fresh JWT attempt to access an endpoint
+    protected by fresh_jwt_required
+    """
+
+    def __init__(self, message, jwt_header: dict, jwt_data: dict) -> None:
+        super().__init__(message)
+        self.jwt_header = jwt_header
+        self.jwt_data = jwt_data
+
+
+class UserLookupError(JWTExtendedException):
+    """
+    Error raised when a user_lookup callback function returns None, indicating
+    that it cannot or will not load a user for the given identity.
+    """
+
+    def __init__(self, message, jwt_header: dict, jwt_data: dict) -> None:
+        super().__init__(message)
+        self.jwt_header = jwt_header
+        self.jwt_data = jwt_data
+
+
+class UserClaimsVerificationError(JWTExtendedException):
+    """
+    Error raised when the claims_verification_callback function returns False,
+    indicating that the expected user claims are invalid
+    """
+
+    def __init__(self, message, jwt_header: dict, jwt_data: dict) -> None:
+        super().__init__(message)
+        self.jwt_header = jwt_header
+        self.jwt_data = jwt_data
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/internal_utils.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/internal_utils.py
new file mode 100644
index 00000000..dfeaa9ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/internal_utils.py
@@ -0,0 +1,98 @@
+import json
+from typing import Any
+from typing import Type
+from typing import TYPE_CHECKING
+
+from flask import current_app
+from flask import Flask
+
+from flask_jwt_extended.exceptions import RevokedTokenError
+from flask_jwt_extended.exceptions import UserClaimsVerificationError
+from flask_jwt_extended.exceptions import WrongTokenError
+
+try:
+    from flask.json.provider import DefaultJSONProvider
+
+    HAS_JSON_PROVIDER = True
+except ModuleNotFoundError:  # pragma: no cover
+    # The flask.json.provider module was added in Flask 2.2.
+    # Further details are handled in get_json_encoder.
+    HAS_JSON_PROVIDER = False
+
+
+if TYPE_CHECKING:  # pragma: no cover
+    from flask_jwt_extended import JWTManager
+
+
+def get_jwt_manager() -> "JWTManager":
+    try:
+        return current_app.extensions["flask-jwt-extended"]
+    except KeyError:  # pragma: no cover
+        raise RuntimeError(
+            "You must initialize a JWTManager with this flask "
+            "application before using this method"
+        ) from None
+
+
+def has_user_lookup() -> bool:
+    jwt_manager = get_jwt_manager()
+    return jwt_manager._user_lookup_callback is not None
+
+
+def user_lookup(*args, **kwargs) -> Any:
+    jwt_manager = get_jwt_manager()
+    return jwt_manager._user_lookup_callback and jwt_manager._user_lookup_callback(
+        *args, **kwargs
+    )
+
+
+def verify_token_type(decoded_token: dict, refresh: bool) -> None:
+    if not refresh and decoded_token["type"] == "refresh":
+        raise WrongTokenError("Only non-refresh tokens are allowed")
+    elif refresh and decoded_token["type"] != "refresh":
+        raise WrongTokenError("Only refresh tokens are allowed")
+
+
+def verify_token_not_blocklisted(jwt_header: dict, jwt_data: dict) -> None:
+    jwt_manager = get_jwt_manager()
+    if jwt_manager._token_in_blocklist_callback(jwt_header, jwt_data):
+        raise RevokedTokenError(jwt_header, jwt_data)
+
+
+def custom_verification_for_token(jwt_header: dict, jwt_data: dict) -> None:
+    jwt_manager = get_jwt_manager()
+    if not jwt_manager._token_verification_callback(jwt_header, jwt_data):
+        error_msg = "User claims verification failed"
+        raise UserClaimsVerificationError(error_msg, jwt_header, jwt_data)
+
+
+class JSONEncoder(json.JSONEncoder):
+    """A JSON encoder which uses the app.json_provider_class for the default"""
+
+    def default(self, o: Any) -> Any:
+        # If the registered JSON provider does not implement a default classmethod
+        # use the method defined by the DefaultJSONProvider
+        default = getattr(
+            current_app.json_provider_class, "default", DefaultJSONProvider.default
+        )
+        return default(o)
+
+
+def get_json_encoder(app: Flask) -> Type[json.JSONEncoder]:
+    """Get the JSON Encoder for the provided flask app
+
+    Starting with flask version 2.2 the flask application provides a
+    interface to register a custom JSON Encoder/Decoder under the json_provider_class.
+    As this interface is not compatible with the standard JSONEncoder, the `default`
+    method of the class is wrapped.
+
+    Lookup Order:
+      - app.json_encoder - For Flask < 2.2
+      - app.json_provider_class.default
+      - flask.json.provider.DefaultJSONProvider.default
+
+    """
+    if not HAS_JSON_PROVIDER:  # pragma: no cover
+        return app.json_encoder  # type: ignore
+
+    return JSONEncoder
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/jwt_manager.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/jwt_manager.py
new file mode 100644
index 00000000..7f0d967d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/jwt_manager.py
@@ -0,0 +1,564 @@
+import datetime
+from typing import Any
+from typing import Callable
+from typing import Optional
+
+import jwt
+from flask import Flask
+from jwt import DecodeError
+from jwt import ExpiredSignatureError
+from jwt import InvalidAudienceError
+from jwt import InvalidIssuerError
+from jwt import InvalidTokenError
+from jwt import MissingRequiredClaimError
+
+from flask_jwt_extended.config import config
+from flask_jwt_extended.default_callbacks import default_additional_claims_callback
+from flask_jwt_extended.default_callbacks import default_blocklist_callback
+from flask_jwt_extended.default_callbacks import default_decode_key_callback
+from flask_jwt_extended.default_callbacks import default_encode_key_callback
+from flask_jwt_extended.default_callbacks import default_expired_token_callback
+from flask_jwt_extended.default_callbacks import default_invalid_token_callback
+from flask_jwt_extended.default_callbacks import default_jwt_headers_callback
+from flask_jwt_extended.default_callbacks import default_needs_fresh_token_callback
+from flask_jwt_extended.default_callbacks import default_revoked_token_callback
+from flask_jwt_extended.default_callbacks import default_token_verification_callback
+from flask_jwt_extended.default_callbacks import (
+    default_token_verification_failed_callback,
+)
+from flask_jwt_extended.default_callbacks import default_unauthorized_callback
+from flask_jwt_extended.default_callbacks import default_user_identity_callback
+from flask_jwt_extended.default_callbacks import default_user_lookup_error_callback
+from flask_jwt_extended.exceptions import CSRFError
+from flask_jwt_extended.exceptions import FreshTokenRequired
+from flask_jwt_extended.exceptions import InvalidHeaderError
+from flask_jwt_extended.exceptions import InvalidQueryParamError
+from flask_jwt_extended.exceptions import JWTDecodeError
+from flask_jwt_extended.exceptions import NoAuthorizationError
+from flask_jwt_extended.exceptions import RevokedTokenError
+from flask_jwt_extended.exceptions import UserClaimsVerificationError
+from flask_jwt_extended.exceptions import UserLookupError
+from flask_jwt_extended.exceptions import WrongTokenError
+from flask_jwt_extended.tokens import _decode_jwt
+from flask_jwt_extended.tokens import _encode_jwt
+from flask_jwt_extended.typing import ExpiresDelta
+from flask_jwt_extended.typing import Fresh
+from flask_jwt_extended.utils import current_user_context_processor
+
+
+class JWTManager(object):
+    """
+    An object used to hold JWT settings and callback functions for the
+    Flask-JWT-Extended extension.
+
+    Instances of :class:`JWTManager` are *not* bound to specific apps, so
+    you can create one in the main body of your code and then bind it
+    to your app in a factory function.
+    """
+
+    def __init__(
+        self, app: Optional[Flask] = None, add_context_processor: bool = False
+    ) -> None:
+        """
+        Create the JWTManager instance. You can either pass a flask application
+        in directly here to register this extension with the flask app, or
+        call init_app after creating this object (in a factory pattern).
+
+        :param app:
+            The Flask Application object
+        :param add_context_processor:
+            Controls if `current_user` is should be added to flasks template
+            context (and thus be available for use in Jinja templates). Defaults
+            to ``False``.
+        """
+        # Register the default error handler callback methods. These can be
+        # overridden with the appropriate loader decorators
+        self._decode_key_callback = default_decode_key_callback
+        self._encode_key_callback = default_encode_key_callback
+        self._expired_token_callback = default_expired_token_callback
+        self._invalid_token_callback = default_invalid_token_callback
+        self._jwt_additional_header_callback = default_jwt_headers_callback
+        self._needs_fresh_token_callback = default_needs_fresh_token_callback
+        self._revoked_token_callback = default_revoked_token_callback
+        self._token_in_blocklist_callback = default_blocklist_callback
+        self._token_verification_callback = default_token_verification_callback
+        self._unauthorized_callback = default_unauthorized_callback
+        self._user_claims_callback = default_additional_claims_callback
+        self._user_identity_callback = default_user_identity_callback
+        self._user_lookup_callback: Optional[Callable] = None
+        self._user_lookup_error_callback = default_user_lookup_error_callback
+        self._token_verification_failed_callback = (
+            default_token_verification_failed_callback
+        )
+
+        # Register this extension with the flask app now (if it is provided)
+        if app is not None:
+            self.init_app(app, add_context_processor)
+
+    def init_app(self, app: Flask, add_context_processor: bool = False) -> None:
+        """
+        Register this extension with the flask app.
+
+        :param app:
+            The Flask Application object
+        :param add_context_processor:
+            Controls if `current_user` is should be added to flasks template
+            context (and thus be available for use in Jinja templates). Defaults
+            to ``False``.
+        """
+        # Save this so we can use it later in the extension
+        if not hasattr(app, "extensions"):  # pragma: no cover
+            app.extensions = {}
+        app.extensions["flask-jwt-extended"] = self
+
+        if add_context_processor:
+            app.context_processor(current_user_context_processor)
+
+        # Set all the default configurations for this extension
+        self._set_default_configuration_options(app)
+        self._set_error_handler_callbacks(app)
+
+    def _set_error_handler_callbacks(self, app: Flask) -> None:
+        @app.errorhandler(CSRFError)
+        def handle_csrf_error(e):
+            return self._unauthorized_callback(str(e))
+
+        @app.errorhandler(DecodeError)
+        def handle_decode_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(ExpiredSignatureError)
+        def handle_expired_error(e):
+            return self._expired_token_callback(e.jwt_header, e.jwt_data)
+
+        @app.errorhandler(FreshTokenRequired)
+        def handle_fresh_token_required(e):
+            return self._needs_fresh_token_callback(e.jwt_header, e.jwt_data)
+
+        @app.errorhandler(MissingRequiredClaimError)
+        def handle_missing_required_claim_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(InvalidAudienceError)
+        def handle_invalid_audience_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(InvalidIssuerError)
+        def handle_invalid_issuer_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(InvalidHeaderError)
+        def handle_invalid_header_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(InvalidTokenError)
+        def handle_invalid_token_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(JWTDecodeError)
+        def handle_jwt_decode_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(NoAuthorizationError)
+        def handle_auth_error(e):
+            return self._unauthorized_callback(str(e))
+
+        @app.errorhandler(InvalidQueryParamError)
+        def handle_invalid_query_param_error(e):
+            return self._invalid_token_callback(str(e))
+
+        @app.errorhandler(RevokedTokenError)
+        def handle_revoked_token_error(e):
+            return self._revoked_token_callback(e.jwt_header, e.jwt_data)
+
+        @app.errorhandler(UserClaimsVerificationError)
+        def handle_failed_token_verification(e):
+            return self._token_verification_failed_callback(e.jwt_header, e.jwt_data)
+
+        @app.errorhandler(UserLookupError)
+        def handler_user_lookup_error(e):
+            return self._user_lookup_error_callback(e.jwt_header, e.jwt_data)
+
+        @app.errorhandler(WrongTokenError)
+        def handle_wrong_token_error(e):
+            return self._invalid_token_callback(str(e))
+
+    @staticmethod
+    def _set_default_configuration_options(app: Flask) -> None:
+        app.config.setdefault(
+            "JWT_ACCESS_TOKEN_EXPIRES", datetime.timedelta(minutes=15)
+        )
+        app.config.setdefault("JWT_ACCESS_COOKIE_NAME", "access_token_cookie")
+        app.config.setdefault("JWT_ACCESS_COOKIE_PATH", "/")
+        app.config.setdefault("JWT_ACCESS_CSRF_COOKIE_NAME", "csrf_access_token")
+        app.config.setdefault("JWT_ACCESS_CSRF_COOKIE_PATH", "/")
+        app.config.setdefault("JWT_ACCESS_CSRF_FIELD_NAME", "csrf_token")
+        app.config.setdefault("JWT_ACCESS_CSRF_HEADER_NAME", "X-CSRF-TOKEN")
+        app.config.setdefault("JWT_ALGORITHM", "HS256")
+        app.config.setdefault("JWT_COOKIE_CSRF_PROTECT", True)
+        app.config.setdefault("JWT_COOKIE_DOMAIN", None)
+        app.config.setdefault("JWT_COOKIE_SAMESITE", None)
+        app.config.setdefault("JWT_COOKIE_SECURE", False)
+        app.config.setdefault("JWT_CSRF_CHECK_FORM", False)
+        app.config.setdefault("JWT_CSRF_IN_COOKIES", True)
+        app.config.setdefault("JWT_CSRF_METHODS", ["POST", "PUT", "PATCH", "DELETE"])
+        app.config.setdefault("JWT_DECODE_ALGORITHMS", None)
+        app.config.setdefault("JWT_DECODE_AUDIENCE", None)
+        app.config.setdefault("JWT_DECODE_ISSUER", None)
+        app.config.setdefault("JWT_DECODE_LEEWAY", 0)
+        app.config.setdefault("JWT_ENCODE_AUDIENCE", None)
+        app.config.setdefault("JWT_ENCODE_ISSUER", None)
+        app.config.setdefault("JWT_ERROR_MESSAGE_KEY", "msg")
+        app.config.setdefault("JWT_HEADER_NAME", "Authorization")
+        app.config.setdefault("JWT_HEADER_TYPE", "Bearer")
+        app.config.setdefault("JWT_IDENTITY_CLAIM", "sub")
+        app.config.setdefault("JWT_JSON_KEY", "access_token")
+        app.config.setdefault("JWT_PRIVATE_KEY", None)
+        app.config.setdefault("JWT_PUBLIC_KEY", None)
+        app.config.setdefault("JWT_QUERY_STRING_NAME", "jwt")
+        app.config.setdefault("JWT_QUERY_STRING_VALUE_PREFIX", "")
+        app.config.setdefault("JWT_REFRESH_COOKIE_NAME", "refresh_token_cookie")
+        app.config.setdefault("JWT_REFRESH_COOKIE_PATH", "/")
+        app.config.setdefault("JWT_REFRESH_CSRF_COOKIE_NAME", "csrf_refresh_token")
+        app.config.setdefault("JWT_REFRESH_CSRF_COOKIE_PATH", "/")
+        app.config.setdefault("JWT_REFRESH_CSRF_FIELD_NAME", "csrf_token")
+        app.config.setdefault("JWT_REFRESH_CSRF_HEADER_NAME", "X-CSRF-TOKEN")
+        app.config.setdefault("JWT_REFRESH_JSON_KEY", "refresh_token")
+        app.config.setdefault("JWT_REFRESH_TOKEN_EXPIRES", datetime.timedelta(days=30))
+        app.config.setdefault("JWT_SECRET_KEY", None)
+        app.config.setdefault("JWT_SESSION_COOKIE", True)
+        app.config.setdefault("JWT_TOKEN_LOCATION", ("headers",))
+        app.config.setdefault("JWT_VERIFY_SUB", True)
+        app.config.setdefault("JWT_ENCODE_NBF", True)
+
+    def additional_claims_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to add additional claims
+        when creating a JWT. The claims returned by this function will be merged
+        with any claims passed in via the ``additional_claims`` argument to
+        :func:`~flask_jwt_extended.create_access_token` or
+        :func:`~flask_jwt_extended.create_refresh_token`.
+
+        The decorated function must take **one** argument.
+
+        The argument is the identity that was used when creating a JWT.
+
+        The decorated function must return a dictionary of claims to add to the JWT.
+        """
+        self._user_claims_callback = callback
+        return callback
+
+    def additional_headers_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to add additional headers
+        when creating a JWT. The headers returned by this function will be merged
+        with any headers passed in via the ``additional_headers`` argument to
+        :func:`~flask_jwt_extended.create_access_token` or
+        :func:`~flask_jwt_extended.create_refresh_token`.
+
+        The decorated function must take **one** argument.
+
+        The argument is the identity that was used when creating a JWT.
+
+        The decorated function must return a dictionary of headers to add to the JWT.
+        """
+        self._jwt_additional_header_callback = callback
+        return callback
+
+    def decode_key_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for dynamically setting the JWT
+        decode key based on the **UNVERIFIED** contents of the token. Think
+        carefully before using this functionality, in most cases you probably
+        don't need it.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the
+        unverified JWT.
+
+        The second argument is a dictionary containing the payload data of the
+        unverified JWT.
+
+        The decorated function must return a *string* that is used to decode and
+        verify the token.
+        """
+        self._decode_key_callback = callback
+        return callback
+
+    def encode_key_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for dynamically setting the JWT
+        encode key based on the tokens identity. Think carefully before using this
+        functionality, in most cases you probably don't need it.
+
+        The decorated function must take **one** argument.
+
+        The argument is the identity used to create this JWT.
+
+        The decorated function must return a *string* which is the secrete key used to
+        encode the JWT.
+        """
+        self._encode_key_callback = callback
+        return callback
+
+    def expired_token_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for returning a custom
+        response when an expired JWT is encountered.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return a Flask Response.
+        """
+        self._expired_token_callback = callback
+        return callback
+
+    def invalid_token_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for returning a custom
+        response when an invalid JWT is encountered.
+
+        This decorator sets the callback function that will be used if an
+        invalid JWT attempts to access a protected endpoint.
+
+        The decorated function must take **one** argument.
+
+        The argument is a string which contains the reason why a token is invalid.
+
+        The decorated function must return a Flask Response.
+        """
+        self._invalid_token_callback = callback
+        return callback
+
+    def needs_fresh_token_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for returning a custom
+        response when a valid and non-fresh token is used on an endpoint
+        that is marked as ``fresh=True``.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return a Flask Response.
+        """
+        self._needs_fresh_token_callback = callback
+        return callback
+
+    def revoked_token_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function for returning a custom
+        response when a revoked token is encountered.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return a Flask Response.
+        """
+        self._revoked_token_callback = callback
+        return callback
+
+    def token_in_blocklist_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to check if a JWT has
+        been revoked.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must be return ``True`` if the token has been
+        revoked, ``False`` otherwise.
+        """
+        self._token_in_blocklist_callback = callback
+        return callback
+
+    def token_verification_failed_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to return a custom
+        response when the claims verification check fails.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return a Flask Response.
+        """
+        self._token_verification_failed_callback = callback
+        return callback
+
+    def token_verification_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used for custom verification
+        of a valid JWT.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return ``True`` if the token is valid, or
+        ``False`` otherwise.
+        """
+        self._token_verification_callback = callback
+        return callback
+
+    def unauthorized_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to return a custom
+        response when no JWT is present.
+
+        The decorated function must take **one** argument.
+
+        The argument is a string that explains why the JWT could not be found.
+
+        The decorated function must return a Flask Response.
+        """
+        self._unauthorized_callback = callback
+        return callback
+
+    def user_identity_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to convert an identity to
+        a string when creating JWTs. This is useful for using objects (such as
+        SQLAlchemy instances) as the identity when creating your tokens.
+
+        The decorated function must take **one** argument.
+
+        The argument is the identity that was used when creating a JWT.
+
+        The decorated function must return a string.
+        """
+        self._user_identity_callback = callback
+        return callback
+
+    def user_lookup_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to convert a JWT into
+        a python object that can be used in a protected endpoint. This is useful
+        for automatically loading a SQLAlchemy instance based on the contents
+        of the JWT.
+
+        The object returned from this function can be accessed via
+        :attr:`~flask_jwt_extended.current_user` or
+        :meth:`~flask_jwt_extended.get_current_user`
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function can return any python object, which can then be
+        accessed in a protected endpoint. If an object cannot be loaded, for
+        example if a user has been deleted from your database, ``None`` must be
+        returned to indicate that an error occurred loading the user.
+        """
+        self._user_lookup_callback = callback
+        return callback
+
+    def user_lookup_error_loader(self, callback: Callable) -> Callable:
+        """
+        This decorator sets the callback function used to return a custom
+        response when loading a user via
+        :meth:`~flask_jwt_extended.JWTManager.user_lookup_loader` fails.
+
+        The decorated function must take **two** arguments.
+
+        The first argument is a dictionary containing the header data of the JWT.
+
+        The second argument is a dictionary containing the payload data of the JWT.
+
+        The decorated function must return a Flask Response.
+        """
+        self._user_lookup_error_callback = callback
+        return callback
+
+    def _encode_jwt_from_config(
+        self,
+        identity: Any,
+        token_type: str,
+        claims=None,
+        fresh: Fresh = False,
+        expires_delta: Optional[ExpiresDelta] = None,
+        headers=None,
+    ) -> str:
+        header_overrides = self._jwt_additional_header_callback(identity)
+        if headers is not None:
+            header_overrides.update(headers)
+
+        claim_overrides = self._user_claims_callback(identity)
+        if claims is not None:
+            claim_overrides.update(claims)
+
+        if expires_delta is None:
+            if token_type == "access":
+                expires_delta = config.access_expires
+            else:
+                expires_delta = config.refresh_expires
+
+        return _encode_jwt(
+            algorithm=config.algorithm,
+            audience=config.encode_audience,
+            claim_overrides=claim_overrides,
+            csrf=config.cookie_csrf_protect,
+            expires_delta=expires_delta,
+            fresh=fresh,
+            header_overrides=header_overrides,
+            identity=self._user_identity_callback(identity),
+            identity_claim_key=config.identity_claim_key,
+            issuer=config.encode_issuer,
+            json_encoder=config.json_encoder,
+            secret=self._encode_key_callback(identity),
+            token_type=token_type,
+            nbf=config.encode_nbf,
+        )
+
+    def _decode_jwt_from_config(
+        self, encoded_token: str, csrf_value=None, allow_expired: bool = False
+    ) -> dict:
+        unverified_claims = jwt.decode(
+            encoded_token,
+            algorithms=config.decode_algorithms,
+            options={"verify_signature": False},
+        )
+        unverified_headers = jwt.get_unverified_header(encoded_token)
+        secret = self._decode_key_callback(unverified_headers, unverified_claims)
+
+        kwargs = {
+            "algorithms": config.decode_algorithms,
+            "audience": config.decode_audience,
+            "csrf_value": csrf_value,
+            "encoded_token": encoded_token,
+            "identity_claim_key": config.identity_claim_key,
+            "issuer": config.decode_issuer,
+            "leeway": config.leeway,
+            "secret": secret,
+            "verify_aud": config.decode_audience is not None,
+            "verify_sub": config.verify_sub,
+        }
+
+        try:
+            return _decode_jwt(**kwargs, allow_expired=allow_expired)
+        except ExpiredSignatureError as e:
+            # TODO: If we ever do another breaking change, don't raise this pyjwt
+            #       error directly, instead raise a custom error of ours from this
+            #       error.
+            e.jwt_header = unverified_headers  # type: ignore
+            e.jwt_data = _decode_jwt(**kwargs, allow_expired=True)  # type: ignore
+            raise
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/py.typed b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/tokens.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/tokens.py
new file mode 100644
index 00000000..836101c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/tokens.py
@@ -0,0 +1,125 @@
+import uuid
+from datetime import datetime
+from datetime import timedelta
+from datetime import timezone
+from hmac import compare_digest
+from json import JSONEncoder
+from typing import Any
+from typing import Iterable
+from typing import List
+from typing import Type
+from typing import Union
+
+import jwt
+
+from flask_jwt_extended.exceptions import CSRFError
+from flask_jwt_extended.exceptions import JWTDecodeError
+from flask_jwt_extended.typing import ExpiresDelta
+from flask_jwt_extended.typing import Fresh
+
+
+def _encode_jwt(
+    algorithm: str,
+    audience: Union[str, Iterable[str]],
+    claim_overrides: dict,
+    csrf: bool,
+    expires_delta: ExpiresDelta,
+    fresh: Fresh,
+    header_overrides: dict,
+    identity: Any,
+    identity_claim_key: str,
+    issuer: str,
+    json_encoder: Type[JSONEncoder],
+    secret: str,
+    token_type: str,
+    nbf: bool,
+) -> str:
+    now = datetime.now(timezone.utc)
+
+    if isinstance(fresh, timedelta):
+        fresh = datetime.timestamp(now + fresh)
+
+    token_data = {
+        "fresh": fresh,
+        "iat": now,
+        "jti": str(uuid.uuid4()),
+        "type": token_type,
+        identity_claim_key: identity,
+    }
+
+    if nbf:
+        token_data["nbf"] = now
+
+    if csrf:
+        token_data["csrf"] = str(uuid.uuid4())
+
+    if audience:
+        token_data["aud"] = audience
+
+    if issuer:
+        token_data["iss"] = issuer
+
+    if expires_delta:
+        token_data["exp"] = now + expires_delta
+
+    if claim_overrides:
+        token_data.update(claim_overrides)
+
+    return jwt.encode(
+        token_data,
+        secret,
+        algorithm,
+        json_encoder=json_encoder,  # type: ignore
+        headers=header_overrides,
+    )
+
+
+def _decode_jwt(
+    algorithms: List,
+    allow_expired: bool,
+    audience: Union[str, Iterable[str]],
+    csrf_value: str,
+    encoded_token: str,
+    identity_claim_key: str,
+    issuer: str,
+    leeway: int,
+    secret: str,
+    verify_aud: bool,
+    verify_sub: bool,
+) -> dict:
+    options = {"verify_aud": verify_aud, "verify_sub": verify_sub}
+    if allow_expired:
+        options["verify_exp"] = False
+
+    # This call verifies the ext, iat, and nbf claims
+    # This optionally verifies the exp and aud claims if enabled
+    decoded_token = jwt.decode(
+        encoded_token,
+        secret,
+        algorithms=algorithms,
+        audience=audience,
+        issuer=issuer,
+        leeway=leeway,
+        options=options,
+    )
+
+    # Make sure that any custom claims we expect in the token are present
+    if identity_claim_key not in decoded_token:
+        raise JWTDecodeError("Missing claim: {}".format(identity_claim_key))
+
+    if "type" not in decoded_token:
+        decoded_token["type"] = "access"
+
+    if "fresh" not in decoded_token:
+        decoded_token["fresh"] = False
+
+    if "jti" not in decoded_token:
+        decoded_token["jti"] = None
+
+    if csrf_value:
+        if "csrf" not in decoded_token:
+            raise JWTDecodeError("Missing claim: csrf")
+        if not compare_digest(decoded_token["csrf"], csrf_value):
+            raise CSRFError("CSRF double submit tokens do not match")
+
+    return decoded_token
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/typing.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/typing.py
new file mode 100644
index 00000000..cc36d7bf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/typing.py
@@ -0,0 +1,6 @@
+from datetime import timedelta
+from typing import Literal
+from typing import Union
+
+ExpiresDelta = Union[Literal[False], timedelta]
+Fresh = Union[bool, float, timedelta]
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/utils.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/utils.py
new file mode 100644
index 00000000..d42f582a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/utils.py
@@ -0,0 +1,461 @@
+from typing import Any
+from typing import Optional
+
+import jwt
+from flask import g
+from flask import Response
+from werkzeug.local import LocalProxy
+
+from flask_jwt_extended.config import config
+from flask_jwt_extended.internal_utils import get_jwt_manager
+from flask_jwt_extended.typing import ExpiresDelta
+from flask_jwt_extended.typing import Fresh
+
+# Proxy to access the current user
+current_user: Any = LocalProxy(lambda: get_current_user())
+
+
+def get_jwt() -> dict:
+    """
+    In a protected endpoint, this will return the python dictionary which has
+    the payload of the JWT that is accessing the endpoint. If no JWT is present
+    due to ``jwt_required(optional=True)``, an empty dictionary is returned.
+
+    :return:
+        The payload (claims) of the JWT in the current request
+    """
+    decoded_jwt = g.get("_jwt_extended_jwt", None)
+    if decoded_jwt is None:
+        raise RuntimeError(
+            "You must call `@jwt_required()` or `verify_jwt_in_request()` "
+            "before using this method"
+        )
+    return decoded_jwt
+
+
+def get_jwt_header() -> dict:
+    """
+    In a protected endpoint, this will return the python dictionary which has
+    the header of the JWT that is accessing the endpoint. If no JWT is present
+    due to ``jwt_required(optional=True)``, an empty dictionary is returned.
+
+    :return:
+        The headers of the JWT in the current request
+    """
+    decoded_header = g.get("_jwt_extended_jwt_header", None)
+    if decoded_header is None:
+        raise RuntimeError(
+            "You must call `@jwt_required()` or `verify_jwt_in_request()` "
+            "before using this method"
+        )
+    return decoded_header
+
+
+def get_jwt_identity() -> Any:
+    """
+    In a protected endpoint, this will return the identity of the JWT that is
+    accessing the endpoint. If no JWT is present due to
+    ``jwt_required(optional=True)``, ``None`` is returned.
+
+    :return:
+        The identity of the JWT in the current request
+    """
+    return get_jwt().get(config.identity_claim_key, None)
+
+
+def get_jwt_request_location() -> Optional[str]:
+    """
+    In a protected endpoint, this will return the "location" at which the JWT
+    that is accessing the endpoint was found--e.g., "cookies", "query-string",
+    "headers", or "json". If no JWT is present due to ``jwt_required(optional=True)``,
+    None is returned.
+
+    :return:
+        The location of the JWT in the current request; e.g., "cookies",
+        "query-string", "headers", or "json"
+    """
+    return g.get("_jwt_extended_jwt_location", None)
+
+
+def get_current_user() -> Any:
+    """
+    In a protected endpoint, this will return the user object for the JWT that
+    is accessing the endpoint.
+
+    This is only usable if :meth:`~flask_jwt_extended.JWTManager.user_lookup_loader`
+    is configured. If the user loader callback is not being used, this will
+    raise an error.
+
+    If no JWT is present due to ``jwt_required(optional=True)``, ``None`` is returned.
+
+    :return:
+        The current user object for the JWT in the current request
+    """
+    get_jwt()  # Raise an error if not in a decorated context
+    jwt_user_dict = g.get("_jwt_extended_jwt_user", None)
+    if jwt_user_dict is None:
+        raise RuntimeError(
+            "You must provide a `@jwt.user_lookup_loader` callback to use "
+            "this method"
+        )
+    return jwt_user_dict["loaded_user"]
+
+
+def decode_token(
+    encoded_token: str, csrf_value: Optional[str] = None, allow_expired: bool = False
+) -> dict:
+    """
+    Returns the decoded token (python dict) from an encoded JWT. This does all
+    the checks to ensure that the decoded token is valid before returning it.
+
+    This will not fire the user loader callbacks, save the token for access
+    in protected endpoints, checked if a token is revoked, etc. This is puerly
+    used to ensure that a JWT is valid.
+
+    :param encoded_token:
+        The encoded JWT to decode.
+
+    :param csrf_value:
+        Expected CSRF double submit value (optional).
+
+    :param allow_expired:
+        If ``True``, do not raise an error if the JWT is expired.  Defaults to ``False``
+
+    :return:
+        Dictionary containing the payload of the JWT decoded JWT.
+    """
+    jwt_manager = get_jwt_manager()
+    return jwt_manager._decode_jwt_from_config(encoded_token, csrf_value, allow_expired)
+
+
+def create_access_token(
+    identity: Any,
+    fresh: Fresh = False,
+    expires_delta: Optional[ExpiresDelta] = None,
+    additional_claims=None,
+    additional_headers=None,
+):
+    """
+    Create a new access token.
+
+    :param identity:
+        The identity of this token. This must either be a string, or you must have
+        defined :meth:`~flask_jwt_extended.JWTManager.user_identity_loader` in order
+        to convert the object you passed in into a string.
+
+    :param fresh:
+        If this token should be marked as fresh, and can thus access endpoints
+        protected with ``@jwt_required(fresh=True)``. Defaults to ``False``.
+
+        This value can also be a ``datetime.timedelta``, which indicate
+        how long this token will be considered fresh.
+
+    :param expires_delta:
+        A ``datetime.timedelta`` for how long this token should last before it
+        expires. Set to False to disable expiration. If this is None, it will use
+        the ``JWT_ACCESS_TOKEN_EXPIRES`` config value (see :ref:`Configuration Options`)
+
+    :param additional_claims:
+        Optional. A hash of claims to include in the access token.  These claims are
+        merged into the default claims (exp, iat, etc) and claims returned from the
+        :meth:`~flask_jwt_extended.JWTManager.additional_claims_loader` callback.
+        On conflict, these claims take precedence.
+
+    :param headers:
+        Optional. A hash of headers to include in the access token. These headers
+        are merged into the default headers (alg, typ) and headers returned from
+        the :meth:`~flask_jwt_extended.JWTManager.additional_headers_loader`
+        callback. On conflict, these headers take precedence.
+
+    :return:
+        An encoded access token
+    """
+    jwt_manager = get_jwt_manager()
+    return jwt_manager._encode_jwt_from_config(
+        claims=additional_claims,
+        expires_delta=expires_delta,
+        fresh=fresh,
+        headers=additional_headers,
+        identity=identity,
+        token_type="access",
+    )
+
+
+def create_refresh_token(
+    identity: Any,
+    expires_delta: Optional[ExpiresDelta] = None,
+    additional_claims=None,
+    additional_headers=None,
+):
+    """
+    Create a new refresh token.
+
+    :param identity:
+        The identity of this token. This must either be a string, or you must have
+        defined :meth:`~flask_jwt_extended.JWTManager.user_identity_loader` in order
+        to convert the object you passed in into a string.
+
+    :param expires_delta:
+        A ``datetime.timedelta`` for how long this token should last before it expires.
+        Set to False to disable expiration. If this is None, it will use the
+        ``JWT_REFRESH_TOKEN_EXPIRES`` config value (see :ref:`Configuration Options`)
+
+    :param additional_claims:
+        Optional. A hash of claims to include in the refresh token. These claims are
+        merged into the default claims (exp, iat, etc) and claims returned from the
+        :meth:`~flask_jwt_extended.JWTManager.additional_claims_loader` callback.
+        On conflict, these claims take precedence.
+
+    :param headers:
+        Optional. A hash of headers to include in the refresh token. These headers
+        are merged into the default headers (alg, typ) and headers returned from the
+        :meth:`~flask_jwt_extended.JWTManager.additional_headers_loader` callback.
+        On conflict, these headers take precedence.
+
+    :return:
+        An encoded refresh token
+    """
+    jwt_manager = get_jwt_manager()
+    return jwt_manager._encode_jwt_from_config(
+        claims=additional_claims,
+        expires_delta=expires_delta,
+        fresh=False,
+        headers=additional_headers,
+        identity=identity,
+        token_type="refresh",
+    )
+
+
+def get_unverified_jwt_headers(encoded_token: str) -> dict:
+    """
+    Returns the Headers of an encoded JWT without verifying the signature of the JWT.
+
+    :param encoded_token:
+        The encoded JWT to get the Header from.
+
+    :return:
+        JWT header parameters as python dict()
+    """
+    return jwt.get_unverified_header(encoded_token)
+
+
+def get_jti(encoded_token: str) -> Optional[str]:
+    """
+    Returns the JTI (unique identifier) of an encoded JWT
+
+    :param encoded_token:
+        The encoded JWT to get the JTI from.
+
+    :return:
+        The JTI (unique identifier) of a JWT, if it is present.
+    """
+    return decode_token(encoded_token).get("jti")
+
+
+def get_csrf_token(encoded_token: str) -> str:
+    """
+    Returns the CSRF double submit token from an encoded JWT.
+
+    :param encoded_token:
+        The encoded JWT
+
+    :return:
+        The CSRF double submit token (string)
+    """
+    token = decode_token(encoded_token)
+    return token["csrf"]
+
+
+def set_access_cookies(
+    response: Response, encoded_access_token: str, max_age=None, domain=None
+) -> None:
+    """
+    Modifiy a Flask Response to set a cookie containing the access JWT.
+    Also sets the corresponding CSRF cookies if ``JWT_CSRF_IN_COOKIES`` is ``True``
+    (see :ref:`Configuration Options`)
+
+    :param response:
+        A Flask Response object.
+
+    :param encoded_access_token:
+        The encoded access token to set in the cookies.
+
+    :param max_age:
+        The max age of the cookie. If this is None, it will use the
+        ``JWT_SESSION_COOKIE`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``max-age`` and the JWT_SESSION_COOKIE option
+        will be ignored. Values should be the number of seconds (as an integer).
+
+    :param domain:
+        The domain of the cookie. If this is None, it will use the
+        ``JWT_COOKIE_DOMAIN`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``domain`` and the JWT_COOKIE_DOMAIN option
+        will be ignored.
+    """
+    response.set_cookie(
+        config.access_cookie_name,
+        value=encoded_access_token,
+        max_age=max_age or config.cookie_max_age,
+        secure=config.cookie_secure,
+        httponly=True,
+        domain=domain or config.cookie_domain,
+        path=config.access_cookie_path,
+        samesite=config.cookie_samesite,
+    )
+
+    if config.cookie_csrf_protect and config.csrf_in_cookies:
+        response.set_cookie(
+            config.access_csrf_cookie_name,
+            value=get_csrf_token(encoded_access_token),
+            max_age=max_age or config.cookie_max_age,
+            secure=config.cookie_secure,
+            httponly=False,
+            domain=domain or config.cookie_domain,
+            path=config.access_csrf_cookie_path,
+            samesite=config.cookie_samesite,
+        )
+
+
+def set_refresh_cookies(
+    response: Response,
+    encoded_refresh_token: str,
+    max_age: Optional[int] = None,
+    domain: Optional[str] = None,
+) -> None:
+    """
+    Modifiy a Flask Response to set a cookie containing the refresh JWT.
+    Also sets the corresponding CSRF cookies if ``JWT_CSRF_IN_COOKIES`` is ``True``
+    (see :ref:`Configuration Options`)
+
+    :param response:
+        A Flask Response object.
+
+    :param encoded_refresh_token:
+        The encoded refresh token to set in the cookies.
+
+    :param max_age:
+        The max age of the cookie. If this is None, it will use the
+        ``JWT_SESSION_COOKIE`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``max-age`` and the JWT_SESSION_COOKIE option
+        will be ignored. Values should be the number of seconds (as an integer).
+
+    :param domain:
+        The domain of the cookie. If this is None, it will use the
+        ``JWT_COOKIE_DOMAIN`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``domain`` and the JWT_COOKIE_DOMAIN option
+        will be ignored.
+    """
+    response.set_cookie(
+        config.refresh_cookie_name,
+        value=encoded_refresh_token,
+        max_age=max_age or config.cookie_max_age,
+        secure=config.cookie_secure,
+        httponly=True,
+        domain=domain or config.cookie_domain,
+        path=config.refresh_cookie_path,
+        samesite=config.cookie_samesite,
+    )
+
+    if config.cookie_csrf_protect and config.csrf_in_cookies:
+        response.set_cookie(
+            config.refresh_csrf_cookie_name,
+            value=get_csrf_token(encoded_refresh_token),
+            max_age=max_age or config.cookie_max_age,
+            secure=config.cookie_secure,
+            httponly=False,
+            domain=domain or config.cookie_domain,
+            path=config.refresh_csrf_cookie_path,
+            samesite=config.cookie_samesite,
+        )
+
+
+def unset_jwt_cookies(response: Response, domain: Optional[str] = None) -> None:
+    """
+    Modifiy a Flask Response to delete the cookies containing access or refresh
+    JWTs.  Also deletes the corresponding CSRF cookies if applicable.
+
+    :param response:
+        A Flask Response object
+    """
+    unset_access_cookies(response, domain)
+    unset_refresh_cookies(response, domain)
+
+
+def unset_access_cookies(response: Response, domain: Optional[str] = None) -> None:
+    """
+    Modifiy a Flask Response to delete the cookie containing an access JWT.
+    Also deletes the corresponding CSRF cookie if applicable.
+
+    :param response:
+        A Flask Response object
+
+    :param domain:
+        The domain of the cookie. If this is None, it will use the
+        ``JWT_COOKIE_DOMAIN`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``domain`` and the JWT_COOKIE_DOMAIN option
+        will be ignored.
+    """
+    response.set_cookie(
+        config.access_cookie_name,
+        value="",
+        expires=0,
+        secure=config.cookie_secure,
+        httponly=True,
+        domain=domain or config.cookie_domain,
+        path=config.access_cookie_path,
+        samesite=config.cookie_samesite,
+    )
+
+    if config.cookie_csrf_protect and config.csrf_in_cookies:
+        response.set_cookie(
+            config.access_csrf_cookie_name,
+            value="",
+            expires=0,
+            secure=config.cookie_secure,
+            httponly=False,
+            domain=domain or config.cookie_domain,
+            path=config.access_csrf_cookie_path,
+            samesite=config.cookie_samesite,
+        )
+
+
+def unset_refresh_cookies(response: Response, domain: Optional[str] = None) -> None:
+    """
+    Modifiy a Flask Response to delete the cookie containing a refresh JWT.
+    Also deletes the corresponding CSRF cookie if applicable.
+
+    :param response:
+        A Flask Response object
+
+    :param domain:
+        The domain of the cookie. If this is None, it will use the
+        ``JWT_COOKIE_DOMAIN`` option (see :ref:`Configuration Options`). Otherwise,
+        it will use this as the cookies ``domain`` and the JWT_COOKIE_DOMAIN option
+        will be ignored.
+    """
+    response.set_cookie(
+        config.refresh_cookie_name,
+        value="",
+        expires=0,
+        secure=config.cookie_secure,
+        httponly=True,
+        domain=domain or config.cookie_domain,
+        path=config.refresh_cookie_path,
+        samesite=config.cookie_samesite,
+    )
+
+    if config.cookie_csrf_protect and config.csrf_in_cookies:
+        response.set_cookie(
+            config.refresh_csrf_cookie_name,
+            value="",
+            expires=0,
+            secure=config.cookie_secure,
+            httponly=False,
+            domain=domain or config.cookie_domain,
+            path=config.refresh_csrf_cookie_path,
+            samesite=config.cookie_samesite,
+        )
+
+
+def current_user_context_processor() -> Any:
+    return {"current_user": get_current_user()}
diff --git a/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/view_decorators.py b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/view_decorators.py
new file mode 100644
index 00000000..407bea4d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_jwt_extended/view_decorators.py
@@ -0,0 +1,372 @@
+from datetime import datetime
+from datetime import timezone
+from functools import wraps
+from re import split
+from typing import Any
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Union
+
+from flask import current_app
+from flask import g
+from flask import request
+from werkzeug.exceptions import BadRequest
+
+from flask_jwt_extended.config import config
+from flask_jwt_extended.exceptions import CSRFError
+from flask_jwt_extended.exceptions import FreshTokenRequired
+from flask_jwt_extended.exceptions import InvalidHeaderError
+from flask_jwt_extended.exceptions import InvalidQueryParamError
+from flask_jwt_extended.exceptions import NoAuthorizationError
+from flask_jwt_extended.exceptions import UserLookupError
+from flask_jwt_extended.internal_utils import custom_verification_for_token
+from flask_jwt_extended.internal_utils import has_user_lookup
+from flask_jwt_extended.internal_utils import user_lookup
+from flask_jwt_extended.internal_utils import verify_token_not_blocklisted
+from flask_jwt_extended.internal_utils import verify_token_type
+from flask_jwt_extended.utils import decode_token
+from flask_jwt_extended.utils import get_unverified_jwt_headers
+
+LocationType = Union[str, Sequence, None]
+
+
+def _verify_token_is_fresh(jwt_header: dict, jwt_data: dict) -> None:
+    fresh = jwt_data["fresh"]
+    if isinstance(fresh, bool):
+        if not fresh:
+            raise FreshTokenRequired("Fresh token required", jwt_header, jwt_data)
+    else:
+        now = datetime.timestamp(datetime.now(timezone.utc))
+        if fresh < now:
+            raise FreshTokenRequired("Fresh token required", jwt_header, jwt_data)
+
+
+def verify_jwt_in_request(
+    optional: bool = False,
+    fresh: bool = False,
+    refresh: bool = False,
+    locations: Optional[LocationType] = None,
+    verify_type: bool = True,
+    skip_revocation_check: bool = False,
+) -> Optional[Tuple[dict, dict]]:
+    """
+    Verify that a valid JWT is present in the request, unless ``optional=True`` in
+    which case no JWT is also considered valid.
+
+    :param optional:
+        If ``True``, do not raise an error if no JWT is present in the request.
+        Defaults to ``False``.
+
+    :param fresh:
+        If ``True``, require a JWT marked as ``fresh`` in order to be verified.
+        Defaults to ``False``.
+
+    :param refresh:
+        If ``True``, requires a refresh JWT to access this endpoint. If ``False``,
+        requires an access JWT to access this endpoint. Defaults to ``False``
+
+    :param locations:
+        A location or list of locations to look for the JWT in this request, for
+        example ``'headers'`` or ``['headers', 'cookies']``. Defaults to ``None``
+        which indicates that JWTs will be looked for in the locations defined by the
+        ``JWT_TOKEN_LOCATION`` configuration option.
+
+    :param verify_type:
+        If ``True``, the token type (access or refresh) will be checked according
+        to the ``refresh`` argument. If ``False``, type will not be checked and both
+        access and refresh tokens will be accepted.
+
+    :param skip_revocation_check:
+        If ``True``, revocation status of the token will be *not* checked. If ``False``,
+        revocation status of the token will be checked.
+
+    :return:
+        A tuple containing the jwt_header and the jwt_data if a valid JWT is
+        present in the request. If ``optional=True`` and no JWT is in the request,
+        ``None`` will be returned instead. Raise an exception if an invalid JWT
+        is in the request.
+    """
+    if request.method in config.exempt_methods:
+        return None
+
+    try:
+        jwt_data, jwt_header, jwt_location = _decode_jwt_from_request(
+            locations,
+            fresh,
+            refresh=refresh,
+            verify_type=verify_type,
+            skip_revocation_check=skip_revocation_check,
+        )
+
+    except NoAuthorizationError:
+        if not optional:
+            raise
+        g._jwt_extended_jwt = {}
+        g._jwt_extended_jwt_header = {}
+        g._jwt_extended_jwt_user = {"loaded_user": None}
+        g._jwt_extended_jwt_location = None
+        return None
+
+    # Save these at the very end so that they are only saved in the requet
+    # context if the token is valid and all callbacks succeed
+    g._jwt_extended_jwt_user = _load_user(jwt_header, jwt_data)
+    g._jwt_extended_jwt_header = jwt_header
+    g._jwt_extended_jwt = jwt_data
+    g._jwt_extended_jwt_location = jwt_location
+
+    return jwt_header, jwt_data
+
+
+def jwt_required(
+    optional: bool = False,
+    fresh: bool = False,
+    refresh: bool = False,
+    locations: Optional[LocationType] = None,
+    verify_type: bool = True,
+    skip_revocation_check: bool = False,
+) -> Any:
+    """
+    A decorator to protect a Flask endpoint with JSON Web Tokens.
+
+    Any route decorated with this will require a valid JWT to be present in the
+    request (unless optional=True, in which case no JWT is also valid) before the
+    endpoint can be called.
+
+    :param optional:
+        If ``True``, allow the decorated endpoint to be accessed if no JWT is present in
+        the request. Defaults to ``False``.
+
+    :param fresh:
+        If ``True``, require a JWT marked with ``fresh`` to be able to access this
+        endpoint. Defaults to ``False``.
+
+    :param refresh:
+        If ``True``, requires a refresh JWT to access this endpoint. If ``False``,
+        requires an access JWT to access this endpoint. Defaults to ``False``.
+
+    :param locations:
+        A location or list of locations to look for the JWT in this request, for
+        example ``'headers'`` or ``['headers', 'cookies']``. Defaults to ``None``
+        which indicates that JWTs will be looked for in the locations defined by the
+        ``JWT_TOKEN_LOCATION`` configuration option.
+
+    :param verify_type:
+        If ``True``, the token type (access or refresh) will be checked according
+        to the ``refresh`` argument. If ``False``, type will not be checked and both
+        access and refresh tokens will be accepted.
+
+    :param skip_revocation_check:
+        If ``True``, revocation status of the token will be *not* checked. If ``False``,
+        revocation status of the token will be checked.
+    """
+
+    def wrapper(fn):
+        @wraps(fn)
+        def decorator(*args, **kwargs):
+            verify_jwt_in_request(
+                optional, fresh, refresh, locations, verify_type, skip_revocation_check
+            )
+            return current_app.ensure_sync(fn)(*args, **kwargs)
+
+        return decorator
+
+    return wrapper
+
+
+def _load_user(jwt_header: dict, jwt_data: dict) -> Optional[dict]:
+    if not has_user_lookup():
+        return None
+
+    identity = jwt_data[config.identity_claim_key]
+    user = user_lookup(jwt_header, jwt_data)
+    if user is None:
+        error_msg = "user_lookup returned None for {}".format(identity)
+        raise UserLookupError(error_msg, jwt_header, jwt_data)
+    return {"loaded_user": user}
+
+
+def _decode_jwt_from_headers() -> Tuple[str, None]:
+    header_name = config.header_name
+    header_type = config.header_type
+
+    # Verify we have the auth header
+    auth_header = request.headers.get(header_name, "").strip().strip(",")
+    if not auth_header:
+        raise NoAuthorizationError(f"Missing {header_name} Header")
+
+    # Make sure the header is in a valid format that we are expecting, ie
+    # <HeaderName>: <HeaderType(optional)> <JWT>.
+    #
+    # Also handle the fact that the header that can be comma delimited, ie
+    # <HeaderName>: <field> <value>, <field> <value>, etc...
+    if header_type:
+        field_values = split(r",\s*", auth_header)
+        jwt_headers = [s for s in field_values if s.split()[0] == header_type]
+        if len(jwt_headers) != 1:
+            msg = (
+                f"Missing '{header_type}' type in '{header_name}' header. "
+                f"Expected '{header_name}: {header_type} <JWT>'"
+            )
+            raise NoAuthorizationError(msg)
+
+        parts = jwt_headers[0].split()
+        if len(parts) != 2:
+            msg = (
+                f"Bad {header_name} header. "
+                f"Expected '{header_name}: {header_type} <JWT>'"
+            )
+            raise InvalidHeaderError(msg)
+
+        encoded_token = parts[1]
+    else:
+        parts = auth_header.split()
+        if len(parts) != 1:
+            msg = f"Bad {header_name} header. Expected '{header_name}: <JWT>'"
+            raise InvalidHeaderError(msg)
+
+        encoded_token = parts[0]
+
+    return encoded_token, None
+
+
+def _decode_jwt_from_cookies(refresh: bool) -> Tuple[str, Optional[str]]:
+    if refresh:
+        cookie_key = config.refresh_cookie_name
+        csrf_header_key = config.refresh_csrf_header_name
+        csrf_field_key = config.refresh_csrf_field_name
+    else:
+        cookie_key = config.access_cookie_name
+        csrf_header_key = config.access_csrf_header_name
+        csrf_field_key = config.access_csrf_field_name
+
+    encoded_token = request.cookies.get(cookie_key)
+    if not encoded_token:
+        raise NoAuthorizationError('Missing cookie "{}"'.format(cookie_key))
+
+    if config.cookie_csrf_protect and request.method in config.csrf_request_methods:
+        csrf_value = request.headers.get(csrf_header_key, None)
+        if not csrf_value and config.csrf_check_form:
+            csrf_value = request.form.get(csrf_field_key, None)
+        if not csrf_value:
+            raise CSRFError("Missing CSRF token")
+    else:
+        csrf_value = None
+
+    return encoded_token, csrf_value
+
+
+def _decode_jwt_from_query_string() -> Tuple[str, None]:
+    param_name = config.query_string_name
+    prefix = config.query_string_value_prefix
+
+    value = request.args.get(param_name)
+    if not value:
+        raise NoAuthorizationError(f"Missing '{param_name}' query paramater")
+
+    if not value.startswith(prefix):
+        raise InvalidQueryParamError(
+            f"Invalid value for query parameter '{param_name}'. "
+            f"Expected the value to start with '{prefix}'"
+        )
+
+    encoded_token = value[len(prefix) :]  # noqa: E203
+    return encoded_token, None
+
+
+def _decode_jwt_from_json(refresh: bool) -> Tuple[str, None]:
+    if not request.is_json:
+        raise NoAuthorizationError("Invalid content-type. Must be application/json.")
+
+    if refresh:
+        token_key = config.refresh_json_key
+    else:
+        token_key = config.json_key
+
+    try:
+        encoded_token = request.json and request.json.get(token_key, None)
+        if not encoded_token:
+            raise BadRequest()
+    except BadRequest:
+        raise NoAuthorizationError(
+            'Missing "{}" key in json data.'.format(token_key)
+        ) from None
+
+    return encoded_token, None
+
+
+def _decode_jwt_from_request(
+    locations: LocationType,
+    fresh: bool,
+    refresh: bool = False,
+    verify_type: bool = True,
+    skip_revocation_check: bool = False,
+) -> Tuple[dict, dict, str]:
+    # Figure out what locations to look for the JWT in this request
+    if isinstance(locations, str):
+        locations = [locations]
+
+    if not locations:
+        locations = config.token_location
+
+    # Get the decode functions in the order specified by locations.
+    # Each entry in this list is a tuple (<location>, <encoded-token-function>)
+    get_encoded_token_functions = []
+    for location in locations:
+        if location == "cookies":
+            get_encoded_token_functions.append(
+                (location, lambda: _decode_jwt_from_cookies(refresh))
+            )
+        elif location == "query_string":
+            get_encoded_token_functions.append(
+                (location, _decode_jwt_from_query_string)
+            )
+        elif location == "headers":
+            get_encoded_token_functions.append((location, _decode_jwt_from_headers))
+        elif location == "json":
+            get_encoded_token_functions.append(
+                (location, lambda: _decode_jwt_from_json(refresh))
+            )
+        else:
+            raise RuntimeError(f"'{location}' is not a valid location")
+
+    # Try to find the token from one of these locations. It only needs to exist
+    # in one place to be valid (not every location).
+    errors = []
+    decoded_token = None
+    for location, get_encoded_token_function in get_encoded_token_functions:
+        try:
+            encoded_token, csrf_token = get_encoded_token_function()
+            decoded_token = decode_token(encoded_token, csrf_token)
+            jwt_location = location
+            jwt_header = get_unverified_jwt_headers(encoded_token)
+            break
+        except NoAuthorizationError as e:
+            errors.append(str(e))
+
+    # Do some work to make a helpful and human readable error message if no
+    # token was found in any of the expected locations.
+    if not decoded_token:
+        if len(locations) > 1:
+            err_msg = "Missing JWT in {start_locs} or {end_locs} ({details})".format(
+                start_locs=", ".join(locations[:-1]),
+                end_locs=locations[-1],
+                details="; ".join(errors),
+            )
+            raise NoAuthorizationError(err_msg)
+        else:
+            raise NoAuthorizationError(errors[0])
+
+    # Additional verifications provided by this extension
+    if verify_type:
+        verify_token_type(decoded_token, refresh)
+
+    if fresh:
+        _verify_token_is_fresh(jwt_header, decoded_token)
+
+    if not skip_revocation_check:
+        verify_token_not_blocklisted(jwt_header, decoded_token)
+
+    custom_verification_for_token(jwt_header, decoded_token)
+
+    return decoded_token, jwt_header, jwt_location
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/LICENSE.rst b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/LICENSE.rst
new file mode 100644
index 00000000..9d227a0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/LICENSE.rst
@@ -0,0 +1,28 @@
+Copyright 2010 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/METADATA
new file mode 100644
index 00000000..92f239cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/METADATA
@@ -0,0 +1,109 @@
+Metadata-Version: 2.1
+Name: Flask-SQLAlchemy
+Version: 3.1.1
+Summary: Add SQLAlchemy support to your Flask application.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Requires-Dist: flask>=2.2.5
+Requires-Dist: sqlalchemy>=2.0.16
+Project-URL: Changes, https://flask-sqlalchemy.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://flask-sqlalchemy.palletsprojects.com
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Issue Tracker, https://github.com/pallets-eco/flask-sqlalchemy/issues/
+Project-URL: Source Code, https://github.com/pallets-eco/flask-sqlalchemy/
+
+Flask-SQLAlchemy
+================
+
+Flask-SQLAlchemy is an extension for `Flask`_ that adds support for
+`SQLAlchemy`_ to your application. It aims to simplify using SQLAlchemy
+with Flask by providing useful defaults and extra helpers that make it
+easier to accomplish common tasks.
+
+.. _Flask: https://palletsprojects.com/p/flask/
+.. _SQLAlchemy: https://www.sqlalchemy.org
+
+
+Installing
+----------
+
+Install and update using `pip`_:
+
+.. code-block:: text
+
+  $ pip install -U Flask-SQLAlchemy
+
+.. _pip: https://pip.pypa.io/en/stable/getting-started/
+
+
+A Simple Example
+----------------
+
+.. code-block:: python
+
+    from flask import Flask
+    from flask_sqlalchemy import SQLAlchemy
+    from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column
+
+    app = Flask(__name__)
+    app.config["SQLALCHEMY_DATABASE_URI"] = "sqlite:///example.sqlite"
+
+    class Base(DeclarativeBase):
+      pass
+
+    db = SQLAlchemy(app, model_class=Base)
+
+    class User(db.Model):
+        id: Mapped[int] = mapped_column(db.Integer, primary_key=True)
+        username: Mapped[str] = mapped_column(db.String, unique=True, nullable=False)
+
+    with app.app_context():
+        db.create_all()
+
+        db.session.add(User(username="example"))
+        db.session.commit()
+
+        users = db.session.execute(db.select(User)).scalars()
+
+
+Contributing
+------------
+
+For guidance on setting up a development environment and how to make a
+contribution to Flask-SQLAlchemy, see the `contributing guidelines`_.
+
+.. _contributing guidelines: https://github.com/pallets-eco/flask-sqlalchemy/blob/main/CONTRIBUTING.rst
+
+
+Donate
+------
+
+The Pallets organization develops and supports Flask-SQLAlchemy and
+other popular packages. In order to grow the community of contributors
+and users, and allow the maintainers to devote more time to the
+projects, `please donate today`_.
+
+.. _please donate today: https://palletsprojects.com/donate
+
+
+Links
+-----
+
+-   Documentation: https://flask-sqlalchemy.palletsprojects.com/
+-   Changes: https://flask-sqlalchemy.palletsprojects.com/changes/
+-   PyPI Releases: https://pypi.org/project/Flask-SQLAlchemy/
+-   Source Code: https://github.com/pallets-eco/flask-sqlalchemy/
+-   Issue Tracker: https://github.com/pallets-eco/flask-sqlalchemy/issues/
+-   Website: https://palletsprojects.com/
+-   Twitter: https://twitter.com/PalletsTeam
+-   Chat: https://discord.gg/pallets
+
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/RECORD
new file mode 100644
index 00000000..46e3d055
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/RECORD
@@ -0,0 +1,27 @@
+flask_sqlalchemy-3.1.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+flask_sqlalchemy-3.1.1.dist-info/LICENSE.rst,sha256=SJqOEQhQntmKN7uYPhHg9-HTHwvY-Zp5yESOf_N9B-o,1475
+flask_sqlalchemy-3.1.1.dist-info/METADATA,sha256=lBxR1akBt7n9XBjIVTL2OV52OhCfFrb-Mqtoe0DCbR8,3432
+flask_sqlalchemy-3.1.1.dist-info/RECORD,,
+flask_sqlalchemy-3.1.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flask_sqlalchemy-3.1.1.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+flask_sqlalchemy/__init__.py,sha256=he_w4qQQVS2Z1ms5GCTptDTXNOXBXw0n8zSuWCp8n6Y,653
+flask_sqlalchemy/__pycache__/__init__.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/cli.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/extension.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/model.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/pagination.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/query.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/record_queries.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/session.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/table.cpython-312.pyc,,
+flask_sqlalchemy/__pycache__/track_modifications.cpython-312.pyc,,
+flask_sqlalchemy/cli.py,sha256=pg3QDxP36GW2qnwe_CpPtkRhPchyVSGM6zlBNWuNCFE,484
+flask_sqlalchemy/extension.py,sha256=71tP_kNtb5VgZdafy_OH1sWdZOA6PaT7cJqX7tKgZ-k,38261
+flask_sqlalchemy/model.py,sha256=_mSisC2Eni0TgTyFWeN_O4LIexTeP_sVTdxh03yMK50,11461
+flask_sqlalchemy/pagination.py,sha256=JFpllrqkRkwacb8DAmQWaz9wsvQa0dypfSkhUDSC2ws,11119
+flask_sqlalchemy/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+flask_sqlalchemy/query.py,sha256=Uls9qbmnpb9Vba43EDfsRP17eHJ0X4VG7SE22tH5R3g,3748
+flask_sqlalchemy/record_queries.py,sha256=ouS1ayj16h76LJprx13iYdoFZbm6m8OncrOgAVbG1Sk,3520
+flask_sqlalchemy/session.py,sha256=pBbtN8iDc8yuGVt0k18BvZHh2uEI7QPzZXO7eXrRi1g,3426
+flask_sqlalchemy/table.py,sha256=wAPOy8qwyAxpMwOIUJY4iMOultzz2W0D6xvBkQ7U2CE,859
+flask_sqlalchemy/track_modifications.py,sha256=yieyozj7IiVzwnAGZ-ZrgqrzjrUfG0kPrXBfW_hStSU,2755
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy-3.1.1.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__init__.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__init__.py
new file mode 100644
index 00000000..c2fa0591
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__init__.py
@@ -0,0 +1,26 @@
+from __future__ import annotations
+
+import typing as t
+
+from .extension import SQLAlchemy
+
+__all__ = [
+    "SQLAlchemy",
+]
+
+
+def __getattr__(name: str) -> t.Any:
+    if name == "__version__":
+        import importlib.metadata
+        import warnings
+
+        warnings.warn(
+            "The '__version__' attribute is deprecated and will be removed in"
+            " Flask-SQLAlchemy 3.2. Use feature detection or"
+            " 'importlib.metadata.version(\"flask-sqlalchemy\")' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return importlib.metadata.version("flask-sqlalchemy")
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..bfe820c5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/cli.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/cli.cpython-312.pyc
new file mode 100644
index 00000000..73ddbb0a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/cli.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/extension.cpython-312.pyc
new file mode 100644
index 00000000..f820807f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/model.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/model.cpython-312.pyc
new file mode 100644
index 00000000..6b39d020
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/model.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/pagination.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/pagination.cpython-312.pyc
new file mode 100644
index 00000000..c1b413f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/pagination.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/query.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/query.cpython-312.pyc
new file mode 100644
index 00000000..51f5a516
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/query.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/record_queries.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/record_queries.cpython-312.pyc
new file mode 100644
index 00000000..6fe23d8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/record_queries.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/session.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/session.cpython-312.pyc
new file mode 100644
index 00000000..536000e3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/session.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/table.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/table.cpython-312.pyc
new file mode 100644
index 00000000..416e864f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/table.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/track_modifications.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/track_modifications.cpython-312.pyc
new file mode 100644
index 00000000..53c07ddb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/__pycache__/track_modifications.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/cli.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/cli.py
new file mode 100644
index 00000000..d7d7e4be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/cli.py
@@ -0,0 +1,16 @@
+from __future__ import annotations
+
+import typing as t
+
+from flask import current_app
+
+
+def add_models_to_shell() -> dict[str, t.Any]:
+    """Registered with :meth:`~flask.Flask.shell_context_processor` if
+    ``add_models_to_shell`` is enabled. Adds the ``db`` instance and all model classes
+    to ``flask shell``.
+    """
+    db = current_app.extensions["sqlalchemy"]
+    out = {m.class_.__name__: m.class_ for m in db.Model._sa_registry.mappers}
+    out["db"] = db
+    return out
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/extension.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/extension.py
new file mode 100644
index 00000000..43e1b9a4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/extension.py
@@ -0,0 +1,1008 @@
+from __future__ import annotations
+
+import os
+import types
+import typing as t
+import warnings
+from weakref import WeakKeyDictionary
+
+import sqlalchemy as sa
+import sqlalchemy.event as sa_event
+import sqlalchemy.exc as sa_exc
+import sqlalchemy.orm as sa_orm
+from flask import abort
+from flask import current_app
+from flask import Flask
+from flask import has_app_context
+
+from .model import _QueryProperty
+from .model import BindMixin
+from .model import DefaultMeta
+from .model import DefaultMetaNoName
+from .model import Model
+from .model import NameMixin
+from .pagination import Pagination
+from .pagination import SelectPagination
+from .query import Query
+from .session import _app_ctx_id
+from .session import Session
+from .table import _Table
+
+_O = t.TypeVar("_O", bound=object)  # Based on sqlalchemy.orm._typing.py
+
+
+# Type accepted for model_class argument
+_FSA_MCT = t.TypeVar(
+    "_FSA_MCT",
+    bound=t.Union[
+        t.Type[Model],
+        sa_orm.DeclarativeMeta,
+        t.Type[sa_orm.DeclarativeBase],
+        t.Type[sa_orm.DeclarativeBaseNoMeta],
+    ],
+)
+
+
+# Type returned by make_declarative_base
+class _FSAModel(Model):
+    metadata: sa.MetaData
+
+
+def _get_2x_declarative_bases(
+    model_class: _FSA_MCT,
+) -> list[t.Type[t.Union[sa_orm.DeclarativeBase, sa_orm.DeclarativeBaseNoMeta]]]:
+    return [
+        b
+        for b in model_class.__bases__
+        if issubclass(b, (sa_orm.DeclarativeBase, sa_orm.DeclarativeBaseNoMeta))
+    ]
+
+
+class SQLAlchemy:
+    """Integrates SQLAlchemy with Flask. This handles setting up one or more engines,
+    associating tables and models with specific engines, and cleaning up connections and
+    sessions after each request.
+
+    Only the engine configuration is specific to each application, other things like
+    the model, table, metadata, and session are shared for all applications using that
+    extension instance. Call :meth:`init_app` to configure the extension on an
+    application.
+
+    After creating the extension, create model classes by subclassing :attr:`Model`, and
+    table classes with :attr:`Table`. These can be accessed before :meth:`init_app` is
+    called, making it possible to define the models separately from the application.
+
+    Accessing :attr:`session` and :attr:`engine` requires an active Flask application
+    context. This includes methods like :meth:`create_all` which use the engine.
+
+    This class also provides access to names in SQLAlchemy's ``sqlalchemy`` and
+    ``sqlalchemy.orm`` modules. For example, you can use ``db.Column`` and
+    ``db.relationship`` instead of importing ``sqlalchemy.Column`` and
+    ``sqlalchemy.orm.relationship``. This can be convenient when defining models.
+
+    :param app: Call :meth:`init_app` on this Flask application now.
+    :param metadata: Use this as the default :class:`sqlalchemy.schema.MetaData`. Useful
+        for setting a naming convention.
+    :param session_options: Arguments used by :attr:`session` to create each session
+        instance. A ``scopefunc`` key will be passed to the scoped session, not the
+        session instance. See :class:`sqlalchemy.orm.sessionmaker` for a list of
+        arguments.
+    :param query_class: Use this as the default query class for models and dynamic
+        relationships. The query interface is considered legacy in SQLAlchemy.
+    :param model_class: Use this as the model base class when creating the declarative
+        model class :attr:`Model`. Can also be a fully created declarative model class
+        for further customization.
+    :param engine_options: Default arguments used when creating every engine. These are
+        lower precedence than application config. See :func:`sqlalchemy.create_engine`
+        for a list of arguments.
+    :param add_models_to_shell: Add the ``db`` instance and all model classes to
+        ``flask shell``.
+
+    .. versionchanged:: 3.1.0
+        The ``metadata`` parameter can still be used with SQLAlchemy 1.x classes,
+        but is ignored when using SQLAlchemy 2.x style of declarative classes.
+        Instead, specify metadata on your Base class.
+
+    .. versionchanged:: 3.1.0
+        Added the ``disable_autonaming`` parameter.
+
+    .. versionchanged:: 3.1.0
+        Changed ``model_class`` parameter to accepta SQLAlchemy 2.x
+        declarative base subclass.
+
+    .. versionchanged:: 3.0
+        An active Flask application context is always required to access ``session`` and
+        ``engine``.
+
+    .. versionchanged:: 3.0
+        Separate ``metadata`` are used for each bind key.
+
+    .. versionchanged:: 3.0
+        The ``engine_options`` parameter is applied as defaults before per-engine
+        configuration.
+
+    .. versionchanged:: 3.0
+        The session class can be customized in ``session_options``.
+
+    .. versionchanged:: 3.0
+        Added the ``add_models_to_shell`` parameter.
+
+    .. versionchanged:: 3.0
+        Engines are created when calling ``init_app`` rather than the first time they
+        are accessed.
+
+    .. versionchanged:: 3.0
+        All parameters except ``app`` are keyword-only.
+
+    .. versionchanged:: 3.0
+        The extension instance is stored directly as ``app.extensions["sqlalchemy"]``.
+
+    .. versionchanged:: 3.0
+        Setup methods are renamed with a leading underscore. They are considered
+        internal interfaces which may change at any time.
+
+    .. versionchanged:: 3.0
+        Removed the ``use_native_unicode`` parameter and config.
+
+    .. versionchanged:: 2.4
+        Added the ``engine_options`` parameter.
+
+    .. versionchanged:: 2.1
+        Added the ``metadata``, ``query_class``, and ``model_class`` parameters.
+
+    .. versionchanged:: 2.1
+        Use the same query class across ``session``, ``Model.query`` and
+        ``Query``.
+
+    .. versionchanged:: 0.16
+        ``scopefunc`` is accepted in ``session_options``.
+
+    .. versionchanged:: 0.10
+        Added the ``session_options`` parameter.
+    """
+
+    def __init__(
+        self,
+        app: Flask | None = None,
+        *,
+        metadata: sa.MetaData | None = None,
+        session_options: dict[str, t.Any] | None = None,
+        query_class: type[Query] = Query,
+        model_class: _FSA_MCT = Model,  # type: ignore[assignment]
+        engine_options: dict[str, t.Any] | None = None,
+        add_models_to_shell: bool = True,
+        disable_autonaming: bool = False,
+    ):
+        if session_options is None:
+            session_options = {}
+
+        self.Query = query_class
+        """The default query class used by ``Model.query`` and ``lazy="dynamic"``
+        relationships.
+
+        .. warning::
+            The query interface is considered legacy in SQLAlchemy.
+
+        Customize this by passing the ``query_class`` parameter to the extension.
+        """
+
+        self.session = self._make_scoped_session(session_options)
+        """A :class:`sqlalchemy.orm.scoping.scoped_session` that creates instances of
+        :class:`.Session` scoped to the current Flask application context. The session
+        will be removed, returning the engine connection to the pool, when the
+        application context exits.
+
+        Customize this by passing ``session_options`` to the extension.
+
+        This requires that a Flask application context is active.
+
+        .. versionchanged:: 3.0
+            The session is scoped to the current app context.
+        """
+
+        self.metadatas: dict[str | None, sa.MetaData] = {}
+        """Map of bind keys to :class:`sqlalchemy.schema.MetaData` instances. The
+        ``None`` key refers to the default metadata, and is available as
+        :attr:`metadata`.
+
+        Customize the default metadata by passing the ``metadata`` parameter to the
+        extension. This can be used to set a naming convention. When metadata for
+        another bind key is created, it copies the default's naming convention.
+
+        .. versionadded:: 3.0
+        """
+
+        if metadata is not None:
+            if len(_get_2x_declarative_bases(model_class)) > 0:
+                warnings.warn(
+                    "When using SQLAlchemy 2.x style of declarative classes,"
+                    " the `metadata` should be an attribute of the base class."
+                    "The metadata passed into SQLAlchemy() is ignored.",
+                    DeprecationWarning,
+                    stacklevel=2,
+                )
+            else:
+                metadata.info["bind_key"] = None
+                self.metadatas[None] = metadata
+
+        self.Table = self._make_table_class()
+        """A :class:`sqlalchemy.schema.Table` class that chooses a metadata
+        automatically.
+
+        Unlike the base ``Table``, the ``metadata`` argument is not required. If it is
+        not given, it is selected based on the ``bind_key`` argument.
+
+        :param bind_key: Used to select a different metadata.
+        :param args: Arguments passed to the base class. These are typically the table's
+            name, columns, and constraints.
+        :param kwargs: Arguments passed to the base class.
+
+        .. versionchanged:: 3.0
+            This is a subclass of SQLAlchemy's ``Table`` rather than a function.
+        """
+
+        self.Model = self._make_declarative_base(
+            model_class, disable_autonaming=disable_autonaming
+        )
+        """A SQLAlchemy declarative model class. Subclass this to define database
+        models.
+
+        If a model does not set ``__tablename__``, it will be generated by converting
+        the class name from ``CamelCase`` to ``snake_case``. It will not be generated
+        if the model looks like it uses single-table inheritance.
+
+        If a model or parent class sets ``__bind_key__``, it will use that metadata and
+        database engine. Otherwise, it will use the default :attr:`metadata` and
+        :attr:`engine`. This is ignored if the model sets ``metadata`` or ``__table__``.
+
+        For code using the SQLAlchemy 1.x API, customize this model by subclassing
+        :class:`.Model` and passing the ``model_class`` parameter to the extension.
+        A fully created declarative model class can be
+        passed as well, to use a custom metaclass.
+
+        For code using the SQLAlchemy 2.x API, customize this model by subclassing
+        :class:`sqlalchemy.orm.DeclarativeBase` or
+        :class:`sqlalchemy.orm.DeclarativeBaseNoMeta`
+        and passing the ``model_class`` parameter to the extension.
+        """
+
+        if engine_options is None:
+            engine_options = {}
+
+        self._engine_options = engine_options
+        self._app_engines: WeakKeyDictionary[Flask, dict[str | None, sa.engine.Engine]]
+        self._app_engines = WeakKeyDictionary()
+        self._add_models_to_shell = add_models_to_shell
+
+        if app is not None:
+            self.init_app(app)
+
+    def __repr__(self) -> str:
+        if not has_app_context():
+            return f"<{type(self).__name__}>"
+
+        message = f"{type(self).__name__} {self.engine.url}"
+
+        if len(self.engines) > 1:
+            message = f"{message} +{len(self.engines) - 1}"
+
+        return f"<{message}>"
+
+    def init_app(self, app: Flask) -> None:
+        """Initialize a Flask application for use with this extension instance. This
+        must be called before accessing the database engine or session with the app.
+
+        This sets default configuration values, then configures the extension on the
+        application and creates the engines for each bind key. Therefore, this must be
+        called after the application has been configured. Changes to application config
+        after this call will not be reflected.
+
+        The following keys from ``app.config`` are used:
+
+        - :data:`.SQLALCHEMY_DATABASE_URI`
+        - :data:`.SQLALCHEMY_ENGINE_OPTIONS`
+        - :data:`.SQLALCHEMY_ECHO`
+        - :data:`.SQLALCHEMY_BINDS`
+        - :data:`.SQLALCHEMY_RECORD_QUERIES`
+        - :data:`.SQLALCHEMY_TRACK_MODIFICATIONS`
+
+        :param app: The Flask application to initialize.
+        """
+        if "sqlalchemy" in app.extensions:
+            raise RuntimeError(
+                "A 'SQLAlchemy' instance has already been registered on this Flask app."
+                " Import and use that instance instead."
+            )
+
+        app.extensions["sqlalchemy"] = self
+        app.teardown_appcontext(self._teardown_session)
+
+        if self._add_models_to_shell:
+            from .cli import add_models_to_shell
+
+            app.shell_context_processor(add_models_to_shell)
+
+        basic_uri: str | sa.engine.URL | None = app.config.setdefault(
+            "SQLALCHEMY_DATABASE_URI", None
+        )
+        basic_engine_options = self._engine_options.copy()
+        basic_engine_options.update(
+            app.config.setdefault("SQLALCHEMY_ENGINE_OPTIONS", {})
+        )
+        echo: bool = app.config.setdefault("SQLALCHEMY_ECHO", False)
+        config_binds: dict[
+            str | None, str | sa.engine.URL | dict[str, t.Any]
+        ] = app.config.setdefault("SQLALCHEMY_BINDS", {})
+        engine_options: dict[str | None, dict[str, t.Any]] = {}
+
+        # Build the engine config for each bind key.
+        for key, value in config_binds.items():
+            engine_options[key] = self._engine_options.copy()
+
+            if isinstance(value, (str, sa.engine.URL)):
+                engine_options[key]["url"] = value
+            else:
+                engine_options[key].update(value)
+
+        # Build the engine config for the default bind key.
+        if basic_uri is not None:
+            basic_engine_options["url"] = basic_uri
+
+        if "url" in basic_engine_options:
+            engine_options.setdefault(None, {}).update(basic_engine_options)
+
+        if not engine_options:
+            raise RuntimeError(
+                "Either 'SQLALCHEMY_DATABASE_URI' or 'SQLALCHEMY_BINDS' must be set."
+            )
+
+        engines = self._app_engines.setdefault(app, {})
+
+        # Dispose existing engines in case init_app is called again.
+        if engines:
+            for engine in engines.values():
+                engine.dispose()
+
+            engines.clear()
+
+        # Create the metadata and engine for each bind key.
+        for key, options in engine_options.items():
+            self._make_metadata(key)
+            options.setdefault("echo", echo)
+            options.setdefault("echo_pool", echo)
+            self._apply_driver_defaults(options, app)
+            engines[key] = self._make_engine(key, options, app)
+
+        if app.config.setdefault("SQLALCHEMY_RECORD_QUERIES", False):
+            from . import record_queries
+
+            for engine in engines.values():
+                record_queries._listen(engine)
+
+        if app.config.setdefault("SQLALCHEMY_TRACK_MODIFICATIONS", False):
+            from . import track_modifications
+
+            track_modifications._listen(self.session)
+
+    def _make_scoped_session(
+        self, options: dict[str, t.Any]
+    ) -> sa_orm.scoped_session[Session]:
+        """Create a :class:`sqlalchemy.orm.scoping.scoped_session` around the factory
+        from :meth:`_make_session_factory`. The result is available as :attr:`session`.
+
+        The scope function can be customized using the ``scopefunc`` key in the
+        ``session_options`` parameter to the extension. By default it uses the current
+        thread or greenlet id.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param options: The ``session_options`` parameter from ``__init__``. Keyword
+            arguments passed to the session factory. A ``scopefunc`` key is popped.
+
+        .. versionchanged:: 3.0
+            The session is scoped to the current app context.
+
+        .. versionchanged:: 3.0
+            Renamed from ``create_scoped_session``, this method is internal.
+        """
+        scope = options.pop("scopefunc", _app_ctx_id)
+        factory = self._make_session_factory(options)
+        return sa_orm.scoped_session(factory, scope)
+
+    def _make_session_factory(
+        self, options: dict[str, t.Any]
+    ) -> sa_orm.sessionmaker[Session]:
+        """Create the SQLAlchemy :class:`sqlalchemy.orm.sessionmaker` used by
+        :meth:`_make_scoped_session`.
+
+        To customize, pass the ``session_options`` parameter to :class:`SQLAlchemy`. To
+        customize the session class, subclass :class:`.Session` and pass it as the
+        ``class_`` key.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param options: The ``session_options`` parameter from ``__init__``. Keyword
+            arguments passed to the session factory.
+
+        .. versionchanged:: 3.0
+            The session class can be customized.
+
+        .. versionchanged:: 3.0
+            Renamed from ``create_session``, this method is internal.
+        """
+        options.setdefault("class_", Session)
+        options.setdefault("query_cls", self.Query)
+        return sa_orm.sessionmaker(db=self, **options)
+
+    def _teardown_session(self, exc: BaseException | None) -> None:
+        """Remove the current session at the end of the request.
+
+        :meta private:
+
+        .. versionadded:: 3.0
+        """
+        self.session.remove()
+
+    def _make_metadata(self, bind_key: str | None) -> sa.MetaData:
+        """Get or create a :class:`sqlalchemy.schema.MetaData` for the given bind key.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param bind_key: The name of the metadata being created.
+
+        .. versionadded:: 3.0
+        """
+        if bind_key in self.metadatas:
+            return self.metadatas[bind_key]
+
+        if bind_key is not None:
+            # Copy the naming convention from the default metadata.
+            naming_convention = self._make_metadata(None).naming_convention
+        else:
+            naming_convention = None
+
+        # Set the bind key in info to be used by session.get_bind.
+        metadata = sa.MetaData(
+            naming_convention=naming_convention, info={"bind_key": bind_key}
+        )
+        self.metadatas[bind_key] = metadata
+        return metadata
+
+    def _make_table_class(self) -> type[_Table]:
+        """Create a SQLAlchemy :class:`sqlalchemy.schema.Table` class that chooses a
+        metadata automatically based on the ``bind_key``. The result is available as
+        :attr:`Table`.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        .. versionadded:: 3.0
+        """
+
+        class Table(_Table):
+            def __new__(
+                cls, *args: t.Any, bind_key: str | None = None, **kwargs: t.Any
+            ) -> Table:
+                # If a metadata arg is passed, go directly to the base Table. Also do
+                # this for no args so the correct error is shown.
+                if not args or (len(args) >= 2 and isinstance(args[1], sa.MetaData)):
+                    return super().__new__(cls, *args, **kwargs)
+
+                metadata = self._make_metadata(bind_key)
+                return super().__new__(cls, *[args[0], metadata, *args[1:]], **kwargs)
+
+        return Table
+
+    def _make_declarative_base(
+        self,
+        model_class: _FSA_MCT,
+        disable_autonaming: bool = False,
+    ) -> t.Type[_FSAModel]:
+        """Create a SQLAlchemy declarative model class. The result is available as
+        :attr:`Model`.
+
+        To customize, subclass :class:`.Model` and pass it as ``model_class`` to
+        :class:`SQLAlchemy`. To customize at the metaclass level, pass an already
+        created declarative model class as ``model_class``.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param model_class: A model base class, or an already created declarative model
+        class.
+
+        :param disable_autonaming: Turns off automatic tablename generation in models.
+
+        .. versionchanged:: 3.1.0
+            Added support for passing SQLAlchemy 2.x base class as model class.
+            Added optional ``disable_autonaming`` parameter.
+
+        .. versionchanged:: 3.0
+            Renamed with a leading underscore, this method is internal.
+
+        .. versionchanged:: 2.3
+            ``model`` can be an already created declarative model class.
+        """
+        model: t.Type[_FSAModel]
+        declarative_bases = _get_2x_declarative_bases(model_class)
+        if len(declarative_bases) > 1:
+            # raise error if more than one declarative base is found
+            raise ValueError(
+                "Only one declarative base can be passed to SQLAlchemy."
+                " Got: {}".format(model_class.__bases__)
+            )
+        elif len(declarative_bases) == 1:
+            body = dict(model_class.__dict__)
+            body["__fsa__"] = self
+            mixin_classes = [BindMixin, NameMixin, Model]
+            if disable_autonaming:
+                mixin_classes.remove(NameMixin)
+            model = types.new_class(
+                "FlaskSQLAlchemyBase",
+                (*mixin_classes, *model_class.__bases__),
+                {"metaclass": type(declarative_bases[0])},
+                lambda ns: ns.update(body),
+            )
+        elif not isinstance(model_class, sa_orm.DeclarativeMeta):
+            metadata = self._make_metadata(None)
+            metaclass = DefaultMetaNoName if disable_autonaming else DefaultMeta
+            model = sa_orm.declarative_base(
+                metadata=metadata, cls=model_class, name="Model", metaclass=metaclass
+            )
+        else:
+            model = model_class  # type: ignore[assignment]
+
+        if None not in self.metadatas:
+            # Use the model's metadata as the default metadata.
+            model.metadata.info["bind_key"] = None
+            self.metadatas[None] = model.metadata
+        else:
+            # Use the passed in default metadata as the model's metadata.
+            model.metadata = self.metadatas[None]
+
+        model.query_class = self.Query
+        model.query = _QueryProperty()  # type: ignore[assignment]
+        model.__fsa__ = self
+        return model
+
+    def _apply_driver_defaults(self, options: dict[str, t.Any], app: Flask) -> None:
+        """Apply driver-specific configuration to an engine.
+
+        SQLite in-memory databases use ``StaticPool`` and disable ``check_same_thread``.
+        File paths are relative to the app's :attr:`~flask.Flask.instance_path`,
+        which is created if it doesn't exist.
+
+        MySQL sets ``charset="utf8mb4"``, and ``pool_timeout`` defaults to 2 hours.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param options: Arguments passed to the engine.
+        :param app: The application that the engine configuration belongs to.
+
+        .. versionchanged:: 3.0
+            SQLite paths are relative to ``app.instance_path``. It does not use
+            ``NullPool`` if ``pool_size`` is 0. Driver-level URIs are supported.
+
+        .. versionchanged:: 3.0
+            MySQL sets ``charset="utf8mb4". It does not set ``pool_size`` to 10. It
+            does not set ``pool_recycle`` if not using a queue pool.
+
+        .. versionchanged:: 3.0
+            Renamed from ``apply_driver_hacks``, this method is internal. It does not
+            return anything.
+
+        .. versionchanged:: 2.5
+            Returns ``(sa_url, options)``.
+        """
+        url = sa.engine.make_url(options["url"])
+
+        if url.drivername in {"sqlite", "sqlite+pysqlite"}:
+            if url.database is None or url.database in {"", ":memory:"}:
+                options["poolclass"] = sa.pool.StaticPool
+
+                if "connect_args" not in options:
+                    options["connect_args"] = {}
+
+                options["connect_args"]["check_same_thread"] = False
+            else:
+                # the url might look like sqlite:///file:path?uri=true
+                is_uri = url.query.get("uri", False)
+
+                if is_uri:
+                    db_str = url.database[5:]
+                else:
+                    db_str = url.database
+
+                if not os.path.isabs(db_str):
+                    os.makedirs(app.instance_path, exist_ok=True)
+                    db_str = os.path.join(app.instance_path, db_str)
+
+                    if is_uri:
+                        db_str = f"file:{db_str}"
+
+                    options["url"] = url.set(database=db_str)
+        elif url.drivername.startswith("mysql"):
+            # set queue defaults only when using queue pool
+            if (
+                "pool_class" not in options
+                or options["pool_class"] is sa.pool.QueuePool
+            ):
+                options.setdefault("pool_recycle", 7200)
+
+            if "charset" not in url.query:
+                options["url"] = url.update_query_dict({"charset": "utf8mb4"})
+
+    def _make_engine(
+        self, bind_key: str | None, options: dict[str, t.Any], app: Flask
+    ) -> sa.engine.Engine:
+        """Create the :class:`sqlalchemy.engine.Engine` for the given bind key and app.
+
+        To customize, use :data:`.SQLALCHEMY_ENGINE_OPTIONS` or
+        :data:`.SQLALCHEMY_BINDS` config. Pass ``engine_options`` to :class:`SQLAlchemy`
+        to set defaults for all engines.
+
+        This method is used for internal setup. Its signature may change at any time.
+
+        :meta private:
+
+        :param bind_key: The name of the engine being created.
+        :param options: Arguments passed to the engine.
+        :param app: The application that the engine configuration belongs to.
+
+        .. versionchanged:: 3.0
+            Renamed from ``create_engine``, this method is internal.
+        """
+        return sa.engine_from_config(options, prefix="")
+
+    @property
+    def metadata(self) -> sa.MetaData:
+        """The default metadata used by :attr:`Model` and :attr:`Table` if no bind key
+        is set.
+        """
+        return self.metadatas[None]
+
+    @property
+    def engines(self) -> t.Mapping[str | None, sa.engine.Engine]:
+        """Map of bind keys to :class:`sqlalchemy.engine.Engine` instances for current
+        application. The ``None`` key refers to the default engine, and is available as
+        :attr:`engine`.
+
+        To customize, set the :data:`.SQLALCHEMY_BINDS` config, and set defaults by
+        passing the ``engine_options`` parameter to the extension.
+
+        This requires that a Flask application context is active.
+
+        .. versionadded:: 3.0
+        """
+        app = current_app._get_current_object()  # type: ignore[attr-defined]
+
+        if app not in self._app_engines:
+            raise RuntimeError(
+                "The current Flask app is not registered with this 'SQLAlchemy'"
+                " instance. Did you forget to call 'init_app', or did you create"
+                " multiple 'SQLAlchemy' instances?"
+            )
+
+        return self._app_engines[app]
+
+    @property
+    def engine(self) -> sa.engine.Engine:
+        """The default :class:`~sqlalchemy.engine.Engine` for the current application,
+        used by :attr:`session` if the :attr:`Model` or :attr:`Table` being queried does
+        not set a bind key.
+
+        To customize, set the :data:`.SQLALCHEMY_ENGINE_OPTIONS` config, and set
+        defaults by passing the ``engine_options`` parameter to the extension.
+
+        This requires that a Flask application context is active.
+        """
+        return self.engines[None]
+
+    def get_engine(
+        self, bind_key: str | None = None, **kwargs: t.Any
+    ) -> sa.engine.Engine:
+        """Get the engine for the given bind key for the current application.
+        This requires that a Flask application context is active.
+
+        :param bind_key: The name of the engine.
+
+        .. deprecated:: 3.0
+            Will be removed in Flask-SQLAlchemy 3.2. Use ``engines[key]`` instead.
+
+        .. versionchanged:: 3.0
+            Renamed the ``bind`` parameter to ``bind_key``. Removed the ``app``
+            parameter.
+        """
+        warnings.warn(
+            "'get_engine' is deprecated and will be removed in Flask-SQLAlchemy"
+            " 3.2. Use 'engine' or 'engines[key]' instead. If you're using"
+            " Flask-Migrate or Alembic, you'll need to update your 'env.py' file.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        if "bind" in kwargs:
+            bind_key = kwargs.pop("bind")
+
+        return self.engines[bind_key]
+
+    def get_or_404(
+        self,
+        entity: type[_O],
+        ident: t.Any,
+        *,
+        description: str | None = None,
+        **kwargs: t.Any,
+    ) -> _O:
+        """Like :meth:`session.get() <sqlalchemy.orm.Session.get>` but aborts with a
+        ``404 Not Found`` error instead of returning ``None``.
+
+        :param entity: The model class to query.
+        :param ident: The primary key to query.
+        :param description: A custom message to show on the error page.
+        :param kwargs: Extra arguments passed to ``session.get()``.
+
+        .. versionchanged:: 3.1
+            Pass extra keyword arguments to ``session.get()``.
+
+        .. versionadded:: 3.0
+        """
+        value = self.session.get(entity, ident, **kwargs)
+
+        if value is None:
+            abort(404, description=description)
+
+        return value
+
+    def first_or_404(
+        self, statement: sa.sql.Select[t.Any], *, description: str | None = None
+    ) -> t.Any:
+        """Like :meth:`Result.scalar() <sqlalchemy.engine.Result.scalar>`, but aborts
+        with a ``404 Not Found`` error instead of returning ``None``.
+
+        :param statement: The ``select`` statement to execute.
+        :param description: A custom message to show on the error page.
+
+        .. versionadded:: 3.0
+        """
+        value = self.session.execute(statement).scalar()
+
+        if value is None:
+            abort(404, description=description)
+
+        return value
+
+    def one_or_404(
+        self, statement: sa.sql.Select[t.Any], *, description: str | None = None
+    ) -> t.Any:
+        """Like :meth:`Result.scalar_one() <sqlalchemy.engine.Result.scalar_one>`,
+        but aborts with a ``404 Not Found`` error instead of raising ``NoResultFound``
+        or ``MultipleResultsFound``.
+
+        :param statement: The ``select`` statement to execute.
+        :param description: A custom message to show on the error page.
+
+        .. versionadded:: 3.0
+        """
+        try:
+            return self.session.execute(statement).scalar_one()
+        except (sa_exc.NoResultFound, sa_exc.MultipleResultsFound):
+            abort(404, description=description)
+
+    def paginate(
+        self,
+        select: sa.sql.Select[t.Any],
+        *,
+        page: int | None = None,
+        per_page: int | None = None,
+        max_per_page: int | None = None,
+        error_out: bool = True,
+        count: bool = True,
+    ) -> Pagination:
+        """Apply an offset and limit to a select statment based on the current page and
+        number of items per page, returning a :class:`.Pagination` object.
+
+        The statement should select a model class, like ``select(User)``. This applies
+        ``unique()`` and ``scalars()`` modifiers to the result, so compound selects will
+        not return the expected results.
+
+        :param select: The ``select`` statement to paginate.
+        :param page: The current page, used to calculate the offset. Defaults to the
+            ``page`` query arg during a request, or 1 otherwise.
+        :param per_page: The maximum number of items on a page, used to calculate the
+            offset and limit. Defaults to the ``per_page`` query arg during a request,
+            or 20 otherwise.
+        :param max_per_page: The maximum allowed value for ``per_page``, to limit a
+            user-provided value. Use ``None`` for no limit. Defaults to 100.
+        :param error_out: Abort with a ``404 Not Found`` error if no items are returned
+            and ``page`` is not 1, or if ``page`` or ``per_page`` is less than 1, or if
+            either are not ints.
+        :param count: Calculate the total number of values by issuing an extra count
+            query. For very complex queries this may be inaccurate or slow, so it can be
+            disabled and set manually if necessary.
+
+        .. versionchanged:: 3.0
+            The ``count`` query is more efficient.
+
+        .. versionadded:: 3.0
+        """
+        return SelectPagination(
+            select=select,
+            session=self.session(),
+            page=page,
+            per_page=per_page,
+            max_per_page=max_per_page,
+            error_out=error_out,
+            count=count,
+        )
+
+    def _call_for_binds(
+        self, bind_key: str | None | list[str | None], op_name: str
+    ) -> None:
+        """Call a method on each metadata.
+
+        :meta private:
+
+        :param bind_key: A bind key or list of keys. Defaults to all binds.
+        :param op_name: The name of the method to call.
+
+        .. versionchanged:: 3.0
+            Renamed from ``_execute_for_all_tables``.
+        """
+        if bind_key == "__all__":
+            keys: list[str | None] = list(self.metadatas)
+        elif bind_key is None or isinstance(bind_key, str):
+            keys = [bind_key]
+        else:
+            keys = bind_key
+
+        for key in keys:
+            try:
+                engine = self.engines[key]
+            except KeyError:
+                message = f"Bind key '{key}' is not in 'SQLALCHEMY_BINDS' config."
+
+                if key is None:
+                    message = f"'SQLALCHEMY_DATABASE_URI' config is not set. {message}"
+
+                raise sa_exc.UnboundExecutionError(message) from None
+
+            metadata = self.metadatas[key]
+            getattr(metadata, op_name)(bind=engine)
+
+    def create_all(self, bind_key: str | None | list[str | None] = "__all__") -> None:
+        """Create tables that do not exist in the database by calling
+        ``metadata.create_all()`` for all or some bind keys. This does not
+        update existing tables, use a migration library for that.
+
+        This requires that a Flask application context is active.
+
+        :param bind_key: A bind key or list of keys to create the tables for. Defaults
+            to all binds.
+
+        .. versionchanged:: 3.0
+            Renamed the ``bind`` parameter to ``bind_key``. Removed the ``app``
+            parameter.
+
+        .. versionchanged:: 0.12
+            Added the ``bind`` and ``app`` parameters.
+        """
+        self._call_for_binds(bind_key, "create_all")
+
+    def drop_all(self, bind_key: str | None | list[str | None] = "__all__") -> None:
+        """Drop tables by calling ``metadata.drop_all()`` for all or some bind keys.
+
+        This requires that a Flask application context is active.
+
+        :param bind_key: A bind key or list of keys to drop the tables from. Defaults to
+            all binds.
+
+        .. versionchanged:: 3.0
+            Renamed the ``bind`` parameter to ``bind_key``. Removed the ``app``
+            parameter.
+
+        .. versionchanged:: 0.12
+            Added the ``bind`` and ``app`` parameters.
+        """
+        self._call_for_binds(bind_key, "drop_all")
+
+    def reflect(self, bind_key: str | None | list[str | None] = "__all__") -> None:
+        """Load table definitions from the database by calling ``metadata.reflect()``
+        for all or some bind keys.
+
+        This requires that a Flask application context is active.
+
+        :param bind_key: A bind key or list of keys to reflect the tables from. Defaults
+            to all binds.
+
+        .. versionchanged:: 3.0
+            Renamed the ``bind`` parameter to ``bind_key``. Removed the ``app``
+            parameter.
+
+        .. versionchanged:: 0.12
+            Added the ``bind`` and ``app`` parameters.
+        """
+        self._call_for_binds(bind_key, "reflect")
+
+    def _set_rel_query(self, kwargs: dict[str, t.Any]) -> None:
+        """Apply the extension's :attr:`Query` class as the default for relationships
+        and backrefs.
+
+        :meta private:
+        """
+        kwargs.setdefault("query_class", self.Query)
+
+        if "backref" in kwargs:
+            backref = kwargs["backref"]
+
+            if isinstance(backref, str):
+                backref = (backref, {})
+
+            backref[1].setdefault("query_class", self.Query)
+
+    def relationship(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> sa_orm.RelationshipProperty[t.Any]:
+        """A :func:`sqlalchemy.orm.relationship` that applies this extension's
+        :attr:`Query` class for dynamic relationships and backrefs.
+
+        .. versionchanged:: 3.0
+            The :attr:`Query` class is set on ``backref``.
+        """
+        self._set_rel_query(kwargs)
+        return sa_orm.relationship(*args, **kwargs)
+
+    def dynamic_loader(
+        self, argument: t.Any, **kwargs: t.Any
+    ) -> sa_orm.RelationshipProperty[t.Any]:
+        """A :func:`sqlalchemy.orm.dynamic_loader` that applies this extension's
+        :attr:`Query` class for relationships and backrefs.
+
+        .. versionchanged:: 3.0
+            The :attr:`Query` class is set on ``backref``.
+        """
+        self._set_rel_query(kwargs)
+        return sa_orm.dynamic_loader(argument, **kwargs)
+
+    def _relation(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> sa_orm.RelationshipProperty[t.Any]:
+        """A :func:`sqlalchemy.orm.relationship` that applies this extension's
+        :attr:`Query` class for dynamic relationships and backrefs.
+
+        SQLAlchemy 2.0 removes this name, use ``relationship`` instead.
+
+        :meta private:
+
+        .. versionchanged:: 3.0
+            The :attr:`Query` class is set on ``backref``.
+        """
+        self._set_rel_query(kwargs)
+        f = sa_orm.relationship
+        return f(*args, **kwargs)
+
+    def __getattr__(self, name: str) -> t.Any:
+        if name == "relation":
+            return self._relation
+
+        if name == "event":
+            return sa_event
+
+        if name.startswith("_"):
+            raise AttributeError(name)
+
+        for mod in (sa, sa_orm):
+            if hasattr(mod, name):
+                return getattr(mod, name)
+
+        raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/model.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/model.py
new file mode 100644
index 00000000..c6f9e5a9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/model.py
@@ -0,0 +1,330 @@
+from __future__ import annotations
+
+import re
+import typing as t
+
+import sqlalchemy as sa
+import sqlalchemy.orm as sa_orm
+
+from .query import Query
+
+if t.TYPE_CHECKING:
+    from .extension import SQLAlchemy
+
+
+class _QueryProperty:
+    """A class property that creates a query object for a model.
+
+    :meta private:
+    """
+
+    def __get__(self, obj: Model | None, cls: type[Model]) -> Query:
+        return cls.query_class(
+            cls, session=cls.__fsa__.session()  # type: ignore[arg-type]
+        )
+
+
+class Model:
+    """The base class of the :attr:`.SQLAlchemy.Model` declarative model class.
+
+    To define models, subclass :attr:`db.Model <.SQLAlchemy.Model>`, not this. To
+    customize ``db.Model``, subclass this and pass it as ``model_class`` to
+    :class:`.SQLAlchemy`. To customize ``db.Model`` at the metaclass level, pass an
+    already created declarative model class as ``model_class``.
+    """
+
+    __fsa__: t.ClassVar[SQLAlchemy]
+    """Internal reference to the extension object.
+
+    :meta private:
+    """
+
+    query_class: t.ClassVar[type[Query]] = Query
+    """Query class used by :attr:`query`. Defaults to :attr:`.SQLAlchemy.Query`, which
+    defaults to :class:`.Query`.
+    """
+
+    query: t.ClassVar[Query] = _QueryProperty()  # type: ignore[assignment]
+    """A SQLAlchemy query for a model. Equivalent to ``db.session.query(Model)``. Can be
+    customized per-model by overriding :attr:`query_class`.
+
+    .. warning::
+        The query interface is considered legacy in SQLAlchemy. Prefer using
+        ``session.execute(select())`` instead.
+    """
+
+    def __repr__(self) -> str:
+        state = sa.inspect(self)
+        assert state is not None
+
+        if state.transient:
+            pk = f"(transient {id(self)})"
+        elif state.pending:
+            pk = f"(pending {id(self)})"
+        else:
+            pk = ", ".join(map(str, state.identity))
+
+        return f"<{type(self).__name__} {pk}>"
+
+
+class BindMetaMixin(type):
+    """Metaclass mixin that sets a model's ``metadata`` based on its ``__bind_key__``.
+
+    If the model sets ``metadata`` or ``__table__`` directly, ``__bind_key__`` is
+    ignored. If the ``metadata`` is the same as the parent model, it will not be set
+    directly on the child model.
+    """
+
+    __fsa__: SQLAlchemy
+    metadata: sa.MetaData
+
+    def __init__(
+        cls, name: str, bases: tuple[type, ...], d: dict[str, t.Any], **kwargs: t.Any
+    ) -> None:
+        if not ("metadata" in cls.__dict__ or "__table__" in cls.__dict__):
+            bind_key = getattr(cls, "__bind_key__", None)
+            parent_metadata = getattr(cls, "metadata", None)
+            metadata = cls.__fsa__._make_metadata(bind_key)
+
+            if metadata is not parent_metadata:
+                cls.metadata = metadata
+
+        super().__init__(name, bases, d, **kwargs)
+
+
+class BindMixin:
+    """DeclarativeBase mixin to set a model's ``metadata`` based on ``__bind_key__``.
+
+    If no ``__bind_key__`` is specified, the model will use the default metadata
+    provided by ``DeclarativeBase`` or ``DeclarativeBaseNoMeta``.
+    If the model doesn't set ``metadata`` or ``__table__`` directly
+    and does set ``__bind_key__``, the model will use the metadata
+    for the specified bind key.
+    If the ``metadata`` is the same as the parent model, it will not be set
+    directly on the child model.
+
+    .. versionchanged:: 3.1.0
+    """
+
+    __fsa__: SQLAlchemy
+    metadata: sa.MetaData
+
+    @classmethod
+    def __init_subclass__(cls: t.Type[BindMixin], **kwargs: t.Dict[str, t.Any]) -> None:
+        if not ("metadata" in cls.__dict__ or "__table__" in cls.__dict__) and hasattr(
+            cls, "__bind_key__"
+        ):
+            bind_key = getattr(cls, "__bind_key__", None)
+            parent_metadata = getattr(cls, "metadata", None)
+            metadata = cls.__fsa__._make_metadata(bind_key)
+
+            if metadata is not parent_metadata:
+                cls.metadata = metadata
+
+        super().__init_subclass__(**kwargs)
+
+
+class NameMetaMixin(type):
+    """Metaclass mixin that sets a model's ``__tablename__`` by converting the
+    ``CamelCase`` class name to ``snake_case``. A name is set for non-abstract models
+    that do not otherwise define ``__tablename__``. If a model does not define a primary
+    key, it will not generate a name or ``__table__``, for single-table inheritance.
+    """
+
+    metadata: sa.MetaData
+    __tablename__: str
+    __table__: sa.Table
+
+    def __init__(
+        cls, name: str, bases: tuple[type, ...], d: dict[str, t.Any], **kwargs: t.Any
+    ) -> None:
+        if should_set_tablename(cls):
+            cls.__tablename__ = camel_to_snake_case(cls.__name__)
+
+        super().__init__(name, bases, d, **kwargs)
+
+        # __table_cls__ has run. If no table was created, use the parent table.
+        if (
+            "__tablename__" not in cls.__dict__
+            and "__table__" in cls.__dict__
+            and cls.__dict__["__table__"] is None
+        ):
+            del cls.__table__
+
+    def __table_cls__(cls, *args: t.Any, **kwargs: t.Any) -> sa.Table | None:
+        """This is called by SQLAlchemy during mapper setup. It determines the final
+        table object that the model will use.
+
+        If no primary key is found, that indicates single-table inheritance, so no table
+        will be created and ``__tablename__`` will be unset.
+        """
+        schema = kwargs.get("schema")
+
+        if schema is None:
+            key = args[0]
+        else:
+            key = f"{schema}.{args[0]}"
+
+        # Check if a table with this name already exists. Allows reflected tables to be
+        # applied to models by name.
+        if key in cls.metadata.tables:
+            return sa.Table(*args, **kwargs)
+
+        # If a primary key is found, create a table for joined-table inheritance.
+        for arg in args:
+            if (isinstance(arg, sa.Column) and arg.primary_key) or isinstance(
+                arg, sa.PrimaryKeyConstraint
+            ):
+                return sa.Table(*args, **kwargs)
+
+        # If no base classes define a table, return one that's missing a primary key
+        # so SQLAlchemy shows the correct error.
+        for base in cls.__mro__[1:-1]:
+            if "__table__" in base.__dict__:
+                break
+        else:
+            return sa.Table(*args, **kwargs)
+
+        # Single-table inheritance, use the parent table name. __init__ will unset
+        # __table__ based on this.
+        if "__tablename__" in cls.__dict__:
+            del cls.__tablename__
+
+        return None
+
+
+class NameMixin:
+    """DeclarativeBase mixin that sets a model's ``__tablename__`` by converting the
+    ``CamelCase`` class name to ``snake_case``. A name is set for non-abstract models
+    that do not otherwise define ``__tablename__``. If a model does not define a primary
+    key, it will not generate a name or ``__table__``, for single-table inheritance.
+
+    .. versionchanged:: 3.1.0
+    """
+
+    metadata: sa.MetaData
+    __tablename__: str
+    __table__: sa.Table
+
+    @classmethod
+    def __init_subclass__(cls: t.Type[NameMixin], **kwargs: t.Dict[str, t.Any]) -> None:
+        if should_set_tablename(cls):
+            cls.__tablename__ = camel_to_snake_case(cls.__name__)
+
+        super().__init_subclass__(**kwargs)
+
+        # __table_cls__ has run. If no table was created, use the parent table.
+        if (
+            "__tablename__" not in cls.__dict__
+            and "__table__" in cls.__dict__
+            and cls.__dict__["__table__"] is None
+        ):
+            del cls.__table__
+
+    @classmethod
+    def __table_cls__(cls, *args: t.Any, **kwargs: t.Any) -> sa.Table | None:
+        """This is called by SQLAlchemy during mapper setup. It determines the final
+        table object that the model will use.
+
+        If no primary key is found, that indicates single-table inheritance, so no table
+        will be created and ``__tablename__`` will be unset.
+        """
+        schema = kwargs.get("schema")
+
+        if schema is None:
+            key = args[0]
+        else:
+            key = f"{schema}.{args[0]}"
+
+        # Check if a table with this name already exists. Allows reflected tables to be
+        # applied to models by name.
+        if key in cls.metadata.tables:
+            return sa.Table(*args, **kwargs)
+
+        # If a primary key is found, create a table for joined-table inheritance.
+        for arg in args:
+            if (isinstance(arg, sa.Column) and arg.primary_key) or isinstance(
+                arg, sa.PrimaryKeyConstraint
+            ):
+                return sa.Table(*args, **kwargs)
+
+        # If no base classes define a table, return one that's missing a primary key
+        # so SQLAlchemy shows the correct error.
+        for base in cls.__mro__[1:-1]:
+            if "__table__" in base.__dict__:
+                break
+        else:
+            return sa.Table(*args, **kwargs)
+
+        # Single-table inheritance, use the parent table name. __init__ will unset
+        # __table__ based on this.
+        if "__tablename__" in cls.__dict__:
+            del cls.__tablename__
+
+        return None
+
+
+def should_set_tablename(cls: type) -> bool:
+    """Determine whether ``__tablename__`` should be generated for a model.
+
+    -   If no class in the MRO sets a name, one should be generated.
+    -   If a declared attr is found, it should be used instead.
+    -   If a name is found, it should be used if the class is a mixin, otherwise one
+        should be generated.
+    -   Abstract models should not have one generated.
+
+    Later, ``__table_cls__`` will determine if the model looks like single or
+    joined-table inheritance. If no primary key is found, the name will be unset.
+    """
+    if (
+        cls.__dict__.get("__abstract__", False)
+        or (
+            not issubclass(cls, (sa_orm.DeclarativeBase, sa_orm.DeclarativeBaseNoMeta))
+            and not any(isinstance(b, sa_orm.DeclarativeMeta) for b in cls.__mro__[1:])
+        )
+        or any(
+            (b is sa_orm.DeclarativeBase or b is sa_orm.DeclarativeBaseNoMeta)
+            for b in cls.__bases__
+        )
+    ):
+        return False
+
+    for base in cls.__mro__:
+        if "__tablename__" not in base.__dict__:
+            continue
+
+        if isinstance(base.__dict__["__tablename__"], sa_orm.declared_attr):
+            return False
+
+        return not (
+            base is cls
+            or base.__dict__.get("__abstract__", False)
+            or not (
+                # SQLAlchemy 1.x
+                isinstance(base, sa_orm.DeclarativeMeta)
+                # 2.x: DeclarativeBas uses this as metaclass
+                or isinstance(base, sa_orm.decl_api.DeclarativeAttributeIntercept)
+                # 2.x: DeclarativeBaseNoMeta doesn't use a metaclass
+                or issubclass(base, sa_orm.DeclarativeBaseNoMeta)
+            )
+        )
+
+    return True
+
+
+def camel_to_snake_case(name: str) -> str:
+    """Convert a ``CamelCase`` name to ``snake_case``."""
+    name = re.sub(r"((?<=[a-z0-9])[A-Z]|(?!^)[A-Z](?=[a-z]))", r"_\1", name)
+    return name.lower().lstrip("_")
+
+
+class DefaultMeta(BindMetaMixin, NameMetaMixin, sa_orm.DeclarativeMeta):
+    """SQLAlchemy declarative metaclass that provides ``__bind_key__`` and
+    ``__tablename__`` support.
+    """
+
+
+class DefaultMetaNoName(BindMetaMixin, sa_orm.DeclarativeMeta):
+    """SQLAlchemy declarative metaclass that provides ``__bind_key__`` and
+    ``__tablename__`` support.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/pagination.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/pagination.py
new file mode 100644
index 00000000..3d49d6e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/pagination.py
@@ -0,0 +1,364 @@
+from __future__ import annotations
+
+import typing as t
+from math import ceil
+
+import sqlalchemy as sa
+import sqlalchemy.orm as sa_orm
+from flask import abort
+from flask import request
+
+
+class Pagination:
+    """Apply an offset and limit to the query based on the current page and number of
+    items per page.
+
+    Don't create pagination objects manually. They are created by
+    :meth:`.SQLAlchemy.paginate` and :meth:`.Query.paginate`.
+
+    This is a base class, a subclass must implement :meth:`_query_items` and
+    :meth:`_query_count`. Those methods will use arguments passed as ``kwargs`` to
+    perform the queries.
+
+    :param page: The current page, used to calculate the offset. Defaults to the
+        ``page`` query arg during a request, or 1 otherwise.
+    :param per_page: The maximum number of items on a page, used to calculate the
+        offset and limit. Defaults to the ``per_page`` query arg during a request,
+        or 20 otherwise.
+    :param max_per_page: The maximum allowed value for ``per_page``, to limit a
+        user-provided value. Use ``None`` for no limit. Defaults to 100.
+    :param error_out: Abort with a ``404 Not Found`` error if no items are returned
+        and ``page`` is not 1, or if ``page`` or ``per_page`` is less than 1, or if
+        either are not ints.
+    :param count: Calculate the total number of values by issuing an extra count
+        query. For very complex queries this may be inaccurate or slow, so it can be
+        disabled and set manually if necessary.
+    :param kwargs: Information about the query to paginate. Different subclasses will
+        require different arguments.
+
+    .. versionchanged:: 3.0
+        Iterating over a pagination object iterates over its items.
+
+    .. versionchanged:: 3.0
+        Creating instances manually is not a public API.
+    """
+
+    def __init__(
+        self,
+        page: int | None = None,
+        per_page: int | None = None,
+        max_per_page: int | None = 100,
+        error_out: bool = True,
+        count: bool = True,
+        **kwargs: t.Any,
+    ) -> None:
+        self._query_args = kwargs
+        page, per_page = self._prepare_page_args(
+            page=page,
+            per_page=per_page,
+            max_per_page=max_per_page,
+            error_out=error_out,
+        )
+
+        self.page: int = page
+        """The current page."""
+
+        self.per_page: int = per_page
+        """The maximum number of items on a page."""
+
+        self.max_per_page: int | None = max_per_page
+        """The maximum allowed value for ``per_page``."""
+
+        items = self._query_items()
+
+        if not items and page != 1 and error_out:
+            abort(404)
+
+        self.items: list[t.Any] = items
+        """The items on the current page. Iterating over the pagination object is
+        equivalent to iterating over the items.
+        """
+
+        if count:
+            total = self._query_count()
+        else:
+            total = None
+
+        self.total: int | None = total
+        """The total number of items across all pages."""
+
+    @staticmethod
+    def _prepare_page_args(
+        *,
+        page: int | None = None,
+        per_page: int | None = None,
+        max_per_page: int | None = None,
+        error_out: bool = True,
+    ) -> tuple[int, int]:
+        if request:
+            if page is None:
+                try:
+                    page = int(request.args.get("page", 1))
+                except (TypeError, ValueError):
+                    if error_out:
+                        abort(404)
+
+                    page = 1
+
+            if per_page is None:
+                try:
+                    per_page = int(request.args.get("per_page", 20))
+                except (TypeError, ValueError):
+                    if error_out:
+                        abort(404)
+
+                    per_page = 20
+        else:
+            if page is None:
+                page = 1
+
+            if per_page is None:
+                per_page = 20
+
+        if max_per_page is not None:
+            per_page = min(per_page, max_per_page)
+
+        if page < 1:
+            if error_out:
+                abort(404)
+            else:
+                page = 1
+
+        if per_page < 1:
+            if error_out:
+                abort(404)
+            else:
+                per_page = 20
+
+        return page, per_page
+
+    @property
+    def _query_offset(self) -> int:
+        """The index of the first item to query, passed to ``offset()``.
+
+        :meta private:
+
+        .. versionadded:: 3.0
+        """
+        return (self.page - 1) * self.per_page
+
+    def _query_items(self) -> list[t.Any]:
+        """Execute the query to get the items on the current page.
+
+        Uses init arguments stored in :attr:`_query_args`.
+
+        :meta private:
+
+        .. versionadded:: 3.0
+        """
+        raise NotImplementedError
+
+    def _query_count(self) -> int:
+        """Execute the query to get the total number of items.
+
+        Uses init arguments stored in :attr:`_query_args`.
+
+        :meta private:
+
+        .. versionadded:: 3.0
+        """
+        raise NotImplementedError
+
+    @property
+    def first(self) -> int:
+        """The number of the first item on the page, starting from 1, or 0 if there are
+        no items.
+
+        .. versionadded:: 3.0
+        """
+        if len(self.items) == 0:
+            return 0
+
+        return (self.page - 1) * self.per_page + 1
+
+    @property
+    def last(self) -> int:
+        """The number of the last item on the page, starting from 1, inclusive, or 0 if
+        there are no items.
+
+        .. versionadded:: 3.0
+        """
+        first = self.first
+        return max(first, first + len(self.items) - 1)
+
+    @property
+    def pages(self) -> int:
+        """The total number of pages."""
+        if self.total == 0 or self.total is None:
+            return 0
+
+        return ceil(self.total / self.per_page)
+
+    @property
+    def has_prev(self) -> bool:
+        """``True`` if this is not the first page."""
+        return self.page > 1
+
+    @property
+    def prev_num(self) -> int | None:
+        """The previous page number, or ``None`` if this is the first page."""
+        if not self.has_prev:
+            return None
+
+        return self.page - 1
+
+    def prev(self, *, error_out: bool = False) -> Pagination:
+        """Query the :class:`Pagination` object for the previous page.
+
+        :param error_out: Abort with a ``404 Not Found`` error if no items are returned
+            and ``page`` is not 1, or if ``page`` or ``per_page`` is less than 1, or if
+            either are not ints.
+        """
+        p = type(self)(
+            page=self.page - 1,
+            per_page=self.per_page,
+            error_out=error_out,
+            count=False,
+            **self._query_args,
+        )
+        p.total = self.total
+        return p
+
+    @property
+    def has_next(self) -> bool:
+        """``True`` if this is not the last page."""
+        return self.page < self.pages
+
+    @property
+    def next_num(self) -> int | None:
+        """The next page number, or ``None`` if this is the last page."""
+        if not self.has_next:
+            return None
+
+        return self.page + 1
+
+    def next(self, *, error_out: bool = False) -> Pagination:
+        """Query the :class:`Pagination` object for the next page.
+
+        :param error_out: Abort with a ``404 Not Found`` error if no items are returned
+            and ``page`` is not 1, or if ``page`` or ``per_page`` is less than 1, or if
+            either are not ints.
+        """
+        p = type(self)(
+            page=self.page + 1,
+            per_page=self.per_page,
+            max_per_page=self.max_per_page,
+            error_out=error_out,
+            count=False,
+            **self._query_args,
+        )
+        p.total = self.total
+        return p
+
+    def iter_pages(
+        self,
+        *,
+        left_edge: int = 2,
+        left_current: int = 2,
+        right_current: int = 4,
+        right_edge: int = 2,
+    ) -> t.Iterator[int | None]:
+        """Yield page numbers for a pagination widget. Skipped pages between the edges
+        and middle are represented by a ``None``.
+
+        For example, if there are 20 pages and the current page is 7, the following
+        values are yielded.
+
+        .. code-block:: python
+
+            1, 2, None, 5, 6, 7, 8, 9, 10, 11, None, 19, 20
+
+        :param left_edge: How many pages to show from the first page.
+        :param left_current: How many pages to show left of the current page.
+        :param right_current: How many pages to show right of the current page.
+        :param right_edge: How many pages to show from the last page.
+
+        .. versionchanged:: 3.0
+            Improved efficiency of calculating what to yield.
+
+        .. versionchanged:: 3.0
+            ``right_current`` boundary is inclusive.
+
+        .. versionchanged:: 3.0
+            All parameters are keyword-only.
+        """
+        pages_end = self.pages + 1
+
+        if pages_end == 1:
+            return
+
+        left_end = min(1 + left_edge, pages_end)
+        yield from range(1, left_end)
+
+        if left_end == pages_end:
+            return
+
+        mid_start = max(left_end, self.page - left_current)
+        mid_end = min(self.page + right_current + 1, pages_end)
+
+        if mid_start - left_end > 0:
+            yield None
+
+        yield from range(mid_start, mid_end)
+
+        if mid_end == pages_end:
+            return
+
+        right_start = max(mid_end, pages_end - right_edge)
+
+        if right_start - mid_end > 0:
+            yield None
+
+        yield from range(right_start, pages_end)
+
+    def __iter__(self) -> t.Iterator[t.Any]:
+        yield from self.items
+
+
+class SelectPagination(Pagination):
+    """Returned by :meth:`.SQLAlchemy.paginate`. Takes ``select`` and ``session``
+    arguments in addition to the :class:`Pagination` arguments.
+
+    .. versionadded:: 3.0
+    """
+
+    def _query_items(self) -> list[t.Any]:
+        select = self._query_args["select"]
+        select = select.limit(self.per_page).offset(self._query_offset)
+        session = self._query_args["session"]
+        return list(session.execute(select).unique().scalars())
+
+    def _query_count(self) -> int:
+        select = self._query_args["select"]
+        sub = select.options(sa_orm.lazyload("*")).order_by(None).subquery()
+        session = self._query_args["session"]
+        out = session.execute(sa.select(sa.func.count()).select_from(sub)).scalar()
+        return out  # type: ignore[no-any-return]
+
+
+class QueryPagination(Pagination):
+    """Returned by :meth:`.Query.paginate`. Takes a ``query`` argument in addition to
+    the :class:`Pagination` arguments.
+
+    .. versionadded:: 3.0
+    """
+
+    def _query_items(self) -> list[t.Any]:
+        query = self._query_args["query"]
+        out = query.limit(self.per_page).offset(self._query_offset).all()
+        return out  # type: ignore[no-any-return]
+
+    def _query_count(self) -> int:
+        # Query.count automatically disables eager loads
+        out = self._query_args["query"].order_by(None).count()
+        return out  # type: ignore[no-any-return]
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/py.typed b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/query.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/query.py
new file mode 100644
index 00000000..35f927d2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/query.py
@@ -0,0 +1,105 @@
+from __future__ import annotations
+
+import typing as t
+
+import sqlalchemy.exc as sa_exc
+import sqlalchemy.orm as sa_orm
+from flask import abort
+
+from .pagination import Pagination
+from .pagination import QueryPagination
+
+
+class Query(sa_orm.Query):  # type: ignore[type-arg]
+    """SQLAlchemy :class:`~sqlalchemy.orm.query.Query` subclass with some extra methods
+    useful for querying in a web application.
+
+    This is the default query class for :attr:`.Model.query`.
+
+    .. versionchanged:: 3.0
+        Renamed to ``Query`` from ``BaseQuery``.
+    """
+
+    def get_or_404(self, ident: t.Any, description: str | None = None) -> t.Any:
+        """Like :meth:`~sqlalchemy.orm.Query.get` but aborts with a ``404 Not Found``
+        error instead of returning ``None``.
+
+        :param ident: The primary key to query.
+        :param description: A custom message to show on the error page.
+        """
+        rv = self.get(ident)
+
+        if rv is None:
+            abort(404, description=description)
+
+        return rv
+
+    def first_or_404(self, description: str | None = None) -> t.Any:
+        """Like :meth:`~sqlalchemy.orm.Query.first` but aborts with a ``404 Not Found``
+        error instead of returning ``None``.
+
+        :param description: A custom message to show on the error page.
+        """
+        rv = self.first()
+
+        if rv is None:
+            abort(404, description=description)
+
+        return rv
+
+    def one_or_404(self, description: str | None = None) -> t.Any:
+        """Like :meth:`~sqlalchemy.orm.Query.one` but aborts with a ``404 Not Found``
+        error instead of raising ``NoResultFound`` or ``MultipleResultsFound``.
+
+        :param description: A custom message to show on the error page.
+
+        .. versionadded:: 3.0
+        """
+        try:
+            return self.one()
+        except (sa_exc.NoResultFound, sa_exc.MultipleResultsFound):
+            abort(404, description=description)
+
+    def paginate(
+        self,
+        *,
+        page: int | None = None,
+        per_page: int | None = None,
+        max_per_page: int | None = None,
+        error_out: bool = True,
+        count: bool = True,
+    ) -> Pagination:
+        """Apply an offset and limit to the query based on the current page and number
+        of items per page, returning a :class:`.Pagination` object.
+
+        :param page: The current page, used to calculate the offset. Defaults to the
+            ``page`` query arg during a request, or 1 otherwise.
+        :param per_page: The maximum number of items on a page, used to calculate the
+            offset and limit. Defaults to the ``per_page`` query arg during a request,
+            or 20 otherwise.
+        :param max_per_page: The maximum allowed value for ``per_page``, to limit a
+            user-provided value. Use ``None`` for no limit. Defaults to 100.
+        :param error_out: Abort with a ``404 Not Found`` error if no items are returned
+            and ``page`` is not 1, or if ``page`` or ``per_page`` is less than 1, or if
+            either are not ints.
+        :param count: Calculate the total number of values by issuing an extra count
+            query. For very complex queries this may be inaccurate or slow, so it can be
+            disabled and set manually if necessary.
+
+        .. versionchanged:: 3.0
+            All parameters are keyword-only.
+
+        .. versionchanged:: 3.0
+            The ``count`` query is more efficient.
+
+        .. versionchanged:: 3.0
+            ``max_per_page`` defaults to 100.
+        """
+        return QueryPagination(
+            query=self,
+            page=page,
+            per_page=per_page,
+            max_per_page=max_per_page,
+            error_out=error_out,
+            count=count,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/record_queries.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/record_queries.py
new file mode 100644
index 00000000..e8273be9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/record_queries.py
@@ -0,0 +1,117 @@
+from __future__ import annotations
+
+import dataclasses
+import inspect
+import typing as t
+from time import perf_counter
+
+import sqlalchemy as sa
+import sqlalchemy.event as sa_event
+from flask import current_app
+from flask import g
+from flask import has_app_context
+
+
+def get_recorded_queries() -> list[_QueryInfo]:
+    """Get the list of recorded query information for the current session. Queries are
+    recorded if the config :data:`.SQLALCHEMY_RECORD_QUERIES` is enabled.
+
+    Each query info object has the following attributes:
+
+    ``statement``
+        The string of SQL generated by SQLAlchemy with parameter placeholders.
+    ``parameters``
+        The parameters sent with the SQL statement.
+    ``start_time`` / ``end_time``
+        Timing info about when the query started execution and when the results where
+        returned. Accuracy and value depends on the operating system.
+    ``duration``
+        The time the query took in seconds.
+    ``location``
+        A string description of where in your application code the query was executed.
+        This may not be possible to calculate, and the format is not stable.
+
+    .. versionchanged:: 3.0
+        Renamed from ``get_debug_queries``.
+
+    .. versionchanged:: 3.0
+        The info object is a dataclass instead of a tuple.
+
+    .. versionchanged:: 3.0
+        The info object attribute ``context`` is renamed to ``location``.
+
+    .. versionchanged:: 3.0
+        Not enabled automatically in debug or testing mode.
+    """
+    return g.get("_sqlalchemy_queries", [])  # type: ignore[no-any-return]
+
+
+@dataclasses.dataclass
+class _QueryInfo:
+    """Information about an executed query. Returned by :func:`get_recorded_queries`.
+
+    .. versionchanged:: 3.0
+        Renamed from ``_DebugQueryTuple``.
+
+    .. versionchanged:: 3.0
+        Changed to a dataclass instead of a tuple.
+
+    .. versionchanged:: 3.0
+        ``context`` is renamed to ``location``.
+    """
+
+    statement: str | None
+    parameters: t.Any
+    start_time: float
+    end_time: float
+    location: str
+
+    @property
+    def duration(self) -> float:
+        return self.end_time - self.start_time
+
+
+def _listen(engine: sa.engine.Engine) -> None:
+    sa_event.listen(engine, "before_cursor_execute", _record_start, named=True)
+    sa_event.listen(engine, "after_cursor_execute", _record_end, named=True)
+
+
+def _record_start(context: sa.engine.ExecutionContext, **kwargs: t.Any) -> None:
+    if not has_app_context():
+        return
+
+    context._fsa_start_time = perf_counter()  # type: ignore[attr-defined]
+
+
+def _record_end(context: sa.engine.ExecutionContext, **kwargs: t.Any) -> None:
+    if not has_app_context():
+        return
+
+    if "_sqlalchemy_queries" not in g:
+        g._sqlalchemy_queries = []
+
+    import_top = current_app.import_name.partition(".")[0]
+    import_dot = f"{import_top}."
+    frame = inspect.currentframe()
+
+    while frame:
+        name = frame.f_globals.get("__name__")
+
+        if name and (name == import_top or name.startswith(import_dot)):
+            code = frame.f_code
+            location = f"{code.co_filename}:{frame.f_lineno} ({code.co_name})"
+            break
+
+        frame = frame.f_back
+    else:
+        location = "<unknown>"
+
+    g._sqlalchemy_queries.append(
+        _QueryInfo(
+            statement=context.statement,
+            parameters=context.parameters,
+            start_time=context._fsa_start_time,  # type: ignore[attr-defined]
+            end_time=perf_counter(),
+            location=location,
+        )
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/session.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/session.py
new file mode 100644
index 00000000..631fffa8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/session.py
@@ -0,0 +1,111 @@
+from __future__ import annotations
+
+import typing as t
+
+import sqlalchemy as sa
+import sqlalchemy.exc as sa_exc
+import sqlalchemy.orm as sa_orm
+from flask.globals import app_ctx
+
+if t.TYPE_CHECKING:
+    from .extension import SQLAlchemy
+
+
+class Session(sa_orm.Session):
+    """A SQLAlchemy :class:`~sqlalchemy.orm.Session` class that chooses what engine to
+    use based on the bind key associated with the metadata associated with the thing
+    being queried.
+
+    To customize ``db.session``, subclass this and pass it as the ``class_`` key in the
+    ``session_options`` to :class:`.SQLAlchemy`.
+
+    .. versionchanged:: 3.0
+        Renamed from ``SignallingSession``.
+    """
+
+    def __init__(self, db: SQLAlchemy, **kwargs: t.Any) -> None:
+        super().__init__(**kwargs)
+        self._db = db
+        self._model_changes: dict[object, tuple[t.Any, str]] = {}
+
+    def get_bind(
+        self,
+        mapper: t.Any | None = None,
+        clause: t.Any | None = None,
+        bind: sa.engine.Engine | sa.engine.Connection | None = None,
+        **kwargs: t.Any,
+    ) -> sa.engine.Engine | sa.engine.Connection:
+        """Select an engine based on the ``bind_key`` of the metadata associated with
+        the model or table being queried. If no bind key is set, uses the default bind.
+
+        .. versionchanged:: 3.0.3
+            Fix finding the bind for a joined inheritance model.
+
+        .. versionchanged:: 3.0
+            The implementation more closely matches the base SQLAlchemy implementation.
+
+        .. versionchanged:: 2.1
+            Support joining an external transaction.
+        """
+        if bind is not None:
+            return bind
+
+        engines = self._db.engines
+
+        if mapper is not None:
+            try:
+                mapper = sa.inspect(mapper)
+            except sa_exc.NoInspectionAvailable as e:
+                if isinstance(mapper, type):
+                    raise sa_orm.exc.UnmappedClassError(mapper) from e
+
+                raise
+
+            engine = _clause_to_engine(mapper.local_table, engines)
+
+            if engine is not None:
+                return engine
+
+        if clause is not None:
+            engine = _clause_to_engine(clause, engines)
+
+            if engine is not None:
+                return engine
+
+        if None in engines:
+            return engines[None]
+
+        return super().get_bind(mapper=mapper, clause=clause, bind=bind, **kwargs)
+
+
+def _clause_to_engine(
+    clause: sa.ClauseElement | None,
+    engines: t.Mapping[str | None, sa.engine.Engine],
+) -> sa.engine.Engine | None:
+    """If the clause is a table, return the engine associated with the table's
+    metadata's bind key.
+    """
+    table = None
+
+    if clause is not None:
+        if isinstance(clause, sa.Table):
+            table = clause
+        elif isinstance(clause, sa.UpdateBase) and isinstance(clause.table, sa.Table):
+            table = clause.table
+
+    if table is not None and "bind_key" in table.metadata.info:
+        key = table.metadata.info["bind_key"]
+
+        if key not in engines:
+            raise sa_exc.UnboundExecutionError(
+                f"Bind key '{key}' is not in 'SQLALCHEMY_BINDS' config."
+            )
+
+        return engines[key]
+
+    return None
+
+
+def _app_ctx_id() -> int:
+    """Get the id of the current Flask application context for the session scope."""
+    return id(app_ctx._get_current_object())  # type: ignore[attr-defined]
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/table.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/table.py
new file mode 100644
index 00000000..ab08a692
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/table.py
@@ -0,0 +1,39 @@
+from __future__ import annotations
+
+import typing as t
+
+import sqlalchemy as sa
+import sqlalchemy.sql.schema as sa_sql_schema
+
+
+class _Table(sa.Table):
+    @t.overload
+    def __init__(
+        self,
+        name: str,
+        *args: sa_sql_schema.SchemaItem,
+        bind_key: str | None = None,
+        **kwargs: t.Any,
+    ) -> None:
+        ...
+
+    @t.overload
+    def __init__(
+        self,
+        name: str,
+        metadata: sa.MetaData,
+        *args: sa_sql_schema.SchemaItem,
+        **kwargs: t.Any,
+    ) -> None:
+        ...
+
+    @t.overload
+    def __init__(
+        self, name: str, *args: sa_sql_schema.SchemaItem, **kwargs: t.Any
+    ) -> None:
+        ...
+
+    def __init__(
+        self, name: str, *args: sa_sql_schema.SchemaItem, **kwargs: t.Any
+    ) -> None:
+        super().__init__(name, *args, **kwargs)  # type: ignore[arg-type]
diff --git a/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/track_modifications.py b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/track_modifications.py
new file mode 100644
index 00000000..7028b653
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/flask_sqlalchemy/track_modifications.py
@@ -0,0 +1,88 @@
+from __future__ import annotations
+
+import typing as t
+
+import sqlalchemy as sa
+import sqlalchemy.event as sa_event
+import sqlalchemy.orm as sa_orm
+from flask import current_app
+from flask import has_app_context
+from flask.signals import Namespace  # type: ignore[attr-defined]
+
+if t.TYPE_CHECKING:
+    from .session import Session
+
+_signals = Namespace()
+
+models_committed = _signals.signal("models-committed")
+"""This Blinker signal is sent after the session is committed if there were changed
+models in the session.
+
+The sender is the application that emitted the changes. The receiver is passed the
+``changes`` argument with a list of tuples in the form ``(instance, operation)``.
+The operations are ``"insert"``, ``"update"``, and ``"delete"``.
+"""
+
+before_models_committed = _signals.signal("before-models-committed")
+"""This signal works exactly like :data:`models_committed` but is emitted before the
+commit takes place.
+"""
+
+
+def _listen(session: sa_orm.scoped_session[Session]) -> None:
+    sa_event.listen(session, "before_flush", _record_ops, named=True)
+    sa_event.listen(session, "before_commit", _record_ops, named=True)
+    sa_event.listen(session, "before_commit", _before_commit)
+    sa_event.listen(session, "after_commit", _after_commit)
+    sa_event.listen(session, "after_rollback", _after_rollback)
+
+
+def _record_ops(session: Session, **kwargs: t.Any) -> None:
+    if not has_app_context():
+        return
+
+    if not current_app.config["SQLALCHEMY_TRACK_MODIFICATIONS"]:
+        return
+
+    for targets, operation in (
+        (session.new, "insert"),
+        (session.dirty, "update"),
+        (session.deleted, "delete"),
+    ):
+        for target in targets:
+            state = sa.inspect(target)
+            key = state.identity_key if state.has_identity else id(target)
+            session._model_changes[key] = (target, operation)
+
+
+def _before_commit(session: Session) -> None:
+    if not has_app_context():
+        return
+
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+
+    if not app.config["SQLALCHEMY_TRACK_MODIFICATIONS"]:
+        return
+
+    if session._model_changes:
+        changes = list(session._model_changes.values())
+        before_models_committed.send(app, changes=changes)
+
+
+def _after_commit(session: Session) -> None:
+    if not has_app_context():
+        return
+
+    app = current_app._get_current_object()  # type: ignore[attr-defined]
+
+    if not app.config["SQLALCHEMY_TRACK_MODIFICATIONS"]:
+        return
+
+    if session._model_changes:
+        changes = list(session._model_changes.values())
+        models_committed.send(app, changes=changes)
+        session._model_changes.clear()
+
+
+def _after_rollback(session: Session) -> None:
+    session._model_changes.clear()
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__init__.py b/backend/venv/lib/python3.12/site-packages/gherkin/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__main__.py b/backend/venv/lib/python3.12/site-packages/gherkin/__main__.py
new file mode 100644
index 00000000..55d96540
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/__main__.py
@@ -0,0 +1,30 @@
+import os
+from optparse import OptionParser
+import sys
+if sys.version_info < (3, 0):
+    string_type = basestring
+    if os.name != 'nt':
+        import codecs
+        UTF8Writer = codecs.getwriter('utf8')
+        sys.stdout = UTF8Writer(sys.stdout)
+else:
+    string_type = str
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.realpath(__file__))))
+import json
+from gherkin.stream.gherkin_events import GherkinEvents
+from gherkin.stream.source_events import SourceEvents
+
+parser = OptionParser()
+parser.add_option("--no-source",  action="store_false", dest="print_source",  default=True, help="don't print source events")
+parser.add_option("--no-ast",     action="store_false", dest="print_ast",     default=True, help="don't print ast events")
+parser.add_option("--no-pickles", action="store_false", dest="print_pickles", default=True, help="don't print pickle events")
+
+(options, args) = parser.parse_args()
+
+source_events = SourceEvents(args)
+gherkin_events = GherkinEvents(options)
+
+for source_event in source_events.enum():
+    for event in gherkin_events.enum(source_event):
+        print(json.dumps(event))
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6ee34285
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..06bb9091
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_builder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_builder.cpython-312.pyc
new file mode 100644
index 00000000..aa15a22b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_builder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_node.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_node.cpython-312.pyc
new file mode 100644
index 00000000..6ca535b5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/ast_node.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols.cpython-312.pyc
new file mode 100644
index 00000000..93074313
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols_py3_plus.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols_py3_plus.cpython-312.pyc
new file mode 100644
index 00000000..eec5f207
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/count_symbols_py3_plus.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/dialect.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/dialect.cpython-312.pyc
new file mode 100644
index 00000000..820f2573
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/dialect.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/errors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/errors.cpython-312.pyc
new file mode 100644
index 00000000..d5341ded
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/errors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/gherkin_line.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/gherkin_line.cpython-312.pyc
new file mode 100644
index 00000000..c8c9ba36
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/gherkin_line.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/inout.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/inout.cpython-312.pyc
new file mode 100644
index 00000000..c237019e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/inout.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/parser.cpython-312.pyc
new file mode 100644
index 00000000..3d191ea7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token.cpython-312.pyc
new file mode 100644
index 00000000..effa003e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_formatter_builder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_formatter_builder.cpython-312.pyc
new file mode 100644
index 00000000..fc1984c5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_formatter_builder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher.cpython-312.pyc
new file mode 100644
index 00000000..5f5cefde
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher_markdown.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher_markdown.cpython-312.pyc
new file mode 100644
index 00000000..9cc49b30
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_matcher_markdown.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_scanner.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_scanner.cpython-312.pyc
new file mode 100644
index 00000000..27ffbae8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/__pycache__/token_scanner.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/ast_builder.py b/backend/venv/lib/python3.12/site-packages/gherkin/ast_builder.py
new file mode 100644
index 00000000..7c85fef1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/ast_builder.py
@@ -0,0 +1,250 @@
+from .ast_node import AstNode
+from .errors import AstBuilderException
+from .stream.id_generator import IdGenerator
+
+class AstBuilder(object):
+    def __init__(self, id_generator=None):
+        self.id_generator = id_generator
+        if self.id_generator is None:
+            self.id_generator = IdGenerator()
+        self.reset()
+
+    def reset(self):
+        self.stack = [AstNode('None')]
+        self.comments = []
+        self.id_counter = 0
+
+    def start_rule(self, rule_type):
+        self.stack.append(AstNode(rule_type))
+
+    def end_rule(self, rule_type):
+        node = self.stack.pop()
+        self.current_node.add(node.rule_type, self.transform_node(node))
+
+    def build(self, token):
+        if token.matched_type == 'Comment':
+            self.comments.append({
+                'location': self.get_location(token),
+                'text': token.matched_text
+            })
+        else:
+            self.current_node.add(token.matched_type, token)
+
+    def get_result(self):
+        return self.current_node.get_single('GherkinDocument')
+
+    @property
+    def current_node(self):
+        return self.stack[-1]
+
+    def get_location(self, token, column=None):
+        return (token.location if not column else
+                {'line': token.location['line'], 'column': column})
+
+    def get_tags(self, node):
+        tags = []
+        tags_node = node.get_single('Tags')
+        if not tags_node:
+            return tags
+
+        for token in tags_node.get_tokens('TagLine'):
+            tags += [{'id': self.id_generator.get_next_id(),
+                      'location': self.get_location(token, tag_item['column']),
+                      'name': tag_item['text']} for tag_item in token.matched_items]
+
+        return tags
+
+    def get_table_rows(self, node):
+        rows = [{'id': self.id_generator.get_next_id(),
+                 'location': self.get_location(token),
+                 'cells': self.get_cells(token)} for token in node.get_tokens('TableRow')]
+        self.ensure_cell_count(rows)
+        return rows
+
+    def ensure_cell_count(self, rows):
+        if not rows:
+            return
+
+        cell_count = len(rows[0]['cells'])
+        for row in rows:
+            if len(row['cells']) != cell_count:
+                raise AstBuilderException("inconsistent cell count within the table",
+                                          row['location'])
+
+    def get_cells(self, table_row_token):
+        return [self.reject_nones(
+            {'location': self.get_location(table_row_token, cell_item['column']),
+             'value': cell_item['text']}) for cell_item in table_row_token.matched_items]
+
+    def get_description(self, node):
+        return node.get_single('Description', '')
+
+    def get_steps(self, node):
+        return node.get_items('Step')
+
+    def transform_node(self, node):
+        if node.rule_type == 'Step':
+            step_line = node.get_token('StepLine')
+            step_argument_type = 'dummy_type'
+            step_argument = None
+            if node.get_single('DataTable'):
+                step_argument_type = 'dataTable'
+                step_argument = node.get_single('DataTable')
+            elif node.get_single('DocString'):
+                step_argument_type = 'docString'
+                step_argument = node.get_single('DocString')
+
+            return self.reject_nones({
+                'id': self.id_generator.get_next_id(),
+                'location': self.get_location(step_line),
+                'keyword': step_line.matched_keyword,
+                'keywordType': step_line.matched_keyword_type,
+                'text': step_line.matched_text,
+                step_argument_type: step_argument
+            })
+        elif node.rule_type == 'DocString':
+            separator_token = node.get_tokens('DocStringSeparator')[0]
+            media_type = (separator_token.matched_text if len(separator_token.matched_text) > 0
+                            else None)
+            line_tokens = node.get_tokens('Other')
+            content = '\n'.join([t.matched_text for t in line_tokens])
+
+            return self.reject_nones({
+                'location': self.get_location(separator_token),
+                'content': content,
+                'delimiter': separator_token.matched_keyword,
+                'mediaType': media_type
+            })
+        elif node.rule_type == 'DataTable':
+            rows = self.get_table_rows(node)
+            return self.reject_nones({
+                'location': rows[0]['location'],
+                'rows': rows,
+            })
+        elif node.rule_type == 'Background':
+            background_line = node.get_token('BackgroundLine')
+            description = self.get_description(node)
+            steps = self.get_steps(node)
+
+            return self.reject_nones({
+                'id': self.id_generator.get_next_id(),
+                'location': self.get_location(background_line),
+                'keyword': background_line.matched_keyword,
+                'name': background_line.matched_text,
+                'description': description,
+                'steps': steps
+            })
+        elif node.rule_type == 'ScenarioDefinition':
+            tags = self.get_tags(node)
+            scenario_node = node.get_single('Scenario')
+            scenario_line = scenario_node.get_token('ScenarioLine')
+            description = self.get_description(scenario_node)
+            steps = self.get_steps(scenario_node)
+            examples = scenario_node.get_items('ExamplesDefinition')
+
+            return self.reject_nones({
+                'id': self.id_generator.get_next_id(),
+                'tags': tags,
+                'location': self.get_location(scenario_line),
+                'keyword': scenario_line.matched_keyword,
+                'name': scenario_line.matched_text,
+                'description': description,
+                'steps': steps,
+                'examples': examples
+            })
+        elif node.rule_type == 'ExamplesDefinition':
+            tags = self.get_tags(node)
+            examples_node = node.get_single('Examples')
+            examples_line = examples_node.get_token('ExamplesLine')
+            description = self.get_description(examples_node)
+            examples_table_rows = examples_node.get_single('ExamplesTable')
+            table_header = examples_table_rows[0] if examples_table_rows else None
+            table_body = examples_table_rows[1:] if examples_table_rows else []
+
+            return self.reject_nones({
+                'id': self.id_generator.get_next_id(),
+                'tags': tags,
+                'location': self.get_location(examples_line),
+                'keyword': examples_line.matched_keyword,
+                'name': examples_line.matched_text,
+                'description': description,
+                'tableHeader': table_header,
+                'tableBody': table_body
+            })
+        elif node.rule_type == 'ExamplesTable':
+            return self.get_table_rows(node)
+        elif node.rule_type == 'Description':
+            line_tokens = node.get_tokens('Other')
+            # Trim trailing empty lines
+            last_non_empty = next(i for i, j in reversed(list(enumerate(line_tokens)))
+                                  if j.matched_text)
+            description = '\n'.join([token.matched_text for token in
+                                     line_tokens[:last_non_empty + 1]])
+
+            return description
+        elif node.rule_type == 'Rule':
+            header = node.get_single('RuleHeader')
+            if not header:
+                return
+
+            tags = self.get_tags(header)
+            rule_line = header.get_token('RuleLine')
+            if not rule_line:
+                return
+
+            children = []
+            background = node.get_single('Background')
+            if (background):
+                children.append({'background': background})
+            children = children + [{'scenario': i} for i in node.get_items('ScenarioDefinition')]
+            description = self.get_description(header)
+
+            return self.reject_nones({
+                'id': self.id_generator.get_next_id(),
+                'tags': tags,
+                'location': self.get_location(rule_line),
+                'keyword': rule_line.matched_keyword,
+                'name': rule_line.matched_text,
+                'description': description,
+                'children': children
+            })
+        elif node.rule_type == 'Feature':
+            header = node.get_single('FeatureHeader')
+            if not header:
+                return
+
+            tags = self.get_tags(header)
+            feature_line = header.get_token('FeatureLine')
+            if not feature_line:
+                return
+
+            children = []
+            background = node.get_single('Background')
+            if (background):
+                children.append({'background': background})
+            children = children + [{'scenario': i} for i in node.get_items('ScenarioDefinition')]
+            children = children + [{'rule': i} for i in node.get_items('Rule')]
+            description = self.get_description(header)
+            language = feature_line.matched_gherkin_dialect
+
+            return self.reject_nones({
+                'tags': tags,
+                'location': self.get_location(feature_line),
+                'language': language,
+                'keyword': feature_line.matched_keyword,
+                'name': feature_line.matched_text,
+                'description': description,
+                'children': children
+            })
+        elif node.rule_type == 'GherkinDocument':
+            feature = node.get_single('Feature')
+
+            return self.reject_nones({
+                'feature': feature,
+                'comments': self.comments
+            })
+        else:
+            return node
+
+    def reject_nones(self, values):
+        return {k: v for k, v in values.items() if v is not None}
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/ast_node.py b/backend/venv/lib/python3.12/site-packages/gherkin/ast_node.py
new file mode 100644
index 00000000..595d4ba8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/ast_node.py
@@ -0,0 +1,23 @@
+from collections import defaultdict
+
+
+class AstNode(object):
+
+    def __init__(self, rule_type):
+        self.rule_type = rule_type
+        self._sub_items = defaultdict(list)
+
+    def add(self, rule_type, obj):
+        self._sub_items[rule_type].append(obj)
+
+    def get_single(self, rule_type, default_value=None):
+        return self._sub_items[rule_type][0] if self._sub_items[rule_type] else default_value
+
+    def get_items(self, rule_type):
+        return self._sub_items[rule_type]
+
+    def get_token(self, token_type):
+        return self.get_single(token_type)
+
+    def get_tokens(self, token_type):
+        return self._sub_items[token_type]
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols.py b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols.py
new file mode 100644
index 00000000..da00ea1a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols.py
@@ -0,0 +1,6 @@
+import sys
+
+if sys.version_info[0] == 3:
+    from .count_symbols_py3_plus import count_symbols
+else:
+    from .count_symbols_py2 import count_symbols
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py2.py b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py2.py
new file mode 100644
index 00000000..1363da84
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py2.py
@@ -0,0 +1,6 @@
+import re
+
+REGEX_ASTRAL_SYMBOLS = re.compile(ur'[\uD800-\uDBFF][\uDC00-\uDFFF]', re.UNICODE)
+
+def count_symbols(string):
+    return len(REGEX_ASTRAL_SYMBOLS.sub('_', string))
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py3_plus.py b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py3_plus.py
new file mode 100644
index 00000000..d326c150
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/count_symbols_py3_plus.py
@@ -0,0 +1,2 @@
+def count_symbols(string):
+    return len(string)
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/dialect.py b/backend/venv/lib/python3.12/site-packages/gherkin/dialect.py
new file mode 100644
index 00000000..2289bf69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/dialect.py
@@ -0,0 +1,65 @@
+import io
+import os
+import json
+
+
+DIALECT_FILE_PATH = os.path.join(
+    os.path.dirname(__file__),
+    'gherkin-languages.json')
+
+with io.open(DIALECT_FILE_PATH, 'r', encoding='utf-8') as file:
+    DIALECTS = json.load(file)
+
+
+class Dialect(object):
+
+    @classmethod
+    def for_name(cls, name):
+        return cls(DIALECTS[name]) if name in DIALECTS else None
+
+    def __init__(self, spec):
+        self.spec = spec
+
+    @property
+    def feature_keywords(self):
+        return self.spec['feature']
+
+    @property
+    def rule_keywords(self):
+        return self.spec['rule']
+
+    @property
+    def scenario_keywords(self):
+        return self.spec['scenario']
+
+    @property
+    def scenario_outline_keywords(self):
+        return self.spec['scenarioOutline']
+
+    @property
+    def background_keywords(self):
+        return self.spec['background']
+
+    @property
+    def examples_keywords(self):
+        return self.spec['examples']
+
+    @property
+    def given_keywords(self):
+        return self.spec['given']
+
+    @property
+    def when_keywords(self):
+        return self.spec['when']
+
+    @property
+    def then_keywords(self):
+        return self.spec['then']
+
+    @property
+    def and_keywords(self):
+        return self.spec['and']
+
+    @property
+    def but_keywords(self):
+        return self.spec['but']
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/errors.py b/backend/venv/lib/python3.12/site-packages/gherkin/errors.py
new file mode 100644
index 00000000..4c956c42
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/errors.py
@@ -0,0 +1,45 @@
+class ParserError(Exception):
+    pass
+
+
+class ParserException(ParserError):
+    def __init__(self, message, location):
+        self.location = location
+        super(ParserException, self).__init__('(' + str(location['line']) + ':' +
+                                              str(location['column'] if 'column' in
+                                                  location else 0) + '): ' + message)
+
+
+class NoSuchLanguageException(ParserException):
+    def __init__(self, language, location):
+        super(NoSuchLanguageException, self).__init__('Language not supported: ' + language,
+                                                      location)
+
+
+class AstBuilderException(ParserException):
+    pass
+
+
+class UnexpectedEOFException(ParserException):
+    def __init__(self, received_token, expected_token_types, state_comment):
+        message = 'unexpected end of file, expected: ' + ', '.join(expected_token_types)
+        super(UnexpectedEOFException, self).__init__(message, received_token.location)
+
+
+class UnexpectedTokenException(ParserException):
+    def __init__(self, received_token, expected_token_types, state_comment):
+        message = ("expected: " + ', '.join(expected_token_types) + ", got '" +
+                   received_token.token_value().strip() + "'")
+        column = received_token.location['column'] if 'column' in received_token.location else None
+        location = (received_token.location if column else
+                    {'line': received_token.location['line'],
+                     'column': received_token.line.indent + 1})
+        super(UnexpectedTokenException, self).__init__(message, location)
+
+
+class CompositeParserException(ParserError):
+    def __init__(self, errors):
+        self.errors = errors
+        super(CompositeParserException, self).__init__("Parser errors:\n" +
+                                                       '\n'.join([error.args[0] for error in
+                                                                  errors]))
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/gherkin-languages.json b/backend/venv/lib/python3.12/site-packages/gherkin/gherkin-languages.json
new file mode 100644
index 00000000..3953e610
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/gherkin-languages.json
@@ -0,0 +1,3831 @@
+{
+  "af": {
+    "and": [
+      "* ",
+      "En "
+    ],
+    "background": [
+      "Agtergrond"
+    ],
+    "but": [
+      "* ",
+      "Maar "
+    ],
+    "examples": [
+      "Voorbeelde"
+    ],
+    "feature": [
+      "Funksie",
+      "Besigheid Behoefte",
+      "Vermoë"
+    ],
+    "given": [
+      "* ",
+      "Gegewe "
+    ],
+    "name": "Afrikaans",
+    "native": "Afrikaans",
+    "rule": [
+      "Regel"
+    ],
+    "scenario": [
+      "Voorbeeld",
+      "Situasie"
+    ],
+    "scenarioOutline": [
+      "Situasie Uiteensetting"
+    ],
+    "then": [
+      "* ",
+      "Dan "
+    ],
+    "when": [
+      "* ",
+      "Wanneer "
+    ]
+  },
+  "am": {
+    "and": [
+      "* ",
+      "Եվ "
+    ],
+    "background": [
+      "Կոնտեքստ"
+    ],
+    "but": [
+      "* ",
+      "Բայց "
+    ],
+    "examples": [
+      "Օրինակներ"
+    ],
+    "feature": [
+      "Ֆունկցիոնալություն",
+      "Հատկություն"
+    ],
+    "given": [
+      "* ",
+      "Դիցուք "
+    ],
+    "name": "Armenian",
+    "native": "հայերեն",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Օրինակ",
+      "Սցենար"
+    ],
+    "scenarioOutline": [
+      "Սցենարի կառուցվացքը"
+    ],
+    "then": [
+      "* ",
+      "Ապա "
+    ],
+    "when": [
+      "* ",
+      "Եթե ",
+      "Երբ "
+    ]
+  },
+  "an": {
+    "and": [
+      "* ",
+      "Y ",
+      "E "
+    ],
+    "background": [
+      "Antecedents"
+    ],
+    "but": [
+      "* ",
+      "Pero "
+    ],
+    "examples": [
+      "Eixemplos"
+    ],
+    "feature": [
+      "Caracteristica"
+    ],
+    "given": [
+      "* ",
+      "Dau ",
+      "Dada ",
+      "Daus ",
+      "Dadas "
+    ],
+    "name": "Aragonese",
+    "native": "Aragonés",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Eixemplo",
+      "Caso"
+    ],
+    "scenarioOutline": [
+      "Esquema del caso"
+    ],
+    "then": [
+      "* ",
+      "Alavez ",
+      "Allora ",
+      "Antonces "
+    ],
+    "when": [
+      "* ",
+      "Cuan "
+    ]
+  },
+  "ar": {
+    "and": [
+      "* ",
+      "و "
+    ],
+    "background": [
+      "الخلفية"
+    ],
+    "but": [
+      "* ",
+      "لكن "
+    ],
+    "examples": [
+      "امثلة"
+    ],
+    "feature": [
+      "خاصية"
+    ],
+    "given": [
+      "* ",
+      "بفرض "
+    ],
+    "name": "Arabic",
+    "native": "العربية",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "مثال",
+      "سيناريو"
+    ],
+    "scenarioOutline": [
+      "سيناريو مخطط"
+    ],
+    "then": [
+      "* ",
+      "اذاً ",
+      "ثم "
+    ],
+    "when": [
+      "* ",
+      "متى ",
+      "عندما "
+    ]
+  },
+  "ast": {
+    "and": [
+      "* ",
+      "Y ",
+      "Ya "
+    ],
+    "background": [
+      "Antecedentes"
+    ],
+    "but": [
+      "* ",
+      "Peru "
+    ],
+    "examples": [
+      "Exemplos"
+    ],
+    "feature": [
+      "Carauterística"
+    ],
+    "given": [
+      "* ",
+      "Dáu ",
+      "Dada ",
+      "Daos ",
+      "Daes "
+    ],
+    "name": "Asturian",
+    "native": "asturianu",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Exemplo",
+      "Casu"
+    ],
+    "scenarioOutline": [
+      "Esbozu del casu"
+    ],
+    "then": [
+      "* ",
+      "Entós "
+    ],
+    "when": [
+      "* ",
+      "Cuando "
+    ]
+  },
+  "az": {
+    "and": [
+      "* ",
+      "Və ",
+      "Həm "
+    ],
+    "background": [
+      "Keçmiş",
+      "Kontekst"
+    ],
+    "but": [
+      "* ",
+      "Amma ",
+      "Ancaq "
+    ],
+    "examples": [
+      "Nümunələr"
+    ],
+    "feature": [
+      "Özəllik"
+    ],
+    "given": [
+      "* ",
+      "Tutaq ki ",
+      "Verilir "
+    ],
+    "name": "Azerbaijani",
+    "native": "Azərbaycanca",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Nümunə",
+      "Ssenari"
+    ],
+    "scenarioOutline": [
+      "Ssenarinin strukturu"
+    ],
+    "then": [
+      "* ",
+      "O halda "
+    ],
+    "when": [
+      "* ",
+      "Əgər ",
+      "Nə vaxt ki "
+    ]
+  },
+  "be": {
+    "and": [
+      "* ",
+      "I ",
+      "Ды ",
+      "Таксама "
+    ],
+    "background": [
+      "Кантэкст"
+    ],
+    "but": [
+      "* ",
+      "Але ",
+      "Інакш "
+    ],
+    "examples": [
+      "Прыклады"
+    ],
+    "feature": [
+      "Функцыянальнасць",
+      "Фіча"
+    ],
+    "given": [
+      "* ",
+      "Няхай ",
+      "Дадзена "
+    ],
+    "name": "Belarusian",
+    "native": "Беларуская",
+    "rule": [
+      "Правілы"
+    ],
+    "scenario": [
+      "Сцэнарый",
+      "Cцэнар"
+    ],
+    "scenarioOutline": [
+      "Шаблон сцэнарыя",
+      "Узор сцэнара"
+    ],
+    "then": [
+      "* ",
+      "Тады "
+    ],
+    "when": [
+      "* ",
+      "Калі "
+    ]
+  },
+  "bg": {
+    "and": [
+      "* ",
+      "И "
+    ],
+    "background": [
+      "Предистория"
+    ],
+    "but": [
+      "* ",
+      "Но "
+    ],
+    "examples": [
+      "Примери"
+    ],
+    "feature": [
+      "Функционалност"
+    ],
+    "given": [
+      "* ",
+      "Дадено "
+    ],
+    "name": "Bulgarian",
+    "native": "български",
+    "rule": [
+      "Правило"
+    ],
+    "scenario": [
+      "Пример",
+      "Сценарий"
+    ],
+    "scenarioOutline": [
+      "Рамка на сценарий"
+    ],
+    "then": [
+      "* ",
+      "То "
+    ],
+    "when": [
+      "* ",
+      "Когато "
+    ]
+  },
+  "bm": {
+    "and": [
+      "* ",
+      "Dan "
+    ],
+    "background": [
+      "Latar Belakang"
+    ],
+    "but": [
+      "* ",
+      "Tetapi ",
+      "Tapi "
+    ],
+    "examples": [
+      "Contoh"
+    ],
+    "feature": [
+      "Fungsi"
+    ],
+    "given": [
+      "* ",
+      "Diberi ",
+      "Bagi "
+    ],
+    "name": "Malay",
+    "native": "Bahasa Melayu",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Senario",
+      "Situasi",
+      "Keadaan"
+    ],
+    "scenarioOutline": [
+      "Kerangka Senario",
+      "Kerangka Situasi",
+      "Kerangka Keadaan",
+      "Garis Panduan Senario"
+    ],
+    "then": [
+      "* ",
+      "Maka ",
+      "Kemudian "
+    ],
+    "when": [
+      "* ",
+      "Apabila "
+    ]
+  },
+  "bs": {
+    "and": [
+      "* ",
+      "I ",
+      "A "
+    ],
+    "background": [
+      "Pozadina"
+    ],
+    "but": [
+      "* ",
+      "Ali "
+    ],
+    "examples": [
+      "Primjeri"
+    ],
+    "feature": [
+      "Karakteristika"
+    ],
+    "given": [
+      "* ",
+      "Dato "
+    ],
+    "name": "Bosnian",
+    "native": "Bosanski",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Primjer",
+      "Scenariju",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Scenariju-obris",
+      "Scenario-outline"
+    ],
+    "then": [
+      "* ",
+      "Zatim "
+    ],
+    "when": [
+      "* ",
+      "Kada "
+    ]
+  },
+  "ca": {
+    "and": [
+      "* ",
+      "I "
+    ],
+    "background": [
+      "Rerefons",
+      "Antecedents"
+    ],
+    "but": [
+      "* ",
+      "Però "
+    ],
+    "examples": [
+      "Exemples"
+    ],
+    "feature": [
+      "Característica",
+      "Funcionalitat"
+    ],
+    "given": [
+      "* ",
+      "Donat ",
+      "Donada ",
+      "Atès ",
+      "Atesa "
+    ],
+    "name": "Catalan",
+    "native": "català",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Exemple",
+      "Escenari"
+    ],
+    "scenarioOutline": [
+      "Esquema de l'escenari"
+    ],
+    "then": [
+      "* ",
+      "Aleshores ",
+      "Cal "
+    ],
+    "when": [
+      "* ",
+      "Quan "
+    ]
+  },
+  "cs": {
+    "and": [
+      "* ",
+      "A také ",
+      "A "
+    ],
+    "background": [
+      "Pozadí",
+      "Kontext"
+    ],
+    "but": [
+      "* ",
+      "Ale "
+    ],
+    "examples": [
+      "Příklady"
+    ],
+    "feature": [
+      "Požadavek"
+    ],
+    "given": [
+      "* ",
+      "Pokud ",
+      "Za předpokladu "
+    ],
+    "name": "Czech",
+    "native": "Česky",
+    "rule": [
+      "Pravidlo"
+    ],
+    "scenario": [
+      "Příklad",
+      "Scénář"
+    ],
+    "scenarioOutline": [
+      "Náčrt Scénáře",
+      "Osnova scénáře"
+    ],
+    "then": [
+      "* ",
+      "Pak "
+    ],
+    "when": [
+      "* ",
+      "Když "
+    ]
+  },
+  "cy-GB": {
+    "and": [
+      "* ",
+      "A "
+    ],
+    "background": [
+      "Cefndir"
+    ],
+    "but": [
+      "* ",
+      "Ond "
+    ],
+    "examples": [
+      "Enghreifftiau"
+    ],
+    "feature": [
+      "Arwedd"
+    ],
+    "given": [
+      "* ",
+      "Anrhegedig a "
+    ],
+    "name": "Welsh",
+    "native": "Cymraeg",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Enghraifft",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Scenario Amlinellol"
+    ],
+    "then": [
+      "* ",
+      "Yna "
+    ],
+    "when": [
+      "* ",
+      "Pryd "
+    ]
+  },
+  "da": {
+    "and": [
+      "* ",
+      "Og "
+    ],
+    "background": [
+      "Baggrund"
+    ],
+    "but": [
+      "* ",
+      "Men "
+    ],
+    "examples": [
+      "Eksempler"
+    ],
+    "feature": [
+      "Egenskab"
+    ],
+    "given": [
+      "* ",
+      "Givet "
+    ],
+    "name": "Danish",
+    "native": "dansk",
+    "rule": [
+      "Regel"
+    ],
+    "scenario": [
+      "Eksempel",
+      "Scenarie"
+    ],
+    "scenarioOutline": [
+      "Abstrakt Scenario"
+    ],
+    "then": [
+      "* ",
+      "Så "
+    ],
+    "when": [
+      "* ",
+      "Når "
+    ]
+  },
+  "de": {
+    "and": [
+      "* ",
+      "Und "
+    ],
+    "background": [
+      "Grundlage",
+      "Hintergrund",
+      "Voraussetzungen",
+      "Vorbedingungen"
+    ],
+    "but": [
+      "* ",
+      "Aber "
+    ],
+    "examples": [
+      "Beispiele"
+    ],
+    "feature": [
+      "Funktionalität",
+      "Funktion"
+    ],
+    "given": [
+      "* ",
+      "Angenommen ",
+      "Gegeben sei ",
+      "Gegeben seien "
+    ],
+    "name": "German",
+    "native": "Deutsch",
+    "rule": [
+      "Rule",
+      "Regel"
+    ],
+    "scenario": [
+      "Beispiel",
+      "Szenario"
+    ],
+    "scenarioOutline": [
+      "Szenariogrundriss",
+      "Szenarien"
+    ],
+    "then": [
+      "* ",
+      "Dann "
+    ],
+    "when": [
+      "* ",
+      "Wenn "
+    ]
+  },
+  "el": {
+    "and": [
+      "* ",
+      "Και "
+    ],
+    "background": [
+      "Υπόβαθρο"
+    ],
+    "but": [
+      "* ",
+      "Αλλά "
+    ],
+    "examples": [
+      "Παραδείγματα",
+      "Σενάρια"
+    ],
+    "feature": [
+      "Δυνατότητα",
+      "Λειτουργία"
+    ],
+    "given": [
+      "* ",
+      "Δεδομένου "
+    ],
+    "name": "Greek",
+    "native": "Ελληνικά",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Παράδειγμα",
+      "Σενάριο"
+    ],
+    "scenarioOutline": [
+      "Περιγραφή Σεναρίου",
+      "Περίγραμμα Σεναρίου"
+    ],
+    "then": [
+      "* ",
+      "Τότε "
+    ],
+    "when": [
+      "* ",
+      "Όταν "
+    ]
+  },
+  "em": {
+    "and": [
+      "* ",
+      "😂"
+    ],
+    "background": [
+      "💤"
+    ],
+    "but": [
+      "* ",
+      "😔"
+    ],
+    "examples": [
+      "📓"
+    ],
+    "feature": [
+      "📚"
+    ],
+    "given": [
+      "* ",
+      "😐"
+    ],
+    "name": "Emoji",
+    "native": "😀",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "🥒",
+      "📕"
+    ],
+    "scenarioOutline": [
+      "📖"
+    ],
+    "then": [
+      "* ",
+      "🙏"
+    ],
+    "when": [
+      "* ",
+      "🎬"
+    ]
+  },
+  "en": {
+    "and": [
+      "* ",
+      "And "
+    ],
+    "background": [
+      "Background"
+    ],
+    "but": [
+      "* ",
+      "But "
+    ],
+    "examples": [
+      "Examples",
+      "Scenarios"
+    ],
+    "feature": [
+      "Feature",
+      "Business Need",
+      "Ability"
+    ],
+    "given": [
+      "* ",
+      "Given "
+    ],
+    "name": "English",
+    "native": "English",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Example",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Scenario Outline",
+      "Scenario Template"
+    ],
+    "then": [
+      "* ",
+      "Then "
+    ],
+    "when": [
+      "* ",
+      "When "
+    ]
+  },
+  "en-Scouse": {
+    "and": [
+      "* ",
+      "An "
+    ],
+    "background": [
+      "Dis is what went down"
+    ],
+    "but": [
+      "* ",
+      "Buh "
+    ],
+    "examples": [
+      "Examples"
+    ],
+    "feature": [
+      "Feature"
+    ],
+    "given": [
+      "* ",
+      "Givun ",
+      "Youse know when youse got "
+    ],
+    "name": "Scouse",
+    "native": "Scouse",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "The thing of it is"
+    ],
+    "scenarioOutline": [
+      "Wharrimean is"
+    ],
+    "then": [
+      "* ",
+      "Dun ",
+      "Den youse gotta "
+    ],
+    "when": [
+      "* ",
+      "Wun ",
+      "Youse know like when "
+    ]
+  },
+  "en-au": {
+    "and": [
+      "* ",
+      "Too right "
+    ],
+    "background": [
+      "First off"
+    ],
+    "but": [
+      "* ",
+      "Yeah nah "
+    ],
+    "examples": [
+      "You'll wanna"
+    ],
+    "feature": [
+      "Pretty much"
+    ],
+    "given": [
+      "* ",
+      "Y'know "
+    ],
+    "name": "Australian",
+    "native": "Australian",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Awww, look mate"
+    ],
+    "scenarioOutline": [
+      "Reckon it's like"
+    ],
+    "then": [
+      "* ",
+      "But at the end of the day I reckon "
+    ],
+    "when": [
+      "* ",
+      "It's just unbelievable "
+    ]
+  },
+  "en-lol": {
+    "and": [
+      "* ",
+      "AN "
+    ],
+    "background": [
+      "B4"
+    ],
+    "but": [
+      "* ",
+      "BUT "
+    ],
+    "examples": [
+      "EXAMPLZ"
+    ],
+    "feature": [
+      "OH HAI"
+    ],
+    "given": [
+      "* ",
+      "I CAN HAZ "
+    ],
+    "name": "LOLCAT",
+    "native": "LOLCAT",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "MISHUN"
+    ],
+    "scenarioOutline": [
+      "MISHUN SRSLY"
+    ],
+    "then": [
+      "* ",
+      "DEN "
+    ],
+    "when": [
+      "* ",
+      "WEN "
+    ]
+  },
+  "en-old": {
+    "and": [
+      "* ",
+      "Ond ",
+      "7 "
+    ],
+    "background": [
+      "Aer",
+      "Ær"
+    ],
+    "but": [
+      "* ",
+      "Ac "
+    ],
+    "examples": [
+      "Se the",
+      "Se þe",
+      "Se ðe"
+    ],
+    "feature": [
+      "Hwaet",
+      "Hwæt"
+    ],
+    "given": [
+      "* ",
+      "Thurh ",
+      "Þurh ",
+      "Ðurh "
+    ],
+    "name": "Old English",
+    "native": "Englisc",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Swa"
+    ],
+    "scenarioOutline": [
+      "Swa hwaer swa",
+      "Swa hwær swa"
+    ],
+    "then": [
+      "* ",
+      "Tha ",
+      "Þa ",
+      "Ða ",
+      "Tha the ",
+      "Þa þe ",
+      "Ða ðe "
+    ],
+    "when": [
+      "* ",
+      "Bæþsealf ",
+      "Bæþsealfa ",
+      "Bæþsealfe ",
+      "Ciricæw ",
+      "Ciricæwe ",
+      "Ciricæwa "
+    ]
+  },
+  "en-pirate": {
+    "and": [
+      "* ",
+      "Aye "
+    ],
+    "background": [
+      "Yo-ho-ho"
+    ],
+    "but": [
+      "* ",
+      "Avast! "
+    ],
+    "examples": [
+      "Dead men tell no tales"
+    ],
+    "feature": [
+      "Ahoy matey!"
+    ],
+    "given": [
+      "* ",
+      "Gangway! "
+    ],
+    "name": "Pirate",
+    "native": "Pirate",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Heave to"
+    ],
+    "scenarioOutline": [
+      "Shiver me timbers"
+    ],
+    "then": [
+      "* ",
+      "Let go and haul "
+    ],
+    "when": [
+      "* ",
+      "Blimey! "
+    ]
+  },
+    "en-tx": {
+    "and": [
+      "Come hell or high water "
+    ],
+    "background": [
+      "Lemme tell y'all a story"
+    ],
+    "but": [
+      "Well now hold on, I'll you what "
+    ],
+    "examples": [
+      "Now that's a story longer than a cattle drive in July"
+    ],
+    "feature": [
+      "This ain’t my first rodeo",
+      "All gussied up"
+    ],
+    "given": [
+      "Fixin' to ",
+      "All git out "
+    ],
+    "name": "Texas",
+    "native": "Texas",
+    "rule": [
+      "Rule "
+    ],
+    "scenario": [
+      "All hat and no cattle"
+    ],
+    "scenarioOutline": [
+      "Serious as a snake bite",
+      "Busy as a hound in flea season"
+    ],
+    "then": [
+      "There’s no tree but bears some fruit "
+    ],
+    "when": [
+      "Quick out of the chute "
+    ]
+  },
+  "eo": {
+    "and": [
+      "* ",
+      "Kaj "
+    ],
+    "background": [
+      "Fono"
+    ],
+    "but": [
+      "* ",
+      "Sed "
+    ],
+    "examples": [
+      "Ekzemploj"
+    ],
+    "feature": [
+      "Trajto"
+    ],
+    "given": [
+      "* ",
+      "Donitaĵo ",
+      "Komence "
+    ],
+    "name": "Esperanto",
+    "native": "Esperanto",
+    "rule": [
+      "Regulo"
+    ],
+    "scenario": [
+      "Ekzemplo",
+      "Scenaro",
+      "Kazo"
+    ],
+    "scenarioOutline": [
+      "Konturo de la scenaro",
+      "Skizo",
+      "Kazo-skizo"
+    ],
+    "then": [
+      "* ",
+      "Do "
+    ],
+    "when": [
+      "* ",
+      "Se "
+    ]
+  },
+  "es": {
+    "and": [
+      "* ",
+      "Y ",
+      "E "
+    ],
+    "background": [
+      "Antecedentes"
+    ],
+    "but": [
+      "* ",
+      "Pero "
+    ],
+    "examples": [
+      "Ejemplos"
+    ],
+    "feature": [
+      "Característica",
+      "Necesidad del negocio",
+      "Requisito"
+    ],
+    "given": [
+      "* ",
+      "Dado ",
+      "Dada ",
+      "Dados ",
+      "Dadas "
+    ],
+    "name": "Spanish",
+    "native": "español",
+    "rule": [
+      "Regla",
+      "Regla de negocio"
+    ],
+    "scenario": [
+      "Ejemplo",
+      "Escenario"
+    ],
+    "scenarioOutline": [
+      "Esquema del escenario"
+    ],
+    "then": [
+      "* ",
+      "Entonces "
+    ],
+    "when": [
+      "* ",
+      "Cuando "
+    ]
+  },
+  "et": {
+    "and": [
+      "* ",
+      "Ja "
+    ],
+    "background": [
+      "Taust"
+    ],
+    "but": [
+      "* ",
+      "Kuid "
+    ],
+    "examples": [
+      "Juhtumid"
+    ],
+    "feature": [
+      "Omadus"
+    ],
+    "given": [
+      "* ",
+      "Eeldades "
+    ],
+    "name": "Estonian",
+    "native": "eesti keel",
+    "rule": [
+      "Reegel"
+    ],
+    "scenario": [
+      "Juhtum",
+      "Stsenaarium"
+    ],
+    "scenarioOutline": [
+      "Raamjuhtum",
+      "Raamstsenaarium"
+    ],
+    "then": [
+      "* ",
+      "Siis "
+    ],
+    "when": [
+      "* ",
+      "Kui "
+    ]
+  },
+  "fa": {
+    "and": [
+      "* ",
+      "و "
+    ],
+    "background": [
+      "زمینه"
+    ],
+    "but": [
+      "* ",
+      "اما "
+    ],
+    "examples": [
+      "نمونه ها"
+    ],
+    "feature": [
+      "وِیژگی"
+    ],
+    "given": [
+      "* ",
+      "با فرض "
+    ],
+    "name": "Persian",
+    "native": "فارسی",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "مثال",
+      "سناریو"
+    ],
+    "scenarioOutline": [
+      "الگوی سناریو"
+    ],
+    "then": [
+      "* ",
+      "آنگاه "
+    ],
+    "when": [
+      "* ",
+      "هنگامی "
+    ]
+  },
+  "fi": {
+    "and": [
+      "* ",
+      "Ja "
+    ],
+    "background": [
+      "Tausta"
+    ],
+    "but": [
+      "* ",
+      "Mutta "
+    ],
+    "examples": [
+      "Tapaukset"
+    ],
+    "feature": [
+      "Ominaisuus"
+    ],
+    "given": [
+      "* ",
+      "Oletetaan "
+    ],
+    "name": "Finnish",
+    "native": "suomi",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Tapaus"
+    ],
+    "scenarioOutline": [
+      "Tapausaihio"
+    ],
+    "then": [
+      "* ",
+      "Niin "
+    ],
+    "when": [
+      "* ",
+      "Kun "
+    ]
+  },
+  "fr": {
+    "and": [
+      "* ",
+      "Et que ",
+      "Et qu'",
+      "Et "
+    ],
+    "background": [
+      "Contexte"
+    ],
+    "but": [
+      "* ",
+      "Mais que ",
+      "Mais qu'",
+      "Mais "
+    ],
+    "examples": [
+      "Exemples"
+    ],
+    "feature": [
+      "Fonctionnalité"
+    ],
+    "given": [
+      "* ",
+      "Soit ",
+      "Sachant que ",
+      "Sachant qu'",
+      "Sachant ",
+      "Etant donné que ",
+      "Etant donné qu'",
+      "Etant donné ",
+      "Etant donnée ",
+      "Etant donnés ",
+      "Etant données ",
+      "Étant donné que ",
+      "Étant donné qu'",
+      "Étant donné ",
+      "Étant donnée ",
+      "Étant donnés ",
+      "Étant données "
+    ],
+    "name": "French",
+    "native": "français",
+    "rule": [
+      "Règle"
+    ],
+    "scenario": [
+      "Exemple",
+      "Scénario"
+    ],
+    "scenarioOutline": [
+      "Plan du scénario",
+      "Plan du Scénario"
+    ],
+    "then": [
+      "* ",
+      "Alors ",
+      "Donc "
+    ],
+    "when": [
+      "* ",
+      "Quand ",
+      "Lorsque ",
+      "Lorsqu'"
+    ]
+  },
+  "ga": {
+    "and": [
+      "* ",
+      "Agus "
+    ],
+    "background": [
+      "Cúlra"
+    ],
+    "but": [
+      "* ",
+      "Ach "
+    ],
+    "examples": [
+      "Samplaí"
+    ],
+    "feature": [
+      "Gné"
+    ],
+    "given": [
+      "* ",
+      "Cuir i gcás go ",
+      "Cuir i gcás nach ",
+      "Cuir i gcás gur ",
+      "Cuir i gcás nár "
+    ],
+    "name": "Irish",
+    "native": "Gaeilge",
+    "rule": [
+      "Riail"
+    ],
+    "scenario": [
+      "Sampla",
+      "Cás"
+    ],
+    "scenarioOutline": [
+      "Cás Achomair"
+    ],
+    "then": [
+      "* ",
+      "Ansin "
+    ],
+    "when": [
+      "* ",
+      "Nuair a ",
+      "Nuair nach ",
+      "Nuair ba ",
+      "Nuair nár "
+    ]
+  },
+  "gj": {
+    "and": [
+      "* ",
+      "અને "
+    ],
+    "background": [
+      "બેકગ્રાઉન્ડ"
+    ],
+    "but": [
+      "* ",
+      "પણ "
+    ],
+    "examples": [
+      "ઉદાહરણો"
+    ],
+    "feature": [
+      "લક્ષણ",
+      "વ્યાપાર જરૂર",
+      "ક્ષમતા"
+    ],
+    "given": [
+      "* ",
+      "આપેલ છે "
+    ],
+    "name": "Gujarati",
+    "native": "ગુજરાતી",
+    "rule": [
+      "નિયમ"
+    ],
+    "scenario": [
+      "ઉદાહરણ",
+      "સ્થિતિ"
+    ],
+    "scenarioOutline": [
+      "પરિદ્દશ્ય રૂપરેખા",
+      "પરિદ્દશ્ય ઢાંચો"
+    ],
+    "then": [
+      "* ",
+      "પછી "
+    ],
+    "when": [
+      "* ",
+      "ક્યારે "
+    ]
+  },
+  "gl": {
+    "and": [
+      "* ",
+      "E "
+    ],
+    "background": [
+      "Contexto"
+    ],
+    "but": [
+      "* ",
+      "Mais ",
+      "Pero "
+    ],
+    "examples": [
+      "Exemplos"
+    ],
+    "feature": [
+      "Característica"
+    ],
+    "given": [
+      "* ",
+      "Dado ",
+      "Dada ",
+      "Dados ",
+      "Dadas "
+    ],
+    "name": "Galician",
+    "native": "galego",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Exemplo",
+      "Escenario"
+    ],
+    "scenarioOutline": [
+      "Esbozo do escenario"
+    ],
+    "then": [
+      "* ",
+      "Entón ",
+      "Logo "
+    ],
+    "when": [
+      "* ",
+      "Cando "
+    ]
+  },
+  "he": {
+    "and": [
+      "* ",
+      "וגם "
+    ],
+    "background": [
+      "רקע"
+    ],
+    "but": [
+      "* ",
+      "אבל "
+    ],
+    "examples": [
+      "דוגמאות"
+    ],
+    "feature": [
+      "תכונה"
+    ],
+    "given": [
+      "* ",
+      "בהינתן "
+    ],
+    "name": "Hebrew",
+    "native": "עברית",
+    "rule": [
+      "כלל"
+    ],
+    "scenario": [
+      "דוגמא",
+      "תרחיש"
+    ],
+    "scenarioOutline": [
+      "תבנית תרחיש"
+    ],
+    "then": [
+      "* ",
+      "אז ",
+      "אזי "
+    ],
+    "when": [
+      "* ",
+      "כאשר "
+    ]
+  },
+  "hi": {
+    "and": [
+      "* ",
+      "और ",
+      "तथा "
+    ],
+    "background": [
+      "पृष्ठभूमि"
+    ],
+    "but": [
+      "* ",
+      "पर ",
+      "परन्तु ",
+      "किन्तु "
+    ],
+    "examples": [
+      "उदाहरण"
+    ],
+    "feature": [
+      "रूप लेख"
+    ],
+    "given": [
+      "* ",
+      "अगर ",
+      "यदि ",
+      "चूंकि "
+    ],
+    "name": "Hindi",
+    "native": "हिंदी",
+    "rule": [
+      "नियम"
+    ],
+    "scenario": [
+      "परिदृश्य"
+    ],
+    "scenarioOutline": [
+      "परिदृश्य रूपरेखा"
+    ],
+    "then": [
+      "* ",
+      "तब ",
+      "तदा "
+    ],
+    "when": [
+      "* ",
+      "जब ",
+      "कदा "
+    ]
+  },
+  "hr": {
+    "and": [
+      "* ",
+      "I "
+    ],
+    "background": [
+      "Pozadina"
+    ],
+    "but": [
+      "* ",
+      "Ali "
+    ],
+    "examples": [
+      "Primjeri",
+      "Scenariji"
+    ],
+    "feature": [
+      "Osobina",
+      "Mogućnost",
+      "Mogucnost"
+    ],
+    "given": [
+      "* ",
+      "Zadan ",
+      "Zadani ",
+      "Zadano ",
+      "Ukoliko "
+    ],
+    "name": "Croatian",
+    "native": "hrvatski",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Primjer",
+      "Scenarij"
+    ],
+    "scenarioOutline": [
+      "Skica",
+      "Koncept"
+    ],
+    "then": [
+      "* ",
+      "Onda "
+    ],
+    "when": [
+      "* ",
+      "Kada ",
+      "Kad "
+    ]
+  },
+  "ht": {
+    "and": [
+      "* ",
+      "Ak ",
+      "Epi ",
+      "E "
+    ],
+    "background": [
+      "Kontèks",
+      "Istorik"
+    ],
+    "but": [
+      "* ",
+      "Men "
+    ],
+    "examples": [
+      "Egzanp"
+    ],
+    "feature": [
+      "Karakteristik",
+      "Mak",
+      "Fonksyonalite"
+    ],
+    "given": [
+      "* ",
+      "Sipoze ",
+      "Sipoze ke ",
+      "Sipoze Ke "
+    ],
+    "name": "Creole",
+    "native": "kreyòl",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Senaryo"
+    ],
+    "scenarioOutline": [
+      "Plan senaryo",
+      "Plan Senaryo",
+      "Senaryo deskripsyon",
+      "Senaryo Deskripsyon",
+      "Dyagram senaryo",
+      "Dyagram Senaryo"
+    ],
+    "then": [
+      "* ",
+      "Lè sa a ",
+      "Le sa a "
+    ],
+    "when": [
+      "* ",
+      "Lè ",
+      "Le "
+    ]
+  },
+  "hu": {
+    "and": [
+      "* ",
+      "És "
+    ],
+    "background": [
+      "Háttér"
+    ],
+    "but": [
+      "* ",
+      "De "
+    ],
+    "examples": [
+      "Példák"
+    ],
+    "feature": [
+      "Jellemző"
+    ],
+    "given": [
+      "* ",
+      "Amennyiben ",
+      "Adott "
+    ],
+    "name": "Hungarian",
+    "native": "magyar",
+    "rule": [
+      "Szabály"
+    ],
+    "scenario": [
+      "Példa",
+      "Forgatókönyv"
+    ],
+    "scenarioOutline": [
+      "Forgatókönyv vázlat"
+    ],
+    "then": [
+      "* ",
+      "Akkor "
+    ],
+    "when": [
+      "* ",
+      "Majd ",
+      "Ha ",
+      "Amikor "
+    ]
+  },
+  "id": {
+    "and": [
+      "* ",
+      "Dan "
+    ],
+    "background": [
+      "Dasar",
+      "Latar Belakang"
+    ],
+    "but": [
+      "* ",
+      "Tapi ",
+      "Tetapi "
+    ],
+    "examples": [
+      "Contoh",
+      "Misal"
+    ],
+    "feature": [
+      "Fitur"
+    ],
+    "given": [
+      "* ",
+      "Dengan ",
+      "Diketahui ",
+      "Diasumsikan ",
+      "Bila ",
+      "Jika "
+    ],
+    "name": "Indonesian",
+    "native": "Bahasa Indonesia",
+    "rule": [
+      "Rule",
+      "Aturan"
+    ],
+    "scenario": [
+      "Skenario"
+    ],
+    "scenarioOutline": [
+      "Skenario konsep",
+      "Garis-Besar Skenario"
+    ],
+    "then": [
+      "* ",
+      "Maka ",
+      "Kemudian "
+    ],
+    "when": [
+      "* ",
+      "Ketika "
+    ]
+  },
+  "is": {
+    "and": [
+      "* ",
+      "Og "
+    ],
+    "background": [
+      "Bakgrunnur"
+    ],
+    "but": [
+      "* ",
+      "En "
+    ],
+    "examples": [
+      "Dæmi",
+      "Atburðarásir"
+    ],
+    "feature": [
+      "Eiginleiki"
+    ],
+    "given": [
+      "* ",
+      "Ef "
+    ],
+    "name": "Icelandic",
+    "native": "Íslenska",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Atburðarás"
+    ],
+    "scenarioOutline": [
+      "Lýsing Atburðarásar",
+      "Lýsing Dæma"
+    ],
+    "then": [
+      "* ",
+      "Þá "
+    ],
+    "when": [
+      "* ",
+      "Þegar "
+    ]
+  },
+  "it": {
+    "and": [
+      "* ",
+      "E ",
+      "Ed "
+    ],
+    "background": [
+      "Contesto"
+    ],
+    "but": [
+      "* ",
+      "Ma "
+    ],
+    "examples": [
+      "Esempi"
+    ],
+    "feature": [
+      "Funzionalità",
+      "Esigenza di Business",
+      "Abilità"
+    ],
+    "given": [
+      "* ",
+      "Dato ",
+      "Data ",
+      "Dati ",
+      "Date "
+    ],
+    "name": "Italian",
+    "native": "italiano",
+    "rule": [
+      "Regola"
+    ],
+    "scenario": [
+      "Esempio",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Schema dello scenario"
+    ],
+    "then": [
+      "* ",
+      "Allora "
+    ],
+    "when": [
+      "* ",
+      "Quando "
+    ]
+  },
+  "ja": {
+    "and": [
+      "* ",
+      "且つ",
+      "かつ"
+    ],
+    "background": [
+      "背景"
+    ],
+    "but": [
+      "* ",
+      "然し",
+      "しかし",
+      "但し",
+      "ただし"
+    ],
+    "examples": [
+      "例",
+      "サンプル"
+    ],
+    "feature": [
+      "フィーチャ",
+      "機能"
+    ],
+    "given": [
+      "* ",
+      "前提"
+    ],
+    "name": "Japanese",
+    "native": "日本語",
+    "rule": [
+      "ルール"
+    ],
+    "scenario": [
+      "シナリオ"
+    ],
+    "scenarioOutline": [
+      "シナリオアウトライン",
+      "シナリオテンプレート",
+      "テンプレ",
+      "シナリオテンプレ"
+    ],
+    "then": [
+      "* ",
+      "ならば"
+    ],
+    "when": [
+      "* ",
+      "もし"
+    ]
+  },
+  "jv": {
+    "and": [
+      "* ",
+      "Lan "
+    ],
+    "background": [
+      "Dasar"
+    ],
+    "but": [
+      "* ",
+      "Tapi ",
+      "Nanging ",
+      "Ananging "
+    ],
+    "examples": [
+      "Conto",
+      "Contone"
+    ],
+    "feature": [
+      "Fitur"
+    ],
+    "given": [
+      "* ",
+      "Nalika ",
+      "Nalikaning "
+    ],
+    "name": "Javanese",
+    "native": "Basa Jawa",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Skenario"
+    ],
+    "scenarioOutline": [
+      "Konsep skenario"
+    ],
+    "then": [
+      "* ",
+      "Njuk ",
+      "Banjur "
+    ],
+    "when": [
+      "* ",
+      "Manawa ",
+      "Menawa "
+    ]
+  },
+  "ka": {
+    "and": [
+      "* ",
+      "და ",
+      "ასევე "
+    ],
+    "background": [
+      "კონტექსტი"
+    ],
+    "but": [
+      "* ",
+      "მაგრამ ",
+      "თუმცა "
+    ],
+    "examples": [
+      "მაგალითები"
+    ],
+    "feature": [
+      "თვისება",
+      "მოთხოვნა"
+    ],
+    "given": [
+      "* ",
+      "მოცემული ",
+      "მოცემულია ",
+      "ვთქვათ "
+    ],
+    "name": "Georgian",
+    "native": "ქართული",
+    "rule": [
+      "წესი"
+    ],
+    "scenario": [
+      "მაგალითად",
+      "მაგალითი",
+      "მაგ",
+      "სცენარი"
+    ],
+    "scenarioOutline": [
+      "სცენარის ნიმუში",
+      "სცენარის შაბლონი",
+      "ნიმუში",
+      "შაბლონი"
+    ],
+    "then": [
+      "* ",
+      "მაშინ "
+    ],
+    "when": [
+      "* ",
+      "როდესაც ",
+      "როცა ",
+      "როგორც კი ",
+      "თუ "
+    ]
+  },
+  "kn": {
+    "and": [
+      "* ",
+      "ಮತ್ತು "
+    ],
+    "background": [
+      "ಹಿನ್ನೆಲೆ"
+    ],
+    "but": [
+      "* ",
+      "ಆದರೆ "
+    ],
+    "examples": [
+      "ಉದಾಹರಣೆಗಳು"
+    ],
+    "feature": [
+      "ಹೆಚ್ಚಳ"
+    ],
+    "given": [
+      "* ",
+      "ನೀಡಿದ "
+    ],
+    "name": "Kannada",
+    "native": "ಕನ್ನಡ",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "ಉದಾಹರಣೆ",
+      "ಕಥಾಸಾರಾಂಶ"
+    ],
+    "scenarioOutline": [
+      "ವಿವರಣೆ"
+    ],
+    "then": [
+      "* ",
+      "ನಂತರ "
+    ],
+    "when": [
+      "* ",
+      "ಸ್ಥಿತಿಯನ್ನು "
+    ]
+  },
+  "ko": {
+    "and": [
+      "* ",
+      "그리고 "
+    ],
+    "background": [
+      "배경"
+    ],
+    "but": [
+      "* ",
+      "하지만 ",
+      "단 "
+    ],
+    "examples": [
+      "예"
+    ],
+    "feature": [
+      "기능"
+    ],
+    "given": [
+      "* ",
+      "조건 ",
+      "먼저 "
+    ],
+    "name": "Korean",
+    "native": "한국어",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "시나리오"
+    ],
+    "scenarioOutline": [
+      "시나리오 개요"
+    ],
+    "then": [
+      "* ",
+      "그러면 "
+    ],
+    "when": [
+      "* ",
+      "만일 ",
+      "만약 "
+    ]
+  },
+  "lt": {
+    "and": [
+      "* ",
+      "Ir "
+    ],
+    "background": [
+      "Kontekstas"
+    ],
+    "but": [
+      "* ",
+      "Bet "
+    ],
+    "examples": [
+      "Pavyzdžiai",
+      "Scenarijai",
+      "Variantai"
+    ],
+    "feature": [
+      "Savybė"
+    ],
+    "given": [
+      "* ",
+      "Duota "
+    ],
+    "name": "Lithuanian",
+    "native": "lietuvių kalba",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Pavyzdys",
+      "Scenarijus"
+    ],
+    "scenarioOutline": [
+      "Scenarijaus šablonas"
+    ],
+    "then": [
+      "* ",
+      "Tada "
+    ],
+    "when": [
+      "* ",
+      "Kai "
+    ]
+  },
+  "lu": {
+    "and": [
+      "* ",
+      "an ",
+      "a "
+    ],
+    "background": [
+      "Hannergrond"
+    ],
+    "but": [
+      "* ",
+      "awer ",
+      "mä "
+    ],
+    "examples": [
+      "Beispiller"
+    ],
+    "feature": [
+      "Funktionalitéit"
+    ],
+    "given": [
+      "* ",
+      "ugeholl "
+    ],
+    "name": "Luxemburgish",
+    "native": "Lëtzebuergesch",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Beispill",
+      "Szenario"
+    ],
+    "scenarioOutline": [
+      "Plang vum Szenario"
+    ],
+    "then": [
+      "* ",
+      "dann "
+    ],
+    "when": [
+      "* ",
+      "wann "
+    ]
+  },
+  "lv": {
+    "and": [
+      "* ",
+      "Un "
+    ],
+    "background": [
+      "Konteksts",
+      "Situācija"
+    ],
+    "but": [
+      "* ",
+      "Bet "
+    ],
+    "examples": [
+      "Piemēri",
+      "Paraugs"
+    ],
+    "feature": [
+      "Funkcionalitāte",
+      "Fīča"
+    ],
+    "given": [
+      "* ",
+      "Kad "
+    ],
+    "name": "Latvian",
+    "native": "latviešu",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Piemērs",
+      "Scenārijs"
+    ],
+    "scenarioOutline": [
+      "Scenārijs pēc parauga"
+    ],
+    "then": [
+      "* ",
+      "Tad "
+    ],
+    "when": [
+      "* ",
+      "Ja "
+    ]
+  },
+  "mk-Cyrl": {
+    "and": [
+      "* ",
+      "И "
+    ],
+    "background": [
+      "Контекст",
+      "Содржина"
+    ],
+    "but": [
+      "* ",
+      "Но "
+    ],
+    "examples": [
+      "Примери",
+      "Сценарија"
+    ],
+    "feature": [
+      "Функционалност",
+      "Бизнис потреба",
+      "Можност"
+    ],
+    "given": [
+      "* ",
+      "Дадено ",
+      "Дадена "
+    ],
+    "name": "Macedonian",
+    "native": "Македонски",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Пример",
+      "Сценарио",
+      "На пример"
+    ],
+    "scenarioOutline": [
+      "Преглед на сценарија",
+      "Скица",
+      "Концепт"
+    ],
+    "then": [
+      "* ",
+      "Тогаш "
+    ],
+    "when": [
+      "* ",
+      "Кога "
+    ]
+  },
+  "mk-Latn": {
+    "and": [
+      "* ",
+      "I "
+    ],
+    "background": [
+      "Kontekst",
+      "Sodrzhina"
+    ],
+    "but": [
+      "* ",
+      "No "
+    ],
+    "examples": [
+      "Primeri",
+      "Scenaria"
+    ],
+    "feature": [
+      "Funkcionalnost",
+      "Biznis potreba",
+      "Mozhnost"
+    ],
+    "given": [
+      "* ",
+      "Dadeno ",
+      "Dadena "
+    ],
+    "name": "Macedonian (Latin)",
+    "native": "Makedonski (Latinica)",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Scenario",
+      "Na primer"
+    ],
+    "scenarioOutline": [
+      "Pregled na scenarija",
+      "Skica",
+      "Koncept"
+    ],
+    "then": [
+      "* ",
+      "Togash "
+    ],
+    "when": [
+      "* ",
+      "Koga "
+    ]
+  },
+  "mn": {
+    "and": [
+      "* ",
+      "Мөн ",
+      "Тэгээд "
+    ],
+    "background": [
+      "Агуулга"
+    ],
+    "but": [
+      "* ",
+      "Гэхдээ ",
+      "Харин "
+    ],
+    "examples": [
+      "Тухайлбал"
+    ],
+    "feature": [
+      "Функц",
+      "Функционал"
+    ],
+    "given": [
+      "* ",
+      "Өгөгдсөн нь ",
+      "Анх "
+    ],
+    "name": "Mongolian",
+    "native": "монгол",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Сценар"
+    ],
+    "scenarioOutline": [
+      "Сценарын төлөвлөгөө"
+    ],
+    "then": [
+      "* ",
+      "Тэгэхэд ",
+      "Үүний дараа "
+    ],
+    "when": [
+      "* ",
+      "Хэрэв "
+    ]
+  },
+  "ne": {
+    "and": [
+      "* ",
+      "र ",
+      "अनि "
+    ],
+    "background": [
+      "पृष्ठभूमी"
+    ],
+    "but": [
+      "* ",
+      "तर "
+    ],
+    "examples": [
+      "उदाहरण",
+      "उदाहरणहरु"
+    ],
+    "feature": [
+      "सुविधा",
+      "विशेषता"
+    ],
+    "given": [
+      "* ",
+      "दिइएको ",
+      "दिएको ",
+      "यदि "
+    ],
+    "name": "Nepali",
+    "native": "नेपाली",
+    "rule": [
+      "नियम"
+    ],
+    "scenario": [
+      "परिदृश्य"
+    ],
+    "scenarioOutline": [
+      "परिदृश्य रूपरेखा"
+    ],
+    "then": [
+      "* ",
+      "त्यसपछि ",
+      "अनी "
+    ],
+    "when": [
+      "* ",
+      "जब "
+    ]
+  },
+  "nl": {
+    "and": [
+      "* ",
+      "En "
+    ],
+    "background": [
+      "Achtergrond"
+    ],
+    "but": [
+      "* ",
+      "Maar "
+    ],
+    "examples": [
+      "Voorbeelden"
+    ],
+    "feature": [
+      "Functionaliteit"
+    ],
+    "given": [
+      "* ",
+      "Gegeven ",
+      "Stel "
+    ],
+    "name": "Dutch",
+    "native": "Nederlands",
+    "rule": [
+      "Regel"
+    ],
+    "scenario": [
+      "Voorbeeld",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Abstract Scenario"
+    ],
+    "then": [
+      "* ",
+      "Dan "
+    ],
+    "when": [
+      "* ",
+      "Als ",
+      "Wanneer "
+    ]
+  },
+  "no": {
+    "and": [
+      "* ",
+      "Og "
+    ],
+    "background": [
+      "Bakgrunn"
+    ],
+    "but": [
+      "* ",
+      "Men "
+    ],
+    "examples": [
+      "Eksempler"
+    ],
+    "feature": [
+      "Egenskap"
+    ],
+    "given": [
+      "* ",
+      "Gitt "
+    ],
+    "name": "Norwegian",
+    "native": "norsk",
+    "rule": [
+      "Regel"
+    ],
+    "scenario": [
+      "Eksempel",
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Scenariomal",
+      "Abstrakt Scenario"
+    ],
+    "then": [
+      "* ",
+      "Så "
+    ],
+    "when": [
+      "* ",
+      "Når "
+    ]
+  },
+  "pa": {
+    "and": [
+      "* ",
+      "ਅਤੇ "
+    ],
+    "background": [
+      "ਪਿਛੋਕੜ"
+    ],
+    "but": [
+      "* ",
+      "ਪਰ "
+    ],
+    "examples": [
+      "ਉਦਾਹਰਨਾਂ"
+    ],
+    "feature": [
+      "ਖਾਸੀਅਤ",
+      "ਮੁਹਾਂਦਰਾ",
+      "ਨਕਸ਼ ਨੁਹਾਰ"
+    ],
+    "given": [
+      "* ",
+      "ਜੇਕਰ ",
+      "ਜਿਵੇਂ ਕਿ "
+    ],
+    "name": "Panjabi",
+    "native": "ਪੰਜਾਬੀ",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "ਉਦਾਹਰਨ",
+      "ਪਟਕਥਾ"
+    ],
+    "scenarioOutline": [
+      "ਪਟਕਥਾ ਢਾਂਚਾ",
+      "ਪਟਕਥਾ ਰੂਪ ਰੇਖਾ"
+    ],
+    "then": [
+      "* ",
+      "ਤਦ "
+    ],
+    "when": [
+      "* ",
+      "ਜਦੋਂ "
+    ]
+  },
+  "pl": {
+    "and": [
+      "* ",
+      "Oraz ",
+      "I "
+    ],
+    "background": [
+      "Założenia"
+    ],
+    "but": [
+      "* ",
+      "Ale "
+    ],
+    "examples": [
+      "Przykłady"
+    ],
+    "feature": [
+      "Właściwość",
+      "Funkcja",
+      "Aspekt",
+      "Potrzeba biznesowa"
+    ],
+    "given": [
+      "* ",
+      "Zakładając ",
+      "Mając ",
+      "Zakładając, że "
+    ],
+    "name": "Polish",
+    "native": "polski",
+    "rule": [
+      "Zasada",
+      "Reguła"
+    ],
+    "scenario": [
+      "Przykład",
+      "Scenariusz"
+    ],
+    "scenarioOutline": [
+      "Szablon scenariusza"
+    ],
+    "then": [
+      "* ",
+      "Wtedy "
+    ],
+    "when": [
+      "* ",
+      "Jeżeli ",
+      "Jeśli ",
+      "Gdy ",
+      "Kiedy "
+    ]
+  },
+  "pt": {
+    "and": [
+      "* ",
+      "E "
+    ],
+    "background": [
+      "Contexto",
+      "Cenário de Fundo",
+      "Cenario de Fundo",
+      "Fundo"
+    ],
+    "but": [
+      "* ",
+      "Mas "
+    ],
+    "examples": [
+      "Exemplos",
+      "Cenários",
+      "Cenarios"
+    ],
+    "feature": [
+      "Funcionalidade",
+      "Característica",
+      "Caracteristica"
+    ],
+    "given": [
+      "* ",
+      "Dado ",
+      "Dada ",
+      "Dados ",
+      "Dadas "
+    ],
+    "name": "Portuguese",
+    "native": "português",
+    "rule": [
+      "Regra"
+    ],
+    "scenario": [
+      "Exemplo",
+      "Cenário",
+      "Cenario"
+    ],
+    "scenarioOutline": [
+      "Esquema do Cenário",
+      "Esquema do Cenario",
+      "Delineação do Cenário",
+      "Delineacao do Cenario"
+    ],
+    "then": [
+      "* ",
+      "Então ",
+      "Entao "
+    ],
+    "when": [
+      "* ",
+      "Quando "
+    ]
+  },
+  "ro": {
+    "and": [
+      "* ",
+      "Si ",
+      "Și ",
+      "Şi "
+    ],
+    "background": [
+      "Context"
+    ],
+    "but": [
+      "* ",
+      "Dar "
+    ],
+    "examples": [
+      "Exemple"
+    ],
+    "feature": [
+      "Functionalitate",
+      "Funcționalitate",
+      "Funcţionalitate"
+    ],
+    "given": [
+      "* ",
+      "Date fiind ",
+      "Dat fiind ",
+      "Dată fiind",
+      "Dati fiind ",
+      "Dați fiind ",
+      "Daţi fiind "
+    ],
+    "name": "Romanian",
+    "native": "română",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Exemplu",
+      "Scenariu"
+    ],
+    "scenarioOutline": [
+      "Structura scenariu",
+      "Structură scenariu"
+    ],
+    "then": [
+      "* ",
+      "Atunci "
+    ],
+    "when": [
+      "* ",
+      "Cand ",
+      "Când "
+    ]
+  },
+  "ru": {
+    "and": [
+      "* ",
+      "И ",
+      "К тому же ",
+      "Также "
+    ],
+    "background": [
+      "Предыстория",
+      "Контекст"
+    ],
+    "but": [
+      "* ",
+      "Но ",
+      "А ",
+      "Иначе "
+    ],
+    "examples": [
+      "Примеры"
+    ],
+    "feature": [
+      "Функция",
+      "Функциональность",
+      "Функционал",
+      "Свойство",
+      "Фича"
+    ],
+    "given": [
+      "* ",
+      "Допустим ",
+      "Дано ",
+      "Пусть "
+    ],
+    "name": "Russian",
+    "native": "русский",
+    "rule": [
+      "Правило"
+    ],
+    "scenario": [
+      "Пример",
+      "Сценарий"
+    ],
+    "scenarioOutline": [
+      "Структура сценария",
+      "Шаблон сценария"
+    ],
+    "then": [
+      "* ",
+      "То ",
+      "Затем ",
+      "Тогда "
+    ],
+    "when": [
+      "* ",
+      "Когда ",
+      "Если "
+    ]
+  },
+  "sk": {
+    "and": [
+      "* ",
+      "A ",
+      "A tiež ",
+      "A taktiež ",
+      "A zároveň "
+    ],
+    "background": [
+      "Pozadie"
+    ],
+    "but": [
+      "* ",
+      "Ale "
+    ],
+    "examples": [
+      "Príklady"
+    ],
+    "feature": [
+      "Požiadavka",
+      "Funkcia",
+      "Vlastnosť"
+    ],
+    "given": [
+      "* ",
+      "Pokiaľ ",
+      "Za predpokladu "
+    ],
+    "name": "Slovak",
+    "native": "Slovensky",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Príklad",
+      "Scenár"
+    ],
+    "scenarioOutline": [
+      "Náčrt Scenáru",
+      "Náčrt Scenára",
+      "Osnova Scenára"
+    ],
+    "then": [
+      "* ",
+      "Tak ",
+      "Potom "
+    ],
+    "when": [
+      "* ",
+      "Keď ",
+      "Ak "
+    ]
+  },
+  "sl": {
+    "and": [
+      "In ",
+      "Ter "
+    ],
+    "background": [
+      "Kontekst",
+      "Osnova",
+      "Ozadje"
+    ],
+    "but": [
+      "Toda ",
+      "Ampak ",
+      "Vendar "
+    ],
+    "examples": [
+      "Primeri",
+      "Scenariji"
+    ],
+    "feature": [
+      "Funkcionalnost",
+      "Funkcija",
+      "Možnosti",
+      "Moznosti",
+      "Lastnost",
+      "Značilnost"
+    ],
+    "given": [
+      "Dano ",
+      "Podano ",
+      "Zaradi ",
+      "Privzeto "
+    ],
+    "name": "Slovenian",
+    "native": "Slovenski",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Primer",
+      "Scenarij"
+    ],
+    "scenarioOutline": [
+      "Struktura scenarija",
+      "Skica",
+      "Koncept",
+      "Oris scenarija",
+      "Osnutek"
+    ],
+    "then": [
+      "Nato ",
+      "Potem ",
+      "Takrat "
+    ],
+    "when": [
+      "Ko ",
+      "Ce ",
+      "Če ",
+      "Kadar "
+    ]
+  },
+  "sr-Cyrl": {
+    "and": [
+      "* ",
+      "И "
+    ],
+    "background": [
+      "Контекст",
+      "Основа",
+      "Позадина"
+    ],
+    "but": [
+      "* ",
+      "Али "
+    ],
+    "examples": [
+      "Примери",
+      "Сценарији"
+    ],
+    "feature": [
+      "Функционалност",
+      "Могућност",
+      "Особина"
+    ],
+    "given": [
+      "* ",
+      "За дато ",
+      "За дате ",
+      "За дати "
+    ],
+    "name": "Serbian",
+    "native": "Српски",
+    "rule": [
+      "Правило"
+    ],
+    "scenario": [
+      "Сценарио",
+      "Пример"
+    ],
+    "scenarioOutline": [
+      "Структура сценарија",
+      "Скица",
+      "Концепт"
+    ],
+    "then": [
+      "* ",
+      "Онда "
+    ],
+    "when": [
+      "* ",
+      "Када ",
+      "Кад "
+    ]
+  },
+  "sr-Latn": {
+    "and": [
+      "* ",
+      "I "
+    ],
+    "background": [
+      "Kontekst",
+      "Osnova",
+      "Pozadina"
+    ],
+    "but": [
+      "* ",
+      "Ali "
+    ],
+    "examples": [
+      "Primeri",
+      "Scenariji"
+    ],
+    "feature": [
+      "Funkcionalnost",
+      "Mogućnost",
+      "Mogucnost",
+      "Osobina"
+    ],
+    "given": [
+      "* ",
+      "Za dato ",
+      "Za date ",
+      "Za dati "
+    ],
+    "name": "Serbian (Latin)",
+    "native": "Srpski (Latinica)",
+    "rule": [
+      "Pravilo"
+    ],
+    "scenario": [
+      "Scenario",
+      "Primer"
+    ],
+    "scenarioOutline": [
+      "Struktura scenarija",
+      "Skica",
+      "Koncept"
+    ],
+    "then": [
+      "* ",
+      "Onda "
+    ],
+    "when": [
+      "* ",
+      "Kada ",
+      "Kad "
+    ]
+  },
+  "sv": {
+    "and": [
+      "* ",
+      "Och "
+    ],
+    "background": [
+      "Bakgrund"
+    ],
+    "but": [
+      "* ",
+      "Men "
+    ],
+    "examples": [
+      "Exempel"
+    ],
+    "feature": [
+      "Egenskap"
+    ],
+    "given": [
+      "* ",
+      "Givet "
+    ],
+    "name": "Swedish",
+    "native": "Svenska",
+    "rule": [
+      "Regel"
+    ],
+    "scenario": [
+      "Scenario"
+    ],
+    "scenarioOutline": [
+      "Abstrakt Scenario",
+      "Scenariomall"
+    ],
+    "then": [
+      "* ",
+      "Så "
+    ],
+    "when": [
+      "* ",
+      "När "
+    ]
+  },
+  "ta": {
+    "and": [
+      "* ",
+      "மேலும் ",
+      "மற்றும் "
+    ],
+    "background": [
+      "பின்னணி"
+    ],
+    "but": [
+      "* ",
+      "ஆனால் "
+    ],
+    "examples": [
+      "எடுத்துக்காட்டுகள்",
+      "காட்சிகள்",
+      "நிலைமைகளில்"
+    ],
+    "feature": [
+      "அம்சம்",
+      "வணிக தேவை",
+      "திறன்"
+    ],
+    "given": [
+      "* ",
+      "கொடுக்கப்பட்ட "
+    ],
+    "name": "Tamil",
+    "native": "தமிழ்",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "உதாரணமாக",
+      "காட்சி"
+    ],
+    "scenarioOutline": [
+      "காட்சி சுருக்கம்",
+      "காட்சி வார்ப்புரு"
+    ],
+    "then": [
+      "* ",
+      "அப்பொழுது "
+    ],
+    "when": [
+      "* ",
+      "எப்போது "
+    ]
+  },
+  "th": {
+    "and": [
+      "* ",
+      "และ "
+    ],
+    "background": [
+      "แนวคิด"
+    ],
+    "but": [
+      "* ",
+      "แต่ "
+    ],
+    "examples": [
+      "ชุดของตัวอย่าง",
+      "ชุดของเหตุการณ์"
+    ],
+    "feature": [
+      "โครงหลัก",
+      "ความต้องการทางธุรกิจ",
+      "ความสามารถ"
+    ],
+    "given": [
+      "* ",
+      "กำหนดให้ "
+    ],
+    "name": "Thai",
+    "native": "ไทย",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "เหตุการณ์"
+    ],
+    "scenarioOutline": [
+      "สรุปเหตุการณ์",
+      "โครงสร้างของเหตุการณ์"
+    ],
+    "then": [
+      "* ",
+      "ดังนั้น "
+    ],
+    "when": [
+      "* ",
+      "เมื่อ "
+    ]
+  },
+  "te": {
+    "and": [
+      "* ",
+      "మరియు "
+    ],
+    "background": [
+      "నేపథ్యం"
+    ],
+    "but": [
+      "* ",
+      "కాని "
+    ],
+    "examples": [
+      "ఉదాహరణలు"
+    ],
+    "feature": [
+      "గుణము"
+    ],
+    "given": [
+      "* ",
+      "చెప్పబడినది "
+    ],
+    "name": "Telugu",
+    "native": "తెలుగు",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "ఉదాహరణ",
+      "సన్నివేశం"
+    ],
+    "scenarioOutline": [
+      "కథనం"
+    ],
+    "then": [
+      "* ",
+      "అప్పుడు "
+    ],
+    "when": [
+      "* ",
+      "ఈ పరిస్థితిలో "
+    ]
+  },
+  "tlh": {
+    "and": [
+      "* ",
+      "'ej ",
+      "latlh "
+    ],
+    "background": [
+      "mo'"
+    ],
+    "but": [
+      "* ",
+      "'ach ",
+      "'a "
+    ],
+    "examples": [
+      "ghantoH",
+      "lutmey"
+    ],
+    "feature": [
+      "Qap",
+      "Qu'meH 'ut",
+      "perbogh",
+      "poQbogh malja'",
+      "laH"
+    ],
+    "given": [
+      "* ",
+      "ghu' noblu' ",
+      "DaH ghu' bejlu' "
+    ],
+    "name": "Klingon",
+    "native": "tlhIngan",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "lut"
+    ],
+    "scenarioOutline": [
+      "lut chovnatlh"
+    ],
+    "then": [
+      "* ",
+      "vaj "
+    ],
+    "when": [
+      "* ",
+      "qaSDI' "
+    ]
+  },
+  "tr": {
+    "and": [
+      "* ",
+      "Ve "
+    ],
+    "background": [
+      "Geçmiş"
+    ],
+    "but": [
+      "* ",
+      "Fakat ",
+      "Ama "
+    ],
+    "examples": [
+      "Örnekler"
+    ],
+    "feature": [
+      "Özellik"
+    ],
+    "given": [
+      "* ",
+      "Diyelim ki "
+    ],
+    "name": "Turkish",
+    "native": "Türkçe",
+    "rule": [
+      "Kural"
+    ],
+    "scenario": [
+      "Örnek",
+      "Senaryo"
+    ],
+    "scenarioOutline": [
+      "Senaryo taslağı"
+    ],
+    "then": [
+      "* ",
+      "O zaman "
+    ],
+    "when": [
+      "* ",
+      "Eğer ki "
+    ]
+  },
+  "tt": {
+    "and": [
+      "* ",
+      "Һәм ",
+      "Вә "
+    ],
+    "background": [
+      "Кереш"
+    ],
+    "but": [
+      "* ",
+      "Ләкин ",
+      "Әмма "
+    ],
+    "examples": [
+      "Үрнәкләр",
+      "Мисаллар"
+    ],
+    "feature": [
+      "Мөмкинлек",
+      "Үзенчәлеклелек"
+    ],
+    "given": [
+      "* ",
+      "Әйтик "
+    ],
+    "name": "Tatar",
+    "native": "Татарча",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Сценарий"
+    ],
+    "scenarioOutline": [
+      "Сценарийның төзелеше"
+    ],
+    "then": [
+      "* ",
+      "Нәтиҗәдә "
+    ],
+    "when": [
+      "* ",
+      "Әгәр "
+    ]
+  },
+  "uk": {
+    "and": [
+      "* ",
+      "І ",
+      "А також ",
+      "Та "
+    ],
+    "background": [
+      "Передумова"
+    ],
+    "but": [
+      "* ",
+      "Але "
+    ],
+    "examples": [
+      "Приклади"
+    ],
+    "feature": [
+      "Функціонал"
+    ],
+    "given": [
+      "* ",
+      "Припустимо ",
+      "Припустимо, що ",
+      "Нехай ",
+      "Дано "
+    ],
+    "name": "Ukrainian",
+    "native": "Українська",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Приклад",
+      "Сценарій"
+    ],
+    "scenarioOutline": [
+      "Структура сценарію"
+    ],
+    "then": [
+      "* ",
+      "То ",
+      "Тоді "
+    ],
+    "when": [
+      "* ",
+      "Якщо ",
+      "Коли "
+    ]
+  },
+  "ur": {
+    "and": [
+      "* ",
+      "اور "
+    ],
+    "background": [
+      "پس منظر"
+    ],
+    "but": [
+      "* ",
+      "لیکن "
+    ],
+    "examples": [
+      "مثالیں"
+    ],
+    "feature": [
+      "صلاحیت",
+      "کاروبار کی ضرورت",
+      "خصوصیت"
+    ],
+    "given": [
+      "* ",
+      "اگر ",
+      "بالفرض ",
+      "فرض کیا "
+    ],
+    "name": "Urdu",
+    "native": "اردو",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "منظرنامہ"
+    ],
+    "scenarioOutline": [
+      "منظر نامے کا خاکہ"
+    ],
+    "then": [
+      "* ",
+      "پھر ",
+      "تب "
+    ],
+    "when": [
+      "* ",
+      "جب "
+    ]
+  },
+  "uz": {
+    "and": [
+      "* ",
+      "Ва "
+    ],
+    "background": [
+      "Тарих"
+    ],
+    "but": [
+      "* ",
+      "Лекин ",
+      "Бирок ",
+      "Аммо "
+    ],
+    "examples": [
+      "Мисоллар"
+    ],
+    "feature": [
+      "Функционал"
+    ],
+    "given": [
+      "* ",
+      "Belgilangan "
+    ],
+    "name": "Uzbek",
+    "native": "Узбекча",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "Сценарий"
+    ],
+    "scenarioOutline": [
+      "Сценарий структураси"
+    ],
+    "then": [
+      "* ",
+      "Унда "
+    ],
+    "when": [
+      "* ",
+      "Агар "
+    ]
+  },
+  "vi": {
+    "and": [
+      "* ",
+      "Và "
+    ],
+    "background": [
+      "Bối cảnh"
+    ],
+    "but": [
+      "* ",
+      "Nhưng "
+    ],
+    "examples": [
+      "Dữ liệu"
+    ],
+    "feature": [
+      "Tính năng"
+    ],
+    "given": [
+      "* ",
+      "Biết ",
+      "Cho "
+    ],
+    "name": "Vietnamese",
+    "native": "Tiếng Việt",
+    "rule": [
+      "Quy tắc"
+    ],
+    "scenario": [
+      "Tình huống",
+      "Kịch bản"
+    ],
+    "scenarioOutline": [
+      "Khung tình huống",
+      "Khung kịch bản"
+    ],
+    "then": [
+      "* ",
+      "Thì "
+    ],
+    "when": [
+      "* ",
+      "Khi "
+    ]
+  },
+  "zh-CN": {
+    "and": [
+      "* ",
+      "而且",
+      "并且",
+      "同时"
+    ],
+    "background": [
+      "背景"
+    ],
+    "but": [
+      "* ",
+      "但是"
+    ],
+    "examples": [
+      "例子"
+    ],
+    "feature": [
+      "功能"
+    ],
+    "given": [
+      "* ",
+      "假如",
+      "假设",
+      "假定"
+    ],
+    "name": "Chinese simplified",
+    "native": "简体中文",
+    "rule": [
+      "Rule",
+      "规则"
+    ],
+    "scenario": [
+      "场景",
+      "剧本"
+    ],
+    "scenarioOutline": [
+      "场景大纲",
+      "剧本大纲"
+    ],
+    "then": [
+      "* ",
+      "那么"
+    ],
+    "when": [
+      "* ",
+      "当"
+    ]
+  },
+  "ml": {
+    "and": [
+      "* ",
+      "ഒപ്പം"
+    ],
+    "background": [
+      "പശ്ചാത്തലം"
+    ],
+    "but": [
+      "* ",
+      "പക്ഷേ"
+    ],
+    "examples": [
+      "ഉദാഹരണങ്ങൾ"
+    ],
+    "feature": [
+      "സവിശേഷത"
+    ],
+    "given": [
+      "* ",
+      "നൽകിയത്"
+    ],
+    "name": "Malayalam",
+    "native": "മലയാളം",
+    "rule": [
+      "നിയമം"
+    ],
+    "scenario": [
+      "രംഗം"
+    ],
+    "scenarioOutline": [
+      "സാഹചര്യത്തിന്റെ രൂപരേഖ"
+    ],
+    "then": [
+      "* ",
+      "പിന്നെ"
+    ],
+    "when": [
+      "എപ്പോൾ"
+    ]
+  },
+  "zh-TW": {
+    "and": [
+      "* ",
+      "而且",
+      "並且",
+      "同時"
+    ],
+    "background": [
+      "背景"
+    ],
+    "but": [
+      "* ",
+      "但是"
+    ],
+    "examples": [
+      "例子"
+    ],
+    "feature": [
+      "功能"
+    ],
+    "given": [
+      "* ",
+      "假如",
+      "假設",
+      "假定"
+    ],
+    "name": "Chinese traditional",
+    "native": "繁體中文",
+    "rule": [
+      "Rule"
+    ],
+    "scenario": [
+      "場景",
+      "劇本"
+    ],
+    "scenarioOutline": [
+      "場景大綱",
+      "劇本大綱"
+    ],
+    "then": [
+      "* ",
+      "那麼"
+    ],
+    "when": [
+      "* ",
+      "當"
+    ]
+  },
+  "mr": {
+    "and": [
+      "* ",
+      "आणि ",
+      "तसेच "
+    ],
+    "background": [
+      "पार्श्वभूमी"
+    ],
+    "but": [
+      "* ",
+      "पण ",
+      "परंतु "
+    ],
+    "examples": [
+      "उदाहरण"
+    ],
+    "feature": [
+      "वैशिष्ट्य",
+      "सुविधा"
+    ],
+    "given": [
+      "* ",
+      "जर",
+      "दिलेल्या प्रमाणे "
+    ],
+    "name": "Marathi",
+    "native": "मराठी",
+    "rule": [
+      "नियम"
+    ],
+    "scenario": [
+      "परिदृश्य"
+    ],
+    "scenarioOutline": [
+      "परिदृश्य रूपरेखा"
+    ],
+    "then": [
+      "* ",
+      "मग ",
+      "तेव्हा "
+    ],
+    "when": [
+      "* ",
+      "जेव्हा "
+    ]
+  },
+  "amh": {
+    "and": [
+      "* ",
+      "እና "
+    ],
+    "background": [
+      "ቅድመ ሁኔታ",
+      "መነሻ",
+      "መነሻ ሀሳብ"
+    ],
+    "but": [
+      "* ",
+      "ግን "
+    ],
+    "examples": [
+      "ምሳሌዎች",
+      "ሁናቴዎች"
+    ],
+    "feature": [
+      "ስራ",
+      "የተፈለገው ስራ",
+      "የሚፈለገው ድርጊት"
+    ],
+    "given": [
+      "* ",
+      "የተሰጠ "
+    ],
+    "name": "Amharic",
+    "native": "አማርኛ",
+    "rule": [
+      "ህግ"
+    ],
+    "scenario": [
+      "ምሳሌ",
+      "ሁናቴ"
+    ],
+    "scenarioOutline": [
+      "ሁናቴ ዝርዝር",
+      "ሁናቴ አብነት"
+    ],
+    "then": [
+      "* ",
+      "ከዚያ "
+    ],
+    "when": [
+      "* ",
+      "መቼ "
+    ]
+  }
+}
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/gherkin_line.py b/backend/venv/lib/python3.12/site-packages/gherkin/gherkin_line.py
new file mode 100644
index 00000000..8ca94528
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/gherkin_line.py
@@ -0,0 +1,90 @@
+import re
+from .errors import ParserException
+
+
+class GherkinLine(object):
+    def __init__(self, line_text, line_number):
+        self._line_text = line_text
+        self._line_number = line_number
+        self._trimmed_line_text = line_text.lstrip()
+        self.indent = len(line_text) - len(self._trimmed_line_text)
+
+    def get_rest_trimmed(self, length):
+        return self._trimmed_line_text[length:].strip()
+
+    def get_line_text(self, indent_to_remove=-1):
+        if indent_to_remove < 0 or indent_to_remove > self.indent:
+            return self._trimmed_line_text
+        else:
+            return self._line_text[indent_to_remove:]
+
+    def is_empty(self):
+        return not self._trimmed_line_text
+
+    def startswith(self, prefix):
+        return self._trimmed_line_text.startswith(prefix)
+
+    def startswith_title_keyword(self, keyword):
+        return self._trimmed_line_text.startswith(keyword + ':')
+
+    @property
+    def table_cells(self):
+        cells = []
+        for cell, col in self.split_table_cells(self._trimmed_line_text.strip()):
+            lstripped_cell = re.sub(r"^[^\S\n]*", "" , cell, flags=re.U)
+            cell_indent = len(cell) - len(lstripped_cell)
+            cells.append({'column': col + self.indent + cell_indent, 'text': re.sub(r"[^\S\n]*$", "", lstripped_cell, flags=re.U)})
+        return cells
+
+    def split_table_cells(self, row):
+        """
+        An iterator returning all the table cells in a row with their positions,
+        accounting for escaping.
+        """
+
+        row = iter(row)
+        col = 0
+        start_col = col + 1
+        cell = ''
+        first_cell = True
+        while True:
+            char = next(row, None)
+            col += 1
+            if char == '|':
+                if first_cell:
+                    # First cell (content before the first |) is skipped
+                    first_cell = False
+                else:
+                    yield (cell, start_col)
+                cell = ''
+                start_col = col + 1
+            elif char == '\\':
+                char = next(row, "")
+                col += 1
+                if char == 'n':
+                    cell += '\n'
+                else:
+                    if char not in ['|', '\\']:
+                        cell += '\\'
+                    cell += char
+            elif char:
+                cell += char
+            else:
+                break
+        # Last cell (content after the last |) is skipped
+
+    @property
+    def tags(self):
+        column = self.indent + 1
+        uncommented_line = re.split(r"\s#", self._trimmed_line_text.strip(), 2)[0]
+        items = uncommented_line.strip().split('@')
+        tags = []
+        for item in items[1:]:
+            tag_value = '@' + item.strip()
+            if re.search(r"[^\S+]", tag_value) is not None:
+                location = {'line': self._line_number, 'column': column}
+                raise ParserException('A tag may not contain whitespace', location)
+
+            tags.append({'column': column, 'text': tag_value})
+            column += len(item) + 1
+        return tags
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/inout.py b/backend/venv/lib/python3.12/site-packages/gherkin/inout.py
new file mode 100644
index 00000000..bedb1544
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/inout.py
@@ -0,0 +1,56 @@
+from __future__ import print_function
+import json
+from .parser import Parser
+from .token_scanner import TokenScanner
+from .pickles.compiler import compile
+from .errors import ParserException, CompositeParserException
+
+class Inout(object):
+    def __init__(self, print_source, print_ast, print_pickles):
+        self.print_source = print_source
+        self.print_ast = print_ast
+        self.print_pickles = print_pickles
+        self.parser = Parser()
+        self.parser.stop_at_first_error = False
+
+    def process(self, input, output):
+        line = input.readline().rstrip()
+        event = json.loads(line)
+        if (event['type'] == 'source'):
+            uri = event['uri']
+            source = event['data']
+            token_scanner = TokenScanner(source)
+
+            try:
+                gherkin_document = self.parser.parse(token_scanner)
+                if (self.print_source):
+                    print(line, file=output)
+                if (self.print_ast):
+                    print(json.dumps(gherkin_document), file=output)
+                if (self.print_pickles):
+                    pickles = compile(gherkin_document, uri)
+                    for pickle in pickles:
+                        print(json.dumps(pickle), file=output)
+            except CompositeParserException as e:
+                self.print_errors(output, e.errors, uri)
+            except ParserException as e:
+                self.print_errors(output, [e], uri)
+
+    def print_errors(self, output, errors, uri):
+        for error in errors:
+            attachment = {
+                'type': "attachment",
+                'source': {
+                    'uri': uri,
+                    'start': {
+                        'line': error.location['line'],
+                        'column': error.location['column']
+                    }
+                },
+                'data': error.message,
+                'media': {
+                    'encoding': "utf-8",
+                    'type': "text/x.cucumber.stacktrace+plain"
+                }
+            }
+            print(json.dumps(attachment), file=output)
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/parser.py b/backend/venv/lib/python3.12/site-packages/gherkin/parser.py
new file mode 100644
index 00000000..1c0baad9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/parser.py
@@ -0,0 +1,3321 @@
+# This file is generated. Do not edit! Edit gherkin-python.razor instead.
+import sys
+from collections import deque
+from .ast_builder import AstBuilder
+from .token_matcher import TokenMatcher
+from .token_scanner import TokenScanner
+from .errors import UnexpectedEOFException, UnexpectedTokenException, ParserException, CompositeParserException
+
+RULE_TYPE = [
+    'None',
+    '_EOF',  # #EOF
+    '_Empty',  # #Empty
+    '_Comment',  # #Comment
+    '_TagLine',  # #TagLine
+    '_FeatureLine',  # #FeatureLine
+    '_RuleLine',  # #RuleLine
+    '_BackgroundLine',  # #BackgroundLine
+    '_ScenarioLine',  # #ScenarioLine
+    '_ExamplesLine',  # #ExamplesLine
+    '_StepLine',  # #StepLine
+    '_DocStringSeparator',  # #DocStringSeparator
+    '_TableRow',  # #TableRow
+    '_Language',  # #Language
+    '_Other',  # #Other
+    'GherkinDocument',  # GherkinDocument! := Feature?
+    'Feature',  # Feature! := FeatureHeader Background? ScenarioDefinition* Rule*
+    'FeatureHeader',  # FeatureHeader! := #Language? Tags? #FeatureLine DescriptionHelper
+    'Rule',  # Rule! := RuleHeader Background? ScenarioDefinition*
+    'RuleHeader',  # RuleHeader! := Tags? #RuleLine DescriptionHelper
+    'Background',  # Background! := #BackgroundLine DescriptionHelper Step*
+    'ScenarioDefinition',  # ScenarioDefinition! [#Empty|#Comment|#TagLine->#ScenarioLine] := Tags? Scenario
+    'Scenario',  # Scenario! := #ScenarioLine DescriptionHelper Step* ExamplesDefinition*
+    'ExamplesDefinition',  # ExamplesDefinition! [#Empty|#Comment|#TagLine->#ExamplesLine] := Tags? Examples
+    'Examples',  # Examples! := #ExamplesLine DescriptionHelper ExamplesTable?
+    'ExamplesTable',  # ExamplesTable! := #TableRow #TableRow*
+    'Step',  # Step! := #StepLine StepArg?
+    'StepArg',  # StepArg := (DataTable | DocString)
+    'DataTable',  # DataTable! := #TableRow+
+    'DocString',  # DocString! := #DocStringSeparator #Other* #DocStringSeparator
+    'Tags',  # Tags! := #TagLine+
+    'DescriptionHelper',  # DescriptionHelper := #Empty* Description? #Comment*
+    'Description',  # Description! := #Other+
+]
+
+
+class ParserContext(object):
+    def __init__(self, token_scanner, token_matcher, token_queue, errors):
+        self.token_scanner = token_scanner
+        self.token_matcher = token_matcher
+        self.token_queue = token_queue
+        self.errors = errors
+
+
+class Parser(object):
+    def __init__(self, ast_builder=None):
+        self.ast_builder = ast_builder if ast_builder is not None else AstBuilder()
+        self.stop_at_first_error = False
+
+    def parse(self, token_scanner_or_str, token_matcher=None):
+        if sys.version_info < (3, 0):
+            token_scanner = TokenScanner(token_scanner_or_str) if isinstance(token_scanner_or_str, basestring) else token_scanner_or_str
+        else:
+            token_scanner = TokenScanner(token_scanner_or_str) if isinstance(token_scanner_or_str, str) else token_scanner_or_str
+        self.ast_builder.reset()
+        if token_matcher is None:
+            token_matcher = TokenMatcher()
+        token_matcher.reset()
+        context = ParserContext(
+            token_scanner,
+            token_matcher,
+            deque(),
+            [])
+
+        self.start_rule(context, 'GherkinDocument')
+        state = 0
+        token = None
+        while True:
+            token = self.read_token(context)
+            state = self.match_token(state, token, context)
+            if token.eof():
+                break
+
+        self.end_rule(context, 'GherkinDocument')
+
+        if context.errors:
+            raise CompositeParserException(context.errors)
+
+        return self.get_result()
+
+    def build(self, context, token):
+        self.handle_ast_error(context, token, self.ast_builder.build)
+
+    def add_error(self, context, error):
+        if str(error) not in (str(e) for e in context.errors):
+            context.errors.append(error)
+            if len(context.errors) > 10:
+                raise CompositeParserException(context.errors)
+
+    def start_rule(self, context, rule_type):
+        self.handle_ast_error(context, rule_type, self.ast_builder.start_rule)
+
+    def end_rule(self, context, rule_type):
+        self.handle_ast_error(context, rule_type, self.ast_builder.end_rule)
+
+    def get_result(self):
+        return self.ast_builder.get_result()
+
+    def read_token(self, context):
+        if context.token_queue:
+            return context.token_queue.popleft()
+        else:
+            return context.token_scanner.read()
+
+    def match_EOF(self, context, token):
+        return self.handle_external_error(context, False, token, context.token_matcher.match_EOF)
+    def match_Empty(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_Empty)
+    def match_Comment(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_Comment)
+    def match_TagLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_TagLine)
+    def match_FeatureLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_FeatureLine)
+    def match_RuleLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_RuleLine)
+    def match_BackgroundLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_BackgroundLine)
+    def match_ScenarioLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_ScenarioLine)
+    def match_ExamplesLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_ExamplesLine)
+    def match_StepLine(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_StepLine)
+    def match_DocStringSeparator(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_DocStringSeparator)
+    def match_TableRow(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_TableRow)
+    def match_Language(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_Language)
+    def match_Other(self, context, token):
+        if token.eof():
+            return False
+        return self.handle_external_error(context, False, token, context.token_matcher.match_Other)
+    def match_token(self, state, token, context):
+        state_map = {
+            0: self.match_token_at_0,
+            1: self.match_token_at_1,
+            2: self.match_token_at_2,
+            3: self.match_token_at_3,
+            4: self.match_token_at_4,
+            5: self.match_token_at_5,
+            6: self.match_token_at_6,
+            7: self.match_token_at_7,
+            8: self.match_token_at_8,
+            9: self.match_token_at_9,
+            10: self.match_token_at_10,
+            11: self.match_token_at_11,
+            12: self.match_token_at_12,
+            13: self.match_token_at_13,
+            14: self.match_token_at_14,
+            15: self.match_token_at_15,
+            16: self.match_token_at_16,
+            17: self.match_token_at_17,
+            18: self.match_token_at_18,
+            19: self.match_token_at_19,
+            20: self.match_token_at_20,
+            21: self.match_token_at_21,
+            22: self.match_token_at_22,
+            23: self.match_token_at_23,
+            24: self.match_token_at_24,
+            25: self.match_token_at_25,
+            26: self.match_token_at_26,
+            27: self.match_token_at_27,
+            28: self.match_token_at_28,
+            29: self.match_token_at_29,
+            30: self.match_token_at_30,
+            31: self.match_token_at_31,
+            32: self.match_token_at_32,
+            33: self.match_token_at_33,
+            34: self.match_token_at_34,
+            35: self.match_token_at_35,
+            36: self.match_token_at_36,
+            37: self.match_token_at_37,
+            38: self.match_token_at_38,
+            39: self.match_token_at_39,
+            40: self.match_token_at_40,
+            41: self.match_token_at_41,
+            43: self.match_token_at_43,
+            44: self.match_token_at_44,
+            45: self.match_token_at_45,
+            46: self.match_token_at_46,
+            47: self.match_token_at_47,
+            48: self.match_token_at_48,
+            49: self.match_token_at_49,
+            50: self.match_token_at_50,
+        }
+        if state in state_map:
+            return state_map[state](token, context)
+        else:
+            raise RuntimeError("Unknown state: " + str(state))
+
+    # Start
+    def match_token_at_0(self, token, context):
+        if self.match_EOF(context, token):
+                self.build(context, token)
+                return 42
+        if self.match_Language(context, token):
+                self.start_rule(context, 'Feature')
+                self.start_rule(context, 'FeatureHeader')
+                self.build(context, token)
+                return 1
+        if self.match_TagLine(context, token):
+                self.start_rule(context, 'Feature')
+                self.start_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 2
+        if self.match_FeatureLine(context, token):
+                self.start_rule(context, 'Feature')
+                self.start_rule(context, 'FeatureHeader')
+                self.build(context, token)
+                return 3
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 0
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 0
+
+        state_comment = "State: 0 - Start"
+        token.detach
+        expected_tokens = ["#EOF", "#Language", "#TagLine", "#FeatureLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 0
+    # GherkinDocument:0>Feature:0>FeatureHeader:0>#Language:0
+    def match_token_at_1(self, token, context):
+        if self.match_TagLine(context, token):
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 2
+        if self.match_FeatureLine(context, token):
+                self.build(context, token)
+                return 3
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 1
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 1
+
+        state_comment = "State: 1 - GherkinDocument:0>Feature:0>FeatureHeader:0>#Language:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#FeatureLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 1
+    # GherkinDocument:0>Feature:0>FeatureHeader:1>Tags:0>#TagLine:0
+    def match_token_at_2(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 2
+        if self.match_FeatureLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.build(context, token)
+                return 3
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 2
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 2
+
+        state_comment = "State: 2 - GherkinDocument:0>Feature:0>FeatureHeader:1>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#FeatureLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 2
+    # GherkinDocument:0>Feature:0>FeatureHeader:2>#FeatureLine:0
+    def match_token_at_3(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 3
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 5
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 6
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 4
+
+        state_comment = "State: 3 - GherkinDocument:0>Feature:0>FeatureHeader:2>#FeatureLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 3
+    # GherkinDocument:0>Feature:0>FeatureHeader:3>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_4(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 5
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 6
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 4
+
+        state_comment = "State: 4 - GherkinDocument:0>Feature:0>FeatureHeader:3>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 4
+    # GherkinDocument:0>Feature:0>FeatureHeader:3>DescriptionHelper:2>#Comment:0
+    def match_token_at_5(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 5
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 6
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'FeatureHeader')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 5
+
+        state_comment = "State: 5 - GherkinDocument:0>Feature:0>FeatureHeader:3>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 5
+    # GherkinDocument:0>Feature:1>Background:0>#BackgroundLine:0
+    def match_token_at_6(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 6
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 8
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 7
+
+        state_comment = "State: 6 - GherkinDocument:0>Feature:1>Background:0>#BackgroundLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 6
+    # GherkinDocument:0>Feature:1>Background:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_7(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 8
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 7
+
+        state_comment = "State: 7 - GherkinDocument:0>Feature:1>Background:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 7
+    # GherkinDocument:0>Feature:1>Background:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_8(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 8
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 8
+
+        state_comment = "State: 8 - GherkinDocument:0>Feature:1>Background:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 8
+    # GherkinDocument:0>Feature:1>Background:2>Step:0>#StepLine:0
+    def match_token_at_9(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'DataTable')
+                self.build(context, token)
+                return 10
+        if self.match_DocStringSeparator(context, token):
+                self.start_rule(context, 'DocString')
+                self.build(context, token)
+                return 49
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 9
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 9
+
+        state_comment = "State: 9 - GherkinDocument:0>Feature:1>Background:2>Step:0>#StepLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#DocStringSeparator", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 9
+    # GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0
+    def match_token_at_10(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 10
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 10
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 10
+
+        state_comment = "State: 10 - GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 10
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:0>Tags:0>#TagLine:0
+    def match_token_at_11(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 11
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 11
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 11
+
+        state_comment = "State: 11 - GherkinDocument:0>Feature:2>ScenarioDefinition:0>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#ScenarioLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 11
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:0>#ScenarioLine:0
+    def match_token_at_12(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 12
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 14
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 13
+
+        state_comment = "State: 12 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:0>#ScenarioLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 12
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_13(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 14
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 13
+
+        state_comment = "State: 13 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 13
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_14(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 14
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 14
+
+        state_comment = "State: 14 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 14
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:0>#StepLine:0
+    def match_token_at_15(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'DataTable')
+                self.build(context, token)
+                return 16
+        if self.match_DocStringSeparator(context, token):
+                self.start_rule(context, 'DocString')
+                self.build(context, token)
+                return 47
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 15
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 15
+
+        state_comment = "State: 15 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:0>#StepLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#DocStringSeparator", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 15
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0
+    def match_token_at_16(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 16
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 16
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 16
+
+        state_comment = "State: 16 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 16
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:0>Tags:0>#TagLine:0
+    def match_token_at_17(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 17
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 17
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 17
+
+        state_comment = "State: 17 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:0>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#ExamplesLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 17
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:0>#ExamplesLine:0
+    def match_token_at_18(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 18
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 20
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 21
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 19
+
+        state_comment = "State: 18 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:0>#ExamplesLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 18
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_19(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 20
+        if self.match_TableRow(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 21
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 19
+
+        state_comment = "State: 19 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 19
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_20(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 20
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 21
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 20
+
+        state_comment = "State: 20 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 20
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:2>ExamplesTable:0>#TableRow:0
+    def match_token_at_21(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 21
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 21
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 21
+
+        state_comment = "State: 21 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:2>ExamplesTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 21
+    # GherkinDocument:0>Feature:3>Rule:0>RuleHeader:0>Tags:0>#TagLine:0
+    def match_token_at_22(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 22
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 22
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 22
+
+        state_comment = "State: 22 - GherkinDocument:0>Feature:3>Rule:0>RuleHeader:0>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 22
+    # GherkinDocument:0>Feature:3>Rule:0>RuleHeader:1>#RuleLine:0
+    def match_token_at_23(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 25
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 26
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 24
+
+        state_comment = "State: 23 - GherkinDocument:0>Feature:3>Rule:0>RuleHeader:1>#RuleLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 23
+    # GherkinDocument:0>Feature:3>Rule:0>RuleHeader:2>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_24(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 25
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 26
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 24
+
+        state_comment = "State: 24 - GherkinDocument:0>Feature:3>Rule:0>RuleHeader:2>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 24
+    # GherkinDocument:0>Feature:3>Rule:0>RuleHeader:2>DescriptionHelper:2>#Comment:0
+    def match_token_at_25(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 25
+        if self.match_BackgroundLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Background')
+                self.build(context, token)
+                return 26
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'RuleHeader')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 25
+
+        state_comment = "State: 25 - GherkinDocument:0>Feature:3>Rule:0>RuleHeader:2>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#BackgroundLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 25
+    # GherkinDocument:0>Feature:3>Rule:1>Background:0>#BackgroundLine:0
+    def match_token_at_26(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 26
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 28
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 27
+
+        state_comment = "State: 26 - GherkinDocument:0>Feature:3>Rule:1>Background:0>#BackgroundLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 26
+    # GherkinDocument:0>Feature:3>Rule:1>Background:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_27(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 28
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 27
+
+        state_comment = "State: 27 - GherkinDocument:0>Feature:3>Rule:1>Background:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 27
+    # GherkinDocument:0>Feature:3>Rule:1>Background:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_28(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 28
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 28
+
+        state_comment = "State: 28 - GherkinDocument:0>Feature:3>Rule:1>Background:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 28
+    # GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:0>#StepLine:0
+    def match_token_at_29(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'DataTable')
+                self.build(context, token)
+                return 30
+        if self.match_DocStringSeparator(context, token):
+                self.start_rule(context, 'DocString')
+                self.build(context, token)
+                return 45
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 29
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 29
+
+        state_comment = "State: 29 - GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:0>#StepLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#DocStringSeparator", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 29
+    # GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0
+    def match_token_at_30(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 30
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 30
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 30
+
+        state_comment = "State: 30 - GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 30
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:0>Tags:0>#TagLine:0
+    def match_token_at_31(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 31
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 31
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 31
+
+        state_comment = "State: 31 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:0>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#ScenarioLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 31
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:0>#ScenarioLine:0
+    def match_token_at_32(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 32
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 34
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 33
+
+        state_comment = "State: 32 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:0>#ScenarioLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 32
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_33(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 34
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 33
+
+        state_comment = "State: 33 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 33
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_34(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 34
+        if self.match_StepLine(context, token):
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 34
+
+        state_comment = "State: 34 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 34
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:0>#StepLine:0
+    def match_token_at_35(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'DataTable')
+                self.build(context, token)
+                return 36
+        if self.match_DocStringSeparator(context, token):
+                self.start_rule(context, 'DocString')
+                self.build(context, token)
+                return 43
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 35
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 35
+
+        state_comment = "State: 35 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:0>#StepLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#DocStringSeparator", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 35
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0
+    def match_token_at_36(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 36
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DataTable')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 36
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 36
+
+        state_comment = "State: 36 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:0>DataTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 36
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:0>Tags:0>#TagLine:0
+    def match_token_at_37(self, token, context):
+        if self.match_TagLine(context, token):
+                self.build(context, token)
+                return 37
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Tags')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 37
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 37
+
+        state_comment = "State: 37 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:0>Tags:0>#TagLine:0"
+        token.detach
+        expected_tokens = ["#TagLine", "#ExamplesLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 37
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:0>#ExamplesLine:0
+    def match_token_at_38(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 38
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 40
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 41
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.start_rule(context, 'Description')
+                self.build(context, token)
+                return 39
+
+        state_comment = "State: 38 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:0>#ExamplesLine:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Empty", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 38
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:1>Description:0>#Other:0
+    def match_token_at_39(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.end_rule(context, 'Description')
+                self.build(context, token)
+                return 40
+        if self.match_TableRow(context, token):
+                self.end_rule(context, 'Description')
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 41
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Description')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 39
+
+        state_comment = "State: 39 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:1>Description:0>#Other:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 39
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:2>#Comment:0
+    def match_token_at_40(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 40
+        if self.match_TableRow(context, token):
+                self.start_rule(context, 'ExamplesTable')
+                self.build(context, token)
+                return 41
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 40
+
+        state_comment = "State: 40 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:1>DescriptionHelper:2>#Comment:0"
+        token.detach
+        expected_tokens = ["#EOF", "#Comment", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 40
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:2>ExamplesTable:0>#TableRow:0
+    def match_token_at_41(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_TableRow(context, token):
+                self.build(context, token)
+                return 41
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'ExamplesTable')
+                self.end_rule(context, 'Examples')
+                self.end_rule(context, 'ExamplesDefinition')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 41
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 41
+
+        state_comment = "State: 41 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:3>ExamplesDefinition:1>Examples:2>ExamplesTable:0>#TableRow:0"
+        token.detach
+        expected_tokens = ["#EOF", "#TableRow", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 41
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0
+    def match_token_at_43(self, token, context):
+        if self.match_DocStringSeparator(context, token):
+                self.build(context, token)
+                return 44
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 43
+
+        state_comment = "State: 43 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#DocStringSeparator", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 43
+    # GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0
+    def match_token_at_44(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 35
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 37
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 38
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 44
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 44
+
+        state_comment = "State: 44 - GherkinDocument:0>Feature:3>Rule:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#EOF", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 44
+    # GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0
+    def match_token_at_45(self, token, context):
+        if self.match_DocStringSeparator(context, token):
+                self.build(context, token)
+                return 46
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 45
+
+        state_comment = "State: 45 - GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#DocStringSeparator", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 45
+    # GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0
+    def match_token_at_46(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 29
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 31
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 32
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Rule')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 46
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 46
+
+        state_comment = "State: 46 - GherkinDocument:0>Feature:3>Rule:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#EOF", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 46
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0
+    def match_token_at_47(self, token, context):
+        if self.match_DocStringSeparator(context, token):
+                self.build(context, token)
+                return 48
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 47
+
+        state_comment = "State: 47 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#DocStringSeparator", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 47
+    # GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0
+    def match_token_at_48(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 15
+        if self.match_TagLine(context, token):
+            if self.lookahead_1(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 17
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ExamplesLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'ExamplesDefinition')
+                self.start_rule(context, 'Examples')
+                self.build(context, token)
+                return 18
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Scenario')
+                self.end_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 48
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 48
+
+        state_comment = "State: 48 - GherkinDocument:0>Feature:2>ScenarioDefinition:1>Scenario:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#EOF", "#StepLine", "#TagLine", "#ExamplesLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 48
+    # GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0
+    def match_token_at_49(self, token, context):
+        if self.match_DocStringSeparator(context, token):
+                self.build(context, token)
+                return 50
+        if self.match_Other(context, token):
+                self.build(context, token)
+                return 49
+
+        state_comment = "State: 49 - GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:0>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#DocStringSeparator", "#Other"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 49
+    # GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0
+    def match_token_at_50(self, token, context):
+        if self.match_EOF(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.end_rule(context, 'Feature')
+                self.build(context, token)
+                return 42
+        if self.match_StepLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.start_rule(context, 'Step')
+                self.build(context, token)
+                return 9
+        if self.match_TagLine(context, token):
+            if self.lookahead_0(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 11
+        if self.match_TagLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.start_rule(context, 'Tags')
+                self.build(context, token)
+                return 22
+        if self.match_ScenarioLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'ScenarioDefinition')
+                self.start_rule(context, 'Scenario')
+                self.build(context, token)
+                return 12
+        if self.match_RuleLine(context, token):
+                self.end_rule(context, 'DocString')
+                self.end_rule(context, 'Step')
+                self.end_rule(context, 'Background')
+                self.start_rule(context, 'Rule')
+                self.start_rule(context, 'RuleHeader')
+                self.build(context, token)
+                return 23
+        if self.match_Comment(context, token):
+                self.build(context, token)
+                return 50
+        if self.match_Empty(context, token):
+                self.build(context, token)
+                return 50
+
+        state_comment = "State: 50 - GherkinDocument:0>Feature:1>Background:2>Step:1>StepArg:0>__alt0:1>DocString:2>#DocStringSeparator:0"
+        token.detach
+        expected_tokens = ["#EOF", "#StepLine", "#TagLine", "#ScenarioLine", "#RuleLine", "#Comment", "#Empty"]
+        error = UnexpectedEOFException(token, expected_tokens, state_comment) if token.eof() else UnexpectedTokenException(token, expected_tokens, state_comment)
+        if (self.stop_at_first_error):
+            raise error
+        self.add_error(context, error)
+        return 50
+    def lookahead_0(self, context, currentToken):
+        currentToken.detach
+        token = None
+        queue = []
+        match = False
+        while True:
+            token = self.read_token(context)
+            token.detach
+            queue.append(token)
+
+            if (self.match_ScenarioLine(context, token) or False):
+                match = True
+                break
+
+            if not (self.match_Empty(context, token) or self.match_Comment(context, token) or self.match_TagLine(context, token) or False):
+                break
+
+        context.token_queue.extend(queue)
+
+        return match
+    def lookahead_1(self, context, currentToken):
+        currentToken.detach
+        token = None
+        queue = []
+        match = False
+        while True:
+            token = self.read_token(context)
+            token.detach
+            queue.append(token)
+
+            if (self.match_ExamplesLine(context, token) or False):
+                match = True
+                break
+
+            if not (self.match_Empty(context, token) or self.match_Comment(context, token) or self.match_TagLine(context, token) or False):
+                break
+
+        context.token_queue.extend(queue)
+
+        return match
+    # private
+
+    def handle_ast_error(self, context, argument, action):
+        self.handle_external_error(context, True, argument, action)
+
+    def handle_external_error(self, context, default_value, argument, action):
+        if self.stop_at_first_error:
+            return action(argument)
+
+        try:
+            return action(argument)
+        except CompositeParserException as e:
+            for error in e.errors:
+                self.add_error(context, error)
+        except ParserException as e:
+            self.add_error(context, e)
+        return default_value
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__init__.py b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..681302c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/compiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/compiler.cpython-312.pyc
new file mode 100644
index 00000000..c18f3f64
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/__pycache__/compiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/pickles/compiler.py b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/compiler.py
new file mode 100644
index 00000000..07f5ca91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/pickles/compiler.py
@@ -0,0 +1,191 @@
+import re
+
+from ..count_symbols import count_symbols
+from ..stream.id_generator import IdGenerator
+
+
+class Compiler(object):
+    def __init__(self, id_generator=None):
+        self.id_generator = id_generator
+        if self.id_generator is None:
+            self.id_generator = IdGenerator()
+
+    def compile(self, gherkin_document):
+        pickles = []
+        if 'feature' not in gherkin_document:
+            return pickles
+
+        feature = gherkin_document['feature']
+        if not feature['children']:
+            return pickles
+
+        uri = gherkin_document['uri']
+        feature_tags = feature['tags']
+        language = feature['language']
+        background_steps = []
+        for scenario_definition in feature['children']:
+            if 'background' in scenario_definition:
+                if scenario_definition['background']['steps']:
+                    background_steps += scenario_definition['background']['steps']
+            elif 'rule' in scenario_definition:
+                self._compile_rule(uri, feature_tags, background_steps, scenario_definition['rule'], language, pickles)
+            else:
+                scenario = scenario_definition['scenario']
+                args = (uri, feature_tags, background_steps, scenario, language, pickles)
+                if not scenario['examples']:
+                    self._compile_scenario(*args)
+                else:
+                    self._compile_scenario_outline(*args)
+        return pickles
+
+    def _compile_rule(self, uri, feature_tags, feature_background_steps, rule, language, pickles):
+        tags = list(feature_tags) + list(rule['tags'])
+        background_steps = []
+        background_steps += feature_background_steps
+        for scenario_definition in rule['children']:
+            if 'background' in scenario_definition:
+                if scenario_definition['background']['steps']:
+                    background_steps += scenario_definition['background']['steps']
+            else:
+                scenario = scenario_definition['scenario']
+                args = (uri, tags, background_steps, scenario, language, pickles)
+                if not scenario['examples']:
+                    self._compile_scenario(*args)
+                else:
+                    self._compile_scenario_outline(*args)
+        return pickles
+
+    def _compile_scenario(self, uri, inherited_tags, background_steps, scenario, language, pickles):
+        tags = list(inherited_tags) + list(scenario['tags'])
+        last_keyword_type = 'Unknown'
+        steps = list()
+        if scenario['steps']:
+            for step in background_steps + scenario['steps']:
+                last_keyword_type = last_keyword_type if step['keywordType'] == 'Conjunction' else step['keywordType']
+                steps.append(self._pickle_step(step, last_keyword_type))
+
+        pickle = {
+            'astNodeIds': [scenario['id']],
+            'id': self.id_generator.get_next_id(),
+            'tags': self._pickle_tags(tags),
+            'name': scenario['name'],
+            'language': language,
+            'steps': steps,
+            'uri': uri
+        }
+        pickles.append(pickle)
+
+
+    def _compile_scenario_outline(self, uri, inherited_tags, background_steps, scenario, language, pickles):
+        for examples in (e for e in scenario['examples'] if 'tableHeader' in e):
+            variable_cells = examples['tableHeader']['cells']
+
+            for values in examples['tableBody']:
+                value_cells = values['cells']
+                tags = list(inherited_tags) + list(scenario['tags']) + list(examples['tags'])
+                last_keyword_type = None
+                steps = list()
+                if scenario['steps']:
+                    for step in background_steps:
+                        last_keyword_type = last_keyword_type if step['keywordType'] == 'Conjunction' else step['keywordType']
+                        steps.append(self._pickle_step(step, last_keyword_type))
+
+                if scenario['steps']:
+                    for outline_step in scenario['steps']:
+                        last_keyword_type = last_keyword_type if outline_step['keywordType'] == 'Conjunction' else outline_step['keywordType']
+                        step_text = self._interpolate(
+                            outline_step['text'],
+                            variable_cells,
+                            value_cells)
+                        argument = self._create_pickle_arguments(
+                            outline_step,
+                            variable_cells,
+                            value_cells)
+                        _pickle_step = {
+                            'astNodeIds': [outline_step['id'], values['id']],
+                            'id': self.id_generator.get_next_id(),
+                            'type': last_keyword_type,
+                            'text': step_text
+                        }
+                        if argument is not None:
+                            _pickle_step['argument'] = argument
+
+                        steps.append(_pickle_step)
+
+                pickle = {
+                    'astNodeIds': [scenario['id'], values['id']],
+                    'id': self.id_generator.get_next_id(),
+                    'name': self._interpolate(
+                        scenario['name'],
+                        variable_cells,
+                        value_cells),
+                    'language': language,
+                    'steps': steps,
+                    'tags': self._pickle_tags(tags),
+                    'uri': uri
+                }
+                pickles.append(pickle)
+
+    def _create_pickle_arguments(self, step, variables, values):
+        if 'dataTable' in step:
+            table = {'rows': []}
+            for row in step['dataTable']['rows']:
+                cells = [
+                    {
+                        'value': self._interpolate(cell['value'], variables, values)
+                    } for cell in row['cells']
+                ]
+                table['rows'].append({'cells': cells})
+            return {'dataTable': table}
+
+        elif 'docString' in step:
+            argument = step['docString']
+            docstring = {
+                'content': self._interpolate(argument['content'], variables, values)
+            }
+            if 'mediaType' in argument:
+                docstring['mediaType'] = self._interpolate(argument['mediaType'], variables, values)
+            return {'docString': docstring}
+        else:
+            return None
+
+    def _interpolate(self, name, variable_cells, value_cells):
+        if name is None:
+            return name
+
+        for n, variable_cell in enumerate(variable_cells):
+            value_cell = value_cells[n]
+            # For the case of trailing backslash, re-escaping backslashes are needed
+            reescaped_value = re.sub(r'\\', r'\\\\', value_cell['value'])
+            name = re.sub(
+                u'<{0[value]}>'.format(variable_cell),
+                reescaped_value,
+                name
+            )
+        return name
+
+    def _pickle_step(self, step, keyword_type):
+        pickle_step = {
+            'astNodeIds': [step['id']],
+            'id': self.id_generator.get_next_id(),
+            'type': keyword_type,
+            'text': step['text'],
+        }
+        argument = self._create_pickle_arguments(
+            step,
+            [],
+            [])
+        if argument is not None:
+            pickle_step['argument'] = argument
+        return pickle_step
+
+
+    def _pickle_tags(self, tags):
+        return [self._pickle_tag(tag) for tag in tags]
+
+
+    def _pickle_tag(self, tag):
+        return {
+            'astNodeId': tag['id'],
+            'name': tag['name']
+        }
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/__init__.py b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..dfe6c2e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/gherkin_events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/gherkin_events.cpython-312.pyc
new file mode 100644
index 00000000..c8f2c464
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/gherkin_events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/id_generator.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/id_generator.cpython-312.pyc
new file mode 100644
index 00000000..d794ea68
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/id_generator.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/source_events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/source_events.cpython-312.pyc
new file mode 100644
index 00000000..10ee30ff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/gherkin/stream/__pycache__/source_events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/gherkin_events.py b/backend/venv/lib/python3.12/site-packages/gherkin/stream/gherkin_events.py
new file mode 100644
index 00000000..23dcb15d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/stream/gherkin_events.py
@@ -0,0 +1,53 @@
+from gherkin.ast_builder import AstBuilder
+from gherkin.parser import Parser
+from gherkin.pickles.compiler import Compiler
+from gherkin.errors import ParserError, CompositeParserException
+from gherkin.stream.id_generator import IdGenerator
+
+def create_errors(errors, uri):
+    for error in errors:
+        yield {
+            'parseError': {
+            'source': {
+                'uri': uri,
+                'location': error.location
+            },
+            'message': str(error),
+        }}
+
+
+class GherkinEvents:
+    def __init__(self, options):
+        self.options = options
+        self.id_generator = IdGenerator()
+        self.parser = Parser(ast_builder=AstBuilder(self.id_generator))
+        self.compiler = Compiler(self.id_generator)
+
+    def enum(self, source_event):
+        uri = source_event['source']['uri']
+        source = source_event['source']['data']
+
+        try:
+            gherkin_document = self.parser.parse(source)
+            gherkin_document['uri'] = uri
+
+            if (self.options.print_source):
+                yield source_event
+
+            if (self.options.print_ast):
+                yield {
+                    'gherkinDocument': gherkin_document
+                }
+
+            if (self.options.print_pickles):
+                pickles = self.compiler.compile(gherkin_document)
+                for pickle in pickles:
+                    yield {
+                        'pickle': pickle
+                    }
+        except CompositeParserException as e:
+            for event in create_errors(e.errors, uri):
+                yield event
+        except ParserError as e:
+            for event in create_errors([e], uri):
+                yield event
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/id_generator.py b/backend/venv/lib/python3.12/site-packages/gherkin/stream/id_generator.py
new file mode 100644
index 00000000..f172cc8b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/stream/id_generator.py
@@ -0,0 +1,8 @@
+class IdGenerator(object):
+    def __init__(self):
+        self._id_counter = 0
+
+    def get_next_id(self):
+        id = self._id_counter
+        self._id_counter += 1
+        return str(id)
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/stream/source_events.py b/backend/venv/lib/python3.12/site-packages/gherkin/stream/source_events.py
new file mode 100644
index 00000000..946b6bc6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/stream/source_events.py
@@ -0,0 +1,18 @@
+import io
+
+def source_event(path):
+    event = {
+        'source': {
+            'uri': path,
+            'data': io.open(path, 'r', encoding='utf8', newline='').read(),
+            'mediaType': 'text/x.cucumber.gherkin+plain'
+        }
+    }
+    return event
+
+class SourceEvents:
+    def __init__(self, paths):
+        self.paths = paths
+
+    def enum(self):
+        return map(source_event, self.paths)
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/token.py b/backend/venv/lib/python3.12/site-packages/gherkin/token.py
new file mode 100644
index 00000000..4536ecb5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/token.py
@@ -0,0 +1,13 @@
+class Token(object):
+    def __init__(self, gherkin_line, location):
+        self.line = gherkin_line
+        self.location = location
+
+    def eof(self):
+        return not self.line
+
+    def detach(self):
+        pass  # TODO: detach line - is this needed?
+
+    def token_value(self):
+        return "EOF" if self.eof() else self.line.get_line_text()
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/token_formatter_builder.py b/backend/venv/lib/python3.12/site-packages/gherkin/token_formatter_builder.py
new file mode 100644
index 00000000..975a9f27
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/token_formatter_builder.py
@@ -0,0 +1,43 @@
+from .ast_builder import AstBuilder
+
+
+class TokenFormatterBuilder(AstBuilder):
+    def __init__(self):
+        self.reset()
+
+    def reset(self):
+        self._tokens = []
+
+    def build(self, token):
+        self._tokens.append(token)
+
+    def start_rule(self, rule_type):
+        pass
+
+    def end_rule(self, rule_type):
+        pass
+
+    def get_result(self):
+        return '\n'.join([self._format_token(token) for token in self._tokens])
+
+    def _format_token(self, token):
+        if token.eof():
+            return "EOF"
+
+        return ''.join([
+            '(',
+            str(token.location['line']), ':', str(token.location['column']),
+            ')',
+            token.matched_type,
+            ':',
+            ''.join([
+                '(',
+                token.matched_keyword_type if token.matched_keyword_type else '',
+                ')',
+                token.matched_keyword
+            ]) if token.matched_keyword else '',
+            '/',
+            (token.matched_text if token.matched_text else ""),
+            '/',
+            ','.join([str(item['column']) + ':' + item['text'] for item in token.matched_items])
+        ])
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher.py b/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher.py
new file mode 100644
index 00000000..265cf583
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher.py
@@ -0,0 +1,197 @@
+import re
+from collections import defaultdict
+from .dialect import Dialect
+from .errors import NoSuchLanguageException
+
+# Source: https://stackoverflow.com/a/8348914
+try:
+    import textwrap
+    textwrap.indent
+except AttributeError:  # undefined function (wasn't added until Python 3.3)
+    def indent(text, amount, ch=' '):
+        padding = amount * ch
+        return ''.join(padding+line for line in text.splitlines(True))
+else:
+    def indent(text, amount, ch=' '):
+        return textwrap.indent(text, amount * ch)
+
+class TokenMatcher(object):
+    LANGUAGE_RE = re.compile(r"^\s*#\s*language\s*:\s*([a-zA-Z\-_]+)\s*$")
+
+    def __init__(self, dialect_name='en'):
+        self._default_dialect_name = dialect_name
+        self._change_dialect(dialect_name)
+        self.reset()
+
+    def reset(self):
+        if self.dialect_name != self._default_dialect_name:
+            self._change_dialect(self._default_dialect_name)
+        self._indent_to_remove = 0
+        self._active_doc_string_separator = None
+
+    def match_FeatureLine(self, token):
+        return self._match_title_line(token, 'FeatureLine', self.dialect.feature_keywords)
+
+    def match_RuleLine(self, token):
+        return self._match_title_line(token, 'RuleLine', self.dialect.rule_keywords)
+
+    def match_ScenarioLine(self, token):
+        return (self._match_title_line(token, 'ScenarioLine', self.dialect.scenario_keywords) or
+                self._match_title_line(token, 'ScenarioLine',
+                                       self.dialect.scenario_outline_keywords))
+
+    def match_BackgroundLine(self, token):
+        return self._match_title_line(token, 'BackgroundLine', self.dialect.background_keywords)
+
+    def match_ExamplesLine(self, token):
+        return self._match_title_line(token, 'ExamplesLine', self.dialect.examples_keywords)
+
+    def match_TableRow(self, token):
+        if not token.line.startswith('|'):
+            return False
+        # TODO: indent
+        self._set_token_matched(token, 'TableRow', items=token.line.table_cells)
+        return True
+
+    def match_StepLine(self, token):
+        keywords = (self.dialect.given_keywords +
+                    self.dialect.when_keywords +
+                    self.dialect.then_keywords +
+                    self.dialect.and_keywords +
+                    self.dialect.but_keywords)
+        for keyword in (k for k in keywords if token.line.startswith(k)):
+            title = token.line.get_rest_trimmed(len(keyword))
+            keyword_types = self.keyword_types[keyword]
+            if len(keyword_types) == 1:
+                keyword_type = keyword_types[0]
+            else:
+                keyword_type = 'Unknown'
+            self._set_token_matched(token, 'StepLine', title, keyword, keyword_type=keyword_type)
+            return True
+
+        return False
+
+    def match_Comment(self, token):
+        if not token.line.startswith('#'):
+            return False
+
+        text = token.line._line_text  # take the entire line, including leading space
+        self._set_token_matched(token, 'Comment', text, indent=0)
+        return True
+
+    def match_Empty(self, token):
+        if not token.line.is_empty():
+            return False
+
+        self._set_token_matched(token, 'Empty', indent=0)
+        return True
+
+    def match_Language(self, token):
+        match = self.LANGUAGE_RE.match(token.line.get_line_text())
+        if not match:
+            return False
+
+        dialect_name = match.group(1)
+        self._set_token_matched(token, 'Language', dialect_name)
+        self._change_dialect(dialect_name, token.location)
+        return True
+
+    def match_TagLine(self, token):
+        if not token.line.startswith('@'):
+            return False
+
+        self._set_token_matched(token, 'TagLine', items=token.line.tags)
+        return True
+
+    def match_DocStringSeparator(self, token):
+        if not self._active_doc_string_separator:
+            # open
+            return (self._match_DocStringSeparator(token, '"""', True) or
+                    self._match_DocStringSeparator(token, '```', True))
+        else:
+            # close
+            return self._match_DocStringSeparator(token, self._active_doc_string_separator, False)
+
+    @staticmethod
+    def _default_docstring_content_type():
+        return None
+
+    def _match_DocStringSeparator(self, token, separator, is_open):
+        if not token.line.startswith(separator):
+            return False
+
+        content_type = self._default_docstring_content_type()
+        if is_open:
+            content_type = token.line.get_rest_trimmed(len(separator))
+            self._active_doc_string_separator = separator
+            self._indent_to_remove = token.line.indent
+        else:
+            self._active_doc_string_separator = None
+            self._indent_to_remove = 0
+
+        # TODO: Use the separator as keyword. That's needed for pretty printing.
+        self._set_token_matched(token, 'DocStringSeparator', content_type, separator)
+        return True
+
+    def match_Other(self, token):
+        # take the entire line, except removing DocString indents
+        text = token.line.get_line_text(self._indent_to_remove)
+        self._set_token_matched(token, 'Other', self._unescaped_docstring(text), indent=0)
+        return True
+
+    def match_EOF(self, token):
+        if not token.eof():
+            return False
+
+        self._set_token_matched(token, 'EOF')
+        return True
+
+    def _match_title_line(self, token, token_type, keywords):
+        for keyword in (k for k in keywords if token.line.startswith_title_keyword(k)):
+            title = token.line.get_rest_trimmed(len(keyword) + len(':'))
+            self._set_token_matched(token, token_type, title, keyword)
+            return True
+
+        return False
+
+    def _set_token_matched(self, token, matched_type, text=None,
+                           keyword=None, keyword_type=None, indent=None, items=None):
+        if items is None:
+            items = []
+        token.matched_type = matched_type
+        # text == '' should not result in None
+        token.matched_text = text.rstrip('\r\n') if text is not None else None
+        token.matched_keyword = keyword
+        token.matched_keyword_type = keyword_type
+        if indent is not None:
+            token.matched_indent = indent
+        else:
+            token.matched_indent = token.line.indent if token.line else 0
+        token.matched_items = items
+        token.location['column'] = token.matched_indent + 1
+        token.matched_gherkin_dialect = self.dialect_name
+
+    def _change_dialect(self, dialect_name, location=None):
+        dialect = Dialect.for_name(dialect_name)
+        if not dialect:
+            raise NoSuchLanguageException(dialect_name, location)
+
+        self.dialect_name = dialect_name
+        self.dialect = dialect
+        self.keyword_types = defaultdict(list)
+        for keyword in self.dialect.given_keywords:
+            self.keyword_types[keyword].append('Context')
+        for keyword in self.dialect.when_keywords:
+            self.keyword_types[keyword].append('Action')
+        for keyword in self.dialect.then_keywords:
+            self.keyword_types[keyword].append('Outcome')
+        for keyword in self.dialect.and_keywords + self.dialect.but_keywords:
+            self.keyword_types[keyword].append('Conjunction')
+
+    def _unescaped_docstring(self, text):
+        if self._active_doc_string_separator == '"""':
+            return text.replace('\\"\\"\\"', '"""')
+        elif self._active_doc_string_separator == '```':
+            return text.replace('\\`\\`\\`', '```')
+        else:
+            return text
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher_markdown.py b/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher_markdown.py
new file mode 100644
index 00000000..68f0737c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/token_matcher_markdown.py
@@ -0,0 +1,153 @@
+import re
+from .token_matcher import TokenMatcher
+
+KEYWORD_PREFIX_BULLET = '^(\\s*[*+-]\\s*)'
+KEYWORD_PREFIX_HEADER = '^(#{1,6}\\s)'
+
+class GherkinInMarkdownTokenMatcher(TokenMatcher):
+
+    def __init__(self, dialect_name='en'):
+        super(GherkinInMarkdownTokenMatcher, self).__init__(dialect_name)
+
+    def reset(self):
+        super(GherkinInMarkdownTokenMatcher, self).reset()
+        self.matched_feature_line=False
+
+    def match_FeatureLine(self, token):
+
+        if(self.matched_feature_line):
+            self._set_token_matched(token,None)
+
+        # We first try to match "# Feature: blah"
+        result = self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.feature_keywords, ':', token, 'FeatureLine')
+        # If we didn't match "# Feature: blah", we still match this line
+        # as a FeatureLine.
+        # The reason for this is that users may not want to be constrained by having this as their fist line.
+
+        if not result:
+            self._set_token_matched(token,'FeatureLine',token.line.get_line_text())
+        self.matched_feature_line=result
+        return result
+        
+
+
+    def match_RuleLine(self, token):
+        return self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.rule_keywords, ':', token, 'RuleLine')
+
+    def match_ScenarioLine(self, token):
+        return self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.scenario_keywords, ':', token, 'ScenarioLine') or self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.scenario_outline_keywords, ':', token, 'ScenarioLine')
+
+    def match_BackgroundLine(self, token):
+        return self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.background_keywords, ':', token, 'BackgroundLine')
+
+    def match_ExamplesLine(self, token):
+        return self._match_title_line(KEYWORD_PREFIX_HEADER, self.dialect.examples_keywords, ':', token, 'ExamplesLine')
+
+    def match_TableRow(self, token):
+        # Gherkin tables must be indented 2-5 spaces to be distinguishable from non-Gherkin tables
+
+        if re.match('^\\s\\s\\s?\\s?\\s?\\|',token.line.get_line_text(0)):
+            table_cells = token.line.table_cells
+            if(self._is_gfm_table_separator(table_cells)):
+                return False
+        
+            self._set_token_matched(token, 'TableRow', keyword='|',items=token.line.table_cells)
+
+            return True
+        return False
+
+    def _is_gfm_table_separator(self, table_cells):
+        text_of_table_cells = map(lambda x: x['text'], table_cells)
+        separator_values = list(filter(lambda x: re.match('^:?-+:?$',x),text_of_table_cells))
+        return len(separator_values) > 0
+
+
+    def match_StepLine(self, token):
+        nonStarStepKeywords = (self.dialect.given_keywords +
+                    self.dialect.when_keywords +
+                    self.dialect.then_keywords +
+                    self.dialect.and_keywords +
+                    self.dialect.but_keywords)
+        return self._match_title_line(KEYWORD_PREFIX_BULLET, nonStarStepKeywords, '', token, 'StepLine')
+        
+    def match_Comment(self, token):
+        if(token.line.startswith('|')):
+            table_cells = token.line.table_cells
+            if(self._is_gfm_table_separator(table_cells)):
+                return True
+        return self._set_token_matched(token,None,False)
+
+    def match_Empty(self, token):
+
+        result = False
+        if token.line.is_empty():
+            result = True
+        if ( not self.match_TagLine(token) and 
+             not self.match_FeatureLine(token) and
+             not self.match_ScenarioLine(token) and
+             not self.match_BackgroundLine(token) and
+             not self.match_ExamplesLine(token) and
+             not self.match_RuleLine(token) and
+             not self.match_TableRow(token) and
+             not self.match_Comment(token) and
+             not self.match_Language(token) and
+             not self.match_DocStringSeparator(token) and
+             not self.match_EOF(token) and
+             not self.match_StepLine(token)
+           ):
+            # neutered
+            result = True
+
+        if(result):
+            self._set_token_matched(token, 'Empty', indent=0)
+            return result
+        return False
+
+    # We've made a deliberate choice not to support `# language: [ISO 639-1]` headers or similar
+    # in Markdown. Users should specify a language globally. 
+    def match_Language(self, token):
+        if not token:
+            raise ValueError('no token')
+        return False
+
+    def match_TagLine(self, token):
+        tags = []
+        matching_tags = re.finditer('`(@[^`]+)`', token.line.get_line_text())
+        idx=0
+        for match in matching_tags:
+            tags.append({
+                'column': token.line.indent + match.start(idx) + 2,
+                'text': match.group(1)
+            })
+
+        if(len(tags) == 0):
+            return False
+
+        self._set_token_matched(token, 'TagLine', items=tags)
+        return True
+
+    def match_DocStringSeparator(self, token):
+        if not self._active_doc_string_separator:
+            # open
+            return (self._match_DocStringSeparator(token, '"""', True) or
+                    self._match_DocStringSeparator(token, '````', True) or self._match_DocStringSeparator(token, '```', True))
+        else:
+            # close
+            return self._match_DocStringSeparator(token, self._active_doc_string_separator, False)
+
+    @staticmethod
+    def _default_docstring_content_type():
+        return ''
+
+    def _match_title_line(self, prefix, keywords, keywordSuffix, token, token_type):
+        
+        keywords_or_list="|".join(map(lambda x: re.escape(x), keywords))
+        match = re.search(u'{}({}){}(.*)'.format(prefix, keywords_or_list, keywordSuffix), token.line.get_line_text())
+        indent = token.line.indent
+        
+        if(match):
+            matchedKeyword = match.group(2)
+            indent += len(match.group(1))
+            self._set_token_matched(token, token_type, match.group(3).strip(), matchedKeyword, indent=indent)
+            return True
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin/token_scanner.py b/backend/venv/lib/python3.12/site-packages/gherkin/token_scanner.py
new file mode 100644
index 00000000..531b51b2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin/token_scanner.py
@@ -0,0 +1,44 @@
+import io
+import os
+import sys
+from .token import Token
+from .gherkin_line import GherkinLine
+
+
+class TokenScanner(object):
+    """
+    The scanner reads a gherkin doc (typically read from a `.feature` file) and creates a token for
+    each line.
+
+    The tokens are passed to the parser, which outputs an AST (Abstract Syntax Tree).
+
+    If the scanner sees a `#` language header, it will reconfigure itself dynamically to look for
+    Gherkin keywords for the associated language. The keywords are defined in
+    :file:`gherkin-languages.json`.
+    """
+
+    def __init__(self, path_or_str):
+        if os.path.exists(path_or_str):
+            self.io = io.open(path_or_str, 'r', encoding='utf8')
+        else:
+            if sys.version_info < (3, 0):
+                if isinstance(path_or_str, str):
+                    self.io = io.StringIO(unicode(path_or_str, encoding='utf8'))
+                else:
+                    self.io = io.StringIO(path_or_str)
+            else:
+                self.io = io.StringIO(path_or_str)
+        self.line_number = 0
+
+    def read(self):
+        self.line_number += 1
+        location = {'line': self.line_number}
+        line = self.io.readline()
+        return Token((GherkinLine(line, self.line_number) if line else line), location)
+
+    def __del__(self):
+        # close file descriptor if it's still open
+        try:
+            self.io.close()
+        except AttributeError:
+            pass
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/LICENSE
new file mode 100644
index 00000000..58780c68
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2017 Cucumber Ltd, Gaspar Nagy, Björn Rasmusson, Peter Sergeant, and contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/METADATA
new file mode 100644
index 00000000..eb895b05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/METADATA
@@ -0,0 +1,17 @@
+Metadata-Version: 2.1
+Name: gherkin-official
+Version: 29.0.0
+Summary: Gherkin parser (official, by Cucumber team)
+Home-page: https://github.com/cucumber/gherkin
+Download-URL: http://pypi.python.org/pypi/gherkin-official
+Author: Cucumber Ltd and Björn Rasmusson
+Author-email: cukes@googlegroups.com
+License: MIT
+Keywords: gherkin,cucumber,bdd
+Platform: any
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 2
+Classifier: Programming Language :: Python :: 3
+License-File: LICENSE
+
+Gherkin parser (official, by Cucumber team)
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/RECORD
new file mode 100644
index 00000000..f1d14d99
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/RECORD
@@ -0,0 +1,53 @@
+../../../bin/gherkin,sha256=vcYVq87_FUJ8-9F98fBZjepA_GPKUwOhBFIBrS188ck,529
+gherkin/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gherkin/__main__.py,sha256=-JQHYEAGW5prmfgkIpi2LmE_l3pzW64-2IvnRDGQKKc,1125
+gherkin/__pycache__/__init__.cpython-312.pyc,,
+gherkin/__pycache__/__main__.cpython-312.pyc,,
+gherkin/__pycache__/ast_builder.cpython-312.pyc,,
+gherkin/__pycache__/ast_node.cpython-312.pyc,,
+gherkin/__pycache__/count_symbols.cpython-312.pyc,,
+gherkin/__pycache__/count_symbols_py3_plus.cpython-312.pyc,,
+gherkin/__pycache__/dialect.cpython-312.pyc,,
+gherkin/__pycache__/errors.cpython-312.pyc,,
+gherkin/__pycache__/gherkin_line.cpython-312.pyc,,
+gherkin/__pycache__/inout.cpython-312.pyc,,
+gherkin/__pycache__/parser.cpython-312.pyc,,
+gherkin/__pycache__/token.cpython-312.pyc,,
+gherkin/__pycache__/token_formatter_builder.cpython-312.pyc,,
+gherkin/__pycache__/token_matcher.cpython-312.pyc,,
+gherkin/__pycache__/token_matcher_markdown.cpython-312.pyc,,
+gherkin/__pycache__/token_scanner.cpython-312.pyc,,
+gherkin/ast_builder.py,sha256=qhfq54gE9dVNHV53_nKLOCYdnZJxpMBmzZWnqv5LR3o,9932
+gherkin/ast_node.py,sha256=i-BE-tztnyjvapSABjNXt63q6Qyhc-lYSQ-azb9dXnM,653
+gherkin/count_symbols.py,sha256=hNEy9osGUYuP4sRPGhgefXoBv3A2tZ030gFnp-FxEL8,150
+gherkin/count_symbols_py2.py,sha256=B2zTqkEDwOhVkLkBHqVxAgLUyrxw2fvTP0AZ27FcOkg,175
+gherkin/count_symbols_py3_plus.py,sha256=lAxIn-Vc6XI9njfL6OWQCXehr7lvQObOSnSH3CLJbSs,50
+gherkin/dialect.py,sha256=9alHsNWqFjA0TcZL-8AevMlylEeiGVG61LFa66SKtm0,1320
+gherkin/errors.py,sha256=lqmAH8vQY2Myf4l_dWP-DMgpfWtvCU_J3CIY5Y_etQI,1996
+gherkin/gherkin-languages.json,sha256=jfIV8SA1xNmKilyz1ZfR9Xa88vU-Vvb76Y_jczaRdag,58509
+gherkin/gherkin_line.py,sha256=bOqn9GtmiG1LSSohYBXPmGeouYMCVcAmYmsfCrEu0I4,3094
+gherkin/inout.py,sha256=8tpsa1qE_hshVWfWM3BnTProYqgQJPeuK6FJzFBzqrk,2102
+gherkin/parser.py,sha256=DGDQ9gpHI7UJnBQ5puUFmgFtThE1BPtiDe2bIEo9Zeo,161201
+gherkin/pickles/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gherkin/pickles/__pycache__/__init__.cpython-312.pyc,,
+gherkin/pickles/__pycache__/compiler.cpython-312.pyc,,
+gherkin/pickles/compiler.py,sha256=MMKeIHjmZbjXMlqvOj9EhsVcq0hDp8J_9zqDY9qtDmM,7713
+gherkin/stream/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gherkin/stream/__pycache__/__init__.cpython-312.pyc,,
+gherkin/stream/__pycache__/gherkin_events.cpython-312.pyc,,
+gherkin/stream/__pycache__/id_generator.cpython-312.pyc,,
+gherkin/stream/__pycache__/source_events.cpython-312.pyc,,
+gherkin/stream/gherkin_events.py,sha256=Emx-pLBWloYX4Thm2saTR0bvOj-mZ3CSW9CnGarcb3g,1693
+gherkin/stream/id_generator.py,sha256=lD9zb9XXkCPuGrHpa3BTxk-yvi2G0jrVLmrWIya8mQk,191
+gherkin/stream/source_events.py,sha256=zDEgS7vhpfRdABFvDF0B0-vWPCt-FWiqiTkDDz8KuBY,405
+gherkin/token.py,sha256=q86WUR-sCjmAnnhS5_sEXGK9pouFApf4O2wh0kvxkGA,353
+gherkin/token_formatter_builder.py,sha256=65UvAiVzQOO2DTDnL5Cjd4GqdrPXn-iMI3S9S-Ocns4,1154
+gherkin/token_matcher.py,sha256=rgS6G5cd-jQikx7SBwTmYyO-rOWFJwc1BEdkU34MqO8,7615
+gherkin/token_matcher_markdown.py,sha256=tiOuJ-0yEfUAfeU8F1YRfMOP7S8nmfTEGYYaBxti4zY,6090
+gherkin/token_scanner.py,sha256=hlhXs4t0CXz9vxIZAwzQT2u_VqdeGArpllFCEKZRETw,1442
+gherkin_official-29.0.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+gherkin_official-29.0.0.dist-info/LICENSE,sha256=BCvaKILXk8QvzV9NJyV5pDLR1SoOVDYgLl6CPq5HcbY,1134
+gherkin_official-29.0.0.dist-info/METADATA,sha256=U-aHkTcoT7BOBog50NQOFWda2cp4gel8Twb4wHBuihI,563
+gherkin_official-29.0.0.dist-info/RECORD,,
+gherkin_official-29.0.0.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
+gherkin_official-29.0.0.dist-info/top_level.txt,sha256=blDdZiMARjjfHuj8eqwYuzW0M_M_s44LNxU6eSXxuyI,8
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/WHEEL
new file mode 100644
index 00000000..1a9c5358
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (72.1.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/top_level.txt
new file mode 100644
index 00000000..1a503021
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/gherkin_official-29.0.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+gherkin
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/AUTHORS b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/AUTHORS
new file mode 100644
index 00000000..42a5c227
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/AUTHORS
@@ -0,0 +1,51 @@
+Original Authors
+----------------
+* Armin Rigo
+* Christian Tismer
+
+Contributors
+------------
+* Al Stone
+* Alexander Schmidt
+* Alexey Borzenkov
+* Andreas Schwab
+* Armin Ronacher
+* Bin Wang <feisuzhu@163.com>
+* Bob Ippolito
+* ChangBo Guo
+* Christoph Gohlke
+* Denis Bilenko
+* Dirk Mueller
+* Donovan Preston
+* Fantix King
+* Floris Bruynooghe
+* Fredrik Fornwall
+* Gerd Woetzel
+* Giel van Schijndel
+* Gökhan Karabulut
+* Gustavo Niemeyer
+* Guy Rozendorn
+* Hye-Shik Chang
+* Jared Kuolt
+* Jason Madden
+* Josh Snyder
+* Kyle Ambroff
+* Laszlo Boszormenyi
+* Mao Han
+* Marc Abramowitz
+* Marc Schlaich
+* Marcin Bachry
+* Matt Madison
+* Matt Turner
+* Michael Ellerman
+* Michael Matz
+* Ralf Schmitt
+* Robie Basak
+* Ronny Pfannschmidt
+* Samual M. Rushing
+* Tony Bowles
+* Tony Breeds
+* Trevor Bowen
+* Tulio Magno Quites Machado Filho
+* Ulrich Weigand
+* Victor Stinner
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE
new file mode 100644
index 00000000..b73a4a10
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE
@@ -0,0 +1,30 @@
+The following files are derived from Stackless Python and are subject to the
+same license as Stackless Python:
+
+	src/greenlet/slp_platformselect.h
+	files in src/greenlet/platform/ directory
+
+See LICENSE.PSF and http://www.stackless.com/ for details.
+
+Unless otherwise noted, the files in greenlet have been released under the
+following MIT license:
+
+Copyright (c) Armin Rigo, Christian Tismer and contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE.PSF b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE.PSF
new file mode 100644
index 00000000..d3b509a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/LICENSE.PSF
@@ -0,0 +1,47 @@
+PYTHON SOFTWARE FOUNDATION LICENSE VERSION 2
+--------------------------------------------
+
+1. This LICENSE AGREEMENT is between the Python Software Foundation
+("PSF"), and the Individual or Organization ("Licensee") accessing and
+otherwise using this software ("Python") in source or binary form and
+its associated documentation.
+
+2. Subject to the terms and conditions of this License Agreement, PSF hereby
+grants Licensee a nonexclusive, royalty-free, world-wide license to reproduce,
+analyze, test, perform and/or display publicly, prepare derivative works,
+distribute, and otherwise use Python alone or in any derivative version,
+provided, however, that PSF's License Agreement and PSF's notice of copyright,
+i.e., "Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
+2011 Python Software Foundation; All Rights Reserved" are retained in Python
+alone or in any derivative version prepared by Licensee.
+
+3. In the event Licensee prepares a derivative work that is based on
+or incorporates Python or any part thereof, and wants to make
+the derivative work available to others as provided herein, then
+Licensee hereby agrees to include in any such work a brief summary of
+the changes made to Python.
+
+4. PSF is making Python available to Licensee on an "AS IS"
+basis.  PSF MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, PSF MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF PYTHON WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+5. PSF SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF PYTHON
+FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS AS
+A RESULT OF MODIFYING, DISTRIBUTING, OR OTHERWISE USING PYTHON,
+OR ANY DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+6. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+7. Nothing in this License Agreement shall be deemed to create any
+relationship of agency, partnership, or joint venture between PSF and
+Licensee.  This License Agreement does not grant permission to use PSF
+trademarks or trade name in a trademark sense to endorse or promote
+products or services of Licensee, or any third party.
+
+8. By copying, installing or otherwise using Python, Licensee
+agrees to be bound by the terms and conditions of this License
+Agreement.
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/METADATA
new file mode 100644
index 00000000..15294101
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/METADATA
@@ -0,0 +1,103 @@
+Metadata-Version: 2.1
+Name: greenlet
+Version: 3.1.1
+Summary: Lightweight in-process concurrent programming
+Home-page: https://greenlet.readthedocs.io/
+Author: Alexey Borzenkov
+Author-email: snaury@gmail.com
+Maintainer: Jason Madden
+Maintainer-email: jason@seecoresoftware.com
+License: MIT License
+Project-URL: Bug Tracker, https://github.com/python-greenlet/greenlet/issues
+Project-URL: Source Code, https://github.com/python-greenlet/greenlet/
+Project-URL: Documentation, https://greenlet.readthedocs.io/
+Keywords: greenlet coroutine concurrency threads cooperative
+Platform: any
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Natural Language :: English
+Classifier: Programming Language :: C
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.7
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+License-File: LICENSE.PSF
+License-File: AUTHORS
+Provides-Extra: docs
+Requires-Dist: Sphinx ; extra == 'docs'
+Requires-Dist: furo ; extra == 'docs'
+Provides-Extra: test
+Requires-Dist: objgraph ; extra == 'test'
+Requires-Dist: psutil ; extra == 'test'
+
+.. This file is included into docs/history.rst
+
+
+Greenlets are lightweight coroutines for in-process concurrent
+programming.
+
+The "greenlet" package is a spin-off of `Stackless`_, a version of
+CPython that supports micro-threads called "tasklets". Tasklets run
+pseudo-concurrently (typically in a single or a few OS-level threads)
+and are synchronized with data exchanges on "channels".
+
+A "greenlet", on the other hand, is a still more primitive notion of
+micro-thread with no implicit scheduling; coroutines, in other words.
+This is useful when you want to control exactly when your code runs.
+You can build custom scheduled micro-threads on top of greenlet;
+however, it seems that greenlets are useful on their own as a way to
+make advanced control flow structures. For example, we can recreate
+generators; the difference with Python's own generators is that our
+generators can call nested functions and the nested functions can
+yield values too. (Additionally, you don't need a "yield" keyword. See
+the example in `test_generator.py
+<https://github.com/python-greenlet/greenlet/blob/adca19bf1f287b3395896a8f41f3f4fd1797fdc7/src/greenlet/tests/test_generator.py#L1>`_).
+
+Greenlets are provided as a C extension module for the regular unmodified
+interpreter.
+
+.. _`Stackless`: http://www.stackless.com
+
+
+Who is using Greenlet?
+======================
+
+There are several libraries that use Greenlet as a more flexible
+alternative to Python's built in coroutine support:
+
+ - `Concurrence`_
+ - `Eventlet`_
+ - `Gevent`_
+
+.. _Concurrence: http://opensource.hyves.org/concurrence/
+.. _Eventlet: http://eventlet.net/
+.. _Gevent: http://www.gevent.org/
+
+Getting Greenlet
+================
+
+The easiest way to get Greenlet is to install it with pip::
+
+  pip install greenlet
+
+
+Source code archives and binary distributions are available on the
+python package index at https://pypi.org/project/greenlet
+
+The source code repository is hosted on github:
+https://github.com/python-greenlet/greenlet
+
+Documentation is available on readthedocs.org:
+https://greenlet.readthedocs.io
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/RECORD
new file mode 100644
index 00000000..e4dda9b4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/RECORD
@@ -0,0 +1,123 @@
+../../../include/site/python3.12/greenlet/greenlet.h,sha256=sz5pYRSQqedgOt2AMgxLZdTjO-qcr_JMvgiEJR9IAJ8,4755
+greenlet-3.1.1.dist-info/AUTHORS,sha256=swW28t2knVRxRkaEQNZtO7MP9Sgnompb7B6cNgJM8Gk,849
+greenlet-3.1.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+greenlet-3.1.1.dist-info/LICENSE,sha256=dpgx1uXfrywggC-sz_H6-0wgJd2PYlPfpH_K1Z1NCXk,1434
+greenlet-3.1.1.dist-info/LICENSE.PSF,sha256=5f88I8EQ5JTNfXNsEP2W1GJFe6_soxCEDbZScpjH1Gs,2424
+greenlet-3.1.1.dist-info/METADATA,sha256=MoOb0T5ZdGbCymjJygO-CkVIAaeI5KcuwE_JPpH1hp4,3830
+greenlet-3.1.1.dist-info/RECORD,,
+greenlet-3.1.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+greenlet-3.1.1.dist-info/WHEEL,sha256=TYidmGtjaed4MWwiHADTxoCZc9qRgcbwfa4cVq3D8vA,114
+greenlet-3.1.1.dist-info/top_level.txt,sha256=YSnRsCRoO61JGlP57o8iKL6rdLWDWuiyKD8ekpWUsDc,9
+greenlet/CObjects.cpp,sha256=OPej1bWBgc4sRrTRQ2aFFML9pzDYKlKhlJSjsI0X_eU,3508
+greenlet/PyGreenlet.cpp,sha256=ogWsQ5VhSdItWRLLpWOgSuqYuM3QwQ4cVCxOQIgHx6E,23441
+greenlet/PyGreenlet.hpp,sha256=2ZQlOxYNoy7QwD7mppFoOXe_At56NIsJ0eNsE_hoSsw,1463
+greenlet/PyGreenletUnswitchable.cpp,sha256=PQE0fSZa_IOyUM44IESHkJoD2KtGW3dkhkmZSYY3WHs,4375
+greenlet/PyModule.cpp,sha256=J2TH06dGcNEarioS6NbWXkdME8hJY05XVbdqLrfO5w4,8587
+greenlet/TBrokenGreenlet.cpp,sha256=smN26uC7ahAbNYiS10rtWPjCeTG4jevM8siA2sjJiXg,1021
+greenlet/TExceptionState.cpp,sha256=U7Ctw9fBdNraS0d174MoQW7bN-ae209Ta0JuiKpcpVI,1359
+greenlet/TGreenlet.cpp,sha256=HGYGKpmKYqQ842tASW-QaaV8wua4a5XV_quYKPDsV_Y,25731
+greenlet/TGreenlet.hpp,sha256=mMHcb_rSuozdDiGJjX3GgyYkWgVM4kuO1UgbUP84BlU,27869
+greenlet/TGreenletGlobals.cpp,sha256=YyEmDjKf1g32bsL-unIUScFLnnA1fzLWf2gOMd-D0Zw,3264
+greenlet/TMainGreenlet.cpp,sha256=fvgb8HHB-FVTPEKjR1s_ifCZSpp5D5YQByik0CnIABg,3276
+greenlet/TPythonState.cpp,sha256=FxRdi76lTGXaQKWwkq82VaCfIRdF2Z-fh-TlRTMjYqg,15359
+greenlet/TStackState.cpp,sha256=V444I8Jj9DhQz-9leVW_9dtiSRjaE1NMlgDG02Xxq-Y,7381
+greenlet/TThreadState.hpp,sha256=2Jgg7DtGggMYR_x3CLAvAFf1mIdIDtQvSSItcdmX4ZQ,19131
+greenlet/TThreadStateCreator.hpp,sha256=uYTexDWooXSSgUc5uh-Mhm5BQi3-kR6CqpizvNynBFQ,2610
+greenlet/TThreadStateDestroy.cpp,sha256=wt7lQwLI0mi_JtnZB_jB4bUmfCa5b6nQhA7XOmnI1yk,9568
+greenlet/TUserGreenlet.cpp,sha256=uemg0lwKXtYB0yzmvyYdIIAsKnNkifXM1OJ2OlrFP1A,23553
+greenlet/__init__.py,sha256=OOmvT6_vn_SekdPzkj4qm6hjfikXMmdNZYDmGTOaRNo,1723
+greenlet/__pycache__/__init__.cpython-312.pyc,,
+greenlet/_greenlet.cpython-312-darwin.so,sha256=B0OZoDh_PaJNAAsjF5StySigXwgUcG50q_ApgucNMEU,280928
+greenlet/greenlet.cpp,sha256=WdItb1yWL9WNsTqJNf0Iw8ZwDHD49pkDP0rIRGBg2pw,10996
+greenlet/greenlet.h,sha256=sz5pYRSQqedgOt2AMgxLZdTjO-qcr_JMvgiEJR9IAJ8,4755
+greenlet/greenlet_allocator.hpp,sha256=kxyWW4Qdwlrc7ufgdb5vd6Y7jhauQ699Kod0mqiO1iM,1582
+greenlet/greenlet_compiler_compat.hpp,sha256=nRxpLN9iNbnLVyFDeVmOwyeeNm6scQrOed1l7JQYMCM,4346
+greenlet/greenlet_cpython_add_pending.hpp,sha256=apAwIhGlgYrnYn03zWL6Sxy68kltDeb1e0QupZfb3DQ,6043
+greenlet/greenlet_cpython_compat.hpp,sha256=L_jig3dm2bsJWRazrhlokma2NfnwixoQ0cydshh6ce4,3964
+greenlet/greenlet_exceptions.hpp,sha256=06Bx81DtVaJTa6RtiMcV141b-XHv4ppEgVItkblcLWY,4503
+greenlet/greenlet_internal.hpp,sha256=Ajc-_09W4xWzm9XfyXHAeQAFUgKGKsnJwYsTCoNy3ns,2709
+greenlet/greenlet_refs.hpp,sha256=OnbA91yZf3QHH6-eJccvoNDAaN-pQBMMrclFU1Ot3J4,34436
+greenlet/greenlet_slp_switch.hpp,sha256=kM1QHA2iV-gH4cFyN6lfIagHQxvJZjWOVJdIxRE3TlQ,3198
+greenlet/greenlet_thread_support.hpp,sha256=XUJ6ljWjf9OYyuOILiz8e_yHvT3fbaUiHdhiPNQUV4s,867
+greenlet/platform/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+greenlet/platform/__pycache__/__init__.cpython-312.pyc,,
+greenlet/platform/setup_switch_x64_masm.cmd,sha256=ZpClUJeU0ujEPSTWNSepP0W2f9XiYQKA8QKSoVou8EU,143
+greenlet/platform/switch_aarch64_gcc.h,sha256=GKC0yWNXnbK2X--X6aguRCMj2Tg7hDU1Zkl3RljDvC8,4307
+greenlet/platform/switch_alpha_unix.h,sha256=Z-SvF8JQV3oxWT8JRbL9RFu4gRFxPdJ7cviM8YayMmw,671
+greenlet/platform/switch_amd64_unix.h,sha256=EcSFCBlodEBhqhKjcJqY_5Dn_jn7pKpkJlOvp7gFXLI,2748
+greenlet/platform/switch_arm32_gcc.h,sha256=Z3KkHszdgq6uU4YN3BxvKMG2AdDnovwCCNrqGWZ1Lyo,2479
+greenlet/platform/switch_arm32_ios.h,sha256=mm5_R9aXB92hyxzFRwB71M60H6AlvHjrpTrc72Pz3l8,1892
+greenlet/platform/switch_arm64_masm.asm,sha256=4kpTtfy7rfcr8j1CpJLAK21EtZpGDAJXWRU68HEy5A8,1245
+greenlet/platform/switch_arm64_masm.obj,sha256=DmLnIB_icoEHAz1naue_pJPTZgR9ElM7-Nmztr-o9_U,746
+greenlet/platform/switch_arm64_msvc.h,sha256=RqK5MHLmXI3Q-FQ7tm32KWnbDNZKnkJdq8CR89cz640,398
+greenlet/platform/switch_csky_gcc.h,sha256=kDikyiPpewP71KoBZQO_MukDTXTXBiC7x-hF0_2DL0w,1331
+greenlet/platform/switch_loongarch64_linux.h,sha256=7M-Dhc4Q8tRbJCJhalDLwU6S9Mx8MjmN1RbTDgIvQTM,779
+greenlet/platform/switch_m68k_gcc.h,sha256=VSa6NpZhvyyvF-Q58CTIWSpEDo4FKygOyTz00whctlw,928
+greenlet/platform/switch_mips_unix.h,sha256=E0tYsqc5anDY1BhenU1l8DW-nVHC_BElzLgJw3TGtPk,1426
+greenlet/platform/switch_ppc64_aix.h,sha256=_BL0iyRr3ZA5iPlr3uk9SJ5sNRWGYLrXcZ5z-CE9anE,3860
+greenlet/platform/switch_ppc64_linux.h,sha256=0rriT5XyxPb0GqsSSn_bP9iQsnjsPbBmu0yqo5goSyQ,3815
+greenlet/platform/switch_ppc_aix.h,sha256=pHA4slEjUFP3J3SYm1TAlNPhgb2G_PAtax5cO8BEe1A,2941
+greenlet/platform/switch_ppc_linux.h,sha256=YwrlKUzxlXuiKMQqr6MFAV1bPzWnmvk6X1AqJZEpOWU,2759
+greenlet/platform/switch_ppc_macosx.h,sha256=Z6KN_ud0n6nC3ltJrNz2qtvER6vnRAVRNH9mdIDpMxY,2624
+greenlet/platform/switch_ppc_unix.h,sha256=-ZG7MSSPEA5N4qO9PQChtyEJ-Fm6qInhyZm_ZBHTtMg,2652
+greenlet/platform/switch_riscv_unix.h,sha256=Xg0wBen8Je21LWzFtLNLvUUYq6p9n_WY7AUQbiBVyyk,865
+greenlet/platform/switch_s390_unix.h,sha256=RRlGu957ybmq95qNNY4Qw1mcaoT3eBnW5KbVwu48KX8,2763
+greenlet/platform/switch_sh_gcc.h,sha256=mcRJBTu-2UBf4kZtX601qofwuDuy-Y-hnxJtrcaB7do,901
+greenlet/platform/switch_sparc_sun_gcc.h,sha256=xZish9GsMHBienUbUMsX1-ZZ-as7hs36sVhYIE3ew8Y,2797
+greenlet/platform/switch_x32_unix.h,sha256=nM98PKtzTWc1lcM7TRMUZJzskVdR1C69U1UqZRWX0GE,1509
+greenlet/platform/switch_x64_masm.asm,sha256=nu6n2sWyXuXfpPx40d9YmLfHXUc1sHgeTvX1kUzuvEM,1841
+greenlet/platform/switch_x64_masm.obj,sha256=GNtTNxYdo7idFUYsQv-mrXWgyT5EJ93-9q90lN6svtQ,1078
+greenlet/platform/switch_x64_msvc.h,sha256=LIeasyKo_vHzspdMzMHbosRhrBfKI4BkQOh4qcTHyJw,1805
+greenlet/platform/switch_x86_msvc.h,sha256=TtGOwinbFfnn6clxMNkCz8i6OmgB6kVRrShoF5iT9to,12838
+greenlet/platform/switch_x86_unix.h,sha256=VplW9H0FF0cZHw1DhJdIUs5q6YLS4cwb2nYwjF83R1s,3059
+greenlet/slp_platformselect.h,sha256=s-U-BrZ3qwwfI-6W9zWw2rb404OksZYbxYC2w5kSMXM,3841
+greenlet/tests/__init__.py,sha256=cj2-qpMXnlVRLbMLX-rPNNMVJ42ZssdxHd84NSQ3YXw,9246
+greenlet/tests/__pycache__/__init__.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_clearing_run_switches.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_cpp_exception.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_initialstub_already_started.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_slp_switch.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_switch_three_greenlets.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_switch_three_greenlets2.cpython-312.pyc,,
+greenlet/tests/__pycache__/fail_switch_two_greenlets.cpython-312.pyc,,
+greenlet/tests/__pycache__/leakcheck.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_contextvars.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_cpp.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_extension_interface.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_gc.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_generator.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_generator_nested.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_greenlet.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_greenlet_trash.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_leaks.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_stack_saved.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_throw.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_tracing.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_version.cpython-312.pyc,,
+greenlet/tests/__pycache__/test_weakref.cpython-312.pyc,,
+greenlet/tests/_test_extension.c,sha256=vkeGA-6oeJcGILsD7oIrT1qZop2GaTOHXiNT7mcSl-0,5773
+greenlet/tests/_test_extension.cpython-312-darwin.so,sha256=CK5RZ9bJhGDQ6-12FLNroVmqX8OsANBurDdSSI6plpQ,68488
+greenlet/tests/_test_extension_cpp.cpp,sha256=e0kVnaB8CCaEhE9yHtNyfqTjevsPDKKx-zgxk7PPK48,6565
+greenlet/tests/_test_extension_cpp.cpython-312-darwin.so,sha256=lrVcjTbuYux0s3uXfRWxWt8r5gjNw9ci763OeZc70qc,69312
+greenlet/tests/fail_clearing_run_switches.py,sha256=o433oA_nUCtOPaMEGc8VEhZIKa71imVHXFw7TsXaP8M,1263
+greenlet/tests/fail_cpp_exception.py,sha256=o_ZbipWikok8Bjc-vjiQvcb5FHh2nVW-McGKMLcMzh0,985
+greenlet/tests/fail_initialstub_already_started.py,sha256=txENn5IyzGx2p-XR1XB7qXmC8JX_4mKDEA8kYBXUQKc,1961
+greenlet/tests/fail_slp_switch.py,sha256=rJBZcZfTWR3e2ERQtPAud6YKShiDsP84PmwOJbp4ey0,524
+greenlet/tests/fail_switch_three_greenlets.py,sha256=zSitV7rkNnaoHYVzAGGLnxz-yPtohXJJzaE8ehFDQ0M,956
+greenlet/tests/fail_switch_three_greenlets2.py,sha256=FPJensn2EJxoropl03JSTVP3kgP33k04h6aDWWozrOk,1285
+greenlet/tests/fail_switch_two_greenlets.py,sha256=1CaI8s3504VbbF1vj1uBYuy-zxBHVzHPIAd1LIc8ONg,817
+greenlet/tests/leakcheck.py,sha256=inbfM7_oVzd8jIKGxCgo4JqpFZaDAnWPkSULJ8vIE1s,11964
+greenlet/tests/test_contextvars.py,sha256=0n5pR_lbpAppc5wFfK0e1SwYLM-fsSFp72B5_ArLPGE,10348
+greenlet/tests/test_cpp.py,sha256=hpxhFAdKJTpAVZP8CBGs1ZcrKdscI9BaDZk4btkI5d4,2736
+greenlet/tests/test_extension_interface.py,sha256=eJ3cwLacdK2WbsrC-4DgeyHdwLRcG4zx7rrkRtqSzC4,3829
+greenlet/tests/test_gc.py,sha256=PCOaRpIyjNnNlDogGL3FZU_lrdXuM-pv1rxeE5TP5mc,2923
+greenlet/tests/test_generator.py,sha256=tONXiTf98VGm347o1b-810daPiwdla5cbpFg6QI1R1g,1240
+greenlet/tests/test_generator_nested.py,sha256=7v4HOYrf1XZP39dk5IUMubdZ8yc3ynwZcqj9GUJyMSA,3718
+greenlet/tests/test_greenlet.py,sha256=zoAy56MtEyz5P93Iknpt2pPjNO3ePYrgM7SDE8Cw_uI,45990
+greenlet/tests/test_greenlet_trash.py,sha256=n2dBlQfOoEO1ODatFi8QdhboH3fB86YtqzcYMYOXxbw,7947
+greenlet/tests/test_leaks.py,sha256=wskLqCAvqZ3qTZkam_wXzd-E5zelUjlXS5Ss8KshtZY,17465
+greenlet/tests/test_stack_saved.py,sha256=eyzqNY2VCGuGlxhT_In6TvZ6Okb0AXFZVyBEnK1jDwA,446
+greenlet/tests/test_throw.py,sha256=u2TQ_WvvCd6N6JdXWIxVEcXkKu5fepDlz9dktYdmtng,3712
+greenlet/tests/test_tracing.py,sha256=VlwzMU0C1noospZhuUMyB7MHw200emIvGCN_6G2p2ZU,8250
+greenlet/tests/test_version.py,sha256=O9DpAITsOFgiRcjd4odQ7ejmwx_N9Q1zQENVcbtFHIc,1339
+greenlet/tests/test_weakref.py,sha256=F8M23btEF87bIbpptLNBORosbQqNZGiYeKMqYjWrsak,883
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/WHEEL
new file mode 100644
index 00000000..66ca7f71
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.1.0)
+Root-Is-Purelib: false
+Tag: cp312-cp312-macosx_11_0_universal2
+
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/top_level.txt
new file mode 100644
index 00000000..46725be4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet-3.1.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+greenlet
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/CObjects.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/CObjects.cpp
new file mode 100644
index 00000000..c135995b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/CObjects.cpp
@@ -0,0 +1,157 @@
+#ifndef COBJECTS_CPP
+#define COBJECTS_CPP
+/*****************************************************************************
+ * C interface
+ *
+ * These are exported using the CObject API
+ */
+#ifdef __clang__
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wunused-function"
+#endif
+
+#include "greenlet_exceptions.hpp"
+
+#include "greenlet_internal.hpp"
+#include "greenlet_refs.hpp"
+
+
+#include "TThreadStateDestroy.cpp"
+
+#include "PyGreenlet.hpp"
+
+using greenlet::PyErrOccurred;
+using greenlet::Require;
+
+
+
+extern "C" {
+static PyGreenlet*
+PyGreenlet_GetCurrent(void)
+{
+    return GET_THREAD_STATE().state().get_current().relinquish_ownership();
+}
+
+static int
+PyGreenlet_SetParent(PyGreenlet* g, PyGreenlet* nparent)
+{
+    return green_setparent((PyGreenlet*)g, (PyObject*)nparent, NULL);
+}
+
+static PyGreenlet*
+PyGreenlet_New(PyObject* run, PyGreenlet* parent)
+{
+    using greenlet::refs::NewDictReference;
+    // In the past, we didn't use green_new and green_init, but that
+    // was a maintenance issue because we duplicated code. This way is
+    // much safer, but slightly slower. If that's a problem, we could
+    // refactor green_init to separate argument parsing from initialization.
+    OwnedGreenlet g = OwnedGreenlet::consuming(green_new(&PyGreenlet_Type, nullptr, nullptr));
+    if (!g) {
+        return NULL;
+    }
+
+    try {
+        NewDictReference kwargs;
+        if (run) {
+            kwargs.SetItem(mod_globs->str_run, run);
+        }
+        if (parent) {
+            kwargs.SetItem("parent", (PyObject*)parent);
+        }
+
+        Require(green_init(g.borrow(), mod_globs->empty_tuple, kwargs.borrow()));
+    }
+    catch (const PyErrOccurred&) {
+        return nullptr;
+    }
+
+    return g.relinquish_ownership();
+}
+
+static PyObject*
+PyGreenlet_Switch(PyGreenlet* self, PyObject* args, PyObject* kwargs)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return NULL;
+    }
+
+    if (args == NULL) {
+        args = mod_globs->empty_tuple;
+    }
+
+    if (kwargs == NULL || !PyDict_Check(kwargs)) {
+        kwargs = NULL;
+    }
+
+    return green_switch(self, args, kwargs);
+}
+
+static PyObject*
+PyGreenlet_Throw(PyGreenlet* self, PyObject* typ, PyObject* val, PyObject* tb)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return nullptr;
+    }
+    try {
+        PyErrPieces err_pieces(typ, val, tb);
+        return internal_green_throw(self, err_pieces).relinquish_ownership();
+    }
+    catch (const PyErrOccurred&) {
+        return nullptr;
+    }
+}
+
+
+
+static int
+Extern_PyGreenlet_MAIN(PyGreenlet* self)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return -1;
+    }
+    return self->pimpl->main();
+}
+
+static int
+Extern_PyGreenlet_ACTIVE(PyGreenlet* self)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return -1;
+    }
+    return self->pimpl->active();
+}
+
+static int
+Extern_PyGreenlet_STARTED(PyGreenlet* self)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return -1;
+    }
+    return self->pimpl->started();
+}
+
+static PyGreenlet*
+Extern_PyGreenlet_GET_PARENT(PyGreenlet* self)
+{
+    if (!PyGreenlet_Check(self)) {
+        PyErr_BadArgument();
+        return NULL;
+    }
+    // This can return NULL even if there is no exception
+    return self->pimpl->parent().acquire();
+}
+} // extern C.
+
+/** End C API ****************************************************************/
+#ifdef __clang__
+#    pragma clang diagnostic pop
+#endif
+
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.cpp
new file mode 100644
index 00000000..29c0bba0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.cpp
@@ -0,0 +1,738 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+#ifndef PYGREENLET_CPP
+#define PYGREENLET_CPP
+/*****************
+The Python slot functions for TGreenlet.
+ */
+
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+#include "structmember.h" // PyMemberDef
+
+#include "greenlet_internal.hpp"
+#include "TThreadStateDestroy.cpp"
+#include "TGreenlet.hpp"
+// #include "TUserGreenlet.cpp"
+// #include "TMainGreenlet.cpp"
+// #include "TBrokenGreenlet.cpp"
+
+
+#include "greenlet_refs.hpp"
+#include "greenlet_slp_switch.hpp"
+
+#include "greenlet_thread_support.hpp"
+#include "TGreenlet.hpp"
+
+#include "TGreenletGlobals.cpp"
+#include "TThreadStateDestroy.cpp"
+#include "PyGreenlet.hpp"
+// #include "TGreenlet.cpp"
+
+// #include "TExceptionState.cpp"
+// #include "TPythonState.cpp"
+// #include "TStackState.cpp"
+
+using greenlet::LockGuard;
+using greenlet::LockInitError;
+using greenlet::PyErrOccurred;
+using greenlet::Require;
+
+using greenlet::g_handle_exit;
+using greenlet::single_result;
+
+using greenlet::Greenlet;
+using greenlet::UserGreenlet;
+using greenlet::MainGreenlet;
+using greenlet::BrokenGreenlet;
+using greenlet::ThreadState;
+using greenlet::PythonState;
+
+
+
+static PyGreenlet*
+green_new(PyTypeObject* type, PyObject* UNUSED(args), PyObject* UNUSED(kwds))
+{
+    PyGreenlet* o =
+        (PyGreenlet*)PyBaseObject_Type.tp_new(type, mod_globs->empty_tuple, mod_globs->empty_dict);
+    if (o) {
+        new UserGreenlet(o, GET_THREAD_STATE().state().borrow_current());
+        assert(Py_REFCNT(o) == 1);
+    }
+    return o;
+}
+
+
+// green_init is used in the tp_init slot. So it's important that
+// it can be called directly from CPython. Thus, we don't use
+// BorrowedGreenlet and BorrowedObject --- although in theory
+// these should be binary layout compatible, that may not be
+// guaranteed to be the case (32-bit linux ppc possibly).
+static int
+green_init(PyGreenlet* self, PyObject* args, PyObject* kwargs)
+{
+    PyArgParseParam run;
+    PyArgParseParam nparent;
+    static const char* kwlist[] = {
+        "run",
+        "parent",
+        NULL
+    };
+
+    // recall: The O specifier does NOT increase the reference count.
+    if (!PyArg_ParseTupleAndKeywords(
+             args, kwargs, "|OO:green", (char**)kwlist, &run, &nparent)) {
+        return -1;
+    }
+
+    if (run) {
+        if (green_setrun(self, run, NULL)) {
+            return -1;
+        }
+    }
+    if (nparent && !nparent.is_None()) {
+        return green_setparent(self, nparent, NULL);
+    }
+    return 0;
+}
+
+
+
+static int
+green_traverse(PyGreenlet* self, visitproc visit, void* arg)
+{
+    // We must only visit referenced objects, i.e. only objects
+    // Py_INCREF'ed by this greenlet (directly or indirectly):
+    //
+    // - stack_prev is not visited: holds previous stack pointer, but it's not
+    //    referenced
+    // - frames are not visited as we don't strongly reference them;
+    //    alive greenlets are not garbage collected
+    //    anyway. This can be a problem, however, if this greenlet is
+    //    never allowed to finish, and is referenced from the frame: we
+    //    have an uncollectible cycle in that case. Note that the
+    //    frame object itself is also frequently not even tracked by the GC
+    //    starting with Python 3.7 (frames are allocated by the
+    //    interpreter untracked, and only become tracked when their
+    //    evaluation is finished if they have a refcount > 1). All of
+    //    this is to say that we should probably strongly reference
+    //    the frame object. Doing so, while always allowing GC on a
+    //    greenlet, solves several leaks for us.
+
+    Py_VISIT(self->dict);
+    if (!self->pimpl) {
+        // Hmm. I have seen this at interpreter shutdown time,
+        // I think. That's very odd because this doesn't go away until
+        // we're ``green_dealloc()``, at which point we shouldn't be
+        // traversed anymore.
+        return 0;
+    }
+
+    return self->pimpl->tp_traverse(visit, arg);
+}
+
+static int
+green_is_gc(PyObject* _self)
+{
+    BorrowedGreenlet self(_self);
+    int result = 0;
+    /* Main greenlet can be garbage collected since it can only
+       become unreachable if the underlying thread exited.
+       Active greenlets --- including those that are suspended ---
+       cannot be garbage collected, however.
+    */
+    if (self->main() || !self->active()) {
+        result = 1;
+    }
+    // The main greenlet pointer will eventually go away after the thread dies.
+    if (self->was_running_in_dead_thread()) {
+        // Our thread is dead! We can never run again. Might as well
+        // GC us. Note that if a tuple containing only us and other
+        // immutable objects had been scanned before this, when we
+        // would have returned 0, the tuple will take itself out of GC
+        // tracking and never be investigated again. So that could
+        // result in both us and the tuple leaking due to an
+        // unreachable/uncollectible reference. The same goes for
+        // dictionaries.
+        //
+        // It's not a great idea to be changing our GC state on the
+        // fly.
+        result = 1;
+    }
+    return result;
+}
+
+
+static int
+green_clear(PyGreenlet* self)
+{
+    /* Greenlet is only cleared if it is about to be collected.
+       Since active greenlets are not garbage collectable, we can
+       be sure that, even if they are deallocated during clear,
+       nothing they reference is in unreachable or finalizers,
+       so even if it switches we are relatively safe. */
+    // XXX: Are we responsible for clearing weakrefs here?
+    Py_CLEAR(self->dict);
+    return self->pimpl->tp_clear();
+}
+
+/**
+ * Returns 0 on failure (the object was resurrected) or 1 on success.
+ **/
+static int
+_green_dealloc_kill_started_non_main_greenlet(BorrowedGreenlet self)
+{
+    /* Hacks hacks hacks copied from instance_dealloc() */
+    /* Temporarily resurrect the greenlet. */
+    assert(self.REFCNT() == 0);
+    Py_SET_REFCNT(self.borrow(), 1);
+    /* Save the current exception, if any. */
+    PyErrPieces saved_err;
+    try {
+        // BY THE TIME WE GET HERE, the state may actually be going
+        // away
+        // if we're shutting down the interpreter and freeing thread
+        // entries,
+        // this could result in freeing greenlets that were leaked. So
+        // we can't try to read the state.
+        self->deallocing_greenlet_in_thread(
+              self->thread_state()
+              ? static_cast<ThreadState*>(GET_THREAD_STATE())
+              : nullptr);
+    }
+    catch (const PyErrOccurred&) {
+        PyErr_WriteUnraisable(self.borrow_o());
+        /* XXX what else should we do? */
+    }
+    /* Check for no resurrection must be done while we keep
+     * our internal reference, otherwise PyFile_WriteObject
+     * causes recursion if using Py_INCREF/Py_DECREF
+     */
+    if (self.REFCNT() == 1 && self->active()) {
+        /* Not resurrected, but still not dead!
+           XXX what else should we do? we complain. */
+        PyObject* f = PySys_GetObject("stderr");
+        Py_INCREF(self.borrow_o()); /* leak! */
+        if (f != NULL) {
+            PyFile_WriteString("GreenletExit did not kill ", f);
+            PyFile_WriteObject(self.borrow_o(), f, 0);
+            PyFile_WriteString("\n", f);
+        }
+    }
+    /* Restore the saved exception. */
+    saved_err.PyErrRestore();
+    /* Undo the temporary resurrection; can't use DECREF here,
+     * it would cause a recursive call.
+     */
+    assert(self.REFCNT() > 0);
+
+    Py_ssize_t refcnt = self.REFCNT() - 1;
+    Py_SET_REFCNT(self.borrow_o(), refcnt);
+    if (refcnt != 0) {
+        /* Resurrected! */
+        _Py_NewReference(self.borrow_o());
+        Py_SET_REFCNT(self.borrow_o(), refcnt);
+        /* Better to use tp_finalizer slot (PEP 442)
+         * and call ``PyObject_CallFinalizerFromDealloc``,
+         * but that's only supported in Python 3.4+; see
+         * Modules/_io/iobase.c for an example.
+         *
+         * The following approach is copied from iobase.c in CPython 2.7.
+         * (along with much of this function in general). Here's their
+         * comment:
+         *
+         * When called from a heap type's dealloc, the type will be
+         * decref'ed on return (see e.g. subtype_dealloc in typeobject.c). */
+        if (PyType_HasFeature(self.TYPE(), Py_TPFLAGS_HEAPTYPE)) {
+            Py_INCREF(self.TYPE());
+        }
+
+        PyObject_GC_Track((PyObject*)self);
+
+        _Py_DEC_REFTOTAL;
+#ifdef COUNT_ALLOCS
+        --Py_TYPE(self)->tp_frees;
+        --Py_TYPE(self)->tp_allocs;
+#endif /* COUNT_ALLOCS */
+        return 0;
+    }
+    return 1;
+}
+
+
+static void
+green_dealloc(PyGreenlet* self)
+{
+    PyObject_GC_UnTrack(self);
+    BorrowedGreenlet me(self);
+    if (me->active()
+        && me->started()
+        && !me->main()) {
+        if (!_green_dealloc_kill_started_non_main_greenlet(me)) {
+            return;
+        }
+    }
+
+    if (self->weakreflist != NULL) {
+        PyObject_ClearWeakRefs((PyObject*)self);
+    }
+    Py_CLEAR(self->dict);
+
+    if (self->pimpl) {
+        // In case deleting this, which frees some memory,
+        // somehow winds up calling back into us. That's usually a
+        //bug in our code.
+        Greenlet* p = self->pimpl;
+        self->pimpl = nullptr;
+        delete p;
+    }
+    // and finally we're done. self is now invalid.
+    Py_TYPE(self)->tp_free((PyObject*)self);
+}
+
+
+
+static OwnedObject
+internal_green_throw(BorrowedGreenlet self, PyErrPieces& err_pieces)
+{
+    PyObject* result = nullptr;
+    err_pieces.PyErrRestore();
+    assert(PyErr_Occurred());
+    if (self->started() && !self->active()) {
+        /* dead greenlet: turn GreenletExit into a regular return */
+        result = g_handle_exit(OwnedObject()).relinquish_ownership();
+    }
+    self->args() <<= result;
+
+    return single_result(self->g_switch());
+}
+
+
+
+PyDoc_STRVAR(
+    green_switch_doc,
+    "switch(*args, **kwargs)\n"
+    "\n"
+    "Switch execution to this greenlet.\n"
+    "\n"
+    "If this greenlet has never been run, then this greenlet\n"
+    "will be switched to using the body of ``self.run(*args, **kwargs)``.\n"
+    "\n"
+    "If the greenlet is active (has been run, but was switch()'ed\n"
+    "out before leaving its run function), then this greenlet will\n"
+    "be resumed and the return value to its switch call will be\n"
+    "None if no arguments are given, the given argument if one\n"
+    "argument is given, or the args tuple and keyword args dict if\n"
+    "multiple arguments are given.\n"
+    "\n"
+    "If the greenlet is dead, or is the current greenlet then this\n"
+    "function will simply return the arguments using the same rules as\n"
+    "above.\n");
+
+static PyObject*
+green_switch(PyGreenlet* self, PyObject* args, PyObject* kwargs)
+{
+    using greenlet::SwitchingArgs;
+    SwitchingArgs switch_args(OwnedObject::owning(args), OwnedObject::owning(kwargs));
+    self->pimpl->may_switch_away();
+    self->pimpl->args() <<= switch_args;
+
+    // If we're switching out of a greenlet, and that switch is the
+    // last thing the greenlet does, the greenlet ought to be able to
+    // go ahead and die at that point. Currently, someone else must
+    // manually switch back to the greenlet so that we "fall off the
+    // end" and can perform cleanup. You'd think we'd be able to
+    // figure out that this is happening using the frame's ``f_lasti``
+    // member, which is supposed to be an index into
+    // ``frame->f_code->co_code``, the bytecode string. However, in
+    // recent interpreters, ``f_lasti`` tends not to be updated thanks
+    // to things like the PREDICT() macros in ceval.c. So it doesn't
+    // really work to do that in many cases. For example, the Python
+    // code:
+    //     def run():
+    //         greenlet.getcurrent().parent.switch()
+    // produces bytecode of len 16, with the actual call to switch()
+    // being at index 10 (in Python 3.10). However, the reported
+    // ``f_lasti`` we actually see is...5! (Which happens to be the
+    // second byte of the CALL_METHOD op for ``getcurrent()``).
+
+    try {
+        //OwnedObject result = single_result(self->pimpl->g_switch());
+        OwnedObject result(single_result(self->pimpl->g_switch()));
+#ifndef NDEBUG
+        // Note that the current greenlet isn't necessarily self. If self
+        // finished, we went to one of its parents.
+        assert(!self->pimpl->args());
+
+        const BorrowedGreenlet& current = GET_THREAD_STATE().state().borrow_current();
+        // It's possible it's never been switched to.
+        assert(!current->args());
+#endif
+        PyObject* p = result.relinquish_ownership();
+
+        if (!p && !PyErr_Occurred()) {
+            // This shouldn't be happening anymore, so the asserts
+            // are there for debug builds. Non-debug builds
+            // crash "gracefully" in this case, although there is an
+            // argument to be made for killing the process in all
+            // cases --- for this to be the case, our switches
+            // probably nested in an incorrect way, so the state is
+            // suspicious. Nothing should be corrupt though, just
+            // confused at the Python level. Letting this propagate is
+            // probably good enough.
+            assert(p || PyErr_Occurred());
+            throw PyErrOccurred(
+                mod_globs->PyExc_GreenletError,
+                "Greenlet.switch() returned NULL without an exception set."
+            );
+        }
+        return p;
+    }
+    catch(const PyErrOccurred&) {
+        return nullptr;
+    }
+}
+
+PyDoc_STRVAR(
+    green_throw_doc,
+    "Switches execution to this greenlet, but immediately raises the\n"
+    "given exception in this greenlet.  If no argument is provided, the "
+    "exception\n"
+    "defaults to `greenlet.GreenletExit`.  The normal exception\n"
+    "propagation rules apply, as described for `switch`.  Note that calling "
+    "this\n"
+    "method is almost equivalent to the following::\n"
+    "\n"
+    "    def raiser():\n"
+    "        raise typ, val, tb\n"
+    "    g_raiser = greenlet(raiser, parent=g)\n"
+    "    g_raiser.switch()\n"
+    "\n"
+    "except that this trick does not work for the\n"
+    "`greenlet.GreenletExit` exception, which would not propagate\n"
+    "from ``g_raiser`` to ``g``.\n");
+
+static PyObject*
+green_throw(PyGreenlet* self, PyObject* args)
+{
+    PyArgParseParam typ(mod_globs->PyExc_GreenletExit);
+    PyArgParseParam val;
+    PyArgParseParam tb;
+
+    if (!PyArg_ParseTuple(args, "|OOO:throw", &typ, &val, &tb)) {
+        return nullptr;
+    }
+
+    assert(typ.borrow() || val.borrow());
+
+    self->pimpl->may_switch_away();
+    try {
+        // Both normalizing the error and the actual throw_greenlet
+        // could throw PyErrOccurred.
+        PyErrPieces err_pieces(typ.borrow(), val.borrow(), tb.borrow());
+
+        return internal_green_throw(self, err_pieces).relinquish_ownership();
+    }
+    catch (const PyErrOccurred&) {
+        return nullptr;
+    }
+}
+
+static int
+green_bool(PyGreenlet* self)
+{
+    return self->pimpl->active();
+}
+
+/**
+ * CAUTION: Allocates memory, may run GC and arbitrary Python code.
+ */
+static PyObject*
+green_getdict(PyGreenlet* self, void* UNUSED(context))
+{
+    if (self->dict == NULL) {
+        self->dict = PyDict_New();
+        if (self->dict == NULL) {
+            return NULL;
+        }
+    }
+    Py_INCREF(self->dict);
+    return self->dict;
+}
+
+static int
+green_setdict(PyGreenlet* self, PyObject* val, void* UNUSED(context))
+{
+    PyObject* tmp;
+
+    if (val == NULL) {
+        PyErr_SetString(PyExc_TypeError, "__dict__ may not be deleted");
+        return -1;
+    }
+    if (!PyDict_Check(val)) {
+        PyErr_SetString(PyExc_TypeError, "__dict__ must be a dictionary");
+        return -1;
+    }
+    tmp = self->dict;
+    Py_INCREF(val);
+    self->dict = val;
+    Py_XDECREF(tmp);
+    return 0;
+}
+
+static bool
+_green_not_dead(BorrowedGreenlet self)
+{
+    // XXX: Where else should we do this?
+    // Probably on entry to most Python-facing functions?
+    if (self->was_running_in_dead_thread()) {
+        self->deactivate_and_free();
+        return false;
+    }
+    return self->active() || !self->started();
+}
+
+
+static PyObject*
+green_getdead(PyGreenlet* self, void* UNUSED(context))
+{
+    if (_green_not_dead(self)) {
+        Py_RETURN_FALSE;
+    }
+    else {
+        Py_RETURN_TRUE;
+    }
+}
+
+static PyObject*
+green_get_stack_saved(PyGreenlet* self, void* UNUSED(context))
+{
+    return PyLong_FromSsize_t(self->pimpl->stack_saved());
+}
+
+
+static PyObject*
+green_getrun(PyGreenlet* self, void* UNUSED(context))
+{
+    try {
+        OwnedObject result(BorrowedGreenlet(self)->run());
+        return result.relinquish_ownership();
+    }
+    catch(const PyErrOccurred&) {
+        return nullptr;
+    }
+}
+
+
+static int
+green_setrun(PyGreenlet* self, PyObject* nrun, void* UNUSED(context))
+{
+    try {
+        BorrowedGreenlet(self)->run(nrun);
+        return 0;
+    }
+    catch(const PyErrOccurred&) {
+        return -1;
+    }
+}
+
+static PyObject*
+green_getparent(PyGreenlet* self, void* UNUSED(context))
+{
+    return BorrowedGreenlet(self)->parent().acquire_or_None();
+}
+
+
+static int
+green_setparent(PyGreenlet* self, PyObject* nparent, void* UNUSED(context))
+{
+    try {
+        BorrowedGreenlet(self)->parent(nparent);
+    }
+    catch(const PyErrOccurred&) {
+        return -1;
+    }
+    return 0;
+}
+
+
+static PyObject*
+green_getcontext(const PyGreenlet* self, void* UNUSED(context))
+{
+    const Greenlet *const g = self->pimpl;
+    try {
+        OwnedObject result(g->context());
+        return result.relinquish_ownership();
+    }
+    catch(const PyErrOccurred&) {
+        return nullptr;
+    }
+}
+
+static int
+green_setcontext(PyGreenlet* self, PyObject* nctx, void* UNUSED(context))
+{
+    try {
+        BorrowedGreenlet(self)->context(nctx);
+        return 0;
+    }
+    catch(const PyErrOccurred&) {
+        return -1;
+    }
+}
+
+
+static PyObject*
+green_getframe(PyGreenlet* self, void* UNUSED(context))
+{
+    const PythonState::OwnedFrame& top_frame = BorrowedGreenlet(self)->top_frame();
+    return top_frame.acquire_or_None();
+}
+
+
+static PyObject*
+green_getstate(PyGreenlet* self)
+{
+    PyErr_Format(PyExc_TypeError,
+                 "cannot serialize '%s' object",
+                 Py_TYPE(self)->tp_name);
+    return nullptr;
+}
+
+static PyObject*
+green_repr(PyGreenlet* _self)
+{
+    BorrowedGreenlet self(_self);
+    /*
+      Return a string like
+      <greenlet.greenlet at 0xdeadbeef [current][active started]|dead main>
+
+      The handling of greenlets across threads is not super good.
+      We mostly use the internal definitions of these terms, but they
+      generally should make sense to users as well.
+     */
+    PyObject* result;
+    int never_started = !self->started() && !self->active();
+
+    const char* const tp_name = Py_TYPE(self)->tp_name;
+
+    if (_green_not_dead(self)) {
+        /* XXX: The otid= is almost useless because you can't correlate it to
+         any thread identifier exposed to Python. We could use
+         PyThreadState_GET()->thread_id, but we'd need to save that in the
+         greenlet, or save the whole PyThreadState object itself.
+
+         As it stands, its only useful for identifying greenlets from the same thread.
+        */
+        const char* state_in_thread;
+        if (self->was_running_in_dead_thread()) {
+            // The thread it was running in is dead!
+            // This can happen, especially at interpreter shut down.
+            // It complicates debugging output because it may be
+            // impossible to access the current thread state at that
+            // time. Thus, don't access the current thread state.
+            state_in_thread = " (thread exited)";
+        }
+        else {
+            state_in_thread = GET_THREAD_STATE().state().is_current(self)
+                ? " current"
+                : (self->started() ? " suspended" : "");
+        }
+        result = PyUnicode_FromFormat(
+            "<%s object at %p (otid=%p)%s%s%s%s>",
+            tp_name,
+            self.borrow_o(),
+            self->thread_state(),
+            state_in_thread,
+            self->active() ? " active" : "",
+            never_started ? " pending" : " started",
+            self->main() ? " main" : ""
+        );
+    }
+    else {
+        result = PyUnicode_FromFormat(
+            "<%s object at %p (otid=%p) %sdead>",
+            tp_name,
+            self.borrow_o(),
+            self->thread_state(),
+            self->was_running_in_dead_thread()
+            ? "(thread exited) "
+            : ""
+            );
+    }
+
+    return result;
+}
+
+
+static PyMethodDef green_methods[] = {
+    {
+      .ml_name="switch",
+      .ml_meth=reinterpret_cast<PyCFunction>(green_switch),
+      .ml_flags=METH_VARARGS | METH_KEYWORDS,
+      .ml_doc=green_switch_doc
+    },
+    {.ml_name="throw", .ml_meth=(PyCFunction)green_throw, .ml_flags=METH_VARARGS, .ml_doc=green_throw_doc},
+    {.ml_name="__getstate__", .ml_meth=(PyCFunction)green_getstate, .ml_flags=METH_NOARGS, .ml_doc=NULL},
+    {.ml_name=NULL, .ml_meth=NULL} /* sentinel */
+};
+
+static PyGetSetDef green_getsets[] = {
+    /* name, getter, setter, doc, context pointer */
+    {.name="__dict__", .get=(getter)green_getdict, .set=(setter)green_setdict},
+    {.name="run", .get=(getter)green_getrun, .set=(setter)green_setrun},
+    {.name="parent", .get=(getter)green_getparent, .set=(setter)green_setparent},
+    {.name="gr_frame", .get=(getter)green_getframe },
+    {
+      .name="gr_context",
+      .get=(getter)green_getcontext,
+      .set=(setter)green_setcontext
+    },
+    {.name="dead", .get=(getter)green_getdead},
+    {.name="_stack_saved", .get=(getter)green_get_stack_saved},
+    {.name=NULL}
+};
+
+static PyMemberDef green_members[] = {
+    {.name=NULL}
+};
+
+static PyNumberMethods green_as_number = {
+  .nb_bool=(inquiry)green_bool,
+};
+
+
+PyTypeObject PyGreenlet_Type = {
+    .ob_base=PyVarObject_HEAD_INIT(NULL, 0)
+    .tp_name="greenlet.greenlet", /* tp_name */
+    .tp_basicsize=sizeof(PyGreenlet),  /* tp_basicsize */
+    /* methods */
+    .tp_dealloc=(destructor)green_dealloc, /* tp_dealloc */
+    .tp_repr=(reprfunc)green_repr,      /* tp_repr */
+    .tp_as_number=&green_as_number,          /* tp_as _number*/
+    .tp_flags=G_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /* tp_flags */
+    .tp_doc="greenlet(run=None, parent=None) -> greenlet\n\n"
+    "Creates a new greenlet object (without running it).\n\n"
+    " - *run* -- The callable to invoke.\n"
+    " - *parent* -- The parent greenlet. The default is the current "
+    "greenlet.",                        /* tp_doc */
+    .tp_traverse=(traverseproc)green_traverse, /* tp_traverse */
+    .tp_clear=(inquiry)green_clear,         /* tp_clear */
+    .tp_weaklistoffset=offsetof(PyGreenlet, weakreflist),  /* tp_weaklistoffset */
+
+    .tp_methods=green_methods,                      /* tp_methods */
+    .tp_members=green_members,                      /* tp_members */
+    .tp_getset=green_getsets,                      /* tp_getset */
+    .tp_dictoffset=offsetof(PyGreenlet, dict),         /* tp_dictoffset */
+    .tp_init=(initproc)green_init,               /* tp_init */
+    .tp_alloc=PyType_GenericAlloc,                  /* tp_alloc */
+    .tp_new=(newfunc)green_new,                          /* tp_new */
+    .tp_free=PyObject_GC_Del,                   /* tp_free */
+    .tp_is_gc=(inquiry)green_is_gc,         /* tp_is_gc */
+};
+
+#endif
+
+// Local Variables:
+// flycheck-clang-include-path: ("/opt/local/Library/Frameworks/Python.framework/Versions/3.8/include/python3.8")
+// End:
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.hpp
new file mode 100644
index 00000000..df6cd805
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenlet.hpp
@@ -0,0 +1,35 @@
+#ifndef PYGREENLET_HPP
+#define PYGREENLET_HPP
+
+
+#include "greenlet.h"
+#include "greenlet_compiler_compat.hpp"
+#include "greenlet_refs.hpp"
+
+
+using greenlet::refs::OwnedGreenlet;
+using greenlet::refs::BorrowedGreenlet;
+using greenlet::refs::BorrowedObject;;
+using greenlet::refs::OwnedObject;
+using greenlet::refs::PyErrPieces;
+
+
+// XXX: These doesn't really belong here, it's not a Python slot.
+static OwnedObject internal_green_throw(BorrowedGreenlet self, PyErrPieces& err_pieces);
+
+static PyGreenlet* green_new(PyTypeObject* type, PyObject* UNUSED(args), PyObject* UNUSED(kwds));
+static int green_clear(PyGreenlet* self);
+static int green_init(PyGreenlet* self, PyObject* args, PyObject* kwargs);
+static int green_setparent(PyGreenlet* self, PyObject* nparent, void* UNUSED(context));
+static int green_setrun(PyGreenlet* self, PyObject* nrun, void* UNUSED(context));
+static int green_traverse(PyGreenlet* self, visitproc visit, void* arg);
+static void green_dealloc(PyGreenlet* self);
+static PyObject* green_getparent(PyGreenlet* self, void* UNUSED(context));
+
+static int green_is_gc(PyObject* self);
+static PyObject* green_getdead(PyGreenlet* self, void* UNUSED(context));
+static PyObject* green_getrun(PyGreenlet* self, void* UNUSED(context));
+static int green_setcontext(PyGreenlet* self, PyObject* nctx, void* UNUSED(context));
+static PyObject* green_getframe(PyGreenlet* self, void* UNUSED(context));
+static PyObject* green_repr(PyGreenlet* self);
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenletUnswitchable.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenletUnswitchable.cpp
new file mode 100644
index 00000000..1b768ee3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/PyGreenletUnswitchable.cpp
@@ -0,0 +1,147 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+   Implementation of the Python slots for PyGreenletUnswitchable_Type
+*/
+#ifndef PY_GREENLET_UNSWITCHABLE_CPP
+#define PY_GREENLET_UNSWITCHABLE_CPP
+
+
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+#include "structmember.h" // PyMemberDef
+
+#include "greenlet_internal.hpp"
+// Code after this point can assume access to things declared in stdint.h,
+// including the fixed-width types. This goes for the platform-specific switch functions
+// as well.
+#include "greenlet_refs.hpp"
+#include "greenlet_slp_switch.hpp"
+
+#include "greenlet_thread_support.hpp"
+#include "TGreenlet.hpp"
+
+#include "TGreenlet.cpp"
+#include "TGreenletGlobals.cpp"
+#include "TThreadStateDestroy.cpp"
+
+
+using greenlet::LockGuard;
+using greenlet::LockInitError;
+using greenlet::PyErrOccurred;
+using greenlet::Require;
+
+using greenlet::g_handle_exit;
+using greenlet::single_result;
+
+using greenlet::Greenlet;
+using greenlet::UserGreenlet;
+using greenlet::MainGreenlet;
+using greenlet::BrokenGreenlet;
+using greenlet::ThreadState;
+using greenlet::PythonState;
+
+
+#include "PyGreenlet.hpp"
+
+static PyGreenlet*
+green_unswitchable_new(PyTypeObject* type, PyObject* UNUSED(args), PyObject* UNUSED(kwds))
+{
+    PyGreenlet* o =
+        (PyGreenlet*)PyBaseObject_Type.tp_new(type, mod_globs->empty_tuple, mod_globs->empty_dict);
+    if (o) {
+        new BrokenGreenlet(o, GET_THREAD_STATE().state().borrow_current());
+        assert(Py_REFCNT(o) == 1);
+    }
+    return o;
+}
+
+static PyObject*
+green_unswitchable_getforce(PyGreenlet* self, void* UNUSED(context))
+{
+    BrokenGreenlet* broken = dynamic_cast<BrokenGreenlet*>(self->pimpl);
+    return PyBool_FromLong(broken->_force_switch_error);
+}
+
+static int
+green_unswitchable_setforce(PyGreenlet* self, PyObject* nforce, void* UNUSED(context))
+{
+    if (!nforce) {
+        PyErr_SetString(
+            PyExc_AttributeError,
+            "Cannot delete force_switch_error"
+        );
+        return -1;
+    }
+    BrokenGreenlet* broken = dynamic_cast<BrokenGreenlet*>(self->pimpl);
+    int is_true = PyObject_IsTrue(nforce);
+    if (is_true == -1) {
+        return -1;
+    }
+    broken->_force_switch_error = is_true;
+    return 0;
+}
+
+static PyObject*
+green_unswitchable_getforceslp(PyGreenlet* self, void* UNUSED(context))
+{
+    BrokenGreenlet* broken = dynamic_cast<BrokenGreenlet*>(self->pimpl);
+    return PyBool_FromLong(broken->_force_slp_switch_error);
+}
+
+static int
+green_unswitchable_setforceslp(PyGreenlet* self, PyObject* nforce, void* UNUSED(context))
+{
+    if (!nforce) {
+        PyErr_SetString(
+            PyExc_AttributeError,
+            "Cannot delete force_slp_switch_error"
+        );
+        return -1;
+    }
+    BrokenGreenlet* broken = dynamic_cast<BrokenGreenlet*>(self->pimpl);
+    int is_true = PyObject_IsTrue(nforce);
+    if (is_true == -1) {
+        return -1;
+    }
+    broken->_force_slp_switch_error = is_true;
+    return 0;
+}
+
+static PyGetSetDef green_unswitchable_getsets[] = {
+    /* name, getter, setter, doc, closure (context pointer) */
+    {
+      .name="force_switch_error",
+      .get=(getter)green_unswitchable_getforce,
+      .set=(setter)green_unswitchable_setforce,
+      .doc=NULL
+    },
+    {
+      .name="force_slp_switch_error",
+      .get=(getter)green_unswitchable_getforceslp,
+      .set=(setter)green_unswitchable_setforceslp,
+      .doc=nullptr
+    },
+    {.name=nullptr}
+};
+
+PyTypeObject PyGreenletUnswitchable_Type = {
+    .ob_base=PyVarObject_HEAD_INIT(NULL, 0)
+    .tp_name="greenlet._greenlet.UnswitchableGreenlet",
+    .tp_dealloc= (destructor)green_dealloc, /* tp_dealloc */
+    .tp_flags=G_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /* tp_flags */
+    .tp_doc="Undocumented internal class",                        /* tp_doc */
+    .tp_traverse=(traverseproc)green_traverse, /* tp_traverse */
+    .tp_clear=(inquiry)green_clear,         /* tp_clear */
+
+    .tp_getset=green_unswitchable_getsets,                      /* tp_getset */
+    .tp_base=&PyGreenlet_Type,                                  /* tp_base */
+    .tp_init=(initproc)green_init,               /* tp_init */
+    .tp_alloc=PyType_GenericAlloc,                  /* tp_alloc */
+    .tp_new=(newfunc)green_unswitchable_new,                          /* tp_new */
+    .tp_free=PyObject_GC_Del,                   /* tp_free */
+    .tp_is_gc=(inquiry)green_is_gc,         /* tp_is_gc */
+};
+
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/PyModule.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/PyModule.cpp
new file mode 100644
index 00000000..6adcb5c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/PyModule.cpp
@@ -0,0 +1,292 @@
+#ifndef PY_MODULE_CPP
+#define PY_MODULE_CPP
+
+#include "greenlet_internal.hpp"
+
+
+#include "TGreenletGlobals.cpp"
+#include "TMainGreenlet.cpp"
+#include "TThreadStateDestroy.cpp"
+
+using greenlet::LockGuard;
+using greenlet::ThreadState;
+
+#ifdef __clang__
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wunused-function"
+#    pragma clang diagnostic ignored "-Wunused-variable"
+#endif
+
+PyDoc_STRVAR(mod_getcurrent_doc,
+             "getcurrent() -> greenlet\n"
+             "\n"
+             "Returns the current greenlet (i.e. the one which called this "
+             "function).\n");
+
+static PyObject*
+mod_getcurrent(PyObject* UNUSED(module))
+{
+    return GET_THREAD_STATE().state().get_current().relinquish_ownership_o();
+}
+
+PyDoc_STRVAR(mod_settrace_doc,
+             "settrace(callback) -> object\n"
+             "\n"
+             "Sets a new tracing function and returns the previous one.\n");
+static PyObject*
+mod_settrace(PyObject* UNUSED(module), PyObject* args)
+{
+    PyArgParseParam tracefunc;
+    if (!PyArg_ParseTuple(args, "O", &tracefunc)) {
+        return NULL;
+    }
+    ThreadState& state = GET_THREAD_STATE();
+    OwnedObject previous = state.get_tracefunc();
+    if (!previous) {
+        previous = Py_None;
+    }
+
+    state.set_tracefunc(tracefunc);
+
+    return previous.relinquish_ownership();
+}
+
+PyDoc_STRVAR(mod_gettrace_doc,
+             "gettrace() -> object\n"
+             "\n"
+             "Returns the currently set tracing function, or None.\n");
+
+static PyObject*
+mod_gettrace(PyObject* UNUSED(module))
+{
+    OwnedObject tracefunc = GET_THREAD_STATE().state().get_tracefunc();
+    if (!tracefunc) {
+        tracefunc = Py_None;
+    }
+    return tracefunc.relinquish_ownership();
+}
+
+
+
+PyDoc_STRVAR(mod_set_thread_local_doc,
+             "set_thread_local(key, value) -> None\n"
+             "\n"
+             "Set a value in the current thread-local dictionary. Debugging only.\n");
+
+static PyObject*
+mod_set_thread_local(PyObject* UNUSED(module), PyObject* args)
+{
+    PyArgParseParam key;
+    PyArgParseParam value;
+    PyObject* result = NULL;
+
+    if (PyArg_UnpackTuple(args, "set_thread_local", 2, 2, &key, &value)) {
+        if(PyDict_SetItem(
+                          PyThreadState_GetDict(), // borrow
+                          key,
+                          value) == 0 ) {
+            // success
+            Py_INCREF(Py_None);
+            result = Py_None;
+        }
+    }
+    return result;
+}
+
+PyDoc_STRVAR(mod_get_pending_cleanup_count_doc,
+             "get_pending_cleanup_count() -> Integer\n"
+             "\n"
+             "Get the number of greenlet cleanup operations pending. Testing only.\n");
+
+
+static PyObject*
+mod_get_pending_cleanup_count(PyObject* UNUSED(module))
+{
+    LockGuard cleanup_lock(*mod_globs->thread_states_to_destroy_lock);
+    return PyLong_FromSize_t(mod_globs->thread_states_to_destroy.size());
+}
+
+PyDoc_STRVAR(mod_get_total_main_greenlets_doc,
+             "get_total_main_greenlets() -> Integer\n"
+             "\n"
+             "Quickly return the number of main greenlets that exist. Testing only.\n");
+
+static PyObject*
+mod_get_total_main_greenlets(PyObject* UNUSED(module))
+{
+    return PyLong_FromSize_t(G_TOTAL_MAIN_GREENLETS);
+}
+
+
+
+PyDoc_STRVAR(mod_get_clocks_used_doing_optional_cleanup_doc,
+             "get_clocks_used_doing_optional_cleanup() -> Integer\n"
+             "\n"
+             "Get the number of clock ticks the program has used doing optional "
+             "greenlet cleanup.\n"
+             "Beginning in greenlet 2.0, greenlet tries to find and dispose of greenlets\n"
+             "that leaked after a thread exited. This requires invoking Python's garbage collector,\n"
+             "which may have a performance cost proportional to the number of live objects.\n"
+             "This function returns the amount of processor time\n"
+             "greenlet has used to do this. In programs that run with very large amounts of live\n"
+             "objects, this metric can be used to decide whether the cost of doing this cleanup\n"
+             "is worth the memory leak being corrected. If not, you can disable the cleanup\n"
+             "using ``enable_optional_cleanup(False)``.\n"
+             "The units are arbitrary and can only be compared to themselves (similarly to ``time.clock()``);\n"
+             "for example, to see how it scales with your heap. You can attempt to convert them into seconds\n"
+             "by dividing by the value of CLOCKS_PER_SEC."
+             "If cleanup has been disabled, returns None."
+             "\n"
+             "This is an implementation specific, provisional API. It may be changed or removed\n"
+             "in the future.\n"
+             ".. versionadded:: 2.0"
+             );
+static PyObject*
+mod_get_clocks_used_doing_optional_cleanup(PyObject* UNUSED(module))
+{
+    std::clock_t& clocks = ThreadState::clocks_used_doing_gc();
+
+    if (clocks == std::clock_t(-1)) {
+        Py_RETURN_NONE;
+    }
+    // This might not actually work on some implementations; clock_t
+    // is an opaque type.
+    return PyLong_FromSsize_t(clocks);
+}
+
+PyDoc_STRVAR(mod_enable_optional_cleanup_doc,
+             "mod_enable_optional_cleanup(bool) -> None\n"
+             "\n"
+             "Enable or disable optional cleanup operations.\n"
+             "See ``get_clocks_used_doing_optional_cleanup()`` for details.\n"
+             );
+static PyObject*
+mod_enable_optional_cleanup(PyObject* UNUSED(module), PyObject* flag)
+{
+    int is_true = PyObject_IsTrue(flag);
+    if (is_true == -1) {
+        return nullptr;
+    }
+
+    std::clock_t& clocks = ThreadState::clocks_used_doing_gc();
+    if (is_true) {
+        // If we already have a value, we don't want to lose it.
+        if (clocks == std::clock_t(-1)) {
+            clocks = 0;
+        }
+    }
+    else {
+        clocks = std::clock_t(-1);
+    }
+    Py_RETURN_NONE;
+}
+
+
+
+
+#if !GREENLET_PY313
+PyDoc_STRVAR(mod_get_tstate_trash_delete_nesting_doc,
+             "get_tstate_trash_delete_nesting() -> Integer\n"
+             "\n"
+             "Return the 'trash can' nesting level. Testing only.\n");
+static PyObject*
+mod_get_tstate_trash_delete_nesting(PyObject* UNUSED(module))
+{
+    PyThreadState* tstate = PyThreadState_GET();
+
+#if GREENLET_PY312
+    return PyLong_FromLong(tstate->trash.delete_nesting);
+#else
+    return PyLong_FromLong(tstate->trash_delete_nesting);
+#endif
+}
+#endif
+
+
+
+
+static PyMethodDef GreenMethods[] = {
+    {
+      .ml_name="getcurrent",
+      .ml_meth=(PyCFunction)mod_getcurrent,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_getcurrent_doc
+    },
+    {
+      .ml_name="settrace",
+      .ml_meth=(PyCFunction)mod_settrace,
+      .ml_flags=METH_VARARGS,
+      .ml_doc=mod_settrace_doc
+    },
+    {
+      .ml_name="gettrace",
+      .ml_meth=(PyCFunction)mod_gettrace,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_gettrace_doc
+    },
+    {
+      .ml_name="set_thread_local",
+      .ml_meth=(PyCFunction)mod_set_thread_local,
+      .ml_flags=METH_VARARGS,
+      .ml_doc=mod_set_thread_local_doc
+    },
+    {
+      .ml_name="get_pending_cleanup_count",
+      .ml_meth=(PyCFunction)mod_get_pending_cleanup_count,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_get_pending_cleanup_count_doc
+    },
+    {
+      .ml_name="get_total_main_greenlets",
+      .ml_meth=(PyCFunction)mod_get_total_main_greenlets,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_get_total_main_greenlets_doc
+    },
+    {
+      .ml_name="get_clocks_used_doing_optional_cleanup",
+      .ml_meth=(PyCFunction)mod_get_clocks_used_doing_optional_cleanup,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_get_clocks_used_doing_optional_cleanup_doc
+    },
+    {
+      .ml_name="enable_optional_cleanup",
+      .ml_meth=(PyCFunction)mod_enable_optional_cleanup,
+      .ml_flags=METH_O,
+      .ml_doc=mod_enable_optional_cleanup_doc
+    },
+#if !GREENLET_PY313
+    {
+      .ml_name="get_tstate_trash_delete_nesting",
+      .ml_meth=(PyCFunction)mod_get_tstate_trash_delete_nesting,
+      .ml_flags=METH_NOARGS,
+      .ml_doc=mod_get_tstate_trash_delete_nesting_doc
+    },
+#endif
+    {.ml_name=NULL, .ml_meth=NULL} /* Sentinel */
+};
+
+static const char* const copy_on_greentype[] = {
+    "getcurrent",
+    "error",
+    "GreenletExit",
+    "settrace",
+    "gettrace",
+    NULL
+};
+
+static struct PyModuleDef greenlet_module_def = {
+    .m_base=PyModuleDef_HEAD_INIT,
+    .m_name="greenlet._greenlet",
+    .m_doc=NULL,
+    .m_size=-1,
+    .m_methods=GreenMethods,
+};
+
+
+#endif
+
+#ifdef __clang__
+#    pragma clang diagnostic pop
+#elif defined(__GNUC__)
+#    pragma GCC diagnostic pop
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TBrokenGreenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TBrokenGreenlet.cpp
new file mode 100644
index 00000000..7e9ab5be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TBrokenGreenlet.cpp
@@ -0,0 +1,45 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of greenlet::UserGreenlet.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+
+#include "TGreenlet.hpp"
+
+namespace greenlet {
+
+void* BrokenGreenlet::operator new(size_t UNUSED(count))
+{
+    return allocator.allocate(1);
+}
+
+
+void BrokenGreenlet::operator delete(void* ptr)
+{
+    return allocator.deallocate(static_cast<BrokenGreenlet*>(ptr),
+                                1);
+}
+
+greenlet::PythonAllocator<greenlet::BrokenGreenlet> greenlet::BrokenGreenlet::allocator;
+
+bool
+BrokenGreenlet::force_slp_switch_error() const noexcept
+{
+    return this->_force_slp_switch_error;
+}
+
+UserGreenlet::switchstack_result_t BrokenGreenlet::g_switchstack(void)
+{
+  if (this->_force_switch_error) {
+    return switchstack_result_t(-1);
+  }
+  return UserGreenlet::g_switchstack();
+}
+
+}; //namespace greenlet
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TExceptionState.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TExceptionState.cpp
new file mode 100644
index 00000000..08a94ae8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TExceptionState.cpp
@@ -0,0 +1,62 @@
+#ifndef GREENLET_EXCEPTION_STATE_CPP
+#define GREENLET_EXCEPTION_STATE_CPP
+
+#include <Python.h>
+#include "TGreenlet.hpp"
+
+namespace greenlet {
+
+
+ExceptionState::ExceptionState()
+{
+    this->clear();
+}
+
+void ExceptionState::operator<<(const PyThreadState *const tstate) noexcept
+{
+    this->exc_info = tstate->exc_info;
+    this->exc_state = tstate->exc_state;
+}
+
+void ExceptionState::operator>>(PyThreadState *const tstate) noexcept
+{
+    tstate->exc_state = this->exc_state;
+    tstate->exc_info =
+        this->exc_info ? this->exc_info : &tstate->exc_state;
+    this->clear();
+}
+
+void ExceptionState::clear() noexcept
+{
+    this->exc_info = nullptr;
+    this->exc_state.exc_value = nullptr;
+#if !GREENLET_PY311
+    this->exc_state.exc_type = nullptr;
+    this->exc_state.exc_traceback = nullptr;
+#endif
+    this->exc_state.previous_item = nullptr;
+}
+
+int ExceptionState::tp_traverse(visitproc visit, void* arg) noexcept
+{
+    Py_VISIT(this->exc_state.exc_value);
+#if !GREENLET_PY311
+    Py_VISIT(this->exc_state.exc_type);
+    Py_VISIT(this->exc_state.exc_traceback);
+#endif
+    return 0;
+}
+
+void ExceptionState::tp_clear() noexcept
+{
+    Py_CLEAR(this->exc_state.exc_value);
+#if !GREENLET_PY311
+    Py_CLEAR(this->exc_state.exc_type);
+    Py_CLEAR(this->exc_state.exc_traceback);
+#endif
+}
+
+
+}; // namespace greenlet
+
+#endif // GREENLET_EXCEPTION_STATE_CPP
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.cpp
new file mode 100644
index 00000000..4698a178
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.cpp
@@ -0,0 +1,718 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of greenlet::Greenlet.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#ifndef TGREENLET_CPP
+#define TGREENLET_CPP
+#include "greenlet_internal.hpp"
+#include "TGreenlet.hpp"
+
+
+#include "TGreenletGlobals.cpp"
+#include "TThreadStateDestroy.cpp"
+
+namespace greenlet {
+
+Greenlet::Greenlet(PyGreenlet* p)
+    :  Greenlet(p, StackState())
+{
+}
+
+Greenlet::Greenlet(PyGreenlet* p, const StackState& initial_stack)
+    :  _self(p), stack_state(initial_stack)
+{
+    assert(p->pimpl == nullptr);
+    p->pimpl = this;
+}
+
+Greenlet::~Greenlet()
+{
+    // XXX: Can't do this. tp_clear is a virtual function, and by the
+    // time we're here, we've sliced off our child classes.
+    //this->tp_clear();
+    this->_self->pimpl = nullptr;
+}
+
+bool
+Greenlet::force_slp_switch_error() const noexcept
+{
+    return false;
+}
+
+void
+Greenlet::release_args()
+{
+    this->switch_args.CLEAR();
+}
+
+/**
+ * CAUTION: This will allocate memory and may trigger garbage
+ * collection and arbitrary Python code.
+ */
+OwnedObject
+Greenlet::throw_GreenletExit_during_dealloc(const ThreadState& UNUSED(current_thread_state))
+{
+    // If we're killed because we lost all references in the
+    // middle of a switch, that's ok. Don't reset the args/kwargs,
+    // we still want to pass them to the parent.
+    PyErr_SetString(mod_globs->PyExc_GreenletExit,
+                    "Killing the greenlet because all references have vanished.");
+    // To get here it had to have run before
+    return this->g_switch();
+}
+
+inline void
+Greenlet::slp_restore_state() noexcept
+{
+#ifdef SLP_BEFORE_RESTORE_STATE
+    SLP_BEFORE_RESTORE_STATE();
+#endif
+    this->stack_state.copy_heap_to_stack(
+           this->thread_state()->borrow_current()->stack_state);
+}
+
+
+inline int
+Greenlet::slp_save_state(char *const stackref) noexcept
+{
+    // XXX: This used to happen in the middle, before saving, but
+    // after finding the next owner. Does that matter? This is
+    // only defined for Sparc/GCC where it flushes register
+    // windows to the stack (I think)
+#ifdef SLP_BEFORE_SAVE_STATE
+    SLP_BEFORE_SAVE_STATE();
+#endif
+    return this->stack_state.copy_stack_to_heap(stackref,
+                                                this->thread_state()->borrow_current()->stack_state);
+}
+
+/**
+ * CAUTION: This will allocate memory and may trigger garbage
+ * collection and arbitrary Python code.
+ */
+OwnedObject
+Greenlet::on_switchstack_or_initialstub_failure(
+    Greenlet* target,
+    const Greenlet::switchstack_result_t& err,
+    const bool target_was_me,
+    const bool was_initial_stub)
+{
+    // If we get here, either g_initialstub()
+    // failed, or g_switchstack() failed. Either one of those
+    // cases SHOULD leave us in the original greenlet with a valid stack.
+    if (!PyErr_Occurred()) {
+        PyErr_SetString(
+            PyExc_SystemError,
+            was_initial_stub
+            ? "Failed to switch stacks into a greenlet for the first time."
+            : "Failed to switch stacks into a running greenlet.");
+    }
+    this->release_args();
+
+    if (target && !target_was_me) {
+        target->murder_in_place();
+    }
+
+    assert(!err.the_new_current_greenlet);
+    assert(!err.origin_greenlet);
+    return OwnedObject();
+
+}
+
+OwnedGreenlet
+Greenlet::g_switchstack_success() noexcept
+{
+    PyThreadState* tstate = PyThreadState_GET();
+    // restore the saved state
+    this->python_state >> tstate;
+    this->exception_state >> tstate;
+
+    // The thread state hasn't been changed yet.
+    ThreadState* thread_state = this->thread_state();
+    OwnedGreenlet result(thread_state->get_current());
+    thread_state->set_current(this->self());
+    //assert(thread_state->borrow_current().borrow() == this->_self);
+    return result;
+}
+
+Greenlet::switchstack_result_t
+Greenlet::g_switchstack(void)
+{
+    // if any of these assertions fail, it's likely because we
+    // switched away and tried to switch back to us. Early stages of
+    // switching are not reentrant because we re-use ``this->args()``.
+    // Switching away would happen if we trigger a garbage collection
+    // (by just using some Python APIs that happen to allocate Python
+    // objects) and some garbage had weakref callbacks or __del__ that
+    // switches (people don't write code like that by hand, but with
+    // gevent it's possible without realizing it)
+    assert(this->args() || PyErr_Occurred());
+    { /* save state */
+        if (this->thread_state()->is_current(this->self())) {
+            // Hmm, nothing to do.
+            // TODO: Does this bypass trace events that are
+            // important?
+            return switchstack_result_t(0,
+                                        this, this->thread_state()->borrow_current());
+        }
+        BorrowedGreenlet current = this->thread_state()->borrow_current();
+        PyThreadState* tstate = PyThreadState_GET();
+
+        current->python_state << tstate;
+        current->exception_state << tstate;
+        this->python_state.will_switch_from(tstate);
+        switching_thread_state = this;
+        current->expose_frames();
+    }
+    assert(this->args() || PyErr_Occurred());
+    // If this is the first switch into a greenlet, this will
+    // return twice, once with 1 in the new greenlet, once with 0
+    // in the origin.
+    int err;
+    if (this->force_slp_switch_error()) {
+        err = -1;
+    }
+    else {
+        err = slp_switch();
+    }
+
+    if (err < 0) { /* error */
+        // Tested by
+        // test_greenlet.TestBrokenGreenlets.test_failed_to_slp_switch_into_running
+        //
+        // It's not clear if it's worth trying to clean up and
+        // continue here. Failing to switch stacks is a big deal which
+        // may not be recoverable (who knows what state the stack is in).
+        // Also, we've stolen references in preparation for calling
+        // ``g_switchstack_success()`` and we don't have a clean
+        // mechanism for backing that all out.
+        Py_FatalError("greenlet: Failed low-level slp_switch(). The stack is probably corrupt.");
+    }
+
+    // No stack-based variables are valid anymore.
+
+    // But the global is volatile so we can reload it without the
+    // compiler caching it from earlier.
+    Greenlet* greenlet_that_switched_in = switching_thread_state; // aka this
+    switching_thread_state = nullptr;
+    // except that no stack variables are valid, we would:
+    // assert(this == greenlet_that_switched_in);
+
+    // switchstack success is where we restore the exception state,
+    // etc. It returns the origin greenlet because its convenient.
+
+    OwnedGreenlet origin = greenlet_that_switched_in->g_switchstack_success();
+    assert(greenlet_that_switched_in->args() || PyErr_Occurred());
+    return switchstack_result_t(err, greenlet_that_switched_in, origin);
+}
+
+
+inline void
+Greenlet::check_switch_allowed() const
+{
+    // TODO: Make this take a parameter of the current greenlet,
+    // or current main greenlet, to make the check for
+    // cross-thread switching cheaper. Surely somewhere up the
+    // call stack we've already accessed the thread local variable.
+
+    // We expect to always have a main greenlet now; accessing the thread state
+    // created it. However, if we get here and cleanup has already
+    // begun because we're a greenlet that was running in a
+    // (now dead) thread, these invariants will not hold true. In
+    // fact, accessing `this->thread_state` may not even be possible.
+
+    // If the thread this greenlet was running in is dead,
+    // we'll still have a reference to a main greenlet, but the
+    // thread state pointer we have is bogus.
+    // TODO: Give the objects an API to determine if they belong
+    // to a dead thread.
+
+    const BorrowedMainGreenlet main_greenlet = this->find_main_greenlet_in_lineage();
+
+    if (!main_greenlet) {
+        throw PyErrOccurred(mod_globs->PyExc_GreenletError,
+                            "cannot switch to a garbage collected greenlet");
+    }
+
+    if (!main_greenlet->thread_state()) {
+        throw PyErrOccurred(mod_globs->PyExc_GreenletError,
+                            "cannot switch to a different thread (which happens to have exited)");
+    }
+
+    // The main greenlet we found was from the .parent lineage.
+    // That may or may not have any relationship to the main
+    // greenlet of the running thread. We can't actually access
+    // our this->thread_state members to try to check that,
+    // because it could be in the process of getting destroyed,
+    // but setting the main_greenlet->thread_state member to NULL
+    // may not be visible yet. So we need to check against the
+    // current thread state (once the cheaper checks are out of
+    // the way)
+    const BorrowedMainGreenlet current_main_greenlet = GET_THREAD_STATE().state().borrow_main_greenlet();
+    if (
+        // lineage main greenlet is not this thread's greenlet
+        current_main_greenlet != main_greenlet
+        || (
+            // atteched to some thread
+            this->main_greenlet()
+            // XXX: Same condition as above. Was this supposed to be
+            // this->main_greenlet()?
+            && current_main_greenlet != main_greenlet)
+        // switching into a known dead thread (XXX: which, if we get here,
+        // is bad, because we just accessed the thread state, which is
+        // gone!)
+        || (!current_main_greenlet->thread_state())) {
+        // CAUTION: This may trigger memory allocations, gc, and
+        // arbitrary Python code.
+        throw PyErrOccurred(
+            mod_globs->PyExc_GreenletError,
+            "Cannot switch to a different thread\n\tCurrent:  %R\n\tExpected: %R",
+            current_main_greenlet, main_greenlet);
+    }
+}
+
+const OwnedObject
+Greenlet::context() const
+{
+    using greenlet::PythonStateContext;
+    OwnedObject result;
+
+    if (this->is_currently_running_in_some_thread()) {
+        /* Currently running greenlet: context is stored in the thread state,
+           not the greenlet object. */
+        if (GET_THREAD_STATE().state().is_current(this->self())) {
+            result = PythonStateContext::context(PyThreadState_GET());
+        }
+        else {
+            throw ValueError(
+                            "cannot get context of a "
+                            "greenlet that is running in a different thread");
+        }
+    }
+    else {
+        /* Greenlet is not running: just return context. */
+        result = this->python_state.context();
+    }
+    if (!result) {
+        result = OwnedObject::None();
+    }
+    return result;
+}
+
+
+void
+Greenlet::context(BorrowedObject given)
+{
+    using greenlet::PythonStateContext;
+    if (!given) {
+        throw AttributeError("can't delete context attribute");
+    }
+    if (given.is_None()) {
+        /* "Empty context" is stored as NULL, not None. */
+        given = nullptr;
+    }
+
+    //checks type, incrs refcnt
+    greenlet::refs::OwnedContext context(given);
+    PyThreadState* tstate = PyThreadState_GET();
+
+    if (this->is_currently_running_in_some_thread()) {
+        if (!GET_THREAD_STATE().state().is_current(this->self())) {
+            throw ValueError("cannot set context of a greenlet"
+                             " that is running in a different thread");
+        }
+
+        /* Currently running greenlet: context is stored in the thread state,
+           not the greenlet object. */
+        OwnedObject octx = OwnedObject::consuming(PythonStateContext::context(tstate));
+        PythonStateContext::context(tstate, context.relinquish_ownership());
+    }
+    else {
+        /* Greenlet is not running: just set context. Note that the
+           greenlet may be dead.*/
+        this->python_state.context() = context;
+    }
+}
+
+/**
+ * CAUTION: May invoke arbitrary Python code.
+ *
+ * Figure out what the result of ``greenlet.switch(arg, kwargs)``
+ * should be and transfers ownership of it to the left-hand-side.
+ *
+ * If switch() was just passed an arg tuple, then we'll just return that.
+ * If only keyword arguments were passed, then we'll pass the keyword
+ * argument dict. Otherwise, we'll create a tuple of (args, kwargs) and
+ * return both.
+ *
+ * CAUTION: This may allocate a new tuple object, which may
+ * cause the Python garbage collector to run, which in turn may
+ * run arbitrary Python code that switches.
+ */
+OwnedObject& operator<<=(OwnedObject& lhs, greenlet::SwitchingArgs& rhs) noexcept
+{
+    // Because this may invoke arbitrary Python code, which could
+    // result in switching back to us, we need to get the
+    // arguments locally on the stack.
+    assert(rhs);
+    OwnedObject args = rhs.args();
+    OwnedObject kwargs = rhs.kwargs();
+    rhs.CLEAR();
+    // We shouldn't be called twice for the same switch.
+    assert(args || kwargs);
+    assert(!rhs);
+
+    if (!kwargs) {
+        lhs = args;
+    }
+    else if (!PyDict_Size(kwargs.borrow())) {
+        lhs = args;
+    }
+    else if (!PySequence_Length(args.borrow())) {
+        lhs = kwargs;
+    }
+    else {
+        // PyTuple_Pack allocates memory, may GC, may run arbitrary
+        // Python code.
+        lhs = OwnedObject::consuming(PyTuple_Pack(2, args.borrow(), kwargs.borrow()));
+    }
+    return lhs;
+}
+
+static OwnedObject
+g_handle_exit(const OwnedObject& greenlet_result)
+{
+    if (!greenlet_result && mod_globs->PyExc_GreenletExit.PyExceptionMatches()) {
+        /* catch and ignore GreenletExit */
+        PyErrFetchParam val;
+        PyErr_Fetch(PyErrFetchParam(), val, PyErrFetchParam());
+        if (!val) {
+            return OwnedObject::None();
+        }
+        return OwnedObject(val);
+    }
+
+    if (greenlet_result) {
+        // package the result into a 1-tuple
+        // PyTuple_Pack increments the reference of its arguments,
+        // so we always need to decref the greenlet result;
+        // the owner will do that.
+        return OwnedObject::consuming(PyTuple_Pack(1, greenlet_result.borrow()));
+    }
+
+    return OwnedObject();
+}
+
+
+
+/**
+ * May run arbitrary Python code.
+ */
+OwnedObject
+Greenlet::g_switch_finish(const switchstack_result_t& err)
+{
+    assert(err.the_new_current_greenlet == this);
+
+    ThreadState& state = *this->thread_state();
+    // Because calling the trace function could do arbitrary things,
+    // including switching away from this greenlet and then maybe
+    // switching back, we need to capture the arguments now so that
+    // they don't change.
+    OwnedObject result;
+    if (this->args()) {
+        result <<= this->args();
+    }
+    else {
+        assert(PyErr_Occurred());
+    }
+    assert(!this->args());
+    try {
+        // Our only caller handles the bad error case
+        assert(err.status >= 0);
+        assert(state.borrow_current() == this->self());
+        if (OwnedObject tracefunc = state.get_tracefunc()) {
+            assert(result || PyErr_Occurred());
+            g_calltrace(tracefunc,
+                        result ? mod_globs->event_switch : mod_globs->event_throw,
+                        err.origin_greenlet,
+                        this->self());
+        }
+        // The above could have invoked arbitrary Python code, but
+        // it couldn't switch back to this object and *also*
+        // throw an exception, so the args won't have changed.
+
+        if (PyErr_Occurred()) {
+            // We get here if we fell of the end of the run() function
+            // raising an exception. The switch itself was
+            // successful, but the function raised.
+            // valgrind reports that memory allocated here can still
+            // be reached after a test run.
+            throw PyErrOccurred::from_current();
+        }
+        return result;
+    }
+    catch (const PyErrOccurred&) {
+        /* Turn switch errors into switch throws */
+        /* Turn trace errors into switch throws */
+        this->release_args();
+        throw;
+    }
+}
+
+void
+Greenlet::g_calltrace(const OwnedObject& tracefunc,
+                      const greenlet::refs::ImmortalEventName& event,
+                      const BorrowedGreenlet& origin,
+                      const BorrowedGreenlet& target)
+{
+    PyErrPieces saved_exc;
+    try {
+        TracingGuard tracing_guard;
+        // TODO: We have saved the active exception (if any) that's
+        // about to be raised. In the 'throw' case, we could provide
+        // the exception to the tracefunction, which seems very helpful.
+        tracing_guard.CallTraceFunction(tracefunc, event, origin, target);
+    }
+    catch (const PyErrOccurred&) {
+        // In case of exceptions trace function is removed,
+        // and any existing exception is replaced with the tracing
+        // exception.
+        GET_THREAD_STATE().state().set_tracefunc(Py_None);
+        throw;
+    }
+
+    saved_exc.PyErrRestore();
+    assert(
+        (event == mod_globs->event_throw && PyErr_Occurred())
+        || (event == mod_globs->event_switch && !PyErr_Occurred())
+    );
+}
+
+void
+Greenlet::murder_in_place()
+{
+    if (this->active()) {
+        assert(!this->is_currently_running_in_some_thread());
+        this->deactivate_and_free();
+    }
+}
+
+inline void
+Greenlet::deactivate_and_free()
+{
+    if (!this->active()) {
+        return;
+    }
+    // Throw away any saved stack.
+    this->stack_state = StackState();
+    assert(!this->stack_state.active());
+    // Throw away any Python references.
+    // We're holding a borrowed reference to the last
+    // frame we executed. Since we borrowed it, the
+    // normal traversal, clear, and dealloc functions
+    // ignore it, meaning it leaks. (The thread state
+    // object can't find it to clear it when that's
+    // deallocated either, because by definition if we
+    // got an object on this list, it wasn't
+    // running and the thread state doesn't have
+    // this frame.)
+    // So here, we *do* clear it.
+    this->python_state.tp_clear(true);
+}
+
+bool
+Greenlet::belongs_to_thread(const ThreadState* thread_state) const
+{
+    if (!this->thread_state() // not running anywhere, or thread
+                              // exited
+        || !thread_state) { // same, or there is no thread state.
+        return false;
+    }
+    return true;
+}
+
+
+void
+Greenlet::deallocing_greenlet_in_thread(const ThreadState* current_thread_state)
+{
+    /* Cannot raise an exception to kill the greenlet if
+       it is not running in the same thread! */
+    if (this->belongs_to_thread(current_thread_state)) {
+        assert(current_thread_state);
+        // To get here it had to have run before
+        /* Send the greenlet a GreenletExit exception. */
+
+        // We don't care about the return value, only whether an
+        // exception happened.
+        this->throw_GreenletExit_during_dealloc(*current_thread_state);
+        return;
+    }
+
+    // Not the same thread! Temporarily save the greenlet
+    // into its thread's deleteme list, *if* it exists.
+    // If that thread has already exited, and processed its pending
+    // cleanup, we'll never be able to clean everything up: we won't
+    // be able to raise an exception.
+    // That's mostly OK! Since we can't add it to a list, our refcount
+    // won't increase, and we'll go ahead with the DECREFs later.
+    ThreadState *const  thread_state = this->thread_state();
+    if (thread_state) {
+        thread_state->delete_when_thread_running(this->self());
+    }
+    else {
+        // The thread is dead, we can't raise an exception.
+        // We need to make it look non-active, though, so that dealloc
+        // finishes killing it.
+        this->deactivate_and_free();
+    }
+    return;
+}
+
+
+int
+Greenlet::tp_traverse(visitproc visit, void* arg)
+{
+
+    int result;
+    if ((result = this->exception_state.tp_traverse(visit, arg)) != 0) {
+        return result;
+    }
+    //XXX: This is ugly. But so is handling everything having to do
+    //with the top frame.
+    bool visit_top_frame = this->was_running_in_dead_thread();
+    // When true, the thread is dead. Our implicit weak reference to the
+    // frame is now all that's left; we consider ourselves to
+    // strongly own it now.
+    if ((result = this->python_state.tp_traverse(visit, arg, visit_top_frame)) != 0) {
+        return result;
+    }
+    return 0;
+}
+
+int
+Greenlet::tp_clear()
+{
+    bool own_top_frame = this->was_running_in_dead_thread();
+    this->exception_state.tp_clear();
+    this->python_state.tp_clear(own_top_frame);
+    return 0;
+}
+
+bool Greenlet::is_currently_running_in_some_thread() const
+{
+    return this->stack_state.active() && !this->python_state.top_frame();
+}
+
+#if GREENLET_PY312
+void GREENLET_NOINLINE(Greenlet::expose_frames)()
+{
+    if (!this->python_state.top_frame()) {
+        return;
+    }
+
+    _PyInterpreterFrame* last_complete_iframe = nullptr;
+    _PyInterpreterFrame* iframe = this->python_state.top_frame()->f_frame;
+    while (iframe) {
+        // We must make a copy before looking at the iframe contents,
+        // since iframe might point to a portion of the greenlet's C stack
+        // that was spilled when switching greenlets.
+        _PyInterpreterFrame iframe_copy;
+        this->stack_state.copy_from_stack(&iframe_copy, iframe, sizeof(*iframe));
+        if (!_PyFrame_IsIncomplete(&iframe_copy)) {
+            // If the iframe were OWNED_BY_CSTACK then it would always be
+            // incomplete. Since it's not incomplete, it's not on the C stack
+            // and we can access it through the original `iframe` pointer
+            // directly.  This is important since GetFrameObject might
+            // lazily _create_ the frame object and we don't want the
+            // interpreter to lose track of it.
+            assert(iframe_copy.owner != FRAME_OWNED_BY_CSTACK);
+
+            // We really want to just write:
+            //     PyFrameObject* frame = _PyFrame_GetFrameObject(iframe);
+            // but _PyFrame_GetFrameObject calls _PyFrame_MakeAndSetFrameObject
+            // which is not a visible symbol in libpython. The easiest
+            // way to get a public function to call it is using
+            // PyFrame_GetBack, which is defined as follows:
+            //     assert(frame != NULL);
+            //     assert(!_PyFrame_IsIncomplete(frame->f_frame));
+            //     PyFrameObject *back = frame->f_back;
+            //     if (back == NULL) {
+            //         _PyInterpreterFrame *prev = frame->f_frame->previous;
+            //         prev = _PyFrame_GetFirstComplete(prev);
+            //         if (prev) {
+            //             back = _PyFrame_GetFrameObject(prev);
+            //         }
+            //     }
+            //     return (PyFrameObject*)Py_XNewRef(back);
+            if (!iframe->frame_obj) {
+                PyFrameObject dummy_frame;
+                _PyInterpreterFrame dummy_iframe;
+                dummy_frame.f_back = nullptr;
+                dummy_frame.f_frame = &dummy_iframe;
+                // force the iframe to be considered complete without
+                // needing to check its code object:
+                dummy_iframe.owner = FRAME_OWNED_BY_GENERATOR;
+                dummy_iframe.previous = iframe;
+                assert(!_PyFrame_IsIncomplete(&dummy_iframe));
+                // Drop the returned reference immediately; the iframe
+                // continues to hold a strong reference
+                Py_XDECREF(PyFrame_GetBack(&dummy_frame));
+                assert(iframe->frame_obj);
+            }
+
+            // This is a complete frame, so make the last one of those we saw
+            // point at it, bypassing any incomplete frames (which may have
+            // been on the C stack) in between the two. We're overwriting
+            // last_complete_iframe->previous and need that to be reversible,
+            // so we store the original previous ptr in the frame object
+            // (which we must have created on a previous iteration through
+            // this loop). The frame object has a bunch of storage that is
+            // only used when its iframe is OWNED_BY_FRAME_OBJECT, which only
+            // occurs when the frame object outlives the frame's execution,
+            // which can't have happened yet because the frame is currently
+            // executing as far as the interpreter is concerned. So, we can
+            // reuse it for our own purposes.
+            assert(iframe->owner == FRAME_OWNED_BY_THREAD
+                   || iframe->owner == FRAME_OWNED_BY_GENERATOR);
+            if (last_complete_iframe) {
+                assert(last_complete_iframe->frame_obj);
+                memcpy(&last_complete_iframe->frame_obj->_f_frame_data[0],
+                       &last_complete_iframe->previous, sizeof(void *));
+                last_complete_iframe->previous = iframe;
+            }
+            last_complete_iframe = iframe;
+        }
+        // Frames that are OWNED_BY_FRAME_OBJECT are linked via the
+        // frame's f_back while all others are linked via the iframe's
+        // previous ptr. Since all the frames we traverse are running
+        // as far as the interpreter is concerned, we don't have to
+        // worry about the OWNED_BY_FRAME_OBJECT case.
+        iframe = iframe_copy.previous;
+    }
+
+    // Give the outermost complete iframe a null previous pointer to
+    // account for any potential incomplete/C-stack iframes between it
+    // and the actual top-of-stack
+    if (last_complete_iframe) {
+        assert(last_complete_iframe->frame_obj);
+        memcpy(&last_complete_iframe->frame_obj->_f_frame_data[0],
+               &last_complete_iframe->previous, sizeof(void *));
+        last_complete_iframe->previous = nullptr;
+    }
+}
+#else
+void Greenlet::expose_frames()
+{
+
+}
+#endif
+
+}; // namespace greenlet
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.hpp
new file mode 100644
index 00000000..512f7fb3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenlet.hpp
@@ -0,0 +1,813 @@
+#ifndef GREENLET_GREENLET_HPP
+#define GREENLET_GREENLET_HPP
+/*
+ * Declarations of the core data structures.
+*/
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+
+#include "greenlet_compiler_compat.hpp"
+#include "greenlet_refs.hpp"
+#include "greenlet_cpython_compat.hpp"
+#include "greenlet_allocator.hpp"
+
+using greenlet::refs::OwnedObject;
+using greenlet::refs::OwnedGreenlet;
+using greenlet::refs::OwnedMainGreenlet;
+using greenlet::refs::BorrowedGreenlet;
+
+#if PY_VERSION_HEX < 0x30B00A6
+#  define _PyCFrame CFrame
+#  define _PyInterpreterFrame _interpreter_frame
+#endif
+
+#if GREENLET_PY312
+#  define Py_BUILD_CORE
+#  include "internal/pycore_frame.h"
+#endif
+
+// XXX: TODO: Work to remove all virtual functions
+// for speed of calling and size of objects (no vtable).
+// One pattern is the Curiously Recurring Template
+namespace greenlet
+{
+    class ExceptionState
+    {
+    private:
+        G_NO_COPIES_OF_CLS(ExceptionState);
+
+        // Even though these are borrowed objects, we actually own
+        // them, when they're not null.
+        // XXX: Express that in the API.
+    private:
+        _PyErr_StackItem* exc_info;
+        _PyErr_StackItem exc_state;
+    public:
+        ExceptionState();
+        void operator<<(const PyThreadState *const tstate) noexcept;
+        void operator>>(PyThreadState* tstate) noexcept;
+        void clear() noexcept;
+
+        int tp_traverse(visitproc visit, void* arg) noexcept;
+        void tp_clear() noexcept;
+    };
+
+    template<typename T>
+    void operator<<(const PyThreadState *const tstate, T& exc);
+
+    class PythonStateContext
+    {
+    protected:
+        greenlet::refs::OwnedContext _context;
+    public:
+        inline const greenlet::refs::OwnedContext& context() const
+        {
+            return this->_context;
+        }
+        inline greenlet::refs::OwnedContext& context()
+        {
+            return this->_context;
+        }
+
+        inline void tp_clear()
+        {
+            this->_context.CLEAR();
+        }
+
+        template<typename T>
+        inline static PyObject* context(T* tstate)
+        {
+            return tstate->context;
+        }
+
+        template<typename T>
+        inline static void context(T* tstate, PyObject* new_context)
+        {
+            tstate->context = new_context;
+            tstate->context_ver++;
+        }
+    };
+    class SwitchingArgs;
+    class PythonState : public PythonStateContext
+    {
+    public:
+        typedef greenlet::refs::OwnedReference<struct _frame> OwnedFrame;
+    private:
+        G_NO_COPIES_OF_CLS(PythonState);
+        // We own this if we're suspended (although currently we don't
+        // tp_traverse into it; that's a TODO). If we're running, it's
+        // empty. If we get deallocated and *still* have a frame, it
+        // won't be reachable from the place that normally decref's
+        // it, so we need to do it (hence owning it).
+        OwnedFrame _top_frame;
+#if GREENLET_USE_CFRAME
+        _PyCFrame* cframe;
+        int use_tracing;
+#endif
+#if GREENLET_PY312
+        int py_recursion_depth;
+        int c_recursion_depth;
+#else
+        int recursion_depth;
+#endif
+#if GREENLET_PY313
+        PyObject *delete_later;
+#else
+        int trash_delete_nesting;
+#endif
+#if GREENLET_PY311
+        _PyInterpreterFrame* current_frame;
+        _PyStackChunk* datastack_chunk;
+        PyObject** datastack_top;
+        PyObject** datastack_limit;
+#endif
+        // The PyInterpreterFrame list on 3.12+ contains some entries that are
+        // on the C stack, which can't be directly accessed while a greenlet is
+        // suspended. In order to keep greenlet gr_frame introspection working,
+        // we adjust stack switching to rewrite the interpreter frame list
+        // to skip these C-stack frames; we call this "exposing" the greenlet's
+        // frames because it makes them valid to work with in Python. Then when
+        // the greenlet is resumed we need to remember to reverse the operation
+        // we did. The C-stack frames are "entry frames" which are a low-level
+        // interpreter detail; they're not needed for introspection, but do
+        // need to be present for the eval loop to work.
+        void unexpose_frames();
+
+    public:
+
+        PythonState();
+        // You can use this for testing whether we have a frame
+        // or not. It returns const so they can't modify it.
+        const OwnedFrame& top_frame() const noexcept;
+
+        inline void operator<<(const PyThreadState *const tstate) noexcept;
+        inline void operator>>(PyThreadState* tstate) noexcept;
+        void clear() noexcept;
+
+        int tp_traverse(visitproc visit, void* arg, bool visit_top_frame) noexcept;
+        void tp_clear(bool own_top_frame) noexcept;
+        void set_initial_state(const PyThreadState* const tstate) noexcept;
+#if GREENLET_USE_CFRAME
+        void set_new_cframe(_PyCFrame& frame) noexcept;
+#endif
+
+        void may_switch_away() noexcept;
+        inline void will_switch_from(PyThreadState *const origin_tstate) noexcept;
+        void did_finish(PyThreadState* tstate) noexcept;
+    };
+
+    class StackState
+    {
+        // By having only plain C (POD) members, no virtual functions
+        // or bases, we get a trivial assignment operator generated
+        // for us. However, that's not safe since we do manage memory.
+        // So we declare an assignment operator that only works if we
+        // don't have any memory allocated. (We don't use
+        // std::shared_ptr for reference counting just to keep this
+        // object small)
+    private:
+        char* _stack_start;
+        char* stack_stop;
+        char* stack_copy;
+        intptr_t _stack_saved;
+        StackState* stack_prev;
+        inline int copy_stack_to_heap_up_to(const char* const stop) noexcept;
+        inline void free_stack_copy() noexcept;
+
+    public:
+        /**
+         * Creates a started, but inactive, state, using *current*
+         * as the previous.
+         */
+        StackState(void* mark, StackState& current);
+        /**
+         * Creates an inactive, unstarted, state.
+         */
+        StackState();
+        ~StackState();
+        StackState(const StackState& other);
+        StackState& operator=(const StackState& other);
+        inline void copy_heap_to_stack(const StackState& current) noexcept;
+        inline int copy_stack_to_heap(char* const stackref, const StackState& current) noexcept;
+        inline bool started() const noexcept;
+        inline bool main() const noexcept;
+        inline bool active() const noexcept;
+        inline void set_active() noexcept;
+        inline void set_inactive() noexcept;
+        inline intptr_t stack_saved() const noexcept;
+        inline char* stack_start() const noexcept;
+        static inline StackState make_main() noexcept;
+#ifdef GREENLET_USE_STDIO
+        friend std::ostream& operator<<(std::ostream& os, const StackState& s);
+#endif
+
+        // Fill in [dest, dest + n) with the values that would be at
+        // [src, src + n) while this greenlet is running. This is like memcpy
+        // except that if the greenlet is suspended it accounts for the portion
+        // of the greenlet's stack that was spilled to the heap. `src` may
+        // be on this greenlet's stack, or on the heap, but not on a different
+        // greenlet's stack.
+        void copy_from_stack(void* dest, const void* src, size_t n) const;
+    };
+#ifdef GREENLET_USE_STDIO
+    std::ostream& operator<<(std::ostream& os, const StackState& s);
+#endif
+
+    class SwitchingArgs
+    {
+    private:
+        G_NO_ASSIGNMENT_OF_CLS(SwitchingArgs);
+        // If args and kwargs are both false (NULL), this is a *throw*, not a
+        // switch. PyErr_... must have been called already.
+        OwnedObject _args;
+        OwnedObject _kwargs;
+    public:
+
+        SwitchingArgs()
+        {}
+
+        SwitchingArgs(const OwnedObject& args, const OwnedObject& kwargs)
+            : _args(args),
+              _kwargs(kwargs)
+        {}
+
+        SwitchingArgs(const SwitchingArgs& other)
+            : _args(other._args),
+              _kwargs(other._kwargs)
+        {}
+
+        const OwnedObject& args()
+        {
+            return this->_args;
+        }
+
+        const OwnedObject& kwargs()
+        {
+            return this->_kwargs;
+        }
+
+        /**
+         * Moves ownership from the argument to this object.
+         */
+        SwitchingArgs& operator<<=(SwitchingArgs& other)
+        {
+            if (this != &other) {
+                this->_args = other._args;
+                this->_kwargs = other._kwargs;
+                other.CLEAR();
+            }
+            return *this;
+        }
+
+        /**
+         * Acquires ownership of the argument (consumes the reference).
+         */
+        SwitchingArgs& operator<<=(PyObject* args)
+        {
+            this->_args = OwnedObject::consuming(args);
+            this->_kwargs.CLEAR();
+            return *this;
+        }
+
+        /**
+         * Acquires ownership of the argument.
+         *
+         * Sets the args to be the given value; clears the kwargs.
+         */
+        SwitchingArgs& operator<<=(OwnedObject& args)
+        {
+            assert(&args != &this->_args);
+            this->_args = args;
+            this->_kwargs.CLEAR();
+            args.CLEAR();
+
+            return *this;
+        }
+
+        explicit operator bool() const noexcept
+        {
+            return this->_args || this->_kwargs;
+        }
+
+        inline void CLEAR()
+        {
+            this->_args.CLEAR();
+            this->_kwargs.CLEAR();
+        }
+
+        const std::string as_str() const noexcept
+        {
+            return PyUnicode_AsUTF8(
+                OwnedObject::consuming(
+                    PyUnicode_FromFormat(
+                        "SwitchingArgs(args=%R, kwargs=%R)",
+                        this->_args.borrow(),
+                        this->_kwargs.borrow()
+                    )
+                ).borrow()
+            );
+        }
+    };
+
+    class ThreadState;
+
+    class UserGreenlet;
+    class MainGreenlet;
+
+    class Greenlet
+    {
+    private:
+        G_NO_COPIES_OF_CLS(Greenlet);
+        PyGreenlet* const _self;
+    private:
+        // XXX: Work to remove these.
+        friend class ThreadState;
+        friend class UserGreenlet;
+        friend class MainGreenlet;
+    protected:
+        ExceptionState exception_state;
+        SwitchingArgs switch_args;
+        StackState stack_state;
+        PythonState python_state;
+        Greenlet(PyGreenlet* p, const StackState& initial_state);
+    public:
+        // This constructor takes ownership of the PyGreenlet, by
+        // setting ``p->pimpl = this;``.
+        Greenlet(PyGreenlet* p);
+        virtual ~Greenlet();
+
+        const OwnedObject context() const;
+
+        // You MUST call this _very_ early in the switching process to
+        // prepare anything that may need prepared. This might perform
+        // garbage collections or otherwise run arbitrary Python code.
+        //
+        // One specific use of it is for Python 3.11+, preventing
+        // running arbitrary code at unsafe times. See
+        // PythonState::may_switch_away().
+        inline void may_switch_away()
+        {
+            this->python_state.may_switch_away();
+        }
+
+        inline void context(refs::BorrowedObject new_context);
+
+        inline SwitchingArgs& args()
+        {
+            return this->switch_args;
+        }
+
+        virtual const refs::BorrowedMainGreenlet main_greenlet() const = 0;
+
+        inline intptr_t stack_saved() const noexcept
+        {
+            return this->stack_state.stack_saved();
+        }
+
+        // This is used by the macro SLP_SAVE_STATE to compute the
+        // difference in stack sizes. It might be nice to handle the
+        // computation ourself, but the type of the result
+        // varies by platform, so doing it in the macro is the
+        // simplest way.
+        inline const char* stack_start() const noexcept
+        {
+            return this->stack_state.stack_start();
+        }
+
+        virtual OwnedObject throw_GreenletExit_during_dealloc(const ThreadState& current_thread_state);
+        virtual OwnedObject g_switch() = 0;
+        /**
+         * Force the greenlet to appear dead. Used when it's not
+         * possible to throw an exception into a greenlet anymore.
+         *
+         * This losses access to the thread state and the main greenlet.
+         */
+        virtual void murder_in_place();
+
+        /**
+         * Called when somebody notices we were running in a dead
+         * thread to allow cleaning up resources (because we can't
+         * raise GreenletExit into it anymore).
+         * This is very similar to ``murder_in_place()``, except that
+         * it DOES NOT lose the main greenlet or thread state.
+         */
+        inline void deactivate_and_free();
+
+
+        // Called when some thread wants to deallocate a greenlet
+        // object.
+        // The thread may or may not be the same thread the greenlet
+        // was running in.
+        // The thread state will be null if the thread the greenlet
+        // was running in was known to have exited.
+        void deallocing_greenlet_in_thread(const ThreadState* current_state);
+
+        // Must be called on 3.12+ before exposing a suspended greenlet's
+        // frames to user code. This rewrites the linked list of interpreter
+        // frames to skip the ones that are being stored on the C stack (which
+        // can't be safely accessed while the greenlet is suspended because
+        // that stack space might be hosting a different greenlet), and
+        // sets PythonState::frames_were_exposed so we remember to restore
+        // the original list before resuming the greenlet. The C-stack frames
+        // are a low-level interpreter implementation detail; while they're
+        // important to the bytecode eval loop, they're superfluous for
+        // introspection purposes.
+        void expose_frames();
+
+
+        // TODO: Figure out how to make these non-public.
+        inline void slp_restore_state() noexcept;
+        inline int slp_save_state(char *const stackref) noexcept;
+
+        inline bool is_currently_running_in_some_thread() const;
+        virtual bool belongs_to_thread(const ThreadState* state) const;
+
+        inline bool started() const
+        {
+            return this->stack_state.started();
+        }
+        inline bool active() const
+        {
+            return this->stack_state.active();
+        }
+        inline bool main() const
+        {
+            return this->stack_state.main();
+        }
+        virtual refs::BorrowedMainGreenlet find_main_greenlet_in_lineage() const = 0;
+
+        virtual const OwnedGreenlet parent() const = 0;
+        virtual void parent(const refs::BorrowedObject new_parent) = 0;
+
+        inline const PythonState::OwnedFrame& top_frame()
+        {
+            return this->python_state.top_frame();
+        }
+
+        virtual const OwnedObject& run() const = 0;
+        virtual void run(const refs::BorrowedObject nrun) = 0;
+
+
+        virtual int tp_traverse(visitproc visit, void* arg);
+        virtual int tp_clear();
+
+
+        // Return the thread state that the greenlet is running in, or
+        // null if the greenlet is not running or the thread is known
+        // to have exited.
+        virtual ThreadState* thread_state() const noexcept = 0;
+
+        // Return true if the greenlet is known to have been running
+        // (active) in a thread that has now exited.
+        virtual bool was_running_in_dead_thread() const noexcept = 0;
+
+        // Return a borrowed greenlet that is the Python object
+        // this object represents.
+        inline BorrowedGreenlet self() const noexcept
+        {
+            return BorrowedGreenlet(this->_self);
+        }
+
+        // For testing. If this returns true, we should pretend that
+        // slp_switch() failed.
+        virtual bool force_slp_switch_error() const noexcept;
+
+    protected:
+        inline void release_args();
+
+        // The functions that must not be inlined are declared virtual.
+        // We also mark them as protected, not private, so that the
+        // compiler is forced to call them through a function pointer.
+        // (A sufficiently smart compiler could directly call a private
+        // virtual function since it can never be overridden in a
+        // subclass).
+
+        // Also TODO: Switch away from integer error codes and to enums,
+        // or throw exceptions when possible.
+        struct switchstack_result_t
+        {
+            int status;
+            Greenlet* the_new_current_greenlet;
+            OwnedGreenlet origin_greenlet;
+
+            switchstack_result_t()
+                : status(0),
+                  the_new_current_greenlet(nullptr)
+            {}
+
+            switchstack_result_t(int err)
+                : status(err),
+                  the_new_current_greenlet(nullptr)
+            {}
+
+            switchstack_result_t(int err, Greenlet* state, OwnedGreenlet& origin)
+                : status(err),
+                  the_new_current_greenlet(state),
+                  origin_greenlet(origin)
+            {
+            }
+
+            switchstack_result_t(int err, Greenlet* state, const BorrowedGreenlet& origin)
+                : status(err),
+                  the_new_current_greenlet(state),
+                  origin_greenlet(origin)
+            {
+            }
+
+            switchstack_result_t(const switchstack_result_t& other)
+                : status(other.status),
+                  the_new_current_greenlet(other.the_new_current_greenlet),
+                  origin_greenlet(other.origin_greenlet)
+            {}
+
+            switchstack_result_t& operator=(const switchstack_result_t& other)
+            {
+                this->status = other.status;
+                this->the_new_current_greenlet = other.the_new_current_greenlet;
+                this->origin_greenlet = other.origin_greenlet;
+                return *this;
+            }
+        };
+
+        OwnedObject on_switchstack_or_initialstub_failure(
+            Greenlet* target,
+            const switchstack_result_t& err,
+            const bool target_was_me=false,
+            const bool was_initial_stub=false);
+
+        // Returns the previous greenlet we just switched away from.
+        virtual OwnedGreenlet g_switchstack_success() noexcept;
+
+
+        // Check the preconditions for switching to this greenlet; if they
+        // aren't met, throws PyErrOccurred. Most callers will want to
+        // catch this and clear the arguments
+        inline void check_switch_allowed() const;
+        class GreenletStartedWhileInPython : public std::runtime_error
+        {
+        public:
+            GreenletStartedWhileInPython() : std::runtime_error("")
+            {}
+        };
+
+    protected:
+
+
+        /**
+           Perform a stack switch into this greenlet.
+
+           This temporarily sets the global variable
+           ``switching_thread_state`` to this greenlet; as soon as the
+           call to ``slp_switch`` completes, this is reset to NULL.
+           Consequently, this depends on the GIL.
+
+           TODO: Adopt the stackman model and pass ``slp_switch`` a
+           callback function and context pointer; this eliminates the
+           need for global variables altogether.
+
+           Because the stack switch happens in this function, this
+           function can't use its own stack (local) variables, set
+           before the switch, and then accessed after the switch.
+
+           Further, you con't even access ``g_thread_state_global``
+           before and after the switch from the global variable.
+           Because it is thread local some compilers cache it in a
+           register/on the stack, notably new versions of MSVC; this
+           breaks with strange crashes sometime later, because writing
+           to anything in ``g_thread_state_global`` after the switch
+           is actually writing to random memory. For this reason, we
+           call a non-inlined function to finish the operation. (XXX:
+           The ``/GT`` MSVC compiler argument probably fixes that.)
+
+           It is very important that stack switch is 'atomic', i.e. no
+           calls into other Python code allowed (except very few that
+           are safe), because global variables are very fragile. (This
+           should no longer be the case with thread-local variables.)
+
+        */
+        // Made virtual to facilitate subclassing UserGreenlet for testing.
+        virtual switchstack_result_t g_switchstack(void);
+
+class TracingGuard
+{
+private:
+    PyThreadState* tstate;
+public:
+    TracingGuard()
+        : tstate(PyThreadState_GET())
+    {
+        PyThreadState_EnterTracing(this->tstate);
+    }
+
+    ~TracingGuard()
+    {
+        PyThreadState_LeaveTracing(this->tstate);
+        this->tstate = nullptr;
+    }
+
+    inline void CallTraceFunction(const OwnedObject& tracefunc,
+                                  const greenlet::refs::ImmortalEventName& event,
+                                  const BorrowedGreenlet& origin,
+                                  const BorrowedGreenlet& target)
+    {
+        // TODO: This calls tracefunc(event, (origin, target)). Add a shortcut
+        // function for that that's specialized to avoid the Py_BuildValue
+        // string parsing, or start with just using "ON" format with PyTuple_Pack(2,
+        // origin, target). That seems like what the N format is meant
+        // for.
+        // XXX: Why does event not automatically cast back to a PyObject?
+        // It tries to call the "deleted constructor ImmortalEventName
+        // const" instead.
+        assert(tracefunc);
+        assert(event);
+        assert(origin);
+        assert(target);
+        greenlet::refs::NewReference retval(
+            PyObject_CallFunction(
+                tracefunc.borrow(),
+                "O(OO)",
+                event.borrow(),
+                origin.borrow(),
+                target.borrow()
+            ));
+        if (!retval) {
+            throw PyErrOccurred::from_current();
+        }
+    }
+};
+
+      static void
+      g_calltrace(const OwnedObject& tracefunc,
+                  const greenlet::refs::ImmortalEventName& event,
+                  const greenlet::refs::BorrowedGreenlet& origin,
+                  const BorrowedGreenlet& target);
+    private:
+        OwnedObject g_switch_finish(const switchstack_result_t& err);
+
+    };
+
+    class UserGreenlet : public Greenlet
+    {
+    private:
+        static greenlet::PythonAllocator<UserGreenlet> allocator;
+        OwnedMainGreenlet _main_greenlet;
+        OwnedObject _run_callable;
+        OwnedGreenlet _parent;
+    public:
+        static void* operator new(size_t UNUSED(count));
+        static void operator delete(void* ptr);
+
+        UserGreenlet(PyGreenlet* p, BorrowedGreenlet the_parent);
+        virtual ~UserGreenlet();
+
+        virtual refs::BorrowedMainGreenlet find_main_greenlet_in_lineage() const;
+        virtual bool was_running_in_dead_thread() const noexcept;
+        virtual ThreadState* thread_state() const noexcept;
+        virtual OwnedObject g_switch();
+        virtual const OwnedObject& run() const
+        {
+            if (this->started() || !this->_run_callable) {
+                throw AttributeError("run");
+            }
+            return this->_run_callable;
+        }
+        virtual void run(const refs::BorrowedObject nrun);
+
+        virtual const OwnedGreenlet parent() const;
+        virtual void parent(const refs::BorrowedObject new_parent);
+
+        virtual const refs::BorrowedMainGreenlet main_greenlet() const;
+
+        virtual void murder_in_place();
+        virtual bool belongs_to_thread(const ThreadState* state) const;
+        virtual int tp_traverse(visitproc visit, void* arg);
+        virtual int tp_clear();
+        class ParentIsCurrentGuard
+        {
+        private:
+            OwnedGreenlet oldparent;
+            UserGreenlet* greenlet;
+            G_NO_COPIES_OF_CLS(ParentIsCurrentGuard);
+        public:
+            ParentIsCurrentGuard(UserGreenlet* p, const ThreadState& thread_state);
+            ~ParentIsCurrentGuard();
+        };
+        virtual OwnedObject throw_GreenletExit_during_dealloc(const ThreadState& current_thread_state);
+    protected:
+        virtual switchstack_result_t g_initialstub(void* mark);
+    private:
+        // This function isn't meant to return.
+        // This accepts raw pointers and the ownership of them at the
+        // same time. The caller should use ``inner_bootstrap(origin.relinquish_ownership())``.
+        void inner_bootstrap(PyGreenlet* origin_greenlet, PyObject* run);
+    };
+
+    class BrokenGreenlet : public UserGreenlet
+    {
+    private:
+        static greenlet::PythonAllocator<BrokenGreenlet> allocator;
+    public:
+        bool _force_switch_error = false;
+        bool _force_slp_switch_error = false;
+
+        static void* operator new(size_t UNUSED(count));
+        static void operator delete(void* ptr);
+        BrokenGreenlet(PyGreenlet* p, BorrowedGreenlet the_parent)
+            : UserGreenlet(p, the_parent)
+        {}
+        virtual ~BrokenGreenlet()
+        {}
+
+        virtual switchstack_result_t g_switchstack(void);
+        virtual bool force_slp_switch_error() const noexcept;
+
+    };
+
+    class MainGreenlet : public Greenlet
+    {
+    private:
+        static greenlet::PythonAllocator<MainGreenlet> allocator;
+        refs::BorrowedMainGreenlet _self;
+        ThreadState* _thread_state;
+        G_NO_COPIES_OF_CLS(MainGreenlet);
+    public:
+        static void* operator new(size_t UNUSED(count));
+        static void operator delete(void* ptr);
+
+        MainGreenlet(refs::BorrowedMainGreenlet::PyType*, ThreadState*);
+        virtual ~MainGreenlet();
+
+
+        virtual const OwnedObject& run() const;
+        virtual void run(const refs::BorrowedObject nrun);
+
+        virtual const OwnedGreenlet parent() const;
+        virtual void parent(const refs::BorrowedObject new_parent);
+
+        virtual const refs::BorrowedMainGreenlet main_greenlet() const;
+
+        virtual refs::BorrowedMainGreenlet find_main_greenlet_in_lineage() const;
+        virtual bool was_running_in_dead_thread() const noexcept;
+        virtual ThreadState* thread_state() const noexcept;
+        void thread_state(ThreadState*) noexcept;
+        virtual OwnedObject g_switch();
+        virtual int tp_traverse(visitproc visit, void* arg);
+    };
+
+    // Instantiate one on the stack to save the GC state,
+    // and then disable GC. When it goes out of scope, GC will be
+    // restored to its original state. Sadly, these APIs are only
+    // available on 3.10+; luckily, we only need them on 3.11+.
+#if GREENLET_PY310
+    class GCDisabledGuard
+    {
+    private:
+        int was_enabled = 0;
+    public:
+        GCDisabledGuard()
+            : was_enabled(PyGC_IsEnabled())
+        {
+            PyGC_Disable();
+        }
+
+        ~GCDisabledGuard()
+        {
+            if (this->was_enabled) {
+                PyGC_Enable();
+            }
+        }
+    };
+#endif
+
+    OwnedObject& operator<<=(OwnedObject& lhs, greenlet::SwitchingArgs& rhs) noexcept;
+
+    //TODO: Greenlet::g_switch() should call this automatically on its
+    //return value. As it is, the module code is calling it.
+    static inline OwnedObject
+    single_result(const OwnedObject& results)
+    {
+        if (results
+            && PyTuple_Check(results.borrow())
+            && PyTuple_GET_SIZE(results.borrow()) == 1) {
+            PyObject* result = PyTuple_GET_ITEM(results.borrow(), 0);
+            assert(result);
+            return OwnedObject::owning(result);
+        }
+        return results;
+    }
+
+
+    static OwnedObject
+    g_handle_exit(const OwnedObject& greenlet_result);
+
+
+    template<typename T>
+    void operator<<(const PyThreadState *const lhs, T& rhs)
+    {
+        rhs.operator<<(lhs);
+    }
+
+} // namespace greenlet ;
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TGreenletGlobals.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenletGlobals.cpp
new file mode 100644
index 00000000..0087d2ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TGreenletGlobals.cpp
@@ -0,0 +1,94 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of GreenletGlobals.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#ifndef T_GREENLET_GLOBALS
+#define T_GREENLET_GLOBALS
+
+#include "greenlet_refs.hpp"
+#include "greenlet_exceptions.hpp"
+#include "greenlet_thread_support.hpp"
+#include "greenlet_internal.hpp"
+
+namespace greenlet {
+
+// This encapsulates what were previously module global "constants"
+// established at init time.
+// This is a step towards Python3 style module state that allows
+// reloading.
+//
+// In an earlier iteration of this code, we used placement new to be
+// able to allocate this object statically still, so that references
+// to its members don't incur an extra pointer indirection.
+// But under some scenarios, that could result in crashes at
+// shutdown because apparently the destructor was getting run twice?
+class GreenletGlobals
+{
+
+public:
+    const greenlet::refs::ImmortalEventName event_switch;
+    const greenlet::refs::ImmortalEventName event_throw;
+    const greenlet::refs::ImmortalException PyExc_GreenletError;
+    const greenlet::refs::ImmortalException PyExc_GreenletExit;
+    const greenlet::refs::ImmortalObject empty_tuple;
+    const greenlet::refs::ImmortalObject empty_dict;
+    const greenlet::refs::ImmortalString str_run;
+    Mutex* const thread_states_to_destroy_lock;
+    greenlet::cleanup_queue_t thread_states_to_destroy;
+
+    GreenletGlobals() :
+        event_switch("switch"),
+        event_throw("throw"),
+        PyExc_GreenletError("greenlet.error"),
+        PyExc_GreenletExit("greenlet.GreenletExit", PyExc_BaseException),
+        empty_tuple(Require(PyTuple_New(0))),
+        empty_dict(Require(PyDict_New())),
+        str_run("run"),
+        thread_states_to_destroy_lock(new Mutex())
+    {}
+
+    ~GreenletGlobals()
+    {
+        // This object is (currently) effectively immortal, and not
+        // just because of those placement new tricks; if we try to
+        // deallocate the static object we allocated, and overwrote,
+        // we would be doing so at C++ teardown time, which is after
+        // the final Python GIL is released, and we can't use the API
+        // then.
+        // (The members will still be destructed, but they also don't
+        // do any deallocation.)
+    }
+
+    void queue_to_destroy(ThreadState* ts) const
+    {
+        // we're currently accessed through a static const object,
+        // implicitly marking our members as const, so code can't just
+        // call push_back (or pop_back) without casting away the
+        // const.
+        //
+        // Do that for callers.
+        greenlet::cleanup_queue_t& q = const_cast<greenlet::cleanup_queue_t&>(this->thread_states_to_destroy);
+        q.push_back(ts);
+    }
+
+    ThreadState* take_next_to_destroy() const
+    {
+        greenlet::cleanup_queue_t& q = const_cast<greenlet::cleanup_queue_t&>(this->thread_states_to_destroy);
+        ThreadState* result = q.back();
+        q.pop_back();
+        return result;
+    }
+};
+
+}; // namespace greenlet
+
+static const greenlet::GreenletGlobals* mod_globs;
+
+#endif // T_GREENLET_GLOBALS
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TMainGreenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TMainGreenlet.cpp
new file mode 100644
index 00000000..a2a9cfe4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TMainGreenlet.cpp
@@ -0,0 +1,153 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of greenlet::MainGreenlet.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#ifndef T_MAIN_GREENLET_CPP
+#define T_MAIN_GREENLET_CPP
+
+#include "TGreenlet.hpp"
+
+
+
+// Protected by the GIL. Incremented when we create a main greenlet,
+// in a new thread, decremented when it is destroyed.
+static Py_ssize_t G_TOTAL_MAIN_GREENLETS;
+
+namespace greenlet {
+greenlet::PythonAllocator<MainGreenlet> MainGreenlet::allocator;
+
+void* MainGreenlet::operator new(size_t UNUSED(count))
+{
+    return allocator.allocate(1);
+}
+
+
+void MainGreenlet::operator delete(void* ptr)
+{
+    return allocator.deallocate(static_cast<MainGreenlet*>(ptr),
+                                1);
+}
+
+
+MainGreenlet::MainGreenlet(PyGreenlet* p, ThreadState* state)
+    : Greenlet(p, StackState::make_main()),
+      _self(p),
+      _thread_state(state)
+{
+    G_TOTAL_MAIN_GREENLETS++;
+}
+
+MainGreenlet::~MainGreenlet()
+{
+    G_TOTAL_MAIN_GREENLETS--;
+    this->tp_clear();
+}
+
+ThreadState*
+MainGreenlet::thread_state() const noexcept
+{
+    return this->_thread_state;
+}
+
+void
+MainGreenlet::thread_state(ThreadState* t) noexcept
+{
+    assert(!t);
+    this->_thread_state = t;
+}
+
+
+const BorrowedMainGreenlet
+MainGreenlet::main_greenlet() const
+{
+    return this->_self;
+}
+
+BorrowedMainGreenlet
+MainGreenlet::find_main_greenlet_in_lineage() const
+{
+    return BorrowedMainGreenlet(this->_self);
+}
+
+bool
+MainGreenlet::was_running_in_dead_thread() const noexcept
+{
+    return !this->_thread_state;
+}
+
+OwnedObject
+MainGreenlet::g_switch()
+{
+    try {
+        this->check_switch_allowed();
+    }
+    catch (const PyErrOccurred&) {
+        this->release_args();
+        throw;
+    }
+
+    switchstack_result_t err = this->g_switchstack();
+    if (err.status < 0) {
+        // XXX: This code path is untested, but it is shared
+        // with the UserGreenlet path that is tested.
+        return this->on_switchstack_or_initialstub_failure(
+            this,
+            err,
+            true, // target was me
+            false // was initial stub
+        );
+    }
+
+    return err.the_new_current_greenlet->g_switch_finish(err);
+}
+
+int
+MainGreenlet::tp_traverse(visitproc visit, void* arg)
+{
+    if (this->_thread_state) {
+        // we've already traversed main, (self), don't do it again.
+        int result = this->_thread_state->tp_traverse(visit, arg, false);
+        if (result) {
+            return result;
+        }
+    }
+    return Greenlet::tp_traverse(visit, arg);
+}
+
+const OwnedObject&
+MainGreenlet::run() const
+{
+    throw AttributeError("Main greenlets do not have a run attribute.");
+}
+
+void
+MainGreenlet::run(const BorrowedObject UNUSED(nrun))
+{
+   throw AttributeError("Main greenlets do not have a run attribute.");
+}
+
+void
+MainGreenlet::parent(const BorrowedObject raw_new_parent)
+{
+    if (!raw_new_parent) {
+        throw AttributeError("can't delete attribute");
+    }
+    throw AttributeError("cannot set the parent of a main greenlet");
+}
+
+const OwnedGreenlet
+MainGreenlet::parent() const
+{
+    return OwnedGreenlet(); // null becomes None
+}
+
+}; // namespace greenlet
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TPythonState.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TPythonState.cpp
new file mode 100644
index 00000000..cc5dff51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TPythonState.cpp
@@ -0,0 +1,393 @@
+#ifndef GREENLET_PYTHON_STATE_CPP
+#define GREENLET_PYTHON_STATE_CPP
+
+#include <Python.h>
+#include "TGreenlet.hpp"
+
+namespace greenlet {
+
+PythonState::PythonState()
+    : _top_frame()
+#if GREENLET_USE_CFRAME
+    ,cframe(nullptr)
+    ,use_tracing(0)
+#endif
+#if GREENLET_PY312
+    ,py_recursion_depth(0)
+    ,c_recursion_depth(0)
+#else
+    ,recursion_depth(0)
+#endif
+#if GREENLET_PY313
+    ,delete_later(nullptr)
+#else
+    ,trash_delete_nesting(0)
+#endif
+#if GREENLET_PY311
+    ,current_frame(nullptr)
+    ,datastack_chunk(nullptr)
+    ,datastack_top(nullptr)
+    ,datastack_limit(nullptr)
+#endif
+{
+#if GREENLET_USE_CFRAME
+    /*
+      The PyThreadState->cframe pointer usually points to memory on
+      the stack, alloceted in a call into PyEval_EvalFrameDefault.
+
+      Initially, before any evaluation begins, it points to the
+      initial PyThreadState object's ``root_cframe`` object, which is
+      statically allocated for the lifetime of the thread.
+
+      A greenlet can last for longer than a call to
+      PyEval_EvalFrameDefault, so we can't set its ``cframe`` pointer
+      to be the current ``PyThreadState->cframe``; nor could we use
+      one from the greenlet parent for the same reason. Yet a further
+      no: we can't allocate one scoped to the greenlet and then
+      destroy it when the greenlet is deallocated, because inside the
+      interpreter the _PyCFrame objects form a linked list, and that too
+      can result in accessing memory beyond its dynamic lifetime (if
+      the greenlet doesn't actually finish before it dies, its entry
+      could still be in the list).
+
+      Using the ``root_cframe`` is problematic, though, because its
+      members are never modified by the interpreter and are set to 0,
+      meaning that its ``use_tracing`` flag is never updated. We don't
+      want to modify that value in the ``root_cframe`` ourself: it
+      *shouldn't* matter much because we should probably never get
+      back to the point where that's the only cframe on the stack;
+      even if it did matter, the major consequence of an incorrect
+      value for ``use_tracing`` is that if its true the interpreter
+      does some extra work --- however, it's just good code hygiene.
+
+      Our solution: before a greenlet runs, after its initial
+      creation, it uses the ``root_cframe`` just to have something to
+      put there. However, once the greenlet is actually switched to
+      for the first time, ``g_initialstub`` (which doesn't actually
+      "return" while the greenlet is running) stores a new _PyCFrame on
+      its local stack, and copies the appropriate values from the
+      currently running _PyCFrame; this is then made the _PyCFrame for the
+      newly-minted greenlet. ``g_initialstub`` then proceeds to call
+      ``glet.run()``, which results in ``PyEval_...`` adding the
+      _PyCFrame to the list. Switches continue as normal. Finally, when
+      the greenlet finishes, the call to ``glet.run()`` returns and
+      the _PyCFrame is taken out of the linked list and the stack value
+      is now unused and free to expire.
+
+      XXX: I think we can do better. If we're deallocing in the same
+      thread, can't we traverse the list and unlink our frame?
+      Can we just keep a reference to the thread state in case we
+      dealloc in another thread? (Is that even possible if we're still
+      running and haven't returned from g_initialstub?)
+    */
+    this->cframe = &PyThreadState_GET()->root_cframe;
+#endif
+}
+
+
+inline void PythonState::may_switch_away() noexcept
+{
+#if GREENLET_PY311
+    // PyThreadState_GetFrame is probably going to have to allocate a
+    // new frame object. That may trigger garbage collection. Because
+    // we call this during the early phases of a switch (it doesn't
+    // matter to which greenlet, as this has a global effect), if a GC
+    // triggers a switch away, two things can happen, both bad:
+    // - We might not get switched back to, halting forward progress.
+    //   this is pathological, but possible.
+    // - We might get switched back to with a different set of
+    //   arguments or a throw instead of a switch. That would corrupt
+    //   our state (specifically, PyErr_Occurred() and this->args()
+    //   would no longer agree).
+    //
+    // Thus, when we call this API, we need to have GC disabled.
+    // This method serves as a bottleneck we call when maybe beginning
+    // a switch. In this way, it is always safe -- no risk of GC -- to
+    // use ``_GetFrame()`` whenever we need to, just as it was in
+    // <=3.10 (because subsequent calls will be cached and not
+    // allocate memory).
+
+    GCDisabledGuard no_gc;
+    Py_XDECREF(PyThreadState_GetFrame(PyThreadState_GET()));
+#endif
+}
+
+void PythonState::operator<<(const PyThreadState *const tstate) noexcept
+{
+    this->_context.steal(tstate->context);
+#if GREENLET_USE_CFRAME
+    /*
+      IMPORTANT: ``cframe`` is a pointer into the STACK. Thus, because
+      the call to ``slp_switch()`` changes the contents of the stack,
+      you cannot read from ``ts_current->cframe`` after that call and
+      necessarily get the same values you get from reading it here.
+      Anything you need to restore from now to then must be saved in a
+      global/threadlocal variable (because we can't use stack
+      variables here either). For things that need to persist across
+      the switch, use `will_switch_from`.
+    */
+    this->cframe = tstate->cframe;
+  #if !GREENLET_PY312
+    this->use_tracing = tstate->cframe->use_tracing;
+  #endif
+#endif // GREENLET_USE_CFRAME
+#if GREENLET_PY311
+  #if GREENLET_PY312
+    this->py_recursion_depth = tstate->py_recursion_limit - tstate->py_recursion_remaining;
+    this->c_recursion_depth = Py_C_RECURSION_LIMIT - tstate->c_recursion_remaining;
+  #else // not 312
+    this->recursion_depth = tstate->recursion_limit - tstate->recursion_remaining;
+  #endif // GREENLET_PY312
+  #if GREENLET_PY313
+    this->current_frame = tstate->current_frame;
+  #elif GREENLET_USE_CFRAME
+    this->current_frame = tstate->cframe->current_frame;
+  #endif
+    this->datastack_chunk = tstate->datastack_chunk;
+    this->datastack_top = tstate->datastack_top;
+    this->datastack_limit = tstate->datastack_limit;
+
+    PyFrameObject *frame = PyThreadState_GetFrame((PyThreadState *)tstate);
+    Py_XDECREF(frame);  // PyThreadState_GetFrame gives us a new
+                        // reference.
+    this->_top_frame.steal(frame);
+  #if GREENLET_PY313
+    this->delete_later = Py_XNewRef(tstate->delete_later);
+  #elif GREENLET_PY312
+    this->trash_delete_nesting = tstate->trash.delete_nesting;
+  #else // not 312
+    this->trash_delete_nesting = tstate->trash_delete_nesting;
+  #endif // GREENLET_PY312
+#else // Not 311
+    this->recursion_depth = tstate->recursion_depth;
+    this->_top_frame.steal(tstate->frame);
+    this->trash_delete_nesting = tstate->trash_delete_nesting;
+#endif // GREENLET_PY311
+}
+
+#if GREENLET_PY312
+void GREENLET_NOINLINE(PythonState::unexpose_frames)()
+{
+    if (!this->top_frame()) {
+        return;
+    }
+
+    // See GreenletState::expose_frames() and the comment on frames_were_exposed
+    // for more information about this logic.
+    _PyInterpreterFrame *iframe = this->_top_frame->f_frame;
+    while (iframe != nullptr) {
+        _PyInterpreterFrame *prev_exposed = iframe->previous;
+        assert(iframe->frame_obj);
+        memcpy(&iframe->previous, &iframe->frame_obj->_f_frame_data[0],
+               sizeof(void *));
+        iframe = prev_exposed;
+    }
+}
+#else
+void PythonState::unexpose_frames()
+{}
+#endif
+
+void PythonState::operator>>(PyThreadState *const tstate) noexcept
+{
+    tstate->context = this->_context.relinquish_ownership();
+    /* Incrementing this value invalidates the contextvars cache,
+       which would otherwise remain valid across switches */
+    tstate->context_ver++;
+#if GREENLET_USE_CFRAME
+    tstate->cframe = this->cframe;
+    /*
+      If we were tracing, we need to keep tracing.
+      There should never be the possibility of hitting the
+      root_cframe here. See note above about why we can't
+      just copy this from ``origin->cframe->use_tracing``.
+    */
+  #if !GREENLET_PY312
+    tstate->cframe->use_tracing = this->use_tracing;
+  #endif
+#endif // GREENLET_USE_CFRAME
+#if GREENLET_PY311
+  #if GREENLET_PY312
+    tstate->py_recursion_remaining = tstate->py_recursion_limit - this->py_recursion_depth;
+    tstate->c_recursion_remaining = Py_C_RECURSION_LIMIT - this->c_recursion_depth;
+    this->unexpose_frames();
+  #else // \/ 3.11
+    tstate->recursion_remaining = tstate->recursion_limit - this->recursion_depth;
+  #endif // GREENLET_PY312
+  #if GREENLET_PY313
+    tstate->current_frame = this->current_frame;
+  #elif GREENLET_USE_CFRAME
+    tstate->cframe->current_frame = this->current_frame;
+  #endif
+    tstate->datastack_chunk = this->datastack_chunk;
+    tstate->datastack_top = this->datastack_top;
+    tstate->datastack_limit = this->datastack_limit;
+    this->_top_frame.relinquish_ownership();
+  #if GREENLET_PY313
+    Py_XDECREF(tstate->delete_later);
+    tstate->delete_later = this->delete_later;
+    Py_CLEAR(this->delete_later);
+  #elif GREENLET_PY312
+    tstate->trash.delete_nesting = this->trash_delete_nesting;
+  #else // not 3.12
+    tstate->trash_delete_nesting = this->trash_delete_nesting;
+  #endif // GREENLET_PY312
+#else // not 3.11
+    tstate->frame = this->_top_frame.relinquish_ownership();
+    tstate->recursion_depth = this->recursion_depth;
+    tstate->trash_delete_nesting = this->trash_delete_nesting;
+#endif // GREENLET_PY311
+}
+
+inline void PythonState::will_switch_from(PyThreadState *const origin_tstate) noexcept
+{
+#if GREENLET_USE_CFRAME && !GREENLET_PY312
+    // The weird thing is, we don't actually save this for an
+    // effect on the current greenlet, it's saved for an
+    // effect on the target greenlet. That is, we want
+    // continuity of this setting across the greenlet switch.
+    this->use_tracing = origin_tstate->cframe->use_tracing;
+#endif
+}
+
+void PythonState::set_initial_state(const PyThreadState* const tstate) noexcept
+{
+    this->_top_frame = nullptr;
+#if GREENLET_PY312
+    this->py_recursion_depth = tstate->py_recursion_limit - tstate->py_recursion_remaining;
+    // XXX: TODO: Comment from a reviewer:
+    //     Should this be ``Py_C_RECURSION_LIMIT - tstate->c_recursion_remaining``?
+    // But to me it looks more like that might not be the right
+    // initialization either?
+    this->c_recursion_depth = tstate->py_recursion_limit - tstate->py_recursion_remaining;
+#elif GREENLET_PY311
+    this->recursion_depth = tstate->recursion_limit - tstate->recursion_remaining;
+#else
+    this->recursion_depth = tstate->recursion_depth;
+#endif
+}
+// TODO: Better state management about when we own the top frame.
+int PythonState::tp_traverse(visitproc visit, void* arg, bool own_top_frame) noexcept
+{
+    Py_VISIT(this->_context.borrow());
+    if (own_top_frame) {
+        Py_VISIT(this->_top_frame.borrow());
+    }
+    return 0;
+}
+
+void PythonState::tp_clear(bool own_top_frame) noexcept
+{
+    PythonStateContext::tp_clear();
+    // If we get here owning a frame,
+    // we got dealloc'd without being finished. We may or may not be
+    // in the same thread.
+    if (own_top_frame) {
+        this->_top_frame.CLEAR();
+    }
+}
+
+#if GREENLET_USE_CFRAME
+void PythonState::set_new_cframe(_PyCFrame& frame) noexcept
+{
+    frame = *PyThreadState_GET()->cframe;
+    /* Make the target greenlet refer to the stack value. */
+    this->cframe = &frame;
+    /*
+      And restore the link to the previous frame so this one gets
+      unliked appropriately.
+    */
+    this->cframe->previous = &PyThreadState_GET()->root_cframe;
+}
+#endif
+
+const PythonState::OwnedFrame& PythonState::top_frame() const noexcept
+{
+    return this->_top_frame;
+}
+
+void PythonState::did_finish(PyThreadState* tstate) noexcept
+{
+#if GREENLET_PY311
+    // See https://github.com/gevent/gevent/issues/1924 and
+    // https://github.com/python-greenlet/greenlet/issues/328. In
+    // short, Python 3.11 allocates memory for frames as a sort of
+    // linked list that's kept as part of PyThreadState in the
+    // ``datastack_chunk`` member and friends. These are saved and
+    // restored as part of switching greenlets.
+    //
+    // When we initially switch to a greenlet, we set those to NULL.
+    // That causes the frame management code to treat this like a
+    // brand new thread and start a fresh list of chunks, beginning
+    // with a new "root" chunk. As we make calls in this greenlet,
+    // those chunks get added, and as calls return, they get popped.
+    // But the frame code (pystate.c) is careful to make sure that the
+    // root chunk never gets popped.
+    //
+    // Thus, when a greenlet exits for the last time, there will be at
+    // least a single root chunk that we must be responsible for
+    // deallocating.
+    //
+    // The complex part is that these chunks are allocated and freed
+    // using ``_PyObject_VirtualAlloc``/``Free``. Those aren't public
+    // functions, and they aren't exported for linking. It so happens
+    // that we know they are just thin wrappers around the Arena
+    // allocator, so we can use that directly to deallocate in a
+    // compatible way.
+    //
+    // CAUTION: Check this implementation detail on every major version.
+    //
+    // It might be nice to be able to do this in our destructor, but
+    // can we be sure that no one else is using that memory? Plus, as
+    // described below, our pointers may not even be valid anymore. As
+    // a special case, there is one time that we know we can do this,
+    // and that's from the destructor of the associated UserGreenlet
+    // (NOT main greenlet)
+    PyObjectArenaAllocator alloc;
+    _PyStackChunk* chunk = nullptr;
+    if (tstate) {
+        // We really did finish, we can never be switched to again.
+        chunk = tstate->datastack_chunk;
+        // Unfortunately, we can't do much sanity checking. Our
+        // this->datastack_chunk pointer is out of date (evaluation may
+        // have popped down through it already) so we can't verify that
+        // we deallocate it. I don't think we can even check datastack_top
+        // for the same reason.
+
+        PyObject_GetArenaAllocator(&alloc);
+        tstate->datastack_chunk = nullptr;
+        tstate->datastack_limit = nullptr;
+        tstate->datastack_top = nullptr;
+
+    }
+    else if (this->datastack_chunk) {
+        // The UserGreenlet (NOT the main greenlet!) is being deallocated. If we're
+        // still holding a stack chunk, it's garbage because we know
+        // we can never switch back to let cPython clean it up.
+        // Because the last time we got switched away from, and we
+        // haven't run since then, we know our chain is valid and can
+        // be dealloced.
+        chunk = this->datastack_chunk;
+        PyObject_GetArenaAllocator(&alloc);
+    }
+
+    if (alloc.free && chunk) {
+        // In case the arena mechanism has been torn down already.
+        while (chunk) {
+            _PyStackChunk *prev = chunk->previous;
+            chunk->previous = nullptr;
+            alloc.free(alloc.ctx, chunk, chunk->size);
+            chunk = prev;
+        }
+    }
+
+    this->datastack_chunk = nullptr;
+    this->datastack_limit = nullptr;
+    this->datastack_top = nullptr;
+#endif
+}
+
+
+}; // namespace greenlet
+
+#endif // GREENLET_PYTHON_STATE_CPP
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TStackState.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TStackState.cpp
new file mode 100644
index 00000000..9743ab51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TStackState.cpp
@@ -0,0 +1,265 @@
+#ifndef GREENLET_STACK_STATE_CPP
+#define GREENLET_STACK_STATE_CPP
+
+#include "TGreenlet.hpp"
+
+namespace greenlet {
+
+#ifdef GREENLET_USE_STDIO
+#include <iostream>
+using std::cerr;
+using std::endl;
+
+std::ostream& operator<<(std::ostream& os, const StackState& s)
+{
+    os << "StackState(stack_start=" << (void*)s._stack_start
+       << ", stack_stop=" << (void*)s.stack_stop
+       << ", stack_copy=" << (void*)s.stack_copy
+       << ", stack_saved=" << s._stack_saved
+       << ", stack_prev=" << s.stack_prev
+       << ", addr=" << &s
+       << ")";
+    return os;
+}
+#endif
+
+StackState::StackState(void* mark, StackState& current)
+    : _stack_start(nullptr),
+      stack_stop((char*)mark),
+      stack_copy(nullptr),
+      _stack_saved(0),
+      /* Skip a dying greenlet */
+      stack_prev(current._stack_start
+                 ? &current
+                 : current.stack_prev)
+{
+}
+
+StackState::StackState()
+    : _stack_start(nullptr),
+      stack_stop(nullptr),
+      stack_copy(nullptr),
+      _stack_saved(0),
+      stack_prev(nullptr)
+{
+}
+
+StackState::StackState(const StackState& other)
+// can't use a delegating constructor because of
+// MSVC for Python 2.7
+    : _stack_start(nullptr),
+      stack_stop(nullptr),
+      stack_copy(nullptr),
+      _stack_saved(0),
+      stack_prev(nullptr)
+{
+    this->operator=(other);
+}
+
+StackState& StackState::operator=(const StackState& other)
+{
+    if (&other == this) {
+        return *this;
+    }
+    if (other._stack_saved) {
+        throw std::runtime_error("Refusing to steal memory.");
+    }
+
+    //If we have memory allocated, dispose of it
+    this->free_stack_copy();
+
+    this->_stack_start = other._stack_start;
+    this->stack_stop = other.stack_stop;
+    this->stack_copy = other.stack_copy;
+    this->_stack_saved = other._stack_saved;
+    this->stack_prev = other.stack_prev;
+    return *this;
+}
+
+inline void StackState::free_stack_copy() noexcept
+{
+    PyMem_Free(this->stack_copy);
+    this->stack_copy = nullptr;
+    this->_stack_saved = 0;
+}
+
+inline void StackState::copy_heap_to_stack(const StackState& current) noexcept
+{
+
+    /* Restore the heap copy back into the C stack */
+    if (this->_stack_saved != 0) {
+        memcpy(this->_stack_start, this->stack_copy, this->_stack_saved);
+        this->free_stack_copy();
+    }
+    StackState* owner = const_cast<StackState*>(&current);
+    if (!owner->_stack_start) {
+        owner = owner->stack_prev; /* greenlet is dying, skip it */
+    }
+    while (owner && owner->stack_stop <= this->stack_stop) {
+        // cerr << "\tOwner: " << owner << endl;
+        owner = owner->stack_prev; /* find greenlet with more stack */
+    }
+    this->stack_prev = owner;
+    // cerr << "\tFinished with: " << *this << endl;
+}
+
+inline int StackState::copy_stack_to_heap_up_to(const char* const stop) noexcept
+{
+    /* Save more of g's stack into the heap -- at least up to 'stop'
+       g->stack_stop |________|
+                     |        |
+                     |    __ stop       . . . . .
+                     |        |    ==>  .       .
+                     |________|          _______
+                     |        |         |       |
+                     |        |         |       |
+      g->stack_start |        |         |_______| g->stack_copy
+     */
+    intptr_t sz1 = this->_stack_saved;
+    intptr_t sz2 = stop - this->_stack_start;
+    assert(this->_stack_start);
+    if (sz2 > sz1) {
+        char* c = (char*)PyMem_Realloc(this->stack_copy, sz2);
+        if (!c) {
+            PyErr_NoMemory();
+            return -1;
+        }
+        memcpy(c + sz1, this->_stack_start + sz1, sz2 - sz1);
+        this->stack_copy = c;
+        this->_stack_saved = sz2;
+    }
+    return 0;
+}
+
+inline int StackState::copy_stack_to_heap(char* const stackref,
+                                          const StackState& current) noexcept
+{
+    /* must free all the C stack up to target_stop */
+    const char* const target_stop = this->stack_stop;
+
+    StackState* owner = const_cast<StackState*>(&current);
+    assert(owner->_stack_saved == 0); // everything is present on the stack
+    if (!owner->_stack_start) {
+        owner = owner->stack_prev; /* not saved if dying */
+    }
+    else {
+        owner->_stack_start = stackref;
+    }
+
+    while (owner->stack_stop < target_stop) {
+        /* ts_current is entierely within the area to free */
+        if (owner->copy_stack_to_heap_up_to(owner->stack_stop)) {
+            return -1; /* XXX */
+        }
+        owner = owner->stack_prev;
+    }
+    if (owner != this) {
+        if (owner->copy_stack_to_heap_up_to(target_stop)) {
+            return -1; /* XXX */
+        }
+    }
+    return 0;
+}
+
+inline bool StackState::started() const noexcept
+{
+    return this->stack_stop != nullptr;
+}
+
+inline bool StackState::main() const noexcept
+{
+    return this->stack_stop == (char*)-1;
+}
+
+inline bool StackState::active() const noexcept
+{
+    return this->_stack_start != nullptr;
+}
+
+inline void StackState::set_active() noexcept
+{
+    assert(this->_stack_start == nullptr);
+    this->_stack_start = (char*)1;
+}
+
+inline void StackState::set_inactive() noexcept
+{
+    this->_stack_start = nullptr;
+    // XXX: What if we still have memory out there?
+    // That case is actually triggered by
+    // test_issue251_issue252_explicit_reference_not_collectable (greenlet.tests.test_leaks.TestLeaks)
+    // and
+    // test_issue251_issue252_need_to_collect_in_background
+    // (greenlet.tests.test_leaks.TestLeaks)
+    //
+    // Those objects never get deallocated, so the destructor never
+    // runs.
+    // It *seems* safe to clean up the memory here?
+    if (this->_stack_saved) {
+        this->free_stack_copy();
+    }
+}
+
+inline intptr_t StackState::stack_saved() const noexcept
+{
+    return this->_stack_saved;
+}
+
+inline char* StackState::stack_start() const noexcept
+{
+    return this->_stack_start;
+}
+
+
+inline StackState StackState::make_main() noexcept
+{
+    StackState s;
+    s._stack_start = (char*)1;
+    s.stack_stop = (char*)-1;
+    return s;
+}
+
+StackState::~StackState()
+{
+    if (this->_stack_saved != 0) {
+        this->free_stack_copy();
+    }
+}
+
+void StackState::copy_from_stack(void* vdest, const void* vsrc, size_t n) const
+{
+    char* dest = static_cast<char*>(vdest);
+    const char* src = static_cast<const char*>(vsrc);
+    if (src + n <= this->_stack_start
+        || src >= this->_stack_start + this->_stack_saved
+        || this->_stack_saved == 0) {
+        // Nothing we're copying was spilled from the stack
+        memcpy(dest, src, n);
+        return;
+    }
+
+    if (src < this->_stack_start) {
+        // Copy the part before the saved stack.
+        // We know src + n > _stack_start due to the test above.
+        const size_t nbefore = this->_stack_start - src;
+        memcpy(dest, src, nbefore);
+        dest += nbefore;
+        src += nbefore;
+        n -= nbefore;
+    }
+    // We know src >= _stack_start after the before-copy, and
+    // src < _stack_start + _stack_saved due to the first if condition
+    size_t nspilled = std::min<size_t>(n, this->_stack_start + this->_stack_saved - src);
+    memcpy(dest, this->stack_copy + (src - this->_stack_start), nspilled);
+    dest += nspilled;
+    src += nspilled;
+    n -= nspilled;
+    if (n > 0) {
+        // Copy the part after the saved stack
+        memcpy(dest, src, n);
+    }
+}
+
+}; // namespace greenlet
+
+#endif // GREENLET_STACK_STATE_CPP
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TThreadState.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadState.hpp
new file mode 100644
index 00000000..e4e6f6cb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadState.hpp
@@ -0,0 +1,497 @@
+#ifndef GREENLET_THREAD_STATE_HPP
+#define GREENLET_THREAD_STATE_HPP
+
+#include <ctime>
+#include <stdexcept>
+
+#include "greenlet_internal.hpp"
+#include "greenlet_refs.hpp"
+#include "greenlet_thread_support.hpp"
+
+using greenlet::refs::BorrowedObject;
+using greenlet::refs::BorrowedGreenlet;
+using greenlet::refs::BorrowedMainGreenlet;
+using greenlet::refs::OwnedMainGreenlet;
+using greenlet::refs::OwnedObject;
+using greenlet::refs::OwnedGreenlet;
+using greenlet::refs::OwnedList;
+using greenlet::refs::PyErrFetchParam;
+using greenlet::refs::PyArgParseParam;
+using greenlet::refs::ImmortalString;
+using greenlet::refs::CreatedModule;
+using greenlet::refs::PyErrPieces;
+using greenlet::refs::NewReference;
+
+namespace greenlet {
+/**
+ * Thread-local state of greenlets.
+ *
+ * Each native thread will get exactly one of these objects,
+ * automatically accessed through the best available thread-local
+ * mechanism the compiler supports (``thread_local`` for C++11
+ * compilers or ``__thread``/``declspec(thread)`` for older GCC/clang
+ * or MSVC, respectively.)
+ *
+ * Previously, we kept thread-local state mostly in a bunch of
+ * ``static volatile`` variables in the main greenlet file.. This had
+ * the problem of requiring extra checks, loops, and great care
+ * accessing these variables if we potentially invoked any Python code
+ * that could release the GIL, because the state could change out from
+ * under us. Making the variables thread-local solves this problem.
+ *
+ * When we detected that a greenlet API accessing the current greenlet
+ * was invoked from a different thread than the greenlet belonged to,
+ * we stored a reference to the greenlet in the Python thread
+ * dictionary for the thread the greenlet belonged to. This could lead
+ * to memory leaks if the thread then exited (because of a reference
+ * cycle, as greenlets referred to the thread dictionary, and deleting
+ * non-current greenlets leaked their frame plus perhaps arguments on
+ * the C stack). If a thread exited while still having running
+ * greenlet objects (perhaps that had just switched back to the main
+ * greenlet), and did not invoke one of the greenlet APIs *in that
+ * thread, immediately before it exited, without some other thread
+ * then being invoked*, such a leak was guaranteed.
+ *
+ * This can be partly solved by using compiler thread-local variables
+ * instead of the Python thread dictionary, thus avoiding a cycle.
+ *
+ * To fully solve this problem, we need a reliable way to know that a
+ * thread is done and we should clean up the main greenlet. On POSIX,
+ * we can use the destructor function of ``pthread_key_create``, but
+ * there's nothing similar on Windows; a C++11 thread local object
+ * reliably invokes its destructor when the thread it belongs to exits
+ * (non-C++11 compilers offer ``__thread`` or ``declspec(thread)`` to
+ * create thread-local variables, but they can't hold C++ objects that
+ * invoke destructors; the C++11 version is the most portable solution
+ * I found). When the thread exits, we can drop references and
+ * otherwise manipulate greenlets and frames that we know can no
+ * longer be switched to. For compilers that don't support C++11
+ * thread locals, we have a solution that uses the python thread
+ * dictionary, though it may not collect everything as promptly as
+ * other compilers do, if some other library is using the thread
+ * dictionary and has a cycle or extra reference.
+ *
+ * There are two small wrinkles. The first is that when the thread
+ * exits, it is too late to actually invoke Python APIs: the Python
+ * thread state is gone, and the GIL is released. To solve *this*
+ * problem, our destructor uses ``Py_AddPendingCall`` to transfer the
+ * destruction work to the main thread. (This is not an issue for the
+ * dictionary solution.)
+ *
+ * The second is that once the thread exits, the thread local object
+ * is invalid and we can't even access a pointer to it, so we can't
+ * pass it to ``Py_AddPendingCall``. This is handled by actually using
+ * a second object that's thread local (ThreadStateCreator) and having
+ * it dynamically allocate this object so it can live until the
+ * pending call runs.
+ */
+
+
+
+class ThreadState {
+private:
+    // As of commit 08ad1dd7012b101db953f492e0021fb08634afad
+    // this class needed 56 bytes in o Py_DEBUG build
+    // on 64-bit macOS 11.
+    // Adding the vector takes us up to 80 bytes ()
+
+    /* Strong reference to the main greenlet */
+    OwnedMainGreenlet main_greenlet;
+
+    /* Strong reference to the current greenlet. */
+    OwnedGreenlet current_greenlet;
+
+    /* Strong reference to the trace function, if any. */
+    OwnedObject tracefunc;
+
+    typedef std::vector<PyGreenlet*, PythonAllocator<PyGreenlet*> > deleteme_t;
+    /* A vector of raw PyGreenlet pointers representing things that need
+       deleted when this thread is running. The vector owns the
+       references, but you need to manually INCREF/DECREF as you use
+       them. We don't use a vector<refs::OwnedGreenlet> because we
+       make copy of this vector, and that would become O(n) as all the
+       refcounts are incremented in the copy.
+    */
+    deleteme_t deleteme;
+
+#ifdef GREENLET_NEEDS_EXCEPTION_STATE_SAVED
+    void* exception_state;
+#endif
+
+    static std::clock_t _clocks_used_doing_gc;
+    static ImmortalString get_referrers_name;
+    static PythonAllocator<ThreadState> allocator;
+
+    G_NO_COPIES_OF_CLS(ThreadState);
+
+
+    // Allocates a main greenlet for the thread state. If this fails,
+    // exits the process. Called only during constructing a ThreadState.
+    MainGreenlet* alloc_main()
+    {
+        PyGreenlet* gmain;
+
+        /* create the main greenlet for this thread */
+        gmain = reinterpret_cast<PyGreenlet*>(PyType_GenericAlloc(&PyGreenlet_Type, 0));
+        if (gmain == NULL) {
+            throw PyFatalError("alloc_main failed to alloc"); //exits the process
+        }
+
+        MainGreenlet* const main = new MainGreenlet(gmain, this);
+
+        assert(Py_REFCNT(gmain) == 1);
+        assert(gmain->pimpl == main);
+        return main;
+    }
+
+
+public:
+    static void* operator new(size_t UNUSED(count))
+    {
+        return ThreadState::allocator.allocate(1);
+    }
+
+    static void operator delete(void* ptr)
+    {
+        return ThreadState::allocator.deallocate(static_cast<ThreadState*>(ptr),
+                                                 1);
+    }
+
+    static void init()
+    {
+        ThreadState::get_referrers_name = "get_referrers";
+        ThreadState::_clocks_used_doing_gc = 0;
+    }
+
+    ThreadState()
+    {
+
+#ifdef GREENLET_NEEDS_EXCEPTION_STATE_SAVED
+        this->exception_state = slp_get_exception_state();
+#endif
+
+        // XXX: Potentially dangerous, exposing a not fully
+        // constructed object.
+        MainGreenlet* const main = this->alloc_main();
+        this->main_greenlet = OwnedMainGreenlet::consuming(
+            main->self()
+        );
+        assert(this->main_greenlet);
+        this->current_greenlet = main->self();
+        // The main greenlet starts with 1 refs: The returned one. We
+        // then copied it to the current greenlet.
+        assert(this->main_greenlet.REFCNT() == 2);
+    }
+
+    inline void restore_exception_state()
+    {
+#ifdef GREENLET_NEEDS_EXCEPTION_STATE_SAVED
+        // It's probably important this be inlined and only call C
+        // functions to avoid adding an SEH frame.
+        slp_set_exception_state(this->exception_state);
+#endif
+    }
+
+    inline bool has_main_greenlet() const noexcept
+    {
+        return bool(this->main_greenlet);
+    }
+
+    // Called from the ThreadStateCreator when we're in non-standard
+    // threading mode. In that case, there is an object in the Python
+    // thread state dictionary that points to us. The main greenlet
+    // also traverses into us, in which case it's crucial not to
+    // traverse back into the main greenlet.
+    int tp_traverse(visitproc visit, void* arg, bool traverse_main=true)
+    {
+        if (traverse_main) {
+            Py_VISIT(main_greenlet.borrow_o());
+        }
+        if (traverse_main || current_greenlet != main_greenlet) {
+            Py_VISIT(current_greenlet.borrow_o());
+        }
+        Py_VISIT(tracefunc.borrow());
+        return 0;
+    }
+
+    inline BorrowedMainGreenlet borrow_main_greenlet() const noexcept
+    {
+        assert(this->main_greenlet);
+        assert(this->main_greenlet.REFCNT() >= 2);
+        return this->main_greenlet;
+    };
+
+    inline OwnedMainGreenlet get_main_greenlet() const noexcept
+    {
+        return this->main_greenlet;
+    }
+
+    /**
+     * In addition to returning a new reference to the currunt
+     * greenlet, this performs any maintenance needed.
+     */
+    inline OwnedGreenlet get_current()
+    {
+        /* green_dealloc() cannot delete greenlets from other threads, so
+           it stores them in the thread dict; delete them now. */
+        this->clear_deleteme_list();
+        //assert(this->current_greenlet->main_greenlet == this->main_greenlet);
+        //assert(this->main_greenlet->main_greenlet == this->main_greenlet);
+        return this->current_greenlet;
+    }
+
+    /**
+     * As for non-const get_current();
+     */
+    inline BorrowedGreenlet borrow_current()
+    {
+        this->clear_deleteme_list();
+        return this->current_greenlet;
+    }
+
+    /**
+     * Does no maintenance.
+     */
+    inline OwnedGreenlet get_current() const
+    {
+        return this->current_greenlet;
+    }
+
+    template<typename T, refs::TypeChecker TC>
+    inline bool is_current(const refs::PyObjectPointer<T, TC>& obj) const
+    {
+        return this->current_greenlet.borrow_o() == obj.borrow_o();
+    }
+
+    inline void set_current(const OwnedGreenlet& target)
+    {
+        this->current_greenlet = target;
+    }
+
+private:
+    /**
+     * Deref and remove the greenlets from the deleteme list. Must be
+     * holding the GIL.
+     *
+     * If *murder* is true, then we must be called from a different
+     * thread than the one that these greenlets were running in.
+     * In that case, if the greenlet was actually running, we destroy
+     * the frame reference and otherwise make it appear dead before
+     * proceeding; otherwise, we would try (and fail) to raise an
+     * exception in it and wind up right back in this list.
+     */
+    inline void clear_deleteme_list(const bool murder=false)
+    {
+        if (!this->deleteme.empty()) {
+            // It's possible we could add items to this list while
+            // running Python code if there's a thread switch, so we
+            // need to defensively copy it before that can happen.
+            deleteme_t copy = this->deleteme;
+            this->deleteme.clear(); // in case things come back on the list
+            for(deleteme_t::iterator it = copy.begin(), end = copy.end();
+                it != end;
+                ++it ) {
+                PyGreenlet* to_del = *it;
+                if (murder) {
+                    // Force each greenlet to appear dead; we can't raise an
+                    // exception into it anymore anyway.
+                    to_del->pimpl->murder_in_place();
+                }
+
+                // The only reference to these greenlets should be in
+                // this list, decreffing them should let them be
+                // deleted again, triggering calls to green_dealloc()
+                // in the correct thread (if we're not murdering).
+                // This may run arbitrary Python code and switch
+                // threads or greenlets!
+                Py_DECREF(to_del);
+                if (PyErr_Occurred()) {
+                    PyErr_WriteUnraisable(nullptr);
+                    PyErr_Clear();
+                }
+            }
+        }
+    }
+
+public:
+
+    /**
+     * Returns a new reference, or a false object.
+     */
+    inline OwnedObject get_tracefunc() const
+    {
+        return tracefunc;
+    };
+
+
+    inline void set_tracefunc(BorrowedObject tracefunc)
+    {
+        assert(tracefunc);
+        if (tracefunc == BorrowedObject(Py_None)) {
+            this->tracefunc.CLEAR();
+        }
+        else {
+            this->tracefunc = tracefunc;
+        }
+    }
+
+    /**
+     * Given a reference to a greenlet that some other thread
+     * attempted to delete (has a refcount of 0) store it for later
+     * deletion when the thread this state belongs to is current.
+     */
+    inline void delete_when_thread_running(PyGreenlet* to_del)
+    {
+        Py_INCREF(to_del);
+        this->deleteme.push_back(to_del);
+    }
+
+    /**
+     * Set to std::clock_t(-1) to disable.
+     */
+    inline static std::clock_t& clocks_used_doing_gc()
+    {
+        return ThreadState::_clocks_used_doing_gc;
+    }
+
+    ~ThreadState()
+    {
+        if (!PyInterpreterState_Head()) {
+            // We shouldn't get here (our callers protect us)
+            // but if we do, all we can do is bail early.
+            return;
+        }
+
+        // We should not have an "origin" greenlet; that only exists
+        // for the temporary time during a switch, which should not
+        // be in progress as the thread dies.
+        //assert(!this->switching_state.origin);
+
+        this->tracefunc.CLEAR();
+
+        // Forcibly GC as much as we can.
+        this->clear_deleteme_list(true);
+
+        // The pending call did this.
+        assert(this->main_greenlet->thread_state() == nullptr);
+
+        // If the main greenlet is the current greenlet,
+        // then we "fell off the end" and the thread died.
+        // It's possible that there is some other greenlet that
+        // switched to us, leaving a reference to the main greenlet
+        // on the stack, somewhere uncollectible. Try to detect that.
+        if (this->current_greenlet == this->main_greenlet && this->current_greenlet) {
+            assert(this->current_greenlet->is_currently_running_in_some_thread());
+            // Drop one reference we hold.
+            this->current_greenlet.CLEAR();
+            assert(!this->current_greenlet);
+            // Only our reference to the main greenlet should be left,
+            // But hold onto the pointer in case we need to do extra cleanup.
+            PyGreenlet* old_main_greenlet = this->main_greenlet.borrow();
+            Py_ssize_t cnt = this->main_greenlet.REFCNT();
+            this->main_greenlet.CLEAR();
+            if (ThreadState::_clocks_used_doing_gc != std::clock_t(-1)
+                && cnt == 2 && Py_REFCNT(old_main_greenlet) == 1) {
+                // Highly likely that the reference is somewhere on
+                // the stack, not reachable by GC. Verify.
+                // XXX: This is O(n) in the total number of objects.
+                // TODO: Add a way to disable this at runtime, and
+                // another way to report on it.
+                std::clock_t begin = std::clock();
+                NewReference gc(PyImport_ImportModule("gc"));
+                if (gc) {
+                    OwnedObject get_referrers = gc.PyRequireAttr(ThreadState::get_referrers_name);
+                    OwnedList refs(get_referrers.PyCall(old_main_greenlet));
+                    if (refs && refs.empty()) {
+                        assert(refs.REFCNT() == 1);
+                        // We found nothing! So we left a dangling
+                        // reference: Probably the last thing some
+                        // other greenlet did was call
+                        // 'getcurrent().parent.switch()' to switch
+                        // back to us. Clean it up. This will be the
+                        // case on CPython 3.7 and newer, as they use
+                        // an internal calling conversion that avoids
+                        // creating method objects and storing them on
+                        // the stack.
+                        Py_DECREF(old_main_greenlet);
+                    }
+                    else if (refs
+                             && refs.size() == 1
+                             && PyCFunction_Check(refs.at(0))
+                             && Py_REFCNT(refs.at(0)) == 2) {
+                        assert(refs.REFCNT() == 1);
+                        // Ok, we found a C method that refers to the
+                        // main greenlet, and its only referenced
+                        // twice, once in the list we just created,
+                        // once from...somewhere else. If we can't
+                        // find where else, then this is a leak.
+                        // This happens in older versions of CPython
+                        // that create a bound method object somewhere
+                        // on the stack that we'll never get back to.
+                        if (PyCFunction_GetFunction(refs.at(0).borrow()) == (PyCFunction)green_switch) {
+                            BorrowedObject function_w = refs.at(0);
+                            refs.clear(); // destroy the reference
+                                          // from the list.
+                            // back to one reference. Can *it* be
+                            // found?
+                            assert(function_w.REFCNT() == 1);
+                            refs = get_referrers.PyCall(function_w);
+                            if (refs && refs.empty()) {
+                                // Nope, it can't be found so it won't
+                                // ever be GC'd. Drop it.
+                                Py_CLEAR(function_w);
+                            }
+                        }
+                    }
+                    std::clock_t end = std::clock();
+                    ThreadState::_clocks_used_doing_gc += (end - begin);
+                }
+            }
+        }
+
+        // We need to make sure this greenlet appears to be dead,
+        // because otherwise deallocing it would fail to raise an
+        // exception in it (the thread is dead) and put it back in our
+        // deleteme list.
+        if (this->current_greenlet) {
+            this->current_greenlet->murder_in_place();
+            this->current_greenlet.CLEAR();
+        }
+
+        if (this->main_greenlet) {
+            // Couldn't have been the main greenlet that was running
+            // when the thread exited (because we already cleared this
+            // pointer if it was). This shouldn't be possible?
+
+            // If the main greenlet was current when the thread died (it
+            // should be, right?) then we cleared its self pointer above
+            // when we cleared the current greenlet's main greenlet pointer.
+            // assert(this->main_greenlet->main_greenlet == this->main_greenlet
+            //        || !this->main_greenlet->main_greenlet);
+            // // self reference, probably gone
+            // this->main_greenlet->main_greenlet.CLEAR();
+
+            // This will actually go away when the ivar is destructed.
+            this->main_greenlet.CLEAR();
+        }
+
+        if (PyErr_Occurred()) {
+            PyErr_WriteUnraisable(NULL);
+            PyErr_Clear();
+        }
+
+    }
+
+};
+
+ImmortalString ThreadState::get_referrers_name(nullptr);
+PythonAllocator<ThreadState> ThreadState::allocator;
+std::clock_t ThreadState::_clocks_used_doing_gc(0);
+
+
+
+
+
+}; // namespace greenlet
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateCreator.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateCreator.hpp
new file mode 100644
index 00000000..2ec7ab55
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateCreator.hpp
@@ -0,0 +1,102 @@
+#ifndef GREENLET_THREAD_STATE_CREATOR_HPP
+#define GREENLET_THREAD_STATE_CREATOR_HPP
+
+#include <ctime>
+#include <stdexcept>
+
+#include "greenlet_internal.hpp"
+#include "greenlet_refs.hpp"
+#include "greenlet_thread_support.hpp"
+
+#include "TThreadState.hpp"
+
+namespace greenlet {
+
+
+typedef void (*ThreadStateDestructor)(ThreadState* const);
+
+template<ThreadStateDestructor Destructor>
+class ThreadStateCreator
+{
+private:
+    // Initialized to 1, and, if still 1, created on access.
+    // Set to 0 on destruction.
+    ThreadState* _state;
+    G_NO_COPIES_OF_CLS(ThreadStateCreator);
+
+    inline bool has_initialized_state() const noexcept
+    {
+        return this->_state != (ThreadState*)1;
+    }
+
+    inline bool has_state() const noexcept
+    {
+        return this->has_initialized_state() && this->_state != nullptr;
+    }
+
+public:
+
+    // Only one of these, auto created per thread.
+    // Constructing the state constructs the MainGreenlet.
+    ThreadStateCreator() :
+        _state((ThreadState*)1)
+    {
+    }
+
+    ~ThreadStateCreator()
+    {
+        if (this->has_state()) {
+            Destructor(this->_state);
+        }
+
+        this->_state = nullptr;
+    }
+
+    inline ThreadState& state()
+    {
+        // The main greenlet will own this pointer when it is created,
+        // which will be right after this. The plan is to give every
+        // greenlet a pointer to the main greenlet for the thread it
+        // runs in; if we are doing something cross-thread, we need to
+        // access the pointer from the main greenlet. Deleting the
+        // thread, and hence the thread-local storage, will delete the
+        // state pointer in the main greenlet.
+        if (!this->has_initialized_state()) {
+            // XXX: Assuming allocation never fails
+            this->_state = new ThreadState;
+            // For non-standard threading, we need to store an object
+            // in the Python thread state dictionary so that it can be
+            // DECREF'd when the thread ends (ideally; the dict could
+            // last longer) and clean this object up.
+        }
+        if (!this->_state) {
+            throw std::runtime_error("Accessing state after destruction.");
+        }
+        return *this->_state;
+    }
+
+    operator ThreadState&()
+    {
+        return this->state();
+    }
+
+    operator ThreadState*()
+    {
+        return &this->state();
+    }
+
+    inline int tp_traverse(visitproc visit, void* arg)
+    {
+        if (this->has_state()) {
+            return this->_state->tp_traverse(visit, arg);
+        }
+        return 0;
+    }
+
+};
+
+
+
+}; // namespace greenlet
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateDestroy.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateDestroy.cpp
new file mode 100644
index 00000000..37fcc8c1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TThreadStateDestroy.cpp
@@ -0,0 +1,258 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of the ThreadState destructors.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#ifndef T_THREADSTATE_DESTROY
+#define T_THREADSTATE_DESTROY
+
+#include "TGreenlet.hpp"
+
+#include "greenlet_thread_support.hpp"
+#include "greenlet_cpython_add_pending.hpp"
+#include "greenlet_compiler_compat.hpp"
+#include "TGreenletGlobals.cpp"
+#include "TThreadState.hpp"
+#include "TThreadStateCreator.hpp"
+
+namespace greenlet {
+
+extern "C" {
+
+struct ThreadState_DestroyNoGIL
+{
+    /**
+       This function uses the same lock that the PendingCallback does
+     */
+    static void
+    MarkGreenletDeadAndQueueCleanup(ThreadState* const state)
+    {
+#if GREENLET_BROKEN_THREAD_LOCAL_CLEANUP_JUST_LEAK
+        return;
+#endif
+        // We are *NOT* holding the GIL. Our thread is in the middle
+        // of its death throes and the Python thread state is already
+        // gone so we can't use most Python APIs. One that is safe is
+        // ``Py_AddPendingCall``, unless the interpreter itself has
+        // been torn down. There is a limited number of calls that can
+        // be queued: 32 (NPENDINGCALLS) in CPython 3.10, so we
+        // coalesce these calls using our own queue.
+
+        if (!MarkGreenletDeadIfNeeded(state)) {
+            // No state, or no greenlet
+            return;
+        }
+
+        // XXX: Because we don't have the GIL, this is a race condition.
+        if (!PyInterpreterState_Head()) {
+            // We have to leak the thread state, if the
+            // interpreter has shut down when we're getting
+            // deallocated, we can't run the cleanup code that
+            // deleting it would imply.
+            return;
+        }
+
+        AddToCleanupQueue(state);
+
+    }
+
+private:
+
+    // If the state has an allocated main greenlet:
+    // - mark the greenlet as dead by disassociating it from the state;
+    // - return 1
+    // Otherwise, return 0.
+    static bool
+    MarkGreenletDeadIfNeeded(ThreadState* const state)
+    {
+        if (state && state->has_main_greenlet()) {
+            // mark the thread as dead ASAP.
+            // this is racy! If we try to throw or switch to a
+            // greenlet from this thread from some other thread before
+            // we clear the state pointer, it won't realize the state
+            // is dead which can crash the process.
+            PyGreenlet* p(state->borrow_main_greenlet().borrow());
+            assert(p->pimpl->thread_state() == state || p->pimpl->thread_state() == nullptr);
+            dynamic_cast<MainGreenlet*>(p->pimpl)->thread_state(nullptr);
+           return true;
+        }
+        return false;
+    }
+
+    static void
+    AddToCleanupQueue(ThreadState* const state)
+    {
+        assert(state && state->has_main_greenlet());
+
+        // NOTE: Because we're not holding the GIL here, some other
+        // Python thread could run and call ``os.fork()``, which would
+        // be bad if that happened while we are holding the cleanup
+        // lock (it wouldn't function in the child process).
+        // Make a best effort to try to keep the duration we hold the
+        // lock short.
+        // TODO: On platforms that support it, use ``pthread_atfork`` to
+        // drop this lock.
+        LockGuard cleanup_lock(*mod_globs->thread_states_to_destroy_lock);
+
+        mod_globs->queue_to_destroy(state);
+        if (mod_globs->thread_states_to_destroy.size() == 1) {
+            // We added the first item to the queue. We need to schedule
+            // the cleanup.
+
+            // A size greater than 1 means that we have already added the pending call,
+            // and in fact, it may be executing now.
+            // If it is executing, our lock makes sure that it will see the item we just added
+            // to the queue on its next iteration (after we release the lock)
+            //
+            // A size of 1 means there is no pending call, OR the pending call is
+            // currently executing, has dropped the lock, and is deleting the last item
+            // from the queue; its next iteration will go ahead and delete the item we just added.
+            // And the pending call we schedule here will have no work to do.
+            int result = AddPendingCall(
+                           PendingCallback_DestroyQueueWithGIL,
+                            nullptr);
+            if (result < 0) {
+                // Hmm, what can we do here?
+                fprintf(stderr,
+                        "greenlet: WARNING: failed in call to Py_AddPendingCall; "
+                        "expect a memory leak.\n");
+            }
+        }
+    }
+
+    static int
+    PendingCallback_DestroyQueueWithGIL(void* UNUSED(arg))
+    {
+        // We're holding the GIL here, so no Python code should be able to
+        // run to call ``os.fork()``.
+        while (1) {
+            ThreadState* to_destroy;
+            {
+                LockGuard cleanup_lock(*mod_globs->thread_states_to_destroy_lock);
+                if (mod_globs->thread_states_to_destroy.empty()) {
+                    break;
+                }
+                to_destroy = mod_globs->take_next_to_destroy();
+            }
+            assert(to_destroy);
+            assert(to_destroy->has_main_greenlet());
+            // Drop the lock while we do the actual deletion.
+            // This allows other calls to MarkGreenletDeadAndQueueCleanup
+            // to enter and add to our queue.
+            DestroyOneWithGIL(to_destroy);
+        }
+        return 0;
+    }
+
+    static void
+    DestroyOneWithGIL(const ThreadState* const state)
+    {
+        // Holding the GIL.
+        // Passed a non-shared pointer to the actual thread state.
+        // state -> main greenlet
+        assert(state->has_main_greenlet());
+        PyGreenlet* main(state->borrow_main_greenlet());
+        // When we need to do cross-thread operations, we check this.
+        // A NULL value means the thread died some time ago.
+        // We do this here, rather than in a Python dealloc function
+        // for the greenlet, in case there's still a reference out
+        // there.
+        dynamic_cast<MainGreenlet*>(main->pimpl)->thread_state(nullptr);
+
+        delete state; // Deleting this runs the destructor, DECREFs the main greenlet.
+    }
+
+    // ensure this is actually defined.
+    static_assert(GREENLET_BROKEN_PY_ADD_PENDING == 1 || GREENLET_BROKEN_PY_ADD_PENDING == 0,
+                  "GREENLET_BROKEN_PY_ADD_PENDING not defined correctly.");
+
+#if GREENLET_BROKEN_PY_ADD_PENDING
+    static int _push_pending_call(struct _pending_calls *pending,
+                                  int (*func)(void *), void *arg)
+    {
+        int i = pending->last;
+        int j = (i + 1) % NPENDINGCALLS;
+        if (j == pending->first) {
+            return -1; /* Queue full */
+        }
+        pending->calls[i].func = func;
+        pending->calls[i].arg = arg;
+        pending->last = j;
+        return 0;
+    }
+
+    static int AddPendingCall(int (*func)(void *), void *arg)
+    {
+        _PyRuntimeState *runtime = &_PyRuntime;
+        if (!runtime) {
+            // obviously impossible
+            return 0;
+        }
+        struct _pending_calls *pending = &runtime->ceval.pending;
+        if (!pending->lock) {
+            return 0;
+        }
+        int result = 0;
+        PyThread_acquire_lock(pending->lock, WAIT_LOCK);
+        if (!pending->finishing) {
+            result = _push_pending_call(pending, func, arg);
+        }
+        PyThread_release_lock(pending->lock);
+        SIGNAL_PENDING_CALLS(&runtime->ceval);
+        return result;
+    }
+#else
+    // Python < 3.8 or >= 3.9
+    static int AddPendingCall(int (*func)(void*), void* arg)
+    {
+        // If the interpreter is in the middle of finalizing, we can't add a
+        // pending call. Trying to do so will end up in a SIGSEGV, as
+        // Py_AddPendingCall will not be able to get the interpreter and will
+        // try to dereference a NULL pointer. It's possible this can still
+        // segfault if we happen to get context switched, and maybe we should
+        // just always implement our own AddPendingCall, but I'd like to see if
+        // this works first
+#if GREENLET_PY313
+        if (Py_IsFinalizing()) {
+#else
+        if (_Py_IsFinalizing()) {
+#endif
+#ifdef GREENLET_DEBUG
+            // No need to log in the general case. Yes, we'll leak,
+            // but we're shutting down so it should be ok.
+            fprintf(stderr,
+                    "greenlet: WARNING: Interpreter is finalizing. Ignoring "
+                    "call to Py_AddPendingCall; \n");
+#endif
+            return 0;
+        }
+        return Py_AddPendingCall(func, arg);
+    }
+#endif
+
+
+
+
+};
+};
+
+}; // namespace greenlet
+
+// The intent when GET_THREAD_STATE() is needed multiple times in a
+// function is to take a reference to its return value in a local
+// variable, to avoid the thread-local indirection. On some platforms
+// (macOS), accessing a thread-local involves a function call (plus an
+// initial function call in each function that uses a thread local);
+// in contrast, static volatile variables are at some pre-computed
+// offset.
+typedef greenlet::ThreadStateCreator<greenlet::ThreadState_DestroyNoGIL::MarkGreenletDeadAndQueueCleanup> ThreadStateCreator;
+static thread_local ThreadStateCreator g_thread_state_global;
+#define GET_THREAD_STATE() g_thread_state_global
+
+#endif //T_THREADSTATE_DESTROY
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/TUserGreenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/TUserGreenlet.cpp
new file mode 100644
index 00000000..73a81330
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/TUserGreenlet.cpp
@@ -0,0 +1,662 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/**
+ * Implementation of greenlet::UserGreenlet.
+ *
+ * Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#ifndef T_USER_GREENLET_CPP
+#define T_USER_GREENLET_CPP
+
+#include "greenlet_internal.hpp"
+#include "TGreenlet.hpp"
+
+#include "TThreadStateDestroy.cpp"
+
+
+namespace greenlet {
+using greenlet::refs::BorrowedMainGreenlet;
+greenlet::PythonAllocator<UserGreenlet> UserGreenlet::allocator;
+
+void* UserGreenlet::operator new(size_t UNUSED(count))
+{
+    return allocator.allocate(1);
+}
+
+
+void UserGreenlet::operator delete(void* ptr)
+{
+    return allocator.deallocate(static_cast<UserGreenlet*>(ptr),
+                                1);
+}
+
+
+UserGreenlet::UserGreenlet(PyGreenlet* p, BorrowedGreenlet the_parent)
+    : Greenlet(p), _parent(the_parent)
+{
+}
+
+UserGreenlet::~UserGreenlet()
+{
+    // Python 3.11: If we don't clear out the raw frame datastack
+    // when deleting an unfinished greenlet,
+    // TestLeaks.test_untracked_memory_doesnt_increase_unfinished_thread_dealloc_in_main fails.
+    this->python_state.did_finish(nullptr);
+    this->tp_clear();
+}
+
+
+const BorrowedMainGreenlet
+UserGreenlet::main_greenlet() const
+{
+    return this->_main_greenlet;
+}
+
+
+BorrowedMainGreenlet
+UserGreenlet::find_main_greenlet_in_lineage() const
+{
+    if (this->started()) {
+        assert(this->_main_greenlet);
+        return BorrowedMainGreenlet(this->_main_greenlet);
+    }
+
+    if (!this->_parent) {
+        /* garbage collected greenlet in chain */
+        // XXX: WHAT?
+        return BorrowedMainGreenlet(nullptr);
+    }
+
+    return this->_parent->find_main_greenlet_in_lineage();
+}
+
+
+/**
+ * CAUTION: This will allocate memory and may trigger garbage
+ * collection and arbitrary Python code.
+ */
+OwnedObject
+UserGreenlet::throw_GreenletExit_during_dealloc(const ThreadState& current_thread_state)
+{
+    /* The dying greenlet cannot be a parent of ts_current
+       because the 'parent' field chain would hold a
+       reference */
+    UserGreenlet::ParentIsCurrentGuard with_current_parent(this, current_thread_state);
+
+    // We don't care about the return value, only whether an
+    // exception happened. Whether or not an exception happens,
+    // we need to restore the parent in case the greenlet gets
+    // resurrected.
+    return Greenlet::throw_GreenletExit_during_dealloc(current_thread_state);
+}
+
+ThreadState*
+UserGreenlet::thread_state() const noexcept
+{
+    // TODO: maybe make this throw, if the thread state isn't there?
+    // if (!this->main_greenlet) {
+    //     throw std::runtime_error("No thread state"); // TODO: Better exception
+    // }
+    if (!this->_main_greenlet) {
+        return nullptr;
+    }
+    return this->_main_greenlet->thread_state();
+}
+
+
+bool
+UserGreenlet::was_running_in_dead_thread() const noexcept
+{
+    return this->_main_greenlet && !this->thread_state();
+}
+
+OwnedObject
+UserGreenlet::g_switch()
+{
+    assert(this->args() || PyErr_Occurred());
+
+    try {
+        this->check_switch_allowed();
+    }
+    catch (const PyErrOccurred&) {
+        this->release_args();
+        throw;
+    }
+
+    // Switching greenlets used to attempt to clean out ones that need
+    // deleted *if* we detected a thread switch. Should it still do
+    // that?
+    // An issue is that if we delete a greenlet from another thread,
+    // it gets queued to this thread, and ``kill_greenlet()`` switches
+    // back into the greenlet
+
+    /* find the real target by ignoring dead greenlets,
+       and if necessary starting a greenlet. */
+    switchstack_result_t err;
+    Greenlet* target = this;
+    // TODO: probably cleaner to handle the case where we do
+    // switch to ourself separately from the other cases.
+    // This can probably even further be simplified if we keep
+    // track of the switching_state we're going for and just call
+    // into g_switch() if it's not ourself. The main problem with that
+    // is that we would be using more stack space.
+    bool target_was_me = true;
+    bool was_initial_stub = false;
+    while (target) {
+        if (target->active()) {
+            if (!target_was_me) {
+                target->args() <<= this->args();
+                assert(!this->args());
+            }
+            err = target->g_switchstack();
+            break;
+        }
+        if (!target->started()) {
+            // We never encounter a main greenlet that's not started.
+            assert(!target->main());
+            UserGreenlet* real_target = static_cast<UserGreenlet*>(target);
+            assert(real_target);
+            void* dummymarker;
+            was_initial_stub = true;
+            if (!target_was_me) {
+                target->args() <<= this->args();
+                assert(!this->args());
+            }
+            try {
+                // This can only throw back to us while we're
+                // still in this greenlet. Once the new greenlet
+                // is bootstrapped, it has its own exception state.
+                err = real_target->g_initialstub(&dummymarker);
+            }
+            catch (const PyErrOccurred&) {
+                this->release_args();
+                throw;
+            }
+            catch (const GreenletStartedWhileInPython&) {
+                // The greenlet was started sometime before this
+                // greenlet actually switched to it, i.e.,
+                // "concurrent" calls to switch() or throw().
+                // We need to retry the switch.
+                // Note that the current greenlet has been reset
+                // to this one (or we wouldn't be running!)
+                continue;
+            }
+            break;
+        }
+
+        target = target->parent();
+        target_was_me = false;
+    }
+    // The ``this`` pointer and all other stack or register based
+    // variables are invalid now, at least where things succeed
+    // above.
+    // But this one, probably not so much? It's not clear if it's
+    // safe to throw an exception at this point.
+
+    if (err.status < 0) {
+        // If we get here, either g_initialstub()
+        // failed, or g_switchstack() failed. Either one of those
+        // cases SHOULD leave us in the original greenlet with a valid
+        // stack.
+        return this->on_switchstack_or_initialstub_failure(target, err, target_was_me, was_initial_stub);
+    }
+
+    // err.the_new_current_greenlet would be the same as ``target``,
+    // if target wasn't probably corrupt.
+    return err.the_new_current_greenlet->g_switch_finish(err);
+}
+
+
+
+Greenlet::switchstack_result_t
+UserGreenlet::g_initialstub(void* mark)
+{
+    OwnedObject run;
+
+    // We need to grab a reference to the current switch arguments
+    // in case we're entered concurrently during the call to
+    // GetAttr() and have to try again.
+    // We'll restore them when we return in that case.
+    // Scope them tightly to avoid ref leaks.
+    {
+        SwitchingArgs args(this->args());
+
+        /* save exception in case getattr clears it */
+        PyErrPieces saved;
+
+        /*
+          self.run is the object to call in the new greenlet.
+          This could run arbitrary python code and switch greenlets!
+        */
+        run = this->self().PyRequireAttr(mod_globs->str_run);
+        /* restore saved exception */
+        saved.PyErrRestore();
+
+
+        /* recheck that it's safe to switch in case greenlet reparented anywhere above */
+        this->check_switch_allowed();
+
+        /* by the time we got here another start could happen elsewhere,
+         * that means it should now be a regular switch.
+         * This can happen if the Python code is a subclass that implements
+         * __getattribute__ or __getattr__, or makes ``run`` a descriptor;
+         * all of those can run arbitrary code that switches back into
+         * this greenlet.
+         */
+        if (this->stack_state.started()) {
+            // the successful switch cleared these out, we need to
+            // restore our version. They will be copied on up to the
+            // next target.
+            assert(!this->args());
+            this->args() <<= args;
+            throw GreenletStartedWhileInPython();
+        }
+    }
+
+    // Sweet, if we got here, we have the go-ahead and will switch
+    // greenlets.
+    // Nothing we do from here on out should allow for a thread or
+    // greenlet switch: No arbitrary calls to Python, including
+    // decref'ing
+
+#if GREENLET_USE_CFRAME
+    /* OK, we need it, we're about to switch greenlets, save the state. */
+    /*
+      See green_new(). This is a stack-allocated variable used
+      while *self* is in PyObject_Call().
+      We want to defer copying the state info until we're sure
+      we need it and are in a stable place to do so.
+    */
+    _PyCFrame trace_info;
+
+    this->python_state.set_new_cframe(trace_info);
+#endif
+    /* start the greenlet */
+    ThreadState& thread_state = GET_THREAD_STATE().state();
+    this->stack_state = StackState(mark,
+                                   thread_state.borrow_current()->stack_state);
+    this->python_state.set_initial_state(PyThreadState_GET());
+    this->exception_state.clear();
+    this->_main_greenlet = thread_state.get_main_greenlet();
+
+    /* perform the initial switch */
+    switchstack_result_t err = this->g_switchstack();
+    /* returns twice!
+       The 1st time with ``err == 1``: we are in the new greenlet.
+       This one owns a greenlet that used to be current.
+       The 2nd time with ``err <= 0``: back in the caller's
+       greenlet; this happens if the child finishes or switches
+       explicitly to us. Either way, the ``err`` variable is
+       created twice at the same memory location, but possibly
+       having different ``origin`` values. Note that it's not
+       constructed for the second time until the switch actually happens.
+    */
+    if (err.status == 1) {
+        // In the new greenlet.
+
+        // This never returns! Calling inner_bootstrap steals
+        // the contents of our run object within this stack frame, so
+        // it is not valid to do anything with it.
+        try {
+            this->inner_bootstrap(err.origin_greenlet.relinquish_ownership(),
+                                  run.relinquish_ownership());
+        }
+        // Getting a C++ exception here isn't good. It's probably a
+        // bug in the underlying greenlet, meaning it's probably a
+        // C++ extension. We're going to abort anyway, but try to
+        // display some nice information *if* possible. Some obscure
+        // platforms don't properly support this (old 32-bit Arm, see see
+        // https://github.com/python-greenlet/greenlet/issues/385); that's not
+        // great, but should usually be OK because, as mentioned above, we're
+        // terminating anyway.
+        //
+        // The catching is tested by
+        // ``test_cpp.CPPTests.test_unhandled_exception_in_greenlet_aborts``.
+        //
+        // PyErrOccurred can theoretically be thrown by
+        // inner_bootstrap() -> g_switch_finish(), but that should
+        // never make it back to here. It is a std::exception and
+        // would be caught if it is.
+        catch (const std::exception& e) {
+            std::string base = "greenlet: Unhandled C++ exception: ";
+            base += e.what();
+            Py_FatalError(base.c_str());
+        }
+        catch (...) {
+            // Some compilers/runtimes use exceptions internally.
+            // It appears that GCC on Linux with libstdc++ throws an
+            // exception internally at process shutdown time to unwind
+            // stacks and clean up resources. Depending on exactly
+            // where we are when the process exits, that could result
+            // in an unknown exception getting here. If we
+            // Py_FatalError() or abort() here, we interfere with
+            // orderly process shutdown. Throwing the exception on up
+            // is the right thing to do.
+            //
+            // gevent's ``examples/dns_mass_resolve.py`` demonstrates this.
+#ifndef NDEBUG
+            fprintf(stderr,
+                    "greenlet: inner_bootstrap threw unknown exception; "
+                    "is the process terminating?\n");
+#endif
+            throw;
+        }
+        Py_FatalError("greenlet: inner_bootstrap returned with no exception.\n");
+    }
+
+
+    // In contrast, notice that we're keeping the origin greenlet
+    // around as an owned reference; we need it to call the trace
+    // function for the switch back into the parent. It was only
+    // captured at the time the switch actually happened, though,
+    // so we haven't been keeping an extra reference around this
+    // whole time.
+
+    /* back in the parent */
+    if (err.status < 0) {
+        /* start failed badly, restore greenlet state */
+        this->stack_state = StackState();
+        this->_main_greenlet.CLEAR();
+        // CAUTION: This may run arbitrary Python code.
+        run.CLEAR(); // inner_bootstrap didn't run, we own the reference.
+    }
+
+    // In the success case, the spawned code (inner_bootstrap) will
+    // take care of decrefing this, so we relinquish ownership so as
+    // to not double-decref.
+
+    run.relinquish_ownership();
+
+    return err;
+}
+
+
+void
+UserGreenlet::inner_bootstrap(PyGreenlet* origin_greenlet, PyObject* run)
+{
+    // The arguments here would be another great place for move.
+    // As it is, we take them as a reference so that when we clear
+    // them we clear what's on the stack above us. Do that NOW, and
+    // without using a C++ RAII object,
+    // so there's no way that exiting the parent frame can clear it,
+    // or we clear it unexpectedly. This arises in the context of the
+    // interpreter shutting down. See https://github.com/python-greenlet/greenlet/issues/325
+    //PyObject* run = _run.relinquish_ownership();
+
+    /* in the new greenlet */
+    assert(this->thread_state()->borrow_current() == BorrowedGreenlet(this->_self));
+    // C++ exceptions cannot propagate to the parent greenlet from
+    // here. (TODO: Do we need a catch(...) clause, perhaps on the
+    // function itself? ALl we could do is terminate the program.)
+    // NOTE: On 32-bit Windows, the call chain is extremely
+    // important here in ways that are subtle, having to do with
+    // the depth of the SEH list. The call to restore it MUST NOT
+    // add a new SEH handler to the list, or we'll restore it to
+    // the wrong thing.
+    this->thread_state()->restore_exception_state();
+    /* stack variables from above are no good and also will not unwind! */
+    // EXCEPT: That can't be true, we access run, among others, here.
+
+    this->stack_state.set_active(); /* running */
+
+    // We're about to possibly run Python code again, which
+    // could switch back/away to/from us, so we need to grab the
+    // arguments locally.
+    SwitchingArgs args;
+    args <<= this->args();
+    assert(!this->args());
+
+    // XXX: We could clear this much earlier, right?
+    // Or would that introduce the possibility of running Python
+    // code when we don't want to?
+    // CAUTION: This may run arbitrary Python code.
+    this->_run_callable.CLEAR();
+
+
+    // The first switch we need to manually call the trace
+    // function here instead of in g_switch_finish, because we
+    // never return there.
+    if (OwnedObject tracefunc = this->thread_state()->get_tracefunc()) {
+        OwnedGreenlet trace_origin;
+        trace_origin = origin_greenlet;
+        try {
+            g_calltrace(tracefunc,
+                        args ? mod_globs->event_switch : mod_globs->event_throw,
+                        trace_origin,
+                        this->_self);
+        }
+        catch (const PyErrOccurred&) {
+            /* Turn trace errors into switch throws */
+            args.CLEAR();
+        }
+    }
+
+    // We no longer need the origin, it was only here for
+    // tracing.
+    // We may never actually exit this stack frame so we need
+    // to explicitly clear it.
+    // This could run Python code and switch.
+    Py_CLEAR(origin_greenlet);
+
+    OwnedObject result;
+    if (!args) {
+        /* pending exception */
+        result = NULL;
+    }
+    else {
+        /* call g.run(*args, **kwargs) */
+        // This could result in further switches
+        try {
+            //result = run.PyCall(args.args(), args.kwargs());
+            // CAUTION: Just invoking this, before the function even
+            // runs, may cause memory allocations, which may trigger
+            // GC, which may run arbitrary Python code.
+            result = OwnedObject::consuming(PyObject_Call(run, args.args().borrow(), args.kwargs().borrow()));
+        }
+        catch (...) {
+            // Unhandled C++ exception!
+
+            // If we declare ourselves as noexcept, if we don't catch
+            // this here, most platforms will just abort() the
+            // process. But on 64-bit Windows with older versions of
+            // the C runtime, this can actually corrupt memory and
+            // just return. We see this when compiling with the
+            // Windows 7.0 SDK targeting Windows Server 2008, but not
+            // when using the Appveyor Visual Studio 2019 image. So
+            // this currently only affects Python 2.7 on Windows 64.
+            // That is, the tests pass and the runtime aborts
+            // everywhere else.
+            //
+            // However, if we catch it and try to continue with a
+            // Python error, then all Windows 64 bit platforms corrupt
+            // memory. So all we can do is manually abort, hopefully
+            // with a good error message. (Note that the above was
+            // tested WITHOUT the `/EHr` switch being used at compile
+            // time, so MSVC may have "optimized" out important
+            // checking. Using that switch, we may be in a better
+            // place in terms of memory corruption.) But sometimes it
+            // can't be caught here at all, which is confusing but not
+            // terribly surprising; so again, the G_NOEXCEPT_WIN32
+            // plus "/EHr".
+            //
+            // Hopefully the basic C stdlib is still functional enough
+            // for us to at least print an error.
+            //
+            // It gets more complicated than that, though, on some
+            // platforms, specifically at least Linux/gcc/libstdc++. They use
+            // an exception to unwind the stack when a background
+            // thread exits. (See comments about noexcept.) So this
+            // may not actually represent anything untoward. On those
+            // platforms we allow throws of this to propagate, or
+            // attempt to anyway.
+# if defined(WIN32) || defined(_WIN32)
+            Py_FatalError(
+                "greenlet: Unhandled C++ exception from a greenlet run function. "
+                "Because memory is likely corrupted, terminating process.");
+            std::abort();
+#else
+            throw;
+#endif
+        }
+    }
+    // These lines may run arbitrary code
+    args.CLEAR();
+    Py_CLEAR(run);
+
+    if (!result
+        && mod_globs->PyExc_GreenletExit.PyExceptionMatches()
+        && (this->args())) {
+        // This can happen, for example, if our only reference
+        // goes away after we switch back to the parent.
+        // See test_dealloc_switch_args_not_lost
+        PyErrPieces clear_error;
+        result <<= this->args();
+        result = single_result(result);
+    }
+    this->release_args();
+    this->python_state.did_finish(PyThreadState_GET());
+
+    result = g_handle_exit(result);
+    assert(this->thread_state()->borrow_current() == this->_self);
+
+    /* jump back to parent */
+    this->stack_state.set_inactive(); /* dead */
+
+
+    // TODO: Can we decref some things here? Release our main greenlet
+    // and maybe parent?
+    for (Greenlet* parent = this->_parent;
+         parent;
+         parent = parent->parent()) {
+        // We need to somewhere consume a reference to
+        // the result; in most cases we'll never have control
+        // back in this stack frame again. Calling
+        // green_switch actually adds another reference!
+        // This would probably be clearer with a specific API
+        // to hand results to the parent.
+        parent->args() <<= result;
+        assert(!result);
+        // The parent greenlet now owns the result; in the
+        // typical case we'll never get back here to assign to
+        // result and thus release the reference.
+        try {
+            result = parent->g_switch();
+        }
+        catch (const PyErrOccurred&) {
+            // Ignore, keep passing the error on up.
+        }
+
+        /* Return here means switch to parent failed,
+         * in which case we throw *current* exception
+         * to the next parent in chain.
+         */
+        assert(!result);
+    }
+    /* We ran out of parents, cannot continue */
+    PyErr_WriteUnraisable(this->self().borrow_o());
+    Py_FatalError("greenlet: ran out of parent greenlets while propagating exception; "
+                  "cannot continue");
+    std::abort();
+}
+
+void
+UserGreenlet::run(const BorrowedObject nrun)
+{
+    if (this->started()) {
+        throw AttributeError(
+                        "run cannot be set "
+                        "after the start of the greenlet");
+    }
+    this->_run_callable = nrun;
+}
+
+const OwnedGreenlet
+UserGreenlet::parent() const
+{
+    return this->_parent;
+}
+
+void
+UserGreenlet::parent(const BorrowedObject raw_new_parent)
+{
+    if (!raw_new_parent) {
+        throw AttributeError("can't delete attribute");
+    }
+
+    BorrowedMainGreenlet main_greenlet_of_new_parent;
+    BorrowedGreenlet new_parent(raw_new_parent.borrow()); // could
+                                                          // throw
+                                                          // TypeError!
+    for (BorrowedGreenlet p = new_parent; p; p = p->parent()) {
+        if (p == this->self()) {
+            throw ValueError("cyclic parent chain");
+        }
+        main_greenlet_of_new_parent = p->main_greenlet();
+    }
+
+    if (!main_greenlet_of_new_parent) {
+        throw ValueError("parent must not be garbage collected");
+    }
+
+    if (this->started()
+        && this->_main_greenlet != main_greenlet_of_new_parent) {
+        throw ValueError("parent cannot be on a different thread");
+    }
+
+    this->_parent = new_parent;
+}
+
+void
+UserGreenlet::murder_in_place()
+{
+    this->_main_greenlet.CLEAR();
+    Greenlet::murder_in_place();
+}
+
+bool
+UserGreenlet::belongs_to_thread(const ThreadState* thread_state) const
+{
+    return Greenlet::belongs_to_thread(thread_state) && this->_main_greenlet == thread_state->borrow_main_greenlet();
+}
+
+
+int
+UserGreenlet::tp_traverse(visitproc visit, void* arg)
+{
+    Py_VISIT(this->_parent.borrow_o());
+    Py_VISIT(this->_main_greenlet.borrow_o());
+    Py_VISIT(this->_run_callable.borrow_o());
+
+    return Greenlet::tp_traverse(visit, arg);
+}
+
+int
+UserGreenlet::tp_clear()
+{
+    Greenlet::tp_clear();
+    this->_parent.CLEAR();
+    this->_main_greenlet.CLEAR();
+    this->_run_callable.CLEAR();
+    return 0;
+}
+
+UserGreenlet::ParentIsCurrentGuard::ParentIsCurrentGuard(UserGreenlet* p,
+                                                     const ThreadState& thread_state)
+    : oldparent(p->_parent),
+      greenlet(p)
+{
+    p->_parent = thread_state.get_current();
+}
+
+UserGreenlet::ParentIsCurrentGuard::~ParentIsCurrentGuard()
+{
+    this->greenlet->_parent = oldparent;
+    oldparent.CLEAR();
+}
+
+}; //namespace greenlet
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/__init__.py b/backend/venv/lib/python3.12/site-packages/greenlet/__init__.py
new file mode 100644
index 00000000..b2dcc9bd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/__init__.py
@@ -0,0 +1,71 @@
+# -*- coding: utf-8 -*-
+"""
+The root of the greenlet package.
+"""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+__all__ = [
+    '__version__',
+    '_C_API',
+
+    'GreenletExit',
+    'error',
+
+    'getcurrent',
+    'greenlet',
+
+    'gettrace',
+    'settrace',
+]
+
+# pylint:disable=no-name-in-module
+
+###
+# Metadata
+###
+__version__ = '3.1.1'
+from ._greenlet import _C_API # pylint:disable=no-name-in-module
+
+###
+# Exceptions
+###
+from ._greenlet import GreenletExit
+from ._greenlet import error
+
+###
+# greenlets
+###
+from ._greenlet import getcurrent
+from ._greenlet import greenlet
+
+###
+# tracing
+###
+try:
+    from ._greenlet import gettrace
+    from ._greenlet import settrace
+except ImportError:
+    # Tracing wasn't supported.
+    # XXX: The option to disable it was removed in 1.0,
+    # so this branch should be dead code.
+    pass
+
+###
+# Constants
+# These constants aren't documented and aren't recommended.
+# In 1.0, USE_GC and USE_TRACING are always true, and USE_CONTEXT_VARS
+# is the same as ``sys.version_info[:2] >= 3.7``
+###
+from ._greenlet import GREENLET_USE_CONTEXT_VARS # pylint:disable=unused-import
+from ._greenlet import GREENLET_USE_GC # pylint:disable=unused-import
+from ._greenlet import GREENLET_USE_TRACING # pylint:disable=unused-import
+
+# Controlling the use of the gc module. Provisional API for this greenlet
+# implementation in 2.0.
+from ._greenlet import CLOCKS_PER_SEC # pylint:disable=unused-import
+from ._greenlet import enable_optional_cleanup # pylint:disable=unused-import
+from ._greenlet import get_clocks_used_doing_optional_cleanup # pylint:disable=unused-import
+
+# Other APIS in the _greenlet module are for test support.
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..1bd73366
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/_greenlet.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/greenlet/_greenlet.cpython-312-darwin.so
new file mode 100755
index 00000000..0d34002b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/_greenlet.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.cpp
new file mode 100644
index 00000000..e8d92a00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.cpp
@@ -0,0 +1,320 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+/* Format with:
+ *  clang-format -i --style=file src/greenlet/greenlet.c
+ *
+ *
+ * Fix missing braces with:
+ *   clang-tidy src/greenlet/greenlet.c -fix -checks="readability-braces-around-statements"
+*/
+#include <cstdlib>
+#include <string>
+#include <algorithm>
+#include <exception>
+
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+#include "structmember.h" // PyMemberDef
+
+#include "greenlet_internal.hpp"
+// Code after this point can assume access to things declared in stdint.h,
+// including the fixed-width types. This goes for the platform-specific switch functions
+// as well.
+#include "greenlet_refs.hpp"
+#include "greenlet_slp_switch.hpp"
+
+#include "greenlet_thread_support.hpp"
+#include "TGreenlet.hpp"
+
+#include "TGreenletGlobals.cpp"
+
+#include "TGreenlet.cpp"
+#include "TMainGreenlet.cpp"
+#include "TUserGreenlet.cpp"
+#include "TBrokenGreenlet.cpp"
+#include "TExceptionState.cpp"
+#include "TPythonState.cpp"
+#include "TStackState.cpp"
+
+#include "TThreadState.hpp"
+#include "TThreadStateCreator.hpp"
+#include "TThreadStateDestroy.cpp"
+
+#include "PyGreenlet.cpp"
+#include "PyGreenletUnswitchable.cpp"
+#include "CObjects.cpp"
+
+using greenlet::LockGuard;
+using greenlet::LockInitError;
+using greenlet::PyErrOccurred;
+using greenlet::Require;
+
+using greenlet::g_handle_exit;
+using greenlet::single_result;
+
+using greenlet::Greenlet;
+using greenlet::UserGreenlet;
+using greenlet::MainGreenlet;
+using greenlet::BrokenGreenlet;
+using greenlet::ThreadState;
+using greenlet::PythonState;
+
+
+
+// ******* Implementation of things from included files
+template<typename T, greenlet::refs::TypeChecker TC>
+greenlet::refs::_BorrowedGreenlet<T, TC>& greenlet::refs::_BorrowedGreenlet<T, TC>::operator=(const greenlet::refs::BorrowedObject& other)
+{
+    this->_set_raw_pointer(static_cast<PyObject*>(other));
+    return *this;
+}
+
+template <typename T, greenlet::refs::TypeChecker TC>
+inline greenlet::refs::_BorrowedGreenlet<T, TC>::operator Greenlet*() const noexcept
+{
+    if (!this->p) {
+        return nullptr;
+    }
+    return reinterpret_cast<PyGreenlet*>(this->p)->pimpl;
+}
+
+template<typename T, greenlet::refs::TypeChecker TC>
+greenlet::refs::_BorrowedGreenlet<T, TC>::_BorrowedGreenlet(const BorrowedObject& p)
+    : BorrowedReference<T, TC>(nullptr)
+{
+
+    this->_set_raw_pointer(p.borrow());
+}
+
+template <typename T, greenlet::refs::TypeChecker TC>
+inline greenlet::refs::_OwnedGreenlet<T, TC>::operator Greenlet*() const noexcept
+{
+    if (!this->p) {
+        return nullptr;
+    }
+    return reinterpret_cast<PyGreenlet*>(this->p)->pimpl;
+}
+
+
+
+#ifdef __clang__
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wmissing-field-initializers"
+#    pragma clang diagnostic ignored "-Wwritable-strings"
+#elif defined(__GNUC__)
+#    pragma GCC diagnostic push
+//  warning: ISO C++ forbids converting a string constant to ‘char*’
+// (The python APIs aren't const correct and accept writable char*)
+#    pragma GCC diagnostic ignored "-Wwrite-strings"
+#endif
+
+
+/***********************************************************
+
+A PyGreenlet is a range of C stack addresses that must be
+saved and restored in such a way that the full range of the
+stack contains valid data when we switch to it.
+
+Stack layout for a greenlet:
+
+               |     ^^^       |
+               |  older data   |
+               |               |
+  stack_stop . |_______________|
+        .      |               |
+        .      | greenlet data |
+        .      |   in stack    |
+        .    * |_______________| . .  _____________  stack_copy + stack_saved
+        .      |               |     |             |
+        .      |     data      |     |greenlet data|
+        .      |   unrelated   |     |    saved    |
+        .      |      to       |     |   in heap   |
+ stack_start . |     this      | . . |_____________| stack_copy
+               |   greenlet    |
+               |               |
+               |  newer data   |
+               |     vvv       |
+
+
+Note that a greenlet's stack data is typically partly at its correct
+place in the stack, and partly saved away in the heap, but always in
+the above configuration: two blocks, the more recent one in the heap
+and the older one still in the stack (either block may be empty).
+
+Greenlets are chained: each points to the previous greenlet, which is
+the one that owns the data currently in the C stack above my
+stack_stop.  The currently running greenlet is the first element of
+this chain.  The main (initial) greenlet is the last one.  Greenlets
+whose stack is entirely in the heap can be skipped from the chain.
+
+The chain is not related to execution order, but only to the order
+in which bits of C stack happen to belong to greenlets at a particular
+point in time.
+
+The main greenlet doesn't have a stack_stop: it is responsible for the
+complete rest of the C stack, and we don't know where it begins.  We
+use (char*) -1, the largest possible address.
+
+States:
+  stack_stop == NULL && stack_start == NULL:  did not start yet
+  stack_stop != NULL && stack_start == NULL:  already finished
+  stack_stop != NULL && stack_start != NULL:  active
+
+The running greenlet's stack_start is undefined but not NULL.
+
+ ***********************************************************/
+
+
+
+
+/***********************************************************/
+
+/* Some functions must not be inlined:
+   * slp_restore_state, when inlined into slp_switch might cause
+     it to restore stack over its own local variables
+   * slp_save_state, when inlined would add its own local
+     variables to the saved stack, wasting space
+   * slp_switch, cannot be inlined for obvious reasons
+   * g_initialstub, when inlined would receive a pointer into its
+     own stack frame, leading to incomplete stack save/restore
+
+g_initialstub is a member function and declared virtual so that the
+compiler always calls it through a vtable.
+
+slp_save_state and slp_restore_state are also member functions. They
+are called from trampoline functions that themselves are declared as
+not eligible for inlining.
+*/
+
+extern "C" {
+static int GREENLET_NOINLINE(slp_save_state_trampoline)(char* stackref)
+{
+    return switching_thread_state->slp_save_state(stackref);
+}
+static void GREENLET_NOINLINE(slp_restore_state_trampoline)()
+{
+    switching_thread_state->slp_restore_state();
+}
+}
+
+
+/***********************************************************/
+
+
+#include "PyModule.cpp"
+
+
+
+static PyObject*
+greenlet_internal_mod_init() noexcept
+{
+    static void* _PyGreenlet_API[PyGreenlet_API_pointers];
+
+    try {
+        CreatedModule m(greenlet_module_def);
+
+        Require(PyType_Ready(&PyGreenlet_Type));
+        Require(PyType_Ready(&PyGreenletUnswitchable_Type));
+
+        mod_globs = new greenlet::GreenletGlobals;
+        ThreadState::init();
+
+        m.PyAddObject("greenlet", PyGreenlet_Type);
+        m.PyAddObject("UnswitchableGreenlet", PyGreenletUnswitchable_Type);
+        m.PyAddObject("error", mod_globs->PyExc_GreenletError);
+        m.PyAddObject("GreenletExit", mod_globs->PyExc_GreenletExit);
+
+        m.PyAddObject("GREENLET_USE_GC", 1);
+        m.PyAddObject("GREENLET_USE_TRACING", 1);
+        m.PyAddObject("GREENLET_USE_CONTEXT_VARS", 1L);
+        m.PyAddObject("GREENLET_USE_STANDARD_THREADING", 1L);
+
+        OwnedObject clocks_per_sec = OwnedObject::consuming(PyLong_FromSsize_t(CLOCKS_PER_SEC));
+        m.PyAddObject("CLOCKS_PER_SEC", clocks_per_sec);
+
+        /* also publish module-level data as attributes of the greentype. */
+        // XXX: This is weird, and enables a strange pattern of
+        // confusing the class greenlet with the module greenlet; with
+        // the exception of (possibly) ``getcurrent()``, this
+        // shouldn't be encouraged so don't add new items here.
+        for (const char* const* p = copy_on_greentype; *p; p++) {
+            OwnedObject o = m.PyRequireAttr(*p);
+            PyDict_SetItemString(PyGreenlet_Type.tp_dict, *p, o.borrow());
+        }
+
+        /*
+         * Expose C API
+         */
+
+        /* types */
+        _PyGreenlet_API[PyGreenlet_Type_NUM] = (void*)&PyGreenlet_Type;
+
+        /* exceptions */
+        _PyGreenlet_API[PyExc_GreenletError_NUM] = (void*)mod_globs->PyExc_GreenletError;
+        _PyGreenlet_API[PyExc_GreenletExit_NUM] = (void*)mod_globs->PyExc_GreenletExit;
+
+        /* methods */
+        _PyGreenlet_API[PyGreenlet_New_NUM] = (void*)PyGreenlet_New;
+        _PyGreenlet_API[PyGreenlet_GetCurrent_NUM] = (void*)PyGreenlet_GetCurrent;
+        _PyGreenlet_API[PyGreenlet_Throw_NUM] = (void*)PyGreenlet_Throw;
+        _PyGreenlet_API[PyGreenlet_Switch_NUM] = (void*)PyGreenlet_Switch;
+        _PyGreenlet_API[PyGreenlet_SetParent_NUM] = (void*)PyGreenlet_SetParent;
+
+        /* Previously macros, but now need to be functions externally. */
+        _PyGreenlet_API[PyGreenlet_MAIN_NUM] = (void*)Extern_PyGreenlet_MAIN;
+        _PyGreenlet_API[PyGreenlet_STARTED_NUM] = (void*)Extern_PyGreenlet_STARTED;
+        _PyGreenlet_API[PyGreenlet_ACTIVE_NUM] = (void*)Extern_PyGreenlet_ACTIVE;
+        _PyGreenlet_API[PyGreenlet_GET_PARENT_NUM] = (void*)Extern_PyGreenlet_GET_PARENT;
+
+        /* XXX: Note that our module name is ``greenlet._greenlet``, but for
+           backwards compatibility with existing C code, we need the _C_API to
+           be directly in greenlet.
+        */
+        const NewReference c_api_object(Require(
+                                           PyCapsule_New(
+                                               (void*)_PyGreenlet_API,
+                                               "greenlet._C_API",
+                                               NULL)));
+        m.PyAddObject("_C_API", c_api_object);
+        assert(c_api_object.REFCNT() == 2);
+
+        // cerr << "Sizes:"
+        //      << "\n\tGreenlet       : " << sizeof(Greenlet)
+        //      << "\n\tUserGreenlet   : " << sizeof(UserGreenlet)
+        //      << "\n\tMainGreenlet   : " << sizeof(MainGreenlet)
+        //      << "\n\tExceptionState : " << sizeof(greenlet::ExceptionState)
+        //      << "\n\tPythonState    : " << sizeof(greenlet::PythonState)
+        //      << "\n\tStackState     : " << sizeof(greenlet::StackState)
+        //      << "\n\tSwitchingArgs  : " << sizeof(greenlet::SwitchingArgs)
+        //      << "\n\tOwnedObject    : " << sizeof(greenlet::refs::OwnedObject)
+        //      << "\n\tBorrowedObject : " << sizeof(greenlet::refs::BorrowedObject)
+        //      << "\n\tPyGreenlet     : " << sizeof(PyGreenlet)
+        //      << endl;
+
+        return m.borrow(); // But really it's the main reference.
+    }
+    catch (const LockInitError& e) {
+        PyErr_SetString(PyExc_MemoryError, e.what());
+        return NULL;
+    }
+    catch (const PyErrOccurred&) {
+        return NULL;
+    }
+
+}
+
+extern "C" {
+
+PyMODINIT_FUNC
+PyInit__greenlet(void)
+{
+    return greenlet_internal_mod_init();
+}
+
+}; // extern C
+
+#ifdef __clang__
+#    pragma clang diagnostic pop
+#elif defined(__GNUC__)
+#    pragma GCC diagnostic pop
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.h b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.h
new file mode 100644
index 00000000..d02a16e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet.h
@@ -0,0 +1,164 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+
+/* Greenlet object interface */
+
+#ifndef Py_GREENLETOBJECT_H
+#define Py_GREENLETOBJECT_H
+
+
+#include <Python.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* This is deprecated and undocumented. It does not change. */
+#define GREENLET_VERSION "1.0.0"
+
+#ifndef GREENLET_MODULE
+#define implementation_ptr_t void*
+#endif
+
+typedef struct _greenlet {
+    PyObject_HEAD
+    PyObject* weakreflist;
+    PyObject* dict;
+    implementation_ptr_t pimpl;
+} PyGreenlet;
+
+#define PyGreenlet_Check(op) (op && PyObject_TypeCheck(op, &PyGreenlet_Type))
+
+
+/* C API functions */
+
+/* Total number of symbols that are exported */
+#define PyGreenlet_API_pointers 12
+
+#define PyGreenlet_Type_NUM 0
+#define PyExc_GreenletError_NUM 1
+#define PyExc_GreenletExit_NUM 2
+
+#define PyGreenlet_New_NUM 3
+#define PyGreenlet_GetCurrent_NUM 4
+#define PyGreenlet_Throw_NUM 5
+#define PyGreenlet_Switch_NUM 6
+#define PyGreenlet_SetParent_NUM 7
+
+#define PyGreenlet_MAIN_NUM 8
+#define PyGreenlet_STARTED_NUM 9
+#define PyGreenlet_ACTIVE_NUM 10
+#define PyGreenlet_GET_PARENT_NUM 11
+
+#ifndef GREENLET_MODULE
+/* This section is used by modules that uses the greenlet C API */
+static void** _PyGreenlet_API = NULL;
+
+#    define PyGreenlet_Type \
+        (*(PyTypeObject*)_PyGreenlet_API[PyGreenlet_Type_NUM])
+
+#    define PyExc_GreenletError \
+        ((PyObject*)_PyGreenlet_API[PyExc_GreenletError_NUM])
+
+#    define PyExc_GreenletExit \
+        ((PyObject*)_PyGreenlet_API[PyExc_GreenletExit_NUM])
+
+/*
+ * PyGreenlet_New(PyObject *args)
+ *
+ * greenlet.greenlet(run, parent=None)
+ */
+#    define PyGreenlet_New                                        \
+        (*(PyGreenlet * (*)(PyObject * run, PyGreenlet * parent)) \
+             _PyGreenlet_API[PyGreenlet_New_NUM])
+
+/*
+ * PyGreenlet_GetCurrent(void)
+ *
+ * greenlet.getcurrent()
+ */
+#    define PyGreenlet_GetCurrent \
+        (*(PyGreenlet * (*)(void)) _PyGreenlet_API[PyGreenlet_GetCurrent_NUM])
+
+/*
+ * PyGreenlet_Throw(
+ *         PyGreenlet *greenlet,
+ *         PyObject *typ,
+ *         PyObject *val,
+ *         PyObject *tb)
+ *
+ * g.throw(...)
+ */
+#    define PyGreenlet_Throw                 \
+        (*(PyObject * (*)(PyGreenlet * self, \
+                          PyObject * typ,    \
+                          PyObject * val,    \
+                          PyObject * tb))    \
+             _PyGreenlet_API[PyGreenlet_Throw_NUM])
+
+/*
+ * PyGreenlet_Switch(PyGreenlet *greenlet, PyObject *args)
+ *
+ * g.switch(*args, **kwargs)
+ */
+#    define PyGreenlet_Switch                                              \
+        (*(PyObject *                                                      \
+           (*)(PyGreenlet * greenlet, PyObject * args, PyObject * kwargs)) \
+             _PyGreenlet_API[PyGreenlet_Switch_NUM])
+
+/*
+ * PyGreenlet_SetParent(PyObject *greenlet, PyObject *new_parent)
+ *
+ * g.parent = new_parent
+ */
+#    define PyGreenlet_SetParent                                 \
+        (*(int (*)(PyGreenlet * greenlet, PyGreenlet * nparent)) \
+             _PyGreenlet_API[PyGreenlet_SetParent_NUM])
+
+/*
+ * PyGreenlet_GetParent(PyObject* greenlet)
+ *
+ * return greenlet.parent;
+ *
+ * This could return NULL even if there is no exception active.
+ * If it does not return NULL, you are responsible for decrementing the
+ * reference count.
+ */
+#     define PyGreenlet_GetParent                                    \
+    (*(PyGreenlet* (*)(PyGreenlet*))                                 \
+     _PyGreenlet_API[PyGreenlet_GET_PARENT_NUM])
+
+/*
+ * deprecated, undocumented alias.
+ */
+#     define PyGreenlet_GET_PARENT PyGreenlet_GetParent
+
+#     define PyGreenlet_MAIN                                         \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_MAIN_NUM])
+
+#     define PyGreenlet_STARTED                                      \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_STARTED_NUM])
+
+#     define PyGreenlet_ACTIVE                                       \
+    (*(int (*)(PyGreenlet*))                                         \
+     _PyGreenlet_API[PyGreenlet_ACTIVE_NUM])
+
+
+
+
+/* Macro that imports greenlet and initializes C API */
+/* NOTE: This has actually moved to ``greenlet._greenlet._C_API``, but we
+   keep the older definition to be sure older code that might have a copy of
+   the header still works. */
+#    define PyGreenlet_Import()                                               \
+        {                                                                     \
+            _PyGreenlet_API = (void**)PyCapsule_Import("greenlet._C_API", 0); \
+        }
+
+#endif /* GREENLET_MODULE */
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_GREENLETOBJECT_H */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_allocator.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_allocator.hpp
new file mode 100644
index 00000000..b452f544
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_allocator.hpp
@@ -0,0 +1,63 @@
+#ifndef GREENLET_ALLOCATOR_HPP
+#define GREENLET_ALLOCATOR_HPP
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+#include <memory>
+#include "greenlet_compiler_compat.hpp"
+
+
+namespace greenlet
+{
+    // This allocator is stateless; all instances are identical.
+    // It can *ONLY* be used when we're sure we're holding the GIL
+    // (Python's allocators require the GIL).
+    template <class T>
+    struct PythonAllocator : public std::allocator<T> {
+
+        PythonAllocator(const PythonAllocator& UNUSED(other))
+            : std::allocator<T>()
+        {
+        }
+
+        PythonAllocator(const std::allocator<T> other)
+            : std::allocator<T>(other)
+        {}
+
+        template <class U>
+        PythonAllocator(const std::allocator<U>& other)
+            : std::allocator<T>(other)
+        {
+        }
+
+        PythonAllocator() : std::allocator<T>() {}
+
+        T* allocate(size_t number_objects, const void* UNUSED(hint)=0)
+        {
+            void* p;
+            if (number_objects == 1)
+                p = PyObject_Malloc(sizeof(T));
+            else
+                p = PyMem_Malloc(sizeof(T) * number_objects);
+            return static_cast<T*>(p);
+        }
+
+        void deallocate(T* t, size_t n)
+        {
+            void* p = t;
+            if (n == 1) {
+                PyObject_Free(p);
+            }
+            else
+                PyMem_Free(p);
+        }
+        // This member is deprecated in C++17 and removed in C++20
+        template< class U >
+        struct rebind {
+            typedef PythonAllocator<U> other;
+        };
+
+    };
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_compiler_compat.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_compiler_compat.hpp
new file mode 100644
index 00000000..af24bd83
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_compiler_compat.hpp
@@ -0,0 +1,98 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+#ifndef GREENLET_COMPILER_COMPAT_HPP
+#define GREENLET_COMPILER_COMPAT_HPP
+
+/**
+ * Definitions to aid with compatibility with different compilers.
+ *
+ * .. caution:: Use extreme care with noexcept.
+ * Some compilers and runtimes, specifically gcc/libgcc/libstdc++ on
+ * Linux, implement stack unwinding by throwing an uncatchable
+ * exception, one that specifically does not appear to be an active
+ * exception to the rest of the runtime. If this happens while we're in a noexcept function,
+ * we have violated our dynamic exception contract, and so the runtime
+ * will call std::terminate(), which kills the process with the
+ * unhelpful message "terminate called without an active exception".
+ *
+ * This has happened in this scenario: A background thread is running
+ * a greenlet that has made a native call and released the GIL.
+ * Meanwhile, the main thread finishes and starts shutting down the
+ * interpreter. When the background thread is scheduled again and
+ * attempts to obtain the  GIL, it notices that the interpreter is
+ * exiting and calls ``pthread_exit()``. This in turn starts to unwind
+ * the stack by throwing that exception. But we had the ``PyCall``
+ * functions annotated as noexcept, so the runtime terminated us.
+ *
+ * #2  0x00007fab26fec2b7 in std::terminate() () from /lib/x86_64-linux-gnu/libstdc++.so.6
+ * #3  0x00007fab26febb3c in __gxx_personality_v0 () from /lib/x86_64-linux-gnu/libstdc++.so.6
+ * #4  0x00007fab26f34de6 in ?? () from /lib/x86_64-linux-gnu/libgcc_s.so.1
+ * #6  0x00007fab276a34c6 in __GI___pthread_unwind  at ./nptl/unwind.c:130
+ * #7  0x00007fab2769bd3a in __do_cancel () at ../sysdeps/nptl/pthreadP.h:280
+ * #8  __GI___pthread_exit (value=value@entry=0x0) at ./nptl/pthread_exit.c:36
+ * #9  0x000000000052e567 in PyThread_exit_thread () at ../Python/thread_pthread.h:370
+ * #10 0x00000000004d60b5 in take_gil at ../Python/ceval_gil.h:224
+ * #11 0x00000000004d65f9 in PyEval_RestoreThread  at ../Python/ceval.c:467
+ * #12 0x000000000060cce3 in setipaddr  at ../Modules/socketmodule.c:1203
+ * #13 0x00000000006101cd in socket_gethostbyname
+ */
+
+#include <cstdint>
+
+#    define G_NO_COPIES_OF_CLS(Cls) private:     \
+    Cls(const Cls& other) = delete; \
+    Cls& operator=(const Cls& other) = delete
+
+#    define G_NO_ASSIGNMENT_OF_CLS(Cls) private:  \
+    Cls& operator=(const Cls& other) = delete
+
+#    define G_NO_COPY_CONSTRUCTOR_OF_CLS(Cls) private: \
+    Cls(const Cls& other) = delete;
+
+
+// CAUTION: MSVC is stupidly picky:
+//
+// "The compiler ignores, without warning, any __declspec keywords
+// placed after * or & and in front of the variable identifier in a
+// declaration."
+// (https://docs.microsoft.com/en-us/cpp/cpp/declspec?view=msvc-160)
+//
+// So pointer return types must be handled differently (because of the
+// trailing *), or you get inscrutable compiler warnings like "error
+// C2059: syntax error: ''"
+//
+// In C++ 11, there is a standard syntax for attributes, and
+// GCC defines an attribute to use with this: [[gnu:noinline]].
+// In the future, this is expected to become standard.
+
+#if defined(__GNUC__) || defined(__clang__)
+/* We used to check for GCC 4+ or 3.4+, but those compilers are
+   laughably out of date. Just assume they support it. */
+#    define GREENLET_NOINLINE(name) __attribute__((noinline)) name
+#    define GREENLET_NOINLINE_P(rtype, name) rtype __attribute__((noinline)) name
+#    define UNUSED(x) UNUSED_ ## x __attribute__((__unused__))
+#elif defined(_MSC_VER)
+/* We used to check for  && (_MSC_VER >= 1300) but that's also out of date. */
+#    define GREENLET_NOINLINE(name) __declspec(noinline) name
+#    define GREENLET_NOINLINE_P(rtype, name) __declspec(noinline) rtype name
+#    define UNUSED(x) UNUSED_ ## x
+#endif
+
+#if defined(_MSC_VER)
+#    define G_NOEXCEPT_WIN32 noexcept
+#else
+#    define G_NOEXCEPT_WIN32
+#endif
+
+#if defined(__GNUC__) && defined(__POWERPC__) && defined(__APPLE__)
+// 32-bit PPC/MacOSX. Only known to be tested on unreleased versions
+// of macOS 10.6 using a macports build gcc 14. It appears that
+// running C++ destructors of thread-local variables is broken.
+
+// See https://github.com/python-greenlet/greenlet/pull/419
+#     define GREENLET_BROKEN_THREAD_LOCAL_CLEANUP_JUST_LEAK 1
+#else
+#     define GREENLET_BROKEN_THREAD_LOCAL_CLEANUP_JUST_LEAK 0
+#endif
+
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_add_pending.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_add_pending.hpp
new file mode 100644
index 00000000..0d28efd3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_add_pending.hpp
@@ -0,0 +1,172 @@
+#ifndef GREENLET_CPYTHON_ADD_PENDING_HPP
+#define GREENLET_CPYTHON_ADD_PENDING_HPP
+
+#if (PY_VERSION_HEX >= 0x30800A0 && PY_VERSION_HEX < 0x3090000) && !(defined(_WIN32) || defined(WIN32))
+// XXX: From Python 3.8a3 [1] up until Python 3.9a6 [2][3],
+// ``Py_AddPendingCall`` would try to produce a Python exception if
+// the interpreter was in the beginning of shutting down when this
+// function is called. However, ``Py_AddPendingCall`` doesn't require
+// the GIL, and we are absolutely not holding it when we make that
+// call. That means that trying to create the Python exception is
+// using the C API in an undefined state; here the C API detects this
+// and aborts the process with an error ("Fatal Python error: Python
+// memory allocator called without holding the GIL": Add ->
+// PyErr_SetString -> PyUnicode_New -> PyObject_Malloc). This arises
+// (obviously) in multi-threaded programs and happens if one thread is
+// exiting and cleaning up its thread-local data while the other
+// thread is trying to shut down the interpreter. A crash on shutdown
+// is still a crash and could result in data loss (e.g., daemon
+// threads are still running, pending signal handlers may be present,
+// buffers may not be flushed, there may be __del__ that need run,
+// etc), so we have to work around it.
+//
+// Of course, we can (and do) check for whether the interpreter is
+// shutting down before calling ``Py_AddPendingCall``, but that's a
+// race condition since we don't hold the GIL, and so we may not
+// actually get the right answer. Plus, ``Py_FinalizeEx`` actually
+// calls ``_Py_FinishPendingCalls`` (which sets the pending->finishing
+// flag, which is used to gate creating the exceptioen) *before*
+// publishing any other data that would let us detect the shutdown
+// (such as runtime->finalizing). So that point is moot.
+//
+// Our solution for those versions is to inline the same code, without
+// the problematic bit that sets the exception. Unfortunately, all of
+// the structure definitions are private/opaque, *and* we can't
+// actually count on being able to include their definitions from
+// ``internal/pycore_*``, because on some platforms those header files
+// are incomplete (i.e., on macOS with macports 3.8, the includes are
+// fine, but on Ubuntu jammy with 3.8 from ppa:deadsnakes or GitHub
+// Actions 3.8 (I think it's Ubuntu 18.04), they con't be used; at
+// least, I couldn't get them to work). So we need to define the
+// structures and _PyRuntime data member ourself. Yet more
+// unfortunately, _PyRuntime  won't link on Windows, so we can only do
+// this on other platforms.
+//
+// [1] https://github.com/python/cpython/commit/842a2f07f2f08a935ef470bfdaeef40f87490cfc
+// [2] https://github.com/python/cpython/commit/cfc3c2f8b34d3864717ab584c5b6c260014ba55a
+// [3] https://github.com/python/cpython/issues/81308
+# define GREENLET_BROKEN_PY_ADD_PENDING 1
+
+// When defining these structures, the important thing is to get
+// binary compatibility, i.e., structure layout. For that, we only
+// need to define fields up to the ones we use; after that they're
+// irrelevant UNLESS the structure is included in another structure
+// *before* the structure we're interested in --- in that case, it
+// must be complete. Ellipsis indicate elided trailing members.
+// Pointer types are changed to void* to keep from having to define
+// more structures.
+
+// From "internal/pycore_atomic.h"
+
+// There are several different definitions of this, including the
+// plain ``int`` version, a ``volatile int`` and an ``_Atomic int``
+// I don't think any of those change the size/layout.
+typedef struct _Py_atomic_int {
+    volatile int _value;
+} _Py_atomic_int;
+
+// This needs too much infrastructure, so we just do a regular store.
+#define _Py_atomic_store_relaxed(ATOMIC_VAL, NEW_VAL) \
+    (ATOMIC_VAL)->_value = NEW_VAL
+
+
+
+// From "internal/pycore_pymem.h"
+#define NUM_GENERATIONS 3
+
+
+struct gc_generation {
+    PyGC_Head head; // We already have this defined.
+    int threshold;
+    int count;
+};
+struct gc_generation_stats {
+    Py_ssize_t collections;
+    Py_ssize_t collected;
+    Py_ssize_t uncollectable;
+};
+
+struct _gc_runtime_state {
+    void *trash_delete_later;
+    int trash_delete_nesting;
+    int enabled;
+    int debug;
+    struct gc_generation generations[NUM_GENERATIONS];
+    void *generation0;
+    struct gc_generation permanent_generation;
+    struct gc_generation_stats generation_stats[NUM_GENERATIONS];
+    int collecting;
+    void *garbage;
+    void *callbacks;
+    Py_ssize_t long_lived_total;
+    Py_ssize_t long_lived_pending;
+};
+
+// From "internal/pycore_pystate.h"
+struct _pending_calls {
+    int finishing;
+    PyThread_type_lock lock;
+    _Py_atomic_int calls_to_do;
+    int async_exc;
+#define NPENDINGCALLS 32
+    struct {
+        int (*func)(void *);
+        void *arg;
+    } calls[NPENDINGCALLS];
+    int first;
+    int last;
+};
+
+struct _ceval_runtime_state {
+    int recursion_limit;
+    int tracing_possible;
+    _Py_atomic_int eval_breaker;
+    _Py_atomic_int gil_drop_request;
+    struct _pending_calls pending;
+    // ...
+};
+
+typedef struct pyruntimestate {
+    int preinitializing;
+    int preinitialized;
+    int core_initialized;
+    int initialized;
+    void *finalizing;
+
+    struct pyinterpreters {
+        PyThread_type_lock mutex;
+        void *head;
+        void *main;
+        int64_t next_id;
+    } interpreters;
+    // XXX Remove this field once we have a tp_* slot.
+    struct _xidregistry {
+        PyThread_type_lock mutex;
+        void *head;
+    } xidregistry;
+
+    unsigned long main_thread;
+
+#define NEXITFUNCS 32
+    void (*exitfuncs[NEXITFUNCS])(void);
+    int nexitfuncs;
+
+    struct _gc_runtime_state gc;
+    struct _ceval_runtime_state ceval;
+    // ...
+} _PyRuntimeState;
+
+#define SIGNAL_PENDING_CALLS(ceval) \
+    do { \
+        _Py_atomic_store_relaxed(&(ceval)->pending.calls_to_do, 1); \
+        _Py_atomic_store_relaxed(&(ceval)->eval_breaker, 1); \
+    } while (0)
+
+extern _PyRuntimeState _PyRuntime;
+
+#else
+# define GREENLET_BROKEN_PY_ADD_PENDING 0
+#endif
+
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_compat.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_compat.hpp
new file mode 100644
index 00000000..ce5fd882
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_cpython_compat.hpp
@@ -0,0 +1,142 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+#ifndef GREENLET_CPYTHON_COMPAT_H
+#define GREENLET_CPYTHON_COMPAT_H
+
+/**
+ * Helpers for compatibility with multiple versions of CPython.
+ */
+
+#define PY_SSIZE_T_CLEAN
+#include "Python.h"
+
+
+#if PY_VERSION_HEX >= 0x30A00B1
+#    define GREENLET_PY310 1
+#else
+#    define GREENLET_PY310 0
+#endif
+
+/*
+Python 3.10 beta 1 changed tstate->use_tracing to a nested cframe member.
+See https://github.com/python/cpython/pull/25276
+We have to save and restore this as well.
+
+Python 3.13 removed PyThreadState.cframe (GH-108035).
+*/
+#if GREENLET_PY310 && PY_VERSION_HEX < 0x30D0000
+#    define GREENLET_USE_CFRAME 1
+#else
+#    define GREENLET_USE_CFRAME 0
+#endif
+
+
+#if PY_VERSION_HEX >= 0x30B00A4
+/*
+Greenlet won't compile on anything older than Python 3.11 alpha 4 (see
+https://bugs.python.org/issue46090). Summary of breaking internal changes:
+- Python 3.11 alpha 1 changed how frame objects are represented internally.
+  - https://github.com/python/cpython/pull/30122
+- Python 3.11 alpha 3 changed how recursion limits are stored.
+  - https://github.com/python/cpython/pull/29524
+- Python 3.11 alpha 4 changed how exception state is stored. It also includes a
+  change to help greenlet save and restore the interpreter frame "data stack".
+  - https://github.com/python/cpython/pull/30122
+  - https://github.com/python/cpython/pull/30234
+*/
+#    define GREENLET_PY311 1
+#else
+#    define GREENLET_PY311 0
+#endif
+
+
+#if PY_VERSION_HEX >= 0x30C0000
+#    define GREENLET_PY312 1
+#else
+#    define GREENLET_PY312 0
+#endif
+
+#if PY_VERSION_HEX >= 0x30D0000
+#    define GREENLET_PY313 1
+#else
+#    define GREENLET_PY313 0
+#endif
+
+#ifndef Py_SET_REFCNT
+/* Py_REFCNT and Py_SIZE macros are converted to functions
+https://bugs.python.org/issue39573 */
+#    define Py_SET_REFCNT(obj, refcnt) Py_REFCNT(obj) = (refcnt)
+#endif
+
+#ifndef _Py_DEC_REFTOTAL
+/* _Py_DEC_REFTOTAL macro has been removed from Python 3.9 by:
+  https://github.com/python/cpython/commit/49932fec62c616ec88da52642339d83ae719e924
+
+  The symbol we use to replace it was removed by at least 3.12.
+*/
+#    ifdef Py_REF_DEBUG
+#      if GREENLET_PY312
+#         define _Py_DEC_REFTOTAL
+#      else
+#        define _Py_DEC_REFTOTAL _Py_RefTotal--
+#      endif
+#    else
+#        define _Py_DEC_REFTOTAL
+#    endif
+#endif
+// Define these flags like Cython does if we're on an old version.
+#ifndef Py_TPFLAGS_CHECKTYPES
+  #define Py_TPFLAGS_CHECKTYPES 0
+#endif
+#ifndef Py_TPFLAGS_HAVE_INDEX
+  #define Py_TPFLAGS_HAVE_INDEX 0
+#endif
+#ifndef Py_TPFLAGS_HAVE_NEWBUFFER
+  #define Py_TPFLAGS_HAVE_NEWBUFFER 0
+#endif
+
+#ifndef Py_TPFLAGS_HAVE_VERSION_TAG
+   #define Py_TPFLAGS_HAVE_VERSION_TAG 0
+#endif
+
+#define G_TPFLAGS_DEFAULT Py_TPFLAGS_DEFAULT | Py_TPFLAGS_HAVE_VERSION_TAG | Py_TPFLAGS_CHECKTYPES | Py_TPFLAGS_HAVE_NEWBUFFER | Py_TPFLAGS_HAVE_GC
+
+
+#if PY_VERSION_HEX < 0x03090000
+// The official version only became available in 3.9
+#    define PyObject_GC_IsTracked(o) _PyObject_GC_IS_TRACKED(o)
+#endif
+
+
+// bpo-43760 added PyThreadState_EnterTracing() to Python 3.11.0a2
+#if PY_VERSION_HEX < 0x030B00A2 && !defined(PYPY_VERSION)
+static inline void PyThreadState_EnterTracing(PyThreadState *tstate)
+{
+    tstate->tracing++;
+#if PY_VERSION_HEX >= 0x030A00A1
+    tstate->cframe->use_tracing = 0;
+#else
+    tstate->use_tracing = 0;
+#endif
+}
+#endif
+
+// bpo-43760 added PyThreadState_LeaveTracing() to Python 3.11.0a2
+#if PY_VERSION_HEX < 0x030B00A2 && !defined(PYPY_VERSION)
+static inline void PyThreadState_LeaveTracing(PyThreadState *tstate)
+{
+    tstate->tracing--;
+    int use_tracing = (tstate->c_tracefunc != NULL
+                       || tstate->c_profilefunc != NULL);
+#if PY_VERSION_HEX >= 0x030A00A1
+    tstate->cframe->use_tracing = use_tracing;
+#else
+    tstate->use_tracing = use_tracing;
+#endif
+}
+#endif
+
+#if !defined(Py_C_RECURSION_LIMIT) && defined(C_RECURSION_LIMIT)
+#  define Py_C_RECURSION_LIMIT C_RECURSION_LIMIT
+#endif
+
+#endif /* GREENLET_CPYTHON_COMPAT_H */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_exceptions.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_exceptions.hpp
new file mode 100644
index 00000000..617f07c2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_exceptions.hpp
@@ -0,0 +1,171 @@
+#ifndef GREENLET_EXCEPTIONS_HPP
+#define GREENLET_EXCEPTIONS_HPP
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+#include <stdexcept>
+#include <string>
+
+#ifdef __clang__
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wunused-function"
+#endif
+
+namespace greenlet {
+
+    class PyErrOccurred : public std::runtime_error
+    {
+    public:
+
+        // CAUTION: In debug builds, may run arbitrary Python code.
+        static const PyErrOccurred
+        from_current()
+        {
+            assert(PyErr_Occurred());
+#ifndef NDEBUG
+            // This is not exception safe, and
+            // not necessarily safe in general (what if it switches?)
+            // But we only do this in debug mode, where we are in
+            // tight control of what exceptions are getting raised and
+            // can prevent those issues.
+
+            // You can't call PyObject_Str with a pending exception.
+            PyObject* typ;
+            PyObject* val;
+            PyObject* tb;
+
+            PyErr_Fetch(&typ, &val, &tb);
+            PyObject* typs = PyObject_Str(typ);
+            PyObject* vals = PyObject_Str(val ? val : typ);
+            const char* typ_msg = PyUnicode_AsUTF8(typs);
+            const char* val_msg = PyUnicode_AsUTF8(vals);
+            PyErr_Restore(typ, val, tb);
+
+            std::string msg(typ_msg);
+            msg += ": ";
+            msg += val_msg;
+            PyErrOccurred ex(msg);
+            Py_XDECREF(typs);
+            Py_XDECREF(vals);
+
+            return ex;
+#else
+            return PyErrOccurred();
+#endif
+        }
+
+        PyErrOccurred() : std::runtime_error("")
+        {
+            assert(PyErr_Occurred());
+        }
+
+        PyErrOccurred(const std::string& msg) : std::runtime_error(msg)
+        {
+            assert(PyErr_Occurred());
+        }
+
+        PyErrOccurred(PyObject* exc_kind, const char* const msg)
+            : std::runtime_error(msg)
+        {
+            PyErr_SetString(exc_kind, msg);
+        }
+
+        PyErrOccurred(PyObject* exc_kind, const std::string msg)
+            : std::runtime_error(msg)
+        {
+            // This copies the c_str, so we don't have any lifetime
+            // issues to worry about.
+            PyErr_SetString(exc_kind, msg.c_str());
+        }
+
+        PyErrOccurred(PyObject* exc_kind,
+                      const std::string msg, //This is the format
+                                             //string; that's not
+                                             //usually safe!
+
+                      PyObject* borrowed_obj_one, PyObject* borrowed_obj_two)
+            : std::runtime_error(msg)
+        {
+
+            //This is designed specifically for the
+            //``check_switch_allowed`` function.
+
+            // PyObject_Str and PyObject_Repr are safe to call with
+            // NULL pointers; they return the string "<NULL>" in that
+            // case.
+            // This function always returns null.
+            PyErr_Format(exc_kind,
+                         msg.c_str(),
+                         borrowed_obj_one, borrowed_obj_two);
+        }
+    };
+
+    class TypeError : public PyErrOccurred
+    {
+    public:
+        TypeError(const char* const what)
+            : PyErrOccurred(PyExc_TypeError, what)
+        {
+        }
+        TypeError(const std::string what)
+            : PyErrOccurred(PyExc_TypeError, what)
+        {
+        }
+    };
+
+    class ValueError : public PyErrOccurred
+    {
+    public:
+        ValueError(const char* const what)
+            : PyErrOccurred(PyExc_ValueError, what)
+        {
+        }
+    };
+
+    class AttributeError : public PyErrOccurred
+    {
+    public:
+        AttributeError(const char* const what)
+            : PyErrOccurred(PyExc_AttributeError, what)
+        {
+        }
+    };
+
+    /**
+     * Calls `Py_FatalError` when constructed, so you can't actually
+     * throw this. It just makes static analysis easier.
+     */
+    class PyFatalError : public std::runtime_error
+    {
+    public:
+        PyFatalError(const char* const msg)
+            : std::runtime_error(msg)
+        {
+            Py_FatalError(msg);
+        }
+    };
+
+    static inline PyObject*
+    Require(PyObject* p, const std::string& msg="")
+    {
+        if (!p) {
+            throw PyErrOccurred(msg);
+        }
+        return p;
+    };
+
+    static inline void
+    Require(const int retval)
+    {
+        if (retval < 0) {
+            throw PyErrOccurred();
+        }
+    };
+
+
+};
+#ifdef __clang__
+#    pragma clang diagnostic pop
+#endif
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_internal.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_internal.hpp
new file mode 100644
index 00000000..f2b15d5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_internal.hpp
@@ -0,0 +1,107 @@
+/* -*- indent-tabs-mode: nil; tab-width: 4; -*- */
+#ifndef GREENLET_INTERNAL_H
+#define GREENLET_INTERNAL_H
+#ifdef __clang__
+#    pragma clang diagnostic push
+#    pragma clang diagnostic ignored "-Wunused-function"
+#endif
+
+/**
+ * Implementation helpers.
+ *
+ * C++ templates and inline functions should go here.
+ */
+#define PY_SSIZE_T_CLEAN
+#include "greenlet_compiler_compat.hpp"
+#include "greenlet_cpython_compat.hpp"
+#include "greenlet_exceptions.hpp"
+#include "TGreenlet.hpp"
+#include "greenlet_allocator.hpp"
+
+#include <vector>
+#include <string>
+
+#define GREENLET_MODULE
+struct _greenlet;
+typedef struct _greenlet PyGreenlet;
+namespace greenlet {
+
+    class ThreadState;
+    // We can't use the PythonAllocator for this, because we push to it
+    // from the thread state destructor, which doesn't have the GIL,
+    // and Python's allocators can only be called with the GIL.
+    typedef std::vector<ThreadState*> cleanup_queue_t;
+
+};
+
+
+#define implementation_ptr_t greenlet::Greenlet*
+
+
+#include "greenlet.h"
+
+void
+greenlet::refs::MainGreenletExactChecker(void *p)
+{
+    if (!p) {
+        return;
+    }
+    // We control the class of the main greenlet exactly.
+    if (Py_TYPE(p) != &PyGreenlet_Type) {
+        std::string err("MainGreenlet: Expected exactly a greenlet, not a ");
+        err += Py_TYPE(p)->tp_name;
+        throw greenlet::TypeError(err);
+    }
+
+    // Greenlets from dead threads no longer respond to main() with a
+    // true value; so in that case we need to perform an additional
+    // check.
+    Greenlet* g = static_cast<PyGreenlet*>(p)->pimpl;
+    if (g->main()) {
+        return;
+    }
+    if (!dynamic_cast<MainGreenlet*>(g)) {
+        std::string err("MainGreenlet: Expected exactly a main greenlet, not a ");
+        err += Py_TYPE(p)->tp_name;
+        throw greenlet::TypeError(err);
+    }
+}
+
+
+
+template <typename T, greenlet::refs::TypeChecker TC>
+inline greenlet::Greenlet* greenlet::refs::_OwnedGreenlet<T, TC>::operator->() const noexcept
+{
+    return reinterpret_cast<PyGreenlet*>(this->p)->pimpl;
+}
+
+template <typename T, greenlet::refs::TypeChecker TC>
+inline greenlet::Greenlet* greenlet::refs::_BorrowedGreenlet<T, TC>::operator->() const noexcept
+{
+    return reinterpret_cast<PyGreenlet*>(this->p)->pimpl;
+}
+
+#include <memory>
+#include <stdexcept>
+
+
+extern PyTypeObject PyGreenlet_Type;
+
+
+
+/**
+  * Forward declarations needed in multiple files.
+  */
+static PyObject* green_switch(PyGreenlet* self, PyObject* args, PyObject* kwargs);
+
+
+#ifdef __clang__
+#    pragma clang diagnostic pop
+#endif
+
+
+#endif
+
+// Local Variables:
+// flycheck-clang-include-path: ("../../include" "/opt/local/Library/Frameworks/Python.framework/Versions/3.10/include/python3.10")
+// End:
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_refs.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_refs.hpp
new file mode 100644
index 00000000..b7e5e3f2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_refs.hpp
@@ -0,0 +1,1118 @@
+#ifndef GREENLET_REFS_HPP
+#define GREENLET_REFS_HPP
+
+#define PY_SSIZE_T_CLEAN
+#include <Python.h>
+
+#include <string>
+
+//#include "greenlet_internal.hpp"
+#include "greenlet_compiler_compat.hpp"
+#include "greenlet_cpython_compat.hpp"
+#include "greenlet_exceptions.hpp"
+
+struct _greenlet;
+struct _PyMainGreenlet;
+
+typedef struct _greenlet PyGreenlet;
+extern PyTypeObject PyGreenlet_Type;
+
+
+#ifdef  GREENLET_USE_STDIO
+#include <iostream>
+using std::cerr;
+using std::endl;
+#endif
+
+namespace greenlet
+{
+    class Greenlet;
+
+    namespace refs
+    {
+        // Type checkers throw a TypeError if the argument is not
+        // null, and isn't of the required Python type.
+        // (We can't use most of the defined type checkers
+        // like PyList_Check, etc, directly, because they are
+        // implemented as macros.)
+        typedef void (*TypeChecker)(void*);
+
+        void
+        NoOpChecker(void*)
+        {
+            return;
+        }
+
+        void
+        GreenletChecker(void *p)
+        {
+            if (!p) {
+                return;
+            }
+
+            PyTypeObject* typ = Py_TYPE(p);
+            // fast, common path. (PyObject_TypeCheck is a macro or
+            // static inline function, and it also does a
+            // direct comparison of the type pointers, but its fast
+            // path only handles one type)
+            if (typ == &PyGreenlet_Type) {
+                return;
+            }
+
+            if (!PyObject_TypeCheck(p, &PyGreenlet_Type)) {
+                std::string err("GreenletChecker: Expected any type of greenlet, not ");
+                err += Py_TYPE(p)->tp_name;
+                throw TypeError(err);
+            }
+        }
+
+        void
+        MainGreenletExactChecker(void *p);
+
+        template <typename T, TypeChecker>
+        class PyObjectPointer;
+
+        template<typename T, TypeChecker>
+        class OwnedReference;
+
+
+        template<typename T, TypeChecker>
+        class BorrowedReference;
+
+        typedef BorrowedReference<PyObject, NoOpChecker> BorrowedObject;
+        typedef OwnedReference<PyObject, NoOpChecker> OwnedObject;
+
+        class ImmortalObject;
+        class ImmortalString;
+
+        template<typename T, TypeChecker TC>
+        class _OwnedGreenlet;
+
+        typedef _OwnedGreenlet<PyGreenlet, GreenletChecker> OwnedGreenlet;
+        typedef _OwnedGreenlet<PyGreenlet, MainGreenletExactChecker> OwnedMainGreenlet;
+
+        template<typename T, TypeChecker TC>
+        class _BorrowedGreenlet;
+
+        typedef _BorrowedGreenlet<PyGreenlet, GreenletChecker> BorrowedGreenlet;
+
+        void
+        ContextExactChecker(void *p)
+        {
+            if (!p) {
+                return;
+            }
+            if (!PyContext_CheckExact(p)) {
+                throw TypeError(
+                    "greenlet context must be a contextvars.Context or None"
+                );
+            }
+        }
+
+        typedef OwnedReference<PyObject, ContextExactChecker> OwnedContext;
+    }
+}
+
+namespace greenlet {
+
+
+    namespace refs {
+    // A set of classes to make reference counting rules in python
+    // code explicit.
+    //
+    // Rules of use:
+    // (1) Functions returning a new reference that the caller of the
+    // function is expected to dispose of should return a
+    // ``OwnedObject`` object. This object automatically releases its
+    // reference when it goes out of scope. It works like a ``std::shared_ptr``
+    // and can be copied or used as a function parameter (but don't do
+    // that). Note that constructing a ``OwnedObject`` from a
+    // PyObject* steals the reference.
+    // (2) Parameters to functions should be either a
+    // ``OwnedObject&``, or, more generally, a ``PyObjectPointer&``.
+    // If the function needs to create its own new reference, it can
+    // do so by copying to a local ``OwnedObject``.
+    // (3) Functions returning an existing pointer that is NOT
+    // incref'd, and which the caller MUST NOT decref,
+    // should return a ``BorrowedObject``.
+
+    // XXX: The following two paragraphs do not hold for all platforms.
+    // Notably, 32-bit PPC Linux passes structs by reference, not by
+    // value, so this actually doesn't work. (Although that's the only
+    // platform that doesn't work on.) DO NOT ATTEMPT IT. The
+    // unfortunate consequence of that is that the slots which we
+    // *know* are already type safe will wind up calling the type
+    // checker function (when we had the slots accepting
+    // BorrowedGreenlet, this was bypassed), so this slows us down.
+    // TODO: Optimize this again.
+
+    // For a class with a single pointer member, whose constructor
+    // does nothing but copy a pointer parameter into the member, and
+    // which can then be converted back to the pointer type, compilers
+    // generate code that's the same as just passing the pointer.
+    // That is, func(BorrowedObject x) called like ``PyObject* p =
+    // ...; f(p)`` has 0 overhead. Similarly, they "unpack" to the
+    // pointer type with 0 overhead.
+    //
+    // If there are no virtual functions, no complex inheritance (maybe?) and
+    // no destructor, these can be directly used as parameters in
+    // Python callbacks like tp_init: the layout is the same as a
+    // single pointer. Only subclasses with trivial constructors that
+    // do nothing but set the single pointer member are safe to use
+    // that way.
+
+
+    // This is the base class for things that can be done with a
+    // PyObject pointer. It assumes nothing about memory management.
+    // NOTE: Nothing is virtual, so subclasses shouldn't add new
+    // storage fields or try to override these methods.
+    template <typename T=PyObject, TypeChecker TC=NoOpChecker>
+    class PyObjectPointer
+    {
+    public:
+        typedef T PyType;
+    protected:
+        T* p;
+    public:
+        PyObjectPointer(T* it=nullptr) : p(it)
+        {
+            TC(p);
+        }
+
+        // We don't allow automatic casting to PyObject* at this
+        // level, because then we could be passed to Py_DECREF/INCREF,
+        // but we want nothing to do with memory management. If you
+        // know better, then you can use the get() method, like on a
+        // std::shared_ptr. Except we name it borrow() to clarify that
+        // if this is a reference-tracked object, the pointer you get
+        // back will go away when the object does.
+        // TODO: This should probably not exist here, but be moved
+        // down to relevant sub-types.
+
+        T* borrow() const noexcept
+        {
+            return this->p;
+        }
+
+        PyObject* borrow_o() const noexcept
+        {
+            return reinterpret_cast<PyObject*>(this->p);
+        }
+
+         T* operator->() const noexcept
+        {
+            return this->p;
+        }
+
+        bool is_None() const noexcept
+        {
+            return this->p == Py_None;
+        }
+
+        PyObject* acquire_or_None() const noexcept
+        {
+            PyObject* result = this->p ? reinterpret_cast<PyObject*>(this->p) : Py_None;
+            Py_INCREF(result);
+            return result;
+        }
+
+        explicit operator bool() const noexcept
+        {
+            return this->p != nullptr;
+        }
+
+        bool operator!() const noexcept
+        {
+            return this->p == nullptr;
+        }
+
+        Py_ssize_t REFCNT() const noexcept
+        {
+            return p ? Py_REFCNT(p) : -42;
+        }
+
+        PyTypeObject* TYPE() const noexcept
+        {
+            return p ? Py_TYPE(p) : nullptr;
+        }
+
+        inline OwnedObject PyStr() const noexcept;
+        inline const std::string as_str() const noexcept;
+        inline OwnedObject PyGetAttr(const ImmortalObject& name) const noexcept;
+        inline OwnedObject PyRequireAttr(const char* const name) const;
+        inline OwnedObject PyRequireAttr(const ImmortalString& name) const;
+        inline OwnedObject PyCall(const BorrowedObject& arg) const;
+        inline OwnedObject PyCall(PyGreenlet* arg) const ;
+        inline OwnedObject PyCall(PyObject* arg) const ;
+        // PyObject_Call(this, args, kwargs);
+        inline OwnedObject PyCall(const BorrowedObject args,
+                                  const BorrowedObject kwargs) const;
+        inline OwnedObject PyCall(const OwnedObject& args,
+                                  const OwnedObject& kwargs) const;
+
+    protected:
+        void _set_raw_pointer(void* t)
+        {
+            TC(t);
+            p = reinterpret_cast<T*>(t);
+        }
+        void* _get_raw_pointer() const
+        {
+            return p;
+        }
+    };
+
+#ifdef GREENLET_USE_STDIO
+        template<typename T, TypeChecker TC>
+        std::ostream& operator<<(std::ostream& os, const PyObjectPointer<T, TC>& s)
+        {
+            const std::type_info& t = typeid(s);
+            os << t.name()
+               << "(addr=" << s.borrow()
+               << ", refcnt=" << s.REFCNT()
+               << ", value=" << s.as_str()
+               << ")";
+
+            return os;
+        }
+#endif
+
+    template<typename T, TypeChecker TC>
+    inline bool operator==(const PyObjectPointer<T, TC>& lhs, const PyObject* const rhs) noexcept
+    {
+        return static_cast<const void*>(lhs.borrow_o()) == static_cast<const void*>(rhs);
+    }
+
+    template<typename T, TypeChecker TC, typename X, TypeChecker XC>
+    inline bool operator==(const PyObjectPointer<T, TC>& lhs, const PyObjectPointer<X, XC>& rhs) noexcept
+    {
+        return lhs.borrow_o() == rhs.borrow_o();
+    }
+
+    template<typename T, TypeChecker TC, typename X, TypeChecker XC>
+    inline bool operator!=(const PyObjectPointer<T, TC>& lhs,
+                           const PyObjectPointer<X, XC>& rhs) noexcept
+    {
+        return lhs.borrow_o() != rhs.borrow_o();
+    }
+
+    template<typename T=PyObject, TypeChecker TC=NoOpChecker>
+    class OwnedReference : public PyObjectPointer<T, TC>
+    {
+    private:
+        friend class OwnedList;
+
+    protected:
+        explicit OwnedReference(T* it) : PyObjectPointer<T, TC>(it)
+        {
+        }
+
+    public:
+
+        // Constructors
+
+        static OwnedReference<T, TC> consuming(PyObject* p)
+        {
+            return OwnedReference<T, TC>(reinterpret_cast<T*>(p));
+        }
+
+        static OwnedReference<T, TC> owning(T* p)
+        {
+            OwnedReference<T, TC> result(p);
+            Py_XINCREF(result.p);
+            return result;
+        }
+
+        OwnedReference() : PyObjectPointer<T, TC>(nullptr)
+        {}
+
+        explicit OwnedReference(const PyObjectPointer<>& other)
+            : PyObjectPointer<T, TC>(nullptr)
+        {
+            T* op = other.borrow();
+            TC(op);
+            this->p = other.borrow();
+            Py_XINCREF(this->p);
+        }
+
+        // It would be good to make use of the C++11 distinction
+        // between move and copy operations, e.g., constructing from a
+        // pointer should be a move operation.
+        // In the common case of ``OwnedObject x = Py_SomeFunction()``,
+        // the call to the copy constructor will be elided completely.
+        OwnedReference(const OwnedReference<T, TC>& other)
+            : PyObjectPointer<T, TC>(other.p)
+        {
+            Py_XINCREF(this->p);
+        }
+
+        static OwnedReference<PyObject> None()
+        {
+            Py_INCREF(Py_None);
+            return OwnedReference<PyObject>(Py_None);
+        }
+
+        // We can assign from exactly our type without any extra checking
+        OwnedReference<T, TC>& operator=(const OwnedReference<T, TC>& other)
+        {
+            Py_XINCREF(other.p);
+            const T* tmp = this->p;
+            this->p = other.p;
+            Py_XDECREF(tmp);
+            return *this;
+        }
+
+        OwnedReference<T, TC>& operator=(const BorrowedReference<T, TC> other)
+        {
+            return this->operator=(other.borrow());
+        }
+
+        OwnedReference<T, TC>& operator=(T* const other)
+        {
+            TC(other);
+            Py_XINCREF(other);
+            T* tmp = this->p;
+            this->p = other;
+            Py_XDECREF(tmp);
+            return *this;
+        }
+
+        // We can assign from an arbitrary reference type
+        // if it passes our check.
+        template<typename X, TypeChecker XC>
+        OwnedReference<T, TC>& operator=(const OwnedReference<X, XC>& other)
+        {
+            X* op = other.borrow();
+            TC(op);
+            return this->operator=(reinterpret_cast<T*>(op));
+        }
+
+        inline void steal(T* other)
+        {
+            assert(this->p == nullptr);
+            TC(other);
+            this->p = other;
+        }
+
+        T* relinquish_ownership()
+        {
+            T* result = this->p;
+            this->p = nullptr;
+            return result;
+        }
+
+        T* acquire() const
+        {
+            // Return a new reference.
+            // TODO: This may go away when we have reference objects
+            // throughout the code.
+            Py_XINCREF(this->p);
+            return this->p;
+        }
+
+        // Nothing else declares a destructor, we're the leaf, so we
+        // should be able to get away without virtual.
+        ~OwnedReference()
+        {
+            Py_CLEAR(this->p);
+        }
+
+        void CLEAR()
+        {
+            Py_CLEAR(this->p);
+            assert(this->p == nullptr);
+        }
+    };
+
+    static inline
+    void operator<<=(PyObject*& target, OwnedObject& o)
+    {
+        target = o.relinquish_ownership();
+    }
+
+
+    class NewReference : public OwnedObject
+    {
+    private:
+        G_NO_COPIES_OF_CLS(NewReference);
+    public:
+        // Consumes the reference. Only use this
+        // for API return values.
+        NewReference(PyObject* it) : OwnedObject(it)
+        {
+        }
+    };
+
+    class NewDictReference : public NewReference
+    {
+    private:
+        G_NO_COPIES_OF_CLS(NewDictReference);
+    public:
+        NewDictReference() : NewReference(PyDict_New())
+        {
+            if (!this->p) {
+                throw PyErrOccurred();
+            }
+        }
+
+        void SetItem(const char* const key, PyObject* value)
+        {
+            Require(PyDict_SetItemString(this->p, key, value));
+        }
+
+        void SetItem(const PyObjectPointer<>& key, PyObject* value)
+        {
+            Require(PyDict_SetItem(this->p, key.borrow_o(), value));
+        }
+    };
+
+    template<typename T=PyGreenlet, TypeChecker TC=GreenletChecker>
+    class _OwnedGreenlet: public OwnedReference<T, TC>
+    {
+    private:
+    protected:
+        _OwnedGreenlet(T* it) : OwnedReference<T, TC>(it)
+        {}
+
+    public:
+        _OwnedGreenlet() : OwnedReference<T, TC>()
+        {}
+
+        _OwnedGreenlet(const _OwnedGreenlet<T, TC>& other) : OwnedReference<T, TC>(other)
+        {
+        }
+        _OwnedGreenlet(OwnedMainGreenlet& other) :
+            OwnedReference<T, TC>(reinterpret_cast<T*>(other.acquire()))
+        {
+        }
+        _OwnedGreenlet(const BorrowedGreenlet& other);
+        // Steals a reference.
+        static _OwnedGreenlet<T, TC> consuming(PyGreenlet* it)
+        {
+            return _OwnedGreenlet<T, TC>(reinterpret_cast<T*>(it));
+        }
+
+        inline _OwnedGreenlet<T, TC>& operator=(const OwnedGreenlet& other)
+        {
+            return this->operator=(other.borrow());
+        }
+
+        inline _OwnedGreenlet<T, TC>& operator=(const BorrowedGreenlet& other);
+
+        _OwnedGreenlet<T, TC>& operator=(const OwnedMainGreenlet& other)
+        {
+            PyGreenlet* owned = other.acquire();
+            Py_XDECREF(this->p);
+            this->p = reinterpret_cast<T*>(owned);
+            return *this;
+        }
+
+        _OwnedGreenlet<T, TC>& operator=(T* const other)
+        {
+            OwnedReference<T, TC>::operator=(other);
+            return *this;
+        }
+
+        T* relinquish_ownership()
+        {
+            T* result = this->p;
+            this->p = nullptr;
+            return result;
+        }
+
+        PyObject* relinquish_ownership_o()
+        {
+            return reinterpret_cast<PyObject*>(relinquish_ownership());
+        }
+
+        inline Greenlet* operator->() const noexcept;
+        inline operator Greenlet*() const noexcept;
+    };
+
+    template <typename T=PyObject, TypeChecker TC=NoOpChecker>
+    class BorrowedReference : public PyObjectPointer<T, TC>
+    {
+    public:
+        // Allow implicit creation from PyObject* pointers as we
+        // transition to using these classes. Also allow automatic
+        // conversion to PyObject* for passing to C API calls and even
+        // for Py_INCREF/DECREF, because we ourselves do no memory management.
+        BorrowedReference(T* it) : PyObjectPointer<T, TC>(it)
+        {}
+
+        BorrowedReference(const PyObjectPointer<T>& ref) : PyObjectPointer<T, TC>(ref.borrow())
+        {}
+
+        BorrowedReference() : PyObjectPointer<T, TC>(nullptr)
+        {}
+
+        operator T*() const
+        {
+            return this->p;
+        }
+    };
+
+    typedef BorrowedReference<PyObject> BorrowedObject;
+    //typedef BorrowedReference<PyGreenlet> BorrowedGreenlet;
+
+    template<typename T=PyGreenlet, TypeChecker TC=GreenletChecker>
+    class _BorrowedGreenlet : public BorrowedReference<T, TC>
+    {
+    public:
+        _BorrowedGreenlet() :
+            BorrowedReference<T, TC>(nullptr)
+        {}
+
+        _BorrowedGreenlet(T* it) :
+            BorrowedReference<T, TC>(it)
+        {}
+
+        _BorrowedGreenlet(const BorrowedObject& it);
+
+        _BorrowedGreenlet(const OwnedGreenlet& it) :
+            BorrowedReference<T, TC>(it.borrow())
+        {}
+
+        _BorrowedGreenlet<T, TC>& operator=(const BorrowedObject& other);
+
+        // We get one of these for PyGreenlet, but one for PyObject
+        // is handy as well
+        operator PyObject*() const
+        {
+            return reinterpret_cast<PyObject*>(this->p);
+        }
+        Greenlet* operator->() const noexcept;
+        operator Greenlet*() const noexcept;
+    };
+
+    typedef _BorrowedGreenlet<PyGreenlet> BorrowedGreenlet;
+
+    template<typename T, TypeChecker TC>
+    _OwnedGreenlet<T, TC>::_OwnedGreenlet(const BorrowedGreenlet& other)
+        : OwnedReference<T, TC>(reinterpret_cast<T*>(other.borrow()))
+    {
+        Py_XINCREF(this->p);
+    }
+
+
+     class BorrowedMainGreenlet
+            : public _BorrowedGreenlet<PyGreenlet, MainGreenletExactChecker>
+    {
+    public:
+        BorrowedMainGreenlet(const OwnedMainGreenlet& it) :
+            _BorrowedGreenlet<PyGreenlet, MainGreenletExactChecker>(it.borrow())
+        {}
+        BorrowedMainGreenlet(PyGreenlet* it=nullptr)
+            : _BorrowedGreenlet<PyGreenlet, MainGreenletExactChecker>(it)
+        {}
+    };
+
+    template<typename T, TypeChecker TC>
+    _OwnedGreenlet<T, TC>& _OwnedGreenlet<T, TC>::operator=(const BorrowedGreenlet& other)
+    {
+        return this->operator=(other.borrow());
+    }
+
+
+    class ImmortalObject : public PyObjectPointer<>
+    {
+    private:
+        G_NO_ASSIGNMENT_OF_CLS(ImmortalObject);
+    public:
+        explicit ImmortalObject(PyObject* it) : PyObjectPointer<>(it)
+        {
+        }
+
+        ImmortalObject(const ImmortalObject& other)
+            : PyObjectPointer<>(other.p)
+        {
+
+        }
+
+        /**
+         * Become the new owner of the object. Does not change the
+         * reference count.
+         */
+        ImmortalObject& operator=(PyObject* it)
+        {
+            assert(this->p == nullptr);
+            this->p = it;
+            return *this;
+        }
+
+        static ImmortalObject consuming(PyObject* it)
+        {
+            return ImmortalObject(it);
+        }
+
+        inline operator PyObject*() const
+        {
+            return this->p;
+        }
+    };
+
+    class ImmortalString : public ImmortalObject
+    {
+    private:
+        G_NO_COPIES_OF_CLS(ImmortalString);
+        const char* str;
+    public:
+        ImmortalString(const char* const str) :
+            ImmortalObject(str ? Require(PyUnicode_InternFromString(str)) : nullptr)
+        {
+            this->str = str;
+        }
+
+        inline ImmortalString& operator=(const char* const str)
+        {
+            if (!this->p) {
+                this->p = Require(PyUnicode_InternFromString(str));
+                this->str = str;
+            }
+            else {
+                assert(this->str == str);
+            }
+            return *this;
+        }
+
+        inline operator std::string() const
+        {
+            return this->str;
+        }
+
+    };
+
+    class ImmortalEventName : public ImmortalString
+    {
+    private:
+        G_NO_COPIES_OF_CLS(ImmortalEventName);
+    public:
+        ImmortalEventName(const char* const str) : ImmortalString(str)
+        {}
+    };
+
+    class ImmortalException : public ImmortalObject
+    {
+    private:
+        G_NO_COPIES_OF_CLS(ImmortalException);
+    public:
+        ImmortalException(const char* const name, PyObject* base=nullptr) :
+            ImmortalObject(name
+                           // Python 2.7 isn't const correct
+                           ? Require(PyErr_NewException((char*)name, base, nullptr))
+                           : nullptr)
+        {}
+
+        inline bool PyExceptionMatches() const
+        {
+            return PyErr_ExceptionMatches(this->p) > 0;
+        }
+
+    };
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyStr() const noexcept
+    {
+        if (!this->p) {
+            return OwnedObject();
+        }
+        return OwnedObject::consuming(PyObject_Str(reinterpret_cast<PyObject*>(this->p)));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline const std::string PyObjectPointer<T, TC>::as_str() const noexcept
+    {
+        // NOTE: This is not Python exception safe.
+        if (this->p) {
+            // The Python APIs return a cached char* value that's only valid
+            // as long as the original object stays around, and we're
+            // about to (probably) toss it. Hence the copy to std::string.
+            OwnedObject py_str = this->PyStr();
+            if (!py_str) {
+                return "(nil)";
+            }
+            return PyUnicode_AsUTF8(py_str.borrow());
+        }
+        return "(nil)";
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyGetAttr(const ImmortalObject& name) const noexcept
+    {
+        assert(this->p);
+        return OwnedObject::consuming(PyObject_GetAttr(reinterpret_cast<PyObject*>(this->p), name));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyRequireAttr(const char* const name) const
+    {
+        assert(this->p);
+        return OwnedObject::consuming(Require(PyObject_GetAttrString(this->p, name), name));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyRequireAttr(const ImmortalString& name) const
+    {
+        assert(this->p);
+        return OwnedObject::consuming(Require(
+                   PyObject_GetAttr(
+                      reinterpret_cast<PyObject*>(this->p),
+                      name
+                   ),
+                   name
+               ));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyCall(const BorrowedObject& arg) const
+    {
+        return this->PyCall(arg.borrow());
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyCall(PyGreenlet* arg) const
+    {
+        return this->PyCall(reinterpret_cast<PyObject*>(arg));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyCall(PyObject* arg) const
+    {
+        assert(this->p);
+        return OwnedObject::consuming(PyObject_CallFunctionObjArgs(this->p, arg, NULL));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyCall(const BorrowedObject args,
+                                                  const BorrowedObject kwargs) const
+    {
+        assert(this->p);
+        return OwnedObject::consuming(PyObject_Call(this->p, args, kwargs));
+    }
+
+    template<typename T, TypeChecker TC>
+    inline OwnedObject PyObjectPointer<T, TC>::PyCall(const OwnedObject& args,
+                                                  const OwnedObject& kwargs) const
+    {
+        assert(this->p);
+        return OwnedObject::consuming(PyObject_Call(this->p, args.borrow(), kwargs.borrow()));
+    }
+
+    inline void
+    ListChecker(void * p)
+    {
+        if (!p) {
+            return;
+        }
+        if (!PyList_Check(p)) {
+            throw TypeError("Expected a list");
+        }
+    }
+
+    class OwnedList : public OwnedReference<PyObject, ListChecker>
+    {
+    private:
+        G_NO_ASSIGNMENT_OF_CLS(OwnedList);
+    public:
+        // TODO: Would like to use move.
+        explicit OwnedList(const OwnedObject& other)
+            : OwnedReference<PyObject, ListChecker>(other)
+        {
+        }
+
+        OwnedList& operator=(const OwnedObject& other)
+        {
+            if (other && PyList_Check(other.p)) {
+                // Valid list. Own a new reference to it, discard the
+                // reference to what we did own.
+                PyObject* new_ptr = other.p;
+                Py_INCREF(new_ptr);
+                Py_XDECREF(this->p);
+                this->p = new_ptr;
+            }
+            else {
+                // Either the other object was NULL (an error) or it
+                // wasn't a list. Either way, we're now invalidated.
+                Py_XDECREF(this->p);
+                this->p = nullptr;
+            }
+            return *this;
+        }
+
+        inline bool empty() const
+        {
+            return PyList_GET_SIZE(p) == 0;
+        }
+
+        inline Py_ssize_t size() const
+        {
+            return PyList_GET_SIZE(p);
+        }
+
+        inline BorrowedObject at(const Py_ssize_t index) const
+        {
+            return PyList_GET_ITEM(p, index);
+        }
+
+        inline void clear()
+        {
+            PyList_SetSlice(p, 0, PyList_GET_SIZE(p), NULL);
+        }
+    };
+
+    // Use this to represent the module object used at module init
+    // time.
+    // This could either be a borrowed (Py2) or new (Py3) reference;
+    // either way, we don't want to do any memory management
+    // on it here, Python itself will handle that.
+    // XXX: Actually, that's not quite right. On Python 3, if an
+    // exception occurs before we return to the interpreter, this will
+    // leak; but all previous versions also had that problem.
+    class CreatedModule : public PyObjectPointer<>
+    {
+    private:
+        G_NO_COPIES_OF_CLS(CreatedModule);
+    public:
+        CreatedModule(PyModuleDef& mod_def) : PyObjectPointer<>(
+            Require(PyModule_Create(&mod_def)))
+        {
+        }
+
+        // PyAddObject(): Add a reference to the object to the module.
+        // On return, the reference count of the object is unchanged.
+        //
+        // The docs warn that PyModule_AddObject only steals the
+        // reference on success, so if it fails after we've incref'd
+        // or allocated, we're responsible for the decref.
+        void PyAddObject(const char* name, const long new_bool)
+        {
+            OwnedObject p = OwnedObject::consuming(Require(PyBool_FromLong(new_bool)));
+            this->PyAddObject(name, p);
+        }
+
+        void PyAddObject(const char* name, const OwnedObject& new_object)
+        {
+            // The caller already owns a reference they will decref
+            // when their variable goes out of scope, we still need to
+            // incref/decref.
+            this->PyAddObject(name, new_object.borrow());
+        }
+
+        void PyAddObject(const char* name, const ImmortalObject& new_object)
+        {
+            this->PyAddObject(name, new_object.borrow());
+        }
+
+        void PyAddObject(const char* name, PyTypeObject& type)
+        {
+            this->PyAddObject(name, reinterpret_cast<PyObject*>(&type));
+        }
+
+        void PyAddObject(const char* name, PyObject* new_object)
+        {
+            Py_INCREF(new_object);
+            try {
+                Require(PyModule_AddObject(this->p, name, new_object));
+            }
+            catch (const PyErrOccurred&) {
+                Py_DECREF(p);
+                throw;
+            }
+        }
+    };
+
+    class PyErrFetchParam : public PyObjectPointer<>
+    {
+        // Not an owned object, because we can't be initialized with
+        // one, and we only sometimes acquire ownership.
+    private:
+        G_NO_COPIES_OF_CLS(PyErrFetchParam);
+    public:
+        // To allow declaring these and passing them to
+        // PyErr_Fetch we implement the empty constructor,
+        // and the address operator.
+        PyErrFetchParam() : PyObjectPointer<>(nullptr)
+        {
+        }
+
+        PyObject** operator&()
+        {
+            return &this->p;
+        }
+
+        // This allows us to pass one directly without the &,
+        // BUT it has higher precedence than the bool operator
+        // if it's not explicit.
+        operator PyObject**()
+        {
+            return &this->p;
+        }
+
+        // We don't want to be able to pass these to Py_DECREF and
+        // such so we don't have the implicit PyObject* conversion.
+
+        inline PyObject* relinquish_ownership()
+        {
+            PyObject* result = this->p;
+            this->p = nullptr;
+            return result;
+        }
+
+        ~PyErrFetchParam()
+        {
+            Py_XDECREF(p);
+        }
+    };
+
+    class OwnedErrPiece : public OwnedObject
+    {
+    private:
+
+    public:
+        // Unlike OwnedObject, this increments the refcount.
+        OwnedErrPiece(PyObject* p=nullptr) : OwnedObject(p)
+        {
+            this->acquire();
+        }
+
+        PyObject** operator&()
+        {
+            return &this->p;
+        }
+
+        inline operator PyObject*() const
+        {
+            return this->p;
+        }
+
+        operator PyTypeObject*() const
+        {
+            return reinterpret_cast<PyTypeObject*>(this->p);
+        }
+    };
+
+    class PyErrPieces
+    {
+    private:
+        OwnedErrPiece type;
+        OwnedErrPiece instance;
+        OwnedErrPiece traceback;
+        bool restored;
+    public:
+        // Takes new references; if we're destroyed before
+        // restoring the error, we drop the references.
+        PyErrPieces(PyObject* t, PyObject* v, PyObject* tb) :
+            type(t),
+            instance(v),
+            traceback(tb),
+            restored(0)
+        {
+            this->normalize();
+        }
+
+        PyErrPieces() :
+            restored(0)
+        {
+            // PyErr_Fetch transfers ownership to us, so
+            // we don't actually need to INCREF; but we *do*
+            // need to DECREF if we're not restored.
+            PyErrFetchParam t, v, tb;
+            PyErr_Fetch(&t, &v, &tb);
+            type.steal(t.relinquish_ownership());
+            instance.steal(v.relinquish_ownership());
+            traceback.steal(tb.relinquish_ownership());
+        }
+
+        void PyErrRestore()
+        {
+            // can only do this once
+            assert(!this->restored);
+            this->restored = true;
+            PyErr_Restore(
+                this->type.relinquish_ownership(),
+                this->instance.relinquish_ownership(),
+                this->traceback.relinquish_ownership());
+            assert(!this->type && !this->instance && !this->traceback);
+        }
+
+    private:
+        void normalize()
+        {
+            // First, check the traceback argument, replacing None,
+            // with NULL
+            if (traceback.is_None()) {
+                traceback = nullptr;
+            }
+
+            if (traceback && !PyTraceBack_Check(traceback.borrow())) {
+                throw PyErrOccurred(PyExc_TypeError,
+                                    "throw() third argument must be a traceback object");
+            }
+
+            if (PyExceptionClass_Check(type)) {
+                // If we just had a type, we'll now have a type and
+                // instance.
+                // The type's refcount will have gone up by one
+                // because of the instance and the instance will have
+                // a refcount of one. Either way, we owned, and still
+                // do own, exactly one reference.
+                PyErr_NormalizeException(&type, &instance, &traceback);
+
+            }
+            else if (PyExceptionInstance_Check(type)) {
+                /* Raising an instance --- usually that means an
+                   object that is a subclass of BaseException, but on
+                   Python 2, that can also mean an arbitrary old-style
+                   object. The value should be a dummy. */
+                if (instance && !instance.is_None()) {
+                    throw PyErrOccurred(
+                                    PyExc_TypeError,
+                                    "instance exception may not have a separate value");
+                }
+                /* Normalize to raise <class>, <instance> */
+                this->instance = this->type;
+                this->type = PyExceptionInstance_Class(instance.borrow());
+
+                /*
+                  It would be tempting to do this:
+
+                Py_ssize_t type_count = Py_REFCNT(Py_TYPE(instance.borrow()));
+                this->type = PyExceptionInstance_Class(instance.borrow());
+                assert(this->type.REFCNT() == type_count + 1);
+
+                But that doesn't work on Python 2 in the case of
+                old-style instances: The result of Py_TYPE is going to
+                be the global shared <type instance> that all
+                old-style classes have, while the return of Instance_Class()
+                will be the Python-level class object. The two are unrelated.
+                */
+            }
+            else {
+                /* Not something you can raise. throw() fails. */
+                PyErr_Format(PyExc_TypeError,
+                     "exceptions must be classes, or instances, not %s",
+                             Py_TYPE(type.borrow())->tp_name);
+                throw PyErrOccurred();
+            }
+        }
+    };
+
+    // PyArg_Parse's O argument returns a borrowed reference.
+    class PyArgParseParam : public BorrowedObject
+    {
+    private:
+        G_NO_COPIES_OF_CLS(PyArgParseParam);
+    public:
+        explicit PyArgParseParam(PyObject* p=nullptr) : BorrowedObject(p)
+        {
+        }
+
+        inline PyObject** operator&()
+        {
+            return &this->p;
+        }
+    };
+
+};};
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_slp_switch.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_slp_switch.hpp
new file mode 100644
index 00000000..bd4b7ae1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_slp_switch.hpp
@@ -0,0 +1,99 @@
+#ifndef GREENLET_SLP_SWITCH_HPP
+#define GREENLET_SLP_SWITCH_HPP
+
+#include "greenlet_compiler_compat.hpp"
+#include "greenlet_refs.hpp"
+
+/*
+ * the following macros are spliced into the OS/compiler
+ * specific code, in order to simplify maintenance.
+ */
+// We can save about 10% of the time it takes to switch greenlets if
+// we thread the thread state through the slp_save_state() and the
+// following slp_restore_state() calls from
+// slp_switch()->g_switchstack() (which already needs to access it).
+//
+// However:
+//
+// that requires changing the prototypes and implementations of the
+// switching functions. If we just change the prototype of
+// slp_switch() to accept the argument and update the macros, without
+// changing the implementation of slp_switch(), we get crashes on
+// 64-bit Linux and 32-bit x86 (for reasons that aren't 100% clear);
+// on the other hand, 64-bit macOS seems to be fine. Also, 64-bit
+// windows is an issue because slp_switch is written fully in assembly
+// and currently ignores its argument so some code would have to be
+// adjusted there to pass the argument on to the
+// ``slp_save_state_asm()`` function (but interestingly, because of
+// the calling convention, the extra argument is just ignored and
+// things function fine, albeit slower, if we just modify
+// ``slp_save_state_asm`()` to fetch the pointer to pass to the
+// macro.)
+//
+// Our compromise is to use a *glabal*, untracked, weak, pointer
+// to the necessary thread state during the process of switching only.
+// This is safe because we're protected by the GIL, and if we're
+// running this code, the thread isn't exiting. This also nets us a
+// 10-12% speed improvement.
+
+static greenlet::Greenlet* volatile switching_thread_state = nullptr;
+
+
+extern "C" {
+static int GREENLET_NOINLINE(slp_save_state_trampoline)(char* stackref);
+static void GREENLET_NOINLINE(slp_restore_state_trampoline)();
+}
+
+
+#define SLP_SAVE_STATE(stackref, stsizediff) \
+do {                                                    \
+    assert(switching_thread_state);  \
+    stackref += STACK_MAGIC;                 \
+    if (slp_save_state_trampoline((char*)stackref))    \
+        return -1;                                     \
+    if (!switching_thread_state->active()) \
+        return 1;                                      \
+    stsizediff = switching_thread_state->stack_start() - (char*)stackref; \
+} while (0)
+
+#define SLP_RESTORE_STATE() slp_restore_state_trampoline()
+
+#define SLP_EVAL
+extern "C" {
+#define slp_switch GREENLET_NOINLINE(slp_switch)
+#include "slp_platformselect.h"
+}
+#undef slp_switch
+
+#ifndef STACK_MAGIC
+#    error \
+        "greenlet needs to be ported to this platform, or taught how to detect your compiler properly."
+#endif /* !STACK_MAGIC */
+
+
+
+#ifdef EXTERNAL_ASM
+/* CCP addition: Make these functions, to be called from assembler.
+ * The token include file for the given platform should enable the
+ * EXTERNAL_ASM define so that this is included.
+ */
+extern "C" {
+intptr_t
+slp_save_state_asm(intptr_t* ref)
+{
+    intptr_t diff;
+    SLP_SAVE_STATE(ref, diff);
+    return diff;
+}
+
+void
+slp_restore_state_asm(void)
+{
+    SLP_RESTORE_STATE();
+}
+
+extern int slp_switch(void);
+};
+#endif
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_thread_support.hpp b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_thread_support.hpp
new file mode 100644
index 00000000..3ded7d2b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/greenlet_thread_support.hpp
@@ -0,0 +1,31 @@
+#ifndef GREENLET_THREAD_SUPPORT_HPP
+#define GREENLET_THREAD_SUPPORT_HPP
+
+/**
+ * Defines various utility functions to help greenlet integrate well
+ * with threads. This used to be needed when we supported Python
+ * 2.7 on Windows, which used a very old compiler. We wrote an
+ * alternative implementation using Python APIs and POSIX or Windows
+ * APIs, but that's no longer needed. So this file is a shadow of its
+ * former self --- but may be needed in the future.
+ */
+
+#include <stdexcept>
+#include <thread>
+#include <mutex>
+
+#include "greenlet_compiler_compat.hpp"
+
+namespace greenlet {
+    typedef std::mutex Mutex;
+    typedef std::lock_guard<Mutex> LockGuard;
+    class LockInitError : public std::runtime_error
+    {
+    public:
+        LockInitError(const char* what) : std::runtime_error(what)
+        {};
+    };
+};
+
+
+#endif /* GREENLET_THREAD_SUPPORT_HPP */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/__init__.py b/backend/venv/lib/python3.12/site-packages/greenlet/platform/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/platform/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..37bc30ec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/platform/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/setup_switch_x64_masm.cmd b/backend/venv/lib/python3.12/site-packages/greenlet/platform/setup_switch_x64_masm.cmd
new file mode 100644
index 00000000..09285955
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/setup_switch_x64_masm.cmd
@@ -0,0 +1,2 @@
+call "C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat" amd64
+ml64 /nologo /c /Fo switch_x64_masm.obj switch_x64_masm.asm
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_aarch64_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_aarch64_gcc.h
new file mode 100644
index 00000000..058617c4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_aarch64_gcc.h
@@ -0,0 +1,124 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 07-Sep-16 Add clang support using x register naming. Fredrik Fornwall
+ * 13-Apr-13 Add support for strange GCC caller-save decisions
+ * 08-Apr-13 File creation. Michael Matz
+ *
+ * NOTES
+ *
+ * Simply save all callee saved registers
+ *
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+#define REGS_TO_SAVE "x19", "x20", "x21", "x22", "x23", "x24", "x25", "x26", \
+                     "x27", "x28", "x30" /* aka lr */, \
+                     "v8", "v9", "v10", "v11", \
+                     "v12", "v13", "v14", "v15"
+
+/*
+ * Recall:
+   asm asm-qualifiers ( AssemblerTemplate
+                 : OutputOperands
+                 [ : InputOperands
+                 [ : Clobbers ] ])
+
+ or  (if asm-qualifiers contains 'goto')
+
+   asm asm-qualifiers ( AssemblerTemplate
+                      : OutputOperands
+                      : InputOperands
+                      : Clobbers
+                      : GotoLabels)
+
+ and OutputOperands are
+
+   [ [asmSymbolicName] ] constraint (cvariablename)
+
+ When a name is given, refer to it as ``%[the name]``.
+ When not given, ``%i`` where ``i`` is the zero-based index.
+
+ constraints starting with ``=`` means only writing; ``+`` means
+ reading and writing.
+
+ This is followed by ``r`` (must be register) or ``m`` (must be memory)
+ and these can be combined.
+
+ The ``cvariablename`` is actually an lvalue expression.
+
+ In AArch65, 31 general purpose registers. If named X0... they are
+ 64-bit. If named W0... they are the bottom 32 bits of the
+ corresponding 64 bit register.
+
+ XZR and WZR are hardcoded to 0, and ignore writes.
+
+ Arguments are in X0..X7. C++ uses X0 for ``this``. X0 holds simple return
+ values (?)
+
+ Whenever a W register is written, the top half of the X register is zeroed.
+ */
+
+static int
+slp_switch(void)
+{
+	int err;
+	void *fp;
+        /* Windowz uses a 32-bit long on a 64-bit platform, unlike the rest of
+           the world, and in theory we can be compiled with GCC/llvm on 64-bit
+           windows. So we need a fixed-width type.
+        */
+        int64_t *stackref, stsizediff;
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+	__asm__ volatile ("str x29, %0" : "=m"(fp) : : );
+        __asm__ ("mov %0, sp" : "=r" (stackref));
+        {
+                SLP_SAVE_STATE(stackref, stsizediff);
+                __asm__ volatile (
+                    "add sp,sp,%0\n"
+                    "add x29,x29,%0\n"
+                    :
+                    : "r" (stsizediff)
+                    );
+		SLP_RESTORE_STATE();
+		/* SLP_SAVE_STATE macro contains some return statements
+		   (of -1 and 1).  It falls through only when
+		   the return value of slp_save_state() is zero, which
+		   is placed in x0.
+		   In that case we (slp_switch) also want to return zero
+		   (also in x0 of course).
+		   Now, some GCC versions (seen with 4.8) think it's a
+		   good idea to save/restore x0 around the call to
+		   slp_restore_state(), instead of simply zeroing it
+		   at the return below.  But slp_restore_state
+		   writes random values to the stack slot used for this
+		   save/restore (from when it once was saved above in
+		   SLP_SAVE_STATE, when it was still uninitialized), so
+		   "restoring" that precious zero actually makes us
+		   return random values.  There are some ways to make
+		   GCC not use that zero value in the normal return path
+		   (e.g. making err volatile, but that costs a little
+		   stack space), and the simplest is to call a function
+		   that returns an unknown value (which happens to be zero),
+		   so the saved/restored value is unused.
+
+                   Thus, this line stores a 0 into the ``err`` variable
+                   (which must be held in a register for this instruction,
+                    of course). The ``w`` qualifier causes the instruction
+                    to use W0 instead of X0, otherwise we get a warning
+                    about a value size mismatch (because err is an int,
+                    and aarch64 platforms are LP64: 32-bit int, 64 bit long
+                   and pointer).
+                */
+           __asm__ volatile ("mov %w0, #0" : "=r" (err));
+        }
+        __asm__ volatile ("ldr x29, %0" : : "m" (fp) :);
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+        return err;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_alpha_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_alpha_unix.h
new file mode 100644
index 00000000..7e07abfc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_alpha_unix.h
@@ -0,0 +1,30 @@
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "$9", "$10", "$11", "$12", "$13", "$14", "$15", \
+		     "$f2", "$f3", "$f4", "$f5", "$f6", "$f7", "$f8", "$f9"
+
+static int
+slp_switch(void)
+{
+  int ret;
+  long *stackref, stsizediff;
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  __asm__ volatile ("mov $30, %0" : "=r" (stackref) : );
+  {
+      SLP_SAVE_STATE(stackref, stsizediff);
+      __asm__ volatile (
+	  "addq $30, %0, $30\n\t"
+	  : /* no outputs */
+	  : "r" (stsizediff)
+	  );
+      SLP_RESTORE_STATE();
+  }
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  __asm__ volatile ("mov $31, %0" : "=r" (ret) : );
+  return ret;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_amd64_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_amd64_unix.h
new file mode 100644
index 00000000..d4701105
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_amd64_unix.h
@@ -0,0 +1,87 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 3-May-13   Ralf Schmitt  <ralf@systemexit.de>
+ *     Add support for strange GCC caller-save decisions
+ *     (ported from switch_aarch64_gcc.h)
+ * 18-Aug-11  Alexey Borzenkov  <snaury@gmail.com>
+ *      Correctly save rbp, csr and cw
+ * 01-Apr-04  Hye-Shik Chang    <perky@FreeBSD.org>
+ *      Ported from i386 to amd64.
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for spark
+ * 31-Avr-02  Armin Rigo         <arigo@ulb.ac.be>
+ *      Added ebx, esi and edi register-saves.
+ * 01-Mar-02  Samual M. Rushing  <rushing@ironport.com>
+ *      Ported from i386.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+/* #define STACK_MAGIC 3 */
+/* the above works fine with gcc 2.96, but 2.95.3 wants this */
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "r12", "r13", "r14", "r15"
+
+static int
+slp_switch(void)
+{
+    int err;
+    void* rbp;
+    void* rbx;
+    unsigned int csr;
+    unsigned short cw;
+    /* This used to be declared 'register', but that does nothing in
+    modern compilers and is explicitly forbidden in some new
+    standards. */
+    long *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("fstcw %0" : "=m" (cw));
+    __asm__ volatile ("stmxcsr %0" : "=m" (csr));
+    __asm__ volatile ("movq %%rbp, %0" : "=m" (rbp));
+    __asm__ volatile ("movq %%rbx, %0" : "=m" (rbx));
+    __asm__ ("movq %%rsp, %0" : "=g" (stackref));
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "addq %0, %%rsp\n"
+            "addq %0, %%rbp\n"
+            :
+            : "r" (stsizediff)
+            );
+        SLP_RESTORE_STATE();
+        __asm__ volatile ("xorq %%rax, %%rax" : "=a" (err));
+    }
+    __asm__ volatile ("movq %0, %%rbx" : : "m" (rbx));
+    __asm__ volatile ("movq %0, %%rbp" : : "m" (rbp));
+    __asm__ volatile ("ldmxcsr %0" : : "m" (csr));
+    __asm__ volatile ("fldcw %0" : : "m" (cw));
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_gcc.h
new file mode 100644
index 00000000..655003aa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_gcc.h
@@ -0,0 +1,79 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 14-Aug-06 File creation. Ported from Arm Thumb. Sylvain Baro
+ *  3-Sep-06 Commented out saving of r1-r3 (r4 already commented out) as I
+ *           read that these do not need to be saved.  Also added notes and
+ *           errors related to the frame pointer. Richard Tew.
+ *
+ * NOTES
+ *
+ *   It is not possible to detect if fp is used or not, so the supplied
+ *   switch function needs to support it, so that you can remove it if
+ *   it does not apply to you.
+ *
+ * POSSIBLE ERRORS
+ *
+ *   "fp cannot be used in asm here"
+ *
+ *   - Try commenting out "fp" in REGS_TO_SAVE.
+ *
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+#define REG_SP "sp"
+#define REG_SPSP "sp,sp"
+#ifdef __thumb__
+#define REG_FP "r7"
+#define REG_FPFP "r7,r7"
+#define REGS_TO_SAVE_GENERAL "r4", "r5", "r6", "r8", "r9", "r10", "r11", "lr"
+#else
+#define REG_FP "fp"
+#define REG_FPFP "fp,fp"
+#define REGS_TO_SAVE_GENERAL "r4", "r5", "r6", "r7", "r8", "r9", "r10", "lr"
+#endif
+#if defined(__SOFTFP__)
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL
+#elif defined(__VFP_FP__)
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL, "d8", "d9", "d10", "d11", \
+                                           "d12", "d13", "d14", "d15"
+#elif defined(__MAVERICK__)
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL, "mvf4", "mvf5", "mvf6", "mvf7", \
+                                           "mvf8", "mvf9", "mvf10", "mvf11", \
+                                           "mvf12", "mvf13", "mvf14", "mvf15"
+#else
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL, "f4", "f5", "f6", "f7"
+#endif
+
+static int
+#ifdef __GNUC__
+__attribute__((optimize("no-omit-frame-pointer")))
+#endif
+slp_switch(void)
+{
+        void *fp;
+        int *stackref, stsizediff;
+        int result;
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+        __asm__ volatile ("mov r0," REG_FP "\n\tstr r0,%0" : "=m" (fp) : : "r0");
+        __asm__ ("mov %0," REG_SP : "=r" (stackref));
+        {
+                SLP_SAVE_STATE(stackref, stsizediff);
+                __asm__ volatile (
+                    "add " REG_SPSP ",%0\n"
+                    "add " REG_FPFP ",%0\n"
+                    :
+                    : "r" (stsizediff)
+                    );
+                SLP_RESTORE_STATE();
+        }
+        __asm__ volatile ("ldr r0,%1\n\tmov " REG_FP ",r0\n\tmov %0, #0" : "=r" (result) : "m" (fp) : "r0");
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+        return result;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_ios.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_ios.h
new file mode 100644
index 00000000..9e640e15
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm32_ios.h
@@ -0,0 +1,67 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 31-May-15 iOS support. Ported from arm32. Proton <feisuzhu@163.com>
+ *
+ * NOTES
+ *
+ *   It is not possible to detect if fp is used or not, so the supplied
+ *   switch function needs to support it, so that you can remove it if
+ *   it does not apply to you.
+ *
+ * POSSIBLE ERRORS
+ *
+ *   "fp cannot be used in asm here"
+ *
+ *   - Try commenting out "fp" in REGS_TO_SAVE.
+ *
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 0
+#define REG_SP "sp"
+#define REG_SPSP "sp,sp"
+#define REG_FP "r7"
+#define REG_FPFP "r7,r7"
+#define REGS_TO_SAVE_GENERAL "r4", "r5", "r6", "r8", "r10", "r11", "lr"
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL, "d8", "d9", "d10", "d11", \
+                                           "d12", "d13", "d14", "d15"
+
+static int
+#ifdef __GNUC__
+__attribute__((optimize("no-omit-frame-pointer")))
+#endif
+slp_switch(void)
+{
+        void *fp;
+        int *stackref, stsizediff, result;
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+        __asm__ volatile ("str " REG_FP ",%0" : "=m" (fp));
+        __asm__ ("mov %0," REG_SP : "=r" (stackref));
+        {
+                SLP_SAVE_STATE(stackref, stsizediff);
+                __asm__ volatile (
+                    "add " REG_SPSP ",%0\n"
+                    "add " REG_FPFP ",%0\n"
+                    :
+                    : "r" (stsizediff)
+                    : REGS_TO_SAVE /* Clobber registers, force compiler to
+                                    * recalculate address of void *fp from REG_SP or REG_FP */
+                );
+                SLP_RESTORE_STATE();
+        }
+        __asm__ volatile (
+            "ldr " REG_FP ", %1\n\t"
+            "mov %0, #0"
+            : "=r" (result)
+            : "m" (fp)
+            : REGS_TO_SAVE /* Force compiler to restore saved registers after this */
+        );
+        return result;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.asm b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.asm
new file mode 100644
index 00000000..29f9c225
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.asm
@@ -0,0 +1,53 @@
+  AREA switch_arm64_masm, CODE, READONLY;
+  GLOBAL slp_switch [FUNC]
+  EXTERN slp_save_state_asm
+  EXTERN slp_restore_state_asm
+
+slp_switch    
+    ; push callee saved registers to stack
+    stp    x19, x20, [sp, #-16]!
+    stp    x21, x22, [sp, #-16]!
+    stp    x23, x24, [sp, #-16]!
+    stp    x25, x26, [sp, #-16]!
+    stp    x27, x28, [sp, #-16]!
+    stp    x29, x30, [sp, #-16]!
+    stp    d8, d9, [sp, #-16]!
+    stp    d10, d11, [sp, #-16]!
+    stp    d12, d13, [sp, #-16]!
+    stp    d14, d15, [sp, #-16]!
+
+    ; call slp_save_state_asm with stack pointer
+    mov x0, sp
+    bl    slp_save_state_asm
+
+    ; early return for return value of 1 and -1
+    cmp x0, #-1
+    b.eq RETURN
+    cmp x0, #1
+    b.eq RETURN
+
+    ; increment stack and frame pointer
+    add sp, sp, x0
+    add x29, x29, x0
+
+    bl slp_restore_state_asm
+
+    ; store return value for successful completion of routine
+    mov x0, #0
+
+RETURN
+    ; pop registers from stack
+    ldp d14, d15, [sp], #16
+    ldp d12, d13, [sp], #16
+    ldp d10, d11, [sp], #16
+    ldp d8, d9, [sp], #16
+    ldp x29, x30, [sp], #16
+    ldp x27, x28, [sp], #16
+    ldp x25, x26, [sp], #16
+    ldp x23, x24, [sp], #16
+    ldp x21, x22, [sp], #16
+    ldp x19, x20, [sp], #16
+
+    ret
+
+    END
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.obj b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.obj
new file mode 100644
index 00000000..f6f220e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_masm.obj differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_msvc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_msvc.h
new file mode 100644
index 00000000..7ab7f45b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_arm64_msvc.h
@@ -0,0 +1,17 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 21-Oct-21  Niyas Sait  <niyas.sait@linaro.org>
+ *      First version to enable win/arm64 support.
+ */
+
+#define STACK_REFPLUS 1
+#define STACK_MAGIC 0
+
+/* Use the generic support for an external assembly language slp_switch function. */
+#define EXTERNAL_ASM
+
+#ifdef SLP_EVAL
+/* This always uses the external masm assembly file. */
+#endif
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_csky_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_csky_gcc.h
new file mode 100644
index 00000000..ac469d3a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_csky_gcc.h
@@ -0,0 +1,48 @@
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+#define REG_FP "r8"
+#ifdef __CSKYABIV2__
+#define REGS_TO_SAVE_GENERAL "r4", "r5", "r6", "r7", "r9", "r10", "r11", "r15",\
+                             "r16", "r17", "r18", "r19", "r20", "r21", "r22",\
+                             "r23", "r24", "r25"
+
+#if defined (__CSKY_HARD_FLOAT__) || (__CSKY_VDSP__)
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL, "vr8", "vr9", "vr10", "vr11", "vr12",\
+                                           "vr13", "vr14", "vr15"
+#else
+#define REGS_TO_SAVE REGS_TO_SAVE_GENERAL
+#endif
+#else
+#define REGS_TO_SAVE "r9", "r10", "r11", "r12", "r13", "r15"
+#endif
+
+
+static int
+#ifdef __GNUC__
+__attribute__((optimize("no-omit-frame-pointer")))
+#endif
+slp_switch(void)
+{
+        int *stackref, stsizediff;
+        int result;
+
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+        __asm__ ("mov %0, sp" : "=r" (stackref));
+        {
+                SLP_SAVE_STATE(stackref, stsizediff);
+                __asm__ volatile (
+                    "addu sp,%0\n"
+                    "addu "REG_FP",%0\n"
+                    :
+                    : "r" (stsizediff)
+                    );
+		
+                SLP_RESTORE_STATE();
+        }
+        __asm__ volatile ("movi %0, 0" : "=r" (result));
+        __asm__ volatile ("" : : : REGS_TO_SAVE);
+
+        return result;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_loongarch64_linux.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_loongarch64_linux.h
new file mode 100644
index 00000000..9eaf34ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_loongarch64_linux.h
@@ -0,0 +1,31 @@
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "s0", "s1", "s2", "s3", "s4", "s5", \
+                    "s6", "s7", "s8", "fp", \
+                    "f24", "f25", "f26", "f27", "f28", "f29", "f30", "f31"
+
+static int
+slp_switch(void)
+{
+    int ret;
+    long *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("move %0, $sp" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "add.d $sp, $sp, %0\n\t"
+            : /* no outputs */
+            : "r" (stsizediff)
+        );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("move %0, $zero" : "=r" (ret) : );
+    return ret;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_m68k_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_m68k_gcc.h
new file mode 100644
index 00000000..da761c2d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_m68k_gcc.h
@@ -0,0 +1,38 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 2014-01-06  Andreas Schwab  <schwab@linux-m68k.org>
+ *      File created.
+ */
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "%d2", "%d3", "%d4", "%d5", "%d6", "%d7", \
+		     "%a2", "%a3", "%a4"
+
+static int
+slp_switch(void)
+{
+  int err;
+  int *stackref, stsizediff;
+  void *fp, *a5;
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  __asm__ volatile ("move.l %%fp, %0" : "=m"(fp));
+  __asm__ volatile ("move.l %%a5, %0" : "=m"(a5));
+  __asm__ ("move.l %%sp, %0" : "=r"(stackref));
+  {
+    SLP_SAVE_STATE(stackref, stsizediff);
+    __asm__ volatile ("add.l %0, %%sp; add.l %0, %%fp" : : "r"(stsizediff));
+    SLP_RESTORE_STATE();
+    __asm__ volatile ("clr.l %0" : "=g" (err));
+  }
+  __asm__ volatile ("move.l %0, %%a5" : : "m"(a5));
+  __asm__ volatile ("move.l %0, %%fp" : : "m"(fp));
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  return err;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_mips_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_mips_unix.h
new file mode 100644
index 00000000..b9003e94
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_mips_unix.h
@@ -0,0 +1,64 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 20-Sep-14 Matt Madison <madison@bliss-m.org>
+ *      Re-code the saving of the gp register for MIPS64.
+ * 05-Jan-08 Thiemo Seufer  <ths@debian.org>
+ *      Ported from ppc.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "$16", "$17", "$18", "$19", "$20", "$21", "$22", \
+       "$23", "$30"
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+#ifdef __mips64
+    uint64_t gpsave;
+#endif
+    __asm__ __volatile__ ("" : : : REGS_TO_SAVE);
+#ifdef __mips64
+    __asm__ __volatile__ ("sd $28,%0" : "=m" (gpsave) : : );
+#endif
+    __asm__ ("move %0, $29" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ __volatile__ (
+#ifdef __mips64
+            "daddu $29, %0\n"
+#else
+            "addu $29, %0\n"
+#endif
+            : /* no outputs */
+            : "r" (stsizediff)
+            );
+        SLP_RESTORE_STATE();
+    }
+#ifdef __mips64
+    __asm__ __volatile__ ("ld $28,%0" : : "m" (gpsave) : );
+#endif
+    __asm__ __volatile__ ("" : : : REGS_TO_SAVE);
+    __asm__ __volatile__ ("move %0, $0" : "=r" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_aix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_aix.h
new file mode 100644
index 00000000..e7e0b877
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_aix.h
@@ -0,0 +1,103 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 16-Oct-20  Jesse Gorzinski <jgorzins@us.ibm.com>
+ *      Copied from Linux PPC64 implementation
+ * 04-Sep-18  Alexey Borzenkov  <snaury@gmail.com>
+ *      Workaround a gcc bug using manual save/restore of r30
+ * 21-Mar-18  Tulio Magno Quites Machado Filho  <tuliom@linux.vnet.ibm.com>
+ *      Added r30 to the list of saved registers in order to fully comply with
+ *      both ppc64 ELFv1 ABI and the ppc64le ELFv2 ABI, that classify this
+ *      register as a nonvolatile register used for local variables.
+ * 21-Mar-18  Laszlo Boszormenyi  <gcs@debian.org>
+ *      Save r2 (TOC pointer) manually.
+ * 10-Dec-13  Ulrich Weigand  <uweigand@de.ibm.com>
+ *	Support ELFv2 ABI.  Save float/vector registers.
+ * 09-Mar-12 Michael Ellerman <michael@ellerman.id.au>
+ *      64-bit implementation, copied from 32-bit.
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 04-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported from MacOS version.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ * 31-Jul-12  Trevor Bowen    <trevorbowen@gmail.com>
+ *      Changed memory constraints to register only.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 6
+
+#if defined(__ALTIVEC__)
+#define ALTIVEC_REGS \
+       "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", \
+       "v28", "v29", "v30", "v31",
+#else
+#define ALTIVEC_REGS
+#endif
+
+#define REGS_TO_SAVE "r14", "r15", "r16", "r17", "r18", "r19", "r20",  \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29",  \
+       "r31",                                                    \
+       "fr14", "fr15", "fr16", "fr17", "fr18", "fr19", "fr20", "fr21", \
+       "fr22", "fr23", "fr24", "fr25", "fr26", "fr27", "fr28", "fr29", \
+       "fr30", "fr31", \
+       ALTIVEC_REGS \
+       "cr2", "cr3", "cr4"
+
+static int
+slp_switch(void)
+{
+    int err;
+    long *stackref, stsizediff;
+    void * toc;
+    void * r30;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("std 2, %0" : "=m" (toc));
+    __asm__ volatile ("std 30, %0" : "=m" (r30));
+    __asm__ ("mr %0, 1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "mr 11, %0\n"
+            "add 1, 1, 11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("ld 30, %0" : : "m" (r30));
+    __asm__ volatile ("ld 2, %0" : : "m" (toc));
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_linux.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_linux.h
new file mode 100644
index 00000000..3c324d00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc64_linux.h
@@ -0,0 +1,105 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 04-Sep-18  Alexey Borzenkov  <snaury@gmail.com>
+ *      Workaround a gcc bug using manual save/restore of r30
+ * 21-Mar-18  Tulio Magno Quites Machado Filho  <tuliom@linux.vnet.ibm.com>
+ *      Added r30 to the list of saved registers in order to fully comply with
+ *      both ppc64 ELFv1 ABI and the ppc64le ELFv2 ABI, that classify this
+ *      register as a nonvolatile register used for local variables.
+ * 21-Mar-18  Laszlo Boszormenyi  <gcs@debian.org>
+ *      Save r2 (TOC pointer) manually.
+ * 10-Dec-13  Ulrich Weigand  <uweigand@de.ibm.com>
+ *	Support ELFv2 ABI.  Save float/vector registers.
+ * 09-Mar-12 Michael Ellerman <michael@ellerman.id.au>
+ *      64-bit implementation, copied from 32-bit.
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 04-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported from MacOS version.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ * 31-Jul-12  Trevor Bowen    <trevorbowen@gmail.com>
+ *      Changed memory constraints to register only.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#if _CALL_ELF == 2
+#define STACK_MAGIC 4
+#else
+#define STACK_MAGIC 6
+#endif
+
+#if defined(__ALTIVEC__)
+#define ALTIVEC_REGS \
+       "v20", "v21", "v22", "v23", "v24", "v25", "v26", "v27", \
+       "v28", "v29", "v30", "v31",
+#else
+#define ALTIVEC_REGS
+#endif
+
+#define REGS_TO_SAVE "r14", "r15", "r16", "r17", "r18", "r19", "r20",  \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29",  \
+       "r31",                                                    \
+       "fr14", "fr15", "fr16", "fr17", "fr18", "fr19", "fr20", "fr21", \
+       "fr22", "fr23", "fr24", "fr25", "fr26", "fr27", "fr28", "fr29", \
+       "fr30", "fr31", \
+       ALTIVEC_REGS \
+       "cr2", "cr3", "cr4"
+
+static int
+slp_switch(void)
+{
+    int err;
+    long *stackref, stsizediff;
+    void * toc;
+    void * r30;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("std 2, %0" : "=m" (toc));
+    __asm__ volatile ("std 30, %0" : "=m" (r30));
+    __asm__ ("mr %0, 1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "mr 11, %0\n"
+            "add 1, 1, 11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("ld 30, %0" : : "m" (r30));
+    __asm__ volatile ("ld 2, %0" : : "m" (toc));
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_aix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_aix.h
new file mode 100644
index 00000000..6d93c132
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_aix.h
@@ -0,0 +1,87 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 07-Mar-11 Floris Bruynooghe <flub@devork.be>
+ *      Do not add stsizediff to general purpose
+ *      register (GPR) 30 as this is a non-volatile and
+ *      unused by the PowerOpen Environment, therefore
+ *      this was modifying a user register instead of the
+ *      frame pointer (which does not seem to exist).
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 04-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported from MacOS version.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 3
+
+/* !!!!WARNING!!!! need to add "r31" in the next line if this header file
+ * is meant to be compiled non-dynamically!
+ */
+#define REGS_TO_SAVE "r13", "r14", "r15", "r16", "r17", "r18", "r19", "r20", \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29", \
+       "cr2", "cr3", "cr4"
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ ("mr %0, 1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "mr 11, %0\n"
+            "add 1, 1, 11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_linux.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_linux.h
new file mode 100644
index 00000000..e83ad70a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_linux.h
@@ -0,0 +1,84 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 04-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported from MacOS version.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ * 31-Jul-12  Trevor Bowen    <trevorbowen@gmail.com>
+ *      Changed memory constraints to register only.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 3
+
+/* !!!!WARNING!!!! need to add "r31" in the next line if this header file
+ * is meant to be compiled non-dynamically!
+ */
+#define REGS_TO_SAVE "r13", "r14", "r15", "r16", "r17", "r18", "r19", "r20", \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29", \
+       "cr2", "cr3", "cr4"
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ ("mr %0, 1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "mr 11, %0\n"
+            "add 1, 1, 11\n"
+            "add 30, 30, 11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_macosx.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_macosx.h
new file mode 100644
index 00000000..bd414c68
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_macosx.h
@@ -0,0 +1,82 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 3
+
+/* !!!!WARNING!!!! need to add "r31" in the next line if this header file
+ * is meant to be compiled non-dynamically!
+ */
+#define REGS_TO_SAVE "r13", "r14", "r15", "r16", "r17", "r18", "r19", "r20", \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29", \
+       "cr2", "cr3", "cr4"
+
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ ("; asm block 2\n\tmr %0, r1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "; asm block 3\n"
+            "\tmr r11, %0\n"
+            "\tadd r1, r1, r11\n"
+            "\tadd r30, r30, r11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "r11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_unix.h
new file mode 100644
index 00000000..bb188080
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_ppc_unix.h
@@ -0,0 +1,82 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'r31' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 14-Jan-04  Bob Ippolito <bob@redivi.com>
+ *      added cr2-cr4 to the registers to be saved.
+ *      Open questions: Should we save FP registers?
+ *      What about vector registers?
+ *      Differences between darwin and unix?
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 04-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported from MacOS version.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 29-Jun-02  Christian Tismer  <tismer@tismer.com>
+ *      Added register 13-29, 31 saves. The same way as
+ *      Armin Rigo did for the x86_unix version.
+ *      This seems to be now fully functional!
+ * 04-Mar-02  Hye-Shik Chang  <perky@fallin.lv>
+ *      Ported from i386.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 3
+
+/* !!!!WARNING!!!! need to add "r31" in the next line if this header file
+ * is meant to be compiled non-dynamically!
+ */
+#define REGS_TO_SAVE "r13", "r14", "r15", "r16", "r17", "r18", "r19", "r20", \
+       "r21", "r22", "r23", "r24", "r25", "r26", "r27", "r28", "r29", \
+       "cr2", "cr3", "cr4"
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ ("mr %0, 1" : "=r" (stackref) : );
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "mr 11, %0\n"
+            "add 1, 1, 11\n"
+            "add 30, 30, 11\n"
+            : /* no outputs */
+            : "r" (stsizediff)
+            : "11"
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("li %0, 0" : "=r" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_riscv_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_riscv_unix.h
new file mode 100644
index 00000000..e74f37af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_riscv_unix.h
@@ -0,0 +1,36 @@
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "s1", "s2", "s3", "s4", "s5", \
+		     "s6", "s7", "s8", "s9", "s10", "s11", "fs0", "fs1", \
+		     "fs2", "fs3", "fs4", "fs5", "fs6", "fs7", "fs8", "fs9", \
+		     "fs10", "fs11"
+
+static int
+slp_switch(void)
+{
+  long fp;
+  int ret;
+  long *stackref, stsizediff;
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  __asm__ volatile ("mv %0, fp" : "=r" (fp) : );
+  __asm__ volatile ("mv %0, sp" : "=r" (stackref) : );
+  {
+      SLP_SAVE_STATE(stackref, stsizediff);
+      __asm__ volatile (
+	  "add sp, sp, %0\n\t"
+	  "add fp, fp, %0\n\t"
+	  : /* no outputs */
+	  : "r" (stsizediff)
+	  );
+      SLP_RESTORE_STATE();
+  }
+  __asm__ volatile ("" : : : REGS_TO_SAVE);
+  __asm__ volatile ("ld fp, %0" : : "m" (fp));
+  __asm__ volatile ("mv %0, zero" : "=r" (ret) : );
+  return ret;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_s390_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_s390_unix.h
new file mode 100644
index 00000000..9199367f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_s390_unix.h
@@ -0,0 +1,87 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 25-Jan-12  Alexey Borzenkov <snaury@gmail.com>
+ *      Fixed Linux/S390 port to work correctly with
+ *      different optimization options both on 31-bit
+ *      and 64-bit. Thanks to Stefan Raabe for lots
+ *      of testing.
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 06-Oct-02  Gustavo Niemeyer <niemeyer@conectiva.com>
+ *      Ported to Linux/S390.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#ifdef __s390x__
+#define STACK_MAGIC 20 /* 20 * 8 = 160 bytes of function call area */
+#else
+#define STACK_MAGIC 24 /* 24 * 4 = 96 bytes of function call area */
+#endif
+
+/* Technically, r11-r13 also need saving, but function prolog starts
+   with stm(g) and since there are so many saved registers already
+   it won't be optimized, resulting in all r6-r15 being saved */
+#define REGS_TO_SAVE "r6", "r7", "r8", "r9", "r10", "r14", \
+		     "f0", "f1", "f2", "f3", "f4", "f5", "f6", "f7", \
+		     "f8", "f9", "f10", "f11", "f12", "f13", "f14", "f15"
+
+static int
+slp_switch(void)
+{
+    int ret;
+    long *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+#ifdef __s390x__
+    __asm__ volatile ("lgr %0, 15" : "=r" (stackref) : );
+#else
+    __asm__ volatile ("lr %0, 15" : "=r" (stackref) : );
+#endif
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+/* N.B.
+   r11 may be used as the frame pointer, and in that case it cannot be
+   clobbered and needs offsetting just like the stack pointer (but in cases
+   where frame pointer isn't used we might clobber it accidentally). What's
+   scary is that r11 is 2nd (and even 1st when GOT is used) callee saved
+   register that gcc would chose for surviving function calls. However,
+   since r6-r10 are clobbered above, their cost for reuse is reduced, so
+   gcc IRA will chose them over r11 (not seeing r11 is implicitly saved),
+   making it relatively safe to offset in all cases. :) */
+        __asm__ volatile (
+#ifdef __s390x__
+            "agr 15, %0\n\t"
+            "agr 11, %0"
+#else
+            "ar 15, %0\n\t"
+            "ar 11, %0"
+#endif
+            : /* no outputs */
+            : "r" (stsizediff)
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("lhi %0, 0" : "=r" (ret) : );
+    return ret;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/* 
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sh_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sh_gcc.h
new file mode 100644
index 00000000..5ecc3b39
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sh_gcc.h
@@ -0,0 +1,36 @@
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+#define STACK_MAGIC 0
+#define REGS_TO_SAVE "r8", "r9", "r10", "r11", "r13", \
+                     "fr12", "fr13", "fr14", "fr15"
+
+// r12 Global context pointer, GP
+// r14 Frame pointer, FP
+// r15 Stack pointer, SP
+
+static int
+slp_switch(void)
+{
+    int err;
+    void* fp;
+    int *stackref, stsizediff;
+    __asm__ volatile("" : : : REGS_TO_SAVE);
+    __asm__ volatile("mov.l r14, %0" : "=m"(fp) : :);
+    __asm__("mov r15, %0" : "=r"(stackref));
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile(
+            "add %0, r15\n"
+            "add %0, r14\n"
+            : /* no outputs */
+            : "r"(stsizediff));
+        SLP_RESTORE_STATE();
+        __asm__ volatile("mov r0, %0" : "=r"(err) : :);
+    }
+    __asm__ volatile("mov.l %0, r14" : : "m"(fp) :);
+    __asm__ volatile("" : : : REGS_TO_SAVE);
+    return err;
+}
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sparc_sun_gcc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sparc_sun_gcc.h
new file mode 100644
index 00000000..96990c39
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_sparc_sun_gcc.h
@@ -0,0 +1,92 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 16-May-15  Alexey Borzenkov <snaury@gmail.com>
+ *      Move stack spilling code inside save/restore functions
+ * 30-Aug-13  Floris Bruynooghe <flub@devork.be>
+        Clean the register windows again before returning.
+        This does not clobber the PIC register as it leaves
+        the current window intact and is required for multi-
+        threaded code to work correctly.
+ * 08-Mar-11  Floris Bruynooghe <flub@devork.be>
+ *      No need to set return value register explicitly
+ *      before the stack and framepointer are adjusted
+ *      as none of the other registers are influenced by
+ *      this.  Also don't needlessly clean the windows
+ *      ('ta %0" :: "i" (ST_CLEAN_WINDOWS)') as that
+ *      clobbers the gcc PIC register (%l7).
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      added support for SunOS sparc with gcc
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+
+#define STACK_MAGIC 0
+
+
+#if defined(__sparcv9)
+#define SLP_FLUSHW __asm__ volatile ("flushw")
+#else
+#define SLP_FLUSHW __asm__ volatile ("ta 3") /* ST_FLUSH_WINDOWS */
+#endif
+
+/* On sparc we need to spill register windows inside save/restore functions */
+#define SLP_BEFORE_SAVE_STATE() SLP_FLUSHW
+#define SLP_BEFORE_RESTORE_STATE() SLP_FLUSHW
+
+
+static int
+slp_switch(void)
+{
+    int err;
+    int *stackref, stsizediff;
+
+    /* Put current stack pointer into stackref.
+     * Register spilling is done in save/restore.
+     */
+    __asm__ volatile ("mov %%sp, %0" : "=r" (stackref));
+
+    {
+        /* Thou shalt put SLP_SAVE_STATE into a local block */
+        /* Copy the current stack onto the heap */
+        SLP_SAVE_STATE(stackref, stsizediff);
+
+        /* Increment stack and frame pointer by stsizediff */
+        __asm__ volatile (
+            "add %0, %%sp, %%sp\n\t"
+            "add %0, %%fp, %%fp"
+            : : "r" (stsizediff));
+
+        /* Copy new stack from it's save store on the heap */
+        SLP_RESTORE_STATE();
+
+        __asm__ volatile ("mov %1, %0" : "=r" (err) : "i" (0));
+        return err;
+    }
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x32_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x32_unix.h
new file mode 100644
index 00000000..893369c7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x32_unix.h
@@ -0,0 +1,63 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 17-Aug-12  Fantix King <fantix.king@gmail.com>
+ *      Ported from amd64.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 0
+
+#define REGS_TO_SAVE "r12", "r13", "r14", "r15"
+
+
+static int
+slp_switch(void)
+{
+    void* ebp;
+    void* ebx;
+    unsigned int csr;
+    unsigned short cw;
+    int err;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("fstcw %0" : "=m" (cw));
+    __asm__ volatile ("stmxcsr %0" : "=m" (csr));
+    __asm__ volatile ("movl %%ebp, %0" : "=m" (ebp));
+    __asm__ volatile ("movl %%ebx, %0" : "=m" (ebx));
+    __asm__ ("movl %%esp, %0" : "=g" (stackref));
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "addl %0, %%esp\n"
+            "addl %0, %%ebp\n"
+            :
+            : "r" (stsizediff)
+            );
+        SLP_RESTORE_STATE();
+    }
+    __asm__ volatile ("movl %0, %%ebx" : : "m" (ebx));
+    __asm__ volatile ("movl %0, %%ebp" : : "m" (ebp));
+    __asm__ volatile ("ldmxcsr %0" : : "m" (csr));
+    __asm__ volatile ("fldcw %0" : : "m" (cw));
+    __asm__ volatile ("" : : : REGS_TO_SAVE);
+    __asm__ volatile ("xorl %%eax, %%eax" : "=a" (err));
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/*
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.asm b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.asm
new file mode 100644
index 00000000..f5c72a27
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.asm
@@ -0,0 +1,111 @@
+;
+; stack switching code for MASM on x641
+; Kristjan Valur Jonsson, sept 2005
+;
+
+
+;prototypes for our calls
+slp_save_state_asm PROTO
+slp_restore_state_asm PROTO
+
+
+pushxmm MACRO reg
+    sub rsp, 16
+    .allocstack 16
+    movaps [rsp], reg ; faster than movups, but we must be aligned
+    ; .savexmm128 reg, offset  (don't know what offset is, no documentation)
+ENDM
+popxmm MACRO reg
+    movaps reg, [rsp] ; faster than movups, but we must be aligned
+    add rsp, 16
+ENDM
+
+pushreg MACRO reg
+    push reg
+    .pushreg reg
+ENDM
+popreg MACRO reg
+    pop reg
+ENDM
+
+
+.code
+slp_switch PROC FRAME
+    ;realign stack to 16 bytes after return address push, makes the following faster
+    sub rsp,8
+    .allocstack 8
+
+    pushxmm xmm15
+    pushxmm xmm14
+    pushxmm xmm13
+    pushxmm xmm12
+    pushxmm xmm11
+    pushxmm xmm10
+    pushxmm xmm9
+    pushxmm xmm8
+    pushxmm xmm7
+    pushxmm xmm6
+
+    pushreg r15
+    pushreg r14
+    pushreg r13
+    pushreg r12
+
+    pushreg rbp
+    pushreg rbx
+    pushreg rdi
+    pushreg rsi
+
+    sub rsp, 10h ;allocate the singlefunction argument (must be multiple of 16)
+    .allocstack 10h
+.endprolog
+
+    lea rcx, [rsp+10h] ;load stack base that we are saving
+    call slp_save_state_asm ;pass stackpointer, return offset in eax
+    cmp rax, 1
+    je EXIT1
+    cmp rax, -1
+    je EXIT2
+    ;actual stack switch:
+    add rsp, rax
+    call slp_restore_state_asm
+    xor rax, rax ;return 0
+
+EXIT:
+
+    add rsp, 10h
+    popreg rsi
+    popreg rdi
+    popreg rbx
+    popreg rbp
+
+    popreg r12
+    popreg r13
+    popreg r14
+    popreg r15
+
+    popxmm xmm6
+    popxmm xmm7
+    popxmm xmm8
+    popxmm xmm9
+    popxmm xmm10
+    popxmm xmm11
+    popxmm xmm12
+    popxmm xmm13
+    popxmm xmm14
+    popxmm xmm15
+
+    add rsp, 8
+    ret
+
+EXIT1:
+    mov rax, 1
+    jmp EXIT
+
+EXIT2:
+    sar rax, 1
+    jmp EXIT
+
+slp_switch ENDP
+
+END
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.obj b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.obj
new file mode 100644
index 00000000..64e3e6b8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_masm.obj differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_msvc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_msvc.h
new file mode 100644
index 00000000..601ea560
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x64_msvc.h
@@ -0,0 +1,60 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 26-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      again as a result of virtualized stack access,
+ *      the compiler used less registers. Needed to
+ *      explicit mention registers in order to get them saved.
+ *      Thanks to Jeff Senn for pointing this out and help.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 01-Mar-02  Christian Tismer  <tismer@tismer.com>
+ *      Initial final version after lots of iterations for i386.
+ */
+
+/* Avoid alloca redefined warning on mingw64 */
+#ifndef alloca
+#define alloca _alloca
+#endif
+
+#define STACK_REFPLUS 1
+#define STACK_MAGIC 0
+
+/* Use the generic support for an external assembly language slp_switch function. */
+#define EXTERNAL_ASM
+
+#ifdef SLP_EVAL
+/* This always uses the external masm assembly file. */
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/* we have IsBadReadPtr available, so we can peek at objects */
+/*
+#define STACKLESS_SPY
+
+#ifdef IMPLEMENT_STACKLESSMODULE
+#include "Windows.h"
+#define CANNOT_READ_MEM(p, bytes) IsBadReadPtr(p, bytes)
+
+static int IS_ON_STACK(void*p)
+{
+    int stackref;
+    intptr_t stackbase = ((intptr_t)&stackref) & 0xfffff000;
+    return (intptr_t)p >= stackbase && (intptr_t)p < stackbase + 0x00100000;
+}
+
+#endif
+*/
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_msvc.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_msvc.h
new file mode 100644
index 00000000..0f3a59f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_msvc.h
@@ -0,0 +1,326 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 26-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      again as a result of virtualized stack access,
+ *      the compiler used less registers. Needed to
+ *      explicit mention registers in order to get them saved.
+ *      Thanks to Jeff Senn for pointing this out and help.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for sparc
+ * 01-Mar-02  Christian Tismer  <tismer@tismer.com>
+ *      Initial final version after lots of iterations for i386.
+ */
+
+#define alloca _alloca
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+#define STACK_MAGIC 0
+
+/* Some magic to quell warnings and keep slp_switch() from crashing when built
+   with VC90. Disable global optimizations, and the warning: frame pointer
+   register 'ebp' modified by inline assembly code.
+
+   We used to just disable global optimizations ("g") but upstream stackless
+   Python, as well as stackman, turn off all optimizations.
+
+References:
+https://github.com/stackless-dev/stackman/blob/dbc72fe5207a2055e658c819fdeab9731dee78b9/stackman/platforms/switch_x86_msvc.h
+https://github.com/stackless-dev/stackless/blob/main-slp/Stackless/platf/switch_x86_msvc.h
+*/
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+
+#pragma optimize("", off) /* so that autos are stored on the stack */
+#pragma warning(disable:4731)
+#pragma warning(disable:4733) /* disable warning about modifying FS[0] */
+
+/**
+ * Most modern compilers and environments handle C++ exceptions without any
+ * special help from us. MSVC on 32-bit windows is an exception. There, C++
+ * exceptions are dealt with using Windows' Structured Exception Handling
+ * (SEH).
+ *
+ * SEH is implemented as a singly linked list of <function*, prev*> nodes. The
+ * head of this list is stored in the Thread Information Block, which itself
+ * is pointed to from the FS register. It's the first field in the structure,
+ * or offset 0, so we can access it using assembly FS:[0], or the compiler
+ * intrinsics and field offset information from the headers (as we do below).
+ * Somewhat unusually, the tail of the list doesn't have prev == NULL, it has
+ * prev == 0xFFFFFFFF.
+ *
+ * SEH was designed for C, and traditionally uses the MSVC compiler
+ * intrinsincs __try{}/__except{}. It is also utilized for C++ exceptions by
+ * MSVC; there, every throw of a C++ exception raises a SEH error with the
+ * ExceptionCode 0xE06D7363; the SEH handler list is then traversed to
+ * deal with the exception.
+ *
+ * If the SEH list is corrupt, then when a C++ exception is thrown the program
+ * will abruptly exit with exit code 1. This does not use std::terminate(), so
+ * std::set_terminate() is useless to debug this.
+ *
+ * The SEH list is closely tied to the call stack; entering a function that
+ * uses __try{} or most C++ functions will push a new handler onto the front
+ * of the list. Returning from the function will remove the handler. Saving
+ * and restoring the head node of the SEH list (FS:[0]) per-greenlet is NOT
+ * ENOUGH to make SEH or exceptions work.
+ *
+ * Stack switching breaks SEH because the call stack no longer necessarily
+ * matches the SEH list. For example, given greenlet A that switches to
+ * greenlet B, at the moment of entering greenlet B, we will have any SEH
+ * handlers from greenlet A on the SEH list; greenlet B can then add its own
+ * handlers to the SEH list. When greenlet B switches back to greenlet A,
+ * greenlet B's handlers would still be on the SEH stack, but when switch()
+ * returns control to greenlet A, we have replaced the contents of the stack
+ * in memory, so all the address that greenlet B added to the SEH list are now
+ * invalid: part of the call stack has been unwound, but the SEH list was out
+ * of sync with the call stack. The net effect is that exception handling
+ * stops working.
+ *
+ * Thus, when switching greenlets, we need to be sure that the SEH list
+ * matches the effective call stack, "cutting out" any handlers that were
+ * pushed by the greenlet that switched out and which are no longer valid.
+ *
+ * The easiest way to do this is to capture the SEH list at the time the main
+ * greenlet for a thread is created, and, when initially starting a greenlet,
+ * start a new SEH list for it, which contains nothing but the handler
+ * established for the new greenlet itself, with the tail being the handlers
+ * for the main greenlet. If we then save and restore the SEH per-greenlet,
+ * they won't interfere with each others SEH lists. (No greenlet can unwind
+ * the call stack past the handlers established by the main greenlet).
+ *
+ * By observation, a new thread starts with three SEH handlers on the list. By
+ * the time we get around to creating the main greenlet, though, there can be
+ * many more, established by transient calls that lead to the creation of the
+ * main greenlet. Therefore, 3 is a magic constant telling us when to perform
+ * the initial slice.
+ *
+ * All of this can be debugged using a vectored exception handler, which
+ * operates independently of the SEH handler list, and is called first.
+ * Walking the SEH list at key points can also be helpful.
+ *
+ * References:
+ * https://en.wikipedia.org/wiki/Win32_Thread_Information_Block
+ * https://devblogs.microsoft.com/oldnewthing/20100730-00/?p=13273
+ * https://docs.microsoft.com/en-us/cpp/cpp/try-except-statement?view=msvc-160
+ * https://docs.microsoft.com/en-us/cpp/cpp/structured-exception-handling-c-cpp?view=msvc-160
+ * https://docs.microsoft.com/en-us/windows/win32/debug/structured-exception-handling
+ * https://docs.microsoft.com/en-us/windows/win32/debug/using-a-vectored-exception-handler
+ * https://bytepointer.com/resources/pietrek_crash_course_depths_of_win32_seh.htm
+ */
+#define GREENLET_NEEDS_EXCEPTION_STATE_SAVED
+
+
+typedef struct _GExceptionRegistration {
+    struct _GExceptionRegistration* prev;
+    void* handler_f;
+} GExceptionRegistration;
+
+static void
+slp_set_exception_state(const void *const seh_state)
+{
+    // Because the stack from from which we do this is ALSO a handler, and
+    // that one we want to keep, we need to relink the current SEH handler
+    // frame to point to this one, cutting out the middle men, as it were.
+    //
+    // Entering a try block doesn't change the SEH frame, but entering a
+    // function containing a try block does.
+    GExceptionRegistration* current_seh_state = (GExceptionRegistration*)__readfsdword(FIELD_OFFSET(NT_TIB, ExceptionList));
+    current_seh_state->prev = (GExceptionRegistration*)seh_state;
+}
+
+
+static GExceptionRegistration*
+x86_slp_get_third_oldest_handler()
+{
+    GExceptionRegistration* a = NULL; /* Closest to the top */
+    GExceptionRegistration* b = NULL; /* second */
+    GExceptionRegistration* c = NULL;
+    GExceptionRegistration* seh_state = (GExceptionRegistration*)__readfsdword(FIELD_OFFSET(NT_TIB, ExceptionList));
+    a = b = c = seh_state;
+
+    while (seh_state && seh_state != (GExceptionRegistration*)0xFFFFFFFF) {
+        if ((void*)seh_state->prev < (void*)100) {
+            fprintf(stderr, "\tERROR: Broken SEH chain.\n");
+            return NULL;
+        }
+        a = b;
+        b = c;
+        c = seh_state;
+
+        seh_state = seh_state->prev;
+    }
+    return a ? a : (b ? b : c);
+}
+
+
+static void*
+slp_get_exception_state()
+{
+    // XXX: There appear to be three SEH handlers on the stack already at the
+    // start of the thread. Is that a guarantee? Almost certainly not. Yet in
+    // all observed cases it has been three. This is consistent with
+    // faulthandler off or on, and optimizations off or on. It may not be
+    // consistent with other operating system versions, though: we only have
+    // CI on one or two versions (don't ask what there are).
+    // In theory we could capture the number of handlers on the chain when
+    // PyInit__greenlet is called: there are probably only the default
+    // handlers at that point (unless we're embedded and people have used
+    // __try/__except or a C++ handler)?
+    return x86_slp_get_third_oldest_handler();
+}
+
+static int
+slp_switch(void)
+{
+    /* MASM syntax is typically reversed from other assemblers.
+       It is usually <instruction> <destination> <source>
+     */
+    int *stackref, stsizediff;
+    /* store the structured exception state for this stack */
+    DWORD seh_state = __readfsdword(FIELD_OFFSET(NT_TIB, ExceptionList));
+    __asm mov stackref, esp;
+    /* modify EBX, ESI and EDI in order to get them preserved */
+    __asm mov ebx, ebx;
+    __asm xchg esi, edi;
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm {
+            mov     eax, stsizediff
+            add     esp, eax
+            add     ebp, eax
+        }
+        SLP_RESTORE_STATE();
+    }
+    __writefsdword(FIELD_OFFSET(NT_TIB, ExceptionList), seh_state);
+    return 0;
+}
+
+/* re-enable ebp warning and global optimizations. */
+#pragma optimize("", on)
+#pragma warning(default:4731)
+#pragma warning(default:4733) /* disable warning about modifying FS[0] */
+
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/* we have IsBadReadPtr available, so we can peek at objects */
+#define STACKLESS_SPY
+
+#ifdef GREENLET_DEBUG
+
+#define CANNOT_READ_MEM(p, bytes) IsBadReadPtr(p, bytes)
+
+static int IS_ON_STACK(void*p)
+{
+    int stackref;
+    int stackbase = ((int)&stackref) & 0xfffff000;
+    return (int)p >= stackbase && (int)p < stackbase + 0x00100000;
+}
+
+static void
+x86_slp_show_seh_chain()
+{
+    GExceptionRegistration* seh_state = (GExceptionRegistration*)__readfsdword(FIELD_OFFSET(NT_TIB, ExceptionList));
+    fprintf(stderr, "====== SEH Chain ======\n");
+    while (seh_state && seh_state != (GExceptionRegistration*)0xFFFFFFFF) {
+        fprintf(stderr, "\tSEH_chain addr: %p handler: %p prev: %p\n",
+                seh_state,
+                seh_state->handler_f, seh_state->prev);
+        if ((void*)seh_state->prev < (void*)100) {
+            fprintf(stderr, "\tERROR: Broken chain.\n");
+            break;
+        }
+        seh_state = seh_state->prev;
+    }
+    fprintf(stderr, "====== End SEH Chain ======\n");
+    fflush(NULL);
+    return;
+}
+
+//addVectoredExceptionHandler constants:
+//CALL_FIRST means call this exception handler first;
+//CALL_LAST means call this exception handler last
+#define CALL_FIRST 1
+#define CALL_LAST 0
+
+LONG WINAPI
+GreenletVectorHandler(PEXCEPTION_POINTERS ExceptionInfo)
+{
+    // We get one of these for every C++ exception, with code
+    // E06D7363
+    // This is a special value that means "C++ exception from MSVC"
+    // https://devblogs.microsoft.com/oldnewthing/20100730-00/?p=13273
+    //
+    // Install in the module init function with:
+    // AddVectoredExceptionHandler(CALL_FIRST, GreenletVectorHandler);
+    PEXCEPTION_RECORD ExceptionRecord = ExceptionInfo->ExceptionRecord;
+
+    fprintf(stderr,
+            "GOT VECTORED EXCEPTION:\n"
+            "\tExceptionCode   : %p\n"
+            "\tExceptionFlags  : %p\n"
+            "\tExceptionAddr   : %p\n"
+            "\tNumberparams    : %ld\n",
+            ExceptionRecord->ExceptionCode,
+            ExceptionRecord->ExceptionFlags,
+            ExceptionRecord->ExceptionAddress,
+            ExceptionRecord->NumberParameters
+            );
+    if (ExceptionRecord->ExceptionFlags & 1) {
+        fprintf(stderr,  "\t\tEH_NONCONTINUABLE\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 2) {
+        fprintf(stderr,  "\t\tEH_UNWINDING\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 4) {
+        fprintf(stderr, "\t\tEH_EXIT_UNWIND\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 8) {
+        fprintf(stderr,  "\t\tEH_STACK_INVALID\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 0x10) {
+        fprintf(stderr,  "\t\tEH_NESTED_CALL\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 0x20) {
+        fprintf(stderr,  "\t\tEH_TARGET_UNWIND\n" );
+    }
+    if (ExceptionRecord->ExceptionFlags & 0x40) {
+        fprintf(stderr,  "\t\tEH_COLLIDED_UNWIND\n" );
+    }
+    fprintf(stderr, "\n");
+    fflush(NULL);
+    for(DWORD i = 0; i < ExceptionRecord->NumberParameters; i++) {
+        fprintf(stderr, "\t\t\tParam %ld: %lX\n", i, ExceptionRecord->ExceptionInformation[i]);
+    }
+
+    if (ExceptionRecord->NumberParameters == 3) {
+        fprintf(stderr, "\tAbout to traverse SEH chain\n");
+        // C++ Exception records have 3 params.
+        x86_slp_show_seh_chain();
+    }
+
+    return EXCEPTION_CONTINUE_SEARCH;
+}
+
+
+
+
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_unix.h b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_unix.h
new file mode 100644
index 00000000..493fa6ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/platform/switch_x86_unix.h
@@ -0,0 +1,105 @@
+/*
+ * this is the internal transfer function.
+ *
+ * HISTORY
+ * 3-May-13   Ralf Schmitt  <ralf@systemexit.de>
+ *     Add support for strange GCC caller-save decisions
+ *     (ported from switch_aarch64_gcc.h)
+ * 19-Aug-11  Alexey Borzenkov  <snaury@gmail.com>
+ *      Correctly save ebp, ebx and cw
+ * 07-Sep-05 (py-dev mailing list discussion)
+ *      removed 'ebx' from the register-saved.  !!!! WARNING !!!!
+ *      It means that this file can no longer be compiled statically!
+ *      It is now only suitable as part of a dynamic library!
+ * 24-Nov-02  Christian Tismer  <tismer@tismer.com>
+ *      needed to add another magic constant to insure
+ *      that f in slp_eval_frame(PyFrameObject *f)
+ *      STACK_REFPLUS will probably be 1 in most cases.
+ *      gets included into the saved stack area.
+ * 17-Sep-02  Christian Tismer  <tismer@tismer.com>
+ *      after virtualizing stack save/restore, the
+ *      stack size shrunk a bit. Needed to introduce
+ *      an adjustment STACK_MAGIC per platform.
+ * 15-Sep-02  Gerd Woetzel       <gerd.woetzel@GMD.DE>
+ *      slightly changed framework for spark
+ * 31-Avr-02  Armin Rigo         <arigo@ulb.ac.be>
+ *      Added ebx, esi and edi register-saves.
+ * 01-Mar-02  Samual M. Rushing  <rushing@ironport.com>
+ *      Ported from i386.
+ */
+
+#define STACK_REFPLUS 1
+
+#ifdef SLP_EVAL
+
+/* #define STACK_MAGIC 3 */
+/* the above works fine with gcc 2.96, but 2.95.3 wants this */
+#define STACK_MAGIC 0
+
+#if __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)
+# define ATTR_NOCLONE __attribute__((noclone))
+#else
+# define ATTR_NOCLONE
+#endif
+
+static int
+slp_switch(void)
+{
+    int err;
+#ifdef _WIN32
+    void *seh;
+#endif
+    void *ebp, *ebx;
+    unsigned short cw;
+    int *stackref, stsizediff;
+    __asm__ volatile ("" : : : "esi", "edi");
+    __asm__ volatile ("fstcw %0" : "=m" (cw));
+    __asm__ volatile ("movl %%ebp, %0" : "=m" (ebp));
+    __asm__ volatile ("movl %%ebx, %0" : "=m" (ebx));
+#ifdef _WIN32
+    __asm__ volatile (
+        "movl %%fs:0x0, %%eax\n"
+        "movl %%eax, %0\n"
+        : "=m" (seh)
+        :
+        : "eax");
+#endif
+    __asm__ ("movl %%esp, %0" : "=g" (stackref));
+    {
+        SLP_SAVE_STATE(stackref, stsizediff);
+        __asm__ volatile (
+            "addl %0, %%esp\n"
+            "addl %0, %%ebp\n"
+            :
+            : "r" (stsizediff)
+            );
+        SLP_RESTORE_STATE();
+        __asm__ volatile ("xorl %%eax, %%eax" : "=a" (err));
+    }
+#ifdef _WIN32
+    __asm__ volatile (
+        "movl %0, %%eax\n"
+        "movl %%eax, %%fs:0x0\n"
+        :
+        : "m" (seh)
+        : "eax");
+#endif
+    __asm__ volatile ("movl %0, %%ebx" : : "m" (ebx));
+    __asm__ volatile ("movl %0, %%ebp" : : "m" (ebp));
+    __asm__ volatile ("fldcw %0" : : "m" (cw));
+    __asm__ volatile ("" : : : "esi", "edi");
+    return err;
+}
+
+#endif
+
+/*
+ * further self-processing support
+ */
+
+/* 
+ * if you want to add self-inspection tools, place them
+ * here. See the x86_msvc for the necessary defines.
+ * These features are highly experimental und not
+ * essential yet.
+ */
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/slp_platformselect.h b/backend/venv/lib/python3.12/site-packages/greenlet/slp_platformselect.h
new file mode 100644
index 00000000..49456482
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/slp_platformselect.h
@@ -0,0 +1,75 @@
+/*
+ * Platform Selection for Stackless Python
+ */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#if   defined(MS_WIN32) && !defined(MS_WIN64) && defined(_M_IX86) && defined(_MSC_VER)
+# include "platform/switch_x86_msvc.h" /* MS Visual Studio on X86 */
+#elif defined(MS_WIN64) && defined(_M_X64) && defined(_MSC_VER) || defined(__MINGW64__)
+# include "platform/switch_x64_msvc.h" /* MS Visual Studio on X64 */
+#elif  defined(MS_WIN64) && defined(_M_ARM64)
+# include "platform/switch_arm64_msvc.h" /* MS Visual Studio on ARM64 */
+#elif defined(__GNUC__) && defined(__amd64__) && defined(__ILP32__)
+# include "platform/switch_x32_unix.h" /* gcc on amd64 with x32 ABI */
+#elif defined(__GNUC__) && defined(__amd64__)
+# include "platform/switch_amd64_unix.h" /* gcc on amd64 */
+#elif defined(__GNUC__) && defined(__i386__)
+# include "platform/switch_x86_unix.h" /* gcc on X86 */
+#elif defined(__GNUC__) && defined(__powerpc64__) && (defined(__linux__) || defined(__FreeBSD__))
+# include "platform/switch_ppc64_linux.h" /* gcc on PowerPC 64-bit */
+#elif defined(__GNUC__) && defined(__PPC__) && (defined(__linux__) || defined(__FreeBSD__))
+# include "platform/switch_ppc_linux.h" /* gcc on PowerPC */
+#elif defined(__GNUC__) && defined(__POWERPC__) && defined(__APPLE__)
+# include "platform/switch_ppc_macosx.h" /* Apple MacOS X on 32-bit PowerPC */
+#elif defined(__GNUC__) && defined(__powerpc64__) && defined(_AIX)
+# include "platform/switch_ppc64_aix.h" /* gcc on AIX/PowerPC 64-bit */
+#elif defined(__GNUC__) && defined(_ARCH_PPC) && defined(_AIX)
+# include "platform/switch_ppc_aix.h" /* gcc on AIX/PowerPC */
+#elif defined(__GNUC__) && defined(__powerpc__) && defined(__NetBSD__)
+#include "platform/switch_ppc_unix.h" /* gcc on NetBSD/powerpc */
+#elif defined(__GNUC__) && defined(sparc)
+# include "platform/switch_sparc_sun_gcc.h" /* SunOS sparc with gcc */
+#elif defined(__SUNPRO_C) && defined(sparc) && defined(sun)
+# iiclude "platform/switch_sparc_sun_gcc.h" /* SunStudio on amd64 */
+#elif defined(__SUNPRO_C) && defined(__amd64__) && defined(sun)
+# include "platform/switch_amd64_unix.h" /* SunStudio on amd64 */
+#elif defined(__SUNPRO_C) && defined(__i386__) && defined(sun)
+# include "platform/switch_x86_unix.h" /* SunStudio on x86 */
+#elif defined(__GNUC__) && defined(__s390__) && defined(__linux__)
+# include "platform/switch_s390_unix.h" /* Linux/S390 */
+#elif defined(__GNUC__) && defined(__s390x__) && defined(__linux__)
+# include "platform/switch_s390_unix.h" /* Linux/S390 zSeries (64-bit) */
+#elif defined(__GNUC__) && defined(__arm__)
+# ifdef __APPLE__
+#  include <TargetConditionals.h>
+# endif
+# if TARGET_OS_IPHONE
+#  include "platform/switch_arm32_ios.h" /* iPhone OS on arm32 */
+# else
+#  include "platform/switch_arm32_gcc.h" /* gcc using arm32 */
+# endif
+#elif defined(__GNUC__) && defined(__mips__) && defined(__linux__)
+# include "platform/switch_mips_unix.h" /* Linux/MIPS */
+#elif defined(__GNUC__) && defined(__aarch64__)
+# include "platform/switch_aarch64_gcc.h" /* Aarch64 ABI */
+#elif defined(__GNUC__) && defined(__mc68000__)
+# include "platform/switch_m68k_gcc.h" /* gcc on m68k */
+#elif defined(__GNUC__) && defined(__csky__)
+#include "platform/switch_csky_gcc.h" /* gcc on csky */
+# elif defined(__GNUC__) && defined(__riscv)
+# include "platform/switch_riscv_unix.h" /* gcc on RISC-V */
+#elif defined(__GNUC__) && defined(__alpha__)
+# include "platform/switch_alpha_unix.h" /* gcc on DEC Alpha */
+#elif defined(MS_WIN32) && defined(__llvm__) && defined(__aarch64__)
+# include "platform/switch_aarch64_gcc.h" /* LLVM Aarch64 ABI for Windows */
+#elif defined(__GNUC__) && defined(__loongarch64) && defined(__linux__)
+# include "platform/switch_loongarch64_linux.h" /* LoongArch64 */
+#elif defined(__GNUC__) && defined(__sh__)
+# include "platform/switch_sh_gcc.h" /* SuperH */
+#endif
+
+#ifdef __cplusplus
+};
+#endif
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__init__.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__init__.py
new file mode 100644
index 00000000..e69392e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__init__.py
@@ -0,0 +1,240 @@
+# -*- coding: utf-8 -*-
+"""
+Tests for greenlet.
+
+"""
+import os
+import sys
+import unittest
+
+from gc import collect
+from gc import get_objects
+from threading import active_count as active_thread_count
+from time import sleep
+from time import time
+
+import psutil
+
+from greenlet import greenlet as RawGreenlet
+from greenlet import getcurrent
+
+from greenlet._greenlet import get_pending_cleanup_count
+from greenlet._greenlet import get_total_main_greenlets
+
+from . import leakcheck
+
+PY312 = sys.version_info[:2] >= (3, 12)
+PY313 = sys.version_info[:2] >= (3, 13)
+
+WIN = sys.platform.startswith("win")
+RUNNING_ON_GITHUB_ACTIONS = os.environ.get('GITHUB_ACTIONS')
+RUNNING_ON_TRAVIS = os.environ.get('TRAVIS') or RUNNING_ON_GITHUB_ACTIONS
+RUNNING_ON_APPVEYOR = os.environ.get('APPVEYOR')
+RUNNING_ON_CI = RUNNING_ON_TRAVIS or RUNNING_ON_APPVEYOR
+RUNNING_ON_MANYLINUX = os.environ.get('GREENLET_MANYLINUX')
+
+class TestCaseMetaClass(type):
+    # wrap each test method with
+    # a) leak checks
+    def __new__(cls, classname, bases, classDict):
+        # pylint and pep8 fight over what this should be called (mcs or cls).
+        # pylint gets it right, but we can't scope disable pep8, so we go with
+        # its convention.
+        # pylint: disable=bad-mcs-classmethod-argument
+        check_totalrefcount = True
+
+        # Python 3: must copy, we mutate the classDict. Interestingly enough,
+        # it doesn't actually error out, but under 3.6 we wind up wrapping
+        # and re-wrapping the same items over and over and over.
+        for key, value in list(classDict.items()):
+            if key.startswith('test') and callable(value):
+                classDict.pop(key)
+                if check_totalrefcount:
+                    value = leakcheck.wrap_refcount(value)
+                classDict[key] = value
+        return type.__new__(cls, classname, bases, classDict)
+
+
+class TestCase(TestCaseMetaClass(
+        "NewBase",
+        (unittest.TestCase,),
+        {})):
+
+    cleanup_attempt_sleep_duration = 0.001
+    cleanup_max_sleep_seconds = 1
+
+    def wait_for_pending_cleanups(self,
+                                  initial_active_threads=None,
+                                  initial_main_greenlets=None):
+        initial_active_threads = initial_active_threads or self.threads_before_test
+        initial_main_greenlets = initial_main_greenlets or self.main_greenlets_before_test
+        sleep_time = self.cleanup_attempt_sleep_duration
+        # NOTE: This is racy! A Python-level thread object may be dead
+        # and gone, but the C thread may not yet have fired its
+        # destructors and added to the queue. There's no particular
+        # way to know that's about to happen. We try to watch the
+        # Python threads to make sure they, at least, have gone away.
+        # Counting the main greenlets, which we can easily do deterministically,
+        # also helps.
+
+        # Always sleep at least once to let other threads run
+        sleep(sleep_time)
+        quit_after = time() + self.cleanup_max_sleep_seconds
+        # TODO: We could add an API that calls us back when a particular main greenlet is deleted?
+        # It would have to drop the GIL
+        while (
+                get_pending_cleanup_count()
+                or active_thread_count() > initial_active_threads
+                or (not self.expect_greenlet_leak
+                    and get_total_main_greenlets() > initial_main_greenlets)):
+            sleep(sleep_time)
+            if time() > quit_after:
+                print("Time limit exceeded.")
+                print("Threads: Waiting for only", initial_active_threads,
+                      "-->", active_thread_count())
+                print("MGlets : Waiting for only", initial_main_greenlets,
+                      "-->", get_total_main_greenlets())
+                break
+        collect()
+
+    def count_objects(self, kind=list, exact_kind=True):
+        # pylint:disable=unidiomatic-typecheck
+        # Collect the garbage.
+        for _ in range(3):
+            collect()
+        if exact_kind:
+            return sum(
+                1
+                for x in get_objects()
+                if type(x) is kind
+            )
+        # instances
+        return sum(
+            1
+            for x in get_objects()
+            if isinstance(x, kind)
+        )
+
+    greenlets_before_test = 0
+    threads_before_test = 0
+    main_greenlets_before_test = 0
+    expect_greenlet_leak = False
+
+    def count_greenlets(self):
+        """
+        Find all the greenlets and subclasses tracked by the GC.
+        """
+        return self.count_objects(RawGreenlet, False)
+
+    def setUp(self):
+        # Ensure the main greenlet exists, otherwise the first test
+        # gets a false positive leak
+        super().setUp()
+        getcurrent()
+        self.threads_before_test = active_thread_count()
+        self.main_greenlets_before_test = get_total_main_greenlets()
+        self.wait_for_pending_cleanups(self.threads_before_test, self.main_greenlets_before_test)
+        self.greenlets_before_test = self.count_greenlets()
+
+    def tearDown(self):
+        if getattr(self, 'skipTearDown', False):
+            return
+
+        self.wait_for_pending_cleanups(self.threads_before_test, self.main_greenlets_before_test)
+        super().tearDown()
+
+    def get_expected_returncodes_for_aborted_process(self):
+        import signal
+        # The child should be aborted in an unusual way. On POSIX
+        # platforms, this is done with abort() and signal.SIGABRT,
+        # which is reflected in a negative return value; however, on
+        # Windows, even though we observe the child print "Fatal
+        # Python error: Aborted" and in older versions of the C
+        # runtime "This application has requested the Runtime to
+        # terminate it in an unusual way," it always has an exit code
+        # of 3. This is interesting because 3 is the error code for
+        # ERROR_PATH_NOT_FOUND; BUT: the C runtime abort() function
+        # also uses this code.
+        #
+        # If we link to the static C library on Windows, the error
+        # code changes to '0xc0000409' (hex(3221226505)), which
+        # apparently is STATUS_STACK_BUFFER_OVERRUN; but "What this
+        # means is that nowadays when you get a
+        # STATUS_STACK_BUFFER_OVERRUN, it doesn’t actually mean that
+        # there is a stack buffer overrun. It just means that the
+        # application decided to terminate itself with great haste."
+        #
+        #
+        # On windows, we've also seen '0xc0000005' (hex(3221225477)).
+        # That's "Access Violation"
+        #
+        # See
+        # https://devblogs.microsoft.com/oldnewthing/20110519-00/?p=10623
+        # and
+        # https://docs.microsoft.com/en-us/previous-versions/k089yyh0(v=vs.140)?redirectedfrom=MSDN
+        # and
+        # https://devblogs.microsoft.com/oldnewthing/20190108-00/?p=100655
+        expected_exit = (
+            -signal.SIGABRT,
+            # But beginning on Python 3.11, the faulthandler
+            # that prints the C backtraces sometimes segfaults after
+            # reporting the exception but before printing the stack.
+            # This has only been seen on linux/gcc.
+            -signal.SIGSEGV,
+        ) if not WIN else (
+            3,
+            0xc0000409,
+            0xc0000005,
+        )
+        return expected_exit
+
+    def get_process_uss(self):
+        """
+        Return the current process's USS in bytes.
+
+        uss is available on Linux, macOS, Windows. Also known as
+        "Unique Set Size", this is the memory which is unique to a
+        process and which would be freed if the process was terminated
+        right now.
+
+        If this is not supported by ``psutil``, this raises the
+        :exc:`unittest.SkipTest` exception.
+        """
+        try:
+            return psutil.Process().memory_full_info().uss
+        except AttributeError as e:
+            raise unittest.SkipTest("uss not supported") from e
+
+    def run_script(self, script_name, show_output=True):
+        import subprocess
+        script = os.path.join(
+            os.path.dirname(__file__),
+            script_name,
+        )
+
+        try:
+            return subprocess.check_output([sys.executable, script],
+                                           encoding='utf-8',
+                                           stderr=subprocess.STDOUT)
+        except subprocess.CalledProcessError as ex:
+            if show_output:
+                print('-----')
+                print('Failed to run script', script)
+                print('~~~~~')
+                print(ex.output)
+                print('------')
+            raise
+
+
+    def assertScriptRaises(self, script_name, exitcodes=None):
+        import subprocess
+        with self.assertRaises(subprocess.CalledProcessError) as exc:
+            output = self.run_script(script_name, show_output=False)
+            __traceback_info__ = output
+            # We're going to fail the assertion if we get here, at least
+            # preserve the output in the traceback.
+
+        if exitcodes is None:
+            exitcodes = self.get_expected_returncodes_for_aborted_process()
+        self.assertIn(exc.exception.returncode, exitcodes)
+        return exc.exception
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..c35cdfc4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_clearing_run_switches.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_clearing_run_switches.cpython-312.pyc
new file mode 100644
index 00000000..5d7fd574
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_clearing_run_switches.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_cpp_exception.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_cpp_exception.cpython-312.pyc
new file mode 100644
index 00000000..57c712a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_cpp_exception.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_initialstub_already_started.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_initialstub_already_started.cpython-312.pyc
new file mode 100644
index 00000000..b24bca8d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_initialstub_already_started.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_slp_switch.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_slp_switch.cpython-312.pyc
new file mode 100644
index 00000000..dcaf4a6f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_slp_switch.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets.cpython-312.pyc
new file mode 100644
index 00000000..9e81dfa4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets2.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets2.cpython-312.pyc
new file mode 100644
index 00000000..705cc637
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_three_greenlets2.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_two_greenlets.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_two_greenlets.cpython-312.pyc
new file mode 100644
index 00000000..0f64199b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/fail_switch_two_greenlets.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/leakcheck.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/leakcheck.cpython-312.pyc
new file mode 100644
index 00000000..69b63981
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/leakcheck.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_contextvars.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_contextvars.cpython-312.pyc
new file mode 100644
index 00000000..2ab37ebc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_contextvars.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_cpp.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_cpp.cpython-312.pyc
new file mode 100644
index 00000000..5d464d2b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_cpp.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_extension_interface.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_extension_interface.cpython-312.pyc
new file mode 100644
index 00000000..3f754ce9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_extension_interface.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_gc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_gc.cpython-312.pyc
new file mode 100644
index 00000000..9053f387
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_gc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator.cpython-312.pyc
new file mode 100644
index 00000000..3842eb61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator_nested.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator_nested.cpython-312.pyc
new file mode 100644
index 00000000..77d94ad1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_generator_nested.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet.cpython-312.pyc
new file mode 100644
index 00000000..4e0ac2c9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet_trash.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet_trash.cpython-312.pyc
new file mode 100644
index 00000000..089eb1da
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_greenlet_trash.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_leaks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_leaks.cpython-312.pyc
new file mode 100644
index 00000000..522e9ef4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_leaks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_stack_saved.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_stack_saved.cpython-312.pyc
new file mode 100644
index 00000000..f45ecbb5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_stack_saved.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_throw.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_throw.cpython-312.pyc
new file mode 100644
index 00000000..9bc051c2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_throw.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_tracing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_tracing.cpython-312.pyc
new file mode 100644
index 00000000..1cbe8463
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_tracing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_version.cpython-312.pyc
new file mode 100644
index 00000000..fbfebf57
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_weakref.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_weakref.cpython-312.pyc
new file mode 100644
index 00000000..a08d8bbe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/__pycache__/test_weakref.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.c b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.c
new file mode 100644
index 00000000..05e81c03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.c
@@ -0,0 +1,231 @@
+/* This is a set of functions used by test_extension_interface.py to test the
+ * Greenlet C API.
+ */
+
+#include "../greenlet.h"
+
+#ifndef Py_RETURN_NONE
+#    define Py_RETURN_NONE return Py_INCREF(Py_None), Py_None
+#endif
+
+#define TEST_MODULE_NAME "_test_extension"
+
+static PyObject*
+test_switch(PyObject* self, PyObject* greenlet)
+{
+    PyObject* result = NULL;
+
+    if (greenlet == NULL || !PyGreenlet_Check(greenlet)) {
+        PyErr_BadArgument();
+        return NULL;
+    }
+
+    result = PyGreenlet_Switch((PyGreenlet*)greenlet, NULL, NULL);
+    if (result == NULL) {
+        if (!PyErr_Occurred()) {
+            PyErr_SetString(PyExc_AssertionError,
+                            "greenlet.switch() failed for some reason.");
+        }
+        return NULL;
+    }
+    Py_INCREF(result);
+    return result;
+}
+
+static PyObject*
+test_switch_kwargs(PyObject* self, PyObject* args, PyObject* kwargs)
+{
+    PyGreenlet* g = NULL;
+    PyObject* result = NULL;
+
+    PyArg_ParseTuple(args, "O!", &PyGreenlet_Type, &g);
+
+    if (g == NULL || !PyGreenlet_Check(g)) {
+        PyErr_BadArgument();
+        return NULL;
+    }
+
+    result = PyGreenlet_Switch(g, NULL, kwargs);
+    if (result == NULL) {
+        if (!PyErr_Occurred()) {
+            PyErr_SetString(PyExc_AssertionError,
+                            "greenlet.switch() failed for some reason.");
+        }
+        return NULL;
+    }
+    Py_XINCREF(result);
+    return result;
+}
+
+static PyObject*
+test_getcurrent(PyObject* self)
+{
+    PyGreenlet* g = PyGreenlet_GetCurrent();
+    if (g == NULL || !PyGreenlet_Check(g) || !PyGreenlet_ACTIVE(g)) {
+        PyErr_SetString(PyExc_AssertionError,
+                        "getcurrent() returned an invalid greenlet");
+        Py_XDECREF(g);
+        return NULL;
+    }
+    Py_DECREF(g);
+    Py_RETURN_NONE;
+}
+
+static PyObject*
+test_setparent(PyObject* self, PyObject* arg)
+{
+    PyGreenlet* current;
+    PyGreenlet* greenlet = NULL;
+
+    if (arg == NULL || !PyGreenlet_Check(arg)) {
+        PyErr_BadArgument();
+        return NULL;
+    }
+    if ((current = PyGreenlet_GetCurrent()) == NULL) {
+        return NULL;
+    }
+    greenlet = (PyGreenlet*)arg;
+    if (PyGreenlet_SetParent(greenlet, current)) {
+        Py_DECREF(current);
+        return NULL;
+    }
+    Py_DECREF(current);
+    if (PyGreenlet_Switch(greenlet, NULL, NULL) == NULL) {
+        return NULL;
+    }
+    Py_RETURN_NONE;
+}
+
+static PyObject*
+test_new_greenlet(PyObject* self, PyObject* callable)
+{
+    PyObject* result = NULL;
+    PyGreenlet* greenlet = PyGreenlet_New(callable, NULL);
+
+    if (!greenlet) {
+        return NULL;
+    }
+
+    result = PyGreenlet_Switch(greenlet, NULL, NULL);
+    Py_CLEAR(greenlet);
+    if (result == NULL) {
+        return NULL;
+    }
+
+    Py_INCREF(result);
+    return result;
+}
+
+static PyObject*
+test_raise_dead_greenlet(PyObject* self)
+{
+    PyErr_SetString(PyExc_GreenletExit, "test GreenletExit exception.");
+    return NULL;
+}
+
+static PyObject*
+test_raise_greenlet_error(PyObject* self)
+{
+    PyErr_SetString(PyExc_GreenletError, "test greenlet.error exception");
+    return NULL;
+}
+
+static PyObject*
+test_throw(PyObject* self, PyGreenlet* g)
+{
+    const char msg[] = "take that sucka!";
+    PyObject* msg_obj = Py_BuildValue("s", msg);
+    PyGreenlet_Throw(g, PyExc_ValueError, msg_obj, NULL);
+    Py_DECREF(msg_obj);
+    if (PyErr_Occurred()) {
+        return NULL;
+    }
+    Py_RETURN_NONE;
+}
+
+static PyObject*
+test_throw_exact(PyObject* self, PyObject* args)
+{
+    PyGreenlet* g = NULL;
+    PyObject* typ = NULL;
+    PyObject* val = NULL;
+    PyObject* tb = NULL;
+
+    if (!PyArg_ParseTuple(args, "OOOO:throw", &g, &typ, &val, &tb)) {
+        return NULL;
+    }
+
+    PyGreenlet_Throw(g, typ, val, tb);
+    if (PyErr_Occurred()) {
+        return NULL;
+    }
+    Py_RETURN_NONE;
+}
+
+static PyMethodDef test_methods[] = {
+    {"test_switch",
+     (PyCFunction)test_switch,
+     METH_O,
+     "Switch to the provided greenlet sending provided arguments, and \n"
+     "return the results."},
+    {"test_switch_kwargs",
+     (PyCFunction)test_switch_kwargs,
+     METH_VARARGS | METH_KEYWORDS,
+     "Switch to the provided greenlet sending the provided keyword args."},
+    {"test_getcurrent",
+     (PyCFunction)test_getcurrent,
+     METH_NOARGS,
+     "Test PyGreenlet_GetCurrent()"},
+    {"test_setparent",
+     (PyCFunction)test_setparent,
+     METH_O,
+     "Se the parent of the provided greenlet and switch to it."},
+    {"test_new_greenlet",
+     (PyCFunction)test_new_greenlet,
+     METH_O,
+     "Test PyGreenlet_New()"},
+    {"test_raise_dead_greenlet",
+     (PyCFunction)test_raise_dead_greenlet,
+     METH_NOARGS,
+     "Just raise greenlet.GreenletExit"},
+    {"test_raise_greenlet_error",
+     (PyCFunction)test_raise_greenlet_error,
+     METH_NOARGS,
+     "Just raise greenlet.error"},
+    {"test_throw",
+     (PyCFunction)test_throw,
+     METH_O,
+     "Throw a ValueError at the provided greenlet"},
+    {"test_throw_exact",
+     (PyCFunction)test_throw_exact,
+     METH_VARARGS,
+     "Throw exactly the arguments given at the provided greenlet"},
+    {NULL, NULL, 0, NULL}
+};
+
+
+#define INITERROR return NULL
+
+static struct PyModuleDef moduledef = {PyModuleDef_HEAD_INIT,
+                                       TEST_MODULE_NAME,
+                                       NULL,
+                                       0,
+                                       test_methods,
+                                       NULL,
+                                       NULL,
+                                       NULL,
+                                       NULL};
+
+PyMODINIT_FUNC
+PyInit__test_extension(void)
+{
+    PyObject* module = NULL;
+    module = PyModule_Create(&moduledef);
+
+    if (module == NULL) {
+        return NULL;
+    }
+
+    PyGreenlet_Import();
+    return module;
+}
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.cpython-312-darwin.so
new file mode 100755
index 00000000..fd3c93a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpp b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpp
new file mode 100644
index 00000000..5cbe6a76
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpp
@@ -0,0 +1,226 @@
+/* This is a set of functions used to test C++ exceptions are not
+ * broken during greenlet switches
+ */
+
+#include "../greenlet.h"
+#include "../greenlet_compiler_compat.hpp"
+#include <exception>
+#include <stdexcept>
+
+struct exception_t {
+    int depth;
+    exception_t(int depth) : depth(depth) {}
+};
+
+/* Functions are called via pointers to prevent inlining */
+static void (*p_test_exception_throw_nonstd)(int depth);
+static void (*p_test_exception_throw_std)();
+static PyObject* (*p_test_exception_switch_recurse)(int depth, int left);
+
+static void
+test_exception_throw_nonstd(int depth)
+{
+    throw exception_t(depth);
+}
+
+static void
+test_exception_throw_std()
+{
+    throw std::runtime_error("Thrown from an extension.");
+}
+
+static PyObject*
+test_exception_switch_recurse(int depth, int left)
+{
+    if (left > 0) {
+        return p_test_exception_switch_recurse(depth, left - 1);
+    }
+
+    PyObject* result = NULL;
+    PyGreenlet* self = PyGreenlet_GetCurrent();
+    if (self == NULL)
+        return NULL;
+
+    try {
+        if (PyGreenlet_Switch(PyGreenlet_GET_PARENT(self), NULL, NULL) == NULL) {
+            Py_DECREF(self);
+            return NULL;
+        }
+        p_test_exception_throw_nonstd(depth);
+        PyErr_SetString(PyExc_RuntimeError,
+                        "throwing C++ exception didn't work");
+    }
+    catch (const exception_t& e) {
+        if (e.depth != depth)
+            PyErr_SetString(PyExc_AssertionError, "depth mismatch");
+        else
+            result = PyLong_FromLong(depth);
+    }
+    catch (...) {
+        PyErr_SetString(PyExc_RuntimeError, "unexpected C++ exception");
+    }
+
+    Py_DECREF(self);
+    return result;
+}
+
+/* test_exception_switch(int depth)
+ * - recurses depth times
+ * - switches to parent inside try/catch block
+ * - throws an exception that (expected to be caught in the same function)
+ * - verifies depth matches (exceptions shouldn't be caught in other greenlets)
+ */
+static PyObject*
+test_exception_switch(PyObject* UNUSED(self), PyObject* args)
+{
+    int depth;
+    if (!PyArg_ParseTuple(args, "i", &depth))
+        return NULL;
+    return p_test_exception_switch_recurse(depth, depth);
+}
+
+
+static PyObject*
+py_test_exception_throw_nonstd(PyObject* self, PyObject* args)
+{
+    if (!PyArg_ParseTuple(args, ""))
+        return NULL;
+    p_test_exception_throw_nonstd(0);
+    PyErr_SetString(PyExc_AssertionError, "unreachable code running after throw");
+    return NULL;
+}
+
+static PyObject*
+py_test_exception_throw_std(PyObject* self, PyObject* args)
+{
+    if (!PyArg_ParseTuple(args, ""))
+        return NULL;
+    p_test_exception_throw_std();
+    PyErr_SetString(PyExc_AssertionError, "unreachable code running after throw");
+    return NULL;
+}
+
+static PyObject*
+py_test_call(PyObject* self, PyObject* arg)
+{
+    PyObject* noargs = PyTuple_New(0);
+    PyObject* ret = PyObject_Call(arg, noargs, nullptr);
+    Py_DECREF(noargs);
+    return ret;
+}
+
+
+
+/* test_exception_switch_and_do_in_g2(g2func)
+ * - creates new greenlet g2 to run g2func
+ * - switches to g2 inside try/catch block
+ * - verifies that no exception has been caught
+ *
+ * it is used together with test_exception_throw to verify that unhandled
+ * exceptions thrown in one greenlet do not propagate to other greenlet nor
+ * segfault the process.
+ */
+static PyObject*
+test_exception_switch_and_do_in_g2(PyObject* self, PyObject* args)
+{
+    PyObject* g2func = NULL;
+    PyObject* result = NULL;
+
+    if (!PyArg_ParseTuple(args, "O", &g2func))
+        return NULL;
+    PyGreenlet* g2 = PyGreenlet_New(g2func, NULL);
+    if (!g2) {
+        return NULL;
+    }
+
+    try {
+        result = PyGreenlet_Switch(g2, NULL, NULL);
+        if (!result) {
+            return NULL;
+        }
+    }
+    catch (const exception_t& e) {
+        /* if we are here the memory can be already corrupted and the program
+         * might crash before below py-level exception might become printed.
+         * -> print something to stderr to make it clear that we had entered
+         *    this catch block.
+         * See comments in inner_bootstrap()
+         */
+#if defined(WIN32) || defined(_WIN32)
+        fprintf(stderr, "C++ exception unexpectedly caught in g1\n");
+        PyErr_SetString(PyExc_AssertionError, "C++ exception unexpectedly caught in g1");
+        Py_XDECREF(result);
+        return NULL;
+#else
+        throw;
+#endif
+    }
+
+    Py_XDECREF(result);
+    Py_RETURN_NONE;
+}
+
+static PyMethodDef test_methods[] = {
+    {"test_exception_switch",
+     (PyCFunction)&test_exception_switch,
+     METH_VARARGS,
+     "Switches to parent twice, to test exception handling and greenlet "
+     "switching."},
+    {"test_exception_switch_and_do_in_g2",
+     (PyCFunction)&test_exception_switch_and_do_in_g2,
+     METH_VARARGS,
+     "Creates new greenlet g2 to run g2func and switches to it inside try/catch "
+     "block. Used together with test_exception_throw to verify that unhandled "
+     "C++ exceptions thrown in a greenlet doe not corrupt memory."},
+    {"test_exception_throw_nonstd",
+     (PyCFunction)&py_test_exception_throw_nonstd,
+     METH_VARARGS,
+     "Throws non-standard C++ exception. Calling this function directly should abort the process."
+    },
+    {"test_exception_throw_std",
+     (PyCFunction)&py_test_exception_throw_std,
+     METH_VARARGS,
+     "Throws standard C++ exception. Calling this function directly should abort the process."
+    },
+    {"test_call",
+     (PyCFunction)&py_test_call,
+     METH_O,
+     "Call the given callable. Unlike calling it directly, this creates a "
+     "new C-level stack frame, which may be helpful in testing."
+    },
+    {NULL, NULL, 0, NULL}
+};
+
+
+static struct PyModuleDef moduledef = {PyModuleDef_HEAD_INIT,
+                                       "greenlet.tests._test_extension_cpp",
+                                       NULL,
+                                       0,
+                                       test_methods,
+                                       NULL,
+                                       NULL,
+                                       NULL,
+                                       NULL};
+
+PyMODINIT_FUNC
+PyInit__test_extension_cpp(void)
+{
+    PyObject* module = NULL;
+
+    module = PyModule_Create(&moduledef);
+
+    if (module == NULL) {
+        return NULL;
+    }
+
+    PyGreenlet_Import();
+    if (_PyGreenlet_API == NULL) {
+        return NULL;
+    }
+
+    p_test_exception_throw_nonstd = test_exception_throw_nonstd;
+    p_test_exception_throw_std = test_exception_throw_std;
+    p_test_exception_switch_recurse = test_exception_switch_recurse;
+
+    return module;
+}
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpython-312-darwin.so
new file mode 100755
index 00000000..90ff4db7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/greenlet/tests/_test_extension_cpp.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_clearing_run_switches.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_clearing_run_switches.py
new file mode 100644
index 00000000..6dd1492f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_clearing_run_switches.py
@@ -0,0 +1,47 @@
+# -*- coding: utf-8 -*-
+"""
+If we have a run callable passed to the constructor or set as an
+attribute, but we don't actually use that (because ``__getattribute__``
+or the like interferes), then when we clear callable before beginning
+to run, there's an opportunity for Python code to run.
+
+"""
+import greenlet
+
+g = None
+main = greenlet.getcurrent()
+
+results = []
+
+class RunCallable:
+
+    def __del__(self):
+        results.append(('RunCallable', '__del__'))
+        main.switch('from RunCallable')
+
+
+class G(greenlet.greenlet):
+
+    def __getattribute__(self, name):
+        if name == 'run':
+            results.append(('G.__getattribute__', 'run'))
+            return run_func
+        return object.__getattribute__(self, name)
+
+
+def run_func():
+    results.append(('run_func', 'enter'))
+
+
+g = G(RunCallable())
+# Try to start G. It will get to the point where it deletes
+# its run callable C++ variable in inner_bootstrap. That triggers
+# the __del__ method, which switches back to main before g
+# actually even starts running.
+x = g.switch()
+results.append(('main: g.switch()', x))
+# In the C++ code, this results in g->g_switch() appearing to return, even though
+# it has yet to run.
+print('In main with', x, flush=True)
+g.switch()
+print('RESULTS', results)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_cpp_exception.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_cpp_exception.py
new file mode 100644
index 00000000..fa4dc2eb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_cpp_exception.py
@@ -0,0 +1,33 @@
+# -*- coding: utf-8 -*-
+"""
+Helper for testing a C++ exception throw aborts the process.
+
+Takes one argument, the name of the function in :mod:`_test_extension_cpp` to call.
+"""
+import sys
+import greenlet
+from greenlet.tests import _test_extension_cpp
+print('fail_cpp_exception is running')
+
+def run_unhandled_exception_in_greenlet_aborts():
+    def _():
+        _test_extension_cpp.test_exception_switch_and_do_in_g2(
+            _test_extension_cpp.test_exception_throw_nonstd
+        )
+    g1 = greenlet.greenlet(_)
+    g1.switch()
+
+
+func_name = sys.argv[1]
+try:
+    func = getattr(_test_extension_cpp, func_name)
+except AttributeError:
+    if func_name == run_unhandled_exception_in_greenlet_aborts.__name__:
+        func = run_unhandled_exception_in_greenlet_aborts
+    elif func_name == 'run_as_greenlet_target':
+        g = greenlet.greenlet(_test_extension_cpp.test_exception_throw_std)
+        func = g.switch
+    else:
+        raise
+print('raising', func, flush=True)
+func()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_initialstub_already_started.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_initialstub_already_started.py
new file mode 100644
index 00000000..c1a44efd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_initialstub_already_started.py
@@ -0,0 +1,78 @@
+"""
+Testing initialstub throwing an already started exception.
+"""
+
+import greenlet
+
+a = None
+b = None
+c = None
+main = greenlet.getcurrent()
+
+# If we switch into a dead greenlet,
+# we go looking for its parents.
+# if a parent is not yet started, we start it.
+
+results = []
+
+def a_run(*args):
+    #results.append('A')
+    results.append(('Begin A', args))
+
+
+def c_run():
+    results.append('Begin C')
+    b.switch('From C')
+    results.append('C done')
+
+class A(greenlet.greenlet): pass
+
+class B(greenlet.greenlet):
+    doing_it = False
+    def __getattribute__(self, name):
+        if name == 'run' and not self.doing_it:
+            assert greenlet.getcurrent() is c
+            self.doing_it = True
+            results.append('Switch to b from B.__getattribute__ in '
+                           + type(greenlet.getcurrent()).__name__)
+            b.switch()
+            results.append('B.__getattribute__ back from main in '
+                           + type(greenlet.getcurrent()).__name__)
+        if name == 'run':
+            name = '_B_run'
+        return object.__getattribute__(self, name)
+
+    def _B_run(self, *arg):
+        results.append(('Begin B', arg))
+        results.append('_B_run switching to main')
+        main.switch('From B')
+
+class C(greenlet.greenlet):
+    pass
+a = A(a_run)
+b = B(parent=a)
+c = C(c_run, b)
+
+# Start a child; while running, it will start B,
+# but starting B will ALSO start B.
+result = c.switch()
+results.append(('main from c', result))
+
+# Switch back to C, which was in the middle of switching
+# already. This will throw the ``GreenletStartedWhileInPython``
+# exception, which results in parent A getting started (B is finished)
+c.switch()
+
+results.append(('A dead?', a.dead, 'B dead?', b.dead, 'C dead?', c.dead))
+
+# A and B should both be dead now.
+assert a.dead
+assert b.dead
+assert not c.dead
+
+result = c.switch()
+results.append(('main from c.2', result))
+# Now C is dead
+assert c.dead
+
+print("RESULTS:", results)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_slp_switch.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_slp_switch.py
new file mode 100644
index 00000000..09905269
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_slp_switch.py
@@ -0,0 +1,29 @@
+# -*- coding: utf-8 -*-
+"""
+A test helper for seeing what happens when slp_switch()
+fails.
+"""
+# pragma: no cover
+
+import greenlet
+
+
+print('fail_slp_switch is running', flush=True)
+
+runs = []
+def func():
+    runs.append(1)
+    greenlet.getcurrent().parent.switch()
+    runs.append(2)
+    greenlet.getcurrent().parent.switch()
+    runs.append(3)
+
+g = greenlet._greenlet.UnswitchableGreenlet(func)
+g.switch()
+assert runs == [1]
+g.switch()
+assert runs == [1, 2]
+g.force_slp_switch_error = True
+
+# This should crash.
+g.switch()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets.py
new file mode 100644
index 00000000..e151b19a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets.py
@@ -0,0 +1,44 @@
+"""
+Uses a trace function to switch greenlets at unexpected times.
+
+In the trace function, we switch from the current greenlet to another
+greenlet, which switches
+"""
+import greenlet
+
+g1 = None
+g2 = None
+
+switch_to_g2 = False
+
+def tracefunc(*args):
+    print('TRACE', *args)
+    global switch_to_g2
+    if switch_to_g2:
+        switch_to_g2 = False
+        g2.switch()
+    print('\tLEAVE TRACE', *args)
+
+def g1_run():
+    print('In g1_run')
+    global switch_to_g2
+    switch_to_g2 = True
+    from_parent = greenlet.getcurrent().parent.switch()
+    print('Return to g1_run')
+    print('From parent', from_parent)
+
+def g2_run():
+    #g1.switch()
+    greenlet.getcurrent().parent.switch()
+
+greenlet.settrace(tracefunc)
+
+g1 = greenlet.greenlet(g1_run)
+g2 = greenlet.greenlet(g2_run)
+
+# This switch didn't actually finish!
+# And if it did, it would raise TypeError
+# because g1_run() doesn't take any arguments.
+g1.switch(1)
+print('Back in main')
+g1.switch(2)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets2.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets2.py
new file mode 100644
index 00000000..1f6b66bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_three_greenlets2.py
@@ -0,0 +1,55 @@
+"""
+Like fail_switch_three_greenlets, but the call into g1_run would actually be
+valid.
+"""
+import greenlet
+
+g1 = None
+g2 = None
+
+switch_to_g2 = True
+
+results = []
+
+def tracefunc(*args):
+    results.append(('trace', args[0]))
+    print('TRACE', *args)
+    global switch_to_g2
+    if switch_to_g2:
+        switch_to_g2 = False
+        g2.switch('g2 from tracefunc')
+    print('\tLEAVE TRACE', *args)
+
+def g1_run(arg):
+    results.append(('g1 arg', arg))
+    print('In g1_run')
+    from_parent = greenlet.getcurrent().parent.switch('from g1_run')
+    results.append(('g1 from parent', from_parent))
+    return 'g1 done'
+
+def g2_run(arg):
+    #g1.switch()
+    results.append(('g2 arg', arg))
+    parent = greenlet.getcurrent().parent.switch('from g2_run')
+    global switch_to_g2
+    switch_to_g2 = False
+    results.append(('g2 from parent', parent))
+    return 'g2 done'
+
+
+greenlet.settrace(tracefunc)
+
+g1 = greenlet.greenlet(g1_run)
+g2 = greenlet.greenlet(g2_run)
+
+x = g1.switch('g1 from main')
+results.append(('main g1', x))
+print('Back in main', x)
+x = g1.switch('g2 from main')
+results.append(('main g2', x))
+print('back in amain again', x)
+x = g1.switch('g1 from main 2')
+results.append(('main g1.2', x))
+x = g2.switch()
+results.append(('main g2.2', x))
+print("RESULTS:", results)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_two_greenlets.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_two_greenlets.py
new file mode 100644
index 00000000..3e52345a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/fail_switch_two_greenlets.py
@@ -0,0 +1,41 @@
+"""
+Uses a trace function to switch greenlets at unexpected times.
+
+In the trace function, we switch from the current greenlet to another
+greenlet, which switches
+"""
+import greenlet
+
+g1 = None
+g2 = None
+
+switch_to_g2 = False
+
+def tracefunc(*args):
+    print('TRACE', *args)
+    global switch_to_g2
+    if switch_to_g2:
+        switch_to_g2 = False
+        g2.switch()
+    print('\tLEAVE TRACE', *args)
+
+def g1_run():
+    print('In g1_run')
+    global switch_to_g2
+    switch_to_g2 = True
+    greenlet.getcurrent().parent.switch()
+    print('Return to g1_run')
+    print('Falling off end of g1_run')
+
+def g2_run():
+    g1.switch()
+    print('Falling off end of g2')
+
+greenlet.settrace(tracefunc)
+
+g1 = greenlet.greenlet(g1_run)
+g2 = greenlet.greenlet(g2_run)
+
+g1.switch()
+print('Falling off end of main')
+g2.switch()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/leakcheck.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/leakcheck.py
new file mode 100644
index 00000000..a5152fb2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/leakcheck.py
@@ -0,0 +1,319 @@
+# Copyright (c) 2018 gevent community
+# Copyright (c) 2021 greenlet community
+#
+# This was originally part of gevent's test suite. The main author
+# (Jason Madden) vendored a copy of it into greenlet.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+# THE SOFTWARE.
+from __future__ import print_function
+
+import os
+import sys
+import gc
+
+from functools import wraps
+import unittest
+
+
+import objgraph
+
+# graphviz 0.18 (Nov 7 2021), available only on Python 3.6 and newer,
+# has added type hints (sigh). It wants to use ``typing.Literal`` for
+# some stuff, but that's only available on Python 3.9+. If that's not
+# found, it creates a ``unittest.mock.MagicMock`` object and annotates
+# with that. These are GC'able objects, and doing almost *anything*
+# with them results in an explosion of objects. For example, trying to
+# compare them for equality creates new objects. This causes our
+# leakchecks to fail, with reports like:
+#
+# greenlet.tests.leakcheck.LeakCheckError: refcount increased by [337, 1333, 343, 430, 530, 643, 769]
+# _Call          1820      +546
+# dict           4094       +76
+# MagicProxy      585       +73
+# tuple          2693       +66
+# _CallList        24        +3
+# weakref        1441        +1
+# function       5996        +1
+# type            736        +1
+# cell            592        +1
+# MagicMock         8        +1
+#
+# To avoid this, we *could* filter this type of object out early. In
+# principle it could leak, but we don't use mocks in greenlet, so it
+# doesn't leak from us. However, a further issue is that ``MagicMock``
+# objects have subobjects that are also GC'able, like ``_Call``, and
+# those create new mocks of their own too. So we'd have to filter them
+# as well, and they're not public. That's OK, we can workaround the
+# problem by being very careful to never compare by equality or other
+# user-defined operators, only using object identity or other builtin
+# functions.
+
+RUNNING_ON_GITHUB_ACTIONS = os.environ.get('GITHUB_ACTIONS')
+RUNNING_ON_TRAVIS = os.environ.get('TRAVIS') or RUNNING_ON_GITHUB_ACTIONS
+RUNNING_ON_APPVEYOR = os.environ.get('APPVEYOR')
+RUNNING_ON_CI = RUNNING_ON_TRAVIS or RUNNING_ON_APPVEYOR
+RUNNING_ON_MANYLINUX = os.environ.get('GREENLET_MANYLINUX')
+SKIP_LEAKCHECKS = RUNNING_ON_MANYLINUX or os.environ.get('GREENLET_SKIP_LEAKCHECKS')
+SKIP_FAILING_LEAKCHECKS = os.environ.get('GREENLET_SKIP_FAILING_LEAKCHECKS')
+ONLY_FAILING_LEAKCHECKS = os.environ.get('GREENLET_ONLY_FAILING_LEAKCHECKS')
+
+def ignores_leakcheck(func):
+    """
+    Ignore the given object during leakchecks.
+
+    Can be applied to a method, in which case the method will run, but
+    will not be subject to leak checks.
+
+    If applied to a class, the entire class will be skipped during leakchecks. This
+    is intended to be used for classes that are very slow and cause problems such as
+    test timeouts; typically it will be used for classes that are subclasses of a base
+    class and specify variants of behaviour (such as pool sizes).
+    """
+    func.ignore_leakcheck = True
+    return func
+
+def fails_leakcheck(func):
+    """
+    Mark that the function is known to leak.
+    """
+    func.fails_leakcheck = True
+    if SKIP_FAILING_LEAKCHECKS:
+        func = unittest.skip("Skipping known failures")(func)
+    return func
+
+class LeakCheckError(AssertionError):
+    pass
+
+if hasattr(sys, 'getobjects'):
+    # In a Python build with ``--with-trace-refs``, make objgraph
+    # trace *all* the objects, not just those that are tracked by the
+    # GC
+    class _MockGC(object):
+        def get_objects(self):
+            return sys.getobjects(0) # pylint:disable=no-member
+        def __getattr__(self, name):
+            return getattr(gc, name)
+    objgraph.gc = _MockGC()
+    fails_strict_leakcheck = fails_leakcheck
+else:
+    def fails_strict_leakcheck(func):
+        """
+        Decorator for a function that is known to fail when running
+        strict (``sys.getobjects()``) leakchecks.
+
+        This type of leakcheck finds all objects, even those, such as
+        strings, which are not tracked by the garbage collector.
+        """
+        return func
+
+class ignores_types_in_strict_leakcheck(object):
+    def __init__(self, types):
+        self.types = types
+    def __call__(self, func):
+        func.leakcheck_ignore_types = self.types
+        return func
+
+class _RefCountChecker(object):
+
+    # Some builtin things that we ignore
+    # XXX: Those things were ignored by gevent, but they're important here,
+    # presumably.
+    IGNORED_TYPES = () #(tuple, dict, types.FrameType, types.TracebackType)
+
+    def __init__(self, testcase, function):
+        self.testcase = testcase
+        self.function = function
+        self.deltas = []
+        self.peak_stats = {}
+        self.ignored_types = ()
+
+        # The very first time we are called, we have already been
+        # self.setUp() by the test runner, so we don't need to do it again.
+        self.needs_setUp = False
+
+    def _include_object_p(self, obj):
+        # pylint:disable=too-many-return-statements
+        #
+        # See the comment block at the top. We must be careful to
+        # avoid invoking user-defined operations.
+        if obj is self:
+            return False
+        kind = type(obj)
+        # ``self._include_object_p == obj`` returns NotImplemented
+        # for non-function objects, which causes the interpreter
+        # to try to reverse the order of arguments...which leads
+        # to the explosion of mock objects. We don't want that, so we implement
+        # the check manually.
+        if kind == type(self._include_object_p):
+            try:
+                # pylint:disable=not-callable
+                exact_method_equals = self._include_object_p.__eq__(obj)
+            except AttributeError:
+                # Python 2.7 methods may only have __cmp__, and that raises a
+                # TypeError for non-method arguments
+                # pylint:disable=no-member
+                exact_method_equals = self._include_object_p.__cmp__(obj) == 0
+
+            if exact_method_equals is not NotImplemented and exact_method_equals:
+                return False
+
+        # Similarly, we need to check identity in our __dict__ to avoid mock explosions.
+        for x in self.__dict__.values():
+            if obj is x:
+                return False
+
+
+        if kind in self.ignored_types or kind in self.IGNORED_TYPES:
+            return False
+
+        return True
+
+    def _growth(self):
+        return objgraph.growth(limit=None, peak_stats=self.peak_stats,
+                               filter=self._include_object_p)
+
+    def _report_diff(self, growth):
+        if not growth:
+            return "<Unable to calculate growth>"
+
+        lines = []
+        width = max(len(name) for name, _, _ in growth)
+        for name, count, delta in growth:
+            lines.append('%-*s%9d %+9d' % (width, name, count, delta))
+
+        diff = '\n'.join(lines)
+        return diff
+
+
+    def _run_test(self, args, kwargs):
+        gc_enabled = gc.isenabled()
+        gc.disable()
+
+        if self.needs_setUp:
+            self.testcase.setUp()
+            self.testcase.skipTearDown = False
+        try:
+            self.function(self.testcase, *args, **kwargs)
+        finally:
+            self.testcase.tearDown()
+            self.testcase.doCleanups()
+            self.testcase.skipTearDown = True
+            self.needs_setUp = True
+            if gc_enabled:
+                gc.enable()
+
+    def _growth_after(self):
+        # Grab post snapshot
+        # pylint:disable=no-member
+        if 'urlparse' in sys.modules:
+            sys.modules['urlparse'].clear_cache()
+        if 'urllib.parse' in sys.modules:
+            sys.modules['urllib.parse'].clear_cache()
+
+        return self._growth()
+
+    def _check_deltas(self, growth):
+        # Return false when we have decided there is no leak,
+        # true if we should keep looping, raises an assertion
+        # if we have decided there is a leak.
+
+        deltas = self.deltas
+        if not deltas:
+            # We haven't run yet, no data, keep looping
+            return True
+
+        if gc.garbage:
+            raise LeakCheckError("Generated uncollectable garbage %r" % (gc.garbage,))
+
+
+        # the following configurations are classified as "no leak"
+        # [0, 0]
+        # [x, 0, 0]
+        # [... a, b, c, d]  where a+b+c+d = 0
+        #
+        # the following configurations are classified as "leak"
+        # [... z, z, z]  where z > 0
+
+        if deltas[-2:] == [0, 0] and len(deltas) in (2, 3):
+            return False
+
+        if deltas[-3:] == [0, 0, 0]:
+            return False
+
+        if len(deltas) >= 4 and sum(deltas[-4:]) == 0:
+            return False
+
+        if len(deltas) >= 3 and deltas[-1] > 0 and deltas[-1] == deltas[-2] and deltas[-2] == deltas[-3]:
+            diff = self._report_diff(growth)
+            raise LeakCheckError('refcount increased by %r\n%s' % (deltas, diff))
+
+        # OK, we don't know for sure yet. Let's search for more
+        if sum(deltas[-3:]) <= 0 or sum(deltas[-4:]) <= 0 or deltas[-4:].count(0) >= 2:
+            # this is suspicious, so give a few more runs
+            limit = 11
+        else:
+            limit = 7
+        if len(deltas) >= limit:
+            raise LeakCheckError('refcount increased by %r\n%s'
+                                 % (deltas,
+                                    self._report_diff(growth)))
+
+        # We couldn't decide yet, keep going
+        return True
+
+    def __call__(self, args, kwargs):
+        for _ in range(3):
+            gc.collect()
+
+        expect_failure = getattr(self.function, 'fails_leakcheck', False)
+        if expect_failure:
+            self.testcase.expect_greenlet_leak = True
+        self.ignored_types = getattr(self.function, "leakcheck_ignore_types", ())
+
+        # Capture state before; the incremental will be
+        # updated by each call to _growth_after
+        growth = self._growth()
+
+        try:
+            while self._check_deltas(growth):
+                self._run_test(args, kwargs)
+
+                growth = self._growth_after()
+
+                self.deltas.append(sum((stat[2] for stat in growth)))
+        except LeakCheckError:
+            if not expect_failure:
+                raise
+        else:
+            if expect_failure:
+                raise LeakCheckError("Expected %s to leak but it did not." % (self.function,))
+
+def wrap_refcount(method):
+    if getattr(method, 'ignore_leakcheck', False) or SKIP_LEAKCHECKS:
+        return method
+
+    @wraps(method)
+    def wrapper(self, *args, **kwargs): # pylint:disable=too-many-branches
+        if getattr(self, 'ignore_leakcheck', False):
+            raise unittest.SkipTest("This class ignored during leakchecks")
+        if ONLY_FAILING_LEAKCHECKS and not getattr(method, 'fails_leakcheck', False):
+            raise unittest.SkipTest("Only running tests that fail leakchecks.")
+        return _RefCountChecker(self, method)(args, kwargs)
+
+    return wrapper
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_contextvars.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_contextvars.py
new file mode 100644
index 00000000..9a16f671
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_contextvars.py
@@ -0,0 +1,310 @@
+from __future__ import print_function
+
+import gc
+import sys
+import unittest
+
+from functools import partial
+from unittest import skipUnless
+from unittest import skipIf
+
+from greenlet import greenlet
+from greenlet import getcurrent
+from . import TestCase
+
+
+try:
+    from contextvars import Context
+    from contextvars import ContextVar
+    from contextvars import copy_context
+    # From the documentation:
+    #
+    # Important: Context Variables should be created at the top module
+    # level and never in closures. Context objects hold strong
+    # references to context variables which prevents context variables
+    # from being properly garbage collected.
+    ID_VAR = ContextVar("id", default=None)
+    VAR_VAR = ContextVar("var", default=None)
+    ContextVar = None
+except ImportError:
+    Context = ContextVar = copy_context = None
+
+# We don't support testing if greenlet's built-in context var support is disabled.
+@skipUnless(Context is not None, "ContextVar not supported")
+class ContextVarsTests(TestCase):
+    def _new_ctx_run(self, *args, **kwargs):
+        return copy_context().run(*args, **kwargs)
+
+    def _increment(self, greenlet_id, callback, counts, expect):
+        ctx_var = ID_VAR
+        if expect is None:
+            self.assertIsNone(ctx_var.get())
+        else:
+            self.assertEqual(ctx_var.get(), expect)
+        ctx_var.set(greenlet_id)
+        for _ in range(2):
+            counts[ctx_var.get()] += 1
+            callback()
+
+    def _test_context(self, propagate_by):
+        # pylint:disable=too-many-branches
+        ID_VAR.set(0)
+
+        callback = getcurrent().switch
+        counts = dict((i, 0) for i in range(5))
+
+        lets = [
+            greenlet(partial(
+                partial(
+                    copy_context().run,
+                    self._increment
+                ) if propagate_by == "run" else self._increment,
+                greenlet_id=i,
+                callback=callback,
+                counts=counts,
+                expect=(
+                    i - 1 if propagate_by == "share" else
+                    0 if propagate_by in ("set", "run") else None
+                )
+            ))
+            for i in range(1, 5)
+        ]
+
+        for let in lets:
+            if propagate_by == "set":
+                let.gr_context = copy_context()
+            elif propagate_by == "share":
+                let.gr_context = getcurrent().gr_context
+
+        for i in range(2):
+            counts[ID_VAR.get()] += 1
+            for let in lets:
+                let.switch()
+
+        if propagate_by == "run":
+            # Must leave each context.run() in reverse order of entry
+            for let in reversed(lets):
+                let.switch()
+        else:
+            # No context.run(), so fine to exit in any order.
+            for let in lets:
+                let.switch()
+
+        for let in lets:
+            self.assertTrue(let.dead)
+            # When using run(), we leave the run() as the greenlet dies,
+            # and there's no context "underneath". When not using run(),
+            # gr_context still reflects the context the greenlet was
+            # running in.
+            if propagate_by == 'run':
+                self.assertIsNone(let.gr_context)
+            else:
+                self.assertIsNotNone(let.gr_context)
+
+
+        if propagate_by == "share":
+            self.assertEqual(counts, {0: 1, 1: 1, 2: 1, 3: 1, 4: 6})
+        else:
+            self.assertEqual(set(counts.values()), set([2]))
+
+    def test_context_propagated_by_context_run(self):
+        self._new_ctx_run(self._test_context, "run")
+
+    def test_context_propagated_by_setting_attribute(self):
+        self._new_ctx_run(self._test_context, "set")
+
+    def test_context_not_propagated(self):
+        self._new_ctx_run(self._test_context, None)
+
+    def test_context_shared(self):
+        self._new_ctx_run(self._test_context, "share")
+
+    def test_break_ctxvars(self):
+        let1 = greenlet(copy_context().run)
+        let2 = greenlet(copy_context().run)
+        let1.switch(getcurrent().switch)
+        let2.switch(getcurrent().switch)
+        # Since let2 entered the current context and let1 exits its own, the
+        # interpreter emits:
+        # RuntimeError: cannot exit context: thread state references a different context object
+        let1.switch()
+
+    def test_not_broken_if_using_attribute_instead_of_context_run(self):
+        let1 = greenlet(getcurrent().switch)
+        let2 = greenlet(getcurrent().switch)
+        let1.gr_context = copy_context()
+        let2.gr_context = copy_context()
+        let1.switch()
+        let2.switch()
+        let1.switch()
+        let2.switch()
+
+    def test_context_assignment_while_running(self):
+        # pylint:disable=too-many-statements
+        ID_VAR.set(None)
+
+        def target():
+            self.assertIsNone(ID_VAR.get())
+            self.assertIsNone(gr.gr_context)
+
+            # Context is created on first use
+            ID_VAR.set(1)
+            self.assertIsInstance(gr.gr_context, Context)
+            self.assertEqual(ID_VAR.get(), 1)
+            self.assertEqual(gr.gr_context[ID_VAR], 1)
+
+            # Clearing the context makes it get re-created as another
+            # empty context when next used
+            old_context = gr.gr_context
+            gr.gr_context = None  # assign None while running
+            self.assertIsNone(ID_VAR.get())
+            self.assertIsNone(gr.gr_context)
+            ID_VAR.set(2)
+            self.assertIsInstance(gr.gr_context, Context)
+            self.assertEqual(ID_VAR.get(), 2)
+            self.assertEqual(gr.gr_context[ID_VAR], 2)
+
+            new_context = gr.gr_context
+            getcurrent().parent.switch((old_context, new_context))
+            # parent switches us back to old_context
+
+            self.assertEqual(ID_VAR.get(), 1)
+            gr.gr_context = new_context  # assign non-None while running
+            self.assertEqual(ID_VAR.get(), 2)
+
+            getcurrent().parent.switch()
+            # parent switches us back to no context
+            self.assertIsNone(ID_VAR.get())
+            self.assertIsNone(gr.gr_context)
+            gr.gr_context = old_context
+            self.assertEqual(ID_VAR.get(), 1)
+
+            getcurrent().parent.switch()
+            # parent switches us back to no context
+            self.assertIsNone(ID_VAR.get())
+            self.assertIsNone(gr.gr_context)
+
+        gr = greenlet(target)
+
+        with self.assertRaisesRegex(AttributeError, "can't delete context attribute"):
+            del gr.gr_context
+
+        self.assertIsNone(gr.gr_context)
+        old_context, new_context = gr.switch()
+        self.assertIs(new_context, gr.gr_context)
+        self.assertEqual(old_context[ID_VAR], 1)
+        self.assertEqual(new_context[ID_VAR], 2)
+        self.assertEqual(new_context.run(ID_VAR.get), 2)
+        gr.gr_context = old_context  # assign non-None while suspended
+        gr.switch()
+        self.assertIs(gr.gr_context, new_context)
+        gr.gr_context = None  # assign None while suspended
+        gr.switch()
+        self.assertIs(gr.gr_context, old_context)
+        gr.gr_context = None
+        gr.switch()
+        self.assertIsNone(gr.gr_context)
+
+        # Make sure there are no reference leaks
+        gr = None
+        gc.collect()
+        self.assertEqual(sys.getrefcount(old_context), 2)
+        self.assertEqual(sys.getrefcount(new_context), 2)
+
+    def test_context_assignment_different_thread(self):
+        import threading
+        VAR_VAR.set(None)
+        ctx = Context()
+
+        is_running = threading.Event()
+        should_suspend = threading.Event()
+        did_suspend = threading.Event()
+        should_exit = threading.Event()
+        holder = []
+
+        def greenlet_in_thread_fn():
+            VAR_VAR.set(1)
+            is_running.set()
+            should_suspend.wait(10)
+            VAR_VAR.set(2)
+            getcurrent().parent.switch()
+            holder.append(VAR_VAR.get())
+
+        def thread_fn():
+            gr = greenlet(greenlet_in_thread_fn)
+            gr.gr_context = ctx
+            holder.append(gr)
+            gr.switch()
+            did_suspend.set()
+            should_exit.wait(10)
+            gr.switch()
+            del gr
+            greenlet() # trigger cleanup
+
+        thread = threading.Thread(target=thread_fn, daemon=True)
+        thread.start()
+        is_running.wait(10)
+        gr = holder[0]
+
+        # Can't access or modify context if the greenlet is running
+        # in a different thread
+        with self.assertRaisesRegex(ValueError, "running in a different"):
+            getattr(gr, 'gr_context')
+        with self.assertRaisesRegex(ValueError, "running in a different"):
+            gr.gr_context = None
+
+        should_suspend.set()
+        did_suspend.wait(10)
+
+        # OK to access and modify context if greenlet is suspended
+        self.assertIs(gr.gr_context, ctx)
+        self.assertEqual(gr.gr_context[VAR_VAR], 2)
+        gr.gr_context = None
+
+        should_exit.set()
+        thread.join(10)
+
+        self.assertEqual(holder, [gr, None])
+
+        # Context can still be accessed/modified when greenlet is dead:
+        self.assertIsNone(gr.gr_context)
+        gr.gr_context = ctx
+        self.assertIs(gr.gr_context, ctx)
+
+        # Otherwise we leak greenlets on some platforms.
+        # XXX: Should be able to do this automatically
+        del holder[:]
+        gr = None
+        thread = None
+
+    def test_context_assignment_wrong_type(self):
+        g = greenlet()
+        with self.assertRaisesRegex(TypeError,
+                                    "greenlet context must be a contextvars.Context or None"):
+            g.gr_context = self
+
+
+@skipIf(Context is not None, "ContextVar supported")
+class NoContextVarsTests(TestCase):
+    def test_contextvars_errors(self):
+        let1 = greenlet(getcurrent().switch)
+        self.assertFalse(hasattr(let1, 'gr_context'))
+        with self.assertRaises(AttributeError):
+            getattr(let1, 'gr_context')
+
+        with self.assertRaises(AttributeError):
+            let1.gr_context = None
+
+        let1.switch()
+
+        with self.assertRaises(AttributeError):
+            getattr(let1, 'gr_context')
+
+        with self.assertRaises(AttributeError):
+            let1.gr_context = None
+
+        del let1
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_cpp.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_cpp.py
new file mode 100644
index 00000000..2d0cc9c9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_cpp.py
@@ -0,0 +1,73 @@
+from __future__ import print_function
+from __future__ import absolute_import
+
+import subprocess
+import unittest
+
+import greenlet
+from . import _test_extension_cpp
+from . import TestCase
+from . import WIN
+
+class CPPTests(TestCase):
+    def test_exception_switch(self):
+        greenlets = []
+        for i in range(4):
+            g = greenlet.greenlet(_test_extension_cpp.test_exception_switch)
+            g.switch(i)
+            greenlets.append(g)
+        for i, g in enumerate(greenlets):
+            self.assertEqual(g.switch(), i)
+
+    def _do_test_unhandled_exception(self, target):
+        import os
+        import sys
+        script = os.path.join(
+            os.path.dirname(__file__),
+            'fail_cpp_exception.py',
+        )
+        args = [sys.executable, script, target.__name__ if not isinstance(target, str) else target]
+        __traceback_info__ = args
+        with self.assertRaises(subprocess.CalledProcessError) as exc:
+            subprocess.check_output(
+                args,
+                encoding='utf-8',
+                stderr=subprocess.STDOUT
+            )
+
+        ex = exc.exception
+        expected_exit = self.get_expected_returncodes_for_aborted_process()
+        self.assertIn(ex.returncode, expected_exit)
+        self.assertIn('fail_cpp_exception is running', ex.output)
+        return ex.output
+
+
+    def test_unhandled_nonstd_exception_aborts(self):
+        # verify that plain unhandled throw aborts
+        self._do_test_unhandled_exception(_test_extension_cpp.test_exception_throw_nonstd)
+
+    def test_unhandled_std_exception_aborts(self):
+        # verify that plain unhandled throw aborts
+        self._do_test_unhandled_exception(_test_extension_cpp.test_exception_throw_std)
+
+    @unittest.skipIf(WIN, "XXX: This does not crash on Windows")
+    # Meaning the exception is getting lost somewhere...
+    def test_unhandled_std_exception_as_greenlet_function_aborts(self):
+        # verify that plain unhandled throw aborts
+        output = self._do_test_unhandled_exception('run_as_greenlet_target')
+        self.assertIn(
+            # We really expect this to be prefixed with "greenlet: Unhandled C++ exception:"
+            # as added by our handler for std::exception (see TUserGreenlet.cpp), but
+            # that's not correct everywhere --- our handler never runs before std::terminate
+            # gets called (for example, on arm32).
+            'Thrown from an extension.',
+            output
+        )
+
+    def test_unhandled_exception_in_greenlet_aborts(self):
+        # verify that unhandled throw called in greenlet aborts too
+        self._do_test_unhandled_exception('run_unhandled_exception_in_greenlet_aborts')
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_extension_interface.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_extension_interface.py
new file mode 100644
index 00000000..34b66567
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_extension_interface.py
@@ -0,0 +1,115 @@
+from __future__ import print_function
+from __future__ import absolute_import
+
+import sys
+
+import greenlet
+from . import _test_extension
+from . import TestCase
+
+# pylint:disable=c-extension-no-member
+
+class CAPITests(TestCase):
+    def test_switch(self):
+        self.assertEqual(
+            50, _test_extension.test_switch(greenlet.greenlet(lambda: 50)))
+
+    def test_switch_kwargs(self):
+        def adder(x, y):
+            return x * y
+        g = greenlet.greenlet(adder)
+        self.assertEqual(6, _test_extension.test_switch_kwargs(g, x=3, y=2))
+
+    def test_setparent(self):
+        # pylint:disable=disallowed-name
+        def foo():
+            def bar():
+                greenlet.getcurrent().parent.switch()
+
+                # This final switch should go back to the main greenlet, since
+                # the test_setparent() function in the C extension should have
+                # reparented this greenlet.
+                greenlet.getcurrent().parent.switch()
+                raise AssertionError("Should never have reached this code")
+            child = greenlet.greenlet(bar)
+            child.switch()
+            greenlet.getcurrent().parent.switch(child)
+            greenlet.getcurrent().parent.throw(
+                AssertionError("Should never reach this code"))
+        foo_child = greenlet.greenlet(foo).switch()
+        self.assertEqual(None, _test_extension.test_setparent(foo_child))
+
+    def test_getcurrent(self):
+        _test_extension.test_getcurrent()
+
+    def test_new_greenlet(self):
+        self.assertEqual(-15, _test_extension.test_new_greenlet(lambda: -15))
+
+    def test_raise_greenlet_dead(self):
+        self.assertRaises(
+            greenlet.GreenletExit, _test_extension.test_raise_dead_greenlet)
+
+    def test_raise_greenlet_error(self):
+        self.assertRaises(
+            greenlet.error, _test_extension.test_raise_greenlet_error)
+
+    def test_throw(self):
+        seen = []
+
+        def foo():         # pylint:disable=disallowed-name
+            try:
+                greenlet.getcurrent().parent.switch()
+            except ValueError:
+                seen.append(sys.exc_info()[1])
+            except greenlet.GreenletExit:
+                raise AssertionError
+        g = greenlet.greenlet(foo)
+        g.switch()
+        _test_extension.test_throw(g)
+        self.assertEqual(len(seen), 1)
+        self.assertTrue(
+            isinstance(seen[0], ValueError),
+            "ValueError was not raised in foo()")
+        self.assertEqual(
+            str(seen[0]),
+            'take that sucka!',
+            "message doesn't match")
+
+    def test_non_traceback_param(self):
+        with self.assertRaises(TypeError) as exc:
+            _test_extension.test_throw_exact(
+                greenlet.getcurrent(),
+                Exception,
+                Exception(),
+                self
+            )
+        self.assertEqual(str(exc.exception),
+                         "throw() third argument must be a traceback object")
+
+    def test_instance_of_wrong_type(self):
+        with self.assertRaises(TypeError) as exc:
+            _test_extension.test_throw_exact(
+                greenlet.getcurrent(),
+                Exception(),
+                BaseException(),
+                None,
+            )
+
+        self.assertEqual(str(exc.exception),
+                         "instance exception may not have a separate value")
+
+    def test_not_throwable(self):
+        with self.assertRaises(TypeError) as exc:
+            _test_extension.test_throw_exact(
+                greenlet.getcurrent(),
+                "abc",
+                None,
+                None,
+            )
+        self.assertEqual(str(exc.exception),
+                         "exceptions must be classes, or instances, not str")
+
+
+if __name__ == '__main__':
+    import unittest
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_gc.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_gc.py
new file mode 100644
index 00000000..994addb9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_gc.py
@@ -0,0 +1,86 @@
+import gc
+
+import weakref
+
+import greenlet
+
+
+from . import TestCase
+from .leakcheck import fails_leakcheck
+# These only work with greenlet gc support
+# which is no longer optional.
+assert greenlet.GREENLET_USE_GC
+
+class GCTests(TestCase):
+    def test_dead_circular_ref(self):
+        o = weakref.ref(greenlet.greenlet(greenlet.getcurrent).switch())
+        gc.collect()
+        if o() is not None:
+            import sys
+            print("O IS NOT NONE.", sys.getrefcount(o()))
+        self.assertIsNone(o())
+        self.assertFalse(gc.garbage, gc.garbage)
+
+    def test_circular_greenlet(self):
+        class circular_greenlet(greenlet.greenlet):
+            self = None
+        o = circular_greenlet()
+        o.self = o
+        o = weakref.ref(o)
+        gc.collect()
+        self.assertIsNone(o())
+        self.assertFalse(gc.garbage, gc.garbage)
+
+    def test_inactive_ref(self):
+        class inactive_greenlet(greenlet.greenlet):
+            def __init__(self):
+                greenlet.greenlet.__init__(self, run=self.run)
+
+            def run(self):
+                pass
+        o = inactive_greenlet()
+        o = weakref.ref(o)
+        gc.collect()
+        self.assertIsNone(o())
+        self.assertFalse(gc.garbage, gc.garbage)
+
+    @fails_leakcheck
+    def test_finalizer_crash(self):
+        # This test is designed to crash when active greenlets
+        # are made garbage collectable, until the underlying
+        # problem is resolved. How does it work:
+        # - order of object creation is important
+        # - array is created first, so it is moved to unreachable first
+        # - we create a cycle between a greenlet and this array
+        # - we create an object that participates in gc, is only
+        #   referenced by a greenlet, and would corrupt gc lists
+        #   on destruction, the easiest is to use an object with
+        #   a finalizer
+        # - because array is the first object in unreachable it is
+        #   cleared first, which causes all references to greenlet
+        #   to disappear and causes greenlet to be destroyed, but since
+        #   it is still live it causes a switch during gc, which causes
+        #   an object with finalizer to be destroyed, which causes stack
+        #   corruption and then a crash
+
+        class object_with_finalizer(object):
+            def __del__(self):
+                pass
+        array = []
+        parent = greenlet.getcurrent()
+        def greenlet_body():
+            greenlet.getcurrent().object = object_with_finalizer()
+            try:
+                parent.switch()
+            except greenlet.GreenletExit:
+                print("Got greenlet exit!")
+            finally:
+                del greenlet.getcurrent().object
+        g = greenlet.greenlet(greenlet_body)
+        g.array = array
+        array.append(g)
+        g.switch()
+        del array
+        del g
+        greenlet.getcurrent()
+        gc.collect()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator.py
new file mode 100644
index 00000000..ca4a644b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator.py
@@ -0,0 +1,59 @@
+
+from greenlet import greenlet
+
+from . import TestCase
+
+class genlet(greenlet):
+    parent = None
+    def __init__(self, *args, **kwds):
+        self.args = args
+        self.kwds = kwds
+
+    def run(self):
+        fn, = self.fn
+        fn(*self.args, **self.kwds)
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        self.parent = greenlet.getcurrent()
+        result = self.switch()
+        if self:
+            return result
+
+        raise StopIteration
+
+    next = __next__
+
+
+def Yield(value):
+    g = greenlet.getcurrent()
+    while not isinstance(g, genlet):
+        if g is None:
+            raise RuntimeError('yield outside a genlet')
+        g = g.parent
+    g.parent.switch(value)
+
+
+def generator(func):
+    class Generator(genlet):
+        fn = (func,)
+    return Generator
+
+# ____________________________________________________________
+
+
+class GeneratorTests(TestCase):
+    def test_generator(self):
+        seen = []
+
+        def g(n):
+            for i in range(n):
+                seen.append(i)
+                Yield(i)
+        g = generator(g)
+        for _ in range(3):
+            for j in g(5):
+                seen.append(j)
+        self.assertEqual(seen, 3 * [0, 0, 1, 1, 2, 2, 3, 3, 4, 4])
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator_nested.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator_nested.py
new file mode 100644
index 00000000..8d752a63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_generator_nested.py
@@ -0,0 +1,168 @@
+
+from greenlet import greenlet
+from . import TestCase
+from .leakcheck import fails_leakcheck
+
+class genlet(greenlet):
+    parent = None
+    def __init__(self, *args, **kwds):
+        self.args = args
+        self.kwds = kwds
+        self.child = None
+
+    def run(self):
+        # Note the function is packed in a tuple
+        # to avoid creating a bound method for it.
+        fn, = self.fn
+        fn(*self.args, **self.kwds)
+
+    def __iter__(self):
+        return self
+
+    def set_child(self, child):
+        self.child = child
+
+    def __next__(self):
+        if self.child:
+            child = self.child
+            while child.child:
+                tmp = child
+                child = child.child
+                tmp.child = None
+
+            result = child.switch()
+        else:
+            self.parent = greenlet.getcurrent()
+            result = self.switch()
+
+        if self:
+            return result
+
+        raise StopIteration
+
+    next = __next__
+
+def Yield(value, level=1):
+    g = greenlet.getcurrent()
+
+    while level != 0:
+        if not isinstance(g, genlet):
+            raise RuntimeError('yield outside a genlet')
+        if level > 1:
+            g.parent.set_child(g)
+        g = g.parent
+        level -= 1
+
+    g.switch(value)
+
+
+def Genlet(func):
+    class TheGenlet(genlet):
+        fn = (func,)
+    return TheGenlet
+
+# ____________________________________________________________
+
+
+def g1(n, seen):
+    for i in range(n):
+        seen.append(i + 1)
+        yield i
+
+
+def g2(n, seen):
+    for i in range(n):
+        seen.append(i + 1)
+        Yield(i)
+
+g2 = Genlet(g2)
+
+
+def nested(i):
+    Yield(i)
+
+
+def g3(n, seen):
+    for i in range(n):
+        seen.append(i + 1)
+        nested(i)
+g3 = Genlet(g3)
+
+
+def a(n):
+    if n == 0:
+        return
+    for ii in ax(n - 1):
+        Yield(ii)
+    Yield(n)
+ax = Genlet(a)
+
+
+def perms(l):
+    if len(l) > 1:
+        for e in l:
+            # No syntactical sugar for generator expressions
+            x = [Yield([e] + p) for p in perms([x for x in l if x != e])]
+            assert x
+    else:
+        Yield(l)
+perms = Genlet(perms)
+
+
+def gr1(n):
+    for ii in range(1, n):
+        Yield(ii)
+        Yield(ii * ii, 2)
+
+gr1 = Genlet(gr1)
+
+
+def gr2(n, seen):
+    for ii in gr1(n):
+        seen.append(ii)
+
+gr2 = Genlet(gr2)
+
+
+class NestedGeneratorTests(TestCase):
+    def test_layered_genlets(self):
+        seen = []
+        for ii in gr2(5, seen):
+            seen.append(ii)
+        self.assertEqual(seen, [1, 1, 2, 4, 3, 9, 4, 16])
+
+    @fails_leakcheck
+    def test_permutations(self):
+        gen_perms = perms(list(range(4)))
+        permutations = list(gen_perms)
+        self.assertEqual(len(permutations), 4 * 3 * 2 * 1)
+        self.assertIn([0, 1, 2, 3], permutations)
+        self.assertIn([3, 2, 1, 0], permutations)
+        res = []
+        for ii in zip(perms(list(range(4))), perms(list(range(3)))):
+            res.append(ii)
+        self.assertEqual(
+            res,
+            [([0, 1, 2, 3], [0, 1, 2]), ([0, 1, 3, 2], [0, 2, 1]),
+             ([0, 2, 1, 3], [1, 0, 2]), ([0, 2, 3, 1], [1, 2, 0]),
+             ([0, 3, 1, 2], [2, 0, 1]), ([0, 3, 2, 1], [2, 1, 0])])
+        # XXX Test to make sure we are working as a generator expression
+
+    def test_genlet_simple(self):
+        for g in g1, g2, g3:
+            seen = []
+            for _ in range(3):
+                for j in g(5, seen):
+                    seen.append(j)
+            self.assertEqual(seen, 3 * [1, 0, 2, 1, 3, 2, 4, 3, 5, 4])
+
+    def test_genlet_bad(self):
+        try:
+            Yield(10)
+        except RuntimeError:
+            pass
+
+    def test_nested_genlets(self):
+        seen = []
+        for ii in ax(5):
+            seen.append(ii)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet.py
new file mode 100644
index 00000000..c4aabea7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet.py
@@ -0,0 +1,1324 @@
+import gc
+import sys
+import time
+import threading
+import unittest
+
+from abc import ABCMeta
+from abc import abstractmethod
+
+import greenlet
+from greenlet import greenlet as RawGreenlet
+from . import TestCase
+from . import RUNNING_ON_MANYLINUX
+from . import PY313
+from .leakcheck import fails_leakcheck
+
+
+# We manually manage locks in many tests
+# pylint:disable=consider-using-with
+# pylint:disable=too-many-public-methods
+# This module is quite large.
+# TODO: Refactor into separate test files. For example,
+# put all the regression tests that used to produce
+# crashes in test_greenlet_no_crash; put tests that DO deliberately crash
+# the interpreter into test_greenlet_crash.
+# pylint:disable=too-many-lines
+
+class SomeError(Exception):
+    pass
+
+
+def fmain(seen):
+    try:
+        greenlet.getcurrent().parent.switch()
+    except:
+        seen.append(sys.exc_info()[0])
+        raise
+    raise SomeError
+
+
+def send_exception(g, exc):
+    # note: send_exception(g, exc)  can be now done with  g.throw(exc).
+    # the purpose of this test is to explicitly check the propagation rules.
+    def crasher(exc):
+        raise exc
+    g1 = RawGreenlet(crasher, parent=g)
+    g1.switch(exc)
+
+
+class TestGreenlet(TestCase):
+
+    def _do_simple_test(self):
+        lst = []
+
+        def f():
+            lst.append(1)
+            greenlet.getcurrent().parent.switch()
+            lst.append(3)
+        g = RawGreenlet(f)
+        lst.append(0)
+        g.switch()
+        lst.append(2)
+        g.switch()
+        lst.append(4)
+        self.assertEqual(lst, list(range(5)))
+
+    def test_simple(self):
+        self._do_simple_test()
+
+    def test_switch_no_run_raises_AttributeError(self):
+        g = RawGreenlet()
+        with self.assertRaises(AttributeError) as exc:
+            g.switch()
+
+        self.assertIn("run", str(exc.exception))
+
+    def test_throw_no_run_raises_AttributeError(self):
+        g = RawGreenlet()
+        with self.assertRaises(AttributeError) as exc:
+            g.throw(SomeError)
+
+        self.assertIn("run", str(exc.exception))
+
+    def test_parent_equals_None(self):
+        g = RawGreenlet(parent=None)
+        self.assertIsNotNone(g)
+        self.assertIs(g.parent, greenlet.getcurrent())
+
+    def test_run_equals_None(self):
+        g = RawGreenlet(run=None)
+        self.assertIsNotNone(g)
+        self.assertIsNone(g.run)
+
+    def test_two_children(self):
+        lst = []
+
+        def f():
+            lst.append(1)
+            greenlet.getcurrent().parent.switch()
+            lst.extend([1, 1])
+        g = RawGreenlet(f)
+        h = RawGreenlet(f)
+        g.switch()
+        self.assertEqual(len(lst), 1)
+        h.switch()
+        self.assertEqual(len(lst), 2)
+        h.switch()
+        self.assertEqual(len(lst), 4)
+        self.assertEqual(h.dead, True)
+        g.switch()
+        self.assertEqual(len(lst), 6)
+        self.assertEqual(g.dead, True)
+
+    def test_two_recursive_children(self):
+        lst = []
+
+        def f():
+            lst.append('b')
+            greenlet.getcurrent().parent.switch()
+
+        def g():
+            lst.append('a')
+            g = RawGreenlet(f)
+            g.switch()
+            lst.append('c')
+
+        g = RawGreenlet(g)
+        self.assertEqual(sys.getrefcount(g), 2)
+        g.switch()
+        self.assertEqual(lst, ['a', 'b', 'c'])
+        # Just the one in this frame, plus the one on the stack we pass to the function
+        self.assertEqual(sys.getrefcount(g), 2)
+
+    def test_threads(self):
+        success = []
+
+        def f():
+            self._do_simple_test()
+            success.append(True)
+        ths = [threading.Thread(target=f) for i in range(10)]
+        for th in ths:
+            th.start()
+        for th in ths:
+            th.join(10)
+        self.assertEqual(len(success), len(ths))
+
+    def test_exception(self):
+        seen = []
+        g1 = RawGreenlet(fmain)
+        g2 = RawGreenlet(fmain)
+        g1.switch(seen)
+        g2.switch(seen)
+        g2.parent = g1
+
+        self.assertEqual(seen, [])
+        #with self.assertRaises(SomeError):
+        #    p("***Switching back")
+        #    g2.switch()
+        # Creating this as a bound method can reveal bugs that
+        # are hidden on newer versions of Python that avoid creating
+        # bound methods for direct expressions; IOW, don't use the `with`
+        # form!
+        self.assertRaises(SomeError, g2.switch)
+        self.assertEqual(seen, [SomeError])
+
+        value = g2.switch()
+        self.assertEqual(value, ())
+        self.assertEqual(seen, [SomeError])
+
+        value = g2.switch(25)
+        self.assertEqual(value, 25)
+        self.assertEqual(seen, [SomeError])
+
+
+    def test_send_exception(self):
+        seen = []
+        g1 = RawGreenlet(fmain)
+        g1.switch(seen)
+        self.assertRaises(KeyError, send_exception, g1, KeyError)
+        self.assertEqual(seen, [KeyError])
+
+    def test_dealloc(self):
+        seen = []
+        g1 = RawGreenlet(fmain)
+        g2 = RawGreenlet(fmain)
+        g1.switch(seen)
+        g2.switch(seen)
+        self.assertEqual(seen, [])
+        del g1
+        gc.collect()
+        self.assertEqual(seen, [greenlet.GreenletExit])
+        del g2
+        gc.collect()
+        self.assertEqual(seen, [greenlet.GreenletExit, greenlet.GreenletExit])
+
+    def test_dealloc_catches_GreenletExit_throws_other(self):
+        def run():
+            try:
+                greenlet.getcurrent().parent.switch()
+            except greenlet.GreenletExit:
+                raise SomeError from None
+
+        g = RawGreenlet(run)
+        g.switch()
+        # Destroying the only reference to the greenlet causes it
+        # to get GreenletExit; when it in turn raises, even though we're the parent
+        # we don't get the exception, it just gets printed.
+        # When we run on 3.8 only, we can use sys.unraisablehook
+        oldstderr = sys.stderr
+        from io import StringIO
+        stderr = sys.stderr = StringIO()
+        try:
+            del g
+        finally:
+            sys.stderr = oldstderr
+
+        v = stderr.getvalue()
+        self.assertIn("Exception", v)
+        self.assertIn('ignored', v)
+        self.assertIn("SomeError", v)
+
+
+    @unittest.skipIf(
+        PY313 and RUNNING_ON_MANYLINUX,
+        "Sometimes flaky (getting one GreenletExit in the second list)"
+        # Probably due to funky timing interactions?
+        # TODO: FIXME Make that work.
+    )
+
+    def test_dealloc_other_thread(self):
+        seen = []
+        someref = []
+
+        bg_glet_created_running_and_no_longer_ref_in_bg = threading.Event()
+        fg_ref_released = threading.Event()
+        bg_should_be_clear = threading.Event()
+        ok_to_exit_bg_thread = threading.Event()
+
+        def f():
+            g1 = RawGreenlet(fmain)
+            g1.switch(seen)
+            someref.append(g1)
+            del g1
+            gc.collect()
+
+            bg_glet_created_running_and_no_longer_ref_in_bg.set()
+            fg_ref_released.wait(3)
+
+            RawGreenlet()   # trigger release
+            bg_should_be_clear.set()
+            ok_to_exit_bg_thread.wait(3)
+            RawGreenlet() # One more time
+
+        t = threading.Thread(target=f)
+        t.start()
+        bg_glet_created_running_and_no_longer_ref_in_bg.wait(10)
+
+        self.assertEqual(seen, [])
+        self.assertEqual(len(someref), 1)
+        del someref[:]
+        gc.collect()
+        # g1 is not released immediately because it's from another thread
+        self.assertEqual(seen, [])
+        fg_ref_released.set()
+        bg_should_be_clear.wait(3)
+        try:
+            self.assertEqual(seen, [greenlet.GreenletExit])
+        finally:
+            ok_to_exit_bg_thread.set()
+            t.join(10)
+            del seen[:]
+            del someref[:]
+
+    def test_frame(self):
+        def f1():
+            f = sys._getframe(0) # pylint:disable=protected-access
+            self.assertEqual(f.f_back, None)
+            greenlet.getcurrent().parent.switch(f)
+            return "meaning of life"
+        g = RawGreenlet(f1)
+        frame = g.switch()
+        self.assertTrue(frame is g.gr_frame)
+        self.assertTrue(g)
+
+        from_g = g.switch()
+        self.assertFalse(g)
+        self.assertEqual(from_g, 'meaning of life')
+        self.assertEqual(g.gr_frame, None)
+
+    def test_thread_bug(self):
+        def runner(x):
+            g = RawGreenlet(lambda: time.sleep(x))
+            g.switch()
+        t1 = threading.Thread(target=runner, args=(0.2,))
+        t2 = threading.Thread(target=runner, args=(0.3,))
+        t1.start()
+        t2.start()
+        t1.join(10)
+        t2.join(10)
+
+    def test_switch_kwargs(self):
+        def run(a, b):
+            self.assertEqual(a, 4)
+            self.assertEqual(b, 2)
+            return 42
+        x = RawGreenlet(run).switch(a=4, b=2)
+        self.assertEqual(x, 42)
+
+    def test_switch_kwargs_to_parent(self):
+        def run(x):
+            greenlet.getcurrent().parent.switch(x=x)
+            greenlet.getcurrent().parent.switch(2, x=3)
+            return x, x ** 2
+        g = RawGreenlet(run)
+        self.assertEqual({'x': 3}, g.switch(3))
+        self.assertEqual(((2,), {'x': 3}), g.switch())
+        self.assertEqual((3, 9), g.switch())
+
+    def test_switch_to_another_thread(self):
+        data = {}
+        created_event = threading.Event()
+        done_event = threading.Event()
+
+        def run():
+            data['g'] = RawGreenlet(lambda: None)
+            created_event.set()
+            done_event.wait(10)
+        thread = threading.Thread(target=run)
+        thread.start()
+        created_event.wait(10)
+        with self.assertRaises(greenlet.error):
+            data['g'].switch()
+        done_event.set()
+        thread.join(10)
+        # XXX: Should handle this automatically
+        data.clear()
+
+    def test_exc_state(self):
+        def f():
+            try:
+                raise ValueError('fun')
+            except: # pylint:disable=bare-except
+                exc_info = sys.exc_info()
+                RawGreenlet(h).switch()
+                self.assertEqual(exc_info, sys.exc_info())
+
+        def h():
+            self.assertEqual(sys.exc_info(), (None, None, None))
+
+        RawGreenlet(f).switch()
+
+    def test_instance_dict(self):
+        def f():
+            greenlet.getcurrent().test = 42
+        def deldict(g):
+            del g.__dict__
+        def setdict(g, value):
+            g.__dict__ = value
+        g = RawGreenlet(f)
+        self.assertEqual(g.__dict__, {})
+        g.switch()
+        self.assertEqual(g.test, 42)
+        self.assertEqual(g.__dict__, {'test': 42})
+        g.__dict__ = g.__dict__
+        self.assertEqual(g.__dict__, {'test': 42})
+        self.assertRaises(TypeError, deldict, g)
+        self.assertRaises(TypeError, setdict, g, 42)
+
+    def test_running_greenlet_has_no_run(self):
+        has_run = []
+        def func():
+            has_run.append(
+                hasattr(greenlet.getcurrent(), 'run')
+            )
+
+        g = RawGreenlet(func)
+        g.switch()
+        self.assertEqual(has_run, [False])
+
+    def test_deepcopy(self):
+        import copy
+        self.assertRaises(TypeError, copy.copy, RawGreenlet())
+        self.assertRaises(TypeError, copy.deepcopy, RawGreenlet())
+
+    def test_parent_restored_on_kill(self):
+        hub = RawGreenlet(lambda: None)
+        main = greenlet.getcurrent()
+        result = []
+        def worker():
+            try:
+                # Wait to be killed by going back to the test.
+                main.switch()
+            except greenlet.GreenletExit:
+                # Resurrect and switch to parent
+                result.append(greenlet.getcurrent().parent)
+                result.append(greenlet.getcurrent())
+                hub.switch()
+        g = RawGreenlet(worker, parent=hub)
+        g.switch()
+        # delete the only reference, thereby raising GreenletExit
+        del g
+        self.assertTrue(result)
+        self.assertIs(result[0], main)
+        self.assertIs(result[1].parent, hub)
+        # Delete them, thereby breaking the cycle between the greenlet
+        # and the frame, which otherwise would never be collectable
+        # XXX: We should be able to automatically fix this.
+        del result[:]
+        hub = None
+        main = None
+
+    def test_parent_return_failure(self):
+        # No run causes AttributeError on switch
+        g1 = RawGreenlet()
+        # Greenlet that implicitly switches to parent
+        g2 = RawGreenlet(lambda: None, parent=g1)
+        # AttributeError should propagate to us, no fatal errors
+        with self.assertRaises(AttributeError):
+            g2.switch()
+
+    def test_throw_exception_not_lost(self):
+        class mygreenlet(RawGreenlet):
+            def __getattribute__(self, name):
+                try:
+                    raise Exception # pylint:disable=broad-exception-raised
+                except: # pylint:disable=bare-except
+                    pass
+                return RawGreenlet.__getattribute__(self, name)
+        g = mygreenlet(lambda: None)
+        self.assertRaises(SomeError, g.throw, SomeError())
+
+    @fails_leakcheck
+    def _do_test_throw_to_dead_thread_doesnt_crash(self, wait_for_cleanup=False):
+        result = []
+        def worker():
+            greenlet.getcurrent().parent.switch()
+
+        def creator():
+            g = RawGreenlet(worker)
+            g.switch()
+            result.append(g)
+            if wait_for_cleanup:
+                # Let this greenlet eventually be cleaned up.
+                g.switch()
+                greenlet.getcurrent()
+        t = threading.Thread(target=creator)
+        t.start()
+        t.join(10)
+        del t
+        # But, depending on the operating system, the thread
+        # deallocator may not actually have run yet! So we can't be
+        # sure about the error message unless we wait.
+        if wait_for_cleanup:
+            self.wait_for_pending_cleanups()
+        with self.assertRaises(greenlet.error) as exc:
+            result[0].throw(SomeError)
+
+        if not wait_for_cleanup:
+            s = str(exc.exception)
+            self.assertTrue(
+                s == "cannot switch to a different thread (which happens to have exited)"
+                or 'Cannot switch' in s
+            )
+        else:
+            self.assertEqual(
+                str(exc.exception),
+                "cannot switch to a different thread (which happens to have exited)",
+            )
+
+        if hasattr(result[0].gr_frame, 'clear'):
+            # The frame is actually executing (it thinks), we can't clear it.
+            with self.assertRaises(RuntimeError):
+                result[0].gr_frame.clear()
+        # Unfortunately, this doesn't actually clear the references, they're in the
+        # fast local array.
+        if not wait_for_cleanup:
+            # f_locals has no clear method in Python 3.13
+            if hasattr(result[0].gr_frame.f_locals, 'clear'):
+                result[0].gr_frame.f_locals.clear()
+        else:
+            self.assertIsNone(result[0].gr_frame)
+
+        del creator
+        worker = None
+        del result[:]
+        # XXX: we ought to be able to automatically fix this.
+        # See issue 252
+        self.expect_greenlet_leak = True # direct us not to wait for it to go away
+
+    @fails_leakcheck
+    def test_throw_to_dead_thread_doesnt_crash(self):
+        self._do_test_throw_to_dead_thread_doesnt_crash()
+
+    def test_throw_to_dead_thread_doesnt_crash_wait(self):
+        self._do_test_throw_to_dead_thread_doesnt_crash(True)
+
+    @fails_leakcheck
+    def test_recursive_startup(self):
+        class convoluted(RawGreenlet):
+            def __init__(self):
+                RawGreenlet.__init__(self)
+                self.count = 0
+            def __getattribute__(self, name):
+                if name == 'run' and self.count == 0:
+                    self.count = 1
+                    self.switch(43)
+                return RawGreenlet.__getattribute__(self, name)
+            def run(self, value):
+                while True:
+                    self.parent.switch(value)
+        g = convoluted()
+        self.assertEqual(g.switch(42), 43)
+        # Exits the running greenlet, otherwise it leaks
+        # XXX: We should be able to automatically fix this
+        #g.throw(greenlet.GreenletExit)
+        #del g
+        self.expect_greenlet_leak = True
+
+    def test_threaded_updatecurrent(self):
+        # released when main thread should execute
+        lock1 = threading.Lock()
+        lock1.acquire()
+        # released when another thread should execute
+        lock2 = threading.Lock()
+        lock2.acquire()
+        class finalized(object):
+            def __del__(self):
+                # happens while in green_updatecurrent() in main greenlet
+                # should be very careful not to accidentally call it again
+                # at the same time we must make sure another thread executes
+                lock2.release()
+                lock1.acquire()
+                # now ts_current belongs to another thread
+        def deallocator():
+            greenlet.getcurrent().parent.switch()
+        def fthread():
+            lock2.acquire()
+            greenlet.getcurrent()
+            del g[0]
+            lock1.release()
+            lock2.acquire()
+            greenlet.getcurrent()
+            lock1.release()
+        main = greenlet.getcurrent()
+        g = [RawGreenlet(deallocator)]
+        g[0].bomb = finalized()
+        g[0].switch()
+        t = threading.Thread(target=fthread)
+        t.start()
+        # let another thread grab ts_current and deallocate g[0]
+        lock2.release()
+        lock1.acquire()
+        # this is the corner stone
+        # getcurrent() will notice that ts_current belongs to another thread
+        # and start the update process, which would notice that g[0] should
+        # be deallocated, and that will execute an object's finalizer. Now,
+        # that object will let another thread run so it can grab ts_current
+        # again, which would likely crash the interpreter if there's no
+        # check for this case at the end of green_updatecurrent(). This test
+        # passes if getcurrent() returns correct result, but it's likely
+        # to randomly crash if it's not anyway.
+        self.assertEqual(greenlet.getcurrent(), main)
+        # wait for another thread to complete, just in case
+        t.join(10)
+
+    def test_dealloc_switch_args_not_lost(self):
+        seen = []
+        def worker():
+            # wait for the value
+            value = greenlet.getcurrent().parent.switch()
+            # delete all references to ourself
+            del worker[0]
+            initiator.parent = greenlet.getcurrent().parent
+            # switch to main with the value, but because
+            # ts_current is the last reference to us we
+            # return here immediately, where we resurrect ourself.
+            try:
+                greenlet.getcurrent().parent.switch(value)
+            finally:
+                seen.append(greenlet.getcurrent())
+        def initiator():
+            return 42 # implicitly falls thru to parent
+
+        worker = [RawGreenlet(worker)]
+
+        worker[0].switch() # prime worker
+        initiator = RawGreenlet(initiator, worker[0])
+        value = initiator.switch()
+        self.assertTrue(seen)
+        self.assertEqual(value, 42)
+
+    def test_tuple_subclass(self):
+        # The point of this test is to see what happens when a custom
+        # tuple subclass is used as an object passed directly to the C
+        # function ``green_switch``; part of ``green_switch`` checks
+        # the ``len()`` of the ``args`` tuple, and that can call back
+        # into Python. Here, when it calls back into Python, we
+        # recursively enter ``green_switch`` again.
+
+        # This test is really only relevant on Python 2. The builtin
+        # `apply` function directly passes the given args tuple object
+        # to the underlying function, whereas the Python 3 version
+        # unpacks and repacks into an actual tuple. This could still
+        # happen using the C API on Python 3 though. We should write a
+        # builtin version of apply() ourself.
+        def _apply(func, a, k):
+            func(*a, **k)
+
+        class mytuple(tuple):
+            def __len__(self):
+                greenlet.getcurrent().switch()
+                return tuple.__len__(self)
+        args = mytuple()
+        kwargs = dict(a=42)
+        def switchapply():
+            _apply(greenlet.getcurrent().parent.switch, args, kwargs)
+        g = RawGreenlet(switchapply)
+        self.assertEqual(g.switch(), kwargs)
+
+    def test_abstract_subclasses(self):
+        AbstractSubclass = ABCMeta(
+            'AbstractSubclass',
+            (RawGreenlet,),
+            {'run': abstractmethod(lambda self: None)})
+
+        class BadSubclass(AbstractSubclass):
+            pass
+
+        class GoodSubclass(AbstractSubclass):
+            def run(self):
+                pass
+
+        GoodSubclass() # should not raise
+        self.assertRaises(TypeError, BadSubclass)
+
+    def test_implicit_parent_with_threads(self):
+        if not gc.isenabled():
+            return # cannot test with disabled gc
+        N = gc.get_threshold()[0]
+        if N < 50:
+            return # cannot test with such a small N
+        def attempt():
+            lock1 = threading.Lock()
+            lock1.acquire()
+            lock2 = threading.Lock()
+            lock2.acquire()
+            recycled = [False]
+            def another_thread():
+                lock1.acquire() # wait for gc
+                greenlet.getcurrent() # update ts_current
+                lock2.release() # release gc
+            t = threading.Thread(target=another_thread)
+            t.start()
+            class gc_callback(object):
+                def __del__(self):
+                    lock1.release()
+                    lock2.acquire()
+                    recycled[0] = True
+            class garbage(object):
+                def __init__(self):
+                    self.cycle = self
+                    self.callback = gc_callback()
+            l = []
+            x = range(N*2)
+            current = greenlet.getcurrent()
+            g = garbage()
+            for _ in x:
+                g = None # lose reference to garbage
+                if recycled[0]:
+                    # gc callback called prematurely
+                    t.join(10)
+                    return False
+                last = RawGreenlet()
+                if recycled[0]:
+                    break # yes! gc called in green_new
+                l.append(last) # increase allocation counter
+            else:
+                # gc callback not called when expected
+                gc.collect()
+                if recycled[0]:
+                    t.join(10)
+                return False
+            self.assertEqual(last.parent, current)
+            for g in l:
+                self.assertEqual(g.parent, current)
+            return True
+        for _ in range(5):
+            if attempt():
+                break
+
+    def test_issue_245_reference_counting_subclass_no_threads(self):
+        # https://github.com/python-greenlet/greenlet/issues/245
+        # Before the fix, this crashed pretty reliably on
+        # Python 3.10, at least on macOS; but much less reliably on other
+        # interpreters (memory layout must have changed).
+        # The threaded test crashed more reliably on more interpreters.
+        from greenlet import getcurrent
+        from greenlet import GreenletExit
+
+        class Greenlet(RawGreenlet):
+            pass
+
+        initial_refs = sys.getrefcount(Greenlet)
+        # This has to be an instance variable because
+        # Python 2 raises a SyntaxError if we delete a local
+        # variable referenced in an inner scope.
+        self.glets = [] # pylint:disable=attribute-defined-outside-init
+
+        def greenlet_main():
+            try:
+                getcurrent().parent.switch()
+            except GreenletExit:
+                self.glets.append(getcurrent())
+
+        # Before the
+        for _ in range(10):
+            Greenlet(greenlet_main).switch()
+
+        del self.glets
+        self.assertEqual(sys.getrefcount(Greenlet), initial_refs)
+
+    @unittest.skipIf(
+        PY313 and RUNNING_ON_MANYLINUX,
+        "The manylinux images appear to hang on this test on 3.13rc2"
+        # Or perhaps I just got tired of waiting for the 450s timeout.
+        # Still, it shouldn't take anywhere near that long. Does not reproduce in
+        # Ubuntu images, on macOS or Windows.
+    )
+    def test_issue_245_reference_counting_subclass_threads(self):
+        # https://github.com/python-greenlet/greenlet/issues/245
+        from threading import Thread
+        from threading import Event
+
+        from greenlet import getcurrent
+
+        class MyGreenlet(RawGreenlet):
+            pass
+
+        glets = []
+        ref_cleared = Event()
+
+        def greenlet_main():
+            getcurrent().parent.switch()
+
+        def thread_main(greenlet_running_event):
+            mine = MyGreenlet(greenlet_main)
+            glets.append(mine)
+            # The greenlets being deleted must be active
+            mine.switch()
+            # Don't keep any reference to it in this thread
+            del mine
+            # Let main know we published our greenlet.
+            greenlet_running_event.set()
+            # Wait for main to let us know the references are
+            # gone and the greenlet objects no longer reachable
+            ref_cleared.wait(10)
+            # The creating thread must call getcurrent() (or a few other
+            # greenlet APIs) because that's when the thread-local list of dead
+            # greenlets gets cleared.
+            getcurrent()
+
+        # We start with 3 references to the subclass:
+        # - This module
+        # - Its __mro__
+        # - The __subclassess__ attribute of greenlet
+        # - (If we call gc.get_referents(), we find four entries, including
+        #   some other tuple ``(greenlet)`` that I'm not sure about but must be part
+        #   of the machinery.)
+        #
+        # On Python 3.10 it's often enough to just run 3 threads; on Python 2.7,
+        # more threads are needed, and the results are still
+        # non-deterministic. Presumably the memory layouts are different
+        initial_refs = sys.getrefcount(MyGreenlet)
+        thread_ready_events = []
+        for _ in range(
+                initial_refs + 45
+        ):
+            event = Event()
+            thread = Thread(target=thread_main, args=(event,))
+            thread_ready_events.append(event)
+            thread.start()
+
+
+        for done_event in thread_ready_events:
+            done_event.wait(10)
+
+
+        del glets[:]
+        ref_cleared.set()
+        # Let any other thread run; it will crash the interpreter
+        # if not fixed (or silently corrupt memory and we possibly crash
+        # later).
+        self.wait_for_pending_cleanups()
+        self.assertEqual(sys.getrefcount(MyGreenlet), initial_refs)
+
+    def test_falling_off_end_switches_to_unstarted_parent_raises_error(self):
+        def no_args():
+            return 13
+
+        parent_never_started = RawGreenlet(no_args)
+
+        def leaf():
+            return 42
+
+        child = RawGreenlet(leaf, parent_never_started)
+
+        # Because the run function takes to arguments
+        with self.assertRaises(TypeError):
+            child.switch()
+
+    def test_falling_off_end_switches_to_unstarted_parent_works(self):
+        def one_arg(x):
+            return (x, 24)
+
+        parent_never_started = RawGreenlet(one_arg)
+
+        def leaf():
+            return 42
+
+        child = RawGreenlet(leaf, parent_never_started)
+
+        result = child.switch()
+        self.assertEqual(result, (42, 24))
+
+    def test_switch_to_dead_greenlet_with_unstarted_perverse_parent(self):
+        class Parent(RawGreenlet):
+            def __getattribute__(self, name):
+                if name == 'run':
+                    raise SomeError
+
+
+        parent_never_started = Parent()
+        seen = []
+        child = RawGreenlet(lambda: seen.append(42), parent_never_started)
+        # Because we automatically start the parent when the child is
+        # finished
+        with self.assertRaises(SomeError):
+            child.switch()
+
+        self.assertEqual(seen, [42])
+
+        with self.assertRaises(SomeError):
+            child.switch()
+        self.assertEqual(seen, [42])
+
+    def test_switch_to_dead_greenlet_reparent(self):
+        seen = []
+        parent_never_started = RawGreenlet(lambda: seen.append(24))
+        child = RawGreenlet(lambda: seen.append(42))
+
+        child.switch()
+        self.assertEqual(seen, [42])
+
+        child.parent = parent_never_started
+        # This actually is the same as switching to the parent.
+        result = child.switch()
+        self.assertIsNone(result)
+        self.assertEqual(seen, [42, 24])
+
+    def test_can_access_f_back_of_suspended_greenlet(self):
+        # This tests our frame rewriting to work around Python 3.12+ having
+        # some interpreter frames on the C stack. It will crash in the absence
+        # of that logic.
+        main = greenlet.getcurrent()
+
+        def outer():
+            inner()
+
+        def inner():
+            main.switch(sys._getframe(0))
+
+        hub = RawGreenlet(outer)
+        # start it
+        hub.switch()
+
+        # start another greenlet to make sure we aren't relying on
+        # anything in `hub` still being on the C stack
+        unrelated = RawGreenlet(lambda: None)
+        unrelated.switch()
+
+        # now it is suspended
+        self.assertIsNotNone(hub.gr_frame)
+        self.assertEqual(hub.gr_frame.f_code.co_name, "inner")
+        self.assertIsNotNone(hub.gr_frame.f_back)
+        self.assertEqual(hub.gr_frame.f_back.f_code.co_name, "outer")
+        # The next line is what would crash
+        self.assertIsNone(hub.gr_frame.f_back.f_back)
+
+    def test_get_stack_with_nested_c_calls(self):
+        from functools import partial
+        from . import _test_extension_cpp
+
+        def recurse(v):
+            if v > 0:
+                return v * _test_extension_cpp.test_call(partial(recurse, v - 1))
+            return greenlet.getcurrent().parent.switch()
+
+        gr = RawGreenlet(recurse)
+        gr.switch(5)
+        frame = gr.gr_frame
+        for i in range(5):
+            self.assertEqual(frame.f_locals["v"], i)
+            frame = frame.f_back
+        self.assertEqual(frame.f_locals["v"], 5)
+        self.assertIsNone(frame.f_back)
+        self.assertEqual(gr.switch(10), 1200)  # 1200 = 5! * 10
+
+    def test_frames_always_exposed(self):
+        # On Python 3.12 this will crash if we don't set the
+        # gr_frames_always_exposed attribute. More background:
+        # https://github.com/python-greenlet/greenlet/issues/388
+        main = greenlet.getcurrent()
+
+        def outer():
+            inner(sys._getframe(0))
+
+        def inner(frame):
+            main.switch(frame)
+
+        gr = RawGreenlet(outer)
+        frame = gr.switch()
+
+        # Do something else to clobber the part of the C stack used by `gr`,
+        # so we can't skate by on "it just happened to still be there"
+        unrelated = RawGreenlet(lambda: None)
+        unrelated.switch()
+
+        self.assertEqual(frame.f_code.co_name, "outer")
+        # The next line crashes on 3.12 if we haven't exposed the frames.
+        self.assertIsNone(frame.f_back)
+
+
+class TestGreenletSetParentErrors(TestCase):
+    def test_threaded_reparent(self):
+        data = {}
+        created_event = threading.Event()
+        done_event = threading.Event()
+
+        def run():
+            data['g'] = RawGreenlet(lambda: None)
+            created_event.set()
+            done_event.wait(10)
+
+        def blank():
+            greenlet.getcurrent().parent.switch()
+
+        thread = threading.Thread(target=run)
+        thread.start()
+        created_event.wait(10)
+        g = RawGreenlet(blank)
+        g.switch()
+        with self.assertRaises(ValueError) as exc:
+            g.parent = data['g']
+        done_event.set()
+        thread.join(10)
+
+        self.assertEqual(str(exc.exception), "parent cannot be on a different thread")
+
+    def test_unexpected_reparenting(self):
+        another = []
+        def worker():
+            g = RawGreenlet(lambda: None)
+            another.append(g)
+            g.switch()
+        t = threading.Thread(target=worker)
+        t.start()
+        t.join(10)
+        # The first time we switch (running g_initialstub(), which is
+        # when we look up the run attribute) we attempt to change the
+        # parent to one from another thread (which also happens to be
+        # dead). ``g_initialstub()`` should detect this and raise a
+        # greenlet error.
+        #
+        # EXCEPT: With the fix for #252, this is actually detected
+        # sooner, when setting the parent itself. Prior to that fix,
+        # the main greenlet from the background thread kept a valid
+        # value for ``run_info``, and appeared to be a valid parent
+        # until we actually started the greenlet. But now that it's
+        # cleared, this test is catching whether ``green_setparent``
+        # can detect the dead thread.
+        #
+        # Further refactoring once again changes this back to a greenlet.error
+        #
+        # We need to wait for the cleanup to happen, but we're
+        # deliberately leaking a main greenlet here.
+        self.wait_for_pending_cleanups(initial_main_greenlets=self.main_greenlets_before_test + 1)
+
+        class convoluted(RawGreenlet):
+            def __getattribute__(self, name):
+                if name == 'run':
+                    self.parent = another[0] # pylint:disable=attribute-defined-outside-init
+                return RawGreenlet.__getattribute__(self, name)
+        g = convoluted(lambda: None)
+        with self.assertRaises(greenlet.error) as exc:
+            g.switch()
+        self.assertEqual(str(exc.exception),
+                         "cannot switch to a different thread (which happens to have exited)")
+        del another[:]
+
+    def test_unexpected_reparenting_thread_running(self):
+        # Like ``test_unexpected_reparenting``, except the background thread is
+        # actually still alive.
+        another = []
+        switched_to_greenlet = threading.Event()
+        keep_main_alive = threading.Event()
+        def worker():
+            g = RawGreenlet(lambda: None)
+            another.append(g)
+            g.switch()
+            switched_to_greenlet.set()
+            keep_main_alive.wait(10)
+        class convoluted(RawGreenlet):
+            def __getattribute__(self, name):
+                if name == 'run':
+                    self.parent = another[0] # pylint:disable=attribute-defined-outside-init
+                return RawGreenlet.__getattribute__(self, name)
+
+        t = threading.Thread(target=worker)
+        t.start()
+
+        switched_to_greenlet.wait(10)
+        try:
+            g = convoluted(lambda: None)
+
+            with self.assertRaises(greenlet.error) as exc:
+                g.switch()
+            self.assertIn("Cannot switch to a different thread", str(exc.exception))
+            self.assertIn("Expected", str(exc.exception))
+            self.assertIn("Current", str(exc.exception))
+        finally:
+            keep_main_alive.set()
+            t.join(10)
+            # XXX: Should handle this automatically.
+            del another[:]
+
+    def test_cannot_delete_parent(self):
+        worker = RawGreenlet(lambda: None)
+        self.assertIs(worker.parent, greenlet.getcurrent())
+
+        with self.assertRaises(AttributeError) as exc:
+            del worker.parent
+        self.assertEqual(str(exc.exception), "can't delete attribute")
+
+    def test_cannot_delete_parent_of_main(self):
+        with self.assertRaises(AttributeError) as exc:
+            del greenlet.getcurrent().parent
+        self.assertEqual(str(exc.exception), "can't delete attribute")
+
+
+    def test_main_greenlet_parent_is_none(self):
+        # assuming we're in a main greenlet here.
+        self.assertIsNone(greenlet.getcurrent().parent)
+
+    def test_set_parent_wrong_types(self):
+        def bg():
+            # Go back to main.
+            greenlet.getcurrent().parent.switch()
+
+        def check(glet):
+            for p in None, 1, self, "42":
+                with self.assertRaises(TypeError) as exc:
+                    glet.parent = p
+
+                self.assertEqual(
+                    str(exc.exception),
+                    "GreenletChecker: Expected any type of greenlet, not " + type(p).__name__)
+
+        # First, not running
+        g = RawGreenlet(bg)
+        self.assertFalse(g)
+        check(g)
+
+        # Then when running.
+        g.switch()
+        self.assertTrue(g)
+        check(g)
+
+        # Let it finish
+        g.switch()
+
+
+    def test_trivial_cycle(self):
+        glet = RawGreenlet(lambda: None)
+        with self.assertRaises(ValueError) as exc:
+            glet.parent = glet
+        self.assertEqual(str(exc.exception), "cyclic parent chain")
+
+    def test_trivial_cycle_main(self):
+        # This used to produce a ValueError, but we catch it earlier than that now.
+        with self.assertRaises(AttributeError) as exc:
+            greenlet.getcurrent().parent = greenlet.getcurrent()
+        self.assertEqual(str(exc.exception), "cannot set the parent of a main greenlet")
+
+    def test_deeper_cycle(self):
+        g1 = RawGreenlet(lambda: None)
+        g2 = RawGreenlet(lambda: None)
+        g3 = RawGreenlet(lambda: None)
+
+        g1.parent = g2
+        g2.parent = g3
+        with self.assertRaises(ValueError) as exc:
+            g3.parent = g1
+        self.assertEqual(str(exc.exception), "cyclic parent chain")
+
+
+class TestRepr(TestCase):
+
+    def assertEndsWith(self, got, suffix):
+        self.assertTrue(got.endswith(suffix), (got, suffix))
+
+    def test_main_while_running(self):
+        r = repr(greenlet.getcurrent())
+        self.assertEndsWith(r, " current active started main>")
+
+    def test_main_in_background(self):
+        main = greenlet.getcurrent()
+        def run():
+            return repr(main)
+
+        g = RawGreenlet(run)
+        r = g.switch()
+        self.assertEndsWith(r, ' suspended active started main>')
+
+    def test_initial(self):
+        r = repr(RawGreenlet())
+        self.assertEndsWith(r, ' pending>')
+
+    def test_main_from_other_thread(self):
+        main = greenlet.getcurrent()
+
+        class T(threading.Thread):
+            original_main = thread_main = None
+            main_glet = None
+            def run(self):
+                self.original_main = repr(main)
+                self.main_glet = greenlet.getcurrent()
+                self.thread_main = repr(self.main_glet)
+
+        t = T()
+        t.start()
+        t.join(10)
+
+        self.assertEndsWith(t.original_main, ' suspended active started main>')
+        self.assertEndsWith(t.thread_main, ' current active started main>')
+        # give the machinery time to notice the death of the thread,
+        # and clean it up. Note that we don't use
+        # ``expect_greenlet_leak`` or wait_for_pending_cleanups,
+        # because at this point we know we have an extra greenlet
+        # still reachable.
+        for _ in range(3):
+            time.sleep(0.001)
+
+        # In the past, main greenlets, even from dead threads, never
+        # really appear dead. We have fixed that, and we also report
+        # that the thread is dead in the repr. (Do this multiple times
+        # to make sure that we don't self-modify and forget our state
+        # in the C++ code).
+        for _ in range(3):
+            self.assertTrue(t.main_glet.dead)
+            r = repr(t.main_glet)
+            self.assertEndsWith(r, ' (thread exited) dead>')
+
+    def test_dead(self):
+        g = RawGreenlet(lambda: None)
+        g.switch()
+        self.assertEndsWith(repr(g), ' dead>')
+        self.assertNotIn('suspended', repr(g))
+        self.assertNotIn('started', repr(g))
+        self.assertNotIn('active', repr(g))
+
+    def test_formatting_produces_native_str(self):
+        # https://github.com/python-greenlet/greenlet/issues/218
+        # %s formatting on Python 2 was producing unicode, not str.
+
+        g_dead = RawGreenlet(lambda: None)
+        g_not_started = RawGreenlet(lambda: None)
+        g_cur = greenlet.getcurrent()
+
+        for g in g_dead, g_not_started, g_cur:
+
+            self.assertIsInstance(
+                '%s' % (g,),
+                str
+            )
+            self.assertIsInstance(
+                '%r' % (g,),
+                str,
+            )
+
+
+class TestMainGreenlet(TestCase):
+    # Tests some implementation details, and relies on some
+    # implementation details.
+
+    def _check_current_is_main(self):
+        # implementation detail
+        assert 'main' in repr(greenlet.getcurrent())
+
+        t = type(greenlet.getcurrent())
+        assert 'main' not in repr(t)
+        return t
+
+    def test_main_greenlet_type_can_be_subclassed(self):
+        main_type = self._check_current_is_main()
+        subclass = type('subclass', (main_type,), {})
+        self.assertIsNotNone(subclass)
+
+    def test_main_greenlet_is_greenlet(self):
+        self._check_current_is_main()
+        self.assertIsInstance(greenlet.getcurrent(), RawGreenlet)
+
+
+
+class TestBrokenGreenlets(TestCase):
+    # Tests for things that used to, or still do, terminate the interpreter.
+    # This often means doing unsavory things.
+
+    def test_failed_to_initialstub(self):
+        def func():
+            raise AssertionError("Never get here")
+
+
+        g = greenlet._greenlet.UnswitchableGreenlet(func)
+        g.force_switch_error = True
+
+        with self.assertRaisesRegex(SystemError,
+                                    "Failed to switch stacks into a greenlet for the first time."):
+            g.switch()
+
+    def test_failed_to_switch_into_running(self):
+        runs = []
+        def func():
+            runs.append(1)
+            greenlet.getcurrent().parent.switch()
+            runs.append(2)
+            greenlet.getcurrent().parent.switch()
+            runs.append(3) # pragma: no cover
+
+        g = greenlet._greenlet.UnswitchableGreenlet(func)
+        g.switch()
+        self.assertEqual(runs, [1])
+        g.switch()
+        self.assertEqual(runs, [1, 2])
+        g.force_switch_error = True
+
+        with self.assertRaisesRegex(SystemError,
+                                    "Failed to switch stacks into a running greenlet."):
+            g.switch()
+
+        # If we stopped here, we would fail the leakcheck, because we've left
+        # the ``inner_bootstrap()`` C frame and its descendents hanging around,
+        # which have a bunch of Python references. They'll never get cleaned up
+        # if we don't let the greenlet finish.
+        g.force_switch_error = False
+        g.switch()
+        self.assertEqual(runs, [1, 2, 3])
+
+    def test_failed_to_slp_switch_into_running(self):
+        ex = self.assertScriptRaises('fail_slp_switch.py')
+
+        self.assertIn('fail_slp_switch is running', ex.output)
+        self.assertIn(ex.returncode, self.get_expected_returncodes_for_aborted_process())
+
+    def test_reentrant_switch_two_greenlets(self):
+        # Before we started capturing the arguments in g_switch_finish, this could crash.
+        output = self.run_script('fail_switch_two_greenlets.py')
+        self.assertIn('In g1_run', output)
+        self.assertIn('TRACE', output)
+        self.assertIn('LEAVE TRACE', output)
+        self.assertIn('Falling off end of main', output)
+        self.assertIn('Falling off end of g1_run', output)
+        self.assertIn('Falling off end of g2', output)
+
+    def test_reentrant_switch_three_greenlets(self):
+        # On debug builds of greenlet, this used to crash with an assertion error;
+        # on non-debug versions, it ran fine (which it should not do!).
+        # Now it always crashes correctly with a TypeError
+        ex = self.assertScriptRaises('fail_switch_three_greenlets.py', exitcodes=(1,))
+
+        self.assertIn('TypeError', ex.output)
+        self.assertIn('positional arguments', ex.output)
+
+    def test_reentrant_switch_three_greenlets2(self):
+        # This actually passed on debug and non-debug builds. It
+        # should probably have been triggering some debug assertions
+        # but it didn't.
+        #
+        # I think the fixes for the above test also kicked in here.
+        output = self.run_script('fail_switch_three_greenlets2.py')
+        self.assertIn(
+            "RESULTS: [('trace', 'switch'), "
+            "('trace', 'switch'), ('g2 arg', 'g2 from tracefunc'), "
+            "('trace', 'switch'), ('main g1', 'from g2_run'), ('trace', 'switch'), "
+            "('g1 arg', 'g1 from main'), ('trace', 'switch'), ('main g2', 'from g1_run'), "
+            "('trace', 'switch'), ('g1 from parent', 'g1 from main 2'), ('trace', 'switch'), "
+            "('main g1.2', 'g1 done'), ('trace', 'switch'), ('g2 from parent', ()), "
+            "('trace', 'switch'), ('main g2.2', 'g2 done')]",
+            output
+        )
+
+    def test_reentrant_switch_GreenletAlreadyStartedInPython(self):
+        output = self.run_script('fail_initialstub_already_started.py')
+
+        self.assertIn(
+            "RESULTS: ['Begin C', 'Switch to b from B.__getattribute__ in C', "
+            "('Begin B', ()), '_B_run switching to main', ('main from c', 'From B'), "
+            "'B.__getattribute__ back from main in C', ('Begin A', (None,)), "
+            "('A dead?', True, 'B dead?', True, 'C dead?', False), "
+            "'C done', ('main from c.2', None)]",
+            output
+        )
+
+    def test_reentrant_switch_run_callable_has_del(self):
+        output = self.run_script('fail_clearing_run_switches.py')
+        self.assertIn(
+             "RESULTS ["
+            "('G.__getattribute__', 'run'), ('RunCallable', '__del__'), "
+            "('main: g.switch()', 'from RunCallable'), ('run_func', 'enter')"
+            "]",
+            output
+        )
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet_trash.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet_trash.py
new file mode 100644
index 00000000..c1fc1374
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_greenlet_trash.py
@@ -0,0 +1,187 @@
+# -*- coding: utf-8 -*-
+"""
+Tests for greenlets interacting with the CPython trash can API.
+
+The CPython trash can API is not designed to be re-entered from a
+single thread. But this can happen using greenlets, if something
+during the object deallocation process switches greenlets, and this second
+greenlet then causes the trash can to get entered again. Here, we do this
+very explicitly, but in other cases (like gevent) it could be arbitrarily more
+complicated: for example, a weakref callback might try to acquire a lock that's
+already held by another greenlet; that would allow a greenlet switch to occur.
+
+See https://github.com/gevent/gevent/issues/1909
+
+This test is fragile and relies on details of the CPython
+implementation (like most of the rest of this package):
+
+    - We enter the trashcan and deferred deallocation after
+      ``_PyTrash_UNWIND_LEVEL`` calls. This constant, defined in
+      CPython's object.c, is generally 50. That's basically how many objects are required to
+      get us into the deferred deallocation situation.
+
+    - The test fails by hitting an ``assert()`` in object.c; if the
+      build didn't enable assert, then we don't catch this.
+
+    - If the test fails in that way, the interpreter crashes.
+"""
+from __future__ import print_function, absolute_import, division
+
+import unittest
+
+
+class TestTrashCanReEnter(unittest.TestCase):
+
+    def test_it(self):
+        try:
+            # pylint:disable-next=no-name-in-module
+            from greenlet._greenlet import get_tstate_trash_delete_nesting # pylint:disable=unused-import
+        except ImportError:
+            import sys
+            # Python 3.13 has not "trash delete nesting" anymore (but "delete later")
+            assert sys.version_info[:2] >= (3, 13)
+            self.skipTest("get_tstate_trash_delete_nesting is not available.")
+
+        # Try several times to trigger it, because it isn't 100%
+        # reliable.
+        for _ in range(10):
+            self.check_it()
+
+    def check_it(self): # pylint:disable=too-many-statements
+        import greenlet
+        from greenlet._greenlet import get_tstate_trash_delete_nesting # pylint:disable=no-name-in-module
+        main = greenlet.getcurrent()
+
+        assert get_tstate_trash_delete_nesting() == 0
+
+        # We expect to be in deferred deallocation after this many
+        # deallocations have occurred. TODO: I wish we had a better way to do
+        # this --- that was before get_tstate_trash_delete_nesting; perhaps
+        # we can use that API to do better?
+        TRASH_UNWIND_LEVEL = 50
+        # How many objects to put in a container; it's the container that
+        # queues objects for deferred deallocation.
+        OBJECTS_PER_CONTAINER = 500
+
+        class Dealloc: # define the class here because we alter class variables each time we run.
+            """
+            An object with a ``__del__`` method. When it starts getting deallocated
+            from a deferred trash can run, it switches greenlets, allocates more objects
+            which then also go in the trash can. If we don't save state appropriately,
+            nesting gets out of order and we can crash the interpreter.
+            """
+
+            #: Has our deallocation actually run and switched greenlets?
+            #: When it does, this will be set to the current greenlet. This should
+            #: be happening in the main greenlet, so we check that down below.
+            SPAWNED = False
+
+            #: Has the background greenlet run?
+            BG_RAN = False
+
+            BG_GLET = None
+
+            #: How many of these things have ever been allocated.
+            CREATED = 0
+
+            #: How many of these things have ever been deallocated.
+            DESTROYED = 0
+
+            #: How many were destroyed not in the main greenlet. There should always
+            #: be some.
+            #: If the test is broken or things change in the trashcan implementation,
+            #: this may not be correct.
+            DESTROYED_BG = 0
+
+            def __init__(self, sequence_number):
+                """
+                :param sequence_number: The ordinal of this object during
+                   one particular creation run. This is used to detect (guess, really)
+                   when we have entered the trash can's deferred deallocation.
+                """
+                self.i = sequence_number
+                Dealloc.CREATED += 1
+
+            def __del__(self):
+                if self.i == TRASH_UNWIND_LEVEL and not self.SPAWNED:
+                    Dealloc.SPAWNED = greenlet.getcurrent()
+                    other = Dealloc.BG_GLET = greenlet.greenlet(background_greenlet)
+                    x = other.switch()
+                    assert x == 42
+                    # It's important that we don't switch back to the greenlet,
+                    # we leave it hanging there in an incomplete state. But we don't let it
+                    # get collected, either. If we complete it now, while we're still
+                    # in the scope of the initial trash can, things work out and we
+                    # don't see the problem. We need this greenlet to complete
+                    # at some point in the future, after we've exited this trash can invocation.
+                    del other
+                elif self.i == 40 and greenlet.getcurrent() is not main:
+                    Dealloc.BG_RAN = True
+                    try:
+                        main.switch(42)
+                    except greenlet.GreenletExit as ex:
+                        # We expect this; all references to us go away
+                        # while we're still running, and we need to finish deleting
+                        # ourself.
+                        Dealloc.BG_RAN = type(ex)
+                        del ex
+
+                # Record the fact that we're dead last of all. This ensures that
+                # we actually get returned too.
+                Dealloc.DESTROYED += 1
+                if greenlet.getcurrent() is not main:
+                    Dealloc.DESTROYED_BG += 1
+
+
+        def background_greenlet():
+            # We direct through a second function, instead of
+            # directly calling ``make_some()``, so that we have complete
+            # control over when these objects are destroyed: we need them
+            # to be destroyed in the context of the background greenlet
+            t = make_some()
+            del t # Triggere deletion.
+
+        def make_some():
+            t = ()
+            i = OBJECTS_PER_CONTAINER
+            while i:
+                # Nest the tuples; it's the recursion that gets us
+                # into trash.
+                t = (Dealloc(i), t)
+                i -= 1
+            return t
+
+
+        some = make_some()
+        self.assertEqual(Dealloc.CREATED, OBJECTS_PER_CONTAINER)
+        self.assertEqual(Dealloc.DESTROYED, 0)
+
+        # If we're going to crash, it should be on the following line.
+        # We only crash if ``assert()`` is enabled, of course.
+        del some
+
+        # For non-debug builds of CPython, we won't crash. The best we can do is check
+        # the nesting level explicitly.
+        self.assertEqual(0, get_tstate_trash_delete_nesting())
+
+        # Discard this, raising GreenletExit into where it is waiting.
+        Dealloc.BG_GLET = None
+        # The same nesting level maintains.
+        self.assertEqual(0, get_tstate_trash_delete_nesting())
+
+        # We definitely cleaned some up in the background
+        self.assertGreater(Dealloc.DESTROYED_BG, 0)
+
+        # Make sure all the cleanups happened.
+        self.assertIs(Dealloc.SPAWNED, main)
+        self.assertTrue(Dealloc.BG_RAN)
+        self.assertEqual(Dealloc.BG_RAN, greenlet.GreenletExit)
+        self.assertEqual(Dealloc.CREATED, Dealloc.DESTROYED )
+        self.assertEqual(Dealloc.CREATED, OBJECTS_PER_CONTAINER * 2)
+
+        import gc
+        gc.collect()
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_leaks.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_leaks.py
new file mode 100644
index 00000000..ed1fa717
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_leaks.py
@@ -0,0 +1,443 @@
+# -*- coding: utf-8 -*-
+"""
+Testing scenarios that may have leaked.
+"""
+from __future__ import print_function, absolute_import, division
+
+import sys
+import gc
+
+import time
+import weakref
+import threading
+
+
+import greenlet
+from . import TestCase
+from .leakcheck import fails_leakcheck
+from .leakcheck import ignores_leakcheck
+from .leakcheck import RUNNING_ON_MANYLINUX
+
+# pylint:disable=protected-access
+
+assert greenlet.GREENLET_USE_GC # Option to disable this was removed in 1.0
+
+class HasFinalizerTracksInstances(object):
+    EXTANT_INSTANCES = set()
+    def __init__(self, msg):
+        self.msg = sys.intern(msg)
+        self.EXTANT_INSTANCES.add(id(self))
+    def __del__(self):
+        self.EXTANT_INSTANCES.remove(id(self))
+    def __repr__(self):
+        return "<HasFinalizerTracksInstances at 0x%x %r>" % (
+            id(self), self.msg
+        )
+    @classmethod
+    def reset(cls):
+        cls.EXTANT_INSTANCES.clear()
+
+
+class TestLeaks(TestCase):
+
+    def test_arg_refs(self):
+        args = ('a', 'b', 'c')
+        refcount_before = sys.getrefcount(args)
+        # pylint:disable=unnecessary-lambda
+        g = greenlet.greenlet(
+            lambda *args: greenlet.getcurrent().parent.switch(*args))
+        for _ in range(100):
+            g.switch(*args)
+        self.assertEqual(sys.getrefcount(args), refcount_before)
+
+    def test_kwarg_refs(self):
+        kwargs = {}
+        # pylint:disable=unnecessary-lambda
+        g = greenlet.greenlet(
+            lambda **kwargs: greenlet.getcurrent().parent.switch(**kwargs))
+        for _ in range(100):
+            g.switch(**kwargs)
+        self.assertEqual(sys.getrefcount(kwargs), 2)
+
+
+    @staticmethod
+    def __recycle_threads():
+        # By introducing a thread that does sleep we allow other threads,
+        # that have triggered their __block condition, but did not have a
+        # chance to deallocate their thread state yet, to finally do so.
+        # The way it works is by requiring a GIL switch (different thread),
+        # which does a GIL release (sleep), which might do a GIL switch
+        # to finished threads and allow them to clean up.
+        def worker():
+            time.sleep(0.001)
+        t = threading.Thread(target=worker)
+        t.start()
+        time.sleep(0.001)
+        t.join(10)
+
+    def test_threaded_leak(self):
+        gg = []
+        def worker():
+            # only main greenlet present
+            gg.append(weakref.ref(greenlet.getcurrent()))
+        for _ in range(2):
+            t = threading.Thread(target=worker)
+            t.start()
+            t.join(10)
+            del t
+        greenlet.getcurrent() # update ts_current
+        self.__recycle_threads()
+        greenlet.getcurrent() # update ts_current
+        gc.collect()
+        greenlet.getcurrent() # update ts_current
+        for g in gg:
+            self.assertIsNone(g())
+
+    def test_threaded_adv_leak(self):
+        gg = []
+        def worker():
+            # main and additional *finished* greenlets
+            ll = greenlet.getcurrent().ll = []
+            def additional():
+                ll.append(greenlet.getcurrent())
+            for _ in range(2):
+                greenlet.greenlet(additional).switch()
+            gg.append(weakref.ref(greenlet.getcurrent()))
+        for _ in range(2):
+            t = threading.Thread(target=worker)
+            t.start()
+            t.join(10)
+            del t
+        greenlet.getcurrent() # update ts_current
+        self.__recycle_threads()
+        greenlet.getcurrent() # update ts_current
+        gc.collect()
+        greenlet.getcurrent() # update ts_current
+        for g in gg:
+            self.assertIsNone(g())
+
+    def assertClocksUsed(self):
+        used = greenlet._greenlet.get_clocks_used_doing_optional_cleanup()
+        self.assertGreaterEqual(used, 0)
+        # we don't lose the value
+        greenlet._greenlet.enable_optional_cleanup(True)
+        used2 = greenlet._greenlet.get_clocks_used_doing_optional_cleanup()
+        self.assertEqual(used, used2)
+        self.assertGreater(greenlet._greenlet.CLOCKS_PER_SEC, 1)
+
+    def _check_issue251(self,
+                        manually_collect_background=True,
+                        explicit_reference_to_switch=False):
+        # See https://github.com/python-greenlet/greenlet/issues/251
+        # Killing a greenlet (probably not the main one)
+        # in one thread from another thread would
+        # result in leaking a list (the ts_delkey list).
+        # We no longer use lists to hold that stuff, though.
+
+        # For the test to be valid, even empty lists have to be tracked by the
+        # GC
+
+        assert gc.is_tracked([])
+        HasFinalizerTracksInstances.reset()
+        greenlet.getcurrent()
+        greenlets_before = self.count_objects(greenlet.greenlet, exact_kind=False)
+
+        background_glet_running = threading.Event()
+        background_glet_killed = threading.Event()
+        background_greenlets = []
+
+        # XXX: Switching this to a greenlet subclass that overrides
+        # run results in all callers failing the leaktest; that
+        # greenlet instance is leaked. There's a bound method for
+        # run() living on the stack of the greenlet in g_initialstub,
+        # and since we don't manually switch back to the background
+        # greenlet to let it "fall off the end" and exit the
+        # g_initialstub function, it never gets cleaned up. Making the
+        # garbage collector aware of this bound method (making it an
+        # attribute of the greenlet structure and traversing into it)
+        # doesn't help, for some reason.
+        def background_greenlet():
+            # Throw control back to the main greenlet.
+            jd = HasFinalizerTracksInstances("DELETING STACK OBJECT")
+            greenlet._greenlet.set_thread_local(
+                'test_leaks_key',
+                HasFinalizerTracksInstances("DELETING THREAD STATE"))
+            # Explicitly keeping 'switch' in a local variable
+            # breaks this test in all versions
+            if explicit_reference_to_switch:
+                s = greenlet.getcurrent().parent.switch
+                s([jd])
+            else:
+                greenlet.getcurrent().parent.switch([jd])
+
+        bg_main_wrefs = []
+
+        def background_thread():
+            glet = greenlet.greenlet(background_greenlet)
+            bg_main_wrefs.append(weakref.ref(glet.parent))
+
+            background_greenlets.append(glet)
+            glet.switch() # Be sure it's active.
+            # Control is ours again.
+            del glet # Delete one reference from the thread it runs in.
+            background_glet_running.set()
+            background_glet_killed.wait(10)
+
+            # To trigger the background collection of the dead
+            # greenlet, thus clearing out the contents of the list, we
+            # need to run some APIs. See issue 252.
+            if manually_collect_background:
+                greenlet.getcurrent()
+
+
+        t = threading.Thread(target=background_thread)
+        t.start()
+        background_glet_running.wait(10)
+        greenlet.getcurrent()
+        lists_before = self.count_objects(list, exact_kind=True)
+
+        assert len(background_greenlets) == 1
+        self.assertFalse(background_greenlets[0].dead)
+        # Delete the last reference to the background greenlet
+        # from a different thread. This puts it in the background thread's
+        # ts_delkey list.
+        del background_greenlets[:]
+        background_glet_killed.set()
+
+        # Now wait for the background thread to die.
+        t.join(10)
+        del t
+        # As part of the fix for 252, we need to cycle the ceval.c
+        # interpreter loop to be sure it has had a chance to process
+        # the pending call.
+        self.wait_for_pending_cleanups()
+
+        lists_after = self.count_objects(list, exact_kind=True)
+        greenlets_after = self.count_objects(greenlet.greenlet, exact_kind=False)
+
+        # On 2.7, we observe that lists_after is smaller than
+        # lists_before. No idea what lists got cleaned up. All the
+        # Python 3 versions match exactly.
+        self.assertLessEqual(lists_after, lists_before)
+        # On versions after 3.6, we've successfully cleaned up the
+        # greenlet references thanks to the internal "vectorcall"
+        # protocol; prior to that, there is a reference path through
+        # the ``greenlet.switch`` method still on the stack that we
+        # can't reach to clean up. The C code goes through terrific
+        # lengths to clean that up.
+        if not explicit_reference_to_switch \
+           and greenlet._greenlet.get_clocks_used_doing_optional_cleanup() is not None:
+            # If cleanup was disabled, though, we may not find it.
+            self.assertEqual(greenlets_after, greenlets_before)
+            if manually_collect_background:
+                # TODO: Figure out how to make this work!
+                # The one on the stack is still leaking somehow
+                # in the non-manually-collect state.
+                self.assertEqual(HasFinalizerTracksInstances.EXTANT_INSTANCES, set())
+        else:
+            # The explicit reference prevents us from collecting it
+            # and it isn't always found by the GC either for some
+            # reason. The entire frame is leaked somehow, on some
+            # platforms (e.g., MacPorts builds of Python (all
+            # versions!)), but not on other platforms (the linux and
+            # windows builds on GitHub actions and Appveyor). So we'd
+            # like to write a test that proves that the main greenlet
+            # sticks around, and we can on my machine (macOS 11.6,
+            # MacPorts builds of everything) but we can't write that
+            # same test on other platforms. However, hopefully iteration
+            # done by leakcheck will find it.
+            pass
+
+        if greenlet._greenlet.get_clocks_used_doing_optional_cleanup() is not None:
+            self.assertClocksUsed()
+
+    def test_issue251_killing_cross_thread_leaks_list(self):
+        self._check_issue251()
+
+    def test_issue251_with_cleanup_disabled(self):
+        greenlet._greenlet.enable_optional_cleanup(False)
+        try:
+            self._check_issue251()
+        finally:
+            greenlet._greenlet.enable_optional_cleanup(True)
+
+    @fails_leakcheck
+    def test_issue251_issue252_need_to_collect_in_background(self):
+        # Between greenlet 1.1.2 and the next version, this was still
+        # failing because the leak of the list still exists when we
+        # don't call a greenlet API before exiting the thread. The
+        # proximate cause is that neither of the two greenlets from
+        # the background thread are actually being destroyed, even
+        # though the GC is in fact visiting both objects. It's not
+        # clear where that leak is? For some reason the thread-local
+        # dict holding it isn't being cleaned up.
+        #
+        # The leak, I think, is in the CPYthon internal function that
+        # calls into green_switch(). The argument tuple is still on
+        # the C stack somewhere and can't be reached? That doesn't
+        # make sense, because the tuple should be collectable when
+        # this object goes away.
+        #
+        # Note that this test sometimes spuriously passes on Linux,
+        # for some reason, but I've never seen it pass on macOS.
+        self._check_issue251(manually_collect_background=False)
+
+    @fails_leakcheck
+    def test_issue251_issue252_need_to_collect_in_background_cleanup_disabled(self):
+        self.expect_greenlet_leak = True
+        greenlet._greenlet.enable_optional_cleanup(False)
+        try:
+            self._check_issue251(manually_collect_background=False)
+        finally:
+            greenlet._greenlet.enable_optional_cleanup(True)
+
+    @fails_leakcheck
+    def test_issue251_issue252_explicit_reference_not_collectable(self):
+        self._check_issue251(
+            manually_collect_background=False,
+            explicit_reference_to_switch=True)
+
+    UNTRACK_ATTEMPTS = 100
+
+    def _only_test_some_versions(self):
+        # We're only looking for this problem specifically on 3.11,
+        # and this set of tests is relatively fragile, depending on
+        # OS and memory management details. So we want to run it on 3.11+
+        # (obviously) but not every older 3.x version in order to reduce
+        # false negatives. At the moment, those false results seem to have
+        # resolved, so we are actually running this on 3.8+
+        assert sys.version_info[0] >= 3
+        if sys.version_info[:2] < (3, 8):
+            self.skipTest('Only observed on 3.11')
+        if RUNNING_ON_MANYLINUX:
+            self.skipTest("Slow and not worth repeating here")
+
+    @ignores_leakcheck
+    # Because we're just trying to track raw memory, not objects, and running
+    # the leakcheck makes an already slow test slower.
+    def test_untracked_memory_doesnt_increase(self):
+        # See https://github.com/gevent/gevent/issues/1924
+        # and https://github.com/python-greenlet/greenlet/issues/328
+        self._only_test_some_versions()
+        def f():
+            return 1
+
+        ITER = 10000
+        def run_it():
+            for _ in range(ITER):
+                greenlet.greenlet(f).switch()
+
+        # Establish baseline
+        for _ in range(3):
+            run_it()
+
+        # uss: (Linux, macOS, Windows): aka "Unique Set Size", this is
+        # the memory which is unique to a process and which would be
+        # freed if the process was terminated right now.
+        uss_before = self.get_process_uss()
+
+        for count in range(self.UNTRACK_ATTEMPTS):
+            uss_before = max(uss_before, self.get_process_uss())
+            run_it()
+
+            uss_after = self.get_process_uss()
+            if uss_after <= uss_before and count > 1:
+                break
+
+        self.assertLessEqual(uss_after, uss_before)
+
+    def _check_untracked_memory_thread(self, deallocate_in_thread=True):
+        self._only_test_some_versions()
+        # Like the above test, but what if there are a bunch of
+        # unfinished greenlets in a thread that dies?
+        # Does it matter if we deallocate in the thread or not?
+        EXIT_COUNT = [0]
+
+        def f():
+            try:
+                greenlet.getcurrent().parent.switch()
+            except greenlet.GreenletExit:
+                EXIT_COUNT[0] += 1
+                raise
+            return 1
+
+        ITER = 10000
+        def run_it():
+            glets = []
+            for _ in range(ITER):
+                # Greenlet starts, switches back to us.
+                # We keep a strong reference to the greenlet though so it doesn't
+                # get a GreenletExit exception.
+                g = greenlet.greenlet(f)
+                glets.append(g)
+                g.switch()
+
+            return glets
+
+        test = self
+
+        class ThreadFunc:
+            uss_before = uss_after = 0
+            glets = ()
+            ITER = 2
+            def __call__(self):
+                self.uss_before = test.get_process_uss()
+
+                for _ in range(self.ITER):
+                    self.glets += tuple(run_it())
+
+                for g in self.glets:
+                    test.assertIn('suspended active', str(g))
+                # Drop them.
+                if deallocate_in_thread:
+                    self.glets = ()
+                self.uss_after = test.get_process_uss()
+
+        # Establish baseline
+        uss_before = uss_after = None
+        for count in range(self.UNTRACK_ATTEMPTS):
+            EXIT_COUNT[0] = 0
+            thread_func = ThreadFunc()
+            t = threading.Thread(target=thread_func)
+            t.start()
+            t.join(30)
+            self.assertFalse(t.is_alive())
+
+            if uss_before is None:
+                uss_before = thread_func.uss_before
+
+            uss_before = max(uss_before, thread_func.uss_before)
+            if deallocate_in_thread:
+                self.assertEqual(thread_func.glets, ())
+                self.assertEqual(EXIT_COUNT[0], ITER * thread_func.ITER)
+
+            del thread_func # Deallocate the greenlets; but this won't raise into them
+            del t
+            if not deallocate_in_thread:
+                self.assertEqual(EXIT_COUNT[0], 0)
+            if deallocate_in_thread:
+                self.wait_for_pending_cleanups()
+
+            uss_after = self.get_process_uss()
+            # See if we achieve a non-growth state at some point. Break when we do.
+            if uss_after <= uss_before and count > 1:
+                break
+
+        self.wait_for_pending_cleanups()
+        uss_after = self.get_process_uss()
+        self.assertLessEqual(uss_after, uss_before, "after attempts %d" % (count,))
+
+    @ignores_leakcheck
+    # Because we're just trying to track raw memory, not objects, and running
+    # the leakcheck makes an already slow test slower.
+    def test_untracked_memory_doesnt_increase_unfinished_thread_dealloc_in_thread(self):
+        self._check_untracked_memory_thread(deallocate_in_thread=True)
+
+    @ignores_leakcheck
+    # Because the main greenlets from the background threads do not exit in a timely fashion,
+    # we fail the object-based leakchecks.
+    def test_untracked_memory_doesnt_increase_unfinished_thread_dealloc_in_main(self):
+        self._check_untracked_memory_thread(deallocate_in_thread=False)
+
+if __name__ == '__main__':
+    __import__('unittest').main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_stack_saved.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_stack_saved.py
new file mode 100644
index 00000000..b362bf95
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_stack_saved.py
@@ -0,0 +1,19 @@
+import greenlet
+from . import TestCase
+
+
+class Test(TestCase):
+
+    def test_stack_saved(self):
+        main = greenlet.getcurrent()
+        self.assertEqual(main._stack_saved, 0)
+
+        def func():
+            main.switch(main._stack_saved)
+
+        g = greenlet.greenlet(func)
+        x = g.switch()
+        self.assertGreater(x, 0)
+        self.assertGreater(g._stack_saved, 0)
+        g.switch()
+        self.assertEqual(g._stack_saved, 0)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_throw.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_throw.py
new file mode 100644
index 00000000..f4f9a140
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_throw.py
@@ -0,0 +1,128 @@
+import sys
+
+
+from greenlet import greenlet
+from . import TestCase
+
+def switch(*args):
+    return greenlet.getcurrent().parent.switch(*args)
+
+
+class ThrowTests(TestCase):
+    def test_class(self):
+        def f():
+            try:
+                switch("ok")
+            except RuntimeError:
+                switch("ok")
+                return
+            switch("fail")
+        g = greenlet(f)
+        res = g.switch()
+        self.assertEqual(res, "ok")
+        res = g.throw(RuntimeError)
+        self.assertEqual(res, "ok")
+
+    def test_val(self):
+        def f():
+            try:
+                switch("ok")
+            except RuntimeError:
+                val = sys.exc_info()[1]
+                if str(val) == "ciao":
+                    switch("ok")
+                    return
+            switch("fail")
+
+        g = greenlet(f)
+        res = g.switch()
+        self.assertEqual(res, "ok")
+        res = g.throw(RuntimeError("ciao"))
+        self.assertEqual(res, "ok")
+
+        g = greenlet(f)
+        res = g.switch()
+        self.assertEqual(res, "ok")
+        res = g.throw(RuntimeError, "ciao")
+        self.assertEqual(res, "ok")
+
+    def test_kill(self):
+        def f():
+            switch("ok")
+            switch("fail")
+        g = greenlet(f)
+        res = g.switch()
+        self.assertEqual(res, "ok")
+        res = g.throw()
+        self.assertTrue(isinstance(res, greenlet.GreenletExit))
+        self.assertTrue(g.dead)
+        res = g.throw()    # immediately eaten by the already-dead greenlet
+        self.assertTrue(isinstance(res, greenlet.GreenletExit))
+
+    def test_throw_goes_to_original_parent(self):
+        main = greenlet.getcurrent()
+
+        def f1():
+            try:
+                main.switch("f1 ready to catch")
+            except IndexError:
+                return "caught"
+            return "normal exit"
+
+        def f2():
+            main.switch("from f2")
+
+        g1 = greenlet(f1)
+        g2 = greenlet(f2, parent=g1)
+        with self.assertRaises(IndexError):
+            g2.throw(IndexError)
+        self.assertTrue(g2.dead)
+        self.assertTrue(g1.dead)
+
+        g1 = greenlet(f1)
+        g2 = greenlet(f2, parent=g1)
+        res = g1.switch()
+        self.assertEqual(res, "f1 ready to catch")
+        res = g2.throw(IndexError)
+        self.assertEqual(res, "caught")
+        self.assertTrue(g2.dead)
+        self.assertTrue(g1.dead)
+
+        g1 = greenlet(f1)
+        g2 = greenlet(f2, parent=g1)
+        res = g1.switch()
+        self.assertEqual(res, "f1 ready to catch")
+        res = g2.switch()
+        self.assertEqual(res, "from f2")
+        res = g2.throw(IndexError)
+        self.assertEqual(res, "caught")
+        self.assertTrue(g2.dead)
+        self.assertTrue(g1.dead)
+
+    def test_non_traceback_param(self):
+        with self.assertRaises(TypeError) as exc:
+            greenlet.getcurrent().throw(
+                Exception,
+                Exception(),
+                self
+            )
+        self.assertEqual(str(exc.exception),
+                         "throw() third argument must be a traceback object")
+
+    def test_instance_of_wrong_type(self):
+        with self.assertRaises(TypeError) as exc:
+            greenlet.getcurrent().throw(
+                Exception(),
+                BaseException()
+            )
+
+        self.assertEqual(str(exc.exception),
+                         "instance exception may not have a separate value")
+
+    def test_not_throwable(self):
+        with self.assertRaises(TypeError) as exc:
+            greenlet.getcurrent().throw(
+                "abc"
+            )
+        self.assertEqual(str(exc.exception),
+                         "exceptions must be classes, or instances, not str")
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_tracing.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_tracing.py
new file mode 100644
index 00000000..c044d4b6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_tracing.py
@@ -0,0 +1,291 @@
+from __future__ import print_function
+import sys
+import greenlet
+import unittest
+
+from . import TestCase
+from . import PY312
+
+# https://discuss.python.org/t/cpython-3-12-greenlet-and-tracing-profiling-how-to-not-crash-and-get-correct-results/33144/2
+DEBUG_BUILD_PY312 = (
+    PY312 and hasattr(sys, 'gettotalrefcount'),
+    "Broken on debug builds of Python 3.12"
+)
+
+class SomeError(Exception):
+    pass
+
+class GreenletTracer(object):
+    oldtrace = None
+
+    def __init__(self, error_on_trace=False):
+        self.actions = []
+        self.error_on_trace = error_on_trace
+
+    def __call__(self, *args):
+        self.actions.append(args)
+        if self.error_on_trace:
+            raise SomeError
+
+    def __enter__(self):
+        self.oldtrace = greenlet.settrace(self)
+        return self.actions
+
+    def __exit__(self, *args):
+        greenlet.settrace(self.oldtrace)
+
+
+class TestGreenletTracing(TestCase):
+    """
+    Tests of ``greenlet.settrace()``
+    """
+
+    def test_a_greenlet_tracing(self):
+        main = greenlet.getcurrent()
+        def dummy():
+            pass
+        def dummyexc():
+            raise SomeError()
+
+        with GreenletTracer() as actions:
+            g1 = greenlet.greenlet(dummy)
+            g1.switch()
+            g2 = greenlet.greenlet(dummyexc)
+            self.assertRaises(SomeError, g2.switch)
+
+        self.assertEqual(actions, [
+            ('switch', (main, g1)),
+            ('switch', (g1, main)),
+            ('switch', (main, g2)),
+            ('throw', (g2, main)),
+        ])
+
+    def test_b_exception_disables_tracing(self):
+        main = greenlet.getcurrent()
+        def dummy():
+            main.switch()
+        g = greenlet.greenlet(dummy)
+        g.switch()
+        with GreenletTracer(error_on_trace=True) as actions:
+            self.assertRaises(SomeError, g.switch)
+            self.assertEqual(greenlet.gettrace(), None)
+
+        self.assertEqual(actions, [
+            ('switch', (main, g)),
+        ])
+
+    def test_set_same_tracer_twice(self):
+        # https://github.com/python-greenlet/greenlet/issues/332
+        # Our logic in asserting that the tracefunction should
+        # gain a reference was incorrect if the same tracefunction was set
+        # twice.
+        tracer = GreenletTracer()
+        with tracer:
+            greenlet.settrace(tracer)
+
+
+class PythonTracer(object):
+    oldtrace = None
+
+    def __init__(self):
+        self.actions = []
+
+    def __call__(self, frame, event, arg):
+        # Record the co_name so we have an idea what function we're in.
+        self.actions.append((event, frame.f_code.co_name))
+
+    def __enter__(self):
+        self.oldtrace = sys.setprofile(self)
+        return self.actions
+
+    def __exit__(self, *args):
+        sys.setprofile(self.oldtrace)
+
+def tpt_callback():
+    return 42
+
+class TestPythonTracing(TestCase):
+    """
+    Tests of the interaction of ``sys.settrace()``
+    with greenlet facilities.
+
+    NOTE: Most of this is probably CPython specific.
+    """
+
+    maxDiff = None
+
+    def test_trace_events_trivial(self):
+        with PythonTracer() as actions:
+            tpt_callback()
+        # If we use the sys.settrace instead of setprofile, we get
+        # this:
+
+        # self.assertEqual(actions, [
+        #     ('call', 'tpt_callback'),
+        #     ('call', '__exit__'),
+        # ])
+
+        self.assertEqual(actions, [
+            ('return', '__enter__'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('call', '__exit__'),
+            ('c_call', '__exit__'),
+        ])
+
+    def _trace_switch(self, glet):
+        with PythonTracer() as actions:
+            glet.switch()
+        return actions
+
+    def _check_trace_events_func_already_set(self, glet):
+        actions = self._trace_switch(glet)
+        self.assertEqual(actions, [
+            ('return', '__enter__'),
+            ('c_call', '_trace_switch'),
+            ('call', 'run'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('return', 'run'),
+            ('c_return', '_trace_switch'),
+            ('call', '__exit__'),
+            ('c_call', '__exit__'),
+        ])
+
+    def test_trace_events_into_greenlet_func_already_set(self):
+        def run():
+            return tpt_callback()
+
+        self._check_trace_events_func_already_set(greenlet.greenlet(run))
+
+    def test_trace_events_into_greenlet_subclass_already_set(self):
+        class X(greenlet.greenlet):
+            def run(self):
+                return tpt_callback()
+        self._check_trace_events_func_already_set(X())
+
+    def _check_trace_events_from_greenlet_sets_profiler(self, g, tracer):
+        g.switch()
+        tpt_callback()
+        tracer.__exit__()
+        self.assertEqual(tracer.actions, [
+            ('return', '__enter__'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('return', 'run'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('call', '__exit__'),
+            ('c_call', '__exit__'),
+        ])
+
+
+    def test_trace_events_from_greenlet_func_sets_profiler(self):
+        tracer = PythonTracer()
+        def run():
+            tracer.__enter__()
+            return tpt_callback()
+
+        self._check_trace_events_from_greenlet_sets_profiler(greenlet.greenlet(run),
+                                                             tracer)
+
+    def test_trace_events_from_greenlet_subclass_sets_profiler(self):
+        tracer = PythonTracer()
+        class X(greenlet.greenlet):
+            def run(self):
+                tracer.__enter__()
+                return tpt_callback()
+
+        self._check_trace_events_from_greenlet_sets_profiler(X(), tracer)
+
+    @unittest.skipIf(*DEBUG_BUILD_PY312)
+    def test_trace_events_multiple_greenlets_switching(self):
+        tracer = PythonTracer()
+
+        g1 = None
+        g2 = None
+
+        def g1_run():
+            tracer.__enter__()
+            tpt_callback()
+            g2.switch()
+            tpt_callback()
+            return 42
+
+        def g2_run():
+            tpt_callback()
+            tracer.__exit__()
+            tpt_callback()
+            g1.switch()
+
+        g1 = greenlet.greenlet(g1_run)
+        g2 = greenlet.greenlet(g2_run)
+
+        x = g1.switch()
+        self.assertEqual(x, 42)
+        tpt_callback() # ensure not in the trace
+        self.assertEqual(tracer.actions, [
+            ('return', '__enter__'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('c_call', 'g1_run'),
+            ('call', 'g2_run'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('call', '__exit__'),
+            ('c_call', '__exit__'),
+        ])
+
+    @unittest.skipIf(*DEBUG_BUILD_PY312)
+    def test_trace_events_multiple_greenlets_switching_siblings(self):
+        # Like the first version, but get both greenlets running first
+        # as "siblings" and then establish the tracing.
+        tracer = PythonTracer()
+
+        g1 = None
+        g2 = None
+
+        def g1_run():
+            greenlet.getcurrent().parent.switch()
+            tracer.__enter__()
+            tpt_callback()
+            g2.switch()
+            tpt_callback()
+            return 42
+
+        def g2_run():
+            greenlet.getcurrent().parent.switch()
+
+            tpt_callback()
+            tracer.__exit__()
+            tpt_callback()
+            g1.switch()
+
+        g1 = greenlet.greenlet(g1_run)
+        g2 = greenlet.greenlet(g2_run)
+
+        # Start g1
+        g1.switch()
+        # And it immediately returns control to us.
+        # Start g2
+        g2.switch()
+        # Which also returns. Now kick of the real part of the
+        # test.
+        x = g1.switch()
+        self.assertEqual(x, 42)
+
+        tpt_callback() # ensure not in the trace
+        self.assertEqual(tracer.actions, [
+            ('return', '__enter__'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('c_call', 'g1_run'),
+            ('call', 'tpt_callback'),
+            ('return', 'tpt_callback'),
+            ('call', '__exit__'),
+            ('c_call', '__exit__'),
+        ])
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_version.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_version.py
new file mode 100644
index 00000000..96c17cf1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_version.py
@@ -0,0 +1,41 @@
+#! /usr/bin/env python
+from __future__ import absolute_import
+from __future__ import print_function
+
+import sys
+import os
+from unittest import TestCase as NonLeakingTestCase
+
+import greenlet
+
+# No reason to run this multiple times under leakchecks,
+# it doesn't do anything.
+class VersionTests(NonLeakingTestCase):
+    def test_version(self):
+        def find_dominating_file(name):
+            if os.path.exists(name):
+                return name
+
+            tried = []
+            here = os.path.abspath(os.path.dirname(__file__))
+            for i in range(10):
+                up = ['..'] * i
+                path = [here] + up + [name]
+                fname = os.path.join(*path)
+                fname = os.path.abspath(fname)
+                tried.append(fname)
+                if os.path.exists(fname):
+                    return fname
+            raise AssertionError("Could not find file " + name + "; checked " + str(tried))
+
+        try:
+            setup_py = find_dominating_file('setup.py')
+        except AssertionError as e:
+            self.skipTest("Unable to find setup.py; must be out of tree. " + str(e))
+
+
+        invoke_setup = "%s %s --version" % (sys.executable, setup_py)
+        with os.popen(invoke_setup) as f:
+            sversion = f.read().strip()
+
+        self.assertEqual(sversion, greenlet.__version__)
diff --git a/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_weakref.py b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_weakref.py
new file mode 100644
index 00000000..05a38a7f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/greenlet/tests/test_weakref.py
@@ -0,0 +1,35 @@
+import gc
+import weakref
+
+
+import greenlet
+from . import TestCase
+
+class WeakRefTests(TestCase):
+    def test_dead_weakref(self):
+        def _dead_greenlet():
+            g = greenlet.greenlet(lambda: None)
+            g.switch()
+            return g
+        o = weakref.ref(_dead_greenlet())
+        gc.collect()
+        self.assertEqual(o(), None)
+
+    def test_inactive_weakref(self):
+        o = weakref.ref(greenlet.greenlet())
+        gc.collect()
+        self.assertEqual(o(), None)
+
+    def test_dealloc_weakref(self):
+        seen = []
+        def worker():
+            try:
+                greenlet.getcurrent().parent.switch()
+            finally:
+                seen.append(g())
+        g = greenlet.greenlet(worker)
+        g.switch()
+        g2 = greenlet.greenlet(lambda: None, g)
+        g = weakref.ref(g2)
+        g2 = None
+        self.assertEqual(seen, [None])
diff --git a/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/LICENSE.md b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/LICENSE.md
new file mode 100644
index 00000000..19b6b452
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/LICENSE.md
@@ -0,0 +1,31 @@
+BSD 3-Clause License
+
+Copyright (c) 2013-2024, Kim Davies and contributors.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1. Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+3. Neither the name of the copyright holder nor the names of its
+   contributors may be used to endorse or promote products derived from
+   this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/METADATA
new file mode 100644
index 00000000..c42623e9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/METADATA
@@ -0,0 +1,250 @@
+Metadata-Version: 2.1
+Name: idna
+Version: 3.10
+Summary: Internationalized Domain Names in Applications (IDNA)
+Author-email: Kim Davies <kim+pypi@gumleaf.org>
+Requires-Python: >=3.6
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: System Administrators
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Internet :: Name Service (DNS)
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Utilities
+Requires-Dist: ruff >= 0.6.2 ; extra == "all"
+Requires-Dist: mypy >= 1.11.2 ; extra == "all"
+Requires-Dist: pytest >= 8.3.2 ; extra == "all"
+Requires-Dist: flake8 >= 7.1.1 ; extra == "all"
+Project-URL: Changelog, https://github.com/kjd/idna/blob/master/HISTORY.rst
+Project-URL: Issue tracker, https://github.com/kjd/idna/issues
+Project-URL: Source, https://github.com/kjd/idna
+Provides-Extra: all
+
+Internationalized Domain Names in Applications (IDNA)
+=====================================================
+
+Support for the Internationalized Domain Names in
+Applications (IDNA) protocol as specified in `RFC 5891
+<https://tools.ietf.org/html/rfc5891>`_. This is the latest version of
+the protocol and is sometimes referred to as “IDNA 2008”.
+
+This library also provides support for Unicode Technical
+Standard 46, `Unicode IDNA Compatibility Processing
+<https://unicode.org/reports/tr46/>`_.
+
+This acts as a suitable replacement for the “encodings.idna”
+module that comes with the Python standard library, but which
+only supports the older superseded IDNA specification (`RFC 3490
+<https://tools.ietf.org/html/rfc3490>`_).
+
+Basic functions are simply executed:
+
+.. code-block:: pycon
+
+    >>> import idna
+    >>> idna.encode('ドメイン.テスト')
+    b'xn--eckwd4c7c.xn--zckzah'
+    >>> print(idna.decode('xn--eckwd4c7c.xn--zckzah'))
+    ドメイン.テスト
+
+
+Installation
+------------
+
+This package is available for installation from PyPI:
+
+.. code-block:: bash
+
+    $ python3 -m pip install idna
+
+
+Usage
+-----
+
+For typical usage, the ``encode`` and ``decode`` functions will take a
+domain name argument and perform a conversion to A-labels or U-labels
+respectively.
+
+.. code-block:: pycon
+
+    >>> import idna
+    >>> idna.encode('ドメイン.テスト')
+    b'xn--eckwd4c7c.xn--zckzah'
+    >>> print(idna.decode('xn--eckwd4c7c.xn--zckzah'))
+    ドメイン.テスト
+
+You may use the codec encoding and decoding methods using the
+``idna.codec`` module:
+
+.. code-block:: pycon
+
+    >>> import idna.codec
+    >>> print('домен.испытание'.encode('idna2008'))
+    b'xn--d1acufc.xn--80akhbyknj4f'
+    >>> print(b'xn--d1acufc.xn--80akhbyknj4f'.decode('idna2008'))
+    домен.испытание
+
+Conversions can be applied at a per-label basis using the ``ulabel`` or
+``alabel`` functions if necessary:
+
+.. code-block:: pycon
+
+    >>> idna.alabel('测试')
+    b'xn--0zwm56d'
+
+Compatibility Mapping (UTS #46)
++++++++++++++++++++++++++++++++
+
+As described in `RFC 5895 <https://tools.ietf.org/html/rfc5895>`_, the
+IDNA specification does not normalize input from different potential
+ways a user may input a domain name. This functionality, known as
+a “mapping”, is considered by the specification to be a local
+user-interface issue distinct from IDNA conversion functionality.
+
+This library provides one such mapping that was developed by the
+Unicode Consortium. Known as `Unicode IDNA Compatibility Processing
+<https://unicode.org/reports/tr46/>`_, it provides for both a regular
+mapping for typical applications, as well as a transitional mapping to
+help migrate from older IDNA 2003 applications. Strings are
+preprocessed according to Section 4.4 “Preprocessing for IDNA2008”
+prior to the IDNA operations.
+
+For example, “Königsgäßchen” is not a permissible label as *LATIN
+CAPITAL LETTER K* is not allowed (nor are capital letters in general).
+UTS 46 will convert this into lower case prior to applying the IDNA
+conversion.
+
+.. code-block:: pycon
+
+    >>> import idna
+    >>> idna.encode('Königsgäßchen')
+    ...
+    idna.core.InvalidCodepoint: Codepoint U+004B at position 1 of 'Königsgäßchen' not allowed
+    >>> idna.encode('Königsgäßchen', uts46=True)
+    b'xn--knigsgchen-b4a3dun'
+    >>> print(idna.decode('xn--knigsgchen-b4a3dun'))
+    königsgäßchen
+
+Transitional processing provides conversions to help transition from
+the older 2003 standard to the current standard. For example, in the
+original IDNA specification, the *LATIN SMALL LETTER SHARP S* (ß) was
+converted into two *LATIN SMALL LETTER S* (ss), whereas in the current
+IDNA specification this conversion is not performed.
+
+.. code-block:: pycon
+
+    >>> idna.encode('Königsgäßchen', uts46=True, transitional=True)
+    'xn--knigsgsschen-lcb0w'
+
+Implementers should use transitional processing with caution, only in
+rare cases where conversion from legacy labels to current labels must be
+performed (i.e. IDNA implementations that pre-date 2008). For typical
+applications that just need to convert labels, transitional processing
+is unlikely to be beneficial and could produce unexpected incompatible
+results.
+
+``encodings.idna`` Compatibility
+++++++++++++++++++++++++++++++++
+
+Function calls from the Python built-in ``encodings.idna`` module are
+mapped to their IDNA 2008 equivalents using the ``idna.compat`` module.
+Simply substitute the ``import`` clause in your code to refer to the new
+module name.
+
+Exceptions
+----------
+
+All errors raised during the conversion following the specification
+should raise an exception derived from the ``idna.IDNAError`` base
+class.
+
+More specific exceptions that may be generated as ``idna.IDNABidiError``
+when the error reflects an illegal combination of left-to-right and
+right-to-left characters in a label; ``idna.InvalidCodepoint`` when
+a specific codepoint is an illegal character in an IDN label (i.e.
+INVALID); and ``idna.InvalidCodepointContext`` when the codepoint is
+illegal based on its positional context (i.e. it is CONTEXTO or CONTEXTJ
+but the contextual requirements are not satisfied.)
+
+Building and Diagnostics
+------------------------
+
+The IDNA and UTS 46 functionality relies upon pre-calculated lookup
+tables for performance. These tables are derived from computing against
+eligibility criteria in the respective standards. These tables are
+computed using the command-line script ``tools/idna-data``.
+
+This tool will fetch relevant codepoint data from the Unicode repository
+and perform the required calculations to identify eligibility. There are
+three main modes:
+
+* ``idna-data make-libdata``. Generates ``idnadata.py`` and
+  ``uts46data.py``, the pre-calculated lookup tables used for IDNA and
+  UTS 46 conversions. Implementers who wish to track this library against
+  a different Unicode version may use this tool to manually generate a
+  different version of the ``idnadata.py`` and ``uts46data.py`` files.
+
+* ``idna-data make-table``. Generate a table of the IDNA disposition
+  (e.g. PVALID, CONTEXTJ, CONTEXTO) in the format found in Appendix
+  B.1 of RFC 5892 and the pre-computed tables published by `IANA
+  <https://www.iana.org/>`_.
+
+* ``idna-data U+0061``. Prints debugging output on the various
+  properties associated with an individual Unicode codepoint (in this
+  case, U+0061), that are used to assess the IDNA and UTS 46 status of a
+  codepoint. This is helpful in debugging or analysis.
+
+The tool accepts a number of arguments, described using ``idna-data
+-h``. Most notably, the ``--version`` argument allows the specification
+of the version of Unicode to be used in computing the table data. For
+example, ``idna-data --version 9.0.0 make-libdata`` will generate
+library data against Unicode 9.0.0.
+
+
+Additional Notes
+----------------
+
+* **Packages**. The latest tagged release version is published in the
+  `Python Package Index <https://pypi.org/project/idna/>`_.
+
+* **Version support**. This library supports Python 3.6 and higher.
+  As this library serves as a low-level toolkit for a variety of
+  applications, many of which strive for broad compatibility with older
+  Python versions, there is no rush to remove older interpreter support.
+  Removing support for older versions should be well justified in that the
+  maintenance burden has become too high.
+
+* **Python 2**. Python 2 is supported by version 2.x of this library.
+  Use "idna<3" in your requirements file if you need this library for
+  a Python 2 application. Be advised that these versions are no longer
+  actively developed.
+
+* **Testing**. The library has a test suite based on each rule of the
+  IDNA specification, as well as tests that are provided as part of the
+  Unicode Technical Standard 46, `Unicode IDNA Compatibility Processing
+  <https://unicode.org/reports/tr46/>`_.
+
+* **Emoji**. It is an occasional request to support emoji domains in
+  this library. Encoding of symbols like emoji is expressly prohibited by
+  the technical standard IDNA 2008 and emoji domains are broadly phased
+  out across the domain industry due to associated security risks. For
+  now, applications that need to support these non-compliant labels
+  may wish to consider trying the encode/decode operation in this library
+  first, and then falling back to using `encodings.idna`. See `the Github
+  project <https://github.com/kjd/idna/issues/18>`_ for more discussion.
+
diff --git a/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/RECORD
new file mode 100644
index 00000000..9cfce7f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/RECORD
@@ -0,0 +1,22 @@
+idna-3.10.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+idna-3.10.dist-info/LICENSE.md,sha256=pZ8LDvNjWHQQmkRhykT_enDVBpboFHZ7-vch1Mmw2w8,1541
+idna-3.10.dist-info/METADATA,sha256=URR5ZyDfQ1PCEGhkYoojqfi2Ra0tau2--lhwG4XSfjI,10158
+idna-3.10.dist-info/RECORD,,
+idna-3.10.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+idna/__init__.py,sha256=MPqNDLZbXqGaNdXxAFhiqFPKEQXju2jNQhCey6-5eJM,868
+idna/__pycache__/__init__.cpython-312.pyc,,
+idna/__pycache__/codec.cpython-312.pyc,,
+idna/__pycache__/compat.cpython-312.pyc,,
+idna/__pycache__/core.cpython-312.pyc,,
+idna/__pycache__/idnadata.cpython-312.pyc,,
+idna/__pycache__/intranges.cpython-312.pyc,,
+idna/__pycache__/package_data.cpython-312.pyc,,
+idna/__pycache__/uts46data.cpython-312.pyc,,
+idna/codec.py,sha256=PEew3ItwzjW4hymbasnty2N2OXvNcgHB-JjrBuxHPYY,3422
+idna/compat.py,sha256=RzLy6QQCdl9784aFhb2EX9EKGCJjg0P3PilGdeXXcx8,316
+idna/core.py,sha256=YJYyAMnwiQEPjVC4-Fqu_p4CJ6yKKuDGmppBNQNQpFs,13239
+idna/idnadata.py,sha256=W30GcIGvtOWYwAjZj4ZjuouUutC6ffgNuyjJy7fZ-lo,78306
+idna/intranges.py,sha256=amUtkdhYcQG8Zr-CoMM_kVRacxkivC1WgxN1b63KKdU,1898
+idna/package_data.py,sha256=q59S3OXsc5VI8j6vSD0sGBMyk6zZ4vWFREE88yCJYKs,21
+idna/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+idna/uts46data.py,sha256=rt90K9J40gUSwppDPCrhjgi5AA6pWM65dEGRSf6rIhM,239289
diff --git a/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna-3.10.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__init__.py b/backend/venv/lib/python3.12/site-packages/idna/__init__.py
new file mode 100644
index 00000000..cfdc030a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/__init__.py
@@ -0,0 +1,45 @@
+from .core import (
+    IDNABidiError,
+    IDNAError,
+    InvalidCodepoint,
+    InvalidCodepointContext,
+    alabel,
+    check_bidi,
+    check_hyphen_ok,
+    check_initial_combiner,
+    check_label,
+    check_nfc,
+    decode,
+    encode,
+    ulabel,
+    uts46_remap,
+    valid_contextj,
+    valid_contexto,
+    valid_label_length,
+    valid_string_length,
+)
+from .intranges import intranges_contain
+from .package_data import __version__
+
+__all__ = [
+    "__version__",
+    "IDNABidiError",
+    "IDNAError",
+    "InvalidCodepoint",
+    "InvalidCodepointContext",
+    "alabel",
+    "check_bidi",
+    "check_hyphen_ok",
+    "check_initial_combiner",
+    "check_label",
+    "check_nfc",
+    "decode",
+    "encode",
+    "intranges_contain",
+    "ulabel",
+    "uts46_remap",
+    "valid_contextj",
+    "valid_contexto",
+    "valid_label_length",
+    "valid_string_length",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..09d44d4d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/codec.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/codec.cpython-312.pyc
new file mode 100644
index 00000000..fd5bebc0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/codec.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..beeff635
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..46740a90
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/idnadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/idnadata.cpython-312.pyc
new file mode 100644
index 00000000..6d78adb1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/idnadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/intranges.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/intranges.cpython-312.pyc
new file mode 100644
index 00000000..3035d036
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/intranges.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/package_data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/package_data.cpython-312.pyc
new file mode 100644
index 00000000..52b51145
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/package_data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/__pycache__/uts46data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/uts46data.cpython-312.pyc
new file mode 100644
index 00000000..01ecd544
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/idna/__pycache__/uts46data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/idna/codec.py b/backend/venv/lib/python3.12/site-packages/idna/codec.py
new file mode 100644
index 00000000..913abfd6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/codec.py
@@ -0,0 +1,122 @@
+import codecs
+import re
+from typing import Any, Optional, Tuple
+
+from .core import IDNAError, alabel, decode, encode, ulabel
+
+_unicode_dots_re = re.compile("[\u002e\u3002\uff0e\uff61]")
+
+
+class Codec(codecs.Codec):
+    def encode(self, data: str, errors: str = "strict") -> Tuple[bytes, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return b"", 0
+
+        return encode(data), len(data)
+
+    def decode(self, data: bytes, errors: str = "strict") -> Tuple[str, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return "", 0
+
+        return decode(data), len(data)
+
+
+class IncrementalEncoder(codecs.BufferedIncrementalEncoder):
+    def _buffer_encode(self, data: str, errors: str, final: bool) -> Tuple[bytes, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return b"", 0
+
+        labels = _unicode_dots_re.split(data)
+        trailing_dot = b""
+        if labels:
+            if not labels[-1]:
+                trailing_dot = b"."
+                del labels[-1]
+            elif not final:
+                # Keep potentially unfinished label until the next call
+                del labels[-1]
+                if labels:
+                    trailing_dot = b"."
+
+        result = []
+        size = 0
+        for label in labels:
+            result.append(alabel(label))
+            if size:
+                size += 1
+            size += len(label)
+
+        # Join with U+002E
+        result_bytes = b".".join(result) + trailing_dot
+        size += len(trailing_dot)
+        return result_bytes, size
+
+
+class IncrementalDecoder(codecs.BufferedIncrementalDecoder):
+    def _buffer_decode(self, data: Any, errors: str, final: bool) -> Tuple[str, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return ("", 0)
+
+        if not isinstance(data, str):
+            data = str(data, "ascii")
+
+        labels = _unicode_dots_re.split(data)
+        trailing_dot = ""
+        if labels:
+            if not labels[-1]:
+                trailing_dot = "."
+                del labels[-1]
+            elif not final:
+                # Keep potentially unfinished label until the next call
+                del labels[-1]
+                if labels:
+                    trailing_dot = "."
+
+        result = []
+        size = 0
+        for label in labels:
+            result.append(ulabel(label))
+            if size:
+                size += 1
+            size += len(label)
+
+        result_str = ".".join(result) + trailing_dot
+        size += len(trailing_dot)
+        return (result_str, size)
+
+
+class StreamWriter(Codec, codecs.StreamWriter):
+    pass
+
+
+class StreamReader(Codec, codecs.StreamReader):
+    pass
+
+
+def search_function(name: str) -> Optional[codecs.CodecInfo]:
+    if name != "idna2008":
+        return None
+    return codecs.CodecInfo(
+        name=name,
+        encode=Codec().encode,
+        decode=Codec().decode,
+        incrementalencoder=IncrementalEncoder,
+        incrementaldecoder=IncrementalDecoder,
+        streamwriter=StreamWriter,
+        streamreader=StreamReader,
+    )
+
+
+codecs.register(search_function)
diff --git a/backend/venv/lib/python3.12/site-packages/idna/compat.py b/backend/venv/lib/python3.12/site-packages/idna/compat.py
new file mode 100644
index 00000000..1df9f2a7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/compat.py
@@ -0,0 +1,15 @@
+from typing import Any, Union
+
+from .core import decode, encode
+
+
+def ToASCII(label: str) -> bytes:
+    return encode(label)
+
+
+def ToUnicode(label: Union[bytes, bytearray]) -> str:
+    return decode(label)
+
+
+def nameprep(s: Any) -> None:
+    raise NotImplementedError("IDNA 2008 does not utilise nameprep protocol")
diff --git a/backend/venv/lib/python3.12/site-packages/idna/core.py b/backend/venv/lib/python3.12/site-packages/idna/core.py
new file mode 100644
index 00000000..9115f123
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/core.py
@@ -0,0 +1,437 @@
+import bisect
+import re
+import unicodedata
+from typing import Optional, Union
+
+from . import idnadata
+from .intranges import intranges_contain
+
+_virama_combining_class = 9
+_alabel_prefix = b"xn--"
+_unicode_dots_re = re.compile("[\u002e\u3002\uff0e\uff61]")
+
+
+class IDNAError(UnicodeError):
+    """Base exception for all IDNA-encoding related problems"""
+
+    pass
+
+
+class IDNABidiError(IDNAError):
+    """Exception when bidirectional requirements are not satisfied"""
+
+    pass
+
+
+class InvalidCodepoint(IDNAError):
+    """Exception when a disallowed or unallocated codepoint is used"""
+
+    pass
+
+
+class InvalidCodepointContext(IDNAError):
+    """Exception when the codepoint is not valid in the context it is used"""
+
+    pass
+
+
+def _combining_class(cp: int) -> int:
+    v = unicodedata.combining(chr(cp))
+    if v == 0:
+        if not unicodedata.name(chr(cp)):
+            raise ValueError("Unknown character in unicodedata")
+    return v
+
+
+def _is_script(cp: str, script: str) -> bool:
+    return intranges_contain(ord(cp), idnadata.scripts[script])
+
+
+def _punycode(s: str) -> bytes:
+    return s.encode("punycode")
+
+
+def _unot(s: int) -> str:
+    return "U+{:04X}".format(s)
+
+
+def valid_label_length(label: Union[bytes, str]) -> bool:
+    if len(label) > 63:
+        return False
+    return True
+
+
+def valid_string_length(label: Union[bytes, str], trailing_dot: bool) -> bool:
+    if len(label) > (254 if trailing_dot else 253):
+        return False
+    return True
+
+
+def check_bidi(label: str, check_ltr: bool = False) -> bool:
+    # Bidi rules should only be applied if string contains RTL characters
+    bidi_label = False
+    for idx, cp in enumerate(label, 1):
+        direction = unicodedata.bidirectional(cp)
+        if direction == "":
+            # String likely comes from a newer version of Unicode
+            raise IDNABidiError("Unknown directionality in label {} at position {}".format(repr(label), idx))
+        if direction in ["R", "AL", "AN"]:
+            bidi_label = True
+    if not bidi_label and not check_ltr:
+        return True
+
+    # Bidi rule 1
+    direction = unicodedata.bidirectional(label[0])
+    if direction in ["R", "AL"]:
+        rtl = True
+    elif direction == "L":
+        rtl = False
+    else:
+        raise IDNABidiError("First codepoint in label {} must be directionality L, R or AL".format(repr(label)))
+
+    valid_ending = False
+    number_type: Optional[str] = None
+    for idx, cp in enumerate(label, 1):
+        direction = unicodedata.bidirectional(cp)
+
+        if rtl:
+            # Bidi rule 2
+            if direction not in [
+                "R",
+                "AL",
+                "AN",
+                "EN",
+                "ES",
+                "CS",
+                "ET",
+                "ON",
+                "BN",
+                "NSM",
+            ]:
+                raise IDNABidiError("Invalid direction for codepoint at position {} in a right-to-left label".format(idx))
+            # Bidi rule 3
+            if direction in ["R", "AL", "EN", "AN"]:
+                valid_ending = True
+            elif direction != "NSM":
+                valid_ending = False
+            # Bidi rule 4
+            if direction in ["AN", "EN"]:
+                if not number_type:
+                    number_type = direction
+                else:
+                    if number_type != direction:
+                        raise IDNABidiError("Can not mix numeral types in a right-to-left label")
+        else:
+            # Bidi rule 5
+            if direction not in ["L", "EN", "ES", "CS", "ET", "ON", "BN", "NSM"]:
+                raise IDNABidiError("Invalid direction for codepoint at position {} in a left-to-right label".format(idx))
+            # Bidi rule 6
+            if direction in ["L", "EN"]:
+                valid_ending = True
+            elif direction != "NSM":
+                valid_ending = False
+
+    if not valid_ending:
+        raise IDNABidiError("Label ends with illegal codepoint directionality")
+
+    return True
+
+
+def check_initial_combiner(label: str) -> bool:
+    if unicodedata.category(label[0])[0] == "M":
+        raise IDNAError("Label begins with an illegal combining character")
+    return True
+
+
+def check_hyphen_ok(label: str) -> bool:
+    if label[2:4] == "--":
+        raise IDNAError("Label has disallowed hyphens in 3rd and 4th position")
+    if label[0] == "-" or label[-1] == "-":
+        raise IDNAError("Label must not start or end with a hyphen")
+    return True
+
+
+def check_nfc(label: str) -> None:
+    if unicodedata.normalize("NFC", label) != label:
+        raise IDNAError("Label must be in Normalization Form C")
+
+
+def valid_contextj(label: str, pos: int) -> bool:
+    cp_value = ord(label[pos])
+
+    if cp_value == 0x200C:
+        if pos > 0:
+            if _combining_class(ord(label[pos - 1])) == _virama_combining_class:
+                return True
+
+        ok = False
+        for i in range(pos - 1, -1, -1):
+            joining_type = idnadata.joining_types.get(ord(label[i]))
+            if joining_type == ord("T"):
+                continue
+            elif joining_type in [ord("L"), ord("D")]:
+                ok = True
+                break
+            else:
+                break
+
+        if not ok:
+            return False
+
+        ok = False
+        for i in range(pos + 1, len(label)):
+            joining_type = idnadata.joining_types.get(ord(label[i]))
+            if joining_type == ord("T"):
+                continue
+            elif joining_type in [ord("R"), ord("D")]:
+                ok = True
+                break
+            else:
+                break
+        return ok
+
+    if cp_value == 0x200D:
+        if pos > 0:
+            if _combining_class(ord(label[pos - 1])) == _virama_combining_class:
+                return True
+        return False
+
+    else:
+        return False
+
+
+def valid_contexto(label: str, pos: int, exception: bool = False) -> bool:
+    cp_value = ord(label[pos])
+
+    if cp_value == 0x00B7:
+        if 0 < pos < len(label) - 1:
+            if ord(label[pos - 1]) == 0x006C and ord(label[pos + 1]) == 0x006C:
+                return True
+        return False
+
+    elif cp_value == 0x0375:
+        if pos < len(label) - 1 and len(label) > 1:
+            return _is_script(label[pos + 1], "Greek")
+        return False
+
+    elif cp_value == 0x05F3 or cp_value == 0x05F4:
+        if pos > 0:
+            return _is_script(label[pos - 1], "Hebrew")
+        return False
+
+    elif cp_value == 0x30FB:
+        for cp in label:
+            if cp == "\u30fb":
+                continue
+            if _is_script(cp, "Hiragana") or _is_script(cp, "Katakana") or _is_script(cp, "Han"):
+                return True
+        return False
+
+    elif 0x660 <= cp_value <= 0x669:
+        for cp in label:
+            if 0x6F0 <= ord(cp) <= 0x06F9:
+                return False
+        return True
+
+    elif 0x6F0 <= cp_value <= 0x6F9:
+        for cp in label:
+            if 0x660 <= ord(cp) <= 0x0669:
+                return False
+        return True
+
+    return False
+
+
+def check_label(label: Union[str, bytes, bytearray]) -> None:
+    if isinstance(label, (bytes, bytearray)):
+        label = label.decode("utf-8")
+    if len(label) == 0:
+        raise IDNAError("Empty Label")
+
+    check_nfc(label)
+    check_hyphen_ok(label)
+    check_initial_combiner(label)
+
+    for pos, cp in enumerate(label):
+        cp_value = ord(cp)
+        if intranges_contain(cp_value, idnadata.codepoint_classes["PVALID"]):
+            continue
+        elif intranges_contain(cp_value, idnadata.codepoint_classes["CONTEXTJ"]):
+            try:
+                if not valid_contextj(label, pos):
+                    raise InvalidCodepointContext(
+                        "Joiner {} not allowed at position {} in {}".format(_unot(cp_value), pos + 1, repr(label))
+                    )
+            except ValueError:
+                raise IDNAError(
+                    "Unknown codepoint adjacent to joiner {} at position {} in {}".format(
+                        _unot(cp_value), pos + 1, repr(label)
+                    )
+                )
+        elif intranges_contain(cp_value, idnadata.codepoint_classes["CONTEXTO"]):
+            if not valid_contexto(label, pos):
+                raise InvalidCodepointContext(
+                    "Codepoint {} not allowed at position {} in {}".format(_unot(cp_value), pos + 1, repr(label))
+                )
+        else:
+            raise InvalidCodepoint(
+                "Codepoint {} at position {} of {} not allowed".format(_unot(cp_value), pos + 1, repr(label))
+            )
+
+    check_bidi(label)
+
+
+def alabel(label: str) -> bytes:
+    try:
+        label_bytes = label.encode("ascii")
+        ulabel(label_bytes)
+        if not valid_label_length(label_bytes):
+            raise IDNAError("Label too long")
+        return label_bytes
+    except UnicodeEncodeError:
+        pass
+
+    check_label(label)
+    label_bytes = _alabel_prefix + _punycode(label)
+
+    if not valid_label_length(label_bytes):
+        raise IDNAError("Label too long")
+
+    return label_bytes
+
+
+def ulabel(label: Union[str, bytes, bytearray]) -> str:
+    if not isinstance(label, (bytes, bytearray)):
+        try:
+            label_bytes = label.encode("ascii")
+        except UnicodeEncodeError:
+            check_label(label)
+            return label
+    else:
+        label_bytes = label
+
+    label_bytes = label_bytes.lower()
+    if label_bytes.startswith(_alabel_prefix):
+        label_bytes = label_bytes[len(_alabel_prefix) :]
+        if not label_bytes:
+            raise IDNAError("Malformed A-label, no Punycode eligible content found")
+        if label_bytes.decode("ascii")[-1] == "-":
+            raise IDNAError("A-label must not end with a hyphen")
+    else:
+        check_label(label_bytes)
+        return label_bytes.decode("ascii")
+
+    try:
+        label = label_bytes.decode("punycode")
+    except UnicodeError:
+        raise IDNAError("Invalid A-label")
+    check_label(label)
+    return label
+
+
+def uts46_remap(domain: str, std3_rules: bool = True, transitional: bool = False) -> str:
+    """Re-map the characters in the string according to UTS46 processing."""
+    from .uts46data import uts46data
+
+    output = ""
+
+    for pos, char in enumerate(domain):
+        code_point = ord(char)
+        try:
+            uts46row = uts46data[code_point if code_point < 256 else bisect.bisect_left(uts46data, (code_point, "Z")) - 1]
+            status = uts46row[1]
+            replacement: Optional[str] = None
+            if len(uts46row) == 3:
+                replacement = uts46row[2]
+            if (
+                status == "V"
+                or (status == "D" and not transitional)
+                or (status == "3" and not std3_rules and replacement is None)
+            ):
+                output += char
+            elif replacement is not None and (
+                status == "M" or (status == "3" and not std3_rules) or (status == "D" and transitional)
+            ):
+                output += replacement
+            elif status != "I":
+                raise IndexError()
+        except IndexError:
+            raise InvalidCodepoint(
+                "Codepoint {} not allowed at position {} in {}".format(_unot(code_point), pos + 1, repr(domain))
+            )
+
+    return unicodedata.normalize("NFC", output)
+
+
+def encode(
+    s: Union[str, bytes, bytearray],
+    strict: bool = False,
+    uts46: bool = False,
+    std3_rules: bool = False,
+    transitional: bool = False,
+) -> bytes:
+    if not isinstance(s, str):
+        try:
+            s = str(s, "ascii")
+        except UnicodeDecodeError:
+            raise IDNAError("should pass a unicode string to the function rather than a byte string.")
+    if uts46:
+        s = uts46_remap(s, std3_rules, transitional)
+    trailing_dot = False
+    result = []
+    if strict:
+        labels = s.split(".")
+    else:
+        labels = _unicode_dots_re.split(s)
+    if not labels or labels == [""]:
+        raise IDNAError("Empty domain")
+    if labels[-1] == "":
+        del labels[-1]
+        trailing_dot = True
+    for label in labels:
+        s = alabel(label)
+        if s:
+            result.append(s)
+        else:
+            raise IDNAError("Empty label")
+    if trailing_dot:
+        result.append(b"")
+    s = b".".join(result)
+    if not valid_string_length(s, trailing_dot):
+        raise IDNAError("Domain too long")
+    return s
+
+
+def decode(
+    s: Union[str, bytes, bytearray],
+    strict: bool = False,
+    uts46: bool = False,
+    std3_rules: bool = False,
+) -> str:
+    try:
+        if not isinstance(s, str):
+            s = str(s, "ascii")
+    except UnicodeDecodeError:
+        raise IDNAError("Invalid ASCII in A-label")
+    if uts46:
+        s = uts46_remap(s, std3_rules, False)
+    trailing_dot = False
+    result = []
+    if not strict:
+        labels = _unicode_dots_re.split(s)
+    else:
+        labels = s.split(".")
+    if not labels or labels == [""]:
+        raise IDNAError("Empty domain")
+    if not labels[-1]:
+        del labels[-1]
+        trailing_dot = True
+    for label in labels:
+        s = ulabel(label)
+        if s:
+            result.append(s)
+        else:
+            raise IDNAError("Empty label")
+    if trailing_dot:
+        result.append("")
+    return ".".join(result)
diff --git a/backend/venv/lib/python3.12/site-packages/idna/idnadata.py b/backend/venv/lib/python3.12/site-packages/idna/idnadata.py
new file mode 100644
index 00000000..4be60046
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/idnadata.py
@@ -0,0 +1,4243 @@
+# This file is automatically generated by tools/idna-data
+
+__version__ = "15.1.0"
+scripts = {
+    "Greek": (
+        0x37000000374,
+        0x37500000378,
+        0x37A0000037E,
+        0x37F00000380,
+        0x38400000385,
+        0x38600000387,
+        0x3880000038B,
+        0x38C0000038D,
+        0x38E000003A2,
+        0x3A3000003E2,
+        0x3F000000400,
+        0x1D2600001D2B,
+        0x1D5D00001D62,
+        0x1D6600001D6B,
+        0x1DBF00001DC0,
+        0x1F0000001F16,
+        0x1F1800001F1E,
+        0x1F2000001F46,
+        0x1F4800001F4E,
+        0x1F5000001F58,
+        0x1F5900001F5A,
+        0x1F5B00001F5C,
+        0x1F5D00001F5E,
+        0x1F5F00001F7E,
+        0x1F8000001FB5,
+        0x1FB600001FC5,
+        0x1FC600001FD4,
+        0x1FD600001FDC,
+        0x1FDD00001FF0,
+        0x1FF200001FF5,
+        0x1FF600001FFF,
+        0x212600002127,
+        0xAB650000AB66,
+        0x101400001018F,
+        0x101A0000101A1,
+        0x1D2000001D246,
+    ),
+    "Han": (
+        0x2E8000002E9A,
+        0x2E9B00002EF4,
+        0x2F0000002FD6,
+        0x300500003006,
+        0x300700003008,
+        0x30210000302A,
+        0x30380000303C,
+        0x340000004DC0,
+        0x4E000000A000,
+        0xF9000000FA6E,
+        0xFA700000FADA,
+        0x16FE200016FE4,
+        0x16FF000016FF2,
+        0x200000002A6E0,
+        0x2A7000002B73A,
+        0x2B7400002B81E,
+        0x2B8200002CEA2,
+        0x2CEB00002EBE1,
+        0x2EBF00002EE5E,
+        0x2F8000002FA1E,
+        0x300000003134B,
+        0x31350000323B0,
+    ),
+    "Hebrew": (
+        0x591000005C8,
+        0x5D0000005EB,
+        0x5EF000005F5,
+        0xFB1D0000FB37,
+        0xFB380000FB3D,
+        0xFB3E0000FB3F,
+        0xFB400000FB42,
+        0xFB430000FB45,
+        0xFB460000FB50,
+    ),
+    "Hiragana": (
+        0x304100003097,
+        0x309D000030A0,
+        0x1B0010001B120,
+        0x1B1320001B133,
+        0x1B1500001B153,
+        0x1F2000001F201,
+    ),
+    "Katakana": (
+        0x30A1000030FB,
+        0x30FD00003100,
+        0x31F000003200,
+        0x32D0000032FF,
+        0x330000003358,
+        0xFF660000FF70,
+        0xFF710000FF9E,
+        0x1AFF00001AFF4,
+        0x1AFF50001AFFC,
+        0x1AFFD0001AFFF,
+        0x1B0000001B001,
+        0x1B1200001B123,
+        0x1B1550001B156,
+        0x1B1640001B168,
+    ),
+}
+joining_types = {
+    0xAD: 84,
+    0x300: 84,
+    0x301: 84,
+    0x302: 84,
+    0x303: 84,
+    0x304: 84,
+    0x305: 84,
+    0x306: 84,
+    0x307: 84,
+    0x308: 84,
+    0x309: 84,
+    0x30A: 84,
+    0x30B: 84,
+    0x30C: 84,
+    0x30D: 84,
+    0x30E: 84,
+    0x30F: 84,
+    0x310: 84,
+    0x311: 84,
+    0x312: 84,
+    0x313: 84,
+    0x314: 84,
+    0x315: 84,
+    0x316: 84,
+    0x317: 84,
+    0x318: 84,
+    0x319: 84,
+    0x31A: 84,
+    0x31B: 84,
+    0x31C: 84,
+    0x31D: 84,
+    0x31E: 84,
+    0x31F: 84,
+    0x320: 84,
+    0x321: 84,
+    0x322: 84,
+    0x323: 84,
+    0x324: 84,
+    0x325: 84,
+    0x326: 84,
+    0x327: 84,
+    0x328: 84,
+    0x329: 84,
+    0x32A: 84,
+    0x32B: 84,
+    0x32C: 84,
+    0x32D: 84,
+    0x32E: 84,
+    0x32F: 84,
+    0x330: 84,
+    0x331: 84,
+    0x332: 84,
+    0x333: 84,
+    0x334: 84,
+    0x335: 84,
+    0x336: 84,
+    0x337: 84,
+    0x338: 84,
+    0x339: 84,
+    0x33A: 84,
+    0x33B: 84,
+    0x33C: 84,
+    0x33D: 84,
+    0x33E: 84,
+    0x33F: 84,
+    0x340: 84,
+    0x341: 84,
+    0x342: 84,
+    0x343: 84,
+    0x344: 84,
+    0x345: 84,
+    0x346: 84,
+    0x347: 84,
+    0x348: 84,
+    0x349: 84,
+    0x34A: 84,
+    0x34B: 84,
+    0x34C: 84,
+    0x34D: 84,
+    0x34E: 84,
+    0x34F: 84,
+    0x350: 84,
+    0x351: 84,
+    0x352: 84,
+    0x353: 84,
+    0x354: 84,
+    0x355: 84,
+    0x356: 84,
+    0x357: 84,
+    0x358: 84,
+    0x359: 84,
+    0x35A: 84,
+    0x35B: 84,
+    0x35C: 84,
+    0x35D: 84,
+    0x35E: 84,
+    0x35F: 84,
+    0x360: 84,
+    0x361: 84,
+    0x362: 84,
+    0x363: 84,
+    0x364: 84,
+    0x365: 84,
+    0x366: 84,
+    0x367: 84,
+    0x368: 84,
+    0x369: 84,
+    0x36A: 84,
+    0x36B: 84,
+    0x36C: 84,
+    0x36D: 84,
+    0x36E: 84,
+    0x36F: 84,
+    0x483: 84,
+    0x484: 84,
+    0x485: 84,
+    0x486: 84,
+    0x487: 84,
+    0x488: 84,
+    0x489: 84,
+    0x591: 84,
+    0x592: 84,
+    0x593: 84,
+    0x594: 84,
+    0x595: 84,
+    0x596: 84,
+    0x597: 84,
+    0x598: 84,
+    0x599: 84,
+    0x59A: 84,
+    0x59B: 84,
+    0x59C: 84,
+    0x59D: 84,
+    0x59E: 84,
+    0x59F: 84,
+    0x5A0: 84,
+    0x5A1: 84,
+    0x5A2: 84,
+    0x5A3: 84,
+    0x5A4: 84,
+    0x5A5: 84,
+    0x5A6: 84,
+    0x5A7: 84,
+    0x5A8: 84,
+    0x5A9: 84,
+    0x5AA: 84,
+    0x5AB: 84,
+    0x5AC: 84,
+    0x5AD: 84,
+    0x5AE: 84,
+    0x5AF: 84,
+    0x5B0: 84,
+    0x5B1: 84,
+    0x5B2: 84,
+    0x5B3: 84,
+    0x5B4: 84,
+    0x5B5: 84,
+    0x5B6: 84,
+    0x5B7: 84,
+    0x5B8: 84,
+    0x5B9: 84,
+    0x5BA: 84,
+    0x5BB: 84,
+    0x5BC: 84,
+    0x5BD: 84,
+    0x5BF: 84,
+    0x5C1: 84,
+    0x5C2: 84,
+    0x5C4: 84,
+    0x5C5: 84,
+    0x5C7: 84,
+    0x610: 84,
+    0x611: 84,
+    0x612: 84,
+    0x613: 84,
+    0x614: 84,
+    0x615: 84,
+    0x616: 84,
+    0x617: 84,
+    0x618: 84,
+    0x619: 84,
+    0x61A: 84,
+    0x61C: 84,
+    0x620: 68,
+    0x622: 82,
+    0x623: 82,
+    0x624: 82,
+    0x625: 82,
+    0x626: 68,
+    0x627: 82,
+    0x628: 68,
+    0x629: 82,
+    0x62A: 68,
+    0x62B: 68,
+    0x62C: 68,
+    0x62D: 68,
+    0x62E: 68,
+    0x62F: 82,
+    0x630: 82,
+    0x631: 82,
+    0x632: 82,
+    0x633: 68,
+    0x634: 68,
+    0x635: 68,
+    0x636: 68,
+    0x637: 68,
+    0x638: 68,
+    0x639: 68,
+    0x63A: 68,
+    0x63B: 68,
+    0x63C: 68,
+    0x63D: 68,
+    0x63E: 68,
+    0x63F: 68,
+    0x640: 67,
+    0x641: 68,
+    0x642: 68,
+    0x643: 68,
+    0x644: 68,
+    0x645: 68,
+    0x646: 68,
+    0x647: 68,
+    0x648: 82,
+    0x649: 68,
+    0x64A: 68,
+    0x64B: 84,
+    0x64C: 84,
+    0x64D: 84,
+    0x64E: 84,
+    0x64F: 84,
+    0x650: 84,
+    0x651: 84,
+    0x652: 84,
+    0x653: 84,
+    0x654: 84,
+    0x655: 84,
+    0x656: 84,
+    0x657: 84,
+    0x658: 84,
+    0x659: 84,
+    0x65A: 84,
+    0x65B: 84,
+    0x65C: 84,
+    0x65D: 84,
+    0x65E: 84,
+    0x65F: 84,
+    0x66E: 68,
+    0x66F: 68,
+    0x670: 84,
+    0x671: 82,
+    0x672: 82,
+    0x673: 82,
+    0x675: 82,
+    0x676: 82,
+    0x677: 82,
+    0x678: 68,
+    0x679: 68,
+    0x67A: 68,
+    0x67B: 68,
+    0x67C: 68,
+    0x67D: 68,
+    0x67E: 68,
+    0x67F: 68,
+    0x680: 68,
+    0x681: 68,
+    0x682: 68,
+    0x683: 68,
+    0x684: 68,
+    0x685: 68,
+    0x686: 68,
+    0x687: 68,
+    0x688: 82,
+    0x689: 82,
+    0x68A: 82,
+    0x68B: 82,
+    0x68C: 82,
+    0x68D: 82,
+    0x68E: 82,
+    0x68F: 82,
+    0x690: 82,
+    0x691: 82,
+    0x692: 82,
+    0x693: 82,
+    0x694: 82,
+    0x695: 82,
+    0x696: 82,
+    0x697: 82,
+    0x698: 82,
+    0x699: 82,
+    0x69A: 68,
+    0x69B: 68,
+    0x69C: 68,
+    0x69D: 68,
+    0x69E: 68,
+    0x69F: 68,
+    0x6A0: 68,
+    0x6A1: 68,
+    0x6A2: 68,
+    0x6A3: 68,
+    0x6A4: 68,
+    0x6A5: 68,
+    0x6A6: 68,
+    0x6A7: 68,
+    0x6A8: 68,
+    0x6A9: 68,
+    0x6AA: 68,
+    0x6AB: 68,
+    0x6AC: 68,
+    0x6AD: 68,
+    0x6AE: 68,
+    0x6AF: 68,
+    0x6B0: 68,
+    0x6B1: 68,
+    0x6B2: 68,
+    0x6B3: 68,
+    0x6B4: 68,
+    0x6B5: 68,
+    0x6B6: 68,
+    0x6B7: 68,
+    0x6B8: 68,
+    0x6B9: 68,
+    0x6BA: 68,
+    0x6BB: 68,
+    0x6BC: 68,
+    0x6BD: 68,
+    0x6BE: 68,
+    0x6BF: 68,
+    0x6C0: 82,
+    0x6C1: 68,
+    0x6C2: 68,
+    0x6C3: 82,
+    0x6C4: 82,
+    0x6C5: 82,
+    0x6C6: 82,
+    0x6C7: 82,
+    0x6C8: 82,
+    0x6C9: 82,
+    0x6CA: 82,
+    0x6CB: 82,
+    0x6CC: 68,
+    0x6CD: 82,
+    0x6CE: 68,
+    0x6CF: 82,
+    0x6D0: 68,
+    0x6D1: 68,
+    0x6D2: 82,
+    0x6D3: 82,
+    0x6D5: 82,
+    0x6D6: 84,
+    0x6D7: 84,
+    0x6D8: 84,
+    0x6D9: 84,
+    0x6DA: 84,
+    0x6DB: 84,
+    0x6DC: 84,
+    0x6DF: 84,
+    0x6E0: 84,
+    0x6E1: 84,
+    0x6E2: 84,
+    0x6E3: 84,
+    0x6E4: 84,
+    0x6E7: 84,
+    0x6E8: 84,
+    0x6EA: 84,
+    0x6EB: 84,
+    0x6EC: 84,
+    0x6ED: 84,
+    0x6EE: 82,
+    0x6EF: 82,
+    0x6FA: 68,
+    0x6FB: 68,
+    0x6FC: 68,
+    0x6FF: 68,
+    0x70F: 84,
+    0x710: 82,
+    0x711: 84,
+    0x712: 68,
+    0x713: 68,
+    0x714: 68,
+    0x715: 82,
+    0x716: 82,
+    0x717: 82,
+    0x718: 82,
+    0x719: 82,
+    0x71A: 68,
+    0x71B: 68,
+    0x71C: 68,
+    0x71D: 68,
+    0x71E: 82,
+    0x71F: 68,
+    0x720: 68,
+    0x721: 68,
+    0x722: 68,
+    0x723: 68,
+    0x724: 68,
+    0x725: 68,
+    0x726: 68,
+    0x727: 68,
+    0x728: 82,
+    0x729: 68,
+    0x72A: 82,
+    0x72B: 68,
+    0x72C: 82,
+    0x72D: 68,
+    0x72E: 68,
+    0x72F: 82,
+    0x730: 84,
+    0x731: 84,
+    0x732: 84,
+    0x733: 84,
+    0x734: 84,
+    0x735: 84,
+    0x736: 84,
+    0x737: 84,
+    0x738: 84,
+    0x739: 84,
+    0x73A: 84,
+    0x73B: 84,
+    0x73C: 84,
+    0x73D: 84,
+    0x73E: 84,
+    0x73F: 84,
+    0x740: 84,
+    0x741: 84,
+    0x742: 84,
+    0x743: 84,
+    0x744: 84,
+    0x745: 84,
+    0x746: 84,
+    0x747: 84,
+    0x748: 84,
+    0x749: 84,
+    0x74A: 84,
+    0x74D: 82,
+    0x74E: 68,
+    0x74F: 68,
+    0x750: 68,
+    0x751: 68,
+    0x752: 68,
+    0x753: 68,
+    0x754: 68,
+    0x755: 68,
+    0x756: 68,
+    0x757: 68,
+    0x758: 68,
+    0x759: 82,
+    0x75A: 82,
+    0x75B: 82,
+    0x75C: 68,
+    0x75D: 68,
+    0x75E: 68,
+    0x75F: 68,
+    0x760: 68,
+    0x761: 68,
+    0x762: 68,
+    0x763: 68,
+    0x764: 68,
+    0x765: 68,
+    0x766: 68,
+    0x767: 68,
+    0x768: 68,
+    0x769: 68,
+    0x76A: 68,
+    0x76B: 82,
+    0x76C: 82,
+    0x76D: 68,
+    0x76E: 68,
+    0x76F: 68,
+    0x770: 68,
+    0x771: 82,
+    0x772: 68,
+    0x773: 82,
+    0x774: 82,
+    0x775: 68,
+    0x776: 68,
+    0x777: 68,
+    0x778: 82,
+    0x779: 82,
+    0x77A: 68,
+    0x77B: 68,
+    0x77C: 68,
+    0x77D: 68,
+    0x77E: 68,
+    0x77F: 68,
+    0x7A6: 84,
+    0x7A7: 84,
+    0x7A8: 84,
+    0x7A9: 84,
+    0x7AA: 84,
+    0x7AB: 84,
+    0x7AC: 84,
+    0x7AD: 84,
+    0x7AE: 84,
+    0x7AF: 84,
+    0x7B0: 84,
+    0x7CA: 68,
+    0x7CB: 68,
+    0x7CC: 68,
+    0x7CD: 68,
+    0x7CE: 68,
+    0x7CF: 68,
+    0x7D0: 68,
+    0x7D1: 68,
+    0x7D2: 68,
+    0x7D3: 68,
+    0x7D4: 68,
+    0x7D5: 68,
+    0x7D6: 68,
+    0x7D7: 68,
+    0x7D8: 68,
+    0x7D9: 68,
+    0x7DA: 68,
+    0x7DB: 68,
+    0x7DC: 68,
+    0x7DD: 68,
+    0x7DE: 68,
+    0x7DF: 68,
+    0x7E0: 68,
+    0x7E1: 68,
+    0x7E2: 68,
+    0x7E3: 68,
+    0x7E4: 68,
+    0x7E5: 68,
+    0x7E6: 68,
+    0x7E7: 68,
+    0x7E8: 68,
+    0x7E9: 68,
+    0x7EA: 68,
+    0x7EB: 84,
+    0x7EC: 84,
+    0x7ED: 84,
+    0x7EE: 84,
+    0x7EF: 84,
+    0x7F0: 84,
+    0x7F1: 84,
+    0x7F2: 84,
+    0x7F3: 84,
+    0x7FA: 67,
+    0x7FD: 84,
+    0x816: 84,
+    0x817: 84,
+    0x818: 84,
+    0x819: 84,
+    0x81B: 84,
+    0x81C: 84,
+    0x81D: 84,
+    0x81E: 84,
+    0x81F: 84,
+    0x820: 84,
+    0x821: 84,
+    0x822: 84,
+    0x823: 84,
+    0x825: 84,
+    0x826: 84,
+    0x827: 84,
+    0x829: 84,
+    0x82A: 84,
+    0x82B: 84,
+    0x82C: 84,
+    0x82D: 84,
+    0x840: 82,
+    0x841: 68,
+    0x842: 68,
+    0x843: 68,
+    0x844: 68,
+    0x845: 68,
+    0x846: 82,
+    0x847: 82,
+    0x848: 68,
+    0x849: 82,
+    0x84A: 68,
+    0x84B: 68,
+    0x84C: 68,
+    0x84D: 68,
+    0x84E: 68,
+    0x84F: 68,
+    0x850: 68,
+    0x851: 68,
+    0x852: 68,
+    0x853: 68,
+    0x854: 82,
+    0x855: 68,
+    0x856: 82,
+    0x857: 82,
+    0x858: 82,
+    0x859: 84,
+    0x85A: 84,
+    0x85B: 84,
+    0x860: 68,
+    0x862: 68,
+    0x863: 68,
+    0x864: 68,
+    0x865: 68,
+    0x867: 82,
+    0x868: 68,
+    0x869: 82,
+    0x86A: 82,
+    0x870: 82,
+    0x871: 82,
+    0x872: 82,
+    0x873: 82,
+    0x874: 82,
+    0x875: 82,
+    0x876: 82,
+    0x877: 82,
+    0x878: 82,
+    0x879: 82,
+    0x87A: 82,
+    0x87B: 82,
+    0x87C: 82,
+    0x87D: 82,
+    0x87E: 82,
+    0x87F: 82,
+    0x880: 82,
+    0x881: 82,
+    0x882: 82,
+    0x883: 67,
+    0x884: 67,
+    0x885: 67,
+    0x886: 68,
+    0x889: 68,
+    0x88A: 68,
+    0x88B: 68,
+    0x88C: 68,
+    0x88D: 68,
+    0x88E: 82,
+    0x898: 84,
+    0x899: 84,
+    0x89A: 84,
+    0x89B: 84,
+    0x89C: 84,
+    0x89D: 84,
+    0x89E: 84,
+    0x89F: 84,
+    0x8A0: 68,
+    0x8A1: 68,
+    0x8A2: 68,
+    0x8A3: 68,
+    0x8A4: 68,
+    0x8A5: 68,
+    0x8A6: 68,
+    0x8A7: 68,
+    0x8A8: 68,
+    0x8A9: 68,
+    0x8AA: 82,
+    0x8AB: 82,
+    0x8AC: 82,
+    0x8AE: 82,
+    0x8AF: 68,
+    0x8B0: 68,
+    0x8B1: 82,
+    0x8B2: 82,
+    0x8B3: 68,
+    0x8B4: 68,
+    0x8B5: 68,
+    0x8B6: 68,
+    0x8B7: 68,
+    0x8B8: 68,
+    0x8B9: 82,
+    0x8BA: 68,
+    0x8BB: 68,
+    0x8BC: 68,
+    0x8BD: 68,
+    0x8BE: 68,
+    0x8BF: 68,
+    0x8C0: 68,
+    0x8C1: 68,
+    0x8C2: 68,
+    0x8C3: 68,
+    0x8C4: 68,
+    0x8C5: 68,
+    0x8C6: 68,
+    0x8C7: 68,
+    0x8C8: 68,
+    0x8CA: 84,
+    0x8CB: 84,
+    0x8CC: 84,
+    0x8CD: 84,
+    0x8CE: 84,
+    0x8CF: 84,
+    0x8D0: 84,
+    0x8D1: 84,
+    0x8D2: 84,
+    0x8D3: 84,
+    0x8D4: 84,
+    0x8D5: 84,
+    0x8D6: 84,
+    0x8D7: 84,
+    0x8D8: 84,
+    0x8D9: 84,
+    0x8DA: 84,
+    0x8DB: 84,
+    0x8DC: 84,
+    0x8DD: 84,
+    0x8DE: 84,
+    0x8DF: 84,
+    0x8E0: 84,
+    0x8E1: 84,
+    0x8E3: 84,
+    0x8E4: 84,
+    0x8E5: 84,
+    0x8E6: 84,
+    0x8E7: 84,
+    0x8E8: 84,
+    0x8E9: 84,
+    0x8EA: 84,
+    0x8EB: 84,
+    0x8EC: 84,
+    0x8ED: 84,
+    0x8EE: 84,
+    0x8EF: 84,
+    0x8F0: 84,
+    0x8F1: 84,
+    0x8F2: 84,
+    0x8F3: 84,
+    0x8F4: 84,
+    0x8F5: 84,
+    0x8F6: 84,
+    0x8F7: 84,
+    0x8F8: 84,
+    0x8F9: 84,
+    0x8FA: 84,
+    0x8FB: 84,
+    0x8FC: 84,
+    0x8FD: 84,
+    0x8FE: 84,
+    0x8FF: 84,
+    0x900: 84,
+    0x901: 84,
+    0x902: 84,
+    0x93A: 84,
+    0x93C: 84,
+    0x941: 84,
+    0x942: 84,
+    0x943: 84,
+    0x944: 84,
+    0x945: 84,
+    0x946: 84,
+    0x947: 84,
+    0x948: 84,
+    0x94D: 84,
+    0x951: 84,
+    0x952: 84,
+    0x953: 84,
+    0x954: 84,
+    0x955: 84,
+    0x956: 84,
+    0x957: 84,
+    0x962: 84,
+    0x963: 84,
+    0x981: 84,
+    0x9BC: 84,
+    0x9C1: 84,
+    0x9C2: 84,
+    0x9C3: 84,
+    0x9C4: 84,
+    0x9CD: 84,
+    0x9E2: 84,
+    0x9E3: 84,
+    0x9FE: 84,
+    0xA01: 84,
+    0xA02: 84,
+    0xA3C: 84,
+    0xA41: 84,
+    0xA42: 84,
+    0xA47: 84,
+    0xA48: 84,
+    0xA4B: 84,
+    0xA4C: 84,
+    0xA4D: 84,
+    0xA51: 84,
+    0xA70: 84,
+    0xA71: 84,
+    0xA75: 84,
+    0xA81: 84,
+    0xA82: 84,
+    0xABC: 84,
+    0xAC1: 84,
+    0xAC2: 84,
+    0xAC3: 84,
+    0xAC4: 84,
+    0xAC5: 84,
+    0xAC7: 84,
+    0xAC8: 84,
+    0xACD: 84,
+    0xAE2: 84,
+    0xAE3: 84,
+    0xAFA: 84,
+    0xAFB: 84,
+    0xAFC: 84,
+    0xAFD: 84,
+    0xAFE: 84,
+    0xAFF: 84,
+    0xB01: 84,
+    0xB3C: 84,
+    0xB3F: 84,
+    0xB41: 84,
+    0xB42: 84,
+    0xB43: 84,
+    0xB44: 84,
+    0xB4D: 84,
+    0xB55: 84,
+    0xB56: 84,
+    0xB62: 84,
+    0xB63: 84,
+    0xB82: 84,
+    0xBC0: 84,
+    0xBCD: 84,
+    0xC00: 84,
+    0xC04: 84,
+    0xC3C: 84,
+    0xC3E: 84,
+    0xC3F: 84,
+    0xC40: 84,
+    0xC46: 84,
+    0xC47: 84,
+    0xC48: 84,
+    0xC4A: 84,
+    0xC4B: 84,
+    0xC4C: 84,
+    0xC4D: 84,
+    0xC55: 84,
+    0xC56: 84,
+    0xC62: 84,
+    0xC63: 84,
+    0xC81: 84,
+    0xCBC: 84,
+    0xCBF: 84,
+    0xCC6: 84,
+    0xCCC: 84,
+    0xCCD: 84,
+    0xCE2: 84,
+    0xCE3: 84,
+    0xD00: 84,
+    0xD01: 84,
+    0xD3B: 84,
+    0xD3C: 84,
+    0xD41: 84,
+    0xD42: 84,
+    0xD43: 84,
+    0xD44: 84,
+    0xD4D: 84,
+    0xD62: 84,
+    0xD63: 84,
+    0xD81: 84,
+    0xDCA: 84,
+    0xDD2: 84,
+    0xDD3: 84,
+    0xDD4: 84,
+    0xDD6: 84,
+    0xE31: 84,
+    0xE34: 84,
+    0xE35: 84,
+    0xE36: 84,
+    0xE37: 84,
+    0xE38: 84,
+    0xE39: 84,
+    0xE3A: 84,
+    0xE47: 84,
+    0xE48: 84,
+    0xE49: 84,
+    0xE4A: 84,
+    0xE4B: 84,
+    0xE4C: 84,
+    0xE4D: 84,
+    0xE4E: 84,
+    0xEB1: 84,
+    0xEB4: 84,
+    0xEB5: 84,
+    0xEB6: 84,
+    0xEB7: 84,
+    0xEB8: 84,
+    0xEB9: 84,
+    0xEBA: 84,
+    0xEBB: 84,
+    0xEBC: 84,
+    0xEC8: 84,
+    0xEC9: 84,
+    0xECA: 84,
+    0xECB: 84,
+    0xECC: 84,
+    0xECD: 84,
+    0xECE: 84,
+    0xF18: 84,
+    0xF19: 84,
+    0xF35: 84,
+    0xF37: 84,
+    0xF39: 84,
+    0xF71: 84,
+    0xF72: 84,
+    0xF73: 84,
+    0xF74: 84,
+    0xF75: 84,
+    0xF76: 84,
+    0xF77: 84,
+    0xF78: 84,
+    0xF79: 84,
+    0xF7A: 84,
+    0xF7B: 84,
+    0xF7C: 84,
+    0xF7D: 84,
+    0xF7E: 84,
+    0xF80: 84,
+    0xF81: 84,
+    0xF82: 84,
+    0xF83: 84,
+    0xF84: 84,
+    0xF86: 84,
+    0xF87: 84,
+    0xF8D: 84,
+    0xF8E: 84,
+    0xF8F: 84,
+    0xF90: 84,
+    0xF91: 84,
+    0xF92: 84,
+    0xF93: 84,
+    0xF94: 84,
+    0xF95: 84,
+    0xF96: 84,
+    0xF97: 84,
+    0xF99: 84,
+    0xF9A: 84,
+    0xF9B: 84,
+    0xF9C: 84,
+    0xF9D: 84,
+    0xF9E: 84,
+    0xF9F: 84,
+    0xFA0: 84,
+    0xFA1: 84,
+    0xFA2: 84,
+    0xFA3: 84,
+    0xFA4: 84,
+    0xFA5: 84,
+    0xFA6: 84,
+    0xFA7: 84,
+    0xFA8: 84,
+    0xFA9: 84,
+    0xFAA: 84,
+    0xFAB: 84,
+    0xFAC: 84,
+    0xFAD: 84,
+    0xFAE: 84,
+    0xFAF: 84,
+    0xFB0: 84,
+    0xFB1: 84,
+    0xFB2: 84,
+    0xFB3: 84,
+    0xFB4: 84,
+    0xFB5: 84,
+    0xFB6: 84,
+    0xFB7: 84,
+    0xFB8: 84,
+    0xFB9: 84,
+    0xFBA: 84,
+    0xFBB: 84,
+    0xFBC: 84,
+    0xFC6: 84,
+    0x102D: 84,
+    0x102E: 84,
+    0x102F: 84,
+    0x1030: 84,
+    0x1032: 84,
+    0x1033: 84,
+    0x1034: 84,
+    0x1035: 84,
+    0x1036: 84,
+    0x1037: 84,
+    0x1039: 84,
+    0x103A: 84,
+    0x103D: 84,
+    0x103E: 84,
+    0x1058: 84,
+    0x1059: 84,
+    0x105E: 84,
+    0x105F: 84,
+    0x1060: 84,
+    0x1071: 84,
+    0x1072: 84,
+    0x1073: 84,
+    0x1074: 84,
+    0x1082: 84,
+    0x1085: 84,
+    0x1086: 84,
+    0x108D: 84,
+    0x109D: 84,
+    0x135D: 84,
+    0x135E: 84,
+    0x135F: 84,
+    0x1712: 84,
+    0x1713: 84,
+    0x1714: 84,
+    0x1732: 84,
+    0x1733: 84,
+    0x1752: 84,
+    0x1753: 84,
+    0x1772: 84,
+    0x1773: 84,
+    0x17B4: 84,
+    0x17B5: 84,
+    0x17B7: 84,
+    0x17B8: 84,
+    0x17B9: 84,
+    0x17BA: 84,
+    0x17BB: 84,
+    0x17BC: 84,
+    0x17BD: 84,
+    0x17C6: 84,
+    0x17C9: 84,
+    0x17CA: 84,
+    0x17CB: 84,
+    0x17CC: 84,
+    0x17CD: 84,
+    0x17CE: 84,
+    0x17CF: 84,
+    0x17D0: 84,
+    0x17D1: 84,
+    0x17D2: 84,
+    0x17D3: 84,
+    0x17DD: 84,
+    0x1807: 68,
+    0x180A: 67,
+    0x180B: 84,
+    0x180C: 84,
+    0x180D: 84,
+    0x180F: 84,
+    0x1820: 68,
+    0x1821: 68,
+    0x1822: 68,
+    0x1823: 68,
+    0x1824: 68,
+    0x1825: 68,
+    0x1826: 68,
+    0x1827: 68,
+    0x1828: 68,
+    0x1829: 68,
+    0x182A: 68,
+    0x182B: 68,
+    0x182C: 68,
+    0x182D: 68,
+    0x182E: 68,
+    0x182F: 68,
+    0x1830: 68,
+    0x1831: 68,
+    0x1832: 68,
+    0x1833: 68,
+    0x1834: 68,
+    0x1835: 68,
+    0x1836: 68,
+    0x1837: 68,
+    0x1838: 68,
+    0x1839: 68,
+    0x183A: 68,
+    0x183B: 68,
+    0x183C: 68,
+    0x183D: 68,
+    0x183E: 68,
+    0x183F: 68,
+    0x1840: 68,
+    0x1841: 68,
+    0x1842: 68,
+    0x1843: 68,
+    0x1844: 68,
+    0x1845: 68,
+    0x1846: 68,
+    0x1847: 68,
+    0x1848: 68,
+    0x1849: 68,
+    0x184A: 68,
+    0x184B: 68,
+    0x184C: 68,
+    0x184D: 68,
+    0x184E: 68,
+    0x184F: 68,
+    0x1850: 68,
+    0x1851: 68,
+    0x1852: 68,
+    0x1853: 68,
+    0x1854: 68,
+    0x1855: 68,
+    0x1856: 68,
+    0x1857: 68,
+    0x1858: 68,
+    0x1859: 68,
+    0x185A: 68,
+    0x185B: 68,
+    0x185C: 68,
+    0x185D: 68,
+    0x185E: 68,
+    0x185F: 68,
+    0x1860: 68,
+    0x1861: 68,
+    0x1862: 68,
+    0x1863: 68,
+    0x1864: 68,
+    0x1865: 68,
+    0x1866: 68,
+    0x1867: 68,
+    0x1868: 68,
+    0x1869: 68,
+    0x186A: 68,
+    0x186B: 68,
+    0x186C: 68,
+    0x186D: 68,
+    0x186E: 68,
+    0x186F: 68,
+    0x1870: 68,
+    0x1871: 68,
+    0x1872: 68,
+    0x1873: 68,
+    0x1874: 68,
+    0x1875: 68,
+    0x1876: 68,
+    0x1877: 68,
+    0x1878: 68,
+    0x1885: 84,
+    0x1886: 84,
+    0x1887: 68,
+    0x1888: 68,
+    0x1889: 68,
+    0x188A: 68,
+    0x188B: 68,
+    0x188C: 68,
+    0x188D: 68,
+    0x188E: 68,
+    0x188F: 68,
+    0x1890: 68,
+    0x1891: 68,
+    0x1892: 68,
+    0x1893: 68,
+    0x1894: 68,
+    0x1895: 68,
+    0x1896: 68,
+    0x1897: 68,
+    0x1898: 68,
+    0x1899: 68,
+    0x189A: 68,
+    0x189B: 68,
+    0x189C: 68,
+    0x189D: 68,
+    0x189E: 68,
+    0x189F: 68,
+    0x18A0: 68,
+    0x18A1: 68,
+    0x18A2: 68,
+    0x18A3: 68,
+    0x18A4: 68,
+    0x18A5: 68,
+    0x18A6: 68,
+    0x18A7: 68,
+    0x18A8: 68,
+    0x18A9: 84,
+    0x18AA: 68,
+    0x1920: 84,
+    0x1921: 84,
+    0x1922: 84,
+    0x1927: 84,
+    0x1928: 84,
+    0x1932: 84,
+    0x1939: 84,
+    0x193A: 84,
+    0x193B: 84,
+    0x1A17: 84,
+    0x1A18: 84,
+    0x1A1B: 84,
+    0x1A56: 84,
+    0x1A58: 84,
+    0x1A59: 84,
+    0x1A5A: 84,
+    0x1A5B: 84,
+    0x1A5C: 84,
+    0x1A5D: 84,
+    0x1A5E: 84,
+    0x1A60: 84,
+    0x1A62: 84,
+    0x1A65: 84,
+    0x1A66: 84,
+    0x1A67: 84,
+    0x1A68: 84,
+    0x1A69: 84,
+    0x1A6A: 84,
+    0x1A6B: 84,
+    0x1A6C: 84,
+    0x1A73: 84,
+    0x1A74: 84,
+    0x1A75: 84,
+    0x1A76: 84,
+    0x1A77: 84,
+    0x1A78: 84,
+    0x1A79: 84,
+    0x1A7A: 84,
+    0x1A7B: 84,
+    0x1A7C: 84,
+    0x1A7F: 84,
+    0x1AB0: 84,
+    0x1AB1: 84,
+    0x1AB2: 84,
+    0x1AB3: 84,
+    0x1AB4: 84,
+    0x1AB5: 84,
+    0x1AB6: 84,
+    0x1AB7: 84,
+    0x1AB8: 84,
+    0x1AB9: 84,
+    0x1ABA: 84,
+    0x1ABB: 84,
+    0x1ABC: 84,
+    0x1ABD: 84,
+    0x1ABE: 84,
+    0x1ABF: 84,
+    0x1AC0: 84,
+    0x1AC1: 84,
+    0x1AC2: 84,
+    0x1AC3: 84,
+    0x1AC4: 84,
+    0x1AC5: 84,
+    0x1AC6: 84,
+    0x1AC7: 84,
+    0x1AC8: 84,
+    0x1AC9: 84,
+    0x1ACA: 84,
+    0x1ACB: 84,
+    0x1ACC: 84,
+    0x1ACD: 84,
+    0x1ACE: 84,
+    0x1B00: 84,
+    0x1B01: 84,
+    0x1B02: 84,
+    0x1B03: 84,
+    0x1B34: 84,
+    0x1B36: 84,
+    0x1B37: 84,
+    0x1B38: 84,
+    0x1B39: 84,
+    0x1B3A: 84,
+    0x1B3C: 84,
+    0x1B42: 84,
+    0x1B6B: 84,
+    0x1B6C: 84,
+    0x1B6D: 84,
+    0x1B6E: 84,
+    0x1B6F: 84,
+    0x1B70: 84,
+    0x1B71: 84,
+    0x1B72: 84,
+    0x1B73: 84,
+    0x1B80: 84,
+    0x1B81: 84,
+    0x1BA2: 84,
+    0x1BA3: 84,
+    0x1BA4: 84,
+    0x1BA5: 84,
+    0x1BA8: 84,
+    0x1BA9: 84,
+    0x1BAB: 84,
+    0x1BAC: 84,
+    0x1BAD: 84,
+    0x1BE6: 84,
+    0x1BE8: 84,
+    0x1BE9: 84,
+    0x1BED: 84,
+    0x1BEF: 84,
+    0x1BF0: 84,
+    0x1BF1: 84,
+    0x1C2C: 84,
+    0x1C2D: 84,
+    0x1C2E: 84,
+    0x1C2F: 84,
+    0x1C30: 84,
+    0x1C31: 84,
+    0x1C32: 84,
+    0x1C33: 84,
+    0x1C36: 84,
+    0x1C37: 84,
+    0x1CD0: 84,
+    0x1CD1: 84,
+    0x1CD2: 84,
+    0x1CD4: 84,
+    0x1CD5: 84,
+    0x1CD6: 84,
+    0x1CD7: 84,
+    0x1CD8: 84,
+    0x1CD9: 84,
+    0x1CDA: 84,
+    0x1CDB: 84,
+    0x1CDC: 84,
+    0x1CDD: 84,
+    0x1CDE: 84,
+    0x1CDF: 84,
+    0x1CE0: 84,
+    0x1CE2: 84,
+    0x1CE3: 84,
+    0x1CE4: 84,
+    0x1CE5: 84,
+    0x1CE6: 84,
+    0x1CE7: 84,
+    0x1CE8: 84,
+    0x1CED: 84,
+    0x1CF4: 84,
+    0x1CF8: 84,
+    0x1CF9: 84,
+    0x1DC0: 84,
+    0x1DC1: 84,
+    0x1DC2: 84,
+    0x1DC3: 84,
+    0x1DC4: 84,
+    0x1DC5: 84,
+    0x1DC6: 84,
+    0x1DC7: 84,
+    0x1DC8: 84,
+    0x1DC9: 84,
+    0x1DCA: 84,
+    0x1DCB: 84,
+    0x1DCC: 84,
+    0x1DCD: 84,
+    0x1DCE: 84,
+    0x1DCF: 84,
+    0x1DD0: 84,
+    0x1DD1: 84,
+    0x1DD2: 84,
+    0x1DD3: 84,
+    0x1DD4: 84,
+    0x1DD5: 84,
+    0x1DD6: 84,
+    0x1DD7: 84,
+    0x1DD8: 84,
+    0x1DD9: 84,
+    0x1DDA: 84,
+    0x1DDB: 84,
+    0x1DDC: 84,
+    0x1DDD: 84,
+    0x1DDE: 84,
+    0x1DDF: 84,
+    0x1DE0: 84,
+    0x1DE1: 84,
+    0x1DE2: 84,
+    0x1DE3: 84,
+    0x1DE4: 84,
+    0x1DE5: 84,
+    0x1DE6: 84,
+    0x1DE7: 84,
+    0x1DE8: 84,
+    0x1DE9: 84,
+    0x1DEA: 84,
+    0x1DEB: 84,
+    0x1DEC: 84,
+    0x1DED: 84,
+    0x1DEE: 84,
+    0x1DEF: 84,
+    0x1DF0: 84,
+    0x1DF1: 84,
+    0x1DF2: 84,
+    0x1DF3: 84,
+    0x1DF4: 84,
+    0x1DF5: 84,
+    0x1DF6: 84,
+    0x1DF7: 84,
+    0x1DF8: 84,
+    0x1DF9: 84,
+    0x1DFA: 84,
+    0x1DFB: 84,
+    0x1DFC: 84,
+    0x1DFD: 84,
+    0x1DFE: 84,
+    0x1DFF: 84,
+    0x200B: 84,
+    0x200D: 67,
+    0x200E: 84,
+    0x200F: 84,
+    0x202A: 84,
+    0x202B: 84,
+    0x202C: 84,
+    0x202D: 84,
+    0x202E: 84,
+    0x2060: 84,
+    0x2061: 84,
+    0x2062: 84,
+    0x2063: 84,
+    0x2064: 84,
+    0x206A: 84,
+    0x206B: 84,
+    0x206C: 84,
+    0x206D: 84,
+    0x206E: 84,
+    0x206F: 84,
+    0x20D0: 84,
+    0x20D1: 84,
+    0x20D2: 84,
+    0x20D3: 84,
+    0x20D4: 84,
+    0x20D5: 84,
+    0x20D6: 84,
+    0x20D7: 84,
+    0x20D8: 84,
+    0x20D9: 84,
+    0x20DA: 84,
+    0x20DB: 84,
+    0x20DC: 84,
+    0x20DD: 84,
+    0x20DE: 84,
+    0x20DF: 84,
+    0x20E0: 84,
+    0x20E1: 84,
+    0x20E2: 84,
+    0x20E3: 84,
+    0x20E4: 84,
+    0x20E5: 84,
+    0x20E6: 84,
+    0x20E7: 84,
+    0x20E8: 84,
+    0x20E9: 84,
+    0x20EA: 84,
+    0x20EB: 84,
+    0x20EC: 84,
+    0x20ED: 84,
+    0x20EE: 84,
+    0x20EF: 84,
+    0x20F0: 84,
+    0x2CEF: 84,
+    0x2CF0: 84,
+    0x2CF1: 84,
+    0x2D7F: 84,
+    0x2DE0: 84,
+    0x2DE1: 84,
+    0x2DE2: 84,
+    0x2DE3: 84,
+    0x2DE4: 84,
+    0x2DE5: 84,
+    0x2DE6: 84,
+    0x2DE7: 84,
+    0x2DE8: 84,
+    0x2DE9: 84,
+    0x2DEA: 84,
+    0x2DEB: 84,
+    0x2DEC: 84,
+    0x2DED: 84,
+    0x2DEE: 84,
+    0x2DEF: 84,
+    0x2DF0: 84,
+    0x2DF1: 84,
+    0x2DF2: 84,
+    0x2DF3: 84,
+    0x2DF4: 84,
+    0x2DF5: 84,
+    0x2DF6: 84,
+    0x2DF7: 84,
+    0x2DF8: 84,
+    0x2DF9: 84,
+    0x2DFA: 84,
+    0x2DFB: 84,
+    0x2DFC: 84,
+    0x2DFD: 84,
+    0x2DFE: 84,
+    0x2DFF: 84,
+    0x302A: 84,
+    0x302B: 84,
+    0x302C: 84,
+    0x302D: 84,
+    0x3099: 84,
+    0x309A: 84,
+    0xA66F: 84,
+    0xA670: 84,
+    0xA671: 84,
+    0xA672: 84,
+    0xA674: 84,
+    0xA675: 84,
+    0xA676: 84,
+    0xA677: 84,
+    0xA678: 84,
+    0xA679: 84,
+    0xA67A: 84,
+    0xA67B: 84,
+    0xA67C: 84,
+    0xA67D: 84,
+    0xA69E: 84,
+    0xA69F: 84,
+    0xA6F0: 84,
+    0xA6F1: 84,
+    0xA802: 84,
+    0xA806: 84,
+    0xA80B: 84,
+    0xA825: 84,
+    0xA826: 84,
+    0xA82C: 84,
+    0xA840: 68,
+    0xA841: 68,
+    0xA842: 68,
+    0xA843: 68,
+    0xA844: 68,
+    0xA845: 68,
+    0xA846: 68,
+    0xA847: 68,
+    0xA848: 68,
+    0xA849: 68,
+    0xA84A: 68,
+    0xA84B: 68,
+    0xA84C: 68,
+    0xA84D: 68,
+    0xA84E: 68,
+    0xA84F: 68,
+    0xA850: 68,
+    0xA851: 68,
+    0xA852: 68,
+    0xA853: 68,
+    0xA854: 68,
+    0xA855: 68,
+    0xA856: 68,
+    0xA857: 68,
+    0xA858: 68,
+    0xA859: 68,
+    0xA85A: 68,
+    0xA85B: 68,
+    0xA85C: 68,
+    0xA85D: 68,
+    0xA85E: 68,
+    0xA85F: 68,
+    0xA860: 68,
+    0xA861: 68,
+    0xA862: 68,
+    0xA863: 68,
+    0xA864: 68,
+    0xA865: 68,
+    0xA866: 68,
+    0xA867: 68,
+    0xA868: 68,
+    0xA869: 68,
+    0xA86A: 68,
+    0xA86B: 68,
+    0xA86C: 68,
+    0xA86D: 68,
+    0xA86E: 68,
+    0xA86F: 68,
+    0xA870: 68,
+    0xA871: 68,
+    0xA872: 76,
+    0xA8C4: 84,
+    0xA8C5: 84,
+    0xA8E0: 84,
+    0xA8E1: 84,
+    0xA8E2: 84,
+    0xA8E3: 84,
+    0xA8E4: 84,
+    0xA8E5: 84,
+    0xA8E6: 84,
+    0xA8E7: 84,
+    0xA8E8: 84,
+    0xA8E9: 84,
+    0xA8EA: 84,
+    0xA8EB: 84,
+    0xA8EC: 84,
+    0xA8ED: 84,
+    0xA8EE: 84,
+    0xA8EF: 84,
+    0xA8F0: 84,
+    0xA8F1: 84,
+    0xA8FF: 84,
+    0xA926: 84,
+    0xA927: 84,
+    0xA928: 84,
+    0xA929: 84,
+    0xA92A: 84,
+    0xA92B: 84,
+    0xA92C: 84,
+    0xA92D: 84,
+    0xA947: 84,
+    0xA948: 84,
+    0xA949: 84,
+    0xA94A: 84,
+    0xA94B: 84,
+    0xA94C: 84,
+    0xA94D: 84,
+    0xA94E: 84,
+    0xA94F: 84,
+    0xA950: 84,
+    0xA951: 84,
+    0xA980: 84,
+    0xA981: 84,
+    0xA982: 84,
+    0xA9B3: 84,
+    0xA9B6: 84,
+    0xA9B7: 84,
+    0xA9B8: 84,
+    0xA9B9: 84,
+    0xA9BC: 84,
+    0xA9BD: 84,
+    0xA9E5: 84,
+    0xAA29: 84,
+    0xAA2A: 84,
+    0xAA2B: 84,
+    0xAA2C: 84,
+    0xAA2D: 84,
+    0xAA2E: 84,
+    0xAA31: 84,
+    0xAA32: 84,
+    0xAA35: 84,
+    0xAA36: 84,
+    0xAA43: 84,
+    0xAA4C: 84,
+    0xAA7C: 84,
+    0xAAB0: 84,
+    0xAAB2: 84,
+    0xAAB3: 84,
+    0xAAB4: 84,
+    0xAAB7: 84,
+    0xAAB8: 84,
+    0xAABE: 84,
+    0xAABF: 84,
+    0xAAC1: 84,
+    0xAAEC: 84,
+    0xAAED: 84,
+    0xAAF6: 84,
+    0xABE5: 84,
+    0xABE8: 84,
+    0xABED: 84,
+    0xFB1E: 84,
+    0xFE00: 84,
+    0xFE01: 84,
+    0xFE02: 84,
+    0xFE03: 84,
+    0xFE04: 84,
+    0xFE05: 84,
+    0xFE06: 84,
+    0xFE07: 84,
+    0xFE08: 84,
+    0xFE09: 84,
+    0xFE0A: 84,
+    0xFE0B: 84,
+    0xFE0C: 84,
+    0xFE0D: 84,
+    0xFE0E: 84,
+    0xFE0F: 84,
+    0xFE20: 84,
+    0xFE21: 84,
+    0xFE22: 84,
+    0xFE23: 84,
+    0xFE24: 84,
+    0xFE25: 84,
+    0xFE26: 84,
+    0xFE27: 84,
+    0xFE28: 84,
+    0xFE29: 84,
+    0xFE2A: 84,
+    0xFE2B: 84,
+    0xFE2C: 84,
+    0xFE2D: 84,
+    0xFE2E: 84,
+    0xFE2F: 84,
+    0xFEFF: 84,
+    0xFFF9: 84,
+    0xFFFA: 84,
+    0xFFFB: 84,
+    0x101FD: 84,
+    0x102E0: 84,
+    0x10376: 84,
+    0x10377: 84,
+    0x10378: 84,
+    0x10379: 84,
+    0x1037A: 84,
+    0x10A01: 84,
+    0x10A02: 84,
+    0x10A03: 84,
+    0x10A05: 84,
+    0x10A06: 84,
+    0x10A0C: 84,
+    0x10A0D: 84,
+    0x10A0E: 84,
+    0x10A0F: 84,
+    0x10A38: 84,
+    0x10A39: 84,
+    0x10A3A: 84,
+    0x10A3F: 84,
+    0x10AC0: 68,
+    0x10AC1: 68,
+    0x10AC2: 68,
+    0x10AC3: 68,
+    0x10AC4: 68,
+    0x10AC5: 82,
+    0x10AC7: 82,
+    0x10AC9: 82,
+    0x10ACA: 82,
+    0x10ACD: 76,
+    0x10ACE: 82,
+    0x10ACF: 82,
+    0x10AD0: 82,
+    0x10AD1: 82,
+    0x10AD2: 82,
+    0x10AD3: 68,
+    0x10AD4: 68,
+    0x10AD5: 68,
+    0x10AD6: 68,
+    0x10AD7: 76,
+    0x10AD8: 68,
+    0x10AD9: 68,
+    0x10ADA: 68,
+    0x10ADB: 68,
+    0x10ADC: 68,
+    0x10ADD: 82,
+    0x10ADE: 68,
+    0x10ADF: 68,
+    0x10AE0: 68,
+    0x10AE1: 82,
+    0x10AE4: 82,
+    0x10AE5: 84,
+    0x10AE6: 84,
+    0x10AEB: 68,
+    0x10AEC: 68,
+    0x10AED: 68,
+    0x10AEE: 68,
+    0x10AEF: 82,
+    0x10B80: 68,
+    0x10B81: 82,
+    0x10B82: 68,
+    0x10B83: 82,
+    0x10B84: 82,
+    0x10B85: 82,
+    0x10B86: 68,
+    0x10B87: 68,
+    0x10B88: 68,
+    0x10B89: 82,
+    0x10B8A: 68,
+    0x10B8B: 68,
+    0x10B8C: 82,
+    0x10B8D: 68,
+    0x10B8E: 82,
+    0x10B8F: 82,
+    0x10B90: 68,
+    0x10B91: 82,
+    0x10BA9: 82,
+    0x10BAA: 82,
+    0x10BAB: 82,
+    0x10BAC: 82,
+    0x10BAD: 68,
+    0x10BAE: 68,
+    0x10D00: 76,
+    0x10D01: 68,
+    0x10D02: 68,
+    0x10D03: 68,
+    0x10D04: 68,
+    0x10D05: 68,
+    0x10D06: 68,
+    0x10D07: 68,
+    0x10D08: 68,
+    0x10D09: 68,
+    0x10D0A: 68,
+    0x10D0B: 68,
+    0x10D0C: 68,
+    0x10D0D: 68,
+    0x10D0E: 68,
+    0x10D0F: 68,
+    0x10D10: 68,
+    0x10D11: 68,
+    0x10D12: 68,
+    0x10D13: 68,
+    0x10D14: 68,
+    0x10D15: 68,
+    0x10D16: 68,
+    0x10D17: 68,
+    0x10D18: 68,
+    0x10D19: 68,
+    0x10D1A: 68,
+    0x10D1B: 68,
+    0x10D1C: 68,
+    0x10D1D: 68,
+    0x10D1E: 68,
+    0x10D1F: 68,
+    0x10D20: 68,
+    0x10D21: 68,
+    0x10D22: 82,
+    0x10D23: 68,
+    0x10D24: 84,
+    0x10D25: 84,
+    0x10D26: 84,
+    0x10D27: 84,
+    0x10EAB: 84,
+    0x10EAC: 84,
+    0x10EFD: 84,
+    0x10EFE: 84,
+    0x10EFF: 84,
+    0x10F30: 68,
+    0x10F31: 68,
+    0x10F32: 68,
+    0x10F33: 82,
+    0x10F34: 68,
+    0x10F35: 68,
+    0x10F36: 68,
+    0x10F37: 68,
+    0x10F38: 68,
+    0x10F39: 68,
+    0x10F3A: 68,
+    0x10F3B: 68,
+    0x10F3C: 68,
+    0x10F3D: 68,
+    0x10F3E: 68,
+    0x10F3F: 68,
+    0x10F40: 68,
+    0x10F41: 68,
+    0x10F42: 68,
+    0x10F43: 68,
+    0x10F44: 68,
+    0x10F46: 84,
+    0x10F47: 84,
+    0x10F48: 84,
+    0x10F49: 84,
+    0x10F4A: 84,
+    0x10F4B: 84,
+    0x10F4C: 84,
+    0x10F4D: 84,
+    0x10F4E: 84,
+    0x10F4F: 84,
+    0x10F50: 84,
+    0x10F51: 68,
+    0x10F52: 68,
+    0x10F53: 68,
+    0x10F54: 82,
+    0x10F70: 68,
+    0x10F71: 68,
+    0x10F72: 68,
+    0x10F73: 68,
+    0x10F74: 82,
+    0x10F75: 82,
+    0x10F76: 68,
+    0x10F77: 68,
+    0x10F78: 68,
+    0x10F79: 68,
+    0x10F7A: 68,
+    0x10F7B: 68,
+    0x10F7C: 68,
+    0x10F7D: 68,
+    0x10F7E: 68,
+    0x10F7F: 68,
+    0x10F80: 68,
+    0x10F81: 68,
+    0x10F82: 84,
+    0x10F83: 84,
+    0x10F84: 84,
+    0x10F85: 84,
+    0x10FB0: 68,
+    0x10FB2: 68,
+    0x10FB3: 68,
+    0x10FB4: 82,
+    0x10FB5: 82,
+    0x10FB6: 82,
+    0x10FB8: 68,
+    0x10FB9: 82,
+    0x10FBA: 82,
+    0x10FBB: 68,
+    0x10FBC: 68,
+    0x10FBD: 82,
+    0x10FBE: 68,
+    0x10FBF: 68,
+    0x10FC1: 68,
+    0x10FC2: 82,
+    0x10FC3: 82,
+    0x10FC4: 68,
+    0x10FC9: 82,
+    0x10FCA: 68,
+    0x10FCB: 76,
+    0x11001: 84,
+    0x11038: 84,
+    0x11039: 84,
+    0x1103A: 84,
+    0x1103B: 84,
+    0x1103C: 84,
+    0x1103D: 84,
+    0x1103E: 84,
+    0x1103F: 84,
+    0x11040: 84,
+    0x11041: 84,
+    0x11042: 84,
+    0x11043: 84,
+    0x11044: 84,
+    0x11045: 84,
+    0x11046: 84,
+    0x11070: 84,
+    0x11073: 84,
+    0x11074: 84,
+    0x1107F: 84,
+    0x11080: 84,
+    0x11081: 84,
+    0x110B3: 84,
+    0x110B4: 84,
+    0x110B5: 84,
+    0x110B6: 84,
+    0x110B9: 84,
+    0x110BA: 84,
+    0x110C2: 84,
+    0x11100: 84,
+    0x11101: 84,
+    0x11102: 84,
+    0x11127: 84,
+    0x11128: 84,
+    0x11129: 84,
+    0x1112A: 84,
+    0x1112B: 84,
+    0x1112D: 84,
+    0x1112E: 84,
+    0x1112F: 84,
+    0x11130: 84,
+    0x11131: 84,
+    0x11132: 84,
+    0x11133: 84,
+    0x11134: 84,
+    0x11173: 84,
+    0x11180: 84,
+    0x11181: 84,
+    0x111B6: 84,
+    0x111B7: 84,
+    0x111B8: 84,
+    0x111B9: 84,
+    0x111BA: 84,
+    0x111BB: 84,
+    0x111BC: 84,
+    0x111BD: 84,
+    0x111BE: 84,
+    0x111C9: 84,
+    0x111CA: 84,
+    0x111CB: 84,
+    0x111CC: 84,
+    0x111CF: 84,
+    0x1122F: 84,
+    0x11230: 84,
+    0x11231: 84,
+    0x11234: 84,
+    0x11236: 84,
+    0x11237: 84,
+    0x1123E: 84,
+    0x11241: 84,
+    0x112DF: 84,
+    0x112E3: 84,
+    0x112E4: 84,
+    0x112E5: 84,
+    0x112E6: 84,
+    0x112E7: 84,
+    0x112E8: 84,
+    0x112E9: 84,
+    0x112EA: 84,
+    0x11300: 84,
+    0x11301: 84,
+    0x1133B: 84,
+    0x1133C: 84,
+    0x11340: 84,
+    0x11366: 84,
+    0x11367: 84,
+    0x11368: 84,
+    0x11369: 84,
+    0x1136A: 84,
+    0x1136B: 84,
+    0x1136C: 84,
+    0x11370: 84,
+    0x11371: 84,
+    0x11372: 84,
+    0x11373: 84,
+    0x11374: 84,
+    0x11438: 84,
+    0x11439: 84,
+    0x1143A: 84,
+    0x1143B: 84,
+    0x1143C: 84,
+    0x1143D: 84,
+    0x1143E: 84,
+    0x1143F: 84,
+    0x11442: 84,
+    0x11443: 84,
+    0x11444: 84,
+    0x11446: 84,
+    0x1145E: 84,
+    0x114B3: 84,
+    0x114B4: 84,
+    0x114B5: 84,
+    0x114B6: 84,
+    0x114B7: 84,
+    0x114B8: 84,
+    0x114BA: 84,
+    0x114BF: 84,
+    0x114C0: 84,
+    0x114C2: 84,
+    0x114C3: 84,
+    0x115B2: 84,
+    0x115B3: 84,
+    0x115B4: 84,
+    0x115B5: 84,
+    0x115BC: 84,
+    0x115BD: 84,
+    0x115BF: 84,
+    0x115C0: 84,
+    0x115DC: 84,
+    0x115DD: 84,
+    0x11633: 84,
+    0x11634: 84,
+    0x11635: 84,
+    0x11636: 84,
+    0x11637: 84,
+    0x11638: 84,
+    0x11639: 84,
+    0x1163A: 84,
+    0x1163D: 84,
+    0x1163F: 84,
+    0x11640: 84,
+    0x116AB: 84,
+    0x116AD: 84,
+    0x116B0: 84,
+    0x116B1: 84,
+    0x116B2: 84,
+    0x116B3: 84,
+    0x116B4: 84,
+    0x116B5: 84,
+    0x116B7: 84,
+    0x1171D: 84,
+    0x1171E: 84,
+    0x1171F: 84,
+    0x11722: 84,
+    0x11723: 84,
+    0x11724: 84,
+    0x11725: 84,
+    0x11727: 84,
+    0x11728: 84,
+    0x11729: 84,
+    0x1172A: 84,
+    0x1172B: 84,
+    0x1182F: 84,
+    0x11830: 84,
+    0x11831: 84,
+    0x11832: 84,
+    0x11833: 84,
+    0x11834: 84,
+    0x11835: 84,
+    0x11836: 84,
+    0x11837: 84,
+    0x11839: 84,
+    0x1183A: 84,
+    0x1193B: 84,
+    0x1193C: 84,
+    0x1193E: 84,
+    0x11943: 84,
+    0x119D4: 84,
+    0x119D5: 84,
+    0x119D6: 84,
+    0x119D7: 84,
+    0x119DA: 84,
+    0x119DB: 84,
+    0x119E0: 84,
+    0x11A01: 84,
+    0x11A02: 84,
+    0x11A03: 84,
+    0x11A04: 84,
+    0x11A05: 84,
+    0x11A06: 84,
+    0x11A07: 84,
+    0x11A08: 84,
+    0x11A09: 84,
+    0x11A0A: 84,
+    0x11A33: 84,
+    0x11A34: 84,
+    0x11A35: 84,
+    0x11A36: 84,
+    0x11A37: 84,
+    0x11A38: 84,
+    0x11A3B: 84,
+    0x11A3C: 84,
+    0x11A3D: 84,
+    0x11A3E: 84,
+    0x11A47: 84,
+    0x11A51: 84,
+    0x11A52: 84,
+    0x11A53: 84,
+    0x11A54: 84,
+    0x11A55: 84,
+    0x11A56: 84,
+    0x11A59: 84,
+    0x11A5A: 84,
+    0x11A5B: 84,
+    0x11A8A: 84,
+    0x11A8B: 84,
+    0x11A8C: 84,
+    0x11A8D: 84,
+    0x11A8E: 84,
+    0x11A8F: 84,
+    0x11A90: 84,
+    0x11A91: 84,
+    0x11A92: 84,
+    0x11A93: 84,
+    0x11A94: 84,
+    0x11A95: 84,
+    0x11A96: 84,
+    0x11A98: 84,
+    0x11A99: 84,
+    0x11C30: 84,
+    0x11C31: 84,
+    0x11C32: 84,
+    0x11C33: 84,
+    0x11C34: 84,
+    0x11C35: 84,
+    0x11C36: 84,
+    0x11C38: 84,
+    0x11C39: 84,
+    0x11C3A: 84,
+    0x11C3B: 84,
+    0x11C3C: 84,
+    0x11C3D: 84,
+    0x11C3F: 84,
+    0x11C92: 84,
+    0x11C93: 84,
+    0x11C94: 84,
+    0x11C95: 84,
+    0x11C96: 84,
+    0x11C97: 84,
+    0x11C98: 84,
+    0x11C99: 84,
+    0x11C9A: 84,
+    0x11C9B: 84,
+    0x11C9C: 84,
+    0x11C9D: 84,
+    0x11C9E: 84,
+    0x11C9F: 84,
+    0x11CA0: 84,
+    0x11CA1: 84,
+    0x11CA2: 84,
+    0x11CA3: 84,
+    0x11CA4: 84,
+    0x11CA5: 84,
+    0x11CA6: 84,
+    0x11CA7: 84,
+    0x11CAA: 84,
+    0x11CAB: 84,
+    0x11CAC: 84,
+    0x11CAD: 84,
+    0x11CAE: 84,
+    0x11CAF: 84,
+    0x11CB0: 84,
+    0x11CB2: 84,
+    0x11CB3: 84,
+    0x11CB5: 84,
+    0x11CB6: 84,
+    0x11D31: 84,
+    0x11D32: 84,
+    0x11D33: 84,
+    0x11D34: 84,
+    0x11D35: 84,
+    0x11D36: 84,
+    0x11D3A: 84,
+    0x11D3C: 84,
+    0x11D3D: 84,
+    0x11D3F: 84,
+    0x11D40: 84,
+    0x11D41: 84,
+    0x11D42: 84,
+    0x11D43: 84,
+    0x11D44: 84,
+    0x11D45: 84,
+    0x11D47: 84,
+    0x11D90: 84,
+    0x11D91: 84,
+    0x11D95: 84,
+    0x11D97: 84,
+    0x11EF3: 84,
+    0x11EF4: 84,
+    0x11F00: 84,
+    0x11F01: 84,
+    0x11F36: 84,
+    0x11F37: 84,
+    0x11F38: 84,
+    0x11F39: 84,
+    0x11F3A: 84,
+    0x11F40: 84,
+    0x11F42: 84,
+    0x13430: 84,
+    0x13431: 84,
+    0x13432: 84,
+    0x13433: 84,
+    0x13434: 84,
+    0x13435: 84,
+    0x13436: 84,
+    0x13437: 84,
+    0x13438: 84,
+    0x13439: 84,
+    0x1343A: 84,
+    0x1343B: 84,
+    0x1343C: 84,
+    0x1343D: 84,
+    0x1343E: 84,
+    0x1343F: 84,
+    0x13440: 84,
+    0x13447: 84,
+    0x13448: 84,
+    0x13449: 84,
+    0x1344A: 84,
+    0x1344B: 84,
+    0x1344C: 84,
+    0x1344D: 84,
+    0x1344E: 84,
+    0x1344F: 84,
+    0x13450: 84,
+    0x13451: 84,
+    0x13452: 84,
+    0x13453: 84,
+    0x13454: 84,
+    0x13455: 84,
+    0x16AF0: 84,
+    0x16AF1: 84,
+    0x16AF2: 84,
+    0x16AF3: 84,
+    0x16AF4: 84,
+    0x16B30: 84,
+    0x16B31: 84,
+    0x16B32: 84,
+    0x16B33: 84,
+    0x16B34: 84,
+    0x16B35: 84,
+    0x16B36: 84,
+    0x16F4F: 84,
+    0x16F8F: 84,
+    0x16F90: 84,
+    0x16F91: 84,
+    0x16F92: 84,
+    0x16FE4: 84,
+    0x1BC9D: 84,
+    0x1BC9E: 84,
+    0x1BCA0: 84,
+    0x1BCA1: 84,
+    0x1BCA2: 84,
+    0x1BCA3: 84,
+    0x1CF00: 84,
+    0x1CF01: 84,
+    0x1CF02: 84,
+    0x1CF03: 84,
+    0x1CF04: 84,
+    0x1CF05: 84,
+    0x1CF06: 84,
+    0x1CF07: 84,
+    0x1CF08: 84,
+    0x1CF09: 84,
+    0x1CF0A: 84,
+    0x1CF0B: 84,
+    0x1CF0C: 84,
+    0x1CF0D: 84,
+    0x1CF0E: 84,
+    0x1CF0F: 84,
+    0x1CF10: 84,
+    0x1CF11: 84,
+    0x1CF12: 84,
+    0x1CF13: 84,
+    0x1CF14: 84,
+    0x1CF15: 84,
+    0x1CF16: 84,
+    0x1CF17: 84,
+    0x1CF18: 84,
+    0x1CF19: 84,
+    0x1CF1A: 84,
+    0x1CF1B: 84,
+    0x1CF1C: 84,
+    0x1CF1D: 84,
+    0x1CF1E: 84,
+    0x1CF1F: 84,
+    0x1CF20: 84,
+    0x1CF21: 84,
+    0x1CF22: 84,
+    0x1CF23: 84,
+    0x1CF24: 84,
+    0x1CF25: 84,
+    0x1CF26: 84,
+    0x1CF27: 84,
+    0x1CF28: 84,
+    0x1CF29: 84,
+    0x1CF2A: 84,
+    0x1CF2B: 84,
+    0x1CF2C: 84,
+    0x1CF2D: 84,
+    0x1CF30: 84,
+    0x1CF31: 84,
+    0x1CF32: 84,
+    0x1CF33: 84,
+    0x1CF34: 84,
+    0x1CF35: 84,
+    0x1CF36: 84,
+    0x1CF37: 84,
+    0x1CF38: 84,
+    0x1CF39: 84,
+    0x1CF3A: 84,
+    0x1CF3B: 84,
+    0x1CF3C: 84,
+    0x1CF3D: 84,
+    0x1CF3E: 84,
+    0x1CF3F: 84,
+    0x1CF40: 84,
+    0x1CF41: 84,
+    0x1CF42: 84,
+    0x1CF43: 84,
+    0x1CF44: 84,
+    0x1CF45: 84,
+    0x1CF46: 84,
+    0x1D167: 84,
+    0x1D168: 84,
+    0x1D169: 84,
+    0x1D173: 84,
+    0x1D174: 84,
+    0x1D175: 84,
+    0x1D176: 84,
+    0x1D177: 84,
+    0x1D178: 84,
+    0x1D179: 84,
+    0x1D17A: 84,
+    0x1D17B: 84,
+    0x1D17C: 84,
+    0x1D17D: 84,
+    0x1D17E: 84,
+    0x1D17F: 84,
+    0x1D180: 84,
+    0x1D181: 84,
+    0x1D182: 84,
+    0x1D185: 84,
+    0x1D186: 84,
+    0x1D187: 84,
+    0x1D188: 84,
+    0x1D189: 84,
+    0x1D18A: 84,
+    0x1D18B: 84,
+    0x1D1AA: 84,
+    0x1D1AB: 84,
+    0x1D1AC: 84,
+    0x1D1AD: 84,
+    0x1D242: 84,
+    0x1D243: 84,
+    0x1D244: 84,
+    0x1DA00: 84,
+    0x1DA01: 84,
+    0x1DA02: 84,
+    0x1DA03: 84,
+    0x1DA04: 84,
+    0x1DA05: 84,
+    0x1DA06: 84,
+    0x1DA07: 84,
+    0x1DA08: 84,
+    0x1DA09: 84,
+    0x1DA0A: 84,
+    0x1DA0B: 84,
+    0x1DA0C: 84,
+    0x1DA0D: 84,
+    0x1DA0E: 84,
+    0x1DA0F: 84,
+    0x1DA10: 84,
+    0x1DA11: 84,
+    0x1DA12: 84,
+    0x1DA13: 84,
+    0x1DA14: 84,
+    0x1DA15: 84,
+    0x1DA16: 84,
+    0x1DA17: 84,
+    0x1DA18: 84,
+    0x1DA19: 84,
+    0x1DA1A: 84,
+    0x1DA1B: 84,
+    0x1DA1C: 84,
+    0x1DA1D: 84,
+    0x1DA1E: 84,
+    0x1DA1F: 84,
+    0x1DA20: 84,
+    0x1DA21: 84,
+    0x1DA22: 84,
+    0x1DA23: 84,
+    0x1DA24: 84,
+    0x1DA25: 84,
+    0x1DA26: 84,
+    0x1DA27: 84,
+    0x1DA28: 84,
+    0x1DA29: 84,
+    0x1DA2A: 84,
+    0x1DA2B: 84,
+    0x1DA2C: 84,
+    0x1DA2D: 84,
+    0x1DA2E: 84,
+    0x1DA2F: 84,
+    0x1DA30: 84,
+    0x1DA31: 84,
+    0x1DA32: 84,
+    0x1DA33: 84,
+    0x1DA34: 84,
+    0x1DA35: 84,
+    0x1DA36: 84,
+    0x1DA3B: 84,
+    0x1DA3C: 84,
+    0x1DA3D: 84,
+    0x1DA3E: 84,
+    0x1DA3F: 84,
+    0x1DA40: 84,
+    0x1DA41: 84,
+    0x1DA42: 84,
+    0x1DA43: 84,
+    0x1DA44: 84,
+    0x1DA45: 84,
+    0x1DA46: 84,
+    0x1DA47: 84,
+    0x1DA48: 84,
+    0x1DA49: 84,
+    0x1DA4A: 84,
+    0x1DA4B: 84,
+    0x1DA4C: 84,
+    0x1DA4D: 84,
+    0x1DA4E: 84,
+    0x1DA4F: 84,
+    0x1DA50: 84,
+    0x1DA51: 84,
+    0x1DA52: 84,
+    0x1DA53: 84,
+    0x1DA54: 84,
+    0x1DA55: 84,
+    0x1DA56: 84,
+    0x1DA57: 84,
+    0x1DA58: 84,
+    0x1DA59: 84,
+    0x1DA5A: 84,
+    0x1DA5B: 84,
+    0x1DA5C: 84,
+    0x1DA5D: 84,
+    0x1DA5E: 84,
+    0x1DA5F: 84,
+    0x1DA60: 84,
+    0x1DA61: 84,
+    0x1DA62: 84,
+    0x1DA63: 84,
+    0x1DA64: 84,
+    0x1DA65: 84,
+    0x1DA66: 84,
+    0x1DA67: 84,
+    0x1DA68: 84,
+    0x1DA69: 84,
+    0x1DA6A: 84,
+    0x1DA6B: 84,
+    0x1DA6C: 84,
+    0x1DA75: 84,
+    0x1DA84: 84,
+    0x1DA9B: 84,
+    0x1DA9C: 84,
+    0x1DA9D: 84,
+    0x1DA9E: 84,
+    0x1DA9F: 84,
+    0x1DAA1: 84,
+    0x1DAA2: 84,
+    0x1DAA3: 84,
+    0x1DAA4: 84,
+    0x1DAA5: 84,
+    0x1DAA6: 84,
+    0x1DAA7: 84,
+    0x1DAA8: 84,
+    0x1DAA9: 84,
+    0x1DAAA: 84,
+    0x1DAAB: 84,
+    0x1DAAC: 84,
+    0x1DAAD: 84,
+    0x1DAAE: 84,
+    0x1DAAF: 84,
+    0x1E000: 84,
+    0x1E001: 84,
+    0x1E002: 84,
+    0x1E003: 84,
+    0x1E004: 84,
+    0x1E005: 84,
+    0x1E006: 84,
+    0x1E008: 84,
+    0x1E009: 84,
+    0x1E00A: 84,
+    0x1E00B: 84,
+    0x1E00C: 84,
+    0x1E00D: 84,
+    0x1E00E: 84,
+    0x1E00F: 84,
+    0x1E010: 84,
+    0x1E011: 84,
+    0x1E012: 84,
+    0x1E013: 84,
+    0x1E014: 84,
+    0x1E015: 84,
+    0x1E016: 84,
+    0x1E017: 84,
+    0x1E018: 84,
+    0x1E01B: 84,
+    0x1E01C: 84,
+    0x1E01D: 84,
+    0x1E01E: 84,
+    0x1E01F: 84,
+    0x1E020: 84,
+    0x1E021: 84,
+    0x1E023: 84,
+    0x1E024: 84,
+    0x1E026: 84,
+    0x1E027: 84,
+    0x1E028: 84,
+    0x1E029: 84,
+    0x1E02A: 84,
+    0x1E08F: 84,
+    0x1E130: 84,
+    0x1E131: 84,
+    0x1E132: 84,
+    0x1E133: 84,
+    0x1E134: 84,
+    0x1E135: 84,
+    0x1E136: 84,
+    0x1E2AE: 84,
+    0x1E2EC: 84,
+    0x1E2ED: 84,
+    0x1E2EE: 84,
+    0x1E2EF: 84,
+    0x1E4EC: 84,
+    0x1E4ED: 84,
+    0x1E4EE: 84,
+    0x1E4EF: 84,
+    0x1E8D0: 84,
+    0x1E8D1: 84,
+    0x1E8D2: 84,
+    0x1E8D3: 84,
+    0x1E8D4: 84,
+    0x1E8D5: 84,
+    0x1E8D6: 84,
+    0x1E900: 68,
+    0x1E901: 68,
+    0x1E902: 68,
+    0x1E903: 68,
+    0x1E904: 68,
+    0x1E905: 68,
+    0x1E906: 68,
+    0x1E907: 68,
+    0x1E908: 68,
+    0x1E909: 68,
+    0x1E90A: 68,
+    0x1E90B: 68,
+    0x1E90C: 68,
+    0x1E90D: 68,
+    0x1E90E: 68,
+    0x1E90F: 68,
+    0x1E910: 68,
+    0x1E911: 68,
+    0x1E912: 68,
+    0x1E913: 68,
+    0x1E914: 68,
+    0x1E915: 68,
+    0x1E916: 68,
+    0x1E917: 68,
+    0x1E918: 68,
+    0x1E919: 68,
+    0x1E91A: 68,
+    0x1E91B: 68,
+    0x1E91C: 68,
+    0x1E91D: 68,
+    0x1E91E: 68,
+    0x1E91F: 68,
+    0x1E920: 68,
+    0x1E921: 68,
+    0x1E922: 68,
+    0x1E923: 68,
+    0x1E924: 68,
+    0x1E925: 68,
+    0x1E926: 68,
+    0x1E927: 68,
+    0x1E928: 68,
+    0x1E929: 68,
+    0x1E92A: 68,
+    0x1E92B: 68,
+    0x1E92C: 68,
+    0x1E92D: 68,
+    0x1E92E: 68,
+    0x1E92F: 68,
+    0x1E930: 68,
+    0x1E931: 68,
+    0x1E932: 68,
+    0x1E933: 68,
+    0x1E934: 68,
+    0x1E935: 68,
+    0x1E936: 68,
+    0x1E937: 68,
+    0x1E938: 68,
+    0x1E939: 68,
+    0x1E93A: 68,
+    0x1E93B: 68,
+    0x1E93C: 68,
+    0x1E93D: 68,
+    0x1E93E: 68,
+    0x1E93F: 68,
+    0x1E940: 68,
+    0x1E941: 68,
+    0x1E942: 68,
+    0x1E943: 68,
+    0x1E944: 84,
+    0x1E945: 84,
+    0x1E946: 84,
+    0x1E947: 84,
+    0x1E948: 84,
+    0x1E949: 84,
+    0x1E94A: 84,
+    0x1E94B: 84,
+    0xE0001: 84,
+    0xE0020: 84,
+    0xE0021: 84,
+    0xE0022: 84,
+    0xE0023: 84,
+    0xE0024: 84,
+    0xE0025: 84,
+    0xE0026: 84,
+    0xE0027: 84,
+    0xE0028: 84,
+    0xE0029: 84,
+    0xE002A: 84,
+    0xE002B: 84,
+    0xE002C: 84,
+    0xE002D: 84,
+    0xE002E: 84,
+    0xE002F: 84,
+    0xE0030: 84,
+    0xE0031: 84,
+    0xE0032: 84,
+    0xE0033: 84,
+    0xE0034: 84,
+    0xE0035: 84,
+    0xE0036: 84,
+    0xE0037: 84,
+    0xE0038: 84,
+    0xE0039: 84,
+    0xE003A: 84,
+    0xE003B: 84,
+    0xE003C: 84,
+    0xE003D: 84,
+    0xE003E: 84,
+    0xE003F: 84,
+    0xE0040: 84,
+    0xE0041: 84,
+    0xE0042: 84,
+    0xE0043: 84,
+    0xE0044: 84,
+    0xE0045: 84,
+    0xE0046: 84,
+    0xE0047: 84,
+    0xE0048: 84,
+    0xE0049: 84,
+    0xE004A: 84,
+    0xE004B: 84,
+    0xE004C: 84,
+    0xE004D: 84,
+    0xE004E: 84,
+    0xE004F: 84,
+    0xE0050: 84,
+    0xE0051: 84,
+    0xE0052: 84,
+    0xE0053: 84,
+    0xE0054: 84,
+    0xE0055: 84,
+    0xE0056: 84,
+    0xE0057: 84,
+    0xE0058: 84,
+    0xE0059: 84,
+    0xE005A: 84,
+    0xE005B: 84,
+    0xE005C: 84,
+    0xE005D: 84,
+    0xE005E: 84,
+    0xE005F: 84,
+    0xE0060: 84,
+    0xE0061: 84,
+    0xE0062: 84,
+    0xE0063: 84,
+    0xE0064: 84,
+    0xE0065: 84,
+    0xE0066: 84,
+    0xE0067: 84,
+    0xE0068: 84,
+    0xE0069: 84,
+    0xE006A: 84,
+    0xE006B: 84,
+    0xE006C: 84,
+    0xE006D: 84,
+    0xE006E: 84,
+    0xE006F: 84,
+    0xE0070: 84,
+    0xE0071: 84,
+    0xE0072: 84,
+    0xE0073: 84,
+    0xE0074: 84,
+    0xE0075: 84,
+    0xE0076: 84,
+    0xE0077: 84,
+    0xE0078: 84,
+    0xE0079: 84,
+    0xE007A: 84,
+    0xE007B: 84,
+    0xE007C: 84,
+    0xE007D: 84,
+    0xE007E: 84,
+    0xE007F: 84,
+    0xE0100: 84,
+    0xE0101: 84,
+    0xE0102: 84,
+    0xE0103: 84,
+    0xE0104: 84,
+    0xE0105: 84,
+    0xE0106: 84,
+    0xE0107: 84,
+    0xE0108: 84,
+    0xE0109: 84,
+    0xE010A: 84,
+    0xE010B: 84,
+    0xE010C: 84,
+    0xE010D: 84,
+    0xE010E: 84,
+    0xE010F: 84,
+    0xE0110: 84,
+    0xE0111: 84,
+    0xE0112: 84,
+    0xE0113: 84,
+    0xE0114: 84,
+    0xE0115: 84,
+    0xE0116: 84,
+    0xE0117: 84,
+    0xE0118: 84,
+    0xE0119: 84,
+    0xE011A: 84,
+    0xE011B: 84,
+    0xE011C: 84,
+    0xE011D: 84,
+    0xE011E: 84,
+    0xE011F: 84,
+    0xE0120: 84,
+    0xE0121: 84,
+    0xE0122: 84,
+    0xE0123: 84,
+    0xE0124: 84,
+    0xE0125: 84,
+    0xE0126: 84,
+    0xE0127: 84,
+    0xE0128: 84,
+    0xE0129: 84,
+    0xE012A: 84,
+    0xE012B: 84,
+    0xE012C: 84,
+    0xE012D: 84,
+    0xE012E: 84,
+    0xE012F: 84,
+    0xE0130: 84,
+    0xE0131: 84,
+    0xE0132: 84,
+    0xE0133: 84,
+    0xE0134: 84,
+    0xE0135: 84,
+    0xE0136: 84,
+    0xE0137: 84,
+    0xE0138: 84,
+    0xE0139: 84,
+    0xE013A: 84,
+    0xE013B: 84,
+    0xE013C: 84,
+    0xE013D: 84,
+    0xE013E: 84,
+    0xE013F: 84,
+    0xE0140: 84,
+    0xE0141: 84,
+    0xE0142: 84,
+    0xE0143: 84,
+    0xE0144: 84,
+    0xE0145: 84,
+    0xE0146: 84,
+    0xE0147: 84,
+    0xE0148: 84,
+    0xE0149: 84,
+    0xE014A: 84,
+    0xE014B: 84,
+    0xE014C: 84,
+    0xE014D: 84,
+    0xE014E: 84,
+    0xE014F: 84,
+    0xE0150: 84,
+    0xE0151: 84,
+    0xE0152: 84,
+    0xE0153: 84,
+    0xE0154: 84,
+    0xE0155: 84,
+    0xE0156: 84,
+    0xE0157: 84,
+    0xE0158: 84,
+    0xE0159: 84,
+    0xE015A: 84,
+    0xE015B: 84,
+    0xE015C: 84,
+    0xE015D: 84,
+    0xE015E: 84,
+    0xE015F: 84,
+    0xE0160: 84,
+    0xE0161: 84,
+    0xE0162: 84,
+    0xE0163: 84,
+    0xE0164: 84,
+    0xE0165: 84,
+    0xE0166: 84,
+    0xE0167: 84,
+    0xE0168: 84,
+    0xE0169: 84,
+    0xE016A: 84,
+    0xE016B: 84,
+    0xE016C: 84,
+    0xE016D: 84,
+    0xE016E: 84,
+    0xE016F: 84,
+    0xE0170: 84,
+    0xE0171: 84,
+    0xE0172: 84,
+    0xE0173: 84,
+    0xE0174: 84,
+    0xE0175: 84,
+    0xE0176: 84,
+    0xE0177: 84,
+    0xE0178: 84,
+    0xE0179: 84,
+    0xE017A: 84,
+    0xE017B: 84,
+    0xE017C: 84,
+    0xE017D: 84,
+    0xE017E: 84,
+    0xE017F: 84,
+    0xE0180: 84,
+    0xE0181: 84,
+    0xE0182: 84,
+    0xE0183: 84,
+    0xE0184: 84,
+    0xE0185: 84,
+    0xE0186: 84,
+    0xE0187: 84,
+    0xE0188: 84,
+    0xE0189: 84,
+    0xE018A: 84,
+    0xE018B: 84,
+    0xE018C: 84,
+    0xE018D: 84,
+    0xE018E: 84,
+    0xE018F: 84,
+    0xE0190: 84,
+    0xE0191: 84,
+    0xE0192: 84,
+    0xE0193: 84,
+    0xE0194: 84,
+    0xE0195: 84,
+    0xE0196: 84,
+    0xE0197: 84,
+    0xE0198: 84,
+    0xE0199: 84,
+    0xE019A: 84,
+    0xE019B: 84,
+    0xE019C: 84,
+    0xE019D: 84,
+    0xE019E: 84,
+    0xE019F: 84,
+    0xE01A0: 84,
+    0xE01A1: 84,
+    0xE01A2: 84,
+    0xE01A3: 84,
+    0xE01A4: 84,
+    0xE01A5: 84,
+    0xE01A6: 84,
+    0xE01A7: 84,
+    0xE01A8: 84,
+    0xE01A9: 84,
+    0xE01AA: 84,
+    0xE01AB: 84,
+    0xE01AC: 84,
+    0xE01AD: 84,
+    0xE01AE: 84,
+    0xE01AF: 84,
+    0xE01B0: 84,
+    0xE01B1: 84,
+    0xE01B2: 84,
+    0xE01B3: 84,
+    0xE01B4: 84,
+    0xE01B5: 84,
+    0xE01B6: 84,
+    0xE01B7: 84,
+    0xE01B8: 84,
+    0xE01B9: 84,
+    0xE01BA: 84,
+    0xE01BB: 84,
+    0xE01BC: 84,
+    0xE01BD: 84,
+    0xE01BE: 84,
+    0xE01BF: 84,
+    0xE01C0: 84,
+    0xE01C1: 84,
+    0xE01C2: 84,
+    0xE01C3: 84,
+    0xE01C4: 84,
+    0xE01C5: 84,
+    0xE01C6: 84,
+    0xE01C7: 84,
+    0xE01C8: 84,
+    0xE01C9: 84,
+    0xE01CA: 84,
+    0xE01CB: 84,
+    0xE01CC: 84,
+    0xE01CD: 84,
+    0xE01CE: 84,
+    0xE01CF: 84,
+    0xE01D0: 84,
+    0xE01D1: 84,
+    0xE01D2: 84,
+    0xE01D3: 84,
+    0xE01D4: 84,
+    0xE01D5: 84,
+    0xE01D6: 84,
+    0xE01D7: 84,
+    0xE01D8: 84,
+    0xE01D9: 84,
+    0xE01DA: 84,
+    0xE01DB: 84,
+    0xE01DC: 84,
+    0xE01DD: 84,
+    0xE01DE: 84,
+    0xE01DF: 84,
+    0xE01E0: 84,
+    0xE01E1: 84,
+    0xE01E2: 84,
+    0xE01E3: 84,
+    0xE01E4: 84,
+    0xE01E5: 84,
+    0xE01E6: 84,
+    0xE01E7: 84,
+    0xE01E8: 84,
+    0xE01E9: 84,
+    0xE01EA: 84,
+    0xE01EB: 84,
+    0xE01EC: 84,
+    0xE01ED: 84,
+    0xE01EE: 84,
+    0xE01EF: 84,
+}
+codepoint_classes = {
+    "PVALID": (
+        0x2D0000002E,
+        0x300000003A,
+        0x610000007B,
+        0xDF000000F7,
+        0xF800000100,
+        0x10100000102,
+        0x10300000104,
+        0x10500000106,
+        0x10700000108,
+        0x1090000010A,
+        0x10B0000010C,
+        0x10D0000010E,
+        0x10F00000110,
+        0x11100000112,
+        0x11300000114,
+        0x11500000116,
+        0x11700000118,
+        0x1190000011A,
+        0x11B0000011C,
+        0x11D0000011E,
+        0x11F00000120,
+        0x12100000122,
+        0x12300000124,
+        0x12500000126,
+        0x12700000128,
+        0x1290000012A,
+        0x12B0000012C,
+        0x12D0000012E,
+        0x12F00000130,
+        0x13100000132,
+        0x13500000136,
+        0x13700000139,
+        0x13A0000013B,
+        0x13C0000013D,
+        0x13E0000013F,
+        0x14200000143,
+        0x14400000145,
+        0x14600000147,
+        0x14800000149,
+        0x14B0000014C,
+        0x14D0000014E,
+        0x14F00000150,
+        0x15100000152,
+        0x15300000154,
+        0x15500000156,
+        0x15700000158,
+        0x1590000015A,
+        0x15B0000015C,
+        0x15D0000015E,
+        0x15F00000160,
+        0x16100000162,
+        0x16300000164,
+        0x16500000166,
+        0x16700000168,
+        0x1690000016A,
+        0x16B0000016C,
+        0x16D0000016E,
+        0x16F00000170,
+        0x17100000172,
+        0x17300000174,
+        0x17500000176,
+        0x17700000178,
+        0x17A0000017B,
+        0x17C0000017D,
+        0x17E0000017F,
+        0x18000000181,
+        0x18300000184,
+        0x18500000186,
+        0x18800000189,
+        0x18C0000018E,
+        0x19200000193,
+        0x19500000196,
+        0x1990000019C,
+        0x19E0000019F,
+        0x1A1000001A2,
+        0x1A3000001A4,
+        0x1A5000001A6,
+        0x1A8000001A9,
+        0x1AA000001AC,
+        0x1AD000001AE,
+        0x1B0000001B1,
+        0x1B4000001B5,
+        0x1B6000001B7,
+        0x1B9000001BC,
+        0x1BD000001C4,
+        0x1CE000001CF,
+        0x1D0000001D1,
+        0x1D2000001D3,
+        0x1D4000001D5,
+        0x1D6000001D7,
+        0x1D8000001D9,
+        0x1DA000001DB,
+        0x1DC000001DE,
+        0x1DF000001E0,
+        0x1E1000001E2,
+        0x1E3000001E4,
+        0x1E5000001E6,
+        0x1E7000001E8,
+        0x1E9000001EA,
+        0x1EB000001EC,
+        0x1ED000001EE,
+        0x1EF000001F1,
+        0x1F5000001F6,
+        0x1F9000001FA,
+        0x1FB000001FC,
+        0x1FD000001FE,
+        0x1FF00000200,
+        0x20100000202,
+        0x20300000204,
+        0x20500000206,
+        0x20700000208,
+        0x2090000020A,
+        0x20B0000020C,
+        0x20D0000020E,
+        0x20F00000210,
+        0x21100000212,
+        0x21300000214,
+        0x21500000216,
+        0x21700000218,
+        0x2190000021A,
+        0x21B0000021C,
+        0x21D0000021E,
+        0x21F00000220,
+        0x22100000222,
+        0x22300000224,
+        0x22500000226,
+        0x22700000228,
+        0x2290000022A,
+        0x22B0000022C,
+        0x22D0000022E,
+        0x22F00000230,
+        0x23100000232,
+        0x2330000023A,
+        0x23C0000023D,
+        0x23F00000241,
+        0x24200000243,
+        0x24700000248,
+        0x2490000024A,
+        0x24B0000024C,
+        0x24D0000024E,
+        0x24F000002B0,
+        0x2B9000002C2,
+        0x2C6000002D2,
+        0x2EC000002ED,
+        0x2EE000002EF,
+        0x30000000340,
+        0x34200000343,
+        0x3460000034F,
+        0x35000000370,
+        0x37100000372,
+        0x37300000374,
+        0x37700000378,
+        0x37B0000037E,
+        0x39000000391,
+        0x3AC000003CF,
+        0x3D7000003D8,
+        0x3D9000003DA,
+        0x3DB000003DC,
+        0x3DD000003DE,
+        0x3DF000003E0,
+        0x3E1000003E2,
+        0x3E3000003E4,
+        0x3E5000003E6,
+        0x3E7000003E8,
+        0x3E9000003EA,
+        0x3EB000003EC,
+        0x3ED000003EE,
+        0x3EF000003F0,
+        0x3F3000003F4,
+        0x3F8000003F9,
+        0x3FB000003FD,
+        0x43000000460,
+        0x46100000462,
+        0x46300000464,
+        0x46500000466,
+        0x46700000468,
+        0x4690000046A,
+        0x46B0000046C,
+        0x46D0000046E,
+        0x46F00000470,
+        0x47100000472,
+        0x47300000474,
+        0x47500000476,
+        0x47700000478,
+        0x4790000047A,
+        0x47B0000047C,
+        0x47D0000047E,
+        0x47F00000480,
+        0x48100000482,
+        0x48300000488,
+        0x48B0000048C,
+        0x48D0000048E,
+        0x48F00000490,
+        0x49100000492,
+        0x49300000494,
+        0x49500000496,
+        0x49700000498,
+        0x4990000049A,
+        0x49B0000049C,
+        0x49D0000049E,
+        0x49F000004A0,
+        0x4A1000004A2,
+        0x4A3000004A4,
+        0x4A5000004A6,
+        0x4A7000004A8,
+        0x4A9000004AA,
+        0x4AB000004AC,
+        0x4AD000004AE,
+        0x4AF000004B0,
+        0x4B1000004B2,
+        0x4B3000004B4,
+        0x4B5000004B6,
+        0x4B7000004B8,
+        0x4B9000004BA,
+        0x4BB000004BC,
+        0x4BD000004BE,
+        0x4BF000004C0,
+        0x4C2000004C3,
+        0x4C4000004C5,
+        0x4C6000004C7,
+        0x4C8000004C9,
+        0x4CA000004CB,
+        0x4CC000004CD,
+        0x4CE000004D0,
+        0x4D1000004D2,
+        0x4D3000004D4,
+        0x4D5000004D6,
+        0x4D7000004D8,
+        0x4D9000004DA,
+        0x4DB000004DC,
+        0x4DD000004DE,
+        0x4DF000004E0,
+        0x4E1000004E2,
+        0x4E3000004E4,
+        0x4E5000004E6,
+        0x4E7000004E8,
+        0x4E9000004EA,
+        0x4EB000004EC,
+        0x4ED000004EE,
+        0x4EF000004F0,
+        0x4F1000004F2,
+        0x4F3000004F4,
+        0x4F5000004F6,
+        0x4F7000004F8,
+        0x4F9000004FA,
+        0x4FB000004FC,
+        0x4FD000004FE,
+        0x4FF00000500,
+        0x50100000502,
+        0x50300000504,
+        0x50500000506,
+        0x50700000508,
+        0x5090000050A,
+        0x50B0000050C,
+        0x50D0000050E,
+        0x50F00000510,
+        0x51100000512,
+        0x51300000514,
+        0x51500000516,
+        0x51700000518,
+        0x5190000051A,
+        0x51B0000051C,
+        0x51D0000051E,
+        0x51F00000520,
+        0x52100000522,
+        0x52300000524,
+        0x52500000526,
+        0x52700000528,
+        0x5290000052A,
+        0x52B0000052C,
+        0x52D0000052E,
+        0x52F00000530,
+        0x5590000055A,
+        0x56000000587,
+        0x58800000589,
+        0x591000005BE,
+        0x5BF000005C0,
+        0x5C1000005C3,
+        0x5C4000005C6,
+        0x5C7000005C8,
+        0x5D0000005EB,
+        0x5EF000005F3,
+        0x6100000061B,
+        0x62000000640,
+        0x64100000660,
+        0x66E00000675,
+        0x679000006D4,
+        0x6D5000006DD,
+        0x6DF000006E9,
+        0x6EA000006F0,
+        0x6FA00000700,
+        0x7100000074B,
+        0x74D000007B2,
+        0x7C0000007F6,
+        0x7FD000007FE,
+        0x8000000082E,
+        0x8400000085C,
+        0x8600000086B,
+        0x87000000888,
+        0x8890000088F,
+        0x898000008E2,
+        0x8E300000958,
+        0x96000000964,
+        0x96600000970,
+        0x97100000984,
+        0x9850000098D,
+        0x98F00000991,
+        0x993000009A9,
+        0x9AA000009B1,
+        0x9B2000009B3,
+        0x9B6000009BA,
+        0x9BC000009C5,
+        0x9C7000009C9,
+        0x9CB000009CF,
+        0x9D7000009D8,
+        0x9E0000009E4,
+        0x9E6000009F2,
+        0x9FC000009FD,
+        0x9FE000009FF,
+        0xA0100000A04,
+        0xA0500000A0B,
+        0xA0F00000A11,
+        0xA1300000A29,
+        0xA2A00000A31,
+        0xA3200000A33,
+        0xA3500000A36,
+        0xA3800000A3A,
+        0xA3C00000A3D,
+        0xA3E00000A43,
+        0xA4700000A49,
+        0xA4B00000A4E,
+        0xA5100000A52,
+        0xA5C00000A5D,
+        0xA6600000A76,
+        0xA8100000A84,
+        0xA8500000A8E,
+        0xA8F00000A92,
+        0xA9300000AA9,
+        0xAAA00000AB1,
+        0xAB200000AB4,
+        0xAB500000ABA,
+        0xABC00000AC6,
+        0xAC700000ACA,
+        0xACB00000ACE,
+        0xAD000000AD1,
+        0xAE000000AE4,
+        0xAE600000AF0,
+        0xAF900000B00,
+        0xB0100000B04,
+        0xB0500000B0D,
+        0xB0F00000B11,
+        0xB1300000B29,
+        0xB2A00000B31,
+        0xB3200000B34,
+        0xB3500000B3A,
+        0xB3C00000B45,
+        0xB4700000B49,
+        0xB4B00000B4E,
+        0xB5500000B58,
+        0xB5F00000B64,
+        0xB6600000B70,
+        0xB7100000B72,
+        0xB8200000B84,
+        0xB8500000B8B,
+        0xB8E00000B91,
+        0xB9200000B96,
+        0xB9900000B9B,
+        0xB9C00000B9D,
+        0xB9E00000BA0,
+        0xBA300000BA5,
+        0xBA800000BAB,
+        0xBAE00000BBA,
+        0xBBE00000BC3,
+        0xBC600000BC9,
+        0xBCA00000BCE,
+        0xBD000000BD1,
+        0xBD700000BD8,
+        0xBE600000BF0,
+        0xC0000000C0D,
+        0xC0E00000C11,
+        0xC1200000C29,
+        0xC2A00000C3A,
+        0xC3C00000C45,
+        0xC4600000C49,
+        0xC4A00000C4E,
+        0xC5500000C57,
+        0xC5800000C5B,
+        0xC5D00000C5E,
+        0xC6000000C64,
+        0xC6600000C70,
+        0xC8000000C84,
+        0xC8500000C8D,
+        0xC8E00000C91,
+        0xC9200000CA9,
+        0xCAA00000CB4,
+        0xCB500000CBA,
+        0xCBC00000CC5,
+        0xCC600000CC9,
+        0xCCA00000CCE,
+        0xCD500000CD7,
+        0xCDD00000CDF,
+        0xCE000000CE4,
+        0xCE600000CF0,
+        0xCF100000CF4,
+        0xD0000000D0D,
+        0xD0E00000D11,
+        0xD1200000D45,
+        0xD4600000D49,
+        0xD4A00000D4F,
+        0xD5400000D58,
+        0xD5F00000D64,
+        0xD6600000D70,
+        0xD7A00000D80,
+        0xD8100000D84,
+        0xD8500000D97,
+        0xD9A00000DB2,
+        0xDB300000DBC,
+        0xDBD00000DBE,
+        0xDC000000DC7,
+        0xDCA00000DCB,
+        0xDCF00000DD5,
+        0xDD600000DD7,
+        0xDD800000DE0,
+        0xDE600000DF0,
+        0xDF200000DF4,
+        0xE0100000E33,
+        0xE3400000E3B,
+        0xE4000000E4F,
+        0xE5000000E5A,
+        0xE8100000E83,
+        0xE8400000E85,
+        0xE8600000E8B,
+        0xE8C00000EA4,
+        0xEA500000EA6,
+        0xEA700000EB3,
+        0xEB400000EBE,
+        0xEC000000EC5,
+        0xEC600000EC7,
+        0xEC800000ECF,
+        0xED000000EDA,
+        0xEDE00000EE0,
+        0xF0000000F01,
+        0xF0B00000F0C,
+        0xF1800000F1A,
+        0xF2000000F2A,
+        0xF3500000F36,
+        0xF3700000F38,
+        0xF3900000F3A,
+        0xF3E00000F43,
+        0xF4400000F48,
+        0xF4900000F4D,
+        0xF4E00000F52,
+        0xF5300000F57,
+        0xF5800000F5C,
+        0xF5D00000F69,
+        0xF6A00000F6D,
+        0xF7100000F73,
+        0xF7400000F75,
+        0xF7A00000F81,
+        0xF8200000F85,
+        0xF8600000F93,
+        0xF9400000F98,
+        0xF9900000F9D,
+        0xF9E00000FA2,
+        0xFA300000FA7,
+        0xFA800000FAC,
+        0xFAD00000FB9,
+        0xFBA00000FBD,
+        0xFC600000FC7,
+        0x10000000104A,
+        0x10500000109E,
+        0x10D0000010FB,
+        0x10FD00001100,
+        0x120000001249,
+        0x124A0000124E,
+        0x125000001257,
+        0x125800001259,
+        0x125A0000125E,
+        0x126000001289,
+        0x128A0000128E,
+        0x1290000012B1,
+        0x12B2000012B6,
+        0x12B8000012BF,
+        0x12C0000012C1,
+        0x12C2000012C6,
+        0x12C8000012D7,
+        0x12D800001311,
+        0x131200001316,
+        0x13180000135B,
+        0x135D00001360,
+        0x138000001390,
+        0x13A0000013F6,
+        0x14010000166D,
+        0x166F00001680,
+        0x16810000169B,
+        0x16A0000016EB,
+        0x16F1000016F9,
+        0x170000001716,
+        0x171F00001735,
+        0x174000001754,
+        0x17600000176D,
+        0x176E00001771,
+        0x177200001774,
+        0x1780000017B4,
+        0x17B6000017D4,
+        0x17D7000017D8,
+        0x17DC000017DE,
+        0x17E0000017EA,
+        0x18100000181A,
+        0x182000001879,
+        0x1880000018AB,
+        0x18B0000018F6,
+        0x19000000191F,
+        0x19200000192C,
+        0x19300000193C,
+        0x19460000196E,
+        0x197000001975,
+        0x1980000019AC,
+        0x19B0000019CA,
+        0x19D0000019DA,
+        0x1A0000001A1C,
+        0x1A2000001A5F,
+        0x1A6000001A7D,
+        0x1A7F00001A8A,
+        0x1A9000001A9A,
+        0x1AA700001AA8,
+        0x1AB000001ABE,
+        0x1ABF00001ACF,
+        0x1B0000001B4D,
+        0x1B5000001B5A,
+        0x1B6B00001B74,
+        0x1B8000001BF4,
+        0x1C0000001C38,
+        0x1C4000001C4A,
+        0x1C4D00001C7E,
+        0x1CD000001CD3,
+        0x1CD400001CFB,
+        0x1D0000001D2C,
+        0x1D2F00001D30,
+        0x1D3B00001D3C,
+        0x1D4E00001D4F,
+        0x1D6B00001D78,
+        0x1D7900001D9B,
+        0x1DC000001E00,
+        0x1E0100001E02,
+        0x1E0300001E04,
+        0x1E0500001E06,
+        0x1E0700001E08,
+        0x1E0900001E0A,
+        0x1E0B00001E0C,
+        0x1E0D00001E0E,
+        0x1E0F00001E10,
+        0x1E1100001E12,
+        0x1E1300001E14,
+        0x1E1500001E16,
+        0x1E1700001E18,
+        0x1E1900001E1A,
+        0x1E1B00001E1C,
+        0x1E1D00001E1E,
+        0x1E1F00001E20,
+        0x1E2100001E22,
+        0x1E2300001E24,
+        0x1E2500001E26,
+        0x1E2700001E28,
+        0x1E2900001E2A,
+        0x1E2B00001E2C,
+        0x1E2D00001E2E,
+        0x1E2F00001E30,
+        0x1E3100001E32,
+        0x1E3300001E34,
+        0x1E3500001E36,
+        0x1E3700001E38,
+        0x1E3900001E3A,
+        0x1E3B00001E3C,
+        0x1E3D00001E3E,
+        0x1E3F00001E40,
+        0x1E4100001E42,
+        0x1E4300001E44,
+        0x1E4500001E46,
+        0x1E4700001E48,
+        0x1E4900001E4A,
+        0x1E4B00001E4C,
+        0x1E4D00001E4E,
+        0x1E4F00001E50,
+        0x1E5100001E52,
+        0x1E5300001E54,
+        0x1E5500001E56,
+        0x1E5700001E58,
+        0x1E5900001E5A,
+        0x1E5B00001E5C,
+        0x1E5D00001E5E,
+        0x1E5F00001E60,
+        0x1E6100001E62,
+        0x1E6300001E64,
+        0x1E6500001E66,
+        0x1E6700001E68,
+        0x1E6900001E6A,
+        0x1E6B00001E6C,
+        0x1E6D00001E6E,
+        0x1E6F00001E70,
+        0x1E7100001E72,
+        0x1E7300001E74,
+        0x1E7500001E76,
+        0x1E7700001E78,
+        0x1E7900001E7A,
+        0x1E7B00001E7C,
+        0x1E7D00001E7E,
+        0x1E7F00001E80,
+        0x1E8100001E82,
+        0x1E8300001E84,
+        0x1E8500001E86,
+        0x1E8700001E88,
+        0x1E8900001E8A,
+        0x1E8B00001E8C,
+        0x1E8D00001E8E,
+        0x1E8F00001E90,
+        0x1E9100001E92,
+        0x1E9300001E94,
+        0x1E9500001E9A,
+        0x1E9C00001E9E,
+        0x1E9F00001EA0,
+        0x1EA100001EA2,
+        0x1EA300001EA4,
+        0x1EA500001EA6,
+        0x1EA700001EA8,
+        0x1EA900001EAA,
+        0x1EAB00001EAC,
+        0x1EAD00001EAE,
+        0x1EAF00001EB0,
+        0x1EB100001EB2,
+        0x1EB300001EB4,
+        0x1EB500001EB6,
+        0x1EB700001EB8,
+        0x1EB900001EBA,
+        0x1EBB00001EBC,
+        0x1EBD00001EBE,
+        0x1EBF00001EC0,
+        0x1EC100001EC2,
+        0x1EC300001EC4,
+        0x1EC500001EC6,
+        0x1EC700001EC8,
+        0x1EC900001ECA,
+        0x1ECB00001ECC,
+        0x1ECD00001ECE,
+        0x1ECF00001ED0,
+        0x1ED100001ED2,
+        0x1ED300001ED4,
+        0x1ED500001ED6,
+        0x1ED700001ED8,
+        0x1ED900001EDA,
+        0x1EDB00001EDC,
+        0x1EDD00001EDE,
+        0x1EDF00001EE0,
+        0x1EE100001EE2,
+        0x1EE300001EE4,
+        0x1EE500001EE6,
+        0x1EE700001EE8,
+        0x1EE900001EEA,
+        0x1EEB00001EEC,
+        0x1EED00001EEE,
+        0x1EEF00001EF0,
+        0x1EF100001EF2,
+        0x1EF300001EF4,
+        0x1EF500001EF6,
+        0x1EF700001EF8,
+        0x1EF900001EFA,
+        0x1EFB00001EFC,
+        0x1EFD00001EFE,
+        0x1EFF00001F08,
+        0x1F1000001F16,
+        0x1F2000001F28,
+        0x1F3000001F38,
+        0x1F4000001F46,
+        0x1F5000001F58,
+        0x1F6000001F68,
+        0x1F7000001F71,
+        0x1F7200001F73,
+        0x1F7400001F75,
+        0x1F7600001F77,
+        0x1F7800001F79,
+        0x1F7A00001F7B,
+        0x1F7C00001F7D,
+        0x1FB000001FB2,
+        0x1FB600001FB7,
+        0x1FC600001FC7,
+        0x1FD000001FD3,
+        0x1FD600001FD8,
+        0x1FE000001FE3,
+        0x1FE400001FE8,
+        0x1FF600001FF7,
+        0x214E0000214F,
+        0x218400002185,
+        0x2C3000002C60,
+        0x2C6100002C62,
+        0x2C6500002C67,
+        0x2C6800002C69,
+        0x2C6A00002C6B,
+        0x2C6C00002C6D,
+        0x2C7100002C72,
+        0x2C7300002C75,
+        0x2C7600002C7C,
+        0x2C8100002C82,
+        0x2C8300002C84,
+        0x2C8500002C86,
+        0x2C8700002C88,
+        0x2C8900002C8A,
+        0x2C8B00002C8C,
+        0x2C8D00002C8E,
+        0x2C8F00002C90,
+        0x2C9100002C92,
+        0x2C9300002C94,
+        0x2C9500002C96,
+        0x2C9700002C98,
+        0x2C9900002C9A,
+        0x2C9B00002C9C,
+        0x2C9D00002C9E,
+        0x2C9F00002CA0,
+        0x2CA100002CA2,
+        0x2CA300002CA4,
+        0x2CA500002CA6,
+        0x2CA700002CA8,
+        0x2CA900002CAA,
+        0x2CAB00002CAC,
+        0x2CAD00002CAE,
+        0x2CAF00002CB0,
+        0x2CB100002CB2,
+        0x2CB300002CB4,
+        0x2CB500002CB6,
+        0x2CB700002CB8,
+        0x2CB900002CBA,
+        0x2CBB00002CBC,
+        0x2CBD00002CBE,
+        0x2CBF00002CC0,
+        0x2CC100002CC2,
+        0x2CC300002CC4,
+        0x2CC500002CC6,
+        0x2CC700002CC8,
+        0x2CC900002CCA,
+        0x2CCB00002CCC,
+        0x2CCD00002CCE,
+        0x2CCF00002CD0,
+        0x2CD100002CD2,
+        0x2CD300002CD4,
+        0x2CD500002CD6,
+        0x2CD700002CD8,
+        0x2CD900002CDA,
+        0x2CDB00002CDC,
+        0x2CDD00002CDE,
+        0x2CDF00002CE0,
+        0x2CE100002CE2,
+        0x2CE300002CE5,
+        0x2CEC00002CED,
+        0x2CEE00002CF2,
+        0x2CF300002CF4,
+        0x2D0000002D26,
+        0x2D2700002D28,
+        0x2D2D00002D2E,
+        0x2D3000002D68,
+        0x2D7F00002D97,
+        0x2DA000002DA7,
+        0x2DA800002DAF,
+        0x2DB000002DB7,
+        0x2DB800002DBF,
+        0x2DC000002DC7,
+        0x2DC800002DCF,
+        0x2DD000002DD7,
+        0x2DD800002DDF,
+        0x2DE000002E00,
+        0x2E2F00002E30,
+        0x300500003008,
+        0x302A0000302E,
+        0x303C0000303D,
+        0x304100003097,
+        0x30990000309B,
+        0x309D0000309F,
+        0x30A1000030FB,
+        0x30FC000030FF,
+        0x310500003130,
+        0x31A0000031C0,
+        0x31F000003200,
+        0x340000004DC0,
+        0x4E000000A48D,
+        0xA4D00000A4FE,
+        0xA5000000A60D,
+        0xA6100000A62C,
+        0xA6410000A642,
+        0xA6430000A644,
+        0xA6450000A646,
+        0xA6470000A648,
+        0xA6490000A64A,
+        0xA64B0000A64C,
+        0xA64D0000A64E,
+        0xA64F0000A650,
+        0xA6510000A652,
+        0xA6530000A654,
+        0xA6550000A656,
+        0xA6570000A658,
+        0xA6590000A65A,
+        0xA65B0000A65C,
+        0xA65D0000A65E,
+        0xA65F0000A660,
+        0xA6610000A662,
+        0xA6630000A664,
+        0xA6650000A666,
+        0xA6670000A668,
+        0xA6690000A66A,
+        0xA66B0000A66C,
+        0xA66D0000A670,
+        0xA6740000A67E,
+        0xA67F0000A680,
+        0xA6810000A682,
+        0xA6830000A684,
+        0xA6850000A686,
+        0xA6870000A688,
+        0xA6890000A68A,
+        0xA68B0000A68C,
+        0xA68D0000A68E,
+        0xA68F0000A690,
+        0xA6910000A692,
+        0xA6930000A694,
+        0xA6950000A696,
+        0xA6970000A698,
+        0xA6990000A69A,
+        0xA69B0000A69C,
+        0xA69E0000A6E6,
+        0xA6F00000A6F2,
+        0xA7170000A720,
+        0xA7230000A724,
+        0xA7250000A726,
+        0xA7270000A728,
+        0xA7290000A72A,
+        0xA72B0000A72C,
+        0xA72D0000A72E,
+        0xA72F0000A732,
+        0xA7330000A734,
+        0xA7350000A736,
+        0xA7370000A738,
+        0xA7390000A73A,
+        0xA73B0000A73C,
+        0xA73D0000A73E,
+        0xA73F0000A740,
+        0xA7410000A742,
+        0xA7430000A744,
+        0xA7450000A746,
+        0xA7470000A748,
+        0xA7490000A74A,
+        0xA74B0000A74C,
+        0xA74D0000A74E,
+        0xA74F0000A750,
+        0xA7510000A752,
+        0xA7530000A754,
+        0xA7550000A756,
+        0xA7570000A758,
+        0xA7590000A75A,
+        0xA75B0000A75C,
+        0xA75D0000A75E,
+        0xA75F0000A760,
+        0xA7610000A762,
+        0xA7630000A764,
+        0xA7650000A766,
+        0xA7670000A768,
+        0xA7690000A76A,
+        0xA76B0000A76C,
+        0xA76D0000A76E,
+        0xA76F0000A770,
+        0xA7710000A779,
+        0xA77A0000A77B,
+        0xA77C0000A77D,
+        0xA77F0000A780,
+        0xA7810000A782,
+        0xA7830000A784,
+        0xA7850000A786,
+        0xA7870000A789,
+        0xA78C0000A78D,
+        0xA78E0000A790,
+        0xA7910000A792,
+        0xA7930000A796,
+        0xA7970000A798,
+        0xA7990000A79A,
+        0xA79B0000A79C,
+        0xA79D0000A79E,
+        0xA79F0000A7A0,
+        0xA7A10000A7A2,
+        0xA7A30000A7A4,
+        0xA7A50000A7A6,
+        0xA7A70000A7A8,
+        0xA7A90000A7AA,
+        0xA7AF0000A7B0,
+        0xA7B50000A7B6,
+        0xA7B70000A7B8,
+        0xA7B90000A7BA,
+        0xA7BB0000A7BC,
+        0xA7BD0000A7BE,
+        0xA7BF0000A7C0,
+        0xA7C10000A7C2,
+        0xA7C30000A7C4,
+        0xA7C80000A7C9,
+        0xA7CA0000A7CB,
+        0xA7D10000A7D2,
+        0xA7D30000A7D4,
+        0xA7D50000A7D6,
+        0xA7D70000A7D8,
+        0xA7D90000A7DA,
+        0xA7F60000A7F8,
+        0xA7FA0000A828,
+        0xA82C0000A82D,
+        0xA8400000A874,
+        0xA8800000A8C6,
+        0xA8D00000A8DA,
+        0xA8E00000A8F8,
+        0xA8FB0000A8FC,
+        0xA8FD0000A92E,
+        0xA9300000A954,
+        0xA9800000A9C1,
+        0xA9CF0000A9DA,
+        0xA9E00000A9FF,
+        0xAA000000AA37,
+        0xAA400000AA4E,
+        0xAA500000AA5A,
+        0xAA600000AA77,
+        0xAA7A0000AAC3,
+        0xAADB0000AADE,
+        0xAAE00000AAF0,
+        0xAAF20000AAF7,
+        0xAB010000AB07,
+        0xAB090000AB0F,
+        0xAB110000AB17,
+        0xAB200000AB27,
+        0xAB280000AB2F,
+        0xAB300000AB5B,
+        0xAB600000AB69,
+        0xABC00000ABEB,
+        0xABEC0000ABEE,
+        0xABF00000ABFA,
+        0xAC000000D7A4,
+        0xFA0E0000FA10,
+        0xFA110000FA12,
+        0xFA130000FA15,
+        0xFA1F0000FA20,
+        0xFA210000FA22,
+        0xFA230000FA25,
+        0xFA270000FA2A,
+        0xFB1E0000FB1F,
+        0xFE200000FE30,
+        0xFE730000FE74,
+        0x100000001000C,
+        0x1000D00010027,
+        0x100280001003B,
+        0x1003C0001003E,
+        0x1003F0001004E,
+        0x100500001005E,
+        0x10080000100FB,
+        0x101FD000101FE,
+        0x102800001029D,
+        0x102A0000102D1,
+        0x102E0000102E1,
+        0x1030000010320,
+        0x1032D00010341,
+        0x103420001034A,
+        0x103500001037B,
+        0x103800001039E,
+        0x103A0000103C4,
+        0x103C8000103D0,
+        0x104280001049E,
+        0x104A0000104AA,
+        0x104D8000104FC,
+        0x1050000010528,
+        0x1053000010564,
+        0x10597000105A2,
+        0x105A3000105B2,
+        0x105B3000105BA,
+        0x105BB000105BD,
+        0x1060000010737,
+        0x1074000010756,
+        0x1076000010768,
+        0x1078000010781,
+        0x1080000010806,
+        0x1080800010809,
+        0x1080A00010836,
+        0x1083700010839,
+        0x1083C0001083D,
+        0x1083F00010856,
+        0x1086000010877,
+        0x108800001089F,
+        0x108E0000108F3,
+        0x108F4000108F6,
+        0x1090000010916,
+        0x109200001093A,
+        0x10980000109B8,
+        0x109BE000109C0,
+        0x10A0000010A04,
+        0x10A0500010A07,
+        0x10A0C00010A14,
+        0x10A1500010A18,
+        0x10A1900010A36,
+        0x10A3800010A3B,
+        0x10A3F00010A40,
+        0x10A6000010A7D,
+        0x10A8000010A9D,
+        0x10AC000010AC8,
+        0x10AC900010AE7,
+        0x10B0000010B36,
+        0x10B4000010B56,
+        0x10B6000010B73,
+        0x10B8000010B92,
+        0x10C0000010C49,
+        0x10CC000010CF3,
+        0x10D0000010D28,
+        0x10D3000010D3A,
+        0x10E8000010EAA,
+        0x10EAB00010EAD,
+        0x10EB000010EB2,
+        0x10EFD00010F1D,
+        0x10F2700010F28,
+        0x10F3000010F51,
+        0x10F7000010F86,
+        0x10FB000010FC5,
+        0x10FE000010FF7,
+        0x1100000011047,
+        0x1106600011076,
+        0x1107F000110BB,
+        0x110C2000110C3,
+        0x110D0000110E9,
+        0x110F0000110FA,
+        0x1110000011135,
+        0x1113600011140,
+        0x1114400011148,
+        0x1115000011174,
+        0x1117600011177,
+        0x11180000111C5,
+        0x111C9000111CD,
+        0x111CE000111DB,
+        0x111DC000111DD,
+        0x1120000011212,
+        0x1121300011238,
+        0x1123E00011242,
+        0x1128000011287,
+        0x1128800011289,
+        0x1128A0001128E,
+        0x1128F0001129E,
+        0x1129F000112A9,
+        0x112B0000112EB,
+        0x112F0000112FA,
+        0x1130000011304,
+        0x113050001130D,
+        0x1130F00011311,
+        0x1131300011329,
+        0x1132A00011331,
+        0x1133200011334,
+        0x113350001133A,
+        0x1133B00011345,
+        0x1134700011349,
+        0x1134B0001134E,
+        0x1135000011351,
+        0x1135700011358,
+        0x1135D00011364,
+        0x113660001136D,
+        0x1137000011375,
+        0x114000001144B,
+        0x114500001145A,
+        0x1145E00011462,
+        0x11480000114C6,
+        0x114C7000114C8,
+        0x114D0000114DA,
+        0x11580000115B6,
+        0x115B8000115C1,
+        0x115D8000115DE,
+        0x1160000011641,
+        0x1164400011645,
+        0x116500001165A,
+        0x11680000116B9,
+        0x116C0000116CA,
+        0x117000001171B,
+        0x1171D0001172C,
+        0x117300001173A,
+        0x1174000011747,
+        0x118000001183B,
+        0x118C0000118EA,
+        0x118FF00011907,
+        0x119090001190A,
+        0x1190C00011914,
+        0x1191500011917,
+        0x1191800011936,
+        0x1193700011939,
+        0x1193B00011944,
+        0x119500001195A,
+        0x119A0000119A8,
+        0x119AA000119D8,
+        0x119DA000119E2,
+        0x119E3000119E5,
+        0x11A0000011A3F,
+        0x11A4700011A48,
+        0x11A5000011A9A,
+        0x11A9D00011A9E,
+        0x11AB000011AF9,
+        0x11C0000011C09,
+        0x11C0A00011C37,
+        0x11C3800011C41,
+        0x11C5000011C5A,
+        0x11C7200011C90,
+        0x11C9200011CA8,
+        0x11CA900011CB7,
+        0x11D0000011D07,
+        0x11D0800011D0A,
+        0x11D0B00011D37,
+        0x11D3A00011D3B,
+        0x11D3C00011D3E,
+        0x11D3F00011D48,
+        0x11D5000011D5A,
+        0x11D6000011D66,
+        0x11D6700011D69,
+        0x11D6A00011D8F,
+        0x11D9000011D92,
+        0x11D9300011D99,
+        0x11DA000011DAA,
+        0x11EE000011EF7,
+        0x11F0000011F11,
+        0x11F1200011F3B,
+        0x11F3E00011F43,
+        0x11F5000011F5A,
+        0x11FB000011FB1,
+        0x120000001239A,
+        0x1248000012544,
+        0x12F9000012FF1,
+        0x1300000013430,
+        0x1344000013456,
+        0x1440000014647,
+        0x1680000016A39,
+        0x16A4000016A5F,
+        0x16A6000016A6A,
+        0x16A7000016ABF,
+        0x16AC000016ACA,
+        0x16AD000016AEE,
+        0x16AF000016AF5,
+        0x16B0000016B37,
+        0x16B4000016B44,
+        0x16B5000016B5A,
+        0x16B6300016B78,
+        0x16B7D00016B90,
+        0x16E6000016E80,
+        0x16F0000016F4B,
+        0x16F4F00016F88,
+        0x16F8F00016FA0,
+        0x16FE000016FE2,
+        0x16FE300016FE5,
+        0x16FF000016FF2,
+        0x17000000187F8,
+        0x1880000018CD6,
+        0x18D0000018D09,
+        0x1AFF00001AFF4,
+        0x1AFF50001AFFC,
+        0x1AFFD0001AFFF,
+        0x1B0000001B123,
+        0x1B1320001B133,
+        0x1B1500001B153,
+        0x1B1550001B156,
+        0x1B1640001B168,
+        0x1B1700001B2FC,
+        0x1BC000001BC6B,
+        0x1BC700001BC7D,
+        0x1BC800001BC89,
+        0x1BC900001BC9A,
+        0x1BC9D0001BC9F,
+        0x1CF000001CF2E,
+        0x1CF300001CF47,
+        0x1DA000001DA37,
+        0x1DA3B0001DA6D,
+        0x1DA750001DA76,
+        0x1DA840001DA85,
+        0x1DA9B0001DAA0,
+        0x1DAA10001DAB0,
+        0x1DF000001DF1F,
+        0x1DF250001DF2B,
+        0x1E0000001E007,
+        0x1E0080001E019,
+        0x1E01B0001E022,
+        0x1E0230001E025,
+        0x1E0260001E02B,
+        0x1E08F0001E090,
+        0x1E1000001E12D,
+        0x1E1300001E13E,
+        0x1E1400001E14A,
+        0x1E14E0001E14F,
+        0x1E2900001E2AF,
+        0x1E2C00001E2FA,
+        0x1E4D00001E4FA,
+        0x1E7E00001E7E7,
+        0x1E7E80001E7EC,
+        0x1E7ED0001E7EF,
+        0x1E7F00001E7FF,
+        0x1E8000001E8C5,
+        0x1E8D00001E8D7,
+        0x1E9220001E94C,
+        0x1E9500001E95A,
+        0x200000002A6E0,
+        0x2A7000002B73A,
+        0x2B7400002B81E,
+        0x2B8200002CEA2,
+        0x2CEB00002EBE1,
+        0x2EBF00002EE5E,
+        0x300000003134B,
+        0x31350000323B0,
+    ),
+    "CONTEXTJ": (0x200C0000200E,),
+    "CONTEXTO": (
+        0xB7000000B8,
+        0x37500000376,
+        0x5F3000005F5,
+        0x6600000066A,
+        0x6F0000006FA,
+        0x30FB000030FC,
+    ),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/idna/intranges.py b/backend/venv/lib/python3.12/site-packages/idna/intranges.py
new file mode 100644
index 00000000..7bfaa8d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/intranges.py
@@ -0,0 +1,57 @@
+"""
+Given a list of integers, made up of (hopefully) a small number of long runs
+of consecutive integers, compute a representation of the form
+((start1, end1), (start2, end2) ...). Then answer the question "was x present
+in the original list?" in time O(log(# runs)).
+"""
+
+import bisect
+from typing import List, Tuple
+
+
+def intranges_from_list(list_: List[int]) -> Tuple[int, ...]:
+    """Represent a list of integers as a sequence of ranges:
+    ((start_0, end_0), (start_1, end_1), ...), such that the original
+    integers are exactly those x such that start_i <= x < end_i for some i.
+
+    Ranges are encoded as single integers (start << 32 | end), not as tuples.
+    """
+
+    sorted_list = sorted(list_)
+    ranges = []
+    last_write = -1
+    for i in range(len(sorted_list)):
+        if i + 1 < len(sorted_list):
+            if sorted_list[i] == sorted_list[i + 1] - 1:
+                continue
+        current_range = sorted_list[last_write + 1 : i + 1]
+        ranges.append(_encode_range(current_range[0], current_range[-1] + 1))
+        last_write = i
+
+    return tuple(ranges)
+
+
+def _encode_range(start: int, end: int) -> int:
+    return (start << 32) | end
+
+
+def _decode_range(r: int) -> Tuple[int, int]:
+    return (r >> 32), (r & ((1 << 32) - 1))
+
+
+def intranges_contain(int_: int, ranges: Tuple[int, ...]) -> bool:
+    """Determine if `int_` falls into one of the ranges in `ranges`."""
+    tuple_ = _encode_range(int_, 0)
+    pos = bisect.bisect_left(ranges, tuple_)
+    # we could be immediately ahead of a tuple (start, end)
+    # with start < int_ <= end
+    if pos > 0:
+        left, right = _decode_range(ranges[pos - 1])
+        if left <= int_ < right:
+            return True
+    # or we could be immediately behind a tuple (int_, end)
+    if pos < len(ranges):
+        left, _ = _decode_range(ranges[pos])
+        if left == int_:
+            return True
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/idna/package_data.py b/backend/venv/lib/python3.12/site-packages/idna/package_data.py
new file mode 100644
index 00000000..514ff7e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/package_data.py
@@ -0,0 +1 @@
+__version__ = "3.10"
diff --git a/backend/venv/lib/python3.12/site-packages/idna/py.typed b/backend/venv/lib/python3.12/site-packages/idna/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/idna/uts46data.py b/backend/venv/lib/python3.12/site-packages/idna/uts46data.py
new file mode 100644
index 00000000..eb894327
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/idna/uts46data.py
@@ -0,0 +1,8681 @@
+# This file is automatically generated by tools/idna-data
+# vim: set fileencoding=utf-8 :
+
+from typing import List, Tuple, Union
+
+"""IDNA Mapping Table from UTS46."""
+
+
+__version__ = "15.1.0"
+
+
+def _seg_0() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x0, "3"),
+        (0x1, "3"),
+        (0x2, "3"),
+        (0x3, "3"),
+        (0x4, "3"),
+        (0x5, "3"),
+        (0x6, "3"),
+        (0x7, "3"),
+        (0x8, "3"),
+        (0x9, "3"),
+        (0xA, "3"),
+        (0xB, "3"),
+        (0xC, "3"),
+        (0xD, "3"),
+        (0xE, "3"),
+        (0xF, "3"),
+        (0x10, "3"),
+        (0x11, "3"),
+        (0x12, "3"),
+        (0x13, "3"),
+        (0x14, "3"),
+        (0x15, "3"),
+        (0x16, "3"),
+        (0x17, "3"),
+        (0x18, "3"),
+        (0x19, "3"),
+        (0x1A, "3"),
+        (0x1B, "3"),
+        (0x1C, "3"),
+        (0x1D, "3"),
+        (0x1E, "3"),
+        (0x1F, "3"),
+        (0x20, "3"),
+        (0x21, "3"),
+        (0x22, "3"),
+        (0x23, "3"),
+        (0x24, "3"),
+        (0x25, "3"),
+        (0x26, "3"),
+        (0x27, "3"),
+        (0x28, "3"),
+        (0x29, "3"),
+        (0x2A, "3"),
+        (0x2B, "3"),
+        (0x2C, "3"),
+        (0x2D, "V"),
+        (0x2E, "V"),
+        (0x2F, "3"),
+        (0x30, "V"),
+        (0x31, "V"),
+        (0x32, "V"),
+        (0x33, "V"),
+        (0x34, "V"),
+        (0x35, "V"),
+        (0x36, "V"),
+        (0x37, "V"),
+        (0x38, "V"),
+        (0x39, "V"),
+        (0x3A, "3"),
+        (0x3B, "3"),
+        (0x3C, "3"),
+        (0x3D, "3"),
+        (0x3E, "3"),
+        (0x3F, "3"),
+        (0x40, "3"),
+        (0x41, "M", "a"),
+        (0x42, "M", "b"),
+        (0x43, "M", "c"),
+        (0x44, "M", "d"),
+        (0x45, "M", "e"),
+        (0x46, "M", "f"),
+        (0x47, "M", "g"),
+        (0x48, "M", "h"),
+        (0x49, "M", "i"),
+        (0x4A, "M", "j"),
+        (0x4B, "M", "k"),
+        (0x4C, "M", "l"),
+        (0x4D, "M", "m"),
+        (0x4E, "M", "n"),
+        (0x4F, "M", "o"),
+        (0x50, "M", "p"),
+        (0x51, "M", "q"),
+        (0x52, "M", "r"),
+        (0x53, "M", "s"),
+        (0x54, "M", "t"),
+        (0x55, "M", "u"),
+        (0x56, "M", "v"),
+        (0x57, "M", "w"),
+        (0x58, "M", "x"),
+        (0x59, "M", "y"),
+        (0x5A, "M", "z"),
+        (0x5B, "3"),
+        (0x5C, "3"),
+        (0x5D, "3"),
+        (0x5E, "3"),
+        (0x5F, "3"),
+        (0x60, "3"),
+        (0x61, "V"),
+        (0x62, "V"),
+        (0x63, "V"),
+    ]
+
+
+def _seg_1() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x64, "V"),
+        (0x65, "V"),
+        (0x66, "V"),
+        (0x67, "V"),
+        (0x68, "V"),
+        (0x69, "V"),
+        (0x6A, "V"),
+        (0x6B, "V"),
+        (0x6C, "V"),
+        (0x6D, "V"),
+        (0x6E, "V"),
+        (0x6F, "V"),
+        (0x70, "V"),
+        (0x71, "V"),
+        (0x72, "V"),
+        (0x73, "V"),
+        (0x74, "V"),
+        (0x75, "V"),
+        (0x76, "V"),
+        (0x77, "V"),
+        (0x78, "V"),
+        (0x79, "V"),
+        (0x7A, "V"),
+        (0x7B, "3"),
+        (0x7C, "3"),
+        (0x7D, "3"),
+        (0x7E, "3"),
+        (0x7F, "3"),
+        (0x80, "X"),
+        (0x81, "X"),
+        (0x82, "X"),
+        (0x83, "X"),
+        (0x84, "X"),
+        (0x85, "X"),
+        (0x86, "X"),
+        (0x87, "X"),
+        (0x88, "X"),
+        (0x89, "X"),
+        (0x8A, "X"),
+        (0x8B, "X"),
+        (0x8C, "X"),
+        (0x8D, "X"),
+        (0x8E, "X"),
+        (0x8F, "X"),
+        (0x90, "X"),
+        (0x91, "X"),
+        (0x92, "X"),
+        (0x93, "X"),
+        (0x94, "X"),
+        (0x95, "X"),
+        (0x96, "X"),
+        (0x97, "X"),
+        (0x98, "X"),
+        (0x99, "X"),
+        (0x9A, "X"),
+        (0x9B, "X"),
+        (0x9C, "X"),
+        (0x9D, "X"),
+        (0x9E, "X"),
+        (0x9F, "X"),
+        (0xA0, "3", " "),
+        (0xA1, "V"),
+        (0xA2, "V"),
+        (0xA3, "V"),
+        (0xA4, "V"),
+        (0xA5, "V"),
+        (0xA6, "V"),
+        (0xA7, "V"),
+        (0xA8, "3", " ̈"),
+        (0xA9, "V"),
+        (0xAA, "M", "a"),
+        (0xAB, "V"),
+        (0xAC, "V"),
+        (0xAD, "I"),
+        (0xAE, "V"),
+        (0xAF, "3", " ̄"),
+        (0xB0, "V"),
+        (0xB1, "V"),
+        (0xB2, "M", "2"),
+        (0xB3, "M", "3"),
+        (0xB4, "3", " ́"),
+        (0xB5, "M", "μ"),
+        (0xB6, "V"),
+        (0xB7, "V"),
+        (0xB8, "3", " ̧"),
+        (0xB9, "M", "1"),
+        (0xBA, "M", "o"),
+        (0xBB, "V"),
+        (0xBC, "M", "1⁄4"),
+        (0xBD, "M", "1⁄2"),
+        (0xBE, "M", "3⁄4"),
+        (0xBF, "V"),
+        (0xC0, "M", "à"),
+        (0xC1, "M", "á"),
+        (0xC2, "M", "â"),
+        (0xC3, "M", "ã"),
+        (0xC4, "M", "ä"),
+        (0xC5, "M", "å"),
+        (0xC6, "M", "æ"),
+        (0xC7, "M", "ç"),
+    ]
+
+
+def _seg_2() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xC8, "M", "è"),
+        (0xC9, "M", "é"),
+        (0xCA, "M", "ê"),
+        (0xCB, "M", "ë"),
+        (0xCC, "M", "ì"),
+        (0xCD, "M", "í"),
+        (0xCE, "M", "î"),
+        (0xCF, "M", "ï"),
+        (0xD0, "M", "ð"),
+        (0xD1, "M", "ñ"),
+        (0xD2, "M", "ò"),
+        (0xD3, "M", "ó"),
+        (0xD4, "M", "ô"),
+        (0xD5, "M", "õ"),
+        (0xD6, "M", "ö"),
+        (0xD7, "V"),
+        (0xD8, "M", "ø"),
+        (0xD9, "M", "ù"),
+        (0xDA, "M", "ú"),
+        (0xDB, "M", "û"),
+        (0xDC, "M", "ü"),
+        (0xDD, "M", "ý"),
+        (0xDE, "M", "þ"),
+        (0xDF, "D", "ss"),
+        (0xE0, "V"),
+        (0xE1, "V"),
+        (0xE2, "V"),
+        (0xE3, "V"),
+        (0xE4, "V"),
+        (0xE5, "V"),
+        (0xE6, "V"),
+        (0xE7, "V"),
+        (0xE8, "V"),
+        (0xE9, "V"),
+        (0xEA, "V"),
+        (0xEB, "V"),
+        (0xEC, "V"),
+        (0xED, "V"),
+        (0xEE, "V"),
+        (0xEF, "V"),
+        (0xF0, "V"),
+        (0xF1, "V"),
+        (0xF2, "V"),
+        (0xF3, "V"),
+        (0xF4, "V"),
+        (0xF5, "V"),
+        (0xF6, "V"),
+        (0xF7, "V"),
+        (0xF8, "V"),
+        (0xF9, "V"),
+        (0xFA, "V"),
+        (0xFB, "V"),
+        (0xFC, "V"),
+        (0xFD, "V"),
+        (0xFE, "V"),
+        (0xFF, "V"),
+        (0x100, "M", "ā"),
+        (0x101, "V"),
+        (0x102, "M", "ă"),
+        (0x103, "V"),
+        (0x104, "M", "ą"),
+        (0x105, "V"),
+        (0x106, "M", "ć"),
+        (0x107, "V"),
+        (0x108, "M", "ĉ"),
+        (0x109, "V"),
+        (0x10A, "M", "ċ"),
+        (0x10B, "V"),
+        (0x10C, "M", "č"),
+        (0x10D, "V"),
+        (0x10E, "M", "ď"),
+        (0x10F, "V"),
+        (0x110, "M", "đ"),
+        (0x111, "V"),
+        (0x112, "M", "ē"),
+        (0x113, "V"),
+        (0x114, "M", "ĕ"),
+        (0x115, "V"),
+        (0x116, "M", "ė"),
+        (0x117, "V"),
+        (0x118, "M", "ę"),
+        (0x119, "V"),
+        (0x11A, "M", "ě"),
+        (0x11B, "V"),
+        (0x11C, "M", "ĝ"),
+        (0x11D, "V"),
+        (0x11E, "M", "ğ"),
+        (0x11F, "V"),
+        (0x120, "M", "ġ"),
+        (0x121, "V"),
+        (0x122, "M", "ģ"),
+        (0x123, "V"),
+        (0x124, "M", "ĥ"),
+        (0x125, "V"),
+        (0x126, "M", "ħ"),
+        (0x127, "V"),
+        (0x128, "M", "ĩ"),
+        (0x129, "V"),
+        (0x12A, "M", "ī"),
+        (0x12B, "V"),
+    ]
+
+
+def _seg_3() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x12C, "M", "ĭ"),
+        (0x12D, "V"),
+        (0x12E, "M", "į"),
+        (0x12F, "V"),
+        (0x130, "M", "i̇"),
+        (0x131, "V"),
+        (0x132, "M", "ij"),
+        (0x134, "M", "ĵ"),
+        (0x135, "V"),
+        (0x136, "M", "ķ"),
+        (0x137, "V"),
+        (0x139, "M", "ĺ"),
+        (0x13A, "V"),
+        (0x13B, "M", "ļ"),
+        (0x13C, "V"),
+        (0x13D, "M", "ľ"),
+        (0x13E, "V"),
+        (0x13F, "M", "l·"),
+        (0x141, "M", "ł"),
+        (0x142, "V"),
+        (0x143, "M", "ń"),
+        (0x144, "V"),
+        (0x145, "M", "ņ"),
+        (0x146, "V"),
+        (0x147, "M", "ň"),
+        (0x148, "V"),
+        (0x149, "M", "ʼn"),
+        (0x14A, "M", "ŋ"),
+        (0x14B, "V"),
+        (0x14C, "M", "ō"),
+        (0x14D, "V"),
+        (0x14E, "M", "ŏ"),
+        (0x14F, "V"),
+        (0x150, "M", "ő"),
+        (0x151, "V"),
+        (0x152, "M", "œ"),
+        (0x153, "V"),
+        (0x154, "M", "ŕ"),
+        (0x155, "V"),
+        (0x156, "M", "ŗ"),
+        (0x157, "V"),
+        (0x158, "M", "ř"),
+        (0x159, "V"),
+        (0x15A, "M", "ś"),
+        (0x15B, "V"),
+        (0x15C, "M", "ŝ"),
+        (0x15D, "V"),
+        (0x15E, "M", "ş"),
+        (0x15F, "V"),
+        (0x160, "M", "š"),
+        (0x161, "V"),
+        (0x162, "M", "ţ"),
+        (0x163, "V"),
+        (0x164, "M", "ť"),
+        (0x165, "V"),
+        (0x166, "M", "ŧ"),
+        (0x167, "V"),
+        (0x168, "M", "ũ"),
+        (0x169, "V"),
+        (0x16A, "M", "ū"),
+        (0x16B, "V"),
+        (0x16C, "M", "ŭ"),
+        (0x16D, "V"),
+        (0x16E, "M", "ů"),
+        (0x16F, "V"),
+        (0x170, "M", "ű"),
+        (0x171, "V"),
+        (0x172, "M", "ų"),
+        (0x173, "V"),
+        (0x174, "M", "ŵ"),
+        (0x175, "V"),
+        (0x176, "M", "ŷ"),
+        (0x177, "V"),
+        (0x178, "M", "ÿ"),
+        (0x179, "M", "ź"),
+        (0x17A, "V"),
+        (0x17B, "M", "ż"),
+        (0x17C, "V"),
+        (0x17D, "M", "ž"),
+        (0x17E, "V"),
+        (0x17F, "M", "s"),
+        (0x180, "V"),
+        (0x181, "M", "ɓ"),
+        (0x182, "M", "ƃ"),
+        (0x183, "V"),
+        (0x184, "M", "ƅ"),
+        (0x185, "V"),
+        (0x186, "M", "ɔ"),
+        (0x187, "M", "ƈ"),
+        (0x188, "V"),
+        (0x189, "M", "ɖ"),
+        (0x18A, "M", "ɗ"),
+        (0x18B, "M", "ƌ"),
+        (0x18C, "V"),
+        (0x18E, "M", "ǝ"),
+        (0x18F, "M", "ə"),
+        (0x190, "M", "ɛ"),
+        (0x191, "M", "ƒ"),
+        (0x192, "V"),
+        (0x193, "M", "ɠ"),
+    ]
+
+
+def _seg_4() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x194, "M", "ɣ"),
+        (0x195, "V"),
+        (0x196, "M", "ɩ"),
+        (0x197, "M", "ɨ"),
+        (0x198, "M", "ƙ"),
+        (0x199, "V"),
+        (0x19C, "M", "ɯ"),
+        (0x19D, "M", "ɲ"),
+        (0x19E, "V"),
+        (0x19F, "M", "ɵ"),
+        (0x1A0, "M", "ơ"),
+        (0x1A1, "V"),
+        (0x1A2, "M", "ƣ"),
+        (0x1A3, "V"),
+        (0x1A4, "M", "ƥ"),
+        (0x1A5, "V"),
+        (0x1A6, "M", "ʀ"),
+        (0x1A7, "M", "ƨ"),
+        (0x1A8, "V"),
+        (0x1A9, "M", "ʃ"),
+        (0x1AA, "V"),
+        (0x1AC, "M", "ƭ"),
+        (0x1AD, "V"),
+        (0x1AE, "M", "ʈ"),
+        (0x1AF, "M", "ư"),
+        (0x1B0, "V"),
+        (0x1B1, "M", "ʊ"),
+        (0x1B2, "M", "ʋ"),
+        (0x1B3, "M", "ƴ"),
+        (0x1B4, "V"),
+        (0x1B5, "M", "ƶ"),
+        (0x1B6, "V"),
+        (0x1B7, "M", "ʒ"),
+        (0x1B8, "M", "ƹ"),
+        (0x1B9, "V"),
+        (0x1BC, "M", "ƽ"),
+        (0x1BD, "V"),
+        (0x1C4, "M", "dž"),
+        (0x1C7, "M", "lj"),
+        (0x1CA, "M", "nj"),
+        (0x1CD, "M", "ǎ"),
+        (0x1CE, "V"),
+        (0x1CF, "M", "ǐ"),
+        (0x1D0, "V"),
+        (0x1D1, "M", "ǒ"),
+        (0x1D2, "V"),
+        (0x1D3, "M", "ǔ"),
+        (0x1D4, "V"),
+        (0x1D5, "M", "ǖ"),
+        (0x1D6, "V"),
+        (0x1D7, "M", "ǘ"),
+        (0x1D8, "V"),
+        (0x1D9, "M", "ǚ"),
+        (0x1DA, "V"),
+        (0x1DB, "M", "ǜ"),
+        (0x1DC, "V"),
+        (0x1DE, "M", "ǟ"),
+        (0x1DF, "V"),
+        (0x1E0, "M", "ǡ"),
+        (0x1E1, "V"),
+        (0x1E2, "M", "ǣ"),
+        (0x1E3, "V"),
+        (0x1E4, "M", "ǥ"),
+        (0x1E5, "V"),
+        (0x1E6, "M", "ǧ"),
+        (0x1E7, "V"),
+        (0x1E8, "M", "ǩ"),
+        (0x1E9, "V"),
+        (0x1EA, "M", "ǫ"),
+        (0x1EB, "V"),
+        (0x1EC, "M", "ǭ"),
+        (0x1ED, "V"),
+        (0x1EE, "M", "ǯ"),
+        (0x1EF, "V"),
+        (0x1F1, "M", "dz"),
+        (0x1F4, "M", "ǵ"),
+        (0x1F5, "V"),
+        (0x1F6, "M", "ƕ"),
+        (0x1F7, "M", "ƿ"),
+        (0x1F8, "M", "ǹ"),
+        (0x1F9, "V"),
+        (0x1FA, "M", "ǻ"),
+        (0x1FB, "V"),
+        (0x1FC, "M", "ǽ"),
+        (0x1FD, "V"),
+        (0x1FE, "M", "ǿ"),
+        (0x1FF, "V"),
+        (0x200, "M", "ȁ"),
+        (0x201, "V"),
+        (0x202, "M", "ȃ"),
+        (0x203, "V"),
+        (0x204, "M", "ȅ"),
+        (0x205, "V"),
+        (0x206, "M", "ȇ"),
+        (0x207, "V"),
+        (0x208, "M", "ȉ"),
+        (0x209, "V"),
+        (0x20A, "M", "ȋ"),
+        (0x20B, "V"),
+        (0x20C, "M", "ȍ"),
+    ]
+
+
+def _seg_5() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x20D, "V"),
+        (0x20E, "M", "ȏ"),
+        (0x20F, "V"),
+        (0x210, "M", "ȑ"),
+        (0x211, "V"),
+        (0x212, "M", "ȓ"),
+        (0x213, "V"),
+        (0x214, "M", "ȕ"),
+        (0x215, "V"),
+        (0x216, "M", "ȗ"),
+        (0x217, "V"),
+        (0x218, "M", "ș"),
+        (0x219, "V"),
+        (0x21A, "M", "ț"),
+        (0x21B, "V"),
+        (0x21C, "M", "ȝ"),
+        (0x21D, "V"),
+        (0x21E, "M", "ȟ"),
+        (0x21F, "V"),
+        (0x220, "M", "ƞ"),
+        (0x221, "V"),
+        (0x222, "M", "ȣ"),
+        (0x223, "V"),
+        (0x224, "M", "ȥ"),
+        (0x225, "V"),
+        (0x226, "M", "ȧ"),
+        (0x227, "V"),
+        (0x228, "M", "ȩ"),
+        (0x229, "V"),
+        (0x22A, "M", "ȫ"),
+        (0x22B, "V"),
+        (0x22C, "M", "ȭ"),
+        (0x22D, "V"),
+        (0x22E, "M", "ȯ"),
+        (0x22F, "V"),
+        (0x230, "M", "ȱ"),
+        (0x231, "V"),
+        (0x232, "M", "ȳ"),
+        (0x233, "V"),
+        (0x23A, "M", "ⱥ"),
+        (0x23B, "M", "ȼ"),
+        (0x23C, "V"),
+        (0x23D, "M", "ƚ"),
+        (0x23E, "M", "ⱦ"),
+        (0x23F, "V"),
+        (0x241, "M", "ɂ"),
+        (0x242, "V"),
+        (0x243, "M", "ƀ"),
+        (0x244, "M", "ʉ"),
+        (0x245, "M", "ʌ"),
+        (0x246, "M", "ɇ"),
+        (0x247, "V"),
+        (0x248, "M", "ɉ"),
+        (0x249, "V"),
+        (0x24A, "M", "ɋ"),
+        (0x24B, "V"),
+        (0x24C, "M", "ɍ"),
+        (0x24D, "V"),
+        (0x24E, "M", "ɏ"),
+        (0x24F, "V"),
+        (0x2B0, "M", "h"),
+        (0x2B1, "M", "ɦ"),
+        (0x2B2, "M", "j"),
+        (0x2B3, "M", "r"),
+        (0x2B4, "M", "ɹ"),
+        (0x2B5, "M", "ɻ"),
+        (0x2B6, "M", "ʁ"),
+        (0x2B7, "M", "w"),
+        (0x2B8, "M", "y"),
+        (0x2B9, "V"),
+        (0x2D8, "3", " ̆"),
+        (0x2D9, "3", " ̇"),
+        (0x2DA, "3", " ̊"),
+        (0x2DB, "3", " ̨"),
+        (0x2DC, "3", " ̃"),
+        (0x2DD, "3", " ̋"),
+        (0x2DE, "V"),
+        (0x2E0, "M", "ɣ"),
+        (0x2E1, "M", "l"),
+        (0x2E2, "M", "s"),
+        (0x2E3, "M", "x"),
+        (0x2E4, "M", "ʕ"),
+        (0x2E5, "V"),
+        (0x340, "M", "̀"),
+        (0x341, "M", "́"),
+        (0x342, "V"),
+        (0x343, "M", "̓"),
+        (0x344, "M", "̈́"),
+        (0x345, "M", "ι"),
+        (0x346, "V"),
+        (0x34F, "I"),
+        (0x350, "V"),
+        (0x370, "M", "ͱ"),
+        (0x371, "V"),
+        (0x372, "M", "ͳ"),
+        (0x373, "V"),
+        (0x374, "M", "ʹ"),
+        (0x375, "V"),
+        (0x376, "M", "ͷ"),
+        (0x377, "V"),
+    ]
+
+
+def _seg_6() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x378, "X"),
+        (0x37A, "3", " ι"),
+        (0x37B, "V"),
+        (0x37E, "3", ";"),
+        (0x37F, "M", "ϳ"),
+        (0x380, "X"),
+        (0x384, "3", " ́"),
+        (0x385, "3", " ̈́"),
+        (0x386, "M", "ά"),
+        (0x387, "M", "·"),
+        (0x388, "M", "έ"),
+        (0x389, "M", "ή"),
+        (0x38A, "M", "ί"),
+        (0x38B, "X"),
+        (0x38C, "M", "ό"),
+        (0x38D, "X"),
+        (0x38E, "M", "ύ"),
+        (0x38F, "M", "ώ"),
+        (0x390, "V"),
+        (0x391, "M", "α"),
+        (0x392, "M", "β"),
+        (0x393, "M", "γ"),
+        (0x394, "M", "δ"),
+        (0x395, "M", "ε"),
+        (0x396, "M", "ζ"),
+        (0x397, "M", "η"),
+        (0x398, "M", "θ"),
+        (0x399, "M", "ι"),
+        (0x39A, "M", "κ"),
+        (0x39B, "M", "λ"),
+        (0x39C, "M", "μ"),
+        (0x39D, "M", "ν"),
+        (0x39E, "M", "ξ"),
+        (0x39F, "M", "ο"),
+        (0x3A0, "M", "π"),
+        (0x3A1, "M", "ρ"),
+        (0x3A2, "X"),
+        (0x3A3, "M", "σ"),
+        (0x3A4, "M", "τ"),
+        (0x3A5, "M", "υ"),
+        (0x3A6, "M", "φ"),
+        (0x3A7, "M", "χ"),
+        (0x3A8, "M", "ψ"),
+        (0x3A9, "M", "ω"),
+        (0x3AA, "M", "ϊ"),
+        (0x3AB, "M", "ϋ"),
+        (0x3AC, "V"),
+        (0x3C2, "D", "σ"),
+        (0x3C3, "V"),
+        (0x3CF, "M", "ϗ"),
+        (0x3D0, "M", "β"),
+        (0x3D1, "M", "θ"),
+        (0x3D2, "M", "υ"),
+        (0x3D3, "M", "ύ"),
+        (0x3D4, "M", "ϋ"),
+        (0x3D5, "M", "φ"),
+        (0x3D6, "M", "π"),
+        (0x3D7, "V"),
+        (0x3D8, "M", "ϙ"),
+        (0x3D9, "V"),
+        (0x3DA, "M", "ϛ"),
+        (0x3DB, "V"),
+        (0x3DC, "M", "ϝ"),
+        (0x3DD, "V"),
+        (0x3DE, "M", "ϟ"),
+        (0x3DF, "V"),
+        (0x3E0, "M", "ϡ"),
+        (0x3E1, "V"),
+        (0x3E2, "M", "ϣ"),
+        (0x3E3, "V"),
+        (0x3E4, "M", "ϥ"),
+        (0x3E5, "V"),
+        (0x3E6, "M", "ϧ"),
+        (0x3E7, "V"),
+        (0x3E8, "M", "ϩ"),
+        (0x3E9, "V"),
+        (0x3EA, "M", "ϫ"),
+        (0x3EB, "V"),
+        (0x3EC, "M", "ϭ"),
+        (0x3ED, "V"),
+        (0x3EE, "M", "ϯ"),
+        (0x3EF, "V"),
+        (0x3F0, "M", "κ"),
+        (0x3F1, "M", "ρ"),
+        (0x3F2, "M", "σ"),
+        (0x3F3, "V"),
+        (0x3F4, "M", "θ"),
+        (0x3F5, "M", "ε"),
+        (0x3F6, "V"),
+        (0x3F7, "M", "ϸ"),
+        (0x3F8, "V"),
+        (0x3F9, "M", "σ"),
+        (0x3FA, "M", "ϻ"),
+        (0x3FB, "V"),
+        (0x3FD, "M", "ͻ"),
+        (0x3FE, "M", "ͼ"),
+        (0x3FF, "M", "ͽ"),
+        (0x400, "M", "ѐ"),
+        (0x401, "M", "ё"),
+        (0x402, "M", "ђ"),
+    ]
+
+
+def _seg_7() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x403, "M", "ѓ"),
+        (0x404, "M", "є"),
+        (0x405, "M", "ѕ"),
+        (0x406, "M", "і"),
+        (0x407, "M", "ї"),
+        (0x408, "M", "ј"),
+        (0x409, "M", "љ"),
+        (0x40A, "M", "њ"),
+        (0x40B, "M", "ћ"),
+        (0x40C, "M", "ќ"),
+        (0x40D, "M", "ѝ"),
+        (0x40E, "M", "ў"),
+        (0x40F, "M", "џ"),
+        (0x410, "M", "а"),
+        (0x411, "M", "б"),
+        (0x412, "M", "в"),
+        (0x413, "M", "г"),
+        (0x414, "M", "д"),
+        (0x415, "M", "е"),
+        (0x416, "M", "ж"),
+        (0x417, "M", "з"),
+        (0x418, "M", "и"),
+        (0x419, "M", "й"),
+        (0x41A, "M", "к"),
+        (0x41B, "M", "л"),
+        (0x41C, "M", "м"),
+        (0x41D, "M", "н"),
+        (0x41E, "M", "о"),
+        (0x41F, "M", "п"),
+        (0x420, "M", "р"),
+        (0x421, "M", "с"),
+        (0x422, "M", "т"),
+        (0x423, "M", "у"),
+        (0x424, "M", "ф"),
+        (0x425, "M", "х"),
+        (0x426, "M", "ц"),
+        (0x427, "M", "ч"),
+        (0x428, "M", "ш"),
+        (0x429, "M", "щ"),
+        (0x42A, "M", "ъ"),
+        (0x42B, "M", "ы"),
+        (0x42C, "M", "ь"),
+        (0x42D, "M", "э"),
+        (0x42E, "M", "ю"),
+        (0x42F, "M", "я"),
+        (0x430, "V"),
+        (0x460, "M", "ѡ"),
+        (0x461, "V"),
+        (0x462, "M", "ѣ"),
+        (0x463, "V"),
+        (0x464, "M", "ѥ"),
+        (0x465, "V"),
+        (0x466, "M", "ѧ"),
+        (0x467, "V"),
+        (0x468, "M", "ѩ"),
+        (0x469, "V"),
+        (0x46A, "M", "ѫ"),
+        (0x46B, "V"),
+        (0x46C, "M", "ѭ"),
+        (0x46D, "V"),
+        (0x46E, "M", "ѯ"),
+        (0x46F, "V"),
+        (0x470, "M", "ѱ"),
+        (0x471, "V"),
+        (0x472, "M", "ѳ"),
+        (0x473, "V"),
+        (0x474, "M", "ѵ"),
+        (0x475, "V"),
+        (0x476, "M", "ѷ"),
+        (0x477, "V"),
+        (0x478, "M", "ѹ"),
+        (0x479, "V"),
+        (0x47A, "M", "ѻ"),
+        (0x47B, "V"),
+        (0x47C, "M", "ѽ"),
+        (0x47D, "V"),
+        (0x47E, "M", "ѿ"),
+        (0x47F, "V"),
+        (0x480, "M", "ҁ"),
+        (0x481, "V"),
+        (0x48A, "M", "ҋ"),
+        (0x48B, "V"),
+        (0x48C, "M", "ҍ"),
+        (0x48D, "V"),
+        (0x48E, "M", "ҏ"),
+        (0x48F, "V"),
+        (0x490, "M", "ґ"),
+        (0x491, "V"),
+        (0x492, "M", "ғ"),
+        (0x493, "V"),
+        (0x494, "M", "ҕ"),
+        (0x495, "V"),
+        (0x496, "M", "җ"),
+        (0x497, "V"),
+        (0x498, "M", "ҙ"),
+        (0x499, "V"),
+        (0x49A, "M", "қ"),
+        (0x49B, "V"),
+        (0x49C, "M", "ҝ"),
+        (0x49D, "V"),
+    ]
+
+
+def _seg_8() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x49E, "M", "ҟ"),
+        (0x49F, "V"),
+        (0x4A0, "M", "ҡ"),
+        (0x4A1, "V"),
+        (0x4A2, "M", "ң"),
+        (0x4A3, "V"),
+        (0x4A4, "M", "ҥ"),
+        (0x4A5, "V"),
+        (0x4A6, "M", "ҧ"),
+        (0x4A7, "V"),
+        (0x4A8, "M", "ҩ"),
+        (0x4A9, "V"),
+        (0x4AA, "M", "ҫ"),
+        (0x4AB, "V"),
+        (0x4AC, "M", "ҭ"),
+        (0x4AD, "V"),
+        (0x4AE, "M", "ү"),
+        (0x4AF, "V"),
+        (0x4B0, "M", "ұ"),
+        (0x4B1, "V"),
+        (0x4B2, "M", "ҳ"),
+        (0x4B3, "V"),
+        (0x4B4, "M", "ҵ"),
+        (0x4B5, "V"),
+        (0x4B6, "M", "ҷ"),
+        (0x4B7, "V"),
+        (0x4B8, "M", "ҹ"),
+        (0x4B9, "V"),
+        (0x4BA, "M", "һ"),
+        (0x4BB, "V"),
+        (0x4BC, "M", "ҽ"),
+        (0x4BD, "V"),
+        (0x4BE, "M", "ҿ"),
+        (0x4BF, "V"),
+        (0x4C0, "X"),
+        (0x4C1, "M", "ӂ"),
+        (0x4C2, "V"),
+        (0x4C3, "M", "ӄ"),
+        (0x4C4, "V"),
+        (0x4C5, "M", "ӆ"),
+        (0x4C6, "V"),
+        (0x4C7, "M", "ӈ"),
+        (0x4C8, "V"),
+        (0x4C9, "M", "ӊ"),
+        (0x4CA, "V"),
+        (0x4CB, "M", "ӌ"),
+        (0x4CC, "V"),
+        (0x4CD, "M", "ӎ"),
+        (0x4CE, "V"),
+        (0x4D0, "M", "ӑ"),
+        (0x4D1, "V"),
+        (0x4D2, "M", "ӓ"),
+        (0x4D3, "V"),
+        (0x4D4, "M", "ӕ"),
+        (0x4D5, "V"),
+        (0x4D6, "M", "ӗ"),
+        (0x4D7, "V"),
+        (0x4D8, "M", "ә"),
+        (0x4D9, "V"),
+        (0x4DA, "M", "ӛ"),
+        (0x4DB, "V"),
+        (0x4DC, "M", "ӝ"),
+        (0x4DD, "V"),
+        (0x4DE, "M", "ӟ"),
+        (0x4DF, "V"),
+        (0x4E0, "M", "ӡ"),
+        (0x4E1, "V"),
+        (0x4E2, "M", "ӣ"),
+        (0x4E3, "V"),
+        (0x4E4, "M", "ӥ"),
+        (0x4E5, "V"),
+        (0x4E6, "M", "ӧ"),
+        (0x4E7, "V"),
+        (0x4E8, "M", "ө"),
+        (0x4E9, "V"),
+        (0x4EA, "M", "ӫ"),
+        (0x4EB, "V"),
+        (0x4EC, "M", "ӭ"),
+        (0x4ED, "V"),
+        (0x4EE, "M", "ӯ"),
+        (0x4EF, "V"),
+        (0x4F0, "M", "ӱ"),
+        (0x4F1, "V"),
+        (0x4F2, "M", "ӳ"),
+        (0x4F3, "V"),
+        (0x4F4, "M", "ӵ"),
+        (0x4F5, "V"),
+        (0x4F6, "M", "ӷ"),
+        (0x4F7, "V"),
+        (0x4F8, "M", "ӹ"),
+        (0x4F9, "V"),
+        (0x4FA, "M", "ӻ"),
+        (0x4FB, "V"),
+        (0x4FC, "M", "ӽ"),
+        (0x4FD, "V"),
+        (0x4FE, "M", "ӿ"),
+        (0x4FF, "V"),
+        (0x500, "M", "ԁ"),
+        (0x501, "V"),
+        (0x502, "M", "ԃ"),
+    ]
+
+
+def _seg_9() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x503, "V"),
+        (0x504, "M", "ԅ"),
+        (0x505, "V"),
+        (0x506, "M", "ԇ"),
+        (0x507, "V"),
+        (0x508, "M", "ԉ"),
+        (0x509, "V"),
+        (0x50A, "M", "ԋ"),
+        (0x50B, "V"),
+        (0x50C, "M", "ԍ"),
+        (0x50D, "V"),
+        (0x50E, "M", "ԏ"),
+        (0x50F, "V"),
+        (0x510, "M", "ԑ"),
+        (0x511, "V"),
+        (0x512, "M", "ԓ"),
+        (0x513, "V"),
+        (0x514, "M", "ԕ"),
+        (0x515, "V"),
+        (0x516, "M", "ԗ"),
+        (0x517, "V"),
+        (0x518, "M", "ԙ"),
+        (0x519, "V"),
+        (0x51A, "M", "ԛ"),
+        (0x51B, "V"),
+        (0x51C, "M", "ԝ"),
+        (0x51D, "V"),
+        (0x51E, "M", "ԟ"),
+        (0x51F, "V"),
+        (0x520, "M", "ԡ"),
+        (0x521, "V"),
+        (0x522, "M", "ԣ"),
+        (0x523, "V"),
+        (0x524, "M", "ԥ"),
+        (0x525, "V"),
+        (0x526, "M", "ԧ"),
+        (0x527, "V"),
+        (0x528, "M", "ԩ"),
+        (0x529, "V"),
+        (0x52A, "M", "ԫ"),
+        (0x52B, "V"),
+        (0x52C, "M", "ԭ"),
+        (0x52D, "V"),
+        (0x52E, "M", "ԯ"),
+        (0x52F, "V"),
+        (0x530, "X"),
+        (0x531, "M", "ա"),
+        (0x532, "M", "բ"),
+        (0x533, "M", "գ"),
+        (0x534, "M", "դ"),
+        (0x535, "M", "ե"),
+        (0x536, "M", "զ"),
+        (0x537, "M", "է"),
+        (0x538, "M", "ը"),
+        (0x539, "M", "թ"),
+        (0x53A, "M", "ժ"),
+        (0x53B, "M", "ի"),
+        (0x53C, "M", "լ"),
+        (0x53D, "M", "խ"),
+        (0x53E, "M", "ծ"),
+        (0x53F, "M", "կ"),
+        (0x540, "M", "հ"),
+        (0x541, "M", "ձ"),
+        (0x542, "M", "ղ"),
+        (0x543, "M", "ճ"),
+        (0x544, "M", "մ"),
+        (0x545, "M", "յ"),
+        (0x546, "M", "ն"),
+        (0x547, "M", "շ"),
+        (0x548, "M", "ո"),
+        (0x549, "M", "չ"),
+        (0x54A, "M", "պ"),
+        (0x54B, "M", "ջ"),
+        (0x54C, "M", "ռ"),
+        (0x54D, "M", "ս"),
+        (0x54E, "M", "վ"),
+        (0x54F, "M", "տ"),
+        (0x550, "M", "ր"),
+        (0x551, "M", "ց"),
+        (0x552, "M", "ւ"),
+        (0x553, "M", "փ"),
+        (0x554, "M", "ք"),
+        (0x555, "M", "օ"),
+        (0x556, "M", "ֆ"),
+        (0x557, "X"),
+        (0x559, "V"),
+        (0x587, "M", "եւ"),
+        (0x588, "V"),
+        (0x58B, "X"),
+        (0x58D, "V"),
+        (0x590, "X"),
+        (0x591, "V"),
+        (0x5C8, "X"),
+        (0x5D0, "V"),
+        (0x5EB, "X"),
+        (0x5EF, "V"),
+        (0x5F5, "X"),
+        (0x606, "V"),
+        (0x61C, "X"),
+        (0x61D, "V"),
+    ]
+
+
+def _seg_10() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x675, "M", "اٴ"),
+        (0x676, "M", "وٴ"),
+        (0x677, "M", "ۇٴ"),
+        (0x678, "M", "يٴ"),
+        (0x679, "V"),
+        (0x6DD, "X"),
+        (0x6DE, "V"),
+        (0x70E, "X"),
+        (0x710, "V"),
+        (0x74B, "X"),
+        (0x74D, "V"),
+        (0x7B2, "X"),
+        (0x7C0, "V"),
+        (0x7FB, "X"),
+        (0x7FD, "V"),
+        (0x82E, "X"),
+        (0x830, "V"),
+        (0x83F, "X"),
+        (0x840, "V"),
+        (0x85C, "X"),
+        (0x85E, "V"),
+        (0x85F, "X"),
+        (0x860, "V"),
+        (0x86B, "X"),
+        (0x870, "V"),
+        (0x88F, "X"),
+        (0x898, "V"),
+        (0x8E2, "X"),
+        (0x8E3, "V"),
+        (0x958, "M", "क़"),
+        (0x959, "M", "ख़"),
+        (0x95A, "M", "ग़"),
+        (0x95B, "M", "ज़"),
+        (0x95C, "M", "ड़"),
+        (0x95D, "M", "ढ़"),
+        (0x95E, "M", "फ़"),
+        (0x95F, "M", "य़"),
+        (0x960, "V"),
+        (0x984, "X"),
+        (0x985, "V"),
+        (0x98D, "X"),
+        (0x98F, "V"),
+        (0x991, "X"),
+        (0x993, "V"),
+        (0x9A9, "X"),
+        (0x9AA, "V"),
+        (0x9B1, "X"),
+        (0x9B2, "V"),
+        (0x9B3, "X"),
+        (0x9B6, "V"),
+        (0x9BA, "X"),
+        (0x9BC, "V"),
+        (0x9C5, "X"),
+        (0x9C7, "V"),
+        (0x9C9, "X"),
+        (0x9CB, "V"),
+        (0x9CF, "X"),
+        (0x9D7, "V"),
+        (0x9D8, "X"),
+        (0x9DC, "M", "ড়"),
+        (0x9DD, "M", "ঢ়"),
+        (0x9DE, "X"),
+        (0x9DF, "M", "য়"),
+        (0x9E0, "V"),
+        (0x9E4, "X"),
+        (0x9E6, "V"),
+        (0x9FF, "X"),
+        (0xA01, "V"),
+        (0xA04, "X"),
+        (0xA05, "V"),
+        (0xA0B, "X"),
+        (0xA0F, "V"),
+        (0xA11, "X"),
+        (0xA13, "V"),
+        (0xA29, "X"),
+        (0xA2A, "V"),
+        (0xA31, "X"),
+        (0xA32, "V"),
+        (0xA33, "M", "ਲ਼"),
+        (0xA34, "X"),
+        (0xA35, "V"),
+        (0xA36, "M", "ਸ਼"),
+        (0xA37, "X"),
+        (0xA38, "V"),
+        (0xA3A, "X"),
+        (0xA3C, "V"),
+        (0xA3D, "X"),
+        (0xA3E, "V"),
+        (0xA43, "X"),
+        (0xA47, "V"),
+        (0xA49, "X"),
+        (0xA4B, "V"),
+        (0xA4E, "X"),
+        (0xA51, "V"),
+        (0xA52, "X"),
+        (0xA59, "M", "ਖ਼"),
+        (0xA5A, "M", "ਗ਼"),
+        (0xA5B, "M", "ਜ਼"),
+        (0xA5C, "V"),
+        (0xA5D, "X"),
+    ]
+
+
+def _seg_11() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA5E, "M", "ਫ਼"),
+        (0xA5F, "X"),
+        (0xA66, "V"),
+        (0xA77, "X"),
+        (0xA81, "V"),
+        (0xA84, "X"),
+        (0xA85, "V"),
+        (0xA8E, "X"),
+        (0xA8F, "V"),
+        (0xA92, "X"),
+        (0xA93, "V"),
+        (0xAA9, "X"),
+        (0xAAA, "V"),
+        (0xAB1, "X"),
+        (0xAB2, "V"),
+        (0xAB4, "X"),
+        (0xAB5, "V"),
+        (0xABA, "X"),
+        (0xABC, "V"),
+        (0xAC6, "X"),
+        (0xAC7, "V"),
+        (0xACA, "X"),
+        (0xACB, "V"),
+        (0xACE, "X"),
+        (0xAD0, "V"),
+        (0xAD1, "X"),
+        (0xAE0, "V"),
+        (0xAE4, "X"),
+        (0xAE6, "V"),
+        (0xAF2, "X"),
+        (0xAF9, "V"),
+        (0xB00, "X"),
+        (0xB01, "V"),
+        (0xB04, "X"),
+        (0xB05, "V"),
+        (0xB0D, "X"),
+        (0xB0F, "V"),
+        (0xB11, "X"),
+        (0xB13, "V"),
+        (0xB29, "X"),
+        (0xB2A, "V"),
+        (0xB31, "X"),
+        (0xB32, "V"),
+        (0xB34, "X"),
+        (0xB35, "V"),
+        (0xB3A, "X"),
+        (0xB3C, "V"),
+        (0xB45, "X"),
+        (0xB47, "V"),
+        (0xB49, "X"),
+        (0xB4B, "V"),
+        (0xB4E, "X"),
+        (0xB55, "V"),
+        (0xB58, "X"),
+        (0xB5C, "M", "ଡ଼"),
+        (0xB5D, "M", "ଢ଼"),
+        (0xB5E, "X"),
+        (0xB5F, "V"),
+        (0xB64, "X"),
+        (0xB66, "V"),
+        (0xB78, "X"),
+        (0xB82, "V"),
+        (0xB84, "X"),
+        (0xB85, "V"),
+        (0xB8B, "X"),
+        (0xB8E, "V"),
+        (0xB91, "X"),
+        (0xB92, "V"),
+        (0xB96, "X"),
+        (0xB99, "V"),
+        (0xB9B, "X"),
+        (0xB9C, "V"),
+        (0xB9D, "X"),
+        (0xB9E, "V"),
+        (0xBA0, "X"),
+        (0xBA3, "V"),
+        (0xBA5, "X"),
+        (0xBA8, "V"),
+        (0xBAB, "X"),
+        (0xBAE, "V"),
+        (0xBBA, "X"),
+        (0xBBE, "V"),
+        (0xBC3, "X"),
+        (0xBC6, "V"),
+        (0xBC9, "X"),
+        (0xBCA, "V"),
+        (0xBCE, "X"),
+        (0xBD0, "V"),
+        (0xBD1, "X"),
+        (0xBD7, "V"),
+        (0xBD8, "X"),
+        (0xBE6, "V"),
+        (0xBFB, "X"),
+        (0xC00, "V"),
+        (0xC0D, "X"),
+        (0xC0E, "V"),
+        (0xC11, "X"),
+        (0xC12, "V"),
+        (0xC29, "X"),
+        (0xC2A, "V"),
+    ]
+
+
+def _seg_12() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xC3A, "X"),
+        (0xC3C, "V"),
+        (0xC45, "X"),
+        (0xC46, "V"),
+        (0xC49, "X"),
+        (0xC4A, "V"),
+        (0xC4E, "X"),
+        (0xC55, "V"),
+        (0xC57, "X"),
+        (0xC58, "V"),
+        (0xC5B, "X"),
+        (0xC5D, "V"),
+        (0xC5E, "X"),
+        (0xC60, "V"),
+        (0xC64, "X"),
+        (0xC66, "V"),
+        (0xC70, "X"),
+        (0xC77, "V"),
+        (0xC8D, "X"),
+        (0xC8E, "V"),
+        (0xC91, "X"),
+        (0xC92, "V"),
+        (0xCA9, "X"),
+        (0xCAA, "V"),
+        (0xCB4, "X"),
+        (0xCB5, "V"),
+        (0xCBA, "X"),
+        (0xCBC, "V"),
+        (0xCC5, "X"),
+        (0xCC6, "V"),
+        (0xCC9, "X"),
+        (0xCCA, "V"),
+        (0xCCE, "X"),
+        (0xCD5, "V"),
+        (0xCD7, "X"),
+        (0xCDD, "V"),
+        (0xCDF, "X"),
+        (0xCE0, "V"),
+        (0xCE4, "X"),
+        (0xCE6, "V"),
+        (0xCF0, "X"),
+        (0xCF1, "V"),
+        (0xCF4, "X"),
+        (0xD00, "V"),
+        (0xD0D, "X"),
+        (0xD0E, "V"),
+        (0xD11, "X"),
+        (0xD12, "V"),
+        (0xD45, "X"),
+        (0xD46, "V"),
+        (0xD49, "X"),
+        (0xD4A, "V"),
+        (0xD50, "X"),
+        (0xD54, "V"),
+        (0xD64, "X"),
+        (0xD66, "V"),
+        (0xD80, "X"),
+        (0xD81, "V"),
+        (0xD84, "X"),
+        (0xD85, "V"),
+        (0xD97, "X"),
+        (0xD9A, "V"),
+        (0xDB2, "X"),
+        (0xDB3, "V"),
+        (0xDBC, "X"),
+        (0xDBD, "V"),
+        (0xDBE, "X"),
+        (0xDC0, "V"),
+        (0xDC7, "X"),
+        (0xDCA, "V"),
+        (0xDCB, "X"),
+        (0xDCF, "V"),
+        (0xDD5, "X"),
+        (0xDD6, "V"),
+        (0xDD7, "X"),
+        (0xDD8, "V"),
+        (0xDE0, "X"),
+        (0xDE6, "V"),
+        (0xDF0, "X"),
+        (0xDF2, "V"),
+        (0xDF5, "X"),
+        (0xE01, "V"),
+        (0xE33, "M", "ํา"),
+        (0xE34, "V"),
+        (0xE3B, "X"),
+        (0xE3F, "V"),
+        (0xE5C, "X"),
+        (0xE81, "V"),
+        (0xE83, "X"),
+        (0xE84, "V"),
+        (0xE85, "X"),
+        (0xE86, "V"),
+        (0xE8B, "X"),
+        (0xE8C, "V"),
+        (0xEA4, "X"),
+        (0xEA5, "V"),
+        (0xEA6, "X"),
+        (0xEA7, "V"),
+        (0xEB3, "M", "ໍາ"),
+        (0xEB4, "V"),
+    ]
+
+
+def _seg_13() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xEBE, "X"),
+        (0xEC0, "V"),
+        (0xEC5, "X"),
+        (0xEC6, "V"),
+        (0xEC7, "X"),
+        (0xEC8, "V"),
+        (0xECF, "X"),
+        (0xED0, "V"),
+        (0xEDA, "X"),
+        (0xEDC, "M", "ຫນ"),
+        (0xEDD, "M", "ຫມ"),
+        (0xEDE, "V"),
+        (0xEE0, "X"),
+        (0xF00, "V"),
+        (0xF0C, "M", "་"),
+        (0xF0D, "V"),
+        (0xF43, "M", "གྷ"),
+        (0xF44, "V"),
+        (0xF48, "X"),
+        (0xF49, "V"),
+        (0xF4D, "M", "ཌྷ"),
+        (0xF4E, "V"),
+        (0xF52, "M", "དྷ"),
+        (0xF53, "V"),
+        (0xF57, "M", "བྷ"),
+        (0xF58, "V"),
+        (0xF5C, "M", "ཛྷ"),
+        (0xF5D, "V"),
+        (0xF69, "M", "ཀྵ"),
+        (0xF6A, "V"),
+        (0xF6D, "X"),
+        (0xF71, "V"),
+        (0xF73, "M", "ཱི"),
+        (0xF74, "V"),
+        (0xF75, "M", "ཱུ"),
+        (0xF76, "M", "ྲྀ"),
+        (0xF77, "M", "ྲཱྀ"),
+        (0xF78, "M", "ླྀ"),
+        (0xF79, "M", "ླཱྀ"),
+        (0xF7A, "V"),
+        (0xF81, "M", "ཱྀ"),
+        (0xF82, "V"),
+        (0xF93, "M", "ྒྷ"),
+        (0xF94, "V"),
+        (0xF98, "X"),
+        (0xF99, "V"),
+        (0xF9D, "M", "ྜྷ"),
+        (0xF9E, "V"),
+        (0xFA2, "M", "ྡྷ"),
+        (0xFA3, "V"),
+        (0xFA7, "M", "ྦྷ"),
+        (0xFA8, "V"),
+        (0xFAC, "M", "ྫྷ"),
+        (0xFAD, "V"),
+        (0xFB9, "M", "ྐྵ"),
+        (0xFBA, "V"),
+        (0xFBD, "X"),
+        (0xFBE, "V"),
+        (0xFCD, "X"),
+        (0xFCE, "V"),
+        (0xFDB, "X"),
+        (0x1000, "V"),
+        (0x10A0, "X"),
+        (0x10C7, "M", "ⴧ"),
+        (0x10C8, "X"),
+        (0x10CD, "M", "ⴭ"),
+        (0x10CE, "X"),
+        (0x10D0, "V"),
+        (0x10FC, "M", "ნ"),
+        (0x10FD, "V"),
+        (0x115F, "X"),
+        (0x1161, "V"),
+        (0x1249, "X"),
+        (0x124A, "V"),
+        (0x124E, "X"),
+        (0x1250, "V"),
+        (0x1257, "X"),
+        (0x1258, "V"),
+        (0x1259, "X"),
+        (0x125A, "V"),
+        (0x125E, "X"),
+        (0x1260, "V"),
+        (0x1289, "X"),
+        (0x128A, "V"),
+        (0x128E, "X"),
+        (0x1290, "V"),
+        (0x12B1, "X"),
+        (0x12B2, "V"),
+        (0x12B6, "X"),
+        (0x12B8, "V"),
+        (0x12BF, "X"),
+        (0x12C0, "V"),
+        (0x12C1, "X"),
+        (0x12C2, "V"),
+        (0x12C6, "X"),
+        (0x12C8, "V"),
+        (0x12D7, "X"),
+        (0x12D8, "V"),
+        (0x1311, "X"),
+        (0x1312, "V"),
+    ]
+
+
+def _seg_14() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1316, "X"),
+        (0x1318, "V"),
+        (0x135B, "X"),
+        (0x135D, "V"),
+        (0x137D, "X"),
+        (0x1380, "V"),
+        (0x139A, "X"),
+        (0x13A0, "V"),
+        (0x13F6, "X"),
+        (0x13F8, "M", "Ᏸ"),
+        (0x13F9, "M", "Ᏹ"),
+        (0x13FA, "M", "Ᏺ"),
+        (0x13FB, "M", "Ᏻ"),
+        (0x13FC, "M", "Ᏼ"),
+        (0x13FD, "M", "Ᏽ"),
+        (0x13FE, "X"),
+        (0x1400, "V"),
+        (0x1680, "X"),
+        (0x1681, "V"),
+        (0x169D, "X"),
+        (0x16A0, "V"),
+        (0x16F9, "X"),
+        (0x1700, "V"),
+        (0x1716, "X"),
+        (0x171F, "V"),
+        (0x1737, "X"),
+        (0x1740, "V"),
+        (0x1754, "X"),
+        (0x1760, "V"),
+        (0x176D, "X"),
+        (0x176E, "V"),
+        (0x1771, "X"),
+        (0x1772, "V"),
+        (0x1774, "X"),
+        (0x1780, "V"),
+        (0x17B4, "X"),
+        (0x17B6, "V"),
+        (0x17DE, "X"),
+        (0x17E0, "V"),
+        (0x17EA, "X"),
+        (0x17F0, "V"),
+        (0x17FA, "X"),
+        (0x1800, "V"),
+        (0x1806, "X"),
+        (0x1807, "V"),
+        (0x180B, "I"),
+        (0x180E, "X"),
+        (0x180F, "I"),
+        (0x1810, "V"),
+        (0x181A, "X"),
+        (0x1820, "V"),
+        (0x1879, "X"),
+        (0x1880, "V"),
+        (0x18AB, "X"),
+        (0x18B0, "V"),
+        (0x18F6, "X"),
+        (0x1900, "V"),
+        (0x191F, "X"),
+        (0x1920, "V"),
+        (0x192C, "X"),
+        (0x1930, "V"),
+        (0x193C, "X"),
+        (0x1940, "V"),
+        (0x1941, "X"),
+        (0x1944, "V"),
+        (0x196E, "X"),
+        (0x1970, "V"),
+        (0x1975, "X"),
+        (0x1980, "V"),
+        (0x19AC, "X"),
+        (0x19B0, "V"),
+        (0x19CA, "X"),
+        (0x19D0, "V"),
+        (0x19DB, "X"),
+        (0x19DE, "V"),
+        (0x1A1C, "X"),
+        (0x1A1E, "V"),
+        (0x1A5F, "X"),
+        (0x1A60, "V"),
+        (0x1A7D, "X"),
+        (0x1A7F, "V"),
+        (0x1A8A, "X"),
+        (0x1A90, "V"),
+        (0x1A9A, "X"),
+        (0x1AA0, "V"),
+        (0x1AAE, "X"),
+        (0x1AB0, "V"),
+        (0x1ACF, "X"),
+        (0x1B00, "V"),
+        (0x1B4D, "X"),
+        (0x1B50, "V"),
+        (0x1B7F, "X"),
+        (0x1B80, "V"),
+        (0x1BF4, "X"),
+        (0x1BFC, "V"),
+        (0x1C38, "X"),
+        (0x1C3B, "V"),
+        (0x1C4A, "X"),
+        (0x1C4D, "V"),
+        (0x1C80, "M", "в"),
+    ]
+
+
+def _seg_15() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1C81, "M", "д"),
+        (0x1C82, "M", "о"),
+        (0x1C83, "M", "с"),
+        (0x1C84, "M", "т"),
+        (0x1C86, "M", "ъ"),
+        (0x1C87, "M", "ѣ"),
+        (0x1C88, "M", "ꙋ"),
+        (0x1C89, "X"),
+        (0x1C90, "M", "ა"),
+        (0x1C91, "M", "ბ"),
+        (0x1C92, "M", "გ"),
+        (0x1C93, "M", "დ"),
+        (0x1C94, "M", "ე"),
+        (0x1C95, "M", "ვ"),
+        (0x1C96, "M", "ზ"),
+        (0x1C97, "M", "თ"),
+        (0x1C98, "M", "ი"),
+        (0x1C99, "M", "კ"),
+        (0x1C9A, "M", "ლ"),
+        (0x1C9B, "M", "მ"),
+        (0x1C9C, "M", "ნ"),
+        (0x1C9D, "M", "ო"),
+        (0x1C9E, "M", "პ"),
+        (0x1C9F, "M", "ჟ"),
+        (0x1CA0, "M", "რ"),
+        (0x1CA1, "M", "ს"),
+        (0x1CA2, "M", "ტ"),
+        (0x1CA3, "M", "უ"),
+        (0x1CA4, "M", "ფ"),
+        (0x1CA5, "M", "ქ"),
+        (0x1CA6, "M", "ღ"),
+        (0x1CA7, "M", "ყ"),
+        (0x1CA8, "M", "შ"),
+        (0x1CA9, "M", "ჩ"),
+        (0x1CAA, "M", "ც"),
+        (0x1CAB, "M", "ძ"),
+        (0x1CAC, "M", "წ"),
+        (0x1CAD, "M", "ჭ"),
+        (0x1CAE, "M", "ხ"),
+        (0x1CAF, "M", "ჯ"),
+        (0x1CB0, "M", "ჰ"),
+        (0x1CB1, "M", "ჱ"),
+        (0x1CB2, "M", "ჲ"),
+        (0x1CB3, "M", "ჳ"),
+        (0x1CB4, "M", "ჴ"),
+        (0x1CB5, "M", "ჵ"),
+        (0x1CB6, "M", "ჶ"),
+        (0x1CB7, "M", "ჷ"),
+        (0x1CB8, "M", "ჸ"),
+        (0x1CB9, "M", "ჹ"),
+        (0x1CBA, "M", "ჺ"),
+        (0x1CBB, "X"),
+        (0x1CBD, "M", "ჽ"),
+        (0x1CBE, "M", "ჾ"),
+        (0x1CBF, "M", "ჿ"),
+        (0x1CC0, "V"),
+        (0x1CC8, "X"),
+        (0x1CD0, "V"),
+        (0x1CFB, "X"),
+        (0x1D00, "V"),
+        (0x1D2C, "M", "a"),
+        (0x1D2D, "M", "æ"),
+        (0x1D2E, "M", "b"),
+        (0x1D2F, "V"),
+        (0x1D30, "M", "d"),
+        (0x1D31, "M", "e"),
+        (0x1D32, "M", "ǝ"),
+        (0x1D33, "M", "g"),
+        (0x1D34, "M", "h"),
+        (0x1D35, "M", "i"),
+        (0x1D36, "M", "j"),
+        (0x1D37, "M", "k"),
+        (0x1D38, "M", "l"),
+        (0x1D39, "M", "m"),
+        (0x1D3A, "M", "n"),
+        (0x1D3B, "V"),
+        (0x1D3C, "M", "o"),
+        (0x1D3D, "M", "ȣ"),
+        (0x1D3E, "M", "p"),
+        (0x1D3F, "M", "r"),
+        (0x1D40, "M", "t"),
+        (0x1D41, "M", "u"),
+        (0x1D42, "M", "w"),
+        (0x1D43, "M", "a"),
+        (0x1D44, "M", "ɐ"),
+        (0x1D45, "M", "ɑ"),
+        (0x1D46, "M", "ᴂ"),
+        (0x1D47, "M", "b"),
+        (0x1D48, "M", "d"),
+        (0x1D49, "M", "e"),
+        (0x1D4A, "M", "ə"),
+        (0x1D4B, "M", "ɛ"),
+        (0x1D4C, "M", "ɜ"),
+        (0x1D4D, "M", "g"),
+        (0x1D4E, "V"),
+        (0x1D4F, "M", "k"),
+        (0x1D50, "M", "m"),
+        (0x1D51, "M", "ŋ"),
+        (0x1D52, "M", "o"),
+        (0x1D53, "M", "ɔ"),
+    ]
+
+
+def _seg_16() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D54, "M", "ᴖ"),
+        (0x1D55, "M", "ᴗ"),
+        (0x1D56, "M", "p"),
+        (0x1D57, "M", "t"),
+        (0x1D58, "M", "u"),
+        (0x1D59, "M", "ᴝ"),
+        (0x1D5A, "M", "ɯ"),
+        (0x1D5B, "M", "v"),
+        (0x1D5C, "M", "ᴥ"),
+        (0x1D5D, "M", "β"),
+        (0x1D5E, "M", "γ"),
+        (0x1D5F, "M", "δ"),
+        (0x1D60, "M", "φ"),
+        (0x1D61, "M", "χ"),
+        (0x1D62, "M", "i"),
+        (0x1D63, "M", "r"),
+        (0x1D64, "M", "u"),
+        (0x1D65, "M", "v"),
+        (0x1D66, "M", "β"),
+        (0x1D67, "M", "γ"),
+        (0x1D68, "M", "ρ"),
+        (0x1D69, "M", "φ"),
+        (0x1D6A, "M", "χ"),
+        (0x1D6B, "V"),
+        (0x1D78, "M", "н"),
+        (0x1D79, "V"),
+        (0x1D9B, "M", "ɒ"),
+        (0x1D9C, "M", "c"),
+        (0x1D9D, "M", "ɕ"),
+        (0x1D9E, "M", "ð"),
+        (0x1D9F, "M", "ɜ"),
+        (0x1DA0, "M", "f"),
+        (0x1DA1, "M", "ɟ"),
+        (0x1DA2, "M", "ɡ"),
+        (0x1DA3, "M", "ɥ"),
+        (0x1DA4, "M", "ɨ"),
+        (0x1DA5, "M", "ɩ"),
+        (0x1DA6, "M", "ɪ"),
+        (0x1DA7, "M", "ᵻ"),
+        (0x1DA8, "M", "ʝ"),
+        (0x1DA9, "M", "ɭ"),
+        (0x1DAA, "M", "ᶅ"),
+        (0x1DAB, "M", "ʟ"),
+        (0x1DAC, "M", "ɱ"),
+        (0x1DAD, "M", "ɰ"),
+        (0x1DAE, "M", "ɲ"),
+        (0x1DAF, "M", "ɳ"),
+        (0x1DB0, "M", "ɴ"),
+        (0x1DB1, "M", "ɵ"),
+        (0x1DB2, "M", "ɸ"),
+        (0x1DB3, "M", "ʂ"),
+        (0x1DB4, "M", "ʃ"),
+        (0x1DB5, "M", "ƫ"),
+        (0x1DB6, "M", "ʉ"),
+        (0x1DB7, "M", "ʊ"),
+        (0x1DB8, "M", "ᴜ"),
+        (0x1DB9, "M", "ʋ"),
+        (0x1DBA, "M", "ʌ"),
+        (0x1DBB, "M", "z"),
+        (0x1DBC, "M", "ʐ"),
+        (0x1DBD, "M", "ʑ"),
+        (0x1DBE, "M", "ʒ"),
+        (0x1DBF, "M", "θ"),
+        (0x1DC0, "V"),
+        (0x1E00, "M", "ḁ"),
+        (0x1E01, "V"),
+        (0x1E02, "M", "ḃ"),
+        (0x1E03, "V"),
+        (0x1E04, "M", "ḅ"),
+        (0x1E05, "V"),
+        (0x1E06, "M", "ḇ"),
+        (0x1E07, "V"),
+        (0x1E08, "M", "ḉ"),
+        (0x1E09, "V"),
+        (0x1E0A, "M", "ḋ"),
+        (0x1E0B, "V"),
+        (0x1E0C, "M", "ḍ"),
+        (0x1E0D, "V"),
+        (0x1E0E, "M", "ḏ"),
+        (0x1E0F, "V"),
+        (0x1E10, "M", "ḑ"),
+        (0x1E11, "V"),
+        (0x1E12, "M", "ḓ"),
+        (0x1E13, "V"),
+        (0x1E14, "M", "ḕ"),
+        (0x1E15, "V"),
+        (0x1E16, "M", "ḗ"),
+        (0x1E17, "V"),
+        (0x1E18, "M", "ḙ"),
+        (0x1E19, "V"),
+        (0x1E1A, "M", "ḛ"),
+        (0x1E1B, "V"),
+        (0x1E1C, "M", "ḝ"),
+        (0x1E1D, "V"),
+        (0x1E1E, "M", "ḟ"),
+        (0x1E1F, "V"),
+        (0x1E20, "M", "ḡ"),
+        (0x1E21, "V"),
+        (0x1E22, "M", "ḣ"),
+        (0x1E23, "V"),
+    ]
+
+
+def _seg_17() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E24, "M", "ḥ"),
+        (0x1E25, "V"),
+        (0x1E26, "M", "ḧ"),
+        (0x1E27, "V"),
+        (0x1E28, "M", "ḩ"),
+        (0x1E29, "V"),
+        (0x1E2A, "M", "ḫ"),
+        (0x1E2B, "V"),
+        (0x1E2C, "M", "ḭ"),
+        (0x1E2D, "V"),
+        (0x1E2E, "M", "ḯ"),
+        (0x1E2F, "V"),
+        (0x1E30, "M", "ḱ"),
+        (0x1E31, "V"),
+        (0x1E32, "M", "ḳ"),
+        (0x1E33, "V"),
+        (0x1E34, "M", "ḵ"),
+        (0x1E35, "V"),
+        (0x1E36, "M", "ḷ"),
+        (0x1E37, "V"),
+        (0x1E38, "M", "ḹ"),
+        (0x1E39, "V"),
+        (0x1E3A, "M", "ḻ"),
+        (0x1E3B, "V"),
+        (0x1E3C, "M", "ḽ"),
+        (0x1E3D, "V"),
+        (0x1E3E, "M", "ḿ"),
+        (0x1E3F, "V"),
+        (0x1E40, "M", "ṁ"),
+        (0x1E41, "V"),
+        (0x1E42, "M", "ṃ"),
+        (0x1E43, "V"),
+        (0x1E44, "M", "ṅ"),
+        (0x1E45, "V"),
+        (0x1E46, "M", "ṇ"),
+        (0x1E47, "V"),
+        (0x1E48, "M", "ṉ"),
+        (0x1E49, "V"),
+        (0x1E4A, "M", "ṋ"),
+        (0x1E4B, "V"),
+        (0x1E4C, "M", "ṍ"),
+        (0x1E4D, "V"),
+        (0x1E4E, "M", "ṏ"),
+        (0x1E4F, "V"),
+        (0x1E50, "M", "ṑ"),
+        (0x1E51, "V"),
+        (0x1E52, "M", "ṓ"),
+        (0x1E53, "V"),
+        (0x1E54, "M", "ṕ"),
+        (0x1E55, "V"),
+        (0x1E56, "M", "ṗ"),
+        (0x1E57, "V"),
+        (0x1E58, "M", "ṙ"),
+        (0x1E59, "V"),
+        (0x1E5A, "M", "ṛ"),
+        (0x1E5B, "V"),
+        (0x1E5C, "M", "ṝ"),
+        (0x1E5D, "V"),
+        (0x1E5E, "M", "ṟ"),
+        (0x1E5F, "V"),
+        (0x1E60, "M", "ṡ"),
+        (0x1E61, "V"),
+        (0x1E62, "M", "ṣ"),
+        (0x1E63, "V"),
+        (0x1E64, "M", "ṥ"),
+        (0x1E65, "V"),
+        (0x1E66, "M", "ṧ"),
+        (0x1E67, "V"),
+        (0x1E68, "M", "ṩ"),
+        (0x1E69, "V"),
+        (0x1E6A, "M", "ṫ"),
+        (0x1E6B, "V"),
+        (0x1E6C, "M", "ṭ"),
+        (0x1E6D, "V"),
+        (0x1E6E, "M", "ṯ"),
+        (0x1E6F, "V"),
+        (0x1E70, "M", "ṱ"),
+        (0x1E71, "V"),
+        (0x1E72, "M", "ṳ"),
+        (0x1E73, "V"),
+        (0x1E74, "M", "ṵ"),
+        (0x1E75, "V"),
+        (0x1E76, "M", "ṷ"),
+        (0x1E77, "V"),
+        (0x1E78, "M", "ṹ"),
+        (0x1E79, "V"),
+        (0x1E7A, "M", "ṻ"),
+        (0x1E7B, "V"),
+        (0x1E7C, "M", "ṽ"),
+        (0x1E7D, "V"),
+        (0x1E7E, "M", "ṿ"),
+        (0x1E7F, "V"),
+        (0x1E80, "M", "ẁ"),
+        (0x1E81, "V"),
+        (0x1E82, "M", "ẃ"),
+        (0x1E83, "V"),
+        (0x1E84, "M", "ẅ"),
+        (0x1E85, "V"),
+        (0x1E86, "M", "ẇ"),
+        (0x1E87, "V"),
+    ]
+
+
+def _seg_18() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E88, "M", "ẉ"),
+        (0x1E89, "V"),
+        (0x1E8A, "M", "ẋ"),
+        (0x1E8B, "V"),
+        (0x1E8C, "M", "ẍ"),
+        (0x1E8D, "V"),
+        (0x1E8E, "M", "ẏ"),
+        (0x1E8F, "V"),
+        (0x1E90, "M", "ẑ"),
+        (0x1E91, "V"),
+        (0x1E92, "M", "ẓ"),
+        (0x1E93, "V"),
+        (0x1E94, "M", "ẕ"),
+        (0x1E95, "V"),
+        (0x1E9A, "M", "aʾ"),
+        (0x1E9B, "M", "ṡ"),
+        (0x1E9C, "V"),
+        (0x1E9E, "M", "ß"),
+        (0x1E9F, "V"),
+        (0x1EA0, "M", "ạ"),
+        (0x1EA1, "V"),
+        (0x1EA2, "M", "ả"),
+        (0x1EA3, "V"),
+        (0x1EA4, "M", "ấ"),
+        (0x1EA5, "V"),
+        (0x1EA6, "M", "ầ"),
+        (0x1EA7, "V"),
+        (0x1EA8, "M", "ẩ"),
+        (0x1EA9, "V"),
+        (0x1EAA, "M", "ẫ"),
+        (0x1EAB, "V"),
+        (0x1EAC, "M", "ậ"),
+        (0x1EAD, "V"),
+        (0x1EAE, "M", "ắ"),
+        (0x1EAF, "V"),
+        (0x1EB0, "M", "ằ"),
+        (0x1EB1, "V"),
+        (0x1EB2, "M", "ẳ"),
+        (0x1EB3, "V"),
+        (0x1EB4, "M", "ẵ"),
+        (0x1EB5, "V"),
+        (0x1EB6, "M", "ặ"),
+        (0x1EB7, "V"),
+        (0x1EB8, "M", "ẹ"),
+        (0x1EB9, "V"),
+        (0x1EBA, "M", "ẻ"),
+        (0x1EBB, "V"),
+        (0x1EBC, "M", "ẽ"),
+        (0x1EBD, "V"),
+        (0x1EBE, "M", "ế"),
+        (0x1EBF, "V"),
+        (0x1EC0, "M", "ề"),
+        (0x1EC1, "V"),
+        (0x1EC2, "M", "ể"),
+        (0x1EC3, "V"),
+        (0x1EC4, "M", "ễ"),
+        (0x1EC5, "V"),
+        (0x1EC6, "M", "ệ"),
+        (0x1EC7, "V"),
+        (0x1EC8, "M", "ỉ"),
+        (0x1EC9, "V"),
+        (0x1ECA, "M", "ị"),
+        (0x1ECB, "V"),
+        (0x1ECC, "M", "ọ"),
+        (0x1ECD, "V"),
+        (0x1ECE, "M", "ỏ"),
+        (0x1ECF, "V"),
+        (0x1ED0, "M", "ố"),
+        (0x1ED1, "V"),
+        (0x1ED2, "M", "ồ"),
+        (0x1ED3, "V"),
+        (0x1ED4, "M", "ổ"),
+        (0x1ED5, "V"),
+        (0x1ED6, "M", "ỗ"),
+        (0x1ED7, "V"),
+        (0x1ED8, "M", "ộ"),
+        (0x1ED9, "V"),
+        (0x1EDA, "M", "ớ"),
+        (0x1EDB, "V"),
+        (0x1EDC, "M", "ờ"),
+        (0x1EDD, "V"),
+        (0x1EDE, "M", "ở"),
+        (0x1EDF, "V"),
+        (0x1EE0, "M", "ỡ"),
+        (0x1EE1, "V"),
+        (0x1EE2, "M", "ợ"),
+        (0x1EE3, "V"),
+        (0x1EE4, "M", "ụ"),
+        (0x1EE5, "V"),
+        (0x1EE6, "M", "ủ"),
+        (0x1EE7, "V"),
+        (0x1EE8, "M", "ứ"),
+        (0x1EE9, "V"),
+        (0x1EEA, "M", "ừ"),
+        (0x1EEB, "V"),
+        (0x1EEC, "M", "ử"),
+        (0x1EED, "V"),
+        (0x1EEE, "M", "ữ"),
+        (0x1EEF, "V"),
+        (0x1EF0, "M", "ự"),
+    ]
+
+
+def _seg_19() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EF1, "V"),
+        (0x1EF2, "M", "ỳ"),
+        (0x1EF3, "V"),
+        (0x1EF4, "M", "ỵ"),
+        (0x1EF5, "V"),
+        (0x1EF6, "M", "ỷ"),
+        (0x1EF7, "V"),
+        (0x1EF8, "M", "ỹ"),
+        (0x1EF9, "V"),
+        (0x1EFA, "M", "ỻ"),
+        (0x1EFB, "V"),
+        (0x1EFC, "M", "ỽ"),
+        (0x1EFD, "V"),
+        (0x1EFE, "M", "ỿ"),
+        (0x1EFF, "V"),
+        (0x1F08, "M", "ἀ"),
+        (0x1F09, "M", "ἁ"),
+        (0x1F0A, "M", "ἂ"),
+        (0x1F0B, "M", "ἃ"),
+        (0x1F0C, "M", "ἄ"),
+        (0x1F0D, "M", "ἅ"),
+        (0x1F0E, "M", "ἆ"),
+        (0x1F0F, "M", "ἇ"),
+        (0x1F10, "V"),
+        (0x1F16, "X"),
+        (0x1F18, "M", "ἐ"),
+        (0x1F19, "M", "ἑ"),
+        (0x1F1A, "M", "ἒ"),
+        (0x1F1B, "M", "ἓ"),
+        (0x1F1C, "M", "ἔ"),
+        (0x1F1D, "M", "ἕ"),
+        (0x1F1E, "X"),
+        (0x1F20, "V"),
+        (0x1F28, "M", "ἠ"),
+        (0x1F29, "M", "ἡ"),
+        (0x1F2A, "M", "ἢ"),
+        (0x1F2B, "M", "ἣ"),
+        (0x1F2C, "M", "ἤ"),
+        (0x1F2D, "M", "ἥ"),
+        (0x1F2E, "M", "ἦ"),
+        (0x1F2F, "M", "ἧ"),
+        (0x1F30, "V"),
+        (0x1F38, "M", "ἰ"),
+        (0x1F39, "M", "ἱ"),
+        (0x1F3A, "M", "ἲ"),
+        (0x1F3B, "M", "ἳ"),
+        (0x1F3C, "M", "ἴ"),
+        (0x1F3D, "M", "ἵ"),
+        (0x1F3E, "M", "ἶ"),
+        (0x1F3F, "M", "ἷ"),
+        (0x1F40, "V"),
+        (0x1F46, "X"),
+        (0x1F48, "M", "ὀ"),
+        (0x1F49, "M", "ὁ"),
+        (0x1F4A, "M", "ὂ"),
+        (0x1F4B, "M", "ὃ"),
+        (0x1F4C, "M", "ὄ"),
+        (0x1F4D, "M", "ὅ"),
+        (0x1F4E, "X"),
+        (0x1F50, "V"),
+        (0x1F58, "X"),
+        (0x1F59, "M", "ὑ"),
+        (0x1F5A, "X"),
+        (0x1F5B, "M", "ὓ"),
+        (0x1F5C, "X"),
+        (0x1F5D, "M", "ὕ"),
+        (0x1F5E, "X"),
+        (0x1F5F, "M", "ὗ"),
+        (0x1F60, "V"),
+        (0x1F68, "M", "ὠ"),
+        (0x1F69, "M", "ὡ"),
+        (0x1F6A, "M", "ὢ"),
+        (0x1F6B, "M", "ὣ"),
+        (0x1F6C, "M", "ὤ"),
+        (0x1F6D, "M", "ὥ"),
+        (0x1F6E, "M", "ὦ"),
+        (0x1F6F, "M", "ὧ"),
+        (0x1F70, "V"),
+        (0x1F71, "M", "ά"),
+        (0x1F72, "V"),
+        (0x1F73, "M", "έ"),
+        (0x1F74, "V"),
+        (0x1F75, "M", "ή"),
+        (0x1F76, "V"),
+        (0x1F77, "M", "ί"),
+        (0x1F78, "V"),
+        (0x1F79, "M", "ό"),
+        (0x1F7A, "V"),
+        (0x1F7B, "M", "ύ"),
+        (0x1F7C, "V"),
+        (0x1F7D, "M", "ώ"),
+        (0x1F7E, "X"),
+        (0x1F80, "M", "ἀι"),
+        (0x1F81, "M", "ἁι"),
+        (0x1F82, "M", "ἂι"),
+        (0x1F83, "M", "ἃι"),
+        (0x1F84, "M", "ἄι"),
+        (0x1F85, "M", "ἅι"),
+        (0x1F86, "M", "ἆι"),
+        (0x1F87, "M", "ἇι"),
+    ]
+
+
+def _seg_20() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1F88, "M", "ἀι"),
+        (0x1F89, "M", "ἁι"),
+        (0x1F8A, "M", "ἂι"),
+        (0x1F8B, "M", "ἃι"),
+        (0x1F8C, "M", "ἄι"),
+        (0x1F8D, "M", "ἅι"),
+        (0x1F8E, "M", "ἆι"),
+        (0x1F8F, "M", "ἇι"),
+        (0x1F90, "M", "ἠι"),
+        (0x1F91, "M", "ἡι"),
+        (0x1F92, "M", "ἢι"),
+        (0x1F93, "M", "ἣι"),
+        (0x1F94, "M", "ἤι"),
+        (0x1F95, "M", "ἥι"),
+        (0x1F96, "M", "ἦι"),
+        (0x1F97, "M", "ἧι"),
+        (0x1F98, "M", "ἠι"),
+        (0x1F99, "M", "ἡι"),
+        (0x1F9A, "M", "ἢι"),
+        (0x1F9B, "M", "ἣι"),
+        (0x1F9C, "M", "ἤι"),
+        (0x1F9D, "M", "ἥι"),
+        (0x1F9E, "M", "ἦι"),
+        (0x1F9F, "M", "ἧι"),
+        (0x1FA0, "M", "ὠι"),
+        (0x1FA1, "M", "ὡι"),
+        (0x1FA2, "M", "ὢι"),
+        (0x1FA3, "M", "ὣι"),
+        (0x1FA4, "M", "ὤι"),
+        (0x1FA5, "M", "ὥι"),
+        (0x1FA6, "M", "ὦι"),
+        (0x1FA7, "M", "ὧι"),
+        (0x1FA8, "M", "ὠι"),
+        (0x1FA9, "M", "ὡι"),
+        (0x1FAA, "M", "ὢι"),
+        (0x1FAB, "M", "ὣι"),
+        (0x1FAC, "M", "ὤι"),
+        (0x1FAD, "M", "ὥι"),
+        (0x1FAE, "M", "ὦι"),
+        (0x1FAF, "M", "ὧι"),
+        (0x1FB0, "V"),
+        (0x1FB2, "M", "ὰι"),
+        (0x1FB3, "M", "αι"),
+        (0x1FB4, "M", "άι"),
+        (0x1FB5, "X"),
+        (0x1FB6, "V"),
+        (0x1FB7, "M", "ᾶι"),
+        (0x1FB8, "M", "ᾰ"),
+        (0x1FB9, "M", "ᾱ"),
+        (0x1FBA, "M", "ὰ"),
+        (0x1FBB, "M", "ά"),
+        (0x1FBC, "M", "αι"),
+        (0x1FBD, "3", " ̓"),
+        (0x1FBE, "M", "ι"),
+        (0x1FBF, "3", " ̓"),
+        (0x1FC0, "3", " ͂"),
+        (0x1FC1, "3", " ̈͂"),
+        (0x1FC2, "M", "ὴι"),
+        (0x1FC3, "M", "ηι"),
+        (0x1FC4, "M", "ήι"),
+        (0x1FC5, "X"),
+        (0x1FC6, "V"),
+        (0x1FC7, "M", "ῆι"),
+        (0x1FC8, "M", "ὲ"),
+        (0x1FC9, "M", "έ"),
+        (0x1FCA, "M", "ὴ"),
+        (0x1FCB, "M", "ή"),
+        (0x1FCC, "M", "ηι"),
+        (0x1FCD, "3", " ̓̀"),
+        (0x1FCE, "3", " ̓́"),
+        (0x1FCF, "3", " ̓͂"),
+        (0x1FD0, "V"),
+        (0x1FD3, "M", "ΐ"),
+        (0x1FD4, "X"),
+        (0x1FD6, "V"),
+        (0x1FD8, "M", "ῐ"),
+        (0x1FD9, "M", "ῑ"),
+        (0x1FDA, "M", "ὶ"),
+        (0x1FDB, "M", "ί"),
+        (0x1FDC, "X"),
+        (0x1FDD, "3", " ̔̀"),
+        (0x1FDE, "3", " ̔́"),
+        (0x1FDF, "3", " ̔͂"),
+        (0x1FE0, "V"),
+        (0x1FE3, "M", "ΰ"),
+        (0x1FE4, "V"),
+        (0x1FE8, "M", "ῠ"),
+        (0x1FE9, "M", "ῡ"),
+        (0x1FEA, "M", "ὺ"),
+        (0x1FEB, "M", "ύ"),
+        (0x1FEC, "M", "ῥ"),
+        (0x1FED, "3", " ̈̀"),
+        (0x1FEE, "3", " ̈́"),
+        (0x1FEF, "3", "`"),
+        (0x1FF0, "X"),
+        (0x1FF2, "M", "ὼι"),
+        (0x1FF3, "M", "ωι"),
+        (0x1FF4, "M", "ώι"),
+        (0x1FF5, "X"),
+        (0x1FF6, "V"),
+    ]
+
+
+def _seg_21() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1FF7, "M", "ῶι"),
+        (0x1FF8, "M", "ὸ"),
+        (0x1FF9, "M", "ό"),
+        (0x1FFA, "M", "ὼ"),
+        (0x1FFB, "M", "ώ"),
+        (0x1FFC, "M", "ωι"),
+        (0x1FFD, "3", " ́"),
+        (0x1FFE, "3", " ̔"),
+        (0x1FFF, "X"),
+        (0x2000, "3", " "),
+        (0x200B, "I"),
+        (0x200C, "D", ""),
+        (0x200E, "X"),
+        (0x2010, "V"),
+        (0x2011, "M", "‐"),
+        (0x2012, "V"),
+        (0x2017, "3", " ̳"),
+        (0x2018, "V"),
+        (0x2024, "X"),
+        (0x2027, "V"),
+        (0x2028, "X"),
+        (0x202F, "3", " "),
+        (0x2030, "V"),
+        (0x2033, "M", "′′"),
+        (0x2034, "M", "′′′"),
+        (0x2035, "V"),
+        (0x2036, "M", "‵‵"),
+        (0x2037, "M", "‵‵‵"),
+        (0x2038, "V"),
+        (0x203C, "3", "!!"),
+        (0x203D, "V"),
+        (0x203E, "3", " ̅"),
+        (0x203F, "V"),
+        (0x2047, "3", "??"),
+        (0x2048, "3", "?!"),
+        (0x2049, "3", "!?"),
+        (0x204A, "V"),
+        (0x2057, "M", "′′′′"),
+        (0x2058, "V"),
+        (0x205F, "3", " "),
+        (0x2060, "I"),
+        (0x2061, "X"),
+        (0x2064, "I"),
+        (0x2065, "X"),
+        (0x2070, "M", "0"),
+        (0x2071, "M", "i"),
+        (0x2072, "X"),
+        (0x2074, "M", "4"),
+        (0x2075, "M", "5"),
+        (0x2076, "M", "6"),
+        (0x2077, "M", "7"),
+        (0x2078, "M", "8"),
+        (0x2079, "M", "9"),
+        (0x207A, "3", "+"),
+        (0x207B, "M", "−"),
+        (0x207C, "3", "="),
+        (0x207D, "3", "("),
+        (0x207E, "3", ")"),
+        (0x207F, "M", "n"),
+        (0x2080, "M", "0"),
+        (0x2081, "M", "1"),
+        (0x2082, "M", "2"),
+        (0x2083, "M", "3"),
+        (0x2084, "M", "4"),
+        (0x2085, "M", "5"),
+        (0x2086, "M", "6"),
+        (0x2087, "M", "7"),
+        (0x2088, "M", "8"),
+        (0x2089, "M", "9"),
+        (0x208A, "3", "+"),
+        (0x208B, "M", "−"),
+        (0x208C, "3", "="),
+        (0x208D, "3", "("),
+        (0x208E, "3", ")"),
+        (0x208F, "X"),
+        (0x2090, "M", "a"),
+        (0x2091, "M", "e"),
+        (0x2092, "M", "o"),
+        (0x2093, "M", "x"),
+        (0x2094, "M", "ə"),
+        (0x2095, "M", "h"),
+        (0x2096, "M", "k"),
+        (0x2097, "M", "l"),
+        (0x2098, "M", "m"),
+        (0x2099, "M", "n"),
+        (0x209A, "M", "p"),
+        (0x209B, "M", "s"),
+        (0x209C, "M", "t"),
+        (0x209D, "X"),
+        (0x20A0, "V"),
+        (0x20A8, "M", "rs"),
+        (0x20A9, "V"),
+        (0x20C1, "X"),
+        (0x20D0, "V"),
+        (0x20F1, "X"),
+        (0x2100, "3", "a/c"),
+        (0x2101, "3", "a/s"),
+        (0x2102, "M", "c"),
+        (0x2103, "M", "°c"),
+        (0x2104, "V"),
+    ]
+
+
+def _seg_22() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2105, "3", "c/o"),
+        (0x2106, "3", "c/u"),
+        (0x2107, "M", "ɛ"),
+        (0x2108, "V"),
+        (0x2109, "M", "°f"),
+        (0x210A, "M", "g"),
+        (0x210B, "M", "h"),
+        (0x210F, "M", "ħ"),
+        (0x2110, "M", "i"),
+        (0x2112, "M", "l"),
+        (0x2114, "V"),
+        (0x2115, "M", "n"),
+        (0x2116, "M", "no"),
+        (0x2117, "V"),
+        (0x2119, "M", "p"),
+        (0x211A, "M", "q"),
+        (0x211B, "M", "r"),
+        (0x211E, "V"),
+        (0x2120, "M", "sm"),
+        (0x2121, "M", "tel"),
+        (0x2122, "M", "tm"),
+        (0x2123, "V"),
+        (0x2124, "M", "z"),
+        (0x2125, "V"),
+        (0x2126, "M", "ω"),
+        (0x2127, "V"),
+        (0x2128, "M", "z"),
+        (0x2129, "V"),
+        (0x212A, "M", "k"),
+        (0x212B, "M", "å"),
+        (0x212C, "M", "b"),
+        (0x212D, "M", "c"),
+        (0x212E, "V"),
+        (0x212F, "M", "e"),
+        (0x2131, "M", "f"),
+        (0x2132, "X"),
+        (0x2133, "M", "m"),
+        (0x2134, "M", "o"),
+        (0x2135, "M", "א"),
+        (0x2136, "M", "ב"),
+        (0x2137, "M", "ג"),
+        (0x2138, "M", "ד"),
+        (0x2139, "M", "i"),
+        (0x213A, "V"),
+        (0x213B, "M", "fax"),
+        (0x213C, "M", "π"),
+        (0x213D, "M", "γ"),
+        (0x213F, "M", "π"),
+        (0x2140, "M", "∑"),
+        (0x2141, "V"),
+        (0x2145, "M", "d"),
+        (0x2147, "M", "e"),
+        (0x2148, "M", "i"),
+        (0x2149, "M", "j"),
+        (0x214A, "V"),
+        (0x2150, "M", "1⁄7"),
+        (0x2151, "M", "1⁄9"),
+        (0x2152, "M", "1⁄10"),
+        (0x2153, "M", "1⁄3"),
+        (0x2154, "M", "2⁄3"),
+        (0x2155, "M", "1⁄5"),
+        (0x2156, "M", "2⁄5"),
+        (0x2157, "M", "3⁄5"),
+        (0x2158, "M", "4⁄5"),
+        (0x2159, "M", "1⁄6"),
+        (0x215A, "M", "5⁄6"),
+        (0x215B, "M", "1⁄8"),
+        (0x215C, "M", "3⁄8"),
+        (0x215D, "M", "5⁄8"),
+        (0x215E, "M", "7⁄8"),
+        (0x215F, "M", "1⁄"),
+        (0x2160, "M", "i"),
+        (0x2161, "M", "ii"),
+        (0x2162, "M", "iii"),
+        (0x2163, "M", "iv"),
+        (0x2164, "M", "v"),
+        (0x2165, "M", "vi"),
+        (0x2166, "M", "vii"),
+        (0x2167, "M", "viii"),
+        (0x2168, "M", "ix"),
+        (0x2169, "M", "x"),
+        (0x216A, "M", "xi"),
+        (0x216B, "M", "xii"),
+        (0x216C, "M", "l"),
+        (0x216D, "M", "c"),
+        (0x216E, "M", "d"),
+        (0x216F, "M", "m"),
+        (0x2170, "M", "i"),
+        (0x2171, "M", "ii"),
+        (0x2172, "M", "iii"),
+        (0x2173, "M", "iv"),
+        (0x2174, "M", "v"),
+        (0x2175, "M", "vi"),
+        (0x2176, "M", "vii"),
+        (0x2177, "M", "viii"),
+        (0x2178, "M", "ix"),
+        (0x2179, "M", "x"),
+        (0x217A, "M", "xi"),
+        (0x217B, "M", "xii"),
+        (0x217C, "M", "l"),
+    ]
+
+
+def _seg_23() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x217D, "M", "c"),
+        (0x217E, "M", "d"),
+        (0x217F, "M", "m"),
+        (0x2180, "V"),
+        (0x2183, "X"),
+        (0x2184, "V"),
+        (0x2189, "M", "0⁄3"),
+        (0x218A, "V"),
+        (0x218C, "X"),
+        (0x2190, "V"),
+        (0x222C, "M", "∫∫"),
+        (0x222D, "M", "∫∫∫"),
+        (0x222E, "V"),
+        (0x222F, "M", "∮∮"),
+        (0x2230, "M", "∮∮∮"),
+        (0x2231, "V"),
+        (0x2329, "M", "〈"),
+        (0x232A, "M", "〉"),
+        (0x232B, "V"),
+        (0x2427, "X"),
+        (0x2440, "V"),
+        (0x244B, "X"),
+        (0x2460, "M", "1"),
+        (0x2461, "M", "2"),
+        (0x2462, "M", "3"),
+        (0x2463, "M", "4"),
+        (0x2464, "M", "5"),
+        (0x2465, "M", "6"),
+        (0x2466, "M", "7"),
+        (0x2467, "M", "8"),
+        (0x2468, "M", "9"),
+        (0x2469, "M", "10"),
+        (0x246A, "M", "11"),
+        (0x246B, "M", "12"),
+        (0x246C, "M", "13"),
+        (0x246D, "M", "14"),
+        (0x246E, "M", "15"),
+        (0x246F, "M", "16"),
+        (0x2470, "M", "17"),
+        (0x2471, "M", "18"),
+        (0x2472, "M", "19"),
+        (0x2473, "M", "20"),
+        (0x2474, "3", "(1)"),
+        (0x2475, "3", "(2)"),
+        (0x2476, "3", "(3)"),
+        (0x2477, "3", "(4)"),
+        (0x2478, "3", "(5)"),
+        (0x2479, "3", "(6)"),
+        (0x247A, "3", "(7)"),
+        (0x247B, "3", "(8)"),
+        (0x247C, "3", "(9)"),
+        (0x247D, "3", "(10)"),
+        (0x247E, "3", "(11)"),
+        (0x247F, "3", "(12)"),
+        (0x2480, "3", "(13)"),
+        (0x2481, "3", "(14)"),
+        (0x2482, "3", "(15)"),
+        (0x2483, "3", "(16)"),
+        (0x2484, "3", "(17)"),
+        (0x2485, "3", "(18)"),
+        (0x2486, "3", "(19)"),
+        (0x2487, "3", "(20)"),
+        (0x2488, "X"),
+        (0x249C, "3", "(a)"),
+        (0x249D, "3", "(b)"),
+        (0x249E, "3", "(c)"),
+        (0x249F, "3", "(d)"),
+        (0x24A0, "3", "(e)"),
+        (0x24A1, "3", "(f)"),
+        (0x24A2, "3", "(g)"),
+        (0x24A3, "3", "(h)"),
+        (0x24A4, "3", "(i)"),
+        (0x24A5, "3", "(j)"),
+        (0x24A6, "3", "(k)"),
+        (0x24A7, "3", "(l)"),
+        (0x24A8, "3", "(m)"),
+        (0x24A9, "3", "(n)"),
+        (0x24AA, "3", "(o)"),
+        (0x24AB, "3", "(p)"),
+        (0x24AC, "3", "(q)"),
+        (0x24AD, "3", "(r)"),
+        (0x24AE, "3", "(s)"),
+        (0x24AF, "3", "(t)"),
+        (0x24B0, "3", "(u)"),
+        (0x24B1, "3", "(v)"),
+        (0x24B2, "3", "(w)"),
+        (0x24B3, "3", "(x)"),
+        (0x24B4, "3", "(y)"),
+        (0x24B5, "3", "(z)"),
+        (0x24B6, "M", "a"),
+        (0x24B7, "M", "b"),
+        (0x24B8, "M", "c"),
+        (0x24B9, "M", "d"),
+        (0x24BA, "M", "e"),
+        (0x24BB, "M", "f"),
+        (0x24BC, "M", "g"),
+        (0x24BD, "M", "h"),
+        (0x24BE, "M", "i"),
+        (0x24BF, "M", "j"),
+        (0x24C0, "M", "k"),
+    ]
+
+
+def _seg_24() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x24C1, "M", "l"),
+        (0x24C2, "M", "m"),
+        (0x24C3, "M", "n"),
+        (0x24C4, "M", "o"),
+        (0x24C5, "M", "p"),
+        (0x24C6, "M", "q"),
+        (0x24C7, "M", "r"),
+        (0x24C8, "M", "s"),
+        (0x24C9, "M", "t"),
+        (0x24CA, "M", "u"),
+        (0x24CB, "M", "v"),
+        (0x24CC, "M", "w"),
+        (0x24CD, "M", "x"),
+        (0x24CE, "M", "y"),
+        (0x24CF, "M", "z"),
+        (0x24D0, "M", "a"),
+        (0x24D1, "M", "b"),
+        (0x24D2, "M", "c"),
+        (0x24D3, "M", "d"),
+        (0x24D4, "M", "e"),
+        (0x24D5, "M", "f"),
+        (0x24D6, "M", "g"),
+        (0x24D7, "M", "h"),
+        (0x24D8, "M", "i"),
+        (0x24D9, "M", "j"),
+        (0x24DA, "M", "k"),
+        (0x24DB, "M", "l"),
+        (0x24DC, "M", "m"),
+        (0x24DD, "M", "n"),
+        (0x24DE, "M", "o"),
+        (0x24DF, "M", "p"),
+        (0x24E0, "M", "q"),
+        (0x24E1, "M", "r"),
+        (0x24E2, "M", "s"),
+        (0x24E3, "M", "t"),
+        (0x24E4, "M", "u"),
+        (0x24E5, "M", "v"),
+        (0x24E6, "M", "w"),
+        (0x24E7, "M", "x"),
+        (0x24E8, "M", "y"),
+        (0x24E9, "M", "z"),
+        (0x24EA, "M", "0"),
+        (0x24EB, "V"),
+        (0x2A0C, "M", "∫∫∫∫"),
+        (0x2A0D, "V"),
+        (0x2A74, "3", "::="),
+        (0x2A75, "3", "=="),
+        (0x2A76, "3", "==="),
+        (0x2A77, "V"),
+        (0x2ADC, "M", "⫝̸"),
+        (0x2ADD, "V"),
+        (0x2B74, "X"),
+        (0x2B76, "V"),
+        (0x2B96, "X"),
+        (0x2B97, "V"),
+        (0x2C00, "M", "ⰰ"),
+        (0x2C01, "M", "ⰱ"),
+        (0x2C02, "M", "ⰲ"),
+        (0x2C03, "M", "ⰳ"),
+        (0x2C04, "M", "ⰴ"),
+        (0x2C05, "M", "ⰵ"),
+        (0x2C06, "M", "ⰶ"),
+        (0x2C07, "M", "ⰷ"),
+        (0x2C08, "M", "ⰸ"),
+        (0x2C09, "M", "ⰹ"),
+        (0x2C0A, "M", "ⰺ"),
+        (0x2C0B, "M", "ⰻ"),
+        (0x2C0C, "M", "ⰼ"),
+        (0x2C0D, "M", "ⰽ"),
+        (0x2C0E, "M", "ⰾ"),
+        (0x2C0F, "M", "ⰿ"),
+        (0x2C10, "M", "ⱀ"),
+        (0x2C11, "M", "ⱁ"),
+        (0x2C12, "M", "ⱂ"),
+        (0x2C13, "M", "ⱃ"),
+        (0x2C14, "M", "ⱄ"),
+        (0x2C15, "M", "ⱅ"),
+        (0x2C16, "M", "ⱆ"),
+        (0x2C17, "M", "ⱇ"),
+        (0x2C18, "M", "ⱈ"),
+        (0x2C19, "M", "ⱉ"),
+        (0x2C1A, "M", "ⱊ"),
+        (0x2C1B, "M", "ⱋ"),
+        (0x2C1C, "M", "ⱌ"),
+        (0x2C1D, "M", "ⱍ"),
+        (0x2C1E, "M", "ⱎ"),
+        (0x2C1F, "M", "ⱏ"),
+        (0x2C20, "M", "ⱐ"),
+        (0x2C21, "M", "ⱑ"),
+        (0x2C22, "M", "ⱒ"),
+        (0x2C23, "M", "ⱓ"),
+        (0x2C24, "M", "ⱔ"),
+        (0x2C25, "M", "ⱕ"),
+        (0x2C26, "M", "ⱖ"),
+        (0x2C27, "M", "ⱗ"),
+        (0x2C28, "M", "ⱘ"),
+        (0x2C29, "M", "ⱙ"),
+        (0x2C2A, "M", "ⱚ"),
+        (0x2C2B, "M", "ⱛ"),
+        (0x2C2C, "M", "ⱜ"),
+    ]
+
+
+def _seg_25() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2C2D, "M", "ⱝ"),
+        (0x2C2E, "M", "ⱞ"),
+        (0x2C2F, "M", "ⱟ"),
+        (0x2C30, "V"),
+        (0x2C60, "M", "ⱡ"),
+        (0x2C61, "V"),
+        (0x2C62, "M", "ɫ"),
+        (0x2C63, "M", "ᵽ"),
+        (0x2C64, "M", "ɽ"),
+        (0x2C65, "V"),
+        (0x2C67, "M", "ⱨ"),
+        (0x2C68, "V"),
+        (0x2C69, "M", "ⱪ"),
+        (0x2C6A, "V"),
+        (0x2C6B, "M", "ⱬ"),
+        (0x2C6C, "V"),
+        (0x2C6D, "M", "ɑ"),
+        (0x2C6E, "M", "ɱ"),
+        (0x2C6F, "M", "ɐ"),
+        (0x2C70, "M", "ɒ"),
+        (0x2C71, "V"),
+        (0x2C72, "M", "ⱳ"),
+        (0x2C73, "V"),
+        (0x2C75, "M", "ⱶ"),
+        (0x2C76, "V"),
+        (0x2C7C, "M", "j"),
+        (0x2C7D, "M", "v"),
+        (0x2C7E, "M", "ȿ"),
+        (0x2C7F, "M", "ɀ"),
+        (0x2C80, "M", "ⲁ"),
+        (0x2C81, "V"),
+        (0x2C82, "M", "ⲃ"),
+        (0x2C83, "V"),
+        (0x2C84, "M", "ⲅ"),
+        (0x2C85, "V"),
+        (0x2C86, "M", "ⲇ"),
+        (0x2C87, "V"),
+        (0x2C88, "M", "ⲉ"),
+        (0x2C89, "V"),
+        (0x2C8A, "M", "ⲋ"),
+        (0x2C8B, "V"),
+        (0x2C8C, "M", "ⲍ"),
+        (0x2C8D, "V"),
+        (0x2C8E, "M", "ⲏ"),
+        (0x2C8F, "V"),
+        (0x2C90, "M", "ⲑ"),
+        (0x2C91, "V"),
+        (0x2C92, "M", "ⲓ"),
+        (0x2C93, "V"),
+        (0x2C94, "M", "ⲕ"),
+        (0x2C95, "V"),
+        (0x2C96, "M", "ⲗ"),
+        (0x2C97, "V"),
+        (0x2C98, "M", "ⲙ"),
+        (0x2C99, "V"),
+        (0x2C9A, "M", "ⲛ"),
+        (0x2C9B, "V"),
+        (0x2C9C, "M", "ⲝ"),
+        (0x2C9D, "V"),
+        (0x2C9E, "M", "ⲟ"),
+        (0x2C9F, "V"),
+        (0x2CA0, "M", "ⲡ"),
+        (0x2CA1, "V"),
+        (0x2CA2, "M", "ⲣ"),
+        (0x2CA3, "V"),
+        (0x2CA4, "M", "ⲥ"),
+        (0x2CA5, "V"),
+        (0x2CA6, "M", "ⲧ"),
+        (0x2CA7, "V"),
+        (0x2CA8, "M", "ⲩ"),
+        (0x2CA9, "V"),
+        (0x2CAA, "M", "ⲫ"),
+        (0x2CAB, "V"),
+        (0x2CAC, "M", "ⲭ"),
+        (0x2CAD, "V"),
+        (0x2CAE, "M", "ⲯ"),
+        (0x2CAF, "V"),
+        (0x2CB0, "M", "ⲱ"),
+        (0x2CB1, "V"),
+        (0x2CB2, "M", "ⲳ"),
+        (0x2CB3, "V"),
+        (0x2CB4, "M", "ⲵ"),
+        (0x2CB5, "V"),
+        (0x2CB6, "M", "ⲷ"),
+        (0x2CB7, "V"),
+        (0x2CB8, "M", "ⲹ"),
+        (0x2CB9, "V"),
+        (0x2CBA, "M", "ⲻ"),
+        (0x2CBB, "V"),
+        (0x2CBC, "M", "ⲽ"),
+        (0x2CBD, "V"),
+        (0x2CBE, "M", "ⲿ"),
+        (0x2CBF, "V"),
+        (0x2CC0, "M", "ⳁ"),
+        (0x2CC1, "V"),
+        (0x2CC2, "M", "ⳃ"),
+        (0x2CC3, "V"),
+        (0x2CC4, "M", "ⳅ"),
+        (0x2CC5, "V"),
+        (0x2CC6, "M", "ⳇ"),
+    ]
+
+
+def _seg_26() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2CC7, "V"),
+        (0x2CC8, "M", "ⳉ"),
+        (0x2CC9, "V"),
+        (0x2CCA, "M", "ⳋ"),
+        (0x2CCB, "V"),
+        (0x2CCC, "M", "ⳍ"),
+        (0x2CCD, "V"),
+        (0x2CCE, "M", "ⳏ"),
+        (0x2CCF, "V"),
+        (0x2CD0, "M", "ⳑ"),
+        (0x2CD1, "V"),
+        (0x2CD2, "M", "ⳓ"),
+        (0x2CD3, "V"),
+        (0x2CD4, "M", "ⳕ"),
+        (0x2CD5, "V"),
+        (0x2CD6, "M", "ⳗ"),
+        (0x2CD7, "V"),
+        (0x2CD8, "M", "ⳙ"),
+        (0x2CD9, "V"),
+        (0x2CDA, "M", "ⳛ"),
+        (0x2CDB, "V"),
+        (0x2CDC, "M", "ⳝ"),
+        (0x2CDD, "V"),
+        (0x2CDE, "M", "ⳟ"),
+        (0x2CDF, "V"),
+        (0x2CE0, "M", "ⳡ"),
+        (0x2CE1, "V"),
+        (0x2CE2, "M", "ⳣ"),
+        (0x2CE3, "V"),
+        (0x2CEB, "M", "ⳬ"),
+        (0x2CEC, "V"),
+        (0x2CED, "M", "ⳮ"),
+        (0x2CEE, "V"),
+        (0x2CF2, "M", "ⳳ"),
+        (0x2CF3, "V"),
+        (0x2CF4, "X"),
+        (0x2CF9, "V"),
+        (0x2D26, "X"),
+        (0x2D27, "V"),
+        (0x2D28, "X"),
+        (0x2D2D, "V"),
+        (0x2D2E, "X"),
+        (0x2D30, "V"),
+        (0x2D68, "X"),
+        (0x2D6F, "M", "ⵡ"),
+        (0x2D70, "V"),
+        (0x2D71, "X"),
+        (0x2D7F, "V"),
+        (0x2D97, "X"),
+        (0x2DA0, "V"),
+        (0x2DA7, "X"),
+        (0x2DA8, "V"),
+        (0x2DAF, "X"),
+        (0x2DB0, "V"),
+        (0x2DB7, "X"),
+        (0x2DB8, "V"),
+        (0x2DBF, "X"),
+        (0x2DC0, "V"),
+        (0x2DC7, "X"),
+        (0x2DC8, "V"),
+        (0x2DCF, "X"),
+        (0x2DD0, "V"),
+        (0x2DD7, "X"),
+        (0x2DD8, "V"),
+        (0x2DDF, "X"),
+        (0x2DE0, "V"),
+        (0x2E5E, "X"),
+        (0x2E80, "V"),
+        (0x2E9A, "X"),
+        (0x2E9B, "V"),
+        (0x2E9F, "M", "母"),
+        (0x2EA0, "V"),
+        (0x2EF3, "M", "龟"),
+        (0x2EF4, "X"),
+        (0x2F00, "M", "一"),
+        (0x2F01, "M", "丨"),
+        (0x2F02, "M", "丶"),
+        (0x2F03, "M", "丿"),
+        (0x2F04, "M", "乙"),
+        (0x2F05, "M", "亅"),
+        (0x2F06, "M", "二"),
+        (0x2F07, "M", "亠"),
+        (0x2F08, "M", "人"),
+        (0x2F09, "M", "儿"),
+        (0x2F0A, "M", "入"),
+        (0x2F0B, "M", "八"),
+        (0x2F0C, "M", "冂"),
+        (0x2F0D, "M", "冖"),
+        (0x2F0E, "M", "冫"),
+        (0x2F0F, "M", "几"),
+        (0x2F10, "M", "凵"),
+        (0x2F11, "M", "刀"),
+        (0x2F12, "M", "力"),
+        (0x2F13, "M", "勹"),
+        (0x2F14, "M", "匕"),
+        (0x2F15, "M", "匚"),
+        (0x2F16, "M", "匸"),
+        (0x2F17, "M", "十"),
+        (0x2F18, "M", "卜"),
+        (0x2F19, "M", "卩"),
+    ]
+
+
+def _seg_27() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F1A, "M", "厂"),
+        (0x2F1B, "M", "厶"),
+        (0x2F1C, "M", "又"),
+        (0x2F1D, "M", "口"),
+        (0x2F1E, "M", "囗"),
+        (0x2F1F, "M", "土"),
+        (0x2F20, "M", "士"),
+        (0x2F21, "M", "夂"),
+        (0x2F22, "M", "夊"),
+        (0x2F23, "M", "夕"),
+        (0x2F24, "M", "大"),
+        (0x2F25, "M", "女"),
+        (0x2F26, "M", "子"),
+        (0x2F27, "M", "宀"),
+        (0x2F28, "M", "寸"),
+        (0x2F29, "M", "小"),
+        (0x2F2A, "M", "尢"),
+        (0x2F2B, "M", "尸"),
+        (0x2F2C, "M", "屮"),
+        (0x2F2D, "M", "山"),
+        (0x2F2E, "M", "巛"),
+        (0x2F2F, "M", "工"),
+        (0x2F30, "M", "己"),
+        (0x2F31, "M", "巾"),
+        (0x2F32, "M", "干"),
+        (0x2F33, "M", "幺"),
+        (0x2F34, "M", "广"),
+        (0x2F35, "M", "廴"),
+        (0x2F36, "M", "廾"),
+        (0x2F37, "M", "弋"),
+        (0x2F38, "M", "弓"),
+        (0x2F39, "M", "彐"),
+        (0x2F3A, "M", "彡"),
+        (0x2F3B, "M", "彳"),
+        (0x2F3C, "M", "心"),
+        (0x2F3D, "M", "戈"),
+        (0x2F3E, "M", "戶"),
+        (0x2F3F, "M", "手"),
+        (0x2F40, "M", "支"),
+        (0x2F41, "M", "攴"),
+        (0x2F42, "M", "文"),
+        (0x2F43, "M", "斗"),
+        (0x2F44, "M", "斤"),
+        (0x2F45, "M", "方"),
+        (0x2F46, "M", "无"),
+        (0x2F47, "M", "日"),
+        (0x2F48, "M", "曰"),
+        (0x2F49, "M", "月"),
+        (0x2F4A, "M", "木"),
+        (0x2F4B, "M", "欠"),
+        (0x2F4C, "M", "止"),
+        (0x2F4D, "M", "歹"),
+        (0x2F4E, "M", "殳"),
+        (0x2F4F, "M", "毋"),
+        (0x2F50, "M", "比"),
+        (0x2F51, "M", "毛"),
+        (0x2F52, "M", "氏"),
+        (0x2F53, "M", "气"),
+        (0x2F54, "M", "水"),
+        (0x2F55, "M", "火"),
+        (0x2F56, "M", "爪"),
+        (0x2F57, "M", "父"),
+        (0x2F58, "M", "爻"),
+        (0x2F59, "M", "爿"),
+        (0x2F5A, "M", "片"),
+        (0x2F5B, "M", "牙"),
+        (0x2F5C, "M", "牛"),
+        (0x2F5D, "M", "犬"),
+        (0x2F5E, "M", "玄"),
+        (0x2F5F, "M", "玉"),
+        (0x2F60, "M", "瓜"),
+        (0x2F61, "M", "瓦"),
+        (0x2F62, "M", "甘"),
+        (0x2F63, "M", "生"),
+        (0x2F64, "M", "用"),
+        (0x2F65, "M", "田"),
+        (0x2F66, "M", "疋"),
+        (0x2F67, "M", "疒"),
+        (0x2F68, "M", "癶"),
+        (0x2F69, "M", "白"),
+        (0x2F6A, "M", "皮"),
+        (0x2F6B, "M", "皿"),
+        (0x2F6C, "M", "目"),
+        (0x2F6D, "M", "矛"),
+        (0x2F6E, "M", "矢"),
+        (0x2F6F, "M", "石"),
+        (0x2F70, "M", "示"),
+        (0x2F71, "M", "禸"),
+        (0x2F72, "M", "禾"),
+        (0x2F73, "M", "穴"),
+        (0x2F74, "M", "立"),
+        (0x2F75, "M", "竹"),
+        (0x2F76, "M", "米"),
+        (0x2F77, "M", "糸"),
+        (0x2F78, "M", "缶"),
+        (0x2F79, "M", "网"),
+        (0x2F7A, "M", "羊"),
+        (0x2F7B, "M", "羽"),
+        (0x2F7C, "M", "老"),
+        (0x2F7D, "M", "而"),
+    ]
+
+
+def _seg_28() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F7E, "M", "耒"),
+        (0x2F7F, "M", "耳"),
+        (0x2F80, "M", "聿"),
+        (0x2F81, "M", "肉"),
+        (0x2F82, "M", "臣"),
+        (0x2F83, "M", "自"),
+        (0x2F84, "M", "至"),
+        (0x2F85, "M", "臼"),
+        (0x2F86, "M", "舌"),
+        (0x2F87, "M", "舛"),
+        (0x2F88, "M", "舟"),
+        (0x2F89, "M", "艮"),
+        (0x2F8A, "M", "色"),
+        (0x2F8B, "M", "艸"),
+        (0x2F8C, "M", "虍"),
+        (0x2F8D, "M", "虫"),
+        (0x2F8E, "M", "血"),
+        (0x2F8F, "M", "行"),
+        (0x2F90, "M", "衣"),
+        (0x2F91, "M", "襾"),
+        (0x2F92, "M", "見"),
+        (0x2F93, "M", "角"),
+        (0x2F94, "M", "言"),
+        (0x2F95, "M", "谷"),
+        (0x2F96, "M", "豆"),
+        (0x2F97, "M", "豕"),
+        (0x2F98, "M", "豸"),
+        (0x2F99, "M", "貝"),
+        (0x2F9A, "M", "赤"),
+        (0x2F9B, "M", "走"),
+        (0x2F9C, "M", "足"),
+        (0x2F9D, "M", "身"),
+        (0x2F9E, "M", "車"),
+        (0x2F9F, "M", "辛"),
+        (0x2FA0, "M", "辰"),
+        (0x2FA1, "M", "辵"),
+        (0x2FA2, "M", "邑"),
+        (0x2FA3, "M", "酉"),
+        (0x2FA4, "M", "釆"),
+        (0x2FA5, "M", "里"),
+        (0x2FA6, "M", "金"),
+        (0x2FA7, "M", "長"),
+        (0x2FA8, "M", "門"),
+        (0x2FA9, "M", "阜"),
+        (0x2FAA, "M", "隶"),
+        (0x2FAB, "M", "隹"),
+        (0x2FAC, "M", "雨"),
+        (0x2FAD, "M", "靑"),
+        (0x2FAE, "M", "非"),
+        (0x2FAF, "M", "面"),
+        (0x2FB0, "M", "革"),
+        (0x2FB1, "M", "韋"),
+        (0x2FB2, "M", "韭"),
+        (0x2FB3, "M", "音"),
+        (0x2FB4, "M", "頁"),
+        (0x2FB5, "M", "風"),
+        (0x2FB6, "M", "飛"),
+        (0x2FB7, "M", "食"),
+        (0x2FB8, "M", "首"),
+        (0x2FB9, "M", "香"),
+        (0x2FBA, "M", "馬"),
+        (0x2FBB, "M", "骨"),
+        (0x2FBC, "M", "高"),
+        (0x2FBD, "M", "髟"),
+        (0x2FBE, "M", "鬥"),
+        (0x2FBF, "M", "鬯"),
+        (0x2FC0, "M", "鬲"),
+        (0x2FC1, "M", "鬼"),
+        (0x2FC2, "M", "魚"),
+        (0x2FC3, "M", "鳥"),
+        (0x2FC4, "M", "鹵"),
+        (0x2FC5, "M", "鹿"),
+        (0x2FC6, "M", "麥"),
+        (0x2FC7, "M", "麻"),
+        (0x2FC8, "M", "黃"),
+        (0x2FC9, "M", "黍"),
+        (0x2FCA, "M", "黑"),
+        (0x2FCB, "M", "黹"),
+        (0x2FCC, "M", "黽"),
+        (0x2FCD, "M", "鼎"),
+        (0x2FCE, "M", "鼓"),
+        (0x2FCF, "M", "鼠"),
+        (0x2FD0, "M", "鼻"),
+        (0x2FD1, "M", "齊"),
+        (0x2FD2, "M", "齒"),
+        (0x2FD3, "M", "龍"),
+        (0x2FD4, "M", "龜"),
+        (0x2FD5, "M", "龠"),
+        (0x2FD6, "X"),
+        (0x3000, "3", " "),
+        (0x3001, "V"),
+        (0x3002, "M", "."),
+        (0x3003, "V"),
+        (0x3036, "M", "〒"),
+        (0x3037, "V"),
+        (0x3038, "M", "十"),
+        (0x3039, "M", "卄"),
+        (0x303A, "M", "卅"),
+        (0x303B, "V"),
+        (0x3040, "X"),
+    ]
+
+
+def _seg_29() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3041, "V"),
+        (0x3097, "X"),
+        (0x3099, "V"),
+        (0x309B, "3", " ゙"),
+        (0x309C, "3", " ゚"),
+        (0x309D, "V"),
+        (0x309F, "M", "より"),
+        (0x30A0, "V"),
+        (0x30FF, "M", "コト"),
+        (0x3100, "X"),
+        (0x3105, "V"),
+        (0x3130, "X"),
+        (0x3131, "M", "ᄀ"),
+        (0x3132, "M", "ᄁ"),
+        (0x3133, "M", "ᆪ"),
+        (0x3134, "M", "ᄂ"),
+        (0x3135, "M", "ᆬ"),
+        (0x3136, "M", "ᆭ"),
+        (0x3137, "M", "ᄃ"),
+        (0x3138, "M", "ᄄ"),
+        (0x3139, "M", "ᄅ"),
+        (0x313A, "M", "ᆰ"),
+        (0x313B, "M", "ᆱ"),
+        (0x313C, "M", "ᆲ"),
+        (0x313D, "M", "ᆳ"),
+        (0x313E, "M", "ᆴ"),
+        (0x313F, "M", "ᆵ"),
+        (0x3140, "M", "ᄚ"),
+        (0x3141, "M", "ᄆ"),
+        (0x3142, "M", "ᄇ"),
+        (0x3143, "M", "ᄈ"),
+        (0x3144, "M", "ᄡ"),
+        (0x3145, "M", "ᄉ"),
+        (0x3146, "M", "ᄊ"),
+        (0x3147, "M", "ᄋ"),
+        (0x3148, "M", "ᄌ"),
+        (0x3149, "M", "ᄍ"),
+        (0x314A, "M", "ᄎ"),
+        (0x314B, "M", "ᄏ"),
+        (0x314C, "M", "ᄐ"),
+        (0x314D, "M", "ᄑ"),
+        (0x314E, "M", "ᄒ"),
+        (0x314F, "M", "ᅡ"),
+        (0x3150, "M", "ᅢ"),
+        (0x3151, "M", "ᅣ"),
+        (0x3152, "M", "ᅤ"),
+        (0x3153, "M", "ᅥ"),
+        (0x3154, "M", "ᅦ"),
+        (0x3155, "M", "ᅧ"),
+        (0x3156, "M", "ᅨ"),
+        (0x3157, "M", "ᅩ"),
+        (0x3158, "M", "ᅪ"),
+        (0x3159, "M", "ᅫ"),
+        (0x315A, "M", "ᅬ"),
+        (0x315B, "M", "ᅭ"),
+        (0x315C, "M", "ᅮ"),
+        (0x315D, "M", "ᅯ"),
+        (0x315E, "M", "ᅰ"),
+        (0x315F, "M", "ᅱ"),
+        (0x3160, "M", "ᅲ"),
+        (0x3161, "M", "ᅳ"),
+        (0x3162, "M", "ᅴ"),
+        (0x3163, "M", "ᅵ"),
+        (0x3164, "X"),
+        (0x3165, "M", "ᄔ"),
+        (0x3166, "M", "ᄕ"),
+        (0x3167, "M", "ᇇ"),
+        (0x3168, "M", "ᇈ"),
+        (0x3169, "M", "ᇌ"),
+        (0x316A, "M", "ᇎ"),
+        (0x316B, "M", "ᇓ"),
+        (0x316C, "M", "ᇗ"),
+        (0x316D, "M", "ᇙ"),
+        (0x316E, "M", "ᄜ"),
+        (0x316F, "M", "ᇝ"),
+        (0x3170, "M", "ᇟ"),
+        (0x3171, "M", "ᄝ"),
+        (0x3172, "M", "ᄞ"),
+        (0x3173, "M", "ᄠ"),
+        (0x3174, "M", "ᄢ"),
+        (0x3175, "M", "ᄣ"),
+        (0x3176, "M", "ᄧ"),
+        (0x3177, "M", "ᄩ"),
+        (0x3178, "M", "ᄫ"),
+        (0x3179, "M", "ᄬ"),
+        (0x317A, "M", "ᄭ"),
+        (0x317B, "M", "ᄮ"),
+        (0x317C, "M", "ᄯ"),
+        (0x317D, "M", "ᄲ"),
+        (0x317E, "M", "ᄶ"),
+        (0x317F, "M", "ᅀ"),
+        (0x3180, "M", "ᅇ"),
+        (0x3181, "M", "ᅌ"),
+        (0x3182, "M", "ᇱ"),
+        (0x3183, "M", "ᇲ"),
+        (0x3184, "M", "ᅗ"),
+        (0x3185, "M", "ᅘ"),
+        (0x3186, "M", "ᅙ"),
+        (0x3187, "M", "ᆄ"),
+        (0x3188, "M", "ᆅ"),
+    ]
+
+
+def _seg_30() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3189, "M", "ᆈ"),
+        (0x318A, "M", "ᆑ"),
+        (0x318B, "M", "ᆒ"),
+        (0x318C, "M", "ᆔ"),
+        (0x318D, "M", "ᆞ"),
+        (0x318E, "M", "ᆡ"),
+        (0x318F, "X"),
+        (0x3190, "V"),
+        (0x3192, "M", "一"),
+        (0x3193, "M", "二"),
+        (0x3194, "M", "三"),
+        (0x3195, "M", "四"),
+        (0x3196, "M", "上"),
+        (0x3197, "M", "中"),
+        (0x3198, "M", "下"),
+        (0x3199, "M", "甲"),
+        (0x319A, "M", "乙"),
+        (0x319B, "M", "丙"),
+        (0x319C, "M", "丁"),
+        (0x319D, "M", "天"),
+        (0x319E, "M", "地"),
+        (0x319F, "M", "人"),
+        (0x31A0, "V"),
+        (0x31E4, "X"),
+        (0x31F0, "V"),
+        (0x3200, "3", "(ᄀ)"),
+        (0x3201, "3", "(ᄂ)"),
+        (0x3202, "3", "(ᄃ)"),
+        (0x3203, "3", "(ᄅ)"),
+        (0x3204, "3", "(ᄆ)"),
+        (0x3205, "3", "(ᄇ)"),
+        (0x3206, "3", "(ᄉ)"),
+        (0x3207, "3", "(ᄋ)"),
+        (0x3208, "3", "(ᄌ)"),
+        (0x3209, "3", "(ᄎ)"),
+        (0x320A, "3", "(ᄏ)"),
+        (0x320B, "3", "(ᄐ)"),
+        (0x320C, "3", "(ᄑ)"),
+        (0x320D, "3", "(ᄒ)"),
+        (0x320E, "3", "(가)"),
+        (0x320F, "3", "(나)"),
+        (0x3210, "3", "(다)"),
+        (0x3211, "3", "(라)"),
+        (0x3212, "3", "(마)"),
+        (0x3213, "3", "(바)"),
+        (0x3214, "3", "(사)"),
+        (0x3215, "3", "(아)"),
+        (0x3216, "3", "(자)"),
+        (0x3217, "3", "(차)"),
+        (0x3218, "3", "(카)"),
+        (0x3219, "3", "(타)"),
+        (0x321A, "3", "(파)"),
+        (0x321B, "3", "(하)"),
+        (0x321C, "3", "(주)"),
+        (0x321D, "3", "(오전)"),
+        (0x321E, "3", "(오후)"),
+        (0x321F, "X"),
+        (0x3220, "3", "(一)"),
+        (0x3221, "3", "(二)"),
+        (0x3222, "3", "(三)"),
+        (0x3223, "3", "(四)"),
+        (0x3224, "3", "(五)"),
+        (0x3225, "3", "(六)"),
+        (0x3226, "3", "(七)"),
+        (0x3227, "3", "(八)"),
+        (0x3228, "3", "(九)"),
+        (0x3229, "3", "(十)"),
+        (0x322A, "3", "(月)"),
+        (0x322B, "3", "(火)"),
+        (0x322C, "3", "(水)"),
+        (0x322D, "3", "(木)"),
+        (0x322E, "3", "(金)"),
+        (0x322F, "3", "(土)"),
+        (0x3230, "3", "(日)"),
+        (0x3231, "3", "(株)"),
+        (0x3232, "3", "(有)"),
+        (0x3233, "3", "(社)"),
+        (0x3234, "3", "(名)"),
+        (0x3235, "3", "(特)"),
+        (0x3236, "3", "(財)"),
+        (0x3237, "3", "(祝)"),
+        (0x3238, "3", "(労)"),
+        (0x3239, "3", "(代)"),
+        (0x323A, "3", "(呼)"),
+        (0x323B, "3", "(学)"),
+        (0x323C, "3", "(監)"),
+        (0x323D, "3", "(企)"),
+        (0x323E, "3", "(資)"),
+        (0x323F, "3", "(協)"),
+        (0x3240, "3", "(祭)"),
+        (0x3241, "3", "(休)"),
+        (0x3242, "3", "(自)"),
+        (0x3243, "3", "(至)"),
+        (0x3244, "M", "問"),
+        (0x3245, "M", "幼"),
+        (0x3246, "M", "文"),
+        (0x3247, "M", "箏"),
+        (0x3248, "V"),
+        (0x3250, "M", "pte"),
+        (0x3251, "M", "21"),
+    ]
+
+
+def _seg_31() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3252, "M", "22"),
+        (0x3253, "M", "23"),
+        (0x3254, "M", "24"),
+        (0x3255, "M", "25"),
+        (0x3256, "M", "26"),
+        (0x3257, "M", "27"),
+        (0x3258, "M", "28"),
+        (0x3259, "M", "29"),
+        (0x325A, "M", "30"),
+        (0x325B, "M", "31"),
+        (0x325C, "M", "32"),
+        (0x325D, "M", "33"),
+        (0x325E, "M", "34"),
+        (0x325F, "M", "35"),
+        (0x3260, "M", "ᄀ"),
+        (0x3261, "M", "ᄂ"),
+        (0x3262, "M", "ᄃ"),
+        (0x3263, "M", "ᄅ"),
+        (0x3264, "M", "ᄆ"),
+        (0x3265, "M", "ᄇ"),
+        (0x3266, "M", "ᄉ"),
+        (0x3267, "M", "ᄋ"),
+        (0x3268, "M", "ᄌ"),
+        (0x3269, "M", "ᄎ"),
+        (0x326A, "M", "ᄏ"),
+        (0x326B, "M", "ᄐ"),
+        (0x326C, "M", "ᄑ"),
+        (0x326D, "M", "ᄒ"),
+        (0x326E, "M", "가"),
+        (0x326F, "M", "나"),
+        (0x3270, "M", "다"),
+        (0x3271, "M", "라"),
+        (0x3272, "M", "마"),
+        (0x3273, "M", "바"),
+        (0x3274, "M", "사"),
+        (0x3275, "M", "아"),
+        (0x3276, "M", "자"),
+        (0x3277, "M", "차"),
+        (0x3278, "M", "카"),
+        (0x3279, "M", "타"),
+        (0x327A, "M", "파"),
+        (0x327B, "M", "하"),
+        (0x327C, "M", "참고"),
+        (0x327D, "M", "주의"),
+        (0x327E, "M", "우"),
+        (0x327F, "V"),
+        (0x3280, "M", "一"),
+        (0x3281, "M", "二"),
+        (0x3282, "M", "三"),
+        (0x3283, "M", "四"),
+        (0x3284, "M", "五"),
+        (0x3285, "M", "六"),
+        (0x3286, "M", "七"),
+        (0x3287, "M", "八"),
+        (0x3288, "M", "九"),
+        (0x3289, "M", "十"),
+        (0x328A, "M", "月"),
+        (0x328B, "M", "火"),
+        (0x328C, "M", "水"),
+        (0x328D, "M", "木"),
+        (0x328E, "M", "金"),
+        (0x328F, "M", "土"),
+        (0x3290, "M", "日"),
+        (0x3291, "M", "株"),
+        (0x3292, "M", "有"),
+        (0x3293, "M", "社"),
+        (0x3294, "M", "名"),
+        (0x3295, "M", "特"),
+        (0x3296, "M", "財"),
+        (0x3297, "M", "祝"),
+        (0x3298, "M", "労"),
+        (0x3299, "M", "秘"),
+        (0x329A, "M", "男"),
+        (0x329B, "M", "女"),
+        (0x329C, "M", "適"),
+        (0x329D, "M", "優"),
+        (0x329E, "M", "印"),
+        (0x329F, "M", "注"),
+        (0x32A0, "M", "項"),
+        (0x32A1, "M", "休"),
+        (0x32A2, "M", "写"),
+        (0x32A3, "M", "正"),
+        (0x32A4, "M", "上"),
+        (0x32A5, "M", "中"),
+        (0x32A6, "M", "下"),
+        (0x32A7, "M", "左"),
+        (0x32A8, "M", "右"),
+        (0x32A9, "M", "医"),
+        (0x32AA, "M", "宗"),
+        (0x32AB, "M", "学"),
+        (0x32AC, "M", "監"),
+        (0x32AD, "M", "企"),
+        (0x32AE, "M", "資"),
+        (0x32AF, "M", "協"),
+        (0x32B0, "M", "夜"),
+        (0x32B1, "M", "36"),
+        (0x32B2, "M", "37"),
+        (0x32B3, "M", "38"),
+        (0x32B4, "M", "39"),
+        (0x32B5, "M", "40"),
+    ]
+
+
+def _seg_32() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x32B6, "M", "41"),
+        (0x32B7, "M", "42"),
+        (0x32B8, "M", "43"),
+        (0x32B9, "M", "44"),
+        (0x32BA, "M", "45"),
+        (0x32BB, "M", "46"),
+        (0x32BC, "M", "47"),
+        (0x32BD, "M", "48"),
+        (0x32BE, "M", "49"),
+        (0x32BF, "M", "50"),
+        (0x32C0, "M", "1月"),
+        (0x32C1, "M", "2月"),
+        (0x32C2, "M", "3月"),
+        (0x32C3, "M", "4月"),
+        (0x32C4, "M", "5月"),
+        (0x32C5, "M", "6月"),
+        (0x32C6, "M", "7月"),
+        (0x32C7, "M", "8月"),
+        (0x32C8, "M", "9月"),
+        (0x32C9, "M", "10月"),
+        (0x32CA, "M", "11月"),
+        (0x32CB, "M", "12月"),
+        (0x32CC, "M", "hg"),
+        (0x32CD, "M", "erg"),
+        (0x32CE, "M", "ev"),
+        (0x32CF, "M", "ltd"),
+        (0x32D0, "M", "ア"),
+        (0x32D1, "M", "イ"),
+        (0x32D2, "M", "ウ"),
+        (0x32D3, "M", "エ"),
+        (0x32D4, "M", "オ"),
+        (0x32D5, "M", "カ"),
+        (0x32D6, "M", "キ"),
+        (0x32D7, "M", "ク"),
+        (0x32D8, "M", "ケ"),
+        (0x32D9, "M", "コ"),
+        (0x32DA, "M", "サ"),
+        (0x32DB, "M", "シ"),
+        (0x32DC, "M", "ス"),
+        (0x32DD, "M", "セ"),
+        (0x32DE, "M", "ソ"),
+        (0x32DF, "M", "タ"),
+        (0x32E0, "M", "チ"),
+        (0x32E1, "M", "ツ"),
+        (0x32E2, "M", "テ"),
+        (0x32E3, "M", "ト"),
+        (0x32E4, "M", "ナ"),
+        (0x32E5, "M", "ニ"),
+        (0x32E6, "M", "ヌ"),
+        (0x32E7, "M", "ネ"),
+        (0x32E8, "M", "ノ"),
+        (0x32E9, "M", "ハ"),
+        (0x32EA, "M", "ヒ"),
+        (0x32EB, "M", "フ"),
+        (0x32EC, "M", "ヘ"),
+        (0x32ED, "M", "ホ"),
+        (0x32EE, "M", "マ"),
+        (0x32EF, "M", "ミ"),
+        (0x32F0, "M", "ム"),
+        (0x32F1, "M", "メ"),
+        (0x32F2, "M", "モ"),
+        (0x32F3, "M", "ヤ"),
+        (0x32F4, "M", "ユ"),
+        (0x32F5, "M", "ヨ"),
+        (0x32F6, "M", "ラ"),
+        (0x32F7, "M", "リ"),
+        (0x32F8, "M", "ル"),
+        (0x32F9, "M", "レ"),
+        (0x32FA, "M", "ロ"),
+        (0x32FB, "M", "ワ"),
+        (0x32FC, "M", "ヰ"),
+        (0x32FD, "M", "ヱ"),
+        (0x32FE, "M", "ヲ"),
+        (0x32FF, "M", "令和"),
+        (0x3300, "M", "アパート"),
+        (0x3301, "M", "アルファ"),
+        (0x3302, "M", "アンペア"),
+        (0x3303, "M", "アール"),
+        (0x3304, "M", "イニング"),
+        (0x3305, "M", "インチ"),
+        (0x3306, "M", "ウォン"),
+        (0x3307, "M", "エスクード"),
+        (0x3308, "M", "エーカー"),
+        (0x3309, "M", "オンス"),
+        (0x330A, "M", "オーム"),
+        (0x330B, "M", "カイリ"),
+        (0x330C, "M", "カラット"),
+        (0x330D, "M", "カロリー"),
+        (0x330E, "M", "ガロン"),
+        (0x330F, "M", "ガンマ"),
+        (0x3310, "M", "ギガ"),
+        (0x3311, "M", "ギニー"),
+        (0x3312, "M", "キュリー"),
+        (0x3313, "M", "ギルダー"),
+        (0x3314, "M", "キロ"),
+        (0x3315, "M", "キログラム"),
+        (0x3316, "M", "キロメートル"),
+        (0x3317, "M", "キロワット"),
+        (0x3318, "M", "グラム"),
+        (0x3319, "M", "グラムトン"),
+    ]
+
+
+def _seg_33() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x331A, "M", "クルゼイロ"),
+        (0x331B, "M", "クローネ"),
+        (0x331C, "M", "ケース"),
+        (0x331D, "M", "コルナ"),
+        (0x331E, "M", "コーポ"),
+        (0x331F, "M", "サイクル"),
+        (0x3320, "M", "サンチーム"),
+        (0x3321, "M", "シリング"),
+        (0x3322, "M", "センチ"),
+        (0x3323, "M", "セント"),
+        (0x3324, "M", "ダース"),
+        (0x3325, "M", "デシ"),
+        (0x3326, "M", "ドル"),
+        (0x3327, "M", "トン"),
+        (0x3328, "M", "ナノ"),
+        (0x3329, "M", "ノット"),
+        (0x332A, "M", "ハイツ"),
+        (0x332B, "M", "パーセント"),
+        (0x332C, "M", "パーツ"),
+        (0x332D, "M", "バーレル"),
+        (0x332E, "M", "ピアストル"),
+        (0x332F, "M", "ピクル"),
+        (0x3330, "M", "ピコ"),
+        (0x3331, "M", "ビル"),
+        (0x3332, "M", "ファラッド"),
+        (0x3333, "M", "フィート"),
+        (0x3334, "M", "ブッシェル"),
+        (0x3335, "M", "フラン"),
+        (0x3336, "M", "ヘクタール"),
+        (0x3337, "M", "ペソ"),
+        (0x3338, "M", "ペニヒ"),
+        (0x3339, "M", "ヘルツ"),
+        (0x333A, "M", "ペンス"),
+        (0x333B, "M", "ページ"),
+        (0x333C, "M", "ベータ"),
+        (0x333D, "M", "ポイント"),
+        (0x333E, "M", "ボルト"),
+        (0x333F, "M", "ホン"),
+        (0x3340, "M", "ポンド"),
+        (0x3341, "M", "ホール"),
+        (0x3342, "M", "ホーン"),
+        (0x3343, "M", "マイクロ"),
+        (0x3344, "M", "マイル"),
+        (0x3345, "M", "マッハ"),
+        (0x3346, "M", "マルク"),
+        (0x3347, "M", "マンション"),
+        (0x3348, "M", "ミクロン"),
+        (0x3349, "M", "ミリ"),
+        (0x334A, "M", "ミリバール"),
+        (0x334B, "M", "メガ"),
+        (0x334C, "M", "メガトン"),
+        (0x334D, "M", "メートル"),
+        (0x334E, "M", "ヤード"),
+        (0x334F, "M", "ヤール"),
+        (0x3350, "M", "ユアン"),
+        (0x3351, "M", "リットル"),
+        (0x3352, "M", "リラ"),
+        (0x3353, "M", "ルピー"),
+        (0x3354, "M", "ルーブル"),
+        (0x3355, "M", "レム"),
+        (0x3356, "M", "レントゲン"),
+        (0x3357, "M", "ワット"),
+        (0x3358, "M", "0点"),
+        (0x3359, "M", "1点"),
+        (0x335A, "M", "2点"),
+        (0x335B, "M", "3点"),
+        (0x335C, "M", "4点"),
+        (0x335D, "M", "5点"),
+        (0x335E, "M", "6点"),
+        (0x335F, "M", "7点"),
+        (0x3360, "M", "8点"),
+        (0x3361, "M", "9点"),
+        (0x3362, "M", "10点"),
+        (0x3363, "M", "11点"),
+        (0x3364, "M", "12点"),
+        (0x3365, "M", "13点"),
+        (0x3366, "M", "14点"),
+        (0x3367, "M", "15点"),
+        (0x3368, "M", "16点"),
+        (0x3369, "M", "17点"),
+        (0x336A, "M", "18点"),
+        (0x336B, "M", "19点"),
+        (0x336C, "M", "20点"),
+        (0x336D, "M", "21点"),
+        (0x336E, "M", "22点"),
+        (0x336F, "M", "23点"),
+        (0x3370, "M", "24点"),
+        (0x3371, "M", "hpa"),
+        (0x3372, "M", "da"),
+        (0x3373, "M", "au"),
+        (0x3374, "M", "bar"),
+        (0x3375, "M", "ov"),
+        (0x3376, "M", "pc"),
+        (0x3377, "M", "dm"),
+        (0x3378, "M", "dm2"),
+        (0x3379, "M", "dm3"),
+        (0x337A, "M", "iu"),
+        (0x337B, "M", "平成"),
+        (0x337C, "M", "昭和"),
+        (0x337D, "M", "大正"),
+    ]
+
+
+def _seg_34() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x337E, "M", "明治"),
+        (0x337F, "M", "株式会社"),
+        (0x3380, "M", "pa"),
+        (0x3381, "M", "na"),
+        (0x3382, "M", "μa"),
+        (0x3383, "M", "ma"),
+        (0x3384, "M", "ka"),
+        (0x3385, "M", "kb"),
+        (0x3386, "M", "mb"),
+        (0x3387, "M", "gb"),
+        (0x3388, "M", "cal"),
+        (0x3389, "M", "kcal"),
+        (0x338A, "M", "pf"),
+        (0x338B, "M", "nf"),
+        (0x338C, "M", "μf"),
+        (0x338D, "M", "μg"),
+        (0x338E, "M", "mg"),
+        (0x338F, "M", "kg"),
+        (0x3390, "M", "hz"),
+        (0x3391, "M", "khz"),
+        (0x3392, "M", "mhz"),
+        (0x3393, "M", "ghz"),
+        (0x3394, "M", "thz"),
+        (0x3395, "M", "μl"),
+        (0x3396, "M", "ml"),
+        (0x3397, "M", "dl"),
+        (0x3398, "M", "kl"),
+        (0x3399, "M", "fm"),
+        (0x339A, "M", "nm"),
+        (0x339B, "M", "μm"),
+        (0x339C, "M", "mm"),
+        (0x339D, "M", "cm"),
+        (0x339E, "M", "km"),
+        (0x339F, "M", "mm2"),
+        (0x33A0, "M", "cm2"),
+        (0x33A1, "M", "m2"),
+        (0x33A2, "M", "km2"),
+        (0x33A3, "M", "mm3"),
+        (0x33A4, "M", "cm3"),
+        (0x33A5, "M", "m3"),
+        (0x33A6, "M", "km3"),
+        (0x33A7, "M", "m∕s"),
+        (0x33A8, "M", "m∕s2"),
+        (0x33A9, "M", "pa"),
+        (0x33AA, "M", "kpa"),
+        (0x33AB, "M", "mpa"),
+        (0x33AC, "M", "gpa"),
+        (0x33AD, "M", "rad"),
+        (0x33AE, "M", "rad∕s"),
+        (0x33AF, "M", "rad∕s2"),
+        (0x33B0, "M", "ps"),
+        (0x33B1, "M", "ns"),
+        (0x33B2, "M", "μs"),
+        (0x33B3, "M", "ms"),
+        (0x33B4, "M", "pv"),
+        (0x33B5, "M", "nv"),
+        (0x33B6, "M", "μv"),
+        (0x33B7, "M", "mv"),
+        (0x33B8, "M", "kv"),
+        (0x33B9, "M", "mv"),
+        (0x33BA, "M", "pw"),
+        (0x33BB, "M", "nw"),
+        (0x33BC, "M", "μw"),
+        (0x33BD, "M", "mw"),
+        (0x33BE, "M", "kw"),
+        (0x33BF, "M", "mw"),
+        (0x33C0, "M", "kω"),
+        (0x33C1, "M", "mω"),
+        (0x33C2, "X"),
+        (0x33C3, "M", "bq"),
+        (0x33C4, "M", "cc"),
+        (0x33C5, "M", "cd"),
+        (0x33C6, "M", "c∕kg"),
+        (0x33C7, "X"),
+        (0x33C8, "M", "db"),
+        (0x33C9, "M", "gy"),
+        (0x33CA, "M", "ha"),
+        (0x33CB, "M", "hp"),
+        (0x33CC, "M", "in"),
+        (0x33CD, "M", "kk"),
+        (0x33CE, "M", "km"),
+        (0x33CF, "M", "kt"),
+        (0x33D0, "M", "lm"),
+        (0x33D1, "M", "ln"),
+        (0x33D2, "M", "log"),
+        (0x33D3, "M", "lx"),
+        (0x33D4, "M", "mb"),
+        (0x33D5, "M", "mil"),
+        (0x33D6, "M", "mol"),
+        (0x33D7, "M", "ph"),
+        (0x33D8, "X"),
+        (0x33D9, "M", "ppm"),
+        (0x33DA, "M", "pr"),
+        (0x33DB, "M", "sr"),
+        (0x33DC, "M", "sv"),
+        (0x33DD, "M", "wb"),
+        (0x33DE, "M", "v∕m"),
+        (0x33DF, "M", "a∕m"),
+        (0x33E0, "M", "1日"),
+        (0x33E1, "M", "2日"),
+    ]
+
+
+def _seg_35() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x33E2, "M", "3日"),
+        (0x33E3, "M", "4日"),
+        (0x33E4, "M", "5日"),
+        (0x33E5, "M", "6日"),
+        (0x33E6, "M", "7日"),
+        (0x33E7, "M", "8日"),
+        (0x33E8, "M", "9日"),
+        (0x33E9, "M", "10日"),
+        (0x33EA, "M", "11日"),
+        (0x33EB, "M", "12日"),
+        (0x33EC, "M", "13日"),
+        (0x33ED, "M", "14日"),
+        (0x33EE, "M", "15日"),
+        (0x33EF, "M", "16日"),
+        (0x33F0, "M", "17日"),
+        (0x33F1, "M", "18日"),
+        (0x33F2, "M", "19日"),
+        (0x33F3, "M", "20日"),
+        (0x33F4, "M", "21日"),
+        (0x33F5, "M", "22日"),
+        (0x33F6, "M", "23日"),
+        (0x33F7, "M", "24日"),
+        (0x33F8, "M", "25日"),
+        (0x33F9, "M", "26日"),
+        (0x33FA, "M", "27日"),
+        (0x33FB, "M", "28日"),
+        (0x33FC, "M", "29日"),
+        (0x33FD, "M", "30日"),
+        (0x33FE, "M", "31日"),
+        (0x33FF, "M", "gal"),
+        (0x3400, "V"),
+        (0xA48D, "X"),
+        (0xA490, "V"),
+        (0xA4C7, "X"),
+        (0xA4D0, "V"),
+        (0xA62C, "X"),
+        (0xA640, "M", "ꙁ"),
+        (0xA641, "V"),
+        (0xA642, "M", "ꙃ"),
+        (0xA643, "V"),
+        (0xA644, "M", "ꙅ"),
+        (0xA645, "V"),
+        (0xA646, "M", "ꙇ"),
+        (0xA647, "V"),
+        (0xA648, "M", "ꙉ"),
+        (0xA649, "V"),
+        (0xA64A, "M", "ꙋ"),
+        (0xA64B, "V"),
+        (0xA64C, "M", "ꙍ"),
+        (0xA64D, "V"),
+        (0xA64E, "M", "ꙏ"),
+        (0xA64F, "V"),
+        (0xA650, "M", "ꙑ"),
+        (0xA651, "V"),
+        (0xA652, "M", "ꙓ"),
+        (0xA653, "V"),
+        (0xA654, "M", "ꙕ"),
+        (0xA655, "V"),
+        (0xA656, "M", "ꙗ"),
+        (0xA657, "V"),
+        (0xA658, "M", "ꙙ"),
+        (0xA659, "V"),
+        (0xA65A, "M", "ꙛ"),
+        (0xA65B, "V"),
+        (0xA65C, "M", "ꙝ"),
+        (0xA65D, "V"),
+        (0xA65E, "M", "ꙟ"),
+        (0xA65F, "V"),
+        (0xA660, "M", "ꙡ"),
+        (0xA661, "V"),
+        (0xA662, "M", "ꙣ"),
+        (0xA663, "V"),
+        (0xA664, "M", "ꙥ"),
+        (0xA665, "V"),
+        (0xA666, "M", "ꙧ"),
+        (0xA667, "V"),
+        (0xA668, "M", "ꙩ"),
+        (0xA669, "V"),
+        (0xA66A, "M", "ꙫ"),
+        (0xA66B, "V"),
+        (0xA66C, "M", "ꙭ"),
+        (0xA66D, "V"),
+        (0xA680, "M", "ꚁ"),
+        (0xA681, "V"),
+        (0xA682, "M", "ꚃ"),
+        (0xA683, "V"),
+        (0xA684, "M", "ꚅ"),
+        (0xA685, "V"),
+        (0xA686, "M", "ꚇ"),
+        (0xA687, "V"),
+        (0xA688, "M", "ꚉ"),
+        (0xA689, "V"),
+        (0xA68A, "M", "ꚋ"),
+        (0xA68B, "V"),
+        (0xA68C, "M", "ꚍ"),
+        (0xA68D, "V"),
+        (0xA68E, "M", "ꚏ"),
+        (0xA68F, "V"),
+        (0xA690, "M", "ꚑ"),
+        (0xA691, "V"),
+    ]
+
+
+def _seg_36() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA692, "M", "ꚓ"),
+        (0xA693, "V"),
+        (0xA694, "M", "ꚕ"),
+        (0xA695, "V"),
+        (0xA696, "M", "ꚗ"),
+        (0xA697, "V"),
+        (0xA698, "M", "ꚙ"),
+        (0xA699, "V"),
+        (0xA69A, "M", "ꚛ"),
+        (0xA69B, "V"),
+        (0xA69C, "M", "ъ"),
+        (0xA69D, "M", "ь"),
+        (0xA69E, "V"),
+        (0xA6F8, "X"),
+        (0xA700, "V"),
+        (0xA722, "M", "ꜣ"),
+        (0xA723, "V"),
+        (0xA724, "M", "ꜥ"),
+        (0xA725, "V"),
+        (0xA726, "M", "ꜧ"),
+        (0xA727, "V"),
+        (0xA728, "M", "ꜩ"),
+        (0xA729, "V"),
+        (0xA72A, "M", "ꜫ"),
+        (0xA72B, "V"),
+        (0xA72C, "M", "ꜭ"),
+        (0xA72D, "V"),
+        (0xA72E, "M", "ꜯ"),
+        (0xA72F, "V"),
+        (0xA732, "M", "ꜳ"),
+        (0xA733, "V"),
+        (0xA734, "M", "ꜵ"),
+        (0xA735, "V"),
+        (0xA736, "M", "ꜷ"),
+        (0xA737, "V"),
+        (0xA738, "M", "ꜹ"),
+        (0xA739, "V"),
+        (0xA73A, "M", "ꜻ"),
+        (0xA73B, "V"),
+        (0xA73C, "M", "ꜽ"),
+        (0xA73D, "V"),
+        (0xA73E, "M", "ꜿ"),
+        (0xA73F, "V"),
+        (0xA740, "M", "ꝁ"),
+        (0xA741, "V"),
+        (0xA742, "M", "ꝃ"),
+        (0xA743, "V"),
+        (0xA744, "M", "ꝅ"),
+        (0xA745, "V"),
+        (0xA746, "M", "ꝇ"),
+        (0xA747, "V"),
+        (0xA748, "M", "ꝉ"),
+        (0xA749, "V"),
+        (0xA74A, "M", "ꝋ"),
+        (0xA74B, "V"),
+        (0xA74C, "M", "ꝍ"),
+        (0xA74D, "V"),
+        (0xA74E, "M", "ꝏ"),
+        (0xA74F, "V"),
+        (0xA750, "M", "ꝑ"),
+        (0xA751, "V"),
+        (0xA752, "M", "ꝓ"),
+        (0xA753, "V"),
+        (0xA754, "M", "ꝕ"),
+        (0xA755, "V"),
+        (0xA756, "M", "ꝗ"),
+        (0xA757, "V"),
+        (0xA758, "M", "ꝙ"),
+        (0xA759, "V"),
+        (0xA75A, "M", "ꝛ"),
+        (0xA75B, "V"),
+        (0xA75C, "M", "ꝝ"),
+        (0xA75D, "V"),
+        (0xA75E, "M", "ꝟ"),
+        (0xA75F, "V"),
+        (0xA760, "M", "ꝡ"),
+        (0xA761, "V"),
+        (0xA762, "M", "ꝣ"),
+        (0xA763, "V"),
+        (0xA764, "M", "ꝥ"),
+        (0xA765, "V"),
+        (0xA766, "M", "ꝧ"),
+        (0xA767, "V"),
+        (0xA768, "M", "ꝩ"),
+        (0xA769, "V"),
+        (0xA76A, "M", "ꝫ"),
+        (0xA76B, "V"),
+        (0xA76C, "M", "ꝭ"),
+        (0xA76D, "V"),
+        (0xA76E, "M", "ꝯ"),
+        (0xA76F, "V"),
+        (0xA770, "M", "ꝯ"),
+        (0xA771, "V"),
+        (0xA779, "M", "ꝺ"),
+        (0xA77A, "V"),
+        (0xA77B, "M", "ꝼ"),
+        (0xA77C, "V"),
+        (0xA77D, "M", "ᵹ"),
+        (0xA77E, "M", "ꝿ"),
+        (0xA77F, "V"),
+    ]
+
+
+def _seg_37() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA780, "M", "ꞁ"),
+        (0xA781, "V"),
+        (0xA782, "M", "ꞃ"),
+        (0xA783, "V"),
+        (0xA784, "M", "ꞅ"),
+        (0xA785, "V"),
+        (0xA786, "M", "ꞇ"),
+        (0xA787, "V"),
+        (0xA78B, "M", "ꞌ"),
+        (0xA78C, "V"),
+        (0xA78D, "M", "ɥ"),
+        (0xA78E, "V"),
+        (0xA790, "M", "ꞑ"),
+        (0xA791, "V"),
+        (0xA792, "M", "ꞓ"),
+        (0xA793, "V"),
+        (0xA796, "M", "ꞗ"),
+        (0xA797, "V"),
+        (0xA798, "M", "ꞙ"),
+        (0xA799, "V"),
+        (0xA79A, "M", "ꞛ"),
+        (0xA79B, "V"),
+        (0xA79C, "M", "ꞝ"),
+        (0xA79D, "V"),
+        (0xA79E, "M", "ꞟ"),
+        (0xA79F, "V"),
+        (0xA7A0, "M", "ꞡ"),
+        (0xA7A1, "V"),
+        (0xA7A2, "M", "ꞣ"),
+        (0xA7A3, "V"),
+        (0xA7A4, "M", "ꞥ"),
+        (0xA7A5, "V"),
+        (0xA7A6, "M", "ꞧ"),
+        (0xA7A7, "V"),
+        (0xA7A8, "M", "ꞩ"),
+        (0xA7A9, "V"),
+        (0xA7AA, "M", "ɦ"),
+        (0xA7AB, "M", "ɜ"),
+        (0xA7AC, "M", "ɡ"),
+        (0xA7AD, "M", "ɬ"),
+        (0xA7AE, "M", "ɪ"),
+        (0xA7AF, "V"),
+        (0xA7B0, "M", "ʞ"),
+        (0xA7B1, "M", "ʇ"),
+        (0xA7B2, "M", "ʝ"),
+        (0xA7B3, "M", "ꭓ"),
+        (0xA7B4, "M", "ꞵ"),
+        (0xA7B5, "V"),
+        (0xA7B6, "M", "ꞷ"),
+        (0xA7B7, "V"),
+        (0xA7B8, "M", "ꞹ"),
+        (0xA7B9, "V"),
+        (0xA7BA, "M", "ꞻ"),
+        (0xA7BB, "V"),
+        (0xA7BC, "M", "ꞽ"),
+        (0xA7BD, "V"),
+        (0xA7BE, "M", "ꞿ"),
+        (0xA7BF, "V"),
+        (0xA7C0, "M", "ꟁ"),
+        (0xA7C1, "V"),
+        (0xA7C2, "M", "ꟃ"),
+        (0xA7C3, "V"),
+        (0xA7C4, "M", "ꞔ"),
+        (0xA7C5, "M", "ʂ"),
+        (0xA7C6, "M", "ᶎ"),
+        (0xA7C7, "M", "ꟈ"),
+        (0xA7C8, "V"),
+        (0xA7C9, "M", "ꟊ"),
+        (0xA7CA, "V"),
+        (0xA7CB, "X"),
+        (0xA7D0, "M", "ꟑ"),
+        (0xA7D1, "V"),
+        (0xA7D2, "X"),
+        (0xA7D3, "V"),
+        (0xA7D4, "X"),
+        (0xA7D5, "V"),
+        (0xA7D6, "M", "ꟗ"),
+        (0xA7D7, "V"),
+        (0xA7D8, "M", "ꟙ"),
+        (0xA7D9, "V"),
+        (0xA7DA, "X"),
+        (0xA7F2, "M", "c"),
+        (0xA7F3, "M", "f"),
+        (0xA7F4, "M", "q"),
+        (0xA7F5, "M", "ꟶ"),
+        (0xA7F6, "V"),
+        (0xA7F8, "M", "ħ"),
+        (0xA7F9, "M", "œ"),
+        (0xA7FA, "V"),
+        (0xA82D, "X"),
+        (0xA830, "V"),
+        (0xA83A, "X"),
+        (0xA840, "V"),
+        (0xA878, "X"),
+        (0xA880, "V"),
+        (0xA8C6, "X"),
+        (0xA8CE, "V"),
+        (0xA8DA, "X"),
+        (0xA8E0, "V"),
+        (0xA954, "X"),
+    ]
+
+
+def _seg_38() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA95F, "V"),
+        (0xA97D, "X"),
+        (0xA980, "V"),
+        (0xA9CE, "X"),
+        (0xA9CF, "V"),
+        (0xA9DA, "X"),
+        (0xA9DE, "V"),
+        (0xA9FF, "X"),
+        (0xAA00, "V"),
+        (0xAA37, "X"),
+        (0xAA40, "V"),
+        (0xAA4E, "X"),
+        (0xAA50, "V"),
+        (0xAA5A, "X"),
+        (0xAA5C, "V"),
+        (0xAAC3, "X"),
+        (0xAADB, "V"),
+        (0xAAF7, "X"),
+        (0xAB01, "V"),
+        (0xAB07, "X"),
+        (0xAB09, "V"),
+        (0xAB0F, "X"),
+        (0xAB11, "V"),
+        (0xAB17, "X"),
+        (0xAB20, "V"),
+        (0xAB27, "X"),
+        (0xAB28, "V"),
+        (0xAB2F, "X"),
+        (0xAB30, "V"),
+        (0xAB5C, "M", "ꜧ"),
+        (0xAB5D, "M", "ꬷ"),
+        (0xAB5E, "M", "ɫ"),
+        (0xAB5F, "M", "ꭒ"),
+        (0xAB60, "V"),
+        (0xAB69, "M", "ʍ"),
+        (0xAB6A, "V"),
+        (0xAB6C, "X"),
+        (0xAB70, "M", "Ꭰ"),
+        (0xAB71, "M", "Ꭱ"),
+        (0xAB72, "M", "Ꭲ"),
+        (0xAB73, "M", "Ꭳ"),
+        (0xAB74, "M", "Ꭴ"),
+        (0xAB75, "M", "Ꭵ"),
+        (0xAB76, "M", "Ꭶ"),
+        (0xAB77, "M", "Ꭷ"),
+        (0xAB78, "M", "Ꭸ"),
+        (0xAB79, "M", "Ꭹ"),
+        (0xAB7A, "M", "Ꭺ"),
+        (0xAB7B, "M", "Ꭻ"),
+        (0xAB7C, "M", "Ꭼ"),
+        (0xAB7D, "M", "Ꭽ"),
+        (0xAB7E, "M", "Ꭾ"),
+        (0xAB7F, "M", "Ꭿ"),
+        (0xAB80, "M", "Ꮀ"),
+        (0xAB81, "M", "Ꮁ"),
+        (0xAB82, "M", "Ꮂ"),
+        (0xAB83, "M", "Ꮃ"),
+        (0xAB84, "M", "Ꮄ"),
+        (0xAB85, "M", "Ꮅ"),
+        (0xAB86, "M", "Ꮆ"),
+        (0xAB87, "M", "Ꮇ"),
+        (0xAB88, "M", "Ꮈ"),
+        (0xAB89, "M", "Ꮉ"),
+        (0xAB8A, "M", "Ꮊ"),
+        (0xAB8B, "M", "Ꮋ"),
+        (0xAB8C, "M", "Ꮌ"),
+        (0xAB8D, "M", "Ꮍ"),
+        (0xAB8E, "M", "Ꮎ"),
+        (0xAB8F, "M", "Ꮏ"),
+        (0xAB90, "M", "Ꮐ"),
+        (0xAB91, "M", "Ꮑ"),
+        (0xAB92, "M", "Ꮒ"),
+        (0xAB93, "M", "Ꮓ"),
+        (0xAB94, "M", "Ꮔ"),
+        (0xAB95, "M", "Ꮕ"),
+        (0xAB96, "M", "Ꮖ"),
+        (0xAB97, "M", "Ꮗ"),
+        (0xAB98, "M", "Ꮘ"),
+        (0xAB99, "M", "Ꮙ"),
+        (0xAB9A, "M", "Ꮚ"),
+        (0xAB9B, "M", "Ꮛ"),
+        (0xAB9C, "M", "Ꮜ"),
+        (0xAB9D, "M", "Ꮝ"),
+        (0xAB9E, "M", "Ꮞ"),
+        (0xAB9F, "M", "Ꮟ"),
+        (0xABA0, "M", "Ꮠ"),
+        (0xABA1, "M", "Ꮡ"),
+        (0xABA2, "M", "Ꮢ"),
+        (0xABA3, "M", "Ꮣ"),
+        (0xABA4, "M", "Ꮤ"),
+        (0xABA5, "M", "Ꮥ"),
+        (0xABA6, "M", "Ꮦ"),
+        (0xABA7, "M", "Ꮧ"),
+        (0xABA8, "M", "Ꮨ"),
+        (0xABA9, "M", "Ꮩ"),
+        (0xABAA, "M", "Ꮪ"),
+        (0xABAB, "M", "Ꮫ"),
+        (0xABAC, "M", "Ꮬ"),
+        (0xABAD, "M", "Ꮭ"),
+        (0xABAE, "M", "Ꮮ"),
+    ]
+
+
+def _seg_39() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xABAF, "M", "Ꮯ"),
+        (0xABB0, "M", "Ꮰ"),
+        (0xABB1, "M", "Ꮱ"),
+        (0xABB2, "M", "Ꮲ"),
+        (0xABB3, "M", "Ꮳ"),
+        (0xABB4, "M", "Ꮴ"),
+        (0xABB5, "M", "Ꮵ"),
+        (0xABB6, "M", "Ꮶ"),
+        (0xABB7, "M", "Ꮷ"),
+        (0xABB8, "M", "Ꮸ"),
+        (0xABB9, "M", "Ꮹ"),
+        (0xABBA, "M", "Ꮺ"),
+        (0xABBB, "M", "Ꮻ"),
+        (0xABBC, "M", "Ꮼ"),
+        (0xABBD, "M", "Ꮽ"),
+        (0xABBE, "M", "Ꮾ"),
+        (0xABBF, "M", "Ꮿ"),
+        (0xABC0, "V"),
+        (0xABEE, "X"),
+        (0xABF0, "V"),
+        (0xABFA, "X"),
+        (0xAC00, "V"),
+        (0xD7A4, "X"),
+        (0xD7B0, "V"),
+        (0xD7C7, "X"),
+        (0xD7CB, "V"),
+        (0xD7FC, "X"),
+        (0xF900, "M", "豈"),
+        (0xF901, "M", "更"),
+        (0xF902, "M", "車"),
+        (0xF903, "M", "賈"),
+        (0xF904, "M", "滑"),
+        (0xF905, "M", "串"),
+        (0xF906, "M", "句"),
+        (0xF907, "M", "龜"),
+        (0xF909, "M", "契"),
+        (0xF90A, "M", "金"),
+        (0xF90B, "M", "喇"),
+        (0xF90C, "M", "奈"),
+        (0xF90D, "M", "懶"),
+        (0xF90E, "M", "癩"),
+        (0xF90F, "M", "羅"),
+        (0xF910, "M", "蘿"),
+        (0xF911, "M", "螺"),
+        (0xF912, "M", "裸"),
+        (0xF913, "M", "邏"),
+        (0xF914, "M", "樂"),
+        (0xF915, "M", "洛"),
+        (0xF916, "M", "烙"),
+        (0xF917, "M", "珞"),
+        (0xF918, "M", "落"),
+        (0xF919, "M", "酪"),
+        (0xF91A, "M", "駱"),
+        (0xF91B, "M", "亂"),
+        (0xF91C, "M", "卵"),
+        (0xF91D, "M", "欄"),
+        (0xF91E, "M", "爛"),
+        (0xF91F, "M", "蘭"),
+        (0xF920, "M", "鸞"),
+        (0xF921, "M", "嵐"),
+        (0xF922, "M", "濫"),
+        (0xF923, "M", "藍"),
+        (0xF924, "M", "襤"),
+        (0xF925, "M", "拉"),
+        (0xF926, "M", "臘"),
+        (0xF927, "M", "蠟"),
+        (0xF928, "M", "廊"),
+        (0xF929, "M", "朗"),
+        (0xF92A, "M", "浪"),
+        (0xF92B, "M", "狼"),
+        (0xF92C, "M", "郎"),
+        (0xF92D, "M", "來"),
+        (0xF92E, "M", "冷"),
+        (0xF92F, "M", "勞"),
+        (0xF930, "M", "擄"),
+        (0xF931, "M", "櫓"),
+        (0xF932, "M", "爐"),
+        (0xF933, "M", "盧"),
+        (0xF934, "M", "老"),
+        (0xF935, "M", "蘆"),
+        (0xF936, "M", "虜"),
+        (0xF937, "M", "路"),
+        (0xF938, "M", "露"),
+        (0xF939, "M", "魯"),
+        (0xF93A, "M", "鷺"),
+        (0xF93B, "M", "碌"),
+        (0xF93C, "M", "祿"),
+        (0xF93D, "M", "綠"),
+        (0xF93E, "M", "菉"),
+        (0xF93F, "M", "錄"),
+        (0xF940, "M", "鹿"),
+        (0xF941, "M", "論"),
+        (0xF942, "M", "壟"),
+        (0xF943, "M", "弄"),
+        (0xF944, "M", "籠"),
+        (0xF945, "M", "聾"),
+        (0xF946, "M", "牢"),
+        (0xF947, "M", "磊"),
+        (0xF948, "M", "賂"),
+        (0xF949, "M", "雷"),
+    ]
+
+
+def _seg_40() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xF94A, "M", "壘"),
+        (0xF94B, "M", "屢"),
+        (0xF94C, "M", "樓"),
+        (0xF94D, "M", "淚"),
+        (0xF94E, "M", "漏"),
+        (0xF94F, "M", "累"),
+        (0xF950, "M", "縷"),
+        (0xF951, "M", "陋"),
+        (0xF952, "M", "勒"),
+        (0xF953, "M", "肋"),
+        (0xF954, "M", "凜"),
+        (0xF955, "M", "凌"),
+        (0xF956, "M", "稜"),
+        (0xF957, "M", "綾"),
+        (0xF958, "M", "菱"),
+        (0xF959, "M", "陵"),
+        (0xF95A, "M", "讀"),
+        (0xF95B, "M", "拏"),
+        (0xF95C, "M", "樂"),
+        (0xF95D, "M", "諾"),
+        (0xF95E, "M", "丹"),
+        (0xF95F, "M", "寧"),
+        (0xF960, "M", "怒"),
+        (0xF961, "M", "率"),
+        (0xF962, "M", "異"),
+        (0xF963, "M", "北"),
+        (0xF964, "M", "磻"),
+        (0xF965, "M", "便"),
+        (0xF966, "M", "復"),
+        (0xF967, "M", "不"),
+        (0xF968, "M", "泌"),
+        (0xF969, "M", "數"),
+        (0xF96A, "M", "索"),
+        (0xF96B, "M", "參"),
+        (0xF96C, "M", "塞"),
+        (0xF96D, "M", "省"),
+        (0xF96E, "M", "葉"),
+        (0xF96F, "M", "說"),
+        (0xF970, "M", "殺"),
+        (0xF971, "M", "辰"),
+        (0xF972, "M", "沈"),
+        (0xF973, "M", "拾"),
+        (0xF974, "M", "若"),
+        (0xF975, "M", "掠"),
+        (0xF976, "M", "略"),
+        (0xF977, "M", "亮"),
+        (0xF978, "M", "兩"),
+        (0xF979, "M", "凉"),
+        (0xF97A, "M", "梁"),
+        (0xF97B, "M", "糧"),
+        (0xF97C, "M", "良"),
+        (0xF97D, "M", "諒"),
+        (0xF97E, "M", "量"),
+        (0xF97F, "M", "勵"),
+        (0xF980, "M", "呂"),
+        (0xF981, "M", "女"),
+        (0xF982, "M", "廬"),
+        (0xF983, "M", "旅"),
+        (0xF984, "M", "濾"),
+        (0xF985, "M", "礪"),
+        (0xF986, "M", "閭"),
+        (0xF987, "M", "驪"),
+        (0xF988, "M", "麗"),
+        (0xF989, "M", "黎"),
+        (0xF98A, "M", "力"),
+        (0xF98B, "M", "曆"),
+        (0xF98C, "M", "歷"),
+        (0xF98D, "M", "轢"),
+        (0xF98E, "M", "年"),
+        (0xF98F, "M", "憐"),
+        (0xF990, "M", "戀"),
+        (0xF991, "M", "撚"),
+        (0xF992, "M", "漣"),
+        (0xF993, "M", "煉"),
+        (0xF994, "M", "璉"),
+        (0xF995, "M", "秊"),
+        (0xF996, "M", "練"),
+        (0xF997, "M", "聯"),
+        (0xF998, "M", "輦"),
+        (0xF999, "M", "蓮"),
+        (0xF99A, "M", "連"),
+        (0xF99B, "M", "鍊"),
+        (0xF99C, "M", "列"),
+        (0xF99D, "M", "劣"),
+        (0xF99E, "M", "咽"),
+        (0xF99F, "M", "烈"),
+        (0xF9A0, "M", "裂"),
+        (0xF9A1, "M", "說"),
+        (0xF9A2, "M", "廉"),
+        (0xF9A3, "M", "念"),
+        (0xF9A4, "M", "捻"),
+        (0xF9A5, "M", "殮"),
+        (0xF9A6, "M", "簾"),
+        (0xF9A7, "M", "獵"),
+        (0xF9A8, "M", "令"),
+        (0xF9A9, "M", "囹"),
+        (0xF9AA, "M", "寧"),
+        (0xF9AB, "M", "嶺"),
+        (0xF9AC, "M", "怜"),
+        (0xF9AD, "M", "玲"),
+    ]
+
+
+def _seg_41() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xF9AE, "M", "瑩"),
+        (0xF9AF, "M", "羚"),
+        (0xF9B0, "M", "聆"),
+        (0xF9B1, "M", "鈴"),
+        (0xF9B2, "M", "零"),
+        (0xF9B3, "M", "靈"),
+        (0xF9B4, "M", "領"),
+        (0xF9B5, "M", "例"),
+        (0xF9B6, "M", "禮"),
+        (0xF9B7, "M", "醴"),
+        (0xF9B8, "M", "隸"),
+        (0xF9B9, "M", "惡"),
+        (0xF9BA, "M", "了"),
+        (0xF9BB, "M", "僚"),
+        (0xF9BC, "M", "寮"),
+        (0xF9BD, "M", "尿"),
+        (0xF9BE, "M", "料"),
+        (0xF9BF, "M", "樂"),
+        (0xF9C0, "M", "燎"),
+        (0xF9C1, "M", "療"),
+        (0xF9C2, "M", "蓼"),
+        (0xF9C3, "M", "遼"),
+        (0xF9C4, "M", "龍"),
+        (0xF9C5, "M", "暈"),
+        (0xF9C6, "M", "阮"),
+        (0xF9C7, "M", "劉"),
+        (0xF9C8, "M", "杻"),
+        (0xF9C9, "M", "柳"),
+        (0xF9CA, "M", "流"),
+        (0xF9CB, "M", "溜"),
+        (0xF9CC, "M", "琉"),
+        (0xF9CD, "M", "留"),
+        (0xF9CE, "M", "硫"),
+        (0xF9CF, "M", "紐"),
+        (0xF9D0, "M", "類"),
+        (0xF9D1, "M", "六"),
+        (0xF9D2, "M", "戮"),
+        (0xF9D3, "M", "陸"),
+        (0xF9D4, "M", "倫"),
+        (0xF9D5, "M", "崙"),
+        (0xF9D6, "M", "淪"),
+        (0xF9D7, "M", "輪"),
+        (0xF9D8, "M", "律"),
+        (0xF9D9, "M", "慄"),
+        (0xF9DA, "M", "栗"),
+        (0xF9DB, "M", "率"),
+        (0xF9DC, "M", "隆"),
+        (0xF9DD, "M", "利"),
+        (0xF9DE, "M", "吏"),
+        (0xF9DF, "M", "履"),
+        (0xF9E0, "M", "易"),
+        (0xF9E1, "M", "李"),
+        (0xF9E2, "M", "梨"),
+        (0xF9E3, "M", "泥"),
+        (0xF9E4, "M", "理"),
+        (0xF9E5, "M", "痢"),
+        (0xF9E6, "M", "罹"),
+        (0xF9E7, "M", "裏"),
+        (0xF9E8, "M", "裡"),
+        (0xF9E9, "M", "里"),
+        (0xF9EA, "M", "離"),
+        (0xF9EB, "M", "匿"),
+        (0xF9EC, "M", "溺"),
+        (0xF9ED, "M", "吝"),
+        (0xF9EE, "M", "燐"),
+        (0xF9EF, "M", "璘"),
+        (0xF9F0, "M", "藺"),
+        (0xF9F1, "M", "隣"),
+        (0xF9F2, "M", "鱗"),
+        (0xF9F3, "M", "麟"),
+        (0xF9F4, "M", "林"),
+        (0xF9F5, "M", "淋"),
+        (0xF9F6, "M", "臨"),
+        (0xF9F7, "M", "立"),
+        (0xF9F8, "M", "笠"),
+        (0xF9F9, "M", "粒"),
+        (0xF9FA, "M", "狀"),
+        (0xF9FB, "M", "炙"),
+        (0xF9FC, "M", "識"),
+        (0xF9FD, "M", "什"),
+        (0xF9FE, "M", "茶"),
+        (0xF9FF, "M", "刺"),
+        (0xFA00, "M", "切"),
+        (0xFA01, "M", "度"),
+        (0xFA02, "M", "拓"),
+        (0xFA03, "M", "糖"),
+        (0xFA04, "M", "宅"),
+        (0xFA05, "M", "洞"),
+        (0xFA06, "M", "暴"),
+        (0xFA07, "M", "輻"),
+        (0xFA08, "M", "行"),
+        (0xFA09, "M", "降"),
+        (0xFA0A, "M", "見"),
+        (0xFA0B, "M", "廓"),
+        (0xFA0C, "M", "兀"),
+        (0xFA0D, "M", "嗀"),
+        (0xFA0E, "V"),
+        (0xFA10, "M", "塚"),
+        (0xFA11, "V"),
+        (0xFA12, "M", "晴"),
+    ]
+
+
+def _seg_42() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFA13, "V"),
+        (0xFA15, "M", "凞"),
+        (0xFA16, "M", "猪"),
+        (0xFA17, "M", "益"),
+        (0xFA18, "M", "礼"),
+        (0xFA19, "M", "神"),
+        (0xFA1A, "M", "祥"),
+        (0xFA1B, "M", "福"),
+        (0xFA1C, "M", "靖"),
+        (0xFA1D, "M", "精"),
+        (0xFA1E, "M", "羽"),
+        (0xFA1F, "V"),
+        (0xFA20, "M", "蘒"),
+        (0xFA21, "V"),
+        (0xFA22, "M", "諸"),
+        (0xFA23, "V"),
+        (0xFA25, "M", "逸"),
+        (0xFA26, "M", "都"),
+        (0xFA27, "V"),
+        (0xFA2A, "M", "飯"),
+        (0xFA2B, "M", "飼"),
+        (0xFA2C, "M", "館"),
+        (0xFA2D, "M", "鶴"),
+        (0xFA2E, "M", "郞"),
+        (0xFA2F, "M", "隷"),
+        (0xFA30, "M", "侮"),
+        (0xFA31, "M", "僧"),
+        (0xFA32, "M", "免"),
+        (0xFA33, "M", "勉"),
+        (0xFA34, "M", "勤"),
+        (0xFA35, "M", "卑"),
+        (0xFA36, "M", "喝"),
+        (0xFA37, "M", "嘆"),
+        (0xFA38, "M", "器"),
+        (0xFA39, "M", "塀"),
+        (0xFA3A, "M", "墨"),
+        (0xFA3B, "M", "層"),
+        (0xFA3C, "M", "屮"),
+        (0xFA3D, "M", "悔"),
+        (0xFA3E, "M", "慨"),
+        (0xFA3F, "M", "憎"),
+        (0xFA40, "M", "懲"),
+        (0xFA41, "M", "敏"),
+        (0xFA42, "M", "既"),
+        (0xFA43, "M", "暑"),
+        (0xFA44, "M", "梅"),
+        (0xFA45, "M", "海"),
+        (0xFA46, "M", "渚"),
+        (0xFA47, "M", "漢"),
+        (0xFA48, "M", "煮"),
+        (0xFA49, "M", "爫"),
+        (0xFA4A, "M", "琢"),
+        (0xFA4B, "M", "碑"),
+        (0xFA4C, "M", "社"),
+        (0xFA4D, "M", "祉"),
+        (0xFA4E, "M", "祈"),
+        (0xFA4F, "M", "祐"),
+        (0xFA50, "M", "祖"),
+        (0xFA51, "M", "祝"),
+        (0xFA52, "M", "禍"),
+        (0xFA53, "M", "禎"),
+        (0xFA54, "M", "穀"),
+        (0xFA55, "M", "突"),
+        (0xFA56, "M", "節"),
+        (0xFA57, "M", "練"),
+        (0xFA58, "M", "縉"),
+        (0xFA59, "M", "繁"),
+        (0xFA5A, "M", "署"),
+        (0xFA5B, "M", "者"),
+        (0xFA5C, "M", "臭"),
+        (0xFA5D, "M", "艹"),
+        (0xFA5F, "M", "著"),
+        (0xFA60, "M", "褐"),
+        (0xFA61, "M", "視"),
+        (0xFA62, "M", "謁"),
+        (0xFA63, "M", "謹"),
+        (0xFA64, "M", "賓"),
+        (0xFA65, "M", "贈"),
+        (0xFA66, "M", "辶"),
+        (0xFA67, "M", "逸"),
+        (0xFA68, "M", "難"),
+        (0xFA69, "M", "響"),
+        (0xFA6A, "M", "頻"),
+        (0xFA6B, "M", "恵"),
+        (0xFA6C, "M", "𤋮"),
+        (0xFA6D, "M", "舘"),
+        (0xFA6E, "X"),
+        (0xFA70, "M", "並"),
+        (0xFA71, "M", "况"),
+        (0xFA72, "M", "全"),
+        (0xFA73, "M", "侀"),
+        (0xFA74, "M", "充"),
+        (0xFA75, "M", "冀"),
+        (0xFA76, "M", "勇"),
+        (0xFA77, "M", "勺"),
+        (0xFA78, "M", "喝"),
+        (0xFA79, "M", "啕"),
+        (0xFA7A, "M", "喙"),
+        (0xFA7B, "M", "嗢"),
+        (0xFA7C, "M", "塚"),
+    ]
+
+
+def _seg_43() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFA7D, "M", "墳"),
+        (0xFA7E, "M", "奄"),
+        (0xFA7F, "M", "奔"),
+        (0xFA80, "M", "婢"),
+        (0xFA81, "M", "嬨"),
+        (0xFA82, "M", "廒"),
+        (0xFA83, "M", "廙"),
+        (0xFA84, "M", "彩"),
+        (0xFA85, "M", "徭"),
+        (0xFA86, "M", "惘"),
+        (0xFA87, "M", "慎"),
+        (0xFA88, "M", "愈"),
+        (0xFA89, "M", "憎"),
+        (0xFA8A, "M", "慠"),
+        (0xFA8B, "M", "懲"),
+        (0xFA8C, "M", "戴"),
+        (0xFA8D, "M", "揄"),
+        (0xFA8E, "M", "搜"),
+        (0xFA8F, "M", "摒"),
+        (0xFA90, "M", "敖"),
+        (0xFA91, "M", "晴"),
+        (0xFA92, "M", "朗"),
+        (0xFA93, "M", "望"),
+        (0xFA94, "M", "杖"),
+        (0xFA95, "M", "歹"),
+        (0xFA96, "M", "殺"),
+        (0xFA97, "M", "流"),
+        (0xFA98, "M", "滛"),
+        (0xFA99, "M", "滋"),
+        (0xFA9A, "M", "漢"),
+        (0xFA9B, "M", "瀞"),
+        (0xFA9C, "M", "煮"),
+        (0xFA9D, "M", "瞧"),
+        (0xFA9E, "M", "爵"),
+        (0xFA9F, "M", "犯"),
+        (0xFAA0, "M", "猪"),
+        (0xFAA1, "M", "瑱"),
+        (0xFAA2, "M", "甆"),
+        (0xFAA3, "M", "画"),
+        (0xFAA4, "M", "瘝"),
+        (0xFAA5, "M", "瘟"),
+        (0xFAA6, "M", "益"),
+        (0xFAA7, "M", "盛"),
+        (0xFAA8, "M", "直"),
+        (0xFAA9, "M", "睊"),
+        (0xFAAA, "M", "着"),
+        (0xFAAB, "M", "磌"),
+        (0xFAAC, "M", "窱"),
+        (0xFAAD, "M", "節"),
+        (0xFAAE, "M", "类"),
+        (0xFAAF, "M", "絛"),
+        (0xFAB0, "M", "練"),
+        (0xFAB1, "M", "缾"),
+        (0xFAB2, "M", "者"),
+        (0xFAB3, "M", "荒"),
+        (0xFAB4, "M", "華"),
+        (0xFAB5, "M", "蝹"),
+        (0xFAB6, "M", "襁"),
+        (0xFAB7, "M", "覆"),
+        (0xFAB8, "M", "視"),
+        (0xFAB9, "M", "調"),
+        (0xFABA, "M", "諸"),
+        (0xFABB, "M", "請"),
+        (0xFABC, "M", "謁"),
+        (0xFABD, "M", "諾"),
+        (0xFABE, "M", "諭"),
+        (0xFABF, "M", "謹"),
+        (0xFAC0, "M", "變"),
+        (0xFAC1, "M", "贈"),
+        (0xFAC2, "M", "輸"),
+        (0xFAC3, "M", "遲"),
+        (0xFAC4, "M", "醙"),
+        (0xFAC5, "M", "鉶"),
+        (0xFAC6, "M", "陼"),
+        (0xFAC7, "M", "難"),
+        (0xFAC8, "M", "靖"),
+        (0xFAC9, "M", "韛"),
+        (0xFACA, "M", "響"),
+        (0xFACB, "M", "頋"),
+        (0xFACC, "M", "頻"),
+        (0xFACD, "M", "鬒"),
+        (0xFACE, "M", "龜"),
+        (0xFACF, "M", "𢡊"),
+        (0xFAD0, "M", "𢡄"),
+        (0xFAD1, "M", "𣏕"),
+        (0xFAD2, "M", "㮝"),
+        (0xFAD3, "M", "䀘"),
+        (0xFAD4, "M", "䀹"),
+        (0xFAD5, "M", "𥉉"),
+        (0xFAD6, "M", "𥳐"),
+        (0xFAD7, "M", "𧻓"),
+        (0xFAD8, "M", "齃"),
+        (0xFAD9, "M", "龎"),
+        (0xFADA, "X"),
+        (0xFB00, "M", "ff"),
+        (0xFB01, "M", "fi"),
+        (0xFB02, "M", "fl"),
+        (0xFB03, "M", "ffi"),
+        (0xFB04, "M", "ffl"),
+        (0xFB05, "M", "st"),
+    ]
+
+
+def _seg_44() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFB07, "X"),
+        (0xFB13, "M", "մն"),
+        (0xFB14, "M", "մե"),
+        (0xFB15, "M", "մի"),
+        (0xFB16, "M", "վն"),
+        (0xFB17, "M", "մխ"),
+        (0xFB18, "X"),
+        (0xFB1D, "M", "יִ"),
+        (0xFB1E, "V"),
+        (0xFB1F, "M", "ײַ"),
+        (0xFB20, "M", "ע"),
+        (0xFB21, "M", "א"),
+        (0xFB22, "M", "ד"),
+        (0xFB23, "M", "ה"),
+        (0xFB24, "M", "כ"),
+        (0xFB25, "M", "ל"),
+        (0xFB26, "M", "ם"),
+        (0xFB27, "M", "ר"),
+        (0xFB28, "M", "ת"),
+        (0xFB29, "3", "+"),
+        (0xFB2A, "M", "שׁ"),
+        (0xFB2B, "M", "שׂ"),
+        (0xFB2C, "M", "שּׁ"),
+        (0xFB2D, "M", "שּׂ"),
+        (0xFB2E, "M", "אַ"),
+        (0xFB2F, "M", "אָ"),
+        (0xFB30, "M", "אּ"),
+        (0xFB31, "M", "בּ"),
+        (0xFB32, "M", "גּ"),
+        (0xFB33, "M", "דּ"),
+        (0xFB34, "M", "הּ"),
+        (0xFB35, "M", "וּ"),
+        (0xFB36, "M", "זּ"),
+        (0xFB37, "X"),
+        (0xFB38, "M", "טּ"),
+        (0xFB39, "M", "יּ"),
+        (0xFB3A, "M", "ךּ"),
+        (0xFB3B, "M", "כּ"),
+        (0xFB3C, "M", "לּ"),
+        (0xFB3D, "X"),
+        (0xFB3E, "M", "מּ"),
+        (0xFB3F, "X"),
+        (0xFB40, "M", "נּ"),
+        (0xFB41, "M", "סּ"),
+        (0xFB42, "X"),
+        (0xFB43, "M", "ףּ"),
+        (0xFB44, "M", "פּ"),
+        (0xFB45, "X"),
+        (0xFB46, "M", "צּ"),
+        (0xFB47, "M", "קּ"),
+        (0xFB48, "M", "רּ"),
+        (0xFB49, "M", "שּ"),
+        (0xFB4A, "M", "תּ"),
+        (0xFB4B, "M", "וֹ"),
+        (0xFB4C, "M", "בֿ"),
+        (0xFB4D, "M", "כֿ"),
+        (0xFB4E, "M", "פֿ"),
+        (0xFB4F, "M", "אל"),
+        (0xFB50, "M", "ٱ"),
+        (0xFB52, "M", "ٻ"),
+        (0xFB56, "M", "پ"),
+        (0xFB5A, "M", "ڀ"),
+        (0xFB5E, "M", "ٺ"),
+        (0xFB62, "M", "ٿ"),
+        (0xFB66, "M", "ٹ"),
+        (0xFB6A, "M", "ڤ"),
+        (0xFB6E, "M", "ڦ"),
+        (0xFB72, "M", "ڄ"),
+        (0xFB76, "M", "ڃ"),
+        (0xFB7A, "M", "چ"),
+        (0xFB7E, "M", "ڇ"),
+        (0xFB82, "M", "ڍ"),
+        (0xFB84, "M", "ڌ"),
+        (0xFB86, "M", "ڎ"),
+        (0xFB88, "M", "ڈ"),
+        (0xFB8A, "M", "ژ"),
+        (0xFB8C, "M", "ڑ"),
+        (0xFB8E, "M", "ک"),
+        (0xFB92, "M", "گ"),
+        (0xFB96, "M", "ڳ"),
+        (0xFB9A, "M", "ڱ"),
+        (0xFB9E, "M", "ں"),
+        (0xFBA0, "M", "ڻ"),
+        (0xFBA4, "M", "ۀ"),
+        (0xFBA6, "M", "ہ"),
+        (0xFBAA, "M", "ھ"),
+        (0xFBAE, "M", "ے"),
+        (0xFBB0, "M", "ۓ"),
+        (0xFBB2, "V"),
+        (0xFBC3, "X"),
+        (0xFBD3, "M", "ڭ"),
+        (0xFBD7, "M", "ۇ"),
+        (0xFBD9, "M", "ۆ"),
+        (0xFBDB, "M", "ۈ"),
+        (0xFBDD, "M", "ۇٴ"),
+        (0xFBDE, "M", "ۋ"),
+        (0xFBE0, "M", "ۅ"),
+        (0xFBE2, "M", "ۉ"),
+        (0xFBE4, "M", "ې"),
+        (0xFBE8, "M", "ى"),
+    ]
+
+
+def _seg_45() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFBEA, "M", "ئا"),
+        (0xFBEC, "M", "ئە"),
+        (0xFBEE, "M", "ئو"),
+        (0xFBF0, "M", "ئۇ"),
+        (0xFBF2, "M", "ئۆ"),
+        (0xFBF4, "M", "ئۈ"),
+        (0xFBF6, "M", "ئې"),
+        (0xFBF9, "M", "ئى"),
+        (0xFBFC, "M", "ی"),
+        (0xFC00, "M", "ئج"),
+        (0xFC01, "M", "ئح"),
+        (0xFC02, "M", "ئم"),
+        (0xFC03, "M", "ئى"),
+        (0xFC04, "M", "ئي"),
+        (0xFC05, "M", "بج"),
+        (0xFC06, "M", "بح"),
+        (0xFC07, "M", "بخ"),
+        (0xFC08, "M", "بم"),
+        (0xFC09, "M", "بى"),
+        (0xFC0A, "M", "بي"),
+        (0xFC0B, "M", "تج"),
+        (0xFC0C, "M", "تح"),
+        (0xFC0D, "M", "تخ"),
+        (0xFC0E, "M", "تم"),
+        (0xFC0F, "M", "تى"),
+        (0xFC10, "M", "تي"),
+        (0xFC11, "M", "ثج"),
+        (0xFC12, "M", "ثم"),
+        (0xFC13, "M", "ثى"),
+        (0xFC14, "M", "ثي"),
+        (0xFC15, "M", "جح"),
+        (0xFC16, "M", "جم"),
+        (0xFC17, "M", "حج"),
+        (0xFC18, "M", "حم"),
+        (0xFC19, "M", "خج"),
+        (0xFC1A, "M", "خح"),
+        (0xFC1B, "M", "خم"),
+        (0xFC1C, "M", "سج"),
+        (0xFC1D, "M", "سح"),
+        (0xFC1E, "M", "سخ"),
+        (0xFC1F, "M", "سم"),
+        (0xFC20, "M", "صح"),
+        (0xFC21, "M", "صم"),
+        (0xFC22, "M", "ضج"),
+        (0xFC23, "M", "ضح"),
+        (0xFC24, "M", "ضخ"),
+        (0xFC25, "M", "ضم"),
+        (0xFC26, "M", "طح"),
+        (0xFC27, "M", "طم"),
+        (0xFC28, "M", "ظم"),
+        (0xFC29, "M", "عج"),
+        (0xFC2A, "M", "عم"),
+        (0xFC2B, "M", "غج"),
+        (0xFC2C, "M", "غم"),
+        (0xFC2D, "M", "فج"),
+        (0xFC2E, "M", "فح"),
+        (0xFC2F, "M", "فخ"),
+        (0xFC30, "M", "فم"),
+        (0xFC31, "M", "فى"),
+        (0xFC32, "M", "في"),
+        (0xFC33, "M", "قح"),
+        (0xFC34, "M", "قم"),
+        (0xFC35, "M", "قى"),
+        (0xFC36, "M", "قي"),
+        (0xFC37, "M", "كا"),
+        (0xFC38, "M", "كج"),
+        (0xFC39, "M", "كح"),
+        (0xFC3A, "M", "كخ"),
+        (0xFC3B, "M", "كل"),
+        (0xFC3C, "M", "كم"),
+        (0xFC3D, "M", "كى"),
+        (0xFC3E, "M", "كي"),
+        (0xFC3F, "M", "لج"),
+        (0xFC40, "M", "لح"),
+        (0xFC41, "M", "لخ"),
+        (0xFC42, "M", "لم"),
+        (0xFC43, "M", "لى"),
+        (0xFC44, "M", "لي"),
+        (0xFC45, "M", "مج"),
+        (0xFC46, "M", "مح"),
+        (0xFC47, "M", "مخ"),
+        (0xFC48, "M", "مم"),
+        (0xFC49, "M", "مى"),
+        (0xFC4A, "M", "مي"),
+        (0xFC4B, "M", "نج"),
+        (0xFC4C, "M", "نح"),
+        (0xFC4D, "M", "نخ"),
+        (0xFC4E, "M", "نم"),
+        (0xFC4F, "M", "نى"),
+        (0xFC50, "M", "ني"),
+        (0xFC51, "M", "هج"),
+        (0xFC52, "M", "هم"),
+        (0xFC53, "M", "هى"),
+        (0xFC54, "M", "هي"),
+        (0xFC55, "M", "يج"),
+        (0xFC56, "M", "يح"),
+        (0xFC57, "M", "يخ"),
+        (0xFC58, "M", "يم"),
+        (0xFC59, "M", "يى"),
+        (0xFC5A, "M", "يي"),
+    ]
+
+
+def _seg_46() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFC5B, "M", "ذٰ"),
+        (0xFC5C, "M", "رٰ"),
+        (0xFC5D, "M", "ىٰ"),
+        (0xFC5E, "3", " ٌّ"),
+        (0xFC5F, "3", " ٍّ"),
+        (0xFC60, "3", " َّ"),
+        (0xFC61, "3", " ُّ"),
+        (0xFC62, "3", " ِّ"),
+        (0xFC63, "3", " ّٰ"),
+        (0xFC64, "M", "ئر"),
+        (0xFC65, "M", "ئز"),
+        (0xFC66, "M", "ئم"),
+        (0xFC67, "M", "ئن"),
+        (0xFC68, "M", "ئى"),
+        (0xFC69, "M", "ئي"),
+        (0xFC6A, "M", "بر"),
+        (0xFC6B, "M", "بز"),
+        (0xFC6C, "M", "بم"),
+        (0xFC6D, "M", "بن"),
+        (0xFC6E, "M", "بى"),
+        (0xFC6F, "M", "بي"),
+        (0xFC70, "M", "تر"),
+        (0xFC71, "M", "تز"),
+        (0xFC72, "M", "تم"),
+        (0xFC73, "M", "تن"),
+        (0xFC74, "M", "تى"),
+        (0xFC75, "M", "تي"),
+        (0xFC76, "M", "ثر"),
+        (0xFC77, "M", "ثز"),
+        (0xFC78, "M", "ثم"),
+        (0xFC79, "M", "ثن"),
+        (0xFC7A, "M", "ثى"),
+        (0xFC7B, "M", "ثي"),
+        (0xFC7C, "M", "فى"),
+        (0xFC7D, "M", "في"),
+        (0xFC7E, "M", "قى"),
+        (0xFC7F, "M", "قي"),
+        (0xFC80, "M", "كا"),
+        (0xFC81, "M", "كل"),
+        (0xFC82, "M", "كم"),
+        (0xFC83, "M", "كى"),
+        (0xFC84, "M", "كي"),
+        (0xFC85, "M", "لم"),
+        (0xFC86, "M", "لى"),
+        (0xFC87, "M", "لي"),
+        (0xFC88, "M", "ما"),
+        (0xFC89, "M", "مم"),
+        (0xFC8A, "M", "نر"),
+        (0xFC8B, "M", "نز"),
+        (0xFC8C, "M", "نم"),
+        (0xFC8D, "M", "نن"),
+        (0xFC8E, "M", "نى"),
+        (0xFC8F, "M", "ني"),
+        (0xFC90, "M", "ىٰ"),
+        (0xFC91, "M", "ير"),
+        (0xFC92, "M", "يز"),
+        (0xFC93, "M", "يم"),
+        (0xFC94, "M", "ين"),
+        (0xFC95, "M", "يى"),
+        (0xFC96, "M", "يي"),
+        (0xFC97, "M", "ئج"),
+        (0xFC98, "M", "ئح"),
+        (0xFC99, "M", "ئخ"),
+        (0xFC9A, "M", "ئم"),
+        (0xFC9B, "M", "ئه"),
+        (0xFC9C, "M", "بج"),
+        (0xFC9D, "M", "بح"),
+        (0xFC9E, "M", "بخ"),
+        (0xFC9F, "M", "بم"),
+        (0xFCA0, "M", "به"),
+        (0xFCA1, "M", "تج"),
+        (0xFCA2, "M", "تح"),
+        (0xFCA3, "M", "تخ"),
+        (0xFCA4, "M", "تم"),
+        (0xFCA5, "M", "ته"),
+        (0xFCA6, "M", "ثم"),
+        (0xFCA7, "M", "جح"),
+        (0xFCA8, "M", "جم"),
+        (0xFCA9, "M", "حج"),
+        (0xFCAA, "M", "حم"),
+        (0xFCAB, "M", "خج"),
+        (0xFCAC, "M", "خم"),
+        (0xFCAD, "M", "سج"),
+        (0xFCAE, "M", "سح"),
+        (0xFCAF, "M", "سخ"),
+        (0xFCB0, "M", "سم"),
+        (0xFCB1, "M", "صح"),
+        (0xFCB2, "M", "صخ"),
+        (0xFCB3, "M", "صم"),
+        (0xFCB4, "M", "ضج"),
+        (0xFCB5, "M", "ضح"),
+        (0xFCB6, "M", "ضخ"),
+        (0xFCB7, "M", "ضم"),
+        (0xFCB8, "M", "طح"),
+        (0xFCB9, "M", "ظم"),
+        (0xFCBA, "M", "عج"),
+        (0xFCBB, "M", "عم"),
+        (0xFCBC, "M", "غج"),
+        (0xFCBD, "M", "غم"),
+        (0xFCBE, "M", "فج"),
+    ]
+
+
+def _seg_47() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFCBF, "M", "فح"),
+        (0xFCC0, "M", "فخ"),
+        (0xFCC1, "M", "فم"),
+        (0xFCC2, "M", "قح"),
+        (0xFCC3, "M", "قم"),
+        (0xFCC4, "M", "كج"),
+        (0xFCC5, "M", "كح"),
+        (0xFCC6, "M", "كخ"),
+        (0xFCC7, "M", "كل"),
+        (0xFCC8, "M", "كم"),
+        (0xFCC9, "M", "لج"),
+        (0xFCCA, "M", "لح"),
+        (0xFCCB, "M", "لخ"),
+        (0xFCCC, "M", "لم"),
+        (0xFCCD, "M", "له"),
+        (0xFCCE, "M", "مج"),
+        (0xFCCF, "M", "مح"),
+        (0xFCD0, "M", "مخ"),
+        (0xFCD1, "M", "مم"),
+        (0xFCD2, "M", "نج"),
+        (0xFCD3, "M", "نح"),
+        (0xFCD4, "M", "نخ"),
+        (0xFCD5, "M", "نم"),
+        (0xFCD6, "M", "نه"),
+        (0xFCD7, "M", "هج"),
+        (0xFCD8, "M", "هم"),
+        (0xFCD9, "M", "هٰ"),
+        (0xFCDA, "M", "يج"),
+        (0xFCDB, "M", "يح"),
+        (0xFCDC, "M", "يخ"),
+        (0xFCDD, "M", "يم"),
+        (0xFCDE, "M", "يه"),
+        (0xFCDF, "M", "ئم"),
+        (0xFCE0, "M", "ئه"),
+        (0xFCE1, "M", "بم"),
+        (0xFCE2, "M", "به"),
+        (0xFCE3, "M", "تم"),
+        (0xFCE4, "M", "ته"),
+        (0xFCE5, "M", "ثم"),
+        (0xFCE6, "M", "ثه"),
+        (0xFCE7, "M", "سم"),
+        (0xFCE8, "M", "سه"),
+        (0xFCE9, "M", "شم"),
+        (0xFCEA, "M", "شه"),
+        (0xFCEB, "M", "كل"),
+        (0xFCEC, "M", "كم"),
+        (0xFCED, "M", "لم"),
+        (0xFCEE, "M", "نم"),
+        (0xFCEF, "M", "نه"),
+        (0xFCF0, "M", "يم"),
+        (0xFCF1, "M", "يه"),
+        (0xFCF2, "M", "ـَّ"),
+        (0xFCF3, "M", "ـُّ"),
+        (0xFCF4, "M", "ـِّ"),
+        (0xFCF5, "M", "طى"),
+        (0xFCF6, "M", "طي"),
+        (0xFCF7, "M", "عى"),
+        (0xFCF8, "M", "عي"),
+        (0xFCF9, "M", "غى"),
+        (0xFCFA, "M", "غي"),
+        (0xFCFB, "M", "سى"),
+        (0xFCFC, "M", "سي"),
+        (0xFCFD, "M", "شى"),
+        (0xFCFE, "M", "شي"),
+        (0xFCFF, "M", "حى"),
+        (0xFD00, "M", "حي"),
+        (0xFD01, "M", "جى"),
+        (0xFD02, "M", "جي"),
+        (0xFD03, "M", "خى"),
+        (0xFD04, "M", "خي"),
+        (0xFD05, "M", "صى"),
+        (0xFD06, "M", "صي"),
+        (0xFD07, "M", "ضى"),
+        (0xFD08, "M", "ضي"),
+        (0xFD09, "M", "شج"),
+        (0xFD0A, "M", "شح"),
+        (0xFD0B, "M", "شخ"),
+        (0xFD0C, "M", "شم"),
+        (0xFD0D, "M", "شر"),
+        (0xFD0E, "M", "سر"),
+        (0xFD0F, "M", "صر"),
+        (0xFD10, "M", "ضر"),
+        (0xFD11, "M", "طى"),
+        (0xFD12, "M", "طي"),
+        (0xFD13, "M", "عى"),
+        (0xFD14, "M", "عي"),
+        (0xFD15, "M", "غى"),
+        (0xFD16, "M", "غي"),
+        (0xFD17, "M", "سى"),
+        (0xFD18, "M", "سي"),
+        (0xFD19, "M", "شى"),
+        (0xFD1A, "M", "شي"),
+        (0xFD1B, "M", "حى"),
+        (0xFD1C, "M", "حي"),
+        (0xFD1D, "M", "جى"),
+        (0xFD1E, "M", "جي"),
+        (0xFD1F, "M", "خى"),
+        (0xFD20, "M", "خي"),
+        (0xFD21, "M", "صى"),
+        (0xFD22, "M", "صي"),
+    ]
+
+
+def _seg_48() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFD23, "M", "ضى"),
+        (0xFD24, "M", "ضي"),
+        (0xFD25, "M", "شج"),
+        (0xFD26, "M", "شح"),
+        (0xFD27, "M", "شخ"),
+        (0xFD28, "M", "شم"),
+        (0xFD29, "M", "شر"),
+        (0xFD2A, "M", "سر"),
+        (0xFD2B, "M", "صر"),
+        (0xFD2C, "M", "ضر"),
+        (0xFD2D, "M", "شج"),
+        (0xFD2E, "M", "شح"),
+        (0xFD2F, "M", "شخ"),
+        (0xFD30, "M", "شم"),
+        (0xFD31, "M", "سه"),
+        (0xFD32, "M", "شه"),
+        (0xFD33, "M", "طم"),
+        (0xFD34, "M", "سج"),
+        (0xFD35, "M", "سح"),
+        (0xFD36, "M", "سخ"),
+        (0xFD37, "M", "شج"),
+        (0xFD38, "M", "شح"),
+        (0xFD39, "M", "شخ"),
+        (0xFD3A, "M", "طم"),
+        (0xFD3B, "M", "ظم"),
+        (0xFD3C, "M", "اً"),
+        (0xFD3E, "V"),
+        (0xFD50, "M", "تجم"),
+        (0xFD51, "M", "تحج"),
+        (0xFD53, "M", "تحم"),
+        (0xFD54, "M", "تخم"),
+        (0xFD55, "M", "تمج"),
+        (0xFD56, "M", "تمح"),
+        (0xFD57, "M", "تمخ"),
+        (0xFD58, "M", "جمح"),
+        (0xFD5A, "M", "حمي"),
+        (0xFD5B, "M", "حمى"),
+        (0xFD5C, "M", "سحج"),
+        (0xFD5D, "M", "سجح"),
+        (0xFD5E, "M", "سجى"),
+        (0xFD5F, "M", "سمح"),
+        (0xFD61, "M", "سمج"),
+        (0xFD62, "M", "سمم"),
+        (0xFD64, "M", "صحح"),
+        (0xFD66, "M", "صمم"),
+        (0xFD67, "M", "شحم"),
+        (0xFD69, "M", "شجي"),
+        (0xFD6A, "M", "شمخ"),
+        (0xFD6C, "M", "شمم"),
+        (0xFD6E, "M", "ضحى"),
+        (0xFD6F, "M", "ضخم"),
+        (0xFD71, "M", "طمح"),
+        (0xFD73, "M", "طمم"),
+        (0xFD74, "M", "طمي"),
+        (0xFD75, "M", "عجم"),
+        (0xFD76, "M", "عمم"),
+        (0xFD78, "M", "عمى"),
+        (0xFD79, "M", "غمم"),
+        (0xFD7A, "M", "غمي"),
+        (0xFD7B, "M", "غمى"),
+        (0xFD7C, "M", "فخم"),
+        (0xFD7E, "M", "قمح"),
+        (0xFD7F, "M", "قمم"),
+        (0xFD80, "M", "لحم"),
+        (0xFD81, "M", "لحي"),
+        (0xFD82, "M", "لحى"),
+        (0xFD83, "M", "لجج"),
+        (0xFD85, "M", "لخم"),
+        (0xFD87, "M", "لمح"),
+        (0xFD89, "M", "محج"),
+        (0xFD8A, "M", "محم"),
+        (0xFD8B, "M", "محي"),
+        (0xFD8C, "M", "مجح"),
+        (0xFD8D, "M", "مجم"),
+        (0xFD8E, "M", "مخج"),
+        (0xFD8F, "M", "مخم"),
+        (0xFD90, "X"),
+        (0xFD92, "M", "مجخ"),
+        (0xFD93, "M", "همج"),
+        (0xFD94, "M", "همم"),
+        (0xFD95, "M", "نحم"),
+        (0xFD96, "M", "نحى"),
+        (0xFD97, "M", "نجم"),
+        (0xFD99, "M", "نجى"),
+        (0xFD9A, "M", "نمي"),
+        (0xFD9B, "M", "نمى"),
+        (0xFD9C, "M", "يمم"),
+        (0xFD9E, "M", "بخي"),
+        (0xFD9F, "M", "تجي"),
+        (0xFDA0, "M", "تجى"),
+        (0xFDA1, "M", "تخي"),
+        (0xFDA2, "M", "تخى"),
+        (0xFDA3, "M", "تمي"),
+        (0xFDA4, "M", "تمى"),
+        (0xFDA5, "M", "جمي"),
+        (0xFDA6, "M", "جحى"),
+        (0xFDA7, "M", "جمى"),
+        (0xFDA8, "M", "سخى"),
+        (0xFDA9, "M", "صحي"),
+        (0xFDAA, "M", "شحي"),
+    ]
+
+
+def _seg_49() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFDAB, "M", "ضحي"),
+        (0xFDAC, "M", "لجي"),
+        (0xFDAD, "M", "لمي"),
+        (0xFDAE, "M", "يحي"),
+        (0xFDAF, "M", "يجي"),
+        (0xFDB0, "M", "يمي"),
+        (0xFDB1, "M", "ممي"),
+        (0xFDB2, "M", "قمي"),
+        (0xFDB3, "M", "نحي"),
+        (0xFDB4, "M", "قمح"),
+        (0xFDB5, "M", "لحم"),
+        (0xFDB6, "M", "عمي"),
+        (0xFDB7, "M", "كمي"),
+        (0xFDB8, "M", "نجح"),
+        (0xFDB9, "M", "مخي"),
+        (0xFDBA, "M", "لجم"),
+        (0xFDBB, "M", "كمم"),
+        (0xFDBC, "M", "لجم"),
+        (0xFDBD, "M", "نجح"),
+        (0xFDBE, "M", "جحي"),
+        (0xFDBF, "M", "حجي"),
+        (0xFDC0, "M", "مجي"),
+        (0xFDC1, "M", "فمي"),
+        (0xFDC2, "M", "بحي"),
+        (0xFDC3, "M", "كمم"),
+        (0xFDC4, "M", "عجم"),
+        (0xFDC5, "M", "صمم"),
+        (0xFDC6, "M", "سخي"),
+        (0xFDC7, "M", "نجي"),
+        (0xFDC8, "X"),
+        (0xFDCF, "V"),
+        (0xFDD0, "X"),
+        (0xFDF0, "M", "صلے"),
+        (0xFDF1, "M", "قلے"),
+        (0xFDF2, "M", "الله"),
+        (0xFDF3, "M", "اكبر"),
+        (0xFDF4, "M", "محمد"),
+        (0xFDF5, "M", "صلعم"),
+        (0xFDF6, "M", "رسول"),
+        (0xFDF7, "M", "عليه"),
+        (0xFDF8, "M", "وسلم"),
+        (0xFDF9, "M", "صلى"),
+        (0xFDFA, "3", "صلى الله عليه وسلم"),
+        (0xFDFB, "3", "جل جلاله"),
+        (0xFDFC, "M", "ریال"),
+        (0xFDFD, "V"),
+        (0xFE00, "I"),
+        (0xFE10, "3", ","),
+        (0xFE11, "M", "、"),
+        (0xFE12, "X"),
+        (0xFE13, "3", ":"),
+        (0xFE14, "3", ";"),
+        (0xFE15, "3", "!"),
+        (0xFE16, "3", "?"),
+        (0xFE17, "M", "〖"),
+        (0xFE18, "M", "〗"),
+        (0xFE19, "X"),
+        (0xFE20, "V"),
+        (0xFE30, "X"),
+        (0xFE31, "M", "—"),
+        (0xFE32, "M", "–"),
+        (0xFE33, "3", "_"),
+        (0xFE35, "3", "("),
+        (0xFE36, "3", ")"),
+        (0xFE37, "3", "{"),
+        (0xFE38, "3", "}"),
+        (0xFE39, "M", "〔"),
+        (0xFE3A, "M", "〕"),
+        (0xFE3B, "M", "【"),
+        (0xFE3C, "M", "】"),
+        (0xFE3D, "M", "《"),
+        (0xFE3E, "M", "》"),
+        (0xFE3F, "M", "〈"),
+        (0xFE40, "M", "〉"),
+        (0xFE41, "M", "「"),
+        (0xFE42, "M", "」"),
+        (0xFE43, "M", "『"),
+        (0xFE44, "M", "』"),
+        (0xFE45, "V"),
+        (0xFE47, "3", "["),
+        (0xFE48, "3", "]"),
+        (0xFE49, "3", " ̅"),
+        (0xFE4D, "3", "_"),
+        (0xFE50, "3", ","),
+        (0xFE51, "M", "、"),
+        (0xFE52, "X"),
+        (0xFE54, "3", ";"),
+        (0xFE55, "3", ":"),
+        (0xFE56, "3", "?"),
+        (0xFE57, "3", "!"),
+        (0xFE58, "M", "—"),
+        (0xFE59, "3", "("),
+        (0xFE5A, "3", ")"),
+        (0xFE5B, "3", "{"),
+        (0xFE5C, "3", "}"),
+        (0xFE5D, "M", "〔"),
+        (0xFE5E, "M", "〕"),
+        (0xFE5F, "3", "#"),
+        (0xFE60, "3", "&"),
+        (0xFE61, "3", "*"),
+    ]
+
+
+def _seg_50() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFE62, "3", "+"),
+        (0xFE63, "M", "-"),
+        (0xFE64, "3", "<"),
+        (0xFE65, "3", ">"),
+        (0xFE66, "3", "="),
+        (0xFE67, "X"),
+        (0xFE68, "3", "\\"),
+        (0xFE69, "3", "$"),
+        (0xFE6A, "3", "%"),
+        (0xFE6B, "3", "@"),
+        (0xFE6C, "X"),
+        (0xFE70, "3", " ً"),
+        (0xFE71, "M", "ـً"),
+        (0xFE72, "3", " ٌ"),
+        (0xFE73, "V"),
+        (0xFE74, "3", " ٍ"),
+        (0xFE75, "X"),
+        (0xFE76, "3", " َ"),
+        (0xFE77, "M", "ـَ"),
+        (0xFE78, "3", " ُ"),
+        (0xFE79, "M", "ـُ"),
+        (0xFE7A, "3", " ِ"),
+        (0xFE7B, "M", "ـِ"),
+        (0xFE7C, "3", " ّ"),
+        (0xFE7D, "M", "ـّ"),
+        (0xFE7E, "3", " ْ"),
+        (0xFE7F, "M", "ـْ"),
+        (0xFE80, "M", "ء"),
+        (0xFE81, "M", "آ"),
+        (0xFE83, "M", "أ"),
+        (0xFE85, "M", "ؤ"),
+        (0xFE87, "M", "إ"),
+        (0xFE89, "M", "ئ"),
+        (0xFE8D, "M", "ا"),
+        (0xFE8F, "M", "ب"),
+        (0xFE93, "M", "ة"),
+        (0xFE95, "M", "ت"),
+        (0xFE99, "M", "ث"),
+        (0xFE9D, "M", "ج"),
+        (0xFEA1, "M", "ح"),
+        (0xFEA5, "M", "خ"),
+        (0xFEA9, "M", "د"),
+        (0xFEAB, "M", "ذ"),
+        (0xFEAD, "M", "ر"),
+        (0xFEAF, "M", "ز"),
+        (0xFEB1, "M", "س"),
+        (0xFEB5, "M", "ش"),
+        (0xFEB9, "M", "ص"),
+        (0xFEBD, "M", "ض"),
+        (0xFEC1, "M", "ط"),
+        (0xFEC5, "M", "ظ"),
+        (0xFEC9, "M", "ع"),
+        (0xFECD, "M", "غ"),
+        (0xFED1, "M", "ف"),
+        (0xFED5, "M", "ق"),
+        (0xFED9, "M", "ك"),
+        (0xFEDD, "M", "ل"),
+        (0xFEE1, "M", "م"),
+        (0xFEE5, "M", "ن"),
+        (0xFEE9, "M", "ه"),
+        (0xFEED, "M", "و"),
+        (0xFEEF, "M", "ى"),
+        (0xFEF1, "M", "ي"),
+        (0xFEF5, "M", "لآ"),
+        (0xFEF7, "M", "لأ"),
+        (0xFEF9, "M", "لإ"),
+        (0xFEFB, "M", "لا"),
+        (0xFEFD, "X"),
+        (0xFEFF, "I"),
+        (0xFF00, "X"),
+        (0xFF01, "3", "!"),
+        (0xFF02, "3", '"'),
+        (0xFF03, "3", "#"),
+        (0xFF04, "3", "$"),
+        (0xFF05, "3", "%"),
+        (0xFF06, "3", "&"),
+        (0xFF07, "3", "'"),
+        (0xFF08, "3", "("),
+        (0xFF09, "3", ")"),
+        (0xFF0A, "3", "*"),
+        (0xFF0B, "3", "+"),
+        (0xFF0C, "3", ","),
+        (0xFF0D, "M", "-"),
+        (0xFF0E, "M", "."),
+        (0xFF0F, "3", "/"),
+        (0xFF10, "M", "0"),
+        (0xFF11, "M", "1"),
+        (0xFF12, "M", "2"),
+        (0xFF13, "M", "3"),
+        (0xFF14, "M", "4"),
+        (0xFF15, "M", "5"),
+        (0xFF16, "M", "6"),
+        (0xFF17, "M", "7"),
+        (0xFF18, "M", "8"),
+        (0xFF19, "M", "9"),
+        (0xFF1A, "3", ":"),
+        (0xFF1B, "3", ";"),
+        (0xFF1C, "3", "<"),
+        (0xFF1D, "3", "="),
+        (0xFF1E, "3", ">"),
+    ]
+
+
+def _seg_51() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFF1F, "3", "?"),
+        (0xFF20, "3", "@"),
+        (0xFF21, "M", "a"),
+        (0xFF22, "M", "b"),
+        (0xFF23, "M", "c"),
+        (0xFF24, "M", "d"),
+        (0xFF25, "M", "e"),
+        (0xFF26, "M", "f"),
+        (0xFF27, "M", "g"),
+        (0xFF28, "M", "h"),
+        (0xFF29, "M", "i"),
+        (0xFF2A, "M", "j"),
+        (0xFF2B, "M", "k"),
+        (0xFF2C, "M", "l"),
+        (0xFF2D, "M", "m"),
+        (0xFF2E, "M", "n"),
+        (0xFF2F, "M", "o"),
+        (0xFF30, "M", "p"),
+        (0xFF31, "M", "q"),
+        (0xFF32, "M", "r"),
+        (0xFF33, "M", "s"),
+        (0xFF34, "M", "t"),
+        (0xFF35, "M", "u"),
+        (0xFF36, "M", "v"),
+        (0xFF37, "M", "w"),
+        (0xFF38, "M", "x"),
+        (0xFF39, "M", "y"),
+        (0xFF3A, "M", "z"),
+        (0xFF3B, "3", "["),
+        (0xFF3C, "3", "\\"),
+        (0xFF3D, "3", "]"),
+        (0xFF3E, "3", "^"),
+        (0xFF3F, "3", "_"),
+        (0xFF40, "3", "`"),
+        (0xFF41, "M", "a"),
+        (0xFF42, "M", "b"),
+        (0xFF43, "M", "c"),
+        (0xFF44, "M", "d"),
+        (0xFF45, "M", "e"),
+        (0xFF46, "M", "f"),
+        (0xFF47, "M", "g"),
+        (0xFF48, "M", "h"),
+        (0xFF49, "M", "i"),
+        (0xFF4A, "M", "j"),
+        (0xFF4B, "M", "k"),
+        (0xFF4C, "M", "l"),
+        (0xFF4D, "M", "m"),
+        (0xFF4E, "M", "n"),
+        (0xFF4F, "M", "o"),
+        (0xFF50, "M", "p"),
+        (0xFF51, "M", "q"),
+        (0xFF52, "M", "r"),
+        (0xFF53, "M", "s"),
+        (0xFF54, "M", "t"),
+        (0xFF55, "M", "u"),
+        (0xFF56, "M", "v"),
+        (0xFF57, "M", "w"),
+        (0xFF58, "M", "x"),
+        (0xFF59, "M", "y"),
+        (0xFF5A, "M", "z"),
+        (0xFF5B, "3", "{"),
+        (0xFF5C, "3", "|"),
+        (0xFF5D, "3", "}"),
+        (0xFF5E, "3", "~"),
+        (0xFF5F, "M", "⦅"),
+        (0xFF60, "M", "⦆"),
+        (0xFF61, "M", "."),
+        (0xFF62, "M", "「"),
+        (0xFF63, "M", "」"),
+        (0xFF64, "M", "、"),
+        (0xFF65, "M", "・"),
+        (0xFF66, "M", "ヲ"),
+        (0xFF67, "M", "ァ"),
+        (0xFF68, "M", "ィ"),
+        (0xFF69, "M", "ゥ"),
+        (0xFF6A, "M", "ェ"),
+        (0xFF6B, "M", "ォ"),
+        (0xFF6C, "M", "ャ"),
+        (0xFF6D, "M", "ュ"),
+        (0xFF6E, "M", "ョ"),
+        (0xFF6F, "M", "ッ"),
+        (0xFF70, "M", "ー"),
+        (0xFF71, "M", "ア"),
+        (0xFF72, "M", "イ"),
+        (0xFF73, "M", "ウ"),
+        (0xFF74, "M", "エ"),
+        (0xFF75, "M", "オ"),
+        (0xFF76, "M", "カ"),
+        (0xFF77, "M", "キ"),
+        (0xFF78, "M", "ク"),
+        (0xFF79, "M", "ケ"),
+        (0xFF7A, "M", "コ"),
+        (0xFF7B, "M", "サ"),
+        (0xFF7C, "M", "シ"),
+        (0xFF7D, "M", "ス"),
+        (0xFF7E, "M", "セ"),
+        (0xFF7F, "M", "ソ"),
+        (0xFF80, "M", "タ"),
+        (0xFF81, "M", "チ"),
+        (0xFF82, "M", "ツ"),
+    ]
+
+
+def _seg_52() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFF83, "M", "テ"),
+        (0xFF84, "M", "ト"),
+        (0xFF85, "M", "ナ"),
+        (0xFF86, "M", "ニ"),
+        (0xFF87, "M", "ヌ"),
+        (0xFF88, "M", "ネ"),
+        (0xFF89, "M", "ノ"),
+        (0xFF8A, "M", "ハ"),
+        (0xFF8B, "M", "ヒ"),
+        (0xFF8C, "M", "フ"),
+        (0xFF8D, "M", "ヘ"),
+        (0xFF8E, "M", "ホ"),
+        (0xFF8F, "M", "マ"),
+        (0xFF90, "M", "ミ"),
+        (0xFF91, "M", "ム"),
+        (0xFF92, "M", "メ"),
+        (0xFF93, "M", "モ"),
+        (0xFF94, "M", "ヤ"),
+        (0xFF95, "M", "ユ"),
+        (0xFF96, "M", "ヨ"),
+        (0xFF97, "M", "ラ"),
+        (0xFF98, "M", "リ"),
+        (0xFF99, "M", "ル"),
+        (0xFF9A, "M", "レ"),
+        (0xFF9B, "M", "ロ"),
+        (0xFF9C, "M", "ワ"),
+        (0xFF9D, "M", "ン"),
+        (0xFF9E, "M", "゙"),
+        (0xFF9F, "M", "゚"),
+        (0xFFA0, "X"),
+        (0xFFA1, "M", "ᄀ"),
+        (0xFFA2, "M", "ᄁ"),
+        (0xFFA3, "M", "ᆪ"),
+        (0xFFA4, "M", "ᄂ"),
+        (0xFFA5, "M", "ᆬ"),
+        (0xFFA6, "M", "ᆭ"),
+        (0xFFA7, "M", "ᄃ"),
+        (0xFFA8, "M", "ᄄ"),
+        (0xFFA9, "M", "ᄅ"),
+        (0xFFAA, "M", "ᆰ"),
+        (0xFFAB, "M", "ᆱ"),
+        (0xFFAC, "M", "ᆲ"),
+        (0xFFAD, "M", "ᆳ"),
+        (0xFFAE, "M", "ᆴ"),
+        (0xFFAF, "M", "ᆵ"),
+        (0xFFB0, "M", "ᄚ"),
+        (0xFFB1, "M", "ᄆ"),
+        (0xFFB2, "M", "ᄇ"),
+        (0xFFB3, "M", "ᄈ"),
+        (0xFFB4, "M", "ᄡ"),
+        (0xFFB5, "M", "ᄉ"),
+        (0xFFB6, "M", "ᄊ"),
+        (0xFFB7, "M", "ᄋ"),
+        (0xFFB8, "M", "ᄌ"),
+        (0xFFB9, "M", "ᄍ"),
+        (0xFFBA, "M", "ᄎ"),
+        (0xFFBB, "M", "ᄏ"),
+        (0xFFBC, "M", "ᄐ"),
+        (0xFFBD, "M", "ᄑ"),
+        (0xFFBE, "M", "ᄒ"),
+        (0xFFBF, "X"),
+        (0xFFC2, "M", "ᅡ"),
+        (0xFFC3, "M", "ᅢ"),
+        (0xFFC4, "M", "ᅣ"),
+        (0xFFC5, "M", "ᅤ"),
+        (0xFFC6, "M", "ᅥ"),
+        (0xFFC7, "M", "ᅦ"),
+        (0xFFC8, "X"),
+        (0xFFCA, "M", "ᅧ"),
+        (0xFFCB, "M", "ᅨ"),
+        (0xFFCC, "M", "ᅩ"),
+        (0xFFCD, "M", "ᅪ"),
+        (0xFFCE, "M", "ᅫ"),
+        (0xFFCF, "M", "ᅬ"),
+        (0xFFD0, "X"),
+        (0xFFD2, "M", "ᅭ"),
+        (0xFFD3, "M", "ᅮ"),
+        (0xFFD4, "M", "ᅯ"),
+        (0xFFD5, "M", "ᅰ"),
+        (0xFFD6, "M", "ᅱ"),
+        (0xFFD7, "M", "ᅲ"),
+        (0xFFD8, "X"),
+        (0xFFDA, "M", "ᅳ"),
+        (0xFFDB, "M", "ᅴ"),
+        (0xFFDC, "M", "ᅵ"),
+        (0xFFDD, "X"),
+        (0xFFE0, "M", "¢"),
+        (0xFFE1, "M", "£"),
+        (0xFFE2, "M", "¬"),
+        (0xFFE3, "3", " ̄"),
+        (0xFFE4, "M", "¦"),
+        (0xFFE5, "M", "¥"),
+        (0xFFE6, "M", "₩"),
+        (0xFFE7, "X"),
+        (0xFFE8, "M", "│"),
+        (0xFFE9, "M", "←"),
+        (0xFFEA, "M", "↑"),
+        (0xFFEB, "M", "→"),
+        (0xFFEC, "M", "↓"),
+        (0xFFED, "M", "■"),
+    ]
+
+
+def _seg_53() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFFEE, "M", "○"),
+        (0xFFEF, "X"),
+        (0x10000, "V"),
+        (0x1000C, "X"),
+        (0x1000D, "V"),
+        (0x10027, "X"),
+        (0x10028, "V"),
+        (0x1003B, "X"),
+        (0x1003C, "V"),
+        (0x1003E, "X"),
+        (0x1003F, "V"),
+        (0x1004E, "X"),
+        (0x10050, "V"),
+        (0x1005E, "X"),
+        (0x10080, "V"),
+        (0x100FB, "X"),
+        (0x10100, "V"),
+        (0x10103, "X"),
+        (0x10107, "V"),
+        (0x10134, "X"),
+        (0x10137, "V"),
+        (0x1018F, "X"),
+        (0x10190, "V"),
+        (0x1019D, "X"),
+        (0x101A0, "V"),
+        (0x101A1, "X"),
+        (0x101D0, "V"),
+        (0x101FE, "X"),
+        (0x10280, "V"),
+        (0x1029D, "X"),
+        (0x102A0, "V"),
+        (0x102D1, "X"),
+        (0x102E0, "V"),
+        (0x102FC, "X"),
+        (0x10300, "V"),
+        (0x10324, "X"),
+        (0x1032D, "V"),
+        (0x1034B, "X"),
+        (0x10350, "V"),
+        (0x1037B, "X"),
+        (0x10380, "V"),
+        (0x1039E, "X"),
+        (0x1039F, "V"),
+        (0x103C4, "X"),
+        (0x103C8, "V"),
+        (0x103D6, "X"),
+        (0x10400, "M", "𐐨"),
+        (0x10401, "M", "𐐩"),
+        (0x10402, "M", "𐐪"),
+        (0x10403, "M", "𐐫"),
+        (0x10404, "M", "𐐬"),
+        (0x10405, "M", "𐐭"),
+        (0x10406, "M", "𐐮"),
+        (0x10407, "M", "𐐯"),
+        (0x10408, "M", "𐐰"),
+        (0x10409, "M", "𐐱"),
+        (0x1040A, "M", "𐐲"),
+        (0x1040B, "M", "𐐳"),
+        (0x1040C, "M", "𐐴"),
+        (0x1040D, "M", "𐐵"),
+        (0x1040E, "M", "𐐶"),
+        (0x1040F, "M", "𐐷"),
+        (0x10410, "M", "𐐸"),
+        (0x10411, "M", "𐐹"),
+        (0x10412, "M", "𐐺"),
+        (0x10413, "M", "𐐻"),
+        (0x10414, "M", "𐐼"),
+        (0x10415, "M", "𐐽"),
+        (0x10416, "M", "𐐾"),
+        (0x10417, "M", "𐐿"),
+        (0x10418, "M", "𐑀"),
+        (0x10419, "M", "𐑁"),
+        (0x1041A, "M", "𐑂"),
+        (0x1041B, "M", "𐑃"),
+        (0x1041C, "M", "𐑄"),
+        (0x1041D, "M", "𐑅"),
+        (0x1041E, "M", "𐑆"),
+        (0x1041F, "M", "𐑇"),
+        (0x10420, "M", "𐑈"),
+        (0x10421, "M", "𐑉"),
+        (0x10422, "M", "𐑊"),
+        (0x10423, "M", "𐑋"),
+        (0x10424, "M", "𐑌"),
+        (0x10425, "M", "𐑍"),
+        (0x10426, "M", "𐑎"),
+        (0x10427, "M", "𐑏"),
+        (0x10428, "V"),
+        (0x1049E, "X"),
+        (0x104A0, "V"),
+        (0x104AA, "X"),
+        (0x104B0, "M", "𐓘"),
+        (0x104B1, "M", "𐓙"),
+        (0x104B2, "M", "𐓚"),
+        (0x104B3, "M", "𐓛"),
+        (0x104B4, "M", "𐓜"),
+        (0x104B5, "M", "𐓝"),
+        (0x104B6, "M", "𐓞"),
+        (0x104B7, "M", "𐓟"),
+        (0x104B8, "M", "𐓠"),
+        (0x104B9, "M", "𐓡"),
+    ]
+
+
+def _seg_54() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x104BA, "M", "𐓢"),
+        (0x104BB, "M", "𐓣"),
+        (0x104BC, "M", "𐓤"),
+        (0x104BD, "M", "𐓥"),
+        (0x104BE, "M", "𐓦"),
+        (0x104BF, "M", "𐓧"),
+        (0x104C0, "M", "𐓨"),
+        (0x104C1, "M", "𐓩"),
+        (0x104C2, "M", "𐓪"),
+        (0x104C3, "M", "𐓫"),
+        (0x104C4, "M", "𐓬"),
+        (0x104C5, "M", "𐓭"),
+        (0x104C6, "M", "𐓮"),
+        (0x104C7, "M", "𐓯"),
+        (0x104C8, "M", "𐓰"),
+        (0x104C9, "M", "𐓱"),
+        (0x104CA, "M", "𐓲"),
+        (0x104CB, "M", "𐓳"),
+        (0x104CC, "M", "𐓴"),
+        (0x104CD, "M", "𐓵"),
+        (0x104CE, "M", "𐓶"),
+        (0x104CF, "M", "𐓷"),
+        (0x104D0, "M", "𐓸"),
+        (0x104D1, "M", "𐓹"),
+        (0x104D2, "M", "𐓺"),
+        (0x104D3, "M", "𐓻"),
+        (0x104D4, "X"),
+        (0x104D8, "V"),
+        (0x104FC, "X"),
+        (0x10500, "V"),
+        (0x10528, "X"),
+        (0x10530, "V"),
+        (0x10564, "X"),
+        (0x1056F, "V"),
+        (0x10570, "M", "𐖗"),
+        (0x10571, "M", "𐖘"),
+        (0x10572, "M", "𐖙"),
+        (0x10573, "M", "𐖚"),
+        (0x10574, "M", "𐖛"),
+        (0x10575, "M", "𐖜"),
+        (0x10576, "M", "𐖝"),
+        (0x10577, "M", "𐖞"),
+        (0x10578, "M", "𐖟"),
+        (0x10579, "M", "𐖠"),
+        (0x1057A, "M", "𐖡"),
+        (0x1057B, "X"),
+        (0x1057C, "M", "𐖣"),
+        (0x1057D, "M", "𐖤"),
+        (0x1057E, "M", "𐖥"),
+        (0x1057F, "M", "𐖦"),
+        (0x10580, "M", "𐖧"),
+        (0x10581, "M", "𐖨"),
+        (0x10582, "M", "𐖩"),
+        (0x10583, "M", "𐖪"),
+        (0x10584, "M", "𐖫"),
+        (0x10585, "M", "𐖬"),
+        (0x10586, "M", "𐖭"),
+        (0x10587, "M", "𐖮"),
+        (0x10588, "M", "𐖯"),
+        (0x10589, "M", "𐖰"),
+        (0x1058A, "M", "𐖱"),
+        (0x1058B, "X"),
+        (0x1058C, "M", "𐖳"),
+        (0x1058D, "M", "𐖴"),
+        (0x1058E, "M", "𐖵"),
+        (0x1058F, "M", "𐖶"),
+        (0x10590, "M", "𐖷"),
+        (0x10591, "M", "𐖸"),
+        (0x10592, "M", "𐖹"),
+        (0x10593, "X"),
+        (0x10594, "M", "𐖻"),
+        (0x10595, "M", "𐖼"),
+        (0x10596, "X"),
+        (0x10597, "V"),
+        (0x105A2, "X"),
+        (0x105A3, "V"),
+        (0x105B2, "X"),
+        (0x105B3, "V"),
+        (0x105BA, "X"),
+        (0x105BB, "V"),
+        (0x105BD, "X"),
+        (0x10600, "V"),
+        (0x10737, "X"),
+        (0x10740, "V"),
+        (0x10756, "X"),
+        (0x10760, "V"),
+        (0x10768, "X"),
+        (0x10780, "V"),
+        (0x10781, "M", "ː"),
+        (0x10782, "M", "ˑ"),
+        (0x10783, "M", "æ"),
+        (0x10784, "M", "ʙ"),
+        (0x10785, "M", "ɓ"),
+        (0x10786, "X"),
+        (0x10787, "M", "ʣ"),
+        (0x10788, "M", "ꭦ"),
+        (0x10789, "M", "ʥ"),
+        (0x1078A, "M", "ʤ"),
+        (0x1078B, "M", "ɖ"),
+        (0x1078C, "M", "ɗ"),
+    ]
+
+
+def _seg_55() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1078D, "M", "ᶑ"),
+        (0x1078E, "M", "ɘ"),
+        (0x1078F, "M", "ɞ"),
+        (0x10790, "M", "ʩ"),
+        (0x10791, "M", "ɤ"),
+        (0x10792, "M", "ɢ"),
+        (0x10793, "M", "ɠ"),
+        (0x10794, "M", "ʛ"),
+        (0x10795, "M", "ħ"),
+        (0x10796, "M", "ʜ"),
+        (0x10797, "M", "ɧ"),
+        (0x10798, "M", "ʄ"),
+        (0x10799, "M", "ʪ"),
+        (0x1079A, "M", "ʫ"),
+        (0x1079B, "M", "ɬ"),
+        (0x1079C, "M", "𝼄"),
+        (0x1079D, "M", "ꞎ"),
+        (0x1079E, "M", "ɮ"),
+        (0x1079F, "M", "𝼅"),
+        (0x107A0, "M", "ʎ"),
+        (0x107A1, "M", "𝼆"),
+        (0x107A2, "M", "ø"),
+        (0x107A3, "M", "ɶ"),
+        (0x107A4, "M", "ɷ"),
+        (0x107A5, "M", "q"),
+        (0x107A6, "M", "ɺ"),
+        (0x107A7, "M", "𝼈"),
+        (0x107A8, "M", "ɽ"),
+        (0x107A9, "M", "ɾ"),
+        (0x107AA, "M", "ʀ"),
+        (0x107AB, "M", "ʨ"),
+        (0x107AC, "M", "ʦ"),
+        (0x107AD, "M", "ꭧ"),
+        (0x107AE, "M", "ʧ"),
+        (0x107AF, "M", "ʈ"),
+        (0x107B0, "M", "ⱱ"),
+        (0x107B1, "X"),
+        (0x107B2, "M", "ʏ"),
+        (0x107B3, "M", "ʡ"),
+        (0x107B4, "M", "ʢ"),
+        (0x107B5, "M", "ʘ"),
+        (0x107B6, "M", "ǀ"),
+        (0x107B7, "M", "ǁ"),
+        (0x107B8, "M", "ǂ"),
+        (0x107B9, "M", "𝼊"),
+        (0x107BA, "M", "𝼞"),
+        (0x107BB, "X"),
+        (0x10800, "V"),
+        (0x10806, "X"),
+        (0x10808, "V"),
+        (0x10809, "X"),
+        (0x1080A, "V"),
+        (0x10836, "X"),
+        (0x10837, "V"),
+        (0x10839, "X"),
+        (0x1083C, "V"),
+        (0x1083D, "X"),
+        (0x1083F, "V"),
+        (0x10856, "X"),
+        (0x10857, "V"),
+        (0x1089F, "X"),
+        (0x108A7, "V"),
+        (0x108B0, "X"),
+        (0x108E0, "V"),
+        (0x108F3, "X"),
+        (0x108F4, "V"),
+        (0x108F6, "X"),
+        (0x108FB, "V"),
+        (0x1091C, "X"),
+        (0x1091F, "V"),
+        (0x1093A, "X"),
+        (0x1093F, "V"),
+        (0x10940, "X"),
+        (0x10980, "V"),
+        (0x109B8, "X"),
+        (0x109BC, "V"),
+        (0x109D0, "X"),
+        (0x109D2, "V"),
+        (0x10A04, "X"),
+        (0x10A05, "V"),
+        (0x10A07, "X"),
+        (0x10A0C, "V"),
+        (0x10A14, "X"),
+        (0x10A15, "V"),
+        (0x10A18, "X"),
+        (0x10A19, "V"),
+        (0x10A36, "X"),
+        (0x10A38, "V"),
+        (0x10A3B, "X"),
+        (0x10A3F, "V"),
+        (0x10A49, "X"),
+        (0x10A50, "V"),
+        (0x10A59, "X"),
+        (0x10A60, "V"),
+        (0x10AA0, "X"),
+        (0x10AC0, "V"),
+        (0x10AE7, "X"),
+        (0x10AEB, "V"),
+        (0x10AF7, "X"),
+        (0x10B00, "V"),
+    ]
+
+
+def _seg_56() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x10B36, "X"),
+        (0x10B39, "V"),
+        (0x10B56, "X"),
+        (0x10B58, "V"),
+        (0x10B73, "X"),
+        (0x10B78, "V"),
+        (0x10B92, "X"),
+        (0x10B99, "V"),
+        (0x10B9D, "X"),
+        (0x10BA9, "V"),
+        (0x10BB0, "X"),
+        (0x10C00, "V"),
+        (0x10C49, "X"),
+        (0x10C80, "M", "𐳀"),
+        (0x10C81, "M", "𐳁"),
+        (0x10C82, "M", "𐳂"),
+        (0x10C83, "M", "𐳃"),
+        (0x10C84, "M", "𐳄"),
+        (0x10C85, "M", "𐳅"),
+        (0x10C86, "M", "𐳆"),
+        (0x10C87, "M", "𐳇"),
+        (0x10C88, "M", "𐳈"),
+        (0x10C89, "M", "𐳉"),
+        (0x10C8A, "M", "𐳊"),
+        (0x10C8B, "M", "𐳋"),
+        (0x10C8C, "M", "𐳌"),
+        (0x10C8D, "M", "𐳍"),
+        (0x10C8E, "M", "𐳎"),
+        (0x10C8F, "M", "𐳏"),
+        (0x10C90, "M", "𐳐"),
+        (0x10C91, "M", "𐳑"),
+        (0x10C92, "M", "𐳒"),
+        (0x10C93, "M", "𐳓"),
+        (0x10C94, "M", "𐳔"),
+        (0x10C95, "M", "𐳕"),
+        (0x10C96, "M", "𐳖"),
+        (0x10C97, "M", "𐳗"),
+        (0x10C98, "M", "𐳘"),
+        (0x10C99, "M", "𐳙"),
+        (0x10C9A, "M", "𐳚"),
+        (0x10C9B, "M", "𐳛"),
+        (0x10C9C, "M", "𐳜"),
+        (0x10C9D, "M", "𐳝"),
+        (0x10C9E, "M", "𐳞"),
+        (0x10C9F, "M", "𐳟"),
+        (0x10CA0, "M", "𐳠"),
+        (0x10CA1, "M", "𐳡"),
+        (0x10CA2, "M", "𐳢"),
+        (0x10CA3, "M", "𐳣"),
+        (0x10CA4, "M", "𐳤"),
+        (0x10CA5, "M", "𐳥"),
+        (0x10CA6, "M", "𐳦"),
+        (0x10CA7, "M", "𐳧"),
+        (0x10CA8, "M", "𐳨"),
+        (0x10CA9, "M", "𐳩"),
+        (0x10CAA, "M", "𐳪"),
+        (0x10CAB, "M", "𐳫"),
+        (0x10CAC, "M", "𐳬"),
+        (0x10CAD, "M", "𐳭"),
+        (0x10CAE, "M", "𐳮"),
+        (0x10CAF, "M", "𐳯"),
+        (0x10CB0, "M", "𐳰"),
+        (0x10CB1, "M", "𐳱"),
+        (0x10CB2, "M", "𐳲"),
+        (0x10CB3, "X"),
+        (0x10CC0, "V"),
+        (0x10CF3, "X"),
+        (0x10CFA, "V"),
+        (0x10D28, "X"),
+        (0x10D30, "V"),
+        (0x10D3A, "X"),
+        (0x10E60, "V"),
+        (0x10E7F, "X"),
+        (0x10E80, "V"),
+        (0x10EAA, "X"),
+        (0x10EAB, "V"),
+        (0x10EAE, "X"),
+        (0x10EB0, "V"),
+        (0x10EB2, "X"),
+        (0x10EFD, "V"),
+        (0x10F28, "X"),
+        (0x10F30, "V"),
+        (0x10F5A, "X"),
+        (0x10F70, "V"),
+        (0x10F8A, "X"),
+        (0x10FB0, "V"),
+        (0x10FCC, "X"),
+        (0x10FE0, "V"),
+        (0x10FF7, "X"),
+        (0x11000, "V"),
+        (0x1104E, "X"),
+        (0x11052, "V"),
+        (0x11076, "X"),
+        (0x1107F, "V"),
+        (0x110BD, "X"),
+        (0x110BE, "V"),
+        (0x110C3, "X"),
+        (0x110D0, "V"),
+        (0x110E9, "X"),
+        (0x110F0, "V"),
+    ]
+
+
+def _seg_57() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x110FA, "X"),
+        (0x11100, "V"),
+        (0x11135, "X"),
+        (0x11136, "V"),
+        (0x11148, "X"),
+        (0x11150, "V"),
+        (0x11177, "X"),
+        (0x11180, "V"),
+        (0x111E0, "X"),
+        (0x111E1, "V"),
+        (0x111F5, "X"),
+        (0x11200, "V"),
+        (0x11212, "X"),
+        (0x11213, "V"),
+        (0x11242, "X"),
+        (0x11280, "V"),
+        (0x11287, "X"),
+        (0x11288, "V"),
+        (0x11289, "X"),
+        (0x1128A, "V"),
+        (0x1128E, "X"),
+        (0x1128F, "V"),
+        (0x1129E, "X"),
+        (0x1129F, "V"),
+        (0x112AA, "X"),
+        (0x112B0, "V"),
+        (0x112EB, "X"),
+        (0x112F0, "V"),
+        (0x112FA, "X"),
+        (0x11300, "V"),
+        (0x11304, "X"),
+        (0x11305, "V"),
+        (0x1130D, "X"),
+        (0x1130F, "V"),
+        (0x11311, "X"),
+        (0x11313, "V"),
+        (0x11329, "X"),
+        (0x1132A, "V"),
+        (0x11331, "X"),
+        (0x11332, "V"),
+        (0x11334, "X"),
+        (0x11335, "V"),
+        (0x1133A, "X"),
+        (0x1133B, "V"),
+        (0x11345, "X"),
+        (0x11347, "V"),
+        (0x11349, "X"),
+        (0x1134B, "V"),
+        (0x1134E, "X"),
+        (0x11350, "V"),
+        (0x11351, "X"),
+        (0x11357, "V"),
+        (0x11358, "X"),
+        (0x1135D, "V"),
+        (0x11364, "X"),
+        (0x11366, "V"),
+        (0x1136D, "X"),
+        (0x11370, "V"),
+        (0x11375, "X"),
+        (0x11400, "V"),
+        (0x1145C, "X"),
+        (0x1145D, "V"),
+        (0x11462, "X"),
+        (0x11480, "V"),
+        (0x114C8, "X"),
+        (0x114D0, "V"),
+        (0x114DA, "X"),
+        (0x11580, "V"),
+        (0x115B6, "X"),
+        (0x115B8, "V"),
+        (0x115DE, "X"),
+        (0x11600, "V"),
+        (0x11645, "X"),
+        (0x11650, "V"),
+        (0x1165A, "X"),
+        (0x11660, "V"),
+        (0x1166D, "X"),
+        (0x11680, "V"),
+        (0x116BA, "X"),
+        (0x116C0, "V"),
+        (0x116CA, "X"),
+        (0x11700, "V"),
+        (0x1171B, "X"),
+        (0x1171D, "V"),
+        (0x1172C, "X"),
+        (0x11730, "V"),
+        (0x11747, "X"),
+        (0x11800, "V"),
+        (0x1183C, "X"),
+        (0x118A0, "M", "𑣀"),
+        (0x118A1, "M", "𑣁"),
+        (0x118A2, "M", "𑣂"),
+        (0x118A3, "M", "𑣃"),
+        (0x118A4, "M", "𑣄"),
+        (0x118A5, "M", "𑣅"),
+        (0x118A6, "M", "𑣆"),
+        (0x118A7, "M", "𑣇"),
+        (0x118A8, "M", "𑣈"),
+        (0x118A9, "M", "𑣉"),
+        (0x118AA, "M", "𑣊"),
+    ]
+
+
+def _seg_58() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x118AB, "M", "𑣋"),
+        (0x118AC, "M", "𑣌"),
+        (0x118AD, "M", "𑣍"),
+        (0x118AE, "M", "𑣎"),
+        (0x118AF, "M", "𑣏"),
+        (0x118B0, "M", "𑣐"),
+        (0x118B1, "M", "𑣑"),
+        (0x118B2, "M", "𑣒"),
+        (0x118B3, "M", "𑣓"),
+        (0x118B4, "M", "𑣔"),
+        (0x118B5, "M", "𑣕"),
+        (0x118B6, "M", "𑣖"),
+        (0x118B7, "M", "𑣗"),
+        (0x118B8, "M", "𑣘"),
+        (0x118B9, "M", "𑣙"),
+        (0x118BA, "M", "𑣚"),
+        (0x118BB, "M", "𑣛"),
+        (0x118BC, "M", "𑣜"),
+        (0x118BD, "M", "𑣝"),
+        (0x118BE, "M", "𑣞"),
+        (0x118BF, "M", "𑣟"),
+        (0x118C0, "V"),
+        (0x118F3, "X"),
+        (0x118FF, "V"),
+        (0x11907, "X"),
+        (0x11909, "V"),
+        (0x1190A, "X"),
+        (0x1190C, "V"),
+        (0x11914, "X"),
+        (0x11915, "V"),
+        (0x11917, "X"),
+        (0x11918, "V"),
+        (0x11936, "X"),
+        (0x11937, "V"),
+        (0x11939, "X"),
+        (0x1193B, "V"),
+        (0x11947, "X"),
+        (0x11950, "V"),
+        (0x1195A, "X"),
+        (0x119A0, "V"),
+        (0x119A8, "X"),
+        (0x119AA, "V"),
+        (0x119D8, "X"),
+        (0x119DA, "V"),
+        (0x119E5, "X"),
+        (0x11A00, "V"),
+        (0x11A48, "X"),
+        (0x11A50, "V"),
+        (0x11AA3, "X"),
+        (0x11AB0, "V"),
+        (0x11AF9, "X"),
+        (0x11B00, "V"),
+        (0x11B0A, "X"),
+        (0x11C00, "V"),
+        (0x11C09, "X"),
+        (0x11C0A, "V"),
+        (0x11C37, "X"),
+        (0x11C38, "V"),
+        (0x11C46, "X"),
+        (0x11C50, "V"),
+        (0x11C6D, "X"),
+        (0x11C70, "V"),
+        (0x11C90, "X"),
+        (0x11C92, "V"),
+        (0x11CA8, "X"),
+        (0x11CA9, "V"),
+        (0x11CB7, "X"),
+        (0x11D00, "V"),
+        (0x11D07, "X"),
+        (0x11D08, "V"),
+        (0x11D0A, "X"),
+        (0x11D0B, "V"),
+        (0x11D37, "X"),
+        (0x11D3A, "V"),
+        (0x11D3B, "X"),
+        (0x11D3C, "V"),
+        (0x11D3E, "X"),
+        (0x11D3F, "V"),
+        (0x11D48, "X"),
+        (0x11D50, "V"),
+        (0x11D5A, "X"),
+        (0x11D60, "V"),
+        (0x11D66, "X"),
+        (0x11D67, "V"),
+        (0x11D69, "X"),
+        (0x11D6A, "V"),
+        (0x11D8F, "X"),
+        (0x11D90, "V"),
+        (0x11D92, "X"),
+        (0x11D93, "V"),
+        (0x11D99, "X"),
+        (0x11DA0, "V"),
+        (0x11DAA, "X"),
+        (0x11EE0, "V"),
+        (0x11EF9, "X"),
+        (0x11F00, "V"),
+        (0x11F11, "X"),
+        (0x11F12, "V"),
+        (0x11F3B, "X"),
+        (0x11F3E, "V"),
+    ]
+
+
+def _seg_59() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x11F5A, "X"),
+        (0x11FB0, "V"),
+        (0x11FB1, "X"),
+        (0x11FC0, "V"),
+        (0x11FF2, "X"),
+        (0x11FFF, "V"),
+        (0x1239A, "X"),
+        (0x12400, "V"),
+        (0x1246F, "X"),
+        (0x12470, "V"),
+        (0x12475, "X"),
+        (0x12480, "V"),
+        (0x12544, "X"),
+        (0x12F90, "V"),
+        (0x12FF3, "X"),
+        (0x13000, "V"),
+        (0x13430, "X"),
+        (0x13440, "V"),
+        (0x13456, "X"),
+        (0x14400, "V"),
+        (0x14647, "X"),
+        (0x16800, "V"),
+        (0x16A39, "X"),
+        (0x16A40, "V"),
+        (0x16A5F, "X"),
+        (0x16A60, "V"),
+        (0x16A6A, "X"),
+        (0x16A6E, "V"),
+        (0x16ABF, "X"),
+        (0x16AC0, "V"),
+        (0x16ACA, "X"),
+        (0x16AD0, "V"),
+        (0x16AEE, "X"),
+        (0x16AF0, "V"),
+        (0x16AF6, "X"),
+        (0x16B00, "V"),
+        (0x16B46, "X"),
+        (0x16B50, "V"),
+        (0x16B5A, "X"),
+        (0x16B5B, "V"),
+        (0x16B62, "X"),
+        (0x16B63, "V"),
+        (0x16B78, "X"),
+        (0x16B7D, "V"),
+        (0x16B90, "X"),
+        (0x16E40, "M", "𖹠"),
+        (0x16E41, "M", "𖹡"),
+        (0x16E42, "M", "𖹢"),
+        (0x16E43, "M", "𖹣"),
+        (0x16E44, "M", "𖹤"),
+        (0x16E45, "M", "𖹥"),
+        (0x16E46, "M", "𖹦"),
+        (0x16E47, "M", "𖹧"),
+        (0x16E48, "M", "𖹨"),
+        (0x16E49, "M", "𖹩"),
+        (0x16E4A, "M", "𖹪"),
+        (0x16E4B, "M", "𖹫"),
+        (0x16E4C, "M", "𖹬"),
+        (0x16E4D, "M", "𖹭"),
+        (0x16E4E, "M", "𖹮"),
+        (0x16E4F, "M", "𖹯"),
+        (0x16E50, "M", "𖹰"),
+        (0x16E51, "M", "𖹱"),
+        (0x16E52, "M", "𖹲"),
+        (0x16E53, "M", "𖹳"),
+        (0x16E54, "M", "𖹴"),
+        (0x16E55, "M", "𖹵"),
+        (0x16E56, "M", "𖹶"),
+        (0x16E57, "M", "𖹷"),
+        (0x16E58, "M", "𖹸"),
+        (0x16E59, "M", "𖹹"),
+        (0x16E5A, "M", "𖹺"),
+        (0x16E5B, "M", "𖹻"),
+        (0x16E5C, "M", "𖹼"),
+        (0x16E5D, "M", "𖹽"),
+        (0x16E5E, "M", "𖹾"),
+        (0x16E5F, "M", "𖹿"),
+        (0x16E60, "V"),
+        (0x16E9B, "X"),
+        (0x16F00, "V"),
+        (0x16F4B, "X"),
+        (0x16F4F, "V"),
+        (0x16F88, "X"),
+        (0x16F8F, "V"),
+        (0x16FA0, "X"),
+        (0x16FE0, "V"),
+        (0x16FE5, "X"),
+        (0x16FF0, "V"),
+        (0x16FF2, "X"),
+        (0x17000, "V"),
+        (0x187F8, "X"),
+        (0x18800, "V"),
+        (0x18CD6, "X"),
+        (0x18D00, "V"),
+        (0x18D09, "X"),
+        (0x1AFF0, "V"),
+        (0x1AFF4, "X"),
+        (0x1AFF5, "V"),
+        (0x1AFFC, "X"),
+        (0x1AFFD, "V"),
+    ]
+
+
+def _seg_60() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1AFFF, "X"),
+        (0x1B000, "V"),
+        (0x1B123, "X"),
+        (0x1B132, "V"),
+        (0x1B133, "X"),
+        (0x1B150, "V"),
+        (0x1B153, "X"),
+        (0x1B155, "V"),
+        (0x1B156, "X"),
+        (0x1B164, "V"),
+        (0x1B168, "X"),
+        (0x1B170, "V"),
+        (0x1B2FC, "X"),
+        (0x1BC00, "V"),
+        (0x1BC6B, "X"),
+        (0x1BC70, "V"),
+        (0x1BC7D, "X"),
+        (0x1BC80, "V"),
+        (0x1BC89, "X"),
+        (0x1BC90, "V"),
+        (0x1BC9A, "X"),
+        (0x1BC9C, "V"),
+        (0x1BCA0, "I"),
+        (0x1BCA4, "X"),
+        (0x1CF00, "V"),
+        (0x1CF2E, "X"),
+        (0x1CF30, "V"),
+        (0x1CF47, "X"),
+        (0x1CF50, "V"),
+        (0x1CFC4, "X"),
+        (0x1D000, "V"),
+        (0x1D0F6, "X"),
+        (0x1D100, "V"),
+        (0x1D127, "X"),
+        (0x1D129, "V"),
+        (0x1D15E, "M", "𝅗𝅥"),
+        (0x1D15F, "M", "𝅘𝅥"),
+        (0x1D160, "M", "𝅘𝅥𝅮"),
+        (0x1D161, "M", "𝅘𝅥𝅯"),
+        (0x1D162, "M", "𝅘𝅥𝅰"),
+        (0x1D163, "M", "𝅘𝅥𝅱"),
+        (0x1D164, "M", "𝅘𝅥𝅲"),
+        (0x1D165, "V"),
+        (0x1D173, "X"),
+        (0x1D17B, "V"),
+        (0x1D1BB, "M", "𝆹𝅥"),
+        (0x1D1BC, "M", "𝆺𝅥"),
+        (0x1D1BD, "M", "𝆹𝅥𝅮"),
+        (0x1D1BE, "M", "𝆺𝅥𝅮"),
+        (0x1D1BF, "M", "𝆹𝅥𝅯"),
+        (0x1D1C0, "M", "𝆺𝅥𝅯"),
+        (0x1D1C1, "V"),
+        (0x1D1EB, "X"),
+        (0x1D200, "V"),
+        (0x1D246, "X"),
+        (0x1D2C0, "V"),
+        (0x1D2D4, "X"),
+        (0x1D2E0, "V"),
+        (0x1D2F4, "X"),
+        (0x1D300, "V"),
+        (0x1D357, "X"),
+        (0x1D360, "V"),
+        (0x1D379, "X"),
+        (0x1D400, "M", "a"),
+        (0x1D401, "M", "b"),
+        (0x1D402, "M", "c"),
+        (0x1D403, "M", "d"),
+        (0x1D404, "M", "e"),
+        (0x1D405, "M", "f"),
+        (0x1D406, "M", "g"),
+        (0x1D407, "M", "h"),
+        (0x1D408, "M", "i"),
+        (0x1D409, "M", "j"),
+        (0x1D40A, "M", "k"),
+        (0x1D40B, "M", "l"),
+        (0x1D40C, "M", "m"),
+        (0x1D40D, "M", "n"),
+        (0x1D40E, "M", "o"),
+        (0x1D40F, "M", "p"),
+        (0x1D410, "M", "q"),
+        (0x1D411, "M", "r"),
+        (0x1D412, "M", "s"),
+        (0x1D413, "M", "t"),
+        (0x1D414, "M", "u"),
+        (0x1D415, "M", "v"),
+        (0x1D416, "M", "w"),
+        (0x1D417, "M", "x"),
+        (0x1D418, "M", "y"),
+        (0x1D419, "M", "z"),
+        (0x1D41A, "M", "a"),
+        (0x1D41B, "M", "b"),
+        (0x1D41C, "M", "c"),
+        (0x1D41D, "M", "d"),
+        (0x1D41E, "M", "e"),
+        (0x1D41F, "M", "f"),
+        (0x1D420, "M", "g"),
+        (0x1D421, "M", "h"),
+        (0x1D422, "M", "i"),
+        (0x1D423, "M", "j"),
+        (0x1D424, "M", "k"),
+    ]
+
+
+def _seg_61() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D425, "M", "l"),
+        (0x1D426, "M", "m"),
+        (0x1D427, "M", "n"),
+        (0x1D428, "M", "o"),
+        (0x1D429, "M", "p"),
+        (0x1D42A, "M", "q"),
+        (0x1D42B, "M", "r"),
+        (0x1D42C, "M", "s"),
+        (0x1D42D, "M", "t"),
+        (0x1D42E, "M", "u"),
+        (0x1D42F, "M", "v"),
+        (0x1D430, "M", "w"),
+        (0x1D431, "M", "x"),
+        (0x1D432, "M", "y"),
+        (0x1D433, "M", "z"),
+        (0x1D434, "M", "a"),
+        (0x1D435, "M", "b"),
+        (0x1D436, "M", "c"),
+        (0x1D437, "M", "d"),
+        (0x1D438, "M", "e"),
+        (0x1D439, "M", "f"),
+        (0x1D43A, "M", "g"),
+        (0x1D43B, "M", "h"),
+        (0x1D43C, "M", "i"),
+        (0x1D43D, "M", "j"),
+        (0x1D43E, "M", "k"),
+        (0x1D43F, "M", "l"),
+        (0x1D440, "M", "m"),
+        (0x1D441, "M", "n"),
+        (0x1D442, "M", "o"),
+        (0x1D443, "M", "p"),
+        (0x1D444, "M", "q"),
+        (0x1D445, "M", "r"),
+        (0x1D446, "M", "s"),
+        (0x1D447, "M", "t"),
+        (0x1D448, "M", "u"),
+        (0x1D449, "M", "v"),
+        (0x1D44A, "M", "w"),
+        (0x1D44B, "M", "x"),
+        (0x1D44C, "M", "y"),
+        (0x1D44D, "M", "z"),
+        (0x1D44E, "M", "a"),
+        (0x1D44F, "M", "b"),
+        (0x1D450, "M", "c"),
+        (0x1D451, "M", "d"),
+        (0x1D452, "M", "e"),
+        (0x1D453, "M", "f"),
+        (0x1D454, "M", "g"),
+        (0x1D455, "X"),
+        (0x1D456, "M", "i"),
+        (0x1D457, "M", "j"),
+        (0x1D458, "M", "k"),
+        (0x1D459, "M", "l"),
+        (0x1D45A, "M", "m"),
+        (0x1D45B, "M", "n"),
+        (0x1D45C, "M", "o"),
+        (0x1D45D, "M", "p"),
+        (0x1D45E, "M", "q"),
+        (0x1D45F, "M", "r"),
+        (0x1D460, "M", "s"),
+        (0x1D461, "M", "t"),
+        (0x1D462, "M", "u"),
+        (0x1D463, "M", "v"),
+        (0x1D464, "M", "w"),
+        (0x1D465, "M", "x"),
+        (0x1D466, "M", "y"),
+        (0x1D467, "M", "z"),
+        (0x1D468, "M", "a"),
+        (0x1D469, "M", "b"),
+        (0x1D46A, "M", "c"),
+        (0x1D46B, "M", "d"),
+        (0x1D46C, "M", "e"),
+        (0x1D46D, "M", "f"),
+        (0x1D46E, "M", "g"),
+        (0x1D46F, "M", "h"),
+        (0x1D470, "M", "i"),
+        (0x1D471, "M", "j"),
+        (0x1D472, "M", "k"),
+        (0x1D473, "M", "l"),
+        (0x1D474, "M", "m"),
+        (0x1D475, "M", "n"),
+        (0x1D476, "M", "o"),
+        (0x1D477, "M", "p"),
+        (0x1D478, "M", "q"),
+        (0x1D479, "M", "r"),
+        (0x1D47A, "M", "s"),
+        (0x1D47B, "M", "t"),
+        (0x1D47C, "M", "u"),
+        (0x1D47D, "M", "v"),
+        (0x1D47E, "M", "w"),
+        (0x1D47F, "M", "x"),
+        (0x1D480, "M", "y"),
+        (0x1D481, "M", "z"),
+        (0x1D482, "M", "a"),
+        (0x1D483, "M", "b"),
+        (0x1D484, "M", "c"),
+        (0x1D485, "M", "d"),
+        (0x1D486, "M", "e"),
+        (0x1D487, "M", "f"),
+        (0x1D488, "M", "g"),
+    ]
+
+
+def _seg_62() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D489, "M", "h"),
+        (0x1D48A, "M", "i"),
+        (0x1D48B, "M", "j"),
+        (0x1D48C, "M", "k"),
+        (0x1D48D, "M", "l"),
+        (0x1D48E, "M", "m"),
+        (0x1D48F, "M", "n"),
+        (0x1D490, "M", "o"),
+        (0x1D491, "M", "p"),
+        (0x1D492, "M", "q"),
+        (0x1D493, "M", "r"),
+        (0x1D494, "M", "s"),
+        (0x1D495, "M", "t"),
+        (0x1D496, "M", "u"),
+        (0x1D497, "M", "v"),
+        (0x1D498, "M", "w"),
+        (0x1D499, "M", "x"),
+        (0x1D49A, "M", "y"),
+        (0x1D49B, "M", "z"),
+        (0x1D49C, "M", "a"),
+        (0x1D49D, "X"),
+        (0x1D49E, "M", "c"),
+        (0x1D49F, "M", "d"),
+        (0x1D4A0, "X"),
+        (0x1D4A2, "M", "g"),
+        (0x1D4A3, "X"),
+        (0x1D4A5, "M", "j"),
+        (0x1D4A6, "M", "k"),
+        (0x1D4A7, "X"),
+        (0x1D4A9, "M", "n"),
+        (0x1D4AA, "M", "o"),
+        (0x1D4AB, "M", "p"),
+        (0x1D4AC, "M", "q"),
+        (0x1D4AD, "X"),
+        (0x1D4AE, "M", "s"),
+        (0x1D4AF, "M", "t"),
+        (0x1D4B0, "M", "u"),
+        (0x1D4B1, "M", "v"),
+        (0x1D4B2, "M", "w"),
+        (0x1D4B3, "M", "x"),
+        (0x1D4B4, "M", "y"),
+        (0x1D4B5, "M", "z"),
+        (0x1D4B6, "M", "a"),
+        (0x1D4B7, "M", "b"),
+        (0x1D4B8, "M", "c"),
+        (0x1D4B9, "M", "d"),
+        (0x1D4BA, "X"),
+        (0x1D4BB, "M", "f"),
+        (0x1D4BC, "X"),
+        (0x1D4BD, "M", "h"),
+        (0x1D4BE, "M", "i"),
+        (0x1D4BF, "M", "j"),
+        (0x1D4C0, "M", "k"),
+        (0x1D4C1, "M", "l"),
+        (0x1D4C2, "M", "m"),
+        (0x1D4C3, "M", "n"),
+        (0x1D4C4, "X"),
+        (0x1D4C5, "M", "p"),
+        (0x1D4C6, "M", "q"),
+        (0x1D4C7, "M", "r"),
+        (0x1D4C8, "M", "s"),
+        (0x1D4C9, "M", "t"),
+        (0x1D4CA, "M", "u"),
+        (0x1D4CB, "M", "v"),
+        (0x1D4CC, "M", "w"),
+        (0x1D4CD, "M", "x"),
+        (0x1D4CE, "M", "y"),
+        (0x1D4CF, "M", "z"),
+        (0x1D4D0, "M", "a"),
+        (0x1D4D1, "M", "b"),
+        (0x1D4D2, "M", "c"),
+        (0x1D4D3, "M", "d"),
+        (0x1D4D4, "M", "e"),
+        (0x1D4D5, "M", "f"),
+        (0x1D4D6, "M", "g"),
+        (0x1D4D7, "M", "h"),
+        (0x1D4D8, "M", "i"),
+        (0x1D4D9, "M", "j"),
+        (0x1D4DA, "M", "k"),
+        (0x1D4DB, "M", "l"),
+        (0x1D4DC, "M", "m"),
+        (0x1D4DD, "M", "n"),
+        (0x1D4DE, "M", "o"),
+        (0x1D4DF, "M", "p"),
+        (0x1D4E0, "M", "q"),
+        (0x1D4E1, "M", "r"),
+        (0x1D4E2, "M", "s"),
+        (0x1D4E3, "M", "t"),
+        (0x1D4E4, "M", "u"),
+        (0x1D4E5, "M", "v"),
+        (0x1D4E6, "M", "w"),
+        (0x1D4E7, "M", "x"),
+        (0x1D4E8, "M", "y"),
+        (0x1D4E9, "M", "z"),
+        (0x1D4EA, "M", "a"),
+        (0x1D4EB, "M", "b"),
+        (0x1D4EC, "M", "c"),
+        (0x1D4ED, "M", "d"),
+        (0x1D4EE, "M", "e"),
+        (0x1D4EF, "M", "f"),
+    ]
+
+
+def _seg_63() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D4F0, "M", "g"),
+        (0x1D4F1, "M", "h"),
+        (0x1D4F2, "M", "i"),
+        (0x1D4F3, "M", "j"),
+        (0x1D4F4, "M", "k"),
+        (0x1D4F5, "M", "l"),
+        (0x1D4F6, "M", "m"),
+        (0x1D4F7, "M", "n"),
+        (0x1D4F8, "M", "o"),
+        (0x1D4F9, "M", "p"),
+        (0x1D4FA, "M", "q"),
+        (0x1D4FB, "M", "r"),
+        (0x1D4FC, "M", "s"),
+        (0x1D4FD, "M", "t"),
+        (0x1D4FE, "M", "u"),
+        (0x1D4FF, "M", "v"),
+        (0x1D500, "M", "w"),
+        (0x1D501, "M", "x"),
+        (0x1D502, "M", "y"),
+        (0x1D503, "M", "z"),
+        (0x1D504, "M", "a"),
+        (0x1D505, "M", "b"),
+        (0x1D506, "X"),
+        (0x1D507, "M", "d"),
+        (0x1D508, "M", "e"),
+        (0x1D509, "M", "f"),
+        (0x1D50A, "M", "g"),
+        (0x1D50B, "X"),
+        (0x1D50D, "M", "j"),
+        (0x1D50E, "M", "k"),
+        (0x1D50F, "M", "l"),
+        (0x1D510, "M", "m"),
+        (0x1D511, "M", "n"),
+        (0x1D512, "M", "o"),
+        (0x1D513, "M", "p"),
+        (0x1D514, "M", "q"),
+        (0x1D515, "X"),
+        (0x1D516, "M", "s"),
+        (0x1D517, "M", "t"),
+        (0x1D518, "M", "u"),
+        (0x1D519, "M", "v"),
+        (0x1D51A, "M", "w"),
+        (0x1D51B, "M", "x"),
+        (0x1D51C, "M", "y"),
+        (0x1D51D, "X"),
+        (0x1D51E, "M", "a"),
+        (0x1D51F, "M", "b"),
+        (0x1D520, "M", "c"),
+        (0x1D521, "M", "d"),
+        (0x1D522, "M", "e"),
+        (0x1D523, "M", "f"),
+        (0x1D524, "M", "g"),
+        (0x1D525, "M", "h"),
+        (0x1D526, "M", "i"),
+        (0x1D527, "M", "j"),
+        (0x1D528, "M", "k"),
+        (0x1D529, "M", "l"),
+        (0x1D52A, "M", "m"),
+        (0x1D52B, "M", "n"),
+        (0x1D52C, "M", "o"),
+        (0x1D52D, "M", "p"),
+        (0x1D52E, "M", "q"),
+        (0x1D52F, "M", "r"),
+        (0x1D530, "M", "s"),
+        (0x1D531, "M", "t"),
+        (0x1D532, "M", "u"),
+        (0x1D533, "M", "v"),
+        (0x1D534, "M", "w"),
+        (0x1D535, "M", "x"),
+        (0x1D536, "M", "y"),
+        (0x1D537, "M", "z"),
+        (0x1D538, "M", "a"),
+        (0x1D539, "M", "b"),
+        (0x1D53A, "X"),
+        (0x1D53B, "M", "d"),
+        (0x1D53C, "M", "e"),
+        (0x1D53D, "M", "f"),
+        (0x1D53E, "M", "g"),
+        (0x1D53F, "X"),
+        (0x1D540, "M", "i"),
+        (0x1D541, "M", "j"),
+        (0x1D542, "M", "k"),
+        (0x1D543, "M", "l"),
+        (0x1D544, "M", "m"),
+        (0x1D545, "X"),
+        (0x1D546, "M", "o"),
+        (0x1D547, "X"),
+        (0x1D54A, "M", "s"),
+        (0x1D54B, "M", "t"),
+        (0x1D54C, "M", "u"),
+        (0x1D54D, "M", "v"),
+        (0x1D54E, "M", "w"),
+        (0x1D54F, "M", "x"),
+        (0x1D550, "M", "y"),
+        (0x1D551, "X"),
+        (0x1D552, "M", "a"),
+        (0x1D553, "M", "b"),
+        (0x1D554, "M", "c"),
+        (0x1D555, "M", "d"),
+        (0x1D556, "M", "e"),
+    ]
+
+
+def _seg_64() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D557, "M", "f"),
+        (0x1D558, "M", "g"),
+        (0x1D559, "M", "h"),
+        (0x1D55A, "M", "i"),
+        (0x1D55B, "M", "j"),
+        (0x1D55C, "M", "k"),
+        (0x1D55D, "M", "l"),
+        (0x1D55E, "M", "m"),
+        (0x1D55F, "M", "n"),
+        (0x1D560, "M", "o"),
+        (0x1D561, "M", "p"),
+        (0x1D562, "M", "q"),
+        (0x1D563, "M", "r"),
+        (0x1D564, "M", "s"),
+        (0x1D565, "M", "t"),
+        (0x1D566, "M", "u"),
+        (0x1D567, "M", "v"),
+        (0x1D568, "M", "w"),
+        (0x1D569, "M", "x"),
+        (0x1D56A, "M", "y"),
+        (0x1D56B, "M", "z"),
+        (0x1D56C, "M", "a"),
+        (0x1D56D, "M", "b"),
+        (0x1D56E, "M", "c"),
+        (0x1D56F, "M", "d"),
+        (0x1D570, "M", "e"),
+        (0x1D571, "M", "f"),
+        (0x1D572, "M", "g"),
+        (0x1D573, "M", "h"),
+        (0x1D574, "M", "i"),
+        (0x1D575, "M", "j"),
+        (0x1D576, "M", "k"),
+        (0x1D577, "M", "l"),
+        (0x1D578, "M", "m"),
+        (0x1D579, "M", "n"),
+        (0x1D57A, "M", "o"),
+        (0x1D57B, "M", "p"),
+        (0x1D57C, "M", "q"),
+        (0x1D57D, "M", "r"),
+        (0x1D57E, "M", "s"),
+        (0x1D57F, "M", "t"),
+        (0x1D580, "M", "u"),
+        (0x1D581, "M", "v"),
+        (0x1D582, "M", "w"),
+        (0x1D583, "M", "x"),
+        (0x1D584, "M", "y"),
+        (0x1D585, "M", "z"),
+        (0x1D586, "M", "a"),
+        (0x1D587, "M", "b"),
+        (0x1D588, "M", "c"),
+        (0x1D589, "M", "d"),
+        (0x1D58A, "M", "e"),
+        (0x1D58B, "M", "f"),
+        (0x1D58C, "M", "g"),
+        (0x1D58D, "M", "h"),
+        (0x1D58E, "M", "i"),
+        (0x1D58F, "M", "j"),
+        (0x1D590, "M", "k"),
+        (0x1D591, "M", "l"),
+        (0x1D592, "M", "m"),
+        (0x1D593, "M", "n"),
+        (0x1D594, "M", "o"),
+        (0x1D595, "M", "p"),
+        (0x1D596, "M", "q"),
+        (0x1D597, "M", "r"),
+        (0x1D598, "M", "s"),
+        (0x1D599, "M", "t"),
+        (0x1D59A, "M", "u"),
+        (0x1D59B, "M", "v"),
+        (0x1D59C, "M", "w"),
+        (0x1D59D, "M", "x"),
+        (0x1D59E, "M", "y"),
+        (0x1D59F, "M", "z"),
+        (0x1D5A0, "M", "a"),
+        (0x1D5A1, "M", "b"),
+        (0x1D5A2, "M", "c"),
+        (0x1D5A3, "M", "d"),
+        (0x1D5A4, "M", "e"),
+        (0x1D5A5, "M", "f"),
+        (0x1D5A6, "M", "g"),
+        (0x1D5A7, "M", "h"),
+        (0x1D5A8, "M", "i"),
+        (0x1D5A9, "M", "j"),
+        (0x1D5AA, "M", "k"),
+        (0x1D5AB, "M", "l"),
+        (0x1D5AC, "M", "m"),
+        (0x1D5AD, "M", "n"),
+        (0x1D5AE, "M", "o"),
+        (0x1D5AF, "M", "p"),
+        (0x1D5B0, "M", "q"),
+        (0x1D5B1, "M", "r"),
+        (0x1D5B2, "M", "s"),
+        (0x1D5B3, "M", "t"),
+        (0x1D5B4, "M", "u"),
+        (0x1D5B5, "M", "v"),
+        (0x1D5B6, "M", "w"),
+        (0x1D5B7, "M", "x"),
+        (0x1D5B8, "M", "y"),
+        (0x1D5B9, "M", "z"),
+        (0x1D5BA, "M", "a"),
+    ]
+
+
+def _seg_65() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D5BB, "M", "b"),
+        (0x1D5BC, "M", "c"),
+        (0x1D5BD, "M", "d"),
+        (0x1D5BE, "M", "e"),
+        (0x1D5BF, "M", "f"),
+        (0x1D5C0, "M", "g"),
+        (0x1D5C1, "M", "h"),
+        (0x1D5C2, "M", "i"),
+        (0x1D5C3, "M", "j"),
+        (0x1D5C4, "M", "k"),
+        (0x1D5C5, "M", "l"),
+        (0x1D5C6, "M", "m"),
+        (0x1D5C7, "M", "n"),
+        (0x1D5C8, "M", "o"),
+        (0x1D5C9, "M", "p"),
+        (0x1D5CA, "M", "q"),
+        (0x1D5CB, "M", "r"),
+        (0x1D5CC, "M", "s"),
+        (0x1D5CD, "M", "t"),
+        (0x1D5CE, "M", "u"),
+        (0x1D5CF, "M", "v"),
+        (0x1D5D0, "M", "w"),
+        (0x1D5D1, "M", "x"),
+        (0x1D5D2, "M", "y"),
+        (0x1D5D3, "M", "z"),
+        (0x1D5D4, "M", "a"),
+        (0x1D5D5, "M", "b"),
+        (0x1D5D6, "M", "c"),
+        (0x1D5D7, "M", "d"),
+        (0x1D5D8, "M", "e"),
+        (0x1D5D9, "M", "f"),
+        (0x1D5DA, "M", "g"),
+        (0x1D5DB, "M", "h"),
+        (0x1D5DC, "M", "i"),
+        (0x1D5DD, "M", "j"),
+        (0x1D5DE, "M", "k"),
+        (0x1D5DF, "M", "l"),
+        (0x1D5E0, "M", "m"),
+        (0x1D5E1, "M", "n"),
+        (0x1D5E2, "M", "o"),
+        (0x1D5E3, "M", "p"),
+        (0x1D5E4, "M", "q"),
+        (0x1D5E5, "M", "r"),
+        (0x1D5E6, "M", "s"),
+        (0x1D5E7, "M", "t"),
+        (0x1D5E8, "M", "u"),
+        (0x1D5E9, "M", "v"),
+        (0x1D5EA, "M", "w"),
+        (0x1D5EB, "M", "x"),
+        (0x1D5EC, "M", "y"),
+        (0x1D5ED, "M", "z"),
+        (0x1D5EE, "M", "a"),
+        (0x1D5EF, "M", "b"),
+        (0x1D5F0, "M", "c"),
+        (0x1D5F1, "M", "d"),
+        (0x1D5F2, "M", "e"),
+        (0x1D5F3, "M", "f"),
+        (0x1D5F4, "M", "g"),
+        (0x1D5F5, "M", "h"),
+        (0x1D5F6, "M", "i"),
+        (0x1D5F7, "M", "j"),
+        (0x1D5F8, "M", "k"),
+        (0x1D5F9, "M", "l"),
+        (0x1D5FA, "M", "m"),
+        (0x1D5FB, "M", "n"),
+        (0x1D5FC, "M", "o"),
+        (0x1D5FD, "M", "p"),
+        (0x1D5FE, "M", "q"),
+        (0x1D5FF, "M", "r"),
+        (0x1D600, "M", "s"),
+        (0x1D601, "M", "t"),
+        (0x1D602, "M", "u"),
+        (0x1D603, "M", "v"),
+        (0x1D604, "M", "w"),
+        (0x1D605, "M", "x"),
+        (0x1D606, "M", "y"),
+        (0x1D607, "M", "z"),
+        (0x1D608, "M", "a"),
+        (0x1D609, "M", "b"),
+        (0x1D60A, "M", "c"),
+        (0x1D60B, "M", "d"),
+        (0x1D60C, "M", "e"),
+        (0x1D60D, "M", "f"),
+        (0x1D60E, "M", "g"),
+        (0x1D60F, "M", "h"),
+        (0x1D610, "M", "i"),
+        (0x1D611, "M", "j"),
+        (0x1D612, "M", "k"),
+        (0x1D613, "M", "l"),
+        (0x1D614, "M", "m"),
+        (0x1D615, "M", "n"),
+        (0x1D616, "M", "o"),
+        (0x1D617, "M", "p"),
+        (0x1D618, "M", "q"),
+        (0x1D619, "M", "r"),
+        (0x1D61A, "M", "s"),
+        (0x1D61B, "M", "t"),
+        (0x1D61C, "M", "u"),
+        (0x1D61D, "M", "v"),
+        (0x1D61E, "M", "w"),
+    ]
+
+
+def _seg_66() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D61F, "M", "x"),
+        (0x1D620, "M", "y"),
+        (0x1D621, "M", "z"),
+        (0x1D622, "M", "a"),
+        (0x1D623, "M", "b"),
+        (0x1D624, "M", "c"),
+        (0x1D625, "M", "d"),
+        (0x1D626, "M", "e"),
+        (0x1D627, "M", "f"),
+        (0x1D628, "M", "g"),
+        (0x1D629, "M", "h"),
+        (0x1D62A, "M", "i"),
+        (0x1D62B, "M", "j"),
+        (0x1D62C, "M", "k"),
+        (0x1D62D, "M", "l"),
+        (0x1D62E, "M", "m"),
+        (0x1D62F, "M", "n"),
+        (0x1D630, "M", "o"),
+        (0x1D631, "M", "p"),
+        (0x1D632, "M", "q"),
+        (0x1D633, "M", "r"),
+        (0x1D634, "M", "s"),
+        (0x1D635, "M", "t"),
+        (0x1D636, "M", "u"),
+        (0x1D637, "M", "v"),
+        (0x1D638, "M", "w"),
+        (0x1D639, "M", "x"),
+        (0x1D63A, "M", "y"),
+        (0x1D63B, "M", "z"),
+        (0x1D63C, "M", "a"),
+        (0x1D63D, "M", "b"),
+        (0x1D63E, "M", "c"),
+        (0x1D63F, "M", "d"),
+        (0x1D640, "M", "e"),
+        (0x1D641, "M", "f"),
+        (0x1D642, "M", "g"),
+        (0x1D643, "M", "h"),
+        (0x1D644, "M", "i"),
+        (0x1D645, "M", "j"),
+        (0x1D646, "M", "k"),
+        (0x1D647, "M", "l"),
+        (0x1D648, "M", "m"),
+        (0x1D649, "M", "n"),
+        (0x1D64A, "M", "o"),
+        (0x1D64B, "M", "p"),
+        (0x1D64C, "M", "q"),
+        (0x1D64D, "M", "r"),
+        (0x1D64E, "M", "s"),
+        (0x1D64F, "M", "t"),
+        (0x1D650, "M", "u"),
+        (0x1D651, "M", "v"),
+        (0x1D652, "M", "w"),
+        (0x1D653, "M", "x"),
+        (0x1D654, "M", "y"),
+        (0x1D655, "M", "z"),
+        (0x1D656, "M", "a"),
+        (0x1D657, "M", "b"),
+        (0x1D658, "M", "c"),
+        (0x1D659, "M", "d"),
+        (0x1D65A, "M", "e"),
+        (0x1D65B, "M", "f"),
+        (0x1D65C, "M", "g"),
+        (0x1D65D, "M", "h"),
+        (0x1D65E, "M", "i"),
+        (0x1D65F, "M", "j"),
+        (0x1D660, "M", "k"),
+        (0x1D661, "M", "l"),
+        (0x1D662, "M", "m"),
+        (0x1D663, "M", "n"),
+        (0x1D664, "M", "o"),
+        (0x1D665, "M", "p"),
+        (0x1D666, "M", "q"),
+        (0x1D667, "M", "r"),
+        (0x1D668, "M", "s"),
+        (0x1D669, "M", "t"),
+        (0x1D66A, "M", "u"),
+        (0x1D66B, "M", "v"),
+        (0x1D66C, "M", "w"),
+        (0x1D66D, "M", "x"),
+        (0x1D66E, "M", "y"),
+        (0x1D66F, "M", "z"),
+        (0x1D670, "M", "a"),
+        (0x1D671, "M", "b"),
+        (0x1D672, "M", "c"),
+        (0x1D673, "M", "d"),
+        (0x1D674, "M", "e"),
+        (0x1D675, "M", "f"),
+        (0x1D676, "M", "g"),
+        (0x1D677, "M", "h"),
+        (0x1D678, "M", "i"),
+        (0x1D679, "M", "j"),
+        (0x1D67A, "M", "k"),
+        (0x1D67B, "M", "l"),
+        (0x1D67C, "M", "m"),
+        (0x1D67D, "M", "n"),
+        (0x1D67E, "M", "o"),
+        (0x1D67F, "M", "p"),
+        (0x1D680, "M", "q"),
+        (0x1D681, "M", "r"),
+        (0x1D682, "M", "s"),
+    ]
+
+
+def _seg_67() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D683, "M", "t"),
+        (0x1D684, "M", "u"),
+        (0x1D685, "M", "v"),
+        (0x1D686, "M", "w"),
+        (0x1D687, "M", "x"),
+        (0x1D688, "M", "y"),
+        (0x1D689, "M", "z"),
+        (0x1D68A, "M", "a"),
+        (0x1D68B, "M", "b"),
+        (0x1D68C, "M", "c"),
+        (0x1D68D, "M", "d"),
+        (0x1D68E, "M", "e"),
+        (0x1D68F, "M", "f"),
+        (0x1D690, "M", "g"),
+        (0x1D691, "M", "h"),
+        (0x1D692, "M", "i"),
+        (0x1D693, "M", "j"),
+        (0x1D694, "M", "k"),
+        (0x1D695, "M", "l"),
+        (0x1D696, "M", "m"),
+        (0x1D697, "M", "n"),
+        (0x1D698, "M", "o"),
+        (0x1D699, "M", "p"),
+        (0x1D69A, "M", "q"),
+        (0x1D69B, "M", "r"),
+        (0x1D69C, "M", "s"),
+        (0x1D69D, "M", "t"),
+        (0x1D69E, "M", "u"),
+        (0x1D69F, "M", "v"),
+        (0x1D6A0, "M", "w"),
+        (0x1D6A1, "M", "x"),
+        (0x1D6A2, "M", "y"),
+        (0x1D6A3, "M", "z"),
+        (0x1D6A4, "M", "ı"),
+        (0x1D6A5, "M", "ȷ"),
+        (0x1D6A6, "X"),
+        (0x1D6A8, "M", "α"),
+        (0x1D6A9, "M", "β"),
+        (0x1D6AA, "M", "γ"),
+        (0x1D6AB, "M", "δ"),
+        (0x1D6AC, "M", "ε"),
+        (0x1D6AD, "M", "ζ"),
+        (0x1D6AE, "M", "η"),
+        (0x1D6AF, "M", "θ"),
+        (0x1D6B0, "M", "ι"),
+        (0x1D6B1, "M", "κ"),
+        (0x1D6B2, "M", "λ"),
+        (0x1D6B3, "M", "μ"),
+        (0x1D6B4, "M", "ν"),
+        (0x1D6B5, "M", "ξ"),
+        (0x1D6B6, "M", "ο"),
+        (0x1D6B7, "M", "π"),
+        (0x1D6B8, "M", "ρ"),
+        (0x1D6B9, "M", "θ"),
+        (0x1D6BA, "M", "σ"),
+        (0x1D6BB, "M", "τ"),
+        (0x1D6BC, "M", "υ"),
+        (0x1D6BD, "M", "φ"),
+        (0x1D6BE, "M", "χ"),
+        (0x1D6BF, "M", "ψ"),
+        (0x1D6C0, "M", "ω"),
+        (0x1D6C1, "M", "∇"),
+        (0x1D6C2, "M", "α"),
+        (0x1D6C3, "M", "β"),
+        (0x1D6C4, "M", "γ"),
+        (0x1D6C5, "M", "δ"),
+        (0x1D6C6, "M", "ε"),
+        (0x1D6C7, "M", "ζ"),
+        (0x1D6C8, "M", "η"),
+        (0x1D6C9, "M", "θ"),
+        (0x1D6CA, "M", "ι"),
+        (0x1D6CB, "M", "κ"),
+        (0x1D6CC, "M", "λ"),
+        (0x1D6CD, "M", "μ"),
+        (0x1D6CE, "M", "ν"),
+        (0x1D6CF, "M", "ξ"),
+        (0x1D6D0, "M", "ο"),
+        (0x1D6D1, "M", "π"),
+        (0x1D6D2, "M", "ρ"),
+        (0x1D6D3, "M", "σ"),
+        (0x1D6D5, "M", "τ"),
+        (0x1D6D6, "M", "υ"),
+        (0x1D6D7, "M", "φ"),
+        (0x1D6D8, "M", "χ"),
+        (0x1D6D9, "M", "ψ"),
+        (0x1D6DA, "M", "ω"),
+        (0x1D6DB, "M", "∂"),
+        (0x1D6DC, "M", "ε"),
+        (0x1D6DD, "M", "θ"),
+        (0x1D6DE, "M", "κ"),
+        (0x1D6DF, "M", "φ"),
+        (0x1D6E0, "M", "ρ"),
+        (0x1D6E1, "M", "π"),
+        (0x1D6E2, "M", "α"),
+        (0x1D6E3, "M", "β"),
+        (0x1D6E4, "M", "γ"),
+        (0x1D6E5, "M", "δ"),
+        (0x1D6E6, "M", "ε"),
+        (0x1D6E7, "M", "ζ"),
+        (0x1D6E8, "M", "η"),
+    ]
+
+
+def _seg_68() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D6E9, "M", "θ"),
+        (0x1D6EA, "M", "ι"),
+        (0x1D6EB, "M", "κ"),
+        (0x1D6EC, "M", "λ"),
+        (0x1D6ED, "M", "μ"),
+        (0x1D6EE, "M", "ν"),
+        (0x1D6EF, "M", "ξ"),
+        (0x1D6F0, "M", "ο"),
+        (0x1D6F1, "M", "π"),
+        (0x1D6F2, "M", "ρ"),
+        (0x1D6F3, "M", "θ"),
+        (0x1D6F4, "M", "σ"),
+        (0x1D6F5, "M", "τ"),
+        (0x1D6F6, "M", "υ"),
+        (0x1D6F7, "M", "φ"),
+        (0x1D6F8, "M", "χ"),
+        (0x1D6F9, "M", "ψ"),
+        (0x1D6FA, "M", "ω"),
+        (0x1D6FB, "M", "∇"),
+        (0x1D6FC, "M", "α"),
+        (0x1D6FD, "M", "β"),
+        (0x1D6FE, "M", "γ"),
+        (0x1D6FF, "M", "δ"),
+        (0x1D700, "M", "ε"),
+        (0x1D701, "M", "ζ"),
+        (0x1D702, "M", "η"),
+        (0x1D703, "M", "θ"),
+        (0x1D704, "M", "ι"),
+        (0x1D705, "M", "κ"),
+        (0x1D706, "M", "λ"),
+        (0x1D707, "M", "μ"),
+        (0x1D708, "M", "ν"),
+        (0x1D709, "M", "ξ"),
+        (0x1D70A, "M", "ο"),
+        (0x1D70B, "M", "π"),
+        (0x1D70C, "M", "ρ"),
+        (0x1D70D, "M", "σ"),
+        (0x1D70F, "M", "τ"),
+        (0x1D710, "M", "υ"),
+        (0x1D711, "M", "φ"),
+        (0x1D712, "M", "χ"),
+        (0x1D713, "M", "ψ"),
+        (0x1D714, "M", "ω"),
+        (0x1D715, "M", "∂"),
+        (0x1D716, "M", "ε"),
+        (0x1D717, "M", "θ"),
+        (0x1D718, "M", "κ"),
+        (0x1D719, "M", "φ"),
+        (0x1D71A, "M", "ρ"),
+        (0x1D71B, "M", "π"),
+        (0x1D71C, "M", "α"),
+        (0x1D71D, "M", "β"),
+        (0x1D71E, "M", "γ"),
+        (0x1D71F, "M", "δ"),
+        (0x1D720, "M", "ε"),
+        (0x1D721, "M", "ζ"),
+        (0x1D722, "M", "η"),
+        (0x1D723, "M", "θ"),
+        (0x1D724, "M", "ι"),
+        (0x1D725, "M", "κ"),
+        (0x1D726, "M", "λ"),
+        (0x1D727, "M", "μ"),
+        (0x1D728, "M", "ν"),
+        (0x1D729, "M", "ξ"),
+        (0x1D72A, "M", "ο"),
+        (0x1D72B, "M", "π"),
+        (0x1D72C, "M", "ρ"),
+        (0x1D72D, "M", "θ"),
+        (0x1D72E, "M", "σ"),
+        (0x1D72F, "M", "τ"),
+        (0x1D730, "M", "υ"),
+        (0x1D731, "M", "φ"),
+        (0x1D732, "M", "χ"),
+        (0x1D733, "M", "ψ"),
+        (0x1D734, "M", "ω"),
+        (0x1D735, "M", "∇"),
+        (0x1D736, "M", "α"),
+        (0x1D737, "M", "β"),
+        (0x1D738, "M", "γ"),
+        (0x1D739, "M", "δ"),
+        (0x1D73A, "M", "ε"),
+        (0x1D73B, "M", "ζ"),
+        (0x1D73C, "M", "η"),
+        (0x1D73D, "M", "θ"),
+        (0x1D73E, "M", "ι"),
+        (0x1D73F, "M", "κ"),
+        (0x1D740, "M", "λ"),
+        (0x1D741, "M", "μ"),
+        (0x1D742, "M", "ν"),
+        (0x1D743, "M", "ξ"),
+        (0x1D744, "M", "ο"),
+        (0x1D745, "M", "π"),
+        (0x1D746, "M", "ρ"),
+        (0x1D747, "M", "σ"),
+        (0x1D749, "M", "τ"),
+        (0x1D74A, "M", "υ"),
+        (0x1D74B, "M", "φ"),
+        (0x1D74C, "M", "χ"),
+        (0x1D74D, "M", "ψ"),
+        (0x1D74E, "M", "ω"),
+    ]
+
+
+def _seg_69() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D74F, "M", "∂"),
+        (0x1D750, "M", "ε"),
+        (0x1D751, "M", "θ"),
+        (0x1D752, "M", "κ"),
+        (0x1D753, "M", "φ"),
+        (0x1D754, "M", "ρ"),
+        (0x1D755, "M", "π"),
+        (0x1D756, "M", "α"),
+        (0x1D757, "M", "β"),
+        (0x1D758, "M", "γ"),
+        (0x1D759, "M", "δ"),
+        (0x1D75A, "M", "ε"),
+        (0x1D75B, "M", "ζ"),
+        (0x1D75C, "M", "η"),
+        (0x1D75D, "M", "θ"),
+        (0x1D75E, "M", "ι"),
+        (0x1D75F, "M", "κ"),
+        (0x1D760, "M", "λ"),
+        (0x1D761, "M", "μ"),
+        (0x1D762, "M", "ν"),
+        (0x1D763, "M", "ξ"),
+        (0x1D764, "M", "ο"),
+        (0x1D765, "M", "π"),
+        (0x1D766, "M", "ρ"),
+        (0x1D767, "M", "θ"),
+        (0x1D768, "M", "σ"),
+        (0x1D769, "M", "τ"),
+        (0x1D76A, "M", "υ"),
+        (0x1D76B, "M", "φ"),
+        (0x1D76C, "M", "χ"),
+        (0x1D76D, "M", "ψ"),
+        (0x1D76E, "M", "ω"),
+        (0x1D76F, "M", "∇"),
+        (0x1D770, "M", "α"),
+        (0x1D771, "M", "β"),
+        (0x1D772, "M", "γ"),
+        (0x1D773, "M", "δ"),
+        (0x1D774, "M", "ε"),
+        (0x1D775, "M", "ζ"),
+        (0x1D776, "M", "η"),
+        (0x1D777, "M", "θ"),
+        (0x1D778, "M", "ι"),
+        (0x1D779, "M", "κ"),
+        (0x1D77A, "M", "λ"),
+        (0x1D77B, "M", "μ"),
+        (0x1D77C, "M", "ν"),
+        (0x1D77D, "M", "ξ"),
+        (0x1D77E, "M", "ο"),
+        (0x1D77F, "M", "π"),
+        (0x1D780, "M", "ρ"),
+        (0x1D781, "M", "σ"),
+        (0x1D783, "M", "τ"),
+        (0x1D784, "M", "υ"),
+        (0x1D785, "M", "φ"),
+        (0x1D786, "M", "χ"),
+        (0x1D787, "M", "ψ"),
+        (0x1D788, "M", "ω"),
+        (0x1D789, "M", "∂"),
+        (0x1D78A, "M", "ε"),
+        (0x1D78B, "M", "θ"),
+        (0x1D78C, "M", "κ"),
+        (0x1D78D, "M", "φ"),
+        (0x1D78E, "M", "ρ"),
+        (0x1D78F, "M", "π"),
+        (0x1D790, "M", "α"),
+        (0x1D791, "M", "β"),
+        (0x1D792, "M", "γ"),
+        (0x1D793, "M", "δ"),
+        (0x1D794, "M", "ε"),
+        (0x1D795, "M", "ζ"),
+        (0x1D796, "M", "η"),
+        (0x1D797, "M", "θ"),
+        (0x1D798, "M", "ι"),
+        (0x1D799, "M", "κ"),
+        (0x1D79A, "M", "λ"),
+        (0x1D79B, "M", "μ"),
+        (0x1D79C, "M", "ν"),
+        (0x1D79D, "M", "ξ"),
+        (0x1D79E, "M", "ο"),
+        (0x1D79F, "M", "π"),
+        (0x1D7A0, "M", "ρ"),
+        (0x1D7A1, "M", "θ"),
+        (0x1D7A2, "M", "σ"),
+        (0x1D7A3, "M", "τ"),
+        (0x1D7A4, "M", "υ"),
+        (0x1D7A5, "M", "φ"),
+        (0x1D7A6, "M", "χ"),
+        (0x1D7A7, "M", "ψ"),
+        (0x1D7A8, "M", "ω"),
+        (0x1D7A9, "M", "∇"),
+        (0x1D7AA, "M", "α"),
+        (0x1D7AB, "M", "β"),
+        (0x1D7AC, "M", "γ"),
+        (0x1D7AD, "M", "δ"),
+        (0x1D7AE, "M", "ε"),
+        (0x1D7AF, "M", "ζ"),
+        (0x1D7B0, "M", "η"),
+        (0x1D7B1, "M", "θ"),
+        (0x1D7B2, "M", "ι"),
+        (0x1D7B3, "M", "κ"),
+    ]
+
+
+def _seg_70() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D7B4, "M", "λ"),
+        (0x1D7B5, "M", "μ"),
+        (0x1D7B6, "M", "ν"),
+        (0x1D7B7, "M", "ξ"),
+        (0x1D7B8, "M", "ο"),
+        (0x1D7B9, "M", "π"),
+        (0x1D7BA, "M", "ρ"),
+        (0x1D7BB, "M", "σ"),
+        (0x1D7BD, "M", "τ"),
+        (0x1D7BE, "M", "υ"),
+        (0x1D7BF, "M", "φ"),
+        (0x1D7C0, "M", "χ"),
+        (0x1D7C1, "M", "ψ"),
+        (0x1D7C2, "M", "ω"),
+        (0x1D7C3, "M", "∂"),
+        (0x1D7C4, "M", "ε"),
+        (0x1D7C5, "M", "θ"),
+        (0x1D7C6, "M", "κ"),
+        (0x1D7C7, "M", "φ"),
+        (0x1D7C8, "M", "ρ"),
+        (0x1D7C9, "M", "π"),
+        (0x1D7CA, "M", "ϝ"),
+        (0x1D7CC, "X"),
+        (0x1D7CE, "M", "0"),
+        (0x1D7CF, "M", "1"),
+        (0x1D7D0, "M", "2"),
+        (0x1D7D1, "M", "3"),
+        (0x1D7D2, "M", "4"),
+        (0x1D7D3, "M", "5"),
+        (0x1D7D4, "M", "6"),
+        (0x1D7D5, "M", "7"),
+        (0x1D7D6, "M", "8"),
+        (0x1D7D7, "M", "9"),
+        (0x1D7D8, "M", "0"),
+        (0x1D7D9, "M", "1"),
+        (0x1D7DA, "M", "2"),
+        (0x1D7DB, "M", "3"),
+        (0x1D7DC, "M", "4"),
+        (0x1D7DD, "M", "5"),
+        (0x1D7DE, "M", "6"),
+        (0x1D7DF, "M", "7"),
+        (0x1D7E0, "M", "8"),
+        (0x1D7E1, "M", "9"),
+        (0x1D7E2, "M", "0"),
+        (0x1D7E3, "M", "1"),
+        (0x1D7E4, "M", "2"),
+        (0x1D7E5, "M", "3"),
+        (0x1D7E6, "M", "4"),
+        (0x1D7E7, "M", "5"),
+        (0x1D7E8, "M", "6"),
+        (0x1D7E9, "M", "7"),
+        (0x1D7EA, "M", "8"),
+        (0x1D7EB, "M", "9"),
+        (0x1D7EC, "M", "0"),
+        (0x1D7ED, "M", "1"),
+        (0x1D7EE, "M", "2"),
+        (0x1D7EF, "M", "3"),
+        (0x1D7F0, "M", "4"),
+        (0x1D7F1, "M", "5"),
+        (0x1D7F2, "M", "6"),
+        (0x1D7F3, "M", "7"),
+        (0x1D7F4, "M", "8"),
+        (0x1D7F5, "M", "9"),
+        (0x1D7F6, "M", "0"),
+        (0x1D7F7, "M", "1"),
+        (0x1D7F8, "M", "2"),
+        (0x1D7F9, "M", "3"),
+        (0x1D7FA, "M", "4"),
+        (0x1D7FB, "M", "5"),
+        (0x1D7FC, "M", "6"),
+        (0x1D7FD, "M", "7"),
+        (0x1D7FE, "M", "8"),
+        (0x1D7FF, "M", "9"),
+        (0x1D800, "V"),
+        (0x1DA8C, "X"),
+        (0x1DA9B, "V"),
+        (0x1DAA0, "X"),
+        (0x1DAA1, "V"),
+        (0x1DAB0, "X"),
+        (0x1DF00, "V"),
+        (0x1DF1F, "X"),
+        (0x1DF25, "V"),
+        (0x1DF2B, "X"),
+        (0x1E000, "V"),
+        (0x1E007, "X"),
+        (0x1E008, "V"),
+        (0x1E019, "X"),
+        (0x1E01B, "V"),
+        (0x1E022, "X"),
+        (0x1E023, "V"),
+        (0x1E025, "X"),
+        (0x1E026, "V"),
+        (0x1E02B, "X"),
+        (0x1E030, "M", "а"),
+        (0x1E031, "M", "б"),
+        (0x1E032, "M", "в"),
+        (0x1E033, "M", "г"),
+        (0x1E034, "M", "д"),
+        (0x1E035, "M", "е"),
+        (0x1E036, "M", "ж"),
+    ]
+
+
+def _seg_71() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E037, "M", "з"),
+        (0x1E038, "M", "и"),
+        (0x1E039, "M", "к"),
+        (0x1E03A, "M", "л"),
+        (0x1E03B, "M", "м"),
+        (0x1E03C, "M", "о"),
+        (0x1E03D, "M", "п"),
+        (0x1E03E, "M", "р"),
+        (0x1E03F, "M", "с"),
+        (0x1E040, "M", "т"),
+        (0x1E041, "M", "у"),
+        (0x1E042, "M", "ф"),
+        (0x1E043, "M", "х"),
+        (0x1E044, "M", "ц"),
+        (0x1E045, "M", "ч"),
+        (0x1E046, "M", "ш"),
+        (0x1E047, "M", "ы"),
+        (0x1E048, "M", "э"),
+        (0x1E049, "M", "ю"),
+        (0x1E04A, "M", "ꚉ"),
+        (0x1E04B, "M", "ә"),
+        (0x1E04C, "M", "і"),
+        (0x1E04D, "M", "ј"),
+        (0x1E04E, "M", "ө"),
+        (0x1E04F, "M", "ү"),
+        (0x1E050, "M", "ӏ"),
+        (0x1E051, "M", "а"),
+        (0x1E052, "M", "б"),
+        (0x1E053, "M", "в"),
+        (0x1E054, "M", "г"),
+        (0x1E055, "M", "д"),
+        (0x1E056, "M", "е"),
+        (0x1E057, "M", "ж"),
+        (0x1E058, "M", "з"),
+        (0x1E059, "M", "и"),
+        (0x1E05A, "M", "к"),
+        (0x1E05B, "M", "л"),
+        (0x1E05C, "M", "о"),
+        (0x1E05D, "M", "п"),
+        (0x1E05E, "M", "с"),
+        (0x1E05F, "M", "у"),
+        (0x1E060, "M", "ф"),
+        (0x1E061, "M", "х"),
+        (0x1E062, "M", "ц"),
+        (0x1E063, "M", "ч"),
+        (0x1E064, "M", "ш"),
+        (0x1E065, "M", "ъ"),
+        (0x1E066, "M", "ы"),
+        (0x1E067, "M", "ґ"),
+        (0x1E068, "M", "і"),
+        (0x1E069, "M", "ѕ"),
+        (0x1E06A, "M", "џ"),
+        (0x1E06B, "M", "ҫ"),
+        (0x1E06C, "M", "ꙑ"),
+        (0x1E06D, "M", "ұ"),
+        (0x1E06E, "X"),
+        (0x1E08F, "V"),
+        (0x1E090, "X"),
+        (0x1E100, "V"),
+        (0x1E12D, "X"),
+        (0x1E130, "V"),
+        (0x1E13E, "X"),
+        (0x1E140, "V"),
+        (0x1E14A, "X"),
+        (0x1E14E, "V"),
+        (0x1E150, "X"),
+        (0x1E290, "V"),
+        (0x1E2AF, "X"),
+        (0x1E2C0, "V"),
+        (0x1E2FA, "X"),
+        (0x1E2FF, "V"),
+        (0x1E300, "X"),
+        (0x1E4D0, "V"),
+        (0x1E4FA, "X"),
+        (0x1E7E0, "V"),
+        (0x1E7E7, "X"),
+        (0x1E7E8, "V"),
+        (0x1E7EC, "X"),
+        (0x1E7ED, "V"),
+        (0x1E7EF, "X"),
+        (0x1E7F0, "V"),
+        (0x1E7FF, "X"),
+        (0x1E800, "V"),
+        (0x1E8C5, "X"),
+        (0x1E8C7, "V"),
+        (0x1E8D7, "X"),
+        (0x1E900, "M", "𞤢"),
+        (0x1E901, "M", "𞤣"),
+        (0x1E902, "M", "𞤤"),
+        (0x1E903, "M", "𞤥"),
+        (0x1E904, "M", "𞤦"),
+        (0x1E905, "M", "𞤧"),
+        (0x1E906, "M", "𞤨"),
+        (0x1E907, "M", "𞤩"),
+        (0x1E908, "M", "𞤪"),
+        (0x1E909, "M", "𞤫"),
+        (0x1E90A, "M", "𞤬"),
+        (0x1E90B, "M", "𞤭"),
+        (0x1E90C, "M", "𞤮"),
+        (0x1E90D, "M", "𞤯"),
+    ]
+
+
+def _seg_72() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E90E, "M", "𞤰"),
+        (0x1E90F, "M", "𞤱"),
+        (0x1E910, "M", "𞤲"),
+        (0x1E911, "M", "𞤳"),
+        (0x1E912, "M", "𞤴"),
+        (0x1E913, "M", "𞤵"),
+        (0x1E914, "M", "𞤶"),
+        (0x1E915, "M", "𞤷"),
+        (0x1E916, "M", "𞤸"),
+        (0x1E917, "M", "𞤹"),
+        (0x1E918, "M", "𞤺"),
+        (0x1E919, "M", "𞤻"),
+        (0x1E91A, "M", "𞤼"),
+        (0x1E91B, "M", "𞤽"),
+        (0x1E91C, "M", "𞤾"),
+        (0x1E91D, "M", "𞤿"),
+        (0x1E91E, "M", "𞥀"),
+        (0x1E91F, "M", "𞥁"),
+        (0x1E920, "M", "𞥂"),
+        (0x1E921, "M", "𞥃"),
+        (0x1E922, "V"),
+        (0x1E94C, "X"),
+        (0x1E950, "V"),
+        (0x1E95A, "X"),
+        (0x1E95E, "V"),
+        (0x1E960, "X"),
+        (0x1EC71, "V"),
+        (0x1ECB5, "X"),
+        (0x1ED01, "V"),
+        (0x1ED3E, "X"),
+        (0x1EE00, "M", "ا"),
+        (0x1EE01, "M", "ب"),
+        (0x1EE02, "M", "ج"),
+        (0x1EE03, "M", "د"),
+        (0x1EE04, "X"),
+        (0x1EE05, "M", "و"),
+        (0x1EE06, "M", "ز"),
+        (0x1EE07, "M", "ح"),
+        (0x1EE08, "M", "ط"),
+        (0x1EE09, "M", "ي"),
+        (0x1EE0A, "M", "ك"),
+        (0x1EE0B, "M", "ل"),
+        (0x1EE0C, "M", "م"),
+        (0x1EE0D, "M", "ن"),
+        (0x1EE0E, "M", "س"),
+        (0x1EE0F, "M", "ع"),
+        (0x1EE10, "M", "ف"),
+        (0x1EE11, "M", "ص"),
+        (0x1EE12, "M", "ق"),
+        (0x1EE13, "M", "ر"),
+        (0x1EE14, "M", "ش"),
+        (0x1EE15, "M", "ت"),
+        (0x1EE16, "M", "ث"),
+        (0x1EE17, "M", "خ"),
+        (0x1EE18, "M", "ذ"),
+        (0x1EE19, "M", "ض"),
+        (0x1EE1A, "M", "ظ"),
+        (0x1EE1B, "M", "غ"),
+        (0x1EE1C, "M", "ٮ"),
+        (0x1EE1D, "M", "ں"),
+        (0x1EE1E, "M", "ڡ"),
+        (0x1EE1F, "M", "ٯ"),
+        (0x1EE20, "X"),
+        (0x1EE21, "M", "ب"),
+        (0x1EE22, "M", "ج"),
+        (0x1EE23, "X"),
+        (0x1EE24, "M", "ه"),
+        (0x1EE25, "X"),
+        (0x1EE27, "M", "ح"),
+        (0x1EE28, "X"),
+        (0x1EE29, "M", "ي"),
+        (0x1EE2A, "M", "ك"),
+        (0x1EE2B, "M", "ل"),
+        (0x1EE2C, "M", "م"),
+        (0x1EE2D, "M", "ن"),
+        (0x1EE2E, "M", "س"),
+        (0x1EE2F, "M", "ع"),
+        (0x1EE30, "M", "ف"),
+        (0x1EE31, "M", "ص"),
+        (0x1EE32, "M", "ق"),
+        (0x1EE33, "X"),
+        (0x1EE34, "M", "ش"),
+        (0x1EE35, "M", "ت"),
+        (0x1EE36, "M", "ث"),
+        (0x1EE37, "M", "خ"),
+        (0x1EE38, "X"),
+        (0x1EE39, "M", "ض"),
+        (0x1EE3A, "X"),
+        (0x1EE3B, "M", "غ"),
+        (0x1EE3C, "X"),
+        (0x1EE42, "M", "ج"),
+        (0x1EE43, "X"),
+        (0x1EE47, "M", "ح"),
+        (0x1EE48, "X"),
+        (0x1EE49, "M", "ي"),
+        (0x1EE4A, "X"),
+        (0x1EE4B, "M", "ل"),
+        (0x1EE4C, "X"),
+        (0x1EE4D, "M", "ن"),
+        (0x1EE4E, "M", "س"),
+    ]
+
+
+def _seg_73() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EE4F, "M", "ع"),
+        (0x1EE50, "X"),
+        (0x1EE51, "M", "ص"),
+        (0x1EE52, "M", "ق"),
+        (0x1EE53, "X"),
+        (0x1EE54, "M", "ش"),
+        (0x1EE55, "X"),
+        (0x1EE57, "M", "خ"),
+        (0x1EE58, "X"),
+        (0x1EE59, "M", "ض"),
+        (0x1EE5A, "X"),
+        (0x1EE5B, "M", "غ"),
+        (0x1EE5C, "X"),
+        (0x1EE5D, "M", "ں"),
+        (0x1EE5E, "X"),
+        (0x1EE5F, "M", "ٯ"),
+        (0x1EE60, "X"),
+        (0x1EE61, "M", "ب"),
+        (0x1EE62, "M", "ج"),
+        (0x1EE63, "X"),
+        (0x1EE64, "M", "ه"),
+        (0x1EE65, "X"),
+        (0x1EE67, "M", "ح"),
+        (0x1EE68, "M", "ط"),
+        (0x1EE69, "M", "ي"),
+        (0x1EE6A, "M", "ك"),
+        (0x1EE6B, "X"),
+        (0x1EE6C, "M", "م"),
+        (0x1EE6D, "M", "ن"),
+        (0x1EE6E, "M", "س"),
+        (0x1EE6F, "M", "ع"),
+        (0x1EE70, "M", "ف"),
+        (0x1EE71, "M", "ص"),
+        (0x1EE72, "M", "ق"),
+        (0x1EE73, "X"),
+        (0x1EE74, "M", "ش"),
+        (0x1EE75, "M", "ت"),
+        (0x1EE76, "M", "ث"),
+        (0x1EE77, "M", "خ"),
+        (0x1EE78, "X"),
+        (0x1EE79, "M", "ض"),
+        (0x1EE7A, "M", "ظ"),
+        (0x1EE7B, "M", "غ"),
+        (0x1EE7C, "M", "ٮ"),
+        (0x1EE7D, "X"),
+        (0x1EE7E, "M", "ڡ"),
+        (0x1EE7F, "X"),
+        (0x1EE80, "M", "ا"),
+        (0x1EE81, "M", "ب"),
+        (0x1EE82, "M", "ج"),
+        (0x1EE83, "M", "د"),
+        (0x1EE84, "M", "ه"),
+        (0x1EE85, "M", "و"),
+        (0x1EE86, "M", "ز"),
+        (0x1EE87, "M", "ح"),
+        (0x1EE88, "M", "ط"),
+        (0x1EE89, "M", "ي"),
+        (0x1EE8A, "X"),
+        (0x1EE8B, "M", "ل"),
+        (0x1EE8C, "M", "م"),
+        (0x1EE8D, "M", "ن"),
+        (0x1EE8E, "M", "س"),
+        (0x1EE8F, "M", "ع"),
+        (0x1EE90, "M", "ف"),
+        (0x1EE91, "M", "ص"),
+        (0x1EE92, "M", "ق"),
+        (0x1EE93, "M", "ر"),
+        (0x1EE94, "M", "ش"),
+        (0x1EE95, "M", "ت"),
+        (0x1EE96, "M", "ث"),
+        (0x1EE97, "M", "خ"),
+        (0x1EE98, "M", "ذ"),
+        (0x1EE99, "M", "ض"),
+        (0x1EE9A, "M", "ظ"),
+        (0x1EE9B, "M", "غ"),
+        (0x1EE9C, "X"),
+        (0x1EEA1, "M", "ب"),
+        (0x1EEA2, "M", "ج"),
+        (0x1EEA3, "M", "د"),
+        (0x1EEA4, "X"),
+        (0x1EEA5, "M", "و"),
+        (0x1EEA6, "M", "ز"),
+        (0x1EEA7, "M", "ح"),
+        (0x1EEA8, "M", "ط"),
+        (0x1EEA9, "M", "ي"),
+        (0x1EEAA, "X"),
+        (0x1EEAB, "M", "ل"),
+        (0x1EEAC, "M", "م"),
+        (0x1EEAD, "M", "ن"),
+        (0x1EEAE, "M", "س"),
+        (0x1EEAF, "M", "ع"),
+        (0x1EEB0, "M", "ف"),
+        (0x1EEB1, "M", "ص"),
+        (0x1EEB2, "M", "ق"),
+        (0x1EEB3, "M", "ر"),
+        (0x1EEB4, "M", "ش"),
+        (0x1EEB5, "M", "ت"),
+        (0x1EEB6, "M", "ث"),
+        (0x1EEB7, "M", "خ"),
+        (0x1EEB8, "M", "ذ"),
+    ]
+
+
+def _seg_74() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EEB9, "M", "ض"),
+        (0x1EEBA, "M", "ظ"),
+        (0x1EEBB, "M", "غ"),
+        (0x1EEBC, "X"),
+        (0x1EEF0, "V"),
+        (0x1EEF2, "X"),
+        (0x1F000, "V"),
+        (0x1F02C, "X"),
+        (0x1F030, "V"),
+        (0x1F094, "X"),
+        (0x1F0A0, "V"),
+        (0x1F0AF, "X"),
+        (0x1F0B1, "V"),
+        (0x1F0C0, "X"),
+        (0x1F0C1, "V"),
+        (0x1F0D0, "X"),
+        (0x1F0D1, "V"),
+        (0x1F0F6, "X"),
+        (0x1F101, "3", "0,"),
+        (0x1F102, "3", "1,"),
+        (0x1F103, "3", "2,"),
+        (0x1F104, "3", "3,"),
+        (0x1F105, "3", "4,"),
+        (0x1F106, "3", "5,"),
+        (0x1F107, "3", "6,"),
+        (0x1F108, "3", "7,"),
+        (0x1F109, "3", "8,"),
+        (0x1F10A, "3", "9,"),
+        (0x1F10B, "V"),
+        (0x1F110, "3", "(a)"),
+        (0x1F111, "3", "(b)"),
+        (0x1F112, "3", "(c)"),
+        (0x1F113, "3", "(d)"),
+        (0x1F114, "3", "(e)"),
+        (0x1F115, "3", "(f)"),
+        (0x1F116, "3", "(g)"),
+        (0x1F117, "3", "(h)"),
+        (0x1F118, "3", "(i)"),
+        (0x1F119, "3", "(j)"),
+        (0x1F11A, "3", "(k)"),
+        (0x1F11B, "3", "(l)"),
+        (0x1F11C, "3", "(m)"),
+        (0x1F11D, "3", "(n)"),
+        (0x1F11E, "3", "(o)"),
+        (0x1F11F, "3", "(p)"),
+        (0x1F120, "3", "(q)"),
+        (0x1F121, "3", "(r)"),
+        (0x1F122, "3", "(s)"),
+        (0x1F123, "3", "(t)"),
+        (0x1F124, "3", "(u)"),
+        (0x1F125, "3", "(v)"),
+        (0x1F126, "3", "(w)"),
+        (0x1F127, "3", "(x)"),
+        (0x1F128, "3", "(y)"),
+        (0x1F129, "3", "(z)"),
+        (0x1F12A, "M", "〔s〕"),
+        (0x1F12B, "M", "c"),
+        (0x1F12C, "M", "r"),
+        (0x1F12D, "M", "cd"),
+        (0x1F12E, "M", "wz"),
+        (0x1F12F, "V"),
+        (0x1F130, "M", "a"),
+        (0x1F131, "M", "b"),
+        (0x1F132, "M", "c"),
+        (0x1F133, "M", "d"),
+        (0x1F134, "M", "e"),
+        (0x1F135, "M", "f"),
+        (0x1F136, "M", "g"),
+        (0x1F137, "M", "h"),
+        (0x1F138, "M", "i"),
+        (0x1F139, "M", "j"),
+        (0x1F13A, "M", "k"),
+        (0x1F13B, "M", "l"),
+        (0x1F13C, "M", "m"),
+        (0x1F13D, "M", "n"),
+        (0x1F13E, "M", "o"),
+        (0x1F13F, "M", "p"),
+        (0x1F140, "M", "q"),
+        (0x1F141, "M", "r"),
+        (0x1F142, "M", "s"),
+        (0x1F143, "M", "t"),
+        (0x1F144, "M", "u"),
+        (0x1F145, "M", "v"),
+        (0x1F146, "M", "w"),
+        (0x1F147, "M", "x"),
+        (0x1F148, "M", "y"),
+        (0x1F149, "M", "z"),
+        (0x1F14A, "M", "hv"),
+        (0x1F14B, "M", "mv"),
+        (0x1F14C, "M", "sd"),
+        (0x1F14D, "M", "ss"),
+        (0x1F14E, "M", "ppv"),
+        (0x1F14F, "M", "wc"),
+        (0x1F150, "V"),
+        (0x1F16A, "M", "mc"),
+        (0x1F16B, "M", "md"),
+        (0x1F16C, "M", "mr"),
+        (0x1F16D, "V"),
+        (0x1F190, "M", "dj"),
+        (0x1F191, "V"),
+    ]
+
+
+def _seg_75() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1F1AE, "X"),
+        (0x1F1E6, "V"),
+        (0x1F200, "M", "ほか"),
+        (0x1F201, "M", "ココ"),
+        (0x1F202, "M", "サ"),
+        (0x1F203, "X"),
+        (0x1F210, "M", "手"),
+        (0x1F211, "M", "字"),
+        (0x1F212, "M", "双"),
+        (0x1F213, "M", "デ"),
+        (0x1F214, "M", "二"),
+        (0x1F215, "M", "多"),
+        (0x1F216, "M", "解"),
+        (0x1F217, "M", "天"),
+        (0x1F218, "M", "交"),
+        (0x1F219, "M", "映"),
+        (0x1F21A, "M", "無"),
+        (0x1F21B, "M", "料"),
+        (0x1F21C, "M", "前"),
+        (0x1F21D, "M", "後"),
+        (0x1F21E, "M", "再"),
+        (0x1F21F, "M", "新"),
+        (0x1F220, "M", "初"),
+        (0x1F221, "M", "終"),
+        (0x1F222, "M", "生"),
+        (0x1F223, "M", "販"),
+        (0x1F224, "M", "声"),
+        (0x1F225, "M", "吹"),
+        (0x1F226, "M", "演"),
+        (0x1F227, "M", "投"),
+        (0x1F228, "M", "捕"),
+        (0x1F229, "M", "一"),
+        (0x1F22A, "M", "三"),
+        (0x1F22B, "M", "遊"),
+        (0x1F22C, "M", "左"),
+        (0x1F22D, "M", "中"),
+        (0x1F22E, "M", "右"),
+        (0x1F22F, "M", "指"),
+        (0x1F230, "M", "走"),
+        (0x1F231, "M", "打"),
+        (0x1F232, "M", "禁"),
+        (0x1F233, "M", "空"),
+        (0x1F234, "M", "合"),
+        (0x1F235, "M", "満"),
+        (0x1F236, "M", "有"),
+        (0x1F237, "M", "月"),
+        (0x1F238, "M", "申"),
+        (0x1F239, "M", "割"),
+        (0x1F23A, "M", "営"),
+        (0x1F23B, "M", "配"),
+        (0x1F23C, "X"),
+        (0x1F240, "M", "〔本〕"),
+        (0x1F241, "M", "〔三〕"),
+        (0x1F242, "M", "〔二〕"),
+        (0x1F243, "M", "〔安〕"),
+        (0x1F244, "M", "〔点〕"),
+        (0x1F245, "M", "〔打〕"),
+        (0x1F246, "M", "〔盗〕"),
+        (0x1F247, "M", "〔勝〕"),
+        (0x1F248, "M", "〔敗〕"),
+        (0x1F249, "X"),
+        (0x1F250, "M", "得"),
+        (0x1F251, "M", "可"),
+        (0x1F252, "X"),
+        (0x1F260, "V"),
+        (0x1F266, "X"),
+        (0x1F300, "V"),
+        (0x1F6D8, "X"),
+        (0x1F6DC, "V"),
+        (0x1F6ED, "X"),
+        (0x1F6F0, "V"),
+        (0x1F6FD, "X"),
+        (0x1F700, "V"),
+        (0x1F777, "X"),
+        (0x1F77B, "V"),
+        (0x1F7DA, "X"),
+        (0x1F7E0, "V"),
+        (0x1F7EC, "X"),
+        (0x1F7F0, "V"),
+        (0x1F7F1, "X"),
+        (0x1F800, "V"),
+        (0x1F80C, "X"),
+        (0x1F810, "V"),
+        (0x1F848, "X"),
+        (0x1F850, "V"),
+        (0x1F85A, "X"),
+        (0x1F860, "V"),
+        (0x1F888, "X"),
+        (0x1F890, "V"),
+        (0x1F8AE, "X"),
+        (0x1F8B0, "V"),
+        (0x1F8B2, "X"),
+        (0x1F900, "V"),
+        (0x1FA54, "X"),
+        (0x1FA60, "V"),
+        (0x1FA6E, "X"),
+        (0x1FA70, "V"),
+        (0x1FA7D, "X"),
+        (0x1FA80, "V"),
+        (0x1FA89, "X"),
+    ]
+
+
+def _seg_76() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1FA90, "V"),
+        (0x1FABE, "X"),
+        (0x1FABF, "V"),
+        (0x1FAC6, "X"),
+        (0x1FACE, "V"),
+        (0x1FADC, "X"),
+        (0x1FAE0, "V"),
+        (0x1FAE9, "X"),
+        (0x1FAF0, "V"),
+        (0x1FAF9, "X"),
+        (0x1FB00, "V"),
+        (0x1FB93, "X"),
+        (0x1FB94, "V"),
+        (0x1FBCB, "X"),
+        (0x1FBF0, "M", "0"),
+        (0x1FBF1, "M", "1"),
+        (0x1FBF2, "M", "2"),
+        (0x1FBF3, "M", "3"),
+        (0x1FBF4, "M", "4"),
+        (0x1FBF5, "M", "5"),
+        (0x1FBF6, "M", "6"),
+        (0x1FBF7, "M", "7"),
+        (0x1FBF8, "M", "8"),
+        (0x1FBF9, "M", "9"),
+        (0x1FBFA, "X"),
+        (0x20000, "V"),
+        (0x2A6E0, "X"),
+        (0x2A700, "V"),
+        (0x2B73A, "X"),
+        (0x2B740, "V"),
+        (0x2B81E, "X"),
+        (0x2B820, "V"),
+        (0x2CEA2, "X"),
+        (0x2CEB0, "V"),
+        (0x2EBE1, "X"),
+        (0x2EBF0, "V"),
+        (0x2EE5E, "X"),
+        (0x2F800, "M", "丽"),
+        (0x2F801, "M", "丸"),
+        (0x2F802, "M", "乁"),
+        (0x2F803, "M", "𠄢"),
+        (0x2F804, "M", "你"),
+        (0x2F805, "M", "侮"),
+        (0x2F806, "M", "侻"),
+        (0x2F807, "M", "倂"),
+        (0x2F808, "M", "偺"),
+        (0x2F809, "M", "備"),
+        (0x2F80A, "M", "僧"),
+        (0x2F80B, "M", "像"),
+        (0x2F80C, "M", "㒞"),
+        (0x2F80D, "M", "𠘺"),
+        (0x2F80E, "M", "免"),
+        (0x2F80F, "M", "兔"),
+        (0x2F810, "M", "兤"),
+        (0x2F811, "M", "具"),
+        (0x2F812, "M", "𠔜"),
+        (0x2F813, "M", "㒹"),
+        (0x2F814, "M", "內"),
+        (0x2F815, "M", "再"),
+        (0x2F816, "M", "𠕋"),
+        (0x2F817, "M", "冗"),
+        (0x2F818, "M", "冤"),
+        (0x2F819, "M", "仌"),
+        (0x2F81A, "M", "冬"),
+        (0x2F81B, "M", "况"),
+        (0x2F81C, "M", "𩇟"),
+        (0x2F81D, "M", "凵"),
+        (0x2F81E, "M", "刃"),
+        (0x2F81F, "M", "㓟"),
+        (0x2F820, "M", "刻"),
+        (0x2F821, "M", "剆"),
+        (0x2F822, "M", "割"),
+        (0x2F823, "M", "剷"),
+        (0x2F824, "M", "㔕"),
+        (0x2F825, "M", "勇"),
+        (0x2F826, "M", "勉"),
+        (0x2F827, "M", "勤"),
+        (0x2F828, "M", "勺"),
+        (0x2F829, "M", "包"),
+        (0x2F82A, "M", "匆"),
+        (0x2F82B, "M", "北"),
+        (0x2F82C, "M", "卉"),
+        (0x2F82D, "M", "卑"),
+        (0x2F82E, "M", "博"),
+        (0x2F82F, "M", "即"),
+        (0x2F830, "M", "卽"),
+        (0x2F831, "M", "卿"),
+        (0x2F834, "M", "𠨬"),
+        (0x2F835, "M", "灰"),
+        (0x2F836, "M", "及"),
+        (0x2F837, "M", "叟"),
+        (0x2F838, "M", "𠭣"),
+        (0x2F839, "M", "叫"),
+        (0x2F83A, "M", "叱"),
+        (0x2F83B, "M", "吆"),
+        (0x2F83C, "M", "咞"),
+        (0x2F83D, "M", "吸"),
+        (0x2F83E, "M", "呈"),
+        (0x2F83F, "M", "周"),
+        (0x2F840, "M", "咢"),
+    ]
+
+
+def _seg_77() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F841, "M", "哶"),
+        (0x2F842, "M", "唐"),
+        (0x2F843, "M", "啓"),
+        (0x2F844, "M", "啣"),
+        (0x2F845, "M", "善"),
+        (0x2F847, "M", "喙"),
+        (0x2F848, "M", "喫"),
+        (0x2F849, "M", "喳"),
+        (0x2F84A, "M", "嗂"),
+        (0x2F84B, "M", "圖"),
+        (0x2F84C, "M", "嘆"),
+        (0x2F84D, "M", "圗"),
+        (0x2F84E, "M", "噑"),
+        (0x2F84F, "M", "噴"),
+        (0x2F850, "M", "切"),
+        (0x2F851, "M", "壮"),
+        (0x2F852, "M", "城"),
+        (0x2F853, "M", "埴"),
+        (0x2F854, "M", "堍"),
+        (0x2F855, "M", "型"),
+        (0x2F856, "M", "堲"),
+        (0x2F857, "M", "報"),
+        (0x2F858, "M", "墬"),
+        (0x2F859, "M", "𡓤"),
+        (0x2F85A, "M", "売"),
+        (0x2F85B, "M", "壷"),
+        (0x2F85C, "M", "夆"),
+        (0x2F85D, "M", "多"),
+        (0x2F85E, "M", "夢"),
+        (0x2F85F, "M", "奢"),
+        (0x2F860, "M", "𡚨"),
+        (0x2F861, "M", "𡛪"),
+        (0x2F862, "M", "姬"),
+        (0x2F863, "M", "娛"),
+        (0x2F864, "M", "娧"),
+        (0x2F865, "M", "姘"),
+        (0x2F866, "M", "婦"),
+        (0x2F867, "M", "㛮"),
+        (0x2F868, "X"),
+        (0x2F869, "M", "嬈"),
+        (0x2F86A, "M", "嬾"),
+        (0x2F86C, "M", "𡧈"),
+        (0x2F86D, "M", "寃"),
+        (0x2F86E, "M", "寘"),
+        (0x2F86F, "M", "寧"),
+        (0x2F870, "M", "寳"),
+        (0x2F871, "M", "𡬘"),
+        (0x2F872, "M", "寿"),
+        (0x2F873, "M", "将"),
+        (0x2F874, "X"),
+        (0x2F875, "M", "尢"),
+        (0x2F876, "M", "㞁"),
+        (0x2F877, "M", "屠"),
+        (0x2F878, "M", "屮"),
+        (0x2F879, "M", "峀"),
+        (0x2F87A, "M", "岍"),
+        (0x2F87B, "M", "𡷤"),
+        (0x2F87C, "M", "嵃"),
+        (0x2F87D, "M", "𡷦"),
+        (0x2F87E, "M", "嵮"),
+        (0x2F87F, "M", "嵫"),
+        (0x2F880, "M", "嵼"),
+        (0x2F881, "M", "巡"),
+        (0x2F882, "M", "巢"),
+        (0x2F883, "M", "㠯"),
+        (0x2F884, "M", "巽"),
+        (0x2F885, "M", "帨"),
+        (0x2F886, "M", "帽"),
+        (0x2F887, "M", "幩"),
+        (0x2F888, "M", "㡢"),
+        (0x2F889, "M", "𢆃"),
+        (0x2F88A, "M", "㡼"),
+        (0x2F88B, "M", "庰"),
+        (0x2F88C, "M", "庳"),
+        (0x2F88D, "M", "庶"),
+        (0x2F88E, "M", "廊"),
+        (0x2F88F, "M", "𪎒"),
+        (0x2F890, "M", "廾"),
+        (0x2F891, "M", "𢌱"),
+        (0x2F893, "M", "舁"),
+        (0x2F894, "M", "弢"),
+        (0x2F896, "M", "㣇"),
+        (0x2F897, "M", "𣊸"),
+        (0x2F898, "M", "𦇚"),
+        (0x2F899, "M", "形"),
+        (0x2F89A, "M", "彫"),
+        (0x2F89B, "M", "㣣"),
+        (0x2F89C, "M", "徚"),
+        (0x2F89D, "M", "忍"),
+        (0x2F89E, "M", "志"),
+        (0x2F89F, "M", "忹"),
+        (0x2F8A0, "M", "悁"),
+        (0x2F8A1, "M", "㤺"),
+        (0x2F8A2, "M", "㤜"),
+        (0x2F8A3, "M", "悔"),
+        (0x2F8A4, "M", "𢛔"),
+        (0x2F8A5, "M", "惇"),
+        (0x2F8A6, "M", "慈"),
+        (0x2F8A7, "M", "慌"),
+        (0x2F8A8, "M", "慎"),
+    ]
+
+
+def _seg_78() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F8A9, "M", "慌"),
+        (0x2F8AA, "M", "慺"),
+        (0x2F8AB, "M", "憎"),
+        (0x2F8AC, "M", "憲"),
+        (0x2F8AD, "M", "憤"),
+        (0x2F8AE, "M", "憯"),
+        (0x2F8AF, "M", "懞"),
+        (0x2F8B0, "M", "懲"),
+        (0x2F8B1, "M", "懶"),
+        (0x2F8B2, "M", "成"),
+        (0x2F8B3, "M", "戛"),
+        (0x2F8B4, "M", "扝"),
+        (0x2F8B5, "M", "抱"),
+        (0x2F8B6, "M", "拔"),
+        (0x2F8B7, "M", "捐"),
+        (0x2F8B8, "M", "𢬌"),
+        (0x2F8B9, "M", "挽"),
+        (0x2F8BA, "M", "拼"),
+        (0x2F8BB, "M", "捨"),
+        (0x2F8BC, "M", "掃"),
+        (0x2F8BD, "M", "揤"),
+        (0x2F8BE, "M", "𢯱"),
+        (0x2F8BF, "M", "搢"),
+        (0x2F8C0, "M", "揅"),
+        (0x2F8C1, "M", "掩"),
+        (0x2F8C2, "M", "㨮"),
+        (0x2F8C3, "M", "摩"),
+        (0x2F8C4, "M", "摾"),
+        (0x2F8C5, "M", "撝"),
+        (0x2F8C6, "M", "摷"),
+        (0x2F8C7, "M", "㩬"),
+        (0x2F8C8, "M", "敏"),
+        (0x2F8C9, "M", "敬"),
+        (0x2F8CA, "M", "𣀊"),
+        (0x2F8CB, "M", "旣"),
+        (0x2F8CC, "M", "書"),
+        (0x2F8CD, "M", "晉"),
+        (0x2F8CE, "M", "㬙"),
+        (0x2F8CF, "M", "暑"),
+        (0x2F8D0, "M", "㬈"),
+        (0x2F8D1, "M", "㫤"),
+        (0x2F8D2, "M", "冒"),
+        (0x2F8D3, "M", "冕"),
+        (0x2F8D4, "M", "最"),
+        (0x2F8D5, "M", "暜"),
+        (0x2F8D6, "M", "肭"),
+        (0x2F8D7, "M", "䏙"),
+        (0x2F8D8, "M", "朗"),
+        (0x2F8D9, "M", "望"),
+        (0x2F8DA, "M", "朡"),
+        (0x2F8DB, "M", "杞"),
+        (0x2F8DC, "M", "杓"),
+        (0x2F8DD, "M", "𣏃"),
+        (0x2F8DE, "M", "㭉"),
+        (0x2F8DF, "M", "柺"),
+        (0x2F8E0, "M", "枅"),
+        (0x2F8E1, "M", "桒"),
+        (0x2F8E2, "M", "梅"),
+        (0x2F8E3, "M", "𣑭"),
+        (0x2F8E4, "M", "梎"),
+        (0x2F8E5, "M", "栟"),
+        (0x2F8E6, "M", "椔"),
+        (0x2F8E7, "M", "㮝"),
+        (0x2F8E8, "M", "楂"),
+        (0x2F8E9, "M", "榣"),
+        (0x2F8EA, "M", "槪"),
+        (0x2F8EB, "M", "檨"),
+        (0x2F8EC, "M", "𣚣"),
+        (0x2F8ED, "M", "櫛"),
+        (0x2F8EE, "M", "㰘"),
+        (0x2F8EF, "M", "次"),
+        (0x2F8F0, "M", "𣢧"),
+        (0x2F8F1, "M", "歔"),
+        (0x2F8F2, "M", "㱎"),
+        (0x2F8F3, "M", "歲"),
+        (0x2F8F4, "M", "殟"),
+        (0x2F8F5, "M", "殺"),
+        (0x2F8F6, "M", "殻"),
+        (0x2F8F7, "M", "𣪍"),
+        (0x2F8F8, "M", "𡴋"),
+        (0x2F8F9, "M", "𣫺"),
+        (0x2F8FA, "M", "汎"),
+        (0x2F8FB, "M", "𣲼"),
+        (0x2F8FC, "M", "沿"),
+        (0x2F8FD, "M", "泍"),
+        (0x2F8FE, "M", "汧"),
+        (0x2F8FF, "M", "洖"),
+        (0x2F900, "M", "派"),
+        (0x2F901, "M", "海"),
+        (0x2F902, "M", "流"),
+        (0x2F903, "M", "浩"),
+        (0x2F904, "M", "浸"),
+        (0x2F905, "M", "涅"),
+        (0x2F906, "M", "𣴞"),
+        (0x2F907, "M", "洴"),
+        (0x2F908, "M", "港"),
+        (0x2F909, "M", "湮"),
+        (0x2F90A, "M", "㴳"),
+        (0x2F90B, "M", "滋"),
+        (0x2F90C, "M", "滇"),
+    ]
+
+
+def _seg_79() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F90D, "M", "𣻑"),
+        (0x2F90E, "M", "淹"),
+        (0x2F90F, "M", "潮"),
+        (0x2F910, "M", "𣽞"),
+        (0x2F911, "M", "𣾎"),
+        (0x2F912, "M", "濆"),
+        (0x2F913, "M", "瀹"),
+        (0x2F914, "M", "瀞"),
+        (0x2F915, "M", "瀛"),
+        (0x2F916, "M", "㶖"),
+        (0x2F917, "M", "灊"),
+        (0x2F918, "M", "災"),
+        (0x2F919, "M", "灷"),
+        (0x2F91A, "M", "炭"),
+        (0x2F91B, "M", "𠔥"),
+        (0x2F91C, "M", "煅"),
+        (0x2F91D, "M", "𤉣"),
+        (0x2F91E, "M", "熜"),
+        (0x2F91F, "X"),
+        (0x2F920, "M", "爨"),
+        (0x2F921, "M", "爵"),
+        (0x2F922, "M", "牐"),
+        (0x2F923, "M", "𤘈"),
+        (0x2F924, "M", "犀"),
+        (0x2F925, "M", "犕"),
+        (0x2F926, "M", "𤜵"),
+        (0x2F927, "M", "𤠔"),
+        (0x2F928, "M", "獺"),
+        (0x2F929, "M", "王"),
+        (0x2F92A, "M", "㺬"),
+        (0x2F92B, "M", "玥"),
+        (0x2F92C, "M", "㺸"),
+        (0x2F92E, "M", "瑇"),
+        (0x2F92F, "M", "瑜"),
+        (0x2F930, "M", "瑱"),
+        (0x2F931, "M", "璅"),
+        (0x2F932, "M", "瓊"),
+        (0x2F933, "M", "㼛"),
+        (0x2F934, "M", "甤"),
+        (0x2F935, "M", "𤰶"),
+        (0x2F936, "M", "甾"),
+        (0x2F937, "M", "𤲒"),
+        (0x2F938, "M", "異"),
+        (0x2F939, "M", "𢆟"),
+        (0x2F93A, "M", "瘐"),
+        (0x2F93B, "M", "𤾡"),
+        (0x2F93C, "M", "𤾸"),
+        (0x2F93D, "M", "𥁄"),
+        (0x2F93E, "M", "㿼"),
+        (0x2F93F, "M", "䀈"),
+        (0x2F940, "M", "直"),
+        (0x2F941, "M", "𥃳"),
+        (0x2F942, "M", "𥃲"),
+        (0x2F943, "M", "𥄙"),
+        (0x2F944, "M", "𥄳"),
+        (0x2F945, "M", "眞"),
+        (0x2F946, "M", "真"),
+        (0x2F948, "M", "睊"),
+        (0x2F949, "M", "䀹"),
+        (0x2F94A, "M", "瞋"),
+        (0x2F94B, "M", "䁆"),
+        (0x2F94C, "M", "䂖"),
+        (0x2F94D, "M", "𥐝"),
+        (0x2F94E, "M", "硎"),
+        (0x2F94F, "M", "碌"),
+        (0x2F950, "M", "磌"),
+        (0x2F951, "M", "䃣"),
+        (0x2F952, "M", "𥘦"),
+        (0x2F953, "M", "祖"),
+        (0x2F954, "M", "𥚚"),
+        (0x2F955, "M", "𥛅"),
+        (0x2F956, "M", "福"),
+        (0x2F957, "M", "秫"),
+        (0x2F958, "M", "䄯"),
+        (0x2F959, "M", "穀"),
+        (0x2F95A, "M", "穊"),
+        (0x2F95B, "M", "穏"),
+        (0x2F95C, "M", "𥥼"),
+        (0x2F95D, "M", "𥪧"),
+        (0x2F95F, "X"),
+        (0x2F960, "M", "䈂"),
+        (0x2F961, "M", "𥮫"),
+        (0x2F962, "M", "篆"),
+        (0x2F963, "M", "築"),
+        (0x2F964, "M", "䈧"),
+        (0x2F965, "M", "𥲀"),
+        (0x2F966, "M", "糒"),
+        (0x2F967, "M", "䊠"),
+        (0x2F968, "M", "糨"),
+        (0x2F969, "M", "糣"),
+        (0x2F96A, "M", "紀"),
+        (0x2F96B, "M", "𥾆"),
+        (0x2F96C, "M", "絣"),
+        (0x2F96D, "M", "䌁"),
+        (0x2F96E, "M", "緇"),
+        (0x2F96F, "M", "縂"),
+        (0x2F970, "M", "繅"),
+        (0x2F971, "M", "䌴"),
+        (0x2F972, "M", "𦈨"),
+        (0x2F973, "M", "𦉇"),
+    ]
+
+
+def _seg_80() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F974, "M", "䍙"),
+        (0x2F975, "M", "𦋙"),
+        (0x2F976, "M", "罺"),
+        (0x2F977, "M", "𦌾"),
+        (0x2F978, "M", "羕"),
+        (0x2F979, "M", "翺"),
+        (0x2F97A, "M", "者"),
+        (0x2F97B, "M", "𦓚"),
+        (0x2F97C, "M", "𦔣"),
+        (0x2F97D, "M", "聠"),
+        (0x2F97E, "M", "𦖨"),
+        (0x2F97F, "M", "聰"),
+        (0x2F980, "M", "𣍟"),
+        (0x2F981, "M", "䏕"),
+        (0x2F982, "M", "育"),
+        (0x2F983, "M", "脃"),
+        (0x2F984, "M", "䐋"),
+        (0x2F985, "M", "脾"),
+        (0x2F986, "M", "媵"),
+        (0x2F987, "M", "𦞧"),
+        (0x2F988, "M", "𦞵"),
+        (0x2F989, "M", "𣎓"),
+        (0x2F98A, "M", "𣎜"),
+        (0x2F98B, "M", "舁"),
+        (0x2F98C, "M", "舄"),
+        (0x2F98D, "M", "辞"),
+        (0x2F98E, "M", "䑫"),
+        (0x2F98F, "M", "芑"),
+        (0x2F990, "M", "芋"),
+        (0x2F991, "M", "芝"),
+        (0x2F992, "M", "劳"),
+        (0x2F993, "M", "花"),
+        (0x2F994, "M", "芳"),
+        (0x2F995, "M", "芽"),
+        (0x2F996, "M", "苦"),
+        (0x2F997, "M", "𦬼"),
+        (0x2F998, "M", "若"),
+        (0x2F999, "M", "茝"),
+        (0x2F99A, "M", "荣"),
+        (0x2F99B, "M", "莭"),
+        (0x2F99C, "M", "茣"),
+        (0x2F99D, "M", "莽"),
+        (0x2F99E, "M", "菧"),
+        (0x2F99F, "M", "著"),
+        (0x2F9A0, "M", "荓"),
+        (0x2F9A1, "M", "菊"),
+        (0x2F9A2, "M", "菌"),
+        (0x2F9A3, "M", "菜"),
+        (0x2F9A4, "M", "𦰶"),
+        (0x2F9A5, "M", "𦵫"),
+        (0x2F9A6, "M", "𦳕"),
+        (0x2F9A7, "M", "䔫"),
+        (0x2F9A8, "M", "蓱"),
+        (0x2F9A9, "M", "蓳"),
+        (0x2F9AA, "M", "蔖"),
+        (0x2F9AB, "M", "𧏊"),
+        (0x2F9AC, "M", "蕤"),
+        (0x2F9AD, "M", "𦼬"),
+        (0x2F9AE, "M", "䕝"),
+        (0x2F9AF, "M", "䕡"),
+        (0x2F9B0, "M", "𦾱"),
+        (0x2F9B1, "M", "𧃒"),
+        (0x2F9B2, "M", "䕫"),
+        (0x2F9B3, "M", "虐"),
+        (0x2F9B4, "M", "虜"),
+        (0x2F9B5, "M", "虧"),
+        (0x2F9B6, "M", "虩"),
+        (0x2F9B7, "M", "蚩"),
+        (0x2F9B8, "M", "蚈"),
+        (0x2F9B9, "M", "蜎"),
+        (0x2F9BA, "M", "蛢"),
+        (0x2F9BB, "M", "蝹"),
+        (0x2F9BC, "M", "蜨"),
+        (0x2F9BD, "M", "蝫"),
+        (0x2F9BE, "M", "螆"),
+        (0x2F9BF, "X"),
+        (0x2F9C0, "M", "蟡"),
+        (0x2F9C1, "M", "蠁"),
+        (0x2F9C2, "M", "䗹"),
+        (0x2F9C3, "M", "衠"),
+        (0x2F9C4, "M", "衣"),
+        (0x2F9C5, "M", "𧙧"),
+        (0x2F9C6, "M", "裗"),
+        (0x2F9C7, "M", "裞"),
+        (0x2F9C8, "M", "䘵"),
+        (0x2F9C9, "M", "裺"),
+        (0x2F9CA, "M", "㒻"),
+        (0x2F9CB, "M", "𧢮"),
+        (0x2F9CC, "M", "𧥦"),
+        (0x2F9CD, "M", "䚾"),
+        (0x2F9CE, "M", "䛇"),
+        (0x2F9CF, "M", "誠"),
+        (0x2F9D0, "M", "諭"),
+        (0x2F9D1, "M", "變"),
+        (0x2F9D2, "M", "豕"),
+        (0x2F9D3, "M", "𧲨"),
+        (0x2F9D4, "M", "貫"),
+        (0x2F9D5, "M", "賁"),
+        (0x2F9D6, "M", "贛"),
+        (0x2F9D7, "M", "起"),
+    ]
+
+
+def _seg_81() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F9D8, "M", "𧼯"),
+        (0x2F9D9, "M", "𠠄"),
+        (0x2F9DA, "M", "跋"),
+        (0x2F9DB, "M", "趼"),
+        (0x2F9DC, "M", "跰"),
+        (0x2F9DD, "M", "𠣞"),
+        (0x2F9DE, "M", "軔"),
+        (0x2F9DF, "M", "輸"),
+        (0x2F9E0, "M", "𨗒"),
+        (0x2F9E1, "M", "𨗭"),
+        (0x2F9E2, "M", "邔"),
+        (0x2F9E3, "M", "郱"),
+        (0x2F9E4, "M", "鄑"),
+        (0x2F9E5, "M", "𨜮"),
+        (0x2F9E6, "M", "鄛"),
+        (0x2F9E7, "M", "鈸"),
+        (0x2F9E8, "M", "鋗"),
+        (0x2F9E9, "M", "鋘"),
+        (0x2F9EA, "M", "鉼"),
+        (0x2F9EB, "M", "鏹"),
+        (0x2F9EC, "M", "鐕"),
+        (0x2F9ED, "M", "𨯺"),
+        (0x2F9EE, "M", "開"),
+        (0x2F9EF, "M", "䦕"),
+        (0x2F9F0, "M", "閷"),
+        (0x2F9F1, "M", "𨵷"),
+        (0x2F9F2, "M", "䧦"),
+        (0x2F9F3, "M", "雃"),
+        (0x2F9F4, "M", "嶲"),
+        (0x2F9F5, "M", "霣"),
+        (0x2F9F6, "M", "𩅅"),
+        (0x2F9F7, "M", "𩈚"),
+        (0x2F9F8, "M", "䩮"),
+        (0x2F9F9, "M", "䩶"),
+        (0x2F9FA, "M", "韠"),
+        (0x2F9FB, "M", "𩐊"),
+        (0x2F9FC, "M", "䪲"),
+        (0x2F9FD, "M", "𩒖"),
+        (0x2F9FE, "M", "頋"),
+        (0x2FA00, "M", "頩"),
+        (0x2FA01, "M", "𩖶"),
+        (0x2FA02, "M", "飢"),
+        (0x2FA03, "M", "䬳"),
+        (0x2FA04, "M", "餩"),
+        (0x2FA05, "M", "馧"),
+        (0x2FA06, "M", "駂"),
+        (0x2FA07, "M", "駾"),
+        (0x2FA08, "M", "䯎"),
+        (0x2FA09, "M", "𩬰"),
+        (0x2FA0A, "M", "鬒"),
+        (0x2FA0B, "M", "鱀"),
+        (0x2FA0C, "M", "鳽"),
+        (0x2FA0D, "M", "䳎"),
+        (0x2FA0E, "M", "䳭"),
+        (0x2FA0F, "M", "鵧"),
+        (0x2FA10, "M", "𪃎"),
+        (0x2FA11, "M", "䳸"),
+        (0x2FA12, "M", "𪄅"),
+        (0x2FA13, "M", "𪈎"),
+        (0x2FA14, "M", "𪊑"),
+        (0x2FA15, "M", "麻"),
+        (0x2FA16, "M", "䵖"),
+        (0x2FA17, "M", "黹"),
+        (0x2FA18, "M", "黾"),
+        (0x2FA19, "M", "鼅"),
+        (0x2FA1A, "M", "鼏"),
+        (0x2FA1B, "M", "鼖"),
+        (0x2FA1C, "M", "鼻"),
+        (0x2FA1D, "M", "𪘀"),
+        (0x2FA1E, "X"),
+        (0x30000, "V"),
+        (0x3134B, "X"),
+        (0x31350, "V"),
+        (0x323B0, "X"),
+        (0xE0100, "I"),
+        (0xE01F0, "X"),
+    ]
+
+
+uts46data = tuple(
+    _seg_0()
+    + _seg_1()
+    + _seg_2()
+    + _seg_3()
+    + _seg_4()
+    + _seg_5()
+    + _seg_6()
+    + _seg_7()
+    + _seg_8()
+    + _seg_9()
+    + _seg_10()
+    + _seg_11()
+    + _seg_12()
+    + _seg_13()
+    + _seg_14()
+    + _seg_15()
+    + _seg_16()
+    + _seg_17()
+    + _seg_18()
+    + _seg_19()
+    + _seg_20()
+    + _seg_21()
+    + _seg_22()
+    + _seg_23()
+    + _seg_24()
+    + _seg_25()
+    + _seg_26()
+    + _seg_27()
+    + _seg_28()
+    + _seg_29()
+    + _seg_30()
+    + _seg_31()
+    + _seg_32()
+    + _seg_33()
+    + _seg_34()
+    + _seg_35()
+    + _seg_36()
+    + _seg_37()
+    + _seg_38()
+    + _seg_39()
+    + _seg_40()
+    + _seg_41()
+    + _seg_42()
+    + _seg_43()
+    + _seg_44()
+    + _seg_45()
+    + _seg_46()
+    + _seg_47()
+    + _seg_48()
+    + _seg_49()
+    + _seg_50()
+    + _seg_51()
+    + _seg_52()
+    + _seg_53()
+    + _seg_54()
+    + _seg_55()
+    + _seg_56()
+    + _seg_57()
+    + _seg_58()
+    + _seg_59()
+    + _seg_60()
+    + _seg_61()
+    + _seg_62()
+    + _seg_63()
+    + _seg_64()
+    + _seg_65()
+    + _seg_66()
+    + _seg_67()
+    + _seg_68()
+    + _seg_69()
+    + _seg_70()
+    + _seg_71()
+    + _seg_72()
+    + _seg_73()
+    + _seg_74()
+    + _seg_75()
+    + _seg_76()
+    + _seg_77()
+    + _seg_78()
+    + _seg_79()
+    + _seg_80()
+    + _seg_81()
+)  # type: Tuple[Union[Tuple[int, str], Tuple[int, str, str]], ...]
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/METADATA
new file mode 100644
index 00000000..3ea1e01c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/METADATA
@@ -0,0 +1,80 @@
+Metadata-Version: 2.1
+Name: iniconfig
+Version: 2.0.0
+Summary: brain-dead simple config-ini parsing
+Project-URL: Homepage, https://github.com/pytest-dev/iniconfig
+Author-email: Ronny Pfannschmidt <opensource@ronnypfannschmidt.de>, Holger Krekel <holger.krekel@gmail.com>
+License-Expression: MIT
+License-File: LICENSE
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Utilities
+Requires-Python: >=3.7
+Description-Content-Type: text/x-rst
+
+iniconfig: brain-dead simple parsing of ini files
+=======================================================
+
+iniconfig is a small and simple INI-file parser module
+having a unique set of features:
+
+* maintains order of sections and entries
+* supports multi-line values with or without line-continuations
+* supports "#" comments everywhere
+* raises errors with proper line-numbers
+* no bells and whistles like automatic substitutions
+* iniconfig raises an Error if two sections have the same name.
+
+If you encounter issues or have feature wishes please report them to:
+
+    https://github.com/RonnyPfannschmidt/iniconfig/issues
+
+Basic Example
+===================================
+
+If you have an ini file like this:
+
+.. code-block:: ini
+
+    # content of example.ini
+    [section1] # comment
+    name1=value1  # comment
+    name1b=value1,value2  # comment
+
+    [section2]
+    name2=
+        line1
+        line2
+
+then you can do:
+
+.. code-block:: pycon
+
+    >>> import iniconfig
+    >>> ini = iniconfig.IniConfig("example.ini")
+    >>> ini['section1']['name1'] # raises KeyError if not exists
+    'value1'
+    >>> ini.get('section1', 'name1b', [], lambda x: x.split(","))
+    ['value1', 'value2']
+    >>> ini.get('section1', 'notexist', [], lambda x: x.split(","))
+    []
+    >>> [x.name for x in list(ini)]
+    ['section1', 'section2']
+    >>> list(list(ini)[0].items())
+    [('name1', 'value1'), ('name1b', 'value1,value2')]
+    >>> 'section1' in ini
+    True
+    >>> 'inexistendsection' in ini
+    False
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/RECORD
new file mode 100644
index 00000000..6d813d56
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/RECORD
@@ -0,0 +1,14 @@
+iniconfig-2.0.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+iniconfig-2.0.0.dist-info/METADATA,sha256=2KcBd5DEFiZclO-ruP_qzN71qcTL0hNsCw5MCDIPN6I,2599
+iniconfig-2.0.0.dist-info/RECORD,,
+iniconfig-2.0.0.dist-info/WHEEL,sha256=hKi7AIIx6qfnsRbr087vpeJnrVUuDokDHZacPPMW7-Y,87
+iniconfig-2.0.0.dist-info/licenses/LICENSE,sha256=KvaAw570k_uCgwNW0dPfGstaBgM8ui3sehniHKp3qGY,1061
+iniconfig/__init__.py,sha256=ALJSNenAgTD7RNj820NggEQuyaZp2QseTCThGJPavk0,5473
+iniconfig/__pycache__/__init__.cpython-312.pyc,,
+iniconfig/__pycache__/_parse.cpython-312.pyc,,
+iniconfig/__pycache__/_version.cpython-312.pyc,,
+iniconfig/__pycache__/exceptions.cpython-312.pyc,,
+iniconfig/_parse.py,sha256=OWGLbmE8GjxcoMWTvnGbck1RoNsTm5bt5ficIRZqWJ8,2436
+iniconfig/_version.py,sha256=WM8rOXoL5t25aMQJp4qbU2XP09nrDtmDnrAGhHSk0Wk,160
+iniconfig/exceptions.py,sha256=3V2JS5rndwiYUh84PNYS_1zd8H8IB-Rar81ARAA7E9s,501
+iniconfig/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/WHEEL
new file mode 100644
index 00000000..8d5c0ceb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.12.2
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/licenses/LICENSE b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/licenses/LICENSE
new file mode 100644
index 00000000..31ecdfb1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig-2.0.0.dist-info/licenses/LICENSE
@@ -0,0 +1,19 @@
+
+  Permission is hereby granted, free of charge, to any person obtaining a copy
+  of this software and associated documentation files (the "Software"), to deal
+  in the Software without restriction, including without limitation the rights
+  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+  copies of the Software, and to permit persons to whom the Software is
+  furnished to do so, subject to the following conditions:
+     
+  The above copyright notice and this permission notice shall be included in all
+  copies or substantial portions of the Software.
+ 
+  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+  SOFTWARE.
+
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/__init__.py b/backend/venv/lib/python3.12/site-packages/iniconfig/__init__.py
new file mode 100644
index 00000000..c18a8e4b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig/__init__.py
@@ -0,0 +1,216 @@
+""" brain-dead simple parser for ini-style files.
+(C) Ronny Pfannschmidt, Holger Krekel -- MIT licensed
+"""
+from __future__ import annotations
+from typing import (
+    Callable,
+    Iterator,
+    Mapping,
+    Optional,
+    Tuple,
+    TypeVar,
+    Union,
+    TYPE_CHECKING,
+    NoReturn,
+    NamedTuple,
+    overload,
+    cast,
+)
+
+import os
+
+if TYPE_CHECKING:
+    from typing_extensions import Final
+
+__all__ = ["IniConfig", "ParseError", "COMMENTCHARS", "iscommentline"]
+
+from .exceptions import ParseError
+from . import _parse
+from ._parse import COMMENTCHARS, iscommentline
+
+_D = TypeVar("_D")
+_T = TypeVar("_T")
+
+
+class SectionWrapper:
+    config: Final[IniConfig]
+    name: Final[str]
+
+    def __init__(self, config: IniConfig, name: str) -> None:
+        self.config = config
+        self.name = name
+
+    def lineof(self, name: str) -> int | None:
+        return self.config.lineof(self.name, name)
+
+    @overload
+    def get(self, key: str) -> str | None:
+        ...
+
+    @overload
+    def get(
+        self,
+        key: str,
+        convert: Callable[[str], _T],
+    ) -> _T | None:
+        ...
+
+    @overload
+    def get(
+        self,
+        key: str,
+        default: None,
+        convert: Callable[[str], _T],
+    ) -> _T | None:
+        ...
+
+    @overload
+    def get(self, key: str, default: _D, convert: None = None) -> str | _D:
+        ...
+
+    @overload
+    def get(
+        self,
+        key: str,
+        default: _D,
+        convert: Callable[[str], _T],
+    ) -> _T | _D:
+        ...
+
+    # TODO: investigate possible mypy bug wrt matching the passed over data
+    def get(  # type: ignore [misc]
+        self,
+        key: str,
+        default: _D | None = None,
+        convert: Callable[[str], _T] | None = None,
+    ) -> _D | _T | str | None:
+        return self.config.get(self.name, key, convert=convert, default=default)
+
+    def __getitem__(self, key: str) -> str:
+        return self.config.sections[self.name][key]
+
+    def __iter__(self) -> Iterator[str]:
+        section: Mapping[str, str] = self.config.sections.get(self.name, {})
+
+        def lineof(key: str) -> int:
+            return self.config.lineof(self.name, key)  # type: ignore[return-value]
+
+        yield from sorted(section, key=lineof)
+
+    def items(self) -> Iterator[tuple[str, str]]:
+        for name in self:
+            yield name, self[name]
+
+
+class IniConfig:
+    path: Final[str]
+    sections: Final[Mapping[str, Mapping[str, str]]]
+
+    def __init__(
+        self,
+        path: str | os.PathLike[str],
+        data: str | None = None,
+        encoding: str = "utf-8",
+    ) -> None:
+        self.path = os.fspath(path)
+        if data is None:
+            with open(self.path, encoding=encoding) as fp:
+                data = fp.read()
+
+        tokens = _parse.parse_lines(self.path, data.splitlines(True))
+
+        self._sources = {}
+        sections_data: dict[str, dict[str, str]]
+        self.sections = sections_data = {}
+
+        for lineno, section, name, value in tokens:
+            if section is None:
+                raise ParseError(self.path, lineno, "no section header defined")
+            self._sources[section, name] = lineno
+            if name is None:
+                if section in self.sections:
+                    raise ParseError(
+                        self.path, lineno, f"duplicate section {section!r}"
+                    )
+                sections_data[section] = {}
+            else:
+                if name in self.sections[section]:
+                    raise ParseError(self.path, lineno, f"duplicate name {name!r}")
+                assert value is not None
+                sections_data[section][name] = value
+
+    def lineof(self, section: str, name: str | None = None) -> int | None:
+        lineno = self._sources.get((section, name))
+        return None if lineno is None else lineno + 1
+
+    @overload
+    def get(
+        self,
+        section: str,
+        name: str,
+    ) -> str | None:
+        ...
+
+    @overload
+    def get(
+        self,
+        section: str,
+        name: str,
+        convert: Callable[[str], _T],
+    ) -> _T | None:
+        ...
+
+    @overload
+    def get(
+        self,
+        section: str,
+        name: str,
+        default: None,
+        convert: Callable[[str], _T],
+    ) -> _T | None:
+        ...
+
+    @overload
+    def get(
+        self, section: str, name: str, default: _D, convert: None = None
+    ) -> str | _D:
+        ...
+
+    @overload
+    def get(
+        self,
+        section: str,
+        name: str,
+        default: _D,
+        convert: Callable[[str], _T],
+    ) -> _T | _D:
+        ...
+
+    def get(  # type: ignore
+        self,
+        section: str,
+        name: str,
+        default: _D | None = None,
+        convert: Callable[[str], _T] | None = None,
+    ) -> _D | _T | str | None:
+        try:
+            value: str = self.sections[section][name]
+        except KeyError:
+            return default
+        else:
+            if convert is not None:
+                return convert(value)
+            else:
+                return value
+
+    def __getitem__(self, name: str) -> SectionWrapper:
+        if name not in self.sections:
+            raise KeyError(name)
+        return SectionWrapper(self, name)
+
+    def __iter__(self) -> Iterator[SectionWrapper]:
+        for name in sorted(self.sections, key=self.lineof):  # type: ignore
+            yield SectionWrapper(self, name)
+
+    def __contains__(self, arg: str) -> bool:
+        return arg in self.sections
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..92be3e22
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_parse.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_parse.cpython-312.pyc
new file mode 100644
index 00000000..3bd80943
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_parse.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..603168ac
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..0f2ce2b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/iniconfig/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/_parse.py b/backend/venv/lib/python3.12/site-packages/iniconfig/_parse.py
new file mode 100644
index 00000000..2d03437b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig/_parse.py
@@ -0,0 +1,82 @@
+from __future__ import annotations
+from .exceptions import ParseError
+
+from typing import NamedTuple
+
+
+COMMENTCHARS = "#;"
+
+
+class _ParsedLine(NamedTuple):
+    lineno: int
+    section: str | None
+    name: str | None
+    value: str | None
+
+
+def parse_lines(path: str, line_iter: list[str]) -> list[_ParsedLine]:
+    result: list[_ParsedLine] = []
+    section = None
+    for lineno, line in enumerate(line_iter):
+        name, data = _parseline(path, line, lineno)
+        # new value
+        if name is not None and data is not None:
+            result.append(_ParsedLine(lineno, section, name, data))
+        # new section
+        elif name is not None and data is None:
+            if not name:
+                raise ParseError(path, lineno, "empty section name")
+            section = name
+            result.append(_ParsedLine(lineno, section, None, None))
+        # continuation
+        elif name is None and data is not None:
+            if not result:
+                raise ParseError(path, lineno, "unexpected value continuation")
+            last = result.pop()
+            if last.name is None:
+                raise ParseError(path, lineno, "unexpected value continuation")
+
+            if last.value:
+                last = last._replace(value=f"{last.value}\n{data}")
+            else:
+                last = last._replace(value=data)
+            result.append(last)
+    return result
+
+
+def _parseline(path: str, line: str, lineno: int) -> tuple[str | None, str | None]:
+    # blank lines
+    if iscommentline(line):
+        line = ""
+    else:
+        line = line.rstrip()
+    if not line:
+        return None, None
+    # section
+    if line[0] == "[":
+        realline = line
+        for c in COMMENTCHARS:
+            line = line.split(c)[0].rstrip()
+        if line[-1] == "]":
+            return line[1:-1], None
+        return None, realline.strip()
+    # value
+    elif not line[0].isspace():
+        try:
+            name, value = line.split("=", 1)
+            if ":" in name:
+                raise ValueError()
+        except ValueError:
+            try:
+                name, value = line.split(":", 1)
+            except ValueError:
+                raise ParseError(path, lineno, "unexpected line: %r" % line)
+        return name.strip(), value.strip()
+    # continuation
+    else:
+        return None, line.strip()
+
+
+def iscommentline(line: str) -> bool:
+    c = line.lstrip()[:1]
+    return c in COMMENTCHARS
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/_version.py b/backend/venv/lib/python3.12/site-packages/iniconfig/_version.py
new file mode 100644
index 00000000..dd1883d7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig/_version.py
@@ -0,0 +1,4 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+__version__ = version = '2.0.0'
+__version_tuple__ = version_tuple = (2, 0, 0)
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/exceptions.py b/backend/venv/lib/python3.12/site-packages/iniconfig/exceptions.py
new file mode 100644
index 00000000..bc898e68
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/iniconfig/exceptions.py
@@ -0,0 +1,20 @@
+from __future__ import annotations
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing_extensions import Final
+
+
+class ParseError(Exception):
+    path: Final[str]
+    lineno: Final[int]
+    msg: Final[str]
+
+    def __init__(self, path: str, lineno: int, msg: str) -> None:
+        super().__init__(path, lineno, msg)
+        self.path = path
+        self.lineno = lineno
+        self.msg = msg
+
+    def __str__(self) -> str:
+        return f"{self.path}:{self.lineno + 1}: {self.msg}"
diff --git a/backend/venv/lib/python3.12/site-packages/iniconfig/py.typed b/backend/venv/lib/python3.12/site-packages/iniconfig/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/LICENSE.txt
new file mode 100644
index 00000000..7b190ca6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2011 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/METADATA
new file mode 100644
index 00000000..ddf54648
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/METADATA
@@ -0,0 +1,60 @@
+Metadata-Version: 2.1
+Name: itsdangerous
+Version: 2.2.0
+Summary: Safely pass data to untrusted environments and back.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Typing :: Typed
+Project-URL: Changes, https://itsdangerous.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://itsdangerous.palletsprojects.com/
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Source, https://github.com/pallets/itsdangerous/
+
+# ItsDangerous
+
+... so better sign this
+
+Various helpers to pass data to untrusted environments and to get it
+back safe and sound. Data is cryptographically signed to ensure that a
+token has not been tampered with.
+
+It's possible to customize how data is serialized. Data is compressed as
+needed. A timestamp can be added and verified automatically while
+loading a token.
+
+
+## A Simple Example
+
+Here's how you could generate a token for transmitting a user's id and
+name between web requests.
+
+```python
+from itsdangerous import URLSafeSerializer
+auth_s = URLSafeSerializer("secret key", "auth")
+token = auth_s.dumps({"id": 5, "name": "itsdangerous"})
+
+print(token)
+# eyJpZCI6NSwibmFtZSI6Iml0c2Rhbmdlcm91cyJ9.6YP6T0BaO67XP--9UzTrmurXSmg
+
+data = auth_s.loads(token)
+print(data["name"])
+# itsdangerous
+```
+
+
+## Donate
+
+The Pallets organization develops and supports ItsDangerous and other
+popular packages. In order to grow the community of contributors and
+users, and allow the maintainers to devote more time to the projects,
+[please donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
+
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/RECORD
new file mode 100644
index 00000000..5f1a9de9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/RECORD
@@ -0,0 +1,23 @@
+itsdangerous-2.2.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+itsdangerous-2.2.0.dist-info/LICENSE.txt,sha256=Y68JiRtr6K0aQlLtQ68PTvun_JSOIoNnvtfzxa4LCdc,1475
+itsdangerous-2.2.0.dist-info/METADATA,sha256=0rk0-1ZwihuU5DnwJVwPWoEI4yWOyCexih3JyZHblhE,1924
+itsdangerous-2.2.0.dist-info/RECORD,,
+itsdangerous-2.2.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+itsdangerous-2.2.0.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+itsdangerous/__init__.py,sha256=4SK75sCe29xbRgQE1ZQtMHnKUuZYAf3bSpZOrff1IAY,1427
+itsdangerous/__pycache__/__init__.cpython-312.pyc,,
+itsdangerous/__pycache__/_json.cpython-312.pyc,,
+itsdangerous/__pycache__/encoding.cpython-312.pyc,,
+itsdangerous/__pycache__/exc.cpython-312.pyc,,
+itsdangerous/__pycache__/serializer.cpython-312.pyc,,
+itsdangerous/__pycache__/signer.cpython-312.pyc,,
+itsdangerous/__pycache__/timed.cpython-312.pyc,,
+itsdangerous/__pycache__/url_safe.cpython-312.pyc,,
+itsdangerous/_json.py,sha256=wPQGmge2yZ9328EHKF6gadGeyGYCJQKxtU-iLKE6UnA,473
+itsdangerous/encoding.py,sha256=wwTz5q_3zLcaAdunk6_vSoStwGqYWe307Zl_U87aRFM,1409
+itsdangerous/exc.py,sha256=Rr3exo0MRFEcPZltwecyK16VV1bE2K9_F1-d-ljcUn4,3201
+itsdangerous/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+itsdangerous/serializer.py,sha256=PmdwADLqkSyQLZ0jOKAgDsAW4k_H0TlA71Ei3z0C5aI,15601
+itsdangerous/signer.py,sha256=YO0CV7NBvHA6j549REHJFUjUojw2pHqwcUpQnU7yNYQ,9647
+itsdangerous/timed.py,sha256=6RvDMqNumGMxf0-HlpaZdN9PUQQmRvrQGplKhxuivUs,8083
+itsdangerous/url_safe.py,sha256=az4e5fXi_vs-YbWj8YZwn4wiVKfeD--GEKRT5Ueu4P4,2505
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous-2.2.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__init__.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/__init__.py
new file mode 100644
index 00000000..ea55256e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/__init__.py
@@ -0,0 +1,38 @@
+from __future__ import annotations
+
+import typing as t
+
+from .encoding import base64_decode as base64_decode
+from .encoding import base64_encode as base64_encode
+from .encoding import want_bytes as want_bytes
+from .exc import BadData as BadData
+from .exc import BadHeader as BadHeader
+from .exc import BadPayload as BadPayload
+from .exc import BadSignature as BadSignature
+from .exc import BadTimeSignature as BadTimeSignature
+from .exc import SignatureExpired as SignatureExpired
+from .serializer import Serializer as Serializer
+from .signer import HMACAlgorithm as HMACAlgorithm
+from .signer import NoneAlgorithm as NoneAlgorithm
+from .signer import Signer as Signer
+from .timed import TimedSerializer as TimedSerializer
+from .timed import TimestampSigner as TimestampSigner
+from .url_safe import URLSafeSerializer as URLSafeSerializer
+from .url_safe import URLSafeTimedSerializer as URLSafeTimedSerializer
+
+
+def __getattr__(name: str) -> t.Any:
+    if name == "__version__":
+        import importlib.metadata
+        import warnings
+
+        warnings.warn(
+            "The '__version__' attribute is deprecated and will be removed in"
+            " ItsDangerous 2.3. Use feature detection or"
+            " 'importlib.metadata.version(\"itsdangerous\")' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return importlib.metadata.version("itsdangerous")
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ef5fd1a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/_json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/_json.cpython-312.pyc
new file mode 100644
index 00000000..573994b4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/_json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/encoding.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/encoding.cpython-312.pyc
new file mode 100644
index 00000000..783a5319
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/encoding.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/exc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/exc.cpython-312.pyc
new file mode 100644
index 00000000..dece38fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/exc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/serializer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/serializer.cpython-312.pyc
new file mode 100644
index 00000000..390c6b67
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/serializer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/signer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/signer.cpython-312.pyc
new file mode 100644
index 00000000..783a8875
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/signer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/timed.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/timed.cpython-312.pyc
new file mode 100644
index 00000000..2d55278b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/timed.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/url_safe.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/url_safe.cpython-312.pyc
new file mode 100644
index 00000000..f91f1265
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/itsdangerous/__pycache__/url_safe.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/_json.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/_json.py
new file mode 100644
index 00000000..fc23feaa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/_json.py
@@ -0,0 +1,18 @@
+from __future__ import annotations
+
+import json as _json
+import typing as t
+
+
+class _CompactJSON:
+    """Wrapper around json module that strips whitespace."""
+
+    @staticmethod
+    def loads(payload: str | bytes) -> t.Any:
+        return _json.loads(payload)
+
+    @staticmethod
+    def dumps(obj: t.Any, **kwargs: t.Any) -> str:
+        kwargs.setdefault("ensure_ascii", False)
+        kwargs.setdefault("separators", (",", ":"))
+        return _json.dumps(obj, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/encoding.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/encoding.py
new file mode 100644
index 00000000..f5ca80f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/encoding.py
@@ -0,0 +1,54 @@
+from __future__ import annotations
+
+import base64
+import string
+import struct
+import typing as t
+
+from .exc import BadData
+
+
+def want_bytes(
+    s: str | bytes, encoding: str = "utf-8", errors: str = "strict"
+) -> bytes:
+    if isinstance(s, str):
+        s = s.encode(encoding, errors)
+
+    return s
+
+
+def base64_encode(string: str | bytes) -> bytes:
+    """Base64 encode a string of bytes or text. The resulting bytes are
+    safe to use in URLs.
+    """
+    string = want_bytes(string)
+    return base64.urlsafe_b64encode(string).rstrip(b"=")
+
+
+def base64_decode(string: str | bytes) -> bytes:
+    """Base64 decode a URL-safe string of bytes or text. The result is
+    bytes.
+    """
+    string = want_bytes(string, encoding="ascii", errors="ignore")
+    string += b"=" * (-len(string) % 4)
+
+    try:
+        return base64.urlsafe_b64decode(string)
+    except (TypeError, ValueError) as e:
+        raise BadData("Invalid base64-encoded data") from e
+
+
+# The alphabet used by base64.urlsafe_*
+_base64_alphabet = f"{string.ascii_letters}{string.digits}-_=".encode("ascii")
+
+_int64_struct = struct.Struct(">Q")
+_int_to_bytes = _int64_struct.pack
+_bytes_to_int = t.cast("t.Callable[[bytes], tuple[int]]", _int64_struct.unpack)
+
+
+def int_to_bytes(num: int) -> bytes:
+    return _int_to_bytes(num).lstrip(b"\x00")
+
+
+def bytes_to_int(bytestr: bytes) -> int:
+    return _bytes_to_int(bytestr.rjust(8, b"\x00"))[0]
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/exc.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/exc.py
new file mode 100644
index 00000000..a75adcd5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/exc.py
@@ -0,0 +1,106 @@
+from __future__ import annotations
+
+import typing as t
+from datetime import datetime
+
+
+class BadData(Exception):
+    """Raised if bad data of any sort was encountered. This is the base
+    for all exceptions that ItsDangerous defines.
+
+    .. versionadded:: 0.15
+    """
+
+    def __init__(self, message: str):
+        super().__init__(message)
+        self.message = message
+
+    def __str__(self) -> str:
+        return self.message
+
+
+class BadSignature(BadData):
+    """Raised if a signature does not match."""
+
+    def __init__(self, message: str, payload: t.Any | None = None):
+        super().__init__(message)
+
+        #: The payload that failed the signature test. In some
+        #: situations you might still want to inspect this, even if
+        #: you know it was tampered with.
+        #:
+        #: .. versionadded:: 0.14
+        self.payload: t.Any | None = payload
+
+
+class BadTimeSignature(BadSignature):
+    """Raised if a time-based signature is invalid. This is a subclass
+    of :class:`BadSignature`.
+    """
+
+    def __init__(
+        self,
+        message: str,
+        payload: t.Any | None = None,
+        date_signed: datetime | None = None,
+    ):
+        super().__init__(message, payload)
+
+        #: If the signature expired this exposes the date of when the
+        #: signature was created. This can be helpful in order to
+        #: tell the user how long a link has been gone stale.
+        #:
+        #: .. versionchanged:: 2.0
+        #:     The datetime value is timezone-aware rather than naive.
+        #:
+        #: .. versionadded:: 0.14
+        self.date_signed = date_signed
+
+
+class SignatureExpired(BadTimeSignature):
+    """Raised if a signature timestamp is older than ``max_age``. This
+    is a subclass of :exc:`BadTimeSignature`.
+    """
+
+
+class BadHeader(BadSignature):
+    """Raised if a signed header is invalid in some form. This only
+    happens for serializers that have a header that goes with the
+    signature.
+
+    .. versionadded:: 0.24
+    """
+
+    def __init__(
+        self,
+        message: str,
+        payload: t.Any | None = None,
+        header: t.Any | None = None,
+        original_error: Exception | None = None,
+    ):
+        super().__init__(message, payload)
+
+        #: If the header is actually available but just malformed it
+        #: might be stored here.
+        self.header: t.Any | None = header
+
+        #: If available, the error that indicates why the payload was
+        #: not valid. This might be ``None``.
+        self.original_error: Exception | None = original_error
+
+
+class BadPayload(BadData):
+    """Raised if a payload is invalid. This could happen if the payload
+    is loaded despite an invalid signature, or if there is a mismatch
+    between the serializer and deserializer. The original exception
+    that occurred during loading is stored on as :attr:`original_error`.
+
+    .. versionadded:: 0.15
+    """
+
+    def __init__(self, message: str, original_error: Exception | None = None):
+        super().__init__(message)
+
+        #: If available, the error that indicates why the payload was
+        #: not valid. This might be ``None``.
+        self.original_error: Exception | None = original_error
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/py.typed b/backend/venv/lib/python3.12/site-packages/itsdangerous/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/serializer.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/serializer.py
new file mode 100644
index 00000000..5ddf3871
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/serializer.py
@@ -0,0 +1,406 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import json
+import typing as t
+
+from .encoding import want_bytes
+from .exc import BadPayload
+from .exc import BadSignature
+from .signer import _make_keys_list
+from .signer import Signer
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    # This should be either be str or bytes. To avoid having to specify the
+    # bound type, it falls back to a union if structural matching fails.
+    _TSerialized = te.TypeVar(
+        "_TSerialized", bound=t.Union[str, bytes], default=t.Union[str, bytes]
+    )
+else:
+    # Still available at runtime on Python < 3.13, but without the default.
+    _TSerialized = t.TypeVar("_TSerialized", bound=t.Union[str, bytes])
+
+
+class _PDataSerializer(t.Protocol[_TSerialized]):
+    def loads(self, payload: _TSerialized, /) -> t.Any: ...
+    # A signature with additional arguments is not handled correctly by type
+    # checkers right now, so an overload is used below for serializers that
+    # don't match this strict protocol.
+    def dumps(self, obj: t.Any, /) -> _TSerialized: ...
+
+
+# Use TypeIs once it's available in typing_extensions or 3.13.
+def is_text_serializer(
+    serializer: _PDataSerializer[t.Any],
+) -> te.TypeGuard[_PDataSerializer[str]]:
+    """Checks whether a serializer generates text or binary."""
+    return isinstance(serializer.dumps({}), str)
+
+
+class Serializer(t.Generic[_TSerialized]):
+    """A serializer wraps a :class:`~itsdangerous.signer.Signer` to
+    enable serializing and securely signing data other than bytes. It
+    can unsign to verify that the data hasn't been changed.
+
+    The serializer provides :meth:`dumps` and :meth:`loads`, similar to
+    :mod:`json`, and by default uses :mod:`json` internally to serialize
+    the data to bytes.
+
+    The secret key should be a random string of ``bytes`` and should not
+    be saved to code or version control. Different salts should be used
+    to distinguish signing in different contexts. See :doc:`/concepts`
+    for information about the security of the secret key and salt.
+
+    :param secret_key: The secret key to sign and verify with. Can be a
+        list of keys, oldest to newest, to support key rotation.
+    :param salt: Extra key to combine with ``secret_key`` to distinguish
+        signatures in different contexts.
+    :param serializer: An object that provides ``dumps`` and ``loads``
+        methods for serializing data to a string. Defaults to
+        :attr:`default_serializer`, which defaults to :mod:`json`.
+    :param serializer_kwargs: Keyword arguments to pass when calling
+        ``serializer.dumps``.
+    :param signer: A ``Signer`` class to instantiate when signing data.
+        Defaults to :attr:`default_signer`, which defaults to
+        :class:`~itsdangerous.signer.Signer`.
+    :param signer_kwargs: Keyword arguments to pass when instantiating
+        the ``Signer`` class.
+    :param fallback_signers: List of signer parameters to try when
+        unsigning with the default signer fails. Each item can be a dict
+        of ``signer_kwargs``, a ``Signer`` class, or a tuple of
+        ``(signer, signer_kwargs)``. Defaults to
+        :attr:`default_fallback_signers`.
+
+    .. versionchanged:: 2.0
+        Added support for key rotation by passing a list to
+        ``secret_key``.
+
+    .. versionchanged:: 2.0
+        Removed the default SHA-512 fallback signer from
+        ``default_fallback_signers``.
+
+    .. versionchanged:: 1.1
+        Added support for ``fallback_signers`` and configured a default
+        SHA-512 fallback. This fallback is for users who used the yanked
+        1.0.0 release which defaulted to SHA-512.
+
+    .. versionchanged:: 0.14
+        The ``signer`` and ``signer_kwargs`` parameters were added to
+        the constructor.
+    """
+
+    #: The default serialization module to use to serialize data to a
+    #: string internally. The default is :mod:`json`, but can be changed
+    #: to any object that provides ``dumps`` and ``loads`` methods.
+    default_serializer: _PDataSerializer[t.Any] = json
+
+    #: The default ``Signer`` class to instantiate when signing data.
+    #: The default is :class:`itsdangerous.signer.Signer`.
+    default_signer: type[Signer] = Signer
+
+    #: The default fallback signers to try when unsigning fails.
+    default_fallback_signers: list[
+        dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+    ] = []
+
+    # Serializer[str] if no data serializer is provided, or if it returns str.
+    @t.overload
+    def __init__(
+        self: Serializer[str],
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None = b"itsdangerous",
+        serializer: None | _PDataSerializer[str] = None,
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ): ...
+
+    # Serializer[bytes] with a bytes data serializer positional argument.
+    @t.overload
+    def __init__(
+        self: Serializer[bytes],
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None,
+        serializer: _PDataSerializer[bytes],
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ): ...
+
+    # Serializer[bytes] with a bytes data serializer keyword argument.
+    @t.overload
+    def __init__(
+        self: Serializer[bytes],
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None = b"itsdangerous",
+        *,
+        serializer: _PDataSerializer[bytes],
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ): ...
+
+    # Fall back with a positional argument. If the strict signature of
+    # _PDataSerializer doesn't match, fall back to a union, requiring the user
+    # to specify the type.
+    @t.overload
+    def __init__(
+        self,
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None,
+        serializer: t.Any,
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ): ...
+
+    # Fall back with a keyword argument.
+    @t.overload
+    def __init__(
+        self,
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None = b"itsdangerous",
+        *,
+        serializer: t.Any,
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ): ...
+
+    def __init__(
+        self,
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None = b"itsdangerous",
+        serializer: t.Any | None = None,
+        serializer_kwargs: dict[str, t.Any] | None = None,
+        signer: type[Signer] | None = None,
+        signer_kwargs: dict[str, t.Any] | None = None,
+        fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ]
+        | None = None,
+    ):
+        #: The list of secret keys to try for verifying signatures, from
+        #: oldest to newest. The newest (last) key is used for signing.
+        #:
+        #: This allows a key rotation system to keep a list of allowed
+        #: keys and remove expired ones.
+        self.secret_keys: list[bytes] = _make_keys_list(secret_key)
+
+        if salt is not None:
+            salt = want_bytes(salt)
+            # if salt is None then the signer's default is used
+
+        self.salt = salt
+
+        if serializer is None:
+            serializer = self.default_serializer
+
+        self.serializer: _PDataSerializer[_TSerialized] = serializer
+        self.is_text_serializer: bool = is_text_serializer(serializer)
+
+        if signer is None:
+            signer = self.default_signer
+
+        self.signer: type[Signer] = signer
+        self.signer_kwargs: dict[str, t.Any] = signer_kwargs or {}
+
+        if fallback_signers is None:
+            fallback_signers = list(self.default_fallback_signers)
+
+        self.fallback_signers: list[
+            dict[str, t.Any] | tuple[type[Signer], dict[str, t.Any]] | type[Signer]
+        ] = fallback_signers
+        self.serializer_kwargs: dict[str, t.Any] = serializer_kwargs or {}
+
+    @property
+    def secret_key(self) -> bytes:
+        """The newest (last) entry in the :attr:`secret_keys` list. This
+        is for compatibility from before key rotation support was added.
+        """
+        return self.secret_keys[-1]
+
+    def load_payload(
+        self, payload: bytes, serializer: _PDataSerializer[t.Any] | None = None
+    ) -> t.Any:
+        """Loads the encoded object. This function raises
+        :class:`.BadPayload` if the payload is not valid. The
+        ``serializer`` parameter can be used to override the serializer
+        stored on the class. The encoded ``payload`` should always be
+        bytes.
+        """
+        if serializer is None:
+            use_serializer = self.serializer
+            is_text = self.is_text_serializer
+        else:
+            use_serializer = serializer
+            is_text = is_text_serializer(serializer)
+
+        try:
+            if is_text:
+                return use_serializer.loads(payload.decode("utf-8"))  # type: ignore[arg-type]
+
+            return use_serializer.loads(payload)  # type: ignore[arg-type]
+        except Exception as e:
+            raise BadPayload(
+                "Could not load the payload because an exception"
+                " occurred on unserializing the data.",
+                original_error=e,
+            ) from e
+
+    def dump_payload(self, obj: t.Any) -> bytes:
+        """Dumps the encoded object. The return value is always bytes.
+        If the internal serializer returns text, the value will be
+        encoded as UTF-8.
+        """
+        return want_bytes(self.serializer.dumps(obj, **self.serializer_kwargs))
+
+    def make_signer(self, salt: str | bytes | None = None) -> Signer:
+        """Creates a new instance of the signer to be used. The default
+        implementation uses the :class:`.Signer` base class.
+        """
+        if salt is None:
+            salt = self.salt
+
+        return self.signer(self.secret_keys, salt=salt, **self.signer_kwargs)
+
+    def iter_unsigners(self, salt: str | bytes | None = None) -> cabc.Iterator[Signer]:
+        """Iterates over all signers to be tried for unsigning. Starts
+        with the configured signer, then constructs each signer
+        specified in ``fallback_signers``.
+        """
+        if salt is None:
+            salt = self.salt
+
+        yield self.make_signer(salt)
+
+        for fallback in self.fallback_signers:
+            if isinstance(fallback, dict):
+                kwargs = fallback
+                fallback = self.signer
+            elif isinstance(fallback, tuple):
+                fallback, kwargs = fallback
+            else:
+                kwargs = self.signer_kwargs
+
+            for secret_key in self.secret_keys:
+                yield fallback(secret_key, salt=salt, **kwargs)
+
+    def dumps(self, obj: t.Any, salt: str | bytes | None = None) -> _TSerialized:
+        """Returns a signed string serialized with the internal
+        serializer. The return value can be either a byte or unicode
+        string depending on the format of the internal serializer.
+        """
+        payload = want_bytes(self.dump_payload(obj))
+        rv = self.make_signer(salt).sign(payload)
+
+        if self.is_text_serializer:
+            return rv.decode("utf-8")  # type: ignore[return-value]
+
+        return rv  # type: ignore[return-value]
+
+    def dump(self, obj: t.Any, f: t.IO[t.Any], salt: str | bytes | None = None) -> None:
+        """Like :meth:`dumps` but dumps into a file. The file handle has
+        to be compatible with what the internal serializer expects.
+        """
+        f.write(self.dumps(obj, salt))
+
+    def loads(
+        self, s: str | bytes, salt: str | bytes | None = None, **kwargs: t.Any
+    ) -> t.Any:
+        """Reverse of :meth:`dumps`. Raises :exc:`.BadSignature` if the
+        signature validation fails.
+        """
+        s = want_bytes(s)
+        last_exception = None
+
+        for signer in self.iter_unsigners(salt):
+            try:
+                return self.load_payload(signer.unsign(s))
+            except BadSignature as err:
+                last_exception = err
+
+        raise t.cast(BadSignature, last_exception)
+
+    def load(self, f: t.IO[t.Any], salt: str | bytes | None = None) -> t.Any:
+        """Like :meth:`loads` but loads from a file."""
+        return self.loads(f.read(), salt)
+
+    def loads_unsafe(
+        self, s: str | bytes, salt: str | bytes | None = None
+    ) -> tuple[bool, t.Any]:
+        """Like :meth:`loads` but without verifying the signature. This
+        is potentially very dangerous to use depending on how your
+        serializer works. The return value is ``(signature_valid,
+        payload)`` instead of just the payload. The first item will be a
+        boolean that indicates if the signature is valid. This function
+        never fails.
+
+        Use it for debugging only and if you know that your serializer
+        module is not exploitable (for example, do not use it with a
+        pickle serializer).
+
+        .. versionadded:: 0.15
+        """
+        return self._loads_unsafe_impl(s, salt)
+
+    def _loads_unsafe_impl(
+        self,
+        s: str | bytes,
+        salt: str | bytes | None,
+        load_kwargs: dict[str, t.Any] | None = None,
+        load_payload_kwargs: dict[str, t.Any] | None = None,
+    ) -> tuple[bool, t.Any]:
+        """Low level helper function to implement :meth:`loads_unsafe`
+        in serializer subclasses.
+        """
+        if load_kwargs is None:
+            load_kwargs = {}
+
+        try:
+            return True, self.loads(s, salt=salt, **load_kwargs)
+        except BadSignature as e:
+            if e.payload is None:
+                return False, None
+
+            if load_payload_kwargs is None:
+                load_payload_kwargs = {}
+
+            try:
+                return (
+                    False,
+                    self.load_payload(e.payload, **load_payload_kwargs),
+                )
+            except BadPayload:
+                return False, None
+
+    def load_unsafe(
+        self, f: t.IO[t.Any], salt: str | bytes | None = None
+    ) -> tuple[bool, t.Any]:
+        """Like :meth:`loads_unsafe` but loads from a file.
+
+        .. versionadded:: 0.15
+        """
+        return self.loads_unsafe(f.read(), salt=salt)
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/signer.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/signer.py
new file mode 100644
index 00000000..e324dc03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/signer.py
@@ -0,0 +1,266 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import hashlib
+import hmac
+import typing as t
+
+from .encoding import _base64_alphabet
+from .encoding import base64_decode
+from .encoding import base64_encode
+from .encoding import want_bytes
+from .exc import BadSignature
+
+
+class SigningAlgorithm:
+    """Subclasses must implement :meth:`get_signature` to provide
+    signature generation functionality.
+    """
+
+    def get_signature(self, key: bytes, value: bytes) -> bytes:
+        """Returns the signature for the given key and value."""
+        raise NotImplementedError()
+
+    def verify_signature(self, key: bytes, value: bytes, sig: bytes) -> bool:
+        """Verifies the given signature matches the expected
+        signature.
+        """
+        return hmac.compare_digest(sig, self.get_signature(key, value))
+
+
+class NoneAlgorithm(SigningAlgorithm):
+    """Provides an algorithm that does not perform any signing and
+    returns an empty signature.
+    """
+
+    def get_signature(self, key: bytes, value: bytes) -> bytes:
+        return b""
+
+
+def _lazy_sha1(string: bytes = b"") -> t.Any:
+    """Don't access ``hashlib.sha1`` until runtime. FIPS builds may not include
+    SHA-1, in which case the import and use as a default would fail before the
+    developer can configure something else.
+    """
+    return hashlib.sha1(string)
+
+
+class HMACAlgorithm(SigningAlgorithm):
+    """Provides signature generation using HMACs."""
+
+    #: The digest method to use with the MAC algorithm. This defaults to
+    #: SHA1, but can be changed to any other function in the hashlib
+    #: module.
+    default_digest_method: t.Any = staticmethod(_lazy_sha1)
+
+    def __init__(self, digest_method: t.Any = None):
+        if digest_method is None:
+            digest_method = self.default_digest_method
+
+        self.digest_method: t.Any = digest_method
+
+    def get_signature(self, key: bytes, value: bytes) -> bytes:
+        mac = hmac.new(key, msg=value, digestmod=self.digest_method)
+        return mac.digest()
+
+
+def _make_keys_list(
+    secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+) -> list[bytes]:
+    if isinstance(secret_key, (str, bytes)):
+        return [want_bytes(secret_key)]
+
+    return [want_bytes(s) for s in secret_key]  # pyright: ignore
+
+
+class Signer:
+    """A signer securely signs bytes, then unsigns them to verify that
+    the value hasn't been changed.
+
+    The secret key should be a random string of ``bytes`` and should not
+    be saved to code or version control. Different salts should be used
+    to distinguish signing in different contexts. See :doc:`/concepts`
+    for information about the security of the secret key and salt.
+
+    :param secret_key: The secret key to sign and verify with. Can be a
+        list of keys, oldest to newest, to support key rotation.
+    :param salt: Extra key to combine with ``secret_key`` to distinguish
+        signatures in different contexts.
+    :param sep: Separator between the signature and value.
+    :param key_derivation: How to derive the signing key from the secret
+        key and salt. Possible values are ``concat``, ``django-concat``,
+        or ``hmac``. Defaults to :attr:`default_key_derivation`, which
+        defaults to ``django-concat``.
+    :param digest_method: Hash function to use when generating the HMAC
+        signature. Defaults to :attr:`default_digest_method`, which
+        defaults to :func:`hashlib.sha1`. Note that the security of the
+        hash alone doesn't apply when used intermediately in HMAC.
+    :param algorithm: A :class:`SigningAlgorithm` instance to use
+        instead of building a default :class:`HMACAlgorithm` with the
+        ``digest_method``.
+
+    .. versionchanged:: 2.0
+        Added support for key rotation by passing a list to
+        ``secret_key``.
+
+    .. versionchanged:: 0.18
+        ``algorithm`` was added as an argument to the class constructor.
+
+    .. versionchanged:: 0.14
+        ``key_derivation`` and ``digest_method`` were added as arguments
+        to the class constructor.
+    """
+
+    #: The default digest method to use for the signer. The default is
+    #: :func:`hashlib.sha1`, but can be changed to any :mod:`hashlib` or
+    #: compatible object. Note that the security of the hash alone
+    #: doesn't apply when used intermediately in HMAC.
+    #:
+    #: .. versionadded:: 0.14
+    default_digest_method: t.Any = staticmethod(_lazy_sha1)
+
+    #: The default scheme to use to derive the signing key from the
+    #: secret key and salt. The default is ``django-concat``. Possible
+    #: values are ``concat``, ``django-concat``, and ``hmac``.
+    #:
+    #: .. versionadded:: 0.14
+    default_key_derivation: str = "django-concat"
+
+    def __init__(
+        self,
+        secret_key: str | bytes | cabc.Iterable[str] | cabc.Iterable[bytes],
+        salt: str | bytes | None = b"itsdangerous.Signer",
+        sep: str | bytes = b".",
+        key_derivation: str | None = None,
+        digest_method: t.Any | None = None,
+        algorithm: SigningAlgorithm | None = None,
+    ):
+        #: The list of secret keys to try for verifying signatures, from
+        #: oldest to newest. The newest (last) key is used for signing.
+        #:
+        #: This allows a key rotation system to keep a list of allowed
+        #: keys and remove expired ones.
+        self.secret_keys: list[bytes] = _make_keys_list(secret_key)
+        self.sep: bytes = want_bytes(sep)
+
+        if self.sep in _base64_alphabet:
+            raise ValueError(
+                "The given separator cannot be used because it may be"
+                " contained in the signature itself. ASCII letters,"
+                " digits, and '-_=' must not be used."
+            )
+
+        if salt is not None:
+            salt = want_bytes(salt)
+        else:
+            salt = b"itsdangerous.Signer"
+
+        self.salt = salt
+
+        if key_derivation is None:
+            key_derivation = self.default_key_derivation
+
+        self.key_derivation: str = key_derivation
+
+        if digest_method is None:
+            digest_method = self.default_digest_method
+
+        self.digest_method: t.Any = digest_method
+
+        if algorithm is None:
+            algorithm = HMACAlgorithm(self.digest_method)
+
+        self.algorithm: SigningAlgorithm = algorithm
+
+    @property
+    def secret_key(self) -> bytes:
+        """The newest (last) entry in the :attr:`secret_keys` list. This
+        is for compatibility from before key rotation support was added.
+        """
+        return self.secret_keys[-1]
+
+    def derive_key(self, secret_key: str | bytes | None = None) -> bytes:
+        """This method is called to derive the key. The default key
+        derivation choices can be overridden here. Key derivation is not
+        intended to be used as a security method to make a complex key
+        out of a short password. Instead you should use large random
+        secret keys.
+
+        :param secret_key: A specific secret key to derive from.
+            Defaults to the last item in :attr:`secret_keys`.
+
+        .. versionchanged:: 2.0
+            Added the ``secret_key`` parameter.
+        """
+        if secret_key is None:
+            secret_key = self.secret_keys[-1]
+        else:
+            secret_key = want_bytes(secret_key)
+
+        if self.key_derivation == "concat":
+            return t.cast(bytes, self.digest_method(self.salt + secret_key).digest())
+        elif self.key_derivation == "django-concat":
+            return t.cast(
+                bytes, self.digest_method(self.salt + b"signer" + secret_key).digest()
+            )
+        elif self.key_derivation == "hmac":
+            mac = hmac.new(secret_key, digestmod=self.digest_method)
+            mac.update(self.salt)
+            return mac.digest()
+        elif self.key_derivation == "none":
+            return secret_key
+        else:
+            raise TypeError("Unknown key derivation method")
+
+    def get_signature(self, value: str | bytes) -> bytes:
+        """Returns the signature for the given value."""
+        value = want_bytes(value)
+        key = self.derive_key()
+        sig = self.algorithm.get_signature(key, value)
+        return base64_encode(sig)
+
+    def sign(self, value: str | bytes) -> bytes:
+        """Signs the given string."""
+        value = want_bytes(value)
+        return value + self.sep + self.get_signature(value)
+
+    def verify_signature(self, value: str | bytes, sig: str | bytes) -> bool:
+        """Verifies the signature for the given value."""
+        try:
+            sig = base64_decode(sig)
+        except Exception:
+            return False
+
+        value = want_bytes(value)
+
+        for secret_key in reversed(self.secret_keys):
+            key = self.derive_key(secret_key)
+
+            if self.algorithm.verify_signature(key, value, sig):
+                return True
+
+        return False
+
+    def unsign(self, signed_value: str | bytes) -> bytes:
+        """Unsigns the given string."""
+        signed_value = want_bytes(signed_value)
+
+        if self.sep not in signed_value:
+            raise BadSignature(f"No {self.sep!r} found in value")
+
+        value, sig = signed_value.rsplit(self.sep, 1)
+
+        if self.verify_signature(value, sig):
+            return value
+
+        raise BadSignature(f"Signature {sig!r} does not match", payload=value)
+
+    def validate(self, signed_value: str | bytes) -> bool:
+        """Only validates the given signed value. Returns ``True`` if
+        the signature exists and is valid.
+        """
+        try:
+            self.unsign(signed_value)
+            return True
+        except BadSignature:
+            return False
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/timed.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/timed.py
new file mode 100644
index 00000000..73843755
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/timed.py
@@ -0,0 +1,228 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import time
+import typing as t
+from datetime import datetime
+from datetime import timezone
+
+from .encoding import base64_decode
+from .encoding import base64_encode
+from .encoding import bytes_to_int
+from .encoding import int_to_bytes
+from .encoding import want_bytes
+from .exc import BadSignature
+from .exc import BadTimeSignature
+from .exc import SignatureExpired
+from .serializer import _TSerialized
+from .serializer import Serializer
+from .signer import Signer
+
+
+class TimestampSigner(Signer):
+    """Works like the regular :class:`.Signer` but also records the time
+    of the signing and can be used to expire signatures. The
+    :meth:`unsign` method can raise :exc:`.SignatureExpired` if the
+    unsigning failed because the signature is expired.
+    """
+
+    def get_timestamp(self) -> int:
+        """Returns the current timestamp. The function must return an
+        integer.
+        """
+        return int(time.time())
+
+    def timestamp_to_datetime(self, ts: int) -> datetime:
+        """Convert the timestamp from :meth:`get_timestamp` into an
+        aware :class`datetime.datetime` in UTC.
+
+        .. versionchanged:: 2.0
+            The timestamp is returned as a timezone-aware ``datetime``
+            in UTC rather than a naive ``datetime`` assumed to be UTC.
+        """
+        return datetime.fromtimestamp(ts, tz=timezone.utc)
+
+    def sign(self, value: str | bytes) -> bytes:
+        """Signs the given string and also attaches time information."""
+        value = want_bytes(value)
+        timestamp = base64_encode(int_to_bytes(self.get_timestamp()))
+        sep = want_bytes(self.sep)
+        value = value + sep + timestamp
+        return value + sep + self.get_signature(value)
+
+    # Ignore overlapping signatures check, return_timestamp is the only
+    # parameter that affects the return type.
+
+    @t.overload
+    def unsign(  # type: ignore[overload-overlap]
+        self,
+        signed_value: str | bytes,
+        max_age: int | None = None,
+        return_timestamp: t.Literal[False] = False,
+    ) -> bytes: ...
+
+    @t.overload
+    def unsign(
+        self,
+        signed_value: str | bytes,
+        max_age: int | None = None,
+        return_timestamp: t.Literal[True] = True,
+    ) -> tuple[bytes, datetime]: ...
+
+    def unsign(
+        self,
+        signed_value: str | bytes,
+        max_age: int | None = None,
+        return_timestamp: bool = False,
+    ) -> tuple[bytes, datetime] | bytes:
+        """Works like the regular :meth:`.Signer.unsign` but can also
+        validate the time. See the base docstring of the class for
+        the general behavior. If ``return_timestamp`` is ``True`` the
+        timestamp of the signature will be returned as an aware
+        :class:`datetime.datetime` object in UTC.
+
+        .. versionchanged:: 2.0
+            The timestamp is returned as a timezone-aware ``datetime``
+            in UTC rather than a naive ``datetime`` assumed to be UTC.
+        """
+        try:
+            result = super().unsign(signed_value)
+            sig_error = None
+        except BadSignature as e:
+            sig_error = e
+            result = e.payload or b""
+
+        sep = want_bytes(self.sep)
+
+        # If there is no timestamp in the result there is something
+        # seriously wrong. In case there was a signature error, we raise
+        # that one directly, otherwise we have a weird situation in
+        # which we shouldn't have come except someone uses a time-based
+        # serializer on non-timestamp data, so catch that.
+        if sep not in result:
+            if sig_error:
+                raise sig_error
+
+            raise BadTimeSignature("timestamp missing", payload=result)
+
+        value, ts_bytes = result.rsplit(sep, 1)
+        ts_int: int | None = None
+        ts_dt: datetime | None = None
+
+        try:
+            ts_int = bytes_to_int(base64_decode(ts_bytes))
+        except Exception:
+            pass
+
+        # Signature is *not* okay. Raise a proper error now that we have
+        # split the value and the timestamp.
+        if sig_error is not None:
+            if ts_int is not None:
+                try:
+                    ts_dt = self.timestamp_to_datetime(ts_int)
+                except (ValueError, OSError, OverflowError) as exc:
+                    # Windows raises OSError
+                    # 32-bit raises OverflowError
+                    raise BadTimeSignature(
+                        "Malformed timestamp", payload=value
+                    ) from exc
+
+            raise BadTimeSignature(str(sig_error), payload=value, date_signed=ts_dt)
+
+        # Signature was okay but the timestamp is actually not there or
+        # malformed. Should not happen, but we handle it anyway.
+        if ts_int is None:
+            raise BadTimeSignature("Malformed timestamp", payload=value)
+
+        # Check timestamp is not older than max_age
+        if max_age is not None:
+            age = self.get_timestamp() - ts_int
+
+            if age > max_age:
+                raise SignatureExpired(
+                    f"Signature age {age} > {max_age} seconds",
+                    payload=value,
+                    date_signed=self.timestamp_to_datetime(ts_int),
+                )
+
+            if age < 0:
+                raise SignatureExpired(
+                    f"Signature age {age} < 0 seconds",
+                    payload=value,
+                    date_signed=self.timestamp_to_datetime(ts_int),
+                )
+
+        if return_timestamp:
+            return value, self.timestamp_to_datetime(ts_int)
+
+        return value
+
+    def validate(self, signed_value: str | bytes, max_age: int | None = None) -> bool:
+        """Only validates the given signed value. Returns ``True`` if
+        the signature exists and is valid."""
+        try:
+            self.unsign(signed_value, max_age=max_age)
+            return True
+        except BadSignature:
+            return False
+
+
+class TimedSerializer(Serializer[_TSerialized]):
+    """Uses :class:`TimestampSigner` instead of the default
+    :class:`.Signer`.
+    """
+
+    default_signer: type[TimestampSigner] = TimestampSigner
+
+    def iter_unsigners(
+        self, salt: str | bytes | None = None
+    ) -> cabc.Iterator[TimestampSigner]:
+        return t.cast("cabc.Iterator[TimestampSigner]", super().iter_unsigners(salt))
+
+    # TODO: Signature is incompatible because parameters were added
+    #  before salt.
+
+    def loads(  # type: ignore[override]
+        self,
+        s: str | bytes,
+        max_age: int | None = None,
+        return_timestamp: bool = False,
+        salt: str | bytes | None = None,
+    ) -> t.Any:
+        """Reverse of :meth:`dumps`, raises :exc:`.BadSignature` if the
+        signature validation fails. If a ``max_age`` is provided it will
+        ensure the signature is not older than that time in seconds. In
+        case the signature is outdated, :exc:`.SignatureExpired` is
+        raised. All arguments are forwarded to the signer's
+        :meth:`~TimestampSigner.unsign` method.
+        """
+        s = want_bytes(s)
+        last_exception = None
+
+        for signer in self.iter_unsigners(salt):
+            try:
+                base64d, timestamp = signer.unsign(
+                    s, max_age=max_age, return_timestamp=True
+                )
+                payload = self.load_payload(base64d)
+
+                if return_timestamp:
+                    return payload, timestamp
+
+                return payload
+            except SignatureExpired:
+                # The signature was unsigned successfully but was
+                # expired. Do not try the next signer.
+                raise
+            except BadSignature as err:
+                last_exception = err
+
+        raise t.cast(BadSignature, last_exception)
+
+    def loads_unsafe(  # type: ignore[override]
+        self,
+        s: str | bytes,
+        max_age: int | None = None,
+        salt: str | bytes | None = None,
+    ) -> tuple[bool, t.Any]:
+        return self._loads_unsafe_impl(s, salt, load_kwargs={"max_age": max_age})
diff --git a/backend/venv/lib/python3.12/site-packages/itsdangerous/url_safe.py b/backend/venv/lib/python3.12/site-packages/itsdangerous/url_safe.py
new file mode 100644
index 00000000..56a07933
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/itsdangerous/url_safe.py
@@ -0,0 +1,83 @@
+from __future__ import annotations
+
+import typing as t
+import zlib
+
+from ._json import _CompactJSON
+from .encoding import base64_decode
+from .encoding import base64_encode
+from .exc import BadPayload
+from .serializer import _PDataSerializer
+from .serializer import Serializer
+from .timed import TimedSerializer
+
+
+class URLSafeSerializerMixin(Serializer[str]):
+    """Mixed in with a regular serializer it will attempt to zlib
+    compress the string to make it shorter if necessary. It will also
+    base64 encode the string so that it can safely be placed in a URL.
+    """
+
+    default_serializer: _PDataSerializer[str] = _CompactJSON
+
+    def load_payload(
+        self,
+        payload: bytes,
+        *args: t.Any,
+        serializer: t.Any | None = None,
+        **kwargs: t.Any,
+    ) -> t.Any:
+        decompress = False
+
+        if payload.startswith(b"."):
+            payload = payload[1:]
+            decompress = True
+
+        try:
+            json = base64_decode(payload)
+        except Exception as e:
+            raise BadPayload(
+                "Could not base64 decode the payload because of an exception",
+                original_error=e,
+            ) from e
+
+        if decompress:
+            try:
+                json = zlib.decompress(json)
+            except Exception as e:
+                raise BadPayload(
+                    "Could not zlib decompress the payload before decoding the payload",
+                    original_error=e,
+                ) from e
+
+        return super().load_payload(json, *args, **kwargs)
+
+    def dump_payload(self, obj: t.Any) -> bytes:
+        json = super().dump_payload(obj)
+        is_compressed = False
+        compressed = zlib.compress(json)
+
+        if len(compressed) < (len(json) - 1):
+            json = compressed
+            is_compressed = True
+
+        base64d = base64_encode(json)
+
+        if is_compressed:
+            base64d = b"." + base64d
+
+        return base64d
+
+
+class URLSafeSerializer(URLSafeSerializerMixin, Serializer[str]):
+    """Works like :class:`.Serializer` but dumps and loads into a URL
+    safe string consisting of the upper and lowercase character of the
+    alphabet as well as ``'_'``, ``'-'`` and ``'.'``.
+    """
+
+
+class URLSafeTimedSerializer(URLSafeSerializerMixin, TimedSerializer[str]):
+    """Works like :class:`.TimedSerializer` but dumps and loads into a
+    URL safe string consisting of the upper and lowercase character of
+    the alphabet as well as ``'_'``, ``'-'`` and ``'.'``.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/LICENSE.txt
new file mode 100644
index 00000000..c37cae49
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2007 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/METADATA
new file mode 100644
index 00000000..6d0348d9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/METADATA
@@ -0,0 +1,75 @@
+Metadata-Version: 2.3
+Name: Jinja2
+Version: 3.1.5
+Summary: A very fast and expressive template engine.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Classifier: Topic :: Text Processing :: Markup :: HTML
+Classifier: Typing :: Typed
+Requires-Dist: MarkupSafe>=2.0
+Requires-Dist: Babel>=2.7 ; extra == "i18n"
+Project-URL: Changes, https://jinja.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://jinja.palletsprojects.com/
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Source, https://github.com/pallets/jinja/
+Provides-Extra: i18n
+
+# Jinja
+
+Jinja is a fast, expressive, extensible templating engine. Special
+placeholders in the template allow writing code similar to Python
+syntax. Then the template is passed data to render the final document.
+
+It includes:
+
+-   Template inheritance and inclusion.
+-   Define and import macros within templates.
+-   HTML templates can use autoescaping to prevent XSS from untrusted
+    user input.
+-   A sandboxed environment can safely render untrusted templates.
+-   AsyncIO support for generating templates and calling async
+    functions.
+-   I18N support with Babel.
+-   Templates are compiled to optimized Python code just-in-time and
+    cached, or can be compiled ahead-of-time.
+-   Exceptions point to the correct line in templates to make debugging
+    easier.
+-   Extensible filters, tests, functions, and even syntax.
+
+Jinja's philosophy is that while application logic belongs in Python if
+possible, it shouldn't make the template designer's job difficult by
+restricting functionality too much.
+
+
+## In A Nutshell
+
+```jinja
+{% extends "base.html" %}
+{% block title %}Members{% endblock %}
+{% block content %}
+  <ul>
+  {% for user in users %}
+    <li><a href="{{ user.url }}">{{ user.username }}</a></li>
+  {% endfor %}
+  </ul>
+{% endblock %}
+```
+
+## Donate
+
+The Pallets organization develops and supports Jinja and other popular
+packages. In order to grow the community of contributors and users, and
+allow the maintainers to devote more time to the projects, [please
+donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
+
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/RECORD
new file mode 100644
index 00000000..cc15d0b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/RECORD
@@ -0,0 +1,58 @@
+jinja2-3.1.5.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+jinja2-3.1.5.dist-info/LICENSE.txt,sha256=O0nc7kEF6ze6wQ-vG-JgQI_oXSUrjp3y4JefweCUQ3s,1475
+jinja2-3.1.5.dist-info/METADATA,sha256=PJNSUFNBwoqGA2vce2XSP8M_p2EYqAHYI7hoWLABtFo,2593
+jinja2-3.1.5.dist-info/RECORD,,
+jinja2-3.1.5.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+jinja2-3.1.5.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+jinja2-3.1.5.dist-info/entry_points.txt,sha256=OL85gYU1eD8cuPlikifFngXpeBjaxl6rIJ8KkC_3r-I,58
+jinja2/__init__.py,sha256=zpt8UHzpS2eB1c04kn1LkKkaXLXXcKd33klq7UJGIgg,1928
+jinja2/__pycache__/__init__.cpython-312.pyc,,
+jinja2/__pycache__/_identifier.cpython-312.pyc,,
+jinja2/__pycache__/async_utils.cpython-312.pyc,,
+jinja2/__pycache__/bccache.cpython-312.pyc,,
+jinja2/__pycache__/compiler.cpython-312.pyc,,
+jinja2/__pycache__/constants.cpython-312.pyc,,
+jinja2/__pycache__/debug.cpython-312.pyc,,
+jinja2/__pycache__/defaults.cpython-312.pyc,,
+jinja2/__pycache__/environment.cpython-312.pyc,,
+jinja2/__pycache__/exceptions.cpython-312.pyc,,
+jinja2/__pycache__/ext.cpython-312.pyc,,
+jinja2/__pycache__/filters.cpython-312.pyc,,
+jinja2/__pycache__/idtracking.cpython-312.pyc,,
+jinja2/__pycache__/lexer.cpython-312.pyc,,
+jinja2/__pycache__/loaders.cpython-312.pyc,,
+jinja2/__pycache__/meta.cpython-312.pyc,,
+jinja2/__pycache__/nativetypes.cpython-312.pyc,,
+jinja2/__pycache__/nodes.cpython-312.pyc,,
+jinja2/__pycache__/optimizer.cpython-312.pyc,,
+jinja2/__pycache__/parser.cpython-312.pyc,,
+jinja2/__pycache__/runtime.cpython-312.pyc,,
+jinja2/__pycache__/sandbox.cpython-312.pyc,,
+jinja2/__pycache__/tests.cpython-312.pyc,,
+jinja2/__pycache__/utils.cpython-312.pyc,,
+jinja2/__pycache__/visitor.cpython-312.pyc,,
+jinja2/_identifier.py,sha256=_zYctNKzRqlk_murTNlzrju1FFJL7Va_Ijqqd7ii2lU,1958
+jinja2/async_utils.py,sha256=vK-PdsuorOMnWSnEkT3iUJRIkTnYgO2T6MnGxDgHI5o,2834
+jinja2/bccache.py,sha256=gh0qs9rulnXo0PhX5jTJy2UHzI8wFnQ63o_vw7nhzRg,14061
+jinja2/compiler.py,sha256=9RpCQl5X88BHllJiPsHPh295Hh0uApvwFJNQuutULeM,74131
+jinja2/constants.py,sha256=GMoFydBF_kdpaRKPoM5cl5MviquVRLVyZtfp5-16jg0,1433
+jinja2/debug.py,sha256=CnHqCDHd-BVGvti_8ZsTolnXNhA3ECsY-6n_2pwU8Hw,6297
+jinja2/defaults.py,sha256=boBcSw78h-lp20YbaXSJsqkAI2uN_mD_TtCydpeq5wU,1267
+jinja2/environment.py,sha256=9nhrP7Ch-NbGX00wvyr4yy-uhNHq2OCc60ggGrni_fk,61513
+jinja2/exceptions.py,sha256=ioHeHrWwCWNaXX1inHmHVblvc4haO7AXsjCp3GfWvx0,5071
+jinja2/ext.py,sha256=5PF5eHfh8mXAIxXHHRB2xXbXohi8pE3nHSOxa66uS7E,31875
+jinja2/filters.py,sha256=cvRI2pqXNMzw8ba41VOBpgu_wu1r-l1_QxwD6yVoJ5g,55025
+jinja2/idtracking.py,sha256=-ll5lIp73pML3ErUYiIJj7tdmWxcH_IlDv3yA_hiZYo,10555
+jinja2/lexer.py,sha256=LYiYio6br-Tep9nPcupWXsPEtjluw3p1mU-lNBVRUfk,29786
+jinja2/loaders.py,sha256=wIrnxjvcbqh5VwW28NSkfotiDq8qNCxIOSFbGUiSLB4,24055
+jinja2/meta.py,sha256=OTDPkaFvU2Hgvx-6akz7154F8BIWaRmvJcBFvwopHww,4397
+jinja2/nativetypes.py,sha256=7GIGALVJgdyL80oZJdQUaUfwSt5q2lSSZbXt0dNf_M4,4210
+jinja2/nodes.py,sha256=m1Duzcr6qhZI8JQ6VyJgUNinjAf5bQzijSmDnMsvUx8,34579
+jinja2/optimizer.py,sha256=rJnCRlQ7pZsEEmMhsQDgC_pKyDHxP5TPS6zVPGsgcu8,1651
+jinja2/parser.py,sha256=lLOFy3sEmHc5IaEHRiH1sQVnId2moUQzhyeJZTtdY30,40383
+jinja2/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+jinja2/runtime.py,sha256=gDk-GvdriJXqgsGbHgrcKTP0Yp6zPXzhzrIpCFH3jAU,34249
+jinja2/sandbox.py,sha256=Mw2aitlY2I8la7FYhcX2YG9BtUYcLnD0Gh3d29cDWrY,15009
+jinja2/tests.py,sha256=VLsBhVFnWg-PxSBz1MhRnNWgP1ovXk3neO1FLQMeC9Q,5926
+jinja2/utils.py,sha256=rRp3o9e7ZKS4fyrWRbELyLcpuGVTFcnooaOa1qx_FIk,24129
+jinja2/visitor.py,sha256=EcnL1PIwf_4RVCOMxsRNuR8AXHbS1qfAdMOE2ngKJz4,3557
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/entry_points.txt
new file mode 100644
index 00000000..abc3eae3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2-3.1.5.dist-info/entry_points.txt
@@ -0,0 +1,3 @@
+[babel.extractors]
+jinja2=jinja2.ext:babel_extract[i18n]
+
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__init__.py b/backend/venv/lib/python3.12/site-packages/jinja2/__init__.py
new file mode 100644
index 00000000..d669f295
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/__init__.py
@@ -0,0 +1,38 @@
+"""Jinja is a template engine written in pure Python. It provides a
+non-XML syntax that supports inline expressions and an optional
+sandboxed environment.
+"""
+
+from .bccache import BytecodeCache as BytecodeCache
+from .bccache import FileSystemBytecodeCache as FileSystemBytecodeCache
+from .bccache import MemcachedBytecodeCache as MemcachedBytecodeCache
+from .environment import Environment as Environment
+from .environment import Template as Template
+from .exceptions import TemplateAssertionError as TemplateAssertionError
+from .exceptions import TemplateError as TemplateError
+from .exceptions import TemplateNotFound as TemplateNotFound
+from .exceptions import TemplateRuntimeError as TemplateRuntimeError
+from .exceptions import TemplatesNotFound as TemplatesNotFound
+from .exceptions import TemplateSyntaxError as TemplateSyntaxError
+from .exceptions import UndefinedError as UndefinedError
+from .loaders import BaseLoader as BaseLoader
+from .loaders import ChoiceLoader as ChoiceLoader
+from .loaders import DictLoader as DictLoader
+from .loaders import FileSystemLoader as FileSystemLoader
+from .loaders import FunctionLoader as FunctionLoader
+from .loaders import ModuleLoader as ModuleLoader
+from .loaders import PackageLoader as PackageLoader
+from .loaders import PrefixLoader as PrefixLoader
+from .runtime import ChainableUndefined as ChainableUndefined
+from .runtime import DebugUndefined as DebugUndefined
+from .runtime import make_logging_undefined as make_logging_undefined
+from .runtime import StrictUndefined as StrictUndefined
+from .runtime import Undefined as Undefined
+from .utils import clear_caches as clear_caches
+from .utils import is_undefined as is_undefined
+from .utils import pass_context as pass_context
+from .utils import pass_environment as pass_environment
+from .utils import pass_eval_context as pass_eval_context
+from .utils import select_autoescape as select_autoescape
+
+__version__ = "3.1.5"
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..107b844f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/_identifier.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/_identifier.cpython-312.pyc
new file mode 100644
index 00000000..dd6da57e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/_identifier.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/async_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/async_utils.cpython-312.pyc
new file mode 100644
index 00000000..3b986415
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/async_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/bccache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/bccache.cpython-312.pyc
new file mode 100644
index 00000000..c121d6a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/bccache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/compiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/compiler.cpython-312.pyc
new file mode 100644
index 00000000..17b3dc26
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/compiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/constants.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/constants.cpython-312.pyc
new file mode 100644
index 00000000..d8442503
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/constants.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/debug.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/debug.cpython-312.pyc
new file mode 100644
index 00000000..0e589bd3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/debug.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/defaults.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/defaults.cpython-312.pyc
new file mode 100644
index 00000000..187af140
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/defaults.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/environment.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/environment.cpython-312.pyc
new file mode 100644
index 00000000..20ec53ce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/environment.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..82669ad2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/ext.cpython-312.pyc
new file mode 100644
index 00000000..296d0f80
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/filters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/filters.cpython-312.pyc
new file mode 100644
index 00000000..8ed81bfd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/filters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/idtracking.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/idtracking.cpython-312.pyc
new file mode 100644
index 00000000..26b31ec7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/idtracking.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/lexer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/lexer.cpython-312.pyc
new file mode 100644
index 00000000..f98cecfc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/lexer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/loaders.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/loaders.cpython-312.pyc
new file mode 100644
index 00000000..da122ca8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/loaders.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/meta.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/meta.cpython-312.pyc
new file mode 100644
index 00000000..f1414136
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/meta.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nativetypes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nativetypes.cpython-312.pyc
new file mode 100644
index 00000000..54e98a0d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nativetypes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nodes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nodes.cpython-312.pyc
new file mode 100644
index 00000000..46d33230
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/nodes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/optimizer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/optimizer.cpython-312.pyc
new file mode 100644
index 00000000..390712cb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/optimizer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/parser.cpython-312.pyc
new file mode 100644
index 00000000..4767fce3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/runtime.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/runtime.cpython-312.pyc
new file mode 100644
index 00000000..cc665a5a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/runtime.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/sandbox.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/sandbox.cpython-312.pyc
new file mode 100644
index 00000000..3c5e1867
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/sandbox.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/tests.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/tests.cpython-312.pyc
new file mode 100644
index 00000000..8c128be3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/tests.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..6796267f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/visitor.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/visitor.cpython-312.pyc
new file mode 100644
index 00000000..0147d472
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jinja2/__pycache__/visitor.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/_identifier.py b/backend/venv/lib/python3.12/site-packages/jinja2/_identifier.py
new file mode 100644
index 00000000..928c1503
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/_identifier.py
@@ -0,0 +1,6 @@
+import re
+
+# generated by scripts/generate_identifier_pattern.py
+pattern = re.compile(
+    r"[\w·̀-ͯ·҃-֑҇-ׇֽֿׁׂׅׄؐ-ًؚ-ٰٟۖ-ۜ۟-۪ۤۧۨ-ܑۭܰ-݊ަ-ް߫-߽߳ࠖ-࠙ࠛ-ࠣࠥ-ࠧࠩ-࡙࠭-࡛࣓-ࣣ࣡-ःऺ-़ा-ॏ॑-ॗॢॣঁ-ঃ়া-ৄেৈো-্ৗৢৣ৾ਁ-ਃ਼ਾ-ੂੇੈੋ-੍ੑੰੱੵઁ-ઃ઼ા-ૅે-ૉો-્ૢૣૺ-૿ଁ-ଃ଼ା-ୄେୈୋ-୍ୖୗୢୣஂா-ூெ-ைொ-்ௗఀ-ఄా-ౄె-ైొ-్ౕౖౢౣಁ-ಃ಼ಾ-ೄೆ-ೈೊ-್ೕೖೢೣഀ-ഃ഻഼ാ-ൄെ-ൈൊ-്ൗൢൣංඃ්ා-ුූෘ-ෟෲෳัิ-ฺ็-๎ັິ-ູົຼ່-ໍ༹༘༙༵༷༾༿ཱ-྄྆྇ྍ-ྗྙ-ྼ࿆ါ-ှၖ-ၙၞ-ၠၢ-ၤၧ-ၭၱ-ၴႂ-ႍႏႚ-ႝ፝-፟ᜒ-᜔ᜲ-᜴ᝒᝓᝲᝳ឴-៓៝᠋-᠍ᢅᢆᢩᤠ-ᤫᤰ-᤻ᨗ-ᨛᩕ-ᩞ᩠-᩿᩼᪰-᪽ᬀ-ᬄ᬴-᭄᭫-᭳ᮀ-ᮂᮡ-ᮭ᯦-᯳ᰤ-᰷᳐-᳔᳒-᳨᳭ᳲ-᳴᳷-᳹᷀-᷹᷻-᷿‿⁀⁔⃐-⃥⃜⃡-⃰℘℮⳯-⵿⳱ⷠ-〪ⷿ-゙゚〯꙯ꙴ-꙽ꚞꚟ꛰꛱ꠂ꠆ꠋꠣ-ꠧꢀꢁꢴ-ꣅ꣠-꣱ꣿꤦ-꤭ꥇ-꥓ꦀ-ꦃ꦳-꧀ꧥꨩ-ꨶꩃꩌꩍꩻ-ꩽꪰꪲ-ꪴꪷꪸꪾ꪿꫁ꫫ-ꫯꫵ꫶ꯣ-ꯪ꯬꯭ﬞ︀-️︠-︯︳︴﹍-﹏＿𐇽𐋠𐍶-𐍺𐨁-𐨃𐨅𐨆𐨌-𐨏𐨸-𐨿𐨺𐫦𐫥𐴤-𐽆𐴧-𐽐𑀀-𑀂𑀸-𑁆𑁿-𑂂𑂰-𑂺𑄀-𑄂𑄧-𑄴𑅅𑅆𑅳𑆀-𑆂𑆳-𑇀𑇉-𑇌𑈬-𑈷𑈾𑋟-𑋪𑌀-𑌃𑌻𑌼𑌾-𑍄𑍇𑍈𑍋-𑍍𑍗𑍢𑍣𑍦-𑍬𑍰-𑍴𑐵-𑑆𑑞𑒰-𑓃𑖯-𑖵𑖸-𑗀𑗜𑗝𑘰-𑙀𑚫-𑚷𑜝-𑜫𑠬-𑠺𑨁-𑨊𑨳-𑨹𑨻-𑨾𑩇𑩑-𑩛𑪊-𑪙𑰯-𑰶𑰸-𑰿𑲒-𑲧𑲩-𑲶𑴱-𑴶𑴺𑴼𑴽𑴿-𑵅𑵇𑶊-𑶎𑶐𑶑𑶓-𑶗𑻳-𑻶𖫰-𖫴𖬰-𖬶𖽑-𖽾𖾏-𖾒𛲝𛲞𝅥-𝅩𝅭-𝅲𝅻-𝆂𝆅-𝆋𝆪-𝆭𝉂-𝉄𝨀-𝨶𝨻-𝩬𝩵𝪄𝪛-𝪟𝪡-𝪯𞀀-𞀆𞀈-𞀘𞀛-𞀡𞀣𞀤𞀦-𞣐𞀪-𞣖𞥄-𞥊󠄀-󠇯]+"  # noqa: B950
+)
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/async_utils.py b/backend/venv/lib/python3.12/site-packages/jinja2/async_utils.py
new file mode 100644
index 00000000..f0c14020
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/async_utils.py
@@ -0,0 +1,99 @@
+import inspect
+import typing as t
+from functools import WRAPPER_ASSIGNMENTS
+from functools import wraps
+
+from .utils import _PassArg
+from .utils import pass_eval_context
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+V = t.TypeVar("V")
+
+
+def async_variant(normal_func):  # type: ignore
+    def decorator(async_func):  # type: ignore
+        pass_arg = _PassArg.from_obj(normal_func)
+        need_eval_context = pass_arg is None
+
+        if pass_arg is _PassArg.environment:
+
+            def is_async(args: t.Any) -> bool:
+                return t.cast(bool, args[0].is_async)
+
+        else:
+
+            def is_async(args: t.Any) -> bool:
+                return t.cast(bool, args[0].environment.is_async)
+
+        # Take the doc and annotations from the sync function, but the
+        # name from the async function. Pallets-Sphinx-Themes
+        # build_function_directive expects __wrapped__ to point to the
+        # sync function.
+        async_func_attrs = ("__module__", "__name__", "__qualname__")
+        normal_func_attrs = tuple(set(WRAPPER_ASSIGNMENTS).difference(async_func_attrs))
+
+        @wraps(normal_func, assigned=normal_func_attrs)
+        @wraps(async_func, assigned=async_func_attrs, updated=())
+        def wrapper(*args, **kwargs):  # type: ignore
+            b = is_async(args)
+
+            if need_eval_context:
+                args = args[1:]
+
+            if b:
+                return async_func(*args, **kwargs)
+
+            return normal_func(*args, **kwargs)
+
+        if need_eval_context:
+            wrapper = pass_eval_context(wrapper)
+
+        wrapper.jinja_async_variant = True  # type: ignore[attr-defined]
+        return wrapper
+
+    return decorator
+
+
+_common_primitives = {int, float, bool, str, list, dict, tuple, type(None)}
+
+
+async def auto_await(value: t.Union[t.Awaitable["V"], "V"]) -> "V":
+    # Avoid a costly call to isawaitable
+    if type(value) in _common_primitives:
+        return t.cast("V", value)
+
+    if inspect.isawaitable(value):
+        return await t.cast("t.Awaitable[V]", value)
+
+    return value
+
+
+class _IteratorToAsyncIterator(t.Generic[V]):
+    def __init__(self, iterator: "t.Iterator[V]"):
+        self._iterator = iterator
+
+    def __aiter__(self) -> "te.Self":
+        return self
+
+    async def __anext__(self) -> V:
+        try:
+            return next(self._iterator)
+        except StopIteration as e:
+            raise StopAsyncIteration(e.value) from e
+
+
+def auto_aiter(
+    iterable: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+) -> "t.AsyncIterator[V]":
+    if hasattr(iterable, "__aiter__"):
+        return iterable.__aiter__()
+    else:
+        return _IteratorToAsyncIterator(iter(iterable))
+
+
+async def auto_to_list(
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+) -> t.List["V"]:
+    return [x async for x in auto_aiter(value)]
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/bccache.py b/backend/venv/lib/python3.12/site-packages/jinja2/bccache.py
new file mode 100644
index 00000000..ada8b099
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/bccache.py
@@ -0,0 +1,408 @@
+"""The optional bytecode cache system. This is useful if you have very
+complex template situations and the compilation of all those templates
+slows down your application too much.
+
+Situations where this is useful are often forking web applications that
+are initialized on the first request.
+"""
+
+import errno
+import fnmatch
+import marshal
+import os
+import pickle
+import stat
+import sys
+import tempfile
+import typing as t
+from hashlib import sha1
+from io import BytesIO
+from types import CodeType
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+
+    class _MemcachedClient(te.Protocol):
+        def get(self, key: str) -> bytes: ...
+
+        def set(
+            self, key: str, value: bytes, timeout: t.Optional[int] = None
+        ) -> None: ...
+
+
+bc_version = 5
+# Magic bytes to identify Jinja bytecode cache files. Contains the
+# Python major and minor version to avoid loading incompatible bytecode
+# if a project upgrades its Python version.
+bc_magic = (
+    b"j2"
+    + pickle.dumps(bc_version, 2)
+    + pickle.dumps((sys.version_info[0] << 24) | sys.version_info[1], 2)
+)
+
+
+class Bucket:
+    """Buckets are used to store the bytecode for one template.  It's created
+    and initialized by the bytecode cache and passed to the loading functions.
+
+    The buckets get an internal checksum from the cache assigned and use this
+    to automatically reject outdated cache material.  Individual bytecode
+    cache subclasses don't have to care about cache invalidation.
+    """
+
+    def __init__(self, environment: "Environment", key: str, checksum: str) -> None:
+        self.environment = environment
+        self.key = key
+        self.checksum = checksum
+        self.reset()
+
+    def reset(self) -> None:
+        """Resets the bucket (unloads the bytecode)."""
+        self.code: t.Optional[CodeType] = None
+
+    def load_bytecode(self, f: t.BinaryIO) -> None:
+        """Loads bytecode from a file or file like object."""
+        # make sure the magic header is correct
+        magic = f.read(len(bc_magic))
+        if magic != bc_magic:
+            self.reset()
+            return
+        # the source code of the file changed, we need to reload
+        checksum = pickle.load(f)
+        if self.checksum != checksum:
+            self.reset()
+            return
+        # if marshal_load fails then we need to reload
+        try:
+            self.code = marshal.load(f)
+        except (EOFError, ValueError, TypeError):
+            self.reset()
+            return
+
+    def write_bytecode(self, f: t.IO[bytes]) -> None:
+        """Dump the bytecode into the file or file like object passed."""
+        if self.code is None:
+            raise TypeError("can't write empty bucket")
+        f.write(bc_magic)
+        pickle.dump(self.checksum, f, 2)
+        marshal.dump(self.code, f)
+
+    def bytecode_from_string(self, string: bytes) -> None:
+        """Load bytecode from bytes."""
+        self.load_bytecode(BytesIO(string))
+
+    def bytecode_to_string(self) -> bytes:
+        """Return the bytecode as bytes."""
+        out = BytesIO()
+        self.write_bytecode(out)
+        return out.getvalue()
+
+
+class BytecodeCache:
+    """To implement your own bytecode cache you have to subclass this class
+    and override :meth:`load_bytecode` and :meth:`dump_bytecode`.  Both of
+    these methods are passed a :class:`~jinja2.bccache.Bucket`.
+
+    A very basic bytecode cache that saves the bytecode on the file system::
+
+        from os import path
+
+        class MyCache(BytecodeCache):
+
+            def __init__(self, directory):
+                self.directory = directory
+
+            def load_bytecode(self, bucket):
+                filename = path.join(self.directory, bucket.key)
+                if path.exists(filename):
+                    with open(filename, 'rb') as f:
+                        bucket.load_bytecode(f)
+
+            def dump_bytecode(self, bucket):
+                filename = path.join(self.directory, bucket.key)
+                with open(filename, 'wb') as f:
+                    bucket.write_bytecode(f)
+
+    A more advanced version of a filesystem based bytecode cache is part of
+    Jinja.
+    """
+
+    def load_bytecode(self, bucket: Bucket) -> None:
+        """Subclasses have to override this method to load bytecode into a
+        bucket.  If they are not able to find code in the cache for the
+        bucket, it must not do anything.
+        """
+        raise NotImplementedError()
+
+    def dump_bytecode(self, bucket: Bucket) -> None:
+        """Subclasses have to override this method to write the bytecode
+        from a bucket back to the cache.  If it unable to do so it must not
+        fail silently but raise an exception.
+        """
+        raise NotImplementedError()
+
+    def clear(self) -> None:
+        """Clears the cache.  This method is not used by Jinja but should be
+        implemented to allow applications to clear the bytecode cache used
+        by a particular environment.
+        """
+
+    def get_cache_key(
+        self, name: str, filename: t.Optional[t.Union[str]] = None
+    ) -> str:
+        """Returns the unique hash key for this template name."""
+        hash = sha1(name.encode("utf-8"))
+
+        if filename is not None:
+            hash.update(f"|{filename}".encode())
+
+        return hash.hexdigest()
+
+    def get_source_checksum(self, source: str) -> str:
+        """Returns a checksum for the source."""
+        return sha1(source.encode("utf-8")).hexdigest()
+
+    def get_bucket(
+        self,
+        environment: "Environment",
+        name: str,
+        filename: t.Optional[str],
+        source: str,
+    ) -> Bucket:
+        """Return a cache bucket for the given template.  All arguments are
+        mandatory but filename may be `None`.
+        """
+        key = self.get_cache_key(name, filename)
+        checksum = self.get_source_checksum(source)
+        bucket = Bucket(environment, key, checksum)
+        self.load_bytecode(bucket)
+        return bucket
+
+    def set_bucket(self, bucket: Bucket) -> None:
+        """Put the bucket into the cache."""
+        self.dump_bytecode(bucket)
+
+
+class FileSystemBytecodeCache(BytecodeCache):
+    """A bytecode cache that stores bytecode on the filesystem.  It accepts
+    two arguments: The directory where the cache items are stored and a
+    pattern string that is used to build the filename.
+
+    If no directory is specified a default cache directory is selected.  On
+    Windows the user's temp directory is used, on UNIX systems a directory
+    is created for the user in the system temp directory.
+
+    The pattern can be used to have multiple separate caches operate on the
+    same directory.  The default pattern is ``'__jinja2_%s.cache'``.  ``%s``
+    is replaced with the cache key.
+
+    >>> bcc = FileSystemBytecodeCache('/tmp/jinja_cache', '%s.cache')
+
+    This bytecode cache supports clearing of the cache using the clear method.
+    """
+
+    def __init__(
+        self, directory: t.Optional[str] = None, pattern: str = "__jinja2_%s.cache"
+    ) -> None:
+        if directory is None:
+            directory = self._get_default_cache_dir()
+        self.directory = directory
+        self.pattern = pattern
+
+    def _get_default_cache_dir(self) -> str:
+        def _unsafe_dir() -> "te.NoReturn":
+            raise RuntimeError(
+                "Cannot determine safe temp directory.  You "
+                "need to explicitly provide one."
+            )
+
+        tmpdir = tempfile.gettempdir()
+
+        # On windows the temporary directory is used specific unless
+        # explicitly forced otherwise.  We can just use that.
+        if os.name == "nt":
+            return tmpdir
+        if not hasattr(os, "getuid"):
+            _unsafe_dir()
+
+        dirname = f"_jinja2-cache-{os.getuid()}"
+        actual_dir = os.path.join(tmpdir, dirname)
+
+        try:
+            os.mkdir(actual_dir, stat.S_IRWXU)
+        except OSError as e:
+            if e.errno != errno.EEXIST:
+                raise
+        try:
+            os.chmod(actual_dir, stat.S_IRWXU)
+            actual_dir_stat = os.lstat(actual_dir)
+            if (
+                actual_dir_stat.st_uid != os.getuid()
+                or not stat.S_ISDIR(actual_dir_stat.st_mode)
+                or stat.S_IMODE(actual_dir_stat.st_mode) != stat.S_IRWXU
+            ):
+                _unsafe_dir()
+        except OSError as e:
+            if e.errno != errno.EEXIST:
+                raise
+
+        actual_dir_stat = os.lstat(actual_dir)
+        if (
+            actual_dir_stat.st_uid != os.getuid()
+            or not stat.S_ISDIR(actual_dir_stat.st_mode)
+            or stat.S_IMODE(actual_dir_stat.st_mode) != stat.S_IRWXU
+        ):
+            _unsafe_dir()
+
+        return actual_dir
+
+    def _get_cache_filename(self, bucket: Bucket) -> str:
+        return os.path.join(self.directory, self.pattern % (bucket.key,))
+
+    def load_bytecode(self, bucket: Bucket) -> None:
+        filename = self._get_cache_filename(bucket)
+
+        # Don't test for existence before opening the file, since the
+        # file could disappear after the test before the open.
+        try:
+            f = open(filename, "rb")
+        except (FileNotFoundError, IsADirectoryError, PermissionError):
+            # PermissionError can occur on Windows when an operation is
+            # in progress, such as calling clear().
+            return
+
+        with f:
+            bucket.load_bytecode(f)
+
+    def dump_bytecode(self, bucket: Bucket) -> None:
+        # Write to a temporary file, then rename to the real name after
+        # writing. This avoids another process reading the file before
+        # it is fully written.
+        name = self._get_cache_filename(bucket)
+        f = tempfile.NamedTemporaryFile(
+            mode="wb",
+            dir=os.path.dirname(name),
+            prefix=os.path.basename(name),
+            suffix=".tmp",
+            delete=False,
+        )
+
+        def remove_silent() -> None:
+            try:
+                os.remove(f.name)
+            except OSError:
+                # Another process may have called clear(). On Windows,
+                # another program may be holding the file open.
+                pass
+
+        try:
+            with f:
+                bucket.write_bytecode(f)
+        except BaseException:
+            remove_silent()
+            raise
+
+        try:
+            os.replace(f.name, name)
+        except OSError:
+            # Another process may have called clear(). On Windows,
+            # another program may be holding the file open.
+            remove_silent()
+        except BaseException:
+            remove_silent()
+            raise
+
+    def clear(self) -> None:
+        # imported lazily here because google app-engine doesn't support
+        # write access on the file system and the function does not exist
+        # normally.
+        from os import remove
+
+        files = fnmatch.filter(os.listdir(self.directory), self.pattern % ("*",))
+        for filename in files:
+            try:
+                remove(os.path.join(self.directory, filename))
+            except OSError:
+                pass
+
+
+class MemcachedBytecodeCache(BytecodeCache):
+    """This class implements a bytecode cache that uses a memcache cache for
+    storing the information.  It does not enforce a specific memcache library
+    (tummy's memcache or cmemcache) but will accept any class that provides
+    the minimal interface required.
+
+    Libraries compatible with this class:
+
+    -   `cachelib <https://github.com/pallets/cachelib>`_
+    -   `python-memcached <https://pypi.org/project/python-memcached/>`_
+
+    (Unfortunately the django cache interface is not compatible because it
+    does not support storing binary data, only text. You can however pass
+    the underlying cache client to the bytecode cache which is available
+    as `django.core.cache.cache._client`.)
+
+    The minimal interface for the client passed to the constructor is this:
+
+    .. class:: MinimalClientInterface
+
+        .. method:: set(key, value[, timeout])
+
+            Stores the bytecode in the cache.  `value` is a string and
+            `timeout` the timeout of the key.  If timeout is not provided
+            a default timeout or no timeout should be assumed, if it's
+            provided it's an integer with the number of seconds the cache
+            item should exist.
+
+        .. method:: get(key)
+
+            Returns the value for the cache key.  If the item does not
+            exist in the cache the return value must be `None`.
+
+    The other arguments to the constructor are the prefix for all keys that
+    is added before the actual cache key and the timeout for the bytecode in
+    the cache system.  We recommend a high (or no) timeout.
+
+    This bytecode cache does not support clearing of used items in the cache.
+    The clear method is a no-operation function.
+
+    .. versionadded:: 2.7
+       Added support for ignoring memcache errors through the
+       `ignore_memcache_errors` parameter.
+    """
+
+    def __init__(
+        self,
+        client: "_MemcachedClient",
+        prefix: str = "jinja2/bytecode/",
+        timeout: t.Optional[int] = None,
+        ignore_memcache_errors: bool = True,
+    ):
+        self.client = client
+        self.prefix = prefix
+        self.timeout = timeout
+        self.ignore_memcache_errors = ignore_memcache_errors
+
+    def load_bytecode(self, bucket: Bucket) -> None:
+        try:
+            code = self.client.get(self.prefix + bucket.key)
+        except Exception:
+            if not self.ignore_memcache_errors:
+                raise
+        else:
+            bucket.bytecode_from_string(code)
+
+    def dump_bytecode(self, bucket: Bucket) -> None:
+        key = self.prefix + bucket.key
+        value = bucket.bytecode_to_string()
+
+        try:
+            if self.timeout is not None:
+                self.client.set(key, value, self.timeout)
+            else:
+                self.client.set(key, value)
+        except Exception:
+            if not self.ignore_memcache_errors:
+                raise
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/compiler.py b/backend/venv/lib/python3.12/site-packages/jinja2/compiler.py
new file mode 100644
index 00000000..a4ff6a1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/compiler.py
@@ -0,0 +1,1998 @@
+"""Compiles nodes from the parser into Python code."""
+
+import typing as t
+from contextlib import contextmanager
+from functools import update_wrapper
+from io import StringIO
+from itertools import chain
+from keyword import iskeyword as is_python_keyword
+
+from markupsafe import escape
+from markupsafe import Markup
+
+from . import nodes
+from .exceptions import TemplateAssertionError
+from .idtracking import Symbols
+from .idtracking import VAR_LOAD_ALIAS
+from .idtracking import VAR_LOAD_PARAMETER
+from .idtracking import VAR_LOAD_RESOLVE
+from .idtracking import VAR_LOAD_UNDEFINED
+from .nodes import EvalContext
+from .optimizer import Optimizer
+from .utils import _PassArg
+from .utils import concat
+from .visitor import NodeVisitor
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+operators = {
+    "eq": "==",
+    "ne": "!=",
+    "gt": ">",
+    "gteq": ">=",
+    "lt": "<",
+    "lteq": "<=",
+    "in": "in",
+    "notin": "not in",
+}
+
+
+def optimizeconst(f: F) -> F:
+    def new_func(
+        self: "CodeGenerator", node: nodes.Expr, frame: "Frame", **kwargs: t.Any
+    ) -> t.Any:
+        # Only optimize if the frame is not volatile
+        if self.optimizer is not None and not frame.eval_ctx.volatile:
+            new_node = self.optimizer.visit(node, frame.eval_ctx)
+
+            if new_node != node:
+                return self.visit(new_node, frame)
+
+        return f(self, node, frame, **kwargs)
+
+    return update_wrapper(new_func, f)  # type: ignore[return-value]
+
+
+def _make_binop(op: str) -> t.Callable[["CodeGenerator", nodes.BinExpr, "Frame"], None]:
+    @optimizeconst
+    def visitor(self: "CodeGenerator", node: nodes.BinExpr, frame: Frame) -> None:
+        if (
+            self.environment.sandboxed and op in self.environment.intercepted_binops  # type: ignore
+        ):
+            self.write(f"environment.call_binop(context, {op!r}, ")
+            self.visit(node.left, frame)
+            self.write(", ")
+            self.visit(node.right, frame)
+        else:
+            self.write("(")
+            self.visit(node.left, frame)
+            self.write(f" {op} ")
+            self.visit(node.right, frame)
+
+        self.write(")")
+
+    return visitor
+
+
+def _make_unop(
+    op: str,
+) -> t.Callable[["CodeGenerator", nodes.UnaryExpr, "Frame"], None]:
+    @optimizeconst
+    def visitor(self: "CodeGenerator", node: nodes.UnaryExpr, frame: Frame) -> None:
+        if (
+            self.environment.sandboxed and op in self.environment.intercepted_unops  # type: ignore
+        ):
+            self.write(f"environment.call_unop(context, {op!r}, ")
+            self.visit(node.node, frame)
+        else:
+            self.write("(" + op)
+            self.visit(node.node, frame)
+
+        self.write(")")
+
+    return visitor
+
+
+def generate(
+    node: nodes.Template,
+    environment: "Environment",
+    name: t.Optional[str],
+    filename: t.Optional[str],
+    stream: t.Optional[t.TextIO] = None,
+    defer_init: bool = False,
+    optimized: bool = True,
+) -> t.Optional[str]:
+    """Generate the python source for a node tree."""
+    if not isinstance(node, nodes.Template):
+        raise TypeError("Can't compile non template nodes")
+
+    generator = environment.code_generator_class(
+        environment, name, filename, stream, defer_init, optimized
+    )
+    generator.visit(node)
+
+    if stream is None:
+        return generator.stream.getvalue()  # type: ignore
+
+    return None
+
+
+def has_safe_repr(value: t.Any) -> bool:
+    """Does the node have a safe representation?"""
+    if value is None or value is NotImplemented or value is Ellipsis:
+        return True
+
+    if type(value) in {bool, int, float, complex, range, str, Markup}:
+        return True
+
+    if type(value) in {tuple, list, set, frozenset}:
+        return all(has_safe_repr(v) for v in value)
+
+    if type(value) is dict:  # noqa E721
+        return all(has_safe_repr(k) and has_safe_repr(v) for k, v in value.items())
+
+    return False
+
+
+def find_undeclared(
+    nodes: t.Iterable[nodes.Node], names: t.Iterable[str]
+) -> t.Set[str]:
+    """Check if the names passed are accessed undeclared.  The return value
+    is a set of all the undeclared names from the sequence of names found.
+    """
+    visitor = UndeclaredNameVisitor(names)
+    try:
+        for node in nodes:
+            visitor.visit(node)
+    except VisitorExit:
+        pass
+    return visitor.undeclared
+
+
+class MacroRef:
+    def __init__(self, node: t.Union[nodes.Macro, nodes.CallBlock]) -> None:
+        self.node = node
+        self.accesses_caller = False
+        self.accesses_kwargs = False
+        self.accesses_varargs = False
+
+
+class Frame:
+    """Holds compile time information for us."""
+
+    def __init__(
+        self,
+        eval_ctx: EvalContext,
+        parent: t.Optional["Frame"] = None,
+        level: t.Optional[int] = None,
+    ) -> None:
+        self.eval_ctx = eval_ctx
+
+        # the parent of this frame
+        self.parent = parent
+
+        if parent is None:
+            self.symbols = Symbols(level=level)
+
+            # in some dynamic inheritance situations the compiler needs to add
+            # write tests around output statements.
+            self.require_output_check = False
+
+            # inside some tags we are using a buffer rather than yield statements.
+            # this for example affects {% filter %} or {% macro %}.  If a frame
+            # is buffered this variable points to the name of the list used as
+            # buffer.
+            self.buffer: t.Optional[str] = None
+
+            # the name of the block we're in, otherwise None.
+            self.block: t.Optional[str] = None
+
+        else:
+            self.symbols = Symbols(parent.symbols, level=level)
+            self.require_output_check = parent.require_output_check
+            self.buffer = parent.buffer
+            self.block = parent.block
+
+        # a toplevel frame is the root + soft frames such as if conditions.
+        self.toplevel = False
+
+        # the root frame is basically just the outermost frame, so no if
+        # conditions.  This information is used to optimize inheritance
+        # situations.
+        self.rootlevel = False
+
+        # variables set inside of loops and blocks should not affect outer frames,
+        # but they still needs to be kept track of as part of the active context.
+        self.loop_frame = False
+        self.block_frame = False
+
+        # track whether the frame is being used in an if-statement or conditional
+        # expression as it determines which errors should be raised during runtime
+        # or compile time.
+        self.soft_frame = False
+
+    def copy(self) -> "te.Self":
+        """Create a copy of the current one."""
+        rv = object.__new__(self.__class__)
+        rv.__dict__.update(self.__dict__)
+        rv.symbols = self.symbols.copy()
+        return rv
+
+    def inner(self, isolated: bool = False) -> "Frame":
+        """Return an inner frame."""
+        if isolated:
+            return Frame(self.eval_ctx, level=self.symbols.level + 1)
+        return Frame(self.eval_ctx, self)
+
+    def soft(self) -> "te.Self":
+        """Return a soft frame.  A soft frame may not be modified as
+        standalone thing as it shares the resources with the frame it
+        was created of, but it's not a rootlevel frame any longer.
+
+        This is only used to implement if-statements and conditional
+        expressions.
+        """
+        rv = self.copy()
+        rv.rootlevel = False
+        rv.soft_frame = True
+        return rv
+
+    __copy__ = copy
+
+
+class VisitorExit(RuntimeError):
+    """Exception used by the `UndeclaredNameVisitor` to signal a stop."""
+
+
+class DependencyFinderVisitor(NodeVisitor):
+    """A visitor that collects filter and test calls."""
+
+    def __init__(self) -> None:
+        self.filters: t.Set[str] = set()
+        self.tests: t.Set[str] = set()
+
+    def visit_Filter(self, node: nodes.Filter) -> None:
+        self.generic_visit(node)
+        self.filters.add(node.name)
+
+    def visit_Test(self, node: nodes.Test) -> None:
+        self.generic_visit(node)
+        self.tests.add(node.name)
+
+    def visit_Block(self, node: nodes.Block) -> None:
+        """Stop visiting at blocks."""
+
+
+class UndeclaredNameVisitor(NodeVisitor):
+    """A visitor that checks if a name is accessed without being
+    declared.  This is different from the frame visitor as it will
+    not stop at closure frames.
+    """
+
+    def __init__(self, names: t.Iterable[str]) -> None:
+        self.names = set(names)
+        self.undeclared: t.Set[str] = set()
+
+    def visit_Name(self, node: nodes.Name) -> None:
+        if node.ctx == "load" and node.name in self.names:
+            self.undeclared.add(node.name)
+            if self.undeclared == self.names:
+                raise VisitorExit()
+        else:
+            self.names.discard(node.name)
+
+    def visit_Block(self, node: nodes.Block) -> None:
+        """Stop visiting a blocks."""
+
+
+class CompilerExit(Exception):
+    """Raised if the compiler encountered a situation where it just
+    doesn't make sense to further process the code.  Any block that
+    raises such an exception is not further processed.
+    """
+
+
+class CodeGenerator(NodeVisitor):
+    def __init__(
+        self,
+        environment: "Environment",
+        name: t.Optional[str],
+        filename: t.Optional[str],
+        stream: t.Optional[t.TextIO] = None,
+        defer_init: bool = False,
+        optimized: bool = True,
+    ) -> None:
+        if stream is None:
+            stream = StringIO()
+        self.environment = environment
+        self.name = name
+        self.filename = filename
+        self.stream = stream
+        self.created_block_context = False
+        self.defer_init = defer_init
+        self.optimizer: t.Optional[Optimizer] = None
+
+        if optimized:
+            self.optimizer = Optimizer(environment)
+
+        # aliases for imports
+        self.import_aliases: t.Dict[str, str] = {}
+
+        # a registry for all blocks.  Because blocks are moved out
+        # into the global python scope they are registered here
+        self.blocks: t.Dict[str, nodes.Block] = {}
+
+        # the number of extends statements so far
+        self.extends_so_far = 0
+
+        # some templates have a rootlevel extends.  In this case we
+        # can safely assume that we're a child template and do some
+        # more optimizations.
+        self.has_known_extends = False
+
+        # the current line number
+        self.code_lineno = 1
+
+        # registry of all filters and tests (global, not block local)
+        self.tests: t.Dict[str, str] = {}
+        self.filters: t.Dict[str, str] = {}
+
+        # the debug information
+        self.debug_info: t.List[t.Tuple[int, int]] = []
+        self._write_debug_info: t.Optional[int] = None
+
+        # the number of new lines before the next write()
+        self._new_lines = 0
+
+        # the line number of the last written statement
+        self._last_line = 0
+
+        # true if nothing was written so far.
+        self._first_write = True
+
+        # used by the `temporary_identifier` method to get new
+        # unique, temporary identifier
+        self._last_identifier = 0
+
+        # the current indentation
+        self._indentation = 0
+
+        # Tracks toplevel assignments
+        self._assign_stack: t.List[t.Set[str]] = []
+
+        # Tracks parameter definition blocks
+        self._param_def_block: t.List[t.Set[str]] = []
+
+        # Tracks the current context.
+        self._context_reference_stack = ["context"]
+
+    @property
+    def optimized(self) -> bool:
+        return self.optimizer is not None
+
+    # -- Various compilation helpers
+
+    def fail(self, msg: str, lineno: int) -> "te.NoReturn":
+        """Fail with a :exc:`TemplateAssertionError`."""
+        raise TemplateAssertionError(msg, lineno, self.name, self.filename)
+
+    def temporary_identifier(self) -> str:
+        """Get a new unique identifier."""
+        self._last_identifier += 1
+        return f"t_{self._last_identifier}"
+
+    def buffer(self, frame: Frame) -> None:
+        """Enable buffering for the frame from that point onwards."""
+        frame.buffer = self.temporary_identifier()
+        self.writeline(f"{frame.buffer} = []")
+
+    def return_buffer_contents(
+        self, frame: Frame, force_unescaped: bool = False
+    ) -> None:
+        """Return the buffer contents of the frame."""
+        if not force_unescaped:
+            if frame.eval_ctx.volatile:
+                self.writeline("if context.eval_ctx.autoescape:")
+                self.indent()
+                self.writeline(f"return Markup(concat({frame.buffer}))")
+                self.outdent()
+                self.writeline("else:")
+                self.indent()
+                self.writeline(f"return concat({frame.buffer})")
+                self.outdent()
+                return
+            elif frame.eval_ctx.autoescape:
+                self.writeline(f"return Markup(concat({frame.buffer}))")
+                return
+        self.writeline(f"return concat({frame.buffer})")
+
+    def indent(self) -> None:
+        """Indent by one."""
+        self._indentation += 1
+
+    def outdent(self, step: int = 1) -> None:
+        """Outdent by step."""
+        self._indentation -= step
+
+    def start_write(self, frame: Frame, node: t.Optional[nodes.Node] = None) -> None:
+        """Yield or write into the frame buffer."""
+        if frame.buffer is None:
+            self.writeline("yield ", node)
+        else:
+            self.writeline(f"{frame.buffer}.append(", node)
+
+    def end_write(self, frame: Frame) -> None:
+        """End the writing process started by `start_write`."""
+        if frame.buffer is not None:
+            self.write(")")
+
+    def simple_write(
+        self, s: str, frame: Frame, node: t.Optional[nodes.Node] = None
+    ) -> None:
+        """Simple shortcut for start_write + write + end_write."""
+        self.start_write(frame, node)
+        self.write(s)
+        self.end_write(frame)
+
+    def blockvisit(self, nodes: t.Iterable[nodes.Node], frame: Frame) -> None:
+        """Visit a list of nodes as block in a frame.  If the current frame
+        is no buffer a dummy ``if 0: yield None`` is written automatically.
+        """
+        try:
+            self.writeline("pass")
+            for node in nodes:
+                self.visit(node, frame)
+        except CompilerExit:
+            pass
+
+    def write(self, x: str) -> None:
+        """Write a string into the output stream."""
+        if self._new_lines:
+            if not self._first_write:
+                self.stream.write("\n" * self._new_lines)
+                self.code_lineno += self._new_lines
+                if self._write_debug_info is not None:
+                    self.debug_info.append((self._write_debug_info, self.code_lineno))
+                    self._write_debug_info = None
+            self._first_write = False
+            self.stream.write("    " * self._indentation)
+            self._new_lines = 0
+        self.stream.write(x)
+
+    def writeline(
+        self, x: str, node: t.Optional[nodes.Node] = None, extra: int = 0
+    ) -> None:
+        """Combination of newline and write."""
+        self.newline(node, extra)
+        self.write(x)
+
+    def newline(self, node: t.Optional[nodes.Node] = None, extra: int = 0) -> None:
+        """Add one or more newlines before the next write."""
+        self._new_lines = max(self._new_lines, 1 + extra)
+        if node is not None and node.lineno != self._last_line:
+            self._write_debug_info = node.lineno
+            self._last_line = node.lineno
+
+    def signature(
+        self,
+        node: t.Union[nodes.Call, nodes.Filter, nodes.Test],
+        frame: Frame,
+        extra_kwargs: t.Optional[t.Mapping[str, t.Any]] = None,
+    ) -> None:
+        """Writes a function call to the stream for the current node.
+        A leading comma is added automatically.  The extra keyword
+        arguments may not include python keywords otherwise a syntax
+        error could occur.  The extra keyword arguments should be given
+        as python dict.
+        """
+        # if any of the given keyword arguments is a python keyword
+        # we have to make sure that no invalid call is created.
+        kwarg_workaround = any(
+            is_python_keyword(t.cast(str, k))
+            for k in chain((x.key for x in node.kwargs), extra_kwargs or ())
+        )
+
+        for arg in node.args:
+            self.write(", ")
+            self.visit(arg, frame)
+
+        if not kwarg_workaround:
+            for kwarg in node.kwargs:
+                self.write(", ")
+                self.visit(kwarg, frame)
+            if extra_kwargs is not None:
+                for key, value in extra_kwargs.items():
+                    self.write(f", {key}={value}")
+        if node.dyn_args:
+            self.write(", *")
+            self.visit(node.dyn_args, frame)
+
+        if kwarg_workaround:
+            if node.dyn_kwargs is not None:
+                self.write(", **dict({")
+            else:
+                self.write(", **{")
+            for kwarg in node.kwargs:
+                self.write(f"{kwarg.key!r}: ")
+                self.visit(kwarg.value, frame)
+                self.write(", ")
+            if extra_kwargs is not None:
+                for key, value in extra_kwargs.items():
+                    self.write(f"{key!r}: {value}, ")
+            if node.dyn_kwargs is not None:
+                self.write("}, **")
+                self.visit(node.dyn_kwargs, frame)
+                self.write(")")
+            else:
+                self.write("}")
+
+        elif node.dyn_kwargs is not None:
+            self.write(", **")
+            self.visit(node.dyn_kwargs, frame)
+
+    def pull_dependencies(self, nodes: t.Iterable[nodes.Node]) -> None:
+        """Find all filter and test names used in the template and
+        assign them to variables in the compiled namespace. Checking
+        that the names are registered with the environment is done when
+        compiling the Filter and Test nodes. If the node is in an If or
+        CondExpr node, the check is done at runtime instead.
+
+        .. versionchanged:: 3.0
+            Filters and tests in If and CondExpr nodes are checked at
+            runtime instead of compile time.
+        """
+        visitor = DependencyFinderVisitor()
+
+        for node in nodes:
+            visitor.visit(node)
+
+        for id_map, names, dependency in (
+            (self.filters, visitor.filters, "filters"),
+            (
+                self.tests,
+                visitor.tests,
+                "tests",
+            ),
+        ):
+            for name in sorted(names):
+                if name not in id_map:
+                    id_map[name] = self.temporary_identifier()
+
+                # add check during runtime that dependencies used inside of executed
+                # blocks are defined, as this step may be skipped during compile time
+                self.writeline("try:")
+                self.indent()
+                self.writeline(f"{id_map[name]} = environment.{dependency}[{name!r}]")
+                self.outdent()
+                self.writeline("except KeyError:")
+                self.indent()
+                self.writeline("@internalcode")
+                self.writeline(f"def {id_map[name]}(*unused):")
+                self.indent()
+                self.writeline(
+                    f'raise TemplateRuntimeError("No {dependency[:-1]}'
+                    f' named {name!r} found.")'
+                )
+                self.outdent()
+                self.outdent()
+
+    def enter_frame(self, frame: Frame) -> None:
+        undefs = []
+        for target, (action, param) in frame.symbols.loads.items():
+            if action == VAR_LOAD_PARAMETER:
+                pass
+            elif action == VAR_LOAD_RESOLVE:
+                self.writeline(f"{target} = {self.get_resolve_func()}({param!r})")
+            elif action == VAR_LOAD_ALIAS:
+                self.writeline(f"{target} = {param}")
+            elif action == VAR_LOAD_UNDEFINED:
+                undefs.append(target)
+            else:
+                raise NotImplementedError("unknown load instruction")
+        if undefs:
+            self.writeline(f"{' = '.join(undefs)} = missing")
+
+    def leave_frame(self, frame: Frame, with_python_scope: bool = False) -> None:
+        if not with_python_scope:
+            undefs = []
+            for target in frame.symbols.loads:
+                undefs.append(target)
+            if undefs:
+                self.writeline(f"{' = '.join(undefs)} = missing")
+
+    def choose_async(self, async_value: str = "async ", sync_value: str = "") -> str:
+        return async_value if self.environment.is_async else sync_value
+
+    def func(self, name: str) -> str:
+        return f"{self.choose_async()}def {name}"
+
+    def macro_body(
+        self, node: t.Union[nodes.Macro, nodes.CallBlock], frame: Frame
+    ) -> t.Tuple[Frame, MacroRef]:
+        """Dump the function def of a macro or call block."""
+        frame = frame.inner()
+        frame.symbols.analyze_node(node)
+        macro_ref = MacroRef(node)
+
+        explicit_caller = None
+        skip_special_params = set()
+        args = []
+
+        for idx, arg in enumerate(node.args):
+            if arg.name == "caller":
+                explicit_caller = idx
+            if arg.name in ("kwargs", "varargs"):
+                skip_special_params.add(arg.name)
+            args.append(frame.symbols.ref(arg.name))
+
+        undeclared = find_undeclared(node.body, ("caller", "kwargs", "varargs"))
+
+        if "caller" in undeclared:
+            # In older Jinja versions there was a bug that allowed caller
+            # to retain the special behavior even if it was mentioned in
+            # the argument list.  However thankfully this was only really
+            # working if it was the last argument.  So we are explicitly
+            # checking this now and error out if it is anywhere else in
+            # the argument list.
+            if explicit_caller is not None:
+                try:
+                    node.defaults[explicit_caller - len(node.args)]
+                except IndexError:
+                    self.fail(
+                        "When defining macros or call blocks the "
+                        'special "caller" argument must be omitted '
+                        "or be given a default.",
+                        node.lineno,
+                    )
+            else:
+                args.append(frame.symbols.declare_parameter("caller"))
+            macro_ref.accesses_caller = True
+        if "kwargs" in undeclared and "kwargs" not in skip_special_params:
+            args.append(frame.symbols.declare_parameter("kwargs"))
+            macro_ref.accesses_kwargs = True
+        if "varargs" in undeclared and "varargs" not in skip_special_params:
+            args.append(frame.symbols.declare_parameter("varargs"))
+            macro_ref.accesses_varargs = True
+
+        # macros are delayed, they never require output checks
+        frame.require_output_check = False
+        frame.symbols.analyze_node(node)
+        self.writeline(f"{self.func('macro')}({', '.join(args)}):", node)
+        self.indent()
+
+        self.buffer(frame)
+        self.enter_frame(frame)
+
+        self.push_parameter_definitions(frame)
+        for idx, arg in enumerate(node.args):
+            ref = frame.symbols.ref(arg.name)
+            self.writeline(f"if {ref} is missing:")
+            self.indent()
+            try:
+                default = node.defaults[idx - len(node.args)]
+            except IndexError:
+                self.writeline(
+                    f'{ref} = undefined("parameter {arg.name!r} was not provided",'
+                    f" name={arg.name!r})"
+                )
+            else:
+                self.writeline(f"{ref} = ")
+                self.visit(default, frame)
+            self.mark_parameter_stored(ref)
+            self.outdent()
+        self.pop_parameter_definitions()
+
+        self.blockvisit(node.body, frame)
+        self.return_buffer_contents(frame, force_unescaped=True)
+        self.leave_frame(frame, with_python_scope=True)
+        self.outdent()
+
+        return frame, macro_ref
+
+    def macro_def(self, macro_ref: MacroRef, frame: Frame) -> None:
+        """Dump the macro definition for the def created by macro_body."""
+        arg_tuple = ", ".join(repr(x.name) for x in macro_ref.node.args)
+        name = getattr(macro_ref.node, "name", None)
+        if len(macro_ref.node.args) == 1:
+            arg_tuple += ","
+        self.write(
+            f"Macro(environment, macro, {name!r}, ({arg_tuple}),"
+            f" {macro_ref.accesses_kwargs!r}, {macro_ref.accesses_varargs!r},"
+            f" {macro_ref.accesses_caller!r}, context.eval_ctx.autoescape)"
+        )
+
+    def position(self, node: nodes.Node) -> str:
+        """Return a human readable position for the node."""
+        rv = f"line {node.lineno}"
+        if self.name is not None:
+            rv = f"{rv} in {self.name!r}"
+        return rv
+
+    def dump_local_context(self, frame: Frame) -> str:
+        items_kv = ", ".join(
+            f"{name!r}: {target}"
+            for name, target in frame.symbols.dump_stores().items()
+        )
+        return f"{{{items_kv}}}"
+
+    def write_commons(self) -> None:
+        """Writes a common preamble that is used by root and block functions.
+        Primarily this sets up common local helpers and enforces a generator
+        through a dead branch.
+        """
+        self.writeline("resolve = context.resolve_or_missing")
+        self.writeline("undefined = environment.undefined")
+        self.writeline("concat = environment.concat")
+        # always use the standard Undefined class for the implicit else of
+        # conditional expressions
+        self.writeline("cond_expr_undefined = Undefined")
+        self.writeline("if 0: yield None")
+
+    def push_parameter_definitions(self, frame: Frame) -> None:
+        """Pushes all parameter targets from the given frame into a local
+        stack that permits tracking of yet to be assigned parameters.  In
+        particular this enables the optimization from `visit_Name` to skip
+        undefined expressions for parameters in macros as macros can reference
+        otherwise unbound parameters.
+        """
+        self._param_def_block.append(frame.symbols.dump_param_targets())
+
+    def pop_parameter_definitions(self) -> None:
+        """Pops the current parameter definitions set."""
+        self._param_def_block.pop()
+
+    def mark_parameter_stored(self, target: str) -> None:
+        """Marks a parameter in the current parameter definitions as stored.
+        This will skip the enforced undefined checks.
+        """
+        if self._param_def_block:
+            self._param_def_block[-1].discard(target)
+
+    def push_context_reference(self, target: str) -> None:
+        self._context_reference_stack.append(target)
+
+    def pop_context_reference(self) -> None:
+        self._context_reference_stack.pop()
+
+    def get_context_ref(self) -> str:
+        return self._context_reference_stack[-1]
+
+    def get_resolve_func(self) -> str:
+        target = self._context_reference_stack[-1]
+        if target == "context":
+            return "resolve"
+        return f"{target}.resolve"
+
+    def derive_context(self, frame: Frame) -> str:
+        return f"{self.get_context_ref()}.derived({self.dump_local_context(frame)})"
+
+    def parameter_is_undeclared(self, target: str) -> bool:
+        """Checks if a given target is an undeclared parameter."""
+        if not self._param_def_block:
+            return False
+        return target in self._param_def_block[-1]
+
+    def push_assign_tracking(self) -> None:
+        """Pushes a new layer for assignment tracking."""
+        self._assign_stack.append(set())
+
+    def pop_assign_tracking(self, frame: Frame) -> None:
+        """Pops the topmost level for assignment tracking and updates the
+        context variables if necessary.
+        """
+        vars = self._assign_stack.pop()
+        if (
+            not frame.block_frame
+            and not frame.loop_frame
+            and not frame.toplevel
+            or not vars
+        ):
+            return
+        public_names = [x for x in vars if x[:1] != "_"]
+        if len(vars) == 1:
+            name = next(iter(vars))
+            ref = frame.symbols.ref(name)
+            if frame.loop_frame:
+                self.writeline(f"_loop_vars[{name!r}] = {ref}")
+                return
+            if frame.block_frame:
+                self.writeline(f"_block_vars[{name!r}] = {ref}")
+                return
+            self.writeline(f"context.vars[{name!r}] = {ref}")
+        else:
+            if frame.loop_frame:
+                self.writeline("_loop_vars.update({")
+            elif frame.block_frame:
+                self.writeline("_block_vars.update({")
+            else:
+                self.writeline("context.vars.update({")
+            for idx, name in enumerate(sorted(vars)):
+                if idx:
+                    self.write(", ")
+                ref = frame.symbols.ref(name)
+                self.write(f"{name!r}: {ref}")
+            self.write("})")
+        if not frame.block_frame and not frame.loop_frame and public_names:
+            if len(public_names) == 1:
+                self.writeline(f"context.exported_vars.add({public_names[0]!r})")
+            else:
+                names_str = ", ".join(map(repr, sorted(public_names)))
+                self.writeline(f"context.exported_vars.update(({names_str}))")
+
+    # -- Statement Visitors
+
+    def visit_Template(
+        self, node: nodes.Template, frame: t.Optional[Frame] = None
+    ) -> None:
+        assert frame is None, "no root frame allowed"
+        eval_ctx = EvalContext(self.environment, self.name)
+
+        from .runtime import async_exported
+        from .runtime import exported
+
+        if self.environment.is_async:
+            exported_names = sorted(exported + async_exported)
+        else:
+            exported_names = sorted(exported)
+
+        self.writeline("from jinja2.runtime import " + ", ".join(exported_names))
+
+        # if we want a deferred initialization we cannot move the
+        # environment into a local name
+        envenv = "" if self.defer_init else ", environment=environment"
+
+        # do we have an extends tag at all?  If not, we can save some
+        # overhead by just not processing any inheritance code.
+        have_extends = node.find(nodes.Extends) is not None
+
+        # find all blocks
+        for block in node.find_all(nodes.Block):
+            if block.name in self.blocks:
+                self.fail(f"block {block.name!r} defined twice", block.lineno)
+            self.blocks[block.name] = block
+
+        # find all imports and import them
+        for import_ in node.find_all(nodes.ImportedName):
+            if import_.importname not in self.import_aliases:
+                imp = import_.importname
+                self.import_aliases[imp] = alias = self.temporary_identifier()
+                if "." in imp:
+                    module, obj = imp.rsplit(".", 1)
+                    self.writeline(f"from {module} import {obj} as {alias}")
+                else:
+                    self.writeline(f"import {imp} as {alias}")
+
+        # add the load name
+        self.writeline(f"name = {self.name!r}")
+
+        # generate the root render function.
+        self.writeline(
+            f"{self.func('root')}(context, missing=missing{envenv}):", extra=1
+        )
+        self.indent()
+        self.write_commons()
+
+        # process the root
+        frame = Frame(eval_ctx)
+        if "self" in find_undeclared(node.body, ("self",)):
+            ref = frame.symbols.declare_parameter("self")
+            self.writeline(f"{ref} = TemplateReference(context)")
+        frame.symbols.analyze_node(node)
+        frame.toplevel = frame.rootlevel = True
+        frame.require_output_check = have_extends and not self.has_known_extends
+        if have_extends:
+            self.writeline("parent_template = None")
+        self.enter_frame(frame)
+        self.pull_dependencies(node.body)
+        self.blockvisit(node.body, frame)
+        self.leave_frame(frame, with_python_scope=True)
+        self.outdent()
+
+        # make sure that the parent root is called.
+        if have_extends:
+            if not self.has_known_extends:
+                self.indent()
+                self.writeline("if parent_template is not None:")
+            self.indent()
+            if not self.environment.is_async:
+                self.writeline("yield from parent_template.root_render_func(context)")
+            else:
+                self.writeline("agen = parent_template.root_render_func(context)")
+                self.writeline("try:")
+                self.indent()
+                self.writeline("async for event in agen:")
+                self.indent()
+                self.writeline("yield event")
+                self.outdent()
+                self.outdent()
+                self.writeline("finally: await agen.aclose()")
+            self.outdent(1 + (not self.has_known_extends))
+
+        # at this point we now have the blocks collected and can visit them too.
+        for name, block in self.blocks.items():
+            self.writeline(
+                f"{self.func('block_' + name)}(context, missing=missing{envenv}):",
+                block,
+                1,
+            )
+            self.indent()
+            self.write_commons()
+            # It's important that we do not make this frame a child of the
+            # toplevel template.  This would cause a variety of
+            # interesting issues with identifier tracking.
+            block_frame = Frame(eval_ctx)
+            block_frame.block_frame = True
+            undeclared = find_undeclared(block.body, ("self", "super"))
+            if "self" in undeclared:
+                ref = block_frame.symbols.declare_parameter("self")
+                self.writeline(f"{ref} = TemplateReference(context)")
+            if "super" in undeclared:
+                ref = block_frame.symbols.declare_parameter("super")
+                self.writeline(f"{ref} = context.super({name!r}, block_{name})")
+            block_frame.symbols.analyze_node(block)
+            block_frame.block = name
+            self.writeline("_block_vars = {}")
+            self.enter_frame(block_frame)
+            self.pull_dependencies(block.body)
+            self.blockvisit(block.body, block_frame)
+            self.leave_frame(block_frame, with_python_scope=True)
+            self.outdent()
+
+        blocks_kv_str = ", ".join(f"{x!r}: block_{x}" for x in self.blocks)
+        self.writeline(f"blocks = {{{blocks_kv_str}}}", extra=1)
+        debug_kv_str = "&".join(f"{k}={v}" for k, v in self.debug_info)
+        self.writeline(f"debug_info = {debug_kv_str!r}")
+
+    def visit_Block(self, node: nodes.Block, frame: Frame) -> None:
+        """Call a block and register it for the template."""
+        level = 0
+        if frame.toplevel:
+            # if we know that we are a child template, there is no need to
+            # check if we are one
+            if self.has_known_extends:
+                return
+            if self.extends_so_far > 0:
+                self.writeline("if parent_template is None:")
+                self.indent()
+                level += 1
+
+        if node.scoped:
+            context = self.derive_context(frame)
+        else:
+            context = self.get_context_ref()
+
+        if node.required:
+            self.writeline(f"if len(context.blocks[{node.name!r}]) <= 1:", node)
+            self.indent()
+            self.writeline(
+                f'raise TemplateRuntimeError("Required block {node.name!r} not found")',
+                node,
+            )
+            self.outdent()
+
+        if not self.environment.is_async and frame.buffer is None:
+            self.writeline(
+                f"yield from context.blocks[{node.name!r}][0]({context})", node
+            )
+        else:
+            self.writeline(f"gen = context.blocks[{node.name!r}][0]({context})")
+            self.writeline("try:")
+            self.indent()
+            self.writeline(
+                f"{self.choose_async()}for event in gen:",
+                node,
+            )
+            self.indent()
+            self.simple_write("event", frame)
+            self.outdent()
+            self.outdent()
+            self.writeline(
+                f"finally: {self.choose_async('await gen.aclose()', 'gen.close()')}"
+            )
+
+        self.outdent(level)
+
+    def visit_Extends(self, node: nodes.Extends, frame: Frame) -> None:
+        """Calls the extender."""
+        if not frame.toplevel:
+            self.fail("cannot use extend from a non top-level scope", node.lineno)
+
+        # if the number of extends statements in general is zero so
+        # far, we don't have to add a check if something extended
+        # the template before this one.
+        if self.extends_so_far > 0:
+            # if we have a known extends we just add a template runtime
+            # error into the generated code.  We could catch that at compile
+            # time too, but i welcome it not to confuse users by throwing the
+            # same error at different times just "because we can".
+            if not self.has_known_extends:
+                self.writeline("if parent_template is not None:")
+                self.indent()
+            self.writeline('raise TemplateRuntimeError("extended multiple times")')
+
+            # if we have a known extends already we don't need that code here
+            # as we know that the template execution will end here.
+            if self.has_known_extends:
+                raise CompilerExit()
+            else:
+                self.outdent()
+
+        self.writeline("parent_template = environment.get_template(", node)
+        self.visit(node.template, frame)
+        self.write(f", {self.name!r})")
+        self.writeline("for name, parent_block in parent_template.blocks.items():")
+        self.indent()
+        self.writeline("context.blocks.setdefault(name, []).append(parent_block)")
+        self.outdent()
+
+        # if this extends statement was in the root level we can take
+        # advantage of that information and simplify the generated code
+        # in the top level from this point onwards
+        if frame.rootlevel:
+            self.has_known_extends = True
+
+        # and now we have one more
+        self.extends_so_far += 1
+
+    def visit_Include(self, node: nodes.Include, frame: Frame) -> None:
+        """Handles includes."""
+        if node.ignore_missing:
+            self.writeline("try:")
+            self.indent()
+
+        func_name = "get_or_select_template"
+        if isinstance(node.template, nodes.Const):
+            if isinstance(node.template.value, str):
+                func_name = "get_template"
+            elif isinstance(node.template.value, (tuple, list)):
+                func_name = "select_template"
+        elif isinstance(node.template, (nodes.Tuple, nodes.List)):
+            func_name = "select_template"
+
+        self.writeline(f"template = environment.{func_name}(", node)
+        self.visit(node.template, frame)
+        self.write(f", {self.name!r})")
+        if node.ignore_missing:
+            self.outdent()
+            self.writeline("except TemplateNotFound:")
+            self.indent()
+            self.writeline("pass")
+            self.outdent()
+            self.writeline("else:")
+            self.indent()
+
+        def loop_body() -> None:
+            self.indent()
+            self.simple_write("event", frame)
+            self.outdent()
+
+        if node.with_context:
+            self.writeline(
+                f"gen = template.root_render_func("
+                "template.new_context(context.get_all(), True,"
+                f" {self.dump_local_context(frame)}))"
+            )
+            self.writeline("try:")
+            self.indent()
+            self.writeline(f"{self.choose_async()}for event in gen:")
+            loop_body()
+            self.outdent()
+            self.writeline(
+                f"finally: {self.choose_async('await gen.aclose()', 'gen.close()')}"
+            )
+        elif self.environment.is_async:
+            self.writeline(
+                "for event in (await template._get_default_module_async())"
+                "._body_stream:"
+            )
+            loop_body()
+        else:
+            self.writeline("yield from template._get_default_module()._body_stream")
+
+        if node.ignore_missing:
+            self.outdent()
+
+    def _import_common(
+        self, node: t.Union[nodes.Import, nodes.FromImport], frame: Frame
+    ) -> None:
+        self.write(f"{self.choose_async('await ')}environment.get_template(")
+        self.visit(node.template, frame)
+        self.write(f", {self.name!r}).")
+
+        if node.with_context:
+            f_name = f"make_module{self.choose_async('_async')}"
+            self.write(
+                f"{f_name}(context.get_all(), True, {self.dump_local_context(frame)})"
+            )
+        else:
+            self.write(f"_get_default_module{self.choose_async('_async')}(context)")
+
+    def visit_Import(self, node: nodes.Import, frame: Frame) -> None:
+        """Visit regular imports."""
+        self.writeline(f"{frame.symbols.ref(node.target)} = ", node)
+        if frame.toplevel:
+            self.write(f"context.vars[{node.target!r}] = ")
+
+        self._import_common(node, frame)
+
+        if frame.toplevel and not node.target.startswith("_"):
+            self.writeline(f"context.exported_vars.discard({node.target!r})")
+
+    def visit_FromImport(self, node: nodes.FromImport, frame: Frame) -> None:
+        """Visit named imports."""
+        self.newline(node)
+        self.write("included_template = ")
+        self._import_common(node, frame)
+        var_names = []
+        discarded_names = []
+        for name in node.names:
+            if isinstance(name, tuple):
+                name, alias = name
+            else:
+                alias = name
+            self.writeline(
+                f"{frame.symbols.ref(alias)} ="
+                f" getattr(included_template, {name!r}, missing)"
+            )
+            self.writeline(f"if {frame.symbols.ref(alias)} is missing:")
+            self.indent()
+            # The position will contain the template name, and will be formatted
+            # into a string that will be compiled into an f-string. Curly braces
+            # in the name must be replaced with escapes so that they will not be
+            # executed as part of the f-string.
+            position = self.position(node).replace("{", "{{").replace("}", "}}")
+            message = (
+                "the template {included_template.__name__!r}"
+                f" (imported on {position})"
+                f" does not export the requested name {name!r}"
+            )
+            self.writeline(
+                f"{frame.symbols.ref(alias)} = undefined(f{message!r}, name={name!r})"
+            )
+            self.outdent()
+            if frame.toplevel:
+                var_names.append(alias)
+                if not alias.startswith("_"):
+                    discarded_names.append(alias)
+
+        if var_names:
+            if len(var_names) == 1:
+                name = var_names[0]
+                self.writeline(f"context.vars[{name!r}] = {frame.symbols.ref(name)}")
+            else:
+                names_kv = ", ".join(
+                    f"{name!r}: {frame.symbols.ref(name)}" for name in var_names
+                )
+                self.writeline(f"context.vars.update({{{names_kv}}})")
+        if discarded_names:
+            if len(discarded_names) == 1:
+                self.writeline(f"context.exported_vars.discard({discarded_names[0]!r})")
+            else:
+                names_str = ", ".join(map(repr, discarded_names))
+                self.writeline(
+                    f"context.exported_vars.difference_update(({names_str}))"
+                )
+
+    def visit_For(self, node: nodes.For, frame: Frame) -> None:
+        loop_frame = frame.inner()
+        loop_frame.loop_frame = True
+        test_frame = frame.inner()
+        else_frame = frame.inner()
+
+        # try to figure out if we have an extended loop.  An extended loop
+        # is necessary if the loop is in recursive mode if the special loop
+        # variable is accessed in the body if the body is a scoped block.
+        extended_loop = (
+            node.recursive
+            or "loop"
+            in find_undeclared(node.iter_child_nodes(only=("body",)), ("loop",))
+            or any(block.scoped for block in node.find_all(nodes.Block))
+        )
+
+        loop_ref = None
+        if extended_loop:
+            loop_ref = loop_frame.symbols.declare_parameter("loop")
+
+        loop_frame.symbols.analyze_node(node, for_branch="body")
+        if node.else_:
+            else_frame.symbols.analyze_node(node, for_branch="else")
+
+        if node.test:
+            loop_filter_func = self.temporary_identifier()
+            test_frame.symbols.analyze_node(node, for_branch="test")
+            self.writeline(f"{self.func(loop_filter_func)}(fiter):", node.test)
+            self.indent()
+            self.enter_frame(test_frame)
+            self.writeline(self.choose_async("async for ", "for "))
+            self.visit(node.target, loop_frame)
+            self.write(" in ")
+            self.write(self.choose_async("auto_aiter(fiter)", "fiter"))
+            self.write(":")
+            self.indent()
+            self.writeline("if ", node.test)
+            self.visit(node.test, test_frame)
+            self.write(":")
+            self.indent()
+            self.writeline("yield ")
+            self.visit(node.target, loop_frame)
+            self.outdent(3)
+            self.leave_frame(test_frame, with_python_scope=True)
+
+        # if we don't have an recursive loop we have to find the shadowed
+        # variables at that point.  Because loops can be nested but the loop
+        # variable is a special one we have to enforce aliasing for it.
+        if node.recursive:
+            self.writeline(
+                f"{self.func('loop')}(reciter, loop_render_func, depth=0):", node
+            )
+            self.indent()
+            self.buffer(loop_frame)
+
+            # Use the same buffer for the else frame
+            else_frame.buffer = loop_frame.buffer
+
+        # make sure the loop variable is a special one and raise a template
+        # assertion error if a loop tries to write to loop
+        if extended_loop:
+            self.writeline(f"{loop_ref} = missing")
+
+        for name in node.find_all(nodes.Name):
+            if name.ctx == "store" and name.name == "loop":
+                self.fail(
+                    "Can't assign to special loop variable in for-loop target",
+                    name.lineno,
+                )
+
+        if node.else_:
+            iteration_indicator = self.temporary_identifier()
+            self.writeline(f"{iteration_indicator} = 1")
+
+        self.writeline(self.choose_async("async for ", "for "), node)
+        self.visit(node.target, loop_frame)
+        if extended_loop:
+            self.write(f", {loop_ref} in {self.choose_async('Async')}LoopContext(")
+        else:
+            self.write(" in ")
+
+        if node.test:
+            self.write(f"{loop_filter_func}(")
+        if node.recursive:
+            self.write("reciter")
+        else:
+            if self.environment.is_async and not extended_loop:
+                self.write("auto_aiter(")
+            self.visit(node.iter, frame)
+            if self.environment.is_async and not extended_loop:
+                self.write(")")
+        if node.test:
+            self.write(")")
+
+        if node.recursive:
+            self.write(", undefined, loop_render_func, depth):")
+        else:
+            self.write(", undefined):" if extended_loop else ":")
+
+        self.indent()
+        self.enter_frame(loop_frame)
+
+        self.writeline("_loop_vars = {}")
+        self.blockvisit(node.body, loop_frame)
+        if node.else_:
+            self.writeline(f"{iteration_indicator} = 0")
+        self.outdent()
+        self.leave_frame(
+            loop_frame, with_python_scope=node.recursive and not node.else_
+        )
+
+        if node.else_:
+            self.writeline(f"if {iteration_indicator}:")
+            self.indent()
+            self.enter_frame(else_frame)
+            self.blockvisit(node.else_, else_frame)
+            self.leave_frame(else_frame)
+            self.outdent()
+
+        # if the node was recursive we have to return the buffer contents
+        # and start the iteration code
+        if node.recursive:
+            self.return_buffer_contents(loop_frame)
+            self.outdent()
+            self.start_write(frame, node)
+            self.write(f"{self.choose_async('await ')}loop(")
+            if self.environment.is_async:
+                self.write("auto_aiter(")
+            self.visit(node.iter, frame)
+            if self.environment.is_async:
+                self.write(")")
+            self.write(", loop)")
+            self.end_write(frame)
+
+        # at the end of the iteration, clear any assignments made in the
+        # loop from the top level
+        if self._assign_stack:
+            self._assign_stack[-1].difference_update(loop_frame.symbols.stores)
+
+    def visit_If(self, node: nodes.If, frame: Frame) -> None:
+        if_frame = frame.soft()
+        self.writeline("if ", node)
+        self.visit(node.test, if_frame)
+        self.write(":")
+        self.indent()
+        self.blockvisit(node.body, if_frame)
+        self.outdent()
+        for elif_ in node.elif_:
+            self.writeline("elif ", elif_)
+            self.visit(elif_.test, if_frame)
+            self.write(":")
+            self.indent()
+            self.blockvisit(elif_.body, if_frame)
+            self.outdent()
+        if node.else_:
+            self.writeline("else:")
+            self.indent()
+            self.blockvisit(node.else_, if_frame)
+            self.outdent()
+
+    def visit_Macro(self, node: nodes.Macro, frame: Frame) -> None:
+        macro_frame, macro_ref = self.macro_body(node, frame)
+        self.newline()
+        if frame.toplevel:
+            if not node.name.startswith("_"):
+                self.write(f"context.exported_vars.add({node.name!r})")
+            self.writeline(f"context.vars[{node.name!r}] = ")
+        self.write(f"{frame.symbols.ref(node.name)} = ")
+        self.macro_def(macro_ref, macro_frame)
+
+    def visit_CallBlock(self, node: nodes.CallBlock, frame: Frame) -> None:
+        call_frame, macro_ref = self.macro_body(node, frame)
+        self.writeline("caller = ")
+        self.macro_def(macro_ref, call_frame)
+        self.start_write(frame, node)
+        self.visit_Call(node.call, frame, forward_caller=True)
+        self.end_write(frame)
+
+    def visit_FilterBlock(self, node: nodes.FilterBlock, frame: Frame) -> None:
+        filter_frame = frame.inner()
+        filter_frame.symbols.analyze_node(node)
+        self.enter_frame(filter_frame)
+        self.buffer(filter_frame)
+        self.blockvisit(node.body, filter_frame)
+        self.start_write(frame, node)
+        self.visit_Filter(node.filter, filter_frame)
+        self.end_write(frame)
+        self.leave_frame(filter_frame)
+
+    def visit_With(self, node: nodes.With, frame: Frame) -> None:
+        with_frame = frame.inner()
+        with_frame.symbols.analyze_node(node)
+        self.enter_frame(with_frame)
+        for target, expr in zip(node.targets, node.values):
+            self.newline()
+            self.visit(target, with_frame)
+            self.write(" = ")
+            self.visit(expr, frame)
+        self.blockvisit(node.body, with_frame)
+        self.leave_frame(with_frame)
+
+    def visit_ExprStmt(self, node: nodes.ExprStmt, frame: Frame) -> None:
+        self.newline(node)
+        self.visit(node.node, frame)
+
+    class _FinalizeInfo(t.NamedTuple):
+        const: t.Optional[t.Callable[..., str]]
+        src: t.Optional[str]
+
+    @staticmethod
+    def _default_finalize(value: t.Any) -> t.Any:
+        """The default finalize function if the environment isn't
+        configured with one. Or, if the environment has one, this is
+        called on that function's output for constants.
+        """
+        return str(value)
+
+    _finalize: t.Optional[_FinalizeInfo] = None
+
+    def _make_finalize(self) -> _FinalizeInfo:
+        """Build the finalize function to be used on constants and at
+        runtime. Cached so it's only created once for all output nodes.
+
+        Returns a ``namedtuple`` with the following attributes:
+
+        ``const``
+            A function to finalize constant data at compile time.
+
+        ``src``
+            Source code to output around nodes to be evaluated at
+            runtime.
+        """
+        if self._finalize is not None:
+            return self._finalize
+
+        finalize: t.Optional[t.Callable[..., t.Any]]
+        finalize = default = self._default_finalize
+        src = None
+
+        if self.environment.finalize:
+            src = "environment.finalize("
+            env_finalize = self.environment.finalize
+            pass_arg = {
+                _PassArg.context: "context",
+                _PassArg.eval_context: "context.eval_ctx",
+                _PassArg.environment: "environment",
+            }.get(
+                _PassArg.from_obj(env_finalize)  # type: ignore
+            )
+            finalize = None
+
+            if pass_arg is None:
+
+                def finalize(value: t.Any) -> t.Any:  # noqa: F811
+                    return default(env_finalize(value))
+
+            else:
+                src = f"{src}{pass_arg}, "
+
+                if pass_arg == "environment":
+
+                    def finalize(value: t.Any) -> t.Any:  # noqa: F811
+                        return default(env_finalize(self.environment, value))
+
+        self._finalize = self._FinalizeInfo(finalize, src)
+        return self._finalize
+
+    def _output_const_repr(self, group: t.Iterable[t.Any]) -> str:
+        """Given a group of constant values converted from ``Output``
+        child nodes, produce a string to write to the template module
+        source.
+        """
+        return repr(concat(group))
+
+    def _output_child_to_const(
+        self, node: nodes.Expr, frame: Frame, finalize: _FinalizeInfo
+    ) -> str:
+        """Try to optimize a child of an ``Output`` node by trying to
+        convert it to constant, finalized data at compile time.
+
+        If :exc:`Impossible` is raised, the node is not constant and
+        will be evaluated at runtime. Any other exception will also be
+        evaluated at runtime for easier debugging.
+        """
+        const = node.as_const(frame.eval_ctx)
+
+        if frame.eval_ctx.autoescape:
+            const = escape(const)
+
+        # Template data doesn't go through finalize.
+        if isinstance(node, nodes.TemplateData):
+            return str(const)
+
+        return finalize.const(const)  # type: ignore
+
+    def _output_child_pre(
+        self, node: nodes.Expr, frame: Frame, finalize: _FinalizeInfo
+    ) -> None:
+        """Output extra source code before visiting a child of an
+        ``Output`` node.
+        """
+        if frame.eval_ctx.volatile:
+            self.write("(escape if context.eval_ctx.autoescape else str)(")
+        elif frame.eval_ctx.autoescape:
+            self.write("escape(")
+        else:
+            self.write("str(")
+
+        if finalize.src is not None:
+            self.write(finalize.src)
+
+    def _output_child_post(
+        self, node: nodes.Expr, frame: Frame, finalize: _FinalizeInfo
+    ) -> None:
+        """Output extra source code after visiting a child of an
+        ``Output`` node.
+        """
+        self.write(")")
+
+        if finalize.src is not None:
+            self.write(")")
+
+    def visit_Output(self, node: nodes.Output, frame: Frame) -> None:
+        # If an extends is active, don't render outside a block.
+        if frame.require_output_check:
+            # A top-level extends is known to exist at compile time.
+            if self.has_known_extends:
+                return
+
+            self.writeline("if parent_template is None:")
+            self.indent()
+
+        finalize = self._make_finalize()
+        body: t.List[t.Union[t.List[t.Any], nodes.Expr]] = []
+
+        # Evaluate constants at compile time if possible. Each item in
+        # body will be either a list of static data or a node to be
+        # evaluated at runtime.
+        for child in node.nodes:
+            try:
+                if not (
+                    # If the finalize function requires runtime context,
+                    # constants can't be evaluated at compile time.
+                    finalize.const
+                    # Unless it's basic template data that won't be
+                    # finalized anyway.
+                    or isinstance(child, nodes.TemplateData)
+                ):
+                    raise nodes.Impossible()
+
+                const = self._output_child_to_const(child, frame, finalize)
+            except (nodes.Impossible, Exception):
+                # The node was not constant and needs to be evaluated at
+                # runtime. Or another error was raised, which is easier
+                # to debug at runtime.
+                body.append(child)
+                continue
+
+            if body and isinstance(body[-1], list):
+                body[-1].append(const)
+            else:
+                body.append([const])
+
+        if frame.buffer is not None:
+            if len(body) == 1:
+                self.writeline(f"{frame.buffer}.append(")
+            else:
+                self.writeline(f"{frame.buffer}.extend((")
+
+            self.indent()
+
+        for item in body:
+            if isinstance(item, list):
+                # A group of constant data to join and output.
+                val = self._output_const_repr(item)
+
+                if frame.buffer is None:
+                    self.writeline("yield " + val)
+                else:
+                    self.writeline(val + ",")
+            else:
+                if frame.buffer is None:
+                    self.writeline("yield ", item)
+                else:
+                    self.newline(item)
+
+                # A node to be evaluated at runtime.
+                self._output_child_pre(item, frame, finalize)
+                self.visit(item, frame)
+                self._output_child_post(item, frame, finalize)
+
+                if frame.buffer is not None:
+                    self.write(",")
+
+        if frame.buffer is not None:
+            self.outdent()
+            self.writeline(")" if len(body) == 1 else "))")
+
+        if frame.require_output_check:
+            self.outdent()
+
+    def visit_Assign(self, node: nodes.Assign, frame: Frame) -> None:
+        self.push_assign_tracking()
+
+        # ``a.b`` is allowed for assignment, and is parsed as an NSRef. However,
+        # it is only valid if it references a Namespace object. Emit a check for
+        # that for each ref here, before assignment code is emitted. This can't
+        # be done in visit_NSRef as the ref could be in the middle of a tuple.
+        seen_refs: t.Set[str] = set()
+
+        for nsref in node.find_all(nodes.NSRef):
+            if nsref.name in seen_refs:
+                # Only emit the check for each reference once, in case the same
+                # ref is used multiple times in a tuple, `ns.a, ns.b = c, d`.
+                continue
+
+            seen_refs.add(nsref.name)
+            ref = frame.symbols.ref(nsref.name)
+            self.writeline(f"if not isinstance({ref}, Namespace):")
+            self.indent()
+            self.writeline(
+                "raise TemplateRuntimeError"
+                '("cannot assign attribute on non-namespace object")'
+            )
+            self.outdent()
+
+        self.newline(node)
+        self.visit(node.target, frame)
+        self.write(" = ")
+        self.visit(node.node, frame)
+        self.pop_assign_tracking(frame)
+
+    def visit_AssignBlock(self, node: nodes.AssignBlock, frame: Frame) -> None:
+        self.push_assign_tracking()
+        block_frame = frame.inner()
+        # This is a special case.  Since a set block always captures we
+        # will disable output checks.  This way one can use set blocks
+        # toplevel even in extended templates.
+        block_frame.require_output_check = False
+        block_frame.symbols.analyze_node(node)
+        self.enter_frame(block_frame)
+        self.buffer(block_frame)
+        self.blockvisit(node.body, block_frame)
+        self.newline(node)
+        self.visit(node.target, frame)
+        self.write(" = (Markup if context.eval_ctx.autoescape else identity)(")
+        if node.filter is not None:
+            self.visit_Filter(node.filter, block_frame)
+        else:
+            self.write(f"concat({block_frame.buffer})")
+        self.write(")")
+        self.pop_assign_tracking(frame)
+        self.leave_frame(block_frame)
+
+    # -- Expression Visitors
+
+    def visit_Name(self, node: nodes.Name, frame: Frame) -> None:
+        if node.ctx == "store" and (
+            frame.toplevel or frame.loop_frame or frame.block_frame
+        ):
+            if self._assign_stack:
+                self._assign_stack[-1].add(node.name)
+        ref = frame.symbols.ref(node.name)
+
+        # If we are looking up a variable we might have to deal with the
+        # case where it's undefined.  We can skip that case if the load
+        # instruction indicates a parameter which are always defined.
+        if node.ctx == "load":
+            load = frame.symbols.find_load(ref)
+            if not (
+                load is not None
+                and load[0] == VAR_LOAD_PARAMETER
+                and not self.parameter_is_undeclared(ref)
+            ):
+                self.write(
+                    f"(undefined(name={node.name!r}) if {ref} is missing else {ref})"
+                )
+                return
+
+        self.write(ref)
+
+    def visit_NSRef(self, node: nodes.NSRef, frame: Frame) -> None:
+        # NSRef is a dotted assignment target a.b=c, but uses a[b]=c internally.
+        # visit_Assign emits code to validate that each ref is to a Namespace
+        # object only. That can't be emitted here as the ref could be in the
+        # middle of a tuple assignment.
+        ref = frame.symbols.ref(node.name)
+        self.writeline(f"{ref}[{node.attr!r}]")
+
+    def visit_Const(self, node: nodes.Const, frame: Frame) -> None:
+        val = node.as_const(frame.eval_ctx)
+        if isinstance(val, float):
+            self.write(str(val))
+        else:
+            self.write(repr(val))
+
+    def visit_TemplateData(self, node: nodes.TemplateData, frame: Frame) -> None:
+        try:
+            self.write(repr(node.as_const(frame.eval_ctx)))
+        except nodes.Impossible:
+            self.write(
+                f"(Markup if context.eval_ctx.autoescape else identity)({node.data!r})"
+            )
+
+    def visit_Tuple(self, node: nodes.Tuple, frame: Frame) -> None:
+        self.write("(")
+        idx = -1
+        for idx, item in enumerate(node.items):
+            if idx:
+                self.write(", ")
+            self.visit(item, frame)
+        self.write(",)" if idx == 0 else ")")
+
+    def visit_List(self, node: nodes.List, frame: Frame) -> None:
+        self.write("[")
+        for idx, item in enumerate(node.items):
+            if idx:
+                self.write(", ")
+            self.visit(item, frame)
+        self.write("]")
+
+    def visit_Dict(self, node: nodes.Dict, frame: Frame) -> None:
+        self.write("{")
+        for idx, item in enumerate(node.items):
+            if idx:
+                self.write(", ")
+            self.visit(item.key, frame)
+            self.write(": ")
+            self.visit(item.value, frame)
+        self.write("}")
+
+    visit_Add = _make_binop("+")
+    visit_Sub = _make_binop("-")
+    visit_Mul = _make_binop("*")
+    visit_Div = _make_binop("/")
+    visit_FloorDiv = _make_binop("//")
+    visit_Pow = _make_binop("**")
+    visit_Mod = _make_binop("%")
+    visit_And = _make_binop("and")
+    visit_Or = _make_binop("or")
+    visit_Pos = _make_unop("+")
+    visit_Neg = _make_unop("-")
+    visit_Not = _make_unop("not ")
+
+    @optimizeconst
+    def visit_Concat(self, node: nodes.Concat, frame: Frame) -> None:
+        if frame.eval_ctx.volatile:
+            func_name = "(markup_join if context.eval_ctx.volatile else str_join)"
+        elif frame.eval_ctx.autoescape:
+            func_name = "markup_join"
+        else:
+            func_name = "str_join"
+        self.write(f"{func_name}((")
+        for arg in node.nodes:
+            self.visit(arg, frame)
+            self.write(", ")
+        self.write("))")
+
+    @optimizeconst
+    def visit_Compare(self, node: nodes.Compare, frame: Frame) -> None:
+        self.write("(")
+        self.visit(node.expr, frame)
+        for op in node.ops:
+            self.visit(op, frame)
+        self.write(")")
+
+    def visit_Operand(self, node: nodes.Operand, frame: Frame) -> None:
+        self.write(f" {operators[node.op]} ")
+        self.visit(node.expr, frame)
+
+    @optimizeconst
+    def visit_Getattr(self, node: nodes.Getattr, frame: Frame) -> None:
+        if self.environment.is_async:
+            self.write("(await auto_await(")
+
+        self.write("environment.getattr(")
+        self.visit(node.node, frame)
+        self.write(f", {node.attr!r})")
+
+        if self.environment.is_async:
+            self.write("))")
+
+    @optimizeconst
+    def visit_Getitem(self, node: nodes.Getitem, frame: Frame) -> None:
+        # slices bypass the environment getitem method.
+        if isinstance(node.arg, nodes.Slice):
+            self.visit(node.node, frame)
+            self.write("[")
+            self.visit(node.arg, frame)
+            self.write("]")
+        else:
+            if self.environment.is_async:
+                self.write("(await auto_await(")
+
+            self.write("environment.getitem(")
+            self.visit(node.node, frame)
+            self.write(", ")
+            self.visit(node.arg, frame)
+            self.write(")")
+
+            if self.environment.is_async:
+                self.write("))")
+
+    def visit_Slice(self, node: nodes.Slice, frame: Frame) -> None:
+        if node.start is not None:
+            self.visit(node.start, frame)
+        self.write(":")
+        if node.stop is not None:
+            self.visit(node.stop, frame)
+        if node.step is not None:
+            self.write(":")
+            self.visit(node.step, frame)
+
+    @contextmanager
+    def _filter_test_common(
+        self, node: t.Union[nodes.Filter, nodes.Test], frame: Frame, is_filter: bool
+    ) -> t.Iterator[None]:
+        if self.environment.is_async:
+            self.write("(await auto_await(")
+
+        if is_filter:
+            self.write(f"{self.filters[node.name]}(")
+            func = self.environment.filters.get(node.name)
+        else:
+            self.write(f"{self.tests[node.name]}(")
+            func = self.environment.tests.get(node.name)
+
+        # When inside an If or CondExpr frame, allow the filter to be
+        # undefined at compile time and only raise an error if it's
+        # actually called at runtime. See pull_dependencies.
+        if func is None and not frame.soft_frame:
+            type_name = "filter" if is_filter else "test"
+            self.fail(f"No {type_name} named {node.name!r}.", node.lineno)
+
+        pass_arg = {
+            _PassArg.context: "context",
+            _PassArg.eval_context: "context.eval_ctx",
+            _PassArg.environment: "environment",
+        }.get(
+            _PassArg.from_obj(func)  # type: ignore
+        )
+
+        if pass_arg is not None:
+            self.write(f"{pass_arg}, ")
+
+        # Back to the visitor function to handle visiting the target of
+        # the filter or test.
+        yield
+
+        self.signature(node, frame)
+        self.write(")")
+
+        if self.environment.is_async:
+            self.write("))")
+
+    @optimizeconst
+    def visit_Filter(self, node: nodes.Filter, frame: Frame) -> None:
+        with self._filter_test_common(node, frame, True):
+            # if the filter node is None we are inside a filter block
+            # and want to write to the current buffer
+            if node.node is not None:
+                self.visit(node.node, frame)
+            elif frame.eval_ctx.volatile:
+                self.write(
+                    f"(Markup(concat({frame.buffer}))"
+                    f" if context.eval_ctx.autoescape else concat({frame.buffer}))"
+                )
+            elif frame.eval_ctx.autoescape:
+                self.write(f"Markup(concat({frame.buffer}))")
+            else:
+                self.write(f"concat({frame.buffer})")
+
+    @optimizeconst
+    def visit_Test(self, node: nodes.Test, frame: Frame) -> None:
+        with self._filter_test_common(node, frame, False):
+            self.visit(node.node, frame)
+
+    @optimizeconst
+    def visit_CondExpr(self, node: nodes.CondExpr, frame: Frame) -> None:
+        frame = frame.soft()
+
+        def write_expr2() -> None:
+            if node.expr2 is not None:
+                self.visit(node.expr2, frame)
+                return
+
+            self.write(
+                f'cond_expr_undefined("the inline if-expression on'
+                f" {self.position(node)} evaluated to false and no else"
+                f' section was defined.")'
+            )
+
+        self.write("(")
+        self.visit(node.expr1, frame)
+        self.write(" if ")
+        self.visit(node.test, frame)
+        self.write(" else ")
+        write_expr2()
+        self.write(")")
+
+    @optimizeconst
+    def visit_Call(
+        self, node: nodes.Call, frame: Frame, forward_caller: bool = False
+    ) -> None:
+        if self.environment.is_async:
+            self.write("(await auto_await(")
+        if self.environment.sandboxed:
+            self.write("environment.call(context, ")
+        else:
+            self.write("context.call(")
+        self.visit(node.node, frame)
+        extra_kwargs = {"caller": "caller"} if forward_caller else None
+        loop_kwargs = {"_loop_vars": "_loop_vars"} if frame.loop_frame else {}
+        block_kwargs = {"_block_vars": "_block_vars"} if frame.block_frame else {}
+        if extra_kwargs:
+            extra_kwargs.update(loop_kwargs, **block_kwargs)
+        elif loop_kwargs or block_kwargs:
+            extra_kwargs = dict(loop_kwargs, **block_kwargs)
+        self.signature(node, frame, extra_kwargs)
+        self.write(")")
+        if self.environment.is_async:
+            self.write("))")
+
+    def visit_Keyword(self, node: nodes.Keyword, frame: Frame) -> None:
+        self.write(node.key + "=")
+        self.visit(node.value, frame)
+
+    # -- Unused nodes for extensions
+
+    def visit_MarkSafe(self, node: nodes.MarkSafe, frame: Frame) -> None:
+        self.write("Markup(")
+        self.visit(node.expr, frame)
+        self.write(")")
+
+    def visit_MarkSafeIfAutoescape(
+        self, node: nodes.MarkSafeIfAutoescape, frame: Frame
+    ) -> None:
+        self.write("(Markup if context.eval_ctx.autoescape else identity)(")
+        self.visit(node.expr, frame)
+        self.write(")")
+
+    def visit_EnvironmentAttribute(
+        self, node: nodes.EnvironmentAttribute, frame: Frame
+    ) -> None:
+        self.write("environment." + node.name)
+
+    def visit_ExtensionAttribute(
+        self, node: nodes.ExtensionAttribute, frame: Frame
+    ) -> None:
+        self.write(f"environment.extensions[{node.identifier!r}].{node.name}")
+
+    def visit_ImportedName(self, node: nodes.ImportedName, frame: Frame) -> None:
+        self.write(self.import_aliases[node.importname])
+
+    def visit_InternalName(self, node: nodes.InternalName, frame: Frame) -> None:
+        self.write(node.name)
+
+    def visit_ContextReference(
+        self, node: nodes.ContextReference, frame: Frame
+    ) -> None:
+        self.write("context")
+
+    def visit_DerivedContextReference(
+        self, node: nodes.DerivedContextReference, frame: Frame
+    ) -> None:
+        self.write(self.derive_context(frame))
+
+    def visit_Continue(self, node: nodes.Continue, frame: Frame) -> None:
+        self.writeline("continue", node)
+
+    def visit_Break(self, node: nodes.Break, frame: Frame) -> None:
+        self.writeline("break", node)
+
+    def visit_Scope(self, node: nodes.Scope, frame: Frame) -> None:
+        scope_frame = frame.inner()
+        scope_frame.symbols.analyze_node(node)
+        self.enter_frame(scope_frame)
+        self.blockvisit(node.body, scope_frame)
+        self.leave_frame(scope_frame)
+
+    def visit_OverlayScope(self, node: nodes.OverlayScope, frame: Frame) -> None:
+        ctx = self.temporary_identifier()
+        self.writeline(f"{ctx} = {self.derive_context(frame)}")
+        self.writeline(f"{ctx}.vars = ")
+        self.visit(node.context, frame)
+        self.push_context_reference(ctx)
+
+        scope_frame = frame.inner(isolated=True)
+        scope_frame.symbols.analyze_node(node)
+        self.enter_frame(scope_frame)
+        self.blockvisit(node.body, scope_frame)
+        self.leave_frame(scope_frame)
+        self.pop_context_reference()
+
+    def visit_EvalContextModifier(
+        self, node: nodes.EvalContextModifier, frame: Frame
+    ) -> None:
+        for keyword in node.options:
+            self.writeline(f"context.eval_ctx.{keyword.key} = ")
+            self.visit(keyword.value, frame)
+            try:
+                val = keyword.value.as_const(frame.eval_ctx)
+            except nodes.Impossible:
+                frame.eval_ctx.volatile = True
+            else:
+                setattr(frame.eval_ctx, keyword.key, val)
+
+    def visit_ScopedEvalContextModifier(
+        self, node: nodes.ScopedEvalContextModifier, frame: Frame
+    ) -> None:
+        old_ctx_name = self.temporary_identifier()
+        saved_ctx = frame.eval_ctx.save()
+        self.writeline(f"{old_ctx_name} = context.eval_ctx.save()")
+        self.visit_EvalContextModifier(node, frame)
+        for child in node.body:
+            self.visit(child, frame)
+        frame.eval_ctx.revert(saved_ctx)
+        self.writeline(f"context.eval_ctx.revert({old_ctx_name})")
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/constants.py b/backend/venv/lib/python3.12/site-packages/jinja2/constants.py
new file mode 100644
index 00000000..41a1c23b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/constants.py
@@ -0,0 +1,20 @@
+#: list of lorem ipsum words used by the lipsum() helper function
+LOREM_IPSUM_WORDS = """\
+a ac accumsan ad adipiscing aenean aliquam aliquet amet ante aptent arcu at
+auctor augue bibendum blandit class commodo condimentum congue consectetuer
+consequat conubia convallis cras cubilia cum curabitur curae cursus dapibus
+diam dictum dictumst dignissim dis dolor donec dui duis egestas eget eleifend
+elementum elit enim erat eros est et etiam eu euismod facilisi facilisis fames
+faucibus felis fermentum feugiat fringilla fusce gravida habitant habitasse hac
+hendrerit hymenaeos iaculis id imperdiet in inceptos integer interdum ipsum
+justo lacinia lacus laoreet lectus leo libero ligula litora lobortis lorem
+luctus maecenas magna magnis malesuada massa mattis mauris metus mi molestie
+mollis montes morbi mus nam nascetur natoque nec neque netus nibh nisi nisl non
+nonummy nostra nulla nullam nunc odio orci ornare parturient pede pellentesque
+penatibus per pharetra phasellus placerat platea porta porttitor posuere
+potenti praesent pretium primis proin pulvinar purus quam quis quisque rhoncus
+ridiculus risus rutrum sagittis sapien scelerisque sed sem semper senectus sit
+sociis sociosqu sodales sollicitudin suscipit suspendisse taciti tellus tempor
+tempus tincidunt torquent tortor tristique turpis ullamcorper ultrices
+ultricies urna ut varius vehicula vel velit venenatis vestibulum vitae vivamus
+viverra volutpat vulputate"""
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/debug.py b/backend/venv/lib/python3.12/site-packages/jinja2/debug.py
new file mode 100644
index 00000000..eeeeee78
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/debug.py
@@ -0,0 +1,191 @@
+import sys
+import typing as t
+from types import CodeType
+from types import TracebackType
+
+from .exceptions import TemplateSyntaxError
+from .utils import internal_code
+from .utils import missing
+
+if t.TYPE_CHECKING:
+    from .runtime import Context
+
+
+def rewrite_traceback_stack(source: t.Optional[str] = None) -> BaseException:
+    """Rewrite the current exception to replace any tracebacks from
+    within compiled template code with tracebacks that look like they
+    came from the template source.
+
+    This must be called within an ``except`` block.
+
+    :param source: For ``TemplateSyntaxError``, the original source if
+        known.
+    :return: The original exception with the rewritten traceback.
+    """
+    _, exc_value, tb = sys.exc_info()
+    exc_value = t.cast(BaseException, exc_value)
+    tb = t.cast(TracebackType, tb)
+
+    if isinstance(exc_value, TemplateSyntaxError) and not exc_value.translated:
+        exc_value.translated = True
+        exc_value.source = source
+        # Remove the old traceback, otherwise the frames from the
+        # compiler still show up.
+        exc_value.with_traceback(None)
+        # Outside of runtime, so the frame isn't executing template
+        # code, but it still needs to point at the template.
+        tb = fake_traceback(
+            exc_value, None, exc_value.filename or "<unknown>", exc_value.lineno
+        )
+    else:
+        # Skip the frame for the render function.
+        tb = tb.tb_next
+
+    stack = []
+
+    # Build the stack of traceback object, replacing any in template
+    # code with the source file and line information.
+    while tb is not None:
+        # Skip frames decorated with @internalcode. These are internal
+        # calls that aren't useful in template debugging output.
+        if tb.tb_frame.f_code in internal_code:
+            tb = tb.tb_next
+            continue
+
+        template = tb.tb_frame.f_globals.get("__jinja_template__")
+
+        if template is not None:
+            lineno = template.get_corresponding_lineno(tb.tb_lineno)
+            fake_tb = fake_traceback(exc_value, tb, template.filename, lineno)
+            stack.append(fake_tb)
+        else:
+            stack.append(tb)
+
+        tb = tb.tb_next
+
+    tb_next = None
+
+    # Assign tb_next in reverse to avoid circular references.
+    for tb in reversed(stack):
+        tb.tb_next = tb_next
+        tb_next = tb
+
+    return exc_value.with_traceback(tb_next)
+
+
+def fake_traceback(  # type: ignore
+    exc_value: BaseException, tb: t.Optional[TracebackType], filename: str, lineno: int
+) -> TracebackType:
+    """Produce a new traceback object that looks like it came from the
+    template source instead of the compiled code. The filename, line
+    number, and location name will point to the template, and the local
+    variables will be the current template context.
+
+    :param exc_value: The original exception to be re-raised to create
+        the new traceback.
+    :param tb: The original traceback to get the local variables and
+        code info from.
+    :param filename: The template filename.
+    :param lineno: The line number in the template source.
+    """
+    if tb is not None:
+        # Replace the real locals with the context that would be
+        # available at that point in the template.
+        locals = get_template_locals(tb.tb_frame.f_locals)
+        locals.pop("__jinja_exception__", None)
+    else:
+        locals = {}
+
+    globals = {
+        "__name__": filename,
+        "__file__": filename,
+        "__jinja_exception__": exc_value,
+    }
+    # Raise an exception at the correct line number.
+    code: CodeType = compile(
+        "\n" * (lineno - 1) + "raise __jinja_exception__", filename, "exec"
+    )
+
+    # Build a new code object that points to the template file and
+    # replaces the location with a block name.
+    location = "template"
+
+    if tb is not None:
+        function = tb.tb_frame.f_code.co_name
+
+        if function == "root":
+            location = "top-level template code"
+        elif function.startswith("block_"):
+            location = f"block {function[6:]!r}"
+
+    if sys.version_info >= (3, 8):
+        code = code.replace(co_name=location)
+    else:
+        code = CodeType(
+            code.co_argcount,
+            code.co_kwonlyargcount,
+            code.co_nlocals,
+            code.co_stacksize,
+            code.co_flags,
+            code.co_code,
+            code.co_consts,
+            code.co_names,
+            code.co_varnames,
+            code.co_filename,
+            location,
+            code.co_firstlineno,
+            code.co_lnotab,
+            code.co_freevars,
+            code.co_cellvars,
+        )
+
+    # Execute the new code, which is guaranteed to raise, and return
+    # the new traceback without this frame.
+    try:
+        exec(code, globals, locals)
+    except BaseException:
+        return sys.exc_info()[2].tb_next  # type: ignore
+
+
+def get_template_locals(real_locals: t.Mapping[str, t.Any]) -> t.Dict[str, t.Any]:
+    """Based on the runtime locals, get the context that would be
+    available at that point in the template.
+    """
+    # Start with the current template context.
+    ctx: t.Optional[Context] = real_locals.get("context")
+
+    if ctx is not None:
+        data: t.Dict[str, t.Any] = ctx.get_all().copy()
+    else:
+        data = {}
+
+    # Might be in a derived context that only sets local variables
+    # rather than pushing a context. Local variables follow the scheme
+    # l_depth_name. Find the highest-depth local that has a value for
+    # each name.
+    local_overrides: t.Dict[str, t.Tuple[int, t.Any]] = {}
+
+    for name, value in real_locals.items():
+        if not name.startswith("l_") or value is missing:
+            # Not a template variable, or no longer relevant.
+            continue
+
+        try:
+            _, depth_str, name = name.split("_", 2)
+            depth = int(depth_str)
+        except ValueError:
+            continue
+
+        cur_depth = local_overrides.get(name, (-1,))[0]
+
+        if cur_depth < depth:
+            local_overrides[name] = (depth, value)
+
+    # Modify the context with any derived context.
+    for name, (_, value) in local_overrides.items():
+        if value is missing:
+            data.pop(name, None)
+        else:
+            data[name] = value
+
+    return data
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/defaults.py b/backend/venv/lib/python3.12/site-packages/jinja2/defaults.py
new file mode 100644
index 00000000..638cad3d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/defaults.py
@@ -0,0 +1,48 @@
+import typing as t
+
+from .filters import FILTERS as DEFAULT_FILTERS  # noqa: F401
+from .tests import TESTS as DEFAULT_TESTS  # noqa: F401
+from .utils import Cycler
+from .utils import generate_lorem_ipsum
+from .utils import Joiner
+from .utils import Namespace
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+# defaults for the parser / lexer
+BLOCK_START_STRING = "{%"
+BLOCK_END_STRING = "%}"
+VARIABLE_START_STRING = "{{"
+VARIABLE_END_STRING = "}}"
+COMMENT_START_STRING = "{#"
+COMMENT_END_STRING = "#}"
+LINE_STATEMENT_PREFIX: t.Optional[str] = None
+LINE_COMMENT_PREFIX: t.Optional[str] = None
+TRIM_BLOCKS = False
+LSTRIP_BLOCKS = False
+NEWLINE_SEQUENCE: "te.Literal['\\n', '\\r\\n', '\\r']" = "\n"
+KEEP_TRAILING_NEWLINE = False
+
+# default filters, tests and namespace
+
+DEFAULT_NAMESPACE = {
+    "range": range,
+    "dict": dict,
+    "lipsum": generate_lorem_ipsum,
+    "cycler": Cycler,
+    "joiner": Joiner,
+    "namespace": Namespace,
+}
+
+# default policies
+DEFAULT_POLICIES: t.Dict[str, t.Any] = {
+    "compiler.ascii_str": True,
+    "urlize.rel": "noopener",
+    "urlize.target": None,
+    "urlize.extra_schemes": None,
+    "truncate.leeway": 5,
+    "json.dumps_function": None,
+    "json.dumps_kwargs": {"sort_keys": True},
+    "ext.i18n.trimmed": False,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/environment.py b/backend/venv/lib/python3.12/site-packages/jinja2/environment.py
new file mode 100644
index 00000000..0fc6e5be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/environment.py
@@ -0,0 +1,1672 @@
+"""Classes for managing templates and their runtime and compile time
+options.
+"""
+
+import os
+import typing
+import typing as t
+import weakref
+from collections import ChainMap
+from functools import lru_cache
+from functools import partial
+from functools import reduce
+from types import CodeType
+
+from markupsafe import Markup
+
+from . import nodes
+from .compiler import CodeGenerator
+from .compiler import generate
+from .defaults import BLOCK_END_STRING
+from .defaults import BLOCK_START_STRING
+from .defaults import COMMENT_END_STRING
+from .defaults import COMMENT_START_STRING
+from .defaults import DEFAULT_FILTERS  # type: ignore[attr-defined]
+from .defaults import DEFAULT_NAMESPACE
+from .defaults import DEFAULT_POLICIES
+from .defaults import DEFAULT_TESTS  # type: ignore[attr-defined]
+from .defaults import KEEP_TRAILING_NEWLINE
+from .defaults import LINE_COMMENT_PREFIX
+from .defaults import LINE_STATEMENT_PREFIX
+from .defaults import LSTRIP_BLOCKS
+from .defaults import NEWLINE_SEQUENCE
+from .defaults import TRIM_BLOCKS
+from .defaults import VARIABLE_END_STRING
+from .defaults import VARIABLE_START_STRING
+from .exceptions import TemplateNotFound
+from .exceptions import TemplateRuntimeError
+from .exceptions import TemplatesNotFound
+from .exceptions import TemplateSyntaxError
+from .exceptions import UndefinedError
+from .lexer import get_lexer
+from .lexer import Lexer
+from .lexer import TokenStream
+from .nodes import EvalContext
+from .parser import Parser
+from .runtime import Context
+from .runtime import new_context
+from .runtime import Undefined
+from .utils import _PassArg
+from .utils import concat
+from .utils import consume
+from .utils import import_string
+from .utils import internalcode
+from .utils import LRUCache
+from .utils import missing
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .bccache import BytecodeCache
+    from .ext import Extension
+    from .loaders import BaseLoader
+
+_env_bound = t.TypeVar("_env_bound", bound="Environment")
+
+
+# for direct template usage we have up to ten living environments
+@lru_cache(maxsize=10)
+def get_spontaneous_environment(cls: t.Type[_env_bound], *args: t.Any) -> _env_bound:
+    """Return a new spontaneous environment. A spontaneous environment
+    is used for templates created directly rather than through an
+    existing environment.
+
+    :param cls: Environment class to create.
+    :param args: Positional arguments passed to environment.
+    """
+    env = cls(*args)
+    env.shared = True
+    return env
+
+
+def create_cache(
+    size: int,
+) -> t.Optional[t.MutableMapping[t.Tuple["weakref.ref[t.Any]", str], "Template"]]:
+    """Return the cache class for the given size."""
+    if size == 0:
+        return None
+
+    if size < 0:
+        return {}
+
+    return LRUCache(size)  # type: ignore
+
+
+def copy_cache(
+    cache: t.Optional[t.MutableMapping[t.Any, t.Any]],
+) -> t.Optional[t.MutableMapping[t.Tuple["weakref.ref[t.Any]", str], "Template"]]:
+    """Create an empty copy of the given cache."""
+    if cache is None:
+        return None
+
+    if type(cache) is dict:  # noqa E721
+        return {}
+
+    return LRUCache(cache.capacity)  # type: ignore
+
+
+def load_extensions(
+    environment: "Environment",
+    extensions: t.Sequence[t.Union[str, t.Type["Extension"]]],
+) -> t.Dict[str, "Extension"]:
+    """Load the extensions from the list and bind it to the environment.
+    Returns a dict of instantiated extensions.
+    """
+    result = {}
+
+    for extension in extensions:
+        if isinstance(extension, str):
+            extension = t.cast(t.Type["Extension"], import_string(extension))
+
+        result[extension.identifier] = extension(environment)
+
+    return result
+
+
+def _environment_config_check(environment: _env_bound) -> _env_bound:
+    """Perform a sanity check on the environment."""
+    assert issubclass(
+        environment.undefined, Undefined
+    ), "'undefined' must be a subclass of 'jinja2.Undefined'."
+    assert (
+        environment.block_start_string
+        != environment.variable_start_string
+        != environment.comment_start_string
+    ), "block, variable and comment start strings must be different."
+    assert environment.newline_sequence in {
+        "\r",
+        "\r\n",
+        "\n",
+    }, "'newline_sequence' must be one of '\\n', '\\r\\n', or '\\r'."
+    return environment
+
+
+class Environment:
+    r"""The core component of Jinja is the `Environment`.  It contains
+    important shared variables like configuration, filters, tests,
+    globals and others.  Instances of this class may be modified if
+    they are not shared and if no template was loaded so far.
+    Modifications on environments after the first template was loaded
+    will lead to surprising effects and undefined behavior.
+
+    Here are the possible initialization parameters:
+
+        `block_start_string`
+            The string marking the beginning of a block.  Defaults to ``'{%'``.
+
+        `block_end_string`
+            The string marking the end of a block.  Defaults to ``'%}'``.
+
+        `variable_start_string`
+            The string marking the beginning of a print statement.
+            Defaults to ``'{{'``.
+
+        `variable_end_string`
+            The string marking the end of a print statement.  Defaults to
+            ``'}}'``.
+
+        `comment_start_string`
+            The string marking the beginning of a comment.  Defaults to ``'{#'``.
+
+        `comment_end_string`
+            The string marking the end of a comment.  Defaults to ``'#}'``.
+
+        `line_statement_prefix`
+            If given and a string, this will be used as prefix for line based
+            statements.  See also :ref:`line-statements`.
+
+        `line_comment_prefix`
+            If given and a string, this will be used as prefix for line based
+            comments.  See also :ref:`line-statements`.
+
+            .. versionadded:: 2.2
+
+        `trim_blocks`
+            If this is set to ``True`` the first newline after a block is
+            removed (block, not variable tag!).  Defaults to `False`.
+
+        `lstrip_blocks`
+            If this is set to ``True`` leading spaces and tabs are stripped
+            from the start of a line to a block.  Defaults to `False`.
+
+        `newline_sequence`
+            The sequence that starts a newline.  Must be one of ``'\r'``,
+            ``'\n'`` or ``'\r\n'``.  The default is ``'\n'`` which is a
+            useful default for Linux and OS X systems as well as web
+            applications.
+
+        `keep_trailing_newline`
+            Preserve the trailing newline when rendering templates.
+            The default is ``False``, which causes a single newline,
+            if present, to be stripped from the end of the template.
+
+            .. versionadded:: 2.7
+
+        `extensions`
+            List of Jinja extensions to use.  This can either be import paths
+            as strings or extension classes.  For more information have a
+            look at :ref:`the extensions documentation <jinja-extensions>`.
+
+        `optimized`
+            should the optimizer be enabled?  Default is ``True``.
+
+        `undefined`
+            :class:`Undefined` or a subclass of it that is used to represent
+            undefined values in the template.
+
+        `finalize`
+            A callable that can be used to process the result of a variable
+            expression before it is output.  For example one can convert
+            ``None`` implicitly into an empty string here.
+
+        `autoescape`
+            If set to ``True`` the XML/HTML autoescaping feature is enabled by
+            default.  For more details about autoescaping see
+            :class:`~markupsafe.Markup`.  As of Jinja 2.4 this can also
+            be a callable that is passed the template name and has to
+            return ``True`` or ``False`` depending on autoescape should be
+            enabled by default.
+
+            .. versionchanged:: 2.4
+               `autoescape` can now be a function
+
+        `loader`
+            The template loader for this environment.
+
+        `cache_size`
+            The size of the cache.  Per default this is ``400`` which means
+            that if more than 400 templates are loaded the loader will clean
+            out the least recently used template.  If the cache size is set to
+            ``0`` templates are recompiled all the time, if the cache size is
+            ``-1`` the cache will not be cleaned.
+
+            .. versionchanged:: 2.8
+               The cache size was increased to 400 from a low 50.
+
+        `auto_reload`
+            Some loaders load templates from locations where the template
+            sources may change (ie: file system or database).  If
+            ``auto_reload`` is set to ``True`` (default) every time a template is
+            requested the loader checks if the source changed and if yes, it
+            will reload the template.  For higher performance it's possible to
+            disable that.
+
+        `bytecode_cache`
+            If set to a bytecode cache object, this object will provide a
+            cache for the internal Jinja bytecode so that templates don't
+            have to be parsed if they were not changed.
+
+            See :ref:`bytecode-cache` for more information.
+
+        `enable_async`
+            If set to true this enables async template execution which
+            allows using async functions and generators.
+    """
+
+    #: if this environment is sandboxed.  Modifying this variable won't make
+    #: the environment sandboxed though.  For a real sandboxed environment
+    #: have a look at jinja2.sandbox.  This flag alone controls the code
+    #: generation by the compiler.
+    sandboxed = False
+
+    #: True if the environment is just an overlay
+    overlayed = False
+
+    #: the environment this environment is linked to if it is an overlay
+    linked_to: t.Optional["Environment"] = None
+
+    #: shared environments have this set to `True`.  A shared environment
+    #: must not be modified
+    shared = False
+
+    #: the class that is used for code generation.  See
+    #: :class:`~jinja2.compiler.CodeGenerator` for more information.
+    code_generator_class: t.Type["CodeGenerator"] = CodeGenerator
+
+    concat = "".join
+
+    #: the context class that is used for templates.  See
+    #: :class:`~jinja2.runtime.Context` for more information.
+    context_class: t.Type[Context] = Context
+
+    template_class: t.Type["Template"]
+
+    def __init__(
+        self,
+        block_start_string: str = BLOCK_START_STRING,
+        block_end_string: str = BLOCK_END_STRING,
+        variable_start_string: str = VARIABLE_START_STRING,
+        variable_end_string: str = VARIABLE_END_STRING,
+        comment_start_string: str = COMMENT_START_STRING,
+        comment_end_string: str = COMMENT_END_STRING,
+        line_statement_prefix: t.Optional[str] = LINE_STATEMENT_PREFIX,
+        line_comment_prefix: t.Optional[str] = LINE_COMMENT_PREFIX,
+        trim_blocks: bool = TRIM_BLOCKS,
+        lstrip_blocks: bool = LSTRIP_BLOCKS,
+        newline_sequence: "te.Literal['\\n', '\\r\\n', '\\r']" = NEWLINE_SEQUENCE,
+        keep_trailing_newline: bool = KEEP_TRAILING_NEWLINE,
+        extensions: t.Sequence[t.Union[str, t.Type["Extension"]]] = (),
+        optimized: bool = True,
+        undefined: t.Type[Undefined] = Undefined,
+        finalize: t.Optional[t.Callable[..., t.Any]] = None,
+        autoescape: t.Union[bool, t.Callable[[t.Optional[str]], bool]] = False,
+        loader: t.Optional["BaseLoader"] = None,
+        cache_size: int = 400,
+        auto_reload: bool = True,
+        bytecode_cache: t.Optional["BytecodeCache"] = None,
+        enable_async: bool = False,
+    ):
+        # !!Important notice!!
+        #   The constructor accepts quite a few arguments that should be
+        #   passed by keyword rather than position.  However it's important to
+        #   not change the order of arguments because it's used at least
+        #   internally in those cases:
+        #       -   spontaneous environments (i18n extension and Template)
+        #       -   unittests
+        #   If parameter changes are required only add parameters at the end
+        #   and don't change the arguments (or the defaults!) of the arguments
+        #   existing already.
+
+        # lexer / parser information
+        self.block_start_string = block_start_string
+        self.block_end_string = block_end_string
+        self.variable_start_string = variable_start_string
+        self.variable_end_string = variable_end_string
+        self.comment_start_string = comment_start_string
+        self.comment_end_string = comment_end_string
+        self.line_statement_prefix = line_statement_prefix
+        self.line_comment_prefix = line_comment_prefix
+        self.trim_blocks = trim_blocks
+        self.lstrip_blocks = lstrip_blocks
+        self.newline_sequence = newline_sequence
+        self.keep_trailing_newline = keep_trailing_newline
+
+        # runtime information
+        self.undefined: t.Type[Undefined] = undefined
+        self.optimized = optimized
+        self.finalize = finalize
+        self.autoescape = autoescape
+
+        # defaults
+        self.filters = DEFAULT_FILTERS.copy()
+        self.tests = DEFAULT_TESTS.copy()
+        self.globals = DEFAULT_NAMESPACE.copy()
+
+        # set the loader provided
+        self.loader = loader
+        self.cache = create_cache(cache_size)
+        self.bytecode_cache = bytecode_cache
+        self.auto_reload = auto_reload
+
+        # configurable policies
+        self.policies = DEFAULT_POLICIES.copy()
+
+        # load extensions
+        self.extensions = load_extensions(self, extensions)
+
+        self.is_async = enable_async
+        _environment_config_check(self)
+
+    def add_extension(self, extension: t.Union[str, t.Type["Extension"]]) -> None:
+        """Adds an extension after the environment was created.
+
+        .. versionadded:: 2.5
+        """
+        self.extensions.update(load_extensions(self, [extension]))
+
+    def extend(self, **attributes: t.Any) -> None:
+        """Add the items to the instance of the environment if they do not exist
+        yet.  This is used by :ref:`extensions <writing-extensions>` to register
+        callbacks and configuration values without breaking inheritance.
+        """
+        for key, value in attributes.items():
+            if not hasattr(self, key):
+                setattr(self, key, value)
+
+    def overlay(
+        self,
+        block_start_string: str = missing,
+        block_end_string: str = missing,
+        variable_start_string: str = missing,
+        variable_end_string: str = missing,
+        comment_start_string: str = missing,
+        comment_end_string: str = missing,
+        line_statement_prefix: t.Optional[str] = missing,
+        line_comment_prefix: t.Optional[str] = missing,
+        trim_blocks: bool = missing,
+        lstrip_blocks: bool = missing,
+        newline_sequence: "te.Literal['\\n', '\\r\\n', '\\r']" = missing,
+        keep_trailing_newline: bool = missing,
+        extensions: t.Sequence[t.Union[str, t.Type["Extension"]]] = missing,
+        optimized: bool = missing,
+        undefined: t.Type[Undefined] = missing,
+        finalize: t.Optional[t.Callable[..., t.Any]] = missing,
+        autoescape: t.Union[bool, t.Callable[[t.Optional[str]], bool]] = missing,
+        loader: t.Optional["BaseLoader"] = missing,
+        cache_size: int = missing,
+        auto_reload: bool = missing,
+        bytecode_cache: t.Optional["BytecodeCache"] = missing,
+        enable_async: bool = missing,
+    ) -> "te.Self":
+        """Create a new overlay environment that shares all the data with the
+        current environment except for cache and the overridden attributes.
+        Extensions cannot be removed for an overlayed environment.  An overlayed
+        environment automatically gets all the extensions of the environment it
+        is linked to plus optional extra extensions.
+
+        Creating overlays should happen after the initial environment was set
+        up completely.  Not all attributes are truly linked, some are just
+        copied over so modifications on the original environment may not shine
+        through.
+
+        .. versionchanged:: 3.1.5
+            ``enable_async`` is applied correctly.
+
+        .. versionchanged:: 3.1.2
+            Added the ``newline_sequence``, ``keep_trailing_newline``,
+            and ``enable_async`` parameters to match ``__init__``.
+        """
+        args = dict(locals())
+        del args["self"], args["cache_size"], args["extensions"], args["enable_async"]
+
+        rv = object.__new__(self.__class__)
+        rv.__dict__.update(self.__dict__)
+        rv.overlayed = True
+        rv.linked_to = self
+
+        for key, value in args.items():
+            if value is not missing:
+                setattr(rv, key, value)
+
+        if cache_size is not missing:
+            rv.cache = create_cache(cache_size)
+        else:
+            rv.cache = copy_cache(self.cache)
+
+        rv.extensions = {}
+        for key, value in self.extensions.items():
+            rv.extensions[key] = value.bind(rv)
+        if extensions is not missing:
+            rv.extensions.update(load_extensions(rv, extensions))
+
+        if enable_async is not missing:
+            rv.is_async = enable_async
+
+        return _environment_config_check(rv)
+
+    @property
+    def lexer(self) -> Lexer:
+        """The lexer for this environment."""
+        return get_lexer(self)
+
+    def iter_extensions(self) -> t.Iterator["Extension"]:
+        """Iterates over the extensions by priority."""
+        return iter(sorted(self.extensions.values(), key=lambda x: x.priority))
+
+    def getitem(
+        self, obj: t.Any, argument: t.Union[str, t.Any]
+    ) -> t.Union[t.Any, Undefined]:
+        """Get an item or attribute of an object but prefer the item."""
+        try:
+            return obj[argument]
+        except (AttributeError, TypeError, LookupError):
+            if isinstance(argument, str):
+                try:
+                    attr = str(argument)
+                except Exception:
+                    pass
+                else:
+                    try:
+                        return getattr(obj, attr)
+                    except AttributeError:
+                        pass
+            return self.undefined(obj=obj, name=argument)
+
+    def getattr(self, obj: t.Any, attribute: str) -> t.Any:
+        """Get an item or attribute of an object but prefer the attribute.
+        Unlike :meth:`getitem` the attribute *must* be a string.
+        """
+        try:
+            return getattr(obj, attribute)
+        except AttributeError:
+            pass
+        try:
+            return obj[attribute]
+        except (TypeError, LookupError, AttributeError):
+            return self.undefined(obj=obj, name=attribute)
+
+    def _filter_test_common(
+        self,
+        name: t.Union[str, Undefined],
+        value: t.Any,
+        args: t.Optional[t.Sequence[t.Any]],
+        kwargs: t.Optional[t.Mapping[str, t.Any]],
+        context: t.Optional[Context],
+        eval_ctx: t.Optional[EvalContext],
+        is_filter: bool,
+    ) -> t.Any:
+        if is_filter:
+            env_map = self.filters
+            type_name = "filter"
+        else:
+            env_map = self.tests
+            type_name = "test"
+
+        func = env_map.get(name)  # type: ignore
+
+        if func is None:
+            msg = f"No {type_name} named {name!r}."
+
+            if isinstance(name, Undefined):
+                try:
+                    name._fail_with_undefined_error()
+                except Exception as e:
+                    msg = f"{msg} ({e}; did you forget to quote the callable name?)"
+
+            raise TemplateRuntimeError(msg)
+
+        args = [value, *(args if args is not None else ())]
+        kwargs = kwargs if kwargs is not None else {}
+        pass_arg = _PassArg.from_obj(func)
+
+        if pass_arg is _PassArg.context:
+            if context is None:
+                raise TemplateRuntimeError(
+                    f"Attempted to invoke a context {type_name} without context."
+                )
+
+            args.insert(0, context)
+        elif pass_arg is _PassArg.eval_context:
+            if eval_ctx is None:
+                if context is not None:
+                    eval_ctx = context.eval_ctx
+                else:
+                    eval_ctx = EvalContext(self)
+
+            args.insert(0, eval_ctx)
+        elif pass_arg is _PassArg.environment:
+            args.insert(0, self)
+
+        return func(*args, **kwargs)
+
+    def call_filter(
+        self,
+        name: str,
+        value: t.Any,
+        args: t.Optional[t.Sequence[t.Any]] = None,
+        kwargs: t.Optional[t.Mapping[str, t.Any]] = None,
+        context: t.Optional[Context] = None,
+        eval_ctx: t.Optional[EvalContext] = None,
+    ) -> t.Any:
+        """Invoke a filter on a value the same way the compiler does.
+
+        This might return a coroutine if the filter is running from an
+        environment in async mode and the filter supports async
+        execution. It's your responsibility to await this if needed.
+
+        .. versionadded:: 2.7
+        """
+        return self._filter_test_common(
+            name, value, args, kwargs, context, eval_ctx, True
+        )
+
+    def call_test(
+        self,
+        name: str,
+        value: t.Any,
+        args: t.Optional[t.Sequence[t.Any]] = None,
+        kwargs: t.Optional[t.Mapping[str, t.Any]] = None,
+        context: t.Optional[Context] = None,
+        eval_ctx: t.Optional[EvalContext] = None,
+    ) -> t.Any:
+        """Invoke a test on a value the same way the compiler does.
+
+        This might return a coroutine if the test is running from an
+        environment in async mode and the test supports async execution.
+        It's your responsibility to await this if needed.
+
+        .. versionchanged:: 3.0
+            Tests support ``@pass_context``, etc. decorators. Added
+            the ``context`` and ``eval_ctx`` parameters.
+
+        .. versionadded:: 2.7
+        """
+        return self._filter_test_common(
+            name, value, args, kwargs, context, eval_ctx, False
+        )
+
+    @internalcode
+    def parse(
+        self,
+        source: str,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+    ) -> nodes.Template:
+        """Parse the sourcecode and return the abstract syntax tree.  This
+        tree of nodes is used by the compiler to convert the template into
+        executable source- or bytecode.  This is useful for debugging or to
+        extract information from templates.
+
+        If you are :ref:`developing Jinja extensions <writing-extensions>`
+        this gives you a good overview of the node tree generated.
+        """
+        try:
+            return self._parse(source, name, filename)
+        except TemplateSyntaxError:
+            self.handle_exception(source=source)
+
+    def _parse(
+        self, source: str, name: t.Optional[str], filename: t.Optional[str]
+    ) -> nodes.Template:
+        """Internal parsing function used by `parse` and `compile`."""
+        return Parser(self, source, name, filename).parse()
+
+    def lex(
+        self,
+        source: str,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+    ) -> t.Iterator[t.Tuple[int, str, str]]:
+        """Lex the given sourcecode and return a generator that yields
+        tokens as tuples in the form ``(lineno, token_type, value)``.
+        This can be useful for :ref:`extension development <writing-extensions>`
+        and debugging templates.
+
+        This does not perform preprocessing.  If you want the preprocessing
+        of the extensions to be applied you have to filter source through
+        the :meth:`preprocess` method.
+        """
+        source = str(source)
+        try:
+            return self.lexer.tokeniter(source, name, filename)
+        except TemplateSyntaxError:
+            self.handle_exception(source=source)
+
+    def preprocess(
+        self,
+        source: str,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+    ) -> str:
+        """Preprocesses the source with all extensions.  This is automatically
+        called for all parsing and compiling methods but *not* for :meth:`lex`
+        because there you usually only want the actual source tokenized.
+        """
+        return reduce(
+            lambda s, e: e.preprocess(s, name, filename),
+            self.iter_extensions(),
+            str(source),
+        )
+
+    def _tokenize(
+        self,
+        source: str,
+        name: t.Optional[str],
+        filename: t.Optional[str] = None,
+        state: t.Optional[str] = None,
+    ) -> TokenStream:
+        """Called by the parser to do the preprocessing and filtering
+        for all the extensions.  Returns a :class:`~jinja2.lexer.TokenStream`.
+        """
+        source = self.preprocess(source, name, filename)
+        stream = self.lexer.tokenize(source, name, filename, state)
+
+        for ext in self.iter_extensions():
+            stream = ext.filter_stream(stream)  # type: ignore
+
+            if not isinstance(stream, TokenStream):
+                stream = TokenStream(stream, name, filename)
+
+        return stream
+
+    def _generate(
+        self,
+        source: nodes.Template,
+        name: t.Optional[str],
+        filename: t.Optional[str],
+        defer_init: bool = False,
+    ) -> str:
+        """Internal hook that can be overridden to hook a different generate
+        method in.
+
+        .. versionadded:: 2.5
+        """
+        return generate(  # type: ignore
+            source,
+            self,
+            name,
+            filename,
+            defer_init=defer_init,
+            optimized=self.optimized,
+        )
+
+    def _compile(self, source: str, filename: str) -> CodeType:
+        """Internal hook that can be overridden to hook a different compile
+        method in.
+
+        .. versionadded:: 2.5
+        """
+        return compile(source, filename, "exec")
+
+    @typing.overload
+    def compile(
+        self,
+        source: t.Union[str, nodes.Template],
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+        raw: "te.Literal[False]" = False,
+        defer_init: bool = False,
+    ) -> CodeType: ...
+
+    @typing.overload
+    def compile(
+        self,
+        source: t.Union[str, nodes.Template],
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+        raw: "te.Literal[True]" = ...,
+        defer_init: bool = False,
+    ) -> str: ...
+
+    @internalcode
+    def compile(
+        self,
+        source: t.Union[str, nodes.Template],
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+        raw: bool = False,
+        defer_init: bool = False,
+    ) -> t.Union[str, CodeType]:
+        """Compile a node or template source code.  The `name` parameter is
+        the load name of the template after it was joined using
+        :meth:`join_path` if necessary, not the filename on the file system.
+        the `filename` parameter is the estimated filename of the template on
+        the file system.  If the template came from a database or memory this
+        can be omitted.
+
+        The return value of this method is a python code object.  If the `raw`
+        parameter is `True` the return value will be a string with python
+        code equivalent to the bytecode returned otherwise.  This method is
+        mainly used internally.
+
+        `defer_init` is use internally to aid the module code generator.  This
+        causes the generated code to be able to import without the global
+        environment variable to be set.
+
+        .. versionadded:: 2.4
+           `defer_init` parameter added.
+        """
+        source_hint = None
+        try:
+            if isinstance(source, str):
+                source_hint = source
+                source = self._parse(source, name, filename)
+            source = self._generate(source, name, filename, defer_init=defer_init)
+            if raw:
+                return source
+            if filename is None:
+                filename = "<template>"
+            return self._compile(source, filename)
+        except TemplateSyntaxError:
+            self.handle_exception(source=source_hint)
+
+    def compile_expression(
+        self, source: str, undefined_to_none: bool = True
+    ) -> "TemplateExpression":
+        """A handy helper method that returns a callable that accepts keyword
+        arguments that appear as variables in the expression.  If called it
+        returns the result of the expression.
+
+        This is useful if applications want to use the same rules as Jinja
+        in template "configuration files" or similar situations.
+
+        Example usage:
+
+        >>> env = Environment()
+        >>> expr = env.compile_expression('foo == 42')
+        >>> expr(foo=23)
+        False
+        >>> expr(foo=42)
+        True
+
+        Per default the return value is converted to `None` if the
+        expression returns an undefined value.  This can be changed
+        by setting `undefined_to_none` to `False`.
+
+        >>> env.compile_expression('var')() is None
+        True
+        >>> env.compile_expression('var', undefined_to_none=False)()
+        Undefined
+
+        .. versionadded:: 2.1
+        """
+        parser = Parser(self, source, state="variable")
+        try:
+            expr = parser.parse_expression()
+            if not parser.stream.eos:
+                raise TemplateSyntaxError(
+                    "chunk after expression", parser.stream.current.lineno, None, None
+                )
+            expr.set_environment(self)
+        except TemplateSyntaxError:
+            self.handle_exception(source=source)
+
+        body = [nodes.Assign(nodes.Name("result", "store"), expr, lineno=1)]
+        template = self.from_string(nodes.Template(body, lineno=1))
+        return TemplateExpression(template, undefined_to_none)
+
+    def compile_templates(
+        self,
+        target: t.Union[str, "os.PathLike[str]"],
+        extensions: t.Optional[t.Collection[str]] = None,
+        filter_func: t.Optional[t.Callable[[str], bool]] = None,
+        zip: t.Optional[str] = "deflated",
+        log_function: t.Optional[t.Callable[[str], None]] = None,
+        ignore_errors: bool = True,
+    ) -> None:
+        """Finds all the templates the loader can find, compiles them
+        and stores them in `target`.  If `zip` is `None`, instead of in a
+        zipfile, the templates will be stored in a directory.
+        By default a deflate zip algorithm is used. To switch to
+        the stored algorithm, `zip` can be set to ``'stored'``.
+
+        `extensions` and `filter_func` are passed to :meth:`list_templates`.
+        Each template returned will be compiled to the target folder or
+        zipfile.
+
+        By default template compilation errors are ignored.  In case a
+        log function is provided, errors are logged.  If you want template
+        syntax errors to abort the compilation you can set `ignore_errors`
+        to `False` and you will get an exception on syntax errors.
+
+        .. versionadded:: 2.4
+        """
+        from .loaders import ModuleLoader
+
+        if log_function is None:
+
+            def log_function(x: str) -> None:
+                pass
+
+        assert log_function is not None
+        assert self.loader is not None, "No loader configured."
+
+        def write_file(filename: str, data: str) -> None:
+            if zip:
+                info = ZipInfo(filename)
+                info.external_attr = 0o755 << 16
+                zip_file.writestr(info, data)
+            else:
+                with open(os.path.join(target, filename), "wb") as f:
+                    f.write(data.encode("utf8"))
+
+        if zip is not None:
+            from zipfile import ZIP_DEFLATED
+            from zipfile import ZIP_STORED
+            from zipfile import ZipFile
+            from zipfile import ZipInfo
+
+            zip_file = ZipFile(
+                target, "w", dict(deflated=ZIP_DEFLATED, stored=ZIP_STORED)[zip]
+            )
+            log_function(f"Compiling into Zip archive {target!r}")
+        else:
+            if not os.path.isdir(target):
+                os.makedirs(target)
+            log_function(f"Compiling into folder {target!r}")
+
+        try:
+            for name in self.list_templates(extensions, filter_func):
+                source, filename, _ = self.loader.get_source(self, name)
+                try:
+                    code = self.compile(source, name, filename, True, True)
+                except TemplateSyntaxError as e:
+                    if not ignore_errors:
+                        raise
+                    log_function(f'Could not compile "{name}": {e}')
+                    continue
+
+                filename = ModuleLoader.get_module_filename(name)
+
+                write_file(filename, code)
+                log_function(f'Compiled "{name}" as {filename}')
+        finally:
+            if zip:
+                zip_file.close()
+
+        log_function("Finished compiling templates")
+
+    def list_templates(
+        self,
+        extensions: t.Optional[t.Collection[str]] = None,
+        filter_func: t.Optional[t.Callable[[str], bool]] = None,
+    ) -> t.List[str]:
+        """Returns a list of templates for this environment.  This requires
+        that the loader supports the loader's
+        :meth:`~BaseLoader.list_templates` method.
+
+        If there are other files in the template folder besides the
+        actual templates, the returned list can be filtered.  There are two
+        ways: either `extensions` is set to a list of file extensions for
+        templates, or a `filter_func` can be provided which is a callable that
+        is passed a template name and should return `True` if it should end up
+        in the result list.
+
+        If the loader does not support that, a :exc:`TypeError` is raised.
+
+        .. versionadded:: 2.4
+        """
+        assert self.loader is not None, "No loader configured."
+        names = self.loader.list_templates()
+
+        if extensions is not None:
+            if filter_func is not None:
+                raise TypeError(
+                    "either extensions or filter_func can be passed, but not both"
+                )
+
+            def filter_func(x: str) -> bool:
+                return "." in x and x.rsplit(".", 1)[1] in extensions
+
+        if filter_func is not None:
+            names = [name for name in names if filter_func(name)]
+
+        return names
+
+    def handle_exception(self, source: t.Optional[str] = None) -> "te.NoReturn":
+        """Exception handling helper.  This is used internally to either raise
+        rewritten exceptions or return a rendered traceback for the template.
+        """
+        from .debug import rewrite_traceback_stack
+
+        raise rewrite_traceback_stack(source=source)
+
+    def join_path(self, template: str, parent: str) -> str:
+        """Join a template with the parent.  By default all the lookups are
+        relative to the loader root so this method returns the `template`
+        parameter unchanged, but if the paths should be relative to the
+        parent template, this function can be used to calculate the real
+        template name.
+
+        Subclasses may override this method and implement template path
+        joining here.
+        """
+        return template
+
+    @internalcode
+    def _load_template(
+        self, name: str, globals: t.Optional[t.MutableMapping[str, t.Any]]
+    ) -> "Template":
+        if self.loader is None:
+            raise TypeError("no loader for this environment specified")
+        cache_key = (weakref.ref(self.loader), name)
+        if self.cache is not None:
+            template = self.cache.get(cache_key)
+            if template is not None and (
+                not self.auto_reload or template.is_up_to_date
+            ):
+                # template.globals is a ChainMap, modifying it will only
+                # affect the template, not the environment globals.
+                if globals:
+                    template.globals.update(globals)
+
+                return template
+
+        template = self.loader.load(self, name, self.make_globals(globals))
+
+        if self.cache is not None:
+            self.cache[cache_key] = template
+        return template
+
+    @internalcode
+    def get_template(
+        self,
+        name: t.Union[str, "Template"],
+        parent: t.Optional[str] = None,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        """Load a template by name with :attr:`loader` and return a
+        :class:`Template`. If the template does not exist a
+        :exc:`TemplateNotFound` exception is raised.
+
+        :param name: Name of the template to load. When loading
+            templates from the filesystem, "/" is used as the path
+            separator, even on Windows.
+        :param parent: The name of the parent template importing this
+            template. :meth:`join_path` can be used to implement name
+            transformations with this.
+        :param globals: Extend the environment :attr:`globals` with
+            these extra variables available for all renders of this
+            template. If the template has already been loaded and
+            cached, its globals are updated with any new items.
+
+        .. versionchanged:: 3.0
+            If a template is loaded from cache, ``globals`` will update
+            the template's globals instead of ignoring the new values.
+
+        .. versionchanged:: 2.4
+            If ``name`` is a :class:`Template` object it is returned
+            unchanged.
+        """
+        if isinstance(name, Template):
+            return name
+        if parent is not None:
+            name = self.join_path(name, parent)
+
+        return self._load_template(name, globals)
+
+    @internalcode
+    def select_template(
+        self,
+        names: t.Iterable[t.Union[str, "Template"]],
+        parent: t.Optional[str] = None,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        """Like :meth:`get_template`, but tries loading multiple names.
+        If none of the names can be loaded a :exc:`TemplatesNotFound`
+        exception is raised.
+
+        :param names: List of template names to try loading in order.
+        :param parent: The name of the parent template importing this
+            template. :meth:`join_path` can be used to implement name
+            transformations with this.
+        :param globals: Extend the environment :attr:`globals` with
+            these extra variables available for all renders of this
+            template. If the template has already been loaded and
+            cached, its globals are updated with any new items.
+
+        .. versionchanged:: 3.0
+            If a template is loaded from cache, ``globals`` will update
+            the template's globals instead of ignoring the new values.
+
+        .. versionchanged:: 2.11
+            If ``names`` is :class:`Undefined`, an :exc:`UndefinedError`
+            is raised instead. If no templates were found and ``names``
+            contains :class:`Undefined`, the message is more helpful.
+
+        .. versionchanged:: 2.4
+            If ``names`` contains a :class:`Template` object it is
+            returned unchanged.
+
+        .. versionadded:: 2.3
+        """
+        if isinstance(names, Undefined):
+            names._fail_with_undefined_error()
+
+        if not names:
+            raise TemplatesNotFound(
+                message="Tried to select from an empty list of templates."
+            )
+
+        for name in names:
+            if isinstance(name, Template):
+                return name
+            if parent is not None:
+                name = self.join_path(name, parent)
+            try:
+                return self._load_template(name, globals)
+            except (TemplateNotFound, UndefinedError):
+                pass
+        raise TemplatesNotFound(names)  # type: ignore
+
+    @internalcode
+    def get_or_select_template(
+        self,
+        template_name_or_list: t.Union[
+            str, "Template", t.List[t.Union[str, "Template"]]
+        ],
+        parent: t.Optional[str] = None,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        """Use :meth:`select_template` if an iterable of template names
+        is given, or :meth:`get_template` if one name is given.
+
+        .. versionadded:: 2.3
+        """
+        if isinstance(template_name_or_list, (str, Undefined)):
+            return self.get_template(template_name_or_list, parent, globals)
+        elif isinstance(template_name_or_list, Template):
+            return template_name_or_list
+        return self.select_template(template_name_or_list, parent, globals)
+
+    def from_string(
+        self,
+        source: t.Union[str, nodes.Template],
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+        template_class: t.Optional[t.Type["Template"]] = None,
+    ) -> "Template":
+        """Load a template from a source string without using
+        :attr:`loader`.
+
+        :param source: Jinja source to compile into a template.
+        :param globals: Extend the environment :attr:`globals` with
+            these extra variables available for all renders of this
+            template. If the template has already been loaded and
+            cached, its globals are updated with any new items.
+        :param template_class: Return an instance of this
+            :class:`Template` class.
+        """
+        gs = self.make_globals(globals)
+        cls = template_class or self.template_class
+        return cls.from_code(self, self.compile(source), gs, None)
+
+    def make_globals(
+        self, d: t.Optional[t.MutableMapping[str, t.Any]]
+    ) -> t.MutableMapping[str, t.Any]:
+        """Make the globals map for a template. Any given template
+        globals overlay the environment :attr:`globals`.
+
+        Returns a :class:`collections.ChainMap`. This allows any changes
+        to a template's globals to only affect that template, while
+        changes to the environment's globals are still reflected.
+        However, avoid modifying any globals after a template is loaded.
+
+        :param d: Dict of template-specific globals.
+
+        .. versionchanged:: 3.0
+            Use :class:`collections.ChainMap` to always prevent mutating
+            environment globals.
+        """
+        if d is None:
+            d = {}
+
+        return ChainMap(d, self.globals)
+
+
+class Template:
+    """A compiled template that can be rendered.
+
+    Use the methods on :class:`Environment` to create or load templates.
+    The environment is used to configure how templates are compiled and
+    behave.
+
+    It is also possible to create a template object directly. This is
+    not usually recommended. The constructor takes most of the same
+    arguments as :class:`Environment`. All templates created with the
+    same environment arguments share the same ephemeral ``Environment``
+    instance behind the scenes.
+
+    A template object should be considered immutable. Modifications on
+    the object are not supported.
+    """
+
+    #: Type of environment to create when creating a template directly
+    #: rather than through an existing environment.
+    environment_class: t.Type[Environment] = Environment
+
+    environment: Environment
+    globals: t.MutableMapping[str, t.Any]
+    name: t.Optional[str]
+    filename: t.Optional[str]
+    blocks: t.Dict[str, t.Callable[[Context], t.Iterator[str]]]
+    root_render_func: t.Callable[[Context], t.Iterator[str]]
+    _module: t.Optional["TemplateModule"]
+    _debug_info: str
+    _uptodate: t.Optional[t.Callable[[], bool]]
+
+    def __new__(
+        cls,
+        source: t.Union[str, nodes.Template],
+        block_start_string: str = BLOCK_START_STRING,
+        block_end_string: str = BLOCK_END_STRING,
+        variable_start_string: str = VARIABLE_START_STRING,
+        variable_end_string: str = VARIABLE_END_STRING,
+        comment_start_string: str = COMMENT_START_STRING,
+        comment_end_string: str = COMMENT_END_STRING,
+        line_statement_prefix: t.Optional[str] = LINE_STATEMENT_PREFIX,
+        line_comment_prefix: t.Optional[str] = LINE_COMMENT_PREFIX,
+        trim_blocks: bool = TRIM_BLOCKS,
+        lstrip_blocks: bool = LSTRIP_BLOCKS,
+        newline_sequence: "te.Literal['\\n', '\\r\\n', '\\r']" = NEWLINE_SEQUENCE,
+        keep_trailing_newline: bool = KEEP_TRAILING_NEWLINE,
+        extensions: t.Sequence[t.Union[str, t.Type["Extension"]]] = (),
+        optimized: bool = True,
+        undefined: t.Type[Undefined] = Undefined,
+        finalize: t.Optional[t.Callable[..., t.Any]] = None,
+        autoescape: t.Union[bool, t.Callable[[t.Optional[str]], bool]] = False,
+        enable_async: bool = False,
+    ) -> t.Any:  # it returns a `Template`, but this breaks the sphinx build...
+        env = get_spontaneous_environment(
+            cls.environment_class,  # type: ignore
+            block_start_string,
+            block_end_string,
+            variable_start_string,
+            variable_end_string,
+            comment_start_string,
+            comment_end_string,
+            line_statement_prefix,
+            line_comment_prefix,
+            trim_blocks,
+            lstrip_blocks,
+            newline_sequence,
+            keep_trailing_newline,
+            frozenset(extensions),
+            optimized,
+            undefined,  # type: ignore
+            finalize,
+            autoescape,
+            None,
+            0,
+            False,
+            None,
+            enable_async,
+        )
+        return env.from_string(source, template_class=cls)
+
+    @classmethod
+    def from_code(
+        cls,
+        environment: Environment,
+        code: CodeType,
+        globals: t.MutableMapping[str, t.Any],
+        uptodate: t.Optional[t.Callable[[], bool]] = None,
+    ) -> "Template":
+        """Creates a template object from compiled code and the globals.  This
+        is used by the loaders and environment to create a template object.
+        """
+        namespace = {"environment": environment, "__file__": code.co_filename}
+        exec(code, namespace)
+        rv = cls._from_namespace(environment, namespace, globals)
+        rv._uptodate = uptodate
+        return rv
+
+    @classmethod
+    def from_module_dict(
+        cls,
+        environment: Environment,
+        module_dict: t.MutableMapping[str, t.Any],
+        globals: t.MutableMapping[str, t.Any],
+    ) -> "Template":
+        """Creates a template object from a module.  This is used by the
+        module loader to create a template object.
+
+        .. versionadded:: 2.4
+        """
+        return cls._from_namespace(environment, module_dict, globals)
+
+    @classmethod
+    def _from_namespace(
+        cls,
+        environment: Environment,
+        namespace: t.MutableMapping[str, t.Any],
+        globals: t.MutableMapping[str, t.Any],
+    ) -> "Template":
+        t: Template = object.__new__(cls)
+        t.environment = environment
+        t.globals = globals
+        t.name = namespace["name"]
+        t.filename = namespace["__file__"]
+        t.blocks = namespace["blocks"]
+
+        # render function and module
+        t.root_render_func = namespace["root"]
+        t._module = None
+
+        # debug and loader helpers
+        t._debug_info = namespace["debug_info"]
+        t._uptodate = None
+
+        # store the reference
+        namespace["environment"] = environment
+        namespace["__jinja_template__"] = t
+
+        return t
+
+    def render(self, *args: t.Any, **kwargs: t.Any) -> str:
+        """This method accepts the same arguments as the `dict` constructor:
+        A dict, a dict subclass or some keyword arguments.  If no arguments
+        are given the context will be empty.  These two calls do the same::
+
+            template.render(knights='that say nih')
+            template.render({'knights': 'that say nih'})
+
+        This will return the rendered template as a string.
+        """
+        if self.environment.is_async:
+            import asyncio
+
+            return asyncio.run(self.render_async(*args, **kwargs))
+
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            return self.environment.concat(self.root_render_func(ctx))  # type: ignore
+        except Exception:
+            self.environment.handle_exception()
+
+    async def render_async(self, *args: t.Any, **kwargs: t.Any) -> str:
+        """This works similar to :meth:`render` but returns a coroutine
+        that when awaited returns the entire rendered template string.  This
+        requires the async feature to be enabled.
+
+        Example usage::
+
+            await template.render_async(knights='that say nih; asynchronously')
+        """
+        if not self.environment.is_async:
+            raise RuntimeError(
+                "The environment was not created with async mode enabled."
+            )
+
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            return self.environment.concat(  # type: ignore
+                [n async for n in self.root_render_func(ctx)]  # type: ignore
+            )
+        except Exception:
+            return self.environment.handle_exception()
+
+    def stream(self, *args: t.Any, **kwargs: t.Any) -> "TemplateStream":
+        """Works exactly like :meth:`generate` but returns a
+        :class:`TemplateStream`.
+        """
+        return TemplateStream(self.generate(*args, **kwargs))
+
+    def generate(self, *args: t.Any, **kwargs: t.Any) -> t.Iterator[str]:
+        """For very large templates it can be useful to not render the whole
+        template at once but evaluate each statement after another and yield
+        piece for piece.  This method basically does exactly that and returns
+        a generator that yields one item after another as strings.
+
+        It accepts the same arguments as :meth:`render`.
+        """
+        if self.environment.is_async:
+            import asyncio
+
+            async def to_list() -> t.List[str]:
+                return [x async for x in self.generate_async(*args, **kwargs)]
+
+            yield from asyncio.run(to_list())
+            return
+
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            yield from self.root_render_func(ctx)
+        except Exception:
+            yield self.environment.handle_exception()
+
+    async def generate_async(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> t.AsyncGenerator[str, object]:
+        """An async version of :meth:`generate`.  Works very similarly but
+        returns an async iterator instead.
+        """
+        if not self.environment.is_async:
+            raise RuntimeError(
+                "The environment was not created with async mode enabled."
+            )
+
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            agen = self.root_render_func(ctx)
+            try:
+                async for event in agen:  # type: ignore
+                    yield event
+            finally:
+                # we can't use async with aclosing(...) because that's only
+                # in 3.10+
+                await agen.aclose()  # type: ignore
+        except Exception:
+            yield self.environment.handle_exception()
+
+    def new_context(
+        self,
+        vars: t.Optional[t.Dict[str, t.Any]] = None,
+        shared: bool = False,
+        locals: t.Optional[t.Mapping[str, t.Any]] = None,
+    ) -> Context:
+        """Create a new :class:`Context` for this template.  The vars
+        provided will be passed to the template.  Per default the globals
+        are added to the context.  If shared is set to `True` the data
+        is passed as is to the context without adding the globals.
+
+        `locals` can be a dict of local variables for internal usage.
+        """
+        return new_context(
+            self.environment, self.name, self.blocks, vars, shared, self.globals, locals
+        )
+
+    def make_module(
+        self,
+        vars: t.Optional[t.Dict[str, t.Any]] = None,
+        shared: bool = False,
+        locals: t.Optional[t.Mapping[str, t.Any]] = None,
+    ) -> "TemplateModule":
+        """This method works like the :attr:`module` attribute when called
+        without arguments but it will evaluate the template on every call
+        rather than caching it.  It's also possible to provide
+        a dict which is then used as context.  The arguments are the same
+        as for the :meth:`new_context` method.
+        """
+        ctx = self.new_context(vars, shared, locals)
+        return TemplateModule(self, ctx)
+
+    async def make_module_async(
+        self,
+        vars: t.Optional[t.Dict[str, t.Any]] = None,
+        shared: bool = False,
+        locals: t.Optional[t.Mapping[str, t.Any]] = None,
+    ) -> "TemplateModule":
+        """As template module creation can invoke template code for
+        asynchronous executions this method must be used instead of the
+        normal :meth:`make_module` one.  Likewise the module attribute
+        becomes unavailable in async mode.
+        """
+        ctx = self.new_context(vars, shared, locals)
+        return TemplateModule(
+            self,
+            ctx,
+            [x async for x in self.root_render_func(ctx)],  # type: ignore
+        )
+
+    @internalcode
+    def _get_default_module(self, ctx: t.Optional[Context] = None) -> "TemplateModule":
+        """If a context is passed in, this means that the template was
+        imported. Imported templates have access to the current
+        template's globals by default, but they can only be accessed via
+        the context during runtime.
+
+        If there are new globals, we need to create a new module because
+        the cached module is already rendered and will not have access
+        to globals from the current context. This new module is not
+        cached because the template can be imported elsewhere, and it
+        should have access to only the current template's globals.
+        """
+        if self.environment.is_async:
+            raise RuntimeError("Module is not available in async mode.")
+
+        if ctx is not None:
+            keys = ctx.globals_keys - self.globals.keys()
+
+            if keys:
+                return self.make_module({k: ctx.parent[k] for k in keys})
+
+        if self._module is None:
+            self._module = self.make_module()
+
+        return self._module
+
+    async def _get_default_module_async(
+        self, ctx: t.Optional[Context] = None
+    ) -> "TemplateModule":
+        if ctx is not None:
+            keys = ctx.globals_keys - self.globals.keys()
+
+            if keys:
+                return await self.make_module_async({k: ctx.parent[k] for k in keys})
+
+        if self._module is None:
+            self._module = await self.make_module_async()
+
+        return self._module
+
+    @property
+    def module(self) -> "TemplateModule":
+        """The template as module.  This is used for imports in the
+        template runtime but is also useful if one wants to access
+        exported template variables from the Python layer:
+
+        >>> t = Template('{% macro foo() %}42{% endmacro %}23')
+        >>> str(t.module)
+        '23'
+        >>> t.module.foo() == u'42'
+        True
+
+        This attribute is not available if async mode is enabled.
+        """
+        return self._get_default_module()
+
+    def get_corresponding_lineno(self, lineno: int) -> int:
+        """Return the source line number of a line number in the
+        generated bytecode as they are not in sync.
+        """
+        for template_line, code_line in reversed(self.debug_info):
+            if code_line <= lineno:
+                return template_line
+        return 1
+
+    @property
+    def is_up_to_date(self) -> bool:
+        """If this variable is `False` there is a newer version available."""
+        if self._uptodate is None:
+            return True
+        return self._uptodate()
+
+    @property
+    def debug_info(self) -> t.List[t.Tuple[int, int]]:
+        """The debug info mapping."""
+        if self._debug_info:
+            return [
+                tuple(map(int, x.split("=")))  # type: ignore
+                for x in self._debug_info.split("&")
+            ]
+
+        return []
+
+    def __repr__(self) -> str:
+        if self.name is None:
+            name = f"memory:{id(self):x}"
+        else:
+            name = repr(self.name)
+        return f"<{type(self).__name__} {name}>"
+
+
+class TemplateModule:
+    """Represents an imported template.  All the exported names of the
+    template are available as attributes on this object.  Additionally
+    converting it into a string renders the contents.
+    """
+
+    def __init__(
+        self,
+        template: Template,
+        context: Context,
+        body_stream: t.Optional[t.Iterable[str]] = None,
+    ) -> None:
+        if body_stream is None:
+            if context.environment.is_async:
+                raise RuntimeError(
+                    "Async mode requires a body stream to be passed to"
+                    " a template module. Use the async methods of the"
+                    " API you are using."
+                )
+
+            body_stream = list(template.root_render_func(context))
+
+        self._body_stream = body_stream
+        self.__dict__.update(context.get_exported())
+        self.__name__ = template.name
+
+    def __html__(self) -> Markup:
+        return Markup(concat(self._body_stream))
+
+    def __str__(self) -> str:
+        return concat(self._body_stream)
+
+    def __repr__(self) -> str:
+        if self.__name__ is None:
+            name = f"memory:{id(self):x}"
+        else:
+            name = repr(self.__name__)
+        return f"<{type(self).__name__} {name}>"
+
+
+class TemplateExpression:
+    """The :meth:`jinja2.Environment.compile_expression` method returns an
+    instance of this object.  It encapsulates the expression-like access
+    to the template with an expression it wraps.
+    """
+
+    def __init__(self, template: Template, undefined_to_none: bool) -> None:
+        self._template = template
+        self._undefined_to_none = undefined_to_none
+
+    def __call__(self, *args: t.Any, **kwargs: t.Any) -> t.Optional[t.Any]:
+        context = self._template.new_context(dict(*args, **kwargs))
+        consume(self._template.root_render_func(context))
+        rv = context.vars["result"]
+        if self._undefined_to_none and isinstance(rv, Undefined):
+            rv = None
+        return rv
+
+
+class TemplateStream:
+    """A template stream works pretty much like an ordinary python generator
+    but it can buffer multiple items to reduce the number of total iterations.
+    Per default the output is unbuffered which means that for every unbuffered
+    instruction in the template one string is yielded.
+
+    If buffering is enabled with a buffer size of 5, five items are combined
+    into a new string.  This is mainly useful if you are streaming
+    big templates to a client via WSGI which flushes after each iteration.
+    """
+
+    def __init__(self, gen: t.Iterator[str]) -> None:
+        self._gen = gen
+        self.disable_buffering()
+
+    def dump(
+        self,
+        fp: t.Union[str, t.IO[bytes]],
+        encoding: t.Optional[str] = None,
+        errors: t.Optional[str] = "strict",
+    ) -> None:
+        """Dump the complete stream into a file or file-like object.
+        Per default strings are written, if you want to encode
+        before writing specify an `encoding`.
+
+        Example usage::
+
+            Template('Hello {{ name }}!').stream(name='foo').dump('hello.html')
+        """
+        close = False
+
+        if isinstance(fp, str):
+            if encoding is None:
+                encoding = "utf-8"
+
+            real_fp: t.IO[bytes] = open(fp, "wb")
+            close = True
+        else:
+            real_fp = fp
+
+        try:
+            if encoding is not None:
+                iterable = (x.encode(encoding, errors) for x in self)  # type: ignore
+            else:
+                iterable = self  # type: ignore
+
+            if hasattr(real_fp, "writelines"):
+                real_fp.writelines(iterable)
+            else:
+                for item in iterable:
+                    real_fp.write(item)
+        finally:
+            if close:
+                real_fp.close()
+
+    def disable_buffering(self) -> None:
+        """Disable the output buffering."""
+        self._next = partial(next, self._gen)
+        self.buffered = False
+
+    def _buffered_generator(self, size: int) -> t.Iterator[str]:
+        buf: t.List[str] = []
+        c_size = 0
+        push = buf.append
+
+        while True:
+            try:
+                while c_size < size:
+                    c = next(self._gen)
+                    push(c)
+                    if c:
+                        c_size += 1
+            except StopIteration:
+                if not c_size:
+                    return
+            yield concat(buf)
+            del buf[:]
+            c_size = 0
+
+    def enable_buffering(self, size: int = 5) -> None:
+        """Enable buffering.  Buffer `size` items before yielding them."""
+        if size <= 1:
+            raise ValueError("buffer size too small")
+
+        self.buffered = True
+        self._next = partial(next, self._buffered_generator(size))
+
+    def __iter__(self) -> "TemplateStream":
+        return self
+
+    def __next__(self) -> str:
+        return self._next()  # type: ignore
+
+
+# hook in default template class.  if anyone reads this comment: ignore that
+# it's possible to use custom templates ;-)
+Environment.template_class = Template
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/exceptions.py b/backend/venv/lib/python3.12/site-packages/jinja2/exceptions.py
new file mode 100644
index 00000000..082ebe8f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/exceptions.py
@@ -0,0 +1,166 @@
+import typing as t
+
+if t.TYPE_CHECKING:
+    from .runtime import Undefined
+
+
+class TemplateError(Exception):
+    """Baseclass for all template errors."""
+
+    def __init__(self, message: t.Optional[str] = None) -> None:
+        super().__init__(message)
+
+    @property
+    def message(self) -> t.Optional[str]:
+        return self.args[0] if self.args else None
+
+
+class TemplateNotFound(IOError, LookupError, TemplateError):
+    """Raised if a template does not exist.
+
+    .. versionchanged:: 2.11
+        If the given name is :class:`Undefined` and no message was
+        provided, an :exc:`UndefinedError` is raised.
+    """
+
+    # Silence the Python warning about message being deprecated since
+    # it's not valid here.
+    message: t.Optional[str] = None
+
+    def __init__(
+        self,
+        name: t.Optional[t.Union[str, "Undefined"]],
+        message: t.Optional[str] = None,
+    ) -> None:
+        IOError.__init__(self, name)
+
+        if message is None:
+            from .runtime import Undefined
+
+            if isinstance(name, Undefined):
+                name._fail_with_undefined_error()
+
+            message = name
+
+        self.message = message
+        self.name = name
+        self.templates = [name]
+
+    def __str__(self) -> str:
+        return str(self.message)
+
+
+class TemplatesNotFound(TemplateNotFound):
+    """Like :class:`TemplateNotFound` but raised if multiple templates
+    are selected.  This is a subclass of :class:`TemplateNotFound`
+    exception, so just catching the base exception will catch both.
+
+    .. versionchanged:: 2.11
+        If a name in the list of names is :class:`Undefined`, a message
+        about it being undefined is shown rather than the empty string.
+
+    .. versionadded:: 2.2
+    """
+
+    def __init__(
+        self,
+        names: t.Sequence[t.Union[str, "Undefined"]] = (),
+        message: t.Optional[str] = None,
+    ) -> None:
+        if message is None:
+            from .runtime import Undefined
+
+            parts = []
+
+            for name in names:
+                if isinstance(name, Undefined):
+                    parts.append(name._undefined_message)
+                else:
+                    parts.append(name)
+
+            parts_str = ", ".join(map(str, parts))
+            message = f"none of the templates given were found: {parts_str}"
+
+        super().__init__(names[-1] if names else None, message)
+        self.templates = list(names)
+
+
+class TemplateSyntaxError(TemplateError):
+    """Raised to tell the user that there is a problem with the template."""
+
+    def __init__(
+        self,
+        message: str,
+        lineno: int,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+    ) -> None:
+        super().__init__(message)
+        self.lineno = lineno
+        self.name = name
+        self.filename = filename
+        self.source: t.Optional[str] = None
+
+        # this is set to True if the debug.translate_syntax_error
+        # function translated the syntax error into a new traceback
+        self.translated = False
+
+    def __str__(self) -> str:
+        # for translated errors we only return the message
+        if self.translated:
+            return t.cast(str, self.message)
+
+        # otherwise attach some stuff
+        location = f"line {self.lineno}"
+        name = self.filename or self.name
+        if name:
+            location = f'File "{name}", {location}'
+        lines = [t.cast(str, self.message), "  " + location]
+
+        # if the source is set, add the line to the output
+        if self.source is not None:
+            try:
+                line = self.source.splitlines()[self.lineno - 1]
+            except IndexError:
+                pass
+            else:
+                lines.append("    " + line.strip())
+
+        return "\n".join(lines)
+
+    def __reduce__(self):  # type: ignore
+        # https://bugs.python.org/issue1692335 Exceptions that take
+        # multiple required arguments have problems with pickling.
+        # Without this, raises TypeError: __init__() missing 1 required
+        # positional argument: 'lineno'
+        return self.__class__, (self.message, self.lineno, self.name, self.filename)
+
+
+class TemplateAssertionError(TemplateSyntaxError):
+    """Like a template syntax error, but covers cases where something in the
+    template caused an error at compile time that wasn't necessarily caused
+    by a syntax error.  However it's a direct subclass of
+    :exc:`TemplateSyntaxError` and has the same attributes.
+    """
+
+
+class TemplateRuntimeError(TemplateError):
+    """A generic runtime error in the template engine.  Under some situations
+    Jinja may raise this exception.
+    """
+
+
+class UndefinedError(TemplateRuntimeError):
+    """Raised if a template tries to operate on :class:`Undefined`."""
+
+
+class SecurityError(TemplateRuntimeError):
+    """Raised if a template tries to do something insecure if the
+    sandbox is enabled.
+    """
+
+
+class FilterArgumentError(TemplateRuntimeError):
+    """This error is raised if a filter was called with inappropriate
+    arguments
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/ext.py b/backend/venv/lib/python3.12/site-packages/jinja2/ext.py
new file mode 100644
index 00000000..c7af8d45
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/ext.py
@@ -0,0 +1,870 @@
+"""Extension API for adding custom tags and behavior."""
+
+import pprint
+import re
+import typing as t
+
+from markupsafe import Markup
+
+from . import defaults
+from . import nodes
+from .environment import Environment
+from .exceptions import TemplateAssertionError
+from .exceptions import TemplateSyntaxError
+from .runtime import concat  # type: ignore
+from .runtime import Context
+from .runtime import Undefined
+from .utils import import_string
+from .utils import pass_context
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .lexer import Token
+    from .lexer import TokenStream
+    from .parser import Parser
+
+    class _TranslationsBasic(te.Protocol):
+        def gettext(self, message: str) -> str: ...
+
+        def ngettext(self, singular: str, plural: str, n: int) -> str:
+            pass
+
+    class _TranslationsContext(_TranslationsBasic):
+        def pgettext(self, context: str, message: str) -> str: ...
+
+        def npgettext(
+            self, context: str, singular: str, plural: str, n: int
+        ) -> str: ...
+
+    _SupportedTranslations = t.Union[_TranslationsBasic, _TranslationsContext]
+
+
+# I18N functions available in Jinja templates. If the I18N library
+# provides ugettext, it will be assigned to gettext.
+GETTEXT_FUNCTIONS: t.Tuple[str, ...] = (
+    "_",
+    "gettext",
+    "ngettext",
+    "pgettext",
+    "npgettext",
+)
+_ws_re = re.compile(r"\s*\n\s*")
+
+
+class Extension:
+    """Extensions can be used to add extra functionality to the Jinja template
+    system at the parser level.  Custom extensions are bound to an environment
+    but may not store environment specific data on `self`.  The reason for
+    this is that an extension can be bound to another environment (for
+    overlays) by creating a copy and reassigning the `environment` attribute.
+
+    As extensions are created by the environment they cannot accept any
+    arguments for configuration.  One may want to work around that by using
+    a factory function, but that is not possible as extensions are identified
+    by their import name.  The correct way to configure the extension is
+    storing the configuration values on the environment.  Because this way the
+    environment ends up acting as central configuration storage the
+    attributes may clash which is why extensions have to ensure that the names
+    they choose for configuration are not too generic.  ``prefix`` for example
+    is a terrible name, ``fragment_cache_prefix`` on the other hand is a good
+    name as includes the name of the extension (fragment cache).
+    """
+
+    identifier: t.ClassVar[str]
+
+    def __init_subclass__(cls) -> None:
+        cls.identifier = f"{cls.__module__}.{cls.__name__}"
+
+    #: if this extension parses this is the list of tags it's listening to.
+    tags: t.Set[str] = set()
+
+    #: the priority of that extension.  This is especially useful for
+    #: extensions that preprocess values.  A lower value means higher
+    #: priority.
+    #:
+    #: .. versionadded:: 2.4
+    priority = 100
+
+    def __init__(self, environment: Environment) -> None:
+        self.environment = environment
+
+    def bind(self, environment: Environment) -> "te.Self":
+        """Create a copy of this extension bound to another environment."""
+        rv = object.__new__(self.__class__)
+        rv.__dict__.update(self.__dict__)
+        rv.environment = environment
+        return rv
+
+    def preprocess(
+        self, source: str, name: t.Optional[str], filename: t.Optional[str] = None
+    ) -> str:
+        """This method is called before the actual lexing and can be used to
+        preprocess the source.  The `filename` is optional.  The return value
+        must be the preprocessed source.
+        """
+        return source
+
+    def filter_stream(
+        self, stream: "TokenStream"
+    ) -> t.Union["TokenStream", t.Iterable["Token"]]:
+        """It's passed a :class:`~jinja2.lexer.TokenStream` that can be used
+        to filter tokens returned.  This method has to return an iterable of
+        :class:`~jinja2.lexer.Token`\\s, but it doesn't have to return a
+        :class:`~jinja2.lexer.TokenStream`.
+        """
+        return stream
+
+    def parse(self, parser: "Parser") -> t.Union[nodes.Node, t.List[nodes.Node]]:
+        """If any of the :attr:`tags` matched this method is called with the
+        parser as first argument.  The token the parser stream is pointing at
+        is the name token that matched.  This method has to return one or a
+        list of multiple nodes.
+        """
+        raise NotImplementedError()
+
+    def attr(
+        self, name: str, lineno: t.Optional[int] = None
+    ) -> nodes.ExtensionAttribute:
+        """Return an attribute node for the current extension.  This is useful
+        to pass constants on extensions to generated template code.
+
+        ::
+
+            self.attr('_my_attribute', lineno=lineno)
+        """
+        return nodes.ExtensionAttribute(self.identifier, name, lineno=lineno)
+
+    def call_method(
+        self,
+        name: str,
+        args: t.Optional[t.List[nodes.Expr]] = None,
+        kwargs: t.Optional[t.List[nodes.Keyword]] = None,
+        dyn_args: t.Optional[nodes.Expr] = None,
+        dyn_kwargs: t.Optional[nodes.Expr] = None,
+        lineno: t.Optional[int] = None,
+    ) -> nodes.Call:
+        """Call a method of the extension.  This is a shortcut for
+        :meth:`attr` + :class:`jinja2.nodes.Call`.
+        """
+        if args is None:
+            args = []
+        if kwargs is None:
+            kwargs = []
+        return nodes.Call(
+            self.attr(name, lineno=lineno),
+            args,
+            kwargs,
+            dyn_args,
+            dyn_kwargs,
+            lineno=lineno,
+        )
+
+
+@pass_context
+def _gettext_alias(
+    __context: Context, *args: t.Any, **kwargs: t.Any
+) -> t.Union[t.Any, Undefined]:
+    return __context.call(__context.resolve("gettext"), *args, **kwargs)
+
+
+def _make_new_gettext(func: t.Callable[[str], str]) -> t.Callable[..., str]:
+    @pass_context
+    def gettext(__context: Context, __string: str, **variables: t.Any) -> str:
+        rv = __context.call(func, __string)
+        if __context.eval_ctx.autoescape:
+            rv = Markup(rv)
+        # Always treat as a format string, even if there are no
+        # variables. This makes translation strings more consistent
+        # and predictable. This requires escaping
+        return rv % variables  # type: ignore
+
+    return gettext
+
+
+def _make_new_ngettext(func: t.Callable[[str, str, int], str]) -> t.Callable[..., str]:
+    @pass_context
+    def ngettext(
+        __context: Context,
+        __singular: str,
+        __plural: str,
+        __num: int,
+        **variables: t.Any,
+    ) -> str:
+        variables.setdefault("num", __num)
+        rv = __context.call(func, __singular, __plural, __num)
+        if __context.eval_ctx.autoescape:
+            rv = Markup(rv)
+        # Always treat as a format string, see gettext comment above.
+        return rv % variables  # type: ignore
+
+    return ngettext
+
+
+def _make_new_pgettext(func: t.Callable[[str, str], str]) -> t.Callable[..., str]:
+    @pass_context
+    def pgettext(
+        __context: Context, __string_ctx: str, __string: str, **variables: t.Any
+    ) -> str:
+        variables.setdefault("context", __string_ctx)
+        rv = __context.call(func, __string_ctx, __string)
+
+        if __context.eval_ctx.autoescape:
+            rv = Markup(rv)
+
+        # Always treat as a format string, see gettext comment above.
+        return rv % variables  # type: ignore
+
+    return pgettext
+
+
+def _make_new_npgettext(
+    func: t.Callable[[str, str, str, int], str],
+) -> t.Callable[..., str]:
+    @pass_context
+    def npgettext(
+        __context: Context,
+        __string_ctx: str,
+        __singular: str,
+        __plural: str,
+        __num: int,
+        **variables: t.Any,
+    ) -> str:
+        variables.setdefault("context", __string_ctx)
+        variables.setdefault("num", __num)
+        rv = __context.call(func, __string_ctx, __singular, __plural, __num)
+
+        if __context.eval_ctx.autoescape:
+            rv = Markup(rv)
+
+        # Always treat as a format string, see gettext comment above.
+        return rv % variables  # type: ignore
+
+    return npgettext
+
+
+class InternationalizationExtension(Extension):
+    """This extension adds gettext support to Jinja."""
+
+    tags = {"trans"}
+
+    # TODO: the i18n extension is currently reevaluating values in a few
+    # situations.  Take this example:
+    #   {% trans count=something() %}{{ count }} foo{% pluralize
+    #     %}{{ count }} fooss{% endtrans %}
+    # something is called twice here.  One time for the gettext value and
+    # the other time for the n-parameter of the ngettext function.
+
+    def __init__(self, environment: Environment) -> None:
+        super().__init__(environment)
+        environment.globals["_"] = _gettext_alias
+        environment.extend(
+            install_gettext_translations=self._install,
+            install_null_translations=self._install_null,
+            install_gettext_callables=self._install_callables,
+            uninstall_gettext_translations=self._uninstall,
+            extract_translations=self._extract,
+            newstyle_gettext=False,
+        )
+
+    def _install(
+        self, translations: "_SupportedTranslations", newstyle: t.Optional[bool] = None
+    ) -> None:
+        # ugettext and ungettext are preferred in case the I18N library
+        # is providing compatibility with older Python versions.
+        gettext = getattr(translations, "ugettext", None)
+        if gettext is None:
+            gettext = translations.gettext
+        ngettext = getattr(translations, "ungettext", None)
+        if ngettext is None:
+            ngettext = translations.ngettext
+
+        pgettext = getattr(translations, "pgettext", None)
+        npgettext = getattr(translations, "npgettext", None)
+        self._install_callables(
+            gettext, ngettext, newstyle=newstyle, pgettext=pgettext, npgettext=npgettext
+        )
+
+    def _install_null(self, newstyle: t.Optional[bool] = None) -> None:
+        import gettext
+
+        translations = gettext.NullTranslations()
+
+        if hasattr(translations, "pgettext"):
+            # Python < 3.8
+            pgettext = translations.pgettext
+        else:
+
+            def pgettext(c: str, s: str) -> str:  # type: ignore[misc]
+                return s
+
+        if hasattr(translations, "npgettext"):
+            npgettext = translations.npgettext
+        else:
+
+            def npgettext(c: str, s: str, p: str, n: int) -> str:  # type: ignore[misc]
+                return s if n == 1 else p
+
+        self._install_callables(
+            gettext=translations.gettext,
+            ngettext=translations.ngettext,
+            newstyle=newstyle,
+            pgettext=pgettext,
+            npgettext=npgettext,
+        )
+
+    def _install_callables(
+        self,
+        gettext: t.Callable[[str], str],
+        ngettext: t.Callable[[str, str, int], str],
+        newstyle: t.Optional[bool] = None,
+        pgettext: t.Optional[t.Callable[[str, str], str]] = None,
+        npgettext: t.Optional[t.Callable[[str, str, str, int], str]] = None,
+    ) -> None:
+        if newstyle is not None:
+            self.environment.newstyle_gettext = newstyle  # type: ignore
+        if self.environment.newstyle_gettext:  # type: ignore
+            gettext = _make_new_gettext(gettext)
+            ngettext = _make_new_ngettext(ngettext)
+
+            if pgettext is not None:
+                pgettext = _make_new_pgettext(pgettext)
+
+            if npgettext is not None:
+                npgettext = _make_new_npgettext(npgettext)
+
+        self.environment.globals.update(
+            gettext=gettext, ngettext=ngettext, pgettext=pgettext, npgettext=npgettext
+        )
+
+    def _uninstall(self, translations: "_SupportedTranslations") -> None:
+        for key in ("gettext", "ngettext", "pgettext", "npgettext"):
+            self.environment.globals.pop(key, None)
+
+    def _extract(
+        self,
+        source: t.Union[str, nodes.Template],
+        gettext_functions: t.Sequence[str] = GETTEXT_FUNCTIONS,
+    ) -> t.Iterator[
+        t.Tuple[int, str, t.Union[t.Optional[str], t.Tuple[t.Optional[str], ...]]]
+    ]:
+        if isinstance(source, str):
+            source = self.environment.parse(source)
+        return extract_from_ast(source, gettext_functions)
+
+    def parse(self, parser: "Parser") -> t.Union[nodes.Node, t.List[nodes.Node]]:
+        """Parse a translatable tag."""
+        lineno = next(parser.stream).lineno
+
+        context = None
+        context_token = parser.stream.next_if("string")
+
+        if context_token is not None:
+            context = context_token.value
+
+        # find all the variables referenced.  Additionally a variable can be
+        # defined in the body of the trans block too, but this is checked at
+        # a later state.
+        plural_expr: t.Optional[nodes.Expr] = None
+        plural_expr_assignment: t.Optional[nodes.Assign] = None
+        num_called_num = False
+        variables: t.Dict[str, nodes.Expr] = {}
+        trimmed = None
+        while parser.stream.current.type != "block_end":
+            if variables:
+                parser.stream.expect("comma")
+
+            # skip colon for python compatibility
+            if parser.stream.skip_if("colon"):
+                break
+
+            token = parser.stream.expect("name")
+            if token.value in variables:
+                parser.fail(
+                    f"translatable variable {token.value!r} defined twice.",
+                    token.lineno,
+                    exc=TemplateAssertionError,
+                )
+
+            # expressions
+            if parser.stream.current.type == "assign":
+                next(parser.stream)
+                variables[token.value] = var = parser.parse_expression()
+            elif trimmed is None and token.value in ("trimmed", "notrimmed"):
+                trimmed = token.value == "trimmed"
+                continue
+            else:
+                variables[token.value] = var = nodes.Name(token.value, "load")
+
+            if plural_expr is None:
+                if isinstance(var, nodes.Call):
+                    plural_expr = nodes.Name("_trans", "load")
+                    variables[token.value] = plural_expr
+                    plural_expr_assignment = nodes.Assign(
+                        nodes.Name("_trans", "store"), var
+                    )
+                else:
+                    plural_expr = var
+                num_called_num = token.value == "num"
+
+        parser.stream.expect("block_end")
+
+        plural = None
+        have_plural = False
+        referenced = set()
+
+        # now parse until endtrans or pluralize
+        singular_names, singular = self._parse_block(parser, True)
+        if singular_names:
+            referenced.update(singular_names)
+            if plural_expr is None:
+                plural_expr = nodes.Name(singular_names[0], "load")
+                num_called_num = singular_names[0] == "num"
+
+        # if we have a pluralize block, we parse that too
+        if parser.stream.current.test("name:pluralize"):
+            have_plural = True
+            next(parser.stream)
+            if parser.stream.current.type != "block_end":
+                token = parser.stream.expect("name")
+                if token.value not in variables:
+                    parser.fail(
+                        f"unknown variable {token.value!r} for pluralization",
+                        token.lineno,
+                        exc=TemplateAssertionError,
+                    )
+                plural_expr = variables[token.value]
+                num_called_num = token.value == "num"
+            parser.stream.expect("block_end")
+            plural_names, plural = self._parse_block(parser, False)
+            next(parser.stream)
+            referenced.update(plural_names)
+        else:
+            next(parser.stream)
+
+        # register free names as simple name expressions
+        for name in referenced:
+            if name not in variables:
+                variables[name] = nodes.Name(name, "load")
+
+        if not have_plural:
+            plural_expr = None
+        elif plural_expr is None:
+            parser.fail("pluralize without variables", lineno)
+
+        if trimmed is None:
+            trimmed = self.environment.policies["ext.i18n.trimmed"]
+        if trimmed:
+            singular = self._trim_whitespace(singular)
+            if plural:
+                plural = self._trim_whitespace(plural)
+
+        node = self._make_node(
+            singular,
+            plural,
+            context,
+            variables,
+            plural_expr,
+            bool(referenced),
+            num_called_num and have_plural,
+        )
+        node.set_lineno(lineno)
+        if plural_expr_assignment is not None:
+            return [plural_expr_assignment, node]
+        else:
+            return node
+
+    def _trim_whitespace(self, string: str, _ws_re: t.Pattern[str] = _ws_re) -> str:
+        return _ws_re.sub(" ", string.strip())
+
+    def _parse_block(
+        self, parser: "Parser", allow_pluralize: bool
+    ) -> t.Tuple[t.List[str], str]:
+        """Parse until the next block tag with a given name."""
+        referenced = []
+        buf = []
+
+        while True:
+            if parser.stream.current.type == "data":
+                buf.append(parser.stream.current.value.replace("%", "%%"))
+                next(parser.stream)
+            elif parser.stream.current.type == "variable_begin":
+                next(parser.stream)
+                name = parser.stream.expect("name").value
+                referenced.append(name)
+                buf.append(f"%({name})s")
+                parser.stream.expect("variable_end")
+            elif parser.stream.current.type == "block_begin":
+                next(parser.stream)
+                block_name = (
+                    parser.stream.current.value
+                    if parser.stream.current.type == "name"
+                    else None
+                )
+                if block_name == "endtrans":
+                    break
+                elif block_name == "pluralize":
+                    if allow_pluralize:
+                        break
+                    parser.fail(
+                        "a translatable section can have only one pluralize section"
+                    )
+                elif block_name == "trans":
+                    parser.fail(
+                        "trans blocks can't be nested; did you mean `endtrans`?"
+                    )
+                parser.fail(
+                    f"control structures in translatable sections are not allowed; "
+                    f"saw `{block_name}`"
+                )
+            elif parser.stream.eos:
+                parser.fail("unclosed translation block")
+            else:
+                raise RuntimeError("internal parser error")
+
+        return referenced, concat(buf)
+
+    def _make_node(
+        self,
+        singular: str,
+        plural: t.Optional[str],
+        context: t.Optional[str],
+        variables: t.Dict[str, nodes.Expr],
+        plural_expr: t.Optional[nodes.Expr],
+        vars_referenced: bool,
+        num_called_num: bool,
+    ) -> nodes.Output:
+        """Generates a useful node from the data provided."""
+        newstyle = self.environment.newstyle_gettext  # type: ignore
+        node: nodes.Expr
+
+        # no variables referenced?  no need to escape for old style
+        # gettext invocations only if there are vars.
+        if not vars_referenced and not newstyle:
+            singular = singular.replace("%%", "%")
+            if plural:
+                plural = plural.replace("%%", "%")
+
+        func_name = "gettext"
+        func_args: t.List[nodes.Expr] = [nodes.Const(singular)]
+
+        if context is not None:
+            func_args.insert(0, nodes.Const(context))
+            func_name = f"p{func_name}"
+
+        if plural_expr is not None:
+            func_name = f"n{func_name}"
+            func_args.extend((nodes.Const(plural), plural_expr))
+
+        node = nodes.Call(nodes.Name(func_name, "load"), func_args, [], None, None)
+
+        # in case newstyle gettext is used, the method is powerful
+        # enough to handle the variable expansion and autoescape
+        # handling itself
+        if newstyle:
+            for key, value in variables.items():
+                # the function adds that later anyways in case num was
+                # called num, so just skip it.
+                if num_called_num and key == "num":
+                    continue
+                node.kwargs.append(nodes.Keyword(key, value))
+
+        # otherwise do that here
+        else:
+            # mark the return value as safe if we are in an
+            # environment with autoescaping turned on
+            node = nodes.MarkSafeIfAutoescape(node)
+            if variables:
+                node = nodes.Mod(
+                    node,
+                    nodes.Dict(
+                        [
+                            nodes.Pair(nodes.Const(key), value)
+                            for key, value in variables.items()
+                        ]
+                    ),
+                )
+        return nodes.Output([node])
+
+
+class ExprStmtExtension(Extension):
+    """Adds a `do` tag to Jinja that works like the print statement just
+    that it doesn't print the return value.
+    """
+
+    tags = {"do"}
+
+    def parse(self, parser: "Parser") -> nodes.ExprStmt:
+        node = nodes.ExprStmt(lineno=next(parser.stream).lineno)
+        node.node = parser.parse_tuple()
+        return node
+
+
+class LoopControlExtension(Extension):
+    """Adds break and continue to the template engine."""
+
+    tags = {"break", "continue"}
+
+    def parse(self, parser: "Parser") -> t.Union[nodes.Break, nodes.Continue]:
+        token = next(parser.stream)
+        if token.value == "break":
+            return nodes.Break(lineno=token.lineno)
+        return nodes.Continue(lineno=token.lineno)
+
+
+class DebugExtension(Extension):
+    """A ``{% debug %}`` tag that dumps the available variables,
+    filters, and tests.
+
+    .. code-block:: html+jinja
+
+        <pre>{% debug %}</pre>
+
+    .. code-block:: text
+
+        {'context': {'cycler': <class 'jinja2.utils.Cycler'>,
+                     ...,
+                     'namespace': <class 'jinja2.utils.Namespace'>},
+         'filters': ['abs', 'attr', 'batch', 'capitalize', 'center', 'count', 'd',
+                     ..., 'urlencode', 'urlize', 'wordcount', 'wordwrap', 'xmlattr'],
+         'tests': ['!=', '<', '<=', '==', '>', '>=', 'callable', 'defined',
+                   ..., 'odd', 'sameas', 'sequence', 'string', 'undefined', 'upper']}
+
+    .. versionadded:: 2.11.0
+    """
+
+    tags = {"debug"}
+
+    def parse(self, parser: "Parser") -> nodes.Output:
+        lineno = parser.stream.expect("name:debug").lineno
+        context = nodes.ContextReference()
+        result = self.call_method("_render", [context], lineno=lineno)
+        return nodes.Output([result], lineno=lineno)
+
+    def _render(self, context: Context) -> str:
+        result = {
+            "context": context.get_all(),
+            "filters": sorted(self.environment.filters.keys()),
+            "tests": sorted(self.environment.tests.keys()),
+        }
+
+        # Set the depth since the intent is to show the top few names.
+        return pprint.pformat(result, depth=3, compact=True)
+
+
+def extract_from_ast(
+    ast: nodes.Template,
+    gettext_functions: t.Sequence[str] = GETTEXT_FUNCTIONS,
+    babel_style: bool = True,
+) -> t.Iterator[
+    t.Tuple[int, str, t.Union[t.Optional[str], t.Tuple[t.Optional[str], ...]]]
+]:
+    """Extract localizable strings from the given template node.  Per
+    default this function returns matches in babel style that means non string
+    parameters as well as keyword arguments are returned as `None`.  This
+    allows Babel to figure out what you really meant if you are using
+    gettext functions that allow keyword arguments for placeholder expansion.
+    If you don't want that behavior set the `babel_style` parameter to `False`
+    which causes only strings to be returned and parameters are always stored
+    in tuples.  As a consequence invalid gettext calls (calls without a single
+    string parameter or string parameters after non-string parameters) are
+    skipped.
+
+    This example explains the behavior:
+
+    >>> from jinja2 import Environment
+    >>> env = Environment()
+    >>> node = env.parse('{{ (_("foo"), _(), ngettext("foo", "bar", 42)) }}')
+    >>> list(extract_from_ast(node))
+    [(1, '_', 'foo'), (1, '_', ()), (1, 'ngettext', ('foo', 'bar', None))]
+    >>> list(extract_from_ast(node, babel_style=False))
+    [(1, '_', ('foo',)), (1, 'ngettext', ('foo', 'bar'))]
+
+    For every string found this function yields a ``(lineno, function,
+    message)`` tuple, where:
+
+    * ``lineno`` is the number of the line on which the string was found,
+    * ``function`` is the name of the ``gettext`` function used (if the
+      string was extracted from embedded Python code), and
+    *   ``message`` is the string, or a tuple of strings for functions
+         with multiple string arguments.
+
+    This extraction function operates on the AST and is because of that unable
+    to extract any comments.  For comment support you have to use the babel
+    extraction interface or extract comments yourself.
+    """
+    out: t.Union[t.Optional[str], t.Tuple[t.Optional[str], ...]]
+
+    for node in ast.find_all(nodes.Call):
+        if (
+            not isinstance(node.node, nodes.Name)
+            or node.node.name not in gettext_functions
+        ):
+            continue
+
+        strings: t.List[t.Optional[str]] = []
+
+        for arg in node.args:
+            if isinstance(arg, nodes.Const) and isinstance(arg.value, str):
+                strings.append(arg.value)
+            else:
+                strings.append(None)
+
+        for _ in node.kwargs:
+            strings.append(None)
+        if node.dyn_args is not None:
+            strings.append(None)
+        if node.dyn_kwargs is not None:
+            strings.append(None)
+
+        if not babel_style:
+            out = tuple(x for x in strings if x is not None)
+
+            if not out:
+                continue
+        else:
+            if len(strings) == 1:
+                out = strings[0]
+            else:
+                out = tuple(strings)
+
+        yield node.lineno, node.node.name, out
+
+
+class _CommentFinder:
+    """Helper class to find comments in a token stream.  Can only
+    find comments for gettext calls forwards.  Once the comment
+    from line 4 is found, a comment for line 1 will not return a
+    usable value.
+    """
+
+    def __init__(
+        self, tokens: t.Sequence[t.Tuple[int, str, str]], comment_tags: t.Sequence[str]
+    ) -> None:
+        self.tokens = tokens
+        self.comment_tags = comment_tags
+        self.offset = 0
+        self.last_lineno = 0
+
+    def find_backwards(self, offset: int) -> t.List[str]:
+        try:
+            for _, token_type, token_value in reversed(
+                self.tokens[self.offset : offset]
+            ):
+                if token_type in ("comment", "linecomment"):
+                    try:
+                        prefix, comment = token_value.split(None, 1)
+                    except ValueError:
+                        continue
+                    if prefix in self.comment_tags:
+                        return [comment.rstrip()]
+            return []
+        finally:
+            self.offset = offset
+
+    def find_comments(self, lineno: int) -> t.List[str]:
+        if not self.comment_tags or self.last_lineno > lineno:
+            return []
+        for idx, (token_lineno, _, _) in enumerate(self.tokens[self.offset :]):
+            if token_lineno > lineno:
+                return self.find_backwards(self.offset + idx)
+        return self.find_backwards(len(self.tokens))
+
+
+def babel_extract(
+    fileobj: t.BinaryIO,
+    keywords: t.Sequence[str],
+    comment_tags: t.Sequence[str],
+    options: t.Dict[str, t.Any],
+) -> t.Iterator[
+    t.Tuple[
+        int, str, t.Union[t.Optional[str], t.Tuple[t.Optional[str], ...]], t.List[str]
+    ]
+]:
+    """Babel extraction method for Jinja templates.
+
+    .. versionchanged:: 2.3
+       Basic support for translation comments was added.  If `comment_tags`
+       is now set to a list of keywords for extraction, the extractor will
+       try to find the best preceding comment that begins with one of the
+       keywords.  For best results, make sure to not have more than one
+       gettext call in one line of code and the matching comment in the
+       same line or the line before.
+
+    .. versionchanged:: 2.5.1
+       The `newstyle_gettext` flag can be set to `True` to enable newstyle
+       gettext calls.
+
+    .. versionchanged:: 2.7
+       A `silent` option can now be provided.  If set to `False` template
+       syntax errors are propagated instead of being ignored.
+
+    :param fileobj: the file-like object the messages should be extracted from
+    :param keywords: a list of keywords (i.e. function names) that should be
+                     recognized as translation functions
+    :param comment_tags: a list of translator tags to search for and include
+                         in the results.
+    :param options: a dictionary of additional options (optional)
+    :return: an iterator over ``(lineno, funcname, message, comments)`` tuples.
+             (comments will be empty currently)
+    """
+    extensions: t.Dict[t.Type[Extension], None] = {}
+
+    for extension_name in options.get("extensions", "").split(","):
+        extension_name = extension_name.strip()
+
+        if not extension_name:
+            continue
+
+        extensions[import_string(extension_name)] = None
+
+    if InternationalizationExtension not in extensions:
+        extensions[InternationalizationExtension] = None
+
+    def getbool(options: t.Mapping[str, str], key: str, default: bool = False) -> bool:
+        return options.get(key, str(default)).lower() in {"1", "on", "yes", "true"}
+
+    silent = getbool(options, "silent", True)
+    environment = Environment(
+        options.get("block_start_string", defaults.BLOCK_START_STRING),
+        options.get("block_end_string", defaults.BLOCK_END_STRING),
+        options.get("variable_start_string", defaults.VARIABLE_START_STRING),
+        options.get("variable_end_string", defaults.VARIABLE_END_STRING),
+        options.get("comment_start_string", defaults.COMMENT_START_STRING),
+        options.get("comment_end_string", defaults.COMMENT_END_STRING),
+        options.get("line_statement_prefix") or defaults.LINE_STATEMENT_PREFIX,
+        options.get("line_comment_prefix") or defaults.LINE_COMMENT_PREFIX,
+        getbool(options, "trim_blocks", defaults.TRIM_BLOCKS),
+        getbool(options, "lstrip_blocks", defaults.LSTRIP_BLOCKS),
+        defaults.NEWLINE_SEQUENCE,
+        getbool(options, "keep_trailing_newline", defaults.KEEP_TRAILING_NEWLINE),
+        tuple(extensions),
+        cache_size=0,
+        auto_reload=False,
+    )
+
+    if getbool(options, "trimmed"):
+        environment.policies["ext.i18n.trimmed"] = True
+    if getbool(options, "newstyle_gettext"):
+        environment.newstyle_gettext = True  # type: ignore
+
+    source = fileobj.read().decode(options.get("encoding", "utf-8"))
+    try:
+        node = environment.parse(source)
+        tokens = list(environment.lex(environment.preprocess(source)))
+    except TemplateSyntaxError:
+        if not silent:
+            raise
+        # skip templates with syntax errors
+        return
+
+    finder = _CommentFinder(tokens, comment_tags)
+    for lineno, func, message in extract_from_ast(node, keywords):
+        yield lineno, func, message, finder.find_comments(lineno)
+
+
+#: nicer import names
+i18n = InternationalizationExtension
+do = ExprStmtExtension
+loopcontrols = LoopControlExtension
+debug = DebugExtension
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/filters.py b/backend/venv/lib/python3.12/site-packages/jinja2/filters.py
new file mode 100644
index 00000000..e5b5a00c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/filters.py
@@ -0,0 +1,1878 @@
+"""Built-in template filters used with the ``|`` operator."""
+
+import math
+import random
+import re
+import typing
+import typing as t
+from collections import abc
+from itertools import chain
+from itertools import groupby
+
+from markupsafe import escape
+from markupsafe import Markup
+from markupsafe import soft_str
+
+from .async_utils import async_variant
+from .async_utils import auto_aiter
+from .async_utils import auto_await
+from .async_utils import auto_to_list
+from .exceptions import FilterArgumentError
+from .runtime import Undefined
+from .utils import htmlsafe_json_dumps
+from .utils import pass_context
+from .utils import pass_environment
+from .utils import pass_eval_context
+from .utils import pformat
+from .utils import url_quote
+from .utils import urlize
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+    from .nodes import EvalContext
+    from .runtime import Context
+    from .sandbox import SandboxedEnvironment  # noqa: F401
+
+    class HasHTML(te.Protocol):
+        def __html__(self) -> str:
+            pass
+
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+K = t.TypeVar("K")
+V = t.TypeVar("V")
+
+
+def ignore_case(value: V) -> V:
+    """For use as a postprocessor for :func:`make_attrgetter`. Converts strings
+    to lowercase and returns other types as-is."""
+    if isinstance(value, str):
+        return t.cast(V, value.lower())
+
+    return value
+
+
+def make_attrgetter(
+    environment: "Environment",
+    attribute: t.Optional[t.Union[str, int]],
+    postprocess: t.Optional[t.Callable[[t.Any], t.Any]] = None,
+    default: t.Optional[t.Any] = None,
+) -> t.Callable[[t.Any], t.Any]:
+    """Returns a callable that looks up the given attribute from a
+    passed object with the rules of the environment.  Dots are allowed
+    to access attributes of attributes.  Integer parts in paths are
+    looked up as integers.
+    """
+    parts = _prepare_attribute_parts(attribute)
+
+    def attrgetter(item: t.Any) -> t.Any:
+        for part in parts:
+            item = environment.getitem(item, part)
+
+            if default is not None and isinstance(item, Undefined):
+                item = default
+
+        if postprocess is not None:
+            item = postprocess(item)
+
+        return item
+
+    return attrgetter
+
+
+def make_multi_attrgetter(
+    environment: "Environment",
+    attribute: t.Optional[t.Union[str, int]],
+    postprocess: t.Optional[t.Callable[[t.Any], t.Any]] = None,
+) -> t.Callable[[t.Any], t.List[t.Any]]:
+    """Returns a callable that looks up the given comma separated
+    attributes from a passed object with the rules of the environment.
+    Dots are allowed to access attributes of each attribute.  Integer
+    parts in paths are looked up as integers.
+
+    The value returned by the returned callable is a list of extracted
+    attribute values.
+
+    Examples of attribute: "attr1,attr2", "attr1.inner1.0,attr2.inner2.0", etc.
+    """
+    if isinstance(attribute, str):
+        split: t.Sequence[t.Union[str, int, None]] = attribute.split(",")
+    else:
+        split = [attribute]
+
+    parts = [_prepare_attribute_parts(item) for item in split]
+
+    def attrgetter(item: t.Any) -> t.List[t.Any]:
+        items = [None] * len(parts)
+
+        for i, attribute_part in enumerate(parts):
+            item_i = item
+
+            for part in attribute_part:
+                item_i = environment.getitem(item_i, part)
+
+            if postprocess is not None:
+                item_i = postprocess(item_i)
+
+            items[i] = item_i
+
+        return items
+
+    return attrgetter
+
+
+def _prepare_attribute_parts(
+    attr: t.Optional[t.Union[str, int]],
+) -> t.List[t.Union[str, int]]:
+    if attr is None:
+        return []
+
+    if isinstance(attr, str):
+        return [int(x) if x.isdigit() else x for x in attr.split(".")]
+
+    return [attr]
+
+
+def do_forceescape(value: "t.Union[str, HasHTML]") -> Markup:
+    """Enforce HTML escaping.  This will probably double escape variables."""
+    if hasattr(value, "__html__"):
+        value = t.cast("HasHTML", value).__html__()
+
+    return escape(str(value))
+
+
+def do_urlencode(
+    value: t.Union[str, t.Mapping[str, t.Any], t.Iterable[t.Tuple[str, t.Any]]],
+) -> str:
+    """Quote data for use in a URL path or query using UTF-8.
+
+    Basic wrapper around :func:`urllib.parse.quote` when given a
+    string, or :func:`urllib.parse.urlencode` for a dict or iterable.
+
+    :param value: Data to quote. A string will be quoted directly. A
+        dict or iterable of ``(key, value)`` pairs will be joined as a
+        query string.
+
+    When given a string, "/" is not quoted. HTTP servers treat "/" and
+    "%2F" equivalently in paths. If you need quoted slashes, use the
+    ``|replace("/", "%2F")`` filter.
+
+    .. versionadded:: 2.7
+    """
+    if isinstance(value, str) or not isinstance(value, abc.Iterable):
+        return url_quote(value)
+
+    if isinstance(value, dict):
+        items: t.Iterable[t.Tuple[str, t.Any]] = value.items()
+    else:
+        items = value  # type: ignore
+
+    return "&".join(
+        f"{url_quote(k, for_qs=True)}={url_quote(v, for_qs=True)}" for k, v in items
+    )
+
+
+@pass_eval_context
+def do_replace(
+    eval_ctx: "EvalContext", s: str, old: str, new: str, count: t.Optional[int] = None
+) -> str:
+    """Return a copy of the value with all occurrences of a substring
+    replaced with a new one. The first argument is the substring
+    that should be replaced, the second is the replacement string.
+    If the optional third argument ``count`` is given, only the first
+    ``count`` occurrences are replaced:
+
+    .. sourcecode:: jinja
+
+        {{ "Hello World"|replace("Hello", "Goodbye") }}
+            -> Goodbye World
+
+        {{ "aaaaargh"|replace("a", "d'oh, ", 2) }}
+            -> d'oh, d'oh, aaargh
+    """
+    if count is None:
+        count = -1
+
+    if not eval_ctx.autoescape:
+        return str(s).replace(str(old), str(new), count)
+
+    if (
+        hasattr(old, "__html__")
+        or hasattr(new, "__html__")
+        and not hasattr(s, "__html__")
+    ):
+        s = escape(s)
+    else:
+        s = soft_str(s)
+
+    return s.replace(soft_str(old), soft_str(new), count)
+
+
+def do_upper(s: str) -> str:
+    """Convert a value to uppercase."""
+    return soft_str(s).upper()
+
+
+def do_lower(s: str) -> str:
+    """Convert a value to lowercase."""
+    return soft_str(s).lower()
+
+
+def do_items(value: t.Union[t.Mapping[K, V], Undefined]) -> t.Iterator[t.Tuple[K, V]]:
+    """Return an iterator over the ``(key, value)`` items of a mapping.
+
+    ``x|items`` is the same as ``x.items()``, except if ``x`` is
+    undefined an empty iterator is returned.
+
+    This filter is useful if you expect the template to be rendered with
+    an implementation of Jinja in another programming language that does
+    not have a ``.items()`` method on its mapping type.
+
+    .. code-block:: html+jinja
+
+        <dl>
+        {% for key, value in my_dict|items %}
+            <dt>{{ key }}
+            <dd>{{ value }}
+        {% endfor %}
+        </dl>
+
+    .. versionadded:: 3.1
+    """
+    if isinstance(value, Undefined):
+        return
+
+    if not isinstance(value, abc.Mapping):
+        raise TypeError("Can only get item pairs from a mapping.")
+
+    yield from value.items()
+
+
+# Check for characters that would move the parser state from key to value.
+# https://html.spec.whatwg.org/#attribute-name-state
+_attr_key_re = re.compile(r"[\s/>=]", flags=re.ASCII)
+
+
+@pass_eval_context
+def do_xmlattr(
+    eval_ctx: "EvalContext", d: t.Mapping[str, t.Any], autospace: bool = True
+) -> str:
+    """Create an SGML/XML attribute string based on the items in a dict.
+
+    **Values** that are neither ``none`` nor ``undefined`` are automatically
+    escaped, safely allowing untrusted user input.
+
+    User input should not be used as **keys** to this filter. If any key
+    contains a space, ``/`` solidus, ``>`` greater-than sign, or ``=`` equals
+    sign, this fails with a ``ValueError``. Regardless of this, user input
+    should never be used as keys to this filter, or must be separately validated
+    first.
+
+    .. sourcecode:: html+jinja
+
+        <ul{{ {'class': 'my_list', 'missing': none,
+                'id': 'list-%d'|format(variable)}|xmlattr }}>
+        ...
+        </ul>
+
+    Results in something like this:
+
+    .. sourcecode:: html
+
+        <ul class="my_list" id="list-42">
+        ...
+        </ul>
+
+    As you can see it automatically prepends a space in front of the item
+    if the filter returned something unless the second parameter is false.
+
+    .. versionchanged:: 3.1.4
+        Keys with ``/`` solidus, ``>`` greater-than sign, or ``=`` equals sign
+        are not allowed.
+
+    .. versionchanged:: 3.1.3
+        Keys with spaces are not allowed.
+    """
+    items = []
+
+    for key, value in d.items():
+        if value is None or isinstance(value, Undefined):
+            continue
+
+        if _attr_key_re.search(key) is not None:
+            raise ValueError(f"Invalid character in attribute name: {key!r}")
+
+        items.append(f'{escape(key)}="{escape(value)}"')
+
+    rv = " ".join(items)
+
+    if autospace and rv:
+        rv = " " + rv
+
+    if eval_ctx.autoescape:
+        rv = Markup(rv)
+
+    return rv
+
+
+def do_capitalize(s: str) -> str:
+    """Capitalize a value. The first character will be uppercase, all others
+    lowercase.
+    """
+    return soft_str(s).capitalize()
+
+
+_word_beginning_split_re = re.compile(r"([-\s({\[<]+)")
+
+
+def do_title(s: str) -> str:
+    """Return a titlecased version of the value. I.e. words will start with
+    uppercase letters, all remaining characters are lowercase.
+    """
+    return "".join(
+        [
+            item[0].upper() + item[1:].lower()
+            for item in _word_beginning_split_re.split(soft_str(s))
+            if item
+        ]
+    )
+
+
+def do_dictsort(
+    value: t.Mapping[K, V],
+    case_sensitive: bool = False,
+    by: 'te.Literal["key", "value"]' = "key",
+    reverse: bool = False,
+) -> t.List[t.Tuple[K, V]]:
+    """Sort a dict and yield (key, value) pairs. Python dicts may not
+    be in the order you want to display them in, so sort them first.
+
+    .. sourcecode:: jinja
+
+        {% for key, value in mydict|dictsort %}
+            sort the dict by key, case insensitive
+
+        {% for key, value in mydict|dictsort(reverse=true) %}
+            sort the dict by key, case insensitive, reverse order
+
+        {% for key, value in mydict|dictsort(true) %}
+            sort the dict by key, case sensitive
+
+        {% for key, value in mydict|dictsort(false, 'value') %}
+            sort the dict by value, case insensitive
+    """
+    if by == "key":
+        pos = 0
+    elif by == "value":
+        pos = 1
+    else:
+        raise FilterArgumentError('You can only sort by either "key" or "value"')
+
+    def sort_func(item: t.Tuple[t.Any, t.Any]) -> t.Any:
+        value = item[pos]
+
+        if not case_sensitive:
+            value = ignore_case(value)
+
+        return value
+
+    return sorted(value.items(), key=sort_func, reverse=reverse)
+
+
+@pass_environment
+def do_sort(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    reverse: bool = False,
+    case_sensitive: bool = False,
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> "t.List[V]":
+    """Sort an iterable using Python's :func:`sorted`.
+
+    .. sourcecode:: jinja
+
+        {% for city in cities|sort %}
+            ...
+        {% endfor %}
+
+    :param reverse: Sort descending instead of ascending.
+    :param case_sensitive: When sorting strings, sort upper and lower
+        case separately.
+    :param attribute: When sorting objects or dicts, an attribute or
+        key to sort by. Can use dot notation like ``"address.city"``.
+        Can be a list of attributes like ``"age,name"``.
+
+    The sort is stable, it does not change the relative order of
+    elements that compare equal. This makes it is possible to chain
+    sorts on different attributes and ordering.
+
+    .. sourcecode:: jinja
+
+        {% for user in users|sort(attribute="name")
+            |sort(reverse=true, attribute="age") %}
+            ...
+        {% endfor %}
+
+    As a shortcut to chaining when the direction is the same for all
+    attributes, pass a comma separate list of attributes.
+
+    .. sourcecode:: jinja
+
+        {% for user in users|sort(attribute="age,name") %}
+            ...
+        {% endfor %}
+
+    .. versionchanged:: 2.11.0
+        The ``attribute`` parameter can be a comma separated list of
+        attributes, e.g. ``"age,name"``.
+
+    .. versionchanged:: 2.6
+       The ``attribute`` parameter was added.
+    """
+    key_func = make_multi_attrgetter(
+        environment, attribute, postprocess=ignore_case if not case_sensitive else None
+    )
+    return sorted(value, key=key_func, reverse=reverse)
+
+
+@pass_environment
+def sync_do_unique(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    case_sensitive: bool = False,
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> "t.Iterator[V]":
+    """Returns a list of unique items from the given iterable.
+
+    .. sourcecode:: jinja
+
+        {{ ['foo', 'bar', 'foobar', 'FooBar']|unique|list }}
+            -> ['foo', 'bar', 'foobar']
+
+    The unique items are yielded in the same order as their first occurrence in
+    the iterable passed to the filter.
+
+    :param case_sensitive: Treat upper and lower case strings as distinct.
+    :param attribute: Filter objects with unique values for this attribute.
+    """
+    getter = make_attrgetter(
+        environment, attribute, postprocess=ignore_case if not case_sensitive else None
+    )
+    seen = set()
+
+    for item in value:
+        key = getter(item)
+
+        if key not in seen:
+            seen.add(key)
+            yield item
+
+
+@async_variant(sync_do_unique)  # type: ignore
+async def do_unique(
+    environment: "Environment",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    case_sensitive: bool = False,
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> "t.Iterator[V]":
+    return sync_do_unique(
+        environment, await auto_to_list(value), case_sensitive, attribute
+    )
+
+
+def _min_or_max(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    func: "t.Callable[..., V]",
+    case_sensitive: bool,
+    attribute: t.Optional[t.Union[str, int]],
+) -> "t.Union[V, Undefined]":
+    it = iter(value)
+
+    try:
+        first = next(it)
+    except StopIteration:
+        return environment.undefined("No aggregated item, sequence was empty.")
+
+    key_func = make_attrgetter(
+        environment, attribute, postprocess=ignore_case if not case_sensitive else None
+    )
+    return func(chain([first], it), key=key_func)
+
+
+@pass_environment
+def do_min(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    case_sensitive: bool = False,
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> "t.Union[V, Undefined]":
+    """Return the smallest item from the sequence.
+
+    .. sourcecode:: jinja
+
+        {{ [1, 2, 3]|min }}
+            -> 1
+
+    :param case_sensitive: Treat upper and lower case strings as distinct.
+    :param attribute: Get the object with the min value of this attribute.
+    """
+    return _min_or_max(environment, value, min, case_sensitive, attribute)
+
+
+@pass_environment
+def do_max(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    case_sensitive: bool = False,
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> "t.Union[V, Undefined]":
+    """Return the largest item from the sequence.
+
+    .. sourcecode:: jinja
+
+        {{ [1, 2, 3]|max }}
+            -> 3
+
+    :param case_sensitive: Treat upper and lower case strings as distinct.
+    :param attribute: Get the object with the max value of this attribute.
+    """
+    return _min_or_max(environment, value, max, case_sensitive, attribute)
+
+
+def do_default(
+    value: V,
+    default_value: V = "",  # type: ignore
+    boolean: bool = False,
+) -> V:
+    """If the value is undefined it will return the passed default value,
+    otherwise the value of the variable:
+
+    .. sourcecode:: jinja
+
+        {{ my_variable|default('my_variable is not defined') }}
+
+    This will output the value of ``my_variable`` if the variable was
+    defined, otherwise ``'my_variable is not defined'``. If you want
+    to use default with variables that evaluate to false you have to
+    set the second parameter to `true`:
+
+    .. sourcecode:: jinja
+
+        {{ ''|default('the string was empty', true) }}
+
+    .. versionchanged:: 2.11
+       It's now possible to configure the :class:`~jinja2.Environment` with
+       :class:`~jinja2.ChainableUndefined` to make the `default` filter work
+       on nested elements and attributes that may contain undefined values
+       in the chain without getting an :exc:`~jinja2.UndefinedError`.
+    """
+    if isinstance(value, Undefined) or (boolean and not value):
+        return default_value
+
+    return value
+
+
+@pass_eval_context
+def sync_do_join(
+    eval_ctx: "EvalContext",
+    value: t.Iterable[t.Any],
+    d: str = "",
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> str:
+    """Return a string which is the concatenation of the strings in the
+    sequence. The separator between elements is an empty string per
+    default, you can define it with the optional parameter:
+
+    .. sourcecode:: jinja
+
+        {{ [1, 2, 3]|join('|') }}
+            -> 1|2|3
+
+        {{ [1, 2, 3]|join }}
+            -> 123
+
+    It is also possible to join certain attributes of an object:
+
+    .. sourcecode:: jinja
+
+        {{ users|join(', ', attribute='username') }}
+
+    .. versionadded:: 2.6
+       The `attribute` parameter was added.
+    """
+    if attribute is not None:
+        value = map(make_attrgetter(eval_ctx.environment, attribute), value)
+
+    # no automatic escaping?  joining is a lot easier then
+    if not eval_ctx.autoescape:
+        return str(d).join(map(str, value))
+
+    # if the delimiter doesn't have an html representation we check
+    # if any of the items has.  If yes we do a coercion to Markup
+    if not hasattr(d, "__html__"):
+        value = list(value)
+        do_escape = False
+
+        for idx, item in enumerate(value):
+            if hasattr(item, "__html__"):
+                do_escape = True
+            else:
+                value[idx] = str(item)
+
+        if do_escape:
+            d = escape(d)
+        else:
+            d = str(d)
+
+        return d.join(value)
+
+    # no html involved, to normal joining
+    return soft_str(d).join(map(soft_str, value))
+
+
+@async_variant(sync_do_join)  # type: ignore
+async def do_join(
+    eval_ctx: "EvalContext",
+    value: t.Union[t.AsyncIterable[t.Any], t.Iterable[t.Any]],
+    d: str = "",
+    attribute: t.Optional[t.Union[str, int]] = None,
+) -> str:
+    return sync_do_join(eval_ctx, await auto_to_list(value), d, attribute)
+
+
+def do_center(value: str, width: int = 80) -> str:
+    """Centers the value in a field of a given width."""
+    return soft_str(value).center(width)
+
+
+@pass_environment
+def sync_do_first(
+    environment: "Environment", seq: "t.Iterable[V]"
+) -> "t.Union[V, Undefined]":
+    """Return the first item of a sequence."""
+    try:
+        return next(iter(seq))
+    except StopIteration:
+        return environment.undefined("No first item, sequence was empty.")
+
+
+@async_variant(sync_do_first)  # type: ignore
+async def do_first(
+    environment: "Environment", seq: "t.Union[t.AsyncIterable[V], t.Iterable[V]]"
+) -> "t.Union[V, Undefined]":
+    try:
+        return await auto_aiter(seq).__anext__()
+    except StopAsyncIteration:
+        return environment.undefined("No first item, sequence was empty.")
+
+
+@pass_environment
+def do_last(
+    environment: "Environment", seq: "t.Reversible[V]"
+) -> "t.Union[V, Undefined]":
+    """Return the last item of a sequence.
+
+    Note: Does not work with generators. You may want to explicitly
+    convert it to a list:
+
+    .. sourcecode:: jinja
+
+        {{ data | selectattr('name', '==', 'Jinja') | list | last }}
+    """
+    try:
+        return next(iter(reversed(seq)))
+    except StopIteration:
+        return environment.undefined("No last item, sequence was empty.")
+
+
+# No async do_last, it may not be safe in async mode.
+
+
+@pass_context
+def do_random(context: "Context", seq: "t.Sequence[V]") -> "t.Union[V, Undefined]":
+    """Return a random item from the sequence."""
+    try:
+        return random.choice(seq)
+    except IndexError:
+        return context.environment.undefined("No random item, sequence was empty.")
+
+
+def do_filesizeformat(value: t.Union[str, float, int], binary: bool = False) -> str:
+    """Format the value like a 'human-readable' file size (i.e. 13 kB,
+    4.1 MB, 102 Bytes, etc).  Per default decimal prefixes are used (Mega,
+    Giga, etc.), if the second parameter is set to `True` the binary
+    prefixes are used (Mebi, Gibi).
+    """
+    bytes = float(value)
+    base = 1024 if binary else 1000
+    prefixes = [
+        ("KiB" if binary else "kB"),
+        ("MiB" if binary else "MB"),
+        ("GiB" if binary else "GB"),
+        ("TiB" if binary else "TB"),
+        ("PiB" if binary else "PB"),
+        ("EiB" if binary else "EB"),
+        ("ZiB" if binary else "ZB"),
+        ("YiB" if binary else "YB"),
+    ]
+
+    if bytes == 1:
+        return "1 Byte"
+    elif bytes < base:
+        return f"{int(bytes)} Bytes"
+    else:
+        for i, prefix in enumerate(prefixes):
+            unit = base ** (i + 2)
+
+            if bytes < unit:
+                return f"{base * bytes / unit:.1f} {prefix}"
+
+        return f"{base * bytes / unit:.1f} {prefix}"
+
+
+def do_pprint(value: t.Any) -> str:
+    """Pretty print a variable. Useful for debugging."""
+    return pformat(value)
+
+
+_uri_scheme_re = re.compile(r"^([\w.+-]{2,}:(/){0,2})$")
+
+
+@pass_eval_context
+def do_urlize(
+    eval_ctx: "EvalContext",
+    value: str,
+    trim_url_limit: t.Optional[int] = None,
+    nofollow: bool = False,
+    target: t.Optional[str] = None,
+    rel: t.Optional[str] = None,
+    extra_schemes: t.Optional[t.Iterable[str]] = None,
+) -> str:
+    """Convert URLs in text into clickable links.
+
+    This may not recognize links in some situations. Usually, a more
+    comprehensive formatter, such as a Markdown library, is a better
+    choice.
+
+    Works on ``http://``, ``https://``, ``www.``, ``mailto:``, and email
+    addresses. Links with trailing punctuation (periods, commas, closing
+    parentheses) and leading punctuation (opening parentheses) are
+    recognized excluding the punctuation. Email addresses that include
+    header fields are not recognized (for example,
+    ``mailto:address@example.com?cc=copy@example.com``).
+
+    :param value: Original text containing URLs to link.
+    :param trim_url_limit: Shorten displayed URL values to this length.
+    :param nofollow: Add the ``rel=nofollow`` attribute to links.
+    :param target: Add the ``target`` attribute to links.
+    :param rel: Add the ``rel`` attribute to links.
+    :param extra_schemes: Recognize URLs that start with these schemes
+        in addition to the default behavior. Defaults to
+        ``env.policies["urlize.extra_schemes"]``, which defaults to no
+        extra schemes.
+
+    .. versionchanged:: 3.0
+        The ``extra_schemes`` parameter was added.
+
+    .. versionchanged:: 3.0
+        Generate ``https://`` links for URLs without a scheme.
+
+    .. versionchanged:: 3.0
+        The parsing rules were updated. Recognize email addresses with
+        or without the ``mailto:`` scheme. Validate IP addresses. Ignore
+        parentheses and brackets in more cases.
+
+    .. versionchanged:: 2.8
+       The ``target`` parameter was added.
+    """
+    policies = eval_ctx.environment.policies
+    rel_parts = set((rel or "").split())
+
+    if nofollow:
+        rel_parts.add("nofollow")
+
+    rel_parts.update((policies["urlize.rel"] or "").split())
+    rel = " ".join(sorted(rel_parts)) or None
+
+    if target is None:
+        target = policies["urlize.target"]
+
+    if extra_schemes is None:
+        extra_schemes = policies["urlize.extra_schemes"] or ()
+
+    for scheme in extra_schemes:
+        if _uri_scheme_re.fullmatch(scheme) is None:
+            raise FilterArgumentError(f"{scheme!r} is not a valid URI scheme prefix.")
+
+    rv = urlize(
+        value,
+        trim_url_limit=trim_url_limit,
+        rel=rel,
+        target=target,
+        extra_schemes=extra_schemes,
+    )
+
+    if eval_ctx.autoescape:
+        rv = Markup(rv)
+
+    return rv
+
+
+def do_indent(
+    s: str, width: t.Union[int, str] = 4, first: bool = False, blank: bool = False
+) -> str:
+    """Return a copy of the string with each line indented by 4 spaces. The
+    first line and blank lines are not indented by default.
+
+    :param width: Number of spaces, or a string, to indent by.
+    :param first: Don't skip indenting the first line.
+    :param blank: Don't skip indenting empty lines.
+
+    .. versionchanged:: 3.0
+        ``width`` can be a string.
+
+    .. versionchanged:: 2.10
+        Blank lines are not indented by default.
+
+        Rename the ``indentfirst`` argument to ``first``.
+    """
+    if isinstance(width, str):
+        indention = width
+    else:
+        indention = " " * width
+
+    newline = "\n"
+
+    if isinstance(s, Markup):
+        indention = Markup(indention)
+        newline = Markup(newline)
+
+    s += newline  # this quirk is necessary for splitlines method
+
+    if blank:
+        rv = (newline + indention).join(s.splitlines())
+    else:
+        lines = s.splitlines()
+        rv = lines.pop(0)
+
+        if lines:
+            rv += newline + newline.join(
+                indention + line if line else line for line in lines
+            )
+
+    if first:
+        rv = indention + rv
+
+    return rv
+
+
+@pass_environment
+def do_truncate(
+    env: "Environment",
+    s: str,
+    length: int = 255,
+    killwords: bool = False,
+    end: str = "...",
+    leeway: t.Optional[int] = None,
+) -> str:
+    """Return a truncated copy of the string. The length is specified
+    with the first parameter which defaults to ``255``. If the second
+    parameter is ``true`` the filter will cut the text at length. Otherwise
+    it will discard the last word. If the text was in fact
+    truncated it will append an ellipsis sign (``"..."``). If you want a
+    different ellipsis sign than ``"..."`` you can specify it using the
+    third parameter. Strings that only exceed the length by the tolerance
+    margin given in the fourth parameter will not be truncated.
+
+    .. sourcecode:: jinja
+
+        {{ "foo bar baz qux"|truncate(9) }}
+            -> "foo..."
+        {{ "foo bar baz qux"|truncate(9, True) }}
+            -> "foo ba..."
+        {{ "foo bar baz qux"|truncate(11) }}
+            -> "foo bar baz qux"
+        {{ "foo bar baz qux"|truncate(11, False, '...', 0) }}
+            -> "foo bar..."
+
+    The default leeway on newer Jinja versions is 5 and was 0 before but
+    can be reconfigured globally.
+    """
+    if leeway is None:
+        leeway = env.policies["truncate.leeway"]
+
+    assert length >= len(end), f"expected length >= {len(end)}, got {length}"
+    assert leeway >= 0, f"expected leeway >= 0, got {leeway}"
+
+    if len(s) <= length + leeway:
+        return s
+
+    if killwords:
+        return s[: length - len(end)] + end
+
+    result = s[: length - len(end)].rsplit(" ", 1)[0]
+    return result + end
+
+
+@pass_environment
+def do_wordwrap(
+    environment: "Environment",
+    s: str,
+    width: int = 79,
+    break_long_words: bool = True,
+    wrapstring: t.Optional[str] = None,
+    break_on_hyphens: bool = True,
+) -> str:
+    """Wrap a string to the given width. Existing newlines are treated
+    as paragraphs to be wrapped separately.
+
+    :param s: Original text to wrap.
+    :param width: Maximum length of wrapped lines.
+    :param break_long_words: If a word is longer than ``width``, break
+        it across lines.
+    :param break_on_hyphens: If a word contains hyphens, it may be split
+        across lines.
+    :param wrapstring: String to join each wrapped line. Defaults to
+        :attr:`Environment.newline_sequence`.
+
+    .. versionchanged:: 2.11
+        Existing newlines are treated as paragraphs wrapped separately.
+
+    .. versionchanged:: 2.11
+        Added the ``break_on_hyphens`` parameter.
+
+    .. versionchanged:: 2.7
+        Added the ``wrapstring`` parameter.
+    """
+    import textwrap
+
+    if wrapstring is None:
+        wrapstring = environment.newline_sequence
+
+    # textwrap.wrap doesn't consider existing newlines when wrapping.
+    # If the string has a newline before width, wrap will still insert
+    # a newline at width, resulting in a short line. Instead, split and
+    # wrap each paragraph individually.
+    return wrapstring.join(
+        [
+            wrapstring.join(
+                textwrap.wrap(
+                    line,
+                    width=width,
+                    expand_tabs=False,
+                    replace_whitespace=False,
+                    break_long_words=break_long_words,
+                    break_on_hyphens=break_on_hyphens,
+                )
+            )
+            for line in s.splitlines()
+        ]
+    )
+
+
+_word_re = re.compile(r"\w+")
+
+
+def do_wordcount(s: str) -> int:
+    """Count the words in that string."""
+    return len(_word_re.findall(soft_str(s)))
+
+
+def do_int(value: t.Any, default: int = 0, base: int = 10) -> int:
+    """Convert the value into an integer. If the
+    conversion doesn't work it will return ``0``. You can
+    override this default using the first parameter. You
+    can also override the default base (10) in the second
+    parameter, which handles input with prefixes such as
+    0b, 0o and 0x for bases 2, 8 and 16 respectively.
+    The base is ignored for decimal numbers and non-string values.
+    """
+    try:
+        if isinstance(value, str):
+            return int(value, base)
+
+        return int(value)
+    except (TypeError, ValueError):
+        # this quirk is necessary so that "42.23"|int gives 42.
+        try:
+            return int(float(value))
+        except (TypeError, ValueError, OverflowError):
+            return default
+
+
+def do_float(value: t.Any, default: float = 0.0) -> float:
+    """Convert the value into a floating point number. If the
+    conversion doesn't work it will return ``0.0``. You can
+    override this default using the first parameter.
+    """
+    try:
+        return float(value)
+    except (TypeError, ValueError):
+        return default
+
+
+def do_format(value: str, *args: t.Any, **kwargs: t.Any) -> str:
+    """Apply the given values to a `printf-style`_ format string, like
+    ``string % values``.
+
+    .. sourcecode:: jinja
+
+        {{ "%s, %s!"|format(greeting, name) }}
+        Hello, World!
+
+    In most cases it should be more convenient and efficient to use the
+    ``%`` operator or :meth:`str.format`.
+
+    .. code-block:: text
+
+        {{ "%s, %s!" % (greeting, name) }}
+        {{ "{}, {}!".format(greeting, name) }}
+
+    .. _printf-style: https://docs.python.org/library/stdtypes.html
+        #printf-style-string-formatting
+    """
+    if args and kwargs:
+        raise FilterArgumentError(
+            "can't handle positional and keyword arguments at the same time"
+        )
+
+    return soft_str(value) % (kwargs or args)
+
+
+def do_trim(value: str, chars: t.Optional[str] = None) -> str:
+    """Strip leading and trailing characters, by default whitespace."""
+    return soft_str(value).strip(chars)
+
+
+def do_striptags(value: "t.Union[str, HasHTML]") -> str:
+    """Strip SGML/XML tags and replace adjacent whitespace by one space."""
+    if hasattr(value, "__html__"):
+        value = t.cast("HasHTML", value).__html__()
+
+    return Markup(str(value)).striptags()
+
+
+def sync_do_slice(
+    value: "t.Collection[V]", slices: int, fill_with: "t.Optional[V]" = None
+) -> "t.Iterator[t.List[V]]":
+    """Slice an iterator and return a list of lists containing
+    those items. Useful if you want to create a div containing
+    three ul tags that represent columns:
+
+    .. sourcecode:: html+jinja
+
+        <div class="columnwrapper">
+          {%- for column in items|slice(3) %}
+            <ul class="column-{{ loop.index }}">
+            {%- for item in column %}
+              <li>{{ item }}</li>
+            {%- endfor %}
+            </ul>
+          {%- endfor %}
+        </div>
+
+    If you pass it a second argument it's used to fill missing
+    values on the last iteration.
+    """
+    seq = list(value)
+    length = len(seq)
+    items_per_slice = length // slices
+    slices_with_extra = length % slices
+    offset = 0
+
+    for slice_number in range(slices):
+        start = offset + slice_number * items_per_slice
+
+        if slice_number < slices_with_extra:
+            offset += 1
+
+        end = offset + (slice_number + 1) * items_per_slice
+        tmp = seq[start:end]
+
+        if fill_with is not None and slice_number >= slices_with_extra:
+            tmp.append(fill_with)
+
+        yield tmp
+
+
+@async_variant(sync_do_slice)  # type: ignore
+async def do_slice(
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    slices: int,
+    fill_with: t.Optional[t.Any] = None,
+) -> "t.Iterator[t.List[V]]":
+    return sync_do_slice(await auto_to_list(value), slices, fill_with)
+
+
+def do_batch(
+    value: "t.Iterable[V]", linecount: int, fill_with: "t.Optional[V]" = None
+) -> "t.Iterator[t.List[V]]":
+    """
+    A filter that batches items. It works pretty much like `slice`
+    just the other way round. It returns a list of lists with the
+    given number of items. If you provide a second parameter this
+    is used to fill up missing items. See this example:
+
+    .. sourcecode:: html+jinja
+
+        <table>
+        {%- for row in items|batch(3, '&nbsp;') %}
+          <tr>
+          {%- for column in row %}
+            <td>{{ column }}</td>
+          {%- endfor %}
+          </tr>
+        {%- endfor %}
+        </table>
+    """
+    tmp: t.List[V] = []
+
+    for item in value:
+        if len(tmp) == linecount:
+            yield tmp
+            tmp = []
+
+        tmp.append(item)
+
+    if tmp:
+        if fill_with is not None and len(tmp) < linecount:
+            tmp += [fill_with] * (linecount - len(tmp))
+
+        yield tmp
+
+
+def do_round(
+    value: float,
+    precision: int = 0,
+    method: 'te.Literal["common", "ceil", "floor"]' = "common",
+) -> float:
+    """Round the number to a given precision. The first
+    parameter specifies the precision (default is ``0``), the
+    second the rounding method:
+
+    - ``'common'`` rounds either up or down
+    - ``'ceil'`` always rounds up
+    - ``'floor'`` always rounds down
+
+    If you don't specify a method ``'common'`` is used.
+
+    .. sourcecode:: jinja
+
+        {{ 42.55|round }}
+            -> 43.0
+        {{ 42.55|round(1, 'floor') }}
+            -> 42.5
+
+    Note that even if rounded to 0 precision, a float is returned.  If
+    you need a real integer, pipe it through `int`:
+
+    .. sourcecode:: jinja
+
+        {{ 42.55|round|int }}
+            -> 43
+    """
+    if method not in {"common", "ceil", "floor"}:
+        raise FilterArgumentError("method must be common, ceil or floor")
+
+    if method == "common":
+        return round(value, precision)
+
+    func = getattr(math, method)
+    return t.cast(float, func(value * (10**precision)) / (10**precision))
+
+
+class _GroupTuple(t.NamedTuple):
+    grouper: t.Any
+    list: t.List[t.Any]
+
+    # Use the regular tuple repr to hide this subclass if users print
+    # out the value during debugging.
+    def __repr__(self) -> str:
+        return tuple.__repr__(self)
+
+    def __str__(self) -> str:
+        return tuple.__str__(self)
+
+
+@pass_environment
+def sync_do_groupby(
+    environment: "Environment",
+    value: "t.Iterable[V]",
+    attribute: t.Union[str, int],
+    default: t.Optional[t.Any] = None,
+    case_sensitive: bool = False,
+) -> "t.List[_GroupTuple]":
+    """Group a sequence of objects by an attribute using Python's
+    :func:`itertools.groupby`. The attribute can use dot notation for
+    nested access, like ``"address.city"``. Unlike Python's ``groupby``,
+    the values are sorted first so only one group is returned for each
+    unique value.
+
+    For example, a list of ``User`` objects with a ``city`` attribute
+    can be rendered in groups. In this example, ``grouper`` refers to
+    the ``city`` value of the group.
+
+    .. sourcecode:: html+jinja
+
+        <ul>{% for city, items in users|groupby("city") %}
+          <li>{{ city }}
+            <ul>{% for user in items %}
+              <li>{{ user.name }}
+            {% endfor %}</ul>
+          </li>
+        {% endfor %}</ul>
+
+    ``groupby`` yields namedtuples of ``(grouper, list)``, which
+    can be used instead of the tuple unpacking above. ``grouper`` is the
+    value of the attribute, and ``list`` is the items with that value.
+
+    .. sourcecode:: html+jinja
+
+        <ul>{% for group in users|groupby("city") %}
+          <li>{{ group.grouper }}: {{ group.list|join(", ") }}
+        {% endfor %}</ul>
+
+    You can specify a ``default`` value to use if an object in the list
+    does not have the given attribute.
+
+    .. sourcecode:: jinja
+
+        <ul>{% for city, items in users|groupby("city", default="NY") %}
+          <li>{{ city }}: {{ items|map(attribute="name")|join(", ") }}</li>
+        {% endfor %}</ul>
+
+    Like the :func:`~jinja-filters.sort` filter, sorting and grouping is
+    case-insensitive by default. The ``key`` for each group will have
+    the case of the first item in that group of values. For example, if
+    a list of users has cities ``["CA", "NY", "ca"]``, the "CA" group
+    will have two values. This can be disabled by passing
+    ``case_sensitive=True``.
+
+    .. versionchanged:: 3.1
+        Added the ``case_sensitive`` parameter. Sorting and grouping is
+        case-insensitive by default, matching other filters that do
+        comparisons.
+
+    .. versionchanged:: 3.0
+        Added the ``default`` parameter.
+
+    .. versionchanged:: 2.6
+        The attribute supports dot notation for nested access.
+    """
+    expr = make_attrgetter(
+        environment,
+        attribute,
+        postprocess=ignore_case if not case_sensitive else None,
+        default=default,
+    )
+    out = [
+        _GroupTuple(key, list(values))
+        for key, values in groupby(sorted(value, key=expr), expr)
+    ]
+
+    if not case_sensitive:
+        # Return the real key from the first value instead of the lowercase key.
+        output_expr = make_attrgetter(environment, attribute, default=default)
+        out = [_GroupTuple(output_expr(values[0]), values) for _, values in out]
+
+    return out
+
+
+@async_variant(sync_do_groupby)  # type: ignore
+async def do_groupby(
+    environment: "Environment",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    attribute: t.Union[str, int],
+    default: t.Optional[t.Any] = None,
+    case_sensitive: bool = False,
+) -> "t.List[_GroupTuple]":
+    expr = make_attrgetter(
+        environment,
+        attribute,
+        postprocess=ignore_case if not case_sensitive else None,
+        default=default,
+    )
+    out = [
+        _GroupTuple(key, await auto_to_list(values))
+        for key, values in groupby(sorted(await auto_to_list(value), key=expr), expr)
+    ]
+
+    if not case_sensitive:
+        # Return the real key from the first value instead of the lowercase key.
+        output_expr = make_attrgetter(environment, attribute, default=default)
+        out = [_GroupTuple(output_expr(values[0]), values) for _, values in out]
+
+    return out
+
+
+@pass_environment
+def sync_do_sum(
+    environment: "Environment",
+    iterable: "t.Iterable[V]",
+    attribute: t.Optional[t.Union[str, int]] = None,
+    start: V = 0,  # type: ignore
+) -> V:
+    """Returns the sum of a sequence of numbers plus the value of parameter
+    'start' (which defaults to 0).  When the sequence is empty it returns
+    start.
+
+    It is also possible to sum up only certain attributes:
+
+    .. sourcecode:: jinja
+
+        Total: {{ items|sum(attribute='price') }}
+
+    .. versionchanged:: 2.6
+       The ``attribute`` parameter was added to allow summing up over
+       attributes.  Also the ``start`` parameter was moved on to the right.
+    """
+    if attribute is not None:
+        iterable = map(make_attrgetter(environment, attribute), iterable)
+
+    return sum(iterable, start)  # type: ignore[no-any-return, call-overload]
+
+
+@async_variant(sync_do_sum)  # type: ignore
+async def do_sum(
+    environment: "Environment",
+    iterable: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    attribute: t.Optional[t.Union[str, int]] = None,
+    start: V = 0,  # type: ignore
+) -> V:
+    rv = start
+
+    if attribute is not None:
+        func = make_attrgetter(environment, attribute)
+    else:
+
+        def func(x: V) -> V:
+            return x
+
+    async for item in auto_aiter(iterable):
+        rv += func(item)
+
+    return rv
+
+
+def sync_do_list(value: "t.Iterable[V]") -> "t.List[V]":
+    """Convert the value into a list.  If it was a string the returned list
+    will be a list of characters.
+    """
+    return list(value)
+
+
+@async_variant(sync_do_list)  # type: ignore
+async def do_list(value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]") -> "t.List[V]":
+    return await auto_to_list(value)
+
+
+def do_mark_safe(value: str) -> Markup:
+    """Mark the value as safe which means that in an environment with automatic
+    escaping enabled this variable will not be escaped.
+    """
+    return Markup(value)
+
+
+def do_mark_unsafe(value: str) -> str:
+    """Mark a value as unsafe.  This is the reverse operation for :func:`safe`."""
+    return str(value)
+
+
+@typing.overload
+def do_reverse(value: str) -> str: ...
+
+
+@typing.overload
+def do_reverse(value: "t.Iterable[V]") -> "t.Iterable[V]": ...
+
+
+def do_reverse(value: t.Union[str, t.Iterable[V]]) -> t.Union[str, t.Iterable[V]]:
+    """Reverse the object or return an iterator that iterates over it the other
+    way round.
+    """
+    if isinstance(value, str):
+        return value[::-1]
+
+    try:
+        return reversed(value)  # type: ignore
+    except TypeError:
+        try:
+            rv = list(value)
+            rv.reverse()
+            return rv
+        except TypeError as e:
+            raise FilterArgumentError("argument must be iterable") from e
+
+
+@pass_environment
+def do_attr(
+    environment: "Environment", obj: t.Any, name: str
+) -> t.Union[Undefined, t.Any]:
+    """Get an attribute of an object.  ``foo|attr("bar")`` works like
+    ``foo.bar`` just that always an attribute is returned and items are not
+    looked up.
+
+    See :ref:`Notes on subscriptions <notes-on-subscriptions>` for more details.
+    """
+    try:
+        name = str(name)
+    except UnicodeError:
+        pass
+    else:
+        try:
+            value = getattr(obj, name)
+        except AttributeError:
+            pass
+        else:
+            if environment.sandboxed:
+                environment = t.cast("SandboxedEnvironment", environment)
+
+                if not environment.is_safe_attribute(obj, name, value):
+                    return environment.unsafe_undefined(obj, name)
+
+            return value
+
+    return environment.undefined(obj=obj, name=name)
+
+
+@typing.overload
+def sync_do_map(
+    context: "Context",
+    value: t.Iterable[t.Any],
+    name: str,
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> t.Iterable[t.Any]: ...
+
+
+@typing.overload
+def sync_do_map(
+    context: "Context",
+    value: t.Iterable[t.Any],
+    *,
+    attribute: str = ...,
+    default: t.Optional[t.Any] = None,
+) -> t.Iterable[t.Any]: ...
+
+
+@pass_context
+def sync_do_map(
+    context: "Context", value: t.Iterable[t.Any], *args: t.Any, **kwargs: t.Any
+) -> t.Iterable[t.Any]:
+    """Applies a filter on a sequence of objects or looks up an attribute.
+    This is useful when dealing with lists of objects but you are really
+    only interested in a certain value of it.
+
+    The basic usage is mapping on an attribute.  Imagine you have a list
+    of users but you are only interested in a list of usernames:
+
+    .. sourcecode:: jinja
+
+        Users on this page: {{ users|map(attribute='username')|join(', ') }}
+
+    You can specify a ``default`` value to use if an object in the list
+    does not have the given attribute.
+
+    .. sourcecode:: jinja
+
+        {{ users|map(attribute="username", default="Anonymous")|join(", ") }}
+
+    Alternatively you can let it invoke a filter by passing the name of the
+    filter and the arguments afterwards.  A good example would be applying a
+    text conversion filter on a sequence:
+
+    .. sourcecode:: jinja
+
+        Users on this page: {{ titles|map('lower')|join(', ') }}
+
+    Similar to a generator comprehension such as:
+
+    .. code-block:: python
+
+        (u.username for u in users)
+        (getattr(u, "username", "Anonymous") for u in users)
+        (do_lower(x) for x in titles)
+
+    .. versionchanged:: 2.11.0
+        Added the ``default`` parameter.
+
+    .. versionadded:: 2.7
+    """
+    if value:
+        func = prepare_map(context, args, kwargs)
+
+        for item in value:
+            yield func(item)
+
+
+@typing.overload
+def do_map(
+    context: "Context",
+    value: t.Union[t.AsyncIterable[t.Any], t.Iterable[t.Any]],
+    name: str,
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> t.Iterable[t.Any]: ...
+
+
+@typing.overload
+def do_map(
+    context: "Context",
+    value: t.Union[t.AsyncIterable[t.Any], t.Iterable[t.Any]],
+    *,
+    attribute: str = ...,
+    default: t.Optional[t.Any] = None,
+) -> t.Iterable[t.Any]: ...
+
+
+@async_variant(sync_do_map)  # type: ignore
+async def do_map(
+    context: "Context",
+    value: t.Union[t.AsyncIterable[t.Any], t.Iterable[t.Any]],
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> t.AsyncIterable[t.Any]:
+    if value:
+        func = prepare_map(context, args, kwargs)
+
+        async for item in auto_aiter(value):
+            yield await auto_await(func(item))
+
+
+@pass_context
+def sync_do_select(
+    context: "Context", value: "t.Iterable[V]", *args: t.Any, **kwargs: t.Any
+) -> "t.Iterator[V]":
+    """Filters a sequence of objects by applying a test to each object,
+    and only selecting the objects with the test succeeding.
+
+    If no test is specified, each object will be evaluated as a boolean.
+
+    Example usage:
+
+    .. sourcecode:: jinja
+
+        {{ numbers|select("odd") }}
+        {{ numbers|select("odd") }}
+        {{ numbers|select("divisibleby", 3) }}
+        {{ numbers|select("lessthan", 42) }}
+        {{ strings|select("equalto", "mystring") }}
+
+    Similar to a generator comprehension such as:
+
+    .. code-block:: python
+
+        (n for n in numbers if test_odd(n))
+        (n for n in numbers if test_divisibleby(n, 3))
+
+    .. versionadded:: 2.7
+    """
+    return select_or_reject(context, value, args, kwargs, lambda x: x, False)
+
+
+@async_variant(sync_do_select)  # type: ignore
+async def do_select(
+    context: "Context",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> "t.AsyncIterator[V]":
+    return async_select_or_reject(context, value, args, kwargs, lambda x: x, False)
+
+
+@pass_context
+def sync_do_reject(
+    context: "Context", value: "t.Iterable[V]", *args: t.Any, **kwargs: t.Any
+) -> "t.Iterator[V]":
+    """Filters a sequence of objects by applying a test to each object,
+    and rejecting the objects with the test succeeding.
+
+    If no test is specified, each object will be evaluated as a boolean.
+
+    Example usage:
+
+    .. sourcecode:: jinja
+
+        {{ numbers|reject("odd") }}
+
+    Similar to a generator comprehension such as:
+
+    .. code-block:: python
+
+        (n for n in numbers if not test_odd(n))
+
+    .. versionadded:: 2.7
+    """
+    return select_or_reject(context, value, args, kwargs, lambda x: not x, False)
+
+
+@async_variant(sync_do_reject)  # type: ignore
+async def do_reject(
+    context: "Context",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> "t.AsyncIterator[V]":
+    return async_select_or_reject(context, value, args, kwargs, lambda x: not x, False)
+
+
+@pass_context
+def sync_do_selectattr(
+    context: "Context", value: "t.Iterable[V]", *args: t.Any, **kwargs: t.Any
+) -> "t.Iterator[V]":
+    """Filters a sequence of objects by applying a test to the specified
+    attribute of each object, and only selecting the objects with the
+    test succeeding.
+
+    If no test is specified, the attribute's value will be evaluated as
+    a boolean.
+
+    Example usage:
+
+    .. sourcecode:: jinja
+
+        {{ users|selectattr("is_active") }}
+        {{ users|selectattr("email", "none") }}
+
+    Similar to a generator comprehension such as:
+
+    .. code-block:: python
+
+        (user for user in users if user.is_active)
+        (user for user in users if test_none(user.email))
+
+    .. versionadded:: 2.7
+    """
+    return select_or_reject(context, value, args, kwargs, lambda x: x, True)
+
+
+@async_variant(sync_do_selectattr)  # type: ignore
+async def do_selectattr(
+    context: "Context",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> "t.AsyncIterator[V]":
+    return async_select_or_reject(context, value, args, kwargs, lambda x: x, True)
+
+
+@pass_context
+def sync_do_rejectattr(
+    context: "Context", value: "t.Iterable[V]", *args: t.Any, **kwargs: t.Any
+) -> "t.Iterator[V]":
+    """Filters a sequence of objects by applying a test to the specified
+    attribute of each object, and rejecting the objects with the test
+    succeeding.
+
+    If no test is specified, the attribute's value will be evaluated as
+    a boolean.
+
+    .. sourcecode:: jinja
+
+        {{ users|rejectattr("is_active") }}
+        {{ users|rejectattr("email", "none") }}
+
+    Similar to a generator comprehension such as:
+
+    .. code-block:: python
+
+        (user for user in users if not user.is_active)
+        (user for user in users if not test_none(user.email))
+
+    .. versionadded:: 2.7
+    """
+    return select_or_reject(context, value, args, kwargs, lambda x: not x, True)
+
+
+@async_variant(sync_do_rejectattr)  # type: ignore
+async def do_rejectattr(
+    context: "Context",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    *args: t.Any,
+    **kwargs: t.Any,
+) -> "t.AsyncIterator[V]":
+    return async_select_or_reject(context, value, args, kwargs, lambda x: not x, True)
+
+
+@pass_eval_context
+def do_tojson(
+    eval_ctx: "EvalContext", value: t.Any, indent: t.Optional[int] = None
+) -> Markup:
+    """Serialize an object to a string of JSON, and mark it safe to
+    render in HTML. This filter is only for use in HTML documents.
+
+    The returned string is safe to render in HTML documents and
+    ``<script>`` tags. The exception is in HTML attributes that are
+    double quoted; either use single quotes or the ``|forceescape``
+    filter.
+
+    :param value: The object to serialize to JSON.
+    :param indent: The ``indent`` parameter passed to ``dumps``, for
+        pretty-printing the value.
+
+    .. versionadded:: 2.9
+    """
+    policies = eval_ctx.environment.policies
+    dumps = policies["json.dumps_function"]
+    kwargs = policies["json.dumps_kwargs"]
+
+    if indent is not None:
+        kwargs = kwargs.copy()
+        kwargs["indent"] = indent
+
+    return htmlsafe_json_dumps(value, dumps=dumps, **kwargs)
+
+
+def prepare_map(
+    context: "Context", args: t.Tuple[t.Any, ...], kwargs: t.Dict[str, t.Any]
+) -> t.Callable[[t.Any], t.Any]:
+    if not args and "attribute" in kwargs:
+        attribute = kwargs.pop("attribute")
+        default = kwargs.pop("default", None)
+
+        if kwargs:
+            raise FilterArgumentError(
+                f"Unexpected keyword argument {next(iter(kwargs))!r}"
+            )
+
+        func = make_attrgetter(context.environment, attribute, default=default)
+    else:
+        try:
+            name = args[0]
+            args = args[1:]
+        except LookupError:
+            raise FilterArgumentError("map requires a filter argument") from None
+
+        def func(item: t.Any) -> t.Any:
+            return context.environment.call_filter(
+                name, item, args, kwargs, context=context
+            )
+
+    return func
+
+
+def prepare_select_or_reject(
+    context: "Context",
+    args: t.Tuple[t.Any, ...],
+    kwargs: t.Dict[str, t.Any],
+    modfunc: t.Callable[[t.Any], t.Any],
+    lookup_attr: bool,
+) -> t.Callable[[t.Any], t.Any]:
+    if lookup_attr:
+        try:
+            attr = args[0]
+        except LookupError:
+            raise FilterArgumentError("Missing parameter for attribute name") from None
+
+        transfunc = make_attrgetter(context.environment, attr)
+        off = 1
+    else:
+        off = 0
+
+        def transfunc(x: V) -> V:
+            return x
+
+    try:
+        name = args[off]
+        args = args[1 + off :]
+
+        def func(item: t.Any) -> t.Any:
+            return context.environment.call_test(name, item, args, kwargs, context)
+
+    except LookupError:
+        func = bool  # type: ignore
+
+    return lambda item: modfunc(func(transfunc(item)))
+
+
+def select_or_reject(
+    context: "Context",
+    value: "t.Iterable[V]",
+    args: t.Tuple[t.Any, ...],
+    kwargs: t.Dict[str, t.Any],
+    modfunc: t.Callable[[t.Any], t.Any],
+    lookup_attr: bool,
+) -> "t.Iterator[V]":
+    if value:
+        func = prepare_select_or_reject(context, args, kwargs, modfunc, lookup_attr)
+
+        for item in value:
+            if func(item):
+                yield item
+
+
+async def async_select_or_reject(
+    context: "Context",
+    value: "t.Union[t.AsyncIterable[V], t.Iterable[V]]",
+    args: t.Tuple[t.Any, ...],
+    kwargs: t.Dict[str, t.Any],
+    modfunc: t.Callable[[t.Any], t.Any],
+    lookup_attr: bool,
+) -> "t.AsyncIterator[V]":
+    if value:
+        func = prepare_select_or_reject(context, args, kwargs, modfunc, lookup_attr)
+
+        async for item in auto_aiter(value):
+            if func(item):
+                yield item
+
+
+FILTERS = {
+    "abs": abs,
+    "attr": do_attr,
+    "batch": do_batch,
+    "capitalize": do_capitalize,
+    "center": do_center,
+    "count": len,
+    "d": do_default,
+    "default": do_default,
+    "dictsort": do_dictsort,
+    "e": escape,
+    "escape": escape,
+    "filesizeformat": do_filesizeformat,
+    "first": do_first,
+    "float": do_float,
+    "forceescape": do_forceescape,
+    "format": do_format,
+    "groupby": do_groupby,
+    "indent": do_indent,
+    "int": do_int,
+    "join": do_join,
+    "last": do_last,
+    "length": len,
+    "list": do_list,
+    "lower": do_lower,
+    "items": do_items,
+    "map": do_map,
+    "min": do_min,
+    "max": do_max,
+    "pprint": do_pprint,
+    "random": do_random,
+    "reject": do_reject,
+    "rejectattr": do_rejectattr,
+    "replace": do_replace,
+    "reverse": do_reverse,
+    "round": do_round,
+    "safe": do_mark_safe,
+    "select": do_select,
+    "selectattr": do_selectattr,
+    "slice": do_slice,
+    "sort": do_sort,
+    "string": soft_str,
+    "striptags": do_striptags,
+    "sum": do_sum,
+    "title": do_title,
+    "trim": do_trim,
+    "truncate": do_truncate,
+    "unique": do_unique,
+    "upper": do_upper,
+    "urlencode": do_urlencode,
+    "urlize": do_urlize,
+    "wordcount": do_wordcount,
+    "wordwrap": do_wordwrap,
+    "xmlattr": do_xmlattr,
+    "tojson": do_tojson,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/idtracking.py b/backend/venv/lib/python3.12/site-packages/jinja2/idtracking.py
new file mode 100644
index 00000000..e6dd8cd1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/idtracking.py
@@ -0,0 +1,318 @@
+import typing as t
+
+from . import nodes
+from .visitor import NodeVisitor
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+VAR_LOAD_PARAMETER = "param"
+VAR_LOAD_RESOLVE = "resolve"
+VAR_LOAD_ALIAS = "alias"
+VAR_LOAD_UNDEFINED = "undefined"
+
+
+def find_symbols(
+    nodes: t.Iterable[nodes.Node], parent_symbols: t.Optional["Symbols"] = None
+) -> "Symbols":
+    sym = Symbols(parent=parent_symbols)
+    visitor = FrameSymbolVisitor(sym)
+    for node in nodes:
+        visitor.visit(node)
+    return sym
+
+
+def symbols_for_node(
+    node: nodes.Node, parent_symbols: t.Optional["Symbols"] = None
+) -> "Symbols":
+    sym = Symbols(parent=parent_symbols)
+    sym.analyze_node(node)
+    return sym
+
+
+class Symbols:
+    def __init__(
+        self, parent: t.Optional["Symbols"] = None, level: t.Optional[int] = None
+    ) -> None:
+        if level is None:
+            if parent is None:
+                level = 0
+            else:
+                level = parent.level + 1
+
+        self.level: int = level
+        self.parent = parent
+        self.refs: t.Dict[str, str] = {}
+        self.loads: t.Dict[str, t.Any] = {}
+        self.stores: t.Set[str] = set()
+
+    def analyze_node(self, node: nodes.Node, **kwargs: t.Any) -> None:
+        visitor = RootVisitor(self)
+        visitor.visit(node, **kwargs)
+
+    def _define_ref(
+        self, name: str, load: t.Optional[t.Tuple[str, t.Optional[str]]] = None
+    ) -> str:
+        ident = f"l_{self.level}_{name}"
+        self.refs[name] = ident
+        if load is not None:
+            self.loads[ident] = load
+        return ident
+
+    def find_load(self, target: str) -> t.Optional[t.Any]:
+        if target in self.loads:
+            return self.loads[target]
+
+        if self.parent is not None:
+            return self.parent.find_load(target)
+
+        return None
+
+    def find_ref(self, name: str) -> t.Optional[str]:
+        if name in self.refs:
+            return self.refs[name]
+
+        if self.parent is not None:
+            return self.parent.find_ref(name)
+
+        return None
+
+    def ref(self, name: str) -> str:
+        rv = self.find_ref(name)
+        if rv is None:
+            raise AssertionError(
+                "Tried to resolve a name to a reference that was"
+                f" unknown to the frame ({name!r})"
+            )
+        return rv
+
+    def copy(self) -> "te.Self":
+        rv = object.__new__(self.__class__)
+        rv.__dict__.update(self.__dict__)
+        rv.refs = self.refs.copy()
+        rv.loads = self.loads.copy()
+        rv.stores = self.stores.copy()
+        return rv
+
+    def store(self, name: str) -> None:
+        self.stores.add(name)
+
+        # If we have not see the name referenced yet, we need to figure
+        # out what to set it to.
+        if name not in self.refs:
+            # If there is a parent scope we check if the name has a
+            # reference there.  If it does it means we might have to alias
+            # to a variable there.
+            if self.parent is not None:
+                outer_ref = self.parent.find_ref(name)
+                if outer_ref is not None:
+                    self._define_ref(name, load=(VAR_LOAD_ALIAS, outer_ref))
+                    return
+
+            # Otherwise we can just set it to undefined.
+            self._define_ref(name, load=(VAR_LOAD_UNDEFINED, None))
+
+    def declare_parameter(self, name: str) -> str:
+        self.stores.add(name)
+        return self._define_ref(name, load=(VAR_LOAD_PARAMETER, None))
+
+    def load(self, name: str) -> None:
+        if self.find_ref(name) is None:
+            self._define_ref(name, load=(VAR_LOAD_RESOLVE, name))
+
+    def branch_update(self, branch_symbols: t.Sequence["Symbols"]) -> None:
+        stores: t.Set[str] = set()
+
+        for branch in branch_symbols:
+            stores.update(branch.stores)
+
+        stores.difference_update(self.stores)
+
+        for sym in branch_symbols:
+            self.refs.update(sym.refs)
+            self.loads.update(sym.loads)
+            self.stores.update(sym.stores)
+
+        for name in stores:
+            target = self.find_ref(name)
+            assert target is not None, "should not happen"
+
+            if self.parent is not None:
+                outer_target = self.parent.find_ref(name)
+                if outer_target is not None:
+                    self.loads[target] = (VAR_LOAD_ALIAS, outer_target)
+                    continue
+            self.loads[target] = (VAR_LOAD_RESOLVE, name)
+
+    def dump_stores(self) -> t.Dict[str, str]:
+        rv: t.Dict[str, str] = {}
+        node: t.Optional[Symbols] = self
+
+        while node is not None:
+            for name in sorted(node.stores):
+                if name not in rv:
+                    rv[name] = self.find_ref(name)  # type: ignore
+
+            node = node.parent
+
+        return rv
+
+    def dump_param_targets(self) -> t.Set[str]:
+        rv = set()
+        node: t.Optional[Symbols] = self
+
+        while node is not None:
+            for target, (instr, _) in self.loads.items():
+                if instr == VAR_LOAD_PARAMETER:
+                    rv.add(target)
+
+            node = node.parent
+
+        return rv
+
+
+class RootVisitor(NodeVisitor):
+    def __init__(self, symbols: "Symbols") -> None:
+        self.sym_visitor = FrameSymbolVisitor(symbols)
+
+    def _simple_visit(self, node: nodes.Node, **kwargs: t.Any) -> None:
+        for child in node.iter_child_nodes():
+            self.sym_visitor.visit(child)
+
+    visit_Template = _simple_visit
+    visit_Block = _simple_visit
+    visit_Macro = _simple_visit
+    visit_FilterBlock = _simple_visit
+    visit_Scope = _simple_visit
+    visit_If = _simple_visit
+    visit_ScopedEvalContextModifier = _simple_visit
+
+    def visit_AssignBlock(self, node: nodes.AssignBlock, **kwargs: t.Any) -> None:
+        for child in node.body:
+            self.sym_visitor.visit(child)
+
+    def visit_CallBlock(self, node: nodes.CallBlock, **kwargs: t.Any) -> None:
+        for child in node.iter_child_nodes(exclude=("call",)):
+            self.sym_visitor.visit(child)
+
+    def visit_OverlayScope(self, node: nodes.OverlayScope, **kwargs: t.Any) -> None:
+        for child in node.body:
+            self.sym_visitor.visit(child)
+
+    def visit_For(
+        self, node: nodes.For, for_branch: str = "body", **kwargs: t.Any
+    ) -> None:
+        if for_branch == "body":
+            self.sym_visitor.visit(node.target, store_as_param=True)
+            branch = node.body
+        elif for_branch == "else":
+            branch = node.else_
+        elif for_branch == "test":
+            self.sym_visitor.visit(node.target, store_as_param=True)
+            if node.test is not None:
+                self.sym_visitor.visit(node.test)
+            return
+        else:
+            raise RuntimeError("Unknown for branch")
+
+        if branch:
+            for item in branch:
+                self.sym_visitor.visit(item)
+
+    def visit_With(self, node: nodes.With, **kwargs: t.Any) -> None:
+        for target in node.targets:
+            self.sym_visitor.visit(target)
+        for child in node.body:
+            self.sym_visitor.visit(child)
+
+    def generic_visit(self, node: nodes.Node, *args: t.Any, **kwargs: t.Any) -> None:
+        raise NotImplementedError(f"Cannot find symbols for {type(node).__name__!r}")
+
+
+class FrameSymbolVisitor(NodeVisitor):
+    """A visitor for `Frame.inspect`."""
+
+    def __init__(self, symbols: "Symbols") -> None:
+        self.symbols = symbols
+
+    def visit_Name(
+        self, node: nodes.Name, store_as_param: bool = False, **kwargs: t.Any
+    ) -> None:
+        """All assignments to names go through this function."""
+        if store_as_param or node.ctx == "param":
+            self.symbols.declare_parameter(node.name)
+        elif node.ctx == "store":
+            self.symbols.store(node.name)
+        elif node.ctx == "load":
+            self.symbols.load(node.name)
+
+    def visit_NSRef(self, node: nodes.NSRef, **kwargs: t.Any) -> None:
+        self.symbols.load(node.name)
+
+    def visit_If(self, node: nodes.If, **kwargs: t.Any) -> None:
+        self.visit(node.test, **kwargs)
+        original_symbols = self.symbols
+
+        def inner_visit(nodes: t.Iterable[nodes.Node]) -> "Symbols":
+            self.symbols = rv = original_symbols.copy()
+
+            for subnode in nodes:
+                self.visit(subnode, **kwargs)
+
+            self.symbols = original_symbols
+            return rv
+
+        body_symbols = inner_visit(node.body)
+        elif_symbols = inner_visit(node.elif_)
+        else_symbols = inner_visit(node.else_ or ())
+        self.symbols.branch_update([body_symbols, elif_symbols, else_symbols])
+
+    def visit_Macro(self, node: nodes.Macro, **kwargs: t.Any) -> None:
+        self.symbols.store(node.name)
+
+    def visit_Import(self, node: nodes.Import, **kwargs: t.Any) -> None:
+        self.generic_visit(node, **kwargs)
+        self.symbols.store(node.target)
+
+    def visit_FromImport(self, node: nodes.FromImport, **kwargs: t.Any) -> None:
+        self.generic_visit(node, **kwargs)
+
+        for name in node.names:
+            if isinstance(name, tuple):
+                self.symbols.store(name[1])
+            else:
+                self.symbols.store(name)
+
+    def visit_Assign(self, node: nodes.Assign, **kwargs: t.Any) -> None:
+        """Visit assignments in the correct order."""
+        self.visit(node.node, **kwargs)
+        self.visit(node.target, **kwargs)
+
+    def visit_For(self, node: nodes.For, **kwargs: t.Any) -> None:
+        """Visiting stops at for blocks.  However the block sequence
+        is visited as part of the outer scope.
+        """
+        self.visit(node.iter, **kwargs)
+
+    def visit_CallBlock(self, node: nodes.CallBlock, **kwargs: t.Any) -> None:
+        self.visit(node.call, **kwargs)
+
+    def visit_FilterBlock(self, node: nodes.FilterBlock, **kwargs: t.Any) -> None:
+        self.visit(node.filter, **kwargs)
+
+    def visit_With(self, node: nodes.With, **kwargs: t.Any) -> None:
+        for target in node.values:
+            self.visit(target)
+
+    def visit_AssignBlock(self, node: nodes.AssignBlock, **kwargs: t.Any) -> None:
+        """Stop visiting at block assigns."""
+        self.visit(node.target, **kwargs)
+
+    def visit_Scope(self, node: nodes.Scope, **kwargs: t.Any) -> None:
+        """Stop visiting at scopes."""
+
+    def visit_Block(self, node: nodes.Block, **kwargs: t.Any) -> None:
+        """Stop visiting at blocks."""
+
+    def visit_OverlayScope(self, node: nodes.OverlayScope, **kwargs: t.Any) -> None:
+        """Do not visit into overlay scopes."""
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/lexer.py b/backend/venv/lib/python3.12/site-packages/jinja2/lexer.py
new file mode 100644
index 00000000..9b1c9697
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/lexer.py
@@ -0,0 +1,868 @@
+"""Implements a Jinja / Python combination lexer. The ``Lexer`` class
+is used to do some preprocessing. It filters out invalid operators like
+the bitshift operators we don't allow in templates. It separates
+template code and python code in expressions.
+"""
+
+import re
+import typing as t
+from ast import literal_eval
+from collections import deque
+from sys import intern
+
+from ._identifier import pattern as name_re
+from .exceptions import TemplateSyntaxError
+from .utils import LRUCache
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+
+# cache for the lexers. Exists in order to be able to have multiple
+# environments with the same lexer
+_lexer_cache: t.MutableMapping[t.Tuple, "Lexer"] = LRUCache(50)  # type: ignore
+
+# static regular expressions
+whitespace_re = re.compile(r"\s+")
+newline_re = re.compile(r"(\r\n|\r|\n)")
+string_re = re.compile(
+    r"('([^'\\]*(?:\\.[^'\\]*)*)'" r'|"([^"\\]*(?:\\.[^"\\]*)*)")', re.S
+)
+integer_re = re.compile(
+    r"""
+    (
+        0b(_?[0-1])+ # binary
+    |
+        0o(_?[0-7])+ # octal
+    |
+        0x(_?[\da-f])+ # hex
+    |
+        [1-9](_?\d)* # decimal
+    |
+        0(_?0)* # decimal zero
+    )
+    """,
+    re.IGNORECASE | re.VERBOSE,
+)
+float_re = re.compile(
+    r"""
+    (?<!\.)  # doesn't start with a .
+    (\d+_)*\d+  # digits, possibly _ separated
+    (
+        (\.(\d+_)*\d+)?  # optional fractional part
+        e[+\-]?(\d+_)*\d+  # exponent part
+    |
+        \.(\d+_)*\d+  # required fractional part
+    )
+    """,
+    re.IGNORECASE | re.VERBOSE,
+)
+
+# internal the tokens and keep references to them
+TOKEN_ADD = intern("add")
+TOKEN_ASSIGN = intern("assign")
+TOKEN_COLON = intern("colon")
+TOKEN_COMMA = intern("comma")
+TOKEN_DIV = intern("div")
+TOKEN_DOT = intern("dot")
+TOKEN_EQ = intern("eq")
+TOKEN_FLOORDIV = intern("floordiv")
+TOKEN_GT = intern("gt")
+TOKEN_GTEQ = intern("gteq")
+TOKEN_LBRACE = intern("lbrace")
+TOKEN_LBRACKET = intern("lbracket")
+TOKEN_LPAREN = intern("lparen")
+TOKEN_LT = intern("lt")
+TOKEN_LTEQ = intern("lteq")
+TOKEN_MOD = intern("mod")
+TOKEN_MUL = intern("mul")
+TOKEN_NE = intern("ne")
+TOKEN_PIPE = intern("pipe")
+TOKEN_POW = intern("pow")
+TOKEN_RBRACE = intern("rbrace")
+TOKEN_RBRACKET = intern("rbracket")
+TOKEN_RPAREN = intern("rparen")
+TOKEN_SEMICOLON = intern("semicolon")
+TOKEN_SUB = intern("sub")
+TOKEN_TILDE = intern("tilde")
+TOKEN_WHITESPACE = intern("whitespace")
+TOKEN_FLOAT = intern("float")
+TOKEN_INTEGER = intern("integer")
+TOKEN_NAME = intern("name")
+TOKEN_STRING = intern("string")
+TOKEN_OPERATOR = intern("operator")
+TOKEN_BLOCK_BEGIN = intern("block_begin")
+TOKEN_BLOCK_END = intern("block_end")
+TOKEN_VARIABLE_BEGIN = intern("variable_begin")
+TOKEN_VARIABLE_END = intern("variable_end")
+TOKEN_RAW_BEGIN = intern("raw_begin")
+TOKEN_RAW_END = intern("raw_end")
+TOKEN_COMMENT_BEGIN = intern("comment_begin")
+TOKEN_COMMENT_END = intern("comment_end")
+TOKEN_COMMENT = intern("comment")
+TOKEN_LINESTATEMENT_BEGIN = intern("linestatement_begin")
+TOKEN_LINESTATEMENT_END = intern("linestatement_end")
+TOKEN_LINECOMMENT_BEGIN = intern("linecomment_begin")
+TOKEN_LINECOMMENT_END = intern("linecomment_end")
+TOKEN_LINECOMMENT = intern("linecomment")
+TOKEN_DATA = intern("data")
+TOKEN_INITIAL = intern("initial")
+TOKEN_EOF = intern("eof")
+
+# bind operators to token types
+operators = {
+    "+": TOKEN_ADD,
+    "-": TOKEN_SUB,
+    "/": TOKEN_DIV,
+    "//": TOKEN_FLOORDIV,
+    "*": TOKEN_MUL,
+    "%": TOKEN_MOD,
+    "**": TOKEN_POW,
+    "~": TOKEN_TILDE,
+    "[": TOKEN_LBRACKET,
+    "]": TOKEN_RBRACKET,
+    "(": TOKEN_LPAREN,
+    ")": TOKEN_RPAREN,
+    "{": TOKEN_LBRACE,
+    "}": TOKEN_RBRACE,
+    "==": TOKEN_EQ,
+    "!=": TOKEN_NE,
+    ">": TOKEN_GT,
+    ">=": TOKEN_GTEQ,
+    "<": TOKEN_LT,
+    "<=": TOKEN_LTEQ,
+    "=": TOKEN_ASSIGN,
+    ".": TOKEN_DOT,
+    ":": TOKEN_COLON,
+    "|": TOKEN_PIPE,
+    ",": TOKEN_COMMA,
+    ";": TOKEN_SEMICOLON,
+}
+
+reverse_operators = {v: k for k, v in operators.items()}
+assert len(operators) == len(reverse_operators), "operators dropped"
+operator_re = re.compile(
+    f"({'|'.join(re.escape(x) for x in sorted(operators, key=lambda x: -len(x)))})"
+)
+
+ignored_tokens = frozenset(
+    [
+        TOKEN_COMMENT_BEGIN,
+        TOKEN_COMMENT,
+        TOKEN_COMMENT_END,
+        TOKEN_WHITESPACE,
+        TOKEN_LINECOMMENT_BEGIN,
+        TOKEN_LINECOMMENT_END,
+        TOKEN_LINECOMMENT,
+    ]
+)
+ignore_if_empty = frozenset(
+    [TOKEN_WHITESPACE, TOKEN_DATA, TOKEN_COMMENT, TOKEN_LINECOMMENT]
+)
+
+
+def _describe_token_type(token_type: str) -> str:
+    if token_type in reverse_operators:
+        return reverse_operators[token_type]
+
+    return {
+        TOKEN_COMMENT_BEGIN: "begin of comment",
+        TOKEN_COMMENT_END: "end of comment",
+        TOKEN_COMMENT: "comment",
+        TOKEN_LINECOMMENT: "comment",
+        TOKEN_BLOCK_BEGIN: "begin of statement block",
+        TOKEN_BLOCK_END: "end of statement block",
+        TOKEN_VARIABLE_BEGIN: "begin of print statement",
+        TOKEN_VARIABLE_END: "end of print statement",
+        TOKEN_LINESTATEMENT_BEGIN: "begin of line statement",
+        TOKEN_LINESTATEMENT_END: "end of line statement",
+        TOKEN_DATA: "template data / text",
+        TOKEN_EOF: "end of template",
+    }.get(token_type, token_type)
+
+
+def describe_token(token: "Token") -> str:
+    """Returns a description of the token."""
+    if token.type == TOKEN_NAME:
+        return token.value
+
+    return _describe_token_type(token.type)
+
+
+def describe_token_expr(expr: str) -> str:
+    """Like `describe_token` but for token expressions."""
+    if ":" in expr:
+        type, value = expr.split(":", 1)
+
+        if type == TOKEN_NAME:
+            return value
+    else:
+        type = expr
+
+    return _describe_token_type(type)
+
+
+def count_newlines(value: str) -> int:
+    """Count the number of newline characters in the string.  This is
+    useful for extensions that filter a stream.
+    """
+    return len(newline_re.findall(value))
+
+
+def compile_rules(environment: "Environment") -> t.List[t.Tuple[str, str]]:
+    """Compiles all the rules from the environment into a list of rules."""
+    e = re.escape
+    rules = [
+        (
+            len(environment.comment_start_string),
+            TOKEN_COMMENT_BEGIN,
+            e(environment.comment_start_string),
+        ),
+        (
+            len(environment.block_start_string),
+            TOKEN_BLOCK_BEGIN,
+            e(environment.block_start_string),
+        ),
+        (
+            len(environment.variable_start_string),
+            TOKEN_VARIABLE_BEGIN,
+            e(environment.variable_start_string),
+        ),
+    ]
+
+    if environment.line_statement_prefix is not None:
+        rules.append(
+            (
+                len(environment.line_statement_prefix),
+                TOKEN_LINESTATEMENT_BEGIN,
+                r"^[ \t\v]*" + e(environment.line_statement_prefix),
+            )
+        )
+    if environment.line_comment_prefix is not None:
+        rules.append(
+            (
+                len(environment.line_comment_prefix),
+                TOKEN_LINECOMMENT_BEGIN,
+                r"(?:^|(?<=\S))[^\S\r\n]*" + e(environment.line_comment_prefix),
+            )
+        )
+
+    return [x[1:] for x in sorted(rules, reverse=True)]
+
+
+class Failure:
+    """Class that raises a `TemplateSyntaxError` if called.
+    Used by the `Lexer` to specify known errors.
+    """
+
+    def __init__(
+        self, message: str, cls: t.Type[TemplateSyntaxError] = TemplateSyntaxError
+    ) -> None:
+        self.message = message
+        self.error_class = cls
+
+    def __call__(self, lineno: int, filename: t.Optional[str]) -> "te.NoReturn":
+        raise self.error_class(self.message, lineno, filename)
+
+
+class Token(t.NamedTuple):
+    lineno: int
+    type: str
+    value: str
+
+    def __str__(self) -> str:
+        return describe_token(self)
+
+    def test(self, expr: str) -> bool:
+        """Test a token against a token expression.  This can either be a
+        token type or ``'token_type:token_value'``.  This can only test
+        against string values and types.
+        """
+        # here we do a regular string equality check as test_any is usually
+        # passed an iterable of not interned strings.
+        if self.type == expr:
+            return True
+
+        if ":" in expr:
+            return expr.split(":", 1) == [self.type, self.value]
+
+        return False
+
+    def test_any(self, *iterable: str) -> bool:
+        """Test against multiple token expressions."""
+        return any(self.test(expr) for expr in iterable)
+
+
+class TokenStreamIterator:
+    """The iterator for tokenstreams.  Iterate over the stream
+    until the eof token is reached.
+    """
+
+    def __init__(self, stream: "TokenStream") -> None:
+        self.stream = stream
+
+    def __iter__(self) -> "TokenStreamIterator":
+        return self
+
+    def __next__(self) -> Token:
+        token = self.stream.current
+
+        if token.type is TOKEN_EOF:
+            self.stream.close()
+            raise StopIteration
+
+        next(self.stream)
+        return token
+
+
+class TokenStream:
+    """A token stream is an iterable that yields :class:`Token`\\s.  The
+    parser however does not iterate over it but calls :meth:`next` to go
+    one token ahead.  The current active token is stored as :attr:`current`.
+    """
+
+    def __init__(
+        self,
+        generator: t.Iterable[Token],
+        name: t.Optional[str],
+        filename: t.Optional[str],
+    ):
+        self._iter = iter(generator)
+        self._pushed: te.Deque[Token] = deque()
+        self.name = name
+        self.filename = filename
+        self.closed = False
+        self.current = Token(1, TOKEN_INITIAL, "")
+        next(self)
+
+    def __iter__(self) -> TokenStreamIterator:
+        return TokenStreamIterator(self)
+
+    def __bool__(self) -> bool:
+        return bool(self._pushed) or self.current.type is not TOKEN_EOF
+
+    @property
+    def eos(self) -> bool:
+        """Are we at the end of the stream?"""
+        return not self
+
+    def push(self, token: Token) -> None:
+        """Push a token back to the stream."""
+        self._pushed.append(token)
+
+    def look(self) -> Token:
+        """Look at the next token."""
+        old_token = next(self)
+        result = self.current
+        self.push(result)
+        self.current = old_token
+        return result
+
+    def skip(self, n: int = 1) -> None:
+        """Got n tokens ahead."""
+        for _ in range(n):
+            next(self)
+
+    def next_if(self, expr: str) -> t.Optional[Token]:
+        """Perform the token test and return the token if it matched.
+        Otherwise the return value is `None`.
+        """
+        if self.current.test(expr):
+            return next(self)
+
+        return None
+
+    def skip_if(self, expr: str) -> bool:
+        """Like :meth:`next_if` but only returns `True` or `False`."""
+        return self.next_if(expr) is not None
+
+    def __next__(self) -> Token:
+        """Go one token ahead and return the old one.
+
+        Use the built-in :func:`next` instead of calling this directly.
+        """
+        rv = self.current
+
+        if self._pushed:
+            self.current = self._pushed.popleft()
+        elif self.current.type is not TOKEN_EOF:
+            try:
+                self.current = next(self._iter)
+            except StopIteration:
+                self.close()
+
+        return rv
+
+    def close(self) -> None:
+        """Close the stream."""
+        self.current = Token(self.current.lineno, TOKEN_EOF, "")
+        self._iter = iter(())
+        self.closed = True
+
+    def expect(self, expr: str) -> Token:
+        """Expect a given token type and return it.  This accepts the same
+        argument as :meth:`jinja2.lexer.Token.test`.
+        """
+        if not self.current.test(expr):
+            expr = describe_token_expr(expr)
+
+            if self.current.type is TOKEN_EOF:
+                raise TemplateSyntaxError(
+                    f"unexpected end of template, expected {expr!r}.",
+                    self.current.lineno,
+                    self.name,
+                    self.filename,
+                )
+
+            raise TemplateSyntaxError(
+                f"expected token {expr!r}, got {describe_token(self.current)!r}",
+                self.current.lineno,
+                self.name,
+                self.filename,
+            )
+
+        return next(self)
+
+
+def get_lexer(environment: "Environment") -> "Lexer":
+    """Return a lexer which is probably cached."""
+    key = (
+        environment.block_start_string,
+        environment.block_end_string,
+        environment.variable_start_string,
+        environment.variable_end_string,
+        environment.comment_start_string,
+        environment.comment_end_string,
+        environment.line_statement_prefix,
+        environment.line_comment_prefix,
+        environment.trim_blocks,
+        environment.lstrip_blocks,
+        environment.newline_sequence,
+        environment.keep_trailing_newline,
+    )
+    lexer = _lexer_cache.get(key)
+
+    if lexer is None:
+        _lexer_cache[key] = lexer = Lexer(environment)
+
+    return lexer
+
+
+class OptionalLStrip(tuple):  # type: ignore[type-arg]
+    """A special tuple for marking a point in the state that can have
+    lstrip applied.
+    """
+
+    __slots__ = ()
+
+    # Even though it looks like a no-op, creating instances fails
+    # without this.
+    def __new__(cls, *members, **kwargs):  # type: ignore
+        return super().__new__(cls, members)
+
+
+class _Rule(t.NamedTuple):
+    pattern: t.Pattern[str]
+    tokens: t.Union[str, t.Tuple[str, ...], t.Tuple[Failure]]
+    command: t.Optional[str]
+
+
+class Lexer:
+    """Class that implements a lexer for a given environment. Automatically
+    created by the environment class, usually you don't have to do that.
+
+    Note that the lexer is not automatically bound to an environment.
+    Multiple environments can share the same lexer.
+    """
+
+    def __init__(self, environment: "Environment") -> None:
+        # shortcuts
+        e = re.escape
+
+        def c(x: str) -> t.Pattern[str]:
+            return re.compile(x, re.M | re.S)
+
+        # lexing rules for tags
+        tag_rules: t.List[_Rule] = [
+            _Rule(whitespace_re, TOKEN_WHITESPACE, None),
+            _Rule(float_re, TOKEN_FLOAT, None),
+            _Rule(integer_re, TOKEN_INTEGER, None),
+            _Rule(name_re, TOKEN_NAME, None),
+            _Rule(string_re, TOKEN_STRING, None),
+            _Rule(operator_re, TOKEN_OPERATOR, None),
+        ]
+
+        # assemble the root lexing rule. because "|" is ungreedy
+        # we have to sort by length so that the lexer continues working
+        # as expected when we have parsing rules like <% for block and
+        # <%= for variables. (if someone wants asp like syntax)
+        # variables are just part of the rules if variable processing
+        # is required.
+        root_tag_rules = compile_rules(environment)
+
+        block_start_re = e(environment.block_start_string)
+        block_end_re = e(environment.block_end_string)
+        comment_end_re = e(environment.comment_end_string)
+        variable_end_re = e(environment.variable_end_string)
+
+        # block suffix if trimming is enabled
+        block_suffix_re = "\\n?" if environment.trim_blocks else ""
+
+        self.lstrip_blocks = environment.lstrip_blocks
+
+        self.newline_sequence = environment.newline_sequence
+        self.keep_trailing_newline = environment.keep_trailing_newline
+
+        root_raw_re = (
+            rf"(?P<raw_begin>{block_start_re}(\-|\+|)\s*raw\s*"
+            rf"(?:\-{block_end_re}\s*|{block_end_re}))"
+        )
+        root_parts_re = "|".join(
+            [root_raw_re] + [rf"(?P<{n}>{r}(\-|\+|))" for n, r in root_tag_rules]
+        )
+
+        # global lexing rules
+        self.rules: t.Dict[str, t.List[_Rule]] = {
+            "root": [
+                # directives
+                _Rule(
+                    c(rf"(.*?)(?:{root_parts_re})"),
+                    OptionalLStrip(TOKEN_DATA, "#bygroup"),  # type: ignore
+                    "#bygroup",
+                ),
+                # data
+                _Rule(c(".+"), TOKEN_DATA, None),
+            ],
+            # comments
+            TOKEN_COMMENT_BEGIN: [
+                _Rule(
+                    c(
+                        rf"(.*?)((?:\+{comment_end_re}|\-{comment_end_re}\s*"
+                        rf"|{comment_end_re}{block_suffix_re}))"
+                    ),
+                    (TOKEN_COMMENT, TOKEN_COMMENT_END),
+                    "#pop",
+                ),
+                _Rule(c(r"(.)"), (Failure("Missing end of comment tag"),), None),
+            ],
+            # blocks
+            TOKEN_BLOCK_BEGIN: [
+                _Rule(
+                    c(
+                        rf"(?:\+{block_end_re}|\-{block_end_re}\s*"
+                        rf"|{block_end_re}{block_suffix_re})"
+                    ),
+                    TOKEN_BLOCK_END,
+                    "#pop",
+                ),
+            ]
+            + tag_rules,
+            # variables
+            TOKEN_VARIABLE_BEGIN: [
+                _Rule(
+                    c(rf"\-{variable_end_re}\s*|{variable_end_re}"),
+                    TOKEN_VARIABLE_END,
+                    "#pop",
+                )
+            ]
+            + tag_rules,
+            # raw block
+            TOKEN_RAW_BEGIN: [
+                _Rule(
+                    c(
+                        rf"(.*?)((?:{block_start_re}(\-|\+|))\s*endraw\s*"
+                        rf"(?:\+{block_end_re}|\-{block_end_re}\s*"
+                        rf"|{block_end_re}{block_suffix_re}))"
+                    ),
+                    OptionalLStrip(TOKEN_DATA, TOKEN_RAW_END),  # type: ignore
+                    "#pop",
+                ),
+                _Rule(c(r"(.)"), (Failure("Missing end of raw directive"),), None),
+            ],
+            # line statements
+            TOKEN_LINESTATEMENT_BEGIN: [
+                _Rule(c(r"\s*(\n|$)"), TOKEN_LINESTATEMENT_END, "#pop")
+            ]
+            + tag_rules,
+            # line comments
+            TOKEN_LINECOMMENT_BEGIN: [
+                _Rule(
+                    c(r"(.*?)()(?=\n|$)"),
+                    (TOKEN_LINECOMMENT, TOKEN_LINECOMMENT_END),
+                    "#pop",
+                )
+            ],
+        }
+
+    def _normalize_newlines(self, value: str) -> str:
+        """Replace all newlines with the configured sequence in strings
+        and template data.
+        """
+        return newline_re.sub(self.newline_sequence, value)
+
+    def tokenize(
+        self,
+        source: str,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+        state: t.Optional[str] = None,
+    ) -> TokenStream:
+        """Calls tokeniter + tokenize and wraps it in a token stream."""
+        stream = self.tokeniter(source, name, filename, state)
+        return TokenStream(self.wrap(stream, name, filename), name, filename)
+
+    def wrap(
+        self,
+        stream: t.Iterable[t.Tuple[int, str, str]],
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+    ) -> t.Iterator[Token]:
+        """This is called with the stream as returned by `tokenize` and wraps
+        every token in a :class:`Token` and converts the value.
+        """
+        for lineno, token, value_str in stream:
+            if token in ignored_tokens:
+                continue
+
+            value: t.Any = value_str
+
+            if token == TOKEN_LINESTATEMENT_BEGIN:
+                token = TOKEN_BLOCK_BEGIN
+            elif token == TOKEN_LINESTATEMENT_END:
+                token = TOKEN_BLOCK_END
+            # we are not interested in those tokens in the parser
+            elif token in (TOKEN_RAW_BEGIN, TOKEN_RAW_END):
+                continue
+            elif token == TOKEN_DATA:
+                value = self._normalize_newlines(value_str)
+            elif token == "keyword":
+                token = value_str
+            elif token == TOKEN_NAME:
+                value = value_str
+
+                if not value.isidentifier():
+                    raise TemplateSyntaxError(
+                        "Invalid character in identifier", lineno, name, filename
+                    )
+            elif token == TOKEN_STRING:
+                # try to unescape string
+                try:
+                    value = (
+                        self._normalize_newlines(value_str[1:-1])
+                        .encode("ascii", "backslashreplace")
+                        .decode("unicode-escape")
+                    )
+                except Exception as e:
+                    msg = str(e).split(":")[-1].strip()
+                    raise TemplateSyntaxError(msg, lineno, name, filename) from e
+            elif token == TOKEN_INTEGER:
+                value = int(value_str.replace("_", ""), 0)
+            elif token == TOKEN_FLOAT:
+                # remove all "_" first to support more Python versions
+                value = literal_eval(value_str.replace("_", ""))
+            elif token == TOKEN_OPERATOR:
+                token = operators[value_str]
+
+            yield Token(lineno, token, value)
+
+    def tokeniter(
+        self,
+        source: str,
+        name: t.Optional[str],
+        filename: t.Optional[str] = None,
+        state: t.Optional[str] = None,
+    ) -> t.Iterator[t.Tuple[int, str, str]]:
+        """This method tokenizes the text and returns the tokens in a
+        generator. Use this method if you just want to tokenize a template.
+
+        .. versionchanged:: 3.0
+            Only ``\\n``, ``\\r\\n`` and ``\\r`` are treated as line
+            breaks.
+        """
+        lines = newline_re.split(source)[::2]
+
+        if not self.keep_trailing_newline and lines[-1] == "":
+            del lines[-1]
+
+        source = "\n".join(lines)
+        pos = 0
+        lineno = 1
+        stack = ["root"]
+
+        if state is not None and state != "root":
+            assert state in ("variable", "block"), "invalid state"
+            stack.append(state + "_begin")
+
+        statetokens = self.rules[stack[-1]]
+        source_length = len(source)
+        balancing_stack: t.List[str] = []
+        newlines_stripped = 0
+        line_starting = True
+
+        while True:
+            # tokenizer loop
+            for regex, tokens, new_state in statetokens:
+                m = regex.match(source, pos)
+
+                # if no match we try again with the next rule
+                if m is None:
+                    continue
+
+                # we only match blocks and variables if braces / parentheses
+                # are balanced. continue parsing with the lower rule which
+                # is the operator rule. do this only if the end tags look
+                # like operators
+                if balancing_stack and tokens in (
+                    TOKEN_VARIABLE_END,
+                    TOKEN_BLOCK_END,
+                    TOKEN_LINESTATEMENT_END,
+                ):
+                    continue
+
+                # tuples support more options
+                if isinstance(tokens, tuple):
+                    groups: t.Sequence[str] = m.groups()
+
+                    if isinstance(tokens, OptionalLStrip):
+                        # Rule supports lstrip. Match will look like
+                        # text, block type, whitespace control, type, control, ...
+                        text = groups[0]
+                        # Skipping the text and first type, every other group is the
+                        # whitespace control for each type. One of the groups will be
+                        # -, +, or empty string instead of None.
+                        strip_sign = next(g for g in groups[2::2] if g is not None)
+
+                        if strip_sign == "-":
+                            # Strip all whitespace between the text and the tag.
+                            stripped = text.rstrip()
+                            newlines_stripped = text[len(stripped) :].count("\n")
+                            groups = [stripped, *groups[1:]]
+                        elif (
+                            # Not marked for preserving whitespace.
+                            strip_sign != "+"
+                            # lstrip is enabled.
+                            and self.lstrip_blocks
+                            # Not a variable expression.
+                            and not m.groupdict().get(TOKEN_VARIABLE_BEGIN)
+                        ):
+                            # The start of text between the last newline and the tag.
+                            l_pos = text.rfind("\n") + 1
+
+                            if l_pos > 0 or line_starting:
+                                # If there's only whitespace between the newline and the
+                                # tag, strip it.
+                                if whitespace_re.fullmatch(text, l_pos):
+                                    groups = [text[:l_pos], *groups[1:]]
+
+                    for idx, token in enumerate(tokens):
+                        # failure group
+                        if isinstance(token, Failure):
+                            raise token(lineno, filename)
+                        # bygroup is a bit more complex, in that case we
+                        # yield for the current token the first named
+                        # group that matched
+                        elif token == "#bygroup":
+                            for key, value in m.groupdict().items():
+                                if value is not None:
+                                    yield lineno, key, value
+                                    lineno += value.count("\n")
+                                    break
+                            else:
+                                raise RuntimeError(
+                                    f"{regex!r} wanted to resolve the token dynamically"
+                                    " but no group matched"
+                                )
+                        # normal group
+                        else:
+                            data = groups[idx]
+
+                            if data or token not in ignore_if_empty:
+                                yield lineno, token, data  # type: ignore[misc]
+
+                            lineno += data.count("\n") + newlines_stripped
+                            newlines_stripped = 0
+
+                # strings as token just are yielded as it.
+                else:
+                    data = m.group()
+
+                    # update brace/parentheses balance
+                    if tokens == TOKEN_OPERATOR:
+                        if data == "{":
+                            balancing_stack.append("}")
+                        elif data == "(":
+                            balancing_stack.append(")")
+                        elif data == "[":
+                            balancing_stack.append("]")
+                        elif data in ("}", ")", "]"):
+                            if not balancing_stack:
+                                raise TemplateSyntaxError(
+                                    f"unexpected '{data}'", lineno, name, filename
+                                )
+
+                            expected_op = balancing_stack.pop()
+
+                            if expected_op != data:
+                                raise TemplateSyntaxError(
+                                    f"unexpected '{data}', expected '{expected_op}'",
+                                    lineno,
+                                    name,
+                                    filename,
+                                )
+
+                    # yield items
+                    if data or tokens not in ignore_if_empty:
+                        yield lineno, tokens, data
+
+                    lineno += data.count("\n")
+
+                line_starting = m.group()[-1:] == "\n"
+                # fetch new position into new variable so that we can check
+                # if there is a internal parsing error which would result
+                # in an infinite loop
+                pos2 = m.end()
+
+                # handle state changes
+                if new_state is not None:
+                    # remove the uppermost state
+                    if new_state == "#pop":
+                        stack.pop()
+                    # resolve the new state by group checking
+                    elif new_state == "#bygroup":
+                        for key, value in m.groupdict().items():
+                            if value is not None:
+                                stack.append(key)
+                                break
+                        else:
+                            raise RuntimeError(
+                                f"{regex!r} wanted to resolve the new state dynamically"
+                                f" but no group matched"
+                            )
+                    # direct state name given
+                    else:
+                        stack.append(new_state)
+
+                    statetokens = self.rules[stack[-1]]
+                # we are still at the same position and no stack change.
+                # this means a loop without break condition, avoid that and
+                # raise error
+                elif pos2 == pos:
+                    raise RuntimeError(
+                        f"{regex!r} yielded empty string without stack change"
+                    )
+
+                # publish new function and start again
+                pos = pos2
+                break
+            # if loop terminated without break we haven't found a single match
+            # either we are at the end of the file or we have a problem
+            else:
+                # end of text
+                if pos >= source_length:
+                    return
+
+                # something went wrong
+                raise TemplateSyntaxError(
+                    f"unexpected char {source[pos]!r} at {pos}", lineno, name, filename
+                )
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/loaders.py b/backend/venv/lib/python3.12/site-packages/jinja2/loaders.py
new file mode 100644
index 00000000..3913ee51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/loaders.py
@@ -0,0 +1,693 @@
+"""API and implementations for loading templates from different data
+sources.
+"""
+
+import importlib.util
+import os
+import posixpath
+import sys
+import typing as t
+import weakref
+import zipimport
+from collections import abc
+from hashlib import sha1
+from importlib import import_module
+from types import ModuleType
+
+from .exceptions import TemplateNotFound
+from .utils import internalcode
+
+if t.TYPE_CHECKING:
+    from .environment import Environment
+    from .environment import Template
+
+
+def split_template_path(template: str) -> t.List[str]:
+    """Split a path into segments and perform a sanity check.  If it detects
+    '..' in the path it will raise a `TemplateNotFound` error.
+    """
+    pieces = []
+    for piece in template.split("/"):
+        if (
+            os.path.sep in piece
+            or (os.path.altsep and os.path.altsep in piece)
+            or piece == os.path.pardir
+        ):
+            raise TemplateNotFound(template)
+        elif piece and piece != ".":
+            pieces.append(piece)
+    return pieces
+
+
+class BaseLoader:
+    """Baseclass for all loaders.  Subclass this and override `get_source` to
+    implement a custom loading mechanism.  The environment provides a
+    `get_template` method that calls the loader's `load` method to get the
+    :class:`Template` object.
+
+    A very basic example for a loader that looks up templates on the file
+    system could look like this::
+
+        from jinja2 import BaseLoader, TemplateNotFound
+        from os.path import join, exists, getmtime
+
+        class MyLoader(BaseLoader):
+
+            def __init__(self, path):
+                self.path = path
+
+            def get_source(self, environment, template):
+                path = join(self.path, template)
+                if not exists(path):
+                    raise TemplateNotFound(template)
+                mtime = getmtime(path)
+                with open(path) as f:
+                    source = f.read()
+                return source, path, lambda: mtime == getmtime(path)
+    """
+
+    #: if set to `False` it indicates that the loader cannot provide access
+    #: to the source of templates.
+    #:
+    #: .. versionadded:: 2.4
+    has_source_access = True
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
+        """Get the template source, filename and reload helper for a template.
+        It's passed the environment and template name and has to return a
+        tuple in the form ``(source, filename, uptodate)`` or raise a
+        `TemplateNotFound` error if it can't locate the template.
+
+        The source part of the returned tuple must be the source of the
+        template as a string. The filename should be the name of the
+        file on the filesystem if it was loaded from there, otherwise
+        ``None``. The filename is used by Python for the tracebacks
+        if no loader extension is used.
+
+        The last item in the tuple is the `uptodate` function.  If auto
+        reloading is enabled it's always called to check if the template
+        changed.  No arguments are passed so the function must store the
+        old state somewhere (for example in a closure).  If it returns `False`
+        the template will be reloaded.
+        """
+        if not self.has_source_access:
+            raise RuntimeError(
+                f"{type(self).__name__} cannot provide access to the source"
+            )
+        raise TemplateNotFound(template)
+
+    def list_templates(self) -> t.List[str]:
+        """Iterates over all templates.  If the loader does not support that
+        it should raise a :exc:`TypeError` which is the default behavior.
+        """
+        raise TypeError("this loader cannot iterate over all templates")
+
+    @internalcode
+    def load(
+        self,
+        environment: "Environment",
+        name: str,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        """Loads a template.  This method looks up the template in the cache
+        or loads one by calling :meth:`get_source`.  Subclasses should not
+        override this method as loaders working on collections of other
+        loaders (such as :class:`PrefixLoader` or :class:`ChoiceLoader`)
+        will not call this method but `get_source` directly.
+        """
+        code = None
+        if globals is None:
+            globals = {}
+
+        # first we try to get the source for this template together
+        # with the filename and the uptodate function.
+        source, filename, uptodate = self.get_source(environment, name)
+
+        # try to load the code from the bytecode cache if there is a
+        # bytecode cache configured.
+        bcc = environment.bytecode_cache
+        if bcc is not None:
+            bucket = bcc.get_bucket(environment, name, filename, source)
+            code = bucket.code
+
+        # if we don't have code so far (not cached, no longer up to
+        # date) etc. we compile the template
+        if code is None:
+            code = environment.compile(source, name, filename)
+
+        # if the bytecode cache is available and the bucket doesn't
+        # have a code so far, we give the bucket the new code and put
+        # it back to the bytecode cache.
+        if bcc is not None and bucket.code is None:
+            bucket.code = code
+            bcc.set_bucket(bucket)
+
+        return environment.template_class.from_code(
+            environment, code, globals, uptodate
+        )
+
+
+class FileSystemLoader(BaseLoader):
+    """Load templates from a directory in the file system.
+
+    The path can be relative or absolute. Relative paths are relative to
+    the current working directory.
+
+    .. code-block:: python
+
+        loader = FileSystemLoader("templates")
+
+    A list of paths can be given. The directories will be searched in
+    order, stopping at the first matching template.
+
+    .. code-block:: python
+
+        loader = FileSystemLoader(["/override/templates", "/default/templates"])
+
+    :param searchpath: A path, or list of paths, to the directory that
+        contains the templates.
+    :param encoding: Use this encoding to read the text from template
+        files.
+    :param followlinks: Follow symbolic links in the path.
+
+    .. versionchanged:: 2.8
+        Added the ``followlinks`` parameter.
+    """
+
+    def __init__(
+        self,
+        searchpath: t.Union[
+            str, "os.PathLike[str]", t.Sequence[t.Union[str, "os.PathLike[str]"]]
+        ],
+        encoding: str = "utf-8",
+        followlinks: bool = False,
+    ) -> None:
+        if not isinstance(searchpath, abc.Iterable) or isinstance(searchpath, str):
+            searchpath = [searchpath]
+
+        self.searchpath = [os.fspath(p) for p in searchpath]
+        self.encoding = encoding
+        self.followlinks = followlinks
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, str, t.Callable[[], bool]]:
+        pieces = split_template_path(template)
+
+        for searchpath in self.searchpath:
+            # Use posixpath even on Windows to avoid "drive:" or UNC
+            # segments breaking out of the search directory.
+            filename = posixpath.join(searchpath, *pieces)
+
+            if os.path.isfile(filename):
+                break
+        else:
+            plural = "path" if len(self.searchpath) == 1 else "paths"
+            paths_str = ", ".join(repr(p) for p in self.searchpath)
+            raise TemplateNotFound(
+                template,
+                f"{template!r} not found in search {plural}: {paths_str}",
+            )
+
+        with open(filename, encoding=self.encoding) as f:
+            contents = f.read()
+
+        mtime = os.path.getmtime(filename)
+
+        def uptodate() -> bool:
+            try:
+                return os.path.getmtime(filename) == mtime
+            except OSError:
+                return False
+
+        # Use normpath to convert Windows altsep to sep.
+        return contents, os.path.normpath(filename), uptodate
+
+    def list_templates(self) -> t.List[str]:
+        found = set()
+        for searchpath in self.searchpath:
+            walk_dir = os.walk(searchpath, followlinks=self.followlinks)
+            for dirpath, _, filenames in walk_dir:
+                for filename in filenames:
+                    template = (
+                        os.path.join(dirpath, filename)[len(searchpath) :]
+                        .strip(os.path.sep)
+                        .replace(os.path.sep, "/")
+                    )
+                    if template[:2] == "./":
+                        template = template[2:]
+                    if template not in found:
+                        found.add(template)
+        return sorted(found)
+
+
+if sys.version_info >= (3, 13):
+
+    def _get_zipimporter_files(z: t.Any) -> t.Dict[str, object]:
+        try:
+            get_files = z._get_files
+        except AttributeError as e:
+            raise TypeError(
+                "This zip import does not have the required"
+                " metadata to list templates."
+            ) from e
+        return get_files()
+else:
+
+    def _get_zipimporter_files(z: t.Any) -> t.Dict[str, object]:
+        try:
+            files = z._files
+        except AttributeError as e:
+            raise TypeError(
+                "This zip import does not have the required"
+                " metadata to list templates."
+            ) from e
+        return files  # type: ignore[no-any-return]
+
+
+class PackageLoader(BaseLoader):
+    """Load templates from a directory in a Python package.
+
+    :param package_name: Import name of the package that contains the
+        template directory.
+    :param package_path: Directory within the imported package that
+        contains the templates.
+    :param encoding: Encoding of template files.
+
+    The following example looks up templates in the ``pages`` directory
+    within the ``project.ui`` package.
+
+    .. code-block:: python
+
+        loader = PackageLoader("project.ui", "pages")
+
+    Only packages installed as directories (standard pip behavior) or
+    zip/egg files (less common) are supported. The Python API for
+    introspecting data in packages is too limited to support other
+    installation methods the way this loader requires.
+
+    There is limited support for :pep:`420` namespace packages. The
+    template directory is assumed to only be in one namespace
+    contributor. Zip files contributing to a namespace are not
+    supported.
+
+    .. versionchanged:: 3.0
+        No longer uses ``setuptools`` as a dependency.
+
+    .. versionchanged:: 3.0
+        Limited PEP 420 namespace package support.
+    """
+
+    def __init__(
+        self,
+        package_name: str,
+        package_path: "str" = "templates",
+        encoding: str = "utf-8",
+    ) -> None:
+        package_path = os.path.normpath(package_path).rstrip(os.path.sep)
+
+        # normpath preserves ".", which isn't valid in zip paths.
+        if package_path == os.path.curdir:
+            package_path = ""
+        elif package_path[:2] == os.path.curdir + os.path.sep:
+            package_path = package_path[2:]
+
+        self.package_path = package_path
+        self.package_name = package_name
+        self.encoding = encoding
+
+        # Make sure the package exists. This also makes namespace
+        # packages work, otherwise get_loader returns None.
+        import_module(package_name)
+        spec = importlib.util.find_spec(package_name)
+        assert spec is not None, "An import spec was not found for the package."
+        loader = spec.loader
+        assert loader is not None, "A loader was not found for the package."
+        self._loader = loader
+        self._archive = None
+
+        if isinstance(loader, zipimport.zipimporter):
+            self._archive = loader.archive
+            pkgdir = next(iter(spec.submodule_search_locations))  # type: ignore
+            template_root = os.path.join(pkgdir, package_path).rstrip(os.path.sep)
+        else:
+            roots: t.List[str] = []
+
+            # One element for regular packages, multiple for namespace
+            # packages, or None for single module file.
+            if spec.submodule_search_locations:
+                roots.extend(spec.submodule_search_locations)
+            # A single module file, use the parent directory instead.
+            elif spec.origin is not None:
+                roots.append(os.path.dirname(spec.origin))
+
+            if not roots:
+                raise ValueError(
+                    f"The {package_name!r} package was not installed in a"
+                    " way that PackageLoader understands."
+                )
+
+            for root in roots:
+                root = os.path.join(root, package_path)
+
+                if os.path.isdir(root):
+                    template_root = root
+                    break
+            else:
+                raise ValueError(
+                    f"PackageLoader could not find a {package_path!r} directory"
+                    f" in the {package_name!r} package."
+                )
+
+        self._template_root = template_root
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, str, t.Optional[t.Callable[[], bool]]]:
+        # Use posixpath even on Windows to avoid "drive:" or UNC
+        # segments breaking out of the search directory. Use normpath to
+        # convert Windows altsep to sep.
+        p = os.path.normpath(
+            posixpath.join(self._template_root, *split_template_path(template))
+        )
+        up_to_date: t.Optional[t.Callable[[], bool]]
+
+        if self._archive is None:
+            # Package is a directory.
+            if not os.path.isfile(p):
+                raise TemplateNotFound(template)
+
+            with open(p, "rb") as f:
+                source = f.read()
+
+            mtime = os.path.getmtime(p)
+
+            def up_to_date() -> bool:
+                return os.path.isfile(p) and os.path.getmtime(p) == mtime
+
+        else:
+            # Package is a zip file.
+            try:
+                source = self._loader.get_data(p)  # type: ignore
+            except OSError as e:
+                raise TemplateNotFound(template) from e
+
+            # Could use the zip's mtime for all template mtimes, but
+            # would need to safely reload the module if it's out of
+            # date, so just report it as always current.
+            up_to_date = None
+
+        return source.decode(self.encoding), p, up_to_date
+
+    def list_templates(self) -> t.List[str]:
+        results: t.List[str] = []
+
+        if self._archive is None:
+            # Package is a directory.
+            offset = len(self._template_root)
+
+            for dirpath, _, filenames in os.walk(self._template_root):
+                dirpath = dirpath[offset:].lstrip(os.path.sep)
+                results.extend(
+                    os.path.join(dirpath, name).replace(os.path.sep, "/")
+                    for name in filenames
+                )
+        else:
+            files = _get_zipimporter_files(self._loader)
+
+            # Package is a zip file.
+            prefix = (
+                self._template_root[len(self._archive) :].lstrip(os.path.sep)
+                + os.path.sep
+            )
+            offset = len(prefix)
+
+            for name in files:
+                # Find names under the templates directory that aren't directories.
+                if name.startswith(prefix) and name[-1] != os.path.sep:
+                    results.append(name[offset:].replace(os.path.sep, "/"))
+
+        results.sort()
+        return results
+
+
+class DictLoader(BaseLoader):
+    """Loads a template from a Python dict mapping template names to
+    template source.  This loader is useful for unittesting:
+
+    >>> loader = DictLoader({'index.html': 'source here'})
+
+    Because auto reloading is rarely useful this is disabled by default.
+    """
+
+    def __init__(self, mapping: t.Mapping[str, str]) -> None:
+        self.mapping = mapping
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, None, t.Callable[[], bool]]:
+        if template in self.mapping:
+            source = self.mapping[template]
+            return source, None, lambda: source == self.mapping.get(template)
+        raise TemplateNotFound(template)
+
+    def list_templates(self) -> t.List[str]:
+        return sorted(self.mapping)
+
+
+class FunctionLoader(BaseLoader):
+    """A loader that is passed a function which does the loading.  The
+    function receives the name of the template and has to return either
+    a string with the template source, a tuple in the form ``(source,
+    filename, uptodatefunc)`` or `None` if the template does not exist.
+
+    >>> def load_template(name):
+    ...     if name == 'index.html':
+    ...         return '...'
+    ...
+    >>> loader = FunctionLoader(load_template)
+
+    The `uptodatefunc` is a function that is called if autoreload is enabled
+    and has to return `True` if the template is still up to date.  For more
+    details have a look at :meth:`BaseLoader.get_source` which has the same
+    return value.
+    """
+
+    def __init__(
+        self,
+        load_func: t.Callable[
+            [str],
+            t.Optional[
+                t.Union[
+                    str, t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]
+                ]
+            ],
+        ],
+    ) -> None:
+        self.load_func = load_func
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
+        rv = self.load_func(template)
+
+        if rv is None:
+            raise TemplateNotFound(template)
+
+        if isinstance(rv, str):
+            return rv, None, None
+
+        return rv
+
+
+class PrefixLoader(BaseLoader):
+    """A loader that is passed a dict of loaders where each loader is bound
+    to a prefix.  The prefix is delimited from the template by a slash per
+    default, which can be changed by setting the `delimiter` argument to
+    something else::
+
+        loader = PrefixLoader({
+            'app1':     PackageLoader('mypackage.app1'),
+            'app2':     PackageLoader('mypackage.app2')
+        })
+
+    By loading ``'app1/index.html'`` the file from the app1 package is loaded,
+    by loading ``'app2/index.html'`` the file from the second.
+    """
+
+    def __init__(
+        self, mapping: t.Mapping[str, BaseLoader], delimiter: str = "/"
+    ) -> None:
+        self.mapping = mapping
+        self.delimiter = delimiter
+
+    def get_loader(self, template: str) -> t.Tuple[BaseLoader, str]:
+        try:
+            prefix, name = template.split(self.delimiter, 1)
+            loader = self.mapping[prefix]
+        except (ValueError, KeyError) as e:
+            raise TemplateNotFound(template) from e
+        return loader, name
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
+        loader, name = self.get_loader(template)
+        try:
+            return loader.get_source(environment, name)
+        except TemplateNotFound as e:
+            # re-raise the exception with the correct filename here.
+            # (the one that includes the prefix)
+            raise TemplateNotFound(template) from e
+
+    @internalcode
+    def load(
+        self,
+        environment: "Environment",
+        name: str,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        loader, local_name = self.get_loader(name)
+        try:
+            return loader.load(environment, local_name, globals)
+        except TemplateNotFound as e:
+            # re-raise the exception with the correct filename here.
+            # (the one that includes the prefix)
+            raise TemplateNotFound(name) from e
+
+    def list_templates(self) -> t.List[str]:
+        result = []
+        for prefix, loader in self.mapping.items():
+            for template in loader.list_templates():
+                result.append(prefix + self.delimiter + template)
+        return result
+
+
+class ChoiceLoader(BaseLoader):
+    """This loader works like the `PrefixLoader` just that no prefix is
+    specified.  If a template could not be found by one loader the next one
+    is tried.
+
+    >>> loader = ChoiceLoader([
+    ...     FileSystemLoader('/path/to/user/templates'),
+    ...     FileSystemLoader('/path/to/system/templates')
+    ... ])
+
+    This is useful if you want to allow users to override builtin templates
+    from a different location.
+    """
+
+    def __init__(self, loaders: t.Sequence[BaseLoader]) -> None:
+        self.loaders = loaders
+
+    def get_source(
+        self, environment: "Environment", template: str
+    ) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
+        for loader in self.loaders:
+            try:
+                return loader.get_source(environment, template)
+            except TemplateNotFound:
+                pass
+        raise TemplateNotFound(template)
+
+    @internalcode
+    def load(
+        self,
+        environment: "Environment",
+        name: str,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        for loader in self.loaders:
+            try:
+                return loader.load(environment, name, globals)
+            except TemplateNotFound:
+                pass
+        raise TemplateNotFound(name)
+
+    def list_templates(self) -> t.List[str]:
+        found = set()
+        for loader in self.loaders:
+            found.update(loader.list_templates())
+        return sorted(found)
+
+
+class _TemplateModule(ModuleType):
+    """Like a normal module but with support for weak references"""
+
+
+class ModuleLoader(BaseLoader):
+    """This loader loads templates from precompiled templates.
+
+    Example usage:
+
+    >>> loader = ModuleLoader('/path/to/compiled/templates')
+
+    Templates can be precompiled with :meth:`Environment.compile_templates`.
+    """
+
+    has_source_access = False
+
+    def __init__(
+        self,
+        path: t.Union[
+            str, "os.PathLike[str]", t.Sequence[t.Union[str, "os.PathLike[str]"]]
+        ],
+    ) -> None:
+        package_name = f"_jinja2_module_templates_{id(self):x}"
+
+        # create a fake module that looks for the templates in the
+        # path given.
+        mod = _TemplateModule(package_name)
+
+        if not isinstance(path, abc.Iterable) or isinstance(path, str):
+            path = [path]
+
+        mod.__path__ = [os.fspath(p) for p in path]
+
+        sys.modules[package_name] = weakref.proxy(
+            mod, lambda x: sys.modules.pop(package_name, None)
+        )
+
+        # the only strong reference, the sys.modules entry is weak
+        # so that the garbage collector can remove it once the
+        # loader that created it goes out of business.
+        self.module = mod
+        self.package_name = package_name
+
+    @staticmethod
+    def get_template_key(name: str) -> str:
+        return "tmpl_" + sha1(name.encode("utf-8")).hexdigest()
+
+    @staticmethod
+    def get_module_filename(name: str) -> str:
+        return ModuleLoader.get_template_key(name) + ".py"
+
+    @internalcode
+    def load(
+        self,
+        environment: "Environment",
+        name: str,
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ) -> "Template":
+        key = self.get_template_key(name)
+        module = f"{self.package_name}.{key}"
+        mod = getattr(self.module, module, None)
+
+        if mod is None:
+            try:
+                mod = __import__(module, None, None, ["root"])
+            except ImportError as e:
+                raise TemplateNotFound(name) from e
+
+            # remove the entry from sys.modules, we only want the attribute
+            # on the module object we have stored on the loader.
+            sys.modules.pop(module, None)
+
+        if globals is None:
+            globals = {}
+
+        return environment.template_class.from_module_dict(
+            environment, mod.__dict__, globals
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/meta.py b/backend/venv/lib/python3.12/site-packages/jinja2/meta.py
new file mode 100644
index 00000000..298499e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/meta.py
@@ -0,0 +1,112 @@
+"""Functions that expose information about templates that might be
+interesting for introspection.
+"""
+
+import typing as t
+
+from . import nodes
+from .compiler import CodeGenerator
+from .compiler import Frame
+
+if t.TYPE_CHECKING:
+    from .environment import Environment
+
+
+class TrackingCodeGenerator(CodeGenerator):
+    """We abuse the code generator for introspection."""
+
+    def __init__(self, environment: "Environment") -> None:
+        super().__init__(environment, "<introspection>", "<introspection>")
+        self.undeclared_identifiers: t.Set[str] = set()
+
+    def write(self, x: str) -> None:
+        """Don't write."""
+
+    def enter_frame(self, frame: Frame) -> None:
+        """Remember all undeclared identifiers."""
+        super().enter_frame(frame)
+
+        for _, (action, param) in frame.symbols.loads.items():
+            if action == "resolve" and param not in self.environment.globals:
+                self.undeclared_identifiers.add(param)
+
+
+def find_undeclared_variables(ast: nodes.Template) -> t.Set[str]:
+    """Returns a set of all variables in the AST that will be looked up from
+    the context at runtime.  Because at compile time it's not known which
+    variables will be used depending on the path the execution takes at
+    runtime, all variables are returned.
+
+    >>> from jinja2 import Environment, meta
+    >>> env = Environment()
+    >>> ast = env.parse('{% set foo = 42 %}{{ bar + foo }}')
+    >>> meta.find_undeclared_variables(ast) == {'bar'}
+    True
+
+    .. admonition:: Implementation
+
+       Internally the code generator is used for finding undeclared variables.
+       This is good to know because the code generator might raise a
+       :exc:`TemplateAssertionError` during compilation and as a matter of
+       fact this function can currently raise that exception as well.
+    """
+    codegen = TrackingCodeGenerator(ast.environment)  # type: ignore
+    codegen.visit(ast)
+    return codegen.undeclared_identifiers
+
+
+_ref_types = (nodes.Extends, nodes.FromImport, nodes.Import, nodes.Include)
+_RefType = t.Union[nodes.Extends, nodes.FromImport, nodes.Import, nodes.Include]
+
+
+def find_referenced_templates(ast: nodes.Template) -> t.Iterator[t.Optional[str]]:
+    """Finds all the referenced templates from the AST.  This will return an
+    iterator over all the hardcoded template extensions, inclusions and
+    imports.  If dynamic inheritance or inclusion is used, `None` will be
+    yielded.
+
+    >>> from jinja2 import Environment, meta
+    >>> env = Environment()
+    >>> ast = env.parse('{% extends "layout.html" %}{% include helper %}')
+    >>> list(meta.find_referenced_templates(ast))
+    ['layout.html', None]
+
+    This function is useful for dependency tracking.  For example if you want
+    to rebuild parts of the website after a layout template has changed.
+    """
+    template_name: t.Any
+
+    for node in ast.find_all(_ref_types):
+        template: nodes.Expr = node.template  # type: ignore
+
+        if not isinstance(template, nodes.Const):
+            # a tuple with some non consts in there
+            if isinstance(template, (nodes.Tuple, nodes.List)):
+                for template_name in template.items:
+                    # something const, only yield the strings and ignore
+                    # non-string consts that really just make no sense
+                    if isinstance(template_name, nodes.Const):
+                        if isinstance(template_name.value, str):
+                            yield template_name.value
+                    # something dynamic in there
+                    else:
+                        yield None
+            # something dynamic we don't know about here
+            else:
+                yield None
+            continue
+        # constant is a basestring, direct template name
+        if isinstance(template.value, str):
+            yield template.value
+        # a tuple or list (latter *should* not happen) made of consts,
+        # yield the consts that are strings.  We could warn here for
+        # non string values
+        elif isinstance(node, nodes.Include) and isinstance(
+            template.value, (tuple, list)
+        ):
+            for template_name in template.value:
+                if isinstance(template_name, str):
+                    yield template_name
+        # something else we don't care about, we could warn here
+        else:
+            yield None
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/nativetypes.py b/backend/venv/lib/python3.12/site-packages/jinja2/nativetypes.py
new file mode 100644
index 00000000..71db8cc3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/nativetypes.py
@@ -0,0 +1,130 @@
+import typing as t
+from ast import literal_eval
+from ast import parse
+from itertools import chain
+from itertools import islice
+from types import GeneratorType
+
+from . import nodes
+from .compiler import CodeGenerator
+from .compiler import Frame
+from .compiler import has_safe_repr
+from .environment import Environment
+from .environment import Template
+
+
+def native_concat(values: t.Iterable[t.Any]) -> t.Optional[t.Any]:
+    """Return a native Python type from the list of compiled nodes. If
+    the result is a single node, its value is returned. Otherwise, the
+    nodes are concatenated as strings. If the result can be parsed with
+    :func:`ast.literal_eval`, the parsed value is returned. Otherwise,
+    the string is returned.
+
+    :param values: Iterable of outputs to concatenate.
+    """
+    head = list(islice(values, 2))
+
+    if not head:
+        return None
+
+    if len(head) == 1:
+        raw = head[0]
+        if not isinstance(raw, str):
+            return raw
+    else:
+        if isinstance(values, GeneratorType):
+            values = chain(head, values)
+        raw = "".join([str(v) for v in values])
+
+    try:
+        return literal_eval(
+            # In Python 3.10+ ast.literal_eval removes leading spaces/tabs
+            # from the given string. For backwards compatibility we need to
+            # parse the string ourselves without removing leading spaces/tabs.
+            parse(raw, mode="eval")
+        )
+    except (ValueError, SyntaxError, MemoryError):
+        return raw
+
+
+class NativeCodeGenerator(CodeGenerator):
+    """A code generator which renders Python types by not adding
+    ``str()`` around output nodes.
+    """
+
+    @staticmethod
+    def _default_finalize(value: t.Any) -> t.Any:
+        return value
+
+    def _output_const_repr(self, group: t.Iterable[t.Any]) -> str:
+        return repr("".join([str(v) for v in group]))
+
+    def _output_child_to_const(
+        self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
+    ) -> t.Any:
+        const = node.as_const(frame.eval_ctx)
+
+        if not has_safe_repr(const):
+            raise nodes.Impossible()
+
+        if isinstance(node, nodes.TemplateData):
+            return const
+
+        return finalize.const(const)  # type: ignore
+
+    def _output_child_pre(
+        self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
+    ) -> None:
+        if finalize.src is not None:
+            self.write(finalize.src)
+
+    def _output_child_post(
+        self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
+    ) -> None:
+        if finalize.src is not None:
+            self.write(")")
+
+
+class NativeEnvironment(Environment):
+    """An environment that renders templates to native Python types."""
+
+    code_generator_class = NativeCodeGenerator
+    concat = staticmethod(native_concat)  # type: ignore
+
+
+class NativeTemplate(Template):
+    environment_class = NativeEnvironment
+
+    def render(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        """Render the template to produce a native Python type. If the
+        result is a single node, its value is returned. Otherwise, the
+        nodes are concatenated as strings. If the result can be parsed
+        with :func:`ast.literal_eval`, the parsed value is returned.
+        Otherwise, the string is returned.
+        """
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            return self.environment_class.concat(  # type: ignore
+                self.root_render_func(ctx)
+            )
+        except Exception:
+            return self.environment.handle_exception()
+
+    async def render_async(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        if not self.environment.is_async:
+            raise RuntimeError(
+                "The environment was not created with async mode enabled."
+            )
+
+        ctx = self.new_context(dict(*args, **kwargs))
+
+        try:
+            return self.environment_class.concat(  # type: ignore
+                [n async for n in self.root_render_func(ctx)]  # type: ignore
+            )
+        except Exception:
+            return self.environment.handle_exception()
+
+
+NativeEnvironment.template_class = NativeTemplate
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/nodes.py b/backend/venv/lib/python3.12/site-packages/jinja2/nodes.py
new file mode 100644
index 00000000..2f93b90e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/nodes.py
@@ -0,0 +1,1206 @@
+"""AST nodes generated by the parser for the compiler. Also provides
+some node tree helper functions used by the parser and compiler in order
+to normalize nodes.
+"""
+
+import inspect
+import operator
+import typing as t
+from collections import deque
+
+from markupsafe import Markup
+
+from .utils import _PassArg
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+
+_NodeBound = t.TypeVar("_NodeBound", bound="Node")
+
+_binop_to_func: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
+    "*": operator.mul,
+    "/": operator.truediv,
+    "//": operator.floordiv,
+    "**": operator.pow,
+    "%": operator.mod,
+    "+": operator.add,
+    "-": operator.sub,
+}
+
+_uaop_to_func: t.Dict[str, t.Callable[[t.Any], t.Any]] = {
+    "not": operator.not_,
+    "+": operator.pos,
+    "-": operator.neg,
+}
+
+_cmpop_to_func: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
+    "eq": operator.eq,
+    "ne": operator.ne,
+    "gt": operator.gt,
+    "gteq": operator.ge,
+    "lt": operator.lt,
+    "lteq": operator.le,
+    "in": lambda a, b: a in b,
+    "notin": lambda a, b: a not in b,
+}
+
+
+class Impossible(Exception):
+    """Raised if the node could not perform a requested action."""
+
+
+class NodeType(type):
+    """A metaclass for nodes that handles the field and attribute
+    inheritance.  fields and attributes from the parent class are
+    automatically forwarded to the child."""
+
+    def __new__(mcs, name, bases, d):  # type: ignore
+        for attr in "fields", "attributes":
+            storage: t.List[t.Tuple[str, ...]] = []
+            storage.extend(getattr(bases[0] if bases else object, attr, ()))
+            storage.extend(d.get(attr, ()))
+            assert len(bases) <= 1, "multiple inheritance not allowed"
+            assert len(storage) == len(set(storage)), "layout conflict"
+            d[attr] = tuple(storage)
+        d.setdefault("abstract", False)
+        return type.__new__(mcs, name, bases, d)
+
+
+class EvalContext:
+    """Holds evaluation time information.  Custom attributes can be attached
+    to it in extensions.
+    """
+
+    def __init__(
+        self, environment: "Environment", template_name: t.Optional[str] = None
+    ) -> None:
+        self.environment = environment
+        if callable(environment.autoescape):
+            self.autoescape = environment.autoescape(template_name)
+        else:
+            self.autoescape = environment.autoescape
+        self.volatile = False
+
+    def save(self) -> t.Mapping[str, t.Any]:
+        return self.__dict__.copy()
+
+    def revert(self, old: t.Mapping[str, t.Any]) -> None:
+        self.__dict__.clear()
+        self.__dict__.update(old)
+
+
+def get_eval_context(node: "Node", ctx: t.Optional[EvalContext]) -> EvalContext:
+    if ctx is None:
+        if node.environment is None:
+            raise RuntimeError(
+                "if no eval context is passed, the node must have an"
+                " attached environment."
+            )
+        return EvalContext(node.environment)
+    return ctx
+
+
+class Node(metaclass=NodeType):
+    """Baseclass for all Jinja nodes.  There are a number of nodes available
+    of different types.  There are four major types:
+
+    -   :class:`Stmt`: statements
+    -   :class:`Expr`: expressions
+    -   :class:`Helper`: helper nodes
+    -   :class:`Template`: the outermost wrapper node
+
+    All nodes have fields and attributes.  Fields may be other nodes, lists,
+    or arbitrary values.  Fields are passed to the constructor as regular
+    positional arguments, attributes as keyword arguments.  Each node has
+    two attributes: `lineno` (the line number of the node) and `environment`.
+    The `environment` attribute is set at the end of the parsing process for
+    all nodes automatically.
+    """
+
+    fields: t.Tuple[str, ...] = ()
+    attributes: t.Tuple[str, ...] = ("lineno", "environment")
+    abstract = True
+
+    lineno: int
+    environment: t.Optional["Environment"]
+
+    def __init__(self, *fields: t.Any, **attributes: t.Any) -> None:
+        if self.abstract:
+            raise TypeError("abstract nodes are not instantiable")
+        if fields:
+            if len(fields) != len(self.fields):
+                if not self.fields:
+                    raise TypeError(f"{type(self).__name__!r} takes 0 arguments")
+                raise TypeError(
+                    f"{type(self).__name__!r} takes 0 or {len(self.fields)}"
+                    f" argument{'s' if len(self.fields) != 1 else ''}"
+                )
+            for name, arg in zip(self.fields, fields):
+                setattr(self, name, arg)
+        for attr in self.attributes:
+            setattr(self, attr, attributes.pop(attr, None))
+        if attributes:
+            raise TypeError(f"unknown attribute {next(iter(attributes))!r}")
+
+    def iter_fields(
+        self,
+        exclude: t.Optional[t.Container[str]] = None,
+        only: t.Optional[t.Container[str]] = None,
+    ) -> t.Iterator[t.Tuple[str, t.Any]]:
+        """This method iterates over all fields that are defined and yields
+        ``(key, value)`` tuples.  Per default all fields are returned, but
+        it's possible to limit that to some fields by providing the `only`
+        parameter or to exclude some using the `exclude` parameter.  Both
+        should be sets or tuples of field names.
+        """
+        for name in self.fields:
+            if (
+                (exclude is None and only is None)
+                or (exclude is not None and name not in exclude)
+                or (only is not None and name in only)
+            ):
+                try:
+                    yield name, getattr(self, name)
+                except AttributeError:
+                    pass
+
+    def iter_child_nodes(
+        self,
+        exclude: t.Optional[t.Container[str]] = None,
+        only: t.Optional[t.Container[str]] = None,
+    ) -> t.Iterator["Node"]:
+        """Iterates over all direct child nodes of the node.  This iterates
+        over all fields and yields the values of they are nodes.  If the value
+        of a field is a list all the nodes in that list are returned.
+        """
+        for _, item in self.iter_fields(exclude, only):
+            if isinstance(item, list):
+                for n in item:
+                    if isinstance(n, Node):
+                        yield n
+            elif isinstance(item, Node):
+                yield item
+
+    def find(self, node_type: t.Type[_NodeBound]) -> t.Optional[_NodeBound]:
+        """Find the first node of a given type.  If no such node exists the
+        return value is `None`.
+        """
+        for result in self.find_all(node_type):
+            return result
+
+        return None
+
+    def find_all(
+        self, node_type: t.Union[t.Type[_NodeBound], t.Tuple[t.Type[_NodeBound], ...]]
+    ) -> t.Iterator[_NodeBound]:
+        """Find all the nodes of a given type.  If the type is a tuple,
+        the check is performed for any of the tuple items.
+        """
+        for child in self.iter_child_nodes():
+            if isinstance(child, node_type):
+                yield child  # type: ignore
+            yield from child.find_all(node_type)
+
+    def set_ctx(self, ctx: str) -> "Node":
+        """Reset the context of a node and all child nodes.  Per default the
+        parser will all generate nodes that have a 'load' context as it's the
+        most common one.  This method is used in the parser to set assignment
+        targets and other nodes to a store context.
+        """
+        todo = deque([self])
+        while todo:
+            node = todo.popleft()
+            if "ctx" in node.fields:
+                node.ctx = ctx  # type: ignore
+            todo.extend(node.iter_child_nodes())
+        return self
+
+    def set_lineno(self, lineno: int, override: bool = False) -> "Node":
+        """Set the line numbers of the node and children."""
+        todo = deque([self])
+        while todo:
+            node = todo.popleft()
+            if "lineno" in node.attributes:
+                if node.lineno is None or override:
+                    node.lineno = lineno
+            todo.extend(node.iter_child_nodes())
+        return self
+
+    def set_environment(self, environment: "Environment") -> "Node":
+        """Set the environment for all nodes."""
+        todo = deque([self])
+        while todo:
+            node = todo.popleft()
+            node.environment = environment
+            todo.extend(node.iter_child_nodes())
+        return self
+
+    def __eq__(self, other: t.Any) -> bool:
+        if type(self) is not type(other):
+            return NotImplemented
+
+        return tuple(self.iter_fields()) == tuple(other.iter_fields())
+
+    __hash__ = object.__hash__
+
+    def __repr__(self) -> str:
+        args_str = ", ".join(f"{a}={getattr(self, a, None)!r}" for a in self.fields)
+        return f"{type(self).__name__}({args_str})"
+
+    def dump(self) -> str:
+        def _dump(node: t.Union[Node, t.Any]) -> None:
+            if not isinstance(node, Node):
+                buf.append(repr(node))
+                return
+
+            buf.append(f"nodes.{type(node).__name__}(")
+            if not node.fields:
+                buf.append(")")
+                return
+            for idx, field in enumerate(node.fields):
+                if idx:
+                    buf.append(", ")
+                value = getattr(node, field)
+                if isinstance(value, list):
+                    buf.append("[")
+                    for idx, item in enumerate(value):
+                        if idx:
+                            buf.append(", ")
+                        _dump(item)
+                    buf.append("]")
+                else:
+                    _dump(value)
+            buf.append(")")
+
+        buf: t.List[str] = []
+        _dump(self)
+        return "".join(buf)
+
+
+class Stmt(Node):
+    """Base node for all statements."""
+
+    abstract = True
+
+
+class Helper(Node):
+    """Nodes that exist in a specific context only."""
+
+    abstract = True
+
+
+class Template(Node):
+    """Node that represents a template.  This must be the outermost node that
+    is passed to the compiler.
+    """
+
+    fields = ("body",)
+    body: t.List[Node]
+
+
+class Output(Stmt):
+    """A node that holds multiple expressions which are then printed out.
+    This is used both for the `print` statement and the regular template data.
+    """
+
+    fields = ("nodes",)
+    nodes: t.List["Expr"]
+
+
+class Extends(Stmt):
+    """Represents an extends statement."""
+
+    fields = ("template",)
+    template: "Expr"
+
+
+class For(Stmt):
+    """The for loop.  `target` is the target for the iteration (usually a
+    :class:`Name` or :class:`Tuple`), `iter` the iterable.  `body` is a list
+    of nodes that are used as loop-body, and `else_` a list of nodes for the
+    `else` block.  If no else node exists it has to be an empty list.
+
+    For filtered nodes an expression can be stored as `test`, otherwise `None`.
+    """
+
+    fields = ("target", "iter", "body", "else_", "test", "recursive")
+    target: Node
+    iter: Node
+    body: t.List[Node]
+    else_: t.List[Node]
+    test: t.Optional[Node]
+    recursive: bool
+
+
+class If(Stmt):
+    """If `test` is true, `body` is rendered, else `else_`."""
+
+    fields = ("test", "body", "elif_", "else_")
+    test: Node
+    body: t.List[Node]
+    elif_: t.List["If"]
+    else_: t.List[Node]
+
+
+class Macro(Stmt):
+    """A macro definition.  `name` is the name of the macro, `args` a list of
+    arguments and `defaults` a list of defaults if there are any.  `body` is
+    a list of nodes for the macro body.
+    """
+
+    fields = ("name", "args", "defaults", "body")
+    name: str
+    args: t.List["Name"]
+    defaults: t.List["Expr"]
+    body: t.List[Node]
+
+
+class CallBlock(Stmt):
+    """Like a macro without a name but a call instead.  `call` is called with
+    the unnamed macro as `caller` argument this node holds.
+    """
+
+    fields = ("call", "args", "defaults", "body")
+    call: "Call"
+    args: t.List["Name"]
+    defaults: t.List["Expr"]
+    body: t.List[Node]
+
+
+class FilterBlock(Stmt):
+    """Node for filter sections."""
+
+    fields = ("body", "filter")
+    body: t.List[Node]
+    filter: "Filter"
+
+
+class With(Stmt):
+    """Specific node for with statements.  In older versions of Jinja the
+    with statement was implemented on the base of the `Scope` node instead.
+
+    .. versionadded:: 2.9.3
+    """
+
+    fields = ("targets", "values", "body")
+    targets: t.List["Expr"]
+    values: t.List["Expr"]
+    body: t.List[Node]
+
+
+class Block(Stmt):
+    """A node that represents a block.
+
+    .. versionchanged:: 3.0.0
+        the `required` field was added.
+    """
+
+    fields = ("name", "body", "scoped", "required")
+    name: str
+    body: t.List[Node]
+    scoped: bool
+    required: bool
+
+
+class Include(Stmt):
+    """A node that represents the include tag."""
+
+    fields = ("template", "with_context", "ignore_missing")
+    template: "Expr"
+    with_context: bool
+    ignore_missing: bool
+
+
+class Import(Stmt):
+    """A node that represents the import tag."""
+
+    fields = ("template", "target", "with_context")
+    template: "Expr"
+    target: str
+    with_context: bool
+
+
+class FromImport(Stmt):
+    """A node that represents the from import tag.  It's important to not
+    pass unsafe names to the name attribute.  The compiler translates the
+    attribute lookups directly into getattr calls and does *not* use the
+    subscript callback of the interface.  As exported variables may not
+    start with double underscores (which the parser asserts) this is not a
+    problem for regular Jinja code, but if this node is used in an extension
+    extra care must be taken.
+
+    The list of names may contain tuples if aliases are wanted.
+    """
+
+    fields = ("template", "names", "with_context")
+    template: "Expr"
+    names: t.List[t.Union[str, t.Tuple[str, str]]]
+    with_context: bool
+
+
+class ExprStmt(Stmt):
+    """A statement that evaluates an expression and discards the result."""
+
+    fields = ("node",)
+    node: Node
+
+
+class Assign(Stmt):
+    """Assigns an expression to a target."""
+
+    fields = ("target", "node")
+    target: "Expr"
+    node: Node
+
+
+class AssignBlock(Stmt):
+    """Assigns a block to a target."""
+
+    fields = ("target", "filter", "body")
+    target: "Expr"
+    filter: t.Optional["Filter"]
+    body: t.List[Node]
+
+
+class Expr(Node):
+    """Baseclass for all expressions."""
+
+    abstract = True
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        """Return the value of the expression as constant or raise
+        :exc:`Impossible` if this was not possible.
+
+        An :class:`EvalContext` can be provided, if none is given
+        a default context is created which requires the nodes to have
+        an attached environment.
+
+        .. versionchanged:: 2.4
+           the `eval_ctx` parameter was added.
+        """
+        raise Impossible()
+
+    def can_assign(self) -> bool:
+        """Check if it's possible to assign something to this node."""
+        return False
+
+
+class BinExpr(Expr):
+    """Baseclass for all binary expressions."""
+
+    fields = ("left", "right")
+    left: Expr
+    right: Expr
+    operator: str
+    abstract = True
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        # intercepted operators cannot be folded at compile time
+        if (
+            eval_ctx.environment.sandboxed
+            and self.operator in eval_ctx.environment.intercepted_binops  # type: ignore
+        ):
+            raise Impossible()
+        f = _binop_to_func[self.operator]
+        try:
+            return f(self.left.as_const(eval_ctx), self.right.as_const(eval_ctx))
+        except Exception as e:
+            raise Impossible() from e
+
+
+class UnaryExpr(Expr):
+    """Baseclass for all unary expressions."""
+
+    fields = ("node",)
+    node: Expr
+    operator: str
+    abstract = True
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        # intercepted operators cannot be folded at compile time
+        if (
+            eval_ctx.environment.sandboxed
+            and self.operator in eval_ctx.environment.intercepted_unops  # type: ignore
+        ):
+            raise Impossible()
+        f = _uaop_to_func[self.operator]
+        try:
+            return f(self.node.as_const(eval_ctx))
+        except Exception as e:
+            raise Impossible() from e
+
+
+class Name(Expr):
+    """Looks up a name or stores a value in a name.
+    The `ctx` of the node can be one of the following values:
+
+    -   `store`: store a value in the name
+    -   `load`: load that name
+    -   `param`: like `store` but if the name was defined as function parameter.
+    """
+
+    fields = ("name", "ctx")
+    name: str
+    ctx: str
+
+    def can_assign(self) -> bool:
+        return self.name not in {"true", "false", "none", "True", "False", "None"}
+
+
+class NSRef(Expr):
+    """Reference to a namespace value assignment"""
+
+    fields = ("name", "attr")
+    name: str
+    attr: str
+
+    def can_assign(self) -> bool:
+        # We don't need any special checks here; NSRef assignments have a
+        # runtime check to ensure the target is a namespace object which will
+        # have been checked already as it is created using a normal assignment
+        # which goes through a `Name` node.
+        return True
+
+
+class Literal(Expr):
+    """Baseclass for literals."""
+
+    abstract = True
+
+
+class Const(Literal):
+    """All constant values.  The parser will return this node for simple
+    constants such as ``42`` or ``"foo"`` but it can be used to store more
+    complex values such as lists too.  Only constants with a safe
+    representation (objects where ``eval(repr(x)) == x`` is true).
+    """
+
+    fields = ("value",)
+    value: t.Any
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        return self.value
+
+    @classmethod
+    def from_untrusted(
+        cls,
+        value: t.Any,
+        lineno: t.Optional[int] = None,
+        environment: "t.Optional[Environment]" = None,
+    ) -> "Const":
+        """Return a const object if the value is representable as
+        constant value in the generated code, otherwise it will raise
+        an `Impossible` exception.
+        """
+        from .compiler import has_safe_repr
+
+        if not has_safe_repr(value):
+            raise Impossible()
+        return cls(value, lineno=lineno, environment=environment)
+
+
+class TemplateData(Literal):
+    """A constant template string."""
+
+    fields = ("data",)
+    data: str
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> str:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        if eval_ctx.volatile:
+            raise Impossible()
+        if eval_ctx.autoescape:
+            return Markup(self.data)
+        return self.data
+
+
+class Tuple(Literal):
+    """For loop unpacking and some other things like multiple arguments
+    for subscripts.  Like for :class:`Name` `ctx` specifies if the tuple
+    is used for loading the names or storing.
+    """
+
+    fields = ("items", "ctx")
+    items: t.List[Expr]
+    ctx: str
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Tuple[t.Any, ...]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return tuple(x.as_const(eval_ctx) for x in self.items)
+
+    def can_assign(self) -> bool:
+        for item in self.items:
+            if not item.can_assign():
+                return False
+        return True
+
+
+class List(Literal):
+    """Any list literal such as ``[1, 2, 3]``"""
+
+    fields = ("items",)
+    items: t.List[Expr]
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.List[t.Any]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return [x.as_const(eval_ctx) for x in self.items]
+
+
+class Dict(Literal):
+    """Any dict literal such as ``{1: 2, 3: 4}``.  The items must be a list of
+    :class:`Pair` nodes.
+    """
+
+    fields = ("items",)
+    items: t.List["Pair"]
+
+    def as_const(
+        self, eval_ctx: t.Optional[EvalContext] = None
+    ) -> t.Dict[t.Any, t.Any]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return dict(x.as_const(eval_ctx) for x in self.items)
+
+
+class Pair(Helper):
+    """A key, value pair for dicts."""
+
+    fields = ("key", "value")
+    key: Expr
+    value: Expr
+
+    def as_const(
+        self, eval_ctx: t.Optional[EvalContext] = None
+    ) -> t.Tuple[t.Any, t.Any]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return self.key.as_const(eval_ctx), self.value.as_const(eval_ctx)
+
+
+class Keyword(Helper):
+    """A key, value pair for keyword arguments where key is a string."""
+
+    fields = ("key", "value")
+    key: str
+    value: Expr
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Tuple[str, t.Any]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return self.key, self.value.as_const(eval_ctx)
+
+
+class CondExpr(Expr):
+    """A conditional expression (inline if expression).  (``{{
+    foo if bar else baz }}``)
+    """
+
+    fields = ("test", "expr1", "expr2")
+    test: Expr
+    expr1: Expr
+    expr2: t.Optional[Expr]
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        if self.test.as_const(eval_ctx):
+            return self.expr1.as_const(eval_ctx)
+
+        # if we evaluate to an undefined object, we better do that at runtime
+        if self.expr2 is None:
+            raise Impossible()
+
+        return self.expr2.as_const(eval_ctx)
+
+
+def args_as_const(
+    node: t.Union["_FilterTestCommon", "Call"], eval_ctx: t.Optional[EvalContext]
+) -> t.Tuple[t.List[t.Any], t.Dict[t.Any, t.Any]]:
+    args = [x.as_const(eval_ctx) for x in node.args]
+    kwargs = dict(x.as_const(eval_ctx) for x in node.kwargs)
+
+    if node.dyn_args is not None:
+        try:
+            args.extend(node.dyn_args.as_const(eval_ctx))
+        except Exception as e:
+            raise Impossible() from e
+
+    if node.dyn_kwargs is not None:
+        try:
+            kwargs.update(node.dyn_kwargs.as_const(eval_ctx))
+        except Exception as e:
+            raise Impossible() from e
+
+    return args, kwargs
+
+
+class _FilterTestCommon(Expr):
+    fields = ("node", "name", "args", "kwargs", "dyn_args", "dyn_kwargs")
+    node: Expr
+    name: str
+    args: t.List[Expr]
+    kwargs: t.List[Pair]
+    dyn_args: t.Optional[Expr]
+    dyn_kwargs: t.Optional[Expr]
+    abstract = True
+    _is_filter = True
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        if eval_ctx.volatile:
+            raise Impossible()
+
+        if self._is_filter:
+            env_map = eval_ctx.environment.filters
+        else:
+            env_map = eval_ctx.environment.tests
+
+        func = env_map.get(self.name)
+        pass_arg = _PassArg.from_obj(func)  # type: ignore
+
+        if func is None or pass_arg is _PassArg.context:
+            raise Impossible()
+
+        if eval_ctx.environment.is_async and (
+            getattr(func, "jinja_async_variant", False) is True
+            or inspect.iscoroutinefunction(func)
+        ):
+            raise Impossible()
+
+        args, kwargs = args_as_const(self, eval_ctx)
+        args.insert(0, self.node.as_const(eval_ctx))
+
+        if pass_arg is _PassArg.eval_context:
+            args.insert(0, eval_ctx)
+        elif pass_arg is _PassArg.environment:
+            args.insert(0, eval_ctx.environment)
+
+        try:
+            return func(*args, **kwargs)
+        except Exception as e:
+            raise Impossible() from e
+
+
+class Filter(_FilterTestCommon):
+    """Apply a filter to an expression. ``name`` is the name of the
+    filter, the other fields are the same as :class:`Call`.
+
+    If ``node`` is ``None``, the filter is being used in a filter block
+    and is applied to the content of the block.
+    """
+
+    node: t.Optional[Expr]  # type: ignore
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        if self.node is None:
+            raise Impossible()
+
+        return super().as_const(eval_ctx=eval_ctx)
+
+
+class Test(_FilterTestCommon):
+    """Apply a test to an expression. ``name`` is the name of the test,
+    the other field are the same as :class:`Call`.
+
+    .. versionchanged:: 3.0
+        ``as_const`` shares the same logic for filters and tests. Tests
+        check for volatile, async, and ``@pass_context`` etc.
+        decorators.
+    """
+
+    _is_filter = False
+
+
+class Call(Expr):
+    """Calls an expression.  `args` is a list of arguments, `kwargs` a list
+    of keyword arguments (list of :class:`Keyword` nodes), and `dyn_args`
+    and `dyn_kwargs` has to be either `None` or a node that is used as
+    node for dynamic positional (``*args``) or keyword (``**kwargs``)
+    arguments.
+    """
+
+    fields = ("node", "args", "kwargs", "dyn_args", "dyn_kwargs")
+    node: Expr
+    args: t.List[Expr]
+    kwargs: t.List[Keyword]
+    dyn_args: t.Optional[Expr]
+    dyn_kwargs: t.Optional[Expr]
+
+
+class Getitem(Expr):
+    """Get an attribute or item from an expression and prefer the item."""
+
+    fields = ("node", "arg", "ctx")
+    node: Expr
+    arg: Expr
+    ctx: str
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        if self.ctx != "load":
+            raise Impossible()
+
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        try:
+            return eval_ctx.environment.getitem(
+                self.node.as_const(eval_ctx), self.arg.as_const(eval_ctx)
+            )
+        except Exception as e:
+            raise Impossible() from e
+
+
+class Getattr(Expr):
+    """Get an attribute or item from an expression that is a ascii-only
+    bytestring and prefer the attribute.
+    """
+
+    fields = ("node", "attr", "ctx")
+    node: Expr
+    attr: str
+    ctx: str
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        if self.ctx != "load":
+            raise Impossible()
+
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        try:
+            return eval_ctx.environment.getattr(self.node.as_const(eval_ctx), self.attr)
+        except Exception as e:
+            raise Impossible() from e
+
+
+class Slice(Expr):
+    """Represents a slice object.  This must only be used as argument for
+    :class:`Subscript`.
+    """
+
+    fields = ("start", "stop", "step")
+    start: t.Optional[Expr]
+    stop: t.Optional[Expr]
+    step: t.Optional[Expr]
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> slice:
+        eval_ctx = get_eval_context(self, eval_ctx)
+
+        def const(obj: t.Optional[Expr]) -> t.Optional[t.Any]:
+            if obj is None:
+                return None
+            return obj.as_const(eval_ctx)
+
+        return slice(const(self.start), const(self.stop), const(self.step))
+
+
+class Concat(Expr):
+    """Concatenates the list of expressions provided after converting
+    them to strings.
+    """
+
+    fields = ("nodes",)
+    nodes: t.List[Expr]
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> str:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return "".join(str(x.as_const(eval_ctx)) for x in self.nodes)
+
+
+class Compare(Expr):
+    """Compares an expression with some other expressions.  `ops` must be a
+    list of :class:`Operand`\\s.
+    """
+
+    fields = ("expr", "ops")
+    expr: Expr
+    ops: t.List["Operand"]
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        result = value = self.expr.as_const(eval_ctx)
+
+        try:
+            for op in self.ops:
+                new_value = op.expr.as_const(eval_ctx)
+                result = _cmpop_to_func[op.op](value, new_value)
+
+                if not result:
+                    return False
+
+                value = new_value
+        except Exception as e:
+            raise Impossible() from e
+
+        return result
+
+
+class Operand(Helper):
+    """Holds an operator and an expression."""
+
+    fields = ("op", "expr")
+    op: str
+    expr: Expr
+
+
+class Mul(BinExpr):
+    """Multiplies the left with the right node."""
+
+    operator = "*"
+
+
+class Div(BinExpr):
+    """Divides the left by the right node."""
+
+    operator = "/"
+
+
+class FloorDiv(BinExpr):
+    """Divides the left by the right node and converts the
+    result into an integer by truncating.
+    """
+
+    operator = "//"
+
+
+class Add(BinExpr):
+    """Add the left to the right node."""
+
+    operator = "+"
+
+
+class Sub(BinExpr):
+    """Subtract the right from the left node."""
+
+    operator = "-"
+
+
+class Mod(BinExpr):
+    """Left modulo right."""
+
+    operator = "%"
+
+
+class Pow(BinExpr):
+    """Left to the power of right."""
+
+    operator = "**"
+
+
+class And(BinExpr):
+    """Short circuited AND."""
+
+    operator = "and"
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return self.left.as_const(eval_ctx) and self.right.as_const(eval_ctx)
+
+
+class Or(BinExpr):
+    """Short circuited OR."""
+
+    operator = "or"
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return self.left.as_const(eval_ctx) or self.right.as_const(eval_ctx)
+
+
+class Not(UnaryExpr):
+    """Negate the expression."""
+
+    operator = "not"
+
+
+class Neg(UnaryExpr):
+    """Make the expression negative."""
+
+    operator = "-"
+
+
+class Pos(UnaryExpr):
+    """Make the expression positive (noop for most expressions)"""
+
+    operator = "+"
+
+
+# Helpers for extensions
+
+
+class EnvironmentAttribute(Expr):
+    """Loads an attribute from the environment object.  This is useful for
+    extensions that want to call a callback stored on the environment.
+    """
+
+    fields = ("name",)
+    name: str
+
+
+class ExtensionAttribute(Expr):
+    """Returns the attribute of an extension bound to the environment.
+    The identifier is the identifier of the :class:`Extension`.
+
+    This node is usually constructed by calling the
+    :meth:`~jinja2.ext.Extension.attr` method on an extension.
+    """
+
+    fields = ("identifier", "name")
+    identifier: str
+    name: str
+
+
+class ImportedName(Expr):
+    """If created with an import name the import name is returned on node
+    access.  For example ``ImportedName('cgi.escape')`` returns the `escape`
+    function from the cgi module on evaluation.  Imports are optimized by the
+    compiler so there is no need to assign them to local variables.
+    """
+
+    fields = ("importname",)
+    importname: str
+
+
+class InternalName(Expr):
+    """An internal name in the compiler.  You cannot create these nodes
+    yourself but the parser provides a
+    :meth:`~jinja2.parser.Parser.free_identifier` method that creates
+    a new identifier for you.  This identifier is not available from the
+    template and is not treated specially by the compiler.
+    """
+
+    fields = ("name",)
+    name: str
+
+    def __init__(self) -> None:
+        raise TypeError(
+            "Can't create internal names.  Use the "
+            "`free_identifier` method on a parser."
+        )
+
+
+class MarkSafe(Expr):
+    """Mark the wrapped expression as safe (wrap it as `Markup`)."""
+
+    fields = ("expr",)
+    expr: Expr
+
+    def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> Markup:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        return Markup(self.expr.as_const(eval_ctx))
+
+
+class MarkSafeIfAutoescape(Expr):
+    """Mark the wrapped expression as safe (wrap it as `Markup`) but
+    only if autoescaping is active.
+
+    .. versionadded:: 2.5
+    """
+
+    fields = ("expr",)
+    expr: Expr
+
+    def as_const(
+        self, eval_ctx: t.Optional[EvalContext] = None
+    ) -> t.Union[Markup, t.Any]:
+        eval_ctx = get_eval_context(self, eval_ctx)
+        if eval_ctx.volatile:
+            raise Impossible()
+        expr = self.expr.as_const(eval_ctx)
+        if eval_ctx.autoescape:
+            return Markup(expr)
+        return expr
+
+
+class ContextReference(Expr):
+    """Returns the current template context.  It can be used like a
+    :class:`Name` node, with a ``'load'`` ctx and will return the
+    current :class:`~jinja2.runtime.Context` object.
+
+    Here an example that assigns the current template name to a
+    variable named `foo`::
+
+        Assign(Name('foo', ctx='store'),
+               Getattr(ContextReference(), 'name'))
+
+    This is basically equivalent to using the
+    :func:`~jinja2.pass_context` decorator when using the high-level
+    API, which causes a reference to the context to be passed as the
+    first argument to a function.
+    """
+
+
+class DerivedContextReference(Expr):
+    """Return the current template context including locals. Behaves
+    exactly like :class:`ContextReference`, but includes local
+    variables, such as from a ``for`` loop.
+
+    .. versionadded:: 2.11
+    """
+
+
+class Continue(Stmt):
+    """Continue a loop."""
+
+
+class Break(Stmt):
+    """Break a loop."""
+
+
+class Scope(Stmt):
+    """An artificial scope."""
+
+    fields = ("body",)
+    body: t.List[Node]
+
+
+class OverlayScope(Stmt):
+    """An overlay scope for extensions.  This is a largely unoptimized scope
+    that however can be used to introduce completely arbitrary variables into
+    a sub scope from a dictionary or dictionary like object.  The `context`
+    field has to evaluate to a dictionary object.
+
+    Example usage::
+
+        OverlayScope(context=self.call_method('get_context'),
+                     body=[...])
+
+    .. versionadded:: 2.10
+    """
+
+    fields = ("context", "body")
+    context: Expr
+    body: t.List[Node]
+
+
+class EvalContextModifier(Stmt):
+    """Modifies the eval context.  For each option that should be modified,
+    a :class:`Keyword` has to be added to the :attr:`options` list.
+
+    Example to change the `autoescape` setting::
+
+        EvalContextModifier(options=[Keyword('autoescape', Const(True))])
+    """
+
+    fields = ("options",)
+    options: t.List[Keyword]
+
+
+class ScopedEvalContextModifier(EvalContextModifier):
+    """Modifies the eval context and reverts it later.  Works exactly like
+    :class:`EvalContextModifier` but will only modify the
+    :class:`~jinja2.nodes.EvalContext` for nodes in the :attr:`body`.
+    """
+
+    fields = ("body",)
+    body: t.List[Node]
+
+
+# make sure nobody creates custom nodes
+def _failing_new(*args: t.Any, **kwargs: t.Any) -> "te.NoReturn":
+    raise TypeError("can't create custom node types")
+
+
+NodeType.__new__ = staticmethod(_failing_new)  # type: ignore
+del _failing_new
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/optimizer.py b/backend/venv/lib/python3.12/site-packages/jinja2/optimizer.py
new file mode 100644
index 00000000..32d1c717
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/optimizer.py
@@ -0,0 +1,48 @@
+"""The optimizer tries to constant fold expressions and modify the AST
+in place so that it should be faster to evaluate.
+
+Because the AST does not contain all the scoping information and the
+compiler has to find that out, we cannot do all the optimizations we
+want. For example, loop unrolling doesn't work because unrolled loops
+would have a different scope. The solution would be a second syntax tree
+that stored the scoping rules.
+"""
+
+import typing as t
+
+from . import nodes
+from .visitor import NodeTransformer
+
+if t.TYPE_CHECKING:
+    from .environment import Environment
+
+
+def optimize(node: nodes.Node, environment: "Environment") -> nodes.Node:
+    """The context hint can be used to perform an static optimization
+    based on the context given."""
+    optimizer = Optimizer(environment)
+    return t.cast(nodes.Node, optimizer.visit(node))
+
+
+class Optimizer(NodeTransformer):
+    def __init__(self, environment: "t.Optional[Environment]") -> None:
+        self.environment = environment
+
+    def generic_visit(
+        self, node: nodes.Node, *args: t.Any, **kwargs: t.Any
+    ) -> nodes.Node:
+        node = super().generic_visit(node, *args, **kwargs)
+
+        # Do constant folding. Some other nodes besides Expr have
+        # as_const, but folding them causes errors later on.
+        if isinstance(node, nodes.Expr):
+            try:
+                return nodes.Const.from_untrusted(
+                    node.as_const(args[0] if args else None),
+                    lineno=node.lineno,
+                    environment=self.environment,
+                )
+            except nodes.Impossible:
+                pass
+
+        return node
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/parser.py b/backend/venv/lib/python3.12/site-packages/jinja2/parser.py
new file mode 100644
index 00000000..f4117754
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/parser.py
@@ -0,0 +1,1049 @@
+"""Parse tokens from the lexer into nodes for the compiler."""
+
+import typing
+import typing as t
+
+from . import nodes
+from .exceptions import TemplateAssertionError
+from .exceptions import TemplateSyntaxError
+from .lexer import describe_token
+from .lexer import describe_token_expr
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    from .environment import Environment
+
+_ImportInclude = t.TypeVar("_ImportInclude", nodes.Import, nodes.Include)
+_MacroCall = t.TypeVar("_MacroCall", nodes.Macro, nodes.CallBlock)
+
+_statement_keywords = frozenset(
+    [
+        "for",
+        "if",
+        "block",
+        "extends",
+        "print",
+        "macro",
+        "include",
+        "from",
+        "import",
+        "set",
+        "with",
+        "autoescape",
+    ]
+)
+_compare_operators = frozenset(["eq", "ne", "lt", "lteq", "gt", "gteq"])
+
+_math_nodes: t.Dict[str, t.Type[nodes.Expr]] = {
+    "add": nodes.Add,
+    "sub": nodes.Sub,
+    "mul": nodes.Mul,
+    "div": nodes.Div,
+    "floordiv": nodes.FloorDiv,
+    "mod": nodes.Mod,
+}
+
+
+class Parser:
+    """This is the central parsing class Jinja uses.  It's passed to
+    extensions and can be used to parse expressions or statements.
+    """
+
+    def __init__(
+        self,
+        environment: "Environment",
+        source: str,
+        name: t.Optional[str] = None,
+        filename: t.Optional[str] = None,
+        state: t.Optional[str] = None,
+    ) -> None:
+        self.environment = environment
+        self.stream = environment._tokenize(source, name, filename, state)
+        self.name = name
+        self.filename = filename
+        self.closed = False
+        self.extensions: t.Dict[
+            str, t.Callable[[Parser], t.Union[nodes.Node, t.List[nodes.Node]]]
+        ] = {}
+        for extension in environment.iter_extensions():
+            for tag in extension.tags:
+                self.extensions[tag] = extension.parse
+        self._last_identifier = 0
+        self._tag_stack: t.List[str] = []
+        self._end_token_stack: t.List[t.Tuple[str, ...]] = []
+
+    def fail(
+        self,
+        msg: str,
+        lineno: t.Optional[int] = None,
+        exc: t.Type[TemplateSyntaxError] = TemplateSyntaxError,
+    ) -> "te.NoReturn":
+        """Convenience method that raises `exc` with the message, passed
+        line number or last line number as well as the current name and
+        filename.
+        """
+        if lineno is None:
+            lineno = self.stream.current.lineno
+        raise exc(msg, lineno, self.name, self.filename)
+
+    def _fail_ut_eof(
+        self,
+        name: t.Optional[str],
+        end_token_stack: t.List[t.Tuple[str, ...]],
+        lineno: t.Optional[int],
+    ) -> "te.NoReturn":
+        expected: t.Set[str] = set()
+        for exprs in end_token_stack:
+            expected.update(map(describe_token_expr, exprs))
+        if end_token_stack:
+            currently_looking: t.Optional[str] = " or ".join(
+                map(repr, map(describe_token_expr, end_token_stack[-1]))
+            )
+        else:
+            currently_looking = None
+
+        if name is None:
+            message = ["Unexpected end of template."]
+        else:
+            message = [f"Encountered unknown tag {name!r}."]
+
+        if currently_looking:
+            if name is not None and name in expected:
+                message.append(
+                    "You probably made a nesting mistake. Jinja is expecting this tag,"
+                    f" but currently looking for {currently_looking}."
+                )
+            else:
+                message.append(
+                    f"Jinja was looking for the following tags: {currently_looking}."
+                )
+
+        if self._tag_stack:
+            message.append(
+                "The innermost block that needs to be closed is"
+                f" {self._tag_stack[-1]!r}."
+            )
+
+        self.fail(" ".join(message), lineno)
+
+    def fail_unknown_tag(
+        self, name: str, lineno: t.Optional[int] = None
+    ) -> "te.NoReturn":
+        """Called if the parser encounters an unknown tag.  Tries to fail
+        with a human readable error message that could help to identify
+        the problem.
+        """
+        self._fail_ut_eof(name, self._end_token_stack, lineno)
+
+    def fail_eof(
+        self,
+        end_tokens: t.Optional[t.Tuple[str, ...]] = None,
+        lineno: t.Optional[int] = None,
+    ) -> "te.NoReturn":
+        """Like fail_unknown_tag but for end of template situations."""
+        stack = list(self._end_token_stack)
+        if end_tokens is not None:
+            stack.append(end_tokens)
+        self._fail_ut_eof(None, stack, lineno)
+
+    def is_tuple_end(
+        self, extra_end_rules: t.Optional[t.Tuple[str, ...]] = None
+    ) -> bool:
+        """Are we at the end of a tuple?"""
+        if self.stream.current.type in ("variable_end", "block_end", "rparen"):
+            return True
+        elif extra_end_rules is not None:
+            return self.stream.current.test_any(extra_end_rules)  # type: ignore
+        return False
+
+    def free_identifier(self, lineno: t.Optional[int] = None) -> nodes.InternalName:
+        """Return a new free identifier as :class:`~jinja2.nodes.InternalName`."""
+        self._last_identifier += 1
+        rv = object.__new__(nodes.InternalName)
+        nodes.Node.__init__(rv, f"fi{self._last_identifier}", lineno=lineno)
+        return rv
+
+    def parse_statement(self) -> t.Union[nodes.Node, t.List[nodes.Node]]:
+        """Parse a single statement."""
+        token = self.stream.current
+        if token.type != "name":
+            self.fail("tag name expected", token.lineno)
+        self._tag_stack.append(token.value)
+        pop_tag = True
+        try:
+            if token.value in _statement_keywords:
+                f = getattr(self, f"parse_{self.stream.current.value}")
+                return f()  # type: ignore
+            if token.value == "call":
+                return self.parse_call_block()
+            if token.value == "filter":
+                return self.parse_filter_block()
+            ext = self.extensions.get(token.value)
+            if ext is not None:
+                return ext(self)
+
+            # did not work out, remove the token we pushed by accident
+            # from the stack so that the unknown tag fail function can
+            # produce a proper error message.
+            self._tag_stack.pop()
+            pop_tag = False
+            self.fail_unknown_tag(token.value, token.lineno)
+        finally:
+            if pop_tag:
+                self._tag_stack.pop()
+
+    def parse_statements(
+        self, end_tokens: t.Tuple[str, ...], drop_needle: bool = False
+    ) -> t.List[nodes.Node]:
+        """Parse multiple statements into a list until one of the end tokens
+        is reached.  This is used to parse the body of statements as it also
+        parses template data if appropriate.  The parser checks first if the
+        current token is a colon and skips it if there is one.  Then it checks
+        for the block end and parses until if one of the `end_tokens` is
+        reached.  Per default the active token in the stream at the end of
+        the call is the matched end token.  If this is not wanted `drop_needle`
+        can be set to `True` and the end token is removed.
+        """
+        # the first token may be a colon for python compatibility
+        self.stream.skip_if("colon")
+
+        # in the future it would be possible to add whole code sections
+        # by adding some sort of end of statement token and parsing those here.
+        self.stream.expect("block_end")
+        result = self.subparse(end_tokens)
+
+        # we reached the end of the template too early, the subparser
+        # does not check for this, so we do that now
+        if self.stream.current.type == "eof":
+            self.fail_eof(end_tokens)
+
+        if drop_needle:
+            next(self.stream)
+        return result
+
+    def parse_set(self) -> t.Union[nodes.Assign, nodes.AssignBlock]:
+        """Parse an assign statement."""
+        lineno = next(self.stream).lineno
+        target = self.parse_assign_target(with_namespace=True)
+        if self.stream.skip_if("assign"):
+            expr = self.parse_tuple()
+            return nodes.Assign(target, expr, lineno=lineno)
+        filter_node = self.parse_filter(None)
+        body = self.parse_statements(("name:endset",), drop_needle=True)
+        return nodes.AssignBlock(target, filter_node, body, lineno=lineno)
+
+    def parse_for(self) -> nodes.For:
+        """Parse a for loop."""
+        lineno = self.stream.expect("name:for").lineno
+        target = self.parse_assign_target(extra_end_rules=("name:in",))
+        self.stream.expect("name:in")
+        iter = self.parse_tuple(
+            with_condexpr=False, extra_end_rules=("name:recursive",)
+        )
+        test = None
+        if self.stream.skip_if("name:if"):
+            test = self.parse_expression()
+        recursive = self.stream.skip_if("name:recursive")
+        body = self.parse_statements(("name:endfor", "name:else"))
+        if next(self.stream).value == "endfor":
+            else_ = []
+        else:
+            else_ = self.parse_statements(("name:endfor",), drop_needle=True)
+        return nodes.For(target, iter, body, else_, test, recursive, lineno=lineno)
+
+    def parse_if(self) -> nodes.If:
+        """Parse an if construct."""
+        node = result = nodes.If(lineno=self.stream.expect("name:if").lineno)
+        while True:
+            node.test = self.parse_tuple(with_condexpr=False)
+            node.body = self.parse_statements(("name:elif", "name:else", "name:endif"))
+            node.elif_ = []
+            node.else_ = []
+            token = next(self.stream)
+            if token.test("name:elif"):
+                node = nodes.If(lineno=self.stream.current.lineno)
+                result.elif_.append(node)
+                continue
+            elif token.test("name:else"):
+                result.else_ = self.parse_statements(("name:endif",), drop_needle=True)
+            break
+        return result
+
+    def parse_with(self) -> nodes.With:
+        node = nodes.With(lineno=next(self.stream).lineno)
+        targets: t.List[nodes.Expr] = []
+        values: t.List[nodes.Expr] = []
+        while self.stream.current.type != "block_end":
+            if targets:
+                self.stream.expect("comma")
+            target = self.parse_assign_target()
+            target.set_ctx("param")
+            targets.append(target)
+            self.stream.expect("assign")
+            values.append(self.parse_expression())
+        node.targets = targets
+        node.values = values
+        node.body = self.parse_statements(("name:endwith",), drop_needle=True)
+        return node
+
+    def parse_autoescape(self) -> nodes.Scope:
+        node = nodes.ScopedEvalContextModifier(lineno=next(self.stream).lineno)
+        node.options = [nodes.Keyword("autoescape", self.parse_expression())]
+        node.body = self.parse_statements(("name:endautoescape",), drop_needle=True)
+        return nodes.Scope([node])
+
+    def parse_block(self) -> nodes.Block:
+        node = nodes.Block(lineno=next(self.stream).lineno)
+        node.name = self.stream.expect("name").value
+        node.scoped = self.stream.skip_if("name:scoped")
+        node.required = self.stream.skip_if("name:required")
+
+        # common problem people encounter when switching from django
+        # to jinja.  we do not support hyphens in block names, so let's
+        # raise a nicer error message in that case.
+        if self.stream.current.type == "sub":
+            self.fail(
+                "Block names in Jinja have to be valid Python identifiers and may not"
+                " contain hyphens, use an underscore instead."
+            )
+
+        node.body = self.parse_statements(("name:endblock",), drop_needle=True)
+
+        # enforce that required blocks only contain whitespace or comments
+        # by asserting that the body, if not empty, is just TemplateData nodes
+        # with whitespace data
+        if node.required:
+            for body_node in node.body:
+                if not isinstance(body_node, nodes.Output) or any(
+                    not isinstance(output_node, nodes.TemplateData)
+                    or not output_node.data.isspace()
+                    for output_node in body_node.nodes
+                ):
+                    self.fail("Required blocks can only contain comments or whitespace")
+
+        self.stream.skip_if("name:" + node.name)
+        return node
+
+    def parse_extends(self) -> nodes.Extends:
+        node = nodes.Extends(lineno=next(self.stream).lineno)
+        node.template = self.parse_expression()
+        return node
+
+    def parse_import_context(
+        self, node: _ImportInclude, default: bool
+    ) -> _ImportInclude:
+        if self.stream.current.test_any(
+            "name:with", "name:without"
+        ) and self.stream.look().test("name:context"):
+            node.with_context = next(self.stream).value == "with"
+            self.stream.skip()
+        else:
+            node.with_context = default
+        return node
+
+    def parse_include(self) -> nodes.Include:
+        node = nodes.Include(lineno=next(self.stream).lineno)
+        node.template = self.parse_expression()
+        if self.stream.current.test("name:ignore") and self.stream.look().test(
+            "name:missing"
+        ):
+            node.ignore_missing = True
+            self.stream.skip(2)
+        else:
+            node.ignore_missing = False
+        return self.parse_import_context(node, True)
+
+    def parse_import(self) -> nodes.Import:
+        node = nodes.Import(lineno=next(self.stream).lineno)
+        node.template = self.parse_expression()
+        self.stream.expect("name:as")
+        node.target = self.parse_assign_target(name_only=True).name
+        return self.parse_import_context(node, False)
+
+    def parse_from(self) -> nodes.FromImport:
+        node = nodes.FromImport(lineno=next(self.stream).lineno)
+        node.template = self.parse_expression()
+        self.stream.expect("name:import")
+        node.names = []
+
+        def parse_context() -> bool:
+            if self.stream.current.value in {
+                "with",
+                "without",
+            } and self.stream.look().test("name:context"):
+                node.with_context = next(self.stream).value == "with"
+                self.stream.skip()
+                return True
+            return False
+
+        while True:
+            if node.names:
+                self.stream.expect("comma")
+            if self.stream.current.type == "name":
+                if parse_context():
+                    break
+                target = self.parse_assign_target(name_only=True)
+                if target.name.startswith("_"):
+                    self.fail(
+                        "names starting with an underline can not be imported",
+                        target.lineno,
+                        exc=TemplateAssertionError,
+                    )
+                if self.stream.skip_if("name:as"):
+                    alias = self.parse_assign_target(name_only=True)
+                    node.names.append((target.name, alias.name))
+                else:
+                    node.names.append(target.name)
+                if parse_context() or self.stream.current.type != "comma":
+                    break
+            else:
+                self.stream.expect("name")
+        if not hasattr(node, "with_context"):
+            node.with_context = False
+        return node
+
+    def parse_signature(self, node: _MacroCall) -> None:
+        args = node.args = []
+        defaults = node.defaults = []
+        self.stream.expect("lparen")
+        while self.stream.current.type != "rparen":
+            if args:
+                self.stream.expect("comma")
+            arg = self.parse_assign_target(name_only=True)
+            arg.set_ctx("param")
+            if self.stream.skip_if("assign"):
+                defaults.append(self.parse_expression())
+            elif defaults:
+                self.fail("non-default argument follows default argument")
+            args.append(arg)
+        self.stream.expect("rparen")
+
+    def parse_call_block(self) -> nodes.CallBlock:
+        node = nodes.CallBlock(lineno=next(self.stream).lineno)
+        if self.stream.current.type == "lparen":
+            self.parse_signature(node)
+        else:
+            node.args = []
+            node.defaults = []
+
+        call_node = self.parse_expression()
+        if not isinstance(call_node, nodes.Call):
+            self.fail("expected call", node.lineno)
+        node.call = call_node
+        node.body = self.parse_statements(("name:endcall",), drop_needle=True)
+        return node
+
+    def parse_filter_block(self) -> nodes.FilterBlock:
+        node = nodes.FilterBlock(lineno=next(self.stream).lineno)
+        node.filter = self.parse_filter(None, start_inline=True)  # type: ignore
+        node.body = self.parse_statements(("name:endfilter",), drop_needle=True)
+        return node
+
+    def parse_macro(self) -> nodes.Macro:
+        node = nodes.Macro(lineno=next(self.stream).lineno)
+        node.name = self.parse_assign_target(name_only=True).name
+        self.parse_signature(node)
+        node.body = self.parse_statements(("name:endmacro",), drop_needle=True)
+        return node
+
+    def parse_print(self) -> nodes.Output:
+        node = nodes.Output(lineno=next(self.stream).lineno)
+        node.nodes = []
+        while self.stream.current.type != "block_end":
+            if node.nodes:
+                self.stream.expect("comma")
+            node.nodes.append(self.parse_expression())
+        return node
+
+    @typing.overload
+    def parse_assign_target(
+        self, with_tuple: bool = ..., name_only: "te.Literal[True]" = ...
+    ) -> nodes.Name: ...
+
+    @typing.overload
+    def parse_assign_target(
+        self,
+        with_tuple: bool = True,
+        name_only: bool = False,
+        extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
+        with_namespace: bool = False,
+    ) -> t.Union[nodes.NSRef, nodes.Name, nodes.Tuple]: ...
+
+    def parse_assign_target(
+        self,
+        with_tuple: bool = True,
+        name_only: bool = False,
+        extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
+        with_namespace: bool = False,
+    ) -> t.Union[nodes.NSRef, nodes.Name, nodes.Tuple]:
+        """Parse an assignment target.  As Jinja allows assignments to
+        tuples, this function can parse all allowed assignment targets.  Per
+        default assignments to tuples are parsed, that can be disable however
+        by setting `with_tuple` to `False`.  If only assignments to names are
+        wanted `name_only` can be set to `True`.  The `extra_end_rules`
+        parameter is forwarded to the tuple parsing function.  If
+        `with_namespace` is enabled, a namespace assignment may be parsed.
+        """
+        target: nodes.Expr
+
+        if name_only:
+            token = self.stream.expect("name")
+            target = nodes.Name(token.value, "store", lineno=token.lineno)
+        else:
+            if with_tuple:
+                target = self.parse_tuple(
+                    simplified=True,
+                    extra_end_rules=extra_end_rules,
+                    with_namespace=with_namespace,
+                )
+            else:
+                target = self.parse_primary(with_namespace=with_namespace)
+
+            target.set_ctx("store")
+
+        if not target.can_assign():
+            self.fail(
+                f"can't assign to {type(target).__name__.lower()!r}", target.lineno
+            )
+
+        return target  # type: ignore
+
+    def parse_expression(self, with_condexpr: bool = True) -> nodes.Expr:
+        """Parse an expression.  Per default all expressions are parsed, if
+        the optional `with_condexpr` parameter is set to `False` conditional
+        expressions are not parsed.
+        """
+        if with_condexpr:
+            return self.parse_condexpr()
+        return self.parse_or()
+
+    def parse_condexpr(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        expr1 = self.parse_or()
+        expr3: t.Optional[nodes.Expr]
+
+        while self.stream.skip_if("name:if"):
+            expr2 = self.parse_or()
+            if self.stream.skip_if("name:else"):
+                expr3 = self.parse_condexpr()
+            else:
+                expr3 = None
+            expr1 = nodes.CondExpr(expr2, expr1, expr3, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return expr1
+
+    def parse_or(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        left = self.parse_and()
+        while self.stream.skip_if("name:or"):
+            right = self.parse_and()
+            left = nodes.Or(left, right, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return left
+
+    def parse_and(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        left = self.parse_not()
+        while self.stream.skip_if("name:and"):
+            right = self.parse_not()
+            left = nodes.And(left, right, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return left
+
+    def parse_not(self) -> nodes.Expr:
+        if self.stream.current.test("name:not"):
+            lineno = next(self.stream).lineno
+            return nodes.Not(self.parse_not(), lineno=lineno)
+        return self.parse_compare()
+
+    def parse_compare(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        expr = self.parse_math1()
+        ops = []
+        while True:
+            token_type = self.stream.current.type
+            if token_type in _compare_operators:
+                next(self.stream)
+                ops.append(nodes.Operand(token_type, self.parse_math1()))
+            elif self.stream.skip_if("name:in"):
+                ops.append(nodes.Operand("in", self.parse_math1()))
+            elif self.stream.current.test("name:not") and self.stream.look().test(
+                "name:in"
+            ):
+                self.stream.skip(2)
+                ops.append(nodes.Operand("notin", self.parse_math1()))
+            else:
+                break
+            lineno = self.stream.current.lineno
+        if not ops:
+            return expr
+        return nodes.Compare(expr, ops, lineno=lineno)
+
+    def parse_math1(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        left = self.parse_concat()
+        while self.stream.current.type in ("add", "sub"):
+            cls = _math_nodes[self.stream.current.type]
+            next(self.stream)
+            right = self.parse_concat()
+            left = cls(left, right, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return left
+
+    def parse_concat(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        args = [self.parse_math2()]
+        while self.stream.current.type == "tilde":
+            next(self.stream)
+            args.append(self.parse_math2())
+        if len(args) == 1:
+            return args[0]
+        return nodes.Concat(args, lineno=lineno)
+
+    def parse_math2(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        left = self.parse_pow()
+        while self.stream.current.type in ("mul", "div", "floordiv", "mod"):
+            cls = _math_nodes[self.stream.current.type]
+            next(self.stream)
+            right = self.parse_pow()
+            left = cls(left, right, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return left
+
+    def parse_pow(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        left = self.parse_unary()
+        while self.stream.current.type == "pow":
+            next(self.stream)
+            right = self.parse_unary()
+            left = nodes.Pow(left, right, lineno=lineno)
+            lineno = self.stream.current.lineno
+        return left
+
+    def parse_unary(self, with_filter: bool = True) -> nodes.Expr:
+        token_type = self.stream.current.type
+        lineno = self.stream.current.lineno
+        node: nodes.Expr
+
+        if token_type == "sub":
+            next(self.stream)
+            node = nodes.Neg(self.parse_unary(False), lineno=lineno)
+        elif token_type == "add":
+            next(self.stream)
+            node = nodes.Pos(self.parse_unary(False), lineno=lineno)
+        else:
+            node = self.parse_primary()
+        node = self.parse_postfix(node)
+        if with_filter:
+            node = self.parse_filter_expr(node)
+        return node
+
+    def parse_primary(self, with_namespace: bool = False) -> nodes.Expr:
+        """Parse a name or literal value. If ``with_namespace`` is enabled, also
+        parse namespace attr refs, for use in assignments."""
+        token = self.stream.current
+        node: nodes.Expr
+        if token.type == "name":
+            next(self.stream)
+            if token.value in ("true", "false", "True", "False"):
+                node = nodes.Const(token.value in ("true", "True"), lineno=token.lineno)
+            elif token.value in ("none", "None"):
+                node = nodes.Const(None, lineno=token.lineno)
+            elif with_namespace and self.stream.current.type == "dot":
+                # If namespace attributes are allowed at this point, and the next
+                # token is a dot, produce a namespace reference.
+                next(self.stream)
+                attr = self.stream.expect("name")
+                node = nodes.NSRef(token.value, attr.value, lineno=token.lineno)
+            else:
+                node = nodes.Name(token.value, "load", lineno=token.lineno)
+        elif token.type == "string":
+            next(self.stream)
+            buf = [token.value]
+            lineno = token.lineno
+            while self.stream.current.type == "string":
+                buf.append(self.stream.current.value)
+                next(self.stream)
+            node = nodes.Const("".join(buf), lineno=lineno)
+        elif token.type in ("integer", "float"):
+            next(self.stream)
+            node = nodes.Const(token.value, lineno=token.lineno)
+        elif token.type == "lparen":
+            next(self.stream)
+            node = self.parse_tuple(explicit_parentheses=True)
+            self.stream.expect("rparen")
+        elif token.type == "lbracket":
+            node = self.parse_list()
+        elif token.type == "lbrace":
+            node = self.parse_dict()
+        else:
+            self.fail(f"unexpected {describe_token(token)!r}", token.lineno)
+        return node
+
+    def parse_tuple(
+        self,
+        simplified: bool = False,
+        with_condexpr: bool = True,
+        extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
+        explicit_parentheses: bool = False,
+        with_namespace: bool = False,
+    ) -> t.Union[nodes.Tuple, nodes.Expr]:
+        """Works like `parse_expression` but if multiple expressions are
+        delimited by a comma a :class:`~jinja2.nodes.Tuple` node is created.
+        This method could also return a regular expression instead of a tuple
+        if no commas where found.
+
+        The default parsing mode is a full tuple.  If `simplified` is `True`
+        only names and literals are parsed; ``with_namespace`` allows namespace
+        attr refs as well. The `no_condexpr` parameter is forwarded to
+        :meth:`parse_expression`.
+
+        Because tuples do not require delimiters and may end in a bogus comma
+        an extra hint is needed that marks the end of a tuple.  For example
+        for loops support tuples between `for` and `in`.  In that case the
+        `extra_end_rules` is set to ``['name:in']``.
+
+        `explicit_parentheses` is true if the parsing was triggered by an
+        expression in parentheses.  This is used to figure out if an empty
+        tuple is a valid expression or not.
+        """
+        lineno = self.stream.current.lineno
+        if simplified:
+
+            def parse() -> nodes.Expr:
+                return self.parse_primary(with_namespace=with_namespace)
+
+        else:
+
+            def parse() -> nodes.Expr:
+                return self.parse_expression(with_condexpr=with_condexpr)
+
+        args: t.List[nodes.Expr] = []
+        is_tuple = False
+
+        while True:
+            if args:
+                self.stream.expect("comma")
+            if self.is_tuple_end(extra_end_rules):
+                break
+            args.append(parse())
+            if self.stream.current.type == "comma":
+                is_tuple = True
+            else:
+                break
+            lineno = self.stream.current.lineno
+
+        if not is_tuple:
+            if args:
+                return args[0]
+
+            # if we don't have explicit parentheses, an empty tuple is
+            # not a valid expression.  This would mean nothing (literally
+            # nothing) in the spot of an expression would be an empty
+            # tuple.
+            if not explicit_parentheses:
+                self.fail(
+                    "Expected an expression,"
+                    f" got {describe_token(self.stream.current)!r}"
+                )
+
+        return nodes.Tuple(args, "load", lineno=lineno)
+
+    def parse_list(self) -> nodes.List:
+        token = self.stream.expect("lbracket")
+        items: t.List[nodes.Expr] = []
+        while self.stream.current.type != "rbracket":
+            if items:
+                self.stream.expect("comma")
+            if self.stream.current.type == "rbracket":
+                break
+            items.append(self.parse_expression())
+        self.stream.expect("rbracket")
+        return nodes.List(items, lineno=token.lineno)
+
+    def parse_dict(self) -> nodes.Dict:
+        token = self.stream.expect("lbrace")
+        items: t.List[nodes.Pair] = []
+        while self.stream.current.type != "rbrace":
+            if items:
+                self.stream.expect("comma")
+            if self.stream.current.type == "rbrace":
+                break
+            key = self.parse_expression()
+            self.stream.expect("colon")
+            value = self.parse_expression()
+            items.append(nodes.Pair(key, value, lineno=key.lineno))
+        self.stream.expect("rbrace")
+        return nodes.Dict(items, lineno=token.lineno)
+
+    def parse_postfix(self, node: nodes.Expr) -> nodes.Expr:
+        while True:
+            token_type = self.stream.current.type
+            if token_type == "dot" or token_type == "lbracket":
+                node = self.parse_subscript(node)
+            # calls are valid both after postfix expressions (getattr
+            # and getitem) as well as filters and tests
+            elif token_type == "lparen":
+                node = self.parse_call(node)
+            else:
+                break
+        return node
+
+    def parse_filter_expr(self, node: nodes.Expr) -> nodes.Expr:
+        while True:
+            token_type = self.stream.current.type
+            if token_type == "pipe":
+                node = self.parse_filter(node)  # type: ignore
+            elif token_type == "name" and self.stream.current.value == "is":
+                node = self.parse_test(node)
+            # calls are valid both after postfix expressions (getattr
+            # and getitem) as well as filters and tests
+            elif token_type == "lparen":
+                node = self.parse_call(node)
+            else:
+                break
+        return node
+
+    def parse_subscript(
+        self, node: nodes.Expr
+    ) -> t.Union[nodes.Getattr, nodes.Getitem]:
+        token = next(self.stream)
+        arg: nodes.Expr
+
+        if token.type == "dot":
+            attr_token = self.stream.current
+            next(self.stream)
+            if attr_token.type == "name":
+                return nodes.Getattr(
+                    node, attr_token.value, "load", lineno=token.lineno
+                )
+            elif attr_token.type != "integer":
+                self.fail("expected name or number", attr_token.lineno)
+            arg = nodes.Const(attr_token.value, lineno=attr_token.lineno)
+            return nodes.Getitem(node, arg, "load", lineno=token.lineno)
+        if token.type == "lbracket":
+            args: t.List[nodes.Expr] = []
+            while self.stream.current.type != "rbracket":
+                if args:
+                    self.stream.expect("comma")
+                args.append(self.parse_subscribed())
+            self.stream.expect("rbracket")
+            if len(args) == 1:
+                arg = args[0]
+            else:
+                arg = nodes.Tuple(args, "load", lineno=token.lineno)
+            return nodes.Getitem(node, arg, "load", lineno=token.lineno)
+        self.fail("expected subscript expression", token.lineno)
+
+    def parse_subscribed(self) -> nodes.Expr:
+        lineno = self.stream.current.lineno
+        args: t.List[t.Optional[nodes.Expr]]
+
+        if self.stream.current.type == "colon":
+            next(self.stream)
+            args = [None]
+        else:
+            node = self.parse_expression()
+            if self.stream.current.type != "colon":
+                return node
+            next(self.stream)
+            args = [node]
+
+        if self.stream.current.type == "colon":
+            args.append(None)
+        elif self.stream.current.type not in ("rbracket", "comma"):
+            args.append(self.parse_expression())
+        else:
+            args.append(None)
+
+        if self.stream.current.type == "colon":
+            next(self.stream)
+            if self.stream.current.type not in ("rbracket", "comma"):
+                args.append(self.parse_expression())
+            else:
+                args.append(None)
+        else:
+            args.append(None)
+
+        return nodes.Slice(lineno=lineno, *args)  # noqa: B026
+
+    def parse_call_args(
+        self,
+    ) -> t.Tuple[
+        t.List[nodes.Expr],
+        t.List[nodes.Keyword],
+        t.Optional[nodes.Expr],
+        t.Optional[nodes.Expr],
+    ]:
+        token = self.stream.expect("lparen")
+        args = []
+        kwargs = []
+        dyn_args = None
+        dyn_kwargs = None
+        require_comma = False
+
+        def ensure(expr: bool) -> None:
+            if not expr:
+                self.fail("invalid syntax for function call expression", token.lineno)
+
+        while self.stream.current.type != "rparen":
+            if require_comma:
+                self.stream.expect("comma")
+
+                # support for trailing comma
+                if self.stream.current.type == "rparen":
+                    break
+
+            if self.stream.current.type == "mul":
+                ensure(dyn_args is None and dyn_kwargs is None)
+                next(self.stream)
+                dyn_args = self.parse_expression()
+            elif self.stream.current.type == "pow":
+                ensure(dyn_kwargs is None)
+                next(self.stream)
+                dyn_kwargs = self.parse_expression()
+            else:
+                if (
+                    self.stream.current.type == "name"
+                    and self.stream.look().type == "assign"
+                ):
+                    # Parsing a kwarg
+                    ensure(dyn_kwargs is None)
+                    key = self.stream.current.value
+                    self.stream.skip(2)
+                    value = self.parse_expression()
+                    kwargs.append(nodes.Keyword(key, value, lineno=value.lineno))
+                else:
+                    # Parsing an arg
+                    ensure(dyn_args is None and dyn_kwargs is None and not kwargs)
+                    args.append(self.parse_expression())
+
+            require_comma = True
+
+        self.stream.expect("rparen")
+        return args, kwargs, dyn_args, dyn_kwargs
+
+    def parse_call(self, node: nodes.Expr) -> nodes.Call:
+        # The lparen will be expected in parse_call_args, but the lineno
+        # needs to be recorded before the stream is advanced.
+        token = self.stream.current
+        args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
+        return nodes.Call(node, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno)
+
+    def parse_filter(
+        self, node: t.Optional[nodes.Expr], start_inline: bool = False
+    ) -> t.Optional[nodes.Expr]:
+        while self.stream.current.type == "pipe" or start_inline:
+            if not start_inline:
+                next(self.stream)
+            token = self.stream.expect("name")
+            name = token.value
+            while self.stream.current.type == "dot":
+                next(self.stream)
+                name += "." + self.stream.expect("name").value
+            if self.stream.current.type == "lparen":
+                args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
+            else:
+                args = []
+                kwargs = []
+                dyn_args = dyn_kwargs = None
+            node = nodes.Filter(
+                node, name, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno
+            )
+            start_inline = False
+        return node
+
+    def parse_test(self, node: nodes.Expr) -> nodes.Expr:
+        token = next(self.stream)
+        if self.stream.current.test("name:not"):
+            next(self.stream)
+            negated = True
+        else:
+            negated = False
+        name = self.stream.expect("name").value
+        while self.stream.current.type == "dot":
+            next(self.stream)
+            name += "." + self.stream.expect("name").value
+        dyn_args = dyn_kwargs = None
+        kwargs: t.List[nodes.Keyword] = []
+        if self.stream.current.type == "lparen":
+            args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
+        elif self.stream.current.type in {
+            "name",
+            "string",
+            "integer",
+            "float",
+            "lparen",
+            "lbracket",
+            "lbrace",
+        } and not self.stream.current.test_any("name:else", "name:or", "name:and"):
+            if self.stream.current.test("name:is"):
+                self.fail("You cannot chain multiple tests with is")
+            arg_node = self.parse_primary()
+            arg_node = self.parse_postfix(arg_node)
+            args = [arg_node]
+        else:
+            args = []
+        node = nodes.Test(
+            node, name, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno
+        )
+        if negated:
+            node = nodes.Not(node, lineno=token.lineno)
+        return node
+
+    def subparse(
+        self, end_tokens: t.Optional[t.Tuple[str, ...]] = None
+    ) -> t.List[nodes.Node]:
+        body: t.List[nodes.Node] = []
+        data_buffer: t.List[nodes.Node] = []
+        add_data = data_buffer.append
+
+        if end_tokens is not None:
+            self._end_token_stack.append(end_tokens)
+
+        def flush_data() -> None:
+            if data_buffer:
+                lineno = data_buffer[0].lineno
+                body.append(nodes.Output(data_buffer[:], lineno=lineno))
+                del data_buffer[:]
+
+        try:
+            while self.stream:
+                token = self.stream.current
+                if token.type == "data":
+                    if token.value:
+                        add_data(nodes.TemplateData(token.value, lineno=token.lineno))
+                    next(self.stream)
+                elif token.type == "variable_begin":
+                    next(self.stream)
+                    add_data(self.parse_tuple(with_condexpr=True))
+                    self.stream.expect("variable_end")
+                elif token.type == "block_begin":
+                    flush_data()
+                    next(self.stream)
+                    if end_tokens is not None and self.stream.current.test_any(
+                        *end_tokens
+                    ):
+                        return body
+                    rv = self.parse_statement()
+                    if isinstance(rv, list):
+                        body.extend(rv)
+                    else:
+                        body.append(rv)
+                    self.stream.expect("block_end")
+                else:
+                    raise AssertionError("internal parsing error")
+
+            flush_data()
+        finally:
+            if end_tokens is not None:
+                self._end_token_stack.pop()
+        return body
+
+    def parse(self) -> nodes.Template:
+        """Parse the whole template into a `Template` node."""
+        result = nodes.Template(self.subparse(), lineno=1)
+        result.set_environment(self.environment)
+        return result
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/py.typed b/backend/venv/lib/python3.12/site-packages/jinja2/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/runtime.py b/backend/venv/lib/python3.12/site-packages/jinja2/runtime.py
new file mode 100644
index 00000000..09119e2a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/runtime.py
@@ -0,0 +1,1062 @@
+"""The runtime functions and state used by compiled templates."""
+
+import functools
+import sys
+import typing as t
+from collections import abc
+from itertools import chain
+
+from markupsafe import escape  # noqa: F401
+from markupsafe import Markup
+from markupsafe import soft_str
+
+from .async_utils import auto_aiter
+from .async_utils import auto_await  # noqa: F401
+from .exceptions import TemplateNotFound  # noqa: F401
+from .exceptions import TemplateRuntimeError  # noqa: F401
+from .exceptions import UndefinedError
+from .nodes import EvalContext
+from .utils import _PassArg
+from .utils import concat
+from .utils import internalcode
+from .utils import missing
+from .utils import Namespace  # noqa: F401
+from .utils import object_type_repr
+from .utils import pass_eval_context
+
+V = t.TypeVar("V")
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+if t.TYPE_CHECKING:
+    import logging
+
+    import typing_extensions as te
+
+    from .environment import Environment
+
+    class LoopRenderFunc(te.Protocol):
+        def __call__(
+            self,
+            reciter: t.Iterable[V],
+            loop_render_func: "LoopRenderFunc",
+            depth: int = 0,
+        ) -> str: ...
+
+
+# these variables are exported to the template runtime
+exported = [
+    "LoopContext",
+    "TemplateReference",
+    "Macro",
+    "Markup",
+    "TemplateRuntimeError",
+    "missing",
+    "escape",
+    "markup_join",
+    "str_join",
+    "identity",
+    "TemplateNotFound",
+    "Namespace",
+    "Undefined",
+    "internalcode",
+]
+async_exported = [
+    "AsyncLoopContext",
+    "auto_aiter",
+    "auto_await",
+]
+
+
+def identity(x: V) -> V:
+    """Returns its argument. Useful for certain things in the
+    environment.
+    """
+    return x
+
+
+def markup_join(seq: t.Iterable[t.Any]) -> str:
+    """Concatenation that escapes if necessary and converts to string."""
+    buf = []
+    iterator = map(soft_str, seq)
+    for arg in iterator:
+        buf.append(arg)
+        if hasattr(arg, "__html__"):
+            return Markup("").join(chain(buf, iterator))
+    return concat(buf)
+
+
+def str_join(seq: t.Iterable[t.Any]) -> str:
+    """Simple args to string conversion and concatenation."""
+    return concat(map(str, seq))
+
+
+def new_context(
+    environment: "Environment",
+    template_name: t.Optional[str],
+    blocks: t.Dict[str, t.Callable[["Context"], t.Iterator[str]]],
+    vars: t.Optional[t.Dict[str, t.Any]] = None,
+    shared: bool = False,
+    globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    locals: t.Optional[t.Mapping[str, t.Any]] = None,
+) -> "Context":
+    """Internal helper for context creation."""
+    if vars is None:
+        vars = {}
+    if shared:
+        parent = vars
+    else:
+        parent = dict(globals or (), **vars)
+    if locals:
+        # if the parent is shared a copy should be created because
+        # we don't want to modify the dict passed
+        if shared:
+            parent = dict(parent)
+        for key, value in locals.items():
+            if value is not missing:
+                parent[key] = value
+    return environment.context_class(
+        environment, parent, template_name, blocks, globals=globals
+    )
+
+
+class TemplateReference:
+    """The `self` in templates."""
+
+    def __init__(self, context: "Context") -> None:
+        self.__context = context
+
+    def __getitem__(self, name: str) -> t.Any:
+        blocks = self.__context.blocks[name]
+        return BlockReference(name, self.__context, blocks, 0)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.__context.name!r}>"
+
+
+def _dict_method_all(dict_method: F) -> F:
+    @functools.wraps(dict_method)
+    def f_all(self: "Context") -> t.Any:
+        return dict_method(self.get_all())
+
+    return t.cast(F, f_all)
+
+
+@abc.Mapping.register
+class Context:
+    """The template context holds the variables of a template.  It stores the
+    values passed to the template and also the names the template exports.
+    Creating instances is neither supported nor useful as it's created
+    automatically at various stages of the template evaluation and should not
+    be created by hand.
+
+    The context is immutable.  Modifications on :attr:`parent` **must not**
+    happen and modifications on :attr:`vars` are allowed from generated
+    template code only.  Template filters and global functions marked as
+    :func:`pass_context` get the active context passed as first argument
+    and are allowed to access the context read-only.
+
+    The template context supports read only dict operations (`get`,
+    `keys`, `values`, `items`, `iterkeys`, `itervalues`, `iteritems`,
+    `__getitem__`, `__contains__`).  Additionally there is a :meth:`resolve`
+    method that doesn't fail with a `KeyError` but returns an
+    :class:`Undefined` object for missing variables.
+    """
+
+    def __init__(
+        self,
+        environment: "Environment",
+        parent: t.Dict[str, t.Any],
+        name: t.Optional[str],
+        blocks: t.Dict[str, t.Callable[["Context"], t.Iterator[str]]],
+        globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
+    ):
+        self.parent = parent
+        self.vars: t.Dict[str, t.Any] = {}
+        self.environment: Environment = environment
+        self.eval_ctx = EvalContext(self.environment, name)
+        self.exported_vars: t.Set[str] = set()
+        self.name = name
+        self.globals_keys = set() if globals is None else set(globals)
+
+        # create the initial mapping of blocks.  Whenever template inheritance
+        # takes place the runtime will update this mapping with the new blocks
+        # from the template.
+        self.blocks = {k: [v] for k, v in blocks.items()}
+
+    def super(
+        self, name: str, current: t.Callable[["Context"], t.Iterator[str]]
+    ) -> t.Union["BlockReference", "Undefined"]:
+        """Render a parent block."""
+        try:
+            blocks = self.blocks[name]
+            index = blocks.index(current) + 1
+            blocks[index]
+        except LookupError:
+            return self.environment.undefined(
+                f"there is no parent block called {name!r}.", name="super"
+            )
+        return BlockReference(name, self, blocks, index)
+
+    def get(self, key: str, default: t.Any = None) -> t.Any:
+        """Look up a variable by name, or return a default if the key is
+        not found.
+
+        :param key: The variable name to look up.
+        :param default: The value to return if the key is not found.
+        """
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def resolve(self, key: str) -> t.Union[t.Any, "Undefined"]:
+        """Look up a variable by name, or return an :class:`Undefined`
+        object if the key is not found.
+
+        If you need to add custom behavior, override
+        :meth:`resolve_or_missing`, not this method. The various lookup
+        functions use that method, not this one.
+
+        :param key: The variable name to look up.
+        """
+        rv = self.resolve_or_missing(key)
+
+        if rv is missing:
+            return self.environment.undefined(name=key)
+
+        return rv
+
+    def resolve_or_missing(self, key: str) -> t.Any:
+        """Look up a variable by name, or return a ``missing`` sentinel
+        if the key is not found.
+
+        Override this method to add custom lookup behavior.
+        :meth:`resolve`, :meth:`get`, and :meth:`__getitem__` use this
+        method. Don't call this method directly.
+
+        :param key: The variable name to look up.
+        """
+        if key in self.vars:
+            return self.vars[key]
+
+        if key in self.parent:
+            return self.parent[key]
+
+        return missing
+
+    def get_exported(self) -> t.Dict[str, t.Any]:
+        """Get a new dict with the exported variables."""
+        return {k: self.vars[k] for k in self.exported_vars}
+
+    def get_all(self) -> t.Dict[str, t.Any]:
+        """Return the complete context as dict including the exported
+        variables.  For optimizations reasons this might not return an
+        actual copy so be careful with using it.
+        """
+        if not self.vars:
+            return self.parent
+        if not self.parent:
+            return self.vars
+        return dict(self.parent, **self.vars)
+
+    @internalcode
+    def call(
+        __self,
+        __obj: t.Callable[..., t.Any],
+        *args: t.Any,
+        **kwargs: t.Any,  # noqa: B902
+    ) -> t.Union[t.Any, "Undefined"]:
+        """Call the callable with the arguments and keyword arguments
+        provided but inject the active context or environment as first
+        argument if the callable has :func:`pass_context` or
+        :func:`pass_environment`.
+        """
+        if __debug__:
+            __traceback_hide__ = True  # noqa
+
+        # Allow callable classes to take a context
+        if (
+            hasattr(__obj, "__call__")  # noqa: B004
+            and _PassArg.from_obj(__obj.__call__) is not None
+        ):
+            __obj = __obj.__call__
+
+        pass_arg = _PassArg.from_obj(__obj)
+
+        if pass_arg is _PassArg.context:
+            # the active context should have access to variables set in
+            # loops and blocks without mutating the context itself
+            if kwargs.get("_loop_vars"):
+                __self = __self.derived(kwargs["_loop_vars"])
+            if kwargs.get("_block_vars"):
+                __self = __self.derived(kwargs["_block_vars"])
+            args = (__self,) + args
+        elif pass_arg is _PassArg.eval_context:
+            args = (__self.eval_ctx,) + args
+        elif pass_arg is _PassArg.environment:
+            args = (__self.environment,) + args
+
+        kwargs.pop("_block_vars", None)
+        kwargs.pop("_loop_vars", None)
+
+        try:
+            return __obj(*args, **kwargs)
+        except StopIteration:
+            return __self.environment.undefined(
+                "value was undefined because a callable raised a"
+                " StopIteration exception"
+            )
+
+    def derived(self, locals: t.Optional[t.Dict[str, t.Any]] = None) -> "Context":
+        """Internal helper function to create a derived context.  This is
+        used in situations where the system needs a new context in the same
+        template that is independent.
+        """
+        context = new_context(
+            self.environment, self.name, {}, self.get_all(), True, None, locals
+        )
+        context.eval_ctx = self.eval_ctx
+        context.blocks.update((k, list(v)) for k, v in self.blocks.items())
+        return context
+
+    keys = _dict_method_all(dict.keys)
+    values = _dict_method_all(dict.values)
+    items = _dict_method_all(dict.items)
+
+    def __contains__(self, name: str) -> bool:
+        return name in self.vars or name in self.parent
+
+    def __getitem__(self, key: str) -> t.Any:
+        """Look up a variable by name with ``[]`` syntax, or raise a
+        ``KeyError`` if the key is not found.
+        """
+        item = self.resolve_or_missing(key)
+
+        if item is missing:
+            raise KeyError(key)
+
+        return item
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.get_all()!r} of {self.name!r}>"
+
+
+class BlockReference:
+    """One block on a template reference."""
+
+    def __init__(
+        self,
+        name: str,
+        context: "Context",
+        stack: t.List[t.Callable[["Context"], t.Iterator[str]]],
+        depth: int,
+    ) -> None:
+        self.name = name
+        self._context = context
+        self._stack = stack
+        self._depth = depth
+
+    @property
+    def super(self) -> t.Union["BlockReference", "Undefined"]:
+        """Super the block."""
+        if self._depth + 1 >= len(self._stack):
+            return self._context.environment.undefined(
+                f"there is no parent block called {self.name!r}.", name="super"
+            )
+        return BlockReference(self.name, self._context, self._stack, self._depth + 1)
+
+    @internalcode
+    async def _async_call(self) -> str:
+        rv = self._context.environment.concat(  # type: ignore
+            [x async for x in self._stack[self._depth](self._context)]  # type: ignore
+        )
+
+        if self._context.eval_ctx.autoescape:
+            return Markup(rv)
+
+        return rv
+
+    @internalcode
+    def __call__(self) -> str:
+        if self._context.environment.is_async:
+            return self._async_call()  # type: ignore
+
+        rv = self._context.environment.concat(  # type: ignore
+            self._stack[self._depth](self._context)
+        )
+
+        if self._context.eval_ctx.autoescape:
+            return Markup(rv)
+
+        return rv
+
+
+class LoopContext:
+    """A wrapper iterable for dynamic ``for`` loops, with information
+    about the loop and iteration.
+    """
+
+    #: Current iteration of the loop, starting at 0.
+    index0 = -1
+
+    _length: t.Optional[int] = None
+    _after: t.Any = missing
+    _current: t.Any = missing
+    _before: t.Any = missing
+    _last_changed_value: t.Any = missing
+
+    def __init__(
+        self,
+        iterable: t.Iterable[V],
+        undefined: t.Type["Undefined"],
+        recurse: t.Optional["LoopRenderFunc"] = None,
+        depth0: int = 0,
+    ) -> None:
+        """
+        :param iterable: Iterable to wrap.
+        :param undefined: :class:`Undefined` class to use for next and
+            previous items.
+        :param recurse: The function to render the loop body when the
+            loop is marked recursive.
+        :param depth0: Incremented when looping recursively.
+        """
+        self._iterable = iterable
+        self._iterator = self._to_iterator(iterable)
+        self._undefined = undefined
+        self._recurse = recurse
+        #: How many levels deep a recursive loop currently is, starting at 0.
+        self.depth0 = depth0
+
+    @staticmethod
+    def _to_iterator(iterable: t.Iterable[V]) -> t.Iterator[V]:
+        return iter(iterable)
+
+    @property
+    def length(self) -> int:
+        """Length of the iterable.
+
+        If the iterable is a generator or otherwise does not have a
+        size, it is eagerly evaluated to get a size.
+        """
+        if self._length is not None:
+            return self._length
+
+        try:
+            self._length = len(self._iterable)  # type: ignore
+        except TypeError:
+            iterable = list(self._iterator)
+            self._iterator = self._to_iterator(iterable)
+            self._length = len(iterable) + self.index + (self._after is not missing)
+
+        return self._length
+
+    def __len__(self) -> int:
+        return self.length
+
+    @property
+    def depth(self) -> int:
+        """How many levels deep a recursive loop currently is, starting at 1."""
+        return self.depth0 + 1
+
+    @property
+    def index(self) -> int:
+        """Current iteration of the loop, starting at 1."""
+        return self.index0 + 1
+
+    @property
+    def revindex0(self) -> int:
+        """Number of iterations from the end of the loop, ending at 0.
+
+        Requires calculating :attr:`length`.
+        """
+        return self.length - self.index
+
+    @property
+    def revindex(self) -> int:
+        """Number of iterations from the end of the loop, ending at 1.
+
+        Requires calculating :attr:`length`.
+        """
+        return self.length - self.index0
+
+    @property
+    def first(self) -> bool:
+        """Whether this is the first iteration of the loop."""
+        return self.index0 == 0
+
+    def _peek_next(self) -> t.Any:
+        """Return the next element in the iterable, or :data:`missing`
+        if the iterable is exhausted. Only peeks one item ahead, caching
+        the result in :attr:`_last` for use in subsequent checks. The
+        cache is reset when :meth:`__next__` is called.
+        """
+        if self._after is not missing:
+            return self._after
+
+        self._after = next(self._iterator, missing)
+        return self._after
+
+    @property
+    def last(self) -> bool:
+        """Whether this is the last iteration of the loop.
+
+        Causes the iterable to advance early. See
+        :func:`itertools.groupby` for issues this can cause.
+        The :func:`groupby` filter avoids that issue.
+        """
+        return self._peek_next() is missing
+
+    @property
+    def previtem(self) -> t.Union[t.Any, "Undefined"]:
+        """The item in the previous iteration. Undefined during the
+        first iteration.
+        """
+        if self.first:
+            return self._undefined("there is no previous item")
+
+        return self._before
+
+    @property
+    def nextitem(self) -> t.Union[t.Any, "Undefined"]:
+        """The item in the next iteration. Undefined during the last
+        iteration.
+
+        Causes the iterable to advance early. See
+        :func:`itertools.groupby` for issues this can cause.
+        The :func:`jinja-filters.groupby` filter avoids that issue.
+        """
+        rv = self._peek_next()
+
+        if rv is missing:
+            return self._undefined("there is no next item")
+
+        return rv
+
+    def cycle(self, *args: V) -> V:
+        """Return a value from the given args, cycling through based on
+        the current :attr:`index0`.
+
+        :param args: One or more values to cycle through.
+        """
+        if not args:
+            raise TypeError("no items for cycling given")
+
+        return args[self.index0 % len(args)]
+
+    def changed(self, *value: t.Any) -> bool:
+        """Return ``True`` if previously called with a different value
+        (including when called for the first time).
+
+        :param value: One or more values to compare to the last call.
+        """
+        if self._last_changed_value != value:
+            self._last_changed_value = value
+            return True
+
+        return False
+
+    def __iter__(self) -> "LoopContext":
+        return self
+
+    def __next__(self) -> t.Tuple[t.Any, "LoopContext"]:
+        if self._after is not missing:
+            rv = self._after
+            self._after = missing
+        else:
+            rv = next(self._iterator)
+
+        self.index0 += 1
+        self._before = self._current
+        self._current = rv
+        return rv, self
+
+    @internalcode
+    def __call__(self, iterable: t.Iterable[V]) -> str:
+        """When iterating over nested data, render the body of the loop
+        recursively with the given inner iterable data.
+
+        The loop must have the ``recursive`` marker for this to work.
+        """
+        if self._recurse is None:
+            raise TypeError(
+                "The loop must have the 'recursive' marker to be called recursively."
+            )
+
+        return self._recurse(iterable, self._recurse, depth=self.depth)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.index}/{self.length}>"
+
+
+class AsyncLoopContext(LoopContext):
+    _iterator: t.AsyncIterator[t.Any]  # type: ignore
+
+    @staticmethod
+    def _to_iterator(  # type: ignore
+        iterable: t.Union[t.Iterable[V], t.AsyncIterable[V]],
+    ) -> t.AsyncIterator[V]:
+        return auto_aiter(iterable)
+
+    @property
+    async def length(self) -> int:  # type: ignore
+        if self._length is not None:
+            return self._length
+
+        try:
+            self._length = len(self._iterable)  # type: ignore
+        except TypeError:
+            iterable = [x async for x in self._iterator]
+            self._iterator = self._to_iterator(iterable)
+            self._length = len(iterable) + self.index + (self._after is not missing)
+
+        return self._length
+
+    @property
+    async def revindex0(self) -> int:  # type: ignore
+        return await self.length - self.index
+
+    @property
+    async def revindex(self) -> int:  # type: ignore
+        return await self.length - self.index0
+
+    async def _peek_next(self) -> t.Any:
+        if self._after is not missing:
+            return self._after
+
+        try:
+            self._after = await self._iterator.__anext__()
+        except StopAsyncIteration:
+            self._after = missing
+
+        return self._after
+
+    @property
+    async def last(self) -> bool:  # type: ignore
+        return await self._peek_next() is missing
+
+    @property
+    async def nextitem(self) -> t.Union[t.Any, "Undefined"]:
+        rv = await self._peek_next()
+
+        if rv is missing:
+            return self._undefined("there is no next item")
+
+        return rv
+
+    def __aiter__(self) -> "AsyncLoopContext":
+        return self
+
+    async def __anext__(self) -> t.Tuple[t.Any, "AsyncLoopContext"]:
+        if self._after is not missing:
+            rv = self._after
+            self._after = missing
+        else:
+            rv = await self._iterator.__anext__()
+
+        self.index0 += 1
+        self._before = self._current
+        self._current = rv
+        return rv, self
+
+
+class Macro:
+    """Wraps a macro function."""
+
+    def __init__(
+        self,
+        environment: "Environment",
+        func: t.Callable[..., str],
+        name: str,
+        arguments: t.List[str],
+        catch_kwargs: bool,
+        catch_varargs: bool,
+        caller: bool,
+        default_autoescape: t.Optional[bool] = None,
+    ):
+        self._environment = environment
+        self._func = func
+        self._argument_count = len(arguments)
+        self.name = name
+        self.arguments = arguments
+        self.catch_kwargs = catch_kwargs
+        self.catch_varargs = catch_varargs
+        self.caller = caller
+        self.explicit_caller = "caller" in arguments
+
+        if default_autoescape is None:
+            if callable(environment.autoescape):
+                default_autoescape = environment.autoescape(None)
+            else:
+                default_autoescape = environment.autoescape
+
+        self._default_autoescape = default_autoescape
+
+    @internalcode
+    @pass_eval_context
+    def __call__(self, *args: t.Any, **kwargs: t.Any) -> str:
+        # This requires a bit of explanation,  In the past we used to
+        # decide largely based on compile-time information if a macro is
+        # safe or unsafe.  While there was a volatile mode it was largely
+        # unused for deciding on escaping.  This turns out to be
+        # problematic for macros because whether a macro is safe depends not
+        # on the escape mode when it was defined, but rather when it was used.
+        #
+        # Because however we export macros from the module system and
+        # there are historic callers that do not pass an eval context (and
+        # will continue to not pass one), we need to perform an instance
+        # check here.
+        #
+        # This is considered safe because an eval context is not a valid
+        # argument to callables otherwise anyway.  Worst case here is
+        # that if no eval context is passed we fall back to the compile
+        # time autoescape flag.
+        if args and isinstance(args[0], EvalContext):
+            autoescape = args[0].autoescape
+            args = args[1:]
+        else:
+            autoescape = self._default_autoescape
+
+        # try to consume the positional arguments
+        arguments = list(args[: self._argument_count])
+        off = len(arguments)
+
+        # For information why this is necessary refer to the handling
+        # of caller in the `macro_body` handler in the compiler.
+        found_caller = False
+
+        # if the number of arguments consumed is not the number of
+        # arguments expected we start filling in keyword arguments
+        # and defaults.
+        if off != self._argument_count:
+            for name in self.arguments[len(arguments) :]:
+                try:
+                    value = kwargs.pop(name)
+                except KeyError:
+                    value = missing
+                if name == "caller":
+                    found_caller = True
+                arguments.append(value)
+        else:
+            found_caller = self.explicit_caller
+
+        # it's important that the order of these arguments does not change
+        # if not also changed in the compiler's `function_scoping` method.
+        # the order is caller, keyword arguments, positional arguments!
+        if self.caller and not found_caller:
+            caller = kwargs.pop("caller", None)
+            if caller is None:
+                caller = self._environment.undefined("No caller defined", name="caller")
+            arguments.append(caller)
+
+        if self.catch_kwargs:
+            arguments.append(kwargs)
+        elif kwargs:
+            if "caller" in kwargs:
+                raise TypeError(
+                    f"macro {self.name!r} was invoked with two values for the special"
+                    " caller argument. This is most likely a bug."
+                )
+            raise TypeError(
+                f"macro {self.name!r} takes no keyword argument {next(iter(kwargs))!r}"
+            )
+        if self.catch_varargs:
+            arguments.append(args[self._argument_count :])
+        elif len(args) > self._argument_count:
+            raise TypeError(
+                f"macro {self.name!r} takes not more than"
+                f" {len(self.arguments)} argument(s)"
+            )
+
+        return self._invoke(arguments, autoescape)
+
+    async def _async_invoke(self, arguments: t.List[t.Any], autoescape: bool) -> str:
+        rv = await self._func(*arguments)  # type: ignore
+
+        if autoescape:
+            return Markup(rv)
+
+        return rv  # type: ignore
+
+    def _invoke(self, arguments: t.List[t.Any], autoescape: bool) -> str:
+        if self._environment.is_async:
+            return self._async_invoke(arguments, autoescape)  # type: ignore
+
+        rv = self._func(*arguments)
+
+        if autoescape:
+            rv = Markup(rv)
+
+        return rv
+
+    def __repr__(self) -> str:
+        name = "anonymous" if self.name is None else repr(self.name)
+        return f"<{type(self).__name__} {name}>"
+
+
+class Undefined:
+    """The default undefined type. This can be printed, iterated, and treated as
+    a boolean. Any other operation will raise an :exc:`UndefinedError`.
+
+    >>> foo = Undefined(name='foo')
+    >>> str(foo)
+    ''
+    >>> not foo
+    True
+    >>> foo + 42
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+    """
+
+    __slots__ = (
+        "_undefined_hint",
+        "_undefined_obj",
+        "_undefined_name",
+        "_undefined_exception",
+    )
+
+    def __init__(
+        self,
+        hint: t.Optional[str] = None,
+        obj: t.Any = missing,
+        name: t.Optional[str] = None,
+        exc: t.Type[TemplateRuntimeError] = UndefinedError,
+    ) -> None:
+        self._undefined_hint = hint
+        self._undefined_obj = obj
+        self._undefined_name = name
+        self._undefined_exception = exc
+
+    @property
+    def _undefined_message(self) -> str:
+        """Build a message about the undefined value based on how it was
+        accessed.
+        """
+        if self._undefined_hint:
+            return self._undefined_hint
+
+        if self._undefined_obj is missing:
+            return f"{self._undefined_name!r} is undefined"
+
+        if not isinstance(self._undefined_name, str):
+            return (
+                f"{object_type_repr(self._undefined_obj)} has no"
+                f" element {self._undefined_name!r}"
+            )
+
+        return (
+            f"{object_type_repr(self._undefined_obj)!r} has no"
+            f" attribute {self._undefined_name!r}"
+        )
+
+    @internalcode
+    def _fail_with_undefined_error(
+        self, *args: t.Any, **kwargs: t.Any
+    ) -> "te.NoReturn":
+        """Raise an :exc:`UndefinedError` when operations are performed
+        on the undefined value.
+        """
+        raise self._undefined_exception(self._undefined_message)
+
+    @internalcode
+    def __getattr__(self, name: str) -> t.Any:
+        # Raise AttributeError on requests for names that appear to be unimplemented
+        # dunder methods to keep Python's internal protocol probing behaviors working
+        # properly in cases where another exception type could cause unexpected or
+        # difficult-to-diagnose failures.
+        if name[:2] == "__" and name[-2:] == "__":
+            raise AttributeError(name)
+
+        return self._fail_with_undefined_error()
+
+    __add__ = __radd__ = __sub__ = __rsub__ = _fail_with_undefined_error
+    __mul__ = __rmul__ = __div__ = __rdiv__ = _fail_with_undefined_error
+    __truediv__ = __rtruediv__ = _fail_with_undefined_error
+    __floordiv__ = __rfloordiv__ = _fail_with_undefined_error
+    __mod__ = __rmod__ = _fail_with_undefined_error
+    __pos__ = __neg__ = _fail_with_undefined_error
+    __call__ = __getitem__ = _fail_with_undefined_error
+    __lt__ = __le__ = __gt__ = __ge__ = _fail_with_undefined_error
+    __int__ = __float__ = __complex__ = _fail_with_undefined_error
+    __pow__ = __rpow__ = _fail_with_undefined_error
+
+    def __eq__(self, other: t.Any) -> bool:
+        return type(self) is type(other)
+
+    def __ne__(self, other: t.Any) -> bool:
+        return not self.__eq__(other)
+
+    def __hash__(self) -> int:
+        return id(type(self))
+
+    def __str__(self) -> str:
+        return ""
+
+    def __len__(self) -> int:
+        return 0
+
+    def __iter__(self) -> t.Iterator[t.Any]:
+        yield from ()
+
+    async def __aiter__(self) -> t.AsyncIterator[t.Any]:
+        for _ in ():
+            yield
+
+    def __bool__(self) -> bool:
+        return False
+
+    def __repr__(self) -> str:
+        return "Undefined"
+
+
+def make_logging_undefined(
+    logger: t.Optional["logging.Logger"] = None, base: t.Type[Undefined] = Undefined
+) -> t.Type[Undefined]:
+    """Given a logger object this returns a new undefined class that will
+    log certain failures.  It will log iterations and printing.  If no
+    logger is given a default logger is created.
+
+    Example::
+
+        logger = logging.getLogger(__name__)
+        LoggingUndefined = make_logging_undefined(
+            logger=logger,
+            base=Undefined
+        )
+
+    .. versionadded:: 2.8
+
+    :param logger: the logger to use.  If not provided, a default logger
+                   is created.
+    :param base: the base class to add logging functionality to.  This
+                 defaults to :class:`Undefined`.
+    """
+    if logger is None:
+        import logging
+
+        logger = logging.getLogger(__name__)
+        logger.addHandler(logging.StreamHandler(sys.stderr))
+
+    def _log_message(undef: Undefined) -> None:
+        logger.warning("Template variable warning: %s", undef._undefined_message)
+
+    class LoggingUndefined(base):  # type: ignore
+        __slots__ = ()
+
+        def _fail_with_undefined_error(  # type: ignore
+            self, *args: t.Any, **kwargs: t.Any
+        ) -> "te.NoReturn":
+            try:
+                super()._fail_with_undefined_error(*args, **kwargs)
+            except self._undefined_exception as e:
+                logger.error("Template variable error: %s", e)  # type: ignore
+                raise e
+
+        def __str__(self) -> str:
+            _log_message(self)
+            return super().__str__()  # type: ignore
+
+        def __iter__(self) -> t.Iterator[t.Any]:
+            _log_message(self)
+            return super().__iter__()  # type: ignore
+
+        def __bool__(self) -> bool:
+            _log_message(self)
+            return super().__bool__()  # type: ignore
+
+    return LoggingUndefined
+
+
+class ChainableUndefined(Undefined):
+    """An undefined that is chainable, where both ``__getattr__`` and
+    ``__getitem__`` return itself rather than raising an
+    :exc:`UndefinedError`.
+
+    >>> foo = ChainableUndefined(name='foo')
+    >>> str(foo.bar['baz'])
+    ''
+    >>> foo.bar['baz'] + 42
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+
+    .. versionadded:: 2.11.0
+    """
+
+    __slots__ = ()
+
+    def __html__(self) -> str:
+        return str(self)
+
+    def __getattr__(self, name: str) -> "ChainableUndefined":
+        # Raise AttributeError on requests for names that appear to be unimplemented
+        # dunder methods to avoid confusing Python with truthy non-method objects that
+        # do not implement the protocol being probed for. e.g., copy.copy(Undefined())
+        # fails spectacularly if getattr(Undefined(), '__setstate__') returns an
+        # Undefined object instead of raising AttributeError to signal that it does not
+        # support that style of object initialization.
+        if name[:2] == "__" and name[-2:] == "__":
+            raise AttributeError(name)
+
+        return self
+
+    def __getitem__(self, _name: str) -> "ChainableUndefined":  # type: ignore[override]
+        return self
+
+
+class DebugUndefined(Undefined):
+    """An undefined that returns the debug info when printed.
+
+    >>> foo = DebugUndefined(name='foo')
+    >>> str(foo)
+    '{{ foo }}'
+    >>> not foo
+    True
+    >>> foo + 42
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+    """
+
+    __slots__ = ()
+
+    def __str__(self) -> str:
+        if self._undefined_hint:
+            message = f"undefined value printed: {self._undefined_hint}"
+
+        elif self._undefined_obj is missing:
+            message = self._undefined_name  # type: ignore
+
+        else:
+            message = (
+                f"no such element: {object_type_repr(self._undefined_obj)}"
+                f"[{self._undefined_name!r}]"
+            )
+
+        return f"{{{{ {message} }}}}"
+
+
+class StrictUndefined(Undefined):
+    """An undefined that barks on print and iteration as well as boolean
+    tests and all kinds of comparisons.  In other words: you can do nothing
+    with it except checking if it's defined using the `defined` test.
+
+    >>> foo = StrictUndefined(name='foo')
+    >>> str(foo)
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+    >>> not foo
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+    >>> foo + 42
+    Traceback (most recent call last):
+      ...
+    jinja2.exceptions.UndefinedError: 'foo' is undefined
+    """
+
+    __slots__ = ()
+    __iter__ = __str__ = __len__ = Undefined._fail_with_undefined_error
+    __eq__ = __ne__ = __bool__ = __hash__ = Undefined._fail_with_undefined_error
+    __contains__ = Undefined._fail_with_undefined_error
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/sandbox.py b/backend/venv/lib/python3.12/site-packages/jinja2/sandbox.py
new file mode 100644
index 00000000..9c9dae22
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/sandbox.py
@@ -0,0 +1,436 @@
+"""A sandbox layer that ensures unsafe operations cannot be performed.
+Useful when the template itself comes from an untrusted source.
+"""
+
+import operator
+import types
+import typing as t
+from _string import formatter_field_name_split  # type: ignore
+from collections import abc
+from collections import deque
+from functools import update_wrapper
+from string import Formatter
+
+from markupsafe import EscapeFormatter
+from markupsafe import Markup
+
+from .environment import Environment
+from .exceptions import SecurityError
+from .runtime import Context
+from .runtime import Undefined
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+#: maximum number of items a range may produce
+MAX_RANGE = 100000
+
+#: Unsafe function attributes.
+UNSAFE_FUNCTION_ATTRIBUTES: t.Set[str] = set()
+
+#: Unsafe method attributes. Function attributes are unsafe for methods too.
+UNSAFE_METHOD_ATTRIBUTES: t.Set[str] = set()
+
+#: unsafe generator attributes.
+UNSAFE_GENERATOR_ATTRIBUTES = {"gi_frame", "gi_code"}
+
+#: unsafe attributes on coroutines
+UNSAFE_COROUTINE_ATTRIBUTES = {"cr_frame", "cr_code"}
+
+#: unsafe attributes on async generators
+UNSAFE_ASYNC_GENERATOR_ATTRIBUTES = {"ag_code", "ag_frame"}
+
+_mutable_spec: t.Tuple[t.Tuple[t.Type[t.Any], t.FrozenSet[str]], ...] = (
+    (
+        abc.MutableSet,
+        frozenset(
+            [
+                "add",
+                "clear",
+                "difference_update",
+                "discard",
+                "pop",
+                "remove",
+                "symmetric_difference_update",
+                "update",
+            ]
+        ),
+    ),
+    (
+        abc.MutableMapping,
+        frozenset(["clear", "pop", "popitem", "setdefault", "update"]),
+    ),
+    (
+        abc.MutableSequence,
+        frozenset(
+            ["append", "clear", "pop", "reverse", "insert", "sort", "extend", "remove"]
+        ),
+    ),
+    (
+        deque,
+        frozenset(
+            [
+                "append",
+                "appendleft",
+                "clear",
+                "extend",
+                "extendleft",
+                "pop",
+                "popleft",
+                "remove",
+                "rotate",
+            ]
+        ),
+    ),
+)
+
+
+def safe_range(*args: int) -> range:
+    """A range that can't generate ranges with a length of more than
+    MAX_RANGE items.
+    """
+    rng = range(*args)
+
+    if len(rng) > MAX_RANGE:
+        raise OverflowError(
+            "Range too big. The sandbox blocks ranges larger than"
+            f" MAX_RANGE ({MAX_RANGE})."
+        )
+
+    return rng
+
+
+def unsafe(f: F) -> F:
+    """Marks a function or method as unsafe.
+
+    .. code-block: python
+
+        @unsafe
+        def delete(self):
+            pass
+    """
+    f.unsafe_callable = True  # type: ignore
+    return f
+
+
+def is_internal_attribute(obj: t.Any, attr: str) -> bool:
+    """Test if the attribute given is an internal python attribute.  For
+    example this function returns `True` for the `func_code` attribute of
+    python objects.  This is useful if the environment method
+    :meth:`~SandboxedEnvironment.is_safe_attribute` is overridden.
+
+    >>> from jinja2.sandbox import is_internal_attribute
+    >>> is_internal_attribute(str, "mro")
+    True
+    >>> is_internal_attribute(str, "upper")
+    False
+    """
+    if isinstance(obj, types.FunctionType):
+        if attr in UNSAFE_FUNCTION_ATTRIBUTES:
+            return True
+    elif isinstance(obj, types.MethodType):
+        if attr in UNSAFE_FUNCTION_ATTRIBUTES or attr in UNSAFE_METHOD_ATTRIBUTES:
+            return True
+    elif isinstance(obj, type):
+        if attr == "mro":
+            return True
+    elif isinstance(obj, (types.CodeType, types.TracebackType, types.FrameType)):
+        return True
+    elif isinstance(obj, types.GeneratorType):
+        if attr in UNSAFE_GENERATOR_ATTRIBUTES:
+            return True
+    elif hasattr(types, "CoroutineType") and isinstance(obj, types.CoroutineType):
+        if attr in UNSAFE_COROUTINE_ATTRIBUTES:
+            return True
+    elif hasattr(types, "AsyncGeneratorType") and isinstance(
+        obj, types.AsyncGeneratorType
+    ):
+        if attr in UNSAFE_ASYNC_GENERATOR_ATTRIBUTES:
+            return True
+    return attr.startswith("__")
+
+
+def modifies_known_mutable(obj: t.Any, attr: str) -> bool:
+    """This function checks if an attribute on a builtin mutable object
+    (list, dict, set or deque) or the corresponding ABCs would modify it
+    if called.
+
+    >>> modifies_known_mutable({}, "clear")
+    True
+    >>> modifies_known_mutable({}, "keys")
+    False
+    >>> modifies_known_mutable([], "append")
+    True
+    >>> modifies_known_mutable([], "index")
+    False
+
+    If called with an unsupported object, ``False`` is returned.
+
+    >>> modifies_known_mutable("foo", "upper")
+    False
+    """
+    for typespec, unsafe in _mutable_spec:
+        if isinstance(obj, typespec):
+            return attr in unsafe
+    return False
+
+
+class SandboxedEnvironment(Environment):
+    """The sandboxed environment.  It works like the regular environment but
+    tells the compiler to generate sandboxed code.  Additionally subclasses of
+    this environment may override the methods that tell the runtime what
+    attributes or functions are safe to access.
+
+    If the template tries to access insecure code a :exc:`SecurityError` is
+    raised.  However also other exceptions may occur during the rendering so
+    the caller has to ensure that all exceptions are caught.
+    """
+
+    sandboxed = True
+
+    #: default callback table for the binary operators.  A copy of this is
+    #: available on each instance of a sandboxed environment as
+    #: :attr:`binop_table`
+    default_binop_table: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
+        "+": operator.add,
+        "-": operator.sub,
+        "*": operator.mul,
+        "/": operator.truediv,
+        "//": operator.floordiv,
+        "**": operator.pow,
+        "%": operator.mod,
+    }
+
+    #: default callback table for the unary operators.  A copy of this is
+    #: available on each instance of a sandboxed environment as
+    #: :attr:`unop_table`
+    default_unop_table: t.Dict[str, t.Callable[[t.Any], t.Any]] = {
+        "+": operator.pos,
+        "-": operator.neg,
+    }
+
+    #: a set of binary operators that should be intercepted.  Each operator
+    #: that is added to this set (empty by default) is delegated to the
+    #: :meth:`call_binop` method that will perform the operator.  The default
+    #: operator callback is specified by :attr:`binop_table`.
+    #:
+    #: The following binary operators are interceptable:
+    #: ``//``, ``%``, ``+``, ``*``, ``-``, ``/``, and ``**``
+    #:
+    #: The default operation form the operator table corresponds to the
+    #: builtin function.  Intercepted calls are always slower than the native
+    #: operator call, so make sure only to intercept the ones you are
+    #: interested in.
+    #:
+    #: .. versionadded:: 2.6
+    intercepted_binops: t.FrozenSet[str] = frozenset()
+
+    #: a set of unary operators that should be intercepted.  Each operator
+    #: that is added to this set (empty by default) is delegated to the
+    #: :meth:`call_unop` method that will perform the operator.  The default
+    #: operator callback is specified by :attr:`unop_table`.
+    #:
+    #: The following unary operators are interceptable: ``+``, ``-``
+    #:
+    #: The default operation form the operator table corresponds to the
+    #: builtin function.  Intercepted calls are always slower than the native
+    #: operator call, so make sure only to intercept the ones you are
+    #: interested in.
+    #:
+    #: .. versionadded:: 2.6
+    intercepted_unops: t.FrozenSet[str] = frozenset()
+
+    def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:
+        super().__init__(*args, **kwargs)
+        self.globals["range"] = safe_range
+        self.binop_table = self.default_binop_table.copy()
+        self.unop_table = self.default_unop_table.copy()
+
+    def is_safe_attribute(self, obj: t.Any, attr: str, value: t.Any) -> bool:
+        """The sandboxed environment will call this method to check if the
+        attribute of an object is safe to access.  Per default all attributes
+        starting with an underscore are considered private as well as the
+        special attributes of internal python objects as returned by the
+        :func:`is_internal_attribute` function.
+        """
+        return not (attr.startswith("_") or is_internal_attribute(obj, attr))
+
+    def is_safe_callable(self, obj: t.Any) -> bool:
+        """Check if an object is safely callable. By default callables
+        are considered safe unless decorated with :func:`unsafe`.
+
+        This also recognizes the Django convention of setting
+        ``func.alters_data = True``.
+        """
+        return not (
+            getattr(obj, "unsafe_callable", False) or getattr(obj, "alters_data", False)
+        )
+
+    def call_binop(
+        self, context: Context, operator: str, left: t.Any, right: t.Any
+    ) -> t.Any:
+        """For intercepted binary operator calls (:meth:`intercepted_binops`)
+        this function is executed instead of the builtin operator.  This can
+        be used to fine tune the behavior of certain operators.
+
+        .. versionadded:: 2.6
+        """
+        return self.binop_table[operator](left, right)
+
+    def call_unop(self, context: Context, operator: str, arg: t.Any) -> t.Any:
+        """For intercepted unary operator calls (:meth:`intercepted_unops`)
+        this function is executed instead of the builtin operator.  This can
+        be used to fine tune the behavior of certain operators.
+
+        .. versionadded:: 2.6
+        """
+        return self.unop_table[operator](arg)
+
+    def getitem(
+        self, obj: t.Any, argument: t.Union[str, t.Any]
+    ) -> t.Union[t.Any, Undefined]:
+        """Subscribe an object from sandboxed code."""
+        try:
+            return obj[argument]
+        except (TypeError, LookupError):
+            if isinstance(argument, str):
+                try:
+                    attr = str(argument)
+                except Exception:
+                    pass
+                else:
+                    try:
+                        value = getattr(obj, attr)
+                    except AttributeError:
+                        pass
+                    else:
+                        fmt = self.wrap_str_format(value)
+                        if fmt is not None:
+                            return fmt
+                        if self.is_safe_attribute(obj, argument, value):
+                            return value
+                        return self.unsafe_undefined(obj, argument)
+        return self.undefined(obj=obj, name=argument)
+
+    def getattr(self, obj: t.Any, attribute: str) -> t.Union[t.Any, Undefined]:
+        """Subscribe an object from sandboxed code and prefer the
+        attribute.  The attribute passed *must* be a bytestring.
+        """
+        try:
+            value = getattr(obj, attribute)
+        except AttributeError:
+            try:
+                return obj[attribute]
+            except (TypeError, LookupError):
+                pass
+        else:
+            fmt = self.wrap_str_format(value)
+            if fmt is not None:
+                return fmt
+            if self.is_safe_attribute(obj, attribute, value):
+                return value
+            return self.unsafe_undefined(obj, attribute)
+        return self.undefined(obj=obj, name=attribute)
+
+    def unsafe_undefined(self, obj: t.Any, attribute: str) -> Undefined:
+        """Return an undefined object for unsafe attributes."""
+        return self.undefined(
+            f"access to attribute {attribute!r} of"
+            f" {type(obj).__name__!r} object is unsafe.",
+            name=attribute,
+            obj=obj,
+            exc=SecurityError,
+        )
+
+    def wrap_str_format(self, value: t.Any) -> t.Optional[t.Callable[..., str]]:
+        """If the given value is a ``str.format`` or ``str.format_map`` method,
+        return a new function than handles sandboxing. This is done at access
+        rather than in :meth:`call`, so that calls made without ``call`` are
+        also sandboxed.
+        """
+        if not isinstance(
+            value, (types.MethodType, types.BuiltinMethodType)
+        ) or value.__name__ not in ("format", "format_map"):
+            return None
+
+        f_self: t.Any = value.__self__
+
+        if not isinstance(f_self, str):
+            return None
+
+        str_type: t.Type[str] = type(f_self)
+        is_format_map = value.__name__ == "format_map"
+        formatter: SandboxedFormatter
+
+        if isinstance(f_self, Markup):
+            formatter = SandboxedEscapeFormatter(self, escape=f_self.escape)
+        else:
+            formatter = SandboxedFormatter(self)
+
+        vformat = formatter.vformat
+
+        def wrapper(*args: t.Any, **kwargs: t.Any) -> str:
+            if is_format_map:
+                if kwargs:
+                    raise TypeError("format_map() takes no keyword arguments")
+
+                if len(args) != 1:
+                    raise TypeError(
+                        f"format_map() takes exactly one argument ({len(args)} given)"
+                    )
+
+                kwargs = args[0]
+                args = ()
+
+            return str_type(vformat(f_self, args, kwargs))
+
+        return update_wrapper(wrapper, value)
+
+    def call(
+        __self,  # noqa: B902
+        __context: Context,
+        __obj: t.Any,
+        *args: t.Any,
+        **kwargs: t.Any,
+    ) -> t.Any:
+        """Call an object from sandboxed code."""
+
+        # the double prefixes are to avoid double keyword argument
+        # errors when proxying the call.
+        if not __self.is_safe_callable(__obj):
+            raise SecurityError(f"{__obj!r} is not safely callable")
+        return __context.call(__obj, *args, **kwargs)
+
+
+class ImmutableSandboxedEnvironment(SandboxedEnvironment):
+    """Works exactly like the regular `SandboxedEnvironment` but does not
+    permit modifications on the builtin mutable objects `list`, `set`, and
+    `dict` by using the :func:`modifies_known_mutable` function.
+    """
+
+    def is_safe_attribute(self, obj: t.Any, attr: str, value: t.Any) -> bool:
+        if not super().is_safe_attribute(obj, attr, value):
+            return False
+
+        return not modifies_known_mutable(obj, attr)
+
+
+class SandboxedFormatter(Formatter):
+    def __init__(self, env: Environment, **kwargs: t.Any) -> None:
+        self._env = env
+        super().__init__(**kwargs)
+
+    def get_field(
+        self, field_name: str, args: t.Sequence[t.Any], kwargs: t.Mapping[str, t.Any]
+    ) -> t.Tuple[t.Any, str]:
+        first, rest = formatter_field_name_split(field_name)
+        obj = self.get_value(first, args, kwargs)
+        for is_attr, i in rest:
+            if is_attr:
+                obj = self._env.getattr(obj, i)
+            else:
+                obj = self._env.getitem(obj, i)
+        return obj, first
+
+
+class SandboxedEscapeFormatter(SandboxedFormatter, EscapeFormatter):
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/tests.py b/backend/venv/lib/python3.12/site-packages/jinja2/tests.py
new file mode 100644
index 00000000..1a59e370
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/tests.py
@@ -0,0 +1,256 @@
+"""Built-in template tests used with the ``is`` operator."""
+
+import operator
+import typing as t
+from collections import abc
+from numbers import Number
+
+from .runtime import Undefined
+from .utils import pass_environment
+
+if t.TYPE_CHECKING:
+    from .environment import Environment
+
+
+def test_odd(value: int) -> bool:
+    """Return true if the variable is odd."""
+    return value % 2 == 1
+
+
+def test_even(value: int) -> bool:
+    """Return true if the variable is even."""
+    return value % 2 == 0
+
+
+def test_divisibleby(value: int, num: int) -> bool:
+    """Check if a variable is divisible by a number."""
+    return value % num == 0
+
+
+def test_defined(value: t.Any) -> bool:
+    """Return true if the variable is defined:
+
+    .. sourcecode:: jinja
+
+        {% if variable is defined %}
+            value of variable: {{ variable }}
+        {% else %}
+            variable is not defined
+        {% endif %}
+
+    See the :func:`default` filter for a simple way to set undefined
+    variables.
+    """
+    return not isinstance(value, Undefined)
+
+
+def test_undefined(value: t.Any) -> bool:
+    """Like :func:`defined` but the other way round."""
+    return isinstance(value, Undefined)
+
+
+@pass_environment
+def test_filter(env: "Environment", value: str) -> bool:
+    """Check if a filter exists by name. Useful if a filter may be
+    optionally available.
+
+    .. code-block:: jinja
+
+        {% if 'markdown' is filter %}
+            {{ value | markdown }}
+        {% else %}
+            {{ value }}
+        {% endif %}
+
+    .. versionadded:: 3.0
+    """
+    return value in env.filters
+
+
+@pass_environment
+def test_test(env: "Environment", value: str) -> bool:
+    """Check if a test exists by name. Useful if a test may be
+    optionally available.
+
+    .. code-block:: jinja
+
+        {% if 'loud' is test %}
+            {% if value is loud %}
+                {{ value|upper }}
+            {% else %}
+                {{ value|lower }}
+            {% endif %}
+        {% else %}
+            {{ value }}
+        {% endif %}
+
+    .. versionadded:: 3.0
+    """
+    return value in env.tests
+
+
+def test_none(value: t.Any) -> bool:
+    """Return true if the variable is none."""
+    return value is None
+
+
+def test_boolean(value: t.Any) -> bool:
+    """Return true if the object is a boolean value.
+
+    .. versionadded:: 2.11
+    """
+    return value is True or value is False
+
+
+def test_false(value: t.Any) -> bool:
+    """Return true if the object is False.
+
+    .. versionadded:: 2.11
+    """
+    return value is False
+
+
+def test_true(value: t.Any) -> bool:
+    """Return true if the object is True.
+
+    .. versionadded:: 2.11
+    """
+    return value is True
+
+
+# NOTE: The existing 'number' test matches booleans and floats
+def test_integer(value: t.Any) -> bool:
+    """Return true if the object is an integer.
+
+    .. versionadded:: 2.11
+    """
+    return isinstance(value, int) and value is not True and value is not False
+
+
+# NOTE: The existing 'number' test matches booleans and integers
+def test_float(value: t.Any) -> bool:
+    """Return true if the object is a float.
+
+    .. versionadded:: 2.11
+    """
+    return isinstance(value, float)
+
+
+def test_lower(value: str) -> bool:
+    """Return true if the variable is lowercased."""
+    return str(value).islower()
+
+
+def test_upper(value: str) -> bool:
+    """Return true if the variable is uppercased."""
+    return str(value).isupper()
+
+
+def test_string(value: t.Any) -> bool:
+    """Return true if the object is a string."""
+    return isinstance(value, str)
+
+
+def test_mapping(value: t.Any) -> bool:
+    """Return true if the object is a mapping (dict etc.).
+
+    .. versionadded:: 2.6
+    """
+    return isinstance(value, abc.Mapping)
+
+
+def test_number(value: t.Any) -> bool:
+    """Return true if the variable is a number."""
+    return isinstance(value, Number)
+
+
+def test_sequence(value: t.Any) -> bool:
+    """Return true if the variable is a sequence. Sequences are variables
+    that are iterable.
+    """
+    try:
+        len(value)
+        value.__getitem__  # noqa B018
+    except Exception:
+        return False
+
+    return True
+
+
+def test_sameas(value: t.Any, other: t.Any) -> bool:
+    """Check if an object points to the same memory address than another
+    object:
+
+    .. sourcecode:: jinja
+
+        {% if foo.attribute is sameas false %}
+            the foo attribute really is the `False` singleton
+        {% endif %}
+    """
+    return value is other
+
+
+def test_iterable(value: t.Any) -> bool:
+    """Check if it's possible to iterate over an object."""
+    try:
+        iter(value)
+    except TypeError:
+        return False
+
+    return True
+
+
+def test_escaped(value: t.Any) -> bool:
+    """Check if the value is escaped."""
+    return hasattr(value, "__html__")
+
+
+def test_in(value: t.Any, seq: t.Container[t.Any]) -> bool:
+    """Check if value is in seq.
+
+    .. versionadded:: 2.10
+    """
+    return value in seq
+
+
+TESTS = {
+    "odd": test_odd,
+    "even": test_even,
+    "divisibleby": test_divisibleby,
+    "defined": test_defined,
+    "undefined": test_undefined,
+    "filter": test_filter,
+    "test": test_test,
+    "none": test_none,
+    "boolean": test_boolean,
+    "false": test_false,
+    "true": test_true,
+    "integer": test_integer,
+    "float": test_float,
+    "lower": test_lower,
+    "upper": test_upper,
+    "string": test_string,
+    "mapping": test_mapping,
+    "number": test_number,
+    "sequence": test_sequence,
+    "iterable": test_iterable,
+    "callable": callable,
+    "sameas": test_sameas,
+    "escaped": test_escaped,
+    "in": test_in,
+    "==": operator.eq,
+    "eq": operator.eq,
+    "equalto": operator.eq,
+    "!=": operator.ne,
+    "ne": operator.ne,
+    ">": operator.gt,
+    "gt": operator.gt,
+    "greaterthan": operator.gt,
+    "ge": operator.ge,
+    ">=": operator.ge,
+    "<": operator.lt,
+    "lt": operator.lt,
+    "lessthan": operator.lt,
+    "<=": operator.le,
+    "le": operator.le,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/utils.py b/backend/venv/lib/python3.12/site-packages/jinja2/utils.py
new file mode 100644
index 00000000..7c922629
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/utils.py
@@ -0,0 +1,766 @@
+import enum
+import json
+import os
+import re
+import typing as t
+from collections import abc
+from collections import deque
+from random import choice
+from random import randrange
+from threading import Lock
+from types import CodeType
+from urllib.parse import quote_from_bytes
+
+import markupsafe
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+
+class _MissingType:
+    def __repr__(self) -> str:
+        return "missing"
+
+    def __reduce__(self) -> str:
+        return "missing"
+
+
+missing: t.Any = _MissingType()
+"""Special singleton representing missing values for the runtime."""
+
+internal_code: t.MutableSet[CodeType] = set()
+
+concat = "".join
+
+
+def pass_context(f: F) -> F:
+    """Pass the :class:`~jinja2.runtime.Context` as the first argument
+    to the decorated function when called while rendering a template.
+
+    Can be used on functions, filters, and tests.
+
+    If only ``Context.eval_context`` is needed, use
+    :func:`pass_eval_context`. If only ``Context.environment`` is
+    needed, use :func:`pass_environment`.
+
+    .. versionadded:: 3.0.0
+        Replaces ``contextfunction`` and ``contextfilter``.
+    """
+    f.jinja_pass_arg = _PassArg.context  # type: ignore
+    return f
+
+
+def pass_eval_context(f: F) -> F:
+    """Pass the :class:`~jinja2.nodes.EvalContext` as the first argument
+    to the decorated function when called while rendering a template.
+    See :ref:`eval-context`.
+
+    Can be used on functions, filters, and tests.
+
+    If only ``EvalContext.environment`` is needed, use
+    :func:`pass_environment`.
+
+    .. versionadded:: 3.0.0
+        Replaces ``evalcontextfunction`` and ``evalcontextfilter``.
+    """
+    f.jinja_pass_arg = _PassArg.eval_context  # type: ignore
+    return f
+
+
+def pass_environment(f: F) -> F:
+    """Pass the :class:`~jinja2.Environment` as the first argument to
+    the decorated function when called while rendering a template.
+
+    Can be used on functions, filters, and tests.
+
+    .. versionadded:: 3.0.0
+        Replaces ``environmentfunction`` and ``environmentfilter``.
+    """
+    f.jinja_pass_arg = _PassArg.environment  # type: ignore
+    return f
+
+
+class _PassArg(enum.Enum):
+    context = enum.auto()
+    eval_context = enum.auto()
+    environment = enum.auto()
+
+    @classmethod
+    def from_obj(cls, obj: F) -> t.Optional["_PassArg"]:
+        if hasattr(obj, "jinja_pass_arg"):
+            return obj.jinja_pass_arg  # type: ignore
+
+        return None
+
+
+def internalcode(f: F) -> F:
+    """Marks the function as internally used"""
+    internal_code.add(f.__code__)
+    return f
+
+
+def is_undefined(obj: t.Any) -> bool:
+    """Check if the object passed is undefined.  This does nothing more than
+    performing an instance check against :class:`Undefined` but looks nicer.
+    This can be used for custom filters or tests that want to react to
+    undefined variables.  For example a custom default filter can look like
+    this::
+
+        def default(var, default=''):
+            if is_undefined(var):
+                return default
+            return var
+    """
+    from .runtime import Undefined
+
+    return isinstance(obj, Undefined)
+
+
+def consume(iterable: t.Iterable[t.Any]) -> None:
+    """Consumes an iterable without doing anything with it."""
+    for _ in iterable:
+        pass
+
+
+def clear_caches() -> None:
+    """Jinja keeps internal caches for environments and lexers.  These are
+    used so that Jinja doesn't have to recreate environments and lexers all
+    the time.  Normally you don't have to care about that but if you are
+    measuring memory consumption you may want to clean the caches.
+    """
+    from .environment import get_spontaneous_environment
+    from .lexer import _lexer_cache
+
+    get_spontaneous_environment.cache_clear()
+    _lexer_cache.clear()
+
+
+def import_string(import_name: str, silent: bool = False) -> t.Any:
+    """Imports an object based on a string.  This is useful if you want to
+    use import paths as endpoints or something similar.  An import path can
+    be specified either in dotted notation (``xml.sax.saxutils.escape``)
+    or with a colon as object delimiter (``xml.sax.saxutils:escape``).
+
+    If the `silent` is True the return value will be `None` if the import
+    fails.
+
+    :return: imported object
+    """
+    try:
+        if ":" in import_name:
+            module, obj = import_name.split(":", 1)
+        elif "." in import_name:
+            module, _, obj = import_name.rpartition(".")
+        else:
+            return __import__(import_name)
+        return getattr(__import__(module, None, None, [obj]), obj)
+    except (ImportError, AttributeError):
+        if not silent:
+            raise
+
+
+def open_if_exists(filename: str, mode: str = "rb") -> t.Optional[t.IO[t.Any]]:
+    """Returns a file descriptor for the filename if that file exists,
+    otherwise ``None``.
+    """
+    if not os.path.isfile(filename):
+        return None
+
+    return open(filename, mode)
+
+
+def object_type_repr(obj: t.Any) -> str:
+    """Returns the name of the object's type.  For some recognized
+    singletons the name of the object is returned instead. (For
+    example for `None` and `Ellipsis`).
+    """
+    if obj is None:
+        return "None"
+    elif obj is Ellipsis:
+        return "Ellipsis"
+
+    cls = type(obj)
+
+    if cls.__module__ == "builtins":
+        return f"{cls.__name__} object"
+
+    return f"{cls.__module__}.{cls.__name__} object"
+
+
+def pformat(obj: t.Any) -> str:
+    """Format an object using :func:`pprint.pformat`."""
+    from pprint import pformat
+
+    return pformat(obj)
+
+
+_http_re = re.compile(
+    r"""
+    ^
+    (
+        (https?://|www\.)  # scheme or www
+        (([\w%-]+\.)+)?  # subdomain
+        (
+            [a-z]{2,63}  # basic tld
+        |
+            xn--[\w%]{2,59}  # idna tld
+        )
+    |
+        ([\w%-]{2,63}\.)+  # basic domain
+        (com|net|int|edu|gov|org|info|mil)  # basic tld
+    |
+        (https?://)  # scheme
+        (
+            (([\d]{1,3})(\.[\d]{1,3}){3})  # IPv4
+        |
+            (\[([\da-f]{0,4}:){2}([\da-f]{0,4}:?){1,6}])  # IPv6
+        )
+    )
+    (?::[\d]{1,5})?  # port
+    (?:[/?#]\S*)?  # path, query, and fragment
+    $
+    """,
+    re.IGNORECASE | re.VERBOSE,
+)
+_email_re = re.compile(r"^\S+@\w[\w.-]*\.\w+$")
+
+
+def urlize(
+    text: str,
+    trim_url_limit: t.Optional[int] = None,
+    rel: t.Optional[str] = None,
+    target: t.Optional[str] = None,
+    extra_schemes: t.Optional[t.Iterable[str]] = None,
+) -> str:
+    """Convert URLs in text into clickable links.
+
+    This may not recognize links in some situations. Usually, a more
+    comprehensive formatter, such as a Markdown library, is a better
+    choice.
+
+    Works on ``http://``, ``https://``, ``www.``, ``mailto:``, and email
+    addresses. Links with trailing punctuation (periods, commas, closing
+    parentheses) and leading punctuation (opening parentheses) are
+    recognized excluding the punctuation. Email addresses that include
+    header fields are not recognized (for example,
+    ``mailto:address@example.com?cc=copy@example.com``).
+
+    :param text: Original text containing URLs to link.
+    :param trim_url_limit: Shorten displayed URL values to this length.
+    :param target: Add the ``target`` attribute to links.
+    :param rel: Add the ``rel`` attribute to links.
+    :param extra_schemes: Recognize URLs that start with these schemes
+        in addition to the default behavior.
+
+    .. versionchanged:: 3.0
+        The ``extra_schemes`` parameter was added.
+
+    .. versionchanged:: 3.0
+        Generate ``https://`` links for URLs without a scheme.
+
+    .. versionchanged:: 3.0
+        The parsing rules were updated. Recognize email addresses with
+        or without the ``mailto:`` scheme. Validate IP addresses. Ignore
+        parentheses and brackets in more cases.
+    """
+    if trim_url_limit is not None:
+
+        def trim_url(x: str) -> str:
+            if len(x) > trim_url_limit:
+                return f"{x[:trim_url_limit]}..."
+
+            return x
+
+    else:
+
+        def trim_url(x: str) -> str:
+            return x
+
+    words = re.split(r"(\s+)", str(markupsafe.escape(text)))
+    rel_attr = f' rel="{markupsafe.escape(rel)}"' if rel else ""
+    target_attr = f' target="{markupsafe.escape(target)}"' if target else ""
+
+    for i, word in enumerate(words):
+        head, middle, tail = "", word, ""
+        match = re.match(r"^([(<]|&lt;)+", middle)
+
+        if match:
+            head = match.group()
+            middle = middle[match.end() :]
+
+        # Unlike lead, which is anchored to the start of the string,
+        # need to check that the string ends with any of the characters
+        # before trying to match all of them, to avoid backtracking.
+        if middle.endswith((")", ">", ".", ",", "\n", "&gt;")):
+            match = re.search(r"([)>.,\n]|&gt;)+$", middle)
+
+            if match:
+                tail = match.group()
+                middle = middle[: match.start()]
+
+        # Prefer balancing parentheses in URLs instead of ignoring a
+        # trailing character.
+        for start_char, end_char in ("(", ")"), ("<", ">"), ("&lt;", "&gt;"):
+            start_count = middle.count(start_char)
+
+            if start_count <= middle.count(end_char):
+                # Balanced, or lighter on the left
+                continue
+
+            # Move as many as possible from the tail to balance
+            for _ in range(min(start_count, tail.count(end_char))):
+                end_index = tail.index(end_char) + len(end_char)
+                # Move anything in the tail before the end char too
+                middle += tail[:end_index]
+                tail = tail[end_index:]
+
+        if _http_re.match(middle):
+            if middle.startswith("https://") or middle.startswith("http://"):
+                middle = (
+                    f'<a href="{middle}"{rel_attr}{target_attr}>{trim_url(middle)}</a>'
+                )
+            else:
+                middle = (
+                    f'<a href="https://{middle}"{rel_attr}{target_attr}>'
+                    f"{trim_url(middle)}</a>"
+                )
+
+        elif middle.startswith("mailto:") and _email_re.match(middle[7:]):
+            middle = f'<a href="{middle}">{middle[7:]}</a>'
+
+        elif (
+            "@" in middle
+            and not middle.startswith("www.")
+            # ignore values like `@a@b`
+            and not middle.startswith("@")
+            and ":" not in middle
+            and _email_re.match(middle)
+        ):
+            middle = f'<a href="mailto:{middle}">{middle}</a>'
+
+        elif extra_schemes is not None:
+            for scheme in extra_schemes:
+                if middle != scheme and middle.startswith(scheme):
+                    middle = f'<a href="{middle}"{rel_attr}{target_attr}>{middle}</a>'
+
+        words[i] = f"{head}{middle}{tail}"
+
+    return "".join(words)
+
+
+def generate_lorem_ipsum(
+    n: int = 5, html: bool = True, min: int = 20, max: int = 100
+) -> str:
+    """Generate some lorem ipsum for the template."""
+    from .constants import LOREM_IPSUM_WORDS
+
+    words = LOREM_IPSUM_WORDS.split()
+    result = []
+
+    for _ in range(n):
+        next_capitalized = True
+        last_comma = last_fullstop = 0
+        word = None
+        last = None
+        p = []
+
+        # each paragraph contains out of 20 to 100 words.
+        for idx, _ in enumerate(range(randrange(min, max))):
+            while True:
+                word = choice(words)
+                if word != last:
+                    last = word
+                    break
+            if next_capitalized:
+                word = word.capitalize()
+                next_capitalized = False
+            # add commas
+            if idx - randrange(3, 8) > last_comma:
+                last_comma = idx
+                last_fullstop += 2
+                word += ","
+            # add end of sentences
+            if idx - randrange(10, 20) > last_fullstop:
+                last_comma = last_fullstop = idx
+                word += "."
+                next_capitalized = True
+            p.append(word)
+
+        # ensure that the paragraph ends with a dot.
+        p_str = " ".join(p)
+
+        if p_str.endswith(","):
+            p_str = p_str[:-1] + "."
+        elif not p_str.endswith("."):
+            p_str += "."
+
+        result.append(p_str)
+
+    if not html:
+        return "\n\n".join(result)
+    return markupsafe.Markup(
+        "\n".join(f"<p>{markupsafe.escape(x)}</p>" for x in result)
+    )
+
+
+def url_quote(obj: t.Any, charset: str = "utf-8", for_qs: bool = False) -> str:
+    """Quote a string for use in a URL using the given charset.
+
+    :param obj: String or bytes to quote. Other types are converted to
+        string then encoded to bytes using the given charset.
+    :param charset: Encode text to bytes using this charset.
+    :param for_qs: Quote "/" and use "+" for spaces.
+    """
+    if not isinstance(obj, bytes):
+        if not isinstance(obj, str):
+            obj = str(obj)
+
+        obj = obj.encode(charset)
+
+    safe = b"" if for_qs else b"/"
+    rv = quote_from_bytes(obj, safe)
+
+    if for_qs:
+        rv = rv.replace("%20", "+")
+
+    return rv
+
+
+@abc.MutableMapping.register
+class LRUCache:
+    """A simple LRU Cache implementation."""
+
+    # this is fast for small capacities (something below 1000) but doesn't
+    # scale.  But as long as it's only used as storage for templates this
+    # won't do any harm.
+
+    def __init__(self, capacity: int) -> None:
+        self.capacity = capacity
+        self._mapping: t.Dict[t.Any, t.Any] = {}
+        self._queue: te.Deque[t.Any] = deque()
+        self._postinit()
+
+    def _postinit(self) -> None:
+        # alias all queue methods for faster lookup
+        self._popleft = self._queue.popleft
+        self._pop = self._queue.pop
+        self._remove = self._queue.remove
+        self._wlock = Lock()
+        self._append = self._queue.append
+
+    def __getstate__(self) -> t.Mapping[str, t.Any]:
+        return {
+            "capacity": self.capacity,
+            "_mapping": self._mapping,
+            "_queue": self._queue,
+        }
+
+    def __setstate__(self, d: t.Mapping[str, t.Any]) -> None:
+        self.__dict__.update(d)
+        self._postinit()
+
+    def __getnewargs__(self) -> t.Tuple[t.Any, ...]:
+        return (self.capacity,)
+
+    def copy(self) -> "te.Self":
+        """Return a shallow copy of the instance."""
+        rv = self.__class__(self.capacity)
+        rv._mapping.update(self._mapping)
+        rv._queue.extend(self._queue)
+        return rv
+
+    def get(self, key: t.Any, default: t.Any = None) -> t.Any:
+        """Return an item from the cache dict or `default`"""
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def setdefault(self, key: t.Any, default: t.Any = None) -> t.Any:
+        """Set `default` if the key is not in the cache otherwise
+        leave unchanged. Return the value of this key.
+        """
+        try:
+            return self[key]
+        except KeyError:
+            self[key] = default
+            return default
+
+    def clear(self) -> None:
+        """Clear the cache."""
+        with self._wlock:
+            self._mapping.clear()
+            self._queue.clear()
+
+    def __contains__(self, key: t.Any) -> bool:
+        """Check if a key exists in this cache."""
+        return key in self._mapping
+
+    def __len__(self) -> int:
+        """Return the current size of the cache."""
+        return len(self._mapping)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self._mapping!r}>"
+
+    def __getitem__(self, key: t.Any) -> t.Any:
+        """Get an item from the cache. Moves the item up so that it has the
+        highest priority then.
+
+        Raise a `KeyError` if it does not exist.
+        """
+        with self._wlock:
+            rv = self._mapping[key]
+
+            if self._queue[-1] != key:
+                try:
+                    self._remove(key)
+                except ValueError:
+                    # if something removed the key from the container
+                    # when we read, ignore the ValueError that we would
+                    # get otherwise.
+                    pass
+
+                self._append(key)
+
+            return rv
+
+    def __setitem__(self, key: t.Any, value: t.Any) -> None:
+        """Sets the value for an item. Moves the item up so that it
+        has the highest priority then.
+        """
+        with self._wlock:
+            if key in self._mapping:
+                self._remove(key)
+            elif len(self._mapping) == self.capacity:
+                del self._mapping[self._popleft()]
+
+            self._append(key)
+            self._mapping[key] = value
+
+    def __delitem__(self, key: t.Any) -> None:
+        """Remove an item from the cache dict.
+        Raise a `KeyError` if it does not exist.
+        """
+        with self._wlock:
+            del self._mapping[key]
+
+            try:
+                self._remove(key)
+            except ValueError:
+                pass
+
+    def items(self) -> t.Iterable[t.Tuple[t.Any, t.Any]]:
+        """Return a list of items."""
+        result = [(key, self._mapping[key]) for key in list(self._queue)]
+        result.reverse()
+        return result
+
+    def values(self) -> t.Iterable[t.Any]:
+        """Return a list of all values."""
+        return [x[1] for x in self.items()]
+
+    def keys(self) -> t.Iterable[t.Any]:
+        """Return a list of all keys ordered by most recent usage."""
+        return list(self)
+
+    def __iter__(self) -> t.Iterator[t.Any]:
+        return reversed(tuple(self._queue))
+
+    def __reversed__(self) -> t.Iterator[t.Any]:
+        """Iterate over the keys in the cache dict, oldest items
+        coming first.
+        """
+        return iter(tuple(self._queue))
+
+    __copy__ = copy
+
+
+def select_autoescape(
+    enabled_extensions: t.Collection[str] = ("html", "htm", "xml"),
+    disabled_extensions: t.Collection[str] = (),
+    default_for_string: bool = True,
+    default: bool = False,
+) -> t.Callable[[t.Optional[str]], bool]:
+    """Intelligently sets the initial value of autoescaping based on the
+    filename of the template.  This is the recommended way to configure
+    autoescaping if you do not want to write a custom function yourself.
+
+    If you want to enable it for all templates created from strings or
+    for all templates with `.html` and `.xml` extensions::
+
+        from jinja2 import Environment, select_autoescape
+        env = Environment(autoescape=select_autoescape(
+            enabled_extensions=('html', 'xml'),
+            default_for_string=True,
+        ))
+
+    Example configuration to turn it on at all times except if the template
+    ends with `.txt`::
+
+        from jinja2 import Environment, select_autoescape
+        env = Environment(autoescape=select_autoescape(
+            disabled_extensions=('txt',),
+            default_for_string=True,
+            default=True,
+        ))
+
+    The `enabled_extensions` is an iterable of all the extensions that
+    autoescaping should be enabled for.  Likewise `disabled_extensions` is
+    a list of all templates it should be disabled for.  If a template is
+    loaded from a string then the default from `default_for_string` is used.
+    If nothing matches then the initial value of autoescaping is set to the
+    value of `default`.
+
+    For security reasons this function operates case insensitive.
+
+    .. versionadded:: 2.9
+    """
+    enabled_patterns = tuple(f".{x.lstrip('.').lower()}" for x in enabled_extensions)
+    disabled_patterns = tuple(f".{x.lstrip('.').lower()}" for x in disabled_extensions)
+
+    def autoescape(template_name: t.Optional[str]) -> bool:
+        if template_name is None:
+            return default_for_string
+        template_name = template_name.lower()
+        if template_name.endswith(enabled_patterns):
+            return True
+        if template_name.endswith(disabled_patterns):
+            return False
+        return default
+
+    return autoescape
+
+
+def htmlsafe_json_dumps(
+    obj: t.Any, dumps: t.Optional[t.Callable[..., str]] = None, **kwargs: t.Any
+) -> markupsafe.Markup:
+    """Serialize an object to a string of JSON with :func:`json.dumps`,
+    then replace HTML-unsafe characters with Unicode escapes and mark
+    the result safe with :class:`~markupsafe.Markup`.
+
+    This is available in templates as the ``|tojson`` filter.
+
+    The following characters are escaped: ``<``, ``>``, ``&``, ``'``.
+
+    The returned string is safe to render in HTML documents and
+    ``<script>`` tags. The exception is in HTML attributes that are
+    double quoted; either use single quotes or the ``|forceescape``
+    filter.
+
+    :param obj: The object to serialize to JSON.
+    :param dumps: The ``dumps`` function to use. Defaults to
+        ``env.policies["json.dumps_function"]``, which defaults to
+        :func:`json.dumps`.
+    :param kwargs: Extra arguments to pass to ``dumps``. Merged onto
+        ``env.policies["json.dumps_kwargs"]``.
+
+    .. versionchanged:: 3.0
+        The ``dumper`` parameter is renamed to ``dumps``.
+
+    .. versionadded:: 2.9
+    """
+    if dumps is None:
+        dumps = json.dumps
+
+    return markupsafe.Markup(
+        dumps(obj, **kwargs)
+        .replace("<", "\\u003c")
+        .replace(">", "\\u003e")
+        .replace("&", "\\u0026")
+        .replace("'", "\\u0027")
+    )
+
+
+class Cycler:
+    """Cycle through values by yield them one at a time, then restarting
+    once the end is reached. Available as ``cycler`` in templates.
+
+    Similar to ``loop.cycle``, but can be used outside loops or across
+    multiple loops. For example, render a list of folders and files in a
+    list, alternating giving them "odd" and "even" classes.
+
+    .. code-block:: html+jinja
+
+        {% set row_class = cycler("odd", "even") %}
+        <ul class="browser">
+        {% for folder in folders %}
+          <li class="folder {{ row_class.next() }}">{{ folder }}
+        {% endfor %}
+        {% for file in files %}
+          <li class="file {{ row_class.next() }}">{{ file }}
+        {% endfor %}
+        </ul>
+
+    :param items: Each positional argument will be yielded in the order
+        given for each cycle.
+
+    .. versionadded:: 2.1
+    """
+
+    def __init__(self, *items: t.Any) -> None:
+        if not items:
+            raise RuntimeError("at least one item has to be provided")
+        self.items = items
+        self.pos = 0
+
+    def reset(self) -> None:
+        """Resets the current item to the first item."""
+        self.pos = 0
+
+    @property
+    def current(self) -> t.Any:
+        """Return the current item. Equivalent to the item that will be
+        returned next time :meth:`next` is called.
+        """
+        return self.items[self.pos]
+
+    def next(self) -> t.Any:
+        """Return the current item, then advance :attr:`current` to the
+        next item.
+        """
+        rv = self.current
+        self.pos = (self.pos + 1) % len(self.items)
+        return rv
+
+    __next__ = next
+
+
+class Joiner:
+    """A joining helper for templates."""
+
+    def __init__(self, sep: str = ", ") -> None:
+        self.sep = sep
+        self.used = False
+
+    def __call__(self) -> str:
+        if not self.used:
+            self.used = True
+            return ""
+        return self.sep
+
+
+class Namespace:
+    """A namespace object that can hold arbitrary attributes.  It may be
+    initialized from a dictionary or with keyword arguments."""
+
+    def __init__(*args: t.Any, **kwargs: t.Any) -> None:  # noqa: B902
+        self, args = args[0], args[1:]
+        self.__attrs = dict(*args, **kwargs)
+
+    def __getattribute__(self, name: str) -> t.Any:
+        # __class__ is needed for the awaitable check in async mode
+        if name in {"_Namespace__attrs", "__class__"}:
+            return object.__getattribute__(self, name)
+        try:
+            return self.__attrs[name]
+        except KeyError:
+            raise AttributeError(name) from None
+
+    def __setitem__(self, name: str, value: t.Any) -> None:
+        self.__attrs[name] = value
+
+    def __repr__(self) -> str:
+        return f"<Namespace {self.__attrs!r}>"
diff --git a/backend/venv/lib/python3.12/site-packages/jinja2/visitor.py b/backend/venv/lib/python3.12/site-packages/jinja2/visitor.py
new file mode 100644
index 00000000..7b8e1806
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jinja2/visitor.py
@@ -0,0 +1,92 @@
+"""API for traversing the AST nodes. Implemented by the compiler and
+meta introspection.
+"""
+
+import typing as t
+
+from .nodes import Node
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+    class VisitCallable(te.Protocol):
+        def __call__(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any: ...
+
+
+class NodeVisitor:
+    """Walks the abstract syntax tree and call visitor functions for every
+    node found.  The visitor functions may return values which will be
+    forwarded by the `visit` method.
+
+    Per default the visitor functions for the nodes are ``'visit_'`` +
+    class name of the node.  So a `TryFinally` node visit function would
+    be `visit_TryFinally`.  This behavior can be changed by overriding
+    the `get_visitor` function.  If no visitor function exists for a node
+    (return value `None`) the `generic_visit` visitor is used instead.
+    """
+
+    def get_visitor(self, node: Node) -> "t.Optional[VisitCallable]":
+        """Return the visitor function for this node or `None` if no visitor
+        exists for this node.  In that case the generic visit function is
+        used instead.
+        """
+        return getattr(self, f"visit_{type(node).__name__}", None)
+
+    def visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        """Visit a node."""
+        f = self.get_visitor(node)
+
+        if f is not None:
+            return f(node, *args, **kwargs)
+
+        return self.generic_visit(node, *args, **kwargs)
+
+    def generic_visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any:
+        """Called if no explicit visitor function exists for a node."""
+        for child_node in node.iter_child_nodes():
+            self.visit(child_node, *args, **kwargs)
+
+
+class NodeTransformer(NodeVisitor):
+    """Walks the abstract syntax tree and allows modifications of nodes.
+
+    The `NodeTransformer` will walk the AST and use the return value of the
+    visitor functions to replace or remove the old node.  If the return
+    value of the visitor function is `None` the node will be removed
+    from the previous location otherwise it's replaced with the return
+    value.  The return value may be the original node in which case no
+    replacement takes place.
+    """
+
+    def generic_visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> Node:
+        for field, old_value in node.iter_fields():
+            if isinstance(old_value, list):
+                new_values = []
+                for value in old_value:
+                    if isinstance(value, Node):
+                        value = self.visit(value, *args, **kwargs)
+                        if value is None:
+                            continue
+                        elif not isinstance(value, Node):
+                            new_values.extend(value)
+                            continue
+                    new_values.append(value)
+                old_value[:] = new_values
+            elif isinstance(old_value, Node):
+                new_node = self.visit(old_value, *args, **kwargs)
+                if new_node is None:
+                    delattr(node, field)
+                else:
+                    setattr(node, field, new_node)
+        return node
+
+    def visit_list(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.List[Node]:
+        """As transformers may return lists in some places this method
+        can be used to enforce a list as return value.
+        """
+        rv = self.visit(node, *args, **kwargs)
+
+        if not isinstance(rv, list):
+            return [rv]
+
+        return rv
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__init__.py b/backend/venv/lib/python3.12/site-packages/jwt/__init__.py
new file mode 100644
index 00000000..457a4e35
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/__init__.py
@@ -0,0 +1,75 @@
+from .api_jwk import PyJWK, PyJWKSet
+from .api_jws import (
+    PyJWS,
+    get_algorithm_by_name,
+    get_unverified_header,
+    register_algorithm,
+    unregister_algorithm,
+)
+from .api_jwt import PyJWT, decode, decode_complete, encode
+from .exceptions import (
+    DecodeError,
+    ExpiredSignatureError,
+    ImmatureSignatureError,
+    InvalidAlgorithmError,
+    InvalidAudienceError,
+    InvalidIssuedAtError,
+    InvalidIssuerError,
+    InvalidKeyError,
+    InvalidSignatureError,
+    InvalidTokenError,
+    MissingRequiredClaimError,
+    PyJWKClientConnectionError,
+    PyJWKClientError,
+    PyJWKError,
+    PyJWKSetError,
+    PyJWTError,
+)
+from .jwks_client import PyJWKClient
+
+__version__ = "2.10.1"
+
+__title__ = "PyJWT"
+__description__ = "JSON Web Token implementation in Python"
+__url__ = "https://pyjwt.readthedocs.io"
+__uri__ = __url__
+__doc__ = f"{__description__} <{__uri__}>"
+
+__author__ = "José Padilla"
+__email__ = "hello@jpadilla.com"
+
+__license__ = "MIT"
+__copyright__ = "Copyright 2015-2022 José Padilla"
+
+
+__all__ = [
+    "PyJWS",
+    "PyJWT",
+    "PyJWKClient",
+    "PyJWK",
+    "PyJWKSet",
+    "decode",
+    "decode_complete",
+    "encode",
+    "get_unverified_header",
+    "register_algorithm",
+    "unregister_algorithm",
+    "get_algorithm_by_name",
+    # Exceptions
+    "DecodeError",
+    "ExpiredSignatureError",
+    "ImmatureSignatureError",
+    "InvalidAlgorithmError",
+    "InvalidAudienceError",
+    "InvalidIssuedAtError",
+    "InvalidIssuerError",
+    "InvalidKeyError",
+    "InvalidSignatureError",
+    "InvalidTokenError",
+    "MissingRequiredClaimError",
+    "PyJWKClientConnectionError",
+    "PyJWKClientError",
+    "PyJWKError",
+    "PyJWKSetError",
+    "PyJWTError",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2ea50453
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/algorithms.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/algorithms.cpython-312.pyc
new file mode 100644
index 00000000..0d085eed
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/algorithms.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwk.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwk.cpython-312.pyc
new file mode 100644
index 00000000..e930a9a8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwk.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jws.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jws.cpython-312.pyc
new file mode 100644
index 00000000..4953690b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jws.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwt.cpython-312.pyc
new file mode 100644
index 00000000..b742bfcc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/api_jwt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..1c831f69
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/help.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/help.cpython-312.pyc
new file mode 100644
index 00000000..b9f8bad4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/help.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwk_set_cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwk_set_cache.cpython-312.pyc
new file mode 100644
index 00000000..a47cd454
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwk_set_cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwks_client.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwks_client.cpython-312.pyc
new file mode 100644
index 00000000..e8052ce2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/jwks_client.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..6f1592e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..291993ef
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/warnings.cpython-312.pyc
new file mode 100644
index 00000000..5dbcdcb5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/jwt/__pycache__/warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/algorithms.py b/backend/venv/lib/python3.12/site-packages/jwt/algorithms.py
new file mode 100644
index 00000000..ccb1500f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/algorithms.py
@@ -0,0 +1,875 @@
+from __future__ import annotations
+
+import hashlib
+import hmac
+import json
+from abc import ABC, abstractmethod
+from typing import TYPE_CHECKING, Any, ClassVar, Literal, NoReturn, cast, overload
+
+from .exceptions import InvalidKeyError
+from .types import HashlibHash, JWKDict
+from .utils import (
+    base64url_decode,
+    base64url_encode,
+    der_to_raw_signature,
+    force_bytes,
+    from_base64url_uint,
+    is_pem_format,
+    is_ssh_key,
+    raw_to_der_signature,
+    to_base64url_uint,
+)
+
+try:
+    from cryptography.exceptions import InvalidSignature, UnsupportedAlgorithm
+    from cryptography.hazmat.backends import default_backend
+    from cryptography.hazmat.primitives import hashes
+    from cryptography.hazmat.primitives.asymmetric import padding
+    from cryptography.hazmat.primitives.asymmetric.ec import (
+        ECDSA,
+        SECP256K1,
+        SECP256R1,
+        SECP384R1,
+        SECP521R1,
+        EllipticCurve,
+        EllipticCurvePrivateKey,
+        EllipticCurvePrivateNumbers,
+        EllipticCurvePublicKey,
+        EllipticCurvePublicNumbers,
+    )
+    from cryptography.hazmat.primitives.asymmetric.ed448 import (
+        Ed448PrivateKey,
+        Ed448PublicKey,
+    )
+    from cryptography.hazmat.primitives.asymmetric.ed25519 import (
+        Ed25519PrivateKey,
+        Ed25519PublicKey,
+    )
+    from cryptography.hazmat.primitives.asymmetric.rsa import (
+        RSAPrivateKey,
+        RSAPrivateNumbers,
+        RSAPublicKey,
+        RSAPublicNumbers,
+        rsa_crt_dmp1,
+        rsa_crt_dmq1,
+        rsa_crt_iqmp,
+        rsa_recover_prime_factors,
+    )
+    from cryptography.hazmat.primitives.serialization import (
+        Encoding,
+        NoEncryption,
+        PrivateFormat,
+        PublicFormat,
+        load_pem_private_key,
+        load_pem_public_key,
+        load_ssh_public_key,
+    )
+
+    has_crypto = True
+except ModuleNotFoundError:
+    has_crypto = False
+
+
+if TYPE_CHECKING:
+    # Type aliases for convenience in algorithms method signatures
+    AllowedRSAKeys = RSAPrivateKey | RSAPublicKey
+    AllowedECKeys = EllipticCurvePrivateKey | EllipticCurvePublicKey
+    AllowedOKPKeys = (
+        Ed25519PrivateKey | Ed25519PublicKey | Ed448PrivateKey | Ed448PublicKey
+    )
+    AllowedKeys = AllowedRSAKeys | AllowedECKeys | AllowedOKPKeys
+    AllowedPrivateKeys = (
+        RSAPrivateKey | EllipticCurvePrivateKey | Ed25519PrivateKey | Ed448PrivateKey
+    )
+    AllowedPublicKeys = (
+        RSAPublicKey | EllipticCurvePublicKey | Ed25519PublicKey | Ed448PublicKey
+    )
+
+
+requires_cryptography = {
+    "RS256",
+    "RS384",
+    "RS512",
+    "ES256",
+    "ES256K",
+    "ES384",
+    "ES521",
+    "ES512",
+    "PS256",
+    "PS384",
+    "PS512",
+    "EdDSA",
+}
+
+
+def get_default_algorithms() -> dict[str, Algorithm]:
+    """
+    Returns the algorithms that are implemented by the library.
+    """
+    default_algorithms = {
+        "none": NoneAlgorithm(),
+        "HS256": HMACAlgorithm(HMACAlgorithm.SHA256),
+        "HS384": HMACAlgorithm(HMACAlgorithm.SHA384),
+        "HS512": HMACAlgorithm(HMACAlgorithm.SHA512),
+    }
+
+    if has_crypto:
+        default_algorithms.update(
+            {
+                "RS256": RSAAlgorithm(RSAAlgorithm.SHA256),
+                "RS384": RSAAlgorithm(RSAAlgorithm.SHA384),
+                "RS512": RSAAlgorithm(RSAAlgorithm.SHA512),
+                "ES256": ECAlgorithm(ECAlgorithm.SHA256),
+                "ES256K": ECAlgorithm(ECAlgorithm.SHA256),
+                "ES384": ECAlgorithm(ECAlgorithm.SHA384),
+                "ES521": ECAlgorithm(ECAlgorithm.SHA512),
+                "ES512": ECAlgorithm(
+                    ECAlgorithm.SHA512
+                ),  # Backward compat for #219 fix
+                "PS256": RSAPSSAlgorithm(RSAPSSAlgorithm.SHA256),
+                "PS384": RSAPSSAlgorithm(RSAPSSAlgorithm.SHA384),
+                "PS512": RSAPSSAlgorithm(RSAPSSAlgorithm.SHA512),
+                "EdDSA": OKPAlgorithm(),
+            }
+        )
+
+    return default_algorithms
+
+
+class Algorithm(ABC):
+    """
+    The interface for an algorithm used to sign and verify tokens.
+    """
+
+    def compute_hash_digest(self, bytestr: bytes) -> bytes:
+        """
+        Compute a hash digest using the specified algorithm's hash algorithm.
+
+        If there is no hash algorithm, raises a NotImplementedError.
+        """
+        # lookup self.hash_alg if defined in a way that mypy can understand
+        hash_alg = getattr(self, "hash_alg", None)
+        if hash_alg is None:
+            raise NotImplementedError
+
+        if (
+            has_crypto
+            and isinstance(hash_alg, type)
+            and issubclass(hash_alg, hashes.HashAlgorithm)
+        ):
+            digest = hashes.Hash(hash_alg(), backend=default_backend())
+            digest.update(bytestr)
+            return bytes(digest.finalize())
+        else:
+            return bytes(hash_alg(bytestr).digest())
+
+    @abstractmethod
+    def prepare_key(self, key: Any) -> Any:
+        """
+        Performs necessary validation and conversions on the key and returns
+        the key value in the proper format for sign() and verify().
+        """
+
+    @abstractmethod
+    def sign(self, msg: bytes, key: Any) -> bytes:
+        """
+        Returns a digital signature for the specified message
+        using the specified key value.
+        """
+
+    @abstractmethod
+    def verify(self, msg: bytes, key: Any, sig: bytes) -> bool:
+        """
+        Verifies that the specified digital signature is valid
+        for the specified message and key values.
+        """
+
+    @overload
+    @staticmethod
+    @abstractmethod
+    def to_jwk(key_obj, as_dict: Literal[True]) -> JWKDict: ...  # pragma: no cover
+
+    @overload
+    @staticmethod
+    @abstractmethod
+    def to_jwk(key_obj, as_dict: Literal[False] = False) -> str: ...  # pragma: no cover
+
+    @staticmethod
+    @abstractmethod
+    def to_jwk(key_obj, as_dict: bool = False) -> JWKDict | str:
+        """
+        Serializes a given key into a JWK
+        """
+
+    @staticmethod
+    @abstractmethod
+    def from_jwk(jwk: str | JWKDict) -> Any:
+        """
+        Deserializes a given key from JWK back into a key object
+        """
+
+
+class NoneAlgorithm(Algorithm):
+    """
+    Placeholder for use when no signing or verification
+    operations are required.
+    """
+
+    def prepare_key(self, key: str | None) -> None:
+        if key == "":
+            key = None
+
+        if key is not None:
+            raise InvalidKeyError('When alg = "none", key value must be None.')
+
+        return key
+
+    def sign(self, msg: bytes, key: None) -> bytes:
+        return b""
+
+    def verify(self, msg: bytes, key: None, sig: bytes) -> bool:
+        return False
+
+    @staticmethod
+    def to_jwk(key_obj: Any, as_dict: bool = False) -> NoReturn:
+        raise NotImplementedError()
+
+    @staticmethod
+    def from_jwk(jwk: str | JWKDict) -> NoReturn:
+        raise NotImplementedError()
+
+
+class HMACAlgorithm(Algorithm):
+    """
+    Performs signing and verification operations using HMAC
+    and the specified hash function.
+    """
+
+    SHA256: ClassVar[HashlibHash] = hashlib.sha256
+    SHA384: ClassVar[HashlibHash] = hashlib.sha384
+    SHA512: ClassVar[HashlibHash] = hashlib.sha512
+
+    def __init__(self, hash_alg: HashlibHash) -> None:
+        self.hash_alg = hash_alg
+
+    def prepare_key(self, key: str | bytes) -> bytes:
+        key_bytes = force_bytes(key)
+
+        if is_pem_format(key_bytes) or is_ssh_key(key_bytes):
+            raise InvalidKeyError(
+                "The specified key is an asymmetric key or x509 certificate and"
+                " should not be used as an HMAC secret."
+            )
+
+        return key_bytes
+
+    @overload
+    @staticmethod
+    def to_jwk(
+        key_obj: str | bytes, as_dict: Literal[True]
+    ) -> JWKDict: ...  # pragma: no cover
+
+    @overload
+    @staticmethod
+    def to_jwk(
+        key_obj: str | bytes, as_dict: Literal[False] = False
+    ) -> str: ...  # pragma: no cover
+
+    @staticmethod
+    def to_jwk(key_obj: str | bytes, as_dict: bool = False) -> JWKDict | str:
+        jwk = {
+            "k": base64url_encode(force_bytes(key_obj)).decode(),
+            "kty": "oct",
+        }
+
+        if as_dict:
+            return jwk
+        else:
+            return json.dumps(jwk)
+
+    @staticmethod
+    def from_jwk(jwk: str | JWKDict) -> bytes:
+        try:
+            if isinstance(jwk, str):
+                obj: JWKDict = json.loads(jwk)
+            elif isinstance(jwk, dict):
+                obj = jwk
+            else:
+                raise ValueError
+        except ValueError:
+            raise InvalidKeyError("Key is not valid JSON") from None
+
+        if obj.get("kty") != "oct":
+            raise InvalidKeyError("Not an HMAC key")
+
+        return base64url_decode(obj["k"])
+
+    def sign(self, msg: bytes, key: bytes) -> bytes:
+        return hmac.new(key, msg, self.hash_alg).digest()
+
+    def verify(self, msg: bytes, key: bytes, sig: bytes) -> bool:
+        return hmac.compare_digest(sig, self.sign(msg, key))
+
+
+if has_crypto:
+
+    class RSAAlgorithm(Algorithm):
+        """
+        Performs signing and verification operations using
+        RSASSA-PKCS-v1_5 and the specified hash function.
+        """
+
+        SHA256: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA256
+        SHA384: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA384
+        SHA512: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA512
+
+        def __init__(self, hash_alg: type[hashes.HashAlgorithm]) -> None:
+            self.hash_alg = hash_alg
+
+        def prepare_key(self, key: AllowedRSAKeys | str | bytes) -> AllowedRSAKeys:
+            if isinstance(key, (RSAPrivateKey, RSAPublicKey)):
+                return key
+
+            if not isinstance(key, (bytes, str)):
+                raise TypeError("Expecting a PEM-formatted key.")
+
+            key_bytes = force_bytes(key)
+
+            try:
+                if key_bytes.startswith(b"ssh-rsa"):
+                    return cast(RSAPublicKey, load_ssh_public_key(key_bytes))
+                else:
+                    return cast(
+                        RSAPrivateKey, load_pem_private_key(key_bytes, password=None)
+                    )
+            except ValueError:
+                try:
+                    return cast(RSAPublicKey, load_pem_public_key(key_bytes))
+                except (ValueError, UnsupportedAlgorithm):
+                    raise InvalidKeyError(
+                        "Could not parse the provided public key."
+                    ) from None
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key_obj: AllowedRSAKeys, as_dict: Literal[True]
+        ) -> JWKDict: ...  # pragma: no cover
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key_obj: AllowedRSAKeys, as_dict: Literal[False] = False
+        ) -> str: ...  # pragma: no cover
+
+        @staticmethod
+        def to_jwk(key_obj: AllowedRSAKeys, as_dict: bool = False) -> JWKDict | str:
+            obj: dict[str, Any] | None = None
+
+            if hasattr(key_obj, "private_numbers"):
+                # Private key
+                numbers = key_obj.private_numbers()
+
+                obj = {
+                    "kty": "RSA",
+                    "key_ops": ["sign"],
+                    "n": to_base64url_uint(numbers.public_numbers.n).decode(),
+                    "e": to_base64url_uint(numbers.public_numbers.e).decode(),
+                    "d": to_base64url_uint(numbers.d).decode(),
+                    "p": to_base64url_uint(numbers.p).decode(),
+                    "q": to_base64url_uint(numbers.q).decode(),
+                    "dp": to_base64url_uint(numbers.dmp1).decode(),
+                    "dq": to_base64url_uint(numbers.dmq1).decode(),
+                    "qi": to_base64url_uint(numbers.iqmp).decode(),
+                }
+
+            elif hasattr(key_obj, "verify"):
+                # Public key
+                numbers = key_obj.public_numbers()
+
+                obj = {
+                    "kty": "RSA",
+                    "key_ops": ["verify"],
+                    "n": to_base64url_uint(numbers.n).decode(),
+                    "e": to_base64url_uint(numbers.e).decode(),
+                }
+            else:
+                raise InvalidKeyError("Not a public or private key")
+
+            if as_dict:
+                return obj
+            else:
+                return json.dumps(obj)
+
+        @staticmethod
+        def from_jwk(jwk: str | JWKDict) -> AllowedRSAKeys:
+            try:
+                if isinstance(jwk, str):
+                    obj = json.loads(jwk)
+                elif isinstance(jwk, dict):
+                    obj = jwk
+                else:
+                    raise ValueError
+            except ValueError:
+                raise InvalidKeyError("Key is not valid JSON") from None
+
+            if obj.get("kty") != "RSA":
+                raise InvalidKeyError("Not an RSA key") from None
+
+            if "d" in obj and "e" in obj and "n" in obj:
+                # Private key
+                if "oth" in obj:
+                    raise InvalidKeyError(
+                        "Unsupported RSA private key: > 2 primes not supported"
+                    )
+
+                other_props = ["p", "q", "dp", "dq", "qi"]
+                props_found = [prop in obj for prop in other_props]
+                any_props_found = any(props_found)
+
+                if any_props_found and not all(props_found):
+                    raise InvalidKeyError(
+                        "RSA key must include all parameters if any are present besides d"
+                    ) from None
+
+                public_numbers = RSAPublicNumbers(
+                    from_base64url_uint(obj["e"]),
+                    from_base64url_uint(obj["n"]),
+                )
+
+                if any_props_found:
+                    numbers = RSAPrivateNumbers(
+                        d=from_base64url_uint(obj["d"]),
+                        p=from_base64url_uint(obj["p"]),
+                        q=from_base64url_uint(obj["q"]),
+                        dmp1=from_base64url_uint(obj["dp"]),
+                        dmq1=from_base64url_uint(obj["dq"]),
+                        iqmp=from_base64url_uint(obj["qi"]),
+                        public_numbers=public_numbers,
+                    )
+                else:
+                    d = from_base64url_uint(obj["d"])
+                    p, q = rsa_recover_prime_factors(
+                        public_numbers.n, d, public_numbers.e
+                    )
+
+                    numbers = RSAPrivateNumbers(
+                        d=d,
+                        p=p,
+                        q=q,
+                        dmp1=rsa_crt_dmp1(d, p),
+                        dmq1=rsa_crt_dmq1(d, q),
+                        iqmp=rsa_crt_iqmp(p, q),
+                        public_numbers=public_numbers,
+                    )
+
+                return numbers.private_key()
+            elif "n" in obj and "e" in obj:
+                # Public key
+                return RSAPublicNumbers(
+                    from_base64url_uint(obj["e"]),
+                    from_base64url_uint(obj["n"]),
+                ).public_key()
+            else:
+                raise InvalidKeyError("Not a public or private key")
+
+        def sign(self, msg: bytes, key: RSAPrivateKey) -> bytes:
+            return key.sign(msg, padding.PKCS1v15(), self.hash_alg())
+
+        def verify(self, msg: bytes, key: RSAPublicKey, sig: bytes) -> bool:
+            try:
+                key.verify(sig, msg, padding.PKCS1v15(), self.hash_alg())
+                return True
+            except InvalidSignature:
+                return False
+
+    class ECAlgorithm(Algorithm):
+        """
+        Performs signing and verification operations using
+        ECDSA and the specified hash function
+        """
+
+        SHA256: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA256
+        SHA384: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA384
+        SHA512: ClassVar[type[hashes.HashAlgorithm]] = hashes.SHA512
+
+        def __init__(self, hash_alg: type[hashes.HashAlgorithm]) -> None:
+            self.hash_alg = hash_alg
+
+        def prepare_key(self, key: AllowedECKeys | str | bytes) -> AllowedECKeys:
+            if isinstance(key, (EllipticCurvePrivateKey, EllipticCurvePublicKey)):
+                return key
+
+            if not isinstance(key, (bytes, str)):
+                raise TypeError("Expecting a PEM-formatted key.")
+
+            key_bytes = force_bytes(key)
+
+            # Attempt to load key. We don't know if it's
+            # a Signing Key or a Verifying Key, so we try
+            # the Verifying Key first.
+            try:
+                if key_bytes.startswith(b"ecdsa-sha2-"):
+                    crypto_key = load_ssh_public_key(key_bytes)
+                else:
+                    crypto_key = load_pem_public_key(key_bytes)  # type: ignore[assignment]
+            except ValueError:
+                crypto_key = load_pem_private_key(key_bytes, password=None)  # type: ignore[assignment]
+
+            # Explicit check the key to prevent confusing errors from cryptography
+            if not isinstance(
+                crypto_key, (EllipticCurvePrivateKey, EllipticCurvePublicKey)
+            ):
+                raise InvalidKeyError(
+                    "Expecting a EllipticCurvePrivateKey/EllipticCurvePublicKey. Wrong key provided for ECDSA algorithms"
+                ) from None
+
+            return crypto_key
+
+        def sign(self, msg: bytes, key: EllipticCurvePrivateKey) -> bytes:
+            der_sig = key.sign(msg, ECDSA(self.hash_alg()))
+
+            return der_to_raw_signature(der_sig, key.curve)
+
+        def verify(self, msg: bytes, key: AllowedECKeys, sig: bytes) -> bool:
+            try:
+                der_sig = raw_to_der_signature(sig, key.curve)
+            except ValueError:
+                return False
+
+            try:
+                public_key = (
+                    key.public_key()
+                    if isinstance(key, EllipticCurvePrivateKey)
+                    else key
+                )
+                public_key.verify(der_sig, msg, ECDSA(self.hash_alg()))
+                return True
+            except InvalidSignature:
+                return False
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key_obj: AllowedECKeys, as_dict: Literal[True]
+        ) -> JWKDict: ...  # pragma: no cover
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key_obj: AllowedECKeys, as_dict: Literal[False] = False
+        ) -> str: ...  # pragma: no cover
+
+        @staticmethod
+        def to_jwk(key_obj: AllowedECKeys, as_dict: bool = False) -> JWKDict | str:
+            if isinstance(key_obj, EllipticCurvePrivateKey):
+                public_numbers = key_obj.public_key().public_numbers()
+            elif isinstance(key_obj, EllipticCurvePublicKey):
+                public_numbers = key_obj.public_numbers()
+            else:
+                raise InvalidKeyError("Not a public or private key")
+
+            if isinstance(key_obj.curve, SECP256R1):
+                crv = "P-256"
+            elif isinstance(key_obj.curve, SECP384R1):
+                crv = "P-384"
+            elif isinstance(key_obj.curve, SECP521R1):
+                crv = "P-521"
+            elif isinstance(key_obj.curve, SECP256K1):
+                crv = "secp256k1"
+            else:
+                raise InvalidKeyError(f"Invalid curve: {key_obj.curve}")
+
+            obj: dict[str, Any] = {
+                "kty": "EC",
+                "crv": crv,
+                "x": to_base64url_uint(
+                    public_numbers.x,
+                    bit_length=key_obj.curve.key_size,
+                ).decode(),
+                "y": to_base64url_uint(
+                    public_numbers.y,
+                    bit_length=key_obj.curve.key_size,
+                ).decode(),
+            }
+
+            if isinstance(key_obj, EllipticCurvePrivateKey):
+                obj["d"] = to_base64url_uint(
+                    key_obj.private_numbers().private_value,
+                    bit_length=key_obj.curve.key_size,
+                ).decode()
+
+            if as_dict:
+                return obj
+            else:
+                return json.dumps(obj)
+
+        @staticmethod
+        def from_jwk(jwk: str | JWKDict) -> AllowedECKeys:
+            try:
+                if isinstance(jwk, str):
+                    obj = json.loads(jwk)
+                elif isinstance(jwk, dict):
+                    obj = jwk
+                else:
+                    raise ValueError
+            except ValueError:
+                raise InvalidKeyError("Key is not valid JSON") from None
+
+            if obj.get("kty") != "EC":
+                raise InvalidKeyError("Not an Elliptic curve key") from None
+
+            if "x" not in obj or "y" not in obj:
+                raise InvalidKeyError("Not an Elliptic curve key") from None
+
+            x = base64url_decode(obj.get("x"))
+            y = base64url_decode(obj.get("y"))
+
+            curve = obj.get("crv")
+            curve_obj: EllipticCurve
+
+            if curve == "P-256":
+                if len(x) == len(y) == 32:
+                    curve_obj = SECP256R1()
+                else:
+                    raise InvalidKeyError(
+                        "Coords should be 32 bytes for curve P-256"
+                    ) from None
+            elif curve == "P-384":
+                if len(x) == len(y) == 48:
+                    curve_obj = SECP384R1()
+                else:
+                    raise InvalidKeyError(
+                        "Coords should be 48 bytes for curve P-384"
+                    ) from None
+            elif curve == "P-521":
+                if len(x) == len(y) == 66:
+                    curve_obj = SECP521R1()
+                else:
+                    raise InvalidKeyError(
+                        "Coords should be 66 bytes for curve P-521"
+                    ) from None
+            elif curve == "secp256k1":
+                if len(x) == len(y) == 32:
+                    curve_obj = SECP256K1()
+                else:
+                    raise InvalidKeyError(
+                        "Coords should be 32 bytes for curve secp256k1"
+                    )
+            else:
+                raise InvalidKeyError(f"Invalid curve: {curve}")
+
+            public_numbers = EllipticCurvePublicNumbers(
+                x=int.from_bytes(x, byteorder="big"),
+                y=int.from_bytes(y, byteorder="big"),
+                curve=curve_obj,
+            )
+
+            if "d" not in obj:
+                return public_numbers.public_key()
+
+            d = base64url_decode(obj.get("d"))
+            if len(d) != len(x):
+                raise InvalidKeyError(
+                    "D should be {} bytes for curve {}", len(x), curve
+                )
+
+            return EllipticCurvePrivateNumbers(
+                int.from_bytes(d, byteorder="big"), public_numbers
+            ).private_key()
+
+    class RSAPSSAlgorithm(RSAAlgorithm):
+        """
+        Performs a signature using RSASSA-PSS with MGF1
+        """
+
+        def sign(self, msg: bytes, key: RSAPrivateKey) -> bytes:
+            return key.sign(
+                msg,
+                padding.PSS(
+                    mgf=padding.MGF1(self.hash_alg()),
+                    salt_length=self.hash_alg().digest_size,
+                ),
+                self.hash_alg(),
+            )
+
+        def verify(self, msg: bytes, key: RSAPublicKey, sig: bytes) -> bool:
+            try:
+                key.verify(
+                    sig,
+                    msg,
+                    padding.PSS(
+                        mgf=padding.MGF1(self.hash_alg()),
+                        salt_length=self.hash_alg().digest_size,
+                    ),
+                    self.hash_alg(),
+                )
+                return True
+            except InvalidSignature:
+                return False
+
+    class OKPAlgorithm(Algorithm):
+        """
+        Performs signing and verification operations using EdDSA
+
+        This class requires ``cryptography>=2.6`` to be installed.
+        """
+
+        def __init__(self, **kwargs: Any) -> None:
+            pass
+
+        def prepare_key(self, key: AllowedOKPKeys | str | bytes) -> AllowedOKPKeys:
+            if isinstance(key, (bytes, str)):
+                key_str = key.decode("utf-8") if isinstance(key, bytes) else key
+                key_bytes = key.encode("utf-8") if isinstance(key, str) else key
+
+                if "-----BEGIN PUBLIC" in key_str:
+                    key = load_pem_public_key(key_bytes)  # type: ignore[assignment]
+                elif "-----BEGIN PRIVATE" in key_str:
+                    key = load_pem_private_key(key_bytes, password=None)  # type: ignore[assignment]
+                elif key_str[0:4] == "ssh-":
+                    key = load_ssh_public_key(key_bytes)  # type: ignore[assignment]
+
+            # Explicit check the key to prevent confusing errors from cryptography
+            if not isinstance(
+                key,
+                (Ed25519PrivateKey, Ed25519PublicKey, Ed448PrivateKey, Ed448PublicKey),
+            ):
+                raise InvalidKeyError(
+                    "Expecting a EllipticCurvePrivateKey/EllipticCurvePublicKey. Wrong key provided for EdDSA algorithms"
+                )
+
+            return key
+
+        def sign(
+            self, msg: str | bytes, key: Ed25519PrivateKey | Ed448PrivateKey
+        ) -> bytes:
+            """
+            Sign a message ``msg`` using the EdDSA private key ``key``
+            :param str|bytes msg: Message to sign
+            :param Ed25519PrivateKey}Ed448PrivateKey key: A :class:`.Ed25519PrivateKey`
+                or :class:`.Ed448PrivateKey` isinstance
+            :return bytes signature: The signature, as bytes
+            """
+            msg_bytes = msg.encode("utf-8") if isinstance(msg, str) else msg
+            return key.sign(msg_bytes)
+
+        def verify(
+            self, msg: str | bytes, key: AllowedOKPKeys, sig: str | bytes
+        ) -> bool:
+            """
+            Verify a given ``msg`` against a signature ``sig`` using the EdDSA key ``key``
+
+            :param str|bytes sig: EdDSA signature to check ``msg`` against
+            :param str|bytes msg: Message to sign
+            :param Ed25519PrivateKey|Ed25519PublicKey|Ed448PrivateKey|Ed448PublicKey key:
+                A private or public EdDSA key instance
+            :return bool verified: True if signature is valid, False if not.
+            """
+            try:
+                msg_bytes = msg.encode("utf-8") if isinstance(msg, str) else msg
+                sig_bytes = sig.encode("utf-8") if isinstance(sig, str) else sig
+
+                public_key = (
+                    key.public_key()
+                    if isinstance(key, (Ed25519PrivateKey, Ed448PrivateKey))
+                    else key
+                )
+                public_key.verify(sig_bytes, msg_bytes)
+                return True  # If no exception was raised, the signature is valid.
+            except InvalidSignature:
+                return False
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key: AllowedOKPKeys, as_dict: Literal[True]
+        ) -> JWKDict: ...  # pragma: no cover
+
+        @overload
+        @staticmethod
+        def to_jwk(
+            key: AllowedOKPKeys, as_dict: Literal[False] = False
+        ) -> str: ...  # pragma: no cover
+
+        @staticmethod
+        def to_jwk(key: AllowedOKPKeys, as_dict: bool = False) -> JWKDict | str:
+            if isinstance(key, (Ed25519PublicKey, Ed448PublicKey)):
+                x = key.public_bytes(
+                    encoding=Encoding.Raw,
+                    format=PublicFormat.Raw,
+                )
+                crv = "Ed25519" if isinstance(key, Ed25519PublicKey) else "Ed448"
+
+                obj = {
+                    "x": base64url_encode(force_bytes(x)).decode(),
+                    "kty": "OKP",
+                    "crv": crv,
+                }
+
+                if as_dict:
+                    return obj
+                else:
+                    return json.dumps(obj)
+
+            if isinstance(key, (Ed25519PrivateKey, Ed448PrivateKey)):
+                d = key.private_bytes(
+                    encoding=Encoding.Raw,
+                    format=PrivateFormat.Raw,
+                    encryption_algorithm=NoEncryption(),
+                )
+
+                x = key.public_key().public_bytes(
+                    encoding=Encoding.Raw,
+                    format=PublicFormat.Raw,
+                )
+
+                crv = "Ed25519" if isinstance(key, Ed25519PrivateKey) else "Ed448"
+                obj = {
+                    "x": base64url_encode(force_bytes(x)).decode(),
+                    "d": base64url_encode(force_bytes(d)).decode(),
+                    "kty": "OKP",
+                    "crv": crv,
+                }
+
+                if as_dict:
+                    return obj
+                else:
+                    return json.dumps(obj)
+
+            raise InvalidKeyError("Not a public or private key")
+
+        @staticmethod
+        def from_jwk(jwk: str | JWKDict) -> AllowedOKPKeys:
+            try:
+                if isinstance(jwk, str):
+                    obj = json.loads(jwk)
+                elif isinstance(jwk, dict):
+                    obj = jwk
+                else:
+                    raise ValueError
+            except ValueError:
+                raise InvalidKeyError("Key is not valid JSON") from None
+
+            if obj.get("kty") != "OKP":
+                raise InvalidKeyError("Not an Octet Key Pair")
+
+            curve = obj.get("crv")
+            if curve != "Ed25519" and curve != "Ed448":
+                raise InvalidKeyError(f"Invalid curve: {curve}")
+
+            if "x" not in obj:
+                raise InvalidKeyError('OKP should have "x" parameter')
+            x = base64url_decode(obj.get("x"))
+
+            try:
+                if "d" not in obj:
+                    if curve == "Ed25519":
+                        return Ed25519PublicKey.from_public_bytes(x)
+                    return Ed448PublicKey.from_public_bytes(x)
+                d = base64url_decode(obj.get("d"))
+                if curve == "Ed25519":
+                    return Ed25519PrivateKey.from_private_bytes(d)
+                return Ed448PrivateKey.from_private_bytes(d)
+            except ValueError as err:
+                raise InvalidKeyError("Invalid key parameter") from err
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/api_jwk.py b/backend/venv/lib/python3.12/site-packages/jwt/api_jwk.py
new file mode 100644
index 00000000..02f4679c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/api_jwk.py
@@ -0,0 +1,144 @@
+from __future__ import annotations
+
+import json
+import time
+from typing import Any
+
+from .algorithms import get_default_algorithms, has_crypto, requires_cryptography
+from .exceptions import (
+    InvalidKeyError,
+    MissingCryptographyError,
+    PyJWKError,
+    PyJWKSetError,
+    PyJWTError,
+)
+from .types import JWKDict
+
+
+class PyJWK:
+    def __init__(self, jwk_data: JWKDict, algorithm: str | None = None) -> None:
+        self._algorithms = get_default_algorithms()
+        self._jwk_data = jwk_data
+
+        kty = self._jwk_data.get("kty", None)
+        if not kty:
+            raise InvalidKeyError(f"kty is not found: {self._jwk_data}")
+
+        if not algorithm and isinstance(self._jwk_data, dict):
+            algorithm = self._jwk_data.get("alg", None)
+
+        if not algorithm:
+            # Determine alg with kty (and crv).
+            crv = self._jwk_data.get("crv", None)
+            if kty == "EC":
+                if crv == "P-256" or not crv:
+                    algorithm = "ES256"
+                elif crv == "P-384":
+                    algorithm = "ES384"
+                elif crv == "P-521":
+                    algorithm = "ES512"
+                elif crv == "secp256k1":
+                    algorithm = "ES256K"
+                else:
+                    raise InvalidKeyError(f"Unsupported crv: {crv}")
+            elif kty == "RSA":
+                algorithm = "RS256"
+            elif kty == "oct":
+                algorithm = "HS256"
+            elif kty == "OKP":
+                if not crv:
+                    raise InvalidKeyError(f"crv is not found: {self._jwk_data}")
+                if crv == "Ed25519":
+                    algorithm = "EdDSA"
+                else:
+                    raise InvalidKeyError(f"Unsupported crv: {crv}")
+            else:
+                raise InvalidKeyError(f"Unsupported kty: {kty}")
+
+        if not has_crypto and algorithm in requires_cryptography:
+            raise MissingCryptographyError(
+                f"{algorithm} requires 'cryptography' to be installed."
+            )
+
+        self.algorithm_name = algorithm
+
+        if algorithm in self._algorithms:
+            self.Algorithm = self._algorithms[algorithm]
+        else:
+            raise PyJWKError(f"Unable to find an algorithm for key: {self._jwk_data}")
+
+        self.key = self.Algorithm.from_jwk(self._jwk_data)
+
+    @staticmethod
+    def from_dict(obj: JWKDict, algorithm: str | None = None) -> PyJWK:
+        return PyJWK(obj, algorithm)
+
+    @staticmethod
+    def from_json(data: str, algorithm: None = None) -> PyJWK:
+        obj = json.loads(data)
+        return PyJWK.from_dict(obj, algorithm)
+
+    @property
+    def key_type(self) -> str | None:
+        return self._jwk_data.get("kty", None)
+
+    @property
+    def key_id(self) -> str | None:
+        return self._jwk_data.get("kid", None)
+
+    @property
+    def public_key_use(self) -> str | None:
+        return self._jwk_data.get("use", None)
+
+
+class PyJWKSet:
+    def __init__(self, keys: list[JWKDict]) -> None:
+        self.keys = []
+
+        if not keys:
+            raise PyJWKSetError("The JWK Set did not contain any keys")
+
+        if not isinstance(keys, list):
+            raise PyJWKSetError("Invalid JWK Set value")
+
+        for key in keys:
+            try:
+                self.keys.append(PyJWK(key))
+            except PyJWTError as error:
+                if isinstance(error, MissingCryptographyError):
+                    raise error
+                # skip unusable keys
+                continue
+
+        if len(self.keys) == 0:
+            raise PyJWKSetError(
+                "The JWK Set did not contain any usable keys. Perhaps 'cryptography' is not installed?"
+            )
+
+    @staticmethod
+    def from_dict(obj: dict[str, Any]) -> PyJWKSet:
+        keys = obj.get("keys", [])
+        return PyJWKSet(keys)
+
+    @staticmethod
+    def from_json(data: str) -> PyJWKSet:
+        obj = json.loads(data)
+        return PyJWKSet.from_dict(obj)
+
+    def __getitem__(self, kid: str) -> PyJWK:
+        for key in self.keys:
+            if key.key_id == kid:
+                return key
+        raise KeyError(f"keyset has no key for kid: {kid}")
+
+
+class PyJWTSetWithTimestamp:
+    def __init__(self, jwk_set: PyJWKSet):
+        self.jwk_set = jwk_set
+        self.timestamp = time.monotonic()
+
+    def get_jwk_set(self) -> PyJWKSet:
+        return self.jwk_set
+
+    def get_timestamp(self) -> float:
+        return self.timestamp
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/api_jws.py b/backend/venv/lib/python3.12/site-packages/jwt/api_jws.py
new file mode 100644
index 00000000..654ee0b7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/api_jws.py
@@ -0,0 +1,346 @@
+from __future__ import annotations
+
+import binascii
+import json
+import warnings
+from collections.abc import Sequence
+from typing import TYPE_CHECKING, Any
+
+from .algorithms import (
+    Algorithm,
+    get_default_algorithms,
+    has_crypto,
+    requires_cryptography,
+)
+from .api_jwk import PyJWK
+from .exceptions import (
+    DecodeError,
+    InvalidAlgorithmError,
+    InvalidSignatureError,
+    InvalidTokenError,
+)
+from .utils import base64url_decode, base64url_encode
+from .warnings import RemovedInPyjwt3Warning
+
+if TYPE_CHECKING:
+    from .algorithms import AllowedPrivateKeys, AllowedPublicKeys
+
+
+class PyJWS:
+    header_typ = "JWT"
+
+    def __init__(
+        self,
+        algorithms: Sequence[str] | None = None,
+        options: dict[str, Any] | None = None,
+    ) -> None:
+        self._algorithms = get_default_algorithms()
+        self._valid_algs = (
+            set(algorithms) if algorithms is not None else set(self._algorithms)
+        )
+
+        # Remove algorithms that aren't on the whitelist
+        for key in list(self._algorithms.keys()):
+            if key not in self._valid_algs:
+                del self._algorithms[key]
+
+        if options is None:
+            options = {}
+        self.options = {**self._get_default_options(), **options}
+
+    @staticmethod
+    def _get_default_options() -> dict[str, bool]:
+        return {"verify_signature": True}
+
+    def register_algorithm(self, alg_id: str, alg_obj: Algorithm) -> None:
+        """
+        Registers a new Algorithm for use when creating and verifying tokens.
+        """
+        if alg_id in self._algorithms:
+            raise ValueError("Algorithm already has a handler.")
+
+        if not isinstance(alg_obj, Algorithm):
+            raise TypeError("Object is not of type `Algorithm`")
+
+        self._algorithms[alg_id] = alg_obj
+        self._valid_algs.add(alg_id)
+
+    def unregister_algorithm(self, alg_id: str) -> None:
+        """
+        Unregisters an Algorithm for use when creating and verifying tokens
+        Throws KeyError if algorithm is not registered.
+        """
+        if alg_id not in self._algorithms:
+            raise KeyError(
+                "The specified algorithm could not be removed"
+                " because it is not registered."
+            )
+
+        del self._algorithms[alg_id]
+        self._valid_algs.remove(alg_id)
+
+    def get_algorithms(self) -> list[str]:
+        """
+        Returns a list of supported values for the 'alg' parameter.
+        """
+        return list(self._valid_algs)
+
+    def get_algorithm_by_name(self, alg_name: str) -> Algorithm:
+        """
+        For a given string name, return the matching Algorithm object.
+
+        Example usage:
+
+        >>> jws_obj.get_algorithm_by_name("RS256")
+        """
+        try:
+            return self._algorithms[alg_name]
+        except KeyError as e:
+            if not has_crypto and alg_name in requires_cryptography:
+                raise NotImplementedError(
+                    f"Algorithm '{alg_name}' could not be found. Do you have cryptography installed?"
+                ) from e
+            raise NotImplementedError("Algorithm not supported") from e
+
+    def encode(
+        self,
+        payload: bytes,
+        key: AllowedPrivateKeys | PyJWK | str | bytes,
+        algorithm: str | None = None,
+        headers: dict[str, Any] | None = None,
+        json_encoder: type[json.JSONEncoder] | None = None,
+        is_payload_detached: bool = False,
+        sort_headers: bool = True,
+    ) -> str:
+        segments = []
+
+        # declare a new var to narrow the type for type checkers
+        if algorithm is None:
+            if isinstance(key, PyJWK):
+                algorithm_ = key.algorithm_name
+            else:
+                algorithm_ = "HS256"
+        else:
+            algorithm_ = algorithm
+
+        # Prefer headers values if present to function parameters.
+        if headers:
+            headers_alg = headers.get("alg")
+            if headers_alg:
+                algorithm_ = headers["alg"]
+
+            headers_b64 = headers.get("b64")
+            if headers_b64 is False:
+                is_payload_detached = True
+
+        # Header
+        header: dict[str, Any] = {"typ": self.header_typ, "alg": algorithm_}
+
+        if headers:
+            self._validate_headers(headers)
+            header.update(headers)
+
+        if not header["typ"]:
+            del header["typ"]
+
+        if is_payload_detached:
+            header["b64"] = False
+        elif "b64" in header:
+            # True is the standard value for b64, so no need for it
+            del header["b64"]
+
+        json_header = json.dumps(
+            header, separators=(",", ":"), cls=json_encoder, sort_keys=sort_headers
+        ).encode()
+
+        segments.append(base64url_encode(json_header))
+
+        if is_payload_detached:
+            msg_payload = payload
+        else:
+            msg_payload = base64url_encode(payload)
+        segments.append(msg_payload)
+
+        # Segments
+        signing_input = b".".join(segments)
+
+        alg_obj = self.get_algorithm_by_name(algorithm_)
+        if isinstance(key, PyJWK):
+            key = key.key
+        key = alg_obj.prepare_key(key)
+        signature = alg_obj.sign(signing_input, key)
+
+        segments.append(base64url_encode(signature))
+
+        # Don't put the payload content inside the encoded token when detached
+        if is_payload_detached:
+            segments[1] = b""
+        encoded_string = b".".join(segments)
+
+        return encoded_string.decode("utf-8")
+
+    def decode_complete(
+        self,
+        jwt: str | bytes,
+        key: AllowedPublicKeys | PyJWK | str | bytes = "",
+        algorithms: Sequence[str] | None = None,
+        options: dict[str, Any] | None = None,
+        detached_payload: bytes | None = None,
+        **kwargs,
+    ) -> dict[str, Any]:
+        if kwargs:
+            warnings.warn(
+                "passing additional kwargs to decode_complete() is deprecated "
+                "and will be removed in pyjwt version 3. "
+                f"Unsupported kwargs: {tuple(kwargs.keys())}",
+                RemovedInPyjwt3Warning,
+                stacklevel=2,
+            )
+        if options is None:
+            options = {}
+        merged_options = {**self.options, **options}
+        verify_signature = merged_options["verify_signature"]
+
+        if verify_signature and not algorithms and not isinstance(key, PyJWK):
+            raise DecodeError(
+                'It is required that you pass in a value for the "algorithms" argument when calling decode().'
+            )
+
+        payload, signing_input, header, signature = self._load(jwt)
+
+        if header.get("b64", True) is False:
+            if detached_payload is None:
+                raise DecodeError(
+                    'It is required that you pass in a value for the "detached_payload" argument to decode a message having the b64 header set to false.'
+                )
+            payload = detached_payload
+            signing_input = b".".join([signing_input.rsplit(b".", 1)[0], payload])
+
+        if verify_signature:
+            self._verify_signature(signing_input, header, signature, key, algorithms)
+
+        return {
+            "payload": payload,
+            "header": header,
+            "signature": signature,
+        }
+
+    def decode(
+        self,
+        jwt: str | bytes,
+        key: AllowedPublicKeys | PyJWK | str | bytes = "",
+        algorithms: Sequence[str] | None = None,
+        options: dict[str, Any] | None = None,
+        detached_payload: bytes | None = None,
+        **kwargs,
+    ) -> Any:
+        if kwargs:
+            warnings.warn(
+                "passing additional kwargs to decode() is deprecated "
+                "and will be removed in pyjwt version 3. "
+                f"Unsupported kwargs: {tuple(kwargs.keys())}",
+                RemovedInPyjwt3Warning,
+                stacklevel=2,
+            )
+        decoded = self.decode_complete(
+            jwt, key, algorithms, options, detached_payload=detached_payload
+        )
+        return decoded["payload"]
+
+    def get_unverified_header(self, jwt: str | bytes) -> dict[str, Any]:
+        """Returns back the JWT header parameters as a dict()
+
+        Note: The signature is not verified so the header parameters
+        should not be fully trusted until signature verification is complete
+        """
+        headers = self._load(jwt)[2]
+        self._validate_headers(headers)
+
+        return headers
+
+    def _load(self, jwt: str | bytes) -> tuple[bytes, bytes, dict[str, Any], bytes]:
+        if isinstance(jwt, str):
+            jwt = jwt.encode("utf-8")
+
+        if not isinstance(jwt, bytes):
+            raise DecodeError(f"Invalid token type. Token must be a {bytes}")
+
+        try:
+            signing_input, crypto_segment = jwt.rsplit(b".", 1)
+            header_segment, payload_segment = signing_input.split(b".", 1)
+        except ValueError as err:
+            raise DecodeError("Not enough segments") from err
+
+        try:
+            header_data = base64url_decode(header_segment)
+        except (TypeError, binascii.Error) as err:
+            raise DecodeError("Invalid header padding") from err
+
+        try:
+            header = json.loads(header_data)
+        except ValueError as e:
+            raise DecodeError(f"Invalid header string: {e}") from e
+
+        if not isinstance(header, dict):
+            raise DecodeError("Invalid header string: must be a json object")
+
+        try:
+            payload = base64url_decode(payload_segment)
+        except (TypeError, binascii.Error) as err:
+            raise DecodeError("Invalid payload padding") from err
+
+        try:
+            signature = base64url_decode(crypto_segment)
+        except (TypeError, binascii.Error) as err:
+            raise DecodeError("Invalid crypto padding") from err
+
+        return (payload, signing_input, header, signature)
+
+    def _verify_signature(
+        self,
+        signing_input: bytes,
+        header: dict[str, Any],
+        signature: bytes,
+        key: AllowedPublicKeys | PyJWK | str | bytes = "",
+        algorithms: Sequence[str] | None = None,
+    ) -> None:
+        if algorithms is None and isinstance(key, PyJWK):
+            algorithms = [key.algorithm_name]
+        try:
+            alg = header["alg"]
+        except KeyError:
+            raise InvalidAlgorithmError("Algorithm not specified") from None
+
+        if not alg or (algorithms is not None and alg not in algorithms):
+            raise InvalidAlgorithmError("The specified alg value is not allowed")
+
+        if isinstance(key, PyJWK):
+            alg_obj = key.Algorithm
+            prepared_key = key.key
+        else:
+            try:
+                alg_obj = self.get_algorithm_by_name(alg)
+            except NotImplementedError as e:
+                raise InvalidAlgorithmError("Algorithm not supported") from e
+            prepared_key = alg_obj.prepare_key(key)
+
+        if not alg_obj.verify(signing_input, prepared_key, signature):
+            raise InvalidSignatureError("Signature verification failed")
+
+    def _validate_headers(self, headers: dict[str, Any]) -> None:
+        if "kid" in headers:
+            self._validate_kid(headers["kid"])
+
+    def _validate_kid(self, kid: Any) -> None:
+        if not isinstance(kid, str):
+            raise InvalidTokenError("Key ID header parameter must be a string")
+
+
+_jws_global_obj = PyJWS()
+encode = _jws_global_obj.encode
+decode_complete = _jws_global_obj.decode_complete
+decode = _jws_global_obj.decode
+register_algorithm = _jws_global_obj.register_algorithm
+unregister_algorithm = _jws_global_obj.unregister_algorithm
+get_algorithm_by_name = _jws_global_obj.get_algorithm_by_name
+get_unverified_header = _jws_global_obj.get_unverified_header
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/api_jwt.py b/backend/venv/lib/python3.12/site-packages/jwt/api_jwt.py
new file mode 100644
index 00000000..3a201436
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/api_jwt.py
@@ -0,0 +1,433 @@
+from __future__ import annotations
+
+import json
+import warnings
+from calendar import timegm
+from collections.abc import Iterable, Sequence
+from datetime import datetime, timedelta, timezone
+from typing import TYPE_CHECKING, Any
+
+from . import api_jws
+from .exceptions import (
+    DecodeError,
+    ExpiredSignatureError,
+    ImmatureSignatureError,
+    InvalidAudienceError,
+    InvalidIssuedAtError,
+    InvalidIssuerError,
+    InvalidJTIError,
+    InvalidSubjectError,
+    MissingRequiredClaimError,
+)
+from .warnings import RemovedInPyjwt3Warning
+
+if TYPE_CHECKING:
+    from .algorithms import AllowedPrivateKeys, AllowedPublicKeys
+    from .api_jwk import PyJWK
+
+
+class PyJWT:
+    def __init__(self, options: dict[str, Any] | None = None) -> None:
+        if options is None:
+            options = {}
+        self.options: dict[str, Any] = {**self._get_default_options(), **options}
+
+    @staticmethod
+    def _get_default_options() -> dict[str, bool | list[str]]:
+        return {
+            "verify_signature": True,
+            "verify_exp": True,
+            "verify_nbf": True,
+            "verify_iat": True,
+            "verify_aud": True,
+            "verify_iss": True,
+            "verify_sub": True,
+            "verify_jti": True,
+            "require": [],
+        }
+
+    def encode(
+        self,
+        payload: dict[str, Any],
+        key: AllowedPrivateKeys | PyJWK | str | bytes,
+        algorithm: str | None = None,
+        headers: dict[str, Any] | None = None,
+        json_encoder: type[json.JSONEncoder] | None = None,
+        sort_headers: bool = True,
+    ) -> str:
+        # Check that we get a dict
+        if not isinstance(payload, dict):
+            raise TypeError(
+                "Expecting a dict object, as JWT only supports "
+                "JSON objects as payloads."
+            )
+
+        # Payload
+        payload = payload.copy()
+        for time_claim in ["exp", "iat", "nbf"]:
+            # Convert datetime to a intDate value in known time-format claims
+            if isinstance(payload.get(time_claim), datetime):
+                payload[time_claim] = timegm(payload[time_claim].utctimetuple())
+
+        json_payload = self._encode_payload(
+            payload,
+            headers=headers,
+            json_encoder=json_encoder,
+        )
+
+        return api_jws.encode(
+            json_payload,
+            key,
+            algorithm,
+            headers,
+            json_encoder,
+            sort_headers=sort_headers,
+        )
+
+    def _encode_payload(
+        self,
+        payload: dict[str, Any],
+        headers: dict[str, Any] | None = None,
+        json_encoder: type[json.JSONEncoder] | None = None,
+    ) -> bytes:
+        """
+        Encode a given payload to the bytes to be signed.
+
+        This method is intended to be overridden by subclasses that need to
+        encode the payload in a different way, e.g. compress the payload.
+        """
+        return json.dumps(
+            payload,
+            separators=(",", ":"),
+            cls=json_encoder,
+        ).encode("utf-8")
+
+    def decode_complete(
+        self,
+        jwt: str | bytes,
+        key: AllowedPublicKeys | PyJWK | str | bytes = "",
+        algorithms: Sequence[str] | None = None,
+        options: dict[str, Any] | None = None,
+        # deprecated arg, remove in pyjwt3
+        verify: bool | None = None,
+        # could be used as passthrough to api_jws, consider removal in pyjwt3
+        detached_payload: bytes | None = None,
+        # passthrough arguments to _validate_claims
+        # consider putting in options
+        audience: str | Iterable[str] | None = None,
+        issuer: str | Sequence[str] | None = None,
+        subject: str | None = None,
+        leeway: float | timedelta = 0,
+        # kwargs
+        **kwargs: Any,
+    ) -> dict[str, Any]:
+        if kwargs:
+            warnings.warn(
+                "passing additional kwargs to decode_complete() is deprecated "
+                "and will be removed in pyjwt version 3. "
+                f"Unsupported kwargs: {tuple(kwargs.keys())}",
+                RemovedInPyjwt3Warning,
+                stacklevel=2,
+            )
+        options = dict(options or {})  # shallow-copy or initialize an empty dict
+        options.setdefault("verify_signature", True)
+
+        # If the user has set the legacy `verify` argument, and it doesn't match
+        # what the relevant `options` entry for the argument is, inform the user
+        # that they're likely making a mistake.
+        if verify is not None and verify != options["verify_signature"]:
+            warnings.warn(
+                "The `verify` argument to `decode` does nothing in PyJWT 2.0 and newer. "
+                "The equivalent is setting `verify_signature` to False in the `options` dictionary. "
+                "This invocation has a mismatch between the kwarg and the option entry.",
+                category=DeprecationWarning,
+                stacklevel=2,
+            )
+
+        if not options["verify_signature"]:
+            options.setdefault("verify_exp", False)
+            options.setdefault("verify_nbf", False)
+            options.setdefault("verify_iat", False)
+            options.setdefault("verify_aud", False)
+            options.setdefault("verify_iss", False)
+            options.setdefault("verify_sub", False)
+            options.setdefault("verify_jti", False)
+
+        decoded = api_jws.decode_complete(
+            jwt,
+            key=key,
+            algorithms=algorithms,
+            options=options,
+            detached_payload=detached_payload,
+        )
+
+        payload = self._decode_payload(decoded)
+
+        merged_options = {**self.options, **options}
+        self._validate_claims(
+            payload,
+            merged_options,
+            audience=audience,
+            issuer=issuer,
+            leeway=leeway,
+            subject=subject,
+        )
+
+        decoded["payload"] = payload
+        return decoded
+
+    def _decode_payload(self, decoded: dict[str, Any]) -> Any:
+        """
+        Decode the payload from a JWS dictionary (payload, signature, header).
+
+        This method is intended to be overridden by subclasses that need to
+        decode the payload in a different way, e.g. decompress compressed
+        payloads.
+        """
+        try:
+            payload = json.loads(decoded["payload"])
+        except ValueError as e:
+            raise DecodeError(f"Invalid payload string: {e}") from e
+        if not isinstance(payload, dict):
+            raise DecodeError("Invalid payload string: must be a json object")
+        return payload
+
+    def decode(
+        self,
+        jwt: str | bytes,
+        key: AllowedPublicKeys | PyJWK | str | bytes = "",
+        algorithms: Sequence[str] | None = None,
+        options: dict[str, Any] | None = None,
+        # deprecated arg, remove in pyjwt3
+        verify: bool | None = None,
+        # could be used as passthrough to api_jws, consider removal in pyjwt3
+        detached_payload: bytes | None = None,
+        # passthrough arguments to _validate_claims
+        # consider putting in options
+        audience: str | Iterable[str] | None = None,
+        subject: str | None = None,
+        issuer: str | Sequence[str] | None = None,
+        leeway: float | timedelta = 0,
+        # kwargs
+        **kwargs: Any,
+    ) -> Any:
+        if kwargs:
+            warnings.warn(
+                "passing additional kwargs to decode() is deprecated "
+                "and will be removed in pyjwt version 3. "
+                f"Unsupported kwargs: {tuple(kwargs.keys())}",
+                RemovedInPyjwt3Warning,
+                stacklevel=2,
+            )
+        decoded = self.decode_complete(
+            jwt,
+            key,
+            algorithms,
+            options,
+            verify=verify,
+            detached_payload=detached_payload,
+            audience=audience,
+            subject=subject,
+            issuer=issuer,
+            leeway=leeway,
+        )
+        return decoded["payload"]
+
+    def _validate_claims(
+        self,
+        payload: dict[str, Any],
+        options: dict[str, Any],
+        audience=None,
+        issuer=None,
+        subject: str | None = None,
+        leeway: float | timedelta = 0,
+    ) -> None:
+        if isinstance(leeway, timedelta):
+            leeway = leeway.total_seconds()
+
+        if audience is not None and not isinstance(audience, (str, Iterable)):
+            raise TypeError("audience must be a string, iterable or None")
+
+        self._validate_required_claims(payload, options)
+
+        now = datetime.now(tz=timezone.utc).timestamp()
+
+        if "iat" in payload and options["verify_iat"]:
+            self._validate_iat(payload, now, leeway)
+
+        if "nbf" in payload and options["verify_nbf"]:
+            self._validate_nbf(payload, now, leeway)
+
+        if "exp" in payload and options["verify_exp"]:
+            self._validate_exp(payload, now, leeway)
+
+        if options["verify_iss"]:
+            self._validate_iss(payload, issuer)
+
+        if options["verify_aud"]:
+            self._validate_aud(
+                payload, audience, strict=options.get("strict_aud", False)
+            )
+
+        if options["verify_sub"]:
+            self._validate_sub(payload, subject)
+
+        if options["verify_jti"]:
+            self._validate_jti(payload)
+
+    def _validate_required_claims(
+        self,
+        payload: dict[str, Any],
+        options: dict[str, Any],
+    ) -> None:
+        for claim in options["require"]:
+            if payload.get(claim) is None:
+                raise MissingRequiredClaimError(claim)
+
+    def _validate_sub(self, payload: dict[str, Any], subject=None) -> None:
+        """
+        Checks whether "sub" if in the payload is valid ot not.
+        This is an Optional claim
+
+        :param payload(dict): The payload which needs to be validated
+        :param subject(str): The subject of the token
+        """
+
+        if "sub" not in payload:
+            return
+
+        if not isinstance(payload["sub"], str):
+            raise InvalidSubjectError("Subject must be a string")
+
+        if subject is not None:
+            if payload.get("sub") != subject:
+                raise InvalidSubjectError("Invalid subject")
+
+    def _validate_jti(self, payload: dict[str, Any]) -> None:
+        """
+        Checks whether "jti" if in the payload is valid ot not
+        This is an Optional claim
+
+        :param payload(dict): The payload which needs to be validated
+        """
+
+        if "jti" not in payload:
+            return
+
+        if not isinstance(payload.get("jti"), str):
+            raise InvalidJTIError("JWT ID must be a string")
+
+    def _validate_iat(
+        self,
+        payload: dict[str, Any],
+        now: float,
+        leeway: float,
+    ) -> None:
+        try:
+            iat = int(payload["iat"])
+        except ValueError:
+            raise InvalidIssuedAtError(
+                "Issued At claim (iat) must be an integer."
+            ) from None
+        if iat > (now + leeway):
+            raise ImmatureSignatureError("The token is not yet valid (iat)")
+
+    def _validate_nbf(
+        self,
+        payload: dict[str, Any],
+        now: float,
+        leeway: float,
+    ) -> None:
+        try:
+            nbf = int(payload["nbf"])
+        except ValueError:
+            raise DecodeError("Not Before claim (nbf) must be an integer.") from None
+
+        if nbf > (now + leeway):
+            raise ImmatureSignatureError("The token is not yet valid (nbf)")
+
+    def _validate_exp(
+        self,
+        payload: dict[str, Any],
+        now: float,
+        leeway: float,
+    ) -> None:
+        try:
+            exp = int(payload["exp"])
+        except ValueError:
+            raise DecodeError(
+                "Expiration Time claim (exp) must be an integer."
+            ) from None
+
+        if exp <= (now - leeway):
+            raise ExpiredSignatureError("Signature has expired")
+
+    def _validate_aud(
+        self,
+        payload: dict[str, Any],
+        audience: str | Iterable[str] | None,
+        *,
+        strict: bool = False,
+    ) -> None:
+        if audience is None:
+            if "aud" not in payload or not payload["aud"]:
+                return
+            # Application did not specify an audience, but
+            # the token has the 'aud' claim
+            raise InvalidAudienceError("Invalid audience")
+
+        if "aud" not in payload or not payload["aud"]:
+            # Application specified an audience, but it could not be
+            # verified since the token does not contain a claim.
+            raise MissingRequiredClaimError("aud")
+
+        audience_claims = payload["aud"]
+
+        # In strict mode, we forbid list matching: the supplied audience
+        # must be a string, and it must exactly match the audience claim.
+        if strict:
+            # Only a single audience is allowed in strict mode.
+            if not isinstance(audience, str):
+                raise InvalidAudienceError("Invalid audience (strict)")
+
+            # Only a single audience claim is allowed in strict mode.
+            if not isinstance(audience_claims, str):
+                raise InvalidAudienceError("Invalid claim format in token (strict)")
+
+            if audience != audience_claims:
+                raise InvalidAudienceError("Audience doesn't match (strict)")
+
+            return
+
+        if isinstance(audience_claims, str):
+            audience_claims = [audience_claims]
+        if not isinstance(audience_claims, list):
+            raise InvalidAudienceError("Invalid claim format in token")
+        if any(not isinstance(c, str) for c in audience_claims):
+            raise InvalidAudienceError("Invalid claim format in token")
+
+        if isinstance(audience, str):
+            audience = [audience]
+
+        if all(aud not in audience_claims for aud in audience):
+            raise InvalidAudienceError("Audience doesn't match")
+
+    def _validate_iss(self, payload: dict[str, Any], issuer: Any) -> None:
+        if issuer is None:
+            return
+
+        if "iss" not in payload:
+            raise MissingRequiredClaimError("iss")
+
+        if isinstance(issuer, str):
+            if payload["iss"] != issuer:
+                raise InvalidIssuerError("Invalid issuer")
+        else:
+            if payload["iss"] not in issuer:
+                raise InvalidIssuerError("Invalid issuer")
+
+
+_jwt_global_obj = PyJWT()
+encode = _jwt_global_obj.encode
+decode_complete = _jwt_global_obj.decode_complete
+decode = _jwt_global_obj.decode
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/exceptions.py b/backend/venv/lib/python3.12/site-packages/jwt/exceptions.py
new file mode 100644
index 00000000..9b45ae48
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/exceptions.py
@@ -0,0 +1,82 @@
+class PyJWTError(Exception):
+    """
+    Base class for all exceptions
+    """
+
+    pass
+
+
+class InvalidTokenError(PyJWTError):
+    pass
+
+
+class DecodeError(InvalidTokenError):
+    pass
+
+
+class InvalidSignatureError(DecodeError):
+    pass
+
+
+class ExpiredSignatureError(InvalidTokenError):
+    pass
+
+
+class InvalidAudienceError(InvalidTokenError):
+    pass
+
+
+class InvalidIssuerError(InvalidTokenError):
+    pass
+
+
+class InvalidIssuedAtError(InvalidTokenError):
+    pass
+
+
+class ImmatureSignatureError(InvalidTokenError):
+    pass
+
+
+class InvalidKeyError(PyJWTError):
+    pass
+
+
+class InvalidAlgorithmError(InvalidTokenError):
+    pass
+
+
+class MissingRequiredClaimError(InvalidTokenError):
+    def __init__(self, claim: str) -> None:
+        self.claim = claim
+
+    def __str__(self) -> str:
+        return f'Token is missing the "{self.claim}" claim'
+
+
+class PyJWKError(PyJWTError):
+    pass
+
+
+class MissingCryptographyError(PyJWKError):
+    pass
+
+
+class PyJWKSetError(PyJWTError):
+    pass
+
+
+class PyJWKClientError(PyJWTError):
+    pass
+
+
+class PyJWKClientConnectionError(PyJWKClientError):
+    pass
+
+
+class InvalidSubjectError(InvalidTokenError):
+    pass
+
+
+class InvalidJTIError(InvalidTokenError):
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/help.py b/backend/venv/lib/python3.12/site-packages/jwt/help.py
new file mode 100644
index 00000000..8e1c2286
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/help.py
@@ -0,0 +1,67 @@
+import json
+import platform
+import sys
+from typing import Dict
+
+from . import __version__ as pyjwt_version
+
+try:
+    import cryptography
+
+    cryptography_version = cryptography.__version__
+except ModuleNotFoundError:
+    cryptography_version = ""
+
+
+def info() -> Dict[str, Dict[str, str]]:
+    """
+    Generate information for a bug report.
+    Based on the requests package help utility module.
+    """
+    try:
+        platform_info = {
+            "system": platform.system(),
+            "release": platform.release(),
+        }
+    except OSError:
+        platform_info = {"system": "Unknown", "release": "Unknown"}
+
+    implementation = platform.python_implementation()
+
+    if implementation == "CPython":
+        implementation_version = platform.python_version()
+    elif implementation == "PyPy":
+        pypy_version_info = sys.pypy_version_info  # type: ignore[attr-defined]
+        implementation_version = (
+            f"{pypy_version_info.major}."
+            f"{pypy_version_info.minor}."
+            f"{pypy_version_info.micro}"
+        )
+        if pypy_version_info.releaselevel != "final":
+            implementation_version = "".join(
+                [
+                    implementation_version,
+                    pypy_version_info.releaselevel,
+                ]
+            )
+    else:
+        implementation_version = "Unknown"
+
+    return {
+        "platform": platform_info,
+        "implementation": {
+            "name": implementation,
+            "version": implementation_version,
+        },
+        "cryptography": {"version": cryptography_version},
+        "pyjwt": {"version": pyjwt_version},
+    }
+
+
+def main() -> None:
+    """Pretty-print the bug information as JSON."""
+    print(json.dumps(info(), sort_keys=True, indent=2))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/jwk_set_cache.py b/backend/venv/lib/python3.12/site-packages/jwt/jwk_set_cache.py
new file mode 100644
index 00000000..24325630
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/jwk_set_cache.py
@@ -0,0 +1,31 @@
+import time
+from typing import Optional
+
+from .api_jwk import PyJWKSet, PyJWTSetWithTimestamp
+
+
+class JWKSetCache:
+    def __init__(self, lifespan: int) -> None:
+        self.jwk_set_with_timestamp: Optional[PyJWTSetWithTimestamp] = None
+        self.lifespan = lifespan
+
+    def put(self, jwk_set: PyJWKSet) -> None:
+        if jwk_set is not None:
+            self.jwk_set_with_timestamp = PyJWTSetWithTimestamp(jwk_set)
+        else:
+            # clear cache
+            self.jwk_set_with_timestamp = None
+
+    def get(self) -> Optional[PyJWKSet]:
+        if self.jwk_set_with_timestamp is None or self.is_expired():
+            return None
+
+        return self.jwk_set_with_timestamp.get_jwk_set()
+
+    def is_expired(self) -> bool:
+        return (
+            self.jwk_set_with_timestamp is not None
+            and self.lifespan > -1
+            and time.monotonic()
+            > self.jwk_set_with_timestamp.get_timestamp() + self.lifespan
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/jwks_client.py b/backend/venv/lib/python3.12/site-packages/jwt/jwks_client.py
new file mode 100644
index 00000000..9a8992ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/jwks_client.py
@@ -0,0 +1,126 @@
+import json
+import urllib.request
+from functools import lru_cache
+from ssl import SSLContext
+from typing import Any, Dict, List, Optional
+from urllib.error import URLError
+
+from .api_jwk import PyJWK, PyJWKSet
+from .api_jwt import decode_complete as decode_token
+from .exceptions import PyJWKClientConnectionError, PyJWKClientError
+from .jwk_set_cache import JWKSetCache
+
+
+class PyJWKClient:
+    def __init__(
+        self,
+        uri: str,
+        cache_keys: bool = False,
+        max_cached_keys: int = 16,
+        cache_jwk_set: bool = True,
+        lifespan: int = 300,
+        headers: Optional[Dict[str, Any]] = None,
+        timeout: int = 30,
+        ssl_context: Optional[SSLContext] = None,
+    ):
+        if headers is None:
+            headers = {}
+        self.uri = uri
+        self.jwk_set_cache: Optional[JWKSetCache] = None
+        self.headers = headers
+        self.timeout = timeout
+        self.ssl_context = ssl_context
+
+        if cache_jwk_set:
+            # Init jwt set cache with default or given lifespan.
+            # Default lifespan is 300 seconds (5 minutes).
+            if lifespan <= 0:
+                raise PyJWKClientError(
+                    f'Lifespan must be greater than 0, the input is "{lifespan}"'
+                )
+            self.jwk_set_cache = JWKSetCache(lifespan)
+        else:
+            self.jwk_set_cache = None
+
+        if cache_keys:
+            # Cache signing keys
+            # Ignore mypy (https://github.com/python/mypy/issues/2427)
+            self.get_signing_key = lru_cache(maxsize=max_cached_keys)(
+                self.get_signing_key
+            )  # type: ignore
+
+    def fetch_data(self) -> Any:
+        jwk_set: Any = None
+        try:
+            r = urllib.request.Request(url=self.uri, headers=self.headers)
+            with urllib.request.urlopen(
+                r, timeout=self.timeout, context=self.ssl_context
+            ) as response:
+                jwk_set = json.load(response)
+        except (URLError, TimeoutError) as e:
+            raise PyJWKClientConnectionError(
+                f'Fail to fetch data from the url, err: "{e}"'
+            ) from e
+        else:
+            return jwk_set
+        finally:
+            if self.jwk_set_cache is not None:
+                self.jwk_set_cache.put(jwk_set)
+
+    def get_jwk_set(self, refresh: bool = False) -> PyJWKSet:
+        data = None
+        if self.jwk_set_cache is not None and not refresh:
+            data = self.jwk_set_cache.get()
+
+        if data is None:
+            data = self.fetch_data()
+
+        if not isinstance(data, dict):
+            raise PyJWKClientError("The JWKS endpoint did not return a JSON object")
+
+        return PyJWKSet.from_dict(data)
+
+    def get_signing_keys(self, refresh: bool = False) -> List[PyJWK]:
+        jwk_set = self.get_jwk_set(refresh)
+        signing_keys = [
+            jwk_set_key
+            for jwk_set_key in jwk_set.keys
+            if jwk_set_key.public_key_use in ["sig", None] and jwk_set_key.key_id
+        ]
+
+        if not signing_keys:
+            raise PyJWKClientError("The JWKS endpoint did not contain any signing keys")
+
+        return signing_keys
+
+    def get_signing_key(self, kid: str) -> PyJWK:
+        signing_keys = self.get_signing_keys()
+        signing_key = self.match_kid(signing_keys, kid)
+
+        if not signing_key:
+            # If no matching signing key from the jwk set, refresh the jwk set and try again.
+            signing_keys = self.get_signing_keys(refresh=True)
+            signing_key = self.match_kid(signing_keys, kid)
+
+            if not signing_key:
+                raise PyJWKClientError(
+                    f'Unable to find a signing key that matches: "{kid}"'
+                )
+
+        return signing_key
+
+    def get_signing_key_from_jwt(self, token: str) -> PyJWK:
+        unverified = decode_token(token, options={"verify_signature": False})
+        header = unverified["header"]
+        return self.get_signing_key(header.get("kid"))
+
+    @staticmethod
+    def match_kid(signing_keys: List[PyJWK], kid: str) -> Optional[PyJWK]:
+        signing_key = None
+
+        for key in signing_keys:
+            if key.key_id == kid:
+                signing_key = key
+                break
+
+        return signing_key
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/py.typed b/backend/venv/lib/python3.12/site-packages/jwt/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/types.py b/backend/venv/lib/python3.12/site-packages/jwt/types.py
new file mode 100644
index 00000000..7d993520
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/types.py
@@ -0,0 +1,5 @@
+from typing import Any, Callable, Dict
+
+JWKDict = Dict[str, Any]
+
+HashlibHash = Callable[..., Any]
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/utils.py b/backend/venv/lib/python3.12/site-packages/jwt/utils.py
new file mode 100644
index 00000000..56e89bb7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/utils.py
@@ -0,0 +1,142 @@
+import base64
+import binascii
+import re
+from typing import Optional, Union
+
+try:
+    from cryptography.hazmat.primitives.asymmetric.ec import EllipticCurve
+    from cryptography.hazmat.primitives.asymmetric.utils import (
+        decode_dss_signature,
+        encode_dss_signature,
+    )
+except ModuleNotFoundError:
+    pass
+
+
+def force_bytes(value: Union[bytes, str]) -> bytes:
+    if isinstance(value, str):
+        return value.encode("utf-8")
+    elif isinstance(value, bytes):
+        return value
+    else:
+        raise TypeError("Expected a string value")
+
+
+def base64url_decode(input: Union[bytes, str]) -> bytes:
+    input_bytes = force_bytes(input)
+
+    rem = len(input_bytes) % 4
+
+    if rem > 0:
+        input_bytes += b"=" * (4 - rem)
+
+    return base64.urlsafe_b64decode(input_bytes)
+
+
+def base64url_encode(input: bytes) -> bytes:
+    return base64.urlsafe_b64encode(input).replace(b"=", b"")
+
+
+def to_base64url_uint(val: int, *, bit_length: Optional[int] = None) -> bytes:
+    if val < 0:
+        raise ValueError("Must be a positive integer")
+
+    int_bytes = bytes_from_int(val, bit_length=bit_length)
+
+    if len(int_bytes) == 0:
+        int_bytes = b"\x00"
+
+    return base64url_encode(int_bytes)
+
+
+def from_base64url_uint(val: Union[bytes, str]) -> int:
+    data = base64url_decode(force_bytes(val))
+    return int.from_bytes(data, byteorder="big")
+
+
+def number_to_bytes(num: int, num_bytes: int) -> bytes:
+    padded_hex = "%0*x" % (2 * num_bytes, num)
+    return binascii.a2b_hex(padded_hex.encode("ascii"))
+
+
+def bytes_to_number(string: bytes) -> int:
+    return int(binascii.b2a_hex(string), 16)
+
+
+def bytes_from_int(val: int, *, bit_length: Optional[int] = None) -> bytes:
+    if bit_length is None:
+        bit_length = val.bit_length()
+    byte_length = (bit_length + 7) // 8
+
+    return val.to_bytes(byte_length, "big", signed=False)
+
+
+def der_to_raw_signature(der_sig: bytes, curve: "EllipticCurve") -> bytes:
+    num_bits = curve.key_size
+    num_bytes = (num_bits + 7) // 8
+
+    r, s = decode_dss_signature(der_sig)
+
+    return number_to_bytes(r, num_bytes) + number_to_bytes(s, num_bytes)
+
+
+def raw_to_der_signature(raw_sig: bytes, curve: "EllipticCurve") -> bytes:
+    num_bits = curve.key_size
+    num_bytes = (num_bits + 7) // 8
+
+    if len(raw_sig) != 2 * num_bytes:
+        raise ValueError("Invalid signature")
+
+    r = bytes_to_number(raw_sig[:num_bytes])
+    s = bytes_to_number(raw_sig[num_bytes:])
+
+    return bytes(encode_dss_signature(r, s))
+
+
+# Based on https://github.com/hynek/pem/blob/7ad94db26b0bc21d10953f5dbad3acfdfacf57aa/src/pem/_core.py#L224-L252
+_PEMS = {
+    b"CERTIFICATE",
+    b"TRUSTED CERTIFICATE",
+    b"PRIVATE KEY",
+    b"PUBLIC KEY",
+    b"ENCRYPTED PRIVATE KEY",
+    b"OPENSSH PRIVATE KEY",
+    b"DSA PRIVATE KEY",
+    b"RSA PRIVATE KEY",
+    b"RSA PUBLIC KEY",
+    b"EC PRIVATE KEY",
+    b"DH PARAMETERS",
+    b"NEW CERTIFICATE REQUEST",
+    b"CERTIFICATE REQUEST",
+    b"SSH2 PUBLIC KEY",
+    b"SSH2 ENCRYPTED PRIVATE KEY",
+    b"X509 CRL",
+}
+
+_PEM_RE = re.compile(
+    b"----[- ]BEGIN ("
+    + b"|".join(_PEMS)
+    + b""")[- ]----\r?
+.+?\r?
+----[- ]END \\1[- ]----\r?\n?""",
+    re.DOTALL,
+)
+
+
+def is_pem_format(key: bytes) -> bool:
+    return bool(_PEM_RE.search(key))
+
+
+# Based on https://github.com/pyca/cryptography/blob/bcb70852d577b3f490f015378c75cba74986297b/src/cryptography/hazmat/primitives/serialization/ssh.py#L40-L46
+_SSH_KEY_FORMATS = (
+    b"ssh-ed25519",
+    b"ssh-rsa",
+    b"ssh-dss",
+    b"ecdsa-sha2-nistp256",
+    b"ecdsa-sha2-nistp384",
+    b"ecdsa-sha2-nistp521",
+)
+
+
+def is_ssh_key(key: bytes) -> bool:
+    return key.startswith(_SSH_KEY_FORMATS)
diff --git a/backend/venv/lib/python3.12/site-packages/jwt/warnings.py b/backend/venv/lib/python3.12/site-packages/jwt/warnings.py
new file mode 100644
index 00000000..8762a8cb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/jwt/warnings.py
@@ -0,0 +1,2 @@
+class RemovedInPyjwt3Warning(DeprecationWarning):
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__init__.py b/backend/venv/lib/python3.12/site-packages/mako/__init__.py
new file mode 100644
index 00000000..989ddebd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/__init__.py
@@ -0,0 +1,8 @@
+# mako/__init__.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+
+__version__ = "1.3.9"
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..71b1c930
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/_ast_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/_ast_util.cpython-312.pyc
new file mode 100644
index 00000000..5c351b60
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/_ast_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/ast.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/ast.cpython-312.pyc
new file mode 100644
index 00000000..8f453cd9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/ast.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cache.cpython-312.pyc
new file mode 100644
index 00000000..44ced348
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cmd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cmd.cpython-312.pyc
new file mode 100644
index 00000000..f94f087a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/cmd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/codegen.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/codegen.cpython-312.pyc
new file mode 100644
index 00000000..b738146e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/codegen.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..4f763761
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..806e2f13
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/filters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/filters.cpython-312.pyc
new file mode 100644
index 00000000..af829614
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/filters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lexer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lexer.cpython-312.pyc
new file mode 100644
index 00000000..7e80737d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lexer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lookup.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lookup.cpython-312.pyc
new file mode 100644
index 00000000..5aa8797f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/lookup.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/parsetree.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/parsetree.cpython-312.pyc
new file mode 100644
index 00000000..ea0102d5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/parsetree.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pygen.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pygen.cpython-312.pyc
new file mode 100644
index 00000000..e3c94fd6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pygen.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pyparser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pyparser.cpython-312.pyc
new file mode 100644
index 00000000..1b567df0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/pyparser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/runtime.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/runtime.cpython-312.pyc
new file mode 100644
index 00000000..c8ccef61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/runtime.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/template.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/template.cpython-312.pyc
new file mode 100644
index 00000000..1928db18
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/template.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..8dba9809
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/_ast_util.py b/backend/venv/lib/python3.12/site-packages/mako/_ast_util.py
new file mode 100644
index 00000000..385d7a00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/_ast_util.py
@@ -0,0 +1,713 @@
+# mako/_ast_util.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""
+    ast
+    ~~~
+
+    This is a stripped down version of Armin Ronacher's ast module.
+
+    :copyright: Copyright 2008 by Armin Ronacher.
+    :license: Python License.
+"""
+
+
+from _ast import Add
+from _ast import And
+from _ast import AST
+from _ast import BitAnd
+from _ast import BitOr
+from _ast import BitXor
+from _ast import Div
+from _ast import Eq
+from _ast import FloorDiv
+from _ast import Gt
+from _ast import GtE
+from _ast import If
+from _ast import In
+from _ast import Invert
+from _ast import Is
+from _ast import IsNot
+from _ast import LShift
+from _ast import Lt
+from _ast import LtE
+from _ast import Mod
+from _ast import Mult
+from _ast import Name
+from _ast import Not
+from _ast import NotEq
+from _ast import NotIn
+from _ast import Or
+from _ast import PyCF_ONLY_AST
+from _ast import RShift
+from _ast import Sub
+from _ast import UAdd
+from _ast import USub
+
+
+BOOLOP_SYMBOLS = {And: "and", Or: "or"}
+
+BINOP_SYMBOLS = {
+    Add: "+",
+    Sub: "-",
+    Mult: "*",
+    Div: "/",
+    FloorDiv: "//",
+    Mod: "%",
+    LShift: "<<",
+    RShift: ">>",
+    BitOr: "|",
+    BitAnd: "&",
+    BitXor: "^",
+}
+
+CMPOP_SYMBOLS = {
+    Eq: "==",
+    Gt: ">",
+    GtE: ">=",
+    In: "in",
+    Is: "is",
+    IsNot: "is not",
+    Lt: "<",
+    LtE: "<=",
+    NotEq: "!=",
+    NotIn: "not in",
+}
+
+UNARYOP_SYMBOLS = {Invert: "~", Not: "not", UAdd: "+", USub: "-"}
+
+ALL_SYMBOLS = {}
+ALL_SYMBOLS.update(BOOLOP_SYMBOLS)
+ALL_SYMBOLS.update(BINOP_SYMBOLS)
+ALL_SYMBOLS.update(CMPOP_SYMBOLS)
+ALL_SYMBOLS.update(UNARYOP_SYMBOLS)
+
+
+def parse(expr, filename="<unknown>", mode="exec"):
+    """Parse an expression into an AST node."""
+    return compile(expr, filename, mode, PyCF_ONLY_AST)
+
+
+def iter_fields(node):
+    """Iterate over all fields of a node, only yielding existing fields."""
+
+    for field in node._fields:
+        try:
+            yield field, getattr(node, field)
+        except AttributeError:
+            pass
+
+
+class NodeVisitor:
+
+    """
+    Walks the abstract syntax tree and call visitor functions for every node
+    found.  The visitor functions may return values which will be forwarded
+    by the `visit` method.
+
+    Per default the visitor functions for the nodes are ``'visit_'`` +
+    class name of the node.  So a `TryFinally` node visit function would
+    be `visit_TryFinally`.  This behavior can be changed by overriding
+    the `get_visitor` function.  If no visitor function exists for a node
+    (return value `None`) the `generic_visit` visitor is used instead.
+
+    Don't use the `NodeVisitor` if you want to apply changes to nodes during
+    traversing.  For this a special visitor exists (`NodeTransformer`) that
+    allows modifications.
+    """
+
+    def get_visitor(self, node):
+        """
+        Return the visitor function for this node or `None` if no visitor
+        exists for this node.  In that case the generic visit function is
+        used instead.
+        """
+        method = "visit_" + node.__class__.__name__
+        return getattr(self, method, None)
+
+    def visit(self, node):
+        """Visit a node."""
+        f = self.get_visitor(node)
+        if f is not None:
+            return f(node)
+        return self.generic_visit(node)
+
+    def generic_visit(self, node):
+        """Called if no explicit visitor function exists for a node."""
+        for field, value in iter_fields(node):
+            if isinstance(value, list):
+                for item in value:
+                    if isinstance(item, AST):
+                        self.visit(item)
+            elif isinstance(value, AST):
+                self.visit(value)
+
+
+class NodeTransformer(NodeVisitor):
+
+    """
+    Walks the abstract syntax tree and allows modifications of nodes.
+
+    The `NodeTransformer` will walk the AST and use the return value of the
+    visitor functions to replace or remove the old node.  If the return
+    value of the visitor function is `None` the node will be removed
+    from the previous location otherwise it's replaced with the return
+    value.  The return value may be the original node in which case no
+    replacement takes place.
+
+    Here an example transformer that rewrites all `foo` to `data['foo']`::
+
+        class RewriteName(NodeTransformer):
+
+            def visit_Name(self, node):
+                return copy_location(Subscript(
+                    value=Name(id='data', ctx=Load()),
+                    slice=Index(value=Str(s=node.id)),
+                    ctx=node.ctx
+                ), node)
+
+    Keep in mind that if the node you're operating on has child nodes
+    you must either transform the child nodes yourself or call the generic
+    visit function for the node first.
+
+    Nodes that were part of a collection of statements (that applies to
+    all statement nodes) may also return a list of nodes rather than just
+    a single node.
+
+    Usually you use the transformer like this::
+
+        node = YourTransformer().visit(node)
+    """
+
+    def generic_visit(self, node):
+        for field, old_value in iter_fields(node):
+            old_value = getattr(node, field, None)
+            if isinstance(old_value, list):
+                new_values = []
+                for value in old_value:
+                    if isinstance(value, AST):
+                        value = self.visit(value)
+                        if value is None:
+                            continue
+                        elif not isinstance(value, AST):
+                            new_values.extend(value)
+                            continue
+                    new_values.append(value)
+                old_value[:] = new_values
+            elif isinstance(old_value, AST):
+                new_node = self.visit(old_value)
+                if new_node is None:
+                    delattr(node, field)
+                else:
+                    setattr(node, field, new_node)
+        return node
+
+
+class SourceGenerator(NodeVisitor):
+
+    """
+    This visitor is able to transform a well formed syntax tree into python
+    sourcecode.  For more details have a look at the docstring of the
+    `node_to_source` function.
+    """
+
+    def __init__(self, indent_with):
+        self.result = []
+        self.indent_with = indent_with
+        self.indentation = 0
+        self.new_lines = 0
+
+    def write(self, x):
+        if self.new_lines:
+            if self.result:
+                self.result.append("\n" * self.new_lines)
+            self.result.append(self.indent_with * self.indentation)
+            self.new_lines = 0
+        self.result.append(x)
+
+    def newline(self, n=1):
+        self.new_lines = max(self.new_lines, n)
+
+    def body(self, statements):
+        self.new_line = True
+        self.indentation += 1
+        for stmt in statements:
+            self.visit(stmt)
+        self.indentation -= 1
+
+    def body_or_else(self, node):
+        self.body(node.body)
+        if node.orelse:
+            self.newline()
+            self.write("else:")
+            self.body(node.orelse)
+
+    def signature(self, node):
+        want_comma = []
+
+        def write_comma():
+            if want_comma:
+                self.write(", ")
+            else:
+                want_comma.append(True)
+
+        padding = [None] * (len(node.args) - len(node.defaults))
+        for arg, default in zip(node.args, padding + node.defaults):
+            write_comma()
+            self.visit(arg)
+            if default is not None:
+                self.write("=")
+                self.visit(default)
+        if node.vararg is not None:
+            write_comma()
+            self.write("*" + node.vararg.arg)
+        if node.kwarg is not None:
+            write_comma()
+            self.write("**" + node.kwarg.arg)
+
+    def decorators(self, node):
+        for decorator in node.decorator_list:
+            self.newline()
+            self.write("@")
+            self.visit(decorator)
+
+    # Statements
+
+    def visit_Assign(self, node):
+        self.newline()
+        for idx, target in enumerate(node.targets):
+            if idx:
+                self.write(", ")
+            self.visit(target)
+        self.write(" = ")
+        self.visit(node.value)
+
+    def visit_AugAssign(self, node):
+        self.newline()
+        self.visit(node.target)
+        self.write(BINOP_SYMBOLS[type(node.op)] + "=")
+        self.visit(node.value)
+
+    def visit_ImportFrom(self, node):
+        self.newline()
+        self.write("from %s%s import " % ("." * node.level, node.module))
+        for idx, item in enumerate(node.names):
+            if idx:
+                self.write(", ")
+            self.write(item)
+
+    def visit_Import(self, node):
+        self.newline()
+        for item in node.names:
+            self.write("import ")
+            self.visit(item)
+
+    def visit_Expr(self, node):
+        self.newline()
+        self.generic_visit(node)
+
+    def visit_FunctionDef(self, node):
+        self.newline(n=2)
+        self.decorators(node)
+        self.newline()
+        self.write("def %s(" % node.name)
+        self.signature(node.args)
+        self.write("):")
+        self.body(node.body)
+
+    def visit_ClassDef(self, node):
+        have_args = []
+
+        def paren_or_comma():
+            if have_args:
+                self.write(", ")
+            else:
+                have_args.append(True)
+                self.write("(")
+
+        self.newline(n=3)
+        self.decorators(node)
+        self.newline()
+        self.write("class %s" % node.name)
+        for base in node.bases:
+            paren_or_comma()
+            self.visit(base)
+        # XXX: the if here is used to keep this module compatible
+        #      with python 2.6.
+        if hasattr(node, "keywords"):
+            for keyword in node.keywords:
+                paren_or_comma()
+                self.write(keyword.arg + "=")
+                self.visit(keyword.value)
+            if getattr(node, "starargs", None):
+                paren_or_comma()
+                self.write("*")
+                self.visit(node.starargs)
+            if getattr(node, "kwargs", None):
+                paren_or_comma()
+                self.write("**")
+                self.visit(node.kwargs)
+        self.write(have_args and "):" or ":")
+        self.body(node.body)
+
+    def visit_If(self, node):
+        self.newline()
+        self.write("if ")
+        self.visit(node.test)
+        self.write(":")
+        self.body(node.body)
+        while True:
+            else_ = node.orelse
+            if len(else_) == 1 and isinstance(else_[0], If):
+                node = else_[0]
+                self.newline()
+                self.write("elif ")
+                self.visit(node.test)
+                self.write(":")
+                self.body(node.body)
+            else:
+                self.newline()
+                self.write("else:")
+                self.body(else_)
+                break
+
+    def visit_For(self, node):
+        self.newline()
+        self.write("for ")
+        self.visit(node.target)
+        self.write(" in ")
+        self.visit(node.iter)
+        self.write(":")
+        self.body_or_else(node)
+
+    def visit_While(self, node):
+        self.newline()
+        self.write("while ")
+        self.visit(node.test)
+        self.write(":")
+        self.body_or_else(node)
+
+    def visit_With(self, node):
+        self.newline()
+        self.write("with ")
+        self.visit(node.context_expr)
+        if node.optional_vars is not None:
+            self.write(" as ")
+            self.visit(node.optional_vars)
+        self.write(":")
+        self.body(node.body)
+
+    def visit_Pass(self, node):
+        self.newline()
+        self.write("pass")
+
+    def visit_Print(self, node):
+        # XXX: python 2.6 only
+        self.newline()
+        self.write("print ")
+        want_comma = False
+        if node.dest is not None:
+            self.write(" >> ")
+            self.visit(node.dest)
+            want_comma = True
+        for value in node.values:
+            if want_comma:
+                self.write(", ")
+            self.visit(value)
+            want_comma = True
+        if not node.nl:
+            self.write(",")
+
+    def visit_Delete(self, node):
+        self.newline()
+        self.write("del ")
+        for idx, target in enumerate(node):
+            if idx:
+                self.write(", ")
+            self.visit(target)
+
+    def visit_TryExcept(self, node):
+        self.newline()
+        self.write("try:")
+        self.body(node.body)
+        for handler in node.handlers:
+            self.visit(handler)
+
+    def visit_TryFinally(self, node):
+        self.newline()
+        self.write("try:")
+        self.body(node.body)
+        self.newline()
+        self.write("finally:")
+        self.body(node.finalbody)
+
+    def visit_Global(self, node):
+        self.newline()
+        self.write("global " + ", ".join(node.names))
+
+    def visit_Nonlocal(self, node):
+        self.newline()
+        self.write("nonlocal " + ", ".join(node.names))
+
+    def visit_Return(self, node):
+        self.newline()
+        self.write("return ")
+        self.visit(node.value)
+
+    def visit_Break(self, node):
+        self.newline()
+        self.write("break")
+
+    def visit_Continue(self, node):
+        self.newline()
+        self.write("continue")
+
+    def visit_Raise(self, node):
+        # XXX: Python 2.6 / 3.0 compatibility
+        self.newline()
+        self.write("raise")
+        if hasattr(node, "exc") and node.exc is not None:
+            self.write(" ")
+            self.visit(node.exc)
+            if node.cause is not None:
+                self.write(" from ")
+                self.visit(node.cause)
+        elif hasattr(node, "type") and node.type is not None:
+            self.visit(node.type)
+            if node.inst is not None:
+                self.write(", ")
+                self.visit(node.inst)
+            if node.tback is not None:
+                self.write(", ")
+                self.visit(node.tback)
+
+    # Expressions
+
+    def visit_Attribute(self, node):
+        self.visit(node.value)
+        self.write("." + node.attr)
+
+    def visit_Call(self, node):
+        want_comma = []
+
+        def write_comma():
+            if want_comma:
+                self.write(", ")
+            else:
+                want_comma.append(True)
+
+        self.visit(node.func)
+        self.write("(")
+        for arg in node.args:
+            write_comma()
+            self.visit(arg)
+        for keyword in node.keywords:
+            write_comma()
+            self.write(keyword.arg + "=")
+            self.visit(keyword.value)
+        if getattr(node, "starargs", None):
+            write_comma()
+            self.write("*")
+            self.visit(node.starargs)
+        if getattr(node, "kwargs", None):
+            write_comma()
+            self.write("**")
+            self.visit(node.kwargs)
+        self.write(")")
+
+    def visit_Name(self, node):
+        self.write(node.id)
+
+    def visit_NameConstant(self, node):
+        self.write(str(node.value))
+
+    def visit_arg(self, node):
+        self.write(node.arg)
+
+    def visit_Str(self, node):
+        self.write(repr(node.s))
+
+    def visit_Bytes(self, node):
+        self.write(repr(node.s))
+
+    def visit_Num(self, node):
+        self.write(repr(node.n))
+
+    # newly needed in Python 3.8
+    def visit_Constant(self, node):
+        self.write(repr(node.value))
+
+    def visit_Tuple(self, node):
+        self.write("(")
+        idx = -1
+        for idx, item in enumerate(node.elts):
+            if idx:
+                self.write(", ")
+            self.visit(item)
+        self.write(idx and ")" or ",)")
+
+    def sequence_visit(left, right):
+        def visit(self, node):
+            self.write(left)
+            for idx, item in enumerate(node.elts):
+                if idx:
+                    self.write(", ")
+                self.visit(item)
+            self.write(right)
+
+        return visit
+
+    visit_List = sequence_visit("[", "]")
+    visit_Set = sequence_visit("{", "}")
+    del sequence_visit
+
+    def visit_Dict(self, node):
+        self.write("{")
+        for idx, (key, value) in enumerate(zip(node.keys, node.values)):
+            if idx:
+                self.write(", ")
+            self.visit(key)
+            self.write(": ")
+            self.visit(value)
+        self.write("}")
+
+    def visit_BinOp(self, node):
+        self.write("(")
+        self.visit(node.left)
+        self.write(" %s " % BINOP_SYMBOLS[type(node.op)])
+        self.visit(node.right)
+        self.write(")")
+
+    def visit_BoolOp(self, node):
+        self.write("(")
+        for idx, value in enumerate(node.values):
+            if idx:
+                self.write(" %s " % BOOLOP_SYMBOLS[type(node.op)])
+            self.visit(value)
+        self.write(")")
+
+    def visit_Compare(self, node):
+        self.write("(")
+        self.visit(node.left)
+        for op, right in zip(node.ops, node.comparators):
+            self.write(" %s " % CMPOP_SYMBOLS[type(op)])
+            self.visit(right)
+        self.write(")")
+
+    def visit_UnaryOp(self, node):
+        self.write("(")
+        op = UNARYOP_SYMBOLS[type(node.op)]
+        self.write(op)
+        if op == "not":
+            self.write(" ")
+        self.visit(node.operand)
+        self.write(")")
+
+    def visit_Subscript(self, node):
+        self.visit(node.value)
+        self.write("[")
+        self.visit(node.slice)
+        self.write("]")
+
+    def visit_Slice(self, node):
+        if node.lower is not None:
+            self.visit(node.lower)
+        self.write(":")
+        if node.upper is not None:
+            self.visit(node.upper)
+        if node.step is not None:
+            self.write(":")
+            if not (isinstance(node.step, Name) and node.step.id == "None"):
+                self.visit(node.step)
+
+    def visit_ExtSlice(self, node):
+        for idx, item in node.dims:
+            if idx:
+                self.write(", ")
+            self.visit(item)
+
+    def visit_Yield(self, node):
+        self.write("yield ")
+        self.visit(node.value)
+
+    def visit_Lambda(self, node):
+        self.write("lambda ")
+        self.signature(node.args)
+        self.write(": ")
+        self.visit(node.body)
+
+    def visit_Ellipsis(self, node):
+        self.write("Ellipsis")
+
+    def generator_visit(left, right):
+        def visit(self, node):
+            self.write(left)
+            self.visit(node.elt)
+            for comprehension in node.generators:
+                self.visit(comprehension)
+            self.write(right)
+
+        return visit
+
+    visit_ListComp = generator_visit("[", "]")
+    visit_GeneratorExp = generator_visit("(", ")")
+    visit_SetComp = generator_visit("{", "}")
+    del generator_visit
+
+    def visit_DictComp(self, node):
+        self.write("{")
+        self.visit(node.key)
+        self.write(": ")
+        self.visit(node.value)
+        for comprehension in node.generators:
+            self.visit(comprehension)
+        self.write("}")
+
+    def visit_IfExp(self, node):
+        self.visit(node.body)
+        self.write(" if ")
+        self.visit(node.test)
+        self.write(" else ")
+        self.visit(node.orelse)
+
+    def visit_Starred(self, node):
+        self.write("*")
+        self.visit(node.value)
+
+    def visit_Repr(self, node):
+        # XXX: python 2.6 only
+        self.write("`")
+        self.visit(node.value)
+        self.write("`")
+
+    # Helper Nodes
+
+    def visit_alias(self, node):
+        self.write(node.name)
+        if node.asname is not None:
+            self.write(" as " + node.asname)
+
+    def visit_comprehension(self, node):
+        self.write(" for ")
+        self.visit(node.target)
+        self.write(" in ")
+        self.visit(node.iter)
+        if node.ifs:
+            for if_ in node.ifs:
+                self.write(" if ")
+                self.visit(if_)
+
+    def visit_excepthandler(self, node):
+        self.newline()
+        self.write("except")
+        if node.type is not None:
+            self.write(" ")
+            self.visit(node.type)
+            if node.name is not None:
+                self.write(" as ")
+                self.visit(node.name)
+        self.write(":")
+        self.body(node.body)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ast.py b/backend/venv/lib/python3.12/site-packages/mako/ast.py
new file mode 100644
index 00000000..82cfa116
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ast.py
@@ -0,0 +1,202 @@
+# mako/ast.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""utilities for analyzing expressions and blocks of Python
+code, as well as generating Python from AST nodes"""
+
+import re
+
+from mako import exceptions
+from mako import pyparser
+
+
+class PythonCode:
+
+    """represents information about a string containing Python code"""
+
+    def __init__(self, code, **exception_kwargs):
+        self.code = code
+
+        # represents all identifiers which are assigned to at some point in
+        # the code
+        self.declared_identifiers = set()
+
+        # represents all identifiers which are referenced before their
+        # assignment, if any
+        self.undeclared_identifiers = set()
+
+        # note that an identifier can be in both the undeclared and declared
+        # lists.
+
+        # using AST to parse instead of using code.co_varnames,
+        # code.co_names has several advantages:
+        # - we can locate an identifier as "undeclared" even if
+        # its declared later in the same block of code
+        # - AST is less likely to break with version changes
+        # (for example, the behavior of co_names changed a little bit
+        # in python version 2.5)
+        if isinstance(code, str):
+            expr = pyparser.parse(code.lstrip(), "exec", **exception_kwargs)
+        else:
+            expr = code
+
+        f = pyparser.FindIdentifiers(self, **exception_kwargs)
+        f.visit(expr)
+
+
+class ArgumentList:
+
+    """parses a fragment of code as a comma-separated list of expressions"""
+
+    def __init__(self, code, **exception_kwargs):
+        self.codeargs = []
+        self.args = []
+        self.declared_identifiers = set()
+        self.undeclared_identifiers = set()
+        if isinstance(code, str):
+            if re.match(r"\S", code) and not re.match(r",\s*$", code):
+                # if theres text and no trailing comma, insure its parsed
+                # as a tuple by adding a trailing comma
+                code += ","
+            expr = pyparser.parse(code, "exec", **exception_kwargs)
+        else:
+            expr = code
+
+        f = pyparser.FindTuple(self, PythonCode, **exception_kwargs)
+        f.visit(expr)
+
+
+class PythonFragment(PythonCode):
+
+    """extends PythonCode to provide identifier lookups in partial control
+    statements
+
+    e.g.::
+
+        for x in 5:
+        elif y==9:
+        except (MyException, e):
+
+    """
+
+    def __init__(self, code, **exception_kwargs):
+        m = re.match(r"^(\w+)(?:\s+(.*?))?:\s*(#|$)", code.strip(), re.S)
+        if not m:
+            raise exceptions.CompileException(
+                "Fragment '%s' is not a partial control statement" % code,
+                **exception_kwargs,
+            )
+        if m.group(3):
+            code = code[: m.start(3)]
+        (keyword, expr) = m.group(1, 2)
+        if keyword in ["for", "if", "while"]:
+            code = code + "pass"
+        elif keyword == "try":
+            code = code + "pass\nexcept:pass"
+        elif keyword in ["elif", "else"]:
+            code = "if False:pass\n" + code + "pass"
+        elif keyword == "except":
+            code = "try:pass\n" + code + "pass"
+        elif keyword == "with":
+            code = code + "pass"
+        else:
+            raise exceptions.CompileException(
+                "Unsupported control keyword: '%s'" % keyword,
+                **exception_kwargs,
+            )
+        super().__init__(code, **exception_kwargs)
+
+
+class FunctionDecl:
+
+    """function declaration"""
+
+    def __init__(self, code, allow_kwargs=True, **exception_kwargs):
+        self.code = code
+        expr = pyparser.parse(code, "exec", **exception_kwargs)
+
+        f = pyparser.ParseFunc(self, **exception_kwargs)
+        f.visit(expr)
+        if not hasattr(self, "funcname"):
+            raise exceptions.CompileException(
+                "Code '%s' is not a function declaration" % code,
+                **exception_kwargs,
+            )
+        if not allow_kwargs and self.kwargs:
+            raise exceptions.CompileException(
+                "'**%s' keyword argument not allowed here"
+                % self.kwargnames[-1],
+                **exception_kwargs,
+            )
+
+    def get_argument_expressions(self, as_call=False):
+        """Return the argument declarations of this FunctionDecl as a printable
+        list.
+
+        By default the return value is appropriate for writing in a ``def``;
+        set `as_call` to true to build arguments to be passed to the function
+        instead (assuming locals with the same names as the arguments exist).
+        """
+
+        namedecls = []
+
+        # Build in reverse order, since defaults and slurpy args come last
+        argnames = self.argnames[::-1]
+        kwargnames = self.kwargnames[::-1]
+        defaults = self.defaults[::-1]
+        kwdefaults = self.kwdefaults[::-1]
+
+        # Named arguments
+        if self.kwargs:
+            namedecls.append("**" + kwargnames.pop(0))
+
+        for name in kwargnames:
+            # Keyword-only arguments must always be used by name, so even if
+            # this is a call, print out `foo=foo`
+            if as_call:
+                namedecls.append("%s=%s" % (name, name))
+            elif kwdefaults:
+                default = kwdefaults.pop(0)
+                if default is None:
+                    # The AST always gives kwargs a default, since you can do
+                    # `def foo(*, a=1, b, c=3)`
+                    namedecls.append(name)
+                else:
+                    namedecls.append(
+                        "%s=%s"
+                        % (name, pyparser.ExpressionGenerator(default).value())
+                    )
+            else:
+                namedecls.append(name)
+
+        # Positional arguments
+        if self.varargs:
+            namedecls.append("*" + argnames.pop(0))
+
+        for name in argnames:
+            if as_call or not defaults:
+                namedecls.append(name)
+            else:
+                default = defaults.pop(0)
+                namedecls.append(
+                    "%s=%s"
+                    % (name, pyparser.ExpressionGenerator(default).value())
+                )
+
+        namedecls.reverse()
+        return namedecls
+
+    @property
+    def allargnames(self):
+        return tuple(self.argnames) + tuple(self.kwargnames)
+
+
+class FunctionArgs(FunctionDecl):
+
+    """the argument portion of a function declaration"""
+
+    def __init__(self, code, **kwargs):
+        super().__init__("def ANON(%s):pass" % code, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/cache.py b/backend/venv/lib/python3.12/site-packages/mako/cache.py
new file mode 100644
index 00000000..267411c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/cache.py
@@ -0,0 +1,239 @@
+# mako/cache.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+from mako import util
+
+_cache_plugins = util.PluginLoader("mako.cache")
+
+register_plugin = _cache_plugins.register
+register_plugin("beaker", "mako.ext.beaker_cache", "BeakerCacheImpl")
+
+
+class Cache:
+
+    """Represents a data content cache made available to the module
+    space of a specific :class:`.Template` object.
+
+    .. versionadded:: 0.6
+       :class:`.Cache` by itself is mostly a
+       container for a :class:`.CacheImpl` object, which implements
+       a fixed API to provide caching services; specific subclasses exist to
+       implement different
+       caching strategies.   Mako includes a backend that works with
+       the Beaker caching system.   Beaker itself then supports
+       a number of backends (i.e. file, memory, memcached, etc.)
+
+    The construction of a :class:`.Cache` is part of the mechanics
+    of a :class:`.Template`, and programmatic access to this
+    cache is typically via the :attr:`.Template.cache` attribute.
+
+    """
+
+    impl = None
+    """Provide the :class:`.CacheImpl` in use by this :class:`.Cache`.
+
+    This accessor allows a :class:`.CacheImpl` with additional
+    methods beyond that of :class:`.Cache` to be used programmatically.
+
+    """
+
+    id = None
+    """Return the 'id' that identifies this cache.
+
+    This is a value that should be globally unique to the
+    :class:`.Template` associated with this cache, and can
+    be used by a caching system to name a local container
+    for data specific to this template.
+
+    """
+
+    starttime = None
+    """Epochal time value for when the owning :class:`.Template` was
+    first compiled.
+
+    A cache implementation may wish to invalidate data earlier than
+    this timestamp; this has the effect of the cache for a specific
+    :class:`.Template` starting clean any time the :class:`.Template`
+    is recompiled, such as when the original template file changed on
+    the filesystem.
+
+    """
+
+    def __init__(self, template, *args):
+        # check for a stale template calling the
+        # constructor
+        if isinstance(template, str) and args:
+            return
+        self.template = template
+        self.id = template.module.__name__
+        self.starttime = template.module._modified_time
+        self._def_regions = {}
+        self.impl = self._load_impl(self.template.cache_impl)
+
+    def _load_impl(self, name):
+        return _cache_plugins.load(name)(self)
+
+    def get_or_create(self, key, creation_function, **kw):
+        """Retrieve a value from the cache, using the given creation function
+        to generate a new value."""
+
+        return self._ctx_get_or_create(key, creation_function, None, **kw)
+
+    def _ctx_get_or_create(self, key, creation_function, context, **kw):
+        """Retrieve a value from the cache, using the given creation function
+        to generate a new value."""
+
+        if not self.template.cache_enabled:
+            return creation_function()
+
+        return self.impl.get_or_create(
+            key, creation_function, **self._get_cache_kw(kw, context)
+        )
+
+    def set(self, key, value, **kw):
+        r"""Place a value in the cache.
+
+        :param key: the value's key.
+        :param value: the value.
+        :param \**kw: cache configuration arguments.
+
+        """
+
+        self.impl.set(key, value, **self._get_cache_kw(kw, None))
+
+    put = set
+    """A synonym for :meth:`.Cache.set`.
+
+    This is here for backwards compatibility.
+
+    """
+
+    def get(self, key, **kw):
+        r"""Retrieve a value from the cache.
+
+        :param key: the value's key.
+        :param \**kw: cache configuration arguments.  The
+         backend is configured using these arguments upon first request.
+         Subsequent requests that use the same series of configuration
+         values will use that same backend.
+
+        """
+        return self.impl.get(key, **self._get_cache_kw(kw, None))
+
+    def invalidate(self, key, **kw):
+        r"""Invalidate a value in the cache.
+
+        :param key: the value's key.
+        :param \**kw: cache configuration arguments.  The
+         backend is configured using these arguments upon first request.
+         Subsequent requests that use the same series of configuration
+         values will use that same backend.
+
+        """
+        self.impl.invalidate(key, **self._get_cache_kw(kw, None))
+
+    def invalidate_body(self):
+        """Invalidate the cached content of the "body" method for this
+        template.
+
+        """
+        self.invalidate("render_body", __M_defname="render_body")
+
+    def invalidate_def(self, name):
+        """Invalidate the cached content of a particular ``<%def>`` within this
+        template.
+
+        """
+
+        self.invalidate("render_%s" % name, __M_defname="render_%s" % name)
+
+    def invalidate_closure(self, name):
+        """Invalidate a nested ``<%def>`` within this template.
+
+        Caching of nested defs is a blunt tool as there is no
+        management of scope -- nested defs that use cache tags
+        need to have names unique of all other nested defs in the
+        template, else their content will be overwritten by
+        each other.
+
+        """
+
+        self.invalidate(name, __M_defname=name)
+
+    def _get_cache_kw(self, kw, context):
+        defname = kw.pop("__M_defname", None)
+        if not defname:
+            tmpl_kw = self.template.cache_args.copy()
+            tmpl_kw.update(kw)
+        elif defname in self._def_regions:
+            tmpl_kw = self._def_regions[defname]
+        else:
+            tmpl_kw = self.template.cache_args.copy()
+            tmpl_kw.update(kw)
+            self._def_regions[defname] = tmpl_kw
+        if context and self.impl.pass_context:
+            tmpl_kw = tmpl_kw.copy()
+            tmpl_kw.setdefault("context", context)
+        return tmpl_kw
+
+
+class CacheImpl:
+
+    """Provide a cache implementation for use by :class:`.Cache`."""
+
+    def __init__(self, cache):
+        self.cache = cache
+
+    pass_context = False
+    """If ``True``, the :class:`.Context` will be passed to
+    :meth:`get_or_create <.CacheImpl.get_or_create>` as the name ``'context'``.
+    """
+
+    def get_or_create(self, key, creation_function, **kw):
+        r"""Retrieve a value from the cache, using the given creation function
+        to generate a new value.
+
+        This function *must* return a value, either from
+        the cache, or via the given creation function.
+        If the creation function is called, the newly
+        created value should be populated into the cache
+        under the given key before being returned.
+
+        :param key: the value's key.
+        :param creation_function: function that when called generates
+         a new value.
+        :param \**kw: cache configuration arguments.
+
+        """
+        raise NotImplementedError()
+
+    def set(self, key, value, **kw):
+        r"""Place a value in the cache.
+
+        :param key: the value's key.
+        :param value: the value.
+        :param \**kw: cache configuration arguments.
+
+        """
+        raise NotImplementedError()
+
+    def get(self, key, **kw):
+        r"""Retrieve a value from the cache.
+
+        :param key: the value's key.
+        :param \**kw: cache configuration arguments.
+
+        """
+        raise NotImplementedError()
+
+    def invalidate(self, key, **kw):
+        r"""Invalidate a value in the cache.
+
+        :param key: the value's key.
+        :param \**kw: cache configuration arguments.
+
+        """
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/mako/cmd.py b/backend/venv/lib/python3.12/site-packages/mako/cmd.py
new file mode 100644
index 00000000..3fff197b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/cmd.py
@@ -0,0 +1,99 @@
+# mako/cmd.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+from argparse import ArgumentParser
+from os.path import dirname
+from os.path import isfile
+import sys
+
+from mako import exceptions
+from mako.lookup import TemplateLookup
+from mako.template import Template
+
+
+def varsplit(var):
+    if "=" not in var:
+        return (var, "")
+    return var.split("=", 1)
+
+
+def _exit():
+    sys.stderr.write(exceptions.text_error_template().render())
+    sys.exit(1)
+
+
+def cmdline(argv=None):
+    parser = ArgumentParser()
+    parser.add_argument(
+        "--var",
+        default=[],
+        action="append",
+        help="variable (can be used multiple times, use name=value)",
+    )
+    parser.add_argument(
+        "--template-dir",
+        default=[],
+        action="append",
+        help="Directory to use for template lookup (multiple "
+        "directories may be provided). If not given then if the "
+        "template is read from stdin, the value defaults to be "
+        "the current directory, otherwise it defaults to be the "
+        "parent directory of the file provided.",
+    )
+    parser.add_argument(
+        "--output-encoding", default=None, help="force output encoding"
+    )
+    parser.add_argument(
+        "--output-file",
+        default=None,
+        help="Write to file upon successful render instead of stdout",
+    )
+    parser.add_argument("input", nargs="?", default="-")
+
+    options = parser.parse_args(argv)
+
+    output_encoding = options.output_encoding
+    output_file = options.output_file
+
+    if options.input == "-":
+        lookup_dirs = options.template_dir or ["."]
+        lookup = TemplateLookup(lookup_dirs)
+        try:
+            template = Template(
+                sys.stdin.read(),
+                lookup=lookup,
+                output_encoding=output_encoding,
+            )
+        except:
+            _exit()
+    else:
+        filename = options.input
+        if not isfile(filename):
+            raise SystemExit("error: can't find %s" % filename)
+        lookup_dirs = options.template_dir or [dirname(filename)]
+        lookup = TemplateLookup(lookup_dirs)
+        try:
+            template = Template(
+                filename=filename,
+                lookup=lookup,
+                output_encoding=output_encoding,
+            )
+        except:
+            _exit()
+
+    kw = dict(varsplit(var) for var in options.var)
+    try:
+        rendered = template.render(**kw)
+    except:
+        _exit()
+    else:
+        if output_file:
+            open(output_file, "wt", encoding=output_encoding).write(rendered)
+        else:
+            sys.stdout.write(rendered)
+
+
+if __name__ == "__main__":
+    cmdline()
diff --git a/backend/venv/lib/python3.12/site-packages/mako/codegen.py b/backend/venv/lib/python3.12/site-packages/mako/codegen.py
new file mode 100644
index 00000000..b5293f47
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/codegen.py
@@ -0,0 +1,1319 @@
+# mako/codegen.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""provides functionality for rendering a parsetree constructing into module
+source code."""
+
+import json
+import re
+import time
+
+from mako import ast
+from mako import exceptions
+from mako import filters
+from mako import parsetree
+from mako import util
+from mako.pygen import PythonPrinter
+
+
+MAGIC_NUMBER = 10
+
+# names which are hardwired into the
+# template and are not accessed via the
+# context itself
+TOPLEVEL_DECLARED = {"UNDEFINED", "STOP_RENDERING"}
+RESERVED_NAMES = {"context", "loop"}.union(TOPLEVEL_DECLARED)
+
+
+def compile(  # noqa
+    node,
+    uri,
+    filename=None,
+    default_filters=None,
+    buffer_filters=None,
+    imports=None,
+    future_imports=None,
+    source_encoding=None,
+    generate_magic_comment=True,
+    strict_undefined=False,
+    enable_loop=True,
+    reserved_names=frozenset(),
+):
+    """Generate module source code given a parsetree node,
+    uri, and optional source filename"""
+
+    buf = util.FastEncodingBuffer()
+
+    printer = PythonPrinter(buf)
+    _GenerateRenderMethod(
+        printer,
+        _CompileContext(
+            uri,
+            filename,
+            default_filters,
+            buffer_filters,
+            imports,
+            future_imports,
+            source_encoding,
+            generate_magic_comment,
+            strict_undefined,
+            enable_loop,
+            reserved_names,
+        ),
+        node,
+    )
+    return buf.getvalue()
+
+
+class _CompileContext:
+    def __init__(
+        self,
+        uri,
+        filename,
+        default_filters,
+        buffer_filters,
+        imports,
+        future_imports,
+        source_encoding,
+        generate_magic_comment,
+        strict_undefined,
+        enable_loop,
+        reserved_names,
+    ):
+        self.uri = uri
+        self.filename = filename
+        self.default_filters = default_filters
+        self.buffer_filters = buffer_filters
+        self.imports = imports
+        self.future_imports = future_imports
+        self.source_encoding = source_encoding
+        self.generate_magic_comment = generate_magic_comment
+        self.strict_undefined = strict_undefined
+        self.enable_loop = enable_loop
+        self.reserved_names = reserved_names
+
+
+class _GenerateRenderMethod:
+
+    """A template visitor object which generates the
+    full module source for a template.
+
+    """
+
+    def __init__(self, printer, compiler, node):
+        self.printer = printer
+        self.compiler = compiler
+        self.node = node
+        self.identifier_stack = [None]
+        self.in_def = isinstance(node, (parsetree.DefTag, parsetree.BlockTag))
+
+        if self.in_def:
+            name = "render_%s" % node.funcname
+            args = node.get_argument_expressions()
+            filtered = len(node.filter_args.args) > 0
+            buffered = eval(node.attributes.get("buffered", "False"))
+            cached = eval(node.attributes.get("cached", "False"))
+            defs = None
+            pagetag = None
+            if node.is_block and not node.is_anonymous:
+                args += ["**pageargs"]
+        else:
+            defs = self.write_toplevel()
+            pagetag = self.compiler.pagetag
+            name = "render_body"
+            if pagetag is not None:
+                args = pagetag.body_decl.get_argument_expressions()
+                if not pagetag.body_decl.kwargs:
+                    args += ["**pageargs"]
+                cached = eval(pagetag.attributes.get("cached", "False"))
+                self.compiler.enable_loop = self.compiler.enable_loop or eval(
+                    pagetag.attributes.get("enable_loop", "False")
+                )
+            else:
+                args = ["**pageargs"]
+                cached = False
+            buffered = filtered = False
+        if args is None:
+            args = ["context"]
+        else:
+            args = [a for a in ["context"] + args]
+
+        self.write_render_callable(
+            pagetag or node, name, args, buffered, filtered, cached
+        )
+
+        if defs is not None:
+            for node in defs:
+                _GenerateRenderMethod(printer, compiler, node)
+
+        if not self.in_def:
+            self.write_metadata_struct()
+
+    def write_metadata_struct(self):
+        self.printer.source_map[self.printer.lineno] = max(
+            self.printer.source_map
+        )
+        struct = {
+            "filename": self.compiler.filename,
+            "uri": self.compiler.uri,
+            "source_encoding": self.compiler.source_encoding,
+            "line_map": self.printer.source_map,
+        }
+        self.printer.writelines(
+            '"""',
+            "__M_BEGIN_METADATA",
+            json.dumps(struct),
+            "__M_END_METADATA\n" '"""',
+        )
+
+    @property
+    def identifiers(self):
+        return self.identifier_stack[-1]
+
+    def write_toplevel(self):
+        """Traverse a template structure for module-level directives and
+        generate the start of module-level code.
+
+        """
+        inherit = []
+        namespaces = {}
+        module_code = []
+
+        self.compiler.pagetag = None
+
+        class FindTopLevel:
+            def visitInheritTag(s, node):
+                inherit.append(node)
+
+            def visitNamespaceTag(s, node):
+                namespaces[node.name] = node
+
+            def visitPageTag(s, node):
+                self.compiler.pagetag = node
+
+            def visitCode(s, node):
+                if node.ismodule:
+                    module_code.append(node)
+
+        f = FindTopLevel()
+        for n in self.node.nodes:
+            n.accept_visitor(f)
+
+        self.compiler.namespaces = namespaces
+
+        module_ident = set()
+        for n in module_code:
+            module_ident = module_ident.union(n.declared_identifiers())
+
+        module_identifiers = _Identifiers(self.compiler)
+        module_identifiers.declared = module_ident
+
+        # module-level names, python code
+        if (
+            self.compiler.generate_magic_comment
+            and self.compiler.source_encoding
+        ):
+            self.printer.writeline(
+                "# -*- coding:%s -*-" % self.compiler.source_encoding
+            )
+
+        if self.compiler.future_imports:
+            self.printer.writeline(
+                "from __future__ import %s"
+                % (", ".join(self.compiler.future_imports),)
+            )
+        self.printer.writeline("from mako import runtime, filters, cache")
+        self.printer.writeline("UNDEFINED = runtime.UNDEFINED")
+        self.printer.writeline("STOP_RENDERING = runtime.STOP_RENDERING")
+        self.printer.writeline("__M_dict_builtin = dict")
+        self.printer.writeline("__M_locals_builtin = locals")
+        self.printer.writeline("_magic_number = %r" % MAGIC_NUMBER)
+        self.printer.writeline("_modified_time = %r" % time.time())
+        self.printer.writeline("_enable_loop = %r" % self.compiler.enable_loop)
+        self.printer.writeline(
+            "_template_filename = %r" % self.compiler.filename
+        )
+        self.printer.writeline("_template_uri = %r" % self.compiler.uri)
+        self.printer.writeline(
+            "_source_encoding = %r" % self.compiler.source_encoding
+        )
+        if self.compiler.imports:
+            buf = ""
+            for imp in self.compiler.imports:
+                buf += imp + "\n"
+                self.printer.writeline(imp)
+            impcode = ast.PythonCode(
+                buf,
+                source="",
+                lineno=0,
+                pos=0,
+                filename="template defined imports",
+            )
+        else:
+            impcode = None
+
+        main_identifiers = module_identifiers.branch(self.node)
+        mit = module_identifiers.topleveldefs
+        module_identifiers.topleveldefs = mit.union(
+            main_identifiers.topleveldefs
+        )
+        module_identifiers.declared.update(TOPLEVEL_DECLARED)
+        if impcode:
+            module_identifiers.declared.update(impcode.declared_identifiers)
+
+        self.compiler.identifiers = module_identifiers
+        self.printer.writeline(
+            "_exports = %r"
+            % [n.name for n in main_identifiers.topleveldefs.values()]
+        )
+        self.printer.write_blanks(2)
+
+        if len(module_code):
+            self.write_module_code(module_code)
+
+        if len(inherit):
+            self.write_namespaces(namespaces)
+            self.write_inherit(inherit[-1])
+        elif len(namespaces):
+            self.write_namespaces(namespaces)
+
+        return list(main_identifiers.topleveldefs.values())
+
+    def write_render_callable(
+        self, node, name, args, buffered, filtered, cached
+    ):
+        """write a top-level render callable.
+
+        this could be the main render() method or that of a top-level def."""
+
+        if self.in_def:
+            decorator = node.decorator
+            if decorator:
+                self.printer.writeline(
+                    "@runtime._decorate_toplevel(%s)" % decorator
+                )
+
+        self.printer.start_source(node.lineno)
+        self.printer.writelines(
+            "def %s(%s):" % (name, ",".join(args)),
+            # push new frame, assign current frame to __M_caller
+            "__M_caller = context.caller_stack._push_frame()",
+            "try:",
+        )
+        if buffered or filtered or cached:
+            self.printer.writeline("context._push_buffer()")
+
+        self.identifier_stack.append(
+            self.compiler.identifiers.branch(self.node)
+        )
+        if (not self.in_def or self.node.is_block) and "**pageargs" in args:
+            self.identifier_stack[-1].argument_declared.add("pageargs")
+
+        if not self.in_def and (
+            len(self.identifiers.locally_assigned) > 0
+            or len(self.identifiers.argument_declared) > 0
+        ):
+            self.printer.writeline(
+                "__M_locals = __M_dict_builtin(%s)"
+                % ",".join(
+                    [
+                        "%s=%s" % (x, x)
+                        for x in self.identifiers.argument_declared
+                    ]
+                )
+            )
+
+        self.write_variable_declares(self.identifiers, toplevel=True)
+
+        for n in self.node.nodes:
+            n.accept_visitor(self)
+
+        self.write_def_finish(self.node, buffered, filtered, cached)
+        self.printer.writeline(None)
+        self.printer.write_blanks(2)
+        if cached:
+            self.write_cache_decorator(
+                node, name, args, buffered, self.identifiers, toplevel=True
+            )
+
+    def write_module_code(self, module_code):
+        """write module-level template code, i.e. that which
+        is enclosed in <%! %> tags in the template."""
+        for n in module_code:
+            self.printer.write_indented_block(n.text, starting_lineno=n.lineno)
+
+    def write_inherit(self, node):
+        """write the module-level inheritance-determination callable."""
+
+        self.printer.writelines(
+            "def _mako_inherit(template, context):",
+            "_mako_generate_namespaces(context)",
+            "return runtime._inherit_from(context, %s, _template_uri)"
+            % (node.parsed_attributes["file"]),
+            None,
+        )
+
+    def write_namespaces(self, namespaces):
+        """write the module-level namespace-generating callable."""
+        self.printer.writelines(
+            "def _mako_get_namespace(context, name):",
+            "try:",
+            "return context.namespaces[(__name__, name)]",
+            "except KeyError:",
+            "_mako_generate_namespaces(context)",
+            "return context.namespaces[(__name__, name)]",
+            None,
+            None,
+        )
+        self.printer.writeline("def _mako_generate_namespaces(context):")
+
+        for node in namespaces.values():
+            if "import" in node.attributes:
+                self.compiler.has_ns_imports = True
+            self.printer.start_source(node.lineno)
+            if len(node.nodes):
+                self.printer.writeline("def make_namespace():")
+                export = []
+                identifiers = self.compiler.identifiers.branch(node)
+                self.in_def = True
+
+                class NSDefVisitor:
+                    def visitDefTag(s, node):
+                        s.visitDefOrBase(node)
+
+                    def visitBlockTag(s, node):
+                        s.visitDefOrBase(node)
+
+                    def visitDefOrBase(s, node):
+                        if node.is_anonymous:
+                            raise exceptions.CompileException(
+                                "Can't put anonymous blocks inside "
+                                "<%namespace>",
+                                **node.exception_kwargs,
+                            )
+                        self.write_inline_def(node, identifiers, nested=False)
+                        export.append(node.funcname)
+
+                vis = NSDefVisitor()
+                for n in node.nodes:
+                    n.accept_visitor(vis)
+                self.printer.writeline("return [%s]" % (",".join(export)))
+                self.printer.writeline(None)
+                self.in_def = False
+                callable_name = "make_namespace()"
+            else:
+                callable_name = "None"
+
+            if "file" in node.parsed_attributes:
+                self.printer.writeline(
+                    "ns = runtime.TemplateNamespace(%r,"
+                    " context._clean_inheritance_tokens(),"
+                    " templateuri=%s, callables=%s, "
+                    " calling_uri=_template_uri)"
+                    % (
+                        node.name,
+                        node.parsed_attributes.get("file", "None"),
+                        callable_name,
+                    )
+                )
+            elif "module" in node.parsed_attributes:
+                self.printer.writeline(
+                    "ns = runtime.ModuleNamespace(%r,"
+                    " context._clean_inheritance_tokens(),"
+                    " callables=%s, calling_uri=_template_uri,"
+                    " module=%s)"
+                    % (
+                        node.name,
+                        callable_name,
+                        node.parsed_attributes.get("module", "None"),
+                    )
+                )
+            else:
+                self.printer.writeline(
+                    "ns = runtime.Namespace(%r,"
+                    " context._clean_inheritance_tokens(),"
+                    " callables=%s, calling_uri=_template_uri)"
+                    % (node.name, callable_name)
+                )
+            if eval(node.attributes.get("inheritable", "False")):
+                self.printer.writeline("context['self'].%s = ns" % (node.name))
+
+            self.printer.writeline(
+                "context.namespaces[(__name__, %s)] = ns" % repr(node.name)
+            )
+            self.printer.write_blanks(1)
+        if not len(namespaces):
+            self.printer.writeline("pass")
+        self.printer.writeline(None)
+
+    def write_variable_declares(self, identifiers, toplevel=False, limit=None):
+        """write variable declarations at the top of a function.
+
+        the variable declarations are in the form of callable
+        definitions for defs and/or name lookup within the
+        function's context argument. the names declared are based
+        on the names that are referenced in the function body,
+        which don't otherwise have any explicit assignment
+        operation. names that are assigned within the body are
+        assumed to be locally-scoped variables and are not
+        separately declared.
+
+        for def callable definitions, if the def is a top-level
+        callable then a 'stub' callable is generated which wraps
+        the current Context into a closure. if the def is not
+        top-level, it is fully rendered as a local closure.
+
+        """
+
+        # collection of all defs available to us in this scope
+        comp_idents = {c.funcname: c for c in identifiers.defs}
+        to_write = set()
+
+        # write "context.get()" for all variables we are going to
+        # need that arent in the namespace yet
+        to_write = to_write.union(identifiers.undeclared)
+
+        # write closure functions for closures that we define
+        # right here
+        to_write = to_write.union(
+            [c.funcname for c in identifiers.closuredefs.values()]
+        )
+
+        # remove identifiers that are declared in the argument
+        # signature of the callable
+        to_write = to_write.difference(identifiers.argument_declared)
+
+        # remove identifiers that we are going to assign to.
+        # in this way we mimic Python's behavior,
+        # i.e. assignment to a variable within a block
+        # means that variable is now a "locally declared" var,
+        # which cannot be referenced beforehand.
+        to_write = to_write.difference(identifiers.locally_declared)
+
+        if self.compiler.enable_loop:
+            has_loop = "loop" in to_write
+            to_write.discard("loop")
+        else:
+            has_loop = False
+
+        # if a limiting set was sent, constraint to those items in that list
+        # (this is used for the caching decorator)
+        if limit is not None:
+            to_write = to_write.intersection(limit)
+
+        if toplevel and getattr(self.compiler, "has_ns_imports", False):
+            self.printer.writeline("_import_ns = {}")
+            self.compiler.has_imports = True
+            for ident, ns in self.compiler.namespaces.items():
+                if "import" in ns.attributes:
+                    self.printer.writeline(
+                        "_mako_get_namespace(context, %r)."
+                        "_populate(_import_ns, %r)"
+                        % (
+                            ident,
+                            re.split(r"\s*,\s*", ns.attributes["import"]),
+                        )
+                    )
+
+        if has_loop:
+            self.printer.writeline("loop = __M_loop = runtime.LoopStack()")
+
+        for ident in to_write:
+            if ident in comp_idents:
+                comp = comp_idents[ident]
+                if comp.is_block:
+                    if not comp.is_anonymous:
+                        self.write_def_decl(comp, identifiers)
+                    else:
+                        self.write_inline_def(comp, identifiers, nested=True)
+                else:
+                    if comp.is_root():
+                        self.write_def_decl(comp, identifiers)
+                    else:
+                        self.write_inline_def(comp, identifiers, nested=True)
+
+            elif ident in self.compiler.namespaces:
+                self.printer.writeline(
+                    "%s = _mako_get_namespace(context, %r)" % (ident, ident)
+                )
+            else:
+                if getattr(self.compiler, "has_ns_imports", False):
+                    if self.compiler.strict_undefined:
+                        self.printer.writelines(
+                            "%s = _import_ns.get(%r, UNDEFINED)"
+                            % (ident, ident),
+                            "if %s is UNDEFINED:" % ident,
+                            "try:",
+                            "%s = context[%r]" % (ident, ident),
+                            "except KeyError:",
+                            "raise NameError(\"'%s' is not defined\")" % ident,
+                            None,
+                            None,
+                        )
+                    else:
+                        self.printer.writeline(
+                            "%s = _import_ns.get"
+                            "(%r, context.get(%r, UNDEFINED))"
+                            % (ident, ident, ident)
+                        )
+                else:
+                    if self.compiler.strict_undefined:
+                        self.printer.writelines(
+                            "try:",
+                            "%s = context[%r]" % (ident, ident),
+                            "except KeyError:",
+                            "raise NameError(\"'%s' is not defined\")" % ident,
+                            None,
+                        )
+                    else:
+                        self.printer.writeline(
+                            "%s = context.get(%r, UNDEFINED)" % (ident, ident)
+                        )
+
+        self.printer.writeline("__M_writer = context.writer()")
+
+    def write_def_decl(self, node, identifiers):
+        """write a locally-available callable referencing a top-level def"""
+        funcname = node.funcname
+        namedecls = node.get_argument_expressions()
+        nameargs = node.get_argument_expressions(as_call=True)
+
+        if not self.in_def and (
+            len(self.identifiers.locally_assigned) > 0
+            or len(self.identifiers.argument_declared) > 0
+        ):
+            nameargs.insert(0, "context._locals(__M_locals)")
+        else:
+            nameargs.insert(0, "context")
+        self.printer.writeline("def %s(%s):" % (funcname, ",".join(namedecls)))
+        self.printer.writeline(
+            "return render_%s(%s)" % (funcname, ",".join(nameargs))
+        )
+        self.printer.writeline(None)
+
+    def write_inline_def(self, node, identifiers, nested):
+        """write a locally-available def callable inside an enclosing def."""
+
+        namedecls = node.get_argument_expressions()
+
+        decorator = node.decorator
+        if decorator:
+            self.printer.writeline(
+                "@runtime._decorate_inline(context, %s)" % decorator
+            )
+        self.printer.writeline(
+            "def %s(%s):" % (node.funcname, ",".join(namedecls))
+        )
+        filtered = len(node.filter_args.args) > 0
+        buffered = eval(node.attributes.get("buffered", "False"))
+        cached = eval(node.attributes.get("cached", "False"))
+        self.printer.writelines(
+            # push new frame, assign current frame to __M_caller
+            "__M_caller = context.caller_stack._push_frame()",
+            "try:",
+        )
+        if buffered or filtered or cached:
+            self.printer.writelines("context._push_buffer()")
+
+        identifiers = identifiers.branch(node, nested=nested)
+
+        self.write_variable_declares(identifiers)
+
+        self.identifier_stack.append(identifiers)
+        for n in node.nodes:
+            n.accept_visitor(self)
+        self.identifier_stack.pop()
+
+        self.write_def_finish(node, buffered, filtered, cached)
+        self.printer.writeline(None)
+        if cached:
+            self.write_cache_decorator(
+                node,
+                node.funcname,
+                namedecls,
+                False,
+                identifiers,
+                inline=True,
+                toplevel=False,
+            )
+
+    def write_def_finish(
+        self, node, buffered, filtered, cached, callstack=True
+    ):
+        """write the end section of a rendering function, either outermost or
+        inline.
+
+        this takes into account if the rendering function was filtered,
+        buffered, etc.  and closes the corresponding try: block if any, and
+        writes code to retrieve captured content, apply filters, send proper
+        return value."""
+
+        if not buffered and not cached and not filtered:
+            self.printer.writeline("return ''")
+            if callstack:
+                self.printer.writelines(
+                    "finally:", "context.caller_stack._pop_frame()", None
+                )
+
+        if buffered or filtered or cached:
+            if buffered or cached:
+                # in a caching scenario, don't try to get a writer
+                # from the context after popping; assume the caching
+                # implemenation might be using a context with no
+                # extra buffers
+                self.printer.writelines(
+                    "finally:", "__M_buf = context._pop_buffer()"
+                )
+            else:
+                self.printer.writelines(
+                    "finally:",
+                    "__M_buf, __M_writer = context._pop_buffer_and_writer()",
+                )
+
+            if callstack:
+                self.printer.writeline("context.caller_stack._pop_frame()")
+
+            s = "__M_buf.getvalue()"
+            if filtered:
+                s = self.create_filter_callable(
+                    node.filter_args.args, s, False
+                )
+            self.printer.writeline(None)
+            if buffered and not cached:
+                s = self.create_filter_callable(
+                    self.compiler.buffer_filters, s, False
+                )
+            if buffered or cached:
+                self.printer.writeline("return %s" % s)
+            else:
+                self.printer.writelines("__M_writer(%s)" % s, "return ''")
+
+    def write_cache_decorator(
+        self,
+        node_or_pagetag,
+        name,
+        args,
+        buffered,
+        identifiers,
+        inline=False,
+        toplevel=False,
+    ):
+        """write a post-function decorator to replace a rendering
+        callable with a cached version of itself."""
+
+        self.printer.writeline("__M_%s = %s" % (name, name))
+        cachekey = node_or_pagetag.parsed_attributes.get(
+            "cache_key", repr(name)
+        )
+
+        cache_args = {}
+        if self.compiler.pagetag is not None:
+            cache_args.update(
+                (pa[6:], self.compiler.pagetag.parsed_attributes[pa])
+                for pa in self.compiler.pagetag.parsed_attributes
+                if pa.startswith("cache_") and pa != "cache_key"
+            )
+        cache_args.update(
+            (pa[6:], node_or_pagetag.parsed_attributes[pa])
+            for pa in node_or_pagetag.parsed_attributes
+            if pa.startswith("cache_") and pa != "cache_key"
+        )
+        if "timeout" in cache_args:
+            cache_args["timeout"] = int(eval(cache_args["timeout"]))
+
+        self.printer.writeline("def %s(%s):" % (name, ",".join(args)))
+
+        # form "arg1, arg2, arg3=arg3, arg4=arg4", etc.
+        pass_args = [
+            "%s=%s" % ((a.split("=")[0],) * 2) if "=" in a else a for a in args
+        ]
+
+        self.write_variable_declares(
+            identifiers,
+            toplevel=toplevel,
+            limit=node_or_pagetag.undeclared_identifiers(),
+        )
+        if buffered:
+            s = (
+                "context.get('local')."
+                "cache._ctx_get_or_create("
+                "%s, lambda:__M_%s(%s),  context, %s__M_defname=%r)"
+                % (
+                    cachekey,
+                    name,
+                    ",".join(pass_args),
+                    "".join(
+                        ["%s=%s, " % (k, v) for k, v in cache_args.items()]
+                    ),
+                    name,
+                )
+            )
+            # apply buffer_filters
+            s = self.create_filter_callable(
+                self.compiler.buffer_filters, s, False
+            )
+            self.printer.writelines("return " + s, None)
+        else:
+            self.printer.writelines(
+                "__M_writer(context.get('local')."
+                "cache._ctx_get_or_create("
+                "%s, lambda:__M_%s(%s), context, %s__M_defname=%r))"
+                % (
+                    cachekey,
+                    name,
+                    ",".join(pass_args),
+                    "".join(
+                        ["%s=%s, " % (k, v) for k, v in cache_args.items()]
+                    ),
+                    name,
+                ),
+                "return ''",
+                None,
+            )
+
+    def create_filter_callable(self, args, target, is_expression):
+        """write a filter-applying expression based on the filters
+        present in the given filter names, adjusting for the global
+        'default' filter aliases as needed."""
+
+        def locate_encode(name):
+            if re.match(r"decode\..+", name):
+                return "filters." + name
+            else:
+                return filters.DEFAULT_ESCAPES.get(name, name)
+
+        if "n" not in args:
+            if is_expression:
+                if self.compiler.pagetag:
+                    args = self.compiler.pagetag.filter_args.args + args
+                if self.compiler.default_filters and "n" not in args:
+                    args = self.compiler.default_filters + args
+        for e in args:
+            # if filter given as a function, get just the identifier portion
+            if e == "n":
+                continue
+            m = re.match(r"(.+?)(\(.*\))", e)
+            if m:
+                ident, fargs = m.group(1, 2)
+                f = locate_encode(ident)
+                e = f + fargs
+            else:
+                e = locate_encode(e)
+                assert e is not None
+            target = "%s(%s)" % (e, target)
+        return target
+
+    def visitExpression(self, node):
+        self.printer.start_source(node.lineno)
+        if (
+            len(node.escapes)
+            or (
+                self.compiler.pagetag is not None
+                and len(self.compiler.pagetag.filter_args.args)
+            )
+            or len(self.compiler.default_filters)
+        ):
+            s = self.create_filter_callable(
+                node.escapes_code.args, "%s" % node.text, True
+            )
+            self.printer.writeline("__M_writer(%s)" % s)
+        else:
+            self.printer.writeline("__M_writer(%s)" % node.text)
+
+    def visitControlLine(self, node):
+        if node.isend:
+            self.printer.writeline(None)
+            if node.has_loop_context:
+                self.printer.writeline("finally:")
+                self.printer.writeline("loop = __M_loop._exit()")
+                self.printer.writeline(None)
+        else:
+            self.printer.start_source(node.lineno)
+            if self.compiler.enable_loop and node.keyword == "for":
+                text = mangle_mako_loop(node, self.printer)
+            else:
+                text = node.text
+            self.printer.writeline(text)
+            children = node.get_children()
+
+            # this covers the four situations where we want to insert a pass:
+            # 1) a ternary control line with no children,
+            # 2) a primary control line with nothing but its own ternary
+            #       and end control lines, and
+            # 3) any control line with no content other than comments
+            # 4) the first control block with no content other than comments
+            def _search_for_control_line():
+                for c in children:
+                    if isinstance(c, parsetree.Comment):
+                        continue
+                    elif isinstance(c, parsetree.ControlLine):
+                        return True
+                    return False
+
+            if (
+                not children
+                or all(
+                    isinstance(c, (parsetree.Comment, parsetree.ControlLine))
+                    for c in children
+                )
+                and all(
+                    (node.is_ternary(c.keyword) or c.isend)
+                    for c in children
+                    if isinstance(c, parsetree.ControlLine)
+                )
+                or _search_for_control_line()
+            ):
+                self.printer.writeline("pass")
+
+    def visitText(self, node):
+        self.printer.start_source(node.lineno)
+        self.printer.writeline("__M_writer(%s)" % repr(node.content))
+
+    def visitTextTag(self, node):
+        filtered = len(node.filter_args.args) > 0
+        if filtered:
+            self.printer.writelines(
+                "__M_writer = context._push_writer()", "try:"
+            )
+        for n in node.nodes:
+            n.accept_visitor(self)
+        if filtered:
+            self.printer.writelines(
+                "finally:",
+                "__M_buf, __M_writer = context._pop_buffer_and_writer()",
+                "__M_writer(%s)"
+                % self.create_filter_callable(
+                    node.filter_args.args, "__M_buf.getvalue()", False
+                ),
+                None,
+            )
+
+    def visitCode(self, node):
+        if not node.ismodule:
+            self.printer.write_indented_block(
+                node.text, starting_lineno=node.lineno
+            )
+
+            if not self.in_def and len(self.identifiers.locally_assigned) > 0:
+                # if we are the "template" def, fudge locally
+                # declared/modified variables into the "__M_locals" dictionary,
+                # which is used for def calls within the same template,
+                # to simulate "enclosing scope"
+                self.printer.writeline(
+                    "__M_locals_builtin_stored = __M_locals_builtin()"
+                )
+                self.printer.writeline(
+                    "__M_locals.update(__M_dict_builtin([(__M_key,"
+                    " __M_locals_builtin_stored[__M_key]) for __M_key in"
+                    " [%s] if __M_key in __M_locals_builtin_stored]))"
+                    % ",".join([repr(x) for x in node.declared_identifiers()])
+                )
+
+    def visitIncludeTag(self, node):
+        self.printer.start_source(node.lineno)
+        args = node.attributes.get("args")
+        if args:
+            self.printer.writeline(
+                "runtime._include_file(context, %s, _template_uri, %s)"
+                % (node.parsed_attributes["file"], args)
+            )
+        else:
+            self.printer.writeline(
+                "runtime._include_file(context, %s, _template_uri)"
+                % (node.parsed_attributes["file"])
+            )
+
+    def visitNamespaceTag(self, node):
+        pass
+
+    def visitDefTag(self, node):
+        pass
+
+    def visitBlockTag(self, node):
+        if node.is_anonymous:
+            self.printer.writeline("%s()" % node.funcname)
+        else:
+            nameargs = node.get_argument_expressions(as_call=True)
+            nameargs += ["**pageargs"]
+            self.printer.writeline(
+                "if 'parent' not in context._data or "
+                "not hasattr(context._data['parent'], '%s'):" % node.funcname
+            )
+            self.printer.writeline(
+                "context['self'].%s(%s)" % (node.funcname, ",".join(nameargs))
+            )
+            self.printer.writeline("\n")
+
+    def visitCallNamespaceTag(self, node):
+        # TODO: we can put namespace-specific checks here, such
+        # as ensure the given namespace will be imported,
+        # pre-import the namespace, etc.
+        self.visitCallTag(node)
+
+    def visitCallTag(self, node):
+        self.printer.writeline("def ccall(caller):")
+        export = ["body"]
+        callable_identifiers = self.identifiers.branch(node, nested=True)
+        body_identifiers = callable_identifiers.branch(node, nested=False)
+        # we want the 'caller' passed to ccall to be used
+        # for the body() function, but for other non-body()
+        # <%def>s within <%call> we want the current caller
+        # off the call stack (if any)
+        body_identifiers.add_declared("caller")
+
+        self.identifier_stack.append(body_identifiers)
+
+        class DefVisitor:
+            def visitDefTag(s, node):
+                s.visitDefOrBase(node)
+
+            def visitBlockTag(s, node):
+                s.visitDefOrBase(node)
+
+            def visitDefOrBase(s, node):
+                self.write_inline_def(node, callable_identifiers, nested=False)
+                if not node.is_anonymous:
+                    export.append(node.funcname)
+                # remove defs that are within the <%call> from the
+                # "closuredefs" defined in the body, so they dont render twice
+                if node.funcname in body_identifiers.closuredefs:
+                    del body_identifiers.closuredefs[node.funcname]
+
+        vis = DefVisitor()
+        for n in node.nodes:
+            n.accept_visitor(vis)
+        self.identifier_stack.pop()
+
+        bodyargs = node.body_decl.get_argument_expressions()
+        self.printer.writeline("def body(%s):" % ",".join(bodyargs))
+
+        # TODO: figure out best way to specify
+        # buffering/nonbuffering (at call time would be better)
+        buffered = False
+        if buffered:
+            self.printer.writelines("context._push_buffer()", "try:")
+        self.write_variable_declares(body_identifiers)
+        self.identifier_stack.append(body_identifiers)
+
+        for n in node.nodes:
+            n.accept_visitor(self)
+        self.identifier_stack.pop()
+
+        self.write_def_finish(node, buffered, False, False, callstack=False)
+        self.printer.writelines(None, "return [%s]" % (",".join(export)), None)
+
+        self.printer.writelines(
+            # push on caller for nested call
+            "context.caller_stack.nextcaller = "
+            "runtime.Namespace('caller', context, "
+            "callables=ccall(__M_caller))",
+            "try:",
+        )
+        self.printer.start_source(node.lineno)
+        self.printer.writelines(
+            "__M_writer(%s)"
+            % self.create_filter_callable([], node.expression, True),
+            "finally:",
+            "context.caller_stack.nextcaller = None",
+            None,
+        )
+
+
+class _Identifiers:
+
+    """tracks the status of identifier names as template code is rendered."""
+
+    def __init__(self, compiler, node=None, parent=None, nested=False):
+        if parent is not None:
+            # if we are the branch created in write_namespaces(),
+            # we don't share any context from the main body().
+            if isinstance(node, parsetree.NamespaceTag):
+                self.declared = set()
+                self.topleveldefs = util.SetLikeDict()
+            else:
+                # things that have already been declared
+                # in an enclosing namespace (i.e. names we can just use)
+                self.declared = (
+                    set(parent.declared)
+                    .union([c.name for c in parent.closuredefs.values()])
+                    .union(parent.locally_declared)
+                    .union(parent.argument_declared)
+                )
+
+                # if these identifiers correspond to a "nested"
+                # scope, it means whatever the parent identifiers
+                # had as undeclared will have been declared by that parent,
+                # and therefore we have them in our scope.
+                if nested:
+                    self.declared = self.declared.union(parent.undeclared)
+
+                # top level defs that are available
+                self.topleveldefs = util.SetLikeDict(**parent.topleveldefs)
+        else:
+            self.declared = set()
+            self.topleveldefs = util.SetLikeDict()
+
+        self.compiler = compiler
+
+        # things within this level that are referenced before they
+        # are declared (e.g. assigned to)
+        self.undeclared = set()
+
+        # things that are declared locally.  some of these things
+        # could be in the "undeclared" list as well if they are
+        # referenced before declared
+        self.locally_declared = set()
+
+        # assignments made in explicit python blocks.
+        # these will be propagated to
+        # the context of local def calls.
+        self.locally_assigned = set()
+
+        # things that are declared in the argument
+        # signature of the def callable
+        self.argument_declared = set()
+
+        # closure defs that are defined in this level
+        self.closuredefs = util.SetLikeDict()
+
+        self.node = node
+
+        if node is not None:
+            node.accept_visitor(self)
+
+        illegal_names = self.compiler.reserved_names.intersection(
+            self.locally_declared
+        )
+        if illegal_names:
+            raise exceptions.NameConflictError(
+                "Reserved words declared in template: %s"
+                % ", ".join(illegal_names)
+            )
+
+    def branch(self, node, **kwargs):
+        """create a new Identifiers for a new Node, with
+        this Identifiers as the parent."""
+
+        return _Identifiers(self.compiler, node, self, **kwargs)
+
+    @property
+    def defs(self):
+        return set(self.topleveldefs.union(self.closuredefs).values())
+
+    def __repr__(self):
+        return (
+            "Identifiers(declared=%r, locally_declared=%r, "
+            "undeclared=%r, topleveldefs=%r, closuredefs=%r, "
+            "argumentdeclared=%r)"
+            % (
+                list(self.declared),
+                list(self.locally_declared),
+                list(self.undeclared),
+                [c.name for c in self.topleveldefs.values()],
+                [c.name for c in self.closuredefs.values()],
+                self.argument_declared,
+            )
+        )
+
+    def check_declared(self, node):
+        """update the state of this Identifiers with the undeclared
+        and declared identifiers of the given node."""
+
+        for ident in node.undeclared_identifiers():
+            if ident != "context" and ident not in self.declared.union(
+                self.locally_declared
+            ):
+                self.undeclared.add(ident)
+        for ident in node.declared_identifiers():
+            self.locally_declared.add(ident)
+
+    def add_declared(self, ident):
+        self.declared.add(ident)
+        if ident in self.undeclared:
+            self.undeclared.remove(ident)
+
+    def visitExpression(self, node):
+        self.check_declared(node)
+
+    def visitControlLine(self, node):
+        self.check_declared(node)
+
+    def visitCode(self, node):
+        if not node.ismodule:
+            self.check_declared(node)
+            self.locally_assigned = self.locally_assigned.union(
+                node.declared_identifiers()
+            )
+
+    def visitNamespaceTag(self, node):
+        # only traverse into the sub-elements of a
+        # <%namespace> tag if we are the branch created in
+        # write_namespaces()
+        if self.node is node:
+            for n in node.nodes:
+                n.accept_visitor(self)
+
+    def _check_name_exists(self, collection, node):
+        existing = collection.get(node.funcname)
+        collection[node.funcname] = node
+        if (
+            existing is not None
+            and existing is not node
+            and (node.is_block or existing.is_block)
+        ):
+            raise exceptions.CompileException(
+                "%%def or %%block named '%s' already "
+                "exists in this template." % node.funcname,
+                **node.exception_kwargs,
+            )
+
+    def visitDefTag(self, node):
+        if node.is_root() and not node.is_anonymous:
+            self._check_name_exists(self.topleveldefs, node)
+        elif node is not self.node:
+            self._check_name_exists(self.closuredefs, node)
+
+        for ident in node.undeclared_identifiers():
+            if ident != "context" and ident not in self.declared.union(
+                self.locally_declared
+            ):
+                self.undeclared.add(ident)
+
+        # visit defs only one level deep
+        if node is self.node:
+            for ident in node.declared_identifiers():
+                self.argument_declared.add(ident)
+
+            for n in node.nodes:
+                n.accept_visitor(self)
+
+    def visitBlockTag(self, node):
+        if node is not self.node and not node.is_anonymous:
+            if isinstance(self.node, parsetree.DefTag):
+                raise exceptions.CompileException(
+                    "Named block '%s' not allowed inside of def '%s'"
+                    % (node.name, self.node.name),
+                    **node.exception_kwargs,
+                )
+            elif isinstance(
+                self.node, (parsetree.CallTag, parsetree.CallNamespaceTag)
+            ):
+                raise exceptions.CompileException(
+                    "Named block '%s' not allowed inside of <%%call> tag"
+                    % (node.name,),
+                    **node.exception_kwargs,
+                )
+
+        for ident in node.undeclared_identifiers():
+            if ident != "context" and ident not in self.declared.union(
+                self.locally_declared
+            ):
+                self.undeclared.add(ident)
+
+        if not node.is_anonymous:
+            self._check_name_exists(self.topleveldefs, node)
+            self.undeclared.add(node.funcname)
+        elif node is not self.node:
+            self._check_name_exists(self.closuredefs, node)
+        for ident in node.declared_identifiers():
+            self.argument_declared.add(ident)
+        for n in node.nodes:
+            n.accept_visitor(self)
+
+    def visitTextTag(self, node):
+        for ident in node.undeclared_identifiers():
+            if ident != "context" and ident not in self.declared.union(
+                self.locally_declared
+            ):
+                self.undeclared.add(ident)
+
+    def visitIncludeTag(self, node):
+        self.check_declared(node)
+
+    def visitPageTag(self, node):
+        for ident in node.declared_identifiers():
+            self.argument_declared.add(ident)
+        self.check_declared(node)
+
+    def visitCallNamespaceTag(self, node):
+        self.visitCallTag(node)
+
+    def visitCallTag(self, node):
+        if node is self.node:
+            for ident in node.undeclared_identifiers():
+                if ident != "context" and ident not in self.declared.union(
+                    self.locally_declared
+                ):
+                    self.undeclared.add(ident)
+            for ident in node.declared_identifiers():
+                self.argument_declared.add(ident)
+            for n in node.nodes:
+                n.accept_visitor(self)
+        else:
+            for ident in node.undeclared_identifiers():
+                if ident != "context" and ident not in self.declared.union(
+                    self.locally_declared
+                ):
+                    self.undeclared.add(ident)
+
+
+_FOR_LOOP = re.compile(
+    r"^for\s+((?:\(?)\s*"
+    r"(?:\(?)\s*[A-Za-z_][A-Za-z_0-9]*"
+    r"(?:\s*,\s*(?:[A-Za-z_][A-Za-z_0-9]*),??)*\s*(?:\)?)"
+    r"(?:\s*,\s*(?:"
+    r"(?:\(?)\s*[A-Za-z_][A-Za-z_0-9]*"
+    r"(?:\s*,\s*(?:[A-Za-z_][A-Za-z_0-9]*),??)*\s*(?:\)?)"
+    r"),??)*\s*(?:\)?))\s+in\s+(.*):"
+)
+
+
+def mangle_mako_loop(node, printer):
+    """converts a for loop into a context manager wrapped around a for loop
+    when access to the `loop` variable has been detected in the for loop body
+    """
+    loop_variable = LoopVariable()
+    node.accept_visitor(loop_variable)
+    if loop_variable.detected:
+        node.nodes[-1].has_loop_context = True
+        match = _FOR_LOOP.match(node.text)
+        if match:
+            printer.writelines(
+                "loop = __M_loop._enter(%s)" % match.group(2),
+                "try:"
+                # 'with __M_loop(%s) as loop:' % match.group(2)
+            )
+            text = "for %s in loop:" % match.group(1)
+        else:
+            raise SyntaxError("Couldn't apply loop context: %s" % node.text)
+    else:
+        text = node.text
+    return text
+
+
+class LoopVariable:
+
+    """A node visitor which looks for the name 'loop' within undeclared
+    identifiers."""
+
+    def __init__(self):
+        self.detected = False
+
+    def _loop_reference_detected(self, node):
+        if "loop" in node.undeclared_identifiers():
+            self.detected = True
+        else:
+            for n in node.get_children():
+                n.accept_visitor(self)
+
+    def visitControlLine(self, node):
+        self._loop_reference_detected(node)
+
+    def visitCode(self, node):
+        self._loop_reference_detected(node)
+
+    def visitExpression(self, node):
+        self._loop_reference_detected(node)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/compat.py b/backend/venv/lib/python3.12/site-packages/mako/compat.py
new file mode 100644
index 00000000..1c176012
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/compat.py
@@ -0,0 +1,70 @@
+# mako/compat.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+import collections
+from importlib import metadata as importlib_metadata
+from importlib import util
+import inspect
+import sys
+
+win32 = sys.platform.startswith("win")
+pypy = hasattr(sys, "pypy_version_info")
+
+ArgSpec = collections.namedtuple(
+    "ArgSpec", ["args", "varargs", "keywords", "defaults"]
+)
+
+
+def inspect_getargspec(func):
+    """getargspec based on fully vendored getfullargspec from Python 3.3."""
+
+    if inspect.ismethod(func):
+        func = func.__func__
+    if not inspect.isfunction(func):
+        raise TypeError(f"{func!r} is not a Python function")
+
+    co = func.__code__
+    if not inspect.iscode(co):
+        raise TypeError(f"{co!r} is not a code object")
+
+    nargs = co.co_argcount
+    names = co.co_varnames
+    nkwargs = co.co_kwonlyargcount
+    args = list(names[:nargs])
+
+    nargs += nkwargs
+    varargs = None
+    if co.co_flags & inspect.CO_VARARGS:
+        varargs = co.co_varnames[nargs]
+        nargs = nargs + 1
+    varkw = None
+    if co.co_flags & inspect.CO_VARKEYWORDS:
+        varkw = co.co_varnames[nargs]
+
+    return ArgSpec(args, varargs, varkw, func.__defaults__)
+
+
+def load_module(module_id, path):
+    spec = util.spec_from_file_location(module_id, path)
+    module = util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+def exception_as():
+    return sys.exc_info()[1]
+
+
+def exception_name(exc):
+    return exc.__class__.__name__
+
+
+def importlib_metadata_get(group):
+    ep = importlib_metadata.entry_points()
+    if hasattr(ep, "select"):
+        return ep.select(group=group)
+    else:
+        return ep.get(group, ())
diff --git a/backend/venv/lib/python3.12/site-packages/mako/exceptions.py b/backend/venv/lib/python3.12/site-packages/mako/exceptions.py
new file mode 100644
index 00000000..6aeb2d27
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/exceptions.py
@@ -0,0 +1,417 @@
+# mako/exceptions.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""exception classes"""
+
+import sys
+import traceback
+
+from mako import compat
+from mako import util
+
+
+class MakoException(Exception):
+    pass
+
+
+class RuntimeException(MakoException):
+    pass
+
+
+def _format_filepos(lineno, pos, filename):
+    if filename is None:
+        return " at line: %d char: %d" % (lineno, pos)
+    else:
+        return " in file '%s' at line: %d char: %d" % (filename, lineno, pos)
+
+
+class CompileException(MakoException):
+    def __init__(self, message, source, lineno, pos, filename):
+        MakoException.__init__(
+            self, message + _format_filepos(lineno, pos, filename)
+        )
+        self.lineno = lineno
+        self.pos = pos
+        self.filename = filename
+        self.source = source
+
+
+class SyntaxException(MakoException):
+    def __init__(self, message, source, lineno, pos, filename):
+        MakoException.__init__(
+            self, message + _format_filepos(lineno, pos, filename)
+        )
+        self.lineno = lineno
+        self.pos = pos
+        self.filename = filename
+        self.source = source
+
+
+class UnsupportedError(MakoException):
+
+    """raised when a retired feature is used."""
+
+
+class NameConflictError(MakoException):
+
+    """raised when a reserved word is used inappropriately"""
+
+
+class TemplateLookupException(MakoException):
+    pass
+
+
+class TopLevelLookupException(TemplateLookupException):
+    pass
+
+
+class RichTraceback:
+
+    """Pull the current exception from the ``sys`` traceback and extracts
+    Mako-specific template information.
+
+    See the usage examples in :ref:`handling_exceptions`.
+
+    """
+
+    def __init__(self, error=None, traceback=None):
+        self.source, self.lineno = "", 0
+
+        if error is None or traceback is None:
+            t, value, tback = sys.exc_info()
+
+        if error is None:
+            error = value or t
+
+        if traceback is None:
+            traceback = tback
+
+        self.error = error
+        self.records = self._init(traceback)
+
+        if isinstance(self.error, (CompileException, SyntaxException)):
+            self.source = self.error.source
+            self.lineno = self.error.lineno
+            self._has_source = True
+
+        self._init_message()
+
+    @property
+    def errorname(self):
+        return compat.exception_name(self.error)
+
+    def _init_message(self):
+        """Find a unicode representation of self.error"""
+        try:
+            self.message = str(self.error)
+        except UnicodeError:
+            try:
+                self.message = str(self.error)
+            except UnicodeEncodeError:
+                # Fallback to args as neither unicode nor
+                # str(Exception(u'\xe6')) work in Python < 2.6
+                self.message = self.error.args[0]
+        if not isinstance(self.message, str):
+            self.message = str(self.message, "ascii", "replace")
+
+    def _get_reformatted_records(self, records):
+        for rec in records:
+            if rec[6] is not None:
+                yield (rec[4], rec[5], rec[2], rec[6])
+            else:
+                yield tuple(rec[0:4])
+
+    @property
+    def traceback(self):
+        """Return a list of 4-tuple traceback records (i.e. normal python
+        format) with template-corresponding lines remapped to the originating
+        template.
+
+        """
+        return list(self._get_reformatted_records(self.records))
+
+    @property
+    def reverse_records(self):
+        return reversed(self.records)
+
+    @property
+    def reverse_traceback(self):
+        """Return the same data as traceback, except in reverse order."""
+
+        return list(self._get_reformatted_records(self.reverse_records))
+
+    def _init(self, trcback):
+        """format a traceback from sys.exc_info() into 7-item tuples,
+        containing the regular four traceback tuple items, plus the original
+        template filename, the line number adjusted relative to the template
+        source, and code line from that line number of the template."""
+
+        import mako.template
+
+        mods = {}
+        rawrecords = traceback.extract_tb(trcback)
+        new_trcback = []
+        for filename, lineno, function, line in rawrecords:
+            if not line:
+                line = ""
+            try:
+                (line_map, template_lines, template_filename) = mods[filename]
+            except KeyError:
+                try:
+                    info = mako.template._get_module_info(filename)
+                    module_source = info.code
+                    template_source = info.source
+                    template_filename = (
+                        info.template_filename or info.template_uri or filename
+                    )
+                except KeyError:
+                    # A normal .py file (not a Template)
+                    new_trcback.append(
+                        (
+                            filename,
+                            lineno,
+                            function,
+                            line,
+                            None,
+                            None,
+                            None,
+                            None,
+                        )
+                    )
+                    continue
+
+                template_ln = 1
+
+                mtm = mako.template.ModuleInfo
+                source_map = mtm.get_module_source_metadata(
+                    module_source, full_line_map=True
+                )
+                line_map = source_map["full_line_map"]
+
+                template_lines = [
+                    line_ for line_ in template_source.split("\n")
+                ]
+                mods[filename] = (line_map, template_lines, template_filename)
+
+            template_ln = line_map[lineno - 1]
+
+            if template_ln <= len(template_lines):
+                template_line = template_lines[template_ln - 1]
+            else:
+                template_line = None
+            new_trcback.append(
+                (
+                    filename,
+                    lineno,
+                    function,
+                    line,
+                    template_filename,
+                    template_ln,
+                    template_line,
+                    template_source,
+                )
+            )
+        if not self.source:
+            for l in range(len(new_trcback) - 1, 0, -1):
+                if new_trcback[l][5]:
+                    self.source = new_trcback[l][7]
+                    self.lineno = new_trcback[l][5]
+                    break
+            else:
+                if new_trcback:
+                    try:
+                        # A normal .py file (not a Template)
+                        with open(new_trcback[-1][0], "rb") as fp:
+                            encoding = util.parse_encoding(fp)
+                            if not encoding:
+                                encoding = "utf-8"
+                            fp.seek(0)
+                            self.source = fp.read()
+                        if encoding:
+                            self.source = self.source.decode(encoding)
+                    except IOError:
+                        self.source = ""
+                    self.lineno = new_trcback[-1][1]
+        return new_trcback
+
+
+def text_error_template(lookup=None):
+    """Provides a template that renders a stack trace in a similar format to
+    the Python interpreter, substituting source template filenames, line
+    numbers and code for that of the originating source template, as
+    applicable.
+
+    """
+    import mako.template
+
+    return mako.template.Template(
+        r"""
+<%page args="error=None, traceback=None"/>
+<%!
+    from mako.exceptions import RichTraceback
+%>\
+<%
+    tback = RichTraceback(error=error, traceback=traceback)
+%>\
+Traceback (most recent call last):
+% for (filename, lineno, function, line) in tback.traceback:
+  File "${filename}", line ${lineno}, in ${function or '?'}
+    ${line | trim}
+% endfor
+${tback.errorname}: ${tback.message}
+"""
+    )
+
+
+def _install_pygments():
+    global syntax_highlight, pygments_html_formatter
+    from mako.ext.pygmentplugin import syntax_highlight  # noqa
+    from mako.ext.pygmentplugin import pygments_html_formatter  # noqa
+
+
+def _install_fallback():
+    global syntax_highlight, pygments_html_formatter
+    from mako.filters import html_escape
+
+    pygments_html_formatter = None
+
+    def syntax_highlight(filename="", language=None):
+        return html_escape
+
+
+def _install_highlighting():
+    try:
+        _install_pygments()
+    except ImportError:
+        _install_fallback()
+
+
+_install_highlighting()
+
+
+def html_error_template():
+    """Provides a template that renders a stack trace in an HTML format,
+    providing an excerpt of code as well as substituting source template
+    filenames, line numbers and code for that of the originating source
+    template, as applicable.
+
+    The template's default ``encoding_errors`` value is
+    ``'htmlentityreplace'``. The template has two options. With the
+    ``full`` option disabled, only a section of an HTML document is
+    returned. With the ``css`` option disabled, the default stylesheet
+    won't be included.
+
+    """
+    import mako.template
+
+    return mako.template.Template(
+        r"""
+<%!
+    from mako.exceptions import RichTraceback, syntax_highlight,\
+            pygments_html_formatter
+%>
+<%page args="full=True, css=True, error=None, traceback=None"/>
+% if full:
+<html>
+<head>
+    <title>Mako Runtime Error</title>
+% endif
+% if css:
+    <style>
+        body { font-family:verdana; margin:10px 30px 10px 30px;}
+        .stacktrace { margin:5px 5px 5px 5px; }
+        .highlight { padding:0px 10px 0px 10px; background-color:#9F9FDF; }
+        .nonhighlight { padding:0px; background-color:#DFDFDF; }
+        .sample { padding:10px; margin:10px 10px 10px 10px;
+                  font-family:monospace; }
+        .sampleline { padding:0px 10px 0px 10px; }
+        .sourceline { margin:5px 5px 10px 5px; font-family:monospace;}
+        .location { font-size:80%; }
+        .highlight { white-space:pre; }
+        .sampleline { white-space:pre; }
+
+    % if pygments_html_formatter:
+        ${pygments_html_formatter.get_style_defs()}
+        .linenos { min-width: 2.5em; text-align: right; }
+        pre { margin: 0; }
+        .syntax-highlighted { padding: 0 10px; }
+        .syntax-highlightedtable { border-spacing: 1px; }
+        .nonhighlight { border-top: 1px solid #DFDFDF;
+                        border-bottom: 1px solid #DFDFDF; }
+        .stacktrace .nonhighlight { margin: 5px 15px 10px; }
+        .sourceline { margin: 0 0; font-family:monospace; }
+        .code { background-color: #F8F8F8; width: 100%; }
+        .error .code { background-color: #FFBDBD; }
+        .error .syntax-highlighted { background-color: #FFBDBD; }
+    % endif
+
+    </style>
+% endif
+% if full:
+</head>
+<body>
+% endif
+
+<h2>Error !</h2>
+<%
+    tback = RichTraceback(error=error, traceback=traceback)
+    src = tback.source
+    line = tback.lineno
+    if src:
+        lines = src.split('\n')
+    else:
+        lines = None
+%>
+<h3>${tback.errorname}: ${tback.message|h}</h3>
+
+% if lines:
+    <div class="sample">
+    <div class="nonhighlight">
+% for index in range(max(0, line-4),min(len(lines), line+5)):
+    <%
+       if pygments_html_formatter:
+           pygments_html_formatter.linenostart = index + 1
+    %>
+    % if index + 1 == line:
+    <%
+       if pygments_html_formatter:
+           old_cssclass = pygments_html_formatter.cssclass
+           pygments_html_formatter.cssclass = 'error ' + old_cssclass
+    %>
+        ${lines[index] | syntax_highlight(language='mako')}
+    <%
+       if pygments_html_formatter:
+           pygments_html_formatter.cssclass = old_cssclass
+    %>
+    % else:
+        ${lines[index] | syntax_highlight(language='mako')}
+    % endif
+% endfor
+    </div>
+    </div>
+% endif
+
+<div class="stacktrace">
+% for (filename, lineno, function, line) in tback.reverse_traceback:
+    <div class="location">${filename}, line ${lineno}:</div>
+    <div class="nonhighlight">
+    <%
+       if pygments_html_formatter:
+           pygments_html_formatter.linenostart = lineno
+    %>
+      <div class="sourceline">${line | syntax_highlight(filename)}</div>
+    </div>
+% endfor
+</div>
+
+% if full:
+</body>
+</html>
+% endif
+""",
+        output_encoding=sys.getdefaultencoding(),
+        encoding_errors="htmlentityreplace",
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__init__.py b/backend/venv/lib/python3.12/site-packages/mako/ext/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..61005701
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/autohandler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/autohandler.cpython-312.pyc
new file mode 100644
index 00000000..5bf5b426
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/autohandler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/babelplugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/babelplugin.cpython-312.pyc
new file mode 100644
index 00000000..0dd0ce19
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/babelplugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/beaker_cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/beaker_cache.cpython-312.pyc
new file mode 100644
index 00000000..6c79113b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/beaker_cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/extract.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/extract.cpython-312.pyc
new file mode 100644
index 00000000..ad8e8f52
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/extract.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/linguaplugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/linguaplugin.cpython-312.pyc
new file mode 100644
index 00000000..de565f7f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/linguaplugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/preprocessors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/preprocessors.cpython-312.pyc
new file mode 100644
index 00000000..363698e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/preprocessors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/pygmentplugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/pygmentplugin.cpython-312.pyc
new file mode 100644
index 00000000..e8c6f747
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/pygmentplugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/turbogears.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/turbogears.cpython-312.pyc
new file mode 100644
index 00000000..dad0a383
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/ext/__pycache__/turbogears.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/autohandler.py b/backend/venv/lib/python3.12/site-packages/mako/ext/autohandler.py
new file mode 100644
index 00000000..829ed097
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/autohandler.py
@@ -0,0 +1,70 @@
+# ext/autohandler.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""adds autohandler functionality to Mako templates.
+
+requires that the TemplateLookup class is used with templates.
+
+usage::
+
+    <%!
+        from mako.ext.autohandler import autohandler
+    %>
+    <%inherit file="${autohandler(template, context)}"/>
+
+
+or with custom autohandler filename::
+
+    <%!
+        from mako.ext.autohandler import autohandler
+    %>
+    <%inherit file="${autohandler(template, context, name='somefilename')}"/>
+
+"""
+
+import os
+import posixpath
+import re
+
+
+def autohandler(template, context, name="autohandler"):
+    lookup = context.lookup
+    _template_uri = template.module._template_uri
+    if not lookup.filesystem_checks:
+        try:
+            return lookup._uri_cache[(autohandler, _template_uri, name)]
+        except KeyError:
+            pass
+
+    tokens = re.findall(r"([^/]+)", posixpath.dirname(_template_uri)) + [name]
+    while len(tokens):
+        path = "/" + "/".join(tokens)
+        if path != _template_uri and _file_exists(lookup, path):
+            if not lookup.filesystem_checks:
+                return lookup._uri_cache.setdefault(
+                    (autohandler, _template_uri, name), path
+                )
+            else:
+                return path
+        if len(tokens) == 1:
+            break
+        tokens[-2:] = [name]
+
+    if not lookup.filesystem_checks:
+        return lookup._uri_cache.setdefault(
+            (autohandler, _template_uri, name), None
+        )
+    else:
+        return None
+
+
+def _file_exists(lookup, path):
+    psub = re.sub(r"^/", "", path)
+    for d in lookup.directories:
+        if os.path.exists(d + "/" + psub):
+            return True
+    else:
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/babelplugin.py b/backend/venv/lib/python3.12/site-packages/mako/ext/babelplugin.py
new file mode 100644
index 00000000..541ea540
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/babelplugin.py
@@ -0,0 +1,57 @@
+# ext/babelplugin.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""gettext message extraction via Babel: https://pypi.org/project/Babel/"""
+from babel.messages.extract import extract_python
+
+from mako.ext.extract import MessageExtractor
+
+
+class BabelMakoExtractor(MessageExtractor):
+    def __init__(self, keywords, comment_tags, options):
+        self.keywords = keywords
+        self.options = options
+        self.config = {
+            "comment-tags": " ".join(comment_tags),
+            "encoding": options.get(
+                "input_encoding", options.get("encoding", None)
+            ),
+        }
+        super().__init__()
+
+    def __call__(self, fileobj):
+        return self.process_file(fileobj)
+
+    def process_python(self, code, code_lineno, translator_strings):
+        comment_tags = self.config["comment-tags"]
+        for (
+            lineno,
+            funcname,
+            messages,
+            python_translator_comments,
+        ) in extract_python(code, self.keywords, comment_tags, self.options):
+            yield (
+                code_lineno + (lineno - 1),
+                funcname,
+                messages,
+                translator_strings + python_translator_comments,
+            )
+
+
+def extract(fileobj, keywords, comment_tags, options):
+    """Extract messages from Mako templates.
+
+    :param fileobj: the file-like object the messages should be extracted from
+    :param keywords: a list of keywords (i.e. function names) that should be
+                     recognized as translation functions
+    :param comment_tags: a list of translator tags to search for and include
+                         in the results
+    :param options: a dictionary of additional options (optional)
+    :return: an iterator over ``(lineno, funcname, message, comments)`` tuples
+    :rtype: ``iterator``
+    """
+    extractor = BabelMakoExtractor(keywords, comment_tags, options)
+    yield from extractor(fileobj)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/beaker_cache.py b/backend/venv/lib/python3.12/site-packages/mako/ext/beaker_cache.py
new file mode 100644
index 00000000..b1392f30
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/beaker_cache.py
@@ -0,0 +1,82 @@
+# ext/beaker_cache.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""Provide a :class:`.CacheImpl` for the Beaker caching system."""
+
+from mako import exceptions
+from mako.cache import CacheImpl
+
+try:
+    from beaker import cache as beaker_cache
+except:
+    has_beaker = False
+else:
+    has_beaker = True
+
+_beaker_cache = None
+
+
+class BeakerCacheImpl(CacheImpl):
+
+    """A :class:`.CacheImpl` provided for the Beaker caching system.
+
+    This plugin is used by default, based on the default
+    value of ``'beaker'`` for the ``cache_impl`` parameter of the
+    :class:`.Template` or :class:`.TemplateLookup` classes.
+
+    """
+
+    def __init__(self, cache):
+        if not has_beaker:
+            raise exceptions.RuntimeException(
+                "Can't initialize Beaker plugin; Beaker is not installed."
+            )
+        global _beaker_cache
+        if _beaker_cache is None:
+            if "manager" in cache.template.cache_args:
+                _beaker_cache = cache.template.cache_args["manager"]
+            else:
+                _beaker_cache = beaker_cache.CacheManager()
+        super().__init__(cache)
+
+    def _get_cache(self, **kw):
+        expiretime = kw.pop("timeout", None)
+        if "dir" in kw:
+            kw["data_dir"] = kw.pop("dir")
+        elif self.cache.template.module_directory:
+            kw["data_dir"] = self.cache.template.module_directory
+
+        if "manager" in kw:
+            kw.pop("manager")
+
+        if kw.get("type") == "memcached":
+            kw["type"] = "ext:memcached"
+
+        if "region" in kw:
+            region = kw.pop("region")
+            cache = _beaker_cache.get_cache_region(self.cache.id, region, **kw)
+        else:
+            cache = _beaker_cache.get_cache(self.cache.id, **kw)
+        cache_args = {"starttime": self.cache.starttime}
+        if expiretime:
+            cache_args["expiretime"] = expiretime
+        return cache, cache_args
+
+    def get_or_create(self, key, creation_function, **kw):
+        cache, kw = self._get_cache(**kw)
+        return cache.get(key, createfunc=creation_function, **kw)
+
+    def put(self, key, value, **kw):
+        cache, kw = self._get_cache(**kw)
+        cache.put(key, value, **kw)
+
+    def get(self, key, **kw):
+        cache, kw = self._get_cache(**kw)
+        return cache.get(key, **kw)
+
+    def invalidate(self, key, **kw):
+        cache, kw = self._get_cache(**kw)
+        cache.remove_value(key, **kw)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/extract.py b/backend/venv/lib/python3.12/site-packages/mako/ext/extract.py
new file mode 100644
index 00000000..609c5c29
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/extract.py
@@ -0,0 +1,129 @@
+# ext/extract.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+from io import BytesIO
+from io import StringIO
+import re
+
+from mako import lexer
+from mako import parsetree
+
+
+class MessageExtractor:
+    use_bytes = True
+
+    def process_file(self, fileobj):
+        template_node = lexer.Lexer(
+            fileobj.read(), input_encoding=self.config["encoding"]
+        ).parse()
+        yield from self.extract_nodes(template_node.get_children())
+
+    def extract_nodes(self, nodes):
+        translator_comments = []
+        in_translator_comments = False
+        input_encoding = self.config["encoding"] or "ascii"
+        comment_tags = list(
+            filter(None, re.split(r"\s+", self.config["comment-tags"]))
+        )
+
+        for node in nodes:
+            child_nodes = None
+            if (
+                in_translator_comments
+                and isinstance(node, parsetree.Text)
+                and not node.content.strip()
+            ):
+                # Ignore whitespace within translator comments
+                continue
+
+            if isinstance(node, parsetree.Comment):
+                value = node.text.strip()
+                if in_translator_comments:
+                    translator_comments.extend(
+                        self._split_comment(node.lineno, value)
+                    )
+                    continue
+                for comment_tag in comment_tags:
+                    if value.startswith(comment_tag):
+                        in_translator_comments = True
+                        translator_comments.extend(
+                            self._split_comment(node.lineno, value)
+                        )
+                continue
+
+            if isinstance(node, parsetree.DefTag):
+                code = node.function_decl.code
+                child_nodes = node.nodes
+            elif isinstance(node, parsetree.BlockTag):
+                code = node.body_decl.code
+                child_nodes = node.nodes
+            elif isinstance(node, parsetree.CallTag):
+                code = node.code.code
+                child_nodes = node.nodes
+            elif isinstance(node, parsetree.PageTag):
+                code = node.body_decl.code
+            elif isinstance(node, parsetree.CallNamespaceTag):
+                code = node.expression
+                child_nodes = node.nodes
+            elif isinstance(node, parsetree.ControlLine):
+                if node.isend:
+                    in_translator_comments = False
+                    continue
+                code = node.text
+            elif isinstance(node, parsetree.Code):
+                in_translator_comments = False
+                code = node.code.code
+            elif isinstance(node, parsetree.Expression):
+                code = node.code.code
+            else:
+                continue
+
+            # Comments don't apply unless they immediately precede the message
+            if (
+                translator_comments
+                and translator_comments[-1][0] < node.lineno - 1
+            ):
+                translator_comments = []
+
+            translator_strings = [
+                comment[1] for comment in translator_comments
+            ]
+
+            if isinstance(code, str) and self.use_bytes:
+                code = code.encode(input_encoding, "backslashreplace")
+
+            used_translator_comments = False
+            # We add extra newline to work around a pybabel bug
+            # (see python-babel/babel#274, parse_encoding dies if the first
+            # input string of the input is non-ascii)
+            # Also, because we added it, we have to subtract one from
+            # node.lineno
+            if self.use_bytes:
+                code = BytesIO(b"\n" + code)
+            else:
+                code = StringIO("\n" + code)
+
+            for message in self.process_python(
+                code, node.lineno - 1, translator_strings
+            ):
+                yield message
+                used_translator_comments = True
+
+            if used_translator_comments:
+                translator_comments = []
+            in_translator_comments = False
+
+            if child_nodes:
+                yield from self.extract_nodes(child_nodes)
+
+    @staticmethod
+    def _split_comment(lineno, comment):
+        """Return the multiline comment at lineno split into a list of
+        comment line numbers and the accompanying comment line"""
+        return [
+            (lineno + index, line)
+            for index, line in enumerate(comment.splitlines())
+        ]
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/linguaplugin.py b/backend/venv/lib/python3.12/site-packages/mako/ext/linguaplugin.py
new file mode 100644
index 00000000..118668de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/linguaplugin.py
@@ -0,0 +1,57 @@
+# ext/linguaplugin.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+import contextlib
+import io
+
+from lingua.extractors import Extractor
+from lingua.extractors import get_extractor
+from lingua.extractors import Message
+
+from mako.ext.extract import MessageExtractor
+
+
+class LinguaMakoExtractor(Extractor, MessageExtractor):
+    """Mako templates"""
+
+    use_bytes = False
+    extensions = [".mako"]
+    default_config = {"encoding": "utf-8", "comment-tags": ""}
+
+    def __call__(self, filename, options, fileobj=None):
+        self.options = options
+        self.filename = filename
+        self.python_extractor = get_extractor("x.py")
+        if fileobj is None:
+            ctx = open(filename, "r")
+        else:
+            ctx = contextlib.nullcontext(fileobj)
+        with ctx as file_:
+            yield from self.process_file(file_)
+
+    def process_python(self, code, code_lineno, translator_strings):
+        source = code.getvalue().strip()
+        if source.endswith(":"):
+            if source in ("try:", "else:") or source.startswith("except"):
+                source = ""  # Ignore try/except and else
+            elif source.startswith("elif"):
+                source = source[2:]  # Replace "elif" with "if"
+            source += "pass"
+        code = io.StringIO(source)
+        for msg in self.python_extractor(
+            self.filename, self.options, code, code_lineno - 1
+        ):
+            if translator_strings:
+                msg = Message(
+                    msg.msgctxt,
+                    msg.msgid,
+                    msg.msgid_plural,
+                    msg.flags,
+                    " ".join(translator_strings + [msg.comment]),
+                    msg.tcomment,
+                    msg.location,
+                )
+            yield msg
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/preprocessors.py b/backend/venv/lib/python3.12/site-packages/mako/ext/preprocessors.py
new file mode 100644
index 00000000..21f2db54
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/preprocessors.py
@@ -0,0 +1,20 @@
+# ext/preprocessors.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""preprocessing functions, used with the 'preprocessor'
+argument on Template, TemplateLookup"""
+
+import re
+
+
+def convert_comments(text):
+    """preprocess old style comments.
+
+    example:
+
+    from mako.ext.preprocessors import convert_comments
+    t = Template(..., preprocessor=convert_comments)"""
+    return re.sub(r"(?<=\n)\s*#[^#]", "##", text)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/pygmentplugin.py b/backend/venv/lib/python3.12/site-packages/mako/ext/pygmentplugin.py
new file mode 100644
index 00000000..74a8fb9d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/pygmentplugin.py
@@ -0,0 +1,150 @@
+# ext/pygmentplugin.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+from pygments import highlight
+from pygments.formatters.html import HtmlFormatter
+from pygments.lexer import bygroups
+from pygments.lexer import DelegatingLexer
+from pygments.lexer import include
+from pygments.lexer import RegexLexer
+from pygments.lexer import using
+from pygments.lexers.agile import Python3Lexer
+from pygments.lexers.agile import PythonLexer
+from pygments.lexers.web import CssLexer
+from pygments.lexers.web import HtmlLexer
+from pygments.lexers.web import JavascriptLexer
+from pygments.lexers.web import XmlLexer
+from pygments.token import Comment
+from pygments.token import Keyword
+from pygments.token import Name
+from pygments.token import Operator
+from pygments.token import Other
+from pygments.token import String
+from pygments.token import Text
+
+
+class MakoLexer(RegexLexer):
+    name = "Mako"
+    aliases = ["mako"]
+    filenames = ["*.mao"]
+
+    tokens = {
+        "root": [
+            (
+                r"(\s*)(\%)(\s*end(?:\w+))(\n|\Z)",
+                bygroups(Text, Comment.Preproc, Keyword, Other),
+            ),
+            (
+                r"(\s*)(\%(?!%))([^\n]*)(\n|\Z)",
+                bygroups(Text, Comment.Preproc, using(PythonLexer), Other),
+            ),
+            (
+                r"(\s*)(##[^\n]*)(\n|\Z)",
+                bygroups(Text, Comment.Preproc, Other),
+            ),
+            (r"""(?s)<%doc>.*?</%doc>""", Comment.Preproc),
+            (
+                r"(<%)([\w\.\:]+)",
+                bygroups(Comment.Preproc, Name.Builtin),
+                "tag",
+            ),
+            (
+                r"(</%)([\w\.\:]+)(>)",
+                bygroups(Comment.Preproc, Name.Builtin, Comment.Preproc),
+            ),
+            (r"<%(?=([\w\.\:]+))", Comment.Preproc, "ondeftags"),
+            (
+                r"(?s)(<%(?:!?))(.*?)(%>)",
+                bygroups(Comment.Preproc, using(PythonLexer), Comment.Preproc),
+            ),
+            (
+                r"(\$\{)(.*?)(\})",
+                bygroups(Comment.Preproc, using(PythonLexer), Comment.Preproc),
+            ),
+            (
+                r"""(?sx)
+                (.+?)               # anything, followed by:
+                (?:
+                 (?<=\n)(?=%(?!%)|\#\#) |  # an eval or comment line
+                 (?=\#\*) |          # multiline comment
+                 (?=</?%) |         # a python block
+                                    # call start or end
+                 (?=\$\{) |         # a substitution
+                 (?<=\n)(?=\s*%) |
+                                    # - don't consume
+                 (\\\n) |           # an escaped newline
+                 \Z                 # end of string
+                )
+            """,
+                bygroups(Other, Operator),
+            ),
+            (r"\s+", Text),
+        ],
+        "ondeftags": [
+            (r"<%", Comment.Preproc),
+            (r"(?<=<%)(include|inherit|namespace|page)", Name.Builtin),
+            include("tag"),
+        ],
+        "tag": [
+            (r'((?:\w+)\s*=)\s*(".*?")', bygroups(Name.Attribute, String)),
+            (r"/?\s*>", Comment.Preproc, "#pop"),
+            (r"\s+", Text),
+        ],
+        "attr": [
+            ('".*?"', String, "#pop"),
+            ("'.*?'", String, "#pop"),
+            (r"[^\s>]+", String, "#pop"),
+        ],
+    }
+
+
+class MakoHtmlLexer(DelegatingLexer):
+    name = "HTML+Mako"
+    aliases = ["html+mako"]
+
+    def __init__(self, **options):
+        super().__init__(HtmlLexer, MakoLexer, **options)
+
+
+class MakoXmlLexer(DelegatingLexer):
+    name = "XML+Mako"
+    aliases = ["xml+mako"]
+
+    def __init__(self, **options):
+        super().__init__(XmlLexer, MakoLexer, **options)
+
+
+class MakoJavascriptLexer(DelegatingLexer):
+    name = "JavaScript+Mako"
+    aliases = ["js+mako", "javascript+mako"]
+
+    def __init__(self, **options):
+        super().__init__(JavascriptLexer, MakoLexer, **options)
+
+
+class MakoCssLexer(DelegatingLexer):
+    name = "CSS+Mako"
+    aliases = ["css+mako"]
+
+    def __init__(self, **options):
+        super().__init__(CssLexer, MakoLexer, **options)
+
+
+pygments_html_formatter = HtmlFormatter(
+    cssclass="syntax-highlighted", linenos=True
+)
+
+
+def syntax_highlight(filename="", language=None):
+    mako_lexer = MakoLexer()
+    python_lexer = Python3Lexer()
+    if filename.startswith("memory:") or language == "mako":
+        return lambda string: highlight(
+            string, mako_lexer, pygments_html_formatter
+        )
+    return lambda string: highlight(
+        string, python_lexer, pygments_html_formatter
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/mako/ext/turbogears.py b/backend/venv/lib/python3.12/site-packages/mako/ext/turbogears.py
new file mode 100644
index 00000000..25ad35e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/ext/turbogears.py
@@ -0,0 +1,61 @@
+# ext/turbogears.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+from mako import compat
+from mako.lookup import TemplateLookup
+from mako.template import Template
+
+
+class TGPlugin:
+
+    """TurboGears compatible Template Plugin."""
+
+    def __init__(self, extra_vars_func=None, options=None, extension="mak"):
+        self.extra_vars_func = extra_vars_func
+        self.extension = extension
+        if not options:
+            options = {}
+
+        # Pull the options out and initialize the lookup
+        lookup_options = {}
+        for k, v in options.items():
+            if k.startswith("mako."):
+                lookup_options[k[5:]] = v
+            elif k in ["directories", "filesystem_checks", "module_directory"]:
+                lookup_options[k] = v
+        self.lookup = TemplateLookup(**lookup_options)
+
+        self.tmpl_options = {}
+        # transfer lookup args to template args, based on those available
+        # in getargspec
+        for kw in compat.inspect_getargspec(Template.__init__)[0]:
+            if kw in lookup_options:
+                self.tmpl_options[kw] = lookup_options[kw]
+
+    def load_template(self, templatename, template_string=None):
+        """Loads a template from a file or a string"""
+        if template_string is not None:
+            return Template(template_string, **self.tmpl_options)
+        # Translate TG dot notation to normal / template path
+        if "/" not in templatename:
+            templatename = (
+                "/" + templatename.replace(".", "/") + "." + self.extension
+            )
+
+        # Lookup template
+        return self.lookup.get_template(templatename)
+
+    def render(
+        self, info, format="html", fragment=False, template=None  # noqa
+    ):
+        if isinstance(template, str):
+            template = self.load_template(template)
+
+        # Load extra vars func if provided
+        if self.extra_vars_func:
+            info.update(self.extra_vars_func())
+
+        return template.render(**info)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/filters.py b/backend/venv/lib/python3.12/site-packages/mako/filters.py
new file mode 100644
index 00000000..d5338bd9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/filters.py
@@ -0,0 +1,163 @@
+# mako/filters.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+
+import codecs
+from html.entities import codepoint2name
+from html.entities import name2codepoint
+import re
+from urllib.parse import quote_plus
+
+import markupsafe
+
+html_escape = markupsafe.escape
+
+xml_escapes = {
+    "&": "&amp;",
+    ">": "&gt;",
+    "<": "&lt;",
+    '"': "&#34;",  # also &quot; in html-only
+    "'": "&#39;",  # also &apos; in html-only
+}
+
+
+def xml_escape(string):
+    return re.sub(r'([&<"\'>])', lambda m: xml_escapes[m.group()], string)
+
+
+def url_escape(string):
+    # convert into a list of octets
+    string = string.encode("utf8")
+    return quote_plus(string)
+
+
+def trim(string):
+    return string.strip()
+
+
+class Decode:
+    def __getattr__(self, key):
+        def decode(x):
+            if isinstance(x, str):
+                return x
+            elif not isinstance(x, bytes):
+                return decode(str(x))
+            else:
+                return str(x, encoding=key)
+
+        return decode
+
+
+decode = Decode()
+
+
+class XMLEntityEscaper:
+    def __init__(self, codepoint2name, name2codepoint):
+        self.codepoint2entity = {
+            c: str("&%s;" % n) for c, n in codepoint2name.items()
+        }
+        self.name2codepoint = name2codepoint
+
+    def escape_entities(self, text):
+        """Replace characters with their character entity references.
+
+        Only characters corresponding to a named entity are replaced.
+        """
+        return str(text).translate(self.codepoint2entity)
+
+    def __escape(self, m):
+        codepoint = ord(m.group())
+        try:
+            return self.codepoint2entity[codepoint]
+        except (KeyError, IndexError):
+            return "&#x%X;" % codepoint
+
+    __escapable = re.compile(r'["&<>]|[^\x00-\x7f]')
+
+    def escape(self, text):
+        """Replace characters with their character references.
+
+        Replace characters by their named entity references.
+        Non-ASCII characters, if they do not have a named entity reference,
+        are replaced by numerical character references.
+
+        The return value is guaranteed to be ASCII.
+        """
+        return self.__escapable.sub(self.__escape, str(text)).encode("ascii")
+
+    # XXX: This regexp will not match all valid XML entity names__.
+    # (It punts on details involving involving CombiningChars and Extenders.)
+    #
+    # .. __: http://www.w3.org/TR/2000/REC-xml-20001006#NT-EntityRef
+    __characterrefs = re.compile(
+        r"""& (?:
+                                          \#(\d+)
+                                          | \#x([\da-f]+)
+                                          | ( (?!\d) [:\w] [-.:\w]+ )
+                                          ) ;""",
+        re.X | re.UNICODE,
+    )
+
+    def __unescape(self, m):
+        dval, hval, name = m.groups()
+        if dval:
+            codepoint = int(dval)
+        elif hval:
+            codepoint = int(hval, 16)
+        else:
+            codepoint = self.name2codepoint.get(name, 0xFFFD)
+            # U+FFFD = "REPLACEMENT CHARACTER"
+        if codepoint < 128:
+            return chr(codepoint)
+        return chr(codepoint)
+
+    def unescape(self, text):
+        """Unescape character references.
+
+        All character references (both entity references and numerical
+        character references) are unescaped.
+        """
+        return self.__characterrefs.sub(self.__unescape, text)
+
+
+_html_entities_escaper = XMLEntityEscaper(codepoint2name, name2codepoint)
+
+html_entities_escape = _html_entities_escaper.escape_entities
+html_entities_unescape = _html_entities_escaper.unescape
+
+
+def htmlentityreplace_errors(ex):
+    """An encoding error handler.
+
+    This python codecs error handler replaces unencodable
+    characters with HTML entities, or, if no HTML entity exists for
+    the character, XML character references::
+
+        >>> 'The cost was \u20ac12.'.encode('latin1', 'htmlentityreplace')
+        'The cost was &euro;12.'
+    """
+    if isinstance(ex, UnicodeEncodeError):
+        # Handle encoding errors
+        bad_text = ex.object[ex.start : ex.end]
+        text = _html_entities_escaper.escape(bad_text)
+        return (str(text), ex.end)
+    raise ex
+
+
+codecs.register_error("htmlentityreplace", htmlentityreplace_errors)
+
+
+DEFAULT_ESCAPES = {
+    "x": "filters.xml_escape",
+    "h": "filters.html_escape",
+    "u": "filters.url_escape",
+    "trim": "filters.trim",
+    "entity": "filters.html_entities_escape",
+    "unicode": "str",
+    "decode": "decode",
+    "str": "str",
+    "n": "n",
+}
diff --git a/backend/venv/lib/python3.12/site-packages/mako/lexer.py b/backend/venv/lib/python3.12/site-packages/mako/lexer.py
new file mode 100644
index 00000000..37afb4de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/lexer.py
@@ -0,0 +1,481 @@
+# mako/lexer.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""provides the Lexer class for parsing template strings into parse trees."""
+
+import codecs
+import re
+
+from mako import exceptions
+from mako import parsetree
+from mako.pygen import adjust_whitespace
+
+_regexp_cache = {}
+
+
+class Lexer:
+    def __init__(
+        self, text, filename=None, input_encoding=None, preprocessor=None
+    ):
+        self.text = text
+        self.filename = filename
+        self.template = parsetree.TemplateNode(self.filename)
+        self.matched_lineno = 1
+        self.matched_charpos = 0
+        self.lineno = 1
+        self.match_position = 0
+        self.tag = []
+        self.control_line = []
+        self.ternary_stack = []
+        self.encoding = input_encoding
+
+        if preprocessor is None:
+            self.preprocessor = []
+        elif not hasattr(preprocessor, "__iter__"):
+            self.preprocessor = [preprocessor]
+        else:
+            self.preprocessor = preprocessor
+
+    @property
+    def exception_kwargs(self):
+        return {
+            "source": self.text,
+            "lineno": self.matched_lineno,
+            "pos": self.matched_charpos,
+            "filename": self.filename,
+        }
+
+    def match(self, regexp, flags=None):
+        """compile the given regexp, cache the reg, and call match_reg()."""
+
+        try:
+            reg = _regexp_cache[(regexp, flags)]
+        except KeyError:
+            reg = re.compile(regexp, flags) if flags else re.compile(regexp)
+            _regexp_cache[(regexp, flags)] = reg
+
+        return self.match_reg(reg)
+
+    def match_reg(self, reg):
+        """match the given regular expression object to the current text
+        position.
+
+        if a match occurs, update the current text and line position.
+
+        """
+
+        mp = self.match_position
+
+        match = reg.match(self.text, self.match_position)
+        if match:
+            (start, end) = match.span()
+            self.match_position = end + 1 if end == start else end
+            self.matched_lineno = self.lineno
+            cp = mp - 1
+            if cp >= 0 and cp < self.textlength:
+                cp = self.text[: cp + 1].rfind("\n")
+            self.matched_charpos = mp - cp
+            self.lineno += self.text[mp : self.match_position].count("\n")
+        return match
+
+    def parse_until_text(self, watch_nesting, *text):
+        startpos = self.match_position
+        text_re = r"|".join(text)
+        brace_level = 0
+        paren_level = 0
+        bracket_level = 0
+        while True:
+            match = self.match(r"#.*\n")
+            if match:
+                continue
+            match = self.match(
+                r"(\"\"\"|\'\'\'|\"|\')[^\\]*?(\\.[^\\]*?)*\1", re.S
+            )
+            if match:
+                continue
+            match = self.match(r"(%s)" % text_re)
+            if match and not (
+                watch_nesting
+                and (brace_level > 0 or paren_level > 0 or bracket_level > 0)
+            ):
+                return (
+                    self.text[
+                        startpos : self.match_position - len(match.group(1))
+                    ],
+                    match.group(1),
+                )
+            elif not match:
+                match = self.match(r"(.*?)(?=\"|\'|#|%s)" % text_re, re.S)
+            if match:
+                brace_level += match.group(1).count("{")
+                brace_level -= match.group(1).count("}")
+                paren_level += match.group(1).count("(")
+                paren_level -= match.group(1).count(")")
+                bracket_level += match.group(1).count("[")
+                bracket_level -= match.group(1).count("]")
+                continue
+            raise exceptions.SyntaxException(
+                "Expected: %s" % ",".join(text), **self.exception_kwargs
+            )
+
+    def append_node(self, nodecls, *args, **kwargs):
+        kwargs.setdefault("source", self.text)
+        kwargs.setdefault("lineno", self.matched_lineno)
+        kwargs.setdefault("pos", self.matched_charpos)
+        kwargs["filename"] = self.filename
+        node = nodecls(*args, **kwargs)
+        if len(self.tag):
+            self.tag[-1].nodes.append(node)
+        else:
+            self.template.nodes.append(node)
+        # build a set of child nodes for the control line
+        # (used for loop variable detection)
+        # also build a set of child nodes on ternary control lines
+        # (used for determining if a pass needs to be auto-inserted
+        if self.control_line:
+            control_frame = self.control_line[-1]
+            control_frame.nodes.append(node)
+            if (
+                not (
+                    isinstance(node, parsetree.ControlLine)
+                    and control_frame.is_ternary(node.keyword)
+                )
+                and self.ternary_stack
+                and self.ternary_stack[-1]
+            ):
+                self.ternary_stack[-1][-1].nodes.append(node)
+        if isinstance(node, parsetree.Tag):
+            if len(self.tag):
+                node.parent = self.tag[-1]
+            self.tag.append(node)
+        elif isinstance(node, parsetree.ControlLine):
+            if node.isend:
+                self.control_line.pop()
+                self.ternary_stack.pop()
+            elif node.is_primary:
+                self.control_line.append(node)
+                self.ternary_stack.append([])
+            elif self.control_line and self.control_line[-1].is_ternary(
+                node.keyword
+            ):
+                self.ternary_stack[-1].append(node)
+            elif self.control_line and not self.control_line[-1].is_ternary(
+                node.keyword
+            ):
+                raise exceptions.SyntaxException(
+                    "Keyword '%s' not a legal ternary for keyword '%s'"
+                    % (node.keyword, self.control_line[-1].keyword),
+                    **self.exception_kwargs,
+                )
+
+    _coding_re = re.compile(r"#.*coding[:=]\s*([-\w.]+).*\r?\n")
+
+    def decode_raw_stream(self, text, decode_raw, known_encoding, filename):
+        """given string/unicode or bytes/string, determine encoding
+        from magic encoding comment, return body as unicode
+        or raw if decode_raw=False
+
+        """
+        if isinstance(text, str):
+            m = self._coding_re.match(text)
+            encoding = m and m.group(1) or known_encoding or "utf-8"
+            return encoding, text
+
+        if text.startswith(codecs.BOM_UTF8):
+            text = text[len(codecs.BOM_UTF8) :]
+            parsed_encoding = "utf-8"
+            m = self._coding_re.match(text.decode("utf-8", "ignore"))
+            if m is not None and m.group(1) != "utf-8":
+                raise exceptions.CompileException(
+                    "Found utf-8 BOM in file, with conflicting "
+                    "magic encoding comment of '%s'" % m.group(1),
+                    text.decode("utf-8", "ignore"),
+                    0,
+                    0,
+                    filename,
+                )
+        else:
+            m = self._coding_re.match(text.decode("utf-8", "ignore"))
+            parsed_encoding = m.group(1) if m else known_encoding or "utf-8"
+        if decode_raw:
+            try:
+                text = text.decode(parsed_encoding)
+            except UnicodeDecodeError:
+                raise exceptions.CompileException(
+                    "Unicode decode operation of encoding '%s' failed"
+                    % parsed_encoding,
+                    text.decode("utf-8", "ignore"),
+                    0,
+                    0,
+                    filename,
+                )
+
+        return parsed_encoding, text
+
+    def parse(self):
+        self.encoding, self.text = self.decode_raw_stream(
+            self.text, True, self.encoding, self.filename
+        )
+
+        for preproc in self.preprocessor:
+            self.text = preproc(self.text)
+
+        # push the match marker past the
+        # encoding comment.
+        self.match_reg(self._coding_re)
+
+        self.textlength = len(self.text)
+
+        while True:
+            if self.match_position > self.textlength:
+                break
+
+            if self.match_end():
+                break
+            if self.match_expression():
+                continue
+            if self.match_control_line():
+                continue
+            if self.match_comment():
+                continue
+            if self.match_tag_start():
+                continue
+            if self.match_tag_end():
+                continue
+            if self.match_python_block():
+                continue
+            if self.match_percent():
+                continue
+            if self.match_text():
+                continue
+
+            if self.match_position > self.textlength:
+                break
+            # TODO: no coverage here
+            raise exceptions.MakoException("assertion failed")
+
+        if len(self.tag):
+            raise exceptions.SyntaxException(
+                "Unclosed tag: <%%%s>" % self.tag[-1].keyword,
+                **self.exception_kwargs,
+            )
+        if len(self.control_line):
+            raise exceptions.SyntaxException(
+                "Unterminated control keyword: '%s'"
+                % self.control_line[-1].keyword,
+                self.text,
+                self.control_line[-1].lineno,
+                self.control_line[-1].pos,
+                self.filename,
+            )
+        return self.template
+
+    def match_tag_start(self):
+        reg = r"""
+            \<%     # opening tag
+
+            ([\w\.\:]+)   # keyword
+
+            ((?:\s+\w+|\s*=\s*|"[^"]*?"|'[^']*?'|\s*,\s*)*)  # attrname, = \
+                                               #        sign, string expression
+                                               # comma is for backwards compat
+                                               # identified in #366
+
+            \s*     # more whitespace
+
+            (/)?>   # closing
+
+        """
+
+        match = self.match(
+            reg,
+            re.I | re.S | re.X,
+        )
+
+        if not match:
+            return False
+
+        keyword, attr, isend = match.groups()
+        self.keyword = keyword
+        attributes = {}
+        if attr:
+            for att in re.findall(
+                r"\s*(\w+)\s*=\s*(?:'([^']*)'|\"([^\"]*)\")", attr
+            ):
+                key, val1, val2 = att
+                text = val1 or val2
+                text = text.replace("\r\n", "\n")
+                attributes[key] = text
+        self.append_node(parsetree.Tag, keyword, attributes)
+        if isend:
+            self.tag.pop()
+        elif keyword == "text":
+            match = self.match(r"(.*?)(?=\</%text>)", re.S)
+            if not match:
+                raise exceptions.SyntaxException(
+                    "Unclosed tag: <%%%s>" % self.tag[-1].keyword,
+                    **self.exception_kwargs,
+                )
+            self.append_node(parsetree.Text, match.group(1))
+            return self.match_tag_end()
+        return True
+
+    def match_tag_end(self):
+        match = self.match(r"\</%[\t ]*([^\t ]+?)[\t ]*>")
+        if match:
+            if not len(self.tag):
+                raise exceptions.SyntaxException(
+                    "Closing tag without opening tag: </%%%s>"
+                    % match.group(1),
+                    **self.exception_kwargs,
+                )
+            elif self.tag[-1].keyword != match.group(1):
+                raise exceptions.SyntaxException(
+                    "Closing tag </%%%s> does not match tag: <%%%s>"
+                    % (match.group(1), self.tag[-1].keyword),
+                    **self.exception_kwargs,
+                )
+            self.tag.pop()
+            return True
+        else:
+            return False
+
+    def match_end(self):
+        match = self.match(r"\Z", re.S)
+        if not match:
+            return False
+
+        string = match.group()
+        if string:
+            return string
+        else:
+            return True
+
+    def match_percent(self):
+        match = self.match(r"(?<=^)(\s*)%%(%*)", re.M)
+        if match:
+            self.append_node(
+                parsetree.Text, match.group(1) + "%" + match.group(2)
+            )
+            return True
+        else:
+            return False
+
+    def match_text(self):
+        match = self.match(
+            r"""
+                (.*?)         # anything, followed by:
+                (
+                 (?<=\n)(?=[ \t]*(?=%|\#\#))  # an eval or line-based
+                                            # comment, preceded by a
+                                            # consumed newline and whitespace
+                 |
+                 (?=\${)      # an expression
+                 |
+                 (?=</?%)  # a substitution or block or call start or end
+                              # - don't consume
+                 |
+                 (\\\r?\n)    # an escaped newline  - throw away
+                 |
+                 \Z           # end of string
+                )""",
+            re.X | re.S,
+        )
+
+        if match:
+            text = match.group(1)
+            if text:
+                self.append_node(parsetree.Text, text)
+            return True
+        else:
+            return False
+
+    def match_python_block(self):
+        match = self.match(r"<%(!)?")
+        if match:
+            line, pos = self.matched_lineno, self.matched_charpos
+            text, end = self.parse_until_text(False, r"%>")
+            # the trailing newline helps
+            # compiler.parse() not complain about indentation
+            text = adjust_whitespace(text) + "\n"
+            self.append_node(
+                parsetree.Code,
+                text,
+                match.group(1) == "!",
+                lineno=line,
+                pos=pos,
+            )
+            return True
+        else:
+            return False
+
+    def match_expression(self):
+        match = self.match(r"\${")
+        if not match:
+            return False
+
+        line, pos = self.matched_lineno, self.matched_charpos
+        text, end = self.parse_until_text(True, r"\|", r"}")
+        if end == "|":
+            escapes, end = self.parse_until_text(True, r"}")
+        else:
+            escapes = ""
+        text = text.replace("\r\n", "\n")
+        self.append_node(
+            parsetree.Expression,
+            text,
+            escapes.strip(),
+            lineno=line,
+            pos=pos,
+        )
+        return True
+
+    def match_control_line(self):
+        match = self.match(
+            r"(?<=^)[\t ]*(%(?!%)|##)[\t ]*((?:(?:\\\r?\n)|[^\r\n])*)"
+            r"(?:\r?\n|\Z)",
+            re.M,
+        )
+        if not match:
+            return False
+
+        operator = match.group(1)
+        text = match.group(2)
+        if operator == "%":
+            m2 = re.match(r"(end)?(\w+)\s*(.*)", text)
+            if not m2:
+                raise exceptions.SyntaxException(
+                    "Invalid control line: '%s'" % text,
+                    **self.exception_kwargs,
+                )
+            isend, keyword = m2.group(1, 2)
+            isend = isend is not None
+
+            if isend:
+                if not len(self.control_line):
+                    raise exceptions.SyntaxException(
+                        "No starting keyword '%s' for '%s'" % (keyword, text),
+                        **self.exception_kwargs,
+                    )
+                elif self.control_line[-1].keyword != keyword:
+                    raise exceptions.SyntaxException(
+                        "Keyword '%s' doesn't match keyword '%s'"
+                        % (text, self.control_line[-1].keyword),
+                        **self.exception_kwargs,
+                    )
+            self.append_node(parsetree.ControlLine, keyword, isend, text)
+        else:
+            self.append_node(parsetree.Comment, text)
+        return True
+
+    def match_comment(self):
+        """matches the multiline version of a comment"""
+        match = self.match(r"<%doc>(.*?)</%doc>", re.S)
+        if match:
+            self.append_node(parsetree.Comment, match.group(1))
+            return True
+        else:
+            return False
diff --git a/backend/venv/lib/python3.12/site-packages/mako/lookup.py b/backend/venv/lib/python3.12/site-packages/mako/lookup.py
new file mode 100644
index 00000000..1a431c80
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/lookup.py
@@ -0,0 +1,361 @@
+# mako/lookup.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+import os
+import posixpath
+import re
+import stat
+import threading
+
+from mako import exceptions
+from mako import util
+from mako.template import Template
+
+
+class TemplateCollection:
+
+    """Represent a collection of :class:`.Template` objects,
+    identifiable via URI.
+
+    A :class:`.TemplateCollection` is linked to the usage of
+    all template tags that address other templates, such
+    as ``<%include>``, ``<%namespace>``, and ``<%inherit>``.
+    The ``file`` attribute of each of those tags refers
+    to a string URI that is passed to that :class:`.Template`
+    object's :class:`.TemplateCollection` for resolution.
+
+    :class:`.TemplateCollection` is an abstract class,
+    with the usual default implementation being :class:`.TemplateLookup`.
+
+    """
+
+    def has_template(self, uri):
+        """Return ``True`` if this :class:`.TemplateLookup` is
+        capable of returning a :class:`.Template` object for the
+        given ``uri``.
+
+        :param uri: String URI of the template to be resolved.
+
+        """
+        try:
+            self.get_template(uri)
+            return True
+        except exceptions.TemplateLookupException:
+            return False
+
+    def get_template(self, uri, relativeto=None):
+        """Return a :class:`.Template` object corresponding to the given
+        ``uri``.
+
+        The default implementation raises
+        :class:`.NotImplementedError`. Implementations should
+        raise :class:`.TemplateLookupException` if the given ``uri``
+        cannot be resolved.
+
+        :param uri: String URI of the template to be resolved.
+        :param relativeto: if present, the given ``uri`` is assumed to
+         be relative to this URI.
+
+        """
+        raise NotImplementedError()
+
+    def filename_to_uri(self, uri, filename):
+        """Convert the given ``filename`` to a URI relative to
+        this :class:`.TemplateCollection`."""
+
+        return uri
+
+    def adjust_uri(self, uri, filename):
+        """Adjust the given ``uri`` based on the calling ``filename``.
+
+        When this method is called from the runtime, the
+        ``filename`` parameter is taken directly to the ``filename``
+        attribute of the calling template. Therefore a custom
+        :class:`.TemplateCollection` subclass can place any string
+        identifier desired in the ``filename`` parameter of the
+        :class:`.Template` objects it constructs and have them come back
+        here.
+
+        """
+        return uri
+
+
+class TemplateLookup(TemplateCollection):
+
+    """Represent a collection of templates that locates template source files
+    from the local filesystem.
+
+    The primary argument is the ``directories`` argument, the list of
+    directories to search:
+
+    .. sourcecode:: python
+
+        lookup = TemplateLookup(["/path/to/templates"])
+        some_template = lookup.get_template("/index.html")
+
+    The :class:`.TemplateLookup` can also be given :class:`.Template` objects
+    programatically using :meth:`.put_string` or :meth:`.put_template`:
+
+    .. sourcecode:: python
+
+        lookup = TemplateLookup()
+        lookup.put_string("base.html", '''
+            <html><body>${self.next()}</body></html>
+        ''')
+        lookup.put_string("hello.html", '''
+            <%include file='base.html'/>
+
+            Hello, world !
+        ''')
+
+
+    :param directories: A list of directory names which will be
+     searched for a particular template URI. The URI is appended
+     to each directory and the filesystem checked.
+
+    :param collection_size: Approximate size of the collection used
+     to store templates. If left at its default of ``-1``, the size
+     is unbounded, and a plain Python dictionary is used to
+     relate URI strings to :class:`.Template` instances.
+     Otherwise, a least-recently-used cache object is used which
+     will maintain the size of the collection approximately to
+     the number given.
+
+    :param filesystem_checks: When at its default value of ``True``,
+     each call to :meth:`.TemplateLookup.get_template()` will
+     compare the filesystem last modified time to the time in
+     which an existing :class:`.Template` object was created.
+     This allows the :class:`.TemplateLookup` to regenerate a
+     new :class:`.Template` whenever the original source has
+     been updated. Set this to ``False`` for a very minor
+     performance increase.
+
+    :param modulename_callable: A callable which, when present,
+     is passed the path of the source file as well as the
+     requested URI, and then returns the full path of the
+     generated Python module file. This is used to inject
+     alternate schemes for Python module location. If left at
+     its default of ``None``, the built in system of generation
+     based on ``module_directory`` plus ``uri`` is used.
+
+    All other keyword parameters available for
+    :class:`.Template` are mirrored here. When new
+    :class:`.Template` objects are created, the keywords
+    established with this :class:`.TemplateLookup` are passed on
+    to each new :class:`.Template`.
+
+    """
+
+    def __init__(
+        self,
+        directories=None,
+        module_directory=None,
+        filesystem_checks=True,
+        collection_size=-1,
+        format_exceptions=False,
+        error_handler=None,
+        output_encoding=None,
+        encoding_errors="strict",
+        cache_args=None,
+        cache_impl="beaker",
+        cache_enabled=True,
+        cache_type=None,
+        cache_dir=None,
+        cache_url=None,
+        modulename_callable=None,
+        module_writer=None,
+        default_filters=None,
+        buffer_filters=(),
+        strict_undefined=False,
+        imports=None,
+        future_imports=None,
+        enable_loop=True,
+        input_encoding=None,
+        preprocessor=None,
+        lexer_cls=None,
+        include_error_handler=None,
+    ):
+        self.directories = [
+            posixpath.normpath(d) for d in util.to_list(directories, ())
+        ]
+        self.module_directory = module_directory
+        self.modulename_callable = modulename_callable
+        self.filesystem_checks = filesystem_checks
+        self.collection_size = collection_size
+
+        if cache_args is None:
+            cache_args = {}
+        # transfer deprecated cache_* args
+        if cache_dir:
+            cache_args.setdefault("dir", cache_dir)
+        if cache_url:
+            cache_args.setdefault("url", cache_url)
+        if cache_type:
+            cache_args.setdefault("type", cache_type)
+
+        self.template_args = {
+            "format_exceptions": format_exceptions,
+            "error_handler": error_handler,
+            "include_error_handler": include_error_handler,
+            "output_encoding": output_encoding,
+            "cache_impl": cache_impl,
+            "encoding_errors": encoding_errors,
+            "input_encoding": input_encoding,
+            "module_directory": module_directory,
+            "module_writer": module_writer,
+            "cache_args": cache_args,
+            "cache_enabled": cache_enabled,
+            "default_filters": default_filters,
+            "buffer_filters": buffer_filters,
+            "strict_undefined": strict_undefined,
+            "imports": imports,
+            "future_imports": future_imports,
+            "enable_loop": enable_loop,
+            "preprocessor": preprocessor,
+            "lexer_cls": lexer_cls,
+        }
+
+        if collection_size == -1:
+            self._collection = {}
+            self._uri_cache = {}
+        else:
+            self._collection = util.LRUCache(collection_size)
+            self._uri_cache = util.LRUCache(collection_size)
+        self._mutex = threading.Lock()
+
+    def get_template(self, uri):
+        """Return a :class:`.Template` object corresponding to the given
+        ``uri``.
+
+        .. note:: The ``relativeto`` argument is not supported here at
+           the moment.
+
+        """
+
+        try:
+            if self.filesystem_checks:
+                return self._check(uri, self._collection[uri])
+            else:
+                return self._collection[uri]
+        except KeyError as e:
+            u = re.sub(r"^\/+", "", uri)
+            for dir_ in self.directories:
+                # make sure the path seperators are posix - os.altsep is empty
+                # on POSIX and cannot be used.
+                dir_ = dir_.replace(os.path.sep, posixpath.sep)
+                srcfile = posixpath.normpath(posixpath.join(dir_, u))
+                if os.path.isfile(srcfile):
+                    return self._load(srcfile, uri)
+            else:
+                raise exceptions.TopLevelLookupException(
+                    "Can't locate template for uri %r" % uri
+                ) from e
+
+    def adjust_uri(self, uri, relativeto):
+        """Adjust the given ``uri`` based on the given relative URI."""
+
+        key = (uri, relativeto)
+        if key in self._uri_cache:
+            return self._uri_cache[key]
+
+        if uri[0] == "/":
+            v = self._uri_cache[key] = uri
+        elif relativeto is not None:
+            v = self._uri_cache[key] = posixpath.join(
+                posixpath.dirname(relativeto), uri
+            )
+        else:
+            v = self._uri_cache[key] = "/" + uri
+        return v
+
+    def filename_to_uri(self, filename):
+        """Convert the given ``filename`` to a URI relative to
+        this :class:`.TemplateCollection`."""
+
+        try:
+            return self._uri_cache[filename]
+        except KeyError:
+            value = self._relativeize(filename)
+            self._uri_cache[filename] = value
+            return value
+
+    def _relativeize(self, filename):
+        """Return the portion of a filename that is 'relative'
+        to the directories in this lookup.
+
+        """
+
+        filename = posixpath.normpath(filename)
+        for dir_ in self.directories:
+            if filename[0 : len(dir_)] == dir_:
+                return filename[len(dir_) :]
+        else:
+            return None
+
+    def _load(self, filename, uri):
+        self._mutex.acquire()
+        try:
+            try:
+                # try returning from collection one
+                # more time in case concurrent thread already loaded
+                return self._collection[uri]
+            except KeyError:
+                pass
+            try:
+                if self.modulename_callable is not None:
+                    module_filename = self.modulename_callable(filename, uri)
+                else:
+                    module_filename = None
+                self._collection[uri] = template = Template(
+                    uri=uri,
+                    filename=posixpath.normpath(filename),
+                    lookup=self,
+                    module_filename=module_filename,
+                    **self.template_args,
+                )
+                return template
+            except:
+                # if compilation fails etc, ensure
+                # template is removed from collection,
+                # re-raise
+                self._collection.pop(uri, None)
+                raise
+        finally:
+            self._mutex.release()
+
+    def _check(self, uri, template):
+        if template.filename is None:
+            return template
+
+        try:
+            template_stat = os.stat(template.filename)
+            if template.module._modified_time >= template_stat[stat.ST_MTIME]:
+                return template
+            self._collection.pop(uri, None)
+            return self._load(template.filename, uri)
+        except OSError as e:
+            self._collection.pop(uri, None)
+            raise exceptions.TemplateLookupException(
+                "Can't locate template for uri %r" % uri
+            ) from e
+
+    def put_string(self, uri, text):
+        """Place a new :class:`.Template` object into this
+        :class:`.TemplateLookup`, based on the given string of
+        ``text``.
+
+        """
+        self._collection[uri] = Template(
+            text, lookup=self, uri=uri, **self.template_args
+        )
+
+    def put_template(self, uri, template):
+        """Place a new :class:`.Template` object into this
+        :class:`.TemplateLookup`, based on the given
+        :class:`.Template` object.
+
+        """
+        self._collection[uri] = template
diff --git a/backend/venv/lib/python3.12/site-packages/mako/parsetree.py b/backend/venv/lib/python3.12/site-packages/mako/parsetree.py
new file mode 100644
index 00000000..f7bdda8d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/parsetree.py
@@ -0,0 +1,656 @@
+# mako/parsetree.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""defines the parse tree components for Mako templates."""
+
+import re
+
+from mako import ast
+from mako import exceptions
+from mako import filters
+from mako import util
+
+
+class Node:
+
+    """base class for a Node in the parse tree."""
+
+    def __init__(self, source, lineno, pos, filename):
+        self.source = source
+        self.lineno = lineno
+        self.pos = pos
+        self.filename = filename
+
+    @property
+    def exception_kwargs(self):
+        return {
+            "source": self.source,
+            "lineno": self.lineno,
+            "pos": self.pos,
+            "filename": self.filename,
+        }
+
+    def get_children(self):
+        return []
+
+    def accept_visitor(self, visitor):
+        def traverse(node):
+            for n in node.get_children():
+                n.accept_visitor(visitor)
+
+        method = getattr(visitor, "visit" + self.__class__.__name__, traverse)
+        method(self)
+
+
+class TemplateNode(Node):
+
+    """a 'container' node that stores the overall collection of nodes."""
+
+    def __init__(self, filename):
+        super().__init__("", 0, 0, filename)
+        self.nodes = []
+        self.page_attributes = {}
+
+    def get_children(self):
+        return self.nodes
+
+    def __repr__(self):
+        return "TemplateNode(%s, %r)" % (
+            util.sorted_dict_repr(self.page_attributes),
+            self.nodes,
+        )
+
+
+class ControlLine(Node):
+
+    """defines a control line, a line-oriented python line or end tag.
+
+    e.g.::
+
+        % if foo:
+            (markup)
+        % endif
+
+    """
+
+    has_loop_context = False
+
+    def __init__(self, keyword, isend, text, **kwargs):
+        super().__init__(**kwargs)
+        self.text = text
+        self.keyword = keyword
+        self.isend = isend
+        self.is_primary = keyword in ["for", "if", "while", "try", "with"]
+        self.nodes = []
+        if self.isend:
+            self._declared_identifiers = []
+            self._undeclared_identifiers = []
+        else:
+            code = ast.PythonFragment(text, **self.exception_kwargs)
+            self._declared_identifiers = code.declared_identifiers
+            self._undeclared_identifiers = code.undeclared_identifiers
+
+    def get_children(self):
+        return self.nodes
+
+    def declared_identifiers(self):
+        return self._declared_identifiers
+
+    def undeclared_identifiers(self):
+        return self._undeclared_identifiers
+
+    def is_ternary(self, keyword):
+        """return true if the given keyword is a ternary keyword
+        for this ControlLine"""
+
+        cases = {
+            "if": {"else", "elif"},
+            "try": {"except", "finally"},
+            "for": {"else"},
+        }
+
+        return keyword in cases.get(self.keyword, set())
+
+    def __repr__(self):
+        return "ControlLine(%r, %r, %r, %r)" % (
+            self.keyword,
+            self.text,
+            self.isend,
+            (self.lineno, self.pos),
+        )
+
+
+class Text(Node):
+    """defines plain text in the template."""
+
+    def __init__(self, content, **kwargs):
+        super().__init__(**kwargs)
+        self.content = content
+
+    def __repr__(self):
+        return "Text(%r, %r)" % (self.content, (self.lineno, self.pos))
+
+
+class Code(Node):
+    """defines a Python code block, either inline or module level.
+
+    e.g.::
+
+        inline:
+        <%
+            x = 12
+        %>
+
+        module level:
+        <%!
+            import logger
+        %>
+
+    """
+
+    def __init__(self, text, ismodule, **kwargs):
+        super().__init__(**kwargs)
+        self.text = text
+        self.ismodule = ismodule
+        self.code = ast.PythonCode(text, **self.exception_kwargs)
+
+    def declared_identifiers(self):
+        return self.code.declared_identifiers
+
+    def undeclared_identifiers(self):
+        return self.code.undeclared_identifiers
+
+    def __repr__(self):
+        return "Code(%r, %r, %r)" % (
+            self.text,
+            self.ismodule,
+            (self.lineno, self.pos),
+        )
+
+
+class Comment(Node):
+    """defines a comment line.
+
+    # this is a comment
+
+    """
+
+    def __init__(self, text, **kwargs):
+        super().__init__(**kwargs)
+        self.text = text
+
+    def __repr__(self):
+        return "Comment(%r, %r)" % (self.text, (self.lineno, self.pos))
+
+
+class Expression(Node):
+    """defines an inline expression.
+
+    ${x+y}
+
+    """
+
+    def __init__(self, text, escapes, **kwargs):
+        super().__init__(**kwargs)
+        self.text = text
+        self.escapes = escapes
+        self.escapes_code = ast.ArgumentList(escapes, **self.exception_kwargs)
+        self.code = ast.PythonCode(text, **self.exception_kwargs)
+
+    def declared_identifiers(self):
+        return []
+
+    def undeclared_identifiers(self):
+        # TODO: make the "filter" shortcut list configurable at parse/gen time
+        return self.code.undeclared_identifiers.union(
+            self.escapes_code.undeclared_identifiers.difference(
+                filters.DEFAULT_ESCAPES
+            )
+        ).difference(self.code.declared_identifiers)
+
+    def __repr__(self):
+        return "Expression(%r, %r, %r)" % (
+            self.text,
+            self.escapes_code.args,
+            (self.lineno, self.pos),
+        )
+
+
+class _TagMeta(type):
+    """metaclass to allow Tag to produce a subclass according to
+    its keyword"""
+
+    _classmap = {}
+
+    def __init__(cls, clsname, bases, dict_):
+        if getattr(cls, "__keyword__", None) is not None:
+            cls._classmap[cls.__keyword__] = cls
+        super().__init__(clsname, bases, dict_)
+
+    def __call__(cls, keyword, attributes, **kwargs):
+        if ":" in keyword:
+            ns, defname = keyword.split(":")
+            return type.__call__(
+                CallNamespaceTag, ns, defname, attributes, **kwargs
+            )
+
+        try:
+            cls = _TagMeta._classmap[keyword]
+        except KeyError:
+            raise exceptions.CompileException(
+                "No such tag: '%s'" % keyword,
+                source=kwargs["source"],
+                lineno=kwargs["lineno"],
+                pos=kwargs["pos"],
+                filename=kwargs["filename"],
+            )
+        return type.__call__(cls, keyword, attributes, **kwargs)
+
+
+class Tag(Node, metaclass=_TagMeta):
+    """abstract base class for tags.
+
+    e.g.::
+
+        <%sometag/>
+
+        <%someothertag>
+            stuff
+        </%someothertag>
+
+    """
+
+    __keyword__ = None
+
+    def __init__(
+        self,
+        keyword,
+        attributes,
+        expressions,
+        nonexpressions,
+        required,
+        **kwargs,
+    ):
+        r"""construct a new Tag instance.
+
+        this constructor not called directly, and is only called
+        by subclasses.
+
+        :param keyword: the tag keyword
+
+        :param attributes: raw dictionary of attribute key/value pairs
+
+        :param expressions: a set of identifiers that are legal attributes,
+         which can also contain embedded expressions
+
+        :param nonexpressions: a set of identifiers that are legal
+         attributes, which cannot contain embedded expressions
+
+        :param \**kwargs:
+         other arguments passed to the Node superclass (lineno, pos)
+
+        """
+        super().__init__(**kwargs)
+        self.keyword = keyword
+        self.attributes = attributes
+        self._parse_attributes(expressions, nonexpressions)
+        missing = [r for r in required if r not in self.parsed_attributes]
+        if len(missing):
+            raise exceptions.CompileException(
+                (
+                    "Missing attribute(s): %s"
+                    % ",".join(repr(m) for m in missing)
+                ),
+                **self.exception_kwargs,
+            )
+
+        self.parent = None
+        self.nodes = []
+
+    def is_root(self):
+        return self.parent is None
+
+    def get_children(self):
+        return self.nodes
+
+    def _parse_attributes(self, expressions, nonexpressions):
+        undeclared_identifiers = set()
+        self.parsed_attributes = {}
+        for key in self.attributes:
+            if key in expressions:
+                expr = []
+                for x in re.compile(r"(\${(?:[^$]*?{.+|.+?)})", re.S).split(
+                    self.attributes[key]
+                ):
+                    m = re.compile(r"^\${(.+?)}$", re.S).match(x)
+                    if m:
+                        code = ast.PythonCode(
+                            m.group(1).rstrip(), **self.exception_kwargs
+                        )
+                        # we aren't discarding "declared_identifiers" here,
+                        # which we do so that list comprehension-declared
+                        # variables aren't counted.   As yet can't find a
+                        # condition that requires it here.
+                        undeclared_identifiers = undeclared_identifiers.union(
+                            code.undeclared_identifiers
+                        )
+                        expr.append("(%s)" % m.group(1))
+                    elif x:
+                        expr.append(repr(x))
+                self.parsed_attributes[key] = " + ".join(expr) or repr("")
+            elif key in nonexpressions:
+                if re.search(r"\${.+?}", self.attributes[key]):
+                    raise exceptions.CompileException(
+                        "Attribute '%s' in tag '%s' does not allow embedded "
+                        "expressions" % (key, self.keyword),
+                        **self.exception_kwargs,
+                    )
+                self.parsed_attributes[key] = repr(self.attributes[key])
+            else:
+                raise exceptions.CompileException(
+                    "Invalid attribute for tag '%s': '%s'"
+                    % (self.keyword, key),
+                    **self.exception_kwargs,
+                )
+        self.expression_undeclared_identifiers = undeclared_identifiers
+
+    def declared_identifiers(self):
+        return []
+
+    def undeclared_identifiers(self):
+        return self.expression_undeclared_identifiers
+
+    def __repr__(self):
+        return "%s(%r, %s, %r, %r)" % (
+            self.__class__.__name__,
+            self.keyword,
+            util.sorted_dict_repr(self.attributes),
+            (self.lineno, self.pos),
+            self.nodes,
+        )
+
+
+class IncludeTag(Tag):
+    __keyword__ = "include"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        super().__init__(
+            keyword,
+            attributes,
+            ("file", "import", "args"),
+            (),
+            ("file",),
+            **kwargs,
+        )
+        self.page_args = ast.PythonCode(
+            "__DUMMY(%s)" % attributes.get("args", ""), **self.exception_kwargs
+        )
+
+    def declared_identifiers(self):
+        return []
+
+    def undeclared_identifiers(self):
+        identifiers = self.page_args.undeclared_identifiers.difference(
+            {"__DUMMY"}
+        ).difference(self.page_args.declared_identifiers)
+        return identifiers.union(super().undeclared_identifiers())
+
+
+class NamespaceTag(Tag):
+    __keyword__ = "namespace"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        super().__init__(
+            keyword,
+            attributes,
+            ("file",),
+            ("name", "inheritable", "import", "module"),
+            (),
+            **kwargs,
+        )
+
+        self.name = attributes.get("name", "__anon_%s" % hex(abs(id(self))))
+        if "name" not in attributes and "import" not in attributes:
+            raise exceptions.CompileException(
+                "'name' and/or 'import' attributes are required "
+                "for <%namespace>",
+                **self.exception_kwargs,
+            )
+        if "file" in attributes and "module" in attributes:
+            raise exceptions.CompileException(
+                "<%namespace> may only have one of 'file' or 'module'",
+                **self.exception_kwargs,
+            )
+
+    def declared_identifiers(self):
+        return []
+
+
+class TextTag(Tag):
+    __keyword__ = "text"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        super().__init__(keyword, attributes, (), ("filter"), (), **kwargs)
+        self.filter_args = ast.ArgumentList(
+            attributes.get("filter", ""), **self.exception_kwargs
+        )
+
+    def undeclared_identifiers(self):
+        return self.filter_args.undeclared_identifiers.difference(
+            filters.DEFAULT_ESCAPES.keys()
+        ).union(self.expression_undeclared_identifiers)
+
+
+class DefTag(Tag):
+    __keyword__ = "def"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        expressions = ["buffered", "cached"] + [
+            c for c in attributes if c.startswith("cache_")
+        ]
+
+        super().__init__(
+            keyword,
+            attributes,
+            expressions,
+            ("name", "filter", "decorator"),
+            ("name",),
+            **kwargs,
+        )
+        name = attributes["name"]
+        if re.match(r"^[\w_]+$", name):
+            raise exceptions.CompileException(
+                "Missing parenthesis in %def", **self.exception_kwargs
+            )
+        self.function_decl = ast.FunctionDecl(
+            "def " + name + ":pass", **self.exception_kwargs
+        )
+        self.name = self.function_decl.funcname
+        self.decorator = attributes.get("decorator", "")
+        self.filter_args = ast.ArgumentList(
+            attributes.get("filter", ""), **self.exception_kwargs
+        )
+
+    is_anonymous = False
+    is_block = False
+
+    @property
+    def funcname(self):
+        return self.function_decl.funcname
+
+    def get_argument_expressions(self, **kw):
+        return self.function_decl.get_argument_expressions(**kw)
+
+    def declared_identifiers(self):
+        return self.function_decl.allargnames
+
+    def undeclared_identifiers(self):
+        res = []
+        for c in self.function_decl.defaults:
+            res += list(
+                ast.PythonCode(
+                    c, **self.exception_kwargs
+                ).undeclared_identifiers
+            )
+        return (
+            set(res)
+            .union(
+                self.filter_args.undeclared_identifiers.difference(
+                    filters.DEFAULT_ESCAPES.keys()
+                )
+            )
+            .union(self.expression_undeclared_identifiers)
+            .difference(self.function_decl.allargnames)
+        )
+
+
+class BlockTag(Tag):
+    __keyword__ = "block"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        expressions = ["buffered", "cached", "args"] + [
+            c for c in attributes if c.startswith("cache_")
+        ]
+
+        super().__init__(
+            keyword,
+            attributes,
+            expressions,
+            ("name", "filter", "decorator"),
+            (),
+            **kwargs,
+        )
+        name = attributes.get("name")
+        if name and not re.match(r"^[\w_]+$", name):
+            raise exceptions.CompileException(
+                "%block may not specify an argument signature",
+                **self.exception_kwargs,
+            )
+        if not name and attributes.get("args", None):
+            raise exceptions.CompileException(
+                "Only named %blocks may specify args", **self.exception_kwargs
+            )
+        self.body_decl = ast.FunctionArgs(
+            attributes.get("args", ""), **self.exception_kwargs
+        )
+
+        self.name = name
+        self.decorator = attributes.get("decorator", "")
+        self.filter_args = ast.ArgumentList(
+            attributes.get("filter", ""), **self.exception_kwargs
+        )
+
+    is_block = True
+
+    @property
+    def is_anonymous(self):
+        return self.name is None
+
+    @property
+    def funcname(self):
+        return self.name or "__M_anon_%d" % (self.lineno,)
+
+    def get_argument_expressions(self, **kw):
+        return self.body_decl.get_argument_expressions(**kw)
+
+    def declared_identifiers(self):
+        return self.body_decl.allargnames
+
+    def undeclared_identifiers(self):
+        return (
+            self.filter_args.undeclared_identifiers.difference(
+                filters.DEFAULT_ESCAPES.keys()
+            )
+        ).union(self.expression_undeclared_identifiers)
+
+
+class CallTag(Tag):
+    __keyword__ = "call"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        super().__init__(
+            keyword, attributes, ("args"), ("expr",), ("expr",), **kwargs
+        )
+        self.expression = attributes["expr"]
+        self.code = ast.PythonCode(self.expression, **self.exception_kwargs)
+        self.body_decl = ast.FunctionArgs(
+            attributes.get("args", ""), **self.exception_kwargs
+        )
+
+    def declared_identifiers(self):
+        return self.code.declared_identifiers.union(self.body_decl.allargnames)
+
+    def undeclared_identifiers(self):
+        return self.code.undeclared_identifiers.difference(
+            self.code.declared_identifiers
+        )
+
+
+class CallNamespaceTag(Tag):
+    def __init__(self, namespace, defname, attributes, **kwargs):
+        super().__init__(
+            namespace + ":" + defname,
+            attributes,
+            tuple(attributes.keys()) + ("args",),
+            (),
+            (),
+            **kwargs,
+        )
+
+        self.expression = "%s.%s(%s)" % (
+            namespace,
+            defname,
+            ",".join(
+                "%s=%s" % (k, v)
+                for k, v in self.parsed_attributes.items()
+                if k != "args"
+            ),
+        )
+
+        self.code = ast.PythonCode(self.expression, **self.exception_kwargs)
+        self.body_decl = ast.FunctionArgs(
+            attributes.get("args", ""), **self.exception_kwargs
+        )
+
+    def declared_identifiers(self):
+        return self.code.declared_identifiers.union(self.body_decl.allargnames)
+
+    def undeclared_identifiers(self):
+        return self.code.undeclared_identifiers.difference(
+            self.code.declared_identifiers
+        )
+
+
+class InheritTag(Tag):
+    __keyword__ = "inherit"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        super().__init__(
+            keyword, attributes, ("file",), (), ("file",), **kwargs
+        )
+
+
+class PageTag(Tag):
+    __keyword__ = "page"
+
+    def __init__(self, keyword, attributes, **kwargs):
+        expressions = [
+            "cached",
+            "args",
+            "expression_filter",
+            "enable_loop",
+        ] + [c for c in attributes if c.startswith("cache_")]
+
+        super().__init__(keyword, attributes, expressions, (), (), **kwargs)
+        self.body_decl = ast.FunctionArgs(
+            attributes.get("args", ""), **self.exception_kwargs
+        )
+        self.filter_args = ast.ArgumentList(
+            attributes.get("expression_filter", ""), **self.exception_kwargs
+        )
+
+    def declared_identifiers(self):
+        return self.body_decl.allargnames
diff --git a/backend/venv/lib/python3.12/site-packages/mako/pygen.py b/backend/venv/lib/python3.12/site-packages/mako/pygen.py
new file mode 100644
index 00000000..41f9afda
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/pygen.py
@@ -0,0 +1,309 @@
+# mako/pygen.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""utilities for generating and formatting literal Python code."""
+
+import re
+
+from mako import exceptions
+
+
+class PythonPrinter:
+    def __init__(self, stream):
+        # indentation counter
+        self.indent = 0
+
+        # a stack storing information about why we incremented
+        # the indentation counter, to help us determine if we
+        # should decrement it
+        self.indent_detail = []
+
+        # the string of whitespace multiplied by the indent
+        # counter to produce a line
+        self.indentstring = "    "
+
+        # the stream we are writing to
+        self.stream = stream
+
+        # current line number
+        self.lineno = 1
+
+        # a list of lines that represents a buffered "block" of code,
+        # which can be later printed relative to an indent level
+        self.line_buffer = []
+
+        self.in_indent_lines = False
+
+        self._reset_multi_line_flags()
+
+        # mapping of generated python lines to template
+        # source lines
+        self.source_map = {}
+
+        self._re_space_comment = re.compile(r"^\s*#")
+        self._re_space = re.compile(r"^\s*$")
+        self._re_indent = re.compile(r":[ \t]*(?:#.*)?$")
+        self._re_compound = re.compile(r"^\s*(if|try|elif|while|for|with)")
+        self._re_indent_keyword = re.compile(
+            r"^\s*(def|class|else|elif|except|finally)"
+        )
+        self._re_unindentor = re.compile(r"^\s*(else|elif|except|finally).*\:")
+
+    def _update_lineno(self, num):
+        self.lineno += num
+
+    def start_source(self, lineno):
+        if self.lineno not in self.source_map:
+            self.source_map[self.lineno] = lineno
+
+    def write_blanks(self, num):
+        self.stream.write("\n" * num)
+        self._update_lineno(num)
+
+    def write_indented_block(self, block, starting_lineno=None):
+        """print a line or lines of python which already contain indentation.
+
+        The indentation of the total block of lines will be adjusted to that of
+        the current indent level."""
+        self.in_indent_lines = False
+        for i, l in enumerate(re.split(r"\r?\n", block)):
+            self.line_buffer.append(l)
+            if starting_lineno is not None:
+                self.start_source(starting_lineno + i)
+            self._update_lineno(1)
+
+    def writelines(self, *lines):
+        """print a series of lines of python."""
+        for line in lines:
+            self.writeline(line)
+
+    def writeline(self, line):
+        """print a line of python, indenting it according to the current
+        indent level.
+
+        this also adjusts the indentation counter according to the
+        content of the line.
+
+        """
+
+        if not self.in_indent_lines:
+            self._flush_adjusted_lines()
+            self.in_indent_lines = True
+
+        if (
+            line is None
+            or self._re_space_comment.match(line)
+            or self._re_space.match(line)
+        ):
+            hastext = False
+        else:
+            hastext = True
+
+        is_comment = line and len(line) and line[0] == "#"
+
+        # see if this line should decrease the indentation level
+        if (
+            not is_comment
+            and (not hastext or self._is_unindentor(line))
+            and self.indent > 0
+        ):
+            self.indent -= 1
+            # if the indent_detail stack is empty, the user
+            # probably put extra closures - the resulting
+            # module wont compile.
+            if len(self.indent_detail) == 0:
+                # TODO: no coverage here
+                raise exceptions.MakoException("Too many whitespace closures")
+            self.indent_detail.pop()
+
+        if line is None:
+            return
+
+        # write the line
+        self.stream.write(self._indent_line(line) + "\n")
+        self._update_lineno(len(line.split("\n")))
+
+        # see if this line should increase the indentation level.
+        # note that a line can both decrase (before printing) and
+        # then increase (after printing) the indentation level.
+
+        if self._re_indent.search(line):
+            # increment indentation count, and also
+            # keep track of what the keyword was that indented us,
+            # if it is a python compound statement keyword
+            # where we might have to look for an "unindent" keyword
+            match = self._re_compound.match(line)
+            if match:
+                # its a "compound" keyword, so we will check for "unindentors"
+                indentor = match.group(1)
+                self.indent += 1
+                self.indent_detail.append(indentor)
+            else:
+                indentor = None
+                # its not a "compound" keyword.  but lets also
+                # test for valid Python keywords that might be indenting us,
+                # else assume its a non-indenting line
+                m2 = self._re_indent_keyword.match(line)
+                if m2:
+                    self.indent += 1
+                    self.indent_detail.append(indentor)
+
+    def close(self):
+        """close this printer, flushing any remaining lines."""
+        self._flush_adjusted_lines()
+
+    def _is_unindentor(self, line):
+        """return true if the given line is an 'unindentor',
+        relative to the last 'indent' event received.
+
+        """
+
+        # no indentation detail has been pushed on; return False
+        if len(self.indent_detail) == 0:
+            return False
+
+        indentor = self.indent_detail[-1]
+
+        # the last indent keyword we grabbed is not a
+        # compound statement keyword; return False
+        if indentor is None:
+            return False
+
+        # if the current line doesnt have one of the "unindentor" keywords,
+        # return False
+        match = self._re_unindentor.match(line)
+        # if True, whitespace matches up, we have a compound indentor,
+        # and this line has an unindentor, this
+        # is probably good enough
+        return bool(match)
+
+        # should we decide that its not good enough, heres
+        # more stuff to check.
+        # keyword = match.group(1)
+
+        # match the original indent keyword
+        # for crit in [
+        #   (r'if|elif', r'else|elif'),
+        #   (r'try', r'except|finally|else'),
+        #   (r'while|for', r'else'),
+        # ]:
+        #   if re.match(crit[0], indentor) and re.match(crit[1], keyword):
+        #        return True
+
+        # return False
+
+    def _indent_line(self, line, stripspace=""):
+        """indent the given line according to the current indent level.
+
+        stripspace is a string of space that will be truncated from the
+        start of the line before indenting."""
+        if stripspace == "":
+            # Fast path optimization.
+            return self.indentstring * self.indent + line
+
+        return re.sub(
+            r"^%s" % stripspace, self.indentstring * self.indent, line
+        )
+
+    def _reset_multi_line_flags(self):
+        """reset the flags which would indicate we are in a backslashed
+        or triple-quoted section."""
+
+        self.backslashed, self.triplequoted = False, False
+
+    def _in_multi_line(self, line):
+        """return true if the given line is part of a multi-line block,
+        via backslash or triple-quote."""
+
+        # we are only looking for explicitly joined lines here, not
+        # implicit ones (i.e. brackets, braces etc.).  this is just to
+        # guard against the possibility of modifying the space inside of
+        # a literal multiline string with unfortunately placed
+        # whitespace
+
+        current_state = self.backslashed or self.triplequoted
+
+        self.backslashed = bool(re.search(r"\\$", line))
+        triples = len(re.findall(r"\"\"\"|\'\'\'", line))
+        if triples == 1 or triples % 2 != 0:
+            self.triplequoted = not self.triplequoted
+
+        return current_state
+
+    def _flush_adjusted_lines(self):
+        stripspace = None
+        self._reset_multi_line_flags()
+
+        for entry in self.line_buffer:
+            if self._in_multi_line(entry):
+                self.stream.write(entry + "\n")
+            else:
+                entry = entry.expandtabs()
+                if stripspace is None and re.search(r"^[ \t]*[^# \t]", entry):
+                    stripspace = re.match(r"^([ \t]*)", entry).group(1)
+                self.stream.write(self._indent_line(entry, stripspace) + "\n")
+
+        self.line_buffer = []
+        self._reset_multi_line_flags()
+
+
+def adjust_whitespace(text):
+    """remove the left-whitespace margin of a block of Python code."""
+
+    state = [False, False]
+    (backslashed, triplequoted) = (0, 1)
+
+    def in_multi_line(line):
+        start_state = state[backslashed] or state[triplequoted]
+
+        if re.search(r"\\$", line):
+            state[backslashed] = True
+        else:
+            state[backslashed] = False
+
+        def match(reg, t):
+            m = re.match(reg, t)
+            if m:
+                return m, t[len(m.group(0)) :]
+            else:
+                return None, t
+
+        while line:
+            if state[triplequoted]:
+                m, line = match(r"%s" % state[triplequoted], line)
+                if m:
+                    state[triplequoted] = False
+                else:
+                    m, line = match(r".*?(?=%s|$)" % state[triplequoted], line)
+            else:
+                m, line = match(r"#", line)
+                if m:
+                    return start_state
+
+                m, line = match(r"\"\"\"|\'\'\'", line)
+                if m:
+                    state[triplequoted] = m.group(0)
+                    continue
+
+                m, line = match(r".*?(?=\"\"\"|\'\'\'|#|$)", line)
+
+        return start_state
+
+    def _indent_line(line, stripspace=""):
+        return re.sub(r"^%s" % stripspace, "", line)
+
+    lines = []
+    stripspace = None
+
+    for line in re.split(r"\r?\n", text):
+        if in_multi_line(line):
+            lines.append(line)
+        else:
+            line = line.expandtabs()
+            if stripspace is None and re.search(r"^[ \t]*[^# \t]", line):
+                stripspace = re.match(r"^([ \t]*)", line).group(1)
+            lines.append(_indent_line(line, stripspace))
+    return "\n".join(lines)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/pyparser.py b/backend/venv/lib/python3.12/site-packages/mako/pyparser.py
new file mode 100644
index 00000000..bb9e72df
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/pyparser.py
@@ -0,0 +1,233 @@
+# mako/pyparser.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""Handles parsing of Python code.
+
+Parsing to AST is done via _ast on Python > 2.5, otherwise the compiler
+module is used.
+"""
+
+import operator
+
+import _ast
+
+from mako import _ast_util
+from mako import compat
+from mako import exceptions
+from mako import util
+
+# words that cannot be assigned to (notably
+# smaller than the total keys in __builtins__)
+reserved = {"True", "False", "None", "print"}
+
+# the "id" attribute on a function node
+arg_id = operator.attrgetter("arg")
+
+util.restore__ast(_ast)
+
+
+def parse(code, mode="exec", **exception_kwargs):
+    """Parse an expression into AST"""
+
+    try:
+        return _ast_util.parse(code, "<unknown>", mode)
+    except Exception as e:
+        raise exceptions.SyntaxException(
+            "(%s) %s (%r)"
+            % (
+                compat.exception_as().__class__.__name__,
+                compat.exception_as(),
+                code[0:50],
+            ),
+            **exception_kwargs,
+        ) from e
+
+
+class FindIdentifiers(_ast_util.NodeVisitor):
+    def __init__(self, listener, **exception_kwargs):
+        self.in_function = False
+        self.in_assign_targets = False
+        self.local_ident_stack = set()
+        self.listener = listener
+        self.exception_kwargs = exception_kwargs
+
+    def _add_declared(self, name):
+        if not self.in_function:
+            self.listener.declared_identifiers.add(name)
+        else:
+            self.local_ident_stack.add(name)
+
+    def visit_ClassDef(self, node):
+        self._add_declared(node.name)
+
+    def visit_Assign(self, node):
+        # flip around the visiting of Assign so the expression gets
+        # evaluated first, in the case of a clause like "x=x+5" (x
+        # is undeclared)
+
+        self.visit(node.value)
+        in_a = self.in_assign_targets
+        self.in_assign_targets = True
+        for n in node.targets:
+            self.visit(n)
+        self.in_assign_targets = in_a
+
+    def visit_ExceptHandler(self, node):
+        if node.name is not None:
+            self._add_declared(node.name)
+        if node.type is not None:
+            self.visit(node.type)
+        for statement in node.body:
+            self.visit(statement)
+
+    def visit_Lambda(self, node, *args):
+        self._visit_function(node, True)
+
+    def visit_FunctionDef(self, node):
+        self._add_declared(node.name)
+        self._visit_function(node, False)
+
+    def visit_ListComp(self, node):
+        if self.in_function:
+            for comp in node.generators:
+                self.visit(comp.iter)
+        else:
+            self.generic_visit(node)
+
+    visit_SetComp = visit_GeneratorExp = visit_ListComp
+
+    def visit_DictComp(self, node):
+        if self.in_function:
+            for comp in node.generators:
+                self.visit(comp.iter)
+        else:
+            self.generic_visit(node)
+
+    def _expand_tuples(self, args):
+        for arg in args:
+            if isinstance(arg, _ast.Tuple):
+                yield from arg.elts
+            else:
+                yield arg
+
+    def _visit_function(self, node, islambda):
+        # push function state onto stack.  dont log any more
+        # identifiers as "declared" until outside of the function,
+        # but keep logging identifiers as "undeclared". track
+        # argument names in each function header so they arent
+        # counted as "undeclared"
+
+        inf = self.in_function
+        self.in_function = True
+
+        local_ident_stack = self.local_ident_stack
+        self.local_ident_stack = local_ident_stack.union(
+            [arg_id(arg) for arg in self._expand_tuples(node.args.args)]
+        )
+        if islambda:
+            self.visit(node.body)
+        else:
+            for n in node.body:
+                self.visit(n)
+        self.in_function = inf
+        self.local_ident_stack = local_ident_stack
+
+    def visit_For(self, node):
+        # flip around visit
+
+        self.visit(node.iter)
+        self.visit(node.target)
+        for statement in node.body:
+            self.visit(statement)
+        for statement in node.orelse:
+            self.visit(statement)
+
+    def visit_Name(self, node):
+        if isinstance(node.ctx, _ast.Store):
+            # this is eqiuvalent to visit_AssName in
+            # compiler
+            self._add_declared(node.id)
+        elif (
+            node.id not in reserved
+            and node.id not in self.listener.declared_identifiers
+            and node.id not in self.local_ident_stack
+        ):
+            self.listener.undeclared_identifiers.add(node.id)
+
+    def visit_Import(self, node):
+        for name in node.names:
+            if name.asname is not None:
+                self._add_declared(name.asname)
+            else:
+                self._add_declared(name.name.split(".")[0])
+
+    def visit_ImportFrom(self, node):
+        for name in node.names:
+            if name.asname is not None:
+                self._add_declared(name.asname)
+            elif name.name == "*":
+                raise exceptions.CompileException(
+                    "'import *' is not supported, since all identifier "
+                    "names must be explicitly declared.  Please use the "
+                    "form 'from <modulename> import <name1>, <name2>, "
+                    "...' instead.",
+                    **self.exception_kwargs,
+                )
+            else:
+                self._add_declared(name.name)
+
+
+class FindTuple(_ast_util.NodeVisitor):
+    def __init__(self, listener, code_factory, **exception_kwargs):
+        self.listener = listener
+        self.exception_kwargs = exception_kwargs
+        self.code_factory = code_factory
+
+    def visit_Tuple(self, node):
+        for n in node.elts:
+            p = self.code_factory(n, **self.exception_kwargs)
+            self.listener.codeargs.append(p)
+            self.listener.args.append(ExpressionGenerator(n).value())
+            ldi = self.listener.declared_identifiers
+            self.listener.declared_identifiers = ldi.union(
+                p.declared_identifiers
+            )
+            lui = self.listener.undeclared_identifiers
+            self.listener.undeclared_identifiers = lui.union(
+                p.undeclared_identifiers
+            )
+
+
+class ParseFunc(_ast_util.NodeVisitor):
+    def __init__(self, listener, **exception_kwargs):
+        self.listener = listener
+        self.exception_kwargs = exception_kwargs
+
+    def visit_FunctionDef(self, node):
+        self.listener.funcname = node.name
+
+        argnames = [arg_id(arg) for arg in node.args.args]
+        if node.args.vararg:
+            argnames.append(node.args.vararg.arg)
+
+        kwargnames = [arg_id(arg) for arg in node.args.kwonlyargs]
+        if node.args.kwarg:
+            kwargnames.append(node.args.kwarg.arg)
+        self.listener.argnames = argnames
+        self.listener.defaults = node.args.defaults  # ast
+        self.listener.kwargnames = kwargnames
+        self.listener.kwdefaults = node.args.kw_defaults
+        self.listener.varargs = node.args.vararg
+        self.listener.kwargs = node.args.kwarg
+
+
+class ExpressionGenerator:
+    def __init__(self, astnode):
+        self.generator = _ast_util.SourceGenerator(" " * 4)
+        self.generator.visit(astnode)
+
+    def value(self):
+        return "".join(self.generator.result)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/runtime.py b/backend/venv/lib/python3.12/site-packages/mako/runtime.py
new file mode 100644
index 00000000..23401b70
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/runtime.py
@@ -0,0 +1,968 @@
+# mako/runtime.py
+# Copyright 2006-2020 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""provides runtime services for templates, including Context,
+Namespace, and various helper functions."""
+
+import builtins
+import functools
+import sys
+
+from mako import compat
+from mako import exceptions
+from mako import util
+
+
+class Context:
+
+    """Provides runtime namespace, output buffer, and various
+    callstacks for templates.
+
+    See :ref:`runtime_toplevel` for detail on the usage of
+    :class:`.Context`.
+
+    """
+
+    def __init__(self, buffer, **data):
+        self._buffer_stack = [buffer]
+
+        self._data = data
+
+        self._kwargs = data.copy()
+        self._with_template = None
+        self._outputting_as_unicode = None
+        self.namespaces = {}
+
+        # "capture" function which proxies to the
+        # generic "capture" function
+        self._data["capture"] = functools.partial(capture, self)
+
+        # "caller" stack used by def calls with content
+        self.caller_stack = self._data["caller"] = CallerStack()
+
+    def _set_with_template(self, t):
+        self._with_template = t
+        illegal_names = t.reserved_names.intersection(self._data)
+        if illegal_names:
+            raise exceptions.NameConflictError(
+                "Reserved words passed to render(): %s"
+                % ", ".join(illegal_names)
+            )
+
+    @property
+    def lookup(self):
+        """Return the :class:`.TemplateLookup` associated
+        with this :class:`.Context`.
+
+        """
+        return self._with_template.lookup
+
+    @property
+    def kwargs(self):
+        """Return the dictionary of top level keyword arguments associated
+        with this :class:`.Context`.
+
+        This dictionary only includes the top-level arguments passed to
+        :meth:`.Template.render`.  It does not include names produced within
+        the template execution such as local variable names or special names
+        such as ``self``, ``next``, etc.
+
+        The purpose of this dictionary is primarily for the case that
+        a :class:`.Template` accepts arguments via its ``<%page>`` tag,
+        which are normally expected to be passed via :meth:`.Template.render`,
+        except the template is being called in an inheritance context,
+        using the ``body()`` method.   :attr:`.Context.kwargs` can then be
+        used to propagate these arguments to the inheriting template::
+
+            ${next.body(**context.kwargs)}
+
+        """
+        return self._kwargs.copy()
+
+    def push_caller(self, caller):
+        """Push a ``caller`` callable onto the callstack for
+        this :class:`.Context`."""
+
+        self.caller_stack.append(caller)
+
+    def pop_caller(self):
+        """Pop a ``caller`` callable onto the callstack for this
+        :class:`.Context`."""
+
+        del self.caller_stack[-1]
+
+    def keys(self):
+        """Return a list of all names established in this :class:`.Context`."""
+
+        return list(self._data.keys())
+
+    def __getitem__(self, key):
+        if key in self._data:
+            return self._data[key]
+        else:
+            return builtins.__dict__[key]
+
+    def _push_writer(self):
+        """push a capturing buffer onto this Context and return
+        the new writer function."""
+
+        buf = util.FastEncodingBuffer()
+        self._buffer_stack.append(buf)
+        return buf.write
+
+    def _pop_buffer_and_writer(self):
+        """pop the most recent capturing buffer from this Context
+        and return the current writer after the pop.
+
+        """
+
+        buf = self._buffer_stack.pop()
+        return buf, self._buffer_stack[-1].write
+
+    def _push_buffer(self):
+        """push a capturing buffer onto this Context."""
+
+        self._push_writer()
+
+    def _pop_buffer(self):
+        """pop the most recent capturing buffer from this Context."""
+
+        return self._buffer_stack.pop()
+
+    def get(self, key, default=None):
+        """Return a value from this :class:`.Context`."""
+
+        return self._data.get(key, builtins.__dict__.get(key, default))
+
+    def write(self, string):
+        """Write a string to this :class:`.Context` object's
+        underlying output buffer."""
+
+        self._buffer_stack[-1].write(string)
+
+    def writer(self):
+        """Return the current writer function."""
+
+        return self._buffer_stack[-1].write
+
+    def _copy(self):
+        c = Context.__new__(Context)
+        c._buffer_stack = self._buffer_stack
+        c._data = self._data.copy()
+        c._kwargs = self._kwargs
+        c._with_template = self._with_template
+        c._outputting_as_unicode = self._outputting_as_unicode
+        c.namespaces = self.namespaces
+        c.caller_stack = self.caller_stack
+        return c
+
+    def _locals(self, d):
+        """Create a new :class:`.Context` with a copy of this
+        :class:`.Context`'s current state,
+        updated with the given dictionary.
+
+        The :attr:`.Context.kwargs` collection remains
+        unaffected.
+
+
+        """
+
+        if not d:
+            return self
+        c = self._copy()
+        c._data.update(d)
+        return c
+
+    def _clean_inheritance_tokens(self):
+        """create a new copy of this :class:`.Context`. with
+        tokens related to inheritance state removed."""
+
+        c = self._copy()
+        x = c._data
+        x.pop("self", None)
+        x.pop("parent", None)
+        x.pop("next", None)
+        return c
+
+
+class CallerStack(list):
+    def __init__(self):
+        self.nextcaller = None
+
+    def __nonzero__(self):
+        return self.__bool__()
+
+    def __bool__(self):
+        return len(self) and self._get_caller() and True or False
+
+    def _get_caller(self):
+        # this method can be removed once
+        # codegen MAGIC_NUMBER moves past 7
+        return self[-1]
+
+    def __getattr__(self, key):
+        return getattr(self._get_caller(), key)
+
+    def _push_frame(self):
+        frame = self.nextcaller or None
+        self.append(frame)
+        self.nextcaller = None
+        return frame
+
+    def _pop_frame(self):
+        self.nextcaller = self.pop()
+
+
+class Undefined:
+
+    """Represents an undefined value in a template.
+
+    All template modules have a constant value
+    ``UNDEFINED`` present which is an instance of this
+    object.
+
+    """
+
+    def __str__(self):
+        raise NameError("Undefined")
+
+    def __nonzero__(self):
+        return self.__bool__()
+
+    def __bool__(self):
+        return False
+
+
+UNDEFINED = Undefined()
+STOP_RENDERING = ""
+
+
+class LoopStack:
+
+    """a stack for LoopContexts that implements the context manager protocol
+    to automatically pop off the top of the stack on context exit
+    """
+
+    def __init__(self):
+        self.stack = []
+
+    def _enter(self, iterable):
+        self._push(iterable)
+        return self._top
+
+    def _exit(self):
+        self._pop()
+        return self._top
+
+    @property
+    def _top(self):
+        if self.stack:
+            return self.stack[-1]
+        else:
+            return self
+
+    def _pop(self):
+        return self.stack.pop()
+
+    def _push(self, iterable):
+        new = LoopContext(iterable)
+        if self.stack:
+            new.parent = self.stack[-1]
+        return self.stack.append(new)
+
+    def __getattr__(self, key):
+        raise exceptions.RuntimeException("No loop context is established")
+
+    def __iter__(self):
+        return iter(self._top)
+
+
+class LoopContext:
+
+    """A magic loop variable.
+    Automatically accessible in any ``% for`` block.
+
+    See the section :ref:`loop_context` for usage
+    notes.
+
+    :attr:`parent` -> :class:`.LoopContext` or ``None``
+        The parent loop, if one exists.
+    :attr:`index` -> `int`
+        The 0-based iteration count.
+    :attr:`reverse_index` -> `int`
+        The number of iterations remaining.
+    :attr:`first` -> `bool`
+        ``True`` on the first iteration, ``False`` otherwise.
+    :attr:`last` -> `bool`
+        ``True`` on the last iteration, ``False`` otherwise.
+    :attr:`even` -> `bool`
+        ``True`` when ``index`` is even.
+    :attr:`odd` -> `bool`
+        ``True`` when ``index`` is odd.
+    """
+
+    def __init__(self, iterable):
+        self._iterable = iterable
+        self.index = 0
+        self.parent = None
+
+    def __iter__(self):
+        for i in self._iterable:
+            yield i
+            self.index += 1
+
+    @util.memoized_instancemethod
+    def __len__(self):
+        return len(self._iterable)
+
+    @property
+    def reverse_index(self):
+        return len(self) - self.index - 1
+
+    @property
+    def first(self):
+        return self.index == 0
+
+    @property
+    def last(self):
+        return self.index == len(self) - 1
+
+    @property
+    def even(self):
+        return not self.odd
+
+    @property
+    def odd(self):
+        return bool(self.index % 2)
+
+    def cycle(self, *values):
+        """Cycle through values as the loop progresses."""
+        if not values:
+            raise ValueError("You must provide values to cycle through")
+        return values[self.index % len(values)]
+
+
+class _NSAttr:
+    def __init__(self, parent):
+        self.__parent = parent
+
+    def __getattr__(self, key):
+        ns = self.__parent
+        while ns:
+            if hasattr(ns.module, key):
+                return getattr(ns.module, key)
+            else:
+                ns = ns.inherits
+        raise AttributeError(key)
+
+
+class Namespace:
+
+    """Provides access to collections of rendering methods, which
+    can be local, from other templates, or from imported modules.
+
+    To access a particular rendering method referenced by a
+    :class:`.Namespace`, use plain attribute access:
+
+    .. sourcecode:: mako
+
+      ${some_namespace.foo(x, y, z)}
+
+    :class:`.Namespace` also contains several built-in attributes
+    described here.
+
+    """
+
+    def __init__(
+        self,
+        name,
+        context,
+        callables=None,
+        inherits=None,
+        populate_self=True,
+        calling_uri=None,
+    ):
+        self.name = name
+        self.context = context
+        self.inherits = inherits
+        if callables is not None:
+            self.callables = {c.__name__: c for c in callables}
+
+    callables = ()
+
+    module = None
+    """The Python module referenced by this :class:`.Namespace`.
+
+    If the namespace references a :class:`.Template`, then
+    this module is the equivalent of ``template.module``,
+    i.e. the generated module for the template.
+
+    """
+
+    template = None
+    """The :class:`.Template` object referenced by this
+        :class:`.Namespace`, if any.
+
+    """
+
+    context = None
+    """The :class:`.Context` object for this :class:`.Namespace`.
+
+    Namespaces are often created with copies of contexts that
+    contain slightly different data, particularly in inheritance
+    scenarios. Using the :class:`.Context` off of a :class:`.Namespace` one
+    can traverse an entire chain of templates that inherit from
+    one-another.
+
+    """
+
+    filename = None
+    """The path of the filesystem file used for this
+    :class:`.Namespace`'s module or template.
+
+    If this is a pure module-based
+    :class:`.Namespace`, this evaluates to ``module.__file__``. If a
+    template-based namespace, it evaluates to the original
+    template file location.
+
+    """
+
+    uri = None
+    """The URI for this :class:`.Namespace`'s template.
+
+    I.e. whatever was sent to :meth:`.TemplateLookup.get_template()`.
+
+    This is the equivalent of :attr:`.Template.uri`.
+
+    """
+
+    _templateuri = None
+
+    @util.memoized_property
+    def attr(self):
+        """Access module level attributes by name.
+
+        This accessor allows templates to supply "scalar"
+        attributes which are particularly handy in inheritance
+        relationships.
+
+        .. seealso::
+
+            :ref:`inheritance_attr`
+
+            :ref:`namespace_attr_for_includes`
+
+        """
+        return _NSAttr(self)
+
+    def get_namespace(self, uri):
+        """Return a :class:`.Namespace` corresponding to the given ``uri``.
+
+        If the given ``uri`` is a relative URI (i.e. it does not
+        contain a leading slash ``/``), the ``uri`` is adjusted to
+        be relative to the ``uri`` of the namespace itself. This
+        method is therefore mostly useful off of the built-in
+        ``local`` namespace, described in :ref:`namespace_local`.
+
+        In
+        most cases, a template wouldn't need this function, and
+        should instead use the ``<%namespace>`` tag to load
+        namespaces. However, since all ``<%namespace>`` tags are
+        evaluated before the body of a template ever runs,
+        this method can be used to locate namespaces using
+        expressions that were generated within the body code of
+        the template, or to conditionally use a particular
+        namespace.
+
+        """
+        key = (self, uri)
+        if key in self.context.namespaces:
+            return self.context.namespaces[key]
+        ns = TemplateNamespace(
+            uri,
+            self.context._copy(),
+            templateuri=uri,
+            calling_uri=self._templateuri,
+        )
+        self.context.namespaces[key] = ns
+        return ns
+
+    def get_template(self, uri):
+        """Return a :class:`.Template` from the given ``uri``.
+
+        The ``uri`` resolution is relative to the ``uri`` of this
+        :class:`.Namespace` object's :class:`.Template`.
+
+        """
+        return _lookup_template(self.context, uri, self._templateuri)
+
+    def get_cached(self, key, **kwargs):
+        """Return a value from the :class:`.Cache` referenced by this
+        :class:`.Namespace` object's :class:`.Template`.
+
+        The advantage to this method versus direct access to the
+        :class:`.Cache` is that the configuration parameters
+        declared in ``<%page>`` take effect here, thereby calling
+        up the same configured backend as that configured
+        by ``<%page>``.
+
+        """
+
+        return self.cache.get(key, **kwargs)
+
+    @property
+    def cache(self):
+        """Return the :class:`.Cache` object referenced
+        by this :class:`.Namespace` object's
+        :class:`.Template`.
+
+        """
+        return self.template.cache
+
+    def include_file(self, uri, **kwargs):
+        """Include a file at the given ``uri``."""
+
+        _include_file(self.context, uri, self._templateuri, **kwargs)
+
+    def _populate(self, d, l):
+        for ident in l:
+            if ident == "*":
+                for k, v in self._get_star():
+                    d[k] = v
+            else:
+                d[ident] = getattr(self, ident)
+
+    def _get_star(self):
+        if self.callables:
+            for key in self.callables:
+                yield (key, self.callables[key])
+
+    def __getattr__(self, key):
+        if key in self.callables:
+            val = self.callables[key]
+        elif self.inherits:
+            val = getattr(self.inherits, key)
+        else:
+            raise AttributeError(
+                "Namespace '%s' has no member '%s'" % (self.name, key)
+            )
+        setattr(self, key, val)
+        return val
+
+
+class TemplateNamespace(Namespace):
+
+    """A :class:`.Namespace` specific to a :class:`.Template` instance."""
+
+    def __init__(
+        self,
+        name,
+        context,
+        template=None,
+        templateuri=None,
+        callables=None,
+        inherits=None,
+        populate_self=True,
+        calling_uri=None,
+    ):
+        self.name = name
+        self.context = context
+        self.inherits = inherits
+        if callables is not None:
+            self.callables = {c.__name__: c for c in callables}
+
+        if templateuri is not None:
+            self.template = _lookup_template(context, templateuri, calling_uri)
+            self._templateuri = self.template.module._template_uri
+        elif template is not None:
+            self.template = template
+            self._templateuri = template.module._template_uri
+        else:
+            raise TypeError("'template' argument is required.")
+
+        if populate_self:
+            lclcallable, lclcontext = _populate_self_namespace(
+                context, self.template, self_ns=self
+            )
+
+    @property
+    def module(self):
+        """The Python module referenced by this :class:`.Namespace`.
+
+        If the namespace references a :class:`.Template`, then
+        this module is the equivalent of ``template.module``,
+        i.e. the generated module for the template.
+
+        """
+        return self.template.module
+
+    @property
+    def filename(self):
+        """The path of the filesystem file used for this
+        :class:`.Namespace`'s module or template.
+        """
+        return self.template.filename
+
+    @property
+    def uri(self):
+        """The URI for this :class:`.Namespace`'s template.
+
+        I.e. whatever was sent to :meth:`.TemplateLookup.get_template()`.
+
+        This is the equivalent of :attr:`.Template.uri`.
+
+        """
+        return self.template.uri
+
+    def _get_star(self):
+        if self.callables:
+            for key in self.callables:
+                yield (key, self.callables[key])
+
+        def get(key):
+            callable_ = self.template._get_def_callable(key)
+            return functools.partial(callable_, self.context)
+
+        for k in self.template.module._exports:
+            yield (k, get(k))
+
+    def __getattr__(self, key):
+        if key in self.callables:
+            val = self.callables[key]
+        elif self.template.has_def(key):
+            callable_ = self.template._get_def_callable(key)
+            val = functools.partial(callable_, self.context)
+        elif self.inherits:
+            val = getattr(self.inherits, key)
+
+        else:
+            raise AttributeError(
+                "Namespace '%s' has no member '%s'" % (self.name, key)
+            )
+        setattr(self, key, val)
+        return val
+
+
+class ModuleNamespace(Namespace):
+
+    """A :class:`.Namespace` specific to a Python module instance."""
+
+    def __init__(
+        self,
+        name,
+        context,
+        module,
+        callables=None,
+        inherits=None,
+        populate_self=True,
+        calling_uri=None,
+    ):
+        self.name = name
+        self.context = context
+        self.inherits = inherits
+        if callables is not None:
+            self.callables = {c.__name__: c for c in callables}
+
+        mod = __import__(module)
+        for token in module.split(".")[1:]:
+            mod = getattr(mod, token)
+        self.module = mod
+
+    @property
+    def filename(self):
+        """The path of the filesystem file used for this
+        :class:`.Namespace`'s module or template.
+        """
+        return self.module.__file__
+
+    def _get_star(self):
+        if self.callables:
+            for key in self.callables:
+                yield (key, self.callables[key])
+        for key in dir(self.module):
+            if key[0] != "_":
+                callable_ = getattr(self.module, key)
+                if callable(callable_):
+                    yield key, functools.partial(callable_, self.context)
+
+    def __getattr__(self, key):
+        if key in self.callables:
+            val = self.callables[key]
+        elif hasattr(self.module, key):
+            callable_ = getattr(self.module, key)
+            val = functools.partial(callable_, self.context)
+        elif self.inherits:
+            val = getattr(self.inherits, key)
+        else:
+            raise AttributeError(
+                "Namespace '%s' has no member '%s'" % (self.name, key)
+            )
+        setattr(self, key, val)
+        return val
+
+
+def supports_caller(func):
+    """Apply a caller_stack compatibility decorator to a plain
+    Python function.
+
+    See the example in :ref:`namespaces_python_modules`.
+
+    """
+
+    def wrap_stackframe(context, *args, **kwargs):
+        context.caller_stack._push_frame()
+        try:
+            return func(context, *args, **kwargs)
+        finally:
+            context.caller_stack._pop_frame()
+
+    return wrap_stackframe
+
+
+def capture(context, callable_, *args, **kwargs):
+    """Execute the given template def, capturing the output into
+    a buffer.
+
+    See the example in :ref:`namespaces_python_modules`.
+
+    """
+
+    if not callable(callable_):
+        raise exceptions.RuntimeException(
+            "capture() function expects a callable as "
+            "its argument (i.e. capture(func, *args, **kwargs))"
+        )
+    context._push_buffer()
+    try:
+        callable_(*args, **kwargs)
+    finally:
+        buf = context._pop_buffer()
+    return buf.getvalue()
+
+
+def _decorate_toplevel(fn):
+    def decorate_render(render_fn):
+        def go(context, *args, **kw):
+            def y(*args, **kw):
+                return render_fn(context, *args, **kw)
+
+            try:
+                y.__name__ = render_fn.__name__[7:]
+            except TypeError:
+                # < Python 2.4
+                pass
+            return fn(y)(context, *args, **kw)
+
+        return go
+
+    return decorate_render
+
+
+def _decorate_inline(context, fn):
+    def decorate_render(render_fn):
+        dec = fn(render_fn)
+
+        def go(*args, **kw):
+            return dec(context, *args, **kw)
+
+        return go
+
+    return decorate_render
+
+
+def _include_file(context, uri, calling_uri, **kwargs):
+    """locate the template from the given uri and include it in
+    the current output."""
+
+    template = _lookup_template(context, uri, calling_uri)
+    (callable_, ctx) = _populate_self_namespace(
+        context._clean_inheritance_tokens(), template
+    )
+    kwargs = _kwargs_for_include(callable_, context._data, **kwargs)
+    if template.include_error_handler:
+        try:
+            callable_(ctx, **kwargs)
+        except Exception:
+            result = template.include_error_handler(ctx, compat.exception_as())
+            if not result:
+                raise
+    else:
+        callable_(ctx, **kwargs)
+
+
+def _inherit_from(context, uri, calling_uri):
+    """called by the _inherit method in template modules to set
+    up the inheritance chain at the start of a template's
+    execution."""
+
+    if uri is None:
+        return None
+    template = _lookup_template(context, uri, calling_uri)
+    self_ns = context["self"]
+    ih = self_ns
+    while ih.inherits is not None:
+        ih = ih.inherits
+    lclcontext = context._locals({"next": ih})
+    ih.inherits = TemplateNamespace(
+        "self:%s" % template.uri,
+        lclcontext,
+        template=template,
+        populate_self=False,
+    )
+    context._data["parent"] = lclcontext._data["local"] = ih.inherits
+    callable_ = getattr(template.module, "_mako_inherit", None)
+    if callable_ is not None:
+        ret = callable_(template, lclcontext)
+        if ret:
+            return ret
+
+    gen_ns = getattr(template.module, "_mako_generate_namespaces", None)
+    if gen_ns is not None:
+        gen_ns(context)
+    return (template.callable_, lclcontext)
+
+
+def _lookup_template(context, uri, relativeto):
+    lookup = context._with_template.lookup
+    if lookup is None:
+        raise exceptions.TemplateLookupException(
+            "Template '%s' has no TemplateLookup associated"
+            % context._with_template.uri
+        )
+    uri = lookup.adjust_uri(uri, relativeto)
+    try:
+        return lookup.get_template(uri)
+    except exceptions.TopLevelLookupException as e:
+        raise exceptions.TemplateLookupException(
+            str(compat.exception_as())
+        ) from e
+
+
+def _populate_self_namespace(context, template, self_ns=None):
+    if self_ns is None:
+        self_ns = TemplateNamespace(
+            "self:%s" % template.uri,
+            context,
+            template=template,
+            populate_self=False,
+        )
+    context._data["self"] = context._data["local"] = self_ns
+    if hasattr(template.module, "_mako_inherit"):
+        ret = template.module._mako_inherit(template, context)
+        if ret:
+            return ret
+    return (template.callable_, context)
+
+
+def _render(template, callable_, args, data, as_unicode=False):
+    """create a Context and return the string
+    output of the given template and template callable."""
+
+    if as_unicode:
+        buf = util.FastEncodingBuffer()
+    else:
+        buf = util.FastEncodingBuffer(
+            encoding=template.output_encoding, errors=template.encoding_errors
+        )
+    context = Context(buf, **data)
+    context._outputting_as_unicode = as_unicode
+    context._set_with_template(template)
+
+    _render_context(
+        template,
+        callable_,
+        context,
+        *args,
+        **_kwargs_for_callable(callable_, data),
+    )
+    return context._pop_buffer().getvalue()
+
+
+def _kwargs_for_callable(callable_, data):
+    argspec = compat.inspect_getargspec(callable_)
+    # for normal pages, **pageargs is usually present
+    if argspec[2]:
+        return data
+
+    # for rendering defs from the top level, figure out the args
+    namedargs = argspec[0] + [v for v in argspec[1:3] if v is not None]
+    kwargs = {}
+    for arg in namedargs:
+        if arg != "context" and arg in data and arg not in kwargs:
+            kwargs[arg] = data[arg]
+    return kwargs
+
+
+def _kwargs_for_include(callable_, data, **kwargs):
+    argspec = compat.inspect_getargspec(callable_)
+    namedargs = argspec[0] + [v for v in argspec[1:3] if v is not None]
+    for arg in namedargs:
+        if arg != "context" and arg in data and arg not in kwargs:
+            kwargs[arg] = data[arg]
+    return kwargs
+
+
+def _render_context(tmpl, callable_, context, *args, **kwargs):
+    import mako.template as template
+
+    # create polymorphic 'self' namespace for this
+    # template with possibly updated context
+    if not isinstance(tmpl, template.DefTemplate):
+        # if main render method, call from the base of the inheritance stack
+        (inherit, lclcontext) = _populate_self_namespace(context, tmpl)
+        _exec_template(inherit, lclcontext, args=args, kwargs=kwargs)
+    else:
+        # otherwise, call the actual rendering method specified
+        (inherit, lclcontext) = _populate_self_namespace(context, tmpl.parent)
+        _exec_template(callable_, context, args=args, kwargs=kwargs)
+
+
+def _exec_template(callable_, context, args=None, kwargs=None):
+    """execute a rendering callable given the callable, a
+    Context, and optional explicit arguments
+
+    the contextual Template will be located if it exists, and
+    the error handling options specified on that Template will
+    be interpreted here.
+    """
+    template = context._with_template
+    if template is not None and (
+        template.format_exceptions or template.error_handler
+    ):
+        try:
+            callable_(context, *args, **kwargs)
+        except Exception:
+            _render_error(template, context, compat.exception_as())
+        except:
+            e = sys.exc_info()[0]
+            _render_error(template, context, e)
+    else:
+        callable_(context, *args, **kwargs)
+
+
+def _render_error(template, context, error):
+    if template.error_handler:
+        result = template.error_handler(context, error)
+        if not result:
+            tp, value, tb = sys.exc_info()
+            if value and tb:
+                raise value.with_traceback(tb)
+            else:
+                raise error
+    else:
+        error_template = exceptions.html_error_template()
+        if context._outputting_as_unicode:
+            context._buffer_stack[:] = [util.FastEncodingBuffer()]
+        else:
+            context._buffer_stack[:] = [
+                util.FastEncodingBuffer(
+                    error_template.output_encoding,
+                    error_template.encoding_errors,
+                )
+            ]
+
+        context._set_with_template(error_template)
+        error_template.render_context(context, error=error)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/template.py b/backend/venv/lib/python3.12/site-packages/mako/template.py
new file mode 100644
index 00000000..82c7cba8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/template.py
@@ -0,0 +1,711 @@
+# mako/template.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+
+"""Provides the Template class, a facade for parsing, generating and executing
+template strings, as well as template runtime operations."""
+
+import json
+import os
+import re
+import shutil
+import stat
+import tempfile
+import types
+import weakref
+
+from mako import cache
+from mako import codegen
+from mako import compat
+from mako import exceptions
+from mako import runtime
+from mako import util
+from mako.lexer import Lexer
+
+
+class Template:
+    r"""Represents a compiled template.
+
+    :class:`.Template` includes a reference to the original
+    template source (via the :attr:`.source` attribute)
+    as well as the source code of the
+    generated Python module (i.e. the :attr:`.code` attribute),
+    as well as a reference to an actual Python module.
+
+    :class:`.Template` is constructed using either a literal string
+    representing the template text, or a filename representing a filesystem
+    path to a source file.
+
+    :param text: textual template source.  This argument is mutually
+     exclusive versus the ``filename`` parameter.
+
+    :param filename: filename of the source template.  This argument is
+     mutually exclusive versus the ``text`` parameter.
+
+    :param buffer_filters: string list of filters to be applied
+     to the output of ``%def``\ s which are buffered, cached, or otherwise
+     filtered, after all filters
+     defined with the ``%def`` itself have been applied. Allows the
+     creation of default expression filters that let the output
+     of return-valued ``%def``\ s "opt out" of that filtering via
+     passing special attributes or objects.
+
+    :param cache_args: Dictionary of cache configuration arguments that
+     will be passed to the :class:`.CacheImpl`.   See :ref:`caching_toplevel`.
+
+    :param cache_dir:
+
+     .. deprecated:: 0.6
+        Use the ``'dir'`` argument in the ``cache_args`` dictionary.
+        See :ref:`caching_toplevel`.
+
+    :param cache_enabled: Boolean flag which enables caching of this
+     template.  See :ref:`caching_toplevel`.
+
+    :param cache_impl: String name of a :class:`.CacheImpl` caching
+     implementation to use.   Defaults to ``'beaker'``.
+
+    :param cache_type:
+
+     .. deprecated:: 0.6
+        Use the ``'type'`` argument in the ``cache_args`` dictionary.
+        See :ref:`caching_toplevel`.
+
+    :param cache_url:
+
+     .. deprecated:: 0.6
+        Use the ``'url'`` argument in the ``cache_args`` dictionary.
+        See :ref:`caching_toplevel`.
+
+    :param default_filters: List of string filter names that will
+     be applied to all expressions.  See :ref:`filtering_default_filters`.
+
+    :param enable_loop: When ``True``, enable the ``loop`` context variable.
+     This can be set to ``False`` to support templates that may
+     be making usage of the name "``loop``".   Individual templates can
+     re-enable the "loop" context by placing the directive
+     ``enable_loop="True"`` inside the ``<%page>`` tag -- see
+     :ref:`migrating_loop`.
+
+    :param encoding_errors: Error parameter passed to ``encode()`` when
+     string encoding is performed. See :ref:`usage_unicode`.
+
+    :param error_handler: Python callable which is called whenever
+     compile or runtime exceptions occur. The callable is passed
+     the current context as well as the exception. If the
+     callable returns ``True``, the exception is considered to
+     be handled, else it is re-raised after the function
+     completes. Is used to provide custom error-rendering
+     functions.
+
+     .. seealso::
+
+        :paramref:`.Template.include_error_handler` - include-specific
+        error handler function
+
+    :param format_exceptions: if ``True``, exceptions which occur during
+     the render phase of this template will be caught and
+     formatted into an HTML error page, which then becomes the
+     rendered result of the :meth:`.render` call. Otherwise,
+     runtime exceptions are propagated outwards.
+
+    :param imports: String list of Python statements, typically individual
+     "import" lines, which will be placed into the module level
+     preamble of all generated Python modules. See the example
+     in :ref:`filtering_default_filters`.
+
+    :param future_imports: String list of names to import from `__future__`.
+     These will be concatenated into a comma-separated string and inserted
+     into the beginning of the template, e.g. ``futures_imports=['FOO',
+     'BAR']`` results in ``from __future__ import FOO, BAR``.
+
+    :param include_error_handler: An error handler that runs when this template
+     is included within another one via the ``<%include>`` tag, and raises an
+     error.  Compare to the :paramref:`.Template.error_handler` option.
+
+     .. versionadded:: 1.0.6
+
+     .. seealso::
+
+        :paramref:`.Template.error_handler` - top-level error handler function
+
+    :param input_encoding: Encoding of the template's source code.  Can
+     be used in lieu of the coding comment. See
+     :ref:`usage_unicode` as well as :ref:`unicode_toplevel` for
+     details on source encoding.
+
+    :param lookup: a :class:`.TemplateLookup` instance that will be used
+     for all file lookups via the ``<%namespace>``,
+     ``<%include>``, and ``<%inherit>`` tags. See
+     :ref:`usage_templatelookup`.
+
+    :param module_directory: Filesystem location where generated
+     Python module files will be placed.
+
+    :param module_filename: Overrides the filename of the generated
+     Python module file. For advanced usage only.
+
+    :param module_writer: A callable which overrides how the Python
+     module is written entirely.  The callable is passed the
+     encoded source content of the module and the destination
+     path to be written to.   The default behavior of module writing
+     uses a tempfile in conjunction with a file move in order
+     to make the operation atomic.   So a user-defined module
+     writing function that mimics the default behavior would be:
+
+     .. sourcecode:: python
+
+         import tempfile
+         import os
+         import shutil
+
+         def module_writer(source, outputpath):
+             (dest, name) = \\
+                 tempfile.mkstemp(
+                     dir=os.path.dirname(outputpath)
+                 )
+
+             os.write(dest, source)
+             os.close(dest)
+             shutil.move(name, outputpath)
+
+         from mako.template import Template
+         mytemplate = Template(
+                         filename="index.html",
+                         module_directory="/path/to/modules",
+                         module_writer=module_writer
+                     )
+
+     The function is provided for unusual configurations where
+     certain platform-specific permissions or other special
+     steps are needed.
+
+    :param output_encoding: The encoding to use when :meth:`.render`
+     is called.
+     See :ref:`usage_unicode` as well as :ref:`unicode_toplevel`.
+
+    :param preprocessor: Python callable which will be passed
+     the full template source before it is parsed. The return
+     result of the callable will be used as the template source
+     code.
+
+    :param lexer_cls: A :class:`.Lexer` class used to parse
+     the template.   The :class:`.Lexer` class is used by
+     default.
+
+     .. versionadded:: 0.7.4
+
+    :param strict_undefined: Replaces the automatic usage of
+     ``UNDEFINED`` for any undeclared variables not located in
+     the :class:`.Context` with an immediate raise of
+     ``NameError``. The advantage is immediate reporting of
+     missing variables which include the name.
+
+     .. versionadded:: 0.3.6
+
+    :param uri: string URI or other identifier for this template.
+     If not provided, the ``uri`` is generated from the filesystem
+     path, or from the in-memory identity of a non-file-based
+     template. The primary usage of the ``uri`` is to provide a key
+     within :class:`.TemplateLookup`, as well as to generate the
+     file path of the generated Python module file, if
+     ``module_directory`` is specified.
+
+    """
+
+    lexer_cls = Lexer
+
+    def __init__(
+        self,
+        text=None,
+        filename=None,
+        uri=None,
+        format_exceptions=False,
+        error_handler=None,
+        lookup=None,
+        output_encoding=None,
+        encoding_errors="strict",
+        module_directory=None,
+        cache_args=None,
+        cache_impl="beaker",
+        cache_enabled=True,
+        cache_type=None,
+        cache_dir=None,
+        cache_url=None,
+        module_filename=None,
+        input_encoding=None,
+        module_writer=None,
+        default_filters=None,
+        buffer_filters=(),
+        strict_undefined=False,
+        imports=None,
+        future_imports=None,
+        enable_loop=True,
+        preprocessor=None,
+        lexer_cls=None,
+        include_error_handler=None,
+    ):
+        if uri:
+            self.module_id = re.sub(r"\W", "_", uri)
+            self.uri = uri
+        elif filename:
+            self.module_id = re.sub(r"\W", "_", filename)
+            drive, path = os.path.splitdrive(filename)
+            path = os.path.normpath(path).replace(os.path.sep, "/")
+            self.uri = path
+        else:
+            self.module_id = "memory:" + hex(id(self))
+            self.uri = self.module_id
+
+        u_norm = self.uri
+        if u_norm.startswith("/"):
+            u_norm = u_norm[1:]
+        u_norm = os.path.normpath(u_norm)
+        if u_norm.startswith(".."):
+            raise exceptions.TemplateLookupException(
+                'Template uri "%s" is invalid - '
+                "it cannot be relative outside "
+                "of the root path." % self.uri
+            )
+
+        self.input_encoding = input_encoding
+        self.output_encoding = output_encoding
+        self.encoding_errors = encoding_errors
+        self.enable_loop = enable_loop
+        self.strict_undefined = strict_undefined
+        self.module_writer = module_writer
+
+        if default_filters is None:
+            self.default_filters = ["str"]
+        else:
+            self.default_filters = default_filters
+        self.buffer_filters = buffer_filters
+
+        self.imports = imports
+        self.future_imports = future_imports
+        self.preprocessor = preprocessor
+
+        if lexer_cls is not None:
+            self.lexer_cls = lexer_cls
+
+        # if plain text, compile code in memory only
+        if text is not None:
+            (code, module) = _compile_text(self, text, filename)
+            self._code = code
+            self._source = text
+            ModuleInfo(module, None, self, filename, code, text, uri)
+        elif filename is not None:
+            # if template filename and a module directory, load
+            # a filesystem-based module file, generating if needed
+            if module_filename is not None:
+                path = module_filename
+            elif module_directory is not None:
+                path = os.path.abspath(
+                    os.path.join(
+                        os.path.normpath(module_directory), u_norm + ".py"
+                    )
+                )
+            else:
+                path = None
+            module = self._compile_from_file(path, filename)
+        else:
+            raise exceptions.RuntimeException(
+                "Template requires text or filename"
+            )
+
+        self.module = module
+        self.filename = filename
+        self.callable_ = self.module.render_body
+        self.format_exceptions = format_exceptions
+        self.error_handler = error_handler
+        self.include_error_handler = include_error_handler
+        self.lookup = lookup
+
+        self.module_directory = module_directory
+
+        self._setup_cache_args(
+            cache_impl,
+            cache_enabled,
+            cache_args,
+            cache_type,
+            cache_dir,
+            cache_url,
+        )
+
+    @util.memoized_property
+    def reserved_names(self):
+        if self.enable_loop:
+            return codegen.RESERVED_NAMES
+        else:
+            return codegen.RESERVED_NAMES.difference(["loop"])
+
+    def _setup_cache_args(
+        self,
+        cache_impl,
+        cache_enabled,
+        cache_args,
+        cache_type,
+        cache_dir,
+        cache_url,
+    ):
+        self.cache_impl = cache_impl
+        self.cache_enabled = cache_enabled
+        self.cache_args = cache_args or {}
+        # transfer deprecated cache_* args
+        if cache_type:
+            self.cache_args["type"] = cache_type
+        if cache_dir:
+            self.cache_args["dir"] = cache_dir
+        if cache_url:
+            self.cache_args["url"] = cache_url
+
+    def _compile_from_file(self, path, filename):
+        if path is not None:
+            util.verify_directory(os.path.dirname(path))
+            filemtime = os.stat(filename)[stat.ST_MTIME]
+            if (
+                not os.path.exists(path)
+                or os.stat(path)[stat.ST_MTIME] < filemtime
+            ):
+                data = util.read_file(filename)
+                _compile_module_file(
+                    self, data, filename, path, self.module_writer
+                )
+            module = compat.load_module(self.module_id, path)
+            if module._magic_number != codegen.MAGIC_NUMBER:
+                data = util.read_file(filename)
+                _compile_module_file(
+                    self, data, filename, path, self.module_writer
+                )
+                module = compat.load_module(self.module_id, path)
+            ModuleInfo(module, path, self, filename, None, None, None)
+        else:
+            # template filename and no module directory, compile code
+            # in memory
+            data = util.read_file(filename)
+            code, module = _compile_text(self, data, filename)
+            self._source = None
+            self._code = code
+            ModuleInfo(module, None, self, filename, code, None, None)
+        return module
+
+    @property
+    def source(self):
+        """Return the template source code for this :class:`.Template`."""
+
+        return _get_module_info_from_callable(self.callable_).source
+
+    @property
+    def code(self):
+        """Return the module source code for this :class:`.Template`."""
+
+        return _get_module_info_from_callable(self.callable_).code
+
+    @util.memoized_property
+    def cache(self):
+        return cache.Cache(self)
+
+    @property
+    def cache_dir(self):
+        return self.cache_args["dir"]
+
+    @property
+    def cache_url(self):
+        return self.cache_args["url"]
+
+    @property
+    def cache_type(self):
+        return self.cache_args["type"]
+
+    def render(self, *args, **data):
+        """Render the output of this template as a string.
+
+        If the template specifies an output encoding, the string
+        will be encoded accordingly, else the output is raw (raw
+        output uses `StringIO` and can't handle multibyte
+        characters). A :class:`.Context` object is created corresponding
+        to the given data. Arguments that are explicitly declared
+        by this template's internal rendering method are also
+        pulled from the given ``*args``, ``**data`` members.
+
+        """
+        return runtime._render(self, self.callable_, args, data)
+
+    def render_unicode(self, *args, **data):
+        """Render the output of this template as a unicode object."""
+
+        return runtime._render(
+            self, self.callable_, args, data, as_unicode=True
+        )
+
+    def render_context(self, context, *args, **kwargs):
+        """Render this :class:`.Template` with the given context.
+
+        The data is written to the context's buffer.
+
+        """
+        if getattr(context, "_with_template", None) is None:
+            context._set_with_template(self)
+        runtime._render_context(self, self.callable_, context, *args, **kwargs)
+
+    def has_def(self, name):
+        return hasattr(self.module, "render_%s" % name)
+
+    def get_def(self, name):
+        """Return a def of this template as a :class:`.DefTemplate`."""
+
+        return DefTemplate(self, getattr(self.module, "render_%s" % name))
+
+    def list_defs(self):
+        """return a list of defs in the template.
+
+        .. versionadded:: 1.0.4
+
+        """
+        return [i[7:] for i in dir(self.module) if i[:7] == "render_"]
+
+    def _get_def_callable(self, name):
+        return getattr(self.module, "render_%s" % name)
+
+    @property
+    def last_modified(self):
+        return self.module._modified_time
+
+
+class ModuleTemplate(Template):
+
+    """A Template which is constructed given an existing Python module.
+
+    e.g.::
+
+         t = Template("this is a template")
+         f = file("mymodule.py", "w")
+         f.write(t.code)
+         f.close()
+
+         import mymodule
+
+         t = ModuleTemplate(mymodule)
+         print(t.render())
+
+    """
+
+    def __init__(
+        self,
+        module,
+        module_filename=None,
+        template=None,
+        template_filename=None,
+        module_source=None,
+        template_source=None,
+        output_encoding=None,
+        encoding_errors="strict",
+        format_exceptions=False,
+        error_handler=None,
+        lookup=None,
+        cache_args=None,
+        cache_impl="beaker",
+        cache_enabled=True,
+        cache_type=None,
+        cache_dir=None,
+        cache_url=None,
+        include_error_handler=None,
+    ):
+        self.module_id = re.sub(r"\W", "_", module._template_uri)
+        self.uri = module._template_uri
+        self.input_encoding = module._source_encoding
+        self.output_encoding = output_encoding
+        self.encoding_errors = encoding_errors
+        self.enable_loop = module._enable_loop
+
+        self.module = module
+        self.filename = template_filename
+        ModuleInfo(
+            module,
+            module_filename,
+            self,
+            template_filename,
+            module_source,
+            template_source,
+            module._template_uri,
+        )
+
+        self.callable_ = self.module.render_body
+        self.format_exceptions = format_exceptions
+        self.error_handler = error_handler
+        self.include_error_handler = include_error_handler
+        self.lookup = lookup
+        self._setup_cache_args(
+            cache_impl,
+            cache_enabled,
+            cache_args,
+            cache_type,
+            cache_dir,
+            cache_url,
+        )
+
+
+class DefTemplate(Template):
+
+    """A :class:`.Template` which represents a callable def in a parent
+    template."""
+
+    def __init__(self, parent, callable_):
+        self.parent = parent
+        self.callable_ = callable_
+        self.output_encoding = parent.output_encoding
+        self.module = parent.module
+        self.encoding_errors = parent.encoding_errors
+        self.format_exceptions = parent.format_exceptions
+        self.error_handler = parent.error_handler
+        self.include_error_handler = parent.include_error_handler
+        self.enable_loop = parent.enable_loop
+        self.lookup = parent.lookup
+
+    def get_def(self, name):
+        return self.parent.get_def(name)
+
+
+class ModuleInfo:
+
+    """Stores information about a module currently loaded into
+    memory, provides reverse lookups of template source, module
+    source code based on a module's identifier.
+
+    """
+
+    _modules = weakref.WeakValueDictionary()
+
+    def __init__(
+        self,
+        module,
+        module_filename,
+        template,
+        template_filename,
+        module_source,
+        template_source,
+        template_uri,
+    ):
+        self.module = module
+        self.module_filename = module_filename
+        self.template_filename = template_filename
+        self.module_source = module_source
+        self.template_source = template_source
+        self.template_uri = template_uri
+        self._modules[module.__name__] = template._mmarker = self
+        if module_filename:
+            self._modules[module_filename] = self
+
+    @classmethod
+    def get_module_source_metadata(cls, module_source, full_line_map=False):
+        source_map = re.search(
+            r"__M_BEGIN_METADATA(.+?)__M_END_METADATA", module_source, re.S
+        ).group(1)
+        source_map = json.loads(source_map)
+        source_map["line_map"] = {
+            int(k): int(v) for k, v in source_map["line_map"].items()
+        }
+        if full_line_map:
+            f_line_map = source_map["full_line_map"] = []
+            line_map = source_map["line_map"]
+
+            curr_templ_line = 1
+            for mod_line in range(1, max(line_map)):
+                if mod_line in line_map:
+                    curr_templ_line = line_map[mod_line]
+                f_line_map.append(curr_templ_line)
+        return source_map
+
+    @property
+    def code(self):
+        if self.module_source is not None:
+            return self.module_source
+        else:
+            return util.read_python_file(self.module_filename)
+
+    @property
+    def source(self):
+        if self.template_source is None:
+            data = util.read_file(self.template_filename)
+            if self.module._source_encoding:
+                return data.decode(self.module._source_encoding)
+            else:
+                return data
+
+        elif self.module._source_encoding and not isinstance(
+            self.template_source, str
+        ):
+            return self.template_source.decode(self.module._source_encoding)
+        else:
+            return self.template_source
+
+
+def _compile(template, text, filename, generate_magic_comment):
+    lexer = template.lexer_cls(
+        text,
+        filename,
+        input_encoding=template.input_encoding,
+        preprocessor=template.preprocessor,
+    )
+    node = lexer.parse()
+    source = codegen.compile(
+        node,
+        template.uri,
+        filename,
+        default_filters=template.default_filters,
+        buffer_filters=template.buffer_filters,
+        imports=template.imports,
+        future_imports=template.future_imports,
+        source_encoding=lexer.encoding,
+        generate_magic_comment=generate_magic_comment,
+        strict_undefined=template.strict_undefined,
+        enable_loop=template.enable_loop,
+        reserved_names=template.reserved_names,
+    )
+    return source, lexer
+
+
+def _compile_text(template, text, filename):
+    identifier = template.module_id
+    source, lexer = _compile(
+        template, text, filename, generate_magic_comment=False
+    )
+
+    cid = identifier
+    module = types.ModuleType(cid)
+    code = compile(source, cid, "exec")
+
+    # this exec() works for 2.4->3.3.
+    exec(code, module.__dict__, module.__dict__)
+    return (source, module)
+
+
+def _compile_module_file(template, text, filename, outputpath, module_writer):
+    source, lexer = _compile(
+        template, text, filename, generate_magic_comment=True
+    )
+
+    if isinstance(source, str):
+        source = source.encode(lexer.encoding or "ascii")
+
+    if module_writer:
+        module_writer(source, outputpath)
+    else:
+        # make tempfiles in the same location as the ultimate
+        # location.   this ensures they're on the same filesystem,
+        # avoiding synchronization issues.
+        (dest, name) = tempfile.mkstemp(dir=os.path.dirname(outputpath))
+
+        os.write(dest, source)
+        os.close(dest)
+        shutil.move(name, outputpath)
+
+
+def _get_module_info_from_callable(callable_):
+    return _get_module_info(callable_.__globals__["__name__"])
+
+
+def _get_module_info(filename):
+    return ModuleInfo._modules[filename]
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__init__.py b/backend/venv/lib/python3.12/site-packages/mako/testing/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ebba0c42
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/_config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/_config.cpython-312.pyc
new file mode 100644
index 00000000..3d769d61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/_config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/assertions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/assertions.cpython-312.pyc
new file mode 100644
index 00000000..699e0da2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/assertions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..983e8180
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/exclusions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/exclusions.cpython-312.pyc
new file mode 100644
index 00000000..a2af4aa0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/exclusions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/fixtures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/fixtures.cpython-312.pyc
new file mode 100644
index 00000000..4a8ee663
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/fixtures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/helpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/helpers.cpython-312.pyc
new file mode 100644
index 00000000..df8168c6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/mako/testing/__pycache__/helpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/_config.py b/backend/venv/lib/python3.12/site-packages/mako/testing/_config.py
new file mode 100644
index 00000000..4ee3d0a6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/_config.py
@@ -0,0 +1,128 @@
+import configparser
+import dataclasses
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Callable
+from typing import ClassVar
+from typing import Optional
+from typing import Union
+
+from .helpers import make_path
+
+
+class ConfigError(BaseException):
+    pass
+
+
+class MissingConfig(ConfigError):
+    pass
+
+
+class MissingConfigSection(ConfigError):
+    pass
+
+
+class MissingConfigItem(ConfigError):
+    pass
+
+
+class ConfigValueTypeError(ConfigError):
+    pass
+
+
+class _GetterDispatch:
+    def __init__(self, initialdata, default_getter: Callable):
+        self.default_getter = default_getter
+        self.data = initialdata
+
+    def get_fn_for_type(self, type_):
+        return self.data.get(type_, self.default_getter)
+
+    def get_typed_value(self, type_, name):
+        get_fn = self.get_fn_for_type(type_)
+        return get_fn(name)
+
+
+def _parse_cfg_file(filespec: Union[Path, str]):
+    cfg = configparser.ConfigParser()
+    try:
+        filepath = make_path(filespec, check_exists=True)
+    except FileNotFoundError as e:
+        raise MissingConfig(f"No config file found at {filespec}") from e
+    else:
+        with open(filepath, encoding="utf-8") as f:
+            cfg.read_file(f)
+        return cfg
+
+
+def _build_getter(cfg_obj, cfg_section, method, converter=None):
+    def caller(option, **kwargs):
+        try:
+            rv = getattr(cfg_obj, method)(cfg_section, option, **kwargs)
+        except configparser.NoSectionError as nse:
+            raise MissingConfigSection(
+                f"No config section named {cfg_section}"
+            ) from nse
+        except configparser.NoOptionError as noe:
+            raise MissingConfigItem(f"No config item for {option}") from noe
+        except ValueError as ve:
+            # ConfigParser.getboolean, .getint, .getfloat raise ValueError
+            # on bad types
+            raise ConfigValueTypeError(
+                f"Wrong value type for {option}"
+            ) from ve
+        else:
+            if converter:
+                try:
+                    rv = converter(rv)
+                except Exception as e:
+                    raise ConfigValueTypeError(
+                        f"Wrong value type for {option}"
+                    ) from e
+            return rv
+
+    return caller
+
+
+def _build_getter_dispatch(cfg_obj, cfg_section, converters=None):
+    converters = converters or {}
+
+    default_getter = _build_getter(cfg_obj, cfg_section, "get")
+
+    # support ConfigParser builtins
+    getters = {
+        int: _build_getter(cfg_obj, cfg_section, "getint"),
+        bool: _build_getter(cfg_obj, cfg_section, "getboolean"),
+        float: _build_getter(cfg_obj, cfg_section, "getfloat"),
+        str: default_getter,
+    }
+
+    # use ConfigParser.get and convert value
+    getters.update(
+        {
+            type_: _build_getter(
+                cfg_obj, cfg_section, "get", converter=converter_fn
+            )
+            for type_, converter_fn in converters.items()
+        }
+    )
+
+    return _GetterDispatch(getters, default_getter)
+
+
+@dataclass
+class ReadsCfg:
+    section_header: ClassVar[str]
+    converters: ClassVar[Optional[dict]] = None
+
+    @classmethod
+    def from_cfg_file(cls, filespec: Union[Path, str]):
+        cfg = _parse_cfg_file(filespec)
+        dispatch = _build_getter_dispatch(
+            cfg, cls.section_header, converters=cls.converters
+        )
+        kwargs = {
+            field.name: dispatch.get_typed_value(field.type, field.name)
+            for field in dataclasses.fields(cls)
+        }
+        return cls(**kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/assertions.py b/backend/venv/lib/python3.12/site-packages/mako/testing/assertions.py
new file mode 100644
index 00000000..22221cd2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/assertions.py
@@ -0,0 +1,166 @@
+import contextlib
+import re
+import sys
+
+
+def eq_(a, b, msg=None):
+    """Assert a == b, with repr messaging on failure."""
+    assert a == b, msg or "%r != %r" % (a, b)
+
+
+def ne_(a, b, msg=None):
+    """Assert a != b, with repr messaging on failure."""
+    assert a != b, msg or "%r == %r" % (a, b)
+
+
+def in_(a, b, msg=None):
+    """Assert a in b, with repr messaging on failure."""
+    assert a in b, msg or "%r not in %r" % (a, b)
+
+
+def not_in(a, b, msg=None):
+    """Assert a in not b, with repr messaging on failure."""
+    assert a not in b, msg or "%r is in %r" % (a, b)
+
+
+def _assert_proper_exception_context(exception):
+    """assert that any exception we're catching does not have a __context__
+    without a __cause__, and that __suppress_context__ is never set.
+
+    Python 3 will report nested as exceptions as "during the handling of
+    error X, error Y occurred". That's not what we want to do. We want
+    these exceptions in a cause chain.
+
+    """
+
+    if (
+        exception.__context__ is not exception.__cause__
+        and not exception.__suppress_context__
+    ):
+        assert False, (
+            "Exception %r was correctly raised but did not set a cause, "
+            "within context %r as its cause."
+            % (exception, exception.__context__)
+        )
+
+
+def _assert_proper_cause_cls(exception, cause_cls):
+    """assert that any exception we're catching does not have a __context__
+    without a __cause__, and that __suppress_context__ is never set.
+
+    Python 3 will report nested as exceptions as "during the handling of
+    error X, error Y occurred". That's not what we want to do. We want
+    these exceptions in a cause chain.
+
+    """
+    assert isinstance(exception.__cause__, cause_cls), (
+        "Exception %r was correctly raised but has cause %r, which does not "
+        "have the expected cause type %r."
+        % (exception, exception.__cause__, cause_cls)
+    )
+
+
+def assert_raises(except_cls, callable_, *args, **kw):
+    return _assert_raises(except_cls, callable_, args, kw)
+
+
+def assert_raises_with_proper_context(except_cls, callable_, *args, **kw):
+    return _assert_raises(except_cls, callable_, args, kw, check_context=True)
+
+
+def assert_raises_with_given_cause(
+    except_cls, cause_cls, callable_, *args, **kw
+):
+    return _assert_raises(except_cls, callable_, args, kw, cause_cls=cause_cls)
+
+
+def assert_raises_message(except_cls, msg, callable_, *args, **kwargs):
+    return _assert_raises(except_cls, callable_, args, kwargs, msg=msg)
+
+
+def assert_raises_message_with_proper_context(
+    except_cls, msg, callable_, *args, **kwargs
+):
+    return _assert_raises(
+        except_cls, callable_, args, kwargs, msg=msg, check_context=True
+    )
+
+
+def assert_raises_message_with_given_cause(
+    except_cls, msg, cause_cls, callable_, *args, **kwargs
+):
+    return _assert_raises(
+        except_cls, callable_, args, kwargs, msg=msg, cause_cls=cause_cls
+    )
+
+
+def _assert_raises(
+    except_cls,
+    callable_,
+    args,
+    kwargs,
+    msg=None,
+    check_context=False,
+    cause_cls=None,
+):
+    with _expect_raises(except_cls, msg, check_context, cause_cls) as ec:
+        callable_(*args, **kwargs)
+    return ec.error
+
+
+class _ErrorContainer:
+    error = None
+
+
+@contextlib.contextmanager
+def _expect_raises(except_cls, msg=None, check_context=False, cause_cls=None):
+    ec = _ErrorContainer()
+    if check_context:
+        are_we_already_in_a_traceback = sys.exc_info()[0]
+    try:
+        yield ec
+        success = False
+    except except_cls as err:
+        ec.error = err
+        success = True
+        if msg is not None:
+            # I'm often pdbing here, and "err" above isn't
+            # in scope, so assign the string explicitly
+            error_as_string = str(err)
+            assert re.search(msg, error_as_string, re.UNICODE), "%r !~ %s" % (
+                msg,
+                error_as_string,
+            )
+        if cause_cls is not None:
+            _assert_proper_cause_cls(err, cause_cls)
+        if check_context and not are_we_already_in_a_traceback:
+            _assert_proper_exception_context(err)
+        print(str(err).encode("utf-8"))
+
+    # it's generally a good idea to not carry traceback objects outside
+    # of the except: block, but in this case especially we seem to have
+    # hit some bug in either python 3.10.0b2 or greenlet or both which
+    # this seems to fix:
+    # https://github.com/python-greenlet/greenlet/issues/242
+    del ec
+
+    # assert outside the block so it works for AssertionError too !
+    assert success, "Callable did not raise an exception"
+
+
+def expect_raises(except_cls, check_context=False):
+    return _expect_raises(except_cls, check_context=check_context)
+
+
+def expect_raises_message(except_cls, msg, check_context=False):
+    return _expect_raises(except_cls, msg=msg, check_context=check_context)
+
+
+def expect_raises_with_proper_context(except_cls, check_context=True):
+    return _expect_raises(except_cls, check_context=check_context)
+
+
+def expect_raises_message_with_proper_context(
+    except_cls, msg, check_context=True
+):
+    return _expect_raises(except_cls, msg=msg, check_context=check_context)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/config.py b/backend/venv/lib/python3.12/site-packages/mako/testing/config.py
new file mode 100644
index 00000000..b77d0c08
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/config.py
@@ -0,0 +1,17 @@
+from dataclasses import dataclass
+from pathlib import Path
+
+from ._config import ReadsCfg
+from .helpers import make_path
+
+
+@dataclass
+class Config(ReadsCfg):
+    module_base: Path
+    template_base: Path
+
+    section_header = "mako_testing"
+    converters = {Path: make_path}
+
+
+config = Config.from_cfg_file("./setup.cfg")
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/exclusions.py b/backend/venv/lib/python3.12/site-packages/mako/testing/exclusions.py
new file mode 100644
index 00000000..37b2d14a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/exclusions.py
@@ -0,0 +1,80 @@
+import pytest
+
+from mako.ext.beaker_cache import has_beaker
+from mako.util import update_wrapper
+
+
+try:
+    import babel.messages.extract as babel
+except ImportError:
+    babel = None
+
+
+try:
+    import lingua
+except ImportError:
+    lingua = None
+
+
+try:
+    import dogpile.cache  # noqa
+except ImportError:
+    has_dogpile_cache = False
+else:
+    has_dogpile_cache = True
+
+
+requires_beaker = pytest.mark.skipif(
+    not has_beaker, reason="Beaker is required for these tests."
+)
+
+
+requires_babel = pytest.mark.skipif(
+    babel is None, reason="babel not installed: skipping babelplugin test"
+)
+
+
+requires_lingua = pytest.mark.skipif(
+    lingua is None, reason="lingua not installed: skipping linguaplugin test"
+)
+
+
+requires_dogpile_cache = pytest.mark.skipif(
+    not has_dogpile_cache,
+    reason="dogpile.cache is required to run these tests",
+)
+
+
+def _pygments_version():
+    try:
+        import pygments
+
+        version = pygments.__version__
+    except:
+        version = "0"
+    return version
+
+
+requires_pygments_14 = pytest.mark.skipif(
+    _pygments_version() < "1.4", reason="Requires pygments 1.4 or greater"
+)
+
+
+# def requires_pygments_14(fn):
+
+#     return skip_if(
+#         lambda: version < "1.4", "Requires pygments 1.4 or greater"
+#     )(fn)
+
+
+def requires_no_pygments_exceptions(fn):
+    def go(*arg, **kw):
+        from mako import exceptions
+
+        exceptions._install_fallback()
+        try:
+            return fn(*arg, **kw)
+        finally:
+            exceptions._install_highlighting()
+
+    return update_wrapper(go, fn)
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/fixtures.py b/backend/venv/lib/python3.12/site-packages/mako/testing/fixtures.py
new file mode 100644
index 00000000..01e99617
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/fixtures.py
@@ -0,0 +1,119 @@
+import os
+
+from mako.cache import CacheImpl
+from mako.cache import register_plugin
+from mako.template import Template
+from .assertions import eq_
+from .config import config
+
+
+class TemplateTest:
+    def _file_template(self, filename, **kw):
+        filepath = self._file_path(filename)
+        return Template(
+            uri=filename,
+            filename=filepath,
+            module_directory=config.module_base,
+            **kw,
+        )
+
+    def _file_path(self, filename):
+        name, ext = os.path.splitext(filename)
+        py3k_path = os.path.join(config.template_base, name + "_py3k" + ext)
+        if os.path.exists(py3k_path):
+            return py3k_path
+
+        return os.path.join(config.template_base, filename)
+
+    def _do_file_test(
+        self,
+        filename,
+        expected,
+        filters=None,
+        unicode_=True,
+        template_args=None,
+        **kw,
+    ):
+        t1 = self._file_template(filename, **kw)
+        self._do_test(
+            t1,
+            expected,
+            filters=filters,
+            unicode_=unicode_,
+            template_args=template_args,
+        )
+
+    def _do_memory_test(
+        self,
+        source,
+        expected,
+        filters=None,
+        unicode_=True,
+        template_args=None,
+        **kw,
+    ):
+        t1 = Template(text=source, **kw)
+        self._do_test(
+            t1,
+            expected,
+            filters=filters,
+            unicode_=unicode_,
+            template_args=template_args,
+        )
+
+    def _do_test(
+        self,
+        template,
+        expected,
+        filters=None,
+        template_args=None,
+        unicode_=True,
+    ):
+        if template_args is None:
+            template_args = {}
+        if unicode_:
+            output = template.render_unicode(**template_args)
+        else:
+            output = template.render(**template_args)
+
+        if filters:
+            output = filters(output)
+        eq_(output, expected)
+
+    def indicates_unbound_local_error(self, rendered_output, unbound_var):
+        var = f"&#39;{unbound_var}&#39;"
+        error_msgs = (
+            # < 3.11
+            f"local variable {var} referenced before assignment",
+            # >= 3.11
+            f"cannot access local variable {var} where it is not associated",
+        )
+        return any((msg in rendered_output) for msg in error_msgs)
+
+
+class PlainCacheImpl(CacheImpl):
+    """Simple memory cache impl so that tests which
+    use caching can run without beaker."""
+
+    def __init__(self, cache):
+        self.cache = cache
+        self.data = {}
+
+    def get_or_create(self, key, creation_function, **kw):
+        if key in self.data:
+            return self.data[key]
+        else:
+            self.data[key] = data = creation_function(**kw)
+            return data
+
+    def put(self, key, value, **kw):
+        self.data[key] = value
+
+    def get(self, key, **kw):
+        return self.data[key]
+
+    def invalidate(self, key, **kw):
+        del self.data[key]
+
+
+register_plugin("plain", __name__, "PlainCacheImpl")
diff --git a/backend/venv/lib/python3.12/site-packages/mako/testing/helpers.py b/backend/venv/lib/python3.12/site-packages/mako/testing/helpers.py
new file mode 100644
index 00000000..5ae9d38d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/testing/helpers.py
@@ -0,0 +1,71 @@
+import contextlib
+import pathlib
+from pathlib import Path
+import re
+import time
+from typing import Union
+from unittest import mock
+
+
+def flatten_result(result):
+    return re.sub(r"[\s\r\n]+", " ", result).strip()
+
+
+def result_lines(result):
+    return [
+        x.strip()
+        for x in re.split(r"\r?\n", re.sub(r" +", " ", result))
+        if x.strip() != ""
+    ]
+
+
+def result_raw_lines(result):
+    return [x for x in re.split(r"\r?\n", result) if x.strip() != ""]
+
+
+def make_path(
+    filespec: Union[Path, str],
+    make_absolute: bool = True,
+    check_exists: bool = False,
+) -> Path:
+    path = Path(filespec)
+    if make_absolute:
+        path = path.resolve(strict=check_exists)
+    if check_exists and (not path.exists()):
+        raise FileNotFoundError(f"No file or directory at {filespec}")
+    return path
+
+
+def _unlink_path(path, missing_ok=False):
+    # Replicate 3.8+ functionality in 3.7
+    cm = contextlib.nullcontext()
+    if missing_ok:
+        cm = contextlib.suppress(FileNotFoundError)
+
+    with cm:
+        path.unlink()
+
+
+def replace_file_with_dir(pathspec):
+    path = pathlib.Path(pathspec)
+    _unlink_path(path, missing_ok=True)
+    path.mkdir(exist_ok=True)
+    return path
+
+
+def file_with_template_code(filespec):
+    with open(filespec, "w") as f:
+        f.write(
+            """
+i am an artificial template just for you
+"""
+        )
+    return filespec
+
+
+@contextlib.contextmanager
+def rewind_compile_time(hours=1):
+    rewound = time.time() - (hours * 3_600)
+    with mock.patch("mako.codegen.time") as codegen_time:
+        codegen_time.time.return_value = rewound
+        yield
diff --git a/backend/venv/lib/python3.12/site-packages/mako/util.py b/backend/venv/lib/python3.12/site-packages/mako/util.py
new file mode 100644
index 00000000..470713a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/mako/util.py
@@ -0,0 +1,388 @@
+# mako/util.py
+# Copyright 2006-2025 the Mako authors and contributors <see AUTHORS file>
+#
+# This module is part of Mako and is released under
+# the MIT License: http://www.opensource.org/licenses/mit-license.php
+from ast import parse
+import codecs
+import collections
+import operator
+import os
+import re
+import timeit
+
+from .compat import importlib_metadata_get
+
+
+def update_wrapper(decorated, fn):
+    decorated.__wrapped__ = fn
+    decorated.__name__ = fn.__name__
+    return decorated
+
+
+class PluginLoader:
+    def __init__(self, group):
+        self.group = group
+        self.impls = {}
+
+    def load(self, name):
+        if name in self.impls:
+            return self.impls[name]()
+
+        for impl in importlib_metadata_get(self.group):
+            if impl.name == name:
+                self.impls[name] = impl.load
+                return impl.load()
+
+        from mako import exceptions
+
+        raise exceptions.RuntimeException(
+            "Can't load plugin %s %s" % (self.group, name)
+        )
+
+    def register(self, name, modulepath, objname):
+        def load():
+            mod = __import__(modulepath)
+            for token in modulepath.split(".")[1:]:
+                mod = getattr(mod, token)
+            return getattr(mod, objname)
+
+        self.impls[name] = load
+
+
+def verify_directory(dir_):
+    """create and/or verify a filesystem directory."""
+
+    tries = 0
+
+    while not os.path.exists(dir_):
+        try:
+            tries += 1
+            os.makedirs(dir_, 0o755)
+        except:
+            if tries > 5:
+                raise
+
+
+def to_list(x, default=None):
+    if x is None:
+        return default
+    if not isinstance(x, (list, tuple)):
+        return [x]
+    else:
+        return x
+
+
+class memoized_property:
+
+    """A read-only @property that is only evaluated once."""
+
+    def __init__(self, fget, doc=None):
+        self.fget = fget
+        self.__doc__ = doc or fget.__doc__
+        self.__name__ = fget.__name__
+
+    def __get__(self, obj, cls):
+        if obj is None:
+            return self
+        obj.__dict__[self.__name__] = result = self.fget(obj)
+        return result
+
+
+class memoized_instancemethod:
+
+    """Decorate a method memoize its return value.
+
+    Best applied to no-arg methods: memoization is not sensitive to
+    argument values, and will always return the same value even when
+    called with different arguments.
+
+    """
+
+    def __init__(self, fget, doc=None):
+        self.fget = fget
+        self.__doc__ = doc or fget.__doc__
+        self.__name__ = fget.__name__
+
+    def __get__(self, obj, cls):
+        if obj is None:
+            return self
+
+        def oneshot(*args, **kw):
+            result = self.fget(obj, *args, **kw)
+
+            def memo(*a, **kw):
+                return result
+
+            memo.__name__ = self.__name__
+            memo.__doc__ = self.__doc__
+            obj.__dict__[self.__name__] = memo
+            return result
+
+        oneshot.__name__ = self.__name__
+        oneshot.__doc__ = self.__doc__
+        return oneshot
+
+
+class SetLikeDict(dict):
+
+    """a dictionary that has some setlike methods on it"""
+
+    def union(self, other):
+        """produce a 'union' of this dict and another (at the key level).
+
+        values in the second dict take precedence over that of the first"""
+        x = SetLikeDict(**self)
+        x.update(other)
+        return x
+
+
+class FastEncodingBuffer:
+
+    """a very rudimentary buffer that is faster than StringIO,
+    and supports unicode data."""
+
+    def __init__(self, encoding=None, errors="strict"):
+        self.data = collections.deque()
+        self.encoding = encoding
+        self.delim = ""
+        self.errors = errors
+        self.write = self.data.append
+
+    def truncate(self):
+        self.data = collections.deque()
+        self.write = self.data.append
+
+    def getvalue(self):
+        if self.encoding:
+            return self.delim.join(self.data).encode(
+                self.encoding, self.errors
+            )
+        else:
+            return self.delim.join(self.data)
+
+
+class LRUCache(dict):
+
+    """A dictionary-like object that stores a limited number of items,
+    discarding lesser used items periodically.
+
+    this is a rewrite of LRUCache from Myghty to use a periodic timestamp-based
+    paradigm so that synchronization is not really needed.  the size management
+    is inexact.
+    """
+
+    class _Item:
+        def __init__(self, key, value):
+            self.key = key
+            self.value = value
+            self.timestamp = timeit.default_timer()
+
+        def __repr__(self):
+            return repr(self.value)
+
+    def __init__(self, capacity, threshold=0.5):
+        self.capacity = capacity
+        self.threshold = threshold
+
+    def __getitem__(self, key):
+        item = dict.__getitem__(self, key)
+        item.timestamp = timeit.default_timer()
+        return item.value
+
+    def values(self):
+        return [i.value for i in dict.values(self)]
+
+    def setdefault(self, key, value):
+        if key in self:
+            return self[key]
+        self[key] = value
+        return value
+
+    def __setitem__(self, key, value):
+        item = dict.get(self, key)
+        if item is None:
+            item = self._Item(key, value)
+            dict.__setitem__(self, key, item)
+        else:
+            item.value = value
+        self._manage_size()
+
+    def _manage_size(self):
+        while len(self) > self.capacity + self.capacity * self.threshold:
+            bytime = sorted(
+                dict.values(self),
+                key=operator.attrgetter("timestamp"),
+                reverse=True,
+            )
+            for item in bytime[self.capacity :]:
+                try:
+                    del self[item.key]
+                except KeyError:
+                    # if we couldn't find a key, most likely some other thread
+                    # broke in on us. loop around and try again
+                    break
+
+
+# Regexp to match python magic encoding line
+_PYTHON_MAGIC_COMMENT_re = re.compile(
+    r"[ \t\f]* \# .* coding[=:][ \t]*([-\w.]+)", re.VERBOSE
+)
+
+
+def parse_encoding(fp):
+    """Deduce the encoding of a Python source file (binary mode) from magic
+    comment.
+
+    It does this in the same way as the `Python interpreter`__
+
+    .. __: http://docs.python.org/ref/encodings.html
+
+    The ``fp`` argument should be a seekable file object in binary mode.
+    """
+    pos = fp.tell()
+    fp.seek(0)
+    try:
+        line1 = fp.readline()
+        has_bom = line1.startswith(codecs.BOM_UTF8)
+        if has_bom:
+            line1 = line1[len(codecs.BOM_UTF8) :]
+
+        m = _PYTHON_MAGIC_COMMENT_re.match(line1.decode("ascii", "ignore"))
+        if not m:
+            try:
+                parse(line1.decode("ascii", "ignore"))
+            except (ImportError, SyntaxError):
+                # Either it's a real syntax error, in which case the source
+                # is not valid python source, or line2 is a continuation of
+                # line1, in which case we don't want to scan line2 for a magic
+                # comment.
+                pass
+            else:
+                line2 = fp.readline()
+                m = _PYTHON_MAGIC_COMMENT_re.match(
+                    line2.decode("ascii", "ignore")
+                )
+
+        if has_bom:
+            if m:
+                raise SyntaxError(
+                    "python refuses to compile code with both a UTF8"
+                    " byte-order-mark and a magic encoding comment"
+                )
+            return "utf_8"
+        elif m:
+            return m.group(1)
+        else:
+            return None
+    finally:
+        fp.seek(pos)
+
+
+def sorted_dict_repr(d):
+    """repr() a dictionary with the keys in order.
+
+    Used by the lexer unit test to compare parse trees based on strings.
+
+    """
+    keys = list(d.keys())
+    keys.sort()
+    return "{" + ", ".join("%r: %r" % (k, d[k]) for k in keys) + "}"
+
+
+def restore__ast(_ast):
+    """Attempt to restore the required classes to the _ast module if it
+    appears to be missing them
+    """
+    if hasattr(_ast, "AST"):
+        return
+    _ast.PyCF_ONLY_AST = 2 << 9
+    m = compile(
+        """\
+def foo(): pass
+class Bar: pass
+if False: pass
+baz = 'mako'
+1 + 2 - 3 * 4 / 5
+6 // 7 % 8 << 9 >> 10
+11 & 12 ^ 13 | 14
+15 and 16 or 17
+-baz + (not +18) - ~17
+baz and 'foo' or 'bar'
+(mako is baz == baz) is not baz != mako
+mako > baz < mako >= baz <= mako
+mako in baz not in mako""",
+        "<unknown>",
+        "exec",
+        _ast.PyCF_ONLY_AST,
+    )
+    _ast.Module = type(m)
+
+    for cls in _ast.Module.__mro__:
+        if cls.__name__ == "mod":
+            _ast.mod = cls
+        elif cls.__name__ == "AST":
+            _ast.AST = cls
+
+    _ast.FunctionDef = type(m.body[0])
+    _ast.ClassDef = type(m.body[1])
+    _ast.If = type(m.body[2])
+
+    _ast.Name = type(m.body[3].targets[0])
+    _ast.Store = type(m.body[3].targets[0].ctx)
+    _ast.Str = type(m.body[3].value)
+
+    _ast.Sub = type(m.body[4].value.op)
+    _ast.Add = type(m.body[4].value.left.op)
+    _ast.Div = type(m.body[4].value.right.op)
+    _ast.Mult = type(m.body[4].value.right.left.op)
+
+    _ast.RShift = type(m.body[5].value.op)
+    _ast.LShift = type(m.body[5].value.left.op)
+    _ast.Mod = type(m.body[5].value.left.left.op)
+    _ast.FloorDiv = type(m.body[5].value.left.left.left.op)
+
+    _ast.BitOr = type(m.body[6].value.op)
+    _ast.BitXor = type(m.body[6].value.left.op)
+    _ast.BitAnd = type(m.body[6].value.left.left.op)
+
+    _ast.Or = type(m.body[7].value.op)
+    _ast.And = type(m.body[7].value.values[0].op)
+
+    _ast.Invert = type(m.body[8].value.right.op)
+    _ast.Not = type(m.body[8].value.left.right.op)
+    _ast.UAdd = type(m.body[8].value.left.right.operand.op)
+    _ast.USub = type(m.body[8].value.left.left.op)
+
+    _ast.Or = type(m.body[9].value.op)
+    _ast.And = type(m.body[9].value.values[0].op)
+
+    _ast.IsNot = type(m.body[10].value.ops[0])
+    _ast.NotEq = type(m.body[10].value.ops[1])
+    _ast.Is = type(m.body[10].value.left.ops[0])
+    _ast.Eq = type(m.body[10].value.left.ops[1])
+
+    _ast.Gt = type(m.body[11].value.ops[0])
+    _ast.Lt = type(m.body[11].value.ops[1])
+    _ast.GtE = type(m.body[11].value.ops[2])
+    _ast.LtE = type(m.body[11].value.ops[3])
+
+    _ast.In = type(m.body[12].value.ops[0])
+    _ast.NotIn = type(m.body[12].value.ops[1])
+
+
+def read_file(path, mode="rb"):
+    with open(path, mode) as fp:
+        return fp.read()
+
+
+def read_python_file(path):
+    fp = open(path, "rb")
+    try:
+        encoding = parse_encoding(fp)
+        data = fp.read()
+        if encoding:
+            data = data.decode(encoding)
+        return data
+    finally:
+        fp.close()
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/__init__.py b/backend/venv/lib/python3.12/site-packages/markupsafe/__init__.py
new file mode 100644
index 00000000..fee8dc7a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/markupsafe/__init__.py
@@ -0,0 +1,395 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import string
+import typing as t
+
+try:
+    from ._speedups import _escape_inner
+except ImportError:
+    from ._native import _escape_inner
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+
+class _HasHTML(t.Protocol):
+    def __html__(self, /) -> str: ...
+
+
+class _TPEscape(t.Protocol):
+    def __call__(self, s: t.Any, /) -> Markup: ...
+
+
+def escape(s: t.Any, /) -> Markup:
+    """Replace the characters ``&``, ``<``, ``>``, ``'``, and ``"`` in
+    the string with HTML-safe sequences. Use this if you need to display
+    text that might contain such characters in HTML.
+
+    If the object has an ``__html__`` method, it is called and the
+    return value is assumed to already be safe for HTML.
+
+    :param s: An object to be converted to a string and escaped.
+    :return: A :class:`Markup` string with the escaped text.
+    """
+    # If the object is already a plain string, skip __html__ check and string
+    # conversion. This is the most common use case.
+    # Use type(s) instead of s.__class__ because a proxy object may be reporting
+    # the __class__ of the proxied value.
+    if type(s) is str:
+        return Markup(_escape_inner(s))
+
+    if hasattr(s, "__html__"):
+        return Markup(s.__html__())
+
+    return Markup(_escape_inner(str(s)))
+
+
+def escape_silent(s: t.Any | None, /) -> Markup:
+    """Like :func:`escape` but treats ``None`` as the empty string.
+    Useful with optional values, as otherwise you get the string
+    ``'None'`` when the value is ``None``.
+
+    >>> escape(None)
+    Markup('None')
+    >>> escape_silent(None)
+    Markup('')
+    """
+    if s is None:
+        return Markup()
+
+    return escape(s)
+
+
+def soft_str(s: t.Any, /) -> str:
+    """Convert an object to a string if it isn't already. This preserves
+    a :class:`Markup` string rather than converting it back to a basic
+    string, so it will still be marked as safe and won't be escaped
+    again.
+
+    >>> value = escape("<User 1>")
+    >>> value
+    Markup('&lt;User 1&gt;')
+    >>> escape(str(value))
+    Markup('&amp;lt;User 1&amp;gt;')
+    >>> escape(soft_str(value))
+    Markup('&lt;User 1&gt;')
+    """
+    if not isinstance(s, str):
+        return str(s)
+
+    return s
+
+
+class Markup(str):
+    """A string that is ready to be safely inserted into an HTML or XML
+    document, either because it was escaped or because it was marked
+    safe.
+
+    Passing an object to the constructor converts it to text and wraps
+    it to mark it safe without escaping. To escape the text, use the
+    :meth:`escape` class method instead.
+
+    >>> Markup("Hello, <em>World</em>!")
+    Markup('Hello, <em>World</em>!')
+    >>> Markup(42)
+    Markup('42')
+    >>> Markup.escape("Hello, <em>World</em>!")
+    Markup('Hello &lt;em&gt;World&lt;/em&gt;!')
+
+    This implements the ``__html__()`` interface that some frameworks
+    use. Passing an object that implements ``__html__()`` will wrap the
+    output of that method, marking it safe.
+
+    >>> class Foo:
+    ...     def __html__(self):
+    ...         return '<a href="/foo">foo</a>'
+    ...
+    >>> Markup(Foo())
+    Markup('<a href="/foo">foo</a>')
+
+    This is a subclass of :class:`str`. It has the same methods, but
+    escapes their arguments and returns a ``Markup`` instance.
+
+    >>> Markup("<em>%s</em>") % ("foo & bar",)
+    Markup('<em>foo &amp; bar</em>')
+    >>> Markup("<em>Hello</em> ") + "<foo>"
+    Markup('<em>Hello</em> &lt;foo&gt;')
+    """
+
+    __slots__ = ()
+
+    def __new__(
+        cls, object: t.Any = "", encoding: str | None = None, errors: str = "strict"
+    ) -> te.Self:
+        if hasattr(object, "__html__"):
+            object = object.__html__()
+
+        if encoding is None:
+            return super().__new__(cls, object)
+
+        return super().__new__(cls, object, encoding, errors)
+
+    def __html__(self, /) -> te.Self:
+        return self
+
+    def __add__(self, value: str | _HasHTML, /) -> te.Self:
+        if isinstance(value, str) or hasattr(value, "__html__"):
+            return self.__class__(super().__add__(self.escape(value)))
+
+        return NotImplemented
+
+    def __radd__(self, value: str | _HasHTML, /) -> te.Self:
+        if isinstance(value, str) or hasattr(value, "__html__"):
+            return self.escape(value).__add__(self)
+
+        return NotImplemented
+
+    def __mul__(self, value: t.SupportsIndex, /) -> te.Self:
+        return self.__class__(super().__mul__(value))
+
+    def __rmul__(self, value: t.SupportsIndex, /) -> te.Self:
+        return self.__class__(super().__mul__(value))
+
+    def __mod__(self, value: t.Any, /) -> te.Self:
+        if isinstance(value, tuple):
+            # a tuple of arguments, each wrapped
+            value = tuple(_MarkupEscapeHelper(x, self.escape) for x in value)
+        elif hasattr(type(value), "__getitem__") and not isinstance(value, str):
+            # a mapping of arguments, wrapped
+            value = _MarkupEscapeHelper(value, self.escape)
+        else:
+            # a single argument, wrapped with the helper and a tuple
+            value = (_MarkupEscapeHelper(value, self.escape),)
+
+        return self.__class__(super().__mod__(value))
+
+    def __repr__(self, /) -> str:
+        return f"{self.__class__.__name__}({super().__repr__()})"
+
+    def join(self, iterable: cabc.Iterable[str | _HasHTML], /) -> te.Self:
+        return self.__class__(super().join(map(self.escape, iterable)))
+
+    def split(  # type: ignore[override]
+        self, /, sep: str | None = None, maxsplit: t.SupportsIndex = -1
+    ) -> list[te.Self]:
+        return [self.__class__(v) for v in super().split(sep, maxsplit)]
+
+    def rsplit(  # type: ignore[override]
+        self, /, sep: str | None = None, maxsplit: t.SupportsIndex = -1
+    ) -> list[te.Self]:
+        return [self.__class__(v) for v in super().rsplit(sep, maxsplit)]
+
+    def splitlines(  # type: ignore[override]
+        self, /, keepends: bool = False
+    ) -> list[te.Self]:
+        return [self.__class__(v) for v in super().splitlines(keepends)]
+
+    def unescape(self, /) -> str:
+        """Convert escaped markup back into a text string. This replaces
+        HTML entities with the characters they represent.
+
+        >>> Markup("Main &raquo; <em>About</em>").unescape()
+        'Main » <em>About</em>'
+        """
+        from html import unescape
+
+        return unescape(str(self))
+
+    def striptags(self, /) -> str:
+        """:meth:`unescape` the markup, remove tags, and normalize
+        whitespace to single spaces.
+
+        >>> Markup("Main &raquo;\t<em>About</em>").striptags()
+        'Main » About'
+        """
+        value = str(self)
+
+        # Look for comments then tags separately. Otherwise, a comment that
+        # contains a tag would end early, leaving some of the comment behind.
+
+        # keep finding comment start marks
+        while (start := value.find("<!--")) != -1:
+            # find a comment end mark beyond the start, otherwise stop
+            if (end := value.find("-->", start)) == -1:
+                break
+
+            value = f"{value[:start]}{value[end + 3:]}"
+
+        # remove tags using the same method
+        while (start := value.find("<")) != -1:
+            if (end := value.find(">", start)) == -1:
+                break
+
+            value = f"{value[:start]}{value[end + 1:]}"
+
+        # collapse spaces
+        value = " ".join(value.split())
+        return self.__class__(value).unescape()
+
+    @classmethod
+    def escape(cls, s: t.Any, /) -> te.Self:
+        """Escape a string. Calls :func:`escape` and ensures that for
+        subclasses the correct type is returned.
+        """
+        rv = escape(s)
+
+        if rv.__class__ is not cls:
+            return cls(rv)
+
+        return rv  # type: ignore[return-value]
+
+    def __getitem__(self, key: t.SupportsIndex | slice, /) -> te.Self:
+        return self.__class__(super().__getitem__(key))
+
+    def capitalize(self, /) -> te.Self:
+        return self.__class__(super().capitalize())
+
+    def title(self, /) -> te.Self:
+        return self.__class__(super().title())
+
+    def lower(self, /) -> te.Self:
+        return self.__class__(super().lower())
+
+    def upper(self, /) -> te.Self:
+        return self.__class__(super().upper())
+
+    def replace(self, old: str, new: str, count: t.SupportsIndex = -1, /) -> te.Self:
+        return self.__class__(super().replace(old, self.escape(new), count))
+
+    def ljust(self, width: t.SupportsIndex, fillchar: str = " ", /) -> te.Self:
+        return self.__class__(super().ljust(width, self.escape(fillchar)))
+
+    def rjust(self, width: t.SupportsIndex, fillchar: str = " ", /) -> te.Self:
+        return self.__class__(super().rjust(width, self.escape(fillchar)))
+
+    def lstrip(self, chars: str | None = None, /) -> te.Self:
+        return self.__class__(super().lstrip(chars))
+
+    def rstrip(self, chars: str | None = None, /) -> te.Self:
+        return self.__class__(super().rstrip(chars))
+
+    def center(self, width: t.SupportsIndex, fillchar: str = " ", /) -> te.Self:
+        return self.__class__(super().center(width, self.escape(fillchar)))
+
+    def strip(self, chars: str | None = None, /) -> te.Self:
+        return self.__class__(super().strip(chars))
+
+    def translate(
+        self,
+        table: cabc.Mapping[int, str | int | None],  # type: ignore[override]
+        /,
+    ) -> str:
+        return self.__class__(super().translate(table))
+
+    def expandtabs(self, /, tabsize: t.SupportsIndex = 8) -> te.Self:
+        return self.__class__(super().expandtabs(tabsize))
+
+    def swapcase(self, /) -> te.Self:
+        return self.__class__(super().swapcase())
+
+    def zfill(self, width: t.SupportsIndex, /) -> te.Self:
+        return self.__class__(super().zfill(width))
+
+    def casefold(self, /) -> te.Self:
+        return self.__class__(super().casefold())
+
+    def removeprefix(self, prefix: str, /) -> te.Self:
+        return self.__class__(super().removeprefix(prefix))
+
+    def removesuffix(self, suffix: str) -> te.Self:
+        return self.__class__(super().removesuffix(suffix))
+
+    def partition(self, sep: str, /) -> tuple[te.Self, te.Self, te.Self]:
+        left, sep, right = super().partition(sep)
+        cls = self.__class__
+        return cls(left), cls(sep), cls(right)
+
+    def rpartition(self, sep: str, /) -> tuple[te.Self, te.Self, te.Self]:
+        left, sep, right = super().rpartition(sep)
+        cls = self.__class__
+        return cls(left), cls(sep), cls(right)
+
+    def format(self, *args: t.Any, **kwargs: t.Any) -> te.Self:
+        formatter = EscapeFormatter(self.escape)
+        return self.__class__(formatter.vformat(self, args, kwargs))
+
+    def format_map(
+        self,
+        mapping: cabc.Mapping[str, t.Any],  # type: ignore[override]
+        /,
+    ) -> te.Self:
+        formatter = EscapeFormatter(self.escape)
+        return self.__class__(formatter.vformat(self, (), mapping))
+
+    def __html_format__(self, format_spec: str, /) -> te.Self:
+        if format_spec:
+            raise ValueError("Unsupported format specification for Markup.")
+
+        return self
+
+
+class EscapeFormatter(string.Formatter):
+    __slots__ = ("escape",)
+
+    def __init__(self, escape: _TPEscape) -> None:
+        self.escape: _TPEscape = escape
+        super().__init__()
+
+    def format_field(self, value: t.Any, format_spec: str) -> str:
+        if hasattr(value, "__html_format__"):
+            rv = value.__html_format__(format_spec)
+        elif hasattr(value, "__html__"):
+            if format_spec:
+                raise ValueError(
+                    f"Format specifier {format_spec} given, but {type(value)} does not"
+                    " define __html_format__. A class that defines __html__ must define"
+                    " __html_format__ to work with format specifiers."
+                )
+            rv = value.__html__()
+        else:
+            # We need to make sure the format spec is str here as
+            # otherwise the wrong callback methods are invoked.
+            rv = super().format_field(value, str(format_spec))
+        return str(self.escape(rv))
+
+
+class _MarkupEscapeHelper:
+    """Helper for :meth:`Markup.__mod__`."""
+
+    __slots__ = ("obj", "escape")
+
+    def __init__(self, obj: t.Any, escape: _TPEscape) -> None:
+        self.obj: t.Any = obj
+        self.escape: _TPEscape = escape
+
+    def __getitem__(self, key: t.Any, /) -> te.Self:
+        return self.__class__(self.obj[key], self.escape)
+
+    def __str__(self, /) -> str:
+        return str(self.escape(self.obj))
+
+    def __repr__(self, /) -> str:
+        return str(self.escape(repr(self.obj)))
+
+    def __int__(self, /) -> int:
+        return int(self.obj)
+
+    def __float__(self, /) -> float:
+        return float(self.obj)
+
+
+def __getattr__(name: str) -> t.Any:
+    if name == "__version__":
+        import importlib.metadata
+        import warnings
+
+        warnings.warn(
+            "The '__version__' attribute is deprecated and will be removed in"
+            " MarkupSafe 3.1. Use feature detection, or"
+            ' `importlib.metadata.version("markupsafe")`, instead.',
+            stacklevel=2,
+        )
+        return importlib.metadata.version("markupsafe")
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2da35c8e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/_native.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/_native.cpython-312.pyc
new file mode 100644
index 00000000..4b2bcefa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/markupsafe/__pycache__/_native.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/_native.py b/backend/venv/lib/python3.12/site-packages/markupsafe/_native.py
new file mode 100644
index 00000000..088b3bca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/markupsafe/_native.py
@@ -0,0 +1,8 @@
+def _escape_inner(s: str, /) -> str:
+    return (
+        s.replace("&", "&amp;")
+        .replace(">", "&gt;")
+        .replace("<", "&lt;")
+        .replace("'", "&#39;")
+        .replace('"', "&#34;")
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.c b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.c
new file mode 100644
index 00000000..09dd57ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.c
@@ -0,0 +1,204 @@
+#include <Python.h>
+
+#define GET_DELTA(inp, inp_end, delta) \
+	while (inp < inp_end) { \
+		switch (*inp++) { \
+		case '"': \
+		case '\'': \
+		case '&': \
+			delta += 4; \
+			break; \
+		case '<': \
+		case '>': \
+			delta += 3; \
+			break; \
+		} \
+	}
+
+#define DO_ESCAPE(inp, inp_end, outp) \
+	{ \
+		Py_ssize_t ncopy = 0; \
+		while (inp < inp_end) { \
+			switch (*inp) { \
+			case '"': \
+				memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+				outp += ncopy; ncopy = 0; \
+				*outp++ = '&'; \
+				*outp++ = '#'; \
+				*outp++ = '3'; \
+				*outp++ = '4'; \
+				*outp++ = ';'; \
+				break; \
+			case '\'': \
+				memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+				outp += ncopy; ncopy = 0; \
+				*outp++ = '&'; \
+				*outp++ = '#'; \
+				*outp++ = '3'; \
+				*outp++ = '9'; \
+				*outp++ = ';'; \
+				break; \
+			case '&': \
+				memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+				outp += ncopy; ncopy = 0; \
+				*outp++ = '&'; \
+				*outp++ = 'a'; \
+				*outp++ = 'm'; \
+				*outp++ = 'p'; \
+				*outp++ = ';'; \
+				break; \
+			case '<': \
+				memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+				outp += ncopy; ncopy = 0; \
+				*outp++ = '&'; \
+				*outp++ = 'l'; \
+				*outp++ = 't'; \
+				*outp++ = ';'; \
+				break; \
+			case '>': \
+				memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+				outp += ncopy; ncopy = 0; \
+				*outp++ = '&'; \
+				*outp++ = 'g'; \
+				*outp++ = 't'; \
+				*outp++ = ';'; \
+				break; \
+			default: \
+				ncopy++; \
+			} \
+			inp++; \
+		} \
+		memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
+	}
+
+static PyObject*
+escape_unicode_kind1(PyUnicodeObject *in)
+{
+	Py_UCS1 *inp = PyUnicode_1BYTE_DATA(in);
+	Py_UCS1 *inp_end = inp + PyUnicode_GET_LENGTH(in);
+	Py_UCS1 *outp;
+	PyObject *out;
+	Py_ssize_t delta = 0;
+
+	GET_DELTA(inp, inp_end, delta);
+	if (!delta) {
+		Py_INCREF(in);
+		return (PyObject*)in;
+	}
+
+	out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta,
+						PyUnicode_IS_ASCII(in) ? 127 : 255);
+	if (!out)
+		return NULL;
+
+	inp = PyUnicode_1BYTE_DATA(in);
+	outp = PyUnicode_1BYTE_DATA(out);
+	DO_ESCAPE(inp, inp_end, outp);
+	return out;
+}
+
+static PyObject*
+escape_unicode_kind2(PyUnicodeObject *in)
+{
+	Py_UCS2 *inp = PyUnicode_2BYTE_DATA(in);
+	Py_UCS2 *inp_end = inp + PyUnicode_GET_LENGTH(in);
+	Py_UCS2 *outp;
+	PyObject *out;
+	Py_ssize_t delta = 0;
+
+	GET_DELTA(inp, inp_end, delta);
+	if (!delta) {
+		Py_INCREF(in);
+		return (PyObject*)in;
+	}
+
+	out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta, 65535);
+	if (!out)
+		return NULL;
+
+	inp = PyUnicode_2BYTE_DATA(in);
+	outp = PyUnicode_2BYTE_DATA(out);
+	DO_ESCAPE(inp, inp_end, outp);
+	return out;
+}
+
+
+static PyObject*
+escape_unicode_kind4(PyUnicodeObject *in)
+{
+	Py_UCS4 *inp = PyUnicode_4BYTE_DATA(in);
+	Py_UCS4 *inp_end = inp + PyUnicode_GET_LENGTH(in);
+	Py_UCS4 *outp;
+	PyObject *out;
+	Py_ssize_t delta = 0;
+
+	GET_DELTA(inp, inp_end, delta);
+	if (!delta) {
+		Py_INCREF(in);
+		return (PyObject*)in;
+	}
+
+	out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta, 1114111);
+	if (!out)
+		return NULL;
+
+	inp = PyUnicode_4BYTE_DATA(in);
+	outp = PyUnicode_4BYTE_DATA(out);
+	DO_ESCAPE(inp, inp_end, outp);
+	return out;
+}
+
+static PyObject*
+escape_unicode(PyObject *self, PyObject *s)
+{
+	if (!PyUnicode_Check(s))
+		return NULL;
+
+    // This check is no longer needed in Python 3.12.
+	if (PyUnicode_READY(s))
+		return NULL;
+
+	switch (PyUnicode_KIND(s)) {
+	case PyUnicode_1BYTE_KIND:
+		return escape_unicode_kind1((PyUnicodeObject*) s);
+	case PyUnicode_2BYTE_KIND:
+		return escape_unicode_kind2((PyUnicodeObject*) s);
+	case PyUnicode_4BYTE_KIND:
+		return escape_unicode_kind4((PyUnicodeObject*) s);
+	}
+	assert(0);  /* shouldn't happen */
+	return NULL;
+}
+
+static PyMethodDef module_methods[] = {
+	{"_escape_inner", (PyCFunction)escape_unicode, METH_O, NULL},
+	{NULL, NULL, 0, NULL}  /* Sentinel */
+};
+
+static struct PyModuleDef module_definition = {
+	PyModuleDef_HEAD_INIT,
+	"markupsafe._speedups",
+	NULL,
+	-1,
+	module_methods,
+	NULL,
+	NULL,
+	NULL,
+	NULL
+};
+
+PyMODINIT_FUNC
+PyInit__speedups(void)
+{
+	PyObject *m = PyModule_Create(&module_definition);
+
+	if (m == NULL) {
+		return NULL;
+	}
+
+	#ifdef Py_GIL_DISABLED
+	PyUnstable_Module_SetGIL(m, Py_MOD_GIL_NOT_USED);
+	#endif
+
+	return m;
+}
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.cpython-312-darwin.so
new file mode 100755
index 00000000..cccdfaab
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.pyi b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.pyi
new file mode 100644
index 00000000..8c888585
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/markupsafe/_speedups.pyi
@@ -0,0 +1 @@
+def _escape_inner(s: str, /) -> str: ...
diff --git a/backend/venv/lib/python3.12/site-packages/markupsafe/py.typed b/backend/venv/lib/python3.12/site-packages/markupsafe/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE
new file mode 100644
index 00000000..6f62d44e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE
@@ -0,0 +1,3 @@
+This software is made available under the terms of *either* of the licenses
+found in LICENSE.APACHE or LICENSE.BSD. Contributions to this software is made
+under the terms of *both* these licenses.
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.APACHE b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.APACHE
new file mode 100644
index 00000000..f433b1a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.APACHE
@@ -0,0 +1,177 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.BSD b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.BSD
new file mode 100644
index 00000000..42ce7b75
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/LICENSE.BSD
@@ -0,0 +1,23 @@
+Copyright (c) Donald Stufft and individual contributors.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+    1. Redistributions of source code must retain the above copyright notice,
+       this list of conditions and the following disclaimer.
+
+    2. Redistributions in binary form must reproduce the above copyright
+       notice, this list of conditions and the following disclaimer in the
+       documentation and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/METADATA
new file mode 100644
index 00000000..1479c869
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/METADATA
@@ -0,0 +1,102 @@
+Metadata-Version: 2.3
+Name: packaging
+Version: 24.2
+Summary: Core utilities for Python packages
+Author-email: Donald Stufft <donald@stufft.io>
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Typing :: Typed
+Project-URL: Documentation, https://packaging.pypa.io/
+Project-URL: Source, https://github.com/pypa/packaging
+
+packaging
+=========
+
+.. start-intro
+
+Reusable core utilities for various Python Packaging
+`interoperability specifications <https://packaging.python.org/specifications/>`_.
+
+This library provides utilities that implement the interoperability
+specifications which have clearly one correct behaviour (eg: :pep:`440`)
+or benefit greatly from having a single shared implementation (eg: :pep:`425`).
+
+.. end-intro
+
+The ``packaging`` project includes the following: version handling, specifiers,
+markers, requirements, tags, utilities.
+
+Documentation
+-------------
+
+The `documentation`_ provides information and the API for the following:
+
+- Version Handling
+- Specifiers
+- Markers
+- Requirements
+- Tags
+- Utilities
+
+Installation
+------------
+
+Use ``pip`` to install these utilities::
+
+    pip install packaging
+
+The ``packaging`` library uses calendar-based versioning (``YY.N``).
+
+Discussion
+----------
+
+If you run into bugs, you can file them in our `issue tracker`_.
+
+You can also join ``#pypa`` on Freenode to ask questions or get involved.
+
+
+.. _`documentation`: https://packaging.pypa.io/
+.. _`issue tracker`: https://github.com/pypa/packaging/issues
+
+
+Code of Conduct
+---------------
+
+Everyone interacting in the packaging project's codebases, issue trackers, chat
+rooms, and mailing lists is expected to follow the `PSF Code of Conduct`_.
+
+.. _PSF Code of Conduct: https://github.com/pypa/.github/blob/main/CODE_OF_CONDUCT.md
+
+Contributing
+------------
+
+The ``CONTRIBUTING.rst`` file outlines how to contribute to this project as
+well as how to report a potential security issue. The documentation for this
+project also covers information about `project development`_ and `security`_.
+
+.. _`project development`: https://packaging.pypa.io/en/latest/development/
+.. _`security`: https://packaging.pypa.io/en/latest/security/
+
+Project History
+---------------
+
+Please review the ``CHANGELOG.rst`` file or the `Changelog documentation`_ for
+recent changes and project history.
+
+.. _`Changelog documentation`: https://packaging.pypa.io/en/latest/changelog/
+
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/RECORD
new file mode 100644
index 00000000..556139f1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/RECORD
@@ -0,0 +1,40 @@
+packaging-24.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+packaging-24.2.dist-info/LICENSE,sha256=ytHvW9NA1z4HS6YU0m996spceUDD2MNIUuZcSQlobEg,197
+packaging-24.2.dist-info/LICENSE.APACHE,sha256=DVQuDIgE45qn836wDaWnYhSdxoLXgpRRKH4RuTjpRZQ,10174
+packaging-24.2.dist-info/LICENSE.BSD,sha256=tw5-m3QvHMb5SLNMFqo5_-zpQZY2S8iP8NIYDwAo-sU,1344
+packaging-24.2.dist-info/METADATA,sha256=ohH86s6k5mIfQxY2TS0LcSfADeOFa4BiCC-bxZV-pNs,3204
+packaging-24.2.dist-info/RECORD,,
+packaging-24.2.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+packaging/__init__.py,sha256=dk4Ta_vmdVJxYHDcfyhvQNw8V3PgSBomKNXqg-D2JDY,494
+packaging/__pycache__/__init__.cpython-312.pyc,,
+packaging/__pycache__/_elffile.cpython-312.pyc,,
+packaging/__pycache__/_manylinux.cpython-312.pyc,,
+packaging/__pycache__/_musllinux.cpython-312.pyc,,
+packaging/__pycache__/_parser.cpython-312.pyc,,
+packaging/__pycache__/_structures.cpython-312.pyc,,
+packaging/__pycache__/_tokenizer.cpython-312.pyc,,
+packaging/__pycache__/markers.cpython-312.pyc,,
+packaging/__pycache__/metadata.cpython-312.pyc,,
+packaging/__pycache__/requirements.cpython-312.pyc,,
+packaging/__pycache__/specifiers.cpython-312.pyc,,
+packaging/__pycache__/tags.cpython-312.pyc,,
+packaging/__pycache__/utils.cpython-312.pyc,,
+packaging/__pycache__/version.cpython-312.pyc,,
+packaging/_elffile.py,sha256=cflAQAkE25tzhYmq_aCi72QfbT_tn891tPzfpbeHOwE,3306
+packaging/_manylinux.py,sha256=vl5OCoz4kx80H5rwXKeXWjl9WNISGmr4ZgTpTP9lU9c,9612
+packaging/_musllinux.py,sha256=p9ZqNYiOItGee8KcZFeHF_YcdhVwGHdK6r-8lgixvGQ,2694
+packaging/_parser.py,sha256=s_TvTvDNK0NrM2QB3VKThdWFM4Nc0P6JnkObkl3MjpM,10236
+packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+packaging/_tokenizer.py,sha256=J6v5H7Jzvb-g81xp_2QACKwO7LxHQA6ikryMU7zXwN8,5273
+packaging/licenses/__init__.py,sha256=1x5M1nEYjcgwEbLt0dXwz2ukjr18DiCzC0sraQqJ-Ww,5715
+packaging/licenses/__pycache__/__init__.cpython-312.pyc,,
+packaging/licenses/__pycache__/_spdx.cpython-312.pyc,,
+packaging/licenses/_spdx.py,sha256=oAm1ztPFwlsmCKe7lAAsv_OIOfS1cWDu9bNBkeu-2ns,48398
+packaging/markers.py,sha256=c89TNzB7ZdGYhkovm6PYmqGyHxXlYVaLW591PHUNKD8,10561
+packaging/metadata.py,sha256=YJibM7GYe4re8-0a3OlXmGS-XDgTEoO4tlBt2q25Bng,34762
+packaging/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+packaging/requirements.py,sha256=gYyRSAdbrIyKDY66ugIDUQjRMvxkH2ALioTmX3tnL6o,2947
+packaging/specifiers.py,sha256=GG1wPNMcL0fMJO68vF53wKMdwnfehDcaI-r9NpTfilA,40074
+packaging/tags.py,sha256=CFqrJzAzc2XNGexerH__T-Y5Iwq7WbsYXsiLERLWxY0,21014
+packaging/utils.py,sha256=0F3Hh9OFuRgrhTgGZUl5K22Fv1YP2tZl1z_2gO6kJiA,5050
+packaging/version.py,sha256=olfyuk_DPbflNkJ4wBWetXQ17c74x3DB501degUv7DY,16676
diff --git a/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging-24.2.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__init__.py b/backend/venv/lib/python3.12/site-packages/packaging/__init__.py
new file mode 100644
index 00000000..d79f73c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/__init__.py
@@ -0,0 +1,15 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+__title__ = "packaging"
+__summary__ = "Core utilities for Python packages"
+__uri__ = "https://github.com/pypa/packaging"
+
+__version__ = "24.2"
+
+__author__ = "Donald Stufft and individual contributors"
+__email__ = "donald@stufft.io"
+
+__license__ = "BSD-2-Clause or Apache-2.0"
+__copyright__ = f"2014 {__author__}"
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..c32aa614
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_elffile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_elffile.cpython-312.pyc
new file mode 100644
index 00000000..a0124486
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_elffile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_manylinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_manylinux.cpython-312.pyc
new file mode 100644
index 00000000..2dbd7817
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_manylinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_musllinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_musllinux.cpython-312.pyc
new file mode 100644
index 00000000..c6ce0db0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_musllinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..612bd91d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_structures.cpython-312.pyc
new file mode 100644
index 00000000..a5c7ef56
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_tokenizer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_tokenizer.cpython-312.pyc
new file mode 100644
index 00000000..967ff32c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/_tokenizer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/markers.cpython-312.pyc
new file mode 100644
index 00000000..5c58e700
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/metadata.cpython-312.pyc
new file mode 100644
index 00000000..e1cb17e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..11f7f182
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/specifiers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/specifiers.cpython-312.pyc
new file mode 100644
index 00000000..22fb6900
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/specifiers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/tags.cpython-312.pyc
new file mode 100644
index 00000000..762ea698
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..8203b16b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..29d6f4a2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_elffile.py b/backend/venv/lib/python3.12/site-packages/packaging/_elffile.py
new file mode 100644
index 00000000..25f4282c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_elffile.py
@@ -0,0 +1,110 @@
+"""
+ELF file parser.
+
+This provides a class ``ELFFile`` that parses an ELF executable in a similar
+interface to ``ZipFile``. Only the read interface is implemented.
+
+Based on: https://gist.github.com/lyssdod/f51579ae8d93c8657a5564aefc2ffbca
+ELF header: https://refspecs.linuxfoundation.org/elf/gabi4+/ch4.eheader.html
+"""
+
+from __future__ import annotations
+
+import enum
+import os
+import struct
+from typing import IO
+
+
+class ELFInvalid(ValueError):
+    pass
+
+
+class EIClass(enum.IntEnum):
+    C32 = 1
+    C64 = 2
+
+
+class EIData(enum.IntEnum):
+    Lsb = 1
+    Msb = 2
+
+
+class EMachine(enum.IntEnum):
+    I386 = 3
+    S390 = 22
+    Arm = 40
+    X8664 = 62
+    AArc64 = 183
+
+
+class ELFFile:
+    """
+    Representation of an ELF executable.
+    """
+
+    def __init__(self, f: IO[bytes]) -> None:
+        self._f = f
+
+        try:
+            ident = self._read("16B")
+        except struct.error as e:
+            raise ELFInvalid("unable to parse identification") from e
+        magic = bytes(ident[:4])
+        if magic != b"\x7fELF":
+            raise ELFInvalid(f"invalid magic: {magic!r}")
+
+        self.capacity = ident[4]  # Format for program header (bitness).
+        self.encoding = ident[5]  # Data structure encoding (endianness).
+
+        try:
+            # e_fmt: Format for program header.
+            # p_fmt: Format for section header.
+            # p_idx: Indexes to find p_type, p_offset, and p_filesz.
+            e_fmt, self._p_fmt, self._p_idx = {
+                (1, 1): ("<HHIIIIIHHH", "<IIIIIIII", (0, 1, 4)),  # 32-bit LSB.
+                (1, 2): (">HHIIIIIHHH", ">IIIIIIII", (0, 1, 4)),  # 32-bit MSB.
+                (2, 1): ("<HHIQQQIHHH", "<IIQQQQQQ", (0, 2, 5)),  # 64-bit LSB.
+                (2, 2): (">HHIQQQIHHH", ">IIQQQQQQ", (0, 2, 5)),  # 64-bit MSB.
+            }[(self.capacity, self.encoding)]
+        except KeyError as e:
+            raise ELFInvalid(
+                f"unrecognized capacity ({self.capacity}) or "
+                f"encoding ({self.encoding})"
+            ) from e
+
+        try:
+            (
+                _,
+                self.machine,  # Architecture type.
+                _,
+                _,
+                self._e_phoff,  # Offset of program header.
+                _,
+                self.flags,  # Processor-specific flags.
+                _,
+                self._e_phentsize,  # Size of section.
+                self._e_phnum,  # Number of sections.
+            ) = self._read(e_fmt)
+        except struct.error as e:
+            raise ELFInvalid("unable to parse machine and section information") from e
+
+    def _read(self, fmt: str) -> tuple[int, ...]:
+        return struct.unpack(fmt, self._f.read(struct.calcsize(fmt)))
+
+    @property
+    def interpreter(self) -> str | None:
+        """
+        The path recorded in the ``PT_INTERP`` section header.
+        """
+        for index in range(self._e_phnum):
+            self._f.seek(self._e_phoff + self._e_phentsize * index)
+            try:
+                data = self._read(self._p_fmt)
+            except struct.error:
+                continue
+            if data[self._p_idx[0]] != 3:  # Not PT_INTERP.
+                continue
+            self._f.seek(data[self._p_idx[1]])
+            return os.fsdecode(self._f.read(data[self._p_idx[2]])).strip("\0")
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_manylinux.py b/backend/venv/lib/python3.12/site-packages/packaging/_manylinux.py
new file mode 100644
index 00000000..61339a6f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_manylinux.py
@@ -0,0 +1,263 @@
+from __future__ import annotations
+
+import collections
+import contextlib
+import functools
+import os
+import re
+import sys
+import warnings
+from typing import Generator, Iterator, NamedTuple, Sequence
+
+from ._elffile import EIClass, EIData, ELFFile, EMachine
+
+EF_ARM_ABIMASK = 0xFF000000
+EF_ARM_ABI_VER5 = 0x05000000
+EF_ARM_ABI_FLOAT_HARD = 0x00000400
+
+
+# `os.PathLike` not a generic type until Python 3.9, so sticking with `str`
+# as the type for `path` until then.
+@contextlib.contextmanager
+def _parse_elf(path: str) -> Generator[ELFFile | None, None, None]:
+    try:
+        with open(path, "rb") as f:
+            yield ELFFile(f)
+    except (OSError, TypeError, ValueError):
+        yield None
+
+
+def _is_linux_armhf(executable: str) -> bool:
+    # hard-float ABI can be detected from the ELF header of the running
+    # process
+    # https://static.docs.arm.com/ihi0044/g/aaelf32.pdf
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.Arm
+            and f.flags & EF_ARM_ABIMASK == EF_ARM_ABI_VER5
+            and f.flags & EF_ARM_ABI_FLOAT_HARD == EF_ARM_ABI_FLOAT_HARD
+        )
+
+
+def _is_linux_i686(executable: str) -> bool:
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.I386
+        )
+
+
+def _have_compatible_abi(executable: str, archs: Sequence[str]) -> bool:
+    if "armv7l" in archs:
+        return _is_linux_armhf(executable)
+    if "i686" in archs:
+        return _is_linux_i686(executable)
+    allowed_archs = {
+        "x86_64",
+        "aarch64",
+        "ppc64",
+        "ppc64le",
+        "s390x",
+        "loongarch64",
+        "riscv64",
+    }
+    return any(arch in allowed_archs for arch in archs)
+
+
+# If glibc ever changes its major version, we need to know what the last
+# minor version was, so we can build the complete list of all versions.
+# For now, guess what the highest minor version might be, assume it will
+# be 50 for testing. Once this actually happens, update the dictionary
+# with the actual value.
+_LAST_GLIBC_MINOR: dict[int, int] = collections.defaultdict(lambda: 50)
+
+
+class _GLibCVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _glibc_version_string_confstr() -> str | None:
+    """
+    Primary implementation of glibc_version_string using os.confstr.
+    """
+    # os.confstr is quite a bit faster than ctypes.DLL. It's also less likely
+    # to be broken or missing. This strategy is used in the standard library
+    # platform module.
+    # https://github.com/python/cpython/blob/fcf1d003bf4f0100c/Lib/platform.py#L175-L183
+    try:
+        # Should be a string like "glibc 2.17".
+        version_string: str | None = os.confstr("CS_GNU_LIBC_VERSION")
+        assert version_string is not None
+        _, version = version_string.rsplit()
+    except (AssertionError, AttributeError, OSError, ValueError):
+        # os.confstr() or CS_GNU_LIBC_VERSION not available (or a bad value)...
+        return None
+    return version
+
+
+def _glibc_version_string_ctypes() -> str | None:
+    """
+    Fallback implementation of glibc_version_string using ctypes.
+    """
+    try:
+        import ctypes
+    except ImportError:
+        return None
+
+    # ctypes.CDLL(None) internally calls dlopen(NULL), and as the dlopen
+    # manpage says, "If filename is NULL, then the returned handle is for the
+    # main program". This way we can let the linker do the work to figure out
+    # which libc our process is actually using.
+    #
+    # We must also handle the special case where the executable is not a
+    # dynamically linked executable. This can occur when using musl libc,
+    # for example. In this situation, dlopen() will error, leading to an
+    # OSError. Interestingly, at least in the case of musl, there is no
+    # errno set on the OSError. The single string argument used to construct
+    # OSError comes from libc itself and is therefore not portable to
+    # hard code here. In any case, failure to call dlopen() means we
+    # can proceed, so we bail on our attempt.
+    try:
+        process_namespace = ctypes.CDLL(None)
+    except OSError:
+        return None
+
+    try:
+        gnu_get_libc_version = process_namespace.gnu_get_libc_version
+    except AttributeError:
+        # Symbol doesn't exist -> therefore, we are not linked to
+        # glibc.
+        return None
+
+    # Call gnu_get_libc_version, which returns a string like "2.5"
+    gnu_get_libc_version.restype = ctypes.c_char_p
+    version_str: str = gnu_get_libc_version()
+    # py2 / py3 compatibility:
+    if not isinstance(version_str, str):
+        version_str = version_str.decode("ascii")
+
+    return version_str
+
+
+def _glibc_version_string() -> str | None:
+    """Returns glibc version string, or None if not using glibc."""
+    return _glibc_version_string_confstr() or _glibc_version_string_ctypes()
+
+
+def _parse_glibc_version(version_str: str) -> tuple[int, int]:
+    """Parse glibc version.
+
+    We use a regexp instead of str.split because we want to discard any
+    random junk that might come after the minor version -- this might happen
+    in patched/forked versions of glibc (e.g. Linaro's version of glibc
+    uses version strings like "2.20-2014.11"). See gh-3588.
+    """
+    m = re.match(r"(?P<major>[0-9]+)\.(?P<minor>[0-9]+)", version_str)
+    if not m:
+        warnings.warn(
+            f"Expected glibc version with 2 components major.minor,"
+            f" got: {version_str}",
+            RuntimeWarning,
+            stacklevel=2,
+        )
+        return -1, -1
+    return int(m.group("major")), int(m.group("minor"))
+
+
+@functools.lru_cache
+def _get_glibc_version() -> tuple[int, int]:
+    version_str = _glibc_version_string()
+    if version_str is None:
+        return (-1, -1)
+    return _parse_glibc_version(version_str)
+
+
+# From PEP 513, PEP 600
+def _is_compatible(arch: str, version: _GLibCVersion) -> bool:
+    sys_glibc = _get_glibc_version()
+    if sys_glibc < version:
+        return False
+    # Check for presence of _manylinux module.
+    try:
+        import _manylinux
+    except ImportError:
+        return True
+    if hasattr(_manylinux, "manylinux_compatible"):
+        result = _manylinux.manylinux_compatible(version[0], version[1], arch)
+        if result is not None:
+            return bool(result)
+        return True
+    if version == _GLibCVersion(2, 5):
+        if hasattr(_manylinux, "manylinux1_compatible"):
+            return bool(_manylinux.manylinux1_compatible)
+    if version == _GLibCVersion(2, 12):
+        if hasattr(_manylinux, "manylinux2010_compatible"):
+            return bool(_manylinux.manylinux2010_compatible)
+    if version == _GLibCVersion(2, 17):
+        if hasattr(_manylinux, "manylinux2014_compatible"):
+            return bool(_manylinux.manylinux2014_compatible)
+    return True
+
+
+_LEGACY_MANYLINUX_MAP = {
+    # CentOS 7 w/ glibc 2.17 (PEP 599)
+    (2, 17): "manylinux2014",
+    # CentOS 6 w/ glibc 2.12 (PEP 571)
+    (2, 12): "manylinux2010",
+    # CentOS 5 w/ glibc 2.5 (PEP 513)
+    (2, 5): "manylinux1",
+}
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate manylinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be manylinux-compatible.
+
+    :returns: An iterator of compatible manylinux tags.
+    """
+    if not _have_compatible_abi(sys.executable, archs):
+        return
+    # Oldest glibc to be supported regardless of architecture is (2, 17).
+    too_old_glibc2 = _GLibCVersion(2, 16)
+    if set(archs) & {"x86_64", "i686"}:
+        # On x86/i686 also oldest glibc to be supported is (2, 5).
+        too_old_glibc2 = _GLibCVersion(2, 4)
+    current_glibc = _GLibCVersion(*_get_glibc_version())
+    glibc_max_list = [current_glibc]
+    # We can assume compatibility across glibc major versions.
+    # https://sourceware.org/bugzilla/show_bug.cgi?id=24636
+    #
+    # Build a list of maximum glibc versions so that we can
+    # output the canonical list of all glibc from current_glibc
+    # down to too_old_glibc2, including all intermediary versions.
+    for glibc_major in range(current_glibc.major - 1, 1, -1):
+        glibc_minor = _LAST_GLIBC_MINOR[glibc_major]
+        glibc_max_list.append(_GLibCVersion(glibc_major, glibc_minor))
+    for arch in archs:
+        for glibc_max in glibc_max_list:
+            if glibc_max.major == too_old_glibc2.major:
+                min_minor = too_old_glibc2.minor
+            else:
+                # For other glibc major versions oldest supported is (x, 0).
+                min_minor = -1
+            for glibc_minor in range(glibc_max.minor, min_minor, -1):
+                glibc_version = _GLibCVersion(glibc_max.major, glibc_minor)
+                tag = "manylinux_{}_{}".format(*glibc_version)
+                if _is_compatible(arch, glibc_version):
+                    yield f"{tag}_{arch}"
+                # Handle the legacy manylinux1, manylinux2010, manylinux2014 tags.
+                if glibc_version in _LEGACY_MANYLINUX_MAP:
+                    legacy_tag = _LEGACY_MANYLINUX_MAP[glibc_version]
+                    if _is_compatible(arch, glibc_version):
+                        yield f"{legacy_tag}_{arch}"
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_musllinux.py b/backend/venv/lib/python3.12/site-packages/packaging/_musllinux.py
new file mode 100644
index 00000000..d2bf30b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_musllinux.py
@@ -0,0 +1,85 @@
+"""PEP 656 support.
+
+This module implements logic to detect if the currently running Python is
+linked against musl, and what musl version is used.
+"""
+
+from __future__ import annotations
+
+import functools
+import re
+import subprocess
+import sys
+from typing import Iterator, NamedTuple, Sequence
+
+from ._elffile import ELFFile
+
+
+class _MuslVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _parse_musl_version(output: str) -> _MuslVersion | None:
+    lines = [n for n in (n.strip() for n in output.splitlines()) if n]
+    if len(lines) < 2 or lines[0][:4] != "musl":
+        return None
+    m = re.match(r"Version (\d+)\.(\d+)", lines[1])
+    if not m:
+        return None
+    return _MuslVersion(major=int(m.group(1)), minor=int(m.group(2)))
+
+
+@functools.lru_cache
+def _get_musl_version(executable: str) -> _MuslVersion | None:
+    """Detect currently-running musl runtime version.
+
+    This is done by checking the specified executable's dynamic linking
+    information, and invoking the loader to parse its output for a version
+    string. If the loader is musl, the output would be something like::
+
+        musl libc (x86_64)
+        Version 1.2.2
+        Dynamic Program Loader
+    """
+    try:
+        with open(executable, "rb") as f:
+            ld = ELFFile(f).interpreter
+    except (OSError, TypeError, ValueError):
+        return None
+    if ld is None or "musl" not in ld:
+        return None
+    proc = subprocess.run([ld], stderr=subprocess.PIPE, text=True)
+    return _parse_musl_version(proc.stderr)
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate musllinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be musllinux-compatible.
+
+    :returns: An iterator of compatible musllinux tags.
+    """
+    sys_musl = _get_musl_version(sys.executable)
+    if sys_musl is None:  # Python not dynamically linked against musl.
+        return
+    for arch in archs:
+        for minor in range(sys_musl.minor, -1, -1):
+            yield f"musllinux_{sys_musl.major}_{minor}_{arch}"
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sysconfig
+
+    plat = sysconfig.get_platform()
+    assert plat.startswith("linux-"), "not linux"
+
+    print("plat:", plat)
+    print("musl:", _get_musl_version(sys.executable))
+    print("tags:", end=" ")
+    for t in platform_tags(re.sub(r"[.-]", "_", plat.split("-", 1)[-1])):
+        print(t, end="\n      ")
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_parser.py b/backend/venv/lib/python3.12/site-packages/packaging/_parser.py
new file mode 100644
index 00000000..c1238c06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_parser.py
@@ -0,0 +1,354 @@
+"""Handwritten parser of dependency specifiers.
+
+The docstring for each __parse_* function contains EBNF-inspired grammar representing
+the implementation.
+"""
+
+from __future__ import annotations
+
+import ast
+from typing import NamedTuple, Sequence, Tuple, Union
+
+from ._tokenizer import DEFAULT_RULES, Tokenizer
+
+
+class Node:
+    def __init__(self, value: str) -> None:
+        self.value = value
+
+    def __str__(self) -> str:
+        return self.value
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__}('{self}')>"
+
+    def serialize(self) -> str:
+        raise NotImplementedError
+
+
+class Variable(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+class Value(Node):
+    def serialize(self) -> str:
+        return f'"{self}"'
+
+
+class Op(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+MarkerVar = Union[Variable, Value]
+MarkerItem = Tuple[MarkerVar, Op, MarkerVar]
+MarkerAtom = Union[MarkerItem, Sequence["MarkerAtom"]]
+MarkerList = Sequence[Union["MarkerList", MarkerAtom, str]]
+
+
+class ParsedRequirement(NamedTuple):
+    name: str
+    url: str
+    extras: list[str]
+    specifier: str
+    marker: MarkerList | None
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for dependency specifier
+# --------------------------------------------------------------------------------------
+def parse_requirement(source: str) -> ParsedRequirement:
+    return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_requirement(tokenizer: Tokenizer) -> ParsedRequirement:
+    """
+    requirement = WS? IDENTIFIER WS? extras WS? requirement_details
+    """
+    tokenizer.consume("WS")
+
+    name_token = tokenizer.expect(
+        "IDENTIFIER", expected="package name at the start of dependency specifier"
+    )
+    name = name_token.text
+    tokenizer.consume("WS")
+
+    extras = _parse_extras(tokenizer)
+    tokenizer.consume("WS")
+
+    url, specifier, marker = _parse_requirement_details(tokenizer)
+    tokenizer.expect("END", expected="end of dependency specifier")
+
+    return ParsedRequirement(name, url, extras, specifier, marker)
+
+
+def _parse_requirement_details(
+    tokenizer: Tokenizer,
+) -> tuple[str, str, MarkerList | None]:
+    """
+    requirement_details = AT URL (WS requirement_marker?)?
+                        | specifier WS? (requirement_marker)?
+    """
+
+    specifier = ""
+    url = ""
+    marker = None
+
+    if tokenizer.check("AT"):
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        url_start = tokenizer.position
+        url = tokenizer.expect("URL", expected="URL after @").text
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        tokenizer.expect("WS", expected="whitespace after URL")
+
+        # The input might end after whitespace.
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer, span_start=url_start, after="URL and whitespace"
+        )
+    else:
+        specifier_start = tokenizer.position
+        specifier = _parse_specifier(tokenizer)
+        tokenizer.consume("WS")
+
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer,
+            span_start=specifier_start,
+            after=(
+                "version specifier"
+                if specifier
+                else "name and no valid version specifier"
+            ),
+        )
+
+    return (url, specifier, marker)
+
+
+def _parse_requirement_marker(
+    tokenizer: Tokenizer, *, span_start: int, after: str
+) -> MarkerList:
+    """
+    requirement_marker = SEMICOLON marker WS?
+    """
+
+    if not tokenizer.check("SEMICOLON"):
+        tokenizer.raise_syntax_error(
+            f"Expected end or semicolon (after {after})",
+            span_start=span_start,
+        )
+    tokenizer.read()
+
+    marker = _parse_marker(tokenizer)
+    tokenizer.consume("WS")
+
+    return marker
+
+
+def _parse_extras(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras = (LEFT_BRACKET wsp* extras_list? wsp* RIGHT_BRACKET)?
+    """
+    if not tokenizer.check("LEFT_BRACKET", peek=True):
+        return []
+
+    with tokenizer.enclosing_tokens(
+        "LEFT_BRACKET",
+        "RIGHT_BRACKET",
+        around="extras",
+    ):
+        tokenizer.consume("WS")
+        extras = _parse_extras_list(tokenizer)
+        tokenizer.consume("WS")
+
+    return extras
+
+
+def _parse_extras_list(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras_list = identifier (wsp* ',' wsp* identifier)*
+    """
+    extras: list[str] = []
+
+    if not tokenizer.check("IDENTIFIER"):
+        return extras
+
+    extras.append(tokenizer.read().text)
+
+    while True:
+        tokenizer.consume("WS")
+        if tokenizer.check("IDENTIFIER", peek=True):
+            tokenizer.raise_syntax_error("Expected comma between extra names")
+        elif not tokenizer.check("COMMA"):
+            break
+
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        extra_token = tokenizer.expect("IDENTIFIER", expected="extra name after comma")
+        extras.append(extra_token.text)
+
+    return extras
+
+
+def _parse_specifier(tokenizer: Tokenizer) -> str:
+    """
+    specifier = LEFT_PARENTHESIS WS? version_many WS? RIGHT_PARENTHESIS
+              | WS? version_many WS?
+    """
+    with tokenizer.enclosing_tokens(
+        "LEFT_PARENTHESIS",
+        "RIGHT_PARENTHESIS",
+        around="version specifier",
+    ):
+        tokenizer.consume("WS")
+        parsed_specifiers = _parse_version_many(tokenizer)
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+def _parse_version_many(tokenizer: Tokenizer) -> str:
+    """
+    version_many = (SPECIFIER (WS? COMMA WS? SPECIFIER)*)?
+    """
+    parsed_specifiers = ""
+    while tokenizer.check("SPECIFIER"):
+        span_start = tokenizer.position
+        parsed_specifiers += tokenizer.read().text
+        if tokenizer.check("VERSION_PREFIX_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                ".* suffix can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position + 1,
+            )
+        if tokenizer.check("VERSION_LOCAL_LABEL_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                "Local version label can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position,
+            )
+        tokenizer.consume("WS")
+        if not tokenizer.check("COMMA"):
+            break
+        parsed_specifiers += tokenizer.read().text
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for marker expression
+# --------------------------------------------------------------------------------------
+def parse_marker(source: str) -> MarkerList:
+    return _parse_full_marker(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_full_marker(tokenizer: Tokenizer) -> MarkerList:
+    retval = _parse_marker(tokenizer)
+    tokenizer.expect("END", expected="end of marker expression")
+    return retval
+
+
+def _parse_marker(tokenizer: Tokenizer) -> MarkerList:
+    """
+    marker = marker_atom (BOOLOP marker_atom)+
+    """
+    expression = [_parse_marker_atom(tokenizer)]
+    while tokenizer.check("BOOLOP"):
+        token = tokenizer.read()
+        expr_right = _parse_marker_atom(tokenizer)
+        expression.extend((token.text, expr_right))
+    return expression
+
+
+def _parse_marker_atom(tokenizer: Tokenizer) -> MarkerAtom:
+    """
+    marker_atom = WS? LEFT_PARENTHESIS WS? marker WS? RIGHT_PARENTHESIS WS?
+                | WS? marker_item WS?
+    """
+
+    tokenizer.consume("WS")
+    if tokenizer.check("LEFT_PARENTHESIS", peek=True):
+        with tokenizer.enclosing_tokens(
+            "LEFT_PARENTHESIS",
+            "RIGHT_PARENTHESIS",
+            around="marker expression",
+        ):
+            tokenizer.consume("WS")
+            marker: MarkerAtom = _parse_marker(tokenizer)
+            tokenizer.consume("WS")
+    else:
+        marker = _parse_marker_item(tokenizer)
+    tokenizer.consume("WS")
+    return marker
+
+
+def _parse_marker_item(tokenizer: Tokenizer) -> MarkerItem:
+    """
+    marker_item = WS? marker_var WS? marker_op WS? marker_var WS?
+    """
+    tokenizer.consume("WS")
+    marker_var_left = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    marker_op = _parse_marker_op(tokenizer)
+    tokenizer.consume("WS")
+    marker_var_right = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    return (marker_var_left, marker_op, marker_var_right)
+
+
+def _parse_marker_var(tokenizer: Tokenizer) -> MarkerVar:
+    """
+    marker_var = VARIABLE | QUOTED_STRING
+    """
+    if tokenizer.check("VARIABLE"):
+        return process_env_var(tokenizer.read().text.replace(".", "_"))
+    elif tokenizer.check("QUOTED_STRING"):
+        return process_python_str(tokenizer.read().text)
+    else:
+        tokenizer.raise_syntax_error(
+            message="Expected a marker variable or quoted string"
+        )
+
+
+def process_env_var(env_var: str) -> Variable:
+    if env_var in ("platform_python_implementation", "python_implementation"):
+        return Variable("platform_python_implementation")
+    else:
+        return Variable(env_var)
+
+
+def process_python_str(python_str: str) -> Value:
+    value = ast.literal_eval(python_str)
+    return Value(str(value))
+
+
+def _parse_marker_op(tokenizer: Tokenizer) -> Op:
+    """
+    marker_op = IN | NOT IN | OP
+    """
+    if tokenizer.check("IN"):
+        tokenizer.read()
+        return Op("in")
+    elif tokenizer.check("NOT"):
+        tokenizer.read()
+        tokenizer.expect("WS", expected="whitespace after 'not'")
+        tokenizer.expect("IN", expected="'in' after 'not'")
+        return Op("not in")
+    elif tokenizer.check("OP"):
+        return Op(tokenizer.read().text)
+    else:
+        return tokenizer.raise_syntax_error(
+            "Expected marker operator, one of "
+            "<=, <, !=, ==, >=, >, ~=, ===, in, not in"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_structures.py b/backend/venv/lib/python3.12/site-packages/packaging/_structures.py
new file mode 100644
index 00000000..90a6465f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_structures.py
@@ -0,0 +1,61 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+
+class InfinityType:
+    def __repr__(self) -> str:
+        return "Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return False
+
+    def __le__(self, other: object) -> bool:
+        return False
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return True
+
+    def __ge__(self, other: object) -> bool:
+        return True
+
+    def __neg__(self: object) -> "NegativeInfinityType":
+        return NegativeInfinity
+
+
+Infinity = InfinityType()
+
+
+class NegativeInfinityType:
+    def __repr__(self) -> str:
+        return "-Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return True
+
+    def __le__(self, other: object) -> bool:
+        return True
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return False
+
+    def __ge__(self, other: object) -> bool:
+        return False
+
+    def __neg__(self: object) -> InfinityType:
+        return Infinity
+
+
+NegativeInfinity = NegativeInfinityType()
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/_tokenizer.py b/backend/venv/lib/python3.12/site-packages/packaging/_tokenizer.py
new file mode 100644
index 00000000..89d04160
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/_tokenizer.py
@@ -0,0 +1,194 @@
+from __future__ import annotations
+
+import contextlib
+import re
+from dataclasses import dataclass
+from typing import Iterator, NoReturn
+
+from .specifiers import Specifier
+
+
+@dataclass
+class Token:
+    name: str
+    text: str
+    position: int
+
+
+class ParserSyntaxError(Exception):
+    """The provided source text could not be parsed correctly."""
+
+    def __init__(
+        self,
+        message: str,
+        *,
+        source: str,
+        span: tuple[int, int],
+    ) -> None:
+        self.span = span
+        self.message = message
+        self.source = source
+
+        super().__init__()
+
+    def __str__(self) -> str:
+        marker = " " * self.span[0] + "~" * (self.span[1] - self.span[0]) + "^"
+        return "\n    ".join([self.message, self.source, marker])
+
+
+DEFAULT_RULES: dict[str, str | re.Pattern[str]] = {
+    "LEFT_PARENTHESIS": r"\(",
+    "RIGHT_PARENTHESIS": r"\)",
+    "LEFT_BRACKET": r"\[",
+    "RIGHT_BRACKET": r"\]",
+    "SEMICOLON": r";",
+    "COMMA": r",",
+    "QUOTED_STRING": re.compile(
+        r"""
+            (
+                ('[^']*')
+                |
+                ("[^"]*")
+            )
+        """,
+        re.VERBOSE,
+    ),
+    "OP": r"(===|==|~=|!=|<=|>=|<|>)",
+    "BOOLOP": r"\b(or|and)\b",
+    "IN": r"\bin\b",
+    "NOT": r"\bnot\b",
+    "VARIABLE": re.compile(
+        r"""
+            \b(
+                python_version
+                |python_full_version
+                |os[._]name
+                |sys[._]platform
+                |platform_(release|system)
+                |platform[._](version|machine|python_implementation)
+                |python_implementation
+                |implementation_(name|version)
+                |extra
+            )\b
+        """,
+        re.VERBOSE,
+    ),
+    "SPECIFIER": re.compile(
+        Specifier._operator_regex_str + Specifier._version_regex_str,
+        re.VERBOSE | re.IGNORECASE,
+    ),
+    "AT": r"\@",
+    "URL": r"[^ \t]+",
+    "IDENTIFIER": r"\b[a-zA-Z0-9][a-zA-Z0-9._-]*\b",
+    "VERSION_PREFIX_TRAIL": r"\.\*",
+    "VERSION_LOCAL_LABEL_TRAIL": r"\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*",
+    "WS": r"[ \t]+",
+    "END": r"$",
+}
+
+
+class Tokenizer:
+    """Context-sensitive token parsing.
+
+    Provides methods to examine the input stream to check whether the next token
+    matches.
+    """
+
+    def __init__(
+        self,
+        source: str,
+        *,
+        rules: dict[str, str | re.Pattern[str]],
+    ) -> None:
+        self.source = source
+        self.rules: dict[str, re.Pattern[str]] = {
+            name: re.compile(pattern) for name, pattern in rules.items()
+        }
+        self.next_token: Token | None = None
+        self.position = 0
+
+    def consume(self, name: str) -> None:
+        """Move beyond provided token name, if at current position."""
+        if self.check(name):
+            self.read()
+
+    def check(self, name: str, *, peek: bool = False) -> bool:
+        """Check whether the next token has the provided name.
+
+        By default, if the check succeeds, the token *must* be read before
+        another check. If `peek` is set to `True`, the token is not loaded and
+        would need to be checked again.
+        """
+        assert (
+            self.next_token is None
+        ), f"Cannot check for {name!r}, already have {self.next_token!r}"
+        assert name in self.rules, f"Unknown token name: {name!r}"
+
+        expression = self.rules[name]
+
+        match = expression.match(self.source, self.position)
+        if match is None:
+            return False
+        if not peek:
+            self.next_token = Token(name, match[0], self.position)
+        return True
+
+    def expect(self, name: str, *, expected: str) -> Token:
+        """Expect a certain token name next, failing with a syntax error otherwise.
+
+        The token is *not* read.
+        """
+        if not self.check(name):
+            raise self.raise_syntax_error(f"Expected {expected}")
+        return self.read()
+
+    def read(self) -> Token:
+        """Consume the next token and return it."""
+        token = self.next_token
+        assert token is not None
+
+        self.position += len(token.text)
+        self.next_token = None
+
+        return token
+
+    def raise_syntax_error(
+        self,
+        message: str,
+        *,
+        span_start: int | None = None,
+        span_end: int | None = None,
+    ) -> NoReturn:
+        """Raise ParserSyntaxError at the given position."""
+        span = (
+            self.position if span_start is None else span_start,
+            self.position if span_end is None else span_end,
+        )
+        raise ParserSyntaxError(
+            message,
+            source=self.source,
+            span=span,
+        )
+
+    @contextlib.contextmanager
+    def enclosing_tokens(
+        self, open_token: str, close_token: str, *, around: str
+    ) -> Iterator[None]:
+        if self.check(open_token):
+            open_position = self.position
+            self.read()
+        else:
+            open_position = None
+
+        yield
+
+        if open_position is None:
+            return
+
+        if not self.check(close_token):
+            self.raise_syntax_error(
+                f"Expected matching {close_token} for {open_token}, after {around}",
+                span_start=open_position,
+            )
+
+        self.read()
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/licenses/__init__.py b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__init__.py
new file mode 100644
index 00000000..569156d6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__init__.py
@@ -0,0 +1,145 @@
+#######################################################################################
+#
+# Adapted from:
+#  https://github.com/pypa/hatch/blob/5352e44/backend/src/hatchling/licenses/parse.py
+#
+# MIT License
+#
+# Copyright (c) 2017-present Ofek Lev <oss@ofek.dev>
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this
+# software and associated documentation files (the "Software"), to deal in the Software
+# without restriction, including without limitation the rights to use, copy, modify,
+# merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to the following
+# conditions:
+#
+# The above copyright notice and this permission notice shall be included in all copies
+# or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+# PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
+# CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
+# OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+#
+#
+# With additional allowance of arbitrary `LicenseRef-` identifiers, not just
+# `LicenseRef-Public-Domain` and `LicenseRef-Proprietary`.
+#
+#######################################################################################
+from __future__ import annotations
+
+import re
+from typing import NewType, cast
+
+from packaging.licenses._spdx import EXCEPTIONS, LICENSES
+
+__all__ = [
+    "NormalizedLicenseExpression",
+    "InvalidLicenseExpression",
+    "canonicalize_license_expression",
+]
+
+license_ref_allowed = re.compile("^[A-Za-z0-9.-]*$")
+
+NormalizedLicenseExpression = NewType("NormalizedLicenseExpression", str)
+
+
+class InvalidLicenseExpression(ValueError):
+    """Raised when a license-expression string is invalid
+
+    >>> canonicalize_license_expression("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.licenses.InvalidLicenseExpression: Invalid license expression: 'invalid'
+    """
+
+
+def canonicalize_license_expression(
+    raw_license_expression: str,
+) -> NormalizedLicenseExpression:
+    if not raw_license_expression:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message)
+
+    # Pad any parentheses so tokenization can be achieved by merely splitting on
+    # whitespace.
+    license_expression = raw_license_expression.replace("(", " ( ").replace(")", " ) ")
+    licenseref_prefix = "LicenseRef-"
+    license_refs = {
+        ref.lower(): "LicenseRef-" + ref[len(licenseref_prefix) :]
+        for ref in license_expression.split()
+        if ref.lower().startswith(licenseref_prefix.lower())
+    }
+
+    # Normalize to lower case so we can look up licenses/exceptions
+    # and so boolean operators are Python-compatible.
+    license_expression = license_expression.lower()
+
+    tokens = license_expression.split()
+
+    # Rather than implementing boolean logic, we create an expression that Python can
+    # parse. Everything that is not involved with the grammar itself is treated as
+    # `False` and the expression should evaluate as such.
+    python_tokens = []
+    for token in tokens:
+        if token not in {"or", "and", "with", "(", ")"}:
+            python_tokens.append("False")
+        elif token == "with":
+            python_tokens.append("or")
+        elif token == "(" and python_tokens and python_tokens[-1] not in {"or", "and"}:
+            message = f"Invalid license expression: {raw_license_expression!r}"
+            raise InvalidLicenseExpression(message)
+        else:
+            python_tokens.append(token)
+
+    python_expression = " ".join(python_tokens)
+    try:
+        invalid = eval(python_expression, globals(), locals())
+    except Exception:
+        invalid = True
+
+    if invalid is not False:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message) from None
+
+    # Take a final pass to check for unknown licenses/exceptions.
+    normalized_tokens = []
+    for token in tokens:
+        if token in {"or", "and", "with", "(", ")"}:
+            normalized_tokens.append(token.upper())
+            continue
+
+        if normalized_tokens and normalized_tokens[-1] == "WITH":
+            if token not in EXCEPTIONS:
+                message = f"Unknown license exception: {token!r}"
+                raise InvalidLicenseExpression(message)
+
+            normalized_tokens.append(EXCEPTIONS[token]["id"])
+        else:
+            if token.endswith("+"):
+                final_token = token[:-1]
+                suffix = "+"
+            else:
+                final_token = token
+                suffix = ""
+
+            if final_token.startswith("licenseref-"):
+                if not license_ref_allowed.match(final_token):
+                    message = f"Invalid licenseref: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(license_refs[final_token] + suffix)
+            else:
+                if final_token not in LICENSES:
+                    message = f"Unknown license: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(LICENSES[final_token]["id"] + suffix)
+
+    normalized_expression = " ".join(normalized_tokens)
+
+    return cast(
+        NormalizedLicenseExpression,
+        normalized_expression.replace("( ", "(").replace(" )", ")"),
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..5eb95915
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/_spdx.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/_spdx.cpython-312.pyc
new file mode 100644
index 00000000..226b6f75
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/packaging/licenses/__pycache__/_spdx.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/licenses/_spdx.py b/backend/venv/lib/python3.12/site-packages/packaging/licenses/_spdx.py
new file mode 100644
index 00000000..eac22276
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/licenses/_spdx.py
@@ -0,0 +1,759 @@
+
+from __future__ import annotations
+
+from typing import TypedDict
+
+class SPDXLicense(TypedDict):
+    id: str
+    deprecated: bool
+
+class SPDXException(TypedDict):
+    id: str
+    deprecated: bool
+
+
+VERSION = '3.25.0'
+
+LICENSES: dict[str, SPDXLicense] = {
+    '0bsd': {'id': '0BSD', 'deprecated': False},
+    '3d-slicer-1.0': {'id': '3D-Slicer-1.0', 'deprecated': False},
+    'aal': {'id': 'AAL', 'deprecated': False},
+    'abstyles': {'id': 'Abstyles', 'deprecated': False},
+    'adacore-doc': {'id': 'AdaCore-doc', 'deprecated': False},
+    'adobe-2006': {'id': 'Adobe-2006', 'deprecated': False},
+    'adobe-display-postscript': {'id': 'Adobe-Display-PostScript', 'deprecated': False},
+    'adobe-glyph': {'id': 'Adobe-Glyph', 'deprecated': False},
+    'adobe-utopia': {'id': 'Adobe-Utopia', 'deprecated': False},
+    'adsl': {'id': 'ADSL', 'deprecated': False},
+    'afl-1.1': {'id': 'AFL-1.1', 'deprecated': False},
+    'afl-1.2': {'id': 'AFL-1.2', 'deprecated': False},
+    'afl-2.0': {'id': 'AFL-2.0', 'deprecated': False},
+    'afl-2.1': {'id': 'AFL-2.1', 'deprecated': False},
+    'afl-3.0': {'id': 'AFL-3.0', 'deprecated': False},
+    'afmparse': {'id': 'Afmparse', 'deprecated': False},
+    'agpl-1.0': {'id': 'AGPL-1.0', 'deprecated': True},
+    'agpl-1.0-only': {'id': 'AGPL-1.0-only', 'deprecated': False},
+    'agpl-1.0-or-later': {'id': 'AGPL-1.0-or-later', 'deprecated': False},
+    'agpl-3.0': {'id': 'AGPL-3.0', 'deprecated': True},
+    'agpl-3.0-only': {'id': 'AGPL-3.0-only', 'deprecated': False},
+    'agpl-3.0-or-later': {'id': 'AGPL-3.0-or-later', 'deprecated': False},
+    'aladdin': {'id': 'Aladdin', 'deprecated': False},
+    'amd-newlib': {'id': 'AMD-newlib', 'deprecated': False},
+    'amdplpa': {'id': 'AMDPLPA', 'deprecated': False},
+    'aml': {'id': 'AML', 'deprecated': False},
+    'aml-glslang': {'id': 'AML-glslang', 'deprecated': False},
+    'ampas': {'id': 'AMPAS', 'deprecated': False},
+    'antlr-pd': {'id': 'ANTLR-PD', 'deprecated': False},
+    'antlr-pd-fallback': {'id': 'ANTLR-PD-fallback', 'deprecated': False},
+    'any-osi': {'id': 'any-OSI', 'deprecated': False},
+    'apache-1.0': {'id': 'Apache-1.0', 'deprecated': False},
+    'apache-1.1': {'id': 'Apache-1.1', 'deprecated': False},
+    'apache-2.0': {'id': 'Apache-2.0', 'deprecated': False},
+    'apafml': {'id': 'APAFML', 'deprecated': False},
+    'apl-1.0': {'id': 'APL-1.0', 'deprecated': False},
+    'app-s2p': {'id': 'App-s2p', 'deprecated': False},
+    'apsl-1.0': {'id': 'APSL-1.0', 'deprecated': False},
+    'apsl-1.1': {'id': 'APSL-1.1', 'deprecated': False},
+    'apsl-1.2': {'id': 'APSL-1.2', 'deprecated': False},
+    'apsl-2.0': {'id': 'APSL-2.0', 'deprecated': False},
+    'arphic-1999': {'id': 'Arphic-1999', 'deprecated': False},
+    'artistic-1.0': {'id': 'Artistic-1.0', 'deprecated': False},
+    'artistic-1.0-cl8': {'id': 'Artistic-1.0-cl8', 'deprecated': False},
+    'artistic-1.0-perl': {'id': 'Artistic-1.0-Perl', 'deprecated': False},
+    'artistic-2.0': {'id': 'Artistic-2.0', 'deprecated': False},
+    'aswf-digital-assets-1.0': {'id': 'ASWF-Digital-Assets-1.0', 'deprecated': False},
+    'aswf-digital-assets-1.1': {'id': 'ASWF-Digital-Assets-1.1', 'deprecated': False},
+    'baekmuk': {'id': 'Baekmuk', 'deprecated': False},
+    'bahyph': {'id': 'Bahyph', 'deprecated': False},
+    'barr': {'id': 'Barr', 'deprecated': False},
+    'bcrypt-solar-designer': {'id': 'bcrypt-Solar-Designer', 'deprecated': False},
+    'beerware': {'id': 'Beerware', 'deprecated': False},
+    'bitstream-charter': {'id': 'Bitstream-Charter', 'deprecated': False},
+    'bitstream-vera': {'id': 'Bitstream-Vera', 'deprecated': False},
+    'bittorrent-1.0': {'id': 'BitTorrent-1.0', 'deprecated': False},
+    'bittorrent-1.1': {'id': 'BitTorrent-1.1', 'deprecated': False},
+    'blessing': {'id': 'blessing', 'deprecated': False},
+    'blueoak-1.0.0': {'id': 'BlueOak-1.0.0', 'deprecated': False},
+    'boehm-gc': {'id': 'Boehm-GC', 'deprecated': False},
+    'borceux': {'id': 'Borceux', 'deprecated': False},
+    'brian-gladman-2-clause': {'id': 'Brian-Gladman-2-Clause', 'deprecated': False},
+    'brian-gladman-3-clause': {'id': 'Brian-Gladman-3-Clause', 'deprecated': False},
+    'bsd-1-clause': {'id': 'BSD-1-Clause', 'deprecated': False},
+    'bsd-2-clause': {'id': 'BSD-2-Clause', 'deprecated': False},
+    'bsd-2-clause-darwin': {'id': 'BSD-2-Clause-Darwin', 'deprecated': False},
+    'bsd-2-clause-first-lines': {'id': 'BSD-2-Clause-first-lines', 'deprecated': False},
+    'bsd-2-clause-freebsd': {'id': 'BSD-2-Clause-FreeBSD', 'deprecated': True},
+    'bsd-2-clause-netbsd': {'id': 'BSD-2-Clause-NetBSD', 'deprecated': True},
+    'bsd-2-clause-patent': {'id': 'BSD-2-Clause-Patent', 'deprecated': False},
+    'bsd-2-clause-views': {'id': 'BSD-2-Clause-Views', 'deprecated': False},
+    'bsd-3-clause': {'id': 'BSD-3-Clause', 'deprecated': False},
+    'bsd-3-clause-acpica': {'id': 'BSD-3-Clause-acpica', 'deprecated': False},
+    'bsd-3-clause-attribution': {'id': 'BSD-3-Clause-Attribution', 'deprecated': False},
+    'bsd-3-clause-clear': {'id': 'BSD-3-Clause-Clear', 'deprecated': False},
+    'bsd-3-clause-flex': {'id': 'BSD-3-Clause-flex', 'deprecated': False},
+    'bsd-3-clause-hp': {'id': 'BSD-3-Clause-HP', 'deprecated': False},
+    'bsd-3-clause-lbnl': {'id': 'BSD-3-Clause-LBNL', 'deprecated': False},
+    'bsd-3-clause-modification': {'id': 'BSD-3-Clause-Modification', 'deprecated': False},
+    'bsd-3-clause-no-military-license': {'id': 'BSD-3-Clause-No-Military-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license': {'id': 'BSD-3-Clause-No-Nuclear-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license-2014': {'id': 'BSD-3-Clause-No-Nuclear-License-2014', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-warranty': {'id': 'BSD-3-Clause-No-Nuclear-Warranty', 'deprecated': False},
+    'bsd-3-clause-open-mpi': {'id': 'BSD-3-Clause-Open-MPI', 'deprecated': False},
+    'bsd-3-clause-sun': {'id': 'BSD-3-Clause-Sun', 'deprecated': False},
+    'bsd-4-clause': {'id': 'BSD-4-Clause', 'deprecated': False},
+    'bsd-4-clause-shortened': {'id': 'BSD-4-Clause-Shortened', 'deprecated': False},
+    'bsd-4-clause-uc': {'id': 'BSD-4-Clause-UC', 'deprecated': False},
+    'bsd-4.3reno': {'id': 'BSD-4.3RENO', 'deprecated': False},
+    'bsd-4.3tahoe': {'id': 'BSD-4.3TAHOE', 'deprecated': False},
+    'bsd-advertising-acknowledgement': {'id': 'BSD-Advertising-Acknowledgement', 'deprecated': False},
+    'bsd-attribution-hpnd-disclaimer': {'id': 'BSD-Attribution-HPND-disclaimer', 'deprecated': False},
+    'bsd-inferno-nettverk': {'id': 'BSD-Inferno-Nettverk', 'deprecated': False},
+    'bsd-protection': {'id': 'BSD-Protection', 'deprecated': False},
+    'bsd-source-beginning-file': {'id': 'BSD-Source-beginning-file', 'deprecated': False},
+    'bsd-source-code': {'id': 'BSD-Source-Code', 'deprecated': False},
+    'bsd-systemics': {'id': 'BSD-Systemics', 'deprecated': False},
+    'bsd-systemics-w3works': {'id': 'BSD-Systemics-W3Works', 'deprecated': False},
+    'bsl-1.0': {'id': 'BSL-1.0', 'deprecated': False},
+    'busl-1.1': {'id': 'BUSL-1.1', 'deprecated': False},
+    'bzip2-1.0.5': {'id': 'bzip2-1.0.5', 'deprecated': True},
+    'bzip2-1.0.6': {'id': 'bzip2-1.0.6', 'deprecated': False},
+    'c-uda-1.0': {'id': 'C-UDA-1.0', 'deprecated': False},
+    'cal-1.0': {'id': 'CAL-1.0', 'deprecated': False},
+    'cal-1.0-combined-work-exception': {'id': 'CAL-1.0-Combined-Work-Exception', 'deprecated': False},
+    'caldera': {'id': 'Caldera', 'deprecated': False},
+    'caldera-no-preamble': {'id': 'Caldera-no-preamble', 'deprecated': False},
+    'catharon': {'id': 'Catharon', 'deprecated': False},
+    'catosl-1.1': {'id': 'CATOSL-1.1', 'deprecated': False},
+    'cc-by-1.0': {'id': 'CC-BY-1.0', 'deprecated': False},
+    'cc-by-2.0': {'id': 'CC-BY-2.0', 'deprecated': False},
+    'cc-by-2.5': {'id': 'CC-BY-2.5', 'deprecated': False},
+    'cc-by-2.5-au': {'id': 'CC-BY-2.5-AU', 'deprecated': False},
+    'cc-by-3.0': {'id': 'CC-BY-3.0', 'deprecated': False},
+    'cc-by-3.0-at': {'id': 'CC-BY-3.0-AT', 'deprecated': False},
+    'cc-by-3.0-au': {'id': 'CC-BY-3.0-AU', 'deprecated': False},
+    'cc-by-3.0-de': {'id': 'CC-BY-3.0-DE', 'deprecated': False},
+    'cc-by-3.0-igo': {'id': 'CC-BY-3.0-IGO', 'deprecated': False},
+    'cc-by-3.0-nl': {'id': 'CC-BY-3.0-NL', 'deprecated': False},
+    'cc-by-3.0-us': {'id': 'CC-BY-3.0-US', 'deprecated': False},
+    'cc-by-4.0': {'id': 'CC-BY-4.0', 'deprecated': False},
+    'cc-by-nc-1.0': {'id': 'CC-BY-NC-1.0', 'deprecated': False},
+    'cc-by-nc-2.0': {'id': 'CC-BY-NC-2.0', 'deprecated': False},
+    'cc-by-nc-2.5': {'id': 'CC-BY-NC-2.5', 'deprecated': False},
+    'cc-by-nc-3.0': {'id': 'CC-BY-NC-3.0', 'deprecated': False},
+    'cc-by-nc-3.0-de': {'id': 'CC-BY-NC-3.0-DE', 'deprecated': False},
+    'cc-by-nc-4.0': {'id': 'CC-BY-NC-4.0', 'deprecated': False},
+    'cc-by-nc-nd-1.0': {'id': 'CC-BY-NC-ND-1.0', 'deprecated': False},
+    'cc-by-nc-nd-2.0': {'id': 'CC-BY-NC-ND-2.0', 'deprecated': False},
+    'cc-by-nc-nd-2.5': {'id': 'CC-BY-NC-ND-2.5', 'deprecated': False},
+    'cc-by-nc-nd-3.0': {'id': 'CC-BY-NC-ND-3.0', 'deprecated': False},
+    'cc-by-nc-nd-3.0-de': {'id': 'CC-BY-NC-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nc-nd-3.0-igo': {'id': 'CC-BY-NC-ND-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-nd-4.0': {'id': 'CC-BY-NC-ND-4.0', 'deprecated': False},
+    'cc-by-nc-sa-1.0': {'id': 'CC-BY-NC-SA-1.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0': {'id': 'CC-BY-NC-SA-2.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0-de': {'id': 'CC-BY-NC-SA-2.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-2.0-fr': {'id': 'CC-BY-NC-SA-2.0-FR', 'deprecated': False},
+    'cc-by-nc-sa-2.0-uk': {'id': 'CC-BY-NC-SA-2.0-UK', 'deprecated': False},
+    'cc-by-nc-sa-2.5': {'id': 'CC-BY-NC-SA-2.5', 'deprecated': False},
+    'cc-by-nc-sa-3.0': {'id': 'CC-BY-NC-SA-3.0', 'deprecated': False},
+    'cc-by-nc-sa-3.0-de': {'id': 'CC-BY-NC-SA-3.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-3.0-igo': {'id': 'CC-BY-NC-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-sa-4.0': {'id': 'CC-BY-NC-SA-4.0', 'deprecated': False},
+    'cc-by-nd-1.0': {'id': 'CC-BY-ND-1.0', 'deprecated': False},
+    'cc-by-nd-2.0': {'id': 'CC-BY-ND-2.0', 'deprecated': False},
+    'cc-by-nd-2.5': {'id': 'CC-BY-ND-2.5', 'deprecated': False},
+    'cc-by-nd-3.0': {'id': 'CC-BY-ND-3.0', 'deprecated': False},
+    'cc-by-nd-3.0-de': {'id': 'CC-BY-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nd-4.0': {'id': 'CC-BY-ND-4.0', 'deprecated': False},
+    'cc-by-sa-1.0': {'id': 'CC-BY-SA-1.0', 'deprecated': False},
+    'cc-by-sa-2.0': {'id': 'CC-BY-SA-2.0', 'deprecated': False},
+    'cc-by-sa-2.0-uk': {'id': 'CC-BY-SA-2.0-UK', 'deprecated': False},
+    'cc-by-sa-2.1-jp': {'id': 'CC-BY-SA-2.1-JP', 'deprecated': False},
+    'cc-by-sa-2.5': {'id': 'CC-BY-SA-2.5', 'deprecated': False},
+    'cc-by-sa-3.0': {'id': 'CC-BY-SA-3.0', 'deprecated': False},
+    'cc-by-sa-3.0-at': {'id': 'CC-BY-SA-3.0-AT', 'deprecated': False},
+    'cc-by-sa-3.0-de': {'id': 'CC-BY-SA-3.0-DE', 'deprecated': False},
+    'cc-by-sa-3.0-igo': {'id': 'CC-BY-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-sa-4.0': {'id': 'CC-BY-SA-4.0', 'deprecated': False},
+    'cc-pddc': {'id': 'CC-PDDC', 'deprecated': False},
+    'cc0-1.0': {'id': 'CC0-1.0', 'deprecated': False},
+    'cddl-1.0': {'id': 'CDDL-1.0', 'deprecated': False},
+    'cddl-1.1': {'id': 'CDDL-1.1', 'deprecated': False},
+    'cdl-1.0': {'id': 'CDL-1.0', 'deprecated': False},
+    'cdla-permissive-1.0': {'id': 'CDLA-Permissive-1.0', 'deprecated': False},
+    'cdla-permissive-2.0': {'id': 'CDLA-Permissive-2.0', 'deprecated': False},
+    'cdla-sharing-1.0': {'id': 'CDLA-Sharing-1.0', 'deprecated': False},
+    'cecill-1.0': {'id': 'CECILL-1.0', 'deprecated': False},
+    'cecill-1.1': {'id': 'CECILL-1.1', 'deprecated': False},
+    'cecill-2.0': {'id': 'CECILL-2.0', 'deprecated': False},
+    'cecill-2.1': {'id': 'CECILL-2.1', 'deprecated': False},
+    'cecill-b': {'id': 'CECILL-B', 'deprecated': False},
+    'cecill-c': {'id': 'CECILL-C', 'deprecated': False},
+    'cern-ohl-1.1': {'id': 'CERN-OHL-1.1', 'deprecated': False},
+    'cern-ohl-1.2': {'id': 'CERN-OHL-1.2', 'deprecated': False},
+    'cern-ohl-p-2.0': {'id': 'CERN-OHL-P-2.0', 'deprecated': False},
+    'cern-ohl-s-2.0': {'id': 'CERN-OHL-S-2.0', 'deprecated': False},
+    'cern-ohl-w-2.0': {'id': 'CERN-OHL-W-2.0', 'deprecated': False},
+    'cfitsio': {'id': 'CFITSIO', 'deprecated': False},
+    'check-cvs': {'id': 'check-cvs', 'deprecated': False},
+    'checkmk': {'id': 'checkmk', 'deprecated': False},
+    'clartistic': {'id': 'ClArtistic', 'deprecated': False},
+    'clips': {'id': 'Clips', 'deprecated': False},
+    'cmu-mach': {'id': 'CMU-Mach', 'deprecated': False},
+    'cmu-mach-nodoc': {'id': 'CMU-Mach-nodoc', 'deprecated': False},
+    'cnri-jython': {'id': 'CNRI-Jython', 'deprecated': False},
+    'cnri-python': {'id': 'CNRI-Python', 'deprecated': False},
+    'cnri-python-gpl-compatible': {'id': 'CNRI-Python-GPL-Compatible', 'deprecated': False},
+    'coil-1.0': {'id': 'COIL-1.0', 'deprecated': False},
+    'community-spec-1.0': {'id': 'Community-Spec-1.0', 'deprecated': False},
+    'condor-1.1': {'id': 'Condor-1.1', 'deprecated': False},
+    'copyleft-next-0.3.0': {'id': 'copyleft-next-0.3.0', 'deprecated': False},
+    'copyleft-next-0.3.1': {'id': 'copyleft-next-0.3.1', 'deprecated': False},
+    'cornell-lossless-jpeg': {'id': 'Cornell-Lossless-JPEG', 'deprecated': False},
+    'cpal-1.0': {'id': 'CPAL-1.0', 'deprecated': False},
+    'cpl-1.0': {'id': 'CPL-1.0', 'deprecated': False},
+    'cpol-1.02': {'id': 'CPOL-1.02', 'deprecated': False},
+    'cronyx': {'id': 'Cronyx', 'deprecated': False},
+    'crossword': {'id': 'Crossword', 'deprecated': False},
+    'crystalstacker': {'id': 'CrystalStacker', 'deprecated': False},
+    'cua-opl-1.0': {'id': 'CUA-OPL-1.0', 'deprecated': False},
+    'cube': {'id': 'Cube', 'deprecated': False},
+    'curl': {'id': 'curl', 'deprecated': False},
+    'cve-tou': {'id': 'cve-tou', 'deprecated': False},
+    'd-fsl-1.0': {'id': 'D-FSL-1.0', 'deprecated': False},
+    'dec-3-clause': {'id': 'DEC-3-Clause', 'deprecated': False},
+    'diffmark': {'id': 'diffmark', 'deprecated': False},
+    'dl-de-by-2.0': {'id': 'DL-DE-BY-2.0', 'deprecated': False},
+    'dl-de-zero-2.0': {'id': 'DL-DE-ZERO-2.0', 'deprecated': False},
+    'doc': {'id': 'DOC', 'deprecated': False},
+    'docbook-schema': {'id': 'DocBook-Schema', 'deprecated': False},
+    'docbook-xml': {'id': 'DocBook-XML', 'deprecated': False},
+    'dotseqn': {'id': 'Dotseqn', 'deprecated': False},
+    'drl-1.0': {'id': 'DRL-1.0', 'deprecated': False},
+    'drl-1.1': {'id': 'DRL-1.1', 'deprecated': False},
+    'dsdp': {'id': 'DSDP', 'deprecated': False},
+    'dtoa': {'id': 'dtoa', 'deprecated': False},
+    'dvipdfm': {'id': 'dvipdfm', 'deprecated': False},
+    'ecl-1.0': {'id': 'ECL-1.0', 'deprecated': False},
+    'ecl-2.0': {'id': 'ECL-2.0', 'deprecated': False},
+    'ecos-2.0': {'id': 'eCos-2.0', 'deprecated': True},
+    'efl-1.0': {'id': 'EFL-1.0', 'deprecated': False},
+    'efl-2.0': {'id': 'EFL-2.0', 'deprecated': False},
+    'egenix': {'id': 'eGenix', 'deprecated': False},
+    'elastic-2.0': {'id': 'Elastic-2.0', 'deprecated': False},
+    'entessa': {'id': 'Entessa', 'deprecated': False},
+    'epics': {'id': 'EPICS', 'deprecated': False},
+    'epl-1.0': {'id': 'EPL-1.0', 'deprecated': False},
+    'epl-2.0': {'id': 'EPL-2.0', 'deprecated': False},
+    'erlpl-1.1': {'id': 'ErlPL-1.1', 'deprecated': False},
+    'etalab-2.0': {'id': 'etalab-2.0', 'deprecated': False},
+    'eudatagrid': {'id': 'EUDatagrid', 'deprecated': False},
+    'eupl-1.0': {'id': 'EUPL-1.0', 'deprecated': False},
+    'eupl-1.1': {'id': 'EUPL-1.1', 'deprecated': False},
+    'eupl-1.2': {'id': 'EUPL-1.2', 'deprecated': False},
+    'eurosym': {'id': 'Eurosym', 'deprecated': False},
+    'fair': {'id': 'Fair', 'deprecated': False},
+    'fbm': {'id': 'FBM', 'deprecated': False},
+    'fdk-aac': {'id': 'FDK-AAC', 'deprecated': False},
+    'ferguson-twofish': {'id': 'Ferguson-Twofish', 'deprecated': False},
+    'frameworx-1.0': {'id': 'Frameworx-1.0', 'deprecated': False},
+    'freebsd-doc': {'id': 'FreeBSD-DOC', 'deprecated': False},
+    'freeimage': {'id': 'FreeImage', 'deprecated': False},
+    'fsfap': {'id': 'FSFAP', 'deprecated': False},
+    'fsfap-no-warranty-disclaimer': {'id': 'FSFAP-no-warranty-disclaimer', 'deprecated': False},
+    'fsful': {'id': 'FSFUL', 'deprecated': False},
+    'fsfullr': {'id': 'FSFULLR', 'deprecated': False},
+    'fsfullrwd': {'id': 'FSFULLRWD', 'deprecated': False},
+    'ftl': {'id': 'FTL', 'deprecated': False},
+    'furuseth': {'id': 'Furuseth', 'deprecated': False},
+    'fwlw': {'id': 'fwlw', 'deprecated': False},
+    'gcr-docs': {'id': 'GCR-docs', 'deprecated': False},
+    'gd': {'id': 'GD', 'deprecated': False},
+    'gfdl-1.1': {'id': 'GFDL-1.1', 'deprecated': True},
+    'gfdl-1.1-invariants-only': {'id': 'GFDL-1.1-invariants-only', 'deprecated': False},
+    'gfdl-1.1-invariants-or-later': {'id': 'GFDL-1.1-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-no-invariants-only': {'id': 'GFDL-1.1-no-invariants-only', 'deprecated': False},
+    'gfdl-1.1-no-invariants-or-later': {'id': 'GFDL-1.1-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-only': {'id': 'GFDL-1.1-only', 'deprecated': False},
+    'gfdl-1.1-or-later': {'id': 'GFDL-1.1-or-later', 'deprecated': False},
+    'gfdl-1.2': {'id': 'GFDL-1.2', 'deprecated': True},
+    'gfdl-1.2-invariants-only': {'id': 'GFDL-1.2-invariants-only', 'deprecated': False},
+    'gfdl-1.2-invariants-or-later': {'id': 'GFDL-1.2-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-no-invariants-only': {'id': 'GFDL-1.2-no-invariants-only', 'deprecated': False},
+    'gfdl-1.2-no-invariants-or-later': {'id': 'GFDL-1.2-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-only': {'id': 'GFDL-1.2-only', 'deprecated': False},
+    'gfdl-1.2-or-later': {'id': 'GFDL-1.2-or-later', 'deprecated': False},
+    'gfdl-1.3': {'id': 'GFDL-1.3', 'deprecated': True},
+    'gfdl-1.3-invariants-only': {'id': 'GFDL-1.3-invariants-only', 'deprecated': False},
+    'gfdl-1.3-invariants-or-later': {'id': 'GFDL-1.3-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-no-invariants-only': {'id': 'GFDL-1.3-no-invariants-only', 'deprecated': False},
+    'gfdl-1.3-no-invariants-or-later': {'id': 'GFDL-1.3-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-only': {'id': 'GFDL-1.3-only', 'deprecated': False},
+    'gfdl-1.3-or-later': {'id': 'GFDL-1.3-or-later', 'deprecated': False},
+    'giftware': {'id': 'Giftware', 'deprecated': False},
+    'gl2ps': {'id': 'GL2PS', 'deprecated': False},
+    'glide': {'id': 'Glide', 'deprecated': False},
+    'glulxe': {'id': 'Glulxe', 'deprecated': False},
+    'glwtpl': {'id': 'GLWTPL', 'deprecated': False},
+    'gnuplot': {'id': 'gnuplot', 'deprecated': False},
+    'gpl-1.0': {'id': 'GPL-1.0', 'deprecated': True},
+    'gpl-1.0+': {'id': 'GPL-1.0+', 'deprecated': True},
+    'gpl-1.0-only': {'id': 'GPL-1.0-only', 'deprecated': False},
+    'gpl-1.0-or-later': {'id': 'GPL-1.0-or-later', 'deprecated': False},
+    'gpl-2.0': {'id': 'GPL-2.0', 'deprecated': True},
+    'gpl-2.0+': {'id': 'GPL-2.0+', 'deprecated': True},
+    'gpl-2.0-only': {'id': 'GPL-2.0-only', 'deprecated': False},
+    'gpl-2.0-or-later': {'id': 'GPL-2.0-or-later', 'deprecated': False},
+    'gpl-2.0-with-autoconf-exception': {'id': 'GPL-2.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-2.0-with-bison-exception': {'id': 'GPL-2.0-with-bison-exception', 'deprecated': True},
+    'gpl-2.0-with-classpath-exception': {'id': 'GPL-2.0-with-classpath-exception', 'deprecated': True},
+    'gpl-2.0-with-font-exception': {'id': 'GPL-2.0-with-font-exception', 'deprecated': True},
+    'gpl-2.0-with-gcc-exception': {'id': 'GPL-2.0-with-GCC-exception', 'deprecated': True},
+    'gpl-3.0': {'id': 'GPL-3.0', 'deprecated': True},
+    'gpl-3.0+': {'id': 'GPL-3.0+', 'deprecated': True},
+    'gpl-3.0-only': {'id': 'GPL-3.0-only', 'deprecated': False},
+    'gpl-3.0-or-later': {'id': 'GPL-3.0-or-later', 'deprecated': False},
+    'gpl-3.0-with-autoconf-exception': {'id': 'GPL-3.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-3.0-with-gcc-exception': {'id': 'GPL-3.0-with-GCC-exception', 'deprecated': True},
+    'graphics-gems': {'id': 'Graphics-Gems', 'deprecated': False},
+    'gsoap-1.3b': {'id': 'gSOAP-1.3b', 'deprecated': False},
+    'gtkbook': {'id': 'gtkbook', 'deprecated': False},
+    'gutmann': {'id': 'Gutmann', 'deprecated': False},
+    'haskellreport': {'id': 'HaskellReport', 'deprecated': False},
+    'hdparm': {'id': 'hdparm', 'deprecated': False},
+    'hidapi': {'id': 'HIDAPI', 'deprecated': False},
+    'hippocratic-2.1': {'id': 'Hippocratic-2.1', 'deprecated': False},
+    'hp-1986': {'id': 'HP-1986', 'deprecated': False},
+    'hp-1989': {'id': 'HP-1989', 'deprecated': False},
+    'hpnd': {'id': 'HPND', 'deprecated': False},
+    'hpnd-dec': {'id': 'HPND-DEC', 'deprecated': False},
+    'hpnd-doc': {'id': 'HPND-doc', 'deprecated': False},
+    'hpnd-doc-sell': {'id': 'HPND-doc-sell', 'deprecated': False},
+    'hpnd-export-us': {'id': 'HPND-export-US', 'deprecated': False},
+    'hpnd-export-us-acknowledgement': {'id': 'HPND-export-US-acknowledgement', 'deprecated': False},
+    'hpnd-export-us-modify': {'id': 'HPND-export-US-modify', 'deprecated': False},
+    'hpnd-export2-us': {'id': 'HPND-export2-US', 'deprecated': False},
+    'hpnd-fenneberg-livingston': {'id': 'HPND-Fenneberg-Livingston', 'deprecated': False},
+    'hpnd-inria-imag': {'id': 'HPND-INRIA-IMAG', 'deprecated': False},
+    'hpnd-intel': {'id': 'HPND-Intel', 'deprecated': False},
+    'hpnd-kevlin-henney': {'id': 'HPND-Kevlin-Henney', 'deprecated': False},
+    'hpnd-markus-kuhn': {'id': 'HPND-Markus-Kuhn', 'deprecated': False},
+    'hpnd-merchantability-variant': {'id': 'HPND-merchantability-variant', 'deprecated': False},
+    'hpnd-mit-disclaimer': {'id': 'HPND-MIT-disclaimer', 'deprecated': False},
+    'hpnd-netrek': {'id': 'HPND-Netrek', 'deprecated': False},
+    'hpnd-pbmplus': {'id': 'HPND-Pbmplus', 'deprecated': False},
+    'hpnd-sell-mit-disclaimer-xserver': {'id': 'HPND-sell-MIT-disclaimer-xserver', 'deprecated': False},
+    'hpnd-sell-regexpr': {'id': 'HPND-sell-regexpr', 'deprecated': False},
+    'hpnd-sell-variant': {'id': 'HPND-sell-variant', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer': {'id': 'HPND-sell-variant-MIT-disclaimer', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer-rev': {'id': 'HPND-sell-variant-MIT-disclaimer-rev', 'deprecated': False},
+    'hpnd-uc': {'id': 'HPND-UC', 'deprecated': False},
+    'hpnd-uc-export-us': {'id': 'HPND-UC-export-US', 'deprecated': False},
+    'htmltidy': {'id': 'HTMLTIDY', 'deprecated': False},
+    'ibm-pibs': {'id': 'IBM-pibs', 'deprecated': False},
+    'icu': {'id': 'ICU', 'deprecated': False},
+    'iec-code-components-eula': {'id': 'IEC-Code-Components-EULA', 'deprecated': False},
+    'ijg': {'id': 'IJG', 'deprecated': False},
+    'ijg-short': {'id': 'IJG-short', 'deprecated': False},
+    'imagemagick': {'id': 'ImageMagick', 'deprecated': False},
+    'imatix': {'id': 'iMatix', 'deprecated': False},
+    'imlib2': {'id': 'Imlib2', 'deprecated': False},
+    'info-zip': {'id': 'Info-ZIP', 'deprecated': False},
+    'inner-net-2.0': {'id': 'Inner-Net-2.0', 'deprecated': False},
+    'intel': {'id': 'Intel', 'deprecated': False},
+    'intel-acpi': {'id': 'Intel-ACPI', 'deprecated': False},
+    'interbase-1.0': {'id': 'Interbase-1.0', 'deprecated': False},
+    'ipa': {'id': 'IPA', 'deprecated': False},
+    'ipl-1.0': {'id': 'IPL-1.0', 'deprecated': False},
+    'isc': {'id': 'ISC', 'deprecated': False},
+    'isc-veillard': {'id': 'ISC-Veillard', 'deprecated': False},
+    'jam': {'id': 'Jam', 'deprecated': False},
+    'jasper-2.0': {'id': 'JasPer-2.0', 'deprecated': False},
+    'jpl-image': {'id': 'JPL-image', 'deprecated': False},
+    'jpnic': {'id': 'JPNIC', 'deprecated': False},
+    'json': {'id': 'JSON', 'deprecated': False},
+    'kastrup': {'id': 'Kastrup', 'deprecated': False},
+    'kazlib': {'id': 'Kazlib', 'deprecated': False},
+    'knuth-ctan': {'id': 'Knuth-CTAN', 'deprecated': False},
+    'lal-1.2': {'id': 'LAL-1.2', 'deprecated': False},
+    'lal-1.3': {'id': 'LAL-1.3', 'deprecated': False},
+    'latex2e': {'id': 'Latex2e', 'deprecated': False},
+    'latex2e-translated-notice': {'id': 'Latex2e-translated-notice', 'deprecated': False},
+    'leptonica': {'id': 'Leptonica', 'deprecated': False},
+    'lgpl-2.0': {'id': 'LGPL-2.0', 'deprecated': True},
+    'lgpl-2.0+': {'id': 'LGPL-2.0+', 'deprecated': True},
+    'lgpl-2.0-only': {'id': 'LGPL-2.0-only', 'deprecated': False},
+    'lgpl-2.0-or-later': {'id': 'LGPL-2.0-or-later', 'deprecated': False},
+    'lgpl-2.1': {'id': 'LGPL-2.1', 'deprecated': True},
+    'lgpl-2.1+': {'id': 'LGPL-2.1+', 'deprecated': True},
+    'lgpl-2.1-only': {'id': 'LGPL-2.1-only', 'deprecated': False},
+    'lgpl-2.1-or-later': {'id': 'LGPL-2.1-or-later', 'deprecated': False},
+    'lgpl-3.0': {'id': 'LGPL-3.0', 'deprecated': True},
+    'lgpl-3.0+': {'id': 'LGPL-3.0+', 'deprecated': True},
+    'lgpl-3.0-only': {'id': 'LGPL-3.0-only', 'deprecated': False},
+    'lgpl-3.0-or-later': {'id': 'LGPL-3.0-or-later', 'deprecated': False},
+    'lgpllr': {'id': 'LGPLLR', 'deprecated': False},
+    'libpng': {'id': 'Libpng', 'deprecated': False},
+    'libpng-2.0': {'id': 'libpng-2.0', 'deprecated': False},
+    'libselinux-1.0': {'id': 'libselinux-1.0', 'deprecated': False},
+    'libtiff': {'id': 'libtiff', 'deprecated': False},
+    'libutil-david-nugent': {'id': 'libutil-David-Nugent', 'deprecated': False},
+    'liliq-p-1.1': {'id': 'LiLiQ-P-1.1', 'deprecated': False},
+    'liliq-r-1.1': {'id': 'LiLiQ-R-1.1', 'deprecated': False},
+    'liliq-rplus-1.1': {'id': 'LiLiQ-Rplus-1.1', 'deprecated': False},
+    'linux-man-pages-1-para': {'id': 'Linux-man-pages-1-para', 'deprecated': False},
+    'linux-man-pages-copyleft': {'id': 'Linux-man-pages-copyleft', 'deprecated': False},
+    'linux-man-pages-copyleft-2-para': {'id': 'Linux-man-pages-copyleft-2-para', 'deprecated': False},
+    'linux-man-pages-copyleft-var': {'id': 'Linux-man-pages-copyleft-var', 'deprecated': False},
+    'linux-openib': {'id': 'Linux-OpenIB', 'deprecated': False},
+    'loop': {'id': 'LOOP', 'deprecated': False},
+    'lpd-document': {'id': 'LPD-document', 'deprecated': False},
+    'lpl-1.0': {'id': 'LPL-1.0', 'deprecated': False},
+    'lpl-1.02': {'id': 'LPL-1.02', 'deprecated': False},
+    'lppl-1.0': {'id': 'LPPL-1.0', 'deprecated': False},
+    'lppl-1.1': {'id': 'LPPL-1.1', 'deprecated': False},
+    'lppl-1.2': {'id': 'LPPL-1.2', 'deprecated': False},
+    'lppl-1.3a': {'id': 'LPPL-1.3a', 'deprecated': False},
+    'lppl-1.3c': {'id': 'LPPL-1.3c', 'deprecated': False},
+    'lsof': {'id': 'lsof', 'deprecated': False},
+    'lucida-bitmap-fonts': {'id': 'Lucida-Bitmap-Fonts', 'deprecated': False},
+    'lzma-sdk-9.11-to-9.20': {'id': 'LZMA-SDK-9.11-to-9.20', 'deprecated': False},
+    'lzma-sdk-9.22': {'id': 'LZMA-SDK-9.22', 'deprecated': False},
+    'mackerras-3-clause': {'id': 'Mackerras-3-Clause', 'deprecated': False},
+    'mackerras-3-clause-acknowledgment': {'id': 'Mackerras-3-Clause-acknowledgment', 'deprecated': False},
+    'magaz': {'id': 'magaz', 'deprecated': False},
+    'mailprio': {'id': 'mailprio', 'deprecated': False},
+    'makeindex': {'id': 'MakeIndex', 'deprecated': False},
+    'martin-birgmeier': {'id': 'Martin-Birgmeier', 'deprecated': False},
+    'mcphee-slideshow': {'id': 'McPhee-slideshow', 'deprecated': False},
+    'metamail': {'id': 'metamail', 'deprecated': False},
+    'minpack': {'id': 'Minpack', 'deprecated': False},
+    'miros': {'id': 'MirOS', 'deprecated': False},
+    'mit': {'id': 'MIT', 'deprecated': False},
+    'mit-0': {'id': 'MIT-0', 'deprecated': False},
+    'mit-advertising': {'id': 'MIT-advertising', 'deprecated': False},
+    'mit-cmu': {'id': 'MIT-CMU', 'deprecated': False},
+    'mit-enna': {'id': 'MIT-enna', 'deprecated': False},
+    'mit-feh': {'id': 'MIT-feh', 'deprecated': False},
+    'mit-festival': {'id': 'MIT-Festival', 'deprecated': False},
+    'mit-khronos-old': {'id': 'MIT-Khronos-old', 'deprecated': False},
+    'mit-modern-variant': {'id': 'MIT-Modern-Variant', 'deprecated': False},
+    'mit-open-group': {'id': 'MIT-open-group', 'deprecated': False},
+    'mit-testregex': {'id': 'MIT-testregex', 'deprecated': False},
+    'mit-wu': {'id': 'MIT-Wu', 'deprecated': False},
+    'mitnfa': {'id': 'MITNFA', 'deprecated': False},
+    'mmixware': {'id': 'MMIXware', 'deprecated': False},
+    'motosoto': {'id': 'Motosoto', 'deprecated': False},
+    'mpeg-ssg': {'id': 'MPEG-SSG', 'deprecated': False},
+    'mpi-permissive': {'id': 'mpi-permissive', 'deprecated': False},
+    'mpich2': {'id': 'mpich2', 'deprecated': False},
+    'mpl-1.0': {'id': 'MPL-1.0', 'deprecated': False},
+    'mpl-1.1': {'id': 'MPL-1.1', 'deprecated': False},
+    'mpl-2.0': {'id': 'MPL-2.0', 'deprecated': False},
+    'mpl-2.0-no-copyleft-exception': {'id': 'MPL-2.0-no-copyleft-exception', 'deprecated': False},
+    'mplus': {'id': 'mplus', 'deprecated': False},
+    'ms-lpl': {'id': 'MS-LPL', 'deprecated': False},
+    'ms-pl': {'id': 'MS-PL', 'deprecated': False},
+    'ms-rl': {'id': 'MS-RL', 'deprecated': False},
+    'mtll': {'id': 'MTLL', 'deprecated': False},
+    'mulanpsl-1.0': {'id': 'MulanPSL-1.0', 'deprecated': False},
+    'mulanpsl-2.0': {'id': 'MulanPSL-2.0', 'deprecated': False},
+    'multics': {'id': 'Multics', 'deprecated': False},
+    'mup': {'id': 'Mup', 'deprecated': False},
+    'naist-2003': {'id': 'NAIST-2003', 'deprecated': False},
+    'nasa-1.3': {'id': 'NASA-1.3', 'deprecated': False},
+    'naumen': {'id': 'Naumen', 'deprecated': False},
+    'nbpl-1.0': {'id': 'NBPL-1.0', 'deprecated': False},
+    'ncbi-pd': {'id': 'NCBI-PD', 'deprecated': False},
+    'ncgl-uk-2.0': {'id': 'NCGL-UK-2.0', 'deprecated': False},
+    'ncl': {'id': 'NCL', 'deprecated': False},
+    'ncsa': {'id': 'NCSA', 'deprecated': False},
+    'net-snmp': {'id': 'Net-SNMP', 'deprecated': True},
+    'netcdf': {'id': 'NetCDF', 'deprecated': False},
+    'newsletr': {'id': 'Newsletr', 'deprecated': False},
+    'ngpl': {'id': 'NGPL', 'deprecated': False},
+    'nicta-1.0': {'id': 'NICTA-1.0', 'deprecated': False},
+    'nist-pd': {'id': 'NIST-PD', 'deprecated': False},
+    'nist-pd-fallback': {'id': 'NIST-PD-fallback', 'deprecated': False},
+    'nist-software': {'id': 'NIST-Software', 'deprecated': False},
+    'nlod-1.0': {'id': 'NLOD-1.0', 'deprecated': False},
+    'nlod-2.0': {'id': 'NLOD-2.0', 'deprecated': False},
+    'nlpl': {'id': 'NLPL', 'deprecated': False},
+    'nokia': {'id': 'Nokia', 'deprecated': False},
+    'nosl': {'id': 'NOSL', 'deprecated': False},
+    'noweb': {'id': 'Noweb', 'deprecated': False},
+    'npl-1.0': {'id': 'NPL-1.0', 'deprecated': False},
+    'npl-1.1': {'id': 'NPL-1.1', 'deprecated': False},
+    'nposl-3.0': {'id': 'NPOSL-3.0', 'deprecated': False},
+    'nrl': {'id': 'NRL', 'deprecated': False},
+    'ntp': {'id': 'NTP', 'deprecated': False},
+    'ntp-0': {'id': 'NTP-0', 'deprecated': False},
+    'nunit': {'id': 'Nunit', 'deprecated': True},
+    'o-uda-1.0': {'id': 'O-UDA-1.0', 'deprecated': False},
+    'oar': {'id': 'OAR', 'deprecated': False},
+    'occt-pl': {'id': 'OCCT-PL', 'deprecated': False},
+    'oclc-2.0': {'id': 'OCLC-2.0', 'deprecated': False},
+    'odbl-1.0': {'id': 'ODbL-1.0', 'deprecated': False},
+    'odc-by-1.0': {'id': 'ODC-By-1.0', 'deprecated': False},
+    'offis': {'id': 'OFFIS', 'deprecated': False},
+    'ofl-1.0': {'id': 'OFL-1.0', 'deprecated': False},
+    'ofl-1.0-no-rfn': {'id': 'OFL-1.0-no-RFN', 'deprecated': False},
+    'ofl-1.0-rfn': {'id': 'OFL-1.0-RFN', 'deprecated': False},
+    'ofl-1.1': {'id': 'OFL-1.1', 'deprecated': False},
+    'ofl-1.1-no-rfn': {'id': 'OFL-1.1-no-RFN', 'deprecated': False},
+    'ofl-1.1-rfn': {'id': 'OFL-1.1-RFN', 'deprecated': False},
+    'ogc-1.0': {'id': 'OGC-1.0', 'deprecated': False},
+    'ogdl-taiwan-1.0': {'id': 'OGDL-Taiwan-1.0', 'deprecated': False},
+    'ogl-canada-2.0': {'id': 'OGL-Canada-2.0', 'deprecated': False},
+    'ogl-uk-1.0': {'id': 'OGL-UK-1.0', 'deprecated': False},
+    'ogl-uk-2.0': {'id': 'OGL-UK-2.0', 'deprecated': False},
+    'ogl-uk-3.0': {'id': 'OGL-UK-3.0', 'deprecated': False},
+    'ogtsl': {'id': 'OGTSL', 'deprecated': False},
+    'oldap-1.1': {'id': 'OLDAP-1.1', 'deprecated': False},
+    'oldap-1.2': {'id': 'OLDAP-1.2', 'deprecated': False},
+    'oldap-1.3': {'id': 'OLDAP-1.3', 'deprecated': False},
+    'oldap-1.4': {'id': 'OLDAP-1.4', 'deprecated': False},
+    'oldap-2.0': {'id': 'OLDAP-2.0', 'deprecated': False},
+    'oldap-2.0.1': {'id': 'OLDAP-2.0.1', 'deprecated': False},
+    'oldap-2.1': {'id': 'OLDAP-2.1', 'deprecated': False},
+    'oldap-2.2': {'id': 'OLDAP-2.2', 'deprecated': False},
+    'oldap-2.2.1': {'id': 'OLDAP-2.2.1', 'deprecated': False},
+    'oldap-2.2.2': {'id': 'OLDAP-2.2.2', 'deprecated': False},
+    'oldap-2.3': {'id': 'OLDAP-2.3', 'deprecated': False},
+    'oldap-2.4': {'id': 'OLDAP-2.4', 'deprecated': False},
+    'oldap-2.5': {'id': 'OLDAP-2.5', 'deprecated': False},
+    'oldap-2.6': {'id': 'OLDAP-2.6', 'deprecated': False},
+    'oldap-2.7': {'id': 'OLDAP-2.7', 'deprecated': False},
+    'oldap-2.8': {'id': 'OLDAP-2.8', 'deprecated': False},
+    'olfl-1.3': {'id': 'OLFL-1.3', 'deprecated': False},
+    'oml': {'id': 'OML', 'deprecated': False},
+    'openpbs-2.3': {'id': 'OpenPBS-2.3', 'deprecated': False},
+    'openssl': {'id': 'OpenSSL', 'deprecated': False},
+    'openssl-standalone': {'id': 'OpenSSL-standalone', 'deprecated': False},
+    'openvision': {'id': 'OpenVision', 'deprecated': False},
+    'opl-1.0': {'id': 'OPL-1.0', 'deprecated': False},
+    'opl-uk-3.0': {'id': 'OPL-UK-3.0', 'deprecated': False},
+    'opubl-1.0': {'id': 'OPUBL-1.0', 'deprecated': False},
+    'oset-pl-2.1': {'id': 'OSET-PL-2.1', 'deprecated': False},
+    'osl-1.0': {'id': 'OSL-1.0', 'deprecated': False},
+    'osl-1.1': {'id': 'OSL-1.1', 'deprecated': False},
+    'osl-2.0': {'id': 'OSL-2.0', 'deprecated': False},
+    'osl-2.1': {'id': 'OSL-2.1', 'deprecated': False},
+    'osl-3.0': {'id': 'OSL-3.0', 'deprecated': False},
+    'padl': {'id': 'PADL', 'deprecated': False},
+    'parity-6.0.0': {'id': 'Parity-6.0.0', 'deprecated': False},
+    'parity-7.0.0': {'id': 'Parity-7.0.0', 'deprecated': False},
+    'pddl-1.0': {'id': 'PDDL-1.0', 'deprecated': False},
+    'php-3.0': {'id': 'PHP-3.0', 'deprecated': False},
+    'php-3.01': {'id': 'PHP-3.01', 'deprecated': False},
+    'pixar': {'id': 'Pixar', 'deprecated': False},
+    'pkgconf': {'id': 'pkgconf', 'deprecated': False},
+    'plexus': {'id': 'Plexus', 'deprecated': False},
+    'pnmstitch': {'id': 'pnmstitch', 'deprecated': False},
+    'polyform-noncommercial-1.0.0': {'id': 'PolyForm-Noncommercial-1.0.0', 'deprecated': False},
+    'polyform-small-business-1.0.0': {'id': 'PolyForm-Small-Business-1.0.0', 'deprecated': False},
+    'postgresql': {'id': 'PostgreSQL', 'deprecated': False},
+    'ppl': {'id': 'PPL', 'deprecated': False},
+    'psf-2.0': {'id': 'PSF-2.0', 'deprecated': False},
+    'psfrag': {'id': 'psfrag', 'deprecated': False},
+    'psutils': {'id': 'psutils', 'deprecated': False},
+    'python-2.0': {'id': 'Python-2.0', 'deprecated': False},
+    'python-2.0.1': {'id': 'Python-2.0.1', 'deprecated': False},
+    'python-ldap': {'id': 'python-ldap', 'deprecated': False},
+    'qhull': {'id': 'Qhull', 'deprecated': False},
+    'qpl-1.0': {'id': 'QPL-1.0', 'deprecated': False},
+    'qpl-1.0-inria-2004': {'id': 'QPL-1.0-INRIA-2004', 'deprecated': False},
+    'radvd': {'id': 'radvd', 'deprecated': False},
+    'rdisc': {'id': 'Rdisc', 'deprecated': False},
+    'rhecos-1.1': {'id': 'RHeCos-1.1', 'deprecated': False},
+    'rpl-1.1': {'id': 'RPL-1.1', 'deprecated': False},
+    'rpl-1.5': {'id': 'RPL-1.5', 'deprecated': False},
+    'rpsl-1.0': {'id': 'RPSL-1.0', 'deprecated': False},
+    'rsa-md': {'id': 'RSA-MD', 'deprecated': False},
+    'rscpl': {'id': 'RSCPL', 'deprecated': False},
+    'ruby': {'id': 'Ruby', 'deprecated': False},
+    'ruby-pty': {'id': 'Ruby-pty', 'deprecated': False},
+    'sax-pd': {'id': 'SAX-PD', 'deprecated': False},
+    'sax-pd-2.0': {'id': 'SAX-PD-2.0', 'deprecated': False},
+    'saxpath': {'id': 'Saxpath', 'deprecated': False},
+    'scea': {'id': 'SCEA', 'deprecated': False},
+    'schemereport': {'id': 'SchemeReport', 'deprecated': False},
+    'sendmail': {'id': 'Sendmail', 'deprecated': False},
+    'sendmail-8.23': {'id': 'Sendmail-8.23', 'deprecated': False},
+    'sgi-b-1.0': {'id': 'SGI-B-1.0', 'deprecated': False},
+    'sgi-b-1.1': {'id': 'SGI-B-1.1', 'deprecated': False},
+    'sgi-b-2.0': {'id': 'SGI-B-2.0', 'deprecated': False},
+    'sgi-opengl': {'id': 'SGI-OpenGL', 'deprecated': False},
+    'sgp4': {'id': 'SGP4', 'deprecated': False},
+    'shl-0.5': {'id': 'SHL-0.5', 'deprecated': False},
+    'shl-0.51': {'id': 'SHL-0.51', 'deprecated': False},
+    'simpl-2.0': {'id': 'SimPL-2.0', 'deprecated': False},
+    'sissl': {'id': 'SISSL', 'deprecated': False},
+    'sissl-1.2': {'id': 'SISSL-1.2', 'deprecated': False},
+    'sl': {'id': 'SL', 'deprecated': False},
+    'sleepycat': {'id': 'Sleepycat', 'deprecated': False},
+    'smlnj': {'id': 'SMLNJ', 'deprecated': False},
+    'smppl': {'id': 'SMPPL', 'deprecated': False},
+    'snia': {'id': 'SNIA', 'deprecated': False},
+    'snprintf': {'id': 'snprintf', 'deprecated': False},
+    'softsurfer': {'id': 'softSurfer', 'deprecated': False},
+    'soundex': {'id': 'Soundex', 'deprecated': False},
+    'spencer-86': {'id': 'Spencer-86', 'deprecated': False},
+    'spencer-94': {'id': 'Spencer-94', 'deprecated': False},
+    'spencer-99': {'id': 'Spencer-99', 'deprecated': False},
+    'spl-1.0': {'id': 'SPL-1.0', 'deprecated': False},
+    'ssh-keyscan': {'id': 'ssh-keyscan', 'deprecated': False},
+    'ssh-openssh': {'id': 'SSH-OpenSSH', 'deprecated': False},
+    'ssh-short': {'id': 'SSH-short', 'deprecated': False},
+    'ssleay-standalone': {'id': 'SSLeay-standalone', 'deprecated': False},
+    'sspl-1.0': {'id': 'SSPL-1.0', 'deprecated': False},
+    'standardml-nj': {'id': 'StandardML-NJ', 'deprecated': True},
+    'sugarcrm-1.1.3': {'id': 'SugarCRM-1.1.3', 'deprecated': False},
+    'sun-ppp': {'id': 'Sun-PPP', 'deprecated': False},
+    'sun-ppp-2000': {'id': 'Sun-PPP-2000', 'deprecated': False},
+    'sunpro': {'id': 'SunPro', 'deprecated': False},
+    'swl': {'id': 'SWL', 'deprecated': False},
+    'swrule': {'id': 'swrule', 'deprecated': False},
+    'symlinks': {'id': 'Symlinks', 'deprecated': False},
+    'tapr-ohl-1.0': {'id': 'TAPR-OHL-1.0', 'deprecated': False},
+    'tcl': {'id': 'TCL', 'deprecated': False},
+    'tcp-wrappers': {'id': 'TCP-wrappers', 'deprecated': False},
+    'termreadkey': {'id': 'TermReadKey', 'deprecated': False},
+    'tgppl-1.0': {'id': 'TGPPL-1.0', 'deprecated': False},
+    'threeparttable': {'id': 'threeparttable', 'deprecated': False},
+    'tmate': {'id': 'TMate', 'deprecated': False},
+    'torque-1.1': {'id': 'TORQUE-1.1', 'deprecated': False},
+    'tosl': {'id': 'TOSL', 'deprecated': False},
+    'tpdl': {'id': 'TPDL', 'deprecated': False},
+    'tpl-1.0': {'id': 'TPL-1.0', 'deprecated': False},
+    'ttwl': {'id': 'TTWL', 'deprecated': False},
+    'ttyp0': {'id': 'TTYP0', 'deprecated': False},
+    'tu-berlin-1.0': {'id': 'TU-Berlin-1.0', 'deprecated': False},
+    'tu-berlin-2.0': {'id': 'TU-Berlin-2.0', 'deprecated': False},
+    'ubuntu-font-1.0': {'id': 'Ubuntu-font-1.0', 'deprecated': False},
+    'ucar': {'id': 'UCAR', 'deprecated': False},
+    'ucl-1.0': {'id': 'UCL-1.0', 'deprecated': False},
+    'ulem': {'id': 'ulem', 'deprecated': False},
+    'umich-merit': {'id': 'UMich-Merit', 'deprecated': False},
+    'unicode-3.0': {'id': 'Unicode-3.0', 'deprecated': False},
+    'unicode-dfs-2015': {'id': 'Unicode-DFS-2015', 'deprecated': False},
+    'unicode-dfs-2016': {'id': 'Unicode-DFS-2016', 'deprecated': False},
+    'unicode-tou': {'id': 'Unicode-TOU', 'deprecated': False},
+    'unixcrypt': {'id': 'UnixCrypt', 'deprecated': False},
+    'unlicense': {'id': 'Unlicense', 'deprecated': False},
+    'upl-1.0': {'id': 'UPL-1.0', 'deprecated': False},
+    'urt-rle': {'id': 'URT-RLE', 'deprecated': False},
+    'vim': {'id': 'Vim', 'deprecated': False},
+    'vostrom': {'id': 'VOSTROM', 'deprecated': False},
+    'vsl-1.0': {'id': 'VSL-1.0', 'deprecated': False},
+    'w3c': {'id': 'W3C', 'deprecated': False},
+    'w3c-19980720': {'id': 'W3C-19980720', 'deprecated': False},
+    'w3c-20150513': {'id': 'W3C-20150513', 'deprecated': False},
+    'w3m': {'id': 'w3m', 'deprecated': False},
+    'watcom-1.0': {'id': 'Watcom-1.0', 'deprecated': False},
+    'widget-workshop': {'id': 'Widget-Workshop', 'deprecated': False},
+    'wsuipa': {'id': 'Wsuipa', 'deprecated': False},
+    'wtfpl': {'id': 'WTFPL', 'deprecated': False},
+    'wxwindows': {'id': 'wxWindows', 'deprecated': True},
+    'x11': {'id': 'X11', 'deprecated': False},
+    'x11-distribute-modifications-variant': {'id': 'X11-distribute-modifications-variant', 'deprecated': False},
+    'x11-swapped': {'id': 'X11-swapped', 'deprecated': False},
+    'xdebug-1.03': {'id': 'Xdebug-1.03', 'deprecated': False},
+    'xerox': {'id': 'Xerox', 'deprecated': False},
+    'xfig': {'id': 'Xfig', 'deprecated': False},
+    'xfree86-1.1': {'id': 'XFree86-1.1', 'deprecated': False},
+    'xinetd': {'id': 'xinetd', 'deprecated': False},
+    'xkeyboard-config-zinoviev': {'id': 'xkeyboard-config-Zinoviev', 'deprecated': False},
+    'xlock': {'id': 'xlock', 'deprecated': False},
+    'xnet': {'id': 'Xnet', 'deprecated': False},
+    'xpp': {'id': 'xpp', 'deprecated': False},
+    'xskat': {'id': 'XSkat', 'deprecated': False},
+    'xzoom': {'id': 'xzoom', 'deprecated': False},
+    'ypl-1.0': {'id': 'YPL-1.0', 'deprecated': False},
+    'ypl-1.1': {'id': 'YPL-1.1', 'deprecated': False},
+    'zed': {'id': 'Zed', 'deprecated': False},
+    'zeeff': {'id': 'Zeeff', 'deprecated': False},
+    'zend-2.0': {'id': 'Zend-2.0', 'deprecated': False},
+    'zimbra-1.3': {'id': 'Zimbra-1.3', 'deprecated': False},
+    'zimbra-1.4': {'id': 'Zimbra-1.4', 'deprecated': False},
+    'zlib': {'id': 'Zlib', 'deprecated': False},
+    'zlib-acknowledgement': {'id': 'zlib-acknowledgement', 'deprecated': False},
+    'zpl-1.1': {'id': 'ZPL-1.1', 'deprecated': False},
+    'zpl-2.0': {'id': 'ZPL-2.0', 'deprecated': False},
+    'zpl-2.1': {'id': 'ZPL-2.1', 'deprecated': False},
+}
+
+EXCEPTIONS: dict[str, SPDXException] = {
+    '389-exception': {'id': '389-exception', 'deprecated': False},
+    'asterisk-exception': {'id': 'Asterisk-exception', 'deprecated': False},
+    'asterisk-linking-protocols-exception': {'id': 'Asterisk-linking-protocols-exception', 'deprecated': False},
+    'autoconf-exception-2.0': {'id': 'Autoconf-exception-2.0', 'deprecated': False},
+    'autoconf-exception-3.0': {'id': 'Autoconf-exception-3.0', 'deprecated': False},
+    'autoconf-exception-generic': {'id': 'Autoconf-exception-generic', 'deprecated': False},
+    'autoconf-exception-generic-3.0': {'id': 'Autoconf-exception-generic-3.0', 'deprecated': False},
+    'autoconf-exception-macro': {'id': 'Autoconf-exception-macro', 'deprecated': False},
+    'bison-exception-1.24': {'id': 'Bison-exception-1.24', 'deprecated': False},
+    'bison-exception-2.2': {'id': 'Bison-exception-2.2', 'deprecated': False},
+    'bootloader-exception': {'id': 'Bootloader-exception', 'deprecated': False},
+    'classpath-exception-2.0': {'id': 'Classpath-exception-2.0', 'deprecated': False},
+    'clisp-exception-2.0': {'id': 'CLISP-exception-2.0', 'deprecated': False},
+    'cryptsetup-openssl-exception': {'id': 'cryptsetup-OpenSSL-exception', 'deprecated': False},
+    'digirule-foss-exception': {'id': 'DigiRule-FOSS-exception', 'deprecated': False},
+    'ecos-exception-2.0': {'id': 'eCos-exception-2.0', 'deprecated': False},
+    'erlang-otp-linking-exception': {'id': 'erlang-otp-linking-exception', 'deprecated': False},
+    'fawkes-runtime-exception': {'id': 'Fawkes-Runtime-exception', 'deprecated': False},
+    'fltk-exception': {'id': 'FLTK-exception', 'deprecated': False},
+    'fmt-exception': {'id': 'fmt-exception', 'deprecated': False},
+    'font-exception-2.0': {'id': 'Font-exception-2.0', 'deprecated': False},
+    'freertos-exception-2.0': {'id': 'freertos-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0': {'id': 'GCC-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0-note': {'id': 'GCC-exception-2.0-note', 'deprecated': False},
+    'gcc-exception-3.1': {'id': 'GCC-exception-3.1', 'deprecated': False},
+    'gmsh-exception': {'id': 'Gmsh-exception', 'deprecated': False},
+    'gnat-exception': {'id': 'GNAT-exception', 'deprecated': False},
+    'gnome-examples-exception': {'id': 'GNOME-examples-exception', 'deprecated': False},
+    'gnu-compiler-exception': {'id': 'GNU-compiler-exception', 'deprecated': False},
+    'gnu-javamail-exception': {'id': 'gnu-javamail-exception', 'deprecated': False},
+    'gpl-3.0-interface-exception': {'id': 'GPL-3.0-interface-exception', 'deprecated': False},
+    'gpl-3.0-linking-exception': {'id': 'GPL-3.0-linking-exception', 'deprecated': False},
+    'gpl-3.0-linking-source-exception': {'id': 'GPL-3.0-linking-source-exception', 'deprecated': False},
+    'gpl-cc-1.0': {'id': 'GPL-CC-1.0', 'deprecated': False},
+    'gstreamer-exception-2005': {'id': 'GStreamer-exception-2005', 'deprecated': False},
+    'gstreamer-exception-2008': {'id': 'GStreamer-exception-2008', 'deprecated': False},
+    'i2p-gpl-java-exception': {'id': 'i2p-gpl-java-exception', 'deprecated': False},
+    'kicad-libraries-exception': {'id': 'KiCad-libraries-exception', 'deprecated': False},
+    'lgpl-3.0-linking-exception': {'id': 'LGPL-3.0-linking-exception', 'deprecated': False},
+    'libpri-openh323-exception': {'id': 'libpri-OpenH323-exception', 'deprecated': False},
+    'libtool-exception': {'id': 'Libtool-exception', 'deprecated': False},
+    'linux-syscall-note': {'id': 'Linux-syscall-note', 'deprecated': False},
+    'llgpl': {'id': 'LLGPL', 'deprecated': False},
+    'llvm-exception': {'id': 'LLVM-exception', 'deprecated': False},
+    'lzma-exception': {'id': 'LZMA-exception', 'deprecated': False},
+    'mif-exception': {'id': 'mif-exception', 'deprecated': False},
+    'nokia-qt-exception-1.1': {'id': 'Nokia-Qt-exception-1.1', 'deprecated': True},
+    'ocaml-lgpl-linking-exception': {'id': 'OCaml-LGPL-linking-exception', 'deprecated': False},
+    'occt-exception-1.0': {'id': 'OCCT-exception-1.0', 'deprecated': False},
+    'openjdk-assembly-exception-1.0': {'id': 'OpenJDK-assembly-exception-1.0', 'deprecated': False},
+    'openvpn-openssl-exception': {'id': 'openvpn-openssl-exception', 'deprecated': False},
+    'pcre2-exception': {'id': 'PCRE2-exception', 'deprecated': False},
+    'ps-or-pdf-font-exception-20170817': {'id': 'PS-or-PDF-font-exception-20170817', 'deprecated': False},
+    'qpl-1.0-inria-2004-exception': {'id': 'QPL-1.0-INRIA-2004-exception', 'deprecated': False},
+    'qt-gpl-exception-1.0': {'id': 'Qt-GPL-exception-1.0', 'deprecated': False},
+    'qt-lgpl-exception-1.1': {'id': 'Qt-LGPL-exception-1.1', 'deprecated': False},
+    'qwt-exception-1.0': {'id': 'Qwt-exception-1.0', 'deprecated': False},
+    'romic-exception': {'id': 'romic-exception', 'deprecated': False},
+    'rrdtool-floss-exception-2.0': {'id': 'RRDtool-FLOSS-exception-2.0', 'deprecated': False},
+    'sane-exception': {'id': 'SANE-exception', 'deprecated': False},
+    'shl-2.0': {'id': 'SHL-2.0', 'deprecated': False},
+    'shl-2.1': {'id': 'SHL-2.1', 'deprecated': False},
+    'stunnel-exception': {'id': 'stunnel-exception', 'deprecated': False},
+    'swi-exception': {'id': 'SWI-exception', 'deprecated': False},
+    'swift-exception': {'id': 'Swift-exception', 'deprecated': False},
+    'texinfo-exception': {'id': 'Texinfo-exception', 'deprecated': False},
+    'u-boot-exception-2.0': {'id': 'u-boot-exception-2.0', 'deprecated': False},
+    'ubdl-exception': {'id': 'UBDL-exception', 'deprecated': False},
+    'universal-foss-exception-1.0': {'id': 'Universal-FOSS-exception-1.0', 'deprecated': False},
+    'vsftpd-openssl-exception': {'id': 'vsftpd-openssl-exception', 'deprecated': False},
+    'wxwindows-exception-3.1': {'id': 'WxWindows-exception-3.1', 'deprecated': False},
+    'x11vnc-openssl-exception': {'id': 'x11vnc-openssl-exception', 'deprecated': False},
+}
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/markers.py b/backend/venv/lib/python3.12/site-packages/packaging/markers.py
new file mode 100644
index 00000000..fb7f49cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/markers.py
@@ -0,0 +1,331 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import operator
+import os
+import platform
+import sys
+from typing import Any, Callable, TypedDict, cast
+
+from ._parser import MarkerAtom, MarkerList, Op, Value, Variable
+from ._parser import parse_marker as _parse_marker
+from ._tokenizer import ParserSyntaxError
+from .specifiers import InvalidSpecifier, Specifier
+from .utils import canonicalize_name
+
+__all__ = [
+    "InvalidMarker",
+    "Marker",
+    "UndefinedComparison",
+    "UndefinedEnvironmentName",
+    "default_environment",
+]
+
+Operator = Callable[[str, str], bool]
+
+
+class InvalidMarker(ValueError):
+    """
+    An invalid marker was found, users should refer to PEP 508.
+    """
+
+
+class UndefinedComparison(ValueError):
+    """
+    An invalid operation was attempted on a value that doesn't support it.
+    """
+
+
+class UndefinedEnvironmentName(ValueError):
+    """
+    A name was attempted to be used that does not exist inside of the
+    environment.
+    """
+
+
+class Environment(TypedDict):
+    implementation_name: str
+    """The implementation's identifier, e.g. ``'cpython'``."""
+
+    implementation_version: str
+    """
+    The implementation's version, e.g. ``'3.13.0a2'`` for CPython 3.13.0a2, or
+    ``'7.3.13'`` for PyPy3.10 v7.3.13.
+    """
+
+    os_name: str
+    """
+    The value of :py:data:`os.name`. The name of the operating system dependent module
+    imported, e.g. ``'posix'``.
+    """
+
+    platform_machine: str
+    """
+    Returns the machine type, e.g. ``'i386'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_release: str
+    """
+    The system's release, e.g. ``'2.2.0'`` or ``'NT'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_system: str
+    """
+    The system/OS name, e.g. ``'Linux'``, ``'Windows'`` or ``'Java'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_version: str
+    """
+    The system's release version, e.g. ``'#3 on degas'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    python_full_version: str
+    """
+    The Python version as string ``'major.minor.patchlevel'``.
+
+    Note that unlike the Python :py:data:`sys.version`, this value will always include
+    the patchlevel (it defaults to 0).
+    """
+
+    platform_python_implementation: str
+    """
+    A string identifying the Python implementation, e.g. ``'CPython'``.
+    """
+
+    python_version: str
+    """The Python version as string ``'major.minor'``."""
+
+    sys_platform: str
+    """
+    This string contains a platform identifier that can be used to append
+    platform-specific components to :py:data:`sys.path`, for instance.
+
+    For Unix systems, except on Linux and AIX, this is the lowercased OS name as
+    returned by ``uname -s`` with the first part of the version as returned by
+    ``uname -r`` appended, e.g. ``'sunos5'`` or ``'freebsd8'``, at the time when Python
+    was built.
+    """
+
+
+def _normalize_extra_values(results: Any) -> Any:
+    """
+    Normalize extra values.
+    """
+    if isinstance(results[0], tuple):
+        lhs, op, rhs = results[0]
+        if isinstance(lhs, Variable) and lhs.value == "extra":
+            normalized_extra = canonicalize_name(rhs.value)
+            rhs = Value(normalized_extra)
+        elif isinstance(rhs, Variable) and rhs.value == "extra":
+            normalized_extra = canonicalize_name(lhs.value)
+            lhs = Value(normalized_extra)
+        results[0] = lhs, op, rhs
+    return results
+
+
+def _format_marker(
+    marker: list[str] | MarkerAtom | str, first: bool | None = True
+) -> str:
+    assert isinstance(marker, (list, tuple, str))
+
+    # Sometimes we have a structure like [[...]] which is a single item list
+    # where the single item is itself it's own list. In that case we want skip
+    # the rest of this function so that we don't get extraneous () on the
+    # outside.
+    if (
+        isinstance(marker, list)
+        and len(marker) == 1
+        and isinstance(marker[0], (list, tuple))
+    ):
+        return _format_marker(marker[0])
+
+    if isinstance(marker, list):
+        inner = (_format_marker(m, first=False) for m in marker)
+        if first:
+            return " ".join(inner)
+        else:
+            return "(" + " ".join(inner) + ")"
+    elif isinstance(marker, tuple):
+        return " ".join([m.serialize() for m in marker])
+    else:
+        return marker
+
+
+_operators: dict[str, Operator] = {
+    "in": lambda lhs, rhs: lhs in rhs,
+    "not in": lambda lhs, rhs: lhs not in rhs,
+    "<": operator.lt,
+    "<=": operator.le,
+    "==": operator.eq,
+    "!=": operator.ne,
+    ">=": operator.ge,
+    ">": operator.gt,
+}
+
+
+def _eval_op(lhs: str, op: Op, rhs: str) -> bool:
+    try:
+        spec = Specifier("".join([op.serialize(), rhs]))
+    except InvalidSpecifier:
+        pass
+    else:
+        return spec.contains(lhs, prereleases=True)
+
+    oper: Operator | None = _operators.get(op.serialize())
+    if oper is None:
+        raise UndefinedComparison(f"Undefined {op!r} on {lhs!r} and {rhs!r}.")
+
+    return oper(lhs, rhs)
+
+
+def _normalize(*values: str, key: str) -> tuple[str, ...]:
+    # PEP 685 – Comparison of extra names for optional distribution dependencies
+    # https://peps.python.org/pep-0685/
+    # > When comparing extra names, tools MUST normalize the names being
+    # > compared using the semantics outlined in PEP 503 for names
+    if key == "extra":
+        return tuple(canonicalize_name(v) for v in values)
+
+    # other environment markers don't have such standards
+    return values
+
+
+def _evaluate_markers(markers: MarkerList, environment: dict[str, str]) -> bool:
+    groups: list[list[bool]] = [[]]
+
+    for marker in markers:
+        assert isinstance(marker, (list, tuple, str))
+
+        if isinstance(marker, list):
+            groups[-1].append(_evaluate_markers(marker, environment))
+        elif isinstance(marker, tuple):
+            lhs, op, rhs = marker
+
+            if isinstance(lhs, Variable):
+                environment_key = lhs.value
+                lhs_value = environment[environment_key]
+                rhs_value = rhs.value
+            else:
+                lhs_value = lhs.value
+                environment_key = rhs.value
+                rhs_value = environment[environment_key]
+
+            lhs_value, rhs_value = _normalize(lhs_value, rhs_value, key=environment_key)
+            groups[-1].append(_eval_op(lhs_value, op, rhs_value))
+        else:
+            assert marker in ["and", "or"]
+            if marker == "or":
+                groups.append([])
+
+    return any(all(item) for item in groups)
+
+
+def format_full_version(info: sys._version_info) -> str:
+    version = f"{info.major}.{info.minor}.{info.micro}"
+    kind = info.releaselevel
+    if kind != "final":
+        version += kind[0] + str(info.serial)
+    return version
+
+
+def default_environment() -> Environment:
+    iver = format_full_version(sys.implementation.version)
+    implementation_name = sys.implementation.name
+    return {
+        "implementation_name": implementation_name,
+        "implementation_version": iver,
+        "os_name": os.name,
+        "platform_machine": platform.machine(),
+        "platform_release": platform.release(),
+        "platform_system": platform.system(),
+        "platform_version": platform.version(),
+        "python_full_version": platform.python_version(),
+        "platform_python_implementation": platform.python_implementation(),
+        "python_version": ".".join(platform.python_version_tuple()[:2]),
+        "sys_platform": sys.platform,
+    }
+
+
+class Marker:
+    def __init__(self, marker: str) -> None:
+        # Note: We create a Marker object without calling this constructor in
+        #       packaging.requirements.Requirement. If any additional logic is
+        #       added here, make sure to mirror/adapt Requirement.
+        try:
+            self._markers = _normalize_extra_values(_parse_marker(marker))
+            # The attribute `_markers` can be described in terms of a recursive type:
+            # MarkerList = List[Union[Tuple[Node, ...], str, MarkerList]]
+            #
+            # For example, the following expression:
+            # python_version > "3.6" or (python_version == "3.6" and os_name == "unix")
+            #
+            # is parsed into:
+            # [
+            #     (<Variable('python_version')>, <Op('>')>, <Value('3.6')>),
+            #     'and',
+            #     [
+            #         (<Variable('python_version')>, <Op('==')>, <Value('3.6')>),
+            #         'or',
+            #         (<Variable('os_name')>, <Op('==')>, <Value('unix')>)
+            #     ]
+            # ]
+        except ParserSyntaxError as e:
+            raise InvalidMarker(str(e)) from e
+
+    def __str__(self) -> str:
+        return _format_marker(self._markers)
+
+    def __repr__(self) -> str:
+        return f"<Marker('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash((self.__class__.__name__, str(self)))
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Marker):
+            return NotImplemented
+
+        return str(self) == str(other)
+
+    def evaluate(self, environment: dict[str, str] | None = None) -> bool:
+        """Evaluate a marker.
+
+        Return the boolean from evaluating the given marker against the
+        environment. environment is an optional argument to override all or
+        part of the determined environment.
+
+        The environment is determined from the current Python process.
+        """
+        current_environment = cast("dict[str, str]", default_environment())
+        current_environment["extra"] = ""
+        if environment is not None:
+            current_environment.update(environment)
+            # The API used to allow setting extra to None. We need to handle this
+            # case for backwards compatibility.
+            if current_environment["extra"] is None:
+                current_environment["extra"] = ""
+
+        return _evaluate_markers(
+            self._markers, _repair_python_full_version(current_environment)
+        )
+
+
+def _repair_python_full_version(env: dict[str, str]) -> dict[str, str]:
+    """
+    Work around platform.python_version() returning something that is not PEP 440
+    compliant for non-tagged Python builds.
+    """
+    if env["python_full_version"].endswith("+"):
+        env["python_full_version"] += "local"
+    return env
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/metadata.py b/backend/venv/lib/python3.12/site-packages/packaging/metadata.py
new file mode 100644
index 00000000..721f411c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/metadata.py
@@ -0,0 +1,863 @@
+from __future__ import annotations
+
+import email.feedparser
+import email.header
+import email.message
+import email.parser
+import email.policy
+import pathlib
+import sys
+import typing
+from typing import (
+    Any,
+    Callable,
+    Generic,
+    Literal,
+    TypedDict,
+    cast,
+)
+
+from . import licenses, requirements, specifiers, utils
+from . import version as version_module
+from .licenses import NormalizedLicenseExpression
+
+T = typing.TypeVar("T")
+
+
+if sys.version_info >= (3, 11):  # pragma: no cover
+    ExceptionGroup = ExceptionGroup
+else:  # pragma: no cover
+
+    class ExceptionGroup(Exception):
+        """A minimal implementation of :external:exc:`ExceptionGroup` from Python 3.11.
+
+        If :external:exc:`ExceptionGroup` is already defined by Python itself,
+        that version is used instead.
+        """
+
+        message: str
+        exceptions: list[Exception]
+
+        def __init__(self, message: str, exceptions: list[Exception]) -> None:
+            self.message = message
+            self.exceptions = exceptions
+
+        def __repr__(self) -> str:
+            return f"{self.__class__.__name__}({self.message!r}, {self.exceptions!r})"
+
+
+class InvalidMetadata(ValueError):
+    """A metadata field contains invalid data."""
+
+    field: str
+    """The name of the field that contains invalid data."""
+
+    def __init__(self, field: str, message: str) -> None:
+        self.field = field
+        super().__init__(message)
+
+
+# The RawMetadata class attempts to make as few assumptions about the underlying
+# serialization formats as possible. The idea is that as long as a serialization
+# formats offer some very basic primitives in *some* way then we can support
+# serializing to and from that format.
+class RawMetadata(TypedDict, total=False):
+    """A dictionary of raw core metadata.
+
+    Each field in core metadata maps to a key of this dictionary (when data is
+    provided). The key is lower-case and underscores are used instead of dashes
+    compared to the equivalent core metadata field. Any core metadata field that
+    can be specified multiple times or can hold multiple values in a single
+    field have a key with a plural name. See :class:`Metadata` whose attributes
+    match the keys of this dictionary.
+
+    Core metadata fields that can be specified multiple times are stored as a
+    list or dict depending on which is appropriate for the field. Any fields
+    which hold multiple values in a single field are stored as a list.
+
+    """
+
+    # Metadata 1.0 - PEP 241
+    metadata_version: str
+    name: str
+    version: str
+    platforms: list[str]
+    summary: str
+    description: str
+    keywords: list[str]
+    home_page: str
+    author: str
+    author_email: str
+    license: str
+
+    # Metadata 1.1 - PEP 314
+    supported_platforms: list[str]
+    download_url: str
+    classifiers: list[str]
+    requires: list[str]
+    provides: list[str]
+    obsoletes: list[str]
+
+    # Metadata 1.2 - PEP 345
+    maintainer: str
+    maintainer_email: str
+    requires_dist: list[str]
+    provides_dist: list[str]
+    obsoletes_dist: list[str]
+    requires_python: str
+    requires_external: list[str]
+    project_urls: dict[str, str]
+
+    # Metadata 2.0
+    # PEP 426 attempted to completely revamp the metadata format
+    # but got stuck without ever being able to build consensus on
+    # it and ultimately ended up withdrawn.
+    #
+    # However, a number of tools had started emitting METADATA with
+    # `2.0` Metadata-Version, so for historical reasons, this version
+    # was skipped.
+
+    # Metadata 2.1 - PEP 566
+    description_content_type: str
+    provides_extra: list[str]
+
+    # Metadata 2.2 - PEP 643
+    dynamic: list[str]
+
+    # Metadata 2.3 - PEP 685
+    # No new fields were added in PEP 685, just some edge case were
+    # tightened up to provide better interoptability.
+
+    # Metadata 2.4 - PEP 639
+    license_expression: str
+    license_files: list[str]
+
+
+_STRING_FIELDS = {
+    "author",
+    "author_email",
+    "description",
+    "description_content_type",
+    "download_url",
+    "home_page",
+    "license",
+    "license_expression",
+    "maintainer",
+    "maintainer_email",
+    "metadata_version",
+    "name",
+    "requires_python",
+    "summary",
+    "version",
+}
+
+_LIST_FIELDS = {
+    "classifiers",
+    "dynamic",
+    "license_files",
+    "obsoletes",
+    "obsoletes_dist",
+    "platforms",
+    "provides",
+    "provides_dist",
+    "provides_extra",
+    "requires",
+    "requires_dist",
+    "requires_external",
+    "supported_platforms",
+}
+
+_DICT_FIELDS = {
+    "project_urls",
+}
+
+
+def _parse_keywords(data: str) -> list[str]:
+    """Split a string of comma-separated keywords into a list of keywords."""
+    return [k.strip() for k in data.split(",")]
+
+
+def _parse_project_urls(data: list[str]) -> dict[str, str]:
+    """Parse a list of label/URL string pairings separated by a comma."""
+    urls = {}
+    for pair in data:
+        # Our logic is slightly tricky here as we want to try and do
+        # *something* reasonable with malformed data.
+        #
+        # The main thing that we have to worry about, is data that does
+        # not have a ',' at all to split the label from the Value. There
+        # isn't a singular right answer here, and we will fail validation
+        # later on (if the caller is validating) so it doesn't *really*
+        # matter, but since the missing value has to be an empty str
+        # and our return value is dict[str, str], if we let the key
+        # be the missing value, then they'd have multiple '' values that
+        # overwrite each other in a accumulating dict.
+        #
+        # The other potentional issue is that it's possible to have the
+        # same label multiple times in the metadata, with no solid "right"
+        # answer with what to do in that case. As such, we'll do the only
+        # thing we can, which is treat the field as unparseable and add it
+        # to our list of unparsed fields.
+        parts = [p.strip() for p in pair.split(",", 1)]
+        parts.extend([""] * (max(0, 2 - len(parts))))  # Ensure 2 items
+
+        # TODO: The spec doesn't say anything about if the keys should be
+        #       considered case sensitive or not... logically they should
+        #       be case-preserving and case-insensitive, but doing that
+        #       would open up more cases where we might have duplicate
+        #       entries.
+        label, url = parts
+        if label in urls:
+            # The label already exists in our set of urls, so this field
+            # is unparseable, and we can just add the whole thing to our
+            # unparseable data and stop processing it.
+            raise KeyError("duplicate labels in project urls")
+        urls[label] = url
+
+    return urls
+
+
+def _get_payload(msg: email.message.Message, source: bytes | str) -> str:
+    """Get the body of the message."""
+    # If our source is a str, then our caller has managed encodings for us,
+    # and we don't need to deal with it.
+    if isinstance(source, str):
+        payload = msg.get_payload()
+        assert isinstance(payload, str)
+        return payload
+    # If our source is a bytes, then we're managing the encoding and we need
+    # to deal with it.
+    else:
+        bpayload = msg.get_payload(decode=True)
+        assert isinstance(bpayload, bytes)
+        try:
+            return bpayload.decode("utf8", "strict")
+        except UnicodeDecodeError as exc:
+            raise ValueError("payload in an invalid encoding") from exc
+
+
+# The various parse_FORMAT functions here are intended to be as lenient as
+# possible in their parsing, while still returning a correctly typed
+# RawMetadata.
+#
+# To aid in this, we also generally want to do as little touching of the
+# data as possible, except where there are possibly some historic holdovers
+# that make valid data awkward to work with.
+#
+# While this is a lower level, intermediate format than our ``Metadata``
+# class, some light touch ups can make a massive difference in usability.
+
+# Map METADATA fields to RawMetadata.
+_EMAIL_TO_RAW_MAPPING = {
+    "author": "author",
+    "author-email": "author_email",
+    "classifier": "classifiers",
+    "description": "description",
+    "description-content-type": "description_content_type",
+    "download-url": "download_url",
+    "dynamic": "dynamic",
+    "home-page": "home_page",
+    "keywords": "keywords",
+    "license": "license",
+    "license-expression": "license_expression",
+    "license-file": "license_files",
+    "maintainer": "maintainer",
+    "maintainer-email": "maintainer_email",
+    "metadata-version": "metadata_version",
+    "name": "name",
+    "obsoletes": "obsoletes",
+    "obsoletes-dist": "obsoletes_dist",
+    "platform": "platforms",
+    "project-url": "project_urls",
+    "provides": "provides",
+    "provides-dist": "provides_dist",
+    "provides-extra": "provides_extra",
+    "requires": "requires",
+    "requires-dist": "requires_dist",
+    "requires-external": "requires_external",
+    "requires-python": "requires_python",
+    "summary": "summary",
+    "supported-platform": "supported_platforms",
+    "version": "version",
+}
+_RAW_TO_EMAIL_MAPPING = {raw: email for email, raw in _EMAIL_TO_RAW_MAPPING.items()}
+
+
+def parse_email(data: bytes | str) -> tuple[RawMetadata, dict[str, list[str]]]:
+    """Parse a distribution's metadata stored as email headers (e.g. from ``METADATA``).
+
+    This function returns a two-item tuple of dicts. The first dict is of
+    recognized fields from the core metadata specification. Fields that can be
+    parsed and translated into Python's built-in types are converted
+    appropriately. All other fields are left as-is. Fields that are allowed to
+    appear multiple times are stored as lists.
+
+    The second dict contains all other fields from the metadata. This includes
+    any unrecognized fields. It also includes any fields which are expected to
+    be parsed into a built-in type but were not formatted appropriately. Finally,
+    any fields that are expected to appear only once but are repeated are
+    included in this dict.
+
+    """
+    raw: dict[str, str | list[str] | dict[str, str]] = {}
+    unparsed: dict[str, list[str]] = {}
+
+    if isinstance(data, str):
+        parsed = email.parser.Parser(policy=email.policy.compat32).parsestr(data)
+    else:
+        parsed = email.parser.BytesParser(policy=email.policy.compat32).parsebytes(data)
+
+    # We have to wrap parsed.keys() in a set, because in the case of multiple
+    # values for a key (a list), the key will appear multiple times in the
+    # list of keys, but we're avoiding that by using get_all().
+    for name in frozenset(parsed.keys()):
+        # Header names in RFC are case insensitive, so we'll normalize to all
+        # lower case to make comparisons easier.
+        name = name.lower()
+
+        # We use get_all() here, even for fields that aren't multiple use,
+        # because otherwise someone could have e.g. two Name fields, and we
+        # would just silently ignore it rather than doing something about it.
+        headers = parsed.get_all(name) or []
+
+        # The way the email module works when parsing bytes is that it
+        # unconditionally decodes the bytes as ascii using the surrogateescape
+        # handler. When you pull that data back out (such as with get_all() ),
+        # it looks to see if the str has any surrogate escapes, and if it does
+        # it wraps it in a Header object instead of returning the string.
+        #
+        # As such, we'll look for those Header objects, and fix up the encoding.
+        value = []
+        # Flag if we have run into any issues processing the headers, thus
+        # signalling that the data belongs in 'unparsed'.
+        valid_encoding = True
+        for h in headers:
+            # It's unclear if this can return more types than just a Header or
+            # a str, so we'll just assert here to make sure.
+            assert isinstance(h, (email.header.Header, str))
+
+            # If it's a header object, we need to do our little dance to get
+            # the real data out of it. In cases where there is invalid data
+            # we're going to end up with mojibake, but there's no obvious, good
+            # way around that without reimplementing parts of the Header object
+            # ourselves.
+            #
+            # That should be fine since, if mojibacked happens, this key is
+            # going into the unparsed dict anyways.
+            if isinstance(h, email.header.Header):
+                # The Header object stores it's data as chunks, and each chunk
+                # can be independently encoded, so we'll need to check each
+                # of them.
+                chunks: list[tuple[bytes, str | None]] = []
+                for bin, encoding in email.header.decode_header(h):
+                    try:
+                        bin.decode("utf8", "strict")
+                    except UnicodeDecodeError:
+                        # Enable mojibake.
+                        encoding = "latin1"
+                        valid_encoding = False
+                    else:
+                        encoding = "utf8"
+                    chunks.append((bin, encoding))
+
+                # Turn our chunks back into a Header object, then let that
+                # Header object do the right thing to turn them into a
+                # string for us.
+                value.append(str(email.header.make_header(chunks)))
+            # This is already a string, so just add it.
+            else:
+                value.append(h)
+
+        # We've processed all of our values to get them into a list of str,
+        # but we may have mojibake data, in which case this is an unparsed
+        # field.
+        if not valid_encoding:
+            unparsed[name] = value
+            continue
+
+        raw_name = _EMAIL_TO_RAW_MAPPING.get(name)
+        if raw_name is None:
+            # This is a bit of a weird situation, we've encountered a key that
+            # we don't know what it means, so we don't know whether it's meant
+            # to be a list or not.
+            #
+            # Since we can't really tell one way or another, we'll just leave it
+            # as a list, even though it may be a single item list, because that's
+            # what makes the most sense for email headers.
+            unparsed[name] = value
+            continue
+
+        # If this is one of our string fields, then we'll check to see if our
+        # value is a list of a single item. If it is then we'll assume that
+        # it was emitted as a single string, and unwrap the str from inside
+        # the list.
+        #
+        # If it's any other kind of data, then we haven't the faintest clue
+        # what we should parse it as, and we have to just add it to our list
+        # of unparsed stuff.
+        if raw_name in _STRING_FIELDS and len(value) == 1:
+            raw[raw_name] = value[0]
+        # If this is one of our list of string fields, then we can just assign
+        # the value, since email *only* has strings, and our get_all() call
+        # above ensures that this is a list.
+        elif raw_name in _LIST_FIELDS:
+            raw[raw_name] = value
+        # Special Case: Keywords
+        # The keywords field is implemented in the metadata spec as a str,
+        # but it conceptually is a list of strings, and is serialized using
+        # ", ".join(keywords), so we'll do some light data massaging to turn
+        # this into what it logically is.
+        elif raw_name == "keywords" and len(value) == 1:
+            raw[raw_name] = _parse_keywords(value[0])
+        # Special Case: Project-URL
+        # The project urls is implemented in the metadata spec as a list of
+        # specially-formatted strings that represent a key and a value, which
+        # is fundamentally a mapping, however the email format doesn't support
+        # mappings in a sane way, so it was crammed into a list of strings
+        # instead.
+        #
+        # We will do a little light data massaging to turn this into a map as
+        # it logically should be.
+        elif raw_name == "project_urls":
+            try:
+                raw[raw_name] = _parse_project_urls(value)
+            except KeyError:
+                unparsed[name] = value
+        # Nothing that we've done has managed to parse this, so it'll just
+        # throw it in our unparseable data and move on.
+        else:
+            unparsed[name] = value
+
+    # We need to support getting the Description from the message payload in
+    # addition to getting it from the the headers. This does mean, though, there
+    # is the possibility of it being set both ways, in which case we put both
+    # in 'unparsed' since we don't know which is right.
+    try:
+        payload = _get_payload(parsed, data)
+    except ValueError:
+        unparsed.setdefault("description", []).append(
+            parsed.get_payload(decode=isinstance(data, bytes))  # type: ignore[call-overload]
+        )
+    else:
+        if payload:
+            # Check to see if we've already got a description, if so then both
+            # it, and this body move to unparseable.
+            if "description" in raw:
+                description_header = cast(str, raw.pop("description"))
+                unparsed.setdefault("description", []).extend(
+                    [description_header, payload]
+                )
+            elif "description" in unparsed:
+                unparsed["description"].append(payload)
+            else:
+                raw["description"] = payload
+
+    # We need to cast our `raw` to a metadata, because a TypedDict only support
+    # literal key names, but we're computing our key names on purpose, but the
+    # way this function is implemented, our `TypedDict` can only have valid key
+    # names.
+    return cast(RawMetadata, raw), unparsed
+
+
+_NOT_FOUND = object()
+
+
+# Keep the two values in sync.
+_VALID_METADATA_VERSIONS = ["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+_MetadataVersion = Literal["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+
+_REQUIRED_ATTRS = frozenset(["metadata_version", "name", "version"])
+
+
+class _Validator(Generic[T]):
+    """Validate a metadata field.
+
+    All _process_*() methods correspond to a core metadata field. The method is
+    called with the field's raw value. If the raw value is valid it is returned
+    in its "enriched" form (e.g. ``version.Version`` for the ``Version`` field).
+    If the raw value is invalid, :exc:`InvalidMetadata` is raised (with a cause
+    as appropriate).
+    """
+
+    name: str
+    raw_name: str
+    added: _MetadataVersion
+
+    def __init__(
+        self,
+        *,
+        added: _MetadataVersion = "1.0",
+    ) -> None:
+        self.added = added
+
+    def __set_name__(self, _owner: Metadata, name: str) -> None:
+        self.name = name
+        self.raw_name = _RAW_TO_EMAIL_MAPPING[name]
+
+    def __get__(self, instance: Metadata, _owner: type[Metadata]) -> T:
+        # With Python 3.8, the caching can be replaced with functools.cached_property().
+        # No need to check the cache as attribute lookup will resolve into the
+        # instance's __dict__ before __get__ is called.
+        cache = instance.__dict__
+        value = instance._raw.get(self.name)
+
+        # To make the _process_* methods easier, we'll check if the value is None
+        # and if this field is NOT a required attribute, and if both of those
+        # things are true, we'll skip the the converter. This will mean that the
+        # converters never have to deal with the None union.
+        if self.name in _REQUIRED_ATTRS or value is not None:
+            try:
+                converter: Callable[[Any], T] = getattr(self, f"_process_{self.name}")
+            except AttributeError:
+                pass
+            else:
+                value = converter(value)
+
+        cache[self.name] = value
+        try:
+            del instance._raw[self.name]  # type: ignore[misc]
+        except KeyError:
+            pass
+
+        return cast(T, value)
+
+    def _invalid_metadata(
+        self, msg: str, cause: Exception | None = None
+    ) -> InvalidMetadata:
+        exc = InvalidMetadata(
+            self.raw_name, msg.format_map({"field": repr(self.raw_name)})
+        )
+        exc.__cause__ = cause
+        return exc
+
+    def _process_metadata_version(self, value: str) -> _MetadataVersion:
+        # Implicitly makes Metadata-Version required.
+        if value not in _VALID_METADATA_VERSIONS:
+            raise self._invalid_metadata(f"{value!r} is not a valid metadata version")
+        return cast(_MetadataVersion, value)
+
+    def _process_name(self, value: str) -> str:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        # Validate the name as a side-effect.
+        try:
+            utils.canonicalize_name(value, validate=True)
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return value
+
+    def _process_version(self, value: str) -> version_module.Version:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        try:
+            return version_module.parse(value)
+        except version_module.InvalidVersion as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_summary(self, value: str) -> str:
+        """Check the field contains no newlines."""
+        if "\n" in value:
+            raise self._invalid_metadata("{field} must be a single line")
+        return value
+
+    def _process_description_content_type(self, value: str) -> str:
+        content_types = {"text/plain", "text/x-rst", "text/markdown"}
+        message = email.message.EmailMessage()
+        message["content-type"] = value
+
+        content_type, parameters = (
+            # Defaults to `text/plain` if parsing failed.
+            message.get_content_type().lower(),
+            message["content-type"].params,
+        )
+        # Check if content-type is valid or defaulted to `text/plain` and thus was
+        # not parseable.
+        if content_type not in content_types or content_type not in value.lower():
+            raise self._invalid_metadata(
+                f"{{field}} must be one of {list(content_types)}, not {value!r}"
+            )
+
+        charset = parameters.get("charset", "UTF-8")
+        if charset != "UTF-8":
+            raise self._invalid_metadata(
+                f"{{field}} can only specify the UTF-8 charset, not {list(charset)}"
+            )
+
+        markdown_variants = {"GFM", "CommonMark"}
+        variant = parameters.get("variant", "GFM")  # Use an acceptable default.
+        if content_type == "text/markdown" and variant not in markdown_variants:
+            raise self._invalid_metadata(
+                f"valid Markdown variants for {{field}} are {list(markdown_variants)}, "
+                f"not {variant!r}",
+            )
+        return value
+
+    def _process_dynamic(self, value: list[str]) -> list[str]:
+        for dynamic_field in map(str.lower, value):
+            if dynamic_field in {"name", "version", "metadata-version"}:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not allowed as a dynamic field"
+                )
+            elif dynamic_field not in _EMAIL_TO_RAW_MAPPING:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not a valid dynamic field"
+                )
+        return list(map(str.lower, value))
+
+    def _process_provides_extra(
+        self,
+        value: list[str],
+    ) -> list[utils.NormalizedName]:
+        normalized_names = []
+        try:
+            for name in value:
+                normalized_names.append(utils.canonicalize_name(name, validate=True))
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{name!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return normalized_names
+
+    def _process_requires_python(self, value: str) -> specifiers.SpecifierSet:
+        try:
+            return specifiers.SpecifierSet(value)
+        except specifiers.InvalidSpecifier as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_requires_dist(
+        self,
+        value: list[str],
+    ) -> list[requirements.Requirement]:
+        reqs = []
+        try:
+            for req in value:
+                reqs.append(requirements.Requirement(req))
+        except requirements.InvalidRequirement as exc:
+            raise self._invalid_metadata(
+                f"{req!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return reqs
+
+    def _process_license_expression(
+        self, value: str
+    ) -> NormalizedLicenseExpression | None:
+        try:
+            return licenses.canonicalize_license_expression(value)
+        except ValueError as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_license_files(self, value: list[str]) -> list[str]:
+        paths = []
+        for path in value:
+            if ".." in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "parent directory indicators are not allowed"
+                )
+            if "*" in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be resolved"
+                )
+            if (
+                pathlib.PurePosixPath(path).is_absolute()
+                or pathlib.PureWindowsPath(path).is_absolute()
+            ):
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be relative"
+                )
+            if pathlib.PureWindowsPath(path).as_posix() != path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "paths must use '/' delimiter"
+                )
+            paths.append(path)
+        return paths
+
+
+class Metadata:
+    """Representation of distribution metadata.
+
+    Compared to :class:`RawMetadata`, this class provides objects representing
+    metadata fields instead of only using built-in types. Any invalid metadata
+    will cause :exc:`InvalidMetadata` to be raised (with a
+    :py:attr:`~BaseException.__cause__` attribute as appropriate).
+    """
+
+    _raw: RawMetadata
+
+    @classmethod
+    def from_raw(cls, data: RawMetadata, *, validate: bool = True) -> Metadata:
+        """Create an instance from :class:`RawMetadata`.
+
+        If *validate* is true, all metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        ins = cls()
+        ins._raw = data.copy()  # Mutations occur due to caching enriched values.
+
+        if validate:
+            exceptions: list[Exception] = []
+            try:
+                metadata_version = ins.metadata_version
+                metadata_age = _VALID_METADATA_VERSIONS.index(metadata_version)
+            except InvalidMetadata as metadata_version_exc:
+                exceptions.append(metadata_version_exc)
+                metadata_version = None
+
+            # Make sure to check for the fields that are present, the required
+            # fields (so their absence can be reported).
+            fields_to_check = frozenset(ins._raw) | _REQUIRED_ATTRS
+            # Remove fields that have already been checked.
+            fields_to_check -= {"metadata_version"}
+
+            for key in fields_to_check:
+                try:
+                    if metadata_version:
+                        # Can't use getattr() as that triggers descriptor protocol which
+                        # will fail due to no value for the instance argument.
+                        try:
+                            field_metadata_version = cls.__dict__[key].added
+                        except KeyError:
+                            exc = InvalidMetadata(key, f"unrecognized field: {key!r}")
+                            exceptions.append(exc)
+                            continue
+                        field_age = _VALID_METADATA_VERSIONS.index(
+                            field_metadata_version
+                        )
+                        if field_age > metadata_age:
+                            field = _RAW_TO_EMAIL_MAPPING[key]
+                            exc = InvalidMetadata(
+                                field,
+                                f"{field} introduced in metadata version "
+                                f"{field_metadata_version}, not {metadata_version}",
+                            )
+                            exceptions.append(exc)
+                            continue
+                    getattr(ins, key)
+                except InvalidMetadata as exc:
+                    exceptions.append(exc)
+
+            if exceptions:
+                raise ExceptionGroup("invalid metadata", exceptions)
+
+        return ins
+
+    @classmethod
+    def from_email(cls, data: bytes | str, *, validate: bool = True) -> Metadata:
+        """Parse metadata from email headers.
+
+        If *validate* is true, the metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        raw, unparsed = parse_email(data)
+
+        if validate:
+            exceptions: list[Exception] = []
+            for unparsed_key in unparsed:
+                if unparsed_key in _EMAIL_TO_RAW_MAPPING:
+                    message = f"{unparsed_key!r} has invalid data"
+                else:
+                    message = f"unrecognized field: {unparsed_key!r}"
+                exceptions.append(InvalidMetadata(unparsed_key, message))
+
+            if exceptions:
+                raise ExceptionGroup("unparsed", exceptions)
+
+        try:
+            return cls.from_raw(raw, validate=validate)
+        except ExceptionGroup as exc_group:
+            raise ExceptionGroup(
+                "invalid or unparsed metadata", exc_group.exceptions
+            ) from None
+
+    metadata_version: _Validator[_MetadataVersion] = _Validator()
+    """:external:ref:`core-metadata-metadata-version`
+    (required; validated to be a valid metadata version)"""
+    # `name` is not normalized/typed to NormalizedName so as to provide access to
+    # the original/raw name.
+    name: _Validator[str] = _Validator()
+    """:external:ref:`core-metadata-name`
+    (required; validated using :func:`~packaging.utils.canonicalize_name` and its
+    *validate* parameter)"""
+    version: _Validator[version_module.Version] = _Validator()
+    """:external:ref:`core-metadata-version` (required)"""
+    dynamic: _Validator[list[str] | None] = _Validator(
+        added="2.2",
+    )
+    """:external:ref:`core-metadata-dynamic`
+    (validated against core metadata field names and lowercased)"""
+    platforms: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-platform`"""
+    supported_platforms: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-supported-platform`"""
+    summary: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-summary` (validated to contain no newlines)"""
+    description: _Validator[str | None] = _Validator()  # TODO 2.1: can be in body
+    """:external:ref:`core-metadata-description`"""
+    description_content_type: _Validator[str | None] = _Validator(added="2.1")
+    """:external:ref:`core-metadata-description-content-type` (validated)"""
+    keywords: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-keywords`"""
+    home_page: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-home-page`"""
+    download_url: _Validator[str | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-download-url`"""
+    author: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author`"""
+    author_email: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author-email`"""
+    maintainer: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer`"""
+    maintainer_email: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer-email`"""
+    license: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-license`"""
+    license_expression: _Validator[NormalizedLicenseExpression | None] = _Validator(
+        added="2.4"
+    )
+    """:external:ref:`core-metadata-license-expression`"""
+    license_files: _Validator[list[str] | None] = _Validator(added="2.4")
+    """:external:ref:`core-metadata-license-file`"""
+    classifiers: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-classifier`"""
+    requires_dist: _Validator[list[requirements.Requirement] | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-dist`"""
+    requires_python: _Validator[specifiers.SpecifierSet | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-python`"""
+    # Because `Requires-External` allows for non-PEP 440 version specifiers, we
+    # don't do any processing on the values.
+    requires_external: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-requires-external`"""
+    project_urls: _Validator[dict[str, str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-project-url`"""
+    # PEP 685 lets us raise an error if an extra doesn't pass `Name` validation
+    # regardless of metadata version.
+    provides_extra: _Validator[list[utils.NormalizedName] | None] = _Validator(
+        added="2.1",
+    )
+    """:external:ref:`core-metadata-provides-extra`"""
+    provides_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-provides-dist`"""
+    obsoletes_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-obsoletes-dist`"""
+    requires: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Requires`` (deprecated)"""
+    provides: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Provides`` (deprecated)"""
+    obsoletes: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Obsoletes`` (deprecated)"""
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/py.typed b/backend/venv/lib/python3.12/site-packages/packaging/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/requirements.py b/backend/venv/lib/python3.12/site-packages/packaging/requirements.py
new file mode 100644
index 00000000..4e068c95
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/requirements.py
@@ -0,0 +1,91 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+from __future__ import annotations
+
+from typing import Any, Iterator
+
+from ._parser import parse_requirement as _parse_requirement
+from ._tokenizer import ParserSyntaxError
+from .markers import Marker, _normalize_extra_values
+from .specifiers import SpecifierSet
+from .utils import canonicalize_name
+
+
+class InvalidRequirement(ValueError):
+    """
+    An invalid requirement was found, users should refer to PEP 508.
+    """
+
+
+class Requirement:
+    """Parse a requirement.
+
+    Parse a given requirement string into its parts, such as name, specifier,
+    URL, and extras. Raises InvalidRequirement on a badly-formed requirement
+    string.
+    """
+
+    # TODO: Can we test whether something is contained within a requirement?
+    #       If so how do we do that? Do we need to test against the _name_ of
+    #       the thing as well as the version? What about the markers?
+    # TODO: Can we normalize the name and extra name?
+
+    def __init__(self, requirement_string: str) -> None:
+        try:
+            parsed = _parse_requirement(requirement_string)
+        except ParserSyntaxError as e:
+            raise InvalidRequirement(str(e)) from e
+
+        self.name: str = parsed.name
+        self.url: str | None = parsed.url or None
+        self.extras: set[str] = set(parsed.extras or [])
+        self.specifier: SpecifierSet = SpecifierSet(parsed.specifier)
+        self.marker: Marker | None = None
+        if parsed.marker is not None:
+            self.marker = Marker.__new__(Marker)
+            self.marker._markers = _normalize_extra_values(parsed.marker)
+
+    def _iter_parts(self, name: str) -> Iterator[str]:
+        yield name
+
+        if self.extras:
+            formatted_extras = ",".join(sorted(self.extras))
+            yield f"[{formatted_extras}]"
+
+        if self.specifier:
+            yield str(self.specifier)
+
+        if self.url:
+            yield f"@ {self.url}"
+            if self.marker:
+                yield " "
+
+        if self.marker:
+            yield f"; {self.marker}"
+
+    def __str__(self) -> str:
+        return "".join(self._iter_parts(self.name))
+
+    def __repr__(self) -> str:
+        return f"<Requirement('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash(
+            (
+                self.__class__.__name__,
+                *self._iter_parts(canonicalize_name(self.name)),
+            )
+        )
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Requirement):
+            return NotImplemented
+
+        return (
+            canonicalize_name(self.name) == canonicalize_name(other.name)
+            and self.extras == other.extras
+            and self.specifier == other.specifier
+            and self.url == other.url
+            and self.marker == other.marker
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/specifiers.py b/backend/venv/lib/python3.12/site-packages/packaging/specifiers.py
new file mode 100644
index 00000000..b30926af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/specifiers.py
@@ -0,0 +1,1020 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.specifiers import Specifier, SpecifierSet, InvalidSpecifier
+    from packaging.version import Version
+"""
+
+from __future__ import annotations
+
+import abc
+import itertools
+import re
+from typing import Callable, Iterable, Iterator, TypeVar, Union
+
+from .utils import canonicalize_version
+from .version import Version
+
+UnparsedVersion = Union[Version, str]
+UnparsedVersionVar = TypeVar("UnparsedVersionVar", bound=UnparsedVersion)
+CallableOperator = Callable[[Version, str], bool]
+
+
+def _coerce_version(version: UnparsedVersion) -> Version:
+    if not isinstance(version, Version):
+        version = Version(version)
+    return version
+
+
+class InvalidSpecifier(ValueError):
+    """
+    Raised when attempting to create a :class:`Specifier` with a specifier
+    string that is invalid.
+
+    >>> Specifier("lolwat")
+    Traceback (most recent call last):
+        ...
+    packaging.specifiers.InvalidSpecifier: Invalid specifier: 'lolwat'
+    """
+
+
+class BaseSpecifier(metaclass=abc.ABCMeta):
+    @abc.abstractmethod
+    def __str__(self) -> str:
+        """
+        Returns the str representation of this Specifier-like object. This
+        should be representative of the Specifier itself.
+        """
+
+    @abc.abstractmethod
+    def __hash__(self) -> int:
+        """
+        Returns a hash value for this Specifier-like object.
+        """
+
+    @abc.abstractmethod
+    def __eq__(self, other: object) -> bool:
+        """
+        Returns a boolean representing whether or not the two Specifier-like
+        objects are equal.
+
+        :param other: The other object to check against.
+        """
+
+    @property
+    @abc.abstractmethod
+    def prereleases(self) -> bool | None:
+        """Whether or not pre-releases as a whole are allowed.
+
+        This can be set to either ``True`` or ``False`` to explicitly enable or disable
+        prereleases or it can be set to ``None`` (the default) to use default semantics.
+        """
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        """Setter for :attr:`prereleases`.
+
+        :param value: The value to set.
+        """
+
+    @abc.abstractmethod
+    def contains(self, item: str, prereleases: bool | None = None) -> bool:
+        """
+        Determines if the given item is contained within this specifier.
+        """
+
+    @abc.abstractmethod
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """
+        Takes an iterable of items and filters them so that only items which
+        are contained within this specifier are allowed in it.
+        """
+
+
+class Specifier(BaseSpecifier):
+    """This class abstracts handling of version specifiers.
+
+    .. tip::
+
+        It is generally not required to instantiate this manually. You should instead
+        prefer to work with :class:`SpecifierSet` instead, which can parse
+        comma-separated version specifiers (which is what package metadata contains).
+    """
+
+    _operator_regex_str = r"""
+        (?P<operator>(~=|==|!=|<=|>=|<|>|===))
+        """
+    _version_regex_str = r"""
+        (?P<version>
+            (?:
+                # The identity operators allow for an escape hatch that will
+                # do an exact string match of the version you wish to install.
+                # This will not be parsed by PEP 440 and we cannot determine
+                # any semantic meaning from it. This operator is discouraged
+                # but included entirely as an escape hatch.
+                (?<====)  # Only match for the identity operator
+                \s*
+                [^\s;)]*  # The arbitrary version can be just about anything,
+                          # we match everything except for whitespace, a
+                          # semi-colon for marker support, and a closing paren
+                          # since versions can be enclosed in them.
+            )
+            |
+            (?:
+                # The (non)equality operators allow for wild card and local
+                # versions to be specified so we have to define these two
+                # operators separately to enable that.
+                (?<===|!=)            # Only match for equals and not equals
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+
+                # You cannot use a wild card and a pre-release, post-release, a dev or
+                # local version together so group them with a | and make them optional.
+                (?:
+                    \.\*  # Wild card syntax of .*
+                    |
+                    (?:                                  # pre release
+                        [-_\.]?
+                        (alpha|beta|preview|pre|a|b|c|rc)
+                        [-_\.]?
+                        [0-9]*
+                    )?
+                    (?:                                  # post release
+                        (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                    )?
+                    (?:[-_\.]?dev[-_\.]?[0-9]*)?         # dev release
+                    (?:\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*)? # local
+                )?
+            )
+            |
+            (?:
+                # The compatible operator requires at least two digits in the
+                # release segment.
+                (?<=~=)               # Only match for the compatible operator
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)+   # release  (We have a + instead of a *)
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+            |
+            (?:
+                # All other operators only allow a sub set of what the
+                # (non)equality operators do. Specifically they do not allow
+                # local versions to be specified nor do they allow the prefix
+                # matching wild cards.
+                (?<!==|!=|~=)         # We have special cases for these
+                                      # operators so we want to make sure they
+                                      # don't match here.
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+        )
+        """
+
+    _regex = re.compile(
+        r"^\s*" + _operator_regex_str + _version_regex_str + r"\s*$",
+        re.VERBOSE | re.IGNORECASE,
+    )
+
+    _operators = {
+        "~=": "compatible",
+        "==": "equal",
+        "!=": "not_equal",
+        "<=": "less_than_equal",
+        ">=": "greater_than_equal",
+        "<": "less_than",
+        ">": "greater_than",
+        "===": "arbitrary",
+    }
+
+    def __init__(self, spec: str = "", prereleases: bool | None = None) -> None:
+        """Initialize a Specifier instance.
+
+        :param spec:
+            The string representation of a specifier which will be parsed and
+            normalized before use.
+        :param prereleases:
+            This tells the specifier if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+        :raises InvalidSpecifier:
+            If the given specifier is invalid (i.e. bad syntax).
+        """
+        match = self._regex.search(spec)
+        if not match:
+            raise InvalidSpecifier(f"Invalid specifier: {spec!r}")
+
+        self._spec: tuple[str, str] = (
+            match.group("operator").strip(),
+            match.group("version").strip(),
+        )
+
+        # Store whether or not this Specifier should accept prereleases
+        self._prereleases = prereleases
+
+    # https://github.com/python/mypy/pull/13475#pullrequestreview-1079784515
+    @property  # type: ignore[override]
+    def prereleases(self) -> bool:
+        # If there is an explicit prereleases set for this, then we'll just
+        # blindly use that.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # Look at all of our specifiers and determine if they are inclusive
+        # operators, and if they are if they are including an explicit
+        # prerelease.
+        operator, version = self._spec
+        if operator in ["==", ">=", "<=", "~=", "===", ">", "<"]:
+            # The == specifier can include a trailing .*, if it does we
+            # want to remove before parsing.
+            if operator == "==" and version.endswith(".*"):
+                version = version[:-2]
+
+            # Parse the version, and if it is a pre-release than this
+            # specifier allows pre-releases.
+            if Version(version).is_prerelease:
+                return True
+
+        return False
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    @property
+    def operator(self) -> str:
+        """The operator of this specifier.
+
+        >>> Specifier("==1.2.3").operator
+        '=='
+        """
+        return self._spec[0]
+
+    @property
+    def version(self) -> str:
+        """The version of this specifier.
+
+        >>> Specifier("==1.2.3").version
+        '1.2.3'
+        """
+        return self._spec[1]
+
+    def __repr__(self) -> str:
+        """A representation of the Specifier that shows all internal state.
+
+        >>> Specifier('>=1.0.0')
+        <Specifier('>=1.0.0')>
+        >>> Specifier('>=1.0.0', prereleases=False)
+        <Specifier('>=1.0.0', prereleases=False)>
+        >>> Specifier('>=1.0.0', prereleases=True)
+        <Specifier('>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<{self.__class__.__name__}({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the Specifier that can be round-tripped.
+
+        >>> str(Specifier('>=1.0.0'))
+        '>=1.0.0'
+        >>> str(Specifier('>=1.0.0', prereleases=False))
+        '>=1.0.0'
+        """
+        return "{}{}".format(*self._spec)
+
+    @property
+    def _canonical_spec(self) -> tuple[str, str]:
+        canonical_version = canonicalize_version(
+            self._spec[1],
+            strip_trailing_zero=(self._spec[0] != "~="),
+        )
+        return self._spec[0], canonical_version
+
+    def __hash__(self) -> int:
+        return hash(self._canonical_spec)
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two Specifier-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> Specifier("==1.2.3") == Specifier("== 1.2.3.0")
+        True
+        >>> (Specifier("==1.2.3", prereleases=False) ==
+        ...  Specifier("==1.2.3", prereleases=True))
+        True
+        >>> Specifier("==1.2.3") == "==1.2.3"
+        True
+        >>> Specifier("==1.2.3") == Specifier("==1.2.4")
+        False
+        >>> Specifier("==1.2.3") == Specifier("~=1.2.3")
+        False
+        """
+        if isinstance(other, str):
+            try:
+                other = self.__class__(str(other))
+            except InvalidSpecifier:
+                return NotImplemented
+        elif not isinstance(other, self.__class__):
+            return NotImplemented
+
+        return self._canonical_spec == other._canonical_spec
+
+    def _get_operator(self, op: str) -> CallableOperator:
+        operator_callable: CallableOperator = getattr(
+            self, f"_compare_{self._operators[op]}"
+        )
+        return operator_callable
+
+    def _compare_compatible(self, prospective: Version, spec: str) -> bool:
+        # Compatible releases have an equivalent combination of >= and ==. That
+        # is that ~=2.2 is equivalent to >=2.2,==2.*. This allows us to
+        # implement this in terms of the other specifiers instead of
+        # implementing it ourselves. The only thing we need to do is construct
+        # the other specifiers.
+
+        # We want everything but the last item in the version, but we want to
+        # ignore suffix segments.
+        prefix = _version_join(
+            list(itertools.takewhile(_is_not_suffix, _version_split(spec)))[:-1]
+        )
+
+        # Add the prefix notation to the end of our string
+        prefix += ".*"
+
+        return self._get_operator(">=")(prospective, spec) and self._get_operator("==")(
+            prospective, prefix
+        )
+
+    def _compare_equal(self, prospective: Version, spec: str) -> bool:
+        # We need special logic to handle prefix matching
+        if spec.endswith(".*"):
+            # In the case of prefix matching we want to ignore local segment.
+            normalized_prospective = canonicalize_version(
+                prospective.public, strip_trailing_zero=False
+            )
+            # Get the normalized version string ignoring the trailing .*
+            normalized_spec = canonicalize_version(spec[:-2], strip_trailing_zero=False)
+            # Split the spec out by bangs and dots, and pretend that there is
+            # an implicit dot in between a release segment and a pre-release segment.
+            split_spec = _version_split(normalized_spec)
+
+            # Split the prospective version out by bangs and dots, and pretend
+            # that there is an implicit dot in between a release segment and
+            # a pre-release segment.
+            split_prospective = _version_split(normalized_prospective)
+
+            # 0-pad the prospective version before shortening it to get the correct
+            # shortened version.
+            padded_prospective, _ = _pad_version(split_prospective, split_spec)
+
+            # Shorten the prospective version to be the same length as the spec
+            # so that we can determine if the specifier is a prefix of the
+            # prospective version or not.
+            shortened_prospective = padded_prospective[: len(split_spec)]
+
+            return shortened_prospective == split_spec
+        else:
+            # Convert our spec string into a Version
+            spec_version = Version(spec)
+
+            # If the specifier does not have a local segment, then we want to
+            # act as if the prospective version also does not have a local
+            # segment.
+            if not spec_version.local:
+                prospective = Version(prospective.public)
+
+            return prospective == spec_version
+
+    def _compare_not_equal(self, prospective: Version, spec: str) -> bool:
+        return not self._compare_equal(prospective, spec)
+
+    def _compare_less_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) <= Version(spec)
+
+    def _compare_greater_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) >= Version(spec)
+
+    def _compare_less_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is less than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective < spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a pre-release version, that we do not accept pre-release
+        # versions for the version mentioned in the specifier (e.g. <3.1 should
+        # not match 3.1.dev0, but should match 3.0.dev0).
+        if not spec.is_prerelease and prospective.is_prerelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # less than the spec version *and* it's not a pre-release of the same
+        # version in the spec.
+        return True
+
+    def _compare_greater_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is greater than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective > spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a post-release version, that we do not accept
+        # post-release versions for the version mentioned in the specifier
+        # (e.g. >3.1 should not match 3.0.post0, but should match 3.2.post0).
+        if not spec.is_postrelease and prospective.is_postrelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # Ensure that we do not allow a local version of the version mentioned
+        # in the specifier, which is technically greater than, to match.
+        if prospective.local is not None:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # greater than the spec version *and* it's not a pre-release of the
+        # same version in the spec.
+        return True
+
+    def _compare_arbitrary(self, prospective: Version, spec: str) -> bool:
+        return str(prospective).lower() == str(spec).lower()
+
+    def __contains__(self, item: str | Version) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in Specifier(">=1.2.3")
+        True
+        >>> Version("1.2.3") in Specifier(">=1.2.3")
+        True
+        >>> "1.0.0" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(self, item: UnparsedVersion, prereleases: bool | None = None) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this Specifier. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> Specifier(">=1.2.3").contains("1.2.3")
+        True
+        >>> Specifier(">=1.2.3").contains(Version("1.2.3"))
+        True
+        >>> Specifier(">=1.2.3").contains("1.0.0")
+        False
+        >>> Specifier(">=1.2.3").contains("1.3.0a1")
+        False
+        >>> Specifier(">=1.2.3", prereleases=True).contains("1.3.0a1")
+        True
+        >>> Specifier(">=1.2.3").contains("1.3.0a1", prereleases=True)
+        True
+        """
+
+        # Determine if prereleases are to be allowed or not.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # Normalize item to a Version, this allows us to have a shortcut for
+        # "2.0" in Specifier(">=2")
+        normalized_item = _coerce_version(item)
+
+        # Determine if we should be supporting prereleases in this specifier
+        # or not, if we do not support prereleases than we can short circuit
+        # logic if this version is a prereleases.
+        if normalized_item.is_prerelease and not prereleases:
+            return False
+
+        # Actually do the comparison to determine if this item is contained
+        # within this Specifier or not.
+        operator_callable: CallableOperator = self._get_operator(self.operator)
+        return operator_callable(normalized_item, self.version)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifier.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(Specifier().contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.2.3", "1.3", Version("1.4")]))
+        ['1.2.3', '1.3', <Version('1.4')>]
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.5a1"]))
+        ['1.5a1']
+        >>> list(Specifier(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(Specifier(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        """
+
+        yielded = False
+        found_prereleases = []
+
+        kw = {"prereleases": prereleases if prereleases is not None else True}
+
+        # Attempt to iterate over all the values in the iterable and if any of
+        # them match, yield them.
+        for version in iterable:
+            parsed_version = _coerce_version(version)
+
+            if self.contains(parsed_version, **kw):
+                # If our version is a prerelease, and we were not set to allow
+                # prereleases, then we'll store it for later in case nothing
+                # else matches this specifier.
+                if parsed_version.is_prerelease and not (
+                    prereleases or self.prereleases
+                ):
+                    found_prereleases.append(version)
+                # Either this is not a prerelease, or we should have been
+                # accepting prereleases from the beginning.
+                else:
+                    yielded = True
+                    yield version
+
+        # Now that we've iterated over everything, determine if we've yielded
+        # any values, and if we have not and we have any prereleases stored up
+        # then we will go ahead and yield the prereleases.
+        if not yielded and found_prereleases:
+            for version in found_prereleases:
+                yield version
+
+
+_prefix_regex = re.compile(r"^([0-9]+)((?:a|b|c|rc)[0-9]+)$")
+
+
+def _version_split(version: str) -> list[str]:
+    """Split version into components.
+
+    The split components are intended for version comparison. The logic does
+    not attempt to retain the original version string, so joining the
+    components back with :func:`_version_join` may not produce the original
+    version string.
+    """
+    result: list[str] = []
+
+    epoch, _, rest = version.rpartition("!")
+    result.append(epoch or "0")
+
+    for item in rest.split("."):
+        match = _prefix_regex.search(item)
+        if match:
+            result.extend(match.groups())
+        else:
+            result.append(item)
+    return result
+
+
+def _version_join(components: list[str]) -> str:
+    """Join split version components into a version string.
+
+    This function assumes the input came from :func:`_version_split`, where the
+    first component must be the epoch (either empty or numeric), and all other
+    components numeric.
+    """
+    epoch, *rest = components
+    return f"{epoch}!{'.'.join(rest)}"
+
+
+def _is_not_suffix(segment: str) -> bool:
+    return not any(
+        segment.startswith(prefix) for prefix in ("dev", "a", "b", "rc", "post")
+    )
+
+
+def _pad_version(left: list[str], right: list[str]) -> tuple[list[str], list[str]]:
+    left_split, right_split = [], []
+
+    # Get the release segment of our versions
+    left_split.append(list(itertools.takewhile(lambda x: x.isdigit(), left)))
+    right_split.append(list(itertools.takewhile(lambda x: x.isdigit(), right)))
+
+    # Get the rest of our versions
+    left_split.append(left[len(left_split[0]) :])
+    right_split.append(right[len(right_split[0]) :])
+
+    # Insert our padding
+    left_split.insert(1, ["0"] * max(0, len(right_split[0]) - len(left_split[0])))
+    right_split.insert(1, ["0"] * max(0, len(left_split[0]) - len(right_split[0])))
+
+    return (
+        list(itertools.chain.from_iterable(left_split)),
+        list(itertools.chain.from_iterable(right_split)),
+    )
+
+
+class SpecifierSet(BaseSpecifier):
+    """This class abstracts handling of a set of version specifiers.
+
+    It can be passed a single specifier (``>=3.0``), a comma-separated list of
+    specifiers (``>=3.0,!=3.1``), or no specifier at all.
+    """
+
+    def __init__(
+        self,
+        specifiers: str | Iterable[Specifier] = "",
+        prereleases: bool | None = None,
+    ) -> None:
+        """Initialize a SpecifierSet instance.
+
+        :param specifiers:
+            The string representation of a specifier or a comma-separated list of
+            specifiers which will be parsed and normalized before use.
+            May also be an iterable of ``Specifier`` instances, which will be used
+            as is.
+        :param prereleases:
+            This tells the SpecifierSet if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+
+        :raises InvalidSpecifier:
+            If the given ``specifiers`` are not parseable than this exception will be
+            raised.
+        """
+
+        if isinstance(specifiers, str):
+            # Split on `,` to break each individual specifier into its own item, and
+            # strip each item to remove leading/trailing whitespace.
+            split_specifiers = [s.strip() for s in specifiers.split(",") if s.strip()]
+
+            # Make each individual specifier a Specifier and save in a frozen set
+            # for later.
+            self._specs = frozenset(map(Specifier, split_specifiers))
+        else:
+            # Save the supplied specifiers in a frozen set.
+            self._specs = frozenset(specifiers)
+
+        # Store our prereleases value so we can use it later to determine if
+        # we accept prereleases or not.
+        self._prereleases = prereleases
+
+    @property
+    def prereleases(self) -> bool | None:
+        # If we have been given an explicit prerelease modifier, then we'll
+        # pass that through here.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # If we don't have any specifiers, and we don't have a forced value,
+        # then we'll just return None since we don't know if this should have
+        # pre-releases or not.
+        if not self._specs:
+            return None
+
+        # Otherwise we'll see if any of the given specifiers accept
+        # prereleases, if any of them do we'll return True, otherwise False.
+        return any(s.prereleases for s in self._specs)
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    def __repr__(self) -> str:
+        """A representation of the specifier set that shows all internal state.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> SpecifierSet('>=1.0.0,!=2.0.0')
+        <SpecifierSet('!=2.0.0,>=1.0.0')>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=False)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=False)>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=True)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<SpecifierSet({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the specifier set that can be round-tripped.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1"))
+        '!=1.0.1,>=1.0.0'
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False))
+        '!=1.0.1,>=1.0.0'
+        """
+        return ",".join(sorted(str(s) for s in self._specs))
+
+    def __hash__(self) -> int:
+        return hash(self._specs)
+
+    def __and__(self, other: SpecifierSet | str) -> SpecifierSet:
+        """Return a SpecifierSet which is a combination of the two sets.
+
+        :param other: The other object to combine with.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & '<=2.0.0,!=2.0.1'
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & SpecifierSet('<=2.0.0,!=2.0.1')
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        """
+        if isinstance(other, str):
+            other = SpecifierSet(other)
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        specifier = SpecifierSet()
+        specifier._specs = frozenset(self._specs | other._specs)
+
+        if self._prereleases is None and other._prereleases is not None:
+            specifier._prereleases = other._prereleases
+        elif self._prereleases is not None and other._prereleases is None:
+            specifier._prereleases = self._prereleases
+        elif self._prereleases == other._prereleases:
+            specifier._prereleases = self._prereleases
+        else:
+            raise ValueError(
+                "Cannot combine SpecifierSets with True and False prerelease "
+                "overrides."
+            )
+
+        return specifier
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two SpecifierSet-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> (SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False) ==
+        ...  SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == ">=1.0.0,!=1.0.1"
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.2")
+        False
+        """
+        if isinstance(other, (str, Specifier)):
+            other = SpecifierSet(str(other))
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        return self._specs == other._specs
+
+    def __len__(self) -> int:
+        """Returns the number of specifiers in this specifier set."""
+        return len(self._specs)
+
+    def __iter__(self) -> Iterator[Specifier]:
+        """
+        Returns an iterator over all the underlying :class:`Specifier` instances
+        in this specifier set.
+
+        >>> sorted(SpecifierSet(">=1.0.0,!=1.0.1"), key=str)
+        [<Specifier('!=1.0.1')>, <Specifier('>=1.0.0')>]
+        """
+        return iter(self._specs)
+
+    def __contains__(self, item: UnparsedVersion) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> Version("1.2.3") in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> "1.0.1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(
+        self,
+        item: UnparsedVersion,
+        prereleases: bool | None = None,
+        installed: bool | None = None,
+    ) -> bool:
+        """Return whether or not the item is contained in this SpecifierSet.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this SpecifierSet. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.2.3")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains(Version("1.2.3"))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.0.1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True).contains("1.3.0a1")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1", prereleases=True)
+        True
+        """
+        # Ensure that our item is a Version instance.
+        if not isinstance(item, Version):
+            item = Version(item)
+
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # We can determine if we're going to allow pre-releases by looking to
+        # see if any of the underlying items supports them. If none of them do
+        # and this item is a pre-release then we do not allow it and we can
+        # short circuit that here.
+        # Note: This means that 1.0.dev1 would not be contained in something
+        #       like >=1.0.devabc however it would be in >=1.0.debabc,>0.0.dev0
+        if not prereleases and item.is_prerelease:
+            return False
+
+        if installed and item.is_prerelease:
+            item = Version(item.base_version)
+
+        # We simply dispatch to the underlying specs here to make sure that the
+        # given version is contained within all of them.
+        # Note: This use of all() here means that an empty set of specifiers
+        #       will always return True, this is an explicit design decision.
+        return all(s.contains(item, prereleases=prereleases) for s in self._specs)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifiers in this set.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(SpecifierSet(...).contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", Version("1.4")]))
+        ['1.3', <Version('1.4')>]
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.5a1"]))
+        []
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+
+        An "empty" SpecifierSet will filter items based on the presence of prerelease
+        versions in the set.
+
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet("").filter(["1.5a1"]))
+        ['1.5a1']
+        >>> list(SpecifierSet("", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        """
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # If we have any specifiers, then we want to wrap our iterable in the
+        # filter method for each one, this will act as a logical AND amongst
+        # each specifier.
+        if self._specs:
+            for spec in self._specs:
+                iterable = spec.filter(iterable, prereleases=bool(prereleases))
+            return iter(iterable)
+        # If we do not have any specifiers, then we need to have a rough filter
+        # which will filter out any pre-releases, unless there are no final
+        # releases.
+        else:
+            filtered: list[UnparsedVersionVar] = []
+            found_prereleases: list[UnparsedVersionVar] = []
+
+            for item in iterable:
+                parsed_version = _coerce_version(item)
+
+                # Store any item which is a pre-release for later unless we've
+                # already found a final version or we are accepting prereleases
+                if parsed_version.is_prerelease and not prereleases:
+                    if not filtered:
+                        found_prereleases.append(item)
+                else:
+                    filtered.append(item)
+
+            # If we've found no items except for pre-releases, then we'll go
+            # ahead and use the pre-releases
+            if not filtered and found_prereleases and prereleases is None:
+                return iter(found_prereleases)
+
+            return iter(filtered)
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/tags.py b/backend/venv/lib/python3.12/site-packages/packaging/tags.py
new file mode 100644
index 00000000..f5903402
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/tags.py
@@ -0,0 +1,617 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import logging
+import platform
+import re
+import struct
+import subprocess
+import sys
+import sysconfig
+from importlib.machinery import EXTENSION_SUFFIXES
+from typing import (
+    Iterable,
+    Iterator,
+    Sequence,
+    Tuple,
+    cast,
+)
+
+from . import _manylinux, _musllinux
+
+logger = logging.getLogger(__name__)
+
+PythonVersion = Sequence[int]
+AppleVersion = Tuple[int, int]
+
+INTERPRETER_SHORT_NAMES: dict[str, str] = {
+    "python": "py",  # Generic.
+    "cpython": "cp",
+    "pypy": "pp",
+    "ironpython": "ip",
+    "jython": "jy",
+}
+
+
+_32_BIT_INTERPRETER = struct.calcsize("P") == 4
+
+
+class Tag:
+    """
+    A representation of the tag triple for a wheel.
+
+    Instances are considered immutable and thus are hashable. Equality checking
+    is also supported.
+    """
+
+    __slots__ = ["_abi", "_hash", "_interpreter", "_platform"]
+
+    def __init__(self, interpreter: str, abi: str, platform: str) -> None:
+        self._interpreter = interpreter.lower()
+        self._abi = abi.lower()
+        self._platform = platform.lower()
+        # The __hash__ of every single element in a Set[Tag] will be evaluated each time
+        # that a set calls its `.disjoint()` method, which may be called hundreds of
+        # times when scanning a page of links for packages with tags matching that
+        # Set[Tag]. Pre-computing the value here produces significant speedups for
+        # downstream consumers.
+        self._hash = hash((self._interpreter, self._abi, self._platform))
+
+    @property
+    def interpreter(self) -> str:
+        return self._interpreter
+
+    @property
+    def abi(self) -> str:
+        return self._abi
+
+    @property
+    def platform(self) -> str:
+        return self._platform
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Tag):
+            return NotImplemented
+
+        return (
+            (self._hash == other._hash)  # Short-circuit ASAP for perf reasons.
+            and (self._platform == other._platform)
+            and (self._abi == other._abi)
+            and (self._interpreter == other._interpreter)
+        )
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __str__(self) -> str:
+        return f"{self._interpreter}-{self._abi}-{self._platform}"
+
+    def __repr__(self) -> str:
+        return f"<{self} @ {id(self)}>"
+
+
+def parse_tag(tag: str) -> frozenset[Tag]:
+    """
+    Parses the provided tag (e.g. `py3-none-any`) into a frozenset of Tag instances.
+
+    Returning a set is required due to the possibility that the tag is a
+    compressed tag set.
+    """
+    tags = set()
+    interpreters, abis, platforms = tag.split("-")
+    for interpreter in interpreters.split("."):
+        for abi in abis.split("."):
+            for platform_ in platforms.split("."):
+                tags.add(Tag(interpreter, abi, platform_))
+    return frozenset(tags)
+
+
+def _get_config_var(name: str, warn: bool = False) -> int | str | None:
+    value: int | str | None = sysconfig.get_config_var(name)
+    if value is None and warn:
+        logger.debug(
+            "Config variable '%s' is unset, Python ABI tag may be incorrect", name
+        )
+    return value
+
+
+def _normalize_string(string: str) -> str:
+    return string.replace(".", "_").replace("-", "_").replace(" ", "_")
+
+
+def _is_threaded_cpython(abis: list[str]) -> bool:
+    """
+    Determine if the ABI corresponds to a threaded (`--disable-gil`) build.
+
+    The threaded builds are indicated by a "t" in the abiflags.
+    """
+    if len(abis) == 0:
+        return False
+    # expect e.g., cp313
+    m = re.match(r"cp\d+(.*)", abis[0])
+    if not m:
+        return False
+    abiflags = m.group(1)
+    return "t" in abiflags
+
+
+def _abi3_applies(python_version: PythonVersion, threading: bool) -> bool:
+    """
+    Determine if the Python version supports abi3.
+
+    PEP 384 was first implemented in Python 3.2. The threaded (`--disable-gil`)
+    builds do not support abi3.
+    """
+    return len(python_version) > 1 and tuple(python_version) >= (3, 2) and not threading
+
+
+def _cpython_abis(py_version: PythonVersion, warn: bool = False) -> list[str]:
+    py_version = tuple(py_version)  # To allow for version comparison.
+    abis = []
+    version = _version_nodot(py_version[:2])
+    threading = debug = pymalloc = ucs4 = ""
+    with_debug = _get_config_var("Py_DEBUG", warn)
+    has_refcount = hasattr(sys, "gettotalrefcount")
+    # Windows doesn't set Py_DEBUG, so checking for support of debug-compiled
+    # extension modules is the best option.
+    # https://github.com/pypa/pip/issues/3383#issuecomment-173267692
+    has_ext = "_d.pyd" in EXTENSION_SUFFIXES
+    if with_debug or (with_debug is None and (has_refcount or has_ext)):
+        debug = "d"
+    if py_version >= (3, 13) and _get_config_var("Py_GIL_DISABLED", warn):
+        threading = "t"
+    if py_version < (3, 8):
+        with_pymalloc = _get_config_var("WITH_PYMALLOC", warn)
+        if with_pymalloc or with_pymalloc is None:
+            pymalloc = "m"
+        if py_version < (3, 3):
+            unicode_size = _get_config_var("Py_UNICODE_SIZE", warn)
+            if unicode_size == 4 or (
+                unicode_size is None and sys.maxunicode == 0x10FFFF
+            ):
+                ucs4 = "u"
+    elif debug:
+        # Debug builds can also load "normal" extension modules.
+        # We can also assume no UCS-4 or pymalloc requirement.
+        abis.append(f"cp{version}{threading}")
+    abis.insert(0, f"cp{version}{threading}{debug}{pymalloc}{ucs4}")
+    return abis
+
+
+def cpython_tags(
+    python_version: PythonVersion | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a CPython interpreter.
+
+    The tags consist of:
+    - cp<python_version>-<abi>-<platform>
+    - cp<python_version>-abi3-<platform>
+    - cp<python_version>-none-<platform>
+    - cp<less than python_version>-abi3-<platform>  # Older Python versions down to 3.2.
+
+    If python_version only specifies a major version then user-provided ABIs and
+    the 'none' ABItag will be used.
+
+    If 'abi3' or 'none' are specified in 'abis' then they will be yielded at
+    their normal position and not at the beginning.
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+
+    interpreter = f"cp{_version_nodot(python_version[:2])}"
+
+    if abis is None:
+        if len(python_version) > 1:
+            abis = _cpython_abis(python_version, warn)
+        else:
+            abis = []
+    abis = list(abis)
+    # 'abi3' and 'none' are explicitly handled later.
+    for explicit_abi in ("abi3", "none"):
+        try:
+            abis.remove(explicit_abi)
+        except ValueError:
+            pass
+
+    platforms = list(platforms or platform_tags())
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+    threading = _is_threaded_cpython(abis)
+    use_abi3 = _abi3_applies(python_version, threading)
+    if use_abi3:
+        yield from (Tag(interpreter, "abi3", platform_) for platform_ in platforms)
+    yield from (Tag(interpreter, "none", platform_) for platform_ in platforms)
+
+    if use_abi3:
+        for minor_version in range(python_version[1] - 1, 1, -1):
+            for platform_ in platforms:
+                version = _version_nodot((python_version[0], minor_version))
+                interpreter = f"cp{version}"
+                yield Tag(interpreter, "abi3", platform_)
+
+
+def _generic_abi() -> list[str]:
+    """
+    Return the ABI tag based on EXT_SUFFIX.
+    """
+    # The following are examples of `EXT_SUFFIX`.
+    # We want to keep the parts which are related to the ABI and remove the
+    # parts which are related to the platform:
+    # - linux:   '.cpython-310-x86_64-linux-gnu.so' => cp310
+    # - mac:     '.cpython-310-darwin.so'           => cp310
+    # - win:     '.cp310-win_amd64.pyd'             => cp310
+    # - win:     '.pyd'                             => cp37 (uses _cpython_abis())
+    # - pypy:    '.pypy38-pp73-x86_64-linux-gnu.so' => pypy38_pp73
+    # - graalpy: '.graalpy-38-native-x86_64-darwin.dylib'
+    #                                               => graalpy_38_native
+
+    ext_suffix = _get_config_var("EXT_SUFFIX", warn=True)
+    if not isinstance(ext_suffix, str) or ext_suffix[0] != ".":
+        raise SystemError("invalid sysconfig.get_config_var('EXT_SUFFIX')")
+    parts = ext_suffix.split(".")
+    if len(parts) < 3:
+        # CPython3.7 and earlier uses ".pyd" on Windows.
+        return _cpython_abis(sys.version_info[:2])
+    soabi = parts[1]
+    if soabi.startswith("cpython"):
+        # non-windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi.startswith("cp"):
+        # windows
+        abi = soabi.split("-")[0]
+    elif soabi.startswith("pypy"):
+        abi = "-".join(soabi.split("-")[:2])
+    elif soabi.startswith("graalpy"):
+        abi = "-".join(soabi.split("-")[:3])
+    elif soabi:
+        # pyston, ironpython, others?
+        abi = soabi
+    else:
+        return []
+    return [_normalize_string(abi)]
+
+
+def generic_tags(
+    interpreter: str | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a generic interpreter.
+
+    The tags consist of:
+    - <interpreter>-<abi>-<platform>
+
+    The "none" ABI will be added if it was not explicitly provided.
+    """
+    if not interpreter:
+        interp_name = interpreter_name()
+        interp_version = interpreter_version(warn=warn)
+        interpreter = "".join([interp_name, interp_version])
+    if abis is None:
+        abis = _generic_abi()
+    else:
+        abis = list(abis)
+    platforms = list(platforms or platform_tags())
+    if "none" not in abis:
+        abis.append("none")
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+
+def _py_interpreter_range(py_version: PythonVersion) -> Iterator[str]:
+    """
+    Yields Python versions in descending order.
+
+    After the latest version, the major-only version will be yielded, and then
+    all previous versions of that major version.
+    """
+    if len(py_version) > 1:
+        yield f"py{_version_nodot(py_version[:2])}"
+    yield f"py{py_version[0]}"
+    if len(py_version) > 1:
+        for minor in range(py_version[1] - 1, -1, -1):
+            yield f"py{_version_nodot((py_version[0], minor))}"
+
+
+def compatible_tags(
+    python_version: PythonVersion | None = None,
+    interpreter: str | None = None,
+    platforms: Iterable[str] | None = None,
+) -> Iterator[Tag]:
+    """
+    Yields the sequence of tags that are compatible with a specific version of Python.
+
+    The tags consist of:
+    - py*-none-<platform>
+    - <interpreter>-none-any  # ... if `interpreter` is provided.
+    - py*-none-any
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+    platforms = list(platforms or platform_tags())
+    for version in _py_interpreter_range(python_version):
+        for platform_ in platforms:
+            yield Tag(version, "none", platform_)
+    if interpreter:
+        yield Tag(interpreter, "none", "any")
+    for version in _py_interpreter_range(python_version):
+        yield Tag(version, "none", "any")
+
+
+def _mac_arch(arch: str, is_32bit: bool = _32_BIT_INTERPRETER) -> str:
+    if not is_32bit:
+        return arch
+
+    if arch.startswith("ppc"):
+        return "ppc"
+
+    return "i386"
+
+
+def _mac_binary_formats(version: AppleVersion, cpu_arch: str) -> list[str]:
+    formats = [cpu_arch]
+    if cpu_arch == "x86_64":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat64", "fat32"])
+
+    elif cpu_arch == "i386":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat32", "fat"])
+
+    elif cpu_arch == "ppc64":
+        # TODO: Need to care about 32-bit PPC for ppc64 through 10.2?
+        if version > (10, 5) or version < (10, 4):
+            return []
+        formats.append("fat64")
+
+    elif cpu_arch == "ppc":
+        if version > (10, 6):
+            return []
+        formats.extend(["fat32", "fat"])
+
+    if cpu_arch in {"arm64", "x86_64"}:
+        formats.append("universal2")
+
+    if cpu_arch in {"x86_64", "i386", "ppc64", "ppc", "intel"}:
+        formats.append("universal")
+
+    return formats
+
+
+def mac_platforms(
+    version: AppleVersion | None = None, arch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for a macOS system.
+
+    The `version` parameter is a two-item tuple specifying the macOS version to
+    generate platform tags for. The `arch` parameter is the CPU architecture to
+    generate platform tags for. Both parameters default to the appropriate value
+    for the current system.
+    """
+    version_str, _, cpu_arch = platform.mac_ver()
+    if version is None:
+        version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+        if version == (10, 16):
+            # When built against an older macOS SDK, Python will report macOS 10.16
+            # instead of the real version.
+            version_str = subprocess.run(
+                [
+                    sys.executable,
+                    "-sS",
+                    "-c",
+                    "import platform; print(platform.mac_ver()[0])",
+                ],
+                check=True,
+                env={"SYSTEM_VERSION_COMPAT": "0"},
+                stdout=subprocess.PIPE,
+                text=True,
+            ).stdout
+            version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+    else:
+        version = version
+    if arch is None:
+        arch = _mac_arch(cpu_arch)
+    else:
+        arch = arch
+
+    if (10, 0) <= version and version < (11, 0):
+        # Prior to Mac OS 11, each yearly release of Mac OS bumped the
+        # "minor" version number.  The major version was always 10.
+        major_version = 10
+        for minor_version in range(version[1], -1, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Starting with Mac OS 11, each yearly release bumps the major version
+        # number.   The minor versions are now the midyear updates.
+        minor_version = 0
+        for major_version in range(version[0], 10, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Mac OS 11 on x86_64 is compatible with binaries from previous releases.
+        # Arm64 support was introduced in 11.0, so no Arm binaries from previous
+        # releases exist.
+        #
+        # However, the "universal2" binary format can have a
+        # macOS version earlier than 11.0 when the x86_64 part of the binary supports
+        # that version of macOS.
+        major_version = 10
+        if arch == "x86_64":
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_formats = _mac_binary_formats(compat_version, arch)
+                for binary_format in binary_formats:
+                    yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+        else:
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_format = "universal2"
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+
+def ios_platforms(
+    version: AppleVersion | None = None, multiarch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for an iOS system.
+
+    :param version: A two-item tuple specifying the iOS version to generate
+        platform tags for. Defaults to the current iOS version.
+    :param multiarch: The CPU architecture+ABI to generate platform tags for -
+        (the value used by `sys.implementation._multiarch` e.g.,
+        `arm64_iphoneos` or `x84_64_iphonesimulator`). Defaults to the current
+        multiarch value.
+    """
+    if version is None:
+        # if iOS is the current platform, ios_ver *must* be defined. However,
+        # it won't exist for CPython versions before 3.13, which causes a mypy
+        # error.
+        _, release, _, _ = platform.ios_ver()  # type: ignore[attr-defined, unused-ignore]
+        version = cast("AppleVersion", tuple(map(int, release.split(".")[:2])))
+
+    if multiarch is None:
+        multiarch = sys.implementation._multiarch
+    multiarch = multiarch.replace("-", "_")
+
+    ios_platform_template = "ios_{major}_{minor}_{multiarch}"
+
+    # Consider any iOS major.minor version from the version requested, down to
+    # 12.0. 12.0 is the first iOS version that is known to have enough features
+    # to support CPython. Consider every possible minor release up to X.9. There
+    # highest the minor has ever gone is 8 (14.8 and 15.8) but having some extra
+    # candidates that won't ever match doesn't really hurt, and it saves us from
+    # having to keep an explicit list of known iOS versions in the code. Return
+    # the results descending order of version number.
+
+    # If the requested major version is less than 12, there won't be any matches.
+    if version[0] < 12:
+        return
+
+    # Consider the actual X.Y version that was requested.
+    yield ios_platform_template.format(
+        major=version[0], minor=version[1], multiarch=multiarch
+    )
+
+    # Consider every minor version from X.0 to the minor version prior to the
+    # version requested by the platform.
+    for minor in range(version[1] - 1, -1, -1):
+        yield ios_platform_template.format(
+            major=version[0], minor=minor, multiarch=multiarch
+        )
+
+    for major in range(version[0] - 1, 11, -1):
+        for minor in range(9, -1, -1):
+            yield ios_platform_template.format(
+                major=major, minor=minor, multiarch=multiarch
+            )
+
+
+def _linux_platforms(is_32bit: bool = _32_BIT_INTERPRETER) -> Iterator[str]:
+    linux = _normalize_string(sysconfig.get_platform())
+    if not linux.startswith("linux_"):
+        # we should never be here, just yield the sysconfig one and return
+        yield linux
+        return
+    if is_32bit:
+        if linux == "linux_x86_64":
+            linux = "linux_i686"
+        elif linux == "linux_aarch64":
+            linux = "linux_armv8l"
+    _, arch = linux.split("_", 1)
+    archs = {"armv8l": ["armv8l", "armv7l"]}.get(arch, [arch])
+    yield from _manylinux.platform_tags(archs)
+    yield from _musllinux.platform_tags(archs)
+    for arch in archs:
+        yield f"linux_{arch}"
+
+
+def _generic_platforms() -> Iterator[str]:
+    yield _normalize_string(sysconfig.get_platform())
+
+
+def platform_tags() -> Iterator[str]:
+    """
+    Provides the platform tags for this installation.
+    """
+    if platform.system() == "Darwin":
+        return mac_platforms()
+    elif platform.system() == "iOS":
+        return ios_platforms()
+    elif platform.system() == "Linux":
+        return _linux_platforms()
+    else:
+        return _generic_platforms()
+
+
+def interpreter_name() -> str:
+    """
+    Returns the name of the running interpreter.
+
+    Some implementations have a reserved, two-letter abbreviation which will
+    be returned when appropriate.
+    """
+    name = sys.implementation.name
+    return INTERPRETER_SHORT_NAMES.get(name) or name
+
+
+def interpreter_version(*, warn: bool = False) -> str:
+    """
+    Returns the version of the running interpreter.
+    """
+    version = _get_config_var("py_version_nodot", warn=warn)
+    if version:
+        version = str(version)
+    else:
+        version = _version_nodot(sys.version_info[:2])
+    return version
+
+
+def _version_nodot(version: PythonVersion) -> str:
+    return "".join(map(str, version))
+
+
+def sys_tags(*, warn: bool = False) -> Iterator[Tag]:
+    """
+    Returns the sequence of tag triples for the running interpreter.
+
+    The order of the sequence corresponds to priority order for the
+    interpreter, from most to least important.
+    """
+
+    interp_name = interpreter_name()
+    if interp_name == "cp":
+        yield from cpython_tags(warn=warn)
+    else:
+        yield from generic_tags()
+
+    if interp_name == "pp":
+        interp = "pp3"
+    elif interp_name == "cp":
+        interp = "cp" + interpreter_version(warn=warn)
+    else:
+        interp = None
+    yield from compatible_tags(interpreter=interp)
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/utils.py b/backend/venv/lib/python3.12/site-packages/packaging/utils.py
new file mode 100644
index 00000000..23450953
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/utils.py
@@ -0,0 +1,163 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import functools
+import re
+from typing import NewType, Tuple, Union, cast
+
+from .tags import Tag, parse_tag
+from .version import InvalidVersion, Version, _TrimmedRelease
+
+BuildTag = Union[Tuple[()], Tuple[int, str]]
+NormalizedName = NewType("NormalizedName", str)
+
+
+class InvalidName(ValueError):
+    """
+    An invalid distribution name; users should refer to the packaging user guide.
+    """
+
+
+class InvalidWheelFilename(ValueError):
+    """
+    An invalid wheel filename was found, users should refer to PEP 427.
+    """
+
+
+class InvalidSdistFilename(ValueError):
+    """
+    An invalid sdist filename was found, users should refer to the packaging user guide.
+    """
+
+
+# Core metadata spec for `Name`
+_validate_regex = re.compile(
+    r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.IGNORECASE
+)
+_canonicalize_regex = re.compile(r"[-_.]+")
+_normalized_regex = re.compile(r"^([a-z0-9]|[a-z0-9]([a-z0-9-](?!--))*[a-z0-9])$")
+# PEP 427: The build number must start with a digit.
+_build_tag_regex = re.compile(r"(\d+)(.*)")
+
+
+def canonicalize_name(name: str, *, validate: bool = False) -> NormalizedName:
+    if validate and not _validate_regex.match(name):
+        raise InvalidName(f"name is invalid: {name!r}")
+    # This is taken from PEP 503.
+    value = _canonicalize_regex.sub("-", name).lower()
+    return cast(NormalizedName, value)
+
+
+def is_normalized_name(name: str) -> bool:
+    return _normalized_regex.match(name) is not None
+
+
+@functools.singledispatch
+def canonicalize_version(
+    version: Version | str, *, strip_trailing_zero: bool = True
+) -> str:
+    """
+    Return a canonical form of a version as a string.
+
+    >>> canonicalize_version('1.0.1')
+    '1.0.1'
+
+    Per PEP 625, versions may have multiple canonical forms, differing
+    only by trailing zeros.
+
+    >>> canonicalize_version('1.0.0')
+    '1'
+    >>> canonicalize_version('1.0.0', strip_trailing_zero=False)
+    '1.0.0'
+
+    Invalid versions are returned unaltered.
+
+    >>> canonicalize_version('foo bar baz')
+    'foo bar baz'
+    """
+    return str(_TrimmedRelease(str(version)) if strip_trailing_zero else version)
+
+
+@canonicalize_version.register
+def _(version: str, *, strip_trailing_zero: bool = True) -> str:
+    try:
+        parsed = Version(version)
+    except InvalidVersion:
+        # Legacy versions cannot be normalized
+        return version
+    return canonicalize_version(parsed, strip_trailing_zero=strip_trailing_zero)
+
+
+def parse_wheel_filename(
+    filename: str,
+) -> tuple[NormalizedName, Version, BuildTag, frozenset[Tag]]:
+    if not filename.endswith(".whl"):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (extension must be '.whl'): {filename!r}"
+        )
+
+    filename = filename[:-4]
+    dashes = filename.count("-")
+    if dashes not in (4, 5):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (wrong number of parts): {filename!r}"
+        )
+
+    parts = filename.split("-", dashes - 2)
+    name_part = parts[0]
+    # See PEP 427 for the rules on escaping the project name.
+    if "__" in name_part or re.match(r"^[\w\d._]*$", name_part, re.UNICODE) is None:
+        raise InvalidWheelFilename(f"Invalid project name: {filename!r}")
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(parts[1])
+    except InvalidVersion as e:
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (invalid version): {filename!r}"
+        ) from e
+
+    if dashes == 5:
+        build_part = parts[2]
+        build_match = _build_tag_regex.match(build_part)
+        if build_match is None:
+            raise InvalidWheelFilename(
+                f"Invalid build number: {build_part} in {filename!r}"
+            )
+        build = cast(BuildTag, (int(build_match.group(1)), build_match.group(2)))
+    else:
+        build = ()
+    tags = parse_tag(parts[-1])
+    return (name, version, build, tags)
+
+
+def parse_sdist_filename(filename: str) -> tuple[NormalizedName, Version]:
+    if filename.endswith(".tar.gz"):
+        file_stem = filename[: -len(".tar.gz")]
+    elif filename.endswith(".zip"):
+        file_stem = filename[: -len(".zip")]
+    else:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (extension must be '.tar.gz' or '.zip'):"
+            f" {filename!r}"
+        )
+
+    # We are requiring a PEP 440 version, which cannot contain dashes,
+    # so we split on the last dash.
+    name_part, sep, version_part = file_stem.rpartition("-")
+    if not sep:
+        raise InvalidSdistFilename(f"Invalid sdist filename: {filename!r}")
+
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(version_part)
+    except InvalidVersion as e:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (invalid version): {filename!r}"
+        ) from e
+
+    return (name, version)
diff --git a/backend/venv/lib/python3.12/site-packages/packaging/version.py b/backend/venv/lib/python3.12/site-packages/packaging/version.py
new file mode 100644
index 00000000..c9bbda20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/packaging/version.py
@@ -0,0 +1,582 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.version import parse, Version
+"""
+
+from __future__ import annotations
+
+import itertools
+import re
+from typing import Any, Callable, NamedTuple, SupportsInt, Tuple, Union
+
+from ._structures import Infinity, InfinityType, NegativeInfinity, NegativeInfinityType
+
+__all__ = ["VERSION_PATTERN", "InvalidVersion", "Version", "parse"]
+
+LocalType = Tuple[Union[int, str], ...]
+
+CmpPrePostDevType = Union[InfinityType, NegativeInfinityType, Tuple[str, int]]
+CmpLocalType = Union[
+    NegativeInfinityType,
+    Tuple[Union[Tuple[int, str], Tuple[NegativeInfinityType, Union[int, str]]], ...],
+]
+CmpKey = Tuple[
+    int,
+    Tuple[int, ...],
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpLocalType,
+]
+VersionComparisonMethod = Callable[[CmpKey, CmpKey], bool]
+
+
+class _Version(NamedTuple):
+    epoch: int
+    release: tuple[int, ...]
+    dev: tuple[str, int] | None
+    pre: tuple[str, int] | None
+    post: tuple[str, int] | None
+    local: LocalType | None
+
+
+def parse(version: str) -> Version:
+    """Parse the given version string.
+
+    >>> parse('1.0.dev1')
+    <Version('1.0.dev1')>
+
+    :param version: The version string to parse.
+    :raises InvalidVersion: When the version string is not a valid version.
+    """
+    return Version(version)
+
+
+class InvalidVersion(ValueError):
+    """Raised when a version string is not a valid version.
+
+    >>> Version("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.version.InvalidVersion: Invalid version: 'invalid'
+    """
+
+
+class _BaseVersion:
+    _key: tuple[Any, ...]
+
+    def __hash__(self) -> int:
+        return hash(self._key)
+
+    # Please keep the duplicated `isinstance` check
+    # in the six comparisons hereunder
+    # unless you find a way to avoid adding overhead function calls.
+    def __lt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key < other._key
+
+    def __le__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key <= other._key
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key == other._key
+
+    def __ge__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key >= other._key
+
+    def __gt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key > other._key
+
+    def __ne__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key != other._key
+
+
+# Deliberately not anchored to the start and end of the string, to make it
+# easier for 3rd party code to reuse
+_VERSION_PATTERN = r"""
+    v?
+    (?:
+        (?:(?P<epoch>[0-9]+)!)?                           # epoch
+        (?P<release>[0-9]+(?:\.[0-9]+)*)                  # release segment
+        (?P<pre>                                          # pre-release
+            [-_\.]?
+            (?P<pre_l>alpha|a|beta|b|preview|pre|c|rc)
+            [-_\.]?
+            (?P<pre_n>[0-9]+)?
+        )?
+        (?P<post>                                         # post release
+            (?:-(?P<post_n1>[0-9]+))
+            |
+            (?:
+                [-_\.]?
+                (?P<post_l>post|rev|r)
+                [-_\.]?
+                (?P<post_n2>[0-9]+)?
+            )
+        )?
+        (?P<dev>                                          # dev release
+            [-_\.]?
+            (?P<dev_l>dev)
+            [-_\.]?
+            (?P<dev_n>[0-9]+)?
+        )?
+    )
+    (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
+"""
+
+VERSION_PATTERN = _VERSION_PATTERN
+"""
+A string containing the regular expression used to match a valid version.
+
+The pattern is not anchored at either end, and is intended for embedding in larger
+expressions (for example, matching a version number as part of a file name). The
+regular expression should be compiled with the ``re.VERBOSE`` and ``re.IGNORECASE``
+flags set.
+
+:meta hide-value:
+"""
+
+
+class Version(_BaseVersion):
+    """This class abstracts handling of a project's versions.
+
+    A :class:`Version` instance is comparison aware and can be compared and
+    sorted using the standard Python interfaces.
+
+    >>> v1 = Version("1.0a5")
+    >>> v2 = Version("1.0")
+    >>> v1
+    <Version('1.0a5')>
+    >>> v2
+    <Version('1.0')>
+    >>> v1 < v2
+    True
+    >>> v1 == v2
+    False
+    >>> v1 > v2
+    False
+    >>> v1 >= v2
+    False
+    >>> v1 <= v2
+    True
+    """
+
+    _regex = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.VERBOSE | re.IGNORECASE)
+    _key: CmpKey
+
+    def __init__(self, version: str) -> None:
+        """Initialize a Version object.
+
+        :param version:
+            The string representation of a version which will be parsed and normalized
+            before use.
+        :raises InvalidVersion:
+            If the ``version`` does not conform to PEP 440 in any way then this
+            exception will be raised.
+        """
+
+        # Validate the version and parse it into pieces
+        match = self._regex.search(version)
+        if not match:
+            raise InvalidVersion(f"Invalid version: {version!r}")
+
+        # Store the parsed out pieces of the version
+        self._version = _Version(
+            epoch=int(match.group("epoch")) if match.group("epoch") else 0,
+            release=tuple(int(i) for i in match.group("release").split(".")),
+            pre=_parse_letter_version(match.group("pre_l"), match.group("pre_n")),
+            post=_parse_letter_version(
+                match.group("post_l"), match.group("post_n1") or match.group("post_n2")
+            ),
+            dev=_parse_letter_version(match.group("dev_l"), match.group("dev_n")),
+            local=_parse_local_version(match.group("local")),
+        )
+
+        # Generate a key which will be used for sorting
+        self._key = _cmpkey(
+            self._version.epoch,
+            self._version.release,
+            self._version.pre,
+            self._version.post,
+            self._version.dev,
+            self._version.local,
+        )
+
+    def __repr__(self) -> str:
+        """A representation of the Version that shows all internal state.
+
+        >>> Version('1.0.0')
+        <Version('1.0.0')>
+        """
+        return f"<Version('{self}')>"
+
+    def __str__(self) -> str:
+        """A string representation of the version that can be round-tripped.
+
+        >>> str(Version("1.0a5"))
+        '1.0a5'
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        # Pre-release
+        if self.pre is not None:
+            parts.append("".join(str(x) for x in self.pre))
+
+        # Post-release
+        if self.post is not None:
+            parts.append(f".post{self.post}")
+
+        # Development release
+        if self.dev is not None:
+            parts.append(f".dev{self.dev}")
+
+        # Local version segment
+        if self.local is not None:
+            parts.append(f"+{self.local}")
+
+        return "".join(parts)
+
+    @property
+    def epoch(self) -> int:
+        """The epoch of the version.
+
+        >>> Version("2.0.0").epoch
+        0
+        >>> Version("1!2.0.0").epoch
+        1
+        """
+        return self._version.epoch
+
+    @property
+    def release(self) -> tuple[int, ...]:
+        """The components of the "release" segment of the version.
+
+        >>> Version("1.2.3").release
+        (1, 2, 3)
+        >>> Version("2.0.0").release
+        (2, 0, 0)
+        >>> Version("1!2.0.0.post0").release
+        (2, 0, 0)
+
+        Includes trailing zeroes but not the epoch or any pre-release / development /
+        post-release suffixes.
+        """
+        return self._version.release
+
+    @property
+    def pre(self) -> tuple[str, int] | None:
+        """The pre-release segment of the version.
+
+        >>> print(Version("1.2.3").pre)
+        None
+        >>> Version("1.2.3a1").pre
+        ('a', 1)
+        >>> Version("1.2.3b1").pre
+        ('b', 1)
+        >>> Version("1.2.3rc1").pre
+        ('rc', 1)
+        """
+        return self._version.pre
+
+    @property
+    def post(self) -> int | None:
+        """The post-release number of the version.
+
+        >>> print(Version("1.2.3").post)
+        None
+        >>> Version("1.2.3.post1").post
+        1
+        """
+        return self._version.post[1] if self._version.post else None
+
+    @property
+    def dev(self) -> int | None:
+        """The development number of the version.
+
+        >>> print(Version("1.2.3").dev)
+        None
+        >>> Version("1.2.3.dev1").dev
+        1
+        """
+        return self._version.dev[1] if self._version.dev else None
+
+    @property
+    def local(self) -> str | None:
+        """The local version segment of the version.
+
+        >>> print(Version("1.2.3").local)
+        None
+        >>> Version("1.2.3+abc").local
+        'abc'
+        """
+        if self._version.local:
+            return ".".join(str(x) for x in self._version.local)
+        else:
+            return None
+
+    @property
+    def public(self) -> str:
+        """The public portion of the version.
+
+        >>> Version("1.2.3").public
+        '1.2.3'
+        >>> Version("1.2.3+abc").public
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").public
+        '1!1.2.3.dev1'
+        """
+        return str(self).split("+", 1)[0]
+
+    @property
+    def base_version(self) -> str:
+        """The "base version" of the version.
+
+        >>> Version("1.2.3").base_version
+        '1.2.3'
+        >>> Version("1.2.3+abc").base_version
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").base_version
+        '1!1.2.3'
+
+        The "base version" is the public version of the project without any pre or post
+        release markers.
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        return "".join(parts)
+
+    @property
+    def is_prerelease(self) -> bool:
+        """Whether this version is a pre-release.
+
+        >>> Version("1.2.3").is_prerelease
+        False
+        >>> Version("1.2.3a1").is_prerelease
+        True
+        >>> Version("1.2.3b1").is_prerelease
+        True
+        >>> Version("1.2.3rc1").is_prerelease
+        True
+        >>> Version("1.2.3dev1").is_prerelease
+        True
+        """
+        return self.dev is not None or self.pre is not None
+
+    @property
+    def is_postrelease(self) -> bool:
+        """Whether this version is a post-release.
+
+        >>> Version("1.2.3").is_postrelease
+        False
+        >>> Version("1.2.3.post1").is_postrelease
+        True
+        """
+        return self.post is not None
+
+    @property
+    def is_devrelease(self) -> bool:
+        """Whether this version is a development release.
+
+        >>> Version("1.2.3").is_devrelease
+        False
+        >>> Version("1.2.3.dev1").is_devrelease
+        True
+        """
+        return self.dev is not None
+
+    @property
+    def major(self) -> int:
+        """The first item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").major
+        1
+        """
+        return self.release[0] if len(self.release) >= 1 else 0
+
+    @property
+    def minor(self) -> int:
+        """The second item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").minor
+        2
+        >>> Version("1").minor
+        0
+        """
+        return self.release[1] if len(self.release) >= 2 else 0
+
+    @property
+    def micro(self) -> int:
+        """The third item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").micro
+        3
+        >>> Version("1").micro
+        0
+        """
+        return self.release[2] if len(self.release) >= 3 else 0
+
+
+class _TrimmedRelease(Version):
+    @property
+    def release(self) -> tuple[int, ...]:
+        """
+        Release segment without any trailing zeros.
+
+        >>> _TrimmedRelease('1.0.0').release
+        (1,)
+        >>> _TrimmedRelease('0.0').release
+        (0,)
+        """
+        rel = super().release
+        nonzeros = (index for index, val in enumerate(rel) if val)
+        last_nonzero = max(nonzeros, default=0)
+        return rel[: last_nonzero + 1]
+
+
+def _parse_letter_version(
+    letter: str | None, number: str | bytes | SupportsInt | None
+) -> tuple[str, int] | None:
+    if letter:
+        # We consider there to be an implicit 0 in a pre-release if there is
+        # not a numeral associated with it.
+        if number is None:
+            number = 0
+
+        # We normalize any letters to their lower case form
+        letter = letter.lower()
+
+        # We consider some words to be alternate spellings of other words and
+        # in those cases we want to normalize the spellings to our preferred
+        # spelling.
+        if letter == "alpha":
+            letter = "a"
+        elif letter == "beta":
+            letter = "b"
+        elif letter in ["c", "pre", "preview"]:
+            letter = "rc"
+        elif letter in ["rev", "r"]:
+            letter = "post"
+
+        return letter, int(number)
+
+    assert not letter
+    if number:
+        # We assume if we are given a number, but we are not given a letter
+        # then this is using the implicit post release syntax (e.g. 1.0-1)
+        letter = "post"
+
+        return letter, int(number)
+
+    return None
+
+
+_local_version_separators = re.compile(r"[\._-]")
+
+
+def _parse_local_version(local: str | None) -> LocalType | None:
+    """
+    Takes a string like abc.1.twelve and turns it into ("abc", 1, "twelve").
+    """
+    if local is not None:
+        return tuple(
+            part.lower() if not part.isdigit() else int(part)
+            for part in _local_version_separators.split(local)
+        )
+    return None
+
+
+def _cmpkey(
+    epoch: int,
+    release: tuple[int, ...],
+    pre: tuple[str, int] | None,
+    post: tuple[str, int] | None,
+    dev: tuple[str, int] | None,
+    local: LocalType | None,
+) -> CmpKey:
+    # When we compare a release version, we want to compare it with all of the
+    # trailing zeros removed. So we'll use a reverse the list, drop all the now
+    # leading zeros until we come to something non zero, then take the rest
+    # re-reverse it back into the correct order and make it a tuple and use
+    # that for our sorting key.
+    _release = tuple(
+        reversed(list(itertools.dropwhile(lambda x: x == 0, reversed(release))))
+    )
+
+    # We need to "trick" the sorting algorithm to put 1.0.dev0 before 1.0a0.
+    # We'll do this by abusing the pre segment, but we _only_ want to do this
+    # if there is not a pre or a post segment. If we have one of those then
+    # the normal sorting rules will handle this case correctly.
+    if pre is None and post is None and dev is not None:
+        _pre: CmpPrePostDevType = NegativeInfinity
+    # Versions without a pre-release (except as noted above) should sort after
+    # those with one.
+    elif pre is None:
+        _pre = Infinity
+    else:
+        _pre = pre
+
+    # Versions without a post segment should sort before those with one.
+    if post is None:
+        _post: CmpPrePostDevType = NegativeInfinity
+
+    else:
+        _post = post
+
+    # Versions without a development segment should sort after those with one.
+    if dev is None:
+        _dev: CmpPrePostDevType = Infinity
+
+    else:
+        _dev = dev
+
+    if local is None:
+        # Versions without a local segment should sort before those with one.
+        _local: CmpLocalType = NegativeInfinity
+    else:
+        # Versions with a local segment need that segment parsed to implement
+        # the sorting rules in PEP440.
+        # - Alpha numeric segments sort before numeric segments
+        # - Alpha numeric segments sort lexicographically
+        # - Numeric segments sort numerically
+        # - Shorter versions sort before longer versions when the prefixes
+        #   match exactly
+        _local = tuple(
+            (i, "") if isinstance(i, int) else (NegativeInfinity, i) for i in local
+        )
+
+    return epoch, _release, _pre, _post, _dev, _local
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/LICENSE
new file mode 100644
index 00000000..6c73b16c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/LICENSE
@@ -0,0 +1,19 @@
+Copyright (c) 2012-2019 Richard Jones <richard@python.org>
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+ The above copyright notice and this permission notice shall be included in
+ all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/METADATA
new file mode 100644
index 00000000..6794e695
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/METADATA
@@ -0,0 +1,564 @@
+Metadata-Version: 2.1
+Name: parse
+Version: 1.20.2
+Summary: parse() is the opposite of format()
+Author-email: Richard Jones <richard@python.org>
+Maintainer-email: Wim Glenn <hey@wimglenn.com>
+License: Copyright (c) 2012-2019 Richard Jones <richard@python.org>
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+         The above copyright notice and this permission notice shall be included in
+         all copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: homepage, https://github.com/r1chardj0n3s/parse
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Software Development :: Code Generators
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: License :: OSI Approved :: MIT License
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+
+Installation
+------------
+
+.. code-block:: pycon
+
+    pip install parse
+
+Usage
+-----
+
+Parse strings using a specification based on the Python `format()`_ syntax.
+
+   ``parse()`` is the opposite of ``format()``
+
+The module is set up to only export ``parse()``, ``search()``, ``findall()``,
+and ``with_pattern()`` when ``import *`` is used:
+
+>>> from parse import *
+
+From there it's a simple thing to parse a string:
+
+.. code-block:: pycon
+
+    >>> parse("It's {}, I love it!", "It's spam, I love it!")
+    <Result ('spam',) {}>
+    >>> _[0]
+    'spam'
+
+Or to search a string for some pattern:
+
+.. code-block:: pycon
+
+    >>> search('Age: {:d}\n', 'Name: Rufus\nAge: 42\nColor: red\n')
+    <Result (42,) {}>
+
+Or find all the occurrences of some pattern in a string:
+
+.. code-block:: pycon
+
+    >>> ''.join(r[0] for r in findall(">{}<", "<p>the <b>bold</b> text</p>"))
+    'the bold text'
+
+If you're going to use the same pattern to match lots of strings you can
+compile it once:
+
+.. code-block:: pycon
+
+    >>> from parse import compile
+    >>> p = compile("It's {}, I love it!")
+    >>> print(p)
+    <Parser "It's {}, I love it!">
+    >>> p.parse("It's spam, I love it!")
+    <Result ('spam',) {}>
+
+("compile" is not exported for ``import *`` usage as it would override the
+built-in ``compile()`` function)
+
+The default behaviour is to match strings case insensitively. You may match with
+case by specifying `case_sensitive=True`:
+
+.. code-block:: pycon
+
+    >>> parse('SPAM', 'spam', case_sensitive=True) is None
+    True
+
+.. _format():
+  https://docs.python.org/3/library/stdtypes.html#str.format
+
+
+Format Syntax
+-------------
+
+A basic version of the `Format String Syntax`_ is supported with anonymous
+(fixed-position), named and formatted fields::
+
+   {[field name]:[format spec]}
+
+Field names must be a valid Python identifiers, including dotted names;
+element indexes imply dictionaries (see below for example).
+
+Numbered fields are also not supported: the result of parsing will include
+the parsed fields in the order they are parsed.
+
+The conversion of fields to types other than strings is done based on the
+type in the format specification, which mirrors the ``format()`` behaviour.
+There are no "!" field conversions like ``format()`` has.
+
+Some simple parse() format string examples:
+
+.. code-block:: pycon
+
+    >>> parse("Bring me a {}", "Bring me a shrubbery")
+    <Result ('shrubbery',) {}>
+    >>> r = parse("The {} who {} {}", "The knights who say Ni!")
+    >>> print(r)
+    <Result ('knights', 'say', 'Ni!') {}>
+    >>> print(r.fixed)
+    ('knights', 'say', 'Ni!')
+    >>> print(r[0])
+    knights
+    >>> print(r[1:])
+    ('say', 'Ni!')
+    >>> r = parse("Bring out the holy {item}", "Bring out the holy hand grenade")
+    >>> print(r)
+    <Result () {'item': 'hand grenade'}>
+    >>> print(r.named)
+    {'item': 'hand grenade'}
+    >>> print(r['item'])
+    hand grenade
+    >>> 'item' in r
+    True
+
+Note that `in` only works if you have named fields.
+
+Dotted names and indexes are possible with some limits. Only word identifiers
+are supported (ie. no numeric indexes) and the application must make additional
+sense of the result:
+
+.. code-block:: pycon
+
+    >>> r = parse("Mmm, {food.type}, I love it!", "Mmm, spam, I love it!")
+    >>> print(r)
+    <Result () {'food.type': 'spam'}>
+    >>> print(r.named)
+    {'food.type': 'spam'}
+    >>> print(r['food.type'])
+    spam
+    >>> r = parse("My quest is {quest[name]}", "My quest is to seek the holy grail!")
+    >>> print(r)
+    <Result () {'quest': {'name': 'to seek the holy grail!'}}>
+    >>> print(r['quest'])
+    {'name': 'to seek the holy grail!'}
+    >>> print(r['quest']['name'])
+    to seek the holy grail!
+
+If the text you're matching has braces in it you can match those by including
+a double-brace ``{{`` or ``}}`` in your format string, just like format() does.
+
+
+Format Specification
+--------------------
+
+Most often a straight format-less ``{}`` will suffice where a more complex
+format specification might have been used.
+
+Most of `format()`'s `Format Specification Mini-Language`_ is supported:
+
+   [[fill]align][sign][0][width][.precision][type]
+
+The differences between `parse()` and `format()` are:
+
+- The align operators will cause spaces (or specified fill character) to be
+  stripped from the parsed value. The width is not enforced; it just indicates
+  there may be whitespace or "0"s to strip.
+- Numeric parsing will automatically handle a "0b", "0o" or "0x" prefix.
+  That is, the "#" format character is handled automatically by d, b, o
+  and x formats. For "d" any will be accepted, but for the others the correct
+  prefix must be present if at all.
+- Numeric sign is handled automatically.  A sign specifier can be given, but
+  has no effect.
+- The thousands separator is handled automatically if the "n" type is used.
+- The types supported are a slightly different mix to the format() types.  Some
+  format() types come directly over: "d", "n", "%", "f", "e", "b", "o" and "x".
+  In addition some regular expression character group types "D", "w", "W", "s"
+  and "S" are also available.
+- The "e" and "g" types are case-insensitive so there is not need for
+  the "E" or "G" types. The "e" type handles Fortran formatted numbers (no
+  leading 0 before the decimal point).
+
+===== =========================================== ========
+Type  Characters Matched                          Output
+===== =========================================== ========
+l     Letters (ASCII)                             str
+w     Letters, numbers and underscore             str
+W     Not letters, numbers and underscore         str
+s     Whitespace                                  str
+S     Non-whitespace                              str
+d     Digits (effectively integer numbers)        int
+D     Non-digit                                   str
+n     Numbers with thousands separators (, or .)  int
+%     Percentage (converted to value/100.0)       float
+f     Fixed-point numbers                         float
+F     Decimal numbers                             Decimal
+e     Floating-point numbers with exponent        float
+      e.g. 1.1e-10, NAN (all case insensitive)
+g     General number format (either d, f or e)    float
+b     Binary numbers                              int
+o     Octal numbers                               int
+x     Hexadecimal numbers (lower and upper case)  int
+ti    ISO 8601 format date/time                   datetime
+      e.g. 1972-01-20T10:21:36Z ("T" and "Z"
+      optional)
+te    RFC2822 e-mail format date/time             datetime
+      e.g. Mon, 20 Jan 1972 10:21:36 +1000
+tg    Global (day/month) format date/time         datetime
+      e.g. 20/1/1972 10:21:36 AM +1:00
+ta    US (month/day) format date/time             datetime
+      e.g. 1/20/1972 10:21:36 PM +10:30
+tc    ctime() format date/time                    datetime
+      e.g. Sun Sep 16 01:03:52 1973
+th    HTTP log format date/time                   datetime
+      e.g. 21/Nov/2011:00:07:11 +0000
+ts    Linux system log format date/time           datetime
+      e.g. Nov  9 03:37:44
+tt    Time                                        time
+      e.g. 10:21:36 PM -5:30
+===== =========================================== ========
+
+The type can also be a datetime format string, following the
+`1989 C standard format codes`_, e.g. ``%Y-%m-%d``. Depending on the
+directives contained in the format string, parsed output may be an instance
+of ``datetime.datetime``, ``datetime.time``, or ``datetime.date``.
+
+.. code-block:: pycon
+
+    >>> parse("{:%Y-%m-%d %H:%M:%S}", "2023-11-23 12:56:47")
+    <Result (datetime.datetime(2023, 11, 23, 12, 56, 47),) {}>
+    >>> parse("{:%H:%M}", "10:26")
+    <Result (datetime.time(10, 26),) {}>
+    >>> parse("{:%Y/%m/%d}", "2023/11/25")
+    <Result (datetime.date(2023, 11, 25),) {}>
+
+
+Some examples of typed parsing with ``None`` returned if the typing
+does not match:
+
+.. code-block:: pycon
+
+    >>> parse('Our {:d} {:w} are...', 'Our 3 weapons are...')
+    <Result (3, 'weapons') {}>
+    >>> parse('Our {:d} {:w} are...', 'Our three weapons are...')
+    >>> parse('Meet at {:tg}', 'Meet at 1/2/2011 11:00 PM')
+    <Result (datetime.datetime(2011, 2, 1, 23, 0),) {}>
+
+And messing about with alignment:
+
+.. code-block:: pycon
+
+    >>> parse('with {:>} herring', 'with     a herring')
+    <Result ('a',) {}>
+    >>> parse('spam {:^} spam', 'spam    lovely     spam')
+    <Result ('lovely',) {}>
+
+Note that the "center" alignment does not test to make sure the value is
+centered - it just strips leading and trailing whitespace.
+
+Width and precision may be used to restrict the size of matched text
+from the input. Width specifies a minimum size and precision specifies
+a maximum. For example:
+
+.. code-block:: pycon
+
+    >>> parse('{:.2}{:.2}', 'look')           # specifying precision
+    <Result ('lo', 'ok') {}>
+    >>> parse('{:4}{:4}', 'look at that')     # specifying width
+    <Result ('look', 'at that') {}>
+    >>> parse('{:4}{:.4}', 'look at that')    # specifying both
+    <Result ('look at ', 'that') {}>
+    >>> parse('{:2d}{:2d}', '0440')           # parsing two contiguous numbers
+    <Result (4, 40) {}>
+
+Some notes for the special date and time types:
+
+- the presence of the time part is optional (including ISO 8601, starting
+  at the "T"). A full datetime object will always be returned; the time
+  will be set to 00:00:00. You may also specify a time without seconds.
+- when a seconds amount is present in the input fractions will be parsed
+  to give microseconds.
+- except in ISO 8601 the day and month digits may be 0-padded.
+- the date separator for the tg and ta formats may be "-" or "/".
+- named months (abbreviations or full names) may be used in the ta and tg
+  formats in place of numeric months.
+- as per RFC 2822 the e-mail format may omit the day (and comma), and the
+  seconds but nothing else.
+- hours greater than 12 will be happily accepted.
+- the AM/PM are optional, and if PM is found then 12 hours will be added
+  to the datetime object's hours amount - even if the hour is greater
+  than 12 (for consistency.)
+- in ISO 8601 the "Z" (UTC) timezone part may be a numeric offset
+- timezones are specified as "+HH:MM" or "-HH:MM". The hour may be one or two
+  digits (0-padded is OK.) Also, the ":" is optional.
+- the timezone is optional in all except the e-mail format (it defaults to
+  UTC.)
+- named timezones are not handled yet.
+
+Note: attempting to match too many datetime fields in a single parse() will
+currently result in a resource allocation issue. A TooManyFields exception
+will be raised in this instance. The current limit is about 15. It is hoped
+that this limit will be removed one day.
+
+.. _`Format String Syntax`:
+  https://docs.python.org/3/library/string.html#format-string-syntax
+.. _`Format Specification Mini-Language`:
+  https://docs.python.org/3/library/string.html#format-specification-mini-language
+.. _`1989 C standard format codes`:
+  https://docs.python.org/3/library/datetime.html#strftime-and-strptime-format-codes
+
+
+
+Result and Match Objects
+------------------------
+
+The result of a ``parse()`` and ``search()`` operation is either ``None`` (no match), a
+``Result`` instance or a ``Match`` instance if ``evaluate_result`` is False.
+
+The ``Result`` instance has three attributes:
+
+``fixed``
+   A tuple of the fixed-position, anonymous fields extracted from the input.
+``named``
+   A dictionary of the named fields extracted from the input.
+``spans``
+   A dictionary mapping the names and fixed position indices matched to a
+   2-tuple slice range of where the match occurred in the input.
+   The span does not include any stripped padding (alignment or width).
+
+The ``Match`` instance has one method:
+
+``evaluate_result()``
+   Generates and returns a ``Result`` instance for this ``Match`` object.
+
+
+
+Custom Type Conversions
+-----------------------
+
+If you wish to have matched fields automatically converted to your own type you
+may pass in a dictionary of type conversion information to ``parse()`` and
+``compile()``.
+
+The converter will be passed the field string matched. Whatever it returns
+will be substituted in the ``Result`` instance for that field.
+
+Your custom type conversions may override the builtin types if you supply one
+with the same identifier:
+
+.. code-block:: pycon
+
+    >>> def shouty(string):
+    ...    return string.upper()
+    ...
+    >>> parse('{:shouty} world', 'hello world', {"shouty": shouty})
+    <Result ('HELLO',) {}>
+
+If the type converter has the optional ``pattern`` attribute, it is used as
+regular expression for better pattern matching (instead of the default one):
+
+.. code-block:: pycon
+
+    >>> def parse_number(text):
+    ...    return int(text)
+    >>> parse_number.pattern = r'\d+'
+    >>> parse('Answer: {number:Number}', 'Answer: 42', {"Number": parse_number})
+    <Result () {'number': 42}>
+    >>> _ = parse('Answer: {:Number}', 'Answer: Alice', {"Number": parse_number})
+    >>> assert _ is None, "MISMATCH"
+
+You can also use the ``with_pattern(pattern)`` decorator to add this
+information to a type converter function:
+
+.. code-block:: pycon
+
+    >>> from parse import with_pattern
+    >>> @with_pattern(r'\d+')
+    ... def parse_number(text):
+    ...    return int(text)
+    >>> parse('Answer: {number:Number}', 'Answer: 42', {"Number": parse_number})
+    <Result () {'number': 42}>
+
+A more complete example of a custom type might be:
+
+.. code-block:: pycon
+
+    >>> yesno_mapping = {
+    ...     "yes":  True,   "no":    False,
+    ...     "on":   True,   "off":   False,
+    ...     "true": True,   "false": False,
+    ... }
+    >>> @with_pattern(r"|".join(yesno_mapping))
+    ... def parse_yesno(text):
+    ...     return yesno_mapping[text.lower()]
+
+
+If the type converter ``pattern`` uses regex-grouping (with parenthesis),
+you should indicate this by using the optional ``regex_group_count`` parameter
+in the ``with_pattern()`` decorator:
+
+.. code-block:: pycon
+
+    >>> @with_pattern(r'((\d+))', regex_group_count=2)
+    ... def parse_number2(text):
+    ...    return int(text)
+    >>> parse('Answer: {:Number2} {:Number2}', 'Answer: 42 43', {"Number2": parse_number2})
+    <Result (42, 43) {}>
+
+Otherwise, this may cause parsing problems with unnamed/fixed parameters.
+
+
+Potential Gotchas
+-----------------
+
+``parse()`` will always match the shortest text necessary (from left to right)
+to fulfil the parse pattern, so for example:
+
+
+.. code-block:: pycon
+
+    >>> pattern = '{dir1}/{dir2}'
+    >>> data = 'root/parent/subdir'
+    >>> sorted(parse(pattern, data).named.items())
+    [('dir1', 'root'), ('dir2', 'parent/subdir')]
+
+So, even though `{'dir1': 'root/parent', 'dir2': 'subdir'}` would also fit
+the pattern, the actual match represents the shortest successful match for
+``dir1``.
+
+Developers
+----------
+
+Want to contribute to parse? Fork the repo to your own GitHub account, and create a pull-request.
+
+.. code-block:: bash
+
+   git clone git@github.com:r1chardj0n3s/parse.git
+   git remote rename origin upstream
+   git remote add origin git@github.com:YOURUSERNAME/parse.git
+   git checkout -b myfeature
+
+To run the tests locally:
+
+.. code-block:: bash
+
+   python -m venv .venv
+   source .venv/bin/activate
+   pip install -r tests/requirements.txt
+   pip install -e .
+   pytest
+
+----
+
+Changelog
+---------
+
+- 1.20.2 Template field names can now contain - character i.e. HYPHEN-MINUS, chr(0x2d)
+- 1.20.1 The `%f` directive accepts 1-6 digits, like strptime (thanks @bbertincourt)
+- 1.20.0 Added support for strptime codes (thanks @bendichter)
+- 1.19.1 Added support for sign specifiers in number formats (thanks @anntzer)
+- 1.19.0 Added slice access to fixed results (thanks @jonathangjertsen).
+  Also corrected matching of *full string* vs. *full line* (thanks @giladreti)
+  Fix issue with using digit field numbering and types
+- 1.18.0 Correct bug in int parsing introduced in 1.16.0 (thanks @maxxk)
+- 1.17.0 Make left- and center-aligned search consume up to next space
+- 1.16.0 Make compiled parse objects pickleable (thanks @martinResearch)
+- 1.15.0 Several fixes for parsing non-base 10 numbers (thanks @vladikcomper)
+- 1.14.0 More broad acceptance of Fortran number format (thanks @purpleskyfall)
+- 1.13.1 Project metadata correction.
+- 1.13.0 Handle Fortran formatted numbers with no leading 0 before decimal
+  point (thanks @purpleskyfall).
+  Handle comparison of FixedTzOffset with other types of object.
+- 1.12.1 Actually use the `case_sensitive` arg in compile (thanks @jacquev6)
+- 1.12.0 Do not assume closing brace when an opening one is found (thanks @mattsep)
+- 1.11.1 Revert having unicode char in docstring, it breaks Bamboo builds(?!)
+- 1.11.0 Implement `__contains__` for Result instances.
+- 1.10.0 Introduce a "letters" matcher, since "w" matches numbers
+  also.
+- 1.9.1 Fix deprecation warnings around backslashes in regex strings
+  (thanks Mickael Schoentgen). Also fix some documentation formatting
+  issues.
+- 1.9.0 We now honor precision and width specifiers when parsing numbers
+  and strings, allowing parsing of concatenated elements of fixed width
+  (thanks Julia Signell)
+- 1.8.4 Add LICENSE file at request of packagers.
+  Correct handling of AM/PM to follow most common interpretation.
+  Correct parsing of hexadecimal that looks like a binary prefix.
+  Add ability to parse case sensitively.
+  Add parsing of numbers to Decimal with "F" (thanks John Vandenberg)
+- 1.8.3 Add regex_group_count to with_pattern() decorator to support
+  user-defined types that contain brackets/parenthesis (thanks Jens Engel)
+- 1.8.2 add documentation for including braces in format string
+- 1.8.1 ensure bare hexadecimal digits are not matched
+- 1.8.0 support manual control over result evaluation (thanks Timo Furrer)
+- 1.7.0 parse dict fields (thanks Mark Visser) and adapted to allow
+  more than 100 re groups in Python 3.5+ (thanks David King)
+- 1.6.6 parse Linux system log dates (thanks Alex Cowan)
+- 1.6.5 handle precision in float format (thanks Levi Kilcher)
+- 1.6.4 handle pipe "|" characters in parse string (thanks Martijn Pieters)
+- 1.6.3 handle repeated instances of named fields, fix bug in PM time
+  overflow
+- 1.6.2 fix logging to use local, not root logger (thanks Necku)
+- 1.6.1 be more flexible regarding matched ISO datetimes and timezones in
+  general, fix bug in timezones without ":" and improve docs
+- 1.6.0 add support for optional ``pattern`` attribute in user-defined types
+  (thanks Jens Engel)
+- 1.5.3 fix handling of question marks
+- 1.5.2 fix type conversion error with dotted names (thanks Sebastian Thiel)
+- 1.5.1 implement handling of named datetime fields
+- 1.5 add handling of dotted field names (thanks Sebastian Thiel)
+- 1.4.1 fix parsing of "0" in int conversion (thanks James Rowe)
+- 1.4 add __getitem__ convenience access on Result.
+- 1.3.3 fix Python 2.5 setup.py issue.
+- 1.3.2 fix Python 3.2 setup.py issue.
+- 1.3.1 fix a couple of Python 3.2 compatibility issues.
+- 1.3 added search() and findall(); removed compile() from ``import *``
+  export as it overwrites builtin.
+- 1.2 added ability for custom and override type conversions to be
+  provided; some cleanup
+- 1.1.9 to keep things simpler number sign is handled automatically;
+  significant robustification in the face of edge-case input.
+- 1.1.8 allow "d" fields to have number base "0x" etc. prefixes;
+  fix up some field type interactions after stress-testing the parser;
+  implement "%" type.
+- 1.1.7 Python 3 compatibility tweaks (2.5 to 2.7 and 3.2 are supported).
+- 1.1.6 add "e" and "g" field types; removed redundant "h" and "X";
+  removed need for explicit "#".
+- 1.1.5 accept textual dates in more places; Result now holds match span
+  positions.
+- 1.1.4 fixes to some int type conversion; implemented "=" alignment; added
+  date/time parsing with a variety of formats handled.
+- 1.1.3 type conversion is automatic based on specified field types. Also added
+  "f" and "n" types.
+- 1.1.2 refactored, added compile() and limited ``from parse import *``
+- 1.1.1 documentation improvements
+- 1.1.0 implemented more of the `Format Specification Mini-Language`_
+  and removed the restriction on mixing fixed-position and named fields
+- 1.0.0 initial release
+
+This code is copyright 2012-2021 Richard Jones <richard@python.org>
+See the end of the source file for the license of use.
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/RECORD
new file mode 100644
index 00000000..c79e1485
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/RECORD
@@ -0,0 +1,8 @@
+__pycache__/parse.cpython-312.pyc,,
+parse-1.20.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+parse-1.20.2.dist-info/LICENSE,sha256=LMmUL-P1KmacH7tOC66aJzANVlZKOC6VSEXDJWxMq-I,1085
+parse-1.20.2.dist-info/METADATA,sha256=-hBJcA8T1uHy0cqvnEmSXWoXnHfgF12VwboHOom_MqM,22134
+parse-1.20.2.dist-info/RECORD,,
+parse-1.20.2.dist-info/WHEEL,sha256=DZajD4pwLWue70CAfc7YaxT1wLUciNBvN_TTcvXpltE,110
+parse-1.20.2.dist-info/top_level.txt,sha256=EQ7GYrwhphlri0fKN9z9qBS7eZZMRhj7jeRlmhrawN8,6
+parse.py,sha256=4dF5Mct5SWR9D_xXsnCgUcxa9Cx-TQA7LAYed29bCWU,35530
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/WHEEL
new file mode 100644
index 00000000..832be111
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/WHEEL
@@ -0,0 +1,6 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/top_level.txt
new file mode 100644
index 00000000..d8dc22d1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse-1.20.2.dist-info/top_level.txt
@@ -0,0 +1 @@
+parse
diff --git a/backend/venv/lib/python3.12/site-packages/parse.py b/backend/venv/lib/python3.12/site-packages/parse.py
new file mode 100644
index 00000000..51a6953c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse.py
@@ -0,0 +1,1081 @@
+from __future__ import absolute_import
+
+import logging
+import re
+import sys
+from datetime import datetime
+from datetime import time
+from datetime import timedelta
+from datetime import tzinfo
+from decimal import Decimal
+from functools import partial
+
+
+__version__ = "1.20.2"
+__all__ = ["parse", "search", "findall", "with_pattern"]
+
+log = logging.getLogger(__name__)
+
+
+def with_pattern(pattern, regex_group_count=None):
+    r"""Attach a regular expression pattern matcher to a custom type converter
+    function.
+
+    This annotates the type converter with the :attr:`pattern` attribute.
+
+    EXAMPLE:
+        >>> import parse
+        >>> @parse.with_pattern(r"\d+")
+        ... def parse_number(text):
+        ...     return int(text)
+
+    is equivalent to:
+
+        >>> def parse_number(text):
+        ...     return int(text)
+        >>> parse_number.pattern = r"\d+"
+
+    :param pattern: regular expression pattern (as text)
+    :param regex_group_count: Indicates how many regex-groups are in pattern.
+    :return: wrapped function
+    """
+
+    def decorator(func):
+        func.pattern = pattern
+        func.regex_group_count = regex_group_count
+        return func
+
+    return decorator
+
+
+class int_convert:
+    """Convert a string to an integer.
+
+    The string may start with a sign.
+
+    It may be of a base other than 2, 8, 10 or 16.
+
+    If base isn't specified, it will be detected automatically based
+    on a string format. When string starts with a base indicator, 0#nnnn,
+    it overrides the default base of 10.
+
+    It may also have other non-numeric characters that we can ignore.
+    """
+
+    CHARS = "0123456789abcdefghijklmnopqrstuvwxyz"
+
+    def __init__(self, base=None):
+        self.base = base
+
+    def __call__(self, string, match):
+        if string[0] == "-":
+            sign = -1
+            number_start = 1
+        elif string[0] == "+":
+            sign = 1
+            number_start = 1
+        else:
+            sign = 1
+            number_start = 0
+
+        base = self.base
+        # If base wasn't specified, detect it automatically
+        if base is None:
+            # Assume decimal number, unless different base is detected
+            base = 10
+
+            # For number formats starting with 0b, 0o, 0x, use corresponding base ...
+            if string[number_start] == "0" and len(string) - number_start > 2:
+                if string[number_start + 1] in "bB":
+                    base = 2
+                elif string[number_start + 1] in "oO":
+                    base = 8
+                elif string[number_start + 1] in "xX":
+                    base = 16
+
+        chars = int_convert.CHARS[:base]
+        string = re.sub("[^%s]" % chars, "", string.lower())
+        return sign * int(string, base)
+
+
+class convert_first:
+    """Convert the first element of a pair.
+    This equivalent to lambda s,m: converter(s). But unlike a lambda function, it can be pickled
+    """
+
+    def __init__(self, converter):
+        self.converter = converter
+
+    def __call__(self, string, match):
+        return self.converter(string)
+
+
+def percentage(string, match):
+    return float(string[:-1]) / 100.0
+
+
+class FixedTzOffset(tzinfo):
+    """Fixed offset in minutes east from UTC."""
+
+    ZERO = timedelta(0)
+
+    def __init__(self, offset, name):
+        self._offset = timedelta(minutes=offset)
+        self._name = name
+
+    def __repr__(self):
+        return "<%s %s %s>" % (self.__class__.__name__, self._name, self._offset)
+
+    def utcoffset(self, dt):
+        return self._offset
+
+    def tzname(self, dt):
+        return self._name
+
+    def dst(self, dt):
+        return self.ZERO
+
+    def __eq__(self, other):
+        if not isinstance(other, FixedTzOffset):
+            return NotImplemented
+        return self._name == other._name and self._offset == other._offset
+
+
+MONTHS_MAP = {
+    "Jan": 1,
+    "January": 1,
+    "Feb": 2,
+    "February": 2,
+    "Mar": 3,
+    "March": 3,
+    "Apr": 4,
+    "April": 4,
+    "May": 5,
+    "Jun": 6,
+    "June": 6,
+    "Jul": 7,
+    "July": 7,
+    "Aug": 8,
+    "August": 8,
+    "Sep": 9,
+    "September": 9,
+    "Oct": 10,
+    "October": 10,
+    "Nov": 11,
+    "November": 11,
+    "Dec": 12,
+    "December": 12,
+}
+DAYS_PAT = r"(Mon|Tue|Wed|Thu|Fri|Sat|Sun)"
+MONTHS_PAT = r"(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)"
+ALL_MONTHS_PAT = r"(%s)" % "|".join(MONTHS_MAP)
+TIME_PAT = r"(\d{1,2}:\d{1,2}(:\d{1,2}(\.\d+)?)?)"
+AM_PAT = r"(\s+[AP]M)"
+TZ_PAT = r"(\s+[-+]\d\d?:?\d\d)"
+
+
+def date_convert(
+    string,
+    match,
+    ymd=None,
+    mdy=None,
+    dmy=None,
+    d_m_y=None,
+    hms=None,
+    am=None,
+    tz=None,
+    mm=None,
+    dd=None,
+):
+    """Convert the incoming string containing some date / time info into a
+    datetime instance.
+    """
+    groups = match.groups()
+    time_only = False
+    if mm and dd:
+        y = datetime.today().year
+        m = groups[mm]
+        d = groups[dd]
+    elif ymd is not None:
+        y, m, d = re.split(r"[-/\s]", groups[ymd])
+    elif mdy is not None:
+        m, d, y = re.split(r"[-/\s]", groups[mdy])
+    elif dmy is not None:
+        d, m, y = re.split(r"[-/\s]", groups[dmy])
+    elif d_m_y is not None:
+        d, m, y = d_m_y
+        d = groups[d]
+        m = groups[m]
+        y = groups[y]
+    else:
+        time_only = True
+
+    H = M = S = u = 0
+    if hms is not None and groups[hms]:
+        t = groups[hms].split(":")
+        if len(t) == 2:
+            H, M = t
+        else:
+            H, M, S = t
+            if "." in S:
+                S, u = S.split(".")
+                u = int(float("." + u) * 1000000)
+            S = int(S)
+        H = int(H)
+        M = int(M)
+
+    if am is not None:
+        am = groups[am]
+        if am:
+            am = am.strip()
+        if am == "AM" and H == 12:
+            # correction for "12" hour functioning as "0" hour: 12:15 AM = 00:15 by 24 hr clock
+            H -= 12
+        elif am == "PM" and H == 12:
+            # no correction needed: 12PM is midday, 12:00 by 24 hour clock
+            pass
+        elif am == "PM":
+            H += 12
+
+    if tz is not None:
+        tz = groups[tz]
+    if tz == "Z":
+        tz = FixedTzOffset(0, "UTC")
+    elif tz:
+        tz = tz.strip()
+        if tz.isupper():
+            # TODO use the awesome python TZ module?
+            pass
+        else:
+            sign = tz[0]
+            if ":" in tz:
+                tzh, tzm = tz[1:].split(":")
+            elif len(tz) == 4:  # 'snnn'
+                tzh, tzm = tz[1], tz[2:4]
+            else:
+                tzh, tzm = tz[1:3], tz[3:5]
+            offset = int(tzm) + int(tzh) * 60
+            if sign == "-":
+                offset = -offset
+            tz = FixedTzOffset(offset, tz)
+
+    if time_only:
+        d = time(H, M, S, u, tzinfo=tz)
+    else:
+        y = int(y)
+        if m.isdigit():
+            m = int(m)
+        else:
+            m = MONTHS_MAP[m]
+        d = int(d)
+        d = datetime(y, m, d, H, M, S, u, tzinfo=tz)
+
+    return d
+
+
+def strf_date_convert(x, _, type):
+    is_date = any("%" + x in type for x in "aAwdbBmyYjUW")
+    is_time = any("%" + x in type for x in "HIpMSfz")
+
+    dt = datetime.strptime(x, type)
+    if "%y" not in type and "%Y" not in type:  # year not specified
+        dt = dt.replace(year=datetime.today().year)
+
+    if is_date and is_time:
+        return dt
+    elif is_date:
+        return dt.date()
+    elif is_time:
+        return dt.time()
+    else:
+        ValueError("Datetime not a date nor a time?")
+
+
+# ref: https://docs.python.org/3/library/datetime.html#strftime-and-strptime-format-codes
+dt_format_to_regex = {
+    "%a": "(?:Sun|Mon|Tue|Wed|Thu|Fri|Sat)",
+    "%A": "(?:Sunday|Monday|Tuesday|Wednesday|Thursday|Friday|Saturday)",
+    "%w": "[0-6]",
+    "%d": "[0-9]{1,2}",
+    "%b": "(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)",
+    "%B": "(?:January|February|March|April|May|June|July|August|September|October|November|December)",
+    "%m": "[0-9]{1,2}",
+    "%y": "[0-9]{2}",
+    "%Y": "[0-9]{4}",
+    "%H": "[0-9]{1,2}",
+    "%I": "[0-9]{1,2}",
+    "%p": "(?:AM|PM)",
+    "%M": "[0-9]{2}",
+    "%S": "[0-9]{2}",
+    "%f": "[0-9]{1,6}",
+    "%z": "[+|-][0-9]{2}(:?[0-9]{2})?(:?[0-9]{2})?",
+    # "%Z": punt
+    "%j": "[0-9]{1,3}",
+    "%U": "[0-9]{1,2}",
+    "%W": "[0-9]{1,2}",
+}
+
+# Compile a regular expression pattern that matches any date/time format symbol.
+dt_format_symbols_re = re.compile("|".join(dt_format_to_regex))
+
+
+def get_regex_for_datetime_format(format_):
+    """
+    Generate a regex pattern for a given datetime format string.
+
+    Parameters:
+        format_ (str): The datetime format string.
+
+    Returns:
+        str: A regex pattern corresponding to the datetime format string.
+    """
+    # Replace all format symbols with their regex patterns.
+    return dt_format_symbols_re.sub(lambda m: dt_format_to_regex[m.group(0)], format_)
+
+
+class TooManyFields(ValueError):
+    pass
+
+
+class RepeatedNameError(ValueError):
+    pass
+
+
+# note: {} are handled separately
+REGEX_SAFETY = re.compile(r"([?\\.[\]()*+^$!|])")
+
+# allowed field types
+ALLOWED_TYPES = set(list("nbox%fFegwWdDsSl") + ["t" + c for c in "ieahgcts"])
+
+
+def extract_format(format, extra_types):
+    """Pull apart the format [[fill]align][sign][0][width][.precision][type]"""
+    fill = align = None
+    if format[0] in "<>=^":
+        align = format[0]
+        format = format[1:]
+    elif len(format) > 1 and format[1] in "<>=^":
+        fill = format[0]
+        align = format[1]
+        format = format[2:]
+
+    if format.startswith(("+", "-", " ")):
+        format = format[1:]
+
+    zero = False
+    if format and format[0] == "0":
+        zero = True
+        format = format[1:]
+
+    width = ""
+    while format:
+        if not format[0].isdigit():
+            break
+        width += format[0]
+        format = format[1:]
+
+    if format.startswith("."):
+        # Precision isn't needed but we need to capture it so that
+        # the ValueError isn't raised.
+        format = format[1:]  # drop the '.'
+        precision = ""
+        while format:
+            if not format[0].isdigit():
+                break
+            precision += format[0]
+            format = format[1:]
+
+    # the rest is the type, if present
+    type = format
+    if (
+        type
+        and type not in ALLOWED_TYPES
+        and type not in extra_types
+        and not any(k in type for k in dt_format_to_regex)
+    ):
+        raise ValueError("format spec %r not recognised" % type)
+
+    return locals()
+
+
+PARSE_RE = re.compile(r"({{|}}|{[\w-]*(?:\.[\w-]+|\[[^]]+])*(?::[^}]+)?})")
+
+
+class Parser(object):
+    """Encapsulate a format string that may be used to parse other strings."""
+
+    def __init__(self, format, extra_types=None, case_sensitive=False):
+        # a mapping of a name as in {hello.world} to a regex-group compatible
+        # name, like hello__world. It's used to prevent the transformation of
+        # name-to-group and group to name to fail subtly, such as in:
+        # hello_.world-> hello___world->hello._world
+        self._group_to_name_map = {}
+        # also store the original field name to group name mapping to allow
+        # multiple instances of a name in the format string
+        self._name_to_group_map = {}
+        # and to sanity check the repeated instances store away the first
+        # field type specification for the named field
+        self._name_types = {}
+
+        self._format = format
+        if extra_types is None:
+            extra_types = {}
+        self._extra_types = extra_types
+        if case_sensitive:
+            self._re_flags = re.DOTALL
+        else:
+            self._re_flags = re.IGNORECASE | re.DOTALL
+        self._fixed_fields = []
+        self._named_fields = []
+        self._group_index = 0
+        self._type_conversions = {}
+        self._expression = self._generate_expression()
+        self.__search_re = None
+        self.__match_re = None
+
+        log.debug("format %r -> %r", format, self._expression)
+
+    def __repr__(self):
+        if len(self._format) > 20:
+            return "<%s %r>" % (self.__class__.__name__, self._format[:17] + "...")
+        return "<%s %r>" % (self.__class__.__name__, self._format)
+
+    @property
+    def _search_re(self):
+        if self.__search_re is None:
+            try:
+                self.__search_re = re.compile(self._expression, self._re_flags)
+            except AssertionError:
+                # access error through sys to keep py3k and backward compat
+                e = str(sys.exc_info()[1])
+                if e.endswith("this version only supports 100 named groups"):
+                    raise TooManyFields(
+                        "sorry, you are attempting to parse too many complex fields"
+                    )
+        return self.__search_re
+
+    @property
+    def _match_re(self):
+        if self.__match_re is None:
+            expression = r"\A%s\Z" % self._expression
+            try:
+                self.__match_re = re.compile(expression, self._re_flags)
+            except AssertionError:
+                # access error through sys to keep py3k and backward compat
+                e = str(sys.exc_info()[1])
+                if e.endswith("this version only supports 100 named groups"):
+                    raise TooManyFields(
+                        "sorry, you are attempting to parse too many complex fields"
+                    )
+            except re.error:
+                raise NotImplementedError(
+                    "Group names (e.g. (?P<name>) can "
+                    "cause failure, as they are not escaped properly: '%s'" % expression
+                )
+        return self.__match_re
+
+    @property
+    def named_fields(self):
+        return self._named_fields[:]
+
+    @property
+    def fixed_fields(self):
+        return self._fixed_fields[:]
+
+    @property
+    def format(self):
+        return self._format
+
+    def parse(self, string, evaluate_result=True):
+        """Match my format to the string exactly.
+
+        Return a Result or Match instance or None if there's no match.
+        """
+        m = self._match_re.match(string)
+        if m is None:
+            return None
+
+        if evaluate_result:
+            return self.evaluate_result(m)
+        else:
+            return Match(self, m)
+
+    def search(self, string, pos=0, endpos=None, evaluate_result=True):
+        """Search the string for my format.
+
+        Optionally start the search at "pos" character index and limit the
+        search to a maximum index of endpos - equivalent to
+        search(string[:endpos]).
+
+        If the ``evaluate_result`` argument is set to ``False`` a
+        Match instance is returned instead of the actual Result instance.
+
+        Return either a Result instance or None if there's no match.
+        """
+        if endpos is None:
+            endpos = len(string)
+        m = self._search_re.search(string, pos, endpos)
+        if m is None:
+            return None
+
+        if evaluate_result:
+            return self.evaluate_result(m)
+        else:
+            return Match(self, m)
+
+    def findall(
+        self, string, pos=0, endpos=None, extra_types=None, evaluate_result=True
+    ):
+        """Search "string" for all occurrences of "format".
+
+        Optionally start the search at "pos" character index and limit the
+        search to a maximum index of endpos - equivalent to
+        search(string[:endpos]).
+
+        Returns an iterator that holds Result or Match instances for each format match
+        found.
+        """
+        if endpos is None:
+            endpos = len(string)
+        return ResultIterator(
+            self, string, pos, endpos, evaluate_result=evaluate_result
+        )
+
+    def _expand_named_fields(self, named_fields):
+        result = {}
+        for field, value in named_fields.items():
+            # split 'aaa[bbb][ccc]...' into 'aaa' and '[bbb][ccc]...'
+            n = field.find("[")
+            if n == -1:
+                basename, subkeys = field, ""
+            else:
+                basename, subkeys = field[:n], field[n:]
+
+            # create nested dictionaries {'aaa': {'bbb': {'ccc': ...}}}
+            d = result
+            k = basename
+
+            if subkeys:
+                for subkey in re.findall(r"\[[^]]+]", subkeys):
+                    d = d.setdefault(k, {})
+                    k = subkey[1:-1]
+
+            # assign the value to the last key
+            d[k] = value
+
+        return result
+
+    def evaluate_result(self, m):
+        """Generate a Result instance for the given regex match object"""
+        # ok, figure the fixed fields we've pulled out and type convert them
+        fixed_fields = list(m.groups())
+        for n in self._fixed_fields:
+            if n in self._type_conversions:
+                fixed_fields[n] = self._type_conversions[n](fixed_fields[n], m)
+        fixed_fields = tuple(fixed_fields[n] for n in self._fixed_fields)
+
+        # grab the named fields, converting where requested
+        groupdict = m.groupdict()
+        named_fields = {}
+        name_map = {}
+        for k in self._named_fields:
+            korig = self._group_to_name_map[k]
+            name_map[korig] = k
+            if k in self._type_conversions:
+                value = self._type_conversions[k](groupdict[k], m)
+            else:
+                value = groupdict[k]
+
+            named_fields[korig] = value
+
+        # now figure the match spans
+        spans = {n: m.span(name_map[n]) for n in named_fields}
+        spans.update((i, m.span(n + 1)) for i, n in enumerate(self._fixed_fields))
+
+        # and that's our result
+        return Result(fixed_fields, self._expand_named_fields(named_fields), spans)
+
+    def _regex_replace(self, match):
+        return "\\" + match.group(1)
+
+    def _generate_expression(self):
+        # turn my _format attribute into the _expression attribute
+        e = []
+        for part in PARSE_RE.split(self._format):
+            if not part:
+                continue
+            elif part == "{{":
+                e.append(r"\{")
+            elif part == "}}":
+                e.append(r"\}")
+            elif part[0] == "{" and part[-1] == "}":
+                # this will be a braces-delimited field to handle
+                e.append(self._handle_field(part))
+            else:
+                # just some text to match
+                e.append(REGEX_SAFETY.sub(self._regex_replace, part))
+        return "".join(e)
+
+    def _to_group_name(self, field):
+        # return a version of field which can be used as capture group, even
+        # though it might contain '.'
+        group = field.replace(".", "_").replace("[", "_").replace("]", "_").replace("-", "_")
+
+        # make sure we don't collide ("a.b" colliding with "a_b")
+        n = 1
+        while group in self._group_to_name_map:
+            n += 1
+            if "." in field:
+                group = field.replace(".", "_" * n)
+            elif "_" in field:
+                group = field.replace("_", "_" * n)
+            elif "-" in field:
+                group = field.replace("-", "_" * n)
+            else:
+                raise KeyError("duplicated group name %r" % (field,))
+
+        # save off the mapping
+        self._group_to_name_map[group] = field
+        self._name_to_group_map[field] = group
+        return group
+
+    def _handle_field(self, field):
+        # first: lose the braces
+        field = field[1:-1]
+
+        # now figure whether this is an anonymous or named field, and whether
+        # there's any format specification
+        format = ""
+
+        if ":" in field:
+            name, format = field.split(":", 1)
+        else:
+            name = field
+
+        # This *should* be more flexible, but parsing complicated structures
+        # out of the string is hard (and not necessarily useful) ... and I'm
+        # being lazy. So for now `identifier` is "anything starting with a
+        # letter" and digit args don't get attribute or element stuff.
+        if name and name[0].isalpha():
+            if name in self._name_to_group_map:
+                if self._name_types[name] != format:
+                    raise RepeatedNameError(
+                        'field type %r for field "%s" '
+                        "does not match previous seen type %r"
+                        % (format, name, self._name_types[name])
+                    )
+                group = self._name_to_group_map[name]
+                # match previously-seen value
+                return r"(?P=%s)" % group
+            else:
+                group = self._to_group_name(name)
+                self._name_types[name] = format
+            self._named_fields.append(group)
+            # this will become a group, which must not contain dots
+            wrap = r"(?P<%s>%%s)" % group
+        else:
+            self._fixed_fields.append(self._group_index)
+            wrap = r"(%s)"
+            group = self._group_index
+
+        # simplest case: no type specifier ({} or {name})
+        if not format:
+            self._group_index += 1
+            return wrap % r".+?"
+
+        # decode the format specification
+        format = extract_format(format, self._extra_types)
+
+        # figure type conversions, if any
+        type = format["type"]
+        is_numeric = type and type in "n%fegdobx"
+        conv = self._type_conversions
+        if type in self._extra_types:
+            type_converter = self._extra_types[type]
+            s = getattr(type_converter, "pattern", r".+?")
+            regex_group_count = getattr(type_converter, "regex_group_count", 0)
+            if regex_group_count is None:
+                regex_group_count = 0
+            self._group_index += regex_group_count
+            conv[group] = convert_first(type_converter)
+        elif type == "n":
+            s = r"\d{1,3}([,.]\d{3})*"
+            self._group_index += 1
+            conv[group] = int_convert(10)
+        elif type == "b":
+            s = r"(0[bB])?[01]+"
+            conv[group] = int_convert(2)
+            self._group_index += 1
+        elif type == "o":
+            s = r"(0[oO])?[0-7]+"
+            conv[group] = int_convert(8)
+            self._group_index += 1
+        elif type == "x":
+            s = r"(0[xX])?[0-9a-fA-F]+"
+            conv[group] = int_convert(16)
+            self._group_index += 1
+        elif type == "%":
+            s = r"\d+(\.\d+)?%"
+            self._group_index += 1
+            conv[group] = percentage
+        elif type == "f":
+            s = r"\d*\.\d+"
+            conv[group] = convert_first(float)
+        elif type == "F":
+            s = r"\d*\.\d+"
+            conv[group] = convert_first(Decimal)
+        elif type == "e":
+            s = r"\d*\.\d+[eE][-+]?\d+|nan|NAN|[-+]?inf|[-+]?INF"
+            conv[group] = convert_first(float)
+        elif type == "g":
+            s = r"\d+(\.\d+)?([eE][-+]?\d+)?|nan|NAN|[-+]?inf|[-+]?INF"
+            self._group_index += 2
+            conv[group] = convert_first(float)
+        elif type == "d":
+            if format.get("width"):
+                width = r"{1,%s}" % int(format["width"])
+            else:
+                width = "+"
+            s = r"\d{w}|[-+ ]?0[xX][0-9a-fA-F]{w}|[-+ ]?0[bB][01]{w}|[-+ ]?0[oO][0-7]{w}".format(
+                w=width
+            )
+            conv[group] = int_convert()
+            # do not specify number base, determine it automatically
+        elif any(k in type for k in dt_format_to_regex):
+            s = get_regex_for_datetime_format(type)
+            conv[group] = partial(strf_date_convert, type=type)
+        elif type == "ti":
+            s = r"(\d{4}-\d\d-\d\d)((\s+|T)%s)?(Z|\s*[-+]\d\d:?\d\d)?" % TIME_PAT
+            n = self._group_index
+            conv[group] = partial(date_convert, ymd=n + 1, hms=n + 4, tz=n + 7)
+            self._group_index += 7
+        elif type == "tg":
+            s = r"(\d{1,2}[-/](\d{1,2}|%s)[-/]\d{4})(\s+%s)?%s?%s?"
+            s %= (ALL_MONTHS_PAT, TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(
+                date_convert, dmy=n + 1, hms=n + 5, am=n + 8, tz=n + 9
+            )
+            self._group_index += 9
+        elif type == "ta":
+            s = r"((\d{1,2}|%s)[-/]\d{1,2}[-/]\d{4})(\s+%s)?%s?%s?"
+            s %= (ALL_MONTHS_PAT, TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(
+                date_convert, mdy=n + 1, hms=n + 5, am=n + 8, tz=n + 9
+            )
+            self._group_index += 9
+        elif type == "te":
+            # this will allow microseconds through if they're present, but meh
+            s = r"(%s,\s+)?(\d{1,2}\s+%s\s+\d{4})\s+%s%s"
+            s %= (DAYS_PAT, MONTHS_PAT, TIME_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, dmy=n + 3, hms=n + 5, tz=n + 8)
+            self._group_index += 8
+        elif type == "th":
+            # slight flexibility here from the stock Apache format
+            s = r"(\d{1,2}[-/]%s[-/]\d{4}):%s%s" % (MONTHS_PAT, TIME_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, dmy=n + 1, hms=n + 3, tz=n + 6)
+            self._group_index += 6
+        elif type == "tc":
+            s = r"(%s)\s+%s\s+(\d{1,2})\s+%s\s+(\d{4})"
+            s %= (DAYS_PAT, MONTHS_PAT, TIME_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, d_m_y=(n + 4, n + 3, n + 8), hms=n + 5)
+            self._group_index += 8
+        elif type == "tt":
+            s = r"%s?%s?%s?" % (TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, hms=n + 1, am=n + 4, tz=n + 5)
+            self._group_index += 5
+        elif type == "ts":
+            s = r"%s(\s+)(\d+)(\s+)(\d{1,2}:\d{1,2}:\d{1,2})?" % MONTHS_PAT
+            n = self._group_index
+            conv[group] = partial(date_convert, mm=n + 1, dd=n + 3, hms=n + 5)
+            self._group_index += 5
+        elif type == "l":
+            s = r"[A-Za-z]+"
+        elif type:
+            s = r"\%s+" % type
+        elif format.get("precision"):
+            if format.get("width"):
+                s = r".{%s,%s}?" % (format["width"], format["precision"])
+            else:
+                s = r".{1,%s}?" % format["precision"]
+        elif format.get("width"):
+            s = r".{%s,}?" % format["width"]
+        else:
+            s = r".+?"
+
+        align = format["align"]
+        fill = format["fill"]
+
+        # handle some numeric-specific things like fill and sign
+        if is_numeric:
+            # prefix with something (align "=" trumps zero)
+            if align == "=":
+                # special case - align "=" acts like the zero above but with
+                # configurable fill defaulting to "0"
+                if not fill:
+                    fill = "0"
+                s = r"%s*" % fill + s
+
+            # allow numbers to be prefixed with a sign
+            s = r"[-+ ]?" + s
+
+        if not fill:
+            fill = " "
+
+        # Place into a group now - this captures the value we want to keep.
+        # Everything else from now is just padding to be stripped off
+        if wrap:
+            s = wrap % s
+            self._group_index += 1
+
+        if format["width"]:
+            # all we really care about is that if the format originally
+            # specified a width then there will probably be padding - without
+            # an explicit alignment that'll mean right alignment with spaces
+            # padding
+            if not align:
+                align = ">"
+
+        if fill in r".\+?*[](){}^$":
+            fill = "\\" + fill
+
+        # align "=" has been handled
+        if align == "<":
+            s = "%s%s*" % (s, fill)
+        elif align == ">":
+            s = "%s*%s" % (fill, s)
+        elif align == "^":
+            s = "%s*%s%s*" % (fill, s, fill)
+
+        return s
+
+
+class Result(object):
+    """The result of a parse() or search().
+
+    Fixed results may be looked up using `result[index]`.
+    Slices of fixed results may also be looked up.
+
+    Named results may be looked up using `result['name']`.
+
+    Named results may be tested for existence using `'name' in result`.
+    """
+
+    def __init__(self, fixed, named, spans):
+        self.fixed = fixed
+        self.named = named
+        self.spans = spans
+
+    def __getitem__(self, item):
+        if isinstance(item, (int, slice)):
+            return self.fixed[item]
+        return self.named[item]
+
+    def __repr__(self):
+        return "<%s %r %r>" % (self.__class__.__name__, self.fixed, self.named)
+
+    def __contains__(self, name):
+        return name in self.named
+
+
+class Match(object):
+    """The result of a parse() or search() if no results are generated.
+
+    This class is only used to expose internal used regex match objects
+    to the user and use them for external Parser.evaluate_result calls.
+    """
+
+    def __init__(self, parser, match):
+        self.parser = parser
+        self.match = match
+
+    def evaluate_result(self):
+        """Generate results for this Match"""
+        return self.parser.evaluate_result(self.match)
+
+
+class ResultIterator(object):
+    """The result of a findall() operation.
+
+    Each element is a Result instance.
+    """
+
+    def __init__(self, parser, string, pos, endpos, evaluate_result=True):
+        self.parser = parser
+        self.string = string
+        self.pos = pos
+        self.endpos = endpos
+        self.evaluate_result = evaluate_result
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        m = self.parser._search_re.search(self.string, self.pos, self.endpos)
+        if m is None:
+            raise StopIteration()
+        self.pos = m.end()
+
+        if self.evaluate_result:
+            return self.parser.evaluate_result(m)
+        else:
+            return Match(self.parser, m)
+
+    # pre-py3k compat
+    next = __next__
+
+
+def parse(format, string, extra_types=None, evaluate_result=True, case_sensitive=False):
+    """Using "format" attempt to pull values from "string".
+
+    The format must match the string contents exactly. If the value
+    you're looking for is instead just a part of the string use
+    search().
+
+    If ``evaluate_result`` is True the return value will be an Result instance with two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False the return value will be a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+
+    In the case there is no match parse() will return None.
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.parse(string, evaluate_result=evaluate_result)
+
+
+def search(
+    format,
+    string,
+    pos=0,
+    endpos=None,
+    extra_types=None,
+    evaluate_result=True,
+    case_sensitive=False,
+):
+    """Search "string" for the first occurrence of "format".
+
+    The format may occur anywhere within the string. If
+    instead you wish for the format to exactly match the string
+    use parse().
+
+    Optionally start the search at "pos" character index and limit the search
+    to a maximum index of endpos - equivalent to search(string[:endpos]).
+
+    If ``evaluate_result`` is True the return value will be an Result instance with two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False the return value will be a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+
+    In the case there is no match parse() will return None.
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.search(string, pos, endpos, evaluate_result=evaluate_result)
+
+
+def findall(
+    format,
+    string,
+    pos=0,
+    endpos=None,
+    extra_types=None,
+    evaluate_result=True,
+    case_sensitive=False,
+):
+    """Search "string" for all occurrences of "format".
+
+    You will be returned an iterator that holds Result instances
+    for each format match found.
+
+    Optionally start the search at "pos" character index and limit the search
+    to a maximum index of endpos - equivalent to search(string[:endpos]).
+
+    If ``evaluate_result`` is True each returned Result instance has two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False each returned value is a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.findall(string, pos, endpos, evaluate_result=evaluate_result)
+
+
+def compile(format, extra_types=None, case_sensitive=False):
+    """Create a Parser instance to parse "format".
+
+    The resultant Parser has a method .parse(string) which
+    behaves in the same manner as parse(format, string).
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    Use this function if you intend to parse many strings
+    with the same format.
+
+    See the module documentation for the use of "extra_types".
+
+    Returns a Parser instance.
+    """
+    return Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+
+
+# Copyright (c) 2012-2020 Richard Jones <richard@python.org>
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+#  The above copyright notice and this permission notice shall be included in
+#  all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# vim: set filetype=python ts=4 sw=4 et si tw=75
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/LICENSE
new file mode 100644
index 00000000..ce0d0494
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2013-2023 jenisys
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/METADATA
new file mode 100644
index 00000000..83178608
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/METADATA
@@ -0,0 +1,345 @@
+Metadata-Version: 2.1
+Name: parse_type
+Version: 0.6.4
+Summary: Simplifies to build parse types based on the parse module
+Home-page: https://github.com/jenisys/parse_type
+Download-URL: http://pypi.python.org/pypi/parse_type
+Author: Jens Engel
+Author-email: Jens Engel <jenisys@noreply.github.com>
+License: MIT
+Project-URL: Homepage, https://github.com/jenisys/parse_type
+Project-URL: Download, https://pypi.org/project/parse_type/
+Project-URL: Source Code, https://github.com/jenisys/parse_type
+Project-URL: Issue Tracker, https://github.com/jenisys/parse_type/issues/
+Keywords: parse,parsing
+Platform: any
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Console
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 2.7
+Classifier: Programming Language :: Python :: 3.2
+Classifier: Programming Language :: Python :: 3.3
+Classifier: Programming Language :: Python :: 3.4
+Classifier: Programming Language :: Python :: 3.5
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Software Development :: Code Generators
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=2.7, !=3.0.*, !=3.1.*
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: six>=1.15
+Requires-Dist: enum34; python_version < "3.4"
+Requires-Dist: parse>=1.13.1; python_version <= "2.7"
+Requires-Dist: parse>=1.18.0; python_version >= "3.0"
+Provides-Extra: develop
+Requires-Dist: setuptools; extra == "develop"
+Requires-Dist: setuptools-scm; extra == "develop"
+Requires-Dist: wheel; extra == "develop"
+Requires-Dist: build>=0.5.1; extra == "develop"
+Requires-Dist: twine>=1.13.0; extra == "develop"
+Requires-Dist: coverage>=4.4; extra == "develop"
+Requires-Dist: pytest-html>=1.19.0; extra == "develop"
+Requires-Dist: pytest-cov; extra == "develop"
+Requires-Dist: tox<4.0,>=2.8; extra == "develop"
+Requires-Dist: pylint; extra == "develop"
+Requires-Dist: pytest<5.0; python_version < "3.0" and extra == "develop"
+Requires-Dist: virtualenv<20.22.0; python_version <= "3.6" and extra == "develop"
+Requires-Dist: virtualenv>=20.0.0; python_version > "3.6" and extra == "develop"
+Requires-Dist: pytest>=5.0; python_version >= "3.0" and extra == "develop"
+Requires-Dist: ruff; python_version >= "3.7" and extra == "develop"
+Provides-Extra: docs
+Requires-Dist: Sphinx>=1.6; extra == "docs"
+Requires-Dist: sphinx-bootstrap-theme>=0.6.0; extra == "docs"
+Provides-Extra: testing
+Requires-Dist: pytest-html>=1.19.0; extra == "testing"
+Requires-Dist: pytest<5.0; python_version < "3.0" and extra == "testing"
+Requires-Dist: pytest>=5.0; python_version >= "3.0" and extra == "testing"
+
+===============================================================================
+parse_type
+===============================================================================
+
+.. |badge.CI_status| image:: https://github.com/jenisys/parse_type/actions/workflows/test.yml/badge.svg
+    :target: https://github.com/jenisys/parse_type/actions/workflows/test.yml
+    :alt: CI Build Status
+
+.. |badge.latest_version| image:: https://img.shields.io/pypi/v/parse_type.svg
+    :target: https://pypi.python.org/pypi/parse_type
+    :alt: Latest Version
+
+.. |badge.downloads| image:: https://img.shields.io/pypi/dm/parse_type.svg
+    :target: https://pypi.python.org/pypi/parse_type
+    :alt: Downloads
+
+.. |badge.license| image:: https://img.shields.io/pypi/l/parse_type.svg
+    :target: https://pypi.python.org/pypi/parse_type/
+    :alt: License
+
+|badge.CI_status| |badge.latest_version| |badge.license| |badge.downloads|
+
+`parse_type`_ extends the `parse`_ module (opposite of `string.format()`_)
+with the following features:
+
+* build type converters for common use cases (enum/mapping, choice)
+* build a type converter with a cardinality constraint (0..1, 0..*, 1..*)
+    from the type converter with cardinality=1.
+* compose a type converter from other type converters
+* an extended parser that supports the CardinalityField naming schema
+    and creates missing type variants (0..1, 0..*, 1..*) from the
+    primary type converter
+
+.. _parse_type: http://pypi.python.org/pypi/parse_type
+.. _parse:      http://pypi.python.org/pypi/parse
+.. _`string.format()`: http://docs.python.org/library/string.html#format-string-syntax
+
+
+Definitions
+-------------------------------------------------------------------------------
+
+*type converter*
+    A type converter function that converts a textual representation
+    of a value type into instance of this value type.
+    In addition, a type converter function is often annotated with attributes
+    that allows the `parse`_ module to use it in a generic way.
+    A type converter is also called a *parse_type* (a definition used here).
+
+*cardinality field*
+    A naming convention for related types that differ in cardinality.
+    A cardinality field is a type name suffix in the format of a field.
+    It allows parse format expression, ala::
+
+        "{person:Person}"     #< Cardinality: 1    (one; the normal case)
+        "{person:Person?}"    #< Cardinality: 0..1 (zero or one  = optional)
+        "{persons:Person*}"   #< Cardinality: 0..* (zero or more = many0)
+        "{persons:Person+}"   #< Cardinality: 1..* (one  or more = many)
+
+    This naming convention mimics the relationship descriptions in UML diagrams.
+
+
+Basic Example
+-------------------------------------------------------------------------------
+
+Define an own type converter for numbers (integers):
+
+.. code-block:: python
+
+    # -- USE CASE:
+    def parse_number(text):
+        return int(text)
+    parse_number.pattern = r"\d+"  # -- REGULAR EXPRESSION pattern for type.
+
+This is equivalent to:
+
+.. code-block:: python
+
+    import parse
+
+    @parse.with_pattern(r"\d+")
+    def parse_number(text):
+         return int(text)
+    assert hasattr(parse_number, "pattern")
+    assert parse_number.pattern == r"\d+"
+
+
+.. code-block:: python
+
+    # -- USE CASE: Use the type converter with the parse module.
+    schema = "Hello {number:Number}"
+    parser = parse.Parser(schema, dict(Number=parse_number))
+    result = parser.parse("Hello 42")
+    assert result is not None, "REQUIRE: text matches the schema."
+    assert result["number"] == 42
+
+    result = parser.parse("Hello XXX")
+    assert result is None, "MISMATCH: text does not match the schema."
+
+.. hint::
+
+    The described functionality above is standard functionality
+    of the `parse`_ module. It serves as introduction for the remaining cases.
+
+
+Cardinality
+-------------------------------------------------------------------------------
+
+Create an type converter for "ManyNumbers" (List, separated with commas)
+with cardinality "1..* = 1+" (many) from the type converter for a "Number".
+
+.. code-block:: python
+
+    # -- USE CASE: Create new type converter with a cardinality constraint.
+    # CARDINALITY: many := one or more (1..*)
+    from parse import Parser
+    from parse_type import TypeBuilder
+    parse_numbers = TypeBuilder.with_many(parse_number, listsep=",")
+
+    schema = "List: {numbers:ManyNumbers}"
+    parser = Parser(schema, dict(ManyNumbers=parse_numbers))
+    result = parser.parse("List: 1, 2, 3")
+    assert result["numbers"] == [1, 2, 3]
+
+
+Create an type converter for an "OptionalNumbers" with cardinality "0..1 = ?"
+(optional) from the type converter for a "Number".
+
+.. code-block:: python
+
+    # -- USE CASE: Create new type converter with cardinality constraint.
+    # CARDINALITY: optional := zero or one (0..1)
+    from parse import Parser
+    from parse_type import TypeBuilder
+
+    parse_optional_number = TypeBuilder.with_optional(parse_number)
+    schema = "Optional: {number:OptionalNumber}"
+    parser = Parser(schema, dict(OptionalNumber=parse_optional_number))
+    result = parser.parse("Optional: 42")
+    assert result["number"] == 42
+    result = parser.parse("Optional: ")
+    assert result["number"] == None
+
+
+Enumeration (Name-to-Value Mapping)
+-------------------------------------------------------------------------------
+
+Create an type converter for an "Enumeration" from the description of
+the mapping as dictionary.
+
+.. code-block:: python
+
+    # -- USE CASE: Create a type converter for an enumeration.
+    from parse import Parser
+    from parse_type import TypeBuilder
+
+    parse_enum_yesno = TypeBuilder.make_enum({"yes": True, "no": False})
+    parser = Parser("Answer: {answer:YesNo}", dict(YesNo=parse_enum_yesno))
+    result = parser.parse("Answer: yes")
+    assert result["answer"] == True
+
+
+Create an type converter for an "Enumeration" from the description of
+the mapping as an enumeration class (`Python 3.4 enum`_ or the `enum34`_
+backport; see also: `PEP-0435`_).
+
+.. code-block:: python
+
+    # -- USE CASE: Create a type converter for enum34 enumeration class.
+    # NOTE: Use Python 3.4 or enum34 backport.
+    from parse import Parser
+    from parse_type import TypeBuilder
+    from enum import Enum
+
+    class Color(Enum):
+        red   = 1
+        green = 2
+        blue  = 3
+
+    parse_enum_color = TypeBuilder.make_enum(Color)
+    parser = Parser("Select: {color:Color}", dict(Color=parse_enum_color))
+    result = parser.parse("Select: red")
+    assert result["color"] is Color.red
+
+.. _`Python 3.4 enum`: http://docs.python.org/3.4/library/enum.html#module-enum
+.. _enum34:   http://pypi.python.org/pypi/enum34
+.. _PEP-0435: http://www.python.org/dev/peps/pep-0435
+
+
+Choice (Name Enumeration)
+-------------------------------------------------------------------------------
+
+A Choice data type allows to select one of several strings.
+
+Create an type converter for an "Choice" list, a list of unique names
+(as string).
+
+.. code-block:: python
+
+    from parse import Parser
+    from parse_type import TypeBuilder
+
+    parse_choice_yesno = TypeBuilder.make_choice(["yes", "no"])
+    schema = "Answer: {answer:ChoiceYesNo}"
+    parser = Parser(schema, dict(ChoiceYesNo=parse_choice_yesno))
+    result = parser.parse("Answer: yes")
+    assert result["answer"] == "yes"
+
+
+Variant (Type Alternatives)
+-------------------------------------------------------------------------------
+
+Sometimes you need a type converter that can accept text for multiple
+type converter alternatives. This is normally called a "variant" (or: union).
+
+Create an type converter for an "Variant" type that accepts:
+
+* Numbers (positive numbers, as integer)
+* Color enum values (by name)
+
+.. code-block:: python
+
+    from parse import Parser, with_pattern
+    from parse_type import TypeBuilder
+    from enum import Enum
+
+    class Color(Enum):
+        red   = 1
+        green = 2
+        blue  = 3
+
+    @with_pattern(r"\d+")
+    def parse_number(text):
+        return int(text)
+
+    # -- MAKE VARIANT: Alternatives of different type converters.
+    parse_color = TypeBuilder.make_enum(Color)
+    parse_variant = TypeBuilder.make_variant([parse_number, parse_color])
+    schema = "Variant: {variant:Number_or_Color}"
+    parser = Parser(schema, dict(Number_or_Color=parse_variant))
+
+    # -- TEST VARIANT: With number, color and mismatch.
+    result = parser.parse("Variant: 42")
+    assert result["variant"] == 42
+    result = parser.parse("Variant: blue")
+    assert result["variant"] is Color.blue
+    result = parser.parse("Variant: __MISMATCH__")
+    assert not result
+
+
+
+Extended Parser with CardinalityField support
+-------------------------------------------------------------------------------
+
+The parser extends the ``parse.Parser`` and adds the following functionality:
+
+* supports the CardinalityField naming scheme
+* automatically creates missing type variants for types with
+  a CardinalityField by using the primary type converter for cardinality=1
+* extends the provide type converter dictionary with new type variants.
+
+Example:
+
+.. code-block:: python
+
+    # -- USE CASE: Parser with CardinalityField support.
+    # NOTE: Automatically adds missing type variants with CardinalityField part.
+    # USE:  parse_number() type converter from above.
+    from parse_type.cfparse import Parser
+
+    # -- PREPARE: parser, adds missing type variant for cardinality 1..* (many)
+    type_dict = dict(Number=parse_number)
+    schema = "List: {numbers:Number+}"
+    parser = Parser(schema, type_dict)
+    assert "Number+" in type_dict, "Created missing type variant based on: Number"
+
+    # -- USE: parser.
+    result = parser.parse("List: 1, 2, 3")
+    assert result["numbers"] == [1, 2, 3]
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/RECORD
new file mode 100644
index 00000000..ae1855eb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/RECORD
@@ -0,0 +1,23 @@
+parse_type-0.6.4.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+parse_type-0.6.4.dist-info/LICENSE,sha256=dqKclzCv7yWim0qEr84NxLaRzru1iFehLyFqMY74jF4,1069
+parse_type-0.6.4.dist-info/METADATA,sha256=6hLV2emeuYLLoy5duNYoChsr-Pkl306UPEBXfiH2j20,12933
+parse_type-0.6.4.dist-info/RECORD,,
+parse_type-0.6.4.dist-info/WHEEL,sha256=AHX6tWk3qWuce7vKLrj7lnulVHEdWoltgauo8bgCXgU,109
+parse_type-0.6.4.dist-info/top_level.txt,sha256=y83B6aMw8N7s8yay5hdzfOGDHug2gi3NNxOLdaipdoU,11
+parse_type-0.6.4.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+parse_type/__init__.py,sha256=hIV3NhoX1MlAFT6C1_Ohtpz1igYfU44ohN5BLFKOf6o,412
+parse_type/__pycache__/__init__.cpython-312.pyc,,
+parse_type/__pycache__/_version.cpython-312.pyc,,
+parse_type/__pycache__/builder.cpython-312.pyc,,
+parse_type/__pycache__/cardinality.cpython-312.pyc,,
+parse_type/__pycache__/cardinality_field.cpython-312.pyc,,
+parse_type/__pycache__/cfparse.cpython-312.pyc,,
+parse_type/__pycache__/parse.cpython-312.pyc,,
+parse_type/__pycache__/parse_util.cpython-312.pyc,,
+parse_type/_version.py,sha256=YeKWh9qHxBSK-fURNyCejbICk3LuDCjfwlZrUuEgWgM,411
+parse_type/builder.py,sha256=eh0OGBdzuWA9C-bmV8EwWwH4N0ORc6ETyw-EURkr_CE,12159
+parse_type/cardinality.py,sha256=0xnp4GXrvd8Bw_3LTcCyaWwcCB1PN5dzFWZ_rS88_2E,8560
+parse_type/cardinality_field.py,sha256=6F-xsjurUFNPwSH7_Qg8KLcPGtlJst6b_CNEX_Nzpx0,6830
+parse_type/cfparse.py,sha256=Fui8_VeHaHzq593Z8TJ57Jg0ae1XIhKBz-PJ_biAzF8,3773
+parse_type/parse.py,sha256=aJjN_45uFaPKkLrvVVtObCOTnEVjQ47LfkmYfTgc4CI,36112
+parse_type/parse_util.py,sha256=zY3cxtcbY3bmouX9L0NLCiwFbC-nCONC9mHGGNG_Vtg,6176
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/WHEEL
new file mode 100644
index 00000000..f57cfecf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/WHEEL
@@ -0,0 +1,6 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.1.0)
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/top_level.txt
new file mode 100644
index 00000000..5bc13af6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/top_level.txt
@@ -0,0 +1 @@
+parse_type
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/zip-safe b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/zip-safe
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type-0.6.4.dist-info/zip-safe
@@ -0,0 +1 @@
+
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__init__.py b/backend/venv/lib/python3.12/site-packages/parse_type/__init__.py
new file mode 100644
index 00000000..898c5b43
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/__init__.py
@@ -0,0 +1,13 @@
+# -*- coding: UTF-8 -*-
+# Copyright 2013 - 2023, jenisys
+# SPDX-License-Identifier: MIT
+"""
+This module extends the :mod:`parse` to build and derive additional
+parse-types from other, existing types.
+"""
+
+from __future__ import absolute_import
+from parse_type.cardinality import Cardinality
+from parse_type.builder import TypeBuilder, build_type_dict
+
+__all__ = ["Cardinality", "TypeBuilder", "build_type_dict"]
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..1b7dfe82
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..5191aa40
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/builder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/builder.cpython-312.pyc
new file mode 100644
index 00000000..4eca6e2a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/builder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality.cpython-312.pyc
new file mode 100644
index 00000000..c88b5b94
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality_field.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality_field.cpython-312.pyc
new file mode 100644
index 00000000..13dcfabe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cardinality_field.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cfparse.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cfparse.cpython-312.pyc
new file mode 100644
index 00000000..ca325a5a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/cfparse.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse.cpython-312.pyc
new file mode 100644
index 00000000..4d9db523
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse_util.cpython-312.pyc
new file mode 100644
index 00000000..110e0950
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/parse_type/__pycache__/parse_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/_version.py b/backend/venv/lib/python3.12/site-packages/parse_type/_version.py
new file mode 100644
index 00000000..55e6017c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/_version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '0.6.4'
+__version_tuple__ = version_tuple = (0, 6, 4)
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/builder.py b/backend/venv/lib/python3.12/site-packages/parse_type/builder.py
new file mode 100644
index 00000000..ed5cda18
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/builder.py
@@ -0,0 +1,312 @@
+# -*- coding: utf-8 -*-
+# pylint: disable=missing-docstring
+r"""
+Provides support to compose user-defined parse types.
+
+Cardinality
+------------
+
+It is often useful to constrain how often a data type occurs.
+This is also called the cardinality of a data type (in a context).
+The supported cardinality are:
+
+  * 0..1  zero_or_one,  optional<T>: T or None
+  * 0..N  zero_or_more, list_of<T>
+  * 1..N  one_or_more,  list_of<T> (many)
+
+
+.. doctest:: cardinality
+
+    >>> from parse_type import TypeBuilder
+    >>> from parse import Parser
+
+    >>> def parse_number(text):
+    ...     return int(text)
+    >>> parse_number.pattern = r"\d+"
+
+    >>> parse_many_numbers = TypeBuilder.with_many(parse_number)
+    >>> more_types = { "Numbers": parse_many_numbers }
+    >>> parser = Parser("List: {numbers:Numbers}", more_types)
+    >>> parser.parse("List: 1, 2, 3")
+    <Result () {'numbers': [1, 2, 3]}>
+
+
+Enumeration Type (Name-to-Value Mappings)
+-----------------------------------------
+
+An Enumeration data type allows to select one of several enum values by using
+its name. The converter function returns the selected enum value.
+
+.. doctest:: make_enum
+
+    >>> parse_enum_yesno = TypeBuilder.make_enum({"yes": True, "no": False})
+    >>> more_types = { "YesNo": parse_enum_yesno }
+    >>> parser = Parser("Answer: {answer:YesNo}", more_types)
+    >>> parser.parse("Answer: yes")
+    <Result () {'answer': True}>
+
+
+Choice (Name Enumerations)
+-----------------------------
+
+A Choice data type allows to select one of several strings.
+
+.. doctest:: make_choice
+
+    >>> parse_choice_yesno = TypeBuilder.make_choice(["yes", "no"])
+    >>> more_types = { "ChoiceYesNo": parse_choice_yesno }
+    >>> parser = Parser("Answer: {answer:ChoiceYesNo}", more_types)
+    >>> parser.parse("Answer: yes")
+    <Result () {'answer': 'yes'}>
+
+"""
+
+from __future__ import absolute_import
+import inspect
+import re
+import enum
+from parse_type.cardinality import pattern_group_count, \
+    Cardinality, TypeBuilder as CardinalityTypeBuilder
+
+__all__ = ["TypeBuilder", "build_type_dict", "parse_anything"]
+
+
+class TypeBuilder(CardinalityTypeBuilder):
+    """
+    Provides a utility class to build type-converters (parse_types) for
+    the :mod:`parse` module.
+    """
+    default_strict = True
+    default_re_opts = (re.IGNORECASE | re.DOTALL)
+
+    @classmethod
+    def make_list(cls, item_converter=None, listsep=','):
+        """
+        Create a type converter for a list of items (many := 1..*).
+        The parser accepts anything and the converter needs to fail on errors.
+
+        :param item_converter:  Type converter for an item.
+        :param listsep:  List separator to use (as string).
+        :return: Type converter function object for the list.
+        """
+        if not item_converter:
+            item_converter = parse_anything
+        return cls.with_cardinality(Cardinality.many, item_converter,
+                                    pattern=cls.anything_pattern,
+                                    listsep=listsep)
+
+    @staticmethod
+    def make_enum(enum_mappings):
+        """
+        Creates a type converter for an enumeration or text-to-value mapping.
+
+        :param enum_mappings: Defines enumeration names and values.
+        :return: Type converter function object for the enum/mapping.
+        """
+        if (inspect.isclass(enum_mappings) and
+                issubclass(enum_mappings, enum.Enum)):
+            enum_class = enum_mappings
+            enum_mappings = enum_class.__members__
+
+        def convert_enum(text):
+            if text not in convert_enum.mappings:
+                text = text.lower()     # REQUIRED-BY: parse re.IGNORECASE
+            return convert_enum.mappings[text]    #< text.lower() ???
+        convert_enum.pattern = r"|".join(enum_mappings.keys())
+        convert_enum.mappings = enum_mappings
+        return convert_enum
+
+    @staticmethod
+    def _normalize_choices(choices, transform):
+        assert transform is None or callable(transform)
+        if transform:
+            choices = [transform(value)  for value in choices]
+        else:
+            choices = list(choices)
+        return choices
+
+    @classmethod
+    def make_choice(cls, choices, transform=None, strict=None):
+        """
+        Creates a type-converter function to select one from a list of strings.
+        The type-converter function returns the selected choice_text.
+        The :param:`transform()` function is applied in the type converter.
+        It can be used to enforce the case (because parser uses re.IGNORECASE).
+
+        :param choices: List of strings as choice.
+        :param transform: Optional, initial transform function for parsed text.
+        :return: Type converter function object for this choices.
+        """
+        # -- NOTE: Parser uses re.IGNORECASE flag
+        #    => transform may enforce case.
+        choices = cls._normalize_choices(choices, transform)
+        if strict is None:
+            strict = cls.default_strict
+
+        def convert_choice(text):
+            if transform:
+                text = transform(text)
+            if strict and text not in convert_choice.choices:
+                values = ", ".join(convert_choice.choices)
+                raise ValueError("%s not in: %s" % (text, values))
+            return text
+        convert_choice.pattern = r"|".join(choices)
+        convert_choice.choices = choices
+        return convert_choice
+
+    @classmethod
+    def make_choice2(cls, choices, transform=None, strict=None):
+        """
+        Creates a type converter to select one item from a list of strings.
+        The type converter function returns a tuple (index, choice_text).
+
+        :param choices: List of strings as choice.
+        :param transform: Optional, initial transform function for parsed text.
+        :return: Type converter function object for this choices.
+        """
+        choices = cls._normalize_choices(choices, transform)
+        if strict is None:
+            strict = cls.default_strict
+
+        def convert_choice2(text):
+            if transform:
+                text = transform(text)
+            if strict and text not in convert_choice2.choices:
+                values = ", ".join(convert_choice2.choices)
+                raise ValueError("%s not in: %s" % (text, values))
+            index = convert_choice2.choices.index(text)
+            return index, text
+        convert_choice2.pattern = r"|".join(choices)
+        convert_choice2.choices = choices
+        return convert_choice2
+
+    @classmethod
+    def make_variant(cls, converters, re_opts=None, compiled=False, strict=True):
+        """
+        Creates a type converter for a number of type converter alternatives.
+        The first matching type converter is used.
+
+        REQUIRES: type_converter.pattern attribute
+
+        :param converters: List of type converters as alternatives.
+        :param re_opts:  Regular expression options zu use (=default_re_opts).
+        :param compiled: Use compiled regexp matcher, if true (=False).
+        :param strict:   Enable assertion checks.
+        :return: Type converter function object.
+
+        .. note::
+
+            Works only with named fields in :class:`parse.Parser`.
+            Parser needs group_index delta for unnamed/fixed fields.
+            This is not supported for user-defined types.
+            Otherwise, you need to use :class:`parse_type.parse.Parser`
+            (patched version of the :mod:`parse` module).
+        """
+        # -- NOTE: Uses double-dispatch with regex pattern rematch because
+        #          match is not passed through to primary type converter.
+        assert converters, "REQUIRE: Non-empty list."
+        if len(converters) == 1:
+            return converters[0]
+        if re_opts is None:
+            re_opts = cls.default_re_opts
+
+        pattern = r")|(".join([tc.pattern for tc in converters])
+        pattern = r"("+ pattern + ")"
+        group_count = len(converters)
+        for converter in converters:
+            group_count += pattern_group_count(converter.pattern)
+
+        if compiled:
+            convert_variant = cls.__create_convert_variant_compiled(converters,
+                                                                    re_opts,
+                                                                    strict)
+        else:
+            convert_variant = cls.__create_convert_variant(re_opts, strict)
+        convert_variant.pattern = pattern
+        convert_variant.converters = tuple(converters)
+        convert_variant.regex_group_count = group_count
+        return convert_variant
+
+    @staticmethod
+    def __create_convert_variant(re_opts, strict):
+        # -- USE: Regular expression pattern (compiled on use).
+        def convert_variant(text, m=None):
+            # pylint: disable=invalid-name, unused-argument, missing-docstring
+            for converter in convert_variant.converters:
+                if re.match(converter.pattern, text, re_opts):
+                    return converter(text)
+            # -- pragma: no cover
+            assert not strict, "OOPS-VARIANT-MISMATCH: %s" % text
+            return None
+        return convert_variant
+
+    @staticmethod
+    def __create_convert_variant_compiled(converters, re_opts, strict):
+        # -- USE: Compiled regular expression matcher.
+        for converter in converters:
+            matcher = getattr(converter, "matcher", None)
+            if not matcher:
+                converter.matcher = re.compile(converter.pattern, re_opts)
+
+        def convert_variant(text, m=None):
+            # pylint: disable=invalid-name, unused-argument, missing-docstring
+            for converter in convert_variant.converters:
+                if converter.matcher.match(text):
+                    return converter(text)
+            # -- pragma: no cover
+            assert not strict, "OOPS-VARIANT-MISMATCH: %s" % text
+            return None
+        return convert_variant
+
+
+def build_type_dict(converters):
+    """
+    Builds type dictionary for user-defined type converters,
+    used by :mod:`parse` module.
+    This requires that each type converter has a "name" attribute.
+
+    :param converters: List of type converters (parse_types)
+    :return: Type converter dictionary
+    """
+    more_types = {}
+    for converter in converters:
+        assert callable(converter)
+        more_types[converter.name] = converter
+    return more_types
+
+# -----------------------------------------------------------------------------
+# COMMON TYPE CONVERTERS
+# -----------------------------------------------------------------------------
+def parse_anything(text, match=None, match_start=0):
+    """
+    Provides a generic type converter that accepts anything and returns
+    the text (unchanged).
+
+    :param text:  Text to convert (as string).
+    :return: Same text (as string).
+    """
+    # pylint: disable=unused-argument
+    return text
+parse_anything.pattern = TypeBuilder.anything_pattern
+
+
+# -----------------------------------------------------------------------------
+# Copyright (c) 2012-2020 by Jens Engel (https://github/jenisys/parse_type)
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+#  The above copyright notice and this permission notice shall be included in
+#  all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/cardinality.py b/backend/venv/lib/python3.12/site-packages/parse_type/cardinality.py
new file mode 100644
index 00000000..68577671
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/cardinality.py
@@ -0,0 +1,207 @@
+# -*- coding: utf-8 -*-
+"""
+This module simplifies to build parse types and regular expressions
+for a data type with the specified cardinality.
+"""
+
+# -- USE: enum34
+from __future__ import absolute_import
+from enum import Enum
+
+
+# -----------------------------------------------------------------------------
+# FUNCTIONS:
+# -----------------------------------------------------------------------------
+def pattern_group_count(pattern):
+    """Count the pattern-groups within a regex-pattern (as text)."""
+    return pattern.replace(r"\(", "").count("(")
+
+
+# -----------------------------------------------------------------------------
+# CLASS: Cardinality (Enum Class)
+# -----------------------------------------------------------------------------
+class Cardinality(Enum):
+    """Cardinality enumeration class to simplify building regular expression
+    patterns for a data type with the specified cardinality.
+    """
+    # pylint: disable=bad-whitespace
+    __order__ = "one, zero_or_one, zero_or_more, one_or_more"
+    one          = (None, 0)
+    zero_or_one  = (r"(%s)?", 1)                 # SCHEMA: pattern
+    zero_or_more = (r"(%s)?(\s*%s\s*(%s))*", 3)  # SCHEMA: pattern sep pattern
+    one_or_more  = (r"(%s)(\s*%s\s*(%s))*",  3)  # SCHEMA: pattern sep pattern
+
+    # -- ALIASES:
+    optional = zero_or_one
+    many0 = zero_or_more
+    many  = one_or_more
+
+    def __init__(self, schema, group_count=0):
+        self.schema = schema
+        self.group_count = group_count  #< Number of match groups.
+
+    def is_many(self):
+        """Checks for a more general interpretation of "many".
+
+        :return: True, if Cardinality.zero_or_more or Cardinality.one_or_more.
+        """
+        return ((self is Cardinality.zero_or_more) or
+                (self is Cardinality.one_or_more))
+
+    def make_pattern(self, pattern, listsep=','):
+        """Make pattern for a data type with the specified cardinality.
+
+        .. code-block:: python
+
+            yes_no_pattern = r"yes|no"
+            many_yes_no = Cardinality.one_or_more.make_pattern(yes_no_pattern)
+
+        :param pattern:  Regular expression for type (as string).
+        :param listsep:  List separator for multiple items (as string, optional)
+        :return: Regular expression pattern for type with cardinality.
+        """
+        if self is Cardinality.one:
+            return pattern
+        elif self is Cardinality.zero_or_one:
+            return self.schema % pattern
+        # -- OTHERWISE:
+        return self.schema % (pattern, listsep, pattern)
+
+    def compute_group_count(self, pattern):
+        """Compute the number of regexp match groups when the pattern is provided
+        to the :func:`Cardinality.make_pattern()` method.
+
+        :param pattern: Item regexp pattern (as string).
+        :return: Number of regexp match groups in the cardinality pattern.
+        """
+        group_count = self.group_count
+        pattern_repeated = 1
+        if self.is_many():
+            pattern_repeated = 2
+        return group_count + pattern_repeated * pattern_group_count(pattern)
+
+
+# -----------------------------------------------------------------------------
+# CLASS: TypeBuilder
+# -----------------------------------------------------------------------------
+class TypeBuilder(object):
+    """Provides a utility class to build type-converters (parse_types) for parse.
+    It supports to build new type-converters for different cardinality
+    based on the type-converter for cardinality one.
+    """
+    anything_pattern = r".+?"
+    default_pattern = anything_pattern
+
+    @classmethod
+    def with_cardinality(cls, cardinality, converter, pattern=None,
+                         listsep=','):
+        """Creates a type converter for the specified cardinality
+        by using the type converter for T.
+
+        :param cardinality: Cardinality to use (0..1, 0..*, 1..*).
+        :param converter: Type converter (function) for data type T.
+        :param pattern:  Regexp pattern for an item (=converter.pattern).
+        :return: type-converter for optional<T> (T or None).
+        """
+        if cardinality is Cardinality.one:
+            return converter
+        # -- NORMAL-CASE
+        builder_func = getattr(cls, "with_%s" % cardinality.name)
+        if cardinality is Cardinality.zero_or_one:
+            return builder_func(converter, pattern)
+        # -- MANY CASE: 0..*, 1..*
+        return builder_func(converter, pattern, listsep=listsep)
+
+    @classmethod
+    def with_zero_or_one(cls, converter, pattern=None):
+        """Creates a type converter for a T with 0..1 times
+        by using the type converter for one item of T.
+
+        :param converter: Type converter (function) for data type T.
+        :param pattern:  Regexp pattern for an item (=converter.pattern).
+        :return: type-converter for optional<T> (T or None).
+        """
+        cardinality = Cardinality.zero_or_one
+        if not pattern:
+            pattern = getattr(converter, "pattern", cls.default_pattern)
+        optional_pattern = cardinality.make_pattern(pattern)
+        group_count = cardinality.compute_group_count(pattern)
+
+        def convert_optional(text, m=None):
+            # pylint: disable=invalid-name, unused-argument, missing-docstring
+            if text:
+                text = text.strip()
+            if not text:
+                return None
+            return converter(text)
+        convert_optional.pattern = optional_pattern
+        convert_optional.regex_group_count = group_count
+        return convert_optional
+
+    @classmethod
+    def with_zero_or_more(cls, converter, pattern=None, listsep=","):
+        """Creates a type converter function for a list<T> with 0..N items
+        by using the type converter for one item of T.
+
+        :param converter: Type converter (function) for data type T.
+        :param pattern:  Regexp pattern for an item (=converter.pattern).
+        :param listsep:  Optional list separator between items (default: ',')
+        :return: type-converter for list<T>
+        """
+        cardinality = Cardinality.zero_or_more
+        if not pattern:
+            pattern = getattr(converter, "pattern", cls.default_pattern)
+        many0_pattern = cardinality.make_pattern(pattern, listsep)
+        group_count = cardinality.compute_group_count(pattern)
+
+        def convert_list0(text, m=None):
+            # pylint: disable=invalid-name, unused-argument, missing-docstring
+            if text:
+                text = text.strip()
+            if not text:
+                return []
+            return [converter(part.strip()) for part in text.split(listsep)]
+        convert_list0.pattern = many0_pattern
+        # OLD convert_list0.group_count = group_count
+        convert_list0.regex_group_count = group_count
+        return convert_list0
+
+    @classmethod
+    def with_one_or_more(cls, converter, pattern=None, listsep=","):
+        """Creates a type converter function for a list<T> with 1..N items
+        by using the type converter for one item of T.
+
+        :param converter: Type converter (function) for data type T.
+        :param pattern:  Regexp pattern for an item (=converter.pattern).
+        :param listsep:  Optional list separator between items (default: ',')
+        :return: Type converter for list<T>
+        """
+        cardinality = Cardinality.one_or_more
+        if not pattern:
+            pattern = getattr(converter, "pattern", cls.default_pattern)
+        many_pattern = cardinality.make_pattern(pattern, listsep)
+        group_count = cardinality.compute_group_count(pattern)
+
+        def convert_list(text, m=None):
+            # pylint: disable=invalid-name, unused-argument, missing-docstring
+            return [converter(part.strip()) for part in text.split(listsep)]
+        convert_list.pattern = many_pattern
+        # OLD: convert_list.group_count = group_count
+        convert_list.regex_group_count = group_count
+        return convert_list
+
+    # -- ALIAS METHODS:
+    @classmethod
+    def with_optional(cls, converter, pattern=None):
+        """Alias for :py:meth:`with_zero_or_one()` method."""
+        return cls.with_zero_or_one(converter, pattern)
+
+    @classmethod
+    def with_many(cls, converter, pattern=None, listsep=','):
+        """Alias for :py:meth:`with_one_or_more()` method."""
+        return cls.with_one_or_more(converter, pattern, listsep)
+
+    @classmethod
+    def with_many0(cls, converter, pattern=None, listsep=','):
+        """Alias for :py:meth:`with_zero_or_more()` method."""
+        return cls.with_zero_or_more(converter, pattern, listsep)
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/cardinality_field.py b/backend/venv/lib/python3.12/site-packages/parse_type/cardinality_field.py
new file mode 100644
index 00000000..4e892edd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/cardinality_field.py
@@ -0,0 +1,171 @@
+# -*- coding: utf-8 -*-
+"""
+Provides support for cardinality fields.
+A cardinality field is a type suffix for parse format expression, ala:
+
+    "{person:Person?}"   #< Cardinality: 0..1 = zero or one  = optional
+    "{persons:Person*}"  #< Cardinality: 0..* = zero or more = many0
+    "{persons:Person+}"  #< Cardinality: 1..* = one  or more = many
+"""
+
+from __future__ import absolute_import
+import six
+from parse_type.cardinality import Cardinality, TypeBuilder
+
+
+class MissingTypeError(KeyError):   # pylint: disable=missing-docstring
+    pass
+
+# -----------------------------------------------------------------------------
+# CLASS: Cardinality (Field Part)
+# -----------------------------------------------------------------------------
+class CardinalityField(object):
+    """Cardinality field for parse format expression, ala:
+
+        "{person:Person?}"   #< Cardinality: 0..1 = zero or one  = optional
+        "{persons:Person*}"  #< Cardinality: 0..* = zero or more = many0
+        "{persons:Person+}"  #< Cardinality: 1..* = one  or more = many
+    """
+
+    # -- MAPPING SUPPORT:
+    pattern_chars = "?*+"
+    from_char_map = {
+        '?': Cardinality.zero_or_one,
+        '*': Cardinality.zero_or_more,
+        '+': Cardinality.one_or_more,
+    }
+    to_char_map = dict([(value, key)  for key, value in from_char_map.items()])
+
+    @classmethod
+    def matches_type(cls, type_name):
+        """Checks if a type name uses the CardinalityField naming scheme.
+
+        :param type_name:  Type name to check (as string).
+        :return: True, if type name has CardinalityField name suffix.
+        """
+        return type_name and type_name[-1] in CardinalityField.pattern_chars
+
+    @classmethod
+    def split_type(cls, type_name):
+        """Split type of a type name with CardinalityField suffix into its parts.
+
+        :param type_name:  Type name (as string).
+        :return: Tuple (type_basename, cardinality)
+        """
+        if cls.matches_type(type_name):
+            basename = type_name[:-1]
+            cardinality = cls.from_char_map[type_name[-1]]
+        else:
+            # -- ASSUME: Cardinality.one
+            cardinality = Cardinality.one
+            basename = type_name
+        return (basename, cardinality)
+
+    @classmethod
+    def make_type(cls, basename, cardinality):
+        """Build new type name according to CardinalityField naming scheme.
+
+        :param basename:  Type basename of primary type (as string).
+        :param cardinality: Cardinality of the new type (as Cardinality item).
+        :return: Type name with CardinalityField suffix (if needed)
+        """
+        if cardinality is Cardinality.one:
+            # -- POSTCONDITION: assert not cls.make_type(type_name)
+            return basename
+        # -- NORMAL CASE: type with CardinalityField suffix.
+        type_name = "%s%s" % (basename, cls.to_char_map[cardinality])
+        # -- POSTCONDITION: assert cls.make_type(type_name)
+        return type_name
+
+
+# -----------------------------------------------------------------------------
+# CLASS: CardinalityFieldTypeBuilder
+# -----------------------------------------------------------------------------
+class CardinalityFieldTypeBuilder(object):
+    """Utility class to create type converters based on:
+
+      * the CardinalityField naming scheme and
+      * type converter for cardinality=1
+    """
+
+    listsep = ','
+
+    @classmethod
+    def create_type_variant(cls, type_name, type_converter):
+        r"""Create type variants for types with a cardinality field.
+        The new type converters are based on the type converter with
+        cardinality=1.
+
+        .. code-block:: python
+
+            import parse
+
+            @parse.with_pattern(r'\d+')
+            def parse_number(text):
+                return int(text)
+
+            new_type = CardinalityFieldTypeBuilder.create_type_variant(
+                                    "Number+", parse_number)
+            new_type = CardinalityFieldTypeBuilder.create_type_variant(
+                                    "Number+", dict(Number=parse_number))
+
+        :param type_name:  Type name with cardinality field suffix.
+        :param type_converter:  Type converter or type dictionary.
+        :return: Type converter variant (function).
+        :raises: ValueError, if type_name does not end with CardinalityField
+        :raises: MissingTypeError, if type_converter is missing in type_dict
+        """
+        assert isinstance(type_name, six.string_types)
+        if not CardinalityField.matches_type(type_name):
+            message = "type_name='%s' has no CardinalityField" % type_name
+            raise ValueError(message)
+
+        primary_name, cardinality = CardinalityField.split_type(type_name)
+        if isinstance(type_converter, dict):
+            type_dict = type_converter
+            type_converter = type_dict.get(primary_name, None)
+            if not type_converter:
+                raise MissingTypeError(primary_name)
+
+        assert callable(type_converter)
+        type_variant = TypeBuilder.with_cardinality(cardinality,
+                                                    type_converter,
+                                                    listsep=cls.listsep)
+        type_variant.name = type_name
+        return type_variant
+
+
+    @classmethod
+    def create_type_variants(cls, type_names, type_dict):
+        """Create type variants for types with a cardinality field.
+        The new type converters are based on the type converter with
+        cardinality=1.
+
+        .. code-block:: python
+
+            # -- USE: parse_number() type converter function.
+            new_types = CardinalityFieldTypeBuilder.create_type_variants(
+                            ["Number?", "Number+"], dict(Number=parse_number))
+
+        :param type_names: List of type names with cardinality field suffix.
+        :param type_dict:  Type dictionary with named type converters.
+        :return: Type dictionary with type converter variants.
+        """
+        type_variant_dict = {}
+        for type_name in type_names:
+            type_variant = cls.create_type_variant(type_name, type_dict)
+            type_variant_dict[type_name] = type_variant
+        return type_variant_dict
+
+    # MAYBE: Check if really needed.
+    @classmethod
+    def create_missing_type_variants(cls, type_names, type_dict):
+        """Create missing type variants for types with a cardinality field.
+
+        :param type_names: List of type names with cardinality field suffix.
+        :param type_dict:  Type dictionary with named type converters.
+        :return: Type dictionary with missing type converter variants.
+        """
+        missing_type_names = [name for name in type_names
+                              if name not in type_dict]
+        return cls.create_type_variants(missing_type_names, type_dict)
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/cfparse.py b/backend/venv/lib/python3.12/site-packages/parse_type/cfparse.py
new file mode 100644
index 00000000..1032a874
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/cfparse.py
@@ -0,0 +1,87 @@
+# -*- coding: utf-8 -*-
+"""
+Provides an extended :class:`parse.Parser` class that supports the
+cardinality fields in (user-defined) types.
+"""
+
+from __future__ import absolute_import
+import logging
+import parse
+from .cardinality_field import CardinalityField, CardinalityFieldTypeBuilder
+from .parse_util import FieldParser
+
+
+log = logging.getLogger(__name__)   # pylint: disable=invalid-name
+
+
+class Parser(parse.Parser):
+    """Provides an extended :class:`parse.Parser` with cardinality field support.
+    A cardinality field is a type suffix for parse format expression, ala:
+
+        "... {person:Person?} ..."   -- OPTIONAL: Cardinality zero or one, 0..1
+        "... {persons:Person*} ..."  -- MANY0: Cardinality zero or more, 0..
+        "... {persons:Person+} ..."  -- MANY:  Cardinality one  or more, 1..
+
+    When the primary type converter for cardinality=1 is provided,
+    the type variants for the other cardinality cases can be derived from it.
+
+    This parser class automatically creates missing type variants for types
+    with a cardinality field and passes the extended type dictionary
+    to its base class.
+    """
+    # -- TYPE-BUILDER: For missing types in Fields with CardinalityField part.
+    type_builder = CardinalityFieldTypeBuilder
+
+    def __init__(self, schema, extra_types=None, case_sensitive=False,
+                 type_builder=None):
+        """Creates a parser with CardinalityField part support.
+
+        :param schema:  Parse schema (or format) for parser (as string).
+        :param extra_types:  Type dictionary with type converters (or None).
+        :param case_sensitive: Indicates if case-sensitive regexp are used.
+        :param type_builder: Type builder to use for missing types.
+        """
+        if extra_types is None:
+            extra_types = {}
+        missing = self.create_missing_types(schema, extra_types, type_builder)
+        if missing:
+            # pylint: disable=logging-not-lazy
+            log.debug("MISSING TYPES: %s" % ",".join(missing.keys()))
+            extra_types.update(missing)
+
+        # -- FINALLY: Delegate to base class.
+        super(Parser, self).__init__(schema, extra_types,
+                                     case_sensitive=case_sensitive)
+
+    @classmethod
+    def create_missing_types(cls, schema, type_dict, type_builder=None):
+        """Creates missing types for fields with a CardinalityField part.
+        It is assumed that the primary type converter for cardinality=1
+        is registered in the type dictionary.
+
+        :param schema:  Parse schema (or format) for parser (as string).
+        :param type_dict:  Type dictionary with type converters.
+        :param type_builder: Type builder to use for missing types.
+        :return: Type dictionary with missing types. Empty, if none.
+        :raises: MissingTypeError,
+                if a primary type converter with cardinality=1 is missing.
+        """
+        if not type_builder:
+            type_builder = cls.type_builder
+
+        missing = cls.extract_missing_special_type_names(schema, type_dict)
+        return type_builder.create_type_variants(missing, type_dict)
+
+    @staticmethod
+    def extract_missing_special_type_names(schema, type_dict):
+        # pylint: disable=invalid-name
+        """Extract the type names for fields with CardinalityField part.
+        Selects only the missing type names that are not in the type dictionary.
+
+        :param schema:     Parse schema to use (as string).
+        :param type_dict:  Type dictionary with type converters.
+        :return: Generator with missing type names (as string).
+        """
+        for name in FieldParser.extract_types(schema):
+            if CardinalityField.matches_type(name) and (name not in type_dict):
+                yield name
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/parse.py b/backend/venv/lib/python3.12/site-packages/parse_type/parse.py
new file mode 100644
index 00000000..fd1a01d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/parse.py
@@ -0,0 +1,1094 @@
+# -*- coding: UTF-8 -*-
+# BASED-ON: https://github.com/r1chardj0n3s/parse/parse.py
+# VERSION:  parse 1.20.2
+# Same as original parse modules.
+#
+# pylint: disable=line-too-long, invalid-name, too-many-locals, too-many-arguments
+# pylint: disable=redefined-builtin, too-few-public-methods, no-else-return
+# pylint: disable=unused-variable, no-self-use, missing-docstring
+# pylint: disable=unused-argument, unused-variable
+# pylint: disable=too-many-branches, too-many-statements
+# pylint: disable=all
+#
+#  -- ORIGINAL-CODE STARTS-HERE ------------------------------------------------
+from __future__ import absolute_import
+
+import logging
+import re
+import sys
+from datetime import datetime
+from datetime import time
+from datetime import timedelta
+from datetime import tzinfo
+from decimal import Decimal
+from functools import partial
+
+
+__version__ = "1.20.2"
+__all__ = ["parse", "search", "findall", "with_pattern"]
+
+log = logging.getLogger(__name__)
+
+
+def with_pattern(pattern, regex_group_count=None):
+    r"""Attach a regular expression pattern matcher to a custom type converter
+    function.
+
+    This annotates the type converter with the :attr:`pattern` attribute.
+
+    EXAMPLE:
+        >>> import parse
+        >>> @parse.with_pattern(r"\d+")
+        ... def parse_number(text):
+        ...     return int(text)
+
+    is equivalent to:
+
+        >>> def parse_number(text):
+        ...     return int(text)
+        >>> parse_number.pattern = r"\d+"
+
+    :param pattern: regular expression pattern (as text)
+    :param regex_group_count: Indicates how many regex-groups are in pattern.
+    :return: wrapped function
+    """
+
+    def decorator(func):
+        func.pattern = pattern
+        func.regex_group_count = regex_group_count
+        return func
+
+    return decorator
+
+
+class int_convert:
+    """Convert a string to an integer.
+
+    The string may start with a sign.
+
+    It may be of a base other than 2, 8, 10 or 16.
+
+    If base isn't specified, it will be detected automatically based
+    on a string format. When string starts with a base indicator, 0#nnnn,
+    it overrides the default base of 10.
+
+    It may also have other non-numeric characters that we can ignore.
+    """
+
+    CHARS = "0123456789abcdefghijklmnopqrstuvwxyz"
+
+    def __init__(self, base=None):
+        self.base = base
+
+    def __call__(self, string, match):
+        if string[0] == "-":
+            sign = -1
+            number_start = 1
+        elif string[0] == "+":
+            sign = 1
+            number_start = 1
+        else:
+            sign = 1
+            number_start = 0
+
+        base = self.base
+        # If base wasn't specified, detect it automatically
+        if base is None:
+            # Assume decimal number, unless different base is detected
+            base = 10
+
+            # For number formats starting with 0b, 0o, 0x, use corresponding base ...
+            if string[number_start] == "0" and len(string) - number_start > 2:
+                if string[number_start + 1] in "bB":
+                    base = 2
+                elif string[number_start + 1] in "oO":
+                    base = 8
+                elif string[number_start + 1] in "xX":
+                    base = 16
+
+        chars = int_convert.CHARS[:base]
+        string = re.sub("[^%s]" % chars, "", string.lower())
+        return sign * int(string, base)
+
+
+class convert_first:
+    """Convert the first element of a pair.
+    This equivalent to lambda s,m: converter(s). But unlike a lambda function, it can be pickled
+    """
+
+    def __init__(self, converter):
+        self.converter = converter
+
+    def __call__(self, string, match):
+        return self.converter(string)
+
+
+def percentage(string, match):
+    return float(string[:-1]) / 100.0
+
+
+class FixedTzOffset(tzinfo):
+    """Fixed offset in minutes east from UTC."""
+
+    ZERO = timedelta(0)
+
+    def __init__(self, offset, name):
+        self._offset = timedelta(minutes=offset)
+        self._name = name
+
+    def __repr__(self):
+        return "<%s %s %s>" % (self.__class__.__name__, self._name, self._offset)
+
+    def utcoffset(self, dt):
+        return self._offset
+
+    def tzname(self, dt):
+        return self._name
+
+    def dst(self, dt):
+        return self.ZERO
+
+    def __eq__(self, other):
+        if not isinstance(other, FixedTzOffset):
+            return NotImplemented
+        return self._name == other._name and self._offset == other._offset
+
+
+MONTHS_MAP = {
+    "Jan": 1,
+    "January": 1,
+    "Feb": 2,
+    "February": 2,
+    "Mar": 3,
+    "March": 3,
+    "Apr": 4,
+    "April": 4,
+    "May": 5,
+    "Jun": 6,
+    "June": 6,
+    "Jul": 7,
+    "July": 7,
+    "Aug": 8,
+    "August": 8,
+    "Sep": 9,
+    "September": 9,
+    "Oct": 10,
+    "October": 10,
+    "Nov": 11,
+    "November": 11,
+    "Dec": 12,
+    "December": 12,
+}
+DAYS_PAT = r"(Mon|Tue|Wed|Thu|Fri|Sat|Sun)"
+MONTHS_PAT = r"(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)"
+ALL_MONTHS_PAT = r"(%s)" % "|".join(MONTHS_MAP)
+TIME_PAT = r"(\d{1,2}:\d{1,2}(:\d{1,2}(\.\d+)?)?)"
+AM_PAT = r"(\s+[AP]M)"
+TZ_PAT = r"(\s+[-+]\d\d?:?\d\d)"
+
+
+def date_convert(
+    string,
+    match,
+    ymd=None,
+    mdy=None,
+    dmy=None,
+    d_m_y=None,
+    hms=None,
+    am=None,
+    tz=None,
+    mm=None,
+    dd=None,
+):
+    """Convert the incoming string containing some date / time info into a
+    datetime instance.
+    """
+    groups = match.groups()
+    time_only = False
+    if mm and dd:
+        y = datetime.today().year
+        m = groups[mm]
+        d = groups[dd]
+    elif ymd is not None:
+        y, m, d = re.split(r"[-/\s]", groups[ymd])
+    elif mdy is not None:
+        m, d, y = re.split(r"[-/\s]", groups[mdy])
+    elif dmy is not None:
+        d, m, y = re.split(r"[-/\s]", groups[dmy])
+    elif d_m_y is not None:
+        d, m, y = d_m_y
+        d = groups[d]
+        m = groups[m]
+        y = groups[y]
+    else:
+        time_only = True
+
+    H = M = S = u = 0
+    if hms is not None and groups[hms]:
+        t = groups[hms].split(":")
+        if len(t) == 2:
+            H, M = t
+        else:
+            H, M, S = t
+            if "." in S:
+                S, u = S.split(".")
+                u = int(float("." + u) * 1000000)
+            S = int(S)
+        H = int(H)
+        M = int(M)
+
+    if am is not None:
+        am = groups[am]
+        if am:
+            am = am.strip()
+        if am == "AM" and H == 12:
+            # correction for "12" hour functioning as "0" hour: 12:15 AM = 00:15 by 24 hr clock
+            H -= 12
+        elif am == "PM" and H == 12:
+            # no correction needed: 12PM is midday, 12:00 by 24 hour clock
+            pass
+        elif am == "PM":
+            H += 12
+
+    if tz is not None:
+        tz = groups[tz]
+    if tz == "Z":
+        tz = FixedTzOffset(0, "UTC")
+    elif tz:
+        tz = tz.strip()
+        if tz.isupper():
+            # TODO use the awesome python TZ module?
+            pass
+        else:
+            sign = tz[0]
+            if ":" in tz:
+                tzh, tzm = tz[1:].split(":")
+            elif len(tz) == 4:  # 'snnn'
+                tzh, tzm = tz[1], tz[2:4]
+            else:
+                tzh, tzm = tz[1:3], tz[3:5]
+            offset = int(tzm) + int(tzh) * 60
+            if sign == "-":
+                offset = -offset
+            tz = FixedTzOffset(offset, tz)
+
+    if time_only:
+        d = time(H, M, S, u, tzinfo=tz)
+    else:
+        y = int(y)
+        if m.isdigit():
+            m = int(m)
+        else:
+            m = MONTHS_MAP[m]
+        d = int(d)
+        d = datetime(y, m, d, H, M, S, u, tzinfo=tz)
+
+    return d
+
+
+def strf_date_convert(x, _, type):
+    is_date = any("%" + x in type for x in "aAwdbBmyYjUW")
+    is_time = any("%" + x in type for x in "HIpMSfz")
+
+    dt = datetime.strptime(x, type)
+    if "%y" not in type and "%Y" not in type:  # year not specified
+        dt = dt.replace(year=datetime.today().year)
+
+    if is_date and is_time:
+        return dt
+    elif is_date:
+        return dt.date()
+    elif is_time:
+        return dt.time()
+    else:
+        ValueError("Datetime not a date nor a time?")
+
+
+# ref: https://docs.python.org/3/library/datetime.html#strftime-and-strptime-format-codes
+dt_format_to_regex = {
+    "%a": "(?:Sun|Mon|Tue|Wed|Thu|Fri|Sat)",
+    "%A": "(?:Sunday|Monday|Tuesday|Wednesday|Thursday|Friday|Saturday)",
+    "%w": "[0-6]",
+    "%d": "[0-9]{1,2}",
+    "%b": "(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)",
+    "%B": "(?:January|February|March|April|May|June|July|August|September|October|November|December)",
+    "%m": "[0-9]{1,2}",
+    "%y": "[0-9]{2}",
+    "%Y": "[0-9]{4}",
+    "%H": "[0-9]{1,2}",
+    "%I": "[0-9]{1,2}",
+    "%p": "(?:AM|PM)",
+    "%M": "[0-9]{2}",
+    "%S": "[0-9]{2}",
+    "%f": "[0-9]{1,6}",
+    "%z": "[+|-][0-9]{2}(:?[0-9]{2})?(:?[0-9]{2})?",
+    # "%Z": punt
+    "%j": "[0-9]{1,3}",
+    "%U": "[0-9]{1,2}",
+    "%W": "[0-9]{1,2}",
+}
+
+# Compile a regular expression pattern that matches any date/time format symbol.
+dt_format_symbols_re = re.compile("|".join(dt_format_to_regex))
+
+
+def get_regex_for_datetime_format(format_):
+    """
+    Generate a regex pattern for a given datetime format string.
+
+    Parameters:
+        format_ (str): The datetime format string.
+
+    Returns:
+        str: A regex pattern corresponding to the datetime format string.
+    """
+    # Replace all format symbols with their regex patterns.
+    return dt_format_symbols_re.sub(lambda m: dt_format_to_regex[m.group(0)], format_)
+
+
+class TooManyFields(ValueError):
+    pass
+
+
+class RepeatedNameError(ValueError):
+    pass
+
+
+# note: {} are handled separately
+REGEX_SAFETY = re.compile(r"([?\\.[\]()*+^$!|])")
+
+# allowed field types
+ALLOWED_TYPES = set(list("nbox%fFegwWdDsSl") + ["t" + c for c in "ieahgcts"])
+
+
+def extract_format(format, extra_types):
+    """Pull apart the format [[fill]align][sign][0][width][.precision][type]"""
+    fill = align = None
+    if format[0] in "<>=^":
+        align = format[0]
+        format = format[1:]
+    elif len(format) > 1 and format[1] in "<>=^":
+        fill = format[0]
+        align = format[1]
+        format = format[2:]
+
+    if format.startswith(("+", "-", " ")):
+        format = format[1:]
+
+    zero = False
+    if format and format[0] == "0":
+        zero = True
+        format = format[1:]
+
+    width = ""
+    while format:
+        if not format[0].isdigit():
+            break
+        width += format[0]
+        format = format[1:]
+
+    if format.startswith("."):
+        # Precision isn't needed but we need to capture it so that
+        # the ValueError isn't raised.
+        format = format[1:]  # drop the '.'
+        precision = ""
+        while format:
+            if not format[0].isdigit():
+                break
+            precision += format[0]
+            format = format[1:]
+
+    # the rest is the type, if present
+    type = format
+    if (
+        type
+        and type not in ALLOWED_TYPES
+        and type not in extra_types
+        and not any(k in type for k in dt_format_to_regex)
+    ):
+        raise ValueError("format spec %r not recognised" % type)
+
+    return locals()
+
+
+PARSE_RE = re.compile(r"({{|}}|{[\w-]*(?:\.[\w-]+|\[[^]]+])*(?::[^}]+)?})")
+
+
+class Parser(object):
+    """Encapsulate a format string that may be used to parse other strings."""
+
+    def __init__(self, format, extra_types=None, case_sensitive=False):
+        # a mapping of a name as in {hello.world} to a regex-group compatible
+        # name, like hello__world. It's used to prevent the transformation of
+        # name-to-group and group to name to fail subtly, such as in:
+        # hello_.world-> hello___world->hello._world
+        self._group_to_name_map = {}
+        # also store the original field name to group name mapping to allow
+        # multiple instances of a name in the format string
+        self._name_to_group_map = {}
+        # and to sanity check the repeated instances store away the first
+        # field type specification for the named field
+        self._name_types = {}
+
+        self._format = format
+        if extra_types is None:
+            extra_types = {}
+        self._extra_types = extra_types
+        if case_sensitive:
+            self._re_flags = re.DOTALL
+        else:
+            self._re_flags = re.IGNORECASE | re.DOTALL
+        self._fixed_fields = []
+        self._named_fields = []
+        self._group_index = 0
+        self._type_conversions = {}
+        self._expression = self._generate_expression()
+        self.__search_re = None
+        self.__match_re = None
+
+        log.debug("format %r -> %r", format, self._expression)
+
+    def __repr__(self):
+        if len(self._format) > 20:
+            return "<%s %r>" % (self.__class__.__name__, self._format[:17] + "...")
+        return "<%s %r>" % (self.__class__.__name__, self._format)
+
+    @property
+    def _search_re(self):
+        if self.__search_re is None:
+            try:
+                self.__search_re = re.compile(self._expression, self._re_flags)
+            except AssertionError:
+                # access error through sys to keep py3k and backward compat
+                e = str(sys.exc_info()[1])
+                if e.endswith("this version only supports 100 named groups"):
+                    raise TooManyFields(
+                        "sorry, you are attempting to parse too many complex fields"
+                    )
+        return self.__search_re
+
+    @property
+    def _match_re(self):
+        if self.__match_re is None:
+            expression = r"\A%s\Z" % self._expression
+            try:
+                self.__match_re = re.compile(expression, self._re_flags)
+            except AssertionError:
+                # access error through sys to keep py3k and backward compat
+                e = str(sys.exc_info()[1])
+                if e.endswith("this version only supports 100 named groups"):
+                    raise TooManyFields(
+                        "sorry, you are attempting to parse too many complex fields"
+                    )
+            except re.error:
+                raise NotImplementedError(
+                    "Group names (e.g. (?P<name>) can "
+                    "cause failure, as they are not escaped properly: '%s'" % expression
+                )
+        return self.__match_re
+
+    @property
+    def named_fields(self):
+        return self._named_fields[:]
+
+    @property
+    def fixed_fields(self):
+        return self._fixed_fields[:]
+
+    @property
+    def format(self):
+        return self._format
+
+    def parse(self, string, evaluate_result=True):
+        """Match my format to the string exactly.
+
+        Return a Result or Match instance or None if there's no match.
+        """
+        m = self._match_re.match(string)
+        if m is None:
+            return None
+
+        if evaluate_result:
+            return self.evaluate_result(m)
+        else:
+            return Match(self, m)
+
+    def search(self, string, pos=0, endpos=None, evaluate_result=True):
+        """Search the string for my format.
+
+        Optionally start the search at "pos" character index and limit the
+        search to a maximum index of endpos - equivalent to
+        search(string[:endpos]).
+
+        If the ``evaluate_result`` argument is set to ``False`` a
+        Match instance is returned instead of the actual Result instance.
+
+        Return either a Result instance or None if there's no match.
+        """
+        if endpos is None:
+            endpos = len(string)
+        m = self._search_re.search(string, pos, endpos)
+        if m is None:
+            return None
+
+        if evaluate_result:
+            return self.evaluate_result(m)
+        else:
+            return Match(self, m)
+
+    def findall(
+        self, string, pos=0, endpos=None, extra_types=None, evaluate_result=True
+    ):
+        """Search "string" for all occurrences of "format".
+
+        Optionally start the search at "pos" character index and limit the
+        search to a maximum index of endpos - equivalent to
+        search(string[:endpos]).
+
+        Returns an iterator that holds Result or Match instances for each format match
+        found.
+        """
+        if endpos is None:
+            endpos = len(string)
+        return ResultIterator(
+            self, string, pos, endpos, evaluate_result=evaluate_result
+        )
+
+    def _expand_named_fields(self, named_fields):
+        result = {}
+        for field, value in named_fields.items():
+            # split 'aaa[bbb][ccc]...' into 'aaa' and '[bbb][ccc]...'
+            n = field.find("[")
+            if n == -1:
+                basename, subkeys = field, ""
+            else:
+                basename, subkeys = field[:n], field[n:]
+
+            # create nested dictionaries {'aaa': {'bbb': {'ccc': ...}}}
+            d = result
+            k = basename
+
+            if subkeys:
+                for subkey in re.findall(r"\[[^]]+]", subkeys):
+                    d = d.setdefault(k, {})
+                    k = subkey[1:-1]
+
+            # assign the value to the last key
+            d[k] = value
+
+        return result
+
+    def evaluate_result(self, m):
+        """Generate a Result instance for the given regex match object"""
+        # ok, figure the fixed fields we've pulled out and type convert them
+        fixed_fields = list(m.groups())
+        for n in self._fixed_fields:
+            if n in self._type_conversions:
+                fixed_fields[n] = self._type_conversions[n](fixed_fields[n], m)
+        fixed_fields = tuple(fixed_fields[n] for n in self._fixed_fields)
+
+        # grab the named fields, converting where requested
+        groupdict = m.groupdict()
+        named_fields = {}
+        name_map = {}
+        for k in self._named_fields:
+            korig = self._group_to_name_map[k]
+            name_map[korig] = k
+            if k in self._type_conversions:
+                value = self._type_conversions[k](groupdict[k], m)
+            else:
+                value = groupdict[k]
+
+            named_fields[korig] = value
+
+        # now figure the match spans
+        spans = {n: m.span(name_map[n]) for n in named_fields}
+        spans.update((i, m.span(n + 1)) for i, n in enumerate(self._fixed_fields))
+
+        # and that's our result
+        return Result(fixed_fields, self._expand_named_fields(named_fields), spans)
+
+    def _regex_replace(self, match):
+        return "\\" + match.group(1)
+
+    def _generate_expression(self):
+        # turn my _format attribute into the _expression attribute
+        e = []
+        for part in PARSE_RE.split(self._format):
+            if not part:
+                continue
+            elif part == "{{":
+                e.append(r"\{")
+            elif part == "}}":
+                e.append(r"\}")
+            elif part[0] == "{" and part[-1] == "}":
+                # this will be a braces-delimited field to handle
+                e.append(self._handle_field(part))
+            else:
+                # just some text to match
+                e.append(REGEX_SAFETY.sub(self._regex_replace, part))
+        return "".join(e)
+
+    def _to_group_name(self, field):
+        # return a version of field which can be used as capture group, even
+        # though it might contain '.'
+        group = field.replace(".", "_").replace("[", "_").replace("]", "_").replace("-", "_")
+
+        # make sure we don't collide ("a.b" colliding with "a_b")
+        n = 1
+        while group in self._group_to_name_map:
+            n += 1
+            if "." in field:
+                group = field.replace(".", "_" * n)
+            elif "_" in field:
+                group = field.replace("_", "_" * n)
+            elif "-" in field:
+                group = field.replace("-", "_" * n)
+            else:
+                raise KeyError("duplicated group name %r" % (field,))
+
+        # save off the mapping
+        self._group_to_name_map[group] = field
+        self._name_to_group_map[field] = group
+        return group
+
+    def _handle_field(self, field):
+        # first: lose the braces
+        field = field[1:-1]
+
+        # now figure whether this is an anonymous or named field, and whether
+        # there's any format specification
+        format = ""
+
+        if ":" in field:
+            name, format = field.split(":", 1)
+        else:
+            name = field
+
+        # This *should* be more flexible, but parsing complicated structures
+        # out of the string is hard (and not necessarily useful) ... and I'm
+        # being lazy. So for now `identifier` is "anything starting with a
+        # letter" and digit args don't get attribute or element stuff.
+        if name and name[0].isalpha():
+            if name in self._name_to_group_map:
+                if self._name_types[name] != format:
+                    raise RepeatedNameError(
+                        'field type %r for field "%s" '
+                        "does not match previous seen type %r"
+                        % (format, name, self._name_types[name])
+                    )
+                group = self._name_to_group_map[name]
+                # match previously-seen value
+                return r"(?P=%s)" % group
+            else:
+                group = self._to_group_name(name)
+                self._name_types[name] = format
+            self._named_fields.append(group)
+            # this will become a group, which must not contain dots
+            wrap = r"(?P<%s>%%s)" % group
+        else:
+            self._fixed_fields.append(self._group_index)
+            wrap = r"(%s)"
+            group = self._group_index
+
+        # simplest case: no type specifier ({} or {name})
+        if not format:
+            self._group_index += 1
+            return wrap % r".+?"
+
+        # decode the format specification
+        format = extract_format(format, self._extra_types)
+
+        # figure type conversions, if any
+        type = format["type"]
+        is_numeric = type and type in "n%fegdobx"
+        conv = self._type_conversions
+        if type in self._extra_types:
+            type_converter = self._extra_types[type]
+            s = getattr(type_converter, "pattern", r".+?")
+            regex_group_count = getattr(type_converter, "regex_group_count", 0)
+            if regex_group_count is None:
+                regex_group_count = 0
+            self._group_index += regex_group_count
+            conv[group] = convert_first(type_converter)
+        elif type == "n":
+            s = r"\d{1,3}([,.]\d{3})*"
+            self._group_index += 1
+            conv[group] = int_convert(10)
+        elif type == "b":
+            s = r"(0[bB])?[01]+"
+            conv[group] = int_convert(2)
+            self._group_index += 1
+        elif type == "o":
+            s = r"(0[oO])?[0-7]+"
+            conv[group] = int_convert(8)
+            self._group_index += 1
+        elif type == "x":
+            s = r"(0[xX])?[0-9a-fA-F]+"
+            conv[group] = int_convert(16)
+            self._group_index += 1
+        elif type == "%":
+            s = r"\d+(\.\d+)?%"
+            self._group_index += 1
+            conv[group] = percentage
+        elif type == "f":
+            s = r"\d*\.\d+"
+            conv[group] = convert_first(float)
+        elif type == "F":
+            s = r"\d*\.\d+"
+            conv[group] = convert_first(Decimal)
+        elif type == "e":
+            s = r"\d*\.\d+[eE][-+]?\d+|nan|NAN|[-+]?inf|[-+]?INF"
+            conv[group] = convert_first(float)
+        elif type == "g":
+            s = r"\d+(\.\d+)?([eE][-+]?\d+)?|nan|NAN|[-+]?inf|[-+]?INF"
+            self._group_index += 2
+            conv[group] = convert_first(float)
+        elif type == "d":
+            if format.get("width"):
+                width = r"{1,%s}" % int(format["width"])
+            else:
+                width = "+"
+            s = r"\d{w}|[-+ ]?0[xX][0-9a-fA-F]{w}|[-+ ]?0[bB][01]{w}|[-+ ]?0[oO][0-7]{w}".format(
+                w=width
+            )
+            conv[group] = int_convert()
+            # do not specify number base, determine it automatically
+        elif any(k in type for k in dt_format_to_regex):
+            s = get_regex_for_datetime_format(type)
+            conv[group] = partial(strf_date_convert, type=type)
+        elif type == "ti":
+            s = r"(\d{4}-\d\d-\d\d)((\s+|T)%s)?(Z|\s*[-+]\d\d:?\d\d)?" % TIME_PAT
+            n = self._group_index
+            conv[group] = partial(date_convert, ymd=n + 1, hms=n + 4, tz=n + 7)
+            self._group_index += 7
+        elif type == "tg":
+            s = r"(\d{1,2}[-/](\d{1,2}|%s)[-/]\d{4})(\s+%s)?%s?%s?"
+            s %= (ALL_MONTHS_PAT, TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(
+                date_convert, dmy=n + 1, hms=n + 5, am=n + 8, tz=n + 9
+            )
+            self._group_index += 9
+        elif type == "ta":
+            s = r"((\d{1,2}|%s)[-/]\d{1,2}[-/]\d{4})(\s+%s)?%s?%s?"
+            s %= (ALL_MONTHS_PAT, TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(
+                date_convert, mdy=n + 1, hms=n + 5, am=n + 8, tz=n + 9
+            )
+            self._group_index += 9
+        elif type == "te":
+            # this will allow microseconds through if they're present, but meh
+            s = r"(%s,\s+)?(\d{1,2}\s+%s\s+\d{4})\s+%s%s"
+            s %= (DAYS_PAT, MONTHS_PAT, TIME_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, dmy=n + 3, hms=n + 5, tz=n + 8)
+            self._group_index += 8
+        elif type == "th":
+            # slight flexibility here from the stock Apache format
+            s = r"(\d{1,2}[-/]%s[-/]\d{4}):%s%s" % (MONTHS_PAT, TIME_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, dmy=n + 1, hms=n + 3, tz=n + 6)
+            self._group_index += 6
+        elif type == "tc":
+            s = r"(%s)\s+%s\s+(\d{1,2})\s+%s\s+(\d{4})"
+            s %= (DAYS_PAT, MONTHS_PAT, TIME_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, d_m_y=(n + 4, n + 3, n + 8), hms=n + 5)
+            self._group_index += 8
+        elif type == "tt":
+            s = r"%s?%s?%s?" % (TIME_PAT, AM_PAT, TZ_PAT)
+            n = self._group_index
+            conv[group] = partial(date_convert, hms=n + 1, am=n + 4, tz=n + 5)
+            self._group_index += 5
+        elif type == "ts":
+            s = r"%s(\s+)(\d+)(\s+)(\d{1,2}:\d{1,2}:\d{1,2})?" % MONTHS_PAT
+            n = self._group_index
+            conv[group] = partial(date_convert, mm=n + 1, dd=n + 3, hms=n + 5)
+            self._group_index += 5
+        elif type == "l":
+            s = r"[A-Za-z]+"
+        elif type:
+            s = r"\%s+" % type
+        elif format.get("precision"):
+            if format.get("width"):
+                s = r".{%s,%s}?" % (format["width"], format["precision"])
+            else:
+                s = r".{1,%s}?" % format["precision"]
+        elif format.get("width"):
+            s = r".{%s,}?" % format["width"]
+        else:
+            s = r".+?"
+
+        align = format["align"]
+        fill = format["fill"]
+
+        # handle some numeric-specific things like fill and sign
+        if is_numeric:
+            # prefix with something (align "=" trumps zero)
+            if align == "=":
+                # special case - align "=" acts like the zero above but with
+                # configurable fill defaulting to "0"
+                if not fill:
+                    fill = "0"
+                s = r"%s*" % fill + s
+
+            # allow numbers to be prefixed with a sign
+            s = r"[-+ ]?" + s
+
+        if not fill:
+            fill = " "
+
+        # Place into a group now - this captures the value we want to keep.
+        # Everything else from now is just padding to be stripped off
+        if wrap:
+            s = wrap % s
+            self._group_index += 1
+
+        if format["width"]:
+            # all we really care about is that if the format originally
+            # specified a width then there will probably be padding - without
+            # an explicit alignment that'll mean right alignment with spaces
+            # padding
+            if not align:
+                align = ">"
+
+        if fill in r".\+?*[](){}^$":
+            fill = "\\" + fill
+
+        # align "=" has been handled
+        if align == "<":
+            s = "%s%s*" % (s, fill)
+        elif align == ">":
+            s = "%s*%s" % (fill, s)
+        elif align == "^":
+            s = "%s*%s%s*" % (fill, s, fill)
+
+        return s
+
+
+class Result(object):
+    """The result of a parse() or search().
+
+    Fixed results may be looked up using `result[index]`.
+    Slices of fixed results may also be looked up.
+
+    Named results may be looked up using `result['name']`.
+
+    Named results may be tested for existence using `'name' in result`.
+    """
+
+    def __init__(self, fixed, named, spans):
+        self.fixed = fixed
+        self.named = named
+        self.spans = spans
+
+    def __getitem__(self, item):
+        if isinstance(item, (int, slice)):
+            return self.fixed[item]
+        return self.named[item]
+
+    def __repr__(self):
+        return "<%s %r %r>" % (self.__class__.__name__, self.fixed, self.named)
+
+    def __contains__(self, name):
+        return name in self.named
+
+
+class Match(object):
+    """The result of a parse() or search() if no results are generated.
+
+    This class is only used to expose internal used regex match objects
+    to the user and use them for external Parser.evaluate_result calls.
+    """
+
+    def __init__(self, parser, match):
+        self.parser = parser
+        self.match = match
+
+    def evaluate_result(self):
+        """Generate results for this Match"""
+        return self.parser.evaluate_result(self.match)
+
+
+class ResultIterator(object):
+    """The result of a findall() operation.
+
+    Each element is a Result instance.
+    """
+
+    def __init__(self, parser, string, pos, endpos, evaluate_result=True):
+        self.parser = parser
+        self.string = string
+        self.pos = pos
+        self.endpos = endpos
+        self.evaluate_result = evaluate_result
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        m = self.parser._search_re.search(self.string, self.pos, self.endpos)
+        if m is None:
+            raise StopIteration()
+        self.pos = m.end()
+
+        if self.evaluate_result:
+            return self.parser.evaluate_result(m)
+        else:
+            return Match(self.parser, m)
+
+    # pre-py3k compat
+    next = __next__
+
+
+def parse(format, string, extra_types=None, evaluate_result=True, case_sensitive=False):
+    """Using "format" attempt to pull values from "string".
+
+    The format must match the string contents exactly. If the value
+    you're looking for is instead just a part of the string use
+    search().
+
+    If ``evaluate_result`` is True the return value will be an Result instance with two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False the return value will be a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+
+    In the case there is no match parse() will return None.
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.parse(string, evaluate_result=evaluate_result)
+
+
+def search(
+    format,
+    string,
+    pos=0,
+    endpos=None,
+    extra_types=None,
+    evaluate_result=True,
+    case_sensitive=False,
+):
+    """Search "string" for the first occurrence of "format".
+
+    The format may occur anywhere within the string. If
+    instead you wish for the format to exactly match the string
+    use parse().
+
+    Optionally start the search at "pos" character index and limit the search
+    to a maximum index of endpos - equivalent to search(string[:endpos]).
+
+    If ``evaluate_result`` is True the return value will be an Result instance with two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False the return value will be a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+
+    In the case there is no match parse() will return None.
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.search(string, pos, endpos, evaluate_result=evaluate_result)
+
+
+def findall(
+    format,
+    string,
+    pos=0,
+    endpos=None,
+    extra_types=None,
+    evaluate_result=True,
+    case_sensitive=False,
+):
+    """Search "string" for all occurrences of "format".
+
+    You will be returned an iterator that holds Result instances
+    for each format match found.
+
+    Optionally start the search at "pos" character index and limit the search
+    to a maximum index of endpos - equivalent to search(string[:endpos]).
+
+    If ``evaluate_result`` is True each returned Result instance has two attributes:
+
+     .fixed - tuple of fixed-position values from the string
+     .named - dict of named values from the string
+
+    If ``evaluate_result`` is False each returned value is a Match instance with one method:
+
+     .evaluate_result() - This will return a Result instance like you would get
+                          with ``evaluate_result`` set to True
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    If the format is invalid a ValueError will be raised.
+
+    See the module documentation for the use of "extra_types".
+    """
+    p = Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+    return p.findall(string, pos, endpos, evaluate_result=evaluate_result)
+
+
+def compile(format, extra_types=None, case_sensitive=False):
+    """Create a Parser instance to parse "format".
+
+    The resultant Parser has a method .parse(string) which
+    behaves in the same manner as parse(format, string).
+
+    The default behaviour is to match strings case insensitively. You may match with
+    case by specifying case_sensitive=True.
+
+    Use this function if you intend to parse many strings
+    with the same format.
+
+    See the module documentation for the use of "extra_types".
+
+    Returns a Parser instance.
+    """
+    return Parser(format, extra_types=extra_types, case_sensitive=case_sensitive)
+
+
+# Copyright (c) 2012-2020 Richard Jones <richard@python.org>
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+#  The above copyright notice and this permission notice shall be included in
+#  all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# vim: set filetype=python ts=4 sw=4 et si tw=75
diff --git a/backend/venv/lib/python3.12/site-packages/parse_type/parse_util.py b/backend/venv/lib/python3.12/site-packages/parse_type/parse_util.py
new file mode 100644
index 00000000..0e5ee735
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/parse_type/parse_util.py
@@ -0,0 +1,198 @@
+# -*- coding: utf-8 -*-
+# pylint: disable=missing-docstring
+"""
+Provides generic utility classes for the :class:`parse.Parser` class.
+"""
+
+from __future__ import absolute_import
+from collections import namedtuple
+import parse
+import six
+
+
+# -- HELPER-CLASS: For format part in a Field.
+# REQUIRES: Python 2.6 or newer.
+# pylint: disable=redefined-builtin, too-many-arguments
+FormatSpec = namedtuple("FormatSpec",
+                        ["type", "width", "zero", "align", "fill", "precision"])
+
+def make_format_spec(type=None, width="", zero=False, align=None, fill=None,
+                     precision=None):
+    return FormatSpec(type, width, zero, align, fill, precision)
+# pylint: enable=redefined-builtin
+
+class Field(object):
+    """
+    Provides a ValueObject for a Field in a parse expression.
+
+    Examples:
+        * "{}"
+        * "{name}"
+        * "{:format}"
+        * "{name:format}"
+
+    Format specification: [[fill]align][0][width][.precision][type]
+    """
+    # pylint: disable=redefined-builtin
+    ALIGN_CHARS = '<>=^'
+
+    def __init__(self, name="", format=None):
+        self.name = name
+        self.format = format
+        self._format_spec = None
+
+    def set_format(self, format):
+        self.format = format
+        self._format_spec = None
+
+    @property
+    def has_format(self):
+        return bool(self.format)
+
+    @property
+    def format_spec(self):
+        if not self._format_spec and self.format:
+            self._format_spec = self.extract_format_spec(self.format)
+        return self._format_spec
+
+    def __str__(self):
+        name = self.name or ""
+        if self.has_format:
+            return "{%s:%s}" % (name, self.format)
+        return "{%s}" % name
+
+    def __eq__(self, other):
+        if isinstance(other, Field):
+            format1 = self.format or ""
+            format2 = other.format or ""
+            return (self.name == other.name) and (format1 == format2)
+        elif isinstance(other, six.string_types):
+            return str(self) == other
+        else:
+            raise ValueError(other)
+
+    def __ne__(self, other):
+        return not self.__eq__(other)
+
+    @staticmethod
+    def make_format(format_spec):
+        """Build format string from a format specification.
+
+        :param format_spec:     Format specification (as FormatSpec object).
+        :return: Composed format (as string).
+        """
+        fill = ''
+        align = ''
+        zero = ''
+        width = format_spec.width
+        if format_spec.align:
+            align = format_spec.align[0]
+            if format_spec.fill:
+                fill = format_spec.fill[0]
+        if format_spec.zero:
+            zero = '0'
+
+        precision_part = ""
+        if format_spec.precision:
+            precision_part = ".%s" % format_spec.precision
+
+        # -- FORMAT-SPEC: [[fill]align][0][width][.precision][type]
+        return "%s%s%s%s%s%s" % (fill, align, zero, width,
+                                 precision_part, format_spec.type)
+
+
+    @classmethod
+    def extract_format_spec(cls, format):
+        """Pull apart the format: [[fill]align][0][width][.precision][type]"""
+        # -- BASED-ON: parse.extract_format()
+        # pylint: disable=redefined-builtin, unsubscriptable-object
+        if not format:
+            raise ValueError("INVALID-FORMAT: %s (empty-string)" % format)
+
+        orig_format = format
+        fill = align = None
+        if format[0] in cls.ALIGN_CHARS:
+            align = format[0]
+            format = format[1:]
+        elif len(format) > 1 and format[1] in cls.ALIGN_CHARS:
+            fill = format[0]
+            align = format[1]
+            format = format[2:]
+
+        zero = False
+        if format and format[0] == '0':
+            zero = True
+            format = format[1:]
+
+        width = ''
+        while format:
+            if not format[0].isdigit():
+                break
+            width += format[0]
+            format = format[1:]
+
+        precision = None
+        if format.startswith('.'):
+            # Precision isn't needed but we need to capture it so that
+            # the ValueError isn't raised.
+            format = format[1:]  # drop the '.'
+            precision = ''
+            while format:
+                if not format[0].isdigit():
+                    break
+                precision += format[0]
+                format = format[1:]
+
+        # the rest is the type, if present
+        type = format
+        if not type:
+            raise ValueError("INVALID-FORMAT: %s (without type)" % orig_format)
+        return FormatSpec(type, width, zero, align, fill, precision)
+
+
+class FieldParser(object):
+    """
+    Utility class that parses/extracts fields in parse expressions.
+    """
+
+    @classmethod
+    def parse(cls, text):
+        if not (text.startswith('{') and text.endswith('}')):
+            message = "FIELD-SCHEMA MISMATCH: text='%s' (missing braces)" % text
+            raise ValueError(message)
+
+        # first: lose the braces
+        text = text[1:-1]
+        if ':' in text:
+            # -- CASE: Typed field with format.
+            name, format_ = text.split(':')
+        else:
+            name = text
+            format_ = None
+        return Field(name, format_)
+
+    @classmethod
+    def extract_fields(cls, schema):
+        """Extract fields in a parse expression schema.
+
+        :param schema: Parse expression schema/format to use (as string).
+        :return: Generator for fields in schema (as Field objects).
+        """
+        # -- BASED-ON: parse.Parser._generate_expression()
+        for part in parse.PARSE_RE.split(schema):
+            if not part or part == '{{' or part == '}}':
+                continue
+            elif part[0] == '{':
+                # this will be a braces-delimited field to handle
+                yield cls.parse(part)
+
+    @classmethod
+    def extract_types(cls, schema):
+        """Extract types (names) for typed fields (with format/type part).
+
+        :param schema: Parser schema/format to use.
+        :return: Generator for type names (as string).
+        """
+        for field in cls.extract_fields(schema):
+            if field.has_format:
+                yield field.format_spec.type
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/LICENSE
new file mode 100644
index 00000000..cf56bed0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/LICENSE
@@ -0,0 +1,1343 @@
+The Python Imaging Library (PIL) is
+
+    Copyright © 1997-2011 by Secret Labs AB
+    Copyright © 1995-2011 by Fredrik Lundh and contributors
+
+Pillow is the friendly PIL fork. It is
+
+    Copyright © 2010 by Jeffrey A. Clark and contributors
+
+Like PIL, Pillow is licensed under the open source MIT-CMU License:
+
+By obtaining, using, and/or copying this software and/or its associated
+documentation, you agree that you have read, understood, and will comply
+with the following terms and conditions:
+
+Permission to use, copy, modify and distribute this software and its
+documentation for any purpose and without fee is hereby granted,
+provided that the above copyright notice appears in all copies, and that
+both that copyright notice and this permission notice appear in supporting
+documentation, and that the name of Secret Labs AB or the author not be
+used in advertising or publicity pertaining to distribution of the software
+without specific, written prior permission.
+
+SECRET LABS AB AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS
+SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
+IN NO EVENT SHALL SECRET LABS AB OR THE AUTHOR BE LIABLE FOR ANY SPECIAL,
+INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
+OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+PERFORMANCE OF THIS SOFTWARE.
+
+
+----
+
+BROTLI
+
+Copyright (c) 2009, 2010, 2013-2016 by the Brotli Authors.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.
+
+
+----
+
+BZIP2
+
+
+--------------------------------------------------------------------------
+
+This program, "bzip2", the associated library "libbzip2", and all
+documentation, are copyright (C) 1996-2019 Julian R Seward.  All
+rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions
+are met:
+
+1. Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+2. The origin of this software must not be misrepresented; you must
+   not claim that you wrote the original software.  If you use this
+   software in a product, an acknowledgment in the product
+   documentation would be appreciated but is not required.
+
+3. Altered source versions must be plainly marked as such, and must
+   not be misrepresented as being the original software.
+
+4. The name of the author may not be used to endorse or promote
+   products derived from this software without specific prior written
+   permission.
+
+THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS
+OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
+DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE
+GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+Julian Seward, jseward@acm.org
+bzip2/libbzip2 version 1.0.8 of 13 July 2019
+
+--------------------------------------------------------------------------
+
+
+----
+
+FREETYPE2
+
+The FreeType  2 font  engine is  copyrighted work  and cannot  be used
+legally without  a software  license.  In order  to make  this project
+usable to  a vast majority of  developers, we distribute it  under two
+mutually exclusive open-source licenses.
+
+This means that *you* must choose  *one* of the two licenses described
+below, then obey all its terms and conditions when using FreeType 2 in
+any of your projects or products.
+
+  - The FreeType License,  found in the file  `docs/FTL.TXT`, which is
+    similar to the  original BSD license *with*  an advertising clause
+    that forces  you to explicitly  cite the FreeType project  in your
+    product's  documentation.  All  details are  in the  license file.
+    This license is suited to products which don't use the GNU General
+    Public License.
+
+    Note that  this license  is compatible to  the GNU  General Public
+    License version 3, but not version 2.
+
+  - The   GNU   General   Public   License   version   2,   found   in
+    `docs/GPLv2.TXT`  (any  later  version  can  be  used  also),  for
+    programs  which  already  use  the  GPL.  Note  that  the  FTL  is
+    incompatible with GPLv2 due to its advertisement clause.
+
+The contributed  BDF and PCF  drivers come  with a license  similar to
+that  of the  X Window  System.   It is  compatible to  the above  two
+licenses (see files `src/bdf/README`  and `src/pcf/README`).  The same
+holds   for   the   source    code   files   `src/base/fthash.c`   and
+`include/freetype/internal/fthash.h`; they were part of the BDF driver
+in earlier FreeType versions.
+
+The gzip  module uses the  zlib license (see  `src/gzip/zlib.h`) which
+too is compatible to the above two licenses.
+
+The files `src/autofit/ft-hb.c` and `src/autofit/ft-hb.h` contain code
+taken almost  verbatim from the  HarfBuzz file `hb-ft.cc`,  which uses
+the 'Old MIT' license, compatible to the above two licenses.
+
+The  MD5 checksum  support  (only used  for  debugging in  development
+builds) is in the public domain.
+
+--------------------------------------------------------------------------
+
+                    The FreeType Project LICENSE
+                    ----------------------------
+
+                            2006-Jan-27
+
+                    Copyright 1996-2002, 2006 by
+          David Turner, Robert Wilhelm, and Werner Lemberg
+
+
+
+Introduction
+============
+
+  The FreeType  Project is distributed in  several archive packages;
+  some of them may contain, in addition to the FreeType font engine,
+  various tools and  contributions which rely on, or  relate to, the
+  FreeType Project.
+
+  This  license applies  to all  files found  in such  packages, and
+  which do not  fall under their own explicit  license.  The license
+  affects  thus  the  FreeType   font  engine,  the  test  programs,
+  documentation and makefiles, at the very least.
+
+  This  license   was  inspired  by  the  BSD,   Artistic,  and  IJG
+  (Independent JPEG  Group) licenses, which  all encourage inclusion
+  and  use of  free  software in  commercial  and freeware  products
+  alike.  As a consequence, its main points are that:
+
+    o We don't promise that this software works. However, we will be
+      interested in any kind of bug reports. (`as is' distribution)
+
+    o You can  use this software for whatever you  want, in parts or
+      full form, without having to pay us. (`royalty-free' usage)
+
+    o You may not pretend that  you wrote this software.  If you use
+      it, or  only parts of it,  in a program,  you must acknowledge
+      somewhere  in  your  documentation  that  you  have  used  the
+      FreeType code. (`credits')
+
+  We  specifically  permit  and  encourage  the  inclusion  of  this
+  software, with  or without modifications,  in commercial products.
+  We  disclaim  all warranties  covering  The  FreeType Project  and
+  assume no liability related to The FreeType Project.
+
+
+  Finally,  many  people  asked  us  for  a  preferred  form  for  a
+  credit/disclaimer to use in compliance with this license.  We thus
+  encourage you to use the following text:
+
+   """
+    Portions of this software are copyright © <year> The FreeType
+    Project (www.freetype.org).  All rights reserved.
+   """
+
+  Please replace <year> with the value from the FreeType version you
+  actually use.
+
+
+Legal Terms
+===========
+
+0. Definitions
+--------------
+
+  Throughout this license,  the terms `package', `FreeType Project',
+  and  `FreeType  archive' refer  to  the  set  of files  originally
+  distributed  by the  authors  (David Turner,  Robert Wilhelm,  and
+  Werner Lemberg) as the `FreeType Project', be they named as alpha,
+  beta or final release.
+
+  `You' refers to  the licensee, or person using  the project, where
+  `using' is a generic term including compiling the project's source
+  code as  well as linking it  to form a  `program' or `executable'.
+  This  program is  referred to  as  `a program  using the  FreeType
+  engine'.
+
+  This  license applies  to all  files distributed  in  the original
+  FreeType  Project,   including  all  source   code,  binaries  and
+  documentation,  unless  otherwise  stated   in  the  file  in  its
+  original, unmodified form as  distributed in the original archive.
+  If you are  unsure whether or not a particular  file is covered by
+  this license, you must contact us to verify this.
+
+  The FreeType  Project is copyright (C) 1996-2000  by David Turner,
+  Robert Wilhelm, and Werner Lemberg.  All rights reserved except as
+  specified below.
+
+1. No Warranty
+--------------
+
+  THE FREETYPE PROJECT  IS PROVIDED `AS IS' WITHOUT  WARRANTY OF ANY
+  KIND, EITHER  EXPRESS OR IMPLIED,  INCLUDING, BUT NOT  LIMITED TO,
+  WARRANTIES  OF  MERCHANTABILITY   AND  FITNESS  FOR  A  PARTICULAR
+  PURPOSE.  IN NO EVENT WILL ANY OF THE AUTHORS OR COPYRIGHT HOLDERS
+  BE LIABLE  FOR ANY DAMAGES CAUSED  BY THE USE OR  THE INABILITY TO
+  USE, OF THE FREETYPE PROJECT.
+
+2. Redistribution
+-----------------
+
+  This  license  grants  a  worldwide, royalty-free,  perpetual  and
+  irrevocable right  and license to use,  execute, perform, compile,
+  display,  copy,   create  derivative  works   of,  distribute  and
+  sublicense the  FreeType Project (in  both source and  object code
+  forms)  and  derivative works  thereof  for  any  purpose; and  to
+  authorize others  to exercise  some or all  of the  rights granted
+  herein, subject to the following conditions:
+
+    o Redistribution of  source code  must retain this  license file
+      (`FTL.TXT') unaltered; any  additions, deletions or changes to
+      the original  files must be clearly  indicated in accompanying
+      documentation.   The  copyright   notices  of  the  unaltered,
+      original  files must  be  preserved in  all  copies of  source
+      files.
+
+    o Redistribution in binary form must provide a  disclaimer  that
+      states  that  the software is based in part of the work of the
+      FreeType Team,  in  the  distribution  documentation.  We also
+      encourage you to put an URL to the FreeType web page  in  your
+      documentation, though this isn't mandatory.
+
+  These conditions  apply to any  software derived from or  based on
+  the FreeType Project,  not just the unmodified files.   If you use
+  our work, you  must acknowledge us.  However, no  fee need be paid
+  to us.
+
+3. Advertising
+--------------
+
+  Neither the  FreeType authors and  contributors nor you  shall use
+  the name of the  other for commercial, advertising, or promotional
+  purposes without specific prior written permission.
+
+  We suggest,  but do not require, that  you use one or  more of the
+  following phrases to refer  to this software in your documentation
+  or advertising  materials: `FreeType Project',  `FreeType Engine',
+  `FreeType library', or `FreeType Distribution'.
+
+  As  you have  not signed  this license,  you are  not  required to
+  accept  it.   However,  as  the FreeType  Project  is  copyrighted
+  material, only  this license, or  another one contracted  with the
+  authors, grants you  the right to use, distribute,  and modify it.
+  Therefore,  by  using,  distributing,  or modifying  the  FreeType
+  Project, you indicate that you understand and accept all the terms
+  of this license.
+
+4. Contacts
+-----------
+
+  There are two mailing lists related to FreeType:
+
+    o freetype@nongnu.org
+
+      Discusses general use and applications of FreeType, as well as
+      future and  wanted additions to the  library and distribution.
+      If  you are looking  for support,  start in  this list  if you
+      haven't found anything to help you in the documentation.
+
+    o freetype-devel@nongnu.org
+
+      Discusses bugs,  as well  as engine internals,  design issues,
+      specific licenses, porting, etc.
+
+  Our home page can be found at
+
+    https://www.freetype.org
+
+
+--- end of FTL.TXT ---
+
+--------------------------------------------------------------------------
+
+            GNU GENERAL PUBLIC LICENSE
+               Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.
+     51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Library General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+            GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+             END OF TERMS AND CONDITIONS
+
+        How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year  name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Library General
+Public License instead of this License.
+
+--------------------------------------------------------------------------
+
+The following license details are part of `src/bdf/README`:
+
+```
+License
+*******
+
+Copyright (C) 2001-2002 by Francesco Zappa Nardelli
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+*** Portions of the driver (that is, bdflib.c and bdf.h):
+
+Copyright 2000 Computing Research Labs, New Mexico State University
+Copyright 2001-2002, 2011 Francesco Zappa Nardelli
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and associated documentation files (the "Software"),
+to deal in the Software without restriction, including without limitation
+the rights to use, copy, modify, merge, publish, distribute, sublicense,
+and/or sell copies of the Software, and to permit persons to whom the
+Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+THE COMPUTING RESEARCH LAB OR NEW MEXICO STATE UNIVERSITY BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT
+OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR
+THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+
+Credits
+*******
+
+This driver is based on excellent Mark Leisher's bdf library.  If you
+find something good in this driver you should probably thank him, not
+me.
+```
+
+The following license details are part of `src/pcf/README`:
+
+```
+License
+*******
+
+Copyright (C) 2000 by Francesco Zappa Nardelli
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+
+Credits
+*******
+
+Keith Packard wrote the pcf driver found in XFree86.  His work is at
+the same time the specification and the sample implementation of the
+PCF format.  Undoubtedly, this driver is inspired from his work.
+```
+
+
+----
+
+HARFBUZZ
+
+HarfBuzz is licensed under the so-called "Old MIT" license.  Details follow.
+For parts of HarfBuzz that are licensed under different licenses see individual
+files names COPYING in subdirectories where applicable.
+
+Copyright © 2010-2022  Google, Inc.
+Copyright © 2015-2020  Ebrahim Byagowi
+Copyright © 2019,2020  Facebook, Inc.
+Copyright © 2012,2015  Mozilla Foundation
+Copyright © 2011  Codethink Limited
+Copyright © 2008,2010  Nokia Corporation and/or its subsidiary(-ies)
+Copyright © 2009  Keith Stribley
+Copyright © 2011  Martin Hosken and SIL International
+Copyright © 2007  Chris Wilson
+Copyright © 2005,2006,2020,2021,2022,2023  Behdad Esfahbod
+Copyright © 2004,2007,2008,2009,2010,2013,2021,2022,2023  Red Hat, Inc.
+Copyright © 1998-2005  David Turner and Werner Lemberg
+Copyright © 2016  Igalia S.L.
+Copyright © 2022  Matthias Clasen
+Copyright © 2018,2021  Khaled Hosny
+Copyright © 2018,2019,2020  Adobe, Inc
+Copyright © 2013-2015  Alexei Podtelezhnikov
+
+For full copyright notices consult the individual files in the package.
+
+
+Permission is hereby granted, without written agreement and without
+license or royalty fees, to use, copy, modify, and distribute this
+software and its documentation for any purpose, provided that the
+above copyright notice and the following two paragraphs appear in
+all copies of this software.
+
+IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
+DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
+ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
+IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
+DAMAGE.
+
+THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
+BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
+FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
+ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
+PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
+
+
+----
+
+LCMS2
+
+Little CMS
+Copyright (c) 1998-2020 Marti Maria Saguer
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+
+----
+
+LIBJPEG
+
+1. We don't promise that this software works.  (But if you find any bugs,
+   please let us know!)
+2. You can use this software for whatever you want.  You don't have to pay us.
+3. You may not pretend that you wrote this software.  If you use it in a
+   program, you must acknowledge somewhere in your documentation that
+   you've used the IJG code.
+
+In legalese:
+
+The authors make NO WARRANTY or representation, either express or implied,
+with respect to this software, its quality, accuracy, merchantability, or
+fitness for a particular purpose.  This software is provided "AS IS", and you,
+its user, assume the entire risk as to its quality and accuracy.
+
+This software is copyright (C) 1991-2020, Thomas G. Lane, Guido Vollbeding.
+All Rights Reserved except as specified below.
+
+Permission is hereby granted to use, copy, modify, and distribute this
+software (or portions thereof) for any purpose, without fee, subject to these
+conditions:
+(1) If any part of the source code for this software is distributed, then this
+README file must be included, with this copyright and no-warranty notice
+unaltered; and any additions, deletions, or changes to the original files
+must be clearly indicated in accompanying documentation.
+(2) If only executable code is distributed, then the accompanying
+documentation must state that "this software is based in part on the work of
+the Independent JPEG Group".
+(3) Permission for use of this software is granted only if the user accepts
+full responsibility for any undesirable consequences; the authors accept
+NO LIABILITY for damages of any kind.
+
+These conditions apply to any software derived from or based on the IJG code,
+not just to the unmodified library.  If you use our work, you ought to
+acknowledge us.
+
+Permission is NOT granted for the use of any IJG author's name or company name
+in advertising or publicity relating to this software or products derived from
+it.  This software may be referred to only as "the Independent JPEG Group's
+software".
+
+We specifically permit and encourage the use of this software as the basis of
+commercial products, provided that all warranty or liability claims are
+assumed by the product vendor.
+
+
+----
+
+LIBLZMA
+
+XZ Utils Licensing
+==================
+
+    Different licenses apply to different files in this package. Here
+    is a rough summary of which licenses apply to which parts of this
+    package (but check the individual files to be sure!):
+
+      - liblzma is in the public domain.
+
+      - xz, xzdec, and lzmadec command line tools are in the public
+        domain unless GNU getopt_long had to be compiled and linked
+        in from the lib directory. The getopt_long code is under
+        GNU LGPLv2.1+.
+
+      - The scripts to grep, diff, and view compressed files have been
+        adapted from gzip. These scripts and their documentation are
+        under GNU GPLv2+.
+
+      - All the documentation in the doc directory and most of the
+        XZ Utils specific documentation files in other directories
+        are in the public domain.
+
+      - Translated messages are in the public domain.
+
+      - The build system contains public domain files, and files that
+        are under GNU GPLv2+ or GNU GPLv3+. None of these files end up
+        in the binaries being built.
+
+      - Test files and test code in the tests directory, and debugging
+        utilities in the debug directory are in the public domain.
+
+      - The extra directory may contain public domain files, and files
+        that are under various free software licenses.
+
+    You can do whatever you want with the files that have been put into
+    the public domain. If you find public domain legally problematic,
+    take the previous sentence as a license grant. If you still find
+    the lack of copyright legally problematic, you have too many
+    lawyers.
+
+    As usual, this software is provided "as is", without any warranty.
+
+    If you copy significant amounts of public domain code from XZ Utils
+    into your project, acknowledging this somewhere in your software is
+    polite (especially if it is proprietary, non-free software), but
+    naturally it is not legally required. Here is an example of a good
+    notice to put into "about box" or into documentation:
+
+        This software includes code from XZ Utils <http://tukaani.org/xz/>.
+
+    The following license texts are included in the following files:
+      - COPYING.LGPLv2.1: GNU Lesser General Public License version 2.1
+      - COPYING.GPLv2: GNU General Public License version 2
+      - COPYING.GPLv3: GNU General Public License version 3
+
+    Note that the toolchain (compiler, linker etc.) may add some code
+    pieces that are copyrighted. Thus, it is possible that e.g. liblzma
+    binary wouldn't actually be in the public domain in its entirety
+    even though it contains no copyrighted code from the XZ Utils source
+    package.
+
+    If you have questions, don't hesitate to ask the author(s) for more
+    information.
+
+
+----
+
+LIBPNG
+
+COPYRIGHT NOTICE, DISCLAIMER, and LICENSE
+=========================================
+
+PNG Reference Library License version 2
+---------------------------------------
+
+ * Copyright (c) 1995-2022 The PNG Reference Library Authors.
+ * Copyright (c) 2018-2022 Cosmin Truta.
+ * Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson.
+ * Copyright (c) 1996-1997 Andreas Dilger.
+ * Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.
+
+The software is supplied "as is", without warranty of any kind,
+express or implied, including, without limitation, the warranties
+of merchantability, fitness for a particular purpose, title, and
+non-infringement.  In no event shall the Copyright owners, or
+anyone distributing the software, be liable for any damages or
+other liability, whether in contract, tort or otherwise, arising
+from, out of, or in connection with the software, or the use or
+other dealings in the software, even if advised of the possibility
+of such damage.
+
+Permission is hereby granted to use, copy, modify, and distribute
+this software, or portions hereof, for any purpose, without fee,
+subject to the following restrictions:
+
+ 1. The origin of this software must not be misrepresented; you
+    must not claim that you wrote the original software.  If you
+    use this software in a product, an acknowledgment in the product
+    documentation would be appreciated, but is not required.
+
+ 2. Altered source versions must be plainly marked as such, and must
+    not be misrepresented as being the original software.
+
+ 3. This Copyright notice may not be removed or altered from any
+    source or altered source distribution.
+
+
+PNG Reference Library License version 1 (for libpng 0.5 through 1.6.35)
+-----------------------------------------------------------------------
+
+libpng versions 1.0.7, July 1, 2000, through 1.6.35, July 15, 2018 are
+Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson, are
+derived from libpng-1.0.6, and are distributed according to the same
+disclaimer and license as libpng-1.0.6 with the following individuals
+added to the list of Contributing Authors:
+
+    Simon-Pierre Cadieux
+    Eric S. Raymond
+    Mans Rullgard
+    Cosmin Truta
+    Gilles Vollant
+    James Yu
+    Mandar Sahastrabuddhe
+    Google Inc.
+    Vadim Barkov
+
+and with the following additions to the disclaimer:
+
+    There is no warranty against interference with your enjoyment of
+    the library or against infringement.  There is no warranty that our
+    efforts or the library will fulfill any of your particular purposes
+    or needs.  This library is provided with all faults, and the entire
+    risk of satisfactory quality, performance, accuracy, and effort is
+    with the user.
+
+Some files in the "contrib" directory and some configure-generated
+files that are distributed with libpng have other copyright owners, and
+are released under other open source licenses.
+
+libpng versions 0.97, January 1998, through 1.0.6, March 20, 2000, are
+Copyright (c) 1998-2000 Glenn Randers-Pehrson, are derived from
+libpng-0.96, and are distributed according to the same disclaimer and
+license as libpng-0.96, with the following individuals added to the
+list of Contributing Authors:
+
+    Tom Lane
+    Glenn Randers-Pehrson
+    Willem van Schaik
+
+libpng versions 0.89, June 1996, through 0.96, May 1997, are
+Copyright (c) 1996-1997 Andreas Dilger, are derived from libpng-0.88,
+and are distributed according to the same disclaimer and license as
+libpng-0.88, with the following individuals added to the list of
+Contributing Authors:
+
+    John Bowler
+    Kevin Bracey
+    Sam Bushell
+    Magnus Holmgren
+    Greg Roelofs
+    Tom Tanner
+
+Some files in the "scripts" directory have other copyright owners,
+but are released under this license.
+
+libpng versions 0.5, May 1995, through 0.88, January 1996, are
+Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.
+
+For the purposes of this copyright and license, "Contributing Authors"
+is defined as the following set of individuals:
+
+    Andreas Dilger
+    Dave Martindale
+    Guy Eric Schalnat
+    Paul Schmidt
+    Tim Wegner
+
+The PNG Reference Library is supplied "AS IS".  The Contributing
+Authors and Group 42, Inc. disclaim all warranties, expressed or
+implied, including, without limitation, the warranties of
+merchantability and of fitness for any purpose.  The Contributing
+Authors and Group 42, Inc. assume no liability for direct, indirect,
+incidental, special, exemplary, or consequential damages, which may
+result from the use of the PNG Reference Library, even if advised of
+the possibility of such damage.
+
+Permission is hereby granted to use, copy, modify, and distribute this
+source code, or portions hereof, for any purpose, without fee, subject
+to the following restrictions:
+
+ 1. The origin of this source code must not be misrepresented.
+
+ 2. Altered versions must be plainly marked as such and must not
+    be misrepresented as being the original source.
+
+ 3. This Copyright notice may not be removed or altered from any
+    source or altered source distribution.
+
+The Contributing Authors and Group 42, Inc. specifically permit,
+without fee, and encourage the use of this source code as a component
+to supporting the PNG file format in commercial products.  If you use
+this source code in a product, acknowledgment is not required but would
+be appreciated.
+
+
+----
+
+LIBTIFF
+
+Copyright (c) 1988-1997 Sam Leffler
+Copyright (c) 1991-1997 Silicon Graphics, Inc.
+
+Permission to use, copy, modify, distribute, and sell this software and
+its documentation for any purpose is hereby granted without fee, provided
+that (i) the above copyright notices and this permission notice appear in
+all copies of the software and related documentation, and (ii) the names of
+Sam Leffler and Silicon Graphics may not be used in any advertising or
+publicity relating to the software without the specific, prior written
+permission of Sam Leffler and Silicon Graphics.
+
+THE SOFTWARE IS PROVIDED "AS-IS" AND WITHOUT WARRANTY OF ANY KIND,
+EXPRESS, IMPLIED OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY
+WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
+
+IN NO EVENT SHALL SAM LEFFLER OR SILICON GRAPHICS BE LIABLE FOR
+ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND,
+OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
+WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF
+LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+OF THIS SOFTWARE.
+
+
+----
+
+LIBWEBP
+
+Copyright (c) 2010, Google Inc. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+  * Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+  * Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in
+    the documentation and/or other materials provided with the
+    distribution.
+
+  * Neither the name of Google nor the names of its contributors may
+    be used to endorse or promote products derived from this software
+    without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+----
+
+OPENJPEG
+
+*
+ * The copyright in this software is being made available under the 2-clauses
+ * BSD License, included below. This software may be subject to other third
+ * party and contributor rights, including patent rights, and no such rights
+ * are granted under this license.
+ *
+ * Copyright (c) 2002-2014, Universite catholique de Louvain (UCL), Belgium
+ * Copyright (c) 2002-2014, Professor Benoit Macq
+ * Copyright (c) 2003-2014, Antonin Descampe
+ * Copyright (c) 2003-2009, Francois-Olivier Devaux
+ * Copyright (c) 2005, Herve Drolon, FreeImage Team
+ * Copyright (c) 2002-2003, Yannick Verschueren
+ * Copyright (c) 2001-2003, David Janssens
+ * Copyright (c) 2011-2012, Centre National d'Etudes Spatiales (CNES), France
+ * Copyright (c) 2012, CS Systemes d'Information, France
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS `AS IS'
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+
+----
+
+RAQM
+
+The MIT License (MIT)
+
+Copyright © 2015 Information Technology Authority (ITA) <foss@ita.gov.om>
+Copyright © 2016 Khaled Hosny <khaledhosny@eglug.org>
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+
+----
+
+XAU
+
+Copyright 1988, 1993, 1994, 1998  The Open Group
+
+Permission to use, copy, modify, distribute, and sell this software and its
+documentation for any purpose is hereby granted without fee, provided that
+the above copyright notice appear in all copies and that both that
+copyright notice and this permission notice appear in supporting
+documentation.
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+OPEN GROUP BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+Except as contained in this notice, the name of The Open Group shall not be
+used in advertising or otherwise to promote the sale, use or other dealings
+in this Software without prior written authorization from The Open Group.
+
+
+----
+
+XCB
+
+Copyright (C) 2001-2006 Bart Massey, Jamey Sharp, and Josh Triplett.
+All Rights Reserved.
+
+Permission is hereby granted, free of charge, to any person
+obtaining a copy of this software and associated
+documentation files (the "Software"), to deal in the
+Software without restriction, including without limitation
+the rights to use, copy, modify, merge, publish, distribute,
+sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall
+be included in all copies or substantial portions of the
+Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
+KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
+WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS
+BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+OTHER DEALINGS IN THE SOFTWARE.
+
+Except as contained in this notice, the names of the authors
+or their institutions shall not be used in advertising or
+otherwise to promote the sale, use or other dealings in this
+Software without prior written authorization from the
+authors.
+
+
+----
+
+XDMCP
+
+Copyright 1989, 1998  The Open Group
+
+Permission to use, copy, modify, distribute, and sell this software and its
+documentation for any purpose is hereby granted without fee, provided that
+the above copyright notice appear in all copies and that both that
+copyright notice and this permission notice appear in supporting
+documentation.
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+OPEN GROUP BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+Except as contained in this notice, the name of The Open Group shall not be
+used in advertising or otherwise to promote the sale, use or other dealings
+in this Software without prior written authorization from The Open Group.
+
+Author:  Keith Packard, MIT X Consortium
+
+
+----
+
+ZLIB
+
+ (C) 1995-2017 Jean-loup Gailly and Mark Adler
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+
+  Jean-loup Gailly        Mark Adler
+  jloup@gzip.org          madler@alumni.caltech.edu
+
+If you use the zlib library in a product, we would appreciate *not* receiving
+lengthy legal documents to sign.  The sources are provided for free but without
+warranty of any kind.  The library has been entirely written by Jean-loup
+Gailly and Mark Adler; it does not include third-party code.
+
+If you redistribute modified sources, we would appreciate that you include in
+the file ChangeLog history information documenting your changes.  Please read
+the FAQ for more information on the distribution of modified source versions.
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/METADATA
new file mode 100644
index 00000000..8813a52d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/METADATA
@@ -0,0 +1,176 @@
+Metadata-Version: 2.1
+Name: pillow
+Version: 11.1.0
+Summary: Python Imaging Library (Fork)
+Author-email: "Jeffrey A. Clark" <aclark@aclark.net>
+License: MIT-CMU
+Project-URL: Changelog, https://github.com/python-pillow/Pillow/releases
+Project-URL: Documentation, https://pillow.readthedocs.io
+Project-URL: Funding, https://tidelift.com/subscription/pkg/pypi-pillow?utm_source=pypi-pillow&utm_medium=pypi
+Project-URL: Homepage, https://python-pillow.github.io
+Project-URL: Mastodon, https://fosstodon.org/@pillow
+Project-URL: Release notes, https://pillow.readthedocs.io/en/stable/releasenotes/index.html
+Project-URL: Source, https://github.com/python-pillow/Pillow
+Keywords: Imaging
+Classifier: Development Status :: 6 - Mature
+Classifier: License :: OSI Approved :: CMU License (MIT-CMU)
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Multimedia :: Graphics
+Classifier: Topic :: Multimedia :: Graphics :: Capture :: Digital Camera
+Classifier: Topic :: Multimedia :: Graphics :: Capture :: Screen Capture
+Classifier: Topic :: Multimedia :: Graphics :: Graphics Conversion
+Classifier: Topic :: Multimedia :: Graphics :: Viewers
+Classifier: Typing :: Typed
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: docs
+Requires-Dist: furo; extra == "docs"
+Requires-Dist: olefile; extra == "docs"
+Requires-Dist: sphinx>=8.1; extra == "docs"
+Requires-Dist: sphinx-copybutton; extra == "docs"
+Requires-Dist: sphinx-inline-tabs; extra == "docs"
+Requires-Dist: sphinxext-opengraph; extra == "docs"
+Provides-Extra: fpx
+Requires-Dist: olefile; extra == "fpx"
+Provides-Extra: mic
+Requires-Dist: olefile; extra == "mic"
+Provides-Extra: tests
+Requires-Dist: check-manifest; extra == "tests"
+Requires-Dist: coverage>=7.4.2; extra == "tests"
+Requires-Dist: defusedxml; extra == "tests"
+Requires-Dist: markdown2; extra == "tests"
+Requires-Dist: olefile; extra == "tests"
+Requires-Dist: packaging; extra == "tests"
+Requires-Dist: pyroma; extra == "tests"
+Requires-Dist: pytest; extra == "tests"
+Requires-Dist: pytest-cov; extra == "tests"
+Requires-Dist: pytest-timeout; extra == "tests"
+Requires-Dist: trove-classifiers>=2024.10.12; extra == "tests"
+Provides-Extra: typing
+Requires-Dist: typing-extensions; python_version < "3.10" and extra == "typing"
+Provides-Extra: xmp
+Requires-Dist: defusedxml; extra == "xmp"
+
+<p align="center">
+    <img width="248" height="250" src="https://raw.githubusercontent.com/python-pillow/pillow-logo/main/pillow-logo-248x250.png" alt="Pillow logo">
+</p>
+
+# Pillow
+
+## Python Imaging Library (Fork)
+
+Pillow is the friendly PIL fork by [Jeffrey A. Clark and
+contributors](https://github.com/python-pillow/Pillow/graphs/contributors).
+PIL is the Python Imaging Library by Fredrik Lundh and contributors.
+As of 2019, Pillow development is
+[supported by Tidelift](https://tidelift.com/subscription/pkg/pypi-pillow?utm_source=pypi-pillow&utm_medium=readme&utm_campaign=enterprise).
+
+<table>
+    <tr>
+        <th>docs</th>
+        <td>
+            <a href="https://pillow.readthedocs.io/?badge=latest"><img
+                alt="Documentation Status"
+                src="https://readthedocs.org/projects/pillow/badge/?version=latest"></a>
+        </td>
+    </tr>
+    <tr>
+        <th>tests</th>
+        <td>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/lint.yml"><img
+                alt="GitHub Actions build status (Lint)"
+                src="https://github.com/python-pillow/Pillow/workflows/Lint/badge.svg"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/test.yml"><img
+                alt="GitHub Actions build status (Test Linux and macOS)"
+                src="https://github.com/python-pillow/Pillow/workflows/Test/badge.svg"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/test-windows.yml"><img
+                alt="GitHub Actions build status (Test Windows)"
+                src="https://github.com/python-pillow/Pillow/workflows/Test%20Windows/badge.svg"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/test-mingw.yml"><img
+                alt="GitHub Actions build status (Test MinGW)"
+                src="https://github.com/python-pillow/Pillow/workflows/Test%20MinGW/badge.svg"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/test-cygwin.yml"><img
+                alt="GitHub Actions build status (Test Cygwin)"
+                src="https://github.com/python-pillow/Pillow/workflows/Test%20Cygwin/badge.svg"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/test-docker.yml"><img
+                alt="GitHub Actions build status (Test Docker)"
+                src="https://github.com/python-pillow/Pillow/workflows/Test%20Docker/badge.svg"></a>
+            <a href="https://ci.appveyor.com/project/python-pillow/Pillow"><img
+                alt="AppVeyor CI build status (Windows)"
+                src="https://img.shields.io/appveyor/build/python-pillow/Pillow/main.svg?label=Windows%20build"></a>
+            <a href="https://github.com/python-pillow/Pillow/actions/workflows/wheels.yml"><img
+                alt="GitHub Actions build status (Wheels)"
+                src="https://github.com/python-pillow/Pillow/workflows/Wheels/badge.svg"></a>
+            <a href="https://app.codecov.io/gh/python-pillow/Pillow"><img
+                alt="Code coverage"
+                src="https://codecov.io/gh/python-pillow/Pillow/branch/main/graph/badge.svg"></a>
+            <a href="https://issues.oss-fuzz.com/issues?q=title:pillow"><img
+                alt="Fuzzing Status"
+                src="https://oss-fuzz-build-logs.storage.googleapis.com/badges/pillow.svg"></a>
+        </td>
+    </tr>
+    <tr>
+        <th>package</th>
+        <td>
+            <a href="https://zenodo.org/badge/latestdoi/17549/python-pillow/Pillow"><img
+                alt="Zenodo"
+                src="https://zenodo.org/badge/17549/python-pillow/Pillow.svg"></a>
+            <a href="https://tidelift.com/subscription/pkg/pypi-pillow?utm_source=pypi-pillow&utm_medium=badge"><img
+                alt="Tidelift"
+                src="https://tidelift.com/badges/package/pypi/pillow?style=flat"></a>
+            <a href="https://pypi.org/project/pillow/"><img
+                alt="Newest PyPI version"
+                src="https://img.shields.io/pypi/v/pillow.svg"></a>
+            <a href="https://pypi.org/project/pillow/"><img
+                alt="Number of PyPI downloads"
+                src="https://img.shields.io/pypi/dm/pillow.svg"></a>
+            <a href="https://www.bestpractices.dev/projects/6331"><img
+                alt="OpenSSF Best Practices"
+                src="https://www.bestpractices.dev/projects/6331/badge"></a>
+        </td>
+    </tr>
+    <tr>
+        <th>social</th>
+        <td>
+            <a href="https://gitter.im/python-pillow/Pillow?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge"><img
+                alt="Join the chat at https://gitter.im/python-pillow/Pillow"
+                src="https://badges.gitter.im/python-pillow/Pillow.svg"></a>
+            <a href="https://fosstodon.org/@pillow"><img
+                alt="Follow on https://fosstodon.org/@pillow"
+                src="https://img.shields.io/badge/publish-on%20Mastodon-595aff.svg"
+                rel="me"></a>
+        </td>
+    </tr>
+</table>
+
+## Overview
+
+The Python Imaging Library adds image processing capabilities to your Python interpreter.
+
+This library provides extensive file format support, an efficient internal representation, and fairly powerful image processing capabilities.
+
+The core image library is designed for fast access to data stored in a few basic pixel formats. It should provide a solid foundation for a general image processing tool.
+
+## More Information
+
+- [Documentation](https://pillow.readthedocs.io/)
+  - [Installation](https://pillow.readthedocs.io/en/latest/installation/basic-installation.html)
+  - [Handbook](https://pillow.readthedocs.io/en/latest/handbook/index.html)
+- [Contribute](https://github.com/python-pillow/Pillow/blob/main/.github/CONTRIBUTING.md)
+  - [Issues](https://github.com/python-pillow/Pillow/issues)
+  - [Pull requests](https://github.com/python-pillow/Pillow/pulls)
+- [Release notes](https://pillow.readthedocs.io/en/stable/releasenotes/index.html)
+- [Changelog](https://github.com/python-pillow/Pillow/releases)
+  - [Pre-fork](https://github.com/python-pillow/Pillow/blob/main/CHANGES.rst#pre-fork)
+
+## Report a Vulnerability
+
+To report a security vulnerability, please follow the procedure described in the [Tidelift security policy](https://tidelift.com/docs/security).
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/RECORD
new file mode 100644
index 00000000..c3375eb1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/RECORD
@@ -0,0 +1,229 @@
+PIL/.dylibs/libXau.6.dylib,sha256=yguEvFLSeCoS37j4E4J6TQPmRXzKbcMznx8IlMbToi4,70048
+PIL/.dylibs/libbrotlicommon.1.1.0.dylib,sha256=neNJ4ideqXAuP2nuW64mRekfacBHTWZ04LEm6-Om0wU,201200
+PIL/.dylibs/libbrotlidec.1.1.0.dylib,sha256=KWwh9jBZ--LEJrSr_n7ZJ0eXpb_4db1Y3XXjYsESHFY,104576
+PIL/.dylibs/libfreetype.6.dylib,sha256=4b5NvXmBn2_c4mxm1DEv3DqZz1Lv1muLzlCFEVqfHZM,1208416
+PIL/.dylibs/libharfbuzz.0.dylib,sha256=_b3NQ0Uo47rRMi8qMTb0TzGD6njjtH4ec9zt5AzOF9o,1351296
+PIL/.dylibs/libjpeg.62.4.0.dylib,sha256=m4isGb9aExku62NOxl5fG95O8KTWYpMedH0gOrn_Uho,637072
+PIL/.dylibs/liblcms2.2.dylib,sha256=nFXEMxOj44QqyD9eKAWbJtCkUFyLVQE8QHgOuFt7pSw,557344
+PIL/.dylibs/liblzma.5.dylib,sha256=S1-lCOwmPmZgKAT5wCbs-I2yryP4wMKDbtEqXWLnS1w,341168
+PIL/.dylibs/libopenjp2.2.5.3.dylib,sha256=e9LwA1LlL4j3A9CrXyjsTOuA3RTBTQVH0U1c8bjTYmY,716096
+PIL/.dylibs/libpng16.16.dylib,sha256=XVnfb9thR2triFjBttWt7orzDEZphUIt2V9gaNrAPvc,361456
+PIL/.dylibs/libsharpyuv.0.dylib,sha256=v8SQPMlrHdiJZsKE4f1U8d1BMmVnS2D9REAMyQQvgM8,86256
+PIL/.dylibs/libtiff.6.dylib,sha256=e_IqtXCsWB91Cc3y633vvYxyZ0emt5JimLAvAmZNri8,802784
+PIL/.dylibs/libwebp.7.dylib,sha256=ammPv5a7pAw12xGWHT_MlTshIR0cv1PA6vfXnYF9mvU,498848
+PIL/.dylibs/libwebpdemux.2.dylib,sha256=mGgB9C3DcJtufwSF3tF8lXWuAddVRz25pTGYTdXyvIs,69904
+PIL/.dylibs/libwebpmux.3.dylib,sha256=dYIZkSmVeZJxa4FfwTN8bObCWn-eFnc_GVOLxnMckzU,106400
+PIL/.dylibs/libxcb.1.1.0.dylib,sha256=lKmyXERKkIBMzVryCiVENzAMer4-IrWKTE5yZfZtMu4,277696
+PIL/.dylibs/libz.1.3.1.zlib-ng.dylib,sha256=DR27XaFKcRnOt7gorBTVRSpSfYTn2d0FIDQ_ei1ANjY,177248
+PIL/BdfFontFile.py,sha256=Hnlsd7gTtt7gaY1ilXHP-co5ehkBQ8xml942d06MEno,3477
+PIL/BlpImagePlugin.py,sha256=7Ye8a8ojr4sW0a9bp6oYeWXsD1zkCIRHo6UqENlcOpU,16683
+PIL/BmpImagePlugin.py,sha256=3li_kaQt9Q6-xvIFQ6tP_g_UkpbOadqdKDNlqBzT9Fg,19758
+PIL/BufrStubImagePlugin.py,sha256=PcwNafLqxrQ_RTsz7hAqfpLWcQHT-kXKYPLaB46ZtUA,1753
+PIL/ContainerIO.py,sha256=wkBqL2GDAb5fh3wrtfTGUfqioJipCl-lg2GxbjQrTZw,4604
+PIL/CurImagePlugin.py,sha256=xTou-ULQPpI5lVdqpQ2-4Pdjnh7j1GddH-K22DQbdk0,1792
+PIL/DcxImagePlugin.py,sha256=K70qkwz-e3nkgjftD5pIElMcU2klHWFTKbINlwgeQ24,2034
+PIL/DdsImagePlugin.py,sha256=vv1dNj9T67YmoiBcKyLeDcMDv42WWdOXSDXg7THhA7s,16938
+PIL/EpsImagePlugin.py,sha256=a8ErgTj7iJcPkzT1Iu7Dclkm-ms9WuQ3h0Gca505mh0,16365
+PIL/ExifTags.py,sha256=zW6kVikCosiyoCo7J7R62evD3hoxjKPchnVh8po7CZc,9931
+PIL/FitsImagePlugin.py,sha256=QVW0dCJpPCrV2N6qSHVS9nXpjYPHnbHWxSC88LcYDN0,4639
+PIL/FliImagePlugin.py,sha256=pf-F5OmW9ehUQGUcZDcjlQMdJBxdjEC0_1VprHActeM,4675
+PIL/FontFile.py,sha256=St7MxO5Q-oakCLWn3ZrgrtaT3wSsmAarxm8AU-G8Moc,3577
+PIL/FpxImagePlugin.py,sha256=OAEpvvnalEIU2KeyXX7S-LEay_qyQNKu2dN76_xUoXk,7288
+PIL/FtexImagePlugin.py,sha256=Eh5ZFg-js6TjVqTe5EUehRzsTKMPpr2Hs66kXs8pgw4,3527
+PIL/GbrImagePlugin.py,sha256=5t0UfLubTPQcuDDbafwC78OLR7IsD5hjpvhUZ5g8z4A,3006
+PIL/GdImageFile.py,sha256=sBT3em6cOKea3atNYE8u4L0ugaFlH6zcmt_7JHOOCYw,2802
+PIL/GifImagePlugin.py,sha256=mRuxEiCPGp_CGet7DgXbWagA_JtPV-gdEpWxN_wLZCs,41456
+PIL/GimpGradientFile.py,sha256=ABNhtD50Gv82Dn1VxbdgfSIz3Q2_nPiv_zDquOYyVAw,3898
+PIL/GimpPaletteFile.py,sha256=mK8RqdS7Ae9W7gZ7NB7MkyzuOqhC6Z09_OsLkxCJoak,1427
+PIL/GribStubImagePlugin.py,sha256=hzsipSut7wvQ14P5fx7mkGtLj2TWKZk7OwSiFstB194,1747
+PIL/Hdf5StubImagePlugin.py,sha256=6bSeB8RJaWqdU3-xwnJIWtGBZjpM0QnpcM8974UWN90,1750
+PIL/IcnsImagePlugin.py,sha256=mRXkUrNlnKB8vyW9UphHJ4JefgIrpzIqaUamf76uPec,12953
+PIL/IcoImagePlugin.py,sha256=-6GillVbyW9nWlC83IKV7F-99O2aF01U7B1EuxGZpgY,12468
+PIL/ImImagePlugin.py,sha256=HIbI1-XpN2RxNyrmMjFB9RKyb9GchBGaFlokk1DPYok,11438
+PIL/Image.py,sha256=S3kgQ2R7nEP9b5ljsr3jbV-ybq-8TwZdSpSJ9LU2X0Y,146129
+PIL/ImageChops.py,sha256=GEjlymcoDtA5OOeIxQVIX96BD-s6AXhb7TmSLYn2tUg,7946
+PIL/ImageCms.py,sha256=wpVg1Kmp5WfeCNbEfGUCZsjcWVreg3HZqMHyTttlz1s,42010
+PIL/ImageColor.py,sha256=IGA9C2umeED_EzS2Cvj6KsU0VutC9RstWIYPe8uDsVk,9441
+PIL/ImageDraw.py,sha256=7TZ0miXREA8vFh0yCCAy3k0olUfYz8erDjM4-AH586o,42275
+PIL/ImageDraw2.py,sha256=pdVMW7bVw3KwhXvRZh28Md4y-2xFfuo5fHcDnaYqVK4,7227
+PIL/ImageEnhance.py,sha256=4Elhz_lyyxLmx0GkSHrwOAmNJ2TkqVQPHejzGihZUMI,3627
+PIL/ImageFile.py,sha256=3Rkbo6XqFlg2QgbpVSf5jS8K0xps8-3j1w0LIAAZwKw,26125
+PIL/ImageFilter.py,sha256=X3a-7xf1loq3j_MFDha0nAglsHImezBV8D4C0mMntEE,18710
+PIL/ImageFont.py,sha256=LLt1wvLhLfAND_XruWyykRP3InLsyyCpXdwH4a9XQtA,64261
+PIL/ImageGrab.py,sha256=-9RS8qP6nark5RW31IrRXlrPNrTno-gRJb0zOujyi48,6002
+PIL/ImageMath.py,sha256=qDVyqP24n4FnCgJRgW_DVcRFoTdZFJLQd5qxAZS4EG4,11942
+PIL/ImageMode.py,sha256=5yOxODAZ7jG03DsUFrt7eQayTtIpWPgvfyhlXDWwcv8,2681
+PIL/ImageMorph.py,sha256=TowXnk1Q2wX9AXVBDWRRQhCfAbFOUWGMo00vq4yn-fU,8563
+PIL/ImageOps.py,sha256=g68uCpv-BJIvZr2vHUEBZG98WRJDhEOj3jEYKurOdco,25091
+PIL/ImagePalette.py,sha256=wTokkN4dylYRAzxirCmh6M_GyyqbkDazwfX2tEoohCs,9002
+PIL/ImagePath.py,sha256=5yUG5XCUil1KKTTA_8PgGhcmg-mnue-GK0FwTBlhjw4,371
+PIL/ImageQt.py,sha256=maxKjoYomf3nqO9dWhg9kXGqg0eRz5erAa1yT2_d9oo,6834
+PIL/ImageSequence.py,sha256=gx2EvywPBEjxNJujCqdpbfAm2BpyNV2_f1IaO3niubw,2200
+PIL/ImageShow.py,sha256=LBGhPR3k5Z20S7vDyCsL0BftIX5tvTvCd5xdCvA7lTc,9993
+PIL/ImageStat.py,sha256=S43FZ89r_u4hKCj59lVuWpyVJfhbUy3igXkp9DwaMgM,5325
+PIL/ImageTk.py,sha256=JuzOgUMKiAhR8JAYCSY1Il3iwQ8Hx-vwC4ng_KRKfCQ,8997
+PIL/ImageTransform.py,sha256=okpZipXf2u7tDB3dticLozrOKI8QNIsniCq_J4CxQC0,3886
+PIL/ImageWin.py,sha256=LT05w8_vTfRrC3n9S9pM0TNbXrzZLEJHlCJil7Xv80k,8085
+PIL/ImtImagePlugin.py,sha256=JZnVN1bWNiptIrDmiUQYMJQL_aduaGrSCEjRIHFuoEA,2665
+PIL/IptcImagePlugin.py,sha256=zMOEYveSc8ph1WdJtC-tUJEueDcInpVUviCcnqKXq0Q,6669
+PIL/Jpeg2KImagePlugin.py,sha256=Zv0PhFk5dx0c1H85__4PWyDC-xHjjnb9h3iIe32KtRU,13885
+PIL/JpegImagePlugin.py,sha256=tOi0VgbxWSQqChM0ZBSya05bH4FJHz8F1kT_6vRHUoY,31800
+PIL/JpegPresets.py,sha256=lnqWHo4DLIHIulcdHp0NJ7CWexHt8T3w51kIKlLfkIA,12379
+PIL/McIdasImagePlugin.py,sha256=G_sNQybucqISdmU-XGCtdIcm4jZTI59tcSz96WiUwDI,1938
+PIL/MicImagePlugin.py,sha256=8EqZ-Vm9FK23XB-5thR3dWEcq4j3XbMG961-ILRnC0g,2680
+PIL/MpegImagePlugin.py,sha256=AplKMsPC9Y9ExJfqTE_xlm8TW-CSx578NGQWyVcmIiQ,2100
+PIL/MpoImagePlugin.py,sha256=QVkZnrOGAlPCARpraeNSS6Q-ymQXfQGKAUAfRWTDZMA,6220
+PIL/MspImagePlugin.py,sha256=VCFo9otTuMvH_xnWu43Cm1gn5Lv0-C5utBOKkxo1JB8,5882
+PIL/PSDraw.py,sha256=3hY8wDQamJr5X5dS8pwQ9eUMJAV835of7aX3t8kM5Q8,6909
+PIL/PaletteFile.py,sha256=rC4YrlwwpJtl7RdPDnfl21HR4Vge3cAUw5Z6zicBqIk,1211
+PIL/PalmImagePlugin.py,sha256=Lz2yNR9099-cjTs4NY-0XvHxxCDBSYJkqXJltcZkNXQ,9351
+PIL/PcdImagePlugin.py,sha256=8LEZLY4sjRfsL5mQQF_V9YkoiQlrXoD1dRpzM1CtcWg,1592
+PIL/PcfFontFile.py,sha256=NPZQ0XkbGB8uTlGqgmIPGkwuLMYBdykDeVuvFgIC7JU,7147
+PIL/PcxImagePlugin.py,sha256=egtz8QOKgBGIXxLdUxd8CfD3vBVxRcWjkt6P9HtUpSY,6247
+PIL/PdfImagePlugin.py,sha256=AbJA2f4qzH8G1olfmk18SzQlcx3WsipUYDc5bcR8Wvk,9349
+PIL/PdfParser.py,sha256=0p4yxf90wHEx1jDRnjpKxjwfDqUYO463kaYS31PJpYY,37980
+PIL/PixarImagePlugin.py,sha256=MZ3iR3IwjVKRIRs2CSl1xExkPizhYoAFNBGeufY0fLc,1753
+PIL/PngImagePlugin.py,sha256=OWYk88lLD3eBYiXQf6FBLBjLgCwggK5zukTBYhGHL-w,50861
+PIL/PpmImagePlugin.py,sha256=m2PDVO97GAn5kISON3-PJENWU3WZOiwRbPjiUp_rK0M,12354
+PIL/PsdImagePlugin.py,sha256=5g-l_HrIWMG7xSAb4ofhgKVhsnK2yqh6ee9xE-Z1620,8621
+PIL/QoiImagePlugin.py,sha256=xZVSfZVW9nVxjQjNDcOB6q7F0H1Iq--vOr_irCp65w0,4183
+PIL/SgiImagePlugin.py,sha256=wjO3mgTO7AYC2Bs6RJBEKafm49wgFkCXZuVoBD6UWxc,6732
+PIL/SpiderImagePlugin.py,sha256=a9S-wC7lckgN5x412LhPuCvpwg5kz9VaISDbAewpcV4,10133
+PIL/SunImagePlugin.py,sha256=Hdxkhk0pxpBGxYhPJfCDLwsYcO1KjxjtplNMFYibIvk,4589
+PIL/TarIO.py,sha256=uQ5Zh47x67H9fq8iGLSeCfk22i0E7Ae06fVC2bf1LcU,1376
+PIL/TgaImagePlugin.py,sha256=2vDsFTcBUBHw1V80wpVv4tgpLDbPr6yVHi6Fvaqf0HY,6980
+PIL/TiffImagePlugin.py,sha256=OuJZgFdoUnGvZ-NYZbhLuQ7VBz9UqeQSPsr2OGoSSqo,83398
+PIL/TiffTags.py,sha256=-gbXLZ5rlHD6crwtY6TkafDm2tamlc5v8e7FjS8PcIg,17082
+PIL/WalImageFile.py,sha256=Lfuq_WZ_V_onwucfUc6GWfvY7z_K4s-5EdaQGu_2DD4,5704
+PIL/WebPImagePlugin.py,sha256=cewKnQP9W43Im-mG6htOV26w6mS1KuNHjAyuuGETGjQ,10061
+PIL/WmfImagePlugin.py,sha256=XbuT349Oa-9JNil6wB-gMMEQkyyMTvdQ1POxDzt3UPk,5137
+PIL/XVThumbImagePlugin.py,sha256=BoEiffKzyGEYnjZAKCICSCF7En7SaYHt77mqTW6R-qs,2110
+PIL/XbmImagePlugin.py,sha256=dzrVoHrTR41zGlNCuy2U9za2WbWuR1G4S_JPA33wtJk,2664
+PIL/XpmImagePlugin.py,sha256=ioqboXOmeB3LBqe6ACECiKYbISM4VZmv8DPy148kR3U,3226
+PIL/__init__.py,sha256=fJUwPGhI8_mcB8jNWD-hUw7MiMJyWgqVX_nFtzIj1Zs,2008
+PIL/__main__.py,sha256=Lpj4vef8mI7jA1sRCUAoVYaeePD_Uc898xF5c7XLx1A,133
+PIL/__pycache__/BdfFontFile.cpython-312.pyc,,
+PIL/__pycache__/BlpImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/BmpImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/BufrStubImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/ContainerIO.cpython-312.pyc,,
+PIL/__pycache__/CurImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/DcxImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/DdsImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/EpsImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/ExifTags.cpython-312.pyc,,
+PIL/__pycache__/FitsImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/FliImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/FontFile.cpython-312.pyc,,
+PIL/__pycache__/FpxImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/FtexImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/GbrImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/GdImageFile.cpython-312.pyc,,
+PIL/__pycache__/GifImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/GimpGradientFile.cpython-312.pyc,,
+PIL/__pycache__/GimpPaletteFile.cpython-312.pyc,,
+PIL/__pycache__/GribStubImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/Hdf5StubImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/IcnsImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/IcoImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/ImImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/Image.cpython-312.pyc,,
+PIL/__pycache__/ImageChops.cpython-312.pyc,,
+PIL/__pycache__/ImageCms.cpython-312.pyc,,
+PIL/__pycache__/ImageColor.cpython-312.pyc,,
+PIL/__pycache__/ImageDraw.cpython-312.pyc,,
+PIL/__pycache__/ImageDraw2.cpython-312.pyc,,
+PIL/__pycache__/ImageEnhance.cpython-312.pyc,,
+PIL/__pycache__/ImageFile.cpython-312.pyc,,
+PIL/__pycache__/ImageFilter.cpython-312.pyc,,
+PIL/__pycache__/ImageFont.cpython-312.pyc,,
+PIL/__pycache__/ImageGrab.cpython-312.pyc,,
+PIL/__pycache__/ImageMath.cpython-312.pyc,,
+PIL/__pycache__/ImageMode.cpython-312.pyc,,
+PIL/__pycache__/ImageMorph.cpython-312.pyc,,
+PIL/__pycache__/ImageOps.cpython-312.pyc,,
+PIL/__pycache__/ImagePalette.cpython-312.pyc,,
+PIL/__pycache__/ImagePath.cpython-312.pyc,,
+PIL/__pycache__/ImageQt.cpython-312.pyc,,
+PIL/__pycache__/ImageSequence.cpython-312.pyc,,
+PIL/__pycache__/ImageShow.cpython-312.pyc,,
+PIL/__pycache__/ImageStat.cpython-312.pyc,,
+PIL/__pycache__/ImageTk.cpython-312.pyc,,
+PIL/__pycache__/ImageTransform.cpython-312.pyc,,
+PIL/__pycache__/ImageWin.cpython-312.pyc,,
+PIL/__pycache__/ImtImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/IptcImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/Jpeg2KImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/JpegImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/JpegPresets.cpython-312.pyc,,
+PIL/__pycache__/McIdasImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/MicImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/MpegImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/MpoImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/MspImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PSDraw.cpython-312.pyc,,
+PIL/__pycache__/PaletteFile.cpython-312.pyc,,
+PIL/__pycache__/PalmImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PcdImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PcfFontFile.cpython-312.pyc,,
+PIL/__pycache__/PcxImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PdfImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PdfParser.cpython-312.pyc,,
+PIL/__pycache__/PixarImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PngImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PpmImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/PsdImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/QoiImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/SgiImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/SpiderImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/SunImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/TarIO.cpython-312.pyc,,
+PIL/__pycache__/TgaImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/TiffImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/TiffTags.cpython-312.pyc,,
+PIL/__pycache__/WalImageFile.cpython-312.pyc,,
+PIL/__pycache__/WebPImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/WmfImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/XVThumbImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/XbmImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/XpmImagePlugin.cpython-312.pyc,,
+PIL/__pycache__/__init__.cpython-312.pyc,,
+PIL/__pycache__/__main__.cpython-312.pyc,,
+PIL/__pycache__/_binary.cpython-312.pyc,,
+PIL/__pycache__/_deprecate.cpython-312.pyc,,
+PIL/__pycache__/_tkinter_finder.cpython-312.pyc,,
+PIL/__pycache__/_typing.cpython-312.pyc,,
+PIL/__pycache__/_util.cpython-312.pyc,,
+PIL/__pycache__/_version.cpython-312.pyc,,
+PIL/__pycache__/features.cpython-312.pyc,,
+PIL/__pycache__/report.cpython-312.pyc,,
+PIL/_binary.py,sha256=pcM6AL04GxgmGeLfcH1V1BZHENwIrQH0uxhJ7r0HIL0,2550
+PIL/_deprecate.py,sha256=SLU2p8O9ImHYHsD4VFGKLTkewh_Eda0axfIWUCnkKSg,1936
+PIL/_imaging.cpython-312-darwin.so,sha256=xT9Jbc8FXA4aZldQlWz4cABXIEcDOh6gth5cjm9Ui2E,553728
+PIL/_imaging.pyi,sha256=StMbXUZS32AegATP1sUHfs5P05A3TD_BiQKsDHQBW40,868
+PIL/_imagingcms.cpython-312-darwin.so,sha256=cIMWtd85iK8hf7FttAGDnkHr0wPnbAdAzMACZritYhE,80800
+PIL/_imagingcms.pyi,sha256=brpjxRoiY_2ItyfTrjhKeGEsExe4GPG-25q9AQP8Jp8,4389
+PIL/_imagingft.cpython-312-darwin.so,sha256=hHDs88NRLwh1amJ--gUrgJOvAbOgijEcnOob-QkUyBs,116464
+PIL/_imagingft.pyi,sha256=62nD4AzNDHddKXgcxblOrxKwu8w7TJIHNM-mqC2fue0,1789
+PIL/_imagingmath.cpython-312-darwin.so,sha256=D8WHOsXQJ6NQnHeAkpio34nqL0JgVcruzBf8U2PE0iA,55760
+PIL/_imagingmath.pyi,sha256=3fBxcSppJr6EOEcUojvflG3Eegg7lv2Qp0dNQQILrP4,63
+PIL/_imagingmorph.cpython-312-darwin.so,sha256=x7XYoi3GJJuvzLYUvVbhjCb4jE0pBMrcNGsGo9cwB6g,51424
+PIL/_imagingmorph.pyi,sha256=3fBxcSppJr6EOEcUojvflG3Eegg7lv2Qp0dNQQILrP4,63
+PIL/_imagingtk.cpython-312-darwin.so,sha256=nt5JYhIO5zBM1FsgQFEA-ldHgp7VqXyV4FLHVQCz9yQ,52720
+PIL/_imagingtk.pyi,sha256=3fBxcSppJr6EOEcUojvflG3Eegg7lv2Qp0dNQQILrP4,63
+PIL/_tkinter_finder.py,sha256=CECvYrzWNc7BuzzR_mWZZKjPdADg6iRG8ilJToyjD3w,540
+PIL/_typing.py,sha256=5nIQp0No7js6KhThjafVIrrddGbd_6EGvrj4BzElP-Q,1244
+PIL/_util.py,sha256=E76J1WLAe6Xg5yNWYztQwYzxUT_sR_VQxFJu7IZ3S3k,635
+PIL/_version.py,sha256=huIl0a2frkcR-Wem-VhDvWm-Vl4arrQvHgkDSBYVuTs,87
+PIL/_webp.cpython-312-darwin.so,sha256=ce6uCYbVwKJ9bPawML0ltFPQ7ef_ByR_jLMPtslG9m0,75072
+PIL/_webp.pyi,sha256=3fBxcSppJr6EOEcUojvflG3Eegg7lv2Qp0dNQQILrP4,63
+PIL/features.py,sha256=rwlbHZYj_OM2-qhscJczwnr3UnA1QlEJg6CBmGcKETA,11262
+PIL/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+PIL/report.py,sha256=4JY6-IU7sH1RKuRbOvy1fUt0dAoi79FX4tYJN3p1DT0,100
+pillow-11.1.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pillow-11.1.0.dist-info/LICENSE,sha256=9XbTk30hFh6sWbC294B6L18w_nfXKr_g1bTY6TmzEog,60071
+pillow-11.1.0.dist-info/METADATA,sha256=UIIl--QvedtNB7RSKhfc5b6W1htrqa7JadqYx4sFRrA,9137
+pillow-11.1.0.dist-info/RECORD,,
+pillow-11.1.0.dist-info/WHEEL,sha256=7Wd-yga4fjSiXpUH443rsPZpiZ4h8-uNrXJrYRW_e14,109
+pillow-11.1.0.dist-info/top_level.txt,sha256=riZqrk-hyZqh5f1Z0Zwii3dKfxEsByhu9cU9IODF-NY,4
+pillow-11.1.0.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/WHEEL
new file mode 100644
index 00000000..ad235f41
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: false
+Tag: cp312-cp312-macosx_11_0_arm64
+
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/top_level.txt
new file mode 100644
index 00000000..b338169c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+PIL
diff --git a/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/zip-safe b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/zip-safe
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pillow-11.1.0.dist-info/zip-safe
@@ -0,0 +1 @@
+
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/AUTHORS.txt b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/AUTHORS.txt
new file mode 100644
index 00000000..f42daec0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/AUTHORS.txt
@@ -0,0 +1,806 @@
+@Switch01
+A_Rog
+Aakanksha Agrawal
+Abhinav Sagar
+ABHYUDAY PRATAP SINGH
+abs51295
+AceGentile
+Adam Chainz
+Adam Tse
+Adam Wentz
+admin
+Adolfo Ochagavía
+Adrien Morison
+Agus
+ahayrapetyan
+Ahilya
+AinsworthK
+Akash Srivastava
+Alan Yee
+Albert Tugushev
+Albert-Guan
+albertg
+Alberto Sottile
+Aleks Bunin
+Ales Erjavec
+Alethea Flowers
+Alex Gaynor
+Alex Grönholm
+Alex Hedges
+Alex Loosley
+Alex Morega
+Alex Stachowiak
+Alexander Shtyrov
+Alexandre Conrad
+Alexey Popravka
+Aleš Erjavec
+Alli
+Ami Fischman
+Ananya Maiti
+Anatoly Techtonik
+Anders Kaseorg
+Andre Aguiar
+Andreas Lutro
+Andrei Geacar
+Andrew Gaul
+Andrew Shymanel
+Andrey Bienkowski
+Andrey Bulgakov
+Andrés Delfino
+Andy Freeland
+Andy Kluger
+Ani Hayrapetyan
+Aniruddha Basak
+Anish Tambe
+Anrs Hu
+Anthony Sottile
+Antoine Musso
+Anton Ovchinnikov
+Anton Patrushev
+Anton Zelenov
+Antonio Alvarado Hernandez
+Antony Lee
+Antti Kaihola
+Anubhav Patel
+Anudit Nagar
+Anuj Godase
+AQNOUCH Mohammed
+AraHaan
+arena
+arenasys
+Arindam Choudhury
+Armin Ronacher
+Arnon Yaari
+Artem
+Arun Babu Neelicattu
+Ashley Manton
+Ashwin Ramaswami
+atse
+Atsushi Odagiri
+Avinash Karhana
+Avner Cohen
+Awit (Ah-Wit) Ghirmai
+Baptiste Mispelon
+Barney Gale
+barneygale
+Bartek Ogryczak
+Bastian Venthur
+Ben Bodenmiller
+Ben Darnell
+Ben Hoyt
+Ben Mares
+Ben Rosser
+Bence Nagy
+Benjamin Peterson
+Benjamin VanEvery
+Benoit Pierre
+Berker Peksag
+Bernard
+Bernard Tyers
+Bernardo B. Marques
+Bernhard M. Wiedemann
+Bertil Hatt
+Bhavam Vidyarthi
+Blazej Michalik
+Bogdan Opanchuk
+BorisZZZ
+Brad Erickson
+Bradley Ayers
+Branch Vincent
+Brandon L. Reiss
+Brandt Bucher
+Brannon Dorsey
+Brett Randall
+Brett Rosen
+Brian Cristante
+Brian Rosner
+briantracy
+BrownTruck
+Bruno Oliveira
+Bruno Renié
+Bruno S
+Bstrdsmkr
+Buck Golemon
+burrows
+Bussonnier Matthias
+bwoodsend
+c22
+Caleb Brown
+Caleb Martinez
+Calvin Smith
+Carl Meyer
+Carlos Liam
+Carol Willing
+Carter Thayer
+Cass
+Chandrasekhar Atina
+Charlie Marsh
+charwick
+Chih-Hsuan Yen
+Chris Brinker
+Chris Hunt
+Chris Jerdonek
+Chris Kuehl
+Chris Markiewicz
+Chris McDonough
+Chris Pawley
+Chris Pryer
+Chris Wolfe
+Christian Clauss
+Christian Heimes
+Christian Oudard
+Christoph Reiter
+Christopher Hunt
+Christopher Snyder
+chrysle
+cjc7373
+Clark Boylan
+Claudio Jolowicz
+Clay McClure
+Cody
+Cody Soyland
+Colin Watson
+Collin Anderson
+Connor Osborn
+Cooper Lees
+Cooper Ry Lees
+Cory Benfield
+Cory Wright
+Craig Kerstiens
+Cristian Sorinel
+Cristina
+Cristina Muñoz
+ctg123
+Curtis Doty
+cytolentino
+Daan De Meyer
+Dale
+Damian
+Damian Quiroga
+Damian Shaw
+Dan Black
+Dan Savilonis
+Dan Sully
+Dane Hillard
+daniel
+Daniel Collins
+Daniel Hahler
+Daniel Holth
+Daniel Jost
+Daniel Katz
+Daniel Shaulov
+Daniele Esposti
+Daniele Nicolodi
+Daniele Procida
+Daniil Konovalenko
+Danny Hermes
+Danny McClanahan
+Darren Kavanagh
+Dav Clark
+Dave Abrahams
+Dave Jones
+David Aguilar
+David Black
+David Bordeynik
+David Caro
+David D Lowe
+David Evans
+David Hewitt
+David Linke
+David Poggi
+David Poznik
+David Pursehouse
+David Runge
+David Tucker
+David Wales
+Davidovich
+ddelange
+Deepak Sharma
+Deepyaman Datta
+Denise Yu
+dependabot[bot]
+derwolfe
+Desetude
+Devesh Kumar Singh
+devsagul
+Diego Caraballo
+Diego Ramirez
+DiegoCaraballo
+Dimitri Merejkowsky
+Dimitri Papadopoulos
+Dimitri Papadopoulos Orfanos
+Dirk Stolle
+Dmitry Gladkov
+Dmitry Volodin
+Domen Kožar
+Dominic Davis-Foster
+Donald Stufft
+Dongweiming
+doron zarhi
+Dos Moonen
+Douglas Thor
+DrFeathers
+Dustin Ingram
+Dustin Rodrigues
+Dwayne Bailey
+Ed Morley
+Edgar Ramírez
+Edgar Ramírez Mondragón
+Ee Durbin
+Efflam Lemaillet
+efflamlemaillet
+Eitan Adler
+ekristina
+elainechan
+Eli Schwartz
+Elisha Hollander
+Ellen Marie Dash
+Emil Burzo
+Emil Styrke
+Emmanuel Arias
+Endoh Takanao
+enoch
+Erdinc Mutlu
+Eric Cousineau
+Eric Gillingham
+Eric Hanchrow
+Eric Hopper
+Erik M. Bray
+Erik Rose
+Erwin Janssen
+Eugene Vereshchagin
+everdimension
+Federico
+Felipe Peter
+Felix Yan
+fiber-space
+Filip Kokosiński
+Filipe Laíns
+Finn Womack
+finnagin
+Flavio Amurrio
+Florian Briand
+Florian Rathgeber
+Francesco
+Francesco Montesano
+Fredrik Orderud
+Frost Ming
+Gabriel Curio
+Gabriel de Perthuis
+Garry Polley
+gavin
+gdanielson
+Geoffrey Sneddon
+George Song
+Georgi Valkov
+Georgy Pchelkin
+ghost
+Giftlin Rajaiah
+gizmoguy1
+gkdoc
+Godefroid Chapelle
+Gopinath M
+GOTO Hayato
+gousaiyang
+gpiks
+Greg Roodt
+Greg Ward
+Guilherme Espada
+Guillaume Seguin
+gutsytechster
+Guy Rozendorn
+Guy Tuval
+gzpan123
+Hanjun Kim
+Hari Charan
+Harsh Vardhan
+harupy
+Harutaka Kawamura
+hauntsaninja
+Henrich Hartzer
+Henry Schreiner
+Herbert Pfennig
+Holly Stotelmyer
+Honnix
+Hsiaoming Yang
+Hugo Lopes Tavares
+Hugo van Kemenade
+Hugues Bruant
+Hynek Schlawack
+Ian Bicking
+Ian Cordasco
+Ian Lee
+Ian Stapleton Cordasco
+Ian Wienand
+Igor Kuzmitshov
+Igor Sobreira
+Ikko Ashimine
+Ilan Schnell
+Illia Volochii
+Ilya Baryshev
+Inada Naoki
+Ionel Cristian Mărieș
+Ionel Maries Cristian
+Itamar Turner-Trauring
+Ivan Pozdeev
+J. Nick Koston
+Jacob Kim
+Jacob Walls
+Jaime Sanz
+jakirkham
+Jakub Kuczys
+Jakub Stasiak
+Jakub Vysoky
+Jakub Wilk
+James Cleveland
+James Curtin
+James Firth
+James Gerity
+James Polley
+Jan Pokorný
+Jannis Leidel
+Jarek Potiuk
+jarondl
+Jason Curtis
+Jason R. Coombs
+JasonMo
+JasonMo1
+Jay Graves
+Jean Abou Samra
+Jean-Christophe Fillion-Robin
+Jeff Barber
+Jeff Dairiki
+Jeff Widman
+Jelmer Vernooij
+jenix21
+Jeremy Fleischman
+Jeremy Stanley
+Jeremy Zafran
+Jesse Rittner
+Jiashuo Li
+Jim Fisher
+Jim Garrison
+Jinzhe Zeng
+Jiun Bae
+Jivan Amara
+Joe Bylund
+Joe Michelini
+John Paton
+John Sirois
+John T. Wodder II
+John-Scott Atlakson
+johnthagen
+Jon Banafato
+Jon Dufresne
+Jon Parise
+Jonas Nockert
+Jonathan Herbert
+Joonatan Partanen
+Joost Molenaar
+Jorge Niedbalski
+Joseph Bylund
+Joseph Long
+Josh Bronson
+Josh Cannon
+Josh Hansen
+Josh Schneier
+Joshua
+JoshuaPerdue
+Juan Luis Cano Rodríguez
+Juanjo Bazán
+Judah Rand
+Julian Berman
+Julian Gethmann
+Julien Demoor
+July Tikhonov
+Jussi Kukkonen
+Justin van Heek
+jwg4
+Jyrki Pulliainen
+Kai Chen
+Kai Mueller
+Kamal Bin Mustafa
+Karolina Surma
+kasium
+kaustav haldar
+keanemind
+Keith Maxwell
+Kelsey Hightower
+Kenneth Belitzky
+Kenneth Reitz
+Kevin Burke
+Kevin Carter
+Kevin Frommelt
+Kevin R Patterson
+Kexuan Sun
+Kit Randel
+Klaas van Schelven
+KOLANICH
+konstin
+kpinc
+Krishna Oza
+Kumar McMillan
+Kuntal Majumder
+Kurt McKee
+Kyle Persohn
+lakshmanaram
+Laszlo Kiss-Kollar
+Laurent Bristiel
+Laurent LAPORTE
+Laurie O
+Laurie Opperman
+layday
+Leon Sasson
+Lev Givon
+Lincoln de Sousa
+Lipis
+lorddavidiii
+Loren Carvalho
+Lucas Cimon
+Ludovic Gasc
+Luis Medel
+Lukas Geiger
+Lukas Juhrich
+Luke Macken
+Luo Jiebin
+luojiebin
+luz.paz
+László Kiss Kollár
+M00nL1ght
+Marc Abramowitz
+Marc Tamlyn
+Marcus Smith
+Mariatta
+Mark Kohler
+Mark McLoughlin
+Mark Williams
+Markus Hametner
+Martey Dodoo
+Martin Fischer
+Martin Häcker
+Martin Pavlasek
+Masaki
+Masklinn
+Matej Stuchlik
+Mathew Jennings
+Mathieu Bridon
+Mathieu Kniewallner
+Matt Bacchi
+Matt Good
+Matt Maker
+Matt Robenolt
+Matt Wozniski
+matthew
+Matthew Einhorn
+Matthew Feickert
+Matthew Gilliard
+Matthew Hughes
+Matthew Iversen
+Matthew Treinish
+Matthew Trumbell
+Matthew Willson
+Matthias Bussonnier
+mattip
+Maurits van Rees
+Max W Chase
+Maxim Kurnikov
+Maxime Rouyrre
+mayeut
+mbaluna
+mdebi
+memoselyk
+meowmeowcat
+Michael
+Michael Aquilina
+Michael E. Karpeles
+Michael Klich
+Michael Mintz
+Michael Williamson
+michaelpacer
+Michał Górny
+Mickaël Schoentgen
+Miguel Araujo Perez
+Mihir Singh
+Mike
+Mike Hendricks
+Min RK
+MinRK
+Miro Hrončok
+Monica Baluna
+montefra
+Monty Taylor
+morotti
+mrKazzila
+Muha Ajjan
+Nadav Wexler
+Nahuel Ambrosini
+Nate Coraor
+Nate Prewitt
+Nathan Houghton
+Nathaniel J. Smith
+Nehal J Wani
+Neil Botelho
+Nguyễn Gia Phong
+Nicholas Serra
+Nick Coghlan
+Nick Stenning
+Nick Timkovich
+Nicolas Bock
+Nicole Harris
+Nikhil Benesch
+Nikhil Ladha
+Nikita Chepanov
+Nikolay Korolev
+Nipunn Koorapati
+Nitesh Sharma
+Niyas Sait
+Noah
+Noah Gorny
+Nowell Strite
+NtaleGrey
+nvdv
+OBITORASU
+Ofek Lev
+ofrinevo
+Oliver Freund
+Oliver Jeeves
+Oliver Mannion
+Oliver Tonnhofer
+Olivier Girardot
+Olivier Grisel
+Ollie Rutherfurd
+OMOTO Kenji
+Omry Yadan
+onlinejudge95
+Oren Held
+Oscar Benjamin
+Oz N Tiram
+Pachwenko
+Patrick Dubroy
+Patrick Jenkins
+Patrick Lawson
+patricktokeeffe
+Patrik Kopkan
+Paul Ganssle
+Paul Kehrer
+Paul Moore
+Paul Nasrat
+Paul Oswald
+Paul van der Linden
+Paulus Schoutsen
+Pavel Safronov
+Pavithra Eswaramoorthy
+Pawel Jasinski
+Paweł Szramowski
+Pekka Klärck
+Peter Gessler
+Peter Lisák
+Peter Shen
+Peter Waller
+Petr Viktorin
+petr-tik
+Phaneendra Chiruvella
+Phil Elson
+Phil Freo
+Phil Pennock
+Phil Whelan
+Philip Jägenstedt
+Philip Molloy
+Philippe Ombredanne
+Pi Delport
+Pierre-Yves Rofes
+Pieter Degroote
+pip
+Prabakaran Kumaresshan
+Prabhjyotsing Surjit Singh Sodhi
+Prabhu Marappan
+Pradyun Gedam
+Prashant Sharma
+Pratik Mallya
+pre-commit-ci[bot]
+Preet Thakkar
+Preston Holmes
+Przemek Wrzos
+Pulkit Goyal
+q0w
+Qiangning Hong
+Qiming Xu
+Quentin Lee
+Quentin Pradet
+R. David Murray
+Rafael Caricio
+Ralf Schmitt
+Ran Benita
+Randy Döring
+Razzi Abuissa
+rdb
+Reece Dunham
+Remi Rampin
+Rene Dudfield
+Riccardo Magliocchetti
+Riccardo Schirone
+Richard Jones
+Richard Si
+Ricky Ng-Adam
+Rishi
+rmorotti
+RobberPhex
+Robert Collins
+Robert McGibbon
+Robert Pollak
+Robert T. McGibbon
+robin elisha robinson
+Roey Berman
+Rohan Jain
+Roman Bogorodskiy
+Roman Donchenko
+Romuald Brunet
+ronaudinho
+Ronny Pfannschmidt
+Rory McCann
+Ross Brattain
+Roy Wellington Ⅳ
+Ruairidh MacLeod
+Russell Keith-Magee
+Ryan Shepherd
+Ryan Wooden
+ryneeverett
+S. Guliaev
+Sachi King
+Salvatore Rinchiera
+sandeepkiran-js
+Sander Van Balen
+Savio Jomton
+schlamar
+Scott Kitterman
+Sean
+seanj
+Sebastian Jordan
+Sebastian Schaetz
+Segev Finer
+SeongSoo Cho
+Sergey Vasilyev
+Seth Michael Larson
+Seth Woodworth
+Shahar Epstein
+Shantanu
+shenxianpeng
+shireenrao
+Shivansh-007
+Shixian Sheng
+Shlomi Fish
+Shovan Maity
+Simeon Visser
+Simon Cross
+Simon Pichugin
+sinoroc
+sinscary
+snook92
+socketubs
+Sorin Sbarnea
+Srinivas Nyayapati
+Srishti Hegde
+Stavros Korokithakis
+Stefan Scherfke
+Stefano Rivera
+Stephan Erb
+Stephen Rosen
+stepshal
+Steve (Gadget) Barnes
+Steve Barnes
+Steve Dower
+Steve Kowalik
+Steven Myint
+Steven Silvester
+stonebig
+studioj
+Stéphane Bidoul
+Stéphane Bidoul (ACSONE)
+Stéphane Klein
+Sumana Harihareswara
+Surbhi Sharma
+Sviatoslav Sydorenko
+Sviatoslav Sydorenko (Святослав Сидоренко)
+Swat009
+Sylvain
+Takayuki SHIMIZUKAWA
+Taneli Hukkinen
+tbeswick
+Thiago
+Thijs Triemstra
+Thomas Fenzl
+Thomas Grainger
+Thomas Guettler
+Thomas Johansson
+Thomas Kluyver
+Thomas Smith
+Thomas VINCENT
+Tim D. Smith
+Tim Gates
+Tim Harder
+Tim Heap
+tim smith
+tinruufu
+Tobias Hermann
+Tom Forbes
+Tom Freudenheim
+Tom V
+Tomas Hrnciar
+Tomas Orsava
+Tomer Chachamu
+Tommi Enenkel | AnB
+Tomáš Hrnčiar
+Tony Beswick
+Tony Narlock
+Tony Zhaocheng Tan
+TonyBeswick
+toonarmycaptain
+Toshio Kuratomi
+toxinu
+Travis Swicegood
+Tushar Sadhwani
+Tzu-ping Chung
+Valentin Haenel
+Victor Stinner
+victorvpaulo
+Vikram - Google
+Viktor Szépe
+Ville Skyttä
+Vinay Sajip
+Vincent Philippon
+Vinicyus Macedo
+Vipul Kumar
+Vitaly Babiy
+Vladimir Fokow
+Vladimir Rutsky
+W. Trevor King
+Wil Tan
+Wilfred Hughes
+William Edwards
+William ML Leslie
+William T Olson
+William Woodruff
+Wilson Mo
+wim glenn
+Winson Luk
+Wolfgang Maier
+Wu Zhenyu
+XAMES3
+Xavier Fernandez
+Xianpeng Shen
+xoviat
+xtreak
+YAMAMOTO Takashi
+Yen Chi Hsuan
+Yeray Diaz Diaz
+Yoval P
+Yu Jian
+Yuan Jing Vincent Yan
+Yusuke Hayashi
+Zearin
+Zhiping Deng
+ziebam
+Zvezdan Petkovic
+Łukasz Langa
+Роман Донченко
+Семён Марьясин
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/LICENSE.txt
new file mode 100644
index 00000000..8e7b65ea
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/LICENSE.txt
@@ -0,0 +1,20 @@
+Copyright (c) 2008-present The pip developers (see AUTHORS.txt file)
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice shall be
+included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/METADATA
new file mode 100644
index 00000000..3315c063
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/METADATA
@@ -0,0 +1,90 @@
+Metadata-Version: 2.2
+Name: pip
+Version: 25.0.1
+Summary: The PyPA recommended tool for installing Python packages.
+Author-email: The pip developers <distutils-sig@python.org>
+License: MIT
+Project-URL: Homepage, https://pip.pypa.io/
+Project-URL: Documentation, https://pip.pypa.io
+Project-URL: Source, https://github.com/pypa/pip
+Project-URL: Changelog, https://pip.pypa.io/en/stable/news/
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Topic :: Software Development :: Build Tools
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE.txt
+License-File: AUTHORS.txt
+
+pip - The Python Package Installer
+==================================
+
+.. |pypi-version| image:: https://img.shields.io/pypi/v/pip.svg
+   :target: https://pypi.org/project/pip/
+   :alt: PyPI
+
+.. |python-versions| image:: https://img.shields.io/pypi/pyversions/pip
+   :target: https://pypi.org/project/pip
+   :alt: PyPI - Python Version
+
+.. |docs-badge| image:: https://readthedocs.org/projects/pip/badge/?version=latest
+   :target: https://pip.pypa.io/en/latest
+   :alt: Documentation
+
+|pypi-version| |python-versions| |docs-badge|
+
+pip is the `package installer`_ for Python. You can use pip to install packages from the `Python Package Index`_ and other indexes.
+
+Please take a look at our documentation for how to install and use pip:
+
+* `Installation`_
+* `Usage`_
+
+We release updates regularly, with a new version every 3 months. Find more details in our documentation:
+
+* `Release notes`_
+* `Release process`_
+
+If you find bugs, need help, or want to talk to the developers, please use our mailing lists or chat rooms:
+
+* `Issue tracking`_
+* `Discourse channel`_
+* `User IRC`_
+
+If you want to get involved head over to GitHub to get the source code, look at our development documentation and feel free to jump on the developer mailing lists and chat rooms:
+
+* `GitHub page`_
+* `Development documentation`_
+* `Development IRC`_
+
+Code of Conduct
+---------------
+
+Everyone interacting in the pip project's codebases, issue trackers, chat
+rooms, and mailing lists is expected to follow the `PSF Code of Conduct`_.
+
+.. _package installer: https://packaging.python.org/guides/tool-recommendations/
+.. _Python Package Index: https://pypi.org
+.. _Installation: https://pip.pypa.io/en/stable/installation/
+.. _Usage: https://pip.pypa.io/en/stable/
+.. _Release notes: https://pip.pypa.io/en/stable/news.html
+.. _Release process: https://pip.pypa.io/en/latest/development/release-process/
+.. _GitHub page: https://github.com/pypa/pip
+.. _Development documentation: https://pip.pypa.io/en/latest/development
+.. _Issue tracking: https://github.com/pypa/pip/issues
+.. _Discourse channel: https://discuss.python.org/c/packaging
+.. _User IRC: https://kiwiirc.com/nextclient/#ircs://irc.libera.chat:+6697/pypa
+.. _Development IRC: https://kiwiirc.com/nextclient/#ircs://irc.libera.chat:+6697/pypa-dev
+.. _PSF Code of Conduct: https://github.com/pypa/.github/blob/main/CODE_OF_CONDUCT.md
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/RECORD
new file mode 100644
index 00000000..db8653e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/RECORD
@@ -0,0 +1,854 @@
+../../../bin/pip,sha256=TjceC8jpgAf8Mfahc2DMI3Z_a0WUekMulWp-jf0ExYM,278
+../../../bin/pip3,sha256=TjceC8jpgAf8Mfahc2DMI3Z_a0WUekMulWp-jf0ExYM,278
+../../../bin/pip3.12,sha256=TjceC8jpgAf8Mfahc2DMI3Z_a0WUekMulWp-jf0ExYM,278
+pip-25.0.1.dist-info/AUTHORS.txt,sha256=HqzpBVLfT1lBthqQfiDlVeFkg65hJ7ZQvvWhoq-BAsA,11018
+pip-25.0.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pip-25.0.1.dist-info/LICENSE.txt,sha256=Y0MApmnUmurmWxLGxIySTFGkzfPR_whtw0VtyLyqIQQ,1093
+pip-25.0.1.dist-info/METADATA,sha256=T6cxjPMPl523zsRcEsu8K0-IoV8S7vv1mmKR0KA6-SY,3677
+pip-25.0.1.dist-info/RECORD,,
+pip-25.0.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip-25.0.1.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+pip-25.0.1.dist-info/entry_points.txt,sha256=eeIjuzfnfR2PrhbjnbzFU6MnSS70kZLxwaHHq6M-bD0,87
+pip-25.0.1.dist-info/top_level.txt,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pip/__init__.py,sha256=aKiv_sTe7UbE7qmtCinJutFjqN0MndZQZ1fKLNwFFLE,357
+pip/__main__.py,sha256=WzbhHXTbSE6gBY19mNN9m4s5o_365LOvTYSgqgbdBhE,854
+pip/__pip-runner__.py,sha256=cPPWuJ6NK_k-GzfvlejLFgwzmYUROmpAR6QC3Q-vkXQ,1450
+pip/__pycache__/__init__.cpython-312.pyc,,
+pip/__pycache__/__main__.cpython-312.pyc,,
+pip/__pycache__/__pip-runner__.cpython-312.pyc,,
+pip/_internal/__init__.py,sha256=MfcoOluDZ8QMCFYal04IqOJ9q6m2V7a0aOsnI-WOxUo,513
+pip/_internal/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/__pycache__/build_env.cpython-312.pyc,,
+pip/_internal/__pycache__/cache.cpython-312.pyc,,
+pip/_internal/__pycache__/configuration.cpython-312.pyc,,
+pip/_internal/__pycache__/exceptions.cpython-312.pyc,,
+pip/_internal/__pycache__/main.cpython-312.pyc,,
+pip/_internal/__pycache__/pyproject.cpython-312.pyc,,
+pip/_internal/__pycache__/self_outdated_check.cpython-312.pyc,,
+pip/_internal/__pycache__/wheel_builder.cpython-312.pyc,,
+pip/_internal/build_env.py,sha256=Dv4UCClSg4uNaal_hL-trg5-zl3Is9CuIDxkChCkXF4,10700
+pip/_internal/cache.py,sha256=Jb698p5PNigRtpW5o26wQNkkUv4MnQ94mc471wL63A0,10369
+pip/_internal/cli/__init__.py,sha256=FkHBgpxxb-_gd6r1FjnNhfMOzAUYyXoXKJ6abijfcFU,132
+pip/_internal/cli/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/autocompletion.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/base_command.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/cmdoptions.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/command_context.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/index_command.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/main.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/main_parser.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/parser.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/progress_bars.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/req_command.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/spinners.cpython-312.pyc,,
+pip/_internal/cli/__pycache__/status_codes.cpython-312.pyc,,
+pip/_internal/cli/autocompletion.py,sha256=Lli3Mr6aDNu7ZkJJFFvwD2-hFxNI6Avz8OwMyS5TVrs,6865
+pip/_internal/cli/base_command.py,sha256=NZin6KMzW9NSYzKk4Tc8isb_TQYKR4CKd5j9mSm46PI,8625
+pip/_internal/cli/cmdoptions.py,sha256=V3BB22F4_v_RkHaZ5onWnszhbBtjYZvNhbn9M0NO0HI,30116
+pip/_internal/cli/command_context.py,sha256=RHgIPwtObh5KhMrd3YZTkl8zbVG-6Okml7YbFX4Ehg0,774
+pip/_internal/cli/index_command.py,sha256=i_sgNlPmXC5iHUaY-dmmrHKKTgc5O4hWzisr5Al1rr0,5677
+pip/_internal/cli/main.py,sha256=BDZef-bWe9g9Jpr4OVs4dDf-845HJsKw835T7AqEnAc,2817
+pip/_internal/cli/main_parser.py,sha256=laDpsuBDl6kyfywp9eMMA9s84jfH2TJJn-vmL0GG90w,4338
+pip/_internal/cli/parser.py,sha256=VCMtduzECUV87KaHNu-xJ-wLNL82yT3x16V4XBxOAqI,10825
+pip/_internal/cli/progress_bars.py,sha256=9GcgusWtwfqou2zhAQp1XNbQHIDslqyyz9UwLzw7Jgc,2717
+pip/_internal/cli/req_command.py,sha256=DqeFhmUMs6o6Ev8qawAcOoYNdAZsfyKS0MZI5jsJYwQ,12250
+pip/_internal/cli/spinners.py,sha256=hIJ83GerdFgFCdobIA23Jggetegl_uC4Sp586nzFbPE,5118
+pip/_internal/cli/status_codes.py,sha256=sEFHUaUJbqv8iArL3HAtcztWZmGOFX01hTesSytDEh0,116
+pip/_internal/commands/__init__.py,sha256=5oRO9O3dM2vGuh0bFw4HOVletryrz5HHMmmPWwJrH9U,3882
+pip/_internal/commands/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/cache.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/check.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/completion.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/configuration.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/debug.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/download.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/freeze.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/hash.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/help.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/index.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/inspect.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/install.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/list.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/search.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/show.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/uninstall.cpython-312.pyc,,
+pip/_internal/commands/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/commands/cache.py,sha256=IOezTicHjGE5sWdBx2nwPVgbjuJHM3s-BZEkpZLemuY,8107
+pip/_internal/commands/check.py,sha256=Hr_4eiMd9cgVDgEvjtIdw915NmL7ROIWW8enkr8slPQ,2268
+pip/_internal/commands/completion.py,sha256=HT4lD0bgsflHq2IDgYfiEdp7IGGtE7s6MgI3xn0VQEw,4287
+pip/_internal/commands/configuration.py,sha256=n98enwp6y0b5G6fiRQjaZo43FlJKYve_daMhN-4BRNc,9766
+pip/_internal/commands/debug.py,sha256=DNDRgE9YsKrbYzU0s3VKi8rHtKF4X13CJ_br_8PUXO0,6797
+pip/_internal/commands/download.py,sha256=0qB0nys6ZEPsog451lDsjL5Bx7Z97t-B80oFZKhpzKM,5273
+pip/_internal/commands/freeze.py,sha256=2Vt72BYTSm9rzue6d8dNzt8idxWK4Db6Hd-anq7GQ80,3203
+pip/_internal/commands/hash.py,sha256=EVVOuvGtoPEdFi8SNnmdqlCQrhCxV-kJsdwtdcCnXGQ,1703
+pip/_internal/commands/help.py,sha256=gcc6QDkcgHMOuAn5UxaZwAStsRBrnGSn_yxjS57JIoM,1132
+pip/_internal/commands/index.py,sha256=RAXxmJwFhVb5S1BYzb5ifX3sn9Na8v2CCVYwSMP8pao,4731
+pip/_internal/commands/inspect.py,sha256=PGrY9TRTRCM3y5Ml8Bdk8DEOXquWRfscr4DRo1LOTPc,3189
+pip/_internal/commands/install.py,sha256=r3yHQUxvxt7gD5j9n6zRDslAvtx9CT_whLuQJcktp6M,29390
+pip/_internal/commands/list.py,sha256=oiIzSjLP6__d7dIS3q0Xb5ywsaOThBWRqMyjjKzkPdM,12769
+pip/_internal/commands/search.py,sha256=fWkUQVx_gm8ebbFAlCgqtxKXT9rNahpJ-BI__3HNZpg,5626
+pip/_internal/commands/show.py,sha256=0YBhCga3PAd81vT3l7UWflktSpB5-aYqQcJxBVPazVM,7857
+pip/_internal/commands/uninstall.py,sha256=7pOR7enK76gimyxQbzxcG1OsyLXL3DvX939xmM8Fvtg,3892
+pip/_internal/commands/wheel.py,sha256=eJRhr_qoNNxWAkkdJCNiQM7CXd4E1_YyQhsqJnBPGGg,6414
+pip/_internal/configuration.py,sha256=-KOok6jh3hFzXMPQFPJ1_EFjBpAsge-RSreQuLHLmzo,14005
+pip/_internal/distributions/__init__.py,sha256=Hq6kt6gXBgjNit5hTTWLAzeCNOKoB-N0pGYSqehrli8,858
+pip/_internal/distributions/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/distributions/__pycache__/base.cpython-312.pyc,,
+pip/_internal/distributions/__pycache__/installed.cpython-312.pyc,,
+pip/_internal/distributions/__pycache__/sdist.cpython-312.pyc,,
+pip/_internal/distributions/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/distributions/base.py,sha256=QeB9qvKXDIjLdPBDE5fMgpfGqMMCr-govnuoQnGuiF8,1783
+pip/_internal/distributions/installed.py,sha256=QinHFbWAQ8oE0pbD8MFZWkwlnfU1QYTccA1vnhrlYOU,842
+pip/_internal/distributions/sdist.py,sha256=PlcP4a6-R6c98XnOM-b6Lkb3rsvh9iG4ok8shaanrzs,6751
+pip/_internal/distributions/wheel.py,sha256=THBYfnv7VVt8mYhMYUtH13S1E7FDwtDyDfmUcl8ai0E,1317
+pip/_internal/exceptions.py,sha256=2_byISIv3kSnI_9T-Esfxrt0LnTRgcUHyxu0twsHjQY,26481
+pip/_internal/index/__init__.py,sha256=vpt-JeTZefh8a-FC22ZeBSXFVbuBcXSGiILhQZJaNpQ,30
+pip/_internal/index/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/index/__pycache__/collector.cpython-312.pyc,,
+pip/_internal/index/__pycache__/package_finder.cpython-312.pyc,,
+pip/_internal/index/__pycache__/sources.cpython-312.pyc,,
+pip/_internal/index/collector.py,sha256=RdPO0JLAlmyBWPAWYHPyRoGjz3GNAeTngCNkbGey_mE,16265
+pip/_internal/index/package_finder.py,sha256=mJHAljlHeHuclyuxtjvBZO6DtovKjsZjF_tCh_wux5E,38076
+pip/_internal/index/sources.py,sha256=lPBLK5Xiy8Q6IQMio26Wl7ocfZOKkgGklIBNyUJ23fI,8632
+pip/_internal/locations/__init__.py,sha256=UaAxeZ_f93FyouuFf4p7SXYF-4WstXuEvd3LbmPCAno,14925
+pip/_internal/locations/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/locations/__pycache__/_distutils.cpython-312.pyc,,
+pip/_internal/locations/__pycache__/_sysconfig.cpython-312.pyc,,
+pip/_internal/locations/__pycache__/base.cpython-312.pyc,,
+pip/_internal/locations/_distutils.py,sha256=x6nyVLj7X11Y4khIdf-mFlxMl2FWadtVEgeb8upc_WI,6013
+pip/_internal/locations/_sysconfig.py,sha256=IGzds60qsFneRogC-oeBaY7bEh3lPt_v47kMJChQXsU,7724
+pip/_internal/locations/base.py,sha256=RQiPi1d4FVM2Bxk04dQhXZ2PqkeljEL2fZZ9SYqIQ78,2556
+pip/_internal/main.py,sha256=r-UnUe8HLo5XFJz8inTcOOTiu_sxNhgHb6VwlGUllOI,340
+pip/_internal/metadata/__init__.py,sha256=CU8jK1TZso7jOLdr0sX9xDjrcs5iy8d7IRK-hvaIO5Y,4337
+pip/_internal/metadata/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/metadata/__pycache__/_json.cpython-312.pyc,,
+pip/_internal/metadata/__pycache__/base.cpython-312.pyc,,
+pip/_internal/metadata/__pycache__/pkg_resources.cpython-312.pyc,,
+pip/_internal/metadata/_json.py,sha256=ezrIYazHCINM2QUk1eA9wEAMj3aeGWeDVgGalgUzKpc,2707
+pip/_internal/metadata/base.py,sha256=ft0K5XNgI4ETqZnRv2-CtvgYiMOMAeGMAzxT-f6VLJA,25298
+pip/_internal/metadata/importlib/__init__.py,sha256=jUUidoxnHcfITHHaAWG1G2i5fdBYklv_uJcjo2x7VYE,135
+pip/_internal/metadata/importlib/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/metadata/importlib/__pycache__/_compat.cpython-312.pyc,,
+pip/_internal/metadata/importlib/__pycache__/_dists.cpython-312.pyc,,
+pip/_internal/metadata/importlib/__pycache__/_envs.cpython-312.pyc,,
+pip/_internal/metadata/importlib/_compat.py,sha256=c6av8sP8BBjAZuFSJow1iWfygUXNM3xRTCn5nqw6B9M,2796
+pip/_internal/metadata/importlib/_dists.py,sha256=ftmYiyfUGUIjnVwt6W-Ijsimy5c28KgmXly5Q5IQ2P4,8279
+pip/_internal/metadata/importlib/_envs.py,sha256=UUB980XSrDWrMpQ1_G45i0r8Hqlg_tg3IPQ63mEqbNc,7431
+pip/_internal/metadata/pkg_resources.py,sha256=U07ETAINSGeSRBfWUG93E4tZZbaW_f7PGzEqZN0hulc,10542
+pip/_internal/models/__init__.py,sha256=3DHUd_qxpPozfzouoqa9g9ts1Czr5qaHfFxbnxriepM,63
+pip/_internal/models/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/models/__pycache__/candidate.cpython-312.pyc,,
+pip/_internal/models/__pycache__/direct_url.cpython-312.pyc,,
+pip/_internal/models/__pycache__/format_control.cpython-312.pyc,,
+pip/_internal/models/__pycache__/index.cpython-312.pyc,,
+pip/_internal/models/__pycache__/installation_report.cpython-312.pyc,,
+pip/_internal/models/__pycache__/link.cpython-312.pyc,,
+pip/_internal/models/__pycache__/scheme.cpython-312.pyc,,
+pip/_internal/models/__pycache__/search_scope.cpython-312.pyc,,
+pip/_internal/models/__pycache__/selection_prefs.cpython-312.pyc,,
+pip/_internal/models/__pycache__/target_python.cpython-312.pyc,,
+pip/_internal/models/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/models/candidate.py,sha256=zzgFRuw_kWPjKpGw7LC0ZUMD2CQ2EberUIYs8izjdCA,753
+pip/_internal/models/direct_url.py,sha256=uBtY2HHd3TO9cKQJWh0ThvE5FRr-MWRYChRU4IG9HZE,6578
+pip/_internal/models/format_control.py,sha256=wtsQqSK9HaUiNxQEuB-C62eVimw6G4_VQFxV9-_KDBE,2486
+pip/_internal/models/index.py,sha256=tYnL8oxGi4aSNWur0mG8DAP7rC6yuha_MwJO8xw0crI,1030
+pip/_internal/models/installation_report.py,sha256=zRVZoaz-2vsrezj_H3hLOhMZCK9c7TbzWgC-jOalD00,2818
+pip/_internal/models/link.py,sha256=GQ8hq7x-FDFPv25Nbn2veIM-MlBrGZDGLd7aZeF4Xrg,21448
+pip/_internal/models/scheme.py,sha256=PakmHJM3e8OOWSZFtfz1Az7f1meONJnkGuQxFlt3wBE,575
+pip/_internal/models/search_scope.py,sha256=67NEnsYY84784S-MM7ekQuo9KXLH-7MzFntXjapvAo0,4531
+pip/_internal/models/selection_prefs.py,sha256=qaFfDs3ciqoXPg6xx45N1jPLqccLJw4N0s4P0PyHTQ8,2015
+pip/_internal/models/target_python.py,sha256=2XaH2rZ5ZF-K5wcJbEMGEl7SqrTToDDNkrtQ2v_v_-Q,4271
+pip/_internal/models/wheel.py,sha256=G7dND_s4ebPkEL7RJ1qCY0QhUUWIIK6AnjWgRATF5no,4539
+pip/_internal/network/__init__.py,sha256=jf6Tt5nV_7zkARBrKojIXItgejvoegVJVKUbhAa5Ioc,50
+pip/_internal/network/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/network/__pycache__/auth.cpython-312.pyc,,
+pip/_internal/network/__pycache__/cache.cpython-312.pyc,,
+pip/_internal/network/__pycache__/download.cpython-312.pyc,,
+pip/_internal/network/__pycache__/lazy_wheel.cpython-312.pyc,,
+pip/_internal/network/__pycache__/session.cpython-312.pyc,,
+pip/_internal/network/__pycache__/utils.cpython-312.pyc,,
+pip/_internal/network/__pycache__/xmlrpc.cpython-312.pyc,,
+pip/_internal/network/auth.py,sha256=D4gASjUrqoDFlSt6gQ767KAAjv6PUyJU0puDlhXNVRE,20809
+pip/_internal/network/cache.py,sha256=0yGMA3Eet59xBSLtbPAenvI53dl29oUOeqZ2c0QL2Ss,4614
+pip/_internal/network/download.py,sha256=FLOP29dPYECBiAi7eEjvAbNkyzaKNqbyjOT2m8HPW8U,6048
+pip/_internal/network/lazy_wheel.py,sha256=PBdoMoNQQIA84Fhgne38jWF52W4x_KtsHjxgv4dkRKA,7622
+pip/_internal/network/session.py,sha256=msM4es16LmmNEYNkrYyg8fTc7gAHbKFltawfKP27LOI,18771
+pip/_internal/network/utils.py,sha256=Inaxel-NxBu4PQWkjyErdnfewsFCcgHph7dzR1-FboY,4088
+pip/_internal/network/xmlrpc.py,sha256=sAxzOacJ-N1NXGPvap9jC3zuYWSnnv3GXtgR2-E2APA,1838
+pip/_internal/operations/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/operations/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/operations/__pycache__/check.cpython-312.pyc,,
+pip/_internal/operations/__pycache__/freeze.cpython-312.pyc,,
+pip/_internal/operations/__pycache__/prepare.cpython-312.pyc,,
+pip/_internal/operations/build/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/operations/build/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/build_tracker.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/metadata.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/metadata_editable.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/metadata_legacy.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/wheel_editable.cpython-312.pyc,,
+pip/_internal/operations/build/__pycache__/wheel_legacy.cpython-312.pyc,,
+pip/_internal/operations/build/build_tracker.py,sha256=-ARW_TcjHCOX7D2NUOGntB4Fgc6b4aolsXkAK6BWL7w,4774
+pip/_internal/operations/build/metadata.py,sha256=9S0CUD8U3QqZeXp-Zyt8HxwU90lE4QrnYDgrqZDzBnc,1422
+pip/_internal/operations/build/metadata_editable.py,sha256=xlAwcP9q_8_fmv_3I39w9EZ7SQV9hnJZr9VuTsq2Y68,1510
+pip/_internal/operations/build/metadata_legacy.py,sha256=8i6i1QZX9m_lKPStEFsHKM0MT4a-CD408JOw99daLmo,2190
+pip/_internal/operations/build/wheel.py,sha256=sT12FBLAxDC6wyrDorh8kvcZ1jG5qInCRWzzP-UkJiQ,1075
+pip/_internal/operations/build/wheel_editable.py,sha256=yOtoH6zpAkoKYEUtr8FhzrYnkNHQaQBjWQ2HYae1MQg,1417
+pip/_internal/operations/build/wheel_legacy.py,sha256=K-6kNhmj-1xDF45ny1yheMerF0ui4EoQCLzEoHh6-tc,3045
+pip/_internal/operations/check.py,sha256=L24vRL8VWbyywdoeAhM89WCd8zLTnjIbULlKelUgIec,5912
+pip/_internal/operations/freeze.py,sha256=1_M79jAQKnCxWr-KCCmHuVXOVFGaUJHmoWLfFzgh7K4,9843
+pip/_internal/operations/install/__init__.py,sha256=mX7hyD2GNBO2mFGokDQ30r_GXv7Y_PLdtxcUv144e-s,51
+pip/_internal/operations/install/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/operations/install/__pycache__/editable_legacy.cpython-312.pyc,,
+pip/_internal/operations/install/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/operations/install/editable_legacy.py,sha256=PoEsNEPGbIZ2yQphPsmYTKLOCMs4gv5OcCdzW124NcA,1283
+pip/_internal/operations/install/wheel.py,sha256=X5Iz9yUg5LlK5VNQ9g2ikc6dcRu8EPi_SUi5iuEDRgo,27615
+pip/_internal/operations/prepare.py,sha256=joWJwPkuqGscQgVNImLK71e9hRapwKvRCM8HclysmvU,28118
+pip/_internal/pyproject.py,sha256=GLJ6rWRS5_2noKdajohoLyDty57Z7QXhcUAYghmTnWc,7286
+pip/_internal/req/__init__.py,sha256=HxBFtZy_BbCclLgr26waMtpzYdO5T3vxePvpGAXSt5s,2653
+pip/_internal/req/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/req/__pycache__/constructors.cpython-312.pyc,,
+pip/_internal/req/__pycache__/req_file.cpython-312.pyc,,
+pip/_internal/req/__pycache__/req_install.cpython-312.pyc,,
+pip/_internal/req/__pycache__/req_set.cpython-312.pyc,,
+pip/_internal/req/__pycache__/req_uninstall.cpython-312.pyc,,
+pip/_internal/req/constructors.py,sha256=v1qzCN1mIldwx-nCrPc8JO4lxkm3Fv8M5RWvt8LISjc,18430
+pip/_internal/req/req_file.py,sha256=eys82McgaICOGic2UZRHjD720piKJPwmeSYdXlWwl6w,20234
+pip/_internal/req/req_install.py,sha256=BMptxHYg2uG_b-7HFEULPb3nuw0FMAbuea8zTq2rE7w,35786
+pip/_internal/req/req_set.py,sha256=j3esG0s6SzoVReX9rWn4rpYNtyET_fwxbwJPRimvRxo,2858
+pip/_internal/req/req_uninstall.py,sha256=qzDIxJo-OETWqGais7tSMCDcWbATYABT-Tid3ityF0s,23853
+pip/_internal/resolution/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/resolution/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/resolution/__pycache__/base.cpython-312.pyc,,
+pip/_internal/resolution/base.py,sha256=qlmh325SBVfvG6Me9gc5Nsh5sdwHBwzHBq6aEXtKsLA,583
+pip/_internal/resolution/legacy/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/resolution/legacy/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/resolution/legacy/__pycache__/resolver.cpython-312.pyc,,
+pip/_internal/resolution/legacy/resolver.py,sha256=3HZiJBRd1FTN6jQpI4qRO8-TbLYeIbUTS6PFvXnXs2w,24068
+pip/_internal/resolution/resolvelib/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/resolution/resolvelib/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/base.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/candidates.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/factory.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/found_candidates.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/provider.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/reporter.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/requirements.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/__pycache__/resolver.cpython-312.pyc,,
+pip/_internal/resolution/resolvelib/base.py,sha256=DCf669FsqyQY5uqXeePDHQY1e4QO-pBzWH8O0s9-K94,5023
+pip/_internal/resolution/resolvelib/candidates.py,sha256=5UZ1upNnmqsP-nmEZaDYxaBgCoejw_e2WVGmmAvBxXc,20001
+pip/_internal/resolution/resolvelib/factory.py,sha256=MJOLSZJY8_28PPdcutoQ6gjJ_1eBDt6Z1edtfTJyR4E,32659
+pip/_internal/resolution/resolvelib/found_candidates.py,sha256=9hrTyQqFvl9I7Tji79F1AxHv39Qh1rkJ_7deSHSMfQc,6383
+pip/_internal/resolution/resolvelib/provider.py,sha256=bcsFnYvlmtB80cwVdW1fIwgol8ZNr1f1VHyRTkz47SM,9935
+pip/_internal/resolution/resolvelib/reporter.py,sha256=00JtoXEkTlw0-rl_sl54d71avwOsJHt9GGHcrj5Sza0,3168
+pip/_internal/resolution/resolvelib/requirements.py,sha256=7JG4Z72e5Yk4vU0S5ulGvbqTy4FMQGYhY5zQhX9zTtY,8065
+pip/_internal/resolution/resolvelib/resolver.py,sha256=nLJOsVMEVi2gQUVJoUFKMZAeu2f7GRMjGMvNSWyz0Bc,12592
+pip/_internal/self_outdated_check.py,sha256=1PFtttvLAeyCVR3tPoBq2sOlPD0IJ-KSqU6bc1HUk9c,8318
+pip/_internal/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_internal/utils/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/_jaraco_text.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/_log.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/appdirs.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/compat.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/compatibility_tags.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/datetime.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/deprecation.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/direct_url_helpers.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/egg_link.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/entrypoints.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/filesystem.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/filetypes.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/glibc.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/hashes.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/logging.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/misc.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/packaging.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/retry.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/setuptools_build.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/subprocess.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/temp_dir.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/unpacking.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/urls.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/virtualenv.cpython-312.pyc,,
+pip/_internal/utils/__pycache__/wheel.cpython-312.pyc,,
+pip/_internal/utils/_jaraco_text.py,sha256=M15uUPIh5NpP1tdUGBxRau6q1ZAEtI8-XyLEETscFfE,3350
+pip/_internal/utils/_log.py,sha256=-jHLOE_THaZz5BFcCnoSL9EYAtJ0nXem49s9of4jvKw,1015
+pip/_internal/utils/appdirs.py,sha256=swgcTKOm3daLeXTW6v5BUS2Ti2RvEnGRQYH_yDXklAo,1665
+pip/_internal/utils/compat.py,sha256=ckkFveBiYQjRWjkNsajt_oWPS57tJvE8XxoC4OIYgCY,2399
+pip/_internal/utils/compatibility_tags.py,sha256=OWq5axHpW-MEEPztGdvgADrgJPAcV9a88Rxm4Z8VBs8,6272
+pip/_internal/utils/datetime.py,sha256=m21Y3wAtQc-ji6Veb6k_M5g6A0ZyFI4egchTdnwh-pQ,242
+pip/_internal/utils/deprecation.py,sha256=k7Qg_UBAaaTdyq82YVARA6D7RmcGTXGv7fnfcgigj4Q,3707
+pip/_internal/utils/direct_url_helpers.py,sha256=r2MRtkVDACv9AGqYODBUC9CjwgtsUU1s68hmgfCJMtA,3196
+pip/_internal/utils/egg_link.py,sha256=0FePZoUYKv4RGQ2t6x7w5Z427wbA_Uo3WZnAkrgsuqo,2463
+pip/_internal/utils/entrypoints.py,sha256=YlhLTRl2oHBAuqhc-zmL7USS67TPWVHImjeAQHreZTQ,3064
+pip/_internal/utils/filesystem.py,sha256=ajvA-q4ocliW9kPp8Yquh-4vssXbu-UKbo5FV9V4X64,4950
+pip/_internal/utils/filetypes.py,sha256=i8XAQ0eFCog26Fw9yV0Yb1ygAqKYB1w9Cz9n0fj8gZU,716
+pip/_internal/utils/glibc.py,sha256=vUkWq_1pJuzcYNcGKLlQmABoUiisK8noYY1yc8Wq4w4,3734
+pip/_internal/utils/hashes.py,sha256=XGGLL0AG8-RhWnyz87xF6MFZ--BKadHU35D47eApCKI,4972
+pip/_internal/utils/logging.py,sha256=ONfbrhaD248akkosK79if97n20EABxwjOxp5dE5RCRY,11845
+pip/_internal/utils/misc.py,sha256=DWnYxBUItjRp7hhxEg4ih6P6YpKrykM86dbi_EcU8SQ,23450
+pip/_internal/utils/packaging.py,sha256=cm-X_0HVHV_jRwUVZh6AuEWqSitzf8EpaJ7Uv2UGu6A,2142
+pip/_internal/utils/retry.py,sha256=mhFbykXjhTnZfgzeuy-vl9c8nECnYn_CMtwNJX2tYzQ,1392
+pip/_internal/utils/setuptools_build.py,sha256=ouXpud-jeS8xPyTPsXJ-m34NPvK5os45otAzdSV_IJE,4435
+pip/_internal/utils/subprocess.py,sha256=EsvqSRiSMHF98T8Txmu6NLU3U--MpTTQjtNgKP0P--M,8988
+pip/_internal/utils/temp_dir.py,sha256=5qOXe8M4JeY6vaFQM867d5zkp1bSwMZ-KT5jymmP0Zg,9310
+pip/_internal/utils/unpacking.py,sha256=_gVdyzTRDMYktpnYljn4OoxrZTtMCf4xknSm4rK0WaA,11967
+pip/_internal/utils/urls.py,sha256=qceSOZb5lbNDrHNsv7_S4L4Ytszja5NwPKUMnZHbYnM,1599
+pip/_internal/utils/virtualenv.py,sha256=S6f7csYorRpiD6cvn3jISZYc3I8PJC43H5iMFpRAEDU,3456
+pip/_internal/utils/wheel.py,sha256=b442jkydFHjXzDy6cMR7MpzWBJ1Q82hR5F33cmcHV3g,4494
+pip/_internal/vcs/__init__.py,sha256=UAqvzpbi0VbZo3Ub6skEeZAw-ooIZR-zX_WpCbxyCoU,596
+pip/_internal/vcs/__pycache__/__init__.cpython-312.pyc,,
+pip/_internal/vcs/__pycache__/bazaar.cpython-312.pyc,,
+pip/_internal/vcs/__pycache__/git.cpython-312.pyc,,
+pip/_internal/vcs/__pycache__/mercurial.cpython-312.pyc,,
+pip/_internal/vcs/__pycache__/subversion.cpython-312.pyc,,
+pip/_internal/vcs/__pycache__/versioncontrol.cpython-312.pyc,,
+pip/_internal/vcs/bazaar.py,sha256=EKStcQaKpNu0NK4p5Q10Oc4xb3DUxFw024XrJy40bFQ,3528
+pip/_internal/vcs/git.py,sha256=3tpc9LQA_J4IVW5r5NvWaaSeDzcmJOrSFZN0J8vIKfU,18177
+pip/_internal/vcs/mercurial.py,sha256=oULOhzJ2Uie-06d1omkL-_Gc6meGaUkyogvqG9ZCyPs,5249
+pip/_internal/vcs/subversion.py,sha256=ddTugHBqHzV3ebKlU5QXHPN4gUqlyXbOx8q8NgXKvs8,11735
+pip/_internal/vcs/versioncontrol.py,sha256=cvf_-hnTAjQLXJ3d17FMNhQfcO1AcKWUF10tfrYyP-c,22440
+pip/_internal/wheel_builder.py,sha256=DL3A8LKeRj_ACp11WS5wSgASgPFqeyAeXJKdXfmaWXU,11799
+pip/_vendor/__init__.py,sha256=JYuAXvClhInxIrA2FTp5p-uuWVL7WV6-vEpTs46-Qh4,4873
+pip/_vendor/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/__pycache__/typing_extensions.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__init__.py,sha256=LMC5CBe94ZRL5xhlzwyPDmHXvBD0p7lT4R3Z73D6a_I,677
+pip/_vendor/cachecontrol/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/_cmd.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/adapter.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/cache.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/controller.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/filewrapper.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/heuristics.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/serialize.cpython-312.pyc,,
+pip/_vendor/cachecontrol/__pycache__/wrapper.cpython-312.pyc,,
+pip/_vendor/cachecontrol/_cmd.py,sha256=iist2EpzJvDVIhMAxXq8iFnTBsiZAd6iplxfmNboNyk,1737
+pip/_vendor/cachecontrol/adapter.py,sha256=febjY4LV87iiCIK3jcl8iH58iaSA7b9WkovsByIDK0Y,6348
+pip/_vendor/cachecontrol/cache.py,sha256=OXwv7Fn2AwnKNiahJHnjtvaKLndvVLv_-zO-ltlV9qI,1953
+pip/_vendor/cachecontrol/caches/__init__.py,sha256=dtrrroK5BnADR1GWjCZ19aZ0tFsMfvFBtLQQU1sp_ag,303
+pip/_vendor/cachecontrol/caches/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/cachecontrol/caches/__pycache__/file_cache.cpython-312.pyc,,
+pip/_vendor/cachecontrol/caches/__pycache__/redis_cache.cpython-312.pyc,,
+pip/_vendor/cachecontrol/caches/file_cache.py,sha256=b7oMgsRSqPmEsonVJw6uFEYUlFgD6GF8TyacOGG1x3M,5399
+pip/_vendor/cachecontrol/caches/redis_cache.py,sha256=9rmqwtYu_ljVkW6_oLqbC7EaX_a8YT_yLuna-eS0dgo,1386
+pip/_vendor/cachecontrol/controller.py,sha256=glbPj2iZlGqdBg8z09D2DtQOzoOGXnWvy7K2LEyBsEQ,18576
+pip/_vendor/cachecontrol/filewrapper.py,sha256=2ktXNPE0KqnyzF24aOsKCA58HQq1xeC6l2g6_zwjghc,4291
+pip/_vendor/cachecontrol/heuristics.py,sha256=gqMXU8w0gQuEQiSdu3Yg-0vd9kW7nrWKbLca75rheGE,4881
+pip/_vendor/cachecontrol/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/cachecontrol/serialize.py,sha256=HQd2IllQ05HzPkVLMXTF2uX5mjEQjDBkxCqUJUODpZk,5163
+pip/_vendor/cachecontrol/wrapper.py,sha256=hsGc7g8QGQTT-4f8tgz3AM5qwScg6FO0BSdLSRdEvpU,1417
+pip/_vendor/certifi/__init__.py,sha256=p_GYZrjUwPBUhpLlCZoGb0miKBKSqDAyZC5DvIuqbHQ,94
+pip/_vendor/certifi/__main__.py,sha256=1k3Cr95vCxxGRGDljrW3wMdpZdL3Nhf0u1n-k2qdsCY,255
+pip/_vendor/certifi/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/certifi/__pycache__/__main__.cpython-312.pyc,,
+pip/_vendor/certifi/__pycache__/core.cpython-312.pyc,,
+pip/_vendor/certifi/cacert.pem,sha256=lO3rZukXdPyuk6BWUJFOKQliWaXH6HGh9l1GGrUgG0c,299427
+pip/_vendor/certifi/core.py,sha256=2SRT5rIcQChFDbe37BQa-kULxAgJ8qN6l1jfqTp4HIs,4486
+pip/_vendor/certifi/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/distlib/__init__.py,sha256=dcwgYGYGQqAEawBXPDtIx80DO_3cOmFv8HTc8JMzknQ,625
+pip/_vendor/distlib/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/compat.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/database.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/index.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/locators.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/manifest.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/markers.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/metadata.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/resources.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/scripts.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/util.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/version.cpython-312.pyc,,
+pip/_vendor/distlib/__pycache__/wheel.cpython-312.pyc,,
+pip/_vendor/distlib/compat.py,sha256=2jRSjRI4o-vlXeTK2BCGIUhkc6e9ZGhSsacRM5oseTw,41467
+pip/_vendor/distlib/database.py,sha256=mHy_LxiXIsIVRb-T0-idBrVLw3Ffij5teHCpbjmJ9YU,51160
+pip/_vendor/distlib/index.py,sha256=lTbw268rRhj8dw1sib3VZ_0EhSGgoJO3FKJzSFMOaeA,20797
+pip/_vendor/distlib/locators.py,sha256=oBeAZpFuPQSY09MgNnLfQGGAXXvVO96BFpZyKMuK4tM,51026
+pip/_vendor/distlib/manifest.py,sha256=3qfmAmVwxRqU1o23AlfXrQGZzh6g_GGzTAP_Hb9C5zQ,14168
+pip/_vendor/distlib/markers.py,sha256=X6sDvkFGcYS8gUW8hfsWuKEKAqhQZAJ7iXOMLxRYjYk,5164
+pip/_vendor/distlib/metadata.py,sha256=zil3sg2EUfLXVigljY2d_03IJt-JSs7nX-73fECMX2s,38724
+pip/_vendor/distlib/resources.py,sha256=LwbPksc0A1JMbi6XnuPdMBUn83X7BPuFNWqPGEKI698,10820
+pip/_vendor/distlib/scripts.py,sha256=BJliaDAZaVB7WAkwokgC3HXwLD2iWiHaVI50H7C6eG8,18608
+pip/_vendor/distlib/t32.exe,sha256=a0GV5kCoWsMutvliiCKmIgV98eRZ33wXoS-XrqvJQVs,97792
+pip/_vendor/distlib/t64-arm.exe,sha256=68TAa32V504xVBnufojh0PcenpR3U4wAqTqf-MZqbPw,182784
+pip/_vendor/distlib/t64.exe,sha256=gaYY8hy4fbkHYTTnA4i26ct8IQZzkBG2pRdy0iyuBrc,108032
+pip/_vendor/distlib/util.py,sha256=vMPGvsS4j9hF6Y9k3Tyom1aaHLb0rFmZAEyzeAdel9w,66682
+pip/_vendor/distlib/version.py,sha256=s5VIs8wBn0fxzGxWM_aA2ZZyx525HcZbMvcTlTyZ3Rg,23727
+pip/_vendor/distlib/w32.exe,sha256=R4csx3-OGM9kL4aPIzQKRo5TfmRSHZo6QWyLhDhNBks,91648
+pip/_vendor/distlib/w64-arm.exe,sha256=xdyYhKj0WDcVUOCb05blQYvzdYIKMbmJn2SZvzkcey4,168448
+pip/_vendor/distlib/w64.exe,sha256=ejGf-rojoBfXseGLpya6bFTFPWRG21X5KvU8J5iU-K0,101888
+pip/_vendor/distlib/wheel.py,sha256=DFIVguEQHCdxnSdAO0dfFsgMcvVZitg7bCOuLwZ7A_s,43979
+pip/_vendor/distro/__init__.py,sha256=2fHjF-SfgPvjyNZ1iHh_wjqWdR_Yo5ODHwZC0jLBPhc,981
+pip/_vendor/distro/__main__.py,sha256=bu9d3TifoKciZFcqRBuygV3GSuThnVD_m2IK4cz96Vs,64
+pip/_vendor/distro/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/distro/__pycache__/__main__.cpython-312.pyc,,
+pip/_vendor/distro/__pycache__/distro.cpython-312.pyc,,
+pip/_vendor/distro/distro.py,sha256=XqbefacAhDT4zr_trnbA15eY8vdK4GTghgmvUGrEM_4,49430
+pip/_vendor/distro/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/idna/__init__.py,sha256=MPqNDLZbXqGaNdXxAFhiqFPKEQXju2jNQhCey6-5eJM,868
+pip/_vendor/idna/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/codec.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/compat.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/core.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/idnadata.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/intranges.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/package_data.cpython-312.pyc,,
+pip/_vendor/idna/__pycache__/uts46data.cpython-312.pyc,,
+pip/_vendor/idna/codec.py,sha256=PEew3ItwzjW4hymbasnty2N2OXvNcgHB-JjrBuxHPYY,3422
+pip/_vendor/idna/compat.py,sha256=RzLy6QQCdl9784aFhb2EX9EKGCJjg0P3PilGdeXXcx8,316
+pip/_vendor/idna/core.py,sha256=YJYyAMnwiQEPjVC4-Fqu_p4CJ6yKKuDGmppBNQNQpFs,13239
+pip/_vendor/idna/idnadata.py,sha256=W30GcIGvtOWYwAjZj4ZjuouUutC6ffgNuyjJy7fZ-lo,78306
+pip/_vendor/idna/intranges.py,sha256=amUtkdhYcQG8Zr-CoMM_kVRacxkivC1WgxN1b63KKdU,1898
+pip/_vendor/idna/package_data.py,sha256=q59S3OXsc5VI8j6vSD0sGBMyk6zZ4vWFREE88yCJYKs,21
+pip/_vendor/idna/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/idna/uts46data.py,sha256=rt90K9J40gUSwppDPCrhjgi5AA6pWM65dEGRSf6rIhM,239289
+pip/_vendor/msgpack/__init__.py,sha256=reRaiOtEzSjPnr7TpxjgIvbfln5pV66FhricAs2eC-g,1109
+pip/_vendor/msgpack/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/msgpack/__pycache__/exceptions.cpython-312.pyc,,
+pip/_vendor/msgpack/__pycache__/ext.cpython-312.pyc,,
+pip/_vendor/msgpack/__pycache__/fallback.cpython-312.pyc,,
+pip/_vendor/msgpack/exceptions.py,sha256=dCTWei8dpkrMsQDcjQk74ATl9HsIBH0ybt8zOPNqMYc,1081
+pip/_vendor/msgpack/ext.py,sha256=kteJv03n9tYzd5oo3xYopVTo4vRaAxonBQQJhXohZZo,5726
+pip/_vendor/msgpack/fallback.py,sha256=0g1Pzp0vtmBEmJ5w9F3s_-JMVURP8RS4G1cc5TRaAsI,32390
+pip/_vendor/packaging/__init__.py,sha256=dk4Ta_vmdVJxYHDcfyhvQNw8V3PgSBomKNXqg-D2JDY,494
+pip/_vendor/packaging/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_parser.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_structures.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/markers.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/metadata.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/requirements.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/tags.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/utils.cpython-312.pyc,,
+pip/_vendor/packaging/__pycache__/version.cpython-312.pyc,,
+pip/_vendor/packaging/_elffile.py,sha256=cflAQAkE25tzhYmq_aCi72QfbT_tn891tPzfpbeHOwE,3306
+pip/_vendor/packaging/_manylinux.py,sha256=vl5OCoz4kx80H5rwXKeXWjl9WNISGmr4ZgTpTP9lU9c,9612
+pip/_vendor/packaging/_musllinux.py,sha256=p9ZqNYiOItGee8KcZFeHF_YcdhVwGHdK6r-8lgixvGQ,2694
+pip/_vendor/packaging/_parser.py,sha256=s_TvTvDNK0NrM2QB3VKThdWFM4Nc0P6JnkObkl3MjpM,10236
+pip/_vendor/packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+pip/_vendor/packaging/_tokenizer.py,sha256=J6v5H7Jzvb-g81xp_2QACKwO7LxHQA6ikryMU7zXwN8,5273
+pip/_vendor/packaging/licenses/__init__.py,sha256=A116-FU49_Dz4162M4y1uAiZN4Rgdc83FxNd8EjlfqI,5727
+pip/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc,,
+pip/_vendor/packaging/licenses/_spdx.py,sha256=oAm1ztPFwlsmCKe7lAAsv_OIOfS1cWDu9bNBkeu-2ns,48398
+pip/_vendor/packaging/markers.py,sha256=c89TNzB7ZdGYhkovm6PYmqGyHxXlYVaLW591PHUNKD8,10561
+pip/_vendor/packaging/metadata.py,sha256=YJibM7GYe4re8-0a3OlXmGS-XDgTEoO4tlBt2q25Bng,34762
+pip/_vendor/packaging/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/packaging/requirements.py,sha256=gYyRSAdbrIyKDY66ugIDUQjRMvxkH2ALioTmX3tnL6o,2947
+pip/_vendor/packaging/specifiers.py,sha256=hGU6kuCd77bL-msIL6yLCp6MNT75RSMUKZDuju26c8U,40098
+pip/_vendor/packaging/tags.py,sha256=CFqrJzAzc2XNGexerH__T-Y5Iwq7WbsYXsiLERLWxY0,21014
+pip/_vendor/packaging/utils.py,sha256=0F3Hh9OFuRgrhTgGZUl5K22Fv1YP2tZl1z_2gO6kJiA,5050
+pip/_vendor/packaging/version.py,sha256=oiHqzTUv_p12hpjgsLDVcaF5hT7pDaSOViUNMD4GTW0,16688
+pip/_vendor/pkg_resources/__init__.py,sha256=jrhDRbOubP74QuPXxd7U7Po42PH2l-LZ2XfcO7llpZ4,124463
+pip/_vendor/pkg_resources/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/platformdirs/__init__.py,sha256=JueR2cRLkxY7iwik-qNWJCwKOrAlBgVgcZ_IHQzqGLE,22344
+pip/_vendor/platformdirs/__main__.py,sha256=jBJ8zb7Mpx5ebcqF83xrpO94MaeCpNGHVf9cvDN2JLg,1505
+pip/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/android.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/api.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/version.cpython-312.pyc,,
+pip/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc,,
+pip/_vendor/platformdirs/android.py,sha256=kV5oL3V3DZ6WZKu9yFiQupv18yp_jlSV2ChH1TmPcds,9007
+pip/_vendor/platformdirs/api.py,sha256=2dfUDNbEXeDhDKarqtR5NY7oUikUZ4RZhs3ozstmhBQ,9246
+pip/_vendor/platformdirs/macos.py,sha256=UlbyFZ8Rzu3xndCqQEHrfsYTeHwYdFap1Ioz-yxveT4,6154
+pip/_vendor/platformdirs/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/platformdirs/unix.py,sha256=uRPJWRyQEtv7yOSvU94rUmsblo5XKDLA1SzFg55kbK0,10393
+pip/_vendor/platformdirs/version.py,sha256=oH4KgTfK4AklbTYVcV_yynvJ9JLI3pyvDVay0hRsLCs,411
+pip/_vendor/platformdirs/windows.py,sha256=IFpiohUBwxPtCzlyKwNtxyW4Jk8haa6W8o59mfrDXVo,10125
+pip/_vendor/pygments/__init__.py,sha256=7N1oiaWulw_nCsTY4EEixYLz15pWY5u4uPAFFi-ielU,2983
+pip/_vendor/pygments/__main__.py,sha256=isIhBxLg65nLlXukG4VkMuPfNdd7gFzTZ_R_z3Q8diY,353
+pip/_vendor/pygments/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/__main__.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/cmdline.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/console.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/filter.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/formatter.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/lexer.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/modeline.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/plugin.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/regexopt.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/scanner.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/sphinxext.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/style.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/token.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/unistring.cpython-312.pyc,,
+pip/_vendor/pygments/__pycache__/util.cpython-312.pyc,,
+pip/_vendor/pygments/cmdline.py,sha256=LIVzmAunlk9sRJJp54O4KRy9GDIN4Wu13v9p9QzfGPM,23656
+pip/_vendor/pygments/console.py,sha256=yhP9UsLAVmWKVQf2446JJewkA7AiXeeTf4Ieg3Oi2fU,1718
+pip/_vendor/pygments/filter.py,sha256=_ADNPCskD8_GmodHi6_LoVgPU3Zh336aBCT5cOeTMs0,1910
+pip/_vendor/pygments/filters/__init__.py,sha256=RdedK2KWKXlKwR7cvkfr3NUj9YiZQgMgilRMFUg2jPA,40392
+pip/_vendor/pygments/filters/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pygments/formatter.py,sha256=jDWBTndlBH2Z5IYZFVDnP0qn1CaTQjTWt7iAGtCnJEg,4390
+pip/_vendor/pygments/formatters/__init__.py,sha256=8No-NUs8rBTSSBJIv4hSEQt2M0cFB4hwAT0snVc2QGE,5385
+pip/_vendor/pygments/formatters/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/_mapping.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/bbcode.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/groff.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/html.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/img.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/irc.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/latex.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/other.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/pangomarkup.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/rtf.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/svg.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/terminal.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/__pycache__/terminal256.cpython-312.pyc,,
+pip/_vendor/pygments/formatters/_mapping.py,sha256=1Cw37FuQlNacnxRKmtlPX4nyLoX9_ttko5ZwscNUZZ4,4176
+pip/_vendor/pygments/formatters/bbcode.py,sha256=3JQLI45tcrQ_kRUMjuab6C7Hb0XUsbVWqqbSn9cMjkI,3320
+pip/_vendor/pygments/formatters/groff.py,sha256=M39k0PaSSZRnxWjqBSVPkF0mu1-Vr7bm6RsFvs-CNN4,5106
+pip/_vendor/pygments/formatters/html.py,sha256=SE2jc3YCqbMS3rZW9EAmDlAUhdVxJ52gA4dileEvCGU,35669
+pip/_vendor/pygments/formatters/img.py,sha256=MwA4xWPLOwh6j7Yc6oHzjuqSPt0M1fh5r-5BTIIUfsU,23287
+pip/_vendor/pygments/formatters/irc.py,sha256=dp1Z0l_ObJ5NFh9MhqLGg5ptG5hgJqedT2Vkutt9v0M,4981
+pip/_vendor/pygments/formatters/latex.py,sha256=XMmhOCqUKDBQtG5mGJNAFYxApqaC5puo5cMmPfK3944,19306
+pip/_vendor/pygments/formatters/other.py,sha256=56PMJOliin-rAUdnRM0i1wsV1GdUPd_dvQq0_UPfF9c,5034
+pip/_vendor/pygments/formatters/pangomarkup.py,sha256=y16U00aVYYEFpeCfGXlYBSMacG425CbfoG8oKbKegIg,2218
+pip/_vendor/pygments/formatters/rtf.py,sha256=ZT90dmcKyJboIB0mArhL7IhE467GXRN0G7QAUgG03To,11957
+pip/_vendor/pygments/formatters/svg.py,sha256=KKsiophPupHuxm0So-MsbQEWOT54IAiSF7hZPmxtKXE,7174
+pip/_vendor/pygments/formatters/terminal.py,sha256=AojNG4MlKq2L6IsC_VnXHu4AbHCBn9Otog6u45XvxeI,4674
+pip/_vendor/pygments/formatters/terminal256.py,sha256=kGkNUVo3FpwjytIDS0if79EuUoroAprcWt3igrcIqT0,11753
+pip/_vendor/pygments/lexer.py,sha256=TYHDt___gNW4axTl2zvPZff-VQi8fPaIh5OKRcVSjUM,35349
+pip/_vendor/pygments/lexers/__init__.py,sha256=pIlxyQJuu_syh9lE080cq8ceVbEVcKp0osAFU5fawJU,12115
+pip/_vendor/pygments/lexers/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pygments/lexers/__pycache__/_mapping.cpython-312.pyc,,
+pip/_vendor/pygments/lexers/__pycache__/python.cpython-312.pyc,,
+pip/_vendor/pygments/lexers/_mapping.py,sha256=61-h3zr103m01OS5BUq_AfUiL9YI06Ves9ipQ7k4vr4,76097
+pip/_vendor/pygments/lexers/python.py,sha256=2J_YJrPTr_A6fJY_qKiKv0GpgPwHMrlMSeo59qN3fe4,53687
+pip/_vendor/pygments/modeline.py,sha256=gtRYZBS-CKOCDXHhGZqApboHBaZwGH8gznN3O6nuxj4,1005
+pip/_vendor/pygments/plugin.py,sha256=ioeJ3QeoJ-UQhZpY9JL7vbxsTVuwwM7BCu-Jb8nN0AU,1891
+pip/_vendor/pygments/regexopt.py,sha256=Hky4EB13rIXEHQUNkwmCrYqtIlnXDehNR3MztafZ43w,3072
+pip/_vendor/pygments/scanner.py,sha256=NDy3ofK_fHRFK4hIDvxpamG871aewqcsIb6sgTi7Fhk,3092
+pip/_vendor/pygments/sphinxext.py,sha256=iOptJBcqOGPwMEJ2p70PvwpZPIGdvdZ8dxvq6kzxDgA,7981
+pip/_vendor/pygments/style.py,sha256=rSCZWFpg1_DwFMXDU0nEVmAcBHpuQGf9RxvOPPQvKLQ,6420
+pip/_vendor/pygments/styles/__init__.py,sha256=qUk6_1z5KmT8EdJFZYgESmG6P_HJF_2vVrDD7HSCGYY,2042
+pip/_vendor/pygments/styles/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pygments/styles/__pycache__/_mapping.cpython-312.pyc,,
+pip/_vendor/pygments/styles/_mapping.py,sha256=6lovFUE29tz6EsV3XYY4hgozJ7q1JL7cfO3UOlgnS8w,3312
+pip/_vendor/pygments/token.py,sha256=qZwT7LSPy5YBY3JgDjut642CCy7JdQzAfmqD9NmT5j0,6226
+pip/_vendor/pygments/unistring.py,sha256=p5c1i-HhoIhWemy9CUsaN9o39oomYHNxXll0Xfw6tEA,63208
+pip/_vendor/pygments/util.py,sha256=2tj2nS1X9_OpcuSjf8dOET2bDVZhs8cEKd_uT6-Fgg8,10031
+pip/_vendor/pyproject_hooks/__init__.py,sha256=cPB_a9LXz5xvsRbX1o2qyAdjLatZJdQ_Lc5McNX-X7Y,691
+pip/_vendor/pyproject_hooks/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pyproject_hooks/__pycache__/_impl.cpython-312.pyc,,
+pip/_vendor/pyproject_hooks/_impl.py,sha256=jY-raxnmyRyB57ruAitrJRUzEexuAhGTpgMygqx67Z4,14936
+pip/_vendor/pyproject_hooks/_in_process/__init__.py,sha256=MJNPpfIxcO-FghxpBbxkG1rFiQf6HOUbV4U5mq0HFns,557
+pip/_vendor/pyproject_hooks/_in_process/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/pyproject_hooks/_in_process/__pycache__/_in_process.cpython-312.pyc,,
+pip/_vendor/pyproject_hooks/_in_process/_in_process.py,sha256=qcXMhmx__MIJq10gGHW3mA4Tl8dy8YzHMccwnNoKlw0,12216
+pip/_vendor/pyproject_hooks/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/requests/__init__.py,sha256=HlB_HzhrzGtfD_aaYUwUh1zWXLZ75_YCLyit75d0Vz8,5057
+pip/_vendor/requests/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/__version__.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/_internal_utils.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/adapters.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/api.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/auth.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/certs.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/compat.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/cookies.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/exceptions.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/help.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/hooks.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/models.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/packages.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/sessions.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/status_codes.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/structures.cpython-312.pyc,,
+pip/_vendor/requests/__pycache__/utils.cpython-312.pyc,,
+pip/_vendor/requests/__version__.py,sha256=FVfglgZmNQnmYPXpOohDU58F5EUb_-VnSTaAesS187g,435
+pip/_vendor/requests/_internal_utils.py,sha256=nMQymr4hs32TqVo5AbCrmcJEhvPUh7xXlluyqwslLiQ,1495
+pip/_vendor/requests/adapters.py,sha256=J7VeVxKBvawbtlX2DERVo05J9BXTcWYLMHNd1Baa-bk,27607
+pip/_vendor/requests/api.py,sha256=_Zb9Oa7tzVIizTKwFrPjDEY9ejtm_OnSRERnADxGsQs,6449
+pip/_vendor/requests/auth.py,sha256=kF75tqnLctZ9Mf_hm9TZIj4cQWnN5uxRz8oWsx5wmR0,10186
+pip/_vendor/requests/certs.py,sha256=kHDlkK_beuHXeMPc5jta2wgl8gdKeUWt5f2nTDVrvt8,441
+pip/_vendor/requests/compat.py,sha256=Mo9f9xZpefod8Zm-n9_StJcVTmwSukXR2p3IQyyVXvU,1485
+pip/_vendor/requests/cookies.py,sha256=bNi-iqEj4NPZ00-ob-rHvzkvObzN3lEpgw3g6paS3Xw,18590
+pip/_vendor/requests/exceptions.py,sha256=D1wqzYWne1mS2rU43tP9CeN1G7QAy7eqL9o1god6Ejw,4272
+pip/_vendor/requests/help.py,sha256=hRKaf9u0G7fdwrqMHtF3oG16RKktRf6KiwtSq2Fo1_0,3813
+pip/_vendor/requests/hooks.py,sha256=CiuysiHA39V5UfcCBXFIx83IrDpuwfN9RcTUgv28ftQ,733
+pip/_vendor/requests/models.py,sha256=x4K4CmH-lC0l2Kb-iPfMN4dRXxHEcbOaEWBL_i09AwI,35483
+pip/_vendor/requests/packages.py,sha256=_ZQDCJTJ8SP3kVWunSqBsRZNPzj2c1WFVqbdr08pz3U,1057
+pip/_vendor/requests/sessions.py,sha256=ykTI8UWGSltOfH07HKollH7kTBGw4WhiBVaQGmckTw4,30495
+pip/_vendor/requests/status_codes.py,sha256=iJUAeA25baTdw-6PfD0eF4qhpINDJRJI-yaMqxs4LEI,4322
+pip/_vendor/requests/structures.py,sha256=-IbmhVz06S-5aPSZuUthZ6-6D9XOjRuTXHOabY041XM,2912
+pip/_vendor/requests/utils.py,sha256=L79vnFbzJ3SFLKtJwpoWe41Tozi3RlZv94pY1TFIyow,33631
+pip/_vendor/resolvelib/__init__.py,sha256=h509TdEcpb5-44JonaU3ex2TM15GVBLjM9CNCPwnTTs,537
+pip/_vendor/resolvelib/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/resolvelib/__pycache__/providers.cpython-312.pyc,,
+pip/_vendor/resolvelib/__pycache__/reporters.cpython-312.pyc,,
+pip/_vendor/resolvelib/__pycache__/resolvers.cpython-312.pyc,,
+pip/_vendor/resolvelib/__pycache__/structs.cpython-312.pyc,,
+pip/_vendor/resolvelib/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/resolvelib/compat/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/resolvelib/compat/__pycache__/collections_abc.cpython-312.pyc,,
+pip/_vendor/resolvelib/compat/collections_abc.py,sha256=uy8xUZ-NDEw916tugUXm8HgwCGiMO0f-RcdnpkfXfOs,156
+pip/_vendor/resolvelib/providers.py,sha256=fuuvVrCetu5gsxPB43ERyjfO8aReS3rFQHpDgiItbs4,5871
+pip/_vendor/resolvelib/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/resolvelib/reporters.py,sha256=TSbRmWzTc26w0ggsV1bxVpeWDB8QNIre6twYl7GIZBE,1601
+pip/_vendor/resolvelib/resolvers.py,sha256=G8rsLZSq64g5VmIq-lB7UcIJ1gjAxIQJmTF4REZleQ0,20511
+pip/_vendor/resolvelib/structs.py,sha256=0_1_XO8z_CLhegP3Vpf9VJ3zJcfLm0NOHRM-i0Ykz3o,4963
+pip/_vendor/rich/__init__.py,sha256=dRxjIL-SbFVY0q3IjSMrfgBTHrm1LZDgLOygVBwiYZc,6090
+pip/_vendor/rich/__main__.py,sha256=eO7Cq8JnrgG8zVoeImiAs92q3hXNMIfp0w5lMsO7Q2Y,8477
+pip/_vendor/rich/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/__main__.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_cell_widths.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_emoji_codes.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_emoji_replace.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_export_format.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_extension.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_fileno.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_inspect.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_log_render.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_loop.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_null_file.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_palettes.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_pick.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_ratio.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_spinners.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_stack.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_timer.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_win32_console.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_windows.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_windows_renderer.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/_wrap.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/abc.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/align.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/ansi.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/bar.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/box.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/cells.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/color.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/color_triplet.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/columns.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/console.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/constrain.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/containers.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/control.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/default_styles.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/diagnose.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/emoji.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/errors.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/file_proxy.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/filesize.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/highlighter.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/json.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/jupyter.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/layout.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/live.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/live_render.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/logging.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/markup.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/measure.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/padding.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/pager.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/palette.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/panel.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/pretty.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/progress.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/progress_bar.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/prompt.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/protocol.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/region.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/repr.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/rule.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/scope.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/screen.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/segment.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/spinner.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/status.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/style.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/styled.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/syntax.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/table.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/terminal_theme.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/text.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/theme.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/themes.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/traceback.cpython-312.pyc,,
+pip/_vendor/rich/__pycache__/tree.cpython-312.pyc,,
+pip/_vendor/rich/_cell_widths.py,sha256=fbmeyetEdHjzE_Vx2l1uK7tnPOhMs2X1lJfO3vsKDpA,10209
+pip/_vendor/rich/_emoji_codes.py,sha256=hu1VL9nbVdppJrVoijVshRlcRRe_v3dju3Mmd2sKZdY,140235
+pip/_vendor/rich/_emoji_replace.py,sha256=n-kcetsEUx2ZUmhQrfeMNc-teeGhpuSQ5F8VPBsyvDo,1064
+pip/_vendor/rich/_export_format.py,sha256=RI08pSrm5tBSzPMvnbTqbD9WIalaOoN5d4M1RTmLq1Y,2128
+pip/_vendor/rich/_extension.py,sha256=Xt47QacCKwYruzjDi-gOBq724JReDj9Cm9xUi5fr-34,265
+pip/_vendor/rich/_fileno.py,sha256=HWZxP5C2ajMbHryvAQZseflVfQoGzsKOHzKGsLD8ynQ,799
+pip/_vendor/rich/_inspect.py,sha256=QM05lEFnFoTaFqpnbx-zBEI6k8oIKrD3cvjEOQNhKig,9655
+pip/_vendor/rich/_log_render.py,sha256=1ByI0PA1ZpxZY3CGJOK54hjlq4X-Bz_boIjIqCd8Kns,3225
+pip/_vendor/rich/_loop.py,sha256=hV_6CLdoPm0va22Wpw4zKqM0RYsz3TZxXj0PoS-9eDQ,1236
+pip/_vendor/rich/_null_file.py,sha256=ADGKp1yt-k70FMKV6tnqCqecB-rSJzp-WQsD7LPL-kg,1394
+pip/_vendor/rich/_palettes.py,sha256=cdev1JQKZ0JvlguV9ipHgznTdnvlIzUFDBb0It2PzjI,7063
+pip/_vendor/rich/_pick.py,sha256=evDt8QN4lF5CiwrUIXlOJCntitBCOsI3ZLPEIAVRLJU,423
+pip/_vendor/rich/_ratio.py,sha256=Zt58apszI6hAAcXPpgdWKpu3c31UBWebOeR4mbyptvU,5471
+pip/_vendor/rich/_spinners.py,sha256=U2r1_g_1zSjsjiUdAESc2iAMc3i4ri_S8PYP6kQ5z1I,19919
+pip/_vendor/rich/_stack.py,sha256=-C8OK7rxn3sIUdVwxZBBpeHhIzX0eI-VM3MemYfaXm0,351
+pip/_vendor/rich/_timer.py,sha256=zelxbT6oPFZnNrwWPpc1ktUeAT-Vc4fuFcRZLQGLtMI,417
+pip/_vendor/rich/_win32_console.py,sha256=BSaDRIMwBLITn_m0mTRLPqME5q-quGdSMuYMpYeYJwc,22755
+pip/_vendor/rich/_windows.py,sha256=aBwaD_S56SbgopIvayVmpk0Y28uwY2C5Bab1wl3Bp-I,1925
+pip/_vendor/rich/_windows_renderer.py,sha256=t74ZL3xuDCP3nmTp9pH1L5LiI2cakJuQRQleHCJerlk,2783
+pip/_vendor/rich/_wrap.py,sha256=FlSsom5EX0LVkA3KWy34yHnCfLtqX-ZIepXKh-70rpc,3404
+pip/_vendor/rich/abc.py,sha256=ON-E-ZqSSheZ88VrKX2M3PXpFbGEUUZPMa_Af0l-4f0,890
+pip/_vendor/rich/align.py,sha256=Rh-3adnDaN1Ao07EjR2PhgE62PGLPgO8SMwJBku1urQ,10469
+pip/_vendor/rich/ansi.py,sha256=Avs1LHbSdcyOvDOdpELZUoULcBiYewY76eNBp6uFBhs,6921
+pip/_vendor/rich/bar.py,sha256=ldbVHOzKJOnflVNuv1xS7g6dLX2E3wMnXkdPbpzJTcs,3263
+pip/_vendor/rich/box.py,sha256=nr5fYIUghB_iUCEq6y0Z3LlCT8gFPDrzN9u2kn7tJl4,10831
+pip/_vendor/rich/cells.py,sha256=KrQkj5-LghCCpJLSNQIyAZjndc4bnEqOEmi5YuZ9UCY,5130
+pip/_vendor/rich/color.py,sha256=3HSULVDj7qQkXUdFWv78JOiSZzfy5y1nkcYhna296V0,18211
+pip/_vendor/rich/color_triplet.py,sha256=3lhQkdJbvWPoLDO-AnYImAWmJvV5dlgYNCVZ97ORaN4,1054
+pip/_vendor/rich/columns.py,sha256=HUX0KcMm9dsKNi11fTbiM_h2iDtl8ySCaVcxlalEzq8,7131
+pip/_vendor/rich/console.py,sha256=nKjrEx_7xy8KGmDVT-BgNII0R5hm1cexhAHDwdwNVqg,100156
+pip/_vendor/rich/constrain.py,sha256=1VIPuC8AgtKWrcncQrjBdYqA3JVWysu6jZo1rrh7c7Q,1288
+pip/_vendor/rich/containers.py,sha256=c_56TxcedGYqDepHBMTuZdUIijitAQgnox-Qde0Z1qo,5502
+pip/_vendor/rich/control.py,sha256=DSkHTUQLorfSERAKE_oTAEUFefZnZp4bQb4q8rHbKws,6630
+pip/_vendor/rich/default_styles.py,sha256=dZxgaSD9VUy7SXQShO33aLYiAWspCr2sCQZFX_JK1j4,8159
+pip/_vendor/rich/diagnose.py,sha256=an6uouwhKPAlvQhYpNNpGq9EJysfMIOvvCbO3oSoR24,972
+pip/_vendor/rich/emoji.py,sha256=omTF9asaAnsM4yLY94eR_9dgRRSm1lHUszX20D1yYCQ,2501
+pip/_vendor/rich/errors.py,sha256=5pP3Kc5d4QJ_c0KFsxrfyhjiPVe7J1zOqSFbFAzcV-Y,642
+pip/_vendor/rich/file_proxy.py,sha256=Tl9THMDZ-Pk5Wm8sI1gGg_U5DhusmxD-FZ0fUbcU0W0,1683
+pip/_vendor/rich/filesize.py,sha256=_iz9lIpRgvW7MNSeCZnLg-HwzbP4GETg543WqD8SFs0,2484
+pip/_vendor/rich/highlighter.py,sha256=G_sn-8DKjM1sEjLG_oc4ovkWmiUpWvj8bXi0yed2LnY,9586
+pip/_vendor/rich/json.py,sha256=vVEoKdawoJRjAFayPwXkMBPLy7RSTs-f44wSQDR2nJ0,5031
+pip/_vendor/rich/jupyter.py,sha256=QyoKoE_8IdCbrtiSHp9TsTSNyTHY0FO5whE7jOTd9UE,3252
+pip/_vendor/rich/layout.py,sha256=ajkSFAtEVv9EFTcFs-w4uZfft7nEXhNzL7ZVdgrT5rI,14004
+pip/_vendor/rich/live.py,sha256=DhzAPEnjTxQuq9_0Y2xh2MUwQcP_aGPkenLfKETslwM,14270
+pip/_vendor/rich/live_render.py,sha256=zJtB471jGziBtEwxc54x12wEQtH4BuQr1SA8v9kU82w,3666
+pip/_vendor/rich/logging.py,sha256=ZgpKMMBY_BuMAI_BYzo-UtXak6t5oH9VK8m9Q2Lm0f4,12458
+pip/_vendor/rich/markup.py,sha256=3euGKP5s41NCQwaSjTnJxus5iZMHjxpIM0W6fCxra38,8451
+pip/_vendor/rich/measure.py,sha256=HmrIJX8sWRTHbgh8MxEay_83VkqNW_70s8aKP5ZcYI8,5305
+pip/_vendor/rich/padding.py,sha256=KVEI3tOwo9sgK1YNSuH__M1_jUWmLZwRVV_KmOtVzyM,4908
+pip/_vendor/rich/pager.py,sha256=SO_ETBFKbg3n_AgOzXm41Sv36YxXAyI3_R-KOY2_uSc,828
+pip/_vendor/rich/palette.py,sha256=lInvR1ODDT2f3UZMfL1grq7dY_pDdKHw4bdUgOGaM4Y,3396
+pip/_vendor/rich/panel.py,sha256=fFRHcviXvWhk3V3zx5Zwmsb_RL9KJ3esD-sU0NYEVyw,11235
+pip/_vendor/rich/pretty.py,sha256=gy3S72u4FRg2ytoo7N1ZDWDIvB4unbzd5iUGdgm-8fc,36391
+pip/_vendor/rich/progress.py,sha256=MtmCjTk5zYU_XtRHxRHTAEHG6hF9PeF7EMWbEPleIC0,60357
+pip/_vendor/rich/progress_bar.py,sha256=mZTPpJUwcfcdgQCTTz3kyY-fc79ddLwtx6Ghhxfo064,8162
+pip/_vendor/rich/prompt.py,sha256=l0RhQU-0UVTV9e08xW1BbIj0Jq2IXyChX4lC0lFNzt4,12447
+pip/_vendor/rich/protocol.py,sha256=5hHHDDNHckdk8iWH5zEbi-zuIVSF5hbU2jIo47R7lTE,1391
+pip/_vendor/rich/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/rich/region.py,sha256=rNT9xZrVZTYIXZC0NYn41CJQwYNbR-KecPOxTgQvB8Y,166
+pip/_vendor/rich/repr.py,sha256=5MZJZmONgC6kud-QW-_m1okXwL2aR6u6y-pUcUCJz28,4431
+pip/_vendor/rich/rule.py,sha256=0fNaS_aERa3UMRc3T5WMpN_sumtDxfaor2y3of1ftBk,4602
+pip/_vendor/rich/scope.py,sha256=TMUU8qo17thyqQCPqjDLYpg_UU1k5qVd-WwiJvnJVas,2843
+pip/_vendor/rich/screen.py,sha256=YoeReESUhx74grqb0mSSb9lghhysWmFHYhsbMVQjXO8,1591
+pip/_vendor/rich/segment.py,sha256=otnKeKGEV-WRlQVosfJVeFDcDxAKHpvJ_hLzSu5lumM,24743
+pip/_vendor/rich/spinner.py,sha256=PT5qgXPG3ZpqRj7n3EZQ6NW56mx3ldZqZCU7gEMyZk4,4364
+pip/_vendor/rich/status.py,sha256=kkPph3YeAZBo-X-4wPp8gTqZyU466NLwZBA4PZTTewo,4424
+pip/_vendor/rich/style.py,sha256=aSoUNbVgfP1PAnduAqgbbl4AMQy668qs2S1FEwr3Oqs,27067
+pip/_vendor/rich/styled.py,sha256=eZNnzGrI4ki_54pgY3Oj0T-x3lxdXTYh4_ryDB24wBU,1258
+pip/_vendor/rich/syntax.py,sha256=qqAnEUZ4K57Po81_5RBxnsuU4KRzSdvDPAhKw8ma_3E,35763
+pip/_vendor/rich/table.py,sha256=yXYUr0YsPpG466N50HCAw2bpb5ZUuuzdc-G66Zk-oTc,40103
+pip/_vendor/rich/terminal_theme.py,sha256=1j5-ufJfnvlAo5Qsi_ACZiXDmwMXzqgmFByObT9-yJY,3370
+pip/_vendor/rich/text.py,sha256=AO7JPCz6-gaN1thVLXMBntEmDPVYFgFNG1oM61_sanU,47552
+pip/_vendor/rich/theme.py,sha256=oNyhXhGagtDlbDye3tVu3esWOWk0vNkuxFw-_unlaK0,3771
+pip/_vendor/rich/themes.py,sha256=0xgTLozfabebYtcJtDdC5QkX5IVUEaviqDUJJh4YVFk,102
+pip/_vendor/rich/traceback.py,sha256=z8UoN7NbTQKW6YDDUVwOh7F8snZf6gYnUWtOrKsLE1w,31797
+pip/_vendor/rich/tree.py,sha256=yWnQ6rAvRGJ3qZGqBrxS2SW2TKBTNrP0SdY8QxOFPuw,9451
+pip/_vendor/tomli/__init__.py,sha256=PhNw_eyLgdn7McJ6nrAN8yIm3dXC75vr1sVGVVwDSpA,314
+pip/_vendor/tomli/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/tomli/__pycache__/_parser.cpython-312.pyc,,
+pip/_vendor/tomli/__pycache__/_re.cpython-312.pyc,,
+pip/_vendor/tomli/__pycache__/_types.cpython-312.pyc,,
+pip/_vendor/tomli/_parser.py,sha256=9w8LG0jB7fwmZZWB0vVXbeejDHcl4ANIJxB2scEnDlA,25591
+pip/_vendor/tomli/_re.py,sha256=sh4sBDRgO94KJZwNIrgdcyV_qQast50YvzOAUGpRDKA,3171
+pip/_vendor/tomli/_types.py,sha256=-GTG2VUqkpxwMqzmVO4F7ybKddIbAnuAHXfmWQcTi3Q,254
+pip/_vendor/tomli/py.typed,sha256=8PjyZ1aVoQpRVvt71muvuq5qE-jTFZkK-GLHkhdebmc,26
+pip/_vendor/truststore/__init__.py,sha256=WIDeyzWm7EVX44g354M25vpRXbeY1lsPH6EmUJUcq4o,1264
+pip/_vendor/truststore/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/truststore/__pycache__/_api.cpython-312.pyc,,
+pip/_vendor/truststore/__pycache__/_macos.cpython-312.pyc,,
+pip/_vendor/truststore/__pycache__/_openssl.cpython-312.pyc,,
+pip/_vendor/truststore/__pycache__/_ssl_constants.cpython-312.pyc,,
+pip/_vendor/truststore/__pycache__/_windows.cpython-312.pyc,,
+pip/_vendor/truststore/_api.py,sha256=GeXRNTlxPZ3kif4kNoh6JY0oE4QRzTGcgXr6l_X_Gk0,10555
+pip/_vendor/truststore/_macos.py,sha256=nZlLkOmszUE0g6ryRwBVGY5COzPyudcsiJtDWarM5LQ,20503
+pip/_vendor/truststore/_openssl.py,sha256=LLUZ7ZGaio-i5dpKKjKCSeSufmn6T8pi9lDcFnvSyq0,2324
+pip/_vendor/truststore/_ssl_constants.py,sha256=NUD4fVKdSD02ri7-db0tnO0VqLP9aHuzmStcW7tAl08,1130
+pip/_vendor/truststore/_windows.py,sha256=rAHyKYD8M7t-bXfG8VgOVa3TpfhVhbt4rZQlO45YuP8,17993
+pip/_vendor/truststore/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/typing_extensions.py,sha256=78hFl0HpDY-ylHUVCnWdU5nTHxUP2-S-3wEZk6CQmLk,134499
+pip/_vendor/urllib3/__init__.py,sha256=iXLcYiJySn0GNbWOOZDDApgBL1JgP44EZ8i1760S8Mc,3333
+pip/_vendor/urllib3/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/_collections.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/_version.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/connection.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/connectionpool.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/exceptions.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/fields.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/filepost.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/poolmanager.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/request.cpython-312.pyc,,
+pip/_vendor/urllib3/__pycache__/response.cpython-312.pyc,,
+pip/_vendor/urllib3/_collections.py,sha256=pyASJJhW7wdOpqJj9QJA8FyGRfr8E8uUUhqUvhF0728,11372
+pip/_vendor/urllib3/_version.py,sha256=t9wGB6ooOTXXgiY66K1m6BZS1CJyXHAU8EoWDTe6Shk,64
+pip/_vendor/urllib3/connection.py,sha256=ttIA909BrbTUzwkqEe_TzZVh4JOOj7g61Ysei2mrwGg,20314
+pip/_vendor/urllib3/connectionpool.py,sha256=e2eiAwNbFNCKxj4bwDKNK-w7HIdSz3OmMxU_TIt-evQ,40408
+pip/_vendor/urllib3/contrib/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/urllib3/contrib/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/_appengine_environ.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/appengine.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/ntlmpool.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/securetransport.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/__pycache__/socks.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/_appengine_environ.py,sha256=bDbyOEhW2CKLJcQqAKAyrEHN-aklsyHFKq6vF8ZFsmk,957
+pip/_vendor/urllib3/contrib/_securetransport/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/urllib3/contrib/_securetransport/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/_securetransport/__pycache__/bindings.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/_securetransport/__pycache__/low_level.cpython-312.pyc,,
+pip/_vendor/urllib3/contrib/_securetransport/bindings.py,sha256=4Xk64qIkPBt09A5q-RIFUuDhNc9mXilVapm7WnYnzRw,17632
+pip/_vendor/urllib3/contrib/_securetransport/low_level.py,sha256=B2JBB2_NRP02xK6DCa1Pa9IuxrPwxzDzZbixQkb7U9M,13922
+pip/_vendor/urllib3/contrib/appengine.py,sha256=VR68eAVE137lxTgjBDwCna5UiBZTOKa01Aj_-5BaCz4,11036
+pip/_vendor/urllib3/contrib/ntlmpool.py,sha256=NlfkW7WMdW8ziqudopjHoW299og1BTWi0IeIibquFwk,4528
+pip/_vendor/urllib3/contrib/pyopenssl.py,sha256=hDJh4MhyY_p-oKlFcYcQaVQRDv6GMmBGuW9yjxyeejM,17081
+pip/_vendor/urllib3/contrib/securetransport.py,sha256=Fef1IIUUFHqpevzXiDPbIGkDKchY2FVKeVeLGR1Qq3g,34446
+pip/_vendor/urllib3/contrib/socks.py,sha256=aRi9eWXo9ZEb95XUxef4Z21CFlnnjbEiAo9HOseoMt4,7097
+pip/_vendor/urllib3/exceptions.py,sha256=0Mnno3KHTNfXRfY7638NufOPkUb6mXOm-Lqj-4x2w8A,8217
+pip/_vendor/urllib3/fields.py,sha256=kvLDCg_JmH1lLjUUEY_FLS8UhY7hBvDPuVETbY8mdrM,8579
+pip/_vendor/urllib3/filepost.py,sha256=5b_qqgRHVlL7uLtdAYBzBh-GHmU5AfJVt_2N0XS3PeY,2440
+pip/_vendor/urllib3/packages/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/urllib3/packages/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/packages/__pycache__/six.cpython-312.pyc,,
+pip/_vendor/urllib3/packages/backports/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pip/_vendor/urllib3/packages/backports/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/packages/backports/__pycache__/makefile.cpython-312.pyc,,
+pip/_vendor/urllib3/packages/backports/__pycache__/weakref_finalize.cpython-312.pyc,,
+pip/_vendor/urllib3/packages/backports/makefile.py,sha256=nbzt3i0agPVP07jqqgjhaYjMmuAi_W5E0EywZivVO8E,1417
+pip/_vendor/urllib3/packages/backports/weakref_finalize.py,sha256=tRCal5OAhNSRyb0DhHp-38AtIlCsRP8BxF3NX-6rqIA,5343
+pip/_vendor/urllib3/packages/six.py,sha256=b9LM0wBXv7E7SrbCjAm4wwN-hrH-iNxv18LgWNMMKPo,34665
+pip/_vendor/urllib3/poolmanager.py,sha256=aWyhXRtNO4JUnCSVVqKTKQd8EXTvUm1VN9pgs2bcONo,19990
+pip/_vendor/urllib3/request.py,sha256=YTWFNr7QIwh7E1W9dde9LM77v2VWTJ5V78XuTTw7D1A,6691
+pip/_vendor/urllib3/response.py,sha256=fmDJAFkG71uFTn-sVSTh2Iw0WmcXQYqkbRjihvwBjU8,30641
+pip/_vendor/urllib3/util/__init__.py,sha256=JEmSmmqqLyaw8P51gUImZh8Gwg9i1zSe-DoqAitn2nc,1155
+pip/_vendor/urllib3/util/__pycache__/__init__.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/connection.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/proxy.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/queue.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/request.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/response.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/retry.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/ssl_.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/ssltransport.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/timeout.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/url.cpython-312.pyc,,
+pip/_vendor/urllib3/util/__pycache__/wait.cpython-312.pyc,,
+pip/_vendor/urllib3/util/connection.py,sha256=5Lx2B1PW29KxBn2T0xkN1CBgRBa3gGVJBKoQoRogEVk,4901
+pip/_vendor/urllib3/util/proxy.py,sha256=zUvPPCJrp6dOF0N4GAVbOcl6o-4uXKSrGiTkkr5vUS4,1605
+pip/_vendor/urllib3/util/queue.py,sha256=nRgX8_eX-_VkvxoX096QWoz8Ps0QHUAExILCY_7PncM,498
+pip/_vendor/urllib3/util/request.py,sha256=C0OUt2tcU6LRiQJ7YYNP9GvPrSvl7ziIBekQ-5nlBZk,3997
+pip/_vendor/urllib3/util/response.py,sha256=GJpg3Egi9qaJXRwBh5wv-MNuRWan5BIu40oReoxWP28,3510
+pip/_vendor/urllib3/util/retry.py,sha256=6ENvOZ8PBDzh8kgixpql9lIrb2dxH-k7ZmBanJF2Ng4,22050
+pip/_vendor/urllib3/util/ssl_.py,sha256=QDuuTxPSCj1rYtZ4xpD7Ux-r20TD50aHyqKyhQ7Bq4A,17460
+pip/_vendor/urllib3/util/ssl_match_hostname.py,sha256=Ir4cZVEjmAk8gUAIHWSi7wtOO83UCYABY2xFD1Ql_WA,5758
+pip/_vendor/urllib3/util/ssltransport.py,sha256=NA-u5rMTrDFDFC8QzRKUEKMG0561hOD4qBTr3Z4pv6E,6895
+pip/_vendor/urllib3/util/timeout.py,sha256=cwq4dMk87mJHSBktK1miYJ-85G-3T3RmT20v7SFCpno,10168
+pip/_vendor/urllib3/util/url.py,sha256=lCAE7M5myA8EDdW0sJuyyZhVB9K_j38ljWhHAnFaWoE,14296
+pip/_vendor/urllib3/util/wait.py,sha256=fOX0_faozG2P7iVojQoE1mbydweNyTcm-hXEfFrTtLI,5403
+pip/_vendor/vendor.txt,sha256=EW-E3cE5XEAtVFzGInikArOMDxGP0DLUWzXpY4RZfFY,333
+pip/py.typed,sha256=EBVvvPRTn_eIpz5e5QztSCdrMX7Qwd7VP93RSoIlZ2I,286
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/WHEEL
new file mode 100644
index 00000000..505164bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/entry_points.txt
new file mode 100644
index 00000000..25fcf7e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/entry_points.txt
@@ -0,0 +1,3 @@
+[console_scripts]
+pip = pip._internal.cli.main:main
+pip3 = pip._internal.cli.main:main
diff --git a/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/top_level.txt
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip-25.0.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/__init__.py
new file mode 100644
index 00000000..d628f93e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/__init__.py
@@ -0,0 +1,13 @@
+from typing import List, Optional
+
+__version__ = "25.0.1"
+
+
+def main(args: Optional[List[str]] = None) -> int:
+    """This is an internal API only meant for use by pip's own console scripts.
+
+    For additional details, see https://github.com/pypa/pip/issues/7498.
+    """
+    from pip._internal.utils.entrypoints import _wrapper
+
+    return _wrapper(args)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/__main__.py
new file mode 100644
index 00000000..59913261
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/__main__.py
@@ -0,0 +1,24 @@
+import os
+import sys
+
+# Remove '' and current working directory from the first entry
+# of sys.path, if present to avoid using current directory
+# in pip commands check, freeze, install, list and show,
+# when invoked as python -m pip <command>
+if sys.path[0] in ("", os.getcwd()):
+    sys.path.pop(0)
+
+# If we are running from a wheel, add the wheel to sys.path
+# This allows the usage python pip-*.whl/pip install pip-*.whl
+if __package__ == "":
+    # __file__ is pip-*.whl/pip/__main__.py
+    # first dirname call strips of '/__main__.py', second strips off '/pip'
+    # Resulting path is the name of the wheel itself
+    # Add that to sys.path so we can import pip
+    path = os.path.dirname(os.path.dirname(__file__))
+    sys.path.insert(0, path)
+
+if __name__ == "__main__":
+    from pip._internal.cli.main import main as _main
+
+    sys.exit(_main())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__pip-runner__.py b/backend/venv/lib/python3.12/site-packages/pip/__pip-runner__.py
new file mode 100644
index 00000000..c633787f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/__pip-runner__.py
@@ -0,0 +1,50 @@
+"""Execute exactly this copy of pip, within a different environment.
+
+This file is named as it is, to ensure that this module can't be imported via
+an import statement.
+"""
+
+# /!\ This version compatibility check section must be Python 2 compatible. /!\
+
+import sys
+
+# Copied from pyproject.toml
+PYTHON_REQUIRES = (3, 8)
+
+
+def version_str(version):  # type: ignore
+    return ".".join(str(v) for v in version)
+
+
+if sys.version_info[:2] < PYTHON_REQUIRES:
+    raise SystemExit(
+        "This version of pip does not support python {} (requires >={}).".format(
+            version_str(sys.version_info[:2]), version_str(PYTHON_REQUIRES)
+        )
+    )
+
+# From here on, we can use Python 3 features, but the syntax must remain
+# Python 2 compatible.
+
+import runpy  # noqa: E402
+from importlib.machinery import PathFinder  # noqa: E402
+from os.path import dirname  # noqa: E402
+
+PIP_SOURCES_ROOT = dirname(dirname(__file__))
+
+
+class PipImportRedirectingFinder:
+    @classmethod
+    def find_spec(self, fullname, path=None, target=None):  # type: ignore
+        if fullname != "pip":
+            return None
+
+        spec = PathFinder.find_spec(fullname, [PIP_SOURCES_ROOT], target)
+        assert spec, (PIP_SOURCES_ROOT, fullname)
+        return spec
+
+
+sys.meta_path.insert(0, PipImportRedirectingFinder())
+
+assert __name__ == "__main__", "Cannot run __pip-runner__.py as a non-main module"
+runpy.run_module("pip", run_name="__main__", alter_sys=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..248c46e8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..6a5ac8db
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__pip-runner__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__pip-runner__.cpython-312.pyc
new file mode 100644
index 00000000..8c5a7d95
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/__pycache__/__pip-runner__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/__init__.py
new file mode 100644
index 00000000..1a5b7f87
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/__init__.py
@@ -0,0 +1,18 @@
+from typing import List, Optional
+
+from pip._internal.utils import _log
+
+# init_logging() must be called before any call to logging.getLogger()
+# which happens at import of most modules.
+_log.init_logging()
+
+
+def main(args: Optional[List[str]] = None) -> int:
+    """This is preserved for old console scripts that may still be referencing
+    it.
+
+    For additional details, see https://github.com/pypa/pip/issues/7498.
+    """
+    from pip._internal.utils.entrypoints import _wrapper
+
+    return _wrapper(args)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0922ea8d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/build_env.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/build_env.cpython-312.pyc
new file mode 100644
index 00000000..abcaca89
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/build_env.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/cache.cpython-312.pyc
new file mode 100644
index 00000000..444f21c9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/configuration.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/configuration.cpython-312.pyc
new file mode 100644
index 00000000..5992d8da
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/configuration.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..9fb441b5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/main.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/main.cpython-312.pyc
new file mode 100644
index 00000000..d297b6f0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/main.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/pyproject.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/pyproject.cpython-312.pyc
new file mode 100644
index 00000000..c7d721ff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/pyproject.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/self_outdated_check.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/self_outdated_check.cpython-312.pyc
new file mode 100644
index 00000000..0c9c0872
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/self_outdated_check.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/wheel_builder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/wheel_builder.cpython-312.pyc
new file mode 100644
index 00000000..ebd11f2f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/__pycache__/wheel_builder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/build_env.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/build_env.py
new file mode 100644
index 00000000..e8d1aca0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/build_env.py
@@ -0,0 +1,322 @@
+"""Build Environment used for isolation during sdist building
+"""
+
+import logging
+import os
+import pathlib
+import site
+import sys
+import textwrap
+from collections import OrderedDict
+from types import TracebackType
+from typing import TYPE_CHECKING, Iterable, List, Optional, Set, Tuple, Type, Union
+
+from pip._vendor.packaging.version import Version
+
+from pip import __file__ as pip_location
+from pip._internal.cli.spinners import open_spinner
+from pip._internal.locations import get_platlib, get_purelib, get_scheme
+from pip._internal.metadata import get_default_environment, get_environment
+from pip._internal.utils.logging import VERBOSE
+from pip._internal.utils.packaging import get_requirement
+from pip._internal.utils.subprocess import call_subprocess
+from pip._internal.utils.temp_dir import TempDirectory, tempdir_kinds
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+
+logger = logging.getLogger(__name__)
+
+
+def _dedup(a: str, b: str) -> Union[Tuple[str], Tuple[str, str]]:
+    return (a, b) if a != b else (a,)
+
+
+class _Prefix:
+    def __init__(self, path: str) -> None:
+        self.path = path
+        self.setup = False
+        scheme = get_scheme("", prefix=path)
+        self.bin_dir = scheme.scripts
+        self.lib_dirs = _dedup(scheme.purelib, scheme.platlib)
+
+
+def get_runnable_pip() -> str:
+    """Get a file to pass to a Python executable, to run the currently-running pip.
+
+    This is used to run a pip subprocess, for installing requirements into the build
+    environment.
+    """
+    source = pathlib.Path(pip_location).resolve().parent
+
+    if not source.is_dir():
+        # This would happen if someone is using pip from inside a zip file. In that
+        # case, we can use that directly.
+        return str(source)
+
+    return os.fsdecode(source / "__pip-runner__.py")
+
+
+def _get_system_sitepackages() -> Set[str]:
+    """Get system site packages
+
+    Usually from site.getsitepackages,
+    but fallback on `get_purelib()/get_platlib()` if unavailable
+    (e.g. in a virtualenv created by virtualenv<20)
+
+    Returns normalized set of strings.
+    """
+    if hasattr(site, "getsitepackages"):
+        system_sites = site.getsitepackages()
+    else:
+        # virtualenv < 20 overwrites site.py without getsitepackages
+        # fallback on get_purelib/get_platlib.
+        # this is known to miss things, but shouldn't in the cases
+        # where getsitepackages() has been removed (inside a virtualenv)
+        system_sites = [get_purelib(), get_platlib()]
+    return {os.path.normcase(path) for path in system_sites}
+
+
+class BuildEnvironment:
+    """Creates and manages an isolated environment to install build deps"""
+
+    def __init__(self) -> None:
+        temp_dir = TempDirectory(kind=tempdir_kinds.BUILD_ENV, globally_managed=True)
+
+        self._prefixes = OrderedDict(
+            (name, _Prefix(os.path.join(temp_dir.path, name)))
+            for name in ("normal", "overlay")
+        )
+
+        self._bin_dirs: List[str] = []
+        self._lib_dirs: List[str] = []
+        for prefix in reversed(list(self._prefixes.values())):
+            self._bin_dirs.append(prefix.bin_dir)
+            self._lib_dirs.extend(prefix.lib_dirs)
+
+        # Customize site to:
+        # - ensure .pth files are honored
+        # - prevent access to system site packages
+        system_sites = _get_system_sitepackages()
+
+        self._site_dir = os.path.join(temp_dir.path, "site")
+        if not os.path.exists(self._site_dir):
+            os.mkdir(self._site_dir)
+        with open(
+            os.path.join(self._site_dir, "sitecustomize.py"), "w", encoding="utf-8"
+        ) as fp:
+            fp.write(
+                textwrap.dedent(
+                    """
+                import os, site, sys
+
+                # First, drop system-sites related paths.
+                original_sys_path = sys.path[:]
+                known_paths = set()
+                for path in {system_sites!r}:
+                    site.addsitedir(path, known_paths=known_paths)
+                system_paths = set(
+                    os.path.normcase(path)
+                    for path in sys.path[len(original_sys_path):]
+                )
+                original_sys_path = [
+                    path for path in original_sys_path
+                    if os.path.normcase(path) not in system_paths
+                ]
+                sys.path = original_sys_path
+
+                # Second, add lib directories.
+                # ensuring .pth file are processed.
+                for path in {lib_dirs!r}:
+                    assert not path in sys.path
+                    site.addsitedir(path)
+                """
+                ).format(system_sites=system_sites, lib_dirs=self._lib_dirs)
+            )
+
+    def __enter__(self) -> None:
+        self._save_env = {
+            name: os.environ.get(name, None)
+            for name in ("PATH", "PYTHONNOUSERSITE", "PYTHONPATH")
+        }
+
+        path = self._bin_dirs[:]
+        old_path = self._save_env["PATH"]
+        if old_path:
+            path.extend(old_path.split(os.pathsep))
+
+        pythonpath = [self._site_dir]
+
+        os.environ.update(
+            {
+                "PATH": os.pathsep.join(path),
+                "PYTHONNOUSERSITE": "1",
+                "PYTHONPATH": os.pathsep.join(pythonpath),
+            }
+        )
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        for varname, old_value in self._save_env.items():
+            if old_value is None:
+                os.environ.pop(varname, None)
+            else:
+                os.environ[varname] = old_value
+
+    def check_requirements(
+        self, reqs: Iterable[str]
+    ) -> Tuple[Set[Tuple[str, str]], Set[str]]:
+        """Return 2 sets:
+        - conflicting requirements: set of (installed, wanted) reqs tuples
+        - missing requirements: set of reqs
+        """
+        missing = set()
+        conflicting = set()
+        if reqs:
+            env = (
+                get_environment(self._lib_dirs)
+                if hasattr(self, "_lib_dirs")
+                else get_default_environment()
+            )
+            for req_str in reqs:
+                req = get_requirement(req_str)
+                # We're explicitly evaluating with an empty extra value, since build
+                # environments are not provided any mechanism to select specific extras.
+                if req.marker is not None and not req.marker.evaluate({"extra": ""}):
+                    continue
+                dist = env.get_distribution(req.name)
+                if not dist:
+                    missing.add(req_str)
+                    continue
+                if isinstance(dist.version, Version):
+                    installed_req_str = f"{req.name}=={dist.version}"
+                else:
+                    installed_req_str = f"{req.name}==={dist.version}"
+                if not req.specifier.contains(dist.version, prereleases=True):
+                    conflicting.add((installed_req_str, req_str))
+                # FIXME: Consider direct URL?
+        return conflicting, missing
+
+    def install_requirements(
+        self,
+        finder: "PackageFinder",
+        requirements: Iterable[str],
+        prefix_as_string: str,
+        *,
+        kind: str,
+    ) -> None:
+        prefix = self._prefixes[prefix_as_string]
+        assert not prefix.setup
+        prefix.setup = True
+        if not requirements:
+            return
+        self._install_requirements(
+            get_runnable_pip(),
+            finder,
+            requirements,
+            prefix,
+            kind=kind,
+        )
+
+    @staticmethod
+    def _install_requirements(
+        pip_runnable: str,
+        finder: "PackageFinder",
+        requirements: Iterable[str],
+        prefix: _Prefix,
+        *,
+        kind: str,
+    ) -> None:
+        args: List[str] = [
+            sys.executable,
+            pip_runnable,
+            "install",
+            "--ignore-installed",
+            "--no-user",
+            "--prefix",
+            prefix.path,
+            "--no-warn-script-location",
+            "--disable-pip-version-check",
+            # The prefix specified two lines above, thus
+            # target from config file or env var should be ignored
+            "--target",
+            "",
+        ]
+        if logger.getEffectiveLevel() <= logging.DEBUG:
+            args.append("-vv")
+        elif logger.getEffectiveLevel() <= VERBOSE:
+            args.append("-v")
+        for format_control in ("no_binary", "only_binary"):
+            formats = getattr(finder.format_control, format_control)
+            args.extend(
+                (
+                    "--" + format_control.replace("_", "-"),
+                    ",".join(sorted(formats or {":none:"})),
+                )
+            )
+
+        index_urls = finder.index_urls
+        if index_urls:
+            args.extend(["-i", index_urls[0]])
+            for extra_index in index_urls[1:]:
+                args.extend(["--extra-index-url", extra_index])
+        else:
+            args.append("--no-index")
+        for link in finder.find_links:
+            args.extend(["--find-links", link])
+
+        if finder.proxy:
+            args.extend(["--proxy", finder.proxy])
+        for host in finder.trusted_hosts:
+            args.extend(["--trusted-host", host])
+        if finder.custom_cert:
+            args.extend(["--cert", finder.custom_cert])
+        if finder.client_cert:
+            args.extend(["--client-cert", finder.client_cert])
+        if finder.allow_all_prereleases:
+            args.append("--pre")
+        if finder.prefer_binary:
+            args.append("--prefer-binary")
+        args.append("--")
+        args.extend(requirements)
+        with open_spinner(f"Installing {kind}") as spinner:
+            call_subprocess(
+                args,
+                command_desc=f"pip subprocess to install {kind}",
+                spinner=spinner,
+            )
+
+
+class NoOpBuildEnvironment(BuildEnvironment):
+    """A no-op drop-in replacement for BuildEnvironment"""
+
+    def __init__(self) -> None:
+        pass
+
+    def __enter__(self) -> None:
+        pass
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        pass
+
+    def cleanup(self) -> None:
+        pass
+
+    def install_requirements(
+        self,
+        finder: "PackageFinder",
+        requirements: Iterable[str],
+        prefix_as_string: str,
+        *,
+        kind: str,
+    ) -> None:
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cache.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cache.py
new file mode 100644
index 00000000..6b451267
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cache.py
@@ -0,0 +1,290 @@
+"""Cache Management
+"""
+
+import hashlib
+import json
+import logging
+import os
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+
+from pip._vendor.packaging.tags import Tag, interpreter_name, interpreter_version
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.exceptions import InvalidWheelFilename
+from pip._internal.models.direct_url import DirectUrl
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.utils.temp_dir import TempDirectory, tempdir_kinds
+from pip._internal.utils.urls import path_to_url
+
+logger = logging.getLogger(__name__)
+
+ORIGIN_JSON_NAME = "origin.json"
+
+
+def _hash_dict(d: Dict[str, str]) -> str:
+    """Return a stable sha224 of a dictionary."""
+    s = json.dumps(d, sort_keys=True, separators=(",", ":"), ensure_ascii=True)
+    return hashlib.sha224(s.encode("ascii")).hexdigest()
+
+
+class Cache:
+    """An abstract class - provides cache directories for data from links
+
+    :param cache_dir: The root of the cache.
+    """
+
+    def __init__(self, cache_dir: str) -> None:
+        super().__init__()
+        assert not cache_dir or os.path.isabs(cache_dir)
+        self.cache_dir = cache_dir or None
+
+    def _get_cache_path_parts(self, link: Link) -> List[str]:
+        """Get parts of part that must be os.path.joined with cache_dir"""
+
+        # We want to generate an url to use as our cache key, we don't want to
+        # just reuse the URL because it might have other items in the fragment
+        # and we don't care about those.
+        key_parts = {"url": link.url_without_fragment}
+        if link.hash_name is not None and link.hash is not None:
+            key_parts[link.hash_name] = link.hash
+        if link.subdirectory_fragment:
+            key_parts["subdirectory"] = link.subdirectory_fragment
+
+        # Include interpreter name, major and minor version in cache key
+        # to cope with ill-behaved sdists that build a different wheel
+        # depending on the python version their setup.py is being run on,
+        # and don't encode the difference in compatibility tags.
+        # https://github.com/pypa/pip/issues/7296
+        key_parts["interpreter_name"] = interpreter_name()
+        key_parts["interpreter_version"] = interpreter_version()
+
+        # Encode our key url with sha224, we'll use this because it has similar
+        # security properties to sha256, but with a shorter total output (and
+        # thus less secure). However the differences don't make a lot of
+        # difference for our use case here.
+        hashed = _hash_dict(key_parts)
+
+        # We want to nest the directories some to prevent having a ton of top
+        # level directories where we might run out of sub directories on some
+        # FS.
+        parts = [hashed[:2], hashed[2:4], hashed[4:6], hashed[6:]]
+
+        return parts
+
+    def _get_candidates(self, link: Link, canonical_package_name: str) -> List[Any]:
+        can_not_cache = not self.cache_dir or not canonical_package_name or not link
+        if can_not_cache:
+            return []
+
+        path = self.get_path_for_link(link)
+        if os.path.isdir(path):
+            return [(candidate, path) for candidate in os.listdir(path)]
+        return []
+
+    def get_path_for_link(self, link: Link) -> str:
+        """Return a directory to store cached items in for link."""
+        raise NotImplementedError()
+
+    def get(
+        self,
+        link: Link,
+        package_name: Optional[str],
+        supported_tags: List[Tag],
+    ) -> Link:
+        """Returns a link to a cached item if it exists, otherwise returns the
+        passed link.
+        """
+        raise NotImplementedError()
+
+
+class SimpleWheelCache(Cache):
+    """A cache of wheels for future installs."""
+
+    def __init__(self, cache_dir: str) -> None:
+        super().__init__(cache_dir)
+
+    def get_path_for_link(self, link: Link) -> str:
+        """Return a directory to store cached wheels for link
+
+        Because there are M wheels for any one sdist, we provide a directory
+        to cache them in, and then consult that directory when looking up
+        cache hits.
+
+        We only insert things into the cache if they have plausible version
+        numbers, so that we don't contaminate the cache with things that were
+        not unique. E.g. ./package might have dozens of installs done for it
+        and build a version of 0.0...and if we built and cached a wheel, we'd
+        end up using the same wheel even if the source has been edited.
+
+        :param link: The link of the sdist for which this will cache wheels.
+        """
+        parts = self._get_cache_path_parts(link)
+        assert self.cache_dir
+        # Store wheels within the root cache_dir
+        return os.path.join(self.cache_dir, "wheels", *parts)
+
+    def get(
+        self,
+        link: Link,
+        package_name: Optional[str],
+        supported_tags: List[Tag],
+    ) -> Link:
+        candidates = []
+
+        if not package_name:
+            return link
+
+        canonical_package_name = canonicalize_name(package_name)
+        for wheel_name, wheel_dir in self._get_candidates(link, canonical_package_name):
+            try:
+                wheel = Wheel(wheel_name)
+            except InvalidWheelFilename:
+                continue
+            if canonicalize_name(wheel.name) != canonical_package_name:
+                logger.debug(
+                    "Ignoring cached wheel %s for %s as it "
+                    "does not match the expected distribution name %s.",
+                    wheel_name,
+                    link,
+                    package_name,
+                )
+                continue
+            if not wheel.supported(supported_tags):
+                # Built for a different python/arch/etc
+                continue
+            candidates.append(
+                (
+                    wheel.support_index_min(supported_tags),
+                    wheel_name,
+                    wheel_dir,
+                )
+            )
+
+        if not candidates:
+            return link
+
+        _, wheel_name, wheel_dir = min(candidates)
+        return Link(path_to_url(os.path.join(wheel_dir, wheel_name)))
+
+
+class EphemWheelCache(SimpleWheelCache):
+    """A SimpleWheelCache that creates it's own temporary cache directory"""
+
+    def __init__(self) -> None:
+        self._temp_dir = TempDirectory(
+            kind=tempdir_kinds.EPHEM_WHEEL_CACHE,
+            globally_managed=True,
+        )
+
+        super().__init__(self._temp_dir.path)
+
+
+class CacheEntry:
+    def __init__(
+        self,
+        link: Link,
+        persistent: bool,
+    ):
+        self.link = link
+        self.persistent = persistent
+        self.origin: Optional[DirectUrl] = None
+        origin_direct_url_path = Path(self.link.file_path).parent / ORIGIN_JSON_NAME
+        if origin_direct_url_path.exists():
+            try:
+                self.origin = DirectUrl.from_json(
+                    origin_direct_url_path.read_text(encoding="utf-8")
+                )
+            except Exception as e:
+                logger.warning(
+                    "Ignoring invalid cache entry origin file %s for %s (%s)",
+                    origin_direct_url_path,
+                    link.filename,
+                    e,
+                )
+
+
+class WheelCache(Cache):
+    """Wraps EphemWheelCache and SimpleWheelCache into a single Cache
+
+    This Cache allows for gracefully degradation, using the ephem wheel cache
+    when a certain link is not found in the simple wheel cache first.
+    """
+
+    def __init__(self, cache_dir: str) -> None:
+        super().__init__(cache_dir)
+        self._wheel_cache = SimpleWheelCache(cache_dir)
+        self._ephem_cache = EphemWheelCache()
+
+    def get_path_for_link(self, link: Link) -> str:
+        return self._wheel_cache.get_path_for_link(link)
+
+    def get_ephem_path_for_link(self, link: Link) -> str:
+        return self._ephem_cache.get_path_for_link(link)
+
+    def get(
+        self,
+        link: Link,
+        package_name: Optional[str],
+        supported_tags: List[Tag],
+    ) -> Link:
+        cache_entry = self.get_cache_entry(link, package_name, supported_tags)
+        if cache_entry is None:
+            return link
+        return cache_entry.link
+
+    def get_cache_entry(
+        self,
+        link: Link,
+        package_name: Optional[str],
+        supported_tags: List[Tag],
+    ) -> Optional[CacheEntry]:
+        """Returns a CacheEntry with a link to a cached item if it exists or
+        None. The cache entry indicates if the item was found in the persistent
+        or ephemeral cache.
+        """
+        retval = self._wheel_cache.get(
+            link=link,
+            package_name=package_name,
+            supported_tags=supported_tags,
+        )
+        if retval is not link:
+            return CacheEntry(retval, persistent=True)
+
+        retval = self._ephem_cache.get(
+            link=link,
+            package_name=package_name,
+            supported_tags=supported_tags,
+        )
+        if retval is not link:
+            return CacheEntry(retval, persistent=False)
+
+        return None
+
+    @staticmethod
+    def record_download_origin(cache_dir: str, download_info: DirectUrl) -> None:
+        origin_path = Path(cache_dir) / ORIGIN_JSON_NAME
+        if origin_path.exists():
+            try:
+                origin = DirectUrl.from_json(origin_path.read_text(encoding="utf-8"))
+            except Exception as e:
+                logger.warning(
+                    "Could not read origin file %s in cache entry (%s). "
+                    "Will attempt to overwrite it.",
+                    origin_path,
+                    e,
+                )
+            else:
+                # TODO: use DirectUrl.equivalent when
+                # https://github.com/pypa/pip/pull/10564 is merged.
+                if origin.url != download_info.url:
+                    logger.warning(
+                        "Origin URL %s in cache entry %s does not match download URL "
+                        "%s. This is likely a pip bug or a cache corruption issue. "
+                        "Will overwrite it with the new value.",
+                        origin.url,
+                        cache_dir,
+                        download_info.url,
+                    )
+        origin_path.write_text(download_info.to_json(), encoding="utf-8")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__init__.py
new file mode 100644
index 00000000..e589bb91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__init__.py
@@ -0,0 +1,4 @@
+"""Subpackage containing all of pip's command line interface related code
+"""
+
+# This file intentionally does not import submodules
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..12f5d610
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/autocompletion.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/autocompletion.cpython-312.pyc
new file mode 100644
index 00000000..4d992d27
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/autocompletion.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/base_command.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/base_command.cpython-312.pyc
new file mode 100644
index 00000000..c476875b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/base_command.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/cmdoptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/cmdoptions.cpython-312.pyc
new file mode 100644
index 00000000..727459d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/cmdoptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/command_context.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/command_context.cpython-312.pyc
new file mode 100644
index 00000000..4f8f2af9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/command_context.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/index_command.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/index_command.cpython-312.pyc
new file mode 100644
index 00000000..336283d3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/index_command.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main.cpython-312.pyc
new file mode 100644
index 00000000..110c578a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main_parser.cpython-312.pyc
new file mode 100644
index 00000000..62a56d9f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/main_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/parser.cpython-312.pyc
new file mode 100644
index 00000000..963b475b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/progress_bars.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/progress_bars.cpython-312.pyc
new file mode 100644
index 00000000..d94258cf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/progress_bars.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/req_command.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/req_command.cpython-312.pyc
new file mode 100644
index 00000000..a00a8a6f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/req_command.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/spinners.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/spinners.cpython-312.pyc
new file mode 100644
index 00000000..8c3a49ae
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/spinners.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/status_codes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/status_codes.cpython-312.pyc
new file mode 100644
index 00000000..1aa96714
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/__pycache__/status_codes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/autocompletion.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/autocompletion.py
new file mode 100644
index 00000000..f3f70ac8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/autocompletion.py
@@ -0,0 +1,176 @@
+"""Logic that powers autocompletion installed by ``pip completion``.
+"""
+
+import optparse
+import os
+import sys
+from itertools import chain
+from typing import Any, Iterable, List, Optional
+
+from pip._internal.cli.main_parser import create_main_parser
+from pip._internal.commands import commands_dict, create_command
+from pip._internal.metadata import get_default_environment
+
+
+def autocomplete() -> None:
+    """Entry Point for completion of main and subcommand options."""
+    # Don't complete if user hasn't sourced bash_completion file.
+    if "PIP_AUTO_COMPLETE" not in os.environ:
+        return
+    # Don't complete if autocompletion environment variables
+    # are not present
+    if not os.environ.get("COMP_WORDS") or not os.environ.get("COMP_CWORD"):
+        return
+    cwords = os.environ["COMP_WORDS"].split()[1:]
+    cword = int(os.environ["COMP_CWORD"])
+    try:
+        current = cwords[cword - 1]
+    except IndexError:
+        current = ""
+
+    parser = create_main_parser()
+    subcommands = list(commands_dict)
+    options = []
+
+    # subcommand
+    subcommand_name: Optional[str] = None
+    for word in cwords:
+        if word in subcommands:
+            subcommand_name = word
+            break
+    # subcommand options
+    if subcommand_name is not None:
+        # special case: 'help' subcommand has no options
+        if subcommand_name == "help":
+            sys.exit(1)
+        # special case: list locally installed dists for show and uninstall
+        should_list_installed = not current.startswith("-") and subcommand_name in [
+            "show",
+            "uninstall",
+        ]
+        if should_list_installed:
+            env = get_default_environment()
+            lc = current.lower()
+            installed = [
+                dist.canonical_name
+                for dist in env.iter_installed_distributions(local_only=True)
+                if dist.canonical_name.startswith(lc)
+                and dist.canonical_name not in cwords[1:]
+            ]
+            # if there are no dists installed, fall back to option completion
+            if installed:
+                for dist in installed:
+                    print(dist)
+                sys.exit(1)
+
+        should_list_installables = (
+            not current.startswith("-") and subcommand_name == "install"
+        )
+        if should_list_installables:
+            for path in auto_complete_paths(current, "path"):
+                print(path)
+            sys.exit(1)
+
+        subcommand = create_command(subcommand_name)
+
+        for opt in subcommand.parser.option_list_all:
+            if opt.help != optparse.SUPPRESS_HELP:
+                options += [
+                    (opt_str, opt.nargs) for opt_str in opt._long_opts + opt._short_opts
+                ]
+
+        # filter out previously specified options from available options
+        prev_opts = [x.split("=")[0] for x in cwords[1 : cword - 1]]
+        options = [(x, v) for (x, v) in options if x not in prev_opts]
+        # filter options by current input
+        options = [(k, v) for k, v in options if k.startswith(current)]
+        # get completion type given cwords and available subcommand options
+        completion_type = get_path_completion_type(
+            cwords,
+            cword,
+            subcommand.parser.option_list_all,
+        )
+        # get completion files and directories if ``completion_type`` is
+        # ``<file>``, ``<dir>`` or ``<path>``
+        if completion_type:
+            paths = auto_complete_paths(current, completion_type)
+            options = [(path, 0) for path in paths]
+        for option in options:
+            opt_label = option[0]
+            # append '=' to options which require args
+            if option[1] and option[0][:2] == "--":
+                opt_label += "="
+            print(opt_label)
+    else:
+        # show main parser options only when necessary
+
+        opts = [i.option_list for i in parser.option_groups]
+        opts.append(parser.option_list)
+        flattened_opts = chain.from_iterable(opts)
+        if current.startswith("-"):
+            for opt in flattened_opts:
+                if opt.help != optparse.SUPPRESS_HELP:
+                    subcommands += opt._long_opts + opt._short_opts
+        else:
+            # get completion type given cwords and all available options
+            completion_type = get_path_completion_type(cwords, cword, flattened_opts)
+            if completion_type:
+                subcommands = list(auto_complete_paths(current, completion_type))
+
+        print(" ".join([x for x in subcommands if x.startswith(current)]))
+    sys.exit(1)
+
+
+def get_path_completion_type(
+    cwords: List[str], cword: int, opts: Iterable[Any]
+) -> Optional[str]:
+    """Get the type of path completion (``file``, ``dir``, ``path`` or None)
+
+    :param cwords: same as the environmental variable ``COMP_WORDS``
+    :param cword: same as the environmental variable ``COMP_CWORD``
+    :param opts: The available options to check
+    :return: path completion type (``file``, ``dir``, ``path`` or None)
+    """
+    if cword < 2 or not cwords[cword - 2].startswith("-"):
+        return None
+    for opt in opts:
+        if opt.help == optparse.SUPPRESS_HELP:
+            continue
+        for o in str(opt).split("/"):
+            if cwords[cword - 2].split("=")[0] == o:
+                if not opt.metavar or any(
+                    x in ("path", "file", "dir") for x in opt.metavar.split("/")
+                ):
+                    return opt.metavar
+    return None
+
+
+def auto_complete_paths(current: str, completion_type: str) -> Iterable[str]:
+    """If ``completion_type`` is ``file`` or ``path``, list all regular files
+    and directories starting with ``current``; otherwise only list directories
+    starting with ``current``.
+
+    :param current: The word to be completed
+    :param completion_type: path completion type(``file``, ``path`` or ``dir``)
+    :return: A generator of regular files and/or directories
+    """
+    directory, filename = os.path.split(current)
+    current_path = os.path.abspath(directory)
+    # Don't complete paths if they can't be accessed
+    if not os.access(current_path, os.R_OK):
+        return
+    filename = os.path.normcase(filename)
+    # list all files that start with ``filename``
+    file_list = (
+        x for x in os.listdir(current_path) if os.path.normcase(x).startswith(filename)
+    )
+    for f in file_list:
+        opt = os.path.join(current_path, f)
+        comp_file = os.path.normcase(os.path.join(directory, f))
+        # complete regular files when there is not ``<dir>`` after option
+        # complete directories when there is ``<file>``, ``<path>`` or
+        # ``<dir>``after option
+        if completion_type != "dir" and os.path.isfile(opt):
+            yield comp_file
+        elif os.path.isdir(opt):
+            yield os.path.join(comp_file, "")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/base_command.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/base_command.py
new file mode 100644
index 00000000..362f84b6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/base_command.py
@@ -0,0 +1,240 @@
+"""Base Command class, and related routines"""
+
+import logging
+import logging.config
+import optparse
+import os
+import sys
+import traceback
+from optparse import Values
+from typing import List, Optional, Tuple
+
+from pip._vendor.rich import reconfigure
+from pip._vendor.rich import traceback as rich_traceback
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.command_context import CommandContextMixIn
+from pip._internal.cli.parser import ConfigOptionParser, UpdatingDefaultsHelpFormatter
+from pip._internal.cli.status_codes import (
+    ERROR,
+    PREVIOUS_BUILD_DIR_ERROR,
+    UNKNOWN_ERROR,
+    VIRTUALENV_NOT_FOUND,
+)
+from pip._internal.exceptions import (
+    BadCommand,
+    CommandError,
+    DiagnosticPipError,
+    InstallationError,
+    NetworkConnectionError,
+    PreviousBuildDirError,
+)
+from pip._internal.utils.deprecation import deprecated
+from pip._internal.utils.filesystem import check_path_owner
+from pip._internal.utils.logging import BrokenStdoutLoggingError, setup_logging
+from pip._internal.utils.misc import get_prog, normalize_path
+from pip._internal.utils.temp_dir import TempDirectoryTypeRegistry as TempDirRegistry
+from pip._internal.utils.temp_dir import global_tempdir_manager, tempdir_registry
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+__all__ = ["Command"]
+
+logger = logging.getLogger(__name__)
+
+
+class Command(CommandContextMixIn):
+    usage: str = ""
+    ignore_require_venv: bool = False
+
+    def __init__(self, name: str, summary: str, isolated: bool = False) -> None:
+        super().__init__()
+
+        self.name = name
+        self.summary = summary
+        self.parser = ConfigOptionParser(
+            usage=self.usage,
+            prog=f"{get_prog()} {name}",
+            formatter=UpdatingDefaultsHelpFormatter(),
+            add_help_option=False,
+            name=name,
+            description=self.__doc__,
+            isolated=isolated,
+        )
+
+        self.tempdir_registry: Optional[TempDirRegistry] = None
+
+        # Commands should add options to this option group
+        optgroup_name = f"{self.name.capitalize()} Options"
+        self.cmd_opts = optparse.OptionGroup(self.parser, optgroup_name)
+
+        # Add the general options
+        gen_opts = cmdoptions.make_option_group(
+            cmdoptions.general_group,
+            self.parser,
+        )
+        self.parser.add_option_group(gen_opts)
+
+        self.add_options()
+
+    def add_options(self) -> None:
+        pass
+
+    def handle_pip_version_check(self, options: Values) -> None:
+        """
+        This is a no-op so that commands by default do not do the pip version
+        check.
+        """
+        # Make sure we do the pip version check if the index_group options
+        # are present.
+        assert not hasattr(options, "no_index")
+
+    def run(self, options: Values, args: List[str]) -> int:
+        raise NotImplementedError
+
+    def _run_wrapper(self, level_number: int, options: Values, args: List[str]) -> int:
+        def _inner_run() -> int:
+            try:
+                return self.run(options, args)
+            finally:
+                self.handle_pip_version_check(options)
+
+        if options.debug_mode:
+            rich_traceback.install(show_locals=True)
+            return _inner_run()
+
+        try:
+            status = _inner_run()
+            assert isinstance(status, int)
+            return status
+        except DiagnosticPipError as exc:
+            logger.error("%s", exc, extra={"rich": True})
+            logger.debug("Exception information:", exc_info=True)
+
+            return ERROR
+        except PreviousBuildDirError as exc:
+            logger.critical(str(exc))
+            logger.debug("Exception information:", exc_info=True)
+
+            return PREVIOUS_BUILD_DIR_ERROR
+        except (
+            InstallationError,
+            BadCommand,
+            NetworkConnectionError,
+        ) as exc:
+            logger.critical(str(exc))
+            logger.debug("Exception information:", exc_info=True)
+
+            return ERROR
+        except CommandError as exc:
+            logger.critical("%s", exc)
+            logger.debug("Exception information:", exc_info=True)
+
+            return ERROR
+        except BrokenStdoutLoggingError:
+            # Bypass our logger and write any remaining messages to
+            # stderr because stdout no longer works.
+            print("ERROR: Pipe to stdout was broken", file=sys.stderr)
+            if level_number <= logging.DEBUG:
+                traceback.print_exc(file=sys.stderr)
+
+            return ERROR
+        except KeyboardInterrupt:
+            logger.critical("Operation cancelled by user")
+            logger.debug("Exception information:", exc_info=True)
+
+            return ERROR
+        except BaseException:
+            logger.critical("Exception:", exc_info=True)
+
+            return UNKNOWN_ERROR
+
+    def parse_args(self, args: List[str]) -> Tuple[Values, List[str]]:
+        # factored out for testability
+        return self.parser.parse_args(args)
+
+    def main(self, args: List[str]) -> int:
+        try:
+            with self.main_context():
+                return self._main(args)
+        finally:
+            logging.shutdown()
+
+    def _main(self, args: List[str]) -> int:
+        # We must initialize this before the tempdir manager, otherwise the
+        # configuration would not be accessible by the time we clean up the
+        # tempdir manager.
+        self.tempdir_registry = self.enter_context(tempdir_registry())
+        # Intentionally set as early as possible so globally-managed temporary
+        # directories are available to the rest of the code.
+        self.enter_context(global_tempdir_manager())
+
+        options, args = self.parse_args(args)
+
+        # Set verbosity so that it can be used elsewhere.
+        self.verbosity = options.verbose - options.quiet
+
+        reconfigure(no_color=options.no_color)
+        level_number = setup_logging(
+            verbosity=self.verbosity,
+            no_color=options.no_color,
+            user_log_file=options.log,
+        )
+
+        always_enabled_features = set(options.features_enabled) & set(
+            cmdoptions.ALWAYS_ENABLED_FEATURES
+        )
+        if always_enabled_features:
+            logger.warning(
+                "The following features are always enabled: %s. ",
+                ", ".join(sorted(always_enabled_features)),
+            )
+
+        # Make sure that the --python argument isn't specified after the
+        # subcommand. We can tell, because if --python was specified,
+        # we should only reach this point if we're running in the created
+        # subprocess, which has the _PIP_RUNNING_IN_SUBPROCESS environment
+        # variable set.
+        if options.python and "_PIP_RUNNING_IN_SUBPROCESS" not in os.environ:
+            logger.critical(
+                "The --python option must be placed before the pip subcommand name"
+            )
+            sys.exit(ERROR)
+
+        # TODO: Try to get these passing down from the command?
+        #       without resorting to os.environ to hold these.
+        #       This also affects isolated builds and it should.
+
+        if options.no_input:
+            os.environ["PIP_NO_INPUT"] = "1"
+
+        if options.exists_action:
+            os.environ["PIP_EXISTS_ACTION"] = " ".join(options.exists_action)
+
+        if options.require_venv and not self.ignore_require_venv:
+            # If a venv is required check if it can really be found
+            if not running_under_virtualenv():
+                logger.critical("Could not find an activated virtualenv (required).")
+                sys.exit(VIRTUALENV_NOT_FOUND)
+
+        if options.cache_dir:
+            options.cache_dir = normalize_path(options.cache_dir)
+            if not check_path_owner(options.cache_dir):
+                logger.warning(
+                    "The directory '%s' or its parent directory is not owned "
+                    "or is not writable by the current user. The cache "
+                    "has been disabled. Check the permissions and owner of "
+                    "that directory. If executing pip with sudo, you should "
+                    "use sudo's -H flag.",
+                    options.cache_dir,
+                )
+                options.cache_dir = None
+
+        if options.no_python_version_warning:
+            deprecated(
+                reason="--no-python-version-warning is deprecated.",
+                replacement="to remove the flag as it's a no-op",
+                gone_in="25.1",
+                issue=13154,
+            )
+
+        return self._run_wrapper(level_number, options, args)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/cmdoptions.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/cmdoptions.py
new file mode 100644
index 00000000..eeb7e651
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/cmdoptions.py
@@ -0,0 +1,1075 @@
+"""
+shared options and groups
+
+The principle here is to define options once, but *not* instantiate them
+globally. One reason being that options with action='append' can carry state
+between parses. pip parses general options twice internally, and shouldn't
+pass on state. To be consistent, all options will follow this design.
+"""
+
+# The following comment should be removed at some point in the future.
+# mypy: strict-optional=False
+
+import importlib.util
+import logging
+import os
+import textwrap
+from functools import partial
+from optparse import SUPPRESS_HELP, Option, OptionGroup, OptionParser, Values
+from textwrap import dedent
+from typing import Any, Callable, Dict, Optional, Tuple
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.cli.parser import ConfigOptionParser
+from pip._internal.exceptions import CommandError
+from pip._internal.locations import USER_CACHE_DIR, get_src_prefix
+from pip._internal.models.format_control import FormatControl
+from pip._internal.models.index import PyPI
+from pip._internal.models.target_python import TargetPython
+from pip._internal.utils.hashes import STRONG_HASHES
+from pip._internal.utils.misc import strtobool
+
+logger = logging.getLogger(__name__)
+
+
+def raise_option_error(parser: OptionParser, option: Option, msg: str) -> None:
+    """
+    Raise an option parsing error using parser.error().
+
+    Args:
+      parser: an OptionParser instance.
+      option: an Option instance.
+      msg: the error text.
+    """
+    msg = f"{option} error: {msg}"
+    msg = textwrap.fill(" ".join(msg.split()))
+    parser.error(msg)
+
+
+def make_option_group(group: Dict[str, Any], parser: ConfigOptionParser) -> OptionGroup:
+    """
+    Return an OptionGroup object
+    group  -- assumed to be dict with 'name' and 'options' keys
+    parser -- an optparse Parser
+    """
+    option_group = OptionGroup(parser, group["name"])
+    for option in group["options"]:
+        option_group.add_option(option())
+    return option_group
+
+
+def check_dist_restriction(options: Values, check_target: bool = False) -> None:
+    """Function for determining if custom platform options are allowed.
+
+    :param options: The OptionParser options.
+    :param check_target: Whether or not to check if --target is being used.
+    """
+    dist_restriction_set = any(
+        [
+            options.python_version,
+            options.platforms,
+            options.abis,
+            options.implementation,
+        ]
+    )
+
+    binary_only = FormatControl(set(), {":all:"})
+    sdist_dependencies_allowed = (
+        options.format_control != binary_only and not options.ignore_dependencies
+    )
+
+    # Installations or downloads using dist restrictions must not combine
+    # source distributions and dist-specific wheels, as they are not
+    # guaranteed to be locally compatible.
+    if dist_restriction_set and sdist_dependencies_allowed:
+        raise CommandError(
+            "When restricting platform and interpreter constraints using "
+            "--python-version, --platform, --abi, or --implementation, "
+            "either --no-deps must be set, or --only-binary=:all: must be "
+            "set and --no-binary must not be set (or must be set to "
+            ":none:)."
+        )
+
+    if check_target:
+        if not options.dry_run and dist_restriction_set and not options.target_dir:
+            raise CommandError(
+                "Can not use any platform or abi specific options unless "
+                "installing via '--target' or using '--dry-run'"
+            )
+
+
+def _path_option_check(option: Option, opt: str, value: str) -> str:
+    return os.path.expanduser(value)
+
+
+def _package_name_option_check(option: Option, opt: str, value: str) -> str:
+    return canonicalize_name(value)
+
+
+class PipOption(Option):
+    TYPES = Option.TYPES + ("path", "package_name")
+    TYPE_CHECKER = Option.TYPE_CHECKER.copy()
+    TYPE_CHECKER["package_name"] = _package_name_option_check
+    TYPE_CHECKER["path"] = _path_option_check
+
+
+###########
+# options #
+###########
+
+help_: Callable[..., Option] = partial(
+    Option,
+    "-h",
+    "--help",
+    dest="help",
+    action="help",
+    help="Show help.",
+)
+
+debug_mode: Callable[..., Option] = partial(
+    Option,
+    "--debug",
+    dest="debug_mode",
+    action="store_true",
+    default=False,
+    help=(
+        "Let unhandled exceptions propagate outside the main subroutine, "
+        "instead of logging them to stderr."
+    ),
+)
+
+isolated_mode: Callable[..., Option] = partial(
+    Option,
+    "--isolated",
+    dest="isolated_mode",
+    action="store_true",
+    default=False,
+    help=(
+        "Run pip in an isolated mode, ignoring environment variables and user "
+        "configuration."
+    ),
+)
+
+require_virtualenv: Callable[..., Option] = partial(
+    Option,
+    "--require-virtualenv",
+    "--require-venv",
+    dest="require_venv",
+    action="store_true",
+    default=False,
+    help=(
+        "Allow pip to only run in a virtual environment; "
+        "exit with an error otherwise."
+    ),
+)
+
+override_externally_managed: Callable[..., Option] = partial(
+    Option,
+    "--break-system-packages",
+    dest="override_externally_managed",
+    action="store_true",
+    help="Allow pip to modify an EXTERNALLY-MANAGED Python installation",
+)
+
+python: Callable[..., Option] = partial(
+    Option,
+    "--python",
+    dest="python",
+    help="Run pip with the specified Python interpreter.",
+)
+
+verbose: Callable[..., Option] = partial(
+    Option,
+    "-v",
+    "--verbose",
+    dest="verbose",
+    action="count",
+    default=0,
+    help="Give more output. Option is additive, and can be used up to 3 times.",
+)
+
+no_color: Callable[..., Option] = partial(
+    Option,
+    "--no-color",
+    dest="no_color",
+    action="store_true",
+    default=False,
+    help="Suppress colored output.",
+)
+
+version: Callable[..., Option] = partial(
+    Option,
+    "-V",
+    "--version",
+    dest="version",
+    action="store_true",
+    help="Show version and exit.",
+)
+
+quiet: Callable[..., Option] = partial(
+    Option,
+    "-q",
+    "--quiet",
+    dest="quiet",
+    action="count",
+    default=0,
+    help=(
+        "Give less output. Option is additive, and can be used up to 3"
+        " times (corresponding to WARNING, ERROR, and CRITICAL logging"
+        " levels)."
+    ),
+)
+
+progress_bar: Callable[..., Option] = partial(
+    Option,
+    "--progress-bar",
+    dest="progress_bar",
+    type="choice",
+    choices=["on", "off", "raw"],
+    default="on",
+    help="Specify whether the progress bar should be used [on, off, raw] (default: on)",
+)
+
+log: Callable[..., Option] = partial(
+    PipOption,
+    "--log",
+    "--log-file",
+    "--local-log",
+    dest="log",
+    metavar="path",
+    type="path",
+    help="Path to a verbose appending log.",
+)
+
+no_input: Callable[..., Option] = partial(
+    Option,
+    # Don't ask for input
+    "--no-input",
+    dest="no_input",
+    action="store_true",
+    default=False,
+    help="Disable prompting for input.",
+)
+
+keyring_provider: Callable[..., Option] = partial(
+    Option,
+    "--keyring-provider",
+    dest="keyring_provider",
+    choices=["auto", "disabled", "import", "subprocess"],
+    default="auto",
+    help=(
+        "Enable the credential lookup via the keyring library if user input is allowed."
+        " Specify which mechanism to use [auto, disabled, import, subprocess]."
+        " (default: %default)"
+    ),
+)
+
+proxy: Callable[..., Option] = partial(
+    Option,
+    "--proxy",
+    dest="proxy",
+    type="str",
+    default="",
+    help="Specify a proxy in the form scheme://[user:passwd@]proxy.server:port.",
+)
+
+retries: Callable[..., Option] = partial(
+    Option,
+    "--retries",
+    dest="retries",
+    type="int",
+    default=5,
+    help="Maximum number of retries each connection should attempt "
+    "(default %default times).",
+)
+
+timeout: Callable[..., Option] = partial(
+    Option,
+    "--timeout",
+    "--default-timeout",
+    metavar="sec",
+    dest="timeout",
+    type="float",
+    default=15,
+    help="Set the socket timeout (default %default seconds).",
+)
+
+
+def exists_action() -> Option:
+    return Option(
+        # Option when path already exist
+        "--exists-action",
+        dest="exists_action",
+        type="choice",
+        choices=["s", "i", "w", "b", "a"],
+        default=[],
+        action="append",
+        metavar="action",
+        help="Default action when a path already exists: "
+        "(s)witch, (i)gnore, (w)ipe, (b)ackup, (a)bort.",
+    )
+
+
+cert: Callable[..., Option] = partial(
+    PipOption,
+    "--cert",
+    dest="cert",
+    type="path",
+    metavar="path",
+    help=(
+        "Path to PEM-encoded CA certificate bundle. "
+        "If provided, overrides the default. "
+        "See 'SSL Certificate Verification' in pip documentation "
+        "for more information."
+    ),
+)
+
+client_cert: Callable[..., Option] = partial(
+    PipOption,
+    "--client-cert",
+    dest="client_cert",
+    type="path",
+    default=None,
+    metavar="path",
+    help="Path to SSL client certificate, a single file containing the "
+    "private key and the certificate in PEM format.",
+)
+
+index_url: Callable[..., Option] = partial(
+    Option,
+    "-i",
+    "--index-url",
+    "--pypi-url",
+    dest="index_url",
+    metavar="URL",
+    default=PyPI.simple_url,
+    help="Base URL of the Python Package Index (default %default). "
+    "This should point to a repository compliant with PEP 503 "
+    "(the simple repository API) or a local directory laid out "
+    "in the same format.",
+)
+
+
+def extra_index_url() -> Option:
+    return Option(
+        "--extra-index-url",
+        dest="extra_index_urls",
+        metavar="URL",
+        action="append",
+        default=[],
+        help="Extra URLs of package indexes to use in addition to "
+        "--index-url. Should follow the same rules as "
+        "--index-url.",
+    )
+
+
+no_index: Callable[..., Option] = partial(
+    Option,
+    "--no-index",
+    dest="no_index",
+    action="store_true",
+    default=False,
+    help="Ignore package index (only looking at --find-links URLs instead).",
+)
+
+
+def find_links() -> Option:
+    return Option(
+        "-f",
+        "--find-links",
+        dest="find_links",
+        action="append",
+        default=[],
+        metavar="url",
+        help="If a URL or path to an html file, then parse for links to "
+        "archives such as sdist (.tar.gz) or wheel (.whl) files. "
+        "If a local path or file:// URL that's a directory, "
+        "then look for archives in the directory listing. "
+        "Links to VCS project URLs are not supported.",
+    )
+
+
+def trusted_host() -> Option:
+    return Option(
+        "--trusted-host",
+        dest="trusted_hosts",
+        action="append",
+        metavar="HOSTNAME",
+        default=[],
+        help="Mark this host or host:port pair as trusted, even though it "
+        "does not have valid or any HTTPS.",
+    )
+
+
+def constraints() -> Option:
+    return Option(
+        "-c",
+        "--constraint",
+        dest="constraints",
+        action="append",
+        default=[],
+        metavar="file",
+        help="Constrain versions using the given constraints file. "
+        "This option can be used multiple times.",
+    )
+
+
+def requirements() -> Option:
+    return Option(
+        "-r",
+        "--requirement",
+        dest="requirements",
+        action="append",
+        default=[],
+        metavar="file",
+        help="Install from the given requirements file. "
+        "This option can be used multiple times.",
+    )
+
+
+def editable() -> Option:
+    return Option(
+        "-e",
+        "--editable",
+        dest="editables",
+        action="append",
+        default=[],
+        metavar="path/url",
+        help=(
+            "Install a project in editable mode (i.e. setuptools "
+            '"develop mode") from a local project path or a VCS url.'
+        ),
+    )
+
+
+def _handle_src(option: Option, opt_str: str, value: str, parser: OptionParser) -> None:
+    value = os.path.abspath(value)
+    setattr(parser.values, option.dest, value)
+
+
+src: Callable[..., Option] = partial(
+    PipOption,
+    "--src",
+    "--source",
+    "--source-dir",
+    "--source-directory",
+    dest="src_dir",
+    type="path",
+    metavar="dir",
+    default=get_src_prefix(),
+    action="callback",
+    callback=_handle_src,
+    help="Directory to check out editable projects into. "
+    'The default in a virtualenv is "<venv path>/src". '
+    'The default for global installs is "<current dir>/src".',
+)
+
+
+def _get_format_control(values: Values, option: Option) -> Any:
+    """Get a format_control object."""
+    return getattr(values, option.dest)
+
+
+def _handle_no_binary(
+    option: Option, opt_str: str, value: str, parser: OptionParser
+) -> None:
+    existing = _get_format_control(parser.values, option)
+    FormatControl.handle_mutual_excludes(
+        value,
+        existing.no_binary,
+        existing.only_binary,
+    )
+
+
+def _handle_only_binary(
+    option: Option, opt_str: str, value: str, parser: OptionParser
+) -> None:
+    existing = _get_format_control(parser.values, option)
+    FormatControl.handle_mutual_excludes(
+        value,
+        existing.only_binary,
+        existing.no_binary,
+    )
+
+
+def no_binary() -> Option:
+    format_control = FormatControl(set(), set())
+    return Option(
+        "--no-binary",
+        dest="format_control",
+        action="callback",
+        callback=_handle_no_binary,
+        type="str",
+        default=format_control,
+        help="Do not use binary packages. Can be supplied multiple times, and "
+        'each time adds to the existing value. Accepts either ":all:" to '
+        'disable all binary packages, ":none:" to empty the set (notice '
+        "the colons), or one or more package names with commas between "
+        "them (no colons). Note that some packages are tricky to compile "
+        "and may fail to install when this option is used on them.",
+    )
+
+
+def only_binary() -> Option:
+    format_control = FormatControl(set(), set())
+    return Option(
+        "--only-binary",
+        dest="format_control",
+        action="callback",
+        callback=_handle_only_binary,
+        type="str",
+        default=format_control,
+        help="Do not use source packages. Can be supplied multiple times, and "
+        'each time adds to the existing value. Accepts either ":all:" to '
+        'disable all source packages, ":none:" to empty the set, or one '
+        "or more package names with commas between them. Packages "
+        "without binary distributions will fail to install when this "
+        "option is used on them.",
+    )
+
+
+platforms: Callable[..., Option] = partial(
+    Option,
+    "--platform",
+    dest="platforms",
+    metavar="platform",
+    action="append",
+    default=None,
+    help=(
+        "Only use wheels compatible with <platform>. Defaults to the "
+        "platform of the running system. Use this option multiple times to "
+        "specify multiple platforms supported by the target interpreter."
+    ),
+)
+
+
+# This was made a separate function for unit-testing purposes.
+def _convert_python_version(value: str) -> Tuple[Tuple[int, ...], Optional[str]]:
+    """
+    Convert a version string like "3", "37", or "3.7.3" into a tuple of ints.
+
+    :return: A 2-tuple (version_info, error_msg), where `error_msg` is
+        non-None if and only if there was a parsing error.
+    """
+    if not value:
+        # The empty string is the same as not providing a value.
+        return (None, None)
+
+    parts = value.split(".")
+    if len(parts) > 3:
+        return ((), "at most three version parts are allowed")
+
+    if len(parts) == 1:
+        # Then we are in the case of "3" or "37".
+        value = parts[0]
+        if len(value) > 1:
+            parts = [value[0], value[1:]]
+
+    try:
+        version_info = tuple(int(part) for part in parts)
+    except ValueError:
+        return ((), "each version part must be an integer")
+
+    return (version_info, None)
+
+
+def _handle_python_version(
+    option: Option, opt_str: str, value: str, parser: OptionParser
+) -> None:
+    """
+    Handle a provided --python-version value.
+    """
+    version_info, error_msg = _convert_python_version(value)
+    if error_msg is not None:
+        msg = f"invalid --python-version value: {value!r}: {error_msg}"
+        raise_option_error(parser, option=option, msg=msg)
+
+    parser.values.python_version = version_info
+
+
+python_version: Callable[..., Option] = partial(
+    Option,
+    "--python-version",
+    dest="python_version",
+    metavar="python_version",
+    action="callback",
+    callback=_handle_python_version,
+    type="str",
+    default=None,
+    help=dedent(
+        """\
+    The Python interpreter version to use for wheel and "Requires-Python"
+    compatibility checks. Defaults to a version derived from the running
+    interpreter. The version can be specified using up to three dot-separated
+    integers (e.g. "3" for 3.0.0, "3.7" for 3.7.0, or "3.7.3"). A major-minor
+    version can also be given as a string without dots (e.g. "37" for 3.7.0).
+    """
+    ),
+)
+
+
+implementation: Callable[..., Option] = partial(
+    Option,
+    "--implementation",
+    dest="implementation",
+    metavar="implementation",
+    default=None,
+    help=(
+        "Only use wheels compatible with Python "
+        "implementation <implementation>, e.g. 'pp', 'jy', 'cp', "
+        " or 'ip'. If not specified, then the current "
+        "interpreter implementation is used.  Use 'py' to force "
+        "implementation-agnostic wheels."
+    ),
+)
+
+
+abis: Callable[..., Option] = partial(
+    Option,
+    "--abi",
+    dest="abis",
+    metavar="abi",
+    action="append",
+    default=None,
+    help=(
+        "Only use wheels compatible with Python abi <abi>, e.g. 'pypy_41'. "
+        "If not specified, then the current interpreter abi tag is used. "
+        "Use this option multiple times to specify multiple abis supported "
+        "by the target interpreter. Generally you will need to specify "
+        "--implementation, --platform, and --python-version when using this "
+        "option."
+    ),
+)
+
+
+def add_target_python_options(cmd_opts: OptionGroup) -> None:
+    cmd_opts.add_option(platforms())
+    cmd_opts.add_option(python_version())
+    cmd_opts.add_option(implementation())
+    cmd_opts.add_option(abis())
+
+
+def make_target_python(options: Values) -> TargetPython:
+    target_python = TargetPython(
+        platforms=options.platforms,
+        py_version_info=options.python_version,
+        abis=options.abis,
+        implementation=options.implementation,
+    )
+
+    return target_python
+
+
+def prefer_binary() -> Option:
+    return Option(
+        "--prefer-binary",
+        dest="prefer_binary",
+        action="store_true",
+        default=False,
+        help=(
+            "Prefer binary packages over source packages, even if the "
+            "source packages are newer."
+        ),
+    )
+
+
+cache_dir: Callable[..., Option] = partial(
+    PipOption,
+    "--cache-dir",
+    dest="cache_dir",
+    default=USER_CACHE_DIR,
+    metavar="dir",
+    type="path",
+    help="Store the cache data in <dir>.",
+)
+
+
+def _handle_no_cache_dir(
+    option: Option, opt: str, value: str, parser: OptionParser
+) -> None:
+    """
+    Process a value provided for the --no-cache-dir option.
+
+    This is an optparse.Option callback for the --no-cache-dir option.
+    """
+    # The value argument will be None if --no-cache-dir is passed via the
+    # command-line, since the option doesn't accept arguments.  However,
+    # the value can be non-None if the option is triggered e.g. by an
+    # environment variable, like PIP_NO_CACHE_DIR=true.
+    if value is not None:
+        # Then parse the string value to get argument error-checking.
+        try:
+            strtobool(value)
+        except ValueError as exc:
+            raise_option_error(parser, option=option, msg=str(exc))
+
+    # Originally, setting PIP_NO_CACHE_DIR to a value that strtobool()
+    # converted to 0 (like "false" or "no") caused cache_dir to be disabled
+    # rather than enabled (logic would say the latter).  Thus, we disable
+    # the cache directory not just on values that parse to True, but (for
+    # backwards compatibility reasons) also on values that parse to False.
+    # In other words, always set it to False if the option is provided in
+    # some (valid) form.
+    parser.values.cache_dir = False
+
+
+no_cache: Callable[..., Option] = partial(
+    Option,
+    "--no-cache-dir",
+    dest="cache_dir",
+    action="callback",
+    callback=_handle_no_cache_dir,
+    help="Disable the cache.",
+)
+
+no_deps: Callable[..., Option] = partial(
+    Option,
+    "--no-deps",
+    "--no-dependencies",
+    dest="ignore_dependencies",
+    action="store_true",
+    default=False,
+    help="Don't install package dependencies.",
+)
+
+ignore_requires_python: Callable[..., Option] = partial(
+    Option,
+    "--ignore-requires-python",
+    dest="ignore_requires_python",
+    action="store_true",
+    help="Ignore the Requires-Python information.",
+)
+
+no_build_isolation: Callable[..., Option] = partial(
+    Option,
+    "--no-build-isolation",
+    dest="build_isolation",
+    action="store_false",
+    default=True,
+    help="Disable isolation when building a modern source distribution. "
+    "Build dependencies specified by PEP 518 must be already installed "
+    "if this option is used.",
+)
+
+check_build_deps: Callable[..., Option] = partial(
+    Option,
+    "--check-build-dependencies",
+    dest="check_build_deps",
+    action="store_true",
+    default=False,
+    help="Check the build dependencies when PEP517 is used.",
+)
+
+
+def _handle_no_use_pep517(
+    option: Option, opt: str, value: str, parser: OptionParser
+) -> None:
+    """
+    Process a value provided for the --no-use-pep517 option.
+
+    This is an optparse.Option callback for the no_use_pep517 option.
+    """
+    # Since --no-use-pep517 doesn't accept arguments, the value argument
+    # will be None if --no-use-pep517 is passed via the command-line.
+    # However, the value can be non-None if the option is triggered e.g.
+    # by an environment variable, for example "PIP_NO_USE_PEP517=true".
+    if value is not None:
+        msg = """A value was passed for --no-use-pep517,
+        probably using either the PIP_NO_USE_PEP517 environment variable
+        or the "no-use-pep517" config file option. Use an appropriate value
+        of the PIP_USE_PEP517 environment variable or the "use-pep517"
+        config file option instead.
+        """
+        raise_option_error(parser, option=option, msg=msg)
+
+    # If user doesn't wish to use pep517, we check if setuptools and wheel are installed
+    # and raise error if it is not.
+    packages = ("setuptools", "wheel")
+    if not all(importlib.util.find_spec(package) for package in packages):
+        msg = (
+            f"It is not possible to use --no-use-pep517 "
+            f"without {' and '.join(packages)} installed."
+        )
+        raise_option_error(parser, option=option, msg=msg)
+
+    # Otherwise, --no-use-pep517 was passed via the command-line.
+    parser.values.use_pep517 = False
+
+
+use_pep517: Any = partial(
+    Option,
+    "--use-pep517",
+    dest="use_pep517",
+    action="store_true",
+    default=None,
+    help="Use PEP 517 for building source distributions "
+    "(use --no-use-pep517 to force legacy behaviour).",
+)
+
+no_use_pep517: Any = partial(
+    Option,
+    "--no-use-pep517",
+    dest="use_pep517",
+    action="callback",
+    callback=_handle_no_use_pep517,
+    default=None,
+    help=SUPPRESS_HELP,
+)
+
+
+def _handle_config_settings(
+    option: Option, opt_str: str, value: str, parser: OptionParser
+) -> None:
+    key, sep, val = value.partition("=")
+    if sep != "=":
+        parser.error(f"Arguments to {opt_str} must be of the form KEY=VAL")
+    dest = getattr(parser.values, option.dest)
+    if dest is None:
+        dest = {}
+        setattr(parser.values, option.dest, dest)
+    if key in dest:
+        if isinstance(dest[key], list):
+            dest[key].append(val)
+        else:
+            dest[key] = [dest[key], val]
+    else:
+        dest[key] = val
+
+
+config_settings: Callable[..., Option] = partial(
+    Option,
+    "-C",
+    "--config-settings",
+    dest="config_settings",
+    type=str,
+    action="callback",
+    callback=_handle_config_settings,
+    metavar="settings",
+    help="Configuration settings to be passed to the PEP 517 build backend. "
+    "Settings take the form KEY=VALUE. Use multiple --config-settings options "
+    "to pass multiple keys to the backend.",
+)
+
+build_options: Callable[..., Option] = partial(
+    Option,
+    "--build-option",
+    dest="build_options",
+    metavar="options",
+    action="append",
+    help="Extra arguments to be supplied to 'setup.py bdist_wheel'.",
+)
+
+global_options: Callable[..., Option] = partial(
+    Option,
+    "--global-option",
+    dest="global_options",
+    action="append",
+    metavar="options",
+    help="Extra global options to be supplied to the setup.py "
+    "call before the install or bdist_wheel command.",
+)
+
+no_clean: Callable[..., Option] = partial(
+    Option,
+    "--no-clean",
+    action="store_true",
+    default=False,
+    help="Don't clean up build directories.",
+)
+
+pre: Callable[..., Option] = partial(
+    Option,
+    "--pre",
+    action="store_true",
+    default=False,
+    help="Include pre-release and development versions. By default, "
+    "pip only finds stable versions.",
+)
+
+disable_pip_version_check: Callable[..., Option] = partial(
+    Option,
+    "--disable-pip-version-check",
+    dest="disable_pip_version_check",
+    action="store_true",
+    default=False,
+    help="Don't periodically check PyPI to determine whether a new version "
+    "of pip is available for download. Implied with --no-index.",
+)
+
+root_user_action: Callable[..., Option] = partial(
+    Option,
+    "--root-user-action",
+    dest="root_user_action",
+    default="warn",
+    choices=["warn", "ignore"],
+    help="Action if pip is run as a root user [warn, ignore] (default: warn)",
+)
+
+
+def _handle_merge_hash(
+    option: Option, opt_str: str, value: str, parser: OptionParser
+) -> None:
+    """Given a value spelled "algo:digest", append the digest to a list
+    pointed to in a dict by the algo name."""
+    if not parser.values.hashes:
+        parser.values.hashes = {}
+    try:
+        algo, digest = value.split(":", 1)
+    except ValueError:
+        parser.error(
+            f"Arguments to {opt_str} must be a hash name "
+            "followed by a value, like --hash=sha256:"
+            "abcde..."
+        )
+    if algo not in STRONG_HASHES:
+        parser.error(
+            "Allowed hash algorithms for {} are {}.".format(
+                opt_str, ", ".join(STRONG_HASHES)
+            )
+        )
+    parser.values.hashes.setdefault(algo, []).append(digest)
+
+
+hash: Callable[..., Option] = partial(
+    Option,
+    "--hash",
+    # Hash values eventually end up in InstallRequirement.hashes due to
+    # __dict__ copying in process_line().
+    dest="hashes",
+    action="callback",
+    callback=_handle_merge_hash,
+    type="string",
+    help="Verify that the package's archive matches this "
+    "hash before installing. Example: --hash=sha256:abcdef...",
+)
+
+
+require_hashes: Callable[..., Option] = partial(
+    Option,
+    "--require-hashes",
+    dest="require_hashes",
+    action="store_true",
+    default=False,
+    help="Require a hash to check each requirement against, for "
+    "repeatable installs. This option is implied when any package in a "
+    "requirements file has a --hash option.",
+)
+
+
+list_path: Callable[..., Option] = partial(
+    PipOption,
+    "--path",
+    dest="path",
+    type="path",
+    action="append",
+    help="Restrict to the specified installation path for listing "
+    "packages (can be used multiple times).",
+)
+
+
+def check_list_path_option(options: Values) -> None:
+    if options.path and (options.user or options.local):
+        raise CommandError("Cannot combine '--path' with '--user' or '--local'")
+
+
+list_exclude: Callable[..., Option] = partial(
+    PipOption,
+    "--exclude",
+    dest="excludes",
+    action="append",
+    metavar="package",
+    type="package_name",
+    help="Exclude specified package from the output",
+)
+
+
+no_python_version_warning: Callable[..., Option] = partial(
+    Option,
+    "--no-python-version-warning",
+    dest="no_python_version_warning",
+    action="store_true",
+    default=False,
+    help="Silence deprecation warnings for upcoming unsupported Pythons.",
+)
+
+
+# Features that are now always on. A warning is printed if they are used.
+ALWAYS_ENABLED_FEATURES = [
+    "truststore",  # always on since 24.2
+    "no-binary-enable-wheel-cache",  # always on since 23.1
+]
+
+use_new_feature: Callable[..., Option] = partial(
+    Option,
+    "--use-feature",
+    dest="features_enabled",
+    metavar="feature",
+    action="append",
+    default=[],
+    choices=[
+        "fast-deps",
+    ]
+    + ALWAYS_ENABLED_FEATURES,
+    help="Enable new functionality, that may be backward incompatible.",
+)
+
+use_deprecated_feature: Callable[..., Option] = partial(
+    Option,
+    "--use-deprecated",
+    dest="deprecated_features_enabled",
+    metavar="feature",
+    action="append",
+    default=[],
+    choices=[
+        "legacy-resolver",
+        "legacy-certs",
+    ],
+    help=("Enable deprecated functionality, that will be removed in the future."),
+)
+
+
+##########
+# groups #
+##########
+
+general_group: Dict[str, Any] = {
+    "name": "General Options",
+    "options": [
+        help_,
+        debug_mode,
+        isolated_mode,
+        require_virtualenv,
+        python,
+        verbose,
+        version,
+        quiet,
+        log,
+        no_input,
+        keyring_provider,
+        proxy,
+        retries,
+        timeout,
+        exists_action,
+        trusted_host,
+        cert,
+        client_cert,
+        cache_dir,
+        no_cache,
+        disable_pip_version_check,
+        no_color,
+        no_python_version_warning,
+        use_new_feature,
+        use_deprecated_feature,
+    ],
+}
+
+index_group: Dict[str, Any] = {
+    "name": "Package Index Options",
+    "options": [
+        index_url,
+        extra_index_url,
+        no_index,
+        find_links,
+    ],
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/command_context.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/command_context.py
new file mode 100644
index 00000000..139995ac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/command_context.py
@@ -0,0 +1,27 @@
+from contextlib import ExitStack, contextmanager
+from typing import ContextManager, Generator, TypeVar
+
+_T = TypeVar("_T", covariant=True)
+
+
+class CommandContextMixIn:
+    def __init__(self) -> None:
+        super().__init__()
+        self._in_main_context = False
+        self._main_context = ExitStack()
+
+    @contextmanager
+    def main_context(self) -> Generator[None, None, None]:
+        assert not self._in_main_context
+
+        self._in_main_context = True
+        try:
+            with self._main_context:
+                yield
+        finally:
+            self._in_main_context = False
+
+    def enter_context(self, context_provider: ContextManager[_T]) -> _T:
+        assert self._in_main_context
+
+        return self._main_context.enter_context(context_provider)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/index_command.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/index_command.py
new file mode 100644
index 00000000..295108ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/index_command.py
@@ -0,0 +1,171 @@
+"""
+Contains command classes which may interact with an index / the network.
+
+Unlike its sister module, req_command, this module still uses lazy imports
+so commands which don't always hit the network (e.g. list w/o --outdated or
+--uptodate) don't need waste time importing PipSession and friends.
+"""
+
+import logging
+import os
+import sys
+from optparse import Values
+from typing import TYPE_CHECKING, List, Optional
+
+from pip._vendor import certifi
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.command_context import CommandContextMixIn
+
+if TYPE_CHECKING:
+    from ssl import SSLContext
+
+    from pip._internal.network.session import PipSession
+
+logger = logging.getLogger(__name__)
+
+
+def _create_truststore_ssl_context() -> Optional["SSLContext"]:
+    if sys.version_info < (3, 10):
+        logger.debug("Disabling truststore because Python version isn't 3.10+")
+        return None
+
+    try:
+        import ssl
+    except ImportError:
+        logger.warning("Disabling truststore since ssl support is missing")
+        return None
+
+    try:
+        from pip._vendor import truststore
+    except ImportError:
+        logger.warning("Disabling truststore because platform isn't supported")
+        return None
+
+    ctx = truststore.SSLContext(ssl.PROTOCOL_TLS_CLIENT)
+    ctx.load_verify_locations(certifi.where())
+    return ctx
+
+
+class SessionCommandMixin(CommandContextMixIn):
+    """
+    A class mixin for command classes needing _build_session().
+    """
+
+    def __init__(self) -> None:
+        super().__init__()
+        self._session: Optional[PipSession] = None
+
+    @classmethod
+    def _get_index_urls(cls, options: Values) -> Optional[List[str]]:
+        """Return a list of index urls from user-provided options."""
+        index_urls = []
+        if not getattr(options, "no_index", False):
+            url = getattr(options, "index_url", None)
+            if url:
+                index_urls.append(url)
+        urls = getattr(options, "extra_index_urls", None)
+        if urls:
+            index_urls.extend(urls)
+        # Return None rather than an empty list
+        return index_urls or None
+
+    def get_default_session(self, options: Values) -> "PipSession":
+        """Get a default-managed session."""
+        if self._session is None:
+            self._session = self.enter_context(self._build_session(options))
+            # there's no type annotation on requests.Session, so it's
+            # automatically ContextManager[Any] and self._session becomes Any,
+            # then https://github.com/python/mypy/issues/7696 kicks in
+            assert self._session is not None
+        return self._session
+
+    def _build_session(
+        self,
+        options: Values,
+        retries: Optional[int] = None,
+        timeout: Optional[int] = None,
+    ) -> "PipSession":
+        from pip._internal.network.session import PipSession
+
+        cache_dir = options.cache_dir
+        assert not cache_dir or os.path.isabs(cache_dir)
+
+        if "legacy-certs" not in options.deprecated_features_enabled:
+            ssl_context = _create_truststore_ssl_context()
+        else:
+            ssl_context = None
+
+        session = PipSession(
+            cache=os.path.join(cache_dir, "http-v2") if cache_dir else None,
+            retries=retries if retries is not None else options.retries,
+            trusted_hosts=options.trusted_hosts,
+            index_urls=self._get_index_urls(options),
+            ssl_context=ssl_context,
+        )
+
+        # Handle custom ca-bundles from the user
+        if options.cert:
+            session.verify = options.cert
+
+        # Handle SSL client certificate
+        if options.client_cert:
+            session.cert = options.client_cert
+
+        # Handle timeouts
+        if options.timeout or timeout:
+            session.timeout = timeout if timeout is not None else options.timeout
+
+        # Handle configured proxies
+        if options.proxy:
+            session.proxies = {
+                "http": options.proxy,
+                "https": options.proxy,
+            }
+            session.trust_env = False
+            session.pip_proxy = options.proxy
+
+        # Determine if we can prompt the user for authentication or not
+        session.auth.prompting = not options.no_input
+        session.auth.keyring_provider = options.keyring_provider
+
+        return session
+
+
+def _pip_self_version_check(session: "PipSession", options: Values) -> None:
+    from pip._internal.self_outdated_check import pip_self_version_check as check
+
+    check(session, options)
+
+
+class IndexGroupCommand(Command, SessionCommandMixin):
+    """
+    Abstract base class for commands with the index_group options.
+
+    This also corresponds to the commands that permit the pip version check.
+    """
+
+    def handle_pip_version_check(self, options: Values) -> None:
+        """
+        Do the pip version check if not disabled.
+
+        This overrides the default behavior of not doing the check.
+        """
+        # Make sure the index_group options are present.
+        assert hasattr(options, "no_index")
+
+        if options.disable_pip_version_check or options.no_index:
+            return
+
+        try:
+            # Otherwise, check if we're using the latest version of pip available.
+            session = self._build_session(
+                options,
+                retries=0,
+                timeout=min(5, options.timeout),
+            )
+            with session:
+                _pip_self_version_check(session, options)
+        except Exception:
+            logger.warning("There was an error checking the latest version of pip.")
+            logger.debug("See below for error", exc_info=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main.py
new file mode 100644
index 00000000..563ac79c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main.py
@@ -0,0 +1,80 @@
+"""Primary application entrypoint.
+"""
+
+import locale
+import logging
+import os
+import sys
+import warnings
+from typing import List, Optional
+
+from pip._internal.cli.autocompletion import autocomplete
+from pip._internal.cli.main_parser import parse_command
+from pip._internal.commands import create_command
+from pip._internal.exceptions import PipError
+from pip._internal.utils import deprecation
+
+logger = logging.getLogger(__name__)
+
+
+# Do not import and use main() directly! Using it directly is actively
+# discouraged by pip's maintainers. The name, location and behavior of
+# this function is subject to change, so calling it directly is not
+# portable across different pip versions.
+
+# In addition, running pip in-process is unsupported and unsafe. This is
+# elaborated in detail at
+# https://pip.pypa.io/en/stable/user_guide/#using-pip-from-your-program.
+# That document also provides suggestions that should work for nearly
+# all users that are considering importing and using main() directly.
+
+# However, we know that certain users will still want to invoke pip
+# in-process. If you understand and accept the implications of using pip
+# in an unsupported manner, the best approach is to use runpy to avoid
+# depending on the exact location of this entry point.
+
+# The following example shows how to use runpy to invoke pip in that
+# case:
+#
+#     sys.argv = ["pip", your, args, here]
+#     runpy.run_module("pip", run_name="__main__")
+#
+# Note that this will exit the process after running, unlike a direct
+# call to main. As it is not safe to do any processing after calling
+# main, this should not be an issue in practice.
+
+
+def main(args: Optional[List[str]] = None) -> int:
+    if args is None:
+        args = sys.argv[1:]
+
+    # Suppress the pkg_resources deprecation warning
+    # Note - we use a module of .*pkg_resources to cover
+    # the normal case (pip._vendor.pkg_resources) and the
+    # devendored case (a bare pkg_resources)
+    warnings.filterwarnings(
+        action="ignore", category=DeprecationWarning, module=".*pkg_resources"
+    )
+
+    # Configure our deprecation warnings to be sent through loggers
+    deprecation.install_warning_logger()
+
+    autocomplete()
+
+    try:
+        cmd_name, cmd_args = parse_command(args)
+    except PipError as exc:
+        sys.stderr.write(f"ERROR: {exc}")
+        sys.stderr.write(os.linesep)
+        sys.exit(1)
+
+    # Needed for locale.getpreferredencoding(False) to work
+    # in pip._internal.utils.encoding.auto_decode
+    try:
+        locale.setlocale(locale.LC_ALL, "")
+    except locale.Error as e:
+        # setlocale can apparently crash if locale are uninitialized
+        logger.debug("Ignoring error %s when setting locale", e)
+    command = create_command(cmd_name, isolated=("--isolated" in cmd_args))
+
+    return command.main(cmd_args)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main_parser.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main_parser.py
new file mode 100644
index 00000000..5ade356b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/main_parser.py
@@ -0,0 +1,134 @@
+"""A single place for constructing and exposing the main parser
+"""
+
+import os
+import subprocess
+import sys
+from typing import List, Optional, Tuple
+
+from pip._internal.build_env import get_runnable_pip
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.parser import ConfigOptionParser, UpdatingDefaultsHelpFormatter
+from pip._internal.commands import commands_dict, get_similar_commands
+from pip._internal.exceptions import CommandError
+from pip._internal.utils.misc import get_pip_version, get_prog
+
+__all__ = ["create_main_parser", "parse_command"]
+
+
+def create_main_parser() -> ConfigOptionParser:
+    """Creates and returns the main parser for pip's CLI"""
+
+    parser = ConfigOptionParser(
+        usage="\n%prog <command> [options]",
+        add_help_option=False,
+        formatter=UpdatingDefaultsHelpFormatter(),
+        name="global",
+        prog=get_prog(),
+    )
+    parser.disable_interspersed_args()
+
+    parser.version = get_pip_version()
+
+    # add the general options
+    gen_opts = cmdoptions.make_option_group(cmdoptions.general_group, parser)
+    parser.add_option_group(gen_opts)
+
+    # so the help formatter knows
+    parser.main = True  # type: ignore
+
+    # create command listing for description
+    description = [""] + [
+        f"{name:27} {command_info.summary}"
+        for name, command_info in commands_dict.items()
+    ]
+    parser.description = "\n".join(description)
+
+    return parser
+
+
+def identify_python_interpreter(python: str) -> Optional[str]:
+    # If the named file exists, use it.
+    # If it's a directory, assume it's a virtual environment and
+    # look for the environment's Python executable.
+    if os.path.exists(python):
+        if os.path.isdir(python):
+            # bin/python for Unix, Scripts/python.exe for Windows
+            # Try both in case of odd cases like cygwin.
+            for exe in ("bin/python", "Scripts/python.exe"):
+                py = os.path.join(python, exe)
+                if os.path.exists(py):
+                    return py
+        else:
+            return python
+
+    # Could not find the interpreter specified
+    return None
+
+
+def parse_command(args: List[str]) -> Tuple[str, List[str]]:
+    parser = create_main_parser()
+
+    # Note: parser calls disable_interspersed_args(), so the result of this
+    # call is to split the initial args into the general options before the
+    # subcommand and everything else.
+    # For example:
+    #  args: ['--timeout=5', 'install', '--user', 'INITools']
+    #  general_options: ['--timeout==5']
+    #  args_else: ['install', '--user', 'INITools']
+    general_options, args_else = parser.parse_args(args)
+
+    # --python
+    if general_options.python and "_PIP_RUNNING_IN_SUBPROCESS" not in os.environ:
+        # Re-invoke pip using the specified Python interpreter
+        interpreter = identify_python_interpreter(general_options.python)
+        if interpreter is None:
+            raise CommandError(
+                f"Could not locate Python interpreter {general_options.python}"
+            )
+
+        pip_cmd = [
+            interpreter,
+            get_runnable_pip(),
+        ]
+        pip_cmd.extend(args)
+
+        # Set a flag so the child doesn't re-invoke itself, causing
+        # an infinite loop.
+        os.environ["_PIP_RUNNING_IN_SUBPROCESS"] = "1"
+        returncode = 0
+        try:
+            proc = subprocess.run(pip_cmd)
+            returncode = proc.returncode
+        except (subprocess.SubprocessError, OSError) as exc:
+            raise CommandError(f"Failed to run pip under {interpreter}: {exc}")
+        sys.exit(returncode)
+
+    # --version
+    if general_options.version:
+        sys.stdout.write(parser.version)
+        sys.stdout.write(os.linesep)
+        sys.exit()
+
+    # pip || pip help -> print_help()
+    if not args_else or (args_else[0] == "help" and len(args_else) == 1):
+        parser.print_help()
+        sys.exit()
+
+    # the subcommand name
+    cmd_name = args_else[0]
+
+    if cmd_name not in commands_dict:
+        guess = get_similar_commands(cmd_name)
+
+        msg = [f'unknown command "{cmd_name}"']
+        if guess:
+            msg.append(f'maybe you meant "{guess}"')
+
+        raise CommandError(" - ".join(msg))
+
+    # all the args without the subcommand
+    cmd_args = args[:]
+    cmd_args.remove(cmd_name)
+
+    return cmd_name, cmd_args
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/parser.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/parser.py
new file mode 100644
index 00000000..bc4aca03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/parser.py
@@ -0,0 +1,294 @@
+"""Base option parser setup"""
+
+import logging
+import optparse
+import shutil
+import sys
+import textwrap
+from contextlib import suppress
+from typing import Any, Dict, Generator, List, NoReturn, Optional, Tuple
+
+from pip._internal.cli.status_codes import UNKNOWN_ERROR
+from pip._internal.configuration import Configuration, ConfigurationError
+from pip._internal.utils.misc import redact_auth_from_url, strtobool
+
+logger = logging.getLogger(__name__)
+
+
+class PrettyHelpFormatter(optparse.IndentedHelpFormatter):
+    """A prettier/less verbose help formatter for optparse."""
+
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        # help position must be aligned with __init__.parseopts.description
+        kwargs["max_help_position"] = 30
+        kwargs["indent_increment"] = 1
+        kwargs["width"] = shutil.get_terminal_size()[0] - 2
+        super().__init__(*args, **kwargs)
+
+    def format_option_strings(self, option: optparse.Option) -> str:
+        return self._format_option_strings(option)
+
+    def _format_option_strings(
+        self, option: optparse.Option, mvarfmt: str = " <{}>", optsep: str = ", "
+    ) -> str:
+        """
+        Return a comma-separated list of option strings and metavars.
+
+        :param option:  tuple of (short opt, long opt), e.g: ('-f', '--format')
+        :param mvarfmt: metavar format string
+        :param optsep:  separator
+        """
+        opts = []
+
+        if option._short_opts:
+            opts.append(option._short_opts[0])
+        if option._long_opts:
+            opts.append(option._long_opts[0])
+        if len(opts) > 1:
+            opts.insert(1, optsep)
+
+        if option.takes_value():
+            assert option.dest is not None
+            metavar = option.metavar or option.dest.lower()
+            opts.append(mvarfmt.format(metavar.lower()))
+
+        return "".join(opts)
+
+    def format_heading(self, heading: str) -> str:
+        if heading == "Options":
+            return ""
+        return heading + ":\n"
+
+    def format_usage(self, usage: str) -> str:
+        """
+        Ensure there is only one newline between usage and the first heading
+        if there is no description.
+        """
+        msg = "\nUsage: {}\n".format(self.indent_lines(textwrap.dedent(usage), "  "))
+        return msg
+
+    def format_description(self, description: Optional[str]) -> str:
+        # leave full control over description to us
+        if description:
+            if hasattr(self.parser, "main"):
+                label = "Commands"
+            else:
+                label = "Description"
+            # some doc strings have initial newlines, some don't
+            description = description.lstrip("\n")
+            # some doc strings have final newlines and spaces, some don't
+            description = description.rstrip()
+            # dedent, then reindent
+            description = self.indent_lines(textwrap.dedent(description), "  ")
+            description = f"{label}:\n{description}\n"
+            return description
+        else:
+            return ""
+
+    def format_epilog(self, epilog: Optional[str]) -> str:
+        # leave full control over epilog to us
+        if epilog:
+            return epilog
+        else:
+            return ""
+
+    def indent_lines(self, text: str, indent: str) -> str:
+        new_lines = [indent + line for line in text.split("\n")]
+        return "\n".join(new_lines)
+
+
+class UpdatingDefaultsHelpFormatter(PrettyHelpFormatter):
+    """Custom help formatter for use in ConfigOptionParser.
+
+    This is updates the defaults before expanding them, allowing
+    them to show up correctly in the help listing.
+
+    Also redact auth from url type options
+    """
+
+    def expand_default(self, option: optparse.Option) -> str:
+        default_values = None
+        if self.parser is not None:
+            assert isinstance(self.parser, ConfigOptionParser)
+            self.parser._update_defaults(self.parser.defaults)
+            assert option.dest is not None
+            default_values = self.parser.defaults.get(option.dest)
+        help_text = super().expand_default(option)
+
+        if default_values and option.metavar == "URL":
+            if isinstance(default_values, str):
+                default_values = [default_values]
+
+            # If its not a list, we should abort and just return the help text
+            if not isinstance(default_values, list):
+                default_values = []
+
+            for val in default_values:
+                help_text = help_text.replace(val, redact_auth_from_url(val))
+
+        return help_text
+
+
+class CustomOptionParser(optparse.OptionParser):
+    def insert_option_group(
+        self, idx: int, *args: Any, **kwargs: Any
+    ) -> optparse.OptionGroup:
+        """Insert an OptionGroup at a given position."""
+        group = self.add_option_group(*args, **kwargs)
+
+        self.option_groups.pop()
+        self.option_groups.insert(idx, group)
+
+        return group
+
+    @property
+    def option_list_all(self) -> List[optparse.Option]:
+        """Get a list of all options, including those in option groups."""
+        res = self.option_list[:]
+        for i in self.option_groups:
+            res.extend(i.option_list)
+
+        return res
+
+
+class ConfigOptionParser(CustomOptionParser):
+    """Custom option parser which updates its defaults by checking the
+    configuration files and environmental variables"""
+
+    def __init__(
+        self,
+        *args: Any,
+        name: str,
+        isolated: bool = False,
+        **kwargs: Any,
+    ) -> None:
+        self.name = name
+        self.config = Configuration(isolated)
+
+        assert self.name
+        super().__init__(*args, **kwargs)
+
+    def check_default(self, option: optparse.Option, key: str, val: Any) -> Any:
+        try:
+            return option.check_value(key, val)
+        except optparse.OptionValueError as exc:
+            print(f"An error occurred during configuration: {exc}")
+            sys.exit(3)
+
+    def _get_ordered_configuration_items(
+        self,
+    ) -> Generator[Tuple[str, Any], None, None]:
+        # Configuration gives keys in an unordered manner. Order them.
+        override_order = ["global", self.name, ":env:"]
+
+        # Pool the options into different groups
+        section_items: Dict[str, List[Tuple[str, Any]]] = {
+            name: [] for name in override_order
+        }
+        for section_key, val in self.config.items():
+            # ignore empty values
+            if not val:
+                logger.debug(
+                    "Ignoring configuration key '%s' as it's value is empty.",
+                    section_key,
+                )
+                continue
+
+            section, key = section_key.split(".", 1)
+            if section in override_order:
+                section_items[section].append((key, val))
+
+        # Yield each group in their override order
+        for section in override_order:
+            for key, val in section_items[section]:
+                yield key, val
+
+    def _update_defaults(self, defaults: Dict[str, Any]) -> Dict[str, Any]:
+        """Updates the given defaults with values from the config files and
+        the environ. Does a little special handling for certain types of
+        options (lists)."""
+
+        # Accumulate complex default state.
+        self.values = optparse.Values(self.defaults)
+        late_eval = set()
+        # Then set the options with those values
+        for key, val in self._get_ordered_configuration_items():
+            # '--' because configuration supports only long names
+            option = self.get_option("--" + key)
+
+            # Ignore options not present in this parser. E.g. non-globals put
+            # in [global] by users that want them to apply to all applicable
+            # commands.
+            if option is None:
+                continue
+
+            assert option.dest is not None
+
+            if option.action in ("store_true", "store_false"):
+                try:
+                    val = strtobool(val)
+                except ValueError:
+                    self.error(
+                        f"{val} is not a valid value for {key} option, "
+                        "please specify a boolean value like yes/no, "
+                        "true/false or 1/0 instead."
+                    )
+            elif option.action == "count":
+                with suppress(ValueError):
+                    val = strtobool(val)
+                with suppress(ValueError):
+                    val = int(val)
+                if not isinstance(val, int) or val < 0:
+                    self.error(
+                        f"{val} is not a valid value for {key} option, "
+                        "please instead specify either a non-negative integer "
+                        "or a boolean value like yes/no or false/true "
+                        "which is equivalent to 1/0."
+                    )
+            elif option.action == "append":
+                val = val.split()
+                val = [self.check_default(option, key, v) for v in val]
+            elif option.action == "callback":
+                assert option.callback is not None
+                late_eval.add(option.dest)
+                opt_str = option.get_opt_string()
+                val = option.convert_value(opt_str, val)
+                # From take_action
+                args = option.callback_args or ()
+                kwargs = option.callback_kwargs or {}
+                option.callback(option, opt_str, val, self, *args, **kwargs)
+            else:
+                val = self.check_default(option, key, val)
+
+            defaults[option.dest] = val
+
+        for key in late_eval:
+            defaults[key] = getattr(self.values, key)
+        self.values = None
+        return defaults
+
+    def get_default_values(self) -> optparse.Values:
+        """Overriding to make updating the defaults after instantiation of
+        the option parser possible, _update_defaults() does the dirty work."""
+        if not self.process_default_values:
+            # Old, pre-Optik 1.5 behaviour.
+            return optparse.Values(self.defaults)
+
+        # Load the configuration, or error out in case of an error
+        try:
+            self.config.load()
+        except ConfigurationError as err:
+            self.exit(UNKNOWN_ERROR, str(err))
+
+        defaults = self._update_defaults(self.defaults.copy())  # ours
+        for option in self._get_all_options():
+            assert option.dest is not None
+            default = defaults.get(option.dest)
+            if isinstance(default, str):
+                opt_str = option.get_opt_string()
+                defaults[option.dest] = option.check_value(opt_str, default)
+        return optparse.Values(defaults)
+
+    def error(self, msg: str) -> NoReturn:
+        self.print_usage(sys.stderr)
+        self.exit(UNKNOWN_ERROR, f"{msg}\n")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/progress_bars.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/progress_bars.py
new file mode 100644
index 00000000..3d9dde8e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/progress_bars.py
@@ -0,0 +1,94 @@
+import functools
+import sys
+from typing import Callable, Generator, Iterable, Iterator, Optional, Tuple
+
+from pip._vendor.rich.progress import (
+    BarColumn,
+    DownloadColumn,
+    FileSizeColumn,
+    Progress,
+    ProgressColumn,
+    SpinnerColumn,
+    TextColumn,
+    TimeElapsedColumn,
+    TimeRemainingColumn,
+    TransferSpeedColumn,
+)
+
+from pip._internal.cli.spinners import RateLimiter
+from pip._internal.utils.logging import get_indentation
+
+DownloadProgressRenderer = Callable[[Iterable[bytes]], Iterator[bytes]]
+
+
+def _rich_progress_bar(
+    iterable: Iterable[bytes],
+    *,
+    bar_type: str,
+    size: Optional[int],
+) -> Generator[bytes, None, None]:
+    assert bar_type == "on", "This should only be used in the default mode."
+
+    if not size:
+        total = float("inf")
+        columns: Tuple[ProgressColumn, ...] = (
+            TextColumn("[progress.description]{task.description}"),
+            SpinnerColumn("line", speed=1.5),
+            FileSizeColumn(),
+            TransferSpeedColumn(),
+            TimeElapsedColumn(),
+        )
+    else:
+        total = size
+        columns = (
+            TextColumn("[progress.description]{task.description}"),
+            BarColumn(),
+            DownloadColumn(),
+            TransferSpeedColumn(),
+            TextColumn("eta"),
+            TimeRemainingColumn(),
+        )
+
+    progress = Progress(*columns, refresh_per_second=5)
+    task_id = progress.add_task(" " * (get_indentation() + 2), total=total)
+    with progress:
+        for chunk in iterable:
+            yield chunk
+            progress.update(task_id, advance=len(chunk))
+
+
+def _raw_progress_bar(
+    iterable: Iterable[bytes],
+    *,
+    size: Optional[int],
+) -> Generator[bytes, None, None]:
+    def write_progress(current: int, total: int) -> None:
+        sys.stdout.write(f"Progress {current} of {total}\n")
+        sys.stdout.flush()
+
+    current = 0
+    total = size or 0
+    rate_limiter = RateLimiter(0.25)
+
+    write_progress(current, total)
+    for chunk in iterable:
+        current += len(chunk)
+        if rate_limiter.ready() or current == total:
+            write_progress(current, total)
+            rate_limiter.reset()
+        yield chunk
+
+
+def get_download_progress_renderer(
+    *, bar_type: str, size: Optional[int] = None
+) -> DownloadProgressRenderer:
+    """Get an object that can be used to render the download progress.
+
+    Returns a callable, that takes an iterable to "wrap".
+    """
+    if bar_type == "on":
+        return functools.partial(_rich_progress_bar, bar_type=bar_type, size=size)
+    elif bar_type == "raw":
+        return functools.partial(_raw_progress_bar, size=size)
+    else:
+        return iter  # no-op, when passed an iterator
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/req_command.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/req_command.py
new file mode 100644
index 00000000..92900f94
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/req_command.py
@@ -0,0 +1,329 @@
+"""Contains the RequirementCommand base class.
+
+This class is in a separate module so the commands that do not always
+need PackageFinder capability don't unnecessarily import the
+PackageFinder machinery and all its vendored dependencies, etc.
+"""
+
+import logging
+from functools import partial
+from optparse import Values
+from typing import Any, List, Optional, Tuple
+
+from pip._internal.cache import WheelCache
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.index_command import IndexGroupCommand
+from pip._internal.cli.index_command import SessionCommandMixin as SessionCommandMixin
+from pip._internal.exceptions import CommandError, PreviousBuildDirError
+from pip._internal.index.collector import LinkCollector
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.models.selection_prefs import SelectionPreferences
+from pip._internal.models.target_python import TargetPython
+from pip._internal.network.session import PipSession
+from pip._internal.operations.build.build_tracker import BuildTracker
+from pip._internal.operations.prepare import RequirementPreparer
+from pip._internal.req.constructors import (
+    install_req_from_editable,
+    install_req_from_line,
+    install_req_from_parsed_requirement,
+    install_req_from_req_string,
+)
+from pip._internal.req.req_file import parse_requirements
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.resolution.base import BaseResolver
+from pip._internal.utils.temp_dir import (
+    TempDirectory,
+    TempDirectoryTypeRegistry,
+    tempdir_kinds,
+)
+
+logger = logging.getLogger(__name__)
+
+
+KEEPABLE_TEMPDIR_TYPES = [
+    tempdir_kinds.BUILD_ENV,
+    tempdir_kinds.EPHEM_WHEEL_CACHE,
+    tempdir_kinds.REQ_BUILD,
+]
+
+
+def with_cleanup(func: Any) -> Any:
+    """Decorator for common logic related to managing temporary
+    directories.
+    """
+
+    def configure_tempdir_registry(registry: TempDirectoryTypeRegistry) -> None:
+        for t in KEEPABLE_TEMPDIR_TYPES:
+            registry.set_delete(t, False)
+
+    def wrapper(
+        self: RequirementCommand, options: Values, args: List[Any]
+    ) -> Optional[int]:
+        assert self.tempdir_registry is not None
+        if options.no_clean:
+            configure_tempdir_registry(self.tempdir_registry)
+
+        try:
+            return func(self, options, args)
+        except PreviousBuildDirError:
+            # This kind of conflict can occur when the user passes an explicit
+            # build directory with a pre-existing folder. In that case we do
+            # not want to accidentally remove it.
+            configure_tempdir_registry(self.tempdir_registry)
+            raise
+
+    return wrapper
+
+
+class RequirementCommand(IndexGroupCommand):
+    def __init__(self, *args: Any, **kw: Any) -> None:
+        super().__init__(*args, **kw)
+
+        self.cmd_opts.add_option(cmdoptions.no_clean())
+
+    @staticmethod
+    def determine_resolver_variant(options: Values) -> str:
+        """Determines which resolver should be used, based on the given options."""
+        if "legacy-resolver" in options.deprecated_features_enabled:
+            return "legacy"
+
+        return "resolvelib"
+
+    @classmethod
+    def make_requirement_preparer(
+        cls,
+        temp_build_dir: TempDirectory,
+        options: Values,
+        build_tracker: BuildTracker,
+        session: PipSession,
+        finder: PackageFinder,
+        use_user_site: bool,
+        download_dir: Optional[str] = None,
+        verbosity: int = 0,
+    ) -> RequirementPreparer:
+        """
+        Create a RequirementPreparer instance for the given parameters.
+        """
+        temp_build_dir_path = temp_build_dir.path
+        assert temp_build_dir_path is not None
+        legacy_resolver = False
+
+        resolver_variant = cls.determine_resolver_variant(options)
+        if resolver_variant == "resolvelib":
+            lazy_wheel = "fast-deps" in options.features_enabled
+            if lazy_wheel:
+                logger.warning(
+                    "pip is using lazily downloaded wheels using HTTP "
+                    "range requests to obtain dependency information. "
+                    "This experimental feature is enabled through "
+                    "--use-feature=fast-deps and it is not ready for "
+                    "production."
+                )
+        else:
+            legacy_resolver = True
+            lazy_wheel = False
+            if "fast-deps" in options.features_enabled:
+                logger.warning(
+                    "fast-deps has no effect when used with the legacy resolver."
+                )
+
+        return RequirementPreparer(
+            build_dir=temp_build_dir_path,
+            src_dir=options.src_dir,
+            download_dir=download_dir,
+            build_isolation=options.build_isolation,
+            check_build_deps=options.check_build_deps,
+            build_tracker=build_tracker,
+            session=session,
+            progress_bar=options.progress_bar,
+            finder=finder,
+            require_hashes=options.require_hashes,
+            use_user_site=use_user_site,
+            lazy_wheel=lazy_wheel,
+            verbosity=verbosity,
+            legacy_resolver=legacy_resolver,
+        )
+
+    @classmethod
+    def make_resolver(
+        cls,
+        preparer: RequirementPreparer,
+        finder: PackageFinder,
+        options: Values,
+        wheel_cache: Optional[WheelCache] = None,
+        use_user_site: bool = False,
+        ignore_installed: bool = True,
+        ignore_requires_python: bool = False,
+        force_reinstall: bool = False,
+        upgrade_strategy: str = "to-satisfy-only",
+        use_pep517: Optional[bool] = None,
+        py_version_info: Optional[Tuple[int, ...]] = None,
+    ) -> BaseResolver:
+        """
+        Create a Resolver instance for the given parameters.
+        """
+        make_install_req = partial(
+            install_req_from_req_string,
+            isolated=options.isolated_mode,
+            use_pep517=use_pep517,
+        )
+        resolver_variant = cls.determine_resolver_variant(options)
+        # The long import name and duplicated invocation is needed to convince
+        # Mypy into correctly typechecking. Otherwise it would complain the
+        # "Resolver" class being redefined.
+        if resolver_variant == "resolvelib":
+            import pip._internal.resolution.resolvelib.resolver
+
+            return pip._internal.resolution.resolvelib.resolver.Resolver(
+                preparer=preparer,
+                finder=finder,
+                wheel_cache=wheel_cache,
+                make_install_req=make_install_req,
+                use_user_site=use_user_site,
+                ignore_dependencies=options.ignore_dependencies,
+                ignore_installed=ignore_installed,
+                ignore_requires_python=ignore_requires_python,
+                force_reinstall=force_reinstall,
+                upgrade_strategy=upgrade_strategy,
+                py_version_info=py_version_info,
+            )
+        import pip._internal.resolution.legacy.resolver
+
+        return pip._internal.resolution.legacy.resolver.Resolver(
+            preparer=preparer,
+            finder=finder,
+            wheel_cache=wheel_cache,
+            make_install_req=make_install_req,
+            use_user_site=use_user_site,
+            ignore_dependencies=options.ignore_dependencies,
+            ignore_installed=ignore_installed,
+            ignore_requires_python=ignore_requires_python,
+            force_reinstall=force_reinstall,
+            upgrade_strategy=upgrade_strategy,
+            py_version_info=py_version_info,
+        )
+
+    def get_requirements(
+        self,
+        args: List[str],
+        options: Values,
+        finder: PackageFinder,
+        session: PipSession,
+    ) -> List[InstallRequirement]:
+        """
+        Parse command-line arguments into the corresponding requirements.
+        """
+        requirements: List[InstallRequirement] = []
+        for filename in options.constraints:
+            for parsed_req in parse_requirements(
+                filename,
+                constraint=True,
+                finder=finder,
+                options=options,
+                session=session,
+            ):
+                req_to_add = install_req_from_parsed_requirement(
+                    parsed_req,
+                    isolated=options.isolated_mode,
+                    user_supplied=False,
+                )
+                requirements.append(req_to_add)
+
+        for req in args:
+            req_to_add = install_req_from_line(
+                req,
+                comes_from=None,
+                isolated=options.isolated_mode,
+                use_pep517=options.use_pep517,
+                user_supplied=True,
+                config_settings=getattr(options, "config_settings", None),
+            )
+            requirements.append(req_to_add)
+
+        for req in options.editables:
+            req_to_add = install_req_from_editable(
+                req,
+                user_supplied=True,
+                isolated=options.isolated_mode,
+                use_pep517=options.use_pep517,
+                config_settings=getattr(options, "config_settings", None),
+            )
+            requirements.append(req_to_add)
+
+        # NOTE: options.require_hashes may be set if --require-hashes is True
+        for filename in options.requirements:
+            for parsed_req in parse_requirements(
+                filename, finder=finder, options=options, session=session
+            ):
+                req_to_add = install_req_from_parsed_requirement(
+                    parsed_req,
+                    isolated=options.isolated_mode,
+                    use_pep517=options.use_pep517,
+                    user_supplied=True,
+                    config_settings=(
+                        parsed_req.options.get("config_settings")
+                        if parsed_req.options
+                        else None
+                    ),
+                )
+                requirements.append(req_to_add)
+
+        # If any requirement has hash options, enable hash checking.
+        if any(req.has_hash_options for req in requirements):
+            options.require_hashes = True
+
+        if not (args or options.editables or options.requirements):
+            opts = {"name": self.name}
+            if options.find_links:
+                raise CommandError(
+                    "You must give at least one requirement to {name} "
+                    '(maybe you meant "pip {name} {links}"?)'.format(
+                        **dict(opts, links=" ".join(options.find_links))
+                    )
+                )
+            else:
+                raise CommandError(
+                    "You must give at least one requirement to {name} "
+                    '(see "pip help {name}")'.format(**opts)
+                )
+
+        return requirements
+
+    @staticmethod
+    def trace_basic_info(finder: PackageFinder) -> None:
+        """
+        Trace basic information about the provided objects.
+        """
+        # Display where finder is looking for packages
+        search_scope = finder.search_scope
+        locations = search_scope.get_formatted_locations()
+        if locations:
+            logger.info(locations)
+
+    def _build_package_finder(
+        self,
+        options: Values,
+        session: PipSession,
+        target_python: Optional[TargetPython] = None,
+        ignore_requires_python: Optional[bool] = None,
+    ) -> PackageFinder:
+        """
+        Create a package finder appropriate to this requirement command.
+
+        :param ignore_requires_python: Whether to ignore incompatible
+            "Requires-Python" values in links. Defaults to False.
+        """
+        link_collector = LinkCollector.create(session, options=options)
+        selection_prefs = SelectionPreferences(
+            allow_yanked=True,
+            format_control=options.format_control,
+            allow_all_prereleases=options.pre,
+            prefer_binary=options.prefer_binary,
+            ignore_requires_python=ignore_requires_python,
+        )
+
+        return PackageFinder.create(
+            link_collector=link_collector,
+            selection_prefs=selection_prefs,
+            target_python=target_python,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/spinners.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/spinners.py
new file mode 100644
index 00000000..cf2b976f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/spinners.py
@@ -0,0 +1,159 @@
+import contextlib
+import itertools
+import logging
+import sys
+import time
+from typing import IO, Generator, Optional
+
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.logging import get_indentation
+
+logger = logging.getLogger(__name__)
+
+
+class SpinnerInterface:
+    def spin(self) -> None:
+        raise NotImplementedError()
+
+    def finish(self, final_status: str) -> None:
+        raise NotImplementedError()
+
+
+class InteractiveSpinner(SpinnerInterface):
+    def __init__(
+        self,
+        message: str,
+        file: Optional[IO[str]] = None,
+        spin_chars: str = "-\\|/",
+        # Empirically, 8 updates/second looks nice
+        min_update_interval_seconds: float = 0.125,
+    ):
+        self._message = message
+        if file is None:
+            file = sys.stdout
+        self._file = file
+        self._rate_limiter = RateLimiter(min_update_interval_seconds)
+        self._finished = False
+
+        self._spin_cycle = itertools.cycle(spin_chars)
+
+        self._file.write(" " * get_indentation() + self._message + " ... ")
+        self._width = 0
+
+    def _write(self, status: str) -> None:
+        assert not self._finished
+        # Erase what we wrote before by backspacing to the beginning, writing
+        # spaces to overwrite the old text, and then backspacing again
+        backup = "\b" * self._width
+        self._file.write(backup + " " * self._width + backup)
+        # Now we have a blank slate to add our status
+        self._file.write(status)
+        self._width = len(status)
+        self._file.flush()
+        self._rate_limiter.reset()
+
+    def spin(self) -> None:
+        if self._finished:
+            return
+        if not self._rate_limiter.ready():
+            return
+        self._write(next(self._spin_cycle))
+
+    def finish(self, final_status: str) -> None:
+        if self._finished:
+            return
+        self._write(final_status)
+        self._file.write("\n")
+        self._file.flush()
+        self._finished = True
+
+
+# Used for dumb terminals, non-interactive installs (no tty), etc.
+# We still print updates occasionally (once every 60 seconds by default) to
+# act as a keep-alive for systems like Travis-CI that take lack-of-output as
+# an indication that a task has frozen.
+class NonInteractiveSpinner(SpinnerInterface):
+    def __init__(self, message: str, min_update_interval_seconds: float = 60.0) -> None:
+        self._message = message
+        self._finished = False
+        self._rate_limiter = RateLimiter(min_update_interval_seconds)
+        self._update("started")
+
+    def _update(self, status: str) -> None:
+        assert not self._finished
+        self._rate_limiter.reset()
+        logger.info("%s: %s", self._message, status)
+
+    def spin(self) -> None:
+        if self._finished:
+            return
+        if not self._rate_limiter.ready():
+            return
+        self._update("still running...")
+
+    def finish(self, final_status: str) -> None:
+        if self._finished:
+            return
+        self._update(f"finished with status '{final_status}'")
+        self._finished = True
+
+
+class RateLimiter:
+    def __init__(self, min_update_interval_seconds: float) -> None:
+        self._min_update_interval_seconds = min_update_interval_seconds
+        self._last_update: float = 0
+
+    def ready(self) -> bool:
+        now = time.time()
+        delta = now - self._last_update
+        return delta >= self._min_update_interval_seconds
+
+    def reset(self) -> None:
+        self._last_update = time.time()
+
+
+@contextlib.contextmanager
+def open_spinner(message: str) -> Generator[SpinnerInterface, None, None]:
+    # Interactive spinner goes directly to sys.stdout rather than being routed
+    # through the logging system, but it acts like it has level INFO,
+    # i.e. it's only displayed if we're at level INFO or better.
+    # Non-interactive spinner goes through the logging system, so it is always
+    # in sync with logging configuration.
+    if sys.stdout.isatty() and logger.getEffectiveLevel() <= logging.INFO:
+        spinner: SpinnerInterface = InteractiveSpinner(message)
+    else:
+        spinner = NonInteractiveSpinner(message)
+    try:
+        with hidden_cursor(sys.stdout):
+            yield spinner
+    except KeyboardInterrupt:
+        spinner.finish("canceled")
+        raise
+    except Exception:
+        spinner.finish("error")
+        raise
+    else:
+        spinner.finish("done")
+
+
+HIDE_CURSOR = "\x1b[?25l"
+SHOW_CURSOR = "\x1b[?25h"
+
+
+@contextlib.contextmanager
+def hidden_cursor(file: IO[str]) -> Generator[None, None, None]:
+    # The Windows terminal does not support the hide/show cursor ANSI codes,
+    # even via colorama. So don't even try.
+    if WINDOWS:
+        yield
+    # We don't want to clutter the output with control characters if we're
+    # writing to a file, or if the user is running with --quiet.
+    # See https://github.com/pypa/pip/issues/3418
+    elif not file.isatty() or logger.getEffectiveLevel() > logging.INFO:
+        yield
+    else:
+        file.write(HIDE_CURSOR)
+        try:
+            yield
+        finally:
+            file.write(SHOW_CURSOR)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/status_codes.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/status_codes.py
new file mode 100644
index 00000000..5e29502c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/cli/status_codes.py
@@ -0,0 +1,6 @@
+SUCCESS = 0
+ERROR = 1
+UNKNOWN_ERROR = 2
+VIRTUALENV_NOT_FOUND = 3
+PREVIOUS_BUILD_DIR_ERROR = 4
+NO_MATCHES_FOUND = 23
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__init__.py
new file mode 100644
index 00000000..858a4101
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__init__.py
@@ -0,0 +1,132 @@
+"""
+Package containing all pip commands
+"""
+
+import importlib
+from collections import namedtuple
+from typing import Any, Dict, Optional
+
+from pip._internal.cli.base_command import Command
+
+CommandInfo = namedtuple("CommandInfo", "module_path, class_name, summary")
+
+# This dictionary does a bunch of heavy lifting for help output:
+# - Enables avoiding additional (costly) imports for presenting `--help`.
+# - The ordering matters for help display.
+#
+# Even though the module path starts with the same "pip._internal.commands"
+# prefix, the full path makes testing easier (specifically when modifying
+# `commands_dict` in test setup / teardown).
+commands_dict: Dict[str, CommandInfo] = {
+    "install": CommandInfo(
+        "pip._internal.commands.install",
+        "InstallCommand",
+        "Install packages.",
+    ),
+    "download": CommandInfo(
+        "pip._internal.commands.download",
+        "DownloadCommand",
+        "Download packages.",
+    ),
+    "uninstall": CommandInfo(
+        "pip._internal.commands.uninstall",
+        "UninstallCommand",
+        "Uninstall packages.",
+    ),
+    "freeze": CommandInfo(
+        "pip._internal.commands.freeze",
+        "FreezeCommand",
+        "Output installed packages in requirements format.",
+    ),
+    "inspect": CommandInfo(
+        "pip._internal.commands.inspect",
+        "InspectCommand",
+        "Inspect the python environment.",
+    ),
+    "list": CommandInfo(
+        "pip._internal.commands.list",
+        "ListCommand",
+        "List installed packages.",
+    ),
+    "show": CommandInfo(
+        "pip._internal.commands.show",
+        "ShowCommand",
+        "Show information about installed packages.",
+    ),
+    "check": CommandInfo(
+        "pip._internal.commands.check",
+        "CheckCommand",
+        "Verify installed packages have compatible dependencies.",
+    ),
+    "config": CommandInfo(
+        "pip._internal.commands.configuration",
+        "ConfigurationCommand",
+        "Manage local and global configuration.",
+    ),
+    "search": CommandInfo(
+        "pip._internal.commands.search",
+        "SearchCommand",
+        "Search PyPI for packages.",
+    ),
+    "cache": CommandInfo(
+        "pip._internal.commands.cache",
+        "CacheCommand",
+        "Inspect and manage pip's wheel cache.",
+    ),
+    "index": CommandInfo(
+        "pip._internal.commands.index",
+        "IndexCommand",
+        "Inspect information available from package indexes.",
+    ),
+    "wheel": CommandInfo(
+        "pip._internal.commands.wheel",
+        "WheelCommand",
+        "Build wheels from your requirements.",
+    ),
+    "hash": CommandInfo(
+        "pip._internal.commands.hash",
+        "HashCommand",
+        "Compute hashes of package archives.",
+    ),
+    "completion": CommandInfo(
+        "pip._internal.commands.completion",
+        "CompletionCommand",
+        "A helper command used for command completion.",
+    ),
+    "debug": CommandInfo(
+        "pip._internal.commands.debug",
+        "DebugCommand",
+        "Show information useful for debugging.",
+    ),
+    "help": CommandInfo(
+        "pip._internal.commands.help",
+        "HelpCommand",
+        "Show help for commands.",
+    ),
+}
+
+
+def create_command(name: str, **kwargs: Any) -> Command:
+    """
+    Create an instance of the Command class with the given name.
+    """
+    module_path, class_name, summary = commands_dict[name]
+    module = importlib.import_module(module_path)
+    command_class = getattr(module, class_name)
+    command = command_class(name=name, summary=summary, **kwargs)
+
+    return command
+
+
+def get_similar_commands(name: str) -> Optional[str]:
+    """Command name auto-correct."""
+    from difflib import get_close_matches
+
+    name = name.lower()
+
+    close_commands = get_close_matches(name, commands_dict.keys())
+
+    if close_commands:
+        return close_commands[0]
+    else:
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..8dc0711d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/cache.cpython-312.pyc
new file mode 100644
index 00000000..291fb838
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/check.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/check.cpython-312.pyc
new file mode 100644
index 00000000..9a2345c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/check.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/completion.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/completion.cpython-312.pyc
new file mode 100644
index 00000000..276a7438
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/completion.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/configuration.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/configuration.cpython-312.pyc
new file mode 100644
index 00000000..a5fbe760
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/configuration.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/debug.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/debug.cpython-312.pyc
new file mode 100644
index 00000000..eaab0323
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/debug.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/download.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/download.cpython-312.pyc
new file mode 100644
index 00000000..7c7fcf6f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/download.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/freeze.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/freeze.cpython-312.pyc
new file mode 100644
index 00000000..fd36a62b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/freeze.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/hash.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/hash.cpython-312.pyc
new file mode 100644
index 00000000..8a11afbe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/hash.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/help.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/help.cpython-312.pyc
new file mode 100644
index 00000000..4879ca2d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/help.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/index.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/index.cpython-312.pyc
new file mode 100644
index 00000000..e0447c2c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/index.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/inspect.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/inspect.cpython-312.pyc
new file mode 100644
index 00000000..1ef1743f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/inspect.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/install.cpython-312.pyc
new file mode 100644
index 00000000..879a8eee
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/list.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/list.cpython-312.pyc
new file mode 100644
index 00000000..93451d9a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/list.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/search.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/search.cpython-312.pyc
new file mode 100644
index 00000000..67d6fa3a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/search.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/show.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/show.cpython-312.pyc
new file mode 100644
index 00000000..c9de8a3a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/show.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/uninstall.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/uninstall.cpython-312.pyc
new file mode 100644
index 00000000..4ecf46b5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/uninstall.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..dba2356d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/cache.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/cache.py
new file mode 100644
index 00000000..ad65641e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/cache.py
@@ -0,0 +1,228 @@
+import os
+import textwrap
+from optparse import Values
+from typing import Any, List
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.exceptions import CommandError, PipError
+from pip._internal.utils import filesystem
+from pip._internal.utils.logging import getLogger
+from pip._internal.utils.misc import format_size
+
+logger = getLogger(__name__)
+
+
+class CacheCommand(Command):
+    """
+    Inspect and manage pip's wheel cache.
+
+    Subcommands:
+
+    - dir: Show the cache directory.
+    - info: Show information about the cache.
+    - list: List filenames of packages stored in the cache.
+    - remove: Remove one or more package from the cache.
+    - purge: Remove all items from the cache.
+
+    ``<pattern>`` can be a glob expression or a package name.
+    """
+
+    ignore_require_venv = True
+    usage = """
+        %prog dir
+        %prog info
+        %prog list [<pattern>] [--format=[human, abspath]]
+        %prog remove <pattern>
+        %prog purge
+    """
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "--format",
+            action="store",
+            dest="list_format",
+            default="human",
+            choices=("human", "abspath"),
+            help="Select the output format among: human (default) or abspath",
+        )
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        handlers = {
+            "dir": self.get_cache_dir,
+            "info": self.get_cache_info,
+            "list": self.list_cache_items,
+            "remove": self.remove_cache_items,
+            "purge": self.purge_cache,
+        }
+
+        if not options.cache_dir:
+            logger.error("pip cache commands can not function since cache is disabled.")
+            return ERROR
+
+        # Determine action
+        if not args or args[0] not in handlers:
+            logger.error(
+                "Need an action (%s) to perform.",
+                ", ".join(sorted(handlers)),
+            )
+            return ERROR
+
+        action = args[0]
+
+        # Error handling happens here, not in the action-handlers.
+        try:
+            handlers[action](options, args[1:])
+        except PipError as e:
+            logger.error(e.args[0])
+            return ERROR
+
+        return SUCCESS
+
+    def get_cache_dir(self, options: Values, args: List[Any]) -> None:
+        if args:
+            raise CommandError("Too many arguments")
+
+        logger.info(options.cache_dir)
+
+    def get_cache_info(self, options: Values, args: List[Any]) -> None:
+        if args:
+            raise CommandError("Too many arguments")
+
+        num_http_files = len(self._find_http_files(options))
+        num_packages = len(self._find_wheels(options, "*"))
+
+        http_cache_location = self._cache_dir(options, "http-v2")
+        old_http_cache_location = self._cache_dir(options, "http")
+        wheels_cache_location = self._cache_dir(options, "wheels")
+        http_cache_size = filesystem.format_size(
+            filesystem.directory_size(http_cache_location)
+            + filesystem.directory_size(old_http_cache_location)
+        )
+        wheels_cache_size = filesystem.format_directory_size(wheels_cache_location)
+
+        message = (
+            textwrap.dedent(
+                """
+                    Package index page cache location (pip v23.3+): {http_cache_location}
+                    Package index page cache location (older pips): {old_http_cache_location}
+                    Package index page cache size: {http_cache_size}
+                    Number of HTTP files: {num_http_files}
+                    Locally built wheels location: {wheels_cache_location}
+                    Locally built wheels size: {wheels_cache_size}
+                    Number of locally built wheels: {package_count}
+                """  # noqa: E501
+            )
+            .format(
+                http_cache_location=http_cache_location,
+                old_http_cache_location=old_http_cache_location,
+                http_cache_size=http_cache_size,
+                num_http_files=num_http_files,
+                wheels_cache_location=wheels_cache_location,
+                package_count=num_packages,
+                wheels_cache_size=wheels_cache_size,
+            )
+            .strip()
+        )
+
+        logger.info(message)
+
+    def list_cache_items(self, options: Values, args: List[Any]) -> None:
+        if len(args) > 1:
+            raise CommandError("Too many arguments")
+
+        if args:
+            pattern = args[0]
+        else:
+            pattern = "*"
+
+        files = self._find_wheels(options, pattern)
+        if options.list_format == "human":
+            self.format_for_human(files)
+        else:
+            self.format_for_abspath(files)
+
+    def format_for_human(self, files: List[str]) -> None:
+        if not files:
+            logger.info("No locally built wheels cached.")
+            return
+
+        results = []
+        for filename in files:
+            wheel = os.path.basename(filename)
+            size = filesystem.format_file_size(filename)
+            results.append(f" - {wheel} ({size})")
+        logger.info("Cache contents:\n")
+        logger.info("\n".join(sorted(results)))
+
+    def format_for_abspath(self, files: List[str]) -> None:
+        if files:
+            logger.info("\n".join(sorted(files)))
+
+    def remove_cache_items(self, options: Values, args: List[Any]) -> None:
+        if len(args) > 1:
+            raise CommandError("Too many arguments")
+
+        if not args:
+            raise CommandError("Please provide a pattern")
+
+        files = self._find_wheels(options, args[0])
+
+        no_matching_msg = "No matching packages"
+        if args[0] == "*":
+            # Only fetch http files if no specific pattern given
+            files += self._find_http_files(options)
+        else:
+            # Add the pattern to the log message
+            no_matching_msg += f' for pattern "{args[0]}"'
+
+        if not files:
+            logger.warning(no_matching_msg)
+
+        bytes_removed = 0
+        for filename in files:
+            bytes_removed += os.stat(filename).st_size
+            os.unlink(filename)
+            logger.verbose("Removed %s", filename)
+        logger.info("Files removed: %s (%s)", len(files), format_size(bytes_removed))
+
+    def purge_cache(self, options: Values, args: List[Any]) -> None:
+        if args:
+            raise CommandError("Too many arguments")
+
+        return self.remove_cache_items(options, ["*"])
+
+    def _cache_dir(self, options: Values, subdir: str) -> str:
+        return os.path.join(options.cache_dir, subdir)
+
+    def _find_http_files(self, options: Values) -> List[str]:
+        old_http_dir = self._cache_dir(options, "http")
+        new_http_dir = self._cache_dir(options, "http-v2")
+        return filesystem.find_files(old_http_dir, "*") + filesystem.find_files(
+            new_http_dir, "*"
+        )
+
+    def _find_wheels(self, options: Values, pattern: str) -> List[str]:
+        wheel_dir = self._cache_dir(options, "wheels")
+
+        # The wheel filename format, as specified in PEP 427, is:
+        #     {distribution}-{version}(-{build})?-{python}-{abi}-{platform}.whl
+        #
+        # Additionally, non-alphanumeric values in the distribution are
+        # normalized to underscores (_), meaning hyphens can never occur
+        # before `-{version}`.
+        #
+        # Given that information:
+        # - If the pattern we're given contains a hyphen (-), the user is
+        #   providing at least the version. Thus, we can just append `*.whl`
+        #   to match the rest of it.
+        # - If the pattern we're given doesn't contain a hyphen (-), the
+        #   user is only providing the name. Thus, we append `-*.whl` to
+        #   match the hyphen before the version, followed by anything else.
+        #
+        # PEP 427: https://www.python.org/dev/peps/pep-0427/
+        pattern = pattern + ("*.whl" if "-" in pattern else "-*.whl")
+
+        return filesystem.find_files(wheel_dir, pattern)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/check.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/check.py
new file mode 100644
index 00000000..f54a16dc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/check.py
@@ -0,0 +1,67 @@
+import logging
+from optparse import Values
+from typing import List
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.metadata import get_default_environment
+from pip._internal.operations.check import (
+    check_package_set,
+    check_unsupported,
+    create_package_set_from_installed,
+)
+from pip._internal.utils.compatibility_tags import get_supported
+from pip._internal.utils.misc import write_output
+
+logger = logging.getLogger(__name__)
+
+
+class CheckCommand(Command):
+    """Verify installed packages have compatible dependencies."""
+
+    ignore_require_venv = True
+    usage = """
+      %prog [options]"""
+
+    def run(self, options: Values, args: List[str]) -> int:
+        package_set, parsing_probs = create_package_set_from_installed()
+        missing, conflicting = check_package_set(package_set)
+        unsupported = list(
+            check_unsupported(
+                get_default_environment().iter_installed_distributions(),
+                get_supported(),
+            )
+        )
+
+        for project_name in missing:
+            version = package_set[project_name].version
+            for dependency in missing[project_name]:
+                write_output(
+                    "%s %s requires %s, which is not installed.",
+                    project_name,
+                    version,
+                    dependency[0],
+                )
+
+        for project_name in conflicting:
+            version = package_set[project_name].version
+            for dep_name, dep_version, req in conflicting[project_name]:
+                write_output(
+                    "%s %s has requirement %s, but you have %s %s.",
+                    project_name,
+                    version,
+                    req,
+                    dep_name,
+                    dep_version,
+                )
+        for package in unsupported:
+            write_output(
+                "%s %s is not supported on this platform",
+                package.raw_name,
+                package.version,
+            )
+        if missing or conflicting or parsing_probs or unsupported:
+            return ERROR
+        else:
+            write_output("No broken requirements found.")
+            return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/completion.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/completion.py
new file mode 100644
index 00000000..9e89e279
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/completion.py
@@ -0,0 +1,130 @@
+import sys
+import textwrap
+from optparse import Values
+from typing import List
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.utils.misc import get_prog
+
+BASE_COMPLETION = """
+# pip {shell} completion start{script}# pip {shell} completion end
+"""
+
+COMPLETION_SCRIPTS = {
+    "bash": """
+        _pip_completion()
+        {{
+            COMPREPLY=( $( COMP_WORDS="${{COMP_WORDS[*]}}" \\
+                           COMP_CWORD=$COMP_CWORD \\
+                           PIP_AUTO_COMPLETE=1 $1 2>/dev/null ) )
+        }}
+        complete -o default -F _pip_completion {prog}
+    """,
+    "zsh": """
+        #compdef -P pip[0-9.]#
+        __pip() {{
+          compadd $( COMP_WORDS="$words[*]" \\
+                     COMP_CWORD=$((CURRENT-1)) \\
+                     PIP_AUTO_COMPLETE=1 $words[1] 2>/dev/null )
+        }}
+        if [[ $zsh_eval_context[-1] == loadautofunc ]]; then
+          # autoload from fpath, call function directly
+          __pip "$@"
+        else
+          # eval/source/. command, register function for later
+          compdef __pip -P 'pip[0-9.]#'
+        fi
+    """,
+    "fish": """
+        function __fish_complete_pip
+            set -lx COMP_WORDS (commandline -o) ""
+            set -lx COMP_CWORD ( \\
+                math (contains -i -- (commandline -t) $COMP_WORDS)-1 \\
+            )
+            set -lx PIP_AUTO_COMPLETE 1
+            string split \\  -- (eval $COMP_WORDS[1])
+        end
+        complete -fa "(__fish_complete_pip)" -c {prog}
+    """,
+    "powershell": """
+        if ((Test-Path Function:\\TabExpansion) -and -not `
+            (Test-Path Function:\\_pip_completeBackup)) {{
+            Rename-Item Function:\\TabExpansion _pip_completeBackup
+        }}
+        function TabExpansion($line, $lastWord) {{
+            $lastBlock = [regex]::Split($line, '[|;]')[-1].TrimStart()
+            if ($lastBlock.StartsWith("{prog} ")) {{
+                $Env:COMP_WORDS=$lastBlock
+                $Env:COMP_CWORD=$lastBlock.Split().Length - 1
+                $Env:PIP_AUTO_COMPLETE=1
+                (& {prog}).Split()
+                Remove-Item Env:COMP_WORDS
+                Remove-Item Env:COMP_CWORD
+                Remove-Item Env:PIP_AUTO_COMPLETE
+            }}
+            elseif (Test-Path Function:\\_pip_completeBackup) {{
+                # Fall back on existing tab expansion
+                _pip_completeBackup $line $lastWord
+            }}
+        }}
+    """,
+}
+
+
+class CompletionCommand(Command):
+    """A helper command to be used for command completion."""
+
+    ignore_require_venv = True
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "--bash",
+            "-b",
+            action="store_const",
+            const="bash",
+            dest="shell",
+            help="Emit completion code for bash",
+        )
+        self.cmd_opts.add_option(
+            "--zsh",
+            "-z",
+            action="store_const",
+            const="zsh",
+            dest="shell",
+            help="Emit completion code for zsh",
+        )
+        self.cmd_opts.add_option(
+            "--fish",
+            "-f",
+            action="store_const",
+            const="fish",
+            dest="shell",
+            help="Emit completion code for fish",
+        )
+        self.cmd_opts.add_option(
+            "--powershell",
+            "-p",
+            action="store_const",
+            const="powershell",
+            dest="shell",
+            help="Emit completion code for powershell",
+        )
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        """Prints the completion code of the given shell"""
+        shells = COMPLETION_SCRIPTS.keys()
+        shell_options = ["--" + shell for shell in sorted(shells)]
+        if options.shell in shells:
+            script = textwrap.dedent(
+                COMPLETION_SCRIPTS.get(options.shell, "").format(prog=get_prog())
+            )
+            print(BASE_COMPLETION.format(script=script, shell=options.shell))
+            return SUCCESS
+        else:
+            sys.stderr.write(
+                "ERROR: You must pass {}\n".format(" or ".join(shell_options))
+            )
+            return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/configuration.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/configuration.py
new file mode 100644
index 00000000..1a1dc6b6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/configuration.py
@@ -0,0 +1,280 @@
+import logging
+import os
+import subprocess
+from optparse import Values
+from typing import Any, List, Optional
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.configuration import (
+    Configuration,
+    Kind,
+    get_configuration_files,
+    kinds,
+)
+from pip._internal.exceptions import PipError
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import get_prog, write_output
+
+logger = logging.getLogger(__name__)
+
+
+class ConfigurationCommand(Command):
+    """
+    Manage local and global configuration.
+
+    Subcommands:
+
+    - list: List the active configuration (or from the file specified)
+    - edit: Edit the configuration file in an editor
+    - get: Get the value associated with command.option
+    - set: Set the command.option=value
+    - unset: Unset the value associated with command.option
+    - debug: List the configuration files and values defined under them
+
+    Configuration keys should be dot separated command and option name,
+    with the special prefix "global" affecting any command. For example,
+    "pip config set global.index-url https://example.org/" would configure
+    the index url for all commands, but "pip config set download.timeout 10"
+    would configure a 10 second timeout only for "pip download" commands.
+
+    If none of --user, --global and --site are passed, a virtual
+    environment configuration file is used if one is active and the file
+    exists. Otherwise, all modifications happen to the user file by
+    default.
+    """
+
+    ignore_require_venv = True
+    usage = """
+        %prog [<file-option>] list
+        %prog [<file-option>] [--editor <editor-path>] edit
+
+        %prog [<file-option>] get command.option
+        %prog [<file-option>] set command.option value
+        %prog [<file-option>] unset command.option
+        %prog [<file-option>] debug
+    """
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "--editor",
+            dest="editor",
+            action="store",
+            default=None,
+            help=(
+                "Editor to use to edit the file. Uses VISUAL or EDITOR "
+                "environment variables if not provided."
+            ),
+        )
+
+        self.cmd_opts.add_option(
+            "--global",
+            dest="global_file",
+            action="store_true",
+            default=False,
+            help="Use the system-wide configuration file only",
+        )
+
+        self.cmd_opts.add_option(
+            "--user",
+            dest="user_file",
+            action="store_true",
+            default=False,
+            help="Use the user configuration file only",
+        )
+
+        self.cmd_opts.add_option(
+            "--site",
+            dest="site_file",
+            action="store_true",
+            default=False,
+            help="Use the current environment configuration file only",
+        )
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        handlers = {
+            "list": self.list_values,
+            "edit": self.open_in_editor,
+            "get": self.get_name,
+            "set": self.set_name_value,
+            "unset": self.unset_name,
+            "debug": self.list_config_values,
+        }
+
+        # Determine action
+        if not args or args[0] not in handlers:
+            logger.error(
+                "Need an action (%s) to perform.",
+                ", ".join(sorted(handlers)),
+            )
+            return ERROR
+
+        action = args[0]
+
+        # Determine which configuration files are to be loaded
+        #    Depends on whether the command is modifying.
+        try:
+            load_only = self._determine_file(
+                options, need_value=(action in ["get", "set", "unset", "edit"])
+            )
+        except PipError as e:
+            logger.error(e.args[0])
+            return ERROR
+
+        # Load a new configuration
+        self.configuration = Configuration(
+            isolated=options.isolated_mode, load_only=load_only
+        )
+        self.configuration.load()
+
+        # Error handling happens here, not in the action-handlers.
+        try:
+            handlers[action](options, args[1:])
+        except PipError as e:
+            logger.error(e.args[0])
+            return ERROR
+
+        return SUCCESS
+
+    def _determine_file(self, options: Values, need_value: bool) -> Optional[Kind]:
+        file_options = [
+            key
+            for key, value in (
+                (kinds.USER, options.user_file),
+                (kinds.GLOBAL, options.global_file),
+                (kinds.SITE, options.site_file),
+            )
+            if value
+        ]
+
+        if not file_options:
+            if not need_value:
+                return None
+            # Default to user, unless there's a site file.
+            elif any(
+                os.path.exists(site_config_file)
+                for site_config_file in get_configuration_files()[kinds.SITE]
+            ):
+                return kinds.SITE
+            else:
+                return kinds.USER
+        elif len(file_options) == 1:
+            return file_options[0]
+
+        raise PipError(
+            "Need exactly one file to operate upon "
+            "(--user, --site, --global) to perform."
+        )
+
+    def list_values(self, options: Values, args: List[str]) -> None:
+        self._get_n_args(args, "list", n=0)
+
+        for key, value in sorted(self.configuration.items()):
+            write_output("%s=%r", key, value)
+
+    def get_name(self, options: Values, args: List[str]) -> None:
+        key = self._get_n_args(args, "get [name]", n=1)
+        value = self.configuration.get_value(key)
+
+        write_output("%s", value)
+
+    def set_name_value(self, options: Values, args: List[str]) -> None:
+        key, value = self._get_n_args(args, "set [name] [value]", n=2)
+        self.configuration.set_value(key, value)
+
+        self._save_configuration()
+
+    def unset_name(self, options: Values, args: List[str]) -> None:
+        key = self._get_n_args(args, "unset [name]", n=1)
+        self.configuration.unset_value(key)
+
+        self._save_configuration()
+
+    def list_config_values(self, options: Values, args: List[str]) -> None:
+        """List config key-value pairs across different config files"""
+        self._get_n_args(args, "debug", n=0)
+
+        self.print_env_var_values()
+        # Iterate over config files and print if they exist, and the
+        # key-value pairs present in them if they do
+        for variant, files in sorted(self.configuration.iter_config_files()):
+            write_output("%s:", variant)
+            for fname in files:
+                with indent_log():
+                    file_exists = os.path.exists(fname)
+                    write_output("%s, exists: %r", fname, file_exists)
+                    if file_exists:
+                        self.print_config_file_values(variant)
+
+    def print_config_file_values(self, variant: Kind) -> None:
+        """Get key-value pairs from the file of a variant"""
+        for name, value in self.configuration.get_values_in_config(variant).items():
+            with indent_log():
+                write_output("%s: %s", name, value)
+
+    def print_env_var_values(self) -> None:
+        """Get key-values pairs present as environment variables"""
+        write_output("%s:", "env_var")
+        with indent_log():
+            for key, value in sorted(self.configuration.get_environ_vars()):
+                env_var = f"PIP_{key.upper()}"
+                write_output("%s=%r", env_var, value)
+
+    def open_in_editor(self, options: Values, args: List[str]) -> None:
+        editor = self._determine_editor(options)
+
+        fname = self.configuration.get_file_to_edit()
+        if fname is None:
+            raise PipError("Could not determine appropriate file.")
+        elif '"' in fname:
+            # This shouldn't happen, unless we see a username like that.
+            # If that happens, we'd appreciate a pull request fixing this.
+            raise PipError(
+                f'Can not open an editor for a file name containing "\n{fname}'
+            )
+
+        try:
+            subprocess.check_call(f'{editor} "{fname}"', shell=True)
+        except FileNotFoundError as e:
+            if not e.filename:
+                e.filename = editor
+            raise
+        except subprocess.CalledProcessError as e:
+            raise PipError(f"Editor Subprocess exited with exit code {e.returncode}")
+
+    def _get_n_args(self, args: List[str], example: str, n: int) -> Any:
+        """Helper to make sure the command got the right number of arguments"""
+        if len(args) != n:
+            msg = (
+                f"Got unexpected number of arguments, expected {n}. "
+                f'(example: "{get_prog()} config {example}")'
+            )
+            raise PipError(msg)
+
+        if n == 1:
+            return args[0]
+        else:
+            return args
+
+    def _save_configuration(self) -> None:
+        # We successfully ran a modifying command. Need to save the
+        # configuration.
+        try:
+            self.configuration.save()
+        except Exception:
+            logger.exception(
+                "Unable to save configuration. Please report this as a bug."
+            )
+            raise PipError("Internal Error.")
+
+    def _determine_editor(self, options: Values) -> str:
+        if options.editor is not None:
+            return options.editor
+        elif "VISUAL" in os.environ:
+            return os.environ["VISUAL"]
+        elif "EDITOR" in os.environ:
+            return os.environ["EDITOR"]
+        else:
+            raise PipError("Could not determine editor to use.")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/debug.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/debug.py
new file mode 100644
index 00000000..567ca967
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/debug.py
@@ -0,0 +1,201 @@
+import locale
+import logging
+import os
+import sys
+from optparse import Values
+from types import ModuleType
+from typing import Any, Dict, List, Optional
+
+import pip._vendor
+from pip._vendor.certifi import where
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.cmdoptions import make_target_python
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.configuration import Configuration
+from pip._internal.metadata import get_environment
+from pip._internal.utils.compat import open_text_resource
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import get_pip_version
+
+logger = logging.getLogger(__name__)
+
+
+def show_value(name: str, value: Any) -> None:
+    logger.info("%s: %s", name, value)
+
+
+def show_sys_implementation() -> None:
+    logger.info("sys.implementation:")
+    implementation_name = sys.implementation.name
+    with indent_log():
+        show_value("name", implementation_name)
+
+
+def create_vendor_txt_map() -> Dict[str, str]:
+    with open_text_resource("pip._vendor", "vendor.txt") as f:
+        # Purge non version specifying lines.
+        # Also, remove any space prefix or suffixes (including comments).
+        lines = [
+            line.strip().split(" ", 1)[0] for line in f.readlines() if "==" in line
+        ]
+
+    # Transform into "module" -> version dict.
+    return dict(line.split("==", 1) for line in lines)
+
+
+def get_module_from_module_name(module_name: str) -> Optional[ModuleType]:
+    # Module name can be uppercase in vendor.txt for some reason...
+    module_name = module_name.lower().replace("-", "_")
+    # PATCH: setuptools is actually only pkg_resources.
+    if module_name == "setuptools":
+        module_name = "pkg_resources"
+
+    try:
+        __import__(f"pip._vendor.{module_name}", globals(), locals(), level=0)
+        return getattr(pip._vendor, module_name)
+    except ImportError:
+        # We allow 'truststore' to fail to import due
+        # to being unavailable on Python 3.9 and earlier.
+        if module_name == "truststore" and sys.version_info < (3, 10):
+            return None
+        raise
+
+
+def get_vendor_version_from_module(module_name: str) -> Optional[str]:
+    module = get_module_from_module_name(module_name)
+    version = getattr(module, "__version__", None)
+
+    if module and not version:
+        # Try to find version in debundled module info.
+        assert module.__file__ is not None
+        env = get_environment([os.path.dirname(module.__file__)])
+        dist = env.get_distribution(module_name)
+        if dist:
+            version = str(dist.version)
+
+    return version
+
+
+def show_actual_vendor_versions(vendor_txt_versions: Dict[str, str]) -> None:
+    """Log the actual version and print extra info if there is
+    a conflict or if the actual version could not be imported.
+    """
+    for module_name, expected_version in vendor_txt_versions.items():
+        extra_message = ""
+        actual_version = get_vendor_version_from_module(module_name)
+        if not actual_version:
+            extra_message = (
+                " (Unable to locate actual module version, using"
+                " vendor.txt specified version)"
+            )
+            actual_version = expected_version
+        elif parse_version(actual_version) != parse_version(expected_version):
+            extra_message = (
+                " (CONFLICT: vendor.txt suggests version should"
+                f" be {expected_version})"
+            )
+        logger.info("%s==%s%s", module_name, actual_version, extra_message)
+
+
+def show_vendor_versions() -> None:
+    logger.info("vendored library versions:")
+
+    vendor_txt_versions = create_vendor_txt_map()
+    with indent_log():
+        show_actual_vendor_versions(vendor_txt_versions)
+
+
+def show_tags(options: Values) -> None:
+    tag_limit = 10
+
+    target_python = make_target_python(options)
+    tags = target_python.get_sorted_tags()
+
+    # Display the target options that were explicitly provided.
+    formatted_target = target_python.format_given()
+    suffix = ""
+    if formatted_target:
+        suffix = f" (target: {formatted_target})"
+
+    msg = f"Compatible tags: {len(tags)}{suffix}"
+    logger.info(msg)
+
+    if options.verbose < 1 and len(tags) > tag_limit:
+        tags_limited = True
+        tags = tags[:tag_limit]
+    else:
+        tags_limited = False
+
+    with indent_log():
+        for tag in tags:
+            logger.info(str(tag))
+
+        if tags_limited:
+            msg = f"...\n[First {tag_limit} tags shown. Pass --verbose to show all.]"
+            logger.info(msg)
+
+
+def ca_bundle_info(config: Configuration) -> str:
+    levels = {key.split(".", 1)[0] for key, _ in config.items()}
+    if not levels:
+        return "Not specified"
+
+    levels_that_override_global = ["install", "wheel", "download"]
+    global_overriding_level = [
+        level for level in levels if level in levels_that_override_global
+    ]
+    if not global_overriding_level:
+        return "global"
+
+    if "global" in levels:
+        levels.remove("global")
+    return ", ".join(levels)
+
+
+class DebugCommand(Command):
+    """
+    Display debug information.
+    """
+
+    usage = """
+      %prog <options>"""
+    ignore_require_venv = True
+
+    def add_options(self) -> None:
+        cmdoptions.add_target_python_options(self.cmd_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+        self.parser.config.load()
+
+    def run(self, options: Values, args: List[str]) -> int:
+        logger.warning(
+            "This command is only meant for debugging. "
+            "Do not use this with automation for parsing and getting these "
+            "details, since the output and options of this command may "
+            "change without notice."
+        )
+        show_value("pip version", get_pip_version())
+        show_value("sys.version", sys.version)
+        show_value("sys.executable", sys.executable)
+        show_value("sys.getdefaultencoding", sys.getdefaultencoding())
+        show_value("sys.getfilesystemencoding", sys.getfilesystemencoding())
+        show_value(
+            "locale.getpreferredencoding",
+            locale.getpreferredencoding(),
+        )
+        show_value("sys.platform", sys.platform)
+        show_sys_implementation()
+
+        show_value("'cert' config value", ca_bundle_info(self.parser.config))
+        show_value("REQUESTS_CA_BUNDLE", os.environ.get("REQUESTS_CA_BUNDLE"))
+        show_value("CURL_CA_BUNDLE", os.environ.get("CURL_CA_BUNDLE"))
+        show_value("pip._vendor.certifi.where()", where())
+        show_value("pip._vendor.DEBUNDLED", pip._vendor.DEBUNDLED)
+
+        show_vendor_versions()
+
+        show_tags(options)
+
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/download.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/download.py
new file mode 100644
index 00000000..917bbb91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/download.py
@@ -0,0 +1,146 @@
+import logging
+import os
+from optparse import Values
+from typing import List
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.cmdoptions import make_target_python
+from pip._internal.cli.req_command import RequirementCommand, with_cleanup
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.operations.build.build_tracker import get_build_tracker
+from pip._internal.req.req_install import check_legacy_setup_py_options
+from pip._internal.utils.misc import ensure_dir, normalize_path, write_output
+from pip._internal.utils.temp_dir import TempDirectory
+
+logger = logging.getLogger(__name__)
+
+
+class DownloadCommand(RequirementCommand):
+    """
+    Download packages from:
+
+    - PyPI (and other indexes) using requirement specifiers.
+    - VCS project urls.
+    - Local project directories.
+    - Local or remote source archives.
+
+    pip also supports downloading from "requirements files", which provide
+    an easy way to specify a whole environment to be downloaded.
+    """
+
+    usage = """
+      %prog [options] <requirement specifier> [package-index-options] ...
+      %prog [options] -r <requirements file> [package-index-options] ...
+      %prog [options] <vcs project url> ...
+      %prog [options] <local project path> ...
+      %prog [options] <archive url/path> ..."""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(cmdoptions.constraints())
+        self.cmd_opts.add_option(cmdoptions.requirements())
+        self.cmd_opts.add_option(cmdoptions.no_deps())
+        self.cmd_opts.add_option(cmdoptions.global_options())
+        self.cmd_opts.add_option(cmdoptions.no_binary())
+        self.cmd_opts.add_option(cmdoptions.only_binary())
+        self.cmd_opts.add_option(cmdoptions.prefer_binary())
+        self.cmd_opts.add_option(cmdoptions.src())
+        self.cmd_opts.add_option(cmdoptions.pre())
+        self.cmd_opts.add_option(cmdoptions.require_hashes())
+        self.cmd_opts.add_option(cmdoptions.progress_bar())
+        self.cmd_opts.add_option(cmdoptions.no_build_isolation())
+        self.cmd_opts.add_option(cmdoptions.use_pep517())
+        self.cmd_opts.add_option(cmdoptions.no_use_pep517())
+        self.cmd_opts.add_option(cmdoptions.check_build_deps())
+        self.cmd_opts.add_option(cmdoptions.ignore_requires_python())
+
+        self.cmd_opts.add_option(
+            "-d",
+            "--dest",
+            "--destination-dir",
+            "--destination-directory",
+            dest="download_dir",
+            metavar="dir",
+            default=os.curdir,
+            help="Download packages into <dir>.",
+        )
+
+        cmdoptions.add_target_python_options(self.cmd_opts)
+
+        index_opts = cmdoptions.make_option_group(
+            cmdoptions.index_group,
+            self.parser,
+        )
+
+        self.parser.insert_option_group(0, index_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    @with_cleanup
+    def run(self, options: Values, args: List[str]) -> int:
+        options.ignore_installed = True
+        # editable doesn't really make sense for `pip download`, but the bowels
+        # of the RequirementSet code require that property.
+        options.editables = []
+
+        cmdoptions.check_dist_restriction(options)
+
+        options.download_dir = normalize_path(options.download_dir)
+        ensure_dir(options.download_dir)
+
+        session = self.get_default_session(options)
+
+        target_python = make_target_python(options)
+        finder = self._build_package_finder(
+            options=options,
+            session=session,
+            target_python=target_python,
+            ignore_requires_python=options.ignore_requires_python,
+        )
+
+        build_tracker = self.enter_context(get_build_tracker())
+
+        directory = TempDirectory(
+            delete=not options.no_clean,
+            kind="download",
+            globally_managed=True,
+        )
+
+        reqs = self.get_requirements(args, options, finder, session)
+        check_legacy_setup_py_options(options, reqs)
+
+        preparer = self.make_requirement_preparer(
+            temp_build_dir=directory,
+            options=options,
+            build_tracker=build_tracker,
+            session=session,
+            finder=finder,
+            download_dir=options.download_dir,
+            use_user_site=False,
+            verbosity=self.verbosity,
+        )
+
+        resolver = self.make_resolver(
+            preparer=preparer,
+            finder=finder,
+            options=options,
+            ignore_requires_python=options.ignore_requires_python,
+            use_pep517=options.use_pep517,
+            py_version_info=options.python_version,
+        )
+
+        self.trace_basic_info(finder)
+
+        requirement_set = resolver.resolve(reqs, check_supported_wheels=True)
+
+        downloaded: List[str] = []
+        for req in requirement_set.requirements.values():
+            if req.satisfied_by is None:
+                assert req.name is not None
+                preparer.save_linked_requirement(req)
+                downloaded.append(req.name)
+
+        preparer.prepare_linked_requirements_more(requirement_set.requirements.values())
+
+        if downloaded:
+            write_output("Successfully downloaded %s", " ".join(downloaded))
+
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/freeze.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/freeze.py
new file mode 100644
index 00000000..885fdfeb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/freeze.py
@@ -0,0 +1,109 @@
+import sys
+from optparse import Values
+from typing import AbstractSet, List
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.operations.freeze import freeze
+from pip._internal.utils.compat import stdlib_pkgs
+
+
+def _should_suppress_build_backends() -> bool:
+    return sys.version_info < (3, 12)
+
+
+def _dev_pkgs() -> AbstractSet[str]:
+    pkgs = {"pip"}
+
+    if _should_suppress_build_backends():
+        pkgs |= {"setuptools", "distribute", "wheel"}
+
+    return pkgs
+
+
+class FreezeCommand(Command):
+    """
+    Output installed packages in requirements format.
+
+    packages are listed in a case-insensitive sorted order.
+    """
+
+    ignore_require_venv = True
+    usage = """
+      %prog [options]"""
+    log_streams = ("ext://sys.stderr", "ext://sys.stderr")
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-r",
+            "--requirement",
+            dest="requirements",
+            action="append",
+            default=[],
+            metavar="file",
+            help=(
+                "Use the order in the given requirements file and its "
+                "comments when generating output. This option can be "
+                "used multiple times."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "-l",
+            "--local",
+            dest="local",
+            action="store_true",
+            default=False,
+            help=(
+                "If in a virtualenv that has global access, do not output "
+                "globally-installed packages."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "--user",
+            dest="user",
+            action="store_true",
+            default=False,
+            help="Only output packages installed in user-site.",
+        )
+        self.cmd_opts.add_option(cmdoptions.list_path())
+        self.cmd_opts.add_option(
+            "--all",
+            dest="freeze_all",
+            action="store_true",
+            help=(
+                "Do not skip these packages in the output:"
+                " {}".format(", ".join(_dev_pkgs()))
+            ),
+        )
+        self.cmd_opts.add_option(
+            "--exclude-editable",
+            dest="exclude_editable",
+            action="store_true",
+            help="Exclude editable package from output.",
+        )
+        self.cmd_opts.add_option(cmdoptions.list_exclude())
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        skip = set(stdlib_pkgs)
+        if not options.freeze_all:
+            skip.update(_dev_pkgs())
+
+        if options.excludes:
+            skip.update(options.excludes)
+
+        cmdoptions.check_list_path_option(options)
+
+        for line in freeze(
+            requirement=options.requirements,
+            local_only=options.local,
+            user_only=options.user,
+            paths=options.path,
+            isolated=options.isolated_mode,
+            skip=skip,
+            exclude_editable=options.exclude_editable,
+        ):
+            sys.stdout.write(line + "\n")
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/hash.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/hash.py
new file mode 100644
index 00000000..042dac81
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/hash.py
@@ -0,0 +1,59 @@
+import hashlib
+import logging
+import sys
+from optparse import Values
+from typing import List
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.utils.hashes import FAVORITE_HASH, STRONG_HASHES
+from pip._internal.utils.misc import read_chunks, write_output
+
+logger = logging.getLogger(__name__)
+
+
+class HashCommand(Command):
+    """
+    Compute a hash of a local package archive.
+
+    These can be used with --hash in a requirements file to do repeatable
+    installs.
+    """
+
+    usage = "%prog [options] <file> ..."
+    ignore_require_venv = True
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-a",
+            "--algorithm",
+            dest="algorithm",
+            choices=STRONG_HASHES,
+            action="store",
+            default=FAVORITE_HASH,
+            help="The hash algorithm to use: one of {}".format(
+                ", ".join(STRONG_HASHES)
+            ),
+        )
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        if not args:
+            self.parser.print_usage(sys.stderr)
+            return ERROR
+
+        algorithm = options.algorithm
+        for path in args:
+            write_output(
+                "%s:\n--hash=%s:%s", path, algorithm, _hash_of_file(path, algorithm)
+            )
+        return SUCCESS
+
+
+def _hash_of_file(path: str, algorithm: str) -> str:
+    """Return the hash digest of a file."""
+    with open(path, "rb") as archive:
+        hash = hashlib.new(algorithm)
+        for chunk in read_chunks(archive):
+            hash.update(chunk)
+    return hash.hexdigest()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/help.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/help.py
new file mode 100644
index 00000000..62066318
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/help.py
@@ -0,0 +1,41 @@
+from optparse import Values
+from typing import List
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.exceptions import CommandError
+
+
+class HelpCommand(Command):
+    """Show help for commands"""
+
+    usage = """
+      %prog <command>"""
+    ignore_require_venv = True
+
+    def run(self, options: Values, args: List[str]) -> int:
+        from pip._internal.commands import (
+            commands_dict,
+            create_command,
+            get_similar_commands,
+        )
+
+        try:
+            # 'pip help' with no args is handled by pip.__init__.parseopt()
+            cmd_name = args[0]  # the command we need help for
+        except IndexError:
+            return SUCCESS
+
+        if cmd_name not in commands_dict:
+            guess = get_similar_commands(cmd_name)
+
+            msg = [f'unknown command "{cmd_name}"']
+            if guess:
+                msg.append(f'maybe you meant "{guess}"')
+
+            raise CommandError(" - ".join(msg))
+
+        command = create_command(cmd_name)
+        command.parser.print_help()
+
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/index.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/index.py
new file mode 100644
index 00000000..2e2661bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/index.py
@@ -0,0 +1,139 @@
+import logging
+from optparse import Values
+from typing import Any, Iterable, List, Optional
+
+from pip._vendor.packaging.version import Version
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.req_command import IndexGroupCommand
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.commands.search import print_dist_installation_info
+from pip._internal.exceptions import CommandError, DistributionNotFound, PipError
+from pip._internal.index.collector import LinkCollector
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.models.selection_prefs import SelectionPreferences
+from pip._internal.models.target_python import TargetPython
+from pip._internal.network.session import PipSession
+from pip._internal.utils.misc import write_output
+
+logger = logging.getLogger(__name__)
+
+
+class IndexCommand(IndexGroupCommand):
+    """
+    Inspect information available from package indexes.
+    """
+
+    ignore_require_venv = True
+    usage = """
+        %prog versions <package>
+    """
+
+    def add_options(self) -> None:
+        cmdoptions.add_target_python_options(self.cmd_opts)
+
+        self.cmd_opts.add_option(cmdoptions.ignore_requires_python())
+        self.cmd_opts.add_option(cmdoptions.pre())
+        self.cmd_opts.add_option(cmdoptions.no_binary())
+        self.cmd_opts.add_option(cmdoptions.only_binary())
+
+        index_opts = cmdoptions.make_option_group(
+            cmdoptions.index_group,
+            self.parser,
+        )
+
+        self.parser.insert_option_group(0, index_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        handlers = {
+            "versions": self.get_available_package_versions,
+        }
+
+        logger.warning(
+            "pip index is currently an experimental command. "
+            "It may be removed/changed in a future release "
+            "without prior warning."
+        )
+
+        # Determine action
+        if not args or args[0] not in handlers:
+            logger.error(
+                "Need an action (%s) to perform.",
+                ", ".join(sorted(handlers)),
+            )
+            return ERROR
+
+        action = args[0]
+
+        # Error handling happens here, not in the action-handlers.
+        try:
+            handlers[action](options, args[1:])
+        except PipError as e:
+            logger.error(e.args[0])
+            return ERROR
+
+        return SUCCESS
+
+    def _build_package_finder(
+        self,
+        options: Values,
+        session: PipSession,
+        target_python: Optional[TargetPython] = None,
+        ignore_requires_python: Optional[bool] = None,
+    ) -> PackageFinder:
+        """
+        Create a package finder appropriate to the index command.
+        """
+        link_collector = LinkCollector.create(session, options=options)
+
+        # Pass allow_yanked=False to ignore yanked versions.
+        selection_prefs = SelectionPreferences(
+            allow_yanked=False,
+            allow_all_prereleases=options.pre,
+            ignore_requires_python=ignore_requires_python,
+        )
+
+        return PackageFinder.create(
+            link_collector=link_collector,
+            selection_prefs=selection_prefs,
+            target_python=target_python,
+        )
+
+    def get_available_package_versions(self, options: Values, args: List[Any]) -> None:
+        if len(args) != 1:
+            raise CommandError("You need to specify exactly one argument")
+
+        target_python = cmdoptions.make_target_python(options)
+        query = args[0]
+
+        with self._build_session(options) as session:
+            finder = self._build_package_finder(
+                options=options,
+                session=session,
+                target_python=target_python,
+                ignore_requires_python=options.ignore_requires_python,
+            )
+
+            versions: Iterable[Version] = (
+                candidate.version for candidate in finder.find_all_candidates(query)
+            )
+
+            if not options.pre:
+                # Remove prereleases
+                versions = (
+                    version for version in versions if not version.is_prerelease
+                )
+            versions = set(versions)
+
+            if not versions:
+                raise DistributionNotFound(
+                    f"No matching distribution found for {query}"
+                )
+
+            formatted_versions = [str(ver) for ver in sorted(versions, reverse=True)]
+            latest = formatted_versions[0]
+
+        write_output(f"{query} ({latest})")
+        write_output("Available versions: {}".format(", ".join(formatted_versions)))
+        print_dist_installation_info(query, latest)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/inspect.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/inspect.py
new file mode 100644
index 00000000..e810c131
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/inspect.py
@@ -0,0 +1,92 @@
+import logging
+from optparse import Values
+from typing import Any, Dict, List
+
+from pip._vendor.packaging.markers import default_environment
+from pip._vendor.rich import print_json
+
+from pip import __version__
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.metadata import BaseDistribution, get_environment
+from pip._internal.utils.compat import stdlib_pkgs
+from pip._internal.utils.urls import path_to_url
+
+logger = logging.getLogger(__name__)
+
+
+class InspectCommand(Command):
+    """
+    Inspect the content of a Python environment and produce a report in JSON format.
+    """
+
+    ignore_require_venv = True
+    usage = """
+      %prog [options]"""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "--local",
+            action="store_true",
+            default=False,
+            help=(
+                "If in a virtualenv that has global access, do not list "
+                "globally-installed packages."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "--user",
+            dest="user",
+            action="store_true",
+            default=False,
+            help="Only output packages installed in user-site.",
+        )
+        self.cmd_opts.add_option(cmdoptions.list_path())
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        cmdoptions.check_list_path_option(options)
+        dists = get_environment(options.path).iter_installed_distributions(
+            local_only=options.local,
+            user_only=options.user,
+            skip=set(stdlib_pkgs),
+        )
+        output = {
+            "version": "1",
+            "pip_version": __version__,
+            "installed": [self._dist_to_dict(dist) for dist in dists],
+            "environment": default_environment(),
+            # TODO tags? scheme?
+        }
+        print_json(data=output)
+        return SUCCESS
+
+    def _dist_to_dict(self, dist: BaseDistribution) -> Dict[str, Any]:
+        res: Dict[str, Any] = {
+            "metadata": dist.metadata_dict,
+            "metadata_location": dist.info_location,
+        }
+        # direct_url. Note that we don't have download_info (as in the installation
+        # report) since it is not recorded in installed metadata.
+        direct_url = dist.direct_url
+        if direct_url is not None:
+            res["direct_url"] = direct_url.to_dict()
+        else:
+            # Emulate direct_url for legacy editable installs.
+            editable_project_location = dist.editable_project_location
+            if editable_project_location is not None:
+                res["direct_url"] = {
+                    "url": path_to_url(editable_project_location),
+                    "dir_info": {
+                        "editable": True,
+                    },
+                }
+        # installer
+        installer = dist.installer
+        if dist.installer:
+            res["installer"] = installer
+        # requested
+        if dist.installed_with_dist_info:
+            res["requested"] = dist.requested
+        return res
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/install.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/install.py
new file mode 100644
index 00000000..232a34a6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/install.py
@@ -0,0 +1,784 @@
+import errno
+import json
+import operator
+import os
+import shutil
+import site
+from optparse import SUPPRESS_HELP, Values
+from typing import List, Optional
+
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.rich import print_json
+
+# Eagerly import self_outdated_check to avoid crashes. Otherwise,
+# this module would be imported *after* pip was replaced, resulting
+# in crashes if the new self_outdated_check module was incompatible
+# with the rest of pip that's already imported, or allowing a
+# wheel to execute arbitrary code on install by replacing
+# self_outdated_check.
+import pip._internal.self_outdated_check  # noqa: F401
+from pip._internal.cache import WheelCache
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.cmdoptions import make_target_python
+from pip._internal.cli.req_command import (
+    RequirementCommand,
+    with_cleanup,
+)
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.exceptions import CommandError, InstallationError
+from pip._internal.locations import get_scheme
+from pip._internal.metadata import get_environment
+from pip._internal.models.installation_report import InstallationReport
+from pip._internal.operations.build.build_tracker import get_build_tracker
+from pip._internal.operations.check import ConflictDetails, check_install_conflicts
+from pip._internal.req import install_given_reqs
+from pip._internal.req.req_install import (
+    InstallRequirement,
+    check_legacy_setup_py_options,
+)
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.filesystem import test_writable_dir
+from pip._internal.utils.logging import getLogger
+from pip._internal.utils.misc import (
+    check_externally_managed,
+    ensure_dir,
+    get_pip_version,
+    protect_pip_from_modification_on_windows,
+    warn_if_run_as_root,
+    write_output,
+)
+from pip._internal.utils.temp_dir import TempDirectory
+from pip._internal.utils.virtualenv import (
+    running_under_virtualenv,
+    virtualenv_no_global,
+)
+from pip._internal.wheel_builder import build, should_build_for_install_command
+
+logger = getLogger(__name__)
+
+
+class InstallCommand(RequirementCommand):
+    """
+    Install packages from:
+
+    - PyPI (and other indexes) using requirement specifiers.
+    - VCS project urls.
+    - Local project directories.
+    - Local or remote source archives.
+
+    pip also supports installing from "requirements files", which provide
+    an easy way to specify a whole environment to be installed.
+    """
+
+    usage = """
+      %prog [options] <requirement specifier> [package-index-options] ...
+      %prog [options] -r <requirements file> [package-index-options] ...
+      %prog [options] [-e] <vcs project url> ...
+      %prog [options] [-e] <local project path> ...
+      %prog [options] <archive url/path> ..."""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(cmdoptions.requirements())
+        self.cmd_opts.add_option(cmdoptions.constraints())
+        self.cmd_opts.add_option(cmdoptions.no_deps())
+        self.cmd_opts.add_option(cmdoptions.pre())
+
+        self.cmd_opts.add_option(cmdoptions.editable())
+        self.cmd_opts.add_option(
+            "--dry-run",
+            action="store_true",
+            dest="dry_run",
+            default=False,
+            help=(
+                "Don't actually install anything, just print what would be. "
+                "Can be used in combination with --ignore-installed "
+                "to 'resolve' the requirements."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "-t",
+            "--target",
+            dest="target_dir",
+            metavar="dir",
+            default=None,
+            help=(
+                "Install packages into <dir>. "
+                "By default this will not replace existing files/folders in "
+                "<dir>. Use --upgrade to replace existing packages in <dir> "
+                "with new versions."
+            ),
+        )
+        cmdoptions.add_target_python_options(self.cmd_opts)
+
+        self.cmd_opts.add_option(
+            "--user",
+            dest="use_user_site",
+            action="store_true",
+            help=(
+                "Install to the Python user install directory for your "
+                "platform. Typically ~/.local/, or %APPDATA%\\Python on "
+                "Windows. (See the Python documentation for site.USER_BASE "
+                "for full details.)"
+            ),
+        )
+        self.cmd_opts.add_option(
+            "--no-user",
+            dest="use_user_site",
+            action="store_false",
+            help=SUPPRESS_HELP,
+        )
+        self.cmd_opts.add_option(
+            "--root",
+            dest="root_path",
+            metavar="dir",
+            default=None,
+            help="Install everything relative to this alternate root directory.",
+        )
+        self.cmd_opts.add_option(
+            "--prefix",
+            dest="prefix_path",
+            metavar="dir",
+            default=None,
+            help=(
+                "Installation prefix where lib, bin and other top-level "
+                "folders are placed. Note that the resulting installation may "
+                "contain scripts and other resources which reference the "
+                "Python interpreter of pip, and not that of ``--prefix``. "
+                "See also the ``--python`` option if the intention is to "
+                "install packages into another (possibly pip-free) "
+                "environment."
+            ),
+        )
+
+        self.cmd_opts.add_option(cmdoptions.src())
+
+        self.cmd_opts.add_option(
+            "-U",
+            "--upgrade",
+            dest="upgrade",
+            action="store_true",
+            help=(
+                "Upgrade all specified packages to the newest available "
+                "version. The handling of dependencies depends on the "
+                "upgrade-strategy used."
+            ),
+        )
+
+        self.cmd_opts.add_option(
+            "--upgrade-strategy",
+            dest="upgrade_strategy",
+            default="only-if-needed",
+            choices=["only-if-needed", "eager"],
+            help=(
+                "Determines how dependency upgrading should be handled "
+                "[default: %default]. "
+                '"eager" - dependencies are upgraded regardless of '
+                "whether the currently installed version satisfies the "
+                "requirements of the upgraded package(s). "
+                '"only-if-needed" -  are upgraded only when they do not '
+                "satisfy the requirements of the upgraded package(s)."
+            ),
+        )
+
+        self.cmd_opts.add_option(
+            "--force-reinstall",
+            dest="force_reinstall",
+            action="store_true",
+            help="Reinstall all packages even if they are already up-to-date.",
+        )
+
+        self.cmd_opts.add_option(
+            "-I",
+            "--ignore-installed",
+            dest="ignore_installed",
+            action="store_true",
+            help=(
+                "Ignore the installed packages, overwriting them. "
+                "This can break your system if the existing package "
+                "is of a different version or was installed "
+                "with a different package manager!"
+            ),
+        )
+
+        self.cmd_opts.add_option(cmdoptions.ignore_requires_python())
+        self.cmd_opts.add_option(cmdoptions.no_build_isolation())
+        self.cmd_opts.add_option(cmdoptions.use_pep517())
+        self.cmd_opts.add_option(cmdoptions.no_use_pep517())
+        self.cmd_opts.add_option(cmdoptions.check_build_deps())
+        self.cmd_opts.add_option(cmdoptions.override_externally_managed())
+
+        self.cmd_opts.add_option(cmdoptions.config_settings())
+        self.cmd_opts.add_option(cmdoptions.global_options())
+
+        self.cmd_opts.add_option(
+            "--compile",
+            action="store_true",
+            dest="compile",
+            default=True,
+            help="Compile Python source files to bytecode",
+        )
+
+        self.cmd_opts.add_option(
+            "--no-compile",
+            action="store_false",
+            dest="compile",
+            help="Do not compile Python source files to bytecode",
+        )
+
+        self.cmd_opts.add_option(
+            "--no-warn-script-location",
+            action="store_false",
+            dest="warn_script_location",
+            default=True,
+            help="Do not warn when installing scripts outside PATH",
+        )
+        self.cmd_opts.add_option(
+            "--no-warn-conflicts",
+            action="store_false",
+            dest="warn_about_conflicts",
+            default=True,
+            help="Do not warn about broken dependencies",
+        )
+        self.cmd_opts.add_option(cmdoptions.no_binary())
+        self.cmd_opts.add_option(cmdoptions.only_binary())
+        self.cmd_opts.add_option(cmdoptions.prefer_binary())
+        self.cmd_opts.add_option(cmdoptions.require_hashes())
+        self.cmd_opts.add_option(cmdoptions.progress_bar())
+        self.cmd_opts.add_option(cmdoptions.root_user_action())
+
+        index_opts = cmdoptions.make_option_group(
+            cmdoptions.index_group,
+            self.parser,
+        )
+
+        self.parser.insert_option_group(0, index_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+        self.cmd_opts.add_option(
+            "--report",
+            dest="json_report_file",
+            metavar="file",
+            default=None,
+            help=(
+                "Generate a JSON file describing what pip did to install "
+                "the provided requirements. "
+                "Can be used in combination with --dry-run and --ignore-installed "
+                "to 'resolve' the requirements. "
+                "When - is used as file name it writes to stdout. "
+                "When writing to stdout, please combine with the --quiet option "
+                "to avoid mixing pip logging output with JSON output."
+            ),
+        )
+
+    @with_cleanup
+    def run(self, options: Values, args: List[str]) -> int:
+        if options.use_user_site and options.target_dir is not None:
+            raise CommandError("Can not combine '--user' and '--target'")
+
+        # Check whether the environment we're installing into is externally
+        # managed, as specified in PEP 668. Specifying --root, --target, or
+        # --prefix disables the check, since there's no reliable way to locate
+        # the EXTERNALLY-MANAGED file for those cases. An exception is also
+        # made specifically for "--dry-run --report" for convenience.
+        installing_into_current_environment = (
+            not (options.dry_run and options.json_report_file)
+            and options.root_path is None
+            and options.target_dir is None
+            and options.prefix_path is None
+        )
+        if (
+            installing_into_current_environment
+            and not options.override_externally_managed
+        ):
+            check_externally_managed()
+
+        upgrade_strategy = "to-satisfy-only"
+        if options.upgrade:
+            upgrade_strategy = options.upgrade_strategy
+
+        cmdoptions.check_dist_restriction(options, check_target=True)
+
+        logger.verbose("Using %s", get_pip_version())
+        options.use_user_site = decide_user_install(
+            options.use_user_site,
+            prefix_path=options.prefix_path,
+            target_dir=options.target_dir,
+            root_path=options.root_path,
+            isolated_mode=options.isolated_mode,
+        )
+
+        target_temp_dir: Optional[TempDirectory] = None
+        target_temp_dir_path: Optional[str] = None
+        if options.target_dir:
+            options.ignore_installed = True
+            options.target_dir = os.path.abspath(options.target_dir)
+            if (
+                # fmt: off
+                os.path.exists(options.target_dir) and
+                not os.path.isdir(options.target_dir)
+                # fmt: on
+            ):
+                raise CommandError(
+                    "Target path exists but is not a directory, will not continue."
+                )
+
+            # Create a target directory for using with the target option
+            target_temp_dir = TempDirectory(kind="target")
+            target_temp_dir_path = target_temp_dir.path
+            self.enter_context(target_temp_dir)
+
+        global_options = options.global_options or []
+
+        session = self.get_default_session(options)
+
+        target_python = make_target_python(options)
+        finder = self._build_package_finder(
+            options=options,
+            session=session,
+            target_python=target_python,
+            ignore_requires_python=options.ignore_requires_python,
+        )
+        build_tracker = self.enter_context(get_build_tracker())
+
+        directory = TempDirectory(
+            delete=not options.no_clean,
+            kind="install",
+            globally_managed=True,
+        )
+
+        try:
+            reqs = self.get_requirements(args, options, finder, session)
+            check_legacy_setup_py_options(options, reqs)
+
+            wheel_cache = WheelCache(options.cache_dir)
+
+            # Only when installing is it permitted to use PEP 660.
+            # In other circumstances (pip wheel, pip download) we generate
+            # regular (i.e. non editable) metadata and wheels.
+            for req in reqs:
+                req.permit_editable_wheels = True
+
+            preparer = self.make_requirement_preparer(
+                temp_build_dir=directory,
+                options=options,
+                build_tracker=build_tracker,
+                session=session,
+                finder=finder,
+                use_user_site=options.use_user_site,
+                verbosity=self.verbosity,
+            )
+            resolver = self.make_resolver(
+                preparer=preparer,
+                finder=finder,
+                options=options,
+                wheel_cache=wheel_cache,
+                use_user_site=options.use_user_site,
+                ignore_installed=options.ignore_installed,
+                ignore_requires_python=options.ignore_requires_python,
+                force_reinstall=options.force_reinstall,
+                upgrade_strategy=upgrade_strategy,
+                use_pep517=options.use_pep517,
+                py_version_info=options.python_version,
+            )
+
+            self.trace_basic_info(finder)
+
+            requirement_set = resolver.resolve(
+                reqs, check_supported_wheels=not options.target_dir
+            )
+
+            if options.json_report_file:
+                report = InstallationReport(requirement_set.requirements_to_install)
+                if options.json_report_file == "-":
+                    print_json(data=report.to_dict())
+                else:
+                    with open(options.json_report_file, "w", encoding="utf-8") as f:
+                        json.dump(report.to_dict(), f, indent=2, ensure_ascii=False)
+
+            if options.dry_run:
+                would_install_items = sorted(
+                    (r.metadata["name"], r.metadata["version"])
+                    for r in requirement_set.requirements_to_install
+                )
+                if would_install_items:
+                    write_output(
+                        "Would install %s",
+                        " ".join("-".join(item) for item in would_install_items),
+                    )
+                return SUCCESS
+
+            try:
+                pip_req = requirement_set.get_requirement("pip")
+            except KeyError:
+                modifying_pip = False
+            else:
+                # If we're not replacing an already installed pip,
+                # we're not modifying it.
+                modifying_pip = pip_req.satisfied_by is None
+            protect_pip_from_modification_on_windows(modifying_pip=modifying_pip)
+
+            reqs_to_build = [
+                r
+                for r in requirement_set.requirements.values()
+                if should_build_for_install_command(r)
+            ]
+
+            _, build_failures = build(
+                reqs_to_build,
+                wheel_cache=wheel_cache,
+                verify=True,
+                build_options=[],
+                global_options=global_options,
+            )
+
+            if build_failures:
+                raise InstallationError(
+                    "Failed to build installable wheels for some "
+                    "pyproject.toml based projects ({})".format(
+                        ", ".join(r.name for r in build_failures)  # type: ignore
+                    )
+                )
+
+            to_install = resolver.get_installation_order(requirement_set)
+
+            # Check for conflicts in the package set we're installing.
+            conflicts: Optional[ConflictDetails] = None
+            should_warn_about_conflicts = (
+                not options.ignore_dependencies and options.warn_about_conflicts
+            )
+            if should_warn_about_conflicts:
+                conflicts = self._determine_conflicts(to_install)
+
+            # Don't warn about script install locations if
+            # --target or --prefix has been specified
+            warn_script_location = options.warn_script_location
+            if options.target_dir or options.prefix_path:
+                warn_script_location = False
+
+            installed = install_given_reqs(
+                to_install,
+                global_options,
+                root=options.root_path,
+                home=target_temp_dir_path,
+                prefix=options.prefix_path,
+                warn_script_location=warn_script_location,
+                use_user_site=options.use_user_site,
+                pycompile=options.compile,
+            )
+
+            lib_locations = get_lib_location_guesses(
+                user=options.use_user_site,
+                home=target_temp_dir_path,
+                root=options.root_path,
+                prefix=options.prefix_path,
+                isolated=options.isolated_mode,
+            )
+            env = get_environment(lib_locations)
+
+            # Display a summary of installed packages, with extra care to
+            # display a package name as it was requested by the user.
+            installed.sort(key=operator.attrgetter("name"))
+            summary = []
+            installed_versions = {}
+            for distribution in env.iter_all_distributions():
+                installed_versions[distribution.canonical_name] = distribution.version
+            for package in installed:
+                display_name = package.name
+                version = installed_versions.get(canonicalize_name(display_name), None)
+                if version:
+                    text = f"{display_name}-{version}"
+                else:
+                    text = display_name
+                summary.append(text)
+
+            if conflicts is not None:
+                self._warn_about_conflicts(
+                    conflicts,
+                    resolver_variant=self.determine_resolver_variant(options),
+                )
+
+            installed_desc = " ".join(summary)
+            if installed_desc:
+                write_output(
+                    "Successfully installed %s",
+                    installed_desc,
+                )
+        except OSError as error:
+            show_traceback = self.verbosity >= 1
+
+            message = create_os_error_message(
+                error,
+                show_traceback,
+                options.use_user_site,
+            )
+            logger.error(message, exc_info=show_traceback)
+
+            return ERROR
+
+        if options.target_dir:
+            assert target_temp_dir
+            self._handle_target_dir(
+                options.target_dir, target_temp_dir, options.upgrade
+            )
+        if options.root_user_action == "warn":
+            warn_if_run_as_root()
+        return SUCCESS
+
+    def _handle_target_dir(
+        self, target_dir: str, target_temp_dir: TempDirectory, upgrade: bool
+    ) -> None:
+        ensure_dir(target_dir)
+
+        # Checking both purelib and platlib directories for installed
+        # packages to be moved to target directory
+        lib_dir_list = []
+
+        # Checking both purelib and platlib directories for installed
+        # packages to be moved to target directory
+        scheme = get_scheme("", home=target_temp_dir.path)
+        purelib_dir = scheme.purelib
+        platlib_dir = scheme.platlib
+        data_dir = scheme.data
+
+        if os.path.exists(purelib_dir):
+            lib_dir_list.append(purelib_dir)
+        if os.path.exists(platlib_dir) and platlib_dir != purelib_dir:
+            lib_dir_list.append(platlib_dir)
+        if os.path.exists(data_dir):
+            lib_dir_list.append(data_dir)
+
+        for lib_dir in lib_dir_list:
+            for item in os.listdir(lib_dir):
+                if lib_dir == data_dir:
+                    ddir = os.path.join(data_dir, item)
+                    if any(s.startswith(ddir) for s in lib_dir_list[:-1]):
+                        continue
+                target_item_dir = os.path.join(target_dir, item)
+                if os.path.exists(target_item_dir):
+                    if not upgrade:
+                        logger.warning(
+                            "Target directory %s already exists. Specify "
+                            "--upgrade to force replacement.",
+                            target_item_dir,
+                        )
+                        continue
+                    if os.path.islink(target_item_dir):
+                        logger.warning(
+                            "Target directory %s already exists and is "
+                            "a link. pip will not automatically replace "
+                            "links, please remove if replacement is "
+                            "desired.",
+                            target_item_dir,
+                        )
+                        continue
+                    if os.path.isdir(target_item_dir):
+                        shutil.rmtree(target_item_dir)
+                    else:
+                        os.remove(target_item_dir)
+
+                shutil.move(os.path.join(lib_dir, item), target_item_dir)
+
+    def _determine_conflicts(
+        self, to_install: List[InstallRequirement]
+    ) -> Optional[ConflictDetails]:
+        try:
+            return check_install_conflicts(to_install)
+        except Exception:
+            logger.exception(
+                "Error while checking for conflicts. Please file an issue on "
+                "pip's issue tracker: https://github.com/pypa/pip/issues/new"
+            )
+            return None
+
+    def _warn_about_conflicts(
+        self, conflict_details: ConflictDetails, resolver_variant: str
+    ) -> None:
+        package_set, (missing, conflicting) = conflict_details
+        if not missing and not conflicting:
+            return
+
+        parts: List[str] = []
+        if resolver_variant == "legacy":
+            parts.append(
+                "pip's legacy dependency resolver does not consider dependency "
+                "conflicts when selecting packages. This behaviour is the "
+                "source of the following dependency conflicts."
+            )
+        else:
+            assert resolver_variant == "resolvelib"
+            parts.append(
+                "pip's dependency resolver does not currently take into account "
+                "all the packages that are installed. This behaviour is the "
+                "source of the following dependency conflicts."
+            )
+
+        # NOTE: There is some duplication here, with commands/check.py
+        for project_name in missing:
+            version = package_set[project_name][0]
+            for dependency in missing[project_name]:
+                message = (
+                    f"{project_name} {version} requires {dependency[1]}, "
+                    "which is not installed."
+                )
+                parts.append(message)
+
+        for project_name in conflicting:
+            version = package_set[project_name][0]
+            for dep_name, dep_version, req in conflicting[project_name]:
+                message = (
+                    "{name} {version} requires {requirement}, but {you} have "
+                    "{dep_name} {dep_version} which is incompatible."
+                ).format(
+                    name=project_name,
+                    version=version,
+                    requirement=req,
+                    dep_name=dep_name,
+                    dep_version=dep_version,
+                    you=("you" if resolver_variant == "resolvelib" else "you'll"),
+                )
+                parts.append(message)
+
+        logger.critical("\n".join(parts))
+
+
+def get_lib_location_guesses(
+    user: bool = False,
+    home: Optional[str] = None,
+    root: Optional[str] = None,
+    isolated: bool = False,
+    prefix: Optional[str] = None,
+) -> List[str]:
+    scheme = get_scheme(
+        "",
+        user=user,
+        home=home,
+        root=root,
+        isolated=isolated,
+        prefix=prefix,
+    )
+    return [scheme.purelib, scheme.platlib]
+
+
+def site_packages_writable(root: Optional[str], isolated: bool) -> bool:
+    return all(
+        test_writable_dir(d)
+        for d in set(get_lib_location_guesses(root=root, isolated=isolated))
+    )
+
+
+def decide_user_install(
+    use_user_site: Optional[bool],
+    prefix_path: Optional[str] = None,
+    target_dir: Optional[str] = None,
+    root_path: Optional[str] = None,
+    isolated_mode: bool = False,
+) -> bool:
+    """Determine whether to do a user install based on the input options.
+
+    If use_user_site is False, no additional checks are done.
+    If use_user_site is True, it is checked for compatibility with other
+    options.
+    If use_user_site is None, the default behaviour depends on the environment,
+    which is provided by the other arguments.
+    """
+    # In some cases (config from tox), use_user_site can be set to an integer
+    # rather than a bool, which 'use_user_site is False' wouldn't catch.
+    if (use_user_site is not None) and (not use_user_site):
+        logger.debug("Non-user install by explicit request")
+        return False
+
+    if use_user_site:
+        if prefix_path:
+            raise CommandError(
+                "Can not combine '--user' and '--prefix' as they imply "
+                "different installation locations"
+            )
+        if virtualenv_no_global():
+            raise InstallationError(
+                "Can not perform a '--user' install. User site-packages "
+                "are not visible in this virtualenv."
+            )
+        logger.debug("User install by explicit request")
+        return True
+
+    # If we are here, user installs have not been explicitly requested/avoided
+    assert use_user_site is None
+
+    # user install incompatible with --prefix/--target
+    if prefix_path or target_dir:
+        logger.debug("Non-user install due to --prefix or --target option")
+        return False
+
+    # If user installs are not enabled, choose a non-user install
+    if not site.ENABLE_USER_SITE:
+        logger.debug("Non-user install because user site-packages disabled")
+        return False
+
+    # If we have permission for a non-user install, do that,
+    # otherwise do a user install.
+    if site_packages_writable(root=root_path, isolated=isolated_mode):
+        logger.debug("Non-user install because site-packages writeable")
+        return False
+
+    logger.info(
+        "Defaulting to user installation because normal site-packages "
+        "is not writeable"
+    )
+    return True
+
+
+def create_os_error_message(
+    error: OSError, show_traceback: bool, using_user_site: bool
+) -> str:
+    """Format an error message for an OSError
+
+    It may occur anytime during the execution of the install command.
+    """
+    parts = []
+
+    # Mention the error if we are not going to show a traceback
+    parts.append("Could not install packages due to an OSError")
+    if not show_traceback:
+        parts.append(": ")
+        parts.append(str(error))
+    else:
+        parts.append(".")
+
+    # Spilt the error indication from a helper message (if any)
+    parts[-1] += "\n"
+
+    # Suggest useful actions to the user:
+    #  (1) using user site-packages or (2) verifying the permissions
+    if error.errno == errno.EACCES:
+        user_option_part = "Consider using the `--user` option"
+        permissions_part = "Check the permissions"
+
+        if not running_under_virtualenv() and not using_user_site:
+            parts.extend(
+                [
+                    user_option_part,
+                    " or ",
+                    permissions_part.lower(),
+                ]
+            )
+        else:
+            parts.append(permissions_part)
+        parts.append(".\n")
+
+    # Suggest the user to enable Long Paths if path length is
+    # more than 260
+    if (
+        WINDOWS
+        and error.errno == errno.ENOENT
+        and error.filename
+        and len(error.filename) > 260
+    ):
+        parts.append(
+            "HINT: This error might have occurred since "
+            "this system does not have Windows Long Path "
+            "support enabled. You can find information on "
+            "how to enable this at "
+            "https://pip.pypa.io/warnings/enable-long-paths\n"
+        )
+
+    return "".join(parts).strip() + "\n"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/list.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/list.py
new file mode 100644
index 00000000..84943702
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/list.py
@@ -0,0 +1,375 @@
+import json
+import logging
+from optparse import Values
+from typing import TYPE_CHECKING, Generator, List, Optional, Sequence, Tuple, cast
+
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.packaging.version import Version
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.index_command import IndexGroupCommand
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.exceptions import CommandError
+from pip._internal.metadata import BaseDistribution, get_environment
+from pip._internal.models.selection_prefs import SelectionPreferences
+from pip._internal.utils.compat import stdlib_pkgs
+from pip._internal.utils.misc import tabulate, write_output
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+    from pip._internal.network.session import PipSession
+
+    class _DistWithLatestInfo(BaseDistribution):
+        """Give the distribution object a couple of extra fields.
+
+        These will be populated during ``get_outdated()``. This is dirty but
+        makes the rest of the code much cleaner.
+        """
+
+        latest_version: Version
+        latest_filetype: str
+
+    _ProcessedDists = Sequence[_DistWithLatestInfo]
+
+
+logger = logging.getLogger(__name__)
+
+
+class ListCommand(IndexGroupCommand):
+    """
+    List installed packages, including editables.
+
+    Packages are listed in a case-insensitive sorted order.
+    """
+
+    ignore_require_venv = True
+    usage = """
+      %prog [options]"""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-o",
+            "--outdated",
+            action="store_true",
+            default=False,
+            help="List outdated packages",
+        )
+        self.cmd_opts.add_option(
+            "-u",
+            "--uptodate",
+            action="store_true",
+            default=False,
+            help="List uptodate packages",
+        )
+        self.cmd_opts.add_option(
+            "-e",
+            "--editable",
+            action="store_true",
+            default=False,
+            help="List editable projects.",
+        )
+        self.cmd_opts.add_option(
+            "-l",
+            "--local",
+            action="store_true",
+            default=False,
+            help=(
+                "If in a virtualenv that has global access, do not list "
+                "globally-installed packages."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "--user",
+            dest="user",
+            action="store_true",
+            default=False,
+            help="Only output packages installed in user-site.",
+        )
+        self.cmd_opts.add_option(cmdoptions.list_path())
+        self.cmd_opts.add_option(
+            "--pre",
+            action="store_true",
+            default=False,
+            help=(
+                "Include pre-release and development versions. By default, "
+                "pip only finds stable versions."
+            ),
+        )
+
+        self.cmd_opts.add_option(
+            "--format",
+            action="store",
+            dest="list_format",
+            default="columns",
+            choices=("columns", "freeze", "json"),
+            help=(
+                "Select the output format among: columns (default), freeze, or json. "
+                "The 'freeze' format cannot be used with the --outdated option."
+            ),
+        )
+
+        self.cmd_opts.add_option(
+            "--not-required",
+            action="store_true",
+            dest="not_required",
+            help="List packages that are not dependencies of installed packages.",
+        )
+
+        self.cmd_opts.add_option(
+            "--exclude-editable",
+            action="store_false",
+            dest="include_editable",
+            help="Exclude editable package from output.",
+        )
+        self.cmd_opts.add_option(
+            "--include-editable",
+            action="store_true",
+            dest="include_editable",
+            help="Include editable package from output.",
+            default=True,
+        )
+        self.cmd_opts.add_option(cmdoptions.list_exclude())
+        index_opts = cmdoptions.make_option_group(cmdoptions.index_group, self.parser)
+
+        self.parser.insert_option_group(0, index_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def handle_pip_version_check(self, options: Values) -> None:
+        if options.outdated or options.uptodate:
+            super().handle_pip_version_check(options)
+
+    def _build_package_finder(
+        self, options: Values, session: "PipSession"
+    ) -> "PackageFinder":
+        """
+        Create a package finder appropriate to this list command.
+        """
+        # Lazy import the heavy index modules as most list invocations won't need 'em.
+        from pip._internal.index.collector import LinkCollector
+        from pip._internal.index.package_finder import PackageFinder
+
+        link_collector = LinkCollector.create(session, options=options)
+
+        # Pass allow_yanked=False to ignore yanked versions.
+        selection_prefs = SelectionPreferences(
+            allow_yanked=False,
+            allow_all_prereleases=options.pre,
+        )
+
+        return PackageFinder.create(
+            link_collector=link_collector,
+            selection_prefs=selection_prefs,
+        )
+
+    def run(self, options: Values, args: List[str]) -> int:
+        if options.outdated and options.uptodate:
+            raise CommandError("Options --outdated and --uptodate cannot be combined.")
+
+        if options.outdated and options.list_format == "freeze":
+            raise CommandError(
+                "List format 'freeze' cannot be used with the --outdated option."
+            )
+
+        cmdoptions.check_list_path_option(options)
+
+        skip = set(stdlib_pkgs)
+        if options.excludes:
+            skip.update(canonicalize_name(n) for n in options.excludes)
+
+        packages: _ProcessedDists = [
+            cast("_DistWithLatestInfo", d)
+            for d in get_environment(options.path).iter_installed_distributions(
+                local_only=options.local,
+                user_only=options.user,
+                editables_only=options.editable,
+                include_editables=options.include_editable,
+                skip=skip,
+            )
+        ]
+
+        # get_not_required must be called firstly in order to find and
+        # filter out all dependencies correctly. Otherwise a package
+        # can't be identified as requirement because some parent packages
+        # could be filtered out before.
+        if options.not_required:
+            packages = self.get_not_required(packages, options)
+
+        if options.outdated:
+            packages = self.get_outdated(packages, options)
+        elif options.uptodate:
+            packages = self.get_uptodate(packages, options)
+
+        self.output_package_listing(packages, options)
+        return SUCCESS
+
+    def get_outdated(
+        self, packages: "_ProcessedDists", options: Values
+    ) -> "_ProcessedDists":
+        return [
+            dist
+            for dist in self.iter_packages_latest_infos(packages, options)
+            if dist.latest_version > dist.version
+        ]
+
+    def get_uptodate(
+        self, packages: "_ProcessedDists", options: Values
+    ) -> "_ProcessedDists":
+        return [
+            dist
+            for dist in self.iter_packages_latest_infos(packages, options)
+            if dist.latest_version == dist.version
+        ]
+
+    def get_not_required(
+        self, packages: "_ProcessedDists", options: Values
+    ) -> "_ProcessedDists":
+        dep_keys = {
+            canonicalize_name(dep.name)
+            for dist in packages
+            for dep in (dist.iter_dependencies() or ())
+        }
+
+        # Create a set to remove duplicate packages, and cast it to a list
+        # to keep the return type consistent with get_outdated and
+        # get_uptodate
+        return list({pkg for pkg in packages if pkg.canonical_name not in dep_keys})
+
+    def iter_packages_latest_infos(
+        self, packages: "_ProcessedDists", options: Values
+    ) -> Generator["_DistWithLatestInfo", None, None]:
+        with self._build_session(options) as session:
+            finder = self._build_package_finder(options, session)
+
+            def latest_info(
+                dist: "_DistWithLatestInfo",
+            ) -> Optional["_DistWithLatestInfo"]:
+                all_candidates = finder.find_all_candidates(dist.canonical_name)
+                if not options.pre:
+                    # Remove prereleases
+                    all_candidates = [
+                        candidate
+                        for candidate in all_candidates
+                        if not candidate.version.is_prerelease
+                    ]
+
+                evaluator = finder.make_candidate_evaluator(
+                    project_name=dist.canonical_name,
+                )
+                best_candidate = evaluator.sort_best_candidate(all_candidates)
+                if best_candidate is None:
+                    return None
+
+                remote_version = best_candidate.version
+                if best_candidate.link.is_wheel:
+                    typ = "wheel"
+                else:
+                    typ = "sdist"
+                dist.latest_version = remote_version
+                dist.latest_filetype = typ
+                return dist
+
+            for dist in map(latest_info, packages):
+                if dist is not None:
+                    yield dist
+
+    def output_package_listing(
+        self, packages: "_ProcessedDists", options: Values
+    ) -> None:
+        packages = sorted(
+            packages,
+            key=lambda dist: dist.canonical_name,
+        )
+        if options.list_format == "columns" and packages:
+            data, header = format_for_columns(packages, options)
+            self.output_package_listing_columns(data, header)
+        elif options.list_format == "freeze":
+            for dist in packages:
+                if options.verbose >= 1:
+                    write_output(
+                        "%s==%s (%s)", dist.raw_name, dist.version, dist.location
+                    )
+                else:
+                    write_output("%s==%s", dist.raw_name, dist.version)
+        elif options.list_format == "json":
+            write_output(format_for_json(packages, options))
+
+    def output_package_listing_columns(
+        self, data: List[List[str]], header: List[str]
+    ) -> None:
+        # insert the header first: we need to know the size of column names
+        if len(data) > 0:
+            data.insert(0, header)
+
+        pkg_strings, sizes = tabulate(data)
+
+        # Create and add a separator.
+        if len(data) > 0:
+            pkg_strings.insert(1, " ".join("-" * x for x in sizes))
+
+        for val in pkg_strings:
+            write_output(val)
+
+
+def format_for_columns(
+    pkgs: "_ProcessedDists", options: Values
+) -> Tuple[List[List[str]], List[str]]:
+    """
+    Convert the package data into something usable
+    by output_package_listing_columns.
+    """
+    header = ["Package", "Version"]
+
+    running_outdated = options.outdated
+    if running_outdated:
+        header.extend(["Latest", "Type"])
+
+    has_editables = any(x.editable for x in pkgs)
+    if has_editables:
+        header.append("Editable project location")
+
+    if options.verbose >= 1:
+        header.append("Location")
+    if options.verbose >= 1:
+        header.append("Installer")
+
+    data = []
+    for proj in pkgs:
+        # if we're working on the 'outdated' list, separate out the
+        # latest_version and type
+        row = [proj.raw_name, proj.raw_version]
+
+        if running_outdated:
+            row.append(str(proj.latest_version))
+            row.append(proj.latest_filetype)
+
+        if has_editables:
+            row.append(proj.editable_project_location or "")
+
+        if options.verbose >= 1:
+            row.append(proj.location or "")
+        if options.verbose >= 1:
+            row.append(proj.installer)
+
+        data.append(row)
+
+    return data, header
+
+
+def format_for_json(packages: "_ProcessedDists", options: Values) -> str:
+    data = []
+    for dist in packages:
+        info = {
+            "name": dist.raw_name,
+            "version": str(dist.version),
+        }
+        if options.verbose >= 1:
+            info["location"] = dist.location or ""
+            info["installer"] = dist.installer
+        if options.outdated:
+            info["latest_version"] = str(dist.latest_version)
+            info["latest_filetype"] = dist.latest_filetype
+        editable_project_location = dist.editable_project_location
+        if editable_project_location:
+            info["editable_project_location"] = editable_project_location
+        data.append(info)
+    return json.dumps(data)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/search.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/search.py
new file mode 100644
index 00000000..74b8d656
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/search.py
@@ -0,0 +1,172 @@
+import logging
+import shutil
+import sys
+import textwrap
+import xmlrpc.client
+from collections import OrderedDict
+from optparse import Values
+from typing import TYPE_CHECKING, Dict, List, Optional, TypedDict
+
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.req_command import SessionCommandMixin
+from pip._internal.cli.status_codes import NO_MATCHES_FOUND, SUCCESS
+from pip._internal.exceptions import CommandError
+from pip._internal.metadata import get_default_environment
+from pip._internal.models.index import PyPI
+from pip._internal.network.xmlrpc import PipXmlrpcTransport
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import write_output
+
+if TYPE_CHECKING:
+
+    class TransformedHit(TypedDict):
+        name: str
+        summary: str
+        versions: List[str]
+
+
+logger = logging.getLogger(__name__)
+
+
+class SearchCommand(Command, SessionCommandMixin):
+    """Search for PyPI packages whose name or summary contains <query>."""
+
+    usage = """
+      %prog [options] <query>"""
+    ignore_require_venv = True
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-i",
+            "--index",
+            dest="index",
+            metavar="URL",
+            default=PyPI.pypi_url,
+            help="Base URL of Python Package Index (default %default)",
+        )
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        if not args:
+            raise CommandError("Missing required argument (search query).")
+        query = args
+        pypi_hits = self.search(query, options)
+        hits = transform_hits(pypi_hits)
+
+        terminal_width = None
+        if sys.stdout.isatty():
+            terminal_width = shutil.get_terminal_size()[0]
+
+        print_results(hits, terminal_width=terminal_width)
+        if pypi_hits:
+            return SUCCESS
+        return NO_MATCHES_FOUND
+
+    def search(self, query: List[str], options: Values) -> List[Dict[str, str]]:
+        index_url = options.index
+
+        session = self.get_default_session(options)
+
+        transport = PipXmlrpcTransport(index_url, session)
+        pypi = xmlrpc.client.ServerProxy(index_url, transport)
+        try:
+            hits = pypi.search({"name": query, "summary": query}, "or")
+        except xmlrpc.client.Fault as fault:
+            message = (
+                f"XMLRPC request failed [code: {fault.faultCode}]\n{fault.faultString}"
+            )
+            raise CommandError(message)
+        assert isinstance(hits, list)
+        return hits
+
+
+def transform_hits(hits: List[Dict[str, str]]) -> List["TransformedHit"]:
+    """
+    The list from pypi is really a list of versions. We want a list of
+    packages with the list of versions stored inline. This converts the
+    list from pypi into one we can use.
+    """
+    packages: Dict[str, TransformedHit] = OrderedDict()
+    for hit in hits:
+        name = hit["name"]
+        summary = hit["summary"]
+        version = hit["version"]
+
+        if name not in packages.keys():
+            packages[name] = {
+                "name": name,
+                "summary": summary,
+                "versions": [version],
+            }
+        else:
+            packages[name]["versions"].append(version)
+
+            # if this is the highest version, replace summary and score
+            if version == highest_version(packages[name]["versions"]):
+                packages[name]["summary"] = summary
+
+    return list(packages.values())
+
+
+def print_dist_installation_info(name: str, latest: str) -> None:
+    env = get_default_environment()
+    dist = env.get_distribution(name)
+    if dist is not None:
+        with indent_log():
+            if dist.version == latest:
+                write_output("INSTALLED: %s (latest)", dist.version)
+            else:
+                write_output("INSTALLED: %s", dist.version)
+                if parse_version(latest).pre:
+                    write_output(
+                        "LATEST:    %s (pre-release; install"
+                        " with `pip install --pre`)",
+                        latest,
+                    )
+                else:
+                    write_output("LATEST:    %s", latest)
+
+
+def print_results(
+    hits: List["TransformedHit"],
+    name_column_width: Optional[int] = None,
+    terminal_width: Optional[int] = None,
+) -> None:
+    if not hits:
+        return
+    if name_column_width is None:
+        name_column_width = (
+            max(
+                [
+                    len(hit["name"]) + len(highest_version(hit.get("versions", ["-"])))
+                    for hit in hits
+                ]
+            )
+            + 4
+        )
+
+    for hit in hits:
+        name = hit["name"]
+        summary = hit["summary"] or ""
+        latest = highest_version(hit.get("versions", ["-"]))
+        if terminal_width is not None:
+            target_width = terminal_width - name_column_width - 5
+            if target_width > 10:
+                # wrap and indent summary to fit terminal
+                summary_lines = textwrap.wrap(summary, target_width)
+                summary = ("\n" + " " * (name_column_width + 3)).join(summary_lines)
+
+        name_latest = f"{name} ({latest})"
+        line = f"{name_latest:{name_column_width}} - {summary}"
+        try:
+            write_output(line)
+            print_dist_installation_info(name, latest)
+        except UnicodeEncodeError:
+            pass
+
+
+def highest_version(versions: List[str]) -> str:
+    return max(versions, key=parse_version)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/show.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/show.py
new file mode 100644
index 00000000..b47500cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/show.py
@@ -0,0 +1,224 @@
+import logging
+from optparse import Values
+from typing import Generator, Iterable, Iterator, List, NamedTuple, Optional
+
+from pip._vendor.packaging.requirements import InvalidRequirement
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.status_codes import ERROR, SUCCESS
+from pip._internal.metadata import BaseDistribution, get_default_environment
+from pip._internal.utils.misc import write_output
+
+logger = logging.getLogger(__name__)
+
+
+class ShowCommand(Command):
+    """
+    Show information about one or more installed packages.
+
+    The output is in RFC-compliant mail header format.
+    """
+
+    usage = """
+      %prog [options] <package> ..."""
+    ignore_require_venv = True
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-f",
+            "--files",
+            dest="files",
+            action="store_true",
+            default=False,
+            help="Show the full list of installed files for each package.",
+        )
+
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        if not args:
+            logger.warning("ERROR: Please provide a package name or names.")
+            return ERROR
+        query = args
+
+        results = search_packages_info(query)
+        if not print_results(
+            results, list_files=options.files, verbose=options.verbose
+        ):
+            return ERROR
+        return SUCCESS
+
+
+class _PackageInfo(NamedTuple):
+    name: str
+    version: str
+    location: str
+    editable_project_location: Optional[str]
+    requires: List[str]
+    required_by: List[str]
+    installer: str
+    metadata_version: str
+    classifiers: List[str]
+    summary: str
+    homepage: str
+    project_urls: List[str]
+    author: str
+    author_email: str
+    license: str
+    license_expression: str
+    entry_points: List[str]
+    files: Optional[List[str]]
+
+
+def search_packages_info(query: List[str]) -> Generator[_PackageInfo, None, None]:
+    """
+    Gather details from installed distributions. Print distribution name,
+    version, location, and installed files. Installed files requires a
+    pip generated 'installed-files.txt' in the distributions '.egg-info'
+    directory.
+    """
+    env = get_default_environment()
+
+    installed = {dist.canonical_name: dist for dist in env.iter_all_distributions()}
+    query_names = [canonicalize_name(name) for name in query]
+    missing = sorted(
+        [name for name, pkg in zip(query, query_names) if pkg not in installed]
+    )
+    if missing:
+        logger.warning("Package(s) not found: %s", ", ".join(missing))
+
+    def _get_requiring_packages(current_dist: BaseDistribution) -> Iterator[str]:
+        return (
+            dist.metadata["Name"] or "UNKNOWN"
+            for dist in installed.values()
+            if current_dist.canonical_name
+            in {canonicalize_name(d.name) for d in dist.iter_dependencies()}
+        )
+
+    for query_name in query_names:
+        try:
+            dist = installed[query_name]
+        except KeyError:
+            continue
+
+        try:
+            requires = sorted(
+                # Avoid duplicates in requirements (e.g. due to environment markers).
+                {req.name for req in dist.iter_dependencies()},
+                key=str.lower,
+            )
+        except InvalidRequirement:
+            requires = sorted(dist.iter_raw_dependencies(), key=str.lower)
+
+        try:
+            required_by = sorted(_get_requiring_packages(dist), key=str.lower)
+        except InvalidRequirement:
+            required_by = ["#N/A"]
+
+        try:
+            entry_points_text = dist.read_text("entry_points.txt")
+            entry_points = entry_points_text.splitlines(keepends=False)
+        except FileNotFoundError:
+            entry_points = []
+
+        files_iter = dist.iter_declared_entries()
+        if files_iter is None:
+            files: Optional[List[str]] = None
+        else:
+            files = sorted(files_iter)
+
+        metadata = dist.metadata
+
+        project_urls = metadata.get_all("Project-URL", [])
+        homepage = metadata.get("Home-page", "")
+        if not homepage:
+            # It's common that there is a "homepage" Project-URL, but Home-page
+            # remains unset (especially as PEP 621 doesn't surface the field).
+            #
+            # This logic was taken from PyPI's codebase.
+            for url in project_urls:
+                url_label, url = url.split(",", maxsplit=1)
+                normalized_label = (
+                    url_label.casefold().replace("-", "").replace("_", "").strip()
+                )
+                if normalized_label == "homepage":
+                    homepage = url.strip()
+                    break
+
+        yield _PackageInfo(
+            name=dist.raw_name,
+            version=dist.raw_version,
+            location=dist.location or "",
+            editable_project_location=dist.editable_project_location,
+            requires=requires,
+            required_by=required_by,
+            installer=dist.installer,
+            metadata_version=dist.metadata_version or "",
+            classifiers=metadata.get_all("Classifier", []),
+            summary=metadata.get("Summary", ""),
+            homepage=homepage,
+            project_urls=project_urls,
+            author=metadata.get("Author", ""),
+            author_email=metadata.get("Author-email", ""),
+            license=metadata.get("License", ""),
+            license_expression=metadata.get("License-Expression", ""),
+            entry_points=entry_points,
+            files=files,
+        )
+
+
+def print_results(
+    distributions: Iterable[_PackageInfo],
+    list_files: bool,
+    verbose: bool,
+) -> bool:
+    """
+    Print the information from installed distributions found.
+    """
+    results_printed = False
+    for i, dist in enumerate(distributions):
+        results_printed = True
+        if i > 0:
+            write_output("---")
+
+        metadata_version_tuple = tuple(map(int, dist.metadata_version.split(".")))
+
+        write_output("Name: %s", dist.name)
+        write_output("Version: %s", dist.version)
+        write_output("Summary: %s", dist.summary)
+        write_output("Home-page: %s", dist.homepage)
+        write_output("Author: %s", dist.author)
+        write_output("Author-email: %s", dist.author_email)
+        if metadata_version_tuple >= (2, 4) and dist.license_expression:
+            write_output("License-Expression: %s", dist.license_expression)
+        else:
+            write_output("License: %s", dist.license)
+        write_output("Location: %s", dist.location)
+        if dist.editable_project_location is not None:
+            write_output(
+                "Editable project location: %s", dist.editable_project_location
+            )
+        write_output("Requires: %s", ", ".join(dist.requires))
+        write_output("Required-by: %s", ", ".join(dist.required_by))
+
+        if verbose:
+            write_output("Metadata-Version: %s", dist.metadata_version)
+            write_output("Installer: %s", dist.installer)
+            write_output("Classifiers:")
+            for classifier in dist.classifiers:
+                write_output("  %s", classifier)
+            write_output("Entry-points:")
+            for entry in dist.entry_points:
+                write_output("  %s", entry.strip())
+            write_output("Project-URLs:")
+            for project_url in dist.project_urls:
+                write_output("  %s", project_url)
+        if list_files:
+            write_output("Files:")
+            if dist.files is None:
+                write_output("Cannot locate RECORD or installed-files.txt")
+            else:
+                for line in dist.files:
+                    write_output("  %s", line.strip())
+    return results_printed
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/uninstall.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/uninstall.py
new file mode 100644
index 00000000..bc0edeac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/uninstall.py
@@ -0,0 +1,114 @@
+import logging
+from optparse import Values
+from typing import List
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.base_command import Command
+from pip._internal.cli.index_command import SessionCommandMixin
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.exceptions import InstallationError
+from pip._internal.req import parse_requirements
+from pip._internal.req.constructors import (
+    install_req_from_line,
+    install_req_from_parsed_requirement,
+)
+from pip._internal.utils.misc import (
+    check_externally_managed,
+    protect_pip_from_modification_on_windows,
+    warn_if_run_as_root,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class UninstallCommand(Command, SessionCommandMixin):
+    """
+    Uninstall packages.
+
+    pip is able to uninstall most installed packages. Known exceptions are:
+
+    - Pure distutils packages installed with ``python setup.py install``, which
+      leave behind no metadata to determine what files were installed.
+    - Script wrappers installed by ``python setup.py develop``.
+    """
+
+    usage = """
+      %prog [options] <package> ...
+      %prog [options] -r <requirements file> ..."""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-r",
+            "--requirement",
+            dest="requirements",
+            action="append",
+            default=[],
+            metavar="file",
+            help=(
+                "Uninstall all the packages listed in the given requirements "
+                "file.  This option can be used multiple times."
+            ),
+        )
+        self.cmd_opts.add_option(
+            "-y",
+            "--yes",
+            dest="yes",
+            action="store_true",
+            help="Don't ask for confirmation of uninstall deletions.",
+        )
+        self.cmd_opts.add_option(cmdoptions.root_user_action())
+        self.cmd_opts.add_option(cmdoptions.override_externally_managed())
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    def run(self, options: Values, args: List[str]) -> int:
+        session = self.get_default_session(options)
+
+        reqs_to_uninstall = {}
+        for name in args:
+            req = install_req_from_line(
+                name,
+                isolated=options.isolated_mode,
+            )
+            if req.name:
+                reqs_to_uninstall[canonicalize_name(req.name)] = req
+            else:
+                logger.warning(
+                    "Invalid requirement: %r ignored -"
+                    " the uninstall command expects named"
+                    " requirements.",
+                    name,
+                )
+        for filename in options.requirements:
+            for parsed_req in parse_requirements(
+                filename, options=options, session=session
+            ):
+                req = install_req_from_parsed_requirement(
+                    parsed_req, isolated=options.isolated_mode
+                )
+                if req.name:
+                    reqs_to_uninstall[canonicalize_name(req.name)] = req
+        if not reqs_to_uninstall:
+            raise InstallationError(
+                f"You must give at least one requirement to {self.name} (see "
+                f'"pip help {self.name}")'
+            )
+
+        if not options.override_externally_managed:
+            check_externally_managed()
+
+        protect_pip_from_modification_on_windows(
+            modifying_pip="pip" in reqs_to_uninstall
+        )
+
+        for req in reqs_to_uninstall.values():
+            uninstall_pathset = req.uninstall(
+                auto_confirm=options.yes,
+                verbose=self.verbosity > 0,
+            )
+            if uninstall_pathset:
+                uninstall_pathset.commit()
+        if options.root_user_action == "warn":
+            warn_if_run_as_root()
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/wheel.py
new file mode 100644
index 00000000..278719f4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/commands/wheel.py
@@ -0,0 +1,182 @@
+import logging
+import os
+import shutil
+from optparse import Values
+from typing import List
+
+from pip._internal.cache import WheelCache
+from pip._internal.cli import cmdoptions
+from pip._internal.cli.req_command import RequirementCommand, with_cleanup
+from pip._internal.cli.status_codes import SUCCESS
+from pip._internal.exceptions import CommandError
+from pip._internal.operations.build.build_tracker import get_build_tracker
+from pip._internal.req.req_install import (
+    InstallRequirement,
+    check_legacy_setup_py_options,
+)
+from pip._internal.utils.misc import ensure_dir, normalize_path
+from pip._internal.utils.temp_dir import TempDirectory
+from pip._internal.wheel_builder import build, should_build_for_wheel_command
+
+logger = logging.getLogger(__name__)
+
+
+class WheelCommand(RequirementCommand):
+    """
+    Build Wheel archives for your requirements and dependencies.
+
+    Wheel is a built-package format, and offers the advantage of not
+    recompiling your software during every install. For more details, see the
+    wheel docs: https://wheel.readthedocs.io/en/latest/
+
+    'pip wheel' uses the build system interface as described here:
+    https://pip.pypa.io/en/stable/reference/build-system/
+
+    """
+
+    usage = """
+      %prog [options] <requirement specifier> ...
+      %prog [options] -r <requirements file> ...
+      %prog [options] [-e] <vcs project url> ...
+      %prog [options] [-e] <local project path> ...
+      %prog [options] <archive url/path> ..."""
+
+    def add_options(self) -> None:
+        self.cmd_opts.add_option(
+            "-w",
+            "--wheel-dir",
+            dest="wheel_dir",
+            metavar="dir",
+            default=os.curdir,
+            help=(
+                "Build wheels into <dir>, where the default is the "
+                "current working directory."
+            ),
+        )
+        self.cmd_opts.add_option(cmdoptions.no_binary())
+        self.cmd_opts.add_option(cmdoptions.only_binary())
+        self.cmd_opts.add_option(cmdoptions.prefer_binary())
+        self.cmd_opts.add_option(cmdoptions.no_build_isolation())
+        self.cmd_opts.add_option(cmdoptions.use_pep517())
+        self.cmd_opts.add_option(cmdoptions.no_use_pep517())
+        self.cmd_opts.add_option(cmdoptions.check_build_deps())
+        self.cmd_opts.add_option(cmdoptions.constraints())
+        self.cmd_opts.add_option(cmdoptions.editable())
+        self.cmd_opts.add_option(cmdoptions.requirements())
+        self.cmd_opts.add_option(cmdoptions.src())
+        self.cmd_opts.add_option(cmdoptions.ignore_requires_python())
+        self.cmd_opts.add_option(cmdoptions.no_deps())
+        self.cmd_opts.add_option(cmdoptions.progress_bar())
+
+        self.cmd_opts.add_option(
+            "--no-verify",
+            dest="no_verify",
+            action="store_true",
+            default=False,
+            help="Don't verify if built wheel is valid.",
+        )
+
+        self.cmd_opts.add_option(cmdoptions.config_settings())
+        self.cmd_opts.add_option(cmdoptions.build_options())
+        self.cmd_opts.add_option(cmdoptions.global_options())
+
+        self.cmd_opts.add_option(
+            "--pre",
+            action="store_true",
+            default=False,
+            help=(
+                "Include pre-release and development versions. By default, "
+                "pip only finds stable versions."
+            ),
+        )
+
+        self.cmd_opts.add_option(cmdoptions.require_hashes())
+
+        index_opts = cmdoptions.make_option_group(
+            cmdoptions.index_group,
+            self.parser,
+        )
+
+        self.parser.insert_option_group(0, index_opts)
+        self.parser.insert_option_group(0, self.cmd_opts)
+
+    @with_cleanup
+    def run(self, options: Values, args: List[str]) -> int:
+        session = self.get_default_session(options)
+
+        finder = self._build_package_finder(options, session)
+
+        options.wheel_dir = normalize_path(options.wheel_dir)
+        ensure_dir(options.wheel_dir)
+
+        build_tracker = self.enter_context(get_build_tracker())
+
+        directory = TempDirectory(
+            delete=not options.no_clean,
+            kind="wheel",
+            globally_managed=True,
+        )
+
+        reqs = self.get_requirements(args, options, finder, session)
+        check_legacy_setup_py_options(options, reqs)
+
+        wheel_cache = WheelCache(options.cache_dir)
+
+        preparer = self.make_requirement_preparer(
+            temp_build_dir=directory,
+            options=options,
+            build_tracker=build_tracker,
+            session=session,
+            finder=finder,
+            download_dir=options.wheel_dir,
+            use_user_site=False,
+            verbosity=self.verbosity,
+        )
+
+        resolver = self.make_resolver(
+            preparer=preparer,
+            finder=finder,
+            options=options,
+            wheel_cache=wheel_cache,
+            ignore_requires_python=options.ignore_requires_python,
+            use_pep517=options.use_pep517,
+        )
+
+        self.trace_basic_info(finder)
+
+        requirement_set = resolver.resolve(reqs, check_supported_wheels=True)
+
+        reqs_to_build: List[InstallRequirement] = []
+        for req in requirement_set.requirements.values():
+            if req.is_wheel:
+                preparer.save_linked_requirement(req)
+            elif should_build_for_wheel_command(req):
+                reqs_to_build.append(req)
+
+        preparer.prepare_linked_requirements_more(requirement_set.requirements.values())
+
+        # build wheels
+        build_successes, build_failures = build(
+            reqs_to_build,
+            wheel_cache=wheel_cache,
+            verify=(not options.no_verify),
+            build_options=options.build_options or [],
+            global_options=options.global_options or [],
+        )
+        for req in build_successes:
+            assert req.link and req.link.is_wheel
+            assert req.local_file_path
+            # copy from cache to target directory
+            try:
+                shutil.copy(req.local_file_path, options.wheel_dir)
+            except OSError as e:
+                logger.warning(
+                    "Building wheel for %s failed: %s",
+                    req.name,
+                    e,
+                )
+                build_failures.append(req)
+        if len(build_failures) != 0:
+            raise CommandError("Failed to build one or more wheels")
+
+        return SUCCESS
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/configuration.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/configuration.py
new file mode 100644
index 00000000..ffeda1d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/configuration.py
@@ -0,0 +1,383 @@
+"""Configuration management setup
+
+Some terminology:
+- name
+  As written in config files.
+- value
+  Value associated with a name
+- key
+  Name combined with it's section (section.name)
+- variant
+  A single word describing where the configuration key-value pair came from
+"""
+
+import configparser
+import locale
+import os
+import sys
+from typing import Any, Dict, Iterable, List, NewType, Optional, Tuple
+
+from pip._internal.exceptions import (
+    ConfigurationError,
+    ConfigurationFileCouldNotBeLoaded,
+)
+from pip._internal.utils import appdirs
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.logging import getLogger
+from pip._internal.utils.misc import ensure_dir, enum
+
+RawConfigParser = configparser.RawConfigParser  # Shorthand
+Kind = NewType("Kind", str)
+
+CONFIG_BASENAME = "pip.ini" if WINDOWS else "pip.conf"
+ENV_NAMES_IGNORED = "version", "help"
+
+# The kinds of configurations there are.
+kinds = enum(
+    USER="user",  # User Specific
+    GLOBAL="global",  # System Wide
+    SITE="site",  # [Virtual] Environment Specific
+    ENV="env",  # from PIP_CONFIG_FILE
+    ENV_VAR="env-var",  # from Environment Variables
+)
+OVERRIDE_ORDER = kinds.GLOBAL, kinds.USER, kinds.SITE, kinds.ENV, kinds.ENV_VAR
+VALID_LOAD_ONLY = kinds.USER, kinds.GLOBAL, kinds.SITE
+
+logger = getLogger(__name__)
+
+
+# NOTE: Maybe use the optionx attribute to normalize keynames.
+def _normalize_name(name: str) -> str:
+    """Make a name consistent regardless of source (environment or file)"""
+    name = name.lower().replace("_", "-")
+    if name.startswith("--"):
+        name = name[2:]  # only prefer long opts
+    return name
+
+
+def _disassemble_key(name: str) -> List[str]:
+    if "." not in name:
+        error_message = (
+            "Key does not contain dot separated section and key. "
+            f"Perhaps you wanted to use 'global.{name}' instead?"
+        )
+        raise ConfigurationError(error_message)
+    return name.split(".", 1)
+
+
+def get_configuration_files() -> Dict[Kind, List[str]]:
+    global_config_files = [
+        os.path.join(path, CONFIG_BASENAME) for path in appdirs.site_config_dirs("pip")
+    ]
+
+    site_config_file = os.path.join(sys.prefix, CONFIG_BASENAME)
+    legacy_config_file = os.path.join(
+        os.path.expanduser("~"),
+        "pip" if WINDOWS else ".pip",
+        CONFIG_BASENAME,
+    )
+    new_config_file = os.path.join(appdirs.user_config_dir("pip"), CONFIG_BASENAME)
+    return {
+        kinds.GLOBAL: global_config_files,
+        kinds.SITE: [site_config_file],
+        kinds.USER: [legacy_config_file, new_config_file],
+    }
+
+
+class Configuration:
+    """Handles management of configuration.
+
+    Provides an interface to accessing and managing configuration files.
+
+    This class converts provides an API that takes "section.key-name" style
+    keys and stores the value associated with it as "key-name" under the
+    section "section".
+
+    This allows for a clean interface wherein the both the section and the
+    key-name are preserved in an easy to manage form in the configuration files
+    and the data stored is also nice.
+    """
+
+    def __init__(self, isolated: bool, load_only: Optional[Kind] = None) -> None:
+        super().__init__()
+
+        if load_only is not None and load_only not in VALID_LOAD_ONLY:
+            raise ConfigurationError(
+                "Got invalid value for load_only - should be one of {}".format(
+                    ", ".join(map(repr, VALID_LOAD_ONLY))
+                )
+            )
+        self.isolated = isolated
+        self.load_only = load_only
+
+        # Because we keep track of where we got the data from
+        self._parsers: Dict[Kind, List[Tuple[str, RawConfigParser]]] = {
+            variant: [] for variant in OVERRIDE_ORDER
+        }
+        self._config: Dict[Kind, Dict[str, Any]] = {
+            variant: {} for variant in OVERRIDE_ORDER
+        }
+        self._modified_parsers: List[Tuple[str, RawConfigParser]] = []
+
+    def load(self) -> None:
+        """Loads configuration from configuration files and environment"""
+        self._load_config_files()
+        if not self.isolated:
+            self._load_environment_vars()
+
+    def get_file_to_edit(self) -> Optional[str]:
+        """Returns the file with highest priority in configuration"""
+        assert self.load_only is not None, "Need to be specified a file to be editing"
+
+        try:
+            return self._get_parser_to_modify()[0]
+        except IndexError:
+            return None
+
+    def items(self) -> Iterable[Tuple[str, Any]]:
+        """Returns key-value pairs like dict.items() representing the loaded
+        configuration
+        """
+        return self._dictionary.items()
+
+    def get_value(self, key: str) -> Any:
+        """Get a value from the configuration."""
+        orig_key = key
+        key = _normalize_name(key)
+        try:
+            return self._dictionary[key]
+        except KeyError:
+            # disassembling triggers a more useful error message than simply
+            # "No such key" in the case that the key isn't in the form command.option
+            _disassemble_key(key)
+            raise ConfigurationError(f"No such key - {orig_key}")
+
+    def set_value(self, key: str, value: Any) -> None:
+        """Modify a value in the configuration."""
+        key = _normalize_name(key)
+        self._ensure_have_load_only()
+
+        assert self.load_only
+        fname, parser = self._get_parser_to_modify()
+
+        if parser is not None:
+            section, name = _disassemble_key(key)
+
+            # Modify the parser and the configuration
+            if not parser.has_section(section):
+                parser.add_section(section)
+            parser.set(section, name, value)
+
+        self._config[self.load_only][key] = value
+        self._mark_as_modified(fname, parser)
+
+    def unset_value(self, key: str) -> None:
+        """Unset a value in the configuration."""
+        orig_key = key
+        key = _normalize_name(key)
+        self._ensure_have_load_only()
+
+        assert self.load_only
+        if key not in self._config[self.load_only]:
+            raise ConfigurationError(f"No such key - {orig_key}")
+
+        fname, parser = self._get_parser_to_modify()
+
+        if parser is not None:
+            section, name = _disassemble_key(key)
+            if not (
+                parser.has_section(section) and parser.remove_option(section, name)
+            ):
+                # The option was not removed.
+                raise ConfigurationError(
+                    "Fatal Internal error [id=1]. Please report as a bug."
+                )
+
+            # The section may be empty after the option was removed.
+            if not parser.items(section):
+                parser.remove_section(section)
+            self._mark_as_modified(fname, parser)
+
+        del self._config[self.load_only][key]
+
+    def save(self) -> None:
+        """Save the current in-memory state."""
+        self._ensure_have_load_only()
+
+        for fname, parser in self._modified_parsers:
+            logger.info("Writing to %s", fname)
+
+            # Ensure directory exists.
+            ensure_dir(os.path.dirname(fname))
+
+            # Ensure directory's permission(need to be writeable)
+            try:
+                with open(fname, "w") as f:
+                    parser.write(f)
+            except OSError as error:
+                raise ConfigurationError(
+                    f"An error occurred while writing to the configuration file "
+                    f"{fname}: {error}"
+                )
+
+    #
+    # Private routines
+    #
+
+    def _ensure_have_load_only(self) -> None:
+        if self.load_only is None:
+            raise ConfigurationError("Needed a specific file to be modifying.")
+        logger.debug("Will be working with %s variant only", self.load_only)
+
+    @property
+    def _dictionary(self) -> Dict[str, Any]:
+        """A dictionary representing the loaded configuration."""
+        # NOTE: Dictionaries are not populated if not loaded. So, conditionals
+        #       are not needed here.
+        retval = {}
+
+        for variant in OVERRIDE_ORDER:
+            retval.update(self._config[variant])
+
+        return retval
+
+    def _load_config_files(self) -> None:
+        """Loads configuration from configuration files"""
+        config_files = dict(self.iter_config_files())
+        if config_files[kinds.ENV][0:1] == [os.devnull]:
+            logger.debug(
+                "Skipping loading configuration files due to "
+                "environment's PIP_CONFIG_FILE being os.devnull"
+            )
+            return
+
+        for variant, files in config_files.items():
+            for fname in files:
+                # If there's specific variant set in `load_only`, load only
+                # that variant, not the others.
+                if self.load_only is not None and variant != self.load_only:
+                    logger.debug("Skipping file '%s' (variant: %s)", fname, variant)
+                    continue
+
+                parser = self._load_file(variant, fname)
+
+                # Keeping track of the parsers used
+                self._parsers[variant].append((fname, parser))
+
+    def _load_file(self, variant: Kind, fname: str) -> RawConfigParser:
+        logger.verbose("For variant '%s', will try loading '%s'", variant, fname)
+        parser = self._construct_parser(fname)
+
+        for section in parser.sections():
+            items = parser.items(section)
+            self._config[variant].update(self._normalized_keys(section, items))
+
+        return parser
+
+    def _construct_parser(self, fname: str) -> RawConfigParser:
+        parser = configparser.RawConfigParser()
+        # If there is no such file, don't bother reading it but create the
+        # parser anyway, to hold the data.
+        # Doing this is useful when modifying and saving files, where we don't
+        # need to construct a parser.
+        if os.path.exists(fname):
+            locale_encoding = locale.getpreferredencoding(False)
+            try:
+                parser.read(fname, encoding=locale_encoding)
+            except UnicodeDecodeError:
+                # See https://github.com/pypa/pip/issues/4963
+                raise ConfigurationFileCouldNotBeLoaded(
+                    reason=f"contains invalid {locale_encoding} characters",
+                    fname=fname,
+                )
+            except configparser.Error as error:
+                # See https://github.com/pypa/pip/issues/4893
+                raise ConfigurationFileCouldNotBeLoaded(error=error)
+        return parser
+
+    def _load_environment_vars(self) -> None:
+        """Loads configuration from environment variables"""
+        self._config[kinds.ENV_VAR].update(
+            self._normalized_keys(":env:", self.get_environ_vars())
+        )
+
+    def _normalized_keys(
+        self, section: str, items: Iterable[Tuple[str, Any]]
+    ) -> Dict[str, Any]:
+        """Normalizes items to construct a dictionary with normalized keys.
+
+        This routine is where the names become keys and are made the same
+        regardless of source - configuration files or environment.
+        """
+        normalized = {}
+        for name, val in items:
+            key = section + "." + _normalize_name(name)
+            normalized[key] = val
+        return normalized
+
+    def get_environ_vars(self) -> Iterable[Tuple[str, str]]:
+        """Returns a generator with all environmental vars with prefix PIP_"""
+        for key, val in os.environ.items():
+            if key.startswith("PIP_"):
+                name = key[4:].lower()
+                if name not in ENV_NAMES_IGNORED:
+                    yield name, val
+
+    # XXX: This is patched in the tests.
+    def iter_config_files(self) -> Iterable[Tuple[Kind, List[str]]]:
+        """Yields variant and configuration files associated with it.
+
+        This should be treated like items of a dictionary. The order
+        here doesn't affect what gets overridden. That is controlled
+        by OVERRIDE_ORDER. However this does control the order they are
+        displayed to the user. It's probably most ergonomic to display
+        things in the same order as OVERRIDE_ORDER
+        """
+        # SMELL: Move the conditions out of this function
+
+        env_config_file = os.environ.get("PIP_CONFIG_FILE", None)
+        config_files = get_configuration_files()
+
+        yield kinds.GLOBAL, config_files[kinds.GLOBAL]
+
+        # per-user config is not loaded when env_config_file exists
+        should_load_user_config = not self.isolated and not (
+            env_config_file and os.path.exists(env_config_file)
+        )
+        if should_load_user_config:
+            # The legacy config file is overridden by the new config file
+            yield kinds.USER, config_files[kinds.USER]
+
+        # virtualenv config
+        yield kinds.SITE, config_files[kinds.SITE]
+
+        if env_config_file is not None:
+            yield kinds.ENV, [env_config_file]
+        else:
+            yield kinds.ENV, []
+
+    def get_values_in_config(self, variant: Kind) -> Dict[str, Any]:
+        """Get values present in a config file"""
+        return self._config[variant]
+
+    def _get_parser_to_modify(self) -> Tuple[str, RawConfigParser]:
+        # Determine which parser to modify
+        assert self.load_only
+        parsers = self._parsers[self.load_only]
+        if not parsers:
+            # This should not happen if everything works correctly.
+            raise ConfigurationError(
+                "Fatal Internal error [id=2]. Please report as a bug."
+            )
+
+        # Use the highest priority parser.
+        return parsers[-1]
+
+    # XXX: This is patched in the tests.
+    def _mark_as_modified(self, fname: str, parser: RawConfigParser) -> None:
+        file_parser_tuple = (fname, parser)
+        if file_parser_tuple not in self._modified_parsers:
+            self._modified_parsers.append(file_parser_tuple)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self._dictionary!r})"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__init__.py
new file mode 100644
index 00000000..9a89a838
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__init__.py
@@ -0,0 +1,21 @@
+from pip._internal.distributions.base import AbstractDistribution
+from pip._internal.distributions.sdist import SourceDistribution
+from pip._internal.distributions.wheel import WheelDistribution
+from pip._internal.req.req_install import InstallRequirement
+
+
+def make_distribution_for_install_requirement(
+    install_req: InstallRequirement,
+) -> AbstractDistribution:
+    """Returns a Distribution for the given InstallRequirement"""
+    # Editable requirements will always be source distributions. They use the
+    # legacy logic until we create a modern standard for them.
+    if install_req.editable:
+        return SourceDistribution(install_req)
+
+    # If it's a wheel, it's a WheelDistribution
+    if install_req.is_wheel:
+        return WheelDistribution(install_req)
+
+    # Otherwise, a SourceDistribution
+    return SourceDistribution(install_req)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..897a3033
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..2f959562
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/installed.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/installed.cpython-312.pyc
new file mode 100644
index 00000000..a2504604
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/installed.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/sdist.cpython-312.pyc
new file mode 100644
index 00000000..9d451b5b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..38104726
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/base.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/base.py
new file mode 100644
index 00000000..6e4d0c91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/base.py
@@ -0,0 +1,53 @@
+import abc
+from typing import TYPE_CHECKING, Optional
+
+from pip._internal.metadata.base import BaseDistribution
+from pip._internal.req import InstallRequirement
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+
+
+class AbstractDistribution(metaclass=abc.ABCMeta):
+    """A base class for handling installable artifacts.
+
+    The requirements for anything installable are as follows:
+
+     - we must be able to determine the requirement name
+       (or we can't correctly handle the non-upgrade case).
+
+     - for packages with setup requirements, we must also be able
+       to determine their requirements without installing additional
+       packages (for the same reason as run-time dependencies)
+
+     - we must be able to create a Distribution object exposing the
+       above metadata.
+
+     - if we need to do work in the build tracker, we must be able to generate a unique
+       string to identify the requirement in the build tracker.
+    """
+
+    def __init__(self, req: InstallRequirement) -> None:
+        super().__init__()
+        self.req = req
+
+    @abc.abstractproperty
+    def build_tracker_id(self) -> Optional[str]:
+        """A string that uniquely identifies this requirement to the build tracker.
+
+        If None, then this dist has no work to do in the build tracker, and
+        ``.prepare_distribution_metadata()`` will not be called."""
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def get_metadata_distribution(self) -> BaseDistribution:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def prepare_distribution_metadata(
+        self,
+        finder: "PackageFinder",
+        build_isolation: bool,
+        check_build_deps: bool,
+    ) -> None:
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/installed.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/installed.py
new file mode 100644
index 00000000..ab8d53be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/installed.py
@@ -0,0 +1,29 @@
+from typing import Optional
+
+from pip._internal.distributions.base import AbstractDistribution
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.metadata import BaseDistribution
+
+
+class InstalledDistribution(AbstractDistribution):
+    """Represents an installed package.
+
+    This does not need any preparation as the required information has already
+    been computed.
+    """
+
+    @property
+    def build_tracker_id(self) -> Optional[str]:
+        return None
+
+    def get_metadata_distribution(self) -> BaseDistribution:
+        assert self.req.satisfied_by is not None, "not actually installed"
+        return self.req.satisfied_by
+
+    def prepare_distribution_metadata(
+        self,
+        finder: PackageFinder,
+        build_isolation: bool,
+        check_build_deps: bool,
+    ) -> None:
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/sdist.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/sdist.py
new file mode 100644
index 00000000..28ea5cea
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/sdist.py
@@ -0,0 +1,158 @@
+import logging
+from typing import TYPE_CHECKING, Iterable, Optional, Set, Tuple
+
+from pip._internal.build_env import BuildEnvironment
+from pip._internal.distributions.base import AbstractDistribution
+from pip._internal.exceptions import InstallationError
+from pip._internal.metadata import BaseDistribution
+from pip._internal.utils.subprocess import runner_with_spinner_message
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+
+logger = logging.getLogger(__name__)
+
+
+class SourceDistribution(AbstractDistribution):
+    """Represents a source distribution.
+
+    The preparation step for these needs metadata for the packages to be
+    generated, either using PEP 517 or using the legacy `setup.py egg_info`.
+    """
+
+    @property
+    def build_tracker_id(self) -> Optional[str]:
+        """Identify this requirement uniquely by its link."""
+        assert self.req.link
+        return self.req.link.url_without_fragment
+
+    def get_metadata_distribution(self) -> BaseDistribution:
+        return self.req.get_dist()
+
+    def prepare_distribution_metadata(
+        self,
+        finder: "PackageFinder",
+        build_isolation: bool,
+        check_build_deps: bool,
+    ) -> None:
+        # Load pyproject.toml, to determine whether PEP 517 is to be used
+        self.req.load_pyproject_toml()
+
+        # Set up the build isolation, if this requirement should be isolated
+        should_isolate = self.req.use_pep517 and build_isolation
+        if should_isolate:
+            # Setup an isolated environment and install the build backend static
+            # requirements in it.
+            self._prepare_build_backend(finder)
+            # Check that if the requirement is editable, it either supports PEP 660 or
+            # has a setup.py or a setup.cfg. This cannot be done earlier because we need
+            # to setup the build backend to verify it supports build_editable, nor can
+            # it be done later, because we want to avoid installing build requirements
+            # needlessly. Doing it here also works around setuptools generating
+            # UNKNOWN.egg-info when running get_requires_for_build_wheel on a directory
+            # without setup.py nor setup.cfg.
+            self.req.isolated_editable_sanity_check()
+            # Install the dynamic build requirements.
+            self._install_build_reqs(finder)
+        # Check if the current environment provides build dependencies
+        should_check_deps = self.req.use_pep517 and check_build_deps
+        if should_check_deps:
+            pyproject_requires = self.req.pyproject_requires
+            assert pyproject_requires is not None
+            conflicting, missing = self.req.build_env.check_requirements(
+                pyproject_requires
+            )
+            if conflicting:
+                self._raise_conflicts("the backend dependencies", conflicting)
+            if missing:
+                self._raise_missing_reqs(missing)
+        self.req.prepare_metadata()
+
+    def _prepare_build_backend(self, finder: "PackageFinder") -> None:
+        # Isolate in a BuildEnvironment and install the build-time
+        # requirements.
+        pyproject_requires = self.req.pyproject_requires
+        assert pyproject_requires is not None
+
+        self.req.build_env = BuildEnvironment()
+        self.req.build_env.install_requirements(
+            finder, pyproject_requires, "overlay", kind="build dependencies"
+        )
+        conflicting, missing = self.req.build_env.check_requirements(
+            self.req.requirements_to_check
+        )
+        if conflicting:
+            self._raise_conflicts("PEP 517/518 supported requirements", conflicting)
+        if missing:
+            logger.warning(
+                "Missing build requirements in pyproject.toml for %s.",
+                self.req,
+            )
+            logger.warning(
+                "The project does not specify a build backend, and "
+                "pip cannot fall back to setuptools without %s.",
+                " and ".join(map(repr, sorted(missing))),
+            )
+
+    def _get_build_requires_wheel(self) -> Iterable[str]:
+        with self.req.build_env:
+            runner = runner_with_spinner_message("Getting requirements to build wheel")
+            backend = self.req.pep517_backend
+            assert backend is not None
+            with backend.subprocess_runner(runner):
+                return backend.get_requires_for_build_wheel()
+
+    def _get_build_requires_editable(self) -> Iterable[str]:
+        with self.req.build_env:
+            runner = runner_with_spinner_message(
+                "Getting requirements to build editable"
+            )
+            backend = self.req.pep517_backend
+            assert backend is not None
+            with backend.subprocess_runner(runner):
+                return backend.get_requires_for_build_editable()
+
+    def _install_build_reqs(self, finder: "PackageFinder") -> None:
+        # Install any extra build dependencies that the backend requests.
+        # This must be done in a second pass, as the pyproject.toml
+        # dependencies must be installed before we can call the backend.
+        if (
+            self.req.editable
+            and self.req.permit_editable_wheels
+            and self.req.supports_pyproject_editable
+        ):
+            build_reqs = self._get_build_requires_editable()
+        else:
+            build_reqs = self._get_build_requires_wheel()
+        conflicting, missing = self.req.build_env.check_requirements(build_reqs)
+        if conflicting:
+            self._raise_conflicts("the backend dependencies", conflicting)
+        self.req.build_env.install_requirements(
+            finder, missing, "normal", kind="backend dependencies"
+        )
+
+    def _raise_conflicts(
+        self, conflicting_with: str, conflicting_reqs: Set[Tuple[str, str]]
+    ) -> None:
+        format_string = (
+            "Some build dependencies for {requirement} "
+            "conflict with {conflicting_with}: {description}."
+        )
+        error_message = format_string.format(
+            requirement=self.req,
+            conflicting_with=conflicting_with,
+            description=", ".join(
+                f"{installed} is incompatible with {wanted}"
+                for installed, wanted in sorted(conflicting_reqs)
+            ),
+        )
+        raise InstallationError(error_message)
+
+    def _raise_missing_reqs(self, missing: Set[str]) -> None:
+        format_string = (
+            "Some build dependencies for {requirement} are missing: {missing}."
+        )
+        error_message = format_string.format(
+            requirement=self.req, missing=", ".join(map(repr, sorted(missing)))
+        )
+        raise InstallationError(error_message)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/wheel.py
new file mode 100644
index 00000000..bfadd39d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/distributions/wheel.py
@@ -0,0 +1,42 @@
+from typing import TYPE_CHECKING, Optional
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.distributions.base import AbstractDistribution
+from pip._internal.metadata import (
+    BaseDistribution,
+    FilesystemWheel,
+    get_wheel_distribution,
+)
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+
+
+class WheelDistribution(AbstractDistribution):
+    """Represents a wheel distribution.
+
+    This does not need any preparation as wheels can be directly unpacked.
+    """
+
+    @property
+    def build_tracker_id(self) -> Optional[str]:
+        return None
+
+    def get_metadata_distribution(self) -> BaseDistribution:
+        """Loads the metadata from the wheel file into memory and returns a
+        Distribution that uses it, not relying on the wheel file or
+        requirement.
+        """
+        assert self.req.local_file_path, "Set as part of preparation during download"
+        assert self.req.name, "Wheels are never unnamed"
+        wheel = FilesystemWheel(self.req.local_file_path)
+        return get_wheel_distribution(wheel, canonicalize_name(self.req.name))
+
+    def prepare_distribution_metadata(
+        self,
+        finder: "PackageFinder",
+        build_isolation: bool,
+        check_build_deps: bool,
+    ) -> None:
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/exceptions.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/exceptions.py
new file mode 100644
index 00000000..45a876a8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/exceptions.py
@@ -0,0 +1,809 @@
+"""Exceptions used throughout package.
+
+This module MUST NOT try to import from anything within `pip._internal` to
+operate. This is expected to be importable from any/all files within the
+subpackage and, thus, should not depend on them.
+"""
+
+import configparser
+import contextlib
+import locale
+import logging
+import pathlib
+import re
+import sys
+from itertools import chain, groupby, repeat
+from typing import TYPE_CHECKING, Dict, Iterator, List, Literal, Optional, Union
+
+from pip._vendor.packaging.requirements import InvalidRequirement
+from pip._vendor.packaging.version import InvalidVersion
+from pip._vendor.rich.console import Console, ConsoleOptions, RenderResult
+from pip._vendor.rich.markup import escape
+from pip._vendor.rich.text import Text
+
+if TYPE_CHECKING:
+    from hashlib import _Hash
+
+    from pip._vendor.requests.models import Request, Response
+
+    from pip._internal.metadata import BaseDistribution
+    from pip._internal.req.req_install import InstallRequirement
+
+logger = logging.getLogger(__name__)
+
+
+#
+# Scaffolding
+#
+def _is_kebab_case(s: str) -> bool:
+    return re.match(r"^[a-z]+(-[a-z]+)*$", s) is not None
+
+
+def _prefix_with_indent(
+    s: Union[Text, str],
+    console: Console,
+    *,
+    prefix: str,
+    indent: str,
+) -> Text:
+    if isinstance(s, Text):
+        text = s
+    else:
+        text = console.render_str(s)
+
+    return console.render_str(prefix, overflow="ignore") + console.render_str(
+        f"\n{indent}", overflow="ignore"
+    ).join(text.split(allow_blank=True))
+
+
+class PipError(Exception):
+    """The base pip error."""
+
+
+class DiagnosticPipError(PipError):
+    """An error, that presents diagnostic information to the user.
+
+    This contains a bunch of logic, to enable pretty presentation of our error
+    messages. Each error gets a unique reference. Each error can also include
+    additional context, a hint and/or a note -- which are presented with the
+    main error message in a consistent style.
+
+    This is adapted from the error output styling in `sphinx-theme-builder`.
+    """
+
+    reference: str
+
+    def __init__(
+        self,
+        *,
+        kind: 'Literal["error", "warning"]' = "error",
+        reference: Optional[str] = None,
+        message: Union[str, Text],
+        context: Optional[Union[str, Text]],
+        hint_stmt: Optional[Union[str, Text]],
+        note_stmt: Optional[Union[str, Text]] = None,
+        link: Optional[str] = None,
+    ) -> None:
+        # Ensure a proper reference is provided.
+        if reference is None:
+            assert hasattr(self, "reference"), "error reference not provided!"
+            reference = self.reference
+        assert _is_kebab_case(reference), "error reference must be kebab-case!"
+
+        self.kind = kind
+        self.reference = reference
+
+        self.message = message
+        self.context = context
+
+        self.note_stmt = note_stmt
+        self.hint_stmt = hint_stmt
+
+        self.link = link
+
+        super().__init__(f"<{self.__class__.__name__}: {self.reference}>")
+
+    def __repr__(self) -> str:
+        return (
+            f"<{self.__class__.__name__}("
+            f"reference={self.reference!r}, "
+            f"message={self.message!r}, "
+            f"context={self.context!r}, "
+            f"note_stmt={self.note_stmt!r}, "
+            f"hint_stmt={self.hint_stmt!r}"
+            ")>"
+        )
+
+    def __rich_console__(
+        self,
+        console: Console,
+        options: ConsoleOptions,
+    ) -> RenderResult:
+        colour = "red" if self.kind == "error" else "yellow"
+
+        yield f"[{colour} bold]{self.kind}[/]: [bold]{self.reference}[/]"
+        yield ""
+
+        if not options.ascii_only:
+            # Present the main message, with relevant context indented.
+            if self.context is not None:
+                yield _prefix_with_indent(
+                    self.message,
+                    console,
+                    prefix=f"[{colour}]×[/] ",
+                    indent=f"[{colour}]│[/] ",
+                )
+                yield _prefix_with_indent(
+                    self.context,
+                    console,
+                    prefix=f"[{colour}]╰─>[/] ",
+                    indent=f"[{colour}]   [/] ",
+                )
+            else:
+                yield _prefix_with_indent(
+                    self.message,
+                    console,
+                    prefix="[red]×[/] ",
+                    indent="  ",
+                )
+        else:
+            yield self.message
+            if self.context is not None:
+                yield ""
+                yield self.context
+
+        if self.note_stmt is not None or self.hint_stmt is not None:
+            yield ""
+
+        if self.note_stmt is not None:
+            yield _prefix_with_indent(
+                self.note_stmt,
+                console,
+                prefix="[magenta bold]note[/]: ",
+                indent="      ",
+            )
+        if self.hint_stmt is not None:
+            yield _prefix_with_indent(
+                self.hint_stmt,
+                console,
+                prefix="[cyan bold]hint[/]: ",
+                indent="      ",
+            )
+
+        if self.link is not None:
+            yield ""
+            yield f"Link: {self.link}"
+
+
+#
+# Actual Errors
+#
+class ConfigurationError(PipError):
+    """General exception in configuration"""
+
+
+class InstallationError(PipError):
+    """General exception during installation"""
+
+
+class MissingPyProjectBuildRequires(DiagnosticPipError):
+    """Raised when pyproject.toml has `build-system`, but no `build-system.requires`."""
+
+    reference = "missing-pyproject-build-system-requires"
+
+    def __init__(self, *, package: str) -> None:
+        super().__init__(
+            message=f"Can not process {escape(package)}",
+            context=Text(
+                "This package has an invalid pyproject.toml file.\n"
+                "The [build-system] table is missing the mandatory `requires` key."
+            ),
+            note_stmt="This is an issue with the package mentioned above, not pip.",
+            hint_stmt=Text("See PEP 518 for the detailed specification."),
+        )
+
+
+class InvalidPyProjectBuildRequires(DiagnosticPipError):
+    """Raised when pyproject.toml an invalid `build-system.requires`."""
+
+    reference = "invalid-pyproject-build-system-requires"
+
+    def __init__(self, *, package: str, reason: str) -> None:
+        super().__init__(
+            message=f"Can not process {escape(package)}",
+            context=Text(
+                "This package has an invalid `build-system.requires` key in "
+                f"pyproject.toml.\n{reason}"
+            ),
+            note_stmt="This is an issue with the package mentioned above, not pip.",
+            hint_stmt=Text("See PEP 518 for the detailed specification."),
+        )
+
+
+class NoneMetadataError(PipError):
+    """Raised when accessing a Distribution's "METADATA" or "PKG-INFO".
+
+    This signifies an inconsistency, when the Distribution claims to have
+    the metadata file (if not, raise ``FileNotFoundError`` instead), but is
+    not actually able to produce its content. This may be due to permission
+    errors.
+    """
+
+    def __init__(
+        self,
+        dist: "BaseDistribution",
+        metadata_name: str,
+    ) -> None:
+        """
+        :param dist: A Distribution object.
+        :param metadata_name: The name of the metadata being accessed
+            (can be "METADATA" or "PKG-INFO").
+        """
+        self.dist = dist
+        self.metadata_name = metadata_name
+
+    def __str__(self) -> str:
+        # Use `dist` in the error message because its stringification
+        # includes more information, like the version and location.
+        return f"None {self.metadata_name} metadata found for distribution: {self.dist}"
+
+
+class UserInstallationInvalid(InstallationError):
+    """A --user install is requested on an environment without user site."""
+
+    def __str__(self) -> str:
+        return "User base directory is not specified"
+
+
+class InvalidSchemeCombination(InstallationError):
+    def __str__(self) -> str:
+        before = ", ".join(str(a) for a in self.args[:-1])
+        return f"Cannot set {before} and {self.args[-1]} together"
+
+
+class DistributionNotFound(InstallationError):
+    """Raised when a distribution cannot be found to satisfy a requirement"""
+
+
+class RequirementsFileParseError(InstallationError):
+    """Raised when a general error occurs parsing a requirements file line."""
+
+
+class BestVersionAlreadyInstalled(PipError):
+    """Raised when the most up-to-date version of a package is already
+    installed."""
+
+
+class BadCommand(PipError):
+    """Raised when virtualenv or a command is not found"""
+
+
+class CommandError(PipError):
+    """Raised when there is an error in command-line arguments"""
+
+
+class PreviousBuildDirError(PipError):
+    """Raised when there's a previous conflicting build directory"""
+
+
+class NetworkConnectionError(PipError):
+    """HTTP connection error"""
+
+    def __init__(
+        self,
+        error_msg: str,
+        response: Optional["Response"] = None,
+        request: Optional["Request"] = None,
+    ) -> None:
+        """
+        Initialize NetworkConnectionError with  `request` and `response`
+        objects.
+        """
+        self.response = response
+        self.request = request
+        self.error_msg = error_msg
+        if (
+            self.response is not None
+            and not self.request
+            and hasattr(response, "request")
+        ):
+            self.request = self.response.request
+        super().__init__(error_msg, response, request)
+
+    def __str__(self) -> str:
+        return str(self.error_msg)
+
+
+class InvalidWheelFilename(InstallationError):
+    """Invalid wheel filename."""
+
+
+class UnsupportedWheel(InstallationError):
+    """Unsupported wheel."""
+
+
+class InvalidWheel(InstallationError):
+    """Invalid (e.g. corrupt) wheel."""
+
+    def __init__(self, location: str, name: str):
+        self.location = location
+        self.name = name
+
+    def __str__(self) -> str:
+        return f"Wheel '{self.name}' located at {self.location} is invalid."
+
+
+class MetadataInconsistent(InstallationError):
+    """Built metadata contains inconsistent information.
+
+    This is raised when the metadata contains values (e.g. name and version)
+    that do not match the information previously obtained from sdist filename,
+    user-supplied ``#egg=`` value, or an install requirement name.
+    """
+
+    def __init__(
+        self, ireq: "InstallRequirement", field: str, f_val: str, m_val: str
+    ) -> None:
+        self.ireq = ireq
+        self.field = field
+        self.f_val = f_val
+        self.m_val = m_val
+
+    def __str__(self) -> str:
+        return (
+            f"Requested {self.ireq} has inconsistent {self.field}: "
+            f"expected {self.f_val!r}, but metadata has {self.m_val!r}"
+        )
+
+
+class MetadataInvalid(InstallationError):
+    """Metadata is invalid."""
+
+    def __init__(self, ireq: "InstallRequirement", error: str) -> None:
+        self.ireq = ireq
+        self.error = error
+
+    def __str__(self) -> str:
+        return f"Requested {self.ireq} has invalid metadata: {self.error}"
+
+
+class InstallationSubprocessError(DiagnosticPipError, InstallationError):
+    """A subprocess call failed."""
+
+    reference = "subprocess-exited-with-error"
+
+    def __init__(
+        self,
+        *,
+        command_description: str,
+        exit_code: int,
+        output_lines: Optional[List[str]],
+    ) -> None:
+        if output_lines is None:
+            output_prompt = Text("See above for output.")
+        else:
+            output_prompt = (
+                Text.from_markup(f"[red][{len(output_lines)} lines of output][/]\n")
+                + Text("".join(output_lines))
+                + Text.from_markup(R"[red]\[end of output][/]")
+            )
+
+        super().__init__(
+            message=(
+                f"[green]{escape(command_description)}[/] did not run successfully.\n"
+                f"exit code: {exit_code}"
+            ),
+            context=output_prompt,
+            hint_stmt=None,
+            note_stmt=(
+                "This error originates from a subprocess, and is likely not a "
+                "problem with pip."
+            ),
+        )
+
+        self.command_description = command_description
+        self.exit_code = exit_code
+
+    def __str__(self) -> str:
+        return f"{self.command_description} exited with {self.exit_code}"
+
+
+class MetadataGenerationFailed(InstallationSubprocessError, InstallationError):
+    reference = "metadata-generation-failed"
+
+    def __init__(
+        self,
+        *,
+        package_details: str,
+    ) -> None:
+        super(InstallationSubprocessError, self).__init__(
+            message="Encountered error while generating package metadata.",
+            context=escape(package_details),
+            hint_stmt="See above for details.",
+            note_stmt="This is an issue with the package mentioned above, not pip.",
+        )
+
+    def __str__(self) -> str:
+        return "metadata generation failed"
+
+
+class HashErrors(InstallationError):
+    """Multiple HashError instances rolled into one for reporting"""
+
+    def __init__(self) -> None:
+        self.errors: List[HashError] = []
+
+    def append(self, error: "HashError") -> None:
+        self.errors.append(error)
+
+    def __str__(self) -> str:
+        lines = []
+        self.errors.sort(key=lambda e: e.order)
+        for cls, errors_of_cls in groupby(self.errors, lambda e: e.__class__):
+            lines.append(cls.head)
+            lines.extend(e.body() for e in errors_of_cls)
+        if lines:
+            return "\n".join(lines)
+        return ""
+
+    def __bool__(self) -> bool:
+        return bool(self.errors)
+
+
+class HashError(InstallationError):
+    """
+    A failure to verify a package against known-good hashes
+
+    :cvar order: An int sorting hash exception classes by difficulty of
+        recovery (lower being harder), so the user doesn't bother fretting
+        about unpinned packages when he has deeper issues, like VCS
+        dependencies, to deal with. Also keeps error reports in a
+        deterministic order.
+    :cvar head: A section heading for display above potentially many
+        exceptions of this kind
+    :ivar req: The InstallRequirement that triggered this error. This is
+        pasted on after the exception is instantiated, because it's not
+        typically available earlier.
+
+    """
+
+    req: Optional["InstallRequirement"] = None
+    head = ""
+    order: int = -1
+
+    def body(self) -> str:
+        """Return a summary of me for display under the heading.
+
+        This default implementation simply prints a description of the
+        triggering requirement.
+
+        :param req: The InstallRequirement that provoked this error, with
+            its link already populated by the resolver's _populate_link().
+
+        """
+        return f"    {self._requirement_name()}"
+
+    def __str__(self) -> str:
+        return f"{self.head}\n{self.body()}"
+
+    def _requirement_name(self) -> str:
+        """Return a description of the requirement that triggered me.
+
+        This default implementation returns long description of the req, with
+        line numbers
+
+        """
+        return str(self.req) if self.req else "unknown package"
+
+
+class VcsHashUnsupported(HashError):
+    """A hash was provided for a version-control-system-based requirement, but
+    we don't have a method for hashing those."""
+
+    order = 0
+    head = (
+        "Can't verify hashes for these requirements because we don't "
+        "have a way to hash version control repositories:"
+    )
+
+
+class DirectoryUrlHashUnsupported(HashError):
+    """A hash was provided for a version-control-system-based requirement, but
+    we don't have a method for hashing those."""
+
+    order = 1
+    head = (
+        "Can't verify hashes for these file:// requirements because they "
+        "point to directories:"
+    )
+
+
+class HashMissing(HashError):
+    """A hash was needed for a requirement but is absent."""
+
+    order = 2
+    head = (
+        "Hashes are required in --require-hashes mode, but they are "
+        "missing from some requirements. Here is a list of those "
+        "requirements along with the hashes their downloaded archives "
+        "actually had. Add lines like these to your requirements files to "
+        "prevent tampering. (If you did not enable --require-hashes "
+        "manually, note that it turns on automatically when any package "
+        "has a hash.)"
+    )
+
+    def __init__(self, gotten_hash: str) -> None:
+        """
+        :param gotten_hash: The hash of the (possibly malicious) archive we
+            just downloaded
+        """
+        self.gotten_hash = gotten_hash
+
+    def body(self) -> str:
+        # Dodge circular import.
+        from pip._internal.utils.hashes import FAVORITE_HASH
+
+        package = None
+        if self.req:
+            # In the case of URL-based requirements, display the original URL
+            # seen in the requirements file rather than the package name,
+            # so the output can be directly copied into the requirements file.
+            package = (
+                self.req.original_link
+                if self.req.is_direct
+                # In case someone feeds something downright stupid
+                # to InstallRequirement's constructor.
+                else getattr(self.req, "req", None)
+            )
+        return "    {} --hash={}:{}".format(
+            package or "unknown package", FAVORITE_HASH, self.gotten_hash
+        )
+
+
+class HashUnpinned(HashError):
+    """A requirement had a hash specified but was not pinned to a specific
+    version."""
+
+    order = 3
+    head = (
+        "In --require-hashes mode, all requirements must have their "
+        "versions pinned with ==. These do not:"
+    )
+
+
+class HashMismatch(HashError):
+    """
+    Distribution file hash values don't match.
+
+    :ivar package_name: The name of the package that triggered the hash
+        mismatch. Feel free to write to this after the exception is raise to
+        improve its error message.
+
+    """
+
+    order = 4
+    head = (
+        "THESE PACKAGES DO NOT MATCH THE HASHES FROM THE REQUIREMENTS "
+        "FILE. If you have updated the package versions, please update "
+        "the hashes. Otherwise, examine the package contents carefully; "
+        "someone may have tampered with them."
+    )
+
+    def __init__(self, allowed: Dict[str, List[str]], gots: Dict[str, "_Hash"]) -> None:
+        """
+        :param allowed: A dict of algorithm names pointing to lists of allowed
+            hex digests
+        :param gots: A dict of algorithm names pointing to hashes we
+            actually got from the files under suspicion
+        """
+        self.allowed = allowed
+        self.gots = gots
+
+    def body(self) -> str:
+        return f"    {self._requirement_name()}:\n{self._hash_comparison()}"
+
+    def _hash_comparison(self) -> str:
+        """
+        Return a comparison of actual and expected hash values.
+
+        Example::
+
+               Expected sha256 abcdeabcdeabcdeabcdeabcdeabcdeabcdeabcdeabcde
+                            or 123451234512345123451234512345123451234512345
+                    Got        bcdefbcdefbcdefbcdefbcdefbcdefbcdefbcdefbcdef
+
+        """
+
+        def hash_then_or(hash_name: str) -> "chain[str]":
+            # For now, all the decent hashes have 6-char names, so we can get
+            # away with hard-coding space literals.
+            return chain([hash_name], repeat("    or"))
+
+        lines: List[str] = []
+        for hash_name, expecteds in self.allowed.items():
+            prefix = hash_then_or(hash_name)
+            lines.extend((f"        Expected {next(prefix)} {e}") for e in expecteds)
+            lines.append(
+                f"             Got        {self.gots[hash_name].hexdigest()}\n"
+            )
+        return "\n".join(lines)
+
+
+class UnsupportedPythonVersion(InstallationError):
+    """Unsupported python version according to Requires-Python package
+    metadata."""
+
+
+class ConfigurationFileCouldNotBeLoaded(ConfigurationError):
+    """When there are errors while loading a configuration file"""
+
+    def __init__(
+        self,
+        reason: str = "could not be loaded",
+        fname: Optional[str] = None,
+        error: Optional[configparser.Error] = None,
+    ) -> None:
+        super().__init__(error)
+        self.reason = reason
+        self.fname = fname
+        self.error = error
+
+    def __str__(self) -> str:
+        if self.fname is not None:
+            message_part = f" in {self.fname}."
+        else:
+            assert self.error is not None
+            message_part = f".\n{self.error}\n"
+        return f"Configuration file {self.reason}{message_part}"
+
+
+_DEFAULT_EXTERNALLY_MANAGED_ERROR = f"""\
+The Python environment under {sys.prefix} is managed externally, and may not be
+manipulated by the user. Please use specific tooling from the distributor of
+the Python installation to interact with this environment instead.
+"""
+
+
+class ExternallyManagedEnvironment(DiagnosticPipError):
+    """The current environment is externally managed.
+
+    This is raised when the current environment is externally managed, as
+    defined by `PEP 668`_. The ``EXTERNALLY-MANAGED`` configuration is checked
+    and displayed when the error is bubbled up to the user.
+
+    :param error: The error message read from ``EXTERNALLY-MANAGED``.
+    """
+
+    reference = "externally-managed-environment"
+
+    def __init__(self, error: Optional[str]) -> None:
+        if error is None:
+            context = Text(_DEFAULT_EXTERNALLY_MANAGED_ERROR)
+        else:
+            context = Text(error)
+        super().__init__(
+            message="This environment is externally managed",
+            context=context,
+            note_stmt=(
+                "If you believe this is a mistake, please contact your "
+                "Python installation or OS distribution provider. "
+                "You can override this, at the risk of breaking your Python "
+                "installation or OS, by passing --break-system-packages."
+            ),
+            hint_stmt=Text("See PEP 668 for the detailed specification."),
+        )
+
+    @staticmethod
+    def _iter_externally_managed_error_keys() -> Iterator[str]:
+        # LC_MESSAGES is in POSIX, but not the C standard. The most common
+        # platform that does not implement this category is Windows, where
+        # using other categories for console message localization is equally
+        # unreliable, so we fall back to the locale-less vendor message. This
+        # can always be re-evaluated when a vendor proposes a new alternative.
+        try:
+            category = locale.LC_MESSAGES
+        except AttributeError:
+            lang: Optional[str] = None
+        else:
+            lang, _ = locale.getlocale(category)
+        if lang is not None:
+            yield f"Error-{lang}"
+            for sep in ("-", "_"):
+                before, found, _ = lang.partition(sep)
+                if not found:
+                    continue
+                yield f"Error-{before}"
+        yield "Error"
+
+    @classmethod
+    def from_config(
+        cls,
+        config: Union[pathlib.Path, str],
+    ) -> "ExternallyManagedEnvironment":
+        parser = configparser.ConfigParser(interpolation=None)
+        try:
+            parser.read(config, encoding="utf-8")
+            section = parser["externally-managed"]
+            for key in cls._iter_externally_managed_error_keys():
+                with contextlib.suppress(KeyError):
+                    return cls(section[key])
+        except KeyError:
+            pass
+        except (OSError, UnicodeDecodeError, configparser.ParsingError):
+            from pip._internal.utils._log import VERBOSE
+
+            exc_info = logger.isEnabledFor(VERBOSE)
+            logger.warning("Failed to read %s", config, exc_info=exc_info)
+        return cls(None)
+
+
+class UninstallMissingRecord(DiagnosticPipError):
+    reference = "uninstall-no-record-file"
+
+    def __init__(self, *, distribution: "BaseDistribution") -> None:
+        installer = distribution.installer
+        if not installer or installer == "pip":
+            dep = f"{distribution.raw_name}=={distribution.version}"
+            hint = Text.assemble(
+                "You might be able to recover from this via: ",
+                (f"pip install --force-reinstall --no-deps {dep}", "green"),
+            )
+        else:
+            hint = Text(
+                f"The package was installed by {installer}. "
+                "You should check if it can uninstall the package."
+            )
+
+        super().__init__(
+            message=Text(f"Cannot uninstall {distribution}"),
+            context=(
+                "The package's contents are unknown: "
+                f"no RECORD file was found for {distribution.raw_name}."
+            ),
+            hint_stmt=hint,
+        )
+
+
+class LegacyDistutilsInstall(DiagnosticPipError):
+    reference = "uninstall-distutils-installed-package"
+
+    def __init__(self, *, distribution: "BaseDistribution") -> None:
+        super().__init__(
+            message=Text(f"Cannot uninstall {distribution}"),
+            context=(
+                "It is a distutils installed project and thus we cannot accurately "
+                "determine which files belong to it which would lead to only a partial "
+                "uninstall."
+            ),
+            hint_stmt=None,
+        )
+
+
+class InvalidInstalledPackage(DiagnosticPipError):
+    reference = "invalid-installed-package"
+
+    def __init__(
+        self,
+        *,
+        dist: "BaseDistribution",
+        invalid_exc: Union[InvalidRequirement, InvalidVersion],
+    ) -> None:
+        installed_location = dist.installed_location
+
+        if isinstance(invalid_exc, InvalidRequirement):
+            invalid_type = "requirement"
+        else:
+            invalid_type = "version"
+
+        super().__init__(
+            message=Text(
+                f"Cannot process installed package {dist} "
+                + (f"in {installed_location!r} " if installed_location else "")
+                + f"because it has an invalid {invalid_type}:\n{invalid_exc.args[0]}"
+            ),
+            context=(
+                "Starting with pip 24.1, packages with invalid "
+                f"{invalid_type}s can not be processed."
+            ),
+            hint_stmt="To proceed this package must be uninstalled.",
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__init__.py
new file mode 100644
index 00000000..7a17b7b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__init__.py
@@ -0,0 +1,2 @@
+"""Index interaction code
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7e75c301
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/collector.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/collector.cpython-312.pyc
new file mode 100644
index 00000000..dc590006
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/collector.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/package_finder.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/package_finder.cpython-312.pyc
new file mode 100644
index 00000000..f2b759c6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/package_finder.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/sources.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/sources.cpython-312.pyc
new file mode 100644
index 00000000..e1852bec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/__pycache__/sources.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/collector.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/collector.py
new file mode 100644
index 00000000..5f8fdee3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/collector.py
@@ -0,0 +1,494 @@
+"""
+The main purpose of this module is to expose LinkCollector.collect_sources().
+"""
+
+import collections
+import email.message
+import functools
+import itertools
+import json
+import logging
+import os
+import urllib.parse
+import urllib.request
+from dataclasses import dataclass
+from html.parser import HTMLParser
+from optparse import Values
+from typing import (
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    MutableMapping,
+    NamedTuple,
+    Optional,
+    Protocol,
+    Sequence,
+    Tuple,
+    Union,
+)
+
+from pip._vendor import requests
+from pip._vendor.requests import Response
+from pip._vendor.requests.exceptions import RetryError, SSLError
+
+from pip._internal.exceptions import NetworkConnectionError
+from pip._internal.models.link import Link
+from pip._internal.models.search_scope import SearchScope
+from pip._internal.network.session import PipSession
+from pip._internal.network.utils import raise_for_status
+from pip._internal.utils.filetypes import is_archive_file
+from pip._internal.utils.misc import redact_auth_from_url
+from pip._internal.vcs import vcs
+
+from .sources import CandidatesFromPage, LinkSource, build_source
+
+logger = logging.getLogger(__name__)
+
+ResponseHeaders = MutableMapping[str, str]
+
+
+def _match_vcs_scheme(url: str) -> Optional[str]:
+    """Look for VCS schemes in the URL.
+
+    Returns the matched VCS scheme, or None if there's no match.
+    """
+    for scheme in vcs.schemes:
+        if url.lower().startswith(scheme) and url[len(scheme)] in "+:":
+            return scheme
+    return None
+
+
+class _NotAPIContent(Exception):
+    def __init__(self, content_type: str, request_desc: str) -> None:
+        super().__init__(content_type, request_desc)
+        self.content_type = content_type
+        self.request_desc = request_desc
+
+
+def _ensure_api_header(response: Response) -> None:
+    """
+    Check the Content-Type header to ensure the response contains a Simple
+    API Response.
+
+    Raises `_NotAPIContent` if the content type is not a valid content-type.
+    """
+    content_type = response.headers.get("Content-Type", "Unknown")
+
+    content_type_l = content_type.lower()
+    if content_type_l.startswith(
+        (
+            "text/html",
+            "application/vnd.pypi.simple.v1+html",
+            "application/vnd.pypi.simple.v1+json",
+        )
+    ):
+        return
+
+    raise _NotAPIContent(content_type, response.request.method)
+
+
+class _NotHTTP(Exception):
+    pass
+
+
+def _ensure_api_response(url: str, session: PipSession) -> None:
+    """
+    Send a HEAD request to the URL, and ensure the response contains a simple
+    API Response.
+
+    Raises `_NotHTTP` if the URL is not available for a HEAD request, or
+    `_NotAPIContent` if the content type is not a valid content type.
+    """
+    scheme, netloc, path, query, fragment = urllib.parse.urlsplit(url)
+    if scheme not in {"http", "https"}:
+        raise _NotHTTP()
+
+    resp = session.head(url, allow_redirects=True)
+    raise_for_status(resp)
+
+    _ensure_api_header(resp)
+
+
+def _get_simple_response(url: str, session: PipSession) -> Response:
+    """Access an Simple API response with GET, and return the response.
+
+    This consists of three parts:
+
+    1. If the URL looks suspiciously like an archive, send a HEAD first to
+       check the Content-Type is HTML or Simple API, to avoid downloading a
+       large file. Raise `_NotHTTP` if the content type cannot be determined, or
+       `_NotAPIContent` if it is not HTML or a Simple API.
+    2. Actually perform the request. Raise HTTP exceptions on network failures.
+    3. Check the Content-Type header to make sure we got a Simple API response,
+       and raise `_NotAPIContent` otherwise.
+    """
+    if is_archive_file(Link(url).filename):
+        _ensure_api_response(url, session=session)
+
+    logger.debug("Getting page %s", redact_auth_from_url(url))
+
+    resp = session.get(
+        url,
+        headers={
+            "Accept": ", ".join(
+                [
+                    "application/vnd.pypi.simple.v1+json",
+                    "application/vnd.pypi.simple.v1+html; q=0.1",
+                    "text/html; q=0.01",
+                ]
+            ),
+            # We don't want to blindly returned cached data for
+            # /simple/, because authors generally expecting that
+            # twine upload && pip install will function, but if
+            # they've done a pip install in the last ~10 minutes
+            # it won't. Thus by setting this to zero we will not
+            # blindly use any cached data, however the benefit of
+            # using max-age=0 instead of no-cache, is that we will
+            # still support conditional requests, so we will still
+            # minimize traffic sent in cases where the page hasn't
+            # changed at all, we will just always incur the round
+            # trip for the conditional GET now instead of only
+            # once per 10 minutes.
+            # For more information, please see pypa/pip#5670.
+            "Cache-Control": "max-age=0",
+        },
+    )
+    raise_for_status(resp)
+
+    # The check for archives above only works if the url ends with
+    # something that looks like an archive. However that is not a
+    # requirement of an url. Unless we issue a HEAD request on every
+    # url we cannot know ahead of time for sure if something is a
+    # Simple API response or not. However we can check after we've
+    # downloaded it.
+    _ensure_api_header(resp)
+
+    logger.debug(
+        "Fetched page %s as %s",
+        redact_auth_from_url(url),
+        resp.headers.get("Content-Type", "Unknown"),
+    )
+
+    return resp
+
+
+def _get_encoding_from_headers(headers: ResponseHeaders) -> Optional[str]:
+    """Determine if we have any encoding information in our headers."""
+    if headers and "Content-Type" in headers:
+        m = email.message.Message()
+        m["content-type"] = headers["Content-Type"]
+        charset = m.get_param("charset")
+        if charset:
+            return str(charset)
+    return None
+
+
+class CacheablePageContent:
+    def __init__(self, page: "IndexContent") -> None:
+        assert page.cache_link_parsing
+        self.page = page
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, type(self)) and self.page.url == other.page.url
+
+    def __hash__(self) -> int:
+        return hash(self.page.url)
+
+
+class ParseLinks(Protocol):
+    def __call__(self, page: "IndexContent") -> Iterable[Link]: ...
+
+
+def with_cached_index_content(fn: ParseLinks) -> ParseLinks:
+    """
+    Given a function that parses an Iterable[Link] from an IndexContent, cache the
+    function's result (keyed by CacheablePageContent), unless the IndexContent
+    `page` has `page.cache_link_parsing == False`.
+    """
+
+    @functools.lru_cache(maxsize=None)
+    def wrapper(cacheable_page: CacheablePageContent) -> List[Link]:
+        return list(fn(cacheable_page.page))
+
+    @functools.wraps(fn)
+    def wrapper_wrapper(page: "IndexContent") -> List[Link]:
+        if page.cache_link_parsing:
+            return wrapper(CacheablePageContent(page))
+        return list(fn(page))
+
+    return wrapper_wrapper
+
+
+@with_cached_index_content
+def parse_links(page: "IndexContent") -> Iterable[Link]:
+    """
+    Parse a Simple API's Index Content, and yield its anchor elements as Link objects.
+    """
+
+    content_type_l = page.content_type.lower()
+    if content_type_l.startswith("application/vnd.pypi.simple.v1+json"):
+        data = json.loads(page.content)
+        for file in data.get("files", []):
+            link = Link.from_json(file, page.url)
+            if link is None:
+                continue
+            yield link
+        return
+
+    parser = HTMLLinkParser(page.url)
+    encoding = page.encoding or "utf-8"
+    parser.feed(page.content.decode(encoding))
+
+    url = page.url
+    base_url = parser.base_url or url
+    for anchor in parser.anchors:
+        link = Link.from_element(anchor, page_url=url, base_url=base_url)
+        if link is None:
+            continue
+        yield link
+
+
+@dataclass(frozen=True)
+class IndexContent:
+    """Represents one response (or page), along with its URL.
+
+    :param encoding: the encoding to decode the given content.
+    :param url: the URL from which the HTML was downloaded.
+    :param cache_link_parsing: whether links parsed from this page's url
+                               should be cached. PyPI index urls should
+                               have this set to False, for example.
+    """
+
+    content: bytes
+    content_type: str
+    encoding: Optional[str]
+    url: str
+    cache_link_parsing: bool = True
+
+    def __str__(self) -> str:
+        return redact_auth_from_url(self.url)
+
+
+class HTMLLinkParser(HTMLParser):
+    """
+    HTMLParser that keeps the first base HREF and a list of all anchor
+    elements' attributes.
+    """
+
+    def __init__(self, url: str) -> None:
+        super().__init__(convert_charrefs=True)
+
+        self.url: str = url
+        self.base_url: Optional[str] = None
+        self.anchors: List[Dict[str, Optional[str]]] = []
+
+    def handle_starttag(self, tag: str, attrs: List[Tuple[str, Optional[str]]]) -> None:
+        if tag == "base" and self.base_url is None:
+            href = self.get_href(attrs)
+            if href is not None:
+                self.base_url = href
+        elif tag == "a":
+            self.anchors.append(dict(attrs))
+
+    def get_href(self, attrs: List[Tuple[str, Optional[str]]]) -> Optional[str]:
+        for name, value in attrs:
+            if name == "href":
+                return value
+        return None
+
+
+def _handle_get_simple_fail(
+    link: Link,
+    reason: Union[str, Exception],
+    meth: Optional[Callable[..., None]] = None,
+) -> None:
+    if meth is None:
+        meth = logger.debug
+    meth("Could not fetch URL %s: %s - skipping", link, reason)
+
+
+def _make_index_content(
+    response: Response, cache_link_parsing: bool = True
+) -> IndexContent:
+    encoding = _get_encoding_from_headers(response.headers)
+    return IndexContent(
+        response.content,
+        response.headers["Content-Type"],
+        encoding=encoding,
+        url=response.url,
+        cache_link_parsing=cache_link_parsing,
+    )
+
+
+def _get_index_content(link: Link, *, session: PipSession) -> Optional["IndexContent"]:
+    url = link.url.split("#", 1)[0]
+
+    # Check for VCS schemes that do not support lookup as web pages.
+    vcs_scheme = _match_vcs_scheme(url)
+    if vcs_scheme:
+        logger.warning(
+            "Cannot look at %s URL %s because it does not support lookup as web pages.",
+            vcs_scheme,
+            link,
+        )
+        return None
+
+    # Tack index.html onto file:// URLs that point to directories
+    scheme, _, path, _, _, _ = urllib.parse.urlparse(url)
+    if scheme == "file" and os.path.isdir(urllib.request.url2pathname(path)):
+        # add trailing slash if not present so urljoin doesn't trim
+        # final segment
+        if not url.endswith("/"):
+            url += "/"
+        # TODO: In the future, it would be nice if pip supported PEP 691
+        #       style responses in the file:// URLs, however there's no
+        #       standard file extension for application/vnd.pypi.simple.v1+json
+        #       so we'll need to come up with something on our own.
+        url = urllib.parse.urljoin(url, "index.html")
+        logger.debug(" file: URL is directory, getting %s", url)
+
+    try:
+        resp = _get_simple_response(url, session=session)
+    except _NotHTTP:
+        logger.warning(
+            "Skipping page %s because it looks like an archive, and cannot "
+            "be checked by a HTTP HEAD request.",
+            link,
+        )
+    except _NotAPIContent as exc:
+        logger.warning(
+            "Skipping page %s because the %s request got Content-Type: %s. "
+            "The only supported Content-Types are application/vnd.pypi.simple.v1+json, "
+            "application/vnd.pypi.simple.v1+html, and text/html",
+            link,
+            exc.request_desc,
+            exc.content_type,
+        )
+    except NetworkConnectionError as exc:
+        _handle_get_simple_fail(link, exc)
+    except RetryError as exc:
+        _handle_get_simple_fail(link, exc)
+    except SSLError as exc:
+        reason = "There was a problem confirming the ssl certificate: "
+        reason += str(exc)
+        _handle_get_simple_fail(link, reason, meth=logger.info)
+    except requests.ConnectionError as exc:
+        _handle_get_simple_fail(link, f"connection error: {exc}")
+    except requests.Timeout:
+        _handle_get_simple_fail(link, "timed out")
+    else:
+        return _make_index_content(resp, cache_link_parsing=link.cache_link_parsing)
+    return None
+
+
+class CollectedSources(NamedTuple):
+    find_links: Sequence[Optional[LinkSource]]
+    index_urls: Sequence[Optional[LinkSource]]
+
+
+class LinkCollector:
+    """
+    Responsible for collecting Link objects from all configured locations,
+    making network requests as needed.
+
+    The class's main method is its collect_sources() method.
+    """
+
+    def __init__(
+        self,
+        session: PipSession,
+        search_scope: SearchScope,
+    ) -> None:
+        self.search_scope = search_scope
+        self.session = session
+
+    @classmethod
+    def create(
+        cls,
+        session: PipSession,
+        options: Values,
+        suppress_no_index: bool = False,
+    ) -> "LinkCollector":
+        """
+        :param session: The Session to use to make requests.
+        :param suppress_no_index: Whether to ignore the --no-index option
+            when constructing the SearchScope object.
+        """
+        index_urls = [options.index_url] + options.extra_index_urls
+        if options.no_index and not suppress_no_index:
+            logger.debug(
+                "Ignoring indexes: %s",
+                ",".join(redact_auth_from_url(url) for url in index_urls),
+            )
+            index_urls = []
+
+        # Make sure find_links is a list before passing to create().
+        find_links = options.find_links or []
+
+        search_scope = SearchScope.create(
+            find_links=find_links,
+            index_urls=index_urls,
+            no_index=options.no_index,
+        )
+        link_collector = LinkCollector(
+            session=session,
+            search_scope=search_scope,
+        )
+        return link_collector
+
+    @property
+    def find_links(self) -> List[str]:
+        return self.search_scope.find_links
+
+    def fetch_response(self, location: Link) -> Optional[IndexContent]:
+        """
+        Fetch an HTML page containing package links.
+        """
+        return _get_index_content(location, session=self.session)
+
+    def collect_sources(
+        self,
+        project_name: str,
+        candidates_from_page: CandidatesFromPage,
+    ) -> CollectedSources:
+        # The OrderedDict calls deduplicate sources by URL.
+        index_url_sources = collections.OrderedDict(
+            build_source(
+                loc,
+                candidates_from_page=candidates_from_page,
+                page_validator=self.session.is_secure_origin,
+                expand_dir=False,
+                cache_link_parsing=False,
+                project_name=project_name,
+            )
+            for loc in self.search_scope.get_index_urls_locations(project_name)
+        ).values()
+        find_links_sources = collections.OrderedDict(
+            build_source(
+                loc,
+                candidates_from_page=candidates_from_page,
+                page_validator=self.session.is_secure_origin,
+                expand_dir=True,
+                cache_link_parsing=True,
+                project_name=project_name,
+            )
+            for loc in self.find_links
+        ).values()
+
+        if logger.isEnabledFor(logging.DEBUG):
+            lines = [
+                f"* {s.link}"
+                for s in itertools.chain(find_links_sources, index_url_sources)
+                if s is not None and s.link is not None
+            ]
+            lines = [
+                f"{len(lines)} location(s) to search "
+                f"for versions of {project_name}:"
+            ] + lines
+            logger.debug("\n".join(lines))
+
+        return CollectedSources(
+            find_links=list(find_links_sources),
+            index_urls=list(index_url_sources),
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/package_finder.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/package_finder.py
new file mode 100644
index 00000000..85628ee5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/package_finder.py
@@ -0,0 +1,1029 @@
+"""Routines related to PyPI, indexes"""
+
+import enum
+import functools
+import itertools
+import logging
+import re
+from dataclasses import dataclass
+from typing import TYPE_CHECKING, FrozenSet, Iterable, List, Optional, Set, Tuple, Union
+
+from pip._vendor.packaging import specifiers
+from pip._vendor.packaging.tags import Tag
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.packaging.version import InvalidVersion, _BaseVersion
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.exceptions import (
+    BestVersionAlreadyInstalled,
+    DistributionNotFound,
+    InvalidWheelFilename,
+    UnsupportedWheel,
+)
+from pip._internal.index.collector import LinkCollector, parse_links
+from pip._internal.models.candidate import InstallationCandidate
+from pip._internal.models.format_control import FormatControl
+from pip._internal.models.link import Link
+from pip._internal.models.search_scope import SearchScope
+from pip._internal.models.selection_prefs import SelectionPreferences
+from pip._internal.models.target_python import TargetPython
+from pip._internal.models.wheel import Wheel
+from pip._internal.req import InstallRequirement
+from pip._internal.utils._log import getLogger
+from pip._internal.utils.filetypes import WHEEL_EXTENSION
+from pip._internal.utils.hashes import Hashes
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import build_netloc
+from pip._internal.utils.packaging import check_requires_python
+from pip._internal.utils.unpacking import SUPPORTED_EXTENSIONS
+
+if TYPE_CHECKING:
+    from pip._vendor.typing_extensions import TypeGuard
+
+__all__ = ["FormatControl", "BestCandidateResult", "PackageFinder"]
+
+
+logger = getLogger(__name__)
+
+BuildTag = Union[Tuple[()], Tuple[int, str]]
+CandidateSortingKey = Tuple[int, int, int, _BaseVersion, Optional[int], BuildTag]
+
+
+def _check_link_requires_python(
+    link: Link,
+    version_info: Tuple[int, int, int],
+    ignore_requires_python: bool = False,
+) -> bool:
+    """
+    Return whether the given Python version is compatible with a link's
+    "Requires-Python" value.
+
+    :param version_info: A 3-tuple of ints representing the Python
+        major-minor-micro version to check.
+    :param ignore_requires_python: Whether to ignore the "Requires-Python"
+        value if the given Python version isn't compatible.
+    """
+    try:
+        is_compatible = check_requires_python(
+            link.requires_python,
+            version_info=version_info,
+        )
+    except specifiers.InvalidSpecifier:
+        logger.debug(
+            "Ignoring invalid Requires-Python (%r) for link: %s",
+            link.requires_python,
+            link,
+        )
+    else:
+        if not is_compatible:
+            version = ".".join(map(str, version_info))
+            if not ignore_requires_python:
+                logger.verbose(
+                    "Link requires a different Python (%s not in: %r): %s",
+                    version,
+                    link.requires_python,
+                    link,
+                )
+                return False
+
+            logger.debug(
+                "Ignoring failed Requires-Python check (%s not in: %r) for link: %s",
+                version,
+                link.requires_python,
+                link,
+            )
+
+    return True
+
+
+class LinkType(enum.Enum):
+    candidate = enum.auto()
+    different_project = enum.auto()
+    yanked = enum.auto()
+    format_unsupported = enum.auto()
+    format_invalid = enum.auto()
+    platform_mismatch = enum.auto()
+    requires_python_mismatch = enum.auto()
+
+
+class LinkEvaluator:
+    """
+    Responsible for evaluating links for a particular project.
+    """
+
+    _py_version_re = re.compile(r"-py([123]\.?[0-9]?)$")
+
+    # Don't include an allow_yanked default value to make sure each call
+    # site considers whether yanked releases are allowed. This also causes
+    # that decision to be made explicit in the calling code, which helps
+    # people when reading the code.
+    def __init__(
+        self,
+        project_name: str,
+        canonical_name: str,
+        formats: FrozenSet[str],
+        target_python: TargetPython,
+        allow_yanked: bool,
+        ignore_requires_python: Optional[bool] = None,
+    ) -> None:
+        """
+        :param project_name: The user supplied package name.
+        :param canonical_name: The canonical package name.
+        :param formats: The formats allowed for this package. Should be a set
+            with 'binary' or 'source' or both in it.
+        :param target_python: The target Python interpreter to use when
+            evaluating link compatibility. This is used, for example, to
+            check wheel compatibility, as well as when checking the Python
+            version, e.g. the Python version embedded in a link filename
+            (or egg fragment) and against an HTML link's optional PEP 503
+            "data-requires-python" attribute.
+        :param allow_yanked: Whether files marked as yanked (in the sense
+            of PEP 592) are permitted to be candidates for install.
+        :param ignore_requires_python: Whether to ignore incompatible
+            PEP 503 "data-requires-python" values in HTML links. Defaults
+            to False.
+        """
+        if ignore_requires_python is None:
+            ignore_requires_python = False
+
+        self._allow_yanked = allow_yanked
+        self._canonical_name = canonical_name
+        self._ignore_requires_python = ignore_requires_python
+        self._formats = formats
+        self._target_python = target_python
+
+        self.project_name = project_name
+
+    def evaluate_link(self, link: Link) -> Tuple[LinkType, str]:
+        """
+        Determine whether a link is a candidate for installation.
+
+        :return: A tuple (result, detail), where *result* is an enum
+            representing whether the evaluation found a candidate, or the reason
+            why one is not found. If a candidate is found, *detail* will be the
+            candidate's version string; if one is not found, it contains the
+            reason the link fails to qualify.
+        """
+        version = None
+        if link.is_yanked and not self._allow_yanked:
+            reason = link.yanked_reason or "<none given>"
+            return (LinkType.yanked, f"yanked for reason: {reason}")
+
+        if link.egg_fragment:
+            egg_info = link.egg_fragment
+            ext = link.ext
+        else:
+            egg_info, ext = link.splitext()
+            if not ext:
+                return (LinkType.format_unsupported, "not a file")
+            if ext not in SUPPORTED_EXTENSIONS:
+                return (
+                    LinkType.format_unsupported,
+                    f"unsupported archive format: {ext}",
+                )
+            if "binary" not in self._formats and ext == WHEEL_EXTENSION:
+                reason = f"No binaries permitted for {self.project_name}"
+                return (LinkType.format_unsupported, reason)
+            if "macosx10" in link.path and ext == ".zip":
+                return (LinkType.format_unsupported, "macosx10 one")
+            if ext == WHEEL_EXTENSION:
+                try:
+                    wheel = Wheel(link.filename)
+                except InvalidWheelFilename:
+                    return (
+                        LinkType.format_invalid,
+                        "invalid wheel filename",
+                    )
+                if canonicalize_name(wheel.name) != self._canonical_name:
+                    reason = f"wrong project name (not {self.project_name})"
+                    return (LinkType.different_project, reason)
+
+                supported_tags = self._target_python.get_unsorted_tags()
+                if not wheel.supported(supported_tags):
+                    # Include the wheel's tags in the reason string to
+                    # simplify troubleshooting compatibility issues.
+                    file_tags = ", ".join(wheel.get_formatted_file_tags())
+                    reason = (
+                        f"none of the wheel's tags ({file_tags}) are compatible "
+                        f"(run pip debug --verbose to show compatible tags)"
+                    )
+                    return (LinkType.platform_mismatch, reason)
+
+                version = wheel.version
+
+        # This should be up by the self.ok_binary check, but see issue 2700.
+        if "source" not in self._formats and ext != WHEEL_EXTENSION:
+            reason = f"No sources permitted for {self.project_name}"
+            return (LinkType.format_unsupported, reason)
+
+        if not version:
+            version = _extract_version_from_fragment(
+                egg_info,
+                self._canonical_name,
+            )
+        if not version:
+            reason = f"Missing project version for {self.project_name}"
+            return (LinkType.format_invalid, reason)
+
+        match = self._py_version_re.search(version)
+        if match:
+            version = version[: match.start()]
+            py_version = match.group(1)
+            if py_version != self._target_python.py_version:
+                return (
+                    LinkType.platform_mismatch,
+                    "Python version is incorrect",
+                )
+
+        supports_python = _check_link_requires_python(
+            link,
+            version_info=self._target_python.py_version_info,
+            ignore_requires_python=self._ignore_requires_python,
+        )
+        if not supports_python:
+            reason = f"{version} Requires-Python {link.requires_python}"
+            return (LinkType.requires_python_mismatch, reason)
+
+        logger.debug("Found link %s, version: %s", link, version)
+
+        return (LinkType.candidate, version)
+
+
+def filter_unallowed_hashes(
+    candidates: List[InstallationCandidate],
+    hashes: Optional[Hashes],
+    project_name: str,
+) -> List[InstallationCandidate]:
+    """
+    Filter out candidates whose hashes aren't allowed, and return a new
+    list of candidates.
+
+    If at least one candidate has an allowed hash, then all candidates with
+    either an allowed hash or no hash specified are returned.  Otherwise,
+    the given candidates are returned.
+
+    Including the candidates with no hash specified when there is a match
+    allows a warning to be logged if there is a more preferred candidate
+    with no hash specified.  Returning all candidates in the case of no
+    matches lets pip report the hash of the candidate that would otherwise
+    have been installed (e.g. permitting the user to more easily update
+    their requirements file with the desired hash).
+    """
+    if not hashes:
+        logger.debug(
+            "Given no hashes to check %s links for project %r: "
+            "discarding no candidates",
+            len(candidates),
+            project_name,
+        )
+        # Make sure we're not returning back the given value.
+        return list(candidates)
+
+    matches_or_no_digest = []
+    # Collect the non-matches for logging purposes.
+    non_matches = []
+    match_count = 0
+    for candidate in candidates:
+        link = candidate.link
+        if not link.has_hash:
+            pass
+        elif link.is_hash_allowed(hashes=hashes):
+            match_count += 1
+        else:
+            non_matches.append(candidate)
+            continue
+
+        matches_or_no_digest.append(candidate)
+
+    if match_count:
+        filtered = matches_or_no_digest
+    else:
+        # Make sure we're not returning back the given value.
+        filtered = list(candidates)
+
+    if len(filtered) == len(candidates):
+        discard_message = "discarding no candidates"
+    else:
+        discard_message = "discarding {} non-matches:\n  {}".format(
+            len(non_matches),
+            "\n  ".join(str(candidate.link) for candidate in non_matches),
+        )
+
+    logger.debug(
+        "Checked %s links for project %r against %s hashes "
+        "(%s matches, %s no digest): %s",
+        len(candidates),
+        project_name,
+        hashes.digest_count,
+        match_count,
+        len(matches_or_no_digest) - match_count,
+        discard_message,
+    )
+
+    return filtered
+
+
+@dataclass
+class CandidatePreferences:
+    """
+    Encapsulates some of the preferences for filtering and sorting
+    InstallationCandidate objects.
+    """
+
+    prefer_binary: bool = False
+    allow_all_prereleases: bool = False
+
+
+@dataclass(frozen=True)
+class BestCandidateResult:
+    """A collection of candidates, returned by `PackageFinder.find_best_candidate`.
+
+    This class is only intended to be instantiated by CandidateEvaluator's
+    `compute_best_candidate()` method.
+
+    :param all_candidates: A sequence of all available candidates found.
+    :param applicable_candidates: The applicable candidates.
+    :param best_candidate: The most preferred candidate found, or None
+        if no applicable candidates were found.
+    """
+
+    all_candidates: List[InstallationCandidate]
+    applicable_candidates: List[InstallationCandidate]
+    best_candidate: Optional[InstallationCandidate]
+
+    def __post_init__(self) -> None:
+        assert set(self.applicable_candidates) <= set(self.all_candidates)
+
+        if self.best_candidate is None:
+            assert not self.applicable_candidates
+        else:
+            assert self.best_candidate in self.applicable_candidates
+
+
+class CandidateEvaluator:
+    """
+    Responsible for filtering and sorting candidates for installation based
+    on what tags are valid.
+    """
+
+    @classmethod
+    def create(
+        cls,
+        project_name: str,
+        target_python: Optional[TargetPython] = None,
+        prefer_binary: bool = False,
+        allow_all_prereleases: bool = False,
+        specifier: Optional[specifiers.BaseSpecifier] = None,
+        hashes: Optional[Hashes] = None,
+    ) -> "CandidateEvaluator":
+        """Create a CandidateEvaluator object.
+
+        :param target_python: The target Python interpreter to use when
+            checking compatibility. If None (the default), a TargetPython
+            object will be constructed from the running Python.
+        :param specifier: An optional object implementing `filter`
+            (e.g. `packaging.specifiers.SpecifierSet`) to filter applicable
+            versions.
+        :param hashes: An optional collection of allowed hashes.
+        """
+        if target_python is None:
+            target_python = TargetPython()
+        if specifier is None:
+            specifier = specifiers.SpecifierSet()
+
+        supported_tags = target_python.get_sorted_tags()
+
+        return cls(
+            project_name=project_name,
+            supported_tags=supported_tags,
+            specifier=specifier,
+            prefer_binary=prefer_binary,
+            allow_all_prereleases=allow_all_prereleases,
+            hashes=hashes,
+        )
+
+    def __init__(
+        self,
+        project_name: str,
+        supported_tags: List[Tag],
+        specifier: specifiers.BaseSpecifier,
+        prefer_binary: bool = False,
+        allow_all_prereleases: bool = False,
+        hashes: Optional[Hashes] = None,
+    ) -> None:
+        """
+        :param supported_tags: The PEP 425 tags supported by the target
+            Python in order of preference (most preferred first).
+        """
+        self._allow_all_prereleases = allow_all_prereleases
+        self._hashes = hashes
+        self._prefer_binary = prefer_binary
+        self._project_name = project_name
+        self._specifier = specifier
+        self._supported_tags = supported_tags
+        # Since the index of the tag in the _supported_tags list is used
+        # as a priority, precompute a map from tag to index/priority to be
+        # used in wheel.find_most_preferred_tag.
+        self._wheel_tag_preferences = {
+            tag: idx for idx, tag in enumerate(supported_tags)
+        }
+
+    def get_applicable_candidates(
+        self,
+        candidates: List[InstallationCandidate],
+    ) -> List[InstallationCandidate]:
+        """
+        Return the applicable candidates from a list of candidates.
+        """
+        # Using None infers from the specifier instead.
+        allow_prereleases = self._allow_all_prereleases or None
+        specifier = self._specifier
+
+        # We turn the version object into a str here because otherwise
+        # when we're debundled but setuptools isn't, Python will see
+        # packaging.version.Version and
+        # pkg_resources._vendor.packaging.version.Version as different
+        # types. This way we'll use a str as a common data interchange
+        # format. If we stop using the pkg_resources provided specifier
+        # and start using our own, we can drop the cast to str().
+        candidates_and_versions = [(c, str(c.version)) for c in candidates]
+        versions = set(
+            specifier.filter(
+                (v for _, v in candidates_and_versions),
+                prereleases=allow_prereleases,
+            )
+        )
+
+        applicable_candidates = [c for c, v in candidates_and_versions if v in versions]
+        filtered_applicable_candidates = filter_unallowed_hashes(
+            candidates=applicable_candidates,
+            hashes=self._hashes,
+            project_name=self._project_name,
+        )
+
+        return sorted(filtered_applicable_candidates, key=self._sort_key)
+
+    def _sort_key(self, candidate: InstallationCandidate) -> CandidateSortingKey:
+        """
+        Function to pass as the `key` argument to a call to sorted() to sort
+        InstallationCandidates by preference.
+
+        Returns a tuple such that tuples sorting as greater using Python's
+        default comparison operator are more preferred.
+
+        The preference is as follows:
+
+        First and foremost, candidates with allowed (matching) hashes are
+        always preferred over candidates without matching hashes. This is
+        because e.g. if the only candidate with an allowed hash is yanked,
+        we still want to use that candidate.
+
+        Second, excepting hash considerations, candidates that have been
+        yanked (in the sense of PEP 592) are always less preferred than
+        candidates that haven't been yanked. Then:
+
+        If not finding wheels, they are sorted by version only.
+        If finding wheels, then the sort order is by version, then:
+          1. existing installs
+          2. wheels ordered via Wheel.support_index_min(self._supported_tags)
+          3. source archives
+        If prefer_binary was set, then all wheels are sorted above sources.
+
+        Note: it was considered to embed this logic into the Link
+              comparison operators, but then different sdist links
+              with the same version, would have to be considered equal
+        """
+        valid_tags = self._supported_tags
+        support_num = len(valid_tags)
+        build_tag: BuildTag = ()
+        binary_preference = 0
+        link = candidate.link
+        if link.is_wheel:
+            # can raise InvalidWheelFilename
+            wheel = Wheel(link.filename)
+            try:
+                pri = -(
+                    wheel.find_most_preferred_tag(
+                        valid_tags, self._wheel_tag_preferences
+                    )
+                )
+            except ValueError:
+                raise UnsupportedWheel(
+                    f"{wheel.filename} is not a supported wheel for this platform. It "
+                    "can't be sorted."
+                )
+            if self._prefer_binary:
+                binary_preference = 1
+            if wheel.build_tag is not None:
+                match = re.match(r"^(\d+)(.*)$", wheel.build_tag)
+                assert match is not None, "guaranteed by filename validation"
+                build_tag_groups = match.groups()
+                build_tag = (int(build_tag_groups[0]), build_tag_groups[1])
+        else:  # sdist
+            pri = -(support_num)
+        has_allowed_hash = int(link.is_hash_allowed(self._hashes))
+        yank_value = -1 * int(link.is_yanked)  # -1 for yanked.
+        return (
+            has_allowed_hash,
+            yank_value,
+            binary_preference,
+            candidate.version,
+            pri,
+            build_tag,
+        )
+
+    def sort_best_candidate(
+        self,
+        candidates: List[InstallationCandidate],
+    ) -> Optional[InstallationCandidate]:
+        """
+        Return the best candidate per the instance's sort order, or None if
+        no candidate is acceptable.
+        """
+        if not candidates:
+            return None
+        best_candidate = max(candidates, key=self._sort_key)
+        return best_candidate
+
+    def compute_best_candidate(
+        self,
+        candidates: List[InstallationCandidate],
+    ) -> BestCandidateResult:
+        """
+        Compute and return a `BestCandidateResult` instance.
+        """
+        applicable_candidates = self.get_applicable_candidates(candidates)
+
+        best_candidate = self.sort_best_candidate(applicable_candidates)
+
+        return BestCandidateResult(
+            candidates,
+            applicable_candidates=applicable_candidates,
+            best_candidate=best_candidate,
+        )
+
+
+class PackageFinder:
+    """This finds packages.
+
+    This is meant to match easy_install's technique for looking for
+    packages, by reading pages and looking for appropriate links.
+    """
+
+    def __init__(
+        self,
+        link_collector: LinkCollector,
+        target_python: TargetPython,
+        allow_yanked: bool,
+        format_control: Optional[FormatControl] = None,
+        candidate_prefs: Optional[CandidatePreferences] = None,
+        ignore_requires_python: Optional[bool] = None,
+    ) -> None:
+        """
+        This constructor is primarily meant to be used by the create() class
+        method and from tests.
+
+        :param format_control: A FormatControl object, used to control
+            the selection of source packages / binary packages when consulting
+            the index and links.
+        :param candidate_prefs: Options to use when creating a
+            CandidateEvaluator object.
+        """
+        if candidate_prefs is None:
+            candidate_prefs = CandidatePreferences()
+
+        format_control = format_control or FormatControl(set(), set())
+
+        self._allow_yanked = allow_yanked
+        self._candidate_prefs = candidate_prefs
+        self._ignore_requires_python = ignore_requires_python
+        self._link_collector = link_collector
+        self._target_python = target_python
+
+        self.format_control = format_control
+
+        # These are boring links that have already been logged somehow.
+        self._logged_links: Set[Tuple[Link, LinkType, str]] = set()
+
+    # Don't include an allow_yanked default value to make sure each call
+    # site considers whether yanked releases are allowed. This also causes
+    # that decision to be made explicit in the calling code, which helps
+    # people when reading the code.
+    @classmethod
+    def create(
+        cls,
+        link_collector: LinkCollector,
+        selection_prefs: SelectionPreferences,
+        target_python: Optional[TargetPython] = None,
+    ) -> "PackageFinder":
+        """Create a PackageFinder.
+
+        :param selection_prefs: The candidate selection preferences, as a
+            SelectionPreferences object.
+        :param target_python: The target Python interpreter to use when
+            checking compatibility. If None (the default), a TargetPython
+            object will be constructed from the running Python.
+        """
+        if target_python is None:
+            target_python = TargetPython()
+
+        candidate_prefs = CandidatePreferences(
+            prefer_binary=selection_prefs.prefer_binary,
+            allow_all_prereleases=selection_prefs.allow_all_prereleases,
+        )
+
+        return cls(
+            candidate_prefs=candidate_prefs,
+            link_collector=link_collector,
+            target_python=target_python,
+            allow_yanked=selection_prefs.allow_yanked,
+            format_control=selection_prefs.format_control,
+            ignore_requires_python=selection_prefs.ignore_requires_python,
+        )
+
+    @property
+    def target_python(self) -> TargetPython:
+        return self._target_python
+
+    @property
+    def search_scope(self) -> SearchScope:
+        return self._link_collector.search_scope
+
+    @search_scope.setter
+    def search_scope(self, search_scope: SearchScope) -> None:
+        self._link_collector.search_scope = search_scope
+
+    @property
+    def find_links(self) -> List[str]:
+        return self._link_collector.find_links
+
+    @property
+    def index_urls(self) -> List[str]:
+        return self.search_scope.index_urls
+
+    @property
+    def proxy(self) -> Optional[str]:
+        return self._link_collector.session.pip_proxy
+
+    @property
+    def trusted_hosts(self) -> Iterable[str]:
+        for host_port in self._link_collector.session.pip_trusted_origins:
+            yield build_netloc(*host_port)
+
+    @property
+    def custom_cert(self) -> Optional[str]:
+        # session.verify is either a boolean (use default bundle/no SSL
+        # verification) or a string path to a custom CA bundle to use. We only
+        # care about the latter.
+        verify = self._link_collector.session.verify
+        return verify if isinstance(verify, str) else None
+
+    @property
+    def client_cert(self) -> Optional[str]:
+        cert = self._link_collector.session.cert
+        assert not isinstance(cert, tuple), "pip only supports PEM client certs"
+        return cert
+
+    @property
+    def allow_all_prereleases(self) -> bool:
+        return self._candidate_prefs.allow_all_prereleases
+
+    def set_allow_all_prereleases(self) -> None:
+        self._candidate_prefs.allow_all_prereleases = True
+
+    @property
+    def prefer_binary(self) -> bool:
+        return self._candidate_prefs.prefer_binary
+
+    def set_prefer_binary(self) -> None:
+        self._candidate_prefs.prefer_binary = True
+
+    def requires_python_skipped_reasons(self) -> List[str]:
+        reasons = {
+            detail
+            for _, result, detail in self._logged_links
+            if result == LinkType.requires_python_mismatch
+        }
+        return sorted(reasons)
+
+    def make_link_evaluator(self, project_name: str) -> LinkEvaluator:
+        canonical_name = canonicalize_name(project_name)
+        formats = self.format_control.get_allowed_formats(canonical_name)
+
+        return LinkEvaluator(
+            project_name=project_name,
+            canonical_name=canonical_name,
+            formats=formats,
+            target_python=self._target_python,
+            allow_yanked=self._allow_yanked,
+            ignore_requires_python=self._ignore_requires_python,
+        )
+
+    def _sort_links(self, links: Iterable[Link]) -> List[Link]:
+        """
+        Returns elements of links in order, non-egg links first, egg links
+        second, while eliminating duplicates
+        """
+        eggs, no_eggs = [], []
+        seen: Set[Link] = set()
+        for link in links:
+            if link not in seen:
+                seen.add(link)
+                if link.egg_fragment:
+                    eggs.append(link)
+                else:
+                    no_eggs.append(link)
+        return no_eggs + eggs
+
+    def _log_skipped_link(self, link: Link, result: LinkType, detail: str) -> None:
+        # This is a hot method so don't waste time hashing links unless we're
+        # actually going to log 'em.
+        if not logger.isEnabledFor(logging.DEBUG):
+            return
+
+        entry = (link, result, detail)
+        if entry not in self._logged_links:
+            # Put the link at the end so the reason is more visible and because
+            # the link string is usually very long.
+            logger.debug("Skipping link: %s: %s", detail, link)
+            self._logged_links.add(entry)
+
+    def get_install_candidate(
+        self, link_evaluator: LinkEvaluator, link: Link
+    ) -> Optional[InstallationCandidate]:
+        """
+        If the link is a candidate for install, convert it to an
+        InstallationCandidate and return it. Otherwise, return None.
+        """
+        result, detail = link_evaluator.evaluate_link(link)
+        if result != LinkType.candidate:
+            self._log_skipped_link(link, result, detail)
+            return None
+
+        try:
+            return InstallationCandidate(
+                name=link_evaluator.project_name,
+                link=link,
+                version=detail,
+            )
+        except InvalidVersion:
+            return None
+
+    def evaluate_links(
+        self, link_evaluator: LinkEvaluator, links: Iterable[Link]
+    ) -> List[InstallationCandidate]:
+        """
+        Convert links that are candidates to InstallationCandidate objects.
+        """
+        candidates = []
+        for link in self._sort_links(links):
+            candidate = self.get_install_candidate(link_evaluator, link)
+            if candidate is not None:
+                candidates.append(candidate)
+
+        return candidates
+
+    def process_project_url(
+        self, project_url: Link, link_evaluator: LinkEvaluator
+    ) -> List[InstallationCandidate]:
+        logger.debug(
+            "Fetching project page and analyzing links: %s",
+            project_url,
+        )
+        index_response = self._link_collector.fetch_response(project_url)
+        if index_response is None:
+            return []
+
+        page_links = list(parse_links(index_response))
+
+        with indent_log():
+            package_links = self.evaluate_links(
+                link_evaluator,
+                links=page_links,
+            )
+
+        return package_links
+
+    @functools.lru_cache(maxsize=None)
+    def find_all_candidates(self, project_name: str) -> List[InstallationCandidate]:
+        """Find all available InstallationCandidate for project_name
+
+        This checks index_urls and find_links.
+        All versions found are returned as an InstallationCandidate list.
+
+        See LinkEvaluator.evaluate_link() for details on which files
+        are accepted.
+        """
+        link_evaluator = self.make_link_evaluator(project_name)
+
+        collected_sources = self._link_collector.collect_sources(
+            project_name=project_name,
+            candidates_from_page=functools.partial(
+                self.process_project_url,
+                link_evaluator=link_evaluator,
+            ),
+        )
+
+        page_candidates_it = itertools.chain.from_iterable(
+            source.page_candidates()
+            for sources in collected_sources
+            for source in sources
+            if source is not None
+        )
+        page_candidates = list(page_candidates_it)
+
+        file_links_it = itertools.chain.from_iterable(
+            source.file_links()
+            for sources in collected_sources
+            for source in sources
+            if source is not None
+        )
+        file_candidates = self.evaluate_links(
+            link_evaluator,
+            sorted(file_links_it, reverse=True),
+        )
+
+        if logger.isEnabledFor(logging.DEBUG) and file_candidates:
+            paths = []
+            for candidate in file_candidates:
+                assert candidate.link.url  # we need to have a URL
+                try:
+                    paths.append(candidate.link.file_path)
+                except Exception:
+                    paths.append(candidate.link.url)  # it's not a local file
+
+            logger.debug("Local files found: %s", ", ".join(paths))
+
+        # This is an intentional priority ordering
+        return file_candidates + page_candidates
+
+    def make_candidate_evaluator(
+        self,
+        project_name: str,
+        specifier: Optional[specifiers.BaseSpecifier] = None,
+        hashes: Optional[Hashes] = None,
+    ) -> CandidateEvaluator:
+        """Create a CandidateEvaluator object to use."""
+        candidate_prefs = self._candidate_prefs
+        return CandidateEvaluator.create(
+            project_name=project_name,
+            target_python=self._target_python,
+            prefer_binary=candidate_prefs.prefer_binary,
+            allow_all_prereleases=candidate_prefs.allow_all_prereleases,
+            specifier=specifier,
+            hashes=hashes,
+        )
+
+    @functools.lru_cache(maxsize=None)
+    def find_best_candidate(
+        self,
+        project_name: str,
+        specifier: Optional[specifiers.BaseSpecifier] = None,
+        hashes: Optional[Hashes] = None,
+    ) -> BestCandidateResult:
+        """Find matches for the given project and specifier.
+
+        :param specifier: An optional object implementing `filter`
+            (e.g. `packaging.specifiers.SpecifierSet`) to filter applicable
+            versions.
+
+        :return: A `BestCandidateResult` instance.
+        """
+        candidates = self.find_all_candidates(project_name)
+        candidate_evaluator = self.make_candidate_evaluator(
+            project_name=project_name,
+            specifier=specifier,
+            hashes=hashes,
+        )
+        return candidate_evaluator.compute_best_candidate(candidates)
+
+    def find_requirement(
+        self, req: InstallRequirement, upgrade: bool
+    ) -> Optional[InstallationCandidate]:
+        """Try to find a Link matching req
+
+        Expects req, an InstallRequirement and upgrade, a boolean
+        Returns a InstallationCandidate if found,
+        Raises DistributionNotFound or BestVersionAlreadyInstalled otherwise
+        """
+        hashes = req.hashes(trust_internet=False)
+        best_candidate_result = self.find_best_candidate(
+            req.name,
+            specifier=req.specifier,
+            hashes=hashes,
+        )
+        best_candidate = best_candidate_result.best_candidate
+
+        installed_version: Optional[_BaseVersion] = None
+        if req.satisfied_by is not None:
+            installed_version = req.satisfied_by.version
+
+        def _format_versions(cand_iter: Iterable[InstallationCandidate]) -> str:
+            # This repeated parse_version and str() conversion is needed to
+            # handle different vendoring sources from pip and pkg_resources.
+            # If we stop using the pkg_resources provided specifier and start
+            # using our own, we can drop the cast to str().
+            return (
+                ", ".join(
+                    sorted(
+                        {str(c.version) for c in cand_iter},
+                        key=parse_version,
+                    )
+                )
+                or "none"
+            )
+
+        if installed_version is None and best_candidate is None:
+            logger.critical(
+                "Could not find a version that satisfies the requirement %s "
+                "(from versions: %s)",
+                req,
+                _format_versions(best_candidate_result.all_candidates),
+            )
+
+            raise DistributionNotFound(f"No matching distribution found for {req}")
+
+        def _should_install_candidate(
+            candidate: Optional[InstallationCandidate],
+        ) -> "TypeGuard[InstallationCandidate]":
+            if installed_version is None:
+                return True
+            if best_candidate is None:
+                return False
+            return best_candidate.version > installed_version
+
+        if not upgrade and installed_version is not None:
+            if _should_install_candidate(best_candidate):
+                logger.debug(
+                    "Existing installed version (%s) satisfies requirement "
+                    "(most up-to-date version is %s)",
+                    installed_version,
+                    best_candidate.version,
+                )
+            else:
+                logger.debug(
+                    "Existing installed version (%s) is most up-to-date and "
+                    "satisfies requirement",
+                    installed_version,
+                )
+            return None
+
+        if _should_install_candidate(best_candidate):
+            logger.debug(
+                "Using version %s (newest of versions: %s)",
+                best_candidate.version,
+                _format_versions(best_candidate_result.applicable_candidates),
+            )
+            return best_candidate
+
+        # We have an existing version, and its the best version
+        logger.debug(
+            "Installed version (%s) is most up-to-date (past versions: %s)",
+            installed_version,
+            _format_versions(best_candidate_result.applicable_candidates),
+        )
+        raise BestVersionAlreadyInstalled
+
+
+def _find_name_version_sep(fragment: str, canonical_name: str) -> int:
+    """Find the separator's index based on the package's canonical name.
+
+    :param fragment: A <package>+<version> filename "fragment" (stem) or
+        egg fragment.
+    :param canonical_name: The package's canonical name.
+
+    This function is needed since the canonicalized name does not necessarily
+    have the same length as the egg info's name part. An example::
+
+    >>> fragment = 'foo__bar-1.0'
+    >>> canonical_name = 'foo-bar'
+    >>> _find_name_version_sep(fragment, canonical_name)
+    8
+    """
+    # Project name and version must be separated by one single dash. Find all
+    # occurrences of dashes; if the string in front of it matches the canonical
+    # name, this is the one separating the name and version parts.
+    for i, c in enumerate(fragment):
+        if c != "-":
+            continue
+        if canonicalize_name(fragment[:i]) == canonical_name:
+            return i
+    raise ValueError(f"{fragment} does not match {canonical_name}")
+
+
+def _extract_version_from_fragment(fragment: str, canonical_name: str) -> Optional[str]:
+    """Parse the version string from a <package>+<version> filename
+    "fragment" (stem) or egg fragment.
+
+    :param fragment: The string to parse. E.g. foo-2.1
+    :param canonical_name: The canonicalized name of the package this
+        belongs to.
+    """
+    try:
+        version_start = _find_name_version_sep(fragment, canonical_name) + 1
+    except ValueError:
+        return None
+    version = fragment[version_start:]
+    if not version:
+        return None
+    return version
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/index/sources.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/sources.py
new file mode 100644
index 00000000..3dafb30e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/index/sources.py
@@ -0,0 +1,284 @@
+import logging
+import mimetypes
+import os
+from collections import defaultdict
+from typing import Callable, Dict, Iterable, List, Optional, Tuple
+
+from pip._vendor.packaging.utils import (
+    InvalidSdistFilename,
+    InvalidWheelFilename,
+    canonicalize_name,
+    parse_sdist_filename,
+    parse_wheel_filename,
+)
+
+from pip._internal.models.candidate import InstallationCandidate
+from pip._internal.models.link import Link
+from pip._internal.utils.urls import path_to_url, url_to_path
+from pip._internal.vcs import is_url
+
+logger = logging.getLogger(__name__)
+
+FoundCandidates = Iterable[InstallationCandidate]
+FoundLinks = Iterable[Link]
+CandidatesFromPage = Callable[[Link], Iterable[InstallationCandidate]]
+PageValidator = Callable[[Link], bool]
+
+
+class LinkSource:
+    @property
+    def link(self) -> Optional[Link]:
+        """Returns the underlying link, if there's one."""
+        raise NotImplementedError()
+
+    def page_candidates(self) -> FoundCandidates:
+        """Candidates found by parsing an archive listing HTML file."""
+        raise NotImplementedError()
+
+    def file_links(self) -> FoundLinks:
+        """Links found by specifying archives directly."""
+        raise NotImplementedError()
+
+
+def _is_html_file(file_url: str) -> bool:
+    return mimetypes.guess_type(file_url, strict=False)[0] == "text/html"
+
+
+class _FlatDirectoryToUrls:
+    """Scans directory and caches results"""
+
+    def __init__(self, path: str) -> None:
+        self._path = path
+        self._page_candidates: List[str] = []
+        self._project_name_to_urls: Dict[str, List[str]] = defaultdict(list)
+        self._scanned_directory = False
+
+    def _scan_directory(self) -> None:
+        """Scans directory once and populates both page_candidates
+        and project_name_to_urls at the same time
+        """
+        for entry in os.scandir(self._path):
+            url = path_to_url(entry.path)
+            if _is_html_file(url):
+                self._page_candidates.append(url)
+                continue
+
+            # File must have a valid wheel or sdist name,
+            # otherwise not worth considering as a package
+            try:
+                project_filename = parse_wheel_filename(entry.name)[0]
+            except InvalidWheelFilename:
+                try:
+                    project_filename = parse_sdist_filename(entry.name)[0]
+                except InvalidSdistFilename:
+                    continue
+
+            self._project_name_to_urls[project_filename].append(url)
+        self._scanned_directory = True
+
+    @property
+    def page_candidates(self) -> List[str]:
+        if not self._scanned_directory:
+            self._scan_directory()
+
+        return self._page_candidates
+
+    @property
+    def project_name_to_urls(self) -> Dict[str, List[str]]:
+        if not self._scanned_directory:
+            self._scan_directory()
+
+        return self._project_name_to_urls
+
+
+class _FlatDirectorySource(LinkSource):
+    """Link source specified by ``--find-links=<path-to-dir>``.
+
+    This looks the content of the directory, and returns:
+
+    * ``page_candidates``: Links listed on each HTML file in the directory.
+    * ``file_candidates``: Archives in the directory.
+    """
+
+    _paths_to_urls: Dict[str, _FlatDirectoryToUrls] = {}
+
+    def __init__(
+        self,
+        candidates_from_page: CandidatesFromPage,
+        path: str,
+        project_name: str,
+    ) -> None:
+        self._candidates_from_page = candidates_from_page
+        self._project_name = canonicalize_name(project_name)
+
+        # Get existing instance of _FlatDirectoryToUrls if it exists
+        if path in self._paths_to_urls:
+            self._path_to_urls = self._paths_to_urls[path]
+        else:
+            self._path_to_urls = _FlatDirectoryToUrls(path=path)
+            self._paths_to_urls[path] = self._path_to_urls
+
+    @property
+    def link(self) -> Optional[Link]:
+        return None
+
+    def page_candidates(self) -> FoundCandidates:
+        for url in self._path_to_urls.page_candidates:
+            yield from self._candidates_from_page(Link(url))
+
+    def file_links(self) -> FoundLinks:
+        for url in self._path_to_urls.project_name_to_urls[self._project_name]:
+            yield Link(url)
+
+
+class _LocalFileSource(LinkSource):
+    """``--find-links=<path-or-url>`` or ``--[extra-]index-url=<path-or-url>``.
+
+    If a URL is supplied, it must be a ``file:`` URL. If a path is supplied to
+    the option, it is converted to a URL first. This returns:
+
+    * ``page_candidates``: Links listed on an HTML file.
+    * ``file_candidates``: The non-HTML file.
+    """
+
+    def __init__(
+        self,
+        candidates_from_page: CandidatesFromPage,
+        link: Link,
+    ) -> None:
+        self._candidates_from_page = candidates_from_page
+        self._link = link
+
+    @property
+    def link(self) -> Optional[Link]:
+        return self._link
+
+    def page_candidates(self) -> FoundCandidates:
+        if not _is_html_file(self._link.url):
+            return
+        yield from self._candidates_from_page(self._link)
+
+    def file_links(self) -> FoundLinks:
+        if _is_html_file(self._link.url):
+            return
+        yield self._link
+
+
+class _RemoteFileSource(LinkSource):
+    """``--find-links=<url>`` or ``--[extra-]index-url=<url>``.
+
+    This returns:
+
+    * ``page_candidates``: Links listed on an HTML file.
+    * ``file_candidates``: The non-HTML file.
+    """
+
+    def __init__(
+        self,
+        candidates_from_page: CandidatesFromPage,
+        page_validator: PageValidator,
+        link: Link,
+    ) -> None:
+        self._candidates_from_page = candidates_from_page
+        self._page_validator = page_validator
+        self._link = link
+
+    @property
+    def link(self) -> Optional[Link]:
+        return self._link
+
+    def page_candidates(self) -> FoundCandidates:
+        if not self._page_validator(self._link):
+            return
+        yield from self._candidates_from_page(self._link)
+
+    def file_links(self) -> FoundLinks:
+        yield self._link
+
+
+class _IndexDirectorySource(LinkSource):
+    """``--[extra-]index-url=<path-to-directory>``.
+
+    This is treated like a remote URL; ``candidates_from_page`` contains logic
+    for this by appending ``index.html`` to the link.
+    """
+
+    def __init__(
+        self,
+        candidates_from_page: CandidatesFromPage,
+        link: Link,
+    ) -> None:
+        self._candidates_from_page = candidates_from_page
+        self._link = link
+
+    @property
+    def link(self) -> Optional[Link]:
+        return self._link
+
+    def page_candidates(self) -> FoundCandidates:
+        yield from self._candidates_from_page(self._link)
+
+    def file_links(self) -> FoundLinks:
+        return ()
+
+
+def build_source(
+    location: str,
+    *,
+    candidates_from_page: CandidatesFromPage,
+    page_validator: PageValidator,
+    expand_dir: bool,
+    cache_link_parsing: bool,
+    project_name: str,
+) -> Tuple[Optional[str], Optional[LinkSource]]:
+    path: Optional[str] = None
+    url: Optional[str] = None
+    if os.path.exists(location):  # Is a local path.
+        url = path_to_url(location)
+        path = location
+    elif location.startswith("file:"):  # A file: URL.
+        url = location
+        path = url_to_path(location)
+    elif is_url(location):
+        url = location
+
+    if url is None:
+        msg = (
+            "Location '%s' is ignored: "
+            "it is either a non-existing path or lacks a specific scheme."
+        )
+        logger.warning(msg, location)
+        return (None, None)
+
+    if path is None:
+        source: LinkSource = _RemoteFileSource(
+            candidates_from_page=candidates_from_page,
+            page_validator=page_validator,
+            link=Link(url, cache_link_parsing=cache_link_parsing),
+        )
+        return (url, source)
+
+    if os.path.isdir(path):
+        if expand_dir:
+            source = _FlatDirectorySource(
+                candidates_from_page=candidates_from_page,
+                path=path,
+                project_name=project_name,
+            )
+        else:
+            source = _IndexDirectorySource(
+                candidates_from_page=candidates_from_page,
+                link=Link(url, cache_link_parsing=cache_link_parsing),
+            )
+        return (url, source)
+    elif os.path.isfile(path):
+        source = _LocalFileSource(
+            candidates_from_page=candidates_from_page,
+            link=Link(url, cache_link_parsing=cache_link_parsing),
+        )
+        return (url, source)
+    logger.warning(
+        "Location '%s' is ignored: it is neither a file nor a directory.",
+        location,
+    )
+    return (url, None)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__init__.py
new file mode 100644
index 00000000..32382be7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__init__.py
@@ -0,0 +1,456 @@
+import functools
+import logging
+import os
+import pathlib
+import sys
+import sysconfig
+from typing import Any, Dict, Generator, Optional, Tuple
+
+from pip._internal.models.scheme import SCHEME_KEYS, Scheme
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.deprecation import deprecated
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+from . import _sysconfig
+from .base import (
+    USER_CACHE_DIR,
+    get_major_minor_version,
+    get_src_prefix,
+    is_osx_framework,
+    site_packages,
+    user_site,
+)
+
+__all__ = [
+    "USER_CACHE_DIR",
+    "get_bin_prefix",
+    "get_bin_user",
+    "get_major_minor_version",
+    "get_platlib",
+    "get_purelib",
+    "get_scheme",
+    "get_src_prefix",
+    "site_packages",
+    "user_site",
+]
+
+
+logger = logging.getLogger(__name__)
+
+
+_PLATLIBDIR: str = getattr(sys, "platlibdir", "lib")
+
+_USE_SYSCONFIG_DEFAULT = sys.version_info >= (3, 10)
+
+
+def _should_use_sysconfig() -> bool:
+    """This function determines the value of _USE_SYSCONFIG.
+
+    By default, pip uses sysconfig on Python 3.10+.
+    But Python distributors can override this decision by setting:
+        sysconfig._PIP_USE_SYSCONFIG = True / False
+    Rationale in https://github.com/pypa/pip/issues/10647
+
+    This is a function for testability, but should be constant during any one
+    run.
+    """
+    return bool(getattr(sysconfig, "_PIP_USE_SYSCONFIG", _USE_SYSCONFIG_DEFAULT))
+
+
+_USE_SYSCONFIG = _should_use_sysconfig()
+
+if not _USE_SYSCONFIG:
+    # Import distutils lazily to avoid deprecation warnings,
+    # but import it soon enough that it is in memory and available during
+    # a pip reinstall.
+    from . import _distutils
+
+# Be noisy about incompatibilities if this platforms "should" be using
+# sysconfig, but is explicitly opting out and using distutils instead.
+if _USE_SYSCONFIG_DEFAULT and not _USE_SYSCONFIG:
+    _MISMATCH_LEVEL = logging.WARNING
+else:
+    _MISMATCH_LEVEL = logging.DEBUG
+
+
+def _looks_like_bpo_44860() -> bool:
+    """The resolution to bpo-44860 will change this incorrect platlib.
+
+    See <https://bugs.python.org/issue44860>.
+    """
+    from distutils.command.install import INSTALL_SCHEMES
+
+    try:
+        unix_user_platlib = INSTALL_SCHEMES["unix_user"]["platlib"]
+    except KeyError:
+        return False
+    return unix_user_platlib == "$usersite"
+
+
+def _looks_like_red_hat_patched_platlib_purelib(scheme: Dict[str, str]) -> bool:
+    platlib = scheme["platlib"]
+    if "/$platlibdir/" in platlib:
+        platlib = platlib.replace("/$platlibdir/", f"/{_PLATLIBDIR}/")
+    if "/lib64/" not in platlib:
+        return False
+    unpatched = platlib.replace("/lib64/", "/lib/")
+    return unpatched.replace("$platbase/", "$base/") == scheme["purelib"]
+
+
+@functools.lru_cache(maxsize=None)
+def _looks_like_red_hat_lib() -> bool:
+    """Red Hat patches platlib in unix_prefix and unix_home, but not purelib.
+
+    This is the only way I can see to tell a Red Hat-patched Python.
+    """
+    from distutils.command.install import INSTALL_SCHEMES
+
+    return all(
+        k in INSTALL_SCHEMES
+        and _looks_like_red_hat_patched_platlib_purelib(INSTALL_SCHEMES[k])
+        for k in ("unix_prefix", "unix_home")
+    )
+
+
+@functools.lru_cache(maxsize=None)
+def _looks_like_debian_scheme() -> bool:
+    """Debian adds two additional schemes."""
+    from distutils.command.install import INSTALL_SCHEMES
+
+    return "deb_system" in INSTALL_SCHEMES and "unix_local" in INSTALL_SCHEMES
+
+
+@functools.lru_cache(maxsize=None)
+def _looks_like_red_hat_scheme() -> bool:
+    """Red Hat patches ``sys.prefix`` and ``sys.exec_prefix``.
+
+    Red Hat's ``00251-change-user-install-location.patch`` changes the install
+    command's ``prefix`` and ``exec_prefix`` to append ``"/local"``. This is
+    (fortunately?) done quite unconditionally, so we create a default command
+    object without any configuration to detect this.
+    """
+    from distutils.command.install import install
+    from distutils.dist import Distribution
+
+    cmd: Any = install(Distribution())
+    cmd.finalize_options()
+    return (
+        cmd.exec_prefix == f"{os.path.normpath(sys.exec_prefix)}/local"
+        and cmd.prefix == f"{os.path.normpath(sys.prefix)}/local"
+    )
+
+
+@functools.lru_cache(maxsize=None)
+def _looks_like_slackware_scheme() -> bool:
+    """Slackware patches sysconfig but fails to patch distutils and site.
+
+    Slackware changes sysconfig's user scheme to use ``"lib64"`` for the lib
+    path, but does not do the same to the site module.
+    """
+    if user_site is None:  # User-site not available.
+        return False
+    try:
+        paths = sysconfig.get_paths(scheme="posix_user", expand=False)
+    except KeyError:  # User-site not available.
+        return False
+    return "/lib64/" in paths["purelib"] and "/lib64/" not in user_site
+
+
+@functools.lru_cache(maxsize=None)
+def _looks_like_msys2_mingw_scheme() -> bool:
+    """MSYS2 patches distutils and sysconfig to use a UNIX-like scheme.
+
+    However, MSYS2 incorrectly patches sysconfig ``nt`` scheme. The fix is
+    likely going to be included in their 3.10 release, so we ignore the warning.
+    See msys2/MINGW-packages#9319.
+
+    MSYS2 MINGW's patch uses lowercase ``"lib"`` instead of the usual uppercase,
+    and is missing the final ``"site-packages"``.
+    """
+    paths = sysconfig.get_paths("nt", expand=False)
+    return all(
+        "Lib" not in p and "lib" in p and not p.endswith("site-packages")
+        for p in (paths[key] for key in ("platlib", "purelib"))
+    )
+
+
+def _fix_abiflags(parts: Tuple[str]) -> Generator[str, None, None]:
+    ldversion = sysconfig.get_config_var("LDVERSION")
+    abiflags = getattr(sys, "abiflags", None)
+
+    # LDVERSION does not end with sys.abiflags. Just return the path unchanged.
+    if not ldversion or not abiflags or not ldversion.endswith(abiflags):
+        yield from parts
+        return
+
+    # Strip sys.abiflags from LDVERSION-based path components.
+    for part in parts:
+        if part.endswith(ldversion):
+            part = part[: (0 - len(abiflags))]
+        yield part
+
+
+@functools.lru_cache(maxsize=None)
+def _warn_mismatched(old: pathlib.Path, new: pathlib.Path, *, key: str) -> None:
+    issue_url = "https://github.com/pypa/pip/issues/10151"
+    message = (
+        "Value for %s does not match. Please report this to <%s>"
+        "\ndistutils: %s"
+        "\nsysconfig: %s"
+    )
+    logger.log(_MISMATCH_LEVEL, message, key, issue_url, old, new)
+
+
+def _warn_if_mismatch(old: pathlib.Path, new: pathlib.Path, *, key: str) -> bool:
+    if old == new:
+        return False
+    _warn_mismatched(old, new, key=key)
+    return True
+
+
+@functools.lru_cache(maxsize=None)
+def _log_context(
+    *,
+    user: bool = False,
+    home: Optional[str] = None,
+    root: Optional[str] = None,
+    prefix: Optional[str] = None,
+) -> None:
+    parts = [
+        "Additional context:",
+        "user = %r",
+        "home = %r",
+        "root = %r",
+        "prefix = %r",
+    ]
+
+    logger.log(_MISMATCH_LEVEL, "\n".join(parts), user, home, root, prefix)
+
+
+def get_scheme(
+    dist_name: str,
+    user: bool = False,
+    home: Optional[str] = None,
+    root: Optional[str] = None,
+    isolated: bool = False,
+    prefix: Optional[str] = None,
+) -> Scheme:
+    new = _sysconfig.get_scheme(
+        dist_name,
+        user=user,
+        home=home,
+        root=root,
+        isolated=isolated,
+        prefix=prefix,
+    )
+    if _USE_SYSCONFIG:
+        return new
+
+    old = _distutils.get_scheme(
+        dist_name,
+        user=user,
+        home=home,
+        root=root,
+        isolated=isolated,
+        prefix=prefix,
+    )
+
+    warning_contexts = []
+    for k in SCHEME_KEYS:
+        old_v = pathlib.Path(getattr(old, k))
+        new_v = pathlib.Path(getattr(new, k))
+
+        if old_v == new_v:
+            continue
+
+        # distutils incorrectly put PyPy packages under ``site-packages/python``
+        # in the ``posix_home`` scheme, but PyPy devs said they expect the
+        # directory name to be ``pypy`` instead. So we treat this as a bug fix
+        # and not warn about it. See bpo-43307 and python/cpython#24628.
+        skip_pypy_special_case = (
+            sys.implementation.name == "pypy"
+            and home is not None
+            and k in ("platlib", "purelib")
+            and old_v.parent == new_v.parent
+            and old_v.name.startswith("python")
+            and new_v.name.startswith("pypy")
+        )
+        if skip_pypy_special_case:
+            continue
+
+        # sysconfig's ``osx_framework_user`` does not include ``pythonX.Y`` in
+        # the ``include`` value, but distutils's ``headers`` does. We'll let
+        # CPython decide whether this is a bug or feature. See bpo-43948.
+        skip_osx_framework_user_special_case = (
+            user
+            and is_osx_framework()
+            and k == "headers"
+            and old_v.parent.parent == new_v.parent
+            and old_v.parent.name.startswith("python")
+        )
+        if skip_osx_framework_user_special_case:
+            continue
+
+        # On Red Hat and derived Linux distributions, distutils is patched to
+        # use "lib64" instead of "lib" for platlib.
+        if k == "platlib" and _looks_like_red_hat_lib():
+            continue
+
+        # On Python 3.9+, sysconfig's posix_user scheme sets platlib against
+        # sys.platlibdir, but distutils's unix_user incorrectly coninutes
+        # using the same $usersite for both platlib and purelib. This creates a
+        # mismatch when sys.platlibdir is not "lib".
+        skip_bpo_44860 = (
+            user
+            and k == "platlib"
+            and not WINDOWS
+            and sys.version_info >= (3, 9)
+            and _PLATLIBDIR != "lib"
+            and _looks_like_bpo_44860()
+        )
+        if skip_bpo_44860:
+            continue
+
+        # Slackware incorrectly patches posix_user to use lib64 instead of lib,
+        # but not usersite to match the location.
+        skip_slackware_user_scheme = (
+            user
+            and k in ("platlib", "purelib")
+            and not WINDOWS
+            and _looks_like_slackware_scheme()
+        )
+        if skip_slackware_user_scheme:
+            continue
+
+        # Both Debian and Red Hat patch Python to place the system site under
+        # /usr/local instead of /usr. Debian also places lib in dist-packages
+        # instead of site-packages, but the /usr/local check should cover it.
+        skip_linux_system_special_case = (
+            not (user or home or prefix or running_under_virtualenv())
+            and old_v.parts[1:3] == ("usr", "local")
+            and len(new_v.parts) > 1
+            and new_v.parts[1] == "usr"
+            and (len(new_v.parts) < 3 or new_v.parts[2] != "local")
+            and (_looks_like_red_hat_scheme() or _looks_like_debian_scheme())
+        )
+        if skip_linux_system_special_case:
+            continue
+
+        # MSYS2 MINGW's sysconfig patch does not include the "site-packages"
+        # part of the path. This is incorrect and will be fixed in MSYS.
+        skip_msys2_mingw_bug = (
+            WINDOWS and k in ("platlib", "purelib") and _looks_like_msys2_mingw_scheme()
+        )
+        if skip_msys2_mingw_bug:
+            continue
+
+        # CPython's POSIX install script invokes pip (via ensurepip) against the
+        # interpreter located in the source tree, not the install site. This
+        # triggers special logic in sysconfig that's not present in distutils.
+        # https://github.com/python/cpython/blob/8c21941ddaf/Lib/sysconfig.py#L178-L194
+        skip_cpython_build = (
+            sysconfig.is_python_build(check_home=True)
+            and not WINDOWS
+            and k in ("headers", "include", "platinclude")
+        )
+        if skip_cpython_build:
+            continue
+
+        warning_contexts.append((old_v, new_v, f"scheme.{k}"))
+
+    if not warning_contexts:
+        return old
+
+    # Check if this path mismatch is caused by distutils config files. Those
+    # files will no longer work once we switch to sysconfig, so this raises a
+    # deprecation message for them.
+    default_old = _distutils.distutils_scheme(
+        dist_name,
+        user,
+        home,
+        root,
+        isolated,
+        prefix,
+        ignore_config_files=True,
+    )
+    if any(default_old[k] != getattr(old, k) for k in SCHEME_KEYS):
+        deprecated(
+            reason=(
+                "Configuring installation scheme with distutils config files "
+                "is deprecated and will no longer work in the near future. If you "
+                "are using a Homebrew or Linuxbrew Python, please see discussion "
+                "at https://github.com/Homebrew/homebrew-core/issues/76621"
+            ),
+            replacement=None,
+            gone_in=None,
+        )
+        return old
+
+    # Post warnings about this mismatch so user can report them back.
+    for old_v, new_v, key in warning_contexts:
+        _warn_mismatched(old_v, new_v, key=key)
+    _log_context(user=user, home=home, root=root, prefix=prefix)
+
+    return old
+
+
+def get_bin_prefix() -> str:
+    new = _sysconfig.get_bin_prefix()
+    if _USE_SYSCONFIG:
+        return new
+
+    old = _distutils.get_bin_prefix()
+    if _warn_if_mismatch(pathlib.Path(old), pathlib.Path(new), key="bin_prefix"):
+        _log_context()
+    return old
+
+
+def get_bin_user() -> str:
+    return _sysconfig.get_scheme("", user=True).scripts
+
+
+def _looks_like_deb_system_dist_packages(value: str) -> bool:
+    """Check if the value is Debian's APT-controlled dist-packages.
+
+    Debian's ``distutils.sysconfig.get_python_lib()`` implementation returns the
+    default package path controlled by APT, but does not patch ``sysconfig`` to
+    do the same. This is similar to the bug worked around in ``get_scheme()``,
+    but here the default is ``deb_system`` instead of ``unix_local``. Ultimately
+    we can't do anything about this Debian bug, and this detection allows us to
+    skip the warning when needed.
+    """
+    if not _looks_like_debian_scheme():
+        return False
+    if value == "/usr/lib/python3/dist-packages":
+        return True
+    return False
+
+
+def get_purelib() -> str:
+    """Return the default pure-Python lib location."""
+    new = _sysconfig.get_purelib()
+    if _USE_SYSCONFIG:
+        return new
+
+    old = _distutils.get_purelib()
+    if _looks_like_deb_system_dist_packages(old):
+        return old
+    if _warn_if_mismatch(pathlib.Path(old), pathlib.Path(new), key="purelib"):
+        _log_context()
+    return old
+
+
+def get_platlib() -> str:
+    """Return the default platform-shared lib location."""
+    new = _sysconfig.get_platlib()
+    if _USE_SYSCONFIG:
+        return new
+
+    from . import _distutils
+
+    old = _distutils.get_platlib()
+    if _looks_like_deb_system_dist_packages(old):
+        return old
+    if _warn_if_mismatch(pathlib.Path(old), pathlib.Path(new), key="platlib"):
+        _log_context()
+    return old
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..cad9ed64
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_distutils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_distutils.cpython-312.pyc
new file mode 100644
index 00000000..2e3cbb85
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_distutils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_sysconfig.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_sysconfig.cpython-312.pyc
new file mode 100644
index 00000000..761b3f9a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/_sysconfig.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..4b4d24d7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_distutils.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_distutils.py
new file mode 100644
index 00000000..3d856256
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_distutils.py
@@ -0,0 +1,172 @@
+"""Locations where we look for configs, install stuff, etc"""
+
+# The following comment should be removed at some point in the future.
+# mypy: strict-optional=False
+
+# If pip's going to use distutils, it should not be using the copy that setuptools
+# might have injected into the environment. This is done by removing the injected
+# shim, if it's injected.
+#
+# See https://github.com/pypa/pip/issues/8761 for the original discussion and
+# rationale for why this is done within pip.
+try:
+    __import__("_distutils_hack").remove_shim()
+except (ImportError, AttributeError):
+    pass
+
+import logging
+import os
+import sys
+from distutils.cmd import Command as DistutilsCommand
+from distutils.command.install import SCHEME_KEYS
+from distutils.command.install import install as distutils_install_command
+from distutils.sysconfig import get_python_lib
+from typing import Dict, List, Optional, Union
+
+from pip._internal.models.scheme import Scheme
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+from .base import get_major_minor_version
+
+logger = logging.getLogger(__name__)
+
+
+def distutils_scheme(
+    dist_name: str,
+    user: bool = False,
+    home: Optional[str] = None,
+    root: Optional[str] = None,
+    isolated: bool = False,
+    prefix: Optional[str] = None,
+    *,
+    ignore_config_files: bool = False,
+) -> Dict[str, str]:
+    """
+    Return a distutils install scheme
+    """
+    from distutils.dist import Distribution
+
+    dist_args: Dict[str, Union[str, List[str]]] = {"name": dist_name}
+    if isolated:
+        dist_args["script_args"] = ["--no-user-cfg"]
+
+    d = Distribution(dist_args)
+    if not ignore_config_files:
+        try:
+            d.parse_config_files()
+        except UnicodeDecodeError:
+            paths = d.find_config_files()
+            logger.warning(
+                "Ignore distutils configs in %s due to encoding errors.",
+                ", ".join(os.path.basename(p) for p in paths),
+            )
+    obj: Optional[DistutilsCommand] = None
+    obj = d.get_command_obj("install", create=True)
+    assert obj is not None
+    i: distutils_install_command = obj
+    # NOTE: setting user or home has the side-effect of creating the home dir
+    # or user base for installations during finalize_options()
+    # ideally, we'd prefer a scheme class that has no side-effects.
+    assert not (user and prefix), f"user={user} prefix={prefix}"
+    assert not (home and prefix), f"home={home} prefix={prefix}"
+    i.user = user or i.user
+    if user or home:
+        i.prefix = ""
+    i.prefix = prefix or i.prefix
+    i.home = home or i.home
+    i.root = root or i.root
+    i.finalize_options()
+
+    scheme: Dict[str, str] = {}
+    for key in SCHEME_KEYS:
+        scheme[key] = getattr(i, "install_" + key)
+
+    # install_lib specified in setup.cfg should install *everything*
+    # into there (i.e. it takes precedence over both purelib and
+    # platlib).  Note, i.install_lib is *always* set after
+    # finalize_options(); we only want to override here if the user
+    # has explicitly requested it hence going back to the config
+    if "install_lib" in d.get_option_dict("install"):
+        scheme.update({"purelib": i.install_lib, "platlib": i.install_lib})
+
+    if running_under_virtualenv():
+        if home:
+            prefix = home
+        elif user:
+            prefix = i.install_userbase
+        else:
+            prefix = i.prefix
+        scheme["headers"] = os.path.join(
+            prefix,
+            "include",
+            "site",
+            f"python{get_major_minor_version()}",
+            dist_name,
+        )
+
+        if root is not None:
+            path_no_drive = os.path.splitdrive(os.path.abspath(scheme["headers"]))[1]
+            scheme["headers"] = os.path.join(root, path_no_drive[1:])
+
+    return scheme
+
+
+def get_scheme(
+    dist_name: str,
+    user: bool = False,
+    home: Optional[str] = None,
+    root: Optional[str] = None,
+    isolated: bool = False,
+    prefix: Optional[str] = None,
+) -> Scheme:
+    """
+    Get the "scheme" corresponding to the input parameters. The distutils
+    documentation provides the context for the available schemes:
+    https://docs.python.org/3/install/index.html#alternate-installation
+
+    :param dist_name: the name of the package to retrieve the scheme for, used
+        in the headers scheme path
+    :param user: indicates to use the "user" scheme
+    :param home: indicates to use the "home" scheme and provides the base
+        directory for the same
+    :param root: root under which other directories are re-based
+    :param isolated: equivalent to --no-user-cfg, i.e. do not consider
+        ~/.pydistutils.cfg (posix) or ~/pydistutils.cfg (non-posix) for
+        scheme paths
+    :param prefix: indicates to use the "prefix" scheme and provides the
+        base directory for the same
+    """
+    scheme = distutils_scheme(dist_name, user, home, root, isolated, prefix)
+    return Scheme(
+        platlib=scheme["platlib"],
+        purelib=scheme["purelib"],
+        headers=scheme["headers"],
+        scripts=scheme["scripts"],
+        data=scheme["data"],
+    )
+
+
+def get_bin_prefix() -> str:
+    # XXX: In old virtualenv versions, sys.prefix can contain '..' components,
+    # so we need to call normpath to eliminate them.
+    prefix = os.path.normpath(sys.prefix)
+    if WINDOWS:
+        bin_py = os.path.join(prefix, "Scripts")
+        # buildout uses 'bin' on Windows too?
+        if not os.path.exists(bin_py):
+            bin_py = os.path.join(prefix, "bin")
+        return bin_py
+    # Forcing to use /usr/local/bin for standard macOS framework installs
+    # Also log to ~/Library/Logs/ for use with the Console.app log viewer
+    if sys.platform[:6] == "darwin" and prefix[:16] == "/System/Library/":
+        return "/usr/local/bin"
+    return os.path.join(prefix, "bin")
+
+
+def get_purelib() -> str:
+    return get_python_lib(plat_specific=False)
+
+
+def get_platlib() -> str:
+    return get_python_lib(plat_specific=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_sysconfig.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_sysconfig.py
new file mode 100644
index 00000000..ca860ea5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/_sysconfig.py
@@ -0,0 +1,214 @@
+import logging
+import os
+import sys
+import sysconfig
+import typing
+
+from pip._internal.exceptions import InvalidSchemeCombination, UserInstallationInvalid
+from pip._internal.models.scheme import SCHEME_KEYS, Scheme
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+from .base import change_root, get_major_minor_version, is_osx_framework
+
+logger = logging.getLogger(__name__)
+
+
+# Notes on _infer_* functions.
+# Unfortunately ``get_default_scheme()`` didn't exist before 3.10, so there's no
+# way to ask things like "what is the '_prefix' scheme on this platform". These
+# functions try to answer that with some heuristics while accounting for ad-hoc
+# platforms not covered by CPython's default sysconfig implementation. If the
+# ad-hoc implementation does not fully implement sysconfig, we'll fall back to
+# a POSIX scheme.
+
+_AVAILABLE_SCHEMES = set(sysconfig.get_scheme_names())
+
+_PREFERRED_SCHEME_API = getattr(sysconfig, "get_preferred_scheme", None)
+
+
+def _should_use_osx_framework_prefix() -> bool:
+    """Check for Apple's ``osx_framework_library`` scheme.
+
+    Python distributed by Apple's Command Line Tools has this special scheme
+    that's used when:
+
+    * This is a framework build.
+    * We are installing into the system prefix.
+
+    This does not account for ``pip install --prefix`` (also means we're not
+    installing to the system prefix), which should use ``posix_prefix``, but
+    logic here means ``_infer_prefix()`` outputs ``osx_framework_library``. But
+    since ``prefix`` is not available for ``sysconfig.get_default_scheme()``,
+    which is the stdlib replacement for ``_infer_prefix()``, presumably Apple
+    wouldn't be able to magically switch between ``osx_framework_library`` and
+    ``posix_prefix``. ``_infer_prefix()`` returning ``osx_framework_library``
+    means its behavior is consistent whether we use the stdlib implementation
+    or our own, and we deal with this special case in ``get_scheme()`` instead.
+    """
+    return (
+        "osx_framework_library" in _AVAILABLE_SCHEMES
+        and not running_under_virtualenv()
+        and is_osx_framework()
+    )
+
+
+def _infer_prefix() -> str:
+    """Try to find a prefix scheme for the current platform.
+
+    This tries:
+
+    * A special ``osx_framework_library`` for Python distributed by Apple's
+      Command Line Tools, when not running in a virtual environment.
+    * Implementation + OS, used by PyPy on Windows (``pypy_nt``).
+    * Implementation without OS, used by PyPy on POSIX (``pypy``).
+    * OS + "prefix", used by CPython on POSIX (``posix_prefix``).
+    * Just the OS name, used by CPython on Windows (``nt``).
+
+    If none of the above works, fall back to ``posix_prefix``.
+    """
+    if _PREFERRED_SCHEME_API:
+        return _PREFERRED_SCHEME_API("prefix")
+    if _should_use_osx_framework_prefix():
+        return "osx_framework_library"
+    implementation_suffixed = f"{sys.implementation.name}_{os.name}"
+    if implementation_suffixed in _AVAILABLE_SCHEMES:
+        return implementation_suffixed
+    if sys.implementation.name in _AVAILABLE_SCHEMES:
+        return sys.implementation.name
+    suffixed = f"{os.name}_prefix"
+    if suffixed in _AVAILABLE_SCHEMES:
+        return suffixed
+    if os.name in _AVAILABLE_SCHEMES:  # On Windows, prefx is just called "nt".
+        return os.name
+    return "posix_prefix"
+
+
+def _infer_user() -> str:
+    """Try to find a user scheme for the current platform."""
+    if _PREFERRED_SCHEME_API:
+        return _PREFERRED_SCHEME_API("user")
+    if is_osx_framework() and not running_under_virtualenv():
+        suffixed = "osx_framework_user"
+    else:
+        suffixed = f"{os.name}_user"
+    if suffixed in _AVAILABLE_SCHEMES:
+        return suffixed
+    if "posix_user" not in _AVAILABLE_SCHEMES:  # User scheme unavailable.
+        raise UserInstallationInvalid()
+    return "posix_user"
+
+
+def _infer_home() -> str:
+    """Try to find a home for the current platform."""
+    if _PREFERRED_SCHEME_API:
+        return _PREFERRED_SCHEME_API("home")
+    suffixed = f"{os.name}_home"
+    if suffixed in _AVAILABLE_SCHEMES:
+        return suffixed
+    return "posix_home"
+
+
+# Update these keys if the user sets a custom home.
+_HOME_KEYS = [
+    "installed_base",
+    "base",
+    "installed_platbase",
+    "platbase",
+    "prefix",
+    "exec_prefix",
+]
+if sysconfig.get_config_var("userbase") is not None:
+    _HOME_KEYS.append("userbase")
+
+
+def get_scheme(
+    dist_name: str,
+    user: bool = False,
+    home: typing.Optional[str] = None,
+    root: typing.Optional[str] = None,
+    isolated: bool = False,
+    prefix: typing.Optional[str] = None,
+) -> Scheme:
+    """
+    Get the "scheme" corresponding to the input parameters.
+
+    :param dist_name: the name of the package to retrieve the scheme for, used
+        in the headers scheme path
+    :param user: indicates to use the "user" scheme
+    :param home: indicates to use the "home" scheme
+    :param root: root under which other directories are re-based
+    :param isolated: ignored, but kept for distutils compatibility (where
+        this controls whether the user-site pydistutils.cfg is honored)
+    :param prefix: indicates to use the "prefix" scheme and provides the
+        base directory for the same
+    """
+    if user and prefix:
+        raise InvalidSchemeCombination("--user", "--prefix")
+    if home and prefix:
+        raise InvalidSchemeCombination("--home", "--prefix")
+
+    if home is not None:
+        scheme_name = _infer_home()
+    elif user:
+        scheme_name = _infer_user()
+    else:
+        scheme_name = _infer_prefix()
+
+    # Special case: When installing into a custom prefix, use posix_prefix
+    # instead of osx_framework_library. See _should_use_osx_framework_prefix()
+    # docstring for details.
+    if prefix is not None and scheme_name == "osx_framework_library":
+        scheme_name = "posix_prefix"
+
+    if home is not None:
+        variables = {k: home for k in _HOME_KEYS}
+    elif prefix is not None:
+        variables = {k: prefix for k in _HOME_KEYS}
+    else:
+        variables = {}
+
+    paths = sysconfig.get_paths(scheme=scheme_name, vars=variables)
+
+    # Logic here is very arbitrary, we're doing it for compatibility, don't ask.
+    # 1. Pip historically uses a special header path in virtual environments.
+    # 2. If the distribution name is not known, distutils uses 'UNKNOWN'. We
+    #    only do the same when not running in a virtual environment because
+    #    pip's historical header path logic (see point 1) did not do this.
+    if running_under_virtualenv():
+        if user:
+            base = variables.get("userbase", sys.prefix)
+        else:
+            base = variables.get("base", sys.prefix)
+        python_xy = f"python{get_major_minor_version()}"
+        paths["include"] = os.path.join(base, "include", "site", python_xy)
+    elif not dist_name:
+        dist_name = "UNKNOWN"
+
+    scheme = Scheme(
+        platlib=paths["platlib"],
+        purelib=paths["purelib"],
+        headers=os.path.join(paths["include"], dist_name),
+        scripts=paths["scripts"],
+        data=paths["data"],
+    )
+    if root is not None:
+        converted_keys = {}
+        for key in SCHEME_KEYS:
+            converted_keys[key] = change_root(root, getattr(scheme, key))
+        scheme = Scheme(**converted_keys)
+    return scheme
+
+
+def get_bin_prefix() -> str:
+    # Forcing to use /usr/local/bin for standard macOS framework installs.
+    if sys.platform[:6] == "darwin" and sys.prefix[:16] == "/System/Library/":
+        return "/usr/local/bin"
+    return sysconfig.get_paths()["scripts"]
+
+
+def get_purelib() -> str:
+    return sysconfig.get_paths()["purelib"]
+
+
+def get_platlib() -> str:
+    return sysconfig.get_paths()["platlib"]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/base.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/base.py
new file mode 100644
index 00000000..3f9f896e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/locations/base.py
@@ -0,0 +1,81 @@
+import functools
+import os
+import site
+import sys
+import sysconfig
+import typing
+
+from pip._internal.exceptions import InstallationError
+from pip._internal.utils import appdirs
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+# Application Directories
+USER_CACHE_DIR = appdirs.user_cache_dir("pip")
+
+# FIXME doesn't account for venv linked to global site-packages
+site_packages: str = sysconfig.get_path("purelib")
+
+
+def get_major_minor_version() -> str:
+    """
+    Return the major-minor version of the current Python as a string, e.g.
+    "3.7" or "3.10".
+    """
+    return "{}.{}".format(*sys.version_info)
+
+
+def change_root(new_root: str, pathname: str) -> str:
+    """Return 'pathname' with 'new_root' prepended.
+
+    If 'pathname' is relative, this is equivalent to os.path.join(new_root, pathname).
+    Otherwise, it requires making 'pathname' relative and then joining the
+    two, which is tricky on DOS/Windows and Mac OS.
+
+    This is borrowed from Python's standard library's distutils module.
+    """
+    if os.name == "posix":
+        if not os.path.isabs(pathname):
+            return os.path.join(new_root, pathname)
+        else:
+            return os.path.join(new_root, pathname[1:])
+
+    elif os.name == "nt":
+        (drive, path) = os.path.splitdrive(pathname)
+        if path[0] == "\\":
+            path = path[1:]
+        return os.path.join(new_root, path)
+
+    else:
+        raise InstallationError(
+            f"Unknown platform: {os.name}\n"
+            "Can not change root path prefix on unknown platform."
+        )
+
+
+def get_src_prefix() -> str:
+    if running_under_virtualenv():
+        src_prefix = os.path.join(sys.prefix, "src")
+    else:
+        # FIXME: keep src in cwd for now (it is not a temporary folder)
+        try:
+            src_prefix = os.path.join(os.getcwd(), "src")
+        except OSError:
+            # In case the current working directory has been renamed or deleted
+            sys.exit("The folder you are executing pip from can no longer be found.")
+
+    # under macOS + virtualenv sys.prefix is not properly resolved
+    # it is something like /path/to/python/bin/..
+    return os.path.abspath(src_prefix)
+
+
+try:
+    # Use getusersitepackages if this is present, as it ensures that the
+    # value is initialised properly.
+    user_site: typing.Optional[str] = site.getusersitepackages()
+except AttributeError:
+    user_site = site.USER_SITE
+
+
+@functools.lru_cache(maxsize=None)
+def is_osx_framework() -> bool:
+    return bool(sysconfig.get_config_var("PYTHONFRAMEWORK"))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/main.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/main.py
new file mode 100644
index 00000000..33c6d24c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/main.py
@@ -0,0 +1,12 @@
+from typing import List, Optional
+
+
+def main(args: Optional[List[str]] = None) -> int:
+    """This is preserved for old console scripts that may still be referencing
+    it.
+
+    For additional details, see https://github.com/pypa/pip/issues/7498.
+    """
+    from pip._internal.utils.entrypoints import _wrapper
+
+    return _wrapper(args)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__init__.py
new file mode 100644
index 00000000..1ea1e7fd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__init__.py
@@ -0,0 +1,128 @@
+import contextlib
+import functools
+import os
+import sys
+from typing import TYPE_CHECKING, List, Optional, Type, cast
+
+from pip._internal.utils.misc import strtobool
+
+from .base import BaseDistribution, BaseEnvironment, FilesystemWheel, MemoryWheel, Wheel
+
+if TYPE_CHECKING:
+    from typing import Literal, Protocol
+else:
+    Protocol = object
+
+__all__ = [
+    "BaseDistribution",
+    "BaseEnvironment",
+    "FilesystemWheel",
+    "MemoryWheel",
+    "Wheel",
+    "get_default_environment",
+    "get_environment",
+    "get_wheel_distribution",
+    "select_backend",
+]
+
+
+def _should_use_importlib_metadata() -> bool:
+    """Whether to use the ``importlib.metadata`` or ``pkg_resources`` backend.
+
+    By default, pip uses ``importlib.metadata`` on Python 3.11+, and
+    ``pkg_resources`` otherwise. This can be overridden by a couple of ways:
+
+    * If environment variable ``_PIP_USE_IMPORTLIB_METADATA`` is set, it
+      dictates whether ``importlib.metadata`` is used, regardless of Python
+      version.
+    * On Python 3.11+, Python distributors can patch ``importlib.metadata``
+      to add a global constant ``_PIP_USE_IMPORTLIB_METADATA = False``. This
+      makes pip use ``pkg_resources`` (unless the user set the aforementioned
+      environment variable to *True*).
+    """
+    with contextlib.suppress(KeyError, ValueError):
+        return bool(strtobool(os.environ["_PIP_USE_IMPORTLIB_METADATA"]))
+    if sys.version_info < (3, 11):
+        return False
+    import importlib.metadata
+
+    return bool(getattr(importlib.metadata, "_PIP_USE_IMPORTLIB_METADATA", True))
+
+
+class Backend(Protocol):
+    NAME: 'Literal["importlib", "pkg_resources"]'
+    Distribution: Type[BaseDistribution]
+    Environment: Type[BaseEnvironment]
+
+
+@functools.lru_cache(maxsize=None)
+def select_backend() -> Backend:
+    if _should_use_importlib_metadata():
+        from . import importlib
+
+        return cast(Backend, importlib)
+    from . import pkg_resources
+
+    return cast(Backend, pkg_resources)
+
+
+def get_default_environment() -> BaseEnvironment:
+    """Get the default representation for the current environment.
+
+    This returns an Environment instance from the chosen backend. The default
+    Environment instance should be built from ``sys.path`` and may use caching
+    to share instance state across calls.
+    """
+    return select_backend().Environment.default()
+
+
+def get_environment(paths: Optional[List[str]]) -> BaseEnvironment:
+    """Get a representation of the environment specified by ``paths``.
+
+    This returns an Environment instance from the chosen backend based on the
+    given import paths. The backend must build a fresh instance representing
+    the state of installed distributions when this function is called.
+    """
+    return select_backend().Environment.from_paths(paths)
+
+
+def get_directory_distribution(directory: str) -> BaseDistribution:
+    """Get the distribution metadata representation in the specified directory.
+
+    This returns a Distribution instance from the chosen backend based on
+    the given on-disk ``.dist-info`` directory.
+    """
+    return select_backend().Distribution.from_directory(directory)
+
+
+def get_wheel_distribution(wheel: Wheel, canonical_name: str) -> BaseDistribution:
+    """Get the representation of the specified wheel's distribution metadata.
+
+    This returns a Distribution instance from the chosen backend based on
+    the given wheel's ``.dist-info`` directory.
+
+    :param canonical_name: Normalized project name of the given wheel.
+    """
+    return select_backend().Distribution.from_wheel(wheel, canonical_name)
+
+
+def get_metadata_distribution(
+    metadata_contents: bytes,
+    filename: str,
+    canonical_name: str,
+) -> BaseDistribution:
+    """Get the dist representation of the specified METADATA file contents.
+
+    This returns a Distribution instance from the chosen backend sourced from the data
+    in `metadata_contents`.
+
+    :param metadata_contents: Contents of a METADATA file within a dist, or one served
+                              via PEP 658.
+    :param filename: Filename for the dist this metadata represents.
+    :param canonical_name: Normalized project name of the given dist.
+    """
+    return select_backend().Distribution.from_metadata_file_contents(
+        metadata_contents,
+        filename,
+        canonical_name,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7ff76dcb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/_json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/_json.cpython-312.pyc
new file mode 100644
index 00000000..04705b51
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/_json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..17de1813
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/pkg_resources.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/pkg_resources.cpython-312.pyc
new file mode 100644
index 00000000..d636c9b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/__pycache__/pkg_resources.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/_json.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/_json.py
new file mode 100644
index 00000000..f3aeab32
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/_json.py
@@ -0,0 +1,86 @@
+# Extracted from https://github.com/pfmoore/pkg_metadata
+
+from email.header import Header, decode_header, make_header
+from email.message import Message
+from typing import Any, Dict, List, Union, cast
+
+METADATA_FIELDS = [
+    # Name, Multiple-Use
+    ("Metadata-Version", False),
+    ("Name", False),
+    ("Version", False),
+    ("Dynamic", True),
+    ("Platform", True),
+    ("Supported-Platform", True),
+    ("Summary", False),
+    ("Description", False),
+    ("Description-Content-Type", False),
+    ("Keywords", False),
+    ("Home-page", False),
+    ("Download-URL", False),
+    ("Author", False),
+    ("Author-email", False),
+    ("Maintainer", False),
+    ("Maintainer-email", False),
+    ("License", False),
+    ("License-Expression", False),
+    ("License-File", True),
+    ("Classifier", True),
+    ("Requires-Dist", True),
+    ("Requires-Python", False),
+    ("Requires-External", True),
+    ("Project-URL", True),
+    ("Provides-Extra", True),
+    ("Provides-Dist", True),
+    ("Obsoletes-Dist", True),
+]
+
+
+def json_name(field: str) -> str:
+    return field.lower().replace("-", "_")
+
+
+def msg_to_json(msg: Message) -> Dict[str, Any]:
+    """Convert a Message object into a JSON-compatible dictionary."""
+
+    def sanitise_header(h: Union[Header, str]) -> str:
+        if isinstance(h, Header):
+            chunks = []
+            for bytes, encoding in decode_header(h):
+                if encoding == "unknown-8bit":
+                    try:
+                        # See if UTF-8 works
+                        bytes.decode("utf-8")
+                        encoding = "utf-8"
+                    except UnicodeDecodeError:
+                        # If not, latin1 at least won't fail
+                        encoding = "latin1"
+                chunks.append((bytes, encoding))
+            return str(make_header(chunks))
+        return str(h)
+
+    result = {}
+    for field, multi in METADATA_FIELDS:
+        if field not in msg:
+            continue
+        key = json_name(field)
+        if multi:
+            value: Union[str, List[str]] = [
+                sanitise_header(v) for v in msg.get_all(field)  # type: ignore
+            ]
+        else:
+            value = sanitise_header(msg.get(field))  # type: ignore
+            if key == "keywords":
+                # Accept both comma-separated and space-separated
+                # forms, for better compatibility with old data.
+                if "," in value:
+                    value = [v.strip() for v in value.split(",")]
+                else:
+                    value = value.split()
+        result[key] = value
+
+    payload = cast(str, msg.get_payload())
+    if payload:
+        result["description"] = payload
+
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/base.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/base.py
new file mode 100644
index 00000000..9eabcdb2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/base.py
@@ -0,0 +1,688 @@
+import csv
+import email.message
+import functools
+import json
+import logging
+import pathlib
+import re
+import zipfile
+from typing import (
+    IO,
+    Any,
+    Collection,
+    Container,
+    Dict,
+    Iterable,
+    Iterator,
+    List,
+    NamedTuple,
+    Optional,
+    Protocol,
+    Tuple,
+    Union,
+)
+
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.packaging.specifiers import InvalidSpecifier, SpecifierSet
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import Version
+
+from pip._internal.exceptions import NoneMetadataError
+from pip._internal.locations import site_packages, user_site
+from pip._internal.models.direct_url import (
+    DIRECT_URL_METADATA_NAME,
+    DirectUrl,
+    DirectUrlValidationError,
+)
+from pip._internal.utils.compat import stdlib_pkgs  # TODO: Move definition here.
+from pip._internal.utils.egg_link import egg_link_path_from_sys_path
+from pip._internal.utils.misc import is_local, normalize_path
+from pip._internal.utils.urls import url_to_path
+
+from ._json import msg_to_json
+
+InfoPath = Union[str, pathlib.PurePath]
+
+logger = logging.getLogger(__name__)
+
+
+class BaseEntryPoint(Protocol):
+    @property
+    def name(self) -> str:
+        raise NotImplementedError()
+
+    @property
+    def value(self) -> str:
+        raise NotImplementedError()
+
+    @property
+    def group(self) -> str:
+        raise NotImplementedError()
+
+
+def _convert_installed_files_path(
+    entry: Tuple[str, ...],
+    info: Tuple[str, ...],
+) -> str:
+    """Convert a legacy installed-files.txt path into modern RECORD path.
+
+    The legacy format stores paths relative to the info directory, while the
+    modern format stores paths relative to the package root, e.g. the
+    site-packages directory.
+
+    :param entry: Path parts of the installed-files.txt entry.
+    :param info: Path parts of the egg-info directory relative to package root.
+    :returns: The converted entry.
+
+    For best compatibility with symlinks, this does not use ``abspath()`` or
+    ``Path.resolve()``, but tries to work with path parts:
+
+    1. While ``entry`` starts with ``..``, remove the equal amounts of parts
+       from ``info``; if ``info`` is empty, start appending ``..`` instead.
+    2. Join the two directly.
+    """
+    while entry and entry[0] == "..":
+        if not info or info[-1] == "..":
+            info += ("..",)
+        else:
+            info = info[:-1]
+        entry = entry[1:]
+    return str(pathlib.Path(*info, *entry))
+
+
+class RequiresEntry(NamedTuple):
+    requirement: str
+    extra: str
+    marker: str
+
+
+class BaseDistribution(Protocol):
+    @classmethod
+    def from_directory(cls, directory: str) -> "BaseDistribution":
+        """Load the distribution from a metadata directory.
+
+        :param directory: Path to a metadata directory, e.g. ``.dist-info``.
+        """
+        raise NotImplementedError()
+
+    @classmethod
+    def from_metadata_file_contents(
+        cls,
+        metadata_contents: bytes,
+        filename: str,
+        project_name: str,
+    ) -> "BaseDistribution":
+        """Load the distribution from the contents of a METADATA file.
+
+        This is used to implement PEP 658 by generating a "shallow" dist object that can
+        be used for resolution without downloading or building the actual dist yet.
+
+        :param metadata_contents: The contents of a METADATA file.
+        :param filename: File name for the dist with this metadata.
+        :param project_name: Name of the project this dist represents.
+        """
+        raise NotImplementedError()
+
+    @classmethod
+    def from_wheel(cls, wheel: "Wheel", name: str) -> "BaseDistribution":
+        """Load the distribution from a given wheel.
+
+        :param wheel: A concrete wheel definition.
+        :param name: File name of the wheel.
+
+        :raises InvalidWheel: Whenever loading of the wheel causes a
+            :py:exc:`zipfile.BadZipFile` exception to be thrown.
+        :raises UnsupportedWheel: If the wheel is a valid zip, but malformed
+            internally.
+        """
+        raise NotImplementedError()
+
+    def __repr__(self) -> str:
+        return f"{self.raw_name} {self.raw_version} ({self.location})"
+
+    def __str__(self) -> str:
+        return f"{self.raw_name} {self.raw_version}"
+
+    @property
+    def location(self) -> Optional[str]:
+        """Where the distribution is loaded from.
+
+        A string value is not necessarily a filesystem path, since distributions
+        can be loaded from other sources, e.g. arbitrary zip archives. ``None``
+        means the distribution is created in-memory.
+
+        Do not canonicalize this value with e.g. ``pathlib.Path.resolve()``. If
+        this is a symbolic link, we want to preserve the relative path between
+        it and files in the distribution.
+        """
+        raise NotImplementedError()
+
+    @property
+    def editable_project_location(self) -> Optional[str]:
+        """The project location for editable distributions.
+
+        This is the directory where pyproject.toml or setup.py is located.
+        None if the distribution is not installed in editable mode.
+        """
+        # TODO: this property is relatively costly to compute, memoize it ?
+        direct_url = self.direct_url
+        if direct_url:
+            if direct_url.is_local_editable():
+                return url_to_path(direct_url.url)
+        else:
+            # Search for an .egg-link file by walking sys.path, as it was
+            # done before by dist_is_editable().
+            egg_link_path = egg_link_path_from_sys_path(self.raw_name)
+            if egg_link_path:
+                # TODO: get project location from second line of egg_link file
+                #       (https://github.com/pypa/pip/issues/10243)
+                return self.location
+        return None
+
+    @property
+    def installed_location(self) -> Optional[str]:
+        """The distribution's "installed" location.
+
+        This should generally be a ``site-packages`` directory. This is
+        usually ``dist.location``, except for legacy develop-installed packages,
+        where ``dist.location`` is the source code location, and this is where
+        the ``.egg-link`` file is.
+
+        The returned location is normalized (in particular, with symlinks removed).
+        """
+        raise NotImplementedError()
+
+    @property
+    def info_location(self) -> Optional[str]:
+        """Location of the .[egg|dist]-info directory or file.
+
+        Similarly to ``location``, a string value is not necessarily a
+        filesystem path. ``None`` means the distribution is created in-memory.
+
+        For a modern .dist-info installation on disk, this should be something
+        like ``{location}/{raw_name}-{version}.dist-info``.
+
+        Do not canonicalize this value with e.g. ``pathlib.Path.resolve()``. If
+        this is a symbolic link, we want to preserve the relative path between
+        it and other files in the distribution.
+        """
+        raise NotImplementedError()
+
+    @property
+    def installed_by_distutils(self) -> bool:
+        """Whether this distribution is installed with legacy distutils format.
+
+        A distribution installed with "raw" distutils not patched by setuptools
+        uses one single file at ``info_location`` to store metadata. We need to
+        treat this specially on uninstallation.
+        """
+        info_location = self.info_location
+        if not info_location:
+            return False
+        return pathlib.Path(info_location).is_file()
+
+    @property
+    def installed_as_egg(self) -> bool:
+        """Whether this distribution is installed as an egg.
+
+        This usually indicates the distribution was installed by (older versions
+        of) easy_install.
+        """
+        location = self.location
+        if not location:
+            return False
+        return location.endswith(".egg")
+
+    @property
+    def installed_with_setuptools_egg_info(self) -> bool:
+        """Whether this distribution is installed with the ``.egg-info`` format.
+
+        This usually indicates the distribution was installed with setuptools
+        with an old pip version or with ``single-version-externally-managed``.
+
+        Note that this ensure the metadata store is a directory. distutils can
+        also installs an ``.egg-info``, but as a file, not a directory. This
+        property is *False* for that case. Also see ``installed_by_distutils``.
+        """
+        info_location = self.info_location
+        if not info_location:
+            return False
+        if not info_location.endswith(".egg-info"):
+            return False
+        return pathlib.Path(info_location).is_dir()
+
+    @property
+    def installed_with_dist_info(self) -> bool:
+        """Whether this distribution is installed with the "modern format".
+
+        This indicates a "modern" installation, e.g. storing metadata in the
+        ``.dist-info`` directory. This applies to installations made by
+        setuptools (but through pip, not directly), or anything using the
+        standardized build backend interface (PEP 517).
+        """
+        info_location = self.info_location
+        if not info_location:
+            return False
+        if not info_location.endswith(".dist-info"):
+            return False
+        return pathlib.Path(info_location).is_dir()
+
+    @property
+    def canonical_name(self) -> NormalizedName:
+        raise NotImplementedError()
+
+    @property
+    def version(self) -> Version:
+        raise NotImplementedError()
+
+    @property
+    def raw_version(self) -> str:
+        raise NotImplementedError()
+
+    @property
+    def setuptools_filename(self) -> str:
+        """Convert a project name to its setuptools-compatible filename.
+
+        This is a copy of ``pkg_resources.to_filename()`` for compatibility.
+        """
+        return self.raw_name.replace("-", "_")
+
+    @property
+    def direct_url(self) -> Optional[DirectUrl]:
+        """Obtain a DirectUrl from this distribution.
+
+        Returns None if the distribution has no `direct_url.json` metadata,
+        or if `direct_url.json` is invalid.
+        """
+        try:
+            content = self.read_text(DIRECT_URL_METADATA_NAME)
+        except FileNotFoundError:
+            return None
+        try:
+            return DirectUrl.from_json(content)
+        except (
+            UnicodeDecodeError,
+            json.JSONDecodeError,
+            DirectUrlValidationError,
+        ) as e:
+            logger.warning(
+                "Error parsing %s for %s: %s",
+                DIRECT_URL_METADATA_NAME,
+                self.canonical_name,
+                e,
+            )
+            return None
+
+    @property
+    def installer(self) -> str:
+        try:
+            installer_text = self.read_text("INSTALLER")
+        except (OSError, ValueError, NoneMetadataError):
+            return ""  # Fail silently if the installer file cannot be read.
+        for line in installer_text.splitlines():
+            cleaned_line = line.strip()
+            if cleaned_line:
+                return cleaned_line
+        return ""
+
+    @property
+    def requested(self) -> bool:
+        return self.is_file("REQUESTED")
+
+    @property
+    def editable(self) -> bool:
+        return bool(self.editable_project_location)
+
+    @property
+    def local(self) -> bool:
+        """If distribution is installed in the current virtual environment.
+
+        Always True if we're not in a virtualenv.
+        """
+        if self.installed_location is None:
+            return False
+        return is_local(self.installed_location)
+
+    @property
+    def in_usersite(self) -> bool:
+        if self.installed_location is None or user_site is None:
+            return False
+        return self.installed_location.startswith(normalize_path(user_site))
+
+    @property
+    def in_site_packages(self) -> bool:
+        if self.installed_location is None or site_packages is None:
+            return False
+        return self.installed_location.startswith(normalize_path(site_packages))
+
+    def is_file(self, path: InfoPath) -> bool:
+        """Check whether an entry in the info directory is a file."""
+        raise NotImplementedError()
+
+    def iter_distutils_script_names(self) -> Iterator[str]:
+        """Find distutils 'scripts' entries metadata.
+
+        If 'scripts' is supplied in ``setup.py``, distutils records those in the
+        installed distribution's ``scripts`` directory, a file for each script.
+        """
+        raise NotImplementedError()
+
+    def read_text(self, path: InfoPath) -> str:
+        """Read a file in the info directory.
+
+        :raise FileNotFoundError: If ``path`` does not exist in the directory.
+        :raise NoneMetadataError: If ``path`` exists in the info directory, but
+            cannot be read.
+        """
+        raise NotImplementedError()
+
+    def iter_entry_points(self) -> Iterable[BaseEntryPoint]:
+        raise NotImplementedError()
+
+    def _metadata_impl(self) -> email.message.Message:
+        raise NotImplementedError()
+
+    @functools.cached_property
+    def metadata(self) -> email.message.Message:
+        """Metadata of distribution parsed from e.g. METADATA or PKG-INFO.
+
+        This should return an empty message if the metadata file is unavailable.
+
+        :raises NoneMetadataError: If the metadata file is available, but does
+            not contain valid metadata.
+        """
+        metadata = self._metadata_impl()
+        self._add_egg_info_requires(metadata)
+        return metadata
+
+    @property
+    def metadata_dict(self) -> Dict[str, Any]:
+        """PEP 566 compliant JSON-serializable representation of METADATA or PKG-INFO.
+
+        This should return an empty dict if the metadata file is unavailable.
+
+        :raises NoneMetadataError: If the metadata file is available, but does
+            not contain valid metadata.
+        """
+        return msg_to_json(self.metadata)
+
+    @property
+    def metadata_version(self) -> Optional[str]:
+        """Value of "Metadata-Version:" in distribution metadata, if available."""
+        return self.metadata.get("Metadata-Version")
+
+    @property
+    def raw_name(self) -> str:
+        """Value of "Name:" in distribution metadata."""
+        # The metadata should NEVER be missing the Name: key, but if it somehow
+        # does, fall back to the known canonical name.
+        return self.metadata.get("Name", self.canonical_name)
+
+    @property
+    def requires_python(self) -> SpecifierSet:
+        """Value of "Requires-Python:" in distribution metadata.
+
+        If the key does not exist or contains an invalid value, an empty
+        SpecifierSet should be returned.
+        """
+        value = self.metadata.get("Requires-Python")
+        if value is None:
+            return SpecifierSet()
+        try:
+            # Convert to str to satisfy the type checker; this can be a Header object.
+            spec = SpecifierSet(str(value))
+        except InvalidSpecifier as e:
+            message = "Package %r has an invalid Requires-Python: %s"
+            logger.warning(message, self.raw_name, e)
+            return SpecifierSet()
+        return spec
+
+    def iter_dependencies(self, extras: Collection[str] = ()) -> Iterable[Requirement]:
+        """Dependencies of this distribution.
+
+        For modern .dist-info distributions, this is the collection of
+        "Requires-Dist:" entries in distribution metadata.
+        """
+        raise NotImplementedError()
+
+    def iter_raw_dependencies(self) -> Iterable[str]:
+        """Raw Requires-Dist metadata."""
+        return self.metadata.get_all("Requires-Dist", [])
+
+    def iter_provided_extras(self) -> Iterable[NormalizedName]:
+        """Extras provided by this distribution.
+
+        For modern .dist-info distributions, this is the collection of
+        "Provides-Extra:" entries in distribution metadata.
+
+        The return value of this function is expected to be normalised names,
+        per PEP 685, with the returned value being handled appropriately by
+        `iter_dependencies`.
+        """
+        raise NotImplementedError()
+
+    def _iter_declared_entries_from_record(self) -> Optional[Iterator[str]]:
+        try:
+            text = self.read_text("RECORD")
+        except FileNotFoundError:
+            return None
+        # This extra Path-str cast normalizes entries.
+        return (str(pathlib.Path(row[0])) for row in csv.reader(text.splitlines()))
+
+    def _iter_declared_entries_from_legacy(self) -> Optional[Iterator[str]]:
+        try:
+            text = self.read_text("installed-files.txt")
+        except FileNotFoundError:
+            return None
+        paths = (p for p in text.splitlines(keepends=False) if p)
+        root = self.location
+        info = self.info_location
+        if root is None or info is None:
+            return paths
+        try:
+            info_rel = pathlib.Path(info).relative_to(root)
+        except ValueError:  # info is not relative to root.
+            return paths
+        if not info_rel.parts:  # info *is* root.
+            return paths
+        return (
+            _convert_installed_files_path(pathlib.Path(p).parts, info_rel.parts)
+            for p in paths
+        )
+
+    def iter_declared_entries(self) -> Optional[Iterator[str]]:
+        """Iterate through file entries declared in this distribution.
+
+        For modern .dist-info distributions, this is the files listed in the
+        ``RECORD`` metadata file. For legacy setuptools distributions, this
+        comes from ``installed-files.txt``, with entries normalized to be
+        compatible with the format used by ``RECORD``.
+
+        :return: An iterator for listed entries, or None if the distribution
+            contains neither ``RECORD`` nor ``installed-files.txt``.
+        """
+        return (
+            self._iter_declared_entries_from_record()
+            or self._iter_declared_entries_from_legacy()
+        )
+
+    def _iter_requires_txt_entries(self) -> Iterator[RequiresEntry]:
+        """Parse a ``requires.txt`` in an egg-info directory.
+
+        This is an INI-ish format where an egg-info stores dependencies. A
+        section name describes extra other environment markers, while each entry
+        is an arbitrary string (not a key-value pair) representing a dependency
+        as a requirement string (no markers).
+
+        There is a construct in ``importlib.metadata`` called ``Sectioned`` that
+        does mostly the same, but the format is currently considered private.
+        """
+        try:
+            content = self.read_text("requires.txt")
+        except FileNotFoundError:
+            return
+        extra = marker = ""  # Section-less entries don't have markers.
+        for line in content.splitlines():
+            line = line.strip()
+            if not line or line.startswith("#"):  # Comment; ignored.
+                continue
+            if line.startswith("[") and line.endswith("]"):  # A section header.
+                extra, _, marker = line.strip("[]").partition(":")
+                continue
+            yield RequiresEntry(requirement=line, extra=extra, marker=marker)
+
+    def _iter_egg_info_extras(self) -> Iterable[str]:
+        """Get extras from the egg-info directory."""
+        known_extras = {""}
+        for entry in self._iter_requires_txt_entries():
+            extra = canonicalize_name(entry.extra)
+            if extra in known_extras:
+                continue
+            known_extras.add(extra)
+            yield extra
+
+    def _iter_egg_info_dependencies(self) -> Iterable[str]:
+        """Get distribution dependencies from the egg-info directory.
+
+        To ease parsing, this converts a legacy dependency entry into a PEP 508
+        requirement string. Like ``_iter_requires_txt_entries()``, there is code
+        in ``importlib.metadata`` that does mostly the same, but not do exactly
+        what we need.
+
+        Namely, ``importlib.metadata`` does not normalize the extra name before
+        putting it into the requirement string, which causes marker comparison
+        to fail because the dist-info format do normalize. This is consistent in
+        all currently available PEP 517 backends, although not standardized.
+        """
+        for entry in self._iter_requires_txt_entries():
+            extra = canonicalize_name(entry.extra)
+            if extra and entry.marker:
+                marker = f'({entry.marker}) and extra == "{extra}"'
+            elif extra:
+                marker = f'extra == "{extra}"'
+            elif entry.marker:
+                marker = entry.marker
+            else:
+                marker = ""
+            if marker:
+                yield f"{entry.requirement} ; {marker}"
+            else:
+                yield entry.requirement
+
+    def _add_egg_info_requires(self, metadata: email.message.Message) -> None:
+        """Add egg-info requires.txt information to the metadata."""
+        if not metadata.get_all("Requires-Dist"):
+            for dep in self._iter_egg_info_dependencies():
+                metadata["Requires-Dist"] = dep
+        if not metadata.get_all("Provides-Extra"):
+            for extra in self._iter_egg_info_extras():
+                metadata["Provides-Extra"] = extra
+
+
+class BaseEnvironment:
+    """An environment containing distributions to introspect."""
+
+    @classmethod
+    def default(cls) -> "BaseEnvironment":
+        raise NotImplementedError()
+
+    @classmethod
+    def from_paths(cls, paths: Optional[List[str]]) -> "BaseEnvironment":
+        raise NotImplementedError()
+
+    def get_distribution(self, name: str) -> Optional["BaseDistribution"]:
+        """Given a requirement name, return the installed distributions.
+
+        The name may not be normalized. The implementation must canonicalize
+        it for lookup.
+        """
+        raise NotImplementedError()
+
+    def _iter_distributions(self) -> Iterator["BaseDistribution"]:
+        """Iterate through installed distributions.
+
+        This function should be implemented by subclass, but never called
+        directly. Use the public ``iter_distribution()`` instead, which
+        implements additional logic to make sure the distributions are valid.
+        """
+        raise NotImplementedError()
+
+    def iter_all_distributions(self) -> Iterator[BaseDistribution]:
+        """Iterate through all installed distributions without any filtering."""
+        for dist in self._iter_distributions():
+            # Make sure the distribution actually comes from a valid Python
+            # packaging distribution. Pip's AdjacentTempDirectory leaves folders
+            # e.g. ``~atplotlib.dist-info`` if cleanup was interrupted. The
+            # valid project name pattern is taken from PEP 508.
+            project_name_valid = re.match(
+                r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$",
+                dist.canonical_name,
+                flags=re.IGNORECASE,
+            )
+            if not project_name_valid:
+                logger.warning(
+                    "Ignoring invalid distribution %s (%s)",
+                    dist.canonical_name,
+                    dist.location,
+                )
+                continue
+            yield dist
+
+    def iter_installed_distributions(
+        self,
+        local_only: bool = True,
+        skip: Container[str] = stdlib_pkgs,
+        include_editables: bool = True,
+        editables_only: bool = False,
+        user_only: bool = False,
+    ) -> Iterator[BaseDistribution]:
+        """Return a list of installed distributions.
+
+        This is based on ``iter_all_distributions()`` with additional filtering
+        options. Note that ``iter_installed_distributions()`` without arguments
+        is *not* equal to ``iter_all_distributions()``, since some of the
+        configurations exclude packages by default.
+
+        :param local_only: If True (default), only return installations
+        local to the current virtualenv, if in a virtualenv.
+        :param skip: An iterable of canonicalized project names to ignore;
+            defaults to ``stdlib_pkgs``.
+        :param include_editables: If False, don't report editables.
+        :param editables_only: If True, only report editables.
+        :param user_only: If True, only report installations in the user
+        site directory.
+        """
+        it = self.iter_all_distributions()
+        if local_only:
+            it = (d for d in it if d.local)
+        if not include_editables:
+            it = (d for d in it if not d.editable)
+        if editables_only:
+            it = (d for d in it if d.editable)
+        if user_only:
+            it = (d for d in it if d.in_usersite)
+        return (d for d in it if d.canonical_name not in skip)
+
+
+class Wheel(Protocol):
+    location: str
+
+    def as_zipfile(self) -> zipfile.ZipFile:
+        raise NotImplementedError()
+
+
+class FilesystemWheel(Wheel):
+    def __init__(self, location: str) -> None:
+        self.location = location
+
+    def as_zipfile(self) -> zipfile.ZipFile:
+        return zipfile.ZipFile(self.location, allowZip64=True)
+
+
+class MemoryWheel(Wheel):
+    def __init__(self, location: str, stream: IO[bytes]) -> None:
+        self.location = location
+        self.stream = stream
+
+    def as_zipfile(self) -> zipfile.ZipFile:
+        return zipfile.ZipFile(self.stream, allowZip64=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__init__.py
new file mode 100644
index 00000000..a779138d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__init__.py
@@ -0,0 +1,6 @@
+from ._dists import Distribution
+from ._envs import Environment
+
+__all__ = ["NAME", "Distribution", "Environment"]
+
+NAME = "importlib"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e2036e60
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_compat.cpython-312.pyc
new file mode 100644
index 00000000..a19149c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_dists.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_dists.cpython-312.pyc
new file mode 100644
index 00000000..5df3e1fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_dists.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_envs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_envs.cpython-312.pyc
new file mode 100644
index 00000000..6aedf1ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/__pycache__/_envs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_compat.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_compat.py
new file mode 100644
index 00000000..ec1e815c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_compat.py
@@ -0,0 +1,85 @@
+import importlib.metadata
+import os
+from typing import Any, Optional, Protocol, Tuple, cast
+
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+
+
+class BadMetadata(ValueError):
+    def __init__(self, dist: importlib.metadata.Distribution, *, reason: str) -> None:
+        self.dist = dist
+        self.reason = reason
+
+    def __str__(self) -> str:
+        return f"Bad metadata in {self.dist} ({self.reason})"
+
+
+class BasePath(Protocol):
+    """A protocol that various path objects conform.
+
+    This exists because importlib.metadata uses both ``pathlib.Path`` and
+    ``zipfile.Path``, and we need a common base for type hints (Union does not
+    work well since ``zipfile.Path`` is too new for our linter setup).
+
+    This does not mean to be exhaustive, but only contains things that present
+    in both classes *that we need*.
+    """
+
+    @property
+    def name(self) -> str:
+        raise NotImplementedError()
+
+    @property
+    def parent(self) -> "BasePath":
+        raise NotImplementedError()
+
+
+def get_info_location(d: importlib.metadata.Distribution) -> Optional[BasePath]:
+    """Find the path to the distribution's metadata directory.
+
+    HACK: This relies on importlib.metadata's private ``_path`` attribute. Not
+    all distributions exist on disk, so importlib.metadata is correct to not
+    expose the attribute as public. But pip's code base is old and not as clean,
+    so we do this to avoid having to rewrite too many things. Hopefully we can
+    eliminate this some day.
+    """
+    return getattr(d, "_path", None)
+
+
+def parse_name_and_version_from_info_directory(
+    dist: importlib.metadata.Distribution,
+) -> Tuple[Optional[str], Optional[str]]:
+    """Get a name and version from the metadata directory name.
+
+    This is much faster than reading distribution metadata.
+    """
+    info_location = get_info_location(dist)
+    if info_location is None:
+        return None, None
+
+    stem, suffix = os.path.splitext(info_location.name)
+    if suffix == ".dist-info":
+        name, sep, version = stem.partition("-")
+        if sep:
+            return name, version
+
+    if suffix == ".egg-info":
+        name = stem.split("-", 1)[0]
+        return name, None
+
+    return None, None
+
+
+def get_dist_canonical_name(dist: importlib.metadata.Distribution) -> NormalizedName:
+    """Get the distribution's normalized name.
+
+    The ``name`` attribute is only available in Python 3.10 or later. We are
+    targeting exactly that, but Mypy does not know this.
+    """
+    if name := parse_name_and_version_from_info_directory(dist)[0]:
+        return canonicalize_name(name)
+
+    name = cast(Any, dist).name
+    if not isinstance(name, str):
+        raise BadMetadata(dist, reason="invalid metadata entry 'name'")
+    return canonicalize_name(name)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_dists.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_dists.py
new file mode 100644
index 00000000..d220b616
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_dists.py
@@ -0,0 +1,228 @@
+import email.message
+import importlib.metadata
+import pathlib
+import zipfile
+from os import PathLike
+from typing import (
+    Collection,
+    Dict,
+    Iterable,
+    Iterator,
+    Mapping,
+    Optional,
+    Sequence,
+    Union,
+    cast,
+)
+
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import Version
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.exceptions import InvalidWheel, UnsupportedWheel
+from pip._internal.metadata.base import (
+    BaseDistribution,
+    BaseEntryPoint,
+    InfoPath,
+    Wheel,
+)
+from pip._internal.utils.misc import normalize_path
+from pip._internal.utils.packaging import get_requirement
+from pip._internal.utils.temp_dir import TempDirectory
+from pip._internal.utils.wheel import parse_wheel, read_wheel_metadata_file
+
+from ._compat import (
+    BasePath,
+    get_dist_canonical_name,
+    parse_name_and_version_from_info_directory,
+)
+
+
+class WheelDistribution(importlib.metadata.Distribution):
+    """An ``importlib.metadata.Distribution`` read from a wheel.
+
+    Although ``importlib.metadata.PathDistribution`` accepts ``zipfile.Path``,
+    its implementation is too "lazy" for pip's needs (we can't keep the ZipFile
+    handle open for the entire lifetime of the distribution object).
+
+    This implementation eagerly reads the entire metadata directory into the
+    memory instead, and operates from that.
+    """
+
+    def __init__(
+        self,
+        files: Mapping[pathlib.PurePosixPath, bytes],
+        info_location: pathlib.PurePosixPath,
+    ) -> None:
+        self._files = files
+        self.info_location = info_location
+
+    @classmethod
+    def from_zipfile(
+        cls,
+        zf: zipfile.ZipFile,
+        name: str,
+        location: str,
+    ) -> "WheelDistribution":
+        info_dir, _ = parse_wheel(zf, name)
+        paths = (
+            (name, pathlib.PurePosixPath(name.split("/", 1)[-1]))
+            for name in zf.namelist()
+            if name.startswith(f"{info_dir}/")
+        )
+        files = {
+            relpath: read_wheel_metadata_file(zf, fullpath)
+            for fullpath, relpath in paths
+        }
+        info_location = pathlib.PurePosixPath(location, info_dir)
+        return cls(files, info_location)
+
+    def iterdir(self, path: InfoPath) -> Iterator[pathlib.PurePosixPath]:
+        # Only allow iterating through the metadata directory.
+        if pathlib.PurePosixPath(str(path)) in self._files:
+            return iter(self._files)
+        raise FileNotFoundError(path)
+
+    def read_text(self, filename: str) -> Optional[str]:
+        try:
+            data = self._files[pathlib.PurePosixPath(filename)]
+        except KeyError:
+            return None
+        try:
+            text = data.decode("utf-8")
+        except UnicodeDecodeError as e:
+            wheel = self.info_location.parent
+            error = f"Error decoding metadata for {wheel}: {e} in {filename} file"
+            raise UnsupportedWheel(error)
+        return text
+
+    def locate_file(self, path: Union[str, "PathLike[str]"]) -> pathlib.Path:
+        # This method doesn't make sense for our in-memory wheel, but the API
+        # requires us to define it.
+        raise NotImplementedError
+
+
+class Distribution(BaseDistribution):
+    def __init__(
+        self,
+        dist: importlib.metadata.Distribution,
+        info_location: Optional[BasePath],
+        installed_location: Optional[BasePath],
+    ) -> None:
+        self._dist = dist
+        self._info_location = info_location
+        self._installed_location = installed_location
+
+    @classmethod
+    def from_directory(cls, directory: str) -> BaseDistribution:
+        info_location = pathlib.Path(directory)
+        dist = importlib.metadata.Distribution.at(info_location)
+        return cls(dist, info_location, info_location.parent)
+
+    @classmethod
+    def from_metadata_file_contents(
+        cls,
+        metadata_contents: bytes,
+        filename: str,
+        project_name: str,
+    ) -> BaseDistribution:
+        # Generate temp dir to contain the metadata file, and write the file contents.
+        temp_dir = pathlib.Path(
+            TempDirectory(kind="metadata", globally_managed=True).path
+        )
+        metadata_path = temp_dir / "METADATA"
+        metadata_path.write_bytes(metadata_contents)
+        # Construct dist pointing to the newly created directory.
+        dist = importlib.metadata.Distribution.at(metadata_path.parent)
+        return cls(dist, metadata_path.parent, None)
+
+    @classmethod
+    def from_wheel(cls, wheel: Wheel, name: str) -> BaseDistribution:
+        try:
+            with wheel.as_zipfile() as zf:
+                dist = WheelDistribution.from_zipfile(zf, name, wheel.location)
+        except zipfile.BadZipFile as e:
+            raise InvalidWheel(wheel.location, name) from e
+        return cls(dist, dist.info_location, pathlib.PurePosixPath(wheel.location))
+
+    @property
+    def location(self) -> Optional[str]:
+        if self._info_location is None:
+            return None
+        return str(self._info_location.parent)
+
+    @property
+    def info_location(self) -> Optional[str]:
+        if self._info_location is None:
+            return None
+        return str(self._info_location)
+
+    @property
+    def installed_location(self) -> Optional[str]:
+        if self._installed_location is None:
+            return None
+        return normalize_path(str(self._installed_location))
+
+    @property
+    def canonical_name(self) -> NormalizedName:
+        return get_dist_canonical_name(self._dist)
+
+    @property
+    def version(self) -> Version:
+        if version := parse_name_and_version_from_info_directory(self._dist)[1]:
+            return parse_version(version)
+        return parse_version(self._dist.version)
+
+    @property
+    def raw_version(self) -> str:
+        return self._dist.version
+
+    def is_file(self, path: InfoPath) -> bool:
+        return self._dist.read_text(str(path)) is not None
+
+    def iter_distutils_script_names(self) -> Iterator[str]:
+        # A distutils installation is always "flat" (not in e.g. egg form), so
+        # if this distribution's info location is NOT a pathlib.Path (but e.g.
+        # zipfile.Path), it can never contain any distutils scripts.
+        if not isinstance(self._info_location, pathlib.Path):
+            return
+        for child in self._info_location.joinpath("scripts").iterdir():
+            yield child.name
+
+    def read_text(self, path: InfoPath) -> str:
+        content = self._dist.read_text(str(path))
+        if content is None:
+            raise FileNotFoundError(path)
+        return content
+
+    def iter_entry_points(self) -> Iterable[BaseEntryPoint]:
+        # importlib.metadata's EntryPoint structure satisfies BaseEntryPoint.
+        return self._dist.entry_points
+
+    def _metadata_impl(self) -> email.message.Message:
+        # From Python 3.10+, importlib.metadata declares PackageMetadata as the
+        # return type. This protocol is unfortunately a disaster now and misses
+        # a ton of fields that we need, including get() and get_payload(). We
+        # rely on the implementation that the object is actually a Message now,
+        # until upstream can improve the protocol. (python/cpython#94952)
+        return cast(email.message.Message, self._dist.metadata)
+
+    def iter_provided_extras(self) -> Iterable[NormalizedName]:
+        return [
+            canonicalize_name(extra)
+            for extra in self.metadata.get_all("Provides-Extra", [])
+        ]
+
+    def iter_dependencies(self, extras: Collection[str] = ()) -> Iterable[Requirement]:
+        contexts: Sequence[Dict[str, str]] = [{"extra": e} for e in extras]
+        for req_string in self.metadata.get_all("Requires-Dist", []):
+            # strip() because email.message.Message.get_all() may return a leading \n
+            # in case a long header was wrapped.
+            req = get_requirement(req_string.strip())
+            if not req.marker:
+                yield req
+            elif not extras and req.marker.evaluate({"extra": ""}):
+                yield req
+            elif any(req.marker.evaluate(context) for context in contexts):
+                yield req
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_envs.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_envs.py
new file mode 100644
index 00000000..4d906fd3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/importlib/_envs.py
@@ -0,0 +1,189 @@
+import functools
+import importlib.metadata
+import logging
+import os
+import pathlib
+import sys
+import zipfile
+import zipimport
+from typing import Iterator, List, Optional, Sequence, Set, Tuple
+
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+
+from pip._internal.metadata.base import BaseDistribution, BaseEnvironment
+from pip._internal.models.wheel import Wheel
+from pip._internal.utils.deprecation import deprecated
+from pip._internal.utils.filetypes import WHEEL_EXTENSION
+
+from ._compat import BadMetadata, BasePath, get_dist_canonical_name, get_info_location
+from ._dists import Distribution
+
+logger = logging.getLogger(__name__)
+
+
+def _looks_like_wheel(location: str) -> bool:
+    if not location.endswith(WHEEL_EXTENSION):
+        return False
+    if not os.path.isfile(location):
+        return False
+    if not Wheel.wheel_file_re.match(os.path.basename(location)):
+        return False
+    return zipfile.is_zipfile(location)
+
+
+class _DistributionFinder:
+    """Finder to locate distributions.
+
+    The main purpose of this class is to memoize found distributions' names, so
+    only one distribution is returned for each package name. At lot of pip code
+    assumes this (because it is setuptools's behavior), and not doing the same
+    can potentially cause a distribution in lower precedence path to override a
+    higher precedence one if the caller is not careful.
+
+    Eventually we probably want to make it possible to see lower precedence
+    installations as well. It's useful feature, after all.
+    """
+
+    FoundResult = Tuple[importlib.metadata.Distribution, Optional[BasePath]]
+
+    def __init__(self) -> None:
+        self._found_names: Set[NormalizedName] = set()
+
+    def _find_impl(self, location: str) -> Iterator[FoundResult]:
+        """Find distributions in a location."""
+        # Skip looking inside a wheel. Since a package inside a wheel is not
+        # always valid (due to .data directories etc.), its .dist-info entry
+        # should not be considered an installed distribution.
+        if _looks_like_wheel(location):
+            return
+        # To know exactly where we find a distribution, we have to feed in the
+        # paths one by one, instead of dumping the list to importlib.metadata.
+        for dist in importlib.metadata.distributions(path=[location]):
+            info_location = get_info_location(dist)
+            try:
+                name = get_dist_canonical_name(dist)
+            except BadMetadata as e:
+                logger.warning("Skipping %s due to %s", info_location, e.reason)
+                continue
+            if name in self._found_names:
+                continue
+            self._found_names.add(name)
+            yield dist, info_location
+
+    def find(self, location: str) -> Iterator[BaseDistribution]:
+        """Find distributions in a location.
+
+        The path can be either a directory, or a ZIP archive.
+        """
+        for dist, info_location in self._find_impl(location):
+            if info_location is None:
+                installed_location: Optional[BasePath] = None
+            else:
+                installed_location = info_location.parent
+            yield Distribution(dist, info_location, installed_location)
+
+    def find_linked(self, location: str) -> Iterator[BaseDistribution]:
+        """Read location in egg-link files and return distributions in there.
+
+        The path should be a directory; otherwise this returns nothing. This
+        follows how setuptools does this for compatibility. The first non-empty
+        line in the egg-link is read as a path (resolved against the egg-link's
+        containing directory if relative). Distributions found at that linked
+        location are returned.
+        """
+        path = pathlib.Path(location)
+        if not path.is_dir():
+            return
+        for child in path.iterdir():
+            if child.suffix != ".egg-link":
+                continue
+            with child.open() as f:
+                lines = (line.strip() for line in f)
+                target_rel = next((line for line in lines if line), "")
+            if not target_rel:
+                continue
+            target_location = str(path.joinpath(target_rel))
+            for dist, info_location in self._find_impl(target_location):
+                yield Distribution(dist, info_location, path)
+
+    def _find_eggs_in_dir(self, location: str) -> Iterator[BaseDistribution]:
+        from pip._vendor.pkg_resources import find_distributions
+
+        from pip._internal.metadata import pkg_resources as legacy
+
+        with os.scandir(location) as it:
+            for entry in it:
+                if not entry.name.endswith(".egg"):
+                    continue
+                for dist in find_distributions(entry.path):
+                    yield legacy.Distribution(dist)
+
+    def _find_eggs_in_zip(self, location: str) -> Iterator[BaseDistribution]:
+        from pip._vendor.pkg_resources import find_eggs_in_zip
+
+        from pip._internal.metadata import pkg_resources as legacy
+
+        try:
+            importer = zipimport.zipimporter(location)
+        except zipimport.ZipImportError:
+            return
+        for dist in find_eggs_in_zip(importer, location):
+            yield legacy.Distribution(dist)
+
+    def find_eggs(self, location: str) -> Iterator[BaseDistribution]:
+        """Find eggs in a location.
+
+        This actually uses the old *pkg_resources* backend. We likely want to
+        deprecate this so we can eventually remove the *pkg_resources*
+        dependency entirely. Before that, this should first emit a deprecation
+        warning for some versions when using the fallback since importing
+        *pkg_resources* is slow for those who don't need it.
+        """
+        if os.path.isdir(location):
+            yield from self._find_eggs_in_dir(location)
+        if zipfile.is_zipfile(location):
+            yield from self._find_eggs_in_zip(location)
+
+
+@functools.lru_cache(maxsize=None)  # Warn a distribution exactly once.
+def _emit_egg_deprecation(location: Optional[str]) -> None:
+    deprecated(
+        reason=f"Loading egg at {location} is deprecated.",
+        replacement="to use pip for package installation",
+        gone_in="25.1",
+        issue=12330,
+    )
+
+
+class Environment(BaseEnvironment):
+    def __init__(self, paths: Sequence[str]) -> None:
+        self._paths = paths
+
+    @classmethod
+    def default(cls) -> BaseEnvironment:
+        return cls(sys.path)
+
+    @classmethod
+    def from_paths(cls, paths: Optional[List[str]]) -> BaseEnvironment:
+        if paths is None:
+            return cls(sys.path)
+        return cls(paths)
+
+    def _iter_distributions(self) -> Iterator[BaseDistribution]:
+        finder = _DistributionFinder()
+        for location in self._paths:
+            yield from finder.find(location)
+            for dist in finder.find_eggs(location):
+                _emit_egg_deprecation(dist.location)
+                yield dist
+            # This must go last because that's how pkg_resources tie-breaks.
+            yield from finder.find_linked(location)
+
+    def get_distribution(self, name: str) -> Optional[BaseDistribution]:
+        canonical_name = canonicalize_name(name)
+        matches = (
+            distribution
+            for distribution in self.iter_all_distributions()
+            if distribution.canonical_name == canonical_name
+        )
+        return next(matches, None)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/pkg_resources.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/pkg_resources.py
new file mode 100644
index 00000000..4ea84f93
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/metadata/pkg_resources.py
@@ -0,0 +1,301 @@
+import email.message
+import email.parser
+import logging
+import os
+import zipfile
+from typing import (
+    Collection,
+    Iterable,
+    Iterator,
+    List,
+    Mapping,
+    NamedTuple,
+    Optional,
+)
+
+from pip._vendor import pkg_resources
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import Version
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.exceptions import InvalidWheel, NoneMetadataError, UnsupportedWheel
+from pip._internal.utils.egg_link import egg_link_path_from_location
+from pip._internal.utils.misc import display_path, normalize_path
+from pip._internal.utils.wheel import parse_wheel, read_wheel_metadata_file
+
+from .base import (
+    BaseDistribution,
+    BaseEntryPoint,
+    BaseEnvironment,
+    InfoPath,
+    Wheel,
+)
+
+__all__ = ["NAME", "Distribution", "Environment"]
+
+logger = logging.getLogger(__name__)
+
+NAME = "pkg_resources"
+
+
+class EntryPoint(NamedTuple):
+    name: str
+    value: str
+    group: str
+
+
+class InMemoryMetadata:
+    """IMetadataProvider that reads metadata files from a dictionary.
+
+    This also maps metadata decoding exceptions to our internal exception type.
+    """
+
+    def __init__(self, metadata: Mapping[str, bytes], wheel_name: str) -> None:
+        self._metadata = metadata
+        self._wheel_name = wheel_name
+
+    def has_metadata(self, name: str) -> bool:
+        return name in self._metadata
+
+    def get_metadata(self, name: str) -> str:
+        try:
+            return self._metadata[name].decode()
+        except UnicodeDecodeError as e:
+            # Augment the default error with the origin of the file.
+            raise UnsupportedWheel(
+                f"Error decoding metadata for {self._wheel_name}: {e} in {name} file"
+            )
+
+    def get_metadata_lines(self, name: str) -> Iterable[str]:
+        return pkg_resources.yield_lines(self.get_metadata(name))
+
+    def metadata_isdir(self, name: str) -> bool:
+        return False
+
+    def metadata_listdir(self, name: str) -> List[str]:
+        return []
+
+    def run_script(self, script_name: str, namespace: str) -> None:
+        pass
+
+
+class Distribution(BaseDistribution):
+    def __init__(self, dist: pkg_resources.Distribution) -> None:
+        self._dist = dist
+        # This is populated lazily, to avoid loading metadata for all possible
+        # distributions eagerly.
+        self.__extra_mapping: Optional[Mapping[NormalizedName, str]] = None
+
+    @property
+    def _extra_mapping(self) -> Mapping[NormalizedName, str]:
+        if self.__extra_mapping is None:
+            self.__extra_mapping = {
+                canonicalize_name(extra): extra for extra in self._dist.extras
+            }
+
+        return self.__extra_mapping
+
+    @classmethod
+    def from_directory(cls, directory: str) -> BaseDistribution:
+        dist_dir = directory.rstrip(os.sep)
+
+        # Build a PathMetadata object, from path to metadata. :wink:
+        base_dir, dist_dir_name = os.path.split(dist_dir)
+        metadata = pkg_resources.PathMetadata(base_dir, dist_dir)
+
+        # Determine the correct Distribution object type.
+        if dist_dir.endswith(".egg-info"):
+            dist_cls = pkg_resources.Distribution
+            dist_name = os.path.splitext(dist_dir_name)[0]
+        else:
+            assert dist_dir.endswith(".dist-info")
+            dist_cls = pkg_resources.DistInfoDistribution
+            dist_name = os.path.splitext(dist_dir_name)[0].split("-")[0]
+
+        dist = dist_cls(base_dir, project_name=dist_name, metadata=metadata)
+        return cls(dist)
+
+    @classmethod
+    def from_metadata_file_contents(
+        cls,
+        metadata_contents: bytes,
+        filename: str,
+        project_name: str,
+    ) -> BaseDistribution:
+        metadata_dict = {
+            "METADATA": metadata_contents,
+        }
+        dist = pkg_resources.DistInfoDistribution(
+            location=filename,
+            metadata=InMemoryMetadata(metadata_dict, filename),
+            project_name=project_name,
+        )
+        return cls(dist)
+
+    @classmethod
+    def from_wheel(cls, wheel: Wheel, name: str) -> BaseDistribution:
+        try:
+            with wheel.as_zipfile() as zf:
+                info_dir, _ = parse_wheel(zf, name)
+                metadata_dict = {
+                    path.split("/", 1)[-1]: read_wheel_metadata_file(zf, path)
+                    for path in zf.namelist()
+                    if path.startswith(f"{info_dir}/")
+                }
+        except zipfile.BadZipFile as e:
+            raise InvalidWheel(wheel.location, name) from e
+        except UnsupportedWheel as e:
+            raise UnsupportedWheel(f"{name} has an invalid wheel, {e}")
+        dist = pkg_resources.DistInfoDistribution(
+            location=wheel.location,
+            metadata=InMemoryMetadata(metadata_dict, wheel.location),
+            project_name=name,
+        )
+        return cls(dist)
+
+    @property
+    def location(self) -> Optional[str]:
+        return self._dist.location
+
+    @property
+    def installed_location(self) -> Optional[str]:
+        egg_link = egg_link_path_from_location(self.raw_name)
+        if egg_link:
+            location = egg_link
+        elif self.location:
+            location = self.location
+        else:
+            return None
+        return normalize_path(location)
+
+    @property
+    def info_location(self) -> Optional[str]:
+        return self._dist.egg_info
+
+    @property
+    def installed_by_distutils(self) -> bool:
+        # A distutils-installed distribution is provided by FileMetadata. This
+        # provider has a "path" attribute not present anywhere else. Not the
+        # best introspection logic, but pip has been doing this for a long time.
+        try:
+            return bool(self._dist._provider.path)
+        except AttributeError:
+            return False
+
+    @property
+    def canonical_name(self) -> NormalizedName:
+        return canonicalize_name(self._dist.project_name)
+
+    @property
+    def version(self) -> Version:
+        return parse_version(self._dist.version)
+
+    @property
+    def raw_version(self) -> str:
+        return self._dist.version
+
+    def is_file(self, path: InfoPath) -> bool:
+        return self._dist.has_metadata(str(path))
+
+    def iter_distutils_script_names(self) -> Iterator[str]:
+        yield from self._dist.metadata_listdir("scripts")
+
+    def read_text(self, path: InfoPath) -> str:
+        name = str(path)
+        if not self._dist.has_metadata(name):
+            raise FileNotFoundError(name)
+        content = self._dist.get_metadata(name)
+        if content is None:
+            raise NoneMetadataError(self, name)
+        return content
+
+    def iter_entry_points(self) -> Iterable[BaseEntryPoint]:
+        for group, entries in self._dist.get_entry_map().items():
+            for name, entry_point in entries.items():
+                name, _, value = str(entry_point).partition("=")
+                yield EntryPoint(name=name.strip(), value=value.strip(), group=group)
+
+    def _metadata_impl(self) -> email.message.Message:
+        """
+        :raises NoneMetadataError: if the distribution reports `has_metadata()`
+            True but `get_metadata()` returns None.
+        """
+        if isinstance(self._dist, pkg_resources.DistInfoDistribution):
+            metadata_name = "METADATA"
+        else:
+            metadata_name = "PKG-INFO"
+        try:
+            metadata = self.read_text(metadata_name)
+        except FileNotFoundError:
+            if self.location:
+                displaying_path = display_path(self.location)
+            else:
+                displaying_path = repr(self.location)
+            logger.warning("No metadata found in %s", displaying_path)
+            metadata = ""
+        feed_parser = email.parser.FeedParser()
+        feed_parser.feed(metadata)
+        return feed_parser.close()
+
+    def iter_dependencies(self, extras: Collection[str] = ()) -> Iterable[Requirement]:
+        if extras:
+            relevant_extras = set(self._extra_mapping) & set(
+                map(canonicalize_name, extras)
+            )
+            extras = [self._extra_mapping[extra] for extra in relevant_extras]
+        return self._dist.requires(extras)
+
+    def iter_provided_extras(self) -> Iterable[NormalizedName]:
+        return self._extra_mapping.keys()
+
+
+class Environment(BaseEnvironment):
+    def __init__(self, ws: pkg_resources.WorkingSet) -> None:
+        self._ws = ws
+
+    @classmethod
+    def default(cls) -> BaseEnvironment:
+        return cls(pkg_resources.working_set)
+
+    @classmethod
+    def from_paths(cls, paths: Optional[List[str]]) -> BaseEnvironment:
+        return cls(pkg_resources.WorkingSet(paths))
+
+    def _iter_distributions(self) -> Iterator[BaseDistribution]:
+        for dist in self._ws:
+            yield Distribution(dist)
+
+    def _search_distribution(self, name: str) -> Optional[BaseDistribution]:
+        """Find a distribution matching the ``name`` in the environment.
+
+        This searches from *all* distributions available in the environment, to
+        match the behavior of ``pkg_resources.get_distribution()``.
+        """
+        canonical_name = canonicalize_name(name)
+        for dist in self.iter_all_distributions():
+            if dist.canonical_name == canonical_name:
+                return dist
+        return None
+
+    def get_distribution(self, name: str) -> Optional[BaseDistribution]:
+        # Search the distribution by looking through the working set.
+        dist = self._search_distribution(name)
+        if dist:
+            return dist
+
+        # If distribution could not be found, call working_set.require to
+        # update the working set, and try to find the distribution again.
+        # This might happen for e.g. when you install a package twice, once
+        # using setup.py develop and again using setup.py install. Now when
+        # running pip uninstall twice, the package gets removed from the
+        # working set in the first uninstall, so we have to populate the
+        # working set again so that pip knows about it and the packages gets
+        # picked up and is successfully uninstalled the second time too.
+        try:
+            # We didn't pass in any version specifiers, so this can never
+            # raise pkg_resources.VersionConflict.
+            self._ws.require(name)
+        except pkg_resources.DistributionNotFound:
+            return None
+        return self._search_distribution(name)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__init__.py
new file mode 100644
index 00000000..7855226e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__init__.py
@@ -0,0 +1,2 @@
+"""A package that contains models that represent entities.
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..77afc9ef
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/candidate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/candidate.cpython-312.pyc
new file mode 100644
index 00000000..14329183
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/candidate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/direct_url.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/direct_url.cpython-312.pyc
new file mode 100644
index 00000000..f1ac7799
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/direct_url.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/format_control.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/format_control.cpython-312.pyc
new file mode 100644
index 00000000..5757a495
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/format_control.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/index.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/index.cpython-312.pyc
new file mode 100644
index 00000000..01c728ce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/index.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/installation_report.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/installation_report.cpython-312.pyc
new file mode 100644
index 00000000..4dc154f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/installation_report.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/link.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/link.cpython-312.pyc
new file mode 100644
index 00000000..2d317d8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/link.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/scheme.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/scheme.cpython-312.pyc
new file mode 100644
index 00000000..c88ee2a6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/scheme.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/search_scope.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/search_scope.cpython-312.pyc
new file mode 100644
index 00000000..4dec4548
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/search_scope.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/selection_prefs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/selection_prefs.cpython-312.pyc
new file mode 100644
index 00000000..454cd502
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/selection_prefs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/target_python.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/target_python.cpython-312.pyc
new file mode 100644
index 00000000..622c6d8b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/target_python.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..5723c865
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/candidate.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/candidate.py
new file mode 100644
index 00000000..f27f2831
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/candidate.py
@@ -0,0 +1,25 @@
+from dataclasses import dataclass
+
+from pip._vendor.packaging.version import Version
+from pip._vendor.packaging.version import parse as parse_version
+
+from pip._internal.models.link import Link
+
+
+@dataclass(frozen=True)
+class InstallationCandidate:
+    """Represents a potential "candidate" for installation."""
+
+    __slots__ = ["name", "version", "link"]
+
+    name: str
+    version: Version
+    link: Link
+
+    def __init__(self, name: str, version: str, link: Link) -> None:
+        object.__setattr__(self, "name", name)
+        object.__setattr__(self, "version", parse_version(version))
+        object.__setattr__(self, "link", link)
+
+    def __str__(self) -> str:
+        return f"{self.name!r} candidate (version {self.version} at {self.link})"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/direct_url.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/direct_url.py
new file mode 100644
index 00000000..fc5ec8d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/direct_url.py
@@ -0,0 +1,224 @@
+""" PEP 610 """
+
+import json
+import re
+import urllib.parse
+from dataclasses import dataclass
+from typing import Any, ClassVar, Dict, Iterable, Optional, Type, TypeVar, Union
+
+__all__ = [
+    "DirectUrl",
+    "DirectUrlValidationError",
+    "DirInfo",
+    "ArchiveInfo",
+    "VcsInfo",
+]
+
+T = TypeVar("T")
+
+DIRECT_URL_METADATA_NAME = "direct_url.json"
+ENV_VAR_RE = re.compile(r"^\$\{[A-Za-z0-9-_]+\}(:\$\{[A-Za-z0-9-_]+\})?$")
+
+
+class DirectUrlValidationError(Exception):
+    pass
+
+
+def _get(
+    d: Dict[str, Any], expected_type: Type[T], key: str, default: Optional[T] = None
+) -> Optional[T]:
+    """Get value from dictionary and verify expected type."""
+    if key not in d:
+        return default
+    value = d[key]
+    if not isinstance(value, expected_type):
+        raise DirectUrlValidationError(
+            f"{value!r} has unexpected type for {key} (expected {expected_type})"
+        )
+    return value
+
+
+def _get_required(
+    d: Dict[str, Any], expected_type: Type[T], key: str, default: Optional[T] = None
+) -> T:
+    value = _get(d, expected_type, key, default)
+    if value is None:
+        raise DirectUrlValidationError(f"{key} must have a value")
+    return value
+
+
+def _exactly_one_of(infos: Iterable[Optional["InfoType"]]) -> "InfoType":
+    infos = [info for info in infos if info is not None]
+    if not infos:
+        raise DirectUrlValidationError(
+            "missing one of archive_info, dir_info, vcs_info"
+        )
+    if len(infos) > 1:
+        raise DirectUrlValidationError(
+            "more than one of archive_info, dir_info, vcs_info"
+        )
+    assert infos[0] is not None
+    return infos[0]
+
+
+def _filter_none(**kwargs: Any) -> Dict[str, Any]:
+    """Make dict excluding None values."""
+    return {k: v for k, v in kwargs.items() if v is not None}
+
+
+@dataclass
+class VcsInfo:
+    name: ClassVar = "vcs_info"
+
+    vcs: str
+    commit_id: str
+    requested_revision: Optional[str] = None
+
+    @classmethod
+    def _from_dict(cls, d: Optional[Dict[str, Any]]) -> Optional["VcsInfo"]:
+        if d is None:
+            return None
+        return cls(
+            vcs=_get_required(d, str, "vcs"),
+            commit_id=_get_required(d, str, "commit_id"),
+            requested_revision=_get(d, str, "requested_revision"),
+        )
+
+    def _to_dict(self) -> Dict[str, Any]:
+        return _filter_none(
+            vcs=self.vcs,
+            requested_revision=self.requested_revision,
+            commit_id=self.commit_id,
+        )
+
+
+class ArchiveInfo:
+    name = "archive_info"
+
+    def __init__(
+        self,
+        hash: Optional[str] = None,
+        hashes: Optional[Dict[str, str]] = None,
+    ) -> None:
+        # set hashes before hash, since the hash setter will further populate hashes
+        self.hashes = hashes
+        self.hash = hash
+
+    @property
+    def hash(self) -> Optional[str]:
+        return self._hash
+
+    @hash.setter
+    def hash(self, value: Optional[str]) -> None:
+        if value is not None:
+            # Auto-populate the hashes key to upgrade to the new format automatically.
+            # We don't back-populate the legacy hash key from hashes.
+            try:
+                hash_name, hash_value = value.split("=", 1)
+            except ValueError:
+                raise DirectUrlValidationError(
+                    f"invalid archive_info.hash format: {value!r}"
+                )
+            if self.hashes is None:
+                self.hashes = {hash_name: hash_value}
+            elif hash_name not in self.hashes:
+                self.hashes = self.hashes.copy()
+                self.hashes[hash_name] = hash_value
+        self._hash = value
+
+    @classmethod
+    def _from_dict(cls, d: Optional[Dict[str, Any]]) -> Optional["ArchiveInfo"]:
+        if d is None:
+            return None
+        return cls(hash=_get(d, str, "hash"), hashes=_get(d, dict, "hashes"))
+
+    def _to_dict(self) -> Dict[str, Any]:
+        return _filter_none(hash=self.hash, hashes=self.hashes)
+
+
+@dataclass
+class DirInfo:
+    name: ClassVar = "dir_info"
+
+    editable: bool = False
+
+    @classmethod
+    def _from_dict(cls, d: Optional[Dict[str, Any]]) -> Optional["DirInfo"]:
+        if d is None:
+            return None
+        return cls(editable=_get_required(d, bool, "editable", default=False))
+
+    def _to_dict(self) -> Dict[str, Any]:
+        return _filter_none(editable=self.editable or None)
+
+
+InfoType = Union[ArchiveInfo, DirInfo, VcsInfo]
+
+
+@dataclass
+class DirectUrl:
+    url: str
+    info: InfoType
+    subdirectory: Optional[str] = None
+
+    def _remove_auth_from_netloc(self, netloc: str) -> str:
+        if "@" not in netloc:
+            return netloc
+        user_pass, netloc_no_user_pass = netloc.split("@", 1)
+        if (
+            isinstance(self.info, VcsInfo)
+            and self.info.vcs == "git"
+            and user_pass == "git"
+        ):
+            return netloc
+        if ENV_VAR_RE.match(user_pass):
+            return netloc
+        return netloc_no_user_pass
+
+    @property
+    def redacted_url(self) -> str:
+        """url with user:password part removed unless it is formed with
+        environment variables as specified in PEP 610, or it is ``git``
+        in the case of a git URL.
+        """
+        purl = urllib.parse.urlsplit(self.url)
+        netloc = self._remove_auth_from_netloc(purl.netloc)
+        surl = urllib.parse.urlunsplit(
+            (purl.scheme, netloc, purl.path, purl.query, purl.fragment)
+        )
+        return surl
+
+    def validate(self) -> None:
+        self.from_dict(self.to_dict())
+
+    @classmethod
+    def from_dict(cls, d: Dict[str, Any]) -> "DirectUrl":
+        return DirectUrl(
+            url=_get_required(d, str, "url"),
+            subdirectory=_get(d, str, "subdirectory"),
+            info=_exactly_one_of(
+                [
+                    ArchiveInfo._from_dict(_get(d, dict, "archive_info")),
+                    DirInfo._from_dict(_get(d, dict, "dir_info")),
+                    VcsInfo._from_dict(_get(d, dict, "vcs_info")),
+                ]
+            ),
+        )
+
+    def to_dict(self) -> Dict[str, Any]:
+        res = _filter_none(
+            url=self.redacted_url,
+            subdirectory=self.subdirectory,
+        )
+        res[self.info.name] = self.info._to_dict()
+        return res
+
+    @classmethod
+    def from_json(cls, s: str) -> "DirectUrl":
+        return cls.from_dict(json.loads(s))
+
+    def to_json(self) -> str:
+        return json.dumps(self.to_dict(), sort_keys=True)
+
+    def is_local_editable(self) -> bool:
+        return isinstance(self.info, DirInfo) and self.info.editable
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/format_control.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/format_control.py
new file mode 100644
index 00000000..ccd11272
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/format_control.py
@@ -0,0 +1,78 @@
+from typing import FrozenSet, Optional, Set
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.exceptions import CommandError
+
+
+class FormatControl:
+    """Helper for managing formats from which a package can be installed."""
+
+    __slots__ = ["no_binary", "only_binary"]
+
+    def __init__(
+        self,
+        no_binary: Optional[Set[str]] = None,
+        only_binary: Optional[Set[str]] = None,
+    ) -> None:
+        if no_binary is None:
+            no_binary = set()
+        if only_binary is None:
+            only_binary = set()
+
+        self.no_binary = no_binary
+        self.only_binary = only_binary
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, self.__class__):
+            return NotImplemented
+
+        if self.__slots__ != other.__slots__:
+            return False
+
+        return all(getattr(self, k) == getattr(other, k) for k in self.__slots__)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.no_binary}, {self.only_binary})"
+
+    @staticmethod
+    def handle_mutual_excludes(value: str, target: Set[str], other: Set[str]) -> None:
+        if value.startswith("-"):
+            raise CommandError(
+                "--no-binary / --only-binary option requires 1 argument."
+            )
+        new = value.split(",")
+        while ":all:" in new:
+            other.clear()
+            target.clear()
+            target.add(":all:")
+            del new[: new.index(":all:") + 1]
+            # Without a none, we want to discard everything as :all: covers it
+            if ":none:" not in new:
+                return
+        for name in new:
+            if name == ":none:":
+                target.clear()
+                continue
+            name = canonicalize_name(name)
+            other.discard(name)
+            target.add(name)
+
+    def get_allowed_formats(self, canonical_name: str) -> FrozenSet[str]:
+        result = {"binary", "source"}
+        if canonical_name in self.only_binary:
+            result.discard("source")
+        elif canonical_name in self.no_binary:
+            result.discard("binary")
+        elif ":all:" in self.only_binary:
+            result.discard("source")
+        elif ":all:" in self.no_binary:
+            result.discard("binary")
+        return frozenset(result)
+
+    def disallow_binaries(self) -> None:
+        self.handle_mutual_excludes(
+            ":all:",
+            self.no_binary,
+            self.only_binary,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/index.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/index.py
new file mode 100644
index 00000000..b94c3251
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/index.py
@@ -0,0 +1,28 @@
+import urllib.parse
+
+
+class PackageIndex:
+    """Represents a Package Index and provides easier access to endpoints"""
+
+    __slots__ = ["url", "netloc", "simple_url", "pypi_url", "file_storage_domain"]
+
+    def __init__(self, url: str, file_storage_domain: str) -> None:
+        super().__init__()
+        self.url = url
+        self.netloc = urllib.parse.urlsplit(url).netloc
+        self.simple_url = self._url_for_path("simple")
+        self.pypi_url = self._url_for_path("pypi")
+
+        # This is part of a temporary hack used to block installs of PyPI
+        # packages which depend on external urls only necessary until PyPI can
+        # block such packages themselves
+        self.file_storage_domain = file_storage_domain
+
+    def _url_for_path(self, path: str) -> str:
+        return urllib.parse.urljoin(self.url, path)
+
+
+PyPI = PackageIndex("https://pypi.org/", file_storage_domain="files.pythonhosted.org")
+TestPyPI = PackageIndex(
+    "https://test.pypi.org/", file_storage_domain="test-files.pythonhosted.org"
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/installation_report.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/installation_report.py
new file mode 100644
index 00000000..b9c6330d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/installation_report.py
@@ -0,0 +1,56 @@
+from typing import Any, Dict, Sequence
+
+from pip._vendor.packaging.markers import default_environment
+
+from pip import __version__
+from pip._internal.req.req_install import InstallRequirement
+
+
+class InstallationReport:
+    def __init__(self, install_requirements: Sequence[InstallRequirement]):
+        self._install_requirements = install_requirements
+
+    @classmethod
+    def _install_req_to_dict(cls, ireq: InstallRequirement) -> Dict[str, Any]:
+        assert ireq.download_info, f"No download_info for {ireq}"
+        res = {
+            # PEP 610 json for the download URL. download_info.archive_info.hashes may
+            # be absent when the requirement was installed from the wheel cache
+            # and the cache entry was populated by an older pip version that did not
+            # record origin.json.
+            "download_info": ireq.download_info.to_dict(),
+            # is_direct is true if the requirement was a direct URL reference (which
+            # includes editable requirements), and false if the requirement was
+            # downloaded from a PEP 503 index or --find-links.
+            "is_direct": ireq.is_direct,
+            # is_yanked is true if the requirement was yanked from the index, but
+            # was still selected by pip to conform to PEP 592.
+            "is_yanked": ireq.link.is_yanked if ireq.link else False,
+            # requested is true if the requirement was specified by the user (aka
+            # top level requirement), and false if it was installed as a dependency of a
+            # requirement. https://peps.python.org/pep-0376/#requested
+            "requested": ireq.user_supplied,
+            # PEP 566 json encoding for metadata
+            # https://www.python.org/dev/peps/pep-0566/#json-compatible-metadata
+            "metadata": ireq.get_dist().metadata_dict,
+        }
+        if ireq.user_supplied and ireq.extras:
+            # For top level requirements, the list of requested extras, if any.
+            res["requested_extras"] = sorted(ireq.extras)
+        return res
+
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "version": "1",
+            "pip_version": __version__,
+            "install": [
+                self._install_req_to_dict(ireq) for ireq in self._install_requirements
+            ],
+            # https://peps.python.org/pep-0508/#environment-markers
+            # TODO: currently, the resolver uses the default environment to evaluate
+            # environment markers, so that is what we report here. In the future, it
+            # should also take into account options such as --python-version or
+            # --platform, perhaps under the form of an environment_override field?
+            # https://github.com/pypa/pip/issues/11198
+            "environment": default_environment(),
+        }
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/link.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/link.py
new file mode 100644
index 00000000..27ad0160
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/link.py
@@ -0,0 +1,604 @@
+import functools
+import itertools
+import logging
+import os
+import posixpath
+import re
+import urllib.parse
+from dataclasses import dataclass
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    List,
+    Mapping,
+    NamedTuple,
+    Optional,
+    Tuple,
+    Union,
+)
+
+from pip._internal.utils.deprecation import deprecated
+from pip._internal.utils.filetypes import WHEEL_EXTENSION
+from pip._internal.utils.hashes import Hashes
+from pip._internal.utils.misc import (
+    pairwise,
+    redact_auth_from_url,
+    split_auth_from_netloc,
+    splitext,
+)
+from pip._internal.utils.urls import path_to_url, url_to_path
+
+if TYPE_CHECKING:
+    from pip._internal.index.collector import IndexContent
+
+logger = logging.getLogger(__name__)
+
+
+# Order matters, earlier hashes have a precedence over later hashes for what
+# we will pick to use.
+_SUPPORTED_HASHES = ("sha512", "sha384", "sha256", "sha224", "sha1", "md5")
+
+
+@dataclass(frozen=True)
+class LinkHash:
+    """Links to content may have embedded hash values. This class parses those.
+
+    `name` must be any member of `_SUPPORTED_HASHES`.
+
+    This class can be converted to and from `ArchiveInfo`. While ArchiveInfo intends to
+    be JSON-serializable to conform to PEP 610, this class contains the logic for
+    parsing a hash name and value for correctness, and then checking whether that hash
+    conforms to a schema with `.is_hash_allowed()`."""
+
+    name: str
+    value: str
+
+    _hash_url_fragment_re = re.compile(
+        # NB: we do not validate that the second group (.*) is a valid hex
+        # digest. Instead, we simply keep that string in this class, and then check it
+        # against Hashes when hash-checking is needed. This is easier to debug than
+        # proactively discarding an invalid hex digest, as we handle incorrect hashes
+        # and malformed hashes in the same place.
+        r"[#&]({choices})=([^&]*)".format(
+            choices="|".join(re.escape(hash_name) for hash_name in _SUPPORTED_HASHES)
+        ),
+    )
+
+    def __post_init__(self) -> None:
+        assert self.name in _SUPPORTED_HASHES
+
+    @classmethod
+    @functools.lru_cache(maxsize=None)
+    def find_hash_url_fragment(cls, url: str) -> Optional["LinkHash"]:
+        """Search a string for a checksum algorithm name and encoded output value."""
+        match = cls._hash_url_fragment_re.search(url)
+        if match is None:
+            return None
+        name, value = match.groups()
+        return cls(name=name, value=value)
+
+    def as_dict(self) -> Dict[str, str]:
+        return {self.name: self.value}
+
+    def as_hashes(self) -> Hashes:
+        """Return a Hashes instance which checks only for the current hash."""
+        return Hashes({self.name: [self.value]})
+
+    def is_hash_allowed(self, hashes: Optional[Hashes]) -> bool:
+        """
+        Return True if the current hash is allowed by `hashes`.
+        """
+        if hashes is None:
+            return False
+        return hashes.is_hash_allowed(self.name, hex_digest=self.value)
+
+
+@dataclass(frozen=True)
+class MetadataFile:
+    """Information about a core metadata file associated with a distribution."""
+
+    hashes: Optional[Dict[str, str]]
+
+    def __post_init__(self) -> None:
+        if self.hashes is not None:
+            assert all(name in _SUPPORTED_HASHES for name in self.hashes)
+
+
+def supported_hashes(hashes: Optional[Dict[str, str]]) -> Optional[Dict[str, str]]:
+    # Remove any unsupported hash types from the mapping. If this leaves no
+    # supported hashes, return None
+    if hashes is None:
+        return None
+    hashes = {n: v for n, v in hashes.items() if n in _SUPPORTED_HASHES}
+    if not hashes:
+        return None
+    return hashes
+
+
+def _clean_url_path_part(part: str) -> str:
+    """
+    Clean a "part" of a URL path (i.e. after splitting on "@" characters).
+    """
+    # We unquote prior to quoting to make sure nothing is double quoted.
+    return urllib.parse.quote(urllib.parse.unquote(part))
+
+
+def _clean_file_url_path(part: str) -> str:
+    """
+    Clean the first part of a URL path that corresponds to a local
+    filesystem path (i.e. the first part after splitting on "@" characters).
+    """
+    # We unquote prior to quoting to make sure nothing is double quoted.
+    # Also, on Windows the path part might contain a drive letter which
+    # should not be quoted. On Linux where drive letters do not
+    # exist, the colon should be quoted. We rely on urllib.request
+    # to do the right thing here.
+    return urllib.request.pathname2url(urllib.request.url2pathname(part))
+
+
+# percent-encoded:                   /
+_reserved_chars_re = re.compile("(@|%2F)", re.IGNORECASE)
+
+
+def _clean_url_path(path: str, is_local_path: bool) -> str:
+    """
+    Clean the path portion of a URL.
+    """
+    if is_local_path:
+        clean_func = _clean_file_url_path
+    else:
+        clean_func = _clean_url_path_part
+
+    # Split on the reserved characters prior to cleaning so that
+    # revision strings in VCS URLs are properly preserved.
+    parts = _reserved_chars_re.split(path)
+
+    cleaned_parts = []
+    for to_clean, reserved in pairwise(itertools.chain(parts, [""])):
+        cleaned_parts.append(clean_func(to_clean))
+        # Normalize %xx escapes (e.g. %2f -> %2F)
+        cleaned_parts.append(reserved.upper())
+
+    return "".join(cleaned_parts)
+
+
+def _ensure_quoted_url(url: str) -> str:
+    """
+    Make sure a link is fully quoted.
+    For example, if ' ' occurs in the URL, it will be replaced with "%20",
+    and without double-quoting other characters.
+    """
+    # Split the URL into parts according to the general structure
+    # `scheme://netloc/path?query#fragment`.
+    result = urllib.parse.urlsplit(url)
+    # If the netloc is empty, then the URL refers to a local filesystem path.
+    is_local_path = not result.netloc
+    path = _clean_url_path(result.path, is_local_path=is_local_path)
+    return urllib.parse.urlunsplit(result._replace(path=path))
+
+
+def _absolute_link_url(base_url: str, url: str) -> str:
+    """
+    A faster implementation of urllib.parse.urljoin with a shortcut
+    for absolute http/https URLs.
+    """
+    if url.startswith(("https://", "http://")):
+        return url
+    else:
+        return urllib.parse.urljoin(base_url, url)
+
+
+@functools.total_ordering
+class Link:
+    """Represents a parsed link from a Package Index's simple URL"""
+
+    __slots__ = [
+        "_parsed_url",
+        "_url",
+        "_path",
+        "_hashes",
+        "comes_from",
+        "requires_python",
+        "yanked_reason",
+        "metadata_file_data",
+        "cache_link_parsing",
+        "egg_fragment",
+    ]
+
+    def __init__(
+        self,
+        url: str,
+        comes_from: Optional[Union[str, "IndexContent"]] = None,
+        requires_python: Optional[str] = None,
+        yanked_reason: Optional[str] = None,
+        metadata_file_data: Optional[MetadataFile] = None,
+        cache_link_parsing: bool = True,
+        hashes: Optional[Mapping[str, str]] = None,
+    ) -> None:
+        """
+        :param url: url of the resource pointed to (href of the link)
+        :param comes_from: instance of IndexContent where the link was found,
+            or string.
+        :param requires_python: String containing the `Requires-Python`
+            metadata field, specified in PEP 345. This may be specified by
+            a data-requires-python attribute in the HTML link tag, as
+            described in PEP 503.
+        :param yanked_reason: the reason the file has been yanked, if the
+            file has been yanked, or None if the file hasn't been yanked.
+            This is the value of the "data-yanked" attribute, if present, in
+            a simple repository HTML link. If the file has been yanked but
+            no reason was provided, this should be the empty string. See
+            PEP 592 for more information and the specification.
+        :param metadata_file_data: the metadata attached to the file, or None if
+            no such metadata is provided. This argument, if not None, indicates
+            that a separate metadata file exists, and also optionally supplies
+            hashes for that file.
+        :param cache_link_parsing: A flag that is used elsewhere to determine
+            whether resources retrieved from this link should be cached. PyPI
+            URLs should generally have this set to False, for example.
+        :param hashes: A mapping of hash names to digests to allow us to
+            determine the validity of a download.
+        """
+
+        # The comes_from, requires_python, and metadata_file_data arguments are
+        # only used by classmethods of this class, and are not used in client
+        # code directly.
+
+        # url can be a UNC windows share
+        if url.startswith("\\\\"):
+            url = path_to_url(url)
+
+        self._parsed_url = urllib.parse.urlsplit(url)
+        # Store the url as a private attribute to prevent accidentally
+        # trying to set a new value.
+        self._url = url
+        # The .path property is hot, so calculate its value ahead of time.
+        self._path = urllib.parse.unquote(self._parsed_url.path)
+
+        link_hash = LinkHash.find_hash_url_fragment(url)
+        hashes_from_link = {} if link_hash is None else link_hash.as_dict()
+        if hashes is None:
+            self._hashes = hashes_from_link
+        else:
+            self._hashes = {**hashes, **hashes_from_link}
+
+        self.comes_from = comes_from
+        self.requires_python = requires_python if requires_python else None
+        self.yanked_reason = yanked_reason
+        self.metadata_file_data = metadata_file_data
+
+        self.cache_link_parsing = cache_link_parsing
+        self.egg_fragment = self._egg_fragment()
+
+    @classmethod
+    def from_json(
+        cls,
+        file_data: Dict[str, Any],
+        page_url: str,
+    ) -> Optional["Link"]:
+        """
+        Convert an pypi json document from a simple repository page into a Link.
+        """
+        file_url = file_data.get("url")
+        if file_url is None:
+            return None
+
+        url = _ensure_quoted_url(_absolute_link_url(page_url, file_url))
+        pyrequire = file_data.get("requires-python")
+        yanked_reason = file_data.get("yanked")
+        hashes = file_data.get("hashes", {})
+
+        # PEP 714: Indexes must use the name core-metadata, but
+        # clients should support the old name as a fallback for compatibility.
+        metadata_info = file_data.get("core-metadata")
+        if metadata_info is None:
+            metadata_info = file_data.get("dist-info-metadata")
+
+        # The metadata info value may be a boolean, or a dict of hashes.
+        if isinstance(metadata_info, dict):
+            # The file exists, and hashes have been supplied
+            metadata_file_data = MetadataFile(supported_hashes(metadata_info))
+        elif metadata_info:
+            # The file exists, but there are no hashes
+            metadata_file_data = MetadataFile(None)
+        else:
+            # False or not present: the file does not exist
+            metadata_file_data = None
+
+        # The Link.yanked_reason expects an empty string instead of a boolean.
+        if yanked_reason and not isinstance(yanked_reason, str):
+            yanked_reason = ""
+        # The Link.yanked_reason expects None instead of False.
+        elif not yanked_reason:
+            yanked_reason = None
+
+        return cls(
+            url,
+            comes_from=page_url,
+            requires_python=pyrequire,
+            yanked_reason=yanked_reason,
+            hashes=hashes,
+            metadata_file_data=metadata_file_data,
+        )
+
+    @classmethod
+    def from_element(
+        cls,
+        anchor_attribs: Dict[str, Optional[str]],
+        page_url: str,
+        base_url: str,
+    ) -> Optional["Link"]:
+        """
+        Convert an anchor element's attributes in a simple repository page to a Link.
+        """
+        href = anchor_attribs.get("href")
+        if not href:
+            return None
+
+        url = _ensure_quoted_url(_absolute_link_url(base_url, href))
+        pyrequire = anchor_attribs.get("data-requires-python")
+        yanked_reason = anchor_attribs.get("data-yanked")
+
+        # PEP 714: Indexes must use the name data-core-metadata, but
+        # clients should support the old name as a fallback for compatibility.
+        metadata_info = anchor_attribs.get("data-core-metadata")
+        if metadata_info is None:
+            metadata_info = anchor_attribs.get("data-dist-info-metadata")
+        # The metadata info value may be the string "true", or a string of
+        # the form "hashname=hashval"
+        if metadata_info == "true":
+            # The file exists, but there are no hashes
+            metadata_file_data = MetadataFile(None)
+        elif metadata_info is None:
+            # The file does not exist
+            metadata_file_data = None
+        else:
+            # The file exists, and hashes have been supplied
+            hashname, sep, hashval = metadata_info.partition("=")
+            if sep == "=":
+                metadata_file_data = MetadataFile(supported_hashes({hashname: hashval}))
+            else:
+                # Error - data is wrong. Treat as no hashes supplied.
+                logger.debug(
+                    "Index returned invalid data-dist-info-metadata value: %s",
+                    metadata_info,
+                )
+                metadata_file_data = MetadataFile(None)
+
+        return cls(
+            url,
+            comes_from=page_url,
+            requires_python=pyrequire,
+            yanked_reason=yanked_reason,
+            metadata_file_data=metadata_file_data,
+        )
+
+    def __str__(self) -> str:
+        if self.requires_python:
+            rp = f" (requires-python:{self.requires_python})"
+        else:
+            rp = ""
+        if self.comes_from:
+            return f"{redact_auth_from_url(self._url)} (from {self.comes_from}){rp}"
+        else:
+            return redact_auth_from_url(str(self._url))
+
+    def __repr__(self) -> str:
+        return f"<Link {self}>"
+
+    def __hash__(self) -> int:
+        return hash(self.url)
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Link):
+            return NotImplemented
+        return self.url == other.url
+
+    def __lt__(self, other: Any) -> bool:
+        if not isinstance(other, Link):
+            return NotImplemented
+        return self.url < other.url
+
+    @property
+    def url(self) -> str:
+        return self._url
+
+    @property
+    def filename(self) -> str:
+        path = self.path.rstrip("/")
+        name = posixpath.basename(path)
+        if not name:
+            # Make sure we don't leak auth information if the netloc
+            # includes a username and password.
+            netloc, user_pass = split_auth_from_netloc(self.netloc)
+            return netloc
+
+        name = urllib.parse.unquote(name)
+        assert name, f"URL {self._url!r} produced no filename"
+        return name
+
+    @property
+    def file_path(self) -> str:
+        return url_to_path(self.url)
+
+    @property
+    def scheme(self) -> str:
+        return self._parsed_url.scheme
+
+    @property
+    def netloc(self) -> str:
+        """
+        This can contain auth information.
+        """
+        return self._parsed_url.netloc
+
+    @property
+    def path(self) -> str:
+        return self._path
+
+    def splitext(self) -> Tuple[str, str]:
+        return splitext(posixpath.basename(self.path.rstrip("/")))
+
+    @property
+    def ext(self) -> str:
+        return self.splitext()[1]
+
+    @property
+    def url_without_fragment(self) -> str:
+        scheme, netloc, path, query, fragment = self._parsed_url
+        return urllib.parse.urlunsplit((scheme, netloc, path, query, ""))
+
+    _egg_fragment_re = re.compile(r"[#&]egg=([^&]*)")
+
+    # Per PEP 508.
+    _project_name_re = re.compile(
+        r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.IGNORECASE
+    )
+
+    def _egg_fragment(self) -> Optional[str]:
+        match = self._egg_fragment_re.search(self._url)
+        if not match:
+            return None
+
+        # An egg fragment looks like a PEP 508 project name, along with
+        # an optional extras specifier. Anything else is invalid.
+        project_name = match.group(1)
+        if not self._project_name_re.match(project_name):
+            deprecated(
+                reason=f"{self} contains an egg fragment with a non-PEP 508 name.",
+                replacement="to use the req @ url syntax, and remove the egg fragment",
+                gone_in="25.1",
+                issue=13157,
+            )
+
+        return project_name
+
+    _subdirectory_fragment_re = re.compile(r"[#&]subdirectory=([^&]*)")
+
+    @property
+    def subdirectory_fragment(self) -> Optional[str]:
+        match = self._subdirectory_fragment_re.search(self._url)
+        if not match:
+            return None
+        return match.group(1)
+
+    def metadata_link(self) -> Optional["Link"]:
+        """Return a link to the associated core metadata file (if any)."""
+        if self.metadata_file_data is None:
+            return None
+        metadata_url = f"{self.url_without_fragment}.metadata"
+        if self.metadata_file_data.hashes is None:
+            return Link(metadata_url)
+        return Link(metadata_url, hashes=self.metadata_file_data.hashes)
+
+    def as_hashes(self) -> Hashes:
+        return Hashes({k: [v] for k, v in self._hashes.items()})
+
+    @property
+    def hash(self) -> Optional[str]:
+        return next(iter(self._hashes.values()), None)
+
+    @property
+    def hash_name(self) -> Optional[str]:
+        return next(iter(self._hashes), None)
+
+    @property
+    def show_url(self) -> str:
+        return posixpath.basename(self._url.split("#", 1)[0].split("?", 1)[0])
+
+    @property
+    def is_file(self) -> bool:
+        return self.scheme == "file"
+
+    def is_existing_dir(self) -> bool:
+        return self.is_file and os.path.isdir(self.file_path)
+
+    @property
+    def is_wheel(self) -> bool:
+        return self.ext == WHEEL_EXTENSION
+
+    @property
+    def is_vcs(self) -> bool:
+        from pip._internal.vcs import vcs
+
+        return self.scheme in vcs.all_schemes
+
+    @property
+    def is_yanked(self) -> bool:
+        return self.yanked_reason is not None
+
+    @property
+    def has_hash(self) -> bool:
+        return bool(self._hashes)
+
+    def is_hash_allowed(self, hashes: Optional[Hashes]) -> bool:
+        """
+        Return True if the link has a hash and it is allowed by `hashes`.
+        """
+        if hashes is None:
+            return False
+        return any(hashes.is_hash_allowed(k, v) for k, v in self._hashes.items())
+
+
+class _CleanResult(NamedTuple):
+    """Convert link for equivalency check.
+
+    This is used in the resolver to check whether two URL-specified requirements
+    likely point to the same distribution and can be considered equivalent. This
+    equivalency logic avoids comparing URLs literally, which can be too strict
+    (e.g. "a=1&b=2" vs "b=2&a=1") and produce conflicts unexpecting to users.
+
+    Currently this does three things:
+
+    1. Drop the basic auth part. This is technically wrong since a server can
+       serve different content based on auth, but if it does that, it is even
+       impossible to guarantee two URLs without auth are equivalent, since
+       the user can input different auth information when prompted. So the
+       practical solution is to assume the auth doesn't affect the response.
+    2. Parse the query to avoid the ordering issue. Note that ordering under the
+       same key in the query are NOT cleaned; i.e. "a=1&a=2" and "a=2&a=1" are
+       still considered different.
+    3. Explicitly drop most of the fragment part, except ``subdirectory=`` and
+       hash values, since it should have no impact the downloaded content. Note
+       that this drops the "egg=" part historically used to denote the requested
+       project (and extras), which is wrong in the strictest sense, but too many
+       people are supplying it inconsistently to cause superfluous resolution
+       conflicts, so we choose to also ignore them.
+    """
+
+    parsed: urllib.parse.SplitResult
+    query: Dict[str, List[str]]
+    subdirectory: str
+    hashes: Dict[str, str]
+
+
+def _clean_link(link: Link) -> _CleanResult:
+    parsed = link._parsed_url
+    netloc = parsed.netloc.rsplit("@", 1)[-1]
+    # According to RFC 8089, an empty host in file: means localhost.
+    if parsed.scheme == "file" and not netloc:
+        netloc = "localhost"
+    fragment = urllib.parse.parse_qs(parsed.fragment)
+    if "egg" in fragment:
+        logger.debug("Ignoring egg= fragment in %s", link)
+    try:
+        # If there are multiple subdirectory values, use the first one.
+        # This matches the behavior of Link.subdirectory_fragment.
+        subdirectory = fragment["subdirectory"][0]
+    except (IndexError, KeyError):
+        subdirectory = ""
+    # If there are multiple hash values under the same algorithm, use the
+    # first one. This matches the behavior of Link.hash_value.
+    hashes = {k: fragment[k][0] for k in _SUPPORTED_HASHES if k in fragment}
+    return _CleanResult(
+        parsed=parsed._replace(netloc=netloc, query="", fragment=""),
+        query=urllib.parse.parse_qs(parsed.query),
+        subdirectory=subdirectory,
+        hashes=hashes,
+    )
+
+
+@functools.lru_cache(maxsize=None)
+def links_equivalent(link1: Link, link2: Link) -> bool:
+    return _clean_link(link1) == _clean_link(link2)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/scheme.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/scheme.py
new file mode 100644
index 00000000..06a9a550
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/scheme.py
@@ -0,0 +1,25 @@
+"""
+For types associated with installation schemes.
+
+For a general overview of available schemes and their context, see
+https://docs.python.org/3/install/index.html#alternate-installation.
+"""
+
+from dataclasses import dataclass
+
+SCHEME_KEYS = ["platlib", "purelib", "headers", "scripts", "data"]
+
+
+@dataclass(frozen=True)
+class Scheme:
+    """A Scheme holds paths which are used as the base directories for
+    artifacts associated with a Python package.
+    """
+
+    __slots__ = SCHEME_KEYS
+
+    platlib: str
+    purelib: str
+    headers: str
+    scripts: str
+    data: str
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/search_scope.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/search_scope.py
new file mode 100644
index 00000000..ee7bc862
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/search_scope.py
@@ -0,0 +1,127 @@
+import itertools
+import logging
+import os
+import posixpath
+import urllib.parse
+from dataclasses import dataclass
+from typing import List
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.models.index import PyPI
+from pip._internal.utils.compat import has_tls
+from pip._internal.utils.misc import normalize_path, redact_auth_from_url
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class SearchScope:
+    """
+    Encapsulates the locations that pip is configured to search.
+    """
+
+    __slots__ = ["find_links", "index_urls", "no_index"]
+
+    find_links: List[str]
+    index_urls: List[str]
+    no_index: bool
+
+    @classmethod
+    def create(
+        cls,
+        find_links: List[str],
+        index_urls: List[str],
+        no_index: bool,
+    ) -> "SearchScope":
+        """
+        Create a SearchScope object after normalizing the `find_links`.
+        """
+        # Build find_links. If an argument starts with ~, it may be
+        # a local file relative to a home directory. So try normalizing
+        # it and if it exists, use the normalized version.
+        # This is deliberately conservative - it might be fine just to
+        # blindly normalize anything starting with a ~...
+        built_find_links: List[str] = []
+        for link in find_links:
+            if link.startswith("~"):
+                new_link = normalize_path(link)
+                if os.path.exists(new_link):
+                    link = new_link
+            built_find_links.append(link)
+
+        # If we don't have TLS enabled, then WARN if anyplace we're looking
+        # relies on TLS.
+        if not has_tls():
+            for link in itertools.chain(index_urls, built_find_links):
+                parsed = urllib.parse.urlparse(link)
+                if parsed.scheme == "https":
+                    logger.warning(
+                        "pip is configured with locations that require "
+                        "TLS/SSL, however the ssl module in Python is not "
+                        "available."
+                    )
+                    break
+
+        return cls(
+            find_links=built_find_links,
+            index_urls=index_urls,
+            no_index=no_index,
+        )
+
+    def get_formatted_locations(self) -> str:
+        lines = []
+        redacted_index_urls = []
+        if self.index_urls and self.index_urls != [PyPI.simple_url]:
+            for url in self.index_urls:
+                redacted_index_url = redact_auth_from_url(url)
+
+                # Parse the URL
+                purl = urllib.parse.urlsplit(redacted_index_url)
+
+                # URL is generally invalid if scheme and netloc is missing
+                # there are issues with Python and URL parsing, so this test
+                # is a bit crude. See bpo-20271, bpo-23505. Python doesn't
+                # always parse invalid URLs correctly - it should raise
+                # exceptions for malformed URLs
+                if not purl.scheme and not purl.netloc:
+                    logger.warning(
+                        'The index url "%s" seems invalid, please provide a scheme.',
+                        redacted_index_url,
+                    )
+
+                redacted_index_urls.append(redacted_index_url)
+
+            lines.append(
+                "Looking in indexes: {}".format(", ".join(redacted_index_urls))
+            )
+
+        if self.find_links:
+            lines.append(
+                "Looking in links: {}".format(
+                    ", ".join(redact_auth_from_url(url) for url in self.find_links)
+                )
+            )
+        return "\n".join(lines)
+
+    def get_index_urls_locations(self, project_name: str) -> List[str]:
+        """Returns the locations found via self.index_urls
+
+        Checks the url_name on the main (first in the list) index and
+        use this url_name to produce all locations
+        """
+
+        def mkurl_pypi_url(url: str) -> str:
+            loc = posixpath.join(
+                url, urllib.parse.quote(canonicalize_name(project_name))
+            )
+            # For maximum compatibility with easy_install, ensure the path
+            # ends in a trailing slash.  Although this isn't in the spec
+            # (and PyPI can handle it without the slash) some other index
+            # implementations might break if they relied on easy_install's
+            # behavior.
+            if not loc.endswith("/"):
+                loc = loc + "/"
+            return loc
+
+        return [mkurl_pypi_url(url) for url in self.index_urls]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/selection_prefs.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/selection_prefs.py
new file mode 100644
index 00000000..e9b50aa5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/selection_prefs.py
@@ -0,0 +1,53 @@
+from typing import Optional
+
+from pip._internal.models.format_control import FormatControl
+
+
+# TODO: This needs Python 3.10's improved slots support for dataclasses
+# to be converted into a dataclass.
+class SelectionPreferences:
+    """
+    Encapsulates the candidate selection preferences for downloading
+    and installing files.
+    """
+
+    __slots__ = [
+        "allow_yanked",
+        "allow_all_prereleases",
+        "format_control",
+        "prefer_binary",
+        "ignore_requires_python",
+    ]
+
+    # Don't include an allow_yanked default value to make sure each call
+    # site considers whether yanked releases are allowed. This also causes
+    # that decision to be made explicit in the calling code, which helps
+    # people when reading the code.
+    def __init__(
+        self,
+        allow_yanked: bool,
+        allow_all_prereleases: bool = False,
+        format_control: Optional[FormatControl] = None,
+        prefer_binary: bool = False,
+        ignore_requires_python: Optional[bool] = None,
+    ) -> None:
+        """Create a SelectionPreferences object.
+
+        :param allow_yanked: Whether files marked as yanked (in the sense
+            of PEP 592) are permitted to be candidates for install.
+        :param format_control: A FormatControl object or None. Used to control
+            the selection of source packages / binary packages when consulting
+            the index and links.
+        :param prefer_binary: Whether to prefer an old, but valid, binary
+            dist over a new source dist.
+        :param ignore_requires_python: Whether to ignore incompatible
+            "Requires-Python" values in links. Defaults to False.
+        """
+        if ignore_requires_python is None:
+            ignore_requires_python = False
+
+        self.allow_yanked = allow_yanked
+        self.allow_all_prereleases = allow_all_prereleases
+        self.format_control = format_control
+        self.prefer_binary = prefer_binary
+        self.ignore_requires_python = ignore_requires_python
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/target_python.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/target_python.py
new file mode 100644
index 00000000..88925a9f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/target_python.py
@@ -0,0 +1,121 @@
+import sys
+from typing import List, Optional, Set, Tuple
+
+from pip._vendor.packaging.tags import Tag
+
+from pip._internal.utils.compatibility_tags import get_supported, version_info_to_nodot
+from pip._internal.utils.misc import normalize_version_info
+
+
+class TargetPython:
+    """
+    Encapsulates the properties of a Python interpreter one is targeting
+    for a package install, download, etc.
+    """
+
+    __slots__ = [
+        "_given_py_version_info",
+        "abis",
+        "implementation",
+        "platforms",
+        "py_version",
+        "py_version_info",
+        "_valid_tags",
+        "_valid_tags_set",
+    ]
+
+    def __init__(
+        self,
+        platforms: Optional[List[str]] = None,
+        py_version_info: Optional[Tuple[int, ...]] = None,
+        abis: Optional[List[str]] = None,
+        implementation: Optional[str] = None,
+    ) -> None:
+        """
+        :param platforms: A list of strings or None. If None, searches for
+            packages that are supported by the current system. Otherwise, will
+            find packages that can be built on the platforms passed in. These
+            packages will only be downloaded for distribution: they will
+            not be built locally.
+        :param py_version_info: An optional tuple of ints representing the
+            Python version information to use (e.g. `sys.version_info[:3]`).
+            This can have length 1, 2, or 3 when provided.
+        :param abis: A list of strings or None. This is passed to
+            compatibility_tags.py's get_supported() function as is.
+        :param implementation: A string or None. This is passed to
+            compatibility_tags.py's get_supported() function as is.
+        """
+        # Store the given py_version_info for when we call get_supported().
+        self._given_py_version_info = py_version_info
+
+        if py_version_info is None:
+            py_version_info = sys.version_info[:3]
+        else:
+            py_version_info = normalize_version_info(py_version_info)
+
+        py_version = ".".join(map(str, py_version_info[:2]))
+
+        self.abis = abis
+        self.implementation = implementation
+        self.platforms = platforms
+        self.py_version = py_version
+        self.py_version_info = py_version_info
+
+        # This is used to cache the return value of get_(un)sorted_tags.
+        self._valid_tags: Optional[List[Tag]] = None
+        self._valid_tags_set: Optional[Set[Tag]] = None
+
+    def format_given(self) -> str:
+        """
+        Format the given, non-None attributes for display.
+        """
+        display_version = None
+        if self._given_py_version_info is not None:
+            display_version = ".".join(
+                str(part) for part in self._given_py_version_info
+            )
+
+        key_values = [
+            ("platforms", self.platforms),
+            ("version_info", display_version),
+            ("abis", self.abis),
+            ("implementation", self.implementation),
+        ]
+        return " ".join(
+            f"{key}={value!r}" for key, value in key_values if value is not None
+        )
+
+    def get_sorted_tags(self) -> List[Tag]:
+        """
+        Return the supported PEP 425 tags to check wheel candidates against.
+
+        The tags are returned in order of preference (most preferred first).
+        """
+        if self._valid_tags is None:
+            # Pass versions=None if no py_version_info was given since
+            # versions=None uses special default logic.
+            py_version_info = self._given_py_version_info
+            if py_version_info is None:
+                version = None
+            else:
+                version = version_info_to_nodot(py_version_info)
+
+            tags = get_supported(
+                version=version,
+                platforms=self.platforms,
+                abis=self.abis,
+                impl=self.implementation,
+            )
+            self._valid_tags = tags
+
+        return self._valid_tags
+
+    def get_unsorted_tags(self) -> Set[Tag]:
+        """Exactly the same as get_sorted_tags, but returns a set.
+
+        This is important for performance.
+        """
+        if self._valid_tags_set is None:
+            self._valid_tags_set = set(self.get_sorted_tags())
+
+        return self._valid_tags_set
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/models/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/wheel.py
new file mode 100644
index 00000000..ea856008
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/models/wheel.py
@@ -0,0 +1,118 @@
+"""Represents a wheel file and provides access to the various parts of the
+name that have meaning.
+"""
+
+import re
+from typing import Dict, Iterable, List
+
+from pip._vendor.packaging.tags import Tag
+from pip._vendor.packaging.utils import (
+    InvalidWheelFilename as PackagingInvalidWheelName,
+)
+from pip._vendor.packaging.utils import parse_wheel_filename
+
+from pip._internal.exceptions import InvalidWheelFilename
+from pip._internal.utils.deprecation import deprecated
+
+
+class Wheel:
+    """A wheel file"""
+
+    wheel_file_re = re.compile(
+        r"""^(?P<namever>(?P<name>[^\s-]+?)-(?P<ver>[^\s-]*?))
+        ((-(?P<build>\d[^-]*?))?-(?P<pyver>[^\s-]+?)-(?P<abi>[^\s-]+?)-(?P<plat>[^\s-]+?)
+        \.whl|\.dist-info)$""",
+        re.VERBOSE,
+    )
+
+    def __init__(self, filename: str) -> None:
+        """
+        :raises InvalidWheelFilename: when the filename is invalid for a wheel
+        """
+        wheel_info = self.wheel_file_re.match(filename)
+        if not wheel_info:
+            raise InvalidWheelFilename(f"{filename} is not a valid wheel filename.")
+        self.filename = filename
+        self.name = wheel_info.group("name").replace("_", "-")
+        _version = wheel_info.group("ver")
+        if "_" in _version:
+            try:
+                parse_wheel_filename(filename)
+            except PackagingInvalidWheelName as e:
+                deprecated(
+                    reason=(
+                        f"Wheel filename {filename!r} is not correctly normalised. "
+                        "Future versions of pip will raise the following error:\n"
+                        f"{e.args[0]}\n\n"
+                    ),
+                    replacement=(
+                        "to rename the wheel to use a correctly normalised "
+                        "name (this may require updating the version in "
+                        "the project metadata)"
+                    ),
+                    gone_in="25.1",
+                    issue=12938,
+                )
+
+            _version = _version.replace("_", "-")
+
+        self.version = _version
+        self.build_tag = wheel_info.group("build")
+        self.pyversions = wheel_info.group("pyver").split(".")
+        self.abis = wheel_info.group("abi").split(".")
+        self.plats = wheel_info.group("plat").split(".")
+
+        # All the tag combinations from this file
+        self.file_tags = {
+            Tag(x, y, z) for x in self.pyversions for y in self.abis for z in self.plats
+        }
+
+    def get_formatted_file_tags(self) -> List[str]:
+        """Return the wheel's tags as a sorted list of strings."""
+        return sorted(str(tag) for tag in self.file_tags)
+
+    def support_index_min(self, tags: List[Tag]) -> int:
+        """Return the lowest index that one of the wheel's file_tag combinations
+        achieves in the given list of supported tags.
+
+        For example, if there are 8 supported tags and one of the file tags
+        is first in the list, then return 0.
+
+        :param tags: the PEP 425 tags to check the wheel against, in order
+            with most preferred first.
+
+        :raises ValueError: If none of the wheel's file tags match one of
+            the supported tags.
+        """
+        try:
+            return next(i for i, t in enumerate(tags) if t in self.file_tags)
+        except StopIteration:
+            raise ValueError()
+
+    def find_most_preferred_tag(
+        self, tags: List[Tag], tag_to_priority: Dict[Tag, int]
+    ) -> int:
+        """Return the priority of the most preferred tag that one of the wheel's file
+        tag combinations achieves in the given list of supported tags using the given
+        tag_to_priority mapping, where lower priorities are more-preferred.
+
+        This is used in place of support_index_min in some cases in order to avoid
+        an expensive linear scan of a large list of tags.
+
+        :param tags: the PEP 425 tags to check the wheel against.
+        :param tag_to_priority: a mapping from tag to priority of that tag, where
+            lower is more preferred.
+
+        :raises ValueError: If none of the wheel's file tags match one of
+            the supported tags.
+        """
+        return min(
+            tag_to_priority[tag] for tag in self.file_tags if tag in tag_to_priority
+        )
+
+    def supported(self, tags: Iterable[Tag]) -> bool:
+        """Return whether the wheel is compatible with one of the given tags.
+
+        :param tags: the PEP 425 tags to check the wheel against.
+        """
+        return not self.file_tags.isdisjoint(tags)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__init__.py
new file mode 100644
index 00000000..b51bde91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__init__.py
@@ -0,0 +1,2 @@
+"""Contains purely network-related utilities.
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..57016b91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/auth.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/auth.cpython-312.pyc
new file mode 100644
index 00000000..7b357a10
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/auth.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/cache.cpython-312.pyc
new file mode 100644
index 00000000..24bda888
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/download.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/download.cpython-312.pyc
new file mode 100644
index 00000000..6e06ef2a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/download.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/lazy_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/lazy_wheel.cpython-312.pyc
new file mode 100644
index 00000000..22fc7b04
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/lazy_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/session.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/session.cpython-312.pyc
new file mode 100644
index 00000000..b76feaad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/session.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..3fe8943a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/xmlrpc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/xmlrpc.cpython-312.pyc
new file mode 100644
index 00000000..db3108b7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/__pycache__/xmlrpc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/auth.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/auth.py
new file mode 100644
index 00000000..1a2606ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/auth.py
@@ -0,0 +1,566 @@
+"""Network Authentication Helpers
+
+Contains interface (MultiDomainBasicAuth) and associated glue code for
+providing credentials in the context of network requests.
+"""
+
+import logging
+import os
+import shutil
+import subprocess
+import sysconfig
+import typing
+import urllib.parse
+from abc import ABC, abstractmethod
+from functools import lru_cache
+from os.path import commonprefix
+from pathlib import Path
+from typing import Any, Dict, List, NamedTuple, Optional, Tuple
+
+from pip._vendor.requests.auth import AuthBase, HTTPBasicAuth
+from pip._vendor.requests.models import Request, Response
+from pip._vendor.requests.utils import get_netrc_auth
+
+from pip._internal.utils.logging import getLogger
+from pip._internal.utils.misc import (
+    ask,
+    ask_input,
+    ask_password,
+    remove_auth_from_url,
+    split_auth_netloc_from_url,
+)
+from pip._internal.vcs.versioncontrol import AuthInfo
+
+logger = getLogger(__name__)
+
+KEYRING_DISABLED = False
+
+
+class Credentials(NamedTuple):
+    url: str
+    username: str
+    password: str
+
+
+class KeyRingBaseProvider(ABC):
+    """Keyring base provider interface"""
+
+    has_keyring: bool
+
+    @abstractmethod
+    def get_auth_info(
+        self, url: str, username: Optional[str]
+    ) -> Optional[AuthInfo]: ...
+
+    @abstractmethod
+    def save_auth_info(self, url: str, username: str, password: str) -> None: ...
+
+
+class KeyRingNullProvider(KeyRingBaseProvider):
+    """Keyring null provider"""
+
+    has_keyring = False
+
+    def get_auth_info(self, url: str, username: Optional[str]) -> Optional[AuthInfo]:
+        return None
+
+    def save_auth_info(self, url: str, username: str, password: str) -> None:
+        return None
+
+
+class KeyRingPythonProvider(KeyRingBaseProvider):
+    """Keyring interface which uses locally imported `keyring`"""
+
+    has_keyring = True
+
+    def __init__(self) -> None:
+        import keyring
+
+        self.keyring = keyring
+
+    def get_auth_info(self, url: str, username: Optional[str]) -> Optional[AuthInfo]:
+        # Support keyring's get_credential interface which supports getting
+        # credentials without a username. This is only available for
+        # keyring>=15.2.0.
+        if hasattr(self.keyring, "get_credential"):
+            logger.debug("Getting credentials from keyring for %s", url)
+            cred = self.keyring.get_credential(url, username)
+            if cred is not None:
+                return cred.username, cred.password
+            return None
+
+        if username is not None:
+            logger.debug("Getting password from keyring for %s", url)
+            password = self.keyring.get_password(url, username)
+            if password:
+                return username, password
+        return None
+
+    def save_auth_info(self, url: str, username: str, password: str) -> None:
+        self.keyring.set_password(url, username, password)
+
+
+class KeyRingCliProvider(KeyRingBaseProvider):
+    """Provider which uses `keyring` cli
+
+    Instead of calling the keyring package installed alongside pip
+    we call keyring on the command line which will enable pip to
+    use which ever installation of keyring is available first in
+    PATH.
+    """
+
+    has_keyring = True
+
+    def __init__(self, cmd: str) -> None:
+        self.keyring = cmd
+
+    def get_auth_info(self, url: str, username: Optional[str]) -> Optional[AuthInfo]:
+        # This is the default implementation of keyring.get_credential
+        # https://github.com/jaraco/keyring/blob/97689324abcf01bd1793d49063e7ca01e03d7d07/keyring/backend.py#L134-L139
+        if username is not None:
+            password = self._get_password(url, username)
+            if password is not None:
+                return username, password
+        return None
+
+    def save_auth_info(self, url: str, username: str, password: str) -> None:
+        return self._set_password(url, username, password)
+
+    def _get_password(self, service_name: str, username: str) -> Optional[str]:
+        """Mirror the implementation of keyring.get_password using cli"""
+        if self.keyring is None:
+            return None
+
+        cmd = [self.keyring, "get", service_name, username]
+        env = os.environ.copy()
+        env["PYTHONIOENCODING"] = "utf-8"
+        res = subprocess.run(
+            cmd,
+            stdin=subprocess.DEVNULL,
+            stdout=subprocess.PIPE,
+            env=env,
+        )
+        if res.returncode:
+            return None
+        return res.stdout.decode("utf-8").strip(os.linesep)
+
+    def _set_password(self, service_name: str, username: str, password: str) -> None:
+        """Mirror the implementation of keyring.set_password using cli"""
+        if self.keyring is None:
+            return None
+        env = os.environ.copy()
+        env["PYTHONIOENCODING"] = "utf-8"
+        subprocess.run(
+            [self.keyring, "set", service_name, username],
+            input=f"{password}{os.linesep}".encode(),
+            env=env,
+            check=True,
+        )
+        return None
+
+
+@lru_cache(maxsize=None)
+def get_keyring_provider(provider: str) -> KeyRingBaseProvider:
+    logger.verbose("Keyring provider requested: %s", provider)
+
+    # keyring has previously failed and been disabled
+    if KEYRING_DISABLED:
+        provider = "disabled"
+    if provider in ["import", "auto"]:
+        try:
+            impl = KeyRingPythonProvider()
+            logger.verbose("Keyring provider set: import")
+            return impl
+        except ImportError:
+            pass
+        except Exception as exc:
+            # In the event of an unexpected exception
+            # we should warn the user
+            msg = "Installed copy of keyring fails with exception %s"
+            if provider == "auto":
+                msg = msg + ", trying to find a keyring executable as a fallback"
+            logger.warning(msg, exc, exc_info=logger.isEnabledFor(logging.DEBUG))
+    if provider in ["subprocess", "auto"]:
+        cli = shutil.which("keyring")
+        if cli and cli.startswith(sysconfig.get_path("scripts")):
+            # all code within this function is stolen from shutil.which implementation
+            @typing.no_type_check
+            def PATH_as_shutil_which_determines_it() -> str:
+                path = os.environ.get("PATH", None)
+                if path is None:
+                    try:
+                        path = os.confstr("CS_PATH")
+                    except (AttributeError, ValueError):
+                        # os.confstr() or CS_PATH is not available
+                        path = os.defpath
+                # bpo-35755: Don't use os.defpath if the PATH environment variable is
+                # set to an empty string
+
+                return path
+
+            scripts = Path(sysconfig.get_path("scripts"))
+
+            paths = []
+            for path in PATH_as_shutil_which_determines_it().split(os.pathsep):
+                p = Path(path)
+                try:
+                    if not p.samefile(scripts):
+                        paths.append(path)
+                except FileNotFoundError:
+                    pass
+
+            path = os.pathsep.join(paths)
+
+            cli = shutil.which("keyring", path=path)
+
+        if cli:
+            logger.verbose("Keyring provider set: subprocess with executable %s", cli)
+            return KeyRingCliProvider(cli)
+
+    logger.verbose("Keyring provider set: disabled")
+    return KeyRingNullProvider()
+
+
+class MultiDomainBasicAuth(AuthBase):
+    def __init__(
+        self,
+        prompting: bool = True,
+        index_urls: Optional[List[str]] = None,
+        keyring_provider: str = "auto",
+    ) -> None:
+        self.prompting = prompting
+        self.index_urls = index_urls
+        self.keyring_provider = keyring_provider  # type: ignore[assignment]
+        self.passwords: Dict[str, AuthInfo] = {}
+        # When the user is prompted to enter credentials and keyring is
+        # available, we will offer to save them. If the user accepts,
+        # this value is set to the credentials they entered. After the
+        # request authenticates, the caller should call
+        # ``save_credentials`` to save these.
+        self._credentials_to_save: Optional[Credentials] = None
+
+    @property
+    def keyring_provider(self) -> KeyRingBaseProvider:
+        return get_keyring_provider(self._keyring_provider)
+
+    @keyring_provider.setter
+    def keyring_provider(self, provider: str) -> None:
+        # The free function get_keyring_provider has been decorated with
+        # functools.cache. If an exception occurs in get_keyring_auth that
+        # cache will be cleared and keyring disabled, take that into account
+        # if you want to remove this indirection.
+        self._keyring_provider = provider
+
+    @property
+    def use_keyring(self) -> bool:
+        # We won't use keyring when --no-input is passed unless
+        # a specific provider is requested because it might require
+        # user interaction
+        return self.prompting or self._keyring_provider not in ["auto", "disabled"]
+
+    def _get_keyring_auth(
+        self,
+        url: Optional[str],
+        username: Optional[str],
+    ) -> Optional[AuthInfo]:
+        """Return the tuple auth for a given url from keyring."""
+        # Do nothing if no url was provided
+        if not url:
+            return None
+
+        try:
+            return self.keyring_provider.get_auth_info(url, username)
+        except Exception as exc:
+            # Log the full exception (with stacktrace) at debug, so it'll only
+            # show up when running in verbose mode.
+            logger.debug("Keyring is skipped due to an exception", exc_info=True)
+            # Always log a shortened version of the exception.
+            logger.warning(
+                "Keyring is skipped due to an exception: %s",
+                str(exc),
+            )
+            global KEYRING_DISABLED
+            KEYRING_DISABLED = True
+            get_keyring_provider.cache_clear()
+            return None
+
+    def _get_index_url(self, url: str) -> Optional[str]:
+        """Return the original index URL matching the requested URL.
+
+        Cached or dynamically generated credentials may work against
+        the original index URL rather than just the netloc.
+
+        The provided url should have had its username and password
+        removed already. If the original index url had credentials then
+        they will be included in the return value.
+
+        Returns None if no matching index was found, or if --no-index
+        was specified by the user.
+        """
+        if not url or not self.index_urls:
+            return None
+
+        url = remove_auth_from_url(url).rstrip("/") + "/"
+        parsed_url = urllib.parse.urlsplit(url)
+
+        candidates = []
+
+        for index in self.index_urls:
+            index = index.rstrip("/") + "/"
+            parsed_index = urllib.parse.urlsplit(remove_auth_from_url(index))
+            if parsed_url == parsed_index:
+                return index
+
+            if parsed_url.netloc != parsed_index.netloc:
+                continue
+
+            candidate = urllib.parse.urlsplit(index)
+            candidates.append(candidate)
+
+        if not candidates:
+            return None
+
+        candidates.sort(
+            reverse=True,
+            key=lambda candidate: commonprefix(
+                [
+                    parsed_url.path,
+                    candidate.path,
+                ]
+            ).rfind("/"),
+        )
+
+        return urllib.parse.urlunsplit(candidates[0])
+
+    def _get_new_credentials(
+        self,
+        original_url: str,
+        *,
+        allow_netrc: bool = True,
+        allow_keyring: bool = False,
+    ) -> AuthInfo:
+        """Find and return credentials for the specified URL."""
+        # Split the credentials and netloc from the url.
+        url, netloc, url_user_password = split_auth_netloc_from_url(
+            original_url,
+        )
+
+        # Start with the credentials embedded in the url
+        username, password = url_user_password
+        if username is not None and password is not None:
+            logger.debug("Found credentials in url for %s", netloc)
+            return url_user_password
+
+        # Find a matching index url for this request
+        index_url = self._get_index_url(url)
+        if index_url:
+            # Split the credentials from the url.
+            index_info = split_auth_netloc_from_url(index_url)
+            if index_info:
+                index_url, _, index_url_user_password = index_info
+                logger.debug("Found index url %s", index_url)
+
+        # If an index URL was found, try its embedded credentials
+        if index_url and index_url_user_password[0] is not None:
+            username, password = index_url_user_password
+            if username is not None and password is not None:
+                logger.debug("Found credentials in index url for %s", netloc)
+                return index_url_user_password
+
+        # Get creds from netrc if we still don't have them
+        if allow_netrc:
+            netrc_auth = get_netrc_auth(original_url)
+            if netrc_auth:
+                logger.debug("Found credentials in netrc for %s", netloc)
+                return netrc_auth
+
+        # If we don't have a password and keyring is available, use it.
+        if allow_keyring:
+            # The index url is more specific than the netloc, so try it first
+            # fmt: off
+            kr_auth = (
+                self._get_keyring_auth(index_url, username) or
+                self._get_keyring_auth(netloc, username)
+            )
+            # fmt: on
+            if kr_auth:
+                logger.debug("Found credentials in keyring for %s", netloc)
+                return kr_auth
+
+        return username, password
+
+    def _get_url_and_credentials(
+        self, original_url: str
+    ) -> Tuple[str, Optional[str], Optional[str]]:
+        """Return the credentials to use for the provided URL.
+
+        If allowed, netrc and keyring may be used to obtain the
+        correct credentials.
+
+        Returns (url_without_credentials, username, password). Note
+        that even if the original URL contains credentials, this
+        function may return a different username and password.
+        """
+        url, netloc, _ = split_auth_netloc_from_url(original_url)
+
+        # Try to get credentials from original url
+        username, password = self._get_new_credentials(original_url)
+
+        # If credentials not found, use any stored credentials for this netloc.
+        # Do this if either the username or the password is missing.
+        # This accounts for the situation in which the user has specified
+        # the username in the index url, but the password comes from keyring.
+        if (username is None or password is None) and netloc in self.passwords:
+            un, pw = self.passwords[netloc]
+            # It is possible that the cached credentials are for a different username,
+            # in which case the cache should be ignored.
+            if username is None or username == un:
+                username, password = un, pw
+
+        if username is not None or password is not None:
+            # Convert the username and password if they're None, so that
+            # this netloc will show up as "cached" in the conditional above.
+            # Further, HTTPBasicAuth doesn't accept None, so it makes sense to
+            # cache the value that is going to be used.
+            username = username or ""
+            password = password or ""
+
+            # Store any acquired credentials.
+            self.passwords[netloc] = (username, password)
+
+        assert (
+            # Credentials were found
+            (username is not None and password is not None)
+            # Credentials were not found
+            or (username is None and password is None)
+        ), f"Could not load credentials from url: {original_url}"
+
+        return url, username, password
+
+    def __call__(self, req: Request) -> Request:
+        # Get credentials for this request
+        url, username, password = self._get_url_and_credentials(req.url)
+
+        # Set the url of the request to the url without any credentials
+        req.url = url
+
+        if username is not None and password is not None:
+            # Send the basic auth with this request
+            req = HTTPBasicAuth(username, password)(req)
+
+        # Attach a hook to handle 401 responses
+        req.register_hook("response", self.handle_401)
+
+        return req
+
+    # Factored out to allow for easy patching in tests
+    def _prompt_for_password(
+        self, netloc: str
+    ) -> Tuple[Optional[str], Optional[str], bool]:
+        username = ask_input(f"User for {netloc}: ") if self.prompting else None
+        if not username:
+            return None, None, False
+        if self.use_keyring:
+            auth = self._get_keyring_auth(netloc, username)
+            if auth and auth[0] is not None and auth[1] is not None:
+                return auth[0], auth[1], False
+        password = ask_password("Password: ")
+        return username, password, True
+
+    # Factored out to allow for easy patching in tests
+    def _should_save_password_to_keyring(self) -> bool:
+        if (
+            not self.prompting
+            or not self.use_keyring
+            or not self.keyring_provider.has_keyring
+        ):
+            return False
+        return ask("Save credentials to keyring [y/N]: ", ["y", "n"]) == "y"
+
+    def handle_401(self, resp: Response, **kwargs: Any) -> Response:
+        # We only care about 401 responses, anything else we want to just
+        #   pass through the actual response
+        if resp.status_code != 401:
+            return resp
+
+        username, password = None, None
+
+        # Query the keyring for credentials:
+        if self.use_keyring:
+            username, password = self._get_new_credentials(
+                resp.url,
+                allow_netrc=False,
+                allow_keyring=True,
+            )
+
+        # We are not able to prompt the user so simply return the response
+        if not self.prompting and not username and not password:
+            return resp
+
+        parsed = urllib.parse.urlparse(resp.url)
+
+        # Prompt the user for a new username and password
+        save = False
+        if not username and not password:
+            username, password, save = self._prompt_for_password(parsed.netloc)
+
+        # Store the new username and password to use for future requests
+        self._credentials_to_save = None
+        if username is not None and password is not None:
+            self.passwords[parsed.netloc] = (username, password)
+
+            # Prompt to save the password to keyring
+            if save and self._should_save_password_to_keyring():
+                self._credentials_to_save = Credentials(
+                    url=parsed.netloc,
+                    username=username,
+                    password=password,
+                )
+
+        # Consume content and release the original connection to allow our new
+        #   request to reuse the same one.
+        # The result of the assignment isn't used, it's just needed to consume
+        # the content.
+        _ = resp.content
+        resp.raw.release_conn()
+
+        # Add our new username and password to the request
+        req = HTTPBasicAuth(username or "", password or "")(resp.request)
+        req.register_hook("response", self.warn_on_401)
+
+        # On successful request, save the credentials that were used to
+        # keyring. (Note that if the user responded "no" above, this member
+        # is not set and nothing will be saved.)
+        if self._credentials_to_save:
+            req.register_hook("response", self.save_credentials)
+
+        # Send our new request
+        new_resp = resp.connection.send(req, **kwargs)
+        new_resp.history.append(resp)
+
+        return new_resp
+
+    def warn_on_401(self, resp: Response, **kwargs: Any) -> None:
+        """Response callback to warn about incorrect credentials."""
+        if resp.status_code == 401:
+            logger.warning(
+                "401 Error, Credentials not correct for %s",
+                resp.request.url,
+            )
+
+    def save_credentials(self, resp: Response, **kwargs: Any) -> None:
+        """Response callback to save credentials on success."""
+        assert (
+            self.keyring_provider.has_keyring
+        ), "should never reach here without keyring"
+
+        creds = self._credentials_to_save
+        self._credentials_to_save = None
+        if creds and resp.status_code < 400:
+            try:
+                logger.info("Saving credentials to keyring")
+                self.keyring_provider.save_auth_info(
+                    creds.url, creds.username, creds.password
+                )
+            except Exception:
+                logger.exception("Failed to save credentials")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/cache.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/cache.py
new file mode 100644
index 00000000..fca04e69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/cache.py
@@ -0,0 +1,118 @@
+"""HTTP cache implementation.
+"""
+
+import os
+from contextlib import contextmanager
+from datetime import datetime
+from typing import BinaryIO, Generator, Optional, Union
+
+from pip._vendor.cachecontrol.cache import SeparateBodyBaseCache
+from pip._vendor.cachecontrol.caches import SeparateBodyFileCache
+from pip._vendor.requests.models import Response
+
+from pip._internal.utils.filesystem import adjacent_tmp_file, replace
+from pip._internal.utils.misc import ensure_dir
+
+
+def is_from_cache(response: Response) -> bool:
+    return getattr(response, "from_cache", False)
+
+
+@contextmanager
+def suppressed_cache_errors() -> Generator[None, None, None]:
+    """If we can't access the cache then we can just skip caching and process
+    requests as if caching wasn't enabled.
+    """
+    try:
+        yield
+    except OSError:
+        pass
+
+
+class SafeFileCache(SeparateBodyBaseCache):
+    """
+    A file based cache which is safe to use even when the target directory may
+    not be accessible or writable.
+
+    There is a race condition when two processes try to write and/or read the
+    same entry at the same time, since each entry consists of two separate
+    files (https://github.com/psf/cachecontrol/issues/324).  We therefore have
+    additional logic that makes sure that both files to be present before
+    returning an entry; this fixes the read side of the race condition.
+
+    For the write side, we assume that the server will only ever return the
+    same data for the same URL, which ought to be the case for files pip is
+    downloading.  PyPI does not have a mechanism to swap out a wheel for
+    another wheel, for example.  If this assumption is not true, the
+    CacheControl issue will need to be fixed.
+    """
+
+    def __init__(self, directory: str) -> None:
+        assert directory is not None, "Cache directory must not be None."
+        super().__init__()
+        self.directory = directory
+
+    def _get_cache_path(self, name: str) -> str:
+        # From cachecontrol.caches.file_cache.FileCache._fn, brought into our
+        # class for backwards-compatibility and to avoid using a non-public
+        # method.
+        hashed = SeparateBodyFileCache.encode(name)
+        parts = list(hashed[:5]) + [hashed]
+        return os.path.join(self.directory, *parts)
+
+    def get(self, key: str) -> Optional[bytes]:
+        # The cache entry is only valid if both metadata and body exist.
+        metadata_path = self._get_cache_path(key)
+        body_path = metadata_path + ".body"
+        if not (os.path.exists(metadata_path) and os.path.exists(body_path)):
+            return None
+        with suppressed_cache_errors():
+            with open(metadata_path, "rb") as f:
+                return f.read()
+
+    def _write(self, path: str, data: bytes) -> None:
+        with suppressed_cache_errors():
+            ensure_dir(os.path.dirname(path))
+
+            with adjacent_tmp_file(path) as f:
+                f.write(data)
+                # Inherit the read/write permissions of the cache directory
+                # to enable multi-user cache use-cases.
+                mode = (
+                    os.stat(self.directory).st_mode
+                    & 0o666  # select read/write permissions of cache directory
+                    | 0o600  # set owner read/write permissions
+                )
+                # Change permissions only if there is no risk of following a symlink.
+                if os.chmod in os.supports_fd:
+                    os.chmod(f.fileno(), mode)
+                elif os.chmod in os.supports_follow_symlinks:
+                    os.chmod(f.name, mode, follow_symlinks=False)
+
+            replace(f.name, path)
+
+    def set(
+        self, key: str, value: bytes, expires: Union[int, datetime, None] = None
+    ) -> None:
+        path = self._get_cache_path(key)
+        self._write(path, value)
+
+    def delete(self, key: str) -> None:
+        path = self._get_cache_path(key)
+        with suppressed_cache_errors():
+            os.remove(path)
+        with suppressed_cache_errors():
+            os.remove(path + ".body")
+
+    def get_body(self, key: str) -> Optional[BinaryIO]:
+        # The cache entry is only valid if both metadata and body exist.
+        metadata_path = self._get_cache_path(key)
+        body_path = metadata_path + ".body"
+        if not (os.path.exists(metadata_path) and os.path.exists(body_path)):
+            return None
+        with suppressed_cache_errors():
+            return open(body_path, "rb")
+
+    def set_body(self, key: str, body: bytes) -> None:
+        path = self._get_cache_path(key) + ".body"
+        self._write(path, body)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/download.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/download.py
new file mode 100644
index 00000000..5c3bce3d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/download.py
@@ -0,0 +1,187 @@
+"""Download files with progress indicators.
+"""
+
+import email.message
+import logging
+import mimetypes
+import os
+from typing import Iterable, Optional, Tuple
+
+from pip._vendor.requests.models import Response
+
+from pip._internal.cli.progress_bars import get_download_progress_renderer
+from pip._internal.exceptions import NetworkConnectionError
+from pip._internal.models.index import PyPI
+from pip._internal.models.link import Link
+from pip._internal.network.cache import is_from_cache
+from pip._internal.network.session import PipSession
+from pip._internal.network.utils import HEADERS, raise_for_status, response_chunks
+from pip._internal.utils.misc import format_size, redact_auth_from_url, splitext
+
+logger = logging.getLogger(__name__)
+
+
+def _get_http_response_size(resp: Response) -> Optional[int]:
+    try:
+        return int(resp.headers["content-length"])
+    except (ValueError, KeyError, TypeError):
+        return None
+
+
+def _prepare_download(
+    resp: Response,
+    link: Link,
+    progress_bar: str,
+) -> Iterable[bytes]:
+    total_length = _get_http_response_size(resp)
+
+    if link.netloc == PyPI.file_storage_domain:
+        url = link.show_url
+    else:
+        url = link.url_without_fragment
+
+    logged_url = redact_auth_from_url(url)
+
+    if total_length:
+        logged_url = f"{logged_url} ({format_size(total_length)})"
+
+    if is_from_cache(resp):
+        logger.info("Using cached %s", logged_url)
+    else:
+        logger.info("Downloading %s", logged_url)
+
+    if logger.getEffectiveLevel() > logging.INFO:
+        show_progress = False
+    elif is_from_cache(resp):
+        show_progress = False
+    elif not total_length:
+        show_progress = True
+    elif total_length > (512 * 1024):
+        show_progress = True
+    else:
+        show_progress = False
+
+    chunks = response_chunks(resp)
+
+    if not show_progress:
+        return chunks
+
+    renderer = get_download_progress_renderer(bar_type=progress_bar, size=total_length)
+    return renderer(chunks)
+
+
+def sanitize_content_filename(filename: str) -> str:
+    """
+    Sanitize the "filename" value from a Content-Disposition header.
+    """
+    return os.path.basename(filename)
+
+
+def parse_content_disposition(content_disposition: str, default_filename: str) -> str:
+    """
+    Parse the "filename" value from a Content-Disposition header, and
+    return the default filename if the result is empty.
+    """
+    m = email.message.Message()
+    m["content-type"] = content_disposition
+    filename = m.get_param("filename")
+    if filename:
+        # We need to sanitize the filename to prevent directory traversal
+        # in case the filename contains ".." path parts.
+        filename = sanitize_content_filename(str(filename))
+    return filename or default_filename
+
+
+def _get_http_response_filename(resp: Response, link: Link) -> str:
+    """Get an ideal filename from the given HTTP response, falling back to
+    the link filename if not provided.
+    """
+    filename = link.filename  # fallback
+    # Have a look at the Content-Disposition header for a better guess
+    content_disposition = resp.headers.get("content-disposition")
+    if content_disposition:
+        filename = parse_content_disposition(content_disposition, filename)
+    ext: Optional[str] = splitext(filename)[1]
+    if not ext:
+        ext = mimetypes.guess_extension(resp.headers.get("content-type", ""))
+        if ext:
+            filename += ext
+    if not ext and link.url != resp.url:
+        ext = os.path.splitext(resp.url)[1]
+        if ext:
+            filename += ext
+    return filename
+
+
+def _http_get_download(session: PipSession, link: Link) -> Response:
+    target_url = link.url.split("#", 1)[0]
+    resp = session.get(target_url, headers=HEADERS, stream=True)
+    raise_for_status(resp)
+    return resp
+
+
+class Downloader:
+    def __init__(
+        self,
+        session: PipSession,
+        progress_bar: str,
+    ) -> None:
+        self._session = session
+        self._progress_bar = progress_bar
+
+    def __call__(self, link: Link, location: str) -> Tuple[str, str]:
+        """Download the file given by link into location."""
+        try:
+            resp = _http_get_download(self._session, link)
+        except NetworkConnectionError as e:
+            assert e.response is not None
+            logger.critical(
+                "HTTP error %s while getting %s", e.response.status_code, link
+            )
+            raise
+
+        filename = _get_http_response_filename(resp, link)
+        filepath = os.path.join(location, filename)
+
+        chunks = _prepare_download(resp, link, self._progress_bar)
+        with open(filepath, "wb") as content_file:
+            for chunk in chunks:
+                content_file.write(chunk)
+        content_type = resp.headers.get("Content-Type", "")
+        return filepath, content_type
+
+
+class BatchDownloader:
+    def __init__(
+        self,
+        session: PipSession,
+        progress_bar: str,
+    ) -> None:
+        self._session = session
+        self._progress_bar = progress_bar
+
+    def __call__(
+        self, links: Iterable[Link], location: str
+    ) -> Iterable[Tuple[Link, Tuple[str, str]]]:
+        """Download the files given by links into location."""
+        for link in links:
+            try:
+                resp = _http_get_download(self._session, link)
+            except NetworkConnectionError as e:
+                assert e.response is not None
+                logger.critical(
+                    "HTTP error %s while getting %s",
+                    e.response.status_code,
+                    link,
+                )
+                raise
+
+            filename = _get_http_response_filename(resp, link)
+            filepath = os.path.join(location, filename)
+
+            chunks = _prepare_download(resp, link, self._progress_bar)
+            with open(filepath, "wb") as content_file:
+                for chunk in chunks:
+                    content_file.write(chunk)
+            content_type = resp.headers.get("Content-Type", "")
+            yield link, (filepath, content_type)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/lazy_wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/lazy_wheel.py
new file mode 100644
index 00000000..03f883c1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/lazy_wheel.py
@@ -0,0 +1,210 @@
+"""Lazy ZIP over HTTP"""
+
+__all__ = ["HTTPRangeRequestUnsupported", "dist_from_wheel_url"]
+
+from bisect import bisect_left, bisect_right
+from contextlib import contextmanager
+from tempfile import NamedTemporaryFile
+from typing import Any, Dict, Generator, List, Optional, Tuple
+from zipfile import BadZipFile, ZipFile
+
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.requests.models import CONTENT_CHUNK_SIZE, Response
+
+from pip._internal.metadata import BaseDistribution, MemoryWheel, get_wheel_distribution
+from pip._internal.network.session import PipSession
+from pip._internal.network.utils import HEADERS, raise_for_status, response_chunks
+
+
+class HTTPRangeRequestUnsupported(Exception):
+    pass
+
+
+def dist_from_wheel_url(name: str, url: str, session: PipSession) -> BaseDistribution:
+    """Return a distribution object from the given wheel URL.
+
+    This uses HTTP range requests to only fetch the portion of the wheel
+    containing metadata, just enough for the object to be constructed.
+    If such requests are not supported, HTTPRangeRequestUnsupported
+    is raised.
+    """
+    with LazyZipOverHTTP(url, session) as zf:
+        # For read-only ZIP files, ZipFile only needs methods read,
+        # seek, seekable and tell, not the whole IO protocol.
+        wheel = MemoryWheel(zf.name, zf)  # type: ignore
+        # After context manager exit, wheel.name
+        # is an invalid file by intention.
+        return get_wheel_distribution(wheel, canonicalize_name(name))
+
+
+class LazyZipOverHTTP:
+    """File-like object mapped to a ZIP file over HTTP.
+
+    This uses HTTP range requests to lazily fetch the file's content,
+    which is supposed to be fed to ZipFile.  If such requests are not
+    supported by the server, raise HTTPRangeRequestUnsupported
+    during initialization.
+    """
+
+    def __init__(
+        self, url: str, session: PipSession, chunk_size: int = CONTENT_CHUNK_SIZE
+    ) -> None:
+        head = session.head(url, headers=HEADERS)
+        raise_for_status(head)
+        assert head.status_code == 200
+        self._session, self._url, self._chunk_size = session, url, chunk_size
+        self._length = int(head.headers["Content-Length"])
+        self._file = NamedTemporaryFile()
+        self.truncate(self._length)
+        self._left: List[int] = []
+        self._right: List[int] = []
+        if "bytes" not in head.headers.get("Accept-Ranges", "none"):
+            raise HTTPRangeRequestUnsupported("range request is not supported")
+        self._check_zip()
+
+    @property
+    def mode(self) -> str:
+        """Opening mode, which is always rb."""
+        return "rb"
+
+    @property
+    def name(self) -> str:
+        """Path to the underlying file."""
+        return self._file.name
+
+    def seekable(self) -> bool:
+        """Return whether random access is supported, which is True."""
+        return True
+
+    def close(self) -> None:
+        """Close the file."""
+        self._file.close()
+
+    @property
+    def closed(self) -> bool:
+        """Whether the file is closed."""
+        return self._file.closed
+
+    def read(self, size: int = -1) -> bytes:
+        """Read up to size bytes from the object and return them.
+
+        As a convenience, if size is unspecified or -1,
+        all bytes until EOF are returned.  Fewer than
+        size bytes may be returned if EOF is reached.
+        """
+        download_size = max(size, self._chunk_size)
+        start, length = self.tell(), self._length
+        stop = length if size < 0 else min(start + download_size, length)
+        start = max(0, stop - download_size)
+        self._download(start, stop - 1)
+        return self._file.read(size)
+
+    def readable(self) -> bool:
+        """Return whether the file is readable, which is True."""
+        return True
+
+    def seek(self, offset: int, whence: int = 0) -> int:
+        """Change stream position and return the new absolute position.
+
+        Seek to offset relative position indicated by whence:
+        * 0: Start of stream (the default).  pos should be >= 0;
+        * 1: Current position - pos may be negative;
+        * 2: End of stream - pos usually negative.
+        """
+        return self._file.seek(offset, whence)
+
+    def tell(self) -> int:
+        """Return the current position."""
+        return self._file.tell()
+
+    def truncate(self, size: Optional[int] = None) -> int:
+        """Resize the stream to the given size in bytes.
+
+        If size is unspecified resize to the current position.
+        The current stream position isn't changed.
+
+        Return the new file size.
+        """
+        return self._file.truncate(size)
+
+    def writable(self) -> bool:
+        """Return False."""
+        return False
+
+    def __enter__(self) -> "LazyZipOverHTTP":
+        self._file.__enter__()
+        return self
+
+    def __exit__(self, *exc: Any) -> None:
+        self._file.__exit__(*exc)
+
+    @contextmanager
+    def _stay(self) -> Generator[None, None, None]:
+        """Return a context manager keeping the position.
+
+        At the end of the block, seek back to original position.
+        """
+        pos = self.tell()
+        try:
+            yield
+        finally:
+            self.seek(pos)
+
+    def _check_zip(self) -> None:
+        """Check and download until the file is a valid ZIP."""
+        end = self._length - 1
+        for start in reversed(range(0, end, self._chunk_size)):
+            self._download(start, end)
+            with self._stay():
+                try:
+                    # For read-only ZIP files, ZipFile only needs
+                    # methods read, seek, seekable and tell.
+                    ZipFile(self)
+                except BadZipFile:
+                    pass
+                else:
+                    break
+
+    def _stream_response(
+        self, start: int, end: int, base_headers: Dict[str, str] = HEADERS
+    ) -> Response:
+        """Return HTTP response to a range request from start to end."""
+        headers = base_headers.copy()
+        headers["Range"] = f"bytes={start}-{end}"
+        # TODO: Get range requests to be correctly cached
+        headers["Cache-Control"] = "no-cache"
+        return self._session.get(self._url, headers=headers, stream=True)
+
+    def _merge(
+        self, start: int, end: int, left: int, right: int
+    ) -> Generator[Tuple[int, int], None, None]:
+        """Return a generator of intervals to be fetched.
+
+        Args:
+            start (int): Start of needed interval
+            end (int): End of needed interval
+            left (int): Index of first overlapping downloaded data
+            right (int): Index after last overlapping downloaded data
+        """
+        lslice, rslice = self._left[left:right], self._right[left:right]
+        i = start = min([start] + lslice[:1])
+        end = max([end] + rslice[-1:])
+        for j, k in zip(lslice, rslice):
+            if j > i:
+                yield i, j - 1
+            i = k + 1
+        if i <= end:
+            yield i, end
+        self._left[left:right], self._right[left:right] = [start], [end]
+
+    def _download(self, start: int, end: int) -> None:
+        """Download bytes from start to end inclusively."""
+        with self._stay():
+            left = bisect_left(self._right, start)
+            right = bisect_right(self._left, end)
+            for start, end in self._merge(start, end, left, right):
+                response = self._stream_response(start, end)
+                response.raise_for_status()
+                self.seek(start)
+                for chunk in response_chunks(response, self._chunk_size):
+                    self._file.write(chunk)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/session.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/session.py
new file mode 100644
index 00000000..5e10f8f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/session.py
@@ -0,0 +1,523 @@
+"""PipSession and supporting code, containing all pip-specific
+network request configuration and behavior.
+"""
+
+import email.utils
+import functools
+import io
+import ipaddress
+import json
+import logging
+import mimetypes
+import os
+import platform
+import shutil
+import subprocess
+import sys
+import urllib.parse
+import warnings
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    Generator,
+    List,
+    Mapping,
+    Optional,
+    Sequence,
+    Tuple,
+    Union,
+)
+
+from pip._vendor import requests, urllib3
+from pip._vendor.cachecontrol import CacheControlAdapter as _BaseCacheControlAdapter
+from pip._vendor.requests.adapters import DEFAULT_POOLBLOCK, BaseAdapter
+from pip._vendor.requests.adapters import HTTPAdapter as _BaseHTTPAdapter
+from pip._vendor.requests.models import PreparedRequest, Response
+from pip._vendor.requests.structures import CaseInsensitiveDict
+from pip._vendor.urllib3.connectionpool import ConnectionPool
+from pip._vendor.urllib3.exceptions import InsecureRequestWarning
+
+from pip import __version__
+from pip._internal.metadata import get_default_environment
+from pip._internal.models.link import Link
+from pip._internal.network.auth import MultiDomainBasicAuth
+from pip._internal.network.cache import SafeFileCache
+
+# Import ssl from compat so the initial import occurs in only one place.
+from pip._internal.utils.compat import has_tls
+from pip._internal.utils.glibc import libc_ver
+from pip._internal.utils.misc import build_url_from_netloc, parse_netloc
+from pip._internal.utils.urls import url_to_path
+
+if TYPE_CHECKING:
+    from ssl import SSLContext
+
+    from pip._vendor.urllib3.poolmanager import PoolManager
+
+
+logger = logging.getLogger(__name__)
+
+SecureOrigin = Tuple[str, str, Optional[Union[int, str]]]
+
+
+# Ignore warning raised when using --trusted-host.
+warnings.filterwarnings("ignore", category=InsecureRequestWarning)
+
+
+SECURE_ORIGINS: List[SecureOrigin] = [
+    # protocol, hostname, port
+    # Taken from Chrome's list of secure origins (See: http://bit.ly/1qrySKC)
+    ("https", "*", "*"),
+    ("*", "localhost", "*"),
+    ("*", "127.0.0.0/8", "*"),
+    ("*", "::1/128", "*"),
+    ("file", "*", None),
+    # ssh is always secure.
+    ("ssh", "*", "*"),
+]
+
+
+# These are environment variables present when running under various
+# CI systems.  For each variable, some CI systems that use the variable
+# are indicated.  The collection was chosen so that for each of a number
+# of popular systems, at least one of the environment variables is used.
+# This list is used to provide some indication of and lower bound for
+# CI traffic to PyPI.  Thus, it is okay if the list is not comprehensive.
+# For more background, see: https://github.com/pypa/pip/issues/5499
+CI_ENVIRONMENT_VARIABLES = (
+    # Azure Pipelines
+    "BUILD_BUILDID",
+    # Jenkins
+    "BUILD_ID",
+    # AppVeyor, CircleCI, Codeship, Gitlab CI, Shippable, Travis CI
+    "CI",
+    # Explicit environment variable.
+    "PIP_IS_CI",
+)
+
+
+def looks_like_ci() -> bool:
+    """
+    Return whether it looks like pip is running under CI.
+    """
+    # We don't use the method of checking for a tty (e.g. using isatty())
+    # because some CI systems mimic a tty (e.g. Travis CI).  Thus that
+    # method doesn't provide definitive information in either direction.
+    return any(name in os.environ for name in CI_ENVIRONMENT_VARIABLES)
+
+
+@functools.lru_cache(maxsize=1)
+def user_agent() -> str:
+    """
+    Return a string representing the user agent.
+    """
+    data: Dict[str, Any] = {
+        "installer": {"name": "pip", "version": __version__},
+        "python": platform.python_version(),
+        "implementation": {
+            "name": platform.python_implementation(),
+        },
+    }
+
+    if data["implementation"]["name"] == "CPython":
+        data["implementation"]["version"] = platform.python_version()
+    elif data["implementation"]["name"] == "PyPy":
+        pypy_version_info = sys.pypy_version_info  # type: ignore
+        if pypy_version_info.releaselevel == "final":
+            pypy_version_info = pypy_version_info[:3]
+        data["implementation"]["version"] = ".".join(
+            [str(x) for x in pypy_version_info]
+        )
+    elif data["implementation"]["name"] == "Jython":
+        # Complete Guess
+        data["implementation"]["version"] = platform.python_version()
+    elif data["implementation"]["name"] == "IronPython":
+        # Complete Guess
+        data["implementation"]["version"] = platform.python_version()
+
+    if sys.platform.startswith("linux"):
+        from pip._vendor import distro
+
+        linux_distribution = distro.name(), distro.version(), distro.codename()
+        distro_infos: Dict[str, Any] = dict(
+            filter(
+                lambda x: x[1],
+                zip(["name", "version", "id"], linux_distribution),
+            )
+        )
+        libc = dict(
+            filter(
+                lambda x: x[1],
+                zip(["lib", "version"], libc_ver()),
+            )
+        )
+        if libc:
+            distro_infos["libc"] = libc
+        if distro_infos:
+            data["distro"] = distro_infos
+
+    if sys.platform.startswith("darwin") and platform.mac_ver()[0]:
+        data["distro"] = {"name": "macOS", "version": platform.mac_ver()[0]}
+
+    if platform.system():
+        data.setdefault("system", {})["name"] = platform.system()
+
+    if platform.release():
+        data.setdefault("system", {})["release"] = platform.release()
+
+    if platform.machine():
+        data["cpu"] = platform.machine()
+
+    if has_tls():
+        import _ssl as ssl
+
+        data["openssl_version"] = ssl.OPENSSL_VERSION
+
+    setuptools_dist = get_default_environment().get_distribution("setuptools")
+    if setuptools_dist is not None:
+        data["setuptools_version"] = str(setuptools_dist.version)
+
+    if shutil.which("rustc") is not None:
+        # If for any reason `rustc --version` fails, silently ignore it
+        try:
+            rustc_output = subprocess.check_output(
+                ["rustc", "--version"], stderr=subprocess.STDOUT, timeout=0.5
+            )
+        except Exception:
+            pass
+        else:
+            if rustc_output.startswith(b"rustc "):
+                # The format of `rustc --version` is:
+                # `b'rustc 1.52.1 (9bc8c42bb 2021-05-09)\n'`
+                # We extract just the middle (1.52.1) part
+                data["rustc_version"] = rustc_output.split(b" ")[1].decode()
+
+    # Use None rather than False so as not to give the impression that
+    # pip knows it is not being run under CI.  Rather, it is a null or
+    # inconclusive result.  Also, we include some value rather than no
+    # value to make it easier to know that the check has been run.
+    data["ci"] = True if looks_like_ci() else None
+
+    user_data = os.environ.get("PIP_USER_AGENT_USER_DATA")
+    if user_data is not None:
+        data["user_data"] = user_data
+
+    return "{data[installer][name]}/{data[installer][version]} {json}".format(
+        data=data,
+        json=json.dumps(data, separators=(",", ":"), sort_keys=True),
+    )
+
+
+class LocalFSAdapter(BaseAdapter):
+    def send(
+        self,
+        request: PreparedRequest,
+        stream: bool = False,
+        timeout: Optional[Union[float, Tuple[float, float]]] = None,
+        verify: Union[bool, str] = True,
+        cert: Optional[Union[str, Tuple[str, str]]] = None,
+        proxies: Optional[Mapping[str, str]] = None,
+    ) -> Response:
+        pathname = url_to_path(request.url)
+
+        resp = Response()
+        resp.status_code = 200
+        resp.url = request.url
+
+        try:
+            stats = os.stat(pathname)
+        except OSError as exc:
+            # format the exception raised as a io.BytesIO object,
+            # to return a better error message:
+            resp.status_code = 404
+            resp.reason = type(exc).__name__
+            resp.raw = io.BytesIO(f"{resp.reason}: {exc}".encode())
+        else:
+            modified = email.utils.formatdate(stats.st_mtime, usegmt=True)
+            content_type = mimetypes.guess_type(pathname)[0] or "text/plain"
+            resp.headers = CaseInsensitiveDict(
+                {
+                    "Content-Type": content_type,
+                    "Content-Length": stats.st_size,
+                    "Last-Modified": modified,
+                }
+            )
+
+            resp.raw = open(pathname, "rb")
+            resp.close = resp.raw.close
+
+        return resp
+
+    def close(self) -> None:
+        pass
+
+
+class _SSLContextAdapterMixin:
+    """Mixin to add the ``ssl_context`` constructor argument to HTTP adapters.
+
+    The additional argument is forwarded directly to the pool manager. This allows us
+    to dynamically decide what SSL store to use at runtime, which is used to implement
+    the optional ``truststore`` backend.
+    """
+
+    def __init__(
+        self,
+        *,
+        ssl_context: Optional["SSLContext"] = None,
+        **kwargs: Any,
+    ) -> None:
+        self._ssl_context = ssl_context
+        super().__init__(**kwargs)
+
+    def init_poolmanager(
+        self,
+        connections: int,
+        maxsize: int,
+        block: bool = DEFAULT_POOLBLOCK,
+        **pool_kwargs: Any,
+    ) -> "PoolManager":
+        if self._ssl_context is not None:
+            pool_kwargs.setdefault("ssl_context", self._ssl_context)
+        return super().init_poolmanager(  # type: ignore[misc]
+            connections=connections,
+            maxsize=maxsize,
+            block=block,
+            **pool_kwargs,
+        )
+
+
+class HTTPAdapter(_SSLContextAdapterMixin, _BaseHTTPAdapter):
+    pass
+
+
+class CacheControlAdapter(_SSLContextAdapterMixin, _BaseCacheControlAdapter):
+    pass
+
+
+class InsecureHTTPAdapter(HTTPAdapter):
+    def cert_verify(
+        self,
+        conn: ConnectionPool,
+        url: str,
+        verify: Union[bool, str],
+        cert: Optional[Union[str, Tuple[str, str]]],
+    ) -> None:
+        super().cert_verify(conn=conn, url=url, verify=False, cert=cert)
+
+
+class InsecureCacheControlAdapter(CacheControlAdapter):
+    def cert_verify(
+        self,
+        conn: ConnectionPool,
+        url: str,
+        verify: Union[bool, str],
+        cert: Optional[Union[str, Tuple[str, str]]],
+    ) -> None:
+        super().cert_verify(conn=conn, url=url, verify=False, cert=cert)
+
+
+class PipSession(requests.Session):
+    timeout: Optional[int] = None
+
+    def __init__(
+        self,
+        *args: Any,
+        retries: int = 0,
+        cache: Optional[str] = None,
+        trusted_hosts: Sequence[str] = (),
+        index_urls: Optional[List[str]] = None,
+        ssl_context: Optional["SSLContext"] = None,
+        **kwargs: Any,
+    ) -> None:
+        """
+        :param trusted_hosts: Domains not to emit warnings for when not using
+            HTTPS.
+        """
+        super().__init__(*args, **kwargs)
+
+        # Namespace the attribute with "pip_" just in case to prevent
+        # possible conflicts with the base class.
+        self.pip_trusted_origins: List[Tuple[str, Optional[int]]] = []
+        self.pip_proxy = None
+
+        # Attach our User Agent to the request
+        self.headers["User-Agent"] = user_agent()
+
+        # Attach our Authentication handler to the session
+        self.auth = MultiDomainBasicAuth(index_urls=index_urls)
+
+        # Create our urllib3.Retry instance which will allow us to customize
+        # how we handle retries.
+        retries = urllib3.Retry(
+            # Set the total number of retries that a particular request can
+            # have.
+            total=retries,
+            # A 503 error from PyPI typically means that the Fastly -> Origin
+            # connection got interrupted in some way. A 503 error in general
+            # is typically considered a transient error so we'll go ahead and
+            # retry it.
+            # A 500 may indicate transient error in Amazon S3
+            # A 502 may be a transient error from a CDN like CloudFlare or CloudFront
+            # A 520 or 527 - may indicate transient error in CloudFlare
+            status_forcelist=[500, 502, 503, 520, 527],
+            # Add a small amount of back off between failed requests in
+            # order to prevent hammering the service.
+            backoff_factor=0.25,
+        )  # type: ignore
+
+        # Our Insecure HTTPAdapter disables HTTPS validation. It does not
+        # support caching so we'll use it for all http:// URLs.
+        # If caching is disabled, we will also use it for
+        # https:// hosts that we've marked as ignoring
+        # TLS errors for (trusted-hosts).
+        insecure_adapter = InsecureHTTPAdapter(max_retries=retries)
+
+        # We want to _only_ cache responses on securely fetched origins or when
+        # the host is specified as trusted. We do this because
+        # we can't validate the response of an insecurely/untrusted fetched
+        # origin, and we don't want someone to be able to poison the cache and
+        # require manual eviction from the cache to fix it.
+        if cache:
+            secure_adapter = CacheControlAdapter(
+                cache=SafeFileCache(cache),
+                max_retries=retries,
+                ssl_context=ssl_context,
+            )
+            self._trusted_host_adapter = InsecureCacheControlAdapter(
+                cache=SafeFileCache(cache),
+                max_retries=retries,
+            )
+        else:
+            secure_adapter = HTTPAdapter(max_retries=retries, ssl_context=ssl_context)
+            self._trusted_host_adapter = insecure_adapter
+
+        self.mount("https://", secure_adapter)
+        self.mount("http://", insecure_adapter)
+
+        # Enable file:// urls
+        self.mount("file://", LocalFSAdapter())
+
+        for host in trusted_hosts:
+            self.add_trusted_host(host, suppress_logging=True)
+
+    def update_index_urls(self, new_index_urls: List[str]) -> None:
+        """
+        :param new_index_urls: New index urls to update the authentication
+            handler with.
+        """
+        self.auth.index_urls = new_index_urls
+
+    def add_trusted_host(
+        self, host: str, source: Optional[str] = None, suppress_logging: bool = False
+    ) -> None:
+        """
+        :param host: It is okay to provide a host that has previously been
+            added.
+        :param source: An optional source string, for logging where the host
+            string came from.
+        """
+        if not suppress_logging:
+            msg = f"adding trusted host: {host!r}"
+            if source is not None:
+                msg += f" (from {source})"
+            logger.info(msg)
+
+        parsed_host, parsed_port = parse_netloc(host)
+        if parsed_host is None:
+            raise ValueError(f"Trusted host URL must include a host part: {host!r}")
+        if (parsed_host, parsed_port) not in self.pip_trusted_origins:
+            self.pip_trusted_origins.append((parsed_host, parsed_port))
+
+        self.mount(
+            build_url_from_netloc(host, scheme="http") + "/", self._trusted_host_adapter
+        )
+        self.mount(build_url_from_netloc(host) + "/", self._trusted_host_adapter)
+        if not parsed_port:
+            self.mount(
+                build_url_from_netloc(host, scheme="http") + ":",
+                self._trusted_host_adapter,
+            )
+            # Mount wildcard ports for the same host.
+            self.mount(build_url_from_netloc(host) + ":", self._trusted_host_adapter)
+
+    def iter_secure_origins(self) -> Generator[SecureOrigin, None, None]:
+        yield from SECURE_ORIGINS
+        for host, port in self.pip_trusted_origins:
+            yield ("*", host, "*" if port is None else port)
+
+    def is_secure_origin(self, location: Link) -> bool:
+        # Determine if this url used a secure transport mechanism
+        parsed = urllib.parse.urlparse(str(location))
+        origin_protocol, origin_host, origin_port = (
+            parsed.scheme,
+            parsed.hostname,
+            parsed.port,
+        )
+
+        # The protocol to use to see if the protocol matches.
+        # Don't count the repository type as part of the protocol: in
+        # cases such as "git+ssh", only use "ssh". (I.e., Only verify against
+        # the last scheme.)
+        origin_protocol = origin_protocol.rsplit("+", 1)[-1]
+
+        # Determine if our origin is a secure origin by looking through our
+        # hardcoded list of secure origins, as well as any additional ones
+        # configured on this PackageFinder instance.
+        for secure_origin in self.iter_secure_origins():
+            secure_protocol, secure_host, secure_port = secure_origin
+            if origin_protocol != secure_protocol and secure_protocol != "*":
+                continue
+
+            try:
+                addr = ipaddress.ip_address(origin_host or "")
+                network = ipaddress.ip_network(secure_host)
+            except ValueError:
+                # We don't have both a valid address or a valid network, so
+                # we'll check this origin against hostnames.
+                if (
+                    origin_host
+                    and origin_host.lower() != secure_host.lower()
+                    and secure_host != "*"
+                ):
+                    continue
+            else:
+                # We have a valid address and network, so see if the address
+                # is contained within the network.
+                if addr not in network:
+                    continue
+
+            # Check to see if the port matches.
+            if (
+                origin_port != secure_port
+                and secure_port != "*"
+                and secure_port is not None
+            ):
+                continue
+
+            # If we've gotten here, then this origin matches the current
+            # secure origin and we should return True
+            return True
+
+        # If we've gotten to this point, then the origin isn't secure and we
+        # will not accept it as a valid location to search. We will however
+        # log a warning that we are ignoring it.
+        logger.warning(
+            "The repository located at %s is not a trusted or secure host and "
+            "is being ignored. If this repository is available via HTTPS we "
+            "recommend you use HTTPS instead, otherwise you may silence "
+            "this warning and allow it anyway with '--trusted-host %s'.",
+            origin_host,
+            origin_host,
+        )
+
+        return False
+
+    def request(self, method: str, url: str, *args: Any, **kwargs: Any) -> Response:
+        # Allow setting a default timeout on a session
+        kwargs.setdefault("timeout", self.timeout)
+        # Allow setting a default proxies on a session
+        kwargs.setdefault("proxies", self.proxies)
+
+        # Dispatch the actual request
+        return super().request(method, url, *args, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/utils.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/utils.py
new file mode 100644
index 00000000..bba4c265
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/utils.py
@@ -0,0 +1,98 @@
+from typing import Dict, Generator
+
+from pip._vendor.requests.models import Response
+
+from pip._internal.exceptions import NetworkConnectionError
+
+# The following comments and HTTP headers were originally added by
+# Donald Stufft in git commit 22c562429a61bb77172039e480873fb239dd8c03.
+#
+# We use Accept-Encoding: identity here because requests defaults to
+# accepting compressed responses. This breaks in a variety of ways
+# depending on how the server is configured.
+# - Some servers will notice that the file isn't a compressible file
+#   and will leave the file alone and with an empty Content-Encoding
+# - Some servers will notice that the file is already compressed and
+#   will leave the file alone, adding a Content-Encoding: gzip header
+# - Some servers won't notice anything at all and will take a file
+#   that's already been compressed and compress it again, and set
+#   the Content-Encoding: gzip header
+# By setting this to request only the identity encoding we're hoping
+# to eliminate the third case.  Hopefully there does not exist a server
+# which when given a file will notice it is already compressed and that
+# you're not asking for a compressed file and will then decompress it
+# before sending because if that's the case I don't think it'll ever be
+# possible to make this work.
+HEADERS: Dict[str, str] = {"Accept-Encoding": "identity"}
+
+DOWNLOAD_CHUNK_SIZE = 256 * 1024
+
+
+def raise_for_status(resp: Response) -> None:
+    http_error_msg = ""
+    if isinstance(resp.reason, bytes):
+        # We attempt to decode utf-8 first because some servers
+        # choose to localize their reason strings. If the string
+        # isn't utf-8, we fall back to iso-8859-1 for all other
+        # encodings.
+        try:
+            reason = resp.reason.decode("utf-8")
+        except UnicodeDecodeError:
+            reason = resp.reason.decode("iso-8859-1")
+    else:
+        reason = resp.reason
+
+    if 400 <= resp.status_code < 500:
+        http_error_msg = (
+            f"{resp.status_code} Client Error: {reason} for url: {resp.url}"
+        )
+
+    elif 500 <= resp.status_code < 600:
+        http_error_msg = (
+            f"{resp.status_code} Server Error: {reason} for url: {resp.url}"
+        )
+
+    if http_error_msg:
+        raise NetworkConnectionError(http_error_msg, response=resp)
+
+
+def response_chunks(
+    response: Response, chunk_size: int = DOWNLOAD_CHUNK_SIZE
+) -> Generator[bytes, None, None]:
+    """Given a requests Response, provide the data chunks."""
+    try:
+        # Special case for urllib3.
+        for chunk in response.raw.stream(
+            chunk_size,
+            # We use decode_content=False here because we don't
+            # want urllib3 to mess with the raw bytes we get
+            # from the server. If we decompress inside of
+            # urllib3 then we cannot verify the checksum
+            # because the checksum will be of the compressed
+            # file. This breakage will only occur if the
+            # server adds a Content-Encoding header, which
+            # depends on how the server was configured:
+            # - Some servers will notice that the file isn't a
+            #   compressible file and will leave the file alone
+            #   and with an empty Content-Encoding
+            # - Some servers will notice that the file is
+            #   already compressed and will leave the file
+            #   alone and will add a Content-Encoding: gzip
+            #   header
+            # - Some servers won't notice anything at all and
+            #   will take a file that's already been compressed
+            #   and compress it again and set the
+            #   Content-Encoding: gzip header
+            #
+            # By setting this not to decode automatically we
+            # hope to eliminate problems with the second case.
+            decode_content=False,
+        ):
+            yield chunk
+    except AttributeError:
+        # Standard file-like object.
+        while True:
+            chunk = response.raw.read(chunk_size)
+            if not chunk:
+                break
+            yield chunk
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/network/xmlrpc.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/xmlrpc.py
new file mode 100644
index 00000000..22ec8d2f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/network/xmlrpc.py
@@ -0,0 +1,62 @@
+"""xmlrpclib.Transport implementation
+"""
+
+import logging
+import urllib.parse
+import xmlrpc.client
+from typing import TYPE_CHECKING, Tuple
+
+from pip._internal.exceptions import NetworkConnectionError
+from pip._internal.network.session import PipSession
+from pip._internal.network.utils import raise_for_status
+
+if TYPE_CHECKING:
+    from xmlrpc.client import _HostType, _Marshallable
+
+    from _typeshed import SizedBuffer
+
+logger = logging.getLogger(__name__)
+
+
+class PipXmlrpcTransport(xmlrpc.client.Transport):
+    """Provide a `xmlrpclib.Transport` implementation via a `PipSession`
+    object.
+    """
+
+    def __init__(
+        self, index_url: str, session: PipSession, use_datetime: bool = False
+    ) -> None:
+        super().__init__(use_datetime)
+        index_parts = urllib.parse.urlparse(index_url)
+        self._scheme = index_parts.scheme
+        self._session = session
+
+    def request(
+        self,
+        host: "_HostType",
+        handler: str,
+        request_body: "SizedBuffer",
+        verbose: bool = False,
+    ) -> Tuple["_Marshallable", ...]:
+        assert isinstance(host, str)
+        parts = (self._scheme, host, handler, None, None, None)
+        url = urllib.parse.urlunparse(parts)
+        try:
+            headers = {"Content-Type": "text/xml"}
+            response = self._session.post(
+                url,
+                data=request_body,
+                headers=headers,
+                stream=True,
+            )
+            raise_for_status(response)
+            self.verbose = verbose
+            return self.parse_response(response.raw)
+        except NetworkConnectionError as exc:
+            assert exc.response
+            logger.critical(
+                "HTTP error %s while getting %s",
+                exc.response.status_code,
+                url,
+            )
+            raise
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..753e34fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/check.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/check.cpython-312.pyc
new file mode 100644
index 00000000..10500a6d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/check.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/freeze.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/freeze.cpython-312.pyc
new file mode 100644
index 00000000..63d3f091
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/freeze.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/prepare.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/prepare.cpython-312.pyc
new file mode 100644
index 00000000..03320e83
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/__pycache__/prepare.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/check.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/check.py
new file mode 100644
index 00000000..4b6fbc4c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/check.py
@@ -0,0 +1,181 @@
+"""Validation of dependencies of packages
+"""
+
+import logging
+from contextlib import suppress
+from email.parser import Parser
+from functools import reduce
+from typing import (
+    Callable,
+    Dict,
+    FrozenSet,
+    Generator,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Set,
+    Tuple,
+)
+
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.packaging.tags import Tag, parse_tag
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import Version
+
+from pip._internal.distributions import make_distribution_for_install_requirement
+from pip._internal.metadata import get_default_environment
+from pip._internal.metadata.base import BaseDistribution
+from pip._internal.req.req_install import InstallRequirement
+
+logger = logging.getLogger(__name__)
+
+
+class PackageDetails(NamedTuple):
+    version: Version
+    dependencies: List[Requirement]
+
+
+# Shorthands
+PackageSet = Dict[NormalizedName, PackageDetails]
+Missing = Tuple[NormalizedName, Requirement]
+Conflicting = Tuple[NormalizedName, Version, Requirement]
+
+MissingDict = Dict[NormalizedName, List[Missing]]
+ConflictingDict = Dict[NormalizedName, List[Conflicting]]
+CheckResult = Tuple[MissingDict, ConflictingDict]
+ConflictDetails = Tuple[PackageSet, CheckResult]
+
+
+def create_package_set_from_installed() -> Tuple[PackageSet, bool]:
+    """Converts a list of distributions into a PackageSet."""
+    package_set = {}
+    problems = False
+    env = get_default_environment()
+    for dist in env.iter_installed_distributions(local_only=False, skip=()):
+        name = dist.canonical_name
+        try:
+            dependencies = list(dist.iter_dependencies())
+            package_set[name] = PackageDetails(dist.version, dependencies)
+        except (OSError, ValueError) as e:
+            # Don't crash on unreadable or broken metadata.
+            logger.warning("Error parsing dependencies of %s: %s", name, e)
+            problems = True
+    return package_set, problems
+
+
+def check_package_set(
+    package_set: PackageSet, should_ignore: Optional[Callable[[str], bool]] = None
+) -> CheckResult:
+    """Check if a package set is consistent
+
+    If should_ignore is passed, it should be a callable that takes a
+    package name and returns a boolean.
+    """
+
+    missing = {}
+    conflicting = {}
+
+    for package_name, package_detail in package_set.items():
+        # Info about dependencies of package_name
+        missing_deps: Set[Missing] = set()
+        conflicting_deps: Set[Conflicting] = set()
+
+        if should_ignore and should_ignore(package_name):
+            continue
+
+        for req in package_detail.dependencies:
+            name = canonicalize_name(req.name)
+
+            # Check if it's missing
+            if name not in package_set:
+                missed = True
+                if req.marker is not None:
+                    missed = req.marker.evaluate({"extra": ""})
+                if missed:
+                    missing_deps.add((name, req))
+                continue
+
+            # Check if there's a conflict
+            version = package_set[name].version
+            if not req.specifier.contains(version, prereleases=True):
+                conflicting_deps.add((name, version, req))
+
+        if missing_deps:
+            missing[package_name] = sorted(missing_deps, key=str)
+        if conflicting_deps:
+            conflicting[package_name] = sorted(conflicting_deps, key=str)
+
+    return missing, conflicting
+
+
+def check_install_conflicts(to_install: List[InstallRequirement]) -> ConflictDetails:
+    """For checking if the dependency graph would be consistent after \
+    installing given requirements
+    """
+    # Start from the current state
+    package_set, _ = create_package_set_from_installed()
+    # Install packages
+    would_be_installed = _simulate_installation_of(to_install, package_set)
+
+    # Only warn about directly-dependent packages; create a whitelist of them
+    whitelist = _create_whitelist(would_be_installed, package_set)
+
+    return (
+        package_set,
+        check_package_set(
+            package_set, should_ignore=lambda name: name not in whitelist
+        ),
+    )
+
+
+def check_unsupported(
+    packages: Iterable[BaseDistribution],
+    supported_tags: Iterable[Tag],
+) -> Generator[BaseDistribution, None, None]:
+    for p in packages:
+        with suppress(FileNotFoundError):
+            wheel_file = p.read_text("WHEEL")
+            wheel_tags: FrozenSet[Tag] = reduce(
+                frozenset.union,
+                map(parse_tag, Parser().parsestr(wheel_file).get_all("Tag", [])),
+                frozenset(),
+            )
+            if wheel_tags.isdisjoint(supported_tags):
+                yield p
+
+
+def _simulate_installation_of(
+    to_install: List[InstallRequirement], package_set: PackageSet
+) -> Set[NormalizedName]:
+    """Computes the version of packages after installing to_install."""
+    # Keep track of packages that were installed
+    installed = set()
+
+    # Modify it as installing requirement_set would (assuming no errors)
+    for inst_req in to_install:
+        abstract_dist = make_distribution_for_install_requirement(inst_req)
+        dist = abstract_dist.get_metadata_distribution()
+        name = dist.canonical_name
+        package_set[name] = PackageDetails(dist.version, list(dist.iter_dependencies()))
+
+        installed.add(name)
+
+    return installed
+
+
+def _create_whitelist(
+    would_be_installed: Set[NormalizedName], package_set: PackageSet
+) -> Set[NormalizedName]:
+    packages_affected = set(would_be_installed)
+
+    for package_name in package_set:
+        if package_name in packages_affected:
+            continue
+
+        for req in package_set[package_name].dependencies:
+            if canonicalize_name(req.name) in packages_affected:
+                packages_affected.add(package_name)
+                break
+
+    return packages_affected
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/freeze.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/freeze.py
new file mode 100644
index 00000000..ae5dd37f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/freeze.py
@@ -0,0 +1,256 @@
+import collections
+import logging
+import os
+from dataclasses import dataclass, field
+from typing import Container, Dict, Generator, Iterable, List, NamedTuple, Optional, Set
+
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import InvalidVersion
+
+from pip._internal.exceptions import BadCommand, InstallationError
+from pip._internal.metadata import BaseDistribution, get_environment
+from pip._internal.req.constructors import (
+    install_req_from_editable,
+    install_req_from_line,
+)
+from pip._internal.req.req_file import COMMENT_RE
+from pip._internal.utils.direct_url_helpers import direct_url_as_pep440_direct_reference
+
+logger = logging.getLogger(__name__)
+
+
+class _EditableInfo(NamedTuple):
+    requirement: str
+    comments: List[str]
+
+
+def freeze(
+    requirement: Optional[List[str]] = None,
+    local_only: bool = False,
+    user_only: bool = False,
+    paths: Optional[List[str]] = None,
+    isolated: bool = False,
+    exclude_editable: bool = False,
+    skip: Container[str] = (),
+) -> Generator[str, None, None]:
+    installations: Dict[str, FrozenRequirement] = {}
+
+    dists = get_environment(paths).iter_installed_distributions(
+        local_only=local_only,
+        skip=(),
+        user_only=user_only,
+    )
+    for dist in dists:
+        req = FrozenRequirement.from_dist(dist)
+        if exclude_editable and req.editable:
+            continue
+        installations[req.canonical_name] = req
+
+    if requirement:
+        # the options that don't get turned into an InstallRequirement
+        # should only be emitted once, even if the same option is in multiple
+        # requirements files, so we need to keep track of what has been emitted
+        # so that we don't emit it again if it's seen again
+        emitted_options: Set[str] = set()
+        # keep track of which files a requirement is in so that we can
+        # give an accurate warning if a requirement appears multiple times.
+        req_files: Dict[str, List[str]] = collections.defaultdict(list)
+        for req_file_path in requirement:
+            with open(req_file_path) as req_file:
+                for line in req_file:
+                    if (
+                        not line.strip()
+                        or line.strip().startswith("#")
+                        or line.startswith(
+                            (
+                                "-r",
+                                "--requirement",
+                                "-f",
+                                "--find-links",
+                                "-i",
+                                "--index-url",
+                                "--pre",
+                                "--trusted-host",
+                                "--process-dependency-links",
+                                "--extra-index-url",
+                                "--use-feature",
+                            )
+                        )
+                    ):
+                        line = line.rstrip()
+                        if line not in emitted_options:
+                            emitted_options.add(line)
+                            yield line
+                        continue
+
+                    if line.startswith("-e") or line.startswith("--editable"):
+                        if line.startswith("-e"):
+                            line = line[2:].strip()
+                        else:
+                            line = line[len("--editable") :].strip().lstrip("=")
+                        line_req = install_req_from_editable(
+                            line,
+                            isolated=isolated,
+                        )
+                    else:
+                        line_req = install_req_from_line(
+                            COMMENT_RE.sub("", line).strip(),
+                            isolated=isolated,
+                        )
+
+                    if not line_req.name:
+                        logger.info(
+                            "Skipping line in requirement file [%s] because "
+                            "it's not clear what it would install: %s",
+                            req_file_path,
+                            line.strip(),
+                        )
+                        logger.info(
+                            "  (add #egg=PackageName to the URL to avoid"
+                            " this warning)"
+                        )
+                    else:
+                        line_req_canonical_name = canonicalize_name(line_req.name)
+                        if line_req_canonical_name not in installations:
+                            # either it's not installed, or it is installed
+                            # but has been processed already
+                            if not req_files[line_req.name]:
+                                logger.warning(
+                                    "Requirement file [%s] contains %s, but "
+                                    "package %r is not installed",
+                                    req_file_path,
+                                    COMMENT_RE.sub("", line).strip(),
+                                    line_req.name,
+                                )
+                            else:
+                                req_files[line_req.name].append(req_file_path)
+                        else:
+                            yield str(installations[line_req_canonical_name]).rstrip()
+                            del installations[line_req_canonical_name]
+                            req_files[line_req.name].append(req_file_path)
+
+        # Warn about requirements that were included multiple times (in a
+        # single requirements file or in different requirements files).
+        for name, files in req_files.items():
+            if len(files) > 1:
+                logger.warning(
+                    "Requirement %s included multiple times [%s]",
+                    name,
+                    ", ".join(sorted(set(files))),
+                )
+
+        yield ("## The following requirements were added by pip freeze:")
+    for installation in sorted(installations.values(), key=lambda x: x.name.lower()):
+        if installation.canonical_name not in skip:
+            yield str(installation).rstrip()
+
+
+def _format_as_name_version(dist: BaseDistribution) -> str:
+    try:
+        dist_version = dist.version
+    except InvalidVersion:
+        # legacy version
+        return f"{dist.raw_name}==={dist.raw_version}"
+    else:
+        return f"{dist.raw_name}=={dist_version}"
+
+
+def _get_editable_info(dist: BaseDistribution) -> _EditableInfo:
+    """
+    Compute and return values (req, comments) for use in
+    FrozenRequirement.from_dist().
+    """
+    editable_project_location = dist.editable_project_location
+    assert editable_project_location
+    location = os.path.normcase(os.path.abspath(editable_project_location))
+
+    from pip._internal.vcs import RemoteNotFoundError, RemoteNotValidError, vcs
+
+    vcs_backend = vcs.get_backend_for_dir(location)
+
+    if vcs_backend is None:
+        display = _format_as_name_version(dist)
+        logger.debug(
+            'No VCS found for editable requirement "%s" in: %r',
+            display,
+            location,
+        )
+        return _EditableInfo(
+            requirement=location,
+            comments=[f"# Editable install with no version control ({display})"],
+        )
+
+    vcs_name = type(vcs_backend).__name__
+
+    try:
+        req = vcs_backend.get_src_requirement(location, dist.raw_name)
+    except RemoteNotFoundError:
+        display = _format_as_name_version(dist)
+        return _EditableInfo(
+            requirement=location,
+            comments=[f"# Editable {vcs_name} install with no remote ({display})"],
+        )
+    except RemoteNotValidError as ex:
+        display = _format_as_name_version(dist)
+        return _EditableInfo(
+            requirement=location,
+            comments=[
+                f"# Editable {vcs_name} install ({display}) with either a deleted "
+                f"local remote or invalid URI:",
+                f"# '{ex.url}'",
+            ],
+        )
+    except BadCommand:
+        logger.warning(
+            "cannot determine version of editable source in %s "
+            "(%s command not found in path)",
+            location,
+            vcs_backend.name,
+        )
+        return _EditableInfo(requirement=location, comments=[])
+    except InstallationError as exc:
+        logger.warning("Error when trying to get requirement for VCS system %s", exc)
+    else:
+        return _EditableInfo(requirement=req, comments=[])
+
+    logger.warning("Could not determine repository location of %s", location)
+
+    return _EditableInfo(
+        requirement=location,
+        comments=["## !! Could not determine repository location"],
+    )
+
+
+@dataclass(frozen=True)
+class FrozenRequirement:
+    name: str
+    req: str
+    editable: bool
+    comments: Iterable[str] = field(default_factory=tuple)
+
+    @property
+    def canonical_name(self) -> NormalizedName:
+        return canonicalize_name(self.name)
+
+    @classmethod
+    def from_dist(cls, dist: BaseDistribution) -> "FrozenRequirement":
+        editable = dist.editable
+        if editable:
+            req, comments = _get_editable_info(dist)
+        else:
+            comments = []
+            direct_url = dist.direct_url
+            if direct_url:
+                # if PEP 610 metadata is present, use it
+                req = direct_url_as_pep440_direct_reference(direct_url, dist.raw_name)
+            else:
+                # name==version requirement
+                req = _format_as_name_version(dist)
+
+        return cls(dist.raw_name, req, editable, comments=comments)
+
+    def __str__(self) -> str:
+        req = self.req
+        if self.editable:
+            req = f"-e {req}"
+        return "\n".join(list(self.comments) + [str(req)]) + "\n"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__init__.py
new file mode 100644
index 00000000..24d6a5dd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__init__.py
@@ -0,0 +1,2 @@
+"""For modules related to installing packages.
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..12e63dda
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/editable_legacy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/editable_legacy.cpython-312.pyc
new file mode 100644
index 00000000..f480f700
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/editable_legacy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..99ba81b9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/editable_legacy.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/editable_legacy.py
new file mode 100644
index 00000000..9aaa699a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/editable_legacy.py
@@ -0,0 +1,47 @@
+"""Legacy editable installation process, i.e. `setup.py develop`.
+"""
+
+import logging
+from typing import Optional, Sequence
+
+from pip._internal.build_env import BuildEnvironment
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.setuptools_build import make_setuptools_develop_args
+from pip._internal.utils.subprocess import call_subprocess
+
+logger = logging.getLogger(__name__)
+
+
+def install_editable(
+    *,
+    global_options: Sequence[str],
+    prefix: Optional[str],
+    home: Optional[str],
+    use_user_site: bool,
+    name: str,
+    setup_py_path: str,
+    isolated: bool,
+    build_env: BuildEnvironment,
+    unpacked_source_directory: str,
+) -> None:
+    """Install a package in editable mode. Most arguments are pass-through
+    to setuptools.
+    """
+    logger.info("Running setup.py develop for %s", name)
+
+    args = make_setuptools_develop_args(
+        setup_py_path,
+        global_options=global_options,
+        no_user_config=isolated,
+        prefix=prefix,
+        home=home,
+        use_user_site=use_user_site,
+    )
+
+    with indent_log():
+        with build_env:
+            call_subprocess(
+                args,
+                command_desc="python setup.py develop",
+                cwd=unpacked_source_directory,
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/wheel.py
new file mode 100644
index 00000000..aef42aa9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/install/wheel.py
@@ -0,0 +1,741 @@
+"""Support for installing and building the "wheel" binary package format.
+"""
+
+import collections
+import compileall
+import contextlib
+import csv
+import importlib
+import logging
+import os.path
+import re
+import shutil
+import sys
+import warnings
+from base64 import urlsafe_b64encode
+from email.message import Message
+from itertools import chain, filterfalse, starmap
+from typing import (
+    IO,
+    TYPE_CHECKING,
+    Any,
+    BinaryIO,
+    Callable,
+    Dict,
+    Generator,
+    Iterable,
+    Iterator,
+    List,
+    NewType,
+    Optional,
+    Protocol,
+    Sequence,
+    Set,
+    Tuple,
+    Union,
+    cast,
+)
+from zipfile import ZipFile, ZipInfo
+
+from pip._vendor.distlib.scripts import ScriptMaker
+from pip._vendor.distlib.util import get_export_entry
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.exceptions import InstallationError
+from pip._internal.locations import get_major_minor_version
+from pip._internal.metadata import (
+    BaseDistribution,
+    FilesystemWheel,
+    get_wheel_distribution,
+)
+from pip._internal.models.direct_url import DIRECT_URL_METADATA_NAME, DirectUrl
+from pip._internal.models.scheme import SCHEME_KEYS, Scheme
+from pip._internal.utils.filesystem import adjacent_tmp_file, replace
+from pip._internal.utils.misc import StreamWrapper, ensure_dir, hash_file, partition
+from pip._internal.utils.unpacking import (
+    current_umask,
+    is_within_directory,
+    set_extracted_file_to_default_mode_plus_executable,
+    zip_item_is_executable,
+)
+from pip._internal.utils.wheel import parse_wheel
+
+if TYPE_CHECKING:
+
+    class File(Protocol):
+        src_record_path: "RecordPath"
+        dest_path: str
+        changed: bool
+
+        def save(self) -> None:
+            pass
+
+
+logger = logging.getLogger(__name__)
+
+RecordPath = NewType("RecordPath", str)
+InstalledCSVRow = Tuple[RecordPath, str, Union[int, str]]
+
+
+def rehash(path: str, blocksize: int = 1 << 20) -> Tuple[str, str]:
+    """Return (encoded_digest, length) for path using hashlib.sha256()"""
+    h, length = hash_file(path, blocksize)
+    digest = "sha256=" + urlsafe_b64encode(h.digest()).decode("latin1").rstrip("=")
+    return (digest, str(length))
+
+
+def csv_io_kwargs(mode: str) -> Dict[str, Any]:
+    """Return keyword arguments to properly open a CSV file
+    in the given mode.
+    """
+    return {"mode": mode, "newline": "", "encoding": "utf-8"}
+
+
+def fix_script(path: str) -> bool:
+    """Replace #!python with #!/path/to/python
+    Return True if file was changed.
+    """
+    # XXX RECORD hashes will need to be updated
+    assert os.path.isfile(path)
+
+    with open(path, "rb") as script:
+        firstline = script.readline()
+        if not firstline.startswith(b"#!python"):
+            return False
+        exename = sys.executable.encode(sys.getfilesystemencoding())
+        firstline = b"#!" + exename + os.linesep.encode("ascii")
+        rest = script.read()
+    with open(path, "wb") as script:
+        script.write(firstline)
+        script.write(rest)
+    return True
+
+
+def wheel_root_is_purelib(metadata: Message) -> bool:
+    return metadata.get("Root-Is-Purelib", "").lower() == "true"
+
+
+def get_entrypoints(dist: BaseDistribution) -> Tuple[Dict[str, str], Dict[str, str]]:
+    console_scripts = {}
+    gui_scripts = {}
+    for entry_point in dist.iter_entry_points():
+        if entry_point.group == "console_scripts":
+            console_scripts[entry_point.name] = entry_point.value
+        elif entry_point.group == "gui_scripts":
+            gui_scripts[entry_point.name] = entry_point.value
+    return console_scripts, gui_scripts
+
+
+def message_about_scripts_not_on_PATH(scripts: Sequence[str]) -> Optional[str]:
+    """Determine if any scripts are not on PATH and format a warning.
+    Returns a warning message if one or more scripts are not on PATH,
+    otherwise None.
+    """
+    if not scripts:
+        return None
+
+    # Group scripts by the path they were installed in
+    grouped_by_dir: Dict[str, Set[str]] = collections.defaultdict(set)
+    for destfile in scripts:
+        parent_dir = os.path.dirname(destfile)
+        script_name = os.path.basename(destfile)
+        grouped_by_dir[parent_dir].add(script_name)
+
+    # We don't want to warn for directories that are on PATH.
+    not_warn_dirs = [
+        os.path.normcase(os.path.normpath(i)).rstrip(os.sep)
+        for i in os.environ.get("PATH", "").split(os.pathsep)
+    ]
+    # If an executable sits with sys.executable, we don't warn for it.
+    #     This covers the case of venv invocations without activating the venv.
+    not_warn_dirs.append(
+        os.path.normcase(os.path.normpath(os.path.dirname(sys.executable)))
+    )
+    warn_for: Dict[str, Set[str]] = {
+        parent_dir: scripts
+        for parent_dir, scripts in grouped_by_dir.items()
+        if os.path.normcase(os.path.normpath(parent_dir)) not in not_warn_dirs
+    }
+    if not warn_for:
+        return None
+
+    # Format a message
+    msg_lines = []
+    for parent_dir, dir_scripts in warn_for.items():
+        sorted_scripts: List[str] = sorted(dir_scripts)
+        if len(sorted_scripts) == 1:
+            start_text = f"script {sorted_scripts[0]} is"
+        else:
+            start_text = "scripts {} are".format(
+                ", ".join(sorted_scripts[:-1]) + " and " + sorted_scripts[-1]
+            )
+
+        msg_lines.append(
+            f"The {start_text} installed in '{parent_dir}' which is not on PATH."
+        )
+
+    last_line_fmt = (
+        "Consider adding {} to PATH or, if you prefer "
+        "to suppress this warning, use --no-warn-script-location."
+    )
+    if len(msg_lines) == 1:
+        msg_lines.append(last_line_fmt.format("this directory"))
+    else:
+        msg_lines.append(last_line_fmt.format("these directories"))
+
+    # Add a note if any directory starts with ~
+    warn_for_tilde = any(
+        i[0] == "~" for i in os.environ.get("PATH", "").split(os.pathsep) if i
+    )
+    if warn_for_tilde:
+        tilde_warning_msg = (
+            "NOTE: The current PATH contains path(s) starting with `~`, "
+            "which may not be expanded by all applications."
+        )
+        msg_lines.append(tilde_warning_msg)
+
+    # Returns the formatted multiline message
+    return "\n".join(msg_lines)
+
+
+def _normalized_outrows(
+    outrows: Iterable[InstalledCSVRow],
+) -> List[Tuple[str, str, str]]:
+    """Normalize the given rows of a RECORD file.
+
+    Items in each row are converted into str. Rows are then sorted to make
+    the value more predictable for tests.
+
+    Each row is a 3-tuple (path, hash, size) and corresponds to a record of
+    a RECORD file (see PEP 376 and PEP 427 for details).  For the rows
+    passed to this function, the size can be an integer as an int or string,
+    or the empty string.
+    """
+    # Normally, there should only be one row per path, in which case the
+    # second and third elements don't come into play when sorting.
+    # However, in cases in the wild where a path might happen to occur twice,
+    # we don't want the sort operation to trigger an error (but still want
+    # determinism).  Since the third element can be an int or string, we
+    # coerce each element to a string to avoid a TypeError in this case.
+    # For additional background, see--
+    # https://github.com/pypa/pip/issues/5868
+    return sorted(
+        (record_path, hash_, str(size)) for record_path, hash_, size in outrows
+    )
+
+
+def _record_to_fs_path(record_path: RecordPath, lib_dir: str) -> str:
+    return os.path.join(lib_dir, record_path)
+
+
+def _fs_to_record_path(path: str, lib_dir: str) -> RecordPath:
+    # On Windows, do not handle relative paths if they belong to different
+    # logical disks
+    if os.path.splitdrive(path)[0].lower() == os.path.splitdrive(lib_dir)[0].lower():
+        path = os.path.relpath(path, lib_dir)
+
+    path = path.replace(os.path.sep, "/")
+    return cast("RecordPath", path)
+
+
+def get_csv_rows_for_installed(
+    old_csv_rows: List[List[str]],
+    installed: Dict[RecordPath, RecordPath],
+    changed: Set[RecordPath],
+    generated: List[str],
+    lib_dir: str,
+) -> List[InstalledCSVRow]:
+    """
+    :param installed: A map from archive RECORD path to installation RECORD
+        path.
+    """
+    installed_rows: List[InstalledCSVRow] = []
+    for row in old_csv_rows:
+        if len(row) > 3:
+            logger.warning("RECORD line has more than three elements: %s", row)
+        old_record_path = cast("RecordPath", row[0])
+        new_record_path = installed.pop(old_record_path, old_record_path)
+        if new_record_path in changed:
+            digest, length = rehash(_record_to_fs_path(new_record_path, lib_dir))
+        else:
+            digest = row[1] if len(row) > 1 else ""
+            length = row[2] if len(row) > 2 else ""
+        installed_rows.append((new_record_path, digest, length))
+    for f in generated:
+        path = _fs_to_record_path(f, lib_dir)
+        digest, length = rehash(f)
+        installed_rows.append((path, digest, length))
+    return installed_rows + [
+        (installed_record_path, "", "") for installed_record_path in installed.values()
+    ]
+
+
+def get_console_script_specs(console: Dict[str, str]) -> List[str]:
+    """
+    Given the mapping from entrypoint name to callable, return the relevant
+    console script specs.
+    """
+    # Don't mutate caller's version
+    console = console.copy()
+
+    scripts_to_generate = []
+
+    # Special case pip and setuptools to generate versioned wrappers
+    #
+    # The issue is that some projects (specifically, pip and setuptools) use
+    # code in setup.py to create "versioned" entry points - pip2.7 on Python
+    # 2.7, pip3.3 on Python 3.3, etc. But these entry points are baked into
+    # the wheel metadata at build time, and so if the wheel is installed with
+    # a *different* version of Python the entry points will be wrong. The
+    # correct fix for this is to enhance the metadata to be able to describe
+    # such versioned entry points.
+    # Currently, projects using versioned entry points will either have
+    # incorrect versioned entry points, or they will not be able to distribute
+    # "universal" wheels (i.e., they will need a wheel per Python version).
+    #
+    # Because setuptools and pip are bundled with _ensurepip and virtualenv,
+    # we need to use universal wheels. As a workaround, we
+    # override the versioned entry points in the wheel and generate the
+    # correct ones.
+    #
+    # To add the level of hack in this section of code, in order to support
+    # ensurepip this code will look for an ``ENSUREPIP_OPTIONS`` environment
+    # variable which will control which version scripts get installed.
+    #
+    # ENSUREPIP_OPTIONS=altinstall
+    #   - Only pipX.Y and easy_install-X.Y will be generated and installed
+    # ENSUREPIP_OPTIONS=install
+    #   - pipX.Y, pipX, easy_install-X.Y will be generated and installed. Note
+    #     that this option is technically if ENSUREPIP_OPTIONS is set and is
+    #     not altinstall
+    # DEFAULT
+    #   - The default behavior is to install pip, pipX, pipX.Y, easy_install
+    #     and easy_install-X.Y.
+    pip_script = console.pop("pip", None)
+    if pip_script:
+        if "ENSUREPIP_OPTIONS" not in os.environ:
+            scripts_to_generate.append("pip = " + pip_script)
+
+        if os.environ.get("ENSUREPIP_OPTIONS", "") != "altinstall":
+            scripts_to_generate.append(f"pip{sys.version_info[0]} = {pip_script}")
+
+        scripts_to_generate.append(f"pip{get_major_minor_version()} = {pip_script}")
+        # Delete any other versioned pip entry points
+        pip_ep = [k for k in console if re.match(r"pip(\d+(\.\d+)?)?$", k)]
+        for k in pip_ep:
+            del console[k]
+    easy_install_script = console.pop("easy_install", None)
+    if easy_install_script:
+        if "ENSUREPIP_OPTIONS" not in os.environ:
+            scripts_to_generate.append("easy_install = " + easy_install_script)
+
+        scripts_to_generate.append(
+            f"easy_install-{get_major_minor_version()} = {easy_install_script}"
+        )
+        # Delete any other versioned easy_install entry points
+        easy_install_ep = [
+            k for k in console if re.match(r"easy_install(-\d+\.\d+)?$", k)
+        ]
+        for k in easy_install_ep:
+            del console[k]
+
+    # Generate the console entry points specified in the wheel
+    scripts_to_generate.extend(starmap("{} = {}".format, console.items()))
+
+    return scripts_to_generate
+
+
+class ZipBackedFile:
+    def __init__(
+        self, src_record_path: RecordPath, dest_path: str, zip_file: ZipFile
+    ) -> None:
+        self.src_record_path = src_record_path
+        self.dest_path = dest_path
+        self._zip_file = zip_file
+        self.changed = False
+
+    def _getinfo(self) -> ZipInfo:
+        return self._zip_file.getinfo(self.src_record_path)
+
+    def save(self) -> None:
+        # When we open the output file below, any existing file is truncated
+        # before we start writing the new contents. This is fine in most
+        # cases, but can cause a segfault if pip has loaded a shared
+        # object (e.g. from pyopenssl through its vendored urllib3)
+        # Since the shared object is mmap'd an attempt to call a
+        # symbol in it will then cause a segfault. Unlinking the file
+        # allows writing of new contents while allowing the process to
+        # continue to use the old copy.
+        if os.path.exists(self.dest_path):
+            os.unlink(self.dest_path)
+
+        zipinfo = self._getinfo()
+
+        # optimization: the file is created by open(),
+        # skip the decompression when there is 0 bytes to decompress.
+        with open(self.dest_path, "wb") as dest:
+            if zipinfo.file_size > 0:
+                with self._zip_file.open(zipinfo) as f:
+                    blocksize = min(zipinfo.file_size, 1024 * 1024)
+                    shutil.copyfileobj(f, dest, blocksize)
+
+        if zip_item_is_executable(zipinfo):
+            set_extracted_file_to_default_mode_plus_executable(self.dest_path)
+
+
+class ScriptFile:
+    def __init__(self, file: "File") -> None:
+        self._file = file
+        self.src_record_path = self._file.src_record_path
+        self.dest_path = self._file.dest_path
+        self.changed = False
+
+    def save(self) -> None:
+        self._file.save()
+        self.changed = fix_script(self.dest_path)
+
+
+class MissingCallableSuffix(InstallationError):
+    def __init__(self, entry_point: str) -> None:
+        super().__init__(
+            f"Invalid script entry point: {entry_point} - A callable "
+            "suffix is required. Cf https://packaging.python.org/"
+            "specifications/entry-points/#use-for-scripts for more "
+            "information."
+        )
+
+
+def _raise_for_invalid_entrypoint(specification: str) -> None:
+    entry = get_export_entry(specification)
+    if entry is not None and entry.suffix is None:
+        raise MissingCallableSuffix(str(entry))
+
+
+class PipScriptMaker(ScriptMaker):
+    def make(
+        self, specification: str, options: Optional[Dict[str, Any]] = None
+    ) -> List[str]:
+        _raise_for_invalid_entrypoint(specification)
+        return super().make(specification, options)
+
+
+def _install_wheel(  # noqa: C901, PLR0915 function is too long
+    name: str,
+    wheel_zip: ZipFile,
+    wheel_path: str,
+    scheme: Scheme,
+    pycompile: bool = True,
+    warn_script_location: bool = True,
+    direct_url: Optional[DirectUrl] = None,
+    requested: bool = False,
+) -> None:
+    """Install a wheel.
+
+    :param name: Name of the project to install
+    :param wheel_zip: open ZipFile for wheel being installed
+    :param scheme: Distutils scheme dictating the install directories
+    :param req_description: String used in place of the requirement, for
+        logging
+    :param pycompile: Whether to byte-compile installed Python files
+    :param warn_script_location: Whether to check that scripts are installed
+        into a directory on PATH
+    :raises UnsupportedWheel:
+        * when the directory holds an unpacked wheel with incompatible
+          Wheel-Version
+        * when the .dist-info dir does not match the wheel
+    """
+    info_dir, metadata = parse_wheel(wheel_zip, name)
+
+    if wheel_root_is_purelib(metadata):
+        lib_dir = scheme.purelib
+    else:
+        lib_dir = scheme.platlib
+
+    # Record details of the files moved
+    #   installed = files copied from the wheel to the destination
+    #   changed = files changed while installing (scripts #! line typically)
+    #   generated = files newly generated during the install (script wrappers)
+    installed: Dict[RecordPath, RecordPath] = {}
+    changed: Set[RecordPath] = set()
+    generated: List[str] = []
+
+    def record_installed(
+        srcfile: RecordPath, destfile: str, modified: bool = False
+    ) -> None:
+        """Map archive RECORD paths to installation RECORD paths."""
+        newpath = _fs_to_record_path(destfile, lib_dir)
+        installed[srcfile] = newpath
+        if modified:
+            changed.add(newpath)
+
+    def is_dir_path(path: RecordPath) -> bool:
+        return path.endswith("/")
+
+    def assert_no_path_traversal(dest_dir_path: str, target_path: str) -> None:
+        if not is_within_directory(dest_dir_path, target_path):
+            message = (
+                "The wheel {!r} has a file {!r} trying to install"
+                " outside the target directory {!r}"
+            )
+            raise InstallationError(
+                message.format(wheel_path, target_path, dest_dir_path)
+            )
+
+    def root_scheme_file_maker(
+        zip_file: ZipFile, dest: str
+    ) -> Callable[[RecordPath], "File"]:
+        def make_root_scheme_file(record_path: RecordPath) -> "File":
+            normed_path = os.path.normpath(record_path)
+            dest_path = os.path.join(dest, normed_path)
+            assert_no_path_traversal(dest, dest_path)
+            return ZipBackedFile(record_path, dest_path, zip_file)
+
+        return make_root_scheme_file
+
+    def data_scheme_file_maker(
+        zip_file: ZipFile, scheme: Scheme
+    ) -> Callable[[RecordPath], "File"]:
+        scheme_paths = {key: getattr(scheme, key) for key in SCHEME_KEYS}
+
+        def make_data_scheme_file(record_path: RecordPath) -> "File":
+            normed_path = os.path.normpath(record_path)
+            try:
+                _, scheme_key, dest_subpath = normed_path.split(os.path.sep, 2)
+            except ValueError:
+                message = (
+                    f"Unexpected file in {wheel_path}: {record_path!r}. .data directory"
+                    " contents should be named like: '<scheme key>/<path>'."
+                )
+                raise InstallationError(message)
+
+            try:
+                scheme_path = scheme_paths[scheme_key]
+            except KeyError:
+                valid_scheme_keys = ", ".join(sorted(scheme_paths))
+                message = (
+                    f"Unknown scheme key used in {wheel_path}: {scheme_key} "
+                    f"(for file {record_path!r}). .data directory contents "
+                    f"should be in subdirectories named with a valid scheme "
+                    f"key ({valid_scheme_keys})"
+                )
+                raise InstallationError(message)
+
+            dest_path = os.path.join(scheme_path, dest_subpath)
+            assert_no_path_traversal(scheme_path, dest_path)
+            return ZipBackedFile(record_path, dest_path, zip_file)
+
+        return make_data_scheme_file
+
+    def is_data_scheme_path(path: RecordPath) -> bool:
+        return path.split("/", 1)[0].endswith(".data")
+
+    paths = cast(List[RecordPath], wheel_zip.namelist())
+    file_paths = filterfalse(is_dir_path, paths)
+    root_scheme_paths, data_scheme_paths = partition(is_data_scheme_path, file_paths)
+
+    make_root_scheme_file = root_scheme_file_maker(wheel_zip, lib_dir)
+    files: Iterator[File] = map(make_root_scheme_file, root_scheme_paths)
+
+    def is_script_scheme_path(path: RecordPath) -> bool:
+        parts = path.split("/", 2)
+        return len(parts) > 2 and parts[0].endswith(".data") and parts[1] == "scripts"
+
+    other_scheme_paths, script_scheme_paths = partition(
+        is_script_scheme_path, data_scheme_paths
+    )
+
+    make_data_scheme_file = data_scheme_file_maker(wheel_zip, scheme)
+    other_scheme_files = map(make_data_scheme_file, other_scheme_paths)
+    files = chain(files, other_scheme_files)
+
+    # Get the defined entry points
+    distribution = get_wheel_distribution(
+        FilesystemWheel(wheel_path),
+        canonicalize_name(name),
+    )
+    console, gui = get_entrypoints(distribution)
+
+    def is_entrypoint_wrapper(file: "File") -> bool:
+        # EP, EP.exe and EP-script.py are scripts generated for
+        # entry point EP by setuptools
+        path = file.dest_path
+        name = os.path.basename(path)
+        if name.lower().endswith(".exe"):
+            matchname = name[:-4]
+        elif name.lower().endswith("-script.py"):
+            matchname = name[:-10]
+        elif name.lower().endswith(".pya"):
+            matchname = name[:-4]
+        else:
+            matchname = name
+        # Ignore setuptools-generated scripts
+        return matchname in console or matchname in gui
+
+    script_scheme_files: Iterator[File] = map(
+        make_data_scheme_file, script_scheme_paths
+    )
+    script_scheme_files = filterfalse(is_entrypoint_wrapper, script_scheme_files)
+    script_scheme_files = map(ScriptFile, script_scheme_files)
+    files = chain(files, script_scheme_files)
+
+    existing_parents = set()
+    for file in files:
+        # directory creation is lazy and after file filtering
+        # to ensure we don't install empty dirs; empty dirs can't be
+        # uninstalled.
+        parent_dir = os.path.dirname(file.dest_path)
+        if parent_dir not in existing_parents:
+            ensure_dir(parent_dir)
+            existing_parents.add(parent_dir)
+        file.save()
+        record_installed(file.src_record_path, file.dest_path, file.changed)
+
+    def pyc_source_file_paths() -> Generator[str, None, None]:
+        # We de-duplicate installation paths, since there can be overlap (e.g.
+        # file in .data maps to same location as file in wheel root).
+        # Sorting installation paths makes it easier to reproduce and debug
+        # issues related to permissions on existing files.
+        for installed_path in sorted(set(installed.values())):
+            full_installed_path = os.path.join(lib_dir, installed_path)
+            if not os.path.isfile(full_installed_path):
+                continue
+            if not full_installed_path.endswith(".py"):
+                continue
+            yield full_installed_path
+
+    def pyc_output_path(path: str) -> str:
+        """Return the path the pyc file would have been written to."""
+        return importlib.util.cache_from_source(path)
+
+    # Compile all of the pyc files for the installed files
+    if pycompile:
+        with contextlib.redirect_stdout(
+            StreamWrapper.from_stream(sys.stdout)
+        ) as stdout:
+            with warnings.catch_warnings():
+                warnings.filterwarnings("ignore")
+                for path in pyc_source_file_paths():
+                    success = compileall.compile_file(path, force=True, quiet=True)
+                    if success:
+                        pyc_path = pyc_output_path(path)
+                        assert os.path.exists(pyc_path)
+                        pyc_record_path = cast(
+                            "RecordPath", pyc_path.replace(os.path.sep, "/")
+                        )
+                        record_installed(pyc_record_path, pyc_path)
+        logger.debug(stdout.getvalue())
+
+    maker = PipScriptMaker(None, scheme.scripts)
+
+    # Ensure old scripts are overwritten.
+    # See https://github.com/pypa/pip/issues/1800
+    maker.clobber = True
+
+    # Ensure we don't generate any variants for scripts because this is almost
+    # never what somebody wants.
+    # See https://bitbucket.org/pypa/distlib/issue/35/
+    maker.variants = {""}
+
+    # This is required because otherwise distlib creates scripts that are not
+    # executable.
+    # See https://bitbucket.org/pypa/distlib/issue/32/
+    maker.set_mode = True
+
+    # Generate the console and GUI entry points specified in the wheel
+    scripts_to_generate = get_console_script_specs(console)
+
+    gui_scripts_to_generate = list(starmap("{} = {}".format, gui.items()))
+
+    generated_console_scripts = maker.make_multiple(scripts_to_generate)
+    generated.extend(generated_console_scripts)
+
+    generated.extend(maker.make_multiple(gui_scripts_to_generate, {"gui": True}))
+
+    if warn_script_location:
+        msg = message_about_scripts_not_on_PATH(generated_console_scripts)
+        if msg is not None:
+            logger.warning(msg)
+
+    generated_file_mode = 0o666 & ~current_umask()
+
+    @contextlib.contextmanager
+    def _generate_file(path: str, **kwargs: Any) -> Generator[BinaryIO, None, None]:
+        with adjacent_tmp_file(path, **kwargs) as f:
+            yield f
+        os.chmod(f.name, generated_file_mode)
+        replace(f.name, path)
+
+    dest_info_dir = os.path.join(lib_dir, info_dir)
+
+    # Record pip as the installer
+    installer_path = os.path.join(dest_info_dir, "INSTALLER")
+    with _generate_file(installer_path) as installer_file:
+        installer_file.write(b"pip\n")
+    generated.append(installer_path)
+
+    # Record the PEP 610 direct URL reference
+    if direct_url is not None:
+        direct_url_path = os.path.join(dest_info_dir, DIRECT_URL_METADATA_NAME)
+        with _generate_file(direct_url_path) as direct_url_file:
+            direct_url_file.write(direct_url.to_json().encode("utf-8"))
+        generated.append(direct_url_path)
+
+    # Record the REQUESTED file
+    if requested:
+        requested_path = os.path.join(dest_info_dir, "REQUESTED")
+        with open(requested_path, "wb"):
+            pass
+        generated.append(requested_path)
+
+    record_text = distribution.read_text("RECORD")
+    record_rows = list(csv.reader(record_text.splitlines()))
+
+    rows = get_csv_rows_for_installed(
+        record_rows,
+        installed=installed,
+        changed=changed,
+        generated=generated,
+        lib_dir=lib_dir,
+    )
+
+    # Record details of all files installed
+    record_path = os.path.join(dest_info_dir, "RECORD")
+
+    with _generate_file(record_path, **csv_io_kwargs("w")) as record_file:
+        # Explicitly cast to typing.IO[str] as a workaround for the mypy error:
+        # "writer" has incompatible type "BinaryIO"; expected "_Writer"
+        writer = csv.writer(cast("IO[str]", record_file))
+        writer.writerows(_normalized_outrows(rows))
+
+
+@contextlib.contextmanager
+def req_error_context(req_description: str) -> Generator[None, None, None]:
+    try:
+        yield
+    except InstallationError as e:
+        message = f"For req: {req_description}. {e.args[0]}"
+        raise InstallationError(message) from e
+
+
+def install_wheel(
+    name: str,
+    wheel_path: str,
+    scheme: Scheme,
+    req_description: str,
+    pycompile: bool = True,
+    warn_script_location: bool = True,
+    direct_url: Optional[DirectUrl] = None,
+    requested: bool = False,
+) -> None:
+    with ZipFile(wheel_path, allowZip64=True) as z:
+        with req_error_context(req_description):
+            _install_wheel(
+                name=name,
+                wheel_zip=z,
+                wheel_path=wheel_path,
+                scheme=scheme,
+                pycompile=pycompile,
+                warn_script_location=warn_script_location,
+                direct_url=direct_url,
+                requested=requested,
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/prepare.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/prepare.py
new file mode 100644
index 00000000..e6aa3447
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/operations/prepare.py
@@ -0,0 +1,732 @@
+"""Prepares a distribution for installation
+"""
+
+# The following comment should be removed at some point in the future.
+# mypy: strict-optional=False
+
+import mimetypes
+import os
+import shutil
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Dict, Iterable, List, Optional
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.distributions import make_distribution_for_install_requirement
+from pip._internal.distributions.installed import InstalledDistribution
+from pip._internal.exceptions import (
+    DirectoryUrlHashUnsupported,
+    HashMismatch,
+    HashUnpinned,
+    InstallationError,
+    MetadataInconsistent,
+    NetworkConnectionError,
+    VcsHashUnsupported,
+)
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.metadata import BaseDistribution, get_metadata_distribution
+from pip._internal.models.direct_url import ArchiveInfo
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.network.download import BatchDownloader, Downloader
+from pip._internal.network.lazy_wheel import (
+    HTTPRangeRequestUnsupported,
+    dist_from_wheel_url,
+)
+from pip._internal.network.session import PipSession
+from pip._internal.operations.build.build_tracker import BuildTracker
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.utils._log import getLogger
+from pip._internal.utils.direct_url_helpers import (
+    direct_url_for_editable,
+    direct_url_from_link,
+)
+from pip._internal.utils.hashes import Hashes, MissingHashes
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import (
+    display_path,
+    hash_file,
+    hide_url,
+    redact_auth_from_requirement,
+)
+from pip._internal.utils.temp_dir import TempDirectory
+from pip._internal.utils.unpacking import unpack_file
+from pip._internal.vcs import vcs
+
+logger = getLogger(__name__)
+
+
+def _get_prepared_distribution(
+    req: InstallRequirement,
+    build_tracker: BuildTracker,
+    finder: PackageFinder,
+    build_isolation: bool,
+    check_build_deps: bool,
+) -> BaseDistribution:
+    """Prepare a distribution for installation."""
+    abstract_dist = make_distribution_for_install_requirement(req)
+    tracker_id = abstract_dist.build_tracker_id
+    if tracker_id is not None:
+        with build_tracker.track(req, tracker_id):
+            abstract_dist.prepare_distribution_metadata(
+                finder, build_isolation, check_build_deps
+            )
+    return abstract_dist.get_metadata_distribution()
+
+
+def unpack_vcs_link(link: Link, location: str, verbosity: int) -> None:
+    vcs_backend = vcs.get_backend_for_scheme(link.scheme)
+    assert vcs_backend is not None
+    vcs_backend.unpack(location, url=hide_url(link.url), verbosity=verbosity)
+
+
+@dataclass
+class File:
+    path: str
+    content_type: Optional[str] = None
+
+    def __post_init__(self) -> None:
+        if self.content_type is None:
+            self.content_type = mimetypes.guess_type(self.path)[0]
+
+
+def get_http_url(
+    link: Link,
+    download: Downloader,
+    download_dir: Optional[str] = None,
+    hashes: Optional[Hashes] = None,
+) -> File:
+    temp_dir = TempDirectory(kind="unpack", globally_managed=True)
+    # If a download dir is specified, is the file already downloaded there?
+    already_downloaded_path = None
+    if download_dir:
+        already_downloaded_path = _check_download_dir(link, download_dir, hashes)
+
+    if already_downloaded_path:
+        from_path = already_downloaded_path
+        content_type = None
+    else:
+        # let's download to a tmp dir
+        from_path, content_type = download(link, temp_dir.path)
+        if hashes:
+            hashes.check_against_path(from_path)
+
+    return File(from_path, content_type)
+
+
+def get_file_url(
+    link: Link, download_dir: Optional[str] = None, hashes: Optional[Hashes] = None
+) -> File:
+    """Get file and optionally check its hash."""
+    # If a download dir is specified, is the file already there and valid?
+    already_downloaded_path = None
+    if download_dir:
+        already_downloaded_path = _check_download_dir(link, download_dir, hashes)
+
+    if already_downloaded_path:
+        from_path = already_downloaded_path
+    else:
+        from_path = link.file_path
+
+    # If --require-hashes is off, `hashes` is either empty, the
+    # link's embedded hash, or MissingHashes; it is required to
+    # match. If --require-hashes is on, we are satisfied by any
+    # hash in `hashes` matching: a URL-based or an option-based
+    # one; no internet-sourced hash will be in `hashes`.
+    if hashes:
+        hashes.check_against_path(from_path)
+    return File(from_path, None)
+
+
+def unpack_url(
+    link: Link,
+    location: str,
+    download: Downloader,
+    verbosity: int,
+    download_dir: Optional[str] = None,
+    hashes: Optional[Hashes] = None,
+) -> Optional[File]:
+    """Unpack link into location, downloading if required.
+
+    :param hashes: A Hashes object, one of whose embedded hashes must match,
+        or HashMismatch will be raised. If the Hashes is empty, no matches are
+        required, and unhashable types of requirements (like VCS ones, which
+        would ordinarily raise HashUnsupported) are allowed.
+    """
+    # non-editable vcs urls
+    if link.is_vcs:
+        unpack_vcs_link(link, location, verbosity=verbosity)
+        return None
+
+    assert not link.is_existing_dir()
+
+    # file urls
+    if link.is_file:
+        file = get_file_url(link, download_dir, hashes=hashes)
+
+    # http urls
+    else:
+        file = get_http_url(
+            link,
+            download,
+            download_dir,
+            hashes=hashes,
+        )
+
+    # unpack the archive to the build dir location. even when only downloading
+    # archives, they have to be unpacked to parse dependencies, except wheels
+    if not link.is_wheel:
+        unpack_file(file.path, location, file.content_type)
+
+    return file
+
+
+def _check_download_dir(
+    link: Link,
+    download_dir: str,
+    hashes: Optional[Hashes],
+    warn_on_hash_mismatch: bool = True,
+) -> Optional[str]:
+    """Check download_dir for previously downloaded file with correct hash
+    If a correct file is found return its path else None
+    """
+    download_path = os.path.join(download_dir, link.filename)
+
+    if not os.path.exists(download_path):
+        return None
+
+    # If already downloaded, does its hash match?
+    logger.info("File was already downloaded %s", download_path)
+    if hashes:
+        try:
+            hashes.check_against_path(download_path)
+        except HashMismatch:
+            if warn_on_hash_mismatch:
+                logger.warning(
+                    "Previously-downloaded file %s has bad hash. Re-downloading.",
+                    download_path,
+                )
+            os.unlink(download_path)
+            return None
+    return download_path
+
+
+class RequirementPreparer:
+    """Prepares a Requirement"""
+
+    def __init__(
+        self,
+        build_dir: str,
+        download_dir: Optional[str],
+        src_dir: str,
+        build_isolation: bool,
+        check_build_deps: bool,
+        build_tracker: BuildTracker,
+        session: PipSession,
+        progress_bar: str,
+        finder: PackageFinder,
+        require_hashes: bool,
+        use_user_site: bool,
+        lazy_wheel: bool,
+        verbosity: int,
+        legacy_resolver: bool,
+    ) -> None:
+        super().__init__()
+
+        self.src_dir = src_dir
+        self.build_dir = build_dir
+        self.build_tracker = build_tracker
+        self._session = session
+        self._download = Downloader(session, progress_bar)
+        self._batch_download = BatchDownloader(session, progress_bar)
+        self.finder = finder
+
+        # Where still-packed archives should be written to. If None, they are
+        # not saved, and are deleted immediately after unpacking.
+        self.download_dir = download_dir
+
+        # Is build isolation allowed?
+        self.build_isolation = build_isolation
+
+        # Should check build dependencies?
+        self.check_build_deps = check_build_deps
+
+        # Should hash-checking be required?
+        self.require_hashes = require_hashes
+
+        # Should install in user site-packages?
+        self.use_user_site = use_user_site
+
+        # Should wheels be downloaded lazily?
+        self.use_lazy_wheel = lazy_wheel
+
+        # How verbose should underlying tooling be?
+        self.verbosity = verbosity
+
+        # Are we using the legacy resolver?
+        self.legacy_resolver = legacy_resolver
+
+        # Memoized downloaded files, as mapping of url: path.
+        self._downloaded: Dict[str, str] = {}
+
+        # Previous "header" printed for a link-based InstallRequirement
+        self._previous_requirement_header = ("", "")
+
+    def _log_preparing_link(self, req: InstallRequirement) -> None:
+        """Provide context for the requirement being prepared."""
+        if req.link.is_file and not req.is_wheel_from_cache:
+            message = "Processing %s"
+            information = str(display_path(req.link.file_path))
+        else:
+            message = "Collecting %s"
+            information = redact_auth_from_requirement(req.req) if req.req else str(req)
+
+        # If we used req.req, inject requirement source if available (this
+        # would already be included if we used req directly)
+        if req.req and req.comes_from:
+            if isinstance(req.comes_from, str):
+                comes_from: Optional[str] = req.comes_from
+            else:
+                comes_from = req.comes_from.from_path()
+            if comes_from:
+                information += f" (from {comes_from})"
+
+        if (message, information) != self._previous_requirement_header:
+            self._previous_requirement_header = (message, information)
+            logger.info(message, information)
+
+        if req.is_wheel_from_cache:
+            with indent_log():
+                logger.info("Using cached %s", req.link.filename)
+
+    def _ensure_link_req_src_dir(
+        self, req: InstallRequirement, parallel_builds: bool
+    ) -> None:
+        """Ensure source_dir of a linked InstallRequirement."""
+        # Since source_dir is only set for editable requirements.
+        if req.link.is_wheel:
+            # We don't need to unpack wheels, so no need for a source
+            # directory.
+            return
+        assert req.source_dir is None
+        if req.link.is_existing_dir():
+            # build local directories in-tree
+            req.source_dir = req.link.file_path
+            return
+
+        # We always delete unpacked sdists after pip runs.
+        req.ensure_has_source_dir(
+            self.build_dir,
+            autodelete=True,
+            parallel_builds=parallel_builds,
+        )
+        req.ensure_pristine_source_checkout()
+
+    def _get_linked_req_hashes(self, req: InstallRequirement) -> Hashes:
+        # By the time this is called, the requirement's link should have
+        # been checked so we can tell what kind of requirements req is
+        # and raise some more informative errors than otherwise.
+        # (For example, we can raise VcsHashUnsupported for a VCS URL
+        # rather than HashMissing.)
+        if not self.require_hashes:
+            return req.hashes(trust_internet=True)
+
+        # We could check these first 2 conditions inside unpack_url
+        # and save repetition of conditions, but then we would
+        # report less-useful error messages for unhashable
+        # requirements, complaining that there's no hash provided.
+        if req.link.is_vcs:
+            raise VcsHashUnsupported()
+        if req.link.is_existing_dir():
+            raise DirectoryUrlHashUnsupported()
+
+        # Unpinned packages are asking for trouble when a new version
+        # is uploaded.  This isn't a security check, but it saves users
+        # a surprising hash mismatch in the future.
+        # file:/// URLs aren't pinnable, so don't complain about them
+        # not being pinned.
+        if not req.is_direct and not req.is_pinned:
+            raise HashUnpinned()
+
+        # If known-good hashes are missing for this requirement,
+        # shim it with a facade object that will provoke hash
+        # computation and then raise a HashMissing exception
+        # showing the user what the hash should be.
+        return req.hashes(trust_internet=False) or MissingHashes()
+
+    def _fetch_metadata_only(
+        self,
+        req: InstallRequirement,
+    ) -> Optional[BaseDistribution]:
+        if self.legacy_resolver:
+            logger.debug(
+                "Metadata-only fetching is not used in the legacy resolver",
+            )
+            return None
+        if self.require_hashes:
+            logger.debug(
+                "Metadata-only fetching is not used as hash checking is required",
+            )
+            return None
+        # Try PEP 658 metadata first, then fall back to lazy wheel if unavailable.
+        return self._fetch_metadata_using_link_data_attr(
+            req
+        ) or self._fetch_metadata_using_lazy_wheel(req.link)
+
+    def _fetch_metadata_using_link_data_attr(
+        self,
+        req: InstallRequirement,
+    ) -> Optional[BaseDistribution]:
+        """Fetch metadata from the data-dist-info-metadata attribute, if possible."""
+        # (1) Get the link to the metadata file, if provided by the backend.
+        metadata_link = req.link.metadata_link()
+        if metadata_link is None:
+            return None
+        assert req.req is not None
+        logger.verbose(
+            "Obtaining dependency information for %s from %s",
+            req.req,
+            metadata_link,
+        )
+        # (2) Download the contents of the METADATA file, separate from the dist itself.
+        metadata_file = get_http_url(
+            metadata_link,
+            self._download,
+            hashes=metadata_link.as_hashes(),
+        )
+        with open(metadata_file.path, "rb") as f:
+            metadata_contents = f.read()
+        # (3) Generate a dist just from those file contents.
+        metadata_dist = get_metadata_distribution(
+            metadata_contents,
+            req.link.filename,
+            req.req.name,
+        )
+        # (4) Ensure the Name: field from the METADATA file matches the name from the
+        #     install requirement.
+        #
+        #     NB: raw_name will fall back to the name from the install requirement if
+        #     the Name: field is not present, but it's noted in the raw_name docstring
+        #     that that should NEVER happen anyway.
+        if canonicalize_name(metadata_dist.raw_name) != canonicalize_name(req.req.name):
+            raise MetadataInconsistent(
+                req, "Name", req.req.name, metadata_dist.raw_name
+            )
+        return metadata_dist
+
+    def _fetch_metadata_using_lazy_wheel(
+        self,
+        link: Link,
+    ) -> Optional[BaseDistribution]:
+        """Fetch metadata using lazy wheel, if possible."""
+        # --use-feature=fast-deps must be provided.
+        if not self.use_lazy_wheel:
+            return None
+        if link.is_file or not link.is_wheel:
+            logger.debug(
+                "Lazy wheel is not used as %r does not point to a remote wheel",
+                link,
+            )
+            return None
+
+        wheel = Wheel(link.filename)
+        name = canonicalize_name(wheel.name)
+        logger.info(
+            "Obtaining dependency information from %s %s",
+            name,
+            wheel.version,
+        )
+        url = link.url.split("#", 1)[0]
+        try:
+            return dist_from_wheel_url(name, url, self._session)
+        except HTTPRangeRequestUnsupported:
+            logger.debug("%s does not support range requests", url)
+            return None
+
+    def _complete_partial_requirements(
+        self,
+        partially_downloaded_reqs: Iterable[InstallRequirement],
+        parallel_builds: bool = False,
+    ) -> None:
+        """Download any requirements which were only fetched by metadata."""
+        # Download to a temporary directory. These will be copied over as
+        # needed for downstream 'download', 'wheel', and 'install' commands.
+        temp_dir = TempDirectory(kind="unpack", globally_managed=True).path
+
+        # Map each link to the requirement that owns it. This allows us to set
+        # `req.local_file_path` on the appropriate requirement after passing
+        # all the links at once into BatchDownloader.
+        links_to_fully_download: Dict[Link, InstallRequirement] = {}
+        for req in partially_downloaded_reqs:
+            assert req.link
+            links_to_fully_download[req.link] = req
+
+        batch_download = self._batch_download(
+            links_to_fully_download.keys(),
+            temp_dir,
+        )
+        for link, (filepath, _) in batch_download:
+            logger.debug("Downloading link %s to %s", link, filepath)
+            req = links_to_fully_download[link]
+            # Record the downloaded file path so wheel reqs can extract a Distribution
+            # in .get_dist().
+            req.local_file_path = filepath
+            # Record that the file is downloaded so we don't do it again in
+            # _prepare_linked_requirement().
+            self._downloaded[req.link.url] = filepath
+
+            # If this is an sdist, we need to unpack it after downloading, but the
+            # .source_dir won't be set up until we are in _prepare_linked_requirement().
+            # Add the downloaded archive to the install requirement to unpack after
+            # preparing the source dir.
+            if not req.is_wheel:
+                req.needs_unpacked_archive(Path(filepath))
+
+        # This step is necessary to ensure all lazy wheels are processed
+        # successfully by the 'download', 'wheel', and 'install' commands.
+        for req in partially_downloaded_reqs:
+            self._prepare_linked_requirement(req, parallel_builds)
+
+    def prepare_linked_requirement(
+        self, req: InstallRequirement, parallel_builds: bool = False
+    ) -> BaseDistribution:
+        """Prepare a requirement to be obtained from req.link."""
+        assert req.link
+        self._log_preparing_link(req)
+        with indent_log():
+            # Check if the relevant file is already available
+            # in the download directory
+            file_path = None
+            if self.download_dir is not None and req.link.is_wheel:
+                hashes = self._get_linked_req_hashes(req)
+                file_path = _check_download_dir(
+                    req.link,
+                    self.download_dir,
+                    hashes,
+                    # When a locally built wheel has been found in cache, we don't warn
+                    # about re-downloading when the already downloaded wheel hash does
+                    # not match. This is because the hash must be checked against the
+                    # original link, not the cached link. It that case the already
+                    # downloaded file will be removed and re-fetched from cache (which
+                    # implies a hash check against the cache entry's origin.json).
+                    warn_on_hash_mismatch=not req.is_wheel_from_cache,
+                )
+
+            if file_path is not None:
+                # The file is already available, so mark it as downloaded
+                self._downloaded[req.link.url] = file_path
+            else:
+                # The file is not available, attempt to fetch only metadata
+                metadata_dist = self._fetch_metadata_only(req)
+                if metadata_dist is not None:
+                    req.needs_more_preparation = True
+                    return metadata_dist
+
+            # None of the optimizations worked, fully prepare the requirement
+            return self._prepare_linked_requirement(req, parallel_builds)
+
+    def prepare_linked_requirements_more(
+        self, reqs: Iterable[InstallRequirement], parallel_builds: bool = False
+    ) -> None:
+        """Prepare linked requirements more, if needed."""
+        reqs = [req for req in reqs if req.needs_more_preparation]
+        for req in reqs:
+            # Determine if any of these requirements were already downloaded.
+            if self.download_dir is not None and req.link.is_wheel:
+                hashes = self._get_linked_req_hashes(req)
+                file_path = _check_download_dir(req.link, self.download_dir, hashes)
+                if file_path is not None:
+                    self._downloaded[req.link.url] = file_path
+                    req.needs_more_preparation = False
+
+        # Prepare requirements we found were already downloaded for some
+        # reason. The other downloads will be completed separately.
+        partially_downloaded_reqs: List[InstallRequirement] = []
+        for req in reqs:
+            if req.needs_more_preparation:
+                partially_downloaded_reqs.append(req)
+            else:
+                self._prepare_linked_requirement(req, parallel_builds)
+
+        # TODO: separate this part out from RequirementPreparer when the v1
+        # resolver can be removed!
+        self._complete_partial_requirements(
+            partially_downloaded_reqs,
+            parallel_builds=parallel_builds,
+        )
+
+    def _prepare_linked_requirement(
+        self, req: InstallRequirement, parallel_builds: bool
+    ) -> BaseDistribution:
+        assert req.link
+        link = req.link
+
+        hashes = self._get_linked_req_hashes(req)
+
+        if hashes and req.is_wheel_from_cache:
+            assert req.download_info is not None
+            assert link.is_wheel
+            assert link.is_file
+            # We need to verify hashes, and we have found the requirement in the cache
+            # of locally built wheels.
+            if (
+                isinstance(req.download_info.info, ArchiveInfo)
+                and req.download_info.info.hashes
+                and hashes.has_one_of(req.download_info.info.hashes)
+            ):
+                # At this point we know the requirement was built from a hashable source
+                # artifact, and we verified that the cache entry's hash of the original
+                # artifact matches one of the hashes we expect. We don't verify hashes
+                # against the cached wheel, because the wheel is not the original.
+                hashes = None
+            else:
+                logger.warning(
+                    "The hashes of the source archive found in cache entry "
+                    "don't match, ignoring cached built wheel "
+                    "and re-downloading source."
+                )
+                req.link = req.cached_wheel_source_link
+                link = req.link
+
+        self._ensure_link_req_src_dir(req, parallel_builds)
+
+        if link.is_existing_dir():
+            local_file = None
+        elif link.url not in self._downloaded:
+            try:
+                local_file = unpack_url(
+                    link,
+                    req.source_dir,
+                    self._download,
+                    self.verbosity,
+                    self.download_dir,
+                    hashes,
+                )
+            except NetworkConnectionError as exc:
+                raise InstallationError(
+                    f"Could not install requirement {req} because of HTTP "
+                    f"error {exc} for URL {link}"
+                )
+        else:
+            file_path = self._downloaded[link.url]
+            if hashes:
+                hashes.check_against_path(file_path)
+            local_file = File(file_path, content_type=None)
+
+        # If download_info is set, we got it from the wheel cache.
+        if req.download_info is None:
+            # Editables don't go through this function (see
+            # prepare_editable_requirement).
+            assert not req.editable
+            req.download_info = direct_url_from_link(link, req.source_dir)
+            # Make sure we have a hash in download_info. If we got it as part of the
+            # URL, it will have been verified and we can rely on it. Otherwise we
+            # compute it from the downloaded file.
+            # FIXME: https://github.com/pypa/pip/issues/11943
+            if (
+                isinstance(req.download_info.info, ArchiveInfo)
+                and not req.download_info.info.hashes
+                and local_file
+            ):
+                hash = hash_file(local_file.path)[0].hexdigest()
+                # We populate info.hash for backward compatibility.
+                # This will automatically populate info.hashes.
+                req.download_info.info.hash = f"sha256={hash}"
+
+        # For use in later processing,
+        # preserve the file path on the requirement.
+        if local_file:
+            req.local_file_path = local_file.path
+
+        dist = _get_prepared_distribution(
+            req,
+            self.build_tracker,
+            self.finder,
+            self.build_isolation,
+            self.check_build_deps,
+        )
+        return dist
+
+    def save_linked_requirement(self, req: InstallRequirement) -> None:
+        assert self.download_dir is not None
+        assert req.link is not None
+        link = req.link
+        if link.is_vcs or (link.is_existing_dir() and req.editable):
+            # Make a .zip of the source_dir we already created.
+            req.archive(self.download_dir)
+            return
+
+        if link.is_existing_dir():
+            logger.debug(
+                "Not copying link to destination directory "
+                "since it is a directory: %s",
+                link,
+            )
+            return
+        if req.local_file_path is None:
+            # No distribution was downloaded for this requirement.
+            return
+
+        download_location = os.path.join(self.download_dir, link.filename)
+        if not os.path.exists(download_location):
+            shutil.copy(req.local_file_path, download_location)
+            download_path = display_path(download_location)
+            logger.info("Saved %s", download_path)
+
+    def prepare_editable_requirement(
+        self,
+        req: InstallRequirement,
+    ) -> BaseDistribution:
+        """Prepare an editable requirement."""
+        assert req.editable, "cannot prepare a non-editable req as editable"
+
+        logger.info("Obtaining %s", req)
+
+        with indent_log():
+            if self.require_hashes:
+                raise InstallationError(
+                    f"The editable requirement {req} cannot be installed when "
+                    "requiring hashes, because there is no single file to "
+                    "hash."
+                )
+            req.ensure_has_source_dir(self.src_dir)
+            req.update_editable()
+            assert req.source_dir
+            req.download_info = direct_url_for_editable(req.unpacked_source_directory)
+
+            dist = _get_prepared_distribution(
+                req,
+                self.build_tracker,
+                self.finder,
+                self.build_isolation,
+                self.check_build_deps,
+            )
+
+            req.check_if_exists(self.use_user_site)
+
+        return dist
+
+    def prepare_installed_requirement(
+        self,
+        req: InstallRequirement,
+        skip_reason: str,
+    ) -> BaseDistribution:
+        """Prepare an already-installed requirement."""
+        assert req.satisfied_by, "req should have been satisfied but isn't"
+        assert skip_reason is not None, (
+            "did not get skip reason skipped but req.satisfied_by "
+            f"is set to {req.satisfied_by}"
+        )
+        logger.info(
+            "Requirement %s: %s (%s)", skip_reason, req, req.satisfied_by.version
+        )
+        with indent_log():
+            if self.require_hashes:
+                logger.debug(
+                    "Since it is already installed, we are trusting this "
+                    "package without checking its hash. To ensure a "
+                    "completely repeatable environment, install into an "
+                    "empty virtualenv."
+                )
+            return InstalledDistribution(req).get_metadata_distribution()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/pyproject.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/pyproject.py
new file mode 100644
index 00000000..0e8452f3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/pyproject.py
@@ -0,0 +1,185 @@
+import importlib.util
+import os
+import sys
+from collections import namedtuple
+from typing import Any, List, Optional
+
+if sys.version_info >= (3, 11):
+    import tomllib
+else:
+    from pip._vendor import tomli as tomllib
+
+from pip._vendor.packaging.requirements import InvalidRequirement
+
+from pip._internal.exceptions import (
+    InstallationError,
+    InvalidPyProjectBuildRequires,
+    MissingPyProjectBuildRequires,
+)
+from pip._internal.utils.packaging import get_requirement
+
+
+def _is_list_of_str(obj: Any) -> bool:
+    return isinstance(obj, list) and all(isinstance(item, str) for item in obj)
+
+
+def make_pyproject_path(unpacked_source_directory: str) -> str:
+    return os.path.join(unpacked_source_directory, "pyproject.toml")
+
+
+BuildSystemDetails = namedtuple(
+    "BuildSystemDetails", ["requires", "backend", "check", "backend_path"]
+)
+
+
+def load_pyproject_toml(
+    use_pep517: Optional[bool], pyproject_toml: str, setup_py: str, req_name: str
+) -> Optional[BuildSystemDetails]:
+    """Load the pyproject.toml file.
+
+    Parameters:
+        use_pep517 - Has the user requested PEP 517 processing? None
+                     means the user hasn't explicitly specified.
+        pyproject_toml - Location of the project's pyproject.toml file
+        setup_py - Location of the project's setup.py file
+        req_name - The name of the requirement we're processing (for
+                   error reporting)
+
+    Returns:
+        None if we should use the legacy code path, otherwise a tuple
+        (
+            requirements from pyproject.toml,
+            name of PEP 517 backend,
+            requirements we should check are installed after setting
+                up the build environment
+            directory paths to import the backend from (backend-path),
+                relative to the project root.
+        )
+    """
+    has_pyproject = os.path.isfile(pyproject_toml)
+    has_setup = os.path.isfile(setup_py)
+
+    if not has_pyproject and not has_setup:
+        raise InstallationError(
+            f"{req_name} does not appear to be a Python project: "
+            f"neither 'setup.py' nor 'pyproject.toml' found."
+        )
+
+    if has_pyproject:
+        with open(pyproject_toml, encoding="utf-8") as f:
+            pp_toml = tomllib.loads(f.read())
+        build_system = pp_toml.get("build-system")
+    else:
+        build_system = None
+
+    # The following cases must use PEP 517
+    # We check for use_pep517 being non-None and falsy because that means
+    # the user explicitly requested --no-use-pep517.  The value 0 as
+    # opposed to False can occur when the value is provided via an
+    # environment variable or config file option (due to the quirk of
+    # strtobool() returning an integer in pip's configuration code).
+    if has_pyproject and not has_setup:
+        if use_pep517 is not None and not use_pep517:
+            raise InstallationError(
+                "Disabling PEP 517 processing is invalid: "
+                "project does not have a setup.py"
+            )
+        use_pep517 = True
+    elif build_system and "build-backend" in build_system:
+        if use_pep517 is not None and not use_pep517:
+            raise InstallationError(
+                "Disabling PEP 517 processing is invalid: "
+                "project specifies a build backend of {} "
+                "in pyproject.toml".format(build_system["build-backend"])
+            )
+        use_pep517 = True
+
+    # If we haven't worked out whether to use PEP 517 yet,
+    # and the user hasn't explicitly stated a preference,
+    # we do so if the project has a pyproject.toml file
+    # or if we cannot import setuptools or wheels.
+
+    # We fallback to PEP 517 when without setuptools or without the wheel package,
+    # so setuptools can be installed as a default build backend.
+    # For more info see:
+    # https://discuss.python.org/t/pip-without-setuptools-could-the-experience-be-improved/11810/9
+    # https://github.com/pypa/pip/issues/8559
+    elif use_pep517 is None:
+        use_pep517 = (
+            has_pyproject
+            or not importlib.util.find_spec("setuptools")
+            or not importlib.util.find_spec("wheel")
+        )
+
+    # At this point, we know whether we're going to use PEP 517.
+    assert use_pep517 is not None
+
+    # If we're using the legacy code path, there is nothing further
+    # for us to do here.
+    if not use_pep517:
+        return None
+
+    if build_system is None:
+        # Either the user has a pyproject.toml with no build-system
+        # section, or the user has no pyproject.toml, but has opted in
+        # explicitly via --use-pep517.
+        # In the absence of any explicit backend specification, we
+        # assume the setuptools backend that most closely emulates the
+        # traditional direct setup.py execution, and require wheel and
+        # a version of setuptools that supports that backend.
+
+        build_system = {
+            "requires": ["setuptools>=40.8.0"],
+            "build-backend": "setuptools.build_meta:__legacy__",
+        }
+
+    # If we're using PEP 517, we have build system information (either
+    # from pyproject.toml, or defaulted by the code above).
+    # Note that at this point, we do not know if the user has actually
+    # specified a backend, though.
+    assert build_system is not None
+
+    # Ensure that the build-system section in pyproject.toml conforms
+    # to PEP 518.
+
+    # Specifying the build-system table but not the requires key is invalid
+    if "requires" not in build_system:
+        raise MissingPyProjectBuildRequires(package=req_name)
+
+    # Error out if requires is not a list of strings
+    requires = build_system["requires"]
+    if not _is_list_of_str(requires):
+        raise InvalidPyProjectBuildRequires(
+            package=req_name,
+            reason="It is not a list of strings.",
+        )
+
+    # Each requirement must be valid as per PEP 508
+    for requirement in requires:
+        try:
+            get_requirement(requirement)
+        except InvalidRequirement as error:
+            raise InvalidPyProjectBuildRequires(
+                package=req_name,
+                reason=f"It contains an invalid requirement: {requirement!r}",
+            ) from error
+
+    backend = build_system.get("build-backend")
+    backend_path = build_system.get("backend-path", [])
+    check: List[str] = []
+    if backend is None:
+        # If the user didn't specify a backend, we assume they want to use
+        # the setuptools backend. But we can't be sure they have included
+        # a version of setuptools which supplies the backend. So we
+        # make a note to check that this requirement is present once
+        # we have set up the environment.
+        # This is quite a lot of work to check for a very specific case. But
+        # the problem is, that case is potentially quite common - projects that
+        # adopted PEP 518 early for the ability to specify requirements to
+        # execute setup.py, but never considered needing to mention the build
+        # tools themselves. The original PEP 518 code had a similar check (but
+        # implemented in a different way).
+        backend = "setuptools.build_meta:__legacy__"
+        check = ["setuptools>=40.8.0"]
+
+    return BuildSystemDetails(requires, backend, check, backend_path)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__init__.py
new file mode 100644
index 00000000..422d851d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__init__.py
@@ -0,0 +1,90 @@
+import collections
+import logging
+from dataclasses import dataclass
+from typing import Generator, List, Optional, Sequence, Tuple
+
+from pip._internal.utils.logging import indent_log
+
+from .req_file import parse_requirements
+from .req_install import InstallRequirement
+from .req_set import RequirementSet
+
+__all__ = [
+    "RequirementSet",
+    "InstallRequirement",
+    "parse_requirements",
+    "install_given_reqs",
+]
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class InstallationResult:
+    name: str
+
+
+def _validate_requirements(
+    requirements: List[InstallRequirement],
+) -> Generator[Tuple[str, InstallRequirement], None, None]:
+    for req in requirements:
+        assert req.name, f"invalid to-be-installed requirement: {req}"
+        yield req.name, req
+
+
+def install_given_reqs(
+    requirements: List[InstallRequirement],
+    global_options: Sequence[str],
+    root: Optional[str],
+    home: Optional[str],
+    prefix: Optional[str],
+    warn_script_location: bool,
+    use_user_site: bool,
+    pycompile: bool,
+) -> List[InstallationResult]:
+    """
+    Install everything in the given list.
+
+    (to be called after having downloaded and unpacked the packages)
+    """
+    to_install = collections.OrderedDict(_validate_requirements(requirements))
+
+    if to_install:
+        logger.info(
+            "Installing collected packages: %s",
+            ", ".join(to_install.keys()),
+        )
+
+    installed = []
+
+    with indent_log():
+        for req_name, requirement in to_install.items():
+            if requirement.should_reinstall:
+                logger.info("Attempting uninstall: %s", req_name)
+                with indent_log():
+                    uninstalled_pathset = requirement.uninstall(auto_confirm=True)
+            else:
+                uninstalled_pathset = None
+
+            try:
+                requirement.install(
+                    global_options,
+                    root=root,
+                    home=home,
+                    prefix=prefix,
+                    warn_script_location=warn_script_location,
+                    use_user_site=use_user_site,
+                    pycompile=pycompile,
+                )
+            except Exception:
+                # if install did not succeed, rollback previous uninstall
+                if uninstalled_pathset and not requirement.install_succeeded:
+                    uninstalled_pathset.rollback()
+                raise
+            else:
+                if uninstalled_pathset and requirement.install_succeeded:
+                    uninstalled_pathset.commit()
+
+            installed.append(InstallationResult(req_name))
+
+    return installed
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..edde2d4b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/constructors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/constructors.cpython-312.pyc
new file mode 100644
index 00000000..23a1f799
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/constructors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_file.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_file.cpython-312.pyc
new file mode 100644
index 00000000..3555669a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_file.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_install.cpython-312.pyc
new file mode 100644
index 00000000..9cfbbe66
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_set.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_set.cpython-312.pyc
new file mode 100644
index 00000000..db7e2435
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_set.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_uninstall.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_uninstall.cpython-312.pyc
new file mode 100644
index 00000000..c597d3ec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/__pycache__/req_uninstall.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/constructors.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/constructors.py
new file mode 100644
index 00000000..56a964f3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/constructors.py
@@ -0,0 +1,560 @@
+"""Backing implementation for InstallRequirement's various constructors
+
+The idea here is that these formed a major chunk of InstallRequirement's size
+so, moving them and support code dedicated to them outside of that class
+helps creates for better understandability for the rest of the code.
+
+These are meant to be used elsewhere within pip to create instances of
+InstallRequirement.
+"""
+
+import copy
+import logging
+import os
+import re
+from dataclasses import dataclass
+from typing import Collection, Dict, List, Optional, Set, Tuple, Union
+
+from pip._vendor.packaging.markers import Marker
+from pip._vendor.packaging.requirements import InvalidRequirement, Requirement
+from pip._vendor.packaging.specifiers import Specifier
+
+from pip._internal.exceptions import InstallationError
+from pip._internal.models.index import PyPI, TestPyPI
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.req.req_file import ParsedRequirement
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.utils.filetypes import is_archive_file
+from pip._internal.utils.misc import is_installable_dir
+from pip._internal.utils.packaging import get_requirement
+from pip._internal.utils.urls import path_to_url
+from pip._internal.vcs import is_url, vcs
+
+__all__ = [
+    "install_req_from_editable",
+    "install_req_from_line",
+    "parse_editable",
+]
+
+logger = logging.getLogger(__name__)
+operators = Specifier._operators.keys()
+
+
+def _strip_extras(path: str) -> Tuple[str, Optional[str]]:
+    m = re.match(r"^(.+)(\[[^\]]+\])$", path)
+    extras = None
+    if m:
+        path_no_extras = m.group(1)
+        extras = m.group(2)
+    else:
+        path_no_extras = path
+
+    return path_no_extras, extras
+
+
+def convert_extras(extras: Optional[str]) -> Set[str]:
+    if not extras:
+        return set()
+    return get_requirement("placeholder" + extras.lower()).extras
+
+
+def _set_requirement_extras(req: Requirement, new_extras: Set[str]) -> Requirement:
+    """
+    Returns a new requirement based on the given one, with the supplied extras. If the
+    given requirement already has extras those are replaced (or dropped if no new extras
+    are given).
+    """
+    match: Optional[re.Match[str]] = re.fullmatch(
+        # see https://peps.python.org/pep-0508/#complete-grammar
+        r"([\w\t .-]+)(\[[^\]]*\])?(.*)",
+        str(req),
+        flags=re.ASCII,
+    )
+    # ireq.req is a valid requirement so the regex should always match
+    assert (
+        match is not None
+    ), f"regex match on requirement {req} failed, this should never happen"
+    pre: Optional[str] = match.group(1)
+    post: Optional[str] = match.group(3)
+    assert (
+        pre is not None and post is not None
+    ), f"regex group selection for requirement {req} failed, this should never happen"
+    extras: str = "[{}]".format(",".join(sorted(new_extras)) if new_extras else "")
+    return get_requirement(f"{pre}{extras}{post}")
+
+
+def parse_editable(editable_req: str) -> Tuple[Optional[str], str, Set[str]]:
+    """Parses an editable requirement into:
+        - a requirement name
+        - an URL
+        - extras
+        - editable options
+    Accepted requirements:
+        svn+http://blahblah@rev#egg=Foobar[baz]&subdirectory=version_subdir
+        .[some_extra]
+    """
+
+    url = editable_req
+
+    # If a file path is specified with extras, strip off the extras.
+    url_no_extras, extras = _strip_extras(url)
+
+    if os.path.isdir(url_no_extras):
+        # Treating it as code that has already been checked out
+        url_no_extras = path_to_url(url_no_extras)
+
+    if url_no_extras.lower().startswith("file:"):
+        package_name = Link(url_no_extras).egg_fragment
+        if extras:
+            return (
+                package_name,
+                url_no_extras,
+                get_requirement("placeholder" + extras.lower()).extras,
+            )
+        else:
+            return package_name, url_no_extras, set()
+
+    for version_control in vcs:
+        if url.lower().startswith(f"{version_control}:"):
+            url = f"{version_control}+{url}"
+            break
+
+    link = Link(url)
+
+    if not link.is_vcs:
+        backends = ", ".join(vcs.all_schemes)
+        raise InstallationError(
+            f"{editable_req} is not a valid editable requirement. "
+            f"It should either be a path to a local project or a VCS URL "
+            f"(beginning with {backends})."
+        )
+
+    package_name = link.egg_fragment
+    if not package_name:
+        raise InstallationError(
+            f"Could not detect requirement name for '{editable_req}', "
+            "please specify one with #egg=your_package_name"
+        )
+    return package_name, url, set()
+
+
+def check_first_requirement_in_file(filename: str) -> None:
+    """Check if file is parsable as a requirements file.
+
+    This is heavily based on ``pkg_resources.parse_requirements``, but
+    simplified to just check the first meaningful line.
+
+    :raises InvalidRequirement: If the first meaningful line cannot be parsed
+        as an requirement.
+    """
+    with open(filename, encoding="utf-8", errors="ignore") as f:
+        # Create a steppable iterator, so we can handle \-continuations.
+        lines = (
+            line
+            for line in (line.strip() for line in f)
+            if line and not line.startswith("#")  # Skip blank lines/comments.
+        )
+
+        for line in lines:
+            # Drop comments -- a hash without a space may be in a URL.
+            if " #" in line:
+                line = line[: line.find(" #")]
+            # If there is a line continuation, drop it, and append the next line.
+            if line.endswith("\\"):
+                line = line[:-2].strip() + next(lines, "")
+            get_requirement(line)
+            return
+
+
+def deduce_helpful_msg(req: str) -> str:
+    """Returns helpful msg in case requirements file does not exist,
+    or cannot be parsed.
+
+    :params req: Requirements file path
+    """
+    if not os.path.exists(req):
+        return f" File '{req}' does not exist."
+    msg = " The path does exist. "
+    # Try to parse and check if it is a requirements file.
+    try:
+        check_first_requirement_in_file(req)
+    except InvalidRequirement:
+        logger.debug("Cannot parse '%s' as requirements file", req)
+    else:
+        msg += (
+            f"The argument you provided "
+            f"({req}) appears to be a"
+            f" requirements file. If that is the"
+            f" case, use the '-r' flag to install"
+            f" the packages specified within it."
+        )
+    return msg
+
+
+@dataclass(frozen=True)
+class RequirementParts:
+    requirement: Optional[Requirement]
+    link: Optional[Link]
+    markers: Optional[Marker]
+    extras: Set[str]
+
+
+def parse_req_from_editable(editable_req: str) -> RequirementParts:
+    name, url, extras_override = parse_editable(editable_req)
+
+    if name is not None:
+        try:
+            req: Optional[Requirement] = get_requirement(name)
+        except InvalidRequirement as exc:
+            raise InstallationError(f"Invalid requirement: {name!r}: {exc}")
+    else:
+        req = None
+
+    link = Link(url)
+
+    return RequirementParts(req, link, None, extras_override)
+
+
+# ---- The actual constructors follow ----
+
+
+def install_req_from_editable(
+    editable_req: str,
+    comes_from: Optional[Union[InstallRequirement, str]] = None,
+    *,
+    use_pep517: Optional[bool] = None,
+    isolated: bool = False,
+    global_options: Optional[List[str]] = None,
+    hash_options: Optional[Dict[str, List[str]]] = None,
+    constraint: bool = False,
+    user_supplied: bool = False,
+    permit_editable_wheels: bool = False,
+    config_settings: Optional[Dict[str, Union[str, List[str]]]] = None,
+) -> InstallRequirement:
+    parts = parse_req_from_editable(editable_req)
+
+    return InstallRequirement(
+        parts.requirement,
+        comes_from=comes_from,
+        user_supplied=user_supplied,
+        editable=True,
+        permit_editable_wheels=permit_editable_wheels,
+        link=parts.link,
+        constraint=constraint,
+        use_pep517=use_pep517,
+        isolated=isolated,
+        global_options=global_options,
+        hash_options=hash_options,
+        config_settings=config_settings,
+        extras=parts.extras,
+    )
+
+
+def _looks_like_path(name: str) -> bool:
+    """Checks whether the string "looks like" a path on the filesystem.
+
+    This does not check whether the target actually exists, only judge from the
+    appearance.
+
+    Returns true if any of the following conditions is true:
+    * a path separator is found (either os.path.sep or os.path.altsep);
+    * a dot is found (which represents the current directory).
+    """
+    if os.path.sep in name:
+        return True
+    if os.path.altsep is not None and os.path.altsep in name:
+        return True
+    if name.startswith("."):
+        return True
+    return False
+
+
+def _get_url_from_path(path: str, name: str) -> Optional[str]:
+    """
+    First, it checks whether a provided path is an installable directory. If it
+    is, returns the path.
+
+    If false, check if the path is an archive file (such as a .whl).
+    The function checks if the path is a file. If false, if the path has
+    an @, it will treat it as a PEP 440 URL requirement and return the path.
+    """
+    if _looks_like_path(name) and os.path.isdir(path):
+        if is_installable_dir(path):
+            return path_to_url(path)
+        # TODO: The is_installable_dir test here might not be necessary
+        #       now that it is done in load_pyproject_toml too.
+        raise InstallationError(
+            f"Directory {name!r} is not installable. Neither 'setup.py' "
+            "nor 'pyproject.toml' found."
+        )
+    if not is_archive_file(path):
+        return None
+    if os.path.isfile(path):
+        return path_to_url(path)
+    urlreq_parts = name.split("@", 1)
+    if len(urlreq_parts) >= 2 and not _looks_like_path(urlreq_parts[0]):
+        # If the path contains '@' and the part before it does not look
+        # like a path, try to treat it as a PEP 440 URL req instead.
+        return None
+    logger.warning(
+        "Requirement %r looks like a filename, but the file does not exist",
+        name,
+    )
+    return path_to_url(path)
+
+
+def parse_req_from_line(name: str, line_source: Optional[str]) -> RequirementParts:
+    if is_url(name):
+        marker_sep = "; "
+    else:
+        marker_sep = ";"
+    if marker_sep in name:
+        name, markers_as_string = name.split(marker_sep, 1)
+        markers_as_string = markers_as_string.strip()
+        if not markers_as_string:
+            markers = None
+        else:
+            markers = Marker(markers_as_string)
+    else:
+        markers = None
+    name = name.strip()
+    req_as_string = None
+    path = os.path.normpath(os.path.abspath(name))
+    link = None
+    extras_as_string = None
+
+    if is_url(name):
+        link = Link(name)
+    else:
+        p, extras_as_string = _strip_extras(path)
+        url = _get_url_from_path(p, name)
+        if url is not None:
+            link = Link(url)
+
+    # it's a local file, dir, or url
+    if link:
+        # Handle relative file URLs
+        if link.scheme == "file" and re.search(r"\.\./", link.url):
+            link = Link(path_to_url(os.path.normpath(os.path.abspath(link.path))))
+        # wheel file
+        if link.is_wheel:
+            wheel = Wheel(link.filename)  # can raise InvalidWheelFilename
+            req_as_string = f"{wheel.name}=={wheel.version}"
+        else:
+            # set the req to the egg fragment.  when it's not there, this
+            # will become an 'unnamed' requirement
+            req_as_string = link.egg_fragment
+
+    # a requirement specifier
+    else:
+        req_as_string = name
+
+    extras = convert_extras(extras_as_string)
+
+    def with_source(text: str) -> str:
+        if not line_source:
+            return text
+        return f"{text} (from {line_source})"
+
+    def _parse_req_string(req_as_string: str) -> Requirement:
+        try:
+            return get_requirement(req_as_string)
+        except InvalidRequirement as exc:
+            if os.path.sep in req_as_string:
+                add_msg = "It looks like a path."
+                add_msg += deduce_helpful_msg(req_as_string)
+            elif "=" in req_as_string and not any(
+                op in req_as_string for op in operators
+            ):
+                add_msg = "= is not a valid operator. Did you mean == ?"
+            else:
+                add_msg = ""
+            msg = with_source(f"Invalid requirement: {req_as_string!r}: {exc}")
+            if add_msg:
+                msg += f"\nHint: {add_msg}"
+            raise InstallationError(msg)
+
+    if req_as_string is not None:
+        req: Optional[Requirement] = _parse_req_string(req_as_string)
+    else:
+        req = None
+
+    return RequirementParts(req, link, markers, extras)
+
+
+def install_req_from_line(
+    name: str,
+    comes_from: Optional[Union[str, InstallRequirement]] = None,
+    *,
+    use_pep517: Optional[bool] = None,
+    isolated: bool = False,
+    global_options: Optional[List[str]] = None,
+    hash_options: Optional[Dict[str, List[str]]] = None,
+    constraint: bool = False,
+    line_source: Optional[str] = None,
+    user_supplied: bool = False,
+    config_settings: Optional[Dict[str, Union[str, List[str]]]] = None,
+) -> InstallRequirement:
+    """Creates an InstallRequirement from a name, which might be a
+    requirement, directory containing 'setup.py', filename, or URL.
+
+    :param line_source: An optional string describing where the line is from,
+        for logging purposes in case of an error.
+    """
+    parts = parse_req_from_line(name, line_source)
+
+    return InstallRequirement(
+        parts.requirement,
+        comes_from,
+        link=parts.link,
+        markers=parts.markers,
+        use_pep517=use_pep517,
+        isolated=isolated,
+        global_options=global_options,
+        hash_options=hash_options,
+        config_settings=config_settings,
+        constraint=constraint,
+        extras=parts.extras,
+        user_supplied=user_supplied,
+    )
+
+
+def install_req_from_req_string(
+    req_string: str,
+    comes_from: Optional[InstallRequirement] = None,
+    isolated: bool = False,
+    use_pep517: Optional[bool] = None,
+    user_supplied: bool = False,
+) -> InstallRequirement:
+    try:
+        req = get_requirement(req_string)
+    except InvalidRequirement as exc:
+        raise InstallationError(f"Invalid requirement: {req_string!r}: {exc}")
+
+    domains_not_allowed = [
+        PyPI.file_storage_domain,
+        TestPyPI.file_storage_domain,
+    ]
+    if (
+        req.url
+        and comes_from
+        and comes_from.link
+        and comes_from.link.netloc in domains_not_allowed
+    ):
+        # Explicitly disallow pypi packages that depend on external urls
+        raise InstallationError(
+            "Packages installed from PyPI cannot depend on packages "
+            "which are not also hosted on PyPI.\n"
+            f"{comes_from.name} depends on {req} "
+        )
+
+    return InstallRequirement(
+        req,
+        comes_from,
+        isolated=isolated,
+        use_pep517=use_pep517,
+        user_supplied=user_supplied,
+    )
+
+
+def install_req_from_parsed_requirement(
+    parsed_req: ParsedRequirement,
+    isolated: bool = False,
+    use_pep517: Optional[bool] = None,
+    user_supplied: bool = False,
+    config_settings: Optional[Dict[str, Union[str, List[str]]]] = None,
+) -> InstallRequirement:
+    if parsed_req.is_editable:
+        req = install_req_from_editable(
+            parsed_req.requirement,
+            comes_from=parsed_req.comes_from,
+            use_pep517=use_pep517,
+            constraint=parsed_req.constraint,
+            isolated=isolated,
+            user_supplied=user_supplied,
+            config_settings=config_settings,
+        )
+
+    else:
+        req = install_req_from_line(
+            parsed_req.requirement,
+            comes_from=parsed_req.comes_from,
+            use_pep517=use_pep517,
+            isolated=isolated,
+            global_options=(
+                parsed_req.options.get("global_options", [])
+                if parsed_req.options
+                else []
+            ),
+            hash_options=(
+                parsed_req.options.get("hashes", {}) if parsed_req.options else {}
+            ),
+            constraint=parsed_req.constraint,
+            line_source=parsed_req.line_source,
+            user_supplied=user_supplied,
+            config_settings=config_settings,
+        )
+    return req
+
+
+def install_req_from_link_and_ireq(
+    link: Link, ireq: InstallRequirement
+) -> InstallRequirement:
+    return InstallRequirement(
+        req=ireq.req,
+        comes_from=ireq.comes_from,
+        editable=ireq.editable,
+        link=link,
+        markers=ireq.markers,
+        use_pep517=ireq.use_pep517,
+        isolated=ireq.isolated,
+        global_options=ireq.global_options,
+        hash_options=ireq.hash_options,
+        config_settings=ireq.config_settings,
+        user_supplied=ireq.user_supplied,
+    )
+
+
+def install_req_drop_extras(ireq: InstallRequirement) -> InstallRequirement:
+    """
+    Creates a new InstallationRequirement using the given template but without
+    any extras. Sets the original requirement as the new one's parent
+    (comes_from).
+    """
+    return InstallRequirement(
+        req=(
+            _set_requirement_extras(ireq.req, set()) if ireq.req is not None else None
+        ),
+        comes_from=ireq,
+        editable=ireq.editable,
+        link=ireq.link,
+        markers=ireq.markers,
+        use_pep517=ireq.use_pep517,
+        isolated=ireq.isolated,
+        global_options=ireq.global_options,
+        hash_options=ireq.hash_options,
+        constraint=ireq.constraint,
+        extras=[],
+        config_settings=ireq.config_settings,
+        user_supplied=ireq.user_supplied,
+        permit_editable_wheels=ireq.permit_editable_wheels,
+    )
+
+
+def install_req_extend_extras(
+    ireq: InstallRequirement,
+    extras: Collection[str],
+) -> InstallRequirement:
+    """
+    Returns a copy of an installation requirement with some additional extras.
+    Makes a shallow copy of the ireq object.
+    """
+    result = copy.copy(ireq)
+    result.extras = {*ireq.extras, *extras}
+    result.req = (
+        _set_requirement_extras(ireq.req, result.extras)
+        if ireq.req is not None
+        else None
+    )
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_file.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_file.py
new file mode 100644
index 00000000..f6ba70fe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_file.py
@@ -0,0 +1,623 @@
+"""
+Requirements file parsing
+"""
+
+import codecs
+import locale
+import logging
+import optparse
+import os
+import re
+import shlex
+import sys
+import urllib.parse
+from dataclasses import dataclass
+from optparse import Values
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    Dict,
+    Generator,
+    Iterable,
+    List,
+    NoReturn,
+    Optional,
+    Tuple,
+)
+
+from pip._internal.cli import cmdoptions
+from pip._internal.exceptions import InstallationError, RequirementsFileParseError
+from pip._internal.models.search_scope import SearchScope
+
+if TYPE_CHECKING:
+    from pip._internal.index.package_finder import PackageFinder
+    from pip._internal.network.session import PipSession
+
+__all__ = ["parse_requirements"]
+
+ReqFileLines = Iterable[Tuple[int, str]]
+
+LineParser = Callable[[str], Tuple[str, Values]]
+
+SCHEME_RE = re.compile(r"^(http|https|file):", re.I)
+COMMENT_RE = re.compile(r"(^|\s+)#.*$")
+
+# Matches environment variable-style values in '${MY_VARIABLE_1}' with the
+# variable name consisting of only uppercase letters, digits or the '_'
+# (underscore). This follows the POSIX standard defined in IEEE Std 1003.1,
+# 2013 Edition.
+ENV_VAR_RE = re.compile(r"(?P<var>\$\{(?P<name>[A-Z0-9_]+)\})")
+
+SUPPORTED_OPTIONS: List[Callable[..., optparse.Option]] = [
+    cmdoptions.index_url,
+    cmdoptions.extra_index_url,
+    cmdoptions.no_index,
+    cmdoptions.constraints,
+    cmdoptions.requirements,
+    cmdoptions.editable,
+    cmdoptions.find_links,
+    cmdoptions.no_binary,
+    cmdoptions.only_binary,
+    cmdoptions.prefer_binary,
+    cmdoptions.require_hashes,
+    cmdoptions.pre,
+    cmdoptions.trusted_host,
+    cmdoptions.use_new_feature,
+]
+
+# options to be passed to requirements
+SUPPORTED_OPTIONS_REQ: List[Callable[..., optparse.Option]] = [
+    cmdoptions.global_options,
+    cmdoptions.hash,
+    cmdoptions.config_settings,
+]
+
+SUPPORTED_OPTIONS_EDITABLE_REQ: List[Callable[..., optparse.Option]] = [
+    cmdoptions.config_settings,
+]
+
+
+# the 'dest' string values
+SUPPORTED_OPTIONS_REQ_DEST = [str(o().dest) for o in SUPPORTED_OPTIONS_REQ]
+SUPPORTED_OPTIONS_EDITABLE_REQ_DEST = [
+    str(o().dest) for o in SUPPORTED_OPTIONS_EDITABLE_REQ
+]
+
+# order of BOMS is important: codecs.BOM_UTF16_LE is a prefix of codecs.BOM_UTF32_LE
+# so data.startswith(BOM_UTF16_LE) would be true for UTF32_LE data
+BOMS: List[Tuple[bytes, str]] = [
+    (codecs.BOM_UTF8, "utf-8"),
+    (codecs.BOM_UTF32, "utf-32"),
+    (codecs.BOM_UTF32_BE, "utf-32-be"),
+    (codecs.BOM_UTF32_LE, "utf-32-le"),
+    (codecs.BOM_UTF16, "utf-16"),
+    (codecs.BOM_UTF16_BE, "utf-16-be"),
+    (codecs.BOM_UTF16_LE, "utf-16-le"),
+]
+
+PEP263_ENCODING_RE = re.compile(rb"coding[:=]\s*([-\w.]+)")
+DEFAULT_ENCODING = "utf-8"
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class ParsedRequirement:
+    # TODO: replace this with slots=True when dropping Python 3.9 support.
+    __slots__ = (
+        "requirement",
+        "is_editable",
+        "comes_from",
+        "constraint",
+        "options",
+        "line_source",
+    )
+
+    requirement: str
+    is_editable: bool
+    comes_from: str
+    constraint: bool
+    options: Optional[Dict[str, Any]]
+    line_source: Optional[str]
+
+
+@dataclass(frozen=True)
+class ParsedLine:
+    __slots__ = ("filename", "lineno", "args", "opts", "constraint")
+
+    filename: str
+    lineno: int
+    args: str
+    opts: Values
+    constraint: bool
+
+    @property
+    def is_editable(self) -> bool:
+        return bool(self.opts.editables)
+
+    @property
+    def requirement(self) -> Optional[str]:
+        if self.args:
+            return self.args
+        elif self.is_editable:
+            # We don't support multiple -e on one line
+            return self.opts.editables[0]
+        return None
+
+
+def parse_requirements(
+    filename: str,
+    session: "PipSession",
+    finder: Optional["PackageFinder"] = None,
+    options: Optional[optparse.Values] = None,
+    constraint: bool = False,
+) -> Generator[ParsedRequirement, None, None]:
+    """Parse a requirements file and yield ParsedRequirement instances.
+
+    :param filename:    Path or url of requirements file.
+    :param session:     PipSession instance.
+    :param finder:      Instance of pip.index.PackageFinder.
+    :param options:     cli options.
+    :param constraint:  If true, parsing a constraint file rather than
+        requirements file.
+    """
+    line_parser = get_line_parser(finder)
+    parser = RequirementsFileParser(session, line_parser)
+
+    for parsed_line in parser.parse(filename, constraint):
+        parsed_req = handle_line(
+            parsed_line, options=options, finder=finder, session=session
+        )
+        if parsed_req is not None:
+            yield parsed_req
+
+
+def preprocess(content: str) -> ReqFileLines:
+    """Split, filter, and join lines, and return a line iterator
+
+    :param content: the content of the requirements file
+    """
+    lines_enum: ReqFileLines = enumerate(content.splitlines(), start=1)
+    lines_enum = join_lines(lines_enum)
+    lines_enum = ignore_comments(lines_enum)
+    lines_enum = expand_env_variables(lines_enum)
+    return lines_enum
+
+
+def handle_requirement_line(
+    line: ParsedLine,
+    options: Optional[optparse.Values] = None,
+) -> ParsedRequirement:
+    # preserve for the nested code path
+    line_comes_from = "{} {} (line {})".format(
+        "-c" if line.constraint else "-r",
+        line.filename,
+        line.lineno,
+    )
+
+    assert line.requirement is not None
+
+    # get the options that apply to requirements
+    if line.is_editable:
+        supported_dest = SUPPORTED_OPTIONS_EDITABLE_REQ_DEST
+    else:
+        supported_dest = SUPPORTED_OPTIONS_REQ_DEST
+    req_options = {}
+    for dest in supported_dest:
+        if dest in line.opts.__dict__ and line.opts.__dict__[dest]:
+            req_options[dest] = line.opts.__dict__[dest]
+
+    line_source = f"line {line.lineno} of {line.filename}"
+    return ParsedRequirement(
+        requirement=line.requirement,
+        is_editable=line.is_editable,
+        comes_from=line_comes_from,
+        constraint=line.constraint,
+        options=req_options,
+        line_source=line_source,
+    )
+
+
+def handle_option_line(
+    opts: Values,
+    filename: str,
+    lineno: int,
+    finder: Optional["PackageFinder"] = None,
+    options: Optional[optparse.Values] = None,
+    session: Optional["PipSession"] = None,
+) -> None:
+    if opts.hashes:
+        logger.warning(
+            "%s line %s has --hash but no requirement, and will be ignored.",
+            filename,
+            lineno,
+        )
+
+    if options:
+        # percolate options upward
+        if opts.require_hashes:
+            options.require_hashes = opts.require_hashes
+        if opts.features_enabled:
+            options.features_enabled.extend(
+                f for f in opts.features_enabled if f not in options.features_enabled
+            )
+
+    # set finder options
+    if finder:
+        find_links = finder.find_links
+        index_urls = finder.index_urls
+        no_index = finder.search_scope.no_index
+        if opts.no_index is True:
+            no_index = True
+            index_urls = []
+        if opts.index_url and not no_index:
+            index_urls = [opts.index_url]
+        if opts.extra_index_urls and not no_index:
+            index_urls.extend(opts.extra_index_urls)
+        if opts.find_links:
+            # FIXME: it would be nice to keep track of the source
+            # of the find_links: support a find-links local path
+            # relative to a requirements file.
+            value = opts.find_links[0]
+            req_dir = os.path.dirname(os.path.abspath(filename))
+            relative_to_reqs_file = os.path.join(req_dir, value)
+            if os.path.exists(relative_to_reqs_file):
+                value = relative_to_reqs_file
+            find_links.append(value)
+
+        if session:
+            # We need to update the auth urls in session
+            session.update_index_urls(index_urls)
+
+        search_scope = SearchScope(
+            find_links=find_links,
+            index_urls=index_urls,
+            no_index=no_index,
+        )
+        finder.search_scope = search_scope
+
+        if opts.pre:
+            finder.set_allow_all_prereleases()
+
+        if opts.prefer_binary:
+            finder.set_prefer_binary()
+
+        if session:
+            for host in opts.trusted_hosts or []:
+                source = f"line {lineno} of {filename}"
+                session.add_trusted_host(host, source=source)
+
+
+def handle_line(
+    line: ParsedLine,
+    options: Optional[optparse.Values] = None,
+    finder: Optional["PackageFinder"] = None,
+    session: Optional["PipSession"] = None,
+) -> Optional[ParsedRequirement]:
+    """Handle a single parsed requirements line; This can result in
+    creating/yielding requirements, or updating the finder.
+
+    :param line:        The parsed line to be processed.
+    :param options:     CLI options.
+    :param finder:      The finder - updated by non-requirement lines.
+    :param session:     The session - updated by non-requirement lines.
+
+    Returns a ParsedRequirement object if the line is a requirement line,
+    otherwise returns None.
+
+    For lines that contain requirements, the only options that have an effect
+    are from SUPPORTED_OPTIONS_REQ, and they are scoped to the
+    requirement. Other options from SUPPORTED_OPTIONS may be present, but are
+    ignored.
+
+    For lines that do not contain requirements, the only options that have an
+    effect are from SUPPORTED_OPTIONS. Options from SUPPORTED_OPTIONS_REQ may
+    be present, but are ignored. These lines may contain multiple options
+    (although our docs imply only one is supported), and all our parsed and
+    affect the finder.
+    """
+
+    if line.requirement is not None:
+        parsed_req = handle_requirement_line(line, options)
+        return parsed_req
+    else:
+        handle_option_line(
+            line.opts,
+            line.filename,
+            line.lineno,
+            finder,
+            options,
+            session,
+        )
+        return None
+
+
+class RequirementsFileParser:
+    def __init__(
+        self,
+        session: "PipSession",
+        line_parser: LineParser,
+    ) -> None:
+        self._session = session
+        self._line_parser = line_parser
+
+    def parse(
+        self, filename: str, constraint: bool
+    ) -> Generator[ParsedLine, None, None]:
+        """Parse a given file, yielding parsed lines."""
+        yield from self._parse_and_recurse(
+            filename, constraint, [{os.path.abspath(filename): None}]
+        )
+
+    def _parse_and_recurse(
+        self,
+        filename: str,
+        constraint: bool,
+        parsed_files_stack: List[Dict[str, Optional[str]]],
+    ) -> Generator[ParsedLine, None, None]:
+        for line in self._parse_file(filename, constraint):
+            if line.requirement is None and (
+                line.opts.requirements or line.opts.constraints
+            ):
+                # parse a nested requirements file
+                if line.opts.requirements:
+                    req_path = line.opts.requirements[0]
+                    nested_constraint = False
+                else:
+                    req_path = line.opts.constraints[0]
+                    nested_constraint = True
+
+                # original file is over http
+                if SCHEME_RE.search(filename):
+                    # do a url join so relative paths work
+                    req_path = urllib.parse.urljoin(filename, req_path)
+                # original file and nested file are paths
+                elif not SCHEME_RE.search(req_path):
+                    # do a join so relative paths work
+                    # and then abspath so that we can identify recursive references
+                    req_path = os.path.abspath(
+                        os.path.join(
+                            os.path.dirname(filename),
+                            req_path,
+                        )
+                    )
+                parsed_files = parsed_files_stack[0]
+                if req_path in parsed_files:
+                    initial_file = parsed_files[req_path]
+                    tail = (
+                        f" and again in {initial_file}"
+                        if initial_file is not None
+                        else ""
+                    )
+                    raise RequirementsFileParseError(
+                        f"{req_path} recursively references itself in {filename}{tail}"
+                    )
+                # Keeping a track where was each file first included in
+                new_parsed_files = parsed_files.copy()
+                new_parsed_files[req_path] = filename
+                yield from self._parse_and_recurse(
+                    req_path, nested_constraint, [new_parsed_files, *parsed_files_stack]
+                )
+            else:
+                yield line
+
+    def _parse_file(
+        self, filename: str, constraint: bool
+    ) -> Generator[ParsedLine, None, None]:
+        _, content = get_file_content(filename, self._session)
+
+        lines_enum = preprocess(content)
+
+        for line_number, line in lines_enum:
+            try:
+                args_str, opts = self._line_parser(line)
+            except OptionParsingError as e:
+                # add offending line
+                msg = f"Invalid requirement: {line}\n{e.msg}"
+                raise RequirementsFileParseError(msg)
+
+            yield ParsedLine(
+                filename,
+                line_number,
+                args_str,
+                opts,
+                constraint,
+            )
+
+
+def get_line_parser(finder: Optional["PackageFinder"]) -> LineParser:
+    def parse_line(line: str) -> Tuple[str, Values]:
+        # Build new parser for each line since it accumulates appendable
+        # options.
+        parser = build_parser()
+        defaults = parser.get_default_values()
+        defaults.index_url = None
+        if finder:
+            defaults.format_control = finder.format_control
+
+        args_str, options_str = break_args_options(line)
+
+        try:
+            options = shlex.split(options_str)
+        except ValueError as e:
+            raise OptionParsingError(f"Could not split options: {options_str}") from e
+
+        opts, _ = parser.parse_args(options, defaults)
+
+        return args_str, opts
+
+    return parse_line
+
+
+def break_args_options(line: str) -> Tuple[str, str]:
+    """Break up the line into an args and options string.  We only want to shlex
+    (and then optparse) the options, not the args.  args can contain markers
+    which are corrupted by shlex.
+    """
+    tokens = line.split(" ")
+    args = []
+    options = tokens[:]
+    for token in tokens:
+        if token.startswith("-") or token.startswith("--"):
+            break
+        else:
+            args.append(token)
+            options.pop(0)
+    return " ".join(args), " ".join(options)
+
+
+class OptionParsingError(Exception):
+    def __init__(self, msg: str) -> None:
+        self.msg = msg
+
+
+def build_parser() -> optparse.OptionParser:
+    """
+    Return a parser for parsing requirement lines
+    """
+    parser = optparse.OptionParser(add_help_option=False)
+
+    option_factories = SUPPORTED_OPTIONS + SUPPORTED_OPTIONS_REQ
+    for option_factory in option_factories:
+        option = option_factory()
+        parser.add_option(option)
+
+    # By default optparse sys.exits on parsing errors. We want to wrap
+    # that in our own exception.
+    def parser_exit(self: Any, msg: str) -> "NoReturn":
+        raise OptionParsingError(msg)
+
+    # NOTE: mypy disallows assigning to a method
+    #       https://github.com/python/mypy/issues/2427
+    parser.exit = parser_exit  # type: ignore
+
+    return parser
+
+
+def join_lines(lines_enum: ReqFileLines) -> ReqFileLines:
+    """Joins a line ending in '\' with the previous line (except when following
+    comments).  The joined line takes on the index of the first line.
+    """
+    primary_line_number = None
+    new_line: List[str] = []
+    for line_number, line in lines_enum:
+        if not line.endswith("\\") or COMMENT_RE.match(line):
+            if COMMENT_RE.match(line):
+                # this ensures comments are always matched later
+                line = " " + line
+            if new_line:
+                new_line.append(line)
+                assert primary_line_number is not None
+                yield primary_line_number, "".join(new_line)
+                new_line = []
+            else:
+                yield line_number, line
+        else:
+            if not new_line:
+                primary_line_number = line_number
+            new_line.append(line.strip("\\"))
+
+    # last line contains \
+    if new_line:
+        assert primary_line_number is not None
+        yield primary_line_number, "".join(new_line)
+
+    # TODO: handle space after '\'.
+
+
+def ignore_comments(lines_enum: ReqFileLines) -> ReqFileLines:
+    """
+    Strips comments and filter empty lines.
+    """
+    for line_number, line in lines_enum:
+        line = COMMENT_RE.sub("", line)
+        line = line.strip()
+        if line:
+            yield line_number, line
+
+
+def expand_env_variables(lines_enum: ReqFileLines) -> ReqFileLines:
+    """Replace all environment variables that can be retrieved via `os.getenv`.
+
+    The only allowed format for environment variables defined in the
+    requirement file is `${MY_VARIABLE_1}` to ensure two things:
+
+    1. Strings that contain a `$` aren't accidentally (partially) expanded.
+    2. Ensure consistency across platforms for requirement files.
+
+    These points are the result of a discussion on the `github pull
+    request #3514 <https://github.com/pypa/pip/pull/3514>`_.
+
+    Valid characters in variable names follow the `POSIX standard
+    <http://pubs.opengroup.org/onlinepubs/9699919799/>`_ and are limited
+    to uppercase letter, digits and the `_` (underscore).
+    """
+    for line_number, line in lines_enum:
+        for env_var, var_name in ENV_VAR_RE.findall(line):
+            value = os.getenv(var_name)
+            if not value:
+                continue
+
+            line = line.replace(env_var, value)
+
+        yield line_number, line
+
+
+def get_file_content(url: str, session: "PipSession") -> Tuple[str, str]:
+    """Gets the content of a file; it may be a filename, file: URL, or
+    http: URL.  Returns (location, content).  Content is unicode.
+    Respects # -*- coding: declarations on the retrieved files.
+
+    :param url:         File path or url.
+    :param session:     PipSession instance.
+    """
+    scheme = urllib.parse.urlsplit(url).scheme
+    # Pip has special support for file:// URLs (LocalFSAdapter).
+    if scheme in ["http", "https", "file"]:
+        # Delay importing heavy network modules until absolutely necessary.
+        from pip._internal.network.utils import raise_for_status
+
+        resp = session.get(url)
+        raise_for_status(resp)
+        return resp.url, resp.text
+
+    # Assume this is a bare path.
+    try:
+        with open(url, "rb") as f:
+            raw_content = f.read()
+    except OSError as exc:
+        raise InstallationError(f"Could not open requirements file: {exc}")
+
+    content = _decode_req_file(raw_content, url)
+
+    return url, content
+
+
+def _decode_req_file(data: bytes, url: str) -> str:
+    for bom, encoding in BOMS:
+        if data.startswith(bom):
+            return data[len(bom) :].decode(encoding)
+
+    for line in data.split(b"\n")[:2]:
+        if line[0:1] == b"#":
+            result = PEP263_ENCODING_RE.search(line)
+            if result is not None:
+                encoding = result.groups()[0].decode("ascii")
+                return data.decode(encoding)
+
+    try:
+        return data.decode(DEFAULT_ENCODING)
+    except UnicodeDecodeError:
+        locale_encoding = locale.getpreferredencoding(False) or sys.getdefaultencoding()
+        logging.warning(
+            "unable to decode data from %s with default encoding %s, "
+            "falling back to encoding from locale: %s. "
+            "If this is intentional you should specify the encoding with a "
+            "PEP-263 style comment, e.g. '# -*- coding: %s -*-'",
+            url,
+            DEFAULT_ENCODING,
+            locale_encoding,
+            locale_encoding,
+        )
+        return data.decode(locale_encoding)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_install.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_install.py
new file mode 100644
index 00000000..3262d826
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_install.py
@@ -0,0 +1,934 @@
+import functools
+import logging
+import os
+import shutil
+import sys
+import uuid
+import zipfile
+from optparse import Values
+from pathlib import Path
+from typing import Any, Collection, Dict, Iterable, List, Optional, Sequence, Union
+
+from pip._vendor.packaging.markers import Marker
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.packaging.specifiers import SpecifierSet
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.packaging.version import Version
+from pip._vendor.packaging.version import parse as parse_version
+from pip._vendor.pyproject_hooks import BuildBackendHookCaller
+
+from pip._internal.build_env import BuildEnvironment, NoOpBuildEnvironment
+from pip._internal.exceptions import InstallationError, PreviousBuildDirError
+from pip._internal.locations import get_scheme
+from pip._internal.metadata import (
+    BaseDistribution,
+    get_default_environment,
+    get_directory_distribution,
+    get_wheel_distribution,
+)
+from pip._internal.metadata.base import FilesystemWheel
+from pip._internal.models.direct_url import DirectUrl
+from pip._internal.models.link import Link
+from pip._internal.operations.build.metadata import generate_metadata
+from pip._internal.operations.build.metadata_editable import generate_editable_metadata
+from pip._internal.operations.build.metadata_legacy import (
+    generate_metadata as generate_metadata_legacy,
+)
+from pip._internal.operations.install.editable_legacy import (
+    install_editable as install_editable_legacy,
+)
+from pip._internal.operations.install.wheel import install_wheel
+from pip._internal.pyproject import load_pyproject_toml, make_pyproject_path
+from pip._internal.req.req_uninstall import UninstallPathSet
+from pip._internal.utils.deprecation import deprecated
+from pip._internal.utils.hashes import Hashes
+from pip._internal.utils.misc import (
+    ConfiguredBuildBackendHookCaller,
+    ask_path_exists,
+    backup_dir,
+    display_path,
+    hide_url,
+    is_installable_dir,
+    redact_auth_from_requirement,
+    redact_auth_from_url,
+)
+from pip._internal.utils.packaging import get_requirement
+from pip._internal.utils.subprocess import runner_with_spinner_message
+from pip._internal.utils.temp_dir import TempDirectory, tempdir_kinds
+from pip._internal.utils.unpacking import unpack_file
+from pip._internal.utils.virtualenv import running_under_virtualenv
+from pip._internal.vcs import vcs
+
+logger = logging.getLogger(__name__)
+
+
+class InstallRequirement:
+    """
+    Represents something that may be installed later on, may have information
+    about where to fetch the relevant requirement and also contains logic for
+    installing the said requirement.
+    """
+
+    def __init__(
+        self,
+        req: Optional[Requirement],
+        comes_from: Optional[Union[str, "InstallRequirement"]],
+        editable: bool = False,
+        link: Optional[Link] = None,
+        markers: Optional[Marker] = None,
+        use_pep517: Optional[bool] = None,
+        isolated: bool = False,
+        *,
+        global_options: Optional[List[str]] = None,
+        hash_options: Optional[Dict[str, List[str]]] = None,
+        config_settings: Optional[Dict[str, Union[str, List[str]]]] = None,
+        constraint: bool = False,
+        extras: Collection[str] = (),
+        user_supplied: bool = False,
+        permit_editable_wheels: bool = False,
+    ) -> None:
+        assert req is None or isinstance(req, Requirement), req
+        self.req = req
+        self.comes_from = comes_from
+        self.constraint = constraint
+        self.editable = editable
+        self.permit_editable_wheels = permit_editable_wheels
+
+        # source_dir is the local directory where the linked requirement is
+        # located, or unpacked. In case unpacking is needed, creating and
+        # populating source_dir is done by the RequirementPreparer. Note this
+        # is not necessarily the directory where pyproject.toml or setup.py is
+        # located - that one is obtained via unpacked_source_directory.
+        self.source_dir: Optional[str] = None
+        if self.editable:
+            assert link
+            if link.is_file:
+                self.source_dir = os.path.normpath(os.path.abspath(link.file_path))
+
+        # original_link is the direct URL that was provided by the user for the
+        # requirement, either directly or via a constraints file.
+        if link is None and req and req.url:
+            # PEP 508 URL requirement
+            link = Link(req.url)
+        self.link = self.original_link = link
+
+        # When this InstallRequirement is a wheel obtained from the cache of locally
+        # built wheels, this is the source link corresponding to the cache entry, which
+        # was used to download and build the cached wheel.
+        self.cached_wheel_source_link: Optional[Link] = None
+
+        # Information about the location of the artifact that was downloaded . This
+        # property is guaranteed to be set in resolver results.
+        self.download_info: Optional[DirectUrl] = None
+
+        # Path to any downloaded or already-existing package.
+        self.local_file_path: Optional[str] = None
+        if self.link and self.link.is_file:
+            self.local_file_path = self.link.file_path
+
+        if extras:
+            self.extras = extras
+        elif req:
+            self.extras = req.extras
+        else:
+            self.extras = set()
+        if markers is None and req:
+            markers = req.marker
+        self.markers = markers
+
+        # This holds the Distribution object if this requirement is already installed.
+        self.satisfied_by: Optional[BaseDistribution] = None
+        # Whether the installation process should try to uninstall an existing
+        # distribution before installing this requirement.
+        self.should_reinstall = False
+        # Temporary build location
+        self._temp_build_dir: Optional[TempDirectory] = None
+        # Set to True after successful installation
+        self.install_succeeded: Optional[bool] = None
+        # Supplied options
+        self.global_options = global_options if global_options else []
+        self.hash_options = hash_options if hash_options else {}
+        self.config_settings = config_settings
+        # Set to True after successful preparation of this requirement
+        self.prepared = False
+        # User supplied requirement are explicitly requested for installation
+        # by the user via CLI arguments or requirements files, as opposed to,
+        # e.g. dependencies, extras or constraints.
+        self.user_supplied = user_supplied
+
+        self.isolated = isolated
+        self.build_env: BuildEnvironment = NoOpBuildEnvironment()
+
+        # For PEP 517, the directory where we request the project metadata
+        # gets stored. We need this to pass to build_wheel, so the backend
+        # can ensure that the wheel matches the metadata (see the PEP for
+        # details).
+        self.metadata_directory: Optional[str] = None
+
+        # The static build requirements (from pyproject.toml)
+        self.pyproject_requires: Optional[List[str]] = None
+
+        # Build requirements that we will check are available
+        self.requirements_to_check: List[str] = []
+
+        # The PEP 517 backend we should use to build the project
+        self.pep517_backend: Optional[BuildBackendHookCaller] = None
+
+        # Are we using PEP 517 for this requirement?
+        # After pyproject.toml has been loaded, the only valid values are True
+        # and False. Before loading, None is valid (meaning "use the default").
+        # Setting an explicit value before loading pyproject.toml is supported,
+        # but after loading this flag should be treated as read only.
+        self.use_pep517 = use_pep517
+
+        # If config settings are provided, enforce PEP 517.
+        if self.config_settings:
+            if self.use_pep517 is False:
+                logger.warning(
+                    "--no-use-pep517 ignored for %s "
+                    "because --config-settings are specified.",
+                    self,
+                )
+            self.use_pep517 = True
+
+        # This requirement needs more preparation before it can be built
+        self.needs_more_preparation = False
+
+        # This requirement needs to be unpacked before it can be installed.
+        self._archive_source: Optional[Path] = None
+
+    def __str__(self) -> str:
+        if self.req:
+            s = redact_auth_from_requirement(self.req)
+            if self.link:
+                s += f" from {redact_auth_from_url(self.link.url)}"
+        elif self.link:
+            s = redact_auth_from_url(self.link.url)
+        else:
+            s = "<InstallRequirement>"
+        if self.satisfied_by is not None:
+            if self.satisfied_by.location is not None:
+                location = display_path(self.satisfied_by.location)
+            else:
+                location = "<memory>"
+            s += f" in {location}"
+        if self.comes_from:
+            if isinstance(self.comes_from, str):
+                comes_from: Optional[str] = self.comes_from
+            else:
+                comes_from = self.comes_from.from_path()
+            if comes_from:
+                s += f" (from {comes_from})"
+        return s
+
+    def __repr__(self) -> str:
+        return (
+            f"<{self.__class__.__name__} object: "
+            f"{str(self)} editable={self.editable!r}>"
+        )
+
+    def format_debug(self) -> str:
+        """An un-tested helper for getting state, for debugging."""
+        attributes = vars(self)
+        names = sorted(attributes)
+
+        state = (f"{attr}={attributes[attr]!r}" for attr in sorted(names))
+        return "<{name} object: {{{state}}}>".format(
+            name=self.__class__.__name__,
+            state=", ".join(state),
+        )
+
+    # Things that are valid for all kinds of requirements?
+    @property
+    def name(self) -> Optional[str]:
+        if self.req is None:
+            return None
+        return self.req.name
+
+    @functools.cached_property
+    def supports_pyproject_editable(self) -> bool:
+        if not self.use_pep517:
+            return False
+        assert self.pep517_backend
+        with self.build_env:
+            runner = runner_with_spinner_message(
+                "Checking if build backend supports build_editable"
+            )
+            with self.pep517_backend.subprocess_runner(runner):
+                return "build_editable" in self.pep517_backend._supported_features()
+
+    @property
+    def specifier(self) -> SpecifierSet:
+        assert self.req is not None
+        return self.req.specifier
+
+    @property
+    def is_direct(self) -> bool:
+        """Whether this requirement was specified as a direct URL."""
+        return self.original_link is not None
+
+    @property
+    def is_pinned(self) -> bool:
+        """Return whether I am pinned to an exact version.
+
+        For example, some-package==1.2 is pinned; some-package>1.2 is not.
+        """
+        assert self.req is not None
+        specifiers = self.req.specifier
+        return len(specifiers) == 1 and next(iter(specifiers)).operator in {"==", "==="}
+
+    def match_markers(self, extras_requested: Optional[Iterable[str]] = None) -> bool:
+        if not extras_requested:
+            # Provide an extra to safely evaluate the markers
+            # without matching any extra
+            extras_requested = ("",)
+        if self.markers is not None:
+            return any(
+                self.markers.evaluate({"extra": extra}) for extra in extras_requested
+            )
+        else:
+            return True
+
+    @property
+    def has_hash_options(self) -> bool:
+        """Return whether any known-good hashes are specified as options.
+
+        These activate --require-hashes mode; hashes specified as part of a
+        URL do not.
+
+        """
+        return bool(self.hash_options)
+
+    def hashes(self, trust_internet: bool = True) -> Hashes:
+        """Return a hash-comparer that considers my option- and URL-based
+        hashes to be known-good.
+
+        Hashes in URLs--ones embedded in the requirements file, not ones
+        downloaded from an index server--are almost peers with ones from
+        flags. They satisfy --require-hashes (whether it was implicitly or
+        explicitly activated) but do not activate it. md5 and sha224 are not
+        allowed in flags, which should nudge people toward good algos. We
+        always OR all hashes together, even ones from URLs.
+
+        :param trust_internet: Whether to trust URL-based (#md5=...) hashes
+            downloaded from the internet, as by populate_link()
+
+        """
+        good_hashes = self.hash_options.copy()
+        if trust_internet:
+            link = self.link
+        elif self.is_direct and self.user_supplied:
+            link = self.original_link
+        else:
+            link = None
+        if link and link.hash:
+            assert link.hash_name is not None
+            good_hashes.setdefault(link.hash_name, []).append(link.hash)
+        return Hashes(good_hashes)
+
+    def from_path(self) -> Optional[str]:
+        """Format a nice indicator to show where this "comes from" """
+        if self.req is None:
+            return None
+        s = str(self.req)
+        if self.comes_from:
+            comes_from: Optional[str]
+            if isinstance(self.comes_from, str):
+                comes_from = self.comes_from
+            else:
+                comes_from = self.comes_from.from_path()
+            if comes_from:
+                s += "->" + comes_from
+        return s
+
+    def ensure_build_location(
+        self, build_dir: str, autodelete: bool, parallel_builds: bool
+    ) -> str:
+        assert build_dir is not None
+        if self._temp_build_dir is not None:
+            assert self._temp_build_dir.path
+            return self._temp_build_dir.path
+        if self.req is None:
+            # Some systems have /tmp as a symlink which confuses custom
+            # builds (such as numpy). Thus, we ensure that the real path
+            # is returned.
+            self._temp_build_dir = TempDirectory(
+                kind=tempdir_kinds.REQ_BUILD, globally_managed=True
+            )
+
+            return self._temp_build_dir.path
+
+        # This is the only remaining place where we manually determine the path
+        # for the temporary directory. It is only needed for editables where
+        # it is the value of the --src option.
+
+        # When parallel builds are enabled, add a UUID to the build directory
+        # name so multiple builds do not interfere with each other.
+        dir_name: str = canonicalize_name(self.req.name)
+        if parallel_builds:
+            dir_name = f"{dir_name}_{uuid.uuid4().hex}"
+
+        # FIXME: Is there a better place to create the build_dir? (hg and bzr
+        # need this)
+        if not os.path.exists(build_dir):
+            logger.debug("Creating directory %s", build_dir)
+            os.makedirs(build_dir)
+        actual_build_dir = os.path.join(build_dir, dir_name)
+        # `None` indicates that we respect the globally-configured deletion
+        # settings, which is what we actually want when auto-deleting.
+        delete_arg = None if autodelete else False
+        return TempDirectory(
+            path=actual_build_dir,
+            delete=delete_arg,
+            kind=tempdir_kinds.REQ_BUILD,
+            globally_managed=True,
+        ).path
+
+    def _set_requirement(self) -> None:
+        """Set requirement after generating metadata."""
+        assert self.req is None
+        assert self.metadata is not None
+        assert self.source_dir is not None
+
+        # Construct a Requirement object from the generated metadata
+        if isinstance(parse_version(self.metadata["Version"]), Version):
+            op = "=="
+        else:
+            op = "==="
+
+        self.req = get_requirement(
+            "".join(
+                [
+                    self.metadata["Name"],
+                    op,
+                    self.metadata["Version"],
+                ]
+            )
+        )
+
+    def warn_on_mismatching_name(self) -> None:
+        assert self.req is not None
+        metadata_name = canonicalize_name(self.metadata["Name"])
+        if canonicalize_name(self.req.name) == metadata_name:
+            # Everything is fine.
+            return
+
+        # If we're here, there's a mismatch. Log a warning about it.
+        logger.warning(
+            "Generating metadata for package %s "
+            "produced metadata for project name %s. Fix your "
+            "#egg=%s fragments.",
+            self.name,
+            metadata_name,
+            self.name,
+        )
+        self.req = get_requirement(metadata_name)
+
+    def check_if_exists(self, use_user_site: bool) -> None:
+        """Find an installed distribution that satisfies or conflicts
+        with this requirement, and set self.satisfied_by or
+        self.should_reinstall appropriately.
+        """
+        if self.req is None:
+            return
+        existing_dist = get_default_environment().get_distribution(self.req.name)
+        if not existing_dist:
+            return
+
+        version_compatible = self.req.specifier.contains(
+            existing_dist.version,
+            prereleases=True,
+        )
+        if not version_compatible:
+            self.satisfied_by = None
+            if use_user_site:
+                if existing_dist.in_usersite:
+                    self.should_reinstall = True
+                elif running_under_virtualenv() and existing_dist.in_site_packages:
+                    raise InstallationError(
+                        f"Will not install to the user site because it will "
+                        f"lack sys.path precedence to {existing_dist.raw_name} "
+                        f"in {existing_dist.location}"
+                    )
+            else:
+                self.should_reinstall = True
+        else:
+            if self.editable:
+                self.should_reinstall = True
+                # when installing editables, nothing pre-existing should ever
+                # satisfy
+                self.satisfied_by = None
+            else:
+                self.satisfied_by = existing_dist
+
+    # Things valid for wheels
+    @property
+    def is_wheel(self) -> bool:
+        if not self.link:
+            return False
+        return self.link.is_wheel
+
+    @property
+    def is_wheel_from_cache(self) -> bool:
+        # When True, it means that this InstallRequirement is a local wheel file in the
+        # cache of locally built wheels.
+        return self.cached_wheel_source_link is not None
+
+    # Things valid for sdists
+    @property
+    def unpacked_source_directory(self) -> str:
+        assert self.source_dir, f"No source dir for {self}"
+        return os.path.join(
+            self.source_dir, self.link and self.link.subdirectory_fragment or ""
+        )
+
+    @property
+    def setup_py_path(self) -> str:
+        assert self.source_dir, f"No source dir for {self}"
+        setup_py = os.path.join(self.unpacked_source_directory, "setup.py")
+
+        return setup_py
+
+    @property
+    def setup_cfg_path(self) -> str:
+        assert self.source_dir, f"No source dir for {self}"
+        setup_cfg = os.path.join(self.unpacked_source_directory, "setup.cfg")
+
+        return setup_cfg
+
+    @property
+    def pyproject_toml_path(self) -> str:
+        assert self.source_dir, f"No source dir for {self}"
+        return make_pyproject_path(self.unpacked_source_directory)
+
+    def load_pyproject_toml(self) -> None:
+        """Load the pyproject.toml file.
+
+        After calling this routine, all of the attributes related to PEP 517
+        processing for this requirement have been set. In particular, the
+        use_pep517 attribute can be used to determine whether we should
+        follow the PEP 517 or legacy (setup.py) code path.
+        """
+        pyproject_toml_data = load_pyproject_toml(
+            self.use_pep517, self.pyproject_toml_path, self.setup_py_path, str(self)
+        )
+
+        if pyproject_toml_data is None:
+            assert not self.config_settings
+            self.use_pep517 = False
+            return
+
+        self.use_pep517 = True
+        requires, backend, check, backend_path = pyproject_toml_data
+        self.requirements_to_check = check
+        self.pyproject_requires = requires
+        self.pep517_backend = ConfiguredBuildBackendHookCaller(
+            self,
+            self.unpacked_source_directory,
+            backend,
+            backend_path=backend_path,
+        )
+
+    def isolated_editable_sanity_check(self) -> None:
+        """Check that an editable requirement if valid for use with PEP 517/518.
+
+        This verifies that an editable that has a pyproject.toml either supports PEP 660
+        or as a setup.py or a setup.cfg
+        """
+        if (
+            self.editable
+            and self.use_pep517
+            and not self.supports_pyproject_editable
+            and not os.path.isfile(self.setup_py_path)
+            and not os.path.isfile(self.setup_cfg_path)
+        ):
+            raise InstallationError(
+                f"Project {self} has a 'pyproject.toml' and its build "
+                f"backend is missing the 'build_editable' hook. Since it does not "
+                f"have a 'setup.py' nor a 'setup.cfg', "
+                f"it cannot be installed in editable mode. "
+                f"Consider using a build backend that supports PEP 660."
+            )
+
+    def prepare_metadata(self) -> None:
+        """Ensure that project metadata is available.
+
+        Under PEP 517 and PEP 660, call the backend hook to prepare the metadata.
+        Under legacy processing, call setup.py egg-info.
+        """
+        assert self.source_dir, f"No source dir for {self}"
+        details = self.name or f"from {self.link}"
+
+        if self.use_pep517:
+            assert self.pep517_backend is not None
+            if (
+                self.editable
+                and self.permit_editable_wheels
+                and self.supports_pyproject_editable
+            ):
+                self.metadata_directory = generate_editable_metadata(
+                    build_env=self.build_env,
+                    backend=self.pep517_backend,
+                    details=details,
+                )
+            else:
+                self.metadata_directory = generate_metadata(
+                    build_env=self.build_env,
+                    backend=self.pep517_backend,
+                    details=details,
+                )
+        else:
+            self.metadata_directory = generate_metadata_legacy(
+                build_env=self.build_env,
+                setup_py_path=self.setup_py_path,
+                source_dir=self.unpacked_source_directory,
+                isolated=self.isolated,
+                details=details,
+            )
+
+        # Act on the newly generated metadata, based on the name and version.
+        if not self.name:
+            self._set_requirement()
+        else:
+            self.warn_on_mismatching_name()
+
+        self.assert_source_matches_version()
+
+    @property
+    def metadata(self) -> Any:
+        if not hasattr(self, "_metadata"):
+            self._metadata = self.get_dist().metadata
+
+        return self._metadata
+
+    def get_dist(self) -> BaseDistribution:
+        if self.metadata_directory:
+            return get_directory_distribution(self.metadata_directory)
+        elif self.local_file_path and self.is_wheel:
+            assert self.req is not None
+            return get_wheel_distribution(
+                FilesystemWheel(self.local_file_path),
+                canonicalize_name(self.req.name),
+            )
+        raise AssertionError(
+            f"InstallRequirement {self} has no metadata directory and no wheel: "
+            f"can't make a distribution."
+        )
+
+    def assert_source_matches_version(self) -> None:
+        assert self.source_dir, f"No source dir for {self}"
+        version = self.metadata["version"]
+        if self.req and self.req.specifier and version not in self.req.specifier:
+            logger.warning(
+                "Requested %s, but installing version %s",
+                self,
+                version,
+            )
+        else:
+            logger.debug(
+                "Source in %s has version %s, which satisfies requirement %s",
+                display_path(self.source_dir),
+                version,
+                self,
+            )
+
+    # For both source distributions and editables
+    def ensure_has_source_dir(
+        self,
+        parent_dir: str,
+        autodelete: bool = False,
+        parallel_builds: bool = False,
+    ) -> None:
+        """Ensure that a source_dir is set.
+
+        This will create a temporary build dir if the name of the requirement
+        isn't known yet.
+
+        :param parent_dir: The ideal pip parent_dir for the source_dir.
+            Generally src_dir for editables and build_dir for sdists.
+        :return: self.source_dir
+        """
+        if self.source_dir is None:
+            self.source_dir = self.ensure_build_location(
+                parent_dir,
+                autodelete=autodelete,
+                parallel_builds=parallel_builds,
+            )
+
+    def needs_unpacked_archive(self, archive_source: Path) -> None:
+        assert self._archive_source is None
+        self._archive_source = archive_source
+
+    def ensure_pristine_source_checkout(self) -> None:
+        """Ensure the source directory has not yet been built in."""
+        assert self.source_dir is not None
+        if self._archive_source is not None:
+            unpack_file(str(self._archive_source), self.source_dir)
+        elif is_installable_dir(self.source_dir):
+            # If a checkout exists, it's unwise to keep going.
+            # version inconsistencies are logged later, but do not fail
+            # the installation.
+            raise PreviousBuildDirError(
+                f"pip can't proceed with requirements '{self}' due to a "
+                f"pre-existing build directory ({self.source_dir}). This is likely "
+                "due to a previous installation that failed . pip is "
+                "being responsible and not assuming it can delete this. "
+                "Please delete it and try again."
+            )
+
+    # For editable installations
+    def update_editable(self) -> None:
+        if not self.link:
+            logger.debug(
+                "Cannot update repository at %s; repository location is unknown",
+                self.source_dir,
+            )
+            return
+        assert self.editable
+        assert self.source_dir
+        if self.link.scheme == "file":
+            # Static paths don't get updated
+            return
+        vcs_backend = vcs.get_backend_for_scheme(self.link.scheme)
+        # Editable requirements are validated in Requirement constructors.
+        # So here, if it's neither a path nor a valid VCS URL, it's a bug.
+        assert vcs_backend, f"Unsupported VCS URL {self.link.url}"
+        hidden_url = hide_url(self.link.url)
+        vcs_backend.obtain(self.source_dir, url=hidden_url, verbosity=0)
+
+    # Top-level Actions
+    def uninstall(
+        self, auto_confirm: bool = False, verbose: bool = False
+    ) -> Optional[UninstallPathSet]:
+        """
+        Uninstall the distribution currently satisfying this requirement.
+
+        Prompts before removing or modifying files unless
+        ``auto_confirm`` is True.
+
+        Refuses to delete or modify files outside of ``sys.prefix`` -
+        thus uninstallation within a virtual environment can only
+        modify that virtual environment, even if the virtualenv is
+        linked to global site-packages.
+
+        """
+        assert self.req
+        dist = get_default_environment().get_distribution(self.req.name)
+        if not dist:
+            logger.warning("Skipping %s as it is not installed.", self.name)
+            return None
+        logger.info("Found existing installation: %s", dist)
+
+        uninstalled_pathset = UninstallPathSet.from_dist(dist)
+        uninstalled_pathset.remove(auto_confirm, verbose)
+        return uninstalled_pathset
+
+    def _get_archive_name(self, path: str, parentdir: str, rootdir: str) -> str:
+        def _clean_zip_name(name: str, prefix: str) -> str:
+            assert name.startswith(
+                prefix + os.path.sep
+            ), f"name {name!r} doesn't start with prefix {prefix!r}"
+            name = name[len(prefix) + 1 :]
+            name = name.replace(os.path.sep, "/")
+            return name
+
+        assert self.req is not None
+        path = os.path.join(parentdir, path)
+        name = _clean_zip_name(path, rootdir)
+        return self.req.name + "/" + name
+
+    def archive(self, build_dir: Optional[str]) -> None:
+        """Saves archive to provided build_dir.
+
+        Used for saving downloaded VCS requirements as part of `pip download`.
+        """
+        assert self.source_dir
+        if build_dir is None:
+            return
+
+        create_archive = True
+        archive_name = "{}-{}.zip".format(self.name, self.metadata["version"])
+        archive_path = os.path.join(build_dir, archive_name)
+
+        if os.path.exists(archive_path):
+            response = ask_path_exists(
+                f"The file {display_path(archive_path)} exists. (i)gnore, (w)ipe, "
+                "(b)ackup, (a)bort ",
+                ("i", "w", "b", "a"),
+            )
+            if response == "i":
+                create_archive = False
+            elif response == "w":
+                logger.warning("Deleting %s", display_path(archive_path))
+                os.remove(archive_path)
+            elif response == "b":
+                dest_file = backup_dir(archive_path)
+                logger.warning(
+                    "Backing up %s to %s",
+                    display_path(archive_path),
+                    display_path(dest_file),
+                )
+                shutil.move(archive_path, dest_file)
+            elif response == "a":
+                sys.exit(-1)
+
+        if not create_archive:
+            return
+
+        zip_output = zipfile.ZipFile(
+            archive_path,
+            "w",
+            zipfile.ZIP_DEFLATED,
+            allowZip64=True,
+        )
+        with zip_output:
+            dir = os.path.normcase(os.path.abspath(self.unpacked_source_directory))
+            for dirpath, dirnames, filenames in os.walk(dir):
+                for dirname in dirnames:
+                    dir_arcname = self._get_archive_name(
+                        dirname,
+                        parentdir=dirpath,
+                        rootdir=dir,
+                    )
+                    zipdir = zipfile.ZipInfo(dir_arcname + "/")
+                    zipdir.external_attr = 0x1ED << 16  # 0o755
+                    zip_output.writestr(zipdir, "")
+                for filename in filenames:
+                    file_arcname = self._get_archive_name(
+                        filename,
+                        parentdir=dirpath,
+                        rootdir=dir,
+                    )
+                    filename = os.path.join(dirpath, filename)
+                    zip_output.write(filename, file_arcname)
+
+        logger.info("Saved %s", display_path(archive_path))
+
+    def install(
+        self,
+        global_options: Optional[Sequence[str]] = None,
+        root: Optional[str] = None,
+        home: Optional[str] = None,
+        prefix: Optional[str] = None,
+        warn_script_location: bool = True,
+        use_user_site: bool = False,
+        pycompile: bool = True,
+    ) -> None:
+        assert self.req is not None
+        scheme = get_scheme(
+            self.req.name,
+            user=use_user_site,
+            home=home,
+            root=root,
+            isolated=self.isolated,
+            prefix=prefix,
+        )
+
+        if self.editable and not self.is_wheel:
+            deprecated(
+                reason=(
+                    f"Legacy editable install of {self} (setup.py develop) "
+                    "is deprecated."
+                ),
+                replacement=(
+                    "to add a pyproject.toml or enable --use-pep517, "
+                    "and use setuptools >= 64. "
+                    "If the resulting installation is not behaving as expected, "
+                    "try using --config-settings editable_mode=compat. "
+                    "Please consult the setuptools documentation for more information"
+                ),
+                gone_in="25.1",
+                issue=11457,
+            )
+            if self.config_settings:
+                logger.warning(
+                    "--config-settings ignored for legacy editable install of %s. "
+                    "Consider upgrading to a version of setuptools "
+                    "that supports PEP 660 (>= 64).",
+                    self,
+                )
+            install_editable_legacy(
+                global_options=global_options if global_options is not None else [],
+                prefix=prefix,
+                home=home,
+                use_user_site=use_user_site,
+                name=self.req.name,
+                setup_py_path=self.setup_py_path,
+                isolated=self.isolated,
+                build_env=self.build_env,
+                unpacked_source_directory=self.unpacked_source_directory,
+            )
+            self.install_succeeded = True
+            return
+
+        assert self.is_wheel
+        assert self.local_file_path
+
+        install_wheel(
+            self.req.name,
+            self.local_file_path,
+            scheme=scheme,
+            req_description=str(self.req),
+            pycompile=pycompile,
+            warn_script_location=warn_script_location,
+            direct_url=self.download_info if self.is_direct else None,
+            requested=self.user_supplied,
+        )
+        self.install_succeeded = True
+
+
+def check_invalid_constraint_type(req: InstallRequirement) -> str:
+    # Check for unsupported forms
+    problem = ""
+    if not req.name:
+        problem = "Unnamed requirements are not allowed as constraints"
+    elif req.editable:
+        problem = "Editable requirements are not allowed as constraints"
+    elif req.extras:
+        problem = "Constraints cannot have extras"
+
+    if problem:
+        deprecated(
+            reason=(
+                "Constraints are only allowed to take the form of a package "
+                "name and a version specifier. Other forms were originally "
+                "permitted as an accident of the implementation, but were "
+                "undocumented. The new implementation of the resolver no "
+                "longer supports these forms."
+            ),
+            replacement="replacing the constraint with a requirement",
+            # No plan yet for when the new resolver becomes default
+            gone_in=None,
+            issue=8210,
+        )
+
+    return problem
+
+
+def _has_option(options: Values, reqs: List[InstallRequirement], option: str) -> bool:
+    if getattr(options, option, None):
+        return True
+    for req in reqs:
+        if getattr(req, option, None):
+            return True
+    return False
+
+
+def check_legacy_setup_py_options(
+    options: Values,
+    reqs: List[InstallRequirement],
+) -> None:
+    has_build_options = _has_option(options, reqs, "build_options")
+    has_global_options = _has_option(options, reqs, "global_options")
+    if has_build_options or has_global_options:
+        deprecated(
+            reason="--build-option and --global-option are deprecated.",
+            issue=11859,
+            replacement="to use --config-settings",
+            gone_in=None,
+        )
+        logger.warning(
+            "Implying --no-binary=:all: due to the presence of "
+            "--build-option / --global-option. "
+        )
+        options.format_control.disallow_binaries()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_set.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_set.py
new file mode 100644
index 00000000..ec7a6e07
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_set.py
@@ -0,0 +1,82 @@
+import logging
+from collections import OrderedDict
+from typing import Dict, List
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.req.req_install import InstallRequirement
+
+logger = logging.getLogger(__name__)
+
+
+class RequirementSet:
+    def __init__(self, check_supported_wheels: bool = True) -> None:
+        """Create a RequirementSet."""
+
+        self.requirements: Dict[str, InstallRequirement] = OrderedDict()
+        self.check_supported_wheels = check_supported_wheels
+
+        self.unnamed_requirements: List[InstallRequirement] = []
+
+    def __str__(self) -> str:
+        requirements = sorted(
+            (req for req in self.requirements.values() if not req.comes_from),
+            key=lambda req: canonicalize_name(req.name or ""),
+        )
+        return " ".join(str(req.req) for req in requirements)
+
+    def __repr__(self) -> str:
+        requirements = sorted(
+            self.requirements.values(),
+            key=lambda req: canonicalize_name(req.name or ""),
+        )
+
+        format_string = "<{classname} object; {count} requirement(s): {reqs}>"
+        return format_string.format(
+            classname=self.__class__.__name__,
+            count=len(requirements),
+            reqs=", ".join(str(req.req) for req in requirements),
+        )
+
+    def add_unnamed_requirement(self, install_req: InstallRequirement) -> None:
+        assert not install_req.name
+        self.unnamed_requirements.append(install_req)
+
+    def add_named_requirement(self, install_req: InstallRequirement) -> None:
+        assert install_req.name
+
+        project_name = canonicalize_name(install_req.name)
+        self.requirements[project_name] = install_req
+
+    def has_requirement(self, name: str) -> bool:
+        project_name = canonicalize_name(name)
+
+        return (
+            project_name in self.requirements
+            and not self.requirements[project_name].constraint
+        )
+
+    def get_requirement(self, name: str) -> InstallRequirement:
+        project_name = canonicalize_name(name)
+
+        if project_name in self.requirements:
+            return self.requirements[project_name]
+
+        raise KeyError(f"No project with the name {name!r}")
+
+    @property
+    def all_requirements(self) -> List[InstallRequirement]:
+        return self.unnamed_requirements + list(self.requirements.values())
+
+    @property
+    def requirements_to_install(self) -> List[InstallRequirement]:
+        """Return the list of requirements that need to be installed.
+
+        TODO remove this property together with the legacy resolver, since the new
+             resolver only returns requirements that need to be installed.
+        """
+        return [
+            install_req
+            for install_req in self.all_requirements
+            if not install_req.constraint and not install_req.satisfied_by
+        ]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_uninstall.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_uninstall.py
new file mode 100644
index 00000000..26df2084
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/req/req_uninstall.py
@@ -0,0 +1,633 @@
+import functools
+import os
+import sys
+import sysconfig
+from importlib.util import cache_from_source
+from typing import Any, Callable, Dict, Generator, Iterable, List, Optional, Set, Tuple
+
+from pip._internal.exceptions import LegacyDistutilsInstall, UninstallMissingRecord
+from pip._internal.locations import get_bin_prefix, get_bin_user
+from pip._internal.metadata import BaseDistribution
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.egg_link import egg_link_path_from_location
+from pip._internal.utils.logging import getLogger, indent_log
+from pip._internal.utils.misc import ask, normalize_path, renames, rmtree
+from pip._internal.utils.temp_dir import AdjacentTempDirectory, TempDirectory
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+logger = getLogger(__name__)
+
+
+def _script_names(
+    bin_dir: str, script_name: str, is_gui: bool
+) -> Generator[str, None, None]:
+    """Create the fully qualified name of the files created by
+    {console,gui}_scripts for the given ``dist``.
+    Returns the list of file names
+    """
+    exe_name = os.path.join(bin_dir, script_name)
+    yield exe_name
+    if not WINDOWS:
+        return
+    yield f"{exe_name}.exe"
+    yield f"{exe_name}.exe.manifest"
+    if is_gui:
+        yield f"{exe_name}-script.pyw"
+    else:
+        yield f"{exe_name}-script.py"
+
+
+def _unique(
+    fn: Callable[..., Generator[Any, None, None]]
+) -> Callable[..., Generator[Any, None, None]]:
+    @functools.wraps(fn)
+    def unique(*args: Any, **kw: Any) -> Generator[Any, None, None]:
+        seen: Set[Any] = set()
+        for item in fn(*args, **kw):
+            if item not in seen:
+                seen.add(item)
+                yield item
+
+    return unique
+
+
+@_unique
+def uninstallation_paths(dist: BaseDistribution) -> Generator[str, None, None]:
+    """
+    Yield all the uninstallation paths for dist based on RECORD-without-.py[co]
+
+    Yield paths to all the files in RECORD. For each .py file in RECORD, add
+    the .pyc and .pyo in the same directory.
+
+    UninstallPathSet.add() takes care of the __pycache__ .py[co].
+
+    If RECORD is not found, raises an error,
+    with possible information from the INSTALLER file.
+
+    https://packaging.python.org/specifications/recording-installed-packages/
+    """
+    location = dist.location
+    assert location is not None, "not installed"
+
+    entries = dist.iter_declared_entries()
+    if entries is None:
+        raise UninstallMissingRecord(distribution=dist)
+
+    for entry in entries:
+        path = os.path.join(location, entry)
+        yield path
+        if path.endswith(".py"):
+            dn, fn = os.path.split(path)
+            base = fn[:-3]
+            path = os.path.join(dn, base + ".pyc")
+            yield path
+            path = os.path.join(dn, base + ".pyo")
+            yield path
+
+
+def compact(paths: Iterable[str]) -> Set[str]:
+    """Compact a path set to contain the minimal number of paths
+    necessary to contain all paths in the set. If /a/path/ and
+    /a/path/to/a/file.txt are both in the set, leave only the
+    shorter path."""
+
+    sep = os.path.sep
+    short_paths: Set[str] = set()
+    for path in sorted(paths, key=len):
+        should_skip = any(
+            path.startswith(shortpath.rstrip("*"))
+            and path[len(shortpath.rstrip("*").rstrip(sep))] == sep
+            for shortpath in short_paths
+        )
+        if not should_skip:
+            short_paths.add(path)
+    return short_paths
+
+
+def compress_for_rename(paths: Iterable[str]) -> Set[str]:
+    """Returns a set containing the paths that need to be renamed.
+
+    This set may include directories when the original sequence of paths
+    included every file on disk.
+    """
+    case_map = {os.path.normcase(p): p for p in paths}
+    remaining = set(case_map)
+    unchecked = sorted({os.path.split(p)[0] for p in case_map.values()}, key=len)
+    wildcards: Set[str] = set()
+
+    def norm_join(*a: str) -> str:
+        return os.path.normcase(os.path.join(*a))
+
+    for root in unchecked:
+        if any(os.path.normcase(root).startswith(w) for w in wildcards):
+            # This directory has already been handled.
+            continue
+
+        all_files: Set[str] = set()
+        all_subdirs: Set[str] = set()
+        for dirname, subdirs, files in os.walk(root):
+            all_subdirs.update(norm_join(root, dirname, d) for d in subdirs)
+            all_files.update(norm_join(root, dirname, f) for f in files)
+        # If all the files we found are in our remaining set of files to
+        # remove, then remove them from the latter set and add a wildcard
+        # for the directory.
+        if not (all_files - remaining):
+            remaining.difference_update(all_files)
+            wildcards.add(root + os.sep)
+
+    return set(map(case_map.__getitem__, remaining)) | wildcards
+
+
+def compress_for_output_listing(paths: Iterable[str]) -> Tuple[Set[str], Set[str]]:
+    """Returns a tuple of 2 sets of which paths to display to user
+
+    The first set contains paths that would be deleted. Files of a package
+    are not added and the top-level directory of the package has a '*' added
+    at the end - to signify that all it's contents are removed.
+
+    The second set contains files that would have been skipped in the above
+    folders.
+    """
+
+    will_remove = set(paths)
+    will_skip = set()
+
+    # Determine folders and files
+    folders = set()
+    files = set()
+    for path in will_remove:
+        if path.endswith(".pyc"):
+            continue
+        if path.endswith("__init__.py") or ".dist-info" in path:
+            folders.add(os.path.dirname(path))
+        files.add(path)
+
+    _normcased_files = set(map(os.path.normcase, files))
+
+    folders = compact(folders)
+
+    # This walks the tree using os.walk to not miss extra folders
+    # that might get added.
+    for folder in folders:
+        for dirpath, _, dirfiles in os.walk(folder):
+            for fname in dirfiles:
+                if fname.endswith(".pyc"):
+                    continue
+
+                file_ = os.path.join(dirpath, fname)
+                if (
+                    os.path.isfile(file_)
+                    and os.path.normcase(file_) not in _normcased_files
+                ):
+                    # We are skipping this file. Add it to the set.
+                    will_skip.add(file_)
+
+    will_remove = files | {os.path.join(folder, "*") for folder in folders}
+
+    return will_remove, will_skip
+
+
+class StashedUninstallPathSet:
+    """A set of file rename operations to stash files while
+    tentatively uninstalling them."""
+
+    def __init__(self) -> None:
+        # Mapping from source file root to [Adjacent]TempDirectory
+        # for files under that directory.
+        self._save_dirs: Dict[str, TempDirectory] = {}
+        # (old path, new path) tuples for each move that may need
+        # to be undone.
+        self._moves: List[Tuple[str, str]] = []
+
+    def _get_directory_stash(self, path: str) -> str:
+        """Stashes a directory.
+
+        Directories are stashed adjacent to their original location if
+        possible, or else moved/copied into the user's temp dir."""
+
+        try:
+            save_dir: TempDirectory = AdjacentTempDirectory(path)
+        except OSError:
+            save_dir = TempDirectory(kind="uninstall")
+        self._save_dirs[os.path.normcase(path)] = save_dir
+
+        return save_dir.path
+
+    def _get_file_stash(self, path: str) -> str:
+        """Stashes a file.
+
+        If no root has been provided, one will be created for the directory
+        in the user's temp directory."""
+        path = os.path.normcase(path)
+        head, old_head = os.path.dirname(path), None
+        save_dir = None
+
+        while head != old_head:
+            try:
+                save_dir = self._save_dirs[head]
+                break
+            except KeyError:
+                pass
+            head, old_head = os.path.dirname(head), head
+        else:
+            # Did not find any suitable root
+            head = os.path.dirname(path)
+            save_dir = TempDirectory(kind="uninstall")
+            self._save_dirs[head] = save_dir
+
+        relpath = os.path.relpath(path, head)
+        if relpath and relpath != os.path.curdir:
+            return os.path.join(save_dir.path, relpath)
+        return save_dir.path
+
+    def stash(self, path: str) -> str:
+        """Stashes the directory or file and returns its new location.
+        Handle symlinks as files to avoid modifying the symlink targets.
+        """
+        path_is_dir = os.path.isdir(path) and not os.path.islink(path)
+        if path_is_dir:
+            new_path = self._get_directory_stash(path)
+        else:
+            new_path = self._get_file_stash(path)
+
+        self._moves.append((path, new_path))
+        if path_is_dir and os.path.isdir(new_path):
+            # If we're moving a directory, we need to
+            # remove the destination first or else it will be
+            # moved to inside the existing directory.
+            # We just created new_path ourselves, so it will
+            # be removable.
+            os.rmdir(new_path)
+        renames(path, new_path)
+        return new_path
+
+    def commit(self) -> None:
+        """Commits the uninstall by removing stashed files."""
+        for save_dir in self._save_dirs.values():
+            save_dir.cleanup()
+        self._moves = []
+        self._save_dirs = {}
+
+    def rollback(self) -> None:
+        """Undoes the uninstall by moving stashed files back."""
+        for p in self._moves:
+            logger.info("Moving to %s\n from %s", *p)
+
+        for new_path, path in self._moves:
+            try:
+                logger.debug("Replacing %s from %s", new_path, path)
+                if os.path.isfile(new_path) or os.path.islink(new_path):
+                    os.unlink(new_path)
+                elif os.path.isdir(new_path):
+                    rmtree(new_path)
+                renames(path, new_path)
+            except OSError as ex:
+                logger.error("Failed to restore %s", new_path)
+                logger.debug("Exception: %s", ex)
+
+        self.commit()
+
+    @property
+    def can_rollback(self) -> bool:
+        return bool(self._moves)
+
+
+class UninstallPathSet:
+    """A set of file paths to be removed in the uninstallation of a
+    requirement."""
+
+    def __init__(self, dist: BaseDistribution) -> None:
+        self._paths: Set[str] = set()
+        self._refuse: Set[str] = set()
+        self._pth: Dict[str, UninstallPthEntries] = {}
+        self._dist = dist
+        self._moved_paths = StashedUninstallPathSet()
+        # Create local cache of normalize_path results. Creating an UninstallPathSet
+        # can result in hundreds/thousands of redundant calls to normalize_path with
+        # the same args, which hurts performance.
+        self._normalize_path_cached = functools.lru_cache(normalize_path)
+
+    def _permitted(self, path: str) -> bool:
+        """
+        Return True if the given path is one we are permitted to
+        remove/modify, False otherwise.
+
+        """
+        # aka is_local, but caching normalized sys.prefix
+        if not running_under_virtualenv():
+            return True
+        return path.startswith(self._normalize_path_cached(sys.prefix))
+
+    def add(self, path: str) -> None:
+        head, tail = os.path.split(path)
+
+        # we normalize the head to resolve parent directory symlinks, but not
+        # the tail, since we only want to uninstall symlinks, not their targets
+        path = os.path.join(self._normalize_path_cached(head), os.path.normcase(tail))
+
+        if not os.path.exists(path):
+            return
+        if self._permitted(path):
+            self._paths.add(path)
+        else:
+            self._refuse.add(path)
+
+        # __pycache__ files can show up after 'installed-files.txt' is created,
+        # due to imports
+        if os.path.splitext(path)[1] == ".py":
+            self.add(cache_from_source(path))
+
+    def add_pth(self, pth_file: str, entry: str) -> None:
+        pth_file = self._normalize_path_cached(pth_file)
+        if self._permitted(pth_file):
+            if pth_file not in self._pth:
+                self._pth[pth_file] = UninstallPthEntries(pth_file)
+            self._pth[pth_file].add(entry)
+        else:
+            self._refuse.add(pth_file)
+
+    def remove(self, auto_confirm: bool = False, verbose: bool = False) -> None:
+        """Remove paths in ``self._paths`` with confirmation (unless
+        ``auto_confirm`` is True)."""
+
+        if not self._paths:
+            logger.info(
+                "Can't uninstall '%s'. No files were found to uninstall.",
+                self._dist.raw_name,
+            )
+            return
+
+        dist_name_version = f"{self._dist.raw_name}-{self._dist.raw_version}"
+        logger.info("Uninstalling %s:", dist_name_version)
+
+        with indent_log():
+            if auto_confirm or self._allowed_to_proceed(verbose):
+                moved = self._moved_paths
+
+                for_rename = compress_for_rename(self._paths)
+
+                for path in sorted(compact(for_rename)):
+                    moved.stash(path)
+                    logger.verbose("Removing file or directory %s", path)
+
+                for pth in self._pth.values():
+                    pth.remove()
+
+                logger.info("Successfully uninstalled %s", dist_name_version)
+
+    def _allowed_to_proceed(self, verbose: bool) -> bool:
+        """Display which files would be deleted and prompt for confirmation"""
+
+        def _display(msg: str, paths: Iterable[str]) -> None:
+            if not paths:
+                return
+
+            logger.info(msg)
+            with indent_log():
+                for path in sorted(compact(paths)):
+                    logger.info(path)
+
+        if not verbose:
+            will_remove, will_skip = compress_for_output_listing(self._paths)
+        else:
+            # In verbose mode, display all the files that are going to be
+            # deleted.
+            will_remove = set(self._paths)
+            will_skip = set()
+
+        _display("Would remove:", will_remove)
+        _display("Would not remove (might be manually added):", will_skip)
+        _display("Would not remove (outside of prefix):", self._refuse)
+        if verbose:
+            _display("Will actually move:", compress_for_rename(self._paths))
+
+        return ask("Proceed (Y/n)? ", ("y", "n", "")) != "n"
+
+    def rollback(self) -> None:
+        """Rollback the changes previously made by remove()."""
+        if not self._moved_paths.can_rollback:
+            logger.error(
+                "Can't roll back %s; was not uninstalled",
+                self._dist.raw_name,
+            )
+            return
+        logger.info("Rolling back uninstall of %s", self._dist.raw_name)
+        self._moved_paths.rollback()
+        for pth in self._pth.values():
+            pth.rollback()
+
+    def commit(self) -> None:
+        """Remove temporary save dir: rollback will no longer be possible."""
+        self._moved_paths.commit()
+
+    @classmethod
+    def from_dist(cls, dist: BaseDistribution) -> "UninstallPathSet":
+        dist_location = dist.location
+        info_location = dist.info_location
+        if dist_location is None:
+            logger.info(
+                "Not uninstalling %s since it is not installed",
+                dist.canonical_name,
+            )
+            return cls(dist)
+
+        normalized_dist_location = normalize_path(dist_location)
+        if not dist.local:
+            logger.info(
+                "Not uninstalling %s at %s, outside environment %s",
+                dist.canonical_name,
+                normalized_dist_location,
+                sys.prefix,
+            )
+            return cls(dist)
+
+        if normalized_dist_location in {
+            p
+            for p in {sysconfig.get_path("stdlib"), sysconfig.get_path("platstdlib")}
+            if p
+        }:
+            logger.info(
+                "Not uninstalling %s at %s, as it is in the standard library.",
+                dist.canonical_name,
+                normalized_dist_location,
+            )
+            return cls(dist)
+
+        paths_to_remove = cls(dist)
+        develop_egg_link = egg_link_path_from_location(dist.raw_name)
+
+        # Distribution is installed with metadata in a "flat" .egg-info
+        # directory. This means it is not a modern .dist-info installation, an
+        # egg, or legacy editable.
+        setuptools_flat_installation = (
+            dist.installed_with_setuptools_egg_info
+            and info_location is not None
+            and os.path.exists(info_location)
+            # If dist is editable and the location points to a ``.egg-info``,
+            # we are in fact in the legacy editable case.
+            and not info_location.endswith(f"{dist.setuptools_filename}.egg-info")
+        )
+
+        # Uninstall cases order do matter as in the case of 2 installs of the
+        # same package, pip needs to uninstall the currently detected version
+        if setuptools_flat_installation:
+            if info_location is not None:
+                paths_to_remove.add(info_location)
+            installed_files = dist.iter_declared_entries()
+            if installed_files is not None:
+                for installed_file in installed_files:
+                    paths_to_remove.add(os.path.join(dist_location, installed_file))
+            # FIXME: need a test for this elif block
+            # occurs with --single-version-externally-managed/--record outside
+            # of pip
+            elif dist.is_file("top_level.txt"):
+                try:
+                    namespace_packages = dist.read_text("namespace_packages.txt")
+                except FileNotFoundError:
+                    namespaces = []
+                else:
+                    namespaces = namespace_packages.splitlines(keepends=False)
+                for top_level_pkg in [
+                    p
+                    for p in dist.read_text("top_level.txt").splitlines()
+                    if p and p not in namespaces
+                ]:
+                    path = os.path.join(dist_location, top_level_pkg)
+                    paths_to_remove.add(path)
+                    paths_to_remove.add(f"{path}.py")
+                    paths_to_remove.add(f"{path}.pyc")
+                    paths_to_remove.add(f"{path}.pyo")
+
+        elif dist.installed_by_distutils:
+            raise LegacyDistutilsInstall(distribution=dist)
+
+        elif dist.installed_as_egg:
+            # package installed by easy_install
+            # We cannot match on dist.egg_name because it can slightly vary
+            # i.e. setuptools-0.6c11-py2.6.egg vs setuptools-0.6rc11-py2.6.egg
+            paths_to_remove.add(dist_location)
+            easy_install_egg = os.path.split(dist_location)[1]
+            easy_install_pth = os.path.join(
+                os.path.dirname(dist_location),
+                "easy-install.pth",
+            )
+            paths_to_remove.add_pth(easy_install_pth, "./" + easy_install_egg)
+
+        elif dist.installed_with_dist_info:
+            for path in uninstallation_paths(dist):
+                paths_to_remove.add(path)
+
+        elif develop_egg_link:
+            # PEP 660 modern editable is handled in the ``.dist-info`` case
+            # above, so this only covers the setuptools-style editable.
+            with open(develop_egg_link) as fh:
+                link_pointer = os.path.normcase(fh.readline().strip())
+                normalized_link_pointer = paths_to_remove._normalize_path_cached(
+                    link_pointer
+                )
+            assert os.path.samefile(
+                normalized_link_pointer, normalized_dist_location
+            ), (
+                f"Egg-link {develop_egg_link} (to {link_pointer}) does not match "
+                f"installed location of {dist.raw_name} (at {dist_location})"
+            )
+            paths_to_remove.add(develop_egg_link)
+            easy_install_pth = os.path.join(
+                os.path.dirname(develop_egg_link), "easy-install.pth"
+            )
+            paths_to_remove.add_pth(easy_install_pth, dist_location)
+
+        else:
+            logger.debug(
+                "Not sure how to uninstall: %s - Check: %s",
+                dist,
+                dist_location,
+            )
+
+        if dist.in_usersite:
+            bin_dir = get_bin_user()
+        else:
+            bin_dir = get_bin_prefix()
+
+        # find distutils scripts= scripts
+        try:
+            for script in dist.iter_distutils_script_names():
+                paths_to_remove.add(os.path.join(bin_dir, script))
+                if WINDOWS:
+                    paths_to_remove.add(os.path.join(bin_dir, f"{script}.bat"))
+        except (FileNotFoundError, NotADirectoryError):
+            pass
+
+        # find console_scripts and gui_scripts
+        def iter_scripts_to_remove(
+            dist: BaseDistribution,
+            bin_dir: str,
+        ) -> Generator[str, None, None]:
+            for entry_point in dist.iter_entry_points():
+                if entry_point.group == "console_scripts":
+                    yield from _script_names(bin_dir, entry_point.name, False)
+                elif entry_point.group == "gui_scripts":
+                    yield from _script_names(bin_dir, entry_point.name, True)
+
+        for s in iter_scripts_to_remove(dist, bin_dir):
+            paths_to_remove.add(s)
+
+        return paths_to_remove
+
+
+class UninstallPthEntries:
+    def __init__(self, pth_file: str) -> None:
+        self.file = pth_file
+        self.entries: Set[str] = set()
+        self._saved_lines: Optional[List[bytes]] = None
+
+    def add(self, entry: str) -> None:
+        entry = os.path.normcase(entry)
+        # On Windows, os.path.normcase converts the entry to use
+        # backslashes.  This is correct for entries that describe absolute
+        # paths outside of site-packages, but all the others use forward
+        # slashes.
+        # os.path.splitdrive is used instead of os.path.isabs because isabs
+        # treats non-absolute paths with drive letter markings like c:foo\bar
+        # as absolute paths. It also does not recognize UNC paths if they don't
+        # have more than "\\sever\share". Valid examples: "\\server\share\" or
+        # "\\server\share\folder".
+        if WINDOWS and not os.path.splitdrive(entry)[0]:
+            entry = entry.replace("\\", "/")
+        self.entries.add(entry)
+
+    def remove(self) -> None:
+        logger.verbose("Removing pth entries from %s:", self.file)
+
+        # If the file doesn't exist, log a warning and return
+        if not os.path.isfile(self.file):
+            logger.warning("Cannot remove entries from nonexistent file %s", self.file)
+            return
+        with open(self.file, "rb") as fh:
+            # windows uses '\r\n' with py3k, but uses '\n' with py2.x
+            lines = fh.readlines()
+            self._saved_lines = lines
+        if any(b"\r\n" in line for line in lines):
+            endline = "\r\n"
+        else:
+            endline = "\n"
+        # handle missing trailing newline
+        if lines and not lines[-1].endswith(endline.encode("utf-8")):
+            lines[-1] = lines[-1] + endline.encode("utf-8")
+        for entry in self.entries:
+            try:
+                logger.verbose("Removing entry: %s", entry)
+                lines.remove((entry + endline).encode("utf-8"))
+            except ValueError:
+                pass
+        with open(self.file, "wb") as fh:
+            fh.writelines(lines)
+
+    def rollback(self) -> bool:
+        if self._saved_lines is None:
+            logger.error("Cannot roll back changes to %s, none were made", self.file)
+            return False
+        logger.debug("Rolling %s back to previous state", self.file)
+        with open(self.file, "wb") as fh:
+            fh.writelines(self._saved_lines)
+        return True
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ab0c3f1c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..86d3cd6d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/base.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/base.py
new file mode 100644
index 00000000..42dade18
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/base.py
@@ -0,0 +1,20 @@
+from typing import Callable, List, Optional
+
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.req.req_set import RequirementSet
+
+InstallRequirementProvider = Callable[
+    [str, Optional[InstallRequirement]], InstallRequirement
+]
+
+
+class BaseResolver:
+    def resolve(
+        self, root_reqs: List[InstallRequirement], check_supported_wheels: bool
+    ) -> RequirementSet:
+        raise NotImplementedError()
+
+    def get_installation_order(
+        self, req_set: RequirementSet
+    ) -> List[InstallRequirement]:
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b2b70f6c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/resolver.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/resolver.cpython-312.pyc
new file mode 100644
index 00000000..be856e21
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/__pycache__/resolver.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/resolver.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/resolver.py
new file mode 100644
index 00000000..1dd0d704
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/legacy/resolver.py
@@ -0,0 +1,597 @@
+"""Dependency Resolution
+
+The dependency resolution in pip is performed as follows:
+
+for top-level requirements:
+    a. only one spec allowed per project, regardless of conflicts or not.
+       otherwise a "double requirement" exception is raised
+    b. they override sub-dependency requirements.
+for sub-dependencies
+    a. "first found, wins" (where the order is breadth first)
+"""
+
+import logging
+import sys
+from collections import defaultdict
+from itertools import chain
+from typing import DefaultDict, Iterable, List, Optional, Set, Tuple
+
+from pip._vendor.packaging import specifiers
+from pip._vendor.packaging.requirements import Requirement
+
+from pip._internal.cache import WheelCache
+from pip._internal.exceptions import (
+    BestVersionAlreadyInstalled,
+    DistributionNotFound,
+    HashError,
+    HashErrors,
+    InstallationError,
+    NoneMetadataError,
+    UnsupportedPythonVersion,
+)
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.metadata import BaseDistribution
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.operations.prepare import RequirementPreparer
+from pip._internal.req.req_install import (
+    InstallRequirement,
+    check_invalid_constraint_type,
+)
+from pip._internal.req.req_set import RequirementSet
+from pip._internal.resolution.base import BaseResolver, InstallRequirementProvider
+from pip._internal.utils import compatibility_tags
+from pip._internal.utils.compatibility_tags import get_supported
+from pip._internal.utils.direct_url_helpers import direct_url_from_link
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import normalize_version_info
+from pip._internal.utils.packaging import check_requires_python
+
+logger = logging.getLogger(__name__)
+
+DiscoveredDependencies = DefaultDict[Optional[str], List[InstallRequirement]]
+
+
+def _check_dist_requires_python(
+    dist: BaseDistribution,
+    version_info: Tuple[int, int, int],
+    ignore_requires_python: bool = False,
+) -> None:
+    """
+    Check whether the given Python version is compatible with a distribution's
+    "Requires-Python" value.
+
+    :param version_info: A 3-tuple of ints representing the Python
+        major-minor-micro version to check.
+    :param ignore_requires_python: Whether to ignore the "Requires-Python"
+        value if the given Python version isn't compatible.
+
+    :raises UnsupportedPythonVersion: When the given Python version isn't
+        compatible.
+    """
+    # This idiosyncratically converts the SpecifierSet to str and let
+    # check_requires_python then parse it again into SpecifierSet. But this
+    # is the legacy resolver so I'm just not going to bother refactoring.
+    try:
+        requires_python = str(dist.requires_python)
+    except FileNotFoundError as e:
+        raise NoneMetadataError(dist, str(e))
+    try:
+        is_compatible = check_requires_python(
+            requires_python,
+            version_info=version_info,
+        )
+    except specifiers.InvalidSpecifier as exc:
+        logger.warning(
+            "Package %r has an invalid Requires-Python: %s", dist.raw_name, exc
+        )
+        return
+
+    if is_compatible:
+        return
+
+    version = ".".join(map(str, version_info))
+    if ignore_requires_python:
+        logger.debug(
+            "Ignoring failed Requires-Python check for package %r: %s not in %r",
+            dist.raw_name,
+            version,
+            requires_python,
+        )
+        return
+
+    raise UnsupportedPythonVersion(
+        f"Package {dist.raw_name!r} requires a different Python: "
+        f"{version} not in {requires_python!r}"
+    )
+
+
+class Resolver(BaseResolver):
+    """Resolves which packages need to be installed/uninstalled to perform \
+    the requested operation without breaking the requirements of any package.
+    """
+
+    _allowed_strategies = {"eager", "only-if-needed", "to-satisfy-only"}
+
+    def __init__(
+        self,
+        preparer: RequirementPreparer,
+        finder: PackageFinder,
+        wheel_cache: Optional[WheelCache],
+        make_install_req: InstallRequirementProvider,
+        use_user_site: bool,
+        ignore_dependencies: bool,
+        ignore_installed: bool,
+        ignore_requires_python: bool,
+        force_reinstall: bool,
+        upgrade_strategy: str,
+        py_version_info: Optional[Tuple[int, ...]] = None,
+    ) -> None:
+        super().__init__()
+        assert upgrade_strategy in self._allowed_strategies
+
+        if py_version_info is None:
+            py_version_info = sys.version_info[:3]
+        else:
+            py_version_info = normalize_version_info(py_version_info)
+
+        self._py_version_info = py_version_info
+
+        self.preparer = preparer
+        self.finder = finder
+        self.wheel_cache = wheel_cache
+
+        self.upgrade_strategy = upgrade_strategy
+        self.force_reinstall = force_reinstall
+        self.ignore_dependencies = ignore_dependencies
+        self.ignore_installed = ignore_installed
+        self.ignore_requires_python = ignore_requires_python
+        self.use_user_site = use_user_site
+        self._make_install_req = make_install_req
+
+        self._discovered_dependencies: DiscoveredDependencies = defaultdict(list)
+
+    def resolve(
+        self, root_reqs: List[InstallRequirement], check_supported_wheels: bool
+    ) -> RequirementSet:
+        """Resolve what operations need to be done
+
+        As a side-effect of this method, the packages (and their dependencies)
+        are downloaded, unpacked and prepared for installation. This
+        preparation is done by ``pip.operations.prepare``.
+
+        Once PyPI has static dependency metadata available, it would be
+        possible to move the preparation to become a step separated from
+        dependency resolution.
+        """
+        requirement_set = RequirementSet(check_supported_wheels=check_supported_wheels)
+        for req in root_reqs:
+            if req.constraint:
+                check_invalid_constraint_type(req)
+            self._add_requirement_to_set(requirement_set, req)
+
+        # Actually prepare the files, and collect any exceptions. Most hash
+        # exceptions cannot be checked ahead of time, because
+        # _populate_link() needs to be called before we can make decisions
+        # based on link type.
+        discovered_reqs: List[InstallRequirement] = []
+        hash_errors = HashErrors()
+        for req in chain(requirement_set.all_requirements, discovered_reqs):
+            try:
+                discovered_reqs.extend(self._resolve_one(requirement_set, req))
+            except HashError as exc:
+                exc.req = req
+                hash_errors.append(exc)
+
+        if hash_errors:
+            raise hash_errors
+
+        return requirement_set
+
+    def _add_requirement_to_set(
+        self,
+        requirement_set: RequirementSet,
+        install_req: InstallRequirement,
+        parent_req_name: Optional[str] = None,
+        extras_requested: Optional[Iterable[str]] = None,
+    ) -> Tuple[List[InstallRequirement], Optional[InstallRequirement]]:
+        """Add install_req as a requirement to install.
+
+        :param parent_req_name: The name of the requirement that needed this
+            added. The name is used because when multiple unnamed requirements
+            resolve to the same name, we could otherwise end up with dependency
+            links that point outside the Requirements set. parent_req must
+            already be added. Note that None implies that this is a user
+            supplied requirement, vs an inferred one.
+        :param extras_requested: an iterable of extras used to evaluate the
+            environment markers.
+        :return: Additional requirements to scan. That is either [] if
+            the requirement is not applicable, or [install_req] if the
+            requirement is applicable and has just been added.
+        """
+        # If the markers do not match, ignore this requirement.
+        if not install_req.match_markers(extras_requested):
+            logger.info(
+                "Ignoring %s: markers '%s' don't match your environment",
+                install_req.name,
+                install_req.markers,
+            )
+            return [], None
+
+        # If the wheel is not supported, raise an error.
+        # Should check this after filtering out based on environment markers to
+        # allow specifying different wheels based on the environment/OS, in a
+        # single requirements file.
+        if install_req.link and install_req.link.is_wheel:
+            wheel = Wheel(install_req.link.filename)
+            tags = compatibility_tags.get_supported()
+            if requirement_set.check_supported_wheels and not wheel.supported(tags):
+                raise InstallationError(
+                    f"{wheel.filename} is not a supported wheel on this platform."
+                )
+
+        # This next bit is really a sanity check.
+        assert (
+            not install_req.user_supplied or parent_req_name is None
+        ), "a user supplied req shouldn't have a parent"
+
+        # Unnamed requirements are scanned again and the requirement won't be
+        # added as a dependency until after scanning.
+        if not install_req.name:
+            requirement_set.add_unnamed_requirement(install_req)
+            return [install_req], None
+
+        try:
+            existing_req: Optional[InstallRequirement] = (
+                requirement_set.get_requirement(install_req.name)
+            )
+        except KeyError:
+            existing_req = None
+
+        has_conflicting_requirement = (
+            parent_req_name is None
+            and existing_req
+            and not existing_req.constraint
+            and existing_req.extras == install_req.extras
+            and existing_req.req
+            and install_req.req
+            and existing_req.req.specifier != install_req.req.specifier
+        )
+        if has_conflicting_requirement:
+            raise InstallationError(
+                f"Double requirement given: {install_req} "
+                f"(already in {existing_req}, name={install_req.name!r})"
+            )
+
+        # When no existing requirement exists, add the requirement as a
+        # dependency and it will be scanned again after.
+        if not existing_req:
+            requirement_set.add_named_requirement(install_req)
+            # We'd want to rescan this requirement later
+            return [install_req], install_req
+
+        # Assume there's no need to scan, and that we've already
+        # encountered this for scanning.
+        if install_req.constraint or not existing_req.constraint:
+            return [], existing_req
+
+        does_not_satisfy_constraint = install_req.link and not (
+            existing_req.link and install_req.link.path == existing_req.link.path
+        )
+        if does_not_satisfy_constraint:
+            raise InstallationError(
+                f"Could not satisfy constraints for '{install_req.name}': "
+                "installation from path or url cannot be "
+                "constrained to a version"
+            )
+        # If we're now installing a constraint, mark the existing
+        # object for real installation.
+        existing_req.constraint = False
+        # If we're now installing a user supplied requirement,
+        # mark the existing object as such.
+        if install_req.user_supplied:
+            existing_req.user_supplied = True
+        existing_req.extras = tuple(
+            sorted(set(existing_req.extras) | set(install_req.extras))
+        )
+        logger.debug(
+            "Setting %s extras to: %s",
+            existing_req,
+            existing_req.extras,
+        )
+        # Return the existing requirement for addition to the parent and
+        # scanning again.
+        return [existing_req], existing_req
+
+    def _is_upgrade_allowed(self, req: InstallRequirement) -> bool:
+        if self.upgrade_strategy == "to-satisfy-only":
+            return False
+        elif self.upgrade_strategy == "eager":
+            return True
+        else:
+            assert self.upgrade_strategy == "only-if-needed"
+            return req.user_supplied or req.constraint
+
+    def _set_req_to_reinstall(self, req: InstallRequirement) -> None:
+        """
+        Set a requirement to be installed.
+        """
+        # Don't uninstall the conflict if doing a user install and the
+        # conflict is not a user install.
+        assert req.satisfied_by is not None
+        if not self.use_user_site or req.satisfied_by.in_usersite:
+            req.should_reinstall = True
+        req.satisfied_by = None
+
+    def _check_skip_installed(
+        self, req_to_install: InstallRequirement
+    ) -> Optional[str]:
+        """Check if req_to_install should be skipped.
+
+        This will check if the req is installed, and whether we should upgrade
+        or reinstall it, taking into account all the relevant user options.
+
+        After calling this req_to_install will only have satisfied_by set to
+        None if the req_to_install is to be upgraded/reinstalled etc. Any
+        other value will be a dist recording the current thing installed that
+        satisfies the requirement.
+
+        Note that for vcs urls and the like we can't assess skipping in this
+        routine - we simply identify that we need to pull the thing down,
+        then later on it is pulled down and introspected to assess upgrade/
+        reinstalls etc.
+
+        :return: A text reason for why it was skipped, or None.
+        """
+        if self.ignore_installed:
+            return None
+
+        req_to_install.check_if_exists(self.use_user_site)
+        if not req_to_install.satisfied_by:
+            return None
+
+        if self.force_reinstall:
+            self._set_req_to_reinstall(req_to_install)
+            return None
+
+        if not self._is_upgrade_allowed(req_to_install):
+            if self.upgrade_strategy == "only-if-needed":
+                return "already satisfied, skipping upgrade"
+            return "already satisfied"
+
+        # Check for the possibility of an upgrade.  For link-based
+        # requirements we have to pull the tree down and inspect to assess
+        # the version #, so it's handled way down.
+        if not req_to_install.link:
+            try:
+                self.finder.find_requirement(req_to_install, upgrade=True)
+            except BestVersionAlreadyInstalled:
+                # Then the best version is installed.
+                return "already up-to-date"
+            except DistributionNotFound:
+                # No distribution found, so we squash the error.  It will
+                # be raised later when we re-try later to do the install.
+                # Why don't we just raise here?
+                pass
+
+        self._set_req_to_reinstall(req_to_install)
+        return None
+
+    def _find_requirement_link(self, req: InstallRequirement) -> Optional[Link]:
+        upgrade = self._is_upgrade_allowed(req)
+        best_candidate = self.finder.find_requirement(req, upgrade)
+        if not best_candidate:
+            return None
+
+        # Log a warning per PEP 592 if necessary before returning.
+        link = best_candidate.link
+        if link.is_yanked:
+            reason = link.yanked_reason or "<none given>"
+            msg = (
+                # Mark this as a unicode string to prevent
+                # "UnicodeEncodeError: 'ascii' codec can't encode character"
+                # in Python 2 when the reason contains non-ascii characters.
+                "The candidate selected for download or install is a "
+                f"yanked version: {best_candidate}\n"
+                f"Reason for being yanked: {reason}"
+            )
+            logger.warning(msg)
+
+        return link
+
+    def _populate_link(self, req: InstallRequirement) -> None:
+        """Ensure that if a link can be found for this, that it is found.
+
+        Note that req.link may still be None - if the requirement is already
+        installed and not needed to be upgraded based on the return value of
+        _is_upgrade_allowed().
+
+        If preparer.require_hashes is True, don't use the wheel cache, because
+        cached wheels, always built locally, have different hashes than the
+        files downloaded from the index server and thus throw false hash
+        mismatches. Furthermore, cached wheels at present have undeterministic
+        contents due to file modification times.
+        """
+        if req.link is None:
+            req.link = self._find_requirement_link(req)
+
+        if self.wheel_cache is None or self.preparer.require_hashes:
+            return
+
+        assert req.link is not None, "_find_requirement_link unexpectedly returned None"
+        cache_entry = self.wheel_cache.get_cache_entry(
+            link=req.link,
+            package_name=req.name,
+            supported_tags=get_supported(),
+        )
+        if cache_entry is not None:
+            logger.debug("Using cached wheel link: %s", cache_entry.link)
+            if req.link is req.original_link and cache_entry.persistent:
+                req.cached_wheel_source_link = req.link
+            if cache_entry.origin is not None:
+                req.download_info = cache_entry.origin
+            else:
+                # Legacy cache entry that does not have origin.json.
+                # download_info may miss the archive_info.hashes field.
+                req.download_info = direct_url_from_link(
+                    req.link, link_is_in_wheel_cache=cache_entry.persistent
+                )
+            req.link = cache_entry.link
+
+    def _get_dist_for(self, req: InstallRequirement) -> BaseDistribution:
+        """Takes a InstallRequirement and returns a single AbstractDist \
+        representing a prepared variant of the same.
+        """
+        if req.editable:
+            return self.preparer.prepare_editable_requirement(req)
+
+        # satisfied_by is only evaluated by calling _check_skip_installed,
+        # so it must be None here.
+        assert req.satisfied_by is None
+        skip_reason = self._check_skip_installed(req)
+
+        if req.satisfied_by:
+            return self.preparer.prepare_installed_requirement(req, skip_reason)
+
+        # We eagerly populate the link, since that's our "legacy" behavior.
+        self._populate_link(req)
+        dist = self.preparer.prepare_linked_requirement(req)
+
+        # NOTE
+        # The following portion is for determining if a certain package is
+        # going to be re-installed/upgraded or not and reporting to the user.
+        # This should probably get cleaned up in a future refactor.
+
+        # req.req is only avail after unpack for URL
+        # pkgs repeat check_if_exists to uninstall-on-upgrade
+        # (#14)
+        if not self.ignore_installed:
+            req.check_if_exists(self.use_user_site)
+
+        if req.satisfied_by:
+            should_modify = (
+                self.upgrade_strategy != "to-satisfy-only"
+                or self.force_reinstall
+                or self.ignore_installed
+                or req.link.scheme == "file"
+            )
+            if should_modify:
+                self._set_req_to_reinstall(req)
+            else:
+                logger.info(
+                    "Requirement already satisfied (use --upgrade to upgrade): %s",
+                    req,
+                )
+        return dist
+
+    def _resolve_one(
+        self,
+        requirement_set: RequirementSet,
+        req_to_install: InstallRequirement,
+    ) -> List[InstallRequirement]:
+        """Prepare a single requirements file.
+
+        :return: A list of additional InstallRequirements to also install.
+        """
+        # Tell user what we are doing for this requirement:
+        # obtain (editable), skipping, processing (local url), collecting
+        # (remote url or package name)
+        if req_to_install.constraint or req_to_install.prepared:
+            return []
+
+        req_to_install.prepared = True
+
+        # Parse and return dependencies
+        dist = self._get_dist_for(req_to_install)
+        # This will raise UnsupportedPythonVersion if the given Python
+        # version isn't compatible with the distribution's Requires-Python.
+        _check_dist_requires_python(
+            dist,
+            version_info=self._py_version_info,
+            ignore_requires_python=self.ignore_requires_python,
+        )
+
+        more_reqs: List[InstallRequirement] = []
+
+        def add_req(subreq: Requirement, extras_requested: Iterable[str]) -> None:
+            # This idiosyncratically converts the Requirement to str and let
+            # make_install_req then parse it again into Requirement. But this is
+            # the legacy resolver so I'm just not going to bother refactoring.
+            sub_install_req = self._make_install_req(str(subreq), req_to_install)
+            parent_req_name = req_to_install.name
+            to_scan_again, add_to_parent = self._add_requirement_to_set(
+                requirement_set,
+                sub_install_req,
+                parent_req_name=parent_req_name,
+                extras_requested=extras_requested,
+            )
+            if parent_req_name and add_to_parent:
+                self._discovered_dependencies[parent_req_name].append(add_to_parent)
+            more_reqs.extend(to_scan_again)
+
+        with indent_log():
+            # We add req_to_install before its dependencies, so that we
+            # can refer to it when adding dependencies.
+            assert req_to_install.name is not None
+            if not requirement_set.has_requirement(req_to_install.name):
+                # 'unnamed' requirements will get added here
+                # 'unnamed' requirements can only come from being directly
+                # provided by the user.
+                assert req_to_install.user_supplied
+                self._add_requirement_to_set(
+                    requirement_set, req_to_install, parent_req_name=None
+                )
+
+            if not self.ignore_dependencies:
+                if req_to_install.extras:
+                    logger.debug(
+                        "Installing extra requirements: %r",
+                        ",".join(req_to_install.extras),
+                    )
+                missing_requested = sorted(
+                    set(req_to_install.extras) - set(dist.iter_provided_extras())
+                )
+                for missing in missing_requested:
+                    logger.warning(
+                        "%s %s does not provide the extra '%s'",
+                        dist.raw_name,
+                        dist.version,
+                        missing,
+                    )
+
+                available_requested = sorted(
+                    set(dist.iter_provided_extras()) & set(req_to_install.extras)
+                )
+                for subreq in dist.iter_dependencies(available_requested):
+                    add_req(subreq, extras_requested=available_requested)
+
+        return more_reqs
+
+    def get_installation_order(
+        self, req_set: RequirementSet
+    ) -> List[InstallRequirement]:
+        """Create the installation order.
+
+        The installation order is topological - requirements are installed
+        before the requiring thing. We break cycles at an arbitrary point,
+        and make no other guarantees.
+        """
+        # The current implementation, which we may change at any point
+        # installs the user specified things in the order given, except when
+        # dependencies must come earlier to achieve topological order.
+        order = []
+        ordered_reqs: Set[InstallRequirement] = set()
+
+        def schedule(req: InstallRequirement) -> None:
+            if req.satisfied_by or req in ordered_reqs:
+                return
+            if req.constraint:
+                return
+            ordered_reqs.add(req)
+            for dep in self._discovered_dependencies[req.name]:
+                schedule(dep)
+            order.append(req)
+
+        for install_req in req_set.requirements.values():
+            schedule(install_req)
+        return order
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ee558fc5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..d9657666
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/candidates.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/candidates.cpython-312.pyc
new file mode 100644
index 00000000..75908906
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/candidates.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/factory.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/factory.cpython-312.pyc
new file mode 100644
index 00000000..4e1c47c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/factory.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/found_candidates.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/found_candidates.cpython-312.pyc
new file mode 100644
index 00000000..d937729a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/found_candidates.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/provider.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/provider.cpython-312.pyc
new file mode 100644
index 00000000..100cbc4c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/provider.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/reporter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/reporter.cpython-312.pyc
new file mode 100644
index 00000000..7d2b206c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/reporter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..733c756f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/resolver.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/resolver.cpython-312.pyc
new file mode 100644
index 00000000..4e391c95
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/__pycache__/resolver.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/base.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/base.py
new file mode 100644
index 00000000..0f31dc9b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/base.py
@@ -0,0 +1,139 @@
+from dataclasses import dataclass
+from typing import FrozenSet, Iterable, Optional, Tuple
+
+from pip._vendor.packaging.specifiers import SpecifierSet
+from pip._vendor.packaging.utils import NormalizedName
+from pip._vendor.packaging.version import Version
+
+from pip._internal.models.link import Link, links_equivalent
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.utils.hashes import Hashes
+
+CandidateLookup = Tuple[Optional["Candidate"], Optional[InstallRequirement]]
+
+
+def format_name(project: NormalizedName, extras: FrozenSet[NormalizedName]) -> str:
+    if not extras:
+        return project
+    extras_expr = ",".join(sorted(extras))
+    return f"{project}[{extras_expr}]"
+
+
+@dataclass(frozen=True)
+class Constraint:
+    specifier: SpecifierSet
+    hashes: Hashes
+    links: FrozenSet[Link]
+
+    @classmethod
+    def empty(cls) -> "Constraint":
+        return Constraint(SpecifierSet(), Hashes(), frozenset())
+
+    @classmethod
+    def from_ireq(cls, ireq: InstallRequirement) -> "Constraint":
+        links = frozenset([ireq.link]) if ireq.link else frozenset()
+        return Constraint(ireq.specifier, ireq.hashes(trust_internet=False), links)
+
+    def __bool__(self) -> bool:
+        return bool(self.specifier) or bool(self.hashes) or bool(self.links)
+
+    def __and__(self, other: InstallRequirement) -> "Constraint":
+        if not isinstance(other, InstallRequirement):
+            return NotImplemented
+        specifier = self.specifier & other.specifier
+        hashes = self.hashes & other.hashes(trust_internet=False)
+        links = self.links
+        if other.link:
+            links = links.union([other.link])
+        return Constraint(specifier, hashes, links)
+
+    def is_satisfied_by(self, candidate: "Candidate") -> bool:
+        # Reject if there are any mismatched URL constraints on this package.
+        if self.links and not all(_match_link(link, candidate) for link in self.links):
+            return False
+        # We can safely always allow prereleases here since PackageFinder
+        # already implements the prerelease logic, and would have filtered out
+        # prerelease candidates if the user does not expect them.
+        return self.specifier.contains(candidate.version, prereleases=True)
+
+
+class Requirement:
+    @property
+    def project_name(self) -> NormalizedName:
+        """The "project name" of a requirement.
+
+        This is different from ``name`` if this requirement contains extras,
+        in which case ``name`` would contain the ``[...]`` part, while this
+        refers to the name of the project.
+        """
+        raise NotImplementedError("Subclass should override")
+
+    @property
+    def name(self) -> str:
+        """The name identifying this requirement in the resolver.
+
+        This is different from ``project_name`` if this requirement contains
+        extras, where ``project_name`` would not contain the ``[...]`` part.
+        """
+        raise NotImplementedError("Subclass should override")
+
+    def is_satisfied_by(self, candidate: "Candidate") -> bool:
+        return False
+
+    def get_candidate_lookup(self) -> CandidateLookup:
+        raise NotImplementedError("Subclass should override")
+
+    def format_for_error(self) -> str:
+        raise NotImplementedError("Subclass should override")
+
+
+def _match_link(link: Link, candidate: "Candidate") -> bool:
+    if candidate.source_link:
+        return links_equivalent(link, candidate.source_link)
+    return False
+
+
+class Candidate:
+    @property
+    def project_name(self) -> NormalizedName:
+        """The "project name" of the candidate.
+
+        This is different from ``name`` if this candidate contains extras,
+        in which case ``name`` would contain the ``[...]`` part, while this
+        refers to the name of the project.
+        """
+        raise NotImplementedError("Override in subclass")
+
+    @property
+    def name(self) -> str:
+        """The name identifying this candidate in the resolver.
+
+        This is different from ``project_name`` if this candidate contains
+        extras, where ``project_name`` would not contain the ``[...]`` part.
+        """
+        raise NotImplementedError("Override in subclass")
+
+    @property
+    def version(self) -> Version:
+        raise NotImplementedError("Override in subclass")
+
+    @property
+    def is_installed(self) -> bool:
+        raise NotImplementedError("Override in subclass")
+
+    @property
+    def is_editable(self) -> bool:
+        raise NotImplementedError("Override in subclass")
+
+    @property
+    def source_link(self) -> Optional[Link]:
+        raise NotImplementedError("Override in subclass")
+
+    def iter_dependencies(self, with_requires: bool) -> Iterable[Optional[Requirement]]:
+        raise NotImplementedError("Override in subclass")
+
+    def get_install_requirement(self) -> Optional[InstallRequirement]:
+        raise NotImplementedError("Override in subclass")
+
+    def format_for_error(self) -> str:
+        raise NotImplementedError("Subclass should override")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/candidates.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/candidates.py
new file mode 100644
index 00000000..6617644f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/candidates.py
@@ -0,0 +1,574 @@
+import logging
+import sys
+from typing import TYPE_CHECKING, Any, FrozenSet, Iterable, Optional, Tuple, Union, cast
+
+from pip._vendor.packaging.requirements import InvalidRequirement
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import Version
+
+from pip._internal.exceptions import (
+    HashError,
+    InstallationSubprocessError,
+    InvalidInstalledPackage,
+    MetadataInconsistent,
+    MetadataInvalid,
+)
+from pip._internal.metadata import BaseDistribution
+from pip._internal.models.link import Link, links_equivalent
+from pip._internal.models.wheel import Wheel
+from pip._internal.req.constructors import (
+    install_req_from_editable,
+    install_req_from_line,
+)
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.utils.direct_url_helpers import direct_url_from_link
+from pip._internal.utils.misc import normalize_version_info
+
+from .base import Candidate, Requirement, format_name
+
+if TYPE_CHECKING:
+    from .factory import Factory
+
+logger = logging.getLogger(__name__)
+
+BaseCandidate = Union[
+    "AlreadyInstalledCandidate",
+    "EditableCandidate",
+    "LinkCandidate",
+]
+
+# Avoid conflicting with the PyPI package "Python".
+REQUIRES_PYTHON_IDENTIFIER = cast(NormalizedName, "<Python from Requires-Python>")
+
+
+def as_base_candidate(candidate: Candidate) -> Optional[BaseCandidate]:
+    """The runtime version of BaseCandidate."""
+    base_candidate_classes = (
+        AlreadyInstalledCandidate,
+        EditableCandidate,
+        LinkCandidate,
+    )
+    if isinstance(candidate, base_candidate_classes):
+        return candidate
+    return None
+
+
+def make_install_req_from_link(
+    link: Link, template: InstallRequirement
+) -> InstallRequirement:
+    assert not template.editable, "template is editable"
+    if template.req:
+        line = str(template.req)
+    else:
+        line = link.url
+    ireq = install_req_from_line(
+        line,
+        user_supplied=template.user_supplied,
+        comes_from=template.comes_from,
+        use_pep517=template.use_pep517,
+        isolated=template.isolated,
+        constraint=template.constraint,
+        global_options=template.global_options,
+        hash_options=template.hash_options,
+        config_settings=template.config_settings,
+    )
+    ireq.original_link = template.original_link
+    ireq.link = link
+    ireq.extras = template.extras
+    return ireq
+
+
+def make_install_req_from_editable(
+    link: Link, template: InstallRequirement
+) -> InstallRequirement:
+    assert template.editable, "template not editable"
+    ireq = install_req_from_editable(
+        link.url,
+        user_supplied=template.user_supplied,
+        comes_from=template.comes_from,
+        use_pep517=template.use_pep517,
+        isolated=template.isolated,
+        constraint=template.constraint,
+        permit_editable_wheels=template.permit_editable_wheels,
+        global_options=template.global_options,
+        hash_options=template.hash_options,
+        config_settings=template.config_settings,
+    )
+    ireq.extras = template.extras
+    return ireq
+
+
+def _make_install_req_from_dist(
+    dist: BaseDistribution, template: InstallRequirement
+) -> InstallRequirement:
+    if template.req:
+        line = str(template.req)
+    elif template.link:
+        line = f"{dist.canonical_name} @ {template.link.url}"
+    else:
+        line = f"{dist.canonical_name}=={dist.version}"
+    ireq = install_req_from_line(
+        line,
+        user_supplied=template.user_supplied,
+        comes_from=template.comes_from,
+        use_pep517=template.use_pep517,
+        isolated=template.isolated,
+        constraint=template.constraint,
+        global_options=template.global_options,
+        hash_options=template.hash_options,
+        config_settings=template.config_settings,
+    )
+    ireq.satisfied_by = dist
+    return ireq
+
+
+class _InstallRequirementBackedCandidate(Candidate):
+    """A candidate backed by an ``InstallRequirement``.
+
+    This represents a package request with the target not being already
+    in the environment, and needs to be fetched and installed. The backing
+    ``InstallRequirement`` is responsible for most of the leg work; this
+    class exposes appropriate information to the resolver.
+
+    :param link: The link passed to the ``InstallRequirement``. The backing
+        ``InstallRequirement`` will use this link to fetch the distribution.
+    :param source_link: The link this candidate "originates" from. This is
+        different from ``link`` when the link is found in the wheel cache.
+        ``link`` would point to the wheel cache, while this points to the
+        found remote link (e.g. from pypi.org).
+    """
+
+    dist: BaseDistribution
+    is_installed = False
+
+    def __init__(
+        self,
+        link: Link,
+        source_link: Link,
+        ireq: InstallRequirement,
+        factory: "Factory",
+        name: Optional[NormalizedName] = None,
+        version: Optional[Version] = None,
+    ) -> None:
+        self._link = link
+        self._source_link = source_link
+        self._factory = factory
+        self._ireq = ireq
+        self._name = name
+        self._version = version
+        self.dist = self._prepare()
+        self._hash: Optional[int] = None
+
+    def __str__(self) -> str:
+        return f"{self.name} {self.version}"
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({str(self._link)!r})"
+
+    def __hash__(self) -> int:
+        if self._hash is not None:
+            return self._hash
+
+        self._hash = hash((self.__class__, self._link))
+        return self._hash
+
+    def __eq__(self, other: Any) -> bool:
+        if isinstance(other, self.__class__):
+            return links_equivalent(self._link, other._link)
+        return False
+
+    @property
+    def source_link(self) -> Optional[Link]:
+        return self._source_link
+
+    @property
+    def project_name(self) -> NormalizedName:
+        """The normalised name of the project the candidate refers to"""
+        if self._name is None:
+            self._name = self.dist.canonical_name
+        return self._name
+
+    @property
+    def name(self) -> str:
+        return self.project_name
+
+    @property
+    def version(self) -> Version:
+        if self._version is None:
+            self._version = self.dist.version
+        return self._version
+
+    def format_for_error(self) -> str:
+        return (
+            f"{self.name} {self.version} "
+            f"(from {self._link.file_path if self._link.is_file else self._link})"
+        )
+
+    def _prepare_distribution(self) -> BaseDistribution:
+        raise NotImplementedError("Override in subclass")
+
+    def _check_metadata_consistency(self, dist: BaseDistribution) -> None:
+        """Check for consistency of project name and version of dist."""
+        if self._name is not None and self._name != dist.canonical_name:
+            raise MetadataInconsistent(
+                self._ireq,
+                "name",
+                self._name,
+                dist.canonical_name,
+            )
+        if self._version is not None and self._version != dist.version:
+            raise MetadataInconsistent(
+                self._ireq,
+                "version",
+                str(self._version),
+                str(dist.version),
+            )
+        # check dependencies are valid
+        # TODO performance: this means we iterate the dependencies at least twice,
+        # we may want to cache parsed Requires-Dist
+        try:
+            list(dist.iter_dependencies(list(dist.iter_provided_extras())))
+        except InvalidRequirement as e:
+            raise MetadataInvalid(self._ireq, str(e))
+
+    def _prepare(self) -> BaseDistribution:
+        try:
+            dist = self._prepare_distribution()
+        except HashError as e:
+            # Provide HashError the underlying ireq that caused it. This
+            # provides context for the resulting error message to show the
+            # offending line to the user.
+            e.req = self._ireq
+            raise
+        except InstallationSubprocessError as exc:
+            # The output has been presented already, so don't duplicate it.
+            exc.context = "See above for output."
+            raise
+
+        self._check_metadata_consistency(dist)
+        return dist
+
+    def iter_dependencies(self, with_requires: bool) -> Iterable[Optional[Requirement]]:
+        requires = self.dist.iter_dependencies() if with_requires else ()
+        for r in requires:
+            yield from self._factory.make_requirements_from_spec(str(r), self._ireq)
+        yield self._factory.make_requires_python_requirement(self.dist.requires_python)
+
+    def get_install_requirement(self) -> Optional[InstallRequirement]:
+        return self._ireq
+
+
+class LinkCandidate(_InstallRequirementBackedCandidate):
+    is_editable = False
+
+    def __init__(
+        self,
+        link: Link,
+        template: InstallRequirement,
+        factory: "Factory",
+        name: Optional[NormalizedName] = None,
+        version: Optional[Version] = None,
+    ) -> None:
+        source_link = link
+        cache_entry = factory.get_wheel_cache_entry(source_link, name)
+        if cache_entry is not None:
+            logger.debug("Using cached wheel link: %s", cache_entry.link)
+            link = cache_entry.link
+        ireq = make_install_req_from_link(link, template)
+        assert ireq.link == link
+        if ireq.link.is_wheel and not ireq.link.is_file:
+            wheel = Wheel(ireq.link.filename)
+            wheel_name = canonicalize_name(wheel.name)
+            assert name == wheel_name, f"{name!r} != {wheel_name!r} for wheel"
+            # Version may not be present for PEP 508 direct URLs
+            if version is not None:
+                wheel_version = Version(wheel.version)
+                assert (
+                    version == wheel_version
+                ), f"{version!r} != {wheel_version!r} for wheel {name}"
+
+        if cache_entry is not None:
+            assert ireq.link.is_wheel
+            assert ireq.link.is_file
+            if cache_entry.persistent and template.link is template.original_link:
+                ireq.cached_wheel_source_link = source_link
+            if cache_entry.origin is not None:
+                ireq.download_info = cache_entry.origin
+            else:
+                # Legacy cache entry that does not have origin.json.
+                # download_info may miss the archive_info.hashes field.
+                ireq.download_info = direct_url_from_link(
+                    source_link, link_is_in_wheel_cache=cache_entry.persistent
+                )
+
+        super().__init__(
+            link=link,
+            source_link=source_link,
+            ireq=ireq,
+            factory=factory,
+            name=name,
+            version=version,
+        )
+
+    def _prepare_distribution(self) -> BaseDistribution:
+        preparer = self._factory.preparer
+        return preparer.prepare_linked_requirement(self._ireq, parallel_builds=True)
+
+
+class EditableCandidate(_InstallRequirementBackedCandidate):
+    is_editable = True
+
+    def __init__(
+        self,
+        link: Link,
+        template: InstallRequirement,
+        factory: "Factory",
+        name: Optional[NormalizedName] = None,
+        version: Optional[Version] = None,
+    ) -> None:
+        super().__init__(
+            link=link,
+            source_link=link,
+            ireq=make_install_req_from_editable(link, template),
+            factory=factory,
+            name=name,
+            version=version,
+        )
+
+    def _prepare_distribution(self) -> BaseDistribution:
+        return self._factory.preparer.prepare_editable_requirement(self._ireq)
+
+
+class AlreadyInstalledCandidate(Candidate):
+    is_installed = True
+    source_link = None
+
+    def __init__(
+        self,
+        dist: BaseDistribution,
+        template: InstallRequirement,
+        factory: "Factory",
+    ) -> None:
+        self.dist = dist
+        self._ireq = _make_install_req_from_dist(dist, template)
+        self._factory = factory
+        self._version = None
+
+        # This is just logging some messages, so we can do it eagerly.
+        # The returned dist would be exactly the same as self.dist because we
+        # set satisfied_by in _make_install_req_from_dist.
+        # TODO: Supply reason based on force_reinstall and upgrade_strategy.
+        skip_reason = "already satisfied"
+        factory.preparer.prepare_installed_requirement(self._ireq, skip_reason)
+
+    def __str__(self) -> str:
+        return str(self.dist)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.dist!r})"
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, AlreadyInstalledCandidate):
+            return NotImplemented
+        return self.name == other.name and self.version == other.version
+
+    def __hash__(self) -> int:
+        return hash((self.name, self.version))
+
+    @property
+    def project_name(self) -> NormalizedName:
+        return self.dist.canonical_name
+
+    @property
+    def name(self) -> str:
+        return self.project_name
+
+    @property
+    def version(self) -> Version:
+        if self._version is None:
+            self._version = self.dist.version
+        return self._version
+
+    @property
+    def is_editable(self) -> bool:
+        return self.dist.editable
+
+    def format_for_error(self) -> str:
+        return f"{self.name} {self.version} (Installed)"
+
+    def iter_dependencies(self, with_requires: bool) -> Iterable[Optional[Requirement]]:
+        if not with_requires:
+            return
+
+        try:
+            for r in self.dist.iter_dependencies():
+                yield from self._factory.make_requirements_from_spec(str(r), self._ireq)
+        except InvalidRequirement as exc:
+            raise InvalidInstalledPackage(dist=self.dist, invalid_exc=exc) from None
+
+    def get_install_requirement(self) -> Optional[InstallRequirement]:
+        return None
+
+
+class ExtrasCandidate(Candidate):
+    """A candidate that has 'extras', indicating additional dependencies.
+
+    Requirements can be for a project with dependencies, something like
+    foo[extra].  The extras don't affect the project/version being installed
+    directly, but indicate that we need additional dependencies. We model that
+    by having an artificial ExtrasCandidate that wraps the "base" candidate.
+
+    The ExtrasCandidate differs from the base in the following ways:
+
+    1. It has a unique name, of the form foo[extra]. This causes the resolver
+       to treat it as a separate node in the dependency graph.
+    2. When we're getting the candidate's dependencies,
+       a) We specify that we want the extra dependencies as well.
+       b) We add a dependency on the base candidate.
+          See below for why this is needed.
+    3. We return None for the underlying InstallRequirement, as the base
+       candidate will provide it, and we don't want to end up with duplicates.
+
+    The dependency on the base candidate is needed so that the resolver can't
+    decide that it should recommend foo[extra1] version 1.0 and foo[extra2]
+    version 2.0. Having those candidates depend on foo=1.0 and foo=2.0
+    respectively forces the resolver to recognise that this is a conflict.
+    """
+
+    def __init__(
+        self,
+        base: BaseCandidate,
+        extras: FrozenSet[str],
+        *,
+        comes_from: Optional[InstallRequirement] = None,
+    ) -> None:
+        """
+        :param comes_from: the InstallRequirement that led to this candidate if it
+            differs from the base's InstallRequirement. This will often be the
+            case in the sense that this candidate's requirement has the extras
+            while the base's does not. Unlike the InstallRequirement backed
+            candidates, this requirement is used solely for reporting purposes,
+            it does not do any leg work.
+        """
+        self.base = base
+        self.extras = frozenset(canonicalize_name(e) for e in extras)
+        self._comes_from = comes_from if comes_from is not None else self.base._ireq
+
+    def __str__(self) -> str:
+        name, rest = str(self.base).split(" ", 1)
+        return "{}[{}] {}".format(name, ",".join(self.extras), rest)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}(base={self.base!r}, extras={self.extras!r})"
+
+    def __hash__(self) -> int:
+        return hash((self.base, self.extras))
+
+    def __eq__(self, other: Any) -> bool:
+        if isinstance(other, self.__class__):
+            return self.base == other.base and self.extras == other.extras
+        return False
+
+    @property
+    def project_name(self) -> NormalizedName:
+        return self.base.project_name
+
+    @property
+    def name(self) -> str:
+        """The normalised name of the project the candidate refers to"""
+        return format_name(self.base.project_name, self.extras)
+
+    @property
+    def version(self) -> Version:
+        return self.base.version
+
+    def format_for_error(self) -> str:
+        return "{} [{}]".format(
+            self.base.format_for_error(), ", ".join(sorted(self.extras))
+        )
+
+    @property
+    def is_installed(self) -> bool:
+        return self.base.is_installed
+
+    @property
+    def is_editable(self) -> bool:
+        return self.base.is_editable
+
+    @property
+    def source_link(self) -> Optional[Link]:
+        return self.base.source_link
+
+    def iter_dependencies(self, with_requires: bool) -> Iterable[Optional[Requirement]]:
+        factory = self.base._factory
+
+        # Add a dependency on the exact base
+        # (See note 2b in the class docstring)
+        yield factory.make_requirement_from_candidate(self.base)
+        if not with_requires:
+            return
+
+        # The user may have specified extras that the candidate doesn't
+        # support. We ignore any unsupported extras here.
+        valid_extras = self.extras.intersection(self.base.dist.iter_provided_extras())
+        invalid_extras = self.extras.difference(self.base.dist.iter_provided_extras())
+        for extra in sorted(invalid_extras):
+            logger.warning(
+                "%s %s does not provide the extra '%s'",
+                self.base.name,
+                self.version,
+                extra,
+            )
+
+        for r in self.base.dist.iter_dependencies(valid_extras):
+            yield from factory.make_requirements_from_spec(
+                str(r),
+                self._comes_from,
+                valid_extras,
+            )
+
+    def get_install_requirement(self) -> Optional[InstallRequirement]:
+        # We don't return anything here, because we always
+        # depend on the base candidate, and we'll get the
+        # install requirement from that.
+        return None
+
+
+class RequiresPythonCandidate(Candidate):
+    is_installed = False
+    source_link = None
+
+    def __init__(self, py_version_info: Optional[Tuple[int, ...]]) -> None:
+        if py_version_info is not None:
+            version_info = normalize_version_info(py_version_info)
+        else:
+            version_info = sys.version_info[:3]
+        self._version = Version(".".join(str(c) for c in version_info))
+
+    # We don't need to implement __eq__() and __ne__() since there is always
+    # only one RequiresPythonCandidate in a resolution, i.e. the host Python.
+    # The built-in object.__eq__() and object.__ne__() do exactly what we want.
+
+    def __str__(self) -> str:
+        return f"Python {self._version}"
+
+    @property
+    def project_name(self) -> NormalizedName:
+        return REQUIRES_PYTHON_IDENTIFIER
+
+    @property
+    def name(self) -> str:
+        return REQUIRES_PYTHON_IDENTIFIER
+
+    @property
+    def version(self) -> Version:
+        return self._version
+
+    def format_for_error(self) -> str:
+        return f"Python {self.version}"
+
+    def iter_dependencies(self, with_requires: bool) -> Iterable[Optional[Requirement]]:
+        return ()
+
+    def get_install_requirement(self) -> Optional[InstallRequirement]:
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/factory.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/factory.py
new file mode 100644
index 00000000..6c273eb8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/factory.py
@@ -0,0 +1,823 @@
+import contextlib
+import functools
+import logging
+from typing import (
+    TYPE_CHECKING,
+    Callable,
+    Dict,
+    FrozenSet,
+    Iterable,
+    Iterator,
+    List,
+    Mapping,
+    NamedTuple,
+    Optional,
+    Protocol,
+    Sequence,
+    Set,
+    Tuple,
+    TypeVar,
+    cast,
+)
+
+from pip._vendor.packaging.requirements import InvalidRequirement
+from pip._vendor.packaging.specifiers import SpecifierSet
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+from pip._vendor.packaging.version import InvalidVersion, Version
+from pip._vendor.resolvelib import ResolutionImpossible
+
+from pip._internal.cache import CacheEntry, WheelCache
+from pip._internal.exceptions import (
+    DistributionNotFound,
+    InstallationError,
+    InvalidInstalledPackage,
+    MetadataInconsistent,
+    MetadataInvalid,
+    UnsupportedPythonVersion,
+    UnsupportedWheel,
+)
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.metadata import BaseDistribution, get_default_environment
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.operations.prepare import RequirementPreparer
+from pip._internal.req.constructors import (
+    install_req_drop_extras,
+    install_req_from_link_and_ireq,
+)
+from pip._internal.req.req_install import (
+    InstallRequirement,
+    check_invalid_constraint_type,
+)
+from pip._internal.resolution.base import InstallRequirementProvider
+from pip._internal.utils.compatibility_tags import get_supported
+from pip._internal.utils.hashes import Hashes
+from pip._internal.utils.packaging import get_requirement
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+from .base import Candidate, Constraint, Requirement
+from .candidates import (
+    AlreadyInstalledCandidate,
+    BaseCandidate,
+    EditableCandidate,
+    ExtrasCandidate,
+    LinkCandidate,
+    RequiresPythonCandidate,
+    as_base_candidate,
+)
+from .found_candidates import FoundCandidates, IndexCandidateInfo
+from .requirements import (
+    ExplicitRequirement,
+    RequiresPythonRequirement,
+    SpecifierRequirement,
+    SpecifierWithoutExtrasRequirement,
+    UnsatisfiableRequirement,
+)
+
+if TYPE_CHECKING:
+
+    class ConflictCause(Protocol):
+        requirement: RequiresPythonRequirement
+        parent: Candidate
+
+
+logger = logging.getLogger(__name__)
+
+C = TypeVar("C")
+Cache = Dict[Link, C]
+
+
+class CollectedRootRequirements(NamedTuple):
+    requirements: List[Requirement]
+    constraints: Dict[str, Constraint]
+    user_requested: Dict[str, int]
+
+
+class Factory:
+    def __init__(
+        self,
+        finder: PackageFinder,
+        preparer: RequirementPreparer,
+        make_install_req: InstallRequirementProvider,
+        wheel_cache: Optional[WheelCache],
+        use_user_site: bool,
+        force_reinstall: bool,
+        ignore_installed: bool,
+        ignore_requires_python: bool,
+        py_version_info: Optional[Tuple[int, ...]] = None,
+    ) -> None:
+        self._finder = finder
+        self.preparer = preparer
+        self._wheel_cache = wheel_cache
+        self._python_candidate = RequiresPythonCandidate(py_version_info)
+        self._make_install_req_from_spec = make_install_req
+        self._use_user_site = use_user_site
+        self._force_reinstall = force_reinstall
+        self._ignore_requires_python = ignore_requires_python
+
+        self._build_failures: Cache[InstallationError] = {}
+        self._link_candidate_cache: Cache[LinkCandidate] = {}
+        self._editable_candidate_cache: Cache[EditableCandidate] = {}
+        self._installed_candidate_cache: Dict[str, AlreadyInstalledCandidate] = {}
+        self._extras_candidate_cache: Dict[
+            Tuple[int, FrozenSet[NormalizedName]], ExtrasCandidate
+        ] = {}
+        self._supported_tags_cache = get_supported()
+
+        if not ignore_installed:
+            env = get_default_environment()
+            self._installed_dists = {
+                dist.canonical_name: dist
+                for dist in env.iter_installed_distributions(local_only=False)
+            }
+        else:
+            self._installed_dists = {}
+
+    @property
+    def force_reinstall(self) -> bool:
+        return self._force_reinstall
+
+    def _fail_if_link_is_unsupported_wheel(self, link: Link) -> None:
+        if not link.is_wheel:
+            return
+        wheel = Wheel(link.filename)
+        if wheel.supported(self._finder.target_python.get_unsorted_tags()):
+            return
+        msg = f"{link.filename} is not a supported wheel on this platform."
+        raise UnsupportedWheel(msg)
+
+    def _make_extras_candidate(
+        self,
+        base: BaseCandidate,
+        extras: FrozenSet[str],
+        *,
+        comes_from: Optional[InstallRequirement] = None,
+    ) -> ExtrasCandidate:
+        cache_key = (id(base), frozenset(canonicalize_name(e) for e in extras))
+        try:
+            candidate = self._extras_candidate_cache[cache_key]
+        except KeyError:
+            candidate = ExtrasCandidate(base, extras, comes_from=comes_from)
+            self._extras_candidate_cache[cache_key] = candidate
+        return candidate
+
+    def _make_candidate_from_dist(
+        self,
+        dist: BaseDistribution,
+        extras: FrozenSet[str],
+        template: InstallRequirement,
+    ) -> Candidate:
+        try:
+            base = self._installed_candidate_cache[dist.canonical_name]
+        except KeyError:
+            base = AlreadyInstalledCandidate(dist, template, factory=self)
+            self._installed_candidate_cache[dist.canonical_name] = base
+        if not extras:
+            return base
+        return self._make_extras_candidate(base, extras, comes_from=template)
+
+    def _make_candidate_from_link(
+        self,
+        link: Link,
+        extras: FrozenSet[str],
+        template: InstallRequirement,
+        name: Optional[NormalizedName],
+        version: Optional[Version],
+    ) -> Optional[Candidate]:
+        base: Optional[BaseCandidate] = self._make_base_candidate_from_link(
+            link, template, name, version
+        )
+        if not extras or base is None:
+            return base
+        return self._make_extras_candidate(base, extras, comes_from=template)
+
+    def _make_base_candidate_from_link(
+        self,
+        link: Link,
+        template: InstallRequirement,
+        name: Optional[NormalizedName],
+        version: Optional[Version],
+    ) -> Optional[BaseCandidate]:
+        # TODO: Check already installed candidate, and use it if the link and
+        # editable flag match.
+
+        if link in self._build_failures:
+            # We already tried this candidate before, and it does not build.
+            # Don't bother trying again.
+            return None
+
+        if template.editable:
+            if link not in self._editable_candidate_cache:
+                try:
+                    self._editable_candidate_cache[link] = EditableCandidate(
+                        link,
+                        template,
+                        factory=self,
+                        name=name,
+                        version=version,
+                    )
+                except (MetadataInconsistent, MetadataInvalid) as e:
+                    logger.info(
+                        "Discarding [blue underline]%s[/]: [yellow]%s[reset]",
+                        link,
+                        e,
+                        extra={"markup": True},
+                    )
+                    self._build_failures[link] = e
+                    return None
+
+            return self._editable_candidate_cache[link]
+        else:
+            if link not in self._link_candidate_cache:
+                try:
+                    self._link_candidate_cache[link] = LinkCandidate(
+                        link,
+                        template,
+                        factory=self,
+                        name=name,
+                        version=version,
+                    )
+                except MetadataInconsistent as e:
+                    logger.info(
+                        "Discarding [blue underline]%s[/]: [yellow]%s[reset]",
+                        link,
+                        e,
+                        extra={"markup": True},
+                    )
+                    self._build_failures[link] = e
+                    return None
+            return self._link_candidate_cache[link]
+
+    def _iter_found_candidates(
+        self,
+        ireqs: Sequence[InstallRequirement],
+        specifier: SpecifierSet,
+        hashes: Hashes,
+        prefers_installed: bool,
+        incompatible_ids: Set[int],
+    ) -> Iterable[Candidate]:
+        if not ireqs:
+            return ()
+
+        # The InstallRequirement implementation requires us to give it a
+        # "template". Here we just choose the first requirement to represent
+        # all of them.
+        # Hopefully the Project model can correct this mismatch in the future.
+        template = ireqs[0]
+        assert template.req, "Candidates found on index must be PEP 508"
+        name = canonicalize_name(template.req.name)
+
+        extras: FrozenSet[str] = frozenset()
+        for ireq in ireqs:
+            assert ireq.req, "Candidates found on index must be PEP 508"
+            specifier &= ireq.req.specifier
+            hashes &= ireq.hashes(trust_internet=False)
+            extras |= frozenset(ireq.extras)
+
+        def _get_installed_candidate() -> Optional[Candidate]:
+            """Get the candidate for the currently-installed version."""
+            # If --force-reinstall is set, we want the version from the index
+            # instead, so we "pretend" there is nothing installed.
+            if self._force_reinstall:
+                return None
+            try:
+                installed_dist = self._installed_dists[name]
+            except KeyError:
+                return None
+
+            try:
+                # Don't use the installed distribution if its version
+                # does not fit the current dependency graph.
+                if not specifier.contains(installed_dist.version, prereleases=True):
+                    return None
+            except InvalidVersion as e:
+                raise InvalidInstalledPackage(dist=installed_dist, invalid_exc=e)
+
+            candidate = self._make_candidate_from_dist(
+                dist=installed_dist,
+                extras=extras,
+                template=template,
+            )
+            # The candidate is a known incompatibility. Don't use it.
+            if id(candidate) in incompatible_ids:
+                return None
+            return candidate
+
+        def iter_index_candidate_infos() -> Iterator[IndexCandidateInfo]:
+            result = self._finder.find_best_candidate(
+                project_name=name,
+                specifier=specifier,
+                hashes=hashes,
+            )
+            icans = result.applicable_candidates
+
+            # PEP 592: Yanked releases are ignored unless the specifier
+            # explicitly pins a version (via '==' or '===') that can be
+            # solely satisfied by a yanked release.
+            all_yanked = all(ican.link.is_yanked for ican in icans)
+
+            def is_pinned(specifier: SpecifierSet) -> bool:
+                for sp in specifier:
+                    if sp.operator == "===":
+                        return True
+                    if sp.operator != "==":
+                        continue
+                    if sp.version.endswith(".*"):
+                        continue
+                    return True
+                return False
+
+            pinned = is_pinned(specifier)
+
+            # PackageFinder returns earlier versions first, so we reverse.
+            for ican in reversed(icans):
+                if not (all_yanked and pinned) and ican.link.is_yanked:
+                    continue
+                func = functools.partial(
+                    self._make_candidate_from_link,
+                    link=ican.link,
+                    extras=extras,
+                    template=template,
+                    name=name,
+                    version=ican.version,
+                )
+                yield ican.version, func
+
+        return FoundCandidates(
+            iter_index_candidate_infos,
+            _get_installed_candidate(),
+            prefers_installed,
+            incompatible_ids,
+        )
+
+    def _iter_explicit_candidates_from_base(
+        self,
+        base_requirements: Iterable[Requirement],
+        extras: FrozenSet[str],
+    ) -> Iterator[Candidate]:
+        """Produce explicit candidates from the base given an extra-ed package.
+
+        :param base_requirements: Requirements known to the resolver. The
+            requirements are guaranteed to not have extras.
+        :param extras: The extras to inject into the explicit requirements'
+            candidates.
+        """
+        for req in base_requirements:
+            lookup_cand, _ = req.get_candidate_lookup()
+            if lookup_cand is None:  # Not explicit.
+                continue
+            # We've stripped extras from the identifier, and should always
+            # get a BaseCandidate here, unless there's a bug elsewhere.
+            base_cand = as_base_candidate(lookup_cand)
+            assert base_cand is not None, "no extras here"
+            yield self._make_extras_candidate(base_cand, extras)
+
+    def _iter_candidates_from_constraints(
+        self,
+        identifier: str,
+        constraint: Constraint,
+        template: InstallRequirement,
+    ) -> Iterator[Candidate]:
+        """Produce explicit candidates from constraints.
+
+        This creates "fake" InstallRequirement objects that are basically clones
+        of what "should" be the template, but with original_link set to link.
+        """
+        for link in constraint.links:
+            self._fail_if_link_is_unsupported_wheel(link)
+            candidate = self._make_base_candidate_from_link(
+                link,
+                template=install_req_from_link_and_ireq(link, template),
+                name=canonicalize_name(identifier),
+                version=None,
+            )
+            if candidate:
+                yield candidate
+
+    def find_candidates(
+        self,
+        identifier: str,
+        requirements: Mapping[str, Iterable[Requirement]],
+        incompatibilities: Mapping[str, Iterator[Candidate]],
+        constraint: Constraint,
+        prefers_installed: bool,
+        is_satisfied_by: Callable[[Requirement, Candidate], bool],
+    ) -> Iterable[Candidate]:
+        # Collect basic lookup information from the requirements.
+        explicit_candidates: Set[Candidate] = set()
+        ireqs: List[InstallRequirement] = []
+        for req in requirements[identifier]:
+            cand, ireq = req.get_candidate_lookup()
+            if cand is not None:
+                explicit_candidates.add(cand)
+            if ireq is not None:
+                ireqs.append(ireq)
+
+        # If the current identifier contains extras, add requires and explicit
+        # candidates from entries from extra-less identifier.
+        with contextlib.suppress(InvalidRequirement):
+            parsed_requirement = get_requirement(identifier)
+            if parsed_requirement.name != identifier:
+                explicit_candidates.update(
+                    self._iter_explicit_candidates_from_base(
+                        requirements.get(parsed_requirement.name, ()),
+                        frozenset(parsed_requirement.extras),
+                    ),
+                )
+                for req in requirements.get(parsed_requirement.name, []):
+                    _, ireq = req.get_candidate_lookup()
+                    if ireq is not None:
+                        ireqs.append(ireq)
+
+        # Add explicit candidates from constraints. We only do this if there are
+        # known ireqs, which represent requirements not already explicit. If
+        # there are no ireqs, we're constraining already-explicit requirements,
+        # which is handled later when we return the explicit candidates.
+        if ireqs:
+            try:
+                explicit_candidates.update(
+                    self._iter_candidates_from_constraints(
+                        identifier,
+                        constraint,
+                        template=ireqs[0],
+                    ),
+                )
+            except UnsupportedWheel:
+                # If we're constrained to install a wheel incompatible with the
+                # target architecture, no candidates will ever be valid.
+                return ()
+
+        # Since we cache all the candidates, incompatibility identification
+        # can be made quicker by comparing only the id() values.
+        incompat_ids = {id(c) for c in incompatibilities.get(identifier, ())}
+
+        # If none of the requirements want an explicit candidate, we can ask
+        # the finder for candidates.
+        if not explicit_candidates:
+            return self._iter_found_candidates(
+                ireqs,
+                constraint.specifier,
+                constraint.hashes,
+                prefers_installed,
+                incompat_ids,
+            )
+
+        return (
+            c
+            for c in explicit_candidates
+            if id(c) not in incompat_ids
+            and constraint.is_satisfied_by(c)
+            and all(is_satisfied_by(req, c) for req in requirements[identifier])
+        )
+
+    def _make_requirements_from_install_req(
+        self, ireq: InstallRequirement, requested_extras: Iterable[str]
+    ) -> Iterator[Requirement]:
+        """
+        Returns requirement objects associated with the given InstallRequirement. In
+        most cases this will be a single object but the following special cases exist:
+            - the InstallRequirement has markers that do not apply -> result is empty
+            - the InstallRequirement has both a constraint (or link) and extras
+                -> result is split in two requirement objects: one with the constraint
+                (or link) and one with the extra. This allows centralized constraint
+                handling for the base, resulting in fewer candidate rejections.
+        """
+        if not ireq.match_markers(requested_extras):
+            logger.info(
+                "Ignoring %s: markers '%s' don't match your environment",
+                ireq.name,
+                ireq.markers,
+            )
+        elif not ireq.link:
+            if ireq.extras and ireq.req is not None and ireq.req.specifier:
+                yield SpecifierWithoutExtrasRequirement(ireq)
+            yield SpecifierRequirement(ireq)
+        else:
+            self._fail_if_link_is_unsupported_wheel(ireq.link)
+            # Always make the link candidate for the base requirement to make it
+            # available to `find_candidates` for explicit candidate lookup for any
+            # set of extras.
+            # The extras are required separately via a second requirement.
+            cand = self._make_base_candidate_from_link(
+                ireq.link,
+                template=install_req_drop_extras(ireq) if ireq.extras else ireq,
+                name=canonicalize_name(ireq.name) if ireq.name else None,
+                version=None,
+            )
+            if cand is None:
+                # There's no way we can satisfy a URL requirement if the underlying
+                # candidate fails to build. An unnamed URL must be user-supplied, so
+                # we fail eagerly. If the URL is named, an unsatisfiable requirement
+                # can make the resolver do the right thing, either backtrack (and
+                # maybe find some other requirement that's buildable) or raise a
+                # ResolutionImpossible eventually.
+                if not ireq.name:
+                    raise self._build_failures[ireq.link]
+                yield UnsatisfiableRequirement(canonicalize_name(ireq.name))
+            else:
+                # require the base from the link
+                yield self.make_requirement_from_candidate(cand)
+                if ireq.extras:
+                    # require the extras on top of the base candidate
+                    yield self.make_requirement_from_candidate(
+                        self._make_extras_candidate(cand, frozenset(ireq.extras))
+                    )
+
+    def collect_root_requirements(
+        self, root_ireqs: List[InstallRequirement]
+    ) -> CollectedRootRequirements:
+        collected = CollectedRootRequirements([], {}, {})
+        for i, ireq in enumerate(root_ireqs):
+            if ireq.constraint:
+                # Ensure we only accept valid constraints
+                problem = check_invalid_constraint_type(ireq)
+                if problem:
+                    raise InstallationError(problem)
+                if not ireq.match_markers():
+                    continue
+                assert ireq.name, "Constraint must be named"
+                name = canonicalize_name(ireq.name)
+                if name in collected.constraints:
+                    collected.constraints[name] &= ireq
+                else:
+                    collected.constraints[name] = Constraint.from_ireq(ireq)
+            else:
+                reqs = list(
+                    self._make_requirements_from_install_req(
+                        ireq,
+                        requested_extras=(),
+                    )
+                )
+                if not reqs:
+                    continue
+                template = reqs[0]
+                if ireq.user_supplied and template.name not in collected.user_requested:
+                    collected.user_requested[template.name] = i
+                collected.requirements.extend(reqs)
+        # Put requirements with extras at the end of the root requires. This does not
+        # affect resolvelib's picking preference but it does affect its initial criteria
+        # population: by putting extras at the end we enable the candidate finder to
+        # present resolvelib with a smaller set of candidates to resolvelib, already
+        # taking into account any non-transient constraints on the associated base. This
+        # means resolvelib will have fewer candidates to visit and reject.
+        # Python's list sort is stable, meaning relative order is kept for objects with
+        # the same key.
+        collected.requirements.sort(key=lambda r: r.name != r.project_name)
+        return collected
+
+    def make_requirement_from_candidate(
+        self, candidate: Candidate
+    ) -> ExplicitRequirement:
+        return ExplicitRequirement(candidate)
+
+    def make_requirements_from_spec(
+        self,
+        specifier: str,
+        comes_from: Optional[InstallRequirement],
+        requested_extras: Iterable[str] = (),
+    ) -> Iterator[Requirement]:
+        """
+        Returns requirement objects associated with the given specifier. In most cases
+        this will be a single object but the following special cases exist:
+            - the specifier has markers that do not apply -> result is empty
+            - the specifier has both a constraint and extras -> result is split
+                in two requirement objects: one with the constraint and one with the
+                extra. This allows centralized constraint handling for the base,
+                resulting in fewer candidate rejections.
+        """
+        ireq = self._make_install_req_from_spec(specifier, comes_from)
+        return self._make_requirements_from_install_req(ireq, requested_extras)
+
+    def make_requires_python_requirement(
+        self,
+        specifier: SpecifierSet,
+    ) -> Optional[Requirement]:
+        if self._ignore_requires_python:
+            return None
+        # Don't bother creating a dependency for an empty Requires-Python.
+        if not str(specifier):
+            return None
+        return RequiresPythonRequirement(specifier, self._python_candidate)
+
+    def get_wheel_cache_entry(
+        self, link: Link, name: Optional[str]
+    ) -> Optional[CacheEntry]:
+        """Look up the link in the wheel cache.
+
+        If ``preparer.require_hashes`` is True, don't use the wheel cache,
+        because cached wheels, always built locally, have different hashes
+        than the files downloaded from the index server and thus throw false
+        hash mismatches. Furthermore, cached wheels at present have
+        nondeterministic contents due to file modification times.
+        """
+        if self._wheel_cache is None:
+            return None
+        return self._wheel_cache.get_cache_entry(
+            link=link,
+            package_name=name,
+            supported_tags=self._supported_tags_cache,
+        )
+
+    def get_dist_to_uninstall(self, candidate: Candidate) -> Optional[BaseDistribution]:
+        # TODO: Are there more cases this needs to return True? Editable?
+        dist = self._installed_dists.get(candidate.project_name)
+        if dist is None:  # Not installed, no uninstallation required.
+            return None
+
+        # We're installing into global site. The current installation must
+        # be uninstalled, no matter it's in global or user site, because the
+        # user site installation has precedence over global.
+        if not self._use_user_site:
+            return dist
+
+        # We're installing into user site. Remove the user site installation.
+        if dist.in_usersite:
+            return dist
+
+        # We're installing into user site, but the installed incompatible
+        # package is in global site. We can't uninstall that, and would let
+        # the new user installation to "shadow" it. But shadowing won't work
+        # in virtual environments, so we error out.
+        if running_under_virtualenv() and dist.in_site_packages:
+            message = (
+                f"Will not install to the user site because it will lack "
+                f"sys.path precedence to {dist.raw_name} in {dist.location}"
+            )
+            raise InstallationError(message)
+        return None
+
+    def _report_requires_python_error(
+        self, causes: Sequence["ConflictCause"]
+    ) -> UnsupportedPythonVersion:
+        assert causes, "Requires-Python error reported with no cause"
+
+        version = self._python_candidate.version
+
+        if len(causes) == 1:
+            specifier = str(causes[0].requirement.specifier)
+            message = (
+                f"Package {causes[0].parent.name!r} requires a different "
+                f"Python: {version} not in {specifier!r}"
+            )
+            return UnsupportedPythonVersion(message)
+
+        message = f"Packages require a different Python. {version} not in:"
+        for cause in causes:
+            package = cause.parent.format_for_error()
+            specifier = str(cause.requirement.specifier)
+            message += f"\n{specifier!r} (required by {package})"
+        return UnsupportedPythonVersion(message)
+
+    def _report_single_requirement_conflict(
+        self, req: Requirement, parent: Optional[Candidate]
+    ) -> DistributionNotFound:
+        if parent is None:
+            req_disp = str(req)
+        else:
+            req_disp = f"{req} (from {parent.name})"
+
+        cands = self._finder.find_all_candidates(req.project_name)
+        skipped_by_requires_python = self._finder.requires_python_skipped_reasons()
+
+        versions_set: Set[Version] = set()
+        yanked_versions_set: Set[Version] = set()
+        for c in cands:
+            is_yanked = c.link.is_yanked if c.link else False
+            if is_yanked:
+                yanked_versions_set.add(c.version)
+            else:
+                versions_set.add(c.version)
+
+        versions = [str(v) for v in sorted(versions_set)]
+        yanked_versions = [str(v) for v in sorted(yanked_versions_set)]
+
+        if yanked_versions:
+            # Saying "version X is yanked" isn't entirely accurate.
+            # https://github.com/pypa/pip/issues/11745#issuecomment-1402805842
+            logger.critical(
+                "Ignored the following yanked versions: %s",
+                ", ".join(yanked_versions) or "none",
+            )
+        if skipped_by_requires_python:
+            logger.critical(
+                "Ignored the following versions that require a different python "
+                "version: %s",
+                "; ".join(skipped_by_requires_python) or "none",
+            )
+        logger.critical(
+            "Could not find a version that satisfies the requirement %s "
+            "(from versions: %s)",
+            req_disp,
+            ", ".join(versions) or "none",
+        )
+        if str(req) == "requirements.txt":
+            logger.info(
+                "HINT: You are attempting to install a package literally "
+                'named "requirements.txt" (which cannot exist). Consider '
+                "using the '-r' flag to install the packages listed in "
+                "requirements.txt"
+            )
+
+        return DistributionNotFound(f"No matching distribution found for {req}")
+
+    def get_installation_error(
+        self,
+        e: "ResolutionImpossible[Requirement, Candidate]",
+        constraints: Dict[str, Constraint],
+    ) -> InstallationError:
+        assert e.causes, "Installation error reported with no cause"
+
+        # If one of the things we can't solve is "we need Python X.Y",
+        # that is what we report.
+        requires_python_causes = [
+            cause
+            for cause in e.causes
+            if isinstance(cause.requirement, RequiresPythonRequirement)
+            and not cause.requirement.is_satisfied_by(self._python_candidate)
+        ]
+        if requires_python_causes:
+            # The comprehension above makes sure all Requirement instances are
+            # RequiresPythonRequirement, so let's cast for convenience.
+            return self._report_requires_python_error(
+                cast("Sequence[ConflictCause]", requires_python_causes),
+            )
+
+        # Otherwise, we have a set of causes which can't all be satisfied
+        # at once.
+
+        # The simplest case is when we have *one* cause that can't be
+        # satisfied. We just report that case.
+        if len(e.causes) == 1:
+            req, parent = e.causes[0]
+            if req.name not in constraints:
+                return self._report_single_requirement_conflict(req, parent)
+
+        # OK, we now have a list of requirements that can't all be
+        # satisfied at once.
+
+        # A couple of formatting helpers
+        def text_join(parts: List[str]) -> str:
+            if len(parts) == 1:
+                return parts[0]
+
+            return ", ".join(parts[:-1]) + " and " + parts[-1]
+
+        def describe_trigger(parent: Candidate) -> str:
+            ireq = parent.get_install_requirement()
+            if not ireq or not ireq.comes_from:
+                return f"{parent.name}=={parent.version}"
+            if isinstance(ireq.comes_from, InstallRequirement):
+                return str(ireq.comes_from.name)
+            return str(ireq.comes_from)
+
+        triggers = set()
+        for req, parent in e.causes:
+            if parent is None:
+                # This is a root requirement, so we can report it directly
+                trigger = req.format_for_error()
+            else:
+                trigger = describe_trigger(parent)
+            triggers.add(trigger)
+
+        if triggers:
+            info = text_join(sorted(triggers))
+        else:
+            info = "the requested packages"
+
+        msg = (
+            f"Cannot install {info} because these package versions "
+            "have conflicting dependencies."
+        )
+        logger.critical(msg)
+        msg = "\nThe conflict is caused by:"
+
+        relevant_constraints = set()
+        for req, parent in e.causes:
+            if req.name in constraints:
+                relevant_constraints.add(req.name)
+            msg = msg + "\n    "
+            if parent:
+                msg = msg + f"{parent.name} {parent.version} depends on "
+            else:
+                msg = msg + "The user requested "
+            msg = msg + req.format_for_error()
+        for key in relevant_constraints:
+            spec = constraints[key].specifier
+            msg += f"\n    The user requested (constraint) {key}{spec}"
+
+        msg = (
+            msg
+            + "\n\n"
+            + "To fix this you could try to:\n"
+            + "1. loosen the range of package versions you've specified\n"
+            + "2. remove package versions to allow pip to attempt to solve "
+            + "the dependency conflict\n"
+        )
+
+        logger.info(msg)
+
+        return DistributionNotFound(
+            "ResolutionImpossible: for help visit "
+            "https://pip.pypa.io/en/latest/topics/dependency-resolution/"
+            "#dealing-with-dependency-conflicts"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/found_candidates.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/found_candidates.py
new file mode 100644
index 00000000..a1d57e0f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/found_candidates.py
@@ -0,0 +1,174 @@
+"""Utilities to lazily create and visit candidates found.
+
+Creating and visiting a candidate is a *very* costly operation. It involves
+fetching, extracting, potentially building modules from source, and verifying
+distribution metadata. It is therefore crucial for performance to keep
+everything here lazy all the way down, so we only touch candidates that we
+absolutely need, and not "download the world" when we only need one version of
+something.
+"""
+
+import functools
+import logging
+from collections.abc import Sequence
+from typing import TYPE_CHECKING, Any, Callable, Iterator, Optional, Set, Tuple
+
+from pip._vendor.packaging.version import _BaseVersion
+
+from pip._internal.exceptions import MetadataInvalid
+
+from .base import Candidate
+
+logger = logging.getLogger(__name__)
+
+IndexCandidateInfo = Tuple[_BaseVersion, Callable[[], Optional[Candidate]]]
+
+if TYPE_CHECKING:
+    SequenceCandidate = Sequence[Candidate]
+else:
+    # For compatibility: Python before 3.9 does not support using [] on the
+    # Sequence class.
+    #
+    # >>> from collections.abc import Sequence
+    # >>> Sequence[str]
+    # Traceback (most recent call last):
+    #   File "<stdin>", line 1, in <module>
+    # TypeError: 'ABCMeta' object is not subscriptable
+    #
+    # TODO: Remove this block after dropping Python 3.8 support.
+    SequenceCandidate = Sequence
+
+
+def _iter_built(infos: Iterator[IndexCandidateInfo]) -> Iterator[Candidate]:
+    """Iterator for ``FoundCandidates``.
+
+    This iterator is used when the package is not already installed. Candidates
+    from index come later in their normal ordering.
+    """
+    versions_found: Set[_BaseVersion] = set()
+    for version, func in infos:
+        if version in versions_found:
+            continue
+        try:
+            candidate = func()
+        except MetadataInvalid as e:
+            logger.warning(
+                "Ignoring version %s of %s since it has invalid metadata:\n"
+                "%s\n"
+                "Please use pip<24.1 if you need to use this version.",
+                version,
+                e.ireq.name,
+                e,
+            )
+            # Mark version as found to avoid trying other candidates with the same
+            # version, since they most likely have invalid metadata as well.
+            versions_found.add(version)
+        else:
+            if candidate is None:
+                continue
+            yield candidate
+            versions_found.add(version)
+
+
+def _iter_built_with_prepended(
+    installed: Candidate, infos: Iterator[IndexCandidateInfo]
+) -> Iterator[Candidate]:
+    """Iterator for ``FoundCandidates``.
+
+    This iterator is used when the resolver prefers the already-installed
+    candidate and NOT to upgrade. The installed candidate is therefore
+    always yielded first, and candidates from index come later in their
+    normal ordering, except skipped when the version is already installed.
+    """
+    yield installed
+    versions_found: Set[_BaseVersion] = {installed.version}
+    for version, func in infos:
+        if version in versions_found:
+            continue
+        candidate = func()
+        if candidate is None:
+            continue
+        yield candidate
+        versions_found.add(version)
+
+
+def _iter_built_with_inserted(
+    installed: Candidate, infos: Iterator[IndexCandidateInfo]
+) -> Iterator[Candidate]:
+    """Iterator for ``FoundCandidates``.
+
+    This iterator is used when the resolver prefers to upgrade an
+    already-installed package. Candidates from index are returned in their
+    normal ordering, except replaced when the version is already installed.
+
+    The implementation iterates through and yields other candidates, inserting
+    the installed candidate exactly once before we start yielding older or
+    equivalent candidates, or after all other candidates if they are all newer.
+    """
+    versions_found: Set[_BaseVersion] = set()
+    for version, func in infos:
+        if version in versions_found:
+            continue
+        # If the installed candidate is better, yield it first.
+        if installed.version >= version:
+            yield installed
+            versions_found.add(installed.version)
+        candidate = func()
+        if candidate is None:
+            continue
+        yield candidate
+        versions_found.add(version)
+
+    # If the installed candidate is older than all other candidates.
+    if installed.version not in versions_found:
+        yield installed
+
+
+class FoundCandidates(SequenceCandidate):
+    """A lazy sequence to provide candidates to the resolver.
+
+    The intended usage is to return this from `find_matches()` so the resolver
+    can iterate through the sequence multiple times, but only access the index
+    page when remote packages are actually needed. This improve performances
+    when suitable candidates are already installed on disk.
+    """
+
+    def __init__(
+        self,
+        get_infos: Callable[[], Iterator[IndexCandidateInfo]],
+        installed: Optional[Candidate],
+        prefers_installed: bool,
+        incompatible_ids: Set[int],
+    ):
+        self._get_infos = get_infos
+        self._installed = installed
+        self._prefers_installed = prefers_installed
+        self._incompatible_ids = incompatible_ids
+
+    def __getitem__(self, index: Any) -> Any:
+        # Implemented to satisfy the ABC check. This is not needed by the
+        # resolver, and should not be used by the provider either (for
+        # performance reasons).
+        raise NotImplementedError("don't do this")
+
+    def __iter__(self) -> Iterator[Candidate]:
+        infos = self._get_infos()
+        if not self._installed:
+            iterator = _iter_built(infos)
+        elif self._prefers_installed:
+            iterator = _iter_built_with_prepended(self._installed, infos)
+        else:
+            iterator = _iter_built_with_inserted(self._installed, infos)
+        return (c for c in iterator if id(c) not in self._incompatible_ids)
+
+    def __len__(self) -> int:
+        # Implemented to satisfy the ABC check. This is not needed by the
+        # resolver, and should not be used by the provider either (for
+        # performance reasons).
+        raise NotImplementedError("don't do this")
+
+    @functools.lru_cache(maxsize=1)
+    def __bool__(self) -> bool:
+        if self._prefers_installed and self._installed:
+            return True
+        return any(self)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/provider.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/provider.py
new file mode 100644
index 00000000..fb0dd85f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/provider.py
@@ -0,0 +1,258 @@
+import collections
+import math
+from functools import lru_cache
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    Iterable,
+    Iterator,
+    Mapping,
+    Sequence,
+    TypeVar,
+    Union,
+)
+
+from pip._vendor.resolvelib.providers import AbstractProvider
+
+from .base import Candidate, Constraint, Requirement
+from .candidates import REQUIRES_PYTHON_IDENTIFIER
+from .factory import Factory
+
+if TYPE_CHECKING:
+    from pip._vendor.resolvelib.providers import Preference
+    from pip._vendor.resolvelib.resolvers import RequirementInformation
+
+    PreferenceInformation = RequirementInformation[Requirement, Candidate]
+
+    _ProviderBase = AbstractProvider[Requirement, Candidate, str]
+else:
+    _ProviderBase = AbstractProvider
+
+# Notes on the relationship between the provider, the factory, and the
+# candidate and requirement classes.
+#
+# The provider is a direct implementation of the resolvelib class. Its role
+# is to deliver the API that resolvelib expects.
+#
+# Rather than work with completely abstract "requirement" and "candidate"
+# concepts as resolvelib does, pip has concrete classes implementing these two
+# ideas. The API of Requirement and Candidate objects are defined in the base
+# classes, but essentially map fairly directly to the equivalent provider
+# methods. In particular, `find_matches` and `is_satisfied_by` are
+# requirement methods, and `get_dependencies` is a candidate method.
+#
+# The factory is the interface to pip's internal mechanisms. It is stateless,
+# and is created by the resolver and held as a property of the provider. It is
+# responsible for creating Requirement and Candidate objects, and provides
+# services to those objects (access to pip's finder and preparer).
+
+
+D = TypeVar("D")
+V = TypeVar("V")
+
+
+def _get_with_identifier(
+    mapping: Mapping[str, V],
+    identifier: str,
+    default: D,
+) -> Union[D, V]:
+    """Get item from a package name lookup mapping with a resolver identifier.
+
+    This extra logic is needed when the target mapping is keyed by package
+    name, which cannot be directly looked up with an identifier (which may
+    contain requested extras). Additional logic is added to also look up a value
+    by "cleaning up" the extras from the identifier.
+    """
+    if identifier in mapping:
+        return mapping[identifier]
+    # HACK: Theoretically we should check whether this identifier is a valid
+    # "NAME[EXTRAS]" format, and parse out the name part with packaging or
+    # some regular expression. But since pip's resolver only spits out three
+    # kinds of identifiers: normalized PEP 503 names, normalized names plus
+    # extras, and Requires-Python, we can cheat a bit here.
+    name, open_bracket, _ = identifier.partition("[")
+    if open_bracket and name in mapping:
+        return mapping[name]
+    return default
+
+
+class PipProvider(_ProviderBase):
+    """Pip's provider implementation for resolvelib.
+
+    :params constraints: A mapping of constraints specified by the user. Keys
+        are canonicalized project names.
+    :params ignore_dependencies: Whether the user specified ``--no-deps``.
+    :params upgrade_strategy: The user-specified upgrade strategy.
+    :params user_requested: A set of canonicalized package names that the user
+        supplied for pip to install/upgrade.
+    """
+
+    def __init__(
+        self,
+        factory: Factory,
+        constraints: Dict[str, Constraint],
+        ignore_dependencies: bool,
+        upgrade_strategy: str,
+        user_requested: Dict[str, int],
+    ) -> None:
+        self._factory = factory
+        self._constraints = constraints
+        self._ignore_dependencies = ignore_dependencies
+        self._upgrade_strategy = upgrade_strategy
+        self._user_requested = user_requested
+        self._known_depths: Dict[str, float] = collections.defaultdict(lambda: math.inf)
+
+    def identify(self, requirement_or_candidate: Union[Requirement, Candidate]) -> str:
+        return requirement_or_candidate.name
+
+    def get_preference(
+        self,
+        identifier: str,
+        resolutions: Mapping[str, Candidate],
+        candidates: Mapping[str, Iterator[Candidate]],
+        information: Mapping[str, Iterable["PreferenceInformation"]],
+        backtrack_causes: Sequence["PreferenceInformation"],
+    ) -> "Preference":
+        """Produce a sort key for given requirement based on preference.
+
+        The lower the return value is, the more preferred this group of
+        arguments is.
+
+        Currently pip considers the following in order:
+
+        * Prefer if any of the known requirements is "direct", e.g. points to an
+          explicit URL.
+        * If equal, prefer if any requirement is "pinned", i.e. contains
+          operator ``===`` or ``==``.
+        * If equal, calculate an approximate "depth" and resolve requirements
+          closer to the user-specified requirements first. If the depth cannot
+          by determined (eg: due to no matching parents), it is considered
+          infinite.
+        * Order user-specified requirements by the order they are specified.
+        * If equal, prefers "non-free" requirements, i.e. contains at least one
+          operator, such as ``>=`` or ``<``.
+        * If equal, order alphabetically for consistency (helps debuggability).
+        """
+        try:
+            next(iter(information[identifier]))
+        except StopIteration:
+            # There is no information for this identifier, so there's no known
+            # candidates.
+            has_information = False
+        else:
+            has_information = True
+
+        if has_information:
+            lookups = (r.get_candidate_lookup() for r, _ in information[identifier])
+            candidate, ireqs = zip(*lookups)
+        else:
+            candidate, ireqs = None, ()
+
+        operators = [
+            specifier.operator
+            for specifier_set in (ireq.specifier for ireq in ireqs if ireq)
+            for specifier in specifier_set
+        ]
+
+        direct = candidate is not None
+        pinned = any(op[:2] == "==" for op in operators)
+        unfree = bool(operators)
+
+        try:
+            requested_order: Union[int, float] = self._user_requested[identifier]
+        except KeyError:
+            requested_order = math.inf
+            if has_information:
+                parent_depths = (
+                    self._known_depths[parent.name] if parent is not None else 0.0
+                    for _, parent in information[identifier]
+                )
+                inferred_depth = min(d for d in parent_depths) + 1.0
+            else:
+                inferred_depth = math.inf
+        else:
+            inferred_depth = 1.0
+        self._known_depths[identifier] = inferred_depth
+
+        requested_order = self._user_requested.get(identifier, math.inf)
+
+        # Requires-Python has only one candidate and the check is basically
+        # free, so we always do it first to avoid needless work if it fails.
+        requires_python = identifier == REQUIRES_PYTHON_IDENTIFIER
+
+        # Prefer the causes of backtracking on the assumption that the problem
+        # resolving the dependency tree is related to the failures that caused
+        # the backtracking
+        backtrack_cause = self.is_backtrack_cause(identifier, backtrack_causes)
+
+        return (
+            not requires_python,
+            not direct,
+            not pinned,
+            not backtrack_cause,
+            inferred_depth,
+            requested_order,
+            not unfree,
+            identifier,
+        )
+
+    def find_matches(
+        self,
+        identifier: str,
+        requirements: Mapping[str, Iterator[Requirement]],
+        incompatibilities: Mapping[str, Iterator[Candidate]],
+    ) -> Iterable[Candidate]:
+        def _eligible_for_upgrade(identifier: str) -> bool:
+            """Are upgrades allowed for this project?
+
+            This checks the upgrade strategy, and whether the project was one
+            that the user specified in the command line, in order to decide
+            whether we should upgrade if there's a newer version available.
+
+            (Note that we don't need access to the `--upgrade` flag, because
+            an upgrade strategy of "to-satisfy-only" means that `--upgrade`
+            was not specified).
+            """
+            if self._upgrade_strategy == "eager":
+                return True
+            elif self._upgrade_strategy == "only-if-needed":
+                user_order = _get_with_identifier(
+                    self._user_requested,
+                    identifier,
+                    default=None,
+                )
+                return user_order is not None
+            return False
+
+        constraint = _get_with_identifier(
+            self._constraints,
+            identifier,
+            default=Constraint.empty(),
+        )
+        return self._factory.find_candidates(
+            identifier=identifier,
+            requirements=requirements,
+            constraint=constraint,
+            prefers_installed=(not _eligible_for_upgrade(identifier)),
+            incompatibilities=incompatibilities,
+            is_satisfied_by=self.is_satisfied_by,
+        )
+
+    @lru_cache(maxsize=None)
+    def is_satisfied_by(self, requirement: Requirement, candidate: Candidate) -> bool:
+        return requirement.is_satisfied_by(candidate)
+
+    def get_dependencies(self, candidate: Candidate) -> Sequence[Requirement]:
+        with_requires = not self._ignore_dependencies
+        return [r for r in candidate.iter_dependencies(with_requires) if r is not None]
+
+    @staticmethod
+    def is_backtrack_cause(
+        identifier: str, backtrack_causes: Sequence["PreferenceInformation"]
+    ) -> bool:
+        for backtrack_cause in backtrack_causes:
+            if identifier == backtrack_cause.requirement.name:
+                return True
+            if backtrack_cause.parent and identifier == backtrack_cause.parent.name:
+                return True
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/reporter.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/reporter.py
new file mode 100644
index 00000000..0594569d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/reporter.py
@@ -0,0 +1,81 @@
+from collections import defaultdict
+from logging import getLogger
+from typing import Any, DefaultDict
+
+from pip._vendor.resolvelib.reporters import BaseReporter
+
+from .base import Candidate, Requirement
+
+logger = getLogger(__name__)
+
+
+class PipReporter(BaseReporter):
+    def __init__(self) -> None:
+        self.reject_count_by_package: DefaultDict[str, int] = defaultdict(int)
+
+        self._messages_at_reject_count = {
+            1: (
+                "pip is looking at multiple versions of {package_name} to "
+                "determine which version is compatible with other "
+                "requirements. This could take a while."
+            ),
+            8: (
+                "pip is still looking at multiple versions of {package_name} to "
+                "determine which version is compatible with other "
+                "requirements. This could take a while."
+            ),
+            13: (
+                "This is taking longer than usual. You might need to provide "
+                "the dependency resolver with stricter constraints to reduce "
+                "runtime. See https://pip.pypa.io/warnings/backtracking for "
+                "guidance. If you want to abort this run, press Ctrl + C."
+            ),
+        }
+
+    def rejecting_candidate(self, criterion: Any, candidate: Candidate) -> None:
+        self.reject_count_by_package[candidate.name] += 1
+
+        count = self.reject_count_by_package[candidate.name]
+        if count not in self._messages_at_reject_count:
+            return
+
+        message = self._messages_at_reject_count[count]
+        logger.info("INFO: %s", message.format(package_name=candidate.name))
+
+        msg = "Will try a different candidate, due to conflict:"
+        for req_info in criterion.information:
+            req, parent = req_info.requirement, req_info.parent
+            # Inspired by Factory.get_installation_error
+            msg += "\n    "
+            if parent:
+                msg += f"{parent.name} {parent.version} depends on "
+            else:
+                msg += "The user requested "
+            msg += req.format_for_error()
+        logger.debug(msg)
+
+
+class PipDebuggingReporter(BaseReporter):
+    """A reporter that does an info log for every event it sees."""
+
+    def starting(self) -> None:
+        logger.info("Reporter.starting()")
+
+    def starting_round(self, index: int) -> None:
+        logger.info("Reporter.starting_round(%r)", index)
+
+    def ending_round(self, index: int, state: Any) -> None:
+        logger.info("Reporter.ending_round(%r, state)", index)
+        logger.debug("Reporter.ending_round(%r, %r)", index, state)
+
+    def ending(self, state: Any) -> None:
+        logger.info("Reporter.ending(%r)", state)
+
+    def adding_requirement(self, requirement: Requirement, parent: Candidate) -> None:
+        logger.info("Reporter.adding_requirement(%r, %r)", requirement, parent)
+
+    def rejecting_candidate(self, criterion: Any, candidate: Candidate) -> None:
+        logger.info("Reporter.rejecting_candidate(%r, %r)", criterion, candidate)
+
+    def pinning(self, candidate: Candidate) -> None:
+        logger.info("Reporter.pinning(%r)", candidate)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/requirements.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/requirements.py
new file mode 100644
index 00000000..b04f41b2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/requirements.py
@@ -0,0 +1,245 @@
+from typing import Any, Optional
+
+from pip._vendor.packaging.specifiers import SpecifierSet
+from pip._vendor.packaging.utils import NormalizedName, canonicalize_name
+
+from pip._internal.req.constructors import install_req_drop_extras
+from pip._internal.req.req_install import InstallRequirement
+
+from .base import Candidate, CandidateLookup, Requirement, format_name
+
+
+class ExplicitRequirement(Requirement):
+    def __init__(self, candidate: Candidate) -> None:
+        self.candidate = candidate
+
+    def __str__(self) -> str:
+        return str(self.candidate)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.candidate!r})"
+
+    def __hash__(self) -> int:
+        return hash(self.candidate)
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, ExplicitRequirement):
+            return False
+        return self.candidate == other.candidate
+
+    @property
+    def project_name(self) -> NormalizedName:
+        # No need to canonicalize - the candidate did this
+        return self.candidate.project_name
+
+    @property
+    def name(self) -> str:
+        # No need to canonicalize - the candidate did this
+        return self.candidate.name
+
+    def format_for_error(self) -> str:
+        return self.candidate.format_for_error()
+
+    def get_candidate_lookup(self) -> CandidateLookup:
+        return self.candidate, None
+
+    def is_satisfied_by(self, candidate: Candidate) -> bool:
+        return candidate == self.candidate
+
+
+class SpecifierRequirement(Requirement):
+    def __init__(self, ireq: InstallRequirement) -> None:
+        assert ireq.link is None, "This is a link, not a specifier"
+        self._ireq = ireq
+        self._equal_cache: Optional[str] = None
+        self._hash: Optional[int] = None
+        self._extras = frozenset(canonicalize_name(e) for e in self._ireq.extras)
+
+    @property
+    def _equal(self) -> str:
+        if self._equal_cache is not None:
+            return self._equal_cache
+
+        self._equal_cache = str(self._ireq)
+        return self._equal_cache
+
+    def __str__(self) -> str:
+        return str(self._ireq.req)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({str(self._ireq.req)!r})"
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, SpecifierRequirement):
+            return NotImplemented
+        return self._equal == other._equal
+
+    def __hash__(self) -> int:
+        if self._hash is not None:
+            return self._hash
+
+        self._hash = hash(self._equal)
+        return self._hash
+
+    @property
+    def project_name(self) -> NormalizedName:
+        assert self._ireq.req, "Specifier-backed ireq is always PEP 508"
+        return canonicalize_name(self._ireq.req.name)
+
+    @property
+    def name(self) -> str:
+        return format_name(self.project_name, self._extras)
+
+    def format_for_error(self) -> str:
+        # Convert comma-separated specifiers into "A, B, ..., F and G"
+        # This makes the specifier a bit more "human readable", without
+        # risking a change in meaning. (Hopefully! Not all edge cases have
+        # been checked)
+        parts = [s.strip() for s in str(self).split(",")]
+        if len(parts) == 0:
+            return ""
+        elif len(parts) == 1:
+            return parts[0]
+
+        return ", ".join(parts[:-1]) + " and " + parts[-1]
+
+    def get_candidate_lookup(self) -> CandidateLookup:
+        return None, self._ireq
+
+    def is_satisfied_by(self, candidate: Candidate) -> bool:
+        assert candidate.name == self.name, (
+            f"Internal issue: Candidate is not for this requirement "
+            f"{candidate.name} vs {self.name}"
+        )
+        # We can safely always allow prereleases here since PackageFinder
+        # already implements the prerelease logic, and would have filtered out
+        # prerelease candidates if the user does not expect them.
+        assert self._ireq.req, "Specifier-backed ireq is always PEP 508"
+        spec = self._ireq.req.specifier
+        return spec.contains(candidate.version, prereleases=True)
+
+
+class SpecifierWithoutExtrasRequirement(SpecifierRequirement):
+    """
+    Requirement backed by an install requirement on a base package.
+    Trims extras from its install requirement if there are any.
+    """
+
+    def __init__(self, ireq: InstallRequirement) -> None:
+        assert ireq.link is None, "This is a link, not a specifier"
+        self._ireq = install_req_drop_extras(ireq)
+        self._equal_cache: Optional[str] = None
+        self._hash: Optional[int] = None
+        self._extras = frozenset(canonicalize_name(e) for e in self._ireq.extras)
+
+    @property
+    def _equal(self) -> str:
+        if self._equal_cache is not None:
+            return self._equal_cache
+
+        self._equal_cache = str(self._ireq)
+        return self._equal_cache
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, SpecifierWithoutExtrasRequirement):
+            return NotImplemented
+        return self._equal == other._equal
+
+    def __hash__(self) -> int:
+        if self._hash is not None:
+            return self._hash
+
+        self._hash = hash(self._equal)
+        return self._hash
+
+
+class RequiresPythonRequirement(Requirement):
+    """A requirement representing Requires-Python metadata."""
+
+    def __init__(self, specifier: SpecifierSet, match: Candidate) -> None:
+        self.specifier = specifier
+        self._specifier_string = str(specifier)  # for faster __eq__
+        self._hash: Optional[int] = None
+        self._candidate = match
+
+    def __str__(self) -> str:
+        return f"Python {self.specifier}"
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({str(self.specifier)!r})"
+
+    def __hash__(self) -> int:
+        if self._hash is not None:
+            return self._hash
+
+        self._hash = hash((self._specifier_string, self._candidate))
+        return self._hash
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, RequiresPythonRequirement):
+            return False
+        return (
+            self._specifier_string == other._specifier_string
+            and self._candidate == other._candidate
+        )
+
+    @property
+    def project_name(self) -> NormalizedName:
+        return self._candidate.project_name
+
+    @property
+    def name(self) -> str:
+        return self._candidate.name
+
+    def format_for_error(self) -> str:
+        return str(self)
+
+    def get_candidate_lookup(self) -> CandidateLookup:
+        if self.specifier.contains(self._candidate.version, prereleases=True):
+            return self._candidate, None
+        return None, None
+
+    def is_satisfied_by(self, candidate: Candidate) -> bool:
+        assert candidate.name == self._candidate.name, "Not Python candidate"
+        # We can safely always allow prereleases here since PackageFinder
+        # already implements the prerelease logic, and would have filtered out
+        # prerelease candidates if the user does not expect them.
+        return self.specifier.contains(candidate.version, prereleases=True)
+
+
+class UnsatisfiableRequirement(Requirement):
+    """A requirement that cannot be satisfied."""
+
+    def __init__(self, name: NormalizedName) -> None:
+        self._name = name
+
+    def __str__(self) -> str:
+        return f"{self._name} (unavailable)"
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({str(self._name)!r})"
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, UnsatisfiableRequirement):
+            return NotImplemented
+        return self._name == other._name
+
+    def __hash__(self) -> int:
+        return hash(self._name)
+
+    @property
+    def project_name(self) -> NormalizedName:
+        return self._name
+
+    @property
+    def name(self) -> str:
+        return self._name
+
+    def format_for_error(self) -> str:
+        return str(self)
+
+    def get_candidate_lookup(self) -> CandidateLookup:
+        return None, None
+
+    def is_satisfied_by(self, candidate: Candidate) -> bool:
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/resolver.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/resolver.py
new file mode 100644
index 00000000..c12beef0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/resolution/resolvelib/resolver.py
@@ -0,0 +1,317 @@
+import contextlib
+import functools
+import logging
+import os
+from typing import TYPE_CHECKING, Dict, List, Optional, Set, Tuple, cast
+
+from pip._vendor.packaging.utils import canonicalize_name
+from pip._vendor.resolvelib import BaseReporter, ResolutionImpossible
+from pip._vendor.resolvelib import Resolver as RLResolver
+from pip._vendor.resolvelib.structs import DirectedGraph
+
+from pip._internal.cache import WheelCache
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.operations.prepare import RequirementPreparer
+from pip._internal.req.constructors import install_req_extend_extras
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.req.req_set import RequirementSet
+from pip._internal.resolution.base import BaseResolver, InstallRequirementProvider
+from pip._internal.resolution.resolvelib.provider import PipProvider
+from pip._internal.resolution.resolvelib.reporter import (
+    PipDebuggingReporter,
+    PipReporter,
+)
+from pip._internal.utils.packaging import get_requirement
+
+from .base import Candidate, Requirement
+from .factory import Factory
+
+if TYPE_CHECKING:
+    from pip._vendor.resolvelib.resolvers import Result as RLResult
+
+    Result = RLResult[Requirement, Candidate, str]
+
+
+logger = logging.getLogger(__name__)
+
+
+class Resolver(BaseResolver):
+    _allowed_strategies = {"eager", "only-if-needed", "to-satisfy-only"}
+
+    def __init__(
+        self,
+        preparer: RequirementPreparer,
+        finder: PackageFinder,
+        wheel_cache: Optional[WheelCache],
+        make_install_req: InstallRequirementProvider,
+        use_user_site: bool,
+        ignore_dependencies: bool,
+        ignore_installed: bool,
+        ignore_requires_python: bool,
+        force_reinstall: bool,
+        upgrade_strategy: str,
+        py_version_info: Optional[Tuple[int, ...]] = None,
+    ):
+        super().__init__()
+        assert upgrade_strategy in self._allowed_strategies
+
+        self.factory = Factory(
+            finder=finder,
+            preparer=preparer,
+            make_install_req=make_install_req,
+            wheel_cache=wheel_cache,
+            use_user_site=use_user_site,
+            force_reinstall=force_reinstall,
+            ignore_installed=ignore_installed,
+            ignore_requires_python=ignore_requires_python,
+            py_version_info=py_version_info,
+        )
+        self.ignore_dependencies = ignore_dependencies
+        self.upgrade_strategy = upgrade_strategy
+        self._result: Optional[Result] = None
+
+    def resolve(
+        self, root_reqs: List[InstallRequirement], check_supported_wheels: bool
+    ) -> RequirementSet:
+        collected = self.factory.collect_root_requirements(root_reqs)
+        provider = PipProvider(
+            factory=self.factory,
+            constraints=collected.constraints,
+            ignore_dependencies=self.ignore_dependencies,
+            upgrade_strategy=self.upgrade_strategy,
+            user_requested=collected.user_requested,
+        )
+        if "PIP_RESOLVER_DEBUG" in os.environ:
+            reporter: BaseReporter = PipDebuggingReporter()
+        else:
+            reporter = PipReporter()
+        resolver: RLResolver[Requirement, Candidate, str] = RLResolver(
+            provider,
+            reporter,
+        )
+
+        try:
+            limit_how_complex_resolution_can_be = 200000
+            result = self._result = resolver.resolve(
+                collected.requirements, max_rounds=limit_how_complex_resolution_can_be
+            )
+
+        except ResolutionImpossible as e:
+            error = self.factory.get_installation_error(
+                cast("ResolutionImpossible[Requirement, Candidate]", e),
+                collected.constraints,
+            )
+            raise error from e
+
+        req_set = RequirementSet(check_supported_wheels=check_supported_wheels)
+        # process candidates with extras last to ensure their base equivalent is
+        # already in the req_set if appropriate.
+        # Python's sort is stable so using a binary key function keeps relative order
+        # within both subsets.
+        for candidate in sorted(
+            result.mapping.values(), key=lambda c: c.name != c.project_name
+        ):
+            ireq = candidate.get_install_requirement()
+            if ireq is None:
+                if candidate.name != candidate.project_name:
+                    # extend existing req's extras
+                    with contextlib.suppress(KeyError):
+                        req = req_set.get_requirement(candidate.project_name)
+                        req_set.add_named_requirement(
+                            install_req_extend_extras(
+                                req, get_requirement(candidate.name).extras
+                            )
+                        )
+                continue
+
+            # Check if there is already an installation under the same name,
+            # and set a flag for later stages to uninstall it, if needed.
+            installed_dist = self.factory.get_dist_to_uninstall(candidate)
+            if installed_dist is None:
+                # There is no existing installation -- nothing to uninstall.
+                ireq.should_reinstall = False
+            elif self.factory.force_reinstall:
+                # The --force-reinstall flag is set -- reinstall.
+                ireq.should_reinstall = True
+            elif installed_dist.version != candidate.version:
+                # The installation is different in version -- reinstall.
+                ireq.should_reinstall = True
+            elif candidate.is_editable or installed_dist.editable:
+                # The incoming distribution is editable, or different in
+                # editable-ness to installation -- reinstall.
+                ireq.should_reinstall = True
+            elif candidate.source_link and candidate.source_link.is_file:
+                # The incoming distribution is under file://
+                if candidate.source_link.is_wheel:
+                    # is a local wheel -- do nothing.
+                    logger.info(
+                        "%s is already installed with the same version as the "
+                        "provided wheel. Use --force-reinstall to force an "
+                        "installation of the wheel.",
+                        ireq.name,
+                    )
+                    continue
+
+                # is a local sdist or path -- reinstall
+                ireq.should_reinstall = True
+            else:
+                continue
+
+            link = candidate.source_link
+            if link and link.is_yanked:
+                # The reason can contain non-ASCII characters, Unicode
+                # is required for Python 2.
+                msg = (
+                    "The candidate selected for download or install is a "
+                    "yanked version: {name!r} candidate (version {version} "
+                    "at {link})\nReason for being yanked: {reason}"
+                ).format(
+                    name=candidate.name,
+                    version=candidate.version,
+                    link=link,
+                    reason=link.yanked_reason or "<none given>",
+                )
+                logger.warning(msg)
+
+            req_set.add_named_requirement(ireq)
+
+        reqs = req_set.all_requirements
+        self.factory.preparer.prepare_linked_requirements_more(reqs)
+        for req in reqs:
+            req.prepared = True
+            req.needs_more_preparation = False
+        return req_set
+
+    def get_installation_order(
+        self, req_set: RequirementSet
+    ) -> List[InstallRequirement]:
+        """Get order for installation of requirements in RequirementSet.
+
+        The returned list contains a requirement before another that depends on
+        it. This helps ensure that the environment is kept consistent as they
+        get installed one-by-one.
+
+        The current implementation creates a topological ordering of the
+        dependency graph, giving more weight to packages with less
+        or no dependencies, while breaking any cycles in the graph at
+        arbitrary points. We make no guarantees about where the cycle
+        would be broken, other than it *would* be broken.
+        """
+        assert self._result is not None, "must call resolve() first"
+
+        if not req_set.requirements:
+            # Nothing is left to install, so we do not need an order.
+            return []
+
+        graph = self._result.graph
+        weights = get_topological_weights(graph, set(req_set.requirements.keys()))
+
+        sorted_items = sorted(
+            req_set.requirements.items(),
+            key=functools.partial(_req_set_item_sorter, weights=weights),
+            reverse=True,
+        )
+        return [ireq for _, ireq in sorted_items]
+
+
+def get_topological_weights(
+    graph: "DirectedGraph[Optional[str]]", requirement_keys: Set[str]
+) -> Dict[Optional[str], int]:
+    """Assign weights to each node based on how "deep" they are.
+
+    This implementation may change at any point in the future without prior
+    notice.
+
+    We first simplify the dependency graph by pruning any leaves and giving them
+    the highest weight: a package without any dependencies should be installed
+    first. This is done again and again in the same way, giving ever less weight
+    to the newly found leaves. The loop stops when no leaves are left: all
+    remaining packages have at least one dependency left in the graph.
+
+    Then we continue with the remaining graph, by taking the length for the
+    longest path to any node from root, ignoring any paths that contain a single
+    node twice (i.e. cycles). This is done through a depth-first search through
+    the graph, while keeping track of the path to the node.
+
+    Cycles in the graph result would result in node being revisited while also
+    being on its own path. In this case, take no action. This helps ensure we
+    don't get stuck in a cycle.
+
+    When assigning weight, the longer path (i.e. larger length) is preferred.
+
+    We are only interested in the weights of packages that are in the
+    requirement_keys.
+    """
+    path: Set[Optional[str]] = set()
+    weights: Dict[Optional[str], int] = {}
+
+    def visit(node: Optional[str]) -> None:
+        if node in path:
+            # We hit a cycle, so we'll break it here.
+            return
+
+        # Time to visit the children!
+        path.add(node)
+        for child in graph.iter_children(node):
+            visit(child)
+        path.remove(node)
+
+        if node not in requirement_keys:
+            return
+
+        last_known_parent_count = weights.get(node, 0)
+        weights[node] = max(last_known_parent_count, len(path))
+
+    # Simplify the graph, pruning leaves that have no dependencies.
+    # This is needed for large graphs (say over 200 packages) because the
+    # `visit` function is exponentially slower then, taking minutes.
+    # See https://github.com/pypa/pip/issues/10557
+    # We will loop until we explicitly break the loop.
+    while True:
+        leaves = set()
+        for key in graph:
+            if key is None:
+                continue
+            for _child in graph.iter_children(key):
+                # This means we have at least one child
+                break
+            else:
+                # No child.
+                leaves.add(key)
+        if not leaves:
+            # We are done simplifying.
+            break
+        # Calculate the weight for the leaves.
+        weight = len(graph) - 1
+        for leaf in leaves:
+            if leaf not in requirement_keys:
+                continue
+            weights[leaf] = weight
+        # Remove the leaves from the graph, making it simpler.
+        for leaf in leaves:
+            graph.remove(leaf)
+
+    # Visit the remaining graph.
+    # `None` is guaranteed to be the root node by resolvelib.
+    visit(None)
+
+    # Sanity check: all requirement keys should be in the weights,
+    # and no other keys should be in the weights.
+    difference = set(weights.keys()).difference(requirement_keys)
+    assert not difference, difference
+
+    return weights
+
+
+def _req_set_item_sorter(
+    item: Tuple[str, InstallRequirement],
+    weights: Dict[Optional[str], int],
+) -> Tuple[int, str]:
+    """Key function used to sort install requirements for installation.
+
+    Based on the "weight" mapping calculated in ``get_installation_order()``.
+    The canonical package name is returned as the second member as a tie-
+    breaker to ensure the result is predictable, which is useful in tests.
+    """
+    name = canonicalize_name(item[0])
+    return weights[name], name
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/self_outdated_check.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/self_outdated_check.py
new file mode 100644
index 00000000..2e0e3df3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/self_outdated_check.py
@@ -0,0 +1,252 @@
+import datetime
+import functools
+import hashlib
+import json
+import logging
+import optparse
+import os.path
+import sys
+from dataclasses import dataclass
+from typing import Any, Callable, Dict, Optional
+
+from pip._vendor.packaging.version import Version
+from pip._vendor.packaging.version import parse as parse_version
+from pip._vendor.rich.console import Group
+from pip._vendor.rich.markup import escape
+from pip._vendor.rich.text import Text
+
+from pip._internal.index.collector import LinkCollector
+from pip._internal.index.package_finder import PackageFinder
+from pip._internal.metadata import get_default_environment
+from pip._internal.models.selection_prefs import SelectionPreferences
+from pip._internal.network.session import PipSession
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.entrypoints import (
+    get_best_invocation_for_this_pip,
+    get_best_invocation_for_this_python,
+)
+from pip._internal.utils.filesystem import adjacent_tmp_file, check_path_owner, replace
+from pip._internal.utils.misc import (
+    ExternallyManagedEnvironment,
+    check_externally_managed,
+    ensure_dir,
+)
+
+_WEEK = datetime.timedelta(days=7)
+
+logger = logging.getLogger(__name__)
+
+
+def _get_statefile_name(key: str) -> str:
+    key_bytes = key.encode()
+    name = hashlib.sha224(key_bytes).hexdigest()
+    return name
+
+
+def _convert_date(isodate: str) -> datetime.datetime:
+    """Convert an ISO format string to a date.
+
+    Handles the format 2020-01-22T14:24:01Z (trailing Z)
+    which is not supported by older versions of fromisoformat.
+    """
+    return datetime.datetime.fromisoformat(isodate.replace("Z", "+00:00"))
+
+
+class SelfCheckState:
+    def __init__(self, cache_dir: str) -> None:
+        self._state: Dict[str, Any] = {}
+        self._statefile_path = None
+
+        # Try to load the existing state
+        if cache_dir:
+            self._statefile_path = os.path.join(
+                cache_dir, "selfcheck", _get_statefile_name(self.key)
+            )
+            try:
+                with open(self._statefile_path, encoding="utf-8") as statefile:
+                    self._state = json.load(statefile)
+            except (OSError, ValueError, KeyError):
+                # Explicitly suppressing exceptions, since we don't want to
+                # error out if the cache file is invalid.
+                pass
+
+    @property
+    def key(self) -> str:
+        return sys.prefix
+
+    def get(self, current_time: datetime.datetime) -> Optional[str]:
+        """Check if we have a not-outdated version loaded already."""
+        if not self._state:
+            return None
+
+        if "last_check" not in self._state:
+            return None
+
+        if "pypi_version" not in self._state:
+            return None
+
+        # Determine if we need to refresh the state
+        last_check = _convert_date(self._state["last_check"])
+        time_since_last_check = current_time - last_check
+        if time_since_last_check > _WEEK:
+            return None
+
+        return self._state["pypi_version"]
+
+    def set(self, pypi_version: str, current_time: datetime.datetime) -> None:
+        # If we do not have a path to cache in, don't bother saving.
+        if not self._statefile_path:
+            return
+
+        # Check to make sure that we own the directory
+        if not check_path_owner(os.path.dirname(self._statefile_path)):
+            return
+
+        # Now that we've ensured the directory is owned by this user, we'll go
+        # ahead and make sure that all our directories are created.
+        ensure_dir(os.path.dirname(self._statefile_path))
+
+        state = {
+            # Include the key so it's easy to tell which pip wrote the
+            # file.
+            "key": self.key,
+            "last_check": current_time.isoformat(),
+            "pypi_version": pypi_version,
+        }
+
+        text = json.dumps(state, sort_keys=True, separators=(",", ":"))
+
+        with adjacent_tmp_file(self._statefile_path) as f:
+            f.write(text.encode())
+
+        try:
+            # Since we have a prefix-specific state file, we can just
+            # overwrite whatever is there, no need to check.
+            replace(f.name, self._statefile_path)
+        except OSError:
+            # Best effort.
+            pass
+
+
+@dataclass
+class UpgradePrompt:
+    old: str
+    new: str
+
+    def __rich__(self) -> Group:
+        if WINDOWS:
+            pip_cmd = f"{get_best_invocation_for_this_python()} -m pip"
+        else:
+            pip_cmd = get_best_invocation_for_this_pip()
+
+        notice = "[bold][[reset][blue]notice[reset][bold]][reset]"
+        return Group(
+            Text(),
+            Text.from_markup(
+                f"{notice} A new release of pip is available: "
+                f"[red]{self.old}[reset] -> [green]{self.new}[reset]"
+            ),
+            Text.from_markup(
+                f"{notice} To update, run: "
+                f"[green]{escape(pip_cmd)} install --upgrade pip"
+            ),
+        )
+
+
+def was_installed_by_pip(pkg: str) -> bool:
+    """Checks whether pkg was installed by pip
+
+    This is used not to display the upgrade message when pip is in fact
+    installed by system package manager, such as dnf on Fedora.
+    """
+    dist = get_default_environment().get_distribution(pkg)
+    return dist is not None and "pip" == dist.installer
+
+
+def _get_current_remote_pip_version(
+    session: PipSession, options: optparse.Values
+) -> Optional[str]:
+    # Lets use PackageFinder to see what the latest pip version is
+    link_collector = LinkCollector.create(
+        session,
+        options=options,
+        suppress_no_index=True,
+    )
+
+    # Pass allow_yanked=False so we don't suggest upgrading to a
+    # yanked version.
+    selection_prefs = SelectionPreferences(
+        allow_yanked=False,
+        allow_all_prereleases=False,  # Explicitly set to False
+    )
+
+    finder = PackageFinder.create(
+        link_collector=link_collector,
+        selection_prefs=selection_prefs,
+    )
+    best_candidate = finder.find_best_candidate("pip").best_candidate
+    if best_candidate is None:
+        return None
+
+    return str(best_candidate.version)
+
+
+def _self_version_check_logic(
+    *,
+    state: SelfCheckState,
+    current_time: datetime.datetime,
+    local_version: Version,
+    get_remote_version: Callable[[], Optional[str]],
+) -> Optional[UpgradePrompt]:
+    remote_version_str = state.get(current_time)
+    if remote_version_str is None:
+        remote_version_str = get_remote_version()
+        if remote_version_str is None:
+            logger.debug("No remote pip version found")
+            return None
+        state.set(remote_version_str, current_time)
+
+    remote_version = parse_version(remote_version_str)
+    logger.debug("Remote version of pip: %s", remote_version)
+    logger.debug("Local version of pip:  %s", local_version)
+
+    pip_installed_by_pip = was_installed_by_pip("pip")
+    logger.debug("Was pip installed by pip? %s", pip_installed_by_pip)
+    if not pip_installed_by_pip:
+        return None  # Only suggest upgrade if pip is installed by pip.
+
+    local_version_is_older = (
+        local_version < remote_version
+        and local_version.base_version != remote_version.base_version
+    )
+    if local_version_is_older:
+        return UpgradePrompt(old=str(local_version), new=remote_version_str)
+
+    return None
+
+
+def pip_self_version_check(session: PipSession, options: optparse.Values) -> None:
+    """Check for an update for pip.
+
+    Limit the frequency of checks to once per week. State is stored either in
+    the active virtualenv or in the user's USER_CACHE_DIR keyed off the prefix
+    of the pip script path.
+    """
+    installed_dist = get_default_environment().get_distribution("pip")
+    if not installed_dist:
+        return
+    try:
+        check_externally_managed()
+    except ExternallyManagedEnvironment:
+        return
+
+    upgrade_prompt = _self_version_check_logic(
+        state=SelfCheckState(cache_dir=options.cache_dir),
+        current_time=datetime.datetime.now(datetime.timezone.utc),
+        local_version=installed_dist.version,
+        get_remote_version=functools.partial(
+            _get_current_remote_pip_version, session, options
+        ),
+    )
+    if upgrade_prompt is not None:
+        logger.warning("%s", upgrade_prompt, extra={"rich": True})
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2948b6f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_jaraco_text.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_jaraco_text.cpython-312.pyc
new file mode 100644
index 00000000..e518e86c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_jaraco_text.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_log.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_log.cpython-312.pyc
new file mode 100644
index 00000000..ce4892fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/_log.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/appdirs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/appdirs.cpython-312.pyc
new file mode 100644
index 00000000..5cfd8b7e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/appdirs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..14db0b94
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compatibility_tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compatibility_tags.cpython-312.pyc
new file mode 100644
index 00000000..26d08c03
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/compatibility_tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/datetime.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/datetime.cpython-312.pyc
new file mode 100644
index 00000000..1ba39d60
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/datetime.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/deprecation.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/deprecation.cpython-312.pyc
new file mode 100644
index 00000000..ad7c5f19
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/deprecation.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/direct_url_helpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/direct_url_helpers.cpython-312.pyc
new file mode 100644
index 00000000..9d8cda02
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/direct_url_helpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/egg_link.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/egg_link.cpython-312.pyc
new file mode 100644
index 00000000..4ee8c413
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/egg_link.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/entrypoints.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/entrypoints.cpython-312.pyc
new file mode 100644
index 00000000..97d50ca6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/entrypoints.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filesystem.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filesystem.cpython-312.pyc
new file mode 100644
index 00000000..806a5a25
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filesystem.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filetypes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filetypes.cpython-312.pyc
new file mode 100644
index 00000000..f3f8c496
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/filetypes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/glibc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/glibc.cpython-312.pyc
new file mode 100644
index 00000000..8bedb9c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/glibc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/hashes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/hashes.cpython-312.pyc
new file mode 100644
index 00000000..f6da73b8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/hashes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/logging.cpython-312.pyc
new file mode 100644
index 00000000..99478ff7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/misc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/misc.cpython-312.pyc
new file mode 100644
index 00000000..44ee595a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/misc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/packaging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/packaging.cpython-312.pyc
new file mode 100644
index 00000000..7526f59f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/packaging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/retry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/retry.cpython-312.pyc
new file mode 100644
index 00000000..2c5a431d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/retry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/setuptools_build.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/setuptools_build.cpython-312.pyc
new file mode 100644
index 00000000..066bf8f5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/setuptools_build.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/subprocess.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/subprocess.cpython-312.pyc
new file mode 100644
index 00000000..c96feae3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/subprocess.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/temp_dir.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/temp_dir.cpython-312.pyc
new file mode 100644
index 00000000..5e760910
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/temp_dir.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/unpacking.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/unpacking.cpython-312.pyc
new file mode 100644
index 00000000..fb6febf7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/unpacking.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/urls.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/urls.cpython-312.pyc
new file mode 100644
index 00000000..82eef8ec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/urls.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/virtualenv.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/virtualenv.cpython-312.pyc
new file mode 100644
index 00000000..ced57e41
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/virtualenv.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..c8594cf9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_jaraco_text.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_jaraco_text.py
new file mode 100644
index 00000000..6ccf53b7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_jaraco_text.py
@@ -0,0 +1,109 @@
+"""Functions brought over from jaraco.text.
+
+These functions are not supposed to be used within `pip._internal`. These are
+helper functions brought over from `jaraco.text` to enable vendoring newer
+copies of `pkg_resources` without having to vendor `jaraco.text` and its entire
+dependency cone; something that our vendoring setup is not currently capable of
+handling.
+
+License reproduced from original source below:
+
+Copyright Jason R. Coombs
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
+"""
+
+import functools
+import itertools
+
+
+def _nonblank(str):
+    return str and not str.startswith("#")
+
+
+@functools.singledispatch
+def yield_lines(iterable):
+    r"""
+    Yield valid lines of a string or iterable.
+
+    >>> list(yield_lines(''))
+    []
+    >>> list(yield_lines(['foo', 'bar']))
+    ['foo', 'bar']
+    >>> list(yield_lines('foo\nbar'))
+    ['foo', 'bar']
+    >>> list(yield_lines('\nfoo\n#bar\nbaz #comment'))
+    ['foo', 'baz #comment']
+    >>> list(yield_lines(['foo\nbar', 'baz', 'bing\n\n\n']))
+    ['foo', 'bar', 'baz', 'bing']
+    """
+    return itertools.chain.from_iterable(map(yield_lines, iterable))
+
+
+@yield_lines.register(str)
+def _(text):
+    return filter(_nonblank, map(str.strip, text.splitlines()))
+
+
+def drop_comment(line):
+    """
+    Drop comments.
+
+    >>> drop_comment('foo # bar')
+    'foo'
+
+    A hash without a space may be in a URL.
+
+    >>> drop_comment('http://example.com/foo#bar')
+    'http://example.com/foo#bar'
+    """
+    return line.partition(" #")[0]
+
+
+def join_continuation(lines):
+    r"""
+    Join lines continued by a trailing backslash.
+
+    >>> list(join_continuation(['foo \\', 'bar', 'baz']))
+    ['foobar', 'baz']
+    >>> list(join_continuation(['foo \\', 'bar', 'baz']))
+    ['foobar', 'baz']
+    >>> list(join_continuation(['foo \\', 'bar \\', 'baz']))
+    ['foobarbaz']
+
+    Not sure why, but...
+    The character preceding the backslash is also elided.
+
+    >>> list(join_continuation(['goo\\', 'dly']))
+    ['godly']
+
+    A terrible idea, but...
+    If no line is available to continue, suppress the lines.
+
+    >>> list(join_continuation(['foo', 'bar\\', 'baz\\']))
+    ['foo']
+    """
+    lines = iter(lines)
+    for item in lines:
+        while item.endswith("\\"):
+            try:
+                item = item[:-2].strip() + next(lines)
+            except StopIteration:
+                return
+        yield item
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_log.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_log.py
new file mode 100644
index 00000000..92c4c6a1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/_log.py
@@ -0,0 +1,38 @@
+"""Customize logging
+
+Defines custom logger class for the `logger.verbose(...)` method.
+
+init_logging() must be called before any other modules that call logging.getLogger.
+"""
+
+import logging
+from typing import Any, cast
+
+# custom log level for `--verbose` output
+# between DEBUG and INFO
+VERBOSE = 15
+
+
+class VerboseLogger(logging.Logger):
+    """Custom Logger, defining a verbose log-level
+
+    VERBOSE is between INFO and DEBUG.
+    """
+
+    def verbose(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        return self.log(VERBOSE, msg, *args, **kwargs)
+
+
+def getLogger(name: str) -> VerboseLogger:
+    """logging.getLogger, but ensures our VerboseLogger class is returned"""
+    return cast(VerboseLogger, logging.getLogger(name))
+
+
+def init_logging() -> None:
+    """Register our VerboseLogger and VERBOSE log level.
+
+    Should be called before any calls to getLogger(),
+    i.e. in pip._internal.__init__
+    """
+    logging.setLoggerClass(VerboseLogger)
+    logging.addLevelName(VERBOSE, "VERBOSE")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/appdirs.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/appdirs.py
new file mode 100644
index 00000000..16933bf8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/appdirs.py
@@ -0,0 +1,52 @@
+"""
+This code wraps the vendored appdirs module to so the return values are
+compatible for the current pip code base.
+
+The intention is to rewrite current usages gradually, keeping the tests pass,
+and eventually drop this after all usages are changed.
+"""
+
+import os
+import sys
+from typing import List
+
+from pip._vendor import platformdirs as _appdirs
+
+
+def user_cache_dir(appname: str) -> str:
+    return _appdirs.user_cache_dir(appname, appauthor=False)
+
+
+def _macos_user_config_dir(appname: str, roaming: bool = True) -> str:
+    # Use ~/Application Support/pip, if the directory exists.
+    path = _appdirs.user_data_dir(appname, appauthor=False, roaming=roaming)
+    if os.path.isdir(path):
+        return path
+
+    # Use a Linux-like ~/.config/pip, by default.
+    linux_like_path = "~/.config/"
+    if appname:
+        linux_like_path = os.path.join(linux_like_path, appname)
+
+    return os.path.expanduser(linux_like_path)
+
+
+def user_config_dir(appname: str, roaming: bool = True) -> str:
+    if sys.platform == "darwin":
+        return _macos_user_config_dir(appname, roaming)
+
+    return _appdirs.user_config_dir(appname, appauthor=False, roaming=roaming)
+
+
+# for the discussion regarding site_config_dir locations
+# see <https://github.com/pypa/pip/issues/1733>
+def site_config_dirs(appname: str) -> List[str]:
+    if sys.platform == "darwin":
+        return [_appdirs.site_data_dir(appname, appauthor=False, multipath=True)]
+
+    dirval = _appdirs.site_config_dir(appname, appauthor=False, multipath=True)
+    if sys.platform == "win32":
+        return [dirval]
+
+    # Unix-y system. Look in /etc as well.
+    return dirval.split(os.pathsep) + ["/etc"]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compat.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compat.py
new file mode 100644
index 00000000..d8b54e4e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compat.py
@@ -0,0 +1,79 @@
+"""Stuff that differs in different Python versions and platform
+distributions."""
+
+import importlib.resources
+import logging
+import os
+import sys
+from typing import IO
+
+__all__ = ["get_path_uid", "stdlib_pkgs", "WINDOWS"]
+
+
+logger = logging.getLogger(__name__)
+
+
+def has_tls() -> bool:
+    try:
+        import _ssl  # noqa: F401  # ignore unused
+
+        return True
+    except ImportError:
+        pass
+
+    from pip._vendor.urllib3.util import IS_PYOPENSSL
+
+    return IS_PYOPENSSL
+
+
+def get_path_uid(path: str) -> int:
+    """
+    Return path's uid.
+
+    Does not follow symlinks:
+        https://github.com/pypa/pip/pull/935#discussion_r5307003
+
+    Placed this function in compat due to differences on AIX and
+    Jython, that should eventually go away.
+
+    :raises OSError: When path is a symlink or can't be read.
+    """
+    if hasattr(os, "O_NOFOLLOW"):
+        fd = os.open(path, os.O_RDONLY | os.O_NOFOLLOW)
+        file_uid = os.fstat(fd).st_uid
+        os.close(fd)
+    else:  # AIX and Jython
+        # WARNING: time of check vulnerability, but best we can do w/o NOFOLLOW
+        if not os.path.islink(path):
+            # older versions of Jython don't have `os.fstat`
+            file_uid = os.stat(path).st_uid
+        else:
+            # raise OSError for parity with os.O_NOFOLLOW above
+            raise OSError(f"{path} is a symlink; Will not return uid for symlinks")
+    return file_uid
+
+
+# The importlib.resources.open_text function was deprecated in 3.11 with suggested
+# replacement we use below.
+if sys.version_info < (3, 11):
+    open_text_resource = importlib.resources.open_text
+else:
+
+    def open_text_resource(
+        package: str, resource: str, encoding: str = "utf-8", errors: str = "strict"
+    ) -> IO[str]:
+        return (importlib.resources.files(package) / resource).open(
+            "r", encoding=encoding, errors=errors
+        )
+
+
+# packages in the stdlib that may have installation metadata, but should not be
+# considered 'installed'.  this theoretically could be determined based on
+# dist.location (py27:`sysconfig.get_paths()['stdlib']`,
+# py26:sysconfig.get_config_vars('LIBDEST')), but fear platform variation may
+# make this ineffective, so hard-coding
+stdlib_pkgs = {"python", "wsgiref", "argparse"}
+
+
+# windows detection, covers cpython and ironpython
+WINDOWS = sys.platform.startswith("win") or (sys.platform == "cli" and os.name == "nt")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compatibility_tags.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compatibility_tags.py
new file mode 100644
index 00000000..2e7b7450
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/compatibility_tags.py
@@ -0,0 +1,188 @@
+"""Generate and work with PEP 425 Compatibility Tags.
+"""
+
+import re
+from typing import List, Optional, Tuple
+
+from pip._vendor.packaging.tags import (
+    PythonVersion,
+    Tag,
+    compatible_tags,
+    cpython_tags,
+    generic_tags,
+    interpreter_name,
+    interpreter_version,
+    ios_platforms,
+    mac_platforms,
+)
+
+_apple_arch_pat = re.compile(r"(.+)_(\d+)_(\d+)_(.+)")
+
+
+def version_info_to_nodot(version_info: Tuple[int, ...]) -> str:
+    # Only use up to the first two numbers.
+    return "".join(map(str, version_info[:2]))
+
+
+def _mac_platforms(arch: str) -> List[str]:
+    match = _apple_arch_pat.match(arch)
+    if match:
+        name, major, minor, actual_arch = match.groups()
+        mac_version = (int(major), int(minor))
+        arches = [
+            # Since we have always only checked that the platform starts
+            # with "macosx", for backwards-compatibility we extract the
+            # actual prefix provided by the user in case they provided
+            # something like "macosxcustom_". It may be good to remove
+            # this as undocumented or deprecate it in the future.
+            "{}_{}".format(name, arch[len("macosx_") :])
+            for arch in mac_platforms(mac_version, actual_arch)
+        ]
+    else:
+        # arch pattern didn't match (?!)
+        arches = [arch]
+    return arches
+
+
+def _ios_platforms(arch: str) -> List[str]:
+    match = _apple_arch_pat.match(arch)
+    if match:
+        name, major, minor, actual_multiarch = match.groups()
+        ios_version = (int(major), int(minor))
+        arches = [
+            # Since we have always only checked that the platform starts
+            # with "ios", for backwards-compatibility we extract the
+            # actual prefix provided by the user in case they provided
+            # something like "ioscustom_". It may be good to remove
+            # this as undocumented or deprecate it in the future.
+            "{}_{}".format(name, arch[len("ios_") :])
+            for arch in ios_platforms(ios_version, actual_multiarch)
+        ]
+    else:
+        # arch pattern didn't match (?!)
+        arches = [arch]
+    return arches
+
+
+def _custom_manylinux_platforms(arch: str) -> List[str]:
+    arches = [arch]
+    arch_prefix, arch_sep, arch_suffix = arch.partition("_")
+    if arch_prefix == "manylinux2014":
+        # manylinux1/manylinux2010 wheels run on most manylinux2014 systems
+        # with the exception of wheels depending on ncurses. PEP 599 states
+        # manylinux1/manylinux2010 wheels should be considered
+        # manylinux2014 wheels:
+        # https://www.python.org/dev/peps/pep-0599/#backwards-compatibility-with-manylinux2010-wheels
+        if arch_suffix in {"i686", "x86_64"}:
+            arches.append("manylinux2010" + arch_sep + arch_suffix)
+            arches.append("manylinux1" + arch_sep + arch_suffix)
+    elif arch_prefix == "manylinux2010":
+        # manylinux1 wheels run on most manylinux2010 systems with the
+        # exception of wheels depending on ncurses. PEP 571 states
+        # manylinux1 wheels should be considered manylinux2010 wheels:
+        # https://www.python.org/dev/peps/pep-0571/#backwards-compatibility-with-manylinux1-wheels
+        arches.append("manylinux1" + arch_sep + arch_suffix)
+    return arches
+
+
+def _get_custom_platforms(arch: str) -> List[str]:
+    arch_prefix, arch_sep, arch_suffix = arch.partition("_")
+    if arch.startswith("macosx"):
+        arches = _mac_platforms(arch)
+    elif arch.startswith("ios"):
+        arches = _ios_platforms(arch)
+    elif arch_prefix in ["manylinux2014", "manylinux2010"]:
+        arches = _custom_manylinux_platforms(arch)
+    else:
+        arches = [arch]
+    return arches
+
+
+def _expand_allowed_platforms(platforms: Optional[List[str]]) -> Optional[List[str]]:
+    if not platforms:
+        return None
+
+    seen = set()
+    result = []
+
+    for p in platforms:
+        if p in seen:
+            continue
+        additions = [c for c in _get_custom_platforms(p) if c not in seen]
+        seen.update(additions)
+        result.extend(additions)
+
+    return result
+
+
+def _get_python_version(version: str) -> PythonVersion:
+    if len(version) > 1:
+        return int(version[0]), int(version[1:])
+    else:
+        return (int(version[0]),)
+
+
+def _get_custom_interpreter(
+    implementation: Optional[str] = None, version: Optional[str] = None
+) -> str:
+    if implementation is None:
+        implementation = interpreter_name()
+    if version is None:
+        version = interpreter_version()
+    return f"{implementation}{version}"
+
+
+def get_supported(
+    version: Optional[str] = None,
+    platforms: Optional[List[str]] = None,
+    impl: Optional[str] = None,
+    abis: Optional[List[str]] = None,
+) -> List[Tag]:
+    """Return a list of supported tags for each version specified in
+    `versions`.
+
+    :param version: a string version, of the form "33" or "32",
+        or None. The version will be assumed to support our ABI.
+    :param platform: specify a list of platforms you want valid
+        tags for, or None. If None, use the local system platform.
+    :param impl: specify the exact implementation you want valid
+        tags for, or None. If None, use the local interpreter impl.
+    :param abis: specify a list of abis you want valid
+        tags for, or None. If None, use the local interpreter abi.
+    """
+    supported: List[Tag] = []
+
+    python_version: Optional[PythonVersion] = None
+    if version is not None:
+        python_version = _get_python_version(version)
+
+    interpreter = _get_custom_interpreter(impl, version)
+
+    platforms = _expand_allowed_platforms(platforms)
+
+    is_cpython = (impl or interpreter_name()) == "cp"
+    if is_cpython:
+        supported.extend(
+            cpython_tags(
+                python_version=python_version,
+                abis=abis,
+                platforms=platforms,
+            )
+        )
+    else:
+        supported.extend(
+            generic_tags(
+                interpreter=interpreter,
+                abis=abis,
+                platforms=platforms,
+            )
+        )
+    supported.extend(
+        compatible_tags(
+            python_version=python_version,
+            interpreter=interpreter,
+            platforms=platforms,
+        )
+    )
+
+    return supported
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/datetime.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/datetime.py
new file mode 100644
index 00000000..8668b3b0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/datetime.py
@@ -0,0 +1,11 @@
+"""For when pip wants to check the date or time.
+"""
+
+import datetime
+
+
+def today_is_later_than(year: int, month: int, day: int) -> bool:
+    today = datetime.date.today()
+    given = datetime.date(year, month, day)
+
+    return today > given
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/deprecation.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/deprecation.py
new file mode 100644
index 00000000..0911147e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/deprecation.py
@@ -0,0 +1,124 @@
+"""
+A module that implements tooling to enable easy warnings about deprecations.
+"""
+
+import logging
+import warnings
+from typing import Any, Optional, TextIO, Type, Union
+
+from pip._vendor.packaging.version import parse
+
+from pip import __version__ as current_version  # NOTE: tests patch this name.
+
+DEPRECATION_MSG_PREFIX = "DEPRECATION: "
+
+
+class PipDeprecationWarning(Warning):
+    pass
+
+
+_original_showwarning: Any = None
+
+
+# Warnings <-> Logging Integration
+def _showwarning(
+    message: Union[Warning, str],
+    category: Type[Warning],
+    filename: str,
+    lineno: int,
+    file: Optional[TextIO] = None,
+    line: Optional[str] = None,
+) -> None:
+    if file is not None:
+        if _original_showwarning is not None:
+            _original_showwarning(message, category, filename, lineno, file, line)
+    elif issubclass(category, PipDeprecationWarning):
+        # We use a specially named logger which will handle all of the
+        # deprecation messages for pip.
+        logger = logging.getLogger("pip._internal.deprecations")
+        logger.warning(message)
+    else:
+        _original_showwarning(message, category, filename, lineno, file, line)
+
+
+def install_warning_logger() -> None:
+    # Enable our Deprecation Warnings
+    warnings.simplefilter("default", PipDeprecationWarning, append=True)
+
+    global _original_showwarning
+
+    if _original_showwarning is None:
+        _original_showwarning = warnings.showwarning
+        warnings.showwarning = _showwarning
+
+
+def deprecated(
+    *,
+    reason: str,
+    replacement: Optional[str],
+    gone_in: Optional[str],
+    feature_flag: Optional[str] = None,
+    issue: Optional[int] = None,
+) -> None:
+    """Helper to deprecate existing functionality.
+
+    reason:
+        Textual reason shown to the user about why this functionality has
+        been deprecated. Should be a complete sentence.
+    replacement:
+        Textual suggestion shown to the user about what alternative
+        functionality they can use.
+    gone_in:
+        The version of pip does this functionality should get removed in.
+        Raises an error if pip's current version is greater than or equal to
+        this.
+    feature_flag:
+        Command-line flag of the form --use-feature={feature_flag} for testing
+        upcoming functionality.
+    issue:
+        Issue number on the tracker that would serve as a useful place for
+        users to find related discussion and provide feedback.
+    """
+
+    # Determine whether or not the feature is already gone in this version.
+    is_gone = gone_in is not None and parse(current_version) >= parse(gone_in)
+
+    message_parts = [
+        (reason, f"{DEPRECATION_MSG_PREFIX}{{}}"),
+        (
+            gone_in,
+            (
+                "pip {} will enforce this behaviour change."
+                if not is_gone
+                else "Since pip {}, this is no longer supported."
+            ),
+        ),
+        (
+            replacement,
+            "A possible replacement is {}.",
+        ),
+        (
+            feature_flag,
+            (
+                "You can use the flag --use-feature={} to test the upcoming behaviour."
+                if not is_gone
+                else None
+            ),
+        ),
+        (
+            issue,
+            "Discussion can be found at https://github.com/pypa/pip/issues/{}",
+        ),
+    ]
+
+    message = " ".join(
+        format_str.format(value)
+        for value, format_str in message_parts
+        if format_str is not None and value is not None
+    )
+
+    # Raise as an error if this behaviour is deprecated.
+    if is_gone:
+        raise PipDeprecationWarning(message)
+
+    warnings.warn(message, category=PipDeprecationWarning, stacklevel=2)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/direct_url_helpers.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/direct_url_helpers.py
new file mode 100644
index 00000000..66020d39
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/direct_url_helpers.py
@@ -0,0 +1,87 @@
+from typing import Optional
+
+from pip._internal.models.direct_url import ArchiveInfo, DirectUrl, DirInfo, VcsInfo
+from pip._internal.models.link import Link
+from pip._internal.utils.urls import path_to_url
+from pip._internal.vcs import vcs
+
+
+def direct_url_as_pep440_direct_reference(direct_url: DirectUrl, name: str) -> str:
+    """Convert a DirectUrl to a pip requirement string."""
+    direct_url.validate()  # if invalid, this is a pip bug
+    requirement = name + " @ "
+    fragments = []
+    if isinstance(direct_url.info, VcsInfo):
+        requirement += (
+            f"{direct_url.info.vcs}+{direct_url.url}@{direct_url.info.commit_id}"
+        )
+    elif isinstance(direct_url.info, ArchiveInfo):
+        requirement += direct_url.url
+        if direct_url.info.hash:
+            fragments.append(direct_url.info.hash)
+    else:
+        assert isinstance(direct_url.info, DirInfo)
+        requirement += direct_url.url
+    if direct_url.subdirectory:
+        fragments.append("subdirectory=" + direct_url.subdirectory)
+    if fragments:
+        requirement += "#" + "&".join(fragments)
+    return requirement
+
+
+def direct_url_for_editable(source_dir: str) -> DirectUrl:
+    return DirectUrl(
+        url=path_to_url(source_dir),
+        info=DirInfo(editable=True),
+    )
+
+
+def direct_url_from_link(
+    link: Link, source_dir: Optional[str] = None, link_is_in_wheel_cache: bool = False
+) -> DirectUrl:
+    if link.is_vcs:
+        vcs_backend = vcs.get_backend_for_scheme(link.scheme)
+        assert vcs_backend
+        url, requested_revision, _ = vcs_backend.get_url_rev_and_auth(
+            link.url_without_fragment
+        )
+        # For VCS links, we need to find out and add commit_id.
+        if link_is_in_wheel_cache:
+            # If the requested VCS link corresponds to a cached
+            # wheel, it means the requested revision was an
+            # immutable commit hash, otherwise it would not have
+            # been cached. In that case we don't have a source_dir
+            # with the VCS checkout.
+            assert requested_revision
+            commit_id = requested_revision
+        else:
+            # If the wheel was not in cache, it means we have
+            # had to checkout from VCS to build and we have a source_dir
+            # which we can inspect to find out the commit id.
+            assert source_dir
+            commit_id = vcs_backend.get_revision(source_dir)
+        return DirectUrl(
+            url=url,
+            info=VcsInfo(
+                vcs=vcs_backend.name,
+                commit_id=commit_id,
+                requested_revision=requested_revision,
+            ),
+            subdirectory=link.subdirectory_fragment,
+        )
+    elif link.is_existing_dir():
+        return DirectUrl(
+            url=link.url_without_fragment,
+            info=DirInfo(),
+            subdirectory=link.subdirectory_fragment,
+        )
+    else:
+        hash = None
+        hash_name = link.hash_name
+        if hash_name:
+            hash = f"{hash_name}={link.hash}"
+        return DirectUrl(
+            url=link.url_without_fragment,
+            info=ArchiveInfo(hash=hash),
+            subdirectory=link.subdirectory_fragment,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/egg_link.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/egg_link.py
new file mode 100644
index 00000000..4a384a63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/egg_link.py
@@ -0,0 +1,80 @@
+import os
+import re
+import sys
+from typing import List, Optional
+
+from pip._internal.locations import site_packages, user_site
+from pip._internal.utils.virtualenv import (
+    running_under_virtualenv,
+    virtualenv_no_global,
+)
+
+__all__ = [
+    "egg_link_path_from_sys_path",
+    "egg_link_path_from_location",
+]
+
+
+def _egg_link_names(raw_name: str) -> List[str]:
+    """
+    Convert a Name metadata value to a .egg-link name, by applying
+    the same substitution as pkg_resources's safe_name function.
+    Note: we cannot use canonicalize_name because it has a different logic.
+
+    We also look for the raw name (without normalization) as setuptools 69 changed
+    the way it names .egg-link files (https://github.com/pypa/setuptools/issues/4167).
+    """
+    return [
+        re.sub("[^A-Za-z0-9.]+", "-", raw_name) + ".egg-link",
+        f"{raw_name}.egg-link",
+    ]
+
+
+def egg_link_path_from_sys_path(raw_name: str) -> Optional[str]:
+    """
+    Look for a .egg-link file for project name, by walking sys.path.
+    """
+    egg_link_names = _egg_link_names(raw_name)
+    for path_item in sys.path:
+        for egg_link_name in egg_link_names:
+            egg_link = os.path.join(path_item, egg_link_name)
+            if os.path.isfile(egg_link):
+                return egg_link
+    return None
+
+
+def egg_link_path_from_location(raw_name: str) -> Optional[str]:
+    """
+    Return the path for the .egg-link file if it exists, otherwise, None.
+
+    There's 3 scenarios:
+    1) not in a virtualenv
+       try to find in site.USER_SITE, then site_packages
+    2) in a no-global virtualenv
+       try to find in site_packages
+    3) in a yes-global virtualenv
+       try to find in site_packages, then site.USER_SITE
+       (don't look in global location)
+
+    For #1 and #3, there could be odd cases, where there's an egg-link in 2
+    locations.
+
+    This method will just return the first one found.
+    """
+    sites: List[str] = []
+    if running_under_virtualenv():
+        sites.append(site_packages)
+        if not virtualenv_no_global() and user_site:
+            sites.append(user_site)
+    else:
+        if user_site:
+            sites.append(user_site)
+        sites.append(site_packages)
+
+    egg_link_names = _egg_link_names(raw_name)
+    for site in sites:
+        for egg_link_name in egg_link_names:
+            egglink = os.path.join(site, egg_link_name)
+            if os.path.isfile(egglink):
+                return egglink
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/entrypoints.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/entrypoints.py
new file mode 100644
index 00000000..15013693
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/entrypoints.py
@@ -0,0 +1,84 @@
+import itertools
+import os
+import shutil
+import sys
+from typing import List, Optional
+
+from pip._internal.cli.main import main
+from pip._internal.utils.compat import WINDOWS
+
+_EXECUTABLE_NAMES = [
+    "pip",
+    f"pip{sys.version_info.major}",
+    f"pip{sys.version_info.major}.{sys.version_info.minor}",
+]
+if WINDOWS:
+    _allowed_extensions = {"", ".exe"}
+    _EXECUTABLE_NAMES = [
+        "".join(parts)
+        for parts in itertools.product(_EXECUTABLE_NAMES, _allowed_extensions)
+    ]
+
+
+def _wrapper(args: Optional[List[str]] = None) -> int:
+    """Central wrapper for all old entrypoints.
+
+    Historically pip has had several entrypoints defined. Because of issues
+    arising from PATH, sys.path, multiple Pythons, their interactions, and most
+    of them having a pip installed, users suffer every time an entrypoint gets
+    moved.
+
+    To alleviate this pain, and provide a mechanism for warning users and
+    directing them to an appropriate place for help, we now define all of
+    our old entrypoints as wrappers for the current one.
+    """
+    sys.stderr.write(
+        "WARNING: pip is being invoked by an old script wrapper. This will "
+        "fail in a future version of pip.\n"
+        "Please see https://github.com/pypa/pip/issues/5599 for advice on "
+        "fixing the underlying issue.\n"
+        "To avoid this problem you can invoke Python with '-m pip' instead of "
+        "running pip directly.\n"
+    )
+    return main(args)
+
+
+def get_best_invocation_for_this_pip() -> str:
+    """Try to figure out the best way to invoke pip in the current environment."""
+    binary_directory = "Scripts" if WINDOWS else "bin"
+    binary_prefix = os.path.join(sys.prefix, binary_directory)
+
+    # Try to use pip[X[.Y]] names, if those executables for this environment are
+    # the first on PATH with that name.
+    path_parts = os.path.normcase(os.environ.get("PATH", "")).split(os.pathsep)
+    exe_are_in_PATH = os.path.normcase(binary_prefix) in path_parts
+    if exe_are_in_PATH:
+        for exe_name in _EXECUTABLE_NAMES:
+            found_executable = shutil.which(exe_name)
+            binary_executable = os.path.join(binary_prefix, exe_name)
+            if (
+                found_executable
+                and os.path.exists(binary_executable)
+                and os.path.samefile(
+                    found_executable,
+                    binary_executable,
+                )
+            ):
+                return exe_name
+
+    # Use the `-m` invocation, if there's no "nice" invocation.
+    return f"{get_best_invocation_for_this_python()} -m pip"
+
+
+def get_best_invocation_for_this_python() -> str:
+    """Try to figure out the best way to invoke the current Python."""
+    exe = sys.executable
+    exe_name = os.path.basename(exe)
+
+    # Try to use the basename, if it's the first executable.
+    found_executable = shutil.which(exe_name)
+    if found_executable and os.path.samefile(found_executable, exe):
+        return exe_name
+
+    # Use the full executable name, because we couldn't find something simpler.
+    return exe
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filesystem.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filesystem.py
new file mode 100644
index 00000000..22e356cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filesystem.py
@@ -0,0 +1,149 @@
+import fnmatch
+import os
+import os.path
+import random
+import sys
+from contextlib import contextmanager
+from tempfile import NamedTemporaryFile
+from typing import Any, BinaryIO, Generator, List, Union, cast
+
+from pip._internal.utils.compat import get_path_uid
+from pip._internal.utils.misc import format_size
+from pip._internal.utils.retry import retry
+
+
+def check_path_owner(path: str) -> bool:
+    # If we don't have a way to check the effective uid of this process, then
+    # we'll just assume that we own the directory.
+    if sys.platform == "win32" or not hasattr(os, "geteuid"):
+        return True
+
+    assert os.path.isabs(path)
+
+    previous = None
+    while path != previous:
+        if os.path.lexists(path):
+            # Check if path is writable by current user.
+            if os.geteuid() == 0:
+                # Special handling for root user in order to handle properly
+                # cases where users use sudo without -H flag.
+                try:
+                    path_uid = get_path_uid(path)
+                except OSError:
+                    return False
+                return path_uid == 0
+            else:
+                return os.access(path, os.W_OK)
+        else:
+            previous, path = path, os.path.dirname(path)
+    return False  # assume we don't own the path
+
+
+@contextmanager
+def adjacent_tmp_file(path: str, **kwargs: Any) -> Generator[BinaryIO, None, None]:
+    """Return a file-like object pointing to a tmp file next to path.
+
+    The file is created securely and is ensured to be written to disk
+    after the context reaches its end.
+
+    kwargs will be passed to tempfile.NamedTemporaryFile to control
+    the way the temporary file will be opened.
+    """
+    with NamedTemporaryFile(
+        delete=False,
+        dir=os.path.dirname(path),
+        prefix=os.path.basename(path),
+        suffix=".tmp",
+        **kwargs,
+    ) as f:
+        result = cast(BinaryIO, f)
+        try:
+            yield result
+        finally:
+            result.flush()
+            os.fsync(result.fileno())
+
+
+replace = retry(stop_after_delay=1, wait=0.25)(os.replace)
+
+
+# test_writable_dir and _test_writable_dir_win are copied from Flit,
+# with the author's agreement to also place them under pip's license.
+def test_writable_dir(path: str) -> bool:
+    """Check if a directory is writable.
+
+    Uses os.access() on POSIX, tries creating files on Windows.
+    """
+    # If the directory doesn't exist, find the closest parent that does.
+    while not os.path.isdir(path):
+        parent = os.path.dirname(path)
+        if parent == path:
+            break  # Should never get here, but infinite loops are bad
+        path = parent
+
+    if os.name == "posix":
+        return os.access(path, os.W_OK)
+
+    return _test_writable_dir_win(path)
+
+
+def _test_writable_dir_win(path: str) -> bool:
+    # os.access doesn't work on Windows: http://bugs.python.org/issue2528
+    # and we can't use tempfile: http://bugs.python.org/issue22107
+    basename = "accesstest_deleteme_fishfingers_custard_"
+    alphabet = "abcdefghijklmnopqrstuvwxyz0123456789"
+    for _ in range(10):
+        name = basename + "".join(random.choice(alphabet) for _ in range(6))
+        file = os.path.join(path, name)
+        try:
+            fd = os.open(file, os.O_RDWR | os.O_CREAT | os.O_EXCL)
+        except FileExistsError:
+            pass
+        except PermissionError:
+            # This could be because there's a directory with the same name.
+            # But it's highly unlikely there's a directory called that,
+            # so we'll assume it's because the parent dir is not writable.
+            # This could as well be because the parent dir is not readable,
+            # due to non-privileged user access.
+            return False
+        else:
+            os.close(fd)
+            os.unlink(file)
+            return True
+
+    # This should never be reached
+    raise OSError("Unexpected condition testing for writable directory")
+
+
+def find_files(path: str, pattern: str) -> List[str]:
+    """Returns a list of absolute paths of files beneath path, recursively,
+    with filenames which match the UNIX-style shell glob pattern."""
+    result: List[str] = []
+    for root, _, files in os.walk(path):
+        matches = fnmatch.filter(files, pattern)
+        result.extend(os.path.join(root, f) for f in matches)
+    return result
+
+
+def file_size(path: str) -> Union[int, float]:
+    # If it's a symlink, return 0.
+    if os.path.islink(path):
+        return 0
+    return os.path.getsize(path)
+
+
+def format_file_size(path: str) -> str:
+    return format_size(file_size(path))
+
+
+def directory_size(path: str) -> Union[int, float]:
+    size = 0.0
+    for root, _dirs, files in os.walk(path):
+        for filename in files:
+            file_path = os.path.join(root, filename)
+            size += file_size(file_path)
+    return size
+
+
+def format_directory_size(path: str) -> str:
+    return format_size(directory_size(path))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filetypes.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filetypes.py
new file mode 100644
index 00000000..59485701
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/filetypes.py
@@ -0,0 +1,27 @@
+"""Filetype information.
+"""
+
+from typing import Tuple
+
+from pip._internal.utils.misc import splitext
+
+WHEEL_EXTENSION = ".whl"
+BZ2_EXTENSIONS: Tuple[str, ...] = (".tar.bz2", ".tbz")
+XZ_EXTENSIONS: Tuple[str, ...] = (
+    ".tar.xz",
+    ".txz",
+    ".tlz",
+    ".tar.lz",
+    ".tar.lzma",
+)
+ZIP_EXTENSIONS: Tuple[str, ...] = (".zip", WHEEL_EXTENSION)
+TAR_EXTENSIONS: Tuple[str, ...] = (".tar.gz", ".tgz", ".tar")
+ARCHIVE_EXTENSIONS = ZIP_EXTENSIONS + BZ2_EXTENSIONS + TAR_EXTENSIONS + XZ_EXTENSIONS
+
+
+def is_archive_file(name: str) -> bool:
+    """Return True if `name` is a considered as an archive file."""
+    ext = splitext(name)[1].lower()
+    if ext in ARCHIVE_EXTENSIONS:
+        return True
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/glibc.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/glibc.py
new file mode 100644
index 00000000..998868ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/glibc.py
@@ -0,0 +1,101 @@
+import os
+import sys
+from typing import Optional, Tuple
+
+
+def glibc_version_string() -> Optional[str]:
+    "Returns glibc version string, or None if not using glibc."
+    return glibc_version_string_confstr() or glibc_version_string_ctypes()
+
+
+def glibc_version_string_confstr() -> Optional[str]:
+    "Primary implementation of glibc_version_string using os.confstr."
+    # os.confstr is quite a bit faster than ctypes.DLL. It's also less likely
+    # to be broken or missing. This strategy is used in the standard library
+    # platform module:
+    # https://github.com/python/cpython/blob/fcf1d003bf4f0100c9d0921ff3d70e1127ca1b71/Lib/platform.py#L175-L183
+    if sys.platform == "win32":
+        return None
+    try:
+        gnu_libc_version = os.confstr("CS_GNU_LIBC_VERSION")
+        if gnu_libc_version is None:
+            return None
+        # os.confstr("CS_GNU_LIBC_VERSION") returns a string like "glibc 2.17":
+        _, version = gnu_libc_version.split()
+    except (AttributeError, OSError, ValueError):
+        # os.confstr() or CS_GNU_LIBC_VERSION not available (or a bad value)...
+        return None
+    return version
+
+
+def glibc_version_string_ctypes() -> Optional[str]:
+    "Fallback implementation of glibc_version_string using ctypes."
+
+    try:
+        import ctypes
+    except ImportError:
+        return None
+
+    # ctypes.CDLL(None) internally calls dlopen(NULL), and as the dlopen
+    # manpage says, "If filename is NULL, then the returned handle is for the
+    # main program". This way we can let the linker do the work to figure out
+    # which libc our process is actually using.
+    #
+    # We must also handle the special case where the executable is not a
+    # dynamically linked executable. This can occur when using musl libc,
+    # for example. In this situation, dlopen() will error, leading to an
+    # OSError. Interestingly, at least in the case of musl, there is no
+    # errno set on the OSError. The single string argument used to construct
+    # OSError comes from libc itself and is therefore not portable to
+    # hard code here. In any case, failure to call dlopen() means we
+    # can't proceed, so we bail on our attempt.
+    try:
+        process_namespace = ctypes.CDLL(None)
+    except OSError:
+        return None
+
+    try:
+        gnu_get_libc_version = process_namespace.gnu_get_libc_version
+    except AttributeError:
+        # Symbol doesn't exist -> therefore, we are not linked to
+        # glibc.
+        return None
+
+    # Call gnu_get_libc_version, which returns a string like "2.5"
+    gnu_get_libc_version.restype = ctypes.c_char_p
+    version_str: str = gnu_get_libc_version()
+    # py2 / py3 compatibility:
+    if not isinstance(version_str, str):
+        version_str = version_str.decode("ascii")
+
+    return version_str
+
+
+# platform.libc_ver regularly returns completely nonsensical glibc
+# versions. E.g. on my computer, platform says:
+#
+#   ~$ python2.7 -c 'import platform; print(platform.libc_ver())'
+#   ('glibc', '2.7')
+#   ~$ python3.5 -c 'import platform; print(platform.libc_ver())'
+#   ('glibc', '2.9')
+#
+# But the truth is:
+#
+#   ~$ ldd --version
+#   ldd (Debian GLIBC 2.22-11) 2.22
+#
+# This is unfortunate, because it means that the linehaul data on libc
+# versions that was generated by pip 8.1.2 and earlier is useless and
+# misleading. Solution: instead of using platform, use our code that actually
+# works.
+def libc_ver() -> Tuple[str, str]:
+    """Try to determine the glibc version
+
+    Returns a tuple of strings (lib, version) which default to empty strings
+    in case the lookup fails.
+    """
+    glibc_version = glibc_version_string()
+    if glibc_version is None:
+        return ("", "")
+    else:
+        return ("glibc", glibc_version)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/hashes.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/hashes.py
new file mode 100644
index 00000000..535e94fc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/hashes.py
@@ -0,0 +1,147 @@
+import hashlib
+from typing import TYPE_CHECKING, BinaryIO, Dict, Iterable, List, NoReturn, Optional
+
+from pip._internal.exceptions import HashMismatch, HashMissing, InstallationError
+from pip._internal.utils.misc import read_chunks
+
+if TYPE_CHECKING:
+    from hashlib import _Hash
+
+
+# The recommended hash algo of the moment. Change this whenever the state of
+# the art changes; it won't hurt backward compatibility.
+FAVORITE_HASH = "sha256"
+
+
+# Names of hashlib algorithms allowed by the --hash option and ``pip hash``
+# Currently, those are the ones at least as collision-resistant as sha256.
+STRONG_HASHES = ["sha256", "sha384", "sha512"]
+
+
+class Hashes:
+    """A wrapper that builds multiple hashes at once and checks them against
+    known-good values
+
+    """
+
+    def __init__(self, hashes: Optional[Dict[str, List[str]]] = None) -> None:
+        """
+        :param hashes: A dict of algorithm names pointing to lists of allowed
+            hex digests
+        """
+        allowed = {}
+        if hashes is not None:
+            for alg, keys in hashes.items():
+                # Make sure values are always sorted (to ease equality checks)
+                allowed[alg] = [k.lower() for k in sorted(keys)]
+        self._allowed = allowed
+
+    def __and__(self, other: "Hashes") -> "Hashes":
+        if not isinstance(other, Hashes):
+            return NotImplemented
+
+        # If either of the Hashes object is entirely empty (i.e. no hash
+        # specified at all), all hashes from the other object are allowed.
+        if not other:
+            return self
+        if not self:
+            return other
+
+        # Otherwise only hashes that present in both objects are allowed.
+        new = {}
+        for alg, values in other._allowed.items():
+            if alg not in self._allowed:
+                continue
+            new[alg] = [v for v in values if v in self._allowed[alg]]
+        return Hashes(new)
+
+    @property
+    def digest_count(self) -> int:
+        return sum(len(digests) for digests in self._allowed.values())
+
+    def is_hash_allowed(self, hash_name: str, hex_digest: str) -> bool:
+        """Return whether the given hex digest is allowed."""
+        return hex_digest in self._allowed.get(hash_name, [])
+
+    def check_against_chunks(self, chunks: Iterable[bytes]) -> None:
+        """Check good hashes against ones built from iterable of chunks of
+        data.
+
+        Raise HashMismatch if none match.
+
+        """
+        gots = {}
+        for hash_name in self._allowed.keys():
+            try:
+                gots[hash_name] = hashlib.new(hash_name)
+            except (ValueError, TypeError):
+                raise InstallationError(f"Unknown hash name: {hash_name}")
+
+        for chunk in chunks:
+            for hash in gots.values():
+                hash.update(chunk)
+
+        for hash_name, got in gots.items():
+            if got.hexdigest() in self._allowed[hash_name]:
+                return
+        self._raise(gots)
+
+    def _raise(self, gots: Dict[str, "_Hash"]) -> "NoReturn":
+        raise HashMismatch(self._allowed, gots)
+
+    def check_against_file(self, file: BinaryIO) -> None:
+        """Check good hashes against a file-like object
+
+        Raise HashMismatch if none match.
+
+        """
+        return self.check_against_chunks(read_chunks(file))
+
+    def check_against_path(self, path: str) -> None:
+        with open(path, "rb") as file:
+            return self.check_against_file(file)
+
+    def has_one_of(self, hashes: Dict[str, str]) -> bool:
+        """Return whether any of the given hashes are allowed."""
+        for hash_name, hex_digest in hashes.items():
+            if self.is_hash_allowed(hash_name, hex_digest):
+                return True
+        return False
+
+    def __bool__(self) -> bool:
+        """Return whether I know any known-good hashes."""
+        return bool(self._allowed)
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Hashes):
+            return NotImplemented
+        return self._allowed == other._allowed
+
+    def __hash__(self) -> int:
+        return hash(
+            ",".join(
+                sorted(
+                    ":".join((alg, digest))
+                    for alg, digest_list in self._allowed.items()
+                    for digest in digest_list
+                )
+            )
+        )
+
+
+class MissingHashes(Hashes):
+    """A workalike for Hashes used when we're missing a hash for a requirement
+
+    It computes the actual hash of the requirement and raises a HashMissing
+    exception showing it to the user.
+
+    """
+
+    def __init__(self) -> None:
+        """Don't offer the ``hashes`` kwarg."""
+        # Pass our favorite hash in to generate a "gotten hash". With the
+        # empty list, it will never match, so an error will always raise.
+        super().__init__(hashes={FAVORITE_HASH: []})
+
+    def _raise(self, gots: Dict[str, "_Hash"]) -> "NoReturn":
+        raise HashMissing(gots[FAVORITE_HASH].hexdigest())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/logging.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/logging.py
new file mode 100644
index 00000000..62035fc4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/logging.py
@@ -0,0 +1,354 @@
+import contextlib
+import errno
+import logging
+import logging.handlers
+import os
+import sys
+import threading
+from dataclasses import dataclass
+from io import TextIOWrapper
+from logging import Filter
+from typing import Any, ClassVar, Generator, List, Optional, TextIO, Type
+
+from pip._vendor.rich.console import (
+    Console,
+    ConsoleOptions,
+    ConsoleRenderable,
+    RenderableType,
+    RenderResult,
+    RichCast,
+)
+from pip._vendor.rich.highlighter import NullHighlighter
+from pip._vendor.rich.logging import RichHandler
+from pip._vendor.rich.segment import Segment
+from pip._vendor.rich.style import Style
+
+from pip._internal.utils._log import VERBOSE, getLogger
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.deprecation import DEPRECATION_MSG_PREFIX
+from pip._internal.utils.misc import ensure_dir
+
+_log_state = threading.local()
+subprocess_logger = getLogger("pip.subprocessor")
+
+
+class BrokenStdoutLoggingError(Exception):
+    """
+    Raised if BrokenPipeError occurs for the stdout stream while logging.
+    """
+
+
+def _is_broken_pipe_error(exc_class: Type[BaseException], exc: BaseException) -> bool:
+    if exc_class is BrokenPipeError:
+        return True
+
+    # On Windows, a broken pipe can show up as EINVAL rather than EPIPE:
+    # https://bugs.python.org/issue19612
+    # https://bugs.python.org/issue30418
+    if not WINDOWS:
+        return False
+
+    return isinstance(exc, OSError) and exc.errno in (errno.EINVAL, errno.EPIPE)
+
+
+@contextlib.contextmanager
+def indent_log(num: int = 2) -> Generator[None, None, None]:
+    """
+    A context manager which will cause the log output to be indented for any
+    log messages emitted inside it.
+    """
+    # For thread-safety
+    _log_state.indentation = get_indentation()
+    _log_state.indentation += num
+    try:
+        yield
+    finally:
+        _log_state.indentation -= num
+
+
+def get_indentation() -> int:
+    return getattr(_log_state, "indentation", 0)
+
+
+class IndentingFormatter(logging.Formatter):
+    default_time_format = "%Y-%m-%dT%H:%M:%S"
+
+    def __init__(
+        self,
+        *args: Any,
+        add_timestamp: bool = False,
+        **kwargs: Any,
+    ) -> None:
+        """
+        A logging.Formatter that obeys the indent_log() context manager.
+
+        :param add_timestamp: A bool indicating output lines should be prefixed
+            with their record's timestamp.
+        """
+        self.add_timestamp = add_timestamp
+        super().__init__(*args, **kwargs)
+
+    def get_message_start(self, formatted: str, levelno: int) -> str:
+        """
+        Return the start of the formatted log message (not counting the
+        prefix to add to each line).
+        """
+        if levelno < logging.WARNING:
+            return ""
+        if formatted.startswith(DEPRECATION_MSG_PREFIX):
+            # Then the message already has a prefix.  We don't want it to
+            # look like "WARNING: DEPRECATION: ...."
+            return ""
+        if levelno < logging.ERROR:
+            return "WARNING: "
+
+        return "ERROR: "
+
+    def format(self, record: logging.LogRecord) -> str:
+        """
+        Calls the standard formatter, but will indent all of the log message
+        lines by our current indentation level.
+        """
+        formatted = super().format(record)
+        message_start = self.get_message_start(formatted, record.levelno)
+        formatted = message_start + formatted
+
+        prefix = ""
+        if self.add_timestamp:
+            prefix = f"{self.formatTime(record)} "
+        prefix += " " * get_indentation()
+        formatted = "".join([prefix + line for line in formatted.splitlines(True)])
+        return formatted
+
+
+@dataclass
+class IndentedRenderable:
+    renderable: RenderableType
+    indent: int
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        segments = console.render(self.renderable, options)
+        lines = Segment.split_lines(segments)
+        for line in lines:
+            yield Segment(" " * self.indent)
+            yield from line
+            yield Segment("\n")
+
+
+class PipConsole(Console):
+    def on_broken_pipe(self) -> None:
+        # Reraise the original exception, rich 13.8.0+ exits by default
+        # instead, preventing our handler from firing.
+        raise BrokenPipeError() from None
+
+
+class RichPipStreamHandler(RichHandler):
+    KEYWORDS: ClassVar[Optional[List[str]]] = []
+
+    def __init__(self, stream: Optional[TextIO], no_color: bool) -> None:
+        super().__init__(
+            console=PipConsole(file=stream, no_color=no_color, soft_wrap=True),
+            show_time=False,
+            show_level=False,
+            show_path=False,
+            highlighter=NullHighlighter(),
+        )
+
+    # Our custom override on Rich's logger, to make things work as we need them to.
+    def emit(self, record: logging.LogRecord) -> None:
+        style: Optional[Style] = None
+
+        # If we are given a diagnostic error to present, present it with indentation.
+        if getattr(record, "rich", False):
+            assert isinstance(record.args, tuple)
+            (rich_renderable,) = record.args
+            assert isinstance(
+                rich_renderable, (ConsoleRenderable, RichCast, str)
+            ), f"{rich_renderable} is not rich-console-renderable"
+
+            renderable: RenderableType = IndentedRenderable(
+                rich_renderable, indent=get_indentation()
+            )
+        else:
+            message = self.format(record)
+            renderable = self.render_message(record, message)
+            if record.levelno is not None:
+                if record.levelno >= logging.ERROR:
+                    style = Style(color="red")
+                elif record.levelno >= logging.WARNING:
+                    style = Style(color="yellow")
+
+        try:
+            self.console.print(renderable, overflow="ignore", crop=False, style=style)
+        except Exception:
+            self.handleError(record)
+
+    def handleError(self, record: logging.LogRecord) -> None:
+        """Called when logging is unable to log some output."""
+
+        exc_class, exc = sys.exc_info()[:2]
+        # If a broken pipe occurred while calling write() or flush() on the
+        # stdout stream in logging's Handler.emit(), then raise our special
+        # exception so we can handle it in main() instead of logging the
+        # broken pipe error and continuing.
+        if (
+            exc_class
+            and exc
+            and self.console.file is sys.stdout
+            and _is_broken_pipe_error(exc_class, exc)
+        ):
+            raise BrokenStdoutLoggingError()
+
+        return super().handleError(record)
+
+
+class BetterRotatingFileHandler(logging.handlers.RotatingFileHandler):
+    def _open(self) -> TextIOWrapper:
+        ensure_dir(os.path.dirname(self.baseFilename))
+        return super()._open()
+
+
+class MaxLevelFilter(Filter):
+    def __init__(self, level: int) -> None:
+        self.level = level
+
+    def filter(self, record: logging.LogRecord) -> bool:
+        return record.levelno < self.level
+
+
+class ExcludeLoggerFilter(Filter):
+    """
+    A logging Filter that excludes records from a logger (or its children).
+    """
+
+    def filter(self, record: logging.LogRecord) -> bool:
+        # The base Filter class allows only records from a logger (or its
+        # children).
+        return not super().filter(record)
+
+
+def setup_logging(verbosity: int, no_color: bool, user_log_file: Optional[str]) -> int:
+    """Configures and sets up all of the logging
+
+    Returns the requested logging level, as its integer value.
+    """
+
+    # Determine the level to be logging at.
+    if verbosity >= 2:
+        level_number = logging.DEBUG
+    elif verbosity == 1:
+        level_number = VERBOSE
+    elif verbosity == -1:
+        level_number = logging.WARNING
+    elif verbosity == -2:
+        level_number = logging.ERROR
+    elif verbosity <= -3:
+        level_number = logging.CRITICAL
+    else:
+        level_number = logging.INFO
+
+    level = logging.getLevelName(level_number)
+
+    # The "root" logger should match the "console" level *unless* we also need
+    # to log to a user log file.
+    include_user_log = user_log_file is not None
+    if include_user_log:
+        additional_log_file = user_log_file
+        root_level = "DEBUG"
+    else:
+        additional_log_file = "/dev/null"
+        root_level = level
+
+    # Disable any logging besides WARNING unless we have DEBUG level logging
+    # enabled for vendored libraries.
+    vendored_log_level = "WARNING" if level in ["INFO", "ERROR"] else "DEBUG"
+
+    # Shorthands for clarity
+    log_streams = {
+        "stdout": "ext://sys.stdout",
+        "stderr": "ext://sys.stderr",
+    }
+    handler_classes = {
+        "stream": "pip._internal.utils.logging.RichPipStreamHandler",
+        "file": "pip._internal.utils.logging.BetterRotatingFileHandler",
+    }
+    handlers = ["console", "console_errors", "console_subprocess"] + (
+        ["user_log"] if include_user_log else []
+    )
+
+    logging.config.dictConfig(
+        {
+            "version": 1,
+            "disable_existing_loggers": False,
+            "filters": {
+                "exclude_warnings": {
+                    "()": "pip._internal.utils.logging.MaxLevelFilter",
+                    "level": logging.WARNING,
+                },
+                "restrict_to_subprocess": {
+                    "()": "logging.Filter",
+                    "name": subprocess_logger.name,
+                },
+                "exclude_subprocess": {
+                    "()": "pip._internal.utils.logging.ExcludeLoggerFilter",
+                    "name": subprocess_logger.name,
+                },
+            },
+            "formatters": {
+                "indent": {
+                    "()": IndentingFormatter,
+                    "format": "%(message)s",
+                },
+                "indent_with_timestamp": {
+                    "()": IndentingFormatter,
+                    "format": "%(message)s",
+                    "add_timestamp": True,
+                },
+            },
+            "handlers": {
+                "console": {
+                    "level": level,
+                    "class": handler_classes["stream"],
+                    "no_color": no_color,
+                    "stream": log_streams["stdout"],
+                    "filters": ["exclude_subprocess", "exclude_warnings"],
+                    "formatter": "indent",
+                },
+                "console_errors": {
+                    "level": "WARNING",
+                    "class": handler_classes["stream"],
+                    "no_color": no_color,
+                    "stream": log_streams["stderr"],
+                    "filters": ["exclude_subprocess"],
+                    "formatter": "indent",
+                },
+                # A handler responsible for logging to the console messages
+                # from the "subprocessor" logger.
+                "console_subprocess": {
+                    "level": level,
+                    "class": handler_classes["stream"],
+                    "stream": log_streams["stderr"],
+                    "no_color": no_color,
+                    "filters": ["restrict_to_subprocess"],
+                    "formatter": "indent",
+                },
+                "user_log": {
+                    "level": "DEBUG",
+                    "class": handler_classes["file"],
+                    "filename": additional_log_file,
+                    "encoding": "utf-8",
+                    "delay": True,
+                    "formatter": "indent_with_timestamp",
+                },
+            },
+            "root": {
+                "level": root_level,
+                "handlers": handlers,
+            },
+            "loggers": {"pip._vendor": {"level": vendored_log_level}},
+        }
+    )
+
+    return level_number
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/misc.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/misc.py
new file mode 100644
index 00000000..44f6a05f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/misc.py
@@ -0,0 +1,773 @@
+import errno
+import getpass
+import hashlib
+import logging
+import os
+import posixpath
+import shutil
+import stat
+import sys
+import sysconfig
+import urllib.parse
+from dataclasses import dataclass
+from functools import partial
+from io import StringIO
+from itertools import filterfalse, tee, zip_longest
+from pathlib import Path
+from types import FunctionType, TracebackType
+from typing import (
+    Any,
+    BinaryIO,
+    Callable,
+    Generator,
+    Iterable,
+    Iterator,
+    List,
+    Mapping,
+    Optional,
+    Sequence,
+    TextIO,
+    Tuple,
+    Type,
+    TypeVar,
+    Union,
+    cast,
+)
+
+from pip._vendor.packaging.requirements import Requirement
+from pip._vendor.pyproject_hooks import BuildBackendHookCaller
+
+from pip import __version__
+from pip._internal.exceptions import CommandError, ExternallyManagedEnvironment
+from pip._internal.locations import get_major_minor_version
+from pip._internal.utils.compat import WINDOWS
+from pip._internal.utils.retry import retry
+from pip._internal.utils.virtualenv import running_under_virtualenv
+
+__all__ = [
+    "rmtree",
+    "display_path",
+    "backup_dir",
+    "ask",
+    "splitext",
+    "format_size",
+    "is_installable_dir",
+    "normalize_path",
+    "renames",
+    "get_prog",
+    "ensure_dir",
+    "remove_auth_from_url",
+    "check_externally_managed",
+    "ConfiguredBuildBackendHookCaller",
+]
+
+logger = logging.getLogger(__name__)
+
+T = TypeVar("T")
+ExcInfo = Tuple[Type[BaseException], BaseException, TracebackType]
+VersionInfo = Tuple[int, int, int]
+NetlocTuple = Tuple[str, Tuple[Optional[str], Optional[str]]]
+OnExc = Callable[[FunctionType, Path, BaseException], Any]
+OnErr = Callable[[FunctionType, Path, ExcInfo], Any]
+
+FILE_CHUNK_SIZE = 1024 * 1024
+
+
+def get_pip_version() -> str:
+    pip_pkg_dir = os.path.join(os.path.dirname(__file__), "..", "..")
+    pip_pkg_dir = os.path.abspath(pip_pkg_dir)
+
+    return f"pip {__version__} from {pip_pkg_dir} (python {get_major_minor_version()})"
+
+
+def normalize_version_info(py_version_info: Tuple[int, ...]) -> Tuple[int, int, int]:
+    """
+    Convert a tuple of ints representing a Python version to one of length
+    three.
+
+    :param py_version_info: a tuple of ints representing a Python version,
+        or None to specify no version. The tuple can have any length.
+
+    :return: a tuple of length three if `py_version_info` is non-None.
+        Otherwise, return `py_version_info` unchanged (i.e. None).
+    """
+    if len(py_version_info) < 3:
+        py_version_info += (3 - len(py_version_info)) * (0,)
+    elif len(py_version_info) > 3:
+        py_version_info = py_version_info[:3]
+
+    return cast("VersionInfo", py_version_info)
+
+
+def ensure_dir(path: str) -> None:
+    """os.path.makedirs without EEXIST."""
+    try:
+        os.makedirs(path)
+    except OSError as e:
+        # Windows can raise spurious ENOTEMPTY errors. See #6426.
+        if e.errno != errno.EEXIST and e.errno != errno.ENOTEMPTY:
+            raise
+
+
+def get_prog() -> str:
+    try:
+        prog = os.path.basename(sys.argv[0])
+        if prog in ("__main__.py", "-c"):
+            return f"{sys.executable} -m pip"
+        else:
+            return prog
+    except (AttributeError, TypeError, IndexError):
+        pass
+    return "pip"
+
+
+# Retry every half second for up to 3 seconds
+@retry(stop_after_delay=3, wait=0.5)
+def rmtree(
+    dir: str, ignore_errors: bool = False, onexc: Optional[OnExc] = None
+) -> None:
+    if ignore_errors:
+        onexc = _onerror_ignore
+    if onexc is None:
+        onexc = _onerror_reraise
+    handler: OnErr = partial(rmtree_errorhandler, onexc=onexc)
+    if sys.version_info >= (3, 12):
+        # See https://docs.python.org/3.12/whatsnew/3.12.html#shutil.
+        shutil.rmtree(dir, onexc=handler)  # type: ignore
+    else:
+        shutil.rmtree(dir, onerror=handler)  # type: ignore
+
+
+def _onerror_ignore(*_args: Any) -> None:
+    pass
+
+
+def _onerror_reraise(*_args: Any) -> None:
+    raise  # noqa: PLE0704 - Bare exception used to reraise existing exception
+
+
+def rmtree_errorhandler(
+    func: FunctionType,
+    path: Path,
+    exc_info: Union[ExcInfo, BaseException],
+    *,
+    onexc: OnExc = _onerror_reraise,
+) -> None:
+    """
+    `rmtree` error handler to 'force' a file remove (i.e. like `rm -f`).
+
+    * If a file is readonly then it's write flag is set and operation is
+      retried.
+
+    * `onerror` is the original callback from `rmtree(... onerror=onerror)`
+      that is chained at the end if the "rm -f" still fails.
+    """
+    try:
+        st_mode = os.stat(path).st_mode
+    except OSError:
+        # it's equivalent to os.path.exists
+        return
+
+    if not st_mode & stat.S_IWRITE:
+        # convert to read/write
+        try:
+            os.chmod(path, st_mode | stat.S_IWRITE)
+        except OSError:
+            pass
+        else:
+            # use the original function to repeat the operation
+            try:
+                func(path)
+                return
+            except OSError:
+                pass
+
+    if not isinstance(exc_info, BaseException):
+        _, exc_info, _ = exc_info
+    onexc(func, path, exc_info)
+
+
+def display_path(path: str) -> str:
+    """Gives the display value for a given path, making it relative to cwd
+    if possible."""
+    path = os.path.normcase(os.path.abspath(path))
+    if path.startswith(os.getcwd() + os.path.sep):
+        path = "." + path[len(os.getcwd()) :]
+    return path
+
+
+def backup_dir(dir: str, ext: str = ".bak") -> str:
+    """Figure out the name of a directory to back up the given dir to
+    (adding .bak, .bak2, etc)"""
+    n = 1
+    extension = ext
+    while os.path.exists(dir + extension):
+        n += 1
+        extension = ext + str(n)
+    return dir + extension
+
+
+def ask_path_exists(message: str, options: Iterable[str]) -> str:
+    for action in os.environ.get("PIP_EXISTS_ACTION", "").split():
+        if action in options:
+            return action
+    return ask(message, options)
+
+
+def _check_no_input(message: str) -> None:
+    """Raise an error if no input is allowed."""
+    if os.environ.get("PIP_NO_INPUT"):
+        raise Exception(
+            f"No input was expected ($PIP_NO_INPUT set); question: {message}"
+        )
+
+
+def ask(message: str, options: Iterable[str]) -> str:
+    """Ask the message interactively, with the given possible responses"""
+    while 1:
+        _check_no_input(message)
+        response = input(message)
+        response = response.strip().lower()
+        if response not in options:
+            print(
+                "Your response ({!r}) was not one of the expected responses: "
+                "{}".format(response, ", ".join(options))
+            )
+        else:
+            return response
+
+
+def ask_input(message: str) -> str:
+    """Ask for input interactively."""
+    _check_no_input(message)
+    return input(message)
+
+
+def ask_password(message: str) -> str:
+    """Ask for a password interactively."""
+    _check_no_input(message)
+    return getpass.getpass(message)
+
+
+def strtobool(val: str) -> int:
+    """Convert a string representation of truth to true (1) or false (0).
+
+    True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values
+    are 'n', 'no', 'f', 'false', 'off', and '0'.  Raises ValueError if
+    'val' is anything else.
+    """
+    val = val.lower()
+    if val in ("y", "yes", "t", "true", "on", "1"):
+        return 1
+    elif val in ("n", "no", "f", "false", "off", "0"):
+        return 0
+    else:
+        raise ValueError(f"invalid truth value {val!r}")
+
+
+def format_size(bytes: float) -> str:
+    if bytes > 1000 * 1000:
+        return f"{bytes / 1000.0 / 1000:.1f} MB"
+    elif bytes > 10 * 1000:
+        return f"{int(bytes / 1000)} kB"
+    elif bytes > 1000:
+        return f"{bytes / 1000.0:.1f} kB"
+    else:
+        return f"{int(bytes)} bytes"
+
+
+def tabulate(rows: Iterable[Iterable[Any]]) -> Tuple[List[str], List[int]]:
+    """Return a list of formatted rows and a list of column sizes.
+
+    For example::
+
+    >>> tabulate([['foobar', 2000], [0xdeadbeef]])
+    (['foobar     2000', '3735928559'], [10, 4])
+    """
+    rows = [tuple(map(str, row)) for row in rows]
+    sizes = [max(map(len, col)) for col in zip_longest(*rows, fillvalue="")]
+    table = [" ".join(map(str.ljust, row, sizes)).rstrip() for row in rows]
+    return table, sizes
+
+
+def is_installable_dir(path: str) -> bool:
+    """Is path is a directory containing pyproject.toml or setup.py?
+
+    If pyproject.toml exists, this is a PEP 517 project. Otherwise we look for
+    a legacy setuptools layout by identifying setup.py. We don't check for the
+    setup.cfg because using it without setup.py is only available for PEP 517
+    projects, which are already covered by the pyproject.toml check.
+    """
+    if not os.path.isdir(path):
+        return False
+    if os.path.isfile(os.path.join(path, "pyproject.toml")):
+        return True
+    if os.path.isfile(os.path.join(path, "setup.py")):
+        return True
+    return False
+
+
+def read_chunks(
+    file: BinaryIO, size: int = FILE_CHUNK_SIZE
+) -> Generator[bytes, None, None]:
+    """Yield pieces of data from a file-like object until EOF."""
+    while True:
+        chunk = file.read(size)
+        if not chunk:
+            break
+        yield chunk
+
+
+def normalize_path(path: str, resolve_symlinks: bool = True) -> str:
+    """
+    Convert a path to its canonical, case-normalized, absolute version.
+
+    """
+    path = os.path.expanduser(path)
+    if resolve_symlinks:
+        path = os.path.realpath(path)
+    else:
+        path = os.path.abspath(path)
+    return os.path.normcase(path)
+
+
+def splitext(path: str) -> Tuple[str, str]:
+    """Like os.path.splitext, but take off .tar too"""
+    base, ext = posixpath.splitext(path)
+    if base.lower().endswith(".tar"):
+        ext = base[-4:] + ext
+        base = base[:-4]
+    return base, ext
+
+
+def renames(old: str, new: str) -> None:
+    """Like os.renames(), but handles renaming across devices."""
+    # Implementation borrowed from os.renames().
+    head, tail = os.path.split(new)
+    if head and tail and not os.path.exists(head):
+        os.makedirs(head)
+
+    shutil.move(old, new)
+
+    head, tail = os.path.split(old)
+    if head and tail:
+        try:
+            os.removedirs(head)
+        except OSError:
+            pass
+
+
+def is_local(path: str) -> bool:
+    """
+    Return True if path is within sys.prefix, if we're running in a virtualenv.
+
+    If we're not in a virtualenv, all paths are considered "local."
+
+    Caution: this function assumes the head of path has been normalized
+    with normalize_path.
+    """
+    if not running_under_virtualenv():
+        return True
+    return path.startswith(normalize_path(sys.prefix))
+
+
+def write_output(msg: Any, *args: Any) -> None:
+    logger.info(msg, *args)
+
+
+class StreamWrapper(StringIO):
+    orig_stream: TextIO
+
+    @classmethod
+    def from_stream(cls, orig_stream: TextIO) -> "StreamWrapper":
+        ret = cls()
+        ret.orig_stream = orig_stream
+        return ret
+
+    # compileall.compile_dir() needs stdout.encoding to print to stdout
+    # type ignore is because TextIOBase.encoding is writeable
+    @property
+    def encoding(self) -> str:  # type: ignore
+        return self.orig_stream.encoding
+
+
+# Simulates an enum
+def enum(*sequential: Any, **named: Any) -> Type[Any]:
+    enums = dict(zip(sequential, range(len(sequential))), **named)
+    reverse = {value: key for key, value in enums.items()}
+    enums["reverse_mapping"] = reverse
+    return type("Enum", (), enums)
+
+
+def build_netloc(host: str, port: Optional[int]) -> str:
+    """
+    Build a netloc from a host-port pair
+    """
+    if port is None:
+        return host
+    if ":" in host:
+        # Only wrap host with square brackets when it is IPv6
+        host = f"[{host}]"
+    return f"{host}:{port}"
+
+
+def build_url_from_netloc(netloc: str, scheme: str = "https") -> str:
+    """
+    Build a full URL from a netloc.
+    """
+    if netloc.count(":") >= 2 and "@" not in netloc and "[" not in netloc:
+        # It must be a bare IPv6 address, so wrap it with brackets.
+        netloc = f"[{netloc}]"
+    return f"{scheme}://{netloc}"
+
+
+def parse_netloc(netloc: str) -> Tuple[Optional[str], Optional[int]]:
+    """
+    Return the host-port pair from a netloc.
+    """
+    url = build_url_from_netloc(netloc)
+    parsed = urllib.parse.urlparse(url)
+    return parsed.hostname, parsed.port
+
+
+def split_auth_from_netloc(netloc: str) -> NetlocTuple:
+    """
+    Parse out and remove the auth information from a netloc.
+
+    Returns: (netloc, (username, password)).
+    """
+    if "@" not in netloc:
+        return netloc, (None, None)
+
+    # Split from the right because that's how urllib.parse.urlsplit()
+    # behaves if more than one @ is present (which can be checked using
+    # the password attribute of urlsplit()'s return value).
+    auth, netloc = netloc.rsplit("@", 1)
+    pw: Optional[str] = None
+    if ":" in auth:
+        # Split from the left because that's how urllib.parse.urlsplit()
+        # behaves if more than one : is present (which again can be checked
+        # using the password attribute of the return value)
+        user, pw = auth.split(":", 1)
+    else:
+        user, pw = auth, None
+
+    user = urllib.parse.unquote(user)
+    if pw is not None:
+        pw = urllib.parse.unquote(pw)
+
+    return netloc, (user, pw)
+
+
+def redact_netloc(netloc: str) -> str:
+    """
+    Replace the sensitive data in a netloc with "****", if it exists.
+
+    For example:
+        - "user:pass@example.com" returns "user:****@example.com"
+        - "accesstoken@example.com" returns "****@example.com"
+    """
+    netloc, (user, password) = split_auth_from_netloc(netloc)
+    if user is None:
+        return netloc
+    if password is None:
+        user = "****"
+        password = ""
+    else:
+        user = urllib.parse.quote(user)
+        password = ":****"
+    return f"{user}{password}@{netloc}"
+
+
+def _transform_url(
+    url: str, transform_netloc: Callable[[str], Tuple[Any, ...]]
+) -> Tuple[str, NetlocTuple]:
+    """Transform and replace netloc in a url.
+
+    transform_netloc is a function taking the netloc and returning a
+    tuple. The first element of this tuple is the new netloc. The
+    entire tuple is returned.
+
+    Returns a tuple containing the transformed url as item 0 and the
+    original tuple returned by transform_netloc as item 1.
+    """
+    purl = urllib.parse.urlsplit(url)
+    netloc_tuple = transform_netloc(purl.netloc)
+    # stripped url
+    url_pieces = (purl.scheme, netloc_tuple[0], purl.path, purl.query, purl.fragment)
+    surl = urllib.parse.urlunsplit(url_pieces)
+    return surl, cast("NetlocTuple", netloc_tuple)
+
+
+def _get_netloc(netloc: str) -> NetlocTuple:
+    return split_auth_from_netloc(netloc)
+
+
+def _redact_netloc(netloc: str) -> Tuple[str]:
+    return (redact_netloc(netloc),)
+
+
+def split_auth_netloc_from_url(
+    url: str,
+) -> Tuple[str, str, Tuple[Optional[str], Optional[str]]]:
+    """
+    Parse a url into separate netloc, auth, and url with no auth.
+
+    Returns: (url_without_auth, netloc, (username, password))
+    """
+    url_without_auth, (netloc, auth) = _transform_url(url, _get_netloc)
+    return url_without_auth, netloc, auth
+
+
+def remove_auth_from_url(url: str) -> str:
+    """Return a copy of url with 'username:password@' removed."""
+    # username/pass params are passed to subversion through flags
+    # and are not recognized in the url.
+    return _transform_url(url, _get_netloc)[0]
+
+
+def redact_auth_from_url(url: str) -> str:
+    """Replace the password in a given url with ****."""
+    return _transform_url(url, _redact_netloc)[0]
+
+
+def redact_auth_from_requirement(req: Requirement) -> str:
+    """Replace the password in a given requirement url with ****."""
+    if not req.url:
+        return str(req)
+    return str(req).replace(req.url, redact_auth_from_url(req.url))
+
+
+@dataclass(frozen=True)
+class HiddenText:
+    secret: str
+    redacted: str
+
+    def __repr__(self) -> str:
+        return f"<HiddenText {str(self)!r}>"
+
+    def __str__(self) -> str:
+        return self.redacted
+
+    # This is useful for testing.
+    def __eq__(self, other: Any) -> bool:
+        if type(self) is not type(other):
+            return False
+
+        # The string being used for redaction doesn't also have to match,
+        # just the raw, original string.
+        return self.secret == other.secret
+
+
+def hide_value(value: str) -> HiddenText:
+    return HiddenText(value, redacted="****")
+
+
+def hide_url(url: str) -> HiddenText:
+    redacted = redact_auth_from_url(url)
+    return HiddenText(url, redacted=redacted)
+
+
+def protect_pip_from_modification_on_windows(modifying_pip: bool) -> None:
+    """Protection of pip.exe from modification on Windows
+
+    On Windows, any operation modifying pip should be run as:
+        python -m pip ...
+    """
+    pip_names = [
+        "pip",
+        f"pip{sys.version_info.major}",
+        f"pip{sys.version_info.major}.{sys.version_info.minor}",
+    ]
+
+    # See https://github.com/pypa/pip/issues/1299 for more discussion
+    should_show_use_python_msg = (
+        modifying_pip and WINDOWS and os.path.basename(sys.argv[0]) in pip_names
+    )
+
+    if should_show_use_python_msg:
+        new_command = [sys.executable, "-m", "pip"] + sys.argv[1:]
+        raise CommandError(
+            "To modify pip, please run the following command:\n{}".format(
+                " ".join(new_command)
+            )
+        )
+
+
+def check_externally_managed() -> None:
+    """Check whether the current environment is externally managed.
+
+    If the ``EXTERNALLY-MANAGED`` config file is found, the current environment
+    is considered externally managed, and an ExternallyManagedEnvironment is
+    raised.
+    """
+    if running_under_virtualenv():
+        return
+    marker = os.path.join(sysconfig.get_path("stdlib"), "EXTERNALLY-MANAGED")
+    if not os.path.isfile(marker):
+        return
+    raise ExternallyManagedEnvironment.from_config(marker)
+
+
+def is_console_interactive() -> bool:
+    """Is this console interactive?"""
+    return sys.stdin is not None and sys.stdin.isatty()
+
+
+def hash_file(path: str, blocksize: int = 1 << 20) -> Tuple[Any, int]:
+    """Return (hash, length) for path using hashlib.sha256()"""
+
+    h = hashlib.sha256()
+    length = 0
+    with open(path, "rb") as f:
+        for block in read_chunks(f, size=blocksize):
+            length += len(block)
+            h.update(block)
+    return h, length
+
+
+def pairwise(iterable: Iterable[Any]) -> Iterator[Tuple[Any, Any]]:
+    """
+    Return paired elements.
+
+    For example:
+        s -> (s0, s1), (s2, s3), (s4, s5), ...
+    """
+    iterable = iter(iterable)
+    return zip_longest(iterable, iterable)
+
+
+def partition(
+    pred: Callable[[T], bool], iterable: Iterable[T]
+) -> Tuple[Iterable[T], Iterable[T]]:
+    """
+    Use a predicate to partition entries into false entries and true entries,
+    like
+
+        partition(is_odd, range(10)) --> 0 2 4 6 8   and  1 3 5 7 9
+    """
+    t1, t2 = tee(iterable)
+    return filterfalse(pred, t1), filter(pred, t2)
+
+
+class ConfiguredBuildBackendHookCaller(BuildBackendHookCaller):
+    def __init__(
+        self,
+        config_holder: Any,
+        source_dir: str,
+        build_backend: str,
+        backend_path: Optional[str] = None,
+        runner: Optional[Callable[..., None]] = None,
+        python_executable: Optional[str] = None,
+    ):
+        super().__init__(
+            source_dir, build_backend, backend_path, runner, python_executable
+        )
+        self.config_holder = config_holder
+
+    def build_wheel(
+        self,
+        wheel_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        metadata_directory: Optional[str] = None,
+    ) -> str:
+        cs = self.config_holder.config_settings
+        return super().build_wheel(
+            wheel_directory, config_settings=cs, metadata_directory=metadata_directory
+        )
+
+    def build_sdist(
+        self,
+        sdist_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+    ) -> str:
+        cs = self.config_holder.config_settings
+        return super().build_sdist(sdist_directory, config_settings=cs)
+
+    def build_editable(
+        self,
+        wheel_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        metadata_directory: Optional[str] = None,
+    ) -> str:
+        cs = self.config_holder.config_settings
+        return super().build_editable(
+            wheel_directory, config_settings=cs, metadata_directory=metadata_directory
+        )
+
+    def get_requires_for_build_wheel(
+        self, config_settings: Optional[Mapping[str, Any]] = None
+    ) -> Sequence[str]:
+        cs = self.config_holder.config_settings
+        return super().get_requires_for_build_wheel(config_settings=cs)
+
+    def get_requires_for_build_sdist(
+        self, config_settings: Optional[Mapping[str, Any]] = None
+    ) -> Sequence[str]:
+        cs = self.config_holder.config_settings
+        return super().get_requires_for_build_sdist(config_settings=cs)
+
+    def get_requires_for_build_editable(
+        self, config_settings: Optional[Mapping[str, Any]] = None
+    ) -> Sequence[str]:
+        cs = self.config_holder.config_settings
+        return super().get_requires_for_build_editable(config_settings=cs)
+
+    def prepare_metadata_for_build_wheel(
+        self,
+        metadata_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        _allow_fallback: bool = True,
+    ) -> str:
+        cs = self.config_holder.config_settings
+        return super().prepare_metadata_for_build_wheel(
+            metadata_directory=metadata_directory,
+            config_settings=cs,
+            _allow_fallback=_allow_fallback,
+        )
+
+    def prepare_metadata_for_build_editable(
+        self,
+        metadata_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        _allow_fallback: bool = True,
+    ) -> Optional[str]:
+        cs = self.config_holder.config_settings
+        return super().prepare_metadata_for_build_editable(
+            metadata_directory=metadata_directory,
+            config_settings=cs,
+            _allow_fallback=_allow_fallback,
+        )
+
+
+def warn_if_run_as_root() -> None:
+    """Output a warning for sudo users on Unix.
+
+    In a virtual environment, sudo pip still writes to virtualenv.
+    On Windows, users may run pip as Administrator without issues.
+    This warning only applies to Unix root users outside of virtualenv.
+    """
+    if running_under_virtualenv():
+        return
+    if not hasattr(os, "getuid"):
+        return
+    # On Windows, there are no "system managed" Python packages. Installing as
+    # Administrator via pip is the correct way of updating system environments.
+    #
+    # We choose sys.platform over utils.compat.WINDOWS here to enable Mypy platform
+    # checks: https://mypy.readthedocs.io/en/stable/common_issues.html
+    if sys.platform == "win32" or sys.platform == "cygwin":
+        return
+
+    if os.getuid() != 0:
+        return
+
+    logger.warning(
+        "Running pip as the 'root' user can result in broken permissions and "
+        "conflicting behaviour with the system package manager, possibly "
+        "rendering your system unusable. "
+        "It is recommended to use a virtual environment instead: "
+        "https://pip.pypa.io/warnings/venv. "
+        "Use the --root-user-action option if you know what you are doing and "
+        "want to suppress this warning."
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/packaging.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/packaging.py
new file mode 100644
index 00000000..caad70f7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/packaging.py
@@ -0,0 +1,58 @@
+import functools
+import logging
+import re
+from typing import NewType, Optional, Tuple, cast
+
+from pip._vendor.packaging import specifiers, version
+from pip._vendor.packaging.requirements import Requirement
+
+NormalizedExtra = NewType("NormalizedExtra", str)
+
+logger = logging.getLogger(__name__)
+
+
+@functools.lru_cache(maxsize=32)
+def check_requires_python(
+    requires_python: Optional[str], version_info: Tuple[int, ...]
+) -> bool:
+    """
+    Check if the given Python version matches a "Requires-Python" specifier.
+
+    :param version_info: A 3-tuple of ints representing a Python
+        major-minor-micro version to check (e.g. `sys.version_info[:3]`).
+
+    :return: `True` if the given Python version satisfies the requirement.
+        Otherwise, return `False`.
+
+    :raises InvalidSpecifier: If `requires_python` has an invalid format.
+    """
+    if requires_python is None:
+        # The package provides no information
+        return True
+    requires_python_specifier = specifiers.SpecifierSet(requires_python)
+
+    python_version = version.parse(".".join(map(str, version_info)))
+    return python_version in requires_python_specifier
+
+
+@functools.lru_cache(maxsize=2048)
+def get_requirement(req_string: str) -> Requirement:
+    """Construct a packaging.Requirement object with caching"""
+    # Parsing requirement strings is expensive, and is also expected to happen
+    # with a low diversity of different arguments (at least relative the number
+    # constructed). This method adds a cache to requirement object creation to
+    # minimize repeated parsing of the same string to construct equivalent
+    # Requirement objects.
+    return Requirement(req_string)
+
+
+def safe_extra(extra: str) -> NormalizedExtra:
+    """Convert an arbitrary string to a standard 'extra' name
+
+    Any runs of non-alphanumeric characters are replaced with a single '_',
+    and the result is always lowercased.
+
+    This function is duplicated from ``pkg_resources``. Note that this is not
+    the same to either ``canonicalize_name`` or ``_egg_link_name``.
+    """
+    return cast(NormalizedExtra, re.sub("[^A-Za-z0-9.-]+", "_", extra).lower())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/retry.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/retry.py
new file mode 100644
index 00000000..abfe0728
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/retry.py
@@ -0,0 +1,42 @@
+import functools
+from time import perf_counter, sleep
+from typing import Callable, TypeVar
+
+from pip._vendor.typing_extensions import ParamSpec
+
+T = TypeVar("T")
+P = ParamSpec("P")
+
+
+def retry(
+    wait: float, stop_after_delay: float
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """Decorator to automatically retry a function on error.
+
+    If the function raises, the function is recalled with the same arguments
+    until it returns or the time limit is reached. When the time limit is
+    surpassed, the last exception raised is reraised.
+
+    :param wait: The time to wait after an error before retrying, in seconds.
+    :param stop_after_delay: The time limit after which retries will cease,
+        in seconds.
+    """
+
+    def wrapper(func: Callable[P, T]) -> Callable[P, T]:
+
+        @functools.wraps(func)
+        def retry_wrapped(*args: P.args, **kwargs: P.kwargs) -> T:
+            # The performance counter is monotonic on all platforms we care
+            # about and has much better resolution than time.monotonic().
+            start_time = perf_counter()
+            while True:
+                try:
+                    return func(*args, **kwargs)
+                except Exception:
+                    if perf_counter() - start_time > stop_after_delay:
+                        raise
+                    sleep(wait)
+
+        return retry_wrapped
+
+    return wrapper
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/setuptools_build.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/setuptools_build.py
new file mode 100644
index 00000000..96d1b246
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/setuptools_build.py
@@ -0,0 +1,146 @@
+import sys
+import textwrap
+from typing import List, Optional, Sequence
+
+# Shim to wrap setup.py invocation with setuptools
+# Note that __file__ is handled via two {!r} *and* %r, to ensure that paths on
+# Windows are correctly handled (it should be "C:\\Users" not "C:\Users").
+_SETUPTOOLS_SHIM = textwrap.dedent(
+    """
+    exec(compile('''
+    # This is <pip-setuptools-caller> -- a caller that pip uses to run setup.py
+    #
+    # - It imports setuptools before invoking setup.py, to enable projects that directly
+    #   import from `distutils.core` to work with newer packaging standards.
+    # - It provides a clear error message when setuptools is not installed.
+    # - It sets `sys.argv[0]` to the underlying `setup.py`, when invoking `setup.py` so
+    #   setuptools doesn't think the script is `-c`. This avoids the following warning:
+    #     manifest_maker: standard file '-c' not found".
+    # - It generates a shim setup.py, for handling setup.cfg-only projects.
+    import os, sys, tokenize
+
+    try:
+        import setuptools
+    except ImportError as error:
+        print(
+            "ERROR: Can not execute `setup.py` since setuptools is not available in "
+            "the build environment.",
+            file=sys.stderr,
+        )
+        sys.exit(1)
+
+    __file__ = %r
+    sys.argv[0] = __file__
+
+    if os.path.exists(__file__):
+        filename = __file__
+        with tokenize.open(__file__) as f:
+            setup_py_code = f.read()
+    else:
+        filename = "<auto-generated setuptools caller>"
+        setup_py_code = "from setuptools import setup; setup()"
+
+    exec(compile(setup_py_code, filename, "exec"))
+    ''' % ({!r},), "<pip-setuptools-caller>", "exec"))
+    """
+).rstrip()
+
+
+def make_setuptools_shim_args(
+    setup_py_path: str,
+    global_options: Optional[Sequence[str]] = None,
+    no_user_config: bool = False,
+    unbuffered_output: bool = False,
+) -> List[str]:
+    """
+    Get setuptools command arguments with shim wrapped setup file invocation.
+
+    :param setup_py_path: The path to setup.py to be wrapped.
+    :param global_options: Additional global options.
+    :param no_user_config: If True, disables personal user configuration.
+    :param unbuffered_output: If True, adds the unbuffered switch to the
+     argument list.
+    """
+    args = [sys.executable]
+    if unbuffered_output:
+        args += ["-u"]
+    args += ["-c", _SETUPTOOLS_SHIM.format(setup_py_path)]
+    if global_options:
+        args += global_options
+    if no_user_config:
+        args += ["--no-user-cfg"]
+    return args
+
+
+def make_setuptools_bdist_wheel_args(
+    setup_py_path: str,
+    global_options: Sequence[str],
+    build_options: Sequence[str],
+    destination_dir: str,
+) -> List[str]:
+    # NOTE: Eventually, we'd want to also -S to the flags here, when we're
+    # isolating. Currently, it breaks Python in virtualenvs, because it
+    # relies on site.py to find parts of the standard library outside the
+    # virtualenv.
+    args = make_setuptools_shim_args(
+        setup_py_path, global_options=global_options, unbuffered_output=True
+    )
+    args += ["bdist_wheel", "-d", destination_dir]
+    args += build_options
+    return args
+
+
+def make_setuptools_clean_args(
+    setup_py_path: str,
+    global_options: Sequence[str],
+) -> List[str]:
+    args = make_setuptools_shim_args(
+        setup_py_path, global_options=global_options, unbuffered_output=True
+    )
+    args += ["clean", "--all"]
+    return args
+
+
+def make_setuptools_develop_args(
+    setup_py_path: str,
+    *,
+    global_options: Sequence[str],
+    no_user_config: bool,
+    prefix: Optional[str],
+    home: Optional[str],
+    use_user_site: bool,
+) -> List[str]:
+    assert not (use_user_site and prefix)
+
+    args = make_setuptools_shim_args(
+        setup_py_path,
+        global_options=global_options,
+        no_user_config=no_user_config,
+    )
+
+    args += ["develop", "--no-deps"]
+
+    if prefix:
+        args += ["--prefix", prefix]
+    if home is not None:
+        args += ["--install-dir", home]
+
+    if use_user_site:
+        args += ["--user", "--prefix="]
+
+    return args
+
+
+def make_setuptools_egg_info_args(
+    setup_py_path: str,
+    egg_info_dir: Optional[str],
+    no_user_config: bool,
+) -> List[str]:
+    args = make_setuptools_shim_args(setup_py_path, no_user_config=no_user_config)
+
+    args += ["egg_info"]
+
+    if egg_info_dir:
+        args += ["--egg-base", egg_info_dir]
+
+    return args
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/subprocess.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/subprocess.py
new file mode 100644
index 00000000..cb2e23f0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/subprocess.py
@@ -0,0 +1,245 @@
+import logging
+import os
+import shlex
+import subprocess
+from typing import Any, Callable, Iterable, List, Literal, Mapping, Optional, Union
+
+from pip._vendor.rich.markup import escape
+
+from pip._internal.cli.spinners import SpinnerInterface, open_spinner
+from pip._internal.exceptions import InstallationSubprocessError
+from pip._internal.utils.logging import VERBOSE, subprocess_logger
+from pip._internal.utils.misc import HiddenText
+
+CommandArgs = List[Union[str, HiddenText]]
+
+
+def make_command(*args: Union[str, HiddenText, CommandArgs]) -> CommandArgs:
+    """
+    Create a CommandArgs object.
+    """
+    command_args: CommandArgs = []
+    for arg in args:
+        # Check for list instead of CommandArgs since CommandArgs is
+        # only known during type-checking.
+        if isinstance(arg, list):
+            command_args.extend(arg)
+        else:
+            # Otherwise, arg is str or HiddenText.
+            command_args.append(arg)
+
+    return command_args
+
+
+def format_command_args(args: Union[List[str], CommandArgs]) -> str:
+    """
+    Format command arguments for display.
+    """
+    # For HiddenText arguments, display the redacted form by calling str().
+    # Also, we don't apply str() to arguments that aren't HiddenText since
+    # this can trigger a UnicodeDecodeError in Python 2 if the argument
+    # has type unicode and includes a non-ascii character.  (The type
+    # checker doesn't ensure the annotations are correct in all cases.)
+    return " ".join(
+        shlex.quote(str(arg)) if isinstance(arg, HiddenText) else shlex.quote(arg)
+        for arg in args
+    )
+
+
+def reveal_command_args(args: Union[List[str], CommandArgs]) -> List[str]:
+    """
+    Return the arguments in their raw, unredacted form.
+    """
+    return [arg.secret if isinstance(arg, HiddenText) else arg for arg in args]
+
+
+def call_subprocess(
+    cmd: Union[List[str], CommandArgs],
+    show_stdout: bool = False,
+    cwd: Optional[str] = None,
+    on_returncode: 'Literal["raise", "warn", "ignore"]' = "raise",
+    extra_ok_returncodes: Optional[Iterable[int]] = None,
+    extra_environ: Optional[Mapping[str, Any]] = None,
+    unset_environ: Optional[Iterable[str]] = None,
+    spinner: Optional[SpinnerInterface] = None,
+    log_failed_cmd: Optional[bool] = True,
+    stdout_only: Optional[bool] = False,
+    *,
+    command_desc: str,
+) -> str:
+    """
+    Args:
+      show_stdout: if true, use INFO to log the subprocess's stderr and
+        stdout streams.  Otherwise, use DEBUG.  Defaults to False.
+      extra_ok_returncodes: an iterable of integer return codes that are
+        acceptable, in addition to 0. Defaults to None, which means [].
+      unset_environ: an iterable of environment variable names to unset
+        prior to calling subprocess.Popen().
+      log_failed_cmd: if false, failed commands are not logged, only raised.
+      stdout_only: if true, return only stdout, else return both. When true,
+        logging of both stdout and stderr occurs when the subprocess has
+        terminated, else logging occurs as subprocess output is produced.
+    """
+    if extra_ok_returncodes is None:
+        extra_ok_returncodes = []
+    if unset_environ is None:
+        unset_environ = []
+    # Most places in pip use show_stdout=False. What this means is--
+    #
+    # - We connect the child's output (combined stderr and stdout) to a
+    #   single pipe, which we read.
+    # - We log this output to stderr at DEBUG level as it is received.
+    # - If DEBUG logging isn't enabled (e.g. if --verbose logging wasn't
+    #   requested), then we show a spinner so the user can still see the
+    #   subprocess is in progress.
+    # - If the subprocess exits with an error, we log the output to stderr
+    #   at ERROR level if it hasn't already been displayed to the console
+    #   (e.g. if --verbose logging wasn't enabled).  This way we don't log
+    #   the output to the console twice.
+    #
+    # If show_stdout=True, then the above is still done, but with DEBUG
+    # replaced by INFO.
+    if show_stdout:
+        # Then log the subprocess output at INFO level.
+        log_subprocess: Callable[..., None] = subprocess_logger.info
+        used_level = logging.INFO
+    else:
+        # Then log the subprocess output using VERBOSE.  This also ensures
+        # it will be logged to the log file (aka user_log), if enabled.
+        log_subprocess = subprocess_logger.verbose
+        used_level = VERBOSE
+
+    # Whether the subprocess will be visible in the console.
+    showing_subprocess = subprocess_logger.getEffectiveLevel() <= used_level
+
+    # Only use the spinner if we're not showing the subprocess output
+    # and we have a spinner.
+    use_spinner = not showing_subprocess and spinner is not None
+
+    log_subprocess("Running command %s", command_desc)
+    env = os.environ.copy()
+    if extra_environ:
+        env.update(extra_environ)
+    for name in unset_environ:
+        env.pop(name, None)
+    try:
+        proc = subprocess.Popen(
+            # Convert HiddenText objects to the underlying str.
+            reveal_command_args(cmd),
+            stdin=subprocess.PIPE,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.STDOUT if not stdout_only else subprocess.PIPE,
+            cwd=cwd,
+            env=env,
+            errors="backslashreplace",
+        )
+    except Exception as exc:
+        if log_failed_cmd:
+            subprocess_logger.critical(
+                "Error %s while executing command %s",
+                exc,
+                command_desc,
+            )
+        raise
+    all_output = []
+    if not stdout_only:
+        assert proc.stdout
+        assert proc.stdin
+        proc.stdin.close()
+        # In this mode, stdout and stderr are in the same pipe.
+        while True:
+            line: str = proc.stdout.readline()
+            if not line:
+                break
+            line = line.rstrip()
+            all_output.append(line + "\n")
+
+            # Show the line immediately.
+            log_subprocess(line)
+            # Update the spinner.
+            if use_spinner:
+                assert spinner
+                spinner.spin()
+        try:
+            proc.wait()
+        finally:
+            if proc.stdout:
+                proc.stdout.close()
+        output = "".join(all_output)
+    else:
+        # In this mode, stdout and stderr are in different pipes.
+        # We must use communicate() which is the only safe way to read both.
+        out, err = proc.communicate()
+        # log line by line to preserve pip log indenting
+        for out_line in out.splitlines():
+            log_subprocess(out_line)
+        all_output.append(out)
+        for err_line in err.splitlines():
+            log_subprocess(err_line)
+        all_output.append(err)
+        output = out
+
+    proc_had_error = proc.returncode and proc.returncode not in extra_ok_returncodes
+    if use_spinner:
+        assert spinner
+        if proc_had_error:
+            spinner.finish("error")
+        else:
+            spinner.finish("done")
+    if proc_had_error:
+        if on_returncode == "raise":
+            error = InstallationSubprocessError(
+                command_description=command_desc,
+                exit_code=proc.returncode,
+                output_lines=all_output if not showing_subprocess else None,
+            )
+            if log_failed_cmd:
+                subprocess_logger.error("%s", error, extra={"rich": True})
+                subprocess_logger.verbose(
+                    "[bold magenta]full command[/]: [blue]%s[/]",
+                    escape(format_command_args(cmd)),
+                    extra={"markup": True},
+                )
+                subprocess_logger.verbose(
+                    "[bold magenta]cwd[/]: %s",
+                    escape(cwd or "[inherit]"),
+                    extra={"markup": True},
+                )
+
+            raise error
+        elif on_returncode == "warn":
+            subprocess_logger.warning(
+                'Command "%s" had error code %s in %s',
+                command_desc,
+                proc.returncode,
+                cwd,
+            )
+        elif on_returncode == "ignore":
+            pass
+        else:
+            raise ValueError(f"Invalid value: on_returncode={on_returncode!r}")
+    return output
+
+
+def runner_with_spinner_message(message: str) -> Callable[..., None]:
+    """Provide a subprocess_runner that shows a spinner message.
+
+    Intended for use with for BuildBackendHookCaller. Thus, the runner has
+    an API that matches what's expected by BuildBackendHookCaller.subprocess_runner.
+    """
+
+    def runner(
+        cmd: List[str],
+        cwd: Optional[str] = None,
+        extra_environ: Optional[Mapping[str, Any]] = None,
+    ) -> None:
+        with open_spinner(message) as spinner:
+            call_subprocess(
+                cmd,
+                command_desc=message,
+                cwd=cwd,
+                extra_environ=extra_environ,
+                spinner=spinner,
+            )
+
+    return runner
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/temp_dir.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/temp_dir.py
new file mode 100644
index 00000000..06668e8a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/temp_dir.py
@@ -0,0 +1,296 @@
+import errno
+import itertools
+import logging
+import os.path
+import tempfile
+import traceback
+from contextlib import ExitStack, contextmanager
+from pathlib import Path
+from typing import (
+    Any,
+    Callable,
+    Dict,
+    Generator,
+    List,
+    Optional,
+    TypeVar,
+    Union,
+)
+
+from pip._internal.utils.misc import enum, rmtree
+
+logger = logging.getLogger(__name__)
+
+_T = TypeVar("_T", bound="TempDirectory")
+
+
+# Kinds of temporary directories. Only needed for ones that are
+# globally-managed.
+tempdir_kinds = enum(
+    BUILD_ENV="build-env",
+    EPHEM_WHEEL_CACHE="ephem-wheel-cache",
+    REQ_BUILD="req-build",
+)
+
+
+_tempdir_manager: Optional[ExitStack] = None
+
+
+@contextmanager
+def global_tempdir_manager() -> Generator[None, None, None]:
+    global _tempdir_manager
+    with ExitStack() as stack:
+        old_tempdir_manager, _tempdir_manager = _tempdir_manager, stack
+        try:
+            yield
+        finally:
+            _tempdir_manager = old_tempdir_manager
+
+
+class TempDirectoryTypeRegistry:
+    """Manages temp directory behavior"""
+
+    def __init__(self) -> None:
+        self._should_delete: Dict[str, bool] = {}
+
+    def set_delete(self, kind: str, value: bool) -> None:
+        """Indicate whether a TempDirectory of the given kind should be
+        auto-deleted.
+        """
+        self._should_delete[kind] = value
+
+    def get_delete(self, kind: str) -> bool:
+        """Get configured auto-delete flag for a given TempDirectory type,
+        default True.
+        """
+        return self._should_delete.get(kind, True)
+
+
+_tempdir_registry: Optional[TempDirectoryTypeRegistry] = None
+
+
+@contextmanager
+def tempdir_registry() -> Generator[TempDirectoryTypeRegistry, None, None]:
+    """Provides a scoped global tempdir registry that can be used to dictate
+    whether directories should be deleted.
+    """
+    global _tempdir_registry
+    old_tempdir_registry = _tempdir_registry
+    _tempdir_registry = TempDirectoryTypeRegistry()
+    try:
+        yield _tempdir_registry
+    finally:
+        _tempdir_registry = old_tempdir_registry
+
+
+class _Default:
+    pass
+
+
+_default = _Default()
+
+
+class TempDirectory:
+    """Helper class that owns and cleans up a temporary directory.
+
+    This class can be used as a context manager or as an OO representation of a
+    temporary directory.
+
+    Attributes:
+        path
+            Location to the created temporary directory
+        delete
+            Whether the directory should be deleted when exiting
+            (when used as a contextmanager)
+
+    Methods:
+        cleanup()
+            Deletes the temporary directory
+
+    When used as a context manager, if the delete attribute is True, on
+    exiting the context the temporary directory is deleted.
+    """
+
+    def __init__(
+        self,
+        path: Optional[str] = None,
+        delete: Union[bool, None, _Default] = _default,
+        kind: str = "temp",
+        globally_managed: bool = False,
+        ignore_cleanup_errors: bool = True,
+    ):
+        super().__init__()
+
+        if delete is _default:
+            if path is not None:
+                # If we were given an explicit directory, resolve delete option
+                # now.
+                delete = False
+            else:
+                # Otherwise, we wait until cleanup and see what
+                # tempdir_registry says.
+                delete = None
+
+        # The only time we specify path is in for editables where it
+        # is the value of the --src option.
+        if path is None:
+            path = self._create(kind)
+
+        self._path = path
+        self._deleted = False
+        self.delete = delete
+        self.kind = kind
+        self.ignore_cleanup_errors = ignore_cleanup_errors
+
+        if globally_managed:
+            assert _tempdir_manager is not None
+            _tempdir_manager.enter_context(self)
+
+    @property
+    def path(self) -> str:
+        assert not self._deleted, f"Attempted to access deleted path: {self._path}"
+        return self._path
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__} {self.path!r}>"
+
+    def __enter__(self: _T) -> _T:
+        return self
+
+    def __exit__(self, exc: Any, value: Any, tb: Any) -> None:
+        if self.delete is not None:
+            delete = self.delete
+        elif _tempdir_registry:
+            delete = _tempdir_registry.get_delete(self.kind)
+        else:
+            delete = True
+
+        if delete:
+            self.cleanup()
+
+    def _create(self, kind: str) -> str:
+        """Create a temporary directory and store its path in self.path"""
+        # We realpath here because some systems have their default tmpdir
+        # symlinked to another directory.  This tends to confuse build
+        # scripts, so we canonicalize the path by traversing potential
+        # symlinks here.
+        path = os.path.realpath(tempfile.mkdtemp(prefix=f"pip-{kind}-"))
+        logger.debug("Created temporary directory: %s", path)
+        return path
+
+    def cleanup(self) -> None:
+        """Remove the temporary directory created and reset state"""
+        self._deleted = True
+        if not os.path.exists(self._path):
+            return
+
+        errors: List[BaseException] = []
+
+        def onerror(
+            func: Callable[..., Any],
+            path: Path,
+            exc_val: BaseException,
+        ) -> None:
+            """Log a warning for a `rmtree` error and continue"""
+            formatted_exc = "\n".join(
+                traceback.format_exception_only(type(exc_val), exc_val)
+            )
+            formatted_exc = formatted_exc.rstrip()  # remove trailing new line
+            if func in (os.unlink, os.remove, os.rmdir):
+                logger.debug(
+                    "Failed to remove a temporary file '%s' due to %s.\n",
+                    path,
+                    formatted_exc,
+                )
+            else:
+                logger.debug("%s failed with %s.", func.__qualname__, formatted_exc)
+            errors.append(exc_val)
+
+        if self.ignore_cleanup_errors:
+            try:
+                # first try with @retry; retrying to handle ephemeral errors
+                rmtree(self._path, ignore_errors=False)
+            except OSError:
+                # last pass ignore/log all errors
+                rmtree(self._path, onexc=onerror)
+            if errors:
+                logger.warning(
+                    "Failed to remove contents in a temporary directory '%s'.\n"
+                    "You can safely remove it manually.",
+                    self._path,
+                )
+        else:
+            rmtree(self._path)
+
+
+class AdjacentTempDirectory(TempDirectory):
+    """Helper class that creates a temporary directory adjacent to a real one.
+
+    Attributes:
+        original
+            The original directory to create a temp directory for.
+        path
+            After calling create() or entering, contains the full
+            path to the temporary directory.
+        delete
+            Whether the directory should be deleted when exiting
+            (when used as a contextmanager)
+
+    """
+
+    # The characters that may be used to name the temp directory
+    # We always prepend a ~ and then rotate through these until
+    # a usable name is found.
+    # pkg_resources raises a different error for .dist-info folder
+    # with leading '-' and invalid metadata
+    LEADING_CHARS = "-~.=%0123456789"
+
+    def __init__(self, original: str, delete: Optional[bool] = None) -> None:
+        self.original = original.rstrip("/\\")
+        super().__init__(delete=delete)
+
+    @classmethod
+    def _generate_names(cls, name: str) -> Generator[str, None, None]:
+        """Generates a series of temporary names.
+
+        The algorithm replaces the leading characters in the name
+        with ones that are valid filesystem characters, but are not
+        valid package names (for both Python and pip definitions of
+        package).
+        """
+        for i in range(1, len(name)):
+            for candidate in itertools.combinations_with_replacement(
+                cls.LEADING_CHARS, i - 1
+            ):
+                new_name = "~" + "".join(candidate) + name[i:]
+                if new_name != name:
+                    yield new_name
+
+        # If we make it this far, we will have to make a longer name
+        for i in range(len(cls.LEADING_CHARS)):
+            for candidate in itertools.combinations_with_replacement(
+                cls.LEADING_CHARS, i
+            ):
+                new_name = "~" + "".join(candidate) + name
+                if new_name != name:
+                    yield new_name
+
+    def _create(self, kind: str) -> str:
+        root, name = os.path.split(self.original)
+        for candidate in self._generate_names(name):
+            path = os.path.join(root, candidate)
+            try:
+                os.mkdir(path)
+            except OSError as ex:
+                # Continue if the name exists already
+                if ex.errno != errno.EEXIST:
+                    raise
+            else:
+                path = os.path.realpath(path)
+                break
+        else:
+            # Final fallback on the default behavior.
+            path = os.path.realpath(tempfile.mkdtemp(prefix=f"pip-{kind}-"))
+
+        logger.debug("Created temporary directory: %s", path)
+        return path
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/unpacking.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/unpacking.py
new file mode 100644
index 00000000..87a6d19a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/unpacking.py
@@ -0,0 +1,337 @@
+"""Utilities related archives.
+"""
+
+import logging
+import os
+import shutil
+import stat
+import sys
+import tarfile
+import zipfile
+from typing import Iterable, List, Optional
+from zipfile import ZipInfo
+
+from pip._internal.exceptions import InstallationError
+from pip._internal.utils.filetypes import (
+    BZ2_EXTENSIONS,
+    TAR_EXTENSIONS,
+    XZ_EXTENSIONS,
+    ZIP_EXTENSIONS,
+)
+from pip._internal.utils.misc import ensure_dir
+
+logger = logging.getLogger(__name__)
+
+
+SUPPORTED_EXTENSIONS = ZIP_EXTENSIONS + TAR_EXTENSIONS
+
+try:
+    import bz2  # noqa
+
+    SUPPORTED_EXTENSIONS += BZ2_EXTENSIONS
+except ImportError:
+    logger.debug("bz2 module is not available")
+
+try:
+    # Only for Python 3.3+
+    import lzma  # noqa
+
+    SUPPORTED_EXTENSIONS += XZ_EXTENSIONS
+except ImportError:
+    logger.debug("lzma module is not available")
+
+
+def current_umask() -> int:
+    """Get the current umask which involves having to set it temporarily."""
+    mask = os.umask(0)
+    os.umask(mask)
+    return mask
+
+
+def split_leading_dir(path: str) -> List[str]:
+    path = path.lstrip("/").lstrip("\\")
+    if "/" in path and (
+        ("\\" in path and path.find("/") < path.find("\\")) or "\\" not in path
+    ):
+        return path.split("/", 1)
+    elif "\\" in path:
+        return path.split("\\", 1)
+    else:
+        return [path, ""]
+
+
+def has_leading_dir(paths: Iterable[str]) -> bool:
+    """Returns true if all the paths have the same leading path name
+    (i.e., everything is in one subdirectory in an archive)"""
+    common_prefix = None
+    for path in paths:
+        prefix, rest = split_leading_dir(path)
+        if not prefix:
+            return False
+        elif common_prefix is None:
+            common_prefix = prefix
+        elif prefix != common_prefix:
+            return False
+    return True
+
+
+def is_within_directory(directory: str, target: str) -> bool:
+    """
+    Return true if the absolute path of target is within the directory
+    """
+    abs_directory = os.path.abspath(directory)
+    abs_target = os.path.abspath(target)
+
+    prefix = os.path.commonprefix([abs_directory, abs_target])
+    return prefix == abs_directory
+
+
+def _get_default_mode_plus_executable() -> int:
+    return 0o777 & ~current_umask() | 0o111
+
+
+def set_extracted_file_to_default_mode_plus_executable(path: str) -> None:
+    """
+    Make file present at path have execute for user/group/world
+    (chmod +x) is no-op on windows per python docs
+    """
+    os.chmod(path, _get_default_mode_plus_executable())
+
+
+def zip_item_is_executable(info: ZipInfo) -> bool:
+    mode = info.external_attr >> 16
+    # if mode and regular file and any execute permissions for
+    # user/group/world?
+    return bool(mode and stat.S_ISREG(mode) and mode & 0o111)
+
+
+def unzip_file(filename: str, location: str, flatten: bool = True) -> None:
+    """
+    Unzip the file (with path `filename`) to the destination `location`.  All
+    files are written based on system defaults and umask (i.e. permissions are
+    not preserved), except that regular file members with any execute
+    permissions (user, group, or world) have "chmod +x" applied after being
+    written. Note that for windows, any execute changes using os.chmod are
+    no-ops per the python docs.
+    """
+    ensure_dir(location)
+    zipfp = open(filename, "rb")
+    try:
+        zip = zipfile.ZipFile(zipfp, allowZip64=True)
+        leading = has_leading_dir(zip.namelist()) and flatten
+        for info in zip.infolist():
+            name = info.filename
+            fn = name
+            if leading:
+                fn = split_leading_dir(name)[1]
+            fn = os.path.join(location, fn)
+            dir = os.path.dirname(fn)
+            if not is_within_directory(location, fn):
+                message = (
+                    "The zip file ({}) has a file ({}) trying to install "
+                    "outside target directory ({})"
+                )
+                raise InstallationError(message.format(filename, fn, location))
+            if fn.endswith("/") or fn.endswith("\\"):
+                # A directory
+                ensure_dir(fn)
+            else:
+                ensure_dir(dir)
+                # Don't use read() to avoid allocating an arbitrarily large
+                # chunk of memory for the file's content
+                fp = zip.open(name)
+                try:
+                    with open(fn, "wb") as destfp:
+                        shutil.copyfileobj(fp, destfp)
+                finally:
+                    fp.close()
+                    if zip_item_is_executable(info):
+                        set_extracted_file_to_default_mode_plus_executable(fn)
+    finally:
+        zipfp.close()
+
+
+def untar_file(filename: str, location: str) -> None:
+    """
+    Untar the file (with path `filename`) to the destination `location`.
+    All files are written based on system defaults and umask (i.e. permissions
+    are not preserved), except that regular file members with any execute
+    permissions (user, group, or world) have "chmod +x" applied on top of the
+    default.  Note that for windows, any execute changes using os.chmod are
+    no-ops per the python docs.
+    """
+    ensure_dir(location)
+    if filename.lower().endswith(".gz") or filename.lower().endswith(".tgz"):
+        mode = "r:gz"
+    elif filename.lower().endswith(BZ2_EXTENSIONS):
+        mode = "r:bz2"
+    elif filename.lower().endswith(XZ_EXTENSIONS):
+        mode = "r:xz"
+    elif filename.lower().endswith(".tar"):
+        mode = "r"
+    else:
+        logger.warning(
+            "Cannot determine compression type for file %s",
+            filename,
+        )
+        mode = "r:*"
+
+    tar = tarfile.open(filename, mode, encoding="utf-8")  # type: ignore
+    try:
+        leading = has_leading_dir([member.name for member in tar.getmembers()])
+
+        # PEP 706 added `tarfile.data_filter`, and made some other changes to
+        # Python's tarfile module (see below). The features were backported to
+        # security releases.
+        try:
+            data_filter = tarfile.data_filter
+        except AttributeError:
+            _untar_without_filter(filename, location, tar, leading)
+        else:
+            default_mode_plus_executable = _get_default_mode_plus_executable()
+
+            if leading:
+                # Strip the leading directory from all files in the archive,
+                # including hardlink targets (which are relative to the
+                # unpack location).
+                for member in tar.getmembers():
+                    name_lead, name_rest = split_leading_dir(member.name)
+                    member.name = name_rest
+                    if member.islnk():
+                        lnk_lead, lnk_rest = split_leading_dir(member.linkname)
+                        if lnk_lead == name_lead:
+                            member.linkname = lnk_rest
+
+            def pip_filter(member: tarfile.TarInfo, path: str) -> tarfile.TarInfo:
+                orig_mode = member.mode
+                try:
+                    try:
+                        member = data_filter(member, location)
+                    except tarfile.LinkOutsideDestinationError:
+                        if sys.version_info[:3] in {
+                            (3, 8, 17),
+                            (3, 9, 17),
+                            (3, 10, 12),
+                            (3, 11, 4),
+                        }:
+                            # The tarfile filter in specific Python versions
+                            # raises LinkOutsideDestinationError on valid input
+                            # (https://github.com/python/cpython/issues/107845)
+                            # Ignore the error there, but do use the
+                            # more lax `tar_filter`
+                            member = tarfile.tar_filter(member, location)
+                        else:
+                            raise
+                except tarfile.TarError as exc:
+                    message = "Invalid member in the tar file {}: {}"
+                    # Filter error messages mention the member name.
+                    # No need to add it here.
+                    raise InstallationError(
+                        message.format(
+                            filename,
+                            exc,
+                        )
+                    )
+                if member.isfile() and orig_mode & 0o111:
+                    member.mode = default_mode_plus_executable
+                else:
+                    # See PEP 706 note above.
+                    # The PEP changed this from `int` to `Optional[int]`,
+                    # where None means "use the default". Mypy doesn't
+                    # know this yet.
+                    member.mode = None  # type: ignore [assignment]
+                return member
+
+            tar.extractall(location, filter=pip_filter)
+
+    finally:
+        tar.close()
+
+
+def _untar_without_filter(
+    filename: str,
+    location: str,
+    tar: tarfile.TarFile,
+    leading: bool,
+) -> None:
+    """Fallback for Python without tarfile.data_filter"""
+    for member in tar.getmembers():
+        fn = member.name
+        if leading:
+            fn = split_leading_dir(fn)[1]
+        path = os.path.join(location, fn)
+        if not is_within_directory(location, path):
+            message = (
+                "The tar file ({}) has a file ({}) trying to install "
+                "outside target directory ({})"
+            )
+            raise InstallationError(message.format(filename, path, location))
+        if member.isdir():
+            ensure_dir(path)
+        elif member.issym():
+            try:
+                tar._extract_member(member, path)
+            except Exception as exc:
+                # Some corrupt tar files seem to produce this
+                # (specifically bad symlinks)
+                logger.warning(
+                    "In the tar file %s the member %s is invalid: %s",
+                    filename,
+                    member.name,
+                    exc,
+                )
+                continue
+        else:
+            try:
+                fp = tar.extractfile(member)
+            except (KeyError, AttributeError) as exc:
+                # Some corrupt tar files seem to produce this
+                # (specifically bad symlinks)
+                logger.warning(
+                    "In the tar file %s the member %s is invalid: %s",
+                    filename,
+                    member.name,
+                    exc,
+                )
+                continue
+            ensure_dir(os.path.dirname(path))
+            assert fp is not None
+            with open(path, "wb") as destfp:
+                shutil.copyfileobj(fp, destfp)
+            fp.close()
+            # Update the timestamp (useful for cython compiled files)
+            tar.utime(member, path)
+            # member have any execute permissions for user/group/world?
+            if member.mode & 0o111:
+                set_extracted_file_to_default_mode_plus_executable(path)
+
+
+def unpack_file(
+    filename: str,
+    location: str,
+    content_type: Optional[str] = None,
+) -> None:
+    filename = os.path.realpath(filename)
+    if (
+        content_type == "application/zip"
+        or filename.lower().endswith(ZIP_EXTENSIONS)
+        or zipfile.is_zipfile(filename)
+    ):
+        unzip_file(filename, location, flatten=not filename.endswith(".whl"))
+    elif (
+        content_type == "application/x-gzip"
+        or tarfile.is_tarfile(filename)
+        or filename.lower().endswith(TAR_EXTENSIONS + BZ2_EXTENSIONS + XZ_EXTENSIONS)
+    ):
+        untar_file(filename, location)
+    else:
+        # FIXME: handle?
+        # FIXME: magic signatures?
+        logger.critical(
+            "Cannot unpack file %s (downloaded from %s, content-type: %s); "
+            "cannot detect archive format",
+            filename,
+            location,
+            content_type,
+        )
+        raise InstallationError(f"Cannot determine archive format of {location}")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/urls.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/urls.py
new file mode 100644
index 00000000..9f34f882
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/urls.py
@@ -0,0 +1,55 @@
+import os
+import string
+import urllib.parse
+import urllib.request
+
+from .compat import WINDOWS
+
+
+def path_to_url(path: str) -> str:
+    """
+    Convert a path to a file: URL.  The path will be made absolute and have
+    quoted path parts.
+    """
+    path = os.path.normpath(os.path.abspath(path))
+    url = urllib.parse.urljoin("file:", urllib.request.pathname2url(path))
+    return url
+
+
+def url_to_path(url: str) -> str:
+    """
+    Convert a file: URL to a path.
+    """
+    assert url.startswith(
+        "file:"
+    ), f"You can only turn file: urls into filenames (not {url!r})"
+
+    _, netloc, path, _, _ = urllib.parse.urlsplit(url)
+
+    if not netloc or netloc == "localhost":
+        # According to RFC 8089, same as empty authority.
+        netloc = ""
+    elif WINDOWS:
+        # If we have a UNC path, prepend UNC share notation.
+        netloc = "\\\\" + netloc
+    else:
+        raise ValueError(
+            f"non-local file URIs are not supported on this platform: {url!r}"
+        )
+
+    path = urllib.request.url2pathname(netloc + path)
+
+    # On Windows, urlsplit parses the path as something like "/C:/Users/foo".
+    # This creates issues for path-related functions like io.open(), so we try
+    # to detect and strip the leading slash.
+    if (
+        WINDOWS
+        and not netloc  # Not UNC.
+        and len(path) >= 3
+        and path[0] == "/"  # Leading slash to strip.
+        and path[1] in string.ascii_letters  # Drive letter.
+        and path[2:4] in (":", ":/")  # Colon + end of string, or colon + absolute path.
+    ):
+        path = path[1:]
+
+    return path
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/virtualenv.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/virtualenv.py
new file mode 100644
index 00000000..882e36f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/virtualenv.py
@@ -0,0 +1,104 @@
+import logging
+import os
+import re
+import site
+import sys
+from typing import List, Optional
+
+logger = logging.getLogger(__name__)
+_INCLUDE_SYSTEM_SITE_PACKAGES_REGEX = re.compile(
+    r"include-system-site-packages\s*=\s*(?P<value>true|false)"
+)
+
+
+def _running_under_venv() -> bool:
+    """Checks if sys.base_prefix and sys.prefix match.
+
+    This handles PEP 405 compliant virtual environments.
+    """
+    return sys.prefix != getattr(sys, "base_prefix", sys.prefix)
+
+
+def _running_under_legacy_virtualenv() -> bool:
+    """Checks if sys.real_prefix is set.
+
+    This handles virtual environments created with pypa's virtualenv.
+    """
+    # pypa/virtualenv case
+    return hasattr(sys, "real_prefix")
+
+
+def running_under_virtualenv() -> bool:
+    """True if we're running inside a virtual environment, False otherwise."""
+    return _running_under_venv() or _running_under_legacy_virtualenv()
+
+
+def _get_pyvenv_cfg_lines() -> Optional[List[str]]:
+    """Reads {sys.prefix}/pyvenv.cfg and returns its contents as list of lines
+
+    Returns None, if it could not read/access the file.
+    """
+    pyvenv_cfg_file = os.path.join(sys.prefix, "pyvenv.cfg")
+    try:
+        # Although PEP 405 does not specify, the built-in venv module always
+        # writes with UTF-8. (pypa/pip#8717)
+        with open(pyvenv_cfg_file, encoding="utf-8") as f:
+            return f.read().splitlines()  # avoids trailing newlines
+    except OSError:
+        return None
+
+
+def _no_global_under_venv() -> bool:
+    """Check `{sys.prefix}/pyvenv.cfg` for system site-packages inclusion
+
+    PEP 405 specifies that when system site-packages are not supposed to be
+    visible from a virtual environment, `pyvenv.cfg` must contain the following
+    line:
+
+        include-system-site-packages = false
+
+    Additionally, log a warning if accessing the file fails.
+    """
+    cfg_lines = _get_pyvenv_cfg_lines()
+    if cfg_lines is None:
+        # We're not in a "sane" venv, so assume there is no system
+        # site-packages access (since that's PEP 405's default state).
+        logger.warning(
+            "Could not access 'pyvenv.cfg' despite a virtual environment "
+            "being active. Assuming global site-packages is not accessible "
+            "in this environment."
+        )
+        return True
+
+    for line in cfg_lines:
+        match = _INCLUDE_SYSTEM_SITE_PACKAGES_REGEX.match(line)
+        if match is not None and match.group("value") == "false":
+            return True
+    return False
+
+
+def _no_global_under_legacy_virtualenv() -> bool:
+    """Check if "no-global-site-packages.txt" exists beside site.py
+
+    This mirrors logic in pypa/virtualenv for determining whether system
+    site-packages are visible in the virtual environment.
+    """
+    site_mod_dir = os.path.dirname(os.path.abspath(site.__file__))
+    no_global_site_packages_file = os.path.join(
+        site_mod_dir,
+        "no-global-site-packages.txt",
+    )
+    return os.path.exists(no_global_site_packages_file)
+
+
+def virtualenv_no_global() -> bool:
+    """Returns a boolean, whether running in venv with no system site-packages."""
+    # PEP 405 compliance needs to be checked first since virtualenv >=20 would
+    # return True for both checks, but is only able to use the PEP 405 config.
+    if _running_under_venv():
+        return _no_global_under_venv()
+
+    if _running_under_legacy_virtualenv():
+        return _no_global_under_legacy_virtualenv()
+
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/wheel.py
new file mode 100644
index 00000000..f85aee8a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/utils/wheel.py
@@ -0,0 +1,134 @@
+"""Support functions for working with wheel files.
+"""
+
+import logging
+from email.message import Message
+from email.parser import Parser
+from typing import Tuple
+from zipfile import BadZipFile, ZipFile
+
+from pip._vendor.packaging.utils import canonicalize_name
+
+from pip._internal.exceptions import UnsupportedWheel
+
+VERSION_COMPATIBLE = (1, 0)
+
+
+logger = logging.getLogger(__name__)
+
+
+def parse_wheel(wheel_zip: ZipFile, name: str) -> Tuple[str, Message]:
+    """Extract information from the provided wheel, ensuring it meets basic
+    standards.
+
+    Returns the name of the .dist-info directory and the parsed WHEEL metadata.
+    """
+    try:
+        info_dir = wheel_dist_info_dir(wheel_zip, name)
+        metadata = wheel_metadata(wheel_zip, info_dir)
+        version = wheel_version(metadata)
+    except UnsupportedWheel as e:
+        raise UnsupportedWheel(f"{name} has an invalid wheel, {e}")
+
+    check_compatibility(version, name)
+
+    return info_dir, metadata
+
+
+def wheel_dist_info_dir(source: ZipFile, name: str) -> str:
+    """Returns the name of the contained .dist-info directory.
+
+    Raises AssertionError or UnsupportedWheel if not found, >1 found, or
+    it doesn't match the provided name.
+    """
+    # Zip file path separators must be /
+    subdirs = {p.split("/", 1)[0] for p in source.namelist()}
+
+    info_dirs = [s for s in subdirs if s.endswith(".dist-info")]
+
+    if not info_dirs:
+        raise UnsupportedWheel(".dist-info directory not found")
+
+    if len(info_dirs) > 1:
+        raise UnsupportedWheel(
+            "multiple .dist-info directories found: {}".format(", ".join(info_dirs))
+        )
+
+    info_dir = info_dirs[0]
+
+    info_dir_name = canonicalize_name(info_dir)
+    canonical_name = canonicalize_name(name)
+    if not info_dir_name.startswith(canonical_name):
+        raise UnsupportedWheel(
+            f".dist-info directory {info_dir!r} does not start with {canonical_name!r}"
+        )
+
+    return info_dir
+
+
+def read_wheel_metadata_file(source: ZipFile, path: str) -> bytes:
+    try:
+        return source.read(path)
+        # BadZipFile for general corruption, KeyError for missing entry,
+        # and RuntimeError for password-protected files
+    except (BadZipFile, KeyError, RuntimeError) as e:
+        raise UnsupportedWheel(f"could not read {path!r} file: {e!r}")
+
+
+def wheel_metadata(source: ZipFile, dist_info_dir: str) -> Message:
+    """Return the WHEEL metadata of an extracted wheel, if possible.
+    Otherwise, raise UnsupportedWheel.
+    """
+    path = f"{dist_info_dir}/WHEEL"
+    # Zip file path separators must be /
+    wheel_contents = read_wheel_metadata_file(source, path)
+
+    try:
+        wheel_text = wheel_contents.decode()
+    except UnicodeDecodeError as e:
+        raise UnsupportedWheel(f"error decoding {path!r}: {e!r}")
+
+    # FeedParser (used by Parser) does not raise any exceptions. The returned
+    # message may have .defects populated, but for backwards-compatibility we
+    # currently ignore them.
+    return Parser().parsestr(wheel_text)
+
+
+def wheel_version(wheel_data: Message) -> Tuple[int, ...]:
+    """Given WHEEL metadata, return the parsed Wheel-Version.
+    Otherwise, raise UnsupportedWheel.
+    """
+    version_text = wheel_data["Wheel-Version"]
+    if version_text is None:
+        raise UnsupportedWheel("WHEEL is missing Wheel-Version")
+
+    version = version_text.strip()
+
+    try:
+        return tuple(map(int, version.split(".")))
+    except ValueError:
+        raise UnsupportedWheel(f"invalid Wheel-Version: {version!r}")
+
+
+def check_compatibility(version: Tuple[int, ...], name: str) -> None:
+    """Raises errors or warns if called with an incompatible Wheel-Version.
+
+    pip should refuse to install a Wheel-Version that's a major series
+    ahead of what it's compatible with (e.g 2.0 > 1.1); and warn when
+    installing a version only minor version ahead (e.g 1.2 > 1.1).
+
+    version: a 2-tuple representing a Wheel-Version (Major, Minor)
+    name: name of wheel or package to raise exception about
+
+    :raises UnsupportedWheel: when an incompatible Wheel-Version is given
+    """
+    if version[0] > VERSION_COMPATIBLE[0]:
+        raise UnsupportedWheel(
+            "{}'s Wheel-Version ({}) is not compatible with this version "
+            "of pip".format(name, ".".join(map(str, version)))
+        )
+    elif version > VERSION_COMPATIBLE:
+        logger.warning(
+            "Installing from a newer Wheel-Version (%s)",
+            ".".join(map(str, version)),
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__init__.py
new file mode 100644
index 00000000..b6beddbe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__init__.py
@@ -0,0 +1,15 @@
+# Expose a limited set of classes and functions so callers outside of
+# the vcs package don't need to import deeper than `pip._internal.vcs`.
+# (The test directory may still need to import from a vcs sub-package.)
+# Import all vcs modules to register each VCS in the VcsSupport object.
+import pip._internal.vcs.bazaar
+import pip._internal.vcs.git
+import pip._internal.vcs.mercurial
+import pip._internal.vcs.subversion  # noqa: F401
+from pip._internal.vcs.versioncontrol import (  # noqa: F401
+    RemoteNotFoundError,
+    RemoteNotValidError,
+    is_url,
+    make_vcs_requirement_url,
+    vcs,
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2fa20e8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/bazaar.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/bazaar.cpython-312.pyc
new file mode 100644
index 00000000..470ff0df
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/bazaar.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/git.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/git.cpython-312.pyc
new file mode 100644
index 00000000..757f5cdc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/git.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/mercurial.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/mercurial.cpython-312.pyc
new file mode 100644
index 00000000..585c5ca3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/mercurial.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/subversion.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/subversion.cpython-312.pyc
new file mode 100644
index 00000000..120d1b15
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/subversion.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/versioncontrol.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/versioncontrol.cpython-312.pyc
new file mode 100644
index 00000000..346ca4a9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/__pycache__/versioncontrol.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/bazaar.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/bazaar.py
new file mode 100644
index 00000000..c754b7cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/bazaar.py
@@ -0,0 +1,112 @@
+import logging
+from typing import List, Optional, Tuple
+
+from pip._internal.utils.misc import HiddenText, display_path
+from pip._internal.utils.subprocess import make_command
+from pip._internal.utils.urls import path_to_url
+from pip._internal.vcs.versioncontrol import (
+    AuthInfo,
+    RemoteNotFoundError,
+    RevOptions,
+    VersionControl,
+    vcs,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class Bazaar(VersionControl):
+    name = "bzr"
+    dirname = ".bzr"
+    repo_name = "branch"
+    schemes = (
+        "bzr+http",
+        "bzr+https",
+        "bzr+ssh",
+        "bzr+sftp",
+        "bzr+ftp",
+        "bzr+lp",
+        "bzr+file",
+    )
+
+    @staticmethod
+    def get_base_rev_args(rev: str) -> List[str]:
+        return ["-r", rev]
+
+    def fetch_new(
+        self, dest: str, url: HiddenText, rev_options: RevOptions, verbosity: int
+    ) -> None:
+        rev_display = rev_options.to_display()
+        logger.info(
+            "Checking out %s%s to %s",
+            url,
+            rev_display,
+            display_path(dest),
+        )
+        if verbosity <= 0:
+            flags = ["--quiet"]
+        elif verbosity == 1:
+            flags = []
+        else:
+            flags = [f"-{'v'*verbosity}"]
+        cmd_args = make_command(
+            "checkout", "--lightweight", *flags, rev_options.to_args(), url, dest
+        )
+        self.run_command(cmd_args)
+
+    def switch(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        self.run_command(make_command("switch", url), cwd=dest)
+
+    def update(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        output = self.run_command(
+            make_command("info"), show_stdout=False, stdout_only=True, cwd=dest
+        )
+        if output.startswith("Standalone "):
+            # Older versions of pip used to create standalone branches.
+            # Convert the standalone branch to a checkout by calling "bzr bind".
+            cmd_args = make_command("bind", "-q", url)
+            self.run_command(cmd_args, cwd=dest)
+
+        cmd_args = make_command("update", "-q", rev_options.to_args())
+        self.run_command(cmd_args, cwd=dest)
+
+    @classmethod
+    def get_url_rev_and_auth(cls, url: str) -> Tuple[str, Optional[str], AuthInfo]:
+        # hotfix the URL scheme after removing bzr+ from bzr+ssh:// re-add it
+        url, rev, user_pass = super().get_url_rev_and_auth(url)
+        if url.startswith("ssh://"):
+            url = "bzr+" + url
+        return url, rev, user_pass
+
+    @classmethod
+    def get_remote_url(cls, location: str) -> str:
+        urls = cls.run_command(
+            ["info"], show_stdout=False, stdout_only=True, cwd=location
+        )
+        for line in urls.splitlines():
+            line = line.strip()
+            for x in ("checkout of branch: ", "parent branch: "):
+                if line.startswith(x):
+                    repo = line.split(x)[1]
+                    if cls._is_local_repository(repo):
+                        return path_to_url(repo)
+                    return repo
+        raise RemoteNotFoundError
+
+    @classmethod
+    def get_revision(cls, location: str) -> str:
+        revision = cls.run_command(
+            ["revno"],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        )
+        return revision.splitlines()[-1]
+
+    @classmethod
+    def is_commit_id_equal(cls, dest: str, name: Optional[str]) -> bool:
+        """Always assume the versions don't match"""
+        return False
+
+
+vcs.register(Bazaar)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/git.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/git.py
new file mode 100644
index 00000000..0425debb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/git.py
@@ -0,0 +1,527 @@
+import logging
+import os.path
+import pathlib
+import re
+import urllib.parse
+import urllib.request
+from dataclasses import replace
+from typing import List, Optional, Tuple
+
+from pip._internal.exceptions import BadCommand, InstallationError
+from pip._internal.utils.misc import HiddenText, display_path, hide_url
+from pip._internal.utils.subprocess import make_command
+from pip._internal.vcs.versioncontrol import (
+    AuthInfo,
+    RemoteNotFoundError,
+    RemoteNotValidError,
+    RevOptions,
+    VersionControl,
+    find_path_to_project_root_from_repo_root,
+    vcs,
+)
+
+urlsplit = urllib.parse.urlsplit
+urlunsplit = urllib.parse.urlunsplit
+
+
+logger = logging.getLogger(__name__)
+
+
+GIT_VERSION_REGEX = re.compile(
+    r"^git version "  # Prefix.
+    r"(\d+)"  # Major.
+    r"\.(\d+)"  # Dot, minor.
+    r"(?:\.(\d+))?"  # Optional dot, patch.
+    r".*$"  # Suffix, including any pre- and post-release segments we don't care about.
+)
+
+HASH_REGEX = re.compile("^[a-fA-F0-9]{40}$")
+
+# SCP (Secure copy protocol) shorthand. e.g. 'git@example.com:foo/bar.git'
+SCP_REGEX = re.compile(
+    r"""^
+    # Optional user, e.g. 'git@'
+    (\w+@)?
+    # Server, e.g. 'github.com'.
+    ([^/:]+):
+    # The server-side path. e.g. 'user/project.git'. Must start with an
+    # alphanumeric character so as not to be confusable with a Windows paths
+    # like 'C:/foo/bar' or 'C:\foo\bar'.
+    (\w[^:]*)
+    $""",
+    re.VERBOSE,
+)
+
+
+def looks_like_hash(sha: str) -> bool:
+    return bool(HASH_REGEX.match(sha))
+
+
+class Git(VersionControl):
+    name = "git"
+    dirname = ".git"
+    repo_name = "clone"
+    schemes = (
+        "git+http",
+        "git+https",
+        "git+ssh",
+        "git+git",
+        "git+file",
+    )
+    # Prevent the user's environment variables from interfering with pip:
+    # https://github.com/pypa/pip/issues/1130
+    unset_environ = ("GIT_DIR", "GIT_WORK_TREE")
+    default_arg_rev = "HEAD"
+
+    @staticmethod
+    def get_base_rev_args(rev: str) -> List[str]:
+        return [rev]
+
+    def is_immutable_rev_checkout(self, url: str, dest: str) -> bool:
+        _, rev_options = self.get_url_rev_options(hide_url(url))
+        if not rev_options.rev:
+            return False
+        if not self.is_commit_id_equal(dest, rev_options.rev):
+            # the current commit is different from rev,
+            # which means rev was something else than a commit hash
+            return False
+        # return False in the rare case rev is both a commit hash
+        # and a tag or a branch; we don't want to cache in that case
+        # because that branch/tag could point to something else in the future
+        is_tag_or_branch = bool(self.get_revision_sha(dest, rev_options.rev)[0])
+        return not is_tag_or_branch
+
+    def get_git_version(self) -> Tuple[int, ...]:
+        version = self.run_command(
+            ["version"],
+            command_desc="git version",
+            show_stdout=False,
+            stdout_only=True,
+        )
+        match = GIT_VERSION_REGEX.match(version)
+        if not match:
+            logger.warning("Can't parse git version: %s", version)
+            return ()
+        return (int(match.group(1)), int(match.group(2)))
+
+    @classmethod
+    def get_current_branch(cls, location: str) -> Optional[str]:
+        """
+        Return the current branch, or None if HEAD isn't at a branch
+        (e.g. detached HEAD).
+        """
+        # git-symbolic-ref exits with empty stdout if "HEAD" is a detached
+        # HEAD rather than a symbolic ref.  In addition, the -q causes the
+        # command to exit with status code 1 instead of 128 in this case
+        # and to suppress the message to stderr.
+        args = ["symbolic-ref", "-q", "HEAD"]
+        output = cls.run_command(
+            args,
+            extra_ok_returncodes=(1,),
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        )
+        ref = output.strip()
+
+        if ref.startswith("refs/heads/"):
+            return ref[len("refs/heads/") :]
+
+        return None
+
+    @classmethod
+    def get_revision_sha(cls, dest: str, rev: str) -> Tuple[Optional[str], bool]:
+        """
+        Return (sha_or_none, is_branch), where sha_or_none is a commit hash
+        if the revision names a remote branch or tag, otherwise None.
+
+        Args:
+          dest: the repository directory.
+          rev: the revision name.
+        """
+        # Pass rev to pre-filter the list.
+        output = cls.run_command(
+            ["show-ref", rev],
+            cwd=dest,
+            show_stdout=False,
+            stdout_only=True,
+            on_returncode="ignore",
+        )
+        refs = {}
+        # NOTE: We do not use splitlines here since that would split on other
+        #       unicode separators, which can be maliciously used to install a
+        #       different revision.
+        for line in output.strip().split("\n"):
+            line = line.rstrip("\r")
+            if not line:
+                continue
+            try:
+                ref_sha, ref_name = line.split(" ", maxsplit=2)
+            except ValueError:
+                # Include the offending line to simplify troubleshooting if
+                # this error ever occurs.
+                raise ValueError(f"unexpected show-ref line: {line!r}")
+
+            refs[ref_name] = ref_sha
+
+        branch_ref = f"refs/remotes/origin/{rev}"
+        tag_ref = f"refs/tags/{rev}"
+
+        sha = refs.get(branch_ref)
+        if sha is not None:
+            return (sha, True)
+
+        sha = refs.get(tag_ref)
+
+        return (sha, False)
+
+    @classmethod
+    def _should_fetch(cls, dest: str, rev: str) -> bool:
+        """
+        Return true if rev is a ref or is a commit that we don't have locally.
+
+        Branches and tags are not considered in this method because they are
+        assumed to be always available locally (which is a normal outcome of
+        ``git clone`` and ``git fetch --tags``).
+        """
+        if rev.startswith("refs/"):
+            # Always fetch remote refs.
+            return True
+
+        if not looks_like_hash(rev):
+            # Git fetch would fail with abbreviated commits.
+            return False
+
+        if cls.has_commit(dest, rev):
+            # Don't fetch if we have the commit locally.
+            return False
+
+        return True
+
+    @classmethod
+    def resolve_revision(
+        cls, dest: str, url: HiddenText, rev_options: RevOptions
+    ) -> RevOptions:
+        """
+        Resolve a revision to a new RevOptions object with the SHA1 of the
+        branch, tag, or ref if found.
+
+        Args:
+          rev_options: a RevOptions object.
+        """
+        rev = rev_options.arg_rev
+        # The arg_rev property's implementation for Git ensures that the
+        # rev return value is always non-None.
+        assert rev is not None
+
+        sha, is_branch = cls.get_revision_sha(dest, rev)
+
+        if sha is not None:
+            rev_options = rev_options.make_new(sha)
+            rev_options = replace(rev_options, branch_name=(rev if is_branch else None))
+
+            return rev_options
+
+        # Do not show a warning for the common case of something that has
+        # the form of a Git commit hash.
+        if not looks_like_hash(rev):
+            logger.warning(
+                "Did not find branch or tag '%s', assuming revision or ref.",
+                rev,
+            )
+
+        if not cls._should_fetch(dest, rev):
+            return rev_options
+
+        # fetch the requested revision
+        cls.run_command(
+            make_command("fetch", "-q", url, rev_options.to_args()),
+            cwd=dest,
+        )
+        # Change the revision to the SHA of the ref we fetched
+        sha = cls.get_revision(dest, rev="FETCH_HEAD")
+        rev_options = rev_options.make_new(sha)
+
+        return rev_options
+
+    @classmethod
+    def is_commit_id_equal(cls, dest: str, name: Optional[str]) -> bool:
+        """
+        Return whether the current commit hash equals the given name.
+
+        Args:
+          dest: the repository directory.
+          name: a string name.
+        """
+        if not name:
+            # Then avoid an unnecessary subprocess call.
+            return False
+
+        return cls.get_revision(dest) == name
+
+    def fetch_new(
+        self, dest: str, url: HiddenText, rev_options: RevOptions, verbosity: int
+    ) -> None:
+        rev_display = rev_options.to_display()
+        logger.info("Cloning %s%s to %s", url, rev_display, display_path(dest))
+        if verbosity <= 0:
+            flags: Tuple[str, ...] = ("--quiet",)
+        elif verbosity == 1:
+            flags = ()
+        else:
+            flags = ("--verbose", "--progress")
+        if self.get_git_version() >= (2, 17):
+            # Git added support for partial clone in 2.17
+            # https://git-scm.com/docs/partial-clone
+            # Speeds up cloning by functioning without a complete copy of repository
+            self.run_command(
+                make_command(
+                    "clone",
+                    "--filter=blob:none",
+                    *flags,
+                    url,
+                    dest,
+                )
+            )
+        else:
+            self.run_command(make_command("clone", *flags, url, dest))
+
+        if rev_options.rev:
+            # Then a specific revision was requested.
+            rev_options = self.resolve_revision(dest, url, rev_options)
+            branch_name = getattr(rev_options, "branch_name", None)
+            logger.debug("Rev options %s, branch_name %s", rev_options, branch_name)
+            if branch_name is None:
+                # Only do a checkout if the current commit id doesn't match
+                # the requested revision.
+                if not self.is_commit_id_equal(dest, rev_options.rev):
+                    cmd_args = make_command(
+                        "checkout",
+                        "-q",
+                        rev_options.to_args(),
+                    )
+                    self.run_command(cmd_args, cwd=dest)
+            elif self.get_current_branch(dest) != branch_name:
+                # Then a specific branch was requested, and that branch
+                # is not yet checked out.
+                track_branch = f"origin/{branch_name}"
+                cmd_args = [
+                    "checkout",
+                    "-b",
+                    branch_name,
+                    "--track",
+                    track_branch,
+                ]
+                self.run_command(cmd_args, cwd=dest)
+        else:
+            sha = self.get_revision(dest)
+            rev_options = rev_options.make_new(sha)
+
+        logger.info("Resolved %s to commit %s", url, rev_options.rev)
+
+        #: repo may contain submodules
+        self.update_submodules(dest)
+
+    def switch(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        self.run_command(
+            make_command("config", "remote.origin.url", url),
+            cwd=dest,
+        )
+        cmd_args = make_command("checkout", "-q", rev_options.to_args())
+        self.run_command(cmd_args, cwd=dest)
+
+        self.update_submodules(dest)
+
+    def update(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        # First fetch changes from the default remote
+        if self.get_git_version() >= (1, 9):
+            # fetch tags in addition to everything else
+            self.run_command(["fetch", "-q", "--tags"], cwd=dest)
+        else:
+            self.run_command(["fetch", "-q"], cwd=dest)
+        # Then reset to wanted revision (maybe even origin/master)
+        rev_options = self.resolve_revision(dest, url, rev_options)
+        cmd_args = make_command("reset", "--hard", "-q", rev_options.to_args())
+        self.run_command(cmd_args, cwd=dest)
+        #: update submodules
+        self.update_submodules(dest)
+
+    @classmethod
+    def get_remote_url(cls, location: str) -> str:
+        """
+        Return URL of the first remote encountered.
+
+        Raises RemoteNotFoundError if the repository does not have a remote
+        url configured.
+        """
+        # We need to pass 1 for extra_ok_returncodes since the command
+        # exits with return code 1 if there are no matching lines.
+        stdout = cls.run_command(
+            ["config", "--get-regexp", r"remote\..*\.url"],
+            extra_ok_returncodes=(1,),
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        )
+        remotes = stdout.splitlines()
+        try:
+            found_remote = remotes[0]
+        except IndexError:
+            raise RemoteNotFoundError
+
+        for remote in remotes:
+            if remote.startswith("remote.origin.url "):
+                found_remote = remote
+                break
+        url = found_remote.split(" ")[1]
+        return cls._git_remote_to_pip_url(url.strip())
+
+    @staticmethod
+    def _git_remote_to_pip_url(url: str) -> str:
+        """
+        Convert a remote url from what git uses to what pip accepts.
+
+        There are 3 legal forms **url** may take:
+
+            1. A fully qualified url: ssh://git@example.com/foo/bar.git
+            2. A local project.git folder: /path/to/bare/repository.git
+            3. SCP shorthand for form 1: git@example.com:foo/bar.git
+
+        Form 1 is output as-is. Form 2 must be converted to URI and form 3 must
+        be converted to form 1.
+
+        See the corresponding test test_git_remote_url_to_pip() for examples of
+        sample inputs/outputs.
+        """
+        if re.match(r"\w+://", url):
+            # This is already valid. Pass it though as-is.
+            return url
+        if os.path.exists(url):
+            # A local bare remote (git clone --mirror).
+            # Needs a file:// prefix.
+            return pathlib.PurePath(url).as_uri()
+        scp_match = SCP_REGEX.match(url)
+        if scp_match:
+            # Add an ssh:// prefix and replace the ':' with a '/'.
+            return scp_match.expand(r"ssh://\1\2/\3")
+        # Otherwise, bail out.
+        raise RemoteNotValidError(url)
+
+    @classmethod
+    def has_commit(cls, location: str, rev: str) -> bool:
+        """
+        Check if rev is a commit that is available in the local repository.
+        """
+        try:
+            cls.run_command(
+                ["rev-parse", "-q", "--verify", "sha^" + rev],
+                cwd=location,
+                log_failed_cmd=False,
+            )
+        except InstallationError:
+            return False
+        else:
+            return True
+
+    @classmethod
+    def get_revision(cls, location: str, rev: Optional[str] = None) -> str:
+        if rev is None:
+            rev = "HEAD"
+        current_rev = cls.run_command(
+            ["rev-parse", rev],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        )
+        return current_rev.strip()
+
+    @classmethod
+    def get_subdirectory(cls, location: str) -> Optional[str]:
+        """
+        Return the path to Python project root, relative to the repo root.
+        Return None if the project root is in the repo root.
+        """
+        # find the repo root
+        git_dir = cls.run_command(
+            ["rev-parse", "--git-dir"],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        ).strip()
+        if not os.path.isabs(git_dir):
+            git_dir = os.path.join(location, git_dir)
+        repo_root = os.path.abspath(os.path.join(git_dir, ".."))
+        return find_path_to_project_root_from_repo_root(location, repo_root)
+
+    @classmethod
+    def get_url_rev_and_auth(cls, url: str) -> Tuple[str, Optional[str], AuthInfo]:
+        """
+        Prefixes stub URLs like 'user@hostname:user/repo.git' with 'ssh://'.
+        That's required because although they use SSH they sometimes don't
+        work with a ssh:// scheme (e.g. GitHub). But we need a scheme for
+        parsing. Hence we remove it again afterwards and return it as a stub.
+        """
+        # Works around an apparent Git bug
+        # (see https://article.gmane.org/gmane.comp.version-control.git/146500)
+        scheme, netloc, path, query, fragment = urlsplit(url)
+        if scheme.endswith("file"):
+            initial_slashes = path[: -len(path.lstrip("/"))]
+            newpath = initial_slashes + urllib.request.url2pathname(path).replace(
+                "\\", "/"
+            ).lstrip("/")
+            after_plus = scheme.find("+") + 1
+            url = scheme[:after_plus] + urlunsplit(
+                (scheme[after_plus:], netloc, newpath, query, fragment),
+            )
+
+        if "://" not in url:
+            assert "file:" not in url
+            url = url.replace("git+", "git+ssh://")
+            url, rev, user_pass = super().get_url_rev_and_auth(url)
+            url = url.replace("ssh://", "")
+        else:
+            url, rev, user_pass = super().get_url_rev_and_auth(url)
+
+        return url, rev, user_pass
+
+    @classmethod
+    def update_submodules(cls, location: str) -> None:
+        if not os.path.exists(os.path.join(location, ".gitmodules")):
+            return
+        cls.run_command(
+            ["submodule", "update", "--init", "--recursive", "-q"],
+            cwd=location,
+        )
+
+    @classmethod
+    def get_repository_root(cls, location: str) -> Optional[str]:
+        loc = super().get_repository_root(location)
+        if loc:
+            return loc
+        try:
+            r = cls.run_command(
+                ["rev-parse", "--show-toplevel"],
+                cwd=location,
+                show_stdout=False,
+                stdout_only=True,
+                on_returncode="raise",
+                log_failed_cmd=False,
+            )
+        except BadCommand:
+            logger.debug(
+                "could not determine if %s is under git control "
+                "because git is not available",
+                location,
+            )
+            return None
+        except InstallationError:
+            return None
+        return os.path.normpath(r.rstrip("\r\n"))
+
+    @staticmethod
+    def should_add_vcs_url_prefix(repo_url: str) -> bool:
+        """In either https or ssh form, requirements must be prefixed with git+."""
+        return True
+
+
+vcs.register(Git)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/mercurial.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/mercurial.py
new file mode 100644
index 00000000..c183d41d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/mercurial.py
@@ -0,0 +1,163 @@
+import configparser
+import logging
+import os
+from typing import List, Optional, Tuple
+
+from pip._internal.exceptions import BadCommand, InstallationError
+from pip._internal.utils.misc import HiddenText, display_path
+from pip._internal.utils.subprocess import make_command
+from pip._internal.utils.urls import path_to_url
+from pip._internal.vcs.versioncontrol import (
+    RevOptions,
+    VersionControl,
+    find_path_to_project_root_from_repo_root,
+    vcs,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class Mercurial(VersionControl):
+    name = "hg"
+    dirname = ".hg"
+    repo_name = "clone"
+    schemes = (
+        "hg+file",
+        "hg+http",
+        "hg+https",
+        "hg+ssh",
+        "hg+static-http",
+    )
+
+    @staticmethod
+    def get_base_rev_args(rev: str) -> List[str]:
+        return [f"--rev={rev}"]
+
+    def fetch_new(
+        self, dest: str, url: HiddenText, rev_options: RevOptions, verbosity: int
+    ) -> None:
+        rev_display = rev_options.to_display()
+        logger.info(
+            "Cloning hg %s%s to %s",
+            url,
+            rev_display,
+            display_path(dest),
+        )
+        if verbosity <= 0:
+            flags: Tuple[str, ...] = ("--quiet",)
+        elif verbosity == 1:
+            flags = ()
+        elif verbosity == 2:
+            flags = ("--verbose",)
+        else:
+            flags = ("--verbose", "--debug")
+        self.run_command(make_command("clone", "--noupdate", *flags, url, dest))
+        self.run_command(
+            make_command("update", *flags, rev_options.to_args()),
+            cwd=dest,
+        )
+
+    def switch(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        repo_config = os.path.join(dest, self.dirname, "hgrc")
+        config = configparser.RawConfigParser()
+        try:
+            config.read(repo_config)
+            config.set("paths", "default", url.secret)
+            with open(repo_config, "w") as config_file:
+                config.write(config_file)
+        except (OSError, configparser.NoSectionError) as exc:
+            logger.warning("Could not switch Mercurial repository to %s: %s", url, exc)
+        else:
+            cmd_args = make_command("update", "-q", rev_options.to_args())
+            self.run_command(cmd_args, cwd=dest)
+
+    def update(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        self.run_command(["pull", "-q"], cwd=dest)
+        cmd_args = make_command("update", "-q", rev_options.to_args())
+        self.run_command(cmd_args, cwd=dest)
+
+    @classmethod
+    def get_remote_url(cls, location: str) -> str:
+        url = cls.run_command(
+            ["showconfig", "paths.default"],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        ).strip()
+        if cls._is_local_repository(url):
+            url = path_to_url(url)
+        return url.strip()
+
+    @classmethod
+    def get_revision(cls, location: str) -> str:
+        """
+        Return the repository-local changeset revision number, as an integer.
+        """
+        current_revision = cls.run_command(
+            ["parents", "--template={rev}"],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        ).strip()
+        return current_revision
+
+    @classmethod
+    def get_requirement_revision(cls, location: str) -> str:
+        """
+        Return the changeset identification hash, as a 40-character
+        hexadecimal string
+        """
+        current_rev_hash = cls.run_command(
+            ["parents", "--template={node}"],
+            show_stdout=False,
+            stdout_only=True,
+            cwd=location,
+        ).strip()
+        return current_rev_hash
+
+    @classmethod
+    def is_commit_id_equal(cls, dest: str, name: Optional[str]) -> bool:
+        """Always assume the versions don't match"""
+        return False
+
+    @classmethod
+    def get_subdirectory(cls, location: str) -> Optional[str]:
+        """
+        Return the path to Python project root, relative to the repo root.
+        Return None if the project root is in the repo root.
+        """
+        # find the repo root
+        repo_root = cls.run_command(
+            ["root"], show_stdout=False, stdout_only=True, cwd=location
+        ).strip()
+        if not os.path.isabs(repo_root):
+            repo_root = os.path.abspath(os.path.join(location, repo_root))
+        return find_path_to_project_root_from_repo_root(location, repo_root)
+
+    @classmethod
+    def get_repository_root(cls, location: str) -> Optional[str]:
+        loc = super().get_repository_root(location)
+        if loc:
+            return loc
+        try:
+            r = cls.run_command(
+                ["root"],
+                cwd=location,
+                show_stdout=False,
+                stdout_only=True,
+                on_returncode="raise",
+                log_failed_cmd=False,
+            )
+        except BadCommand:
+            logger.debug(
+                "could not determine if %s is under hg control "
+                "because hg is not available",
+                location,
+            )
+            return None
+        except InstallationError:
+            return None
+        return os.path.normpath(r.rstrip("\r\n"))
+
+
+vcs.register(Mercurial)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/subversion.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/subversion.py
new file mode 100644
index 00000000..f359266d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/subversion.py
@@ -0,0 +1,324 @@
+import logging
+import os
+import re
+from typing import List, Optional, Tuple
+
+from pip._internal.utils.misc import (
+    HiddenText,
+    display_path,
+    is_console_interactive,
+    is_installable_dir,
+    split_auth_from_netloc,
+)
+from pip._internal.utils.subprocess import CommandArgs, make_command
+from pip._internal.vcs.versioncontrol import (
+    AuthInfo,
+    RemoteNotFoundError,
+    RevOptions,
+    VersionControl,
+    vcs,
+)
+
+logger = logging.getLogger(__name__)
+
+_svn_xml_url_re = re.compile('url="([^"]+)"')
+_svn_rev_re = re.compile(r'committed-rev="(\d+)"')
+_svn_info_xml_rev_re = re.compile(r'\s*revision="(\d+)"')
+_svn_info_xml_url_re = re.compile(r"<url>(.*)</url>")
+
+
+class Subversion(VersionControl):
+    name = "svn"
+    dirname = ".svn"
+    repo_name = "checkout"
+    schemes = ("svn+ssh", "svn+http", "svn+https", "svn+svn", "svn+file")
+
+    @classmethod
+    def should_add_vcs_url_prefix(cls, remote_url: str) -> bool:
+        return True
+
+    @staticmethod
+    def get_base_rev_args(rev: str) -> List[str]:
+        return ["-r", rev]
+
+    @classmethod
+    def get_revision(cls, location: str) -> str:
+        """
+        Return the maximum revision for all files under a given location
+        """
+        # Note: taken from setuptools.command.egg_info
+        revision = 0
+
+        for base, dirs, _ in os.walk(location):
+            if cls.dirname not in dirs:
+                dirs[:] = []
+                continue  # no sense walking uncontrolled subdirs
+            dirs.remove(cls.dirname)
+            entries_fn = os.path.join(base, cls.dirname, "entries")
+            if not os.path.exists(entries_fn):
+                # FIXME: should we warn?
+                continue
+
+            dirurl, localrev = cls._get_svn_url_rev(base)
+
+            if base == location:
+                assert dirurl is not None
+                base = dirurl + "/"  # save the root url
+            elif not dirurl or not dirurl.startswith(base):
+                dirs[:] = []
+                continue  # not part of the same svn tree, skip it
+            revision = max(revision, localrev)
+        return str(revision)
+
+    @classmethod
+    def get_netloc_and_auth(
+        cls, netloc: str, scheme: str
+    ) -> Tuple[str, Tuple[Optional[str], Optional[str]]]:
+        """
+        This override allows the auth information to be passed to svn via the
+        --username and --password options instead of via the URL.
+        """
+        if scheme == "ssh":
+            # The --username and --password options can't be used for
+            # svn+ssh URLs, so keep the auth information in the URL.
+            return super().get_netloc_and_auth(netloc, scheme)
+
+        return split_auth_from_netloc(netloc)
+
+    @classmethod
+    def get_url_rev_and_auth(cls, url: str) -> Tuple[str, Optional[str], AuthInfo]:
+        # hotfix the URL scheme after removing svn+ from svn+ssh:// re-add it
+        url, rev, user_pass = super().get_url_rev_and_auth(url)
+        if url.startswith("ssh://"):
+            url = "svn+" + url
+        return url, rev, user_pass
+
+    @staticmethod
+    def make_rev_args(
+        username: Optional[str], password: Optional[HiddenText]
+    ) -> CommandArgs:
+        extra_args: CommandArgs = []
+        if username:
+            extra_args += ["--username", username]
+        if password:
+            extra_args += ["--password", password]
+
+        return extra_args
+
+    @classmethod
+    def get_remote_url(cls, location: str) -> str:
+        # In cases where the source is in a subdirectory, we have to look up in
+        # the location until we find a valid project root.
+        orig_location = location
+        while not is_installable_dir(location):
+            last_location = location
+            location = os.path.dirname(location)
+            if location == last_location:
+                # We've traversed up to the root of the filesystem without
+                # finding a Python project.
+                logger.warning(
+                    "Could not find Python project for directory %s (tried all "
+                    "parent directories)",
+                    orig_location,
+                )
+                raise RemoteNotFoundError
+
+        url, _rev = cls._get_svn_url_rev(location)
+        if url is None:
+            raise RemoteNotFoundError
+
+        return url
+
+    @classmethod
+    def _get_svn_url_rev(cls, location: str) -> Tuple[Optional[str], int]:
+        from pip._internal.exceptions import InstallationError
+
+        entries_path = os.path.join(location, cls.dirname, "entries")
+        if os.path.exists(entries_path):
+            with open(entries_path) as f:
+                data = f.read()
+        else:  # subversion >= 1.7 does not have the 'entries' file
+            data = ""
+
+        url = None
+        if data.startswith("8") or data.startswith("9") or data.startswith("10"):
+            entries = list(map(str.splitlines, data.split("\n\x0c\n")))
+            del entries[0][0]  # get rid of the '8'
+            url = entries[0][3]
+            revs = [int(d[9]) for d in entries if len(d) > 9 and d[9]] + [0]
+        elif data.startswith("<?xml"):
+            match = _svn_xml_url_re.search(data)
+            if not match:
+                raise ValueError(f"Badly formatted data: {data!r}")
+            url = match.group(1)  # get repository URL
+            revs = [int(m.group(1)) for m in _svn_rev_re.finditer(data)] + [0]
+        else:
+            try:
+                # subversion >= 1.7
+                # Note that using get_remote_call_options is not necessary here
+                # because `svn info` is being run against a local directory.
+                # We don't need to worry about making sure interactive mode
+                # is being used to prompt for passwords, because passwords
+                # are only potentially needed for remote server requests.
+                xml = cls.run_command(
+                    ["info", "--xml", location],
+                    show_stdout=False,
+                    stdout_only=True,
+                )
+                match = _svn_info_xml_url_re.search(xml)
+                assert match is not None
+                url = match.group(1)
+                revs = [int(m.group(1)) for m in _svn_info_xml_rev_re.finditer(xml)]
+            except InstallationError:
+                url, revs = None, []
+
+        if revs:
+            rev = max(revs)
+        else:
+            rev = 0
+
+        return url, rev
+
+    @classmethod
+    def is_commit_id_equal(cls, dest: str, name: Optional[str]) -> bool:
+        """Always assume the versions don't match"""
+        return False
+
+    def __init__(self, use_interactive: Optional[bool] = None) -> None:
+        if use_interactive is None:
+            use_interactive = is_console_interactive()
+        self.use_interactive = use_interactive
+
+        # This member is used to cache the fetched version of the current
+        # ``svn`` client.
+        # Special value definitions:
+        #   None: Not evaluated yet.
+        #   Empty tuple: Could not parse version.
+        self._vcs_version: Optional[Tuple[int, ...]] = None
+
+        super().__init__()
+
+    def call_vcs_version(self) -> Tuple[int, ...]:
+        """Query the version of the currently installed Subversion client.
+
+        :return: A tuple containing the parts of the version information or
+            ``()`` if the version returned from ``svn`` could not be parsed.
+        :raises: BadCommand: If ``svn`` is not installed.
+        """
+        # Example versions:
+        #   svn, version 1.10.3 (r1842928)
+        #      compiled Feb 25 2019, 14:20:39 on x86_64-apple-darwin17.0.0
+        #   svn, version 1.7.14 (r1542130)
+        #      compiled Mar 28 2018, 08:49:13 on x86_64-pc-linux-gnu
+        #   svn, version 1.12.0-SlikSvn (SlikSvn/1.12.0)
+        #      compiled May 28 2019, 13:44:56 on x86_64-microsoft-windows6.2
+        version_prefix = "svn, version "
+        version = self.run_command(["--version"], show_stdout=False, stdout_only=True)
+        if not version.startswith(version_prefix):
+            return ()
+
+        version = version[len(version_prefix) :].split()[0]
+        version_list = version.partition("-")[0].split(".")
+        try:
+            parsed_version = tuple(map(int, version_list))
+        except ValueError:
+            return ()
+
+        return parsed_version
+
+    def get_vcs_version(self) -> Tuple[int, ...]:
+        """Return the version of the currently installed Subversion client.
+
+        If the version of the Subversion client has already been queried,
+        a cached value will be used.
+
+        :return: A tuple containing the parts of the version information or
+            ``()`` if the version returned from ``svn`` could not be parsed.
+        :raises: BadCommand: If ``svn`` is not installed.
+        """
+        if self._vcs_version is not None:
+            # Use cached version, if available.
+            # If parsing the version failed previously (empty tuple),
+            # do not attempt to parse it again.
+            return self._vcs_version
+
+        vcs_version = self.call_vcs_version()
+        self._vcs_version = vcs_version
+        return vcs_version
+
+    def get_remote_call_options(self) -> CommandArgs:
+        """Return options to be used on calls to Subversion that contact the server.
+
+        These options are applicable for the following ``svn`` subcommands used
+        in this class.
+
+            - checkout
+            - switch
+            - update
+
+        :return: A list of command line arguments to pass to ``svn``.
+        """
+        if not self.use_interactive:
+            # --non-interactive switch is available since Subversion 0.14.4.
+            # Subversion < 1.8 runs in interactive mode by default.
+            return ["--non-interactive"]
+
+        svn_version = self.get_vcs_version()
+        # By default, Subversion >= 1.8 runs in non-interactive mode if
+        # stdin is not a TTY. Since that is how pip invokes SVN, in
+        # call_subprocess(), pip must pass --force-interactive to ensure
+        # the user can be prompted for a password, if required.
+        #   SVN added the --force-interactive option in SVN 1.8. Since
+        # e.g. RHEL/CentOS 7, which is supported until 2024, ships with
+        # SVN 1.7, pip should continue to support SVN 1.7. Therefore, pip
+        # can't safely add the option if the SVN version is < 1.8 (or unknown).
+        if svn_version >= (1, 8):
+            return ["--force-interactive"]
+
+        return []
+
+    def fetch_new(
+        self, dest: str, url: HiddenText, rev_options: RevOptions, verbosity: int
+    ) -> None:
+        rev_display = rev_options.to_display()
+        logger.info(
+            "Checking out %s%s to %s",
+            url,
+            rev_display,
+            display_path(dest),
+        )
+        if verbosity <= 0:
+            flags = ["--quiet"]
+        else:
+            flags = []
+        cmd_args = make_command(
+            "checkout",
+            *flags,
+            self.get_remote_call_options(),
+            rev_options.to_args(),
+            url,
+            dest,
+        )
+        self.run_command(cmd_args)
+
+    def switch(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        cmd_args = make_command(
+            "switch",
+            self.get_remote_call_options(),
+            rev_options.to_args(),
+            url,
+            dest,
+        )
+        self.run_command(cmd_args)
+
+    def update(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        cmd_args = make_command(
+            "update",
+            self.get_remote_call_options(),
+            rev_options.to_args(),
+            dest,
+        )
+        self.run_command(cmd_args)
+
+
+vcs.register(Subversion)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/versioncontrol.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/versioncontrol.py
new file mode 100644
index 00000000..a4133165
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/vcs/versioncontrol.py
@@ -0,0 +1,688 @@
+"""Handles all VCS (version control) support"""
+
+import logging
+import os
+import shutil
+import sys
+import urllib.parse
+from dataclasses import dataclass, field
+from typing import (
+    Any,
+    Dict,
+    Iterable,
+    Iterator,
+    List,
+    Literal,
+    Mapping,
+    Optional,
+    Tuple,
+    Type,
+    Union,
+)
+
+from pip._internal.cli.spinners import SpinnerInterface
+from pip._internal.exceptions import BadCommand, InstallationError
+from pip._internal.utils.misc import (
+    HiddenText,
+    ask_path_exists,
+    backup_dir,
+    display_path,
+    hide_url,
+    hide_value,
+    is_installable_dir,
+    rmtree,
+)
+from pip._internal.utils.subprocess import (
+    CommandArgs,
+    call_subprocess,
+    format_command_args,
+    make_command,
+)
+
+__all__ = ["vcs"]
+
+
+logger = logging.getLogger(__name__)
+
+AuthInfo = Tuple[Optional[str], Optional[str]]
+
+
+def is_url(name: str) -> bool:
+    """
+    Return true if the name looks like a URL.
+    """
+    scheme = urllib.parse.urlsplit(name).scheme
+    if not scheme:
+        return False
+    return scheme in ["http", "https", "file", "ftp"] + vcs.all_schemes
+
+
+def make_vcs_requirement_url(
+    repo_url: str, rev: str, project_name: str, subdir: Optional[str] = None
+) -> str:
+    """
+    Return the URL for a VCS requirement.
+
+    Args:
+      repo_url: the remote VCS url, with any needed VCS prefix (e.g. "git+").
+      project_name: the (unescaped) project name.
+    """
+    egg_project_name = project_name.replace("-", "_")
+    req = f"{repo_url}@{rev}#egg={egg_project_name}"
+    if subdir:
+        req += f"&subdirectory={subdir}"
+
+    return req
+
+
+def find_path_to_project_root_from_repo_root(
+    location: str, repo_root: str
+) -> Optional[str]:
+    """
+    Find the the Python project's root by searching up the filesystem from
+    `location`. Return the path to project root relative to `repo_root`.
+    Return None if the project root is `repo_root`, or cannot be found.
+    """
+    # find project root.
+    orig_location = location
+    while not is_installable_dir(location):
+        last_location = location
+        location = os.path.dirname(location)
+        if location == last_location:
+            # We've traversed up to the root of the filesystem without
+            # finding a Python project.
+            logger.warning(
+                "Could not find a Python project for directory %s (tried all "
+                "parent directories)",
+                orig_location,
+            )
+            return None
+
+    if os.path.samefile(repo_root, location):
+        return None
+
+    return os.path.relpath(location, repo_root)
+
+
+class RemoteNotFoundError(Exception):
+    pass
+
+
+class RemoteNotValidError(Exception):
+    def __init__(self, url: str):
+        super().__init__(url)
+        self.url = url
+
+
+@dataclass(frozen=True)
+class RevOptions:
+    """
+    Encapsulates a VCS-specific revision to install, along with any VCS
+    install options.
+
+    Args:
+        vc_class: a VersionControl subclass.
+        rev: the name of the revision to install.
+        extra_args: a list of extra options.
+    """
+
+    vc_class: Type["VersionControl"]
+    rev: Optional[str] = None
+    extra_args: CommandArgs = field(default_factory=list)
+    branch_name: Optional[str] = None
+
+    def __repr__(self) -> str:
+        return f"<RevOptions {self.vc_class.name}: rev={self.rev!r}>"
+
+    @property
+    def arg_rev(self) -> Optional[str]:
+        if self.rev is None:
+            return self.vc_class.default_arg_rev
+
+        return self.rev
+
+    def to_args(self) -> CommandArgs:
+        """
+        Return the VCS-specific command arguments.
+        """
+        args: CommandArgs = []
+        rev = self.arg_rev
+        if rev is not None:
+            args += self.vc_class.get_base_rev_args(rev)
+        args += self.extra_args
+
+        return args
+
+    def to_display(self) -> str:
+        if not self.rev:
+            return ""
+
+        return f" (to revision {self.rev})"
+
+    def make_new(self, rev: str) -> "RevOptions":
+        """
+        Make a copy of the current instance, but with a new rev.
+
+        Args:
+          rev: the name of the revision for the new object.
+        """
+        return self.vc_class.make_rev_options(rev, extra_args=self.extra_args)
+
+
+class VcsSupport:
+    _registry: Dict[str, "VersionControl"] = {}
+    schemes = ["ssh", "git", "hg", "bzr", "sftp", "svn"]
+
+    def __init__(self) -> None:
+        # Register more schemes with urlparse for various version control
+        # systems
+        urllib.parse.uses_netloc.extend(self.schemes)
+        super().__init__()
+
+    def __iter__(self) -> Iterator[str]:
+        return self._registry.__iter__()
+
+    @property
+    def backends(self) -> List["VersionControl"]:
+        return list(self._registry.values())
+
+    @property
+    def dirnames(self) -> List[str]:
+        return [backend.dirname for backend in self.backends]
+
+    @property
+    def all_schemes(self) -> List[str]:
+        schemes: List[str] = []
+        for backend in self.backends:
+            schemes.extend(backend.schemes)
+        return schemes
+
+    def register(self, cls: Type["VersionControl"]) -> None:
+        if not hasattr(cls, "name"):
+            logger.warning("Cannot register VCS %s", cls.__name__)
+            return
+        if cls.name not in self._registry:
+            self._registry[cls.name] = cls()
+            logger.debug("Registered VCS backend: %s", cls.name)
+
+    def unregister(self, name: str) -> None:
+        if name in self._registry:
+            del self._registry[name]
+
+    def get_backend_for_dir(self, location: str) -> Optional["VersionControl"]:
+        """
+        Return a VersionControl object if a repository of that type is found
+        at the given directory.
+        """
+        vcs_backends = {}
+        for vcs_backend in self._registry.values():
+            repo_path = vcs_backend.get_repository_root(location)
+            if not repo_path:
+                continue
+            logger.debug("Determine that %s uses VCS: %s", location, vcs_backend.name)
+            vcs_backends[repo_path] = vcs_backend
+
+        if not vcs_backends:
+            return None
+
+        # Choose the VCS in the inner-most directory. Since all repository
+        # roots found here would be either `location` or one of its
+        # parents, the longest path should have the most path components,
+        # i.e. the backend representing the inner-most repository.
+        inner_most_repo_path = max(vcs_backends, key=len)
+        return vcs_backends[inner_most_repo_path]
+
+    def get_backend_for_scheme(self, scheme: str) -> Optional["VersionControl"]:
+        """
+        Return a VersionControl object or None.
+        """
+        for vcs_backend in self._registry.values():
+            if scheme in vcs_backend.schemes:
+                return vcs_backend
+        return None
+
+    def get_backend(self, name: str) -> Optional["VersionControl"]:
+        """
+        Return a VersionControl object or None.
+        """
+        name = name.lower()
+        return self._registry.get(name)
+
+
+vcs = VcsSupport()
+
+
+class VersionControl:
+    name = ""
+    dirname = ""
+    repo_name = ""
+    # List of supported schemes for this Version Control
+    schemes: Tuple[str, ...] = ()
+    # Iterable of environment variable names to pass to call_subprocess().
+    unset_environ: Tuple[str, ...] = ()
+    default_arg_rev: Optional[str] = None
+
+    @classmethod
+    def should_add_vcs_url_prefix(cls, remote_url: str) -> bool:
+        """
+        Return whether the vcs prefix (e.g. "git+") should be added to a
+        repository's remote url when used in a requirement.
+        """
+        return not remote_url.lower().startswith(f"{cls.name}:")
+
+    @classmethod
+    def get_subdirectory(cls, location: str) -> Optional[str]:
+        """
+        Return the path to Python project root, relative to the repo root.
+        Return None if the project root is in the repo root.
+        """
+        return None
+
+    @classmethod
+    def get_requirement_revision(cls, repo_dir: str) -> str:
+        """
+        Return the revision string that should be used in a requirement.
+        """
+        return cls.get_revision(repo_dir)
+
+    @classmethod
+    def get_src_requirement(cls, repo_dir: str, project_name: str) -> str:
+        """
+        Return the requirement string to use to redownload the files
+        currently at the given repository directory.
+
+        Args:
+          project_name: the (unescaped) project name.
+
+        The return value has a form similar to the following:
+
+            {repository_url}@{revision}#egg={project_name}
+        """
+        repo_url = cls.get_remote_url(repo_dir)
+
+        if cls.should_add_vcs_url_prefix(repo_url):
+            repo_url = f"{cls.name}+{repo_url}"
+
+        revision = cls.get_requirement_revision(repo_dir)
+        subdir = cls.get_subdirectory(repo_dir)
+        req = make_vcs_requirement_url(repo_url, revision, project_name, subdir=subdir)
+
+        return req
+
+    @staticmethod
+    def get_base_rev_args(rev: str) -> List[str]:
+        """
+        Return the base revision arguments for a vcs command.
+
+        Args:
+          rev: the name of a revision to install.  Cannot be None.
+        """
+        raise NotImplementedError
+
+    def is_immutable_rev_checkout(self, url: str, dest: str) -> bool:
+        """
+        Return true if the commit hash checked out at dest matches
+        the revision in url.
+
+        Always return False, if the VCS does not support immutable commit
+        hashes.
+
+        This method does not check if there are local uncommitted changes
+        in dest after checkout, as pip currently has no use case for that.
+        """
+        return False
+
+    @classmethod
+    def make_rev_options(
+        cls, rev: Optional[str] = None, extra_args: Optional[CommandArgs] = None
+    ) -> RevOptions:
+        """
+        Return a RevOptions object.
+
+        Args:
+          rev: the name of a revision to install.
+          extra_args: a list of extra options.
+        """
+        return RevOptions(cls, rev, extra_args=extra_args or [])
+
+    @classmethod
+    def _is_local_repository(cls, repo: str) -> bool:
+        """
+        posix absolute paths start with os.path.sep,
+        win32 ones start with drive (like c:\\folder)
+        """
+        drive, tail = os.path.splitdrive(repo)
+        return repo.startswith(os.path.sep) or bool(drive)
+
+    @classmethod
+    def get_netloc_and_auth(
+        cls, netloc: str, scheme: str
+    ) -> Tuple[str, Tuple[Optional[str], Optional[str]]]:
+        """
+        Parse the repository URL's netloc, and return the new netloc to use
+        along with auth information.
+
+        Args:
+          netloc: the original repository URL netloc.
+          scheme: the repository URL's scheme without the vcs prefix.
+
+        This is mainly for the Subversion class to override, so that auth
+        information can be provided via the --username and --password options
+        instead of through the URL.  For other subclasses like Git without
+        such an option, auth information must stay in the URL.
+
+        Returns: (netloc, (username, password)).
+        """
+        return netloc, (None, None)
+
+    @classmethod
+    def get_url_rev_and_auth(cls, url: str) -> Tuple[str, Optional[str], AuthInfo]:
+        """
+        Parse the repository URL to use, and return the URL, revision,
+        and auth info to use.
+
+        Returns: (url, rev, (username, password)).
+        """
+        scheme, netloc, path, query, frag = urllib.parse.urlsplit(url)
+        if "+" not in scheme:
+            raise ValueError(
+                f"Sorry, {url!r} is a malformed VCS url. "
+                "The format is <vcs>+<protocol>://<url>, "
+                "e.g. svn+http://myrepo/svn/MyApp#egg=MyApp"
+            )
+        # Remove the vcs prefix.
+        scheme = scheme.split("+", 1)[1]
+        netloc, user_pass = cls.get_netloc_and_auth(netloc, scheme)
+        rev = None
+        if "@" in path:
+            path, rev = path.rsplit("@", 1)
+            if not rev:
+                raise InstallationError(
+                    f"The URL {url!r} has an empty revision (after @) "
+                    "which is not supported. Include a revision after @ "
+                    "or remove @ from the URL."
+                )
+        url = urllib.parse.urlunsplit((scheme, netloc, path, query, ""))
+        return url, rev, user_pass
+
+    @staticmethod
+    def make_rev_args(
+        username: Optional[str], password: Optional[HiddenText]
+    ) -> CommandArgs:
+        """
+        Return the RevOptions "extra arguments" to use in obtain().
+        """
+        return []
+
+    def get_url_rev_options(self, url: HiddenText) -> Tuple[HiddenText, RevOptions]:
+        """
+        Return the URL and RevOptions object to use in obtain(),
+        as a tuple (url, rev_options).
+        """
+        secret_url, rev, user_pass = self.get_url_rev_and_auth(url.secret)
+        username, secret_password = user_pass
+        password: Optional[HiddenText] = None
+        if secret_password is not None:
+            password = hide_value(secret_password)
+        extra_args = self.make_rev_args(username, password)
+        rev_options = self.make_rev_options(rev, extra_args=extra_args)
+
+        return hide_url(secret_url), rev_options
+
+    @staticmethod
+    def normalize_url(url: str) -> str:
+        """
+        Normalize a URL for comparison by unquoting it and removing any
+        trailing slash.
+        """
+        return urllib.parse.unquote(url).rstrip("/")
+
+    @classmethod
+    def compare_urls(cls, url1: str, url2: str) -> bool:
+        """
+        Compare two repo URLs for identity, ignoring incidental differences.
+        """
+        return cls.normalize_url(url1) == cls.normalize_url(url2)
+
+    def fetch_new(
+        self, dest: str, url: HiddenText, rev_options: RevOptions, verbosity: int
+    ) -> None:
+        """
+        Fetch a revision from a repository, in the case that this is the
+        first fetch from the repository.
+
+        Args:
+          dest: the directory to fetch the repository to.
+          rev_options: a RevOptions object.
+          verbosity: verbosity level.
+        """
+        raise NotImplementedError
+
+    def switch(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        """
+        Switch the repo at ``dest`` to point to ``URL``.
+
+        Args:
+          rev_options: a RevOptions object.
+        """
+        raise NotImplementedError
+
+    def update(self, dest: str, url: HiddenText, rev_options: RevOptions) -> None:
+        """
+        Update an already-existing repo to the given ``rev_options``.
+
+        Args:
+          rev_options: a RevOptions object.
+        """
+        raise NotImplementedError
+
+    @classmethod
+    def is_commit_id_equal(cls, dest: str, name: Optional[str]) -> bool:
+        """
+        Return whether the id of the current commit equals the given name.
+
+        Args:
+          dest: the repository directory.
+          name: a string name.
+        """
+        raise NotImplementedError
+
+    def obtain(self, dest: str, url: HiddenText, verbosity: int) -> None:
+        """
+        Install or update in editable mode the package represented by this
+        VersionControl object.
+
+        :param dest: the repository directory in which to install or update.
+        :param url: the repository URL starting with a vcs prefix.
+        :param verbosity: verbosity level.
+        """
+        url, rev_options = self.get_url_rev_options(url)
+
+        if not os.path.exists(dest):
+            self.fetch_new(dest, url, rev_options, verbosity=verbosity)
+            return
+
+        rev_display = rev_options.to_display()
+        if self.is_repository_directory(dest):
+            existing_url = self.get_remote_url(dest)
+            if self.compare_urls(existing_url, url.secret):
+                logger.debug(
+                    "%s in %s exists, and has correct URL (%s)",
+                    self.repo_name.title(),
+                    display_path(dest),
+                    url,
+                )
+                if not self.is_commit_id_equal(dest, rev_options.rev):
+                    logger.info(
+                        "Updating %s %s%s",
+                        display_path(dest),
+                        self.repo_name,
+                        rev_display,
+                    )
+                    self.update(dest, url, rev_options)
+                else:
+                    logger.info("Skipping because already up-to-date.")
+                return
+
+            logger.warning(
+                "%s %s in %s exists with URL %s",
+                self.name,
+                self.repo_name,
+                display_path(dest),
+                existing_url,
+            )
+            prompt = ("(s)witch, (i)gnore, (w)ipe, (b)ackup ", ("s", "i", "w", "b"))
+        else:
+            logger.warning(
+                "Directory %s already exists, and is not a %s %s.",
+                dest,
+                self.name,
+                self.repo_name,
+            )
+            # https://github.com/python/mypy/issues/1174
+            prompt = ("(i)gnore, (w)ipe, (b)ackup ", ("i", "w", "b"))  # type: ignore
+
+        logger.warning(
+            "The plan is to install the %s repository %s",
+            self.name,
+            url,
+        )
+        response = ask_path_exists(f"What to do?  {prompt[0]}", prompt[1])
+
+        if response == "a":
+            sys.exit(-1)
+
+        if response == "w":
+            logger.warning("Deleting %s", display_path(dest))
+            rmtree(dest)
+            self.fetch_new(dest, url, rev_options, verbosity=verbosity)
+            return
+
+        if response == "b":
+            dest_dir = backup_dir(dest)
+            logger.warning("Backing up %s to %s", display_path(dest), dest_dir)
+            shutil.move(dest, dest_dir)
+            self.fetch_new(dest, url, rev_options, verbosity=verbosity)
+            return
+
+        # Do nothing if the response is "i".
+        if response == "s":
+            logger.info(
+                "Switching %s %s to %s%s",
+                self.repo_name,
+                display_path(dest),
+                url,
+                rev_display,
+            )
+            self.switch(dest, url, rev_options)
+
+    def unpack(self, location: str, url: HiddenText, verbosity: int) -> None:
+        """
+        Clean up current location and download the url repository
+        (and vcs infos) into location
+
+        :param url: the repository URL starting with a vcs prefix.
+        :param verbosity: verbosity level.
+        """
+        if os.path.exists(location):
+            rmtree(location)
+        self.obtain(location, url=url, verbosity=verbosity)
+
+    @classmethod
+    def get_remote_url(cls, location: str) -> str:
+        """
+        Return the url used at location
+
+        Raises RemoteNotFoundError if the repository does not have a remote
+        url configured.
+        """
+        raise NotImplementedError
+
+    @classmethod
+    def get_revision(cls, location: str) -> str:
+        """
+        Return the current commit id of the files at the given location.
+        """
+        raise NotImplementedError
+
+    @classmethod
+    def run_command(
+        cls,
+        cmd: Union[List[str], CommandArgs],
+        show_stdout: bool = True,
+        cwd: Optional[str] = None,
+        on_returncode: 'Literal["raise", "warn", "ignore"]' = "raise",
+        extra_ok_returncodes: Optional[Iterable[int]] = None,
+        command_desc: Optional[str] = None,
+        extra_environ: Optional[Mapping[str, Any]] = None,
+        spinner: Optional[SpinnerInterface] = None,
+        log_failed_cmd: bool = True,
+        stdout_only: bool = False,
+    ) -> str:
+        """
+        Run a VCS subcommand
+        This is simply a wrapper around call_subprocess that adds the VCS
+        command name, and checks that the VCS is available
+        """
+        cmd = make_command(cls.name, *cmd)
+        if command_desc is None:
+            command_desc = format_command_args(cmd)
+        try:
+            return call_subprocess(
+                cmd,
+                show_stdout,
+                cwd,
+                on_returncode=on_returncode,
+                extra_ok_returncodes=extra_ok_returncodes,
+                command_desc=command_desc,
+                extra_environ=extra_environ,
+                unset_environ=cls.unset_environ,
+                spinner=spinner,
+                log_failed_cmd=log_failed_cmd,
+                stdout_only=stdout_only,
+            )
+        except NotADirectoryError:
+            raise BadCommand(f"Cannot find command {cls.name!r} - invalid PATH")
+        except FileNotFoundError:
+            # errno.ENOENT = no such file or directory
+            # In other words, the VCS executable isn't available
+            raise BadCommand(
+                f"Cannot find command {cls.name!r} - do you have "
+                f"{cls.name!r} installed and in your PATH?"
+            )
+        except PermissionError:
+            # errno.EACCES = Permission denied
+            # This error occurs, for instance, when the command is installed
+            # only for another user. So, the current user don't have
+            # permission to call the other user command.
+            raise BadCommand(
+                f"No permission to execute {cls.name!r} - install it "
+                f"locally, globally (ask admin), or check your PATH. "
+                f"See possible solutions at "
+                f"https://pip.pypa.io/en/latest/reference/pip_freeze/"
+                f"#fixing-permission-denied."
+            )
+
+    @classmethod
+    def is_repository_directory(cls, path: str) -> bool:
+        """
+        Return whether a directory path is a repository directory.
+        """
+        logger.debug("Checking in %s for %s (%s)...", path, cls.dirname, cls.name)
+        return os.path.exists(os.path.join(path, cls.dirname))
+
+    @classmethod
+    def get_repository_root(cls, location: str) -> Optional[str]:
+        """
+        Return the "root" (top-level) directory controlled by the vcs,
+        or `None` if the directory is not in any.
+
+        It is meant to be overridden to implement smarter detection
+        mechanisms for specific vcs.
+
+        This can do more than is_repository_directory() alone. For
+        example, the Git override checks that Git is actually available.
+        """
+        if cls.is_repository_directory(location):
+            return location
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_internal/wheel_builder.py b/backend/venv/lib/python3.12/site-packages/pip/_internal/wheel_builder.py
new file mode 100644
index 00000000..93f8e1f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_internal/wheel_builder.py
@@ -0,0 +1,354 @@
+"""Orchestrator for building wheels from InstallRequirements.
+"""
+
+import logging
+import os.path
+import re
+import shutil
+from typing import Iterable, List, Optional, Tuple
+
+from pip._vendor.packaging.utils import canonicalize_name, canonicalize_version
+from pip._vendor.packaging.version import InvalidVersion, Version
+
+from pip._internal.cache import WheelCache
+from pip._internal.exceptions import InvalidWheelFilename, UnsupportedWheel
+from pip._internal.metadata import FilesystemWheel, get_wheel_distribution
+from pip._internal.models.link import Link
+from pip._internal.models.wheel import Wheel
+from pip._internal.operations.build.wheel import build_wheel_pep517
+from pip._internal.operations.build.wheel_editable import build_wheel_editable
+from pip._internal.operations.build.wheel_legacy import build_wheel_legacy
+from pip._internal.req.req_install import InstallRequirement
+from pip._internal.utils.logging import indent_log
+from pip._internal.utils.misc import ensure_dir, hash_file
+from pip._internal.utils.setuptools_build import make_setuptools_clean_args
+from pip._internal.utils.subprocess import call_subprocess
+from pip._internal.utils.temp_dir import TempDirectory
+from pip._internal.utils.urls import path_to_url
+from pip._internal.vcs import vcs
+
+logger = logging.getLogger(__name__)
+
+_egg_info_re = re.compile(r"([a-z0-9_.]+)-([a-z0-9_.!+-]+)", re.IGNORECASE)
+
+BuildResult = Tuple[List[InstallRequirement], List[InstallRequirement]]
+
+
+def _contains_egg_info(s: str) -> bool:
+    """Determine whether the string looks like an egg_info.
+
+    :param s: The string to parse. E.g. foo-2.1
+    """
+    return bool(_egg_info_re.search(s))
+
+
+def _should_build(
+    req: InstallRequirement,
+    need_wheel: bool,
+) -> bool:
+    """Return whether an InstallRequirement should be built into a wheel."""
+    if req.constraint:
+        # never build requirements that are merely constraints
+        return False
+    if req.is_wheel:
+        if need_wheel:
+            logger.info(
+                "Skipping %s, due to already being wheel.",
+                req.name,
+            )
+        return False
+
+    if need_wheel:
+        # i.e. pip wheel, not pip install
+        return True
+
+    # From this point, this concerns the pip install command only
+    # (need_wheel=False).
+
+    if not req.source_dir:
+        return False
+
+    if req.editable:
+        # we only build PEP 660 editable requirements
+        return req.supports_pyproject_editable
+
+    return True
+
+
+def should_build_for_wheel_command(
+    req: InstallRequirement,
+) -> bool:
+    return _should_build(req, need_wheel=True)
+
+
+def should_build_for_install_command(
+    req: InstallRequirement,
+) -> bool:
+    return _should_build(req, need_wheel=False)
+
+
+def _should_cache(
+    req: InstallRequirement,
+) -> Optional[bool]:
+    """
+    Return whether a built InstallRequirement can be stored in the persistent
+    wheel cache, assuming the wheel cache is available, and _should_build()
+    has determined a wheel needs to be built.
+    """
+    if req.editable or not req.source_dir:
+        # never cache editable requirements
+        return False
+
+    if req.link and req.link.is_vcs:
+        # VCS checkout. Do not cache
+        # unless it points to an immutable commit hash.
+        assert not req.editable
+        assert req.source_dir
+        vcs_backend = vcs.get_backend_for_scheme(req.link.scheme)
+        assert vcs_backend
+        if vcs_backend.is_immutable_rev_checkout(req.link.url, req.source_dir):
+            return True
+        return False
+
+    assert req.link
+    base, ext = req.link.splitext()
+    if _contains_egg_info(base):
+        return True
+
+    # Otherwise, do not cache.
+    return False
+
+
+def _get_cache_dir(
+    req: InstallRequirement,
+    wheel_cache: WheelCache,
+) -> str:
+    """Return the persistent or temporary cache directory where the built
+    wheel need to be stored.
+    """
+    cache_available = bool(wheel_cache.cache_dir)
+    assert req.link
+    if cache_available and _should_cache(req):
+        cache_dir = wheel_cache.get_path_for_link(req.link)
+    else:
+        cache_dir = wheel_cache.get_ephem_path_for_link(req.link)
+    return cache_dir
+
+
+def _verify_one(req: InstallRequirement, wheel_path: str) -> None:
+    canonical_name = canonicalize_name(req.name or "")
+    w = Wheel(os.path.basename(wheel_path))
+    if canonicalize_name(w.name) != canonical_name:
+        raise InvalidWheelFilename(
+            f"Wheel has unexpected file name: expected {canonical_name!r}, "
+            f"got {w.name!r}",
+        )
+    dist = get_wheel_distribution(FilesystemWheel(wheel_path), canonical_name)
+    dist_verstr = str(dist.version)
+    if canonicalize_version(dist_verstr) != canonicalize_version(w.version):
+        raise InvalidWheelFilename(
+            f"Wheel has unexpected file name: expected {dist_verstr!r}, "
+            f"got {w.version!r}",
+        )
+    metadata_version_value = dist.metadata_version
+    if metadata_version_value is None:
+        raise UnsupportedWheel("Missing Metadata-Version")
+    try:
+        metadata_version = Version(metadata_version_value)
+    except InvalidVersion:
+        msg = f"Invalid Metadata-Version: {metadata_version_value}"
+        raise UnsupportedWheel(msg)
+    if metadata_version >= Version("1.2") and not isinstance(dist.version, Version):
+        raise UnsupportedWheel(
+            f"Metadata 1.2 mandates PEP 440 version, but {dist_verstr!r} is not"
+        )
+
+
+def _build_one(
+    req: InstallRequirement,
+    output_dir: str,
+    verify: bool,
+    build_options: List[str],
+    global_options: List[str],
+    editable: bool,
+) -> Optional[str]:
+    """Build one wheel.
+
+    :return: The filename of the built wheel, or None if the build failed.
+    """
+    artifact = "editable" if editable else "wheel"
+    try:
+        ensure_dir(output_dir)
+    except OSError as e:
+        logger.warning(
+            "Building %s for %s failed: %s",
+            artifact,
+            req.name,
+            e,
+        )
+        return None
+
+    # Install build deps into temporary directory (PEP 518)
+    with req.build_env:
+        wheel_path = _build_one_inside_env(
+            req, output_dir, build_options, global_options, editable
+        )
+    if wheel_path and verify:
+        try:
+            _verify_one(req, wheel_path)
+        except (InvalidWheelFilename, UnsupportedWheel) as e:
+            logger.warning("Built %s for %s is invalid: %s", artifact, req.name, e)
+            return None
+    return wheel_path
+
+
+def _build_one_inside_env(
+    req: InstallRequirement,
+    output_dir: str,
+    build_options: List[str],
+    global_options: List[str],
+    editable: bool,
+) -> Optional[str]:
+    with TempDirectory(kind="wheel") as temp_dir:
+        assert req.name
+        if req.use_pep517:
+            assert req.metadata_directory
+            assert req.pep517_backend
+            if global_options:
+                logger.warning(
+                    "Ignoring --global-option when building %s using PEP 517", req.name
+                )
+            if build_options:
+                logger.warning(
+                    "Ignoring --build-option when building %s using PEP 517", req.name
+                )
+            if editable:
+                wheel_path = build_wheel_editable(
+                    name=req.name,
+                    backend=req.pep517_backend,
+                    metadata_directory=req.metadata_directory,
+                    tempd=temp_dir.path,
+                )
+            else:
+                wheel_path = build_wheel_pep517(
+                    name=req.name,
+                    backend=req.pep517_backend,
+                    metadata_directory=req.metadata_directory,
+                    tempd=temp_dir.path,
+                )
+        else:
+            wheel_path = build_wheel_legacy(
+                name=req.name,
+                setup_py_path=req.setup_py_path,
+                source_dir=req.unpacked_source_directory,
+                global_options=global_options,
+                build_options=build_options,
+                tempd=temp_dir.path,
+            )
+
+        if wheel_path is not None:
+            wheel_name = os.path.basename(wheel_path)
+            dest_path = os.path.join(output_dir, wheel_name)
+            try:
+                wheel_hash, length = hash_file(wheel_path)
+                shutil.move(wheel_path, dest_path)
+                logger.info(
+                    "Created wheel for %s: filename=%s size=%d sha256=%s",
+                    req.name,
+                    wheel_name,
+                    length,
+                    wheel_hash.hexdigest(),
+                )
+                logger.info("Stored in directory: %s", output_dir)
+                return dest_path
+            except Exception as e:
+                logger.warning(
+                    "Building wheel for %s failed: %s",
+                    req.name,
+                    e,
+                )
+        # Ignore return, we can't do anything else useful.
+        if not req.use_pep517:
+            _clean_one_legacy(req, global_options)
+        return None
+
+
+def _clean_one_legacy(req: InstallRequirement, global_options: List[str]) -> bool:
+    clean_args = make_setuptools_clean_args(
+        req.setup_py_path,
+        global_options=global_options,
+    )
+
+    logger.info("Running setup.py clean for %s", req.name)
+    try:
+        call_subprocess(
+            clean_args, command_desc="python setup.py clean", cwd=req.source_dir
+        )
+        return True
+    except Exception:
+        logger.error("Failed cleaning build dir for %s", req.name)
+        return False
+
+
+def build(
+    requirements: Iterable[InstallRequirement],
+    wheel_cache: WheelCache,
+    verify: bool,
+    build_options: List[str],
+    global_options: List[str],
+) -> BuildResult:
+    """Build wheels.
+
+    :return: The list of InstallRequirement that succeeded to build and
+        the list of InstallRequirement that failed to build.
+    """
+    if not requirements:
+        return [], []
+
+    # Build the wheels.
+    logger.info(
+        "Building wheels for collected packages: %s",
+        ", ".join(req.name for req in requirements),  # type: ignore
+    )
+
+    with indent_log():
+        build_successes, build_failures = [], []
+        for req in requirements:
+            assert req.name
+            cache_dir = _get_cache_dir(req, wheel_cache)
+            wheel_file = _build_one(
+                req,
+                cache_dir,
+                verify,
+                build_options,
+                global_options,
+                req.editable and req.permit_editable_wheels,
+            )
+            if wheel_file:
+                # Record the download origin in the cache
+                if req.download_info is not None:
+                    # download_info is guaranteed to be set because when we build an
+                    # InstallRequirement it has been through the preparer before, but
+                    # let's be cautious.
+                    wheel_cache.record_download_origin(cache_dir, req.download_info)
+                # Update the link for this.
+                req.link = Link(path_to_url(wheel_file))
+                req.local_file_path = req.link.file_path
+                assert req.link.is_wheel
+                build_successes.append(req)
+            else:
+                build_failures.append(req)
+
+    # notify success/failure
+    if build_successes:
+        logger.info(
+            "Successfully built %s",
+            " ".join([req.name for req in build_successes]),  # type: ignore
+        )
+    if build_failures:
+        logger.info(
+            "Failed to build %s",
+            " ".join([req.name for req in build_failures]),  # type: ignore
+        )
+    # Return a list of requirements that failed to build
+    return build_successes, build_failures
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__init__.py
new file mode 100644
index 00000000..561089cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__init__.py
@@ -0,0 +1,116 @@
+"""
+pip._vendor is for vendoring dependencies of pip to prevent needing pip to
+depend on something external.
+
+Files inside of pip._vendor should be considered immutable and should only be
+updated to versions from upstream.
+"""
+from __future__ import absolute_import
+
+import glob
+import os.path
+import sys
+
+# Downstream redistributors which have debundled our dependencies should also
+# patch this value to be true. This will trigger the additional patching
+# to cause things like "six" to be available as pip.
+DEBUNDLED = False
+
+# By default, look in this directory for a bunch of .whl files which we will
+# add to the beginning of sys.path before attempting to import anything. This
+# is done to support downstream re-distributors like Debian and Fedora who
+# wish to create their own Wheels for our dependencies to aid in debundling.
+WHEEL_DIR = os.path.abspath(os.path.dirname(__file__))
+
+
+# Define a small helper function to alias our vendored modules to the real ones
+# if the vendored ones do not exist. This idea of this was taken from
+# https://github.com/kennethreitz/requests/pull/2567.
+def vendored(modulename):
+    vendored_name = "{0}.{1}".format(__name__, modulename)
+
+    try:
+        __import__(modulename, globals(), locals(), level=0)
+    except ImportError:
+        # We can just silently allow import failures to pass here. If we
+        # got to this point it means that ``import pip._vendor.whatever``
+        # failed and so did ``import whatever``. Since we're importing this
+        # upfront in an attempt to alias imports, not erroring here will
+        # just mean we get a regular import error whenever pip *actually*
+        # tries to import one of these modules to use it, which actually
+        # gives us a better error message than we would have otherwise
+        # gotten.
+        pass
+    else:
+        sys.modules[vendored_name] = sys.modules[modulename]
+        base, head = vendored_name.rsplit(".", 1)
+        setattr(sys.modules[base], head, sys.modules[modulename])
+
+
+# If we're operating in a debundled setup, then we want to go ahead and trigger
+# the aliasing of our vendored libraries as well as looking for wheels to add
+# to our sys.path. This will cause all of this code to be a no-op typically
+# however downstream redistributors can enable it in a consistent way across
+# all platforms.
+if DEBUNDLED:
+    # Actually look inside of WHEEL_DIR to find .whl files and add them to the
+    # front of our sys.path.
+    sys.path[:] = glob.glob(os.path.join(WHEEL_DIR, "*.whl")) + sys.path
+
+    # Actually alias all of our vendored dependencies.
+    vendored("cachecontrol")
+    vendored("certifi")
+    vendored("distlib")
+    vendored("distro")
+    vendored("packaging")
+    vendored("packaging.version")
+    vendored("packaging.specifiers")
+    vendored("pkg_resources")
+    vendored("platformdirs")
+    vendored("progress")
+    vendored("pyproject_hooks")
+    vendored("requests")
+    vendored("requests.exceptions")
+    vendored("requests.packages")
+    vendored("requests.packages.urllib3")
+    vendored("requests.packages.urllib3._collections")
+    vendored("requests.packages.urllib3.connection")
+    vendored("requests.packages.urllib3.connectionpool")
+    vendored("requests.packages.urllib3.contrib")
+    vendored("requests.packages.urllib3.contrib.ntlmpool")
+    vendored("requests.packages.urllib3.contrib.pyopenssl")
+    vendored("requests.packages.urllib3.exceptions")
+    vendored("requests.packages.urllib3.fields")
+    vendored("requests.packages.urllib3.filepost")
+    vendored("requests.packages.urllib3.packages")
+    vendored("requests.packages.urllib3.packages.ordered_dict")
+    vendored("requests.packages.urllib3.packages.six")
+    vendored("requests.packages.urllib3.packages.ssl_match_hostname")
+    vendored("requests.packages.urllib3.packages.ssl_match_hostname."
+             "_implementation")
+    vendored("requests.packages.urllib3.poolmanager")
+    vendored("requests.packages.urllib3.request")
+    vendored("requests.packages.urllib3.response")
+    vendored("requests.packages.urllib3.util")
+    vendored("requests.packages.urllib3.util.connection")
+    vendored("requests.packages.urllib3.util.request")
+    vendored("requests.packages.urllib3.util.response")
+    vendored("requests.packages.urllib3.util.retry")
+    vendored("requests.packages.urllib3.util.ssl_")
+    vendored("requests.packages.urllib3.util.timeout")
+    vendored("requests.packages.urllib3.util.url")
+    vendored("resolvelib")
+    vendored("rich")
+    vendored("rich.console")
+    vendored("rich.highlighter")
+    vendored("rich.logging")
+    vendored("rich.markup")
+    vendored("rich.progress")
+    vendored("rich.segment")
+    vendored("rich.style")
+    vendored("rich.text")
+    vendored("rich.traceback")
+    if sys.version_info < (3, 11):
+        vendored("tomli")
+    vendored("truststore")
+    vendored("urllib3")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e15d157b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/typing_extensions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/typing_extensions.cpython-312.pyc
new file mode 100644
index 00000000..252902b2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/__pycache__/typing_extensions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__init__.py
new file mode 100644
index 00000000..21916243
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__init__.py
@@ -0,0 +1,29 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
+"""CacheControl import Interface.
+
+Make it easy to import from cachecontrol without long namespaces.
+"""
+
+__author__ = "Eric Larson"
+__email__ = "eric@ionrock.org"
+__version__ = "0.14.1"
+
+from pip._vendor.cachecontrol.adapter import CacheControlAdapter
+from pip._vendor.cachecontrol.controller import CacheController
+from pip._vendor.cachecontrol.wrapper import CacheControl
+
+__all__ = [
+    "__author__",
+    "__email__",
+    "__version__",
+    "CacheControlAdapter",
+    "CacheController",
+    "CacheControl",
+]
+
+import logging
+
+logging.getLogger(__name__).addHandler(logging.NullHandler())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..16621942
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/_cmd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/_cmd.cpython-312.pyc
new file mode 100644
index 00000000..eefef031
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/_cmd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/adapter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/adapter.cpython-312.pyc
new file mode 100644
index 00000000..b51213d0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/adapter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/cache.cpython-312.pyc
new file mode 100644
index 00000000..30c19bbf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/controller.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/controller.cpython-312.pyc
new file mode 100644
index 00000000..6a11b6a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/controller.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/filewrapper.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/filewrapper.cpython-312.pyc
new file mode 100644
index 00000000..009d5dc0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/filewrapper.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/heuristics.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/heuristics.cpython-312.pyc
new file mode 100644
index 00000000..fef7daec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/heuristics.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/serialize.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/serialize.cpython-312.pyc
new file mode 100644
index 00000000..e088fda6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/serialize.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/wrapper.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/wrapper.cpython-312.pyc
new file mode 100644
index 00000000..26757eba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/__pycache__/wrapper.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/_cmd.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/_cmd.py
new file mode 100644
index 00000000..2c84208a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/_cmd.py
@@ -0,0 +1,70 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import logging
+from argparse import ArgumentParser
+from typing import TYPE_CHECKING
+
+from pip._vendor import requests
+
+from pip._vendor.cachecontrol.adapter import CacheControlAdapter
+from pip._vendor.cachecontrol.cache import DictCache
+from pip._vendor.cachecontrol.controller import logger
+
+if TYPE_CHECKING:
+    from argparse import Namespace
+
+    from pip._vendor.cachecontrol.controller import CacheController
+
+
+def setup_logging() -> None:
+    logger.setLevel(logging.DEBUG)
+    handler = logging.StreamHandler()
+    logger.addHandler(handler)
+
+
+def get_session() -> requests.Session:
+    adapter = CacheControlAdapter(
+        DictCache(), cache_etags=True, serializer=None, heuristic=None
+    )
+    sess = requests.Session()
+    sess.mount("http://", adapter)
+    sess.mount("https://", adapter)
+
+    sess.cache_controller = adapter.controller  # type: ignore[attr-defined]
+    return sess
+
+
+def get_args() -> Namespace:
+    parser = ArgumentParser()
+    parser.add_argument("url", help="The URL to try and cache")
+    return parser.parse_args()
+
+
+def main() -> None:
+    args = get_args()
+    sess = get_session()
+
+    # Make a request to get a response
+    resp = sess.get(args.url)
+
+    # Turn on logging
+    setup_logging()
+
+    # try setting the cache
+    cache_controller: CacheController = (
+        sess.cache_controller  # type: ignore[attr-defined]
+    )
+    cache_controller.cache_response(resp.request, resp.raw)
+
+    # Now try to get it
+    if cache_controller.cached_request(resp.request):
+        print("Cached!")
+    else:
+        print("Not cached :(")
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/adapter.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/adapter.py
new file mode 100644
index 00000000..34a9eb82
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/adapter.py
@@ -0,0 +1,161 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import functools
+import types
+import zlib
+from typing import TYPE_CHECKING, Any, Collection, Mapping
+
+from pip._vendor.requests.adapters import HTTPAdapter
+
+from pip._vendor.cachecontrol.cache import DictCache
+from pip._vendor.cachecontrol.controller import PERMANENT_REDIRECT_STATUSES, CacheController
+from pip._vendor.cachecontrol.filewrapper import CallbackFileWrapper
+
+if TYPE_CHECKING:
+    from pip._vendor.requests import PreparedRequest, Response
+    from pip._vendor.urllib3 import HTTPResponse
+
+    from pip._vendor.cachecontrol.cache import BaseCache
+    from pip._vendor.cachecontrol.heuristics import BaseHeuristic
+    from pip._vendor.cachecontrol.serialize import Serializer
+
+
+class CacheControlAdapter(HTTPAdapter):
+    invalidating_methods = {"PUT", "PATCH", "DELETE"}
+
+    def __init__(
+        self,
+        cache: BaseCache | None = None,
+        cache_etags: bool = True,
+        controller_class: type[CacheController] | None = None,
+        serializer: Serializer | None = None,
+        heuristic: BaseHeuristic | None = None,
+        cacheable_methods: Collection[str] | None = None,
+        *args: Any,
+        **kw: Any,
+    ) -> None:
+        super().__init__(*args, **kw)
+        self.cache = DictCache() if cache is None else cache
+        self.heuristic = heuristic
+        self.cacheable_methods = cacheable_methods or ("GET",)
+
+        controller_factory = controller_class or CacheController
+        self.controller = controller_factory(
+            self.cache, cache_etags=cache_etags, serializer=serializer
+        )
+
+    def send(
+        self,
+        request: PreparedRequest,
+        stream: bool = False,
+        timeout: None | float | tuple[float, float] | tuple[float, None] = None,
+        verify: bool | str = True,
+        cert: (None | bytes | str | tuple[bytes | str, bytes | str]) = None,
+        proxies: Mapping[str, str] | None = None,
+        cacheable_methods: Collection[str] | None = None,
+    ) -> Response:
+        """
+        Send a request. Use the request information to see if it
+        exists in the cache and cache the response if we need to and can.
+        """
+        cacheable = cacheable_methods or self.cacheable_methods
+        if request.method in cacheable:
+            try:
+                cached_response = self.controller.cached_request(request)
+            except zlib.error:
+                cached_response = None
+            if cached_response:
+                return self.build_response(request, cached_response, from_cache=True)
+
+            # check for etags and add headers if appropriate
+            request.headers.update(self.controller.conditional_headers(request))
+
+        resp = super().send(request, stream, timeout, verify, cert, proxies)
+
+        return resp
+
+    def build_response(  # type: ignore[override]
+        self,
+        request: PreparedRequest,
+        response: HTTPResponse,
+        from_cache: bool = False,
+        cacheable_methods: Collection[str] | None = None,
+    ) -> Response:
+        """
+        Build a response by making a request or using the cache.
+
+        This will end up calling send and returning a potentially
+        cached response
+        """
+        cacheable = cacheable_methods or self.cacheable_methods
+        if not from_cache and request.method in cacheable:
+            # Check for any heuristics that might update headers
+            # before trying to cache.
+            if self.heuristic:
+                response = self.heuristic.apply(response)
+
+            # apply any expiration heuristics
+            if response.status == 304:
+                # We must have sent an ETag request. This could mean
+                # that we've been expired already or that we simply
+                # have an etag. In either case, we want to try and
+                # update the cache if that is the case.
+                cached_response = self.controller.update_cached_response(
+                    request, response
+                )
+
+                if cached_response is not response:
+                    from_cache = True
+
+                # We are done with the server response, read a
+                # possible response body (compliant servers will
+                # not return one, but we cannot be 100% sure) and
+                # release the connection back to the pool.
+                response.read(decode_content=False)
+                response.release_conn()
+
+                response = cached_response
+
+            # We always cache the 301 responses
+            elif int(response.status) in PERMANENT_REDIRECT_STATUSES:
+                self.controller.cache_response(request, response)
+            else:
+                # Wrap the response file with a wrapper that will cache the
+                #   response when the stream has been consumed.
+                response._fp = CallbackFileWrapper(  # type: ignore[assignment]
+                    response._fp,  # type: ignore[arg-type]
+                    functools.partial(
+                        self.controller.cache_response, request, response
+                    ),
+                )
+                if response.chunked:
+                    super_update_chunk_length = response._update_chunk_length
+
+                    def _update_chunk_length(self: HTTPResponse) -> None:
+                        super_update_chunk_length()
+                        if self.chunk_left == 0:
+                            self._fp._close()  # type: ignore[union-attr]
+
+                    response._update_chunk_length = types.MethodType(  # type: ignore[method-assign]
+                        _update_chunk_length, response
+                    )
+
+        resp: Response = super().build_response(request, response)
+
+        # See if we should invalidate the cache.
+        if request.method in self.invalidating_methods and resp.ok:
+            assert request.url is not None
+            cache_url = self.controller.cache_url(request.url)
+            self.cache.delete(cache_url)
+
+        # Give the request a from_cache attr to let people use it
+        resp.from_cache = from_cache  # type: ignore[attr-defined]
+
+        return resp
+
+    def close(self) -> None:
+        self.cache.close()
+        super().close()  # type: ignore[no-untyped-call]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/cache.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/cache.py
new file mode 100644
index 00000000..91598e92
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/cache.py
@@ -0,0 +1,75 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
+"""
+The cache object API for implementing caches. The default is a thread
+safe in-memory dictionary.
+"""
+
+from __future__ import annotations
+
+from threading import Lock
+from typing import IO, TYPE_CHECKING, MutableMapping
+
+if TYPE_CHECKING:
+    from datetime import datetime
+
+
+class BaseCache:
+    def get(self, key: str) -> bytes | None:
+        raise NotImplementedError()
+
+    def set(
+        self, key: str, value: bytes, expires: int | datetime | None = None
+    ) -> None:
+        raise NotImplementedError()
+
+    def delete(self, key: str) -> None:
+        raise NotImplementedError()
+
+    def close(self) -> None:
+        pass
+
+
+class DictCache(BaseCache):
+    def __init__(self, init_dict: MutableMapping[str, bytes] | None = None) -> None:
+        self.lock = Lock()
+        self.data = init_dict or {}
+
+    def get(self, key: str) -> bytes | None:
+        return self.data.get(key, None)
+
+    def set(
+        self, key: str, value: bytes, expires: int | datetime | None = None
+    ) -> None:
+        with self.lock:
+            self.data.update({key: value})
+
+    def delete(self, key: str) -> None:
+        with self.lock:
+            if key in self.data:
+                self.data.pop(key)
+
+
+class SeparateBodyBaseCache(BaseCache):
+    """
+    In this variant, the body is not stored mixed in with the metadata, but is
+    passed in (as a bytes-like object) in a separate call to ``set_body()``.
+
+    That is, the expected interaction pattern is::
+
+        cache.set(key, serialized_metadata)
+        cache.set_body(key)
+
+    Similarly, the body should be loaded separately via ``get_body()``.
+    """
+
+    def set_body(self, key: str, body: bytes) -> None:
+        raise NotImplementedError()
+
+    def get_body(self, key: str) -> IO[bytes] | None:
+        """
+        Return the body as file-like object.
+        """
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__init__.py
new file mode 100644
index 00000000..24ff469f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__init__.py
@@ -0,0 +1,8 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
+from pip._vendor.cachecontrol.caches.file_cache import FileCache, SeparateBodyFileCache
+from pip._vendor.cachecontrol.caches.redis_cache import RedisCache
+
+__all__ = ["FileCache", "SeparateBodyFileCache", "RedisCache"]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..a44e9725
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/file_cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/file_cache.cpython-312.pyc
new file mode 100644
index 00000000..b47705a4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/file_cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/redis_cache.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/redis_cache.cpython-312.pyc
new file mode 100644
index 00000000..a1133863
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/__pycache__/redis_cache.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/file_cache.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/file_cache.py
new file mode 100644
index 00000000..81d2ef46
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/file_cache.py
@@ -0,0 +1,182 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import hashlib
+import os
+from textwrap import dedent
+from typing import IO, TYPE_CHECKING
+from pathlib import Path
+
+from pip._vendor.cachecontrol.cache import BaseCache, SeparateBodyBaseCache
+from pip._vendor.cachecontrol.controller import CacheController
+
+if TYPE_CHECKING:
+    from datetime import datetime
+
+    from filelock import BaseFileLock
+
+
+def _secure_open_write(filename: str, fmode: int) -> IO[bytes]:
+    # We only want to write to this file, so open it in write only mode
+    flags = os.O_WRONLY
+
+    # os.O_CREAT | os.O_EXCL will fail if the file already exists, so we only
+    #  will open *new* files.
+    # We specify this because we want to ensure that the mode we pass is the
+    # mode of the file.
+    flags |= os.O_CREAT | os.O_EXCL
+
+    # Do not follow symlinks to prevent someone from making a symlink that
+    # we follow and insecurely open a cache file.
+    if hasattr(os, "O_NOFOLLOW"):
+        flags |= os.O_NOFOLLOW
+
+    # On Windows we'll mark this file as binary
+    if hasattr(os, "O_BINARY"):
+        flags |= os.O_BINARY
+
+    # Before we open our file, we want to delete any existing file that is
+    # there
+    try:
+        os.remove(filename)
+    except OSError:
+        # The file must not exist already, so we can just skip ahead to opening
+        pass
+
+    # Open our file, the use of os.O_CREAT | os.O_EXCL will ensure that if a
+    # race condition happens between the os.remove and this line, that an
+    # error will be raised. Because we utilize a lockfile this should only
+    # happen if someone is attempting to attack us.
+    fd = os.open(filename, flags, fmode)
+    try:
+        return os.fdopen(fd, "wb")
+
+    except:
+        # An error occurred wrapping our FD in a file object
+        os.close(fd)
+        raise
+
+
+class _FileCacheMixin:
+    """Shared implementation for both FileCache variants."""
+
+    def __init__(
+        self,
+        directory: str | Path,
+        forever: bool = False,
+        filemode: int = 0o0600,
+        dirmode: int = 0o0700,
+        lock_class: type[BaseFileLock] | None = None,
+    ) -> None:
+        try:
+            if lock_class is None:
+                from filelock import FileLock
+
+                lock_class = FileLock
+        except ImportError:
+            notice = dedent(
+                """
+            NOTE: In order to use the FileCache you must have
+            filelock installed. You can install it via pip:
+              pip install cachecontrol[filecache]
+            """
+            )
+            raise ImportError(notice)
+
+        self.directory = directory
+        self.forever = forever
+        self.filemode = filemode
+        self.dirmode = dirmode
+        self.lock_class = lock_class
+
+    @staticmethod
+    def encode(x: str) -> str:
+        return hashlib.sha224(x.encode()).hexdigest()
+
+    def _fn(self, name: str) -> str:
+        # NOTE: This method should not change as some may depend on it.
+        #       See: https://github.com/ionrock/cachecontrol/issues/63
+        hashed = self.encode(name)
+        parts = list(hashed[:5]) + [hashed]
+        return os.path.join(self.directory, *parts)
+
+    def get(self, key: str) -> bytes | None:
+        name = self._fn(key)
+        try:
+            with open(name, "rb") as fh:
+                return fh.read()
+
+        except FileNotFoundError:
+            return None
+
+    def set(
+        self, key: str, value: bytes, expires: int | datetime | None = None
+    ) -> None:
+        name = self._fn(key)
+        self._write(name, value)
+
+    def _write(self, path: str, data: bytes) -> None:
+        """
+        Safely write the data to the given path.
+        """
+        # Make sure the directory exists
+        try:
+            os.makedirs(os.path.dirname(path), self.dirmode)
+        except OSError:
+            pass
+
+        with self.lock_class(path + ".lock"):
+            # Write our actual file
+            with _secure_open_write(path, self.filemode) as fh:
+                fh.write(data)
+
+    def _delete(self, key: str, suffix: str) -> None:
+        name = self._fn(key) + suffix
+        if not self.forever:
+            try:
+                os.remove(name)
+            except FileNotFoundError:
+                pass
+
+
+class FileCache(_FileCacheMixin, BaseCache):
+    """
+    Traditional FileCache: body is stored in memory, so not suitable for large
+    downloads.
+    """
+
+    def delete(self, key: str) -> None:
+        self._delete(key, "")
+
+
+class SeparateBodyFileCache(_FileCacheMixin, SeparateBodyBaseCache):
+    """
+    Memory-efficient FileCache: body is stored in a separate file, reducing
+    peak memory usage.
+    """
+
+    def get_body(self, key: str) -> IO[bytes] | None:
+        name = self._fn(key) + ".body"
+        try:
+            return open(name, "rb")
+        except FileNotFoundError:
+            return None
+
+    def set_body(self, key: str, body: bytes) -> None:
+        name = self._fn(key) + ".body"
+        self._write(name, body)
+
+    def delete(self, key: str) -> None:
+        self._delete(key, "")
+        self._delete(key, ".body")
+
+
+def url_to_file_path(url: str, filecache: FileCache) -> str:
+    """Return the file cache path based on the URL.
+
+    This does not ensure the file exists!
+    """
+    key = CacheController.cache_url(url)
+    return filecache._fn(key)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/redis_cache.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/redis_cache.py
new file mode 100644
index 00000000..f4f68c47
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/caches/redis_cache.py
@@ -0,0 +1,48 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+
+from datetime import datetime, timezone
+from typing import TYPE_CHECKING
+
+from pip._vendor.cachecontrol.cache import BaseCache
+
+if TYPE_CHECKING:
+    from redis import Redis
+
+
+class RedisCache(BaseCache):
+    def __init__(self, conn: Redis[bytes]) -> None:
+        self.conn = conn
+
+    def get(self, key: str) -> bytes | None:
+        return self.conn.get(key)
+
+    def set(
+        self, key: str, value: bytes, expires: int | datetime | None = None
+    ) -> None:
+        if not expires:
+            self.conn.set(key, value)
+        elif isinstance(expires, datetime):
+            now_utc = datetime.now(timezone.utc)
+            if expires.tzinfo is None:
+                now_utc = now_utc.replace(tzinfo=None)
+            delta = expires - now_utc
+            self.conn.setex(key, int(delta.total_seconds()), value)
+        else:
+            self.conn.setex(key, expires, value)
+
+    def delete(self, key: str) -> None:
+        self.conn.delete(key)
+
+    def clear(self) -> None:
+        """Helper for clearing all the keys in a database. Use with
+        caution!"""
+        for key in self.conn.keys():
+            self.conn.delete(key)
+
+    def close(self) -> None:
+        """Redis uses connection pooling, no need to close the connection."""
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/controller.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/controller.py
new file mode 100644
index 00000000..f0ff6e1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/controller.py
@@ -0,0 +1,500 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
+"""
+The httplib2 algorithms ported for use with requests.
+"""
+
+from __future__ import annotations
+
+import calendar
+import logging
+import re
+import time
+from email.utils import parsedate_tz
+from typing import TYPE_CHECKING, Collection, Mapping
+
+from pip._vendor.requests.structures import CaseInsensitiveDict
+
+from pip._vendor.cachecontrol.cache import DictCache, SeparateBodyBaseCache
+from pip._vendor.cachecontrol.serialize import Serializer
+
+if TYPE_CHECKING:
+    from typing import Literal
+
+    from pip._vendor.requests import PreparedRequest
+    from pip._vendor.urllib3 import HTTPResponse
+
+    from pip._vendor.cachecontrol.cache import BaseCache
+
+logger = logging.getLogger(__name__)
+
+URI = re.compile(r"^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?")
+
+PERMANENT_REDIRECT_STATUSES = (301, 308)
+
+
+def parse_uri(uri: str) -> tuple[str, str, str, str, str]:
+    """Parses a URI using the regex given in Appendix B of RFC 3986.
+
+    (scheme, authority, path, query, fragment) = parse_uri(uri)
+    """
+    match = URI.match(uri)
+    assert match is not None
+    groups = match.groups()
+    return (groups[1], groups[3], groups[4], groups[6], groups[8])
+
+
+class CacheController:
+    """An interface to see if request should cached or not."""
+
+    def __init__(
+        self,
+        cache: BaseCache | None = None,
+        cache_etags: bool = True,
+        serializer: Serializer | None = None,
+        status_codes: Collection[int] | None = None,
+    ):
+        self.cache = DictCache() if cache is None else cache
+        self.cache_etags = cache_etags
+        self.serializer = serializer or Serializer()
+        self.cacheable_status_codes = status_codes or (200, 203, 300, 301, 308)
+
+    @classmethod
+    def _urlnorm(cls, uri: str) -> str:
+        """Normalize the URL to create a safe key for the cache"""
+        (scheme, authority, path, query, fragment) = parse_uri(uri)
+        if not scheme or not authority:
+            raise Exception("Only absolute URIs are allowed. uri = %s" % uri)
+
+        scheme = scheme.lower()
+        authority = authority.lower()
+
+        if not path:
+            path = "/"
+
+        # Could do syntax based normalization of the URI before
+        # computing the digest. See Section 6.2.2 of Std 66.
+        request_uri = query and "?".join([path, query]) or path
+        defrag_uri = scheme + "://" + authority + request_uri
+
+        return defrag_uri
+
+    @classmethod
+    def cache_url(cls, uri: str) -> str:
+        return cls._urlnorm(uri)
+
+    def parse_cache_control(self, headers: Mapping[str, str]) -> dict[str, int | None]:
+        known_directives = {
+            # https://tools.ietf.org/html/rfc7234#section-5.2
+            "max-age": (int, True),
+            "max-stale": (int, False),
+            "min-fresh": (int, True),
+            "no-cache": (None, False),
+            "no-store": (None, False),
+            "no-transform": (None, False),
+            "only-if-cached": (None, False),
+            "must-revalidate": (None, False),
+            "public": (None, False),
+            "private": (None, False),
+            "proxy-revalidate": (None, False),
+            "s-maxage": (int, True),
+        }
+
+        cc_headers = headers.get("cache-control", headers.get("Cache-Control", ""))
+
+        retval: dict[str, int | None] = {}
+
+        for cc_directive in cc_headers.split(","):
+            if not cc_directive.strip():
+                continue
+
+            parts = cc_directive.split("=", 1)
+            directive = parts[0].strip()
+
+            try:
+                typ, required = known_directives[directive]
+            except KeyError:
+                logger.debug("Ignoring unknown cache-control directive: %s", directive)
+                continue
+
+            if not typ or not required:
+                retval[directive] = None
+            if typ:
+                try:
+                    retval[directive] = typ(parts[1].strip())
+                except IndexError:
+                    if required:
+                        logger.debug(
+                            "Missing value for cache-control " "directive: %s",
+                            directive,
+                        )
+                except ValueError:
+                    logger.debug(
+                        "Invalid value for cache-control directive " "%s, must be %s",
+                        directive,
+                        typ.__name__,
+                    )
+
+        return retval
+
+    def _load_from_cache(self, request: PreparedRequest) -> HTTPResponse | None:
+        """
+        Load a cached response, or return None if it's not available.
+        """
+        # We do not support caching of partial content: so if the request contains a
+        # Range header then we don't want to load anything from the cache.
+        if "Range" in request.headers:
+            return None
+
+        cache_url = request.url
+        assert cache_url is not None
+        cache_data = self.cache.get(cache_url)
+        if cache_data is None:
+            logger.debug("No cache entry available")
+            return None
+
+        if isinstance(self.cache, SeparateBodyBaseCache):
+            body_file = self.cache.get_body(cache_url)
+        else:
+            body_file = None
+
+        result = self.serializer.loads(request, cache_data, body_file)
+        if result is None:
+            logger.warning("Cache entry deserialization failed, entry ignored")
+        return result
+
+    def cached_request(self, request: PreparedRequest) -> HTTPResponse | Literal[False]:
+        """
+        Return a cached response if it exists in the cache, otherwise
+        return False.
+        """
+        assert request.url is not None
+        cache_url = self.cache_url(request.url)
+        logger.debug('Looking up "%s" in the cache', cache_url)
+        cc = self.parse_cache_control(request.headers)
+
+        # Bail out if the request insists on fresh data
+        if "no-cache" in cc:
+            logger.debug('Request header has "no-cache", cache bypassed')
+            return False
+
+        if "max-age" in cc and cc["max-age"] == 0:
+            logger.debug('Request header has "max_age" as 0, cache bypassed')
+            return False
+
+        # Check whether we can load the response from the cache:
+        resp = self._load_from_cache(request)
+        if not resp:
+            return False
+
+        # If we have a cached permanent redirect, return it immediately. We
+        # don't need to test our response for other headers b/c it is
+        # intrinsically "cacheable" as it is Permanent.
+        #
+        # See:
+        #   https://tools.ietf.org/html/rfc7231#section-6.4.2
+        #
+        # Client can try to refresh the value by repeating the request
+        # with cache busting headers as usual (ie no-cache).
+        if int(resp.status) in PERMANENT_REDIRECT_STATUSES:
+            msg = (
+                "Returning cached permanent redirect response "
+                "(ignoring date and etag information)"
+            )
+            logger.debug(msg)
+            return resp
+
+        headers: CaseInsensitiveDict[str] = CaseInsensitiveDict(resp.headers)
+        if not headers or "date" not in headers:
+            if "etag" not in headers:
+                # Without date or etag, the cached response can never be used
+                # and should be deleted.
+                logger.debug("Purging cached response: no date or etag")
+                self.cache.delete(cache_url)
+            logger.debug("Ignoring cached response: no date")
+            return False
+
+        now = time.time()
+        time_tuple = parsedate_tz(headers["date"])
+        assert time_tuple is not None
+        date = calendar.timegm(time_tuple[:6])
+        current_age = max(0, now - date)
+        logger.debug("Current age based on date: %i", current_age)
+
+        # TODO: There is an assumption that the result will be a
+        #       urllib3 response object. This may not be best since we
+        #       could probably avoid instantiating or constructing the
+        #       response until we know we need it.
+        resp_cc = self.parse_cache_control(headers)
+
+        # determine freshness
+        freshness_lifetime = 0
+
+        # Check the max-age pragma in the cache control header
+        max_age = resp_cc.get("max-age")
+        if max_age is not None:
+            freshness_lifetime = max_age
+            logger.debug("Freshness lifetime from max-age: %i", freshness_lifetime)
+
+        # If there isn't a max-age, check for an expires header
+        elif "expires" in headers:
+            expires = parsedate_tz(headers["expires"])
+            if expires is not None:
+                expire_time = calendar.timegm(expires[:6]) - date
+                freshness_lifetime = max(0, expire_time)
+                logger.debug("Freshness lifetime from expires: %i", freshness_lifetime)
+
+        # Determine if we are setting freshness limit in the
+        # request. Note, this overrides what was in the response.
+        max_age = cc.get("max-age")
+        if max_age is not None:
+            freshness_lifetime = max_age
+            logger.debug(
+                "Freshness lifetime from request max-age: %i", freshness_lifetime
+            )
+
+        min_fresh = cc.get("min-fresh")
+        if min_fresh is not None:
+            # adjust our current age by our min fresh
+            current_age += min_fresh
+            logger.debug("Adjusted current age from min-fresh: %i", current_age)
+
+        # Return entry if it is fresh enough
+        if freshness_lifetime > current_age:
+            logger.debug('The response is "fresh", returning cached response')
+            logger.debug("%i > %i", freshness_lifetime, current_age)
+            return resp
+
+        # we're not fresh. If we don't have an Etag, clear it out
+        if "etag" not in headers:
+            logger.debug('The cached response is "stale" with no etag, purging')
+            self.cache.delete(cache_url)
+
+        # return the original handler
+        return False
+
+    def conditional_headers(self, request: PreparedRequest) -> dict[str, str]:
+        resp = self._load_from_cache(request)
+        new_headers = {}
+
+        if resp:
+            headers: CaseInsensitiveDict[str] = CaseInsensitiveDict(resp.headers)
+
+            if "etag" in headers:
+                new_headers["If-None-Match"] = headers["ETag"]
+
+            if "last-modified" in headers:
+                new_headers["If-Modified-Since"] = headers["Last-Modified"]
+
+        return new_headers
+
+    def _cache_set(
+        self,
+        cache_url: str,
+        request: PreparedRequest,
+        response: HTTPResponse,
+        body: bytes | None = None,
+        expires_time: int | None = None,
+    ) -> None:
+        """
+        Store the data in the cache.
+        """
+        if isinstance(self.cache, SeparateBodyBaseCache):
+            # We pass in the body separately; just put a placeholder empty
+            # string in the metadata.
+            self.cache.set(
+                cache_url,
+                self.serializer.dumps(request, response, b""),
+                expires=expires_time,
+            )
+            # body is None can happen when, for example, we're only updating
+            # headers, as is the case in update_cached_response().
+            if body is not None:
+                self.cache.set_body(cache_url, body)
+        else:
+            self.cache.set(
+                cache_url,
+                self.serializer.dumps(request, response, body),
+                expires=expires_time,
+            )
+
+    def cache_response(
+        self,
+        request: PreparedRequest,
+        response: HTTPResponse,
+        body: bytes | None = None,
+        status_codes: Collection[int] | None = None,
+    ) -> None:
+        """
+        Algorithm for caching requests.
+
+        This assumes a requests Response object.
+        """
+        # From httplib2: Don't cache 206's since we aren't going to
+        #                handle byte range requests
+        cacheable_status_codes = status_codes or self.cacheable_status_codes
+        if response.status not in cacheable_status_codes:
+            logger.debug(
+                "Status code %s not in %s", response.status, cacheable_status_codes
+            )
+            return
+
+        response_headers: CaseInsensitiveDict[str] = CaseInsensitiveDict(
+            response.headers
+        )
+
+        if "date" in response_headers:
+            time_tuple = parsedate_tz(response_headers["date"])
+            assert time_tuple is not None
+            date = calendar.timegm(time_tuple[:6])
+        else:
+            date = 0
+
+        # If we've been given a body, our response has a Content-Length, that
+        # Content-Length is valid then we can check to see if the body we've
+        # been given matches the expected size, and if it doesn't we'll just
+        # skip trying to cache it.
+        if (
+            body is not None
+            and "content-length" in response_headers
+            and response_headers["content-length"].isdigit()
+            and int(response_headers["content-length"]) != len(body)
+        ):
+            return
+
+        cc_req = self.parse_cache_control(request.headers)
+        cc = self.parse_cache_control(response_headers)
+
+        assert request.url is not None
+        cache_url = self.cache_url(request.url)
+        logger.debug('Updating cache with response from "%s"', cache_url)
+
+        # Delete it from the cache if we happen to have it stored there
+        no_store = False
+        if "no-store" in cc:
+            no_store = True
+            logger.debug('Response header has "no-store"')
+        if "no-store" in cc_req:
+            no_store = True
+            logger.debug('Request header has "no-store"')
+        if no_store and self.cache.get(cache_url):
+            logger.debug('Purging existing cache entry to honor "no-store"')
+            self.cache.delete(cache_url)
+        if no_store:
+            return
+
+        # https://tools.ietf.org/html/rfc7234#section-4.1:
+        # A Vary header field-value of "*" always fails to match.
+        # Storing such a response leads to a deserialization warning
+        # during cache lookup and is not allowed to ever be served,
+        # so storing it can be avoided.
+        if "*" in response_headers.get("vary", ""):
+            logger.debug('Response header has "Vary: *"')
+            return
+
+        # If we've been given an etag, then keep the response
+        if self.cache_etags and "etag" in response_headers:
+            expires_time = 0
+            if response_headers.get("expires"):
+                expires = parsedate_tz(response_headers["expires"])
+                if expires is not None:
+                    expires_time = calendar.timegm(expires[:6]) - date
+
+            expires_time = max(expires_time, 14 * 86400)
+
+            logger.debug(f"etag object cached for {expires_time} seconds")
+            logger.debug("Caching due to etag")
+            self._cache_set(cache_url, request, response, body, expires_time)
+
+        # Add to the cache any permanent redirects. We do this before looking
+        # that the Date headers.
+        elif int(response.status) in PERMANENT_REDIRECT_STATUSES:
+            logger.debug("Caching permanent redirect")
+            self._cache_set(cache_url, request, response, b"")
+
+        # Add to the cache if the response headers demand it. If there
+        # is no date header then we can't do anything about expiring
+        # the cache.
+        elif "date" in response_headers:
+            time_tuple = parsedate_tz(response_headers["date"])
+            assert time_tuple is not None
+            date = calendar.timegm(time_tuple[:6])
+            # cache when there is a max-age > 0
+            max_age = cc.get("max-age")
+            if max_age is not None and max_age > 0:
+                logger.debug("Caching b/c date exists and max-age > 0")
+                expires_time = max_age
+                self._cache_set(
+                    cache_url,
+                    request,
+                    response,
+                    body,
+                    expires_time,
+                )
+
+            # If the request can expire, it means we should cache it
+            # in the meantime.
+            elif "expires" in response_headers:
+                if response_headers["expires"]:
+                    expires = parsedate_tz(response_headers["expires"])
+                    if expires is not None:
+                        expires_time = calendar.timegm(expires[:6]) - date
+                    else:
+                        expires_time = None
+
+                    logger.debug(
+                        "Caching b/c of expires header. expires in {} seconds".format(
+                            expires_time
+                        )
+                    )
+                    self._cache_set(
+                        cache_url,
+                        request,
+                        response,
+                        body,
+                        expires_time,
+                    )
+
+    def update_cached_response(
+        self, request: PreparedRequest, response: HTTPResponse
+    ) -> HTTPResponse:
+        """On a 304 we will get a new set of headers that we want to
+        update our cached value with, assuming we have one.
+
+        This should only ever be called when we've sent an ETag and
+        gotten a 304 as the response.
+        """
+        assert request.url is not None
+        cache_url = self.cache_url(request.url)
+        cached_response = self._load_from_cache(request)
+
+        if not cached_response:
+            # we didn't have a cached response
+            return response
+
+        # Lets update our headers with the headers from the new request:
+        # http://tools.ietf.org/html/draft-ietf-httpbis-p4-conditional-26#section-4.1
+        #
+        # The server isn't supposed to send headers that would make
+        # the cached body invalid. But... just in case, we'll be sure
+        # to strip out ones we know that might be problmatic due to
+        # typical assumptions.
+        excluded_headers = ["content-length"]
+
+        cached_response.headers.update(
+            {
+                k: v
+                for k, v in response.headers.items()
+                if k.lower() not in excluded_headers
+            }
+        )
+
+        # we want a 200 b/c we have content via the cache
+        cached_response.status = 200
+
+        # update our cache
+        self._cache_set(cache_url, request, cached_response)
+
+        return cached_response
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/filewrapper.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/filewrapper.py
new file mode 100644
index 00000000..37d2fa59
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/filewrapper.py
@@ -0,0 +1,119 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import mmap
+from tempfile import NamedTemporaryFile
+from typing import TYPE_CHECKING, Any, Callable
+
+if TYPE_CHECKING:
+    from http.client import HTTPResponse
+
+
+class CallbackFileWrapper:
+    """
+    Small wrapper around a fp object which will tee everything read into a
+    buffer, and when that file is closed it will execute a callback with the
+    contents of that buffer.
+
+    All attributes are proxied to the underlying file object.
+
+    This class uses members with a double underscore (__) leading prefix so as
+    not to accidentally shadow an attribute.
+
+    The data is stored in a temporary file until it is all available.  As long
+    as the temporary files directory is disk-based (sometimes it's a
+    memory-backed-``tmpfs`` on Linux), data will be unloaded to disk if memory
+    pressure is high.  For small files the disk usually won't be used at all,
+    it'll all be in the filesystem memory cache, so there should be no
+    performance impact.
+    """
+
+    def __init__(
+        self, fp: HTTPResponse, callback: Callable[[bytes], None] | None
+    ) -> None:
+        self.__buf = NamedTemporaryFile("rb+", delete=True)
+        self.__fp = fp
+        self.__callback = callback
+
+    def __getattr__(self, name: str) -> Any:
+        # The vagaries of garbage collection means that self.__fp is
+        # not always set.  By using __getattribute__ and the private
+        # name[0] allows looking up the attribute value and raising an
+        # AttributeError when it doesn't exist. This stop things from
+        # infinitely recursing calls to getattr in the case where
+        # self.__fp hasn't been set.
+        #
+        # [0] https://docs.python.org/2/reference/expressions.html#atom-identifiers
+        fp = self.__getattribute__("_CallbackFileWrapper__fp")
+        return getattr(fp, name)
+
+    def __is_fp_closed(self) -> bool:
+        try:
+            return self.__fp.fp is None
+
+        except AttributeError:
+            pass
+
+        try:
+            closed: bool = self.__fp.closed
+            return closed
+
+        except AttributeError:
+            pass
+
+        # We just don't cache it then.
+        # TODO: Add some logging here...
+        return False
+
+    def _close(self) -> None:
+        if self.__callback:
+            if self.__buf.tell() == 0:
+                # Empty file:
+                result = b""
+            else:
+                # Return the data without actually loading it into memory,
+                # relying on Python's buffer API and mmap(). mmap() just gives
+                # a view directly into the filesystem's memory cache, so it
+                # doesn't result in duplicate memory use.
+                self.__buf.seek(0, 0)
+                result = memoryview(
+                    mmap.mmap(self.__buf.fileno(), 0, access=mmap.ACCESS_READ)
+                )
+            self.__callback(result)
+
+        # We assign this to None here, because otherwise we can get into
+        # really tricky problems where the CPython interpreter dead locks
+        # because the callback is holding a reference to something which
+        # has a __del__ method. Setting this to None breaks the cycle
+        # and allows the garbage collector to do it's thing normally.
+        self.__callback = None
+
+        # Closing the temporary file releases memory and frees disk space.
+        # Important when caching big files.
+        self.__buf.close()
+
+    def read(self, amt: int | None = None) -> bytes:
+        data: bytes = self.__fp.read(amt)
+        if data:
+            # We may be dealing with b'', a sign that things are over:
+            # it's passed e.g. after we've already closed self.__buf.
+            self.__buf.write(data)
+        if self.__is_fp_closed():
+            self._close()
+
+        return data
+
+    def _safe_read(self, amt: int) -> bytes:
+        data: bytes = self.__fp._safe_read(amt)  # type: ignore[attr-defined]
+        if amt == 2 and data == b"\r\n":
+            # urllib executes this read to toss the CRLF at the end
+            # of the chunk.
+            return data
+
+        self.__buf.write(data)
+        if self.__is_fp_closed():
+            self._close()
+
+        return data
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/heuristics.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/heuristics.py
new file mode 100644
index 00000000..b778c4f3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/heuristics.py
@@ -0,0 +1,157 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import calendar
+import time
+from datetime import datetime, timedelta, timezone
+from email.utils import formatdate, parsedate, parsedate_tz
+from typing import TYPE_CHECKING, Any, Mapping
+
+if TYPE_CHECKING:
+    from pip._vendor.urllib3 import HTTPResponse
+
+TIME_FMT = "%a, %d %b %Y %H:%M:%S GMT"
+
+
+def expire_after(delta: timedelta, date: datetime | None = None) -> datetime:
+    date = date or datetime.now(timezone.utc)
+    return date + delta
+
+
+def datetime_to_header(dt: datetime) -> str:
+    return formatdate(calendar.timegm(dt.timetuple()))
+
+
+class BaseHeuristic:
+    def warning(self, response: HTTPResponse) -> str | None:
+        """
+        Return a valid 1xx warning header value describing the cache
+        adjustments.
+
+        The response is provided too allow warnings like 113
+        http://tools.ietf.org/html/rfc7234#section-5.5.4 where we need
+        to explicitly say response is over 24 hours old.
+        """
+        return '110 - "Response is Stale"'
+
+    def update_headers(self, response: HTTPResponse) -> dict[str, str]:
+        """Update the response headers with any new headers.
+
+        NOTE: This SHOULD always include some Warning header to
+              signify that the response was cached by the client, not
+              by way of the provided headers.
+        """
+        return {}
+
+    def apply(self, response: HTTPResponse) -> HTTPResponse:
+        updated_headers = self.update_headers(response)
+
+        if updated_headers:
+            response.headers.update(updated_headers)
+            warning_header_value = self.warning(response)
+            if warning_header_value is not None:
+                response.headers.update({"Warning": warning_header_value})
+
+        return response
+
+
+class OneDayCache(BaseHeuristic):
+    """
+    Cache the response by providing an expires 1 day in the
+    future.
+    """
+
+    def update_headers(self, response: HTTPResponse) -> dict[str, str]:
+        headers = {}
+
+        if "expires" not in response.headers:
+            date = parsedate(response.headers["date"])
+            expires = expire_after(
+                timedelta(days=1),
+                date=datetime(*date[:6], tzinfo=timezone.utc),  # type: ignore[index,misc]
+            )
+            headers["expires"] = datetime_to_header(expires)
+            headers["cache-control"] = "public"
+        return headers
+
+
+class ExpiresAfter(BaseHeuristic):
+    """
+    Cache **all** requests for a defined time period.
+    """
+
+    def __init__(self, **kw: Any) -> None:
+        self.delta = timedelta(**kw)
+
+    def update_headers(self, response: HTTPResponse) -> dict[str, str]:
+        expires = expire_after(self.delta)
+        return {"expires": datetime_to_header(expires), "cache-control": "public"}
+
+    def warning(self, response: HTTPResponse) -> str | None:
+        tmpl = "110 - Automatically cached for %s. Response might be stale"
+        return tmpl % self.delta
+
+
+class LastModified(BaseHeuristic):
+    """
+    If there is no Expires header already, fall back on Last-Modified
+    using the heuristic from
+    http://tools.ietf.org/html/rfc7234#section-4.2.2
+    to calculate a reasonable value.
+
+    Firefox also does something like this per
+    https://developer.mozilla.org/en-US/docs/Web/HTTP/Caching_FAQ
+    http://lxr.mozilla.org/mozilla-release/source/netwerk/protocol/http/nsHttpResponseHead.cpp#397
+    Unlike mozilla we limit this to 24-hr.
+    """
+
+    cacheable_by_default_statuses = {
+        200,
+        203,
+        204,
+        206,
+        300,
+        301,
+        404,
+        405,
+        410,
+        414,
+        501,
+    }
+
+    def update_headers(self, resp: HTTPResponse) -> dict[str, str]:
+        headers: Mapping[str, str] = resp.headers
+
+        if "expires" in headers:
+            return {}
+
+        if "cache-control" in headers and headers["cache-control"] != "public":
+            return {}
+
+        if resp.status not in self.cacheable_by_default_statuses:
+            return {}
+
+        if "date" not in headers or "last-modified" not in headers:
+            return {}
+
+        time_tuple = parsedate_tz(headers["date"])
+        assert time_tuple is not None
+        date = calendar.timegm(time_tuple[:6])
+        last_modified = parsedate(headers["last-modified"])
+        if last_modified is None:
+            return {}
+
+        now = time.time()
+        current_age = max(0, now - date)
+        delta = date - calendar.timegm(last_modified)
+        freshness_lifetime = max(0, min(delta / 10, 24 * 3600))
+        if freshness_lifetime <= current_age:
+            return {}
+
+        expires = date + freshness_lifetime
+        return {"expires": time.strftime(TIME_FMT, time.gmtime(expires))}
+
+    def warning(self, resp: HTTPResponse) -> str | None:
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/serialize.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/serialize.py
new file mode 100644
index 00000000..a49487a1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/serialize.py
@@ -0,0 +1,146 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import io
+from typing import IO, TYPE_CHECKING, Any, Mapping, cast
+
+from pip._vendor import msgpack
+from pip._vendor.requests.structures import CaseInsensitiveDict
+from pip._vendor.urllib3 import HTTPResponse
+
+if TYPE_CHECKING:
+    from pip._vendor.requests import PreparedRequest
+
+
+class Serializer:
+    serde_version = "4"
+
+    def dumps(
+        self,
+        request: PreparedRequest,
+        response: HTTPResponse,
+        body: bytes | None = None,
+    ) -> bytes:
+        response_headers: CaseInsensitiveDict[str] = CaseInsensitiveDict(
+            response.headers
+        )
+
+        if body is None:
+            # When a body isn't passed in, we'll read the response. We
+            # also update the response with a new file handler to be
+            # sure it acts as though it was never read.
+            body = response.read(decode_content=False)
+            response._fp = io.BytesIO(body)  # type: ignore[assignment]
+            response.length_remaining = len(body)
+
+        data = {
+            "response": {
+                "body": body,  # Empty bytestring if body is stored separately
+                "headers": {str(k): str(v) for k, v in response.headers.items()},
+                "status": response.status,
+                "version": response.version,
+                "reason": str(response.reason),
+                "decode_content": response.decode_content,
+            }
+        }
+
+        # Construct our vary headers
+        data["vary"] = {}
+        if "vary" in response_headers:
+            varied_headers = response_headers["vary"].split(",")
+            for header in varied_headers:
+                header = str(header).strip()
+                header_value = request.headers.get(header, None)
+                if header_value is not None:
+                    header_value = str(header_value)
+                data["vary"][header] = header_value
+
+        return b",".join([f"cc={self.serde_version}".encode(), self.serialize(data)])
+
+    def serialize(self, data: dict[str, Any]) -> bytes:
+        return cast(bytes, msgpack.dumps(data, use_bin_type=True))
+
+    def loads(
+        self,
+        request: PreparedRequest,
+        data: bytes,
+        body_file: IO[bytes] | None = None,
+    ) -> HTTPResponse | None:
+        # Short circuit if we've been given an empty set of data
+        if not data:
+            return None
+
+        # Previous versions of this library supported other serialization
+        # formats, but these have all been removed.
+        if not data.startswith(f"cc={self.serde_version},".encode()):
+            return None
+
+        data = data[5:]
+        return self._loads_v4(request, data, body_file)
+
+    def prepare_response(
+        self,
+        request: PreparedRequest,
+        cached: Mapping[str, Any],
+        body_file: IO[bytes] | None = None,
+    ) -> HTTPResponse | None:
+        """Verify our vary headers match and construct a real urllib3
+        HTTPResponse object.
+        """
+        # Special case the '*' Vary value as it means we cannot actually
+        # determine if the cached response is suitable for this request.
+        # This case is also handled in the controller code when creating
+        # a cache entry, but is left here for backwards compatibility.
+        if "*" in cached.get("vary", {}):
+            return None
+
+        # Ensure that the Vary headers for the cached response match our
+        # request
+        for header, value in cached.get("vary", {}).items():
+            if request.headers.get(header, None) != value:
+                return None
+
+        body_raw = cached["response"].pop("body")
+
+        headers: CaseInsensitiveDict[str] = CaseInsensitiveDict(
+            data=cached["response"]["headers"]
+        )
+        if headers.get("transfer-encoding", "") == "chunked":
+            headers.pop("transfer-encoding")
+
+        cached["response"]["headers"] = headers
+
+        try:
+            body: IO[bytes]
+            if body_file is None:
+                body = io.BytesIO(body_raw)
+            else:
+                body = body_file
+        except TypeError:
+            # This can happen if cachecontrol serialized to v1 format (pickle)
+            # using Python 2. A Python 2 str(byte string) will be unpickled as
+            # a Python 3 str (unicode string), which will cause the above to
+            # fail with:
+            #
+            #     TypeError: 'str' does not support the buffer interface
+            body = io.BytesIO(body_raw.encode("utf8"))
+
+        # Discard any `strict` parameter serialized by older version of cachecontrol.
+        cached["response"].pop("strict", None)
+
+        return HTTPResponse(body=body, preload_content=False, **cached["response"])
+
+    def _loads_v4(
+        self,
+        request: PreparedRequest,
+        data: bytes,
+        body_file: IO[bytes] | None = None,
+    ) -> HTTPResponse | None:
+        try:
+            cached = msgpack.loads(data, raw=False)
+        except ValueError:
+            return None
+
+        return self.prepare_response(request, cached, body_file)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/wrapper.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/wrapper.py
new file mode 100644
index 00000000..f618bc36
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/cachecontrol/wrapper.py
@@ -0,0 +1,43 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Collection
+
+from pip._vendor.cachecontrol.adapter import CacheControlAdapter
+from pip._vendor.cachecontrol.cache import DictCache
+
+if TYPE_CHECKING:
+    from pip._vendor import requests
+
+    from pip._vendor.cachecontrol.cache import BaseCache
+    from pip._vendor.cachecontrol.controller import CacheController
+    from pip._vendor.cachecontrol.heuristics import BaseHeuristic
+    from pip._vendor.cachecontrol.serialize import Serializer
+
+
+def CacheControl(
+    sess: requests.Session,
+    cache: BaseCache | None = None,
+    cache_etags: bool = True,
+    serializer: Serializer | None = None,
+    heuristic: BaseHeuristic | None = None,
+    controller_class: type[CacheController] | None = None,
+    adapter_class: type[CacheControlAdapter] | None = None,
+    cacheable_methods: Collection[str] | None = None,
+) -> requests.Session:
+    cache = DictCache() if cache is None else cache
+    adapter_class = adapter_class or CacheControlAdapter
+    adapter = adapter_class(
+        cache,
+        cache_etags=cache_etags,
+        serializer=serializer,
+        heuristic=heuristic,
+        controller_class=controller_class,
+        cacheable_methods=cacheable_methods,
+    )
+    sess.mount("http://", adapter)
+    sess.mount("https://", adapter)
+
+    return sess
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__init__.py
new file mode 100644
index 00000000..f61d77fa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__init__.py
@@ -0,0 +1,4 @@
+from .core import contents, where
+
+__all__ = ["contents", "where"]
+__version__ = "2024.08.30"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__main__.py
new file mode 100644
index 00000000..00376349
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__main__.py
@@ -0,0 +1,12 @@
+import argparse
+
+from pip._vendor.certifi import contents, where
+
+parser = argparse.ArgumentParser()
+parser.add_argument("-c", "--contents", action="store_true")
+args = parser.parse_args()
+
+if args.contents:
+    print(contents())
+else:
+    print(where())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..1d3f451d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..18d7c94c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..35936370
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/cacert.pem b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/cacert.pem
new file mode 100644
index 00000000..3c165a1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/cacert.pem
@@ -0,0 +1,4929 @@
+
+# Issuer: CN=GlobalSign Root CA O=GlobalSign nv-sa OU=Root CA
+# Subject: CN=GlobalSign Root CA O=GlobalSign nv-sa OU=Root CA
+# Label: "GlobalSign Root CA"
+# Serial: 4835703278459707669005204
+# MD5 Fingerprint: 3e:45:52:15:09:51:92:e1:b7:5d:37:9f:b1:87:29:8a
+# SHA1 Fingerprint: b1:bc:96:8b:d4:f4:9d:62:2a:a8:9a:81:f2:15:01:52:a4:1d:82:9c
+# SHA256 Fingerprint: eb:d4:10:40:e4:bb:3e:c7:42:c9:e3:81:d3:1e:f2:a4:1a:48:b6:68:5c:96:e7:ce:f3:c1:df:6c:d4:33:1c:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust.net Certification Authority (2048) O=Entrust.net OU=www.entrust.net/CPS_2048 incorp. by ref. (limits liab.)/(c) 1999 Entrust.net Limited
+# Subject: CN=Entrust.net Certification Authority (2048) O=Entrust.net OU=www.entrust.net/CPS_2048 incorp. by ref. (limits liab.)/(c) 1999 Entrust.net Limited
+# Label: "Entrust.net Premium 2048 Secure Server CA"
+# Serial: 946069240
+# MD5 Fingerprint: ee:29:31:bc:32:7e:9a:e6:e8:b5:f7:51:b4:34:71:90
+# SHA1 Fingerprint: 50:30:06:09:1d:97:d4:f5:ae:39:f7:cb:e7:92:7d:7d:65:2d:34:31
+# SHA256 Fingerprint: 6d:c4:71:72:e0:1c:bc:b0:bf:62:58:0d:89:5f:e2:b8:ac:9a:d4:f8:73:80:1e:0c:10:b9:c8:37:d2:1e:b1:77
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Baltimore CyberTrust Root O=Baltimore OU=CyberTrust
+# Subject: CN=Baltimore CyberTrust Root O=Baltimore OU=CyberTrust
+# Label: "Baltimore CyberTrust Root"
+# Serial: 33554617
+# MD5 Fingerprint: ac:b6:94:a5:9c:17:e0:d7:91:52:9b:b1:97:06:a6:e4
+# SHA1 Fingerprint: d4:de:20:d0:5e:66:fc:53:fe:1a:50:88:2c:78:db:28:52:ca:e4:74
+# SHA256 Fingerprint: 16:af:57:a9:f6:76:b0:ab:12:60:95:aa:5e:ba:de:f2:2a:b3:11:19:d6:44:ac:95:cd:4b:93:db:f3:f2:6a:eb
+-----BEGIN CERTIFICATE-----
+MIIDdzCCAl+gAwIBAgIEAgAAuTANBgkqhkiG9w0BAQUFADBaMQswCQYDVQQGEwJJ
+RTESMBAGA1UEChMJQmFsdGltb3JlMRMwEQYDVQQLEwpDeWJlclRydXN0MSIwIAYD
+VQQDExlCYWx0aW1vcmUgQ3liZXJUcnVzdCBSb290MB4XDTAwMDUxMjE4NDYwMFoX
+DTI1MDUxMjIzNTkwMFowWjELMAkGA1UEBhMCSUUxEjAQBgNVBAoTCUJhbHRpbW9y
+ZTETMBEGA1UECxMKQ3liZXJUcnVzdDEiMCAGA1UEAxMZQmFsdGltb3JlIEN5YmVy
+VHJ1c3QgUm9vdDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAKMEuyKr
+mD1X6CZymrV51Cni4eiVgLGw41uOKymaZN+hXe2wCQVt2yguzmKiYv60iNoS6zjr
+IZ3AQSsBUnuId9Mcj8e6uYi1agnnc+gRQKfRzMpijS3ljwumUNKoUMMo6vWrJYeK
+mpYcqWe4PwzV9/lSEy/CG9VwcPCPwBLKBsua4dnKM3p31vjsufFoREJIE9LAwqSu
+XmD+tqYF/LTdB1kC1FkYmGP1pWPgkAx9XbIGevOF6uvUA65ehD5f/xXtabz5OTZy
+dc93Uk3zyZAsuT3lySNTPx8kmCFcB5kpvcY67Oduhjprl3RjM71oGDHweI12v/ye
+jl0qhqdNkNwnGjkCAwEAAaNFMEMwHQYDVR0OBBYEFOWdWTCCR1jMrPoIVDaGezq1
+BE3wMBIGA1UdEwEB/wQIMAYBAf8CAQMwDgYDVR0PAQH/BAQDAgEGMA0GCSqGSIb3
+DQEBBQUAA4IBAQCFDF2O5G9RaEIFoN27TyclhAO992T9Ldcw46QQF+vaKSm2eT92
+9hkTI7gQCvlYpNRhcL0EYWoSihfVCr3FvDB81ukMJY2GQE/szKN+OMY3EU/t3Wgx
+jkzSswF07r51XgdIGn9w/xZchMB5hbgF/X++ZRGjD8ACtPhSNzkE1akxehi/oCr0
+Epn3o0WC4zxe9Z2etciefC7IpJ5OCBRLbf1wbWsaY71k5h+3zvDyny67G7fyUIhz
+ksLi4xaNmjICq44Y3ekQEe5+NauQrz4wlHrQMz2nZQ/1/I6eYs9HRCwBXbsdtTLS
+R9I4LtD+gdwyah617jzV/OeBHRnDJELqYzmp
+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority O=Entrust, Inc. OU=www.entrust.net/CPS is incorporated by reference/(c) 2006 Entrust, Inc.
+# Subject: CN=Entrust Root Certification Authority O=Entrust, Inc. OU=www.entrust.net/CPS is incorporated by reference/(c) 2006 Entrust, Inc.
+# Label: "Entrust Root Certification Authority"
+# Serial: 1164660820
+# MD5 Fingerprint: d6:a5:c3:ed:5d:dd:3e:00:c1:3d:87:92:1f:1d:3f:e4
+# SHA1 Fingerprint: b3:1e:b1:b7:40:e3:6c:84:02:da:dc:37:d4:4d:f5:d4:67:49:52:f9
+# SHA256 Fingerprint: 73:c1:76:43:4f:1b:c6:d5:ad:f4:5b:0e:76:e7:27:28:7c:8d:e5:76:16:c1:e6:e6:14:1a:2b:2c:bc:7d:8e:4c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AAA Certificate Services O=Comodo CA Limited
+# Subject: CN=AAA Certificate Services O=Comodo CA Limited
+# Label: "Comodo AAA Services root"
+# Serial: 1
+# MD5 Fingerprint: 49:79:04:b0:eb:87:19:ac:47:b0:bc:11:51:9b:74:d0
+# SHA1 Fingerprint: d1:eb:23:a4:6d:17:d6:8f:d9:25:64:c2:f1:f1:60:17:64:d8:e3:49
+# SHA256 Fingerprint: d7:a7:a0:fb:5d:7e:27:31:d7:71:e9:48:4e:bc:de:f7:1d:5f:0c:3e:0a:29:48:78:2b:c8:3e:e0:ea:69:9e:f4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 2 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 2 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 2"
+# Serial: 1289
+# MD5 Fingerprint: 5e:39:7b:dd:f8:ba:ec:82:e9:ac:62:ba:0c:54:00:2b
+# SHA1 Fingerprint: ca:3a:fb:cf:12:40:36:4b:44:b2:16:20:88:80:48:39:19:93:7c:f7
+# SHA256 Fingerprint: 85:a0:dd:7d:d7:20:ad:b7:ff:05:f8:3d:54:2b:20:9d:c7:ff:45:28:f7:d6:77:b1:83:89:fe:a5:e5:c4:9e:86
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 3"
+# Serial: 1478
+# MD5 Fingerprint: 31:85:3c:62:94:97:63:b9:aa:fd:89:4e:af:6f:e0:cf
+# SHA1 Fingerprint: 1f:49:14:f7:d8:74:95:1d:dd:ae:02:c0:be:fd:3a:2d:82:75:51:85
+# SHA256 Fingerprint: 18:f1:fc:7f:20:5d:f8:ad:dd:eb:7f:e0:07:dd:57:e3:af:37:5a:9c:4d:8d:73:54:6b:f4:f1:fe:d1:e1:8d:35
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=XRamp Global Certification Authority O=XRamp Security Services Inc OU=www.xrampsecurity.com
+# Subject: CN=XRamp Global Certification Authority O=XRamp Security Services Inc OU=www.xrampsecurity.com
+# Label: "XRamp Global CA Root"
+# Serial: 107108908803651509692980124233745014957
+# MD5 Fingerprint: a1:0b:44:b3:ca:10:d8:00:6e:9d:0f:d8:0f:92:0a:d1
+# SHA1 Fingerprint: b8:01:86:d1:eb:9c:86:a5:41:04:cf:30:54:f3:4c:52:b7:e5:58:c6
+# SHA256 Fingerprint: ce:cd:dc:90:50:99:d8:da:df:c5:b1:d2:09:b7:37:cb:e2:c1:8c:fb:2c:10:c0:ff:0b:cf:0d:32:86:fc:1a:a2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=The Go Daddy Group, Inc. OU=Go Daddy Class 2 Certification Authority
+# Subject: O=The Go Daddy Group, Inc. OU=Go Daddy Class 2 Certification Authority
+# Label: "Go Daddy Class 2 CA"
+# Serial: 0
+# MD5 Fingerprint: 91:de:06:25:ab:da:fd:32:17:0c:bb:25:17:2a:84:67
+# SHA1 Fingerprint: 27:96:ba:e6:3f:18:01:e2:77:26:1b:a0:d7:77:70:02:8f:20:ee:e4
+# SHA256 Fingerprint: c3:84:6b:f2:4b:9e:93:ca:64:27:4c:0e:c6:7c:1e:cc:5e:02:4f:fc:ac:d2:d7:40:19:35:0e:81:fe:54:6a:e4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=Starfield Technologies, Inc. OU=Starfield Class 2 Certification Authority
+# Subject: O=Starfield Technologies, Inc. OU=Starfield Class 2 Certification Authority
+# Label: "Starfield Class 2 CA"
+# Serial: 0
+# MD5 Fingerprint: 32:4a:4b:bb:c8:63:69:9b:be:74:9a:c6:dd:1d:46:24
+# SHA1 Fingerprint: ad:7e:1c:28:b0:64:ef:8f:60:03:40:20:14:c3:d0:e3:37:0e:b5:8a
+# SHA256 Fingerprint: 14:65:fa:20:53:97:b8:76:fa:a6:f0:a9:95:8e:55:90:e4:0f:cc:7f:aa:4f:b7:c2:c8:67:75:21:fb:5f:b6:58
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root CA"
+# Serial: 17154717934120587862167794914071425081
+# MD5 Fingerprint: 87:ce:0b:7b:2a:0e:49:00:e1:58:71:9b:37:a8:93:72
+# SHA1 Fingerprint: 05:63:b8:63:0d:62:d7:5a:bb:c8:ab:1e:4b:df:b5:a8:99:b2:4d:43
+# SHA256 Fingerprint: 3e:90:99:b5:01:5e:8f:48:6c:00:bc:ea:9d:11:1e:e7:21:fa:ba:35:5a:89:bc:f1:df:69:56:1e:3d:c6:32:5c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root CA"
+# Serial: 10944719598952040374951832963794454346
+# MD5 Fingerprint: 79:e4:a9:84:0d:7d:3a:96:d7:c0:4f:e2:43:4c:89:2e
+# SHA1 Fingerprint: a8:98:5d:3a:65:e5:e5:c4:b2:d7:d6:6d:40:c6:dd:2f:b1:9c:54:36
+# SHA256 Fingerprint: 43:48:a0:e9:44:4c:78:cb:26:5e:05:8d:5e:89:44:b4:d8:4f:96:62:bd:26:db:25:7f:89:34:a4:43:c7:01:61
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert High Assurance EV Root CA O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert High Assurance EV Root CA O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert High Assurance EV Root CA"
+# Serial: 3553400076410547919724730734378100087
+# MD5 Fingerprint: d4:74:de:57:5c:39:b2:d3:9c:85:83:c5:c0:65:49:8a
+# SHA1 Fingerprint: 5f:b7:ee:06:33:e2:59:db:ad:0c:4c:9a:e6:d3:8f:1a:61:c7:dc:25
+# SHA256 Fingerprint: 74:31:e5:f4:c3:c1:ce:46:90:77:4f:0b:61:e0:54:40:88:3b:a9:a0:1e:d0:0b:a6:ab:d7:80:6e:d3:b1:18:cf
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SwissSign Gold CA - G2 O=SwissSign AG
+# Subject: CN=SwissSign Gold CA - G2 O=SwissSign AG
+# Label: "SwissSign Gold CA - G2"
+# Serial: 13492815561806991280
+# MD5 Fingerprint: 24:77:d9:a8:91:d1:3b:fa:88:2d:c2:ff:f8:cd:33:93
+# SHA1 Fingerprint: d8:c5:38:8a:b7:30:1b:1b:6e:d4:7a:e6:45:25:3a:6f:9f:1a:27:61
+# SHA256 Fingerprint: 62:dd:0b:e9:b9:f5:0a:16:3e:a0:f8:e7:5c:05:3b:1e:ca:57:ea:55:c8:68:8f:64:7c:68:81:f2:c8:35:7b:95
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SwissSign Silver CA - G2 O=SwissSign AG
+# Subject: CN=SwissSign Silver CA - G2 O=SwissSign AG
+# Label: "SwissSign Silver CA - G2"
+# Serial: 5700383053117599563
+# MD5 Fingerprint: e0:06:a1:c9:7d:cf:c9:fc:0d:c0:56:75:96:d8:62:13
+# SHA1 Fingerprint: 9b:aa:e5:9f:56:ee:21:cb:43:5a:be:25:93:df:a7:f0:40:d1:1d:cb
+# SHA256 Fingerprint: be:6c:4d:a2:bb:b9:ba:59:b6:f3:93:97:68:37:42:46:c3:c0:05:99:3f:a9:8f:02:0d:1d:ed:be:d4:8a:81:d5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureTrust CA O=SecureTrust Corporation
+# Subject: CN=SecureTrust CA O=SecureTrust Corporation
+# Label: "SecureTrust CA"
+# Serial: 17199774589125277788362757014266862032
+# MD5 Fingerprint: dc:32:c3:a7:6d:25:57:c7:68:09:9d:ea:2d:a9:a2:d1
+# SHA1 Fingerprint: 87:82:c6:c3:04:35:3b:cf:d2:96:92:d2:59:3e:7d:44:d9:34:ff:11
+# SHA256 Fingerprint: f1:c1:b5:0a:e5:a2:0d:d8:03:0e:c9:f6:bc:24:82:3d:d3:67:b5:25:57:59:b4:e7:1b:61:fc:e9:f7:37:5d:73
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=Secure Global CA O=SecureTrust Corporation
+# Subject: CN=Secure Global CA O=SecureTrust Corporation
+# Label: "Secure Global CA"
+# Serial: 9751836167731051554232119481456978597
+# MD5 Fingerprint: cf:f4:27:0d:d4:ed:dc:65:16:49:6d:3d:da:bf:6e:de
+# SHA1 Fingerprint: 3a:44:73:5a:e5:81:90:1f:24:86:61:46:1e:3b:9c:c4:5f:f5:3a:1b
+# SHA256 Fingerprint: 42:00:f5:04:3a:c8:59:0e:bb:52:7d:20:9e:d1:50:30:29:fb:cb:d4:1c:a1:b5:06:ec:27:f1:5a:de:7d:ac:69
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO Certification Authority O=COMODO CA Limited
+# Label: "COMODO Certification Authority"
+# Serial: 104350513648249232941998508985834464573
+# MD5 Fingerprint: 5c:48:dc:f7:42:72:ec:56:94:6d:1c:cc:71:35:80:75
+# SHA1 Fingerprint: 66:31:bf:9e:f7:4f:9e:b6:c9:d5:a6:0c:ba:6a:be:d1:f7:bd:ef:7b
+# SHA256 Fingerprint: 0c:2c:d6:3d:f7:80:6f:a3:99:ed:e8:09:11:6b:57:5b:f8:79:89:f0:65:18:f9:80:8c:86:05:03:17:8b:af:66
+-----BEGIN CERTIFICATE-----
+MIIEHTCCAwWgAwIBAgIQToEtioJl4AsC7j41AkblPTANBgkqhkiG9w0BAQUFADCB
+gTELMAkGA1UEBhMCR0IxGzAZBgNVBAgTEkdyZWF0ZXIgTWFuY2hlc3RlcjEQMA4G
+A1UEBxMHU2FsZm9yZDEaMBgGA1UEChMRQ09NT0RPIENBIExpbWl0ZWQxJzAlBgNV
+BAMTHkNPTU9ETyBDZXJ0aWZpY2F0aW9uIEF1dGhvcml0eTAeFw0wNjEyMDEwMDAw
+MDBaFw0yOTEyMzEyMzU5NTlaMIGBMQswCQYDVQQGEwJHQjEbMBkGA1UECBMSR3Jl
+YXRlciBNYW5jaGVzdGVyMRAwDgYDVQQHEwdTYWxmb3JkMRowGAYDVQQKExFDT01P
+RE8gQ0EgTGltaXRlZDEnMCUGA1UEAxMeQ09NT0RPIENlcnRpZmljYXRpb24gQXV0
+aG9yaXR5MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEA0ECLi3LjkRv3
+UcEbVASY06m/weaKXTuH+7uIzg3jLz8GlvCiKVCZrts7oVewdFFxze1CkU1B/qnI
+2GqGd0S7WWaXUF601CxwRM/aN5VCaTwwxHGzUvAhTaHYujl8HJ6jJJ3ygxaYqhZ8
+Q5sVW7euNJH+1GImGEaaP+vB+fGQV+useg2L23IwambV4EajcNxo2f8ESIl33rXp
++2dtQem8Ob0y2WIC8bGoPW43nOIv4tOiJovGuFVDiOEjPqXSJDlqR6sA1KGzqSX+
+DT+nHbrTUcELpNqsOO9VUCQFZUaTNE8tja3G1CEZ0o7KBWFxB3NH5YoZEr0ETc5O
+nKVIrLsm9wIDAQABo4GOMIGLMB0GA1UdDgQWBBQLWOWLxkwVN6RAqTCpIb5HNlpW
+/zAOBgNVHQ8BAf8EBAMCAQYwDwYDVR0TAQH/BAUwAwEB/zBJBgNVHR8EQjBAMD6g
+PKA6hjhodHRwOi8vY3JsLmNvbW9kb2NhLmNvbS9DT01PRE9DZXJ0aWZpY2F0aW9u
+QXV0aG9yaXR5LmNybDANBgkqhkiG9w0BAQUFAAOCAQEAPpiem/Yb6dc5t3iuHXIY
+SdOH5EOC6z/JqvWote9VfCFSZfnVDeFs9D6Mk3ORLgLETgdxb8CPOGEIqB6BCsAv
+IC9Bi5HcSEW88cbeunZrM8gALTFGTO3nnc+IlP8zwFboJIYmuNg4ON8qa90SzMc/
+RxdMosIGlgnW2/4/PEZB31jiVg88O8EckzXZOFKs7sjsLjBOlDW0JB9LeGna8gI4
+zJVSk/BwJVmcIGfE7vmLV2H0knZ9P4SNVbfo5azV8fUZVqZa+5Acr5Pr5RzUZ5dd
+BA6+C4OmF4O5MBKgxTMVBbkN+8cFduPYSo38NBejxiEovjBFMR7HeL5YYTisO+IB
+ZQ==
+-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO ECC Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO ECC Certification Authority O=COMODO CA Limited
+# Label: "COMODO ECC Certification Authority"
+# Serial: 41578283867086692638256921589707938090
+# MD5 Fingerprint: 7c:62:ff:74:9d:31:53:5e:68:4a:d5:78:aa:1e:bf:23
+# SHA1 Fingerprint: 9f:74:4e:9f:2b:4d:ba:ec:0f:31:2c:50:b6:56:3b:8e:2d:93:c3:11
+# SHA256 Fingerprint: 17:93:92:7a:06:14:54:97:89:ad:ce:2f:8f:34:f7:f0:b6:6d:0f:3a:e3:a3:b8:4d:21:ec:15:db:ba:4f:ad:c7
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certigna O=Dhimyotis
+# Subject: CN=Certigna O=Dhimyotis
+# Label: "Certigna"
+# Serial: 18364802974209362175
+# MD5 Fingerprint: ab:57:a6:5b:7d:42:82:19:b5:d8:58:26:28:5e:fd:ff
+# SHA1 Fingerprint: b1:2e:13:63:45:86:a4:6f:1a:b2:60:68:37:58:2d:c4:ac:fd:94:97
+# SHA256 Fingerprint: e3:b6:a2:db:2e:d7:ce:48:84:2f:7a:c5:32:41:c7:b7:1d:54:14:4b:fb:40:c1:1f:3f:1d:0b:42:f5:ee:a1:2d
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=Chunghwa Telecom Co., Ltd. OU=ePKI Root Certification Authority
+# Subject: O=Chunghwa Telecom Co., Ltd. OU=ePKI Root Certification Authority
+# Label: "ePKI Root Certification Authority"
+# Serial: 28956088682735189655030529057352760477
+# MD5 Fingerprint: 1b:2e:00:ca:26:06:90:3d:ad:fe:6f:15:68:d3:6b:b3
+# SHA1 Fingerprint: 67:65:0d:f1:7e:8e:7e:5b:82:40:a4:f4:56:4b:cf:e2:3d:69:c6:f0
+# SHA256 Fingerprint: c0:a6:f4:dc:63:a2:4b:fd:cf:54:ef:2a:6a:08:2a:0a:72:de:35:80:3e:2f:f5:ff:52:7a:e5:d8:72:06:df:d5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=certSIGN OU=certSIGN ROOT CA
+# Subject: O=certSIGN OU=certSIGN ROOT CA
+# Label: "certSIGN ROOT CA"
+# Serial: 35210227249154
+# MD5 Fingerprint: 18:98:c0:d6:e9:3a:fc:f9:b0:f5:0c:f7:4b:01:44:17
+# SHA1 Fingerprint: fa:b7:ee:36:97:26:62:fb:2d:b0:2a:f6:bf:03:fd:e8:7c:4b:2f:9b
+# SHA256 Fingerprint: ea:a9:62:c4:fa:4a:6b:af:eb:e4:15:19:6d:35:1c:cd:88:8d:4f:53:f3:fa:8a:e6:d7:c4:66:a9:4e:60:42:bb
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny O=NetLock Kft. OU=Tan\xfas\xedtv\xe1nykiad\xf3k (Certification Services)
+# Subject: CN=NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny O=NetLock Kft. OU=Tan\xfas\xedtv\xe1nykiad\xf3k (Certification Services)
+# Label: "NetLock Arany (Class Gold) F\u0151tan\xfas\xedtv\xe1ny"
+# Serial: 80544274841616
+# MD5 Fingerprint: c5:a1:b7:ff:73:dd:d6:d7:34:32:18:df:fc:3c:ad:88
+# SHA1 Fingerprint: 06:08:3f:59:3f:15:a1:04:a0:69:a4:6b:a9:03:d0:06:b7:97:09:91
+# SHA256 Fingerprint: 6c:61:da:c3:a2:de:f0:31:50:6b:e0:36:d2:a6:fe:40:19:94:fb:d1:3d:f9:c8:d4:66:59:92:74:c4:46:ec:98
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign RootCA11 O=Japan Certification Services, Inc.
+# Subject: CN=SecureSign RootCA11 O=Japan Certification Services, Inc.
+# Label: "SecureSign RootCA11"
+# Serial: 1
+# MD5 Fingerprint: b7:52:74:e2:92:b4:80:93:f2:75:e4:cc:d7:f2:ea:26
+# SHA1 Fingerprint: 3b:c4:9f:48:f8:f3:73:a0:9c:1e:bd:f8:5b:b1:c3:65:c7:d8:11:b3
+# SHA256 Fingerprint: bf:0f:ee:fb:9e:3a:58:1a:d5:f9:e9:db:75:89:98:57:43:d2:61:08:5c:4d:31:4f:6f:5d:72:59:aa:42:16:12
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsec e-Szigno Root CA 2009 O=Microsec Ltd.
+# Subject: CN=Microsec e-Szigno Root CA 2009 O=Microsec Ltd.
+# Label: "Microsec e-Szigno Root CA 2009"
+# Serial: 14014712776195784473
+# MD5 Fingerprint: f8:49:f4:03:bc:44:2d:83:be:48:69:7d:29:64:fc:b1
+# SHA1 Fingerprint: 89:df:74:fe:5c:f4:0f:4a:80:f9:e3:37:7d:54:da:91:e1:01:31:8e
+# SHA256 Fingerprint: 3c:5f:81:fe:a5:fa:b8:2c:64:bf:a2:ea:ec:af:cd:e8:e0:77:fc:86:20:a7:ca:e5:37:16:3d:f3:6e:db:f3:78
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R3
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R3
+# Label: "GlobalSign Root CA - R3"
+# Serial: 4835703278459759426209954
+# MD5 Fingerprint: c5:df:b8:49:ca:05:13:55:ee:2d:ba:1a:c3:3e:b0:28
+# SHA1 Fingerprint: d6:9b:56:11:48:f0:1c:77:c5:45:78:c1:09:26:df:5b:85:69:76:ad
+# SHA256 Fingerprint: cb:b5:22:d7:b7:f1:27:ad:6a:01:13:86:5b:df:1c:d4:10:2e:7d:07:59:af:63:5a:7c:f4:72:0d:c9:63:c5:3b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Izenpe.com O=IZENPE S.A.
+# Subject: CN=Izenpe.com O=IZENPE S.A.
+# Label: "Izenpe.com"
+# Serial: 917563065490389241595536686991402621
+# MD5 Fingerprint: a6:b0:cd:85:80:da:5c:50:34:a3:39:90:2f:55:67:73
+# SHA1 Fingerprint: 2f:78:3d:25:52:18:a7:4a:65:39:71:b5:2c:a2:9c:45:15:6f:e9:19
+# SHA256 Fingerprint: 25:30:cc:8e:98:32:15:02:ba:d9:6f:9b:1f:ba:1b:09:9e:2d:29:9e:0f:45:48:bb:91:4f:36:3b:c0:d4:53:1f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Go Daddy Root Certificate Authority - G2 O=GoDaddy.com, Inc.
+# Subject: CN=Go Daddy Root Certificate Authority - G2 O=GoDaddy.com, Inc.
+# Label: "Go Daddy Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: 80:3a:bc:22:c1:e6:fb:8d:9b:3b:27:4a:32:1b:9a:01
+# SHA1 Fingerprint: 47:be:ab:c9:22:ea:e8:0e:78:78:34:62:a7:9f:45:c2:54:fd:e6:8b
+# SHA256 Fingerprint: 45:14:0b:32:47:eb:9c:c8:c5:b4:f0:d7:b5:30:91:f7:32:92:08:9e:6e:5a:63:e2:74:9d:d3:ac:a9:19:8e:da
+-----BEGIN CERTIFICATE-----
+MIIDxTCCAq2gAwIBAgIBADANBgkqhkiG9w0BAQsFADCBgzELMAkGA1UEBhMCVVMx
+EDAOBgNVBAgTB0FyaXpvbmExEzARBgNVBAcTClNjb3R0c2RhbGUxGjAYBgNVBAoT
+EUdvRGFkZHkuY29tLCBJbmMuMTEwLwYDVQQDEyhHbyBEYWRkeSBSb290IENlcnRp
+ZmljYXRlIEF1dGhvcml0eSAtIEcyMB4XDTA5MDkwMTAwMDAwMFoXDTM3MTIzMTIz
+NTk1OVowgYMxCzAJBgNVBAYTAlVTMRAwDgYDVQQIEwdBcml6b25hMRMwEQYDVQQH
+EwpTY290dHNkYWxlMRowGAYDVQQKExFHb0RhZGR5LmNvbSwgSW5jLjExMC8GA1UE
+AxMoR28gRGFkZHkgUm9vdCBDZXJ0aWZpY2F0ZSBBdXRob3JpdHkgLSBHMjCCASIw
+DQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAL9xYgjx+lk09xvJGKP3gElY6SKD
+E6bFIEMBO4Tx5oVJnyfq9oQbTqC023CYxzIBsQU+B07u9PpPL1kwIuerGVZr4oAH
+/PMWdYA5UXvl+TW2dE6pjYIT5LY/qQOD+qK+ihVqf94Lw7YZFAXK6sOoBJQ7Rnwy
+DfMAZiLIjWltNowRGLfTshxgtDj6AozO091GB94KPutdfMh8+7ArU6SSYmlRJQVh
+GkSBjCypQ5Yj36w6gZoOKcUcqeldHraenjAKOc7xiID7S13MMuyFYkMlNAJWJwGR
+tDtwKj9useiciAF9n9T521NtYJ2/LOdYq7hfRvzOxBsDPAnrSTFcaUaz4EcCAwEA
+AaNCMEAwDwYDVR0TAQH/BAUwAwEB/zAOBgNVHQ8BAf8EBAMCAQYwHQYDVR0OBBYE
+FDqahQcQZyi27/a9BUFuIMGU2g/eMA0GCSqGSIb3DQEBCwUAA4IBAQCZ21151fmX
+WWcDYfF+OwYxdS2hII5PZYe096acvNjpL9DbWu7PdIxztDhC2gV7+AJ1uP2lsdeu
+9tfeE8tTEH6KRtGX+rcuKxGrkLAngPnon1rpN5+r5N9ss4UXnT3ZJE95kTXWXwTr
+gIOrmgIttRD02JDHBHNA7XIloKmf7J6raBKZV8aPEjoJpL1E/QYVN8Gb5DKj7Tjo
+2GTzLH4U/ALqn83/B2gX2yKQOC16jdFU8WnjXzPKej17CuPKf1855eJ1usV2GDPO
+LPAvTK33sefOT6jEm0pUBsV/fdUID+Ic/n4XuKxe9tQWskMJDE32p2u0mYRlynqI
+4uJEvlz36hz1
+-----END CERTIFICATE-----
+
+# Issuer: CN=Starfield Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Subject: CN=Starfield Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Label: "Starfield Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: d6:39:81:c6:52:7e:96:69:fc:fc:ca:66:ed:05:f2:96
+# SHA1 Fingerprint: b5:1c:06:7c:ee:2b:0c:3d:f8:55:ab:2d:92:f4:fe:39:d4:e7:0f:0e
+# SHA256 Fingerprint: 2c:e1:cb:0b:f9:d2:f9:e1:02:99:3f:be:21:51:52:c3:b2:dd:0c:ab:de:1c:68:e5:31:9b:83:91:54:db:b7:f5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Starfield Services Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Subject: CN=Starfield Services Root Certificate Authority - G2 O=Starfield Technologies, Inc.
+# Label: "Starfield Services Root Certificate Authority - G2"
+# Serial: 0
+# MD5 Fingerprint: 17:35:74:af:7b:61:1c:eb:f4:f9:3c:e2:ee:40:f9:a2
+# SHA1 Fingerprint: 92:5a:8f:8d:2c:6d:04:e0:66:5f:59:6a:ff:22:d8:63:e8:25:6f:3f
+# SHA256 Fingerprint: 56:8d:69:05:a2:c8:87:08:a4:b3:02:51:90:ed:cf:ed:b1:97:4a:60:6a:13:c6:e5:29:0f:cb:2a:e6:3e:da:b5
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Commercial O=AffirmTrust
+# Subject: CN=AffirmTrust Commercial O=AffirmTrust
+# Label: "AffirmTrust Commercial"
+# Serial: 8608355977964138876
+# MD5 Fingerprint: 82:92:ba:5b:ef:cd:8a:6f:a6:3d:55:f9:84:f6:d6:b7
+# SHA1 Fingerprint: f9:b5:b6:32:45:5f:9c:be:ec:57:5f:80:dc:e9:6e:2c:c7:b2:78:b7
+# SHA256 Fingerprint: 03:76:ab:1d:54:c5:f9:80:3c:e4:b2:e2:01:a0:ee:7e:ef:7b:57:b6:36:e8:a9:3c:9b:8d:48:60:c9:6f:5f:a7
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Networking O=AffirmTrust
+# Subject: CN=AffirmTrust Networking O=AffirmTrust
+# Label: "AffirmTrust Networking"
+# Serial: 8957382827206547757
+# MD5 Fingerprint: 42:65:ca:be:01:9a:9a:4c:a9:8c:41:49:cd:c0:d5:7f
+# SHA1 Fingerprint: 29:36:21:02:8b:20:ed:02:f5:66:c5:32:d1:d6:ed:90:9f:45:00:2f
+# SHA256 Fingerprint: 0a:81:ec:5a:92:97:77:f1:45:90:4a:f3:8d:5d:50:9f:66:b5:e2:c5:8f:cd:b5:31:05:8b:0e:17:f3:f0:b4:1b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Premium O=AffirmTrust
+# Subject: CN=AffirmTrust Premium O=AffirmTrust
+# Label: "AffirmTrust Premium"
+# Serial: 7893706540734352110
+# MD5 Fingerprint: c4:5d:0e:48:b6:ac:28:30:4e:0a:bc:f9:38:16:87:57
+# SHA1 Fingerprint: d8:a6:33:2c:e0:03:6f:b1:85:f6:63:4f:7d:6a:06:65:26:32:28:27
+# SHA256 Fingerprint: 70:a7:3f:7f:37:6b:60:07:42:48:90:45:34:b1:14:82:d5:bf:0e:69:8e:cc:49:8d:f5:25:77:eb:f2:e9:3b:9a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AffirmTrust Premium ECC O=AffirmTrust
+# Subject: CN=AffirmTrust Premium ECC O=AffirmTrust
+# Label: "AffirmTrust Premium ECC"
+# Serial: 8401224907861490260
+# MD5 Fingerprint: 64:b0:09:55:cf:b1:d5:99:e2:be:13:ab:a6:5d:ea:4d
+# SHA1 Fingerprint: b8:23:6b:00:2f:1d:16:86:53:01:55:6c:11:a4:37:ca:eb:ff:c3:bb
+# SHA256 Fingerprint: bd:71:fd:f6:da:97:e4:cf:62:d1:64:7a:dd:25:81:b0:7d:79:ad:f8:39:7e:b4:ec:ba:9c:5e:84:88:82:14:23
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Network CA O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Network CA O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Network CA"
+# Serial: 279744
+# MD5 Fingerprint: d5:e9:81:40:c5:18:69:fc:46:2c:89:75:62:0f:aa:78
+# SHA1 Fingerprint: 07:e0:32:e0:20:b7:2c:3f:19:2f:06:28:a2:59:3a:19:a7:0f:06:9e
+# SHA256 Fingerprint: 5c:58:46:8d:55:f5:8e:49:7e:74:39:82:d2:b5:00:10:b6:d1:65:37:4a:cf:83:a7:d4:a3:2d:b7:68:c4:40:8e
+-----BEGIN CERTIFICATE-----
+MIIDuzCCAqOgAwIBAgIDBETAMA0GCSqGSIb3DQEBBQUAMH4xCzAJBgNVBAYTAlBM
+MSIwIAYDVQQKExlVbml6ZXRvIFRlY2hub2xvZ2llcyBTLkEuMScwJQYDVQQLEx5D
+ZXJ0dW0gQ2VydGlmaWNhdGlvbiBBdXRob3JpdHkxIjAgBgNVBAMTGUNlcnR1bSBU
+cnVzdGVkIE5ldHdvcmsgQ0EwHhcNMDgxMDIyMTIwNzM3WhcNMjkxMjMxMTIwNzM3
+WjB+MQswCQYDVQQGEwJQTDEiMCAGA1UEChMZVW5pemV0byBUZWNobm9sb2dpZXMg
+Uy5BLjEnMCUGA1UECxMeQ2VydHVtIENlcnRpZmljYXRpb24gQXV0aG9yaXR5MSIw
+IAYDVQQDExlDZXJ0dW0gVHJ1c3RlZCBOZXR3b3JrIENBMIIBIjANBgkqhkiG9w0B
+AQEFAAOCAQ8AMIIBCgKCAQEA4/t9o3K6wvDJFIf1awFO4W5AB7ptJ11/91sts1rH
+UV+rpDKmYYe2bg+G0jACl/jXaVehGDldamR5xgFZrDwxSjh80gTSSyjoIF87B6LM
+TXPb865Px1bVWqeWifrzq2jUI4ZZJ88JJ7ysbnKDHDBy3+Ci6dLhdHUZvSqeexVU
+BBvXQzmtVSjF4hq79MDkrjhJM8x2hZ85RdKknvISjFH4fOQtf/WsX+sWn7Et0brM
+kUJ3TCXJkDhv2/DM+44el1k+1WBO5gUo7Ul5E0u6SNsv+XLTOcr+H9g0cvW0QM8x
+AcPs3hEtF10fuFDRXhmnad4HMyjKUJX5p1TLVIZQRan5SQIDAQABo0IwQDAPBgNV
+HRMBAf8EBTADAQH/MB0GA1UdDgQWBBQIds3LB/8k9sXN7buQvOKEN0Z19zAOBgNV
+HQ8BAf8EBAMCAQYwDQYJKoZIhvcNAQEFBQADggEBAKaorSLOAT2mo/9i0Eidi15y
+sHhE49wcrwn9I0j6vSrEuVUEtRCjjSfeC4Jj0O7eDDd5QVsisrCaQVymcODU0HfL
+I9MA4GxWL+FpDQ3Zqr8hgVDZBqWo/5U30Kr+4rP1mS1FhIrlQgnXdAIv94nYmem8
+J9RHjboNRhx3zxSkHLmkMcScKHQDNP8zGSal6Q10tz6XxnboJ5ajZt3hrvJBW8qY
+VoNzcOSGGtIxQbovvi0TWnZvTuhOgQ4/WwMioBK+ZlgRSssDxLQqKi2WF+A5VLxI
+03YnnZotBqbJ7DnSq9ufmgsnAjUpsUCV5/nonFWIGUbWtzT1fs45mtk48VH3Tyw=
+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA Root Certification Authority O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA Root Certification Authority O=TAIWAN-CA OU=Root CA
+# Label: "TWCA Root Certification Authority"
+# Serial: 1
+# MD5 Fingerprint: aa:08:8f:f6:f9:7b:b7:f2:b1:a7:1e:9b:ea:ea:bd:79
+# SHA1 Fingerprint: cf:9e:87:6d:d3:eb:fc:42:26:97:a3:b5:a3:7a:a0:76:a9:06:23:48
+# SHA256 Fingerprint: bf:d8:8f:e1:10:1c:41:ae:3e:80:1b:f8:be:56:35:0e:e9:ba:d1:a6:b9:bd:51:5e:dc:5c:6d:5b:87:11:ac:44
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=SECOM Trust Systems CO.,LTD. OU=Security Communication RootCA2
+# Subject: O=SECOM Trust Systems CO.,LTD. OU=Security Communication RootCA2
+# Label: "Security Communication RootCA2"
+# Serial: 0
+# MD5 Fingerprint: 6c:39:7d:a4:0e:55:59:b2:3f:d6:41:b1:12:50:de:43
+# SHA1 Fingerprint: 5f:3b:8c:f2:f8:10:b3:7d:78:b4:ce:ec:19:19:c3:73:34:b9:c7:74
+# SHA256 Fingerprint: 51:3b:2c:ec:b8:10:d4:cd:e5:dd:85:39:1a:df:c6:c2:dd:60:d8:7b:b7:36:d2:b5:21:48:4a:a4:7a:0e:be:f6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Actalis Authentication Root CA O=Actalis S.p.A./03358520967
+# Subject: CN=Actalis Authentication Root CA O=Actalis S.p.A./03358520967
+# Label: "Actalis Authentication Root CA"
+# Serial: 6271844772424770508
+# MD5 Fingerprint: 69:c1:0d:4f:07:a3:1b:c3:fe:56:3d:04:bc:11:f6:a6
+# SHA1 Fingerprint: f3:73:b3:87:06:5a:28:84:8a:f2:f3:4a:ce:19:2b:dd:c7:8e:9c:ac
+# SHA256 Fingerprint: 55:92:60:84:ec:96:3a:64:b9:6e:2a:be:01:ce:0b:a8:6a:64:fb:fe:bc:c7:aa:b5:af:c1:55:b3:7f:d7:60:66
+-----BEGIN CERTIFICATE-----
+MIIFuzCCA6OgAwIBAgIIVwoRl0LE48wwDQYJKoZIhvcNAQELBQAwazELMAkGA1UE
+BhMCSVQxDjAMBgNVBAcMBU1pbGFuMSMwIQYDVQQKDBpBY3RhbGlzIFMucC5BLi8w
+MzM1ODUyMDk2NzEnMCUGA1UEAwweQWN0YWxpcyBBdXRoZW50aWNhdGlvbiBSb290
+IENBMB4XDTExMDkyMjExMjIwMloXDTMwMDkyMjExMjIwMlowazELMAkGA1UEBhMC
+SVQxDjAMBgNVBAcMBU1pbGFuMSMwIQYDVQQKDBpBY3RhbGlzIFMucC5BLi8wMzM1
+ODUyMDk2NzEnMCUGA1UEAwweQWN0YWxpcyBBdXRoZW50aWNhdGlvbiBSb290IENB
+MIICIjANBgkqhkiG9w0BAQEFAAOCAg8AMIICCgKCAgEAp8bEpSmkLO/lGMWwUKNv
+UTufClrJwkg4CsIcoBh/kbWHuUA/3R1oHwiD1S0eiKD4j1aPbZkCkpAW1V8IbInX
+4ay8IMKx4INRimlNAJZaby/ARH6jDuSRzVju3PvHHkVH3Se5CAGfpiEd9UEtL0z9
+KK3giq0itFZljoZUj5NDKd45RnijMCO6zfB9E1fAXdKDa0hMxKufgFpbOr3JpyI/
+gCczWw63igxdBzcIy2zSekciRDXFzMwujt0q7bd9Zg1fYVEiVRvjRuPjPdA1Yprb
+rxTIW6HMiRvhMCb8oJsfgadHHwTrozmSBp+Z07/T6k9QnBn+locePGX2oxgkg4YQ
+51Q+qDp2JE+BIcXjDwL4k5RHILv+1A7TaLndxHqEguNTVHnd25zS8gebLra8Pu2F
+be8lEfKXGkJh90qX6IuxEAf6ZYGyojnP9zz/GPvG8VqLWeICrHuS0E4UT1lF9gxe
+KF+w6D9Fz8+vm2/7hNN3WpVvrJSEnu68wEqPSpP4RCHiMUVhUE4Q2OM1fEwZtN4F
+v6MGn8i1zeQf1xcGDXqVdFUNaBr8EBtiZJ1t4JWgw5QHVw0U5r0F+7if5t+L4sbn
+fpb2U8WANFAoWPASUHEXMLrmeGO89LKtmyuy/uE5jF66CyCU3nuDuP/jVo23Eek7
+jPKxwV2dpAtMK9myGPW1n0sCAwEAAaNjMGEwHQYDVR0OBBYEFFLYiDrIn3hm7Ynz
+ezhwlMkCAjbQMA8GA1UdEwEB/wQFMAMBAf8wHwYDVR0jBBgwFoAUUtiIOsifeGbt
+ifN7OHCUyQICNtAwDgYDVR0PAQH/BAQDAgEGMA0GCSqGSIb3DQEBCwUAA4ICAQAL
+e3KHwGCmSUyIWOYdiPcUZEim2FgKDk8TNd81HdTtBjHIgT5q1d07GjLukD0R0i70
+jsNjLiNmsGe+b7bAEzlgqqI0JZN1Ut6nna0Oh4lScWoWPBkdg/iaKWW+9D+a2fDz
+WochcYBNy+A4mz+7+uAwTc+G02UQGRjRlwKxK3JCaKygvU5a2hi/a5iB0P2avl4V
+SM0RFbnAKVy06Ij3Pjaut2L9HmLecHgQHEhb2rykOLpn7VU+Xlff1ANATIGk0k9j
+pwlCCRT8AKnCgHNPLsBA2RF7SOp6AsDT6ygBJlh0wcBzIm2Tlf05fbsq4/aC4yyX
+X04fkZT6/iyj2HYauE2yOE+b+h1IYHkm4vP9qdCa6HCPSXrW5b0KDtst842/6+Ok
+fcvHlXHo2qN8xcL4dJIEG4aspCJTQLas/kx2z/uUMsA1n3Y/buWQbqCmJqK4LL7R
+K4X9p2jIugErsWx0Hbhzlefut8cl8ABMALJ+tguLHPPAUJ4lueAI3jZm/zel0btU
+ZCzJJ7VLkn5l/9Mt4blOvH+kQSGQQXemOR/qnuOf0GZvBeyqdn6/axag67XH/JJU
+LysRJyU3eExRarDzzFhdFPFqSBX/wge2sY0PjlxQRrM9vwGYT7JZVEc+NHt4bVaT
+LnPqZih4zR0Uv6CPLy64Lo7yFIrM6bV8+2ydDKXhlg==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Buypass Class 2 Root CA O=Buypass AS-983163327
+# Subject: CN=Buypass Class 2 Root CA O=Buypass AS-983163327
+# Label: "Buypass Class 2 Root CA"
+# Serial: 2
+# MD5 Fingerprint: 46:a7:d2:fe:45:fb:64:5a:a8:59:90:9b:78:44:9b:29
+# SHA1 Fingerprint: 49:0a:75:74:de:87:0a:47:fe:58:ee:f6:c7:6b:eb:c6:0b:12:40:99
+# SHA256 Fingerprint: 9a:11:40:25:19:7c:5b:b9:5d:94:e6:3d:55:cd:43:79:08:47:b6:46:b2:3c:df:11:ad:a4:a0:0e:ff:15:fb:48
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Buypass Class 3 Root CA O=Buypass AS-983163327
+# Subject: CN=Buypass Class 3 Root CA O=Buypass AS-983163327
+# Label: "Buypass Class 3 Root CA"
+# Serial: 2
+# MD5 Fingerprint: 3d:3b:18:9e:2c:64:5a:e8:d5:88:ce:0e:f9:37:c2:ec
+# SHA1 Fingerprint: da:fa:f7:fa:66:84:ec:06:8f:14:50:bd:c7:c2:81:a5:bc:a9:64:57
+# SHA256 Fingerprint: ed:f7:eb:bc:a2:7a:2a:38:4d:38:7b:7d:40:10:c6:66:e2:ed:b4:84:3e:4c:29:b4:ae:1d:5b:93:32:e6:b2:4d
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=T-TeleSec GlobalRoot Class 3 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Subject: CN=T-TeleSec GlobalRoot Class 3 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Label: "T-TeleSec GlobalRoot Class 3"
+# Serial: 1
+# MD5 Fingerprint: ca:fb:40:a8:4e:39:92:8a:1d:fe:8e:2f:c4:27:ea:ef
+# SHA1 Fingerprint: 55:a6:72:3e:cb:f2:ec:cd:c3:23:74:70:19:9d:2a:be:11:e3:81:d1
+# SHA256 Fingerprint: fd:73:da:d3:1c:64:4f:f1:b4:3b:ef:0c:cd:da:96:71:0b:9c:d9:87:5e:ca:7e:31:70:7a:f3:e9:6d:52:2b:bd
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST Root Class 3 CA 2 2009 O=D-Trust GmbH
+# Subject: CN=D-TRUST Root Class 3 CA 2 2009 O=D-Trust GmbH
+# Label: "D-TRUST Root Class 3 CA 2 2009"
+# Serial: 623603
+# MD5 Fingerprint: cd:e0:25:69:8d:47:ac:9c:89:35:90:f7:fd:51:3d:2f
+# SHA1 Fingerprint: 58:e8:ab:b0:36:15:33:fb:80:f7:9b:1b:6d:29:d3:ff:8d:5f:00:f0
+# SHA256 Fingerprint: 49:e7:a4:42:ac:f0:ea:62:87:05:00:54:b5:25:64:b6:50:e4:f4:9e:42:e3:48:d6:aa:38:e0:39:e9:57:b1:c1
+-----BEGIN CERTIFICATE-----
+MIIEMzCCAxugAwIBAgIDCYPzMA0GCSqGSIb3DQEBCwUAME0xCzAJBgNVBAYTAkRF
+MRUwEwYDVQQKDAxELVRydXN0IEdtYkgxJzAlBgNVBAMMHkQtVFJVU1QgUm9vdCBD
+bGFzcyAzIENBIDIgMjAwOTAeFw0wOTExMDUwODM1NThaFw0yOTExMDUwODM1NTha
+ME0xCzAJBgNVBAYTAkRFMRUwEwYDVQQKDAxELVRydXN0IEdtYkgxJzAlBgNVBAMM
+HkQtVFJVU1QgUm9vdCBDbGFzcyAzIENBIDIgMjAwOTCCASIwDQYJKoZIhvcNAQEB
+BQADggEPADCCAQoCggEBANOySs96R+91myP6Oi/WUEWJNTrGa9v+2wBoqOADER03
+UAifTUpolDWzU9GUY6cgVq/eUXjsKj3zSEhQPgrfRlWLJ23DEE0NkVJD2IfgXU42
+tSHKXzlABF9bfsyjxiupQB7ZNoTWSPOSHjRGICTBpFGOShrvUD9pXRl/RcPHAY9R
+ySPocq60vFYJfxLLHLGvKZAKyVXMD9O0Gu1HNVpK7ZxzBCHQqr0ME7UAyiZsxGsM
+lFqVlNpQmvH/pStmMaTJOKDfHR+4CS7zp+hnUquVH+BGPtikw8paxTGA6Eian5Rp
+/hnd2HN8gcqW3o7tszIFZYQ05ub9VxC1X3a/L7AQDcUCAwEAAaOCARowggEWMA8G
+A1UdEwEB/wQFMAMBAf8wHQYDVR0OBBYEFP3aFMSfMN4hvR5COfyrYyNJ4PGEMA4G
+A1UdDwEB/wQEAwIBBjCB0wYDVR0fBIHLMIHIMIGAoH6gfIZ6bGRhcDovL2RpcmVj
+dG9yeS5kLXRydXN0Lm5ldC9DTj1ELVRSVVNUJTIwUm9vdCUyMENsYXNzJTIwMyUy
+MENBJTIwMiUyMDIwMDksTz1ELVRydXN0JTIwR21iSCxDPURFP2NlcnRpZmljYXRl
+cmV2b2NhdGlvbmxpc3QwQ6BBoD+GPWh0dHA6Ly93d3cuZC10cnVzdC5uZXQvY3Js
+L2QtdHJ1c3Rfcm9vdF9jbGFzc18zX2NhXzJfMjAwOS5jcmwwDQYJKoZIhvcNAQEL
+BQADggEBAH+X2zDI36ScfSF6gHDOFBJpiBSVYEQBrLLpME+bUMJm2H6NMLVwMeni
+acfzcNsgFYbQDfC+rAF1hM5+n02/t2A7nPPKHeJeaNijnZflQGDSNiH+0LS4F9p0
+o3/U37CYAqxva2ssJSRyoWXuJVrl5jLn8t+rSfrzkGkj2wTZ51xY/GXUl77M/C4K
+zCUqNQT4YJEVdT1B/yMfGchs64JTBKbkTCJNjYy6zltz7GRUUG3RnFX7acM2w4y8
+PIWmawomDeCTmGCufsYkl4phX5GOZpIJhzbNi5stPvZR1FDUWSi9g/LMKHtThm3Y
+Johw1+qRzT65ysCQblrGXnRl11z+o+I=
+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST Root Class 3 CA 2 EV 2009 O=D-Trust GmbH
+# Subject: CN=D-TRUST Root Class 3 CA 2 EV 2009 O=D-Trust GmbH
+# Label: "D-TRUST Root Class 3 CA 2 EV 2009"
+# Serial: 623604
+# MD5 Fingerprint: aa:c6:43:2c:5e:2d:cd:c4:34:c0:50:4f:11:02:4f:b6
+# SHA1 Fingerprint: 96:c9:1b:0b:95:b4:10:98:42:fa:d0:d8:22:79:fe:60:fa:b9:16:83
+# SHA256 Fingerprint: ee:c5:49:6b:98:8c:e9:86:25:b9:34:09:2e:ec:29:08:be:d0:b0:f3:16:c2:d4:73:0c:84:ea:f1:f3:d3:48:81
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CA Disig Root R2 O=Disig a.s.
+# Subject: CN=CA Disig Root R2 O=Disig a.s.
+# Label: "CA Disig Root R2"
+# Serial: 10572350602393338211
+# MD5 Fingerprint: 26:01:fb:d8:27:a7:17:9a:45:54:38:1a:43:01:3b:03
+# SHA1 Fingerprint: b5:61:eb:ea:a4:de:e4:25:4b:69:1a:98:a5:57:47:c2:34:c7:d9:71
+# SHA256 Fingerprint: e2:3d:4a:03:6d:7b:70:e9:f5:95:b1:42:20:79:d2:b9:1e:df:bb:1f:b6:51:a0:63:3e:aa:8a:9d:c5:f8:07:03
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=ACCVRAIZ1 O=ACCV OU=PKIACCV
+# Subject: CN=ACCVRAIZ1 O=ACCV OU=PKIACCV
+# Label: "ACCVRAIZ1"
+# Serial: 6828503384748696800
+# MD5 Fingerprint: d0:a0:5a:ee:05:b6:09:94:21:a1:7d:f1:b2:29:82:02
+# SHA1 Fingerprint: 93:05:7a:88:15:c6:4f:ce:88:2f:fa:91:16:52:28:78:bc:53:64:17
+# SHA256 Fingerprint: 9a:6e:c0:12:e1:a7:da:9d:be:34:19:4d:47:8a:d7:c0:db:18:22:fb:07:1d:f1:29:81:49:6e:d1:04:38:41:13
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA Global Root CA O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA Global Root CA O=TAIWAN-CA OU=Root CA
+# Label: "TWCA Global Root CA"
+# Serial: 3262
+# MD5 Fingerprint: f9:03:7e:cf:e6:9e:3c:73:7a:2a:90:07:69:ff:2b:96
+# SHA1 Fingerprint: 9c:bb:48:53:f6:a4:f6:d3:52:a4:e8:32:52:55:60:13:f5:ad:af:65
+# SHA256 Fingerprint: 59:76:90:07:f7:68:5d:0f:cd:50:87:2f:9f:95:d5:75:5a:5b:2b:45:7d:81:f3:69:2b:61:0a:98:67:2f:0e:1b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TeliaSonera Root CA v1 O=TeliaSonera
+# Subject: CN=TeliaSonera Root CA v1 O=TeliaSonera
+# Label: "TeliaSonera Root CA v1"
+# Serial: 199041966741090107964904287217786801558
+# MD5 Fingerprint: 37:41:49:1b:18:56:9a:26:f5:ad:c2:66:fb:40:a5:4c
+# SHA1 Fingerprint: 43:13:bb:96:f1:d5:86:9b:c1:4e:6a:92:f6:cf:f6:34:69:87:82:37
+# SHA256 Fingerprint: dd:69:36:fe:21:f8:f0:77:c1:23:a1:a5:21:c1:22:24:f7:22:55:b7:3e:03:a7:26:06:93:e8:a2:4b:0f:a3:89
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=T-TeleSec GlobalRoot Class 2 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Subject: CN=T-TeleSec GlobalRoot Class 2 O=T-Systems Enterprise Services GmbH OU=T-Systems Trust Center
+# Label: "T-TeleSec GlobalRoot Class 2"
+# Serial: 1
+# MD5 Fingerprint: 2b:9b:9e:e4:7b:6c:1f:00:72:1a:cc:c1:77:79:df:6a
+# SHA1 Fingerprint: 59:0d:2d:7d:88:4f:40:2e:61:7e:a5:62:32:17:65:cf:17:d8:94:e9
+# SHA256 Fingerprint: 91:e2:f5:78:8d:58:10:eb:a7:ba:58:73:7d:e1:54:8a:8e:ca:cd:01:45:98:bc:0b:14:3e:04:1b:17:05:25:52
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot 2011 O=Atos
+# Subject: CN=Atos TrustedRoot 2011 O=Atos
+# Label: "Atos TrustedRoot 2011"
+# Serial: 6643877497813316402
+# MD5 Fingerprint: ae:b9:c4:32:4b:ac:7f:5d:66:cc:77:94:bb:2a:77:56
+# SHA1 Fingerprint: 2b:b1:f5:3e:55:0c:1d:c5:f1:d4:e6:b7:6a:46:4b:55:06:02:ac:21
+# SHA256 Fingerprint: f3:56:be:a2:44:b7:a9:1e:b3:5d:53:ca:9a:d7:86:4a:ce:01:8e:2d:35:d5:f8:f9:6d:df:68:a6:f4:1a:a4:74
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 1 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 1 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 1 G3"
+# Serial: 687049649626669250736271037606554624078720034195
+# MD5 Fingerprint: a4:bc:5b:3f:fe:37:9a:fa:64:f0:e2:fa:05:3d:0b:ab
+# SHA1 Fingerprint: 1b:8e:ea:57:96:29:1a:c9:39:ea:b8:0a:81:1a:73:73:c0:93:79:67
+# SHA256 Fingerprint: 8a:86:6f:d1:b2:76:b5:7e:57:8e:92:1c:65:82:8a:2b:ed:58:e9:f2:f2:88:05:41:34:b7:f1:f4:bf:c9:cc:74
+-----BEGIN CERTIFICATE-----
+MIIFYDCCA0igAwIBAgIUeFhfLq0sGUvjNwc1NBMotZbUZZMwDQYJKoZIhvcNAQEL
+BQAwSDELMAkGA1UEBhMCQk0xGTAXBgNVBAoTEFF1b1ZhZGlzIExpbWl0ZWQxHjAc
+BgNVBAMTFVF1b1ZhZGlzIFJvb3QgQ0EgMSBHMzAeFw0xMjAxMTIxNzI3NDRaFw00
+MjAxMTIxNzI3NDRaMEgxCzAJBgNVBAYTAkJNMRkwFwYDVQQKExBRdW9WYWRpcyBM
+aW1pdGVkMR4wHAYDVQQDExVRdW9WYWRpcyBSb290IENBIDEgRzMwggIiMA0GCSqG
+SIb3DQEBAQUAA4ICDwAwggIKAoICAQCgvlAQjunybEC0BJyFuTHK3C3kEakEPBtV
+wedYMB0ktMPvhd6MLOHBPd+C5k+tR4ds7FtJwUrVu4/sh6x/gpqG7D0DmVIB0jWe
+rNrwU8lmPNSsAgHaJNM7qAJGr6Qc4/hzWHa39g6QDbXwz8z6+cZM5cOGMAqNF341
+68Xfuw6cwI2H44g4hWf6Pser4BOcBRiYz5P1sZK0/CPTz9XEJ0ngnjybCKOLXSoh
+4Pw5qlPafX7PGglTvF0FBM+hSo+LdoINofjSxxR3W5A2B4GbPgb6Ul5jxaYA/qXp
+UhtStZI5cgMJYr2wYBZupt0lwgNm3fME0UDiTouG9G/lg6AnhF4EwfWQvTA9xO+o
+abw4m6SkltFi2mnAAZauy8RRNOoMqv8hjlmPSlzkYZqn0ukqeI1RPToV7qJZjqlc
+3sX5kCLliEVx3ZGZbHqfPT2YfF72vhZooF6uCyP8Wg+qInYtyaEQHeTTRCOQiJ/G
+KubX9ZqzWB4vMIkIG1SitZgj7Ah3HJVdYdHLiZxfokqRmu8hqkkWCKi9YSgxyXSt
+hfbZxbGL0eUQMk1fiyA6PEkfM4VZDdvLCXVDaXP7a3F98N/ETH3Goy7IlXnLc6KO
+Tk0k+17kBL5yG6YnLUlamXrXXAkgt3+UuU/xDRxeiEIbEbfnkduebPRq34wGmAOt
+zCjvpUfzUwIDAQABo0IwQDAPBgNVHRMBAf8EBTADAQH/MA4GA1UdDwEB/wQEAwIB
+BjAdBgNVHQ4EFgQUo5fW816iEOGrRZ88F2Q87gFwnMwwDQYJKoZIhvcNAQELBQAD
+ggIBABj6W3X8PnrHX3fHyt/PX8MSxEBd1DKquGrX1RUVRpgjpeaQWxiZTOOtQqOC
+MTaIzen7xASWSIsBx40Bz1szBpZGZnQdT+3Btrm0DWHMY37XLneMlhwqI2hrhVd2
+cDMT/uFPpiN3GPoajOi9ZcnPP/TJF9zrx7zABC4tRi9pZsMbj/7sPtPKlL92CiUN
+qXsCHKnQO18LwIE6PWThv6ctTr1NxNgpxiIY0MWscgKCP6o6ojoilzHdCGPDdRS5
+YCgtW2jgFqlmgiNR9etT2DGbe+m3nUvriBbP+V04ikkwj+3x6xn0dxoxGE1nVGwv
+b2X52z3sIexe9PSLymBlVNFxZPT5pqOBMzYzcfCkeF9OrYMh3jRJjehZrJ3ydlo2
+8hP0r+AJx2EqbPfgna67hkooby7utHnNkDPDs3b69fBsnQGQ+p6Q9pxyz0fawx/k
+NSBT8lTR32GDpgLiJTjehTItXnOQUl1CxM49S+H5GYQd1aJQzEH7QRTDvdbJWqNj
+ZgKAvQU6O0ec7AAmTPWIUb+oI38YB7AL7YsmoWTTYUrrXJ/es69nA7Mf3W1daWhp
+q1467HxpvMc7hU6eFbm0FU/DlXpY18ls6Wy58yljXrQs8C097Vpl4KlbQMJImYFt
+nh8GKjwStIsPm6Ik8KaN1nrgS7ZklmOVhMJKzRwuJIczYOXD
+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 2 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 2 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 2 G3"
+# Serial: 390156079458959257446133169266079962026824725800
+# MD5 Fingerprint: af:0c:86:6e:bf:40:2d:7f:0b:3e:12:50:ba:12:3d:06
+# SHA1 Fingerprint: 09:3c:61:f3:8b:8b:dc:7d:55:df:75:38:02:05:00:e1:25:f5:c8:36
+# SHA256 Fingerprint: 8f:e4:fb:0a:f9:3a:4d:0d:67:db:0b:eb:b2:3e:37:c7:1b:f3:25:dc:bc:dd:24:0e:a0:4d:af:58:b4:7e:18:40
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=QuoVadis Root CA 3 G3 O=QuoVadis Limited
+# Subject: CN=QuoVadis Root CA 3 G3 O=QuoVadis Limited
+# Label: "QuoVadis Root CA 3 G3"
+# Serial: 268090761170461462463995952157327242137089239581
+# MD5 Fingerprint: df:7d:b9:ad:54:6f:68:a1:df:89:57:03:97:43:b0:d7
+# SHA1 Fingerprint: 48:12:bd:92:3c:a8:c4:39:06:e7:30:6d:27:96:e6:a4:cf:22:2e:7d
+# SHA256 Fingerprint: 88:ef:81:de:20:2e:b0:18:45:2e:43:f8:64:72:5c:ea:5f:bd:1f:c2:d9:d2:05:73:07:09:c5:d8:b8:69:0f:46
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root G2 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root G2 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root G2"
+# Serial: 15385348160840213938643033620894905419
+# MD5 Fingerprint: 92:38:b9:f8:63:24:82:65:2c:57:33:e6:fe:81:8f:9d
+# SHA1 Fingerprint: a1:4b:48:d9:43:ee:0a:0e:40:90:4f:3c:e0:a4:c0:91:93:51:5d:3f
+# SHA256 Fingerprint: 7d:05:eb:b6:82:33:9f:8c:94:51:ee:09:4e:eb:fe:fa:79:53:a1:14:ed:b2:f4:49:49:45:2f:ab:7d:2f:c1:85
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Assured ID Root G3 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Assured ID Root G3 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Assured ID Root G3"
+# Serial: 15459312981008553731928384953135426796
+# MD5 Fingerprint: 7c:7f:65:31:0c:81:df:8d:ba:3e:99:e2:5c:ad:6e:fb
+# SHA1 Fingerprint: f5:17:a2:4f:9a:48:c6:c9:f8:a2:00:26:9f:dc:0f:48:2c:ab:30:89
+# SHA256 Fingerprint: 7e:37:cb:8b:4c:47:09:0c:ab:36:55:1b:a6:f4:5d:b8:40:68:0f:ba:16:6a:95:2d:b1:00:71:7f:43:05:3f:c2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root G2 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root G2 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root G2"
+# Serial: 4293743540046975378534879503202253541
+# MD5 Fingerprint: e4:a6:8a:c8:54:ac:52:42:46:0a:fd:72:48:1b:2a:44
+# SHA1 Fingerprint: df:3c:24:f9:bf:d6:66:76:1b:26:80:73:fe:06:d1:cc:8d:4f:82:a4
+# SHA256 Fingerprint: cb:3c:cb:b7:60:31:e5:e0:13:8f:8d:d3:9a:23:f9:de:47:ff:c3:5e:43:c1:14:4c:ea:27:d4:6a:5a:b1:cb:5f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Global Root G3 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Global Root G3 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Global Root G3"
+# Serial: 7089244469030293291760083333884364146
+# MD5 Fingerprint: f5:5d:a4:50:a5:fb:28:7e:1e:0f:0d:cc:96:57:56:ca
+# SHA1 Fingerprint: 7e:04:de:89:6a:3e:66:6d:00:e6:87:d3:3f:fa:d9:3b:e8:3d:34:9e
+# SHA256 Fingerprint: 31:ad:66:48:f8:10:41:38:c7:38:f3:9e:a4:32:01:33:39:3e:3a:18:cc:02:29:6e:f9:7c:2a:c9:ef:67:31:d0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert Trusted Root G4 O=DigiCert Inc OU=www.digicert.com
+# Subject: CN=DigiCert Trusted Root G4 O=DigiCert Inc OU=www.digicert.com
+# Label: "DigiCert Trusted Root G4"
+# Serial: 7451500558977370777930084869016614236
+# MD5 Fingerprint: 78:f2:fc:aa:60:1f:2f:b4:eb:c9:37:ba:53:2e:75:49
+# SHA1 Fingerprint: dd:fb:16:cd:49:31:c9:73:a2:03:7d:3f:c8:3a:4d:7d:77:5d:05:e4
+# SHA256 Fingerprint: 55:2f:7b:dc:f1:a7:af:9e:6c:e6:72:01:7f:4f:12:ab:f7:72:40:c7:8e:76:1a:c2:03:d1:d9:d2:0a:c8:99:88
+-----BEGIN CERTIFICATE-----
+MIIFkDCCA3igAwIBAgIQBZsbV56OITLiOQe9p3d1XDANBgkqhkiG9w0BAQwFADBi
+MQswCQYDVQQGEwJVUzEVMBMGA1UEChMMRGlnaUNlcnQgSW5jMRkwFwYDVQQLExB3
+d3cuZGlnaWNlcnQuY29tMSEwHwYDVQQDExhEaWdpQ2VydCBUcnVzdGVkIFJvb3Qg
+RzQwHhcNMTMwODAxMTIwMDAwWhcNMzgwMTE1MTIwMDAwWjBiMQswCQYDVQQGEwJV
+UzEVMBMGA1UEChMMRGlnaUNlcnQgSW5jMRkwFwYDVQQLExB3d3cuZGlnaWNlcnQu
+Y29tMSEwHwYDVQQDExhEaWdpQ2VydCBUcnVzdGVkIFJvb3QgRzQwggIiMA0GCSqG
+SIb3DQEBAQUAA4ICDwAwggIKAoICAQC/5pBzaN675F1KPDAiMGkz7MKnJS7JIT3y
+ithZwuEppz1Yq3aaza57G4QNxDAf8xukOBbrVsaXbR2rsnnyyhHS5F/WBTxSD1If
+xp4VpX6+n6lXFllVcq9ok3DCsrp1mWpzMpTREEQQLt+C8weE5nQ7bXHiLQwb7iDV
+ySAdYyktzuxeTsiT+CFhmzTrBcZe7FsavOvJz82sNEBfsXpm7nfISKhmV1efVFiO
+DCu3T6cw2Vbuyntd463JT17lNecxy9qTXtyOj4DatpGYQJB5w3jHtrHEtWoYOAMQ
+jdjUN6QuBX2I9YI+EJFwq1WCQTLX2wRzKm6RAXwhTNS8rhsDdV14Ztk6MUSaM0C/
+CNdaSaTC5qmgZ92kJ7yhTzm1EVgX9yRcRo9k98FpiHaYdj1ZXUJ2h4mXaXpI8OCi
+EhtmmnTK3kse5w5jrubU75KSOp493ADkRSWJtppEGSt+wJS00mFt6zPZxd9LBADM
+fRyVw4/3IbKyEbe7f/LVjHAsQWCqsWMYRJUadmJ+9oCw++hkpjPRiQfhvbfmQ6QY
+uKZ3AeEPlAwhHbJUKSWJbOUOUlFHdL4mrLZBdd56rF+NP8m800ERElvlEFDrMcXK
+chYiCd98THU/Y+whX8QgUWtvsauGi0/C1kVfnSD8oR7FwI+isX4KJpn15GkvmB0t
+9dmpsh3lGwIDAQABo0IwQDAPBgNVHRMBAf8EBTADAQH/MA4GA1UdDwEB/wQEAwIB
+hjAdBgNVHQ4EFgQU7NfjgtJxXWRM3y5nP+e6mK4cD08wDQYJKoZIhvcNAQEMBQAD
+ggIBALth2X2pbL4XxJEbw6GiAI3jZGgPVs93rnD5/ZpKmbnJeFwMDF/k5hQpVgs2
+SV1EY+CtnJYYZhsjDT156W1r1lT40jzBQ0CuHVD1UvyQO7uYmWlrx8GnqGikJ9yd
++SeuMIW59mdNOj6PWTkiU0TryF0Dyu1Qen1iIQqAyHNm0aAFYF/opbSnr6j3bTWc
+fFqK1qI4mfN4i/RN0iAL3gTujJtHgXINwBQy7zBZLq7gcfJW5GqXb5JQbZaNaHqa
+sjYUegbyJLkJEVDXCLG4iXqEI2FCKeWjzaIgQdfRnGTZ6iahixTXTBmyUEFxPT9N
+cCOGDErcgdLMMpSEDQgJlxxPwO5rIHQw0uA5NBCFIRUBCOhVMt5xSdkoF1BN5r5N
+0XWs0Mr7QbhDparTwwVETyw2m+L64kW4I1NsBm9nVX9GtUw/bihaeSbSpKhil9Ie
+4u1Ki7wb/UdKDd9nZn6yW0HQO+T0O/QEY+nvwlQAUaCKKsnOeMzV6ocEGLPOr0mI
+r/OSmbaz5mEP0oUA51Aa5BuVnRmhuZyxm7EAHu/QD09CbMkKvO5D+jpxpchNJqU1
+/YldvIViHTLSoCtU7ZpXwdv6EM8Zt4tKG48BtieVU+i2iW1bvGjUI+iLUaJW+fCm
+gKDWHrO8Dw9TdSmq6hN35N6MgSGtBxBHEa2HPQfRdbzP82Z+
+-----END CERTIFICATE-----
+
+# Issuer: CN=COMODO RSA Certification Authority O=COMODO CA Limited
+# Subject: CN=COMODO RSA Certification Authority O=COMODO CA Limited
+# Label: "COMODO RSA Certification Authority"
+# Serial: 101909084537582093308941363524873193117
+# MD5 Fingerprint: 1b:31:b0:71:40:36:cc:14:36:91:ad:c4:3e:fd:ec:18
+# SHA1 Fingerprint: af:e5:d2:44:a8:d1:19:42:30:ff:47:9f:e2:f8:97:bb:cd:7a:8c:b4
+# SHA256 Fingerprint: 52:f0:e1:c4:e5:8e:c6:29:29:1b:60:31:7f:07:46:71:b8:5d:7e:a8:0d:5b:07:27:34:63:53:4b:32:b4:02:34
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=USERTrust RSA Certification Authority O=The USERTRUST Network
+# Subject: CN=USERTrust RSA Certification Authority O=The USERTRUST Network
+# Label: "USERTrust RSA Certification Authority"
+# Serial: 2645093764781058787591871645665788717
+# MD5 Fingerprint: 1b:fe:69:d1:91:b7:19:33:a3:72:a8:0f:e1:55:e5:b5
+# SHA1 Fingerprint: 2b:8f:1b:57:33:0d:bb:a2:d0:7a:6c:51:f7:0e:e9:0d:da:b9:ad:8e
+# SHA256 Fingerprint: e7:93:c9:b0:2f:d8:aa:13:e2:1c:31:22:8a:cc:b0:81:19:64:3b:74:9c:89:89:64:b1:74:6d:46:c3:d4:cb:d2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=USERTrust ECC Certification Authority O=The USERTRUST Network
+# Subject: CN=USERTrust ECC Certification Authority O=The USERTRUST Network
+# Label: "USERTrust ECC Certification Authority"
+# Serial: 123013823720199481456569720443997572134
+# MD5 Fingerprint: fa:68:bc:d9:b5:7f:ad:fd:c9:1d:06:83:28:cc:24:c1
+# SHA1 Fingerprint: d1:cb:ca:5d:b2:d5:2a:7f:69:3b:67:4d:e5:f0:5a:1d:0c:95:7d:f0
+# SHA256 Fingerprint: 4f:f4:60:d5:4b:9c:86:da:bf:bc:fc:57:12:e0:40:0d:2b:ed:3f:bc:4d:4f:bd:aa:86:e0:6a:dc:d2:a9:ad:7a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R5
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R5
+# Label: "GlobalSign ECC Root CA - R5"
+# Serial: 32785792099990507226680698011560947931244
+# MD5 Fingerprint: 9f:ad:3b:1c:02:1e:8a:ba:17:74:38:81:0c:a2:bc:08
+# SHA1 Fingerprint: 1f:24:c6:30:cd:a4:18:ef:20:69:ff:ad:4f:dd:5f:46:3a:1b:69:aa
+# SHA256 Fingerprint: 17:9f:bc:14:8a:3d:d0:0f:d2:4e:a1:34:58:cc:43:bf:a7:f5:9c:81:82:d7:83:a5:13:f6:eb:ec:10:0c:89:24
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=IdenTrust Commercial Root CA 1 O=IdenTrust
+# Subject: CN=IdenTrust Commercial Root CA 1 O=IdenTrust
+# Label: "IdenTrust Commercial Root CA 1"
+# Serial: 13298821034946342390520003877796839426
+# MD5 Fingerprint: b3:3e:77:73:75:ee:a0:d3:e3:7e:49:63:49:59:bb:c7
+# SHA1 Fingerprint: df:71:7e:aa:4a:d9:4e:c9:55:84:99:60:2d:48:de:5f:bc:f0:3a:25
+# SHA256 Fingerprint: 5d:56:49:9b:e4:d2:e0:8b:cf:ca:d0:8a:3e:38:72:3d:50:50:3b:de:70:69:48:e4:2f:55:60:30:19:e5:28:ae
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=IdenTrust Public Sector Root CA 1 O=IdenTrust
+# Subject: CN=IdenTrust Public Sector Root CA 1 O=IdenTrust
+# Label: "IdenTrust Public Sector Root CA 1"
+# Serial: 13298821034946342390521976156843933698
+# MD5 Fingerprint: 37:06:a5:b0:fc:89:9d:ba:f4:6b:8c:1a:64:cd:d5:ba
+# SHA1 Fingerprint: ba:29:41:60:77:98:3f:f4:f3:ef:f2:31:05:3b:2e:ea:6d:4d:45:fd
+# SHA256 Fingerprint: 30:d0:89:5a:9a:44:8a:26:20:91:63:55:22:d1:f5:20:10:b5:86:7a:ca:e1:2c:78:ef:95:8f:d4:f4:38:9f:2f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority - G2 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2009 Entrust, Inc. - for authorized use only
+# Subject: CN=Entrust Root Certification Authority - G2 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2009 Entrust, Inc. - for authorized use only
+# Label: "Entrust Root Certification Authority - G2"
+# Serial: 1246989352
+# MD5 Fingerprint: 4b:e2:c9:91:96:65:0c:f4:0e:5a:93:92:a0:0a:fe:b2
+# SHA1 Fingerprint: 8c:f4:27:fd:79:0c:3a:d1:66:06:8d:e8:1e:57:ef:bb:93:22:72:d4
+# SHA256 Fingerprint: 43:df:57:74:b0:3e:7f:ef:5f:e4:0d:93:1a:7b:ed:f1:bb:2e:6b:42:73:8c:4e:6d:38:41:10:3d:3a:a7:f3:39
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority - EC1 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2012 Entrust, Inc. - for authorized use only
+# Subject: CN=Entrust Root Certification Authority - EC1 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2012 Entrust, Inc. - for authorized use only
+# Label: "Entrust Root Certification Authority - EC1"
+# Serial: 51543124481930649114116133369
+# MD5 Fingerprint: b6:7e:1d:f0:58:c5:49:6c:24:3b:3d:ed:98:18:ed:bc
+# SHA1 Fingerprint: 20:d8:06:40:df:9b:25:f5:12:25:3a:11:ea:f7:59:8a:eb:14:b5:47
+# SHA256 Fingerprint: 02:ed:0e:b2:8c:14:da:45:16:5c:56:67:91:70:0d:64:51:d7:fb:56:f0:b2:ab:1d:3b:8e:b0:70:e5:6e:df:f5
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=CFCA EV ROOT O=China Financial Certification Authority
+# Subject: CN=CFCA EV ROOT O=China Financial Certification Authority
+# Label: "CFCA EV ROOT"
+# Serial: 407555286
+# MD5 Fingerprint: 74:e1:b6:ed:26:7a:7a:44:30:33:94:ab:7b:27:81:30
+# SHA1 Fingerprint: e2:b8:29:4b:55:84:ab:6b:58:c2:90:46:6c:ac:3f:b8:39:8f:84:83
+# SHA256 Fingerprint: 5c:c3:d7:8e:4e:1d:5e:45:54:7a:04:e6:87:3e:64:f9:0c:f9:53:6d:1c:cc:2e:f8:00:f3:55:c4:c5:fd:70:fd
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=OISTE WISeKey Global Root GB CA O=WISeKey OU=OISTE Foundation Endorsed
+# Subject: CN=OISTE WISeKey Global Root GB CA O=WISeKey OU=OISTE Foundation Endorsed
+# Label: "OISTE WISeKey Global Root GB CA"
+# Serial: 157768595616588414422159278966750757568
+# MD5 Fingerprint: a4:eb:b9:61:28:2e:b7:2f:98:b0:35:26:90:99:51:1d
+# SHA1 Fingerprint: 0f:f9:40:76:18:d3:d7:6a:4b:98:f0:a8:35:9e:0c:fd:27:ac:cc:ed
+# SHA256 Fingerprint: 6b:9c:08:e8:6e:b0:f7:67:cf:ad:65:cd:98:b6:21:49:e5:49:4a:67:f5:84:5e:7b:d1:ed:01:9f:27:b8:6b:d6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SZAFIR ROOT CA2 O=Krajowa Izba Rozliczeniowa S.A.
+# Subject: CN=SZAFIR ROOT CA2 O=Krajowa Izba Rozliczeniowa S.A.
+# Label: "SZAFIR ROOT CA2"
+# Serial: 357043034767186914217277344587386743377558296292
+# MD5 Fingerprint: 11:64:c1:89:b0:24:b1:8c:b1:07:7e:89:9e:51:9e:99
+# SHA1 Fingerprint: e2:52:fa:95:3f:ed:db:24:60:bd:6e:28:f3:9c:cc:cf:5e:b3:3f:de
+# SHA256 Fingerprint: a1:33:9d:33:28:1a:0b:56:e5:57:d3:d3:2b:1c:e7:f9:36:7e:b0:94:bd:5f:a7:2a:7e:50:04:c8:de:d7:ca:fe
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Network CA 2 O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Network CA 2 O=Unizeto Technologies S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Network CA 2"
+# Serial: 44979900017204383099463764357512596969
+# MD5 Fingerprint: 6d:46:9e:d9:25:6d:08:23:5b:5e:74:7d:1e:27:db:f2
+# SHA1 Fingerprint: d3:dd:48:3e:2b:bf:4c:05:e8:af:10:f5:fa:76:26:cf:d3:dc:30:92
+# SHA256 Fingerprint: b6:76:f2:ed:da:e8:77:5c:d3:6c:b0:f6:3c:d1:d4:60:39:61:f4:9e:62:65:ba:01:3a:2f:03:07:b6:d0:b8:04
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hellenic Academic and Research Institutions RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Subject: CN=Hellenic Academic and Research Institutions RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Label: "Hellenic Academic and Research Institutions RootCA 2015"
+# Serial: 0
+# MD5 Fingerprint: ca:ff:e2:db:03:d9:cb:4b:e9:0f:ad:84:fd:7b:18:ce
+# SHA1 Fingerprint: 01:0c:06:95:a6:98:19:14:ff:bf:5f:c6:b0:b6:95:ea:29:e9:12:a6
+# SHA256 Fingerprint: a0:40:92:9a:02:ce:53:b4:ac:f4:f2:ff:c6:98:1c:e4:49:6f:75:5e:6d:45:fe:0b:2a:69:2b:cd:52:52:3f:36
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hellenic Academic and Research Institutions ECC RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Subject: CN=Hellenic Academic and Research Institutions ECC RootCA 2015 O=Hellenic Academic and Research Institutions Cert. Authority
+# Label: "Hellenic Academic and Research Institutions ECC RootCA 2015"
+# Serial: 0
+# MD5 Fingerprint: 81:e5:b4:17:eb:c2:f5:e1:4b:0d:41:7b:49:92:fe:ef
+# SHA1 Fingerprint: 9f:f1:71:8d:92:d5:9a:f3:7d:74:97:b4:bc:6f:84:68:0b:ba:b6:66
+# SHA256 Fingerprint: 44:b5:45:aa:8a:25:e6:5a:73:ca:15:dc:27:fc:36:d2:4c:1c:b9:95:3a:06:65:39:b1:15:82:dc:48:7b:48:33
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=ISRG Root X1 O=Internet Security Research Group
+# Subject: CN=ISRG Root X1 O=Internet Security Research Group
+# Label: "ISRG Root X1"
+# Serial: 172886928669790476064670243504169061120
+# MD5 Fingerprint: 0c:d2:f9:e0:da:17:73:e9:ed:86:4d:a5:e3:70:e7:4e
+# SHA1 Fingerprint: ca:bd:2a:79:a1:07:6a:31:f2:1d:25:36:35:cb:03:9d:43:29:a5:e8
+# SHA256 Fingerprint: 96:bc:ec:06:26:49:76:f3:74:60:77:9a:cf:28:c5:a7:cf:e8:a3:c0:aa:e1:1a:8f:fc:ee:05:c0:bd:df:08:c6
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: O=FNMT-RCM OU=AC RAIZ FNMT-RCM
+# Subject: O=FNMT-RCM OU=AC RAIZ FNMT-RCM
+# Label: "AC RAIZ FNMT-RCM"
+# Serial: 485876308206448804701554682760554759
+# MD5 Fingerprint: e2:09:04:b4:d3:bd:d1:a0:14:fd:1a:d2:47:c4:57:1d
+# SHA1 Fingerprint: ec:50:35:07:b2:15:c4:95:62:19:e2:a8:9a:5b:42:99:2c:4c:2c:20
+# SHA256 Fingerprint: eb:c5:57:0c:29:01:8c:4d:67:b1:aa:12:7b:af:12:f7:03:b4:61:1e:bc:17:b7:da:b5:57:38:94:17:9b:93:fa
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 1 O=Amazon
+# Subject: CN=Amazon Root CA 1 O=Amazon
+# Label: "Amazon Root CA 1"
+# Serial: 143266978916655856878034712317230054538369994
+# MD5 Fingerprint: 43:c6:bf:ae:ec:fe:ad:2f:18:c6:88:68:30:fc:c8:e6
+# SHA1 Fingerprint: 8d:a7:f9:65:ec:5e:fc:37:91:0f:1c:6e:59:fd:c1:cc:6a:6e:de:16
+# SHA256 Fingerprint: 8e:cd:e6:88:4f:3d:87:b1:12:5b:a3:1a:c3:fc:b1:3d:70:16:de:7f:57:cc:90:4f:e1:cb:97:c6:ae:98:19:6e
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 2 O=Amazon
+# Subject: CN=Amazon Root CA 2 O=Amazon
+# Label: "Amazon Root CA 2"
+# Serial: 143266982885963551818349160658925006970653239
+# MD5 Fingerprint: c8:e5:8d:ce:a8:42:e2:7a:c0:2a:5c:7c:9e:26:bf:66
+# SHA1 Fingerprint: 5a:8c:ef:45:d7:a6:98:59:76:7a:8c:8b:44:96:b5:78:cf:47:4b:1a
+# SHA256 Fingerprint: 1b:a5:b2:aa:8c:65:40:1a:82:96:01:18:f8:0b:ec:4f:62:30:4d:83:ce:c4:71:3a:19:c3:9c:01:1e:a4:6d:b4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 3 O=Amazon
+# Subject: CN=Amazon Root CA 3 O=Amazon
+# Label: "Amazon Root CA 3"
+# Serial: 143266986699090766294700635381230934788665930
+# MD5 Fingerprint: a0:d4:ef:0b:f7:b5:d8:49:95:2a:ec:f5:c4:fc:81:87
+# SHA1 Fingerprint: 0d:44:dd:8c:3c:8c:1a:1a:58:75:64:81:e9:0f:2e:2a:ff:b3:d2:6e
+# SHA256 Fingerprint: 18:ce:6c:fe:7b:f1:4e:60:b2:e3:47:b8:df:e8:68:cb:31:d0:2e:bb:3a:da:27:15:69:f5:03:43:b4:6d:b3:a4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Amazon Root CA 4 O=Amazon
+# Subject: CN=Amazon Root CA 4 O=Amazon
+# Label: "Amazon Root CA 4"
+# Serial: 143266989758080763974105200630763877849284878
+# MD5 Fingerprint: 89:bc:27:d5:eb:17:8d:06:6a:69:d5:fd:89:47:b4:cd
+# SHA1 Fingerprint: f6:10:84:07:d6:f8:bb:67:98:0c:c2:e2:44:c2:eb:ae:1c:ef:63:be
+# SHA256 Fingerprint: e3:5d:28:41:9e:d0:20:25:cf:a6:90:38:cd:62:39:62:45:8d:a5:c6:95:fb:de:a3:c2:2b:0b:fb:25:89:70:92
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1 O=Turkiye Bilimsel ve Teknolojik Arastirma Kurumu - TUBITAK OU=Kamu Sertifikasyon Merkezi - Kamu SM
+# Subject: CN=TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1 O=Turkiye Bilimsel ve Teknolojik Arastirma Kurumu - TUBITAK OU=Kamu Sertifikasyon Merkezi - Kamu SM
+# Label: "TUBITAK Kamu SM SSL Kok Sertifikasi - Surum 1"
+# Serial: 1
+# MD5 Fingerprint: dc:00:81:dc:69:2f:3e:2f:b0:3b:f6:3d:5a:91:8e:49
+# SHA1 Fingerprint: 31:43:64:9b:ec:ce:27:ec:ed:3a:3f:0b:8f:0d:e4:e8:91:dd:ee:ca
+# SHA256 Fingerprint: 46:ed:c3:68:90:46:d5:3a:45:3f:b3:10:4a:b8:0d:ca:ec:65:8b:26:60:ea:16:29:dd:7e:86:79:90:64:87:16
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GDCA TrustAUTH R5 ROOT O=GUANG DONG CERTIFICATE AUTHORITY CO.,LTD.
+# Subject: CN=GDCA TrustAUTH R5 ROOT O=GUANG DONG CERTIFICATE AUTHORITY CO.,LTD.
+# Label: "GDCA TrustAUTH R5 ROOT"
+# Serial: 9009899650740120186
+# MD5 Fingerprint: 63:cc:d9:3d:34:35:5c:6f:53:a3:e2:08:70:48:1f:b4
+# SHA1 Fingerprint: 0f:36:38:5b:81:1a:25:c3:9b:31:4e:83:ca:e9:34:66:70:cc:74:b4
+# SHA256 Fingerprint: bf:ff:8f:d0:44:33:48:7d:6a:8a:a6:0c:1a:29:76:7a:9f:c2:bb:b0:5e:42:0f:71:3a:13:b9:92:89:1d:38:93
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com Root Certification Authority RSA O=SSL Corporation
+# Subject: CN=SSL.com Root Certification Authority RSA O=SSL Corporation
+# Label: "SSL.com Root Certification Authority RSA"
+# Serial: 8875640296558310041
+# MD5 Fingerprint: 86:69:12:c0:70:f1:ec:ac:ac:c2:d5:bc:a5:5b:a1:29
+# SHA1 Fingerprint: b7:ab:33:08:d1:ea:44:77:ba:14:80:12:5a:6f:bd:a9:36:49:0c:bb
+# SHA256 Fingerprint: 85:66:6a:56:2e:e0:be:5c:e9:25:c1:d8:89:0a:6f:76:a8:7e:c1:6d:4d:7d:5f:29:ea:74:19:cf:20:12:3b:69
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com Root Certification Authority ECC O=SSL Corporation
+# Subject: CN=SSL.com Root Certification Authority ECC O=SSL Corporation
+# Label: "SSL.com Root Certification Authority ECC"
+# Serial: 8495723813297216424
+# MD5 Fingerprint: 2e:da:e4:39:7f:9c:8f:37:d1:70:9f:26:17:51:3a:8e
+# SHA1 Fingerprint: c3:19:7c:39:24:e6:54:af:1b:c4:ab:20:95:7a:e2:c3:0e:13:02:6a
+# SHA256 Fingerprint: 34:17:bb:06:cc:60:07:da:1b:96:1c:92:0b:8a:b4:ce:3f:ad:82:0e:4a:a3:0b:9a:cb:c4:a7:4e:bd:ce:bc:65
+-----BEGIN CERTIFICATE-----
+MIICjTCCAhSgAwIBAgIIdebfy8FoW6gwCgYIKoZIzj0EAwIwfDELMAkGA1UEBhMC
+VVMxDjAMBgNVBAgMBVRleGFzMRAwDgYDVQQHDAdIb3VzdG9uMRgwFgYDVQQKDA9T
+U0wgQ29ycG9yYXRpb24xMTAvBgNVBAMMKFNTTC5jb20gUm9vdCBDZXJ0aWZpY2F0
+aW9uIEF1dGhvcml0eSBFQ0MwHhcNMTYwMjEyMTgxNDAzWhcNNDEwMjEyMTgxNDAz
+WjB8MQswCQYDVQQGEwJVUzEOMAwGA1UECAwFVGV4YXMxEDAOBgNVBAcMB0hvdXN0
+b24xGDAWBgNVBAoMD1NTTCBDb3Jwb3JhdGlvbjExMC8GA1UEAwwoU1NMLmNvbSBS
+b290IENlcnRpZmljYXRpb24gQXV0aG9yaXR5IEVDQzB2MBAGByqGSM49AgEGBSuB
+BAAiA2IABEVuqVDEpiM2nl8ojRfLliJkP9x6jh3MCLOicSS6jkm5BBtHllirLZXI
+7Z4INcgn64mMU1jrYor+8FsPazFSY0E7ic3s7LaNGdM0B9y7xgZ/wkWV7Mt/qCPg
+CemB+vNH06NjMGEwHQYDVR0OBBYEFILRhXMw5zUE044CkvvlpNHEIejNMA8GA1Ud
+EwEB/wQFMAMBAf8wHwYDVR0jBBgwFoAUgtGFczDnNQTTjgKS++Wk0cQh6M0wDgYD
+VR0PAQH/BAQDAgGGMAoGCCqGSM49BAMCA2cAMGQCMG/n61kRpGDPYbCWe+0F+S8T
+kdzt5fxQaxFGRrMcIQBiu77D5+jNB5n5DQtdcj7EqgIwH7y6C+IwJPt8bYBVCpk+
+gA0z5Wajs6O7pdWLjwkspl1+4vAHCGht0nxpbl/f5Wpl
+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com EV Root Certification Authority RSA R2 O=SSL Corporation
+# Subject: CN=SSL.com EV Root Certification Authority RSA R2 O=SSL Corporation
+# Label: "SSL.com EV Root Certification Authority RSA R2"
+# Serial: 6248227494352943350
+# MD5 Fingerprint: e1:1e:31:58:1a:ae:54:53:02:f6:17:6a:11:7b:4d:95
+# SHA1 Fingerprint: 74:3a:f0:52:9b:d0:32:a0:f4:4a:83:cd:d4:ba:a9:7b:7c:2e:c4:9a
+# SHA256 Fingerprint: 2e:7b:f1:6c:c2:24:85:a7:bb:e2:aa:86:96:75:07:61:b0:ae:39:be:3b:2f:e9:d0:cc:6d:4e:f7:34:91:42:5c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com EV Root Certification Authority ECC O=SSL Corporation
+# Subject: CN=SSL.com EV Root Certification Authority ECC O=SSL Corporation
+# Label: "SSL.com EV Root Certification Authority ECC"
+# Serial: 3182246526754555285
+# MD5 Fingerprint: 59:53:22:65:83:42:01:54:c0:ce:42:b9:5a:7c:f2:90
+# SHA1 Fingerprint: 4c:dd:51:a3:d1:f5:20:32:14:b0:c6:c5:32:23:03:91:c7:46:42:6d
+# SHA256 Fingerprint: 22:a2:c1:f7:bd:ed:70:4c:c1:e7:01:b5:f4:08:c3:10:88:0f:e9:56:b5:de:2a:4a:44:f9:9c:87:3a:25:a7:c8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R6
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign Root CA - R6
+# Label: "GlobalSign Root CA - R6"
+# Serial: 1417766617973444989252670301619537
+# MD5 Fingerprint: 4f:dd:07:e4:d4:22:64:39:1e:0c:37:42:ea:d1:c6:ae
+# SHA1 Fingerprint: 80:94:64:0e:b5:a7:a1:ca:11:9c:1f:dd:d5:9f:81:02:63:a7:fb:d1
+# SHA256 Fingerprint: 2c:ab:ea:fe:37:d0:6c:a2:2a:ba:73:91:c0:03:3d:25:98:29:52:c4:53:64:73:49:76:3a:3a:b5:ad:6c:cf:69
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=OISTE WISeKey Global Root GC CA O=WISeKey OU=OISTE Foundation Endorsed
+# Subject: CN=OISTE WISeKey Global Root GC CA O=WISeKey OU=OISTE Foundation Endorsed
+# Label: "OISTE WISeKey Global Root GC CA"
+# Serial: 44084345621038548146064804565436152554
+# MD5 Fingerprint: a9:d6:b9:2d:2f:93:64:f8:a5:69:ca:91:e9:68:07:23
+# SHA1 Fingerprint: e0:11:84:5e:34:de:be:88:81:b9:9c:f6:16:26:d1:96:1f:c3:b9:31
+# SHA256 Fingerprint: 85:60:f9:1c:36:24:da:ba:95:70:b5:fe:a0:db:e3:6f:f1:1a:83:23:be:94:86:85:4f:b3:f3:4a:55:71:19:8d
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=UCA Global G2 Root O=UniTrust
+# Subject: CN=UCA Global G2 Root O=UniTrust
+# Label: "UCA Global G2 Root"
+# Serial: 124779693093741543919145257850076631279
+# MD5 Fingerprint: 80:fe:f0:c4:4a:f0:5c:62:32:9f:1c:ba:78:a9:50:f8
+# SHA1 Fingerprint: 28:f9:78:16:19:7a:ff:18:25:18:aa:44:fe:c1:a0:ce:5c:b6:4c:8a
+# SHA256 Fingerprint: 9b:ea:11:c9:76:fe:01:47:64:c1:be:56:a6:f9:14:b5:a5:60:31:7a:bd:99:88:39:33:82:e5:16:1a:a0:49:3c
+-----BEGIN CERTIFICATE-----
+MIIFRjCCAy6gAwIBAgIQXd+x2lqj7V2+WmUgZQOQ7zANBgkqhkiG9w0BAQsFADA9
+MQswCQYDVQQGEwJDTjERMA8GA1UECgwIVW5pVHJ1c3QxGzAZBgNVBAMMElVDQSBH
+bG9iYWwgRzIgUm9vdDAeFw0xNjAzMTEwMDAwMDBaFw00MDEyMzEwMDAwMDBaMD0x
+CzAJBgNVBAYTAkNOMREwDwYDVQQKDAhVbmlUcnVzdDEbMBkGA1UEAwwSVUNBIEds
+b2JhbCBHMiBSb290MIICIjANBgkqhkiG9w0BAQEFAAOCAg8AMIICCgKCAgEAxeYr
+b3zvJgUno4Ek2m/LAfmZmqkywiKHYUGRO8vDaBsGxUypK8FnFyIdK+35KYmToni9
+kmugow2ifsqTs6bRjDXVdfkX9s9FxeV67HeToI8jrg4aA3++1NDtLnurRiNb/yzm
+VHqUwCoV8MmNsHo7JOHXaOIxPAYzRrZUEaalLyJUKlgNAQLx+hVRZ2zA+te2G3/R
+VogvGjqNO7uCEeBHANBSh6v7hn4PJGtAnTRnvI3HLYZveT6OqTwXS3+wmeOwcWDc
+C/Vkw85DvG1xudLeJ1uK6NjGruFZfc8oLTW4lVYa8bJYS7cSN8h8s+1LgOGN+jIj
+tm+3SJUIsUROhYw6AlQgL9+/V087OpAh18EmNVQg7Mc/R+zvWr9LesGtOxdQXGLY
+D0tK3Cv6brxzks3sx1DoQZbXqX5t2Okdj4q1uViSukqSKwxW/YDrCPBeKW4bHAyv
+j5OJrdu9o54hyokZ7N+1wxrrFv54NkzWbtA+FxyQF2smuvt6L78RHBgOLXMDj6Dl
+NaBa4kx1HXHhOThTeEDMg5PXCp6dW4+K5OXgSORIskfNTip1KnvyIvbJvgmRlld6
+iIis7nCs+dwp4wwcOxJORNanTrAmyPPZGpeRaOrvjUYG0lZFWJo8DA+DuAUlwznP
+O6Q0ibd5Ei9Hxeepl2n8pndntd978XplFeRhVmUCAwEAAaNCMEAwDgYDVR0PAQH/
+BAQDAgEGMA8GA1UdEwEB/wQFMAMBAf8wHQYDVR0OBBYEFIHEjMz15DD/pQwIX4wV
+ZyF0Ad/fMA0GCSqGSIb3DQEBCwUAA4ICAQATZSL1jiutROTL/7lo5sOASD0Ee/oj
+L3rtNtqyzm325p7lX1iPyzcyochltq44PTUbPrw7tgTQvPlJ9Zv3hcU2tsu8+Mg5
+1eRfB70VVJd0ysrtT7q6ZHafgbiERUlMjW+i67HM0cOU2kTC5uLqGOiiHycFutfl
+1qnN3e92mI0ADs0b+gO3joBYDic/UvuUospeZcnWhNq5NXHzJsBPd+aBJ9J3O5oU
+b3n09tDh05S60FdRvScFDcH9yBIw7m+NESsIndTUv4BFFJqIRNow6rSn4+7vW4LV
+PtateJLbXDzz2K36uGt/xDYotgIVilQsnLAXc47QN6MUPJiVAAwpBVueSUmxX8fj
+y88nZY41F7dXyDDZQVu5FLbowg+UMaeUmMxq67XhJ/UQqAHojhJi6IjMtX9Gl8Cb
+EGY4GjZGXyJoPd/JxhMnq1MGrKI8hgZlb7F+sSlEmqO6SWkoaY/X5V+tBIZkbxqg
+DMUIYs6Ao9Dz7GjevjPHF1t/gMRMTLGmhIrDO7gJzRSBuhjjVFc2/tsvfEehOjPI
++Vg7RE+xygKJBJYoaMVLuCaJu9YzL1DV/pqJuhgyklTGW+Cd+V7lDSKb9triyCGy
+YiGqhkCyLmTTX8jjfhFnRR8F/uOi77Oos/N9j/gMHyIfLXC0uAE0djAA5SN4p1bX
+UB+K+wb1whnw0A==
+-----END CERTIFICATE-----
+
+# Issuer: CN=UCA Extended Validation Root O=UniTrust
+# Subject: CN=UCA Extended Validation Root O=UniTrust
+# Label: "UCA Extended Validation Root"
+# Serial: 106100277556486529736699587978573607008
+# MD5 Fingerprint: a1:f3:5f:43:c6:34:9b:da:bf:8c:7e:05:53:ad:96:e2
+# SHA1 Fingerprint: a3:a1:b0:6f:24:61:23:4a:e3:36:a5:c2:37:fc:a6:ff:dd:f0:d7:3a
+# SHA256 Fingerprint: d4:3a:f9:b3:54:73:75:5c:96:84:fc:06:d7:d8:cb:70:ee:5c:28:e7:73:fb:29:4e:b4:1e:e7:17:22:92:4d:24
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certigna Root CA O=Dhimyotis OU=0002 48146308100036
+# Subject: CN=Certigna Root CA O=Dhimyotis OU=0002 48146308100036
+# Label: "Certigna Root CA"
+# Serial: 269714418870597844693661054334862075617
+# MD5 Fingerprint: 0e:5c:30:62:27:eb:5b:bc:d7:ae:62:ba:e9:d5:df:77
+# SHA1 Fingerprint: 2d:0d:52:14:ff:9e:ad:99:24:01:74:20:47:6e:6c:85:27:27:f5:43
+# SHA256 Fingerprint: d4:8d:3d:23:ee:db:50:a4:59:e5:51:97:60:1c:27:77:4b:9d:7b:18:c9:4d:5a:05:95:11:a1:02:50:b9:31:68
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign Root CA - G1 O=eMudhra Technologies Limited OU=emSign PKI
+# Subject: CN=emSign Root CA - G1 O=eMudhra Technologies Limited OU=emSign PKI
+# Label: "emSign Root CA - G1"
+# Serial: 235931866688319308814040
+# MD5 Fingerprint: 9c:42:84:57:dd:cb:0b:a7:2e:95:ad:b6:f3:da:bc:ac
+# SHA1 Fingerprint: 8a:c7:ad:8f:73:ac:4e:c1:b5:75:4d:a5:40:f4:fc:cf:7c:b5:8e:8c
+# SHA256 Fingerprint: 40:f6:af:03:46:a9:9a:a1:cd:1d:55:5a:4e:9c:ce:62:c7:f9:63:46:03:ee:40:66:15:83:3d:c8:c8:d0:03:67
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign ECC Root CA - G3 O=eMudhra Technologies Limited OU=emSign PKI
+# Subject: CN=emSign ECC Root CA - G3 O=eMudhra Technologies Limited OU=emSign PKI
+# Label: "emSign ECC Root CA - G3"
+# Serial: 287880440101571086945156
+# MD5 Fingerprint: ce:0b:72:d1:9f:88:8e:d0:50:03:e8:e3:b8:8b:67:40
+# SHA1 Fingerprint: 30:43:fa:4f:f2:57:dc:a0:c3:80:ee:2e:58:ea:78:b2:3f:e6:bb:c1
+# SHA256 Fingerprint: 86:a1:ec:ba:08:9c:4a:8d:3b:be:27:34:c6:12:ba:34:1d:81:3e:04:3c:f9:e8:a8:62:cd:5c:57:a3:6b:be:6b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign Root CA - C1 O=eMudhra Inc OU=emSign PKI
+# Subject: CN=emSign Root CA - C1 O=eMudhra Inc OU=emSign PKI
+# Label: "emSign Root CA - C1"
+# Serial: 825510296613316004955058
+# MD5 Fingerprint: d8:e3:5d:01:21:fa:78:5a:b0:df:ba:d2:ee:2a:5f:68
+# SHA1 Fingerprint: e7:2e:f1:df:fc:b2:09:28:cf:5d:d4:d5:67:37:b1:51:cb:86:4f:01
+# SHA256 Fingerprint: 12:56:09:aa:30:1d:a0:a2:49:b9:7a:82:39:cb:6a:34:21:6f:44:dc:ac:9f:39:54:b1:42:92:f2:e8:c8:60:8f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=emSign ECC Root CA - C3 O=eMudhra Inc OU=emSign PKI
+# Subject: CN=emSign ECC Root CA - C3 O=eMudhra Inc OU=emSign PKI
+# Label: "emSign ECC Root CA - C3"
+# Serial: 582948710642506000014504
+# MD5 Fingerprint: 3e:53:b3:a3:81:ee:d7:10:f8:d3:b0:1d:17:92:f5:d5
+# SHA1 Fingerprint: b6:af:43:c2:9b:81:53:7d:f6:ef:6b:c3:1f:1f:60:15:0c:ee:48:66
+# SHA256 Fingerprint: bc:4d:80:9b:15:18:9d:78:db:3e:1d:8c:f4:f9:72:6a:79:5d:a1:64:3c:a5:f1:35:8e:1d:db:0e:dc:0d:7e:b3
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Hongkong Post Root CA 3 O=Hongkong Post
+# Subject: CN=Hongkong Post Root CA 3 O=Hongkong Post
+# Label: "Hongkong Post Root CA 3"
+# Serial: 46170865288971385588281144162979347873371282084
+# MD5 Fingerprint: 11:fc:9f:bd:73:30:02:8a:fd:3f:f3:58:b9:cb:20:f0
+# SHA1 Fingerprint: 58:a2:d0:ec:20:52:81:5b:c1:f3:f8:64:02:24:4e:c2:8e:02:4b:02
+# SHA256 Fingerprint: 5a:2f:c0:3f:0c:83:b0:90:bb:fa:40:60:4b:09:88:44:6c:76:36:18:3d:f9:84:6e:17:10:1a:44:7f:b8:ef:d6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Entrust Root Certification Authority - G4 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2015 Entrust, Inc. - for authorized use only
+# Subject: CN=Entrust Root Certification Authority - G4 O=Entrust, Inc. OU=See www.entrust.net/legal-terms/(c) 2015 Entrust, Inc. - for authorized use only
+# Label: "Entrust Root Certification Authority - G4"
+# Serial: 289383649854506086828220374796556676440
+# MD5 Fingerprint: 89:53:f1:83:23:b7:7c:8e:05:f1:8c:71:38:4e:1f:88
+# SHA1 Fingerprint: 14:88:4e:86:26:37:b0:26:af:59:62:5c:40:77:ec:35:29:ba:96:01
+# SHA256 Fingerprint: db:35:17:d1:f6:73:2a:2d:5a:b9:7c:53:3e:c7:07:79:ee:32:70:a6:2f:b4:ac:42:38:37:24:60:e6:f0:1e:88
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsoft ECC Root Certificate Authority 2017 O=Microsoft Corporation
+# Subject: CN=Microsoft ECC Root Certificate Authority 2017 O=Microsoft Corporation
+# Label: "Microsoft ECC Root Certificate Authority 2017"
+# Serial: 136839042543790627607696632466672567020
+# MD5 Fingerprint: dd:a1:03:e6:4a:93:10:d1:bf:f0:19:42:cb:fe:ed:67
+# SHA1 Fingerprint: 99:9a:64:c3:7f:f4:7d:9f:ab:95:f1:47:69:89:14:60:ee:c4:c3:c5
+# SHA256 Fingerprint: 35:8d:f3:9d:76:4a:f9:e1:b7:66:e9:c9:72:df:35:2e:e1:5c:fa:c2:27:af:6a:d1:d7:0e:8e:4a:6e:dc:ba:02
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Microsoft RSA Root Certificate Authority 2017 O=Microsoft Corporation
+# Subject: CN=Microsoft RSA Root Certificate Authority 2017 O=Microsoft Corporation
+# Label: "Microsoft RSA Root Certificate Authority 2017"
+# Serial: 40975477897264996090493496164228220339
+# MD5 Fingerprint: 10:ff:00:ff:cf:c9:f8:c7:7a:c0:ee:35:8e:c9:0f:47
+# SHA1 Fingerprint: 73:a5:e6:4a:3b:ff:83:16:ff:0e:dc:cc:61:8a:90:6e:4e:ae:4d:74
+# SHA256 Fingerprint: c7:41:f7:0f:4b:2a:8d:88:bf:2e:71:c1:41:22:ef:53:ef:10:eb:a0:cf:a5:e6:4c:fa:20:f4:18:85:30:73:e0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=e-Szigno Root CA 2017 O=Microsec Ltd.
+# Subject: CN=e-Szigno Root CA 2017 O=Microsec Ltd.
+# Label: "e-Szigno Root CA 2017"
+# Serial: 411379200276854331539784714
+# MD5 Fingerprint: de:1f:f6:9e:84:ae:a7:b4:21:ce:1e:58:7d:d1:84:98
+# SHA1 Fingerprint: 89:d4:83:03:4f:9e:9a:48:80:5f:72:37:d4:a9:a6:ef:cb:7c:1f:d1
+# SHA256 Fingerprint: be:b0:0b:30:83:9b:9b:c3:2c:32:e4:44:79:05:95:06:41:f2:64:21:b1:5e:d0:89:19:8b:51:8a:e2:ea:1b:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: O=CERTSIGN SA OU=certSIGN ROOT CA G2
+# Subject: O=CERTSIGN SA OU=certSIGN ROOT CA G2
+# Label: "certSIGN Root CA G2"
+# Serial: 313609486401300475190
+# MD5 Fingerprint: 8c:f1:75:8a:c6:19:cf:94:b7:f7:65:20:87:c3:97:c7
+# SHA1 Fingerprint: 26:f9:93:b4:ed:3d:28:27:b0:b9:4b:a7:e9:15:1d:a3:8d:92:e5:32
+# SHA256 Fingerprint: 65:7c:fe:2f:a7:3f:aa:38:46:25:71:f3:32:a2:36:3a:46:fc:e7:02:09:51:71:07:02:cd:fb:b6:ee:da:33:05
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global Certification Authority"
+# Serial: 1846098327275375458322922162
+# MD5 Fingerprint: f8:1c:18:2d:2f:ba:5f:6d:a1:6c:bc:c7:ab:91:c7:0e
+# SHA1 Fingerprint: 2f:8f:36:4f:e1:58:97:44:21:59:87:a5:2a:9a:d0:69:95:26:7f:b5
+# SHA256 Fingerprint: 97:55:20:15:f5:dd:fc:3c:87:88:c0:06:94:45:55:40:88:94:45:00:84:f1:00:86:70:86:bc:1a:2b:b5:8d:c8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global ECC P256 Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global ECC P256 Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global ECC P256 Certification Authority"
+# Serial: 4151900041497450638097112925
+# MD5 Fingerprint: 5b:44:e3:8d:5d:36:86:26:e8:0d:05:d2:59:a7:83:54
+# SHA1 Fingerprint: b4:90:82:dd:45:0c:be:8b:5b:b1:66:d3:e2:a4:08:26:cd:ed:42:cf
+# SHA256 Fingerprint: 94:5b:bc:82:5e:a5:54:f4:89:d1:fd:51:a7:3d:df:2e:a6:24:ac:70:19:a0:52:05:22:5c:22:a7:8c:cf:a8:b4
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Trustwave Global ECC P384 Certification Authority O=Trustwave Holdings, Inc.
+# Subject: CN=Trustwave Global ECC P384 Certification Authority O=Trustwave Holdings, Inc.
+# Label: "Trustwave Global ECC P384 Certification Authority"
+# Serial: 2704997926503831671788816187
+# MD5 Fingerprint: ea:cf:60:c4:3b:b9:15:29:40:a1:97:ed:78:27:93:d6
+# SHA1 Fingerprint: e7:f3:a3:c8:cf:6f:c3:04:2e:6d:0e:67:32:c5:9e:68:95:0d:5e:d2
+# SHA256 Fingerprint: 55:90:38:59:c8:c0:c3:eb:b8:75:9e:ce:4e:25:57:22:5f:f5:75:8b:bd:38:eb:d4:82:76:60:1e:1b:d5:80:97
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=NAVER Global Root Certification Authority O=NAVER BUSINESS PLATFORM Corp.
+# Subject: CN=NAVER Global Root Certification Authority O=NAVER BUSINESS PLATFORM Corp.
+# Label: "NAVER Global Root Certification Authority"
+# Serial: 9013692873798656336226253319739695165984492813
+# MD5 Fingerprint: c8:7e:41:f6:25:3b:f5:09:b3:17:e8:46:3d:bf:d0:9b
+# SHA1 Fingerprint: 8f:6b:f2:a9:27:4a:da:14:a0:c4:f4:8e:61:27:f9:c0:1e:78:5d:d1
+# SHA256 Fingerprint: 88:f4:38:dc:f8:ff:d1:fa:8f:42:91:15:ff:e5:f8:2a:e1:e0:6e:0c:70:c3:75:fa:ad:71:7b:34:a4:9e:72:65
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=AC RAIZ FNMT-RCM SERVIDORES SEGUROS O=FNMT-RCM OU=Ceres
+# Subject: CN=AC RAIZ FNMT-RCM SERVIDORES SEGUROS O=FNMT-RCM OU=Ceres
+# Label: "AC RAIZ FNMT-RCM SERVIDORES SEGUROS"
+# Serial: 131542671362353147877283741781055151509
+# MD5 Fingerprint: 19:36:9c:52:03:2f:d2:d1:bb:23:cc:dd:1e:12:55:bb
+# SHA1 Fingerprint: 62:ff:d9:9e:c0:65:0d:03:ce:75:93:d2:ed:3f:2d:32:c9:e3:e5:4a
+# SHA256 Fingerprint: 55:41:53:b1:3d:2c:f9:dd:b7:53:bf:be:1a:4e:0a:e0:8d:0a:a4:18:70:58:fe:60:a2:b8:62:b2:e4:b8:7b:cb
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign Root R46 O=GlobalSign nv-sa
+# Subject: CN=GlobalSign Root R46 O=GlobalSign nv-sa
+# Label: "GlobalSign Root R46"
+# Serial: 1552617688466950547958867513931858518042577
+# MD5 Fingerprint: c4:14:30:e4:fa:66:43:94:2a:6a:1b:24:5f:19:d0:ef
+# SHA1 Fingerprint: 53:a2:b0:4b:ca:6b:d6:45:e6:39:8a:8e:c4:0d:d2:bf:77:c3:a2:90
+# SHA256 Fingerprint: 4f:a3:12:6d:8d:3a:11:d1:c4:85:5a:4f:80:7c:ba:d6:cf:91:9d:3a:5a:88:b0:3b:ea:2c:63:72:d9:3c:40:c9
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign Root E46 O=GlobalSign nv-sa
+# Subject: CN=GlobalSign Root E46 O=GlobalSign nv-sa
+# Label: "GlobalSign Root E46"
+# Serial: 1552617690338932563915843282459653771421763
+# MD5 Fingerprint: b5:b8:66:ed:de:08:83:e3:c9:e2:01:34:06:ac:51:6f
+# SHA1 Fingerprint: 39:b4:6c:d5:fe:80:06:eb:e2:2f:4a:bb:08:33:a0:af:db:b9:dd:84
+# SHA256 Fingerprint: cb:b9:c4:4d:84:b8:04:3e:10:50:ea:31:a6:9f:51:49:55:d7:bf:d2:e2:c6:b4:93:01:01:9a:d6:1d:9f:50:58
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=ANF Secure Server Root CA O=ANF Autoridad de Certificacion OU=ANF CA Raiz
+# Subject: CN=ANF Secure Server Root CA O=ANF Autoridad de Certificacion OU=ANF CA Raiz
+# Label: "ANF Secure Server Root CA"
+# Serial: 996390341000653745
+# MD5 Fingerprint: 26:a6:44:5a:d9:af:4e:2f:b2:1d:b6:65:b0:4e:e8:96
+# SHA1 Fingerprint: 5b:6e:68:d0:cc:15:b6:a0:5f:1e:c1:5f:ae:02:fc:6b:2f:5d:6f:74
+# SHA256 Fingerprint: fb:8f:ec:75:91:69:b9:10:6b:1e:51:16:44:c6:18:c5:13:04:37:3f:6c:06:43:08:8d:8b:ef:fd:1b:99:75:99
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum EC-384 CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Subject: CN=Certum EC-384 CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Label: "Certum EC-384 CA"
+# Serial: 160250656287871593594747141429395092468
+# MD5 Fingerprint: b6:65:b3:96:60:97:12:a1:ec:4e:e1:3d:a3:c6:c9:f1
+# SHA1 Fingerprint: f3:3e:78:3c:ac:df:f4:a2:cc:ac:67:55:69:56:d7:e5:16:3c:e1:ed
+# SHA256 Fingerprint: 6b:32:80:85:62:53:18:aa:50:d1:73:c9:8d:8b:da:09:d5:7e:27:41:3d:11:4c:f7:87:a0:f5:d0:6c:03:0c:f6
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certum Trusted Root CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Subject: CN=Certum Trusted Root CA O=Asseco Data Systems S.A. OU=Certum Certification Authority
+# Label: "Certum Trusted Root CA"
+# Serial: 40870380103424195783807378461123655149
+# MD5 Fingerprint: 51:e1:c2:e7:fe:4c:84:af:59:0e:2f:f4:54:6f:ea:29
+# SHA1 Fingerprint: c8:83:44:c0:18:ae:9f:cc:f1:87:b7:8f:22:d1:c5:d7:45:84:ba:e5
+# SHA256 Fingerprint: fe:76:96:57:38:55:77:3e:37:a9:5e:7a:d4:d9:cc:96:c3:01:57:c1:5d:31:76:5b:a9:b1:57:04:e1:ae:78:fd
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TunTrust Root CA O=Agence Nationale de Certification Electronique
+# Subject: CN=TunTrust Root CA O=Agence Nationale de Certification Electronique
+# Label: "TunTrust Root CA"
+# Serial: 108534058042236574382096126452369648152337120275
+# MD5 Fingerprint: 85:13:b9:90:5b:36:5c:b6:5e:b8:5a:f8:e0:31:57:b4
+# SHA1 Fingerprint: cf:e9:70:84:0f:e0:73:0f:9d:f6:0c:7f:2c:4b:ee:20:46:34:9c:bb
+# SHA256 Fingerprint: 2e:44:10:2a:b5:8c:b8:54:19:45:1c:8e:19:d9:ac:f3:66:2c:af:bc:61:4b:6a:53:96:0a:30:f7:d0:e2:eb:41
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=HARICA TLS RSA Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Subject: CN=HARICA TLS RSA Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Label: "HARICA TLS RSA Root CA 2021"
+# Serial: 76817823531813593706434026085292783742
+# MD5 Fingerprint: 65:47:9b:58:86:dd:2c:f0:fc:a2:84:1f:1e:96:c4:91
+# SHA1 Fingerprint: 02:2d:05:82:fa:88:ce:14:0c:06:79:de:7f:14:10:e9:45:d7:a5:6d
+# SHA256 Fingerprint: d9:5d:0e:8e:da:79:52:5b:f9:be:b1:1b:14:d2:10:0d:32:94:98:5f:0c:62:d9:fa:bd:9c:d9:99:ec:cb:7b:1d
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=HARICA TLS ECC Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Subject: CN=HARICA TLS ECC Root CA 2021 O=Hellenic Academic and Research Institutions CA
+# Label: "HARICA TLS ECC Root CA 2021"
+# Serial: 137515985548005187474074462014555733966
+# MD5 Fingerprint: ae:f7:4c:e5:66:35:d1:b7:9b:8c:22:93:74:d3:4b:b0
+# SHA1 Fingerprint: bc:b0:c1:9d:e9:98:92:70:19:38:57:e9:8d:a7:b4:5d:6e:ee:01:48
+# SHA256 Fingerprint: 3f:99:cc:47:4a:cf:ce:4d:fe:d5:87:94:66:5e:47:8d:15:47:73:9f:2e:78:0f:1b:b4:ca:9b:13:30:97:d4:01
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Autoridad de Certificacion Firmaprofesional CIF A62634068
+# Subject: CN=Autoridad de Certificacion Firmaprofesional CIF A62634068
+# Label: "Autoridad de Certificacion Firmaprofesional CIF A62634068"
+# Serial: 1977337328857672817
+# MD5 Fingerprint: 4e:6e:9b:54:4c:ca:b7:fa:48:e4:90:b1:15:4b:1c:a3
+# SHA1 Fingerprint: 0b:be:c2:27:22:49:cb:39:aa:db:35:5c:53:e3:8c:ae:78:ff:b6:fe
+# SHA256 Fingerprint: 57:de:05:83:ef:d2:b2:6e:03:61:da:99:da:9d:f4:64:8d:ef:7e:e8:44:1c:3b:72:8a:fa:9b:cd:e0:f9:b2:6a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=vTrus ECC Root CA O=iTrusChina Co.,Ltd.
+# Subject: CN=vTrus ECC Root CA O=iTrusChina Co.,Ltd.
+# Label: "vTrus ECC Root CA"
+# Serial: 630369271402956006249506845124680065938238527194
+# MD5 Fingerprint: de:4b:c1:f5:52:8c:9b:43:e1:3e:8f:55:54:17:8d:85
+# SHA1 Fingerprint: f6:9c:db:b0:fc:f6:02:13:b6:52:32:a6:a3:91:3f:16:70:da:c3:e1
+# SHA256 Fingerprint: 30:fb:ba:2c:32:23:8e:2a:98:54:7a:f9:79:31:e5:50:42:8b:9b:3f:1c:8e:eb:66:33:dc:fa:86:c5:b2:7d:d3
+-----BEGIN CERTIFICATE-----
+MIICDzCCAZWgAwIBAgIUbmq8WapTvpg5Z6LSa6Q75m0c1towCgYIKoZIzj0EAwMw
+RzELMAkGA1UEBhMCQ04xHDAaBgNVBAoTE2lUcnVzQ2hpbmEgQ28uLEx0ZC4xGjAY
+BgNVBAMTEXZUcnVzIEVDQyBSb290IENBMB4XDTE4MDczMTA3MjY0NFoXDTQzMDcz
+MTA3MjY0NFowRzELMAkGA1UEBhMCQ04xHDAaBgNVBAoTE2lUcnVzQ2hpbmEgQ28u
+LEx0ZC4xGjAYBgNVBAMTEXZUcnVzIEVDQyBSb290IENBMHYwEAYHKoZIzj0CAQYF
+K4EEACIDYgAEZVBKrox5lkqqHAjDo6LN/llWQXf9JpRCux3NCNtzslt188+cToL0
+v/hhJoVs1oVbcnDS/dtitN9Ti72xRFhiQgnH+n9bEOf+QP3A2MMrMudwpremIFUd
+e4BdS49nTPEQo0IwQDAdBgNVHQ4EFgQUmDnNvtiyjPeyq+GtJK97fKHbH88wDwYD
+VR0TAQH/BAUwAwEB/zAOBgNVHQ8BAf8EBAMCAQYwCgYIKoZIzj0EAwMDaAAwZQIw
+V53dVvHH4+m4SVBrm2nDb+zDfSXkV5UTQJtS0zvzQBm8JsctBp61ezaf9SXUY2sA
+AjEA6dPGnlaaKsyh2j/IZivTWJwghfqrkYpwcBE4YGQLYgmRWAD5Tfs0aNoJrSEG
+GJTO
+-----END CERTIFICATE-----
+
+# Issuer: CN=vTrus Root CA O=iTrusChina Co.,Ltd.
+# Subject: CN=vTrus Root CA O=iTrusChina Co.,Ltd.
+# Label: "vTrus Root CA"
+# Serial: 387574501246983434957692974888460947164905180485
+# MD5 Fingerprint: b8:c9:37:df:fa:6b:31:84:64:c5:ea:11:6a:1b:75:fc
+# SHA1 Fingerprint: 84:1a:69:fb:f5:cd:1a:25:34:13:3d:e3:f8:fc:b8:99:d0:c9:14:b7
+# SHA256 Fingerprint: 8a:71:de:65:59:33:6f:42:6c:26:e5:38:80:d0:0d:88:a1:8d:a4:c6:a9:1f:0d:cb:61:94:e2:06:c5:c9:63:87
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=ISRG Root X2 O=Internet Security Research Group
+# Subject: CN=ISRG Root X2 O=Internet Security Research Group
+# Label: "ISRG Root X2"
+# Serial: 87493402998870891108772069816698636114
+# MD5 Fingerprint: d3:9e:c4:1e:23:3c:a6:df:cf:a3:7e:6d:e0:14:e6:e5
+# SHA1 Fingerprint: bd:b1:b9:3c:d5:97:8d:45:c6:26:14:55:f8:db:95:c7:5a:d1:53:af
+# SHA256 Fingerprint: 69:72:9b:8e:15:a8:6e:fc:17:7a:57:af:b7:17:1d:fc:64:ad:d2:8c:2f:ca:8c:f1:50:7e:34:45:3c:cb:14:70
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=HiPKI Root CA - G1 O=Chunghwa Telecom Co., Ltd.
+# Subject: CN=HiPKI Root CA - G1 O=Chunghwa Telecom Co., Ltd.
+# Label: "HiPKI Root CA - G1"
+# Serial: 60966262342023497858655262305426234976
+# MD5 Fingerprint: 69:45:df:16:65:4b:e8:68:9a:8f:76:5f:ff:80:9e:d3
+# SHA1 Fingerprint: 6a:92:e4:a8:ee:1b:ec:96:45:37:e3:29:57:49:cd:96:e3:e5:d2:60
+# SHA256 Fingerprint: f0:15:ce:3c:c2:39:bf:ef:06:4b:e9:f1:d2:c4:17:e1:a0:26:4a:0a:94:be:1f:0c:8d:12:18:64:eb:69:49:cc
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R4
+# Subject: CN=GlobalSign O=GlobalSign OU=GlobalSign ECC Root CA - R4
+# Label: "GlobalSign ECC Root CA - R4"
+# Serial: 159662223612894884239637590694
+# MD5 Fingerprint: 26:29:f8:6d:e1:88:bf:a2:65:7f:aa:c4:cd:0f:7f:fc
+# SHA1 Fingerprint: 6b:a0:b0:98:e1:71:ef:5a:ad:fe:48:15:80:77:10:f4:bd:6f:0b:28
+# SHA256 Fingerprint: b0:85:d7:0b:96:4f:19:1a:73:e4:af:0d:54:ae:7a:0e:07:aa:fd:af:9b:71:dd:08:62:13:8a:b7:32:5a:24:a2
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R1 O=Google Trust Services LLC
+# Subject: CN=GTS Root R1 O=Google Trust Services LLC
+# Label: "GTS Root R1"
+# Serial: 159662320309726417404178440727
+# MD5 Fingerprint: 05:fe:d0:bf:71:a8:a3:76:63:da:01:e0:d8:52:dc:40
+# SHA1 Fingerprint: e5:8c:1c:c4:91:3b:38:63:4b:e9:10:6e:e3:ad:8e:6b:9d:d9:81:4a
+# SHA256 Fingerprint: d9:47:43:2a:bd:e7:b7:fa:90:fc:2e:6b:59:10:1b:12:80:e0:e1:c7:e4:e4:0f:a3:c6:88:7f:ff:57:a7:f4:cf
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R2 O=Google Trust Services LLC
+# Subject: CN=GTS Root R2 O=Google Trust Services LLC
+# Label: "GTS Root R2"
+# Serial: 159662449406622349769042896298
+# MD5 Fingerprint: 1e:39:c0:53:e6:1e:29:82:0b:ca:52:55:36:5d:57:dc
+# SHA1 Fingerprint: 9a:44:49:76:32:db:de:fa:d0:bc:fb:5a:7b:17:bd:9e:56:09:24:94
+# SHA256 Fingerprint: 8d:25:cd:97:22:9d:bf:70:35:6b:da:4e:b3:cc:73:40:31:e2:4c:f0:0f:af:cf:d3:2d:c7:6e:b5:84:1c:7e:a8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R3 O=Google Trust Services LLC
+# Subject: CN=GTS Root R3 O=Google Trust Services LLC
+# Label: "GTS Root R3"
+# Serial: 159662495401136852707857743206
+# MD5 Fingerprint: 3e:e7:9d:58:02:94:46:51:94:e5:e0:22:4a:8b:e7:73
+# SHA1 Fingerprint: ed:e5:71:80:2b:c8:92:b9:5b:83:3c:d2:32:68:3f:09:cd:a0:1e:46
+# SHA256 Fingerprint: 34:d8:a7:3e:e2:08:d9:bc:db:0d:95:65:20:93:4b:4e:40:e6:94:82:59:6e:8b:6f:73:c8:42:6b:01:0a:6f:48
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=GTS Root R4 O=Google Trust Services LLC
+# Subject: CN=GTS Root R4 O=Google Trust Services LLC
+# Label: "GTS Root R4"
+# Serial: 159662532700760215368942768210
+# MD5 Fingerprint: 43:96:83:77:19:4d:76:b3:9d:65:52:e4:1d:22:a5:e8
+# SHA1 Fingerprint: 77:d3:03:67:b5:e0:0c:15:f6:0c:38:61:df:7c:e1:3b:92:46:4d:47
+# SHA256 Fingerprint: 34:9d:fa:40:58:c5:e2:63:12:3b:39:8a:e7:95:57:3c:4e:13:13:c8:3f:e6:8f:93:55:6c:d5:e8:03:1b:3c:7d
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=Telia Root CA v2 O=Telia Finland Oyj
+# Subject: CN=Telia Root CA v2 O=Telia Finland Oyj
+# Label: "Telia Root CA v2"
+# Serial: 7288924052977061235122729490515358
+# MD5 Fingerprint: 0e:8f:ac:aa:82:df:85:b1:f4:dc:10:1c:fc:99:d9:48
+# SHA1 Fingerprint: b9:99:cd:d1:73:50:8a:c4:47:05:08:9c:8c:88:fb:be:a0:2b:40:cd
+# SHA256 Fingerprint: 24:2b:69:74:2f:cb:1e:5b:2a:bf:98:89:8b:94:57:21:87:54:4e:5b:4d:99:11:78:65:73:62:1f:6a:74:b8:2c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST BR Root CA 1 2020 O=D-Trust GmbH
+# Subject: CN=D-TRUST BR Root CA 1 2020 O=D-Trust GmbH
+# Label: "D-TRUST BR Root CA 1 2020"
+# Serial: 165870826978392376648679885835942448534
+# MD5 Fingerprint: b5:aa:4b:d5:ed:f7:e3:55:2e:8f:72:0a:f3:75:b8:ed
+# SHA1 Fingerprint: 1f:5b:98:f0:e3:b5:f7:74:3c:ed:e6:b0:36:7d:32:cd:f4:09:41:67
+# SHA256 Fingerprint: e5:9a:aa:81:60:09:c2:2b:ff:5b:25:ba:d3:7d:f3:06:f0:49:79:7c:1f:81:d8:5a:b0:89:e6:57:bd:8f:00:44
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=D-TRUST EV Root CA 1 2020 O=D-Trust GmbH
+# Subject: CN=D-TRUST EV Root CA 1 2020 O=D-Trust GmbH
+# Label: "D-TRUST EV Root CA 1 2020"
+# Serial: 126288379621884218666039612629459926992
+# MD5 Fingerprint: 8c:2d:9d:70:9f:48:99:11:06:11:fb:e9:cb:30:c0:6e
+# SHA1 Fingerprint: 61:db:8c:21:59:69:03:90:d8:7c:9c:12:86:54:cf:9d:3d:f4:dd:07
+# SHA256 Fingerprint: 08:17:0d:1a:a3:64:53:90:1a:2f:95:92:45:e3:47:db:0c:8d:37:ab:aa:bc:56:b8:1a:a1:00:dc:95:89:70:db
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert TLS ECC P384 Root G5 O=DigiCert, Inc.
+# Subject: CN=DigiCert TLS ECC P384 Root G5 O=DigiCert, Inc.
+# Label: "DigiCert TLS ECC P384 Root G5"
+# Serial: 13129116028163249804115411775095713523
+# MD5 Fingerprint: d3:71:04:6a:43:1c:db:a6:59:e1:a8:a3:aa:c5:71:ed
+# SHA1 Fingerprint: 17:f3:de:5e:9f:0f:19:e9:8e:f6:1f:32:26:6e:20:c4:07:ae:30:ee
+# SHA256 Fingerprint: 01:8e:13:f0:77:25:32:cf:80:9b:d1:b1:72:81:86:72:83:fc:48:c6:e1:3b:e9:c6:98:12:85:4a:49:0c:1b:05
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=DigiCert TLS RSA4096 Root G5 O=DigiCert, Inc.
+# Subject: CN=DigiCert TLS RSA4096 Root G5 O=DigiCert, Inc.
+# Label: "DigiCert TLS RSA4096 Root G5"
+# Serial: 11930366277458970227240571539258396554
+# MD5 Fingerprint: ac:fe:f7:34:96:a9:f2:b3:b4:12:4b:e4:27:41:6f:e1
+# SHA1 Fingerprint: a7:88:49:dc:5d:7c:75:8c:8c:de:39:98:56:b3:aa:d0:b2:a5:71:35
+# SHA256 Fingerprint: 37:1a:00:dc:05:33:b3:72:1a:7e:eb:40:e8:41:9e:70:79:9d:2b:0a:0f:2c:1d:80:69:31:65:f7:ce:c4:ad:75
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certainly Root R1 O=Certainly
+# Subject: CN=Certainly Root R1 O=Certainly
+# Label: "Certainly Root R1"
+# Serial: 188833316161142517227353805653483829216
+# MD5 Fingerprint: 07:70:d4:3e:82:87:a0:fa:33:36:13:f4:fa:33:e7:12
+# SHA1 Fingerprint: a0:50:ee:0f:28:71:f4:27:b2:12:6d:6f:50:96:25:ba:cc:86:42:af
+# SHA256 Fingerprint: 77:b8:2c:d8:64:4c:43:05:f7:ac:c5:cb:15:6b:45:67:50:04:03:3d:51:c6:0c:62:02:a8:e0:c3:34:67:d3:a0
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Certainly Root E1 O=Certainly
+# Subject: CN=Certainly Root E1 O=Certainly
+# Label: "Certainly Root E1"
+# Serial: 8168531406727139161245376702891150584
+# MD5 Fingerprint: 0a:9e:ca:cd:3e:52:50:c6:36:f3:4b:a3:ed:a7:53:e9
+# SHA1 Fingerprint: f9:e1:6d:dc:01:89:cf:d5:82:45:63:3e:c5:37:7d:c2:eb:93:6f:2b
+# SHA256 Fingerprint: b4:58:5f:22:e4:ac:75:6a:4e:86:12:a1:36:1c:5d:9d:03:1a:93:fd:84:fe:bb:77:8f:a3:06:8b:0f:c4:2d:c2
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Security Communication RootCA3 O=SECOM Trust Systems CO.,LTD.
+# Subject: CN=Security Communication RootCA3 O=SECOM Trust Systems CO.,LTD.
+# Label: "Security Communication RootCA3"
+# Serial: 16247922307909811815
+# MD5 Fingerprint: 1c:9a:16:ff:9e:5c:e0:4d:8a:14:01:f4:35:5d:29:26
+# SHA1 Fingerprint: c3:03:c8:22:74:92:e5:61:a2:9c:5f:79:91:2b:1e:44:13:91:30:3a
+# SHA256 Fingerprint: 24:a5:5c:2a:b0:51:44:2d:06:17:76:65:41:23:9a:4a:d0:32:d7:c5:51:75:aa:34:ff:de:2f:bc:4f:5c:52:94
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Security Communication ECC RootCA1 O=SECOM Trust Systems CO.,LTD.
+# Subject: CN=Security Communication ECC RootCA1 O=SECOM Trust Systems CO.,LTD.
+# Label: "Security Communication ECC RootCA1"
+# Serial: 15446673492073852651
+# MD5 Fingerprint: 7e:43:b0:92:68:ec:05:43:4c:98:ab:5d:35:2e:7e:86
+# SHA1 Fingerprint: b8:0e:26:a9:bf:d2:b2:3b:c0:ef:46:c9:ba:c7:bb:f6:1d:0d:41:41
+# SHA256 Fingerprint: e7:4f:bd:a5:5b:d5:64:c4:73:a3:6b:44:1a:a7:99:c8:a6:8e:07:74:40:e8:28:8b:9f:a1:e5:0e:4b:ba:ca:11
+-----BEGIN CERTIFICATE-----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=
+-----END CERTIFICATE-----
+
+# Issuer: CN=BJCA Global Root CA1 O=BEIJING CERTIFICATE AUTHORITY
+# Subject: CN=BJCA Global Root CA1 O=BEIJING CERTIFICATE AUTHORITY
+# Label: "BJCA Global Root CA1"
+# Serial: 113562791157148395269083148143378328608
+# MD5 Fingerprint: 42:32:99:76:43:33:36:24:35:07:82:9b:28:f9:d0:90
+# SHA1 Fingerprint: d5:ec:8d:7b:4c:ba:79:f4:e7:e8:cb:9d:6b:ae:77:83:10:03:21:6a
+# SHA256 Fingerprint: f3:89:6f:88:fe:7c:0a:88:27:66:a7:fa:6a:d2:74:9f:b5:7a:7f:3e:98:fb:76:9c:1f:a7:b0:9c:2c:44:d5:ae
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=BJCA Global Root CA2 O=BEIJING CERTIFICATE AUTHORITY
+# Subject: CN=BJCA Global Root CA2 O=BEIJING CERTIFICATE AUTHORITY
+# Label: "BJCA Global Root CA2"
+# Serial: 58605626836079930195615843123109055211
+# MD5 Fingerprint: 5e:0a:f6:47:5f:a6:14:e8:11:01:95:3f:4d:01:eb:3c
+# SHA1 Fingerprint: f4:27:86:eb:6e:b8:6d:88:31:67:02:fb:ba:66:a4:53:00:aa:7a:a6
+# SHA256 Fingerprint: 57:4d:f6:93:1e:27:80:39:66:7b:72:0a:fd:c1:60:0f:c2:7e:b6:6d:d3:09:29:79:fb:73:85:64:87:21:28:82
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Sectigo Public Server Authentication Root E46 O=Sectigo Limited
+# Subject: CN=Sectigo Public Server Authentication Root E46 O=Sectigo Limited
+# Label: "Sectigo Public Server Authentication Root E46"
+# Serial: 88989738453351742415770396670917916916
+# MD5 Fingerprint: 28:23:f8:b2:98:5c:37:16:3b:3e:46:13:4e:b0:b3:01
+# SHA1 Fingerprint: ec:8a:39:6c:40:f0:2e:bc:42:75:d4:9f:ab:1c:1a:5b:67:be:d2:9a
+# SHA256 Fingerprint: c9:0f:26:f0:fb:1b:40:18:b2:22:27:51:9b:5c:a2:b5:3e:2c:a5:b3:be:5c:f1:8e:fe:1b:ef:47:38:0c:53:83
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Sectigo Public Server Authentication Root R46 O=Sectigo Limited
+# Subject: CN=Sectigo Public Server Authentication Root R46 O=Sectigo Limited
+# Label: "Sectigo Public Server Authentication Root R46"
+# Serial: 156256931880233212765902055439220583700
+# MD5 Fingerprint: 32:10:09:52:00:d5:7e:6c:43:df:15:c0:b1:16:93:e5
+# SHA1 Fingerprint: ad:98:f9:f3:e4:7d:75:3b:65:d4:82:b3:a4:52:17:bb:6e:f5:e4:38
+# SHA256 Fingerprint: 7b:b6:47:a6:2a:ee:ac:88:bf:25:7a:a5:22:d0:1f:fe:a3:95:e0:ab:45:c7:3f:93:f6:56:54:ec:38:f2:5a:06
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com TLS RSA Root CA 2022 O=SSL Corporation
+# Subject: CN=SSL.com TLS RSA Root CA 2022 O=SSL Corporation
+# Label: "SSL.com TLS RSA Root CA 2022"
+# Serial: 148535279242832292258835760425842727825
+# MD5 Fingerprint: d8:4e:c6:59:30:d8:fe:a0:d6:7a:5a:2c:2c:69:78:da
+# SHA1 Fingerprint: ec:2c:83:40:72:af:26:95:10:ff:0e:f2:03:ee:31:70:f6:78:9d:ca
+# SHA256 Fingerprint: 8f:af:7d:2e:2c:b4:70:9b:b8:e0:b3:36:66:bf:75:a5:dd:45:b5:de:48:0f:8e:a8:d4:bf:e6:be:bc:17:f2:ed
+-----BEGIN CERTIFICATE-----
+MIIFiTCCA3GgAwIBAgIQb77arXO9CEDii02+1PdbkTANBgkqhkiG9w0BAQsFADBO
+MQswCQYDVQQGEwJVUzEYMBYGA1UECgwPU1NMIENvcnBvcmF0aW9uMSUwIwYDVQQD
+DBxTU0wuY29tIFRMUyBSU0EgUm9vdCBDQSAyMDIyMB4XDTIyMDgyNTE2MzQyMloX
+DTQ2MDgxOTE2MzQyMVowTjELMAkGA1UEBhMCVVMxGDAWBgNVBAoMD1NTTCBDb3Jw
+b3JhdGlvbjElMCMGA1UEAwwcU1NMLmNvbSBUTFMgUlNBIFJvb3QgQ0EgMjAyMjCC
+AiIwDQYJKoZIhvcNAQEBBQADggIPADCCAgoCggIBANCkCXJPQIgSYT41I57u9nTP
+L3tYPc48DRAokC+X94xI2KDYJbFMsBFMF3NQ0CJKY7uB0ylu1bUJPiYYf7ISf5OY
+t6/wNr/y7hienDtSxUcZXXTzZGbVXcdotL8bHAajvI9AI7YexoS9UcQbOcGV0ins
+S657Lb85/bRi3pZ7QcacoOAGcvvwB5cJOYF0r/c0WRFXCsJbwST0MXMwgsadugL3
+PnxEX4MN8/HdIGkWCVDi1FW24IBydm5MR7d1VVm0U3TZlMZBrViKMWYPHqIbKUBO
+L9975hYsLfy/7PO0+r4Y9ptJ1O4Fbtk085zx7AGL0SDGD6C1vBdOSHtRwvzpXGk3
+R2azaPgVKPC506QVzFpPulJwoxJF3ca6TvvC0PeoUidtbnm1jPx7jMEWTO6Af77w
+dr5BUxIzrlo4QqvXDz5BjXYHMtWrifZOZ9mxQnUjbvPNQrL8VfVThxc7wDNY8VLS
++YCk8OjwO4s4zKTGkH8PnP2L0aPP2oOnaclQNtVcBdIKQXTbYxE3waWglksejBYS
+d66UNHsef8JmAOSqg+qKkK3ONkRN0VHpvB/zagX9wHQfJRlAUW7qglFA35u5CCoG
+AtUjHBPW6dvbxrB6y3snm/vg1UYk7RBLY0ulBY+6uB0rpvqR4pJSvezrZ5dtmi2f
+gTIFZzL7SAg/2SW4BCUvAgMBAAGjYzBhMA8GA1UdEwEB/wQFMAMBAf8wHwYDVR0j
+BBgwFoAU+y437uOEeicuzRk1sTN8/9REQrkwHQYDVR0OBBYEFPsuN+7jhHonLs0Z
+NbEzfP/UREK5MA4GA1UdDwEB/wQEAwIBhjANBgkqhkiG9w0BAQsFAAOCAgEAjYlt
+hEUY8U+zoO9opMAdrDC8Z2awms22qyIZZtM7QbUQnRC6cm4pJCAcAZli05bg4vsM
+QtfhWsSWTVTNj8pDU/0quOr4ZcoBwq1gaAafORpR2eCNJvkLTqVTJXojpBzOCBvf
+R4iyrT7gJ4eLSYwfqUdYe5byiB0YrrPRpgqU+tvT5TgKa3kSM/tKWTcWQA673vWJ
+DPFs0/dRa1419dvAJuoSc06pkZCmF8NsLzjUo3KUQyxi4U5cMj29TH0ZR6LDSeeW
+P4+a0zvkEdiLA9z2tmBVGKaBUfPhqBVq6+AL8BQx1rmMRTqoENjwuSfr98t67wVy
+lrXEj5ZzxOhWc5y8aVFjvO9nHEMaX3cZHxj4HCUp+UmZKbaSPaKDN7EgkaibMOlq
+bLQjk2UEqxHzDh1TJElTHaE/nUiSEeJ9DU/1172iWD54nR4fK/4huxoTtrEoZP2w
+AgDHbICivRZQIA9ygV/MlP+7mea6kMvq+cYMwq7FGc4zoWtcu358NFcXrfA/rs3q
+r5nsLFR+jM4uElZI7xc7P0peYNLcdDa8pUNjyw9bowJWCZ4kLOGGgYz+qxcs+sji
+Mho6/4UIyYOf8kpIEFR3N+2ivEC+5BB09+Rbu7nzifmPQdjH5FCQNYA+HLhNkNPU
+98OwoX6EyneSMSy4kLGCenROmxMmtNVQZlR4rmA=
+-----END CERTIFICATE-----
+
+# Issuer: CN=SSL.com TLS ECC Root CA 2022 O=SSL Corporation
+# Subject: CN=SSL.com TLS ECC Root CA 2022 O=SSL Corporation
+# Label: "SSL.com TLS ECC Root CA 2022"
+# Serial: 26605119622390491762507526719404364228
+# MD5 Fingerprint: 99:d7:5c:f1:51:36:cc:e9:ce:d9:19:2e:77:71:56:c5
+# SHA1 Fingerprint: 9f:5f:d9:1a:54:6d:f5:0c:71:f0:ee:7a:bd:17:49:98:84:73:e2:39
+# SHA256 Fingerprint: c3:2f:fd:9f:46:f9:36:d1:6c:36:73:99:09:59:43:4b:9a:d6:0a:af:bb:9e:7c:f3:36:54:f1:44:cc:1b:a1:43
+-----BEGIN CERTIFICATE-----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==
+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot Root CA ECC TLS 2021 O=Atos
+# Subject: CN=Atos TrustedRoot Root CA ECC TLS 2021 O=Atos
+# Label: "Atos TrustedRoot Root CA ECC TLS 2021"
+# Serial: 81873346711060652204712539181482831616
+# MD5 Fingerprint: 16:9f:ad:f1:70:ad:79:d6:ed:29:b4:d1:c5:79:70:a8
+# SHA1 Fingerprint: 9e:bc:75:10:42:b3:02:f3:81:f4:f7:30:62:d4:8f:c3:a7:51:b2:dd
+# SHA256 Fingerprint: b2:fa:e5:3e:14:cc:d7:ab:92:12:06:47:01:ae:27:9c:1d:89:88:fa:cb:77:5f:a8:a0:08:91:4e:66:39:88:a8
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Atos TrustedRoot Root CA RSA TLS 2021 O=Atos
+# Subject: CN=Atos TrustedRoot Root CA RSA TLS 2021 O=Atos
+# Label: "Atos TrustedRoot Root CA RSA TLS 2021"
+# Serial: 111436099570196163832749341232207667876
+# MD5 Fingerprint: d4:d3:46:b8:9a:c0:9c:76:5d:9e:3a:c3:b9:99:31:d2
+# SHA1 Fingerprint: 18:52:3b:0d:06:37:e4:d6:3a:df:23:e4:98:fb:5b:16:fb:86:74:48
+# SHA256 Fingerprint: 81:a9:08:8e:a5:9f:b3:64:c5:48:a6:f8:55:59:09:9b:6f:04:05:ef:bf:18:e5:32:4e:c9:f4:57:ba:00:11:2f
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TrustAsia Global Root CA G3 O=TrustAsia Technologies, Inc.
+# Subject: CN=TrustAsia Global Root CA G3 O=TrustAsia Technologies, Inc.
+# Label: "TrustAsia Global Root CA G3"
+# Serial: 576386314500428537169965010905813481816650257167
+# MD5 Fingerprint: 30:42:1b:b7:bb:81:75:35:e4:16:4f:53:d2:94:de:04
+# SHA1 Fingerprint: 63:cf:b6:c1:27:2b:56:e4:88:8e:1c:23:9a:b6:2e:81:47:24:c3:c7
+# SHA256 Fingerprint: e0:d3:22:6a:eb:11:63:c2:e4:8f:f9:be:3b:50:b4:c6:43:1b:e7:bb:1e:ac:c5:c3:6b:5d:5e:c5:09:03:9a:08
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TrustAsia Global Root CA G4 O=TrustAsia Technologies, Inc.
+# Subject: CN=TrustAsia Global Root CA G4 O=TrustAsia Technologies, Inc.
+# Label: "TrustAsia Global Root CA G4"
+# Serial: 451799571007117016466790293371524403291602933463
+# MD5 Fingerprint: 54:dd:b2:d7:5f:d8:3e:ed:7c:e0:0b:2e:cc:ed:eb:eb
+# SHA1 Fingerprint: 57:73:a5:61:5d:80:b2:e6:ac:38:82:fc:68:07:31:ac:9f:b5:92:5a
+# SHA256 Fingerprint: be:4b:56:cb:50:56:c0:13:6a:52:6d:f4:44:50:8d:aa:36:a0:b5:4f:42:e4:ac:38:f7:2a:f4:70:e4:79:65:4c
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust ECC Root-01 O=CommScope
+# Subject: CN=CommScope Public Trust ECC Root-01 O=CommScope
+# Label: "CommScope Public Trust ECC Root-01"
+# Serial: 385011430473757362783587124273108818652468453534
+# MD5 Fingerprint: 3a:40:a7:fc:03:8c:9c:38:79:2f:3a:a2:6c:b6:0a:16
+# SHA1 Fingerprint: 07:86:c0:d8:dd:8e:c0:80:98:06:98:d0:58:7a:ef:de:a6:cc:a2:5d
+# SHA256 Fingerprint: 11:43:7c:da:7b:b4:5e:41:36:5f:45:b3:9a:38:98:6b:0d:e0:0d:ef:34:8e:0c:7b:b0:87:36:33:80:0b:c3:8b
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust ECC Root-02 O=CommScope
+# Subject: CN=CommScope Public Trust ECC Root-02 O=CommScope
+# Label: "CommScope Public Trust ECC Root-02"
+# Serial: 234015080301808452132356021271193974922492992893
+# MD5 Fingerprint: 59:b0:44:d5:65:4d:b8:5c:55:19:92:02:b6:d1:94:b2
+# SHA1 Fingerprint: 3c:3f:ef:57:0f:fe:65:93:86:9e:a0:fe:b0:f6:ed:8e:d1:13:c7:e5
+# SHA256 Fingerprint: 2f:fb:7f:81:3b:bb:b3:c8:9a:b4:e8:16:2d:0f:16:d7:15:09:a8:30:cc:9d:73:c2:62:e5:14:08:75:d1:ad:4a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust RSA Root-01 O=CommScope
+# Subject: CN=CommScope Public Trust RSA Root-01 O=CommScope
+# Label: "CommScope Public Trust RSA Root-01"
+# Serial: 354030733275608256394402989253558293562031411421
+# MD5 Fingerprint: 0e:b4:15:bc:87:63:5d:5d:02:73:d4:26:38:68:73:d8
+# SHA1 Fingerprint: 6d:0a:5f:f7:b4:23:06:b4:85:b3:b7:97:64:fc:ac:75:f5:33:f2:93
+# SHA256 Fingerprint: 02:bd:f9:6e:2a:45:dd:9b:f1:8f:c7:e1:db:df:21:a0:37:9b:a3:c9:c2:61:03:44:cf:d8:d6:06:fe:c1:ed:81
+-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
+
+# Issuer: CN=CommScope Public Trust RSA Root-02 O=CommScope
+# Subject: CN=CommScope Public Trust RSA Root-02 O=CommScope
+# Label: "CommScope Public Trust RSA Root-02"
+# Serial: 480062499834624527752716769107743131258796508494
+# MD5 Fingerprint: e1:29:f9:62:7b:76:e2:96:6d:f3:d4:d7:0f:ae:1f:aa
+# SHA1 Fingerprint: ea:b0:e2:52:1b:89:93:4c:11:68:f2:d8:9a:ac:22:4c:a3:8a:57:ae
+# SHA256 Fingerprint: ff:e9:43:d7:93:42:4b:4f:7c:44:0c:1c:3d:64:8d:53:63:f3:4b:82:dc:87:aa:7a:9f:11:8f:c5:de:e1:01:f1
+-----BEGIN CERTIFICATE-----
+MIIFbDCCA1SgAwIBAgIUVBa/O345lXGN0aoApYYNK496BU4wDQYJKoZIhvcNAQEL
+BQAwTjELMAkGA1UEBhMCVVMxEjAQBgNVBAoMCUNvbW1TY29wZTErMCkGA1UEAwwi
+Q29tbVNjb3BlIFB1YmxpYyBUcnVzdCBSU0EgUm9vdC0wMjAeFw0yMTA0MjgxNzE2
+NDNaFw00NjA0MjgxNzE2NDJaME4xCzAJBgNVBAYTAlVTMRIwEAYDVQQKDAlDb21t
+U2NvcGUxKzApBgNVBAMMIkNvbW1TY29wZSBQdWJsaWMgVHJ1c3QgUlNBIFJvb3Qt
+MDIwggIiMA0GCSqGSIb3DQEBAQUAA4ICDwAwggIKAoICAQDh+g77aAASyE3VrCLE
+NQE7xVTlWXZjpX/rwcRqmL0yjReA61260WI9JSMZNRTpf4mnG2I81lDnNJUDMrG0
+kyI9p+Kx7eZ7Ti6Hmw0zdQreqjXnfuU2mKKuJZ6VszKWpCtYHu8//mI0SFHRtI1C
+rWDaSWqVcN3SAOLMV2MCe5bdSZdbkk6V0/nLKR8YSvgBKtJjCW4k6YnS5cciTNxz
+hkcAqg2Ijq6FfUrpuzNPDlJwnZXjfG2WWy09X6GDRl224yW4fKcZgBzqZUPckXk2
+LHR88mcGyYnJ27/aaL8j7dxrrSiDeS/sOKUNNwFnJ5rpM9kzXzehxfCrPfp4sOcs
+n/Y+n2Dg70jpkEUeBVF4GiwSLFworA2iI540jwXmojPOEXcT1A6kHkIfhs1w/tku
+FT0du7jyU1fbzMZ0KZwYszZ1OC4PVKH4kh+Jlk+71O6d6Ts2QrUKOyrUZHk2EOH5
+kQMreyBUzQ0ZGshBMjTRsJnhkB4BQDa1t/qp5Xd1pCKBXbCL5CcSD1SIxtuFdOa3
+wNemKfrb3vOTlycEVS8KbzfFPROvCgCpLIscgSjX74Yxqa7ybrjKaixUR9gqiC6v
+wQcQeKwRoi9C8DfF8rhW3Q5iLc4tVn5V8qdE9isy9COoR+jUKgF4z2rDN6ieZdIs
+5fq6M8EGRPbmz6UNp2YINIos8wIDAQABo0IwQDAPBgNVHRMBAf8EBTADAQH/MA4G
+A1UdDwEB/wQEAwIBBjAdBgNVHQ4EFgQUR9DnsSL/nSz12Vdgs7GxcJXvYXowDQYJ
+KoZIhvcNAQELBQADggIBAIZpsU0v6Z9PIpNojuQhmaPORVMbc0RTAIFhzTHjCLqB
+KCh6krm2qMhDnscTJk3C2OVVnJJdUNjCK9v+5qiXz1I6JMNlZFxHMaNlNRPDk7n3
++VGXu6TwYofF1gbTl4MgqX67tiHCpQ2EAOHyJxCDut0DgdXdaMNmEMjRdrSzbyme
+APnCKfWxkxlSaRosTKCL4BWaMS/TiJVZbuXEs1DIFAhKm4sTg7GkcrI7djNB3Nyq
+pgdvHSQSn8h2vS/ZjvQs7rfSOBAkNlEv41xdgSGn2rtO/+YHqP65DSdsu3BaVXoT
+6fEqSWnHX4dXTEN5bTpl6TBcQe7rd6VzEojov32u5cSoHw2OHG1QAk8mGEPej1WF
+sQs3BWDJVTkSBKEqz3EWnzZRSb9wO55nnPt7eck5HHisd5FUmrh1CoFSl+NmYWvt
+PjgelmFV4ZFUjO2MJB+ByRCac5krFk5yAD9UG/iNuovnFNa2RU9g7Jauwy8CTl2d
+lklyALKrdVwPaFsdZcJfMw8eD/A7hvWwTruc9+olBdytoptLFwG+Qt81IR2tq670
+v64fG9PiO/yzcnMcmyiQiRM9HcEARwmWmjgb3bHPDcK0RPOWlc4yOo80nOAXx17O
+rg3bhzjlP1v9mxnhMUF6cKojawHhRUzNlM47ni3niAIi9G7oyOzWPPO5std3eqx7
+-----END CERTIFICATE-----
+
+# Issuer: CN=Telekom Security TLS ECC Root 2020 O=Deutsche Telekom Security GmbH
+# Subject: CN=Telekom Security TLS ECC Root 2020 O=Deutsche Telekom Security GmbH
+# Label: "Telekom Security TLS ECC Root 2020"
+# Serial: 72082518505882327255703894282316633856
+# MD5 Fingerprint: c1:ab:fe:6a:10:2c:03:8d:bc:1c:22:32:c0:85:a7:fd
+# SHA1 Fingerprint: c0:f8:96:c5:a9:3b:01:06:21:07:da:18:42:48:bc:e9:9d:88:d5:ec
+# SHA256 Fingerprint: 57:8a:f4:de:d0:85:3f:4e:59:98:db:4a:ea:f9:cb:ea:8d:94:5f:60:b6:20:a3:8d:1a:3c:13:b2:bc:7b:a8:e1
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=Telekom Security TLS RSA Root 2023 O=Deutsche Telekom Security GmbH
+# Subject: CN=Telekom Security TLS RSA Root 2023 O=Deutsche Telekom Security GmbH
+# Label: "Telekom Security TLS RSA Root 2023"
+# Serial: 44676229530606711399881795178081572759
+# MD5 Fingerprint: bf:5b:eb:54:40:cd:48:71:c4:20:8d:7d:de:0a:42:f2
+# SHA1 Fingerprint: 54:d3:ac:b3:bd:57:56:f6:85:9d:ce:e5:c3:21:e2:d4:ad:83:d0:93
+# SHA256 Fingerprint: ef:c6:5c:ad:bb:59:ad:b6:ef:e8:4d:a2:23:11:b3:56:24:b7:1b:3b:1e:a0:da:8b:66:55:17:4e:c8:97:86:46
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=FIRMAPROFESIONAL CA ROOT-A WEB O=Firmaprofesional SA
+# Subject: CN=FIRMAPROFESIONAL CA ROOT-A WEB O=Firmaprofesional SA
+# Label: "FIRMAPROFESIONAL CA ROOT-A WEB"
+# Serial: 65916896770016886708751106294915943533
+# MD5 Fingerprint: 82:b2:ad:45:00:82:b0:66:63:f8:5f:c3:67:4e:ce:a3
+# SHA1 Fingerprint: a8:31:11:74:a6:14:15:0d:ca:77:dd:0e:e4:0c:5d:58:fc:a0:72:a5
+# SHA256 Fingerprint: be:f2:56:da:f2:6e:9c:69:bd:ec:16:02:35:97:98:f3:ca:f7:18:21:a0:3e:01:82:57:c5:3c:65:61:7f:3d:4a
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=TWCA CYBER Root CA O=TAIWAN-CA OU=Root CA
+# Subject: CN=TWCA CYBER Root CA O=TAIWAN-CA OU=Root CA
+# Label: "TWCA CYBER Root CA"
+# Serial: 85076849864375384482682434040119489222
+# MD5 Fingerprint: 0b:33:a0:97:52:95:d4:a9:fd:bb:db:6e:a3:55:5b:51
+# SHA1 Fingerprint: f6:b1:1c:1a:83:38:e9:7b:db:b3:a8:c8:33:24:e0:2d:9c:7f:26:66
+# SHA256 Fingerprint: 3f:63:bb:28:14:be:17:4e:c8:b6:43:9c:f0:8d:6d:56:f0:b7:c4:05:88:3a:56:48:a3:34:42:4d:6b:3e:c5:58
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA12 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA12 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA12"
+# Serial: 587887345431707215246142177076162061960426065942
+# MD5 Fingerprint: c6:89:ca:64:42:9b:62:08:49:0b:1e:7f:e9:07:3d:e8
+# SHA1 Fingerprint: 7a:22:1e:3d:de:1b:06:ac:9e:c8:47:70:16:8e:3c:e5:f7:6b:06:f4
+# SHA256 Fingerprint: 3f:03:4b:b5:70:4d:44:b2:d0:85:45:a0:20:57:de:93:eb:f3:90:5f:ce:72:1a:cb:c7:30:c0:6d:da:ee:90:4e
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA14 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA14 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA14"
+# Serial: 575790784512929437950770173562378038616896959179
+# MD5 Fingerprint: 71:0d:72:fa:92:19:65:5e:89:04:ac:16:33:f0:bc:d5
+# SHA1 Fingerprint: dd:50:c0:f7:79:b3:64:2e:74:a2:b8:9d:9f:d3:40:dd:bb:f0:f2:4f
+# SHA256 Fingerprint: 4b:00:9c:10:34:49:4f:9a:b5:6b:ba:3b:a1:d6:27:31:fc:4d:20:d8:95:5a:dc:ec:10:a9:25:60:72:61:e3:38
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----
+
+# Issuer: CN=SecureSign Root CA15 O=Cybertrust Japan Co., Ltd.
+# Subject: CN=SecureSign Root CA15 O=Cybertrust Japan Co., Ltd.
+# Label: "SecureSign Root CA15"
+# Serial: 126083514594751269499665114766174399806381178503
+# MD5 Fingerprint: 13:30:fc:c4:62:a6:a9:de:b5:c1:68:af:b5:d2:31:47
+# SHA1 Fingerprint: cb:ba:83:c8:c1:5a:5d:f1:f9:73:6f:ca:d7:ef:28:13:06:4a:07:7d
+# SHA256 Fingerprint: e7:78:f0:f0:95:fe:84:37:29:cd:1a:00:82:17:9e:53:14:a9:c2:91:44:28:05:e1:fb:1d:8f:b6:b8:88:6c:3a
+-----BEGIN CERTIFICATE-----
+MIICIzCCAamgAwIBAgIUFhXHw9hJp75pDIqI7fBw+d23PocwCgYIKoZIzj0EAwMw
+UTELMAkGA1UEBhMCSlAxIzAhBgNVBAoTGkN5YmVydHJ1c3QgSmFwYW4gQ28uLCBM
+dGQuMR0wGwYDVQQDExRTZWN1cmVTaWduIFJvb3QgQ0ExNTAeFw0yMDA0MDgwODMy
+NTZaFw00NTA0MDgwODMyNTZaMFExCzAJBgNVBAYTAkpQMSMwIQYDVQQKExpDeWJl
+cnRydXN0IEphcGFuIENvLiwgTHRkLjEdMBsGA1UEAxMUU2VjdXJlU2lnbiBSb290
+IENBMTUwdjAQBgcqhkjOPQIBBgUrgQQAIgNiAAQLUHSNZDKZmbPSYAi4Io5GdCx4
+wCtELW1fHcmuS1Iggz24FG1Th2CeX2yF2wYUleDHKP+dX+Sq8bOLbe1PL0vJSpSR
+ZHX+AezB2Ot6lHhWGENfa4HL9rzatAy2KZMIaY+jQjBAMA8GA1UdEwEB/wQFMAMB
+Af8wDgYDVR0PAQH/BAQDAgEGMB0GA1UdDgQWBBTrQciu/NWeUUj1vYv0hyCTQSvT
+9DAKBggqhkjOPQQDAwNoADBlAjEA2S6Jfl5OpBEHvVnCB96rMjhTKkZEBhd6zlHp
+4P9mLQlO4E/0BdGF9jVg3PVys0Z9AjBEmEYagoUeYWmJSwdLZrWeqrqgHkHZAXQ6
+bkU6iYAZezKYVWOr62Nuk22rGwlgMU4=
+-----END CERTIFICATE-----
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/core.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/core.py
new file mode 100644
index 00000000..70e0c3bd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/core.py
@@ -0,0 +1,114 @@
+"""
+certifi.py
+~~~~~~~~~~
+
+This module returns the installation location of cacert.pem or its contents.
+"""
+import sys
+import atexit
+
+def exit_cacert_ctx() -> None:
+    _CACERT_CTX.__exit__(None, None, None)  # type: ignore[union-attr]
+
+
+if sys.version_info >= (3, 11):
+
+    from importlib.resources import as_file, files
+
+    _CACERT_CTX = None
+    _CACERT_PATH = None
+
+    def where() -> str:
+        # This is slightly terrible, but we want to delay extracting the file
+        # in cases where we're inside of a zipimport situation until someone
+        # actually calls where(), but we don't want to re-extract the file
+        # on every call of where(), so we'll do it once then store it in a
+        # global variable.
+        global _CACERT_CTX
+        global _CACERT_PATH
+        if _CACERT_PATH is None:
+            # This is slightly janky, the importlib.resources API wants you to
+            # manage the cleanup of this file, so it doesn't actually return a
+            # path, it returns a context manager that will give you the path
+            # when you enter it and will do any cleanup when you leave it. In
+            # the common case of not needing a temporary file, it will just
+            # return the file system location and the __exit__() is a no-op.
+            #
+            # We also have to hold onto the actual context manager, because
+            # it will do the cleanup whenever it gets garbage collected, so
+            # we will also store that at the global level as well.
+            _CACERT_CTX = as_file(files("pip._vendor.certifi").joinpath("cacert.pem"))
+            _CACERT_PATH = str(_CACERT_CTX.__enter__())
+            atexit.register(exit_cacert_ctx)
+
+        return _CACERT_PATH
+
+    def contents() -> str:
+        return files("pip._vendor.certifi").joinpath("cacert.pem").read_text(encoding="ascii")
+
+elif sys.version_info >= (3, 7):
+
+    from importlib.resources import path as get_path, read_text
+
+    _CACERT_CTX = None
+    _CACERT_PATH = None
+
+    def where() -> str:
+        # This is slightly terrible, but we want to delay extracting the
+        # file in cases where we're inside of a zipimport situation until
+        # someone actually calls where(), but we don't want to re-extract
+        # the file on every call of where(), so we'll do it once then store
+        # it in a global variable.
+        global _CACERT_CTX
+        global _CACERT_PATH
+        if _CACERT_PATH is None:
+            # This is slightly janky, the importlib.resources API wants you
+            # to manage the cleanup of this file, so it doesn't actually
+            # return a path, it returns a context manager that will give
+            # you the path when you enter it and will do any cleanup when
+            # you leave it. In the common case of not needing a temporary
+            # file, it will just return the file system location and the
+            # __exit__() is a no-op.
+            #
+            # We also have to hold onto the actual context manager, because
+            # it will do the cleanup whenever it gets garbage collected, so
+            # we will also store that at the global level as well.
+            _CACERT_CTX = get_path("pip._vendor.certifi", "cacert.pem")
+            _CACERT_PATH = str(_CACERT_CTX.__enter__())
+            atexit.register(exit_cacert_ctx)
+
+        return _CACERT_PATH
+
+    def contents() -> str:
+        return read_text("pip._vendor.certifi", "cacert.pem", encoding="ascii")
+
+else:
+    import os
+    import types
+    from typing import Union
+
+    Package = Union[types.ModuleType, str]
+    Resource = Union[str, "os.PathLike"]
+
+    # This fallback will work for Python versions prior to 3.7 that lack the
+    # importlib.resources module but relies on the existing `where` function
+    # so won't address issues with environments like PyOxidizer that don't set
+    # __file__ on modules.
+    def read_text(
+        package: Package,
+        resource: Resource,
+        encoding: str = 'utf-8',
+        errors: str = 'strict'
+    ) -> str:
+        with open(where(), encoding=encoding) as data:
+            return data.read()
+
+    # If we don't have importlib.resources, then we will just do the old logic
+    # of assuming we're on the filesystem and munge the path directly.
+    def where() -> str:
+        f = os.path.dirname(__file__)
+
+        return os.path.join(f, "cacert.pem")
+
+    def contents() -> str:
+        return read_text("pip._vendor.certifi", "cacert.pem", encoding="ascii")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/certifi/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__init__.py
new file mode 100644
index 00000000..bf0d6c6d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__init__.py
@@ -0,0 +1,33 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+import logging
+
+__version__ = '0.3.9'
+
+
+class DistlibException(Exception):
+    pass
+
+
+try:
+    from logging import NullHandler
+except ImportError:  # pragma: no cover
+
+    class NullHandler(logging.Handler):
+
+        def handle(self, record):
+            pass
+
+        def emit(self, record):
+            pass
+
+        def createLock(self):
+            self.lock = None
+
+
+logger = logging.getLogger(__name__)
+logger.addHandler(NullHandler())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9f65cb23
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..8d2c071c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/database.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/database.cpython-312.pyc
new file mode 100644
index 00000000..5a212bbb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/database.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/index.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/index.cpython-312.pyc
new file mode 100644
index 00000000..45a03381
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/index.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/locators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/locators.cpython-312.pyc
new file mode 100644
index 00000000..3bc4b614
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/locators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/manifest.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/manifest.cpython-312.pyc
new file mode 100644
index 00000000..9ceb9922
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/manifest.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/markers.cpython-312.pyc
new file mode 100644
index 00000000..c34b6d14
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/metadata.cpython-312.pyc
new file mode 100644
index 00000000..80b0e092
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/resources.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/resources.cpython-312.pyc
new file mode 100644
index 00000000..4f187146
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/resources.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/scripts.cpython-312.pyc
new file mode 100644
index 00000000..b79100ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..e5d33e42
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..7a1bf173
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..85c98c45
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/compat.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/compat.py
new file mode 100644
index 00000000..ca561dd2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/compat.py
@@ -0,0 +1,1137 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2013-2017 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+from __future__ import absolute_import
+
+import os
+import re
+import shutil
+import sys
+
+try:
+    import ssl
+except ImportError:  # pragma: no cover
+    ssl = None
+
+if sys.version_info[0] < 3:  # pragma: no cover
+    from StringIO import StringIO
+    string_types = basestring,
+    text_type = unicode
+    from types import FileType as file_type
+    import __builtin__ as builtins
+    import ConfigParser as configparser
+    from urlparse import urlparse, urlunparse, urljoin, urlsplit, urlunsplit
+    from urllib import (urlretrieve, quote as _quote, unquote, url2pathname,
+                        pathname2url, ContentTooShortError, splittype)
+
+    def quote(s):
+        if isinstance(s, unicode):
+            s = s.encode('utf-8')
+        return _quote(s)
+
+    import urllib2
+    from urllib2 import (Request, urlopen, URLError, HTTPError,
+                         HTTPBasicAuthHandler, HTTPPasswordMgr, HTTPHandler,
+                         HTTPRedirectHandler, build_opener)
+    if ssl:
+        from urllib2 import HTTPSHandler
+    import httplib
+    import xmlrpclib
+    import Queue as queue
+    from HTMLParser import HTMLParser
+    import htmlentitydefs
+    raw_input = raw_input
+    from itertools import ifilter as filter
+    from itertools import ifilterfalse as filterfalse
+
+    # Leaving this around for now, in case it needs resurrecting in some way
+    # _userprog = None
+    # def splituser(host):
+    # """splituser('user[:passwd]@host[:port]') --> 'user[:passwd]', 'host[:port]'."""
+    # global _userprog
+    # if _userprog is None:
+    # import re
+    # _userprog = re.compile('^(.*)@(.*)$')
+
+    # match = _userprog.match(host)
+    # if match: return match.group(1, 2)
+    # return None, host
+
+else:  # pragma: no cover
+    from io import StringIO
+    string_types = str,
+    text_type = str
+    from io import TextIOWrapper as file_type
+    import builtins
+    import configparser
+    from urllib.parse import (urlparse, urlunparse, urljoin, quote, unquote,
+                              urlsplit, urlunsplit, splittype)
+    from urllib.request import (urlopen, urlretrieve, Request, url2pathname,
+                                pathname2url, HTTPBasicAuthHandler,
+                                HTTPPasswordMgr, HTTPHandler,
+                                HTTPRedirectHandler, build_opener)
+    if ssl:
+        from urllib.request import HTTPSHandler
+    from urllib.error import HTTPError, URLError, ContentTooShortError
+    import http.client as httplib
+    import urllib.request as urllib2
+    import xmlrpc.client as xmlrpclib
+    import queue
+    from html.parser import HTMLParser
+    import html.entities as htmlentitydefs
+    raw_input = input
+    from itertools import filterfalse
+    filter = filter
+
+try:
+    from ssl import match_hostname, CertificateError
+except ImportError:  # pragma: no cover
+
+    class CertificateError(ValueError):
+        pass
+
+    def _dnsname_match(dn, hostname, max_wildcards=1):
+        """Matching according to RFC 6125, section 6.4.3
+
+        http://tools.ietf.org/html/rfc6125#section-6.4.3
+        """
+        pats = []
+        if not dn:
+            return False
+
+        parts = dn.split('.')
+        leftmost, remainder = parts[0], parts[1:]
+
+        wildcards = leftmost.count('*')
+        if wildcards > max_wildcards:
+            # Issue #17980: avoid denials of service by refusing more
+            # than one wildcard per fragment.  A survey of established
+            # policy among SSL implementations showed it to be a
+            # reasonable choice.
+            raise CertificateError(
+                "too many wildcards in certificate DNS name: " + repr(dn))
+
+        # speed up common case w/o wildcards
+        if not wildcards:
+            return dn.lower() == hostname.lower()
+
+        # RFC 6125, section 6.4.3, subitem 1.
+        # The client SHOULD NOT attempt to match a presented identifier in which
+        # the wildcard character comprises a label other than the left-most label.
+        if leftmost == '*':
+            # When '*' is a fragment by itself, it matches a non-empty dotless
+            # fragment.
+            pats.append('[^.]+')
+        elif leftmost.startswith('xn--') or hostname.startswith('xn--'):
+            # RFC 6125, section 6.4.3, subitem 3.
+            # The client SHOULD NOT attempt to match a presented identifier
+            # where the wildcard character is embedded within an A-label or
+            # U-label of an internationalized domain name.
+            pats.append(re.escape(leftmost))
+        else:
+            # Otherwise, '*' matches any dotless string, e.g. www*
+            pats.append(re.escape(leftmost).replace(r'\*', '[^.]*'))
+
+        # add the remaining fragments, ignore any wildcards
+        for frag in remainder:
+            pats.append(re.escape(frag))
+
+        pat = re.compile(r'\A' + r'\.'.join(pats) + r'\Z', re.IGNORECASE)
+        return pat.match(hostname)
+
+    def match_hostname(cert, hostname):
+        """Verify that *cert* (in decoded format as returned by
+        SSLSocket.getpeercert()) matches the *hostname*.  RFC 2818 and RFC 6125
+        rules are followed, but IP addresses are not accepted for *hostname*.
+
+        CertificateError is raised on failure. On success, the function
+        returns nothing.
+        """
+        if not cert:
+            raise ValueError("empty or no certificate, match_hostname needs a "
+                             "SSL socket or SSL context with either "
+                             "CERT_OPTIONAL or CERT_REQUIRED")
+        dnsnames = []
+        san = cert.get('subjectAltName', ())
+        for key, value in san:
+            if key == 'DNS':
+                if _dnsname_match(value, hostname):
+                    return
+                dnsnames.append(value)
+        if not dnsnames:
+            # The subject is only checked when there is no dNSName entry
+            # in subjectAltName
+            for sub in cert.get('subject', ()):
+                for key, value in sub:
+                    # XXX according to RFC 2818, the most specific Common Name
+                    # must be used.
+                    if key == 'commonName':
+                        if _dnsname_match(value, hostname):
+                            return
+                        dnsnames.append(value)
+        if len(dnsnames) > 1:
+            raise CertificateError("hostname %r "
+                                   "doesn't match either of %s" %
+                                   (hostname, ', '.join(map(repr, dnsnames))))
+        elif len(dnsnames) == 1:
+            raise CertificateError("hostname %r "
+                                   "doesn't match %r" %
+                                   (hostname, dnsnames[0]))
+        else:
+            raise CertificateError("no appropriate commonName or "
+                                   "subjectAltName fields were found")
+
+
+try:
+    from types import SimpleNamespace as Container
+except ImportError:  # pragma: no cover
+
+    class Container(object):
+        """
+        A generic container for when multiple values need to be returned
+        """
+
+        def __init__(self, **kwargs):
+            self.__dict__.update(kwargs)
+
+
+try:
+    from shutil import which
+except ImportError:  # pragma: no cover
+    # Implementation from Python 3.3
+    def which(cmd, mode=os.F_OK | os.X_OK, path=None):
+        """Given a command, mode, and a PATH string, return the path which
+        conforms to the given mode on the PATH, or None if there is no such
+        file.
+
+        `mode` defaults to os.F_OK | os.X_OK. `path` defaults to the result
+        of os.environ.get("PATH"), or can be overridden with a custom search
+        path.
+
+        """
+
+        # Check that a given file can be accessed with the correct mode.
+        # Additionally check that `file` is not a directory, as on Windows
+        # directories pass the os.access check.
+        def _access_check(fn, mode):
+            return (os.path.exists(fn) and os.access(fn, mode) and not os.path.isdir(fn))
+
+        # If we're given a path with a directory part, look it up directly rather
+        # than referring to PATH directories. This includes checking relative to the
+        # current directory, e.g. ./script
+        if os.path.dirname(cmd):
+            if _access_check(cmd, mode):
+                return cmd
+            return None
+
+        if path is None:
+            path = os.environ.get("PATH", os.defpath)
+        if not path:
+            return None
+        path = path.split(os.pathsep)
+
+        if sys.platform == "win32":
+            # The current directory takes precedence on Windows.
+            if os.curdir not in path:
+                path.insert(0, os.curdir)
+
+            # PATHEXT is necessary to check on Windows.
+            pathext = os.environ.get("PATHEXT", "").split(os.pathsep)
+            # See if the given file matches any of the expected path extensions.
+            # This will allow us to short circuit when given "python.exe".
+            # If it does match, only test that one, otherwise we have to try
+            # others.
+            if any(cmd.lower().endswith(ext.lower()) for ext in pathext):
+                files = [cmd]
+            else:
+                files = [cmd + ext for ext in pathext]
+        else:
+            # On other platforms you don't have things like PATHEXT to tell you
+            # what file suffixes are executable, so just pass on cmd as-is.
+            files = [cmd]
+
+        seen = set()
+        for dir in path:
+            normdir = os.path.normcase(dir)
+            if normdir not in seen:
+                seen.add(normdir)
+                for thefile in files:
+                    name = os.path.join(dir, thefile)
+                    if _access_check(name, mode):
+                        return name
+        return None
+
+
+# ZipFile is a context manager in 2.7, but not in 2.6
+
+from zipfile import ZipFile as BaseZipFile
+
+if hasattr(BaseZipFile, '__enter__'):  # pragma: no cover
+    ZipFile = BaseZipFile
+else:  # pragma: no cover
+    from zipfile import ZipExtFile as BaseZipExtFile
+
+    class ZipExtFile(BaseZipExtFile):
+
+        def __init__(self, base):
+            self.__dict__.update(base.__dict__)
+
+        def __enter__(self):
+            return self
+
+        def __exit__(self, *exc_info):
+            self.close()
+            # return None, so if an exception occurred, it will propagate
+
+    class ZipFile(BaseZipFile):
+
+        def __enter__(self):
+            return self
+
+        def __exit__(self, *exc_info):
+            self.close()
+            # return None, so if an exception occurred, it will propagate
+
+        def open(self, *args, **kwargs):
+            base = BaseZipFile.open(self, *args, **kwargs)
+            return ZipExtFile(base)
+
+
+try:
+    from platform import python_implementation
+except ImportError:  # pragma: no cover
+
+    def python_implementation():
+        """Return a string identifying the Python implementation."""
+        if 'PyPy' in sys.version:
+            return 'PyPy'
+        if os.name == 'java':
+            return 'Jython'
+        if sys.version.startswith('IronPython'):
+            return 'IronPython'
+        return 'CPython'
+
+
+import sysconfig
+
+try:
+    callable = callable
+except NameError:  # pragma: no cover
+    from collections.abc import Callable
+
+    def callable(obj):
+        return isinstance(obj, Callable)
+
+
+try:
+    fsencode = os.fsencode
+    fsdecode = os.fsdecode
+except AttributeError:  # pragma: no cover
+    # Issue #99: on some systems (e.g. containerised),
+    # sys.getfilesystemencoding() returns None, and we need a real value,
+    # so fall back to utf-8. From the CPython 2.7 docs relating to Unix and
+    # sys.getfilesystemencoding(): the return value is "the user’s preference
+    # according to the result of nl_langinfo(CODESET), or None if the
+    # nl_langinfo(CODESET) failed."
+    _fsencoding = sys.getfilesystemencoding() or 'utf-8'
+    if _fsencoding == 'mbcs':
+        _fserrors = 'strict'
+    else:
+        _fserrors = 'surrogateescape'
+
+    def fsencode(filename):
+        if isinstance(filename, bytes):
+            return filename
+        elif isinstance(filename, text_type):
+            return filename.encode(_fsencoding, _fserrors)
+        else:
+            raise TypeError("expect bytes or str, not %s" %
+                            type(filename).__name__)
+
+    def fsdecode(filename):
+        if isinstance(filename, text_type):
+            return filename
+        elif isinstance(filename, bytes):
+            return filename.decode(_fsencoding, _fserrors)
+        else:
+            raise TypeError("expect bytes or str, not %s" %
+                            type(filename).__name__)
+
+
+try:
+    from tokenize import detect_encoding
+except ImportError:  # pragma: no cover
+    from codecs import BOM_UTF8, lookup
+
+    cookie_re = re.compile(r"coding[:=]\s*([-\w.]+)")
+
+    def _get_normal_name(orig_enc):
+        """Imitates get_normal_name in tokenizer.c."""
+        # Only care about the first 12 characters.
+        enc = orig_enc[:12].lower().replace("_", "-")
+        if enc == "utf-8" or enc.startswith("utf-8-"):
+            return "utf-8"
+        if enc in ("latin-1", "iso-8859-1", "iso-latin-1") or \
+           enc.startswith(("latin-1-", "iso-8859-1-", "iso-latin-1-")):
+            return "iso-8859-1"
+        return orig_enc
+
+    def detect_encoding(readline):
+        """
+        The detect_encoding() function is used to detect the encoding that should
+        be used to decode a Python source file.  It requires one argument, readline,
+        in the same way as the tokenize() generator.
+
+        It will call readline a maximum of twice, and return the encoding used
+        (as a string) and a list of any lines (left as bytes) it has read in.
+
+        It detects the encoding from the presence of a utf-8 bom or an encoding
+        cookie as specified in pep-0263.  If both a bom and a cookie are present,
+        but disagree, a SyntaxError will be raised.  If the encoding cookie is an
+        invalid charset, raise a SyntaxError.  Note that if a utf-8 bom is found,
+        'utf-8-sig' is returned.
+
+        If no encoding is specified, then the default of 'utf-8' will be returned.
+        """
+        try:
+            filename = readline.__self__.name
+        except AttributeError:
+            filename = None
+        bom_found = False
+        encoding = None
+        default = 'utf-8'
+
+        def read_or_stop():
+            try:
+                return readline()
+            except StopIteration:
+                return b''
+
+        def find_cookie(line):
+            try:
+                # Decode as UTF-8. Either the line is an encoding declaration,
+                # in which case it should be pure ASCII, or it must be UTF-8
+                # per default encoding.
+                line_string = line.decode('utf-8')
+            except UnicodeDecodeError:
+                msg = "invalid or missing encoding declaration"
+                if filename is not None:
+                    msg = '{} for {!r}'.format(msg, filename)
+                raise SyntaxError(msg)
+
+            matches = cookie_re.findall(line_string)
+            if not matches:
+                return None
+            encoding = _get_normal_name(matches[0])
+            try:
+                codec = lookup(encoding)
+            except LookupError:
+                # This behaviour mimics the Python interpreter
+                if filename is None:
+                    msg = "unknown encoding: " + encoding
+                else:
+                    msg = "unknown encoding for {!r}: {}".format(
+                        filename, encoding)
+                raise SyntaxError(msg)
+
+            if bom_found:
+                if codec.name != 'utf-8':
+                    # This behaviour mimics the Python interpreter
+                    if filename is None:
+                        msg = 'encoding problem: utf-8'
+                    else:
+                        msg = 'encoding problem for {!r}: utf-8'.format(
+                            filename)
+                    raise SyntaxError(msg)
+                encoding += '-sig'
+            return encoding
+
+        first = read_or_stop()
+        if first.startswith(BOM_UTF8):
+            bom_found = True
+            first = first[3:]
+            default = 'utf-8-sig'
+        if not first:
+            return default, []
+
+        encoding = find_cookie(first)
+        if encoding:
+            return encoding, [first]
+
+        second = read_or_stop()
+        if not second:
+            return default, [first]
+
+        encoding = find_cookie(second)
+        if encoding:
+            return encoding, [first, second]
+
+        return default, [first, second]
+
+
+# For converting & <-> &amp; etc.
+try:
+    from html import escape
+except ImportError:
+    from cgi import escape
+if sys.version_info[:2] < (3, 4):
+    unescape = HTMLParser().unescape
+else:
+    from html import unescape
+
+try:
+    from collections import ChainMap
+except ImportError:  # pragma: no cover
+    from collections import MutableMapping
+
+    try:
+        from reprlib import recursive_repr as _recursive_repr
+    except ImportError:
+
+        def _recursive_repr(fillvalue='...'):
+            '''
+            Decorator to make a repr function return fillvalue for a recursive
+            call
+            '''
+
+            def decorating_function(user_function):
+                repr_running = set()
+
+                def wrapper(self):
+                    key = id(self), get_ident()
+                    if key in repr_running:
+                        return fillvalue
+                    repr_running.add(key)
+                    try:
+                        result = user_function(self)
+                    finally:
+                        repr_running.discard(key)
+                    return result
+
+                # Can't use functools.wraps() here because of bootstrap issues
+                wrapper.__module__ = getattr(user_function, '__module__')
+                wrapper.__doc__ = getattr(user_function, '__doc__')
+                wrapper.__name__ = getattr(user_function, '__name__')
+                wrapper.__annotations__ = getattr(user_function,
+                                                  '__annotations__', {})
+                return wrapper
+
+            return decorating_function
+
+    class ChainMap(MutableMapping):
+        '''
+        A ChainMap groups multiple dicts (or other mappings) together
+        to create a single, updateable view.
+
+        The underlying mappings are stored in a list.  That list is public and can
+        accessed or updated using the *maps* attribute.  There is no other state.
+
+        Lookups search the underlying mappings successively until a key is found.
+        In contrast, writes, updates, and deletions only operate on the first
+        mapping.
+        '''
+
+        def __init__(self, *maps):
+            '''Initialize a ChainMap by setting *maps* to the given mappings.
+            If no mappings are provided, a single empty dictionary is used.
+
+            '''
+            self.maps = list(maps) or [{}]  # always at least one map
+
+        def __missing__(self, key):
+            raise KeyError(key)
+
+        def __getitem__(self, key):
+            for mapping in self.maps:
+                try:
+                    return mapping[
+                        key]  # can't use 'key in mapping' with defaultdict
+                except KeyError:
+                    pass
+            return self.__missing__(
+                key)  # support subclasses that define __missing__
+
+        def get(self, key, default=None):
+            return self[key] if key in self else default
+
+        def __len__(self):
+            return len(set().union(
+                *self.maps))  # reuses stored hash values if possible
+
+        def __iter__(self):
+            return iter(set().union(*self.maps))
+
+        def __contains__(self, key):
+            return any(key in m for m in self.maps)
+
+        def __bool__(self):
+            return any(self.maps)
+
+        @_recursive_repr()
+        def __repr__(self):
+            return '{0.__class__.__name__}({1})'.format(
+                self, ', '.join(map(repr, self.maps)))
+
+        @classmethod
+        def fromkeys(cls, iterable, *args):
+            'Create a ChainMap with a single dict created from the iterable.'
+            return cls(dict.fromkeys(iterable, *args))
+
+        def copy(self):
+            'New ChainMap or subclass with a new copy of maps[0] and refs to maps[1:]'
+            return self.__class__(self.maps[0].copy(), *self.maps[1:])
+
+        __copy__ = copy
+
+        def new_child(self):  # like Django's Context.push()
+            'New ChainMap with a new dict followed by all previous maps.'
+            return self.__class__({}, *self.maps)
+
+        @property
+        def parents(self):  # like Django's Context.pop()
+            'New ChainMap from maps[1:].'
+            return self.__class__(*self.maps[1:])
+
+        def __setitem__(self, key, value):
+            self.maps[0][key] = value
+
+        def __delitem__(self, key):
+            try:
+                del self.maps[0][key]
+            except KeyError:
+                raise KeyError(
+                    'Key not found in the first mapping: {!r}'.format(key))
+
+        def popitem(self):
+            'Remove and return an item pair from maps[0]. Raise KeyError is maps[0] is empty.'
+            try:
+                return self.maps[0].popitem()
+            except KeyError:
+                raise KeyError('No keys found in the first mapping.')
+
+        def pop(self, key, *args):
+            'Remove *key* from maps[0] and return its value. Raise KeyError if *key* not in maps[0].'
+            try:
+                return self.maps[0].pop(key, *args)
+            except KeyError:
+                raise KeyError(
+                    'Key not found in the first mapping: {!r}'.format(key))
+
+        def clear(self):
+            'Clear maps[0], leaving maps[1:] intact.'
+            self.maps[0].clear()
+
+
+try:
+    from importlib.util import cache_from_source  # Python >= 3.4
+except ImportError:  # pragma: no cover
+
+    def cache_from_source(path, debug_override=None):
+        assert path.endswith('.py')
+        if debug_override is None:
+            debug_override = __debug__
+        if debug_override:
+            suffix = 'c'
+        else:
+            suffix = 'o'
+        return path + suffix
+
+
+try:
+    from collections import OrderedDict
+except ImportError:  # pragma: no cover
+    # {{{ http://code.activestate.com/recipes/576693/ (r9)
+    # Backport of OrderedDict() class that runs on Python 2.4, 2.5, 2.6, 2.7 and pypy.
+    # Passes Python2.7's test suite and incorporates all the latest updates.
+    try:
+        from thread import get_ident as _get_ident
+    except ImportError:
+        from dummy_thread import get_ident as _get_ident
+
+    try:
+        from _abcoll import KeysView, ValuesView, ItemsView
+    except ImportError:
+        pass
+
+    class OrderedDict(dict):
+        'Dictionary that remembers insertion order'
+
+        # An inherited dict maps keys to values.
+        # The inherited dict provides __getitem__, __len__, __contains__, and get.
+        # The remaining methods are order-aware.
+        # Big-O running times for all methods are the same as for regular dictionaries.
+
+        # The internal self.__map dictionary maps keys to links in a doubly linked list.
+        # The circular doubly linked list starts and ends with a sentinel element.
+        # The sentinel element never gets deleted (this simplifies the algorithm).
+        # Each link is stored as a list of length three:  [PREV, NEXT, KEY].
+
+        def __init__(self, *args, **kwds):
+            '''Initialize an ordered dictionary.  Signature is the same as for
+            regular dictionaries, but keyword arguments are not recommended
+            because their insertion order is arbitrary.
+
+            '''
+            if len(args) > 1:
+                raise TypeError('expected at most 1 arguments, got %d' %
+                                len(args))
+            try:
+                self.__root
+            except AttributeError:
+                self.__root = root = []  # sentinel node
+                root[:] = [root, root, None]
+                self.__map = {}
+            self.__update(*args, **kwds)
+
+        def __setitem__(self, key, value, dict_setitem=dict.__setitem__):
+            'od.__setitem__(i, y) <==> od[i]=y'
+            # Setting a new item creates a new link which goes at the end of the linked
+            # list, and the inherited dictionary is updated with the new key/value pair.
+            if key not in self:
+                root = self.__root
+                last = root[0]
+                last[1] = root[0] = self.__map[key] = [last, root, key]
+            dict_setitem(self, key, value)
+
+        def __delitem__(self, key, dict_delitem=dict.__delitem__):
+            'od.__delitem__(y) <==> del od[y]'
+            # Deleting an existing item uses self.__map to find the link which is
+            # then removed by updating the links in the predecessor and successor nodes.
+            dict_delitem(self, key)
+            link_prev, link_next, key = self.__map.pop(key)
+            link_prev[1] = link_next
+            link_next[0] = link_prev
+
+        def __iter__(self):
+            'od.__iter__() <==> iter(od)'
+            root = self.__root
+            curr = root[1]
+            while curr is not root:
+                yield curr[2]
+                curr = curr[1]
+
+        def __reversed__(self):
+            'od.__reversed__() <==> reversed(od)'
+            root = self.__root
+            curr = root[0]
+            while curr is not root:
+                yield curr[2]
+                curr = curr[0]
+
+        def clear(self):
+            'od.clear() -> None.  Remove all items from od.'
+            try:
+                for node in self.__map.itervalues():
+                    del node[:]
+                root = self.__root
+                root[:] = [root, root, None]
+                self.__map.clear()
+            except AttributeError:
+                pass
+            dict.clear(self)
+
+        def popitem(self, last=True):
+            '''od.popitem() -> (k, v), return and remove a (key, value) pair.
+            Pairs are returned in LIFO order if last is true or FIFO order if false.
+
+            '''
+            if not self:
+                raise KeyError('dictionary is empty')
+            root = self.__root
+            if last:
+                link = root[0]
+                link_prev = link[0]
+                link_prev[1] = root
+                root[0] = link_prev
+            else:
+                link = root[1]
+                link_next = link[1]
+                root[1] = link_next
+                link_next[0] = root
+            key = link[2]
+            del self.__map[key]
+            value = dict.pop(self, key)
+            return key, value
+
+        # -- the following methods do not depend on the internal structure --
+
+        def keys(self):
+            'od.keys() -> list of keys in od'
+            return list(self)
+
+        def values(self):
+            'od.values() -> list of values in od'
+            return [self[key] for key in self]
+
+        def items(self):
+            'od.items() -> list of (key, value) pairs in od'
+            return [(key, self[key]) for key in self]
+
+        def iterkeys(self):
+            'od.iterkeys() -> an iterator over the keys in od'
+            return iter(self)
+
+        def itervalues(self):
+            'od.itervalues -> an iterator over the values in od'
+            for k in self:
+                yield self[k]
+
+        def iteritems(self):
+            'od.iteritems -> an iterator over the (key, value) items in od'
+            for k in self:
+                yield (k, self[k])
+
+        def update(*args, **kwds):
+            '''od.update(E, **F) -> None.  Update od from dict/iterable E and F.
+
+            If E is a dict instance, does:           for k in E: od[k] = E[k]
+            If E has a .keys() method, does:         for k in E.keys(): od[k] = E[k]
+            Or if E is an iterable of items, does:   for k, v in E: od[k] = v
+            In either case, this is followed by:     for k, v in F.items(): od[k] = v
+
+            '''
+            if len(args) > 2:
+                raise TypeError('update() takes at most 2 positional '
+                                'arguments (%d given)' % (len(args), ))
+            elif not args:
+                raise TypeError('update() takes at least 1 argument (0 given)')
+            self = args[0]
+            # Make progressively weaker assumptions about "other"
+            other = ()
+            if len(args) == 2:
+                other = args[1]
+            if isinstance(other, dict):
+                for key in other:
+                    self[key] = other[key]
+            elif hasattr(other, 'keys'):
+                for key in other.keys():
+                    self[key] = other[key]
+            else:
+                for key, value in other:
+                    self[key] = value
+            for key, value in kwds.items():
+                self[key] = value
+
+        __update = update  # let subclasses override update without breaking __init__
+
+        __marker = object()
+
+        def pop(self, key, default=__marker):
+            '''od.pop(k[,d]) -> v, remove specified key and return the corresponding value.
+            If key is not found, d is returned if given, otherwise KeyError is raised.
+
+            '''
+            if key in self:
+                result = self[key]
+                del self[key]
+                return result
+            if default is self.__marker:
+                raise KeyError(key)
+            return default
+
+        def setdefault(self, key, default=None):
+            'od.setdefault(k[,d]) -> od.get(k,d), also set od[k]=d if k not in od'
+            if key in self:
+                return self[key]
+            self[key] = default
+            return default
+
+        def __repr__(self, _repr_running=None):
+            'od.__repr__() <==> repr(od)'
+            if not _repr_running:
+                _repr_running = {}
+            call_key = id(self), _get_ident()
+            if call_key in _repr_running:
+                return '...'
+            _repr_running[call_key] = 1
+            try:
+                if not self:
+                    return '%s()' % (self.__class__.__name__, )
+                return '%s(%r)' % (self.__class__.__name__, self.items())
+            finally:
+                del _repr_running[call_key]
+
+        def __reduce__(self):
+            'Return state information for pickling'
+            items = [[k, self[k]] for k in self]
+            inst_dict = vars(self).copy()
+            for k in vars(OrderedDict()):
+                inst_dict.pop(k, None)
+            if inst_dict:
+                return (self.__class__, (items, ), inst_dict)
+            return self.__class__, (items, )
+
+        def copy(self):
+            'od.copy() -> a shallow copy of od'
+            return self.__class__(self)
+
+        @classmethod
+        def fromkeys(cls, iterable, value=None):
+            '''OD.fromkeys(S[, v]) -> New ordered dictionary with keys from S
+            and values equal to v (which defaults to None).
+
+            '''
+            d = cls()
+            for key in iterable:
+                d[key] = value
+            return d
+
+        def __eq__(self, other):
+            '''od.__eq__(y) <==> od==y.  Comparison to another OD is order-sensitive
+            while comparison to a regular mapping is order-insensitive.
+
+            '''
+            if isinstance(other, OrderedDict):
+                return len(self) == len(
+                    other) and self.items() == other.items()
+            return dict.__eq__(self, other)
+
+        def __ne__(self, other):
+            return not self == other
+
+        # -- the following methods are only used in Python 2.7 --
+
+        def viewkeys(self):
+            "od.viewkeys() -> a set-like object providing a view on od's keys"
+            return KeysView(self)
+
+        def viewvalues(self):
+            "od.viewvalues() -> an object providing a view on od's values"
+            return ValuesView(self)
+
+        def viewitems(self):
+            "od.viewitems() -> a set-like object providing a view on od's items"
+            return ItemsView(self)
+
+
+try:
+    from logging.config import BaseConfigurator, valid_ident
+except ImportError:  # pragma: no cover
+    IDENTIFIER = re.compile('^[a-z_][a-z0-9_]*$', re.I)
+
+    def valid_ident(s):
+        m = IDENTIFIER.match(s)
+        if not m:
+            raise ValueError('Not a valid Python identifier: %r' % s)
+        return True
+
+    # The ConvertingXXX classes are wrappers around standard Python containers,
+    # and they serve to convert any suitable values in the container. The
+    # conversion converts base dicts, lists and tuples to their wrapped
+    # equivalents, whereas strings which match a conversion format are converted
+    # appropriately.
+    #
+    # Each wrapper should have a configurator attribute holding the actual
+    # configurator to use for conversion.
+
+    class ConvertingDict(dict):
+        """A converting dictionary wrapper."""
+
+        def __getitem__(self, key):
+            value = dict.__getitem__(self, key)
+            result = self.configurator.convert(value)
+            # If the converted value is different, save for next time
+            if value is not result:
+                self[key] = result
+                if type(result) in (ConvertingDict, ConvertingList,
+                                    ConvertingTuple):
+                    result.parent = self
+                    result.key = key
+            return result
+
+        def get(self, key, default=None):
+            value = dict.get(self, key, default)
+            result = self.configurator.convert(value)
+            # If the converted value is different, save for next time
+            if value is not result:
+                self[key] = result
+                if type(result) in (ConvertingDict, ConvertingList,
+                                    ConvertingTuple):
+                    result.parent = self
+                    result.key = key
+            return result
+
+    def pop(self, key, default=None):
+        value = dict.pop(self, key, default)
+        result = self.configurator.convert(value)
+        if value is not result:
+            if type(result) in (ConvertingDict, ConvertingList,
+                                ConvertingTuple):
+                result.parent = self
+                result.key = key
+        return result
+
+    class ConvertingList(list):
+        """A converting list wrapper."""
+
+        def __getitem__(self, key):
+            value = list.__getitem__(self, key)
+            result = self.configurator.convert(value)
+            # If the converted value is different, save for next time
+            if value is not result:
+                self[key] = result
+                if type(result) in (ConvertingDict, ConvertingList,
+                                    ConvertingTuple):
+                    result.parent = self
+                    result.key = key
+            return result
+
+        def pop(self, idx=-1):
+            value = list.pop(self, idx)
+            result = self.configurator.convert(value)
+            if value is not result:
+                if type(result) in (ConvertingDict, ConvertingList,
+                                    ConvertingTuple):
+                    result.parent = self
+            return result
+
+    class ConvertingTuple(tuple):
+        """A converting tuple wrapper."""
+
+        def __getitem__(self, key):
+            value = tuple.__getitem__(self, key)
+            result = self.configurator.convert(value)
+            if value is not result:
+                if type(result) in (ConvertingDict, ConvertingList,
+                                    ConvertingTuple):
+                    result.parent = self
+                    result.key = key
+            return result
+
+    class BaseConfigurator(object):
+        """
+        The configurator base class which defines some useful defaults.
+        """
+
+        CONVERT_PATTERN = re.compile(r'^(?P<prefix>[a-z]+)://(?P<suffix>.*)$')
+
+        WORD_PATTERN = re.compile(r'^\s*(\w+)\s*')
+        DOT_PATTERN = re.compile(r'^\.\s*(\w+)\s*')
+        INDEX_PATTERN = re.compile(r'^\[\s*(\w+)\s*\]\s*')
+        DIGIT_PATTERN = re.compile(r'^\d+$')
+
+        value_converters = {
+            'ext': 'ext_convert',
+            'cfg': 'cfg_convert',
+        }
+
+        # We might want to use a different one, e.g. importlib
+        importer = staticmethod(__import__)
+
+        def __init__(self, config):
+            self.config = ConvertingDict(config)
+            self.config.configurator = self
+
+        def resolve(self, s):
+            """
+            Resolve strings to objects using standard import and attribute
+            syntax.
+            """
+            name = s.split('.')
+            used = name.pop(0)
+            try:
+                found = self.importer(used)
+                for frag in name:
+                    used += '.' + frag
+                    try:
+                        found = getattr(found, frag)
+                    except AttributeError:
+                        self.importer(used)
+                        found = getattr(found, frag)
+                return found
+            except ImportError:
+                e, tb = sys.exc_info()[1:]
+                v = ValueError('Cannot resolve %r: %s' % (s, e))
+                v.__cause__, v.__traceback__ = e, tb
+                raise v
+
+        def ext_convert(self, value):
+            """Default converter for the ext:// protocol."""
+            return self.resolve(value)
+
+        def cfg_convert(self, value):
+            """Default converter for the cfg:// protocol."""
+            rest = value
+            m = self.WORD_PATTERN.match(rest)
+            if m is None:
+                raise ValueError("Unable to convert %r" % value)
+            else:
+                rest = rest[m.end():]
+                d = self.config[m.groups()[0]]
+                while rest:
+                    m = self.DOT_PATTERN.match(rest)
+                    if m:
+                        d = d[m.groups()[0]]
+                    else:
+                        m = self.INDEX_PATTERN.match(rest)
+                        if m:
+                            idx = m.groups()[0]
+                            if not self.DIGIT_PATTERN.match(idx):
+                                d = d[idx]
+                            else:
+                                try:
+                                    n = int(
+                                        idx
+                                    )  # try as number first (most likely)
+                                    d = d[n]
+                                except TypeError:
+                                    d = d[idx]
+                    if m:
+                        rest = rest[m.end():]
+                    else:
+                        raise ValueError('Unable to convert '
+                                         '%r at %r' % (value, rest))
+            # rest should be empty
+            return d
+
+        def convert(self, value):
+            """
+            Convert values to an appropriate type. dicts, lists and tuples are
+            replaced by their converting alternatives. Strings are checked to
+            see if they have a conversion format and are converted if they do.
+            """
+            if not isinstance(value, ConvertingDict) and isinstance(
+                    value, dict):
+                value = ConvertingDict(value)
+                value.configurator = self
+            elif not isinstance(value, ConvertingList) and isinstance(
+                    value, list):
+                value = ConvertingList(value)
+                value.configurator = self
+            elif not isinstance(value, ConvertingTuple) and isinstance(value, tuple):
+                value = ConvertingTuple(value)
+                value.configurator = self
+            elif isinstance(value, string_types):
+                m = self.CONVERT_PATTERN.match(value)
+                if m:
+                    d = m.groupdict()
+                    prefix = d['prefix']
+                    converter = self.value_converters.get(prefix, None)
+                    if converter:
+                        suffix = d['suffix']
+                        converter = getattr(self, converter)
+                        value = converter(suffix)
+            return value
+
+        def configure_custom(self, config):
+            """Configure an object with a user-supplied factory."""
+            c = config.pop('()')
+            if not callable(c):
+                c = self.resolve(c)
+            props = config.pop('.', None)
+            # Check for valid identifiers
+            kwargs = dict([(k, config[k]) for k in config if valid_ident(k)])
+            result = c(**kwargs)
+            if props:
+                for name, value in props.items():
+                    setattr(result, name, value)
+            return result
+
+        def as_tuple(self, value):
+            """Utility function which converts lists to tuples."""
+            if isinstance(value, list):
+                value = tuple(value)
+            return value
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/database.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/database.py
new file mode 100644
index 00000000..c0f896a7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/database.py
@@ -0,0 +1,1329 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 The Python Software Foundation.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+"""PEP 376 implementation."""
+
+from __future__ import unicode_literals
+
+import base64
+import codecs
+import contextlib
+import hashlib
+import logging
+import os
+import posixpath
+import sys
+import zipimport
+
+from . import DistlibException, resources
+from .compat import StringIO
+from .version import get_scheme, UnsupportedVersionError
+from .metadata import (Metadata, METADATA_FILENAME, WHEEL_METADATA_FILENAME, LEGACY_METADATA_FILENAME)
+from .util import (parse_requirement, cached_property, parse_name_and_version, read_exports, write_exports, CSVReader,
+                   CSVWriter)
+
+__all__ = [
+    'Distribution', 'BaseInstalledDistribution', 'InstalledDistribution', 'EggInfoDistribution', 'DistributionPath'
+]
+
+logger = logging.getLogger(__name__)
+
+EXPORTS_FILENAME = 'pydist-exports.json'
+COMMANDS_FILENAME = 'pydist-commands.json'
+
+DIST_FILES = ('INSTALLER', METADATA_FILENAME, 'RECORD', 'REQUESTED', 'RESOURCES', EXPORTS_FILENAME, 'SHARED')
+
+DISTINFO_EXT = '.dist-info'
+
+
+class _Cache(object):
+    """
+    A simple cache mapping names and .dist-info paths to distributions
+    """
+
+    def __init__(self):
+        """
+        Initialise an instance. There is normally one for each DistributionPath.
+        """
+        self.name = {}
+        self.path = {}
+        self.generated = False
+
+    def clear(self):
+        """
+        Clear the cache, setting it to its initial state.
+        """
+        self.name.clear()
+        self.path.clear()
+        self.generated = False
+
+    def add(self, dist):
+        """
+        Add a distribution to the cache.
+        :param dist: The distribution to add.
+        """
+        if dist.path not in self.path:
+            self.path[dist.path] = dist
+            self.name.setdefault(dist.key, []).append(dist)
+
+
+class DistributionPath(object):
+    """
+    Represents a set of distributions installed on a path (typically sys.path).
+    """
+
+    def __init__(self, path=None, include_egg=False):
+        """
+        Create an instance from a path, optionally including legacy (distutils/
+        setuptools/distribute) distributions.
+        :param path: The path to use, as a list of directories. If not specified,
+                     sys.path is used.
+        :param include_egg: If True, this instance will look for and return legacy
+                            distributions as well as those based on PEP 376.
+        """
+        if path is None:
+            path = sys.path
+        self.path = path
+        self._include_dist = True
+        self._include_egg = include_egg
+
+        self._cache = _Cache()
+        self._cache_egg = _Cache()
+        self._cache_enabled = True
+        self._scheme = get_scheme('default')
+
+    def _get_cache_enabled(self):
+        return self._cache_enabled
+
+    def _set_cache_enabled(self, value):
+        self._cache_enabled = value
+
+    cache_enabled = property(_get_cache_enabled, _set_cache_enabled)
+
+    def clear_cache(self):
+        """
+        Clears the internal cache.
+        """
+        self._cache.clear()
+        self._cache_egg.clear()
+
+    def _yield_distributions(self):
+        """
+        Yield .dist-info and/or .egg(-info) distributions.
+        """
+        # We need to check if we've seen some resources already, because on
+        # some Linux systems (e.g. some Debian/Ubuntu variants) there are
+        # symlinks which alias other files in the environment.
+        seen = set()
+        for path in self.path:
+            finder = resources.finder_for_path(path)
+            if finder is None:
+                continue
+            r = finder.find('')
+            if not r or not r.is_container:
+                continue
+            rset = sorted(r.resources)
+            for entry in rset:
+                r = finder.find(entry)
+                if not r or r.path in seen:
+                    continue
+                try:
+                    if self._include_dist and entry.endswith(DISTINFO_EXT):
+                        possible_filenames = [METADATA_FILENAME, WHEEL_METADATA_FILENAME, LEGACY_METADATA_FILENAME]
+                        for metadata_filename in possible_filenames:
+                            metadata_path = posixpath.join(entry, metadata_filename)
+                            pydist = finder.find(metadata_path)
+                            if pydist:
+                                break
+                        else:
+                            continue
+
+                        with contextlib.closing(pydist.as_stream()) as stream:
+                            metadata = Metadata(fileobj=stream, scheme='legacy')
+                        logger.debug('Found %s', r.path)
+                        seen.add(r.path)
+                        yield new_dist_class(r.path, metadata=metadata, env=self)
+                    elif self._include_egg and entry.endswith(('.egg-info', '.egg')):
+                        logger.debug('Found %s', r.path)
+                        seen.add(r.path)
+                        yield old_dist_class(r.path, self)
+                except Exception as e:
+                    msg = 'Unable to read distribution at %s, perhaps due to bad metadata: %s'
+                    logger.warning(msg, r.path, e)
+                    import warnings
+                    warnings.warn(msg % (r.path, e), stacklevel=2)
+
+    def _generate_cache(self):
+        """
+        Scan the path for distributions and populate the cache with
+        those that are found.
+        """
+        gen_dist = not self._cache.generated
+        gen_egg = self._include_egg and not self._cache_egg.generated
+        if gen_dist or gen_egg:
+            for dist in self._yield_distributions():
+                if isinstance(dist, InstalledDistribution):
+                    self._cache.add(dist)
+                else:
+                    self._cache_egg.add(dist)
+
+            if gen_dist:
+                self._cache.generated = True
+            if gen_egg:
+                self._cache_egg.generated = True
+
+    @classmethod
+    def distinfo_dirname(cls, name, version):
+        """
+        The *name* and *version* parameters are converted into their
+        filename-escaped form, i.e. any ``'-'`` characters are replaced
+        with ``'_'`` other than the one in ``'dist-info'`` and the one
+        separating the name from the version number.
+
+        :parameter name: is converted to a standard distribution name by replacing
+                         any runs of non- alphanumeric characters with a single
+                         ``'-'``.
+        :type name: string
+        :parameter version: is converted to a standard version string. Spaces
+                            become dots, and all other non-alphanumeric characters
+                            (except dots) become dashes, with runs of multiple
+                            dashes condensed to a single dash.
+        :type version: string
+        :returns: directory name
+        :rtype: string"""
+        name = name.replace('-', '_')
+        return '-'.join([name, version]) + DISTINFO_EXT
+
+    def get_distributions(self):
+        """
+        Provides an iterator that looks for distributions and returns
+        :class:`InstalledDistribution` or
+        :class:`EggInfoDistribution` instances for each one of them.
+
+        :rtype: iterator of :class:`InstalledDistribution` and
+                :class:`EggInfoDistribution` instances
+        """
+        if not self._cache_enabled:
+            for dist in self._yield_distributions():
+                yield dist
+        else:
+            self._generate_cache()
+
+            for dist in self._cache.path.values():
+                yield dist
+
+            if self._include_egg:
+                for dist in self._cache_egg.path.values():
+                    yield dist
+
+    def get_distribution(self, name):
+        """
+        Looks for a named distribution on the path.
+
+        This function only returns the first result found, as no more than one
+        value is expected. If nothing is found, ``None`` is returned.
+
+        :rtype: :class:`InstalledDistribution`, :class:`EggInfoDistribution`
+                or ``None``
+        """
+        result = None
+        name = name.lower()
+        if not self._cache_enabled:
+            for dist in self._yield_distributions():
+                if dist.key == name:
+                    result = dist
+                    break
+        else:
+            self._generate_cache()
+
+            if name in self._cache.name:
+                result = self._cache.name[name][0]
+            elif self._include_egg and name in self._cache_egg.name:
+                result = self._cache_egg.name[name][0]
+        return result
+
+    def provides_distribution(self, name, version=None):
+        """
+        Iterates over all distributions to find which distributions provide *name*.
+        If a *version* is provided, it will be used to filter the results.
+
+        This function only returns the first result found, since no more than
+        one values are expected. If the directory is not found, returns ``None``.
+
+        :parameter version: a version specifier that indicates the version
+                            required, conforming to the format in ``PEP-345``
+
+        :type name: string
+        :type version: string
+        """
+        matcher = None
+        if version is not None:
+            try:
+                matcher = self._scheme.matcher('%s (%s)' % (name, version))
+            except ValueError:
+                raise DistlibException('invalid name or version: %r, %r' % (name, version))
+
+        for dist in self.get_distributions():
+            # We hit a problem on Travis where enum34 was installed and doesn't
+            # have a provides attribute ...
+            if not hasattr(dist, 'provides'):
+                logger.debug('No "provides": %s', dist)
+            else:
+                provided = dist.provides
+
+                for p in provided:
+                    p_name, p_ver = parse_name_and_version(p)
+                    if matcher is None:
+                        if p_name == name:
+                            yield dist
+                            break
+                    else:
+                        if p_name == name and matcher.match(p_ver):
+                            yield dist
+                            break
+
+    def get_file_path(self, name, relative_path):
+        """
+        Return the path to a resource file.
+        """
+        dist = self.get_distribution(name)
+        if dist is None:
+            raise LookupError('no distribution named %r found' % name)
+        return dist.get_resource_path(relative_path)
+
+    def get_exported_entries(self, category, name=None):
+        """
+        Return all of the exported entries in a particular category.
+
+        :param category: The category to search for entries.
+        :param name: If specified, only entries with that name are returned.
+        """
+        for dist in self.get_distributions():
+            r = dist.exports
+            if category in r:
+                d = r[category]
+                if name is not None:
+                    if name in d:
+                        yield d[name]
+                else:
+                    for v in d.values():
+                        yield v
+
+
+class Distribution(object):
+    """
+    A base class for distributions, whether installed or from indexes.
+    Either way, it must have some metadata, so that's all that's needed
+    for construction.
+    """
+
+    build_time_dependency = False
+    """
+    Set to True if it's known to be only a build-time dependency (i.e.
+    not needed after installation).
+    """
+
+    requested = False
+    """A boolean that indicates whether the ``REQUESTED`` metadata file is
+    present (in other words, whether the package was installed by user
+    request or it was installed as a dependency)."""
+
+    def __init__(self, metadata):
+        """
+        Initialise an instance.
+        :param metadata: The instance of :class:`Metadata` describing this
+        distribution.
+        """
+        self.metadata = metadata
+        self.name = metadata.name
+        self.key = self.name.lower()  # for case-insensitive comparisons
+        self.version = metadata.version
+        self.locator = None
+        self.digest = None
+        self.extras = None  # additional features requested
+        self.context = None  # environment marker overrides
+        self.download_urls = set()
+        self.digests = {}
+
+    @property
+    def source_url(self):
+        """
+        The source archive download URL for this distribution.
+        """
+        return self.metadata.source_url
+
+    download_url = source_url  # Backward compatibility
+
+    @property
+    def name_and_version(self):
+        """
+        A utility property which displays the name and version in parentheses.
+        """
+        return '%s (%s)' % (self.name, self.version)
+
+    @property
+    def provides(self):
+        """
+        A set of distribution names and versions provided by this distribution.
+        :return: A set of "name (version)" strings.
+        """
+        plist = self.metadata.provides
+        s = '%s (%s)' % (self.name, self.version)
+        if s not in plist:
+            plist.append(s)
+        return plist
+
+    def _get_requirements(self, req_attr):
+        md = self.metadata
+        reqts = getattr(md, req_attr)
+        logger.debug('%s: got requirements %r from metadata: %r', self.name, req_attr, reqts)
+        return set(md.get_requirements(reqts, extras=self.extras, env=self.context))
+
+    @property
+    def run_requires(self):
+        return self._get_requirements('run_requires')
+
+    @property
+    def meta_requires(self):
+        return self._get_requirements('meta_requires')
+
+    @property
+    def build_requires(self):
+        return self._get_requirements('build_requires')
+
+    @property
+    def test_requires(self):
+        return self._get_requirements('test_requires')
+
+    @property
+    def dev_requires(self):
+        return self._get_requirements('dev_requires')
+
+    def matches_requirement(self, req):
+        """
+        Say if this instance matches (fulfills) a requirement.
+        :param req: The requirement to match.
+        :rtype req: str
+        :return: True if it matches, else False.
+        """
+        # Requirement may contain extras - parse to lose those
+        # from what's passed to the matcher
+        r = parse_requirement(req)
+        scheme = get_scheme(self.metadata.scheme)
+        try:
+            matcher = scheme.matcher(r.requirement)
+        except UnsupportedVersionError:
+            # XXX compat-mode if cannot read the version
+            logger.warning('could not read version %r - using name only', req)
+            name = req.split()[0]
+            matcher = scheme.matcher(name)
+
+        name = matcher.key  # case-insensitive
+
+        result = False
+        for p in self.provides:
+            p_name, p_ver = parse_name_and_version(p)
+            if p_name != name:
+                continue
+            try:
+                result = matcher.match(p_ver)
+                break
+            except UnsupportedVersionError:
+                pass
+        return result
+
+    def __repr__(self):
+        """
+        Return a textual representation of this instance,
+        """
+        if self.source_url:
+            suffix = ' [%s]' % self.source_url
+        else:
+            suffix = ''
+        return '<Distribution %s (%s)%s>' % (self.name, self.version, suffix)
+
+    def __eq__(self, other):
+        """
+        See if this distribution is the same as another.
+        :param other: The distribution to compare with. To be equal to one
+                      another. distributions must have the same type, name,
+                      version and source_url.
+        :return: True if it is the same, else False.
+        """
+        if type(other) is not type(self):
+            result = False
+        else:
+            result = (self.name == other.name and self.version == other.version and self.source_url == other.source_url)
+        return result
+
+    def __hash__(self):
+        """
+        Compute hash in a way which matches the equality test.
+        """
+        return hash(self.name) + hash(self.version) + hash(self.source_url)
+
+
+class BaseInstalledDistribution(Distribution):
+    """
+    This is the base class for installed distributions (whether PEP 376 or
+    legacy).
+    """
+
+    hasher = None
+
+    def __init__(self, metadata, path, env=None):
+        """
+        Initialise an instance.
+        :param metadata: An instance of :class:`Metadata` which describes the
+                         distribution. This will normally have been initialised
+                         from a metadata file in the ``path``.
+        :param path:     The path of the ``.dist-info`` or ``.egg-info``
+                         directory for the distribution.
+        :param env:      This is normally the :class:`DistributionPath`
+                         instance where this distribution was found.
+        """
+        super(BaseInstalledDistribution, self).__init__(metadata)
+        self.path = path
+        self.dist_path = env
+
+    def get_hash(self, data, hasher=None):
+        """
+        Get the hash of some data, using a particular hash algorithm, if
+        specified.
+
+        :param data: The data to be hashed.
+        :type data: bytes
+        :param hasher: The name of a hash implementation, supported by hashlib,
+                       or ``None``. Examples of valid values are ``'sha1'``,
+                       ``'sha224'``, ``'sha384'``, '``sha256'``, ``'md5'`` and
+                       ``'sha512'``. If no hasher is specified, the ``hasher``
+                       attribute of the :class:`InstalledDistribution` instance
+                       is used. If the hasher is determined to be ``None``, MD5
+                       is used as the hashing algorithm.
+        :returns: The hash of the data. If a hasher was explicitly specified,
+                  the returned hash will be prefixed with the specified hasher
+                  followed by '='.
+        :rtype: str
+        """
+        if hasher is None:
+            hasher = self.hasher
+        if hasher is None:
+            hasher = hashlib.md5
+            prefix = ''
+        else:
+            hasher = getattr(hashlib, hasher)
+            prefix = '%s=' % self.hasher
+        digest = hasher(data).digest()
+        digest = base64.urlsafe_b64encode(digest).rstrip(b'=').decode('ascii')
+        return '%s%s' % (prefix, digest)
+
+
+class InstalledDistribution(BaseInstalledDistribution):
+    """
+    Created with the *path* of the ``.dist-info`` directory provided to the
+    constructor. It reads the metadata contained in ``pydist.json`` when it is
+    instantiated., or uses a passed in Metadata instance (useful for when
+    dry-run mode is being used).
+    """
+
+    hasher = 'sha256'
+
+    def __init__(self, path, metadata=None, env=None):
+        self.modules = []
+        self.finder = finder = resources.finder_for_path(path)
+        if finder is None:
+            raise ValueError('finder unavailable for %s' % path)
+        if env and env._cache_enabled and path in env._cache.path:
+            metadata = env._cache.path[path].metadata
+        elif metadata is None:
+            r = finder.find(METADATA_FILENAME)
+            # Temporary - for Wheel 0.23 support
+            if r is None:
+                r = finder.find(WHEEL_METADATA_FILENAME)
+            # Temporary - for legacy support
+            if r is None:
+                r = finder.find(LEGACY_METADATA_FILENAME)
+            if r is None:
+                raise ValueError('no %s found in %s' % (METADATA_FILENAME, path))
+            with contextlib.closing(r.as_stream()) as stream:
+                metadata = Metadata(fileobj=stream, scheme='legacy')
+
+        super(InstalledDistribution, self).__init__(metadata, path, env)
+
+        if env and env._cache_enabled:
+            env._cache.add(self)
+
+        r = finder.find('REQUESTED')
+        self.requested = r is not None
+        p = os.path.join(path, 'top_level.txt')
+        if os.path.exists(p):
+            with open(p, 'rb') as f:
+                data = f.read().decode('utf-8')
+            self.modules = data.splitlines()
+
+    def __repr__(self):
+        return '<InstalledDistribution %r %s at %r>' % (self.name, self.version, self.path)
+
+    def __str__(self):
+        return "%s %s" % (self.name, self.version)
+
+    def _get_records(self):
+        """
+        Get the list of installed files for the distribution
+        :return: A list of tuples of path, hash and size. Note that hash and
+                 size might be ``None`` for some entries. The path is exactly
+                 as stored in the file (which is as in PEP 376).
+        """
+        results = []
+        r = self.get_distinfo_resource('RECORD')
+        with contextlib.closing(r.as_stream()) as stream:
+            with CSVReader(stream=stream) as record_reader:
+                # Base location is parent dir of .dist-info dir
+                # base_location = os.path.dirname(self.path)
+                # base_location = os.path.abspath(base_location)
+                for row in record_reader:
+                    missing = [None for i in range(len(row), 3)]
+                    path, checksum, size = row + missing
+                    # if not os.path.isabs(path):
+                    #     path = path.replace('/', os.sep)
+                    #     path = os.path.join(base_location, path)
+                    results.append((path, checksum, size))
+        return results
+
+    @cached_property
+    def exports(self):
+        """
+        Return the information exported by this distribution.
+        :return: A dictionary of exports, mapping an export category to a dict
+                 of :class:`ExportEntry` instances describing the individual
+                 export entries, and keyed by name.
+        """
+        result = {}
+        r = self.get_distinfo_resource(EXPORTS_FILENAME)
+        if r:
+            result = self.read_exports()
+        return result
+
+    def read_exports(self):
+        """
+        Read exports data from a file in .ini format.
+
+        :return: A dictionary of exports, mapping an export category to a list
+                 of :class:`ExportEntry` instances describing the individual
+                 export entries.
+        """
+        result = {}
+        r = self.get_distinfo_resource(EXPORTS_FILENAME)
+        if r:
+            with contextlib.closing(r.as_stream()) as stream:
+                result = read_exports(stream)
+        return result
+
+    def write_exports(self, exports):
+        """
+        Write a dictionary of exports to a file in .ini format.
+        :param exports: A dictionary of exports, mapping an export category to
+                        a list of :class:`ExportEntry` instances describing the
+                        individual export entries.
+        """
+        rf = self.get_distinfo_file(EXPORTS_FILENAME)
+        with open(rf, 'w') as f:
+            write_exports(exports, f)
+
+    def get_resource_path(self, relative_path):
+        """
+        NOTE: This API may change in the future.
+
+        Return the absolute path to a resource file with the given relative
+        path.
+
+        :param relative_path: The path, relative to .dist-info, of the resource
+                              of interest.
+        :return: The absolute path where the resource is to be found.
+        """
+        r = self.get_distinfo_resource('RESOURCES')
+        with contextlib.closing(r.as_stream()) as stream:
+            with CSVReader(stream=stream) as resources_reader:
+                for relative, destination in resources_reader:
+                    if relative == relative_path:
+                        return destination
+        raise KeyError('no resource file with relative path %r '
+                       'is installed' % relative_path)
+
+    def list_installed_files(self):
+        """
+        Iterates over the ``RECORD`` entries and returns a tuple
+        ``(path, hash, size)`` for each line.
+
+        :returns: iterator of (path, hash, size)
+        """
+        for result in self._get_records():
+            yield result
+
+    def write_installed_files(self, paths, prefix, dry_run=False):
+        """
+        Writes the ``RECORD`` file, using the ``paths`` iterable passed in. Any
+        existing ``RECORD`` file is silently overwritten.
+
+        prefix is used to determine when to write absolute paths.
+        """
+        prefix = os.path.join(prefix, '')
+        base = os.path.dirname(self.path)
+        base_under_prefix = base.startswith(prefix)
+        base = os.path.join(base, '')
+        record_path = self.get_distinfo_file('RECORD')
+        logger.info('creating %s', record_path)
+        if dry_run:
+            return None
+        with CSVWriter(record_path) as writer:
+            for path in paths:
+                if os.path.isdir(path) or path.endswith(('.pyc', '.pyo')):
+                    # do not put size and hash, as in PEP-376
+                    hash_value = size = ''
+                else:
+                    size = '%d' % os.path.getsize(path)
+                    with open(path, 'rb') as fp:
+                        hash_value = self.get_hash(fp.read())
+                if path.startswith(base) or (base_under_prefix and path.startswith(prefix)):
+                    path = os.path.relpath(path, base)
+                writer.writerow((path, hash_value, size))
+
+            # add the RECORD file itself
+            if record_path.startswith(base):
+                record_path = os.path.relpath(record_path, base)
+            writer.writerow((record_path, '', ''))
+        return record_path
+
+    def check_installed_files(self):
+        """
+        Checks that the hashes and sizes of the files in ``RECORD`` are
+        matched by the files themselves. Returns a (possibly empty) list of
+        mismatches. Each entry in the mismatch list will be a tuple consisting
+        of the path, 'exists', 'size' or 'hash' according to what didn't match
+        (existence is checked first, then size, then hash), the expected
+        value and the actual value.
+        """
+        mismatches = []
+        base = os.path.dirname(self.path)
+        record_path = self.get_distinfo_file('RECORD')
+        for path, hash_value, size in self.list_installed_files():
+            if not os.path.isabs(path):
+                path = os.path.join(base, path)
+            if path == record_path:
+                continue
+            if not os.path.exists(path):
+                mismatches.append((path, 'exists', True, False))
+            elif os.path.isfile(path):
+                actual_size = str(os.path.getsize(path))
+                if size and actual_size != size:
+                    mismatches.append((path, 'size', size, actual_size))
+                elif hash_value:
+                    if '=' in hash_value:
+                        hasher = hash_value.split('=', 1)[0]
+                    else:
+                        hasher = None
+
+                    with open(path, 'rb') as f:
+                        actual_hash = self.get_hash(f.read(), hasher)
+                        if actual_hash != hash_value:
+                            mismatches.append((path, 'hash', hash_value, actual_hash))
+        return mismatches
+
+    @cached_property
+    def shared_locations(self):
+        """
+        A dictionary of shared locations whose keys are in the set 'prefix',
+        'purelib', 'platlib', 'scripts', 'headers', 'data' and 'namespace'.
+        The corresponding value is the absolute path of that category for
+        this distribution, and takes into account any paths selected by the
+        user at installation time (e.g. via command-line arguments). In the
+        case of the 'namespace' key, this would be a list of absolute paths
+        for the roots of namespace packages in this distribution.
+
+        The first time this property is accessed, the relevant information is
+        read from the SHARED file in the .dist-info directory.
+        """
+        result = {}
+        shared_path = os.path.join(self.path, 'SHARED')
+        if os.path.isfile(shared_path):
+            with codecs.open(shared_path, 'r', encoding='utf-8') as f:
+                lines = f.read().splitlines()
+            for line in lines:
+                key, value = line.split('=', 1)
+                if key == 'namespace':
+                    result.setdefault(key, []).append(value)
+                else:
+                    result[key] = value
+        return result
+
+    def write_shared_locations(self, paths, dry_run=False):
+        """
+        Write shared location information to the SHARED file in .dist-info.
+        :param paths: A dictionary as described in the documentation for
+        :meth:`shared_locations`.
+        :param dry_run: If True, the action is logged but no file is actually
+                        written.
+        :return: The path of the file written to.
+        """
+        shared_path = os.path.join(self.path, 'SHARED')
+        logger.info('creating %s', shared_path)
+        if dry_run:
+            return None
+        lines = []
+        for key in ('prefix', 'lib', 'headers', 'scripts', 'data'):
+            path = paths[key]
+            if os.path.isdir(paths[key]):
+                lines.append('%s=%s' % (key, path))
+        for ns in paths.get('namespace', ()):
+            lines.append('namespace=%s' % ns)
+
+        with codecs.open(shared_path, 'w', encoding='utf-8') as f:
+            f.write('\n'.join(lines))
+        return shared_path
+
+    def get_distinfo_resource(self, path):
+        if path not in DIST_FILES:
+            raise DistlibException('invalid path for a dist-info file: '
+                                   '%r at %r' % (path, self.path))
+        finder = resources.finder_for_path(self.path)
+        if finder is None:
+            raise DistlibException('Unable to get a finder for %s' % self.path)
+        return finder.find(path)
+
+    def get_distinfo_file(self, path):
+        """
+        Returns a path located under the ``.dist-info`` directory. Returns a
+        string representing the path.
+
+        :parameter path: a ``'/'``-separated path relative to the
+                         ``.dist-info`` directory or an absolute path;
+                         If *path* is an absolute path and doesn't start
+                         with the ``.dist-info`` directory path,
+                         a :class:`DistlibException` is raised
+        :type path: str
+        :rtype: str
+        """
+        # Check if it is an absolute path  # XXX use relpath, add tests
+        if path.find(os.sep) >= 0:
+            # it's an absolute path?
+            distinfo_dirname, path = path.split(os.sep)[-2:]
+            if distinfo_dirname != self.path.split(os.sep)[-1]:
+                raise DistlibException('dist-info file %r does not belong to the %r %s '
+                                       'distribution' % (path, self.name, self.version))
+
+        # The file must be relative
+        if path not in DIST_FILES:
+            raise DistlibException('invalid path for a dist-info file: '
+                                   '%r at %r' % (path, self.path))
+
+        return os.path.join(self.path, path)
+
+    def list_distinfo_files(self):
+        """
+        Iterates over the ``RECORD`` entries and returns paths for each line if
+        the path is pointing to a file located in the ``.dist-info`` directory
+        or one of its subdirectories.
+
+        :returns: iterator of paths
+        """
+        base = os.path.dirname(self.path)
+        for path, checksum, size in self._get_records():
+            # XXX add separator or use real relpath algo
+            if not os.path.isabs(path):
+                path = os.path.join(base, path)
+            if path.startswith(self.path):
+                yield path
+
+    def __eq__(self, other):
+        return (isinstance(other, InstalledDistribution) and self.path == other.path)
+
+    # See http://docs.python.org/reference/datamodel#object.__hash__
+    __hash__ = object.__hash__
+
+
+class EggInfoDistribution(BaseInstalledDistribution):
+    """Created with the *path* of the ``.egg-info`` directory or file provided
+    to the constructor. It reads the metadata contained in the file itself, or
+    if the given path happens to be a directory, the metadata is read from the
+    file ``PKG-INFO`` under that directory."""
+
+    requested = True  # as we have no way of knowing, assume it was
+    shared_locations = {}
+
+    def __init__(self, path, env=None):
+
+        def set_name_and_version(s, n, v):
+            s.name = n
+            s.key = n.lower()  # for case-insensitive comparisons
+            s.version = v
+
+        self.path = path
+        self.dist_path = env
+        if env and env._cache_enabled and path in env._cache_egg.path:
+            metadata = env._cache_egg.path[path].metadata
+            set_name_and_version(self, metadata.name, metadata.version)
+        else:
+            metadata = self._get_metadata(path)
+
+            # Need to be set before caching
+            set_name_and_version(self, metadata.name, metadata.version)
+
+            if env and env._cache_enabled:
+                env._cache_egg.add(self)
+        super(EggInfoDistribution, self).__init__(metadata, path, env)
+
+    def _get_metadata(self, path):
+        requires = None
+
+        def parse_requires_data(data):
+            """Create a list of dependencies from a requires.txt file.
+
+            *data*: the contents of a setuptools-produced requires.txt file.
+            """
+            reqs = []
+            lines = data.splitlines()
+            for line in lines:
+                line = line.strip()
+                # sectioned files have bare newlines (separating sections)
+                if not line:  # pragma: no cover
+                    continue
+                if line.startswith('['):  # pragma: no cover
+                    logger.warning('Unexpected line: quitting requirement scan: %r', line)
+                    break
+                r = parse_requirement(line)
+                if not r:  # pragma: no cover
+                    logger.warning('Not recognised as a requirement: %r', line)
+                    continue
+                if r.extras:  # pragma: no cover
+                    logger.warning('extra requirements in requires.txt are '
+                                   'not supported')
+                if not r.constraints:
+                    reqs.append(r.name)
+                else:
+                    cons = ', '.join('%s%s' % c for c in r.constraints)
+                    reqs.append('%s (%s)' % (r.name, cons))
+            return reqs
+
+        def parse_requires_path(req_path):
+            """Create a list of dependencies from a requires.txt file.
+
+            *req_path*: the path to a setuptools-produced requires.txt file.
+            """
+
+            reqs = []
+            try:
+                with codecs.open(req_path, 'r', 'utf-8') as fp:
+                    reqs = parse_requires_data(fp.read())
+            except IOError:
+                pass
+            return reqs
+
+        tl_path = tl_data = None
+        if path.endswith('.egg'):
+            if os.path.isdir(path):
+                p = os.path.join(path, 'EGG-INFO')
+                meta_path = os.path.join(p, 'PKG-INFO')
+                metadata = Metadata(path=meta_path, scheme='legacy')
+                req_path = os.path.join(p, 'requires.txt')
+                tl_path = os.path.join(p, 'top_level.txt')
+                requires = parse_requires_path(req_path)
+            else:
+                # FIXME handle the case where zipfile is not available
+                zipf = zipimport.zipimporter(path)
+                fileobj = StringIO(zipf.get_data('EGG-INFO/PKG-INFO').decode('utf8'))
+                metadata = Metadata(fileobj=fileobj, scheme='legacy')
+                try:
+                    data = zipf.get_data('EGG-INFO/requires.txt')
+                    tl_data = zipf.get_data('EGG-INFO/top_level.txt').decode('utf-8')
+                    requires = parse_requires_data(data.decode('utf-8'))
+                except IOError:
+                    requires = None
+        elif path.endswith('.egg-info'):
+            if os.path.isdir(path):
+                req_path = os.path.join(path, 'requires.txt')
+                requires = parse_requires_path(req_path)
+                path = os.path.join(path, 'PKG-INFO')
+                tl_path = os.path.join(path, 'top_level.txt')
+            metadata = Metadata(path=path, scheme='legacy')
+        else:
+            raise DistlibException('path must end with .egg-info or .egg, '
+                                   'got %r' % path)
+
+        if requires:
+            metadata.add_requirements(requires)
+        # look for top-level modules in top_level.txt, if present
+        if tl_data is None:
+            if tl_path is not None and os.path.exists(tl_path):
+                with open(tl_path, 'rb') as f:
+                    tl_data = f.read().decode('utf-8')
+        if not tl_data:
+            tl_data = []
+        else:
+            tl_data = tl_data.splitlines()
+        self.modules = tl_data
+        return metadata
+
+    def __repr__(self):
+        return '<EggInfoDistribution %r %s at %r>' % (self.name, self.version, self.path)
+
+    def __str__(self):
+        return "%s %s" % (self.name, self.version)
+
+    def check_installed_files(self):
+        """
+        Checks that the hashes and sizes of the files in ``RECORD`` are
+        matched by the files themselves. Returns a (possibly empty) list of
+        mismatches. Each entry in the mismatch list will be a tuple consisting
+        of the path, 'exists', 'size' or 'hash' according to what didn't match
+        (existence is checked first, then size, then hash), the expected
+        value and the actual value.
+        """
+        mismatches = []
+        record_path = os.path.join(self.path, 'installed-files.txt')
+        if os.path.exists(record_path):
+            for path, _, _ in self.list_installed_files():
+                if path == record_path:
+                    continue
+                if not os.path.exists(path):
+                    mismatches.append((path, 'exists', True, False))
+        return mismatches
+
+    def list_installed_files(self):
+        """
+        Iterates over the ``installed-files.txt`` entries and returns a tuple
+        ``(path, hash, size)`` for each line.
+
+        :returns: a list of (path, hash, size)
+        """
+
+        def _md5(path):
+            f = open(path, 'rb')
+            try:
+                content = f.read()
+            finally:
+                f.close()
+            return hashlib.md5(content).hexdigest()
+
+        def _size(path):
+            return os.stat(path).st_size
+
+        record_path = os.path.join(self.path, 'installed-files.txt')
+        result = []
+        if os.path.exists(record_path):
+            with codecs.open(record_path, 'r', encoding='utf-8') as f:
+                for line in f:
+                    line = line.strip()
+                    p = os.path.normpath(os.path.join(self.path, line))
+                    # "./" is present as a marker between installed files
+                    # and installation metadata files
+                    if not os.path.exists(p):
+                        logger.warning('Non-existent file: %s', p)
+                        if p.endswith(('.pyc', '.pyo')):
+                            continue
+                        # otherwise fall through and fail
+                    if not os.path.isdir(p):
+                        result.append((p, _md5(p), _size(p)))
+            result.append((record_path, None, None))
+        return result
+
+    def list_distinfo_files(self, absolute=False):
+        """
+        Iterates over the ``installed-files.txt`` entries and returns paths for
+        each line if the path is pointing to a file located in the
+        ``.egg-info`` directory or one of its subdirectories.
+
+        :parameter absolute: If *absolute* is ``True``, each returned path is
+                          transformed into a local absolute path. Otherwise the
+                          raw value from ``installed-files.txt`` is returned.
+        :type absolute: boolean
+        :returns: iterator of paths
+        """
+        record_path = os.path.join(self.path, 'installed-files.txt')
+        if os.path.exists(record_path):
+            skip = True
+            with codecs.open(record_path, 'r', encoding='utf-8') as f:
+                for line in f:
+                    line = line.strip()
+                    if line == './':
+                        skip = False
+                        continue
+                    if not skip:
+                        p = os.path.normpath(os.path.join(self.path, line))
+                        if p.startswith(self.path):
+                            if absolute:
+                                yield p
+                            else:
+                                yield line
+
+    def __eq__(self, other):
+        return (isinstance(other, EggInfoDistribution) and self.path == other.path)
+
+    # See http://docs.python.org/reference/datamodel#object.__hash__
+    __hash__ = object.__hash__
+
+
+new_dist_class = InstalledDistribution
+old_dist_class = EggInfoDistribution
+
+
+class DependencyGraph(object):
+    """
+    Represents a dependency graph between distributions.
+
+    The dependency relationships are stored in an ``adjacency_list`` that maps
+    distributions to a list of ``(other, label)`` tuples where  ``other``
+    is a distribution and the edge is labeled with ``label`` (i.e. the version
+    specifier, if such was provided). Also, for more efficient traversal, for
+    every distribution ``x``, a list of predecessors is kept in
+    ``reverse_list[x]``. An edge from distribution ``a`` to
+    distribution ``b`` means that ``a`` depends on ``b``. If any missing
+    dependencies are found, they are stored in ``missing``, which is a
+    dictionary that maps distributions to a list of requirements that were not
+    provided by any other distributions.
+    """
+
+    def __init__(self):
+        self.adjacency_list = {}
+        self.reverse_list = {}
+        self.missing = {}
+
+    def add_distribution(self, distribution):
+        """Add the *distribution* to the graph.
+
+        :type distribution: :class:`distutils2.database.InstalledDistribution`
+                            or :class:`distutils2.database.EggInfoDistribution`
+        """
+        self.adjacency_list[distribution] = []
+        self.reverse_list[distribution] = []
+        # self.missing[distribution] = []
+
+    def add_edge(self, x, y, label=None):
+        """Add an edge from distribution *x* to distribution *y* with the given
+        *label*.
+
+        :type x: :class:`distutils2.database.InstalledDistribution` or
+                 :class:`distutils2.database.EggInfoDistribution`
+        :type y: :class:`distutils2.database.InstalledDistribution` or
+                 :class:`distutils2.database.EggInfoDistribution`
+        :type label: ``str`` or ``None``
+        """
+        self.adjacency_list[x].append((y, label))
+        # multiple edges are allowed, so be careful
+        if x not in self.reverse_list[y]:
+            self.reverse_list[y].append(x)
+
+    def add_missing(self, distribution, requirement):
+        """
+        Add a missing *requirement* for the given *distribution*.
+
+        :type distribution: :class:`distutils2.database.InstalledDistribution`
+                            or :class:`distutils2.database.EggInfoDistribution`
+        :type requirement: ``str``
+        """
+        logger.debug('%s missing %r', distribution, requirement)
+        self.missing.setdefault(distribution, []).append(requirement)
+
+    def _repr_dist(self, dist):
+        return '%s %s' % (dist.name, dist.version)
+
+    def repr_node(self, dist, level=1):
+        """Prints only a subgraph"""
+        output = [self._repr_dist(dist)]
+        for other, label in self.adjacency_list[dist]:
+            dist = self._repr_dist(other)
+            if label is not None:
+                dist = '%s [%s]' % (dist, label)
+            output.append('    ' * level + str(dist))
+            suboutput = self.repr_node(other, level + 1)
+            subs = suboutput.split('\n')
+            output.extend(subs[1:])
+        return '\n'.join(output)
+
+    def to_dot(self, f, skip_disconnected=True):
+        """Writes a DOT output for the graph to the provided file *f*.
+
+        If *skip_disconnected* is set to ``True``, then all distributions
+        that are not dependent on any other distribution are skipped.
+
+        :type f: has to support ``file``-like operations
+        :type skip_disconnected: ``bool``
+        """
+        disconnected = []
+
+        f.write("digraph dependencies {\n")
+        for dist, adjs in self.adjacency_list.items():
+            if len(adjs) == 0 and not skip_disconnected:
+                disconnected.append(dist)
+            for other, label in adjs:
+                if label is not None:
+                    f.write('"%s" -> "%s" [label="%s"]\n' % (dist.name, other.name, label))
+                else:
+                    f.write('"%s" -> "%s"\n' % (dist.name, other.name))
+        if not skip_disconnected and len(disconnected) > 0:
+            f.write('subgraph disconnected {\n')
+            f.write('label = "Disconnected"\n')
+            f.write('bgcolor = red\n')
+
+            for dist in disconnected:
+                f.write('"%s"' % dist.name)
+                f.write('\n')
+            f.write('}\n')
+        f.write('}\n')
+
+    def topological_sort(self):
+        """
+        Perform a topological sort of the graph.
+        :return: A tuple, the first element of which is a topologically sorted
+                 list of distributions, and the second element of which is a
+                 list of distributions that cannot be sorted because they have
+                 circular dependencies and so form a cycle.
+        """
+        result = []
+        # Make a shallow copy of the adjacency list
+        alist = {}
+        for k, v in self.adjacency_list.items():
+            alist[k] = v[:]
+        while True:
+            # See what we can remove in this run
+            to_remove = []
+            for k, v in list(alist.items())[:]:
+                if not v:
+                    to_remove.append(k)
+                    del alist[k]
+            if not to_remove:
+                # What's left in alist (if anything) is a cycle.
+                break
+            # Remove from the adjacency list of others
+            for k, v in alist.items():
+                alist[k] = [(d, r) for d, r in v if d not in to_remove]
+            logger.debug('Moving to result: %s', ['%s (%s)' % (d.name, d.version) for d in to_remove])
+            result.extend(to_remove)
+        return result, list(alist.keys())
+
+    def __repr__(self):
+        """Representation of the graph"""
+        output = []
+        for dist, adjs in self.adjacency_list.items():
+            output.append(self.repr_node(dist))
+        return '\n'.join(output)
+
+
+def make_graph(dists, scheme='default'):
+    """Makes a dependency graph from the given distributions.
+
+    :parameter dists: a list of distributions
+    :type dists: list of :class:`distutils2.database.InstalledDistribution` and
+                 :class:`distutils2.database.EggInfoDistribution` instances
+    :rtype: a :class:`DependencyGraph` instance
+    """
+    scheme = get_scheme(scheme)
+    graph = DependencyGraph()
+    provided = {}  # maps names to lists of (version, dist) tuples
+
+    # first, build the graph and find out what's provided
+    for dist in dists:
+        graph.add_distribution(dist)
+
+        for p in dist.provides:
+            name, version = parse_name_and_version(p)
+            logger.debug('Add to provided: %s, %s, %s', name, version, dist)
+            provided.setdefault(name, []).append((version, dist))
+
+    # now make the edges
+    for dist in dists:
+        requires = (dist.run_requires | dist.meta_requires | dist.build_requires | dist.dev_requires)
+        for req in requires:
+            try:
+                matcher = scheme.matcher(req)
+            except UnsupportedVersionError:
+                # XXX compat-mode if cannot read the version
+                logger.warning('could not read version %r - using name only', req)
+                name = req.split()[0]
+                matcher = scheme.matcher(name)
+
+            name = matcher.key  # case-insensitive
+
+            matched = False
+            if name in provided:
+                for version, provider in provided[name]:
+                    try:
+                        match = matcher.match(version)
+                    except UnsupportedVersionError:
+                        match = False
+
+                    if match:
+                        graph.add_edge(dist, provider, req)
+                        matched = True
+                        break
+            if not matched:
+                graph.add_missing(dist, req)
+    return graph
+
+
+def get_dependent_dists(dists, dist):
+    """Recursively generate a list of distributions from *dists* that are
+    dependent on *dist*.
+
+    :param dists: a list of distributions
+    :param dist: a distribution, member of *dists* for which we are interested
+    """
+    if dist not in dists:
+        raise DistlibException('given distribution %r is not a member '
+                               'of the list' % dist.name)
+    graph = make_graph(dists)
+
+    dep = [dist]  # dependent distributions
+    todo = graph.reverse_list[dist]  # list of nodes we should inspect
+
+    while todo:
+        d = todo.pop()
+        dep.append(d)
+        for succ in graph.reverse_list[d]:
+            if succ not in dep:
+                todo.append(succ)
+
+    dep.pop(0)  # remove dist from dep, was there to prevent infinite loops
+    return dep
+
+
+def get_required_dists(dists, dist):
+    """Recursively generate a list of distributions from *dists* that are
+    required by *dist*.
+
+    :param dists: a list of distributions
+    :param dist: a distribution, member of *dists* for which we are interested
+                 in finding the dependencies.
+    """
+    if dist not in dists:
+        raise DistlibException('given distribution %r is not a member '
+                               'of the list' % dist.name)
+    graph = make_graph(dists)
+
+    req = set()  # required distributions
+    todo = graph.adjacency_list[dist]  # list of nodes we should inspect
+    seen = set(t[0] for t in todo)  # already added to todo
+
+    while todo:
+        d = todo.pop()[0]
+        req.add(d)
+        pred_list = graph.adjacency_list[d]
+        for pred in pred_list:
+            d = pred[0]
+            if d not in req and d not in seen:
+                seen.add(d)
+                todo.append(pred)
+    return req
+
+
+def make_dist(name, version, **kwargs):
+    """
+    A convenience method for making a dist given just a name and version.
+    """
+    summary = kwargs.pop('summary', 'Placeholder for summary')
+    md = Metadata(**kwargs)
+    md.name = name
+    md.version = version
+    md.summary = summary or 'Placeholder for summary'
+    return Distribution(md)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/index.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/index.py
new file mode 100644
index 00000000..56cd2867
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/index.py
@@ -0,0 +1,508 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2013-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+import hashlib
+import logging
+import os
+import shutil
+import subprocess
+import tempfile
+try:
+    from threading import Thread
+except ImportError:  # pragma: no cover
+    from dummy_threading import Thread
+
+from . import DistlibException
+from .compat import (HTTPBasicAuthHandler, Request, HTTPPasswordMgr,
+                     urlparse, build_opener, string_types)
+from .util import zip_dir, ServerProxy
+
+logger = logging.getLogger(__name__)
+
+DEFAULT_INDEX = 'https://pypi.org/pypi'
+DEFAULT_REALM = 'pypi'
+
+
+class PackageIndex(object):
+    """
+    This class represents a package index compatible with PyPI, the Python
+    Package Index.
+    """
+
+    boundary = b'----------ThIs_Is_tHe_distlib_index_bouNdaRY_$'
+
+    def __init__(self, url=None):
+        """
+        Initialise an instance.
+
+        :param url: The URL of the index. If not specified, the URL for PyPI is
+                    used.
+        """
+        self.url = url or DEFAULT_INDEX
+        self.read_configuration()
+        scheme, netloc, path, params, query, frag = urlparse(self.url)
+        if params or query or frag or scheme not in ('http', 'https'):
+            raise DistlibException('invalid repository: %s' % self.url)
+        self.password_handler = None
+        self.ssl_verifier = None
+        self.gpg = None
+        self.gpg_home = None
+        with open(os.devnull, 'w') as sink:
+            # Use gpg by default rather than gpg2, as gpg2 insists on
+            # prompting for passwords
+            for s in ('gpg', 'gpg2'):
+                try:
+                    rc = subprocess.check_call([s, '--version'], stdout=sink,
+                                               stderr=sink)
+                    if rc == 0:
+                        self.gpg = s
+                        break
+                except OSError:
+                    pass
+
+    def _get_pypirc_command(self):
+        """
+        Get the distutils command for interacting with PyPI configurations.
+        :return: the command.
+        """
+        from .util import _get_pypirc_command as cmd
+        return cmd()
+
+    def read_configuration(self):
+        """
+        Read the PyPI access configuration as supported by distutils. This populates
+        ``username``, ``password``, ``realm`` and ``url`` attributes from the
+        configuration.
+        """
+        from .util import _load_pypirc
+        cfg = _load_pypirc(self)
+        self.username = cfg.get('username')
+        self.password = cfg.get('password')
+        self.realm = cfg.get('realm', 'pypi')
+        self.url = cfg.get('repository', self.url)
+
+    def save_configuration(self):
+        """
+        Save the PyPI access configuration. You must have set ``username`` and
+        ``password`` attributes before calling this method.
+        """
+        self.check_credentials()
+        from .util import _store_pypirc
+        _store_pypirc(self)
+
+    def check_credentials(self):
+        """
+        Check that ``username`` and ``password`` have been set, and raise an
+        exception if not.
+        """
+        if self.username is None or self.password is None:
+            raise DistlibException('username and password must be set')
+        pm = HTTPPasswordMgr()
+        _, netloc, _, _, _, _ = urlparse(self.url)
+        pm.add_password(self.realm, netloc, self.username, self.password)
+        self.password_handler = HTTPBasicAuthHandler(pm)
+
+    def register(self, metadata):  # pragma: no cover
+        """
+        Register a distribution on PyPI, using the provided metadata.
+
+        :param metadata: A :class:`Metadata` instance defining at least a name
+                         and version number for the distribution to be
+                         registered.
+        :return: The HTTP response received from PyPI upon submission of the
+                request.
+        """
+        self.check_credentials()
+        metadata.validate()
+        d = metadata.todict()
+        d[':action'] = 'verify'
+        request = self.encode_request(d.items(), [])
+        self.send_request(request)
+        d[':action'] = 'submit'
+        request = self.encode_request(d.items(), [])
+        return self.send_request(request)
+
+    def _reader(self, name, stream, outbuf):
+        """
+        Thread runner for reading lines of from a subprocess into a buffer.
+
+        :param name: The logical name of the stream (used for logging only).
+        :param stream: The stream to read from. This will typically a pipe
+                       connected to the output stream of a subprocess.
+        :param outbuf: The list to append the read lines to.
+        """
+        while True:
+            s = stream.readline()
+            if not s:
+                break
+            s = s.decode('utf-8').rstrip()
+            outbuf.append(s)
+            logger.debug('%s: %s' % (name, s))
+        stream.close()
+
+    def get_sign_command(self, filename, signer, sign_password, keystore=None):  # pragma: no cover
+        """
+        Return a suitable command for signing a file.
+
+        :param filename: The pathname to the file to be signed.
+        :param signer: The identifier of the signer of the file.
+        :param sign_password: The passphrase for the signer's
+                              private key used for signing.
+        :param keystore: The path to a directory which contains the keys
+                         used in verification. If not specified, the
+                         instance's ``gpg_home`` attribute is used instead.
+        :return: The signing command as a list suitable to be
+                 passed to :class:`subprocess.Popen`.
+        """
+        cmd = [self.gpg, '--status-fd', '2', '--no-tty']
+        if keystore is None:
+            keystore = self.gpg_home
+        if keystore:
+            cmd.extend(['--homedir', keystore])
+        if sign_password is not None:
+            cmd.extend(['--batch', '--passphrase-fd', '0'])
+        td = tempfile.mkdtemp()
+        sf = os.path.join(td, os.path.basename(filename) + '.asc')
+        cmd.extend(['--detach-sign', '--armor', '--local-user',
+                    signer, '--output', sf, filename])
+        logger.debug('invoking: %s', ' '.join(cmd))
+        return cmd, sf
+
+    def run_command(self, cmd, input_data=None):
+        """
+        Run a command in a child process , passing it any input data specified.
+
+        :param cmd: The command to run.
+        :param input_data: If specified, this must be a byte string containing
+                           data to be sent to the child process.
+        :return: A tuple consisting of the subprocess' exit code, a list of
+                 lines read from the subprocess' ``stdout``, and a list of
+                 lines read from the subprocess' ``stderr``.
+        """
+        kwargs = {
+            'stdout': subprocess.PIPE,
+            'stderr': subprocess.PIPE,
+        }
+        if input_data is not None:
+            kwargs['stdin'] = subprocess.PIPE
+        stdout = []
+        stderr = []
+        p = subprocess.Popen(cmd, **kwargs)
+        # We don't use communicate() here because we may need to
+        # get clever with interacting with the command
+        t1 = Thread(target=self._reader, args=('stdout', p.stdout, stdout))
+        t1.start()
+        t2 = Thread(target=self._reader, args=('stderr', p.stderr, stderr))
+        t2.start()
+        if input_data is not None:
+            p.stdin.write(input_data)
+            p.stdin.close()
+
+        p.wait()
+        t1.join()
+        t2.join()
+        return p.returncode, stdout, stderr
+
+    def sign_file(self, filename, signer, sign_password, keystore=None):  # pragma: no cover
+        """
+        Sign a file.
+
+        :param filename: The pathname to the file to be signed.
+        :param signer: The identifier of the signer of the file.
+        :param sign_password: The passphrase for the signer's
+                              private key used for signing.
+        :param keystore: The path to a directory which contains the keys
+                         used in signing. If not specified, the instance's
+                         ``gpg_home`` attribute is used instead.
+        :return: The absolute pathname of the file where the signature is
+                 stored.
+        """
+        cmd, sig_file = self.get_sign_command(filename, signer, sign_password,
+                                              keystore)
+        rc, stdout, stderr = self.run_command(cmd,
+                                              sign_password.encode('utf-8'))
+        if rc != 0:
+            raise DistlibException('sign command failed with error '
+                                   'code %s' % rc)
+        return sig_file
+
+    def upload_file(self, metadata, filename, signer=None, sign_password=None,
+                    filetype='sdist', pyversion='source', keystore=None):
+        """
+        Upload a release file to the index.
+
+        :param metadata: A :class:`Metadata` instance defining at least a name
+                         and version number for the file to be uploaded.
+        :param filename: The pathname of the file to be uploaded.
+        :param signer: The identifier of the signer of the file.
+        :param sign_password: The passphrase for the signer's
+                              private key used for signing.
+        :param filetype: The type of the file being uploaded. This is the
+                        distutils command which produced that file, e.g.
+                        ``sdist`` or ``bdist_wheel``.
+        :param pyversion: The version of Python which the release relates
+                          to. For code compatible with any Python, this would
+                          be ``source``, otherwise it would be e.g. ``3.2``.
+        :param keystore: The path to a directory which contains the keys
+                         used in signing. If not specified, the instance's
+                         ``gpg_home`` attribute is used instead.
+        :return: The HTTP response received from PyPI upon submission of the
+                request.
+        """
+        self.check_credentials()
+        if not os.path.exists(filename):
+            raise DistlibException('not found: %s' % filename)
+        metadata.validate()
+        d = metadata.todict()
+        sig_file = None
+        if signer:
+            if not self.gpg:
+                logger.warning('no signing program available - not signed')
+            else:
+                sig_file = self.sign_file(filename, signer, sign_password,
+                                          keystore)
+        with open(filename, 'rb') as f:
+            file_data = f.read()
+        md5_digest = hashlib.md5(file_data).hexdigest()
+        sha256_digest = hashlib.sha256(file_data).hexdigest()
+        d.update({
+            ':action': 'file_upload',
+            'protocol_version': '1',
+            'filetype': filetype,
+            'pyversion': pyversion,
+            'md5_digest': md5_digest,
+            'sha256_digest': sha256_digest,
+        })
+        files = [('content', os.path.basename(filename), file_data)]
+        if sig_file:
+            with open(sig_file, 'rb') as f:
+                sig_data = f.read()
+            files.append(('gpg_signature', os.path.basename(sig_file),
+                         sig_data))
+            shutil.rmtree(os.path.dirname(sig_file))
+        request = self.encode_request(d.items(), files)
+        return self.send_request(request)
+
+    def upload_documentation(self, metadata, doc_dir):  # pragma: no cover
+        """
+        Upload documentation to the index.
+
+        :param metadata: A :class:`Metadata` instance defining at least a name
+                         and version number for the documentation to be
+                         uploaded.
+        :param doc_dir: The pathname of the directory which contains the
+                        documentation. This should be the directory that
+                        contains the ``index.html`` for the documentation.
+        :return: The HTTP response received from PyPI upon submission of the
+                request.
+        """
+        self.check_credentials()
+        if not os.path.isdir(doc_dir):
+            raise DistlibException('not a directory: %r' % doc_dir)
+        fn = os.path.join(doc_dir, 'index.html')
+        if not os.path.exists(fn):
+            raise DistlibException('not found: %r' % fn)
+        metadata.validate()
+        name, version = metadata.name, metadata.version
+        zip_data = zip_dir(doc_dir).getvalue()
+        fields = [(':action', 'doc_upload'),
+                  ('name', name), ('version', version)]
+        files = [('content', name, zip_data)]
+        request = self.encode_request(fields, files)
+        return self.send_request(request)
+
+    def get_verify_command(self, signature_filename, data_filename,
+                           keystore=None):
+        """
+        Return a suitable command for verifying a file.
+
+        :param signature_filename: The pathname to the file containing the
+                                   signature.
+        :param data_filename: The pathname to the file containing the
+                              signed data.
+        :param keystore: The path to a directory which contains the keys
+                         used in verification. If not specified, the
+                         instance's ``gpg_home`` attribute is used instead.
+        :return: The verifying command as a list suitable to be
+                 passed to :class:`subprocess.Popen`.
+        """
+        cmd = [self.gpg, '--status-fd', '2', '--no-tty']
+        if keystore is None:
+            keystore = self.gpg_home
+        if keystore:
+            cmd.extend(['--homedir', keystore])
+        cmd.extend(['--verify', signature_filename, data_filename])
+        logger.debug('invoking: %s', ' '.join(cmd))
+        return cmd
+
+    def verify_signature(self, signature_filename, data_filename,
+                         keystore=None):
+        """
+        Verify a signature for a file.
+
+        :param signature_filename: The pathname to the file containing the
+                                   signature.
+        :param data_filename: The pathname to the file containing the
+                              signed data.
+        :param keystore: The path to a directory which contains the keys
+                         used in verification. If not specified, the
+                         instance's ``gpg_home`` attribute is used instead.
+        :return: True if the signature was verified, else False.
+        """
+        if not self.gpg:
+            raise DistlibException('verification unavailable because gpg '
+                                   'unavailable')
+        cmd = self.get_verify_command(signature_filename, data_filename,
+                                      keystore)
+        rc, stdout, stderr = self.run_command(cmd)
+        if rc not in (0, 1):
+            raise DistlibException('verify command failed with error code %s' % rc)
+        return rc == 0
+
+    def download_file(self, url, destfile, digest=None, reporthook=None):
+        """
+        This is a convenience method for downloading a file from an URL.
+        Normally, this will be a file from the index, though currently
+        no check is made for this (i.e. a file can be downloaded from
+        anywhere).
+
+        The method is just like the :func:`urlretrieve` function in the
+        standard library, except that it allows digest computation to be
+        done during download and checking that the downloaded data
+        matched any expected value.
+
+        :param url: The URL of the file to be downloaded (assumed to be
+                    available via an HTTP GET request).
+        :param destfile: The pathname where the downloaded file is to be
+                         saved.
+        :param digest: If specified, this must be a (hasher, value)
+                       tuple, where hasher is the algorithm used (e.g.
+                       ``'md5'``) and ``value`` is the expected value.
+        :param reporthook: The same as for :func:`urlretrieve` in the
+                           standard library.
+        """
+        if digest is None:
+            digester = None
+            logger.debug('No digest specified')
+        else:
+            if isinstance(digest, (list, tuple)):
+                hasher, digest = digest
+            else:
+                hasher = 'md5'
+            digester = getattr(hashlib, hasher)()
+            logger.debug('Digest specified: %s' % digest)
+        # The following code is equivalent to urlretrieve.
+        # We need to do it this way so that we can compute the
+        # digest of the file as we go.
+        with open(destfile, 'wb') as dfp:
+            # addinfourl is not a context manager on 2.x
+            # so we have to use try/finally
+            sfp = self.send_request(Request(url))
+            try:
+                headers = sfp.info()
+                blocksize = 8192
+                size = -1
+                read = 0
+                blocknum = 0
+                if "content-length" in headers:
+                    size = int(headers["Content-Length"])
+                if reporthook:
+                    reporthook(blocknum, blocksize, size)
+                while True:
+                    block = sfp.read(blocksize)
+                    if not block:
+                        break
+                    read += len(block)
+                    dfp.write(block)
+                    if digester:
+                        digester.update(block)
+                    blocknum += 1
+                    if reporthook:
+                        reporthook(blocknum, blocksize, size)
+            finally:
+                sfp.close()
+
+        # check that we got the whole file, if we can
+        if size >= 0 and read < size:
+            raise DistlibException(
+                'retrieval incomplete: got only %d out of %d bytes'
+                % (read, size))
+        # if we have a digest, it must match.
+        if digester:
+            actual = digester.hexdigest()
+            if digest != actual:
+                raise DistlibException('%s digest mismatch for %s: expected '
+                                       '%s, got %s' % (hasher, destfile,
+                                                       digest, actual))
+            logger.debug('Digest verified: %s', digest)
+
+    def send_request(self, req):
+        """
+        Send a standard library :class:`Request` to PyPI and return its
+        response.
+
+        :param req: The request to send.
+        :return: The HTTP response from PyPI (a standard library HTTPResponse).
+        """
+        handlers = []
+        if self.password_handler:
+            handlers.append(self.password_handler)
+        if self.ssl_verifier:
+            handlers.append(self.ssl_verifier)
+        opener = build_opener(*handlers)
+        return opener.open(req)
+
+    def encode_request(self, fields, files):
+        """
+        Encode fields and files for posting to an HTTP server.
+
+        :param fields: The fields to send as a list of (fieldname, value)
+                       tuples.
+        :param files: The files to send as a list of (fieldname, filename,
+                      file_bytes) tuple.
+        """
+        # Adapted from packaging, which in turn was adapted from
+        # http://code.activestate.com/recipes/146306
+
+        parts = []
+        boundary = self.boundary
+        for k, values in fields:
+            if not isinstance(values, (list, tuple)):
+                values = [values]
+
+            for v in values:
+                parts.extend((
+                    b'--' + boundary,
+                    ('Content-Disposition: form-data; name="%s"' %
+                     k).encode('utf-8'),
+                    b'',
+                    v.encode('utf-8')))
+        for key, filename, value in files:
+            parts.extend((
+                b'--' + boundary,
+                ('Content-Disposition: form-data; name="%s"; filename="%s"' %
+                 (key, filename)).encode('utf-8'),
+                b'',
+                value))
+
+        parts.extend((b'--' + boundary + b'--', b''))
+
+        body = b'\r\n'.join(parts)
+        ct = b'multipart/form-data; boundary=' + boundary
+        headers = {
+            'Content-type': ct,
+            'Content-length': str(len(body))
+        }
+        return Request(self.url, body, headers)
+
+    def search(self, terms, operator=None):  # pragma: no cover
+        if isinstance(terms, string_types):
+            terms = {'name': terms}
+        rpc_proxy = ServerProxy(self.url, timeout=3.0)
+        try:
+            return rpc_proxy.search(terms, operator or 'and')
+        finally:
+            rpc_proxy('close')()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/locators.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/locators.py
new file mode 100644
index 00000000..222c1bf3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/locators.py
@@ -0,0 +1,1295 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+
+import gzip
+from io import BytesIO
+import json
+import logging
+import os
+import posixpath
+import re
+try:
+    import threading
+except ImportError:  # pragma: no cover
+    import dummy_threading as threading
+import zlib
+
+from . import DistlibException
+from .compat import (urljoin, urlparse, urlunparse, url2pathname, pathname2url, queue, quote, unescape, build_opener,
+                     HTTPRedirectHandler as BaseRedirectHandler, text_type, Request, HTTPError, URLError)
+from .database import Distribution, DistributionPath, make_dist
+from .metadata import Metadata, MetadataInvalidError
+from .util import (cached_property, ensure_slash, split_filename, get_project_data, parse_requirement,
+                   parse_name_and_version, ServerProxy, normalize_name)
+from .version import get_scheme, UnsupportedVersionError
+from .wheel import Wheel, is_compatible
+
+logger = logging.getLogger(__name__)
+
+HASHER_HASH = re.compile(r'^(\w+)=([a-f0-9]+)')
+CHARSET = re.compile(r';\s*charset\s*=\s*(.*)\s*$', re.I)
+HTML_CONTENT_TYPE = re.compile('text/html|application/x(ht)?ml')
+DEFAULT_INDEX = 'https://pypi.org/pypi'
+
+
+def get_all_distribution_names(url=None):
+    """
+    Return all distribution names known by an index.
+    :param url: The URL of the index.
+    :return: A list of all known distribution names.
+    """
+    if url is None:
+        url = DEFAULT_INDEX
+    client = ServerProxy(url, timeout=3.0)
+    try:
+        return client.list_packages()
+    finally:
+        client('close')()
+
+
+class RedirectHandler(BaseRedirectHandler):
+    """
+    A class to work around a bug in some Python 3.2.x releases.
+    """
+
+    # There's a bug in the base version for some 3.2.x
+    # (e.g. 3.2.2 on Ubuntu Oneiric). If a Location header
+    # returns e.g. /abc, it bails because it says the scheme ''
+    # is bogus, when actually it should use the request's
+    # URL for the scheme. See Python issue #13696.
+    def http_error_302(self, req, fp, code, msg, headers):
+        # Some servers (incorrectly) return multiple Location headers
+        # (so probably same goes for URI).  Use first header.
+        newurl = None
+        for key in ('location', 'uri'):
+            if key in headers:
+                newurl = headers[key]
+                break
+        if newurl is None:  # pragma: no cover
+            return
+        urlparts = urlparse(newurl)
+        if urlparts.scheme == '':
+            newurl = urljoin(req.get_full_url(), newurl)
+            if hasattr(headers, 'replace_header'):
+                headers.replace_header(key, newurl)
+            else:
+                headers[key] = newurl
+        return BaseRedirectHandler.http_error_302(self, req, fp, code, msg, headers)
+
+    http_error_301 = http_error_303 = http_error_307 = http_error_302
+
+
+class Locator(object):
+    """
+    A base class for locators - things that locate distributions.
+    """
+    source_extensions = ('.tar.gz', '.tar.bz2', '.tar', '.zip', '.tgz', '.tbz')
+    binary_extensions = ('.egg', '.exe', '.whl')
+    excluded_extensions = ('.pdf', )
+
+    # A list of tags indicating which wheels you want to match. The default
+    # value of None matches against the tags compatible with the running
+    # Python. If you want to match other values, set wheel_tags on a locator
+    # instance to a list of tuples (pyver, abi, arch) which you want to match.
+    wheel_tags = None
+
+    downloadable_extensions = source_extensions + ('.whl', )
+
+    def __init__(self, scheme='default'):
+        """
+        Initialise an instance.
+        :param scheme: Because locators look for most recent versions, they
+                       need to know the version scheme to use. This specifies
+                       the current PEP-recommended scheme - use ``'legacy'``
+                       if you need to support existing distributions on PyPI.
+        """
+        self._cache = {}
+        self.scheme = scheme
+        # Because of bugs in some of the handlers on some of the platforms,
+        # we use our own opener rather than just using urlopen.
+        self.opener = build_opener(RedirectHandler())
+        # If get_project() is called from locate(), the matcher instance
+        # is set from the requirement passed to locate(). See issue #18 for
+        # why this can be useful to know.
+        self.matcher = None
+        self.errors = queue.Queue()
+
+    def get_errors(self):
+        """
+        Return any errors which have occurred.
+        """
+        result = []
+        while not self.errors.empty():  # pragma: no cover
+            try:
+                e = self.errors.get(False)
+                result.append(e)
+            except self.errors.Empty:
+                continue
+            self.errors.task_done()
+        return result
+
+    def clear_errors(self):
+        """
+        Clear any errors which may have been logged.
+        """
+        # Just get the errors and throw them away
+        self.get_errors()
+
+    def clear_cache(self):
+        self._cache.clear()
+
+    def _get_scheme(self):
+        return self._scheme
+
+    def _set_scheme(self, value):
+        self._scheme = value
+
+    scheme = property(_get_scheme, _set_scheme)
+
+    def _get_project(self, name):
+        """
+        For a given project, get a dictionary mapping available versions to Distribution
+        instances.
+
+        This should be implemented in subclasses.
+
+        If called from a locate() request, self.matcher will be set to a
+        matcher for the requirement to satisfy, otherwise it will be None.
+        """
+        raise NotImplementedError('Please implement in the subclass')
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        raise NotImplementedError('Please implement in the subclass')
+
+    def get_project(self, name):
+        """
+        For a given project, get a dictionary mapping available versions to Distribution
+        instances.
+
+        This calls _get_project to do all the work, and just implements a caching layer on top.
+        """
+        if self._cache is None:  # pragma: no cover
+            result = self._get_project(name)
+        elif name in self._cache:
+            result = self._cache[name]
+        else:
+            self.clear_errors()
+            result = self._get_project(name)
+            self._cache[name] = result
+        return result
+
+    def score_url(self, url):
+        """
+        Give an url a score which can be used to choose preferred URLs
+        for a given project release.
+        """
+        t = urlparse(url)
+        basename = posixpath.basename(t.path)
+        compatible = True
+        is_wheel = basename.endswith('.whl')
+        is_downloadable = basename.endswith(self.downloadable_extensions)
+        if is_wheel:
+            compatible = is_compatible(Wheel(basename), self.wheel_tags)
+        return (t.scheme == 'https', 'pypi.org' in t.netloc, is_downloadable, is_wheel, compatible, basename)
+
+    def prefer_url(self, url1, url2):
+        """
+        Choose one of two URLs where both are candidates for distribution
+        archives for the same version of a distribution (for example,
+        .tar.gz vs. zip).
+
+        The current implementation favours https:// URLs over http://, archives
+        from PyPI over those from other locations, wheel compatibility (if a
+        wheel) and then the archive name.
+        """
+        result = url2
+        if url1:
+            s1 = self.score_url(url1)
+            s2 = self.score_url(url2)
+            if s1 > s2:
+                result = url1
+            if result != url2:
+                logger.debug('Not replacing %r with %r', url1, url2)
+            else:
+                logger.debug('Replacing %r with %r', url1, url2)
+        return result
+
+    def split_filename(self, filename, project_name):
+        """
+        Attempt to split a filename in project name, version and Python version.
+        """
+        return split_filename(filename, project_name)
+
+    def convert_url_to_download_info(self, url, project_name):
+        """
+        See if a URL is a candidate for a download URL for a project (the URL
+        has typically been scraped from an HTML page).
+
+        If it is, a dictionary is returned with keys "name", "version",
+        "filename" and "url"; otherwise, None is returned.
+        """
+
+        def same_project(name1, name2):
+            return normalize_name(name1) == normalize_name(name2)
+
+        result = None
+        scheme, netloc, path, params, query, frag = urlparse(url)
+        if frag.lower().startswith('egg='):  # pragma: no cover
+            logger.debug('%s: version hint in fragment: %r', project_name, frag)
+        m = HASHER_HASH.match(frag)
+        if m:
+            algo, digest = m.groups()
+        else:
+            algo, digest = None, None
+        origpath = path
+        if path and path[-1] == '/':  # pragma: no cover
+            path = path[:-1]
+        if path.endswith('.whl'):
+            try:
+                wheel = Wheel(path)
+                if not is_compatible(wheel, self.wheel_tags):
+                    logger.debug('Wheel not compatible: %s', path)
+                else:
+                    if project_name is None:
+                        include = True
+                    else:
+                        include = same_project(wheel.name, project_name)
+                    if include:
+                        result = {
+                            'name': wheel.name,
+                            'version': wheel.version,
+                            'filename': wheel.filename,
+                            'url': urlunparse((scheme, netloc, origpath, params, query, '')),
+                            'python-version': ', '.join(['.'.join(list(v[2:])) for v in wheel.pyver]),
+                        }
+            except Exception:  # pragma: no cover
+                logger.warning('invalid path for wheel: %s', path)
+        elif not path.endswith(self.downloadable_extensions):  # pragma: no cover
+            logger.debug('Not downloadable: %s', path)
+        else:  # downloadable extension
+            path = filename = posixpath.basename(path)
+            for ext in self.downloadable_extensions:
+                if path.endswith(ext):
+                    path = path[:-len(ext)]
+                    t = self.split_filename(path, project_name)
+                    if not t:  # pragma: no cover
+                        logger.debug('No match for project/version: %s', path)
+                    else:
+                        name, version, pyver = t
+                        if not project_name or same_project(project_name, name):
+                            result = {
+                                'name': name,
+                                'version': version,
+                                'filename': filename,
+                                'url': urlunparse((scheme, netloc, origpath, params, query, '')),
+                            }
+                            if pyver:  # pragma: no cover
+                                result['python-version'] = pyver
+                    break
+        if result and algo:
+            result['%s_digest' % algo] = digest
+        return result
+
+    def _get_digest(self, info):
+        """
+        Get a digest from a dictionary by looking at a "digests" dictionary
+        or keys of the form 'algo_digest'.
+
+        Returns a 2-tuple (algo, digest) if found, else None. Currently
+        looks only for SHA256, then MD5.
+        """
+        result = None
+        if 'digests' in info:
+            digests = info['digests']
+            for algo in ('sha256', 'md5'):
+                if algo in digests:
+                    result = (algo, digests[algo])
+                    break
+        if not result:
+            for algo in ('sha256', 'md5'):
+                key = '%s_digest' % algo
+                if key in info:
+                    result = (algo, info[key])
+                    break
+        return result
+
+    def _update_version_data(self, result, info):
+        """
+        Update a result dictionary (the final result from _get_project) with a
+        dictionary for a specific version, which typically holds information
+        gleaned from a filename or URL for an archive for the distribution.
+        """
+        name = info.pop('name')
+        version = info.pop('version')
+        if version in result:
+            dist = result[version]
+            md = dist.metadata
+        else:
+            dist = make_dist(name, version, scheme=self.scheme)
+            md = dist.metadata
+        dist.digest = digest = self._get_digest(info)
+        url = info['url']
+        result['digests'][url] = digest
+        if md.source_url != info['url']:
+            md.source_url = self.prefer_url(md.source_url, url)
+            result['urls'].setdefault(version, set()).add(url)
+        dist.locator = self
+        result[version] = dist
+
+    def locate(self, requirement, prereleases=False):
+        """
+        Find the most recent distribution which matches the given
+        requirement.
+
+        :param requirement: A requirement of the form 'foo (1.0)' or perhaps
+                            'foo (>= 1.0, < 2.0, != 1.3)'
+        :param prereleases: If ``True``, allow pre-release versions
+                            to be located. Otherwise, pre-release versions
+                            are not returned.
+        :return: A :class:`Distribution` instance, or ``None`` if no such
+                 distribution could be located.
+        """
+        result = None
+        r = parse_requirement(requirement)
+        if r is None:  # pragma: no cover
+            raise DistlibException('Not a valid requirement: %r' % requirement)
+        scheme = get_scheme(self.scheme)
+        self.matcher = matcher = scheme.matcher(r.requirement)
+        logger.debug('matcher: %s (%s)', matcher, type(matcher).__name__)
+        versions = self.get_project(r.name)
+        if len(versions) > 2:  # urls and digests keys are present
+            # sometimes, versions are invalid
+            slist = []
+            vcls = matcher.version_class
+            for k in versions:
+                if k in ('urls', 'digests'):
+                    continue
+                try:
+                    if not matcher.match(k):
+                        pass  # logger.debug('%s did not match %r', matcher, k)
+                    else:
+                        if prereleases or not vcls(k).is_prerelease:
+                            slist.append(k)
+                except Exception:  # pragma: no cover
+                    logger.warning('error matching %s with %r', matcher, k)
+                    pass  # slist.append(k)
+            if len(slist) > 1:
+                slist = sorted(slist, key=scheme.key)
+            if slist:
+                logger.debug('sorted list: %s', slist)
+                version = slist[-1]
+                result = versions[version]
+        if result:
+            if r.extras:
+                result.extras = r.extras
+            result.download_urls = versions.get('urls', {}).get(version, set())
+            d = {}
+            sd = versions.get('digests', {})
+            for url in result.download_urls:
+                if url in sd:  # pragma: no cover
+                    d[url] = sd[url]
+            result.digests = d
+        self.matcher = None
+        return result
+
+
+class PyPIRPCLocator(Locator):
+    """
+    This locator uses XML-RPC to locate distributions. It therefore
+    cannot be used with simple mirrors (that only mirror file content).
+    """
+
+    def __init__(self, url, **kwargs):
+        """
+        Initialise an instance.
+
+        :param url: The URL to use for XML-RPC.
+        :param kwargs: Passed to the superclass constructor.
+        """
+        super(PyPIRPCLocator, self).__init__(**kwargs)
+        self.base_url = url
+        self.client = ServerProxy(url, timeout=3.0)
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        return set(self.client.list_packages())
+
+    def _get_project(self, name):
+        result = {'urls': {}, 'digests': {}}
+        versions = self.client.package_releases(name, True)
+        for v in versions:
+            urls = self.client.release_urls(name, v)
+            data = self.client.release_data(name, v)
+            metadata = Metadata(scheme=self.scheme)
+            metadata.name = data['name']
+            metadata.version = data['version']
+            metadata.license = data.get('license')
+            metadata.keywords = data.get('keywords', [])
+            metadata.summary = data.get('summary')
+            dist = Distribution(metadata)
+            if urls:
+                info = urls[0]
+                metadata.source_url = info['url']
+                dist.digest = self._get_digest(info)
+                dist.locator = self
+                result[v] = dist
+                for info in urls:
+                    url = info['url']
+                    digest = self._get_digest(info)
+                    result['urls'].setdefault(v, set()).add(url)
+                    result['digests'][url] = digest
+        return result
+
+
+class PyPIJSONLocator(Locator):
+    """
+    This locator uses PyPI's JSON interface. It's very limited in functionality
+    and probably not worth using.
+    """
+
+    def __init__(self, url, **kwargs):
+        super(PyPIJSONLocator, self).__init__(**kwargs)
+        self.base_url = ensure_slash(url)
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        raise NotImplementedError('Not available from this locator')
+
+    def _get_project(self, name):
+        result = {'urls': {}, 'digests': {}}
+        url = urljoin(self.base_url, '%s/json' % quote(name))
+        try:
+            resp = self.opener.open(url)
+            data = resp.read().decode()  # for now
+            d = json.loads(data)
+            md = Metadata(scheme=self.scheme)
+            data = d['info']
+            md.name = data['name']
+            md.version = data['version']
+            md.license = data.get('license')
+            md.keywords = data.get('keywords', [])
+            md.summary = data.get('summary')
+            dist = Distribution(md)
+            dist.locator = self
+            # urls = d['urls']
+            result[md.version] = dist
+            for info in d['urls']:
+                url = info['url']
+                dist.download_urls.add(url)
+                dist.digests[url] = self._get_digest(info)
+                result['urls'].setdefault(md.version, set()).add(url)
+                result['digests'][url] = self._get_digest(info)
+            # Now get other releases
+            for version, infos in d['releases'].items():
+                if version == md.version:
+                    continue  # already done
+                omd = Metadata(scheme=self.scheme)
+                omd.name = md.name
+                omd.version = version
+                odist = Distribution(omd)
+                odist.locator = self
+                result[version] = odist
+                for info in infos:
+                    url = info['url']
+                    odist.download_urls.add(url)
+                    odist.digests[url] = self._get_digest(info)
+                    result['urls'].setdefault(version, set()).add(url)
+                    result['digests'][url] = self._get_digest(info)
+
+
+#            for info in urls:
+#                md.source_url = info['url']
+#                dist.digest = self._get_digest(info)
+#                dist.locator = self
+#                for info in urls:
+#                    url = info['url']
+#                    result['urls'].setdefault(md.version, set()).add(url)
+#                    result['digests'][url] = self._get_digest(info)
+        except Exception as e:
+            self.errors.put(text_type(e))
+            logger.exception('JSON fetch failed: %s', e)
+        return result
+
+
+class Page(object):
+    """
+    This class represents a scraped HTML page.
+    """
+    # The following slightly hairy-looking regex just looks for the contents of
+    # an anchor link, which has an attribute "href" either immediately preceded
+    # or immediately followed by a "rel" attribute. The attribute values can be
+    # declared with double quotes, single quotes or no quotes - which leads to
+    # the length of the expression.
+    _href = re.compile(
+        """
+(rel\\s*=\\s*(?:"(?P<rel1>[^"]*)"|'(?P<rel2>[^']*)'|(?P<rel3>[^>\\s\n]*))\\s+)?
+href\\s*=\\s*(?:"(?P<url1>[^"]*)"|'(?P<url2>[^']*)'|(?P<url3>[^>\\s\n]*))
+(\\s+rel\\s*=\\s*(?:"(?P<rel4>[^"]*)"|'(?P<rel5>[^']*)'|(?P<rel6>[^>\\s\n]*)))?
+""", re.I | re.S | re.X)
+    _base = re.compile(r"""<base\s+href\s*=\s*['"]?([^'">]+)""", re.I | re.S)
+
+    def __init__(self, data, url):
+        """
+        Initialise an instance with the Unicode page contents and the URL they
+        came from.
+        """
+        self.data = data
+        self.base_url = self.url = url
+        m = self._base.search(self.data)
+        if m:
+            self.base_url = m.group(1)
+
+    _clean_re = re.compile(r'[^a-z0-9$&+,/:;=?@.#%_\\|-]', re.I)
+
+    @cached_property
+    def links(self):
+        """
+        Return the URLs of all the links on a page together with information
+        about their "rel" attribute, for determining which ones to treat as
+        downloads and which ones to queue for further scraping.
+        """
+
+        def clean(url):
+            "Tidy up an URL."
+            scheme, netloc, path, params, query, frag = urlparse(url)
+            return urlunparse((scheme, netloc, quote(path), params, query, frag))
+
+        result = set()
+        for match in self._href.finditer(self.data):
+            d = match.groupdict('')
+            rel = (d['rel1'] or d['rel2'] or d['rel3'] or d['rel4'] or d['rel5'] or d['rel6'])
+            url = d['url1'] or d['url2'] or d['url3']
+            url = urljoin(self.base_url, url)
+            url = unescape(url)
+            url = self._clean_re.sub(lambda m: '%%%2x' % ord(m.group(0)), url)
+            result.add((url, rel))
+        # We sort the result, hoping to bring the most recent versions
+        # to the front
+        result = sorted(result, key=lambda t: t[0], reverse=True)
+        return result
+
+
+class SimpleScrapingLocator(Locator):
+    """
+    A locator which scrapes HTML pages to locate downloads for a distribution.
+    This runs multiple threads to do the I/O; performance is at least as good
+    as pip's PackageFinder, which works in an analogous fashion.
+    """
+
+    # These are used to deal with various Content-Encoding schemes.
+    decoders = {
+        'deflate': zlib.decompress,
+        'gzip': lambda b: gzip.GzipFile(fileobj=BytesIO(b)).read(),
+        'none': lambda b: b,
+    }
+
+    def __init__(self, url, timeout=None, num_workers=10, **kwargs):
+        """
+        Initialise an instance.
+        :param url: The root URL to use for scraping.
+        :param timeout: The timeout, in seconds, to be applied to requests.
+                        This defaults to ``None`` (no timeout specified).
+        :param num_workers: The number of worker threads you want to do I/O,
+                            This defaults to 10.
+        :param kwargs: Passed to the superclass.
+        """
+        super(SimpleScrapingLocator, self).__init__(**kwargs)
+        self.base_url = ensure_slash(url)
+        self.timeout = timeout
+        self._page_cache = {}
+        self._seen = set()
+        self._to_fetch = queue.Queue()
+        self._bad_hosts = set()
+        self.skip_externals = False
+        self.num_workers = num_workers
+        self._lock = threading.RLock()
+        # See issue #45: we need to be resilient when the locator is used
+        # in a thread, e.g. with concurrent.futures. We can't use self._lock
+        # as it is for coordinating our internal threads - the ones created
+        # in _prepare_threads.
+        self._gplock = threading.RLock()
+        self.platform_check = False  # See issue #112
+
+    def _prepare_threads(self):
+        """
+        Threads are created only when get_project is called, and terminate
+        before it returns. They are there primarily to parallelise I/O (i.e.
+        fetching web pages).
+        """
+        self._threads = []
+        for i in range(self.num_workers):
+            t = threading.Thread(target=self._fetch)
+            t.daemon = True
+            t.start()
+            self._threads.append(t)
+
+    def _wait_threads(self):
+        """
+        Tell all the threads to terminate (by sending a sentinel value) and
+        wait for them to do so.
+        """
+        # Note that you need two loops, since you can't say which
+        # thread will get each sentinel
+        for t in self._threads:
+            self._to_fetch.put(None)  # sentinel
+        for t in self._threads:
+            t.join()
+        self._threads = []
+
+    def _get_project(self, name):
+        result = {'urls': {}, 'digests': {}}
+        with self._gplock:
+            self.result = result
+            self.project_name = name
+            url = urljoin(self.base_url, '%s/' % quote(name))
+            self._seen.clear()
+            self._page_cache.clear()
+            self._prepare_threads()
+            try:
+                logger.debug('Queueing %s', url)
+                self._to_fetch.put(url)
+                self._to_fetch.join()
+            finally:
+                self._wait_threads()
+            del self.result
+        return result
+
+    platform_dependent = re.compile(r'\b(linux_(i\d86|x86_64|arm\w+)|'
+                                    r'win(32|_amd64)|macosx_?\d+)\b', re.I)
+
+    def _is_platform_dependent(self, url):
+        """
+        Does an URL refer to a platform-specific download?
+        """
+        return self.platform_dependent.search(url)
+
+    def _process_download(self, url):
+        """
+        See if an URL is a suitable download for a project.
+
+        If it is, register information in the result dictionary (for
+        _get_project) about the specific version it's for.
+
+        Note that the return value isn't actually used other than as a boolean
+        value.
+        """
+        if self.platform_check and self._is_platform_dependent(url):
+            info = None
+        else:
+            info = self.convert_url_to_download_info(url, self.project_name)
+        logger.debug('process_download: %s -> %s', url, info)
+        if info:
+            with self._lock:  # needed because self.result is shared
+                self._update_version_data(self.result, info)
+        return info
+
+    def _should_queue(self, link, referrer, rel):
+        """
+        Determine whether a link URL from a referring page and with a
+        particular "rel" attribute should be queued for scraping.
+        """
+        scheme, netloc, path, _, _, _ = urlparse(link)
+        if path.endswith(self.source_extensions + self.binary_extensions + self.excluded_extensions):
+            result = False
+        elif self.skip_externals and not link.startswith(self.base_url):
+            result = False
+        elif not referrer.startswith(self.base_url):
+            result = False
+        elif rel not in ('homepage', 'download'):
+            result = False
+        elif scheme not in ('http', 'https', 'ftp'):
+            result = False
+        elif self._is_platform_dependent(link):
+            result = False
+        else:
+            host = netloc.split(':', 1)[0]
+            if host.lower() == 'localhost':
+                result = False
+            else:
+                result = True
+        logger.debug('should_queue: %s (%s) from %s -> %s', link, rel, referrer, result)
+        return result
+
+    def _fetch(self):
+        """
+        Get a URL to fetch from the work queue, get the HTML page, examine its
+        links for download candidates and candidates for further scraping.
+
+        This is a handy method to run in a thread.
+        """
+        while True:
+            url = self._to_fetch.get()
+            try:
+                if url:
+                    page = self.get_page(url)
+                    if page is None:  # e.g. after an error
+                        continue
+                    for link, rel in page.links:
+                        if link not in self._seen:
+                            try:
+                                self._seen.add(link)
+                                if (not self._process_download(link) and self._should_queue(link, url, rel)):
+                                    logger.debug('Queueing %s from %s', link, url)
+                                    self._to_fetch.put(link)
+                            except MetadataInvalidError:  # e.g. invalid versions
+                                pass
+            except Exception as e:  # pragma: no cover
+                self.errors.put(text_type(e))
+            finally:
+                # always do this, to avoid hangs :-)
+                self._to_fetch.task_done()
+            if not url:
+                # logger.debug('Sentinel seen, quitting.')
+                break
+
+    def get_page(self, url):
+        """
+        Get the HTML for an URL, possibly from an in-memory cache.
+
+        XXX TODO Note: this cache is never actually cleared. It's assumed that
+        the data won't get stale over the lifetime of a locator instance (not
+        necessarily true for the default_locator).
+        """
+        # http://peak.telecommunity.com/DevCenter/EasyInstall#package-index-api
+        scheme, netloc, path, _, _, _ = urlparse(url)
+        if scheme == 'file' and os.path.isdir(url2pathname(path)):
+            url = urljoin(ensure_slash(url), 'index.html')
+
+        if url in self._page_cache:
+            result = self._page_cache[url]
+            logger.debug('Returning %s from cache: %s', url, result)
+        else:
+            host = netloc.split(':', 1)[0]
+            result = None
+            if host in self._bad_hosts:
+                logger.debug('Skipping %s due to bad host %s', url, host)
+            else:
+                req = Request(url, headers={'Accept-encoding': 'identity'})
+                try:
+                    logger.debug('Fetching %s', url)
+                    resp = self.opener.open(req, timeout=self.timeout)
+                    logger.debug('Fetched %s', url)
+                    headers = resp.info()
+                    content_type = headers.get('Content-Type', '')
+                    if HTML_CONTENT_TYPE.match(content_type):
+                        final_url = resp.geturl()
+                        data = resp.read()
+                        encoding = headers.get('Content-Encoding')
+                        if encoding:
+                            decoder = self.decoders[encoding]  # fail if not found
+                            data = decoder(data)
+                        encoding = 'utf-8'
+                        m = CHARSET.search(content_type)
+                        if m:
+                            encoding = m.group(1)
+                        try:
+                            data = data.decode(encoding)
+                        except UnicodeError:  # pragma: no cover
+                            data = data.decode('latin-1')  # fallback
+                        result = Page(data, final_url)
+                        self._page_cache[final_url] = result
+                except HTTPError as e:
+                    if e.code != 404:
+                        logger.exception('Fetch failed: %s: %s', url, e)
+                except URLError as e:  # pragma: no cover
+                    logger.exception('Fetch failed: %s: %s', url, e)
+                    with self._lock:
+                        self._bad_hosts.add(host)
+                except Exception as e:  # pragma: no cover
+                    logger.exception('Fetch failed: %s: %s', url, e)
+                finally:
+                    self._page_cache[url] = result  # even if None (failure)
+        return result
+
+    _distname_re = re.compile('<a href=[^>]*>([^<]+)<')
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        result = set()
+        page = self.get_page(self.base_url)
+        if not page:
+            raise DistlibException('Unable to get %s' % self.base_url)
+        for match in self._distname_re.finditer(page.data):
+            result.add(match.group(1))
+        return result
+
+
+class DirectoryLocator(Locator):
+    """
+    This class locates distributions in a directory tree.
+    """
+
+    def __init__(self, path, **kwargs):
+        """
+        Initialise an instance.
+        :param path: The root of the directory tree to search.
+        :param kwargs: Passed to the superclass constructor,
+                       except for:
+                       * recursive - if True (the default), subdirectories are
+                         recursed into. If False, only the top-level directory
+                         is searched,
+        """
+        self.recursive = kwargs.pop('recursive', True)
+        super(DirectoryLocator, self).__init__(**kwargs)
+        path = os.path.abspath(path)
+        if not os.path.isdir(path):  # pragma: no cover
+            raise DistlibException('Not a directory: %r' % path)
+        self.base_dir = path
+
+    def should_include(self, filename, parent):
+        """
+        Should a filename be considered as a candidate for a distribution
+        archive? As well as the filename, the directory which contains it
+        is provided, though not used by the current implementation.
+        """
+        return filename.endswith(self.downloadable_extensions)
+
+    def _get_project(self, name):
+        result = {'urls': {}, 'digests': {}}
+        for root, dirs, files in os.walk(self.base_dir):
+            for fn in files:
+                if self.should_include(fn, root):
+                    fn = os.path.join(root, fn)
+                    url = urlunparse(('file', '', pathname2url(os.path.abspath(fn)), '', '', ''))
+                    info = self.convert_url_to_download_info(url, name)
+                    if info:
+                        self._update_version_data(result, info)
+            if not self.recursive:
+                break
+        return result
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        result = set()
+        for root, dirs, files in os.walk(self.base_dir):
+            for fn in files:
+                if self.should_include(fn, root):
+                    fn = os.path.join(root, fn)
+                    url = urlunparse(('file', '', pathname2url(os.path.abspath(fn)), '', '', ''))
+                    info = self.convert_url_to_download_info(url, None)
+                    if info:
+                        result.add(info['name'])
+            if not self.recursive:
+                break
+        return result
+
+
+class JSONLocator(Locator):
+    """
+    This locator uses special extended metadata (not available on PyPI) and is
+    the basis of performant dependency resolution in distlib. Other locators
+    require archive downloads before dependencies can be determined! As you
+    might imagine, that can be slow.
+    """
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        raise NotImplementedError('Not available from this locator')
+
+    def _get_project(self, name):
+        result = {'urls': {}, 'digests': {}}
+        data = get_project_data(name)
+        if data:
+            for info in data.get('files', []):
+                if info['ptype'] != 'sdist' or info['pyversion'] != 'source':
+                    continue
+                # We don't store summary in project metadata as it makes
+                # the data bigger for no benefit during dependency
+                # resolution
+                dist = make_dist(data['name'],
+                                 info['version'],
+                                 summary=data.get('summary', 'Placeholder for summary'),
+                                 scheme=self.scheme)
+                md = dist.metadata
+                md.source_url = info['url']
+                # TODO SHA256 digest
+                if 'digest' in info and info['digest']:
+                    dist.digest = ('md5', info['digest'])
+                md.dependencies = info.get('requirements', {})
+                dist.exports = info.get('exports', {})
+                result[dist.version] = dist
+                result['urls'].setdefault(dist.version, set()).add(info['url'])
+        return result
+
+
+class DistPathLocator(Locator):
+    """
+    This locator finds installed distributions in a path. It can be useful for
+    adding to an :class:`AggregatingLocator`.
+    """
+
+    def __init__(self, distpath, **kwargs):
+        """
+        Initialise an instance.
+
+        :param distpath: A :class:`DistributionPath` instance to search.
+        """
+        super(DistPathLocator, self).__init__(**kwargs)
+        assert isinstance(distpath, DistributionPath)
+        self.distpath = distpath
+
+    def _get_project(self, name):
+        dist = self.distpath.get_distribution(name)
+        if dist is None:
+            result = {'urls': {}, 'digests': {}}
+        else:
+            result = {
+                dist.version: dist,
+                'urls': {
+                    dist.version: set([dist.source_url])
+                },
+                'digests': {
+                    dist.version: set([None])
+                }
+            }
+        return result
+
+
+class AggregatingLocator(Locator):
+    """
+    This class allows you to chain and/or merge a list of locators.
+    """
+
+    def __init__(self, *locators, **kwargs):
+        """
+        Initialise an instance.
+
+        :param locators: The list of locators to search.
+        :param kwargs: Passed to the superclass constructor,
+                       except for:
+                       * merge - if False (the default), the first successful
+                         search from any of the locators is returned. If True,
+                         the results from all locators are merged (this can be
+                         slow).
+        """
+        self.merge = kwargs.pop('merge', False)
+        self.locators = locators
+        super(AggregatingLocator, self).__init__(**kwargs)
+
+    def clear_cache(self):
+        super(AggregatingLocator, self).clear_cache()
+        for locator in self.locators:
+            locator.clear_cache()
+
+    def _set_scheme(self, value):
+        self._scheme = value
+        for locator in self.locators:
+            locator.scheme = value
+
+    scheme = property(Locator.scheme.fget, _set_scheme)
+
+    def _get_project(self, name):
+        result = {}
+        for locator in self.locators:
+            d = locator.get_project(name)
+            if d:
+                if self.merge:
+                    files = result.get('urls', {})
+                    digests = result.get('digests', {})
+                    # next line could overwrite result['urls'], result['digests']
+                    result.update(d)
+                    df = result.get('urls')
+                    if files and df:
+                        for k, v in files.items():
+                            if k in df:
+                                df[k] |= v
+                            else:
+                                df[k] = v
+                    dd = result.get('digests')
+                    if digests and dd:
+                        dd.update(digests)
+                else:
+                    # See issue #18. If any dists are found and we're looking
+                    # for specific constraints, we only return something if
+                    # a match is found. For example, if a DirectoryLocator
+                    # returns just foo (1.0) while we're looking for
+                    # foo (>= 2.0), we'll pretend there was nothing there so
+                    # that subsequent locators can be queried. Otherwise we
+                    # would just return foo (1.0) which would then lead to a
+                    # failure to find foo (>= 2.0), because other locators
+                    # weren't searched. Note that this only matters when
+                    # merge=False.
+                    if self.matcher is None:
+                        found = True
+                    else:
+                        found = False
+                        for k in d:
+                            if self.matcher.match(k):
+                                found = True
+                                break
+                    if found:
+                        result = d
+                        break
+        return result
+
+    def get_distribution_names(self):
+        """
+        Return all the distribution names known to this locator.
+        """
+        result = set()
+        for locator in self.locators:
+            try:
+                result |= locator.get_distribution_names()
+            except NotImplementedError:
+                pass
+        return result
+
+
+# We use a legacy scheme simply because most of the dists on PyPI use legacy
+# versions which don't conform to PEP 440.
+default_locator = AggregatingLocator(
+    # JSONLocator(), # don't use as PEP 426 is withdrawn
+    SimpleScrapingLocator('https://pypi.org/simple/', timeout=3.0),
+    scheme='legacy')
+
+locate = default_locator.locate
+
+
+class DependencyFinder(object):
+    """
+    Locate dependencies for distributions.
+    """
+
+    def __init__(self, locator=None):
+        """
+        Initialise an instance, using the specified locator
+        to locate distributions.
+        """
+        self.locator = locator or default_locator
+        self.scheme = get_scheme(self.locator.scheme)
+
+    def add_distribution(self, dist):
+        """
+        Add a distribution to the finder. This will update internal information
+        about who provides what.
+        :param dist: The distribution to add.
+        """
+        logger.debug('adding distribution %s', dist)
+        name = dist.key
+        self.dists_by_name[name] = dist
+        self.dists[(name, dist.version)] = dist
+        for p in dist.provides:
+            name, version = parse_name_and_version(p)
+            logger.debug('Add to provided: %s, %s, %s', name, version, dist)
+            self.provided.setdefault(name, set()).add((version, dist))
+
+    def remove_distribution(self, dist):
+        """
+        Remove a distribution from the finder. This will update internal
+        information about who provides what.
+        :param dist: The distribution to remove.
+        """
+        logger.debug('removing distribution %s', dist)
+        name = dist.key
+        del self.dists_by_name[name]
+        del self.dists[(name, dist.version)]
+        for p in dist.provides:
+            name, version = parse_name_and_version(p)
+            logger.debug('Remove from provided: %s, %s, %s', name, version, dist)
+            s = self.provided[name]
+            s.remove((version, dist))
+            if not s:
+                del self.provided[name]
+
+    def get_matcher(self, reqt):
+        """
+        Get a version matcher for a requirement.
+        :param reqt: The requirement
+        :type reqt: str
+        :return: A version matcher (an instance of
+                 :class:`distlib.version.Matcher`).
+        """
+        try:
+            matcher = self.scheme.matcher(reqt)
+        except UnsupportedVersionError:  # pragma: no cover
+            # XXX compat-mode if cannot read the version
+            name = reqt.split()[0]
+            matcher = self.scheme.matcher(name)
+        return matcher
+
+    def find_providers(self, reqt):
+        """
+        Find the distributions which can fulfill a requirement.
+
+        :param reqt: The requirement.
+         :type reqt: str
+        :return: A set of distribution which can fulfill the requirement.
+        """
+        matcher = self.get_matcher(reqt)
+        name = matcher.key  # case-insensitive
+        result = set()
+        provided = self.provided
+        if name in provided:
+            for version, provider in provided[name]:
+                try:
+                    match = matcher.match(version)
+                except UnsupportedVersionError:
+                    match = False
+
+                if match:
+                    result.add(provider)
+                    break
+        return result
+
+    def try_to_replace(self, provider, other, problems):
+        """
+        Attempt to replace one provider with another. This is typically used
+        when resolving dependencies from multiple sources, e.g. A requires
+        (B >= 1.0) while C requires (B >= 1.1).
+
+        For successful replacement, ``provider`` must meet all the requirements
+        which ``other`` fulfills.
+
+        :param provider: The provider we are trying to replace with.
+        :param other: The provider we're trying to replace.
+        :param problems: If False is returned, this will contain what
+                         problems prevented replacement. This is currently
+                         a tuple of the literal string 'cantreplace',
+                         ``provider``, ``other``  and the set of requirements
+                         that ``provider`` couldn't fulfill.
+        :return: True if we can replace ``other`` with ``provider``, else
+                 False.
+        """
+        rlist = self.reqts[other]
+        unmatched = set()
+        for s in rlist:
+            matcher = self.get_matcher(s)
+            if not matcher.match(provider.version):
+                unmatched.add(s)
+        if unmatched:
+            # can't replace other with provider
+            problems.add(('cantreplace', provider, other, frozenset(unmatched)))
+            result = False
+        else:
+            # can replace other with provider
+            self.remove_distribution(other)
+            del self.reqts[other]
+            for s in rlist:
+                self.reqts.setdefault(provider, set()).add(s)
+            self.add_distribution(provider)
+            result = True
+        return result
+
+    def find(self, requirement, meta_extras=None, prereleases=False):
+        """
+        Find a distribution and all distributions it depends on.
+
+        :param requirement: The requirement specifying the distribution to
+                            find, or a Distribution instance.
+        :param meta_extras: A list of meta extras such as :test:, :build: and
+                            so on.
+        :param prereleases: If ``True``, allow pre-release versions to be
+                            returned - otherwise, don't return prereleases
+                            unless they're all that's available.
+
+        Return a set of :class:`Distribution` instances and a set of
+        problems.
+
+        The distributions returned should be such that they have the
+        :attr:`required` attribute set to ``True`` if they were
+        from the ``requirement`` passed to ``find()``, and they have the
+        :attr:`build_time_dependency` attribute set to ``True`` unless they
+        are post-installation dependencies of the ``requirement``.
+
+        The problems should be a tuple consisting of the string
+        ``'unsatisfied'`` and the requirement which couldn't be satisfied
+        by any distribution known to the locator.
+        """
+
+        self.provided = {}
+        self.dists = {}
+        self.dists_by_name = {}
+        self.reqts = {}
+
+        meta_extras = set(meta_extras or [])
+        if ':*:' in meta_extras:
+            meta_extras.remove(':*:')
+            # :meta: and :run: are implicitly included
+            meta_extras |= set([':test:', ':build:', ':dev:'])
+
+        if isinstance(requirement, Distribution):
+            dist = odist = requirement
+            logger.debug('passed %s as requirement', odist)
+        else:
+            dist = odist = self.locator.locate(requirement, prereleases=prereleases)
+            if dist is None:
+                raise DistlibException('Unable to locate %r' % requirement)
+            logger.debug('located %s', odist)
+        dist.requested = True
+        problems = set()
+        todo = set([dist])
+        install_dists = set([odist])
+        while todo:
+            dist = todo.pop()
+            name = dist.key  # case-insensitive
+            if name not in self.dists_by_name:
+                self.add_distribution(dist)
+            else:
+                # import pdb; pdb.set_trace()
+                other = self.dists_by_name[name]
+                if other != dist:
+                    self.try_to_replace(dist, other, problems)
+
+            ireqts = dist.run_requires | dist.meta_requires
+            sreqts = dist.build_requires
+            ereqts = set()
+            if meta_extras and dist in install_dists:
+                for key in ('test', 'build', 'dev'):
+                    e = ':%s:' % key
+                    if e in meta_extras:
+                        ereqts |= getattr(dist, '%s_requires' % key)
+            all_reqts = ireqts | sreqts | ereqts
+            for r in all_reqts:
+                providers = self.find_providers(r)
+                if not providers:
+                    logger.debug('No providers found for %r', r)
+                    provider = self.locator.locate(r, prereleases=prereleases)
+                    # If no provider is found and we didn't consider
+                    # prereleases, consider them now.
+                    if provider is None and not prereleases:
+                        provider = self.locator.locate(r, prereleases=True)
+                    if provider is None:
+                        logger.debug('Cannot satisfy %r', r)
+                        problems.add(('unsatisfied', r))
+                    else:
+                        n, v = provider.key, provider.version
+                        if (n, v) not in self.dists:
+                            todo.add(provider)
+                        providers.add(provider)
+                        if r in ireqts and dist in install_dists:
+                            install_dists.add(provider)
+                            logger.debug('Adding %s to install_dists', provider.name_and_version)
+                for p in providers:
+                    name = p.key
+                    if name not in self.dists_by_name:
+                        self.reqts.setdefault(p, set()).add(r)
+                    else:
+                        other = self.dists_by_name[name]
+                        if other != p:
+                            # see if other can be replaced by p
+                            self.try_to_replace(p, other, problems)
+
+        dists = set(self.dists.values())
+        for dist in dists:
+            dist.build_time_dependency = dist not in install_dists
+            if dist.build_time_dependency:
+                logger.debug('%s is a build-time dependency only.', dist.name_and_version)
+        logger.debug('find done for %s', odist)
+        return dists, problems
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/manifest.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/manifest.py
new file mode 100644
index 00000000..420dcf12
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/manifest.py
@@ -0,0 +1,384 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 Python Software Foundation.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+"""
+Class representing the list of files in a distribution.
+
+Equivalent to distutils.filelist, but fixes some problems.
+"""
+import fnmatch
+import logging
+import os
+import re
+import sys
+
+from . import DistlibException
+from .compat import fsdecode
+from .util import convert_path
+
+
+__all__ = ['Manifest']
+
+logger = logging.getLogger(__name__)
+
+# a \ followed by some spaces + EOL
+_COLLAPSE_PATTERN = re.compile('\\\\w*\n', re.M)
+_COMMENTED_LINE = re.compile('#.*?(?=\n)|\n(?=$)', re.M | re.S)
+
+#
+# Due to the different results returned by fnmatch.translate, we need
+# to do slightly different processing for Python 2.7 and 3.2 ... this needed
+# to be brought in for Python 3.6 onwards.
+#
+_PYTHON_VERSION = sys.version_info[:2]
+
+
+class Manifest(object):
+    """
+    A list of files built by exploring the filesystem and filtered by applying various
+    patterns to what we find there.
+    """
+
+    def __init__(self, base=None):
+        """
+        Initialise an instance.
+
+        :param base: The base directory to explore under.
+        """
+        self.base = os.path.abspath(os.path.normpath(base or os.getcwd()))
+        self.prefix = self.base + os.sep
+        self.allfiles = None
+        self.files = set()
+
+    #
+    # Public API
+    #
+
+    def findall(self):
+        """Find all files under the base and set ``allfiles`` to the absolute
+        pathnames of files found.
+        """
+        from stat import S_ISREG, S_ISDIR, S_ISLNK
+
+        self.allfiles = allfiles = []
+        root = self.base
+        stack = [root]
+        pop = stack.pop
+        push = stack.append
+
+        while stack:
+            root = pop()
+            names = os.listdir(root)
+
+            for name in names:
+                fullname = os.path.join(root, name)
+
+                # Avoid excess stat calls -- just one will do, thank you!
+                stat = os.stat(fullname)
+                mode = stat.st_mode
+                if S_ISREG(mode):
+                    allfiles.append(fsdecode(fullname))
+                elif S_ISDIR(mode) and not S_ISLNK(mode):
+                    push(fullname)
+
+    def add(self, item):
+        """
+        Add a file to the manifest.
+
+        :param item: The pathname to add. This can be relative to the base.
+        """
+        if not item.startswith(self.prefix):
+            item = os.path.join(self.base, item)
+        self.files.add(os.path.normpath(item))
+
+    def add_many(self, items):
+        """
+        Add a list of files to the manifest.
+
+        :param items: The pathnames to add. These can be relative to the base.
+        """
+        for item in items:
+            self.add(item)
+
+    def sorted(self, wantdirs=False):
+        """
+        Return sorted files in directory order
+        """
+
+        def add_dir(dirs, d):
+            dirs.add(d)
+            logger.debug('add_dir added %s', d)
+            if d != self.base:
+                parent, _ = os.path.split(d)
+                assert parent not in ('', '/')
+                add_dir(dirs, parent)
+
+        result = set(self.files)    # make a copy!
+        if wantdirs:
+            dirs = set()
+            for f in result:
+                add_dir(dirs, os.path.dirname(f))
+            result |= dirs
+        return [os.path.join(*path_tuple) for path_tuple in
+                sorted(os.path.split(path) for path in result)]
+
+    def clear(self):
+        """Clear all collected files."""
+        self.files = set()
+        self.allfiles = []
+
+    def process_directive(self, directive):
+        """
+        Process a directive which either adds some files from ``allfiles`` to
+        ``files``, or removes some files from ``files``.
+
+        :param directive: The directive to process. This should be in a format
+                     compatible with distutils ``MANIFEST.in`` files:
+
+                     http://docs.python.org/distutils/sourcedist.html#commands
+        """
+        # Parse the line: split it up, make sure the right number of words
+        # is there, and return the relevant words.  'action' is always
+        # defined: it's the first word of the line.  Which of the other
+        # three are defined depends on the action; it'll be either
+        # patterns, (dir and patterns), or (dirpattern).
+        action, patterns, thedir, dirpattern = self._parse_directive(directive)
+
+        # OK, now we know that the action is valid and we have the
+        # right number of words on the line for that action -- so we
+        # can proceed with minimal error-checking.
+        if action == 'include':
+            for pattern in patterns:
+                if not self._include_pattern(pattern, anchor=True):
+                    logger.warning('no files found matching %r', pattern)
+
+        elif action == 'exclude':
+            for pattern in patterns:
+                self._exclude_pattern(pattern, anchor=True)
+
+        elif action == 'global-include':
+            for pattern in patterns:
+                if not self._include_pattern(pattern, anchor=False):
+                    logger.warning('no files found matching %r '
+                                   'anywhere in distribution', pattern)
+
+        elif action == 'global-exclude':
+            for pattern in patterns:
+                self._exclude_pattern(pattern, anchor=False)
+
+        elif action == 'recursive-include':
+            for pattern in patterns:
+                if not self._include_pattern(pattern, prefix=thedir):
+                    logger.warning('no files found matching %r '
+                                   'under directory %r', pattern, thedir)
+
+        elif action == 'recursive-exclude':
+            for pattern in patterns:
+                self._exclude_pattern(pattern, prefix=thedir)
+
+        elif action == 'graft':
+            if not self._include_pattern(None, prefix=dirpattern):
+                logger.warning('no directories found matching %r',
+                               dirpattern)
+
+        elif action == 'prune':
+            if not self._exclude_pattern(None, prefix=dirpattern):
+                logger.warning('no previously-included directories found '
+                               'matching %r', dirpattern)
+        else:   # pragma: no cover
+            # This should never happen, as it should be caught in
+            # _parse_template_line
+            raise DistlibException(
+                'invalid action %r' % action)
+
+    #
+    # Private API
+    #
+
+    def _parse_directive(self, directive):
+        """
+        Validate a directive.
+        :param directive: The directive to validate.
+        :return: A tuple of action, patterns, thedir, dir_patterns
+        """
+        words = directive.split()
+        if len(words) == 1 and words[0] not in ('include', 'exclude',
+                                                'global-include',
+                                                'global-exclude',
+                                                'recursive-include',
+                                                'recursive-exclude',
+                                                'graft', 'prune'):
+            # no action given, let's use the default 'include'
+            words.insert(0, 'include')
+
+        action = words[0]
+        patterns = thedir = dir_pattern = None
+
+        if action in ('include', 'exclude',
+                      'global-include', 'global-exclude'):
+            if len(words) < 2:
+                raise DistlibException(
+                    '%r expects <pattern1> <pattern2> ...' % action)
+
+            patterns = [convert_path(word) for word in words[1:]]
+
+        elif action in ('recursive-include', 'recursive-exclude'):
+            if len(words) < 3:
+                raise DistlibException(
+                    '%r expects <dir> <pattern1> <pattern2> ...' % action)
+
+            thedir = convert_path(words[1])
+            patterns = [convert_path(word) for word in words[2:]]
+
+        elif action in ('graft', 'prune'):
+            if len(words) != 2:
+                raise DistlibException(
+                    '%r expects a single <dir_pattern>' % action)
+
+            dir_pattern = convert_path(words[1])
+
+        else:
+            raise DistlibException('unknown action %r' % action)
+
+        return action, patterns, thedir, dir_pattern
+
+    def _include_pattern(self, pattern, anchor=True, prefix=None,
+                         is_regex=False):
+        """Select strings (presumably filenames) from 'self.files' that
+        match 'pattern', a Unix-style wildcard (glob) pattern.
+
+        Patterns are not quite the same as implemented by the 'fnmatch'
+        module: '*' and '?'  match non-special characters, where "special"
+        is platform-dependent: slash on Unix; colon, slash, and backslash on
+        DOS/Windows; and colon on Mac OS.
+
+        If 'anchor' is true (the default), then the pattern match is more
+        stringent: "*.py" will match "foo.py" but not "foo/bar.py".  If
+        'anchor' is false, both of these will match.
+
+        If 'prefix' is supplied, then only filenames starting with 'prefix'
+        (itself a pattern) and ending with 'pattern', with anything in between
+        them, will match.  'anchor' is ignored in this case.
+
+        If 'is_regex' is true, 'anchor' and 'prefix' are ignored, and
+        'pattern' is assumed to be either a string containing a regex or a
+        regex object -- no translation is done, the regex is just compiled
+        and used as-is.
+
+        Selected strings will be added to self.files.
+
+        Return True if files are found.
+        """
+        # XXX docstring lying about what the special chars are?
+        found = False
+        pattern_re = self._translate_pattern(pattern, anchor, prefix, is_regex)
+
+        # delayed loading of allfiles list
+        if self.allfiles is None:
+            self.findall()
+
+        for name in self.allfiles:
+            if pattern_re.search(name):
+                self.files.add(name)
+                found = True
+        return found
+
+    def _exclude_pattern(self, pattern, anchor=True, prefix=None,
+                         is_regex=False):
+        """Remove strings (presumably filenames) from 'files' that match
+        'pattern'.
+
+        Other parameters are the same as for 'include_pattern()', above.
+        The list 'self.files' is modified in place. Return True if files are
+        found.
+
+        This API is public to allow e.g. exclusion of SCM subdirs, e.g. when
+        packaging source distributions
+        """
+        found = False
+        pattern_re = self._translate_pattern(pattern, anchor, prefix, is_regex)
+        for f in list(self.files):
+            if pattern_re.search(f):
+                self.files.remove(f)
+                found = True
+        return found
+
+    def _translate_pattern(self, pattern, anchor=True, prefix=None,
+                           is_regex=False):
+        """Translate a shell-like wildcard pattern to a compiled regular
+        expression.
+
+        Return the compiled regex.  If 'is_regex' true,
+        then 'pattern' is directly compiled to a regex (if it's a string)
+        or just returned as-is (assumes it's a regex object).
+        """
+        if is_regex:
+            if isinstance(pattern, str):
+                return re.compile(pattern)
+            else:
+                return pattern
+
+        if _PYTHON_VERSION > (3, 2):
+            # ditch start and end characters
+            start, _, end = self._glob_to_re('_').partition('_')
+
+        if pattern:
+            pattern_re = self._glob_to_re(pattern)
+            if _PYTHON_VERSION > (3, 2):
+                assert pattern_re.startswith(start) and pattern_re.endswith(end)
+        else:
+            pattern_re = ''
+
+        base = re.escape(os.path.join(self.base, ''))
+        if prefix is not None:
+            # ditch end of pattern character
+            if _PYTHON_VERSION <= (3, 2):
+                empty_pattern = self._glob_to_re('')
+                prefix_re = self._glob_to_re(prefix)[:-len(empty_pattern)]
+            else:
+                prefix_re = self._glob_to_re(prefix)
+                assert prefix_re.startswith(start) and prefix_re.endswith(end)
+                prefix_re = prefix_re[len(start): len(prefix_re) - len(end)]
+            sep = os.sep
+            if os.sep == '\\':
+                sep = r'\\'
+            if _PYTHON_VERSION <= (3, 2):
+                pattern_re = '^' + base + sep.join((prefix_re,
+                                                    '.*' + pattern_re))
+            else:
+                pattern_re = pattern_re[len(start): len(pattern_re) - len(end)]
+                pattern_re = r'%s%s%s%s.*%s%s' % (start, base, prefix_re, sep,
+                                                  pattern_re, end)
+        else:  # no prefix -- respect anchor flag
+            if anchor:
+                if _PYTHON_VERSION <= (3, 2):
+                    pattern_re = '^' + base + pattern_re
+                else:
+                    pattern_re = r'%s%s%s' % (start, base, pattern_re[len(start):])
+
+        return re.compile(pattern_re)
+
+    def _glob_to_re(self, pattern):
+        """Translate a shell-like glob pattern to a regular expression.
+
+        Return a string containing the regex.  Differs from
+        'fnmatch.translate()' in that '*' does not match "special characters"
+        (which are platform-specific).
+        """
+        pattern_re = fnmatch.translate(pattern)
+
+        # '?' and '*' in the glob pattern become '.' and '.*' in the RE, which
+        # IMHO is wrong -- '?' and '*' aren't supposed to match slash in Unix,
+        # and by extension they shouldn't match such "special characters" under
+        # any OS.  So change all non-escaped dots in the RE to match any
+        # character except the special characters (currently: just os.sep).
+        sep = os.sep
+        if os.sep == '\\':
+            # we're using a regex to manipulate a regex, so we need
+            # to escape the backslash twice
+            sep = r'\\\\'
+        escaped = r'\1[^%s]' % sep
+        pattern_re = re.sub(r'((?<!\\)(\\\\)*)\.', escaped, pattern_re)
+        return pattern_re
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/markers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/markers.py
new file mode 100644
index 00000000..3f5632be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/markers.py
@@ -0,0 +1,162 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+"""
+Parser for the environment markers micro-language defined in PEP 508.
+"""
+
+# Note: In PEP 345, the micro-language was Python compatible, so the ast
+# module could be used to parse it. However, PEP 508 introduced operators such
+# as ~= and === which aren't in Python, necessitating a different approach.
+
+import os
+import re
+import sys
+import platform
+
+from .compat import string_types
+from .util import in_venv, parse_marker
+from .version import LegacyVersion as LV
+
+__all__ = ['interpret']
+
+_VERSION_PATTERN = re.compile(r'((\d+(\.\d+)*\w*)|\'(\d+(\.\d+)*\w*)\'|\"(\d+(\.\d+)*\w*)\")')
+_VERSION_MARKERS = {'python_version', 'python_full_version'}
+
+
+def _is_version_marker(s):
+    return isinstance(s, string_types) and s in _VERSION_MARKERS
+
+
+def _is_literal(o):
+    if not isinstance(o, string_types) or not o:
+        return False
+    return o[0] in '\'"'
+
+
+def _get_versions(s):
+    return {LV(m.groups()[0]) for m in _VERSION_PATTERN.finditer(s)}
+
+
+class Evaluator(object):
+    """
+    This class is used to evaluate marker expressions.
+    """
+
+    operations = {
+        '==': lambda x, y: x == y,
+        '===': lambda x, y: x == y,
+        '~=': lambda x, y: x == y or x > y,
+        '!=': lambda x, y: x != y,
+        '<': lambda x, y: x < y,
+        '<=': lambda x, y: x == y or x < y,
+        '>': lambda x, y: x > y,
+        '>=': lambda x, y: x == y or x > y,
+        'and': lambda x, y: x and y,
+        'or': lambda x, y: x or y,
+        'in': lambda x, y: x in y,
+        'not in': lambda x, y: x not in y,
+    }
+
+    def evaluate(self, expr, context):
+        """
+        Evaluate a marker expression returned by the :func:`parse_requirement`
+        function in the specified context.
+        """
+        if isinstance(expr, string_types):
+            if expr[0] in '\'"':
+                result = expr[1:-1]
+            else:
+                if expr not in context:
+                    raise SyntaxError('unknown variable: %s' % expr)
+                result = context[expr]
+        else:
+            assert isinstance(expr, dict)
+            op = expr['op']
+            if op not in self.operations:
+                raise NotImplementedError('op not implemented: %s' % op)
+            elhs = expr['lhs']
+            erhs = expr['rhs']
+            if _is_literal(expr['lhs']) and _is_literal(expr['rhs']):
+                raise SyntaxError('invalid comparison: %s %s %s' % (elhs, op, erhs))
+
+            lhs = self.evaluate(elhs, context)
+            rhs = self.evaluate(erhs, context)
+            if ((_is_version_marker(elhs) or _is_version_marker(erhs)) and
+                    op in ('<', '<=', '>', '>=', '===', '==', '!=', '~=')):
+                lhs = LV(lhs)
+                rhs = LV(rhs)
+            elif _is_version_marker(elhs) and op in ('in', 'not in'):
+                lhs = LV(lhs)
+                rhs = _get_versions(rhs)
+            result = self.operations[op](lhs, rhs)
+        return result
+
+
+_DIGITS = re.compile(r'\d+\.\d+')
+
+
+def default_context():
+
+    def format_full_version(info):
+        version = '%s.%s.%s' % (info.major, info.minor, info.micro)
+        kind = info.releaselevel
+        if kind != 'final':
+            version += kind[0] + str(info.serial)
+        return version
+
+    if hasattr(sys, 'implementation'):
+        implementation_version = format_full_version(sys.implementation.version)
+        implementation_name = sys.implementation.name
+    else:
+        implementation_version = '0'
+        implementation_name = ''
+
+    ppv = platform.python_version()
+    m = _DIGITS.match(ppv)
+    pv = m.group(0)
+    result = {
+        'implementation_name': implementation_name,
+        'implementation_version': implementation_version,
+        'os_name': os.name,
+        'platform_machine': platform.machine(),
+        'platform_python_implementation': platform.python_implementation(),
+        'platform_release': platform.release(),
+        'platform_system': platform.system(),
+        'platform_version': platform.version(),
+        'platform_in_venv': str(in_venv()),
+        'python_full_version': ppv,
+        'python_version': pv,
+        'sys_platform': sys.platform,
+    }
+    return result
+
+
+DEFAULT_CONTEXT = default_context()
+del default_context
+
+evaluator = Evaluator()
+
+
+def interpret(marker, execution_context=None):
+    """
+    Interpret a marker and return a result depending on environment.
+
+    :param marker: The marker to interpret.
+    :type marker: str
+    :param execution_context: The context used for name lookup.
+    :type execution_context: mapping
+    """
+    try:
+        expr, rest = parse_marker(marker)
+    except Exception as e:
+        raise SyntaxError('Unable to interpret marker syntax: %s: %s' % (marker, e))
+    if rest and rest[0] != '#':
+        raise SyntaxError('unexpected trailing data in marker: %s: %s' % (marker, rest))
+    context = dict(DEFAULT_CONTEXT)
+    if execution_context:
+        context.update(execution_context)
+    return evaluator.evaluate(expr, context)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/metadata.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/metadata.py
new file mode 100644
index 00000000..ce9a34b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/metadata.py
@@ -0,0 +1,1031 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012 The Python Software Foundation.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+"""Implementation of the Metadata for Python packages PEPs.
+
+Supports all metadata formats (1.0, 1.1, 1.2, 1.3/2.1 and 2.2).
+"""
+from __future__ import unicode_literals
+
+import codecs
+from email import message_from_file
+import json
+import logging
+import re
+
+from . import DistlibException, __version__
+from .compat import StringIO, string_types, text_type
+from .markers import interpret
+from .util import extract_by_key, get_extras
+from .version import get_scheme, PEP440_VERSION_RE
+
+logger = logging.getLogger(__name__)
+
+
+class MetadataMissingError(DistlibException):
+    """A required metadata is missing"""
+
+
+class MetadataConflictError(DistlibException):
+    """Attempt to read or write metadata fields that are conflictual."""
+
+
+class MetadataUnrecognizedVersionError(DistlibException):
+    """Unknown metadata version number."""
+
+
+class MetadataInvalidError(DistlibException):
+    """A metadata value is invalid"""
+
+
+# public API of this module
+__all__ = ['Metadata', 'PKG_INFO_ENCODING', 'PKG_INFO_PREFERRED_VERSION']
+
+# Encoding used for the PKG-INFO files
+PKG_INFO_ENCODING = 'utf-8'
+
+# preferred version. Hopefully will be changed
+# to 1.2 once PEP 345 is supported everywhere
+PKG_INFO_PREFERRED_VERSION = '1.1'
+
+_LINE_PREFIX_1_2 = re.compile('\n       \\|')
+_LINE_PREFIX_PRE_1_2 = re.compile('\n        ')
+_241_FIELDS = ('Metadata-Version', 'Name', 'Version', 'Platform', 'Summary', 'Description', 'Keywords', 'Home-page',
+               'Author', 'Author-email', 'License')
+
+_314_FIELDS = ('Metadata-Version', 'Name', 'Version', 'Platform', 'Supported-Platform', 'Summary', 'Description',
+               'Keywords', 'Home-page', 'Author', 'Author-email', 'License', 'Classifier', 'Download-URL', 'Obsoletes',
+               'Provides', 'Requires')
+
+_314_MARKERS = ('Obsoletes', 'Provides', 'Requires', 'Classifier', 'Download-URL')
+
+_345_FIELDS = ('Metadata-Version', 'Name', 'Version', 'Platform', 'Supported-Platform', 'Summary', 'Description',
+               'Keywords', 'Home-page', 'Author', 'Author-email', 'Maintainer', 'Maintainer-email', 'License',
+               'Classifier', 'Download-URL', 'Obsoletes-Dist', 'Project-URL', 'Provides-Dist', 'Requires-Dist',
+               'Requires-Python', 'Requires-External')
+
+_345_MARKERS = ('Provides-Dist', 'Requires-Dist', 'Requires-Python', 'Obsoletes-Dist', 'Requires-External',
+                'Maintainer', 'Maintainer-email', 'Project-URL')
+
+_426_FIELDS = ('Metadata-Version', 'Name', 'Version', 'Platform', 'Supported-Platform', 'Summary', 'Description',
+               'Keywords', 'Home-page', 'Author', 'Author-email', 'Maintainer', 'Maintainer-email', 'License',
+               'Classifier', 'Download-URL', 'Obsoletes-Dist', 'Project-URL', 'Provides-Dist', 'Requires-Dist',
+               'Requires-Python', 'Requires-External', 'Private-Version', 'Obsoleted-By', 'Setup-Requires-Dist',
+               'Extension', 'Provides-Extra')
+
+_426_MARKERS = ('Private-Version', 'Provides-Extra', 'Obsoleted-By', 'Setup-Requires-Dist', 'Extension')
+
+# See issue #106: Sometimes 'Requires' and 'Provides' occur wrongly in
+# the metadata. Include them in the tuple literal below to allow them
+# (for now).
+# Ditto for Obsoletes - see issue #140.
+_566_FIELDS = _426_FIELDS + ('Description-Content-Type', 'Requires', 'Provides', 'Obsoletes')
+
+_566_MARKERS = ('Description-Content-Type', )
+
+_643_MARKERS = ('Dynamic', 'License-File')
+
+_643_FIELDS = _566_FIELDS + _643_MARKERS
+
+_ALL_FIELDS = set()
+_ALL_FIELDS.update(_241_FIELDS)
+_ALL_FIELDS.update(_314_FIELDS)
+_ALL_FIELDS.update(_345_FIELDS)
+_ALL_FIELDS.update(_426_FIELDS)
+_ALL_FIELDS.update(_566_FIELDS)
+_ALL_FIELDS.update(_643_FIELDS)
+
+EXTRA_RE = re.compile(r'''extra\s*==\s*("([^"]+)"|'([^']+)')''')
+
+
+def _version2fieldlist(version):
+    if version == '1.0':
+        return _241_FIELDS
+    elif version == '1.1':
+        return _314_FIELDS
+    elif version == '1.2':
+        return _345_FIELDS
+    elif version in ('1.3', '2.1'):
+        # avoid adding field names if already there
+        return _345_FIELDS + tuple(f for f in _566_FIELDS if f not in _345_FIELDS)
+    elif version == '2.0':
+        raise ValueError('Metadata 2.0 is withdrawn and not supported')
+        # return _426_FIELDS
+    elif version == '2.2':
+        return _643_FIELDS
+    raise MetadataUnrecognizedVersionError(version)
+
+
+def _best_version(fields):
+    """Detect the best version depending on the fields used."""
+
+    def _has_marker(keys, markers):
+        return any(marker in keys for marker in markers)
+
+    keys = [key for key, value in fields.items() if value not in ([], 'UNKNOWN', None)]
+    possible_versions = ['1.0', '1.1', '1.2', '1.3', '2.1', '2.2']  # 2.0 removed
+
+    # first let's try to see if a field is not part of one of the version
+    for key in keys:
+        if key not in _241_FIELDS and '1.0' in possible_versions:
+            possible_versions.remove('1.0')
+            logger.debug('Removed 1.0 due to %s', key)
+        if key not in _314_FIELDS and '1.1' in possible_versions:
+            possible_versions.remove('1.1')
+            logger.debug('Removed 1.1 due to %s', key)
+        if key not in _345_FIELDS and '1.2' in possible_versions:
+            possible_versions.remove('1.2')
+            logger.debug('Removed 1.2 due to %s', key)
+        if key not in _566_FIELDS and '1.3' in possible_versions:
+            possible_versions.remove('1.3')
+            logger.debug('Removed 1.3 due to %s', key)
+        if key not in _566_FIELDS and '2.1' in possible_versions:
+            if key != 'Description':  # In 2.1, description allowed after headers
+                possible_versions.remove('2.1')
+                logger.debug('Removed 2.1 due to %s', key)
+        if key not in _643_FIELDS and '2.2' in possible_versions:
+            possible_versions.remove('2.2')
+            logger.debug('Removed 2.2 due to %s', key)
+        # if key not in _426_FIELDS and '2.0' in possible_versions:
+        # possible_versions.remove('2.0')
+        # logger.debug('Removed 2.0 due to %s', key)
+
+    # possible_version contains qualified versions
+    if len(possible_versions) == 1:
+        return possible_versions[0]  # found !
+    elif len(possible_versions) == 0:
+        logger.debug('Out of options - unknown metadata set: %s', fields)
+        raise MetadataConflictError('Unknown metadata set')
+
+    # let's see if one unique marker is found
+    is_1_1 = '1.1' in possible_versions and _has_marker(keys, _314_MARKERS)
+    is_1_2 = '1.2' in possible_versions and _has_marker(keys, _345_MARKERS)
+    is_2_1 = '2.1' in possible_versions and _has_marker(keys, _566_MARKERS)
+    # is_2_0 = '2.0' in possible_versions and _has_marker(keys, _426_MARKERS)
+    is_2_2 = '2.2' in possible_versions and _has_marker(keys, _643_MARKERS)
+    if int(is_1_1) + int(is_1_2) + int(is_2_1) + int(is_2_2) > 1:
+        raise MetadataConflictError('You used incompatible 1.1/1.2/2.1/2.2 fields')
+
+    # we have the choice, 1.0, or 1.2, 2.1 or 2.2
+    #   - 1.0 has a broken Summary field but works with all tools
+    #   - 1.1 is to avoid
+    #   - 1.2 fixes Summary but has little adoption
+    #   - 2.1 adds more features
+    #   - 2.2 is the latest
+    if not is_1_1 and not is_1_2 and not is_2_1 and not is_2_2:
+        # we couldn't find any specific marker
+        if PKG_INFO_PREFERRED_VERSION in possible_versions:
+            return PKG_INFO_PREFERRED_VERSION
+    if is_1_1:
+        return '1.1'
+    if is_1_2:
+        return '1.2'
+    if is_2_1:
+        return '2.1'
+    # if is_2_2:
+    # return '2.2'
+
+    return '2.2'
+
+
+# This follows the rules about transforming keys as described in
+# https://www.python.org/dev/peps/pep-0566/#id17
+_ATTR2FIELD = {name.lower().replace("-", "_"): name for name in _ALL_FIELDS}
+_FIELD2ATTR = {field: attr for attr, field in _ATTR2FIELD.items()}
+
+_PREDICATE_FIELDS = ('Requires-Dist', 'Obsoletes-Dist', 'Provides-Dist')
+_VERSIONS_FIELDS = ('Requires-Python', )
+_VERSION_FIELDS = ('Version', )
+_LISTFIELDS = ('Platform', 'Classifier', 'Obsoletes', 'Requires', 'Provides', 'Obsoletes-Dist', 'Provides-Dist',
+               'Requires-Dist', 'Requires-External', 'Project-URL', 'Supported-Platform', 'Setup-Requires-Dist',
+               'Provides-Extra', 'Extension', 'License-File')
+_LISTTUPLEFIELDS = ('Project-URL', )
+
+_ELEMENTSFIELD = ('Keywords', )
+
+_UNICODEFIELDS = ('Author', 'Maintainer', 'Summary', 'Description')
+
+_MISSING = object()
+
+_FILESAFE = re.compile('[^A-Za-z0-9.]+')
+
+
+def _get_name_and_version(name, version, for_filename=False):
+    """Return the distribution name with version.
+
+    If for_filename is true, return a filename-escaped form."""
+    if for_filename:
+        # For both name and version any runs of non-alphanumeric or '.'
+        # characters are replaced with a single '-'.  Additionally any
+        # spaces in the version string become '.'
+        name = _FILESAFE.sub('-', name)
+        version = _FILESAFE.sub('-', version.replace(' ', '.'))
+    return '%s-%s' % (name, version)
+
+
+class LegacyMetadata(object):
+    """The legacy metadata of a release.
+
+    Supports versions 1.0, 1.1, 1.2, 2.0 and 1.3/2.1 (auto-detected). You can
+    instantiate the class with one of these arguments (or none):
+    - *path*, the path to a metadata file
+    - *fileobj* give a file-like object with metadata as content
+    - *mapping* is a dict-like object
+    - *scheme* is a version scheme name
+    """
+
+    # TODO document the mapping API and UNKNOWN default key
+
+    def __init__(self, path=None, fileobj=None, mapping=None, scheme='default'):
+        if [path, fileobj, mapping].count(None) < 2:
+            raise TypeError('path, fileobj and mapping are exclusive')
+        self._fields = {}
+        self.requires_files = []
+        self._dependencies = None
+        self.scheme = scheme
+        if path is not None:
+            self.read(path)
+        elif fileobj is not None:
+            self.read_file(fileobj)
+        elif mapping is not None:
+            self.update(mapping)
+            self.set_metadata_version()
+
+    def set_metadata_version(self):
+        self._fields['Metadata-Version'] = _best_version(self._fields)
+
+    def _write_field(self, fileobj, name, value):
+        fileobj.write('%s: %s\n' % (name, value))
+
+    def __getitem__(self, name):
+        return self.get(name)
+
+    def __setitem__(self, name, value):
+        return self.set(name, value)
+
+    def __delitem__(self, name):
+        field_name = self._convert_name(name)
+        try:
+            del self._fields[field_name]
+        except KeyError:
+            raise KeyError(name)
+
+    def __contains__(self, name):
+        return (name in self._fields or self._convert_name(name) in self._fields)
+
+    def _convert_name(self, name):
+        if name in _ALL_FIELDS:
+            return name
+        name = name.replace('-', '_').lower()
+        return _ATTR2FIELD.get(name, name)
+
+    def _default_value(self, name):
+        if name in _LISTFIELDS or name in _ELEMENTSFIELD:
+            return []
+        return 'UNKNOWN'
+
+    def _remove_line_prefix(self, value):
+        if self.metadata_version in ('1.0', '1.1'):
+            return _LINE_PREFIX_PRE_1_2.sub('\n', value)
+        else:
+            return _LINE_PREFIX_1_2.sub('\n', value)
+
+    def __getattr__(self, name):
+        if name in _ATTR2FIELD:
+            return self[name]
+        raise AttributeError(name)
+
+    #
+    # Public API
+    #
+
+    def get_fullname(self, filesafe=False):
+        """
+        Return the distribution name with version.
+
+        If filesafe is true, return a filename-escaped form.
+        """
+        return _get_name_and_version(self['Name'], self['Version'], filesafe)
+
+    def is_field(self, name):
+        """return True if name is a valid metadata key"""
+        name = self._convert_name(name)
+        return name in _ALL_FIELDS
+
+    def is_multi_field(self, name):
+        name = self._convert_name(name)
+        return name in _LISTFIELDS
+
+    def read(self, filepath):
+        """Read the metadata values from a file path."""
+        fp = codecs.open(filepath, 'r', encoding='utf-8')
+        try:
+            self.read_file(fp)
+        finally:
+            fp.close()
+
+    def read_file(self, fileob):
+        """Read the metadata values from a file object."""
+        msg = message_from_file(fileob)
+        self._fields['Metadata-Version'] = msg['metadata-version']
+
+        # When reading, get all the fields we can
+        for field in _ALL_FIELDS:
+            if field not in msg:
+                continue
+            if field in _LISTFIELDS:
+                # we can have multiple lines
+                values = msg.get_all(field)
+                if field in _LISTTUPLEFIELDS and values is not None:
+                    values = [tuple(value.split(',')) for value in values]
+                self.set(field, values)
+            else:
+                # single line
+                value = msg[field]
+                if value is not None and value != 'UNKNOWN':
+                    self.set(field, value)
+
+        # PEP 566 specifies that the body be used for the description, if
+        # available
+        body = msg.get_payload()
+        self["Description"] = body if body else self["Description"]
+        # logger.debug('Attempting to set metadata for %s', self)
+        # self.set_metadata_version()
+
+    def write(self, filepath, skip_unknown=False):
+        """Write the metadata fields to filepath."""
+        fp = codecs.open(filepath, 'w', encoding='utf-8')
+        try:
+            self.write_file(fp, skip_unknown)
+        finally:
+            fp.close()
+
+    def write_file(self, fileobject, skip_unknown=False):
+        """Write the PKG-INFO format data to a file object."""
+        self.set_metadata_version()
+
+        for field in _version2fieldlist(self['Metadata-Version']):
+            values = self.get(field)
+            if skip_unknown and values in ('UNKNOWN', [], ['UNKNOWN']):
+                continue
+            if field in _ELEMENTSFIELD:
+                self._write_field(fileobject, field, ','.join(values))
+                continue
+            if field not in _LISTFIELDS:
+                if field == 'Description':
+                    if self.metadata_version in ('1.0', '1.1'):
+                        values = values.replace('\n', '\n        ')
+                    else:
+                        values = values.replace('\n', '\n       |')
+                values = [values]
+
+            if field in _LISTTUPLEFIELDS:
+                values = [','.join(value) for value in values]
+
+            for value in values:
+                self._write_field(fileobject, field, value)
+
+    def update(self, other=None, **kwargs):
+        """Set metadata values from the given iterable `other` and kwargs.
+
+        Behavior is like `dict.update`: If `other` has a ``keys`` method,
+        they are looped over and ``self[key]`` is assigned ``other[key]``.
+        Else, ``other`` is an iterable of ``(key, value)`` iterables.
+
+        Keys that don't match a metadata field or that have an empty value are
+        dropped.
+        """
+
+        def _set(key, value):
+            if key in _ATTR2FIELD and value:
+                self.set(self._convert_name(key), value)
+
+        if not other:
+            # other is None or empty container
+            pass
+        elif hasattr(other, 'keys'):
+            for k in other.keys():
+                _set(k, other[k])
+        else:
+            for k, v in other:
+                _set(k, v)
+
+        if kwargs:
+            for k, v in kwargs.items():
+                _set(k, v)
+
+    def set(self, name, value):
+        """Control then set a metadata field."""
+        name = self._convert_name(name)
+
+        if ((name in _ELEMENTSFIELD or name == 'Platform') and not isinstance(value, (list, tuple))):
+            if isinstance(value, string_types):
+                value = [v.strip() for v in value.split(',')]
+            else:
+                value = []
+        elif (name in _LISTFIELDS and not isinstance(value, (list, tuple))):
+            if isinstance(value, string_types):
+                value = [value]
+            else:
+                value = []
+
+        if logger.isEnabledFor(logging.WARNING):
+            project_name = self['Name']
+
+            scheme = get_scheme(self.scheme)
+            if name in _PREDICATE_FIELDS and value is not None:
+                for v in value:
+                    # check that the values are valid
+                    if not scheme.is_valid_matcher(v.split(';')[0]):
+                        logger.warning("'%s': '%s' is not valid (field '%s')", project_name, v, name)
+            # FIXME this rejects UNKNOWN, is that right?
+            elif name in _VERSIONS_FIELDS and value is not None:
+                if not scheme.is_valid_constraint_list(value):
+                    logger.warning("'%s': '%s' is not a valid version (field '%s')", project_name, value, name)
+            elif name in _VERSION_FIELDS and value is not None:
+                if not scheme.is_valid_version(value):
+                    logger.warning("'%s': '%s' is not a valid version (field '%s')", project_name, value, name)
+
+        if name in _UNICODEFIELDS:
+            if name == 'Description':
+                value = self._remove_line_prefix(value)
+
+        self._fields[name] = value
+
+    def get(self, name, default=_MISSING):
+        """Get a metadata field."""
+        name = self._convert_name(name)
+        if name not in self._fields:
+            if default is _MISSING:
+                default = self._default_value(name)
+            return default
+        if name in _UNICODEFIELDS:
+            value = self._fields[name]
+            return value
+        elif name in _LISTFIELDS:
+            value = self._fields[name]
+            if value is None:
+                return []
+            res = []
+            for val in value:
+                if name not in _LISTTUPLEFIELDS:
+                    res.append(val)
+                else:
+                    # That's for Project-URL
+                    res.append((val[0], val[1]))
+            return res
+
+        elif name in _ELEMENTSFIELD:
+            value = self._fields[name]
+            if isinstance(value, string_types):
+                return value.split(',')
+        return self._fields[name]
+
+    def check(self, strict=False):
+        """Check if the metadata is compliant. If strict is True then raise if
+        no Name or Version are provided"""
+        self.set_metadata_version()
+
+        # XXX should check the versions (if the file was loaded)
+        missing, warnings = [], []
+
+        for attr in ('Name', 'Version'):  # required by PEP 345
+            if attr not in self:
+                missing.append(attr)
+
+        if strict and missing != []:
+            msg = 'missing required metadata: %s' % ', '.join(missing)
+            raise MetadataMissingError(msg)
+
+        for attr in ('Home-page', 'Author'):
+            if attr not in self:
+                missing.append(attr)
+
+        # checking metadata 1.2 (XXX needs to check 1.1, 1.0)
+        if self['Metadata-Version'] != '1.2':
+            return missing, warnings
+
+        scheme = get_scheme(self.scheme)
+
+        def are_valid_constraints(value):
+            for v in value:
+                if not scheme.is_valid_matcher(v.split(';')[0]):
+                    return False
+            return True
+
+        for fields, controller in ((_PREDICATE_FIELDS, are_valid_constraints),
+                                   (_VERSIONS_FIELDS, scheme.is_valid_constraint_list), (_VERSION_FIELDS,
+                                                                                         scheme.is_valid_version)):
+            for field in fields:
+                value = self.get(field, None)
+                if value is not None and not controller(value):
+                    warnings.append("Wrong value for '%s': %s" % (field, value))
+
+        return missing, warnings
+
+    def todict(self, skip_missing=False):
+        """Return fields as a dict.
+
+        Field names will be converted to use the underscore-lowercase style
+        instead of hyphen-mixed case (i.e. home_page instead of Home-page).
+        This is as per https://www.python.org/dev/peps/pep-0566/#id17.
+        """
+        self.set_metadata_version()
+
+        fields = _version2fieldlist(self['Metadata-Version'])
+
+        data = {}
+
+        for field_name in fields:
+            if not skip_missing or field_name in self._fields:
+                key = _FIELD2ATTR[field_name]
+                if key != 'project_url':
+                    data[key] = self[field_name]
+                else:
+                    data[key] = [','.join(u) for u in self[field_name]]
+
+        return data
+
+    def add_requirements(self, requirements):
+        if self['Metadata-Version'] == '1.1':
+            # we can't have 1.1 metadata *and* Setuptools requires
+            for field in ('Obsoletes', 'Requires', 'Provides'):
+                if field in self:
+                    del self[field]
+        self['Requires-Dist'] += requirements
+
+    # Mapping API
+    # TODO could add iter* variants
+
+    def keys(self):
+        return list(_version2fieldlist(self['Metadata-Version']))
+
+    def __iter__(self):
+        for key in self.keys():
+            yield key
+
+    def values(self):
+        return [self[key] for key in self.keys()]
+
+    def items(self):
+        return [(key, self[key]) for key in self.keys()]
+
+    def __repr__(self):
+        return '<%s %s %s>' % (self.__class__.__name__, self.name, self.version)
+
+
+METADATA_FILENAME = 'pydist.json'
+WHEEL_METADATA_FILENAME = 'metadata.json'
+LEGACY_METADATA_FILENAME = 'METADATA'
+
+
+class Metadata(object):
+    """
+    The metadata of a release. This implementation uses 2.1
+    metadata where possible. If not possible, it wraps a LegacyMetadata
+    instance which handles the key-value metadata format.
+    """
+
+    METADATA_VERSION_MATCHER = re.compile(r'^\d+(\.\d+)*$')
+
+    NAME_MATCHER = re.compile('^[0-9A-Z]([0-9A-Z_.-]*[0-9A-Z])?$', re.I)
+
+    FIELDNAME_MATCHER = re.compile('^[A-Z]([0-9A-Z-]*[0-9A-Z])?$', re.I)
+
+    VERSION_MATCHER = PEP440_VERSION_RE
+
+    SUMMARY_MATCHER = re.compile('.{1,2047}')
+
+    METADATA_VERSION = '2.0'
+
+    GENERATOR = 'distlib (%s)' % __version__
+
+    MANDATORY_KEYS = {
+        'name': (),
+        'version': (),
+        'summary': ('legacy', ),
+    }
+
+    INDEX_KEYS = ('name version license summary description author '
+                  'author_email keywords platform home_page classifiers '
+                  'download_url')
+
+    DEPENDENCY_KEYS = ('extras run_requires test_requires build_requires '
+                       'dev_requires provides meta_requires obsoleted_by '
+                       'supports_environments')
+
+    SYNTAX_VALIDATORS = {
+        'metadata_version': (METADATA_VERSION_MATCHER, ()),
+        'name': (NAME_MATCHER, ('legacy', )),
+        'version': (VERSION_MATCHER, ('legacy', )),
+        'summary': (SUMMARY_MATCHER, ('legacy', )),
+        'dynamic': (FIELDNAME_MATCHER, ('legacy', )),
+    }
+
+    __slots__ = ('_legacy', '_data', 'scheme')
+
+    def __init__(self, path=None, fileobj=None, mapping=None, scheme='default'):
+        if [path, fileobj, mapping].count(None) < 2:
+            raise TypeError('path, fileobj and mapping are exclusive')
+        self._legacy = None
+        self._data = None
+        self.scheme = scheme
+        # import pdb; pdb.set_trace()
+        if mapping is not None:
+            try:
+                self._validate_mapping(mapping, scheme)
+                self._data = mapping
+            except MetadataUnrecognizedVersionError:
+                self._legacy = LegacyMetadata(mapping=mapping, scheme=scheme)
+                self.validate()
+        else:
+            data = None
+            if path:
+                with open(path, 'rb') as f:
+                    data = f.read()
+            elif fileobj:
+                data = fileobj.read()
+            if data is None:
+                # Initialised with no args - to be added
+                self._data = {
+                    'metadata_version': self.METADATA_VERSION,
+                    'generator': self.GENERATOR,
+                }
+            else:
+                if not isinstance(data, text_type):
+                    data = data.decode('utf-8')
+                try:
+                    self._data = json.loads(data)
+                    self._validate_mapping(self._data, scheme)
+                except ValueError:
+                    # Note: MetadataUnrecognizedVersionError does not
+                    # inherit from ValueError (it's a DistlibException,
+                    # which should not inherit from ValueError).
+                    # The ValueError comes from the json.load - if that
+                    # succeeds and we get a validation error, we want
+                    # that to propagate
+                    self._legacy = LegacyMetadata(fileobj=StringIO(data), scheme=scheme)
+                    self.validate()
+
+    common_keys = set(('name', 'version', 'license', 'keywords', 'summary'))
+
+    none_list = (None, list)
+    none_dict = (None, dict)
+
+    mapped_keys = {
+        'run_requires': ('Requires-Dist', list),
+        'build_requires': ('Setup-Requires-Dist', list),
+        'dev_requires': none_list,
+        'test_requires': none_list,
+        'meta_requires': none_list,
+        'extras': ('Provides-Extra', list),
+        'modules': none_list,
+        'namespaces': none_list,
+        'exports': none_dict,
+        'commands': none_dict,
+        'classifiers': ('Classifier', list),
+        'source_url': ('Download-URL', None),
+        'metadata_version': ('Metadata-Version', None),
+    }
+
+    del none_list, none_dict
+
+    def __getattribute__(self, key):
+        common = object.__getattribute__(self, 'common_keys')
+        mapped = object.__getattribute__(self, 'mapped_keys')
+        if key in mapped:
+            lk, maker = mapped[key]
+            if self._legacy:
+                if lk is None:
+                    result = None if maker is None else maker()
+                else:
+                    result = self._legacy.get(lk)
+            else:
+                value = None if maker is None else maker()
+                if key not in ('commands', 'exports', 'modules', 'namespaces', 'classifiers'):
+                    result = self._data.get(key, value)
+                else:
+                    # special cases for PEP 459
+                    sentinel = object()
+                    result = sentinel
+                    d = self._data.get('extensions')
+                    if d:
+                        if key == 'commands':
+                            result = d.get('python.commands', value)
+                        elif key == 'classifiers':
+                            d = d.get('python.details')
+                            if d:
+                                result = d.get(key, value)
+                        else:
+                            d = d.get('python.exports')
+                            if not d:
+                                d = self._data.get('python.exports')
+                            if d:
+                                result = d.get(key, value)
+                    if result is sentinel:
+                        result = value
+        elif key not in common:
+            result = object.__getattribute__(self, key)
+        elif self._legacy:
+            result = self._legacy.get(key)
+        else:
+            result = self._data.get(key)
+        return result
+
+    def _validate_value(self, key, value, scheme=None):
+        if key in self.SYNTAX_VALIDATORS:
+            pattern, exclusions = self.SYNTAX_VALIDATORS[key]
+            if (scheme or self.scheme) not in exclusions:
+                m = pattern.match(value)
+                if not m:
+                    raise MetadataInvalidError("'%s' is an invalid value for "
+                                               "the '%s' property" % (value, key))
+
+    def __setattr__(self, key, value):
+        self._validate_value(key, value)
+        common = object.__getattribute__(self, 'common_keys')
+        mapped = object.__getattribute__(self, 'mapped_keys')
+        if key in mapped:
+            lk, _ = mapped[key]
+            if self._legacy:
+                if lk is None:
+                    raise NotImplementedError
+                self._legacy[lk] = value
+            elif key not in ('commands', 'exports', 'modules', 'namespaces', 'classifiers'):
+                self._data[key] = value
+            else:
+                # special cases for PEP 459
+                d = self._data.setdefault('extensions', {})
+                if key == 'commands':
+                    d['python.commands'] = value
+                elif key == 'classifiers':
+                    d = d.setdefault('python.details', {})
+                    d[key] = value
+                else:
+                    d = d.setdefault('python.exports', {})
+                    d[key] = value
+        elif key not in common:
+            object.__setattr__(self, key, value)
+        else:
+            if key == 'keywords':
+                if isinstance(value, string_types):
+                    value = value.strip()
+                    if value:
+                        value = value.split()
+                    else:
+                        value = []
+            if self._legacy:
+                self._legacy[key] = value
+            else:
+                self._data[key] = value
+
+    @property
+    def name_and_version(self):
+        return _get_name_and_version(self.name, self.version, True)
+
+    @property
+    def provides(self):
+        if self._legacy:
+            result = self._legacy['Provides-Dist']
+        else:
+            result = self._data.setdefault('provides', [])
+        s = '%s (%s)' % (self.name, self.version)
+        if s not in result:
+            result.append(s)
+        return result
+
+    @provides.setter
+    def provides(self, value):
+        if self._legacy:
+            self._legacy['Provides-Dist'] = value
+        else:
+            self._data['provides'] = value
+
+    def get_requirements(self, reqts, extras=None, env=None):
+        """
+        Base method to get dependencies, given a set of extras
+        to satisfy and an optional environment context.
+        :param reqts: A list of sometimes-wanted dependencies,
+                      perhaps dependent on extras and environment.
+        :param extras: A list of optional components being requested.
+        :param env: An optional environment for marker evaluation.
+        """
+        if self._legacy:
+            result = reqts
+        else:
+            result = []
+            extras = get_extras(extras or [], self.extras)
+            for d in reqts:
+                if 'extra' not in d and 'environment' not in d:
+                    # unconditional
+                    include = True
+                else:
+                    if 'extra' not in d:
+                        # Not extra-dependent - only environment-dependent
+                        include = True
+                    else:
+                        include = d.get('extra') in extras
+                    if include:
+                        # Not excluded because of extras, check environment
+                        marker = d.get('environment')
+                        if marker:
+                            include = interpret(marker, env)
+                if include:
+                    result.extend(d['requires'])
+            for key in ('build', 'dev', 'test'):
+                e = ':%s:' % key
+                if e in extras:
+                    extras.remove(e)
+                    # A recursive call, but it should terminate since 'test'
+                    # has been removed from the extras
+                    reqts = self._data.get('%s_requires' % key, [])
+                    result.extend(self.get_requirements(reqts, extras=extras, env=env))
+        return result
+
+    @property
+    def dictionary(self):
+        if self._legacy:
+            return self._from_legacy()
+        return self._data
+
+    @property
+    def dependencies(self):
+        if self._legacy:
+            raise NotImplementedError
+        else:
+            return extract_by_key(self._data, self.DEPENDENCY_KEYS)
+
+    @dependencies.setter
+    def dependencies(self, value):
+        if self._legacy:
+            raise NotImplementedError
+        else:
+            self._data.update(value)
+
+    def _validate_mapping(self, mapping, scheme):
+        if mapping.get('metadata_version') != self.METADATA_VERSION:
+            raise MetadataUnrecognizedVersionError()
+        missing = []
+        for key, exclusions in self.MANDATORY_KEYS.items():
+            if key not in mapping:
+                if scheme not in exclusions:
+                    missing.append(key)
+        if missing:
+            msg = 'Missing metadata items: %s' % ', '.join(missing)
+            raise MetadataMissingError(msg)
+        for k, v in mapping.items():
+            self._validate_value(k, v, scheme)
+
+    def validate(self):
+        if self._legacy:
+            missing, warnings = self._legacy.check(True)
+            if missing or warnings:
+                logger.warning('Metadata: missing: %s, warnings: %s', missing, warnings)
+        else:
+            self._validate_mapping(self._data, self.scheme)
+
+    def todict(self):
+        if self._legacy:
+            return self._legacy.todict(True)
+        else:
+            result = extract_by_key(self._data, self.INDEX_KEYS)
+            return result
+
+    def _from_legacy(self):
+        assert self._legacy and not self._data
+        result = {
+            'metadata_version': self.METADATA_VERSION,
+            'generator': self.GENERATOR,
+        }
+        lmd = self._legacy.todict(True)  # skip missing ones
+        for k in ('name', 'version', 'license', 'summary', 'description', 'classifier'):
+            if k in lmd:
+                if k == 'classifier':
+                    nk = 'classifiers'
+                else:
+                    nk = k
+                result[nk] = lmd[k]
+        kw = lmd.get('Keywords', [])
+        if kw == ['']:
+            kw = []
+        result['keywords'] = kw
+        keys = (('requires_dist', 'run_requires'), ('setup_requires_dist', 'build_requires'))
+        for ok, nk in keys:
+            if ok in lmd and lmd[ok]:
+                result[nk] = [{'requires': lmd[ok]}]
+        result['provides'] = self.provides
+        # author = {}
+        # maintainer = {}
+        return result
+
+    LEGACY_MAPPING = {
+        'name': 'Name',
+        'version': 'Version',
+        ('extensions', 'python.details', 'license'): 'License',
+        'summary': 'Summary',
+        'description': 'Description',
+        ('extensions', 'python.project', 'project_urls', 'Home'): 'Home-page',
+        ('extensions', 'python.project', 'contacts', 0, 'name'): 'Author',
+        ('extensions', 'python.project', 'contacts', 0, 'email'): 'Author-email',
+        'source_url': 'Download-URL',
+        ('extensions', 'python.details', 'classifiers'): 'Classifier',
+    }
+
+    def _to_legacy(self):
+
+        def process_entries(entries):
+            reqts = set()
+            for e in entries:
+                extra = e.get('extra')
+                env = e.get('environment')
+                rlist = e['requires']
+                for r in rlist:
+                    if not env and not extra:
+                        reqts.add(r)
+                    else:
+                        marker = ''
+                        if extra:
+                            marker = 'extra == "%s"' % extra
+                        if env:
+                            if marker:
+                                marker = '(%s) and %s' % (env, marker)
+                            else:
+                                marker = env
+                        reqts.add(';'.join((r, marker)))
+            return reqts
+
+        assert self._data and not self._legacy
+        result = LegacyMetadata()
+        nmd = self._data
+        # import pdb; pdb.set_trace()
+        for nk, ok in self.LEGACY_MAPPING.items():
+            if not isinstance(nk, tuple):
+                if nk in nmd:
+                    result[ok] = nmd[nk]
+            else:
+                d = nmd
+                found = True
+                for k in nk:
+                    try:
+                        d = d[k]
+                    except (KeyError, IndexError):
+                        found = False
+                        break
+                if found:
+                    result[ok] = d
+        r1 = process_entries(self.run_requires + self.meta_requires)
+        r2 = process_entries(self.build_requires + self.dev_requires)
+        if self.extras:
+            result['Provides-Extra'] = sorted(self.extras)
+        result['Requires-Dist'] = sorted(r1)
+        result['Setup-Requires-Dist'] = sorted(r2)
+        # TODO: any other fields wanted
+        return result
+
+    def write(self, path=None, fileobj=None, legacy=False, skip_unknown=True):
+        if [path, fileobj].count(None) != 1:
+            raise ValueError('Exactly one of path and fileobj is needed')
+        self.validate()
+        if legacy:
+            if self._legacy:
+                legacy_md = self._legacy
+            else:
+                legacy_md = self._to_legacy()
+            if path:
+                legacy_md.write(path, skip_unknown=skip_unknown)
+            else:
+                legacy_md.write_file(fileobj, skip_unknown=skip_unknown)
+        else:
+            if self._legacy:
+                d = self._from_legacy()
+            else:
+                d = self._data
+            if fileobj:
+                json.dump(d, fileobj, ensure_ascii=True, indent=2, sort_keys=True)
+            else:
+                with codecs.open(path, 'w', 'utf-8') as f:
+                    json.dump(d, f, ensure_ascii=True, indent=2, sort_keys=True)
+
+    def add_requirements(self, requirements):
+        if self._legacy:
+            self._legacy.add_requirements(requirements)
+        else:
+            run_requires = self._data.setdefault('run_requires', [])
+            always = None
+            for entry in run_requires:
+                if 'environment' not in entry and 'extra' not in entry:
+                    always = entry
+                    break
+            if always is None:
+                always = {'requires': requirements}
+                run_requires.insert(0, always)
+            else:
+                rset = set(always['requires']) | set(requirements)
+                always['requires'] = sorted(rset)
+
+    def __repr__(self):
+        name = self.name or '(no name)'
+        version = self.version or 'no version'
+        return '<%s %s %s (%s)>' % (self.__class__.__name__, self.metadata_version, name, version)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/resources.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/resources.py
new file mode 100644
index 00000000..fef52aa1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/resources.py
@@ -0,0 +1,358 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2013-2017 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+from __future__ import unicode_literals
+
+import bisect
+import io
+import logging
+import os
+import pkgutil
+import sys
+import types
+import zipimport
+
+from . import DistlibException
+from .util import cached_property, get_cache_base, Cache
+
+logger = logging.getLogger(__name__)
+
+
+cache = None    # created when needed
+
+
+class ResourceCache(Cache):
+    def __init__(self, base=None):
+        if base is None:
+            # Use native string to avoid issues on 2.x: see Python #20140.
+            base = os.path.join(get_cache_base(), str('resource-cache'))
+        super(ResourceCache, self).__init__(base)
+
+    def is_stale(self, resource, path):
+        """
+        Is the cache stale for the given resource?
+
+        :param resource: The :class:`Resource` being cached.
+        :param path: The path of the resource in the cache.
+        :return: True if the cache is stale.
+        """
+        # Cache invalidation is a hard problem :-)
+        return True
+
+    def get(self, resource):
+        """
+        Get a resource into the cache,
+
+        :param resource: A :class:`Resource` instance.
+        :return: The pathname of the resource in the cache.
+        """
+        prefix, path = resource.finder.get_cache_info(resource)
+        if prefix is None:
+            result = path
+        else:
+            result = os.path.join(self.base, self.prefix_to_dir(prefix), path)
+            dirname = os.path.dirname(result)
+            if not os.path.isdir(dirname):
+                os.makedirs(dirname)
+            if not os.path.exists(result):
+                stale = True
+            else:
+                stale = self.is_stale(resource, path)
+            if stale:
+                # write the bytes of the resource to the cache location
+                with open(result, 'wb') as f:
+                    f.write(resource.bytes)
+        return result
+
+
+class ResourceBase(object):
+    def __init__(self, finder, name):
+        self.finder = finder
+        self.name = name
+
+
+class Resource(ResourceBase):
+    """
+    A class representing an in-package resource, such as a data file. This is
+    not normally instantiated by user code, but rather by a
+    :class:`ResourceFinder` which manages the resource.
+    """
+    is_container = False        # Backwards compatibility
+
+    def as_stream(self):
+        """
+        Get the resource as a stream.
+
+        This is not a property to make it obvious that it returns a new stream
+        each time.
+        """
+        return self.finder.get_stream(self)
+
+    @cached_property
+    def file_path(self):
+        global cache
+        if cache is None:
+            cache = ResourceCache()
+        return cache.get(self)
+
+    @cached_property
+    def bytes(self):
+        return self.finder.get_bytes(self)
+
+    @cached_property
+    def size(self):
+        return self.finder.get_size(self)
+
+
+class ResourceContainer(ResourceBase):
+    is_container = True     # Backwards compatibility
+
+    @cached_property
+    def resources(self):
+        return self.finder.get_resources(self)
+
+
+class ResourceFinder(object):
+    """
+    Resource finder for file system resources.
+    """
+
+    if sys.platform.startswith('java'):
+        skipped_extensions = ('.pyc', '.pyo', '.class')
+    else:
+        skipped_extensions = ('.pyc', '.pyo')
+
+    def __init__(self, module):
+        self.module = module
+        self.loader = getattr(module, '__loader__', None)
+        self.base = os.path.dirname(getattr(module, '__file__', ''))
+
+    def _adjust_path(self, path):
+        return os.path.realpath(path)
+
+    def _make_path(self, resource_name):
+        # Issue #50: need to preserve type of path on Python 2.x
+        # like os.path._get_sep
+        if isinstance(resource_name, bytes):    # should only happen on 2.x
+            sep = b'/'
+        else:
+            sep = '/'
+        parts = resource_name.split(sep)
+        parts.insert(0, self.base)
+        result = os.path.join(*parts)
+        return self._adjust_path(result)
+
+    def _find(self, path):
+        return os.path.exists(path)
+
+    def get_cache_info(self, resource):
+        return None, resource.path
+
+    def find(self, resource_name):
+        path = self._make_path(resource_name)
+        if not self._find(path):
+            result = None
+        else:
+            if self._is_directory(path):
+                result = ResourceContainer(self, resource_name)
+            else:
+                result = Resource(self, resource_name)
+            result.path = path
+        return result
+
+    def get_stream(self, resource):
+        return open(resource.path, 'rb')
+
+    def get_bytes(self, resource):
+        with open(resource.path, 'rb') as f:
+            return f.read()
+
+    def get_size(self, resource):
+        return os.path.getsize(resource.path)
+
+    def get_resources(self, resource):
+        def allowed(f):
+            return (f != '__pycache__' and not
+                    f.endswith(self.skipped_extensions))
+        return set([f for f in os.listdir(resource.path) if allowed(f)])
+
+    def is_container(self, resource):
+        return self._is_directory(resource.path)
+
+    _is_directory = staticmethod(os.path.isdir)
+
+    def iterator(self, resource_name):
+        resource = self.find(resource_name)
+        if resource is not None:
+            todo = [resource]
+            while todo:
+                resource = todo.pop(0)
+                yield resource
+                if resource.is_container:
+                    rname = resource.name
+                    for name in resource.resources:
+                        if not rname:
+                            new_name = name
+                        else:
+                            new_name = '/'.join([rname, name])
+                        child = self.find(new_name)
+                        if child.is_container:
+                            todo.append(child)
+                        else:
+                            yield child
+
+
+class ZipResourceFinder(ResourceFinder):
+    """
+    Resource finder for resources in .zip files.
+    """
+    def __init__(self, module):
+        super(ZipResourceFinder, self).__init__(module)
+        archive = self.loader.archive
+        self.prefix_len = 1 + len(archive)
+        # PyPy doesn't have a _files attr on zipimporter, and you can't set one
+        if hasattr(self.loader, '_files'):
+            self._files = self.loader._files
+        else:
+            self._files = zipimport._zip_directory_cache[archive]
+        self.index = sorted(self._files)
+
+    def _adjust_path(self, path):
+        return path
+
+    def _find(self, path):
+        path = path[self.prefix_len:]
+        if path in self._files:
+            result = True
+        else:
+            if path and path[-1] != os.sep:
+                path = path + os.sep
+            i = bisect.bisect(self.index, path)
+            try:
+                result = self.index[i].startswith(path)
+            except IndexError:
+                result = False
+        if not result:
+            logger.debug('_find failed: %r %r', path, self.loader.prefix)
+        else:
+            logger.debug('_find worked: %r %r', path, self.loader.prefix)
+        return result
+
+    def get_cache_info(self, resource):
+        prefix = self.loader.archive
+        path = resource.path[1 + len(prefix):]
+        return prefix, path
+
+    def get_bytes(self, resource):
+        return self.loader.get_data(resource.path)
+
+    def get_stream(self, resource):
+        return io.BytesIO(self.get_bytes(resource))
+
+    def get_size(self, resource):
+        path = resource.path[self.prefix_len:]
+        return self._files[path][3]
+
+    def get_resources(self, resource):
+        path = resource.path[self.prefix_len:]
+        if path and path[-1] != os.sep:
+            path += os.sep
+        plen = len(path)
+        result = set()
+        i = bisect.bisect(self.index, path)
+        while i < len(self.index):
+            if not self.index[i].startswith(path):
+                break
+            s = self.index[i][plen:]
+            result.add(s.split(os.sep, 1)[0])   # only immediate children
+            i += 1
+        return result
+
+    def _is_directory(self, path):
+        path = path[self.prefix_len:]
+        if path and path[-1] != os.sep:
+            path += os.sep
+        i = bisect.bisect(self.index, path)
+        try:
+            result = self.index[i].startswith(path)
+        except IndexError:
+            result = False
+        return result
+
+
+_finder_registry = {
+    type(None): ResourceFinder,
+    zipimport.zipimporter: ZipResourceFinder
+}
+
+try:
+    # In Python 3.6, _frozen_importlib -> _frozen_importlib_external
+    try:
+        import _frozen_importlib_external as _fi
+    except ImportError:
+        import _frozen_importlib as _fi
+    _finder_registry[_fi.SourceFileLoader] = ResourceFinder
+    _finder_registry[_fi.FileFinder] = ResourceFinder
+    # See issue #146
+    _finder_registry[_fi.SourcelessFileLoader] = ResourceFinder
+    del _fi
+except (ImportError, AttributeError):
+    pass
+
+
+def register_finder(loader, finder_maker):
+    _finder_registry[type(loader)] = finder_maker
+
+
+_finder_cache = {}
+
+
+def finder(package):
+    """
+    Return a resource finder for a package.
+    :param package: The name of the package.
+    :return: A :class:`ResourceFinder` instance for the package.
+    """
+    if package in _finder_cache:
+        result = _finder_cache[package]
+    else:
+        if package not in sys.modules:
+            __import__(package)
+        module = sys.modules[package]
+        path = getattr(module, '__path__', None)
+        if path is None:
+            raise DistlibException('You cannot get a finder for a module, '
+                                   'only for a package')
+        loader = getattr(module, '__loader__', None)
+        finder_maker = _finder_registry.get(type(loader))
+        if finder_maker is None:
+            raise DistlibException('Unable to locate finder for %r' % package)
+        result = finder_maker(module)
+        _finder_cache[package] = result
+    return result
+
+
+_dummy_module = types.ModuleType(str('__dummy__'))
+
+
+def finder_for_path(path):
+    """
+    Return a resource finder for a path, which should represent a container.
+
+    :param path: The path.
+    :return: A :class:`ResourceFinder` instance for the path.
+    """
+    result = None
+    # calls any path hooks, gets importer into cache
+    pkgutil.get_importer(path)
+    loader = sys.path_importer_cache.get(path)
+    finder = _finder_registry.get(type(loader))
+    if finder:
+        module = _dummy_module
+        module.__file__ = os.path.join(path, '')
+        module.__loader__ = loader
+        result = finder(module)
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/scripts.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/scripts.py
new file mode 100644
index 00000000..b1fc705b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/scripts.py
@@ -0,0 +1,447 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2013-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+from io import BytesIO
+import logging
+import os
+import re
+import struct
+import sys
+import time
+from zipfile import ZipInfo
+
+from .compat import sysconfig, detect_encoding, ZipFile
+from .resources import finder
+from .util import (FileOperator, get_export_entry, convert_path, get_executable, get_platform, in_venv)
+
+logger = logging.getLogger(__name__)
+
+_DEFAULT_MANIFEST = '''
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">
+ <assemblyIdentity version="1.0.0.0"
+ processorArchitecture="X86"
+ name="%s"
+ type="win32"/>
+
+ <!-- Identify the application security requirements. -->
+ <trustInfo xmlns="urn:schemas-microsoft-com:asm.v3">
+ <security>
+ <requestedPrivileges>
+ <requestedExecutionLevel level="asInvoker" uiAccess="false"/>
+ </requestedPrivileges>
+ </security>
+ </trustInfo>
+</assembly>'''.strip()
+
+# check if Python is called on the first line with this expression
+FIRST_LINE_RE = re.compile(b'^#!.*pythonw?[0-9.]*([ \t].*)?$')
+SCRIPT_TEMPLATE = r'''# -*- coding: utf-8 -*-
+import re
+import sys
+from %(module)s import %(import_name)s
+if __name__ == '__main__':
+    sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])
+    sys.exit(%(func)s())
+'''
+
+# Pre-fetch the contents of all executable wrapper stubs.
+# This is to address https://github.com/pypa/pip/issues/12666.
+# When updating pip, we rename the old pip in place before installing the
+# new version. If we try to fetch a wrapper *after* that rename, the finder
+# machinery will be confused as the package is no longer available at the
+# location where it was imported from. So we load everything into memory in
+# advance.
+
+if os.name == 'nt' or (os.name == 'java' and os._name == 'nt'):
+    # Issue 31: don't hardcode an absolute package name, but
+    # determine it relative to the current package
+    DISTLIB_PACKAGE = __name__.rsplit('.', 1)[0]
+
+    WRAPPERS = {
+        r.name: r.bytes
+        for r in finder(DISTLIB_PACKAGE).iterator("")
+        if r.name.endswith(".exe")
+    }
+
+
+def enquote_executable(executable):
+    if ' ' in executable:
+        # make sure we quote only the executable in case of env
+        # for example /usr/bin/env "/dir with spaces/bin/jython"
+        # instead of "/usr/bin/env /dir with spaces/bin/jython"
+        # otherwise whole
+        if executable.startswith('/usr/bin/env '):
+            env, _executable = executable.split(' ', 1)
+            if ' ' in _executable and not _executable.startswith('"'):
+                executable = '%s "%s"' % (env, _executable)
+        else:
+            if not executable.startswith('"'):
+                executable = '"%s"' % executable
+    return executable
+
+
+# Keep the old name around (for now), as there is at least one project using it!
+_enquote_executable = enquote_executable
+
+
+class ScriptMaker(object):
+    """
+    A class to copy or create scripts from source scripts or callable
+    specifications.
+    """
+    script_template = SCRIPT_TEMPLATE
+
+    executable = None  # for shebangs
+
+    def __init__(self, source_dir, target_dir, add_launchers=True, dry_run=False, fileop=None):
+        self.source_dir = source_dir
+        self.target_dir = target_dir
+        self.add_launchers = add_launchers
+        self.force = False
+        self.clobber = False
+        # It only makes sense to set mode bits on POSIX.
+        self.set_mode = (os.name == 'posix') or (os.name == 'java' and os._name == 'posix')
+        self.variants = set(('', 'X.Y'))
+        self._fileop = fileop or FileOperator(dry_run)
+
+        self._is_nt = os.name == 'nt' or (os.name == 'java' and os._name == 'nt')
+        self.version_info = sys.version_info
+
+    def _get_alternate_executable(self, executable, options):
+        if options.get('gui', False) and self._is_nt:  # pragma: no cover
+            dn, fn = os.path.split(executable)
+            fn = fn.replace('python', 'pythonw')
+            executable = os.path.join(dn, fn)
+        return executable
+
+    if sys.platform.startswith('java'):  # pragma: no cover
+
+        def _is_shell(self, executable):
+            """
+            Determine if the specified executable is a script
+            (contains a #! line)
+            """
+            try:
+                with open(executable) as fp:
+                    return fp.read(2) == '#!'
+            except (OSError, IOError):
+                logger.warning('Failed to open %s', executable)
+                return False
+
+        def _fix_jython_executable(self, executable):
+            if self._is_shell(executable):
+                # Workaround for Jython is not needed on Linux systems.
+                import java
+
+                if java.lang.System.getProperty('os.name') == 'Linux':
+                    return executable
+            elif executable.lower().endswith('jython.exe'):
+                # Use wrapper exe for Jython on Windows
+                return executable
+            return '/usr/bin/env %s' % executable
+
+    def _build_shebang(self, executable, post_interp):
+        """
+        Build a shebang line. In the simple case (on Windows, or a shebang line
+        which is not too long or contains spaces) use a simple formulation for
+        the shebang. Otherwise, use /bin/sh as the executable, with a contrived
+        shebang which allows the script to run either under Python or sh, using
+        suitable quoting. Thanks to Harald Nordgren for his input.
+
+        See also: http://www.in-ulm.de/~mascheck/various/shebang/#length
+                  https://hg.mozilla.org/mozilla-central/file/tip/mach
+        """
+        if os.name != 'posix':
+            simple_shebang = True
+        elif getattr(sys, "cross_compiling", False):
+            # In a cross-compiling environment, the shebang will likely be a
+            # script; this *must* be invoked with the "safe" version of the
+            # shebang, or else using os.exec() to run the entry script will
+            # fail, raising "OSError 8 [Errno 8] Exec format error".
+            simple_shebang = False
+        else:
+            # Add 3 for '#!' prefix and newline suffix.
+            shebang_length = len(executable) + len(post_interp) + 3
+            if sys.platform == 'darwin':
+                max_shebang_length = 512
+            else:
+                max_shebang_length = 127
+            simple_shebang = ((b' ' not in executable) and (shebang_length <= max_shebang_length))
+
+        if simple_shebang:
+            result = b'#!' + executable + post_interp + b'\n'
+        else:
+            result = b'#!/bin/sh\n'
+            result += b"'''exec' " + executable + post_interp + b' "$0" "$@"\n'
+            result += b"' '''\n"
+        return result
+
+    def _get_shebang(self, encoding, post_interp=b'', options=None):
+        enquote = True
+        if self.executable:
+            executable = self.executable
+            enquote = False  # assume this will be taken care of
+        elif not sysconfig.is_python_build():
+            executable = get_executable()
+        elif in_venv():  # pragma: no cover
+            executable = os.path.join(sysconfig.get_path('scripts'), 'python%s' % sysconfig.get_config_var('EXE'))
+        else:  # pragma: no cover
+            if os.name == 'nt':
+                # for Python builds from source on Windows, no Python executables with
+                # a version suffix are created, so we use python.exe
+                executable = os.path.join(sysconfig.get_config_var('BINDIR'),
+                                          'python%s' % (sysconfig.get_config_var('EXE')))
+            else:
+                executable = os.path.join(
+                    sysconfig.get_config_var('BINDIR'),
+                    'python%s%s' % (sysconfig.get_config_var('VERSION'), sysconfig.get_config_var('EXE')))
+        if options:
+            executable = self._get_alternate_executable(executable, options)
+
+        if sys.platform.startswith('java'):  # pragma: no cover
+            executable = self._fix_jython_executable(executable)
+
+        # Normalise case for Windows - COMMENTED OUT
+        # executable = os.path.normcase(executable)
+        # N.B. The normalising operation above has been commented out: See
+        # issue #124. Although paths in Windows are generally case-insensitive,
+        # they aren't always. For example, a path containing a ẞ (which is a
+        # LATIN CAPITAL LETTER SHARP S - U+1E9E) is normcased to ß (which is a
+        # LATIN SMALL LETTER SHARP S' - U+00DF). The two are not considered by
+        # Windows as equivalent in path names.
+
+        # If the user didn't specify an executable, it may be necessary to
+        # cater for executable paths with spaces (not uncommon on Windows)
+        if enquote:
+            executable = enquote_executable(executable)
+        # Issue #51: don't use fsencode, since we later try to
+        # check that the shebang is decodable using utf-8.
+        executable = executable.encode('utf-8')
+        # in case of IronPython, play safe and enable frames support
+        if (sys.platform == 'cli' and '-X:Frames' not in post_interp and
+                '-X:FullFrames' not in post_interp):  # pragma: no cover
+            post_interp += b' -X:Frames'
+        shebang = self._build_shebang(executable, post_interp)
+        # Python parser starts to read a script using UTF-8 until
+        # it gets a #coding:xxx cookie. The shebang has to be the
+        # first line of a file, the #coding:xxx cookie cannot be
+        # written before. So the shebang has to be decodable from
+        # UTF-8.
+        try:
+            shebang.decode('utf-8')
+        except UnicodeDecodeError:  # pragma: no cover
+            raise ValueError('The shebang (%r) is not decodable from utf-8' % shebang)
+        # If the script is encoded to a custom encoding (use a
+        # #coding:xxx cookie), the shebang has to be decodable from
+        # the script encoding too.
+        if encoding != 'utf-8':
+            try:
+                shebang.decode(encoding)
+            except UnicodeDecodeError:  # pragma: no cover
+                raise ValueError('The shebang (%r) is not decodable '
+                                 'from the script encoding (%r)' % (shebang, encoding))
+        return shebang
+
+    def _get_script_text(self, entry):
+        return self.script_template % dict(
+            module=entry.prefix, import_name=entry.suffix.split('.')[0], func=entry.suffix)
+
+    manifest = _DEFAULT_MANIFEST
+
+    def get_manifest(self, exename):
+        base = os.path.basename(exename)
+        return self.manifest % base
+
+    def _write_script(self, names, shebang, script_bytes, filenames, ext):
+        use_launcher = self.add_launchers and self._is_nt
+        if not use_launcher:
+            script_bytes = shebang + script_bytes
+        else:  # pragma: no cover
+            if ext == 'py':
+                launcher = self._get_launcher('t')
+            else:
+                launcher = self._get_launcher('w')
+            stream = BytesIO()
+            with ZipFile(stream, 'w') as zf:
+                source_date_epoch = os.environ.get('SOURCE_DATE_EPOCH')
+                if source_date_epoch:
+                    date_time = time.gmtime(int(source_date_epoch))[:6]
+                    zinfo = ZipInfo(filename='__main__.py', date_time=date_time)
+                    zf.writestr(zinfo, script_bytes)
+                else:
+                    zf.writestr('__main__.py', script_bytes)
+            zip_data = stream.getvalue()
+            script_bytes = launcher + shebang + zip_data
+        for name in names:
+            outname = os.path.join(self.target_dir, name)
+            if use_launcher:  # pragma: no cover
+                n, e = os.path.splitext(outname)
+                if e.startswith('.py'):
+                    outname = n
+                outname = '%s.exe' % outname
+                try:
+                    self._fileop.write_binary_file(outname, script_bytes)
+                except Exception:
+                    # Failed writing an executable - it might be in use.
+                    logger.warning('Failed to write executable - trying to '
+                                   'use .deleteme logic')
+                    dfname = '%s.deleteme' % outname
+                    if os.path.exists(dfname):
+                        os.remove(dfname)  # Not allowed to fail here
+                    os.rename(outname, dfname)  # nor here
+                    self._fileop.write_binary_file(outname, script_bytes)
+                    logger.debug('Able to replace executable using '
+                                 '.deleteme logic')
+                    try:
+                        os.remove(dfname)
+                    except Exception:
+                        pass  # still in use - ignore error
+            else:
+                if self._is_nt and not outname.endswith('.' + ext):  # pragma: no cover
+                    outname = '%s.%s' % (outname, ext)
+                if os.path.exists(outname) and not self.clobber:
+                    logger.warning('Skipping existing file %s', outname)
+                    continue
+                self._fileop.write_binary_file(outname, script_bytes)
+                if self.set_mode:
+                    self._fileop.set_executable_mode([outname])
+            filenames.append(outname)
+
+    variant_separator = '-'
+
+    def get_script_filenames(self, name):
+        result = set()
+        if '' in self.variants:
+            result.add(name)
+        if 'X' in self.variants:
+            result.add('%s%s' % (name, self.version_info[0]))
+        if 'X.Y' in self.variants:
+            result.add('%s%s%s.%s' % (name, self.variant_separator, self.version_info[0], self.version_info[1]))
+        return result
+
+    def _make_script(self, entry, filenames, options=None):
+        post_interp = b''
+        if options:
+            args = options.get('interpreter_args', [])
+            if args:
+                args = ' %s' % ' '.join(args)
+                post_interp = args.encode('utf-8')
+        shebang = self._get_shebang('utf-8', post_interp, options=options)
+        script = self._get_script_text(entry).encode('utf-8')
+        scriptnames = self.get_script_filenames(entry.name)
+        if options and options.get('gui', False):
+            ext = 'pyw'
+        else:
+            ext = 'py'
+        self._write_script(scriptnames, shebang, script, filenames, ext)
+
+    def _copy_script(self, script, filenames):
+        adjust = False
+        script = os.path.join(self.source_dir, convert_path(script))
+        outname = os.path.join(self.target_dir, os.path.basename(script))
+        if not self.force and not self._fileop.newer(script, outname):
+            logger.debug('not copying %s (up-to-date)', script)
+            return
+
+        # Always open the file, but ignore failures in dry-run mode --
+        # that way, we'll get accurate feedback if we can read the
+        # script.
+        try:
+            f = open(script, 'rb')
+        except IOError:  # pragma: no cover
+            if not self.dry_run:
+                raise
+            f = None
+        else:
+            first_line = f.readline()
+            if not first_line:  # pragma: no cover
+                logger.warning('%s is an empty file (skipping)', script)
+                return
+
+            match = FIRST_LINE_RE.match(first_line.replace(b'\r\n', b'\n'))
+            if match:
+                adjust = True
+                post_interp = match.group(1) or b''
+
+        if not adjust:
+            if f:
+                f.close()
+            self._fileop.copy_file(script, outname)
+            if self.set_mode:
+                self._fileop.set_executable_mode([outname])
+            filenames.append(outname)
+        else:
+            logger.info('copying and adjusting %s -> %s', script, self.target_dir)
+            if not self._fileop.dry_run:
+                encoding, lines = detect_encoding(f.readline)
+                f.seek(0)
+                shebang = self._get_shebang(encoding, post_interp)
+                if b'pythonw' in first_line:  # pragma: no cover
+                    ext = 'pyw'
+                else:
+                    ext = 'py'
+                n = os.path.basename(outname)
+                self._write_script([n], shebang, f.read(), filenames, ext)
+            if f:
+                f.close()
+
+    @property
+    def dry_run(self):
+        return self._fileop.dry_run
+
+    @dry_run.setter
+    def dry_run(self, value):
+        self._fileop.dry_run = value
+
+    if os.name == 'nt' or (os.name == 'java' and os._name == 'nt'):  # pragma: no cover
+        # Executable launcher support.
+        # Launchers are from https://bitbucket.org/vinay.sajip/simple_launcher/
+
+        def _get_launcher(self, kind):
+            if struct.calcsize('P') == 8:  # 64-bit
+                bits = '64'
+            else:
+                bits = '32'
+            platform_suffix = '-arm' if get_platform() == 'win-arm64' else ''
+            name = '%s%s%s.exe' % (kind, bits, platform_suffix)
+            if name not in WRAPPERS:
+                msg = ('Unable to find resource %s in package %s' %
+                       (name, DISTLIB_PACKAGE))
+                raise ValueError(msg)
+            return WRAPPERS[name]
+
+    # Public API follows
+
+    def make(self, specification, options=None):
+        """
+        Make a script.
+
+        :param specification: The specification, which is either a valid export
+                              entry specification (to make a script from a
+                              callable) or a filename (to make a script by
+                              copying from a source location).
+        :param options: A dictionary of options controlling script generation.
+        :return: A list of all absolute pathnames written to.
+        """
+        filenames = []
+        entry = get_export_entry(specification)
+        if entry is None:
+            self._copy_script(specification, filenames)
+        else:
+            self._make_script(entry, filenames, options=options)
+        return filenames
+
+    def make_multiple(self, specifications, options=None):
+        """
+        Take a list of specifications and make scripts from them,
+        :param specifications: A list of specifications.
+        :return: A list of all absolute pathnames written to,
+        """
+        filenames = []
+        for specification in specifications:
+            filenames.extend(self.make(specification, options))
+        return filenames
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t32.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t32.exe
new file mode 100644
index 00000000..52154f0b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t32.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64-arm.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64-arm.exe
new file mode 100644
index 00000000..e1ab8f8f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64-arm.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64.exe
new file mode 100644
index 00000000..e8bebdba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/t64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/util.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/util.py
new file mode 100644
index 00000000..0d5bd7a8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/util.py
@@ -0,0 +1,1984 @@
+#
+# Copyright (C) 2012-2023 The Python Software Foundation.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+import codecs
+from collections import deque
+import contextlib
+import csv
+from glob import iglob as std_iglob
+import io
+import json
+import logging
+import os
+import py_compile
+import re
+import socket
+try:
+    import ssl
+except ImportError:  # pragma: no cover
+    ssl = None
+import subprocess
+import sys
+import tarfile
+import tempfile
+import textwrap
+
+try:
+    import threading
+except ImportError:  # pragma: no cover
+    import dummy_threading as threading
+import time
+
+from . import DistlibException
+from .compat import (string_types, text_type, shutil, raw_input, StringIO, cache_from_source, urlopen, urljoin, httplib,
+                     xmlrpclib, HTTPHandler, BaseConfigurator, valid_ident, Container, configparser, URLError, ZipFile,
+                     fsdecode, unquote, urlparse)
+
+logger = logging.getLogger(__name__)
+
+#
+# Requirement parsing code as per PEP 508
+#
+
+IDENTIFIER = re.compile(r'^([\w\.-]+)\s*')
+VERSION_IDENTIFIER = re.compile(r'^([\w\.*+-]+)\s*')
+COMPARE_OP = re.compile(r'^(<=?|>=?|={2,3}|[~!]=)\s*')
+MARKER_OP = re.compile(r'^((<=?)|(>=?)|={2,3}|[~!]=|in|not\s+in)\s*')
+OR = re.compile(r'^or\b\s*')
+AND = re.compile(r'^and\b\s*')
+NON_SPACE = re.compile(r'(\S+)\s*')
+STRING_CHUNK = re.compile(r'([\s\w\.{}()*+#:;,/?!~`@$%^&=|<>\[\]-]+)')
+
+
+def parse_marker(marker_string):
+    """
+    Parse a marker string and return a dictionary containing a marker expression.
+
+    The dictionary will contain keys "op", "lhs" and "rhs" for non-terminals in
+    the expression grammar, or strings. A string contained in quotes is to be
+    interpreted as a literal string, and a string not contained in quotes is a
+    variable (such as os_name).
+    """
+
+    def marker_var(remaining):
+        # either identifier, or literal string
+        m = IDENTIFIER.match(remaining)
+        if m:
+            result = m.groups()[0]
+            remaining = remaining[m.end():]
+        elif not remaining:
+            raise SyntaxError('unexpected end of input')
+        else:
+            q = remaining[0]
+            if q not in '\'"':
+                raise SyntaxError('invalid expression: %s' % remaining)
+            oq = '\'"'.replace(q, '')
+            remaining = remaining[1:]
+            parts = [q]
+            while remaining:
+                # either a string chunk, or oq, or q to terminate
+                if remaining[0] == q:
+                    break
+                elif remaining[0] == oq:
+                    parts.append(oq)
+                    remaining = remaining[1:]
+                else:
+                    m = STRING_CHUNK.match(remaining)
+                    if not m:
+                        raise SyntaxError('error in string literal: %s' % remaining)
+                    parts.append(m.groups()[0])
+                    remaining = remaining[m.end():]
+            else:
+                s = ''.join(parts)
+                raise SyntaxError('unterminated string: %s' % s)
+            parts.append(q)
+            result = ''.join(parts)
+            remaining = remaining[1:].lstrip()  # skip past closing quote
+        return result, remaining
+
+    def marker_expr(remaining):
+        if remaining and remaining[0] == '(':
+            result, remaining = marker(remaining[1:].lstrip())
+            if remaining[0] != ')':
+                raise SyntaxError('unterminated parenthesis: %s' % remaining)
+            remaining = remaining[1:].lstrip()
+        else:
+            lhs, remaining = marker_var(remaining)
+            while remaining:
+                m = MARKER_OP.match(remaining)
+                if not m:
+                    break
+                op = m.groups()[0]
+                remaining = remaining[m.end():]
+                rhs, remaining = marker_var(remaining)
+                lhs = {'op': op, 'lhs': lhs, 'rhs': rhs}
+            result = lhs
+        return result, remaining
+
+    def marker_and(remaining):
+        lhs, remaining = marker_expr(remaining)
+        while remaining:
+            m = AND.match(remaining)
+            if not m:
+                break
+            remaining = remaining[m.end():]
+            rhs, remaining = marker_expr(remaining)
+            lhs = {'op': 'and', 'lhs': lhs, 'rhs': rhs}
+        return lhs, remaining
+
+    def marker(remaining):
+        lhs, remaining = marker_and(remaining)
+        while remaining:
+            m = OR.match(remaining)
+            if not m:
+                break
+            remaining = remaining[m.end():]
+            rhs, remaining = marker_and(remaining)
+            lhs = {'op': 'or', 'lhs': lhs, 'rhs': rhs}
+        return lhs, remaining
+
+    return marker(marker_string)
+
+
+def parse_requirement(req):
+    """
+    Parse a requirement passed in as a string. Return a Container
+    whose attributes contain the various parts of the requirement.
+    """
+    remaining = req.strip()
+    if not remaining or remaining.startswith('#'):
+        return None
+    m = IDENTIFIER.match(remaining)
+    if not m:
+        raise SyntaxError('name expected: %s' % remaining)
+    distname = m.groups()[0]
+    remaining = remaining[m.end():]
+    extras = mark_expr = versions = uri = None
+    if remaining and remaining[0] == '[':
+        i = remaining.find(']', 1)
+        if i < 0:
+            raise SyntaxError('unterminated extra: %s' % remaining)
+        s = remaining[1:i]
+        remaining = remaining[i + 1:].lstrip()
+        extras = []
+        while s:
+            m = IDENTIFIER.match(s)
+            if not m:
+                raise SyntaxError('malformed extra: %s' % s)
+            extras.append(m.groups()[0])
+            s = s[m.end():]
+            if not s:
+                break
+            if s[0] != ',':
+                raise SyntaxError('comma expected in extras: %s' % s)
+            s = s[1:].lstrip()
+        if not extras:
+            extras = None
+    if remaining:
+        if remaining[0] == '@':
+            # it's a URI
+            remaining = remaining[1:].lstrip()
+            m = NON_SPACE.match(remaining)
+            if not m:
+                raise SyntaxError('invalid URI: %s' % remaining)
+            uri = m.groups()[0]
+            t = urlparse(uri)
+            # there are issues with Python and URL parsing, so this test
+            # is a bit crude. See bpo-20271, bpo-23505. Python doesn't
+            # always parse invalid URLs correctly - it should raise
+            # exceptions for malformed URLs
+            if not (t.scheme and t.netloc):
+                raise SyntaxError('Invalid URL: %s' % uri)
+            remaining = remaining[m.end():].lstrip()
+        else:
+
+            def get_versions(ver_remaining):
+                """
+                Return a list of operator, version tuples if any are
+                specified, else None.
+                """
+                m = COMPARE_OP.match(ver_remaining)
+                versions = None
+                if m:
+                    versions = []
+                    while True:
+                        op = m.groups()[0]
+                        ver_remaining = ver_remaining[m.end():]
+                        m = VERSION_IDENTIFIER.match(ver_remaining)
+                        if not m:
+                            raise SyntaxError('invalid version: %s' % ver_remaining)
+                        v = m.groups()[0]
+                        versions.append((op, v))
+                        ver_remaining = ver_remaining[m.end():]
+                        if not ver_remaining or ver_remaining[0] != ',':
+                            break
+                        ver_remaining = ver_remaining[1:].lstrip()
+                        # Some packages have a trailing comma which would break things
+                        # See issue #148
+                        if not ver_remaining:
+                            break
+                        m = COMPARE_OP.match(ver_remaining)
+                        if not m:
+                            raise SyntaxError('invalid constraint: %s' % ver_remaining)
+                    if not versions:
+                        versions = None
+                return versions, ver_remaining
+
+            if remaining[0] != '(':
+                versions, remaining = get_versions(remaining)
+            else:
+                i = remaining.find(')', 1)
+                if i < 0:
+                    raise SyntaxError('unterminated parenthesis: %s' % remaining)
+                s = remaining[1:i]
+                remaining = remaining[i + 1:].lstrip()
+                # As a special diversion from PEP 508, allow a version number
+                # a.b.c in parentheses as a synonym for ~= a.b.c (because this
+                # is allowed in earlier PEPs)
+                if COMPARE_OP.match(s):
+                    versions, _ = get_versions(s)
+                else:
+                    m = VERSION_IDENTIFIER.match(s)
+                    if not m:
+                        raise SyntaxError('invalid constraint: %s' % s)
+                    v = m.groups()[0]
+                    s = s[m.end():].lstrip()
+                    if s:
+                        raise SyntaxError('invalid constraint: %s' % s)
+                    versions = [('~=', v)]
+
+    if remaining:
+        if remaining[0] != ';':
+            raise SyntaxError('invalid requirement: %s' % remaining)
+        remaining = remaining[1:].lstrip()
+
+        mark_expr, remaining = parse_marker(remaining)
+
+    if remaining and remaining[0] != '#':
+        raise SyntaxError('unexpected trailing data: %s' % remaining)
+
+    if not versions:
+        rs = distname
+    else:
+        rs = '%s %s' % (distname, ', '.join(['%s %s' % con for con in versions]))
+    return Container(name=distname, extras=extras, constraints=versions, marker=mark_expr, url=uri, requirement=rs)
+
+
+def get_resources_dests(resources_root, rules):
+    """Find destinations for resources files"""
+
+    def get_rel_path(root, path):
+        # normalizes and returns a lstripped-/-separated path
+        root = root.replace(os.path.sep, '/')
+        path = path.replace(os.path.sep, '/')
+        assert path.startswith(root)
+        return path[len(root):].lstrip('/')
+
+    destinations = {}
+    for base, suffix, dest in rules:
+        prefix = os.path.join(resources_root, base)
+        for abs_base in iglob(prefix):
+            abs_glob = os.path.join(abs_base, suffix)
+            for abs_path in iglob(abs_glob):
+                resource_file = get_rel_path(resources_root, abs_path)
+                if dest is None:  # remove the entry if it was here
+                    destinations.pop(resource_file, None)
+                else:
+                    rel_path = get_rel_path(abs_base, abs_path)
+                    rel_dest = dest.replace(os.path.sep, '/').rstrip('/')
+                    destinations[resource_file] = rel_dest + '/' + rel_path
+    return destinations
+
+
+def in_venv():
+    if hasattr(sys, 'real_prefix'):
+        # virtualenv venvs
+        result = True
+    else:
+        # PEP 405 venvs
+        result = sys.prefix != getattr(sys, 'base_prefix', sys.prefix)
+    return result
+
+
+def get_executable():
+    # The __PYVENV_LAUNCHER__ dance is apparently no longer needed, as
+    # changes to the stub launcher mean that sys.executable always points
+    # to the stub on OS X
+    #    if sys.platform == 'darwin' and ('__PYVENV_LAUNCHER__'
+    #                                     in os.environ):
+    #        result =  os.environ['__PYVENV_LAUNCHER__']
+    #    else:
+    #        result = sys.executable
+    #    return result
+    # Avoid normcasing: see issue #143
+    # result = os.path.normcase(sys.executable)
+    result = sys.executable
+    if not isinstance(result, text_type):
+        result = fsdecode(result)
+    return result
+
+
+def proceed(prompt, allowed_chars, error_prompt=None, default=None):
+    p = prompt
+    while True:
+        s = raw_input(p)
+        p = prompt
+        if not s and default:
+            s = default
+        if s:
+            c = s[0].lower()
+            if c in allowed_chars:
+                break
+            if error_prompt:
+                p = '%c: %s\n%s' % (c, error_prompt, prompt)
+    return c
+
+
+def extract_by_key(d, keys):
+    if isinstance(keys, string_types):
+        keys = keys.split()
+    result = {}
+    for key in keys:
+        if key in d:
+            result[key] = d[key]
+    return result
+
+
+def read_exports(stream):
+    if sys.version_info[0] >= 3:
+        # needs to be a text stream
+        stream = codecs.getreader('utf-8')(stream)
+    # Try to load as JSON, falling back on legacy format
+    data = stream.read()
+    stream = StringIO(data)
+    try:
+        jdata = json.load(stream)
+        result = jdata['extensions']['python.exports']['exports']
+        for group, entries in result.items():
+            for k, v in entries.items():
+                s = '%s = %s' % (k, v)
+                entry = get_export_entry(s)
+                assert entry is not None
+                entries[k] = entry
+        return result
+    except Exception:
+        stream.seek(0, 0)
+
+    def read_stream(cp, stream):
+        if hasattr(cp, 'read_file'):
+            cp.read_file(stream)
+        else:
+            cp.readfp(stream)
+
+    cp = configparser.ConfigParser()
+    try:
+        read_stream(cp, stream)
+    except configparser.MissingSectionHeaderError:
+        stream.close()
+        data = textwrap.dedent(data)
+        stream = StringIO(data)
+        read_stream(cp, stream)
+
+    result = {}
+    for key in cp.sections():
+        result[key] = entries = {}
+        for name, value in cp.items(key):
+            s = '%s = %s' % (name, value)
+            entry = get_export_entry(s)
+            assert entry is not None
+            # entry.dist = self
+            entries[name] = entry
+    return result
+
+
+def write_exports(exports, stream):
+    if sys.version_info[0] >= 3:
+        # needs to be a text stream
+        stream = codecs.getwriter('utf-8')(stream)
+    cp = configparser.ConfigParser()
+    for k, v in exports.items():
+        # TODO check k, v for valid values
+        cp.add_section(k)
+        for entry in v.values():
+            if entry.suffix is None:
+                s = entry.prefix
+            else:
+                s = '%s:%s' % (entry.prefix, entry.suffix)
+            if entry.flags:
+                s = '%s [%s]' % (s, ', '.join(entry.flags))
+            cp.set(k, entry.name, s)
+    cp.write(stream)
+
+
+@contextlib.contextmanager
+def tempdir():
+    td = tempfile.mkdtemp()
+    try:
+        yield td
+    finally:
+        shutil.rmtree(td)
+
+
+@contextlib.contextmanager
+def chdir(d):
+    cwd = os.getcwd()
+    try:
+        os.chdir(d)
+        yield
+    finally:
+        os.chdir(cwd)
+
+
+@contextlib.contextmanager
+def socket_timeout(seconds=15):
+    cto = socket.getdefaulttimeout()
+    try:
+        socket.setdefaulttimeout(seconds)
+        yield
+    finally:
+        socket.setdefaulttimeout(cto)
+
+
+class cached_property(object):
+
+    def __init__(self, func):
+        self.func = func
+        # for attr in ('__name__', '__module__', '__doc__'):
+        #     setattr(self, attr, getattr(func, attr, None))
+
+    def __get__(self, obj, cls=None):
+        if obj is None:
+            return self
+        value = self.func(obj)
+        object.__setattr__(obj, self.func.__name__, value)
+        # obj.__dict__[self.func.__name__] = value = self.func(obj)
+        return value
+
+
+def convert_path(pathname):
+    """Return 'pathname' as a name that will work on the native filesystem.
+
+    The path is split on '/' and put back together again using the current
+    directory separator.  Needed because filenames in the setup script are
+    always supplied in Unix style, and have to be converted to the local
+    convention before we can actually use them in the filesystem.  Raises
+    ValueError on non-Unix-ish systems if 'pathname' either starts or
+    ends with a slash.
+    """
+    if os.sep == '/':
+        return pathname
+    if not pathname:
+        return pathname
+    if pathname[0] == '/':
+        raise ValueError("path '%s' cannot be absolute" % pathname)
+    if pathname[-1] == '/':
+        raise ValueError("path '%s' cannot end with '/'" % pathname)
+
+    paths = pathname.split('/')
+    while os.curdir in paths:
+        paths.remove(os.curdir)
+    if not paths:
+        return os.curdir
+    return os.path.join(*paths)
+
+
+class FileOperator(object):
+
+    def __init__(self, dry_run=False):
+        self.dry_run = dry_run
+        self.ensured = set()
+        self._init_record()
+
+    def _init_record(self):
+        self.record = False
+        self.files_written = set()
+        self.dirs_created = set()
+
+    def record_as_written(self, path):
+        if self.record:
+            self.files_written.add(path)
+
+    def newer(self, source, target):
+        """Tell if the target is newer than the source.
+
+        Returns true if 'source' exists and is more recently modified than
+        'target', or if 'source' exists and 'target' doesn't.
+
+        Returns false if both exist and 'target' is the same age or younger
+        than 'source'. Raise PackagingFileError if 'source' does not exist.
+
+        Note that this test is not very accurate: files created in the same
+        second will have the same "age".
+        """
+        if not os.path.exists(source):
+            raise DistlibException("file '%r' does not exist" % os.path.abspath(source))
+        if not os.path.exists(target):
+            return True
+
+        return os.stat(source).st_mtime > os.stat(target).st_mtime
+
+    def copy_file(self, infile, outfile, check=True):
+        """Copy a file respecting dry-run and force flags.
+        """
+        self.ensure_dir(os.path.dirname(outfile))
+        logger.info('Copying %s to %s', infile, outfile)
+        if not self.dry_run:
+            msg = None
+            if check:
+                if os.path.islink(outfile):
+                    msg = '%s is a symlink' % outfile
+                elif os.path.exists(outfile) and not os.path.isfile(outfile):
+                    msg = '%s is a non-regular file' % outfile
+            if msg:
+                raise ValueError(msg + ' which would be overwritten')
+            shutil.copyfile(infile, outfile)
+        self.record_as_written(outfile)
+
+    def copy_stream(self, instream, outfile, encoding=None):
+        assert not os.path.isdir(outfile)
+        self.ensure_dir(os.path.dirname(outfile))
+        logger.info('Copying stream %s to %s', instream, outfile)
+        if not self.dry_run:
+            if encoding is None:
+                outstream = open(outfile, 'wb')
+            else:
+                outstream = codecs.open(outfile, 'w', encoding=encoding)
+            try:
+                shutil.copyfileobj(instream, outstream)
+            finally:
+                outstream.close()
+        self.record_as_written(outfile)
+
+    def write_binary_file(self, path, data):
+        self.ensure_dir(os.path.dirname(path))
+        if not self.dry_run:
+            if os.path.exists(path):
+                os.remove(path)
+            with open(path, 'wb') as f:
+                f.write(data)
+        self.record_as_written(path)
+
+    def write_text_file(self, path, data, encoding):
+        self.write_binary_file(path, data.encode(encoding))
+
+    def set_mode(self, bits, mask, files):
+        if os.name == 'posix' or (os.name == 'java' and os._name == 'posix'):
+            # Set the executable bits (owner, group, and world) on
+            # all the files specified.
+            for f in files:
+                if self.dry_run:
+                    logger.info("changing mode of %s", f)
+                else:
+                    mode = (os.stat(f).st_mode | bits) & mask
+                    logger.info("changing mode of %s to %o", f, mode)
+                    os.chmod(f, mode)
+
+    set_executable_mode = lambda s, f: s.set_mode(0o555, 0o7777, f)
+
+    def ensure_dir(self, path):
+        path = os.path.abspath(path)
+        if path not in self.ensured and not os.path.exists(path):
+            self.ensured.add(path)
+            d, f = os.path.split(path)
+            self.ensure_dir(d)
+            logger.info('Creating %s' % path)
+            if not self.dry_run:
+                os.mkdir(path)
+            if self.record:
+                self.dirs_created.add(path)
+
+    def byte_compile(self, path, optimize=False, force=False, prefix=None, hashed_invalidation=False):
+        dpath = cache_from_source(path, not optimize)
+        logger.info('Byte-compiling %s to %s', path, dpath)
+        if not self.dry_run:
+            if force or self.newer(path, dpath):
+                if not prefix:
+                    diagpath = None
+                else:
+                    assert path.startswith(prefix)
+                    diagpath = path[len(prefix):]
+            compile_kwargs = {}
+            if hashed_invalidation and hasattr(py_compile, 'PycInvalidationMode'):
+                if not isinstance(hashed_invalidation, py_compile.PycInvalidationMode):
+                    hashed_invalidation = py_compile.PycInvalidationMode.CHECKED_HASH
+                compile_kwargs['invalidation_mode'] = hashed_invalidation
+            py_compile.compile(path, dpath, diagpath, True, **compile_kwargs)  # raise error
+        self.record_as_written(dpath)
+        return dpath
+
+    def ensure_removed(self, path):
+        if os.path.exists(path):
+            if os.path.isdir(path) and not os.path.islink(path):
+                logger.debug('Removing directory tree at %s', path)
+                if not self.dry_run:
+                    shutil.rmtree(path)
+                if self.record:
+                    if path in self.dirs_created:
+                        self.dirs_created.remove(path)
+            else:
+                if os.path.islink(path):
+                    s = 'link'
+                else:
+                    s = 'file'
+                logger.debug('Removing %s %s', s, path)
+                if not self.dry_run:
+                    os.remove(path)
+                if self.record:
+                    if path in self.files_written:
+                        self.files_written.remove(path)
+
+    def is_writable(self, path):
+        result = False
+        while not result:
+            if os.path.exists(path):
+                result = os.access(path, os.W_OK)
+                break
+            parent = os.path.dirname(path)
+            if parent == path:
+                break
+            path = parent
+        return result
+
+    def commit(self):
+        """
+        Commit recorded changes, turn off recording, return
+        changes.
+        """
+        assert self.record
+        result = self.files_written, self.dirs_created
+        self._init_record()
+        return result
+
+    def rollback(self):
+        if not self.dry_run:
+            for f in list(self.files_written):
+                if os.path.exists(f):
+                    os.remove(f)
+            # dirs should all be empty now, except perhaps for
+            # __pycache__ subdirs
+            # reverse so that subdirs appear before their parents
+            dirs = sorted(self.dirs_created, reverse=True)
+            for d in dirs:
+                flist = os.listdir(d)
+                if flist:
+                    assert flist == ['__pycache__']
+                    sd = os.path.join(d, flist[0])
+                    os.rmdir(sd)
+                os.rmdir(d)  # should fail if non-empty
+        self._init_record()
+
+
+def resolve(module_name, dotted_path):
+    if module_name in sys.modules:
+        mod = sys.modules[module_name]
+    else:
+        mod = __import__(module_name)
+    if dotted_path is None:
+        result = mod
+    else:
+        parts = dotted_path.split('.')
+        result = getattr(mod, parts.pop(0))
+        for p in parts:
+            result = getattr(result, p)
+    return result
+
+
+class ExportEntry(object):
+
+    def __init__(self, name, prefix, suffix, flags):
+        self.name = name
+        self.prefix = prefix
+        self.suffix = suffix
+        self.flags = flags
+
+    @cached_property
+    def value(self):
+        return resolve(self.prefix, self.suffix)
+
+    def __repr__(self):  # pragma: no cover
+        return '<ExportEntry %s = %s:%s %s>' % (self.name, self.prefix, self.suffix, self.flags)
+
+    def __eq__(self, other):
+        if not isinstance(other, ExportEntry):
+            result = False
+        else:
+            result = (self.name == other.name and self.prefix == other.prefix and self.suffix == other.suffix and
+                      self.flags == other.flags)
+        return result
+
+    __hash__ = object.__hash__
+
+
+ENTRY_RE = re.compile(
+    r'''(?P<name>([^\[]\S*))
+                      \s*=\s*(?P<callable>(\w+)([:\.]\w+)*)
+                      \s*(\[\s*(?P<flags>[\w-]+(=\w+)?(,\s*\w+(=\w+)?)*)\s*\])?
+                      ''', re.VERBOSE)
+
+
+def get_export_entry(specification):
+    m = ENTRY_RE.search(specification)
+    if not m:
+        result = None
+        if '[' in specification or ']' in specification:
+            raise DistlibException("Invalid specification "
+                                   "'%s'" % specification)
+    else:
+        d = m.groupdict()
+        name = d['name']
+        path = d['callable']
+        colons = path.count(':')
+        if colons == 0:
+            prefix, suffix = path, None
+        else:
+            if colons != 1:
+                raise DistlibException("Invalid specification "
+                                       "'%s'" % specification)
+            prefix, suffix = path.split(':')
+        flags = d['flags']
+        if flags is None:
+            if '[' in specification or ']' in specification:
+                raise DistlibException("Invalid specification "
+                                       "'%s'" % specification)
+            flags = []
+        else:
+            flags = [f.strip() for f in flags.split(',')]
+        result = ExportEntry(name, prefix, suffix, flags)
+    return result
+
+
+def get_cache_base(suffix=None):
+    """
+    Return the default base location for distlib caches. If the directory does
+    not exist, it is created. Use the suffix provided for the base directory,
+    and default to '.distlib' if it isn't provided.
+
+    On Windows, if LOCALAPPDATA is defined in the environment, then it is
+    assumed to be a directory, and will be the parent directory of the result.
+    On POSIX, and on Windows if LOCALAPPDATA is not defined, the user's home
+    directory - using os.expanduser('~') - will be the parent directory of
+    the result.
+
+    The result is just the directory '.distlib' in the parent directory as
+    determined above, or with the name specified with ``suffix``.
+    """
+    if suffix is None:
+        suffix = '.distlib'
+    if os.name == 'nt' and 'LOCALAPPDATA' in os.environ:
+        result = os.path.expandvars('$localappdata')
+    else:
+        # Assume posix, or old Windows
+        result = os.path.expanduser('~')
+    # we use 'isdir' instead of 'exists', because we want to
+    # fail if there's a file with that name
+    if os.path.isdir(result):
+        usable = os.access(result, os.W_OK)
+        if not usable:
+            logger.warning('Directory exists but is not writable: %s', result)
+    else:
+        try:
+            os.makedirs(result)
+            usable = True
+        except OSError:
+            logger.warning('Unable to create %s', result, exc_info=True)
+            usable = False
+    if not usable:
+        result = tempfile.mkdtemp()
+        logger.warning('Default location unusable, using %s', result)
+    return os.path.join(result, suffix)
+
+
+def path_to_cache_dir(path, use_abspath=True):
+    """
+    Convert an absolute path to a directory name for use in a cache.
+
+    The algorithm used is:
+
+    #. On Windows, any ``':'`` in the drive is replaced with ``'---'``.
+    #. Any occurrence of ``os.sep`` is replaced with ``'--'``.
+    #. ``'.cache'`` is appended.
+    """
+    d, p = os.path.splitdrive(os.path.abspath(path) if use_abspath else path)
+    if d:
+        d = d.replace(':', '---')
+    p = p.replace(os.sep, '--')
+    return d + p + '.cache'
+
+
+def ensure_slash(s):
+    if not s.endswith('/'):
+        return s + '/'
+    return s
+
+
+def parse_credentials(netloc):
+    username = password = None
+    if '@' in netloc:
+        prefix, netloc = netloc.rsplit('@', 1)
+        if ':' not in prefix:
+            username = prefix
+        else:
+            username, password = prefix.split(':', 1)
+    if username:
+        username = unquote(username)
+    if password:
+        password = unquote(password)
+    return username, password, netloc
+
+
+def get_process_umask():
+    result = os.umask(0o22)
+    os.umask(result)
+    return result
+
+
+def is_string_sequence(seq):
+    result = True
+    i = None
+    for i, s in enumerate(seq):
+        if not isinstance(s, string_types):
+            result = False
+            break
+    assert i is not None
+    return result
+
+
+PROJECT_NAME_AND_VERSION = re.compile('([a-z0-9_]+([.-][a-z_][a-z0-9_]*)*)-'
+                                      '([a-z0-9_.+-]+)', re.I)
+PYTHON_VERSION = re.compile(r'-py(\d\.?\d?)')
+
+
+def split_filename(filename, project_name=None):
+    """
+    Extract name, version, python version from a filename (no extension)
+
+    Return name, version, pyver or None
+    """
+    result = None
+    pyver = None
+    filename = unquote(filename).replace(' ', '-')
+    m = PYTHON_VERSION.search(filename)
+    if m:
+        pyver = m.group(1)
+        filename = filename[:m.start()]
+    if project_name and len(filename) > len(project_name) + 1:
+        m = re.match(re.escape(project_name) + r'\b', filename)
+        if m:
+            n = m.end()
+            result = filename[:n], filename[n + 1:], pyver
+    if result is None:
+        m = PROJECT_NAME_AND_VERSION.match(filename)
+        if m:
+            result = m.group(1), m.group(3), pyver
+    return result
+
+
+# Allow spaces in name because of legacy dists like "Twisted Core"
+NAME_VERSION_RE = re.compile(r'(?P<name>[\w .-]+)\s*'
+                             r'\(\s*(?P<ver>[^\s)]+)\)$')
+
+
+def parse_name_and_version(p):
+    """
+    A utility method used to get name and version from a string.
+
+    From e.g. a Provides-Dist value.
+
+    :param p: A value in a form 'foo (1.0)'
+    :return: The name and version as a tuple.
+    """
+    m = NAME_VERSION_RE.match(p)
+    if not m:
+        raise DistlibException('Ill-formed name/version string: \'%s\'' % p)
+    d = m.groupdict()
+    return d['name'].strip().lower(), d['ver']
+
+
+def get_extras(requested, available):
+    result = set()
+    requested = set(requested or [])
+    available = set(available or [])
+    if '*' in requested:
+        requested.remove('*')
+        result |= available
+    for r in requested:
+        if r == '-':
+            result.add(r)
+        elif r.startswith('-'):
+            unwanted = r[1:]
+            if unwanted not in available:
+                logger.warning('undeclared extra: %s' % unwanted)
+            if unwanted in result:
+                result.remove(unwanted)
+        else:
+            if r not in available:
+                logger.warning('undeclared extra: %s' % r)
+            result.add(r)
+    return result
+
+
+#
+# Extended metadata functionality
+#
+
+
+def _get_external_data(url):
+    result = {}
+    try:
+        # urlopen might fail if it runs into redirections,
+        # because of Python issue #13696. Fixed in locators
+        # using a custom redirect handler.
+        resp = urlopen(url)
+        headers = resp.info()
+        ct = headers.get('Content-Type')
+        if not ct.startswith('application/json'):
+            logger.debug('Unexpected response for JSON request: %s', ct)
+        else:
+            reader = codecs.getreader('utf-8')(resp)
+            # data = reader.read().decode('utf-8')
+            # result = json.loads(data)
+            result = json.load(reader)
+    except Exception as e:
+        logger.exception('Failed to get external data for %s: %s', url, e)
+    return result
+
+
+_external_data_base_url = 'https://www.red-dove.com/pypi/projects/'
+
+
+def get_project_data(name):
+    url = '%s/%s/project.json' % (name[0].upper(), name)
+    url = urljoin(_external_data_base_url, url)
+    result = _get_external_data(url)
+    return result
+
+
+def get_package_data(name, version):
+    url = '%s/%s/package-%s.json' % (name[0].upper(), name, version)
+    url = urljoin(_external_data_base_url, url)
+    return _get_external_data(url)
+
+
+class Cache(object):
+    """
+    A class implementing a cache for resources that need to live in the file system
+    e.g. shared libraries. This class was moved from resources to here because it
+    could be used by other modules, e.g. the wheel module.
+    """
+
+    def __init__(self, base):
+        """
+        Initialise an instance.
+
+        :param base: The base directory where the cache should be located.
+        """
+        # we use 'isdir' instead of 'exists', because we want to
+        # fail if there's a file with that name
+        if not os.path.isdir(base):  # pragma: no cover
+            os.makedirs(base)
+        if (os.stat(base).st_mode & 0o77) != 0:
+            logger.warning('Directory \'%s\' is not private', base)
+        self.base = os.path.abspath(os.path.normpath(base))
+
+    def prefix_to_dir(self, prefix, use_abspath=True):
+        """
+        Converts a resource prefix to a directory name in the cache.
+        """
+        return path_to_cache_dir(prefix, use_abspath=use_abspath)
+
+    def clear(self):
+        """
+        Clear the cache.
+        """
+        not_removed = []
+        for fn in os.listdir(self.base):
+            fn = os.path.join(self.base, fn)
+            try:
+                if os.path.islink(fn) or os.path.isfile(fn):
+                    os.remove(fn)
+                elif os.path.isdir(fn):
+                    shutil.rmtree(fn)
+            except Exception:
+                not_removed.append(fn)
+        return not_removed
+
+
+class EventMixin(object):
+    """
+    A very simple publish/subscribe system.
+    """
+
+    def __init__(self):
+        self._subscribers = {}
+
+    def add(self, event, subscriber, append=True):
+        """
+        Add a subscriber for an event.
+
+        :param event: The name of an event.
+        :param subscriber: The subscriber to be added (and called when the
+                           event is published).
+        :param append: Whether to append or prepend the subscriber to an
+                       existing subscriber list for the event.
+        """
+        subs = self._subscribers
+        if event not in subs:
+            subs[event] = deque([subscriber])
+        else:
+            sq = subs[event]
+            if append:
+                sq.append(subscriber)
+            else:
+                sq.appendleft(subscriber)
+
+    def remove(self, event, subscriber):
+        """
+        Remove a subscriber for an event.
+
+        :param event: The name of an event.
+        :param subscriber: The subscriber to be removed.
+        """
+        subs = self._subscribers
+        if event not in subs:
+            raise ValueError('No subscribers: %r' % event)
+        subs[event].remove(subscriber)
+
+    def get_subscribers(self, event):
+        """
+        Return an iterator for the subscribers for an event.
+        :param event: The event to return subscribers for.
+        """
+        return iter(self._subscribers.get(event, ()))
+
+    def publish(self, event, *args, **kwargs):
+        """
+        Publish a event and return a list of values returned by its
+        subscribers.
+
+        :param event: The event to publish.
+        :param args: The positional arguments to pass to the event's
+                     subscribers.
+        :param kwargs: The keyword arguments to pass to the event's
+                       subscribers.
+        """
+        result = []
+        for subscriber in self.get_subscribers(event):
+            try:
+                value = subscriber(event, *args, **kwargs)
+            except Exception:
+                logger.exception('Exception during event publication')
+                value = None
+            result.append(value)
+        logger.debug('publish %s: args = %s, kwargs = %s, result = %s', event, args, kwargs, result)
+        return result
+
+
+#
+# Simple sequencing
+#
+class Sequencer(object):
+
+    def __init__(self):
+        self._preds = {}
+        self._succs = {}
+        self._nodes = set()  # nodes with no preds/succs
+
+    def add_node(self, node):
+        self._nodes.add(node)
+
+    def remove_node(self, node, edges=False):
+        if node in self._nodes:
+            self._nodes.remove(node)
+        if edges:
+            for p in set(self._preds.get(node, ())):
+                self.remove(p, node)
+            for s in set(self._succs.get(node, ())):
+                self.remove(node, s)
+            # Remove empties
+            for k, v in list(self._preds.items()):
+                if not v:
+                    del self._preds[k]
+            for k, v in list(self._succs.items()):
+                if not v:
+                    del self._succs[k]
+
+    def add(self, pred, succ):
+        assert pred != succ
+        self._preds.setdefault(succ, set()).add(pred)
+        self._succs.setdefault(pred, set()).add(succ)
+
+    def remove(self, pred, succ):
+        assert pred != succ
+        try:
+            preds = self._preds[succ]
+            succs = self._succs[pred]
+        except KeyError:  # pragma: no cover
+            raise ValueError('%r not a successor of anything' % succ)
+        try:
+            preds.remove(pred)
+            succs.remove(succ)
+        except KeyError:  # pragma: no cover
+            raise ValueError('%r not a successor of %r' % (succ, pred))
+
+    def is_step(self, step):
+        return (step in self._preds or step in self._succs or step in self._nodes)
+
+    def get_steps(self, final):
+        if not self.is_step(final):
+            raise ValueError('Unknown: %r' % final)
+        result = []
+        todo = []
+        seen = set()
+        todo.append(final)
+        while todo:
+            step = todo.pop(0)
+            if step in seen:
+                # if a step was already seen,
+                # move it to the end (so it will appear earlier
+                # when reversed on return) ... but not for the
+                # final step, as that would be confusing for
+                # users
+                if step != final:
+                    result.remove(step)
+                    result.append(step)
+            else:
+                seen.add(step)
+                result.append(step)
+                preds = self._preds.get(step, ())
+                todo.extend(preds)
+        return reversed(result)
+
+    @property
+    def strong_connections(self):
+        # http://en.wikipedia.org/wiki/Tarjan%27s_strongly_connected_components_algorithm
+        index_counter = [0]
+        stack = []
+        lowlinks = {}
+        index = {}
+        result = []
+
+        graph = self._succs
+
+        def strongconnect(node):
+            # set the depth index for this node to the smallest unused index
+            index[node] = index_counter[0]
+            lowlinks[node] = index_counter[0]
+            index_counter[0] += 1
+            stack.append(node)
+
+            # Consider successors
+            try:
+                successors = graph[node]
+            except Exception:
+                successors = []
+            for successor in successors:
+                if successor not in lowlinks:
+                    # Successor has not yet been visited
+                    strongconnect(successor)
+                    lowlinks[node] = min(lowlinks[node], lowlinks[successor])
+                elif successor in stack:
+                    # the successor is in the stack and hence in the current
+                    # strongly connected component (SCC)
+                    lowlinks[node] = min(lowlinks[node], index[successor])
+
+            # If `node` is a root node, pop the stack and generate an SCC
+            if lowlinks[node] == index[node]:
+                connected_component = []
+
+                while True:
+                    successor = stack.pop()
+                    connected_component.append(successor)
+                    if successor == node:
+                        break
+                component = tuple(connected_component)
+                # storing the result
+                result.append(component)
+
+        for node in graph:
+            if node not in lowlinks:
+                strongconnect(node)
+
+        return result
+
+    @property
+    def dot(self):
+        result = ['digraph G {']
+        for succ in self._preds:
+            preds = self._preds[succ]
+            for pred in preds:
+                result.append('  %s -> %s;' % (pred, succ))
+        for node in self._nodes:
+            result.append('  %s;' % node)
+        result.append('}')
+        return '\n'.join(result)
+
+
+#
+# Unarchiving functionality for zip, tar, tgz, tbz, whl
+#
+
+ARCHIVE_EXTENSIONS = ('.tar.gz', '.tar.bz2', '.tar', '.zip', '.tgz', '.tbz', '.whl')
+
+
+def unarchive(archive_filename, dest_dir, format=None, check=True):
+
+    def check_path(path):
+        if not isinstance(path, text_type):
+            path = path.decode('utf-8')
+        p = os.path.abspath(os.path.join(dest_dir, path))
+        if not p.startswith(dest_dir) or p[plen] != os.sep:
+            raise ValueError('path outside destination: %r' % p)
+
+    dest_dir = os.path.abspath(dest_dir)
+    plen = len(dest_dir)
+    archive = None
+    if format is None:
+        if archive_filename.endswith(('.zip', '.whl')):
+            format = 'zip'
+        elif archive_filename.endswith(('.tar.gz', '.tgz')):
+            format = 'tgz'
+            mode = 'r:gz'
+        elif archive_filename.endswith(('.tar.bz2', '.tbz')):
+            format = 'tbz'
+            mode = 'r:bz2'
+        elif archive_filename.endswith('.tar'):
+            format = 'tar'
+            mode = 'r'
+        else:  # pragma: no cover
+            raise ValueError('Unknown format for %r' % archive_filename)
+    try:
+        if format == 'zip':
+            archive = ZipFile(archive_filename, 'r')
+            if check:
+                names = archive.namelist()
+                for name in names:
+                    check_path(name)
+        else:
+            archive = tarfile.open(archive_filename, mode)
+            if check:
+                names = archive.getnames()
+                for name in names:
+                    check_path(name)
+        if format != 'zip' and sys.version_info[0] < 3:
+            # See Python issue 17153. If the dest path contains Unicode,
+            # tarfile extraction fails on Python 2.x if a member path name
+            # contains non-ASCII characters - it leads to an implicit
+            # bytes -> unicode conversion using ASCII to decode.
+            for tarinfo in archive.getmembers():
+                if not isinstance(tarinfo.name, text_type):
+                    tarinfo.name = tarinfo.name.decode('utf-8')
+
+        # Limit extraction of dangerous items, if this Python
+        # allows it easily. If not, just trust the input.
+        # See: https://docs.python.org/3/library/tarfile.html#extraction-filters
+        def extraction_filter(member, path):
+            """Run tarfile.tar_filter, but raise the expected ValueError"""
+            # This is only called if the current Python has tarfile filters
+            try:
+                return tarfile.tar_filter(member, path)
+            except tarfile.FilterError as exc:
+                raise ValueError(str(exc))
+
+        archive.extraction_filter = extraction_filter
+
+        archive.extractall(dest_dir)
+
+    finally:
+        if archive:
+            archive.close()
+
+
+def zip_dir(directory):
+    """zip a directory tree into a BytesIO object"""
+    result = io.BytesIO()
+    dlen = len(directory)
+    with ZipFile(result, "w") as zf:
+        for root, dirs, files in os.walk(directory):
+            for name in files:
+                full = os.path.join(root, name)
+                rel = root[dlen:]
+                dest = os.path.join(rel, name)
+                zf.write(full, dest)
+    return result
+
+
+#
+# Simple progress bar
+#
+
+UNITS = ('', 'K', 'M', 'G', 'T', 'P')
+
+
+class Progress(object):
+    unknown = 'UNKNOWN'
+
+    def __init__(self, minval=0, maxval=100):
+        assert maxval is None or maxval >= minval
+        self.min = self.cur = minval
+        self.max = maxval
+        self.started = None
+        self.elapsed = 0
+        self.done = False
+
+    def update(self, curval):
+        assert self.min <= curval
+        assert self.max is None or curval <= self.max
+        self.cur = curval
+        now = time.time()
+        if self.started is None:
+            self.started = now
+        else:
+            self.elapsed = now - self.started
+
+    def increment(self, incr):
+        assert incr >= 0
+        self.update(self.cur + incr)
+
+    def start(self):
+        self.update(self.min)
+        return self
+
+    def stop(self):
+        if self.max is not None:
+            self.update(self.max)
+        self.done = True
+
+    @property
+    def maximum(self):
+        return self.unknown if self.max is None else self.max
+
+    @property
+    def percentage(self):
+        if self.done:
+            result = '100 %'
+        elif self.max is None:
+            result = ' ?? %'
+        else:
+            v = 100.0 * (self.cur - self.min) / (self.max - self.min)
+            result = '%3d %%' % v
+        return result
+
+    def format_duration(self, duration):
+        if (duration <= 0) and self.max is None or self.cur == self.min:
+            result = '??:??:??'
+        # elif duration < 1:
+        #     result = '--:--:--'
+        else:
+            result = time.strftime('%H:%M:%S', time.gmtime(duration))
+        return result
+
+    @property
+    def ETA(self):
+        if self.done:
+            prefix = 'Done'
+            t = self.elapsed
+            # import pdb; pdb.set_trace()
+        else:
+            prefix = 'ETA '
+            if self.max is None:
+                t = -1
+            elif self.elapsed == 0 or (self.cur == self.min):
+                t = 0
+            else:
+                # import pdb; pdb.set_trace()
+                t = float(self.max - self.min)
+                t /= self.cur - self.min
+                t = (t - 1) * self.elapsed
+        return '%s: %s' % (prefix, self.format_duration(t))
+
+    @property
+    def speed(self):
+        if self.elapsed == 0:
+            result = 0.0
+        else:
+            result = (self.cur - self.min) / self.elapsed
+        for unit in UNITS:
+            if result < 1000:
+                break
+            result /= 1000.0
+        return '%d %sB/s' % (result, unit)
+
+
+#
+# Glob functionality
+#
+
+RICH_GLOB = re.compile(r'\{([^}]*)\}')
+_CHECK_RECURSIVE_GLOB = re.compile(r'[^/\\,{]\*\*|\*\*[^/\\,}]')
+_CHECK_MISMATCH_SET = re.compile(r'^[^{]*\}|\{[^}]*$')
+
+
+def iglob(path_glob):
+    """Extended globbing function that supports ** and {opt1,opt2,opt3}."""
+    if _CHECK_RECURSIVE_GLOB.search(path_glob):
+        msg = """invalid glob %r: recursive glob "**" must be used alone"""
+        raise ValueError(msg % path_glob)
+    if _CHECK_MISMATCH_SET.search(path_glob):
+        msg = """invalid glob %r: mismatching set marker '{' or '}'"""
+        raise ValueError(msg % path_glob)
+    return _iglob(path_glob)
+
+
+def _iglob(path_glob):
+    rich_path_glob = RICH_GLOB.split(path_glob, 1)
+    if len(rich_path_glob) > 1:
+        assert len(rich_path_glob) == 3, rich_path_glob
+        prefix, set, suffix = rich_path_glob
+        for item in set.split(','):
+            for path in _iglob(''.join((prefix, item, suffix))):
+                yield path
+    else:
+        if '**' not in path_glob:
+            for item in std_iglob(path_glob):
+                yield item
+        else:
+            prefix, radical = path_glob.split('**', 1)
+            if prefix == '':
+                prefix = '.'
+            if radical == '':
+                radical = '*'
+            else:
+                # we support both
+                radical = radical.lstrip('/')
+                radical = radical.lstrip('\\')
+            for path, dir, files in os.walk(prefix):
+                path = os.path.normpath(path)
+                for fn in _iglob(os.path.join(path, radical)):
+                    yield fn
+
+
+if ssl:
+    from .compat import (HTTPSHandler as BaseHTTPSHandler, match_hostname, CertificateError)
+
+    #
+    # HTTPSConnection which verifies certificates/matches domains
+    #
+
+    class HTTPSConnection(httplib.HTTPSConnection):
+        ca_certs = None  # set this to the path to the certs file (.pem)
+        check_domain = True  # only used if ca_certs is not None
+
+        # noinspection PyPropertyAccess
+        def connect(self):
+            sock = socket.create_connection((self.host, self.port), self.timeout)
+            if getattr(self, '_tunnel_host', False):
+                self.sock = sock
+                self._tunnel()
+
+            context = ssl.SSLContext(ssl.PROTOCOL_SSLv23)
+            if hasattr(ssl, 'OP_NO_SSLv2'):
+                context.options |= ssl.OP_NO_SSLv2
+            if getattr(self, 'cert_file', None):
+                context.load_cert_chain(self.cert_file, self.key_file)
+            kwargs = {}
+            if self.ca_certs:
+                context.verify_mode = ssl.CERT_REQUIRED
+                context.load_verify_locations(cafile=self.ca_certs)
+                if getattr(ssl, 'HAS_SNI', False):
+                    kwargs['server_hostname'] = self.host
+
+            self.sock = context.wrap_socket(sock, **kwargs)
+            if self.ca_certs and self.check_domain:
+                try:
+                    match_hostname(self.sock.getpeercert(), self.host)
+                    logger.debug('Host verified: %s', self.host)
+                except CertificateError:  # pragma: no cover
+                    self.sock.shutdown(socket.SHUT_RDWR)
+                    self.sock.close()
+                    raise
+
+    class HTTPSHandler(BaseHTTPSHandler):
+
+        def __init__(self, ca_certs, check_domain=True):
+            BaseHTTPSHandler.__init__(self)
+            self.ca_certs = ca_certs
+            self.check_domain = check_domain
+
+        def _conn_maker(self, *args, **kwargs):
+            """
+            This is called to create a connection instance. Normally you'd
+            pass a connection class to do_open, but it doesn't actually check for
+            a class, and just expects a callable. As long as we behave just as a
+            constructor would have, we should be OK. If it ever changes so that
+            we *must* pass a class, we'll create an UnsafeHTTPSConnection class
+            which just sets check_domain to False in the class definition, and
+            choose which one to pass to do_open.
+            """
+            result = HTTPSConnection(*args, **kwargs)
+            if self.ca_certs:
+                result.ca_certs = self.ca_certs
+                result.check_domain = self.check_domain
+            return result
+
+        def https_open(self, req):
+            try:
+                return self.do_open(self._conn_maker, req)
+            except URLError as e:
+                if 'certificate verify failed' in str(e.reason):
+                    raise CertificateError('Unable to verify server certificate '
+                                           'for %s' % req.host)
+                else:
+                    raise
+
+    #
+    # To prevent against mixing HTTP traffic with HTTPS (examples: A Man-In-The-
+    # Middle proxy using HTTP listens on port 443, or an index mistakenly serves
+    # HTML containing a http://xyz link when it should be https://xyz),
+    # you can use the following handler class, which does not allow HTTP traffic.
+    #
+    # It works by inheriting from HTTPHandler - so build_opener won't add a
+    # handler for HTTP itself.
+    #
+    class HTTPSOnlyHandler(HTTPSHandler, HTTPHandler):
+
+        def http_open(self, req):
+            raise URLError('Unexpected HTTP request on what should be a secure '
+                           'connection: %s' % req)
+
+
+#
+# XML-RPC with timeouts
+#
+class Transport(xmlrpclib.Transport):
+
+    def __init__(self, timeout, use_datetime=0):
+        self.timeout = timeout
+        xmlrpclib.Transport.__init__(self, use_datetime)
+
+    def make_connection(self, host):
+        h, eh, x509 = self.get_host_info(host)
+        if not self._connection or host != self._connection[0]:
+            self._extra_headers = eh
+            self._connection = host, httplib.HTTPConnection(h)
+        return self._connection[1]
+
+
+if ssl:
+
+    class SafeTransport(xmlrpclib.SafeTransport):
+
+        def __init__(self, timeout, use_datetime=0):
+            self.timeout = timeout
+            xmlrpclib.SafeTransport.__init__(self, use_datetime)
+
+        def make_connection(self, host):
+            h, eh, kwargs = self.get_host_info(host)
+            if not kwargs:
+                kwargs = {}
+            kwargs['timeout'] = self.timeout
+            if not self._connection or host != self._connection[0]:
+                self._extra_headers = eh
+                self._connection = host, httplib.HTTPSConnection(h, None, **kwargs)
+            return self._connection[1]
+
+
+class ServerProxy(xmlrpclib.ServerProxy):
+
+    def __init__(self, uri, **kwargs):
+        self.timeout = timeout = kwargs.pop('timeout', None)
+        # The above classes only come into play if a timeout
+        # is specified
+        if timeout is not None:
+            # scheme = splittype(uri)  # deprecated as of Python 3.8
+            scheme = urlparse(uri)[0]
+            use_datetime = kwargs.get('use_datetime', 0)
+            if scheme == 'https':
+                tcls = SafeTransport
+            else:
+                tcls = Transport
+            kwargs['transport'] = t = tcls(timeout, use_datetime=use_datetime)
+            self.transport = t
+        xmlrpclib.ServerProxy.__init__(self, uri, **kwargs)
+
+
+#
+# CSV functionality. This is provided because on 2.x, the csv module can't
+# handle Unicode. However, we need to deal with Unicode in e.g. RECORD files.
+#
+
+
+def _csv_open(fn, mode, **kwargs):
+    if sys.version_info[0] < 3:
+        mode += 'b'
+    else:
+        kwargs['newline'] = ''
+        # Python 3 determines encoding from locale. Force 'utf-8'
+        # file encoding to match other forced utf-8 encoding
+        kwargs['encoding'] = 'utf-8'
+    return open(fn, mode, **kwargs)
+
+
+class CSVBase(object):
+    defaults = {
+        'delimiter': str(','),  # The strs are used because we need native
+        'quotechar': str('"'),  # str in the csv API (2.x won't take
+        'lineterminator': str('\n')  # Unicode)
+    }
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *exc_info):
+        self.stream.close()
+
+
+class CSVReader(CSVBase):
+
+    def __init__(self, **kwargs):
+        if 'stream' in kwargs:
+            stream = kwargs['stream']
+            if sys.version_info[0] >= 3:
+                # needs to be a text stream
+                stream = codecs.getreader('utf-8')(stream)
+            self.stream = stream
+        else:
+            self.stream = _csv_open(kwargs['path'], 'r')
+        self.reader = csv.reader(self.stream, **self.defaults)
+
+    def __iter__(self):
+        return self
+
+    def next(self):
+        result = next(self.reader)
+        if sys.version_info[0] < 3:
+            for i, item in enumerate(result):
+                if not isinstance(item, text_type):
+                    result[i] = item.decode('utf-8')
+        return result
+
+    __next__ = next
+
+
+class CSVWriter(CSVBase):
+
+    def __init__(self, fn, **kwargs):
+        self.stream = _csv_open(fn, 'w')
+        self.writer = csv.writer(self.stream, **self.defaults)
+
+    def writerow(self, row):
+        if sys.version_info[0] < 3:
+            r = []
+            for item in row:
+                if isinstance(item, text_type):
+                    item = item.encode('utf-8')
+                r.append(item)
+            row = r
+        self.writer.writerow(row)
+
+
+#
+#   Configurator functionality
+#
+
+
+class Configurator(BaseConfigurator):
+
+    value_converters = dict(BaseConfigurator.value_converters)
+    value_converters['inc'] = 'inc_convert'
+
+    def __init__(self, config, base=None):
+        super(Configurator, self).__init__(config)
+        self.base = base or os.getcwd()
+
+    def configure_custom(self, config):
+
+        def convert(o):
+            if isinstance(o, (list, tuple)):
+                result = type(o)([convert(i) for i in o])
+            elif isinstance(o, dict):
+                if '()' in o:
+                    result = self.configure_custom(o)
+                else:
+                    result = {}
+                    for k in o:
+                        result[k] = convert(o[k])
+            else:
+                result = self.convert(o)
+            return result
+
+        c = config.pop('()')
+        if not callable(c):
+            c = self.resolve(c)
+        props = config.pop('.', None)
+        # Check for valid identifiers
+        args = config.pop('[]', ())
+        if args:
+            args = tuple([convert(o) for o in args])
+        items = [(k, convert(config[k])) for k in config if valid_ident(k)]
+        kwargs = dict(items)
+        result = c(*args, **kwargs)
+        if props:
+            for n, v in props.items():
+                setattr(result, n, convert(v))
+        return result
+
+    def __getitem__(self, key):
+        result = self.config[key]
+        if isinstance(result, dict) and '()' in result:
+            self.config[key] = result = self.configure_custom(result)
+        return result
+
+    def inc_convert(self, value):
+        """Default converter for the inc:// protocol."""
+        if not os.path.isabs(value):
+            value = os.path.join(self.base, value)
+        with codecs.open(value, 'r', encoding='utf-8') as f:
+            result = json.load(f)
+        return result
+
+
+class SubprocessMixin(object):
+    """
+    Mixin for running subprocesses and capturing their output
+    """
+
+    def __init__(self, verbose=False, progress=None):
+        self.verbose = verbose
+        self.progress = progress
+
+    def reader(self, stream, context):
+        """
+        Read lines from a subprocess' output stream and either pass to a progress
+        callable (if specified) or write progress information to sys.stderr.
+        """
+        progress = self.progress
+        verbose = self.verbose
+        while True:
+            s = stream.readline()
+            if not s:
+                break
+            if progress is not None:
+                progress(s, context)
+            else:
+                if not verbose:
+                    sys.stderr.write('.')
+                else:
+                    sys.stderr.write(s.decode('utf-8'))
+                sys.stderr.flush()
+        stream.close()
+
+    def run_command(self, cmd, **kwargs):
+        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, **kwargs)
+        t1 = threading.Thread(target=self.reader, args=(p.stdout, 'stdout'))
+        t1.start()
+        t2 = threading.Thread(target=self.reader, args=(p.stderr, 'stderr'))
+        t2.start()
+        p.wait()
+        t1.join()
+        t2.join()
+        if self.progress is not None:
+            self.progress('done.', 'main')
+        elif self.verbose:
+            sys.stderr.write('done.\n')
+        return p
+
+
+def normalize_name(name):
+    """Normalize a python package name a la PEP 503"""
+    # https://www.python.org/dev/peps/pep-0503/#normalized-names
+    return re.sub('[-_.]+', '-', name).lower()
+
+
+# def _get_pypirc_command():
+# """
+# Get the distutils command for interacting with PyPI configurations.
+# :return: the command.
+# """
+# from distutils.core import Distribution
+# from distutils.config import PyPIRCCommand
+# d = Distribution()
+# return PyPIRCCommand(d)
+
+
+class PyPIRCFile(object):
+
+    DEFAULT_REPOSITORY = 'https://upload.pypi.org/legacy/'
+    DEFAULT_REALM = 'pypi'
+
+    def __init__(self, fn=None, url=None):
+        if fn is None:
+            fn = os.path.join(os.path.expanduser('~'), '.pypirc')
+        self.filename = fn
+        self.url = url
+
+    def read(self):
+        result = {}
+
+        if os.path.exists(self.filename):
+            repository = self.url or self.DEFAULT_REPOSITORY
+
+            config = configparser.RawConfigParser()
+            config.read(self.filename)
+            sections = config.sections()
+            if 'distutils' in sections:
+                # let's get the list of servers
+                index_servers = config.get('distutils', 'index-servers')
+                _servers = [server.strip() for server in index_servers.split('\n') if server.strip() != '']
+                if _servers == []:
+                    # nothing set, let's try to get the default pypi
+                    if 'pypi' in sections:
+                        _servers = ['pypi']
+                else:
+                    for server in _servers:
+                        result = {'server': server}
+                        result['username'] = config.get(server, 'username')
+
+                        # optional params
+                        for key, default in (('repository', self.DEFAULT_REPOSITORY), ('realm', self.DEFAULT_REALM),
+                                             ('password', None)):
+                            if config.has_option(server, key):
+                                result[key] = config.get(server, key)
+                            else:
+                                result[key] = default
+
+                        # work around people having "repository" for the "pypi"
+                        # section of their config set to the HTTP (rather than
+                        # HTTPS) URL
+                        if (server == 'pypi' and repository in (self.DEFAULT_REPOSITORY, 'pypi')):
+                            result['repository'] = self.DEFAULT_REPOSITORY
+                        elif (result['server'] != repository and result['repository'] != repository):
+                            result = {}
+            elif 'server-login' in sections:
+                # old format
+                server = 'server-login'
+                if config.has_option(server, 'repository'):
+                    repository = config.get(server, 'repository')
+                else:
+                    repository = self.DEFAULT_REPOSITORY
+                result = {
+                    'username': config.get(server, 'username'),
+                    'password': config.get(server, 'password'),
+                    'repository': repository,
+                    'server': server,
+                    'realm': self.DEFAULT_REALM
+                }
+        return result
+
+    def update(self, username, password):
+        # import pdb; pdb.set_trace()
+        config = configparser.RawConfigParser()
+        fn = self.filename
+        config.read(fn)
+        if not config.has_section('pypi'):
+            config.add_section('pypi')
+        config.set('pypi', 'username', username)
+        config.set('pypi', 'password', password)
+        with open(fn, 'w') as f:
+            config.write(f)
+
+
+def _load_pypirc(index):
+    """
+    Read the PyPI access configuration as supported by distutils.
+    """
+    return PyPIRCFile(url=index.url).read()
+
+
+def _store_pypirc(index):
+    PyPIRCFile().update(index.username, index.password)
+
+
+#
+# get_platform()/get_host_platform() copied from Python 3.10.a0 source, with some minor
+# tweaks
+#
+
+
+def get_host_platform():
+    """Return a string that identifies the current platform.  This is used mainly to
+    distinguish platform-specific build directories and platform-specific built
+    distributions.  Typically includes the OS name and version and the
+    architecture (as supplied by 'os.uname()'), although the exact information
+    included depends on the OS; eg. on Linux, the kernel version isn't
+    particularly important.
+
+    Examples of returned values:
+       linux-i586
+       linux-alpha (?)
+       solaris-2.6-sun4u
+
+    Windows will return one of:
+       win-amd64 (64bit Windows on AMD64 (aka x86_64, Intel64, EM64T, etc)
+       win32 (all others - specifically, sys.platform is returned)
+
+    For other non-POSIX platforms, currently just returns 'sys.platform'.
+
+    """
+    if os.name == 'nt':
+        if 'amd64' in sys.version.lower():
+            return 'win-amd64'
+        if '(arm)' in sys.version.lower():
+            return 'win-arm32'
+        if '(arm64)' in sys.version.lower():
+            return 'win-arm64'
+        return sys.platform
+
+    # Set for cross builds explicitly
+    if "_PYTHON_HOST_PLATFORM" in os.environ:
+        return os.environ["_PYTHON_HOST_PLATFORM"]
+
+    if os.name != 'posix' or not hasattr(os, 'uname'):
+        # XXX what about the architecture? NT is Intel or Alpha,
+        # Mac OS is M68k or PPC, etc.
+        return sys.platform
+
+    # Try to distinguish various flavours of Unix
+
+    (osname, host, release, version, machine) = os.uname()
+
+    # Convert the OS name to lowercase, remove '/' characters, and translate
+    # spaces (for "Power Macintosh")
+    osname = osname.lower().replace('/', '')
+    machine = machine.replace(' ', '_').replace('/', '-')
+
+    if osname[:5] == 'linux':
+        # At least on Linux/Intel, 'machine' is the processor --
+        # i386, etc.
+        # XXX what about Alpha, SPARC, etc?
+        return "%s-%s" % (osname, machine)
+
+    elif osname[:5] == 'sunos':
+        if release[0] >= '5':  # SunOS 5 == Solaris 2
+            osname = 'solaris'
+            release = '%d.%s' % (int(release[0]) - 3, release[2:])
+            # We can't use 'platform.architecture()[0]' because a
+            # bootstrap problem. We use a dict to get an error
+            # if some suspicious happens.
+            bitness = {2147483647: '32bit', 9223372036854775807: '64bit'}
+            machine += '.%s' % bitness[sys.maxsize]
+        # fall through to standard osname-release-machine representation
+    elif osname[:3] == 'aix':
+        from _aix_support import aix_platform
+        return aix_platform()
+    elif osname[:6] == 'cygwin':
+        osname = 'cygwin'
+        rel_re = re.compile(r'[\d.]+', re.ASCII)
+        m = rel_re.match(release)
+        if m:
+            release = m.group()
+    elif osname[:6] == 'darwin':
+        import _osx_support
+        try:
+            from distutils import sysconfig
+        except ImportError:
+            import sysconfig
+        osname, release, machine = _osx_support.get_platform_osx(sysconfig.get_config_vars(), osname, release, machine)
+
+    return '%s-%s-%s' % (osname, release, machine)
+
+
+_TARGET_TO_PLAT = {
+    'x86': 'win32',
+    'x64': 'win-amd64',
+    'arm': 'win-arm32',
+}
+
+
+def get_platform():
+    if os.name != 'nt':
+        return get_host_platform()
+    cross_compilation_target = os.environ.get('VSCMD_ARG_TGT_ARCH')
+    if cross_compilation_target not in _TARGET_TO_PLAT:
+        return get_host_platform()
+    return _TARGET_TO_PLAT[cross_compilation_target]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/version.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/version.py
new file mode 100644
index 00000000..d70a96ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/version.py
@@ -0,0 +1,750 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2012-2023 The Python Software Foundation.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+"""
+Implementation of a flexible versioning scheme providing support for PEP-440,
+setuptools-compatible and semantic versioning.
+"""
+
+import logging
+import re
+
+from .compat import string_types
+from .util import parse_requirement
+
+__all__ = ['NormalizedVersion', 'NormalizedMatcher',
+           'LegacyVersion', 'LegacyMatcher',
+           'SemanticVersion', 'SemanticMatcher',
+           'UnsupportedVersionError', 'get_scheme']
+
+logger = logging.getLogger(__name__)
+
+
+class UnsupportedVersionError(ValueError):
+    """This is an unsupported version."""
+    pass
+
+
+class Version(object):
+    def __init__(self, s):
+        self._string = s = s.strip()
+        self._parts = parts = self.parse(s)
+        assert isinstance(parts, tuple)
+        assert len(parts) > 0
+
+    def parse(self, s):
+        raise NotImplementedError('please implement in a subclass')
+
+    def _check_compatible(self, other):
+        if type(self) != type(other):
+            raise TypeError('cannot compare %r and %r' % (self, other))
+
+    def __eq__(self, other):
+        self._check_compatible(other)
+        return self._parts == other._parts
+
+    def __ne__(self, other):
+        return not self.__eq__(other)
+
+    def __lt__(self, other):
+        self._check_compatible(other)
+        return self._parts < other._parts
+
+    def __gt__(self, other):
+        return not (self.__lt__(other) or self.__eq__(other))
+
+    def __le__(self, other):
+        return self.__lt__(other) or self.__eq__(other)
+
+    def __ge__(self, other):
+        return self.__gt__(other) or self.__eq__(other)
+
+    # See http://docs.python.org/reference/datamodel#object.__hash__
+    def __hash__(self):
+        return hash(self._parts)
+
+    def __repr__(self):
+        return "%s('%s')" % (self.__class__.__name__, self._string)
+
+    def __str__(self):
+        return self._string
+
+    @property
+    def is_prerelease(self):
+        raise NotImplementedError('Please implement in subclasses.')
+
+
+class Matcher(object):
+    version_class = None
+
+    # value is either a callable or the name of a method
+    _operators = {
+        '<': lambda v, c, p: v < c,
+        '>': lambda v, c, p: v > c,
+        '<=': lambda v, c, p: v == c or v < c,
+        '>=': lambda v, c, p: v == c or v > c,
+        '==': lambda v, c, p: v == c,
+        '===': lambda v, c, p: v == c,
+        # by default, compatible => >=.
+        '~=': lambda v, c, p: v == c or v > c,
+        '!=': lambda v, c, p: v != c,
+    }
+
+    # this is a method only to support alternative implementations
+    # via overriding
+    def parse_requirement(self, s):
+        return parse_requirement(s)
+
+    def __init__(self, s):
+        if self.version_class is None:
+            raise ValueError('Please specify a version class')
+        self._string = s = s.strip()
+        r = self.parse_requirement(s)
+        if not r:
+            raise ValueError('Not valid: %r' % s)
+        self.name = r.name
+        self.key = self.name.lower()    # for case-insensitive comparisons
+        clist = []
+        if r.constraints:
+            # import pdb; pdb.set_trace()
+            for op, s in r.constraints:
+                if s.endswith('.*'):
+                    if op not in ('==', '!='):
+                        raise ValueError('\'.*\' not allowed for '
+                                         '%r constraints' % op)
+                    # Could be a partial version (e.g. for '2.*') which
+                    # won't parse as a version, so keep it as a string
+                    vn, prefix = s[:-2], True
+                    # Just to check that vn is a valid version
+                    self.version_class(vn)
+                else:
+                    # Should parse as a version, so we can create an
+                    # instance for the comparison
+                    vn, prefix = self.version_class(s), False
+                clist.append((op, vn, prefix))
+        self._parts = tuple(clist)
+
+    def match(self, version):
+        """
+        Check if the provided version matches the constraints.
+
+        :param version: The version to match against this instance.
+        :type version: String or :class:`Version` instance.
+        """
+        if isinstance(version, string_types):
+            version = self.version_class(version)
+        for operator, constraint, prefix in self._parts:
+            f = self._operators.get(operator)
+            if isinstance(f, string_types):
+                f = getattr(self, f)
+            if not f:
+                msg = ('%r not implemented '
+                       'for %s' % (operator, self.__class__.__name__))
+                raise NotImplementedError(msg)
+            if not f(version, constraint, prefix):
+                return False
+        return True
+
+    @property
+    def exact_version(self):
+        result = None
+        if len(self._parts) == 1 and self._parts[0][0] in ('==', '==='):
+            result = self._parts[0][1]
+        return result
+
+    def _check_compatible(self, other):
+        if type(self) != type(other) or self.name != other.name:
+            raise TypeError('cannot compare %s and %s' % (self, other))
+
+    def __eq__(self, other):
+        self._check_compatible(other)
+        return self.key == other.key and self._parts == other._parts
+
+    def __ne__(self, other):
+        return not self.__eq__(other)
+
+    # See http://docs.python.org/reference/datamodel#object.__hash__
+    def __hash__(self):
+        return hash(self.key) + hash(self._parts)
+
+    def __repr__(self):
+        return "%s(%r)" % (self.__class__.__name__, self._string)
+
+    def __str__(self):
+        return self._string
+
+
+PEP440_VERSION_RE = re.compile(r'^v?(\d+!)?(\d+(\.\d+)*)((a|alpha|b|beta|c|rc|pre|preview)(\d+)?)?'
+                               r'(\.(post|r|rev)(\d+)?)?([._-]?(dev)(\d+)?)?'
+                               r'(\+([a-zA-Z\d]+(\.[a-zA-Z\d]+)?))?$', re.I)
+
+
+def _pep_440_key(s):
+    s = s.strip()
+    m = PEP440_VERSION_RE.match(s)
+    if not m:
+        raise UnsupportedVersionError('Not a valid version: %s' % s)
+    groups = m.groups()
+    nums = tuple(int(v) for v in groups[1].split('.'))
+    while len(nums) > 1 and nums[-1] == 0:
+        nums = nums[:-1]
+
+    if not groups[0]:
+        epoch = 0
+    else:
+        epoch = int(groups[0][:-1])
+    pre = groups[4:6]
+    post = groups[7:9]
+    dev = groups[10:12]
+    local = groups[13]
+    if pre == (None, None):
+        pre = ()
+    else:
+        if pre[1] is None:
+            pre = pre[0], 0
+        else:
+            pre = pre[0], int(pre[1])
+    if post == (None, None):
+        post = ()
+    else:
+        if post[1] is None:
+            post = post[0], 0
+        else:
+            post = post[0], int(post[1])
+    if dev == (None, None):
+        dev = ()
+    else:
+        if dev[1] is None:
+            dev = dev[0], 0
+        else:
+            dev = dev[0], int(dev[1])
+    if local is None:
+        local = ()
+    else:
+        parts = []
+        for part in local.split('.'):
+            # to ensure that numeric compares as > lexicographic, avoid
+            # comparing them directly, but encode a tuple which ensures
+            # correct sorting
+            if part.isdigit():
+                part = (1, int(part))
+            else:
+                part = (0, part)
+            parts.append(part)
+        local = tuple(parts)
+    if not pre:
+        # either before pre-release, or final release and after
+        if not post and dev:
+            # before pre-release
+            pre = ('a', -1)     # to sort before a0
+        else:
+            pre = ('z',)        # to sort after all pre-releases
+    # now look at the state of post and dev.
+    if not post:
+        post = ('_',)   # sort before 'a'
+    if not dev:
+        dev = ('final',)
+
+    return epoch, nums, pre, post, dev, local
+
+
+_normalized_key = _pep_440_key
+
+
+class NormalizedVersion(Version):
+    """A rational version.
+
+    Good:
+        1.2         # equivalent to "1.2.0"
+        1.2.0
+        1.2a1
+        1.2.3a2
+        1.2.3b1
+        1.2.3c1
+        1.2.3.4
+        TODO: fill this out
+
+    Bad:
+        1           # minimum two numbers
+        1.2a        # release level must have a release serial
+        1.2.3b
+    """
+    def parse(self, s):
+        result = _normalized_key(s)
+        # _normalized_key loses trailing zeroes in the release
+        # clause, since that's needed to ensure that X.Y == X.Y.0 == X.Y.0.0
+        # However, PEP 440 prefix matching needs it: for example,
+        # (~= 1.4.5.0) matches differently to (~= 1.4.5.0.0).
+        m = PEP440_VERSION_RE.match(s)      # must succeed
+        groups = m.groups()
+        self._release_clause = tuple(int(v) for v in groups[1].split('.'))
+        return result
+
+    PREREL_TAGS = set(['a', 'b', 'c', 'rc', 'dev'])
+
+    @property
+    def is_prerelease(self):
+        return any(t[0] in self.PREREL_TAGS for t in self._parts if t)
+
+
+def _match_prefix(x, y):
+    x = str(x)
+    y = str(y)
+    if x == y:
+        return True
+    if not x.startswith(y):
+        return False
+    n = len(y)
+    return x[n] == '.'
+
+
+class NormalizedMatcher(Matcher):
+    version_class = NormalizedVersion
+
+    # value is either a callable or the name of a method
+    _operators = {
+        '~=': '_match_compatible',
+        '<': '_match_lt',
+        '>': '_match_gt',
+        '<=': '_match_le',
+        '>=': '_match_ge',
+        '==': '_match_eq',
+        '===': '_match_arbitrary',
+        '!=': '_match_ne',
+    }
+
+    def _adjust_local(self, version, constraint, prefix):
+        if prefix:
+            strip_local = '+' not in constraint and version._parts[-1]
+        else:
+            # both constraint and version are
+            # NormalizedVersion instances.
+            # If constraint does not have a local component,
+            # ensure the version doesn't, either.
+            strip_local = not constraint._parts[-1] and version._parts[-1]
+        if strip_local:
+            s = version._string.split('+', 1)[0]
+            version = self.version_class(s)
+        return version, constraint
+
+    def _match_lt(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        if version >= constraint:
+            return False
+        release_clause = constraint._release_clause
+        pfx = '.'.join([str(i) for i in release_clause])
+        return not _match_prefix(version, pfx)
+
+    def _match_gt(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        if version <= constraint:
+            return False
+        release_clause = constraint._release_clause
+        pfx = '.'.join([str(i) for i in release_clause])
+        return not _match_prefix(version, pfx)
+
+    def _match_le(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        return version <= constraint
+
+    def _match_ge(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        return version >= constraint
+
+    def _match_eq(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        if not prefix:
+            result = (version == constraint)
+        else:
+            result = _match_prefix(version, constraint)
+        return result
+
+    def _match_arbitrary(self, version, constraint, prefix):
+        return str(version) == str(constraint)
+
+    def _match_ne(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        if not prefix:
+            result = (version != constraint)
+        else:
+            result = not _match_prefix(version, constraint)
+        return result
+
+    def _match_compatible(self, version, constraint, prefix):
+        version, constraint = self._adjust_local(version, constraint, prefix)
+        if version == constraint:
+            return True
+        if version < constraint:
+            return False
+#        if not prefix:
+#            return True
+        release_clause = constraint._release_clause
+        if len(release_clause) > 1:
+            release_clause = release_clause[:-1]
+        pfx = '.'.join([str(i) for i in release_clause])
+        return _match_prefix(version, pfx)
+
+
+_REPLACEMENTS = (
+    (re.compile('[.+-]$'), ''),                     # remove trailing puncts
+    (re.compile(r'^[.](\d)'), r'0.\1'),             # .N -> 0.N at start
+    (re.compile('^[.-]'), ''),                      # remove leading puncts
+    (re.compile(r'^\((.*)\)$'), r'\1'),             # remove parentheses
+    (re.compile(r'^v(ersion)?\s*(\d+)'), r'\2'),    # remove leading v(ersion)
+    (re.compile(r'^r(ev)?\s*(\d+)'), r'\2'),        # remove leading v(ersion)
+    (re.compile('[.]{2,}'), '.'),                   # multiple runs of '.'
+    (re.compile(r'\b(alfa|apha)\b'), 'alpha'),      # misspelt alpha
+    (re.compile(r'\b(pre-alpha|prealpha)\b'),
+        'pre.alpha'),                               # standardise
+    (re.compile(r'\(beta\)$'), 'beta'),             # remove parentheses
+)
+
+_SUFFIX_REPLACEMENTS = (
+    (re.compile('^[:~._+-]+'), ''),                   # remove leading puncts
+    (re.compile('[,*")([\\]]'), ''),                  # remove unwanted chars
+    (re.compile('[~:+_ -]'), '.'),                    # replace illegal chars
+    (re.compile('[.]{2,}'), '.'),                   # multiple runs of '.'
+    (re.compile(r'\.$'), ''),                       # trailing '.'
+)
+
+_NUMERIC_PREFIX = re.compile(r'(\d+(\.\d+)*)')
+
+
+def _suggest_semantic_version(s):
+    """
+    Try to suggest a semantic form for a version for which
+    _suggest_normalized_version couldn't come up with anything.
+    """
+    result = s.strip().lower()
+    for pat, repl in _REPLACEMENTS:
+        result = pat.sub(repl, result)
+    if not result:
+        result = '0.0.0'
+
+    # Now look for numeric prefix, and separate it out from
+    # the rest.
+    # import pdb; pdb.set_trace()
+    m = _NUMERIC_PREFIX.match(result)
+    if not m:
+        prefix = '0.0.0'
+        suffix = result
+    else:
+        prefix = m.groups()[0].split('.')
+        prefix = [int(i) for i in prefix]
+        while len(prefix) < 3:
+            prefix.append(0)
+        if len(prefix) == 3:
+            suffix = result[m.end():]
+        else:
+            suffix = '.'.join([str(i) for i in prefix[3:]]) + result[m.end():]
+            prefix = prefix[:3]
+        prefix = '.'.join([str(i) for i in prefix])
+        suffix = suffix.strip()
+    if suffix:
+        # import pdb; pdb.set_trace()
+        # massage the suffix.
+        for pat, repl in _SUFFIX_REPLACEMENTS:
+            suffix = pat.sub(repl, suffix)
+
+    if not suffix:
+        result = prefix
+    else:
+        sep = '-' if 'dev' in suffix else '+'
+        result = prefix + sep + suffix
+    if not is_semver(result):
+        result = None
+    return result
+
+
+def _suggest_normalized_version(s):
+    """Suggest a normalized version close to the given version string.
+
+    If you have a version string that isn't rational (i.e. NormalizedVersion
+    doesn't like it) then you might be able to get an equivalent (or close)
+    rational version from this function.
+
+    This does a number of simple normalizations to the given string, based
+    on observation of versions currently in use on PyPI. Given a dump of
+    those version during PyCon 2009, 4287 of them:
+    - 2312 (53.93%) match NormalizedVersion without change
+      with the automatic suggestion
+    - 3474 (81.04%) match when using this suggestion method
+
+    @param s {str} An irrational version string.
+    @returns A rational version string, or None, if couldn't determine one.
+    """
+    try:
+        _normalized_key(s)
+        return s   # already rational
+    except UnsupportedVersionError:
+        pass
+
+    rs = s.lower()
+
+    # part of this could use maketrans
+    for orig, repl in (('-alpha', 'a'), ('-beta', 'b'), ('alpha', 'a'),
+                       ('beta', 'b'), ('rc', 'c'), ('-final', ''),
+                       ('-pre', 'c'),
+                       ('-release', ''), ('.release', ''), ('-stable', ''),
+                       ('+', '.'), ('_', '.'), (' ', ''), ('.final', ''),
+                       ('final', '')):
+        rs = rs.replace(orig, repl)
+
+    # if something ends with dev or pre, we add a 0
+    rs = re.sub(r"pre$", r"pre0", rs)
+    rs = re.sub(r"dev$", r"dev0", rs)
+
+    # if we have something like "b-2" or "a.2" at the end of the
+    # version, that is probably beta, alpha, etc
+    # let's remove the dash or dot
+    rs = re.sub(r"([abc]|rc)[\-\.](\d+)$", r"\1\2", rs)
+
+    # 1.0-dev-r371 -> 1.0.dev371
+    # 0.1-dev-r79 -> 0.1.dev79
+    rs = re.sub(r"[\-\.](dev)[\-\.]?r?(\d+)$", r".\1\2", rs)
+
+    # Clean: 2.0.a.3, 2.0.b1, 0.9.0~c1
+    rs = re.sub(r"[.~]?([abc])\.?", r"\1", rs)
+
+    # Clean: v0.3, v1.0
+    if rs.startswith('v'):
+        rs = rs[1:]
+
+    # Clean leading '0's on numbers.
+    # TODO: unintended side-effect on, e.g., "2003.05.09"
+    # PyPI stats: 77 (~2%) better
+    rs = re.sub(r"\b0+(\d+)(?!\d)", r"\1", rs)
+
+    # Clean a/b/c with no version. E.g. "1.0a" -> "1.0a0". Setuptools infers
+    # zero.
+    # PyPI stats: 245 (7.56%) better
+    rs = re.sub(r"(\d+[abc])$", r"\g<1>0", rs)
+
+    # the 'dev-rNNN' tag is a dev tag
+    rs = re.sub(r"\.?(dev-r|dev\.r)\.?(\d+)$", r".dev\2", rs)
+
+    # clean the - when used as a pre delimiter
+    rs = re.sub(r"-(a|b|c)(\d+)$", r"\1\2", rs)
+
+    # a terminal "dev" or "devel" can be changed into ".dev0"
+    rs = re.sub(r"[\.\-](dev|devel)$", r".dev0", rs)
+
+    # a terminal "dev" can be changed into ".dev0"
+    rs = re.sub(r"(?![\.\-])dev$", r".dev0", rs)
+
+    # a terminal "final" or "stable" can be removed
+    rs = re.sub(r"(final|stable)$", "", rs)
+
+    # The 'r' and the '-' tags are post release tags
+    #   0.4a1.r10       ->  0.4a1.post10
+    #   0.9.33-17222    ->  0.9.33.post17222
+    #   0.9.33-r17222   ->  0.9.33.post17222
+    rs = re.sub(r"\.?(r|-|-r)\.?(\d+)$", r".post\2", rs)
+
+    # Clean 'r' instead of 'dev' usage:
+    #   0.9.33+r17222   ->  0.9.33.dev17222
+    #   1.0dev123       ->  1.0.dev123
+    #   1.0.git123      ->  1.0.dev123
+    #   1.0.bzr123      ->  1.0.dev123
+    #   0.1a0dev.123    ->  0.1a0.dev123
+    # PyPI stats:  ~150 (~4%) better
+    rs = re.sub(r"\.?(dev|git|bzr)\.?(\d+)$", r".dev\2", rs)
+
+    # Clean '.pre' (normalized from '-pre' above) instead of 'c' usage:
+    #   0.2.pre1        ->  0.2c1
+    #   0.2-c1         ->  0.2c1
+    #   1.0preview123   ->  1.0c123
+    # PyPI stats: ~21 (0.62%) better
+    rs = re.sub(r"\.?(pre|preview|-c)(\d+)$", r"c\g<2>", rs)
+
+    # Tcl/Tk uses "px" for their post release markers
+    rs = re.sub(r"p(\d+)$", r".post\1", rs)
+
+    try:
+        _normalized_key(rs)
+    except UnsupportedVersionError:
+        rs = None
+    return rs
+
+#
+#   Legacy version processing (distribute-compatible)
+#
+
+
+_VERSION_PART = re.compile(r'([a-z]+|\d+|[\.-])', re.I)
+_VERSION_REPLACE = {
+    'pre': 'c',
+    'preview': 'c',
+    '-': 'final-',
+    'rc': 'c',
+    'dev': '@',
+    '': None,
+    '.': None,
+}
+
+
+def _legacy_key(s):
+    def get_parts(s):
+        result = []
+        for p in _VERSION_PART.split(s.lower()):
+            p = _VERSION_REPLACE.get(p, p)
+            if p:
+                if '0' <= p[:1] <= '9':
+                    p = p.zfill(8)
+                else:
+                    p = '*' + p
+                result.append(p)
+        result.append('*final')
+        return result
+
+    result = []
+    for p in get_parts(s):
+        if p.startswith('*'):
+            if p < '*final':
+                while result and result[-1] == '*final-':
+                    result.pop()
+            while result and result[-1] == '00000000':
+                result.pop()
+        result.append(p)
+    return tuple(result)
+
+
+class LegacyVersion(Version):
+    def parse(self, s):
+        return _legacy_key(s)
+
+    @property
+    def is_prerelease(self):
+        result = False
+        for x in self._parts:
+            if (isinstance(x, string_types) and x.startswith('*') and x < '*final'):
+                result = True
+                break
+        return result
+
+
+class LegacyMatcher(Matcher):
+    version_class = LegacyVersion
+
+    _operators = dict(Matcher._operators)
+    _operators['~='] = '_match_compatible'
+
+    numeric_re = re.compile(r'^(\d+(\.\d+)*)')
+
+    def _match_compatible(self, version, constraint, prefix):
+        if version < constraint:
+            return False
+        m = self.numeric_re.match(str(constraint))
+        if not m:
+            logger.warning('Cannot compute compatible match for version %s '
+                           ' and constraint %s', version, constraint)
+            return True
+        s = m.groups()[0]
+        if '.' in s:
+            s = s.rsplit('.', 1)[0]
+        return _match_prefix(version, s)
+
+#
+#   Semantic versioning
+#
+
+
+_SEMVER_RE = re.compile(r'^(\d+)\.(\d+)\.(\d+)'
+                        r'(-[a-z0-9]+(\.[a-z0-9-]+)*)?'
+                        r'(\+[a-z0-9]+(\.[a-z0-9-]+)*)?$', re.I)
+
+
+def is_semver(s):
+    return _SEMVER_RE.match(s)
+
+
+def _semantic_key(s):
+    def make_tuple(s, absent):
+        if s is None:
+            result = (absent,)
+        else:
+            parts = s[1:].split('.')
+            # We can't compare ints and strings on Python 3, so fudge it
+            # by zero-filling numeric values so simulate a numeric comparison
+            result = tuple([p.zfill(8) if p.isdigit() else p for p in parts])
+        return result
+
+    m = is_semver(s)
+    if not m:
+        raise UnsupportedVersionError(s)
+    groups = m.groups()
+    major, minor, patch = [int(i) for i in groups[:3]]
+    # choose the '|' and '*' so that versions sort correctly
+    pre, build = make_tuple(groups[3], '|'), make_tuple(groups[5], '*')
+    return (major, minor, patch), pre, build
+
+
+class SemanticVersion(Version):
+    def parse(self, s):
+        return _semantic_key(s)
+
+    @property
+    def is_prerelease(self):
+        return self._parts[1][0] != '|'
+
+
+class SemanticMatcher(Matcher):
+    version_class = SemanticVersion
+
+
+class VersionScheme(object):
+    def __init__(self, key, matcher, suggester=None):
+        self.key = key
+        self.matcher = matcher
+        self.suggester = suggester
+
+    def is_valid_version(self, s):
+        try:
+            self.matcher.version_class(s)
+            result = True
+        except UnsupportedVersionError:
+            result = False
+        return result
+
+    def is_valid_matcher(self, s):
+        try:
+            self.matcher(s)
+            result = True
+        except UnsupportedVersionError:
+            result = False
+        return result
+
+    def is_valid_constraint_list(self, s):
+        """
+        Used for processing some metadata fields
+        """
+        # See issue #140. Be tolerant of a single trailing comma.
+        if s.endswith(','):
+            s = s[:-1]
+        return self.is_valid_matcher('dummy_name (%s)' % s)
+
+    def suggest(self, s):
+        if self.suggester is None:
+            result = None
+        else:
+            result = self.suggester(s)
+        return result
+
+
+_SCHEMES = {
+    'normalized': VersionScheme(_normalized_key, NormalizedMatcher,
+                                _suggest_normalized_version),
+    'legacy': VersionScheme(_legacy_key, LegacyMatcher, lambda self, s: s),
+    'semantic': VersionScheme(_semantic_key, SemanticMatcher,
+                              _suggest_semantic_version),
+}
+
+_SCHEMES['default'] = _SCHEMES['normalized']
+
+
+def get_scheme(name):
+    if name not in _SCHEMES:
+        raise ValueError('unknown scheme name: %r' % name)
+    return _SCHEMES[name]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w32.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w32.exe
new file mode 100644
index 00000000..4ee2d3a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w32.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64-arm.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64-arm.exe
new file mode 100644
index 00000000..951d5817
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64-arm.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64.exe b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64.exe
new file mode 100644
index 00000000..5763076d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/w64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/wheel.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/wheel.py
new file mode 100644
index 00000000..62ab10fb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distlib/wheel.py
@@ -0,0 +1,1100 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2013-2023 Vinay Sajip.
+# Licensed to the Python Software Foundation under a contributor agreement.
+# See LICENSE.txt and CONTRIBUTORS.txt.
+#
+from __future__ import unicode_literals
+
+import base64
+import codecs
+import datetime
+from email import message_from_file
+import hashlib
+import json
+import logging
+import os
+import posixpath
+import re
+import shutil
+import sys
+import tempfile
+import zipfile
+
+from . import __version__, DistlibException
+from .compat import sysconfig, ZipFile, fsdecode, text_type, filter
+from .database import InstalledDistribution
+from .metadata import Metadata, WHEEL_METADATA_FILENAME, LEGACY_METADATA_FILENAME
+from .util import (FileOperator, convert_path, CSVReader, CSVWriter, Cache, cached_property, get_cache_base,
+                   read_exports, tempdir, get_platform)
+from .version import NormalizedVersion, UnsupportedVersionError
+
+logger = logging.getLogger(__name__)
+
+cache = None  # created when needed
+
+if hasattr(sys, 'pypy_version_info'):  # pragma: no cover
+    IMP_PREFIX = 'pp'
+elif sys.platform.startswith('java'):  # pragma: no cover
+    IMP_PREFIX = 'jy'
+elif sys.platform == 'cli':  # pragma: no cover
+    IMP_PREFIX = 'ip'
+else:
+    IMP_PREFIX = 'cp'
+
+VER_SUFFIX = sysconfig.get_config_var('py_version_nodot')
+if not VER_SUFFIX:  # pragma: no cover
+    VER_SUFFIX = '%s%s' % sys.version_info[:2]
+PYVER = 'py' + VER_SUFFIX
+IMPVER = IMP_PREFIX + VER_SUFFIX
+
+ARCH = get_platform().replace('-', '_').replace('.', '_')
+
+ABI = sysconfig.get_config_var('SOABI')
+if ABI and ABI.startswith('cpython-'):
+    ABI = ABI.replace('cpython-', 'cp').split('-')[0]
+else:
+
+    def _derive_abi():
+        parts = ['cp', VER_SUFFIX]
+        if sysconfig.get_config_var('Py_DEBUG'):
+            parts.append('d')
+        if IMP_PREFIX == 'cp':
+            vi = sys.version_info[:2]
+            if vi < (3, 8):
+                wpm = sysconfig.get_config_var('WITH_PYMALLOC')
+                if wpm is None:
+                    wpm = True
+                if wpm:
+                    parts.append('m')
+                if vi < (3, 3):
+                    us = sysconfig.get_config_var('Py_UNICODE_SIZE')
+                    if us == 4 or (us is None and sys.maxunicode == 0x10FFFF):
+                        parts.append('u')
+        return ''.join(parts)
+
+    ABI = _derive_abi()
+    del _derive_abi
+
+FILENAME_RE = re.compile(
+    r'''
+(?P<nm>[^-]+)
+-(?P<vn>\d+[^-]*)
+(-(?P<bn>\d+[^-]*))?
+-(?P<py>\w+\d+(\.\w+\d+)*)
+-(?P<bi>\w+)
+-(?P<ar>\w+(\.\w+)*)
+\.whl$
+''', re.IGNORECASE | re.VERBOSE)
+
+NAME_VERSION_RE = re.compile(r'''
+(?P<nm>[^-]+)
+-(?P<vn>\d+[^-]*)
+(-(?P<bn>\d+[^-]*))?$
+''', re.IGNORECASE | re.VERBOSE)
+
+SHEBANG_RE = re.compile(br'\s*#![^\r\n]*')
+SHEBANG_DETAIL_RE = re.compile(br'^(\s*#!("[^"]+"|\S+))\s+(.*)$')
+SHEBANG_PYTHON = b'#!python'
+SHEBANG_PYTHONW = b'#!pythonw'
+
+if os.sep == '/':
+    to_posix = lambda o: o
+else:
+    to_posix = lambda o: o.replace(os.sep, '/')
+
+if sys.version_info[0] < 3:
+    import imp
+else:
+    imp = None
+    import importlib.machinery
+    import importlib.util
+
+
+def _get_suffixes():
+    if imp:
+        return [s[0] for s in imp.get_suffixes()]
+    else:
+        return importlib.machinery.EXTENSION_SUFFIXES
+
+
+def _load_dynamic(name, path):
+    # https://docs.python.org/3/library/importlib.html#importing-a-source-file-directly
+    if imp:
+        return imp.load_dynamic(name, path)
+    else:
+        spec = importlib.util.spec_from_file_location(name, path)
+        module = importlib.util.module_from_spec(spec)
+        sys.modules[name] = module
+        spec.loader.exec_module(module)
+        return module
+
+
+class Mounter(object):
+
+    def __init__(self):
+        self.impure_wheels = {}
+        self.libs = {}
+
+    def add(self, pathname, extensions):
+        self.impure_wheels[pathname] = extensions
+        self.libs.update(extensions)
+
+    def remove(self, pathname):
+        extensions = self.impure_wheels.pop(pathname)
+        for k, v in extensions:
+            if k in self.libs:
+                del self.libs[k]
+
+    def find_module(self, fullname, path=None):
+        if fullname in self.libs:
+            result = self
+        else:
+            result = None
+        return result
+
+    def load_module(self, fullname):
+        if fullname in sys.modules:
+            result = sys.modules[fullname]
+        else:
+            if fullname not in self.libs:
+                raise ImportError('unable to find extension for %s' % fullname)
+            result = _load_dynamic(fullname, self.libs[fullname])
+            result.__loader__ = self
+            parts = fullname.rsplit('.', 1)
+            if len(parts) > 1:
+                result.__package__ = parts[0]
+        return result
+
+
+_hook = Mounter()
+
+
+class Wheel(object):
+    """
+    Class to build and install from Wheel files (PEP 427).
+    """
+
+    wheel_version = (1, 1)
+    hash_kind = 'sha256'
+
+    def __init__(self, filename=None, sign=False, verify=False):
+        """
+        Initialise an instance using a (valid) filename.
+        """
+        self.sign = sign
+        self.should_verify = verify
+        self.buildver = ''
+        self.pyver = [PYVER]
+        self.abi = ['none']
+        self.arch = ['any']
+        self.dirname = os.getcwd()
+        if filename is None:
+            self.name = 'dummy'
+            self.version = '0.1'
+            self._filename = self.filename
+        else:
+            m = NAME_VERSION_RE.match(filename)
+            if m:
+                info = m.groupdict('')
+                self.name = info['nm']
+                # Reinstate the local version separator
+                self.version = info['vn'].replace('_', '-')
+                self.buildver = info['bn']
+                self._filename = self.filename
+            else:
+                dirname, filename = os.path.split(filename)
+                m = FILENAME_RE.match(filename)
+                if not m:
+                    raise DistlibException('Invalid name or '
+                                           'filename: %r' % filename)
+                if dirname:
+                    self.dirname = os.path.abspath(dirname)
+                self._filename = filename
+                info = m.groupdict('')
+                self.name = info['nm']
+                self.version = info['vn']
+                self.buildver = info['bn']
+                self.pyver = info['py'].split('.')
+                self.abi = info['bi'].split('.')
+                self.arch = info['ar'].split('.')
+
+    @property
+    def filename(self):
+        """
+        Build and return a filename from the various components.
+        """
+        if self.buildver:
+            buildver = '-' + self.buildver
+        else:
+            buildver = ''
+        pyver = '.'.join(self.pyver)
+        abi = '.'.join(self.abi)
+        arch = '.'.join(self.arch)
+        # replace - with _ as a local version separator
+        version = self.version.replace('-', '_')
+        return '%s-%s%s-%s-%s-%s.whl' % (self.name, version, buildver, pyver, abi, arch)
+
+    @property
+    def exists(self):
+        path = os.path.join(self.dirname, self.filename)
+        return os.path.isfile(path)
+
+    @property
+    def tags(self):
+        for pyver in self.pyver:
+            for abi in self.abi:
+                for arch in self.arch:
+                    yield pyver, abi, arch
+
+    @cached_property
+    def metadata(self):
+        pathname = os.path.join(self.dirname, self.filename)
+        name_ver = '%s-%s' % (self.name, self.version)
+        info_dir = '%s.dist-info' % name_ver
+        wrapper = codecs.getreader('utf-8')
+        with ZipFile(pathname, 'r') as zf:
+            self.get_wheel_metadata(zf)
+            # wv = wheel_metadata['Wheel-Version'].split('.', 1)
+            # file_version = tuple([int(i) for i in wv])
+            # if file_version < (1, 1):
+            # fns = [WHEEL_METADATA_FILENAME, METADATA_FILENAME,
+            # LEGACY_METADATA_FILENAME]
+            # else:
+            # fns = [WHEEL_METADATA_FILENAME, METADATA_FILENAME]
+            fns = [WHEEL_METADATA_FILENAME, LEGACY_METADATA_FILENAME]
+            result = None
+            for fn in fns:
+                try:
+                    metadata_filename = posixpath.join(info_dir, fn)
+                    with zf.open(metadata_filename) as bf:
+                        wf = wrapper(bf)
+                        result = Metadata(fileobj=wf)
+                        if result:
+                            break
+                except KeyError:
+                    pass
+            if not result:
+                raise ValueError('Invalid wheel, because metadata is '
+                                 'missing: looked in %s' % ', '.join(fns))
+        return result
+
+    def get_wheel_metadata(self, zf):
+        name_ver = '%s-%s' % (self.name, self.version)
+        info_dir = '%s.dist-info' % name_ver
+        metadata_filename = posixpath.join(info_dir, 'WHEEL')
+        with zf.open(metadata_filename) as bf:
+            wf = codecs.getreader('utf-8')(bf)
+            message = message_from_file(wf)
+        return dict(message)
+
+    @cached_property
+    def info(self):
+        pathname = os.path.join(self.dirname, self.filename)
+        with ZipFile(pathname, 'r') as zf:
+            result = self.get_wheel_metadata(zf)
+        return result
+
+    def process_shebang(self, data):
+        m = SHEBANG_RE.match(data)
+        if m:
+            end = m.end()
+            shebang, data_after_shebang = data[:end], data[end:]
+            # Preserve any arguments after the interpreter
+            if b'pythonw' in shebang.lower():
+                shebang_python = SHEBANG_PYTHONW
+            else:
+                shebang_python = SHEBANG_PYTHON
+            m = SHEBANG_DETAIL_RE.match(shebang)
+            if m:
+                args = b' ' + m.groups()[-1]
+            else:
+                args = b''
+            shebang = shebang_python + args
+            data = shebang + data_after_shebang
+        else:
+            cr = data.find(b'\r')
+            lf = data.find(b'\n')
+            if cr < 0 or cr > lf:
+                term = b'\n'
+            else:
+                if data[cr:cr + 2] == b'\r\n':
+                    term = b'\r\n'
+                else:
+                    term = b'\r'
+            data = SHEBANG_PYTHON + term + data
+        return data
+
+    def get_hash(self, data, hash_kind=None):
+        if hash_kind is None:
+            hash_kind = self.hash_kind
+        try:
+            hasher = getattr(hashlib, hash_kind)
+        except AttributeError:
+            raise DistlibException('Unsupported hash algorithm: %r' % hash_kind)
+        result = hasher(data).digest()
+        result = base64.urlsafe_b64encode(result).rstrip(b'=').decode('ascii')
+        return hash_kind, result
+
+    def write_record(self, records, record_path, archive_record_path):
+        records = list(records)  # make a copy, as mutated
+        records.append((archive_record_path, '', ''))
+        with CSVWriter(record_path) as writer:
+            for row in records:
+                writer.writerow(row)
+
+    def write_records(self, info, libdir, archive_paths):
+        records = []
+        distinfo, info_dir = info
+        # hasher = getattr(hashlib, self.hash_kind)
+        for ap, p in archive_paths:
+            with open(p, 'rb') as f:
+                data = f.read()
+            digest = '%s=%s' % self.get_hash(data)
+            size = os.path.getsize(p)
+            records.append((ap, digest, size))
+
+        p = os.path.join(distinfo, 'RECORD')
+        ap = to_posix(os.path.join(info_dir, 'RECORD'))
+        self.write_record(records, p, ap)
+        archive_paths.append((ap, p))
+
+    def build_zip(self, pathname, archive_paths):
+        with ZipFile(pathname, 'w', zipfile.ZIP_DEFLATED) as zf:
+            for ap, p in archive_paths:
+                logger.debug('Wrote %s to %s in wheel', p, ap)
+                zf.write(p, ap)
+
+    def build(self, paths, tags=None, wheel_version=None):
+        """
+        Build a wheel from files in specified paths, and use any specified tags
+        when determining the name of the wheel.
+        """
+        if tags is None:
+            tags = {}
+
+        libkey = list(filter(lambda o: o in paths, ('purelib', 'platlib')))[0]
+        if libkey == 'platlib':
+            is_pure = 'false'
+            default_pyver = [IMPVER]
+            default_abi = [ABI]
+            default_arch = [ARCH]
+        else:
+            is_pure = 'true'
+            default_pyver = [PYVER]
+            default_abi = ['none']
+            default_arch = ['any']
+
+        self.pyver = tags.get('pyver', default_pyver)
+        self.abi = tags.get('abi', default_abi)
+        self.arch = tags.get('arch', default_arch)
+
+        libdir = paths[libkey]
+
+        name_ver = '%s-%s' % (self.name, self.version)
+        data_dir = '%s.data' % name_ver
+        info_dir = '%s.dist-info' % name_ver
+
+        archive_paths = []
+
+        # First, stuff which is not in site-packages
+        for key in ('data', 'headers', 'scripts'):
+            if key not in paths:
+                continue
+            path = paths[key]
+            if os.path.isdir(path):
+                for root, dirs, files in os.walk(path):
+                    for fn in files:
+                        p = fsdecode(os.path.join(root, fn))
+                        rp = os.path.relpath(p, path)
+                        ap = to_posix(os.path.join(data_dir, key, rp))
+                        archive_paths.append((ap, p))
+                        if key == 'scripts' and not p.endswith('.exe'):
+                            with open(p, 'rb') as f:
+                                data = f.read()
+                            data = self.process_shebang(data)
+                            with open(p, 'wb') as f:
+                                f.write(data)
+
+        # Now, stuff which is in site-packages, other than the
+        # distinfo stuff.
+        path = libdir
+        distinfo = None
+        for root, dirs, files in os.walk(path):
+            if root == path:
+                # At the top level only, save distinfo for later
+                # and skip it for now
+                for i, dn in enumerate(dirs):
+                    dn = fsdecode(dn)
+                    if dn.endswith('.dist-info'):
+                        distinfo = os.path.join(root, dn)
+                        del dirs[i]
+                        break
+                assert distinfo, '.dist-info directory expected, not found'
+
+            for fn in files:
+                # comment out next suite to leave .pyc files in
+                if fsdecode(fn).endswith(('.pyc', '.pyo')):
+                    continue
+                p = os.path.join(root, fn)
+                rp = to_posix(os.path.relpath(p, path))
+                archive_paths.append((rp, p))
+
+        # Now distinfo. Assumed to be flat, i.e. os.listdir is enough.
+        files = os.listdir(distinfo)
+        for fn in files:
+            if fn not in ('RECORD', 'INSTALLER', 'SHARED', 'WHEEL'):
+                p = fsdecode(os.path.join(distinfo, fn))
+                ap = to_posix(os.path.join(info_dir, fn))
+                archive_paths.append((ap, p))
+
+        wheel_metadata = [
+            'Wheel-Version: %d.%d' % (wheel_version or self.wheel_version),
+            'Generator: distlib %s' % __version__,
+            'Root-Is-Purelib: %s' % is_pure,
+        ]
+        for pyver, abi, arch in self.tags:
+            wheel_metadata.append('Tag: %s-%s-%s' % (pyver, abi, arch))
+        p = os.path.join(distinfo, 'WHEEL')
+        with open(p, 'w') as f:
+            f.write('\n'.join(wheel_metadata))
+        ap = to_posix(os.path.join(info_dir, 'WHEEL'))
+        archive_paths.append((ap, p))
+
+        # sort the entries by archive path. Not needed by any spec, but it
+        # keeps the archive listing and RECORD tidier than they would otherwise
+        # be. Use the number of path segments to keep directory entries together,
+        # and keep the dist-info stuff at the end.
+        def sorter(t):
+            ap = t[0]
+            n = ap.count('/')
+            if '.dist-info' in ap:
+                n += 10000
+            return (n, ap)
+
+        archive_paths = sorted(archive_paths, key=sorter)
+
+        # Now, at last, RECORD.
+        # Paths in here are archive paths - nothing else makes sense.
+        self.write_records((distinfo, info_dir), libdir, archive_paths)
+        # Now, ready to build the zip file
+        pathname = os.path.join(self.dirname, self.filename)
+        self.build_zip(pathname, archive_paths)
+        return pathname
+
+    def skip_entry(self, arcname):
+        """
+        Determine whether an archive entry should be skipped when verifying
+        or installing.
+        """
+        # The signature file won't be in RECORD,
+        # and we  don't currently don't do anything with it
+        # We also skip directories, as they won't be in RECORD
+        # either. See:
+        #
+        # https://github.com/pypa/wheel/issues/294
+        # https://github.com/pypa/wheel/issues/287
+        # https://github.com/pypa/wheel/pull/289
+        #
+        return arcname.endswith(('/', '/RECORD.jws'))
+
+    def install(self, paths, maker, **kwargs):
+        """
+        Install a wheel to the specified paths. If kwarg ``warner`` is
+        specified, it should be a callable, which will be called with two
+        tuples indicating the wheel version of this software and the wheel
+        version in the file, if there is a discrepancy in the versions.
+        This can be used to issue any warnings to raise any exceptions.
+        If kwarg ``lib_only`` is True, only the purelib/platlib files are
+        installed, and the headers, scripts, data and dist-info metadata are
+        not written. If kwarg ``bytecode_hashed_invalidation`` is True, written
+        bytecode will try to use file-hash based invalidation (PEP-552) on
+        supported interpreter versions (CPython 3.7+).
+
+        The return value is a :class:`InstalledDistribution` instance unless
+        ``options.lib_only`` is True, in which case the return value is ``None``.
+        """
+
+        dry_run = maker.dry_run
+        warner = kwargs.get('warner')
+        lib_only = kwargs.get('lib_only', False)
+        bc_hashed_invalidation = kwargs.get('bytecode_hashed_invalidation', False)
+
+        pathname = os.path.join(self.dirname, self.filename)
+        name_ver = '%s-%s' % (self.name, self.version)
+        data_dir = '%s.data' % name_ver
+        info_dir = '%s.dist-info' % name_ver
+
+        metadata_name = posixpath.join(info_dir, LEGACY_METADATA_FILENAME)
+        wheel_metadata_name = posixpath.join(info_dir, 'WHEEL')
+        record_name = posixpath.join(info_dir, 'RECORD')
+
+        wrapper = codecs.getreader('utf-8')
+
+        with ZipFile(pathname, 'r') as zf:
+            with zf.open(wheel_metadata_name) as bwf:
+                wf = wrapper(bwf)
+                message = message_from_file(wf)
+            wv = message['Wheel-Version'].split('.', 1)
+            file_version = tuple([int(i) for i in wv])
+            if (file_version != self.wheel_version) and warner:
+                warner(self.wheel_version, file_version)
+
+            if message['Root-Is-Purelib'] == 'true':
+                libdir = paths['purelib']
+            else:
+                libdir = paths['platlib']
+
+            records = {}
+            with zf.open(record_name) as bf:
+                with CSVReader(stream=bf) as reader:
+                    for row in reader:
+                        p = row[0]
+                        records[p] = row
+
+            data_pfx = posixpath.join(data_dir, '')
+            info_pfx = posixpath.join(info_dir, '')
+            script_pfx = posixpath.join(data_dir, 'scripts', '')
+
+            # make a new instance rather than a copy of maker's,
+            # as we mutate it
+            fileop = FileOperator(dry_run=dry_run)
+            fileop.record = True  # so we can rollback if needed
+
+            bc = not sys.dont_write_bytecode  # Double negatives. Lovely!
+
+            outfiles = []  # for RECORD writing
+
+            # for script copying/shebang processing
+            workdir = tempfile.mkdtemp()
+            # set target dir later
+            # we default add_launchers to False, as the
+            # Python Launcher should be used instead
+            maker.source_dir = workdir
+            maker.target_dir = None
+            try:
+                for zinfo in zf.infolist():
+                    arcname = zinfo.filename
+                    if isinstance(arcname, text_type):
+                        u_arcname = arcname
+                    else:
+                        u_arcname = arcname.decode('utf-8')
+                    if self.skip_entry(u_arcname):
+                        continue
+                    row = records[u_arcname]
+                    if row[2] and str(zinfo.file_size) != row[2]:
+                        raise DistlibException('size mismatch for '
+                                               '%s' % u_arcname)
+                    if row[1]:
+                        kind, value = row[1].split('=', 1)
+                        with zf.open(arcname) as bf:
+                            data = bf.read()
+                        _, digest = self.get_hash(data, kind)
+                        if digest != value:
+                            raise DistlibException('digest mismatch for '
+                                                   '%s' % arcname)
+
+                    if lib_only and u_arcname.startswith((info_pfx, data_pfx)):
+                        logger.debug('lib_only: skipping %s', u_arcname)
+                        continue
+                    is_script = (u_arcname.startswith(script_pfx) and not u_arcname.endswith('.exe'))
+
+                    if u_arcname.startswith(data_pfx):
+                        _, where, rp = u_arcname.split('/', 2)
+                        outfile = os.path.join(paths[where], convert_path(rp))
+                    else:
+                        # meant for site-packages.
+                        if u_arcname in (wheel_metadata_name, record_name):
+                            continue
+                        outfile = os.path.join(libdir, convert_path(u_arcname))
+                    if not is_script:
+                        with zf.open(arcname) as bf:
+                            fileop.copy_stream(bf, outfile)
+                        # Issue #147: permission bits aren't preserved. Using
+                        # zf.extract(zinfo, libdir) should have worked, but didn't,
+                        # see https://www.thetopsites.net/article/53834422.shtml
+                        # So ... manually preserve permission bits as given in zinfo
+                        if os.name == 'posix':
+                            # just set the normal permission bits
+                            os.chmod(outfile, (zinfo.external_attr >> 16) & 0x1FF)
+                        outfiles.append(outfile)
+                        # Double check the digest of the written file
+                        if not dry_run and row[1]:
+                            with open(outfile, 'rb') as bf:
+                                data = bf.read()
+                                _, newdigest = self.get_hash(data, kind)
+                                if newdigest != digest:
+                                    raise DistlibException('digest mismatch '
+                                                           'on write for '
+                                                           '%s' % outfile)
+                        if bc and outfile.endswith('.py'):
+                            try:
+                                pyc = fileop.byte_compile(outfile, hashed_invalidation=bc_hashed_invalidation)
+                                outfiles.append(pyc)
+                            except Exception:
+                                # Don't give up if byte-compilation fails,
+                                # but log it and perhaps warn the user
+                                logger.warning('Byte-compilation failed', exc_info=True)
+                    else:
+                        fn = os.path.basename(convert_path(arcname))
+                        workname = os.path.join(workdir, fn)
+                        with zf.open(arcname) as bf:
+                            fileop.copy_stream(bf, workname)
+
+                        dn, fn = os.path.split(outfile)
+                        maker.target_dir = dn
+                        filenames = maker.make(fn)
+                        fileop.set_executable_mode(filenames)
+                        outfiles.extend(filenames)
+
+                if lib_only:
+                    logger.debug('lib_only: returning None')
+                    dist = None
+                else:
+                    # Generate scripts
+
+                    # Try to get pydist.json so we can see if there are
+                    # any commands to generate. If this fails (e.g. because
+                    # of a legacy wheel), log a warning but don't give up.
+                    commands = None
+                    file_version = self.info['Wheel-Version']
+                    if file_version == '1.0':
+                        # Use legacy info
+                        ep = posixpath.join(info_dir, 'entry_points.txt')
+                        try:
+                            with zf.open(ep) as bwf:
+                                epdata = read_exports(bwf)
+                            commands = {}
+                            for key in ('console', 'gui'):
+                                k = '%s_scripts' % key
+                                if k in epdata:
+                                    commands['wrap_%s' % key] = d = {}
+                                    for v in epdata[k].values():
+                                        s = '%s:%s' % (v.prefix, v.suffix)
+                                        if v.flags:
+                                            s += ' [%s]' % ','.join(v.flags)
+                                        d[v.name] = s
+                        except Exception:
+                            logger.warning('Unable to read legacy script '
+                                           'metadata, so cannot generate '
+                                           'scripts')
+                    else:
+                        try:
+                            with zf.open(metadata_name) as bwf:
+                                wf = wrapper(bwf)
+                                commands = json.load(wf).get('extensions')
+                                if commands:
+                                    commands = commands.get('python.commands')
+                        except Exception:
+                            logger.warning('Unable to read JSON metadata, so '
+                                           'cannot generate scripts')
+                    if commands:
+                        console_scripts = commands.get('wrap_console', {})
+                        gui_scripts = commands.get('wrap_gui', {})
+                        if console_scripts or gui_scripts:
+                            script_dir = paths.get('scripts', '')
+                            if not os.path.isdir(script_dir):
+                                raise ValueError('Valid script path not '
+                                                 'specified')
+                            maker.target_dir = script_dir
+                            for k, v in console_scripts.items():
+                                script = '%s = %s' % (k, v)
+                                filenames = maker.make(script)
+                                fileop.set_executable_mode(filenames)
+
+                            if gui_scripts:
+                                options = {'gui': True}
+                                for k, v in gui_scripts.items():
+                                    script = '%s = %s' % (k, v)
+                                    filenames = maker.make(script, options)
+                                    fileop.set_executable_mode(filenames)
+
+                    p = os.path.join(libdir, info_dir)
+                    dist = InstalledDistribution(p)
+
+                    # Write SHARED
+                    paths = dict(paths)  # don't change passed in dict
+                    del paths['purelib']
+                    del paths['platlib']
+                    paths['lib'] = libdir
+                    p = dist.write_shared_locations(paths, dry_run)
+                    if p:
+                        outfiles.append(p)
+
+                    # Write RECORD
+                    dist.write_installed_files(outfiles, paths['prefix'], dry_run)
+                return dist
+            except Exception:  # pragma: no cover
+                logger.exception('installation failed.')
+                fileop.rollback()
+                raise
+            finally:
+                shutil.rmtree(workdir)
+
+    def _get_dylib_cache(self):
+        global cache
+        if cache is None:
+            # Use native string to avoid issues on 2.x: see Python #20140.
+            base = os.path.join(get_cache_base(), str('dylib-cache'), '%s.%s' % sys.version_info[:2])
+            cache = Cache(base)
+        return cache
+
+    def _get_extensions(self):
+        pathname = os.path.join(self.dirname, self.filename)
+        name_ver = '%s-%s' % (self.name, self.version)
+        info_dir = '%s.dist-info' % name_ver
+        arcname = posixpath.join(info_dir, 'EXTENSIONS')
+        wrapper = codecs.getreader('utf-8')
+        result = []
+        with ZipFile(pathname, 'r') as zf:
+            try:
+                with zf.open(arcname) as bf:
+                    wf = wrapper(bf)
+                    extensions = json.load(wf)
+                    cache = self._get_dylib_cache()
+                    prefix = cache.prefix_to_dir(self.filename, use_abspath=False)
+                    cache_base = os.path.join(cache.base, prefix)
+                    if not os.path.isdir(cache_base):
+                        os.makedirs(cache_base)
+                    for name, relpath in extensions.items():
+                        dest = os.path.join(cache_base, convert_path(relpath))
+                        if not os.path.exists(dest):
+                            extract = True
+                        else:
+                            file_time = os.stat(dest).st_mtime
+                            file_time = datetime.datetime.fromtimestamp(file_time)
+                            info = zf.getinfo(relpath)
+                            wheel_time = datetime.datetime(*info.date_time)
+                            extract = wheel_time > file_time
+                        if extract:
+                            zf.extract(relpath, cache_base)
+                        result.append((name, dest))
+            except KeyError:
+                pass
+        return result
+
+    def is_compatible(self):
+        """
+        Determine if a wheel is compatible with the running system.
+        """
+        return is_compatible(self)
+
+    def is_mountable(self):
+        """
+        Determine if a wheel is asserted as mountable by its metadata.
+        """
+        return True  # for now - metadata details TBD
+
+    def mount(self, append=False):
+        pathname = os.path.abspath(os.path.join(self.dirname, self.filename))
+        if not self.is_compatible():
+            msg = 'Wheel %s not compatible with this Python.' % pathname
+            raise DistlibException(msg)
+        if not self.is_mountable():
+            msg = 'Wheel %s is marked as not mountable.' % pathname
+            raise DistlibException(msg)
+        if pathname in sys.path:
+            logger.debug('%s already in path', pathname)
+        else:
+            if append:
+                sys.path.append(pathname)
+            else:
+                sys.path.insert(0, pathname)
+            extensions = self._get_extensions()
+            if extensions:
+                if _hook not in sys.meta_path:
+                    sys.meta_path.append(_hook)
+                _hook.add(pathname, extensions)
+
+    def unmount(self):
+        pathname = os.path.abspath(os.path.join(self.dirname, self.filename))
+        if pathname not in sys.path:
+            logger.debug('%s not in path', pathname)
+        else:
+            sys.path.remove(pathname)
+            if pathname in _hook.impure_wheels:
+                _hook.remove(pathname)
+            if not _hook.impure_wheels:
+                if _hook in sys.meta_path:
+                    sys.meta_path.remove(_hook)
+
+    def verify(self):
+        pathname = os.path.join(self.dirname, self.filename)
+        name_ver = '%s-%s' % (self.name, self.version)
+        # data_dir = '%s.data' % name_ver
+        info_dir = '%s.dist-info' % name_ver
+
+        # metadata_name = posixpath.join(info_dir, LEGACY_METADATA_FILENAME)
+        wheel_metadata_name = posixpath.join(info_dir, 'WHEEL')
+        record_name = posixpath.join(info_dir, 'RECORD')
+
+        wrapper = codecs.getreader('utf-8')
+
+        with ZipFile(pathname, 'r') as zf:
+            with zf.open(wheel_metadata_name) as bwf:
+                wf = wrapper(bwf)
+                message_from_file(wf)
+            # wv = message['Wheel-Version'].split('.', 1)
+            # file_version = tuple([int(i) for i in wv])
+            # TODO version verification
+
+            records = {}
+            with zf.open(record_name) as bf:
+                with CSVReader(stream=bf) as reader:
+                    for row in reader:
+                        p = row[0]
+                        records[p] = row
+
+            for zinfo in zf.infolist():
+                arcname = zinfo.filename
+                if isinstance(arcname, text_type):
+                    u_arcname = arcname
+                else:
+                    u_arcname = arcname.decode('utf-8')
+                # See issue #115: some wheels have .. in their entries, but
+                # in the filename ... e.g. __main__..py ! So the check is
+                # updated to look for .. in the directory portions
+                p = u_arcname.split('/')
+                if '..' in p:
+                    raise DistlibException('invalid entry in '
+                                           'wheel: %r' % u_arcname)
+
+                if self.skip_entry(u_arcname):
+                    continue
+                row = records[u_arcname]
+                if row[2] and str(zinfo.file_size) != row[2]:
+                    raise DistlibException('size mismatch for '
+                                           '%s' % u_arcname)
+                if row[1]:
+                    kind, value = row[1].split('=', 1)
+                    with zf.open(arcname) as bf:
+                        data = bf.read()
+                    _, digest = self.get_hash(data, kind)
+                    if digest != value:
+                        raise DistlibException('digest mismatch for '
+                                               '%s' % arcname)
+
+    def update(self, modifier, dest_dir=None, **kwargs):
+        """
+        Update the contents of a wheel in a generic way. The modifier should
+        be a callable which expects a dictionary argument: its keys are
+        archive-entry paths, and its values are absolute filesystem paths
+        where the contents the corresponding archive entries can be found. The
+        modifier is free to change the contents of the files pointed to, add
+        new entries and remove entries, before returning. This method will
+        extract the entire contents of the wheel to a temporary location, call
+        the modifier, and then use the passed (and possibly updated)
+        dictionary to write a new wheel. If ``dest_dir`` is specified, the new
+        wheel is written there -- otherwise, the original wheel is overwritten.
+
+        The modifier should return True if it updated the wheel, else False.
+        This method returns the same value the modifier returns.
+        """
+
+        def get_version(path_map, info_dir):
+            version = path = None
+            key = '%s/%s' % (info_dir, LEGACY_METADATA_FILENAME)
+            if key not in path_map:
+                key = '%s/PKG-INFO' % info_dir
+            if key in path_map:
+                path = path_map[key]
+                version = Metadata(path=path).version
+            return version, path
+
+        def update_version(version, path):
+            updated = None
+            try:
+                NormalizedVersion(version)
+                i = version.find('-')
+                if i < 0:
+                    updated = '%s+1' % version
+                else:
+                    parts = [int(s) for s in version[i + 1:].split('.')]
+                    parts[-1] += 1
+                    updated = '%s+%s' % (version[:i], '.'.join(str(i) for i in parts))
+            except UnsupportedVersionError:
+                logger.debug('Cannot update non-compliant (PEP-440) '
+                             'version %r', version)
+            if updated:
+                md = Metadata(path=path)
+                md.version = updated
+                legacy = path.endswith(LEGACY_METADATA_FILENAME)
+                md.write(path=path, legacy=legacy)
+                logger.debug('Version updated from %r to %r', version, updated)
+
+        pathname = os.path.join(self.dirname, self.filename)
+        name_ver = '%s-%s' % (self.name, self.version)
+        info_dir = '%s.dist-info' % name_ver
+        record_name = posixpath.join(info_dir, 'RECORD')
+        with tempdir() as workdir:
+            with ZipFile(pathname, 'r') as zf:
+                path_map = {}
+                for zinfo in zf.infolist():
+                    arcname = zinfo.filename
+                    if isinstance(arcname, text_type):
+                        u_arcname = arcname
+                    else:
+                        u_arcname = arcname.decode('utf-8')
+                    if u_arcname == record_name:
+                        continue
+                    if '..' in u_arcname:
+                        raise DistlibException('invalid entry in '
+                                               'wheel: %r' % u_arcname)
+                    zf.extract(zinfo, workdir)
+                    path = os.path.join(workdir, convert_path(u_arcname))
+                    path_map[u_arcname] = path
+
+            # Remember the version.
+            original_version, _ = get_version(path_map, info_dir)
+            # Files extracted. Call the modifier.
+            modified = modifier(path_map, **kwargs)
+            if modified:
+                # Something changed - need to build a new wheel.
+                current_version, path = get_version(path_map, info_dir)
+                if current_version and (current_version == original_version):
+                    # Add or update local version to signify changes.
+                    update_version(current_version, path)
+                # Decide where the new wheel goes.
+                if dest_dir is None:
+                    fd, newpath = tempfile.mkstemp(suffix='.whl', prefix='wheel-update-', dir=workdir)
+                    os.close(fd)
+                else:
+                    if not os.path.isdir(dest_dir):
+                        raise DistlibException('Not a directory: %r' % dest_dir)
+                    newpath = os.path.join(dest_dir, self.filename)
+                archive_paths = list(path_map.items())
+                distinfo = os.path.join(workdir, info_dir)
+                info = distinfo, info_dir
+                self.write_records(info, workdir, archive_paths)
+                self.build_zip(newpath, archive_paths)
+                if dest_dir is None:
+                    shutil.copyfile(newpath, pathname)
+        return modified
+
+
+def _get_glibc_version():
+    import platform
+    ver = platform.libc_ver()
+    result = []
+    if ver[0] == 'glibc':
+        for s in ver[1].split('.'):
+            result.append(int(s) if s.isdigit() else 0)
+        result = tuple(result)
+    return result
+
+
+def compatible_tags():
+    """
+    Return (pyver, abi, arch) tuples compatible with this Python.
+    """
+    class _Version:
+        def __init__(self, major, minor):
+            self.major = major
+            self.major_minor = (major, minor)
+            self.string = ''.join((str(major), str(minor)))
+
+        def __str__(self):
+            return self.string
+
+
+    versions = [
+        _Version(sys.version_info.major, minor_version)
+        for minor_version in range(sys.version_info.minor, -1, -1)
+    ]
+    abis = []
+    for suffix in _get_suffixes():
+        if suffix.startswith('.abi'):
+            abis.append(suffix.split('.', 2)[1])
+    abis.sort()
+    if ABI != 'none':
+        abis.insert(0, ABI)
+    abis.append('none')
+    result = []
+
+    arches = [ARCH]
+    if sys.platform == 'darwin':
+        m = re.match(r'(\w+)_(\d+)_(\d+)_(\w+)$', ARCH)
+        if m:
+            name, major, minor, arch = m.groups()
+            minor = int(minor)
+            matches = [arch]
+            if arch in ('i386', 'ppc'):
+                matches.append('fat')
+            if arch in ('i386', 'ppc', 'x86_64'):
+                matches.append('fat3')
+            if arch in ('ppc64', 'x86_64'):
+                matches.append('fat64')
+            if arch in ('i386', 'x86_64'):
+                matches.append('intel')
+            if arch in ('i386', 'x86_64', 'intel', 'ppc', 'ppc64'):
+                matches.append('universal')
+            while minor >= 0:
+                for match in matches:
+                    s = '%s_%s_%s_%s' % (name, major, minor, match)
+                    if s != ARCH:  # already there
+                        arches.append(s)
+                minor -= 1
+
+    # Most specific - our Python version, ABI and arch
+    for i, version_object in enumerate(versions):
+        version = str(version_object)
+        add_abis = []
+
+        if i == 0:
+            add_abis = abis
+
+        if IMP_PREFIX == 'cp' and version_object.major_minor >= (3, 2):
+            limited_api_abi = 'abi' + str(version_object.major)
+            if limited_api_abi not in add_abis:
+                add_abis.append(limited_api_abi)
+
+        for abi in add_abis:
+            for arch in arches:
+                result.append((''.join((IMP_PREFIX, version)), abi, arch))
+                # manylinux
+                if abi != 'none' and sys.platform.startswith('linux'):
+                    arch = arch.replace('linux_', '')
+                    parts = _get_glibc_version()
+                    if len(parts) == 2:
+                        if parts >= (2, 5):
+                            result.append((''.join((IMP_PREFIX, version)), abi, 'manylinux1_%s' % arch))
+                        if parts >= (2, 12):
+                            result.append((''.join((IMP_PREFIX, version)), abi, 'manylinux2010_%s' % arch))
+                        if parts >= (2, 17):
+                            result.append((''.join((IMP_PREFIX, version)), abi, 'manylinux2014_%s' % arch))
+                        result.append((''.join(
+                            (IMP_PREFIX, version)), abi, 'manylinux_%s_%s_%s' % (parts[0], parts[1], arch)))
+
+    # where no ABI / arch dependency, but IMP_PREFIX dependency
+    for i, version_object in enumerate(versions):
+        version = str(version_object)
+        result.append((''.join((IMP_PREFIX, version)), 'none', 'any'))
+        if i == 0:
+            result.append((''.join((IMP_PREFIX, version[0])), 'none', 'any'))
+
+    # no IMP_PREFIX, ABI or arch dependency
+    for i, version_object in enumerate(versions):
+        version = str(version_object)
+        result.append((''.join(('py', version)), 'none', 'any'))
+        if i == 0:
+            result.append((''.join(('py', version[0])), 'none', 'any'))
+
+    return set(result)
+
+
+COMPATIBLE_TAGS = compatible_tags()
+
+del compatible_tags
+
+
+def is_compatible(wheel, tags=None):
+    if not isinstance(wheel, Wheel):
+        wheel = Wheel(wheel)  # assume it's a filename
+    result = False
+    if tags is None:
+        tags = COMPATIBLE_TAGS
+    for ver, abi, arch in tags:
+        if ver in wheel.pyver and abi in wheel.abi and arch in wheel.arch:
+            result = True
+            break
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__init__.py
new file mode 100644
index 00000000..7686fe85
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__init__.py
@@ -0,0 +1,54 @@
+from .distro import (
+    NORMALIZED_DISTRO_ID,
+    NORMALIZED_LSB_ID,
+    NORMALIZED_OS_ID,
+    LinuxDistribution,
+    __version__,
+    build_number,
+    codename,
+    distro_release_attr,
+    distro_release_info,
+    id,
+    info,
+    like,
+    linux_distribution,
+    lsb_release_attr,
+    lsb_release_info,
+    major_version,
+    minor_version,
+    name,
+    os_release_attr,
+    os_release_info,
+    uname_attr,
+    uname_info,
+    version,
+    version_parts,
+)
+
+__all__ = [
+    "NORMALIZED_DISTRO_ID",
+    "NORMALIZED_LSB_ID",
+    "NORMALIZED_OS_ID",
+    "LinuxDistribution",
+    "build_number",
+    "codename",
+    "distro_release_attr",
+    "distro_release_info",
+    "id",
+    "info",
+    "like",
+    "linux_distribution",
+    "lsb_release_attr",
+    "lsb_release_info",
+    "major_version",
+    "minor_version",
+    "name",
+    "os_release_attr",
+    "os_release_info",
+    "uname_attr",
+    "uname_info",
+    "version",
+    "version_parts",
+]
+
+__version__ = __version__
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__main__.py
new file mode 100644
index 00000000..0c01d5b0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__main__.py
@@ -0,0 +1,4 @@
+from .distro import main
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..719c7b7b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..a2df3193
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/distro.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/distro.cpython-312.pyc
new file mode 100644
index 00000000..866cd579
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/__pycache__/distro.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/distro.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/distro.py
new file mode 100644
index 00000000..78ccdfa4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/distro.py
@@ -0,0 +1,1403 @@
+#!/usr/bin/env python
+# Copyright 2015-2021 Nir Cohen
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""
+The ``distro`` package (``distro`` stands for Linux Distribution) provides
+information about the Linux distribution it runs on, such as a reliable
+machine-readable distro ID, or version information.
+
+It is the recommended replacement for Python's original
+:py:func:`platform.linux_distribution` function, but it provides much more
+functionality. An alternative implementation became necessary because Python
+3.5 deprecated this function, and Python 3.8 removed it altogether. Its
+predecessor function :py:func:`platform.dist` was already deprecated since
+Python 2.6 and removed in Python 3.8. Still, there are many cases in which
+access to OS distribution information is needed. See `Python issue 1322
+<https://bugs.python.org/issue1322>`_ for more information.
+"""
+
+import argparse
+import json
+import logging
+import os
+import re
+import shlex
+import subprocess
+import sys
+import warnings
+from typing import (
+    Any,
+    Callable,
+    Dict,
+    Iterable,
+    Optional,
+    Sequence,
+    TextIO,
+    Tuple,
+    Type,
+)
+
+try:
+    from typing import TypedDict
+except ImportError:
+    # Python 3.7
+    TypedDict = dict
+
+__version__ = "1.9.0"
+
+
+class VersionDict(TypedDict):
+    major: str
+    minor: str
+    build_number: str
+
+
+class InfoDict(TypedDict):
+    id: str
+    version: str
+    version_parts: VersionDict
+    like: str
+    codename: str
+
+
+_UNIXCONFDIR = os.environ.get("UNIXCONFDIR", "/etc")
+_UNIXUSRLIBDIR = os.environ.get("UNIXUSRLIBDIR", "/usr/lib")
+_OS_RELEASE_BASENAME = "os-release"
+
+#: Translation table for normalizing the "ID" attribute defined in os-release
+#: files, for use by the :func:`distro.id` method.
+#:
+#: * Key: Value as defined in the os-release file, translated to lower case,
+#:   with blanks translated to underscores.
+#:
+#: * Value: Normalized value.
+NORMALIZED_OS_ID = {
+    "ol": "oracle",  # Oracle Linux
+    "opensuse-leap": "opensuse",  # Newer versions of OpenSuSE report as opensuse-leap
+}
+
+#: Translation table for normalizing the "Distributor ID" attribute returned by
+#: the lsb_release command, for use by the :func:`distro.id` method.
+#:
+#: * Key: Value as returned by the lsb_release command, translated to lower
+#:   case, with blanks translated to underscores.
+#:
+#: * Value: Normalized value.
+NORMALIZED_LSB_ID = {
+    "enterpriseenterpriseas": "oracle",  # Oracle Enterprise Linux 4
+    "enterpriseenterpriseserver": "oracle",  # Oracle Linux 5
+    "redhatenterpriseworkstation": "rhel",  # RHEL 6, 7 Workstation
+    "redhatenterpriseserver": "rhel",  # RHEL 6, 7 Server
+    "redhatenterprisecomputenode": "rhel",  # RHEL 6 ComputeNode
+}
+
+#: Translation table for normalizing the distro ID derived from the file name
+#: of distro release files, for use by the :func:`distro.id` method.
+#:
+#: * Key: Value as derived from the file name of a distro release file,
+#:   translated to lower case, with blanks translated to underscores.
+#:
+#: * Value: Normalized value.
+NORMALIZED_DISTRO_ID = {
+    "redhat": "rhel",  # RHEL 6.x, 7.x
+}
+
+# Pattern for content of distro release file (reversed)
+_DISTRO_RELEASE_CONTENT_REVERSED_PATTERN = re.compile(
+    r"(?:[^)]*\)(.*)\()? *(?:STL )?([\d.+\-a-z]*\d) *(?:esaeler *)?(.+)"
+)
+
+# Pattern for base file name of distro release file
+_DISTRO_RELEASE_BASENAME_PATTERN = re.compile(r"(\w+)[-_](release|version)$")
+
+# Base file names to be looked up for if _UNIXCONFDIR is not readable.
+_DISTRO_RELEASE_BASENAMES = [
+    "SuSE-release",
+    "altlinux-release",
+    "arch-release",
+    "base-release",
+    "centos-release",
+    "fedora-release",
+    "gentoo-release",
+    "mageia-release",
+    "mandrake-release",
+    "mandriva-release",
+    "mandrivalinux-release",
+    "manjaro-release",
+    "oracle-release",
+    "redhat-release",
+    "rocky-release",
+    "sl-release",
+    "slackware-version",
+]
+
+# Base file names to be ignored when searching for distro release file
+_DISTRO_RELEASE_IGNORE_BASENAMES = (
+    "debian_version",
+    "lsb-release",
+    "oem-release",
+    _OS_RELEASE_BASENAME,
+    "system-release",
+    "plesk-release",
+    "iredmail-release",
+    "board-release",
+    "ec2_version",
+)
+
+
+def linux_distribution(full_distribution_name: bool = True) -> Tuple[str, str, str]:
+    """
+    .. deprecated:: 1.6.0
+
+        :func:`distro.linux_distribution()` is deprecated. It should only be
+        used as a compatibility shim with Python's
+        :py:func:`platform.linux_distribution()`. Please use :func:`distro.id`,
+        :func:`distro.version` and :func:`distro.name` instead.
+
+    Return information about the current OS distribution as a tuple
+    ``(id_name, version, codename)`` with items as follows:
+
+    * ``id_name``:  If *full_distribution_name* is false, the result of
+      :func:`distro.id`. Otherwise, the result of :func:`distro.name`.
+
+    * ``version``:  The result of :func:`distro.version`.
+
+    * ``codename``:  The extra item (usually in parentheses) after the
+      os-release version number, or the result of :func:`distro.codename`.
+
+    The interface of this function is compatible with the original
+    :py:func:`platform.linux_distribution` function, supporting a subset of
+    its parameters.
+
+    The data it returns may not exactly be the same, because it uses more data
+    sources than the original function, and that may lead to different data if
+    the OS distribution is not consistent across multiple data sources it
+    provides (there are indeed such distributions ...).
+
+    Another reason for differences is the fact that the :func:`distro.id`
+    method normalizes the distro ID string to a reliable machine-readable value
+    for a number of popular OS distributions.
+    """
+    warnings.warn(
+        "distro.linux_distribution() is deprecated. It should only be used as a "
+        "compatibility shim with Python's platform.linux_distribution(). Please use "
+        "distro.id(), distro.version() and distro.name() instead.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    return _distro.linux_distribution(full_distribution_name)
+
+
+def id() -> str:
+    """
+    Return the distro ID of the current distribution, as a
+    machine-readable string.
+
+    For a number of OS distributions, the returned distro ID value is
+    *reliable*, in the sense that it is documented and that it does not change
+    across releases of the distribution.
+
+    This package maintains the following reliable distro ID values:
+
+    ==============  =========================================
+    Distro ID       Distribution
+    ==============  =========================================
+    "ubuntu"        Ubuntu
+    "debian"        Debian
+    "rhel"          RedHat Enterprise Linux
+    "centos"        CentOS
+    "fedora"        Fedora
+    "sles"          SUSE Linux Enterprise Server
+    "opensuse"      openSUSE
+    "amzn"          Amazon Linux
+    "arch"          Arch Linux
+    "buildroot"     Buildroot
+    "cloudlinux"    CloudLinux OS
+    "exherbo"       Exherbo Linux
+    "gentoo"        GenToo Linux
+    "ibm_powerkvm"  IBM PowerKVM
+    "kvmibm"        KVM for IBM z Systems
+    "linuxmint"     Linux Mint
+    "mageia"        Mageia
+    "mandriva"      Mandriva Linux
+    "parallels"     Parallels
+    "pidora"        Pidora
+    "raspbian"      Raspbian
+    "oracle"        Oracle Linux (and Oracle Enterprise Linux)
+    "scientific"    Scientific Linux
+    "slackware"     Slackware
+    "xenserver"     XenServer
+    "openbsd"       OpenBSD
+    "netbsd"        NetBSD
+    "freebsd"       FreeBSD
+    "midnightbsd"   MidnightBSD
+    "rocky"         Rocky Linux
+    "aix"           AIX
+    "guix"          Guix System
+    "altlinux"      ALT Linux
+    ==============  =========================================
+
+    If you have a need to get distros for reliable IDs added into this set,
+    or if you find that the :func:`distro.id` function returns a different
+    distro ID for one of the listed distros, please create an issue in the
+    `distro issue tracker`_.
+
+    **Lookup hierarchy and transformations:**
+
+    First, the ID is obtained from the following sources, in the specified
+    order. The first available and non-empty value is used:
+
+    * the value of the "ID" attribute of the os-release file,
+
+    * the value of the "Distributor ID" attribute returned by the lsb_release
+      command,
+
+    * the first part of the file name of the distro release file,
+
+    The so determined ID value then passes the following transformations,
+    before it is returned by this method:
+
+    * it is translated to lower case,
+
+    * blanks (which should not be there anyway) are translated to underscores,
+
+    * a normalization of the ID is performed, based upon
+      `normalization tables`_. The purpose of this normalization is to ensure
+      that the ID is as reliable as possible, even across incompatible changes
+      in the OS distributions. A common reason for an incompatible change is
+      the addition of an os-release file, or the addition of the lsb_release
+      command, with ID values that differ from what was previously determined
+      from the distro release file name.
+    """
+    return _distro.id()
+
+
+def name(pretty: bool = False) -> str:
+    """
+    Return the name of the current OS distribution, as a human-readable
+    string.
+
+    If *pretty* is false, the name is returned without version or codename.
+    (e.g. "CentOS Linux")
+
+    If *pretty* is true, the version and codename are appended.
+    (e.g. "CentOS Linux 7.1.1503 (Core)")
+
+    **Lookup hierarchy:**
+
+    The name is obtained from the following sources, in the specified order.
+    The first available and non-empty value is used:
+
+    * If *pretty* is false:
+
+      - the value of the "NAME" attribute of the os-release file,
+
+      - the value of the "Distributor ID" attribute returned by the lsb_release
+        command,
+
+      - the value of the "<name>" field of the distro release file.
+
+    * If *pretty* is true:
+
+      - the value of the "PRETTY_NAME" attribute of the os-release file,
+
+      - the value of the "Description" attribute returned by the lsb_release
+        command,
+
+      - the value of the "<name>" field of the distro release file, appended
+        with the value of the pretty version ("<version_id>" and "<codename>"
+        fields) of the distro release file, if available.
+    """
+    return _distro.name(pretty)
+
+
+def version(pretty: bool = False, best: bool = False) -> str:
+    """
+    Return the version of the current OS distribution, as a human-readable
+    string.
+
+    If *pretty* is false, the version is returned without codename (e.g.
+    "7.0").
+
+    If *pretty* is true, the codename in parenthesis is appended, if the
+    codename is non-empty (e.g. "7.0 (Maipo)").
+
+    Some distributions provide version numbers with different precisions in
+    the different sources of distribution information. Examining the different
+    sources in a fixed priority order does not always yield the most precise
+    version (e.g. for Debian 8.2, or CentOS 7.1).
+
+    Some other distributions may not provide this kind of information. In these
+    cases, an empty string would be returned. This behavior can be observed
+    with rolling releases distributions (e.g. Arch Linux).
+
+    The *best* parameter can be used to control the approach for the returned
+    version:
+
+    If *best* is false, the first non-empty version number in priority order of
+    the examined sources is returned.
+
+    If *best* is true, the most precise version number out of all examined
+    sources is returned.
+
+    **Lookup hierarchy:**
+
+    In all cases, the version number is obtained from the following sources.
+    If *best* is false, this order represents the priority order:
+
+    * the value of the "VERSION_ID" attribute of the os-release file,
+    * the value of the "Release" attribute returned by the lsb_release
+      command,
+    * the version number parsed from the "<version_id>" field of the first line
+      of the distro release file,
+    * the version number parsed from the "PRETTY_NAME" attribute of the
+      os-release file, if it follows the format of the distro release files.
+    * the version number parsed from the "Description" attribute returned by
+      the lsb_release command, if it follows the format of the distro release
+      files.
+    """
+    return _distro.version(pretty, best)
+
+
+def version_parts(best: bool = False) -> Tuple[str, str, str]:
+    """
+    Return the version of the current OS distribution as a tuple
+    ``(major, minor, build_number)`` with items as follows:
+
+    * ``major``:  The result of :func:`distro.major_version`.
+
+    * ``minor``:  The result of :func:`distro.minor_version`.
+
+    * ``build_number``:  The result of :func:`distro.build_number`.
+
+    For a description of the *best* parameter, see the :func:`distro.version`
+    method.
+    """
+    return _distro.version_parts(best)
+
+
+def major_version(best: bool = False) -> str:
+    """
+    Return the major version of the current OS distribution, as a string,
+    if provided.
+    Otherwise, the empty string is returned. The major version is the first
+    part of the dot-separated version string.
+
+    For a description of the *best* parameter, see the :func:`distro.version`
+    method.
+    """
+    return _distro.major_version(best)
+
+
+def minor_version(best: bool = False) -> str:
+    """
+    Return the minor version of the current OS distribution, as a string,
+    if provided.
+    Otherwise, the empty string is returned. The minor version is the second
+    part of the dot-separated version string.
+
+    For a description of the *best* parameter, see the :func:`distro.version`
+    method.
+    """
+    return _distro.minor_version(best)
+
+
+def build_number(best: bool = False) -> str:
+    """
+    Return the build number of the current OS distribution, as a string,
+    if provided.
+    Otherwise, the empty string is returned. The build number is the third part
+    of the dot-separated version string.
+
+    For a description of the *best* parameter, see the :func:`distro.version`
+    method.
+    """
+    return _distro.build_number(best)
+
+
+def like() -> str:
+    """
+    Return a space-separated list of distro IDs of distributions that are
+    closely related to the current OS distribution in regards to packaging
+    and programming interfaces, for example distributions the current
+    distribution is a derivative from.
+
+    **Lookup hierarchy:**
+
+    This information item is only provided by the os-release file.
+    For details, see the description of the "ID_LIKE" attribute in the
+    `os-release man page
+    <http://www.freedesktop.org/software/systemd/man/os-release.html>`_.
+    """
+    return _distro.like()
+
+
+def codename() -> str:
+    """
+    Return the codename for the release of the current OS distribution,
+    as a string.
+
+    If the distribution does not have a codename, an empty string is returned.
+
+    Note that the returned codename is not always really a codename. For
+    example, openSUSE returns "x86_64". This function does not handle such
+    cases in any special way and just returns the string it finds, if any.
+
+    **Lookup hierarchy:**
+
+    * the codename within the "VERSION" attribute of the os-release file, if
+      provided,
+
+    * the value of the "Codename" attribute returned by the lsb_release
+      command,
+
+    * the value of the "<codename>" field of the distro release file.
+    """
+    return _distro.codename()
+
+
+def info(pretty: bool = False, best: bool = False) -> InfoDict:
+    """
+    Return certain machine-readable information items about the current OS
+    distribution in a dictionary, as shown in the following example:
+
+    .. sourcecode:: python
+
+        {
+            'id': 'rhel',
+            'version': '7.0',
+            'version_parts': {
+                'major': '7',
+                'minor': '0',
+                'build_number': ''
+            },
+            'like': 'fedora',
+            'codename': 'Maipo'
+        }
+
+    The dictionary structure and keys are always the same, regardless of which
+    information items are available in the underlying data sources. The values
+    for the various keys are as follows:
+
+    * ``id``:  The result of :func:`distro.id`.
+
+    * ``version``:  The result of :func:`distro.version`.
+
+    * ``version_parts -> major``:  The result of :func:`distro.major_version`.
+
+    * ``version_parts -> minor``:  The result of :func:`distro.minor_version`.
+
+    * ``version_parts -> build_number``:  The result of
+      :func:`distro.build_number`.
+
+    * ``like``:  The result of :func:`distro.like`.
+
+    * ``codename``:  The result of :func:`distro.codename`.
+
+    For a description of the *pretty* and *best* parameters, see the
+    :func:`distro.version` method.
+    """
+    return _distro.info(pretty, best)
+
+
+def os_release_info() -> Dict[str, str]:
+    """
+    Return a dictionary containing key-value pairs for the information items
+    from the os-release file data source of the current OS distribution.
+
+    See `os-release file`_ for details about these information items.
+    """
+    return _distro.os_release_info()
+
+
+def lsb_release_info() -> Dict[str, str]:
+    """
+    Return a dictionary containing key-value pairs for the information items
+    from the lsb_release command data source of the current OS distribution.
+
+    See `lsb_release command output`_ for details about these information
+    items.
+    """
+    return _distro.lsb_release_info()
+
+
+def distro_release_info() -> Dict[str, str]:
+    """
+    Return a dictionary containing key-value pairs for the information items
+    from the distro release file data source of the current OS distribution.
+
+    See `distro release file`_ for details about these information items.
+    """
+    return _distro.distro_release_info()
+
+
+def uname_info() -> Dict[str, str]:
+    """
+    Return a dictionary containing key-value pairs for the information items
+    from the distro release file data source of the current OS distribution.
+    """
+    return _distro.uname_info()
+
+
+def os_release_attr(attribute: str) -> str:
+    """
+    Return a single named information item from the os-release file data source
+    of the current OS distribution.
+
+    Parameters:
+
+    * ``attribute`` (string): Key of the information item.
+
+    Returns:
+
+    * (string): Value of the information item, if the item exists.
+      The empty string, if the item does not exist.
+
+    See `os-release file`_ for details about these information items.
+    """
+    return _distro.os_release_attr(attribute)
+
+
+def lsb_release_attr(attribute: str) -> str:
+    """
+    Return a single named information item from the lsb_release command output
+    data source of the current OS distribution.
+
+    Parameters:
+
+    * ``attribute`` (string): Key of the information item.
+
+    Returns:
+
+    * (string): Value of the information item, if the item exists.
+      The empty string, if the item does not exist.
+
+    See `lsb_release command output`_ for details about these information
+    items.
+    """
+    return _distro.lsb_release_attr(attribute)
+
+
+def distro_release_attr(attribute: str) -> str:
+    """
+    Return a single named information item from the distro release file
+    data source of the current OS distribution.
+
+    Parameters:
+
+    * ``attribute`` (string): Key of the information item.
+
+    Returns:
+
+    * (string): Value of the information item, if the item exists.
+      The empty string, if the item does not exist.
+
+    See `distro release file`_ for details about these information items.
+    """
+    return _distro.distro_release_attr(attribute)
+
+
+def uname_attr(attribute: str) -> str:
+    """
+    Return a single named information item from the distro release file
+    data source of the current OS distribution.
+
+    Parameters:
+
+    * ``attribute`` (string): Key of the information item.
+
+    Returns:
+
+    * (string): Value of the information item, if the item exists.
+                The empty string, if the item does not exist.
+    """
+    return _distro.uname_attr(attribute)
+
+
+try:
+    from functools import cached_property
+except ImportError:
+    # Python < 3.8
+    class cached_property:  # type: ignore
+        """A version of @property which caches the value.  On access, it calls the
+        underlying function and sets the value in `__dict__` so future accesses
+        will not re-call the property.
+        """
+
+        def __init__(self, f: Callable[[Any], Any]) -> None:
+            self._fname = f.__name__
+            self._f = f
+
+        def __get__(self, obj: Any, owner: Type[Any]) -> Any:
+            assert obj is not None, f"call {self._fname} on an instance"
+            ret = obj.__dict__[self._fname] = self._f(obj)
+            return ret
+
+
+class LinuxDistribution:
+    """
+    Provides information about a OS distribution.
+
+    This package creates a private module-global instance of this class with
+    default initialization arguments, that is used by the
+    `consolidated accessor functions`_ and `single source accessor functions`_.
+    By using default initialization arguments, that module-global instance
+    returns data about the current OS distribution (i.e. the distro this
+    package runs on).
+
+    Normally, it is not necessary to create additional instances of this class.
+    However, in situations where control is needed over the exact data sources
+    that are used, instances of this class can be created with a specific
+    distro release file, or a specific os-release file, or without invoking the
+    lsb_release command.
+    """
+
+    def __init__(
+        self,
+        include_lsb: Optional[bool] = None,
+        os_release_file: str = "",
+        distro_release_file: str = "",
+        include_uname: Optional[bool] = None,
+        root_dir: Optional[str] = None,
+        include_oslevel: Optional[bool] = None,
+    ) -> None:
+        """
+        The initialization method of this class gathers information from the
+        available data sources, and stores that in private instance attributes.
+        Subsequent access to the information items uses these private instance
+        attributes, so that the data sources are read only once.
+
+        Parameters:
+
+        * ``include_lsb`` (bool): Controls whether the
+          `lsb_release command output`_ is included as a data source.
+
+          If the lsb_release command is not available in the program execution
+          path, the data source for the lsb_release command will be empty.
+
+        * ``os_release_file`` (string): The path name of the
+          `os-release file`_ that is to be used as a data source.
+
+          An empty string (the default) will cause the default path name to
+          be used (see `os-release file`_ for details).
+
+          If the specified or defaulted os-release file does not exist, the
+          data source for the os-release file will be empty.
+
+        * ``distro_release_file`` (string): The path name of the
+          `distro release file`_ that is to be used as a data source.
+
+          An empty string (the default) will cause a default search algorithm
+          to be used (see `distro release file`_ for details).
+
+          If the specified distro release file does not exist, or if no default
+          distro release file can be found, the data source for the distro
+          release file will be empty.
+
+        * ``include_uname`` (bool): Controls whether uname command output is
+          included as a data source. If the uname command is not available in
+          the program execution path the data source for the uname command will
+          be empty.
+
+        * ``root_dir`` (string): The absolute path to the root directory to use
+          to find distro-related information files. Note that ``include_*``
+          parameters must not be enabled in combination with ``root_dir``.
+
+        * ``include_oslevel`` (bool): Controls whether (AIX) oslevel command
+          output is included as a data source. If the oslevel command is not
+          available in the program execution path the data source will be
+          empty.
+
+        Public instance attributes:
+
+        * ``os_release_file`` (string): The path name of the
+          `os-release file`_ that is actually used as a data source. The
+          empty string if no distro release file is used as a data source.
+
+        * ``distro_release_file`` (string): The path name of the
+          `distro release file`_ that is actually used as a data source. The
+          empty string if no distro release file is used as a data source.
+
+        * ``include_lsb`` (bool): The result of the ``include_lsb`` parameter.
+          This controls whether the lsb information will be loaded.
+
+        * ``include_uname`` (bool): The result of the ``include_uname``
+          parameter. This controls whether the uname information will
+          be loaded.
+
+        * ``include_oslevel`` (bool): The result of the ``include_oslevel``
+          parameter. This controls whether (AIX) oslevel information will be
+          loaded.
+
+        * ``root_dir`` (string): The result of the ``root_dir`` parameter.
+          The absolute path to the root directory to use to find distro-related
+          information files.
+
+        Raises:
+
+        * :py:exc:`ValueError`: Initialization parameters combination is not
+           supported.
+
+        * :py:exc:`OSError`: Some I/O issue with an os-release file or distro
+          release file.
+
+        * :py:exc:`UnicodeError`: A data source has unexpected characters or
+          uses an unexpected encoding.
+        """
+        self.root_dir = root_dir
+        self.etc_dir = os.path.join(root_dir, "etc") if root_dir else _UNIXCONFDIR
+        self.usr_lib_dir = (
+            os.path.join(root_dir, "usr/lib") if root_dir else _UNIXUSRLIBDIR
+        )
+
+        if os_release_file:
+            self.os_release_file = os_release_file
+        else:
+            etc_dir_os_release_file = os.path.join(self.etc_dir, _OS_RELEASE_BASENAME)
+            usr_lib_os_release_file = os.path.join(
+                self.usr_lib_dir, _OS_RELEASE_BASENAME
+            )
+
+            # NOTE: The idea is to respect order **and** have it set
+            #       at all times for API backwards compatibility.
+            if os.path.isfile(etc_dir_os_release_file) or not os.path.isfile(
+                usr_lib_os_release_file
+            ):
+                self.os_release_file = etc_dir_os_release_file
+            else:
+                self.os_release_file = usr_lib_os_release_file
+
+        self.distro_release_file = distro_release_file or ""  # updated later
+
+        is_root_dir_defined = root_dir is not None
+        if is_root_dir_defined and (include_lsb or include_uname or include_oslevel):
+            raise ValueError(
+                "Including subprocess data sources from specific root_dir is disallowed"
+                " to prevent false information"
+            )
+        self.include_lsb = (
+            include_lsb if include_lsb is not None else not is_root_dir_defined
+        )
+        self.include_uname = (
+            include_uname if include_uname is not None else not is_root_dir_defined
+        )
+        self.include_oslevel = (
+            include_oslevel if include_oslevel is not None else not is_root_dir_defined
+        )
+
+    def __repr__(self) -> str:
+        """Return repr of all info"""
+        return (
+            "LinuxDistribution("
+            "os_release_file={self.os_release_file!r}, "
+            "distro_release_file={self.distro_release_file!r}, "
+            "include_lsb={self.include_lsb!r}, "
+            "include_uname={self.include_uname!r}, "
+            "include_oslevel={self.include_oslevel!r}, "
+            "root_dir={self.root_dir!r}, "
+            "_os_release_info={self._os_release_info!r}, "
+            "_lsb_release_info={self._lsb_release_info!r}, "
+            "_distro_release_info={self._distro_release_info!r}, "
+            "_uname_info={self._uname_info!r}, "
+            "_oslevel_info={self._oslevel_info!r})".format(self=self)
+        )
+
+    def linux_distribution(
+        self, full_distribution_name: bool = True
+    ) -> Tuple[str, str, str]:
+        """
+        Return information about the OS distribution that is compatible
+        with Python's :func:`platform.linux_distribution`, supporting a subset
+        of its parameters.
+
+        For details, see :func:`distro.linux_distribution`.
+        """
+        return (
+            self.name() if full_distribution_name else self.id(),
+            self.version(),
+            self._os_release_info.get("release_codename") or self.codename(),
+        )
+
+    def id(self) -> str:
+        """Return the distro ID of the OS distribution, as a string.
+
+        For details, see :func:`distro.id`.
+        """
+
+        def normalize(distro_id: str, table: Dict[str, str]) -> str:
+            distro_id = distro_id.lower().replace(" ", "_")
+            return table.get(distro_id, distro_id)
+
+        distro_id = self.os_release_attr("id")
+        if distro_id:
+            return normalize(distro_id, NORMALIZED_OS_ID)
+
+        distro_id = self.lsb_release_attr("distributor_id")
+        if distro_id:
+            return normalize(distro_id, NORMALIZED_LSB_ID)
+
+        distro_id = self.distro_release_attr("id")
+        if distro_id:
+            return normalize(distro_id, NORMALIZED_DISTRO_ID)
+
+        distro_id = self.uname_attr("id")
+        if distro_id:
+            return normalize(distro_id, NORMALIZED_DISTRO_ID)
+
+        return ""
+
+    def name(self, pretty: bool = False) -> str:
+        """
+        Return the name of the OS distribution, as a string.
+
+        For details, see :func:`distro.name`.
+        """
+        name = (
+            self.os_release_attr("name")
+            or self.lsb_release_attr("distributor_id")
+            or self.distro_release_attr("name")
+            or self.uname_attr("name")
+        )
+        if pretty:
+            name = self.os_release_attr("pretty_name") or self.lsb_release_attr(
+                "description"
+            )
+            if not name:
+                name = self.distro_release_attr("name") or self.uname_attr("name")
+                version = self.version(pretty=True)
+                if version:
+                    name = f"{name} {version}"
+        return name or ""
+
+    def version(self, pretty: bool = False, best: bool = False) -> str:
+        """
+        Return the version of the OS distribution, as a string.
+
+        For details, see :func:`distro.version`.
+        """
+        versions = [
+            self.os_release_attr("version_id"),
+            self.lsb_release_attr("release"),
+            self.distro_release_attr("version_id"),
+            self._parse_distro_release_content(self.os_release_attr("pretty_name")).get(
+                "version_id", ""
+            ),
+            self._parse_distro_release_content(
+                self.lsb_release_attr("description")
+            ).get("version_id", ""),
+            self.uname_attr("release"),
+        ]
+        if self.uname_attr("id").startswith("aix"):
+            # On AIX platforms, prefer oslevel command output.
+            versions.insert(0, self.oslevel_info())
+        elif self.id() == "debian" or "debian" in self.like().split():
+            # On Debian-like, add debian_version file content to candidates list.
+            versions.append(self._debian_version)
+        version = ""
+        if best:
+            # This algorithm uses the last version in priority order that has
+            # the best precision. If the versions are not in conflict, that
+            # does not matter; otherwise, using the last one instead of the
+            # first one might be considered a surprise.
+            for v in versions:
+                if v.count(".") > version.count(".") or version == "":
+                    version = v
+        else:
+            for v in versions:
+                if v != "":
+                    version = v
+                    break
+        if pretty and version and self.codename():
+            version = f"{version} ({self.codename()})"
+        return version
+
+    def version_parts(self, best: bool = False) -> Tuple[str, str, str]:
+        """
+        Return the version of the OS distribution, as a tuple of version
+        numbers.
+
+        For details, see :func:`distro.version_parts`.
+        """
+        version_str = self.version(best=best)
+        if version_str:
+            version_regex = re.compile(r"(\d+)\.?(\d+)?\.?(\d+)?")
+            matches = version_regex.match(version_str)
+            if matches:
+                major, minor, build_number = matches.groups()
+                return major, minor or "", build_number or ""
+        return "", "", ""
+
+    def major_version(self, best: bool = False) -> str:
+        """
+        Return the major version number of the current distribution.
+
+        For details, see :func:`distro.major_version`.
+        """
+        return self.version_parts(best)[0]
+
+    def minor_version(self, best: bool = False) -> str:
+        """
+        Return the minor version number of the current distribution.
+
+        For details, see :func:`distro.minor_version`.
+        """
+        return self.version_parts(best)[1]
+
+    def build_number(self, best: bool = False) -> str:
+        """
+        Return the build number of the current distribution.
+
+        For details, see :func:`distro.build_number`.
+        """
+        return self.version_parts(best)[2]
+
+    def like(self) -> str:
+        """
+        Return the IDs of distributions that are like the OS distribution.
+
+        For details, see :func:`distro.like`.
+        """
+        return self.os_release_attr("id_like") or ""
+
+    def codename(self) -> str:
+        """
+        Return the codename of the OS distribution.
+
+        For details, see :func:`distro.codename`.
+        """
+        try:
+            # Handle os_release specially since distros might purposefully set
+            # this to empty string to have no codename
+            return self._os_release_info["codename"]
+        except KeyError:
+            return (
+                self.lsb_release_attr("codename")
+                or self.distro_release_attr("codename")
+                or ""
+            )
+
+    def info(self, pretty: bool = False, best: bool = False) -> InfoDict:
+        """
+        Return certain machine-readable information about the OS
+        distribution.
+
+        For details, see :func:`distro.info`.
+        """
+        return InfoDict(
+            id=self.id(),
+            version=self.version(pretty, best),
+            version_parts=VersionDict(
+                major=self.major_version(best),
+                minor=self.minor_version(best),
+                build_number=self.build_number(best),
+            ),
+            like=self.like(),
+            codename=self.codename(),
+        )
+
+    def os_release_info(self) -> Dict[str, str]:
+        """
+        Return a dictionary containing key-value pairs for the information
+        items from the os-release file data source of the OS distribution.
+
+        For details, see :func:`distro.os_release_info`.
+        """
+        return self._os_release_info
+
+    def lsb_release_info(self) -> Dict[str, str]:
+        """
+        Return a dictionary containing key-value pairs for the information
+        items from the lsb_release command data source of the OS
+        distribution.
+
+        For details, see :func:`distro.lsb_release_info`.
+        """
+        return self._lsb_release_info
+
+    def distro_release_info(self) -> Dict[str, str]:
+        """
+        Return a dictionary containing key-value pairs for the information
+        items from the distro release file data source of the OS
+        distribution.
+
+        For details, see :func:`distro.distro_release_info`.
+        """
+        return self._distro_release_info
+
+    def uname_info(self) -> Dict[str, str]:
+        """
+        Return a dictionary containing key-value pairs for the information
+        items from the uname command data source of the OS distribution.
+
+        For details, see :func:`distro.uname_info`.
+        """
+        return self._uname_info
+
+    def oslevel_info(self) -> str:
+        """
+        Return AIX' oslevel command output.
+        """
+        return self._oslevel_info
+
+    def os_release_attr(self, attribute: str) -> str:
+        """
+        Return a single named information item from the os-release file data
+        source of the OS distribution.
+
+        For details, see :func:`distro.os_release_attr`.
+        """
+        return self._os_release_info.get(attribute, "")
+
+    def lsb_release_attr(self, attribute: str) -> str:
+        """
+        Return a single named information item from the lsb_release command
+        output data source of the OS distribution.
+
+        For details, see :func:`distro.lsb_release_attr`.
+        """
+        return self._lsb_release_info.get(attribute, "")
+
+    def distro_release_attr(self, attribute: str) -> str:
+        """
+        Return a single named information item from the distro release file
+        data source of the OS distribution.
+
+        For details, see :func:`distro.distro_release_attr`.
+        """
+        return self._distro_release_info.get(attribute, "")
+
+    def uname_attr(self, attribute: str) -> str:
+        """
+        Return a single named information item from the uname command
+        output data source of the OS distribution.
+
+        For details, see :func:`distro.uname_attr`.
+        """
+        return self._uname_info.get(attribute, "")
+
+    @cached_property
+    def _os_release_info(self) -> Dict[str, str]:
+        """
+        Get the information items from the specified os-release file.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        if os.path.isfile(self.os_release_file):
+            with open(self.os_release_file, encoding="utf-8") as release_file:
+                return self._parse_os_release_content(release_file)
+        return {}
+
+    @staticmethod
+    def _parse_os_release_content(lines: TextIO) -> Dict[str, str]:
+        """
+        Parse the lines of an os-release file.
+
+        Parameters:
+
+        * lines: Iterable through the lines in the os-release file.
+                 Each line must be a unicode string or a UTF-8 encoded byte
+                 string.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        props = {}
+        lexer = shlex.shlex(lines, posix=True)
+        lexer.whitespace_split = True
+
+        tokens = list(lexer)
+        for token in tokens:
+            # At this point, all shell-like parsing has been done (i.e.
+            # comments processed, quotes and backslash escape sequences
+            # processed, multi-line values assembled, trailing newlines
+            # stripped, etc.), so the tokens are now either:
+            # * variable assignments: var=value
+            # * commands or their arguments (not allowed in os-release)
+            # Ignore any tokens that are not variable assignments
+            if "=" in token:
+                k, v = token.split("=", 1)
+                props[k.lower()] = v
+
+        if "version" in props:
+            # extract release codename (if any) from version attribute
+            match = re.search(r"\((\D+)\)|,\s*(\D+)", props["version"])
+            if match:
+                release_codename = match.group(1) or match.group(2)
+                props["codename"] = props["release_codename"] = release_codename
+
+        if "version_codename" in props:
+            # os-release added a version_codename field.  Use that in
+            # preference to anything else Note that some distros purposefully
+            # do not have code names.  They should be setting
+            # version_codename=""
+            props["codename"] = props["version_codename"]
+        elif "ubuntu_codename" in props:
+            # Same as above but a non-standard field name used on older Ubuntus
+            props["codename"] = props["ubuntu_codename"]
+
+        return props
+
+    @cached_property
+    def _lsb_release_info(self) -> Dict[str, str]:
+        """
+        Get the information items from the lsb_release command output.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        if not self.include_lsb:
+            return {}
+        try:
+            cmd = ("lsb_release", "-a")
+            stdout = subprocess.check_output(cmd, stderr=subprocess.DEVNULL)
+        # Command not found or lsb_release returned error
+        except (OSError, subprocess.CalledProcessError):
+            return {}
+        content = self._to_str(stdout).splitlines()
+        return self._parse_lsb_release_content(content)
+
+    @staticmethod
+    def _parse_lsb_release_content(lines: Iterable[str]) -> Dict[str, str]:
+        """
+        Parse the output of the lsb_release command.
+
+        Parameters:
+
+        * lines: Iterable through the lines of the lsb_release output.
+                 Each line must be a unicode string or a UTF-8 encoded byte
+                 string.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        props = {}
+        for line in lines:
+            kv = line.strip("\n").split(":", 1)
+            if len(kv) != 2:
+                # Ignore lines without colon.
+                continue
+            k, v = kv
+            props.update({k.replace(" ", "_").lower(): v.strip()})
+        return props
+
+    @cached_property
+    def _uname_info(self) -> Dict[str, str]:
+        if not self.include_uname:
+            return {}
+        try:
+            cmd = ("uname", "-rs")
+            stdout = subprocess.check_output(cmd, stderr=subprocess.DEVNULL)
+        except OSError:
+            return {}
+        content = self._to_str(stdout).splitlines()
+        return self._parse_uname_content(content)
+
+    @cached_property
+    def _oslevel_info(self) -> str:
+        if not self.include_oslevel:
+            return ""
+        try:
+            stdout = subprocess.check_output("oslevel", stderr=subprocess.DEVNULL)
+        except (OSError, subprocess.CalledProcessError):
+            return ""
+        return self._to_str(stdout).strip()
+
+    @cached_property
+    def _debian_version(self) -> str:
+        try:
+            with open(
+                os.path.join(self.etc_dir, "debian_version"), encoding="ascii"
+            ) as fp:
+                return fp.readline().rstrip()
+        except FileNotFoundError:
+            return ""
+
+    @staticmethod
+    def _parse_uname_content(lines: Sequence[str]) -> Dict[str, str]:
+        if not lines:
+            return {}
+        props = {}
+        match = re.search(r"^([^\s]+)\s+([\d\.]+)", lines[0].strip())
+        if match:
+            name, version = match.groups()
+
+            # This is to prevent the Linux kernel version from
+            # appearing as the 'best' version on otherwise
+            # identifiable distributions.
+            if name == "Linux":
+                return {}
+            props["id"] = name.lower()
+            props["name"] = name
+            props["release"] = version
+        return props
+
+    @staticmethod
+    def _to_str(bytestring: bytes) -> str:
+        encoding = sys.getfilesystemencoding()
+        return bytestring.decode(encoding)
+
+    @cached_property
+    def _distro_release_info(self) -> Dict[str, str]:
+        """
+        Get the information items from the specified distro release file.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        if self.distro_release_file:
+            # If it was specified, we use it and parse what we can, even if
+            # its file name or content does not match the expected pattern.
+            distro_info = self._parse_distro_release_file(self.distro_release_file)
+            basename = os.path.basename(self.distro_release_file)
+            # The file name pattern for user-specified distro release files
+            # is somewhat more tolerant (compared to when searching for the
+            # file), because we want to use what was specified as best as
+            # possible.
+            match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
+        else:
+            try:
+                basenames = [
+                    basename
+                    for basename in os.listdir(self.etc_dir)
+                    if basename not in _DISTRO_RELEASE_IGNORE_BASENAMES
+                    and os.path.isfile(os.path.join(self.etc_dir, basename))
+                ]
+                # We sort for repeatability in cases where there are multiple
+                # distro specific files; e.g. CentOS, Oracle, Enterprise all
+                # containing `redhat-release` on top of their own.
+                basenames.sort()
+            except OSError:
+                # This may occur when /etc is not readable but we can't be
+                # sure about the *-release files. Check common entries of
+                # /etc for information. If they turn out to not be there the
+                # error is handled in `_parse_distro_release_file()`.
+                basenames = _DISTRO_RELEASE_BASENAMES
+            for basename in basenames:
+                match = _DISTRO_RELEASE_BASENAME_PATTERN.match(basename)
+                if match is None:
+                    continue
+                filepath = os.path.join(self.etc_dir, basename)
+                distro_info = self._parse_distro_release_file(filepath)
+                # The name is always present if the pattern matches.
+                if "name" not in distro_info:
+                    continue
+                self.distro_release_file = filepath
+                break
+            else:  # the loop didn't "break": no candidate.
+                return {}
+
+        if match is not None:
+            distro_info["id"] = match.group(1)
+
+        # CloudLinux < 7: manually enrich info with proper id.
+        if "cloudlinux" in distro_info.get("name", "").lower():
+            distro_info["id"] = "cloudlinux"
+
+        return distro_info
+
+    def _parse_distro_release_file(self, filepath: str) -> Dict[str, str]:
+        """
+        Parse a distro release file.
+
+        Parameters:
+
+        * filepath: Path name of the distro release file.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        try:
+            with open(filepath, encoding="utf-8") as fp:
+                # Only parse the first line. For instance, on SLES there
+                # are multiple lines. We don't want them...
+                return self._parse_distro_release_content(fp.readline())
+        except OSError:
+            # Ignore not being able to read a specific, seemingly version
+            # related file.
+            # See https://github.com/python-distro/distro/issues/162
+            return {}
+
+    @staticmethod
+    def _parse_distro_release_content(line: str) -> Dict[str, str]:
+        """
+        Parse a line from a distro release file.
+
+        Parameters:
+        * line: Line from the distro release file. Must be a unicode string
+                or a UTF-8 encoded byte string.
+
+        Returns:
+            A dictionary containing all information items.
+        """
+        matches = _DISTRO_RELEASE_CONTENT_REVERSED_PATTERN.match(line.strip()[::-1])
+        distro_info = {}
+        if matches:
+            # regexp ensures non-None
+            distro_info["name"] = matches.group(3)[::-1]
+            if matches.group(2):
+                distro_info["version_id"] = matches.group(2)[::-1]
+            if matches.group(1):
+                distro_info["codename"] = matches.group(1)[::-1]
+        elif line:
+            distro_info["name"] = line.strip()
+        return distro_info
+
+
+_distro = LinuxDistribution()
+
+
+def main() -> None:
+    logger = logging.getLogger(__name__)
+    logger.setLevel(logging.DEBUG)
+    logger.addHandler(logging.StreamHandler(sys.stdout))
+
+    parser = argparse.ArgumentParser(description="OS distro info tool")
+    parser.add_argument(
+        "--json", "-j", help="Output in machine readable format", action="store_true"
+    )
+
+    parser.add_argument(
+        "--root-dir",
+        "-r",
+        type=str,
+        dest="root_dir",
+        help="Path to the root filesystem directory (defaults to /)",
+    )
+
+    args = parser.parse_args()
+
+    if args.root_dir:
+        dist = LinuxDistribution(
+            include_lsb=False,
+            include_uname=False,
+            include_oslevel=False,
+            root_dir=args.root_dir,
+        )
+    else:
+        dist = _distro
+
+    if args.json:
+        logger.info(json.dumps(dist.info(), indent=4, sort_keys=True))
+    else:
+        logger.info("Name: %s", dist.name(pretty=True))
+        distribution_version = dist.version(pretty=True)
+        logger.info("Version: %s", distribution_version)
+        distribution_codename = dist.codename()
+        logger.info("Codename: %s", distribution_codename)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/distro/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__init__.py
new file mode 100644
index 00000000..cfdc030a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__init__.py
@@ -0,0 +1,45 @@
+from .core import (
+    IDNABidiError,
+    IDNAError,
+    InvalidCodepoint,
+    InvalidCodepointContext,
+    alabel,
+    check_bidi,
+    check_hyphen_ok,
+    check_initial_combiner,
+    check_label,
+    check_nfc,
+    decode,
+    encode,
+    ulabel,
+    uts46_remap,
+    valid_contextj,
+    valid_contexto,
+    valid_label_length,
+    valid_string_length,
+)
+from .intranges import intranges_contain
+from .package_data import __version__
+
+__all__ = [
+    "__version__",
+    "IDNABidiError",
+    "IDNAError",
+    "InvalidCodepoint",
+    "InvalidCodepointContext",
+    "alabel",
+    "check_bidi",
+    "check_hyphen_ok",
+    "check_initial_combiner",
+    "check_label",
+    "check_nfc",
+    "decode",
+    "encode",
+    "intranges_contain",
+    "ulabel",
+    "uts46_remap",
+    "valid_contextj",
+    "valid_contexto",
+    "valid_label_length",
+    "valid_string_length",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9292b892
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/codec.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/codec.cpython-312.pyc
new file mode 100644
index 00000000..d474a646
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/codec.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..3734717d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..01f0d920
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/idnadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/idnadata.cpython-312.pyc
new file mode 100644
index 00000000..aa2712d6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/idnadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/intranges.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/intranges.cpython-312.pyc
new file mode 100644
index 00000000..ae4c8dde
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/intranges.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/package_data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/package_data.cpython-312.pyc
new file mode 100644
index 00000000..420b945f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/package_data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/uts46data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/uts46data.cpython-312.pyc
new file mode 100644
index 00000000..8887c033
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/__pycache__/uts46data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/codec.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/codec.py
new file mode 100644
index 00000000..913abfd6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/codec.py
@@ -0,0 +1,122 @@
+import codecs
+import re
+from typing import Any, Optional, Tuple
+
+from .core import IDNAError, alabel, decode, encode, ulabel
+
+_unicode_dots_re = re.compile("[\u002e\u3002\uff0e\uff61]")
+
+
+class Codec(codecs.Codec):
+    def encode(self, data: str, errors: str = "strict") -> Tuple[bytes, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return b"", 0
+
+        return encode(data), len(data)
+
+    def decode(self, data: bytes, errors: str = "strict") -> Tuple[str, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return "", 0
+
+        return decode(data), len(data)
+
+
+class IncrementalEncoder(codecs.BufferedIncrementalEncoder):
+    def _buffer_encode(self, data: str, errors: str, final: bool) -> Tuple[bytes, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return b"", 0
+
+        labels = _unicode_dots_re.split(data)
+        trailing_dot = b""
+        if labels:
+            if not labels[-1]:
+                trailing_dot = b"."
+                del labels[-1]
+            elif not final:
+                # Keep potentially unfinished label until the next call
+                del labels[-1]
+                if labels:
+                    trailing_dot = b"."
+
+        result = []
+        size = 0
+        for label in labels:
+            result.append(alabel(label))
+            if size:
+                size += 1
+            size += len(label)
+
+        # Join with U+002E
+        result_bytes = b".".join(result) + trailing_dot
+        size += len(trailing_dot)
+        return result_bytes, size
+
+
+class IncrementalDecoder(codecs.BufferedIncrementalDecoder):
+    def _buffer_decode(self, data: Any, errors: str, final: bool) -> Tuple[str, int]:
+        if errors != "strict":
+            raise IDNAError('Unsupported error handling "{}"'.format(errors))
+
+        if not data:
+            return ("", 0)
+
+        if not isinstance(data, str):
+            data = str(data, "ascii")
+
+        labels = _unicode_dots_re.split(data)
+        trailing_dot = ""
+        if labels:
+            if not labels[-1]:
+                trailing_dot = "."
+                del labels[-1]
+            elif not final:
+                # Keep potentially unfinished label until the next call
+                del labels[-1]
+                if labels:
+                    trailing_dot = "."
+
+        result = []
+        size = 0
+        for label in labels:
+            result.append(ulabel(label))
+            if size:
+                size += 1
+            size += len(label)
+
+        result_str = ".".join(result) + trailing_dot
+        size += len(trailing_dot)
+        return (result_str, size)
+
+
+class StreamWriter(Codec, codecs.StreamWriter):
+    pass
+
+
+class StreamReader(Codec, codecs.StreamReader):
+    pass
+
+
+def search_function(name: str) -> Optional[codecs.CodecInfo]:
+    if name != "idna2008":
+        return None
+    return codecs.CodecInfo(
+        name=name,
+        encode=Codec().encode,
+        decode=Codec().decode,
+        incrementalencoder=IncrementalEncoder,
+        incrementaldecoder=IncrementalDecoder,
+        streamwriter=StreamWriter,
+        streamreader=StreamReader,
+    )
+
+
+codecs.register(search_function)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/compat.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/compat.py
new file mode 100644
index 00000000..1df9f2a7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/compat.py
@@ -0,0 +1,15 @@
+from typing import Any, Union
+
+from .core import decode, encode
+
+
+def ToASCII(label: str) -> bytes:
+    return encode(label)
+
+
+def ToUnicode(label: Union[bytes, bytearray]) -> str:
+    return decode(label)
+
+
+def nameprep(s: Any) -> None:
+    raise NotImplementedError("IDNA 2008 does not utilise nameprep protocol")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/core.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/core.py
new file mode 100644
index 00000000..9115f123
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/core.py
@@ -0,0 +1,437 @@
+import bisect
+import re
+import unicodedata
+from typing import Optional, Union
+
+from . import idnadata
+from .intranges import intranges_contain
+
+_virama_combining_class = 9
+_alabel_prefix = b"xn--"
+_unicode_dots_re = re.compile("[\u002e\u3002\uff0e\uff61]")
+
+
+class IDNAError(UnicodeError):
+    """Base exception for all IDNA-encoding related problems"""
+
+    pass
+
+
+class IDNABidiError(IDNAError):
+    """Exception when bidirectional requirements are not satisfied"""
+
+    pass
+
+
+class InvalidCodepoint(IDNAError):
+    """Exception when a disallowed or unallocated codepoint is used"""
+
+    pass
+
+
+class InvalidCodepointContext(IDNAError):
+    """Exception when the codepoint is not valid in the context it is used"""
+
+    pass
+
+
+def _combining_class(cp: int) -> int:
+    v = unicodedata.combining(chr(cp))
+    if v == 0:
+        if not unicodedata.name(chr(cp)):
+            raise ValueError("Unknown character in unicodedata")
+    return v
+
+
+def _is_script(cp: str, script: str) -> bool:
+    return intranges_contain(ord(cp), idnadata.scripts[script])
+
+
+def _punycode(s: str) -> bytes:
+    return s.encode("punycode")
+
+
+def _unot(s: int) -> str:
+    return "U+{:04X}".format(s)
+
+
+def valid_label_length(label: Union[bytes, str]) -> bool:
+    if len(label) > 63:
+        return False
+    return True
+
+
+def valid_string_length(label: Union[bytes, str], trailing_dot: bool) -> bool:
+    if len(label) > (254 if trailing_dot else 253):
+        return False
+    return True
+
+
+def check_bidi(label: str, check_ltr: bool = False) -> bool:
+    # Bidi rules should only be applied if string contains RTL characters
+    bidi_label = False
+    for idx, cp in enumerate(label, 1):
+        direction = unicodedata.bidirectional(cp)
+        if direction == "":
+            # String likely comes from a newer version of Unicode
+            raise IDNABidiError("Unknown directionality in label {} at position {}".format(repr(label), idx))
+        if direction in ["R", "AL", "AN"]:
+            bidi_label = True
+    if not bidi_label and not check_ltr:
+        return True
+
+    # Bidi rule 1
+    direction = unicodedata.bidirectional(label[0])
+    if direction in ["R", "AL"]:
+        rtl = True
+    elif direction == "L":
+        rtl = False
+    else:
+        raise IDNABidiError("First codepoint in label {} must be directionality L, R or AL".format(repr(label)))
+
+    valid_ending = False
+    number_type: Optional[str] = None
+    for idx, cp in enumerate(label, 1):
+        direction = unicodedata.bidirectional(cp)
+
+        if rtl:
+            # Bidi rule 2
+            if direction not in [
+                "R",
+                "AL",
+                "AN",
+                "EN",
+                "ES",
+                "CS",
+                "ET",
+                "ON",
+                "BN",
+                "NSM",
+            ]:
+                raise IDNABidiError("Invalid direction for codepoint at position {} in a right-to-left label".format(idx))
+            # Bidi rule 3
+            if direction in ["R", "AL", "EN", "AN"]:
+                valid_ending = True
+            elif direction != "NSM":
+                valid_ending = False
+            # Bidi rule 4
+            if direction in ["AN", "EN"]:
+                if not number_type:
+                    number_type = direction
+                else:
+                    if number_type != direction:
+                        raise IDNABidiError("Can not mix numeral types in a right-to-left label")
+        else:
+            # Bidi rule 5
+            if direction not in ["L", "EN", "ES", "CS", "ET", "ON", "BN", "NSM"]:
+                raise IDNABidiError("Invalid direction for codepoint at position {} in a left-to-right label".format(idx))
+            # Bidi rule 6
+            if direction in ["L", "EN"]:
+                valid_ending = True
+            elif direction != "NSM":
+                valid_ending = False
+
+    if not valid_ending:
+        raise IDNABidiError("Label ends with illegal codepoint directionality")
+
+    return True
+
+
+def check_initial_combiner(label: str) -> bool:
+    if unicodedata.category(label[0])[0] == "M":
+        raise IDNAError("Label begins with an illegal combining character")
+    return True
+
+
+def check_hyphen_ok(label: str) -> bool:
+    if label[2:4] == "--":
+        raise IDNAError("Label has disallowed hyphens in 3rd and 4th position")
+    if label[0] == "-" or label[-1] == "-":
+        raise IDNAError("Label must not start or end with a hyphen")
+    return True
+
+
+def check_nfc(label: str) -> None:
+    if unicodedata.normalize("NFC", label) != label:
+        raise IDNAError("Label must be in Normalization Form C")
+
+
+def valid_contextj(label: str, pos: int) -> bool:
+    cp_value = ord(label[pos])
+
+    if cp_value == 0x200C:
+        if pos > 0:
+            if _combining_class(ord(label[pos - 1])) == _virama_combining_class:
+                return True
+
+        ok = False
+        for i in range(pos - 1, -1, -1):
+            joining_type = idnadata.joining_types.get(ord(label[i]))
+            if joining_type == ord("T"):
+                continue
+            elif joining_type in [ord("L"), ord("D")]:
+                ok = True
+                break
+            else:
+                break
+
+        if not ok:
+            return False
+
+        ok = False
+        for i in range(pos + 1, len(label)):
+            joining_type = idnadata.joining_types.get(ord(label[i]))
+            if joining_type == ord("T"):
+                continue
+            elif joining_type in [ord("R"), ord("D")]:
+                ok = True
+                break
+            else:
+                break
+        return ok
+
+    if cp_value == 0x200D:
+        if pos > 0:
+            if _combining_class(ord(label[pos - 1])) == _virama_combining_class:
+                return True
+        return False
+
+    else:
+        return False
+
+
+def valid_contexto(label: str, pos: int, exception: bool = False) -> bool:
+    cp_value = ord(label[pos])
+
+    if cp_value == 0x00B7:
+        if 0 < pos < len(label) - 1:
+            if ord(label[pos - 1]) == 0x006C and ord(label[pos + 1]) == 0x006C:
+                return True
+        return False
+
+    elif cp_value == 0x0375:
+        if pos < len(label) - 1 and len(label) > 1:
+            return _is_script(label[pos + 1], "Greek")
+        return False
+
+    elif cp_value == 0x05F3 or cp_value == 0x05F4:
+        if pos > 0:
+            return _is_script(label[pos - 1], "Hebrew")
+        return False
+
+    elif cp_value == 0x30FB:
+        for cp in label:
+            if cp == "\u30fb":
+                continue
+            if _is_script(cp, "Hiragana") or _is_script(cp, "Katakana") or _is_script(cp, "Han"):
+                return True
+        return False
+
+    elif 0x660 <= cp_value <= 0x669:
+        for cp in label:
+            if 0x6F0 <= ord(cp) <= 0x06F9:
+                return False
+        return True
+
+    elif 0x6F0 <= cp_value <= 0x6F9:
+        for cp in label:
+            if 0x660 <= ord(cp) <= 0x0669:
+                return False
+        return True
+
+    return False
+
+
+def check_label(label: Union[str, bytes, bytearray]) -> None:
+    if isinstance(label, (bytes, bytearray)):
+        label = label.decode("utf-8")
+    if len(label) == 0:
+        raise IDNAError("Empty Label")
+
+    check_nfc(label)
+    check_hyphen_ok(label)
+    check_initial_combiner(label)
+
+    for pos, cp in enumerate(label):
+        cp_value = ord(cp)
+        if intranges_contain(cp_value, idnadata.codepoint_classes["PVALID"]):
+            continue
+        elif intranges_contain(cp_value, idnadata.codepoint_classes["CONTEXTJ"]):
+            try:
+                if not valid_contextj(label, pos):
+                    raise InvalidCodepointContext(
+                        "Joiner {} not allowed at position {} in {}".format(_unot(cp_value), pos + 1, repr(label))
+                    )
+            except ValueError:
+                raise IDNAError(
+                    "Unknown codepoint adjacent to joiner {} at position {} in {}".format(
+                        _unot(cp_value), pos + 1, repr(label)
+                    )
+                )
+        elif intranges_contain(cp_value, idnadata.codepoint_classes["CONTEXTO"]):
+            if not valid_contexto(label, pos):
+                raise InvalidCodepointContext(
+                    "Codepoint {} not allowed at position {} in {}".format(_unot(cp_value), pos + 1, repr(label))
+                )
+        else:
+            raise InvalidCodepoint(
+                "Codepoint {} at position {} of {} not allowed".format(_unot(cp_value), pos + 1, repr(label))
+            )
+
+    check_bidi(label)
+
+
+def alabel(label: str) -> bytes:
+    try:
+        label_bytes = label.encode("ascii")
+        ulabel(label_bytes)
+        if not valid_label_length(label_bytes):
+            raise IDNAError("Label too long")
+        return label_bytes
+    except UnicodeEncodeError:
+        pass
+
+    check_label(label)
+    label_bytes = _alabel_prefix + _punycode(label)
+
+    if not valid_label_length(label_bytes):
+        raise IDNAError("Label too long")
+
+    return label_bytes
+
+
+def ulabel(label: Union[str, bytes, bytearray]) -> str:
+    if not isinstance(label, (bytes, bytearray)):
+        try:
+            label_bytes = label.encode("ascii")
+        except UnicodeEncodeError:
+            check_label(label)
+            return label
+    else:
+        label_bytes = label
+
+    label_bytes = label_bytes.lower()
+    if label_bytes.startswith(_alabel_prefix):
+        label_bytes = label_bytes[len(_alabel_prefix) :]
+        if not label_bytes:
+            raise IDNAError("Malformed A-label, no Punycode eligible content found")
+        if label_bytes.decode("ascii")[-1] == "-":
+            raise IDNAError("A-label must not end with a hyphen")
+    else:
+        check_label(label_bytes)
+        return label_bytes.decode("ascii")
+
+    try:
+        label = label_bytes.decode("punycode")
+    except UnicodeError:
+        raise IDNAError("Invalid A-label")
+    check_label(label)
+    return label
+
+
+def uts46_remap(domain: str, std3_rules: bool = True, transitional: bool = False) -> str:
+    """Re-map the characters in the string according to UTS46 processing."""
+    from .uts46data import uts46data
+
+    output = ""
+
+    for pos, char in enumerate(domain):
+        code_point = ord(char)
+        try:
+            uts46row = uts46data[code_point if code_point < 256 else bisect.bisect_left(uts46data, (code_point, "Z")) - 1]
+            status = uts46row[1]
+            replacement: Optional[str] = None
+            if len(uts46row) == 3:
+                replacement = uts46row[2]
+            if (
+                status == "V"
+                or (status == "D" and not transitional)
+                or (status == "3" and not std3_rules and replacement is None)
+            ):
+                output += char
+            elif replacement is not None and (
+                status == "M" or (status == "3" and not std3_rules) or (status == "D" and transitional)
+            ):
+                output += replacement
+            elif status != "I":
+                raise IndexError()
+        except IndexError:
+            raise InvalidCodepoint(
+                "Codepoint {} not allowed at position {} in {}".format(_unot(code_point), pos + 1, repr(domain))
+            )
+
+    return unicodedata.normalize("NFC", output)
+
+
+def encode(
+    s: Union[str, bytes, bytearray],
+    strict: bool = False,
+    uts46: bool = False,
+    std3_rules: bool = False,
+    transitional: bool = False,
+) -> bytes:
+    if not isinstance(s, str):
+        try:
+            s = str(s, "ascii")
+        except UnicodeDecodeError:
+            raise IDNAError("should pass a unicode string to the function rather than a byte string.")
+    if uts46:
+        s = uts46_remap(s, std3_rules, transitional)
+    trailing_dot = False
+    result = []
+    if strict:
+        labels = s.split(".")
+    else:
+        labels = _unicode_dots_re.split(s)
+    if not labels or labels == [""]:
+        raise IDNAError("Empty domain")
+    if labels[-1] == "":
+        del labels[-1]
+        trailing_dot = True
+    for label in labels:
+        s = alabel(label)
+        if s:
+            result.append(s)
+        else:
+            raise IDNAError("Empty label")
+    if trailing_dot:
+        result.append(b"")
+    s = b".".join(result)
+    if not valid_string_length(s, trailing_dot):
+        raise IDNAError("Domain too long")
+    return s
+
+
+def decode(
+    s: Union[str, bytes, bytearray],
+    strict: bool = False,
+    uts46: bool = False,
+    std3_rules: bool = False,
+) -> str:
+    try:
+        if not isinstance(s, str):
+            s = str(s, "ascii")
+    except UnicodeDecodeError:
+        raise IDNAError("Invalid ASCII in A-label")
+    if uts46:
+        s = uts46_remap(s, std3_rules, False)
+    trailing_dot = False
+    result = []
+    if not strict:
+        labels = _unicode_dots_re.split(s)
+    else:
+        labels = s.split(".")
+    if not labels or labels == [""]:
+        raise IDNAError("Empty domain")
+    if not labels[-1]:
+        del labels[-1]
+        trailing_dot = True
+    for label in labels:
+        s = ulabel(label)
+        if s:
+            result.append(s)
+        else:
+            raise IDNAError("Empty label")
+    if trailing_dot:
+        result.append("")
+    return ".".join(result)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/idnadata.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/idnadata.py
new file mode 100644
index 00000000..4be60046
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/idnadata.py
@@ -0,0 +1,4243 @@
+# This file is automatically generated by tools/idna-data
+
+__version__ = "15.1.0"
+scripts = {
+    "Greek": (
+        0x37000000374,
+        0x37500000378,
+        0x37A0000037E,
+        0x37F00000380,
+        0x38400000385,
+        0x38600000387,
+        0x3880000038B,
+        0x38C0000038D,
+        0x38E000003A2,
+        0x3A3000003E2,
+        0x3F000000400,
+        0x1D2600001D2B,
+        0x1D5D00001D62,
+        0x1D6600001D6B,
+        0x1DBF00001DC0,
+        0x1F0000001F16,
+        0x1F1800001F1E,
+        0x1F2000001F46,
+        0x1F4800001F4E,
+        0x1F5000001F58,
+        0x1F5900001F5A,
+        0x1F5B00001F5C,
+        0x1F5D00001F5E,
+        0x1F5F00001F7E,
+        0x1F8000001FB5,
+        0x1FB600001FC5,
+        0x1FC600001FD4,
+        0x1FD600001FDC,
+        0x1FDD00001FF0,
+        0x1FF200001FF5,
+        0x1FF600001FFF,
+        0x212600002127,
+        0xAB650000AB66,
+        0x101400001018F,
+        0x101A0000101A1,
+        0x1D2000001D246,
+    ),
+    "Han": (
+        0x2E8000002E9A,
+        0x2E9B00002EF4,
+        0x2F0000002FD6,
+        0x300500003006,
+        0x300700003008,
+        0x30210000302A,
+        0x30380000303C,
+        0x340000004DC0,
+        0x4E000000A000,
+        0xF9000000FA6E,
+        0xFA700000FADA,
+        0x16FE200016FE4,
+        0x16FF000016FF2,
+        0x200000002A6E0,
+        0x2A7000002B73A,
+        0x2B7400002B81E,
+        0x2B8200002CEA2,
+        0x2CEB00002EBE1,
+        0x2EBF00002EE5E,
+        0x2F8000002FA1E,
+        0x300000003134B,
+        0x31350000323B0,
+    ),
+    "Hebrew": (
+        0x591000005C8,
+        0x5D0000005EB,
+        0x5EF000005F5,
+        0xFB1D0000FB37,
+        0xFB380000FB3D,
+        0xFB3E0000FB3F,
+        0xFB400000FB42,
+        0xFB430000FB45,
+        0xFB460000FB50,
+    ),
+    "Hiragana": (
+        0x304100003097,
+        0x309D000030A0,
+        0x1B0010001B120,
+        0x1B1320001B133,
+        0x1B1500001B153,
+        0x1F2000001F201,
+    ),
+    "Katakana": (
+        0x30A1000030FB,
+        0x30FD00003100,
+        0x31F000003200,
+        0x32D0000032FF,
+        0x330000003358,
+        0xFF660000FF70,
+        0xFF710000FF9E,
+        0x1AFF00001AFF4,
+        0x1AFF50001AFFC,
+        0x1AFFD0001AFFF,
+        0x1B0000001B001,
+        0x1B1200001B123,
+        0x1B1550001B156,
+        0x1B1640001B168,
+    ),
+}
+joining_types = {
+    0xAD: 84,
+    0x300: 84,
+    0x301: 84,
+    0x302: 84,
+    0x303: 84,
+    0x304: 84,
+    0x305: 84,
+    0x306: 84,
+    0x307: 84,
+    0x308: 84,
+    0x309: 84,
+    0x30A: 84,
+    0x30B: 84,
+    0x30C: 84,
+    0x30D: 84,
+    0x30E: 84,
+    0x30F: 84,
+    0x310: 84,
+    0x311: 84,
+    0x312: 84,
+    0x313: 84,
+    0x314: 84,
+    0x315: 84,
+    0x316: 84,
+    0x317: 84,
+    0x318: 84,
+    0x319: 84,
+    0x31A: 84,
+    0x31B: 84,
+    0x31C: 84,
+    0x31D: 84,
+    0x31E: 84,
+    0x31F: 84,
+    0x320: 84,
+    0x321: 84,
+    0x322: 84,
+    0x323: 84,
+    0x324: 84,
+    0x325: 84,
+    0x326: 84,
+    0x327: 84,
+    0x328: 84,
+    0x329: 84,
+    0x32A: 84,
+    0x32B: 84,
+    0x32C: 84,
+    0x32D: 84,
+    0x32E: 84,
+    0x32F: 84,
+    0x330: 84,
+    0x331: 84,
+    0x332: 84,
+    0x333: 84,
+    0x334: 84,
+    0x335: 84,
+    0x336: 84,
+    0x337: 84,
+    0x338: 84,
+    0x339: 84,
+    0x33A: 84,
+    0x33B: 84,
+    0x33C: 84,
+    0x33D: 84,
+    0x33E: 84,
+    0x33F: 84,
+    0x340: 84,
+    0x341: 84,
+    0x342: 84,
+    0x343: 84,
+    0x344: 84,
+    0x345: 84,
+    0x346: 84,
+    0x347: 84,
+    0x348: 84,
+    0x349: 84,
+    0x34A: 84,
+    0x34B: 84,
+    0x34C: 84,
+    0x34D: 84,
+    0x34E: 84,
+    0x34F: 84,
+    0x350: 84,
+    0x351: 84,
+    0x352: 84,
+    0x353: 84,
+    0x354: 84,
+    0x355: 84,
+    0x356: 84,
+    0x357: 84,
+    0x358: 84,
+    0x359: 84,
+    0x35A: 84,
+    0x35B: 84,
+    0x35C: 84,
+    0x35D: 84,
+    0x35E: 84,
+    0x35F: 84,
+    0x360: 84,
+    0x361: 84,
+    0x362: 84,
+    0x363: 84,
+    0x364: 84,
+    0x365: 84,
+    0x366: 84,
+    0x367: 84,
+    0x368: 84,
+    0x369: 84,
+    0x36A: 84,
+    0x36B: 84,
+    0x36C: 84,
+    0x36D: 84,
+    0x36E: 84,
+    0x36F: 84,
+    0x483: 84,
+    0x484: 84,
+    0x485: 84,
+    0x486: 84,
+    0x487: 84,
+    0x488: 84,
+    0x489: 84,
+    0x591: 84,
+    0x592: 84,
+    0x593: 84,
+    0x594: 84,
+    0x595: 84,
+    0x596: 84,
+    0x597: 84,
+    0x598: 84,
+    0x599: 84,
+    0x59A: 84,
+    0x59B: 84,
+    0x59C: 84,
+    0x59D: 84,
+    0x59E: 84,
+    0x59F: 84,
+    0x5A0: 84,
+    0x5A1: 84,
+    0x5A2: 84,
+    0x5A3: 84,
+    0x5A4: 84,
+    0x5A5: 84,
+    0x5A6: 84,
+    0x5A7: 84,
+    0x5A8: 84,
+    0x5A9: 84,
+    0x5AA: 84,
+    0x5AB: 84,
+    0x5AC: 84,
+    0x5AD: 84,
+    0x5AE: 84,
+    0x5AF: 84,
+    0x5B0: 84,
+    0x5B1: 84,
+    0x5B2: 84,
+    0x5B3: 84,
+    0x5B4: 84,
+    0x5B5: 84,
+    0x5B6: 84,
+    0x5B7: 84,
+    0x5B8: 84,
+    0x5B9: 84,
+    0x5BA: 84,
+    0x5BB: 84,
+    0x5BC: 84,
+    0x5BD: 84,
+    0x5BF: 84,
+    0x5C1: 84,
+    0x5C2: 84,
+    0x5C4: 84,
+    0x5C5: 84,
+    0x5C7: 84,
+    0x610: 84,
+    0x611: 84,
+    0x612: 84,
+    0x613: 84,
+    0x614: 84,
+    0x615: 84,
+    0x616: 84,
+    0x617: 84,
+    0x618: 84,
+    0x619: 84,
+    0x61A: 84,
+    0x61C: 84,
+    0x620: 68,
+    0x622: 82,
+    0x623: 82,
+    0x624: 82,
+    0x625: 82,
+    0x626: 68,
+    0x627: 82,
+    0x628: 68,
+    0x629: 82,
+    0x62A: 68,
+    0x62B: 68,
+    0x62C: 68,
+    0x62D: 68,
+    0x62E: 68,
+    0x62F: 82,
+    0x630: 82,
+    0x631: 82,
+    0x632: 82,
+    0x633: 68,
+    0x634: 68,
+    0x635: 68,
+    0x636: 68,
+    0x637: 68,
+    0x638: 68,
+    0x639: 68,
+    0x63A: 68,
+    0x63B: 68,
+    0x63C: 68,
+    0x63D: 68,
+    0x63E: 68,
+    0x63F: 68,
+    0x640: 67,
+    0x641: 68,
+    0x642: 68,
+    0x643: 68,
+    0x644: 68,
+    0x645: 68,
+    0x646: 68,
+    0x647: 68,
+    0x648: 82,
+    0x649: 68,
+    0x64A: 68,
+    0x64B: 84,
+    0x64C: 84,
+    0x64D: 84,
+    0x64E: 84,
+    0x64F: 84,
+    0x650: 84,
+    0x651: 84,
+    0x652: 84,
+    0x653: 84,
+    0x654: 84,
+    0x655: 84,
+    0x656: 84,
+    0x657: 84,
+    0x658: 84,
+    0x659: 84,
+    0x65A: 84,
+    0x65B: 84,
+    0x65C: 84,
+    0x65D: 84,
+    0x65E: 84,
+    0x65F: 84,
+    0x66E: 68,
+    0x66F: 68,
+    0x670: 84,
+    0x671: 82,
+    0x672: 82,
+    0x673: 82,
+    0x675: 82,
+    0x676: 82,
+    0x677: 82,
+    0x678: 68,
+    0x679: 68,
+    0x67A: 68,
+    0x67B: 68,
+    0x67C: 68,
+    0x67D: 68,
+    0x67E: 68,
+    0x67F: 68,
+    0x680: 68,
+    0x681: 68,
+    0x682: 68,
+    0x683: 68,
+    0x684: 68,
+    0x685: 68,
+    0x686: 68,
+    0x687: 68,
+    0x688: 82,
+    0x689: 82,
+    0x68A: 82,
+    0x68B: 82,
+    0x68C: 82,
+    0x68D: 82,
+    0x68E: 82,
+    0x68F: 82,
+    0x690: 82,
+    0x691: 82,
+    0x692: 82,
+    0x693: 82,
+    0x694: 82,
+    0x695: 82,
+    0x696: 82,
+    0x697: 82,
+    0x698: 82,
+    0x699: 82,
+    0x69A: 68,
+    0x69B: 68,
+    0x69C: 68,
+    0x69D: 68,
+    0x69E: 68,
+    0x69F: 68,
+    0x6A0: 68,
+    0x6A1: 68,
+    0x6A2: 68,
+    0x6A3: 68,
+    0x6A4: 68,
+    0x6A5: 68,
+    0x6A6: 68,
+    0x6A7: 68,
+    0x6A8: 68,
+    0x6A9: 68,
+    0x6AA: 68,
+    0x6AB: 68,
+    0x6AC: 68,
+    0x6AD: 68,
+    0x6AE: 68,
+    0x6AF: 68,
+    0x6B0: 68,
+    0x6B1: 68,
+    0x6B2: 68,
+    0x6B3: 68,
+    0x6B4: 68,
+    0x6B5: 68,
+    0x6B6: 68,
+    0x6B7: 68,
+    0x6B8: 68,
+    0x6B9: 68,
+    0x6BA: 68,
+    0x6BB: 68,
+    0x6BC: 68,
+    0x6BD: 68,
+    0x6BE: 68,
+    0x6BF: 68,
+    0x6C0: 82,
+    0x6C1: 68,
+    0x6C2: 68,
+    0x6C3: 82,
+    0x6C4: 82,
+    0x6C5: 82,
+    0x6C6: 82,
+    0x6C7: 82,
+    0x6C8: 82,
+    0x6C9: 82,
+    0x6CA: 82,
+    0x6CB: 82,
+    0x6CC: 68,
+    0x6CD: 82,
+    0x6CE: 68,
+    0x6CF: 82,
+    0x6D0: 68,
+    0x6D1: 68,
+    0x6D2: 82,
+    0x6D3: 82,
+    0x6D5: 82,
+    0x6D6: 84,
+    0x6D7: 84,
+    0x6D8: 84,
+    0x6D9: 84,
+    0x6DA: 84,
+    0x6DB: 84,
+    0x6DC: 84,
+    0x6DF: 84,
+    0x6E0: 84,
+    0x6E1: 84,
+    0x6E2: 84,
+    0x6E3: 84,
+    0x6E4: 84,
+    0x6E7: 84,
+    0x6E8: 84,
+    0x6EA: 84,
+    0x6EB: 84,
+    0x6EC: 84,
+    0x6ED: 84,
+    0x6EE: 82,
+    0x6EF: 82,
+    0x6FA: 68,
+    0x6FB: 68,
+    0x6FC: 68,
+    0x6FF: 68,
+    0x70F: 84,
+    0x710: 82,
+    0x711: 84,
+    0x712: 68,
+    0x713: 68,
+    0x714: 68,
+    0x715: 82,
+    0x716: 82,
+    0x717: 82,
+    0x718: 82,
+    0x719: 82,
+    0x71A: 68,
+    0x71B: 68,
+    0x71C: 68,
+    0x71D: 68,
+    0x71E: 82,
+    0x71F: 68,
+    0x720: 68,
+    0x721: 68,
+    0x722: 68,
+    0x723: 68,
+    0x724: 68,
+    0x725: 68,
+    0x726: 68,
+    0x727: 68,
+    0x728: 82,
+    0x729: 68,
+    0x72A: 82,
+    0x72B: 68,
+    0x72C: 82,
+    0x72D: 68,
+    0x72E: 68,
+    0x72F: 82,
+    0x730: 84,
+    0x731: 84,
+    0x732: 84,
+    0x733: 84,
+    0x734: 84,
+    0x735: 84,
+    0x736: 84,
+    0x737: 84,
+    0x738: 84,
+    0x739: 84,
+    0x73A: 84,
+    0x73B: 84,
+    0x73C: 84,
+    0x73D: 84,
+    0x73E: 84,
+    0x73F: 84,
+    0x740: 84,
+    0x741: 84,
+    0x742: 84,
+    0x743: 84,
+    0x744: 84,
+    0x745: 84,
+    0x746: 84,
+    0x747: 84,
+    0x748: 84,
+    0x749: 84,
+    0x74A: 84,
+    0x74D: 82,
+    0x74E: 68,
+    0x74F: 68,
+    0x750: 68,
+    0x751: 68,
+    0x752: 68,
+    0x753: 68,
+    0x754: 68,
+    0x755: 68,
+    0x756: 68,
+    0x757: 68,
+    0x758: 68,
+    0x759: 82,
+    0x75A: 82,
+    0x75B: 82,
+    0x75C: 68,
+    0x75D: 68,
+    0x75E: 68,
+    0x75F: 68,
+    0x760: 68,
+    0x761: 68,
+    0x762: 68,
+    0x763: 68,
+    0x764: 68,
+    0x765: 68,
+    0x766: 68,
+    0x767: 68,
+    0x768: 68,
+    0x769: 68,
+    0x76A: 68,
+    0x76B: 82,
+    0x76C: 82,
+    0x76D: 68,
+    0x76E: 68,
+    0x76F: 68,
+    0x770: 68,
+    0x771: 82,
+    0x772: 68,
+    0x773: 82,
+    0x774: 82,
+    0x775: 68,
+    0x776: 68,
+    0x777: 68,
+    0x778: 82,
+    0x779: 82,
+    0x77A: 68,
+    0x77B: 68,
+    0x77C: 68,
+    0x77D: 68,
+    0x77E: 68,
+    0x77F: 68,
+    0x7A6: 84,
+    0x7A7: 84,
+    0x7A8: 84,
+    0x7A9: 84,
+    0x7AA: 84,
+    0x7AB: 84,
+    0x7AC: 84,
+    0x7AD: 84,
+    0x7AE: 84,
+    0x7AF: 84,
+    0x7B0: 84,
+    0x7CA: 68,
+    0x7CB: 68,
+    0x7CC: 68,
+    0x7CD: 68,
+    0x7CE: 68,
+    0x7CF: 68,
+    0x7D0: 68,
+    0x7D1: 68,
+    0x7D2: 68,
+    0x7D3: 68,
+    0x7D4: 68,
+    0x7D5: 68,
+    0x7D6: 68,
+    0x7D7: 68,
+    0x7D8: 68,
+    0x7D9: 68,
+    0x7DA: 68,
+    0x7DB: 68,
+    0x7DC: 68,
+    0x7DD: 68,
+    0x7DE: 68,
+    0x7DF: 68,
+    0x7E0: 68,
+    0x7E1: 68,
+    0x7E2: 68,
+    0x7E3: 68,
+    0x7E4: 68,
+    0x7E5: 68,
+    0x7E6: 68,
+    0x7E7: 68,
+    0x7E8: 68,
+    0x7E9: 68,
+    0x7EA: 68,
+    0x7EB: 84,
+    0x7EC: 84,
+    0x7ED: 84,
+    0x7EE: 84,
+    0x7EF: 84,
+    0x7F0: 84,
+    0x7F1: 84,
+    0x7F2: 84,
+    0x7F3: 84,
+    0x7FA: 67,
+    0x7FD: 84,
+    0x816: 84,
+    0x817: 84,
+    0x818: 84,
+    0x819: 84,
+    0x81B: 84,
+    0x81C: 84,
+    0x81D: 84,
+    0x81E: 84,
+    0x81F: 84,
+    0x820: 84,
+    0x821: 84,
+    0x822: 84,
+    0x823: 84,
+    0x825: 84,
+    0x826: 84,
+    0x827: 84,
+    0x829: 84,
+    0x82A: 84,
+    0x82B: 84,
+    0x82C: 84,
+    0x82D: 84,
+    0x840: 82,
+    0x841: 68,
+    0x842: 68,
+    0x843: 68,
+    0x844: 68,
+    0x845: 68,
+    0x846: 82,
+    0x847: 82,
+    0x848: 68,
+    0x849: 82,
+    0x84A: 68,
+    0x84B: 68,
+    0x84C: 68,
+    0x84D: 68,
+    0x84E: 68,
+    0x84F: 68,
+    0x850: 68,
+    0x851: 68,
+    0x852: 68,
+    0x853: 68,
+    0x854: 82,
+    0x855: 68,
+    0x856: 82,
+    0x857: 82,
+    0x858: 82,
+    0x859: 84,
+    0x85A: 84,
+    0x85B: 84,
+    0x860: 68,
+    0x862: 68,
+    0x863: 68,
+    0x864: 68,
+    0x865: 68,
+    0x867: 82,
+    0x868: 68,
+    0x869: 82,
+    0x86A: 82,
+    0x870: 82,
+    0x871: 82,
+    0x872: 82,
+    0x873: 82,
+    0x874: 82,
+    0x875: 82,
+    0x876: 82,
+    0x877: 82,
+    0x878: 82,
+    0x879: 82,
+    0x87A: 82,
+    0x87B: 82,
+    0x87C: 82,
+    0x87D: 82,
+    0x87E: 82,
+    0x87F: 82,
+    0x880: 82,
+    0x881: 82,
+    0x882: 82,
+    0x883: 67,
+    0x884: 67,
+    0x885: 67,
+    0x886: 68,
+    0x889: 68,
+    0x88A: 68,
+    0x88B: 68,
+    0x88C: 68,
+    0x88D: 68,
+    0x88E: 82,
+    0x898: 84,
+    0x899: 84,
+    0x89A: 84,
+    0x89B: 84,
+    0x89C: 84,
+    0x89D: 84,
+    0x89E: 84,
+    0x89F: 84,
+    0x8A0: 68,
+    0x8A1: 68,
+    0x8A2: 68,
+    0x8A3: 68,
+    0x8A4: 68,
+    0x8A5: 68,
+    0x8A6: 68,
+    0x8A7: 68,
+    0x8A8: 68,
+    0x8A9: 68,
+    0x8AA: 82,
+    0x8AB: 82,
+    0x8AC: 82,
+    0x8AE: 82,
+    0x8AF: 68,
+    0x8B0: 68,
+    0x8B1: 82,
+    0x8B2: 82,
+    0x8B3: 68,
+    0x8B4: 68,
+    0x8B5: 68,
+    0x8B6: 68,
+    0x8B7: 68,
+    0x8B8: 68,
+    0x8B9: 82,
+    0x8BA: 68,
+    0x8BB: 68,
+    0x8BC: 68,
+    0x8BD: 68,
+    0x8BE: 68,
+    0x8BF: 68,
+    0x8C0: 68,
+    0x8C1: 68,
+    0x8C2: 68,
+    0x8C3: 68,
+    0x8C4: 68,
+    0x8C5: 68,
+    0x8C6: 68,
+    0x8C7: 68,
+    0x8C8: 68,
+    0x8CA: 84,
+    0x8CB: 84,
+    0x8CC: 84,
+    0x8CD: 84,
+    0x8CE: 84,
+    0x8CF: 84,
+    0x8D0: 84,
+    0x8D1: 84,
+    0x8D2: 84,
+    0x8D3: 84,
+    0x8D4: 84,
+    0x8D5: 84,
+    0x8D6: 84,
+    0x8D7: 84,
+    0x8D8: 84,
+    0x8D9: 84,
+    0x8DA: 84,
+    0x8DB: 84,
+    0x8DC: 84,
+    0x8DD: 84,
+    0x8DE: 84,
+    0x8DF: 84,
+    0x8E0: 84,
+    0x8E1: 84,
+    0x8E3: 84,
+    0x8E4: 84,
+    0x8E5: 84,
+    0x8E6: 84,
+    0x8E7: 84,
+    0x8E8: 84,
+    0x8E9: 84,
+    0x8EA: 84,
+    0x8EB: 84,
+    0x8EC: 84,
+    0x8ED: 84,
+    0x8EE: 84,
+    0x8EF: 84,
+    0x8F0: 84,
+    0x8F1: 84,
+    0x8F2: 84,
+    0x8F3: 84,
+    0x8F4: 84,
+    0x8F5: 84,
+    0x8F6: 84,
+    0x8F7: 84,
+    0x8F8: 84,
+    0x8F9: 84,
+    0x8FA: 84,
+    0x8FB: 84,
+    0x8FC: 84,
+    0x8FD: 84,
+    0x8FE: 84,
+    0x8FF: 84,
+    0x900: 84,
+    0x901: 84,
+    0x902: 84,
+    0x93A: 84,
+    0x93C: 84,
+    0x941: 84,
+    0x942: 84,
+    0x943: 84,
+    0x944: 84,
+    0x945: 84,
+    0x946: 84,
+    0x947: 84,
+    0x948: 84,
+    0x94D: 84,
+    0x951: 84,
+    0x952: 84,
+    0x953: 84,
+    0x954: 84,
+    0x955: 84,
+    0x956: 84,
+    0x957: 84,
+    0x962: 84,
+    0x963: 84,
+    0x981: 84,
+    0x9BC: 84,
+    0x9C1: 84,
+    0x9C2: 84,
+    0x9C3: 84,
+    0x9C4: 84,
+    0x9CD: 84,
+    0x9E2: 84,
+    0x9E3: 84,
+    0x9FE: 84,
+    0xA01: 84,
+    0xA02: 84,
+    0xA3C: 84,
+    0xA41: 84,
+    0xA42: 84,
+    0xA47: 84,
+    0xA48: 84,
+    0xA4B: 84,
+    0xA4C: 84,
+    0xA4D: 84,
+    0xA51: 84,
+    0xA70: 84,
+    0xA71: 84,
+    0xA75: 84,
+    0xA81: 84,
+    0xA82: 84,
+    0xABC: 84,
+    0xAC1: 84,
+    0xAC2: 84,
+    0xAC3: 84,
+    0xAC4: 84,
+    0xAC5: 84,
+    0xAC7: 84,
+    0xAC8: 84,
+    0xACD: 84,
+    0xAE2: 84,
+    0xAE3: 84,
+    0xAFA: 84,
+    0xAFB: 84,
+    0xAFC: 84,
+    0xAFD: 84,
+    0xAFE: 84,
+    0xAFF: 84,
+    0xB01: 84,
+    0xB3C: 84,
+    0xB3F: 84,
+    0xB41: 84,
+    0xB42: 84,
+    0xB43: 84,
+    0xB44: 84,
+    0xB4D: 84,
+    0xB55: 84,
+    0xB56: 84,
+    0xB62: 84,
+    0xB63: 84,
+    0xB82: 84,
+    0xBC0: 84,
+    0xBCD: 84,
+    0xC00: 84,
+    0xC04: 84,
+    0xC3C: 84,
+    0xC3E: 84,
+    0xC3F: 84,
+    0xC40: 84,
+    0xC46: 84,
+    0xC47: 84,
+    0xC48: 84,
+    0xC4A: 84,
+    0xC4B: 84,
+    0xC4C: 84,
+    0xC4D: 84,
+    0xC55: 84,
+    0xC56: 84,
+    0xC62: 84,
+    0xC63: 84,
+    0xC81: 84,
+    0xCBC: 84,
+    0xCBF: 84,
+    0xCC6: 84,
+    0xCCC: 84,
+    0xCCD: 84,
+    0xCE2: 84,
+    0xCE3: 84,
+    0xD00: 84,
+    0xD01: 84,
+    0xD3B: 84,
+    0xD3C: 84,
+    0xD41: 84,
+    0xD42: 84,
+    0xD43: 84,
+    0xD44: 84,
+    0xD4D: 84,
+    0xD62: 84,
+    0xD63: 84,
+    0xD81: 84,
+    0xDCA: 84,
+    0xDD2: 84,
+    0xDD3: 84,
+    0xDD4: 84,
+    0xDD6: 84,
+    0xE31: 84,
+    0xE34: 84,
+    0xE35: 84,
+    0xE36: 84,
+    0xE37: 84,
+    0xE38: 84,
+    0xE39: 84,
+    0xE3A: 84,
+    0xE47: 84,
+    0xE48: 84,
+    0xE49: 84,
+    0xE4A: 84,
+    0xE4B: 84,
+    0xE4C: 84,
+    0xE4D: 84,
+    0xE4E: 84,
+    0xEB1: 84,
+    0xEB4: 84,
+    0xEB5: 84,
+    0xEB6: 84,
+    0xEB7: 84,
+    0xEB8: 84,
+    0xEB9: 84,
+    0xEBA: 84,
+    0xEBB: 84,
+    0xEBC: 84,
+    0xEC8: 84,
+    0xEC9: 84,
+    0xECA: 84,
+    0xECB: 84,
+    0xECC: 84,
+    0xECD: 84,
+    0xECE: 84,
+    0xF18: 84,
+    0xF19: 84,
+    0xF35: 84,
+    0xF37: 84,
+    0xF39: 84,
+    0xF71: 84,
+    0xF72: 84,
+    0xF73: 84,
+    0xF74: 84,
+    0xF75: 84,
+    0xF76: 84,
+    0xF77: 84,
+    0xF78: 84,
+    0xF79: 84,
+    0xF7A: 84,
+    0xF7B: 84,
+    0xF7C: 84,
+    0xF7D: 84,
+    0xF7E: 84,
+    0xF80: 84,
+    0xF81: 84,
+    0xF82: 84,
+    0xF83: 84,
+    0xF84: 84,
+    0xF86: 84,
+    0xF87: 84,
+    0xF8D: 84,
+    0xF8E: 84,
+    0xF8F: 84,
+    0xF90: 84,
+    0xF91: 84,
+    0xF92: 84,
+    0xF93: 84,
+    0xF94: 84,
+    0xF95: 84,
+    0xF96: 84,
+    0xF97: 84,
+    0xF99: 84,
+    0xF9A: 84,
+    0xF9B: 84,
+    0xF9C: 84,
+    0xF9D: 84,
+    0xF9E: 84,
+    0xF9F: 84,
+    0xFA0: 84,
+    0xFA1: 84,
+    0xFA2: 84,
+    0xFA3: 84,
+    0xFA4: 84,
+    0xFA5: 84,
+    0xFA6: 84,
+    0xFA7: 84,
+    0xFA8: 84,
+    0xFA9: 84,
+    0xFAA: 84,
+    0xFAB: 84,
+    0xFAC: 84,
+    0xFAD: 84,
+    0xFAE: 84,
+    0xFAF: 84,
+    0xFB0: 84,
+    0xFB1: 84,
+    0xFB2: 84,
+    0xFB3: 84,
+    0xFB4: 84,
+    0xFB5: 84,
+    0xFB6: 84,
+    0xFB7: 84,
+    0xFB8: 84,
+    0xFB9: 84,
+    0xFBA: 84,
+    0xFBB: 84,
+    0xFBC: 84,
+    0xFC6: 84,
+    0x102D: 84,
+    0x102E: 84,
+    0x102F: 84,
+    0x1030: 84,
+    0x1032: 84,
+    0x1033: 84,
+    0x1034: 84,
+    0x1035: 84,
+    0x1036: 84,
+    0x1037: 84,
+    0x1039: 84,
+    0x103A: 84,
+    0x103D: 84,
+    0x103E: 84,
+    0x1058: 84,
+    0x1059: 84,
+    0x105E: 84,
+    0x105F: 84,
+    0x1060: 84,
+    0x1071: 84,
+    0x1072: 84,
+    0x1073: 84,
+    0x1074: 84,
+    0x1082: 84,
+    0x1085: 84,
+    0x1086: 84,
+    0x108D: 84,
+    0x109D: 84,
+    0x135D: 84,
+    0x135E: 84,
+    0x135F: 84,
+    0x1712: 84,
+    0x1713: 84,
+    0x1714: 84,
+    0x1732: 84,
+    0x1733: 84,
+    0x1752: 84,
+    0x1753: 84,
+    0x1772: 84,
+    0x1773: 84,
+    0x17B4: 84,
+    0x17B5: 84,
+    0x17B7: 84,
+    0x17B8: 84,
+    0x17B9: 84,
+    0x17BA: 84,
+    0x17BB: 84,
+    0x17BC: 84,
+    0x17BD: 84,
+    0x17C6: 84,
+    0x17C9: 84,
+    0x17CA: 84,
+    0x17CB: 84,
+    0x17CC: 84,
+    0x17CD: 84,
+    0x17CE: 84,
+    0x17CF: 84,
+    0x17D0: 84,
+    0x17D1: 84,
+    0x17D2: 84,
+    0x17D3: 84,
+    0x17DD: 84,
+    0x1807: 68,
+    0x180A: 67,
+    0x180B: 84,
+    0x180C: 84,
+    0x180D: 84,
+    0x180F: 84,
+    0x1820: 68,
+    0x1821: 68,
+    0x1822: 68,
+    0x1823: 68,
+    0x1824: 68,
+    0x1825: 68,
+    0x1826: 68,
+    0x1827: 68,
+    0x1828: 68,
+    0x1829: 68,
+    0x182A: 68,
+    0x182B: 68,
+    0x182C: 68,
+    0x182D: 68,
+    0x182E: 68,
+    0x182F: 68,
+    0x1830: 68,
+    0x1831: 68,
+    0x1832: 68,
+    0x1833: 68,
+    0x1834: 68,
+    0x1835: 68,
+    0x1836: 68,
+    0x1837: 68,
+    0x1838: 68,
+    0x1839: 68,
+    0x183A: 68,
+    0x183B: 68,
+    0x183C: 68,
+    0x183D: 68,
+    0x183E: 68,
+    0x183F: 68,
+    0x1840: 68,
+    0x1841: 68,
+    0x1842: 68,
+    0x1843: 68,
+    0x1844: 68,
+    0x1845: 68,
+    0x1846: 68,
+    0x1847: 68,
+    0x1848: 68,
+    0x1849: 68,
+    0x184A: 68,
+    0x184B: 68,
+    0x184C: 68,
+    0x184D: 68,
+    0x184E: 68,
+    0x184F: 68,
+    0x1850: 68,
+    0x1851: 68,
+    0x1852: 68,
+    0x1853: 68,
+    0x1854: 68,
+    0x1855: 68,
+    0x1856: 68,
+    0x1857: 68,
+    0x1858: 68,
+    0x1859: 68,
+    0x185A: 68,
+    0x185B: 68,
+    0x185C: 68,
+    0x185D: 68,
+    0x185E: 68,
+    0x185F: 68,
+    0x1860: 68,
+    0x1861: 68,
+    0x1862: 68,
+    0x1863: 68,
+    0x1864: 68,
+    0x1865: 68,
+    0x1866: 68,
+    0x1867: 68,
+    0x1868: 68,
+    0x1869: 68,
+    0x186A: 68,
+    0x186B: 68,
+    0x186C: 68,
+    0x186D: 68,
+    0x186E: 68,
+    0x186F: 68,
+    0x1870: 68,
+    0x1871: 68,
+    0x1872: 68,
+    0x1873: 68,
+    0x1874: 68,
+    0x1875: 68,
+    0x1876: 68,
+    0x1877: 68,
+    0x1878: 68,
+    0x1885: 84,
+    0x1886: 84,
+    0x1887: 68,
+    0x1888: 68,
+    0x1889: 68,
+    0x188A: 68,
+    0x188B: 68,
+    0x188C: 68,
+    0x188D: 68,
+    0x188E: 68,
+    0x188F: 68,
+    0x1890: 68,
+    0x1891: 68,
+    0x1892: 68,
+    0x1893: 68,
+    0x1894: 68,
+    0x1895: 68,
+    0x1896: 68,
+    0x1897: 68,
+    0x1898: 68,
+    0x1899: 68,
+    0x189A: 68,
+    0x189B: 68,
+    0x189C: 68,
+    0x189D: 68,
+    0x189E: 68,
+    0x189F: 68,
+    0x18A0: 68,
+    0x18A1: 68,
+    0x18A2: 68,
+    0x18A3: 68,
+    0x18A4: 68,
+    0x18A5: 68,
+    0x18A6: 68,
+    0x18A7: 68,
+    0x18A8: 68,
+    0x18A9: 84,
+    0x18AA: 68,
+    0x1920: 84,
+    0x1921: 84,
+    0x1922: 84,
+    0x1927: 84,
+    0x1928: 84,
+    0x1932: 84,
+    0x1939: 84,
+    0x193A: 84,
+    0x193B: 84,
+    0x1A17: 84,
+    0x1A18: 84,
+    0x1A1B: 84,
+    0x1A56: 84,
+    0x1A58: 84,
+    0x1A59: 84,
+    0x1A5A: 84,
+    0x1A5B: 84,
+    0x1A5C: 84,
+    0x1A5D: 84,
+    0x1A5E: 84,
+    0x1A60: 84,
+    0x1A62: 84,
+    0x1A65: 84,
+    0x1A66: 84,
+    0x1A67: 84,
+    0x1A68: 84,
+    0x1A69: 84,
+    0x1A6A: 84,
+    0x1A6B: 84,
+    0x1A6C: 84,
+    0x1A73: 84,
+    0x1A74: 84,
+    0x1A75: 84,
+    0x1A76: 84,
+    0x1A77: 84,
+    0x1A78: 84,
+    0x1A79: 84,
+    0x1A7A: 84,
+    0x1A7B: 84,
+    0x1A7C: 84,
+    0x1A7F: 84,
+    0x1AB0: 84,
+    0x1AB1: 84,
+    0x1AB2: 84,
+    0x1AB3: 84,
+    0x1AB4: 84,
+    0x1AB5: 84,
+    0x1AB6: 84,
+    0x1AB7: 84,
+    0x1AB8: 84,
+    0x1AB9: 84,
+    0x1ABA: 84,
+    0x1ABB: 84,
+    0x1ABC: 84,
+    0x1ABD: 84,
+    0x1ABE: 84,
+    0x1ABF: 84,
+    0x1AC0: 84,
+    0x1AC1: 84,
+    0x1AC2: 84,
+    0x1AC3: 84,
+    0x1AC4: 84,
+    0x1AC5: 84,
+    0x1AC6: 84,
+    0x1AC7: 84,
+    0x1AC8: 84,
+    0x1AC9: 84,
+    0x1ACA: 84,
+    0x1ACB: 84,
+    0x1ACC: 84,
+    0x1ACD: 84,
+    0x1ACE: 84,
+    0x1B00: 84,
+    0x1B01: 84,
+    0x1B02: 84,
+    0x1B03: 84,
+    0x1B34: 84,
+    0x1B36: 84,
+    0x1B37: 84,
+    0x1B38: 84,
+    0x1B39: 84,
+    0x1B3A: 84,
+    0x1B3C: 84,
+    0x1B42: 84,
+    0x1B6B: 84,
+    0x1B6C: 84,
+    0x1B6D: 84,
+    0x1B6E: 84,
+    0x1B6F: 84,
+    0x1B70: 84,
+    0x1B71: 84,
+    0x1B72: 84,
+    0x1B73: 84,
+    0x1B80: 84,
+    0x1B81: 84,
+    0x1BA2: 84,
+    0x1BA3: 84,
+    0x1BA4: 84,
+    0x1BA5: 84,
+    0x1BA8: 84,
+    0x1BA9: 84,
+    0x1BAB: 84,
+    0x1BAC: 84,
+    0x1BAD: 84,
+    0x1BE6: 84,
+    0x1BE8: 84,
+    0x1BE9: 84,
+    0x1BED: 84,
+    0x1BEF: 84,
+    0x1BF0: 84,
+    0x1BF1: 84,
+    0x1C2C: 84,
+    0x1C2D: 84,
+    0x1C2E: 84,
+    0x1C2F: 84,
+    0x1C30: 84,
+    0x1C31: 84,
+    0x1C32: 84,
+    0x1C33: 84,
+    0x1C36: 84,
+    0x1C37: 84,
+    0x1CD0: 84,
+    0x1CD1: 84,
+    0x1CD2: 84,
+    0x1CD4: 84,
+    0x1CD5: 84,
+    0x1CD6: 84,
+    0x1CD7: 84,
+    0x1CD8: 84,
+    0x1CD9: 84,
+    0x1CDA: 84,
+    0x1CDB: 84,
+    0x1CDC: 84,
+    0x1CDD: 84,
+    0x1CDE: 84,
+    0x1CDF: 84,
+    0x1CE0: 84,
+    0x1CE2: 84,
+    0x1CE3: 84,
+    0x1CE4: 84,
+    0x1CE5: 84,
+    0x1CE6: 84,
+    0x1CE7: 84,
+    0x1CE8: 84,
+    0x1CED: 84,
+    0x1CF4: 84,
+    0x1CF8: 84,
+    0x1CF9: 84,
+    0x1DC0: 84,
+    0x1DC1: 84,
+    0x1DC2: 84,
+    0x1DC3: 84,
+    0x1DC4: 84,
+    0x1DC5: 84,
+    0x1DC6: 84,
+    0x1DC7: 84,
+    0x1DC8: 84,
+    0x1DC9: 84,
+    0x1DCA: 84,
+    0x1DCB: 84,
+    0x1DCC: 84,
+    0x1DCD: 84,
+    0x1DCE: 84,
+    0x1DCF: 84,
+    0x1DD0: 84,
+    0x1DD1: 84,
+    0x1DD2: 84,
+    0x1DD3: 84,
+    0x1DD4: 84,
+    0x1DD5: 84,
+    0x1DD6: 84,
+    0x1DD7: 84,
+    0x1DD8: 84,
+    0x1DD9: 84,
+    0x1DDA: 84,
+    0x1DDB: 84,
+    0x1DDC: 84,
+    0x1DDD: 84,
+    0x1DDE: 84,
+    0x1DDF: 84,
+    0x1DE0: 84,
+    0x1DE1: 84,
+    0x1DE2: 84,
+    0x1DE3: 84,
+    0x1DE4: 84,
+    0x1DE5: 84,
+    0x1DE6: 84,
+    0x1DE7: 84,
+    0x1DE8: 84,
+    0x1DE9: 84,
+    0x1DEA: 84,
+    0x1DEB: 84,
+    0x1DEC: 84,
+    0x1DED: 84,
+    0x1DEE: 84,
+    0x1DEF: 84,
+    0x1DF0: 84,
+    0x1DF1: 84,
+    0x1DF2: 84,
+    0x1DF3: 84,
+    0x1DF4: 84,
+    0x1DF5: 84,
+    0x1DF6: 84,
+    0x1DF7: 84,
+    0x1DF8: 84,
+    0x1DF9: 84,
+    0x1DFA: 84,
+    0x1DFB: 84,
+    0x1DFC: 84,
+    0x1DFD: 84,
+    0x1DFE: 84,
+    0x1DFF: 84,
+    0x200B: 84,
+    0x200D: 67,
+    0x200E: 84,
+    0x200F: 84,
+    0x202A: 84,
+    0x202B: 84,
+    0x202C: 84,
+    0x202D: 84,
+    0x202E: 84,
+    0x2060: 84,
+    0x2061: 84,
+    0x2062: 84,
+    0x2063: 84,
+    0x2064: 84,
+    0x206A: 84,
+    0x206B: 84,
+    0x206C: 84,
+    0x206D: 84,
+    0x206E: 84,
+    0x206F: 84,
+    0x20D0: 84,
+    0x20D1: 84,
+    0x20D2: 84,
+    0x20D3: 84,
+    0x20D4: 84,
+    0x20D5: 84,
+    0x20D6: 84,
+    0x20D7: 84,
+    0x20D8: 84,
+    0x20D9: 84,
+    0x20DA: 84,
+    0x20DB: 84,
+    0x20DC: 84,
+    0x20DD: 84,
+    0x20DE: 84,
+    0x20DF: 84,
+    0x20E0: 84,
+    0x20E1: 84,
+    0x20E2: 84,
+    0x20E3: 84,
+    0x20E4: 84,
+    0x20E5: 84,
+    0x20E6: 84,
+    0x20E7: 84,
+    0x20E8: 84,
+    0x20E9: 84,
+    0x20EA: 84,
+    0x20EB: 84,
+    0x20EC: 84,
+    0x20ED: 84,
+    0x20EE: 84,
+    0x20EF: 84,
+    0x20F0: 84,
+    0x2CEF: 84,
+    0x2CF0: 84,
+    0x2CF1: 84,
+    0x2D7F: 84,
+    0x2DE0: 84,
+    0x2DE1: 84,
+    0x2DE2: 84,
+    0x2DE3: 84,
+    0x2DE4: 84,
+    0x2DE5: 84,
+    0x2DE6: 84,
+    0x2DE7: 84,
+    0x2DE8: 84,
+    0x2DE9: 84,
+    0x2DEA: 84,
+    0x2DEB: 84,
+    0x2DEC: 84,
+    0x2DED: 84,
+    0x2DEE: 84,
+    0x2DEF: 84,
+    0x2DF0: 84,
+    0x2DF1: 84,
+    0x2DF2: 84,
+    0x2DF3: 84,
+    0x2DF4: 84,
+    0x2DF5: 84,
+    0x2DF6: 84,
+    0x2DF7: 84,
+    0x2DF8: 84,
+    0x2DF9: 84,
+    0x2DFA: 84,
+    0x2DFB: 84,
+    0x2DFC: 84,
+    0x2DFD: 84,
+    0x2DFE: 84,
+    0x2DFF: 84,
+    0x302A: 84,
+    0x302B: 84,
+    0x302C: 84,
+    0x302D: 84,
+    0x3099: 84,
+    0x309A: 84,
+    0xA66F: 84,
+    0xA670: 84,
+    0xA671: 84,
+    0xA672: 84,
+    0xA674: 84,
+    0xA675: 84,
+    0xA676: 84,
+    0xA677: 84,
+    0xA678: 84,
+    0xA679: 84,
+    0xA67A: 84,
+    0xA67B: 84,
+    0xA67C: 84,
+    0xA67D: 84,
+    0xA69E: 84,
+    0xA69F: 84,
+    0xA6F0: 84,
+    0xA6F1: 84,
+    0xA802: 84,
+    0xA806: 84,
+    0xA80B: 84,
+    0xA825: 84,
+    0xA826: 84,
+    0xA82C: 84,
+    0xA840: 68,
+    0xA841: 68,
+    0xA842: 68,
+    0xA843: 68,
+    0xA844: 68,
+    0xA845: 68,
+    0xA846: 68,
+    0xA847: 68,
+    0xA848: 68,
+    0xA849: 68,
+    0xA84A: 68,
+    0xA84B: 68,
+    0xA84C: 68,
+    0xA84D: 68,
+    0xA84E: 68,
+    0xA84F: 68,
+    0xA850: 68,
+    0xA851: 68,
+    0xA852: 68,
+    0xA853: 68,
+    0xA854: 68,
+    0xA855: 68,
+    0xA856: 68,
+    0xA857: 68,
+    0xA858: 68,
+    0xA859: 68,
+    0xA85A: 68,
+    0xA85B: 68,
+    0xA85C: 68,
+    0xA85D: 68,
+    0xA85E: 68,
+    0xA85F: 68,
+    0xA860: 68,
+    0xA861: 68,
+    0xA862: 68,
+    0xA863: 68,
+    0xA864: 68,
+    0xA865: 68,
+    0xA866: 68,
+    0xA867: 68,
+    0xA868: 68,
+    0xA869: 68,
+    0xA86A: 68,
+    0xA86B: 68,
+    0xA86C: 68,
+    0xA86D: 68,
+    0xA86E: 68,
+    0xA86F: 68,
+    0xA870: 68,
+    0xA871: 68,
+    0xA872: 76,
+    0xA8C4: 84,
+    0xA8C5: 84,
+    0xA8E0: 84,
+    0xA8E1: 84,
+    0xA8E2: 84,
+    0xA8E3: 84,
+    0xA8E4: 84,
+    0xA8E5: 84,
+    0xA8E6: 84,
+    0xA8E7: 84,
+    0xA8E8: 84,
+    0xA8E9: 84,
+    0xA8EA: 84,
+    0xA8EB: 84,
+    0xA8EC: 84,
+    0xA8ED: 84,
+    0xA8EE: 84,
+    0xA8EF: 84,
+    0xA8F0: 84,
+    0xA8F1: 84,
+    0xA8FF: 84,
+    0xA926: 84,
+    0xA927: 84,
+    0xA928: 84,
+    0xA929: 84,
+    0xA92A: 84,
+    0xA92B: 84,
+    0xA92C: 84,
+    0xA92D: 84,
+    0xA947: 84,
+    0xA948: 84,
+    0xA949: 84,
+    0xA94A: 84,
+    0xA94B: 84,
+    0xA94C: 84,
+    0xA94D: 84,
+    0xA94E: 84,
+    0xA94F: 84,
+    0xA950: 84,
+    0xA951: 84,
+    0xA980: 84,
+    0xA981: 84,
+    0xA982: 84,
+    0xA9B3: 84,
+    0xA9B6: 84,
+    0xA9B7: 84,
+    0xA9B8: 84,
+    0xA9B9: 84,
+    0xA9BC: 84,
+    0xA9BD: 84,
+    0xA9E5: 84,
+    0xAA29: 84,
+    0xAA2A: 84,
+    0xAA2B: 84,
+    0xAA2C: 84,
+    0xAA2D: 84,
+    0xAA2E: 84,
+    0xAA31: 84,
+    0xAA32: 84,
+    0xAA35: 84,
+    0xAA36: 84,
+    0xAA43: 84,
+    0xAA4C: 84,
+    0xAA7C: 84,
+    0xAAB0: 84,
+    0xAAB2: 84,
+    0xAAB3: 84,
+    0xAAB4: 84,
+    0xAAB7: 84,
+    0xAAB8: 84,
+    0xAABE: 84,
+    0xAABF: 84,
+    0xAAC1: 84,
+    0xAAEC: 84,
+    0xAAED: 84,
+    0xAAF6: 84,
+    0xABE5: 84,
+    0xABE8: 84,
+    0xABED: 84,
+    0xFB1E: 84,
+    0xFE00: 84,
+    0xFE01: 84,
+    0xFE02: 84,
+    0xFE03: 84,
+    0xFE04: 84,
+    0xFE05: 84,
+    0xFE06: 84,
+    0xFE07: 84,
+    0xFE08: 84,
+    0xFE09: 84,
+    0xFE0A: 84,
+    0xFE0B: 84,
+    0xFE0C: 84,
+    0xFE0D: 84,
+    0xFE0E: 84,
+    0xFE0F: 84,
+    0xFE20: 84,
+    0xFE21: 84,
+    0xFE22: 84,
+    0xFE23: 84,
+    0xFE24: 84,
+    0xFE25: 84,
+    0xFE26: 84,
+    0xFE27: 84,
+    0xFE28: 84,
+    0xFE29: 84,
+    0xFE2A: 84,
+    0xFE2B: 84,
+    0xFE2C: 84,
+    0xFE2D: 84,
+    0xFE2E: 84,
+    0xFE2F: 84,
+    0xFEFF: 84,
+    0xFFF9: 84,
+    0xFFFA: 84,
+    0xFFFB: 84,
+    0x101FD: 84,
+    0x102E0: 84,
+    0x10376: 84,
+    0x10377: 84,
+    0x10378: 84,
+    0x10379: 84,
+    0x1037A: 84,
+    0x10A01: 84,
+    0x10A02: 84,
+    0x10A03: 84,
+    0x10A05: 84,
+    0x10A06: 84,
+    0x10A0C: 84,
+    0x10A0D: 84,
+    0x10A0E: 84,
+    0x10A0F: 84,
+    0x10A38: 84,
+    0x10A39: 84,
+    0x10A3A: 84,
+    0x10A3F: 84,
+    0x10AC0: 68,
+    0x10AC1: 68,
+    0x10AC2: 68,
+    0x10AC3: 68,
+    0x10AC4: 68,
+    0x10AC5: 82,
+    0x10AC7: 82,
+    0x10AC9: 82,
+    0x10ACA: 82,
+    0x10ACD: 76,
+    0x10ACE: 82,
+    0x10ACF: 82,
+    0x10AD0: 82,
+    0x10AD1: 82,
+    0x10AD2: 82,
+    0x10AD3: 68,
+    0x10AD4: 68,
+    0x10AD5: 68,
+    0x10AD6: 68,
+    0x10AD7: 76,
+    0x10AD8: 68,
+    0x10AD9: 68,
+    0x10ADA: 68,
+    0x10ADB: 68,
+    0x10ADC: 68,
+    0x10ADD: 82,
+    0x10ADE: 68,
+    0x10ADF: 68,
+    0x10AE0: 68,
+    0x10AE1: 82,
+    0x10AE4: 82,
+    0x10AE5: 84,
+    0x10AE6: 84,
+    0x10AEB: 68,
+    0x10AEC: 68,
+    0x10AED: 68,
+    0x10AEE: 68,
+    0x10AEF: 82,
+    0x10B80: 68,
+    0x10B81: 82,
+    0x10B82: 68,
+    0x10B83: 82,
+    0x10B84: 82,
+    0x10B85: 82,
+    0x10B86: 68,
+    0x10B87: 68,
+    0x10B88: 68,
+    0x10B89: 82,
+    0x10B8A: 68,
+    0x10B8B: 68,
+    0x10B8C: 82,
+    0x10B8D: 68,
+    0x10B8E: 82,
+    0x10B8F: 82,
+    0x10B90: 68,
+    0x10B91: 82,
+    0x10BA9: 82,
+    0x10BAA: 82,
+    0x10BAB: 82,
+    0x10BAC: 82,
+    0x10BAD: 68,
+    0x10BAE: 68,
+    0x10D00: 76,
+    0x10D01: 68,
+    0x10D02: 68,
+    0x10D03: 68,
+    0x10D04: 68,
+    0x10D05: 68,
+    0x10D06: 68,
+    0x10D07: 68,
+    0x10D08: 68,
+    0x10D09: 68,
+    0x10D0A: 68,
+    0x10D0B: 68,
+    0x10D0C: 68,
+    0x10D0D: 68,
+    0x10D0E: 68,
+    0x10D0F: 68,
+    0x10D10: 68,
+    0x10D11: 68,
+    0x10D12: 68,
+    0x10D13: 68,
+    0x10D14: 68,
+    0x10D15: 68,
+    0x10D16: 68,
+    0x10D17: 68,
+    0x10D18: 68,
+    0x10D19: 68,
+    0x10D1A: 68,
+    0x10D1B: 68,
+    0x10D1C: 68,
+    0x10D1D: 68,
+    0x10D1E: 68,
+    0x10D1F: 68,
+    0x10D20: 68,
+    0x10D21: 68,
+    0x10D22: 82,
+    0x10D23: 68,
+    0x10D24: 84,
+    0x10D25: 84,
+    0x10D26: 84,
+    0x10D27: 84,
+    0x10EAB: 84,
+    0x10EAC: 84,
+    0x10EFD: 84,
+    0x10EFE: 84,
+    0x10EFF: 84,
+    0x10F30: 68,
+    0x10F31: 68,
+    0x10F32: 68,
+    0x10F33: 82,
+    0x10F34: 68,
+    0x10F35: 68,
+    0x10F36: 68,
+    0x10F37: 68,
+    0x10F38: 68,
+    0x10F39: 68,
+    0x10F3A: 68,
+    0x10F3B: 68,
+    0x10F3C: 68,
+    0x10F3D: 68,
+    0x10F3E: 68,
+    0x10F3F: 68,
+    0x10F40: 68,
+    0x10F41: 68,
+    0x10F42: 68,
+    0x10F43: 68,
+    0x10F44: 68,
+    0x10F46: 84,
+    0x10F47: 84,
+    0x10F48: 84,
+    0x10F49: 84,
+    0x10F4A: 84,
+    0x10F4B: 84,
+    0x10F4C: 84,
+    0x10F4D: 84,
+    0x10F4E: 84,
+    0x10F4F: 84,
+    0x10F50: 84,
+    0x10F51: 68,
+    0x10F52: 68,
+    0x10F53: 68,
+    0x10F54: 82,
+    0x10F70: 68,
+    0x10F71: 68,
+    0x10F72: 68,
+    0x10F73: 68,
+    0x10F74: 82,
+    0x10F75: 82,
+    0x10F76: 68,
+    0x10F77: 68,
+    0x10F78: 68,
+    0x10F79: 68,
+    0x10F7A: 68,
+    0x10F7B: 68,
+    0x10F7C: 68,
+    0x10F7D: 68,
+    0x10F7E: 68,
+    0x10F7F: 68,
+    0x10F80: 68,
+    0x10F81: 68,
+    0x10F82: 84,
+    0x10F83: 84,
+    0x10F84: 84,
+    0x10F85: 84,
+    0x10FB0: 68,
+    0x10FB2: 68,
+    0x10FB3: 68,
+    0x10FB4: 82,
+    0x10FB5: 82,
+    0x10FB6: 82,
+    0x10FB8: 68,
+    0x10FB9: 82,
+    0x10FBA: 82,
+    0x10FBB: 68,
+    0x10FBC: 68,
+    0x10FBD: 82,
+    0x10FBE: 68,
+    0x10FBF: 68,
+    0x10FC1: 68,
+    0x10FC2: 82,
+    0x10FC3: 82,
+    0x10FC4: 68,
+    0x10FC9: 82,
+    0x10FCA: 68,
+    0x10FCB: 76,
+    0x11001: 84,
+    0x11038: 84,
+    0x11039: 84,
+    0x1103A: 84,
+    0x1103B: 84,
+    0x1103C: 84,
+    0x1103D: 84,
+    0x1103E: 84,
+    0x1103F: 84,
+    0x11040: 84,
+    0x11041: 84,
+    0x11042: 84,
+    0x11043: 84,
+    0x11044: 84,
+    0x11045: 84,
+    0x11046: 84,
+    0x11070: 84,
+    0x11073: 84,
+    0x11074: 84,
+    0x1107F: 84,
+    0x11080: 84,
+    0x11081: 84,
+    0x110B3: 84,
+    0x110B4: 84,
+    0x110B5: 84,
+    0x110B6: 84,
+    0x110B9: 84,
+    0x110BA: 84,
+    0x110C2: 84,
+    0x11100: 84,
+    0x11101: 84,
+    0x11102: 84,
+    0x11127: 84,
+    0x11128: 84,
+    0x11129: 84,
+    0x1112A: 84,
+    0x1112B: 84,
+    0x1112D: 84,
+    0x1112E: 84,
+    0x1112F: 84,
+    0x11130: 84,
+    0x11131: 84,
+    0x11132: 84,
+    0x11133: 84,
+    0x11134: 84,
+    0x11173: 84,
+    0x11180: 84,
+    0x11181: 84,
+    0x111B6: 84,
+    0x111B7: 84,
+    0x111B8: 84,
+    0x111B9: 84,
+    0x111BA: 84,
+    0x111BB: 84,
+    0x111BC: 84,
+    0x111BD: 84,
+    0x111BE: 84,
+    0x111C9: 84,
+    0x111CA: 84,
+    0x111CB: 84,
+    0x111CC: 84,
+    0x111CF: 84,
+    0x1122F: 84,
+    0x11230: 84,
+    0x11231: 84,
+    0x11234: 84,
+    0x11236: 84,
+    0x11237: 84,
+    0x1123E: 84,
+    0x11241: 84,
+    0x112DF: 84,
+    0x112E3: 84,
+    0x112E4: 84,
+    0x112E5: 84,
+    0x112E6: 84,
+    0x112E7: 84,
+    0x112E8: 84,
+    0x112E9: 84,
+    0x112EA: 84,
+    0x11300: 84,
+    0x11301: 84,
+    0x1133B: 84,
+    0x1133C: 84,
+    0x11340: 84,
+    0x11366: 84,
+    0x11367: 84,
+    0x11368: 84,
+    0x11369: 84,
+    0x1136A: 84,
+    0x1136B: 84,
+    0x1136C: 84,
+    0x11370: 84,
+    0x11371: 84,
+    0x11372: 84,
+    0x11373: 84,
+    0x11374: 84,
+    0x11438: 84,
+    0x11439: 84,
+    0x1143A: 84,
+    0x1143B: 84,
+    0x1143C: 84,
+    0x1143D: 84,
+    0x1143E: 84,
+    0x1143F: 84,
+    0x11442: 84,
+    0x11443: 84,
+    0x11444: 84,
+    0x11446: 84,
+    0x1145E: 84,
+    0x114B3: 84,
+    0x114B4: 84,
+    0x114B5: 84,
+    0x114B6: 84,
+    0x114B7: 84,
+    0x114B8: 84,
+    0x114BA: 84,
+    0x114BF: 84,
+    0x114C0: 84,
+    0x114C2: 84,
+    0x114C3: 84,
+    0x115B2: 84,
+    0x115B3: 84,
+    0x115B4: 84,
+    0x115B5: 84,
+    0x115BC: 84,
+    0x115BD: 84,
+    0x115BF: 84,
+    0x115C0: 84,
+    0x115DC: 84,
+    0x115DD: 84,
+    0x11633: 84,
+    0x11634: 84,
+    0x11635: 84,
+    0x11636: 84,
+    0x11637: 84,
+    0x11638: 84,
+    0x11639: 84,
+    0x1163A: 84,
+    0x1163D: 84,
+    0x1163F: 84,
+    0x11640: 84,
+    0x116AB: 84,
+    0x116AD: 84,
+    0x116B0: 84,
+    0x116B1: 84,
+    0x116B2: 84,
+    0x116B3: 84,
+    0x116B4: 84,
+    0x116B5: 84,
+    0x116B7: 84,
+    0x1171D: 84,
+    0x1171E: 84,
+    0x1171F: 84,
+    0x11722: 84,
+    0x11723: 84,
+    0x11724: 84,
+    0x11725: 84,
+    0x11727: 84,
+    0x11728: 84,
+    0x11729: 84,
+    0x1172A: 84,
+    0x1172B: 84,
+    0x1182F: 84,
+    0x11830: 84,
+    0x11831: 84,
+    0x11832: 84,
+    0x11833: 84,
+    0x11834: 84,
+    0x11835: 84,
+    0x11836: 84,
+    0x11837: 84,
+    0x11839: 84,
+    0x1183A: 84,
+    0x1193B: 84,
+    0x1193C: 84,
+    0x1193E: 84,
+    0x11943: 84,
+    0x119D4: 84,
+    0x119D5: 84,
+    0x119D6: 84,
+    0x119D7: 84,
+    0x119DA: 84,
+    0x119DB: 84,
+    0x119E0: 84,
+    0x11A01: 84,
+    0x11A02: 84,
+    0x11A03: 84,
+    0x11A04: 84,
+    0x11A05: 84,
+    0x11A06: 84,
+    0x11A07: 84,
+    0x11A08: 84,
+    0x11A09: 84,
+    0x11A0A: 84,
+    0x11A33: 84,
+    0x11A34: 84,
+    0x11A35: 84,
+    0x11A36: 84,
+    0x11A37: 84,
+    0x11A38: 84,
+    0x11A3B: 84,
+    0x11A3C: 84,
+    0x11A3D: 84,
+    0x11A3E: 84,
+    0x11A47: 84,
+    0x11A51: 84,
+    0x11A52: 84,
+    0x11A53: 84,
+    0x11A54: 84,
+    0x11A55: 84,
+    0x11A56: 84,
+    0x11A59: 84,
+    0x11A5A: 84,
+    0x11A5B: 84,
+    0x11A8A: 84,
+    0x11A8B: 84,
+    0x11A8C: 84,
+    0x11A8D: 84,
+    0x11A8E: 84,
+    0x11A8F: 84,
+    0x11A90: 84,
+    0x11A91: 84,
+    0x11A92: 84,
+    0x11A93: 84,
+    0x11A94: 84,
+    0x11A95: 84,
+    0x11A96: 84,
+    0x11A98: 84,
+    0x11A99: 84,
+    0x11C30: 84,
+    0x11C31: 84,
+    0x11C32: 84,
+    0x11C33: 84,
+    0x11C34: 84,
+    0x11C35: 84,
+    0x11C36: 84,
+    0x11C38: 84,
+    0x11C39: 84,
+    0x11C3A: 84,
+    0x11C3B: 84,
+    0x11C3C: 84,
+    0x11C3D: 84,
+    0x11C3F: 84,
+    0x11C92: 84,
+    0x11C93: 84,
+    0x11C94: 84,
+    0x11C95: 84,
+    0x11C96: 84,
+    0x11C97: 84,
+    0x11C98: 84,
+    0x11C99: 84,
+    0x11C9A: 84,
+    0x11C9B: 84,
+    0x11C9C: 84,
+    0x11C9D: 84,
+    0x11C9E: 84,
+    0x11C9F: 84,
+    0x11CA0: 84,
+    0x11CA1: 84,
+    0x11CA2: 84,
+    0x11CA3: 84,
+    0x11CA4: 84,
+    0x11CA5: 84,
+    0x11CA6: 84,
+    0x11CA7: 84,
+    0x11CAA: 84,
+    0x11CAB: 84,
+    0x11CAC: 84,
+    0x11CAD: 84,
+    0x11CAE: 84,
+    0x11CAF: 84,
+    0x11CB0: 84,
+    0x11CB2: 84,
+    0x11CB3: 84,
+    0x11CB5: 84,
+    0x11CB6: 84,
+    0x11D31: 84,
+    0x11D32: 84,
+    0x11D33: 84,
+    0x11D34: 84,
+    0x11D35: 84,
+    0x11D36: 84,
+    0x11D3A: 84,
+    0x11D3C: 84,
+    0x11D3D: 84,
+    0x11D3F: 84,
+    0x11D40: 84,
+    0x11D41: 84,
+    0x11D42: 84,
+    0x11D43: 84,
+    0x11D44: 84,
+    0x11D45: 84,
+    0x11D47: 84,
+    0x11D90: 84,
+    0x11D91: 84,
+    0x11D95: 84,
+    0x11D97: 84,
+    0x11EF3: 84,
+    0x11EF4: 84,
+    0x11F00: 84,
+    0x11F01: 84,
+    0x11F36: 84,
+    0x11F37: 84,
+    0x11F38: 84,
+    0x11F39: 84,
+    0x11F3A: 84,
+    0x11F40: 84,
+    0x11F42: 84,
+    0x13430: 84,
+    0x13431: 84,
+    0x13432: 84,
+    0x13433: 84,
+    0x13434: 84,
+    0x13435: 84,
+    0x13436: 84,
+    0x13437: 84,
+    0x13438: 84,
+    0x13439: 84,
+    0x1343A: 84,
+    0x1343B: 84,
+    0x1343C: 84,
+    0x1343D: 84,
+    0x1343E: 84,
+    0x1343F: 84,
+    0x13440: 84,
+    0x13447: 84,
+    0x13448: 84,
+    0x13449: 84,
+    0x1344A: 84,
+    0x1344B: 84,
+    0x1344C: 84,
+    0x1344D: 84,
+    0x1344E: 84,
+    0x1344F: 84,
+    0x13450: 84,
+    0x13451: 84,
+    0x13452: 84,
+    0x13453: 84,
+    0x13454: 84,
+    0x13455: 84,
+    0x16AF0: 84,
+    0x16AF1: 84,
+    0x16AF2: 84,
+    0x16AF3: 84,
+    0x16AF4: 84,
+    0x16B30: 84,
+    0x16B31: 84,
+    0x16B32: 84,
+    0x16B33: 84,
+    0x16B34: 84,
+    0x16B35: 84,
+    0x16B36: 84,
+    0x16F4F: 84,
+    0x16F8F: 84,
+    0x16F90: 84,
+    0x16F91: 84,
+    0x16F92: 84,
+    0x16FE4: 84,
+    0x1BC9D: 84,
+    0x1BC9E: 84,
+    0x1BCA0: 84,
+    0x1BCA1: 84,
+    0x1BCA2: 84,
+    0x1BCA3: 84,
+    0x1CF00: 84,
+    0x1CF01: 84,
+    0x1CF02: 84,
+    0x1CF03: 84,
+    0x1CF04: 84,
+    0x1CF05: 84,
+    0x1CF06: 84,
+    0x1CF07: 84,
+    0x1CF08: 84,
+    0x1CF09: 84,
+    0x1CF0A: 84,
+    0x1CF0B: 84,
+    0x1CF0C: 84,
+    0x1CF0D: 84,
+    0x1CF0E: 84,
+    0x1CF0F: 84,
+    0x1CF10: 84,
+    0x1CF11: 84,
+    0x1CF12: 84,
+    0x1CF13: 84,
+    0x1CF14: 84,
+    0x1CF15: 84,
+    0x1CF16: 84,
+    0x1CF17: 84,
+    0x1CF18: 84,
+    0x1CF19: 84,
+    0x1CF1A: 84,
+    0x1CF1B: 84,
+    0x1CF1C: 84,
+    0x1CF1D: 84,
+    0x1CF1E: 84,
+    0x1CF1F: 84,
+    0x1CF20: 84,
+    0x1CF21: 84,
+    0x1CF22: 84,
+    0x1CF23: 84,
+    0x1CF24: 84,
+    0x1CF25: 84,
+    0x1CF26: 84,
+    0x1CF27: 84,
+    0x1CF28: 84,
+    0x1CF29: 84,
+    0x1CF2A: 84,
+    0x1CF2B: 84,
+    0x1CF2C: 84,
+    0x1CF2D: 84,
+    0x1CF30: 84,
+    0x1CF31: 84,
+    0x1CF32: 84,
+    0x1CF33: 84,
+    0x1CF34: 84,
+    0x1CF35: 84,
+    0x1CF36: 84,
+    0x1CF37: 84,
+    0x1CF38: 84,
+    0x1CF39: 84,
+    0x1CF3A: 84,
+    0x1CF3B: 84,
+    0x1CF3C: 84,
+    0x1CF3D: 84,
+    0x1CF3E: 84,
+    0x1CF3F: 84,
+    0x1CF40: 84,
+    0x1CF41: 84,
+    0x1CF42: 84,
+    0x1CF43: 84,
+    0x1CF44: 84,
+    0x1CF45: 84,
+    0x1CF46: 84,
+    0x1D167: 84,
+    0x1D168: 84,
+    0x1D169: 84,
+    0x1D173: 84,
+    0x1D174: 84,
+    0x1D175: 84,
+    0x1D176: 84,
+    0x1D177: 84,
+    0x1D178: 84,
+    0x1D179: 84,
+    0x1D17A: 84,
+    0x1D17B: 84,
+    0x1D17C: 84,
+    0x1D17D: 84,
+    0x1D17E: 84,
+    0x1D17F: 84,
+    0x1D180: 84,
+    0x1D181: 84,
+    0x1D182: 84,
+    0x1D185: 84,
+    0x1D186: 84,
+    0x1D187: 84,
+    0x1D188: 84,
+    0x1D189: 84,
+    0x1D18A: 84,
+    0x1D18B: 84,
+    0x1D1AA: 84,
+    0x1D1AB: 84,
+    0x1D1AC: 84,
+    0x1D1AD: 84,
+    0x1D242: 84,
+    0x1D243: 84,
+    0x1D244: 84,
+    0x1DA00: 84,
+    0x1DA01: 84,
+    0x1DA02: 84,
+    0x1DA03: 84,
+    0x1DA04: 84,
+    0x1DA05: 84,
+    0x1DA06: 84,
+    0x1DA07: 84,
+    0x1DA08: 84,
+    0x1DA09: 84,
+    0x1DA0A: 84,
+    0x1DA0B: 84,
+    0x1DA0C: 84,
+    0x1DA0D: 84,
+    0x1DA0E: 84,
+    0x1DA0F: 84,
+    0x1DA10: 84,
+    0x1DA11: 84,
+    0x1DA12: 84,
+    0x1DA13: 84,
+    0x1DA14: 84,
+    0x1DA15: 84,
+    0x1DA16: 84,
+    0x1DA17: 84,
+    0x1DA18: 84,
+    0x1DA19: 84,
+    0x1DA1A: 84,
+    0x1DA1B: 84,
+    0x1DA1C: 84,
+    0x1DA1D: 84,
+    0x1DA1E: 84,
+    0x1DA1F: 84,
+    0x1DA20: 84,
+    0x1DA21: 84,
+    0x1DA22: 84,
+    0x1DA23: 84,
+    0x1DA24: 84,
+    0x1DA25: 84,
+    0x1DA26: 84,
+    0x1DA27: 84,
+    0x1DA28: 84,
+    0x1DA29: 84,
+    0x1DA2A: 84,
+    0x1DA2B: 84,
+    0x1DA2C: 84,
+    0x1DA2D: 84,
+    0x1DA2E: 84,
+    0x1DA2F: 84,
+    0x1DA30: 84,
+    0x1DA31: 84,
+    0x1DA32: 84,
+    0x1DA33: 84,
+    0x1DA34: 84,
+    0x1DA35: 84,
+    0x1DA36: 84,
+    0x1DA3B: 84,
+    0x1DA3C: 84,
+    0x1DA3D: 84,
+    0x1DA3E: 84,
+    0x1DA3F: 84,
+    0x1DA40: 84,
+    0x1DA41: 84,
+    0x1DA42: 84,
+    0x1DA43: 84,
+    0x1DA44: 84,
+    0x1DA45: 84,
+    0x1DA46: 84,
+    0x1DA47: 84,
+    0x1DA48: 84,
+    0x1DA49: 84,
+    0x1DA4A: 84,
+    0x1DA4B: 84,
+    0x1DA4C: 84,
+    0x1DA4D: 84,
+    0x1DA4E: 84,
+    0x1DA4F: 84,
+    0x1DA50: 84,
+    0x1DA51: 84,
+    0x1DA52: 84,
+    0x1DA53: 84,
+    0x1DA54: 84,
+    0x1DA55: 84,
+    0x1DA56: 84,
+    0x1DA57: 84,
+    0x1DA58: 84,
+    0x1DA59: 84,
+    0x1DA5A: 84,
+    0x1DA5B: 84,
+    0x1DA5C: 84,
+    0x1DA5D: 84,
+    0x1DA5E: 84,
+    0x1DA5F: 84,
+    0x1DA60: 84,
+    0x1DA61: 84,
+    0x1DA62: 84,
+    0x1DA63: 84,
+    0x1DA64: 84,
+    0x1DA65: 84,
+    0x1DA66: 84,
+    0x1DA67: 84,
+    0x1DA68: 84,
+    0x1DA69: 84,
+    0x1DA6A: 84,
+    0x1DA6B: 84,
+    0x1DA6C: 84,
+    0x1DA75: 84,
+    0x1DA84: 84,
+    0x1DA9B: 84,
+    0x1DA9C: 84,
+    0x1DA9D: 84,
+    0x1DA9E: 84,
+    0x1DA9F: 84,
+    0x1DAA1: 84,
+    0x1DAA2: 84,
+    0x1DAA3: 84,
+    0x1DAA4: 84,
+    0x1DAA5: 84,
+    0x1DAA6: 84,
+    0x1DAA7: 84,
+    0x1DAA8: 84,
+    0x1DAA9: 84,
+    0x1DAAA: 84,
+    0x1DAAB: 84,
+    0x1DAAC: 84,
+    0x1DAAD: 84,
+    0x1DAAE: 84,
+    0x1DAAF: 84,
+    0x1E000: 84,
+    0x1E001: 84,
+    0x1E002: 84,
+    0x1E003: 84,
+    0x1E004: 84,
+    0x1E005: 84,
+    0x1E006: 84,
+    0x1E008: 84,
+    0x1E009: 84,
+    0x1E00A: 84,
+    0x1E00B: 84,
+    0x1E00C: 84,
+    0x1E00D: 84,
+    0x1E00E: 84,
+    0x1E00F: 84,
+    0x1E010: 84,
+    0x1E011: 84,
+    0x1E012: 84,
+    0x1E013: 84,
+    0x1E014: 84,
+    0x1E015: 84,
+    0x1E016: 84,
+    0x1E017: 84,
+    0x1E018: 84,
+    0x1E01B: 84,
+    0x1E01C: 84,
+    0x1E01D: 84,
+    0x1E01E: 84,
+    0x1E01F: 84,
+    0x1E020: 84,
+    0x1E021: 84,
+    0x1E023: 84,
+    0x1E024: 84,
+    0x1E026: 84,
+    0x1E027: 84,
+    0x1E028: 84,
+    0x1E029: 84,
+    0x1E02A: 84,
+    0x1E08F: 84,
+    0x1E130: 84,
+    0x1E131: 84,
+    0x1E132: 84,
+    0x1E133: 84,
+    0x1E134: 84,
+    0x1E135: 84,
+    0x1E136: 84,
+    0x1E2AE: 84,
+    0x1E2EC: 84,
+    0x1E2ED: 84,
+    0x1E2EE: 84,
+    0x1E2EF: 84,
+    0x1E4EC: 84,
+    0x1E4ED: 84,
+    0x1E4EE: 84,
+    0x1E4EF: 84,
+    0x1E8D0: 84,
+    0x1E8D1: 84,
+    0x1E8D2: 84,
+    0x1E8D3: 84,
+    0x1E8D4: 84,
+    0x1E8D5: 84,
+    0x1E8D6: 84,
+    0x1E900: 68,
+    0x1E901: 68,
+    0x1E902: 68,
+    0x1E903: 68,
+    0x1E904: 68,
+    0x1E905: 68,
+    0x1E906: 68,
+    0x1E907: 68,
+    0x1E908: 68,
+    0x1E909: 68,
+    0x1E90A: 68,
+    0x1E90B: 68,
+    0x1E90C: 68,
+    0x1E90D: 68,
+    0x1E90E: 68,
+    0x1E90F: 68,
+    0x1E910: 68,
+    0x1E911: 68,
+    0x1E912: 68,
+    0x1E913: 68,
+    0x1E914: 68,
+    0x1E915: 68,
+    0x1E916: 68,
+    0x1E917: 68,
+    0x1E918: 68,
+    0x1E919: 68,
+    0x1E91A: 68,
+    0x1E91B: 68,
+    0x1E91C: 68,
+    0x1E91D: 68,
+    0x1E91E: 68,
+    0x1E91F: 68,
+    0x1E920: 68,
+    0x1E921: 68,
+    0x1E922: 68,
+    0x1E923: 68,
+    0x1E924: 68,
+    0x1E925: 68,
+    0x1E926: 68,
+    0x1E927: 68,
+    0x1E928: 68,
+    0x1E929: 68,
+    0x1E92A: 68,
+    0x1E92B: 68,
+    0x1E92C: 68,
+    0x1E92D: 68,
+    0x1E92E: 68,
+    0x1E92F: 68,
+    0x1E930: 68,
+    0x1E931: 68,
+    0x1E932: 68,
+    0x1E933: 68,
+    0x1E934: 68,
+    0x1E935: 68,
+    0x1E936: 68,
+    0x1E937: 68,
+    0x1E938: 68,
+    0x1E939: 68,
+    0x1E93A: 68,
+    0x1E93B: 68,
+    0x1E93C: 68,
+    0x1E93D: 68,
+    0x1E93E: 68,
+    0x1E93F: 68,
+    0x1E940: 68,
+    0x1E941: 68,
+    0x1E942: 68,
+    0x1E943: 68,
+    0x1E944: 84,
+    0x1E945: 84,
+    0x1E946: 84,
+    0x1E947: 84,
+    0x1E948: 84,
+    0x1E949: 84,
+    0x1E94A: 84,
+    0x1E94B: 84,
+    0xE0001: 84,
+    0xE0020: 84,
+    0xE0021: 84,
+    0xE0022: 84,
+    0xE0023: 84,
+    0xE0024: 84,
+    0xE0025: 84,
+    0xE0026: 84,
+    0xE0027: 84,
+    0xE0028: 84,
+    0xE0029: 84,
+    0xE002A: 84,
+    0xE002B: 84,
+    0xE002C: 84,
+    0xE002D: 84,
+    0xE002E: 84,
+    0xE002F: 84,
+    0xE0030: 84,
+    0xE0031: 84,
+    0xE0032: 84,
+    0xE0033: 84,
+    0xE0034: 84,
+    0xE0035: 84,
+    0xE0036: 84,
+    0xE0037: 84,
+    0xE0038: 84,
+    0xE0039: 84,
+    0xE003A: 84,
+    0xE003B: 84,
+    0xE003C: 84,
+    0xE003D: 84,
+    0xE003E: 84,
+    0xE003F: 84,
+    0xE0040: 84,
+    0xE0041: 84,
+    0xE0042: 84,
+    0xE0043: 84,
+    0xE0044: 84,
+    0xE0045: 84,
+    0xE0046: 84,
+    0xE0047: 84,
+    0xE0048: 84,
+    0xE0049: 84,
+    0xE004A: 84,
+    0xE004B: 84,
+    0xE004C: 84,
+    0xE004D: 84,
+    0xE004E: 84,
+    0xE004F: 84,
+    0xE0050: 84,
+    0xE0051: 84,
+    0xE0052: 84,
+    0xE0053: 84,
+    0xE0054: 84,
+    0xE0055: 84,
+    0xE0056: 84,
+    0xE0057: 84,
+    0xE0058: 84,
+    0xE0059: 84,
+    0xE005A: 84,
+    0xE005B: 84,
+    0xE005C: 84,
+    0xE005D: 84,
+    0xE005E: 84,
+    0xE005F: 84,
+    0xE0060: 84,
+    0xE0061: 84,
+    0xE0062: 84,
+    0xE0063: 84,
+    0xE0064: 84,
+    0xE0065: 84,
+    0xE0066: 84,
+    0xE0067: 84,
+    0xE0068: 84,
+    0xE0069: 84,
+    0xE006A: 84,
+    0xE006B: 84,
+    0xE006C: 84,
+    0xE006D: 84,
+    0xE006E: 84,
+    0xE006F: 84,
+    0xE0070: 84,
+    0xE0071: 84,
+    0xE0072: 84,
+    0xE0073: 84,
+    0xE0074: 84,
+    0xE0075: 84,
+    0xE0076: 84,
+    0xE0077: 84,
+    0xE0078: 84,
+    0xE0079: 84,
+    0xE007A: 84,
+    0xE007B: 84,
+    0xE007C: 84,
+    0xE007D: 84,
+    0xE007E: 84,
+    0xE007F: 84,
+    0xE0100: 84,
+    0xE0101: 84,
+    0xE0102: 84,
+    0xE0103: 84,
+    0xE0104: 84,
+    0xE0105: 84,
+    0xE0106: 84,
+    0xE0107: 84,
+    0xE0108: 84,
+    0xE0109: 84,
+    0xE010A: 84,
+    0xE010B: 84,
+    0xE010C: 84,
+    0xE010D: 84,
+    0xE010E: 84,
+    0xE010F: 84,
+    0xE0110: 84,
+    0xE0111: 84,
+    0xE0112: 84,
+    0xE0113: 84,
+    0xE0114: 84,
+    0xE0115: 84,
+    0xE0116: 84,
+    0xE0117: 84,
+    0xE0118: 84,
+    0xE0119: 84,
+    0xE011A: 84,
+    0xE011B: 84,
+    0xE011C: 84,
+    0xE011D: 84,
+    0xE011E: 84,
+    0xE011F: 84,
+    0xE0120: 84,
+    0xE0121: 84,
+    0xE0122: 84,
+    0xE0123: 84,
+    0xE0124: 84,
+    0xE0125: 84,
+    0xE0126: 84,
+    0xE0127: 84,
+    0xE0128: 84,
+    0xE0129: 84,
+    0xE012A: 84,
+    0xE012B: 84,
+    0xE012C: 84,
+    0xE012D: 84,
+    0xE012E: 84,
+    0xE012F: 84,
+    0xE0130: 84,
+    0xE0131: 84,
+    0xE0132: 84,
+    0xE0133: 84,
+    0xE0134: 84,
+    0xE0135: 84,
+    0xE0136: 84,
+    0xE0137: 84,
+    0xE0138: 84,
+    0xE0139: 84,
+    0xE013A: 84,
+    0xE013B: 84,
+    0xE013C: 84,
+    0xE013D: 84,
+    0xE013E: 84,
+    0xE013F: 84,
+    0xE0140: 84,
+    0xE0141: 84,
+    0xE0142: 84,
+    0xE0143: 84,
+    0xE0144: 84,
+    0xE0145: 84,
+    0xE0146: 84,
+    0xE0147: 84,
+    0xE0148: 84,
+    0xE0149: 84,
+    0xE014A: 84,
+    0xE014B: 84,
+    0xE014C: 84,
+    0xE014D: 84,
+    0xE014E: 84,
+    0xE014F: 84,
+    0xE0150: 84,
+    0xE0151: 84,
+    0xE0152: 84,
+    0xE0153: 84,
+    0xE0154: 84,
+    0xE0155: 84,
+    0xE0156: 84,
+    0xE0157: 84,
+    0xE0158: 84,
+    0xE0159: 84,
+    0xE015A: 84,
+    0xE015B: 84,
+    0xE015C: 84,
+    0xE015D: 84,
+    0xE015E: 84,
+    0xE015F: 84,
+    0xE0160: 84,
+    0xE0161: 84,
+    0xE0162: 84,
+    0xE0163: 84,
+    0xE0164: 84,
+    0xE0165: 84,
+    0xE0166: 84,
+    0xE0167: 84,
+    0xE0168: 84,
+    0xE0169: 84,
+    0xE016A: 84,
+    0xE016B: 84,
+    0xE016C: 84,
+    0xE016D: 84,
+    0xE016E: 84,
+    0xE016F: 84,
+    0xE0170: 84,
+    0xE0171: 84,
+    0xE0172: 84,
+    0xE0173: 84,
+    0xE0174: 84,
+    0xE0175: 84,
+    0xE0176: 84,
+    0xE0177: 84,
+    0xE0178: 84,
+    0xE0179: 84,
+    0xE017A: 84,
+    0xE017B: 84,
+    0xE017C: 84,
+    0xE017D: 84,
+    0xE017E: 84,
+    0xE017F: 84,
+    0xE0180: 84,
+    0xE0181: 84,
+    0xE0182: 84,
+    0xE0183: 84,
+    0xE0184: 84,
+    0xE0185: 84,
+    0xE0186: 84,
+    0xE0187: 84,
+    0xE0188: 84,
+    0xE0189: 84,
+    0xE018A: 84,
+    0xE018B: 84,
+    0xE018C: 84,
+    0xE018D: 84,
+    0xE018E: 84,
+    0xE018F: 84,
+    0xE0190: 84,
+    0xE0191: 84,
+    0xE0192: 84,
+    0xE0193: 84,
+    0xE0194: 84,
+    0xE0195: 84,
+    0xE0196: 84,
+    0xE0197: 84,
+    0xE0198: 84,
+    0xE0199: 84,
+    0xE019A: 84,
+    0xE019B: 84,
+    0xE019C: 84,
+    0xE019D: 84,
+    0xE019E: 84,
+    0xE019F: 84,
+    0xE01A0: 84,
+    0xE01A1: 84,
+    0xE01A2: 84,
+    0xE01A3: 84,
+    0xE01A4: 84,
+    0xE01A5: 84,
+    0xE01A6: 84,
+    0xE01A7: 84,
+    0xE01A8: 84,
+    0xE01A9: 84,
+    0xE01AA: 84,
+    0xE01AB: 84,
+    0xE01AC: 84,
+    0xE01AD: 84,
+    0xE01AE: 84,
+    0xE01AF: 84,
+    0xE01B0: 84,
+    0xE01B1: 84,
+    0xE01B2: 84,
+    0xE01B3: 84,
+    0xE01B4: 84,
+    0xE01B5: 84,
+    0xE01B6: 84,
+    0xE01B7: 84,
+    0xE01B8: 84,
+    0xE01B9: 84,
+    0xE01BA: 84,
+    0xE01BB: 84,
+    0xE01BC: 84,
+    0xE01BD: 84,
+    0xE01BE: 84,
+    0xE01BF: 84,
+    0xE01C0: 84,
+    0xE01C1: 84,
+    0xE01C2: 84,
+    0xE01C3: 84,
+    0xE01C4: 84,
+    0xE01C5: 84,
+    0xE01C6: 84,
+    0xE01C7: 84,
+    0xE01C8: 84,
+    0xE01C9: 84,
+    0xE01CA: 84,
+    0xE01CB: 84,
+    0xE01CC: 84,
+    0xE01CD: 84,
+    0xE01CE: 84,
+    0xE01CF: 84,
+    0xE01D0: 84,
+    0xE01D1: 84,
+    0xE01D2: 84,
+    0xE01D3: 84,
+    0xE01D4: 84,
+    0xE01D5: 84,
+    0xE01D6: 84,
+    0xE01D7: 84,
+    0xE01D8: 84,
+    0xE01D9: 84,
+    0xE01DA: 84,
+    0xE01DB: 84,
+    0xE01DC: 84,
+    0xE01DD: 84,
+    0xE01DE: 84,
+    0xE01DF: 84,
+    0xE01E0: 84,
+    0xE01E1: 84,
+    0xE01E2: 84,
+    0xE01E3: 84,
+    0xE01E4: 84,
+    0xE01E5: 84,
+    0xE01E6: 84,
+    0xE01E7: 84,
+    0xE01E8: 84,
+    0xE01E9: 84,
+    0xE01EA: 84,
+    0xE01EB: 84,
+    0xE01EC: 84,
+    0xE01ED: 84,
+    0xE01EE: 84,
+    0xE01EF: 84,
+}
+codepoint_classes = {
+    "PVALID": (
+        0x2D0000002E,
+        0x300000003A,
+        0x610000007B,
+        0xDF000000F7,
+        0xF800000100,
+        0x10100000102,
+        0x10300000104,
+        0x10500000106,
+        0x10700000108,
+        0x1090000010A,
+        0x10B0000010C,
+        0x10D0000010E,
+        0x10F00000110,
+        0x11100000112,
+        0x11300000114,
+        0x11500000116,
+        0x11700000118,
+        0x1190000011A,
+        0x11B0000011C,
+        0x11D0000011E,
+        0x11F00000120,
+        0x12100000122,
+        0x12300000124,
+        0x12500000126,
+        0x12700000128,
+        0x1290000012A,
+        0x12B0000012C,
+        0x12D0000012E,
+        0x12F00000130,
+        0x13100000132,
+        0x13500000136,
+        0x13700000139,
+        0x13A0000013B,
+        0x13C0000013D,
+        0x13E0000013F,
+        0x14200000143,
+        0x14400000145,
+        0x14600000147,
+        0x14800000149,
+        0x14B0000014C,
+        0x14D0000014E,
+        0x14F00000150,
+        0x15100000152,
+        0x15300000154,
+        0x15500000156,
+        0x15700000158,
+        0x1590000015A,
+        0x15B0000015C,
+        0x15D0000015E,
+        0x15F00000160,
+        0x16100000162,
+        0x16300000164,
+        0x16500000166,
+        0x16700000168,
+        0x1690000016A,
+        0x16B0000016C,
+        0x16D0000016E,
+        0x16F00000170,
+        0x17100000172,
+        0x17300000174,
+        0x17500000176,
+        0x17700000178,
+        0x17A0000017B,
+        0x17C0000017D,
+        0x17E0000017F,
+        0x18000000181,
+        0x18300000184,
+        0x18500000186,
+        0x18800000189,
+        0x18C0000018E,
+        0x19200000193,
+        0x19500000196,
+        0x1990000019C,
+        0x19E0000019F,
+        0x1A1000001A2,
+        0x1A3000001A4,
+        0x1A5000001A6,
+        0x1A8000001A9,
+        0x1AA000001AC,
+        0x1AD000001AE,
+        0x1B0000001B1,
+        0x1B4000001B5,
+        0x1B6000001B7,
+        0x1B9000001BC,
+        0x1BD000001C4,
+        0x1CE000001CF,
+        0x1D0000001D1,
+        0x1D2000001D3,
+        0x1D4000001D5,
+        0x1D6000001D7,
+        0x1D8000001D9,
+        0x1DA000001DB,
+        0x1DC000001DE,
+        0x1DF000001E0,
+        0x1E1000001E2,
+        0x1E3000001E4,
+        0x1E5000001E6,
+        0x1E7000001E8,
+        0x1E9000001EA,
+        0x1EB000001EC,
+        0x1ED000001EE,
+        0x1EF000001F1,
+        0x1F5000001F6,
+        0x1F9000001FA,
+        0x1FB000001FC,
+        0x1FD000001FE,
+        0x1FF00000200,
+        0x20100000202,
+        0x20300000204,
+        0x20500000206,
+        0x20700000208,
+        0x2090000020A,
+        0x20B0000020C,
+        0x20D0000020E,
+        0x20F00000210,
+        0x21100000212,
+        0x21300000214,
+        0x21500000216,
+        0x21700000218,
+        0x2190000021A,
+        0x21B0000021C,
+        0x21D0000021E,
+        0x21F00000220,
+        0x22100000222,
+        0x22300000224,
+        0x22500000226,
+        0x22700000228,
+        0x2290000022A,
+        0x22B0000022C,
+        0x22D0000022E,
+        0x22F00000230,
+        0x23100000232,
+        0x2330000023A,
+        0x23C0000023D,
+        0x23F00000241,
+        0x24200000243,
+        0x24700000248,
+        0x2490000024A,
+        0x24B0000024C,
+        0x24D0000024E,
+        0x24F000002B0,
+        0x2B9000002C2,
+        0x2C6000002D2,
+        0x2EC000002ED,
+        0x2EE000002EF,
+        0x30000000340,
+        0x34200000343,
+        0x3460000034F,
+        0x35000000370,
+        0x37100000372,
+        0x37300000374,
+        0x37700000378,
+        0x37B0000037E,
+        0x39000000391,
+        0x3AC000003CF,
+        0x3D7000003D8,
+        0x3D9000003DA,
+        0x3DB000003DC,
+        0x3DD000003DE,
+        0x3DF000003E0,
+        0x3E1000003E2,
+        0x3E3000003E4,
+        0x3E5000003E6,
+        0x3E7000003E8,
+        0x3E9000003EA,
+        0x3EB000003EC,
+        0x3ED000003EE,
+        0x3EF000003F0,
+        0x3F3000003F4,
+        0x3F8000003F9,
+        0x3FB000003FD,
+        0x43000000460,
+        0x46100000462,
+        0x46300000464,
+        0x46500000466,
+        0x46700000468,
+        0x4690000046A,
+        0x46B0000046C,
+        0x46D0000046E,
+        0x46F00000470,
+        0x47100000472,
+        0x47300000474,
+        0x47500000476,
+        0x47700000478,
+        0x4790000047A,
+        0x47B0000047C,
+        0x47D0000047E,
+        0x47F00000480,
+        0x48100000482,
+        0x48300000488,
+        0x48B0000048C,
+        0x48D0000048E,
+        0x48F00000490,
+        0x49100000492,
+        0x49300000494,
+        0x49500000496,
+        0x49700000498,
+        0x4990000049A,
+        0x49B0000049C,
+        0x49D0000049E,
+        0x49F000004A0,
+        0x4A1000004A2,
+        0x4A3000004A4,
+        0x4A5000004A6,
+        0x4A7000004A8,
+        0x4A9000004AA,
+        0x4AB000004AC,
+        0x4AD000004AE,
+        0x4AF000004B0,
+        0x4B1000004B2,
+        0x4B3000004B4,
+        0x4B5000004B6,
+        0x4B7000004B8,
+        0x4B9000004BA,
+        0x4BB000004BC,
+        0x4BD000004BE,
+        0x4BF000004C0,
+        0x4C2000004C3,
+        0x4C4000004C5,
+        0x4C6000004C7,
+        0x4C8000004C9,
+        0x4CA000004CB,
+        0x4CC000004CD,
+        0x4CE000004D0,
+        0x4D1000004D2,
+        0x4D3000004D4,
+        0x4D5000004D6,
+        0x4D7000004D8,
+        0x4D9000004DA,
+        0x4DB000004DC,
+        0x4DD000004DE,
+        0x4DF000004E0,
+        0x4E1000004E2,
+        0x4E3000004E4,
+        0x4E5000004E6,
+        0x4E7000004E8,
+        0x4E9000004EA,
+        0x4EB000004EC,
+        0x4ED000004EE,
+        0x4EF000004F0,
+        0x4F1000004F2,
+        0x4F3000004F4,
+        0x4F5000004F6,
+        0x4F7000004F8,
+        0x4F9000004FA,
+        0x4FB000004FC,
+        0x4FD000004FE,
+        0x4FF00000500,
+        0x50100000502,
+        0x50300000504,
+        0x50500000506,
+        0x50700000508,
+        0x5090000050A,
+        0x50B0000050C,
+        0x50D0000050E,
+        0x50F00000510,
+        0x51100000512,
+        0x51300000514,
+        0x51500000516,
+        0x51700000518,
+        0x5190000051A,
+        0x51B0000051C,
+        0x51D0000051E,
+        0x51F00000520,
+        0x52100000522,
+        0x52300000524,
+        0x52500000526,
+        0x52700000528,
+        0x5290000052A,
+        0x52B0000052C,
+        0x52D0000052E,
+        0x52F00000530,
+        0x5590000055A,
+        0x56000000587,
+        0x58800000589,
+        0x591000005BE,
+        0x5BF000005C0,
+        0x5C1000005C3,
+        0x5C4000005C6,
+        0x5C7000005C8,
+        0x5D0000005EB,
+        0x5EF000005F3,
+        0x6100000061B,
+        0x62000000640,
+        0x64100000660,
+        0x66E00000675,
+        0x679000006D4,
+        0x6D5000006DD,
+        0x6DF000006E9,
+        0x6EA000006F0,
+        0x6FA00000700,
+        0x7100000074B,
+        0x74D000007B2,
+        0x7C0000007F6,
+        0x7FD000007FE,
+        0x8000000082E,
+        0x8400000085C,
+        0x8600000086B,
+        0x87000000888,
+        0x8890000088F,
+        0x898000008E2,
+        0x8E300000958,
+        0x96000000964,
+        0x96600000970,
+        0x97100000984,
+        0x9850000098D,
+        0x98F00000991,
+        0x993000009A9,
+        0x9AA000009B1,
+        0x9B2000009B3,
+        0x9B6000009BA,
+        0x9BC000009C5,
+        0x9C7000009C9,
+        0x9CB000009CF,
+        0x9D7000009D8,
+        0x9E0000009E4,
+        0x9E6000009F2,
+        0x9FC000009FD,
+        0x9FE000009FF,
+        0xA0100000A04,
+        0xA0500000A0B,
+        0xA0F00000A11,
+        0xA1300000A29,
+        0xA2A00000A31,
+        0xA3200000A33,
+        0xA3500000A36,
+        0xA3800000A3A,
+        0xA3C00000A3D,
+        0xA3E00000A43,
+        0xA4700000A49,
+        0xA4B00000A4E,
+        0xA5100000A52,
+        0xA5C00000A5D,
+        0xA6600000A76,
+        0xA8100000A84,
+        0xA8500000A8E,
+        0xA8F00000A92,
+        0xA9300000AA9,
+        0xAAA00000AB1,
+        0xAB200000AB4,
+        0xAB500000ABA,
+        0xABC00000AC6,
+        0xAC700000ACA,
+        0xACB00000ACE,
+        0xAD000000AD1,
+        0xAE000000AE4,
+        0xAE600000AF0,
+        0xAF900000B00,
+        0xB0100000B04,
+        0xB0500000B0D,
+        0xB0F00000B11,
+        0xB1300000B29,
+        0xB2A00000B31,
+        0xB3200000B34,
+        0xB3500000B3A,
+        0xB3C00000B45,
+        0xB4700000B49,
+        0xB4B00000B4E,
+        0xB5500000B58,
+        0xB5F00000B64,
+        0xB6600000B70,
+        0xB7100000B72,
+        0xB8200000B84,
+        0xB8500000B8B,
+        0xB8E00000B91,
+        0xB9200000B96,
+        0xB9900000B9B,
+        0xB9C00000B9D,
+        0xB9E00000BA0,
+        0xBA300000BA5,
+        0xBA800000BAB,
+        0xBAE00000BBA,
+        0xBBE00000BC3,
+        0xBC600000BC9,
+        0xBCA00000BCE,
+        0xBD000000BD1,
+        0xBD700000BD8,
+        0xBE600000BF0,
+        0xC0000000C0D,
+        0xC0E00000C11,
+        0xC1200000C29,
+        0xC2A00000C3A,
+        0xC3C00000C45,
+        0xC4600000C49,
+        0xC4A00000C4E,
+        0xC5500000C57,
+        0xC5800000C5B,
+        0xC5D00000C5E,
+        0xC6000000C64,
+        0xC6600000C70,
+        0xC8000000C84,
+        0xC8500000C8D,
+        0xC8E00000C91,
+        0xC9200000CA9,
+        0xCAA00000CB4,
+        0xCB500000CBA,
+        0xCBC00000CC5,
+        0xCC600000CC9,
+        0xCCA00000CCE,
+        0xCD500000CD7,
+        0xCDD00000CDF,
+        0xCE000000CE4,
+        0xCE600000CF0,
+        0xCF100000CF4,
+        0xD0000000D0D,
+        0xD0E00000D11,
+        0xD1200000D45,
+        0xD4600000D49,
+        0xD4A00000D4F,
+        0xD5400000D58,
+        0xD5F00000D64,
+        0xD6600000D70,
+        0xD7A00000D80,
+        0xD8100000D84,
+        0xD8500000D97,
+        0xD9A00000DB2,
+        0xDB300000DBC,
+        0xDBD00000DBE,
+        0xDC000000DC7,
+        0xDCA00000DCB,
+        0xDCF00000DD5,
+        0xDD600000DD7,
+        0xDD800000DE0,
+        0xDE600000DF0,
+        0xDF200000DF4,
+        0xE0100000E33,
+        0xE3400000E3B,
+        0xE4000000E4F,
+        0xE5000000E5A,
+        0xE8100000E83,
+        0xE8400000E85,
+        0xE8600000E8B,
+        0xE8C00000EA4,
+        0xEA500000EA6,
+        0xEA700000EB3,
+        0xEB400000EBE,
+        0xEC000000EC5,
+        0xEC600000EC7,
+        0xEC800000ECF,
+        0xED000000EDA,
+        0xEDE00000EE0,
+        0xF0000000F01,
+        0xF0B00000F0C,
+        0xF1800000F1A,
+        0xF2000000F2A,
+        0xF3500000F36,
+        0xF3700000F38,
+        0xF3900000F3A,
+        0xF3E00000F43,
+        0xF4400000F48,
+        0xF4900000F4D,
+        0xF4E00000F52,
+        0xF5300000F57,
+        0xF5800000F5C,
+        0xF5D00000F69,
+        0xF6A00000F6D,
+        0xF7100000F73,
+        0xF7400000F75,
+        0xF7A00000F81,
+        0xF8200000F85,
+        0xF8600000F93,
+        0xF9400000F98,
+        0xF9900000F9D,
+        0xF9E00000FA2,
+        0xFA300000FA7,
+        0xFA800000FAC,
+        0xFAD00000FB9,
+        0xFBA00000FBD,
+        0xFC600000FC7,
+        0x10000000104A,
+        0x10500000109E,
+        0x10D0000010FB,
+        0x10FD00001100,
+        0x120000001249,
+        0x124A0000124E,
+        0x125000001257,
+        0x125800001259,
+        0x125A0000125E,
+        0x126000001289,
+        0x128A0000128E,
+        0x1290000012B1,
+        0x12B2000012B6,
+        0x12B8000012BF,
+        0x12C0000012C1,
+        0x12C2000012C6,
+        0x12C8000012D7,
+        0x12D800001311,
+        0x131200001316,
+        0x13180000135B,
+        0x135D00001360,
+        0x138000001390,
+        0x13A0000013F6,
+        0x14010000166D,
+        0x166F00001680,
+        0x16810000169B,
+        0x16A0000016EB,
+        0x16F1000016F9,
+        0x170000001716,
+        0x171F00001735,
+        0x174000001754,
+        0x17600000176D,
+        0x176E00001771,
+        0x177200001774,
+        0x1780000017B4,
+        0x17B6000017D4,
+        0x17D7000017D8,
+        0x17DC000017DE,
+        0x17E0000017EA,
+        0x18100000181A,
+        0x182000001879,
+        0x1880000018AB,
+        0x18B0000018F6,
+        0x19000000191F,
+        0x19200000192C,
+        0x19300000193C,
+        0x19460000196E,
+        0x197000001975,
+        0x1980000019AC,
+        0x19B0000019CA,
+        0x19D0000019DA,
+        0x1A0000001A1C,
+        0x1A2000001A5F,
+        0x1A6000001A7D,
+        0x1A7F00001A8A,
+        0x1A9000001A9A,
+        0x1AA700001AA8,
+        0x1AB000001ABE,
+        0x1ABF00001ACF,
+        0x1B0000001B4D,
+        0x1B5000001B5A,
+        0x1B6B00001B74,
+        0x1B8000001BF4,
+        0x1C0000001C38,
+        0x1C4000001C4A,
+        0x1C4D00001C7E,
+        0x1CD000001CD3,
+        0x1CD400001CFB,
+        0x1D0000001D2C,
+        0x1D2F00001D30,
+        0x1D3B00001D3C,
+        0x1D4E00001D4F,
+        0x1D6B00001D78,
+        0x1D7900001D9B,
+        0x1DC000001E00,
+        0x1E0100001E02,
+        0x1E0300001E04,
+        0x1E0500001E06,
+        0x1E0700001E08,
+        0x1E0900001E0A,
+        0x1E0B00001E0C,
+        0x1E0D00001E0E,
+        0x1E0F00001E10,
+        0x1E1100001E12,
+        0x1E1300001E14,
+        0x1E1500001E16,
+        0x1E1700001E18,
+        0x1E1900001E1A,
+        0x1E1B00001E1C,
+        0x1E1D00001E1E,
+        0x1E1F00001E20,
+        0x1E2100001E22,
+        0x1E2300001E24,
+        0x1E2500001E26,
+        0x1E2700001E28,
+        0x1E2900001E2A,
+        0x1E2B00001E2C,
+        0x1E2D00001E2E,
+        0x1E2F00001E30,
+        0x1E3100001E32,
+        0x1E3300001E34,
+        0x1E3500001E36,
+        0x1E3700001E38,
+        0x1E3900001E3A,
+        0x1E3B00001E3C,
+        0x1E3D00001E3E,
+        0x1E3F00001E40,
+        0x1E4100001E42,
+        0x1E4300001E44,
+        0x1E4500001E46,
+        0x1E4700001E48,
+        0x1E4900001E4A,
+        0x1E4B00001E4C,
+        0x1E4D00001E4E,
+        0x1E4F00001E50,
+        0x1E5100001E52,
+        0x1E5300001E54,
+        0x1E5500001E56,
+        0x1E5700001E58,
+        0x1E5900001E5A,
+        0x1E5B00001E5C,
+        0x1E5D00001E5E,
+        0x1E5F00001E60,
+        0x1E6100001E62,
+        0x1E6300001E64,
+        0x1E6500001E66,
+        0x1E6700001E68,
+        0x1E6900001E6A,
+        0x1E6B00001E6C,
+        0x1E6D00001E6E,
+        0x1E6F00001E70,
+        0x1E7100001E72,
+        0x1E7300001E74,
+        0x1E7500001E76,
+        0x1E7700001E78,
+        0x1E7900001E7A,
+        0x1E7B00001E7C,
+        0x1E7D00001E7E,
+        0x1E7F00001E80,
+        0x1E8100001E82,
+        0x1E8300001E84,
+        0x1E8500001E86,
+        0x1E8700001E88,
+        0x1E8900001E8A,
+        0x1E8B00001E8C,
+        0x1E8D00001E8E,
+        0x1E8F00001E90,
+        0x1E9100001E92,
+        0x1E9300001E94,
+        0x1E9500001E9A,
+        0x1E9C00001E9E,
+        0x1E9F00001EA0,
+        0x1EA100001EA2,
+        0x1EA300001EA4,
+        0x1EA500001EA6,
+        0x1EA700001EA8,
+        0x1EA900001EAA,
+        0x1EAB00001EAC,
+        0x1EAD00001EAE,
+        0x1EAF00001EB0,
+        0x1EB100001EB2,
+        0x1EB300001EB4,
+        0x1EB500001EB6,
+        0x1EB700001EB8,
+        0x1EB900001EBA,
+        0x1EBB00001EBC,
+        0x1EBD00001EBE,
+        0x1EBF00001EC0,
+        0x1EC100001EC2,
+        0x1EC300001EC4,
+        0x1EC500001EC6,
+        0x1EC700001EC8,
+        0x1EC900001ECA,
+        0x1ECB00001ECC,
+        0x1ECD00001ECE,
+        0x1ECF00001ED0,
+        0x1ED100001ED2,
+        0x1ED300001ED4,
+        0x1ED500001ED6,
+        0x1ED700001ED8,
+        0x1ED900001EDA,
+        0x1EDB00001EDC,
+        0x1EDD00001EDE,
+        0x1EDF00001EE0,
+        0x1EE100001EE2,
+        0x1EE300001EE4,
+        0x1EE500001EE6,
+        0x1EE700001EE8,
+        0x1EE900001EEA,
+        0x1EEB00001EEC,
+        0x1EED00001EEE,
+        0x1EEF00001EF0,
+        0x1EF100001EF2,
+        0x1EF300001EF4,
+        0x1EF500001EF6,
+        0x1EF700001EF8,
+        0x1EF900001EFA,
+        0x1EFB00001EFC,
+        0x1EFD00001EFE,
+        0x1EFF00001F08,
+        0x1F1000001F16,
+        0x1F2000001F28,
+        0x1F3000001F38,
+        0x1F4000001F46,
+        0x1F5000001F58,
+        0x1F6000001F68,
+        0x1F7000001F71,
+        0x1F7200001F73,
+        0x1F7400001F75,
+        0x1F7600001F77,
+        0x1F7800001F79,
+        0x1F7A00001F7B,
+        0x1F7C00001F7D,
+        0x1FB000001FB2,
+        0x1FB600001FB7,
+        0x1FC600001FC7,
+        0x1FD000001FD3,
+        0x1FD600001FD8,
+        0x1FE000001FE3,
+        0x1FE400001FE8,
+        0x1FF600001FF7,
+        0x214E0000214F,
+        0x218400002185,
+        0x2C3000002C60,
+        0x2C6100002C62,
+        0x2C6500002C67,
+        0x2C6800002C69,
+        0x2C6A00002C6B,
+        0x2C6C00002C6D,
+        0x2C7100002C72,
+        0x2C7300002C75,
+        0x2C7600002C7C,
+        0x2C8100002C82,
+        0x2C8300002C84,
+        0x2C8500002C86,
+        0x2C8700002C88,
+        0x2C8900002C8A,
+        0x2C8B00002C8C,
+        0x2C8D00002C8E,
+        0x2C8F00002C90,
+        0x2C9100002C92,
+        0x2C9300002C94,
+        0x2C9500002C96,
+        0x2C9700002C98,
+        0x2C9900002C9A,
+        0x2C9B00002C9C,
+        0x2C9D00002C9E,
+        0x2C9F00002CA0,
+        0x2CA100002CA2,
+        0x2CA300002CA4,
+        0x2CA500002CA6,
+        0x2CA700002CA8,
+        0x2CA900002CAA,
+        0x2CAB00002CAC,
+        0x2CAD00002CAE,
+        0x2CAF00002CB0,
+        0x2CB100002CB2,
+        0x2CB300002CB4,
+        0x2CB500002CB6,
+        0x2CB700002CB8,
+        0x2CB900002CBA,
+        0x2CBB00002CBC,
+        0x2CBD00002CBE,
+        0x2CBF00002CC0,
+        0x2CC100002CC2,
+        0x2CC300002CC4,
+        0x2CC500002CC6,
+        0x2CC700002CC8,
+        0x2CC900002CCA,
+        0x2CCB00002CCC,
+        0x2CCD00002CCE,
+        0x2CCF00002CD0,
+        0x2CD100002CD2,
+        0x2CD300002CD4,
+        0x2CD500002CD6,
+        0x2CD700002CD8,
+        0x2CD900002CDA,
+        0x2CDB00002CDC,
+        0x2CDD00002CDE,
+        0x2CDF00002CE0,
+        0x2CE100002CE2,
+        0x2CE300002CE5,
+        0x2CEC00002CED,
+        0x2CEE00002CF2,
+        0x2CF300002CF4,
+        0x2D0000002D26,
+        0x2D2700002D28,
+        0x2D2D00002D2E,
+        0x2D3000002D68,
+        0x2D7F00002D97,
+        0x2DA000002DA7,
+        0x2DA800002DAF,
+        0x2DB000002DB7,
+        0x2DB800002DBF,
+        0x2DC000002DC7,
+        0x2DC800002DCF,
+        0x2DD000002DD7,
+        0x2DD800002DDF,
+        0x2DE000002E00,
+        0x2E2F00002E30,
+        0x300500003008,
+        0x302A0000302E,
+        0x303C0000303D,
+        0x304100003097,
+        0x30990000309B,
+        0x309D0000309F,
+        0x30A1000030FB,
+        0x30FC000030FF,
+        0x310500003130,
+        0x31A0000031C0,
+        0x31F000003200,
+        0x340000004DC0,
+        0x4E000000A48D,
+        0xA4D00000A4FE,
+        0xA5000000A60D,
+        0xA6100000A62C,
+        0xA6410000A642,
+        0xA6430000A644,
+        0xA6450000A646,
+        0xA6470000A648,
+        0xA6490000A64A,
+        0xA64B0000A64C,
+        0xA64D0000A64E,
+        0xA64F0000A650,
+        0xA6510000A652,
+        0xA6530000A654,
+        0xA6550000A656,
+        0xA6570000A658,
+        0xA6590000A65A,
+        0xA65B0000A65C,
+        0xA65D0000A65E,
+        0xA65F0000A660,
+        0xA6610000A662,
+        0xA6630000A664,
+        0xA6650000A666,
+        0xA6670000A668,
+        0xA6690000A66A,
+        0xA66B0000A66C,
+        0xA66D0000A670,
+        0xA6740000A67E,
+        0xA67F0000A680,
+        0xA6810000A682,
+        0xA6830000A684,
+        0xA6850000A686,
+        0xA6870000A688,
+        0xA6890000A68A,
+        0xA68B0000A68C,
+        0xA68D0000A68E,
+        0xA68F0000A690,
+        0xA6910000A692,
+        0xA6930000A694,
+        0xA6950000A696,
+        0xA6970000A698,
+        0xA6990000A69A,
+        0xA69B0000A69C,
+        0xA69E0000A6E6,
+        0xA6F00000A6F2,
+        0xA7170000A720,
+        0xA7230000A724,
+        0xA7250000A726,
+        0xA7270000A728,
+        0xA7290000A72A,
+        0xA72B0000A72C,
+        0xA72D0000A72E,
+        0xA72F0000A732,
+        0xA7330000A734,
+        0xA7350000A736,
+        0xA7370000A738,
+        0xA7390000A73A,
+        0xA73B0000A73C,
+        0xA73D0000A73E,
+        0xA73F0000A740,
+        0xA7410000A742,
+        0xA7430000A744,
+        0xA7450000A746,
+        0xA7470000A748,
+        0xA7490000A74A,
+        0xA74B0000A74C,
+        0xA74D0000A74E,
+        0xA74F0000A750,
+        0xA7510000A752,
+        0xA7530000A754,
+        0xA7550000A756,
+        0xA7570000A758,
+        0xA7590000A75A,
+        0xA75B0000A75C,
+        0xA75D0000A75E,
+        0xA75F0000A760,
+        0xA7610000A762,
+        0xA7630000A764,
+        0xA7650000A766,
+        0xA7670000A768,
+        0xA7690000A76A,
+        0xA76B0000A76C,
+        0xA76D0000A76E,
+        0xA76F0000A770,
+        0xA7710000A779,
+        0xA77A0000A77B,
+        0xA77C0000A77D,
+        0xA77F0000A780,
+        0xA7810000A782,
+        0xA7830000A784,
+        0xA7850000A786,
+        0xA7870000A789,
+        0xA78C0000A78D,
+        0xA78E0000A790,
+        0xA7910000A792,
+        0xA7930000A796,
+        0xA7970000A798,
+        0xA7990000A79A,
+        0xA79B0000A79C,
+        0xA79D0000A79E,
+        0xA79F0000A7A0,
+        0xA7A10000A7A2,
+        0xA7A30000A7A4,
+        0xA7A50000A7A6,
+        0xA7A70000A7A8,
+        0xA7A90000A7AA,
+        0xA7AF0000A7B0,
+        0xA7B50000A7B6,
+        0xA7B70000A7B8,
+        0xA7B90000A7BA,
+        0xA7BB0000A7BC,
+        0xA7BD0000A7BE,
+        0xA7BF0000A7C0,
+        0xA7C10000A7C2,
+        0xA7C30000A7C4,
+        0xA7C80000A7C9,
+        0xA7CA0000A7CB,
+        0xA7D10000A7D2,
+        0xA7D30000A7D4,
+        0xA7D50000A7D6,
+        0xA7D70000A7D8,
+        0xA7D90000A7DA,
+        0xA7F60000A7F8,
+        0xA7FA0000A828,
+        0xA82C0000A82D,
+        0xA8400000A874,
+        0xA8800000A8C6,
+        0xA8D00000A8DA,
+        0xA8E00000A8F8,
+        0xA8FB0000A8FC,
+        0xA8FD0000A92E,
+        0xA9300000A954,
+        0xA9800000A9C1,
+        0xA9CF0000A9DA,
+        0xA9E00000A9FF,
+        0xAA000000AA37,
+        0xAA400000AA4E,
+        0xAA500000AA5A,
+        0xAA600000AA77,
+        0xAA7A0000AAC3,
+        0xAADB0000AADE,
+        0xAAE00000AAF0,
+        0xAAF20000AAF7,
+        0xAB010000AB07,
+        0xAB090000AB0F,
+        0xAB110000AB17,
+        0xAB200000AB27,
+        0xAB280000AB2F,
+        0xAB300000AB5B,
+        0xAB600000AB69,
+        0xABC00000ABEB,
+        0xABEC0000ABEE,
+        0xABF00000ABFA,
+        0xAC000000D7A4,
+        0xFA0E0000FA10,
+        0xFA110000FA12,
+        0xFA130000FA15,
+        0xFA1F0000FA20,
+        0xFA210000FA22,
+        0xFA230000FA25,
+        0xFA270000FA2A,
+        0xFB1E0000FB1F,
+        0xFE200000FE30,
+        0xFE730000FE74,
+        0x100000001000C,
+        0x1000D00010027,
+        0x100280001003B,
+        0x1003C0001003E,
+        0x1003F0001004E,
+        0x100500001005E,
+        0x10080000100FB,
+        0x101FD000101FE,
+        0x102800001029D,
+        0x102A0000102D1,
+        0x102E0000102E1,
+        0x1030000010320,
+        0x1032D00010341,
+        0x103420001034A,
+        0x103500001037B,
+        0x103800001039E,
+        0x103A0000103C4,
+        0x103C8000103D0,
+        0x104280001049E,
+        0x104A0000104AA,
+        0x104D8000104FC,
+        0x1050000010528,
+        0x1053000010564,
+        0x10597000105A2,
+        0x105A3000105B2,
+        0x105B3000105BA,
+        0x105BB000105BD,
+        0x1060000010737,
+        0x1074000010756,
+        0x1076000010768,
+        0x1078000010781,
+        0x1080000010806,
+        0x1080800010809,
+        0x1080A00010836,
+        0x1083700010839,
+        0x1083C0001083D,
+        0x1083F00010856,
+        0x1086000010877,
+        0x108800001089F,
+        0x108E0000108F3,
+        0x108F4000108F6,
+        0x1090000010916,
+        0x109200001093A,
+        0x10980000109B8,
+        0x109BE000109C0,
+        0x10A0000010A04,
+        0x10A0500010A07,
+        0x10A0C00010A14,
+        0x10A1500010A18,
+        0x10A1900010A36,
+        0x10A3800010A3B,
+        0x10A3F00010A40,
+        0x10A6000010A7D,
+        0x10A8000010A9D,
+        0x10AC000010AC8,
+        0x10AC900010AE7,
+        0x10B0000010B36,
+        0x10B4000010B56,
+        0x10B6000010B73,
+        0x10B8000010B92,
+        0x10C0000010C49,
+        0x10CC000010CF3,
+        0x10D0000010D28,
+        0x10D3000010D3A,
+        0x10E8000010EAA,
+        0x10EAB00010EAD,
+        0x10EB000010EB2,
+        0x10EFD00010F1D,
+        0x10F2700010F28,
+        0x10F3000010F51,
+        0x10F7000010F86,
+        0x10FB000010FC5,
+        0x10FE000010FF7,
+        0x1100000011047,
+        0x1106600011076,
+        0x1107F000110BB,
+        0x110C2000110C3,
+        0x110D0000110E9,
+        0x110F0000110FA,
+        0x1110000011135,
+        0x1113600011140,
+        0x1114400011148,
+        0x1115000011174,
+        0x1117600011177,
+        0x11180000111C5,
+        0x111C9000111CD,
+        0x111CE000111DB,
+        0x111DC000111DD,
+        0x1120000011212,
+        0x1121300011238,
+        0x1123E00011242,
+        0x1128000011287,
+        0x1128800011289,
+        0x1128A0001128E,
+        0x1128F0001129E,
+        0x1129F000112A9,
+        0x112B0000112EB,
+        0x112F0000112FA,
+        0x1130000011304,
+        0x113050001130D,
+        0x1130F00011311,
+        0x1131300011329,
+        0x1132A00011331,
+        0x1133200011334,
+        0x113350001133A,
+        0x1133B00011345,
+        0x1134700011349,
+        0x1134B0001134E,
+        0x1135000011351,
+        0x1135700011358,
+        0x1135D00011364,
+        0x113660001136D,
+        0x1137000011375,
+        0x114000001144B,
+        0x114500001145A,
+        0x1145E00011462,
+        0x11480000114C6,
+        0x114C7000114C8,
+        0x114D0000114DA,
+        0x11580000115B6,
+        0x115B8000115C1,
+        0x115D8000115DE,
+        0x1160000011641,
+        0x1164400011645,
+        0x116500001165A,
+        0x11680000116B9,
+        0x116C0000116CA,
+        0x117000001171B,
+        0x1171D0001172C,
+        0x117300001173A,
+        0x1174000011747,
+        0x118000001183B,
+        0x118C0000118EA,
+        0x118FF00011907,
+        0x119090001190A,
+        0x1190C00011914,
+        0x1191500011917,
+        0x1191800011936,
+        0x1193700011939,
+        0x1193B00011944,
+        0x119500001195A,
+        0x119A0000119A8,
+        0x119AA000119D8,
+        0x119DA000119E2,
+        0x119E3000119E5,
+        0x11A0000011A3F,
+        0x11A4700011A48,
+        0x11A5000011A9A,
+        0x11A9D00011A9E,
+        0x11AB000011AF9,
+        0x11C0000011C09,
+        0x11C0A00011C37,
+        0x11C3800011C41,
+        0x11C5000011C5A,
+        0x11C7200011C90,
+        0x11C9200011CA8,
+        0x11CA900011CB7,
+        0x11D0000011D07,
+        0x11D0800011D0A,
+        0x11D0B00011D37,
+        0x11D3A00011D3B,
+        0x11D3C00011D3E,
+        0x11D3F00011D48,
+        0x11D5000011D5A,
+        0x11D6000011D66,
+        0x11D6700011D69,
+        0x11D6A00011D8F,
+        0x11D9000011D92,
+        0x11D9300011D99,
+        0x11DA000011DAA,
+        0x11EE000011EF7,
+        0x11F0000011F11,
+        0x11F1200011F3B,
+        0x11F3E00011F43,
+        0x11F5000011F5A,
+        0x11FB000011FB1,
+        0x120000001239A,
+        0x1248000012544,
+        0x12F9000012FF1,
+        0x1300000013430,
+        0x1344000013456,
+        0x1440000014647,
+        0x1680000016A39,
+        0x16A4000016A5F,
+        0x16A6000016A6A,
+        0x16A7000016ABF,
+        0x16AC000016ACA,
+        0x16AD000016AEE,
+        0x16AF000016AF5,
+        0x16B0000016B37,
+        0x16B4000016B44,
+        0x16B5000016B5A,
+        0x16B6300016B78,
+        0x16B7D00016B90,
+        0x16E6000016E80,
+        0x16F0000016F4B,
+        0x16F4F00016F88,
+        0x16F8F00016FA0,
+        0x16FE000016FE2,
+        0x16FE300016FE5,
+        0x16FF000016FF2,
+        0x17000000187F8,
+        0x1880000018CD6,
+        0x18D0000018D09,
+        0x1AFF00001AFF4,
+        0x1AFF50001AFFC,
+        0x1AFFD0001AFFF,
+        0x1B0000001B123,
+        0x1B1320001B133,
+        0x1B1500001B153,
+        0x1B1550001B156,
+        0x1B1640001B168,
+        0x1B1700001B2FC,
+        0x1BC000001BC6B,
+        0x1BC700001BC7D,
+        0x1BC800001BC89,
+        0x1BC900001BC9A,
+        0x1BC9D0001BC9F,
+        0x1CF000001CF2E,
+        0x1CF300001CF47,
+        0x1DA000001DA37,
+        0x1DA3B0001DA6D,
+        0x1DA750001DA76,
+        0x1DA840001DA85,
+        0x1DA9B0001DAA0,
+        0x1DAA10001DAB0,
+        0x1DF000001DF1F,
+        0x1DF250001DF2B,
+        0x1E0000001E007,
+        0x1E0080001E019,
+        0x1E01B0001E022,
+        0x1E0230001E025,
+        0x1E0260001E02B,
+        0x1E08F0001E090,
+        0x1E1000001E12D,
+        0x1E1300001E13E,
+        0x1E1400001E14A,
+        0x1E14E0001E14F,
+        0x1E2900001E2AF,
+        0x1E2C00001E2FA,
+        0x1E4D00001E4FA,
+        0x1E7E00001E7E7,
+        0x1E7E80001E7EC,
+        0x1E7ED0001E7EF,
+        0x1E7F00001E7FF,
+        0x1E8000001E8C5,
+        0x1E8D00001E8D7,
+        0x1E9220001E94C,
+        0x1E9500001E95A,
+        0x200000002A6E0,
+        0x2A7000002B73A,
+        0x2B7400002B81E,
+        0x2B8200002CEA2,
+        0x2CEB00002EBE1,
+        0x2EBF00002EE5E,
+        0x300000003134B,
+        0x31350000323B0,
+    ),
+    "CONTEXTJ": (0x200C0000200E,),
+    "CONTEXTO": (
+        0xB7000000B8,
+        0x37500000376,
+        0x5F3000005F5,
+        0x6600000066A,
+        0x6F0000006FA,
+        0x30FB000030FC,
+    ),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/intranges.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/intranges.py
new file mode 100644
index 00000000..7bfaa8d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/intranges.py
@@ -0,0 +1,57 @@
+"""
+Given a list of integers, made up of (hopefully) a small number of long runs
+of consecutive integers, compute a representation of the form
+((start1, end1), (start2, end2) ...). Then answer the question "was x present
+in the original list?" in time O(log(# runs)).
+"""
+
+import bisect
+from typing import List, Tuple
+
+
+def intranges_from_list(list_: List[int]) -> Tuple[int, ...]:
+    """Represent a list of integers as a sequence of ranges:
+    ((start_0, end_0), (start_1, end_1), ...), such that the original
+    integers are exactly those x such that start_i <= x < end_i for some i.
+
+    Ranges are encoded as single integers (start << 32 | end), not as tuples.
+    """
+
+    sorted_list = sorted(list_)
+    ranges = []
+    last_write = -1
+    for i in range(len(sorted_list)):
+        if i + 1 < len(sorted_list):
+            if sorted_list[i] == sorted_list[i + 1] - 1:
+                continue
+        current_range = sorted_list[last_write + 1 : i + 1]
+        ranges.append(_encode_range(current_range[0], current_range[-1] + 1))
+        last_write = i
+
+    return tuple(ranges)
+
+
+def _encode_range(start: int, end: int) -> int:
+    return (start << 32) | end
+
+
+def _decode_range(r: int) -> Tuple[int, int]:
+    return (r >> 32), (r & ((1 << 32) - 1))
+
+
+def intranges_contain(int_: int, ranges: Tuple[int, ...]) -> bool:
+    """Determine if `int_` falls into one of the ranges in `ranges`."""
+    tuple_ = _encode_range(int_, 0)
+    pos = bisect.bisect_left(ranges, tuple_)
+    # we could be immediately ahead of a tuple (start, end)
+    # with start < int_ <= end
+    if pos > 0:
+        left, right = _decode_range(ranges[pos - 1])
+        if left <= int_ < right:
+            return True
+    # or we could be immediately behind a tuple (int_, end)
+    if pos < len(ranges):
+        left, _ = _decode_range(ranges[pos])
+        if left == int_:
+            return True
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/package_data.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/package_data.py
new file mode 100644
index 00000000..514ff7e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/package_data.py
@@ -0,0 +1 @@
+__version__ = "3.10"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/uts46data.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/uts46data.py
new file mode 100644
index 00000000..eb894327
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/idna/uts46data.py
@@ -0,0 +1,8681 @@
+# This file is automatically generated by tools/idna-data
+# vim: set fileencoding=utf-8 :
+
+from typing import List, Tuple, Union
+
+"""IDNA Mapping Table from UTS46."""
+
+
+__version__ = "15.1.0"
+
+
+def _seg_0() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x0, "3"),
+        (0x1, "3"),
+        (0x2, "3"),
+        (0x3, "3"),
+        (0x4, "3"),
+        (0x5, "3"),
+        (0x6, "3"),
+        (0x7, "3"),
+        (0x8, "3"),
+        (0x9, "3"),
+        (0xA, "3"),
+        (0xB, "3"),
+        (0xC, "3"),
+        (0xD, "3"),
+        (0xE, "3"),
+        (0xF, "3"),
+        (0x10, "3"),
+        (0x11, "3"),
+        (0x12, "3"),
+        (0x13, "3"),
+        (0x14, "3"),
+        (0x15, "3"),
+        (0x16, "3"),
+        (0x17, "3"),
+        (0x18, "3"),
+        (0x19, "3"),
+        (0x1A, "3"),
+        (0x1B, "3"),
+        (0x1C, "3"),
+        (0x1D, "3"),
+        (0x1E, "3"),
+        (0x1F, "3"),
+        (0x20, "3"),
+        (0x21, "3"),
+        (0x22, "3"),
+        (0x23, "3"),
+        (0x24, "3"),
+        (0x25, "3"),
+        (0x26, "3"),
+        (0x27, "3"),
+        (0x28, "3"),
+        (0x29, "3"),
+        (0x2A, "3"),
+        (0x2B, "3"),
+        (0x2C, "3"),
+        (0x2D, "V"),
+        (0x2E, "V"),
+        (0x2F, "3"),
+        (0x30, "V"),
+        (0x31, "V"),
+        (0x32, "V"),
+        (0x33, "V"),
+        (0x34, "V"),
+        (0x35, "V"),
+        (0x36, "V"),
+        (0x37, "V"),
+        (0x38, "V"),
+        (0x39, "V"),
+        (0x3A, "3"),
+        (0x3B, "3"),
+        (0x3C, "3"),
+        (0x3D, "3"),
+        (0x3E, "3"),
+        (0x3F, "3"),
+        (0x40, "3"),
+        (0x41, "M", "a"),
+        (0x42, "M", "b"),
+        (0x43, "M", "c"),
+        (0x44, "M", "d"),
+        (0x45, "M", "e"),
+        (0x46, "M", "f"),
+        (0x47, "M", "g"),
+        (0x48, "M", "h"),
+        (0x49, "M", "i"),
+        (0x4A, "M", "j"),
+        (0x4B, "M", "k"),
+        (0x4C, "M", "l"),
+        (0x4D, "M", "m"),
+        (0x4E, "M", "n"),
+        (0x4F, "M", "o"),
+        (0x50, "M", "p"),
+        (0x51, "M", "q"),
+        (0x52, "M", "r"),
+        (0x53, "M", "s"),
+        (0x54, "M", "t"),
+        (0x55, "M", "u"),
+        (0x56, "M", "v"),
+        (0x57, "M", "w"),
+        (0x58, "M", "x"),
+        (0x59, "M", "y"),
+        (0x5A, "M", "z"),
+        (0x5B, "3"),
+        (0x5C, "3"),
+        (0x5D, "3"),
+        (0x5E, "3"),
+        (0x5F, "3"),
+        (0x60, "3"),
+        (0x61, "V"),
+        (0x62, "V"),
+        (0x63, "V"),
+    ]
+
+
+def _seg_1() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x64, "V"),
+        (0x65, "V"),
+        (0x66, "V"),
+        (0x67, "V"),
+        (0x68, "V"),
+        (0x69, "V"),
+        (0x6A, "V"),
+        (0x6B, "V"),
+        (0x6C, "V"),
+        (0x6D, "V"),
+        (0x6E, "V"),
+        (0x6F, "V"),
+        (0x70, "V"),
+        (0x71, "V"),
+        (0x72, "V"),
+        (0x73, "V"),
+        (0x74, "V"),
+        (0x75, "V"),
+        (0x76, "V"),
+        (0x77, "V"),
+        (0x78, "V"),
+        (0x79, "V"),
+        (0x7A, "V"),
+        (0x7B, "3"),
+        (0x7C, "3"),
+        (0x7D, "3"),
+        (0x7E, "3"),
+        (0x7F, "3"),
+        (0x80, "X"),
+        (0x81, "X"),
+        (0x82, "X"),
+        (0x83, "X"),
+        (0x84, "X"),
+        (0x85, "X"),
+        (0x86, "X"),
+        (0x87, "X"),
+        (0x88, "X"),
+        (0x89, "X"),
+        (0x8A, "X"),
+        (0x8B, "X"),
+        (0x8C, "X"),
+        (0x8D, "X"),
+        (0x8E, "X"),
+        (0x8F, "X"),
+        (0x90, "X"),
+        (0x91, "X"),
+        (0x92, "X"),
+        (0x93, "X"),
+        (0x94, "X"),
+        (0x95, "X"),
+        (0x96, "X"),
+        (0x97, "X"),
+        (0x98, "X"),
+        (0x99, "X"),
+        (0x9A, "X"),
+        (0x9B, "X"),
+        (0x9C, "X"),
+        (0x9D, "X"),
+        (0x9E, "X"),
+        (0x9F, "X"),
+        (0xA0, "3", " "),
+        (0xA1, "V"),
+        (0xA2, "V"),
+        (0xA3, "V"),
+        (0xA4, "V"),
+        (0xA5, "V"),
+        (0xA6, "V"),
+        (0xA7, "V"),
+        (0xA8, "3", " ̈"),
+        (0xA9, "V"),
+        (0xAA, "M", "a"),
+        (0xAB, "V"),
+        (0xAC, "V"),
+        (0xAD, "I"),
+        (0xAE, "V"),
+        (0xAF, "3", " ̄"),
+        (0xB0, "V"),
+        (0xB1, "V"),
+        (0xB2, "M", "2"),
+        (0xB3, "M", "3"),
+        (0xB4, "3", " ́"),
+        (0xB5, "M", "μ"),
+        (0xB6, "V"),
+        (0xB7, "V"),
+        (0xB8, "3", " ̧"),
+        (0xB9, "M", "1"),
+        (0xBA, "M", "o"),
+        (0xBB, "V"),
+        (0xBC, "M", "1⁄4"),
+        (0xBD, "M", "1⁄2"),
+        (0xBE, "M", "3⁄4"),
+        (0xBF, "V"),
+        (0xC0, "M", "à"),
+        (0xC1, "M", "á"),
+        (0xC2, "M", "â"),
+        (0xC3, "M", "ã"),
+        (0xC4, "M", "ä"),
+        (0xC5, "M", "å"),
+        (0xC6, "M", "æ"),
+        (0xC7, "M", "ç"),
+    ]
+
+
+def _seg_2() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xC8, "M", "è"),
+        (0xC9, "M", "é"),
+        (0xCA, "M", "ê"),
+        (0xCB, "M", "ë"),
+        (0xCC, "M", "ì"),
+        (0xCD, "M", "í"),
+        (0xCE, "M", "î"),
+        (0xCF, "M", "ï"),
+        (0xD0, "M", "ð"),
+        (0xD1, "M", "ñ"),
+        (0xD2, "M", "ò"),
+        (0xD3, "M", "ó"),
+        (0xD4, "M", "ô"),
+        (0xD5, "M", "õ"),
+        (0xD6, "M", "ö"),
+        (0xD7, "V"),
+        (0xD8, "M", "ø"),
+        (0xD9, "M", "ù"),
+        (0xDA, "M", "ú"),
+        (0xDB, "M", "û"),
+        (0xDC, "M", "ü"),
+        (0xDD, "M", "ý"),
+        (0xDE, "M", "þ"),
+        (0xDF, "D", "ss"),
+        (0xE0, "V"),
+        (0xE1, "V"),
+        (0xE2, "V"),
+        (0xE3, "V"),
+        (0xE4, "V"),
+        (0xE5, "V"),
+        (0xE6, "V"),
+        (0xE7, "V"),
+        (0xE8, "V"),
+        (0xE9, "V"),
+        (0xEA, "V"),
+        (0xEB, "V"),
+        (0xEC, "V"),
+        (0xED, "V"),
+        (0xEE, "V"),
+        (0xEF, "V"),
+        (0xF0, "V"),
+        (0xF1, "V"),
+        (0xF2, "V"),
+        (0xF3, "V"),
+        (0xF4, "V"),
+        (0xF5, "V"),
+        (0xF6, "V"),
+        (0xF7, "V"),
+        (0xF8, "V"),
+        (0xF9, "V"),
+        (0xFA, "V"),
+        (0xFB, "V"),
+        (0xFC, "V"),
+        (0xFD, "V"),
+        (0xFE, "V"),
+        (0xFF, "V"),
+        (0x100, "M", "ā"),
+        (0x101, "V"),
+        (0x102, "M", "ă"),
+        (0x103, "V"),
+        (0x104, "M", "ą"),
+        (0x105, "V"),
+        (0x106, "M", "ć"),
+        (0x107, "V"),
+        (0x108, "M", "ĉ"),
+        (0x109, "V"),
+        (0x10A, "M", "ċ"),
+        (0x10B, "V"),
+        (0x10C, "M", "č"),
+        (0x10D, "V"),
+        (0x10E, "M", "ď"),
+        (0x10F, "V"),
+        (0x110, "M", "đ"),
+        (0x111, "V"),
+        (0x112, "M", "ē"),
+        (0x113, "V"),
+        (0x114, "M", "ĕ"),
+        (0x115, "V"),
+        (0x116, "M", "ė"),
+        (0x117, "V"),
+        (0x118, "M", "ę"),
+        (0x119, "V"),
+        (0x11A, "M", "ě"),
+        (0x11B, "V"),
+        (0x11C, "M", "ĝ"),
+        (0x11D, "V"),
+        (0x11E, "M", "ğ"),
+        (0x11F, "V"),
+        (0x120, "M", "ġ"),
+        (0x121, "V"),
+        (0x122, "M", "ģ"),
+        (0x123, "V"),
+        (0x124, "M", "ĥ"),
+        (0x125, "V"),
+        (0x126, "M", "ħ"),
+        (0x127, "V"),
+        (0x128, "M", "ĩ"),
+        (0x129, "V"),
+        (0x12A, "M", "ī"),
+        (0x12B, "V"),
+    ]
+
+
+def _seg_3() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x12C, "M", "ĭ"),
+        (0x12D, "V"),
+        (0x12E, "M", "į"),
+        (0x12F, "V"),
+        (0x130, "M", "i̇"),
+        (0x131, "V"),
+        (0x132, "M", "ij"),
+        (0x134, "M", "ĵ"),
+        (0x135, "V"),
+        (0x136, "M", "ķ"),
+        (0x137, "V"),
+        (0x139, "M", "ĺ"),
+        (0x13A, "V"),
+        (0x13B, "M", "ļ"),
+        (0x13C, "V"),
+        (0x13D, "M", "ľ"),
+        (0x13E, "V"),
+        (0x13F, "M", "l·"),
+        (0x141, "M", "ł"),
+        (0x142, "V"),
+        (0x143, "M", "ń"),
+        (0x144, "V"),
+        (0x145, "M", "ņ"),
+        (0x146, "V"),
+        (0x147, "M", "ň"),
+        (0x148, "V"),
+        (0x149, "M", "ʼn"),
+        (0x14A, "M", "ŋ"),
+        (0x14B, "V"),
+        (0x14C, "M", "ō"),
+        (0x14D, "V"),
+        (0x14E, "M", "ŏ"),
+        (0x14F, "V"),
+        (0x150, "M", "ő"),
+        (0x151, "V"),
+        (0x152, "M", "œ"),
+        (0x153, "V"),
+        (0x154, "M", "ŕ"),
+        (0x155, "V"),
+        (0x156, "M", "ŗ"),
+        (0x157, "V"),
+        (0x158, "M", "ř"),
+        (0x159, "V"),
+        (0x15A, "M", "ś"),
+        (0x15B, "V"),
+        (0x15C, "M", "ŝ"),
+        (0x15D, "V"),
+        (0x15E, "M", "ş"),
+        (0x15F, "V"),
+        (0x160, "M", "š"),
+        (0x161, "V"),
+        (0x162, "M", "ţ"),
+        (0x163, "V"),
+        (0x164, "M", "ť"),
+        (0x165, "V"),
+        (0x166, "M", "ŧ"),
+        (0x167, "V"),
+        (0x168, "M", "ũ"),
+        (0x169, "V"),
+        (0x16A, "M", "ū"),
+        (0x16B, "V"),
+        (0x16C, "M", "ŭ"),
+        (0x16D, "V"),
+        (0x16E, "M", "ů"),
+        (0x16F, "V"),
+        (0x170, "M", "ű"),
+        (0x171, "V"),
+        (0x172, "M", "ų"),
+        (0x173, "V"),
+        (0x174, "M", "ŵ"),
+        (0x175, "V"),
+        (0x176, "M", "ŷ"),
+        (0x177, "V"),
+        (0x178, "M", "ÿ"),
+        (0x179, "M", "ź"),
+        (0x17A, "V"),
+        (0x17B, "M", "ż"),
+        (0x17C, "V"),
+        (0x17D, "M", "ž"),
+        (0x17E, "V"),
+        (0x17F, "M", "s"),
+        (0x180, "V"),
+        (0x181, "M", "ɓ"),
+        (0x182, "M", "ƃ"),
+        (0x183, "V"),
+        (0x184, "M", "ƅ"),
+        (0x185, "V"),
+        (0x186, "M", "ɔ"),
+        (0x187, "M", "ƈ"),
+        (0x188, "V"),
+        (0x189, "M", "ɖ"),
+        (0x18A, "M", "ɗ"),
+        (0x18B, "M", "ƌ"),
+        (0x18C, "V"),
+        (0x18E, "M", "ǝ"),
+        (0x18F, "M", "ə"),
+        (0x190, "M", "ɛ"),
+        (0x191, "M", "ƒ"),
+        (0x192, "V"),
+        (0x193, "M", "ɠ"),
+    ]
+
+
+def _seg_4() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x194, "M", "ɣ"),
+        (0x195, "V"),
+        (0x196, "M", "ɩ"),
+        (0x197, "M", "ɨ"),
+        (0x198, "M", "ƙ"),
+        (0x199, "V"),
+        (0x19C, "M", "ɯ"),
+        (0x19D, "M", "ɲ"),
+        (0x19E, "V"),
+        (0x19F, "M", "ɵ"),
+        (0x1A0, "M", "ơ"),
+        (0x1A1, "V"),
+        (0x1A2, "M", "ƣ"),
+        (0x1A3, "V"),
+        (0x1A4, "M", "ƥ"),
+        (0x1A5, "V"),
+        (0x1A6, "M", "ʀ"),
+        (0x1A7, "M", "ƨ"),
+        (0x1A8, "V"),
+        (0x1A9, "M", "ʃ"),
+        (0x1AA, "V"),
+        (0x1AC, "M", "ƭ"),
+        (0x1AD, "V"),
+        (0x1AE, "M", "ʈ"),
+        (0x1AF, "M", "ư"),
+        (0x1B0, "V"),
+        (0x1B1, "M", "ʊ"),
+        (0x1B2, "M", "ʋ"),
+        (0x1B3, "M", "ƴ"),
+        (0x1B4, "V"),
+        (0x1B5, "M", "ƶ"),
+        (0x1B6, "V"),
+        (0x1B7, "M", "ʒ"),
+        (0x1B8, "M", "ƹ"),
+        (0x1B9, "V"),
+        (0x1BC, "M", "ƽ"),
+        (0x1BD, "V"),
+        (0x1C4, "M", "dž"),
+        (0x1C7, "M", "lj"),
+        (0x1CA, "M", "nj"),
+        (0x1CD, "M", "ǎ"),
+        (0x1CE, "V"),
+        (0x1CF, "M", "ǐ"),
+        (0x1D0, "V"),
+        (0x1D1, "M", "ǒ"),
+        (0x1D2, "V"),
+        (0x1D3, "M", "ǔ"),
+        (0x1D4, "V"),
+        (0x1D5, "M", "ǖ"),
+        (0x1D6, "V"),
+        (0x1D7, "M", "ǘ"),
+        (0x1D8, "V"),
+        (0x1D9, "M", "ǚ"),
+        (0x1DA, "V"),
+        (0x1DB, "M", "ǜ"),
+        (0x1DC, "V"),
+        (0x1DE, "M", "ǟ"),
+        (0x1DF, "V"),
+        (0x1E0, "M", "ǡ"),
+        (0x1E1, "V"),
+        (0x1E2, "M", "ǣ"),
+        (0x1E3, "V"),
+        (0x1E4, "M", "ǥ"),
+        (0x1E5, "V"),
+        (0x1E6, "M", "ǧ"),
+        (0x1E7, "V"),
+        (0x1E8, "M", "ǩ"),
+        (0x1E9, "V"),
+        (0x1EA, "M", "ǫ"),
+        (0x1EB, "V"),
+        (0x1EC, "M", "ǭ"),
+        (0x1ED, "V"),
+        (0x1EE, "M", "ǯ"),
+        (0x1EF, "V"),
+        (0x1F1, "M", "dz"),
+        (0x1F4, "M", "ǵ"),
+        (0x1F5, "V"),
+        (0x1F6, "M", "ƕ"),
+        (0x1F7, "M", "ƿ"),
+        (0x1F8, "M", "ǹ"),
+        (0x1F9, "V"),
+        (0x1FA, "M", "ǻ"),
+        (0x1FB, "V"),
+        (0x1FC, "M", "ǽ"),
+        (0x1FD, "V"),
+        (0x1FE, "M", "ǿ"),
+        (0x1FF, "V"),
+        (0x200, "M", "ȁ"),
+        (0x201, "V"),
+        (0x202, "M", "ȃ"),
+        (0x203, "V"),
+        (0x204, "M", "ȅ"),
+        (0x205, "V"),
+        (0x206, "M", "ȇ"),
+        (0x207, "V"),
+        (0x208, "M", "ȉ"),
+        (0x209, "V"),
+        (0x20A, "M", "ȋ"),
+        (0x20B, "V"),
+        (0x20C, "M", "ȍ"),
+    ]
+
+
+def _seg_5() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x20D, "V"),
+        (0x20E, "M", "ȏ"),
+        (0x20F, "V"),
+        (0x210, "M", "ȑ"),
+        (0x211, "V"),
+        (0x212, "M", "ȓ"),
+        (0x213, "V"),
+        (0x214, "M", "ȕ"),
+        (0x215, "V"),
+        (0x216, "M", "ȗ"),
+        (0x217, "V"),
+        (0x218, "M", "ș"),
+        (0x219, "V"),
+        (0x21A, "M", "ț"),
+        (0x21B, "V"),
+        (0x21C, "M", "ȝ"),
+        (0x21D, "V"),
+        (0x21E, "M", "ȟ"),
+        (0x21F, "V"),
+        (0x220, "M", "ƞ"),
+        (0x221, "V"),
+        (0x222, "M", "ȣ"),
+        (0x223, "V"),
+        (0x224, "M", "ȥ"),
+        (0x225, "V"),
+        (0x226, "M", "ȧ"),
+        (0x227, "V"),
+        (0x228, "M", "ȩ"),
+        (0x229, "V"),
+        (0x22A, "M", "ȫ"),
+        (0x22B, "V"),
+        (0x22C, "M", "ȭ"),
+        (0x22D, "V"),
+        (0x22E, "M", "ȯ"),
+        (0x22F, "V"),
+        (0x230, "M", "ȱ"),
+        (0x231, "V"),
+        (0x232, "M", "ȳ"),
+        (0x233, "V"),
+        (0x23A, "M", "ⱥ"),
+        (0x23B, "M", "ȼ"),
+        (0x23C, "V"),
+        (0x23D, "M", "ƚ"),
+        (0x23E, "M", "ⱦ"),
+        (0x23F, "V"),
+        (0x241, "M", "ɂ"),
+        (0x242, "V"),
+        (0x243, "M", "ƀ"),
+        (0x244, "M", "ʉ"),
+        (0x245, "M", "ʌ"),
+        (0x246, "M", "ɇ"),
+        (0x247, "V"),
+        (0x248, "M", "ɉ"),
+        (0x249, "V"),
+        (0x24A, "M", "ɋ"),
+        (0x24B, "V"),
+        (0x24C, "M", "ɍ"),
+        (0x24D, "V"),
+        (0x24E, "M", "ɏ"),
+        (0x24F, "V"),
+        (0x2B0, "M", "h"),
+        (0x2B1, "M", "ɦ"),
+        (0x2B2, "M", "j"),
+        (0x2B3, "M", "r"),
+        (0x2B4, "M", "ɹ"),
+        (0x2B5, "M", "ɻ"),
+        (0x2B6, "M", "ʁ"),
+        (0x2B7, "M", "w"),
+        (0x2B8, "M", "y"),
+        (0x2B9, "V"),
+        (0x2D8, "3", " ̆"),
+        (0x2D9, "3", " ̇"),
+        (0x2DA, "3", " ̊"),
+        (0x2DB, "3", " ̨"),
+        (0x2DC, "3", " ̃"),
+        (0x2DD, "3", " ̋"),
+        (0x2DE, "V"),
+        (0x2E0, "M", "ɣ"),
+        (0x2E1, "M", "l"),
+        (0x2E2, "M", "s"),
+        (0x2E3, "M", "x"),
+        (0x2E4, "M", "ʕ"),
+        (0x2E5, "V"),
+        (0x340, "M", "̀"),
+        (0x341, "M", "́"),
+        (0x342, "V"),
+        (0x343, "M", "̓"),
+        (0x344, "M", "̈́"),
+        (0x345, "M", "ι"),
+        (0x346, "V"),
+        (0x34F, "I"),
+        (0x350, "V"),
+        (0x370, "M", "ͱ"),
+        (0x371, "V"),
+        (0x372, "M", "ͳ"),
+        (0x373, "V"),
+        (0x374, "M", "ʹ"),
+        (0x375, "V"),
+        (0x376, "M", "ͷ"),
+        (0x377, "V"),
+    ]
+
+
+def _seg_6() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x378, "X"),
+        (0x37A, "3", " ι"),
+        (0x37B, "V"),
+        (0x37E, "3", ";"),
+        (0x37F, "M", "ϳ"),
+        (0x380, "X"),
+        (0x384, "3", " ́"),
+        (0x385, "3", " ̈́"),
+        (0x386, "M", "ά"),
+        (0x387, "M", "·"),
+        (0x388, "M", "έ"),
+        (0x389, "M", "ή"),
+        (0x38A, "M", "ί"),
+        (0x38B, "X"),
+        (0x38C, "M", "ό"),
+        (0x38D, "X"),
+        (0x38E, "M", "ύ"),
+        (0x38F, "M", "ώ"),
+        (0x390, "V"),
+        (0x391, "M", "α"),
+        (0x392, "M", "β"),
+        (0x393, "M", "γ"),
+        (0x394, "M", "δ"),
+        (0x395, "M", "ε"),
+        (0x396, "M", "ζ"),
+        (0x397, "M", "η"),
+        (0x398, "M", "θ"),
+        (0x399, "M", "ι"),
+        (0x39A, "M", "κ"),
+        (0x39B, "M", "λ"),
+        (0x39C, "M", "μ"),
+        (0x39D, "M", "ν"),
+        (0x39E, "M", "ξ"),
+        (0x39F, "M", "ο"),
+        (0x3A0, "M", "π"),
+        (0x3A1, "M", "ρ"),
+        (0x3A2, "X"),
+        (0x3A3, "M", "σ"),
+        (0x3A4, "M", "τ"),
+        (0x3A5, "M", "υ"),
+        (0x3A6, "M", "φ"),
+        (0x3A7, "M", "χ"),
+        (0x3A8, "M", "ψ"),
+        (0x3A9, "M", "ω"),
+        (0x3AA, "M", "ϊ"),
+        (0x3AB, "M", "ϋ"),
+        (0x3AC, "V"),
+        (0x3C2, "D", "σ"),
+        (0x3C3, "V"),
+        (0x3CF, "M", "ϗ"),
+        (0x3D0, "M", "β"),
+        (0x3D1, "M", "θ"),
+        (0x3D2, "M", "υ"),
+        (0x3D3, "M", "ύ"),
+        (0x3D4, "M", "ϋ"),
+        (0x3D5, "M", "φ"),
+        (0x3D6, "M", "π"),
+        (0x3D7, "V"),
+        (0x3D8, "M", "ϙ"),
+        (0x3D9, "V"),
+        (0x3DA, "M", "ϛ"),
+        (0x3DB, "V"),
+        (0x3DC, "M", "ϝ"),
+        (0x3DD, "V"),
+        (0x3DE, "M", "ϟ"),
+        (0x3DF, "V"),
+        (0x3E0, "M", "ϡ"),
+        (0x3E1, "V"),
+        (0x3E2, "M", "ϣ"),
+        (0x3E3, "V"),
+        (0x3E4, "M", "ϥ"),
+        (0x3E5, "V"),
+        (0x3E6, "M", "ϧ"),
+        (0x3E7, "V"),
+        (0x3E8, "M", "ϩ"),
+        (0x3E9, "V"),
+        (0x3EA, "M", "ϫ"),
+        (0x3EB, "V"),
+        (0x3EC, "M", "ϭ"),
+        (0x3ED, "V"),
+        (0x3EE, "M", "ϯ"),
+        (0x3EF, "V"),
+        (0x3F0, "M", "κ"),
+        (0x3F1, "M", "ρ"),
+        (0x3F2, "M", "σ"),
+        (0x3F3, "V"),
+        (0x3F4, "M", "θ"),
+        (0x3F5, "M", "ε"),
+        (0x3F6, "V"),
+        (0x3F7, "M", "ϸ"),
+        (0x3F8, "V"),
+        (0x3F9, "M", "σ"),
+        (0x3FA, "M", "ϻ"),
+        (0x3FB, "V"),
+        (0x3FD, "M", "ͻ"),
+        (0x3FE, "M", "ͼ"),
+        (0x3FF, "M", "ͽ"),
+        (0x400, "M", "ѐ"),
+        (0x401, "M", "ё"),
+        (0x402, "M", "ђ"),
+    ]
+
+
+def _seg_7() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x403, "M", "ѓ"),
+        (0x404, "M", "є"),
+        (0x405, "M", "ѕ"),
+        (0x406, "M", "і"),
+        (0x407, "M", "ї"),
+        (0x408, "M", "ј"),
+        (0x409, "M", "љ"),
+        (0x40A, "M", "њ"),
+        (0x40B, "M", "ћ"),
+        (0x40C, "M", "ќ"),
+        (0x40D, "M", "ѝ"),
+        (0x40E, "M", "ў"),
+        (0x40F, "M", "џ"),
+        (0x410, "M", "а"),
+        (0x411, "M", "б"),
+        (0x412, "M", "в"),
+        (0x413, "M", "г"),
+        (0x414, "M", "д"),
+        (0x415, "M", "е"),
+        (0x416, "M", "ж"),
+        (0x417, "M", "з"),
+        (0x418, "M", "и"),
+        (0x419, "M", "й"),
+        (0x41A, "M", "к"),
+        (0x41B, "M", "л"),
+        (0x41C, "M", "м"),
+        (0x41D, "M", "н"),
+        (0x41E, "M", "о"),
+        (0x41F, "M", "п"),
+        (0x420, "M", "р"),
+        (0x421, "M", "с"),
+        (0x422, "M", "т"),
+        (0x423, "M", "у"),
+        (0x424, "M", "ф"),
+        (0x425, "M", "х"),
+        (0x426, "M", "ц"),
+        (0x427, "M", "ч"),
+        (0x428, "M", "ш"),
+        (0x429, "M", "щ"),
+        (0x42A, "M", "ъ"),
+        (0x42B, "M", "ы"),
+        (0x42C, "M", "ь"),
+        (0x42D, "M", "э"),
+        (0x42E, "M", "ю"),
+        (0x42F, "M", "я"),
+        (0x430, "V"),
+        (0x460, "M", "ѡ"),
+        (0x461, "V"),
+        (0x462, "M", "ѣ"),
+        (0x463, "V"),
+        (0x464, "M", "ѥ"),
+        (0x465, "V"),
+        (0x466, "M", "ѧ"),
+        (0x467, "V"),
+        (0x468, "M", "ѩ"),
+        (0x469, "V"),
+        (0x46A, "M", "ѫ"),
+        (0x46B, "V"),
+        (0x46C, "M", "ѭ"),
+        (0x46D, "V"),
+        (0x46E, "M", "ѯ"),
+        (0x46F, "V"),
+        (0x470, "M", "ѱ"),
+        (0x471, "V"),
+        (0x472, "M", "ѳ"),
+        (0x473, "V"),
+        (0x474, "M", "ѵ"),
+        (0x475, "V"),
+        (0x476, "M", "ѷ"),
+        (0x477, "V"),
+        (0x478, "M", "ѹ"),
+        (0x479, "V"),
+        (0x47A, "M", "ѻ"),
+        (0x47B, "V"),
+        (0x47C, "M", "ѽ"),
+        (0x47D, "V"),
+        (0x47E, "M", "ѿ"),
+        (0x47F, "V"),
+        (0x480, "M", "ҁ"),
+        (0x481, "V"),
+        (0x48A, "M", "ҋ"),
+        (0x48B, "V"),
+        (0x48C, "M", "ҍ"),
+        (0x48D, "V"),
+        (0x48E, "M", "ҏ"),
+        (0x48F, "V"),
+        (0x490, "M", "ґ"),
+        (0x491, "V"),
+        (0x492, "M", "ғ"),
+        (0x493, "V"),
+        (0x494, "M", "ҕ"),
+        (0x495, "V"),
+        (0x496, "M", "җ"),
+        (0x497, "V"),
+        (0x498, "M", "ҙ"),
+        (0x499, "V"),
+        (0x49A, "M", "қ"),
+        (0x49B, "V"),
+        (0x49C, "M", "ҝ"),
+        (0x49D, "V"),
+    ]
+
+
+def _seg_8() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x49E, "M", "ҟ"),
+        (0x49F, "V"),
+        (0x4A0, "M", "ҡ"),
+        (0x4A1, "V"),
+        (0x4A2, "M", "ң"),
+        (0x4A3, "V"),
+        (0x4A4, "M", "ҥ"),
+        (0x4A5, "V"),
+        (0x4A6, "M", "ҧ"),
+        (0x4A7, "V"),
+        (0x4A8, "M", "ҩ"),
+        (0x4A9, "V"),
+        (0x4AA, "M", "ҫ"),
+        (0x4AB, "V"),
+        (0x4AC, "M", "ҭ"),
+        (0x4AD, "V"),
+        (0x4AE, "M", "ү"),
+        (0x4AF, "V"),
+        (0x4B0, "M", "ұ"),
+        (0x4B1, "V"),
+        (0x4B2, "M", "ҳ"),
+        (0x4B3, "V"),
+        (0x4B4, "M", "ҵ"),
+        (0x4B5, "V"),
+        (0x4B6, "M", "ҷ"),
+        (0x4B7, "V"),
+        (0x4B8, "M", "ҹ"),
+        (0x4B9, "V"),
+        (0x4BA, "M", "һ"),
+        (0x4BB, "V"),
+        (0x4BC, "M", "ҽ"),
+        (0x4BD, "V"),
+        (0x4BE, "M", "ҿ"),
+        (0x4BF, "V"),
+        (0x4C0, "X"),
+        (0x4C1, "M", "ӂ"),
+        (0x4C2, "V"),
+        (0x4C3, "M", "ӄ"),
+        (0x4C4, "V"),
+        (0x4C5, "M", "ӆ"),
+        (0x4C6, "V"),
+        (0x4C7, "M", "ӈ"),
+        (0x4C8, "V"),
+        (0x4C9, "M", "ӊ"),
+        (0x4CA, "V"),
+        (0x4CB, "M", "ӌ"),
+        (0x4CC, "V"),
+        (0x4CD, "M", "ӎ"),
+        (0x4CE, "V"),
+        (0x4D0, "M", "ӑ"),
+        (0x4D1, "V"),
+        (0x4D2, "M", "ӓ"),
+        (0x4D3, "V"),
+        (0x4D4, "M", "ӕ"),
+        (0x4D5, "V"),
+        (0x4D6, "M", "ӗ"),
+        (0x4D7, "V"),
+        (0x4D8, "M", "ә"),
+        (0x4D9, "V"),
+        (0x4DA, "M", "ӛ"),
+        (0x4DB, "V"),
+        (0x4DC, "M", "ӝ"),
+        (0x4DD, "V"),
+        (0x4DE, "M", "ӟ"),
+        (0x4DF, "V"),
+        (0x4E0, "M", "ӡ"),
+        (0x4E1, "V"),
+        (0x4E2, "M", "ӣ"),
+        (0x4E3, "V"),
+        (0x4E4, "M", "ӥ"),
+        (0x4E5, "V"),
+        (0x4E6, "M", "ӧ"),
+        (0x4E7, "V"),
+        (0x4E8, "M", "ө"),
+        (0x4E9, "V"),
+        (0x4EA, "M", "ӫ"),
+        (0x4EB, "V"),
+        (0x4EC, "M", "ӭ"),
+        (0x4ED, "V"),
+        (0x4EE, "M", "ӯ"),
+        (0x4EF, "V"),
+        (0x4F0, "M", "ӱ"),
+        (0x4F1, "V"),
+        (0x4F2, "M", "ӳ"),
+        (0x4F3, "V"),
+        (0x4F4, "M", "ӵ"),
+        (0x4F5, "V"),
+        (0x4F6, "M", "ӷ"),
+        (0x4F7, "V"),
+        (0x4F8, "M", "ӹ"),
+        (0x4F9, "V"),
+        (0x4FA, "M", "ӻ"),
+        (0x4FB, "V"),
+        (0x4FC, "M", "ӽ"),
+        (0x4FD, "V"),
+        (0x4FE, "M", "ӿ"),
+        (0x4FF, "V"),
+        (0x500, "M", "ԁ"),
+        (0x501, "V"),
+        (0x502, "M", "ԃ"),
+    ]
+
+
+def _seg_9() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x503, "V"),
+        (0x504, "M", "ԅ"),
+        (0x505, "V"),
+        (0x506, "M", "ԇ"),
+        (0x507, "V"),
+        (0x508, "M", "ԉ"),
+        (0x509, "V"),
+        (0x50A, "M", "ԋ"),
+        (0x50B, "V"),
+        (0x50C, "M", "ԍ"),
+        (0x50D, "V"),
+        (0x50E, "M", "ԏ"),
+        (0x50F, "V"),
+        (0x510, "M", "ԑ"),
+        (0x511, "V"),
+        (0x512, "M", "ԓ"),
+        (0x513, "V"),
+        (0x514, "M", "ԕ"),
+        (0x515, "V"),
+        (0x516, "M", "ԗ"),
+        (0x517, "V"),
+        (0x518, "M", "ԙ"),
+        (0x519, "V"),
+        (0x51A, "M", "ԛ"),
+        (0x51B, "V"),
+        (0x51C, "M", "ԝ"),
+        (0x51D, "V"),
+        (0x51E, "M", "ԟ"),
+        (0x51F, "V"),
+        (0x520, "M", "ԡ"),
+        (0x521, "V"),
+        (0x522, "M", "ԣ"),
+        (0x523, "V"),
+        (0x524, "M", "ԥ"),
+        (0x525, "V"),
+        (0x526, "M", "ԧ"),
+        (0x527, "V"),
+        (0x528, "M", "ԩ"),
+        (0x529, "V"),
+        (0x52A, "M", "ԫ"),
+        (0x52B, "V"),
+        (0x52C, "M", "ԭ"),
+        (0x52D, "V"),
+        (0x52E, "M", "ԯ"),
+        (0x52F, "V"),
+        (0x530, "X"),
+        (0x531, "M", "ա"),
+        (0x532, "M", "բ"),
+        (0x533, "M", "գ"),
+        (0x534, "M", "դ"),
+        (0x535, "M", "ե"),
+        (0x536, "M", "զ"),
+        (0x537, "M", "է"),
+        (0x538, "M", "ը"),
+        (0x539, "M", "թ"),
+        (0x53A, "M", "ժ"),
+        (0x53B, "M", "ի"),
+        (0x53C, "M", "լ"),
+        (0x53D, "M", "խ"),
+        (0x53E, "M", "ծ"),
+        (0x53F, "M", "կ"),
+        (0x540, "M", "հ"),
+        (0x541, "M", "ձ"),
+        (0x542, "M", "ղ"),
+        (0x543, "M", "ճ"),
+        (0x544, "M", "մ"),
+        (0x545, "M", "յ"),
+        (0x546, "M", "ն"),
+        (0x547, "M", "շ"),
+        (0x548, "M", "ո"),
+        (0x549, "M", "չ"),
+        (0x54A, "M", "պ"),
+        (0x54B, "M", "ջ"),
+        (0x54C, "M", "ռ"),
+        (0x54D, "M", "ս"),
+        (0x54E, "M", "վ"),
+        (0x54F, "M", "տ"),
+        (0x550, "M", "ր"),
+        (0x551, "M", "ց"),
+        (0x552, "M", "ւ"),
+        (0x553, "M", "փ"),
+        (0x554, "M", "ք"),
+        (0x555, "M", "օ"),
+        (0x556, "M", "ֆ"),
+        (0x557, "X"),
+        (0x559, "V"),
+        (0x587, "M", "եւ"),
+        (0x588, "V"),
+        (0x58B, "X"),
+        (0x58D, "V"),
+        (0x590, "X"),
+        (0x591, "V"),
+        (0x5C8, "X"),
+        (0x5D0, "V"),
+        (0x5EB, "X"),
+        (0x5EF, "V"),
+        (0x5F5, "X"),
+        (0x606, "V"),
+        (0x61C, "X"),
+        (0x61D, "V"),
+    ]
+
+
+def _seg_10() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x675, "M", "اٴ"),
+        (0x676, "M", "وٴ"),
+        (0x677, "M", "ۇٴ"),
+        (0x678, "M", "يٴ"),
+        (0x679, "V"),
+        (0x6DD, "X"),
+        (0x6DE, "V"),
+        (0x70E, "X"),
+        (0x710, "V"),
+        (0x74B, "X"),
+        (0x74D, "V"),
+        (0x7B2, "X"),
+        (0x7C0, "V"),
+        (0x7FB, "X"),
+        (0x7FD, "V"),
+        (0x82E, "X"),
+        (0x830, "V"),
+        (0x83F, "X"),
+        (0x840, "V"),
+        (0x85C, "X"),
+        (0x85E, "V"),
+        (0x85F, "X"),
+        (0x860, "V"),
+        (0x86B, "X"),
+        (0x870, "V"),
+        (0x88F, "X"),
+        (0x898, "V"),
+        (0x8E2, "X"),
+        (0x8E3, "V"),
+        (0x958, "M", "क़"),
+        (0x959, "M", "ख़"),
+        (0x95A, "M", "ग़"),
+        (0x95B, "M", "ज़"),
+        (0x95C, "M", "ड़"),
+        (0x95D, "M", "ढ़"),
+        (0x95E, "M", "फ़"),
+        (0x95F, "M", "य़"),
+        (0x960, "V"),
+        (0x984, "X"),
+        (0x985, "V"),
+        (0x98D, "X"),
+        (0x98F, "V"),
+        (0x991, "X"),
+        (0x993, "V"),
+        (0x9A9, "X"),
+        (0x9AA, "V"),
+        (0x9B1, "X"),
+        (0x9B2, "V"),
+        (0x9B3, "X"),
+        (0x9B6, "V"),
+        (0x9BA, "X"),
+        (0x9BC, "V"),
+        (0x9C5, "X"),
+        (0x9C7, "V"),
+        (0x9C9, "X"),
+        (0x9CB, "V"),
+        (0x9CF, "X"),
+        (0x9D7, "V"),
+        (0x9D8, "X"),
+        (0x9DC, "M", "ড়"),
+        (0x9DD, "M", "ঢ়"),
+        (0x9DE, "X"),
+        (0x9DF, "M", "য়"),
+        (0x9E0, "V"),
+        (0x9E4, "X"),
+        (0x9E6, "V"),
+        (0x9FF, "X"),
+        (0xA01, "V"),
+        (0xA04, "X"),
+        (0xA05, "V"),
+        (0xA0B, "X"),
+        (0xA0F, "V"),
+        (0xA11, "X"),
+        (0xA13, "V"),
+        (0xA29, "X"),
+        (0xA2A, "V"),
+        (0xA31, "X"),
+        (0xA32, "V"),
+        (0xA33, "M", "ਲ਼"),
+        (0xA34, "X"),
+        (0xA35, "V"),
+        (0xA36, "M", "ਸ਼"),
+        (0xA37, "X"),
+        (0xA38, "V"),
+        (0xA3A, "X"),
+        (0xA3C, "V"),
+        (0xA3D, "X"),
+        (0xA3E, "V"),
+        (0xA43, "X"),
+        (0xA47, "V"),
+        (0xA49, "X"),
+        (0xA4B, "V"),
+        (0xA4E, "X"),
+        (0xA51, "V"),
+        (0xA52, "X"),
+        (0xA59, "M", "ਖ਼"),
+        (0xA5A, "M", "ਗ਼"),
+        (0xA5B, "M", "ਜ਼"),
+        (0xA5C, "V"),
+        (0xA5D, "X"),
+    ]
+
+
+def _seg_11() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA5E, "M", "ਫ਼"),
+        (0xA5F, "X"),
+        (0xA66, "V"),
+        (0xA77, "X"),
+        (0xA81, "V"),
+        (0xA84, "X"),
+        (0xA85, "V"),
+        (0xA8E, "X"),
+        (0xA8F, "V"),
+        (0xA92, "X"),
+        (0xA93, "V"),
+        (0xAA9, "X"),
+        (0xAAA, "V"),
+        (0xAB1, "X"),
+        (0xAB2, "V"),
+        (0xAB4, "X"),
+        (0xAB5, "V"),
+        (0xABA, "X"),
+        (0xABC, "V"),
+        (0xAC6, "X"),
+        (0xAC7, "V"),
+        (0xACA, "X"),
+        (0xACB, "V"),
+        (0xACE, "X"),
+        (0xAD0, "V"),
+        (0xAD1, "X"),
+        (0xAE0, "V"),
+        (0xAE4, "X"),
+        (0xAE6, "V"),
+        (0xAF2, "X"),
+        (0xAF9, "V"),
+        (0xB00, "X"),
+        (0xB01, "V"),
+        (0xB04, "X"),
+        (0xB05, "V"),
+        (0xB0D, "X"),
+        (0xB0F, "V"),
+        (0xB11, "X"),
+        (0xB13, "V"),
+        (0xB29, "X"),
+        (0xB2A, "V"),
+        (0xB31, "X"),
+        (0xB32, "V"),
+        (0xB34, "X"),
+        (0xB35, "V"),
+        (0xB3A, "X"),
+        (0xB3C, "V"),
+        (0xB45, "X"),
+        (0xB47, "V"),
+        (0xB49, "X"),
+        (0xB4B, "V"),
+        (0xB4E, "X"),
+        (0xB55, "V"),
+        (0xB58, "X"),
+        (0xB5C, "M", "ଡ଼"),
+        (0xB5D, "M", "ଢ଼"),
+        (0xB5E, "X"),
+        (0xB5F, "V"),
+        (0xB64, "X"),
+        (0xB66, "V"),
+        (0xB78, "X"),
+        (0xB82, "V"),
+        (0xB84, "X"),
+        (0xB85, "V"),
+        (0xB8B, "X"),
+        (0xB8E, "V"),
+        (0xB91, "X"),
+        (0xB92, "V"),
+        (0xB96, "X"),
+        (0xB99, "V"),
+        (0xB9B, "X"),
+        (0xB9C, "V"),
+        (0xB9D, "X"),
+        (0xB9E, "V"),
+        (0xBA0, "X"),
+        (0xBA3, "V"),
+        (0xBA5, "X"),
+        (0xBA8, "V"),
+        (0xBAB, "X"),
+        (0xBAE, "V"),
+        (0xBBA, "X"),
+        (0xBBE, "V"),
+        (0xBC3, "X"),
+        (0xBC6, "V"),
+        (0xBC9, "X"),
+        (0xBCA, "V"),
+        (0xBCE, "X"),
+        (0xBD0, "V"),
+        (0xBD1, "X"),
+        (0xBD7, "V"),
+        (0xBD8, "X"),
+        (0xBE6, "V"),
+        (0xBFB, "X"),
+        (0xC00, "V"),
+        (0xC0D, "X"),
+        (0xC0E, "V"),
+        (0xC11, "X"),
+        (0xC12, "V"),
+        (0xC29, "X"),
+        (0xC2A, "V"),
+    ]
+
+
+def _seg_12() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xC3A, "X"),
+        (0xC3C, "V"),
+        (0xC45, "X"),
+        (0xC46, "V"),
+        (0xC49, "X"),
+        (0xC4A, "V"),
+        (0xC4E, "X"),
+        (0xC55, "V"),
+        (0xC57, "X"),
+        (0xC58, "V"),
+        (0xC5B, "X"),
+        (0xC5D, "V"),
+        (0xC5E, "X"),
+        (0xC60, "V"),
+        (0xC64, "X"),
+        (0xC66, "V"),
+        (0xC70, "X"),
+        (0xC77, "V"),
+        (0xC8D, "X"),
+        (0xC8E, "V"),
+        (0xC91, "X"),
+        (0xC92, "V"),
+        (0xCA9, "X"),
+        (0xCAA, "V"),
+        (0xCB4, "X"),
+        (0xCB5, "V"),
+        (0xCBA, "X"),
+        (0xCBC, "V"),
+        (0xCC5, "X"),
+        (0xCC6, "V"),
+        (0xCC9, "X"),
+        (0xCCA, "V"),
+        (0xCCE, "X"),
+        (0xCD5, "V"),
+        (0xCD7, "X"),
+        (0xCDD, "V"),
+        (0xCDF, "X"),
+        (0xCE0, "V"),
+        (0xCE4, "X"),
+        (0xCE6, "V"),
+        (0xCF0, "X"),
+        (0xCF1, "V"),
+        (0xCF4, "X"),
+        (0xD00, "V"),
+        (0xD0D, "X"),
+        (0xD0E, "V"),
+        (0xD11, "X"),
+        (0xD12, "V"),
+        (0xD45, "X"),
+        (0xD46, "V"),
+        (0xD49, "X"),
+        (0xD4A, "V"),
+        (0xD50, "X"),
+        (0xD54, "V"),
+        (0xD64, "X"),
+        (0xD66, "V"),
+        (0xD80, "X"),
+        (0xD81, "V"),
+        (0xD84, "X"),
+        (0xD85, "V"),
+        (0xD97, "X"),
+        (0xD9A, "V"),
+        (0xDB2, "X"),
+        (0xDB3, "V"),
+        (0xDBC, "X"),
+        (0xDBD, "V"),
+        (0xDBE, "X"),
+        (0xDC0, "V"),
+        (0xDC7, "X"),
+        (0xDCA, "V"),
+        (0xDCB, "X"),
+        (0xDCF, "V"),
+        (0xDD5, "X"),
+        (0xDD6, "V"),
+        (0xDD7, "X"),
+        (0xDD8, "V"),
+        (0xDE0, "X"),
+        (0xDE6, "V"),
+        (0xDF0, "X"),
+        (0xDF2, "V"),
+        (0xDF5, "X"),
+        (0xE01, "V"),
+        (0xE33, "M", "ํา"),
+        (0xE34, "V"),
+        (0xE3B, "X"),
+        (0xE3F, "V"),
+        (0xE5C, "X"),
+        (0xE81, "V"),
+        (0xE83, "X"),
+        (0xE84, "V"),
+        (0xE85, "X"),
+        (0xE86, "V"),
+        (0xE8B, "X"),
+        (0xE8C, "V"),
+        (0xEA4, "X"),
+        (0xEA5, "V"),
+        (0xEA6, "X"),
+        (0xEA7, "V"),
+        (0xEB3, "M", "ໍາ"),
+        (0xEB4, "V"),
+    ]
+
+
+def _seg_13() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xEBE, "X"),
+        (0xEC0, "V"),
+        (0xEC5, "X"),
+        (0xEC6, "V"),
+        (0xEC7, "X"),
+        (0xEC8, "V"),
+        (0xECF, "X"),
+        (0xED0, "V"),
+        (0xEDA, "X"),
+        (0xEDC, "M", "ຫນ"),
+        (0xEDD, "M", "ຫມ"),
+        (0xEDE, "V"),
+        (0xEE0, "X"),
+        (0xF00, "V"),
+        (0xF0C, "M", "་"),
+        (0xF0D, "V"),
+        (0xF43, "M", "གྷ"),
+        (0xF44, "V"),
+        (0xF48, "X"),
+        (0xF49, "V"),
+        (0xF4D, "M", "ཌྷ"),
+        (0xF4E, "V"),
+        (0xF52, "M", "དྷ"),
+        (0xF53, "V"),
+        (0xF57, "M", "བྷ"),
+        (0xF58, "V"),
+        (0xF5C, "M", "ཛྷ"),
+        (0xF5D, "V"),
+        (0xF69, "M", "ཀྵ"),
+        (0xF6A, "V"),
+        (0xF6D, "X"),
+        (0xF71, "V"),
+        (0xF73, "M", "ཱི"),
+        (0xF74, "V"),
+        (0xF75, "M", "ཱུ"),
+        (0xF76, "M", "ྲྀ"),
+        (0xF77, "M", "ྲཱྀ"),
+        (0xF78, "M", "ླྀ"),
+        (0xF79, "M", "ླཱྀ"),
+        (0xF7A, "V"),
+        (0xF81, "M", "ཱྀ"),
+        (0xF82, "V"),
+        (0xF93, "M", "ྒྷ"),
+        (0xF94, "V"),
+        (0xF98, "X"),
+        (0xF99, "V"),
+        (0xF9D, "M", "ྜྷ"),
+        (0xF9E, "V"),
+        (0xFA2, "M", "ྡྷ"),
+        (0xFA3, "V"),
+        (0xFA7, "M", "ྦྷ"),
+        (0xFA8, "V"),
+        (0xFAC, "M", "ྫྷ"),
+        (0xFAD, "V"),
+        (0xFB9, "M", "ྐྵ"),
+        (0xFBA, "V"),
+        (0xFBD, "X"),
+        (0xFBE, "V"),
+        (0xFCD, "X"),
+        (0xFCE, "V"),
+        (0xFDB, "X"),
+        (0x1000, "V"),
+        (0x10A0, "X"),
+        (0x10C7, "M", "ⴧ"),
+        (0x10C8, "X"),
+        (0x10CD, "M", "ⴭ"),
+        (0x10CE, "X"),
+        (0x10D0, "V"),
+        (0x10FC, "M", "ნ"),
+        (0x10FD, "V"),
+        (0x115F, "X"),
+        (0x1161, "V"),
+        (0x1249, "X"),
+        (0x124A, "V"),
+        (0x124E, "X"),
+        (0x1250, "V"),
+        (0x1257, "X"),
+        (0x1258, "V"),
+        (0x1259, "X"),
+        (0x125A, "V"),
+        (0x125E, "X"),
+        (0x1260, "V"),
+        (0x1289, "X"),
+        (0x128A, "V"),
+        (0x128E, "X"),
+        (0x1290, "V"),
+        (0x12B1, "X"),
+        (0x12B2, "V"),
+        (0x12B6, "X"),
+        (0x12B8, "V"),
+        (0x12BF, "X"),
+        (0x12C0, "V"),
+        (0x12C1, "X"),
+        (0x12C2, "V"),
+        (0x12C6, "X"),
+        (0x12C8, "V"),
+        (0x12D7, "X"),
+        (0x12D8, "V"),
+        (0x1311, "X"),
+        (0x1312, "V"),
+    ]
+
+
+def _seg_14() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1316, "X"),
+        (0x1318, "V"),
+        (0x135B, "X"),
+        (0x135D, "V"),
+        (0x137D, "X"),
+        (0x1380, "V"),
+        (0x139A, "X"),
+        (0x13A0, "V"),
+        (0x13F6, "X"),
+        (0x13F8, "M", "Ᏸ"),
+        (0x13F9, "M", "Ᏹ"),
+        (0x13FA, "M", "Ᏺ"),
+        (0x13FB, "M", "Ᏻ"),
+        (0x13FC, "M", "Ᏼ"),
+        (0x13FD, "M", "Ᏽ"),
+        (0x13FE, "X"),
+        (0x1400, "V"),
+        (0x1680, "X"),
+        (0x1681, "V"),
+        (0x169D, "X"),
+        (0x16A0, "V"),
+        (0x16F9, "X"),
+        (0x1700, "V"),
+        (0x1716, "X"),
+        (0x171F, "V"),
+        (0x1737, "X"),
+        (0x1740, "V"),
+        (0x1754, "X"),
+        (0x1760, "V"),
+        (0x176D, "X"),
+        (0x176E, "V"),
+        (0x1771, "X"),
+        (0x1772, "V"),
+        (0x1774, "X"),
+        (0x1780, "V"),
+        (0x17B4, "X"),
+        (0x17B6, "V"),
+        (0x17DE, "X"),
+        (0x17E0, "V"),
+        (0x17EA, "X"),
+        (0x17F0, "V"),
+        (0x17FA, "X"),
+        (0x1800, "V"),
+        (0x1806, "X"),
+        (0x1807, "V"),
+        (0x180B, "I"),
+        (0x180E, "X"),
+        (0x180F, "I"),
+        (0x1810, "V"),
+        (0x181A, "X"),
+        (0x1820, "V"),
+        (0x1879, "X"),
+        (0x1880, "V"),
+        (0x18AB, "X"),
+        (0x18B0, "V"),
+        (0x18F6, "X"),
+        (0x1900, "V"),
+        (0x191F, "X"),
+        (0x1920, "V"),
+        (0x192C, "X"),
+        (0x1930, "V"),
+        (0x193C, "X"),
+        (0x1940, "V"),
+        (0x1941, "X"),
+        (0x1944, "V"),
+        (0x196E, "X"),
+        (0x1970, "V"),
+        (0x1975, "X"),
+        (0x1980, "V"),
+        (0x19AC, "X"),
+        (0x19B0, "V"),
+        (0x19CA, "X"),
+        (0x19D0, "V"),
+        (0x19DB, "X"),
+        (0x19DE, "V"),
+        (0x1A1C, "X"),
+        (0x1A1E, "V"),
+        (0x1A5F, "X"),
+        (0x1A60, "V"),
+        (0x1A7D, "X"),
+        (0x1A7F, "V"),
+        (0x1A8A, "X"),
+        (0x1A90, "V"),
+        (0x1A9A, "X"),
+        (0x1AA0, "V"),
+        (0x1AAE, "X"),
+        (0x1AB0, "V"),
+        (0x1ACF, "X"),
+        (0x1B00, "V"),
+        (0x1B4D, "X"),
+        (0x1B50, "V"),
+        (0x1B7F, "X"),
+        (0x1B80, "V"),
+        (0x1BF4, "X"),
+        (0x1BFC, "V"),
+        (0x1C38, "X"),
+        (0x1C3B, "V"),
+        (0x1C4A, "X"),
+        (0x1C4D, "V"),
+        (0x1C80, "M", "в"),
+    ]
+
+
+def _seg_15() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1C81, "M", "д"),
+        (0x1C82, "M", "о"),
+        (0x1C83, "M", "с"),
+        (0x1C84, "M", "т"),
+        (0x1C86, "M", "ъ"),
+        (0x1C87, "M", "ѣ"),
+        (0x1C88, "M", "ꙋ"),
+        (0x1C89, "X"),
+        (0x1C90, "M", "ა"),
+        (0x1C91, "M", "ბ"),
+        (0x1C92, "M", "გ"),
+        (0x1C93, "M", "დ"),
+        (0x1C94, "M", "ე"),
+        (0x1C95, "M", "ვ"),
+        (0x1C96, "M", "ზ"),
+        (0x1C97, "M", "თ"),
+        (0x1C98, "M", "ი"),
+        (0x1C99, "M", "კ"),
+        (0x1C9A, "M", "ლ"),
+        (0x1C9B, "M", "მ"),
+        (0x1C9C, "M", "ნ"),
+        (0x1C9D, "M", "ო"),
+        (0x1C9E, "M", "პ"),
+        (0x1C9F, "M", "ჟ"),
+        (0x1CA0, "M", "რ"),
+        (0x1CA1, "M", "ს"),
+        (0x1CA2, "M", "ტ"),
+        (0x1CA3, "M", "უ"),
+        (0x1CA4, "M", "ფ"),
+        (0x1CA5, "M", "ქ"),
+        (0x1CA6, "M", "ღ"),
+        (0x1CA7, "M", "ყ"),
+        (0x1CA8, "M", "შ"),
+        (0x1CA9, "M", "ჩ"),
+        (0x1CAA, "M", "ც"),
+        (0x1CAB, "M", "ძ"),
+        (0x1CAC, "M", "წ"),
+        (0x1CAD, "M", "ჭ"),
+        (0x1CAE, "M", "ხ"),
+        (0x1CAF, "M", "ჯ"),
+        (0x1CB0, "M", "ჰ"),
+        (0x1CB1, "M", "ჱ"),
+        (0x1CB2, "M", "ჲ"),
+        (0x1CB3, "M", "ჳ"),
+        (0x1CB4, "M", "ჴ"),
+        (0x1CB5, "M", "ჵ"),
+        (0x1CB6, "M", "ჶ"),
+        (0x1CB7, "M", "ჷ"),
+        (0x1CB8, "M", "ჸ"),
+        (0x1CB9, "M", "ჹ"),
+        (0x1CBA, "M", "ჺ"),
+        (0x1CBB, "X"),
+        (0x1CBD, "M", "ჽ"),
+        (0x1CBE, "M", "ჾ"),
+        (0x1CBF, "M", "ჿ"),
+        (0x1CC0, "V"),
+        (0x1CC8, "X"),
+        (0x1CD0, "V"),
+        (0x1CFB, "X"),
+        (0x1D00, "V"),
+        (0x1D2C, "M", "a"),
+        (0x1D2D, "M", "æ"),
+        (0x1D2E, "M", "b"),
+        (0x1D2F, "V"),
+        (0x1D30, "M", "d"),
+        (0x1D31, "M", "e"),
+        (0x1D32, "M", "ǝ"),
+        (0x1D33, "M", "g"),
+        (0x1D34, "M", "h"),
+        (0x1D35, "M", "i"),
+        (0x1D36, "M", "j"),
+        (0x1D37, "M", "k"),
+        (0x1D38, "M", "l"),
+        (0x1D39, "M", "m"),
+        (0x1D3A, "M", "n"),
+        (0x1D3B, "V"),
+        (0x1D3C, "M", "o"),
+        (0x1D3D, "M", "ȣ"),
+        (0x1D3E, "M", "p"),
+        (0x1D3F, "M", "r"),
+        (0x1D40, "M", "t"),
+        (0x1D41, "M", "u"),
+        (0x1D42, "M", "w"),
+        (0x1D43, "M", "a"),
+        (0x1D44, "M", "ɐ"),
+        (0x1D45, "M", "ɑ"),
+        (0x1D46, "M", "ᴂ"),
+        (0x1D47, "M", "b"),
+        (0x1D48, "M", "d"),
+        (0x1D49, "M", "e"),
+        (0x1D4A, "M", "ə"),
+        (0x1D4B, "M", "ɛ"),
+        (0x1D4C, "M", "ɜ"),
+        (0x1D4D, "M", "g"),
+        (0x1D4E, "V"),
+        (0x1D4F, "M", "k"),
+        (0x1D50, "M", "m"),
+        (0x1D51, "M", "ŋ"),
+        (0x1D52, "M", "o"),
+        (0x1D53, "M", "ɔ"),
+    ]
+
+
+def _seg_16() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D54, "M", "ᴖ"),
+        (0x1D55, "M", "ᴗ"),
+        (0x1D56, "M", "p"),
+        (0x1D57, "M", "t"),
+        (0x1D58, "M", "u"),
+        (0x1D59, "M", "ᴝ"),
+        (0x1D5A, "M", "ɯ"),
+        (0x1D5B, "M", "v"),
+        (0x1D5C, "M", "ᴥ"),
+        (0x1D5D, "M", "β"),
+        (0x1D5E, "M", "γ"),
+        (0x1D5F, "M", "δ"),
+        (0x1D60, "M", "φ"),
+        (0x1D61, "M", "χ"),
+        (0x1D62, "M", "i"),
+        (0x1D63, "M", "r"),
+        (0x1D64, "M", "u"),
+        (0x1D65, "M", "v"),
+        (0x1D66, "M", "β"),
+        (0x1D67, "M", "γ"),
+        (0x1D68, "M", "ρ"),
+        (0x1D69, "M", "φ"),
+        (0x1D6A, "M", "χ"),
+        (0x1D6B, "V"),
+        (0x1D78, "M", "н"),
+        (0x1D79, "V"),
+        (0x1D9B, "M", "ɒ"),
+        (0x1D9C, "M", "c"),
+        (0x1D9D, "M", "ɕ"),
+        (0x1D9E, "M", "ð"),
+        (0x1D9F, "M", "ɜ"),
+        (0x1DA0, "M", "f"),
+        (0x1DA1, "M", "ɟ"),
+        (0x1DA2, "M", "ɡ"),
+        (0x1DA3, "M", "ɥ"),
+        (0x1DA4, "M", "ɨ"),
+        (0x1DA5, "M", "ɩ"),
+        (0x1DA6, "M", "ɪ"),
+        (0x1DA7, "M", "ᵻ"),
+        (0x1DA8, "M", "ʝ"),
+        (0x1DA9, "M", "ɭ"),
+        (0x1DAA, "M", "ᶅ"),
+        (0x1DAB, "M", "ʟ"),
+        (0x1DAC, "M", "ɱ"),
+        (0x1DAD, "M", "ɰ"),
+        (0x1DAE, "M", "ɲ"),
+        (0x1DAF, "M", "ɳ"),
+        (0x1DB0, "M", "ɴ"),
+        (0x1DB1, "M", "ɵ"),
+        (0x1DB2, "M", "ɸ"),
+        (0x1DB3, "M", "ʂ"),
+        (0x1DB4, "M", "ʃ"),
+        (0x1DB5, "M", "ƫ"),
+        (0x1DB6, "M", "ʉ"),
+        (0x1DB7, "M", "ʊ"),
+        (0x1DB8, "M", "ᴜ"),
+        (0x1DB9, "M", "ʋ"),
+        (0x1DBA, "M", "ʌ"),
+        (0x1DBB, "M", "z"),
+        (0x1DBC, "M", "ʐ"),
+        (0x1DBD, "M", "ʑ"),
+        (0x1DBE, "M", "ʒ"),
+        (0x1DBF, "M", "θ"),
+        (0x1DC0, "V"),
+        (0x1E00, "M", "ḁ"),
+        (0x1E01, "V"),
+        (0x1E02, "M", "ḃ"),
+        (0x1E03, "V"),
+        (0x1E04, "M", "ḅ"),
+        (0x1E05, "V"),
+        (0x1E06, "M", "ḇ"),
+        (0x1E07, "V"),
+        (0x1E08, "M", "ḉ"),
+        (0x1E09, "V"),
+        (0x1E0A, "M", "ḋ"),
+        (0x1E0B, "V"),
+        (0x1E0C, "M", "ḍ"),
+        (0x1E0D, "V"),
+        (0x1E0E, "M", "ḏ"),
+        (0x1E0F, "V"),
+        (0x1E10, "M", "ḑ"),
+        (0x1E11, "V"),
+        (0x1E12, "M", "ḓ"),
+        (0x1E13, "V"),
+        (0x1E14, "M", "ḕ"),
+        (0x1E15, "V"),
+        (0x1E16, "M", "ḗ"),
+        (0x1E17, "V"),
+        (0x1E18, "M", "ḙ"),
+        (0x1E19, "V"),
+        (0x1E1A, "M", "ḛ"),
+        (0x1E1B, "V"),
+        (0x1E1C, "M", "ḝ"),
+        (0x1E1D, "V"),
+        (0x1E1E, "M", "ḟ"),
+        (0x1E1F, "V"),
+        (0x1E20, "M", "ḡ"),
+        (0x1E21, "V"),
+        (0x1E22, "M", "ḣ"),
+        (0x1E23, "V"),
+    ]
+
+
+def _seg_17() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E24, "M", "ḥ"),
+        (0x1E25, "V"),
+        (0x1E26, "M", "ḧ"),
+        (0x1E27, "V"),
+        (0x1E28, "M", "ḩ"),
+        (0x1E29, "V"),
+        (0x1E2A, "M", "ḫ"),
+        (0x1E2B, "V"),
+        (0x1E2C, "M", "ḭ"),
+        (0x1E2D, "V"),
+        (0x1E2E, "M", "ḯ"),
+        (0x1E2F, "V"),
+        (0x1E30, "M", "ḱ"),
+        (0x1E31, "V"),
+        (0x1E32, "M", "ḳ"),
+        (0x1E33, "V"),
+        (0x1E34, "M", "ḵ"),
+        (0x1E35, "V"),
+        (0x1E36, "M", "ḷ"),
+        (0x1E37, "V"),
+        (0x1E38, "M", "ḹ"),
+        (0x1E39, "V"),
+        (0x1E3A, "M", "ḻ"),
+        (0x1E3B, "V"),
+        (0x1E3C, "M", "ḽ"),
+        (0x1E3D, "V"),
+        (0x1E3E, "M", "ḿ"),
+        (0x1E3F, "V"),
+        (0x1E40, "M", "ṁ"),
+        (0x1E41, "V"),
+        (0x1E42, "M", "ṃ"),
+        (0x1E43, "V"),
+        (0x1E44, "M", "ṅ"),
+        (0x1E45, "V"),
+        (0x1E46, "M", "ṇ"),
+        (0x1E47, "V"),
+        (0x1E48, "M", "ṉ"),
+        (0x1E49, "V"),
+        (0x1E4A, "M", "ṋ"),
+        (0x1E4B, "V"),
+        (0x1E4C, "M", "ṍ"),
+        (0x1E4D, "V"),
+        (0x1E4E, "M", "ṏ"),
+        (0x1E4F, "V"),
+        (0x1E50, "M", "ṑ"),
+        (0x1E51, "V"),
+        (0x1E52, "M", "ṓ"),
+        (0x1E53, "V"),
+        (0x1E54, "M", "ṕ"),
+        (0x1E55, "V"),
+        (0x1E56, "M", "ṗ"),
+        (0x1E57, "V"),
+        (0x1E58, "M", "ṙ"),
+        (0x1E59, "V"),
+        (0x1E5A, "M", "ṛ"),
+        (0x1E5B, "V"),
+        (0x1E5C, "M", "ṝ"),
+        (0x1E5D, "V"),
+        (0x1E5E, "M", "ṟ"),
+        (0x1E5F, "V"),
+        (0x1E60, "M", "ṡ"),
+        (0x1E61, "V"),
+        (0x1E62, "M", "ṣ"),
+        (0x1E63, "V"),
+        (0x1E64, "M", "ṥ"),
+        (0x1E65, "V"),
+        (0x1E66, "M", "ṧ"),
+        (0x1E67, "V"),
+        (0x1E68, "M", "ṩ"),
+        (0x1E69, "V"),
+        (0x1E6A, "M", "ṫ"),
+        (0x1E6B, "V"),
+        (0x1E6C, "M", "ṭ"),
+        (0x1E6D, "V"),
+        (0x1E6E, "M", "ṯ"),
+        (0x1E6F, "V"),
+        (0x1E70, "M", "ṱ"),
+        (0x1E71, "V"),
+        (0x1E72, "M", "ṳ"),
+        (0x1E73, "V"),
+        (0x1E74, "M", "ṵ"),
+        (0x1E75, "V"),
+        (0x1E76, "M", "ṷ"),
+        (0x1E77, "V"),
+        (0x1E78, "M", "ṹ"),
+        (0x1E79, "V"),
+        (0x1E7A, "M", "ṻ"),
+        (0x1E7B, "V"),
+        (0x1E7C, "M", "ṽ"),
+        (0x1E7D, "V"),
+        (0x1E7E, "M", "ṿ"),
+        (0x1E7F, "V"),
+        (0x1E80, "M", "ẁ"),
+        (0x1E81, "V"),
+        (0x1E82, "M", "ẃ"),
+        (0x1E83, "V"),
+        (0x1E84, "M", "ẅ"),
+        (0x1E85, "V"),
+        (0x1E86, "M", "ẇ"),
+        (0x1E87, "V"),
+    ]
+
+
+def _seg_18() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E88, "M", "ẉ"),
+        (0x1E89, "V"),
+        (0x1E8A, "M", "ẋ"),
+        (0x1E8B, "V"),
+        (0x1E8C, "M", "ẍ"),
+        (0x1E8D, "V"),
+        (0x1E8E, "M", "ẏ"),
+        (0x1E8F, "V"),
+        (0x1E90, "M", "ẑ"),
+        (0x1E91, "V"),
+        (0x1E92, "M", "ẓ"),
+        (0x1E93, "V"),
+        (0x1E94, "M", "ẕ"),
+        (0x1E95, "V"),
+        (0x1E9A, "M", "aʾ"),
+        (0x1E9B, "M", "ṡ"),
+        (0x1E9C, "V"),
+        (0x1E9E, "M", "ß"),
+        (0x1E9F, "V"),
+        (0x1EA0, "M", "ạ"),
+        (0x1EA1, "V"),
+        (0x1EA2, "M", "ả"),
+        (0x1EA3, "V"),
+        (0x1EA4, "M", "ấ"),
+        (0x1EA5, "V"),
+        (0x1EA6, "M", "ầ"),
+        (0x1EA7, "V"),
+        (0x1EA8, "M", "ẩ"),
+        (0x1EA9, "V"),
+        (0x1EAA, "M", "ẫ"),
+        (0x1EAB, "V"),
+        (0x1EAC, "M", "ậ"),
+        (0x1EAD, "V"),
+        (0x1EAE, "M", "ắ"),
+        (0x1EAF, "V"),
+        (0x1EB0, "M", "ằ"),
+        (0x1EB1, "V"),
+        (0x1EB2, "M", "ẳ"),
+        (0x1EB3, "V"),
+        (0x1EB4, "M", "ẵ"),
+        (0x1EB5, "V"),
+        (0x1EB6, "M", "ặ"),
+        (0x1EB7, "V"),
+        (0x1EB8, "M", "ẹ"),
+        (0x1EB9, "V"),
+        (0x1EBA, "M", "ẻ"),
+        (0x1EBB, "V"),
+        (0x1EBC, "M", "ẽ"),
+        (0x1EBD, "V"),
+        (0x1EBE, "M", "ế"),
+        (0x1EBF, "V"),
+        (0x1EC0, "M", "ề"),
+        (0x1EC1, "V"),
+        (0x1EC2, "M", "ể"),
+        (0x1EC3, "V"),
+        (0x1EC4, "M", "ễ"),
+        (0x1EC5, "V"),
+        (0x1EC6, "M", "ệ"),
+        (0x1EC7, "V"),
+        (0x1EC8, "M", "ỉ"),
+        (0x1EC9, "V"),
+        (0x1ECA, "M", "ị"),
+        (0x1ECB, "V"),
+        (0x1ECC, "M", "ọ"),
+        (0x1ECD, "V"),
+        (0x1ECE, "M", "ỏ"),
+        (0x1ECF, "V"),
+        (0x1ED0, "M", "ố"),
+        (0x1ED1, "V"),
+        (0x1ED2, "M", "ồ"),
+        (0x1ED3, "V"),
+        (0x1ED4, "M", "ổ"),
+        (0x1ED5, "V"),
+        (0x1ED6, "M", "ỗ"),
+        (0x1ED7, "V"),
+        (0x1ED8, "M", "ộ"),
+        (0x1ED9, "V"),
+        (0x1EDA, "M", "ớ"),
+        (0x1EDB, "V"),
+        (0x1EDC, "M", "ờ"),
+        (0x1EDD, "V"),
+        (0x1EDE, "M", "ở"),
+        (0x1EDF, "V"),
+        (0x1EE0, "M", "ỡ"),
+        (0x1EE1, "V"),
+        (0x1EE2, "M", "ợ"),
+        (0x1EE3, "V"),
+        (0x1EE4, "M", "ụ"),
+        (0x1EE5, "V"),
+        (0x1EE6, "M", "ủ"),
+        (0x1EE7, "V"),
+        (0x1EE8, "M", "ứ"),
+        (0x1EE9, "V"),
+        (0x1EEA, "M", "ừ"),
+        (0x1EEB, "V"),
+        (0x1EEC, "M", "ử"),
+        (0x1EED, "V"),
+        (0x1EEE, "M", "ữ"),
+        (0x1EEF, "V"),
+        (0x1EF0, "M", "ự"),
+    ]
+
+
+def _seg_19() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EF1, "V"),
+        (0x1EF2, "M", "ỳ"),
+        (0x1EF3, "V"),
+        (0x1EF4, "M", "ỵ"),
+        (0x1EF5, "V"),
+        (0x1EF6, "M", "ỷ"),
+        (0x1EF7, "V"),
+        (0x1EF8, "M", "ỹ"),
+        (0x1EF9, "V"),
+        (0x1EFA, "M", "ỻ"),
+        (0x1EFB, "V"),
+        (0x1EFC, "M", "ỽ"),
+        (0x1EFD, "V"),
+        (0x1EFE, "M", "ỿ"),
+        (0x1EFF, "V"),
+        (0x1F08, "M", "ἀ"),
+        (0x1F09, "M", "ἁ"),
+        (0x1F0A, "M", "ἂ"),
+        (0x1F0B, "M", "ἃ"),
+        (0x1F0C, "M", "ἄ"),
+        (0x1F0D, "M", "ἅ"),
+        (0x1F0E, "M", "ἆ"),
+        (0x1F0F, "M", "ἇ"),
+        (0x1F10, "V"),
+        (0x1F16, "X"),
+        (0x1F18, "M", "ἐ"),
+        (0x1F19, "M", "ἑ"),
+        (0x1F1A, "M", "ἒ"),
+        (0x1F1B, "M", "ἓ"),
+        (0x1F1C, "M", "ἔ"),
+        (0x1F1D, "M", "ἕ"),
+        (0x1F1E, "X"),
+        (0x1F20, "V"),
+        (0x1F28, "M", "ἠ"),
+        (0x1F29, "M", "ἡ"),
+        (0x1F2A, "M", "ἢ"),
+        (0x1F2B, "M", "ἣ"),
+        (0x1F2C, "M", "ἤ"),
+        (0x1F2D, "M", "ἥ"),
+        (0x1F2E, "M", "ἦ"),
+        (0x1F2F, "M", "ἧ"),
+        (0x1F30, "V"),
+        (0x1F38, "M", "ἰ"),
+        (0x1F39, "M", "ἱ"),
+        (0x1F3A, "M", "ἲ"),
+        (0x1F3B, "M", "ἳ"),
+        (0x1F3C, "M", "ἴ"),
+        (0x1F3D, "M", "ἵ"),
+        (0x1F3E, "M", "ἶ"),
+        (0x1F3F, "M", "ἷ"),
+        (0x1F40, "V"),
+        (0x1F46, "X"),
+        (0x1F48, "M", "ὀ"),
+        (0x1F49, "M", "ὁ"),
+        (0x1F4A, "M", "ὂ"),
+        (0x1F4B, "M", "ὃ"),
+        (0x1F4C, "M", "ὄ"),
+        (0x1F4D, "M", "ὅ"),
+        (0x1F4E, "X"),
+        (0x1F50, "V"),
+        (0x1F58, "X"),
+        (0x1F59, "M", "ὑ"),
+        (0x1F5A, "X"),
+        (0x1F5B, "M", "ὓ"),
+        (0x1F5C, "X"),
+        (0x1F5D, "M", "ὕ"),
+        (0x1F5E, "X"),
+        (0x1F5F, "M", "ὗ"),
+        (0x1F60, "V"),
+        (0x1F68, "M", "ὠ"),
+        (0x1F69, "M", "ὡ"),
+        (0x1F6A, "M", "ὢ"),
+        (0x1F6B, "M", "ὣ"),
+        (0x1F6C, "M", "ὤ"),
+        (0x1F6D, "M", "ὥ"),
+        (0x1F6E, "M", "ὦ"),
+        (0x1F6F, "M", "ὧ"),
+        (0x1F70, "V"),
+        (0x1F71, "M", "ά"),
+        (0x1F72, "V"),
+        (0x1F73, "M", "έ"),
+        (0x1F74, "V"),
+        (0x1F75, "M", "ή"),
+        (0x1F76, "V"),
+        (0x1F77, "M", "ί"),
+        (0x1F78, "V"),
+        (0x1F79, "M", "ό"),
+        (0x1F7A, "V"),
+        (0x1F7B, "M", "ύ"),
+        (0x1F7C, "V"),
+        (0x1F7D, "M", "ώ"),
+        (0x1F7E, "X"),
+        (0x1F80, "M", "ἀι"),
+        (0x1F81, "M", "ἁι"),
+        (0x1F82, "M", "ἂι"),
+        (0x1F83, "M", "ἃι"),
+        (0x1F84, "M", "ἄι"),
+        (0x1F85, "M", "ἅι"),
+        (0x1F86, "M", "ἆι"),
+        (0x1F87, "M", "ἇι"),
+    ]
+
+
+def _seg_20() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1F88, "M", "ἀι"),
+        (0x1F89, "M", "ἁι"),
+        (0x1F8A, "M", "ἂι"),
+        (0x1F8B, "M", "ἃι"),
+        (0x1F8C, "M", "ἄι"),
+        (0x1F8D, "M", "ἅι"),
+        (0x1F8E, "M", "ἆι"),
+        (0x1F8F, "M", "ἇι"),
+        (0x1F90, "M", "ἠι"),
+        (0x1F91, "M", "ἡι"),
+        (0x1F92, "M", "ἢι"),
+        (0x1F93, "M", "ἣι"),
+        (0x1F94, "M", "ἤι"),
+        (0x1F95, "M", "ἥι"),
+        (0x1F96, "M", "ἦι"),
+        (0x1F97, "M", "ἧι"),
+        (0x1F98, "M", "ἠι"),
+        (0x1F99, "M", "ἡι"),
+        (0x1F9A, "M", "ἢι"),
+        (0x1F9B, "M", "ἣι"),
+        (0x1F9C, "M", "ἤι"),
+        (0x1F9D, "M", "ἥι"),
+        (0x1F9E, "M", "ἦι"),
+        (0x1F9F, "M", "ἧι"),
+        (0x1FA0, "M", "ὠι"),
+        (0x1FA1, "M", "ὡι"),
+        (0x1FA2, "M", "ὢι"),
+        (0x1FA3, "M", "ὣι"),
+        (0x1FA4, "M", "ὤι"),
+        (0x1FA5, "M", "ὥι"),
+        (0x1FA6, "M", "ὦι"),
+        (0x1FA7, "M", "ὧι"),
+        (0x1FA8, "M", "ὠι"),
+        (0x1FA9, "M", "ὡι"),
+        (0x1FAA, "M", "ὢι"),
+        (0x1FAB, "M", "ὣι"),
+        (0x1FAC, "M", "ὤι"),
+        (0x1FAD, "M", "ὥι"),
+        (0x1FAE, "M", "ὦι"),
+        (0x1FAF, "M", "ὧι"),
+        (0x1FB0, "V"),
+        (0x1FB2, "M", "ὰι"),
+        (0x1FB3, "M", "αι"),
+        (0x1FB4, "M", "άι"),
+        (0x1FB5, "X"),
+        (0x1FB6, "V"),
+        (0x1FB7, "M", "ᾶι"),
+        (0x1FB8, "M", "ᾰ"),
+        (0x1FB9, "M", "ᾱ"),
+        (0x1FBA, "M", "ὰ"),
+        (0x1FBB, "M", "ά"),
+        (0x1FBC, "M", "αι"),
+        (0x1FBD, "3", " ̓"),
+        (0x1FBE, "M", "ι"),
+        (0x1FBF, "3", " ̓"),
+        (0x1FC0, "3", " ͂"),
+        (0x1FC1, "3", " ̈͂"),
+        (0x1FC2, "M", "ὴι"),
+        (0x1FC3, "M", "ηι"),
+        (0x1FC4, "M", "ήι"),
+        (0x1FC5, "X"),
+        (0x1FC6, "V"),
+        (0x1FC7, "M", "ῆι"),
+        (0x1FC8, "M", "ὲ"),
+        (0x1FC9, "M", "έ"),
+        (0x1FCA, "M", "ὴ"),
+        (0x1FCB, "M", "ή"),
+        (0x1FCC, "M", "ηι"),
+        (0x1FCD, "3", " ̓̀"),
+        (0x1FCE, "3", " ̓́"),
+        (0x1FCF, "3", " ̓͂"),
+        (0x1FD0, "V"),
+        (0x1FD3, "M", "ΐ"),
+        (0x1FD4, "X"),
+        (0x1FD6, "V"),
+        (0x1FD8, "M", "ῐ"),
+        (0x1FD9, "M", "ῑ"),
+        (0x1FDA, "M", "ὶ"),
+        (0x1FDB, "M", "ί"),
+        (0x1FDC, "X"),
+        (0x1FDD, "3", " ̔̀"),
+        (0x1FDE, "3", " ̔́"),
+        (0x1FDF, "3", " ̔͂"),
+        (0x1FE0, "V"),
+        (0x1FE3, "M", "ΰ"),
+        (0x1FE4, "V"),
+        (0x1FE8, "M", "ῠ"),
+        (0x1FE9, "M", "ῡ"),
+        (0x1FEA, "M", "ὺ"),
+        (0x1FEB, "M", "ύ"),
+        (0x1FEC, "M", "ῥ"),
+        (0x1FED, "3", " ̈̀"),
+        (0x1FEE, "3", " ̈́"),
+        (0x1FEF, "3", "`"),
+        (0x1FF0, "X"),
+        (0x1FF2, "M", "ὼι"),
+        (0x1FF3, "M", "ωι"),
+        (0x1FF4, "M", "ώι"),
+        (0x1FF5, "X"),
+        (0x1FF6, "V"),
+    ]
+
+
+def _seg_21() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1FF7, "M", "ῶι"),
+        (0x1FF8, "M", "ὸ"),
+        (0x1FF9, "M", "ό"),
+        (0x1FFA, "M", "ὼ"),
+        (0x1FFB, "M", "ώ"),
+        (0x1FFC, "M", "ωι"),
+        (0x1FFD, "3", " ́"),
+        (0x1FFE, "3", " ̔"),
+        (0x1FFF, "X"),
+        (0x2000, "3", " "),
+        (0x200B, "I"),
+        (0x200C, "D", ""),
+        (0x200E, "X"),
+        (0x2010, "V"),
+        (0x2011, "M", "‐"),
+        (0x2012, "V"),
+        (0x2017, "3", " ̳"),
+        (0x2018, "V"),
+        (0x2024, "X"),
+        (0x2027, "V"),
+        (0x2028, "X"),
+        (0x202F, "3", " "),
+        (0x2030, "V"),
+        (0x2033, "M", "′′"),
+        (0x2034, "M", "′′′"),
+        (0x2035, "V"),
+        (0x2036, "M", "‵‵"),
+        (0x2037, "M", "‵‵‵"),
+        (0x2038, "V"),
+        (0x203C, "3", "!!"),
+        (0x203D, "V"),
+        (0x203E, "3", " ̅"),
+        (0x203F, "V"),
+        (0x2047, "3", "??"),
+        (0x2048, "3", "?!"),
+        (0x2049, "3", "!?"),
+        (0x204A, "V"),
+        (0x2057, "M", "′′′′"),
+        (0x2058, "V"),
+        (0x205F, "3", " "),
+        (0x2060, "I"),
+        (0x2061, "X"),
+        (0x2064, "I"),
+        (0x2065, "X"),
+        (0x2070, "M", "0"),
+        (0x2071, "M", "i"),
+        (0x2072, "X"),
+        (0x2074, "M", "4"),
+        (0x2075, "M", "5"),
+        (0x2076, "M", "6"),
+        (0x2077, "M", "7"),
+        (0x2078, "M", "8"),
+        (0x2079, "M", "9"),
+        (0x207A, "3", "+"),
+        (0x207B, "M", "−"),
+        (0x207C, "3", "="),
+        (0x207D, "3", "("),
+        (0x207E, "3", ")"),
+        (0x207F, "M", "n"),
+        (0x2080, "M", "0"),
+        (0x2081, "M", "1"),
+        (0x2082, "M", "2"),
+        (0x2083, "M", "3"),
+        (0x2084, "M", "4"),
+        (0x2085, "M", "5"),
+        (0x2086, "M", "6"),
+        (0x2087, "M", "7"),
+        (0x2088, "M", "8"),
+        (0x2089, "M", "9"),
+        (0x208A, "3", "+"),
+        (0x208B, "M", "−"),
+        (0x208C, "3", "="),
+        (0x208D, "3", "("),
+        (0x208E, "3", ")"),
+        (0x208F, "X"),
+        (0x2090, "M", "a"),
+        (0x2091, "M", "e"),
+        (0x2092, "M", "o"),
+        (0x2093, "M", "x"),
+        (0x2094, "M", "ə"),
+        (0x2095, "M", "h"),
+        (0x2096, "M", "k"),
+        (0x2097, "M", "l"),
+        (0x2098, "M", "m"),
+        (0x2099, "M", "n"),
+        (0x209A, "M", "p"),
+        (0x209B, "M", "s"),
+        (0x209C, "M", "t"),
+        (0x209D, "X"),
+        (0x20A0, "V"),
+        (0x20A8, "M", "rs"),
+        (0x20A9, "V"),
+        (0x20C1, "X"),
+        (0x20D0, "V"),
+        (0x20F1, "X"),
+        (0x2100, "3", "a/c"),
+        (0x2101, "3", "a/s"),
+        (0x2102, "M", "c"),
+        (0x2103, "M", "°c"),
+        (0x2104, "V"),
+    ]
+
+
+def _seg_22() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2105, "3", "c/o"),
+        (0x2106, "3", "c/u"),
+        (0x2107, "M", "ɛ"),
+        (0x2108, "V"),
+        (0x2109, "M", "°f"),
+        (0x210A, "M", "g"),
+        (0x210B, "M", "h"),
+        (0x210F, "M", "ħ"),
+        (0x2110, "M", "i"),
+        (0x2112, "M", "l"),
+        (0x2114, "V"),
+        (0x2115, "M", "n"),
+        (0x2116, "M", "no"),
+        (0x2117, "V"),
+        (0x2119, "M", "p"),
+        (0x211A, "M", "q"),
+        (0x211B, "M", "r"),
+        (0x211E, "V"),
+        (0x2120, "M", "sm"),
+        (0x2121, "M", "tel"),
+        (0x2122, "M", "tm"),
+        (0x2123, "V"),
+        (0x2124, "M", "z"),
+        (0x2125, "V"),
+        (0x2126, "M", "ω"),
+        (0x2127, "V"),
+        (0x2128, "M", "z"),
+        (0x2129, "V"),
+        (0x212A, "M", "k"),
+        (0x212B, "M", "å"),
+        (0x212C, "M", "b"),
+        (0x212D, "M", "c"),
+        (0x212E, "V"),
+        (0x212F, "M", "e"),
+        (0x2131, "M", "f"),
+        (0x2132, "X"),
+        (0x2133, "M", "m"),
+        (0x2134, "M", "o"),
+        (0x2135, "M", "א"),
+        (0x2136, "M", "ב"),
+        (0x2137, "M", "ג"),
+        (0x2138, "M", "ד"),
+        (0x2139, "M", "i"),
+        (0x213A, "V"),
+        (0x213B, "M", "fax"),
+        (0x213C, "M", "π"),
+        (0x213D, "M", "γ"),
+        (0x213F, "M", "π"),
+        (0x2140, "M", "∑"),
+        (0x2141, "V"),
+        (0x2145, "M", "d"),
+        (0x2147, "M", "e"),
+        (0x2148, "M", "i"),
+        (0x2149, "M", "j"),
+        (0x214A, "V"),
+        (0x2150, "M", "1⁄7"),
+        (0x2151, "M", "1⁄9"),
+        (0x2152, "M", "1⁄10"),
+        (0x2153, "M", "1⁄3"),
+        (0x2154, "M", "2⁄3"),
+        (0x2155, "M", "1⁄5"),
+        (0x2156, "M", "2⁄5"),
+        (0x2157, "M", "3⁄5"),
+        (0x2158, "M", "4⁄5"),
+        (0x2159, "M", "1⁄6"),
+        (0x215A, "M", "5⁄6"),
+        (0x215B, "M", "1⁄8"),
+        (0x215C, "M", "3⁄8"),
+        (0x215D, "M", "5⁄8"),
+        (0x215E, "M", "7⁄8"),
+        (0x215F, "M", "1⁄"),
+        (0x2160, "M", "i"),
+        (0x2161, "M", "ii"),
+        (0x2162, "M", "iii"),
+        (0x2163, "M", "iv"),
+        (0x2164, "M", "v"),
+        (0x2165, "M", "vi"),
+        (0x2166, "M", "vii"),
+        (0x2167, "M", "viii"),
+        (0x2168, "M", "ix"),
+        (0x2169, "M", "x"),
+        (0x216A, "M", "xi"),
+        (0x216B, "M", "xii"),
+        (0x216C, "M", "l"),
+        (0x216D, "M", "c"),
+        (0x216E, "M", "d"),
+        (0x216F, "M", "m"),
+        (0x2170, "M", "i"),
+        (0x2171, "M", "ii"),
+        (0x2172, "M", "iii"),
+        (0x2173, "M", "iv"),
+        (0x2174, "M", "v"),
+        (0x2175, "M", "vi"),
+        (0x2176, "M", "vii"),
+        (0x2177, "M", "viii"),
+        (0x2178, "M", "ix"),
+        (0x2179, "M", "x"),
+        (0x217A, "M", "xi"),
+        (0x217B, "M", "xii"),
+        (0x217C, "M", "l"),
+    ]
+
+
+def _seg_23() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x217D, "M", "c"),
+        (0x217E, "M", "d"),
+        (0x217F, "M", "m"),
+        (0x2180, "V"),
+        (0x2183, "X"),
+        (0x2184, "V"),
+        (0x2189, "M", "0⁄3"),
+        (0x218A, "V"),
+        (0x218C, "X"),
+        (0x2190, "V"),
+        (0x222C, "M", "∫∫"),
+        (0x222D, "M", "∫∫∫"),
+        (0x222E, "V"),
+        (0x222F, "M", "∮∮"),
+        (0x2230, "M", "∮∮∮"),
+        (0x2231, "V"),
+        (0x2329, "M", "〈"),
+        (0x232A, "M", "〉"),
+        (0x232B, "V"),
+        (0x2427, "X"),
+        (0x2440, "V"),
+        (0x244B, "X"),
+        (0x2460, "M", "1"),
+        (0x2461, "M", "2"),
+        (0x2462, "M", "3"),
+        (0x2463, "M", "4"),
+        (0x2464, "M", "5"),
+        (0x2465, "M", "6"),
+        (0x2466, "M", "7"),
+        (0x2467, "M", "8"),
+        (0x2468, "M", "9"),
+        (0x2469, "M", "10"),
+        (0x246A, "M", "11"),
+        (0x246B, "M", "12"),
+        (0x246C, "M", "13"),
+        (0x246D, "M", "14"),
+        (0x246E, "M", "15"),
+        (0x246F, "M", "16"),
+        (0x2470, "M", "17"),
+        (0x2471, "M", "18"),
+        (0x2472, "M", "19"),
+        (0x2473, "M", "20"),
+        (0x2474, "3", "(1)"),
+        (0x2475, "3", "(2)"),
+        (0x2476, "3", "(3)"),
+        (0x2477, "3", "(4)"),
+        (0x2478, "3", "(5)"),
+        (0x2479, "3", "(6)"),
+        (0x247A, "3", "(7)"),
+        (0x247B, "3", "(8)"),
+        (0x247C, "3", "(9)"),
+        (0x247D, "3", "(10)"),
+        (0x247E, "3", "(11)"),
+        (0x247F, "3", "(12)"),
+        (0x2480, "3", "(13)"),
+        (0x2481, "3", "(14)"),
+        (0x2482, "3", "(15)"),
+        (0x2483, "3", "(16)"),
+        (0x2484, "3", "(17)"),
+        (0x2485, "3", "(18)"),
+        (0x2486, "3", "(19)"),
+        (0x2487, "3", "(20)"),
+        (0x2488, "X"),
+        (0x249C, "3", "(a)"),
+        (0x249D, "3", "(b)"),
+        (0x249E, "3", "(c)"),
+        (0x249F, "3", "(d)"),
+        (0x24A0, "3", "(e)"),
+        (0x24A1, "3", "(f)"),
+        (0x24A2, "3", "(g)"),
+        (0x24A3, "3", "(h)"),
+        (0x24A4, "3", "(i)"),
+        (0x24A5, "3", "(j)"),
+        (0x24A6, "3", "(k)"),
+        (0x24A7, "3", "(l)"),
+        (0x24A8, "3", "(m)"),
+        (0x24A9, "3", "(n)"),
+        (0x24AA, "3", "(o)"),
+        (0x24AB, "3", "(p)"),
+        (0x24AC, "3", "(q)"),
+        (0x24AD, "3", "(r)"),
+        (0x24AE, "3", "(s)"),
+        (0x24AF, "3", "(t)"),
+        (0x24B0, "3", "(u)"),
+        (0x24B1, "3", "(v)"),
+        (0x24B2, "3", "(w)"),
+        (0x24B3, "3", "(x)"),
+        (0x24B4, "3", "(y)"),
+        (0x24B5, "3", "(z)"),
+        (0x24B6, "M", "a"),
+        (0x24B7, "M", "b"),
+        (0x24B8, "M", "c"),
+        (0x24B9, "M", "d"),
+        (0x24BA, "M", "e"),
+        (0x24BB, "M", "f"),
+        (0x24BC, "M", "g"),
+        (0x24BD, "M", "h"),
+        (0x24BE, "M", "i"),
+        (0x24BF, "M", "j"),
+        (0x24C0, "M", "k"),
+    ]
+
+
+def _seg_24() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x24C1, "M", "l"),
+        (0x24C2, "M", "m"),
+        (0x24C3, "M", "n"),
+        (0x24C4, "M", "o"),
+        (0x24C5, "M", "p"),
+        (0x24C6, "M", "q"),
+        (0x24C7, "M", "r"),
+        (0x24C8, "M", "s"),
+        (0x24C9, "M", "t"),
+        (0x24CA, "M", "u"),
+        (0x24CB, "M", "v"),
+        (0x24CC, "M", "w"),
+        (0x24CD, "M", "x"),
+        (0x24CE, "M", "y"),
+        (0x24CF, "M", "z"),
+        (0x24D0, "M", "a"),
+        (0x24D1, "M", "b"),
+        (0x24D2, "M", "c"),
+        (0x24D3, "M", "d"),
+        (0x24D4, "M", "e"),
+        (0x24D5, "M", "f"),
+        (0x24D6, "M", "g"),
+        (0x24D7, "M", "h"),
+        (0x24D8, "M", "i"),
+        (0x24D9, "M", "j"),
+        (0x24DA, "M", "k"),
+        (0x24DB, "M", "l"),
+        (0x24DC, "M", "m"),
+        (0x24DD, "M", "n"),
+        (0x24DE, "M", "o"),
+        (0x24DF, "M", "p"),
+        (0x24E0, "M", "q"),
+        (0x24E1, "M", "r"),
+        (0x24E2, "M", "s"),
+        (0x24E3, "M", "t"),
+        (0x24E4, "M", "u"),
+        (0x24E5, "M", "v"),
+        (0x24E6, "M", "w"),
+        (0x24E7, "M", "x"),
+        (0x24E8, "M", "y"),
+        (0x24E9, "M", "z"),
+        (0x24EA, "M", "0"),
+        (0x24EB, "V"),
+        (0x2A0C, "M", "∫∫∫∫"),
+        (0x2A0D, "V"),
+        (0x2A74, "3", "::="),
+        (0x2A75, "3", "=="),
+        (0x2A76, "3", "==="),
+        (0x2A77, "V"),
+        (0x2ADC, "M", "⫝̸"),
+        (0x2ADD, "V"),
+        (0x2B74, "X"),
+        (0x2B76, "V"),
+        (0x2B96, "X"),
+        (0x2B97, "V"),
+        (0x2C00, "M", "ⰰ"),
+        (0x2C01, "M", "ⰱ"),
+        (0x2C02, "M", "ⰲ"),
+        (0x2C03, "M", "ⰳ"),
+        (0x2C04, "M", "ⰴ"),
+        (0x2C05, "M", "ⰵ"),
+        (0x2C06, "M", "ⰶ"),
+        (0x2C07, "M", "ⰷ"),
+        (0x2C08, "M", "ⰸ"),
+        (0x2C09, "M", "ⰹ"),
+        (0x2C0A, "M", "ⰺ"),
+        (0x2C0B, "M", "ⰻ"),
+        (0x2C0C, "M", "ⰼ"),
+        (0x2C0D, "M", "ⰽ"),
+        (0x2C0E, "M", "ⰾ"),
+        (0x2C0F, "M", "ⰿ"),
+        (0x2C10, "M", "ⱀ"),
+        (0x2C11, "M", "ⱁ"),
+        (0x2C12, "M", "ⱂ"),
+        (0x2C13, "M", "ⱃ"),
+        (0x2C14, "M", "ⱄ"),
+        (0x2C15, "M", "ⱅ"),
+        (0x2C16, "M", "ⱆ"),
+        (0x2C17, "M", "ⱇ"),
+        (0x2C18, "M", "ⱈ"),
+        (0x2C19, "M", "ⱉ"),
+        (0x2C1A, "M", "ⱊ"),
+        (0x2C1B, "M", "ⱋ"),
+        (0x2C1C, "M", "ⱌ"),
+        (0x2C1D, "M", "ⱍ"),
+        (0x2C1E, "M", "ⱎ"),
+        (0x2C1F, "M", "ⱏ"),
+        (0x2C20, "M", "ⱐ"),
+        (0x2C21, "M", "ⱑ"),
+        (0x2C22, "M", "ⱒ"),
+        (0x2C23, "M", "ⱓ"),
+        (0x2C24, "M", "ⱔ"),
+        (0x2C25, "M", "ⱕ"),
+        (0x2C26, "M", "ⱖ"),
+        (0x2C27, "M", "ⱗ"),
+        (0x2C28, "M", "ⱘ"),
+        (0x2C29, "M", "ⱙ"),
+        (0x2C2A, "M", "ⱚ"),
+        (0x2C2B, "M", "ⱛ"),
+        (0x2C2C, "M", "ⱜ"),
+    ]
+
+
+def _seg_25() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2C2D, "M", "ⱝ"),
+        (0x2C2E, "M", "ⱞ"),
+        (0x2C2F, "M", "ⱟ"),
+        (0x2C30, "V"),
+        (0x2C60, "M", "ⱡ"),
+        (0x2C61, "V"),
+        (0x2C62, "M", "ɫ"),
+        (0x2C63, "M", "ᵽ"),
+        (0x2C64, "M", "ɽ"),
+        (0x2C65, "V"),
+        (0x2C67, "M", "ⱨ"),
+        (0x2C68, "V"),
+        (0x2C69, "M", "ⱪ"),
+        (0x2C6A, "V"),
+        (0x2C6B, "M", "ⱬ"),
+        (0x2C6C, "V"),
+        (0x2C6D, "M", "ɑ"),
+        (0x2C6E, "M", "ɱ"),
+        (0x2C6F, "M", "ɐ"),
+        (0x2C70, "M", "ɒ"),
+        (0x2C71, "V"),
+        (0x2C72, "M", "ⱳ"),
+        (0x2C73, "V"),
+        (0x2C75, "M", "ⱶ"),
+        (0x2C76, "V"),
+        (0x2C7C, "M", "j"),
+        (0x2C7D, "M", "v"),
+        (0x2C7E, "M", "ȿ"),
+        (0x2C7F, "M", "ɀ"),
+        (0x2C80, "M", "ⲁ"),
+        (0x2C81, "V"),
+        (0x2C82, "M", "ⲃ"),
+        (0x2C83, "V"),
+        (0x2C84, "M", "ⲅ"),
+        (0x2C85, "V"),
+        (0x2C86, "M", "ⲇ"),
+        (0x2C87, "V"),
+        (0x2C88, "M", "ⲉ"),
+        (0x2C89, "V"),
+        (0x2C8A, "M", "ⲋ"),
+        (0x2C8B, "V"),
+        (0x2C8C, "M", "ⲍ"),
+        (0x2C8D, "V"),
+        (0x2C8E, "M", "ⲏ"),
+        (0x2C8F, "V"),
+        (0x2C90, "M", "ⲑ"),
+        (0x2C91, "V"),
+        (0x2C92, "M", "ⲓ"),
+        (0x2C93, "V"),
+        (0x2C94, "M", "ⲕ"),
+        (0x2C95, "V"),
+        (0x2C96, "M", "ⲗ"),
+        (0x2C97, "V"),
+        (0x2C98, "M", "ⲙ"),
+        (0x2C99, "V"),
+        (0x2C9A, "M", "ⲛ"),
+        (0x2C9B, "V"),
+        (0x2C9C, "M", "ⲝ"),
+        (0x2C9D, "V"),
+        (0x2C9E, "M", "ⲟ"),
+        (0x2C9F, "V"),
+        (0x2CA0, "M", "ⲡ"),
+        (0x2CA1, "V"),
+        (0x2CA2, "M", "ⲣ"),
+        (0x2CA3, "V"),
+        (0x2CA4, "M", "ⲥ"),
+        (0x2CA5, "V"),
+        (0x2CA6, "M", "ⲧ"),
+        (0x2CA7, "V"),
+        (0x2CA8, "M", "ⲩ"),
+        (0x2CA9, "V"),
+        (0x2CAA, "M", "ⲫ"),
+        (0x2CAB, "V"),
+        (0x2CAC, "M", "ⲭ"),
+        (0x2CAD, "V"),
+        (0x2CAE, "M", "ⲯ"),
+        (0x2CAF, "V"),
+        (0x2CB0, "M", "ⲱ"),
+        (0x2CB1, "V"),
+        (0x2CB2, "M", "ⲳ"),
+        (0x2CB3, "V"),
+        (0x2CB4, "M", "ⲵ"),
+        (0x2CB5, "V"),
+        (0x2CB6, "M", "ⲷ"),
+        (0x2CB7, "V"),
+        (0x2CB8, "M", "ⲹ"),
+        (0x2CB9, "V"),
+        (0x2CBA, "M", "ⲻ"),
+        (0x2CBB, "V"),
+        (0x2CBC, "M", "ⲽ"),
+        (0x2CBD, "V"),
+        (0x2CBE, "M", "ⲿ"),
+        (0x2CBF, "V"),
+        (0x2CC0, "M", "ⳁ"),
+        (0x2CC1, "V"),
+        (0x2CC2, "M", "ⳃ"),
+        (0x2CC3, "V"),
+        (0x2CC4, "M", "ⳅ"),
+        (0x2CC5, "V"),
+        (0x2CC6, "M", "ⳇ"),
+    ]
+
+
+def _seg_26() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2CC7, "V"),
+        (0x2CC8, "M", "ⳉ"),
+        (0x2CC9, "V"),
+        (0x2CCA, "M", "ⳋ"),
+        (0x2CCB, "V"),
+        (0x2CCC, "M", "ⳍ"),
+        (0x2CCD, "V"),
+        (0x2CCE, "M", "ⳏ"),
+        (0x2CCF, "V"),
+        (0x2CD0, "M", "ⳑ"),
+        (0x2CD1, "V"),
+        (0x2CD2, "M", "ⳓ"),
+        (0x2CD3, "V"),
+        (0x2CD4, "M", "ⳕ"),
+        (0x2CD5, "V"),
+        (0x2CD6, "M", "ⳗ"),
+        (0x2CD7, "V"),
+        (0x2CD8, "M", "ⳙ"),
+        (0x2CD9, "V"),
+        (0x2CDA, "M", "ⳛ"),
+        (0x2CDB, "V"),
+        (0x2CDC, "M", "ⳝ"),
+        (0x2CDD, "V"),
+        (0x2CDE, "M", "ⳟ"),
+        (0x2CDF, "V"),
+        (0x2CE0, "M", "ⳡ"),
+        (0x2CE1, "V"),
+        (0x2CE2, "M", "ⳣ"),
+        (0x2CE3, "V"),
+        (0x2CEB, "M", "ⳬ"),
+        (0x2CEC, "V"),
+        (0x2CED, "M", "ⳮ"),
+        (0x2CEE, "V"),
+        (0x2CF2, "M", "ⳳ"),
+        (0x2CF3, "V"),
+        (0x2CF4, "X"),
+        (0x2CF9, "V"),
+        (0x2D26, "X"),
+        (0x2D27, "V"),
+        (0x2D28, "X"),
+        (0x2D2D, "V"),
+        (0x2D2E, "X"),
+        (0x2D30, "V"),
+        (0x2D68, "X"),
+        (0x2D6F, "M", "ⵡ"),
+        (0x2D70, "V"),
+        (0x2D71, "X"),
+        (0x2D7F, "V"),
+        (0x2D97, "X"),
+        (0x2DA0, "V"),
+        (0x2DA7, "X"),
+        (0x2DA8, "V"),
+        (0x2DAF, "X"),
+        (0x2DB0, "V"),
+        (0x2DB7, "X"),
+        (0x2DB8, "V"),
+        (0x2DBF, "X"),
+        (0x2DC0, "V"),
+        (0x2DC7, "X"),
+        (0x2DC8, "V"),
+        (0x2DCF, "X"),
+        (0x2DD0, "V"),
+        (0x2DD7, "X"),
+        (0x2DD8, "V"),
+        (0x2DDF, "X"),
+        (0x2DE0, "V"),
+        (0x2E5E, "X"),
+        (0x2E80, "V"),
+        (0x2E9A, "X"),
+        (0x2E9B, "V"),
+        (0x2E9F, "M", "母"),
+        (0x2EA0, "V"),
+        (0x2EF3, "M", "龟"),
+        (0x2EF4, "X"),
+        (0x2F00, "M", "一"),
+        (0x2F01, "M", "丨"),
+        (0x2F02, "M", "丶"),
+        (0x2F03, "M", "丿"),
+        (0x2F04, "M", "乙"),
+        (0x2F05, "M", "亅"),
+        (0x2F06, "M", "二"),
+        (0x2F07, "M", "亠"),
+        (0x2F08, "M", "人"),
+        (0x2F09, "M", "儿"),
+        (0x2F0A, "M", "入"),
+        (0x2F0B, "M", "八"),
+        (0x2F0C, "M", "冂"),
+        (0x2F0D, "M", "冖"),
+        (0x2F0E, "M", "冫"),
+        (0x2F0F, "M", "几"),
+        (0x2F10, "M", "凵"),
+        (0x2F11, "M", "刀"),
+        (0x2F12, "M", "力"),
+        (0x2F13, "M", "勹"),
+        (0x2F14, "M", "匕"),
+        (0x2F15, "M", "匚"),
+        (0x2F16, "M", "匸"),
+        (0x2F17, "M", "十"),
+        (0x2F18, "M", "卜"),
+        (0x2F19, "M", "卩"),
+    ]
+
+
+def _seg_27() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F1A, "M", "厂"),
+        (0x2F1B, "M", "厶"),
+        (0x2F1C, "M", "又"),
+        (0x2F1D, "M", "口"),
+        (0x2F1E, "M", "囗"),
+        (0x2F1F, "M", "土"),
+        (0x2F20, "M", "士"),
+        (0x2F21, "M", "夂"),
+        (0x2F22, "M", "夊"),
+        (0x2F23, "M", "夕"),
+        (0x2F24, "M", "大"),
+        (0x2F25, "M", "女"),
+        (0x2F26, "M", "子"),
+        (0x2F27, "M", "宀"),
+        (0x2F28, "M", "寸"),
+        (0x2F29, "M", "小"),
+        (0x2F2A, "M", "尢"),
+        (0x2F2B, "M", "尸"),
+        (0x2F2C, "M", "屮"),
+        (0x2F2D, "M", "山"),
+        (0x2F2E, "M", "巛"),
+        (0x2F2F, "M", "工"),
+        (0x2F30, "M", "己"),
+        (0x2F31, "M", "巾"),
+        (0x2F32, "M", "干"),
+        (0x2F33, "M", "幺"),
+        (0x2F34, "M", "广"),
+        (0x2F35, "M", "廴"),
+        (0x2F36, "M", "廾"),
+        (0x2F37, "M", "弋"),
+        (0x2F38, "M", "弓"),
+        (0x2F39, "M", "彐"),
+        (0x2F3A, "M", "彡"),
+        (0x2F3B, "M", "彳"),
+        (0x2F3C, "M", "心"),
+        (0x2F3D, "M", "戈"),
+        (0x2F3E, "M", "戶"),
+        (0x2F3F, "M", "手"),
+        (0x2F40, "M", "支"),
+        (0x2F41, "M", "攴"),
+        (0x2F42, "M", "文"),
+        (0x2F43, "M", "斗"),
+        (0x2F44, "M", "斤"),
+        (0x2F45, "M", "方"),
+        (0x2F46, "M", "无"),
+        (0x2F47, "M", "日"),
+        (0x2F48, "M", "曰"),
+        (0x2F49, "M", "月"),
+        (0x2F4A, "M", "木"),
+        (0x2F4B, "M", "欠"),
+        (0x2F4C, "M", "止"),
+        (0x2F4D, "M", "歹"),
+        (0x2F4E, "M", "殳"),
+        (0x2F4F, "M", "毋"),
+        (0x2F50, "M", "比"),
+        (0x2F51, "M", "毛"),
+        (0x2F52, "M", "氏"),
+        (0x2F53, "M", "气"),
+        (0x2F54, "M", "水"),
+        (0x2F55, "M", "火"),
+        (0x2F56, "M", "爪"),
+        (0x2F57, "M", "父"),
+        (0x2F58, "M", "爻"),
+        (0x2F59, "M", "爿"),
+        (0x2F5A, "M", "片"),
+        (0x2F5B, "M", "牙"),
+        (0x2F5C, "M", "牛"),
+        (0x2F5D, "M", "犬"),
+        (0x2F5E, "M", "玄"),
+        (0x2F5F, "M", "玉"),
+        (0x2F60, "M", "瓜"),
+        (0x2F61, "M", "瓦"),
+        (0x2F62, "M", "甘"),
+        (0x2F63, "M", "生"),
+        (0x2F64, "M", "用"),
+        (0x2F65, "M", "田"),
+        (0x2F66, "M", "疋"),
+        (0x2F67, "M", "疒"),
+        (0x2F68, "M", "癶"),
+        (0x2F69, "M", "白"),
+        (0x2F6A, "M", "皮"),
+        (0x2F6B, "M", "皿"),
+        (0x2F6C, "M", "目"),
+        (0x2F6D, "M", "矛"),
+        (0x2F6E, "M", "矢"),
+        (0x2F6F, "M", "石"),
+        (0x2F70, "M", "示"),
+        (0x2F71, "M", "禸"),
+        (0x2F72, "M", "禾"),
+        (0x2F73, "M", "穴"),
+        (0x2F74, "M", "立"),
+        (0x2F75, "M", "竹"),
+        (0x2F76, "M", "米"),
+        (0x2F77, "M", "糸"),
+        (0x2F78, "M", "缶"),
+        (0x2F79, "M", "网"),
+        (0x2F7A, "M", "羊"),
+        (0x2F7B, "M", "羽"),
+        (0x2F7C, "M", "老"),
+        (0x2F7D, "M", "而"),
+    ]
+
+
+def _seg_28() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F7E, "M", "耒"),
+        (0x2F7F, "M", "耳"),
+        (0x2F80, "M", "聿"),
+        (0x2F81, "M", "肉"),
+        (0x2F82, "M", "臣"),
+        (0x2F83, "M", "自"),
+        (0x2F84, "M", "至"),
+        (0x2F85, "M", "臼"),
+        (0x2F86, "M", "舌"),
+        (0x2F87, "M", "舛"),
+        (0x2F88, "M", "舟"),
+        (0x2F89, "M", "艮"),
+        (0x2F8A, "M", "色"),
+        (0x2F8B, "M", "艸"),
+        (0x2F8C, "M", "虍"),
+        (0x2F8D, "M", "虫"),
+        (0x2F8E, "M", "血"),
+        (0x2F8F, "M", "行"),
+        (0x2F90, "M", "衣"),
+        (0x2F91, "M", "襾"),
+        (0x2F92, "M", "見"),
+        (0x2F93, "M", "角"),
+        (0x2F94, "M", "言"),
+        (0x2F95, "M", "谷"),
+        (0x2F96, "M", "豆"),
+        (0x2F97, "M", "豕"),
+        (0x2F98, "M", "豸"),
+        (0x2F99, "M", "貝"),
+        (0x2F9A, "M", "赤"),
+        (0x2F9B, "M", "走"),
+        (0x2F9C, "M", "足"),
+        (0x2F9D, "M", "身"),
+        (0x2F9E, "M", "車"),
+        (0x2F9F, "M", "辛"),
+        (0x2FA0, "M", "辰"),
+        (0x2FA1, "M", "辵"),
+        (0x2FA2, "M", "邑"),
+        (0x2FA3, "M", "酉"),
+        (0x2FA4, "M", "釆"),
+        (0x2FA5, "M", "里"),
+        (0x2FA6, "M", "金"),
+        (0x2FA7, "M", "長"),
+        (0x2FA8, "M", "門"),
+        (0x2FA9, "M", "阜"),
+        (0x2FAA, "M", "隶"),
+        (0x2FAB, "M", "隹"),
+        (0x2FAC, "M", "雨"),
+        (0x2FAD, "M", "靑"),
+        (0x2FAE, "M", "非"),
+        (0x2FAF, "M", "面"),
+        (0x2FB0, "M", "革"),
+        (0x2FB1, "M", "韋"),
+        (0x2FB2, "M", "韭"),
+        (0x2FB3, "M", "音"),
+        (0x2FB4, "M", "頁"),
+        (0x2FB5, "M", "風"),
+        (0x2FB6, "M", "飛"),
+        (0x2FB7, "M", "食"),
+        (0x2FB8, "M", "首"),
+        (0x2FB9, "M", "香"),
+        (0x2FBA, "M", "馬"),
+        (0x2FBB, "M", "骨"),
+        (0x2FBC, "M", "高"),
+        (0x2FBD, "M", "髟"),
+        (0x2FBE, "M", "鬥"),
+        (0x2FBF, "M", "鬯"),
+        (0x2FC0, "M", "鬲"),
+        (0x2FC1, "M", "鬼"),
+        (0x2FC2, "M", "魚"),
+        (0x2FC3, "M", "鳥"),
+        (0x2FC4, "M", "鹵"),
+        (0x2FC5, "M", "鹿"),
+        (0x2FC6, "M", "麥"),
+        (0x2FC7, "M", "麻"),
+        (0x2FC8, "M", "黃"),
+        (0x2FC9, "M", "黍"),
+        (0x2FCA, "M", "黑"),
+        (0x2FCB, "M", "黹"),
+        (0x2FCC, "M", "黽"),
+        (0x2FCD, "M", "鼎"),
+        (0x2FCE, "M", "鼓"),
+        (0x2FCF, "M", "鼠"),
+        (0x2FD0, "M", "鼻"),
+        (0x2FD1, "M", "齊"),
+        (0x2FD2, "M", "齒"),
+        (0x2FD3, "M", "龍"),
+        (0x2FD4, "M", "龜"),
+        (0x2FD5, "M", "龠"),
+        (0x2FD6, "X"),
+        (0x3000, "3", " "),
+        (0x3001, "V"),
+        (0x3002, "M", "."),
+        (0x3003, "V"),
+        (0x3036, "M", "〒"),
+        (0x3037, "V"),
+        (0x3038, "M", "十"),
+        (0x3039, "M", "卄"),
+        (0x303A, "M", "卅"),
+        (0x303B, "V"),
+        (0x3040, "X"),
+    ]
+
+
+def _seg_29() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3041, "V"),
+        (0x3097, "X"),
+        (0x3099, "V"),
+        (0x309B, "3", " ゙"),
+        (0x309C, "3", " ゚"),
+        (0x309D, "V"),
+        (0x309F, "M", "より"),
+        (0x30A0, "V"),
+        (0x30FF, "M", "コト"),
+        (0x3100, "X"),
+        (0x3105, "V"),
+        (0x3130, "X"),
+        (0x3131, "M", "ᄀ"),
+        (0x3132, "M", "ᄁ"),
+        (0x3133, "M", "ᆪ"),
+        (0x3134, "M", "ᄂ"),
+        (0x3135, "M", "ᆬ"),
+        (0x3136, "M", "ᆭ"),
+        (0x3137, "M", "ᄃ"),
+        (0x3138, "M", "ᄄ"),
+        (0x3139, "M", "ᄅ"),
+        (0x313A, "M", "ᆰ"),
+        (0x313B, "M", "ᆱ"),
+        (0x313C, "M", "ᆲ"),
+        (0x313D, "M", "ᆳ"),
+        (0x313E, "M", "ᆴ"),
+        (0x313F, "M", "ᆵ"),
+        (0x3140, "M", "ᄚ"),
+        (0x3141, "M", "ᄆ"),
+        (0x3142, "M", "ᄇ"),
+        (0x3143, "M", "ᄈ"),
+        (0x3144, "M", "ᄡ"),
+        (0x3145, "M", "ᄉ"),
+        (0x3146, "M", "ᄊ"),
+        (0x3147, "M", "ᄋ"),
+        (0x3148, "M", "ᄌ"),
+        (0x3149, "M", "ᄍ"),
+        (0x314A, "M", "ᄎ"),
+        (0x314B, "M", "ᄏ"),
+        (0x314C, "M", "ᄐ"),
+        (0x314D, "M", "ᄑ"),
+        (0x314E, "M", "ᄒ"),
+        (0x314F, "M", "ᅡ"),
+        (0x3150, "M", "ᅢ"),
+        (0x3151, "M", "ᅣ"),
+        (0x3152, "M", "ᅤ"),
+        (0x3153, "M", "ᅥ"),
+        (0x3154, "M", "ᅦ"),
+        (0x3155, "M", "ᅧ"),
+        (0x3156, "M", "ᅨ"),
+        (0x3157, "M", "ᅩ"),
+        (0x3158, "M", "ᅪ"),
+        (0x3159, "M", "ᅫ"),
+        (0x315A, "M", "ᅬ"),
+        (0x315B, "M", "ᅭ"),
+        (0x315C, "M", "ᅮ"),
+        (0x315D, "M", "ᅯ"),
+        (0x315E, "M", "ᅰ"),
+        (0x315F, "M", "ᅱ"),
+        (0x3160, "M", "ᅲ"),
+        (0x3161, "M", "ᅳ"),
+        (0x3162, "M", "ᅴ"),
+        (0x3163, "M", "ᅵ"),
+        (0x3164, "X"),
+        (0x3165, "M", "ᄔ"),
+        (0x3166, "M", "ᄕ"),
+        (0x3167, "M", "ᇇ"),
+        (0x3168, "M", "ᇈ"),
+        (0x3169, "M", "ᇌ"),
+        (0x316A, "M", "ᇎ"),
+        (0x316B, "M", "ᇓ"),
+        (0x316C, "M", "ᇗ"),
+        (0x316D, "M", "ᇙ"),
+        (0x316E, "M", "ᄜ"),
+        (0x316F, "M", "ᇝ"),
+        (0x3170, "M", "ᇟ"),
+        (0x3171, "M", "ᄝ"),
+        (0x3172, "M", "ᄞ"),
+        (0x3173, "M", "ᄠ"),
+        (0x3174, "M", "ᄢ"),
+        (0x3175, "M", "ᄣ"),
+        (0x3176, "M", "ᄧ"),
+        (0x3177, "M", "ᄩ"),
+        (0x3178, "M", "ᄫ"),
+        (0x3179, "M", "ᄬ"),
+        (0x317A, "M", "ᄭ"),
+        (0x317B, "M", "ᄮ"),
+        (0x317C, "M", "ᄯ"),
+        (0x317D, "M", "ᄲ"),
+        (0x317E, "M", "ᄶ"),
+        (0x317F, "M", "ᅀ"),
+        (0x3180, "M", "ᅇ"),
+        (0x3181, "M", "ᅌ"),
+        (0x3182, "M", "ᇱ"),
+        (0x3183, "M", "ᇲ"),
+        (0x3184, "M", "ᅗ"),
+        (0x3185, "M", "ᅘ"),
+        (0x3186, "M", "ᅙ"),
+        (0x3187, "M", "ᆄ"),
+        (0x3188, "M", "ᆅ"),
+    ]
+
+
+def _seg_30() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3189, "M", "ᆈ"),
+        (0x318A, "M", "ᆑ"),
+        (0x318B, "M", "ᆒ"),
+        (0x318C, "M", "ᆔ"),
+        (0x318D, "M", "ᆞ"),
+        (0x318E, "M", "ᆡ"),
+        (0x318F, "X"),
+        (0x3190, "V"),
+        (0x3192, "M", "一"),
+        (0x3193, "M", "二"),
+        (0x3194, "M", "三"),
+        (0x3195, "M", "四"),
+        (0x3196, "M", "上"),
+        (0x3197, "M", "中"),
+        (0x3198, "M", "下"),
+        (0x3199, "M", "甲"),
+        (0x319A, "M", "乙"),
+        (0x319B, "M", "丙"),
+        (0x319C, "M", "丁"),
+        (0x319D, "M", "天"),
+        (0x319E, "M", "地"),
+        (0x319F, "M", "人"),
+        (0x31A0, "V"),
+        (0x31E4, "X"),
+        (0x31F0, "V"),
+        (0x3200, "3", "(ᄀ)"),
+        (0x3201, "3", "(ᄂ)"),
+        (0x3202, "3", "(ᄃ)"),
+        (0x3203, "3", "(ᄅ)"),
+        (0x3204, "3", "(ᄆ)"),
+        (0x3205, "3", "(ᄇ)"),
+        (0x3206, "3", "(ᄉ)"),
+        (0x3207, "3", "(ᄋ)"),
+        (0x3208, "3", "(ᄌ)"),
+        (0x3209, "3", "(ᄎ)"),
+        (0x320A, "3", "(ᄏ)"),
+        (0x320B, "3", "(ᄐ)"),
+        (0x320C, "3", "(ᄑ)"),
+        (0x320D, "3", "(ᄒ)"),
+        (0x320E, "3", "(가)"),
+        (0x320F, "3", "(나)"),
+        (0x3210, "3", "(다)"),
+        (0x3211, "3", "(라)"),
+        (0x3212, "3", "(마)"),
+        (0x3213, "3", "(바)"),
+        (0x3214, "3", "(사)"),
+        (0x3215, "3", "(아)"),
+        (0x3216, "3", "(자)"),
+        (0x3217, "3", "(차)"),
+        (0x3218, "3", "(카)"),
+        (0x3219, "3", "(타)"),
+        (0x321A, "3", "(파)"),
+        (0x321B, "3", "(하)"),
+        (0x321C, "3", "(주)"),
+        (0x321D, "3", "(오전)"),
+        (0x321E, "3", "(오후)"),
+        (0x321F, "X"),
+        (0x3220, "3", "(一)"),
+        (0x3221, "3", "(二)"),
+        (0x3222, "3", "(三)"),
+        (0x3223, "3", "(四)"),
+        (0x3224, "3", "(五)"),
+        (0x3225, "3", "(六)"),
+        (0x3226, "3", "(七)"),
+        (0x3227, "3", "(八)"),
+        (0x3228, "3", "(九)"),
+        (0x3229, "3", "(十)"),
+        (0x322A, "3", "(月)"),
+        (0x322B, "3", "(火)"),
+        (0x322C, "3", "(水)"),
+        (0x322D, "3", "(木)"),
+        (0x322E, "3", "(金)"),
+        (0x322F, "3", "(土)"),
+        (0x3230, "3", "(日)"),
+        (0x3231, "3", "(株)"),
+        (0x3232, "3", "(有)"),
+        (0x3233, "3", "(社)"),
+        (0x3234, "3", "(名)"),
+        (0x3235, "3", "(特)"),
+        (0x3236, "3", "(財)"),
+        (0x3237, "3", "(祝)"),
+        (0x3238, "3", "(労)"),
+        (0x3239, "3", "(代)"),
+        (0x323A, "3", "(呼)"),
+        (0x323B, "3", "(学)"),
+        (0x323C, "3", "(監)"),
+        (0x323D, "3", "(企)"),
+        (0x323E, "3", "(資)"),
+        (0x323F, "3", "(協)"),
+        (0x3240, "3", "(祭)"),
+        (0x3241, "3", "(休)"),
+        (0x3242, "3", "(自)"),
+        (0x3243, "3", "(至)"),
+        (0x3244, "M", "問"),
+        (0x3245, "M", "幼"),
+        (0x3246, "M", "文"),
+        (0x3247, "M", "箏"),
+        (0x3248, "V"),
+        (0x3250, "M", "pte"),
+        (0x3251, "M", "21"),
+    ]
+
+
+def _seg_31() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x3252, "M", "22"),
+        (0x3253, "M", "23"),
+        (0x3254, "M", "24"),
+        (0x3255, "M", "25"),
+        (0x3256, "M", "26"),
+        (0x3257, "M", "27"),
+        (0x3258, "M", "28"),
+        (0x3259, "M", "29"),
+        (0x325A, "M", "30"),
+        (0x325B, "M", "31"),
+        (0x325C, "M", "32"),
+        (0x325D, "M", "33"),
+        (0x325E, "M", "34"),
+        (0x325F, "M", "35"),
+        (0x3260, "M", "ᄀ"),
+        (0x3261, "M", "ᄂ"),
+        (0x3262, "M", "ᄃ"),
+        (0x3263, "M", "ᄅ"),
+        (0x3264, "M", "ᄆ"),
+        (0x3265, "M", "ᄇ"),
+        (0x3266, "M", "ᄉ"),
+        (0x3267, "M", "ᄋ"),
+        (0x3268, "M", "ᄌ"),
+        (0x3269, "M", "ᄎ"),
+        (0x326A, "M", "ᄏ"),
+        (0x326B, "M", "ᄐ"),
+        (0x326C, "M", "ᄑ"),
+        (0x326D, "M", "ᄒ"),
+        (0x326E, "M", "가"),
+        (0x326F, "M", "나"),
+        (0x3270, "M", "다"),
+        (0x3271, "M", "라"),
+        (0x3272, "M", "마"),
+        (0x3273, "M", "바"),
+        (0x3274, "M", "사"),
+        (0x3275, "M", "아"),
+        (0x3276, "M", "자"),
+        (0x3277, "M", "차"),
+        (0x3278, "M", "카"),
+        (0x3279, "M", "타"),
+        (0x327A, "M", "파"),
+        (0x327B, "M", "하"),
+        (0x327C, "M", "참고"),
+        (0x327D, "M", "주의"),
+        (0x327E, "M", "우"),
+        (0x327F, "V"),
+        (0x3280, "M", "一"),
+        (0x3281, "M", "二"),
+        (0x3282, "M", "三"),
+        (0x3283, "M", "四"),
+        (0x3284, "M", "五"),
+        (0x3285, "M", "六"),
+        (0x3286, "M", "七"),
+        (0x3287, "M", "八"),
+        (0x3288, "M", "九"),
+        (0x3289, "M", "十"),
+        (0x328A, "M", "月"),
+        (0x328B, "M", "火"),
+        (0x328C, "M", "水"),
+        (0x328D, "M", "木"),
+        (0x328E, "M", "金"),
+        (0x328F, "M", "土"),
+        (0x3290, "M", "日"),
+        (0x3291, "M", "株"),
+        (0x3292, "M", "有"),
+        (0x3293, "M", "社"),
+        (0x3294, "M", "名"),
+        (0x3295, "M", "特"),
+        (0x3296, "M", "財"),
+        (0x3297, "M", "祝"),
+        (0x3298, "M", "労"),
+        (0x3299, "M", "秘"),
+        (0x329A, "M", "男"),
+        (0x329B, "M", "女"),
+        (0x329C, "M", "適"),
+        (0x329D, "M", "優"),
+        (0x329E, "M", "印"),
+        (0x329F, "M", "注"),
+        (0x32A0, "M", "項"),
+        (0x32A1, "M", "休"),
+        (0x32A2, "M", "写"),
+        (0x32A3, "M", "正"),
+        (0x32A4, "M", "上"),
+        (0x32A5, "M", "中"),
+        (0x32A6, "M", "下"),
+        (0x32A7, "M", "左"),
+        (0x32A8, "M", "右"),
+        (0x32A9, "M", "医"),
+        (0x32AA, "M", "宗"),
+        (0x32AB, "M", "学"),
+        (0x32AC, "M", "監"),
+        (0x32AD, "M", "企"),
+        (0x32AE, "M", "資"),
+        (0x32AF, "M", "協"),
+        (0x32B0, "M", "夜"),
+        (0x32B1, "M", "36"),
+        (0x32B2, "M", "37"),
+        (0x32B3, "M", "38"),
+        (0x32B4, "M", "39"),
+        (0x32B5, "M", "40"),
+    ]
+
+
+def _seg_32() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x32B6, "M", "41"),
+        (0x32B7, "M", "42"),
+        (0x32B8, "M", "43"),
+        (0x32B9, "M", "44"),
+        (0x32BA, "M", "45"),
+        (0x32BB, "M", "46"),
+        (0x32BC, "M", "47"),
+        (0x32BD, "M", "48"),
+        (0x32BE, "M", "49"),
+        (0x32BF, "M", "50"),
+        (0x32C0, "M", "1月"),
+        (0x32C1, "M", "2月"),
+        (0x32C2, "M", "3月"),
+        (0x32C3, "M", "4月"),
+        (0x32C4, "M", "5月"),
+        (0x32C5, "M", "6月"),
+        (0x32C6, "M", "7月"),
+        (0x32C7, "M", "8月"),
+        (0x32C8, "M", "9月"),
+        (0x32C9, "M", "10月"),
+        (0x32CA, "M", "11月"),
+        (0x32CB, "M", "12月"),
+        (0x32CC, "M", "hg"),
+        (0x32CD, "M", "erg"),
+        (0x32CE, "M", "ev"),
+        (0x32CF, "M", "ltd"),
+        (0x32D0, "M", "ア"),
+        (0x32D1, "M", "イ"),
+        (0x32D2, "M", "ウ"),
+        (0x32D3, "M", "エ"),
+        (0x32D4, "M", "オ"),
+        (0x32D5, "M", "カ"),
+        (0x32D6, "M", "キ"),
+        (0x32D7, "M", "ク"),
+        (0x32D8, "M", "ケ"),
+        (0x32D9, "M", "コ"),
+        (0x32DA, "M", "サ"),
+        (0x32DB, "M", "シ"),
+        (0x32DC, "M", "ス"),
+        (0x32DD, "M", "セ"),
+        (0x32DE, "M", "ソ"),
+        (0x32DF, "M", "タ"),
+        (0x32E0, "M", "チ"),
+        (0x32E1, "M", "ツ"),
+        (0x32E2, "M", "テ"),
+        (0x32E3, "M", "ト"),
+        (0x32E4, "M", "ナ"),
+        (0x32E5, "M", "ニ"),
+        (0x32E6, "M", "ヌ"),
+        (0x32E7, "M", "ネ"),
+        (0x32E8, "M", "ノ"),
+        (0x32E9, "M", "ハ"),
+        (0x32EA, "M", "ヒ"),
+        (0x32EB, "M", "フ"),
+        (0x32EC, "M", "ヘ"),
+        (0x32ED, "M", "ホ"),
+        (0x32EE, "M", "マ"),
+        (0x32EF, "M", "ミ"),
+        (0x32F0, "M", "ム"),
+        (0x32F1, "M", "メ"),
+        (0x32F2, "M", "モ"),
+        (0x32F3, "M", "ヤ"),
+        (0x32F4, "M", "ユ"),
+        (0x32F5, "M", "ヨ"),
+        (0x32F6, "M", "ラ"),
+        (0x32F7, "M", "リ"),
+        (0x32F8, "M", "ル"),
+        (0x32F9, "M", "レ"),
+        (0x32FA, "M", "ロ"),
+        (0x32FB, "M", "ワ"),
+        (0x32FC, "M", "ヰ"),
+        (0x32FD, "M", "ヱ"),
+        (0x32FE, "M", "ヲ"),
+        (0x32FF, "M", "令和"),
+        (0x3300, "M", "アパート"),
+        (0x3301, "M", "アルファ"),
+        (0x3302, "M", "アンペア"),
+        (0x3303, "M", "アール"),
+        (0x3304, "M", "イニング"),
+        (0x3305, "M", "インチ"),
+        (0x3306, "M", "ウォン"),
+        (0x3307, "M", "エスクード"),
+        (0x3308, "M", "エーカー"),
+        (0x3309, "M", "オンス"),
+        (0x330A, "M", "オーム"),
+        (0x330B, "M", "カイリ"),
+        (0x330C, "M", "カラット"),
+        (0x330D, "M", "カロリー"),
+        (0x330E, "M", "ガロン"),
+        (0x330F, "M", "ガンマ"),
+        (0x3310, "M", "ギガ"),
+        (0x3311, "M", "ギニー"),
+        (0x3312, "M", "キュリー"),
+        (0x3313, "M", "ギルダー"),
+        (0x3314, "M", "キロ"),
+        (0x3315, "M", "キログラム"),
+        (0x3316, "M", "キロメートル"),
+        (0x3317, "M", "キロワット"),
+        (0x3318, "M", "グラム"),
+        (0x3319, "M", "グラムトン"),
+    ]
+
+
+def _seg_33() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x331A, "M", "クルゼイロ"),
+        (0x331B, "M", "クローネ"),
+        (0x331C, "M", "ケース"),
+        (0x331D, "M", "コルナ"),
+        (0x331E, "M", "コーポ"),
+        (0x331F, "M", "サイクル"),
+        (0x3320, "M", "サンチーム"),
+        (0x3321, "M", "シリング"),
+        (0x3322, "M", "センチ"),
+        (0x3323, "M", "セント"),
+        (0x3324, "M", "ダース"),
+        (0x3325, "M", "デシ"),
+        (0x3326, "M", "ドル"),
+        (0x3327, "M", "トン"),
+        (0x3328, "M", "ナノ"),
+        (0x3329, "M", "ノット"),
+        (0x332A, "M", "ハイツ"),
+        (0x332B, "M", "パーセント"),
+        (0x332C, "M", "パーツ"),
+        (0x332D, "M", "バーレル"),
+        (0x332E, "M", "ピアストル"),
+        (0x332F, "M", "ピクル"),
+        (0x3330, "M", "ピコ"),
+        (0x3331, "M", "ビル"),
+        (0x3332, "M", "ファラッド"),
+        (0x3333, "M", "フィート"),
+        (0x3334, "M", "ブッシェル"),
+        (0x3335, "M", "フラン"),
+        (0x3336, "M", "ヘクタール"),
+        (0x3337, "M", "ペソ"),
+        (0x3338, "M", "ペニヒ"),
+        (0x3339, "M", "ヘルツ"),
+        (0x333A, "M", "ペンス"),
+        (0x333B, "M", "ページ"),
+        (0x333C, "M", "ベータ"),
+        (0x333D, "M", "ポイント"),
+        (0x333E, "M", "ボルト"),
+        (0x333F, "M", "ホン"),
+        (0x3340, "M", "ポンド"),
+        (0x3341, "M", "ホール"),
+        (0x3342, "M", "ホーン"),
+        (0x3343, "M", "マイクロ"),
+        (0x3344, "M", "マイル"),
+        (0x3345, "M", "マッハ"),
+        (0x3346, "M", "マルク"),
+        (0x3347, "M", "マンション"),
+        (0x3348, "M", "ミクロン"),
+        (0x3349, "M", "ミリ"),
+        (0x334A, "M", "ミリバール"),
+        (0x334B, "M", "メガ"),
+        (0x334C, "M", "メガトン"),
+        (0x334D, "M", "メートル"),
+        (0x334E, "M", "ヤード"),
+        (0x334F, "M", "ヤール"),
+        (0x3350, "M", "ユアン"),
+        (0x3351, "M", "リットル"),
+        (0x3352, "M", "リラ"),
+        (0x3353, "M", "ルピー"),
+        (0x3354, "M", "ルーブル"),
+        (0x3355, "M", "レム"),
+        (0x3356, "M", "レントゲン"),
+        (0x3357, "M", "ワット"),
+        (0x3358, "M", "0点"),
+        (0x3359, "M", "1点"),
+        (0x335A, "M", "2点"),
+        (0x335B, "M", "3点"),
+        (0x335C, "M", "4点"),
+        (0x335D, "M", "5点"),
+        (0x335E, "M", "6点"),
+        (0x335F, "M", "7点"),
+        (0x3360, "M", "8点"),
+        (0x3361, "M", "9点"),
+        (0x3362, "M", "10点"),
+        (0x3363, "M", "11点"),
+        (0x3364, "M", "12点"),
+        (0x3365, "M", "13点"),
+        (0x3366, "M", "14点"),
+        (0x3367, "M", "15点"),
+        (0x3368, "M", "16点"),
+        (0x3369, "M", "17点"),
+        (0x336A, "M", "18点"),
+        (0x336B, "M", "19点"),
+        (0x336C, "M", "20点"),
+        (0x336D, "M", "21点"),
+        (0x336E, "M", "22点"),
+        (0x336F, "M", "23点"),
+        (0x3370, "M", "24点"),
+        (0x3371, "M", "hpa"),
+        (0x3372, "M", "da"),
+        (0x3373, "M", "au"),
+        (0x3374, "M", "bar"),
+        (0x3375, "M", "ov"),
+        (0x3376, "M", "pc"),
+        (0x3377, "M", "dm"),
+        (0x3378, "M", "dm2"),
+        (0x3379, "M", "dm3"),
+        (0x337A, "M", "iu"),
+        (0x337B, "M", "平成"),
+        (0x337C, "M", "昭和"),
+        (0x337D, "M", "大正"),
+    ]
+
+
+def _seg_34() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x337E, "M", "明治"),
+        (0x337F, "M", "株式会社"),
+        (0x3380, "M", "pa"),
+        (0x3381, "M", "na"),
+        (0x3382, "M", "μa"),
+        (0x3383, "M", "ma"),
+        (0x3384, "M", "ka"),
+        (0x3385, "M", "kb"),
+        (0x3386, "M", "mb"),
+        (0x3387, "M", "gb"),
+        (0x3388, "M", "cal"),
+        (0x3389, "M", "kcal"),
+        (0x338A, "M", "pf"),
+        (0x338B, "M", "nf"),
+        (0x338C, "M", "μf"),
+        (0x338D, "M", "μg"),
+        (0x338E, "M", "mg"),
+        (0x338F, "M", "kg"),
+        (0x3390, "M", "hz"),
+        (0x3391, "M", "khz"),
+        (0x3392, "M", "mhz"),
+        (0x3393, "M", "ghz"),
+        (0x3394, "M", "thz"),
+        (0x3395, "M", "μl"),
+        (0x3396, "M", "ml"),
+        (0x3397, "M", "dl"),
+        (0x3398, "M", "kl"),
+        (0x3399, "M", "fm"),
+        (0x339A, "M", "nm"),
+        (0x339B, "M", "μm"),
+        (0x339C, "M", "mm"),
+        (0x339D, "M", "cm"),
+        (0x339E, "M", "km"),
+        (0x339F, "M", "mm2"),
+        (0x33A0, "M", "cm2"),
+        (0x33A1, "M", "m2"),
+        (0x33A2, "M", "km2"),
+        (0x33A3, "M", "mm3"),
+        (0x33A4, "M", "cm3"),
+        (0x33A5, "M", "m3"),
+        (0x33A6, "M", "km3"),
+        (0x33A7, "M", "m∕s"),
+        (0x33A8, "M", "m∕s2"),
+        (0x33A9, "M", "pa"),
+        (0x33AA, "M", "kpa"),
+        (0x33AB, "M", "mpa"),
+        (0x33AC, "M", "gpa"),
+        (0x33AD, "M", "rad"),
+        (0x33AE, "M", "rad∕s"),
+        (0x33AF, "M", "rad∕s2"),
+        (0x33B0, "M", "ps"),
+        (0x33B1, "M", "ns"),
+        (0x33B2, "M", "μs"),
+        (0x33B3, "M", "ms"),
+        (0x33B4, "M", "pv"),
+        (0x33B5, "M", "nv"),
+        (0x33B6, "M", "μv"),
+        (0x33B7, "M", "mv"),
+        (0x33B8, "M", "kv"),
+        (0x33B9, "M", "mv"),
+        (0x33BA, "M", "pw"),
+        (0x33BB, "M", "nw"),
+        (0x33BC, "M", "μw"),
+        (0x33BD, "M", "mw"),
+        (0x33BE, "M", "kw"),
+        (0x33BF, "M", "mw"),
+        (0x33C0, "M", "kω"),
+        (0x33C1, "M", "mω"),
+        (0x33C2, "X"),
+        (0x33C3, "M", "bq"),
+        (0x33C4, "M", "cc"),
+        (0x33C5, "M", "cd"),
+        (0x33C6, "M", "c∕kg"),
+        (0x33C7, "X"),
+        (0x33C8, "M", "db"),
+        (0x33C9, "M", "gy"),
+        (0x33CA, "M", "ha"),
+        (0x33CB, "M", "hp"),
+        (0x33CC, "M", "in"),
+        (0x33CD, "M", "kk"),
+        (0x33CE, "M", "km"),
+        (0x33CF, "M", "kt"),
+        (0x33D0, "M", "lm"),
+        (0x33D1, "M", "ln"),
+        (0x33D2, "M", "log"),
+        (0x33D3, "M", "lx"),
+        (0x33D4, "M", "mb"),
+        (0x33D5, "M", "mil"),
+        (0x33D6, "M", "mol"),
+        (0x33D7, "M", "ph"),
+        (0x33D8, "X"),
+        (0x33D9, "M", "ppm"),
+        (0x33DA, "M", "pr"),
+        (0x33DB, "M", "sr"),
+        (0x33DC, "M", "sv"),
+        (0x33DD, "M", "wb"),
+        (0x33DE, "M", "v∕m"),
+        (0x33DF, "M", "a∕m"),
+        (0x33E0, "M", "1日"),
+        (0x33E1, "M", "2日"),
+    ]
+
+
+def _seg_35() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x33E2, "M", "3日"),
+        (0x33E3, "M", "4日"),
+        (0x33E4, "M", "5日"),
+        (0x33E5, "M", "6日"),
+        (0x33E6, "M", "7日"),
+        (0x33E7, "M", "8日"),
+        (0x33E8, "M", "9日"),
+        (0x33E9, "M", "10日"),
+        (0x33EA, "M", "11日"),
+        (0x33EB, "M", "12日"),
+        (0x33EC, "M", "13日"),
+        (0x33ED, "M", "14日"),
+        (0x33EE, "M", "15日"),
+        (0x33EF, "M", "16日"),
+        (0x33F0, "M", "17日"),
+        (0x33F1, "M", "18日"),
+        (0x33F2, "M", "19日"),
+        (0x33F3, "M", "20日"),
+        (0x33F4, "M", "21日"),
+        (0x33F5, "M", "22日"),
+        (0x33F6, "M", "23日"),
+        (0x33F7, "M", "24日"),
+        (0x33F8, "M", "25日"),
+        (0x33F9, "M", "26日"),
+        (0x33FA, "M", "27日"),
+        (0x33FB, "M", "28日"),
+        (0x33FC, "M", "29日"),
+        (0x33FD, "M", "30日"),
+        (0x33FE, "M", "31日"),
+        (0x33FF, "M", "gal"),
+        (0x3400, "V"),
+        (0xA48D, "X"),
+        (0xA490, "V"),
+        (0xA4C7, "X"),
+        (0xA4D0, "V"),
+        (0xA62C, "X"),
+        (0xA640, "M", "ꙁ"),
+        (0xA641, "V"),
+        (0xA642, "M", "ꙃ"),
+        (0xA643, "V"),
+        (0xA644, "M", "ꙅ"),
+        (0xA645, "V"),
+        (0xA646, "M", "ꙇ"),
+        (0xA647, "V"),
+        (0xA648, "M", "ꙉ"),
+        (0xA649, "V"),
+        (0xA64A, "M", "ꙋ"),
+        (0xA64B, "V"),
+        (0xA64C, "M", "ꙍ"),
+        (0xA64D, "V"),
+        (0xA64E, "M", "ꙏ"),
+        (0xA64F, "V"),
+        (0xA650, "M", "ꙑ"),
+        (0xA651, "V"),
+        (0xA652, "M", "ꙓ"),
+        (0xA653, "V"),
+        (0xA654, "M", "ꙕ"),
+        (0xA655, "V"),
+        (0xA656, "M", "ꙗ"),
+        (0xA657, "V"),
+        (0xA658, "M", "ꙙ"),
+        (0xA659, "V"),
+        (0xA65A, "M", "ꙛ"),
+        (0xA65B, "V"),
+        (0xA65C, "M", "ꙝ"),
+        (0xA65D, "V"),
+        (0xA65E, "M", "ꙟ"),
+        (0xA65F, "V"),
+        (0xA660, "M", "ꙡ"),
+        (0xA661, "V"),
+        (0xA662, "M", "ꙣ"),
+        (0xA663, "V"),
+        (0xA664, "M", "ꙥ"),
+        (0xA665, "V"),
+        (0xA666, "M", "ꙧ"),
+        (0xA667, "V"),
+        (0xA668, "M", "ꙩ"),
+        (0xA669, "V"),
+        (0xA66A, "M", "ꙫ"),
+        (0xA66B, "V"),
+        (0xA66C, "M", "ꙭ"),
+        (0xA66D, "V"),
+        (0xA680, "M", "ꚁ"),
+        (0xA681, "V"),
+        (0xA682, "M", "ꚃ"),
+        (0xA683, "V"),
+        (0xA684, "M", "ꚅ"),
+        (0xA685, "V"),
+        (0xA686, "M", "ꚇ"),
+        (0xA687, "V"),
+        (0xA688, "M", "ꚉ"),
+        (0xA689, "V"),
+        (0xA68A, "M", "ꚋ"),
+        (0xA68B, "V"),
+        (0xA68C, "M", "ꚍ"),
+        (0xA68D, "V"),
+        (0xA68E, "M", "ꚏ"),
+        (0xA68F, "V"),
+        (0xA690, "M", "ꚑ"),
+        (0xA691, "V"),
+    ]
+
+
+def _seg_36() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA692, "M", "ꚓ"),
+        (0xA693, "V"),
+        (0xA694, "M", "ꚕ"),
+        (0xA695, "V"),
+        (0xA696, "M", "ꚗ"),
+        (0xA697, "V"),
+        (0xA698, "M", "ꚙ"),
+        (0xA699, "V"),
+        (0xA69A, "M", "ꚛ"),
+        (0xA69B, "V"),
+        (0xA69C, "M", "ъ"),
+        (0xA69D, "M", "ь"),
+        (0xA69E, "V"),
+        (0xA6F8, "X"),
+        (0xA700, "V"),
+        (0xA722, "M", "ꜣ"),
+        (0xA723, "V"),
+        (0xA724, "M", "ꜥ"),
+        (0xA725, "V"),
+        (0xA726, "M", "ꜧ"),
+        (0xA727, "V"),
+        (0xA728, "M", "ꜩ"),
+        (0xA729, "V"),
+        (0xA72A, "M", "ꜫ"),
+        (0xA72B, "V"),
+        (0xA72C, "M", "ꜭ"),
+        (0xA72D, "V"),
+        (0xA72E, "M", "ꜯ"),
+        (0xA72F, "V"),
+        (0xA732, "M", "ꜳ"),
+        (0xA733, "V"),
+        (0xA734, "M", "ꜵ"),
+        (0xA735, "V"),
+        (0xA736, "M", "ꜷ"),
+        (0xA737, "V"),
+        (0xA738, "M", "ꜹ"),
+        (0xA739, "V"),
+        (0xA73A, "M", "ꜻ"),
+        (0xA73B, "V"),
+        (0xA73C, "M", "ꜽ"),
+        (0xA73D, "V"),
+        (0xA73E, "M", "ꜿ"),
+        (0xA73F, "V"),
+        (0xA740, "M", "ꝁ"),
+        (0xA741, "V"),
+        (0xA742, "M", "ꝃ"),
+        (0xA743, "V"),
+        (0xA744, "M", "ꝅ"),
+        (0xA745, "V"),
+        (0xA746, "M", "ꝇ"),
+        (0xA747, "V"),
+        (0xA748, "M", "ꝉ"),
+        (0xA749, "V"),
+        (0xA74A, "M", "ꝋ"),
+        (0xA74B, "V"),
+        (0xA74C, "M", "ꝍ"),
+        (0xA74D, "V"),
+        (0xA74E, "M", "ꝏ"),
+        (0xA74F, "V"),
+        (0xA750, "M", "ꝑ"),
+        (0xA751, "V"),
+        (0xA752, "M", "ꝓ"),
+        (0xA753, "V"),
+        (0xA754, "M", "ꝕ"),
+        (0xA755, "V"),
+        (0xA756, "M", "ꝗ"),
+        (0xA757, "V"),
+        (0xA758, "M", "ꝙ"),
+        (0xA759, "V"),
+        (0xA75A, "M", "ꝛ"),
+        (0xA75B, "V"),
+        (0xA75C, "M", "ꝝ"),
+        (0xA75D, "V"),
+        (0xA75E, "M", "ꝟ"),
+        (0xA75F, "V"),
+        (0xA760, "M", "ꝡ"),
+        (0xA761, "V"),
+        (0xA762, "M", "ꝣ"),
+        (0xA763, "V"),
+        (0xA764, "M", "ꝥ"),
+        (0xA765, "V"),
+        (0xA766, "M", "ꝧ"),
+        (0xA767, "V"),
+        (0xA768, "M", "ꝩ"),
+        (0xA769, "V"),
+        (0xA76A, "M", "ꝫ"),
+        (0xA76B, "V"),
+        (0xA76C, "M", "ꝭ"),
+        (0xA76D, "V"),
+        (0xA76E, "M", "ꝯ"),
+        (0xA76F, "V"),
+        (0xA770, "M", "ꝯ"),
+        (0xA771, "V"),
+        (0xA779, "M", "ꝺ"),
+        (0xA77A, "V"),
+        (0xA77B, "M", "ꝼ"),
+        (0xA77C, "V"),
+        (0xA77D, "M", "ᵹ"),
+        (0xA77E, "M", "ꝿ"),
+        (0xA77F, "V"),
+    ]
+
+
+def _seg_37() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA780, "M", "ꞁ"),
+        (0xA781, "V"),
+        (0xA782, "M", "ꞃ"),
+        (0xA783, "V"),
+        (0xA784, "M", "ꞅ"),
+        (0xA785, "V"),
+        (0xA786, "M", "ꞇ"),
+        (0xA787, "V"),
+        (0xA78B, "M", "ꞌ"),
+        (0xA78C, "V"),
+        (0xA78D, "M", "ɥ"),
+        (0xA78E, "V"),
+        (0xA790, "M", "ꞑ"),
+        (0xA791, "V"),
+        (0xA792, "M", "ꞓ"),
+        (0xA793, "V"),
+        (0xA796, "M", "ꞗ"),
+        (0xA797, "V"),
+        (0xA798, "M", "ꞙ"),
+        (0xA799, "V"),
+        (0xA79A, "M", "ꞛ"),
+        (0xA79B, "V"),
+        (0xA79C, "M", "ꞝ"),
+        (0xA79D, "V"),
+        (0xA79E, "M", "ꞟ"),
+        (0xA79F, "V"),
+        (0xA7A0, "M", "ꞡ"),
+        (0xA7A1, "V"),
+        (0xA7A2, "M", "ꞣ"),
+        (0xA7A3, "V"),
+        (0xA7A4, "M", "ꞥ"),
+        (0xA7A5, "V"),
+        (0xA7A6, "M", "ꞧ"),
+        (0xA7A7, "V"),
+        (0xA7A8, "M", "ꞩ"),
+        (0xA7A9, "V"),
+        (0xA7AA, "M", "ɦ"),
+        (0xA7AB, "M", "ɜ"),
+        (0xA7AC, "M", "ɡ"),
+        (0xA7AD, "M", "ɬ"),
+        (0xA7AE, "M", "ɪ"),
+        (0xA7AF, "V"),
+        (0xA7B0, "M", "ʞ"),
+        (0xA7B1, "M", "ʇ"),
+        (0xA7B2, "M", "ʝ"),
+        (0xA7B3, "M", "ꭓ"),
+        (0xA7B4, "M", "ꞵ"),
+        (0xA7B5, "V"),
+        (0xA7B6, "M", "ꞷ"),
+        (0xA7B7, "V"),
+        (0xA7B8, "M", "ꞹ"),
+        (0xA7B9, "V"),
+        (0xA7BA, "M", "ꞻ"),
+        (0xA7BB, "V"),
+        (0xA7BC, "M", "ꞽ"),
+        (0xA7BD, "V"),
+        (0xA7BE, "M", "ꞿ"),
+        (0xA7BF, "V"),
+        (0xA7C0, "M", "ꟁ"),
+        (0xA7C1, "V"),
+        (0xA7C2, "M", "ꟃ"),
+        (0xA7C3, "V"),
+        (0xA7C4, "M", "ꞔ"),
+        (0xA7C5, "M", "ʂ"),
+        (0xA7C6, "M", "ᶎ"),
+        (0xA7C7, "M", "ꟈ"),
+        (0xA7C8, "V"),
+        (0xA7C9, "M", "ꟊ"),
+        (0xA7CA, "V"),
+        (0xA7CB, "X"),
+        (0xA7D0, "M", "ꟑ"),
+        (0xA7D1, "V"),
+        (0xA7D2, "X"),
+        (0xA7D3, "V"),
+        (0xA7D4, "X"),
+        (0xA7D5, "V"),
+        (0xA7D6, "M", "ꟗ"),
+        (0xA7D7, "V"),
+        (0xA7D8, "M", "ꟙ"),
+        (0xA7D9, "V"),
+        (0xA7DA, "X"),
+        (0xA7F2, "M", "c"),
+        (0xA7F3, "M", "f"),
+        (0xA7F4, "M", "q"),
+        (0xA7F5, "M", "ꟶ"),
+        (0xA7F6, "V"),
+        (0xA7F8, "M", "ħ"),
+        (0xA7F9, "M", "œ"),
+        (0xA7FA, "V"),
+        (0xA82D, "X"),
+        (0xA830, "V"),
+        (0xA83A, "X"),
+        (0xA840, "V"),
+        (0xA878, "X"),
+        (0xA880, "V"),
+        (0xA8C6, "X"),
+        (0xA8CE, "V"),
+        (0xA8DA, "X"),
+        (0xA8E0, "V"),
+        (0xA954, "X"),
+    ]
+
+
+def _seg_38() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xA95F, "V"),
+        (0xA97D, "X"),
+        (0xA980, "V"),
+        (0xA9CE, "X"),
+        (0xA9CF, "V"),
+        (0xA9DA, "X"),
+        (0xA9DE, "V"),
+        (0xA9FF, "X"),
+        (0xAA00, "V"),
+        (0xAA37, "X"),
+        (0xAA40, "V"),
+        (0xAA4E, "X"),
+        (0xAA50, "V"),
+        (0xAA5A, "X"),
+        (0xAA5C, "V"),
+        (0xAAC3, "X"),
+        (0xAADB, "V"),
+        (0xAAF7, "X"),
+        (0xAB01, "V"),
+        (0xAB07, "X"),
+        (0xAB09, "V"),
+        (0xAB0F, "X"),
+        (0xAB11, "V"),
+        (0xAB17, "X"),
+        (0xAB20, "V"),
+        (0xAB27, "X"),
+        (0xAB28, "V"),
+        (0xAB2F, "X"),
+        (0xAB30, "V"),
+        (0xAB5C, "M", "ꜧ"),
+        (0xAB5D, "M", "ꬷ"),
+        (0xAB5E, "M", "ɫ"),
+        (0xAB5F, "M", "ꭒ"),
+        (0xAB60, "V"),
+        (0xAB69, "M", "ʍ"),
+        (0xAB6A, "V"),
+        (0xAB6C, "X"),
+        (0xAB70, "M", "Ꭰ"),
+        (0xAB71, "M", "Ꭱ"),
+        (0xAB72, "M", "Ꭲ"),
+        (0xAB73, "M", "Ꭳ"),
+        (0xAB74, "M", "Ꭴ"),
+        (0xAB75, "M", "Ꭵ"),
+        (0xAB76, "M", "Ꭶ"),
+        (0xAB77, "M", "Ꭷ"),
+        (0xAB78, "M", "Ꭸ"),
+        (0xAB79, "M", "Ꭹ"),
+        (0xAB7A, "M", "Ꭺ"),
+        (0xAB7B, "M", "Ꭻ"),
+        (0xAB7C, "M", "Ꭼ"),
+        (0xAB7D, "M", "Ꭽ"),
+        (0xAB7E, "M", "Ꭾ"),
+        (0xAB7F, "M", "Ꭿ"),
+        (0xAB80, "M", "Ꮀ"),
+        (0xAB81, "M", "Ꮁ"),
+        (0xAB82, "M", "Ꮂ"),
+        (0xAB83, "M", "Ꮃ"),
+        (0xAB84, "M", "Ꮄ"),
+        (0xAB85, "M", "Ꮅ"),
+        (0xAB86, "M", "Ꮆ"),
+        (0xAB87, "M", "Ꮇ"),
+        (0xAB88, "M", "Ꮈ"),
+        (0xAB89, "M", "Ꮉ"),
+        (0xAB8A, "M", "Ꮊ"),
+        (0xAB8B, "M", "Ꮋ"),
+        (0xAB8C, "M", "Ꮌ"),
+        (0xAB8D, "M", "Ꮍ"),
+        (0xAB8E, "M", "Ꮎ"),
+        (0xAB8F, "M", "Ꮏ"),
+        (0xAB90, "M", "Ꮐ"),
+        (0xAB91, "M", "Ꮑ"),
+        (0xAB92, "M", "Ꮒ"),
+        (0xAB93, "M", "Ꮓ"),
+        (0xAB94, "M", "Ꮔ"),
+        (0xAB95, "M", "Ꮕ"),
+        (0xAB96, "M", "Ꮖ"),
+        (0xAB97, "M", "Ꮗ"),
+        (0xAB98, "M", "Ꮘ"),
+        (0xAB99, "M", "Ꮙ"),
+        (0xAB9A, "M", "Ꮚ"),
+        (0xAB9B, "M", "Ꮛ"),
+        (0xAB9C, "M", "Ꮜ"),
+        (0xAB9D, "M", "Ꮝ"),
+        (0xAB9E, "M", "Ꮞ"),
+        (0xAB9F, "M", "Ꮟ"),
+        (0xABA0, "M", "Ꮠ"),
+        (0xABA1, "M", "Ꮡ"),
+        (0xABA2, "M", "Ꮢ"),
+        (0xABA3, "M", "Ꮣ"),
+        (0xABA4, "M", "Ꮤ"),
+        (0xABA5, "M", "Ꮥ"),
+        (0xABA6, "M", "Ꮦ"),
+        (0xABA7, "M", "Ꮧ"),
+        (0xABA8, "M", "Ꮨ"),
+        (0xABA9, "M", "Ꮩ"),
+        (0xABAA, "M", "Ꮪ"),
+        (0xABAB, "M", "Ꮫ"),
+        (0xABAC, "M", "Ꮬ"),
+        (0xABAD, "M", "Ꮭ"),
+        (0xABAE, "M", "Ꮮ"),
+    ]
+
+
+def _seg_39() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xABAF, "M", "Ꮯ"),
+        (0xABB0, "M", "Ꮰ"),
+        (0xABB1, "M", "Ꮱ"),
+        (0xABB2, "M", "Ꮲ"),
+        (0xABB3, "M", "Ꮳ"),
+        (0xABB4, "M", "Ꮴ"),
+        (0xABB5, "M", "Ꮵ"),
+        (0xABB6, "M", "Ꮶ"),
+        (0xABB7, "M", "Ꮷ"),
+        (0xABB8, "M", "Ꮸ"),
+        (0xABB9, "M", "Ꮹ"),
+        (0xABBA, "M", "Ꮺ"),
+        (0xABBB, "M", "Ꮻ"),
+        (0xABBC, "M", "Ꮼ"),
+        (0xABBD, "M", "Ꮽ"),
+        (0xABBE, "M", "Ꮾ"),
+        (0xABBF, "M", "Ꮿ"),
+        (0xABC0, "V"),
+        (0xABEE, "X"),
+        (0xABF0, "V"),
+        (0xABFA, "X"),
+        (0xAC00, "V"),
+        (0xD7A4, "X"),
+        (0xD7B0, "V"),
+        (0xD7C7, "X"),
+        (0xD7CB, "V"),
+        (0xD7FC, "X"),
+        (0xF900, "M", "豈"),
+        (0xF901, "M", "更"),
+        (0xF902, "M", "車"),
+        (0xF903, "M", "賈"),
+        (0xF904, "M", "滑"),
+        (0xF905, "M", "串"),
+        (0xF906, "M", "句"),
+        (0xF907, "M", "龜"),
+        (0xF909, "M", "契"),
+        (0xF90A, "M", "金"),
+        (0xF90B, "M", "喇"),
+        (0xF90C, "M", "奈"),
+        (0xF90D, "M", "懶"),
+        (0xF90E, "M", "癩"),
+        (0xF90F, "M", "羅"),
+        (0xF910, "M", "蘿"),
+        (0xF911, "M", "螺"),
+        (0xF912, "M", "裸"),
+        (0xF913, "M", "邏"),
+        (0xF914, "M", "樂"),
+        (0xF915, "M", "洛"),
+        (0xF916, "M", "烙"),
+        (0xF917, "M", "珞"),
+        (0xF918, "M", "落"),
+        (0xF919, "M", "酪"),
+        (0xF91A, "M", "駱"),
+        (0xF91B, "M", "亂"),
+        (0xF91C, "M", "卵"),
+        (0xF91D, "M", "欄"),
+        (0xF91E, "M", "爛"),
+        (0xF91F, "M", "蘭"),
+        (0xF920, "M", "鸞"),
+        (0xF921, "M", "嵐"),
+        (0xF922, "M", "濫"),
+        (0xF923, "M", "藍"),
+        (0xF924, "M", "襤"),
+        (0xF925, "M", "拉"),
+        (0xF926, "M", "臘"),
+        (0xF927, "M", "蠟"),
+        (0xF928, "M", "廊"),
+        (0xF929, "M", "朗"),
+        (0xF92A, "M", "浪"),
+        (0xF92B, "M", "狼"),
+        (0xF92C, "M", "郎"),
+        (0xF92D, "M", "來"),
+        (0xF92E, "M", "冷"),
+        (0xF92F, "M", "勞"),
+        (0xF930, "M", "擄"),
+        (0xF931, "M", "櫓"),
+        (0xF932, "M", "爐"),
+        (0xF933, "M", "盧"),
+        (0xF934, "M", "老"),
+        (0xF935, "M", "蘆"),
+        (0xF936, "M", "虜"),
+        (0xF937, "M", "路"),
+        (0xF938, "M", "露"),
+        (0xF939, "M", "魯"),
+        (0xF93A, "M", "鷺"),
+        (0xF93B, "M", "碌"),
+        (0xF93C, "M", "祿"),
+        (0xF93D, "M", "綠"),
+        (0xF93E, "M", "菉"),
+        (0xF93F, "M", "錄"),
+        (0xF940, "M", "鹿"),
+        (0xF941, "M", "論"),
+        (0xF942, "M", "壟"),
+        (0xF943, "M", "弄"),
+        (0xF944, "M", "籠"),
+        (0xF945, "M", "聾"),
+        (0xF946, "M", "牢"),
+        (0xF947, "M", "磊"),
+        (0xF948, "M", "賂"),
+        (0xF949, "M", "雷"),
+    ]
+
+
+def _seg_40() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xF94A, "M", "壘"),
+        (0xF94B, "M", "屢"),
+        (0xF94C, "M", "樓"),
+        (0xF94D, "M", "淚"),
+        (0xF94E, "M", "漏"),
+        (0xF94F, "M", "累"),
+        (0xF950, "M", "縷"),
+        (0xF951, "M", "陋"),
+        (0xF952, "M", "勒"),
+        (0xF953, "M", "肋"),
+        (0xF954, "M", "凜"),
+        (0xF955, "M", "凌"),
+        (0xF956, "M", "稜"),
+        (0xF957, "M", "綾"),
+        (0xF958, "M", "菱"),
+        (0xF959, "M", "陵"),
+        (0xF95A, "M", "讀"),
+        (0xF95B, "M", "拏"),
+        (0xF95C, "M", "樂"),
+        (0xF95D, "M", "諾"),
+        (0xF95E, "M", "丹"),
+        (0xF95F, "M", "寧"),
+        (0xF960, "M", "怒"),
+        (0xF961, "M", "率"),
+        (0xF962, "M", "異"),
+        (0xF963, "M", "北"),
+        (0xF964, "M", "磻"),
+        (0xF965, "M", "便"),
+        (0xF966, "M", "復"),
+        (0xF967, "M", "不"),
+        (0xF968, "M", "泌"),
+        (0xF969, "M", "數"),
+        (0xF96A, "M", "索"),
+        (0xF96B, "M", "參"),
+        (0xF96C, "M", "塞"),
+        (0xF96D, "M", "省"),
+        (0xF96E, "M", "葉"),
+        (0xF96F, "M", "說"),
+        (0xF970, "M", "殺"),
+        (0xF971, "M", "辰"),
+        (0xF972, "M", "沈"),
+        (0xF973, "M", "拾"),
+        (0xF974, "M", "若"),
+        (0xF975, "M", "掠"),
+        (0xF976, "M", "略"),
+        (0xF977, "M", "亮"),
+        (0xF978, "M", "兩"),
+        (0xF979, "M", "凉"),
+        (0xF97A, "M", "梁"),
+        (0xF97B, "M", "糧"),
+        (0xF97C, "M", "良"),
+        (0xF97D, "M", "諒"),
+        (0xF97E, "M", "量"),
+        (0xF97F, "M", "勵"),
+        (0xF980, "M", "呂"),
+        (0xF981, "M", "女"),
+        (0xF982, "M", "廬"),
+        (0xF983, "M", "旅"),
+        (0xF984, "M", "濾"),
+        (0xF985, "M", "礪"),
+        (0xF986, "M", "閭"),
+        (0xF987, "M", "驪"),
+        (0xF988, "M", "麗"),
+        (0xF989, "M", "黎"),
+        (0xF98A, "M", "力"),
+        (0xF98B, "M", "曆"),
+        (0xF98C, "M", "歷"),
+        (0xF98D, "M", "轢"),
+        (0xF98E, "M", "年"),
+        (0xF98F, "M", "憐"),
+        (0xF990, "M", "戀"),
+        (0xF991, "M", "撚"),
+        (0xF992, "M", "漣"),
+        (0xF993, "M", "煉"),
+        (0xF994, "M", "璉"),
+        (0xF995, "M", "秊"),
+        (0xF996, "M", "練"),
+        (0xF997, "M", "聯"),
+        (0xF998, "M", "輦"),
+        (0xF999, "M", "蓮"),
+        (0xF99A, "M", "連"),
+        (0xF99B, "M", "鍊"),
+        (0xF99C, "M", "列"),
+        (0xF99D, "M", "劣"),
+        (0xF99E, "M", "咽"),
+        (0xF99F, "M", "烈"),
+        (0xF9A0, "M", "裂"),
+        (0xF9A1, "M", "說"),
+        (0xF9A2, "M", "廉"),
+        (0xF9A3, "M", "念"),
+        (0xF9A4, "M", "捻"),
+        (0xF9A5, "M", "殮"),
+        (0xF9A6, "M", "簾"),
+        (0xF9A7, "M", "獵"),
+        (0xF9A8, "M", "令"),
+        (0xF9A9, "M", "囹"),
+        (0xF9AA, "M", "寧"),
+        (0xF9AB, "M", "嶺"),
+        (0xF9AC, "M", "怜"),
+        (0xF9AD, "M", "玲"),
+    ]
+
+
+def _seg_41() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xF9AE, "M", "瑩"),
+        (0xF9AF, "M", "羚"),
+        (0xF9B0, "M", "聆"),
+        (0xF9B1, "M", "鈴"),
+        (0xF9B2, "M", "零"),
+        (0xF9B3, "M", "靈"),
+        (0xF9B4, "M", "領"),
+        (0xF9B5, "M", "例"),
+        (0xF9B6, "M", "禮"),
+        (0xF9B7, "M", "醴"),
+        (0xF9B8, "M", "隸"),
+        (0xF9B9, "M", "惡"),
+        (0xF9BA, "M", "了"),
+        (0xF9BB, "M", "僚"),
+        (0xF9BC, "M", "寮"),
+        (0xF9BD, "M", "尿"),
+        (0xF9BE, "M", "料"),
+        (0xF9BF, "M", "樂"),
+        (0xF9C0, "M", "燎"),
+        (0xF9C1, "M", "療"),
+        (0xF9C2, "M", "蓼"),
+        (0xF9C3, "M", "遼"),
+        (0xF9C4, "M", "龍"),
+        (0xF9C5, "M", "暈"),
+        (0xF9C6, "M", "阮"),
+        (0xF9C7, "M", "劉"),
+        (0xF9C8, "M", "杻"),
+        (0xF9C9, "M", "柳"),
+        (0xF9CA, "M", "流"),
+        (0xF9CB, "M", "溜"),
+        (0xF9CC, "M", "琉"),
+        (0xF9CD, "M", "留"),
+        (0xF9CE, "M", "硫"),
+        (0xF9CF, "M", "紐"),
+        (0xF9D0, "M", "類"),
+        (0xF9D1, "M", "六"),
+        (0xF9D2, "M", "戮"),
+        (0xF9D3, "M", "陸"),
+        (0xF9D4, "M", "倫"),
+        (0xF9D5, "M", "崙"),
+        (0xF9D6, "M", "淪"),
+        (0xF9D7, "M", "輪"),
+        (0xF9D8, "M", "律"),
+        (0xF9D9, "M", "慄"),
+        (0xF9DA, "M", "栗"),
+        (0xF9DB, "M", "率"),
+        (0xF9DC, "M", "隆"),
+        (0xF9DD, "M", "利"),
+        (0xF9DE, "M", "吏"),
+        (0xF9DF, "M", "履"),
+        (0xF9E0, "M", "易"),
+        (0xF9E1, "M", "李"),
+        (0xF9E2, "M", "梨"),
+        (0xF9E3, "M", "泥"),
+        (0xF9E4, "M", "理"),
+        (0xF9E5, "M", "痢"),
+        (0xF9E6, "M", "罹"),
+        (0xF9E7, "M", "裏"),
+        (0xF9E8, "M", "裡"),
+        (0xF9E9, "M", "里"),
+        (0xF9EA, "M", "離"),
+        (0xF9EB, "M", "匿"),
+        (0xF9EC, "M", "溺"),
+        (0xF9ED, "M", "吝"),
+        (0xF9EE, "M", "燐"),
+        (0xF9EF, "M", "璘"),
+        (0xF9F0, "M", "藺"),
+        (0xF9F1, "M", "隣"),
+        (0xF9F2, "M", "鱗"),
+        (0xF9F3, "M", "麟"),
+        (0xF9F4, "M", "林"),
+        (0xF9F5, "M", "淋"),
+        (0xF9F6, "M", "臨"),
+        (0xF9F7, "M", "立"),
+        (0xF9F8, "M", "笠"),
+        (0xF9F9, "M", "粒"),
+        (0xF9FA, "M", "狀"),
+        (0xF9FB, "M", "炙"),
+        (0xF9FC, "M", "識"),
+        (0xF9FD, "M", "什"),
+        (0xF9FE, "M", "茶"),
+        (0xF9FF, "M", "刺"),
+        (0xFA00, "M", "切"),
+        (0xFA01, "M", "度"),
+        (0xFA02, "M", "拓"),
+        (0xFA03, "M", "糖"),
+        (0xFA04, "M", "宅"),
+        (0xFA05, "M", "洞"),
+        (0xFA06, "M", "暴"),
+        (0xFA07, "M", "輻"),
+        (0xFA08, "M", "行"),
+        (0xFA09, "M", "降"),
+        (0xFA0A, "M", "見"),
+        (0xFA0B, "M", "廓"),
+        (0xFA0C, "M", "兀"),
+        (0xFA0D, "M", "嗀"),
+        (0xFA0E, "V"),
+        (0xFA10, "M", "塚"),
+        (0xFA11, "V"),
+        (0xFA12, "M", "晴"),
+    ]
+
+
+def _seg_42() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFA13, "V"),
+        (0xFA15, "M", "凞"),
+        (0xFA16, "M", "猪"),
+        (0xFA17, "M", "益"),
+        (0xFA18, "M", "礼"),
+        (0xFA19, "M", "神"),
+        (0xFA1A, "M", "祥"),
+        (0xFA1B, "M", "福"),
+        (0xFA1C, "M", "靖"),
+        (0xFA1D, "M", "精"),
+        (0xFA1E, "M", "羽"),
+        (0xFA1F, "V"),
+        (0xFA20, "M", "蘒"),
+        (0xFA21, "V"),
+        (0xFA22, "M", "諸"),
+        (0xFA23, "V"),
+        (0xFA25, "M", "逸"),
+        (0xFA26, "M", "都"),
+        (0xFA27, "V"),
+        (0xFA2A, "M", "飯"),
+        (0xFA2B, "M", "飼"),
+        (0xFA2C, "M", "館"),
+        (0xFA2D, "M", "鶴"),
+        (0xFA2E, "M", "郞"),
+        (0xFA2F, "M", "隷"),
+        (0xFA30, "M", "侮"),
+        (0xFA31, "M", "僧"),
+        (0xFA32, "M", "免"),
+        (0xFA33, "M", "勉"),
+        (0xFA34, "M", "勤"),
+        (0xFA35, "M", "卑"),
+        (0xFA36, "M", "喝"),
+        (0xFA37, "M", "嘆"),
+        (0xFA38, "M", "器"),
+        (0xFA39, "M", "塀"),
+        (0xFA3A, "M", "墨"),
+        (0xFA3B, "M", "層"),
+        (0xFA3C, "M", "屮"),
+        (0xFA3D, "M", "悔"),
+        (0xFA3E, "M", "慨"),
+        (0xFA3F, "M", "憎"),
+        (0xFA40, "M", "懲"),
+        (0xFA41, "M", "敏"),
+        (0xFA42, "M", "既"),
+        (0xFA43, "M", "暑"),
+        (0xFA44, "M", "梅"),
+        (0xFA45, "M", "海"),
+        (0xFA46, "M", "渚"),
+        (0xFA47, "M", "漢"),
+        (0xFA48, "M", "煮"),
+        (0xFA49, "M", "爫"),
+        (0xFA4A, "M", "琢"),
+        (0xFA4B, "M", "碑"),
+        (0xFA4C, "M", "社"),
+        (0xFA4D, "M", "祉"),
+        (0xFA4E, "M", "祈"),
+        (0xFA4F, "M", "祐"),
+        (0xFA50, "M", "祖"),
+        (0xFA51, "M", "祝"),
+        (0xFA52, "M", "禍"),
+        (0xFA53, "M", "禎"),
+        (0xFA54, "M", "穀"),
+        (0xFA55, "M", "突"),
+        (0xFA56, "M", "節"),
+        (0xFA57, "M", "練"),
+        (0xFA58, "M", "縉"),
+        (0xFA59, "M", "繁"),
+        (0xFA5A, "M", "署"),
+        (0xFA5B, "M", "者"),
+        (0xFA5C, "M", "臭"),
+        (0xFA5D, "M", "艹"),
+        (0xFA5F, "M", "著"),
+        (0xFA60, "M", "褐"),
+        (0xFA61, "M", "視"),
+        (0xFA62, "M", "謁"),
+        (0xFA63, "M", "謹"),
+        (0xFA64, "M", "賓"),
+        (0xFA65, "M", "贈"),
+        (0xFA66, "M", "辶"),
+        (0xFA67, "M", "逸"),
+        (0xFA68, "M", "難"),
+        (0xFA69, "M", "響"),
+        (0xFA6A, "M", "頻"),
+        (0xFA6B, "M", "恵"),
+        (0xFA6C, "M", "𤋮"),
+        (0xFA6D, "M", "舘"),
+        (0xFA6E, "X"),
+        (0xFA70, "M", "並"),
+        (0xFA71, "M", "况"),
+        (0xFA72, "M", "全"),
+        (0xFA73, "M", "侀"),
+        (0xFA74, "M", "充"),
+        (0xFA75, "M", "冀"),
+        (0xFA76, "M", "勇"),
+        (0xFA77, "M", "勺"),
+        (0xFA78, "M", "喝"),
+        (0xFA79, "M", "啕"),
+        (0xFA7A, "M", "喙"),
+        (0xFA7B, "M", "嗢"),
+        (0xFA7C, "M", "塚"),
+    ]
+
+
+def _seg_43() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFA7D, "M", "墳"),
+        (0xFA7E, "M", "奄"),
+        (0xFA7F, "M", "奔"),
+        (0xFA80, "M", "婢"),
+        (0xFA81, "M", "嬨"),
+        (0xFA82, "M", "廒"),
+        (0xFA83, "M", "廙"),
+        (0xFA84, "M", "彩"),
+        (0xFA85, "M", "徭"),
+        (0xFA86, "M", "惘"),
+        (0xFA87, "M", "慎"),
+        (0xFA88, "M", "愈"),
+        (0xFA89, "M", "憎"),
+        (0xFA8A, "M", "慠"),
+        (0xFA8B, "M", "懲"),
+        (0xFA8C, "M", "戴"),
+        (0xFA8D, "M", "揄"),
+        (0xFA8E, "M", "搜"),
+        (0xFA8F, "M", "摒"),
+        (0xFA90, "M", "敖"),
+        (0xFA91, "M", "晴"),
+        (0xFA92, "M", "朗"),
+        (0xFA93, "M", "望"),
+        (0xFA94, "M", "杖"),
+        (0xFA95, "M", "歹"),
+        (0xFA96, "M", "殺"),
+        (0xFA97, "M", "流"),
+        (0xFA98, "M", "滛"),
+        (0xFA99, "M", "滋"),
+        (0xFA9A, "M", "漢"),
+        (0xFA9B, "M", "瀞"),
+        (0xFA9C, "M", "煮"),
+        (0xFA9D, "M", "瞧"),
+        (0xFA9E, "M", "爵"),
+        (0xFA9F, "M", "犯"),
+        (0xFAA0, "M", "猪"),
+        (0xFAA1, "M", "瑱"),
+        (0xFAA2, "M", "甆"),
+        (0xFAA3, "M", "画"),
+        (0xFAA4, "M", "瘝"),
+        (0xFAA5, "M", "瘟"),
+        (0xFAA6, "M", "益"),
+        (0xFAA7, "M", "盛"),
+        (0xFAA8, "M", "直"),
+        (0xFAA9, "M", "睊"),
+        (0xFAAA, "M", "着"),
+        (0xFAAB, "M", "磌"),
+        (0xFAAC, "M", "窱"),
+        (0xFAAD, "M", "節"),
+        (0xFAAE, "M", "类"),
+        (0xFAAF, "M", "絛"),
+        (0xFAB0, "M", "練"),
+        (0xFAB1, "M", "缾"),
+        (0xFAB2, "M", "者"),
+        (0xFAB3, "M", "荒"),
+        (0xFAB4, "M", "華"),
+        (0xFAB5, "M", "蝹"),
+        (0xFAB6, "M", "襁"),
+        (0xFAB7, "M", "覆"),
+        (0xFAB8, "M", "視"),
+        (0xFAB9, "M", "調"),
+        (0xFABA, "M", "諸"),
+        (0xFABB, "M", "請"),
+        (0xFABC, "M", "謁"),
+        (0xFABD, "M", "諾"),
+        (0xFABE, "M", "諭"),
+        (0xFABF, "M", "謹"),
+        (0xFAC0, "M", "變"),
+        (0xFAC1, "M", "贈"),
+        (0xFAC2, "M", "輸"),
+        (0xFAC3, "M", "遲"),
+        (0xFAC4, "M", "醙"),
+        (0xFAC5, "M", "鉶"),
+        (0xFAC6, "M", "陼"),
+        (0xFAC7, "M", "難"),
+        (0xFAC8, "M", "靖"),
+        (0xFAC9, "M", "韛"),
+        (0xFACA, "M", "響"),
+        (0xFACB, "M", "頋"),
+        (0xFACC, "M", "頻"),
+        (0xFACD, "M", "鬒"),
+        (0xFACE, "M", "龜"),
+        (0xFACF, "M", "𢡊"),
+        (0xFAD0, "M", "𢡄"),
+        (0xFAD1, "M", "𣏕"),
+        (0xFAD2, "M", "㮝"),
+        (0xFAD3, "M", "䀘"),
+        (0xFAD4, "M", "䀹"),
+        (0xFAD5, "M", "𥉉"),
+        (0xFAD6, "M", "𥳐"),
+        (0xFAD7, "M", "𧻓"),
+        (0xFAD8, "M", "齃"),
+        (0xFAD9, "M", "龎"),
+        (0xFADA, "X"),
+        (0xFB00, "M", "ff"),
+        (0xFB01, "M", "fi"),
+        (0xFB02, "M", "fl"),
+        (0xFB03, "M", "ffi"),
+        (0xFB04, "M", "ffl"),
+        (0xFB05, "M", "st"),
+    ]
+
+
+def _seg_44() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFB07, "X"),
+        (0xFB13, "M", "մն"),
+        (0xFB14, "M", "մե"),
+        (0xFB15, "M", "մի"),
+        (0xFB16, "M", "վն"),
+        (0xFB17, "M", "մխ"),
+        (0xFB18, "X"),
+        (0xFB1D, "M", "יִ"),
+        (0xFB1E, "V"),
+        (0xFB1F, "M", "ײַ"),
+        (0xFB20, "M", "ע"),
+        (0xFB21, "M", "א"),
+        (0xFB22, "M", "ד"),
+        (0xFB23, "M", "ה"),
+        (0xFB24, "M", "כ"),
+        (0xFB25, "M", "ל"),
+        (0xFB26, "M", "ם"),
+        (0xFB27, "M", "ר"),
+        (0xFB28, "M", "ת"),
+        (0xFB29, "3", "+"),
+        (0xFB2A, "M", "שׁ"),
+        (0xFB2B, "M", "שׂ"),
+        (0xFB2C, "M", "שּׁ"),
+        (0xFB2D, "M", "שּׂ"),
+        (0xFB2E, "M", "אַ"),
+        (0xFB2F, "M", "אָ"),
+        (0xFB30, "M", "אּ"),
+        (0xFB31, "M", "בּ"),
+        (0xFB32, "M", "גּ"),
+        (0xFB33, "M", "דּ"),
+        (0xFB34, "M", "הּ"),
+        (0xFB35, "M", "וּ"),
+        (0xFB36, "M", "זּ"),
+        (0xFB37, "X"),
+        (0xFB38, "M", "טּ"),
+        (0xFB39, "M", "יּ"),
+        (0xFB3A, "M", "ךּ"),
+        (0xFB3B, "M", "כּ"),
+        (0xFB3C, "M", "לּ"),
+        (0xFB3D, "X"),
+        (0xFB3E, "M", "מּ"),
+        (0xFB3F, "X"),
+        (0xFB40, "M", "נּ"),
+        (0xFB41, "M", "סּ"),
+        (0xFB42, "X"),
+        (0xFB43, "M", "ףּ"),
+        (0xFB44, "M", "פּ"),
+        (0xFB45, "X"),
+        (0xFB46, "M", "צּ"),
+        (0xFB47, "M", "קּ"),
+        (0xFB48, "M", "רּ"),
+        (0xFB49, "M", "שּ"),
+        (0xFB4A, "M", "תּ"),
+        (0xFB4B, "M", "וֹ"),
+        (0xFB4C, "M", "בֿ"),
+        (0xFB4D, "M", "כֿ"),
+        (0xFB4E, "M", "פֿ"),
+        (0xFB4F, "M", "אל"),
+        (0xFB50, "M", "ٱ"),
+        (0xFB52, "M", "ٻ"),
+        (0xFB56, "M", "پ"),
+        (0xFB5A, "M", "ڀ"),
+        (0xFB5E, "M", "ٺ"),
+        (0xFB62, "M", "ٿ"),
+        (0xFB66, "M", "ٹ"),
+        (0xFB6A, "M", "ڤ"),
+        (0xFB6E, "M", "ڦ"),
+        (0xFB72, "M", "ڄ"),
+        (0xFB76, "M", "ڃ"),
+        (0xFB7A, "M", "چ"),
+        (0xFB7E, "M", "ڇ"),
+        (0xFB82, "M", "ڍ"),
+        (0xFB84, "M", "ڌ"),
+        (0xFB86, "M", "ڎ"),
+        (0xFB88, "M", "ڈ"),
+        (0xFB8A, "M", "ژ"),
+        (0xFB8C, "M", "ڑ"),
+        (0xFB8E, "M", "ک"),
+        (0xFB92, "M", "گ"),
+        (0xFB96, "M", "ڳ"),
+        (0xFB9A, "M", "ڱ"),
+        (0xFB9E, "M", "ں"),
+        (0xFBA0, "M", "ڻ"),
+        (0xFBA4, "M", "ۀ"),
+        (0xFBA6, "M", "ہ"),
+        (0xFBAA, "M", "ھ"),
+        (0xFBAE, "M", "ے"),
+        (0xFBB0, "M", "ۓ"),
+        (0xFBB2, "V"),
+        (0xFBC3, "X"),
+        (0xFBD3, "M", "ڭ"),
+        (0xFBD7, "M", "ۇ"),
+        (0xFBD9, "M", "ۆ"),
+        (0xFBDB, "M", "ۈ"),
+        (0xFBDD, "M", "ۇٴ"),
+        (0xFBDE, "M", "ۋ"),
+        (0xFBE0, "M", "ۅ"),
+        (0xFBE2, "M", "ۉ"),
+        (0xFBE4, "M", "ې"),
+        (0xFBE8, "M", "ى"),
+    ]
+
+
+def _seg_45() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFBEA, "M", "ئا"),
+        (0xFBEC, "M", "ئە"),
+        (0xFBEE, "M", "ئو"),
+        (0xFBF0, "M", "ئۇ"),
+        (0xFBF2, "M", "ئۆ"),
+        (0xFBF4, "M", "ئۈ"),
+        (0xFBF6, "M", "ئې"),
+        (0xFBF9, "M", "ئى"),
+        (0xFBFC, "M", "ی"),
+        (0xFC00, "M", "ئج"),
+        (0xFC01, "M", "ئح"),
+        (0xFC02, "M", "ئم"),
+        (0xFC03, "M", "ئى"),
+        (0xFC04, "M", "ئي"),
+        (0xFC05, "M", "بج"),
+        (0xFC06, "M", "بح"),
+        (0xFC07, "M", "بخ"),
+        (0xFC08, "M", "بم"),
+        (0xFC09, "M", "بى"),
+        (0xFC0A, "M", "بي"),
+        (0xFC0B, "M", "تج"),
+        (0xFC0C, "M", "تح"),
+        (0xFC0D, "M", "تخ"),
+        (0xFC0E, "M", "تم"),
+        (0xFC0F, "M", "تى"),
+        (0xFC10, "M", "تي"),
+        (0xFC11, "M", "ثج"),
+        (0xFC12, "M", "ثم"),
+        (0xFC13, "M", "ثى"),
+        (0xFC14, "M", "ثي"),
+        (0xFC15, "M", "جح"),
+        (0xFC16, "M", "جم"),
+        (0xFC17, "M", "حج"),
+        (0xFC18, "M", "حم"),
+        (0xFC19, "M", "خج"),
+        (0xFC1A, "M", "خح"),
+        (0xFC1B, "M", "خم"),
+        (0xFC1C, "M", "سج"),
+        (0xFC1D, "M", "سح"),
+        (0xFC1E, "M", "سخ"),
+        (0xFC1F, "M", "سم"),
+        (0xFC20, "M", "صح"),
+        (0xFC21, "M", "صم"),
+        (0xFC22, "M", "ضج"),
+        (0xFC23, "M", "ضح"),
+        (0xFC24, "M", "ضخ"),
+        (0xFC25, "M", "ضم"),
+        (0xFC26, "M", "طح"),
+        (0xFC27, "M", "طم"),
+        (0xFC28, "M", "ظم"),
+        (0xFC29, "M", "عج"),
+        (0xFC2A, "M", "عم"),
+        (0xFC2B, "M", "غج"),
+        (0xFC2C, "M", "غم"),
+        (0xFC2D, "M", "فج"),
+        (0xFC2E, "M", "فح"),
+        (0xFC2F, "M", "فخ"),
+        (0xFC30, "M", "فم"),
+        (0xFC31, "M", "فى"),
+        (0xFC32, "M", "في"),
+        (0xFC33, "M", "قح"),
+        (0xFC34, "M", "قم"),
+        (0xFC35, "M", "قى"),
+        (0xFC36, "M", "قي"),
+        (0xFC37, "M", "كا"),
+        (0xFC38, "M", "كج"),
+        (0xFC39, "M", "كح"),
+        (0xFC3A, "M", "كخ"),
+        (0xFC3B, "M", "كل"),
+        (0xFC3C, "M", "كم"),
+        (0xFC3D, "M", "كى"),
+        (0xFC3E, "M", "كي"),
+        (0xFC3F, "M", "لج"),
+        (0xFC40, "M", "لح"),
+        (0xFC41, "M", "لخ"),
+        (0xFC42, "M", "لم"),
+        (0xFC43, "M", "لى"),
+        (0xFC44, "M", "لي"),
+        (0xFC45, "M", "مج"),
+        (0xFC46, "M", "مح"),
+        (0xFC47, "M", "مخ"),
+        (0xFC48, "M", "مم"),
+        (0xFC49, "M", "مى"),
+        (0xFC4A, "M", "مي"),
+        (0xFC4B, "M", "نج"),
+        (0xFC4C, "M", "نح"),
+        (0xFC4D, "M", "نخ"),
+        (0xFC4E, "M", "نم"),
+        (0xFC4F, "M", "نى"),
+        (0xFC50, "M", "ني"),
+        (0xFC51, "M", "هج"),
+        (0xFC52, "M", "هم"),
+        (0xFC53, "M", "هى"),
+        (0xFC54, "M", "هي"),
+        (0xFC55, "M", "يج"),
+        (0xFC56, "M", "يح"),
+        (0xFC57, "M", "يخ"),
+        (0xFC58, "M", "يم"),
+        (0xFC59, "M", "يى"),
+        (0xFC5A, "M", "يي"),
+    ]
+
+
+def _seg_46() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFC5B, "M", "ذٰ"),
+        (0xFC5C, "M", "رٰ"),
+        (0xFC5D, "M", "ىٰ"),
+        (0xFC5E, "3", " ٌّ"),
+        (0xFC5F, "3", " ٍّ"),
+        (0xFC60, "3", " َّ"),
+        (0xFC61, "3", " ُّ"),
+        (0xFC62, "3", " ِّ"),
+        (0xFC63, "3", " ّٰ"),
+        (0xFC64, "M", "ئر"),
+        (0xFC65, "M", "ئز"),
+        (0xFC66, "M", "ئم"),
+        (0xFC67, "M", "ئن"),
+        (0xFC68, "M", "ئى"),
+        (0xFC69, "M", "ئي"),
+        (0xFC6A, "M", "بر"),
+        (0xFC6B, "M", "بز"),
+        (0xFC6C, "M", "بم"),
+        (0xFC6D, "M", "بن"),
+        (0xFC6E, "M", "بى"),
+        (0xFC6F, "M", "بي"),
+        (0xFC70, "M", "تر"),
+        (0xFC71, "M", "تز"),
+        (0xFC72, "M", "تم"),
+        (0xFC73, "M", "تن"),
+        (0xFC74, "M", "تى"),
+        (0xFC75, "M", "تي"),
+        (0xFC76, "M", "ثر"),
+        (0xFC77, "M", "ثز"),
+        (0xFC78, "M", "ثم"),
+        (0xFC79, "M", "ثن"),
+        (0xFC7A, "M", "ثى"),
+        (0xFC7B, "M", "ثي"),
+        (0xFC7C, "M", "فى"),
+        (0xFC7D, "M", "في"),
+        (0xFC7E, "M", "قى"),
+        (0xFC7F, "M", "قي"),
+        (0xFC80, "M", "كا"),
+        (0xFC81, "M", "كل"),
+        (0xFC82, "M", "كم"),
+        (0xFC83, "M", "كى"),
+        (0xFC84, "M", "كي"),
+        (0xFC85, "M", "لم"),
+        (0xFC86, "M", "لى"),
+        (0xFC87, "M", "لي"),
+        (0xFC88, "M", "ما"),
+        (0xFC89, "M", "مم"),
+        (0xFC8A, "M", "نر"),
+        (0xFC8B, "M", "نز"),
+        (0xFC8C, "M", "نم"),
+        (0xFC8D, "M", "نن"),
+        (0xFC8E, "M", "نى"),
+        (0xFC8F, "M", "ني"),
+        (0xFC90, "M", "ىٰ"),
+        (0xFC91, "M", "ير"),
+        (0xFC92, "M", "يز"),
+        (0xFC93, "M", "يم"),
+        (0xFC94, "M", "ين"),
+        (0xFC95, "M", "يى"),
+        (0xFC96, "M", "يي"),
+        (0xFC97, "M", "ئج"),
+        (0xFC98, "M", "ئح"),
+        (0xFC99, "M", "ئخ"),
+        (0xFC9A, "M", "ئم"),
+        (0xFC9B, "M", "ئه"),
+        (0xFC9C, "M", "بج"),
+        (0xFC9D, "M", "بح"),
+        (0xFC9E, "M", "بخ"),
+        (0xFC9F, "M", "بم"),
+        (0xFCA0, "M", "به"),
+        (0xFCA1, "M", "تج"),
+        (0xFCA2, "M", "تح"),
+        (0xFCA3, "M", "تخ"),
+        (0xFCA4, "M", "تم"),
+        (0xFCA5, "M", "ته"),
+        (0xFCA6, "M", "ثم"),
+        (0xFCA7, "M", "جح"),
+        (0xFCA8, "M", "جم"),
+        (0xFCA9, "M", "حج"),
+        (0xFCAA, "M", "حم"),
+        (0xFCAB, "M", "خج"),
+        (0xFCAC, "M", "خم"),
+        (0xFCAD, "M", "سج"),
+        (0xFCAE, "M", "سح"),
+        (0xFCAF, "M", "سخ"),
+        (0xFCB0, "M", "سم"),
+        (0xFCB1, "M", "صح"),
+        (0xFCB2, "M", "صخ"),
+        (0xFCB3, "M", "صم"),
+        (0xFCB4, "M", "ضج"),
+        (0xFCB5, "M", "ضح"),
+        (0xFCB6, "M", "ضخ"),
+        (0xFCB7, "M", "ضم"),
+        (0xFCB8, "M", "طح"),
+        (0xFCB9, "M", "ظم"),
+        (0xFCBA, "M", "عج"),
+        (0xFCBB, "M", "عم"),
+        (0xFCBC, "M", "غج"),
+        (0xFCBD, "M", "غم"),
+        (0xFCBE, "M", "فج"),
+    ]
+
+
+def _seg_47() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFCBF, "M", "فح"),
+        (0xFCC0, "M", "فخ"),
+        (0xFCC1, "M", "فم"),
+        (0xFCC2, "M", "قح"),
+        (0xFCC3, "M", "قم"),
+        (0xFCC4, "M", "كج"),
+        (0xFCC5, "M", "كح"),
+        (0xFCC6, "M", "كخ"),
+        (0xFCC7, "M", "كل"),
+        (0xFCC8, "M", "كم"),
+        (0xFCC9, "M", "لج"),
+        (0xFCCA, "M", "لح"),
+        (0xFCCB, "M", "لخ"),
+        (0xFCCC, "M", "لم"),
+        (0xFCCD, "M", "له"),
+        (0xFCCE, "M", "مج"),
+        (0xFCCF, "M", "مح"),
+        (0xFCD0, "M", "مخ"),
+        (0xFCD1, "M", "مم"),
+        (0xFCD2, "M", "نج"),
+        (0xFCD3, "M", "نح"),
+        (0xFCD4, "M", "نخ"),
+        (0xFCD5, "M", "نم"),
+        (0xFCD6, "M", "نه"),
+        (0xFCD7, "M", "هج"),
+        (0xFCD8, "M", "هم"),
+        (0xFCD9, "M", "هٰ"),
+        (0xFCDA, "M", "يج"),
+        (0xFCDB, "M", "يح"),
+        (0xFCDC, "M", "يخ"),
+        (0xFCDD, "M", "يم"),
+        (0xFCDE, "M", "يه"),
+        (0xFCDF, "M", "ئم"),
+        (0xFCE0, "M", "ئه"),
+        (0xFCE1, "M", "بم"),
+        (0xFCE2, "M", "به"),
+        (0xFCE3, "M", "تم"),
+        (0xFCE4, "M", "ته"),
+        (0xFCE5, "M", "ثم"),
+        (0xFCE6, "M", "ثه"),
+        (0xFCE7, "M", "سم"),
+        (0xFCE8, "M", "سه"),
+        (0xFCE9, "M", "شم"),
+        (0xFCEA, "M", "شه"),
+        (0xFCEB, "M", "كل"),
+        (0xFCEC, "M", "كم"),
+        (0xFCED, "M", "لم"),
+        (0xFCEE, "M", "نم"),
+        (0xFCEF, "M", "نه"),
+        (0xFCF0, "M", "يم"),
+        (0xFCF1, "M", "يه"),
+        (0xFCF2, "M", "ـَّ"),
+        (0xFCF3, "M", "ـُّ"),
+        (0xFCF4, "M", "ـِّ"),
+        (0xFCF5, "M", "طى"),
+        (0xFCF6, "M", "طي"),
+        (0xFCF7, "M", "عى"),
+        (0xFCF8, "M", "عي"),
+        (0xFCF9, "M", "غى"),
+        (0xFCFA, "M", "غي"),
+        (0xFCFB, "M", "سى"),
+        (0xFCFC, "M", "سي"),
+        (0xFCFD, "M", "شى"),
+        (0xFCFE, "M", "شي"),
+        (0xFCFF, "M", "حى"),
+        (0xFD00, "M", "حي"),
+        (0xFD01, "M", "جى"),
+        (0xFD02, "M", "جي"),
+        (0xFD03, "M", "خى"),
+        (0xFD04, "M", "خي"),
+        (0xFD05, "M", "صى"),
+        (0xFD06, "M", "صي"),
+        (0xFD07, "M", "ضى"),
+        (0xFD08, "M", "ضي"),
+        (0xFD09, "M", "شج"),
+        (0xFD0A, "M", "شح"),
+        (0xFD0B, "M", "شخ"),
+        (0xFD0C, "M", "شم"),
+        (0xFD0D, "M", "شر"),
+        (0xFD0E, "M", "سر"),
+        (0xFD0F, "M", "صر"),
+        (0xFD10, "M", "ضر"),
+        (0xFD11, "M", "طى"),
+        (0xFD12, "M", "طي"),
+        (0xFD13, "M", "عى"),
+        (0xFD14, "M", "عي"),
+        (0xFD15, "M", "غى"),
+        (0xFD16, "M", "غي"),
+        (0xFD17, "M", "سى"),
+        (0xFD18, "M", "سي"),
+        (0xFD19, "M", "شى"),
+        (0xFD1A, "M", "شي"),
+        (0xFD1B, "M", "حى"),
+        (0xFD1C, "M", "حي"),
+        (0xFD1D, "M", "جى"),
+        (0xFD1E, "M", "جي"),
+        (0xFD1F, "M", "خى"),
+        (0xFD20, "M", "خي"),
+        (0xFD21, "M", "صى"),
+        (0xFD22, "M", "صي"),
+    ]
+
+
+def _seg_48() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFD23, "M", "ضى"),
+        (0xFD24, "M", "ضي"),
+        (0xFD25, "M", "شج"),
+        (0xFD26, "M", "شح"),
+        (0xFD27, "M", "شخ"),
+        (0xFD28, "M", "شم"),
+        (0xFD29, "M", "شر"),
+        (0xFD2A, "M", "سر"),
+        (0xFD2B, "M", "صر"),
+        (0xFD2C, "M", "ضر"),
+        (0xFD2D, "M", "شج"),
+        (0xFD2E, "M", "شح"),
+        (0xFD2F, "M", "شخ"),
+        (0xFD30, "M", "شم"),
+        (0xFD31, "M", "سه"),
+        (0xFD32, "M", "شه"),
+        (0xFD33, "M", "طم"),
+        (0xFD34, "M", "سج"),
+        (0xFD35, "M", "سح"),
+        (0xFD36, "M", "سخ"),
+        (0xFD37, "M", "شج"),
+        (0xFD38, "M", "شح"),
+        (0xFD39, "M", "شخ"),
+        (0xFD3A, "M", "طم"),
+        (0xFD3B, "M", "ظم"),
+        (0xFD3C, "M", "اً"),
+        (0xFD3E, "V"),
+        (0xFD50, "M", "تجم"),
+        (0xFD51, "M", "تحج"),
+        (0xFD53, "M", "تحم"),
+        (0xFD54, "M", "تخم"),
+        (0xFD55, "M", "تمج"),
+        (0xFD56, "M", "تمح"),
+        (0xFD57, "M", "تمخ"),
+        (0xFD58, "M", "جمح"),
+        (0xFD5A, "M", "حمي"),
+        (0xFD5B, "M", "حمى"),
+        (0xFD5C, "M", "سحج"),
+        (0xFD5D, "M", "سجح"),
+        (0xFD5E, "M", "سجى"),
+        (0xFD5F, "M", "سمح"),
+        (0xFD61, "M", "سمج"),
+        (0xFD62, "M", "سمم"),
+        (0xFD64, "M", "صحح"),
+        (0xFD66, "M", "صمم"),
+        (0xFD67, "M", "شحم"),
+        (0xFD69, "M", "شجي"),
+        (0xFD6A, "M", "شمخ"),
+        (0xFD6C, "M", "شمم"),
+        (0xFD6E, "M", "ضحى"),
+        (0xFD6F, "M", "ضخم"),
+        (0xFD71, "M", "طمح"),
+        (0xFD73, "M", "طمم"),
+        (0xFD74, "M", "طمي"),
+        (0xFD75, "M", "عجم"),
+        (0xFD76, "M", "عمم"),
+        (0xFD78, "M", "عمى"),
+        (0xFD79, "M", "غمم"),
+        (0xFD7A, "M", "غمي"),
+        (0xFD7B, "M", "غمى"),
+        (0xFD7C, "M", "فخم"),
+        (0xFD7E, "M", "قمح"),
+        (0xFD7F, "M", "قمم"),
+        (0xFD80, "M", "لحم"),
+        (0xFD81, "M", "لحي"),
+        (0xFD82, "M", "لحى"),
+        (0xFD83, "M", "لجج"),
+        (0xFD85, "M", "لخم"),
+        (0xFD87, "M", "لمح"),
+        (0xFD89, "M", "محج"),
+        (0xFD8A, "M", "محم"),
+        (0xFD8B, "M", "محي"),
+        (0xFD8C, "M", "مجح"),
+        (0xFD8D, "M", "مجم"),
+        (0xFD8E, "M", "مخج"),
+        (0xFD8F, "M", "مخم"),
+        (0xFD90, "X"),
+        (0xFD92, "M", "مجخ"),
+        (0xFD93, "M", "همج"),
+        (0xFD94, "M", "همم"),
+        (0xFD95, "M", "نحم"),
+        (0xFD96, "M", "نحى"),
+        (0xFD97, "M", "نجم"),
+        (0xFD99, "M", "نجى"),
+        (0xFD9A, "M", "نمي"),
+        (0xFD9B, "M", "نمى"),
+        (0xFD9C, "M", "يمم"),
+        (0xFD9E, "M", "بخي"),
+        (0xFD9F, "M", "تجي"),
+        (0xFDA0, "M", "تجى"),
+        (0xFDA1, "M", "تخي"),
+        (0xFDA2, "M", "تخى"),
+        (0xFDA3, "M", "تمي"),
+        (0xFDA4, "M", "تمى"),
+        (0xFDA5, "M", "جمي"),
+        (0xFDA6, "M", "جحى"),
+        (0xFDA7, "M", "جمى"),
+        (0xFDA8, "M", "سخى"),
+        (0xFDA9, "M", "صحي"),
+        (0xFDAA, "M", "شحي"),
+    ]
+
+
+def _seg_49() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFDAB, "M", "ضحي"),
+        (0xFDAC, "M", "لجي"),
+        (0xFDAD, "M", "لمي"),
+        (0xFDAE, "M", "يحي"),
+        (0xFDAF, "M", "يجي"),
+        (0xFDB0, "M", "يمي"),
+        (0xFDB1, "M", "ممي"),
+        (0xFDB2, "M", "قمي"),
+        (0xFDB3, "M", "نحي"),
+        (0xFDB4, "M", "قمح"),
+        (0xFDB5, "M", "لحم"),
+        (0xFDB6, "M", "عمي"),
+        (0xFDB7, "M", "كمي"),
+        (0xFDB8, "M", "نجح"),
+        (0xFDB9, "M", "مخي"),
+        (0xFDBA, "M", "لجم"),
+        (0xFDBB, "M", "كمم"),
+        (0xFDBC, "M", "لجم"),
+        (0xFDBD, "M", "نجح"),
+        (0xFDBE, "M", "جحي"),
+        (0xFDBF, "M", "حجي"),
+        (0xFDC0, "M", "مجي"),
+        (0xFDC1, "M", "فمي"),
+        (0xFDC2, "M", "بحي"),
+        (0xFDC3, "M", "كمم"),
+        (0xFDC4, "M", "عجم"),
+        (0xFDC5, "M", "صمم"),
+        (0xFDC6, "M", "سخي"),
+        (0xFDC7, "M", "نجي"),
+        (0xFDC8, "X"),
+        (0xFDCF, "V"),
+        (0xFDD0, "X"),
+        (0xFDF0, "M", "صلے"),
+        (0xFDF1, "M", "قلے"),
+        (0xFDF2, "M", "الله"),
+        (0xFDF3, "M", "اكبر"),
+        (0xFDF4, "M", "محمد"),
+        (0xFDF5, "M", "صلعم"),
+        (0xFDF6, "M", "رسول"),
+        (0xFDF7, "M", "عليه"),
+        (0xFDF8, "M", "وسلم"),
+        (0xFDF9, "M", "صلى"),
+        (0xFDFA, "3", "صلى الله عليه وسلم"),
+        (0xFDFB, "3", "جل جلاله"),
+        (0xFDFC, "M", "ریال"),
+        (0xFDFD, "V"),
+        (0xFE00, "I"),
+        (0xFE10, "3", ","),
+        (0xFE11, "M", "、"),
+        (0xFE12, "X"),
+        (0xFE13, "3", ":"),
+        (0xFE14, "3", ";"),
+        (0xFE15, "3", "!"),
+        (0xFE16, "3", "?"),
+        (0xFE17, "M", "〖"),
+        (0xFE18, "M", "〗"),
+        (0xFE19, "X"),
+        (0xFE20, "V"),
+        (0xFE30, "X"),
+        (0xFE31, "M", "—"),
+        (0xFE32, "M", "–"),
+        (0xFE33, "3", "_"),
+        (0xFE35, "3", "("),
+        (0xFE36, "3", ")"),
+        (0xFE37, "3", "{"),
+        (0xFE38, "3", "}"),
+        (0xFE39, "M", "〔"),
+        (0xFE3A, "M", "〕"),
+        (0xFE3B, "M", "【"),
+        (0xFE3C, "M", "】"),
+        (0xFE3D, "M", "《"),
+        (0xFE3E, "M", "》"),
+        (0xFE3F, "M", "〈"),
+        (0xFE40, "M", "〉"),
+        (0xFE41, "M", "「"),
+        (0xFE42, "M", "」"),
+        (0xFE43, "M", "『"),
+        (0xFE44, "M", "』"),
+        (0xFE45, "V"),
+        (0xFE47, "3", "["),
+        (0xFE48, "3", "]"),
+        (0xFE49, "3", " ̅"),
+        (0xFE4D, "3", "_"),
+        (0xFE50, "3", ","),
+        (0xFE51, "M", "、"),
+        (0xFE52, "X"),
+        (0xFE54, "3", ";"),
+        (0xFE55, "3", ":"),
+        (0xFE56, "3", "?"),
+        (0xFE57, "3", "!"),
+        (0xFE58, "M", "—"),
+        (0xFE59, "3", "("),
+        (0xFE5A, "3", ")"),
+        (0xFE5B, "3", "{"),
+        (0xFE5C, "3", "}"),
+        (0xFE5D, "M", "〔"),
+        (0xFE5E, "M", "〕"),
+        (0xFE5F, "3", "#"),
+        (0xFE60, "3", "&"),
+        (0xFE61, "3", "*"),
+    ]
+
+
+def _seg_50() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFE62, "3", "+"),
+        (0xFE63, "M", "-"),
+        (0xFE64, "3", "<"),
+        (0xFE65, "3", ">"),
+        (0xFE66, "3", "="),
+        (0xFE67, "X"),
+        (0xFE68, "3", "\\"),
+        (0xFE69, "3", "$"),
+        (0xFE6A, "3", "%"),
+        (0xFE6B, "3", "@"),
+        (0xFE6C, "X"),
+        (0xFE70, "3", " ً"),
+        (0xFE71, "M", "ـً"),
+        (0xFE72, "3", " ٌ"),
+        (0xFE73, "V"),
+        (0xFE74, "3", " ٍ"),
+        (0xFE75, "X"),
+        (0xFE76, "3", " َ"),
+        (0xFE77, "M", "ـَ"),
+        (0xFE78, "3", " ُ"),
+        (0xFE79, "M", "ـُ"),
+        (0xFE7A, "3", " ِ"),
+        (0xFE7B, "M", "ـِ"),
+        (0xFE7C, "3", " ّ"),
+        (0xFE7D, "M", "ـّ"),
+        (0xFE7E, "3", " ْ"),
+        (0xFE7F, "M", "ـْ"),
+        (0xFE80, "M", "ء"),
+        (0xFE81, "M", "آ"),
+        (0xFE83, "M", "أ"),
+        (0xFE85, "M", "ؤ"),
+        (0xFE87, "M", "إ"),
+        (0xFE89, "M", "ئ"),
+        (0xFE8D, "M", "ا"),
+        (0xFE8F, "M", "ب"),
+        (0xFE93, "M", "ة"),
+        (0xFE95, "M", "ت"),
+        (0xFE99, "M", "ث"),
+        (0xFE9D, "M", "ج"),
+        (0xFEA1, "M", "ح"),
+        (0xFEA5, "M", "خ"),
+        (0xFEA9, "M", "د"),
+        (0xFEAB, "M", "ذ"),
+        (0xFEAD, "M", "ر"),
+        (0xFEAF, "M", "ز"),
+        (0xFEB1, "M", "س"),
+        (0xFEB5, "M", "ش"),
+        (0xFEB9, "M", "ص"),
+        (0xFEBD, "M", "ض"),
+        (0xFEC1, "M", "ط"),
+        (0xFEC5, "M", "ظ"),
+        (0xFEC9, "M", "ع"),
+        (0xFECD, "M", "غ"),
+        (0xFED1, "M", "ف"),
+        (0xFED5, "M", "ق"),
+        (0xFED9, "M", "ك"),
+        (0xFEDD, "M", "ل"),
+        (0xFEE1, "M", "م"),
+        (0xFEE5, "M", "ن"),
+        (0xFEE9, "M", "ه"),
+        (0xFEED, "M", "و"),
+        (0xFEEF, "M", "ى"),
+        (0xFEF1, "M", "ي"),
+        (0xFEF5, "M", "لآ"),
+        (0xFEF7, "M", "لأ"),
+        (0xFEF9, "M", "لإ"),
+        (0xFEFB, "M", "لا"),
+        (0xFEFD, "X"),
+        (0xFEFF, "I"),
+        (0xFF00, "X"),
+        (0xFF01, "3", "!"),
+        (0xFF02, "3", '"'),
+        (0xFF03, "3", "#"),
+        (0xFF04, "3", "$"),
+        (0xFF05, "3", "%"),
+        (0xFF06, "3", "&"),
+        (0xFF07, "3", "'"),
+        (0xFF08, "3", "("),
+        (0xFF09, "3", ")"),
+        (0xFF0A, "3", "*"),
+        (0xFF0B, "3", "+"),
+        (0xFF0C, "3", ","),
+        (0xFF0D, "M", "-"),
+        (0xFF0E, "M", "."),
+        (0xFF0F, "3", "/"),
+        (0xFF10, "M", "0"),
+        (0xFF11, "M", "1"),
+        (0xFF12, "M", "2"),
+        (0xFF13, "M", "3"),
+        (0xFF14, "M", "4"),
+        (0xFF15, "M", "5"),
+        (0xFF16, "M", "6"),
+        (0xFF17, "M", "7"),
+        (0xFF18, "M", "8"),
+        (0xFF19, "M", "9"),
+        (0xFF1A, "3", ":"),
+        (0xFF1B, "3", ";"),
+        (0xFF1C, "3", "<"),
+        (0xFF1D, "3", "="),
+        (0xFF1E, "3", ">"),
+    ]
+
+
+def _seg_51() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFF1F, "3", "?"),
+        (0xFF20, "3", "@"),
+        (0xFF21, "M", "a"),
+        (0xFF22, "M", "b"),
+        (0xFF23, "M", "c"),
+        (0xFF24, "M", "d"),
+        (0xFF25, "M", "e"),
+        (0xFF26, "M", "f"),
+        (0xFF27, "M", "g"),
+        (0xFF28, "M", "h"),
+        (0xFF29, "M", "i"),
+        (0xFF2A, "M", "j"),
+        (0xFF2B, "M", "k"),
+        (0xFF2C, "M", "l"),
+        (0xFF2D, "M", "m"),
+        (0xFF2E, "M", "n"),
+        (0xFF2F, "M", "o"),
+        (0xFF30, "M", "p"),
+        (0xFF31, "M", "q"),
+        (0xFF32, "M", "r"),
+        (0xFF33, "M", "s"),
+        (0xFF34, "M", "t"),
+        (0xFF35, "M", "u"),
+        (0xFF36, "M", "v"),
+        (0xFF37, "M", "w"),
+        (0xFF38, "M", "x"),
+        (0xFF39, "M", "y"),
+        (0xFF3A, "M", "z"),
+        (0xFF3B, "3", "["),
+        (0xFF3C, "3", "\\"),
+        (0xFF3D, "3", "]"),
+        (0xFF3E, "3", "^"),
+        (0xFF3F, "3", "_"),
+        (0xFF40, "3", "`"),
+        (0xFF41, "M", "a"),
+        (0xFF42, "M", "b"),
+        (0xFF43, "M", "c"),
+        (0xFF44, "M", "d"),
+        (0xFF45, "M", "e"),
+        (0xFF46, "M", "f"),
+        (0xFF47, "M", "g"),
+        (0xFF48, "M", "h"),
+        (0xFF49, "M", "i"),
+        (0xFF4A, "M", "j"),
+        (0xFF4B, "M", "k"),
+        (0xFF4C, "M", "l"),
+        (0xFF4D, "M", "m"),
+        (0xFF4E, "M", "n"),
+        (0xFF4F, "M", "o"),
+        (0xFF50, "M", "p"),
+        (0xFF51, "M", "q"),
+        (0xFF52, "M", "r"),
+        (0xFF53, "M", "s"),
+        (0xFF54, "M", "t"),
+        (0xFF55, "M", "u"),
+        (0xFF56, "M", "v"),
+        (0xFF57, "M", "w"),
+        (0xFF58, "M", "x"),
+        (0xFF59, "M", "y"),
+        (0xFF5A, "M", "z"),
+        (0xFF5B, "3", "{"),
+        (0xFF5C, "3", "|"),
+        (0xFF5D, "3", "}"),
+        (0xFF5E, "3", "~"),
+        (0xFF5F, "M", "⦅"),
+        (0xFF60, "M", "⦆"),
+        (0xFF61, "M", "."),
+        (0xFF62, "M", "「"),
+        (0xFF63, "M", "」"),
+        (0xFF64, "M", "、"),
+        (0xFF65, "M", "・"),
+        (0xFF66, "M", "ヲ"),
+        (0xFF67, "M", "ァ"),
+        (0xFF68, "M", "ィ"),
+        (0xFF69, "M", "ゥ"),
+        (0xFF6A, "M", "ェ"),
+        (0xFF6B, "M", "ォ"),
+        (0xFF6C, "M", "ャ"),
+        (0xFF6D, "M", "ュ"),
+        (0xFF6E, "M", "ョ"),
+        (0xFF6F, "M", "ッ"),
+        (0xFF70, "M", "ー"),
+        (0xFF71, "M", "ア"),
+        (0xFF72, "M", "イ"),
+        (0xFF73, "M", "ウ"),
+        (0xFF74, "M", "エ"),
+        (0xFF75, "M", "オ"),
+        (0xFF76, "M", "カ"),
+        (0xFF77, "M", "キ"),
+        (0xFF78, "M", "ク"),
+        (0xFF79, "M", "ケ"),
+        (0xFF7A, "M", "コ"),
+        (0xFF7B, "M", "サ"),
+        (0xFF7C, "M", "シ"),
+        (0xFF7D, "M", "ス"),
+        (0xFF7E, "M", "セ"),
+        (0xFF7F, "M", "ソ"),
+        (0xFF80, "M", "タ"),
+        (0xFF81, "M", "チ"),
+        (0xFF82, "M", "ツ"),
+    ]
+
+
+def _seg_52() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFF83, "M", "テ"),
+        (0xFF84, "M", "ト"),
+        (0xFF85, "M", "ナ"),
+        (0xFF86, "M", "ニ"),
+        (0xFF87, "M", "ヌ"),
+        (0xFF88, "M", "ネ"),
+        (0xFF89, "M", "ノ"),
+        (0xFF8A, "M", "ハ"),
+        (0xFF8B, "M", "ヒ"),
+        (0xFF8C, "M", "フ"),
+        (0xFF8D, "M", "ヘ"),
+        (0xFF8E, "M", "ホ"),
+        (0xFF8F, "M", "マ"),
+        (0xFF90, "M", "ミ"),
+        (0xFF91, "M", "ム"),
+        (0xFF92, "M", "メ"),
+        (0xFF93, "M", "モ"),
+        (0xFF94, "M", "ヤ"),
+        (0xFF95, "M", "ユ"),
+        (0xFF96, "M", "ヨ"),
+        (0xFF97, "M", "ラ"),
+        (0xFF98, "M", "リ"),
+        (0xFF99, "M", "ル"),
+        (0xFF9A, "M", "レ"),
+        (0xFF9B, "M", "ロ"),
+        (0xFF9C, "M", "ワ"),
+        (0xFF9D, "M", "ン"),
+        (0xFF9E, "M", "゙"),
+        (0xFF9F, "M", "゚"),
+        (0xFFA0, "X"),
+        (0xFFA1, "M", "ᄀ"),
+        (0xFFA2, "M", "ᄁ"),
+        (0xFFA3, "M", "ᆪ"),
+        (0xFFA4, "M", "ᄂ"),
+        (0xFFA5, "M", "ᆬ"),
+        (0xFFA6, "M", "ᆭ"),
+        (0xFFA7, "M", "ᄃ"),
+        (0xFFA8, "M", "ᄄ"),
+        (0xFFA9, "M", "ᄅ"),
+        (0xFFAA, "M", "ᆰ"),
+        (0xFFAB, "M", "ᆱ"),
+        (0xFFAC, "M", "ᆲ"),
+        (0xFFAD, "M", "ᆳ"),
+        (0xFFAE, "M", "ᆴ"),
+        (0xFFAF, "M", "ᆵ"),
+        (0xFFB0, "M", "ᄚ"),
+        (0xFFB1, "M", "ᄆ"),
+        (0xFFB2, "M", "ᄇ"),
+        (0xFFB3, "M", "ᄈ"),
+        (0xFFB4, "M", "ᄡ"),
+        (0xFFB5, "M", "ᄉ"),
+        (0xFFB6, "M", "ᄊ"),
+        (0xFFB7, "M", "ᄋ"),
+        (0xFFB8, "M", "ᄌ"),
+        (0xFFB9, "M", "ᄍ"),
+        (0xFFBA, "M", "ᄎ"),
+        (0xFFBB, "M", "ᄏ"),
+        (0xFFBC, "M", "ᄐ"),
+        (0xFFBD, "M", "ᄑ"),
+        (0xFFBE, "M", "ᄒ"),
+        (0xFFBF, "X"),
+        (0xFFC2, "M", "ᅡ"),
+        (0xFFC3, "M", "ᅢ"),
+        (0xFFC4, "M", "ᅣ"),
+        (0xFFC5, "M", "ᅤ"),
+        (0xFFC6, "M", "ᅥ"),
+        (0xFFC7, "M", "ᅦ"),
+        (0xFFC8, "X"),
+        (0xFFCA, "M", "ᅧ"),
+        (0xFFCB, "M", "ᅨ"),
+        (0xFFCC, "M", "ᅩ"),
+        (0xFFCD, "M", "ᅪ"),
+        (0xFFCE, "M", "ᅫ"),
+        (0xFFCF, "M", "ᅬ"),
+        (0xFFD0, "X"),
+        (0xFFD2, "M", "ᅭ"),
+        (0xFFD3, "M", "ᅮ"),
+        (0xFFD4, "M", "ᅯ"),
+        (0xFFD5, "M", "ᅰ"),
+        (0xFFD6, "M", "ᅱ"),
+        (0xFFD7, "M", "ᅲ"),
+        (0xFFD8, "X"),
+        (0xFFDA, "M", "ᅳ"),
+        (0xFFDB, "M", "ᅴ"),
+        (0xFFDC, "M", "ᅵ"),
+        (0xFFDD, "X"),
+        (0xFFE0, "M", "¢"),
+        (0xFFE1, "M", "£"),
+        (0xFFE2, "M", "¬"),
+        (0xFFE3, "3", " ̄"),
+        (0xFFE4, "M", "¦"),
+        (0xFFE5, "M", "¥"),
+        (0xFFE6, "M", "₩"),
+        (0xFFE7, "X"),
+        (0xFFE8, "M", "│"),
+        (0xFFE9, "M", "←"),
+        (0xFFEA, "M", "↑"),
+        (0xFFEB, "M", "→"),
+        (0xFFEC, "M", "↓"),
+        (0xFFED, "M", "■"),
+    ]
+
+
+def _seg_53() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0xFFEE, "M", "○"),
+        (0xFFEF, "X"),
+        (0x10000, "V"),
+        (0x1000C, "X"),
+        (0x1000D, "V"),
+        (0x10027, "X"),
+        (0x10028, "V"),
+        (0x1003B, "X"),
+        (0x1003C, "V"),
+        (0x1003E, "X"),
+        (0x1003F, "V"),
+        (0x1004E, "X"),
+        (0x10050, "V"),
+        (0x1005E, "X"),
+        (0x10080, "V"),
+        (0x100FB, "X"),
+        (0x10100, "V"),
+        (0x10103, "X"),
+        (0x10107, "V"),
+        (0x10134, "X"),
+        (0x10137, "V"),
+        (0x1018F, "X"),
+        (0x10190, "V"),
+        (0x1019D, "X"),
+        (0x101A0, "V"),
+        (0x101A1, "X"),
+        (0x101D0, "V"),
+        (0x101FE, "X"),
+        (0x10280, "V"),
+        (0x1029D, "X"),
+        (0x102A0, "V"),
+        (0x102D1, "X"),
+        (0x102E0, "V"),
+        (0x102FC, "X"),
+        (0x10300, "V"),
+        (0x10324, "X"),
+        (0x1032D, "V"),
+        (0x1034B, "X"),
+        (0x10350, "V"),
+        (0x1037B, "X"),
+        (0x10380, "V"),
+        (0x1039E, "X"),
+        (0x1039F, "V"),
+        (0x103C4, "X"),
+        (0x103C8, "V"),
+        (0x103D6, "X"),
+        (0x10400, "M", "𐐨"),
+        (0x10401, "M", "𐐩"),
+        (0x10402, "M", "𐐪"),
+        (0x10403, "M", "𐐫"),
+        (0x10404, "M", "𐐬"),
+        (0x10405, "M", "𐐭"),
+        (0x10406, "M", "𐐮"),
+        (0x10407, "M", "𐐯"),
+        (0x10408, "M", "𐐰"),
+        (0x10409, "M", "𐐱"),
+        (0x1040A, "M", "𐐲"),
+        (0x1040B, "M", "𐐳"),
+        (0x1040C, "M", "𐐴"),
+        (0x1040D, "M", "𐐵"),
+        (0x1040E, "M", "𐐶"),
+        (0x1040F, "M", "𐐷"),
+        (0x10410, "M", "𐐸"),
+        (0x10411, "M", "𐐹"),
+        (0x10412, "M", "𐐺"),
+        (0x10413, "M", "𐐻"),
+        (0x10414, "M", "𐐼"),
+        (0x10415, "M", "𐐽"),
+        (0x10416, "M", "𐐾"),
+        (0x10417, "M", "𐐿"),
+        (0x10418, "M", "𐑀"),
+        (0x10419, "M", "𐑁"),
+        (0x1041A, "M", "𐑂"),
+        (0x1041B, "M", "𐑃"),
+        (0x1041C, "M", "𐑄"),
+        (0x1041D, "M", "𐑅"),
+        (0x1041E, "M", "𐑆"),
+        (0x1041F, "M", "𐑇"),
+        (0x10420, "M", "𐑈"),
+        (0x10421, "M", "𐑉"),
+        (0x10422, "M", "𐑊"),
+        (0x10423, "M", "𐑋"),
+        (0x10424, "M", "𐑌"),
+        (0x10425, "M", "𐑍"),
+        (0x10426, "M", "𐑎"),
+        (0x10427, "M", "𐑏"),
+        (0x10428, "V"),
+        (0x1049E, "X"),
+        (0x104A0, "V"),
+        (0x104AA, "X"),
+        (0x104B0, "M", "𐓘"),
+        (0x104B1, "M", "𐓙"),
+        (0x104B2, "M", "𐓚"),
+        (0x104B3, "M", "𐓛"),
+        (0x104B4, "M", "𐓜"),
+        (0x104B5, "M", "𐓝"),
+        (0x104B6, "M", "𐓞"),
+        (0x104B7, "M", "𐓟"),
+        (0x104B8, "M", "𐓠"),
+        (0x104B9, "M", "𐓡"),
+    ]
+
+
+def _seg_54() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x104BA, "M", "𐓢"),
+        (0x104BB, "M", "𐓣"),
+        (0x104BC, "M", "𐓤"),
+        (0x104BD, "M", "𐓥"),
+        (0x104BE, "M", "𐓦"),
+        (0x104BF, "M", "𐓧"),
+        (0x104C0, "M", "𐓨"),
+        (0x104C1, "M", "𐓩"),
+        (0x104C2, "M", "𐓪"),
+        (0x104C3, "M", "𐓫"),
+        (0x104C4, "M", "𐓬"),
+        (0x104C5, "M", "𐓭"),
+        (0x104C6, "M", "𐓮"),
+        (0x104C7, "M", "𐓯"),
+        (0x104C8, "M", "𐓰"),
+        (0x104C9, "M", "𐓱"),
+        (0x104CA, "M", "𐓲"),
+        (0x104CB, "M", "𐓳"),
+        (0x104CC, "M", "𐓴"),
+        (0x104CD, "M", "𐓵"),
+        (0x104CE, "M", "𐓶"),
+        (0x104CF, "M", "𐓷"),
+        (0x104D0, "M", "𐓸"),
+        (0x104D1, "M", "𐓹"),
+        (0x104D2, "M", "𐓺"),
+        (0x104D3, "M", "𐓻"),
+        (0x104D4, "X"),
+        (0x104D8, "V"),
+        (0x104FC, "X"),
+        (0x10500, "V"),
+        (0x10528, "X"),
+        (0x10530, "V"),
+        (0x10564, "X"),
+        (0x1056F, "V"),
+        (0x10570, "M", "𐖗"),
+        (0x10571, "M", "𐖘"),
+        (0x10572, "M", "𐖙"),
+        (0x10573, "M", "𐖚"),
+        (0x10574, "M", "𐖛"),
+        (0x10575, "M", "𐖜"),
+        (0x10576, "M", "𐖝"),
+        (0x10577, "M", "𐖞"),
+        (0x10578, "M", "𐖟"),
+        (0x10579, "M", "𐖠"),
+        (0x1057A, "M", "𐖡"),
+        (0x1057B, "X"),
+        (0x1057C, "M", "𐖣"),
+        (0x1057D, "M", "𐖤"),
+        (0x1057E, "M", "𐖥"),
+        (0x1057F, "M", "𐖦"),
+        (0x10580, "M", "𐖧"),
+        (0x10581, "M", "𐖨"),
+        (0x10582, "M", "𐖩"),
+        (0x10583, "M", "𐖪"),
+        (0x10584, "M", "𐖫"),
+        (0x10585, "M", "𐖬"),
+        (0x10586, "M", "𐖭"),
+        (0x10587, "M", "𐖮"),
+        (0x10588, "M", "𐖯"),
+        (0x10589, "M", "𐖰"),
+        (0x1058A, "M", "𐖱"),
+        (0x1058B, "X"),
+        (0x1058C, "M", "𐖳"),
+        (0x1058D, "M", "𐖴"),
+        (0x1058E, "M", "𐖵"),
+        (0x1058F, "M", "𐖶"),
+        (0x10590, "M", "𐖷"),
+        (0x10591, "M", "𐖸"),
+        (0x10592, "M", "𐖹"),
+        (0x10593, "X"),
+        (0x10594, "M", "𐖻"),
+        (0x10595, "M", "𐖼"),
+        (0x10596, "X"),
+        (0x10597, "V"),
+        (0x105A2, "X"),
+        (0x105A3, "V"),
+        (0x105B2, "X"),
+        (0x105B3, "V"),
+        (0x105BA, "X"),
+        (0x105BB, "V"),
+        (0x105BD, "X"),
+        (0x10600, "V"),
+        (0x10737, "X"),
+        (0x10740, "V"),
+        (0x10756, "X"),
+        (0x10760, "V"),
+        (0x10768, "X"),
+        (0x10780, "V"),
+        (0x10781, "M", "ː"),
+        (0x10782, "M", "ˑ"),
+        (0x10783, "M", "æ"),
+        (0x10784, "M", "ʙ"),
+        (0x10785, "M", "ɓ"),
+        (0x10786, "X"),
+        (0x10787, "M", "ʣ"),
+        (0x10788, "M", "ꭦ"),
+        (0x10789, "M", "ʥ"),
+        (0x1078A, "M", "ʤ"),
+        (0x1078B, "M", "ɖ"),
+        (0x1078C, "M", "ɗ"),
+    ]
+
+
+def _seg_55() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1078D, "M", "ᶑ"),
+        (0x1078E, "M", "ɘ"),
+        (0x1078F, "M", "ɞ"),
+        (0x10790, "M", "ʩ"),
+        (0x10791, "M", "ɤ"),
+        (0x10792, "M", "ɢ"),
+        (0x10793, "M", "ɠ"),
+        (0x10794, "M", "ʛ"),
+        (0x10795, "M", "ħ"),
+        (0x10796, "M", "ʜ"),
+        (0x10797, "M", "ɧ"),
+        (0x10798, "M", "ʄ"),
+        (0x10799, "M", "ʪ"),
+        (0x1079A, "M", "ʫ"),
+        (0x1079B, "M", "ɬ"),
+        (0x1079C, "M", "𝼄"),
+        (0x1079D, "M", "ꞎ"),
+        (0x1079E, "M", "ɮ"),
+        (0x1079F, "M", "𝼅"),
+        (0x107A0, "M", "ʎ"),
+        (0x107A1, "M", "𝼆"),
+        (0x107A2, "M", "ø"),
+        (0x107A3, "M", "ɶ"),
+        (0x107A4, "M", "ɷ"),
+        (0x107A5, "M", "q"),
+        (0x107A6, "M", "ɺ"),
+        (0x107A7, "M", "𝼈"),
+        (0x107A8, "M", "ɽ"),
+        (0x107A9, "M", "ɾ"),
+        (0x107AA, "M", "ʀ"),
+        (0x107AB, "M", "ʨ"),
+        (0x107AC, "M", "ʦ"),
+        (0x107AD, "M", "ꭧ"),
+        (0x107AE, "M", "ʧ"),
+        (0x107AF, "M", "ʈ"),
+        (0x107B0, "M", "ⱱ"),
+        (0x107B1, "X"),
+        (0x107B2, "M", "ʏ"),
+        (0x107B3, "M", "ʡ"),
+        (0x107B4, "M", "ʢ"),
+        (0x107B5, "M", "ʘ"),
+        (0x107B6, "M", "ǀ"),
+        (0x107B7, "M", "ǁ"),
+        (0x107B8, "M", "ǂ"),
+        (0x107B9, "M", "𝼊"),
+        (0x107BA, "M", "𝼞"),
+        (0x107BB, "X"),
+        (0x10800, "V"),
+        (0x10806, "X"),
+        (0x10808, "V"),
+        (0x10809, "X"),
+        (0x1080A, "V"),
+        (0x10836, "X"),
+        (0x10837, "V"),
+        (0x10839, "X"),
+        (0x1083C, "V"),
+        (0x1083D, "X"),
+        (0x1083F, "V"),
+        (0x10856, "X"),
+        (0x10857, "V"),
+        (0x1089F, "X"),
+        (0x108A7, "V"),
+        (0x108B0, "X"),
+        (0x108E0, "V"),
+        (0x108F3, "X"),
+        (0x108F4, "V"),
+        (0x108F6, "X"),
+        (0x108FB, "V"),
+        (0x1091C, "X"),
+        (0x1091F, "V"),
+        (0x1093A, "X"),
+        (0x1093F, "V"),
+        (0x10940, "X"),
+        (0x10980, "V"),
+        (0x109B8, "X"),
+        (0x109BC, "V"),
+        (0x109D0, "X"),
+        (0x109D2, "V"),
+        (0x10A04, "X"),
+        (0x10A05, "V"),
+        (0x10A07, "X"),
+        (0x10A0C, "V"),
+        (0x10A14, "X"),
+        (0x10A15, "V"),
+        (0x10A18, "X"),
+        (0x10A19, "V"),
+        (0x10A36, "X"),
+        (0x10A38, "V"),
+        (0x10A3B, "X"),
+        (0x10A3F, "V"),
+        (0x10A49, "X"),
+        (0x10A50, "V"),
+        (0x10A59, "X"),
+        (0x10A60, "V"),
+        (0x10AA0, "X"),
+        (0x10AC0, "V"),
+        (0x10AE7, "X"),
+        (0x10AEB, "V"),
+        (0x10AF7, "X"),
+        (0x10B00, "V"),
+    ]
+
+
+def _seg_56() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x10B36, "X"),
+        (0x10B39, "V"),
+        (0x10B56, "X"),
+        (0x10B58, "V"),
+        (0x10B73, "X"),
+        (0x10B78, "V"),
+        (0x10B92, "X"),
+        (0x10B99, "V"),
+        (0x10B9D, "X"),
+        (0x10BA9, "V"),
+        (0x10BB0, "X"),
+        (0x10C00, "V"),
+        (0x10C49, "X"),
+        (0x10C80, "M", "𐳀"),
+        (0x10C81, "M", "𐳁"),
+        (0x10C82, "M", "𐳂"),
+        (0x10C83, "M", "𐳃"),
+        (0x10C84, "M", "𐳄"),
+        (0x10C85, "M", "𐳅"),
+        (0x10C86, "M", "𐳆"),
+        (0x10C87, "M", "𐳇"),
+        (0x10C88, "M", "𐳈"),
+        (0x10C89, "M", "𐳉"),
+        (0x10C8A, "M", "𐳊"),
+        (0x10C8B, "M", "𐳋"),
+        (0x10C8C, "M", "𐳌"),
+        (0x10C8D, "M", "𐳍"),
+        (0x10C8E, "M", "𐳎"),
+        (0x10C8F, "M", "𐳏"),
+        (0x10C90, "M", "𐳐"),
+        (0x10C91, "M", "𐳑"),
+        (0x10C92, "M", "𐳒"),
+        (0x10C93, "M", "𐳓"),
+        (0x10C94, "M", "𐳔"),
+        (0x10C95, "M", "𐳕"),
+        (0x10C96, "M", "𐳖"),
+        (0x10C97, "M", "𐳗"),
+        (0x10C98, "M", "𐳘"),
+        (0x10C99, "M", "𐳙"),
+        (0x10C9A, "M", "𐳚"),
+        (0x10C9B, "M", "𐳛"),
+        (0x10C9C, "M", "𐳜"),
+        (0x10C9D, "M", "𐳝"),
+        (0x10C9E, "M", "𐳞"),
+        (0x10C9F, "M", "𐳟"),
+        (0x10CA0, "M", "𐳠"),
+        (0x10CA1, "M", "𐳡"),
+        (0x10CA2, "M", "𐳢"),
+        (0x10CA3, "M", "𐳣"),
+        (0x10CA4, "M", "𐳤"),
+        (0x10CA5, "M", "𐳥"),
+        (0x10CA6, "M", "𐳦"),
+        (0x10CA7, "M", "𐳧"),
+        (0x10CA8, "M", "𐳨"),
+        (0x10CA9, "M", "𐳩"),
+        (0x10CAA, "M", "𐳪"),
+        (0x10CAB, "M", "𐳫"),
+        (0x10CAC, "M", "𐳬"),
+        (0x10CAD, "M", "𐳭"),
+        (0x10CAE, "M", "𐳮"),
+        (0x10CAF, "M", "𐳯"),
+        (0x10CB0, "M", "𐳰"),
+        (0x10CB1, "M", "𐳱"),
+        (0x10CB2, "M", "𐳲"),
+        (0x10CB3, "X"),
+        (0x10CC0, "V"),
+        (0x10CF3, "X"),
+        (0x10CFA, "V"),
+        (0x10D28, "X"),
+        (0x10D30, "V"),
+        (0x10D3A, "X"),
+        (0x10E60, "V"),
+        (0x10E7F, "X"),
+        (0x10E80, "V"),
+        (0x10EAA, "X"),
+        (0x10EAB, "V"),
+        (0x10EAE, "X"),
+        (0x10EB0, "V"),
+        (0x10EB2, "X"),
+        (0x10EFD, "V"),
+        (0x10F28, "X"),
+        (0x10F30, "V"),
+        (0x10F5A, "X"),
+        (0x10F70, "V"),
+        (0x10F8A, "X"),
+        (0x10FB0, "V"),
+        (0x10FCC, "X"),
+        (0x10FE0, "V"),
+        (0x10FF7, "X"),
+        (0x11000, "V"),
+        (0x1104E, "X"),
+        (0x11052, "V"),
+        (0x11076, "X"),
+        (0x1107F, "V"),
+        (0x110BD, "X"),
+        (0x110BE, "V"),
+        (0x110C3, "X"),
+        (0x110D0, "V"),
+        (0x110E9, "X"),
+        (0x110F0, "V"),
+    ]
+
+
+def _seg_57() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x110FA, "X"),
+        (0x11100, "V"),
+        (0x11135, "X"),
+        (0x11136, "V"),
+        (0x11148, "X"),
+        (0x11150, "V"),
+        (0x11177, "X"),
+        (0x11180, "V"),
+        (0x111E0, "X"),
+        (0x111E1, "V"),
+        (0x111F5, "X"),
+        (0x11200, "V"),
+        (0x11212, "X"),
+        (0x11213, "V"),
+        (0x11242, "X"),
+        (0x11280, "V"),
+        (0x11287, "X"),
+        (0x11288, "V"),
+        (0x11289, "X"),
+        (0x1128A, "V"),
+        (0x1128E, "X"),
+        (0x1128F, "V"),
+        (0x1129E, "X"),
+        (0x1129F, "V"),
+        (0x112AA, "X"),
+        (0x112B0, "V"),
+        (0x112EB, "X"),
+        (0x112F0, "V"),
+        (0x112FA, "X"),
+        (0x11300, "V"),
+        (0x11304, "X"),
+        (0x11305, "V"),
+        (0x1130D, "X"),
+        (0x1130F, "V"),
+        (0x11311, "X"),
+        (0x11313, "V"),
+        (0x11329, "X"),
+        (0x1132A, "V"),
+        (0x11331, "X"),
+        (0x11332, "V"),
+        (0x11334, "X"),
+        (0x11335, "V"),
+        (0x1133A, "X"),
+        (0x1133B, "V"),
+        (0x11345, "X"),
+        (0x11347, "V"),
+        (0x11349, "X"),
+        (0x1134B, "V"),
+        (0x1134E, "X"),
+        (0x11350, "V"),
+        (0x11351, "X"),
+        (0x11357, "V"),
+        (0x11358, "X"),
+        (0x1135D, "V"),
+        (0x11364, "X"),
+        (0x11366, "V"),
+        (0x1136D, "X"),
+        (0x11370, "V"),
+        (0x11375, "X"),
+        (0x11400, "V"),
+        (0x1145C, "X"),
+        (0x1145D, "V"),
+        (0x11462, "X"),
+        (0x11480, "V"),
+        (0x114C8, "X"),
+        (0x114D0, "V"),
+        (0x114DA, "X"),
+        (0x11580, "V"),
+        (0x115B6, "X"),
+        (0x115B8, "V"),
+        (0x115DE, "X"),
+        (0x11600, "V"),
+        (0x11645, "X"),
+        (0x11650, "V"),
+        (0x1165A, "X"),
+        (0x11660, "V"),
+        (0x1166D, "X"),
+        (0x11680, "V"),
+        (0x116BA, "X"),
+        (0x116C0, "V"),
+        (0x116CA, "X"),
+        (0x11700, "V"),
+        (0x1171B, "X"),
+        (0x1171D, "V"),
+        (0x1172C, "X"),
+        (0x11730, "V"),
+        (0x11747, "X"),
+        (0x11800, "V"),
+        (0x1183C, "X"),
+        (0x118A0, "M", "𑣀"),
+        (0x118A1, "M", "𑣁"),
+        (0x118A2, "M", "𑣂"),
+        (0x118A3, "M", "𑣃"),
+        (0x118A4, "M", "𑣄"),
+        (0x118A5, "M", "𑣅"),
+        (0x118A6, "M", "𑣆"),
+        (0x118A7, "M", "𑣇"),
+        (0x118A8, "M", "𑣈"),
+        (0x118A9, "M", "𑣉"),
+        (0x118AA, "M", "𑣊"),
+    ]
+
+
+def _seg_58() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x118AB, "M", "𑣋"),
+        (0x118AC, "M", "𑣌"),
+        (0x118AD, "M", "𑣍"),
+        (0x118AE, "M", "𑣎"),
+        (0x118AF, "M", "𑣏"),
+        (0x118B0, "M", "𑣐"),
+        (0x118B1, "M", "𑣑"),
+        (0x118B2, "M", "𑣒"),
+        (0x118B3, "M", "𑣓"),
+        (0x118B4, "M", "𑣔"),
+        (0x118B5, "M", "𑣕"),
+        (0x118B6, "M", "𑣖"),
+        (0x118B7, "M", "𑣗"),
+        (0x118B8, "M", "𑣘"),
+        (0x118B9, "M", "𑣙"),
+        (0x118BA, "M", "𑣚"),
+        (0x118BB, "M", "𑣛"),
+        (0x118BC, "M", "𑣜"),
+        (0x118BD, "M", "𑣝"),
+        (0x118BE, "M", "𑣞"),
+        (0x118BF, "M", "𑣟"),
+        (0x118C0, "V"),
+        (0x118F3, "X"),
+        (0x118FF, "V"),
+        (0x11907, "X"),
+        (0x11909, "V"),
+        (0x1190A, "X"),
+        (0x1190C, "V"),
+        (0x11914, "X"),
+        (0x11915, "V"),
+        (0x11917, "X"),
+        (0x11918, "V"),
+        (0x11936, "X"),
+        (0x11937, "V"),
+        (0x11939, "X"),
+        (0x1193B, "V"),
+        (0x11947, "X"),
+        (0x11950, "V"),
+        (0x1195A, "X"),
+        (0x119A0, "V"),
+        (0x119A8, "X"),
+        (0x119AA, "V"),
+        (0x119D8, "X"),
+        (0x119DA, "V"),
+        (0x119E5, "X"),
+        (0x11A00, "V"),
+        (0x11A48, "X"),
+        (0x11A50, "V"),
+        (0x11AA3, "X"),
+        (0x11AB0, "V"),
+        (0x11AF9, "X"),
+        (0x11B00, "V"),
+        (0x11B0A, "X"),
+        (0x11C00, "V"),
+        (0x11C09, "X"),
+        (0x11C0A, "V"),
+        (0x11C37, "X"),
+        (0x11C38, "V"),
+        (0x11C46, "X"),
+        (0x11C50, "V"),
+        (0x11C6D, "X"),
+        (0x11C70, "V"),
+        (0x11C90, "X"),
+        (0x11C92, "V"),
+        (0x11CA8, "X"),
+        (0x11CA9, "V"),
+        (0x11CB7, "X"),
+        (0x11D00, "V"),
+        (0x11D07, "X"),
+        (0x11D08, "V"),
+        (0x11D0A, "X"),
+        (0x11D0B, "V"),
+        (0x11D37, "X"),
+        (0x11D3A, "V"),
+        (0x11D3B, "X"),
+        (0x11D3C, "V"),
+        (0x11D3E, "X"),
+        (0x11D3F, "V"),
+        (0x11D48, "X"),
+        (0x11D50, "V"),
+        (0x11D5A, "X"),
+        (0x11D60, "V"),
+        (0x11D66, "X"),
+        (0x11D67, "V"),
+        (0x11D69, "X"),
+        (0x11D6A, "V"),
+        (0x11D8F, "X"),
+        (0x11D90, "V"),
+        (0x11D92, "X"),
+        (0x11D93, "V"),
+        (0x11D99, "X"),
+        (0x11DA0, "V"),
+        (0x11DAA, "X"),
+        (0x11EE0, "V"),
+        (0x11EF9, "X"),
+        (0x11F00, "V"),
+        (0x11F11, "X"),
+        (0x11F12, "V"),
+        (0x11F3B, "X"),
+        (0x11F3E, "V"),
+    ]
+
+
+def _seg_59() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x11F5A, "X"),
+        (0x11FB0, "V"),
+        (0x11FB1, "X"),
+        (0x11FC0, "V"),
+        (0x11FF2, "X"),
+        (0x11FFF, "V"),
+        (0x1239A, "X"),
+        (0x12400, "V"),
+        (0x1246F, "X"),
+        (0x12470, "V"),
+        (0x12475, "X"),
+        (0x12480, "V"),
+        (0x12544, "X"),
+        (0x12F90, "V"),
+        (0x12FF3, "X"),
+        (0x13000, "V"),
+        (0x13430, "X"),
+        (0x13440, "V"),
+        (0x13456, "X"),
+        (0x14400, "V"),
+        (0x14647, "X"),
+        (0x16800, "V"),
+        (0x16A39, "X"),
+        (0x16A40, "V"),
+        (0x16A5F, "X"),
+        (0x16A60, "V"),
+        (0x16A6A, "X"),
+        (0x16A6E, "V"),
+        (0x16ABF, "X"),
+        (0x16AC0, "V"),
+        (0x16ACA, "X"),
+        (0x16AD0, "V"),
+        (0x16AEE, "X"),
+        (0x16AF0, "V"),
+        (0x16AF6, "X"),
+        (0x16B00, "V"),
+        (0x16B46, "X"),
+        (0x16B50, "V"),
+        (0x16B5A, "X"),
+        (0x16B5B, "V"),
+        (0x16B62, "X"),
+        (0x16B63, "V"),
+        (0x16B78, "X"),
+        (0x16B7D, "V"),
+        (0x16B90, "X"),
+        (0x16E40, "M", "𖹠"),
+        (0x16E41, "M", "𖹡"),
+        (0x16E42, "M", "𖹢"),
+        (0x16E43, "M", "𖹣"),
+        (0x16E44, "M", "𖹤"),
+        (0x16E45, "M", "𖹥"),
+        (0x16E46, "M", "𖹦"),
+        (0x16E47, "M", "𖹧"),
+        (0x16E48, "M", "𖹨"),
+        (0x16E49, "M", "𖹩"),
+        (0x16E4A, "M", "𖹪"),
+        (0x16E4B, "M", "𖹫"),
+        (0x16E4C, "M", "𖹬"),
+        (0x16E4D, "M", "𖹭"),
+        (0x16E4E, "M", "𖹮"),
+        (0x16E4F, "M", "𖹯"),
+        (0x16E50, "M", "𖹰"),
+        (0x16E51, "M", "𖹱"),
+        (0x16E52, "M", "𖹲"),
+        (0x16E53, "M", "𖹳"),
+        (0x16E54, "M", "𖹴"),
+        (0x16E55, "M", "𖹵"),
+        (0x16E56, "M", "𖹶"),
+        (0x16E57, "M", "𖹷"),
+        (0x16E58, "M", "𖹸"),
+        (0x16E59, "M", "𖹹"),
+        (0x16E5A, "M", "𖹺"),
+        (0x16E5B, "M", "𖹻"),
+        (0x16E5C, "M", "𖹼"),
+        (0x16E5D, "M", "𖹽"),
+        (0x16E5E, "M", "𖹾"),
+        (0x16E5F, "M", "𖹿"),
+        (0x16E60, "V"),
+        (0x16E9B, "X"),
+        (0x16F00, "V"),
+        (0x16F4B, "X"),
+        (0x16F4F, "V"),
+        (0x16F88, "X"),
+        (0x16F8F, "V"),
+        (0x16FA0, "X"),
+        (0x16FE0, "V"),
+        (0x16FE5, "X"),
+        (0x16FF0, "V"),
+        (0x16FF2, "X"),
+        (0x17000, "V"),
+        (0x187F8, "X"),
+        (0x18800, "V"),
+        (0x18CD6, "X"),
+        (0x18D00, "V"),
+        (0x18D09, "X"),
+        (0x1AFF0, "V"),
+        (0x1AFF4, "X"),
+        (0x1AFF5, "V"),
+        (0x1AFFC, "X"),
+        (0x1AFFD, "V"),
+    ]
+
+
+def _seg_60() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1AFFF, "X"),
+        (0x1B000, "V"),
+        (0x1B123, "X"),
+        (0x1B132, "V"),
+        (0x1B133, "X"),
+        (0x1B150, "V"),
+        (0x1B153, "X"),
+        (0x1B155, "V"),
+        (0x1B156, "X"),
+        (0x1B164, "V"),
+        (0x1B168, "X"),
+        (0x1B170, "V"),
+        (0x1B2FC, "X"),
+        (0x1BC00, "V"),
+        (0x1BC6B, "X"),
+        (0x1BC70, "V"),
+        (0x1BC7D, "X"),
+        (0x1BC80, "V"),
+        (0x1BC89, "X"),
+        (0x1BC90, "V"),
+        (0x1BC9A, "X"),
+        (0x1BC9C, "V"),
+        (0x1BCA0, "I"),
+        (0x1BCA4, "X"),
+        (0x1CF00, "V"),
+        (0x1CF2E, "X"),
+        (0x1CF30, "V"),
+        (0x1CF47, "X"),
+        (0x1CF50, "V"),
+        (0x1CFC4, "X"),
+        (0x1D000, "V"),
+        (0x1D0F6, "X"),
+        (0x1D100, "V"),
+        (0x1D127, "X"),
+        (0x1D129, "V"),
+        (0x1D15E, "M", "𝅗𝅥"),
+        (0x1D15F, "M", "𝅘𝅥"),
+        (0x1D160, "M", "𝅘𝅥𝅮"),
+        (0x1D161, "M", "𝅘𝅥𝅯"),
+        (0x1D162, "M", "𝅘𝅥𝅰"),
+        (0x1D163, "M", "𝅘𝅥𝅱"),
+        (0x1D164, "M", "𝅘𝅥𝅲"),
+        (0x1D165, "V"),
+        (0x1D173, "X"),
+        (0x1D17B, "V"),
+        (0x1D1BB, "M", "𝆹𝅥"),
+        (0x1D1BC, "M", "𝆺𝅥"),
+        (0x1D1BD, "M", "𝆹𝅥𝅮"),
+        (0x1D1BE, "M", "𝆺𝅥𝅮"),
+        (0x1D1BF, "M", "𝆹𝅥𝅯"),
+        (0x1D1C0, "M", "𝆺𝅥𝅯"),
+        (0x1D1C1, "V"),
+        (0x1D1EB, "X"),
+        (0x1D200, "V"),
+        (0x1D246, "X"),
+        (0x1D2C0, "V"),
+        (0x1D2D4, "X"),
+        (0x1D2E0, "V"),
+        (0x1D2F4, "X"),
+        (0x1D300, "V"),
+        (0x1D357, "X"),
+        (0x1D360, "V"),
+        (0x1D379, "X"),
+        (0x1D400, "M", "a"),
+        (0x1D401, "M", "b"),
+        (0x1D402, "M", "c"),
+        (0x1D403, "M", "d"),
+        (0x1D404, "M", "e"),
+        (0x1D405, "M", "f"),
+        (0x1D406, "M", "g"),
+        (0x1D407, "M", "h"),
+        (0x1D408, "M", "i"),
+        (0x1D409, "M", "j"),
+        (0x1D40A, "M", "k"),
+        (0x1D40B, "M", "l"),
+        (0x1D40C, "M", "m"),
+        (0x1D40D, "M", "n"),
+        (0x1D40E, "M", "o"),
+        (0x1D40F, "M", "p"),
+        (0x1D410, "M", "q"),
+        (0x1D411, "M", "r"),
+        (0x1D412, "M", "s"),
+        (0x1D413, "M", "t"),
+        (0x1D414, "M", "u"),
+        (0x1D415, "M", "v"),
+        (0x1D416, "M", "w"),
+        (0x1D417, "M", "x"),
+        (0x1D418, "M", "y"),
+        (0x1D419, "M", "z"),
+        (0x1D41A, "M", "a"),
+        (0x1D41B, "M", "b"),
+        (0x1D41C, "M", "c"),
+        (0x1D41D, "M", "d"),
+        (0x1D41E, "M", "e"),
+        (0x1D41F, "M", "f"),
+        (0x1D420, "M", "g"),
+        (0x1D421, "M", "h"),
+        (0x1D422, "M", "i"),
+        (0x1D423, "M", "j"),
+        (0x1D424, "M", "k"),
+    ]
+
+
+def _seg_61() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D425, "M", "l"),
+        (0x1D426, "M", "m"),
+        (0x1D427, "M", "n"),
+        (0x1D428, "M", "o"),
+        (0x1D429, "M", "p"),
+        (0x1D42A, "M", "q"),
+        (0x1D42B, "M", "r"),
+        (0x1D42C, "M", "s"),
+        (0x1D42D, "M", "t"),
+        (0x1D42E, "M", "u"),
+        (0x1D42F, "M", "v"),
+        (0x1D430, "M", "w"),
+        (0x1D431, "M", "x"),
+        (0x1D432, "M", "y"),
+        (0x1D433, "M", "z"),
+        (0x1D434, "M", "a"),
+        (0x1D435, "M", "b"),
+        (0x1D436, "M", "c"),
+        (0x1D437, "M", "d"),
+        (0x1D438, "M", "e"),
+        (0x1D439, "M", "f"),
+        (0x1D43A, "M", "g"),
+        (0x1D43B, "M", "h"),
+        (0x1D43C, "M", "i"),
+        (0x1D43D, "M", "j"),
+        (0x1D43E, "M", "k"),
+        (0x1D43F, "M", "l"),
+        (0x1D440, "M", "m"),
+        (0x1D441, "M", "n"),
+        (0x1D442, "M", "o"),
+        (0x1D443, "M", "p"),
+        (0x1D444, "M", "q"),
+        (0x1D445, "M", "r"),
+        (0x1D446, "M", "s"),
+        (0x1D447, "M", "t"),
+        (0x1D448, "M", "u"),
+        (0x1D449, "M", "v"),
+        (0x1D44A, "M", "w"),
+        (0x1D44B, "M", "x"),
+        (0x1D44C, "M", "y"),
+        (0x1D44D, "M", "z"),
+        (0x1D44E, "M", "a"),
+        (0x1D44F, "M", "b"),
+        (0x1D450, "M", "c"),
+        (0x1D451, "M", "d"),
+        (0x1D452, "M", "e"),
+        (0x1D453, "M", "f"),
+        (0x1D454, "M", "g"),
+        (0x1D455, "X"),
+        (0x1D456, "M", "i"),
+        (0x1D457, "M", "j"),
+        (0x1D458, "M", "k"),
+        (0x1D459, "M", "l"),
+        (0x1D45A, "M", "m"),
+        (0x1D45B, "M", "n"),
+        (0x1D45C, "M", "o"),
+        (0x1D45D, "M", "p"),
+        (0x1D45E, "M", "q"),
+        (0x1D45F, "M", "r"),
+        (0x1D460, "M", "s"),
+        (0x1D461, "M", "t"),
+        (0x1D462, "M", "u"),
+        (0x1D463, "M", "v"),
+        (0x1D464, "M", "w"),
+        (0x1D465, "M", "x"),
+        (0x1D466, "M", "y"),
+        (0x1D467, "M", "z"),
+        (0x1D468, "M", "a"),
+        (0x1D469, "M", "b"),
+        (0x1D46A, "M", "c"),
+        (0x1D46B, "M", "d"),
+        (0x1D46C, "M", "e"),
+        (0x1D46D, "M", "f"),
+        (0x1D46E, "M", "g"),
+        (0x1D46F, "M", "h"),
+        (0x1D470, "M", "i"),
+        (0x1D471, "M", "j"),
+        (0x1D472, "M", "k"),
+        (0x1D473, "M", "l"),
+        (0x1D474, "M", "m"),
+        (0x1D475, "M", "n"),
+        (0x1D476, "M", "o"),
+        (0x1D477, "M", "p"),
+        (0x1D478, "M", "q"),
+        (0x1D479, "M", "r"),
+        (0x1D47A, "M", "s"),
+        (0x1D47B, "M", "t"),
+        (0x1D47C, "M", "u"),
+        (0x1D47D, "M", "v"),
+        (0x1D47E, "M", "w"),
+        (0x1D47F, "M", "x"),
+        (0x1D480, "M", "y"),
+        (0x1D481, "M", "z"),
+        (0x1D482, "M", "a"),
+        (0x1D483, "M", "b"),
+        (0x1D484, "M", "c"),
+        (0x1D485, "M", "d"),
+        (0x1D486, "M", "e"),
+        (0x1D487, "M", "f"),
+        (0x1D488, "M", "g"),
+    ]
+
+
+def _seg_62() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D489, "M", "h"),
+        (0x1D48A, "M", "i"),
+        (0x1D48B, "M", "j"),
+        (0x1D48C, "M", "k"),
+        (0x1D48D, "M", "l"),
+        (0x1D48E, "M", "m"),
+        (0x1D48F, "M", "n"),
+        (0x1D490, "M", "o"),
+        (0x1D491, "M", "p"),
+        (0x1D492, "M", "q"),
+        (0x1D493, "M", "r"),
+        (0x1D494, "M", "s"),
+        (0x1D495, "M", "t"),
+        (0x1D496, "M", "u"),
+        (0x1D497, "M", "v"),
+        (0x1D498, "M", "w"),
+        (0x1D499, "M", "x"),
+        (0x1D49A, "M", "y"),
+        (0x1D49B, "M", "z"),
+        (0x1D49C, "M", "a"),
+        (0x1D49D, "X"),
+        (0x1D49E, "M", "c"),
+        (0x1D49F, "M", "d"),
+        (0x1D4A0, "X"),
+        (0x1D4A2, "M", "g"),
+        (0x1D4A3, "X"),
+        (0x1D4A5, "M", "j"),
+        (0x1D4A6, "M", "k"),
+        (0x1D4A7, "X"),
+        (0x1D4A9, "M", "n"),
+        (0x1D4AA, "M", "o"),
+        (0x1D4AB, "M", "p"),
+        (0x1D4AC, "M", "q"),
+        (0x1D4AD, "X"),
+        (0x1D4AE, "M", "s"),
+        (0x1D4AF, "M", "t"),
+        (0x1D4B0, "M", "u"),
+        (0x1D4B1, "M", "v"),
+        (0x1D4B2, "M", "w"),
+        (0x1D4B3, "M", "x"),
+        (0x1D4B4, "M", "y"),
+        (0x1D4B5, "M", "z"),
+        (0x1D4B6, "M", "a"),
+        (0x1D4B7, "M", "b"),
+        (0x1D4B8, "M", "c"),
+        (0x1D4B9, "M", "d"),
+        (0x1D4BA, "X"),
+        (0x1D4BB, "M", "f"),
+        (0x1D4BC, "X"),
+        (0x1D4BD, "M", "h"),
+        (0x1D4BE, "M", "i"),
+        (0x1D4BF, "M", "j"),
+        (0x1D4C0, "M", "k"),
+        (0x1D4C1, "M", "l"),
+        (0x1D4C2, "M", "m"),
+        (0x1D4C3, "M", "n"),
+        (0x1D4C4, "X"),
+        (0x1D4C5, "M", "p"),
+        (0x1D4C6, "M", "q"),
+        (0x1D4C7, "M", "r"),
+        (0x1D4C8, "M", "s"),
+        (0x1D4C9, "M", "t"),
+        (0x1D4CA, "M", "u"),
+        (0x1D4CB, "M", "v"),
+        (0x1D4CC, "M", "w"),
+        (0x1D4CD, "M", "x"),
+        (0x1D4CE, "M", "y"),
+        (0x1D4CF, "M", "z"),
+        (0x1D4D0, "M", "a"),
+        (0x1D4D1, "M", "b"),
+        (0x1D4D2, "M", "c"),
+        (0x1D4D3, "M", "d"),
+        (0x1D4D4, "M", "e"),
+        (0x1D4D5, "M", "f"),
+        (0x1D4D6, "M", "g"),
+        (0x1D4D7, "M", "h"),
+        (0x1D4D8, "M", "i"),
+        (0x1D4D9, "M", "j"),
+        (0x1D4DA, "M", "k"),
+        (0x1D4DB, "M", "l"),
+        (0x1D4DC, "M", "m"),
+        (0x1D4DD, "M", "n"),
+        (0x1D4DE, "M", "o"),
+        (0x1D4DF, "M", "p"),
+        (0x1D4E0, "M", "q"),
+        (0x1D4E1, "M", "r"),
+        (0x1D4E2, "M", "s"),
+        (0x1D4E3, "M", "t"),
+        (0x1D4E4, "M", "u"),
+        (0x1D4E5, "M", "v"),
+        (0x1D4E6, "M", "w"),
+        (0x1D4E7, "M", "x"),
+        (0x1D4E8, "M", "y"),
+        (0x1D4E9, "M", "z"),
+        (0x1D4EA, "M", "a"),
+        (0x1D4EB, "M", "b"),
+        (0x1D4EC, "M", "c"),
+        (0x1D4ED, "M", "d"),
+        (0x1D4EE, "M", "e"),
+        (0x1D4EF, "M", "f"),
+    ]
+
+
+def _seg_63() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D4F0, "M", "g"),
+        (0x1D4F1, "M", "h"),
+        (0x1D4F2, "M", "i"),
+        (0x1D4F3, "M", "j"),
+        (0x1D4F4, "M", "k"),
+        (0x1D4F5, "M", "l"),
+        (0x1D4F6, "M", "m"),
+        (0x1D4F7, "M", "n"),
+        (0x1D4F8, "M", "o"),
+        (0x1D4F9, "M", "p"),
+        (0x1D4FA, "M", "q"),
+        (0x1D4FB, "M", "r"),
+        (0x1D4FC, "M", "s"),
+        (0x1D4FD, "M", "t"),
+        (0x1D4FE, "M", "u"),
+        (0x1D4FF, "M", "v"),
+        (0x1D500, "M", "w"),
+        (0x1D501, "M", "x"),
+        (0x1D502, "M", "y"),
+        (0x1D503, "M", "z"),
+        (0x1D504, "M", "a"),
+        (0x1D505, "M", "b"),
+        (0x1D506, "X"),
+        (0x1D507, "M", "d"),
+        (0x1D508, "M", "e"),
+        (0x1D509, "M", "f"),
+        (0x1D50A, "M", "g"),
+        (0x1D50B, "X"),
+        (0x1D50D, "M", "j"),
+        (0x1D50E, "M", "k"),
+        (0x1D50F, "M", "l"),
+        (0x1D510, "M", "m"),
+        (0x1D511, "M", "n"),
+        (0x1D512, "M", "o"),
+        (0x1D513, "M", "p"),
+        (0x1D514, "M", "q"),
+        (0x1D515, "X"),
+        (0x1D516, "M", "s"),
+        (0x1D517, "M", "t"),
+        (0x1D518, "M", "u"),
+        (0x1D519, "M", "v"),
+        (0x1D51A, "M", "w"),
+        (0x1D51B, "M", "x"),
+        (0x1D51C, "M", "y"),
+        (0x1D51D, "X"),
+        (0x1D51E, "M", "a"),
+        (0x1D51F, "M", "b"),
+        (0x1D520, "M", "c"),
+        (0x1D521, "M", "d"),
+        (0x1D522, "M", "e"),
+        (0x1D523, "M", "f"),
+        (0x1D524, "M", "g"),
+        (0x1D525, "M", "h"),
+        (0x1D526, "M", "i"),
+        (0x1D527, "M", "j"),
+        (0x1D528, "M", "k"),
+        (0x1D529, "M", "l"),
+        (0x1D52A, "M", "m"),
+        (0x1D52B, "M", "n"),
+        (0x1D52C, "M", "o"),
+        (0x1D52D, "M", "p"),
+        (0x1D52E, "M", "q"),
+        (0x1D52F, "M", "r"),
+        (0x1D530, "M", "s"),
+        (0x1D531, "M", "t"),
+        (0x1D532, "M", "u"),
+        (0x1D533, "M", "v"),
+        (0x1D534, "M", "w"),
+        (0x1D535, "M", "x"),
+        (0x1D536, "M", "y"),
+        (0x1D537, "M", "z"),
+        (0x1D538, "M", "a"),
+        (0x1D539, "M", "b"),
+        (0x1D53A, "X"),
+        (0x1D53B, "M", "d"),
+        (0x1D53C, "M", "e"),
+        (0x1D53D, "M", "f"),
+        (0x1D53E, "M", "g"),
+        (0x1D53F, "X"),
+        (0x1D540, "M", "i"),
+        (0x1D541, "M", "j"),
+        (0x1D542, "M", "k"),
+        (0x1D543, "M", "l"),
+        (0x1D544, "M", "m"),
+        (0x1D545, "X"),
+        (0x1D546, "M", "o"),
+        (0x1D547, "X"),
+        (0x1D54A, "M", "s"),
+        (0x1D54B, "M", "t"),
+        (0x1D54C, "M", "u"),
+        (0x1D54D, "M", "v"),
+        (0x1D54E, "M", "w"),
+        (0x1D54F, "M", "x"),
+        (0x1D550, "M", "y"),
+        (0x1D551, "X"),
+        (0x1D552, "M", "a"),
+        (0x1D553, "M", "b"),
+        (0x1D554, "M", "c"),
+        (0x1D555, "M", "d"),
+        (0x1D556, "M", "e"),
+    ]
+
+
+def _seg_64() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D557, "M", "f"),
+        (0x1D558, "M", "g"),
+        (0x1D559, "M", "h"),
+        (0x1D55A, "M", "i"),
+        (0x1D55B, "M", "j"),
+        (0x1D55C, "M", "k"),
+        (0x1D55D, "M", "l"),
+        (0x1D55E, "M", "m"),
+        (0x1D55F, "M", "n"),
+        (0x1D560, "M", "o"),
+        (0x1D561, "M", "p"),
+        (0x1D562, "M", "q"),
+        (0x1D563, "M", "r"),
+        (0x1D564, "M", "s"),
+        (0x1D565, "M", "t"),
+        (0x1D566, "M", "u"),
+        (0x1D567, "M", "v"),
+        (0x1D568, "M", "w"),
+        (0x1D569, "M", "x"),
+        (0x1D56A, "M", "y"),
+        (0x1D56B, "M", "z"),
+        (0x1D56C, "M", "a"),
+        (0x1D56D, "M", "b"),
+        (0x1D56E, "M", "c"),
+        (0x1D56F, "M", "d"),
+        (0x1D570, "M", "e"),
+        (0x1D571, "M", "f"),
+        (0x1D572, "M", "g"),
+        (0x1D573, "M", "h"),
+        (0x1D574, "M", "i"),
+        (0x1D575, "M", "j"),
+        (0x1D576, "M", "k"),
+        (0x1D577, "M", "l"),
+        (0x1D578, "M", "m"),
+        (0x1D579, "M", "n"),
+        (0x1D57A, "M", "o"),
+        (0x1D57B, "M", "p"),
+        (0x1D57C, "M", "q"),
+        (0x1D57D, "M", "r"),
+        (0x1D57E, "M", "s"),
+        (0x1D57F, "M", "t"),
+        (0x1D580, "M", "u"),
+        (0x1D581, "M", "v"),
+        (0x1D582, "M", "w"),
+        (0x1D583, "M", "x"),
+        (0x1D584, "M", "y"),
+        (0x1D585, "M", "z"),
+        (0x1D586, "M", "a"),
+        (0x1D587, "M", "b"),
+        (0x1D588, "M", "c"),
+        (0x1D589, "M", "d"),
+        (0x1D58A, "M", "e"),
+        (0x1D58B, "M", "f"),
+        (0x1D58C, "M", "g"),
+        (0x1D58D, "M", "h"),
+        (0x1D58E, "M", "i"),
+        (0x1D58F, "M", "j"),
+        (0x1D590, "M", "k"),
+        (0x1D591, "M", "l"),
+        (0x1D592, "M", "m"),
+        (0x1D593, "M", "n"),
+        (0x1D594, "M", "o"),
+        (0x1D595, "M", "p"),
+        (0x1D596, "M", "q"),
+        (0x1D597, "M", "r"),
+        (0x1D598, "M", "s"),
+        (0x1D599, "M", "t"),
+        (0x1D59A, "M", "u"),
+        (0x1D59B, "M", "v"),
+        (0x1D59C, "M", "w"),
+        (0x1D59D, "M", "x"),
+        (0x1D59E, "M", "y"),
+        (0x1D59F, "M", "z"),
+        (0x1D5A0, "M", "a"),
+        (0x1D5A1, "M", "b"),
+        (0x1D5A2, "M", "c"),
+        (0x1D5A3, "M", "d"),
+        (0x1D5A4, "M", "e"),
+        (0x1D5A5, "M", "f"),
+        (0x1D5A6, "M", "g"),
+        (0x1D5A7, "M", "h"),
+        (0x1D5A8, "M", "i"),
+        (0x1D5A9, "M", "j"),
+        (0x1D5AA, "M", "k"),
+        (0x1D5AB, "M", "l"),
+        (0x1D5AC, "M", "m"),
+        (0x1D5AD, "M", "n"),
+        (0x1D5AE, "M", "o"),
+        (0x1D5AF, "M", "p"),
+        (0x1D5B0, "M", "q"),
+        (0x1D5B1, "M", "r"),
+        (0x1D5B2, "M", "s"),
+        (0x1D5B3, "M", "t"),
+        (0x1D5B4, "M", "u"),
+        (0x1D5B5, "M", "v"),
+        (0x1D5B6, "M", "w"),
+        (0x1D5B7, "M", "x"),
+        (0x1D5B8, "M", "y"),
+        (0x1D5B9, "M", "z"),
+        (0x1D5BA, "M", "a"),
+    ]
+
+
+def _seg_65() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D5BB, "M", "b"),
+        (0x1D5BC, "M", "c"),
+        (0x1D5BD, "M", "d"),
+        (0x1D5BE, "M", "e"),
+        (0x1D5BF, "M", "f"),
+        (0x1D5C0, "M", "g"),
+        (0x1D5C1, "M", "h"),
+        (0x1D5C2, "M", "i"),
+        (0x1D5C3, "M", "j"),
+        (0x1D5C4, "M", "k"),
+        (0x1D5C5, "M", "l"),
+        (0x1D5C6, "M", "m"),
+        (0x1D5C7, "M", "n"),
+        (0x1D5C8, "M", "o"),
+        (0x1D5C9, "M", "p"),
+        (0x1D5CA, "M", "q"),
+        (0x1D5CB, "M", "r"),
+        (0x1D5CC, "M", "s"),
+        (0x1D5CD, "M", "t"),
+        (0x1D5CE, "M", "u"),
+        (0x1D5CF, "M", "v"),
+        (0x1D5D0, "M", "w"),
+        (0x1D5D1, "M", "x"),
+        (0x1D5D2, "M", "y"),
+        (0x1D5D3, "M", "z"),
+        (0x1D5D4, "M", "a"),
+        (0x1D5D5, "M", "b"),
+        (0x1D5D6, "M", "c"),
+        (0x1D5D7, "M", "d"),
+        (0x1D5D8, "M", "e"),
+        (0x1D5D9, "M", "f"),
+        (0x1D5DA, "M", "g"),
+        (0x1D5DB, "M", "h"),
+        (0x1D5DC, "M", "i"),
+        (0x1D5DD, "M", "j"),
+        (0x1D5DE, "M", "k"),
+        (0x1D5DF, "M", "l"),
+        (0x1D5E0, "M", "m"),
+        (0x1D5E1, "M", "n"),
+        (0x1D5E2, "M", "o"),
+        (0x1D5E3, "M", "p"),
+        (0x1D5E4, "M", "q"),
+        (0x1D5E5, "M", "r"),
+        (0x1D5E6, "M", "s"),
+        (0x1D5E7, "M", "t"),
+        (0x1D5E8, "M", "u"),
+        (0x1D5E9, "M", "v"),
+        (0x1D5EA, "M", "w"),
+        (0x1D5EB, "M", "x"),
+        (0x1D5EC, "M", "y"),
+        (0x1D5ED, "M", "z"),
+        (0x1D5EE, "M", "a"),
+        (0x1D5EF, "M", "b"),
+        (0x1D5F0, "M", "c"),
+        (0x1D5F1, "M", "d"),
+        (0x1D5F2, "M", "e"),
+        (0x1D5F3, "M", "f"),
+        (0x1D5F4, "M", "g"),
+        (0x1D5F5, "M", "h"),
+        (0x1D5F6, "M", "i"),
+        (0x1D5F7, "M", "j"),
+        (0x1D5F8, "M", "k"),
+        (0x1D5F9, "M", "l"),
+        (0x1D5FA, "M", "m"),
+        (0x1D5FB, "M", "n"),
+        (0x1D5FC, "M", "o"),
+        (0x1D5FD, "M", "p"),
+        (0x1D5FE, "M", "q"),
+        (0x1D5FF, "M", "r"),
+        (0x1D600, "M", "s"),
+        (0x1D601, "M", "t"),
+        (0x1D602, "M", "u"),
+        (0x1D603, "M", "v"),
+        (0x1D604, "M", "w"),
+        (0x1D605, "M", "x"),
+        (0x1D606, "M", "y"),
+        (0x1D607, "M", "z"),
+        (0x1D608, "M", "a"),
+        (0x1D609, "M", "b"),
+        (0x1D60A, "M", "c"),
+        (0x1D60B, "M", "d"),
+        (0x1D60C, "M", "e"),
+        (0x1D60D, "M", "f"),
+        (0x1D60E, "M", "g"),
+        (0x1D60F, "M", "h"),
+        (0x1D610, "M", "i"),
+        (0x1D611, "M", "j"),
+        (0x1D612, "M", "k"),
+        (0x1D613, "M", "l"),
+        (0x1D614, "M", "m"),
+        (0x1D615, "M", "n"),
+        (0x1D616, "M", "o"),
+        (0x1D617, "M", "p"),
+        (0x1D618, "M", "q"),
+        (0x1D619, "M", "r"),
+        (0x1D61A, "M", "s"),
+        (0x1D61B, "M", "t"),
+        (0x1D61C, "M", "u"),
+        (0x1D61D, "M", "v"),
+        (0x1D61E, "M", "w"),
+    ]
+
+
+def _seg_66() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D61F, "M", "x"),
+        (0x1D620, "M", "y"),
+        (0x1D621, "M", "z"),
+        (0x1D622, "M", "a"),
+        (0x1D623, "M", "b"),
+        (0x1D624, "M", "c"),
+        (0x1D625, "M", "d"),
+        (0x1D626, "M", "e"),
+        (0x1D627, "M", "f"),
+        (0x1D628, "M", "g"),
+        (0x1D629, "M", "h"),
+        (0x1D62A, "M", "i"),
+        (0x1D62B, "M", "j"),
+        (0x1D62C, "M", "k"),
+        (0x1D62D, "M", "l"),
+        (0x1D62E, "M", "m"),
+        (0x1D62F, "M", "n"),
+        (0x1D630, "M", "o"),
+        (0x1D631, "M", "p"),
+        (0x1D632, "M", "q"),
+        (0x1D633, "M", "r"),
+        (0x1D634, "M", "s"),
+        (0x1D635, "M", "t"),
+        (0x1D636, "M", "u"),
+        (0x1D637, "M", "v"),
+        (0x1D638, "M", "w"),
+        (0x1D639, "M", "x"),
+        (0x1D63A, "M", "y"),
+        (0x1D63B, "M", "z"),
+        (0x1D63C, "M", "a"),
+        (0x1D63D, "M", "b"),
+        (0x1D63E, "M", "c"),
+        (0x1D63F, "M", "d"),
+        (0x1D640, "M", "e"),
+        (0x1D641, "M", "f"),
+        (0x1D642, "M", "g"),
+        (0x1D643, "M", "h"),
+        (0x1D644, "M", "i"),
+        (0x1D645, "M", "j"),
+        (0x1D646, "M", "k"),
+        (0x1D647, "M", "l"),
+        (0x1D648, "M", "m"),
+        (0x1D649, "M", "n"),
+        (0x1D64A, "M", "o"),
+        (0x1D64B, "M", "p"),
+        (0x1D64C, "M", "q"),
+        (0x1D64D, "M", "r"),
+        (0x1D64E, "M", "s"),
+        (0x1D64F, "M", "t"),
+        (0x1D650, "M", "u"),
+        (0x1D651, "M", "v"),
+        (0x1D652, "M", "w"),
+        (0x1D653, "M", "x"),
+        (0x1D654, "M", "y"),
+        (0x1D655, "M", "z"),
+        (0x1D656, "M", "a"),
+        (0x1D657, "M", "b"),
+        (0x1D658, "M", "c"),
+        (0x1D659, "M", "d"),
+        (0x1D65A, "M", "e"),
+        (0x1D65B, "M", "f"),
+        (0x1D65C, "M", "g"),
+        (0x1D65D, "M", "h"),
+        (0x1D65E, "M", "i"),
+        (0x1D65F, "M", "j"),
+        (0x1D660, "M", "k"),
+        (0x1D661, "M", "l"),
+        (0x1D662, "M", "m"),
+        (0x1D663, "M", "n"),
+        (0x1D664, "M", "o"),
+        (0x1D665, "M", "p"),
+        (0x1D666, "M", "q"),
+        (0x1D667, "M", "r"),
+        (0x1D668, "M", "s"),
+        (0x1D669, "M", "t"),
+        (0x1D66A, "M", "u"),
+        (0x1D66B, "M", "v"),
+        (0x1D66C, "M", "w"),
+        (0x1D66D, "M", "x"),
+        (0x1D66E, "M", "y"),
+        (0x1D66F, "M", "z"),
+        (0x1D670, "M", "a"),
+        (0x1D671, "M", "b"),
+        (0x1D672, "M", "c"),
+        (0x1D673, "M", "d"),
+        (0x1D674, "M", "e"),
+        (0x1D675, "M", "f"),
+        (0x1D676, "M", "g"),
+        (0x1D677, "M", "h"),
+        (0x1D678, "M", "i"),
+        (0x1D679, "M", "j"),
+        (0x1D67A, "M", "k"),
+        (0x1D67B, "M", "l"),
+        (0x1D67C, "M", "m"),
+        (0x1D67D, "M", "n"),
+        (0x1D67E, "M", "o"),
+        (0x1D67F, "M", "p"),
+        (0x1D680, "M", "q"),
+        (0x1D681, "M", "r"),
+        (0x1D682, "M", "s"),
+    ]
+
+
+def _seg_67() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D683, "M", "t"),
+        (0x1D684, "M", "u"),
+        (0x1D685, "M", "v"),
+        (0x1D686, "M", "w"),
+        (0x1D687, "M", "x"),
+        (0x1D688, "M", "y"),
+        (0x1D689, "M", "z"),
+        (0x1D68A, "M", "a"),
+        (0x1D68B, "M", "b"),
+        (0x1D68C, "M", "c"),
+        (0x1D68D, "M", "d"),
+        (0x1D68E, "M", "e"),
+        (0x1D68F, "M", "f"),
+        (0x1D690, "M", "g"),
+        (0x1D691, "M", "h"),
+        (0x1D692, "M", "i"),
+        (0x1D693, "M", "j"),
+        (0x1D694, "M", "k"),
+        (0x1D695, "M", "l"),
+        (0x1D696, "M", "m"),
+        (0x1D697, "M", "n"),
+        (0x1D698, "M", "o"),
+        (0x1D699, "M", "p"),
+        (0x1D69A, "M", "q"),
+        (0x1D69B, "M", "r"),
+        (0x1D69C, "M", "s"),
+        (0x1D69D, "M", "t"),
+        (0x1D69E, "M", "u"),
+        (0x1D69F, "M", "v"),
+        (0x1D6A0, "M", "w"),
+        (0x1D6A1, "M", "x"),
+        (0x1D6A2, "M", "y"),
+        (0x1D6A3, "M", "z"),
+        (0x1D6A4, "M", "ı"),
+        (0x1D6A5, "M", "ȷ"),
+        (0x1D6A6, "X"),
+        (0x1D6A8, "M", "α"),
+        (0x1D6A9, "M", "β"),
+        (0x1D6AA, "M", "γ"),
+        (0x1D6AB, "M", "δ"),
+        (0x1D6AC, "M", "ε"),
+        (0x1D6AD, "M", "ζ"),
+        (0x1D6AE, "M", "η"),
+        (0x1D6AF, "M", "θ"),
+        (0x1D6B0, "M", "ι"),
+        (0x1D6B1, "M", "κ"),
+        (0x1D6B2, "M", "λ"),
+        (0x1D6B3, "M", "μ"),
+        (0x1D6B4, "M", "ν"),
+        (0x1D6B5, "M", "ξ"),
+        (0x1D6B6, "M", "ο"),
+        (0x1D6B7, "M", "π"),
+        (0x1D6B8, "M", "ρ"),
+        (0x1D6B9, "M", "θ"),
+        (0x1D6BA, "M", "σ"),
+        (0x1D6BB, "M", "τ"),
+        (0x1D6BC, "M", "υ"),
+        (0x1D6BD, "M", "φ"),
+        (0x1D6BE, "M", "χ"),
+        (0x1D6BF, "M", "ψ"),
+        (0x1D6C0, "M", "ω"),
+        (0x1D6C1, "M", "∇"),
+        (0x1D6C2, "M", "α"),
+        (0x1D6C3, "M", "β"),
+        (0x1D6C4, "M", "γ"),
+        (0x1D6C5, "M", "δ"),
+        (0x1D6C6, "M", "ε"),
+        (0x1D6C7, "M", "ζ"),
+        (0x1D6C8, "M", "η"),
+        (0x1D6C9, "M", "θ"),
+        (0x1D6CA, "M", "ι"),
+        (0x1D6CB, "M", "κ"),
+        (0x1D6CC, "M", "λ"),
+        (0x1D6CD, "M", "μ"),
+        (0x1D6CE, "M", "ν"),
+        (0x1D6CF, "M", "ξ"),
+        (0x1D6D0, "M", "ο"),
+        (0x1D6D1, "M", "π"),
+        (0x1D6D2, "M", "ρ"),
+        (0x1D6D3, "M", "σ"),
+        (0x1D6D5, "M", "τ"),
+        (0x1D6D6, "M", "υ"),
+        (0x1D6D7, "M", "φ"),
+        (0x1D6D8, "M", "χ"),
+        (0x1D6D9, "M", "ψ"),
+        (0x1D6DA, "M", "ω"),
+        (0x1D6DB, "M", "∂"),
+        (0x1D6DC, "M", "ε"),
+        (0x1D6DD, "M", "θ"),
+        (0x1D6DE, "M", "κ"),
+        (0x1D6DF, "M", "φ"),
+        (0x1D6E0, "M", "ρ"),
+        (0x1D6E1, "M", "π"),
+        (0x1D6E2, "M", "α"),
+        (0x1D6E3, "M", "β"),
+        (0x1D6E4, "M", "γ"),
+        (0x1D6E5, "M", "δ"),
+        (0x1D6E6, "M", "ε"),
+        (0x1D6E7, "M", "ζ"),
+        (0x1D6E8, "M", "η"),
+    ]
+
+
+def _seg_68() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D6E9, "M", "θ"),
+        (0x1D6EA, "M", "ι"),
+        (0x1D6EB, "M", "κ"),
+        (0x1D6EC, "M", "λ"),
+        (0x1D6ED, "M", "μ"),
+        (0x1D6EE, "M", "ν"),
+        (0x1D6EF, "M", "ξ"),
+        (0x1D6F0, "M", "ο"),
+        (0x1D6F1, "M", "π"),
+        (0x1D6F2, "M", "ρ"),
+        (0x1D6F3, "M", "θ"),
+        (0x1D6F4, "M", "σ"),
+        (0x1D6F5, "M", "τ"),
+        (0x1D6F6, "M", "υ"),
+        (0x1D6F7, "M", "φ"),
+        (0x1D6F8, "M", "χ"),
+        (0x1D6F9, "M", "ψ"),
+        (0x1D6FA, "M", "ω"),
+        (0x1D6FB, "M", "∇"),
+        (0x1D6FC, "M", "α"),
+        (0x1D6FD, "M", "β"),
+        (0x1D6FE, "M", "γ"),
+        (0x1D6FF, "M", "δ"),
+        (0x1D700, "M", "ε"),
+        (0x1D701, "M", "ζ"),
+        (0x1D702, "M", "η"),
+        (0x1D703, "M", "θ"),
+        (0x1D704, "M", "ι"),
+        (0x1D705, "M", "κ"),
+        (0x1D706, "M", "λ"),
+        (0x1D707, "M", "μ"),
+        (0x1D708, "M", "ν"),
+        (0x1D709, "M", "ξ"),
+        (0x1D70A, "M", "ο"),
+        (0x1D70B, "M", "π"),
+        (0x1D70C, "M", "ρ"),
+        (0x1D70D, "M", "σ"),
+        (0x1D70F, "M", "τ"),
+        (0x1D710, "M", "υ"),
+        (0x1D711, "M", "φ"),
+        (0x1D712, "M", "χ"),
+        (0x1D713, "M", "ψ"),
+        (0x1D714, "M", "ω"),
+        (0x1D715, "M", "∂"),
+        (0x1D716, "M", "ε"),
+        (0x1D717, "M", "θ"),
+        (0x1D718, "M", "κ"),
+        (0x1D719, "M", "φ"),
+        (0x1D71A, "M", "ρ"),
+        (0x1D71B, "M", "π"),
+        (0x1D71C, "M", "α"),
+        (0x1D71D, "M", "β"),
+        (0x1D71E, "M", "γ"),
+        (0x1D71F, "M", "δ"),
+        (0x1D720, "M", "ε"),
+        (0x1D721, "M", "ζ"),
+        (0x1D722, "M", "η"),
+        (0x1D723, "M", "θ"),
+        (0x1D724, "M", "ι"),
+        (0x1D725, "M", "κ"),
+        (0x1D726, "M", "λ"),
+        (0x1D727, "M", "μ"),
+        (0x1D728, "M", "ν"),
+        (0x1D729, "M", "ξ"),
+        (0x1D72A, "M", "ο"),
+        (0x1D72B, "M", "π"),
+        (0x1D72C, "M", "ρ"),
+        (0x1D72D, "M", "θ"),
+        (0x1D72E, "M", "σ"),
+        (0x1D72F, "M", "τ"),
+        (0x1D730, "M", "υ"),
+        (0x1D731, "M", "φ"),
+        (0x1D732, "M", "χ"),
+        (0x1D733, "M", "ψ"),
+        (0x1D734, "M", "ω"),
+        (0x1D735, "M", "∇"),
+        (0x1D736, "M", "α"),
+        (0x1D737, "M", "β"),
+        (0x1D738, "M", "γ"),
+        (0x1D739, "M", "δ"),
+        (0x1D73A, "M", "ε"),
+        (0x1D73B, "M", "ζ"),
+        (0x1D73C, "M", "η"),
+        (0x1D73D, "M", "θ"),
+        (0x1D73E, "M", "ι"),
+        (0x1D73F, "M", "κ"),
+        (0x1D740, "M", "λ"),
+        (0x1D741, "M", "μ"),
+        (0x1D742, "M", "ν"),
+        (0x1D743, "M", "ξ"),
+        (0x1D744, "M", "ο"),
+        (0x1D745, "M", "π"),
+        (0x1D746, "M", "ρ"),
+        (0x1D747, "M", "σ"),
+        (0x1D749, "M", "τ"),
+        (0x1D74A, "M", "υ"),
+        (0x1D74B, "M", "φ"),
+        (0x1D74C, "M", "χ"),
+        (0x1D74D, "M", "ψ"),
+        (0x1D74E, "M", "ω"),
+    ]
+
+
+def _seg_69() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D74F, "M", "∂"),
+        (0x1D750, "M", "ε"),
+        (0x1D751, "M", "θ"),
+        (0x1D752, "M", "κ"),
+        (0x1D753, "M", "φ"),
+        (0x1D754, "M", "ρ"),
+        (0x1D755, "M", "π"),
+        (0x1D756, "M", "α"),
+        (0x1D757, "M", "β"),
+        (0x1D758, "M", "γ"),
+        (0x1D759, "M", "δ"),
+        (0x1D75A, "M", "ε"),
+        (0x1D75B, "M", "ζ"),
+        (0x1D75C, "M", "η"),
+        (0x1D75D, "M", "θ"),
+        (0x1D75E, "M", "ι"),
+        (0x1D75F, "M", "κ"),
+        (0x1D760, "M", "λ"),
+        (0x1D761, "M", "μ"),
+        (0x1D762, "M", "ν"),
+        (0x1D763, "M", "ξ"),
+        (0x1D764, "M", "ο"),
+        (0x1D765, "M", "π"),
+        (0x1D766, "M", "ρ"),
+        (0x1D767, "M", "θ"),
+        (0x1D768, "M", "σ"),
+        (0x1D769, "M", "τ"),
+        (0x1D76A, "M", "υ"),
+        (0x1D76B, "M", "φ"),
+        (0x1D76C, "M", "χ"),
+        (0x1D76D, "M", "ψ"),
+        (0x1D76E, "M", "ω"),
+        (0x1D76F, "M", "∇"),
+        (0x1D770, "M", "α"),
+        (0x1D771, "M", "β"),
+        (0x1D772, "M", "γ"),
+        (0x1D773, "M", "δ"),
+        (0x1D774, "M", "ε"),
+        (0x1D775, "M", "ζ"),
+        (0x1D776, "M", "η"),
+        (0x1D777, "M", "θ"),
+        (0x1D778, "M", "ι"),
+        (0x1D779, "M", "κ"),
+        (0x1D77A, "M", "λ"),
+        (0x1D77B, "M", "μ"),
+        (0x1D77C, "M", "ν"),
+        (0x1D77D, "M", "ξ"),
+        (0x1D77E, "M", "ο"),
+        (0x1D77F, "M", "π"),
+        (0x1D780, "M", "ρ"),
+        (0x1D781, "M", "σ"),
+        (0x1D783, "M", "τ"),
+        (0x1D784, "M", "υ"),
+        (0x1D785, "M", "φ"),
+        (0x1D786, "M", "χ"),
+        (0x1D787, "M", "ψ"),
+        (0x1D788, "M", "ω"),
+        (0x1D789, "M", "∂"),
+        (0x1D78A, "M", "ε"),
+        (0x1D78B, "M", "θ"),
+        (0x1D78C, "M", "κ"),
+        (0x1D78D, "M", "φ"),
+        (0x1D78E, "M", "ρ"),
+        (0x1D78F, "M", "π"),
+        (0x1D790, "M", "α"),
+        (0x1D791, "M", "β"),
+        (0x1D792, "M", "γ"),
+        (0x1D793, "M", "δ"),
+        (0x1D794, "M", "ε"),
+        (0x1D795, "M", "ζ"),
+        (0x1D796, "M", "η"),
+        (0x1D797, "M", "θ"),
+        (0x1D798, "M", "ι"),
+        (0x1D799, "M", "κ"),
+        (0x1D79A, "M", "λ"),
+        (0x1D79B, "M", "μ"),
+        (0x1D79C, "M", "ν"),
+        (0x1D79D, "M", "ξ"),
+        (0x1D79E, "M", "ο"),
+        (0x1D79F, "M", "π"),
+        (0x1D7A0, "M", "ρ"),
+        (0x1D7A1, "M", "θ"),
+        (0x1D7A2, "M", "σ"),
+        (0x1D7A3, "M", "τ"),
+        (0x1D7A4, "M", "υ"),
+        (0x1D7A5, "M", "φ"),
+        (0x1D7A6, "M", "χ"),
+        (0x1D7A7, "M", "ψ"),
+        (0x1D7A8, "M", "ω"),
+        (0x1D7A9, "M", "∇"),
+        (0x1D7AA, "M", "α"),
+        (0x1D7AB, "M", "β"),
+        (0x1D7AC, "M", "γ"),
+        (0x1D7AD, "M", "δ"),
+        (0x1D7AE, "M", "ε"),
+        (0x1D7AF, "M", "ζ"),
+        (0x1D7B0, "M", "η"),
+        (0x1D7B1, "M", "θ"),
+        (0x1D7B2, "M", "ι"),
+        (0x1D7B3, "M", "κ"),
+    ]
+
+
+def _seg_70() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1D7B4, "M", "λ"),
+        (0x1D7B5, "M", "μ"),
+        (0x1D7B6, "M", "ν"),
+        (0x1D7B7, "M", "ξ"),
+        (0x1D7B8, "M", "ο"),
+        (0x1D7B9, "M", "π"),
+        (0x1D7BA, "M", "ρ"),
+        (0x1D7BB, "M", "σ"),
+        (0x1D7BD, "M", "τ"),
+        (0x1D7BE, "M", "υ"),
+        (0x1D7BF, "M", "φ"),
+        (0x1D7C0, "M", "χ"),
+        (0x1D7C1, "M", "ψ"),
+        (0x1D7C2, "M", "ω"),
+        (0x1D7C3, "M", "∂"),
+        (0x1D7C4, "M", "ε"),
+        (0x1D7C5, "M", "θ"),
+        (0x1D7C6, "M", "κ"),
+        (0x1D7C7, "M", "φ"),
+        (0x1D7C8, "M", "ρ"),
+        (0x1D7C9, "M", "π"),
+        (0x1D7CA, "M", "ϝ"),
+        (0x1D7CC, "X"),
+        (0x1D7CE, "M", "0"),
+        (0x1D7CF, "M", "1"),
+        (0x1D7D0, "M", "2"),
+        (0x1D7D1, "M", "3"),
+        (0x1D7D2, "M", "4"),
+        (0x1D7D3, "M", "5"),
+        (0x1D7D4, "M", "6"),
+        (0x1D7D5, "M", "7"),
+        (0x1D7D6, "M", "8"),
+        (0x1D7D7, "M", "9"),
+        (0x1D7D8, "M", "0"),
+        (0x1D7D9, "M", "1"),
+        (0x1D7DA, "M", "2"),
+        (0x1D7DB, "M", "3"),
+        (0x1D7DC, "M", "4"),
+        (0x1D7DD, "M", "5"),
+        (0x1D7DE, "M", "6"),
+        (0x1D7DF, "M", "7"),
+        (0x1D7E0, "M", "8"),
+        (0x1D7E1, "M", "9"),
+        (0x1D7E2, "M", "0"),
+        (0x1D7E3, "M", "1"),
+        (0x1D7E4, "M", "2"),
+        (0x1D7E5, "M", "3"),
+        (0x1D7E6, "M", "4"),
+        (0x1D7E7, "M", "5"),
+        (0x1D7E8, "M", "6"),
+        (0x1D7E9, "M", "7"),
+        (0x1D7EA, "M", "8"),
+        (0x1D7EB, "M", "9"),
+        (0x1D7EC, "M", "0"),
+        (0x1D7ED, "M", "1"),
+        (0x1D7EE, "M", "2"),
+        (0x1D7EF, "M", "3"),
+        (0x1D7F0, "M", "4"),
+        (0x1D7F1, "M", "5"),
+        (0x1D7F2, "M", "6"),
+        (0x1D7F3, "M", "7"),
+        (0x1D7F4, "M", "8"),
+        (0x1D7F5, "M", "9"),
+        (0x1D7F6, "M", "0"),
+        (0x1D7F7, "M", "1"),
+        (0x1D7F8, "M", "2"),
+        (0x1D7F9, "M", "3"),
+        (0x1D7FA, "M", "4"),
+        (0x1D7FB, "M", "5"),
+        (0x1D7FC, "M", "6"),
+        (0x1D7FD, "M", "7"),
+        (0x1D7FE, "M", "8"),
+        (0x1D7FF, "M", "9"),
+        (0x1D800, "V"),
+        (0x1DA8C, "X"),
+        (0x1DA9B, "V"),
+        (0x1DAA0, "X"),
+        (0x1DAA1, "V"),
+        (0x1DAB0, "X"),
+        (0x1DF00, "V"),
+        (0x1DF1F, "X"),
+        (0x1DF25, "V"),
+        (0x1DF2B, "X"),
+        (0x1E000, "V"),
+        (0x1E007, "X"),
+        (0x1E008, "V"),
+        (0x1E019, "X"),
+        (0x1E01B, "V"),
+        (0x1E022, "X"),
+        (0x1E023, "V"),
+        (0x1E025, "X"),
+        (0x1E026, "V"),
+        (0x1E02B, "X"),
+        (0x1E030, "M", "а"),
+        (0x1E031, "M", "б"),
+        (0x1E032, "M", "в"),
+        (0x1E033, "M", "г"),
+        (0x1E034, "M", "д"),
+        (0x1E035, "M", "е"),
+        (0x1E036, "M", "ж"),
+    ]
+
+
+def _seg_71() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E037, "M", "з"),
+        (0x1E038, "M", "и"),
+        (0x1E039, "M", "к"),
+        (0x1E03A, "M", "л"),
+        (0x1E03B, "M", "м"),
+        (0x1E03C, "M", "о"),
+        (0x1E03D, "M", "п"),
+        (0x1E03E, "M", "р"),
+        (0x1E03F, "M", "с"),
+        (0x1E040, "M", "т"),
+        (0x1E041, "M", "у"),
+        (0x1E042, "M", "ф"),
+        (0x1E043, "M", "х"),
+        (0x1E044, "M", "ц"),
+        (0x1E045, "M", "ч"),
+        (0x1E046, "M", "ш"),
+        (0x1E047, "M", "ы"),
+        (0x1E048, "M", "э"),
+        (0x1E049, "M", "ю"),
+        (0x1E04A, "M", "ꚉ"),
+        (0x1E04B, "M", "ә"),
+        (0x1E04C, "M", "і"),
+        (0x1E04D, "M", "ј"),
+        (0x1E04E, "M", "ө"),
+        (0x1E04F, "M", "ү"),
+        (0x1E050, "M", "ӏ"),
+        (0x1E051, "M", "а"),
+        (0x1E052, "M", "б"),
+        (0x1E053, "M", "в"),
+        (0x1E054, "M", "г"),
+        (0x1E055, "M", "д"),
+        (0x1E056, "M", "е"),
+        (0x1E057, "M", "ж"),
+        (0x1E058, "M", "з"),
+        (0x1E059, "M", "и"),
+        (0x1E05A, "M", "к"),
+        (0x1E05B, "M", "л"),
+        (0x1E05C, "M", "о"),
+        (0x1E05D, "M", "п"),
+        (0x1E05E, "M", "с"),
+        (0x1E05F, "M", "у"),
+        (0x1E060, "M", "ф"),
+        (0x1E061, "M", "х"),
+        (0x1E062, "M", "ц"),
+        (0x1E063, "M", "ч"),
+        (0x1E064, "M", "ш"),
+        (0x1E065, "M", "ъ"),
+        (0x1E066, "M", "ы"),
+        (0x1E067, "M", "ґ"),
+        (0x1E068, "M", "і"),
+        (0x1E069, "M", "ѕ"),
+        (0x1E06A, "M", "џ"),
+        (0x1E06B, "M", "ҫ"),
+        (0x1E06C, "M", "ꙑ"),
+        (0x1E06D, "M", "ұ"),
+        (0x1E06E, "X"),
+        (0x1E08F, "V"),
+        (0x1E090, "X"),
+        (0x1E100, "V"),
+        (0x1E12D, "X"),
+        (0x1E130, "V"),
+        (0x1E13E, "X"),
+        (0x1E140, "V"),
+        (0x1E14A, "X"),
+        (0x1E14E, "V"),
+        (0x1E150, "X"),
+        (0x1E290, "V"),
+        (0x1E2AF, "X"),
+        (0x1E2C0, "V"),
+        (0x1E2FA, "X"),
+        (0x1E2FF, "V"),
+        (0x1E300, "X"),
+        (0x1E4D0, "V"),
+        (0x1E4FA, "X"),
+        (0x1E7E0, "V"),
+        (0x1E7E7, "X"),
+        (0x1E7E8, "V"),
+        (0x1E7EC, "X"),
+        (0x1E7ED, "V"),
+        (0x1E7EF, "X"),
+        (0x1E7F0, "V"),
+        (0x1E7FF, "X"),
+        (0x1E800, "V"),
+        (0x1E8C5, "X"),
+        (0x1E8C7, "V"),
+        (0x1E8D7, "X"),
+        (0x1E900, "M", "𞤢"),
+        (0x1E901, "M", "𞤣"),
+        (0x1E902, "M", "𞤤"),
+        (0x1E903, "M", "𞤥"),
+        (0x1E904, "M", "𞤦"),
+        (0x1E905, "M", "𞤧"),
+        (0x1E906, "M", "𞤨"),
+        (0x1E907, "M", "𞤩"),
+        (0x1E908, "M", "𞤪"),
+        (0x1E909, "M", "𞤫"),
+        (0x1E90A, "M", "𞤬"),
+        (0x1E90B, "M", "𞤭"),
+        (0x1E90C, "M", "𞤮"),
+        (0x1E90D, "M", "𞤯"),
+    ]
+
+
+def _seg_72() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1E90E, "M", "𞤰"),
+        (0x1E90F, "M", "𞤱"),
+        (0x1E910, "M", "𞤲"),
+        (0x1E911, "M", "𞤳"),
+        (0x1E912, "M", "𞤴"),
+        (0x1E913, "M", "𞤵"),
+        (0x1E914, "M", "𞤶"),
+        (0x1E915, "M", "𞤷"),
+        (0x1E916, "M", "𞤸"),
+        (0x1E917, "M", "𞤹"),
+        (0x1E918, "M", "𞤺"),
+        (0x1E919, "M", "𞤻"),
+        (0x1E91A, "M", "𞤼"),
+        (0x1E91B, "M", "𞤽"),
+        (0x1E91C, "M", "𞤾"),
+        (0x1E91D, "M", "𞤿"),
+        (0x1E91E, "M", "𞥀"),
+        (0x1E91F, "M", "𞥁"),
+        (0x1E920, "M", "𞥂"),
+        (0x1E921, "M", "𞥃"),
+        (0x1E922, "V"),
+        (0x1E94C, "X"),
+        (0x1E950, "V"),
+        (0x1E95A, "X"),
+        (0x1E95E, "V"),
+        (0x1E960, "X"),
+        (0x1EC71, "V"),
+        (0x1ECB5, "X"),
+        (0x1ED01, "V"),
+        (0x1ED3E, "X"),
+        (0x1EE00, "M", "ا"),
+        (0x1EE01, "M", "ب"),
+        (0x1EE02, "M", "ج"),
+        (0x1EE03, "M", "د"),
+        (0x1EE04, "X"),
+        (0x1EE05, "M", "و"),
+        (0x1EE06, "M", "ز"),
+        (0x1EE07, "M", "ح"),
+        (0x1EE08, "M", "ط"),
+        (0x1EE09, "M", "ي"),
+        (0x1EE0A, "M", "ك"),
+        (0x1EE0B, "M", "ل"),
+        (0x1EE0C, "M", "م"),
+        (0x1EE0D, "M", "ن"),
+        (0x1EE0E, "M", "س"),
+        (0x1EE0F, "M", "ع"),
+        (0x1EE10, "M", "ف"),
+        (0x1EE11, "M", "ص"),
+        (0x1EE12, "M", "ق"),
+        (0x1EE13, "M", "ر"),
+        (0x1EE14, "M", "ش"),
+        (0x1EE15, "M", "ت"),
+        (0x1EE16, "M", "ث"),
+        (0x1EE17, "M", "خ"),
+        (0x1EE18, "M", "ذ"),
+        (0x1EE19, "M", "ض"),
+        (0x1EE1A, "M", "ظ"),
+        (0x1EE1B, "M", "غ"),
+        (0x1EE1C, "M", "ٮ"),
+        (0x1EE1D, "M", "ں"),
+        (0x1EE1E, "M", "ڡ"),
+        (0x1EE1F, "M", "ٯ"),
+        (0x1EE20, "X"),
+        (0x1EE21, "M", "ب"),
+        (0x1EE22, "M", "ج"),
+        (0x1EE23, "X"),
+        (0x1EE24, "M", "ه"),
+        (0x1EE25, "X"),
+        (0x1EE27, "M", "ح"),
+        (0x1EE28, "X"),
+        (0x1EE29, "M", "ي"),
+        (0x1EE2A, "M", "ك"),
+        (0x1EE2B, "M", "ل"),
+        (0x1EE2C, "M", "م"),
+        (0x1EE2D, "M", "ن"),
+        (0x1EE2E, "M", "س"),
+        (0x1EE2F, "M", "ع"),
+        (0x1EE30, "M", "ف"),
+        (0x1EE31, "M", "ص"),
+        (0x1EE32, "M", "ق"),
+        (0x1EE33, "X"),
+        (0x1EE34, "M", "ش"),
+        (0x1EE35, "M", "ت"),
+        (0x1EE36, "M", "ث"),
+        (0x1EE37, "M", "خ"),
+        (0x1EE38, "X"),
+        (0x1EE39, "M", "ض"),
+        (0x1EE3A, "X"),
+        (0x1EE3B, "M", "غ"),
+        (0x1EE3C, "X"),
+        (0x1EE42, "M", "ج"),
+        (0x1EE43, "X"),
+        (0x1EE47, "M", "ح"),
+        (0x1EE48, "X"),
+        (0x1EE49, "M", "ي"),
+        (0x1EE4A, "X"),
+        (0x1EE4B, "M", "ل"),
+        (0x1EE4C, "X"),
+        (0x1EE4D, "M", "ن"),
+        (0x1EE4E, "M", "س"),
+    ]
+
+
+def _seg_73() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EE4F, "M", "ع"),
+        (0x1EE50, "X"),
+        (0x1EE51, "M", "ص"),
+        (0x1EE52, "M", "ق"),
+        (0x1EE53, "X"),
+        (0x1EE54, "M", "ش"),
+        (0x1EE55, "X"),
+        (0x1EE57, "M", "خ"),
+        (0x1EE58, "X"),
+        (0x1EE59, "M", "ض"),
+        (0x1EE5A, "X"),
+        (0x1EE5B, "M", "غ"),
+        (0x1EE5C, "X"),
+        (0x1EE5D, "M", "ں"),
+        (0x1EE5E, "X"),
+        (0x1EE5F, "M", "ٯ"),
+        (0x1EE60, "X"),
+        (0x1EE61, "M", "ب"),
+        (0x1EE62, "M", "ج"),
+        (0x1EE63, "X"),
+        (0x1EE64, "M", "ه"),
+        (0x1EE65, "X"),
+        (0x1EE67, "M", "ح"),
+        (0x1EE68, "M", "ط"),
+        (0x1EE69, "M", "ي"),
+        (0x1EE6A, "M", "ك"),
+        (0x1EE6B, "X"),
+        (0x1EE6C, "M", "م"),
+        (0x1EE6D, "M", "ن"),
+        (0x1EE6E, "M", "س"),
+        (0x1EE6F, "M", "ع"),
+        (0x1EE70, "M", "ف"),
+        (0x1EE71, "M", "ص"),
+        (0x1EE72, "M", "ق"),
+        (0x1EE73, "X"),
+        (0x1EE74, "M", "ش"),
+        (0x1EE75, "M", "ت"),
+        (0x1EE76, "M", "ث"),
+        (0x1EE77, "M", "خ"),
+        (0x1EE78, "X"),
+        (0x1EE79, "M", "ض"),
+        (0x1EE7A, "M", "ظ"),
+        (0x1EE7B, "M", "غ"),
+        (0x1EE7C, "M", "ٮ"),
+        (0x1EE7D, "X"),
+        (0x1EE7E, "M", "ڡ"),
+        (0x1EE7F, "X"),
+        (0x1EE80, "M", "ا"),
+        (0x1EE81, "M", "ب"),
+        (0x1EE82, "M", "ج"),
+        (0x1EE83, "M", "د"),
+        (0x1EE84, "M", "ه"),
+        (0x1EE85, "M", "و"),
+        (0x1EE86, "M", "ز"),
+        (0x1EE87, "M", "ح"),
+        (0x1EE88, "M", "ط"),
+        (0x1EE89, "M", "ي"),
+        (0x1EE8A, "X"),
+        (0x1EE8B, "M", "ل"),
+        (0x1EE8C, "M", "م"),
+        (0x1EE8D, "M", "ن"),
+        (0x1EE8E, "M", "س"),
+        (0x1EE8F, "M", "ع"),
+        (0x1EE90, "M", "ف"),
+        (0x1EE91, "M", "ص"),
+        (0x1EE92, "M", "ق"),
+        (0x1EE93, "M", "ر"),
+        (0x1EE94, "M", "ش"),
+        (0x1EE95, "M", "ت"),
+        (0x1EE96, "M", "ث"),
+        (0x1EE97, "M", "خ"),
+        (0x1EE98, "M", "ذ"),
+        (0x1EE99, "M", "ض"),
+        (0x1EE9A, "M", "ظ"),
+        (0x1EE9B, "M", "غ"),
+        (0x1EE9C, "X"),
+        (0x1EEA1, "M", "ب"),
+        (0x1EEA2, "M", "ج"),
+        (0x1EEA3, "M", "د"),
+        (0x1EEA4, "X"),
+        (0x1EEA5, "M", "و"),
+        (0x1EEA6, "M", "ز"),
+        (0x1EEA7, "M", "ح"),
+        (0x1EEA8, "M", "ط"),
+        (0x1EEA9, "M", "ي"),
+        (0x1EEAA, "X"),
+        (0x1EEAB, "M", "ل"),
+        (0x1EEAC, "M", "م"),
+        (0x1EEAD, "M", "ن"),
+        (0x1EEAE, "M", "س"),
+        (0x1EEAF, "M", "ع"),
+        (0x1EEB0, "M", "ف"),
+        (0x1EEB1, "M", "ص"),
+        (0x1EEB2, "M", "ق"),
+        (0x1EEB3, "M", "ر"),
+        (0x1EEB4, "M", "ش"),
+        (0x1EEB5, "M", "ت"),
+        (0x1EEB6, "M", "ث"),
+        (0x1EEB7, "M", "خ"),
+        (0x1EEB8, "M", "ذ"),
+    ]
+
+
+def _seg_74() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1EEB9, "M", "ض"),
+        (0x1EEBA, "M", "ظ"),
+        (0x1EEBB, "M", "غ"),
+        (0x1EEBC, "X"),
+        (0x1EEF0, "V"),
+        (0x1EEF2, "X"),
+        (0x1F000, "V"),
+        (0x1F02C, "X"),
+        (0x1F030, "V"),
+        (0x1F094, "X"),
+        (0x1F0A0, "V"),
+        (0x1F0AF, "X"),
+        (0x1F0B1, "V"),
+        (0x1F0C0, "X"),
+        (0x1F0C1, "V"),
+        (0x1F0D0, "X"),
+        (0x1F0D1, "V"),
+        (0x1F0F6, "X"),
+        (0x1F101, "3", "0,"),
+        (0x1F102, "3", "1,"),
+        (0x1F103, "3", "2,"),
+        (0x1F104, "3", "3,"),
+        (0x1F105, "3", "4,"),
+        (0x1F106, "3", "5,"),
+        (0x1F107, "3", "6,"),
+        (0x1F108, "3", "7,"),
+        (0x1F109, "3", "8,"),
+        (0x1F10A, "3", "9,"),
+        (0x1F10B, "V"),
+        (0x1F110, "3", "(a)"),
+        (0x1F111, "3", "(b)"),
+        (0x1F112, "3", "(c)"),
+        (0x1F113, "3", "(d)"),
+        (0x1F114, "3", "(e)"),
+        (0x1F115, "3", "(f)"),
+        (0x1F116, "3", "(g)"),
+        (0x1F117, "3", "(h)"),
+        (0x1F118, "3", "(i)"),
+        (0x1F119, "3", "(j)"),
+        (0x1F11A, "3", "(k)"),
+        (0x1F11B, "3", "(l)"),
+        (0x1F11C, "3", "(m)"),
+        (0x1F11D, "3", "(n)"),
+        (0x1F11E, "3", "(o)"),
+        (0x1F11F, "3", "(p)"),
+        (0x1F120, "3", "(q)"),
+        (0x1F121, "3", "(r)"),
+        (0x1F122, "3", "(s)"),
+        (0x1F123, "3", "(t)"),
+        (0x1F124, "3", "(u)"),
+        (0x1F125, "3", "(v)"),
+        (0x1F126, "3", "(w)"),
+        (0x1F127, "3", "(x)"),
+        (0x1F128, "3", "(y)"),
+        (0x1F129, "3", "(z)"),
+        (0x1F12A, "M", "〔s〕"),
+        (0x1F12B, "M", "c"),
+        (0x1F12C, "M", "r"),
+        (0x1F12D, "M", "cd"),
+        (0x1F12E, "M", "wz"),
+        (0x1F12F, "V"),
+        (0x1F130, "M", "a"),
+        (0x1F131, "M", "b"),
+        (0x1F132, "M", "c"),
+        (0x1F133, "M", "d"),
+        (0x1F134, "M", "e"),
+        (0x1F135, "M", "f"),
+        (0x1F136, "M", "g"),
+        (0x1F137, "M", "h"),
+        (0x1F138, "M", "i"),
+        (0x1F139, "M", "j"),
+        (0x1F13A, "M", "k"),
+        (0x1F13B, "M", "l"),
+        (0x1F13C, "M", "m"),
+        (0x1F13D, "M", "n"),
+        (0x1F13E, "M", "o"),
+        (0x1F13F, "M", "p"),
+        (0x1F140, "M", "q"),
+        (0x1F141, "M", "r"),
+        (0x1F142, "M", "s"),
+        (0x1F143, "M", "t"),
+        (0x1F144, "M", "u"),
+        (0x1F145, "M", "v"),
+        (0x1F146, "M", "w"),
+        (0x1F147, "M", "x"),
+        (0x1F148, "M", "y"),
+        (0x1F149, "M", "z"),
+        (0x1F14A, "M", "hv"),
+        (0x1F14B, "M", "mv"),
+        (0x1F14C, "M", "sd"),
+        (0x1F14D, "M", "ss"),
+        (0x1F14E, "M", "ppv"),
+        (0x1F14F, "M", "wc"),
+        (0x1F150, "V"),
+        (0x1F16A, "M", "mc"),
+        (0x1F16B, "M", "md"),
+        (0x1F16C, "M", "mr"),
+        (0x1F16D, "V"),
+        (0x1F190, "M", "dj"),
+        (0x1F191, "V"),
+    ]
+
+
+def _seg_75() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1F1AE, "X"),
+        (0x1F1E6, "V"),
+        (0x1F200, "M", "ほか"),
+        (0x1F201, "M", "ココ"),
+        (0x1F202, "M", "サ"),
+        (0x1F203, "X"),
+        (0x1F210, "M", "手"),
+        (0x1F211, "M", "字"),
+        (0x1F212, "M", "双"),
+        (0x1F213, "M", "デ"),
+        (0x1F214, "M", "二"),
+        (0x1F215, "M", "多"),
+        (0x1F216, "M", "解"),
+        (0x1F217, "M", "天"),
+        (0x1F218, "M", "交"),
+        (0x1F219, "M", "映"),
+        (0x1F21A, "M", "無"),
+        (0x1F21B, "M", "料"),
+        (0x1F21C, "M", "前"),
+        (0x1F21D, "M", "後"),
+        (0x1F21E, "M", "再"),
+        (0x1F21F, "M", "新"),
+        (0x1F220, "M", "初"),
+        (0x1F221, "M", "終"),
+        (0x1F222, "M", "生"),
+        (0x1F223, "M", "販"),
+        (0x1F224, "M", "声"),
+        (0x1F225, "M", "吹"),
+        (0x1F226, "M", "演"),
+        (0x1F227, "M", "投"),
+        (0x1F228, "M", "捕"),
+        (0x1F229, "M", "一"),
+        (0x1F22A, "M", "三"),
+        (0x1F22B, "M", "遊"),
+        (0x1F22C, "M", "左"),
+        (0x1F22D, "M", "中"),
+        (0x1F22E, "M", "右"),
+        (0x1F22F, "M", "指"),
+        (0x1F230, "M", "走"),
+        (0x1F231, "M", "打"),
+        (0x1F232, "M", "禁"),
+        (0x1F233, "M", "空"),
+        (0x1F234, "M", "合"),
+        (0x1F235, "M", "満"),
+        (0x1F236, "M", "有"),
+        (0x1F237, "M", "月"),
+        (0x1F238, "M", "申"),
+        (0x1F239, "M", "割"),
+        (0x1F23A, "M", "営"),
+        (0x1F23B, "M", "配"),
+        (0x1F23C, "X"),
+        (0x1F240, "M", "〔本〕"),
+        (0x1F241, "M", "〔三〕"),
+        (0x1F242, "M", "〔二〕"),
+        (0x1F243, "M", "〔安〕"),
+        (0x1F244, "M", "〔点〕"),
+        (0x1F245, "M", "〔打〕"),
+        (0x1F246, "M", "〔盗〕"),
+        (0x1F247, "M", "〔勝〕"),
+        (0x1F248, "M", "〔敗〕"),
+        (0x1F249, "X"),
+        (0x1F250, "M", "得"),
+        (0x1F251, "M", "可"),
+        (0x1F252, "X"),
+        (0x1F260, "V"),
+        (0x1F266, "X"),
+        (0x1F300, "V"),
+        (0x1F6D8, "X"),
+        (0x1F6DC, "V"),
+        (0x1F6ED, "X"),
+        (0x1F6F0, "V"),
+        (0x1F6FD, "X"),
+        (0x1F700, "V"),
+        (0x1F777, "X"),
+        (0x1F77B, "V"),
+        (0x1F7DA, "X"),
+        (0x1F7E0, "V"),
+        (0x1F7EC, "X"),
+        (0x1F7F0, "V"),
+        (0x1F7F1, "X"),
+        (0x1F800, "V"),
+        (0x1F80C, "X"),
+        (0x1F810, "V"),
+        (0x1F848, "X"),
+        (0x1F850, "V"),
+        (0x1F85A, "X"),
+        (0x1F860, "V"),
+        (0x1F888, "X"),
+        (0x1F890, "V"),
+        (0x1F8AE, "X"),
+        (0x1F8B0, "V"),
+        (0x1F8B2, "X"),
+        (0x1F900, "V"),
+        (0x1FA54, "X"),
+        (0x1FA60, "V"),
+        (0x1FA6E, "X"),
+        (0x1FA70, "V"),
+        (0x1FA7D, "X"),
+        (0x1FA80, "V"),
+        (0x1FA89, "X"),
+    ]
+
+
+def _seg_76() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x1FA90, "V"),
+        (0x1FABE, "X"),
+        (0x1FABF, "V"),
+        (0x1FAC6, "X"),
+        (0x1FACE, "V"),
+        (0x1FADC, "X"),
+        (0x1FAE0, "V"),
+        (0x1FAE9, "X"),
+        (0x1FAF0, "V"),
+        (0x1FAF9, "X"),
+        (0x1FB00, "V"),
+        (0x1FB93, "X"),
+        (0x1FB94, "V"),
+        (0x1FBCB, "X"),
+        (0x1FBF0, "M", "0"),
+        (0x1FBF1, "M", "1"),
+        (0x1FBF2, "M", "2"),
+        (0x1FBF3, "M", "3"),
+        (0x1FBF4, "M", "4"),
+        (0x1FBF5, "M", "5"),
+        (0x1FBF6, "M", "6"),
+        (0x1FBF7, "M", "7"),
+        (0x1FBF8, "M", "8"),
+        (0x1FBF9, "M", "9"),
+        (0x1FBFA, "X"),
+        (0x20000, "V"),
+        (0x2A6E0, "X"),
+        (0x2A700, "V"),
+        (0x2B73A, "X"),
+        (0x2B740, "V"),
+        (0x2B81E, "X"),
+        (0x2B820, "V"),
+        (0x2CEA2, "X"),
+        (0x2CEB0, "V"),
+        (0x2EBE1, "X"),
+        (0x2EBF0, "V"),
+        (0x2EE5E, "X"),
+        (0x2F800, "M", "丽"),
+        (0x2F801, "M", "丸"),
+        (0x2F802, "M", "乁"),
+        (0x2F803, "M", "𠄢"),
+        (0x2F804, "M", "你"),
+        (0x2F805, "M", "侮"),
+        (0x2F806, "M", "侻"),
+        (0x2F807, "M", "倂"),
+        (0x2F808, "M", "偺"),
+        (0x2F809, "M", "備"),
+        (0x2F80A, "M", "僧"),
+        (0x2F80B, "M", "像"),
+        (0x2F80C, "M", "㒞"),
+        (0x2F80D, "M", "𠘺"),
+        (0x2F80E, "M", "免"),
+        (0x2F80F, "M", "兔"),
+        (0x2F810, "M", "兤"),
+        (0x2F811, "M", "具"),
+        (0x2F812, "M", "𠔜"),
+        (0x2F813, "M", "㒹"),
+        (0x2F814, "M", "內"),
+        (0x2F815, "M", "再"),
+        (0x2F816, "M", "𠕋"),
+        (0x2F817, "M", "冗"),
+        (0x2F818, "M", "冤"),
+        (0x2F819, "M", "仌"),
+        (0x2F81A, "M", "冬"),
+        (0x2F81B, "M", "况"),
+        (0x2F81C, "M", "𩇟"),
+        (0x2F81D, "M", "凵"),
+        (0x2F81E, "M", "刃"),
+        (0x2F81F, "M", "㓟"),
+        (0x2F820, "M", "刻"),
+        (0x2F821, "M", "剆"),
+        (0x2F822, "M", "割"),
+        (0x2F823, "M", "剷"),
+        (0x2F824, "M", "㔕"),
+        (0x2F825, "M", "勇"),
+        (0x2F826, "M", "勉"),
+        (0x2F827, "M", "勤"),
+        (0x2F828, "M", "勺"),
+        (0x2F829, "M", "包"),
+        (0x2F82A, "M", "匆"),
+        (0x2F82B, "M", "北"),
+        (0x2F82C, "M", "卉"),
+        (0x2F82D, "M", "卑"),
+        (0x2F82E, "M", "博"),
+        (0x2F82F, "M", "即"),
+        (0x2F830, "M", "卽"),
+        (0x2F831, "M", "卿"),
+        (0x2F834, "M", "𠨬"),
+        (0x2F835, "M", "灰"),
+        (0x2F836, "M", "及"),
+        (0x2F837, "M", "叟"),
+        (0x2F838, "M", "𠭣"),
+        (0x2F839, "M", "叫"),
+        (0x2F83A, "M", "叱"),
+        (0x2F83B, "M", "吆"),
+        (0x2F83C, "M", "咞"),
+        (0x2F83D, "M", "吸"),
+        (0x2F83E, "M", "呈"),
+        (0x2F83F, "M", "周"),
+        (0x2F840, "M", "咢"),
+    ]
+
+
+def _seg_77() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F841, "M", "哶"),
+        (0x2F842, "M", "唐"),
+        (0x2F843, "M", "啓"),
+        (0x2F844, "M", "啣"),
+        (0x2F845, "M", "善"),
+        (0x2F847, "M", "喙"),
+        (0x2F848, "M", "喫"),
+        (0x2F849, "M", "喳"),
+        (0x2F84A, "M", "嗂"),
+        (0x2F84B, "M", "圖"),
+        (0x2F84C, "M", "嘆"),
+        (0x2F84D, "M", "圗"),
+        (0x2F84E, "M", "噑"),
+        (0x2F84F, "M", "噴"),
+        (0x2F850, "M", "切"),
+        (0x2F851, "M", "壮"),
+        (0x2F852, "M", "城"),
+        (0x2F853, "M", "埴"),
+        (0x2F854, "M", "堍"),
+        (0x2F855, "M", "型"),
+        (0x2F856, "M", "堲"),
+        (0x2F857, "M", "報"),
+        (0x2F858, "M", "墬"),
+        (0x2F859, "M", "𡓤"),
+        (0x2F85A, "M", "売"),
+        (0x2F85B, "M", "壷"),
+        (0x2F85C, "M", "夆"),
+        (0x2F85D, "M", "多"),
+        (0x2F85E, "M", "夢"),
+        (0x2F85F, "M", "奢"),
+        (0x2F860, "M", "𡚨"),
+        (0x2F861, "M", "𡛪"),
+        (0x2F862, "M", "姬"),
+        (0x2F863, "M", "娛"),
+        (0x2F864, "M", "娧"),
+        (0x2F865, "M", "姘"),
+        (0x2F866, "M", "婦"),
+        (0x2F867, "M", "㛮"),
+        (0x2F868, "X"),
+        (0x2F869, "M", "嬈"),
+        (0x2F86A, "M", "嬾"),
+        (0x2F86C, "M", "𡧈"),
+        (0x2F86D, "M", "寃"),
+        (0x2F86E, "M", "寘"),
+        (0x2F86F, "M", "寧"),
+        (0x2F870, "M", "寳"),
+        (0x2F871, "M", "𡬘"),
+        (0x2F872, "M", "寿"),
+        (0x2F873, "M", "将"),
+        (0x2F874, "X"),
+        (0x2F875, "M", "尢"),
+        (0x2F876, "M", "㞁"),
+        (0x2F877, "M", "屠"),
+        (0x2F878, "M", "屮"),
+        (0x2F879, "M", "峀"),
+        (0x2F87A, "M", "岍"),
+        (0x2F87B, "M", "𡷤"),
+        (0x2F87C, "M", "嵃"),
+        (0x2F87D, "M", "𡷦"),
+        (0x2F87E, "M", "嵮"),
+        (0x2F87F, "M", "嵫"),
+        (0x2F880, "M", "嵼"),
+        (0x2F881, "M", "巡"),
+        (0x2F882, "M", "巢"),
+        (0x2F883, "M", "㠯"),
+        (0x2F884, "M", "巽"),
+        (0x2F885, "M", "帨"),
+        (0x2F886, "M", "帽"),
+        (0x2F887, "M", "幩"),
+        (0x2F888, "M", "㡢"),
+        (0x2F889, "M", "𢆃"),
+        (0x2F88A, "M", "㡼"),
+        (0x2F88B, "M", "庰"),
+        (0x2F88C, "M", "庳"),
+        (0x2F88D, "M", "庶"),
+        (0x2F88E, "M", "廊"),
+        (0x2F88F, "M", "𪎒"),
+        (0x2F890, "M", "廾"),
+        (0x2F891, "M", "𢌱"),
+        (0x2F893, "M", "舁"),
+        (0x2F894, "M", "弢"),
+        (0x2F896, "M", "㣇"),
+        (0x2F897, "M", "𣊸"),
+        (0x2F898, "M", "𦇚"),
+        (0x2F899, "M", "形"),
+        (0x2F89A, "M", "彫"),
+        (0x2F89B, "M", "㣣"),
+        (0x2F89C, "M", "徚"),
+        (0x2F89D, "M", "忍"),
+        (0x2F89E, "M", "志"),
+        (0x2F89F, "M", "忹"),
+        (0x2F8A0, "M", "悁"),
+        (0x2F8A1, "M", "㤺"),
+        (0x2F8A2, "M", "㤜"),
+        (0x2F8A3, "M", "悔"),
+        (0x2F8A4, "M", "𢛔"),
+        (0x2F8A5, "M", "惇"),
+        (0x2F8A6, "M", "慈"),
+        (0x2F8A7, "M", "慌"),
+        (0x2F8A8, "M", "慎"),
+    ]
+
+
+def _seg_78() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F8A9, "M", "慌"),
+        (0x2F8AA, "M", "慺"),
+        (0x2F8AB, "M", "憎"),
+        (0x2F8AC, "M", "憲"),
+        (0x2F8AD, "M", "憤"),
+        (0x2F8AE, "M", "憯"),
+        (0x2F8AF, "M", "懞"),
+        (0x2F8B0, "M", "懲"),
+        (0x2F8B1, "M", "懶"),
+        (0x2F8B2, "M", "成"),
+        (0x2F8B3, "M", "戛"),
+        (0x2F8B4, "M", "扝"),
+        (0x2F8B5, "M", "抱"),
+        (0x2F8B6, "M", "拔"),
+        (0x2F8B7, "M", "捐"),
+        (0x2F8B8, "M", "𢬌"),
+        (0x2F8B9, "M", "挽"),
+        (0x2F8BA, "M", "拼"),
+        (0x2F8BB, "M", "捨"),
+        (0x2F8BC, "M", "掃"),
+        (0x2F8BD, "M", "揤"),
+        (0x2F8BE, "M", "𢯱"),
+        (0x2F8BF, "M", "搢"),
+        (0x2F8C0, "M", "揅"),
+        (0x2F8C1, "M", "掩"),
+        (0x2F8C2, "M", "㨮"),
+        (0x2F8C3, "M", "摩"),
+        (0x2F8C4, "M", "摾"),
+        (0x2F8C5, "M", "撝"),
+        (0x2F8C6, "M", "摷"),
+        (0x2F8C7, "M", "㩬"),
+        (0x2F8C8, "M", "敏"),
+        (0x2F8C9, "M", "敬"),
+        (0x2F8CA, "M", "𣀊"),
+        (0x2F8CB, "M", "旣"),
+        (0x2F8CC, "M", "書"),
+        (0x2F8CD, "M", "晉"),
+        (0x2F8CE, "M", "㬙"),
+        (0x2F8CF, "M", "暑"),
+        (0x2F8D0, "M", "㬈"),
+        (0x2F8D1, "M", "㫤"),
+        (0x2F8D2, "M", "冒"),
+        (0x2F8D3, "M", "冕"),
+        (0x2F8D4, "M", "最"),
+        (0x2F8D5, "M", "暜"),
+        (0x2F8D6, "M", "肭"),
+        (0x2F8D7, "M", "䏙"),
+        (0x2F8D8, "M", "朗"),
+        (0x2F8D9, "M", "望"),
+        (0x2F8DA, "M", "朡"),
+        (0x2F8DB, "M", "杞"),
+        (0x2F8DC, "M", "杓"),
+        (0x2F8DD, "M", "𣏃"),
+        (0x2F8DE, "M", "㭉"),
+        (0x2F8DF, "M", "柺"),
+        (0x2F8E0, "M", "枅"),
+        (0x2F8E1, "M", "桒"),
+        (0x2F8E2, "M", "梅"),
+        (0x2F8E3, "M", "𣑭"),
+        (0x2F8E4, "M", "梎"),
+        (0x2F8E5, "M", "栟"),
+        (0x2F8E6, "M", "椔"),
+        (0x2F8E7, "M", "㮝"),
+        (0x2F8E8, "M", "楂"),
+        (0x2F8E9, "M", "榣"),
+        (0x2F8EA, "M", "槪"),
+        (0x2F8EB, "M", "檨"),
+        (0x2F8EC, "M", "𣚣"),
+        (0x2F8ED, "M", "櫛"),
+        (0x2F8EE, "M", "㰘"),
+        (0x2F8EF, "M", "次"),
+        (0x2F8F0, "M", "𣢧"),
+        (0x2F8F1, "M", "歔"),
+        (0x2F8F2, "M", "㱎"),
+        (0x2F8F3, "M", "歲"),
+        (0x2F8F4, "M", "殟"),
+        (0x2F8F5, "M", "殺"),
+        (0x2F8F6, "M", "殻"),
+        (0x2F8F7, "M", "𣪍"),
+        (0x2F8F8, "M", "𡴋"),
+        (0x2F8F9, "M", "𣫺"),
+        (0x2F8FA, "M", "汎"),
+        (0x2F8FB, "M", "𣲼"),
+        (0x2F8FC, "M", "沿"),
+        (0x2F8FD, "M", "泍"),
+        (0x2F8FE, "M", "汧"),
+        (0x2F8FF, "M", "洖"),
+        (0x2F900, "M", "派"),
+        (0x2F901, "M", "海"),
+        (0x2F902, "M", "流"),
+        (0x2F903, "M", "浩"),
+        (0x2F904, "M", "浸"),
+        (0x2F905, "M", "涅"),
+        (0x2F906, "M", "𣴞"),
+        (0x2F907, "M", "洴"),
+        (0x2F908, "M", "港"),
+        (0x2F909, "M", "湮"),
+        (0x2F90A, "M", "㴳"),
+        (0x2F90B, "M", "滋"),
+        (0x2F90C, "M", "滇"),
+    ]
+
+
+def _seg_79() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F90D, "M", "𣻑"),
+        (0x2F90E, "M", "淹"),
+        (0x2F90F, "M", "潮"),
+        (0x2F910, "M", "𣽞"),
+        (0x2F911, "M", "𣾎"),
+        (0x2F912, "M", "濆"),
+        (0x2F913, "M", "瀹"),
+        (0x2F914, "M", "瀞"),
+        (0x2F915, "M", "瀛"),
+        (0x2F916, "M", "㶖"),
+        (0x2F917, "M", "灊"),
+        (0x2F918, "M", "災"),
+        (0x2F919, "M", "灷"),
+        (0x2F91A, "M", "炭"),
+        (0x2F91B, "M", "𠔥"),
+        (0x2F91C, "M", "煅"),
+        (0x2F91D, "M", "𤉣"),
+        (0x2F91E, "M", "熜"),
+        (0x2F91F, "X"),
+        (0x2F920, "M", "爨"),
+        (0x2F921, "M", "爵"),
+        (0x2F922, "M", "牐"),
+        (0x2F923, "M", "𤘈"),
+        (0x2F924, "M", "犀"),
+        (0x2F925, "M", "犕"),
+        (0x2F926, "M", "𤜵"),
+        (0x2F927, "M", "𤠔"),
+        (0x2F928, "M", "獺"),
+        (0x2F929, "M", "王"),
+        (0x2F92A, "M", "㺬"),
+        (0x2F92B, "M", "玥"),
+        (0x2F92C, "M", "㺸"),
+        (0x2F92E, "M", "瑇"),
+        (0x2F92F, "M", "瑜"),
+        (0x2F930, "M", "瑱"),
+        (0x2F931, "M", "璅"),
+        (0x2F932, "M", "瓊"),
+        (0x2F933, "M", "㼛"),
+        (0x2F934, "M", "甤"),
+        (0x2F935, "M", "𤰶"),
+        (0x2F936, "M", "甾"),
+        (0x2F937, "M", "𤲒"),
+        (0x2F938, "M", "異"),
+        (0x2F939, "M", "𢆟"),
+        (0x2F93A, "M", "瘐"),
+        (0x2F93B, "M", "𤾡"),
+        (0x2F93C, "M", "𤾸"),
+        (0x2F93D, "M", "𥁄"),
+        (0x2F93E, "M", "㿼"),
+        (0x2F93F, "M", "䀈"),
+        (0x2F940, "M", "直"),
+        (0x2F941, "M", "𥃳"),
+        (0x2F942, "M", "𥃲"),
+        (0x2F943, "M", "𥄙"),
+        (0x2F944, "M", "𥄳"),
+        (0x2F945, "M", "眞"),
+        (0x2F946, "M", "真"),
+        (0x2F948, "M", "睊"),
+        (0x2F949, "M", "䀹"),
+        (0x2F94A, "M", "瞋"),
+        (0x2F94B, "M", "䁆"),
+        (0x2F94C, "M", "䂖"),
+        (0x2F94D, "M", "𥐝"),
+        (0x2F94E, "M", "硎"),
+        (0x2F94F, "M", "碌"),
+        (0x2F950, "M", "磌"),
+        (0x2F951, "M", "䃣"),
+        (0x2F952, "M", "𥘦"),
+        (0x2F953, "M", "祖"),
+        (0x2F954, "M", "𥚚"),
+        (0x2F955, "M", "𥛅"),
+        (0x2F956, "M", "福"),
+        (0x2F957, "M", "秫"),
+        (0x2F958, "M", "䄯"),
+        (0x2F959, "M", "穀"),
+        (0x2F95A, "M", "穊"),
+        (0x2F95B, "M", "穏"),
+        (0x2F95C, "M", "𥥼"),
+        (0x2F95D, "M", "𥪧"),
+        (0x2F95F, "X"),
+        (0x2F960, "M", "䈂"),
+        (0x2F961, "M", "𥮫"),
+        (0x2F962, "M", "篆"),
+        (0x2F963, "M", "築"),
+        (0x2F964, "M", "䈧"),
+        (0x2F965, "M", "𥲀"),
+        (0x2F966, "M", "糒"),
+        (0x2F967, "M", "䊠"),
+        (0x2F968, "M", "糨"),
+        (0x2F969, "M", "糣"),
+        (0x2F96A, "M", "紀"),
+        (0x2F96B, "M", "𥾆"),
+        (0x2F96C, "M", "絣"),
+        (0x2F96D, "M", "䌁"),
+        (0x2F96E, "M", "緇"),
+        (0x2F96F, "M", "縂"),
+        (0x2F970, "M", "繅"),
+        (0x2F971, "M", "䌴"),
+        (0x2F972, "M", "𦈨"),
+        (0x2F973, "M", "𦉇"),
+    ]
+
+
+def _seg_80() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F974, "M", "䍙"),
+        (0x2F975, "M", "𦋙"),
+        (0x2F976, "M", "罺"),
+        (0x2F977, "M", "𦌾"),
+        (0x2F978, "M", "羕"),
+        (0x2F979, "M", "翺"),
+        (0x2F97A, "M", "者"),
+        (0x2F97B, "M", "𦓚"),
+        (0x2F97C, "M", "𦔣"),
+        (0x2F97D, "M", "聠"),
+        (0x2F97E, "M", "𦖨"),
+        (0x2F97F, "M", "聰"),
+        (0x2F980, "M", "𣍟"),
+        (0x2F981, "M", "䏕"),
+        (0x2F982, "M", "育"),
+        (0x2F983, "M", "脃"),
+        (0x2F984, "M", "䐋"),
+        (0x2F985, "M", "脾"),
+        (0x2F986, "M", "媵"),
+        (0x2F987, "M", "𦞧"),
+        (0x2F988, "M", "𦞵"),
+        (0x2F989, "M", "𣎓"),
+        (0x2F98A, "M", "𣎜"),
+        (0x2F98B, "M", "舁"),
+        (0x2F98C, "M", "舄"),
+        (0x2F98D, "M", "辞"),
+        (0x2F98E, "M", "䑫"),
+        (0x2F98F, "M", "芑"),
+        (0x2F990, "M", "芋"),
+        (0x2F991, "M", "芝"),
+        (0x2F992, "M", "劳"),
+        (0x2F993, "M", "花"),
+        (0x2F994, "M", "芳"),
+        (0x2F995, "M", "芽"),
+        (0x2F996, "M", "苦"),
+        (0x2F997, "M", "𦬼"),
+        (0x2F998, "M", "若"),
+        (0x2F999, "M", "茝"),
+        (0x2F99A, "M", "荣"),
+        (0x2F99B, "M", "莭"),
+        (0x2F99C, "M", "茣"),
+        (0x2F99D, "M", "莽"),
+        (0x2F99E, "M", "菧"),
+        (0x2F99F, "M", "著"),
+        (0x2F9A0, "M", "荓"),
+        (0x2F9A1, "M", "菊"),
+        (0x2F9A2, "M", "菌"),
+        (0x2F9A3, "M", "菜"),
+        (0x2F9A4, "M", "𦰶"),
+        (0x2F9A5, "M", "𦵫"),
+        (0x2F9A6, "M", "𦳕"),
+        (0x2F9A7, "M", "䔫"),
+        (0x2F9A8, "M", "蓱"),
+        (0x2F9A9, "M", "蓳"),
+        (0x2F9AA, "M", "蔖"),
+        (0x2F9AB, "M", "𧏊"),
+        (0x2F9AC, "M", "蕤"),
+        (0x2F9AD, "M", "𦼬"),
+        (0x2F9AE, "M", "䕝"),
+        (0x2F9AF, "M", "䕡"),
+        (0x2F9B0, "M", "𦾱"),
+        (0x2F9B1, "M", "𧃒"),
+        (0x2F9B2, "M", "䕫"),
+        (0x2F9B3, "M", "虐"),
+        (0x2F9B4, "M", "虜"),
+        (0x2F9B5, "M", "虧"),
+        (0x2F9B6, "M", "虩"),
+        (0x2F9B7, "M", "蚩"),
+        (0x2F9B8, "M", "蚈"),
+        (0x2F9B9, "M", "蜎"),
+        (0x2F9BA, "M", "蛢"),
+        (0x2F9BB, "M", "蝹"),
+        (0x2F9BC, "M", "蜨"),
+        (0x2F9BD, "M", "蝫"),
+        (0x2F9BE, "M", "螆"),
+        (0x2F9BF, "X"),
+        (0x2F9C0, "M", "蟡"),
+        (0x2F9C1, "M", "蠁"),
+        (0x2F9C2, "M", "䗹"),
+        (0x2F9C3, "M", "衠"),
+        (0x2F9C4, "M", "衣"),
+        (0x2F9C5, "M", "𧙧"),
+        (0x2F9C6, "M", "裗"),
+        (0x2F9C7, "M", "裞"),
+        (0x2F9C8, "M", "䘵"),
+        (0x2F9C9, "M", "裺"),
+        (0x2F9CA, "M", "㒻"),
+        (0x2F9CB, "M", "𧢮"),
+        (0x2F9CC, "M", "𧥦"),
+        (0x2F9CD, "M", "䚾"),
+        (0x2F9CE, "M", "䛇"),
+        (0x2F9CF, "M", "誠"),
+        (0x2F9D0, "M", "諭"),
+        (0x2F9D1, "M", "變"),
+        (0x2F9D2, "M", "豕"),
+        (0x2F9D3, "M", "𧲨"),
+        (0x2F9D4, "M", "貫"),
+        (0x2F9D5, "M", "賁"),
+        (0x2F9D6, "M", "贛"),
+        (0x2F9D7, "M", "起"),
+    ]
+
+
+def _seg_81() -> List[Union[Tuple[int, str], Tuple[int, str, str]]]:
+    return [
+        (0x2F9D8, "M", "𧼯"),
+        (0x2F9D9, "M", "𠠄"),
+        (0x2F9DA, "M", "跋"),
+        (0x2F9DB, "M", "趼"),
+        (0x2F9DC, "M", "跰"),
+        (0x2F9DD, "M", "𠣞"),
+        (0x2F9DE, "M", "軔"),
+        (0x2F9DF, "M", "輸"),
+        (0x2F9E0, "M", "𨗒"),
+        (0x2F9E1, "M", "𨗭"),
+        (0x2F9E2, "M", "邔"),
+        (0x2F9E3, "M", "郱"),
+        (0x2F9E4, "M", "鄑"),
+        (0x2F9E5, "M", "𨜮"),
+        (0x2F9E6, "M", "鄛"),
+        (0x2F9E7, "M", "鈸"),
+        (0x2F9E8, "M", "鋗"),
+        (0x2F9E9, "M", "鋘"),
+        (0x2F9EA, "M", "鉼"),
+        (0x2F9EB, "M", "鏹"),
+        (0x2F9EC, "M", "鐕"),
+        (0x2F9ED, "M", "𨯺"),
+        (0x2F9EE, "M", "開"),
+        (0x2F9EF, "M", "䦕"),
+        (0x2F9F0, "M", "閷"),
+        (0x2F9F1, "M", "𨵷"),
+        (0x2F9F2, "M", "䧦"),
+        (0x2F9F3, "M", "雃"),
+        (0x2F9F4, "M", "嶲"),
+        (0x2F9F5, "M", "霣"),
+        (0x2F9F6, "M", "𩅅"),
+        (0x2F9F7, "M", "𩈚"),
+        (0x2F9F8, "M", "䩮"),
+        (0x2F9F9, "M", "䩶"),
+        (0x2F9FA, "M", "韠"),
+        (0x2F9FB, "M", "𩐊"),
+        (0x2F9FC, "M", "䪲"),
+        (0x2F9FD, "M", "𩒖"),
+        (0x2F9FE, "M", "頋"),
+        (0x2FA00, "M", "頩"),
+        (0x2FA01, "M", "𩖶"),
+        (0x2FA02, "M", "飢"),
+        (0x2FA03, "M", "䬳"),
+        (0x2FA04, "M", "餩"),
+        (0x2FA05, "M", "馧"),
+        (0x2FA06, "M", "駂"),
+        (0x2FA07, "M", "駾"),
+        (0x2FA08, "M", "䯎"),
+        (0x2FA09, "M", "𩬰"),
+        (0x2FA0A, "M", "鬒"),
+        (0x2FA0B, "M", "鱀"),
+        (0x2FA0C, "M", "鳽"),
+        (0x2FA0D, "M", "䳎"),
+        (0x2FA0E, "M", "䳭"),
+        (0x2FA0F, "M", "鵧"),
+        (0x2FA10, "M", "𪃎"),
+        (0x2FA11, "M", "䳸"),
+        (0x2FA12, "M", "𪄅"),
+        (0x2FA13, "M", "𪈎"),
+        (0x2FA14, "M", "𪊑"),
+        (0x2FA15, "M", "麻"),
+        (0x2FA16, "M", "䵖"),
+        (0x2FA17, "M", "黹"),
+        (0x2FA18, "M", "黾"),
+        (0x2FA19, "M", "鼅"),
+        (0x2FA1A, "M", "鼏"),
+        (0x2FA1B, "M", "鼖"),
+        (0x2FA1C, "M", "鼻"),
+        (0x2FA1D, "M", "𪘀"),
+        (0x2FA1E, "X"),
+        (0x30000, "V"),
+        (0x3134B, "X"),
+        (0x31350, "V"),
+        (0x323B0, "X"),
+        (0xE0100, "I"),
+        (0xE01F0, "X"),
+    ]
+
+
+uts46data = tuple(
+    _seg_0()
+    + _seg_1()
+    + _seg_2()
+    + _seg_3()
+    + _seg_4()
+    + _seg_5()
+    + _seg_6()
+    + _seg_7()
+    + _seg_8()
+    + _seg_9()
+    + _seg_10()
+    + _seg_11()
+    + _seg_12()
+    + _seg_13()
+    + _seg_14()
+    + _seg_15()
+    + _seg_16()
+    + _seg_17()
+    + _seg_18()
+    + _seg_19()
+    + _seg_20()
+    + _seg_21()
+    + _seg_22()
+    + _seg_23()
+    + _seg_24()
+    + _seg_25()
+    + _seg_26()
+    + _seg_27()
+    + _seg_28()
+    + _seg_29()
+    + _seg_30()
+    + _seg_31()
+    + _seg_32()
+    + _seg_33()
+    + _seg_34()
+    + _seg_35()
+    + _seg_36()
+    + _seg_37()
+    + _seg_38()
+    + _seg_39()
+    + _seg_40()
+    + _seg_41()
+    + _seg_42()
+    + _seg_43()
+    + _seg_44()
+    + _seg_45()
+    + _seg_46()
+    + _seg_47()
+    + _seg_48()
+    + _seg_49()
+    + _seg_50()
+    + _seg_51()
+    + _seg_52()
+    + _seg_53()
+    + _seg_54()
+    + _seg_55()
+    + _seg_56()
+    + _seg_57()
+    + _seg_58()
+    + _seg_59()
+    + _seg_60()
+    + _seg_61()
+    + _seg_62()
+    + _seg_63()
+    + _seg_64()
+    + _seg_65()
+    + _seg_66()
+    + _seg_67()
+    + _seg_68()
+    + _seg_69()
+    + _seg_70()
+    + _seg_71()
+    + _seg_72()
+    + _seg_73()
+    + _seg_74()
+    + _seg_75()
+    + _seg_76()
+    + _seg_77()
+    + _seg_78()
+    + _seg_79()
+    + _seg_80()
+    + _seg_81()
+)  # type: Tuple[Union[Tuple[int, str], Tuple[int, str, str]], ...]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__init__.py
new file mode 100644
index 00000000..b6151054
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__init__.py
@@ -0,0 +1,55 @@
+# ruff: noqa: F401
+import os
+
+from .exceptions import *  # noqa: F403
+from .ext import ExtType, Timestamp
+
+version = (1, 1, 0)
+__version__ = "1.1.0"
+
+
+if os.environ.get("MSGPACK_PUREPYTHON"):
+    from .fallback import Packer, Unpacker, unpackb
+else:
+    try:
+        from ._cmsgpack import Packer, Unpacker, unpackb
+    except ImportError:
+        from .fallback import Packer, Unpacker, unpackb
+
+
+def pack(o, stream, **kwargs):
+    """
+    Pack object `o` and write it to `stream`
+
+    See :class:`Packer` for options.
+    """
+    packer = Packer(**kwargs)
+    stream.write(packer.pack(o))
+
+
+def packb(o, **kwargs):
+    """
+    Pack object `o` and return packed bytes
+
+    See :class:`Packer` for options.
+    """
+    return Packer(**kwargs).pack(o)
+
+
+def unpack(stream, **kwargs):
+    """
+    Unpack an object from `stream`.
+
+    Raises `ExtraData` when `stream` contains extra bytes.
+    See :class:`Unpacker` for options.
+    """
+    data = stream.read()
+    return unpackb(data, **kwargs)
+
+
+# alias for compatibility to simplejson/marshal/pickle.
+load = unpack
+loads = unpackb
+
+dump = pack
+dumps = packb
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..569bd361
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..976eac65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/ext.cpython-312.pyc
new file mode 100644
index 00000000..06cd7a66
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/fallback.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/fallback.cpython-312.pyc
new file mode 100644
index 00000000..2af5d020
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/__pycache__/fallback.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/exceptions.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/exceptions.py
new file mode 100644
index 00000000..d6d2615c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/exceptions.py
@@ -0,0 +1,48 @@
+class UnpackException(Exception):
+    """Base class for some exceptions raised while unpacking.
+
+    NOTE: unpack may raise exception other than subclass of
+    UnpackException.  If you want to catch all error, catch
+    Exception instead.
+    """
+
+
+class BufferFull(UnpackException):
+    pass
+
+
+class OutOfData(UnpackException):
+    pass
+
+
+class FormatError(ValueError, UnpackException):
+    """Invalid msgpack format"""
+
+
+class StackError(ValueError, UnpackException):
+    """Too nested"""
+
+
+# Deprecated.  Use ValueError instead
+UnpackValueError = ValueError
+
+
+class ExtraData(UnpackValueError):
+    """ExtraData is raised when there is trailing data.
+
+    This exception is raised while only one-shot (not streaming)
+    unpack.
+    """
+
+    def __init__(self, unpacked, extra):
+        self.unpacked = unpacked
+        self.extra = extra
+
+    def __str__(self):
+        return "unpack(b) received extra data."
+
+
+# Deprecated.  Use Exception instead to catch all exception during packing.
+PackException = Exception
+PackValueError = ValueError
+PackOverflowError = OverflowError
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/ext.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/ext.py
new file mode 100644
index 00000000..9694819a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/ext.py
@@ -0,0 +1,170 @@
+import datetime
+import struct
+from collections import namedtuple
+
+
+class ExtType(namedtuple("ExtType", "code data")):
+    """ExtType represents ext type in msgpack."""
+
+    def __new__(cls, code, data):
+        if not isinstance(code, int):
+            raise TypeError("code must be int")
+        if not isinstance(data, bytes):
+            raise TypeError("data must be bytes")
+        if not 0 <= code <= 127:
+            raise ValueError("code must be 0~127")
+        return super().__new__(cls, code, data)
+
+
+class Timestamp:
+    """Timestamp represents the Timestamp extension type in msgpack.
+
+    When built with Cython, msgpack uses C methods to pack and unpack `Timestamp`.
+    When using pure-Python msgpack, :func:`to_bytes` and :func:`from_bytes` are used to pack and
+    unpack `Timestamp`.
+
+    This class is immutable: Do not override seconds and nanoseconds.
+    """
+
+    __slots__ = ["seconds", "nanoseconds"]
+
+    def __init__(self, seconds, nanoseconds=0):
+        """Initialize a Timestamp object.
+
+        :param int seconds:
+            Number of seconds since the UNIX epoch (00:00:00 UTC Jan 1 1970, minus leap seconds).
+            May be negative.
+
+        :param int nanoseconds:
+            Number of nanoseconds to add to `seconds` to get fractional time.
+            Maximum is 999_999_999.  Default is 0.
+
+        Note: Negative times (before the UNIX epoch) are represented as neg. seconds + pos. ns.
+        """
+        if not isinstance(seconds, int):
+            raise TypeError("seconds must be an integer")
+        if not isinstance(nanoseconds, int):
+            raise TypeError("nanoseconds must be an integer")
+        if not (0 <= nanoseconds < 10**9):
+            raise ValueError("nanoseconds must be a non-negative integer less than 999999999.")
+        self.seconds = seconds
+        self.nanoseconds = nanoseconds
+
+    def __repr__(self):
+        """String representation of Timestamp."""
+        return f"Timestamp(seconds={self.seconds}, nanoseconds={self.nanoseconds})"
+
+    def __eq__(self, other):
+        """Check for equality with another Timestamp object"""
+        if type(other) is self.__class__:
+            return self.seconds == other.seconds and self.nanoseconds == other.nanoseconds
+        return False
+
+    def __ne__(self, other):
+        """not-equals method (see :func:`__eq__()`)"""
+        return not self.__eq__(other)
+
+    def __hash__(self):
+        return hash((self.seconds, self.nanoseconds))
+
+    @staticmethod
+    def from_bytes(b):
+        """Unpack bytes into a `Timestamp` object.
+
+        Used for pure-Python msgpack unpacking.
+
+        :param b: Payload from msgpack ext message with code -1
+        :type b: bytes
+
+        :returns: Timestamp object unpacked from msgpack ext payload
+        :rtype: Timestamp
+        """
+        if len(b) == 4:
+            seconds = struct.unpack("!L", b)[0]
+            nanoseconds = 0
+        elif len(b) == 8:
+            data64 = struct.unpack("!Q", b)[0]
+            seconds = data64 & 0x00000003FFFFFFFF
+            nanoseconds = data64 >> 34
+        elif len(b) == 12:
+            nanoseconds, seconds = struct.unpack("!Iq", b)
+        else:
+            raise ValueError(
+                "Timestamp type can only be created from 32, 64, or 96-bit byte objects"
+            )
+        return Timestamp(seconds, nanoseconds)
+
+    def to_bytes(self):
+        """Pack this Timestamp object into bytes.
+
+        Used for pure-Python msgpack packing.
+
+        :returns data: Payload for EXT message with code -1 (timestamp type)
+        :rtype: bytes
+        """
+        if (self.seconds >> 34) == 0:  # seconds is non-negative and fits in 34 bits
+            data64 = self.nanoseconds << 34 | self.seconds
+            if data64 & 0xFFFFFFFF00000000 == 0:
+                # nanoseconds is zero and seconds < 2**32, so timestamp 32
+                data = struct.pack("!L", data64)
+            else:
+                # timestamp 64
+                data = struct.pack("!Q", data64)
+        else:
+            # timestamp 96
+            data = struct.pack("!Iq", self.nanoseconds, self.seconds)
+        return data
+
+    @staticmethod
+    def from_unix(unix_sec):
+        """Create a Timestamp from posix timestamp in seconds.
+
+        :param unix_float: Posix timestamp in seconds.
+        :type unix_float: int or float
+        """
+        seconds = int(unix_sec // 1)
+        nanoseconds = int((unix_sec % 1) * 10**9)
+        return Timestamp(seconds, nanoseconds)
+
+    def to_unix(self):
+        """Get the timestamp as a floating-point value.
+
+        :returns: posix timestamp
+        :rtype: float
+        """
+        return self.seconds + self.nanoseconds / 1e9
+
+    @staticmethod
+    def from_unix_nano(unix_ns):
+        """Create a Timestamp from posix timestamp in nanoseconds.
+
+        :param int unix_ns: Posix timestamp in nanoseconds.
+        :rtype: Timestamp
+        """
+        return Timestamp(*divmod(unix_ns, 10**9))
+
+    def to_unix_nano(self):
+        """Get the timestamp as a unixtime in nanoseconds.
+
+        :returns: posix timestamp in nanoseconds
+        :rtype: int
+        """
+        return self.seconds * 10**9 + self.nanoseconds
+
+    def to_datetime(self):
+        """Get the timestamp as a UTC datetime.
+
+        :rtype: `datetime.datetime`
+        """
+        utc = datetime.timezone.utc
+        return datetime.datetime.fromtimestamp(0, utc) + datetime.timedelta(
+            seconds=self.seconds, microseconds=self.nanoseconds // 1000
+        )
+
+    @staticmethod
+    def from_datetime(dt):
+        """Create a Timestamp from datetime with tzinfo.
+
+        :rtype: Timestamp
+        """
+        return Timestamp(seconds=int(dt.timestamp()), nanoseconds=dt.microsecond * 1000)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/fallback.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/fallback.py
new file mode 100644
index 00000000..b02e47cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/msgpack/fallback.py
@@ -0,0 +1,929 @@
+"""Fallback pure Python implementation of msgpack"""
+
+import struct
+import sys
+from datetime import datetime as _DateTime
+
+if hasattr(sys, "pypy_version_info"):
+    from __pypy__ import newlist_hint
+    from __pypy__.builders import BytesBuilder
+
+    _USING_STRINGBUILDER = True
+
+    class BytesIO:
+        def __init__(self, s=b""):
+            if s:
+                self.builder = BytesBuilder(len(s))
+                self.builder.append(s)
+            else:
+                self.builder = BytesBuilder()
+
+        def write(self, s):
+            if isinstance(s, memoryview):
+                s = s.tobytes()
+            elif isinstance(s, bytearray):
+                s = bytes(s)
+            self.builder.append(s)
+
+        def getvalue(self):
+            return self.builder.build()
+
+else:
+    from io import BytesIO
+
+    _USING_STRINGBUILDER = False
+
+    def newlist_hint(size):
+        return []
+
+
+from .exceptions import BufferFull, ExtraData, FormatError, OutOfData, StackError
+from .ext import ExtType, Timestamp
+
+EX_SKIP = 0
+EX_CONSTRUCT = 1
+EX_READ_ARRAY_HEADER = 2
+EX_READ_MAP_HEADER = 3
+
+TYPE_IMMEDIATE = 0
+TYPE_ARRAY = 1
+TYPE_MAP = 2
+TYPE_RAW = 3
+TYPE_BIN = 4
+TYPE_EXT = 5
+
+DEFAULT_RECURSE_LIMIT = 511
+
+
+def _check_type_strict(obj, t, type=type, tuple=tuple):
+    if type(t) is tuple:
+        return type(obj) in t
+    else:
+        return type(obj) is t
+
+
+def _get_data_from_buffer(obj):
+    view = memoryview(obj)
+    if view.itemsize != 1:
+        raise ValueError("cannot unpack from multi-byte object")
+    return view
+
+
+def unpackb(packed, **kwargs):
+    """
+    Unpack an object from `packed`.
+
+    Raises ``ExtraData`` when *packed* contains extra bytes.
+    Raises ``ValueError`` when *packed* is incomplete.
+    Raises ``FormatError`` when *packed* is not valid msgpack.
+    Raises ``StackError`` when *packed* contains too nested.
+    Other exceptions can be raised during unpacking.
+
+    See :class:`Unpacker` for options.
+    """
+    unpacker = Unpacker(None, max_buffer_size=len(packed), **kwargs)
+    unpacker.feed(packed)
+    try:
+        ret = unpacker._unpack()
+    except OutOfData:
+        raise ValueError("Unpack failed: incomplete input")
+    except RecursionError:
+        raise StackError
+    if unpacker._got_extradata():
+        raise ExtraData(ret, unpacker._get_extradata())
+    return ret
+
+
+_NO_FORMAT_USED = ""
+_MSGPACK_HEADERS = {
+    0xC4: (1, _NO_FORMAT_USED, TYPE_BIN),
+    0xC5: (2, ">H", TYPE_BIN),
+    0xC6: (4, ">I", TYPE_BIN),
+    0xC7: (2, "Bb", TYPE_EXT),
+    0xC8: (3, ">Hb", TYPE_EXT),
+    0xC9: (5, ">Ib", TYPE_EXT),
+    0xCA: (4, ">f"),
+    0xCB: (8, ">d"),
+    0xCC: (1, _NO_FORMAT_USED),
+    0xCD: (2, ">H"),
+    0xCE: (4, ">I"),
+    0xCF: (8, ">Q"),
+    0xD0: (1, "b"),
+    0xD1: (2, ">h"),
+    0xD2: (4, ">i"),
+    0xD3: (8, ">q"),
+    0xD4: (1, "b1s", TYPE_EXT),
+    0xD5: (2, "b2s", TYPE_EXT),
+    0xD6: (4, "b4s", TYPE_EXT),
+    0xD7: (8, "b8s", TYPE_EXT),
+    0xD8: (16, "b16s", TYPE_EXT),
+    0xD9: (1, _NO_FORMAT_USED, TYPE_RAW),
+    0xDA: (2, ">H", TYPE_RAW),
+    0xDB: (4, ">I", TYPE_RAW),
+    0xDC: (2, ">H", TYPE_ARRAY),
+    0xDD: (4, ">I", TYPE_ARRAY),
+    0xDE: (2, ">H", TYPE_MAP),
+    0xDF: (4, ">I", TYPE_MAP),
+}
+
+
+class Unpacker:
+    """Streaming unpacker.
+
+    Arguments:
+
+    :param file_like:
+        File-like object having `.read(n)` method.
+        If specified, unpacker reads serialized data from it and `.feed()` is not usable.
+
+    :param int read_size:
+        Used as `file_like.read(read_size)`. (default: `min(16*1024, max_buffer_size)`)
+
+    :param bool use_list:
+        If true, unpack msgpack array to Python list.
+        Otherwise, unpack to Python tuple. (default: True)
+
+    :param bool raw:
+        If true, unpack msgpack raw to Python bytes.
+        Otherwise, unpack to Python str by decoding with UTF-8 encoding (default).
+
+    :param int timestamp:
+        Control how timestamp type is unpacked:
+
+            0 - Timestamp
+            1 - float  (Seconds from the EPOCH)
+            2 - int  (Nanoseconds from the EPOCH)
+            3 - datetime.datetime  (UTC).
+
+    :param bool strict_map_key:
+        If true (default), only str or bytes are accepted for map (dict) keys.
+
+    :param object_hook:
+        When specified, it should be callable.
+        Unpacker calls it with a dict argument after unpacking msgpack map.
+        (See also simplejson)
+
+    :param object_pairs_hook:
+        When specified, it should be callable.
+        Unpacker calls it with a list of key-value pairs after unpacking msgpack map.
+        (See also simplejson)
+
+    :param str unicode_errors:
+        The error handler for decoding unicode. (default: 'strict')
+        This option should be used only when you have msgpack data which
+        contains invalid UTF-8 string.
+
+    :param int max_buffer_size:
+        Limits size of data waiting unpacked.  0 means 2**32-1.
+        The default value is 100*1024*1024 (100MiB).
+        Raises `BufferFull` exception when it is insufficient.
+        You should set this parameter when unpacking data from untrusted source.
+
+    :param int max_str_len:
+        Deprecated, use *max_buffer_size* instead.
+        Limits max length of str. (default: max_buffer_size)
+
+    :param int max_bin_len:
+        Deprecated, use *max_buffer_size* instead.
+        Limits max length of bin. (default: max_buffer_size)
+
+    :param int max_array_len:
+        Limits max length of array.
+        (default: max_buffer_size)
+
+    :param int max_map_len:
+        Limits max length of map.
+        (default: max_buffer_size//2)
+
+    :param int max_ext_len:
+        Deprecated, use *max_buffer_size* instead.
+        Limits max size of ext type.  (default: max_buffer_size)
+
+    Example of streaming deserialize from file-like object::
+
+        unpacker = Unpacker(file_like)
+        for o in unpacker:
+            process(o)
+
+    Example of streaming deserialize from socket::
+
+        unpacker = Unpacker()
+        while True:
+            buf = sock.recv(1024**2)
+            if not buf:
+                break
+            unpacker.feed(buf)
+            for o in unpacker:
+                process(o)
+
+    Raises ``ExtraData`` when *packed* contains extra bytes.
+    Raises ``OutOfData`` when *packed* is incomplete.
+    Raises ``FormatError`` when *packed* is not valid msgpack.
+    Raises ``StackError`` when *packed* contains too nested.
+    Other exceptions can be raised during unpacking.
+    """
+
+    def __init__(
+        self,
+        file_like=None,
+        *,
+        read_size=0,
+        use_list=True,
+        raw=False,
+        timestamp=0,
+        strict_map_key=True,
+        object_hook=None,
+        object_pairs_hook=None,
+        list_hook=None,
+        unicode_errors=None,
+        max_buffer_size=100 * 1024 * 1024,
+        ext_hook=ExtType,
+        max_str_len=-1,
+        max_bin_len=-1,
+        max_array_len=-1,
+        max_map_len=-1,
+        max_ext_len=-1,
+    ):
+        if unicode_errors is None:
+            unicode_errors = "strict"
+
+        if file_like is None:
+            self._feeding = True
+        else:
+            if not callable(file_like.read):
+                raise TypeError("`file_like.read` must be callable")
+            self.file_like = file_like
+            self._feeding = False
+
+        #: array of bytes fed.
+        self._buffer = bytearray()
+        #: Which position we currently reads
+        self._buff_i = 0
+
+        # When Unpacker is used as an iterable, between the calls to next(),
+        # the buffer is not "consumed" completely, for efficiency sake.
+        # Instead, it is done sloppily.  To make sure we raise BufferFull at
+        # the correct moments, we have to keep track of how sloppy we were.
+        # Furthermore, when the buffer is incomplete (that is: in the case
+        # we raise an OutOfData) we need to rollback the buffer to the correct
+        # state, which _buf_checkpoint records.
+        self._buf_checkpoint = 0
+
+        if not max_buffer_size:
+            max_buffer_size = 2**31 - 1
+        if max_str_len == -1:
+            max_str_len = max_buffer_size
+        if max_bin_len == -1:
+            max_bin_len = max_buffer_size
+        if max_array_len == -1:
+            max_array_len = max_buffer_size
+        if max_map_len == -1:
+            max_map_len = max_buffer_size // 2
+        if max_ext_len == -1:
+            max_ext_len = max_buffer_size
+
+        self._max_buffer_size = max_buffer_size
+        if read_size > self._max_buffer_size:
+            raise ValueError("read_size must be smaller than max_buffer_size")
+        self._read_size = read_size or min(self._max_buffer_size, 16 * 1024)
+        self._raw = bool(raw)
+        self._strict_map_key = bool(strict_map_key)
+        self._unicode_errors = unicode_errors
+        self._use_list = use_list
+        if not (0 <= timestamp <= 3):
+            raise ValueError("timestamp must be 0..3")
+        self._timestamp = timestamp
+        self._list_hook = list_hook
+        self._object_hook = object_hook
+        self._object_pairs_hook = object_pairs_hook
+        self._ext_hook = ext_hook
+        self._max_str_len = max_str_len
+        self._max_bin_len = max_bin_len
+        self._max_array_len = max_array_len
+        self._max_map_len = max_map_len
+        self._max_ext_len = max_ext_len
+        self._stream_offset = 0
+
+        if list_hook is not None and not callable(list_hook):
+            raise TypeError("`list_hook` is not callable")
+        if object_hook is not None and not callable(object_hook):
+            raise TypeError("`object_hook` is not callable")
+        if object_pairs_hook is not None and not callable(object_pairs_hook):
+            raise TypeError("`object_pairs_hook` is not callable")
+        if object_hook is not None and object_pairs_hook is not None:
+            raise TypeError("object_pairs_hook and object_hook are mutually exclusive")
+        if not callable(ext_hook):
+            raise TypeError("`ext_hook` is not callable")
+
+    def feed(self, next_bytes):
+        assert self._feeding
+        view = _get_data_from_buffer(next_bytes)
+        if len(self._buffer) - self._buff_i + len(view) > self._max_buffer_size:
+            raise BufferFull
+
+        # Strip buffer before checkpoint before reading file.
+        if self._buf_checkpoint > 0:
+            del self._buffer[: self._buf_checkpoint]
+            self._buff_i -= self._buf_checkpoint
+            self._buf_checkpoint = 0
+
+        # Use extend here: INPLACE_ADD += doesn't reliably typecast memoryview in jython
+        self._buffer.extend(view)
+        view.release()
+
+    def _consume(self):
+        """Gets rid of the used parts of the buffer."""
+        self._stream_offset += self._buff_i - self._buf_checkpoint
+        self._buf_checkpoint = self._buff_i
+
+    def _got_extradata(self):
+        return self._buff_i < len(self._buffer)
+
+    def _get_extradata(self):
+        return self._buffer[self._buff_i :]
+
+    def read_bytes(self, n):
+        ret = self._read(n, raise_outofdata=False)
+        self._consume()
+        return ret
+
+    def _read(self, n, raise_outofdata=True):
+        # (int) -> bytearray
+        self._reserve(n, raise_outofdata=raise_outofdata)
+        i = self._buff_i
+        ret = self._buffer[i : i + n]
+        self._buff_i = i + len(ret)
+        return ret
+
+    def _reserve(self, n, raise_outofdata=True):
+        remain_bytes = len(self._buffer) - self._buff_i - n
+
+        # Fast path: buffer has n bytes already
+        if remain_bytes >= 0:
+            return
+
+        if self._feeding:
+            self._buff_i = self._buf_checkpoint
+            raise OutOfData
+
+        # Strip buffer before checkpoint before reading file.
+        if self._buf_checkpoint > 0:
+            del self._buffer[: self._buf_checkpoint]
+            self._buff_i -= self._buf_checkpoint
+            self._buf_checkpoint = 0
+
+        # Read from file
+        remain_bytes = -remain_bytes
+        if remain_bytes + len(self._buffer) > self._max_buffer_size:
+            raise BufferFull
+        while remain_bytes > 0:
+            to_read_bytes = max(self._read_size, remain_bytes)
+            read_data = self.file_like.read(to_read_bytes)
+            if not read_data:
+                break
+            assert isinstance(read_data, bytes)
+            self._buffer += read_data
+            remain_bytes -= len(read_data)
+
+        if len(self._buffer) < n + self._buff_i and raise_outofdata:
+            self._buff_i = 0  # rollback
+            raise OutOfData
+
+    def _read_header(self):
+        typ = TYPE_IMMEDIATE
+        n = 0
+        obj = None
+        self._reserve(1)
+        b = self._buffer[self._buff_i]
+        self._buff_i += 1
+        if b & 0b10000000 == 0:
+            obj = b
+        elif b & 0b11100000 == 0b11100000:
+            obj = -1 - (b ^ 0xFF)
+        elif b & 0b11100000 == 0b10100000:
+            n = b & 0b00011111
+            typ = TYPE_RAW
+            if n > self._max_str_len:
+                raise ValueError(f"{n} exceeds max_str_len({self._max_str_len})")
+            obj = self._read(n)
+        elif b & 0b11110000 == 0b10010000:
+            n = b & 0b00001111
+            typ = TYPE_ARRAY
+            if n > self._max_array_len:
+                raise ValueError(f"{n} exceeds max_array_len({self._max_array_len})")
+        elif b & 0b11110000 == 0b10000000:
+            n = b & 0b00001111
+            typ = TYPE_MAP
+            if n > self._max_map_len:
+                raise ValueError(f"{n} exceeds max_map_len({self._max_map_len})")
+        elif b == 0xC0:
+            obj = None
+        elif b == 0xC2:
+            obj = False
+        elif b == 0xC3:
+            obj = True
+        elif 0xC4 <= b <= 0xC6:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            if len(fmt) > 0:
+                n = struct.unpack_from(fmt, self._buffer, self._buff_i)[0]
+            else:
+                n = self._buffer[self._buff_i]
+            self._buff_i += size
+            if n > self._max_bin_len:
+                raise ValueError(f"{n} exceeds max_bin_len({self._max_bin_len})")
+            obj = self._read(n)
+        elif 0xC7 <= b <= 0xC9:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            L, n = struct.unpack_from(fmt, self._buffer, self._buff_i)
+            self._buff_i += size
+            if L > self._max_ext_len:
+                raise ValueError(f"{L} exceeds max_ext_len({self._max_ext_len})")
+            obj = self._read(L)
+        elif 0xCA <= b <= 0xD3:
+            size, fmt = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            if len(fmt) > 0:
+                obj = struct.unpack_from(fmt, self._buffer, self._buff_i)[0]
+            else:
+                obj = self._buffer[self._buff_i]
+            self._buff_i += size
+        elif 0xD4 <= b <= 0xD8:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            if self._max_ext_len < size:
+                raise ValueError(f"{size} exceeds max_ext_len({self._max_ext_len})")
+            self._reserve(size + 1)
+            n, obj = struct.unpack_from(fmt, self._buffer, self._buff_i)
+            self._buff_i += size + 1
+        elif 0xD9 <= b <= 0xDB:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            if len(fmt) > 0:
+                (n,) = struct.unpack_from(fmt, self._buffer, self._buff_i)
+            else:
+                n = self._buffer[self._buff_i]
+            self._buff_i += size
+            if n > self._max_str_len:
+                raise ValueError(f"{n} exceeds max_str_len({self._max_str_len})")
+            obj = self._read(n)
+        elif 0xDC <= b <= 0xDD:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            (n,) = struct.unpack_from(fmt, self._buffer, self._buff_i)
+            self._buff_i += size
+            if n > self._max_array_len:
+                raise ValueError(f"{n} exceeds max_array_len({self._max_array_len})")
+        elif 0xDE <= b <= 0xDF:
+            size, fmt, typ = _MSGPACK_HEADERS[b]
+            self._reserve(size)
+            (n,) = struct.unpack_from(fmt, self._buffer, self._buff_i)
+            self._buff_i += size
+            if n > self._max_map_len:
+                raise ValueError(f"{n} exceeds max_map_len({self._max_map_len})")
+        else:
+            raise FormatError("Unknown header: 0x%x" % b)
+        return typ, n, obj
+
+    def _unpack(self, execute=EX_CONSTRUCT):
+        typ, n, obj = self._read_header()
+
+        if execute == EX_READ_ARRAY_HEADER:
+            if typ != TYPE_ARRAY:
+                raise ValueError("Expected array")
+            return n
+        if execute == EX_READ_MAP_HEADER:
+            if typ != TYPE_MAP:
+                raise ValueError("Expected map")
+            return n
+        # TODO should we eliminate the recursion?
+        if typ == TYPE_ARRAY:
+            if execute == EX_SKIP:
+                for i in range(n):
+                    # TODO check whether we need to call `list_hook`
+                    self._unpack(EX_SKIP)
+                return
+            ret = newlist_hint(n)
+            for i in range(n):
+                ret.append(self._unpack(EX_CONSTRUCT))
+            if self._list_hook is not None:
+                ret = self._list_hook(ret)
+            # TODO is the interaction between `list_hook` and `use_list` ok?
+            return ret if self._use_list else tuple(ret)
+        if typ == TYPE_MAP:
+            if execute == EX_SKIP:
+                for i in range(n):
+                    # TODO check whether we need to call hooks
+                    self._unpack(EX_SKIP)
+                    self._unpack(EX_SKIP)
+                return
+            if self._object_pairs_hook is not None:
+                ret = self._object_pairs_hook(
+                    (self._unpack(EX_CONSTRUCT), self._unpack(EX_CONSTRUCT)) for _ in range(n)
+                )
+            else:
+                ret = {}
+                for _ in range(n):
+                    key = self._unpack(EX_CONSTRUCT)
+                    if self._strict_map_key and type(key) not in (str, bytes):
+                        raise ValueError("%s is not allowed for map key" % str(type(key)))
+                    if isinstance(key, str):
+                        key = sys.intern(key)
+                    ret[key] = self._unpack(EX_CONSTRUCT)
+                if self._object_hook is not None:
+                    ret = self._object_hook(ret)
+            return ret
+        if execute == EX_SKIP:
+            return
+        if typ == TYPE_RAW:
+            if self._raw:
+                obj = bytes(obj)
+            else:
+                obj = obj.decode("utf_8", self._unicode_errors)
+            return obj
+        if typ == TYPE_BIN:
+            return bytes(obj)
+        if typ == TYPE_EXT:
+            if n == -1:  # timestamp
+                ts = Timestamp.from_bytes(bytes(obj))
+                if self._timestamp == 1:
+                    return ts.to_unix()
+                elif self._timestamp == 2:
+                    return ts.to_unix_nano()
+                elif self._timestamp == 3:
+                    return ts.to_datetime()
+                else:
+                    return ts
+            else:
+                return self._ext_hook(n, bytes(obj))
+        assert typ == TYPE_IMMEDIATE
+        return obj
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        try:
+            ret = self._unpack(EX_CONSTRUCT)
+            self._consume()
+            return ret
+        except OutOfData:
+            self._consume()
+            raise StopIteration
+        except RecursionError:
+            raise StackError
+
+    next = __next__
+
+    def skip(self):
+        self._unpack(EX_SKIP)
+        self._consume()
+
+    def unpack(self):
+        try:
+            ret = self._unpack(EX_CONSTRUCT)
+        except RecursionError:
+            raise StackError
+        self._consume()
+        return ret
+
+    def read_array_header(self):
+        ret = self._unpack(EX_READ_ARRAY_HEADER)
+        self._consume()
+        return ret
+
+    def read_map_header(self):
+        ret = self._unpack(EX_READ_MAP_HEADER)
+        self._consume()
+        return ret
+
+    def tell(self):
+        return self._stream_offset
+
+
+class Packer:
+    """
+    MessagePack Packer
+
+    Usage::
+
+        packer = Packer()
+        astream.write(packer.pack(a))
+        astream.write(packer.pack(b))
+
+    Packer's constructor has some keyword arguments:
+
+    :param default:
+        When specified, it should be callable.
+        Convert user type to builtin type that Packer supports.
+        See also simplejson's document.
+
+    :param bool use_single_float:
+        Use single precision float type for float. (default: False)
+
+    :param bool autoreset:
+        Reset buffer after each pack and return its content as `bytes`. (default: True).
+        If set this to false, use `bytes()` to get content and `.reset()` to clear buffer.
+
+    :param bool use_bin_type:
+        Use bin type introduced in msgpack spec 2.0 for bytes.
+        It also enables str8 type for unicode. (default: True)
+
+    :param bool strict_types:
+        If set to true, types will be checked to be exact. Derived classes
+        from serializable types will not be serialized and will be
+        treated as unsupported type and forwarded to default.
+        Additionally tuples will not be serialized as lists.
+        This is useful when trying to implement accurate serialization
+        for python types.
+
+    :param bool datetime:
+        If set to true, datetime with tzinfo is packed into Timestamp type.
+        Note that the tzinfo is stripped in the timestamp.
+        You can get UTC datetime with `timestamp=3` option of the Unpacker.
+
+    :param str unicode_errors:
+        The error handler for encoding unicode. (default: 'strict')
+        DO NOT USE THIS!!  This option is kept for very specific usage.
+
+    :param int buf_size:
+        Internal buffer size. This option is used only for C implementation.
+    """
+
+    def __init__(
+        self,
+        *,
+        default=None,
+        use_single_float=False,
+        autoreset=True,
+        use_bin_type=True,
+        strict_types=False,
+        datetime=False,
+        unicode_errors=None,
+        buf_size=None,
+    ):
+        self._strict_types = strict_types
+        self._use_float = use_single_float
+        self._autoreset = autoreset
+        self._use_bin_type = use_bin_type
+        self._buffer = BytesIO()
+        self._datetime = bool(datetime)
+        self._unicode_errors = unicode_errors or "strict"
+        if default is not None and not callable(default):
+            raise TypeError("default must be callable")
+        self._default = default
+
+    def _pack(
+        self,
+        obj,
+        nest_limit=DEFAULT_RECURSE_LIMIT,
+        check=isinstance,
+        check_type_strict=_check_type_strict,
+    ):
+        default_used = False
+        if self._strict_types:
+            check = check_type_strict
+            list_types = list
+        else:
+            list_types = (list, tuple)
+        while True:
+            if nest_limit < 0:
+                raise ValueError("recursion limit exceeded")
+            if obj is None:
+                return self._buffer.write(b"\xc0")
+            if check(obj, bool):
+                if obj:
+                    return self._buffer.write(b"\xc3")
+                return self._buffer.write(b"\xc2")
+            if check(obj, int):
+                if 0 <= obj < 0x80:
+                    return self._buffer.write(struct.pack("B", obj))
+                if -0x20 <= obj < 0:
+                    return self._buffer.write(struct.pack("b", obj))
+                if 0x80 <= obj <= 0xFF:
+                    return self._buffer.write(struct.pack("BB", 0xCC, obj))
+                if -0x80 <= obj < 0:
+                    return self._buffer.write(struct.pack(">Bb", 0xD0, obj))
+                if 0xFF < obj <= 0xFFFF:
+                    return self._buffer.write(struct.pack(">BH", 0xCD, obj))
+                if -0x8000 <= obj < -0x80:
+                    return self._buffer.write(struct.pack(">Bh", 0xD1, obj))
+                if 0xFFFF < obj <= 0xFFFFFFFF:
+                    return self._buffer.write(struct.pack(">BI", 0xCE, obj))
+                if -0x80000000 <= obj < -0x8000:
+                    return self._buffer.write(struct.pack(">Bi", 0xD2, obj))
+                if 0xFFFFFFFF < obj <= 0xFFFFFFFFFFFFFFFF:
+                    return self._buffer.write(struct.pack(">BQ", 0xCF, obj))
+                if -0x8000000000000000 <= obj < -0x80000000:
+                    return self._buffer.write(struct.pack(">Bq", 0xD3, obj))
+                if not default_used and self._default is not None:
+                    obj = self._default(obj)
+                    default_used = True
+                    continue
+                raise OverflowError("Integer value out of range")
+            if check(obj, (bytes, bytearray)):
+                n = len(obj)
+                if n >= 2**32:
+                    raise ValueError("%s is too large" % type(obj).__name__)
+                self._pack_bin_header(n)
+                return self._buffer.write(obj)
+            if check(obj, str):
+                obj = obj.encode("utf-8", self._unicode_errors)
+                n = len(obj)
+                if n >= 2**32:
+                    raise ValueError("String is too large")
+                self._pack_raw_header(n)
+                return self._buffer.write(obj)
+            if check(obj, memoryview):
+                n = obj.nbytes
+                if n >= 2**32:
+                    raise ValueError("Memoryview is too large")
+                self._pack_bin_header(n)
+                return self._buffer.write(obj)
+            if check(obj, float):
+                if self._use_float:
+                    return self._buffer.write(struct.pack(">Bf", 0xCA, obj))
+                return self._buffer.write(struct.pack(">Bd", 0xCB, obj))
+            if check(obj, (ExtType, Timestamp)):
+                if check(obj, Timestamp):
+                    code = -1
+                    data = obj.to_bytes()
+                else:
+                    code = obj.code
+                    data = obj.data
+                assert isinstance(code, int)
+                assert isinstance(data, bytes)
+                L = len(data)
+                if L == 1:
+                    self._buffer.write(b"\xd4")
+                elif L == 2:
+                    self._buffer.write(b"\xd5")
+                elif L == 4:
+                    self._buffer.write(b"\xd6")
+                elif L == 8:
+                    self._buffer.write(b"\xd7")
+                elif L == 16:
+                    self._buffer.write(b"\xd8")
+                elif L <= 0xFF:
+                    self._buffer.write(struct.pack(">BB", 0xC7, L))
+                elif L <= 0xFFFF:
+                    self._buffer.write(struct.pack(">BH", 0xC8, L))
+                else:
+                    self._buffer.write(struct.pack(">BI", 0xC9, L))
+                self._buffer.write(struct.pack("b", code))
+                self._buffer.write(data)
+                return
+            if check(obj, list_types):
+                n = len(obj)
+                self._pack_array_header(n)
+                for i in range(n):
+                    self._pack(obj[i], nest_limit - 1)
+                return
+            if check(obj, dict):
+                return self._pack_map_pairs(len(obj), obj.items(), nest_limit - 1)
+
+            if self._datetime and check(obj, _DateTime) and obj.tzinfo is not None:
+                obj = Timestamp.from_datetime(obj)
+                default_used = 1
+                continue
+
+            if not default_used and self._default is not None:
+                obj = self._default(obj)
+                default_used = 1
+                continue
+
+            if self._datetime and check(obj, _DateTime):
+                raise ValueError(f"Cannot serialize {obj!r} where tzinfo=None")
+
+            raise TypeError(f"Cannot serialize {obj!r}")
+
+    def pack(self, obj):
+        try:
+            self._pack(obj)
+        except:
+            self._buffer = BytesIO()  # force reset
+            raise
+        if self._autoreset:
+            ret = self._buffer.getvalue()
+            self._buffer = BytesIO()
+            return ret
+
+    def pack_map_pairs(self, pairs):
+        self._pack_map_pairs(len(pairs), pairs)
+        if self._autoreset:
+            ret = self._buffer.getvalue()
+            self._buffer = BytesIO()
+            return ret
+
+    def pack_array_header(self, n):
+        if n >= 2**32:
+            raise ValueError
+        self._pack_array_header(n)
+        if self._autoreset:
+            ret = self._buffer.getvalue()
+            self._buffer = BytesIO()
+            return ret
+
+    def pack_map_header(self, n):
+        if n >= 2**32:
+            raise ValueError
+        self._pack_map_header(n)
+        if self._autoreset:
+            ret = self._buffer.getvalue()
+            self._buffer = BytesIO()
+            return ret
+
+    def pack_ext_type(self, typecode, data):
+        if not isinstance(typecode, int):
+            raise TypeError("typecode must have int type.")
+        if not 0 <= typecode <= 127:
+            raise ValueError("typecode should be 0-127")
+        if not isinstance(data, bytes):
+            raise TypeError("data must have bytes type")
+        L = len(data)
+        if L > 0xFFFFFFFF:
+            raise ValueError("Too large data")
+        if L == 1:
+            self._buffer.write(b"\xd4")
+        elif L == 2:
+            self._buffer.write(b"\xd5")
+        elif L == 4:
+            self._buffer.write(b"\xd6")
+        elif L == 8:
+            self._buffer.write(b"\xd7")
+        elif L == 16:
+            self._buffer.write(b"\xd8")
+        elif L <= 0xFF:
+            self._buffer.write(b"\xc7" + struct.pack("B", L))
+        elif L <= 0xFFFF:
+            self._buffer.write(b"\xc8" + struct.pack(">H", L))
+        else:
+            self._buffer.write(b"\xc9" + struct.pack(">I", L))
+        self._buffer.write(struct.pack("B", typecode))
+        self._buffer.write(data)
+
+    def _pack_array_header(self, n):
+        if n <= 0x0F:
+            return self._buffer.write(struct.pack("B", 0x90 + n))
+        if n <= 0xFFFF:
+            return self._buffer.write(struct.pack(">BH", 0xDC, n))
+        if n <= 0xFFFFFFFF:
+            return self._buffer.write(struct.pack(">BI", 0xDD, n))
+        raise ValueError("Array is too large")
+
+    def _pack_map_header(self, n):
+        if n <= 0x0F:
+            return self._buffer.write(struct.pack("B", 0x80 + n))
+        if n <= 0xFFFF:
+            return self._buffer.write(struct.pack(">BH", 0xDE, n))
+        if n <= 0xFFFFFFFF:
+            return self._buffer.write(struct.pack(">BI", 0xDF, n))
+        raise ValueError("Dict is too large")
+
+    def _pack_map_pairs(self, n, pairs, nest_limit=DEFAULT_RECURSE_LIMIT):
+        self._pack_map_header(n)
+        for k, v in pairs:
+            self._pack(k, nest_limit - 1)
+            self._pack(v, nest_limit - 1)
+
+    def _pack_raw_header(self, n):
+        if n <= 0x1F:
+            self._buffer.write(struct.pack("B", 0xA0 + n))
+        elif self._use_bin_type and n <= 0xFF:
+            self._buffer.write(struct.pack(">BB", 0xD9, n))
+        elif n <= 0xFFFF:
+            self._buffer.write(struct.pack(">BH", 0xDA, n))
+        elif n <= 0xFFFFFFFF:
+            self._buffer.write(struct.pack(">BI", 0xDB, n))
+        else:
+            raise ValueError("Raw is too large")
+
+    def _pack_bin_header(self, n):
+        if not self._use_bin_type:
+            return self._pack_raw_header(n)
+        elif n <= 0xFF:
+            return self._buffer.write(struct.pack(">BB", 0xC4, n))
+        elif n <= 0xFFFF:
+            return self._buffer.write(struct.pack(">BH", 0xC5, n))
+        elif n <= 0xFFFFFFFF:
+            return self._buffer.write(struct.pack(">BI", 0xC6, n))
+        else:
+            raise ValueError("Bin is too large")
+
+    def bytes(self):
+        """Return internal buffer contents as bytes object"""
+        return self._buffer.getvalue()
+
+    def reset(self):
+        """Reset internal buffer.
+
+        This method is useful only when autoreset=False.
+        """
+        self._buffer = BytesIO()
+
+    def getbuffer(self):
+        """Return view of internal buffer."""
+        if _USING_STRINGBUILDER:
+            return memoryview(self.bytes())
+        else:
+            return self._buffer.getbuffer()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__init__.py
new file mode 100644
index 00000000..d79f73c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__init__.py
@@ -0,0 +1,15 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+__title__ = "packaging"
+__summary__ = "Core utilities for Python packages"
+__uri__ = "https://github.com/pypa/packaging"
+
+__version__ = "24.2"
+
+__author__ = "Donald Stufft and individual contributors"
+__email__ = "donald@stufft.io"
+
+__license__ = "BSD-2-Clause or Apache-2.0"
+__copyright__ = f"2014 {__author__}"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..39eae300
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc
new file mode 100644
index 00000000..6e230a3a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc
new file mode 100644
index 00000000..b9f17ce8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc
new file mode 100644
index 00000000..52412997
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..f9039f64
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_structures.cpython-312.pyc
new file mode 100644
index 00000000..15163fb1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc
new file mode 100644
index 00000000..62337d88
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/markers.cpython-312.pyc
new file mode 100644
index 00000000..3b1d6047
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/metadata.cpython-312.pyc
new file mode 100644
index 00000000..c7d12243
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..c94db5a7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc
new file mode 100644
index 00000000..35378317
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/tags.cpython-312.pyc
new file mode 100644
index 00000000..27dc979a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..3e9563f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..d22e6f0b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_elffile.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_elffile.py
new file mode 100644
index 00000000..25f4282c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_elffile.py
@@ -0,0 +1,110 @@
+"""
+ELF file parser.
+
+This provides a class ``ELFFile`` that parses an ELF executable in a similar
+interface to ``ZipFile``. Only the read interface is implemented.
+
+Based on: https://gist.github.com/lyssdod/f51579ae8d93c8657a5564aefc2ffbca
+ELF header: https://refspecs.linuxfoundation.org/elf/gabi4+/ch4.eheader.html
+"""
+
+from __future__ import annotations
+
+import enum
+import os
+import struct
+from typing import IO
+
+
+class ELFInvalid(ValueError):
+    pass
+
+
+class EIClass(enum.IntEnum):
+    C32 = 1
+    C64 = 2
+
+
+class EIData(enum.IntEnum):
+    Lsb = 1
+    Msb = 2
+
+
+class EMachine(enum.IntEnum):
+    I386 = 3
+    S390 = 22
+    Arm = 40
+    X8664 = 62
+    AArc64 = 183
+
+
+class ELFFile:
+    """
+    Representation of an ELF executable.
+    """
+
+    def __init__(self, f: IO[bytes]) -> None:
+        self._f = f
+
+        try:
+            ident = self._read("16B")
+        except struct.error as e:
+            raise ELFInvalid("unable to parse identification") from e
+        magic = bytes(ident[:4])
+        if magic != b"\x7fELF":
+            raise ELFInvalid(f"invalid magic: {magic!r}")
+
+        self.capacity = ident[4]  # Format for program header (bitness).
+        self.encoding = ident[5]  # Data structure encoding (endianness).
+
+        try:
+            # e_fmt: Format for program header.
+            # p_fmt: Format for section header.
+            # p_idx: Indexes to find p_type, p_offset, and p_filesz.
+            e_fmt, self._p_fmt, self._p_idx = {
+                (1, 1): ("<HHIIIIIHHH", "<IIIIIIII", (0, 1, 4)),  # 32-bit LSB.
+                (1, 2): (">HHIIIIIHHH", ">IIIIIIII", (0, 1, 4)),  # 32-bit MSB.
+                (2, 1): ("<HHIQQQIHHH", "<IIQQQQQQ", (0, 2, 5)),  # 64-bit LSB.
+                (2, 2): (">HHIQQQIHHH", ">IIQQQQQQ", (0, 2, 5)),  # 64-bit MSB.
+            }[(self.capacity, self.encoding)]
+        except KeyError as e:
+            raise ELFInvalid(
+                f"unrecognized capacity ({self.capacity}) or "
+                f"encoding ({self.encoding})"
+            ) from e
+
+        try:
+            (
+                _,
+                self.machine,  # Architecture type.
+                _,
+                _,
+                self._e_phoff,  # Offset of program header.
+                _,
+                self.flags,  # Processor-specific flags.
+                _,
+                self._e_phentsize,  # Size of section.
+                self._e_phnum,  # Number of sections.
+            ) = self._read(e_fmt)
+        except struct.error as e:
+            raise ELFInvalid("unable to parse machine and section information") from e
+
+    def _read(self, fmt: str) -> tuple[int, ...]:
+        return struct.unpack(fmt, self._f.read(struct.calcsize(fmt)))
+
+    @property
+    def interpreter(self) -> str | None:
+        """
+        The path recorded in the ``PT_INTERP`` section header.
+        """
+        for index in range(self._e_phnum):
+            self._f.seek(self._e_phoff + self._e_phentsize * index)
+            try:
+                data = self._read(self._p_fmt)
+            except struct.error:
+                continue
+            if data[self._p_idx[0]] != 3:  # Not PT_INTERP.
+                continue
+            self._f.seek(data[self._p_idx[1]])
+            return os.fsdecode(self._f.read(data[self._p_idx[2]])).strip("\0")
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_manylinux.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_manylinux.py
new file mode 100644
index 00000000..61339a6f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_manylinux.py
@@ -0,0 +1,263 @@
+from __future__ import annotations
+
+import collections
+import contextlib
+import functools
+import os
+import re
+import sys
+import warnings
+from typing import Generator, Iterator, NamedTuple, Sequence
+
+from ._elffile import EIClass, EIData, ELFFile, EMachine
+
+EF_ARM_ABIMASK = 0xFF000000
+EF_ARM_ABI_VER5 = 0x05000000
+EF_ARM_ABI_FLOAT_HARD = 0x00000400
+
+
+# `os.PathLike` not a generic type until Python 3.9, so sticking with `str`
+# as the type for `path` until then.
+@contextlib.contextmanager
+def _parse_elf(path: str) -> Generator[ELFFile | None, None, None]:
+    try:
+        with open(path, "rb") as f:
+            yield ELFFile(f)
+    except (OSError, TypeError, ValueError):
+        yield None
+
+
+def _is_linux_armhf(executable: str) -> bool:
+    # hard-float ABI can be detected from the ELF header of the running
+    # process
+    # https://static.docs.arm.com/ihi0044/g/aaelf32.pdf
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.Arm
+            and f.flags & EF_ARM_ABIMASK == EF_ARM_ABI_VER5
+            and f.flags & EF_ARM_ABI_FLOAT_HARD == EF_ARM_ABI_FLOAT_HARD
+        )
+
+
+def _is_linux_i686(executable: str) -> bool:
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.I386
+        )
+
+
+def _have_compatible_abi(executable: str, archs: Sequence[str]) -> bool:
+    if "armv7l" in archs:
+        return _is_linux_armhf(executable)
+    if "i686" in archs:
+        return _is_linux_i686(executable)
+    allowed_archs = {
+        "x86_64",
+        "aarch64",
+        "ppc64",
+        "ppc64le",
+        "s390x",
+        "loongarch64",
+        "riscv64",
+    }
+    return any(arch in allowed_archs for arch in archs)
+
+
+# If glibc ever changes its major version, we need to know what the last
+# minor version was, so we can build the complete list of all versions.
+# For now, guess what the highest minor version might be, assume it will
+# be 50 for testing. Once this actually happens, update the dictionary
+# with the actual value.
+_LAST_GLIBC_MINOR: dict[int, int] = collections.defaultdict(lambda: 50)
+
+
+class _GLibCVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _glibc_version_string_confstr() -> str | None:
+    """
+    Primary implementation of glibc_version_string using os.confstr.
+    """
+    # os.confstr is quite a bit faster than ctypes.DLL. It's also less likely
+    # to be broken or missing. This strategy is used in the standard library
+    # platform module.
+    # https://github.com/python/cpython/blob/fcf1d003bf4f0100c/Lib/platform.py#L175-L183
+    try:
+        # Should be a string like "glibc 2.17".
+        version_string: str | None = os.confstr("CS_GNU_LIBC_VERSION")
+        assert version_string is not None
+        _, version = version_string.rsplit()
+    except (AssertionError, AttributeError, OSError, ValueError):
+        # os.confstr() or CS_GNU_LIBC_VERSION not available (or a bad value)...
+        return None
+    return version
+
+
+def _glibc_version_string_ctypes() -> str | None:
+    """
+    Fallback implementation of glibc_version_string using ctypes.
+    """
+    try:
+        import ctypes
+    except ImportError:
+        return None
+
+    # ctypes.CDLL(None) internally calls dlopen(NULL), and as the dlopen
+    # manpage says, "If filename is NULL, then the returned handle is for the
+    # main program". This way we can let the linker do the work to figure out
+    # which libc our process is actually using.
+    #
+    # We must also handle the special case where the executable is not a
+    # dynamically linked executable. This can occur when using musl libc,
+    # for example. In this situation, dlopen() will error, leading to an
+    # OSError. Interestingly, at least in the case of musl, there is no
+    # errno set on the OSError. The single string argument used to construct
+    # OSError comes from libc itself and is therefore not portable to
+    # hard code here. In any case, failure to call dlopen() means we
+    # can proceed, so we bail on our attempt.
+    try:
+        process_namespace = ctypes.CDLL(None)
+    except OSError:
+        return None
+
+    try:
+        gnu_get_libc_version = process_namespace.gnu_get_libc_version
+    except AttributeError:
+        # Symbol doesn't exist -> therefore, we are not linked to
+        # glibc.
+        return None
+
+    # Call gnu_get_libc_version, which returns a string like "2.5"
+    gnu_get_libc_version.restype = ctypes.c_char_p
+    version_str: str = gnu_get_libc_version()
+    # py2 / py3 compatibility:
+    if not isinstance(version_str, str):
+        version_str = version_str.decode("ascii")
+
+    return version_str
+
+
+def _glibc_version_string() -> str | None:
+    """Returns glibc version string, or None if not using glibc."""
+    return _glibc_version_string_confstr() or _glibc_version_string_ctypes()
+
+
+def _parse_glibc_version(version_str: str) -> tuple[int, int]:
+    """Parse glibc version.
+
+    We use a regexp instead of str.split because we want to discard any
+    random junk that might come after the minor version -- this might happen
+    in patched/forked versions of glibc (e.g. Linaro's version of glibc
+    uses version strings like "2.20-2014.11"). See gh-3588.
+    """
+    m = re.match(r"(?P<major>[0-9]+)\.(?P<minor>[0-9]+)", version_str)
+    if not m:
+        warnings.warn(
+            f"Expected glibc version with 2 components major.minor,"
+            f" got: {version_str}",
+            RuntimeWarning,
+            stacklevel=2,
+        )
+        return -1, -1
+    return int(m.group("major")), int(m.group("minor"))
+
+
+@functools.lru_cache
+def _get_glibc_version() -> tuple[int, int]:
+    version_str = _glibc_version_string()
+    if version_str is None:
+        return (-1, -1)
+    return _parse_glibc_version(version_str)
+
+
+# From PEP 513, PEP 600
+def _is_compatible(arch: str, version: _GLibCVersion) -> bool:
+    sys_glibc = _get_glibc_version()
+    if sys_glibc < version:
+        return False
+    # Check for presence of _manylinux module.
+    try:
+        import _manylinux
+    except ImportError:
+        return True
+    if hasattr(_manylinux, "manylinux_compatible"):
+        result = _manylinux.manylinux_compatible(version[0], version[1], arch)
+        if result is not None:
+            return bool(result)
+        return True
+    if version == _GLibCVersion(2, 5):
+        if hasattr(_manylinux, "manylinux1_compatible"):
+            return bool(_manylinux.manylinux1_compatible)
+    if version == _GLibCVersion(2, 12):
+        if hasattr(_manylinux, "manylinux2010_compatible"):
+            return bool(_manylinux.manylinux2010_compatible)
+    if version == _GLibCVersion(2, 17):
+        if hasattr(_manylinux, "manylinux2014_compatible"):
+            return bool(_manylinux.manylinux2014_compatible)
+    return True
+
+
+_LEGACY_MANYLINUX_MAP = {
+    # CentOS 7 w/ glibc 2.17 (PEP 599)
+    (2, 17): "manylinux2014",
+    # CentOS 6 w/ glibc 2.12 (PEP 571)
+    (2, 12): "manylinux2010",
+    # CentOS 5 w/ glibc 2.5 (PEP 513)
+    (2, 5): "manylinux1",
+}
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate manylinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be manylinux-compatible.
+
+    :returns: An iterator of compatible manylinux tags.
+    """
+    if not _have_compatible_abi(sys.executable, archs):
+        return
+    # Oldest glibc to be supported regardless of architecture is (2, 17).
+    too_old_glibc2 = _GLibCVersion(2, 16)
+    if set(archs) & {"x86_64", "i686"}:
+        # On x86/i686 also oldest glibc to be supported is (2, 5).
+        too_old_glibc2 = _GLibCVersion(2, 4)
+    current_glibc = _GLibCVersion(*_get_glibc_version())
+    glibc_max_list = [current_glibc]
+    # We can assume compatibility across glibc major versions.
+    # https://sourceware.org/bugzilla/show_bug.cgi?id=24636
+    #
+    # Build a list of maximum glibc versions so that we can
+    # output the canonical list of all glibc from current_glibc
+    # down to too_old_glibc2, including all intermediary versions.
+    for glibc_major in range(current_glibc.major - 1, 1, -1):
+        glibc_minor = _LAST_GLIBC_MINOR[glibc_major]
+        glibc_max_list.append(_GLibCVersion(glibc_major, glibc_minor))
+    for arch in archs:
+        for glibc_max in glibc_max_list:
+            if glibc_max.major == too_old_glibc2.major:
+                min_minor = too_old_glibc2.minor
+            else:
+                # For other glibc major versions oldest supported is (x, 0).
+                min_minor = -1
+            for glibc_minor in range(glibc_max.minor, min_minor, -1):
+                glibc_version = _GLibCVersion(glibc_max.major, glibc_minor)
+                tag = "manylinux_{}_{}".format(*glibc_version)
+                if _is_compatible(arch, glibc_version):
+                    yield f"{tag}_{arch}"
+                # Handle the legacy manylinux1, manylinux2010, manylinux2014 tags.
+                if glibc_version in _LEGACY_MANYLINUX_MAP:
+                    legacy_tag = _LEGACY_MANYLINUX_MAP[glibc_version]
+                    if _is_compatible(arch, glibc_version):
+                        yield f"{legacy_tag}_{arch}"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_musllinux.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_musllinux.py
new file mode 100644
index 00000000..d2bf30b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_musllinux.py
@@ -0,0 +1,85 @@
+"""PEP 656 support.
+
+This module implements logic to detect if the currently running Python is
+linked against musl, and what musl version is used.
+"""
+
+from __future__ import annotations
+
+import functools
+import re
+import subprocess
+import sys
+from typing import Iterator, NamedTuple, Sequence
+
+from ._elffile import ELFFile
+
+
+class _MuslVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _parse_musl_version(output: str) -> _MuslVersion | None:
+    lines = [n for n in (n.strip() for n in output.splitlines()) if n]
+    if len(lines) < 2 or lines[0][:4] != "musl":
+        return None
+    m = re.match(r"Version (\d+)\.(\d+)", lines[1])
+    if not m:
+        return None
+    return _MuslVersion(major=int(m.group(1)), minor=int(m.group(2)))
+
+
+@functools.lru_cache
+def _get_musl_version(executable: str) -> _MuslVersion | None:
+    """Detect currently-running musl runtime version.
+
+    This is done by checking the specified executable's dynamic linking
+    information, and invoking the loader to parse its output for a version
+    string. If the loader is musl, the output would be something like::
+
+        musl libc (x86_64)
+        Version 1.2.2
+        Dynamic Program Loader
+    """
+    try:
+        with open(executable, "rb") as f:
+            ld = ELFFile(f).interpreter
+    except (OSError, TypeError, ValueError):
+        return None
+    if ld is None or "musl" not in ld:
+        return None
+    proc = subprocess.run([ld], stderr=subprocess.PIPE, text=True)
+    return _parse_musl_version(proc.stderr)
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate musllinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be musllinux-compatible.
+
+    :returns: An iterator of compatible musllinux tags.
+    """
+    sys_musl = _get_musl_version(sys.executable)
+    if sys_musl is None:  # Python not dynamically linked against musl.
+        return
+    for arch in archs:
+        for minor in range(sys_musl.minor, -1, -1):
+            yield f"musllinux_{sys_musl.major}_{minor}_{arch}"
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sysconfig
+
+    plat = sysconfig.get_platform()
+    assert plat.startswith("linux-"), "not linux"
+
+    print("plat:", plat)
+    print("musl:", _get_musl_version(sys.executable))
+    print("tags:", end=" ")
+    for t in platform_tags(re.sub(r"[.-]", "_", plat.split("-", 1)[-1])):
+        print(t, end="\n      ")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_parser.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_parser.py
new file mode 100644
index 00000000..c1238c06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_parser.py
@@ -0,0 +1,354 @@
+"""Handwritten parser of dependency specifiers.
+
+The docstring for each __parse_* function contains EBNF-inspired grammar representing
+the implementation.
+"""
+
+from __future__ import annotations
+
+import ast
+from typing import NamedTuple, Sequence, Tuple, Union
+
+from ._tokenizer import DEFAULT_RULES, Tokenizer
+
+
+class Node:
+    def __init__(self, value: str) -> None:
+        self.value = value
+
+    def __str__(self) -> str:
+        return self.value
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__}('{self}')>"
+
+    def serialize(self) -> str:
+        raise NotImplementedError
+
+
+class Variable(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+class Value(Node):
+    def serialize(self) -> str:
+        return f'"{self}"'
+
+
+class Op(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+MarkerVar = Union[Variable, Value]
+MarkerItem = Tuple[MarkerVar, Op, MarkerVar]
+MarkerAtom = Union[MarkerItem, Sequence["MarkerAtom"]]
+MarkerList = Sequence[Union["MarkerList", MarkerAtom, str]]
+
+
+class ParsedRequirement(NamedTuple):
+    name: str
+    url: str
+    extras: list[str]
+    specifier: str
+    marker: MarkerList | None
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for dependency specifier
+# --------------------------------------------------------------------------------------
+def parse_requirement(source: str) -> ParsedRequirement:
+    return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_requirement(tokenizer: Tokenizer) -> ParsedRequirement:
+    """
+    requirement = WS? IDENTIFIER WS? extras WS? requirement_details
+    """
+    tokenizer.consume("WS")
+
+    name_token = tokenizer.expect(
+        "IDENTIFIER", expected="package name at the start of dependency specifier"
+    )
+    name = name_token.text
+    tokenizer.consume("WS")
+
+    extras = _parse_extras(tokenizer)
+    tokenizer.consume("WS")
+
+    url, specifier, marker = _parse_requirement_details(tokenizer)
+    tokenizer.expect("END", expected="end of dependency specifier")
+
+    return ParsedRequirement(name, url, extras, specifier, marker)
+
+
+def _parse_requirement_details(
+    tokenizer: Tokenizer,
+) -> tuple[str, str, MarkerList | None]:
+    """
+    requirement_details = AT URL (WS requirement_marker?)?
+                        | specifier WS? (requirement_marker)?
+    """
+
+    specifier = ""
+    url = ""
+    marker = None
+
+    if tokenizer.check("AT"):
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        url_start = tokenizer.position
+        url = tokenizer.expect("URL", expected="URL after @").text
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        tokenizer.expect("WS", expected="whitespace after URL")
+
+        # The input might end after whitespace.
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer, span_start=url_start, after="URL and whitespace"
+        )
+    else:
+        specifier_start = tokenizer.position
+        specifier = _parse_specifier(tokenizer)
+        tokenizer.consume("WS")
+
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer,
+            span_start=specifier_start,
+            after=(
+                "version specifier"
+                if specifier
+                else "name and no valid version specifier"
+            ),
+        )
+
+    return (url, specifier, marker)
+
+
+def _parse_requirement_marker(
+    tokenizer: Tokenizer, *, span_start: int, after: str
+) -> MarkerList:
+    """
+    requirement_marker = SEMICOLON marker WS?
+    """
+
+    if not tokenizer.check("SEMICOLON"):
+        tokenizer.raise_syntax_error(
+            f"Expected end or semicolon (after {after})",
+            span_start=span_start,
+        )
+    tokenizer.read()
+
+    marker = _parse_marker(tokenizer)
+    tokenizer.consume("WS")
+
+    return marker
+
+
+def _parse_extras(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras = (LEFT_BRACKET wsp* extras_list? wsp* RIGHT_BRACKET)?
+    """
+    if not tokenizer.check("LEFT_BRACKET", peek=True):
+        return []
+
+    with tokenizer.enclosing_tokens(
+        "LEFT_BRACKET",
+        "RIGHT_BRACKET",
+        around="extras",
+    ):
+        tokenizer.consume("WS")
+        extras = _parse_extras_list(tokenizer)
+        tokenizer.consume("WS")
+
+    return extras
+
+
+def _parse_extras_list(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras_list = identifier (wsp* ',' wsp* identifier)*
+    """
+    extras: list[str] = []
+
+    if not tokenizer.check("IDENTIFIER"):
+        return extras
+
+    extras.append(tokenizer.read().text)
+
+    while True:
+        tokenizer.consume("WS")
+        if tokenizer.check("IDENTIFIER", peek=True):
+            tokenizer.raise_syntax_error("Expected comma between extra names")
+        elif not tokenizer.check("COMMA"):
+            break
+
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        extra_token = tokenizer.expect("IDENTIFIER", expected="extra name after comma")
+        extras.append(extra_token.text)
+
+    return extras
+
+
+def _parse_specifier(tokenizer: Tokenizer) -> str:
+    """
+    specifier = LEFT_PARENTHESIS WS? version_many WS? RIGHT_PARENTHESIS
+              | WS? version_many WS?
+    """
+    with tokenizer.enclosing_tokens(
+        "LEFT_PARENTHESIS",
+        "RIGHT_PARENTHESIS",
+        around="version specifier",
+    ):
+        tokenizer.consume("WS")
+        parsed_specifiers = _parse_version_many(tokenizer)
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+def _parse_version_many(tokenizer: Tokenizer) -> str:
+    """
+    version_many = (SPECIFIER (WS? COMMA WS? SPECIFIER)*)?
+    """
+    parsed_specifiers = ""
+    while tokenizer.check("SPECIFIER"):
+        span_start = tokenizer.position
+        parsed_specifiers += tokenizer.read().text
+        if tokenizer.check("VERSION_PREFIX_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                ".* suffix can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position + 1,
+            )
+        if tokenizer.check("VERSION_LOCAL_LABEL_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                "Local version label can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position,
+            )
+        tokenizer.consume("WS")
+        if not tokenizer.check("COMMA"):
+            break
+        parsed_specifiers += tokenizer.read().text
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for marker expression
+# --------------------------------------------------------------------------------------
+def parse_marker(source: str) -> MarkerList:
+    return _parse_full_marker(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_full_marker(tokenizer: Tokenizer) -> MarkerList:
+    retval = _parse_marker(tokenizer)
+    tokenizer.expect("END", expected="end of marker expression")
+    return retval
+
+
+def _parse_marker(tokenizer: Tokenizer) -> MarkerList:
+    """
+    marker = marker_atom (BOOLOP marker_atom)+
+    """
+    expression = [_parse_marker_atom(tokenizer)]
+    while tokenizer.check("BOOLOP"):
+        token = tokenizer.read()
+        expr_right = _parse_marker_atom(tokenizer)
+        expression.extend((token.text, expr_right))
+    return expression
+
+
+def _parse_marker_atom(tokenizer: Tokenizer) -> MarkerAtom:
+    """
+    marker_atom = WS? LEFT_PARENTHESIS WS? marker WS? RIGHT_PARENTHESIS WS?
+                | WS? marker_item WS?
+    """
+
+    tokenizer.consume("WS")
+    if tokenizer.check("LEFT_PARENTHESIS", peek=True):
+        with tokenizer.enclosing_tokens(
+            "LEFT_PARENTHESIS",
+            "RIGHT_PARENTHESIS",
+            around="marker expression",
+        ):
+            tokenizer.consume("WS")
+            marker: MarkerAtom = _parse_marker(tokenizer)
+            tokenizer.consume("WS")
+    else:
+        marker = _parse_marker_item(tokenizer)
+    tokenizer.consume("WS")
+    return marker
+
+
+def _parse_marker_item(tokenizer: Tokenizer) -> MarkerItem:
+    """
+    marker_item = WS? marker_var WS? marker_op WS? marker_var WS?
+    """
+    tokenizer.consume("WS")
+    marker_var_left = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    marker_op = _parse_marker_op(tokenizer)
+    tokenizer.consume("WS")
+    marker_var_right = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    return (marker_var_left, marker_op, marker_var_right)
+
+
+def _parse_marker_var(tokenizer: Tokenizer) -> MarkerVar:
+    """
+    marker_var = VARIABLE | QUOTED_STRING
+    """
+    if tokenizer.check("VARIABLE"):
+        return process_env_var(tokenizer.read().text.replace(".", "_"))
+    elif tokenizer.check("QUOTED_STRING"):
+        return process_python_str(tokenizer.read().text)
+    else:
+        tokenizer.raise_syntax_error(
+            message="Expected a marker variable or quoted string"
+        )
+
+
+def process_env_var(env_var: str) -> Variable:
+    if env_var in ("platform_python_implementation", "python_implementation"):
+        return Variable("platform_python_implementation")
+    else:
+        return Variable(env_var)
+
+
+def process_python_str(python_str: str) -> Value:
+    value = ast.literal_eval(python_str)
+    return Value(str(value))
+
+
+def _parse_marker_op(tokenizer: Tokenizer) -> Op:
+    """
+    marker_op = IN | NOT IN | OP
+    """
+    if tokenizer.check("IN"):
+        tokenizer.read()
+        return Op("in")
+    elif tokenizer.check("NOT"):
+        tokenizer.read()
+        tokenizer.expect("WS", expected="whitespace after 'not'")
+        tokenizer.expect("IN", expected="'in' after 'not'")
+        return Op("not in")
+    elif tokenizer.check("OP"):
+        return Op(tokenizer.read().text)
+    else:
+        return tokenizer.raise_syntax_error(
+            "Expected marker operator, one of "
+            "<=, <, !=, ==, >=, >, ~=, ===, in, not in"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_structures.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_structures.py
new file mode 100644
index 00000000..90a6465f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_structures.py
@@ -0,0 +1,61 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+
+class InfinityType:
+    def __repr__(self) -> str:
+        return "Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return False
+
+    def __le__(self, other: object) -> bool:
+        return False
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return True
+
+    def __ge__(self, other: object) -> bool:
+        return True
+
+    def __neg__(self: object) -> "NegativeInfinityType":
+        return NegativeInfinity
+
+
+Infinity = InfinityType()
+
+
+class NegativeInfinityType:
+    def __repr__(self) -> str:
+        return "-Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return True
+
+    def __le__(self, other: object) -> bool:
+        return True
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return False
+
+    def __ge__(self, other: object) -> bool:
+        return False
+
+    def __neg__(self: object) -> InfinityType:
+        return Infinity
+
+
+NegativeInfinity = NegativeInfinityType()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_tokenizer.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_tokenizer.py
new file mode 100644
index 00000000..89d04160
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/_tokenizer.py
@@ -0,0 +1,194 @@
+from __future__ import annotations
+
+import contextlib
+import re
+from dataclasses import dataclass
+from typing import Iterator, NoReturn
+
+from .specifiers import Specifier
+
+
+@dataclass
+class Token:
+    name: str
+    text: str
+    position: int
+
+
+class ParserSyntaxError(Exception):
+    """The provided source text could not be parsed correctly."""
+
+    def __init__(
+        self,
+        message: str,
+        *,
+        source: str,
+        span: tuple[int, int],
+    ) -> None:
+        self.span = span
+        self.message = message
+        self.source = source
+
+        super().__init__()
+
+    def __str__(self) -> str:
+        marker = " " * self.span[0] + "~" * (self.span[1] - self.span[0]) + "^"
+        return "\n    ".join([self.message, self.source, marker])
+
+
+DEFAULT_RULES: dict[str, str | re.Pattern[str]] = {
+    "LEFT_PARENTHESIS": r"\(",
+    "RIGHT_PARENTHESIS": r"\)",
+    "LEFT_BRACKET": r"\[",
+    "RIGHT_BRACKET": r"\]",
+    "SEMICOLON": r";",
+    "COMMA": r",",
+    "QUOTED_STRING": re.compile(
+        r"""
+            (
+                ('[^']*')
+                |
+                ("[^"]*")
+            )
+        """,
+        re.VERBOSE,
+    ),
+    "OP": r"(===|==|~=|!=|<=|>=|<|>)",
+    "BOOLOP": r"\b(or|and)\b",
+    "IN": r"\bin\b",
+    "NOT": r"\bnot\b",
+    "VARIABLE": re.compile(
+        r"""
+            \b(
+                python_version
+                |python_full_version
+                |os[._]name
+                |sys[._]platform
+                |platform_(release|system)
+                |platform[._](version|machine|python_implementation)
+                |python_implementation
+                |implementation_(name|version)
+                |extra
+            )\b
+        """,
+        re.VERBOSE,
+    ),
+    "SPECIFIER": re.compile(
+        Specifier._operator_regex_str + Specifier._version_regex_str,
+        re.VERBOSE | re.IGNORECASE,
+    ),
+    "AT": r"\@",
+    "URL": r"[^ \t]+",
+    "IDENTIFIER": r"\b[a-zA-Z0-9][a-zA-Z0-9._-]*\b",
+    "VERSION_PREFIX_TRAIL": r"\.\*",
+    "VERSION_LOCAL_LABEL_TRAIL": r"\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*",
+    "WS": r"[ \t]+",
+    "END": r"$",
+}
+
+
+class Tokenizer:
+    """Context-sensitive token parsing.
+
+    Provides methods to examine the input stream to check whether the next token
+    matches.
+    """
+
+    def __init__(
+        self,
+        source: str,
+        *,
+        rules: dict[str, str | re.Pattern[str]],
+    ) -> None:
+        self.source = source
+        self.rules: dict[str, re.Pattern[str]] = {
+            name: re.compile(pattern) for name, pattern in rules.items()
+        }
+        self.next_token: Token | None = None
+        self.position = 0
+
+    def consume(self, name: str) -> None:
+        """Move beyond provided token name, if at current position."""
+        if self.check(name):
+            self.read()
+
+    def check(self, name: str, *, peek: bool = False) -> bool:
+        """Check whether the next token has the provided name.
+
+        By default, if the check succeeds, the token *must* be read before
+        another check. If `peek` is set to `True`, the token is not loaded and
+        would need to be checked again.
+        """
+        assert (
+            self.next_token is None
+        ), f"Cannot check for {name!r}, already have {self.next_token!r}"
+        assert name in self.rules, f"Unknown token name: {name!r}"
+
+        expression = self.rules[name]
+
+        match = expression.match(self.source, self.position)
+        if match is None:
+            return False
+        if not peek:
+            self.next_token = Token(name, match[0], self.position)
+        return True
+
+    def expect(self, name: str, *, expected: str) -> Token:
+        """Expect a certain token name next, failing with a syntax error otherwise.
+
+        The token is *not* read.
+        """
+        if not self.check(name):
+            raise self.raise_syntax_error(f"Expected {expected}")
+        return self.read()
+
+    def read(self) -> Token:
+        """Consume the next token and return it."""
+        token = self.next_token
+        assert token is not None
+
+        self.position += len(token.text)
+        self.next_token = None
+
+        return token
+
+    def raise_syntax_error(
+        self,
+        message: str,
+        *,
+        span_start: int | None = None,
+        span_end: int | None = None,
+    ) -> NoReturn:
+        """Raise ParserSyntaxError at the given position."""
+        span = (
+            self.position if span_start is None else span_start,
+            self.position if span_end is None else span_end,
+        )
+        raise ParserSyntaxError(
+            message,
+            source=self.source,
+            span=span,
+        )
+
+    @contextlib.contextmanager
+    def enclosing_tokens(
+        self, open_token: str, close_token: str, *, around: str
+    ) -> Iterator[None]:
+        if self.check(open_token):
+            open_position = self.position
+            self.read()
+        else:
+            open_position = None
+
+        yield
+
+        if open_position is None:
+            return
+
+        if not self.check(close_token):
+            self.raise_syntax_error(
+                f"Expected matching {close_token} for {open_token}, after {around}",
+                span_start=open_position,
+            )
+
+        self.read()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__init__.py
new file mode 100644
index 00000000..71a1a779
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__init__.py
@@ -0,0 +1,145 @@
+#######################################################################################
+#
+# Adapted from:
+#  https://github.com/pypa/hatch/blob/5352e44/backend/src/hatchling/licenses/parse.py
+#
+# MIT License
+#
+# Copyright (c) 2017-present Ofek Lev <oss@ofek.dev>
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this
+# software and associated documentation files (the "Software"), to deal in the Software
+# without restriction, including without limitation the rights to use, copy, modify,
+# merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to the following
+# conditions:
+#
+# The above copyright notice and this permission notice shall be included in all copies
+# or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+# PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
+# CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
+# OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+#
+#
+# With additional allowance of arbitrary `LicenseRef-` identifiers, not just
+# `LicenseRef-Public-Domain` and `LicenseRef-Proprietary`.
+#
+#######################################################################################
+from __future__ import annotations
+
+import re
+from typing import NewType, cast
+
+from pip._vendor.packaging.licenses._spdx import EXCEPTIONS, LICENSES
+
+__all__ = [
+    "NormalizedLicenseExpression",
+    "InvalidLicenseExpression",
+    "canonicalize_license_expression",
+]
+
+license_ref_allowed = re.compile("^[A-Za-z0-9.-]*$")
+
+NormalizedLicenseExpression = NewType("NormalizedLicenseExpression", str)
+
+
+class InvalidLicenseExpression(ValueError):
+    """Raised when a license-expression string is invalid
+
+    >>> canonicalize_license_expression("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.licenses.InvalidLicenseExpression: Invalid license expression: 'invalid'
+    """
+
+
+def canonicalize_license_expression(
+    raw_license_expression: str,
+) -> NormalizedLicenseExpression:
+    if not raw_license_expression:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message)
+
+    # Pad any parentheses so tokenization can be achieved by merely splitting on
+    # whitespace.
+    license_expression = raw_license_expression.replace("(", " ( ").replace(")", " ) ")
+    licenseref_prefix = "LicenseRef-"
+    license_refs = {
+        ref.lower(): "LicenseRef-" + ref[len(licenseref_prefix) :]
+        for ref in license_expression.split()
+        if ref.lower().startswith(licenseref_prefix.lower())
+    }
+
+    # Normalize to lower case so we can look up licenses/exceptions
+    # and so boolean operators are Python-compatible.
+    license_expression = license_expression.lower()
+
+    tokens = license_expression.split()
+
+    # Rather than implementing boolean logic, we create an expression that Python can
+    # parse. Everything that is not involved with the grammar itself is treated as
+    # `False` and the expression should evaluate as such.
+    python_tokens = []
+    for token in tokens:
+        if token not in {"or", "and", "with", "(", ")"}:
+            python_tokens.append("False")
+        elif token == "with":
+            python_tokens.append("or")
+        elif token == "(" and python_tokens and python_tokens[-1] not in {"or", "and"}:
+            message = f"Invalid license expression: {raw_license_expression!r}"
+            raise InvalidLicenseExpression(message)
+        else:
+            python_tokens.append(token)
+
+    python_expression = " ".join(python_tokens)
+    try:
+        invalid = eval(python_expression, globals(), locals())
+    except Exception:
+        invalid = True
+
+    if invalid is not False:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message) from None
+
+    # Take a final pass to check for unknown licenses/exceptions.
+    normalized_tokens = []
+    for token in tokens:
+        if token in {"or", "and", "with", "(", ")"}:
+            normalized_tokens.append(token.upper())
+            continue
+
+        if normalized_tokens and normalized_tokens[-1] == "WITH":
+            if token not in EXCEPTIONS:
+                message = f"Unknown license exception: {token!r}"
+                raise InvalidLicenseExpression(message)
+
+            normalized_tokens.append(EXCEPTIONS[token]["id"])
+        else:
+            if token.endswith("+"):
+                final_token = token[:-1]
+                suffix = "+"
+            else:
+                final_token = token
+                suffix = ""
+
+            if final_token.startswith("licenseref-"):
+                if not license_ref_allowed.match(final_token):
+                    message = f"Invalid licenseref: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(license_refs[final_token] + suffix)
+            else:
+                if final_token not in LICENSES:
+                    message = f"Unknown license: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(LICENSES[final_token]["id"] + suffix)
+
+    normalized_expression = " ".join(normalized_tokens)
+
+    return cast(
+        NormalizedLicenseExpression,
+        normalized_expression.replace("( ", "(").replace(" )", ")"),
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..36b81dff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc
new file mode 100644
index 00000000..f7348679
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/_spdx.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/_spdx.py
new file mode 100644
index 00000000..eac22276
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/licenses/_spdx.py
@@ -0,0 +1,759 @@
+
+from __future__ import annotations
+
+from typing import TypedDict
+
+class SPDXLicense(TypedDict):
+    id: str
+    deprecated: bool
+
+class SPDXException(TypedDict):
+    id: str
+    deprecated: bool
+
+
+VERSION = '3.25.0'
+
+LICENSES: dict[str, SPDXLicense] = {
+    '0bsd': {'id': '0BSD', 'deprecated': False},
+    '3d-slicer-1.0': {'id': '3D-Slicer-1.0', 'deprecated': False},
+    'aal': {'id': 'AAL', 'deprecated': False},
+    'abstyles': {'id': 'Abstyles', 'deprecated': False},
+    'adacore-doc': {'id': 'AdaCore-doc', 'deprecated': False},
+    'adobe-2006': {'id': 'Adobe-2006', 'deprecated': False},
+    'adobe-display-postscript': {'id': 'Adobe-Display-PostScript', 'deprecated': False},
+    'adobe-glyph': {'id': 'Adobe-Glyph', 'deprecated': False},
+    'adobe-utopia': {'id': 'Adobe-Utopia', 'deprecated': False},
+    'adsl': {'id': 'ADSL', 'deprecated': False},
+    'afl-1.1': {'id': 'AFL-1.1', 'deprecated': False},
+    'afl-1.2': {'id': 'AFL-1.2', 'deprecated': False},
+    'afl-2.0': {'id': 'AFL-2.0', 'deprecated': False},
+    'afl-2.1': {'id': 'AFL-2.1', 'deprecated': False},
+    'afl-3.0': {'id': 'AFL-3.0', 'deprecated': False},
+    'afmparse': {'id': 'Afmparse', 'deprecated': False},
+    'agpl-1.0': {'id': 'AGPL-1.0', 'deprecated': True},
+    'agpl-1.0-only': {'id': 'AGPL-1.0-only', 'deprecated': False},
+    'agpl-1.0-or-later': {'id': 'AGPL-1.0-or-later', 'deprecated': False},
+    'agpl-3.0': {'id': 'AGPL-3.0', 'deprecated': True},
+    'agpl-3.0-only': {'id': 'AGPL-3.0-only', 'deprecated': False},
+    'agpl-3.0-or-later': {'id': 'AGPL-3.0-or-later', 'deprecated': False},
+    'aladdin': {'id': 'Aladdin', 'deprecated': False},
+    'amd-newlib': {'id': 'AMD-newlib', 'deprecated': False},
+    'amdplpa': {'id': 'AMDPLPA', 'deprecated': False},
+    'aml': {'id': 'AML', 'deprecated': False},
+    'aml-glslang': {'id': 'AML-glslang', 'deprecated': False},
+    'ampas': {'id': 'AMPAS', 'deprecated': False},
+    'antlr-pd': {'id': 'ANTLR-PD', 'deprecated': False},
+    'antlr-pd-fallback': {'id': 'ANTLR-PD-fallback', 'deprecated': False},
+    'any-osi': {'id': 'any-OSI', 'deprecated': False},
+    'apache-1.0': {'id': 'Apache-1.0', 'deprecated': False},
+    'apache-1.1': {'id': 'Apache-1.1', 'deprecated': False},
+    'apache-2.0': {'id': 'Apache-2.0', 'deprecated': False},
+    'apafml': {'id': 'APAFML', 'deprecated': False},
+    'apl-1.0': {'id': 'APL-1.0', 'deprecated': False},
+    'app-s2p': {'id': 'App-s2p', 'deprecated': False},
+    'apsl-1.0': {'id': 'APSL-1.0', 'deprecated': False},
+    'apsl-1.1': {'id': 'APSL-1.1', 'deprecated': False},
+    'apsl-1.2': {'id': 'APSL-1.2', 'deprecated': False},
+    'apsl-2.0': {'id': 'APSL-2.0', 'deprecated': False},
+    'arphic-1999': {'id': 'Arphic-1999', 'deprecated': False},
+    'artistic-1.0': {'id': 'Artistic-1.0', 'deprecated': False},
+    'artistic-1.0-cl8': {'id': 'Artistic-1.0-cl8', 'deprecated': False},
+    'artistic-1.0-perl': {'id': 'Artistic-1.0-Perl', 'deprecated': False},
+    'artistic-2.0': {'id': 'Artistic-2.0', 'deprecated': False},
+    'aswf-digital-assets-1.0': {'id': 'ASWF-Digital-Assets-1.0', 'deprecated': False},
+    'aswf-digital-assets-1.1': {'id': 'ASWF-Digital-Assets-1.1', 'deprecated': False},
+    'baekmuk': {'id': 'Baekmuk', 'deprecated': False},
+    'bahyph': {'id': 'Bahyph', 'deprecated': False},
+    'barr': {'id': 'Barr', 'deprecated': False},
+    'bcrypt-solar-designer': {'id': 'bcrypt-Solar-Designer', 'deprecated': False},
+    'beerware': {'id': 'Beerware', 'deprecated': False},
+    'bitstream-charter': {'id': 'Bitstream-Charter', 'deprecated': False},
+    'bitstream-vera': {'id': 'Bitstream-Vera', 'deprecated': False},
+    'bittorrent-1.0': {'id': 'BitTorrent-1.0', 'deprecated': False},
+    'bittorrent-1.1': {'id': 'BitTorrent-1.1', 'deprecated': False},
+    'blessing': {'id': 'blessing', 'deprecated': False},
+    'blueoak-1.0.0': {'id': 'BlueOak-1.0.0', 'deprecated': False},
+    'boehm-gc': {'id': 'Boehm-GC', 'deprecated': False},
+    'borceux': {'id': 'Borceux', 'deprecated': False},
+    'brian-gladman-2-clause': {'id': 'Brian-Gladman-2-Clause', 'deprecated': False},
+    'brian-gladman-3-clause': {'id': 'Brian-Gladman-3-Clause', 'deprecated': False},
+    'bsd-1-clause': {'id': 'BSD-1-Clause', 'deprecated': False},
+    'bsd-2-clause': {'id': 'BSD-2-Clause', 'deprecated': False},
+    'bsd-2-clause-darwin': {'id': 'BSD-2-Clause-Darwin', 'deprecated': False},
+    'bsd-2-clause-first-lines': {'id': 'BSD-2-Clause-first-lines', 'deprecated': False},
+    'bsd-2-clause-freebsd': {'id': 'BSD-2-Clause-FreeBSD', 'deprecated': True},
+    'bsd-2-clause-netbsd': {'id': 'BSD-2-Clause-NetBSD', 'deprecated': True},
+    'bsd-2-clause-patent': {'id': 'BSD-2-Clause-Patent', 'deprecated': False},
+    'bsd-2-clause-views': {'id': 'BSD-2-Clause-Views', 'deprecated': False},
+    'bsd-3-clause': {'id': 'BSD-3-Clause', 'deprecated': False},
+    'bsd-3-clause-acpica': {'id': 'BSD-3-Clause-acpica', 'deprecated': False},
+    'bsd-3-clause-attribution': {'id': 'BSD-3-Clause-Attribution', 'deprecated': False},
+    'bsd-3-clause-clear': {'id': 'BSD-3-Clause-Clear', 'deprecated': False},
+    'bsd-3-clause-flex': {'id': 'BSD-3-Clause-flex', 'deprecated': False},
+    'bsd-3-clause-hp': {'id': 'BSD-3-Clause-HP', 'deprecated': False},
+    'bsd-3-clause-lbnl': {'id': 'BSD-3-Clause-LBNL', 'deprecated': False},
+    'bsd-3-clause-modification': {'id': 'BSD-3-Clause-Modification', 'deprecated': False},
+    'bsd-3-clause-no-military-license': {'id': 'BSD-3-Clause-No-Military-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license': {'id': 'BSD-3-Clause-No-Nuclear-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license-2014': {'id': 'BSD-3-Clause-No-Nuclear-License-2014', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-warranty': {'id': 'BSD-3-Clause-No-Nuclear-Warranty', 'deprecated': False},
+    'bsd-3-clause-open-mpi': {'id': 'BSD-3-Clause-Open-MPI', 'deprecated': False},
+    'bsd-3-clause-sun': {'id': 'BSD-3-Clause-Sun', 'deprecated': False},
+    'bsd-4-clause': {'id': 'BSD-4-Clause', 'deprecated': False},
+    'bsd-4-clause-shortened': {'id': 'BSD-4-Clause-Shortened', 'deprecated': False},
+    'bsd-4-clause-uc': {'id': 'BSD-4-Clause-UC', 'deprecated': False},
+    'bsd-4.3reno': {'id': 'BSD-4.3RENO', 'deprecated': False},
+    'bsd-4.3tahoe': {'id': 'BSD-4.3TAHOE', 'deprecated': False},
+    'bsd-advertising-acknowledgement': {'id': 'BSD-Advertising-Acknowledgement', 'deprecated': False},
+    'bsd-attribution-hpnd-disclaimer': {'id': 'BSD-Attribution-HPND-disclaimer', 'deprecated': False},
+    'bsd-inferno-nettverk': {'id': 'BSD-Inferno-Nettverk', 'deprecated': False},
+    'bsd-protection': {'id': 'BSD-Protection', 'deprecated': False},
+    'bsd-source-beginning-file': {'id': 'BSD-Source-beginning-file', 'deprecated': False},
+    'bsd-source-code': {'id': 'BSD-Source-Code', 'deprecated': False},
+    'bsd-systemics': {'id': 'BSD-Systemics', 'deprecated': False},
+    'bsd-systemics-w3works': {'id': 'BSD-Systemics-W3Works', 'deprecated': False},
+    'bsl-1.0': {'id': 'BSL-1.0', 'deprecated': False},
+    'busl-1.1': {'id': 'BUSL-1.1', 'deprecated': False},
+    'bzip2-1.0.5': {'id': 'bzip2-1.0.5', 'deprecated': True},
+    'bzip2-1.0.6': {'id': 'bzip2-1.0.6', 'deprecated': False},
+    'c-uda-1.0': {'id': 'C-UDA-1.0', 'deprecated': False},
+    'cal-1.0': {'id': 'CAL-1.0', 'deprecated': False},
+    'cal-1.0-combined-work-exception': {'id': 'CAL-1.0-Combined-Work-Exception', 'deprecated': False},
+    'caldera': {'id': 'Caldera', 'deprecated': False},
+    'caldera-no-preamble': {'id': 'Caldera-no-preamble', 'deprecated': False},
+    'catharon': {'id': 'Catharon', 'deprecated': False},
+    'catosl-1.1': {'id': 'CATOSL-1.1', 'deprecated': False},
+    'cc-by-1.0': {'id': 'CC-BY-1.0', 'deprecated': False},
+    'cc-by-2.0': {'id': 'CC-BY-2.0', 'deprecated': False},
+    'cc-by-2.5': {'id': 'CC-BY-2.5', 'deprecated': False},
+    'cc-by-2.5-au': {'id': 'CC-BY-2.5-AU', 'deprecated': False},
+    'cc-by-3.0': {'id': 'CC-BY-3.0', 'deprecated': False},
+    'cc-by-3.0-at': {'id': 'CC-BY-3.0-AT', 'deprecated': False},
+    'cc-by-3.0-au': {'id': 'CC-BY-3.0-AU', 'deprecated': False},
+    'cc-by-3.0-de': {'id': 'CC-BY-3.0-DE', 'deprecated': False},
+    'cc-by-3.0-igo': {'id': 'CC-BY-3.0-IGO', 'deprecated': False},
+    'cc-by-3.0-nl': {'id': 'CC-BY-3.0-NL', 'deprecated': False},
+    'cc-by-3.0-us': {'id': 'CC-BY-3.0-US', 'deprecated': False},
+    'cc-by-4.0': {'id': 'CC-BY-4.0', 'deprecated': False},
+    'cc-by-nc-1.0': {'id': 'CC-BY-NC-1.0', 'deprecated': False},
+    'cc-by-nc-2.0': {'id': 'CC-BY-NC-2.0', 'deprecated': False},
+    'cc-by-nc-2.5': {'id': 'CC-BY-NC-2.5', 'deprecated': False},
+    'cc-by-nc-3.0': {'id': 'CC-BY-NC-3.0', 'deprecated': False},
+    'cc-by-nc-3.0-de': {'id': 'CC-BY-NC-3.0-DE', 'deprecated': False},
+    'cc-by-nc-4.0': {'id': 'CC-BY-NC-4.0', 'deprecated': False},
+    'cc-by-nc-nd-1.0': {'id': 'CC-BY-NC-ND-1.0', 'deprecated': False},
+    'cc-by-nc-nd-2.0': {'id': 'CC-BY-NC-ND-2.0', 'deprecated': False},
+    'cc-by-nc-nd-2.5': {'id': 'CC-BY-NC-ND-2.5', 'deprecated': False},
+    'cc-by-nc-nd-3.0': {'id': 'CC-BY-NC-ND-3.0', 'deprecated': False},
+    'cc-by-nc-nd-3.0-de': {'id': 'CC-BY-NC-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nc-nd-3.0-igo': {'id': 'CC-BY-NC-ND-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-nd-4.0': {'id': 'CC-BY-NC-ND-4.0', 'deprecated': False},
+    'cc-by-nc-sa-1.0': {'id': 'CC-BY-NC-SA-1.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0': {'id': 'CC-BY-NC-SA-2.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0-de': {'id': 'CC-BY-NC-SA-2.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-2.0-fr': {'id': 'CC-BY-NC-SA-2.0-FR', 'deprecated': False},
+    'cc-by-nc-sa-2.0-uk': {'id': 'CC-BY-NC-SA-2.0-UK', 'deprecated': False},
+    'cc-by-nc-sa-2.5': {'id': 'CC-BY-NC-SA-2.5', 'deprecated': False},
+    'cc-by-nc-sa-3.0': {'id': 'CC-BY-NC-SA-3.0', 'deprecated': False},
+    'cc-by-nc-sa-3.0-de': {'id': 'CC-BY-NC-SA-3.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-3.0-igo': {'id': 'CC-BY-NC-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-sa-4.0': {'id': 'CC-BY-NC-SA-4.0', 'deprecated': False},
+    'cc-by-nd-1.0': {'id': 'CC-BY-ND-1.0', 'deprecated': False},
+    'cc-by-nd-2.0': {'id': 'CC-BY-ND-2.0', 'deprecated': False},
+    'cc-by-nd-2.5': {'id': 'CC-BY-ND-2.5', 'deprecated': False},
+    'cc-by-nd-3.0': {'id': 'CC-BY-ND-3.0', 'deprecated': False},
+    'cc-by-nd-3.0-de': {'id': 'CC-BY-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nd-4.0': {'id': 'CC-BY-ND-4.0', 'deprecated': False},
+    'cc-by-sa-1.0': {'id': 'CC-BY-SA-1.0', 'deprecated': False},
+    'cc-by-sa-2.0': {'id': 'CC-BY-SA-2.0', 'deprecated': False},
+    'cc-by-sa-2.0-uk': {'id': 'CC-BY-SA-2.0-UK', 'deprecated': False},
+    'cc-by-sa-2.1-jp': {'id': 'CC-BY-SA-2.1-JP', 'deprecated': False},
+    'cc-by-sa-2.5': {'id': 'CC-BY-SA-2.5', 'deprecated': False},
+    'cc-by-sa-3.0': {'id': 'CC-BY-SA-3.0', 'deprecated': False},
+    'cc-by-sa-3.0-at': {'id': 'CC-BY-SA-3.0-AT', 'deprecated': False},
+    'cc-by-sa-3.0-de': {'id': 'CC-BY-SA-3.0-DE', 'deprecated': False},
+    'cc-by-sa-3.0-igo': {'id': 'CC-BY-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-sa-4.0': {'id': 'CC-BY-SA-4.0', 'deprecated': False},
+    'cc-pddc': {'id': 'CC-PDDC', 'deprecated': False},
+    'cc0-1.0': {'id': 'CC0-1.0', 'deprecated': False},
+    'cddl-1.0': {'id': 'CDDL-1.0', 'deprecated': False},
+    'cddl-1.1': {'id': 'CDDL-1.1', 'deprecated': False},
+    'cdl-1.0': {'id': 'CDL-1.0', 'deprecated': False},
+    'cdla-permissive-1.0': {'id': 'CDLA-Permissive-1.0', 'deprecated': False},
+    'cdla-permissive-2.0': {'id': 'CDLA-Permissive-2.0', 'deprecated': False},
+    'cdla-sharing-1.0': {'id': 'CDLA-Sharing-1.0', 'deprecated': False},
+    'cecill-1.0': {'id': 'CECILL-1.0', 'deprecated': False},
+    'cecill-1.1': {'id': 'CECILL-1.1', 'deprecated': False},
+    'cecill-2.0': {'id': 'CECILL-2.0', 'deprecated': False},
+    'cecill-2.1': {'id': 'CECILL-2.1', 'deprecated': False},
+    'cecill-b': {'id': 'CECILL-B', 'deprecated': False},
+    'cecill-c': {'id': 'CECILL-C', 'deprecated': False},
+    'cern-ohl-1.1': {'id': 'CERN-OHL-1.1', 'deprecated': False},
+    'cern-ohl-1.2': {'id': 'CERN-OHL-1.2', 'deprecated': False},
+    'cern-ohl-p-2.0': {'id': 'CERN-OHL-P-2.0', 'deprecated': False},
+    'cern-ohl-s-2.0': {'id': 'CERN-OHL-S-2.0', 'deprecated': False},
+    'cern-ohl-w-2.0': {'id': 'CERN-OHL-W-2.0', 'deprecated': False},
+    'cfitsio': {'id': 'CFITSIO', 'deprecated': False},
+    'check-cvs': {'id': 'check-cvs', 'deprecated': False},
+    'checkmk': {'id': 'checkmk', 'deprecated': False},
+    'clartistic': {'id': 'ClArtistic', 'deprecated': False},
+    'clips': {'id': 'Clips', 'deprecated': False},
+    'cmu-mach': {'id': 'CMU-Mach', 'deprecated': False},
+    'cmu-mach-nodoc': {'id': 'CMU-Mach-nodoc', 'deprecated': False},
+    'cnri-jython': {'id': 'CNRI-Jython', 'deprecated': False},
+    'cnri-python': {'id': 'CNRI-Python', 'deprecated': False},
+    'cnri-python-gpl-compatible': {'id': 'CNRI-Python-GPL-Compatible', 'deprecated': False},
+    'coil-1.0': {'id': 'COIL-1.0', 'deprecated': False},
+    'community-spec-1.0': {'id': 'Community-Spec-1.0', 'deprecated': False},
+    'condor-1.1': {'id': 'Condor-1.1', 'deprecated': False},
+    'copyleft-next-0.3.0': {'id': 'copyleft-next-0.3.0', 'deprecated': False},
+    'copyleft-next-0.3.1': {'id': 'copyleft-next-0.3.1', 'deprecated': False},
+    'cornell-lossless-jpeg': {'id': 'Cornell-Lossless-JPEG', 'deprecated': False},
+    'cpal-1.0': {'id': 'CPAL-1.0', 'deprecated': False},
+    'cpl-1.0': {'id': 'CPL-1.0', 'deprecated': False},
+    'cpol-1.02': {'id': 'CPOL-1.02', 'deprecated': False},
+    'cronyx': {'id': 'Cronyx', 'deprecated': False},
+    'crossword': {'id': 'Crossword', 'deprecated': False},
+    'crystalstacker': {'id': 'CrystalStacker', 'deprecated': False},
+    'cua-opl-1.0': {'id': 'CUA-OPL-1.0', 'deprecated': False},
+    'cube': {'id': 'Cube', 'deprecated': False},
+    'curl': {'id': 'curl', 'deprecated': False},
+    'cve-tou': {'id': 'cve-tou', 'deprecated': False},
+    'd-fsl-1.0': {'id': 'D-FSL-1.0', 'deprecated': False},
+    'dec-3-clause': {'id': 'DEC-3-Clause', 'deprecated': False},
+    'diffmark': {'id': 'diffmark', 'deprecated': False},
+    'dl-de-by-2.0': {'id': 'DL-DE-BY-2.0', 'deprecated': False},
+    'dl-de-zero-2.0': {'id': 'DL-DE-ZERO-2.0', 'deprecated': False},
+    'doc': {'id': 'DOC', 'deprecated': False},
+    'docbook-schema': {'id': 'DocBook-Schema', 'deprecated': False},
+    'docbook-xml': {'id': 'DocBook-XML', 'deprecated': False},
+    'dotseqn': {'id': 'Dotseqn', 'deprecated': False},
+    'drl-1.0': {'id': 'DRL-1.0', 'deprecated': False},
+    'drl-1.1': {'id': 'DRL-1.1', 'deprecated': False},
+    'dsdp': {'id': 'DSDP', 'deprecated': False},
+    'dtoa': {'id': 'dtoa', 'deprecated': False},
+    'dvipdfm': {'id': 'dvipdfm', 'deprecated': False},
+    'ecl-1.0': {'id': 'ECL-1.0', 'deprecated': False},
+    'ecl-2.0': {'id': 'ECL-2.0', 'deprecated': False},
+    'ecos-2.0': {'id': 'eCos-2.0', 'deprecated': True},
+    'efl-1.0': {'id': 'EFL-1.0', 'deprecated': False},
+    'efl-2.0': {'id': 'EFL-2.0', 'deprecated': False},
+    'egenix': {'id': 'eGenix', 'deprecated': False},
+    'elastic-2.0': {'id': 'Elastic-2.0', 'deprecated': False},
+    'entessa': {'id': 'Entessa', 'deprecated': False},
+    'epics': {'id': 'EPICS', 'deprecated': False},
+    'epl-1.0': {'id': 'EPL-1.0', 'deprecated': False},
+    'epl-2.0': {'id': 'EPL-2.0', 'deprecated': False},
+    'erlpl-1.1': {'id': 'ErlPL-1.1', 'deprecated': False},
+    'etalab-2.0': {'id': 'etalab-2.0', 'deprecated': False},
+    'eudatagrid': {'id': 'EUDatagrid', 'deprecated': False},
+    'eupl-1.0': {'id': 'EUPL-1.0', 'deprecated': False},
+    'eupl-1.1': {'id': 'EUPL-1.1', 'deprecated': False},
+    'eupl-1.2': {'id': 'EUPL-1.2', 'deprecated': False},
+    'eurosym': {'id': 'Eurosym', 'deprecated': False},
+    'fair': {'id': 'Fair', 'deprecated': False},
+    'fbm': {'id': 'FBM', 'deprecated': False},
+    'fdk-aac': {'id': 'FDK-AAC', 'deprecated': False},
+    'ferguson-twofish': {'id': 'Ferguson-Twofish', 'deprecated': False},
+    'frameworx-1.0': {'id': 'Frameworx-1.0', 'deprecated': False},
+    'freebsd-doc': {'id': 'FreeBSD-DOC', 'deprecated': False},
+    'freeimage': {'id': 'FreeImage', 'deprecated': False},
+    'fsfap': {'id': 'FSFAP', 'deprecated': False},
+    'fsfap-no-warranty-disclaimer': {'id': 'FSFAP-no-warranty-disclaimer', 'deprecated': False},
+    'fsful': {'id': 'FSFUL', 'deprecated': False},
+    'fsfullr': {'id': 'FSFULLR', 'deprecated': False},
+    'fsfullrwd': {'id': 'FSFULLRWD', 'deprecated': False},
+    'ftl': {'id': 'FTL', 'deprecated': False},
+    'furuseth': {'id': 'Furuseth', 'deprecated': False},
+    'fwlw': {'id': 'fwlw', 'deprecated': False},
+    'gcr-docs': {'id': 'GCR-docs', 'deprecated': False},
+    'gd': {'id': 'GD', 'deprecated': False},
+    'gfdl-1.1': {'id': 'GFDL-1.1', 'deprecated': True},
+    'gfdl-1.1-invariants-only': {'id': 'GFDL-1.1-invariants-only', 'deprecated': False},
+    'gfdl-1.1-invariants-or-later': {'id': 'GFDL-1.1-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-no-invariants-only': {'id': 'GFDL-1.1-no-invariants-only', 'deprecated': False},
+    'gfdl-1.1-no-invariants-or-later': {'id': 'GFDL-1.1-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-only': {'id': 'GFDL-1.1-only', 'deprecated': False},
+    'gfdl-1.1-or-later': {'id': 'GFDL-1.1-or-later', 'deprecated': False},
+    'gfdl-1.2': {'id': 'GFDL-1.2', 'deprecated': True},
+    'gfdl-1.2-invariants-only': {'id': 'GFDL-1.2-invariants-only', 'deprecated': False},
+    'gfdl-1.2-invariants-or-later': {'id': 'GFDL-1.2-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-no-invariants-only': {'id': 'GFDL-1.2-no-invariants-only', 'deprecated': False},
+    'gfdl-1.2-no-invariants-or-later': {'id': 'GFDL-1.2-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-only': {'id': 'GFDL-1.2-only', 'deprecated': False},
+    'gfdl-1.2-or-later': {'id': 'GFDL-1.2-or-later', 'deprecated': False},
+    'gfdl-1.3': {'id': 'GFDL-1.3', 'deprecated': True},
+    'gfdl-1.3-invariants-only': {'id': 'GFDL-1.3-invariants-only', 'deprecated': False},
+    'gfdl-1.3-invariants-or-later': {'id': 'GFDL-1.3-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-no-invariants-only': {'id': 'GFDL-1.3-no-invariants-only', 'deprecated': False},
+    'gfdl-1.3-no-invariants-or-later': {'id': 'GFDL-1.3-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-only': {'id': 'GFDL-1.3-only', 'deprecated': False},
+    'gfdl-1.3-or-later': {'id': 'GFDL-1.3-or-later', 'deprecated': False},
+    'giftware': {'id': 'Giftware', 'deprecated': False},
+    'gl2ps': {'id': 'GL2PS', 'deprecated': False},
+    'glide': {'id': 'Glide', 'deprecated': False},
+    'glulxe': {'id': 'Glulxe', 'deprecated': False},
+    'glwtpl': {'id': 'GLWTPL', 'deprecated': False},
+    'gnuplot': {'id': 'gnuplot', 'deprecated': False},
+    'gpl-1.0': {'id': 'GPL-1.0', 'deprecated': True},
+    'gpl-1.0+': {'id': 'GPL-1.0+', 'deprecated': True},
+    'gpl-1.0-only': {'id': 'GPL-1.0-only', 'deprecated': False},
+    'gpl-1.0-or-later': {'id': 'GPL-1.0-or-later', 'deprecated': False},
+    'gpl-2.0': {'id': 'GPL-2.0', 'deprecated': True},
+    'gpl-2.0+': {'id': 'GPL-2.0+', 'deprecated': True},
+    'gpl-2.0-only': {'id': 'GPL-2.0-only', 'deprecated': False},
+    'gpl-2.0-or-later': {'id': 'GPL-2.0-or-later', 'deprecated': False},
+    'gpl-2.0-with-autoconf-exception': {'id': 'GPL-2.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-2.0-with-bison-exception': {'id': 'GPL-2.0-with-bison-exception', 'deprecated': True},
+    'gpl-2.0-with-classpath-exception': {'id': 'GPL-2.0-with-classpath-exception', 'deprecated': True},
+    'gpl-2.0-with-font-exception': {'id': 'GPL-2.0-with-font-exception', 'deprecated': True},
+    'gpl-2.0-with-gcc-exception': {'id': 'GPL-2.0-with-GCC-exception', 'deprecated': True},
+    'gpl-3.0': {'id': 'GPL-3.0', 'deprecated': True},
+    'gpl-3.0+': {'id': 'GPL-3.0+', 'deprecated': True},
+    'gpl-3.0-only': {'id': 'GPL-3.0-only', 'deprecated': False},
+    'gpl-3.0-or-later': {'id': 'GPL-3.0-or-later', 'deprecated': False},
+    'gpl-3.0-with-autoconf-exception': {'id': 'GPL-3.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-3.0-with-gcc-exception': {'id': 'GPL-3.0-with-GCC-exception', 'deprecated': True},
+    'graphics-gems': {'id': 'Graphics-Gems', 'deprecated': False},
+    'gsoap-1.3b': {'id': 'gSOAP-1.3b', 'deprecated': False},
+    'gtkbook': {'id': 'gtkbook', 'deprecated': False},
+    'gutmann': {'id': 'Gutmann', 'deprecated': False},
+    'haskellreport': {'id': 'HaskellReport', 'deprecated': False},
+    'hdparm': {'id': 'hdparm', 'deprecated': False},
+    'hidapi': {'id': 'HIDAPI', 'deprecated': False},
+    'hippocratic-2.1': {'id': 'Hippocratic-2.1', 'deprecated': False},
+    'hp-1986': {'id': 'HP-1986', 'deprecated': False},
+    'hp-1989': {'id': 'HP-1989', 'deprecated': False},
+    'hpnd': {'id': 'HPND', 'deprecated': False},
+    'hpnd-dec': {'id': 'HPND-DEC', 'deprecated': False},
+    'hpnd-doc': {'id': 'HPND-doc', 'deprecated': False},
+    'hpnd-doc-sell': {'id': 'HPND-doc-sell', 'deprecated': False},
+    'hpnd-export-us': {'id': 'HPND-export-US', 'deprecated': False},
+    'hpnd-export-us-acknowledgement': {'id': 'HPND-export-US-acknowledgement', 'deprecated': False},
+    'hpnd-export-us-modify': {'id': 'HPND-export-US-modify', 'deprecated': False},
+    'hpnd-export2-us': {'id': 'HPND-export2-US', 'deprecated': False},
+    'hpnd-fenneberg-livingston': {'id': 'HPND-Fenneberg-Livingston', 'deprecated': False},
+    'hpnd-inria-imag': {'id': 'HPND-INRIA-IMAG', 'deprecated': False},
+    'hpnd-intel': {'id': 'HPND-Intel', 'deprecated': False},
+    'hpnd-kevlin-henney': {'id': 'HPND-Kevlin-Henney', 'deprecated': False},
+    'hpnd-markus-kuhn': {'id': 'HPND-Markus-Kuhn', 'deprecated': False},
+    'hpnd-merchantability-variant': {'id': 'HPND-merchantability-variant', 'deprecated': False},
+    'hpnd-mit-disclaimer': {'id': 'HPND-MIT-disclaimer', 'deprecated': False},
+    'hpnd-netrek': {'id': 'HPND-Netrek', 'deprecated': False},
+    'hpnd-pbmplus': {'id': 'HPND-Pbmplus', 'deprecated': False},
+    'hpnd-sell-mit-disclaimer-xserver': {'id': 'HPND-sell-MIT-disclaimer-xserver', 'deprecated': False},
+    'hpnd-sell-regexpr': {'id': 'HPND-sell-regexpr', 'deprecated': False},
+    'hpnd-sell-variant': {'id': 'HPND-sell-variant', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer': {'id': 'HPND-sell-variant-MIT-disclaimer', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer-rev': {'id': 'HPND-sell-variant-MIT-disclaimer-rev', 'deprecated': False},
+    'hpnd-uc': {'id': 'HPND-UC', 'deprecated': False},
+    'hpnd-uc-export-us': {'id': 'HPND-UC-export-US', 'deprecated': False},
+    'htmltidy': {'id': 'HTMLTIDY', 'deprecated': False},
+    'ibm-pibs': {'id': 'IBM-pibs', 'deprecated': False},
+    'icu': {'id': 'ICU', 'deprecated': False},
+    'iec-code-components-eula': {'id': 'IEC-Code-Components-EULA', 'deprecated': False},
+    'ijg': {'id': 'IJG', 'deprecated': False},
+    'ijg-short': {'id': 'IJG-short', 'deprecated': False},
+    'imagemagick': {'id': 'ImageMagick', 'deprecated': False},
+    'imatix': {'id': 'iMatix', 'deprecated': False},
+    'imlib2': {'id': 'Imlib2', 'deprecated': False},
+    'info-zip': {'id': 'Info-ZIP', 'deprecated': False},
+    'inner-net-2.0': {'id': 'Inner-Net-2.0', 'deprecated': False},
+    'intel': {'id': 'Intel', 'deprecated': False},
+    'intel-acpi': {'id': 'Intel-ACPI', 'deprecated': False},
+    'interbase-1.0': {'id': 'Interbase-1.0', 'deprecated': False},
+    'ipa': {'id': 'IPA', 'deprecated': False},
+    'ipl-1.0': {'id': 'IPL-1.0', 'deprecated': False},
+    'isc': {'id': 'ISC', 'deprecated': False},
+    'isc-veillard': {'id': 'ISC-Veillard', 'deprecated': False},
+    'jam': {'id': 'Jam', 'deprecated': False},
+    'jasper-2.0': {'id': 'JasPer-2.0', 'deprecated': False},
+    'jpl-image': {'id': 'JPL-image', 'deprecated': False},
+    'jpnic': {'id': 'JPNIC', 'deprecated': False},
+    'json': {'id': 'JSON', 'deprecated': False},
+    'kastrup': {'id': 'Kastrup', 'deprecated': False},
+    'kazlib': {'id': 'Kazlib', 'deprecated': False},
+    'knuth-ctan': {'id': 'Knuth-CTAN', 'deprecated': False},
+    'lal-1.2': {'id': 'LAL-1.2', 'deprecated': False},
+    'lal-1.3': {'id': 'LAL-1.3', 'deprecated': False},
+    'latex2e': {'id': 'Latex2e', 'deprecated': False},
+    'latex2e-translated-notice': {'id': 'Latex2e-translated-notice', 'deprecated': False},
+    'leptonica': {'id': 'Leptonica', 'deprecated': False},
+    'lgpl-2.0': {'id': 'LGPL-2.0', 'deprecated': True},
+    'lgpl-2.0+': {'id': 'LGPL-2.0+', 'deprecated': True},
+    'lgpl-2.0-only': {'id': 'LGPL-2.0-only', 'deprecated': False},
+    'lgpl-2.0-or-later': {'id': 'LGPL-2.0-or-later', 'deprecated': False},
+    'lgpl-2.1': {'id': 'LGPL-2.1', 'deprecated': True},
+    'lgpl-2.1+': {'id': 'LGPL-2.1+', 'deprecated': True},
+    'lgpl-2.1-only': {'id': 'LGPL-2.1-only', 'deprecated': False},
+    'lgpl-2.1-or-later': {'id': 'LGPL-2.1-or-later', 'deprecated': False},
+    'lgpl-3.0': {'id': 'LGPL-3.0', 'deprecated': True},
+    'lgpl-3.0+': {'id': 'LGPL-3.0+', 'deprecated': True},
+    'lgpl-3.0-only': {'id': 'LGPL-3.0-only', 'deprecated': False},
+    'lgpl-3.0-or-later': {'id': 'LGPL-3.0-or-later', 'deprecated': False},
+    'lgpllr': {'id': 'LGPLLR', 'deprecated': False},
+    'libpng': {'id': 'Libpng', 'deprecated': False},
+    'libpng-2.0': {'id': 'libpng-2.0', 'deprecated': False},
+    'libselinux-1.0': {'id': 'libselinux-1.0', 'deprecated': False},
+    'libtiff': {'id': 'libtiff', 'deprecated': False},
+    'libutil-david-nugent': {'id': 'libutil-David-Nugent', 'deprecated': False},
+    'liliq-p-1.1': {'id': 'LiLiQ-P-1.1', 'deprecated': False},
+    'liliq-r-1.1': {'id': 'LiLiQ-R-1.1', 'deprecated': False},
+    'liliq-rplus-1.1': {'id': 'LiLiQ-Rplus-1.1', 'deprecated': False},
+    'linux-man-pages-1-para': {'id': 'Linux-man-pages-1-para', 'deprecated': False},
+    'linux-man-pages-copyleft': {'id': 'Linux-man-pages-copyleft', 'deprecated': False},
+    'linux-man-pages-copyleft-2-para': {'id': 'Linux-man-pages-copyleft-2-para', 'deprecated': False},
+    'linux-man-pages-copyleft-var': {'id': 'Linux-man-pages-copyleft-var', 'deprecated': False},
+    'linux-openib': {'id': 'Linux-OpenIB', 'deprecated': False},
+    'loop': {'id': 'LOOP', 'deprecated': False},
+    'lpd-document': {'id': 'LPD-document', 'deprecated': False},
+    'lpl-1.0': {'id': 'LPL-1.0', 'deprecated': False},
+    'lpl-1.02': {'id': 'LPL-1.02', 'deprecated': False},
+    'lppl-1.0': {'id': 'LPPL-1.0', 'deprecated': False},
+    'lppl-1.1': {'id': 'LPPL-1.1', 'deprecated': False},
+    'lppl-1.2': {'id': 'LPPL-1.2', 'deprecated': False},
+    'lppl-1.3a': {'id': 'LPPL-1.3a', 'deprecated': False},
+    'lppl-1.3c': {'id': 'LPPL-1.3c', 'deprecated': False},
+    'lsof': {'id': 'lsof', 'deprecated': False},
+    'lucida-bitmap-fonts': {'id': 'Lucida-Bitmap-Fonts', 'deprecated': False},
+    'lzma-sdk-9.11-to-9.20': {'id': 'LZMA-SDK-9.11-to-9.20', 'deprecated': False},
+    'lzma-sdk-9.22': {'id': 'LZMA-SDK-9.22', 'deprecated': False},
+    'mackerras-3-clause': {'id': 'Mackerras-3-Clause', 'deprecated': False},
+    'mackerras-3-clause-acknowledgment': {'id': 'Mackerras-3-Clause-acknowledgment', 'deprecated': False},
+    'magaz': {'id': 'magaz', 'deprecated': False},
+    'mailprio': {'id': 'mailprio', 'deprecated': False},
+    'makeindex': {'id': 'MakeIndex', 'deprecated': False},
+    'martin-birgmeier': {'id': 'Martin-Birgmeier', 'deprecated': False},
+    'mcphee-slideshow': {'id': 'McPhee-slideshow', 'deprecated': False},
+    'metamail': {'id': 'metamail', 'deprecated': False},
+    'minpack': {'id': 'Minpack', 'deprecated': False},
+    'miros': {'id': 'MirOS', 'deprecated': False},
+    'mit': {'id': 'MIT', 'deprecated': False},
+    'mit-0': {'id': 'MIT-0', 'deprecated': False},
+    'mit-advertising': {'id': 'MIT-advertising', 'deprecated': False},
+    'mit-cmu': {'id': 'MIT-CMU', 'deprecated': False},
+    'mit-enna': {'id': 'MIT-enna', 'deprecated': False},
+    'mit-feh': {'id': 'MIT-feh', 'deprecated': False},
+    'mit-festival': {'id': 'MIT-Festival', 'deprecated': False},
+    'mit-khronos-old': {'id': 'MIT-Khronos-old', 'deprecated': False},
+    'mit-modern-variant': {'id': 'MIT-Modern-Variant', 'deprecated': False},
+    'mit-open-group': {'id': 'MIT-open-group', 'deprecated': False},
+    'mit-testregex': {'id': 'MIT-testregex', 'deprecated': False},
+    'mit-wu': {'id': 'MIT-Wu', 'deprecated': False},
+    'mitnfa': {'id': 'MITNFA', 'deprecated': False},
+    'mmixware': {'id': 'MMIXware', 'deprecated': False},
+    'motosoto': {'id': 'Motosoto', 'deprecated': False},
+    'mpeg-ssg': {'id': 'MPEG-SSG', 'deprecated': False},
+    'mpi-permissive': {'id': 'mpi-permissive', 'deprecated': False},
+    'mpich2': {'id': 'mpich2', 'deprecated': False},
+    'mpl-1.0': {'id': 'MPL-1.0', 'deprecated': False},
+    'mpl-1.1': {'id': 'MPL-1.1', 'deprecated': False},
+    'mpl-2.0': {'id': 'MPL-2.0', 'deprecated': False},
+    'mpl-2.0-no-copyleft-exception': {'id': 'MPL-2.0-no-copyleft-exception', 'deprecated': False},
+    'mplus': {'id': 'mplus', 'deprecated': False},
+    'ms-lpl': {'id': 'MS-LPL', 'deprecated': False},
+    'ms-pl': {'id': 'MS-PL', 'deprecated': False},
+    'ms-rl': {'id': 'MS-RL', 'deprecated': False},
+    'mtll': {'id': 'MTLL', 'deprecated': False},
+    'mulanpsl-1.0': {'id': 'MulanPSL-1.0', 'deprecated': False},
+    'mulanpsl-2.0': {'id': 'MulanPSL-2.0', 'deprecated': False},
+    'multics': {'id': 'Multics', 'deprecated': False},
+    'mup': {'id': 'Mup', 'deprecated': False},
+    'naist-2003': {'id': 'NAIST-2003', 'deprecated': False},
+    'nasa-1.3': {'id': 'NASA-1.3', 'deprecated': False},
+    'naumen': {'id': 'Naumen', 'deprecated': False},
+    'nbpl-1.0': {'id': 'NBPL-1.0', 'deprecated': False},
+    'ncbi-pd': {'id': 'NCBI-PD', 'deprecated': False},
+    'ncgl-uk-2.0': {'id': 'NCGL-UK-2.0', 'deprecated': False},
+    'ncl': {'id': 'NCL', 'deprecated': False},
+    'ncsa': {'id': 'NCSA', 'deprecated': False},
+    'net-snmp': {'id': 'Net-SNMP', 'deprecated': True},
+    'netcdf': {'id': 'NetCDF', 'deprecated': False},
+    'newsletr': {'id': 'Newsletr', 'deprecated': False},
+    'ngpl': {'id': 'NGPL', 'deprecated': False},
+    'nicta-1.0': {'id': 'NICTA-1.0', 'deprecated': False},
+    'nist-pd': {'id': 'NIST-PD', 'deprecated': False},
+    'nist-pd-fallback': {'id': 'NIST-PD-fallback', 'deprecated': False},
+    'nist-software': {'id': 'NIST-Software', 'deprecated': False},
+    'nlod-1.0': {'id': 'NLOD-1.0', 'deprecated': False},
+    'nlod-2.0': {'id': 'NLOD-2.0', 'deprecated': False},
+    'nlpl': {'id': 'NLPL', 'deprecated': False},
+    'nokia': {'id': 'Nokia', 'deprecated': False},
+    'nosl': {'id': 'NOSL', 'deprecated': False},
+    'noweb': {'id': 'Noweb', 'deprecated': False},
+    'npl-1.0': {'id': 'NPL-1.0', 'deprecated': False},
+    'npl-1.1': {'id': 'NPL-1.1', 'deprecated': False},
+    'nposl-3.0': {'id': 'NPOSL-3.0', 'deprecated': False},
+    'nrl': {'id': 'NRL', 'deprecated': False},
+    'ntp': {'id': 'NTP', 'deprecated': False},
+    'ntp-0': {'id': 'NTP-0', 'deprecated': False},
+    'nunit': {'id': 'Nunit', 'deprecated': True},
+    'o-uda-1.0': {'id': 'O-UDA-1.0', 'deprecated': False},
+    'oar': {'id': 'OAR', 'deprecated': False},
+    'occt-pl': {'id': 'OCCT-PL', 'deprecated': False},
+    'oclc-2.0': {'id': 'OCLC-2.0', 'deprecated': False},
+    'odbl-1.0': {'id': 'ODbL-1.0', 'deprecated': False},
+    'odc-by-1.0': {'id': 'ODC-By-1.0', 'deprecated': False},
+    'offis': {'id': 'OFFIS', 'deprecated': False},
+    'ofl-1.0': {'id': 'OFL-1.0', 'deprecated': False},
+    'ofl-1.0-no-rfn': {'id': 'OFL-1.0-no-RFN', 'deprecated': False},
+    'ofl-1.0-rfn': {'id': 'OFL-1.0-RFN', 'deprecated': False},
+    'ofl-1.1': {'id': 'OFL-1.1', 'deprecated': False},
+    'ofl-1.1-no-rfn': {'id': 'OFL-1.1-no-RFN', 'deprecated': False},
+    'ofl-1.1-rfn': {'id': 'OFL-1.1-RFN', 'deprecated': False},
+    'ogc-1.0': {'id': 'OGC-1.0', 'deprecated': False},
+    'ogdl-taiwan-1.0': {'id': 'OGDL-Taiwan-1.0', 'deprecated': False},
+    'ogl-canada-2.0': {'id': 'OGL-Canada-2.0', 'deprecated': False},
+    'ogl-uk-1.0': {'id': 'OGL-UK-1.0', 'deprecated': False},
+    'ogl-uk-2.0': {'id': 'OGL-UK-2.0', 'deprecated': False},
+    'ogl-uk-3.0': {'id': 'OGL-UK-3.0', 'deprecated': False},
+    'ogtsl': {'id': 'OGTSL', 'deprecated': False},
+    'oldap-1.1': {'id': 'OLDAP-1.1', 'deprecated': False},
+    'oldap-1.2': {'id': 'OLDAP-1.2', 'deprecated': False},
+    'oldap-1.3': {'id': 'OLDAP-1.3', 'deprecated': False},
+    'oldap-1.4': {'id': 'OLDAP-1.4', 'deprecated': False},
+    'oldap-2.0': {'id': 'OLDAP-2.0', 'deprecated': False},
+    'oldap-2.0.1': {'id': 'OLDAP-2.0.1', 'deprecated': False},
+    'oldap-2.1': {'id': 'OLDAP-2.1', 'deprecated': False},
+    'oldap-2.2': {'id': 'OLDAP-2.2', 'deprecated': False},
+    'oldap-2.2.1': {'id': 'OLDAP-2.2.1', 'deprecated': False},
+    'oldap-2.2.2': {'id': 'OLDAP-2.2.2', 'deprecated': False},
+    'oldap-2.3': {'id': 'OLDAP-2.3', 'deprecated': False},
+    'oldap-2.4': {'id': 'OLDAP-2.4', 'deprecated': False},
+    'oldap-2.5': {'id': 'OLDAP-2.5', 'deprecated': False},
+    'oldap-2.6': {'id': 'OLDAP-2.6', 'deprecated': False},
+    'oldap-2.7': {'id': 'OLDAP-2.7', 'deprecated': False},
+    'oldap-2.8': {'id': 'OLDAP-2.8', 'deprecated': False},
+    'olfl-1.3': {'id': 'OLFL-1.3', 'deprecated': False},
+    'oml': {'id': 'OML', 'deprecated': False},
+    'openpbs-2.3': {'id': 'OpenPBS-2.3', 'deprecated': False},
+    'openssl': {'id': 'OpenSSL', 'deprecated': False},
+    'openssl-standalone': {'id': 'OpenSSL-standalone', 'deprecated': False},
+    'openvision': {'id': 'OpenVision', 'deprecated': False},
+    'opl-1.0': {'id': 'OPL-1.0', 'deprecated': False},
+    'opl-uk-3.0': {'id': 'OPL-UK-3.0', 'deprecated': False},
+    'opubl-1.0': {'id': 'OPUBL-1.0', 'deprecated': False},
+    'oset-pl-2.1': {'id': 'OSET-PL-2.1', 'deprecated': False},
+    'osl-1.0': {'id': 'OSL-1.0', 'deprecated': False},
+    'osl-1.1': {'id': 'OSL-1.1', 'deprecated': False},
+    'osl-2.0': {'id': 'OSL-2.0', 'deprecated': False},
+    'osl-2.1': {'id': 'OSL-2.1', 'deprecated': False},
+    'osl-3.0': {'id': 'OSL-3.0', 'deprecated': False},
+    'padl': {'id': 'PADL', 'deprecated': False},
+    'parity-6.0.0': {'id': 'Parity-6.0.0', 'deprecated': False},
+    'parity-7.0.0': {'id': 'Parity-7.0.0', 'deprecated': False},
+    'pddl-1.0': {'id': 'PDDL-1.0', 'deprecated': False},
+    'php-3.0': {'id': 'PHP-3.0', 'deprecated': False},
+    'php-3.01': {'id': 'PHP-3.01', 'deprecated': False},
+    'pixar': {'id': 'Pixar', 'deprecated': False},
+    'pkgconf': {'id': 'pkgconf', 'deprecated': False},
+    'plexus': {'id': 'Plexus', 'deprecated': False},
+    'pnmstitch': {'id': 'pnmstitch', 'deprecated': False},
+    'polyform-noncommercial-1.0.0': {'id': 'PolyForm-Noncommercial-1.0.0', 'deprecated': False},
+    'polyform-small-business-1.0.0': {'id': 'PolyForm-Small-Business-1.0.0', 'deprecated': False},
+    'postgresql': {'id': 'PostgreSQL', 'deprecated': False},
+    'ppl': {'id': 'PPL', 'deprecated': False},
+    'psf-2.0': {'id': 'PSF-2.0', 'deprecated': False},
+    'psfrag': {'id': 'psfrag', 'deprecated': False},
+    'psutils': {'id': 'psutils', 'deprecated': False},
+    'python-2.0': {'id': 'Python-2.0', 'deprecated': False},
+    'python-2.0.1': {'id': 'Python-2.0.1', 'deprecated': False},
+    'python-ldap': {'id': 'python-ldap', 'deprecated': False},
+    'qhull': {'id': 'Qhull', 'deprecated': False},
+    'qpl-1.0': {'id': 'QPL-1.0', 'deprecated': False},
+    'qpl-1.0-inria-2004': {'id': 'QPL-1.0-INRIA-2004', 'deprecated': False},
+    'radvd': {'id': 'radvd', 'deprecated': False},
+    'rdisc': {'id': 'Rdisc', 'deprecated': False},
+    'rhecos-1.1': {'id': 'RHeCos-1.1', 'deprecated': False},
+    'rpl-1.1': {'id': 'RPL-1.1', 'deprecated': False},
+    'rpl-1.5': {'id': 'RPL-1.5', 'deprecated': False},
+    'rpsl-1.0': {'id': 'RPSL-1.0', 'deprecated': False},
+    'rsa-md': {'id': 'RSA-MD', 'deprecated': False},
+    'rscpl': {'id': 'RSCPL', 'deprecated': False},
+    'ruby': {'id': 'Ruby', 'deprecated': False},
+    'ruby-pty': {'id': 'Ruby-pty', 'deprecated': False},
+    'sax-pd': {'id': 'SAX-PD', 'deprecated': False},
+    'sax-pd-2.0': {'id': 'SAX-PD-2.0', 'deprecated': False},
+    'saxpath': {'id': 'Saxpath', 'deprecated': False},
+    'scea': {'id': 'SCEA', 'deprecated': False},
+    'schemereport': {'id': 'SchemeReport', 'deprecated': False},
+    'sendmail': {'id': 'Sendmail', 'deprecated': False},
+    'sendmail-8.23': {'id': 'Sendmail-8.23', 'deprecated': False},
+    'sgi-b-1.0': {'id': 'SGI-B-1.0', 'deprecated': False},
+    'sgi-b-1.1': {'id': 'SGI-B-1.1', 'deprecated': False},
+    'sgi-b-2.0': {'id': 'SGI-B-2.0', 'deprecated': False},
+    'sgi-opengl': {'id': 'SGI-OpenGL', 'deprecated': False},
+    'sgp4': {'id': 'SGP4', 'deprecated': False},
+    'shl-0.5': {'id': 'SHL-0.5', 'deprecated': False},
+    'shl-0.51': {'id': 'SHL-0.51', 'deprecated': False},
+    'simpl-2.0': {'id': 'SimPL-2.0', 'deprecated': False},
+    'sissl': {'id': 'SISSL', 'deprecated': False},
+    'sissl-1.2': {'id': 'SISSL-1.2', 'deprecated': False},
+    'sl': {'id': 'SL', 'deprecated': False},
+    'sleepycat': {'id': 'Sleepycat', 'deprecated': False},
+    'smlnj': {'id': 'SMLNJ', 'deprecated': False},
+    'smppl': {'id': 'SMPPL', 'deprecated': False},
+    'snia': {'id': 'SNIA', 'deprecated': False},
+    'snprintf': {'id': 'snprintf', 'deprecated': False},
+    'softsurfer': {'id': 'softSurfer', 'deprecated': False},
+    'soundex': {'id': 'Soundex', 'deprecated': False},
+    'spencer-86': {'id': 'Spencer-86', 'deprecated': False},
+    'spencer-94': {'id': 'Spencer-94', 'deprecated': False},
+    'spencer-99': {'id': 'Spencer-99', 'deprecated': False},
+    'spl-1.0': {'id': 'SPL-1.0', 'deprecated': False},
+    'ssh-keyscan': {'id': 'ssh-keyscan', 'deprecated': False},
+    'ssh-openssh': {'id': 'SSH-OpenSSH', 'deprecated': False},
+    'ssh-short': {'id': 'SSH-short', 'deprecated': False},
+    'ssleay-standalone': {'id': 'SSLeay-standalone', 'deprecated': False},
+    'sspl-1.0': {'id': 'SSPL-1.0', 'deprecated': False},
+    'standardml-nj': {'id': 'StandardML-NJ', 'deprecated': True},
+    'sugarcrm-1.1.3': {'id': 'SugarCRM-1.1.3', 'deprecated': False},
+    'sun-ppp': {'id': 'Sun-PPP', 'deprecated': False},
+    'sun-ppp-2000': {'id': 'Sun-PPP-2000', 'deprecated': False},
+    'sunpro': {'id': 'SunPro', 'deprecated': False},
+    'swl': {'id': 'SWL', 'deprecated': False},
+    'swrule': {'id': 'swrule', 'deprecated': False},
+    'symlinks': {'id': 'Symlinks', 'deprecated': False},
+    'tapr-ohl-1.0': {'id': 'TAPR-OHL-1.0', 'deprecated': False},
+    'tcl': {'id': 'TCL', 'deprecated': False},
+    'tcp-wrappers': {'id': 'TCP-wrappers', 'deprecated': False},
+    'termreadkey': {'id': 'TermReadKey', 'deprecated': False},
+    'tgppl-1.0': {'id': 'TGPPL-1.0', 'deprecated': False},
+    'threeparttable': {'id': 'threeparttable', 'deprecated': False},
+    'tmate': {'id': 'TMate', 'deprecated': False},
+    'torque-1.1': {'id': 'TORQUE-1.1', 'deprecated': False},
+    'tosl': {'id': 'TOSL', 'deprecated': False},
+    'tpdl': {'id': 'TPDL', 'deprecated': False},
+    'tpl-1.0': {'id': 'TPL-1.0', 'deprecated': False},
+    'ttwl': {'id': 'TTWL', 'deprecated': False},
+    'ttyp0': {'id': 'TTYP0', 'deprecated': False},
+    'tu-berlin-1.0': {'id': 'TU-Berlin-1.0', 'deprecated': False},
+    'tu-berlin-2.0': {'id': 'TU-Berlin-2.0', 'deprecated': False},
+    'ubuntu-font-1.0': {'id': 'Ubuntu-font-1.0', 'deprecated': False},
+    'ucar': {'id': 'UCAR', 'deprecated': False},
+    'ucl-1.0': {'id': 'UCL-1.0', 'deprecated': False},
+    'ulem': {'id': 'ulem', 'deprecated': False},
+    'umich-merit': {'id': 'UMich-Merit', 'deprecated': False},
+    'unicode-3.0': {'id': 'Unicode-3.0', 'deprecated': False},
+    'unicode-dfs-2015': {'id': 'Unicode-DFS-2015', 'deprecated': False},
+    'unicode-dfs-2016': {'id': 'Unicode-DFS-2016', 'deprecated': False},
+    'unicode-tou': {'id': 'Unicode-TOU', 'deprecated': False},
+    'unixcrypt': {'id': 'UnixCrypt', 'deprecated': False},
+    'unlicense': {'id': 'Unlicense', 'deprecated': False},
+    'upl-1.0': {'id': 'UPL-1.0', 'deprecated': False},
+    'urt-rle': {'id': 'URT-RLE', 'deprecated': False},
+    'vim': {'id': 'Vim', 'deprecated': False},
+    'vostrom': {'id': 'VOSTROM', 'deprecated': False},
+    'vsl-1.0': {'id': 'VSL-1.0', 'deprecated': False},
+    'w3c': {'id': 'W3C', 'deprecated': False},
+    'w3c-19980720': {'id': 'W3C-19980720', 'deprecated': False},
+    'w3c-20150513': {'id': 'W3C-20150513', 'deprecated': False},
+    'w3m': {'id': 'w3m', 'deprecated': False},
+    'watcom-1.0': {'id': 'Watcom-1.0', 'deprecated': False},
+    'widget-workshop': {'id': 'Widget-Workshop', 'deprecated': False},
+    'wsuipa': {'id': 'Wsuipa', 'deprecated': False},
+    'wtfpl': {'id': 'WTFPL', 'deprecated': False},
+    'wxwindows': {'id': 'wxWindows', 'deprecated': True},
+    'x11': {'id': 'X11', 'deprecated': False},
+    'x11-distribute-modifications-variant': {'id': 'X11-distribute-modifications-variant', 'deprecated': False},
+    'x11-swapped': {'id': 'X11-swapped', 'deprecated': False},
+    'xdebug-1.03': {'id': 'Xdebug-1.03', 'deprecated': False},
+    'xerox': {'id': 'Xerox', 'deprecated': False},
+    'xfig': {'id': 'Xfig', 'deprecated': False},
+    'xfree86-1.1': {'id': 'XFree86-1.1', 'deprecated': False},
+    'xinetd': {'id': 'xinetd', 'deprecated': False},
+    'xkeyboard-config-zinoviev': {'id': 'xkeyboard-config-Zinoviev', 'deprecated': False},
+    'xlock': {'id': 'xlock', 'deprecated': False},
+    'xnet': {'id': 'Xnet', 'deprecated': False},
+    'xpp': {'id': 'xpp', 'deprecated': False},
+    'xskat': {'id': 'XSkat', 'deprecated': False},
+    'xzoom': {'id': 'xzoom', 'deprecated': False},
+    'ypl-1.0': {'id': 'YPL-1.0', 'deprecated': False},
+    'ypl-1.1': {'id': 'YPL-1.1', 'deprecated': False},
+    'zed': {'id': 'Zed', 'deprecated': False},
+    'zeeff': {'id': 'Zeeff', 'deprecated': False},
+    'zend-2.0': {'id': 'Zend-2.0', 'deprecated': False},
+    'zimbra-1.3': {'id': 'Zimbra-1.3', 'deprecated': False},
+    'zimbra-1.4': {'id': 'Zimbra-1.4', 'deprecated': False},
+    'zlib': {'id': 'Zlib', 'deprecated': False},
+    'zlib-acknowledgement': {'id': 'zlib-acknowledgement', 'deprecated': False},
+    'zpl-1.1': {'id': 'ZPL-1.1', 'deprecated': False},
+    'zpl-2.0': {'id': 'ZPL-2.0', 'deprecated': False},
+    'zpl-2.1': {'id': 'ZPL-2.1', 'deprecated': False},
+}
+
+EXCEPTIONS: dict[str, SPDXException] = {
+    '389-exception': {'id': '389-exception', 'deprecated': False},
+    'asterisk-exception': {'id': 'Asterisk-exception', 'deprecated': False},
+    'asterisk-linking-protocols-exception': {'id': 'Asterisk-linking-protocols-exception', 'deprecated': False},
+    'autoconf-exception-2.0': {'id': 'Autoconf-exception-2.0', 'deprecated': False},
+    'autoconf-exception-3.0': {'id': 'Autoconf-exception-3.0', 'deprecated': False},
+    'autoconf-exception-generic': {'id': 'Autoconf-exception-generic', 'deprecated': False},
+    'autoconf-exception-generic-3.0': {'id': 'Autoconf-exception-generic-3.0', 'deprecated': False},
+    'autoconf-exception-macro': {'id': 'Autoconf-exception-macro', 'deprecated': False},
+    'bison-exception-1.24': {'id': 'Bison-exception-1.24', 'deprecated': False},
+    'bison-exception-2.2': {'id': 'Bison-exception-2.2', 'deprecated': False},
+    'bootloader-exception': {'id': 'Bootloader-exception', 'deprecated': False},
+    'classpath-exception-2.0': {'id': 'Classpath-exception-2.0', 'deprecated': False},
+    'clisp-exception-2.0': {'id': 'CLISP-exception-2.0', 'deprecated': False},
+    'cryptsetup-openssl-exception': {'id': 'cryptsetup-OpenSSL-exception', 'deprecated': False},
+    'digirule-foss-exception': {'id': 'DigiRule-FOSS-exception', 'deprecated': False},
+    'ecos-exception-2.0': {'id': 'eCos-exception-2.0', 'deprecated': False},
+    'erlang-otp-linking-exception': {'id': 'erlang-otp-linking-exception', 'deprecated': False},
+    'fawkes-runtime-exception': {'id': 'Fawkes-Runtime-exception', 'deprecated': False},
+    'fltk-exception': {'id': 'FLTK-exception', 'deprecated': False},
+    'fmt-exception': {'id': 'fmt-exception', 'deprecated': False},
+    'font-exception-2.0': {'id': 'Font-exception-2.0', 'deprecated': False},
+    'freertos-exception-2.0': {'id': 'freertos-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0': {'id': 'GCC-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0-note': {'id': 'GCC-exception-2.0-note', 'deprecated': False},
+    'gcc-exception-3.1': {'id': 'GCC-exception-3.1', 'deprecated': False},
+    'gmsh-exception': {'id': 'Gmsh-exception', 'deprecated': False},
+    'gnat-exception': {'id': 'GNAT-exception', 'deprecated': False},
+    'gnome-examples-exception': {'id': 'GNOME-examples-exception', 'deprecated': False},
+    'gnu-compiler-exception': {'id': 'GNU-compiler-exception', 'deprecated': False},
+    'gnu-javamail-exception': {'id': 'gnu-javamail-exception', 'deprecated': False},
+    'gpl-3.0-interface-exception': {'id': 'GPL-3.0-interface-exception', 'deprecated': False},
+    'gpl-3.0-linking-exception': {'id': 'GPL-3.0-linking-exception', 'deprecated': False},
+    'gpl-3.0-linking-source-exception': {'id': 'GPL-3.0-linking-source-exception', 'deprecated': False},
+    'gpl-cc-1.0': {'id': 'GPL-CC-1.0', 'deprecated': False},
+    'gstreamer-exception-2005': {'id': 'GStreamer-exception-2005', 'deprecated': False},
+    'gstreamer-exception-2008': {'id': 'GStreamer-exception-2008', 'deprecated': False},
+    'i2p-gpl-java-exception': {'id': 'i2p-gpl-java-exception', 'deprecated': False},
+    'kicad-libraries-exception': {'id': 'KiCad-libraries-exception', 'deprecated': False},
+    'lgpl-3.0-linking-exception': {'id': 'LGPL-3.0-linking-exception', 'deprecated': False},
+    'libpri-openh323-exception': {'id': 'libpri-OpenH323-exception', 'deprecated': False},
+    'libtool-exception': {'id': 'Libtool-exception', 'deprecated': False},
+    'linux-syscall-note': {'id': 'Linux-syscall-note', 'deprecated': False},
+    'llgpl': {'id': 'LLGPL', 'deprecated': False},
+    'llvm-exception': {'id': 'LLVM-exception', 'deprecated': False},
+    'lzma-exception': {'id': 'LZMA-exception', 'deprecated': False},
+    'mif-exception': {'id': 'mif-exception', 'deprecated': False},
+    'nokia-qt-exception-1.1': {'id': 'Nokia-Qt-exception-1.1', 'deprecated': True},
+    'ocaml-lgpl-linking-exception': {'id': 'OCaml-LGPL-linking-exception', 'deprecated': False},
+    'occt-exception-1.0': {'id': 'OCCT-exception-1.0', 'deprecated': False},
+    'openjdk-assembly-exception-1.0': {'id': 'OpenJDK-assembly-exception-1.0', 'deprecated': False},
+    'openvpn-openssl-exception': {'id': 'openvpn-openssl-exception', 'deprecated': False},
+    'pcre2-exception': {'id': 'PCRE2-exception', 'deprecated': False},
+    'ps-or-pdf-font-exception-20170817': {'id': 'PS-or-PDF-font-exception-20170817', 'deprecated': False},
+    'qpl-1.0-inria-2004-exception': {'id': 'QPL-1.0-INRIA-2004-exception', 'deprecated': False},
+    'qt-gpl-exception-1.0': {'id': 'Qt-GPL-exception-1.0', 'deprecated': False},
+    'qt-lgpl-exception-1.1': {'id': 'Qt-LGPL-exception-1.1', 'deprecated': False},
+    'qwt-exception-1.0': {'id': 'Qwt-exception-1.0', 'deprecated': False},
+    'romic-exception': {'id': 'romic-exception', 'deprecated': False},
+    'rrdtool-floss-exception-2.0': {'id': 'RRDtool-FLOSS-exception-2.0', 'deprecated': False},
+    'sane-exception': {'id': 'SANE-exception', 'deprecated': False},
+    'shl-2.0': {'id': 'SHL-2.0', 'deprecated': False},
+    'shl-2.1': {'id': 'SHL-2.1', 'deprecated': False},
+    'stunnel-exception': {'id': 'stunnel-exception', 'deprecated': False},
+    'swi-exception': {'id': 'SWI-exception', 'deprecated': False},
+    'swift-exception': {'id': 'Swift-exception', 'deprecated': False},
+    'texinfo-exception': {'id': 'Texinfo-exception', 'deprecated': False},
+    'u-boot-exception-2.0': {'id': 'u-boot-exception-2.0', 'deprecated': False},
+    'ubdl-exception': {'id': 'UBDL-exception', 'deprecated': False},
+    'universal-foss-exception-1.0': {'id': 'Universal-FOSS-exception-1.0', 'deprecated': False},
+    'vsftpd-openssl-exception': {'id': 'vsftpd-openssl-exception', 'deprecated': False},
+    'wxwindows-exception-3.1': {'id': 'WxWindows-exception-3.1', 'deprecated': False},
+    'x11vnc-openssl-exception': {'id': 'x11vnc-openssl-exception', 'deprecated': False},
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/markers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/markers.py
new file mode 100644
index 00000000..fb7f49cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/markers.py
@@ -0,0 +1,331 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import operator
+import os
+import platform
+import sys
+from typing import Any, Callable, TypedDict, cast
+
+from ._parser import MarkerAtom, MarkerList, Op, Value, Variable
+from ._parser import parse_marker as _parse_marker
+from ._tokenizer import ParserSyntaxError
+from .specifiers import InvalidSpecifier, Specifier
+from .utils import canonicalize_name
+
+__all__ = [
+    "InvalidMarker",
+    "Marker",
+    "UndefinedComparison",
+    "UndefinedEnvironmentName",
+    "default_environment",
+]
+
+Operator = Callable[[str, str], bool]
+
+
+class InvalidMarker(ValueError):
+    """
+    An invalid marker was found, users should refer to PEP 508.
+    """
+
+
+class UndefinedComparison(ValueError):
+    """
+    An invalid operation was attempted on a value that doesn't support it.
+    """
+
+
+class UndefinedEnvironmentName(ValueError):
+    """
+    A name was attempted to be used that does not exist inside of the
+    environment.
+    """
+
+
+class Environment(TypedDict):
+    implementation_name: str
+    """The implementation's identifier, e.g. ``'cpython'``."""
+
+    implementation_version: str
+    """
+    The implementation's version, e.g. ``'3.13.0a2'`` for CPython 3.13.0a2, or
+    ``'7.3.13'`` for PyPy3.10 v7.3.13.
+    """
+
+    os_name: str
+    """
+    The value of :py:data:`os.name`. The name of the operating system dependent module
+    imported, e.g. ``'posix'``.
+    """
+
+    platform_machine: str
+    """
+    Returns the machine type, e.g. ``'i386'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_release: str
+    """
+    The system's release, e.g. ``'2.2.0'`` or ``'NT'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_system: str
+    """
+    The system/OS name, e.g. ``'Linux'``, ``'Windows'`` or ``'Java'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_version: str
+    """
+    The system's release version, e.g. ``'#3 on degas'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    python_full_version: str
+    """
+    The Python version as string ``'major.minor.patchlevel'``.
+
+    Note that unlike the Python :py:data:`sys.version`, this value will always include
+    the patchlevel (it defaults to 0).
+    """
+
+    platform_python_implementation: str
+    """
+    A string identifying the Python implementation, e.g. ``'CPython'``.
+    """
+
+    python_version: str
+    """The Python version as string ``'major.minor'``."""
+
+    sys_platform: str
+    """
+    This string contains a platform identifier that can be used to append
+    platform-specific components to :py:data:`sys.path`, for instance.
+
+    For Unix systems, except on Linux and AIX, this is the lowercased OS name as
+    returned by ``uname -s`` with the first part of the version as returned by
+    ``uname -r`` appended, e.g. ``'sunos5'`` or ``'freebsd8'``, at the time when Python
+    was built.
+    """
+
+
+def _normalize_extra_values(results: Any) -> Any:
+    """
+    Normalize extra values.
+    """
+    if isinstance(results[0], tuple):
+        lhs, op, rhs = results[0]
+        if isinstance(lhs, Variable) and lhs.value == "extra":
+            normalized_extra = canonicalize_name(rhs.value)
+            rhs = Value(normalized_extra)
+        elif isinstance(rhs, Variable) and rhs.value == "extra":
+            normalized_extra = canonicalize_name(lhs.value)
+            lhs = Value(normalized_extra)
+        results[0] = lhs, op, rhs
+    return results
+
+
+def _format_marker(
+    marker: list[str] | MarkerAtom | str, first: bool | None = True
+) -> str:
+    assert isinstance(marker, (list, tuple, str))
+
+    # Sometimes we have a structure like [[...]] which is a single item list
+    # where the single item is itself it's own list. In that case we want skip
+    # the rest of this function so that we don't get extraneous () on the
+    # outside.
+    if (
+        isinstance(marker, list)
+        and len(marker) == 1
+        and isinstance(marker[0], (list, tuple))
+    ):
+        return _format_marker(marker[0])
+
+    if isinstance(marker, list):
+        inner = (_format_marker(m, first=False) for m in marker)
+        if first:
+            return " ".join(inner)
+        else:
+            return "(" + " ".join(inner) + ")"
+    elif isinstance(marker, tuple):
+        return " ".join([m.serialize() for m in marker])
+    else:
+        return marker
+
+
+_operators: dict[str, Operator] = {
+    "in": lambda lhs, rhs: lhs in rhs,
+    "not in": lambda lhs, rhs: lhs not in rhs,
+    "<": operator.lt,
+    "<=": operator.le,
+    "==": operator.eq,
+    "!=": operator.ne,
+    ">=": operator.ge,
+    ">": operator.gt,
+}
+
+
+def _eval_op(lhs: str, op: Op, rhs: str) -> bool:
+    try:
+        spec = Specifier("".join([op.serialize(), rhs]))
+    except InvalidSpecifier:
+        pass
+    else:
+        return spec.contains(lhs, prereleases=True)
+
+    oper: Operator | None = _operators.get(op.serialize())
+    if oper is None:
+        raise UndefinedComparison(f"Undefined {op!r} on {lhs!r} and {rhs!r}.")
+
+    return oper(lhs, rhs)
+
+
+def _normalize(*values: str, key: str) -> tuple[str, ...]:
+    # PEP 685 – Comparison of extra names for optional distribution dependencies
+    # https://peps.python.org/pep-0685/
+    # > When comparing extra names, tools MUST normalize the names being
+    # > compared using the semantics outlined in PEP 503 for names
+    if key == "extra":
+        return tuple(canonicalize_name(v) for v in values)
+
+    # other environment markers don't have such standards
+    return values
+
+
+def _evaluate_markers(markers: MarkerList, environment: dict[str, str]) -> bool:
+    groups: list[list[bool]] = [[]]
+
+    for marker in markers:
+        assert isinstance(marker, (list, tuple, str))
+
+        if isinstance(marker, list):
+            groups[-1].append(_evaluate_markers(marker, environment))
+        elif isinstance(marker, tuple):
+            lhs, op, rhs = marker
+
+            if isinstance(lhs, Variable):
+                environment_key = lhs.value
+                lhs_value = environment[environment_key]
+                rhs_value = rhs.value
+            else:
+                lhs_value = lhs.value
+                environment_key = rhs.value
+                rhs_value = environment[environment_key]
+
+            lhs_value, rhs_value = _normalize(lhs_value, rhs_value, key=environment_key)
+            groups[-1].append(_eval_op(lhs_value, op, rhs_value))
+        else:
+            assert marker in ["and", "or"]
+            if marker == "or":
+                groups.append([])
+
+    return any(all(item) for item in groups)
+
+
+def format_full_version(info: sys._version_info) -> str:
+    version = f"{info.major}.{info.minor}.{info.micro}"
+    kind = info.releaselevel
+    if kind != "final":
+        version += kind[0] + str(info.serial)
+    return version
+
+
+def default_environment() -> Environment:
+    iver = format_full_version(sys.implementation.version)
+    implementation_name = sys.implementation.name
+    return {
+        "implementation_name": implementation_name,
+        "implementation_version": iver,
+        "os_name": os.name,
+        "platform_machine": platform.machine(),
+        "platform_release": platform.release(),
+        "platform_system": platform.system(),
+        "platform_version": platform.version(),
+        "python_full_version": platform.python_version(),
+        "platform_python_implementation": platform.python_implementation(),
+        "python_version": ".".join(platform.python_version_tuple()[:2]),
+        "sys_platform": sys.platform,
+    }
+
+
+class Marker:
+    def __init__(self, marker: str) -> None:
+        # Note: We create a Marker object without calling this constructor in
+        #       packaging.requirements.Requirement. If any additional logic is
+        #       added here, make sure to mirror/adapt Requirement.
+        try:
+            self._markers = _normalize_extra_values(_parse_marker(marker))
+            # The attribute `_markers` can be described in terms of a recursive type:
+            # MarkerList = List[Union[Tuple[Node, ...], str, MarkerList]]
+            #
+            # For example, the following expression:
+            # python_version > "3.6" or (python_version == "3.6" and os_name == "unix")
+            #
+            # is parsed into:
+            # [
+            #     (<Variable('python_version')>, <Op('>')>, <Value('3.6')>),
+            #     'and',
+            #     [
+            #         (<Variable('python_version')>, <Op('==')>, <Value('3.6')>),
+            #         'or',
+            #         (<Variable('os_name')>, <Op('==')>, <Value('unix')>)
+            #     ]
+            # ]
+        except ParserSyntaxError as e:
+            raise InvalidMarker(str(e)) from e
+
+    def __str__(self) -> str:
+        return _format_marker(self._markers)
+
+    def __repr__(self) -> str:
+        return f"<Marker('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash((self.__class__.__name__, str(self)))
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Marker):
+            return NotImplemented
+
+        return str(self) == str(other)
+
+    def evaluate(self, environment: dict[str, str] | None = None) -> bool:
+        """Evaluate a marker.
+
+        Return the boolean from evaluating the given marker against the
+        environment. environment is an optional argument to override all or
+        part of the determined environment.
+
+        The environment is determined from the current Python process.
+        """
+        current_environment = cast("dict[str, str]", default_environment())
+        current_environment["extra"] = ""
+        if environment is not None:
+            current_environment.update(environment)
+            # The API used to allow setting extra to None. We need to handle this
+            # case for backwards compatibility.
+            if current_environment["extra"] is None:
+                current_environment["extra"] = ""
+
+        return _evaluate_markers(
+            self._markers, _repair_python_full_version(current_environment)
+        )
+
+
+def _repair_python_full_version(env: dict[str, str]) -> dict[str, str]:
+    """
+    Work around platform.python_version() returning something that is not PEP 440
+    compliant for non-tagged Python builds.
+    """
+    if env["python_full_version"].endswith("+"):
+        env["python_full_version"] += "local"
+    return env
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/metadata.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/metadata.py
new file mode 100644
index 00000000..721f411c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/metadata.py
@@ -0,0 +1,863 @@
+from __future__ import annotations
+
+import email.feedparser
+import email.header
+import email.message
+import email.parser
+import email.policy
+import pathlib
+import sys
+import typing
+from typing import (
+    Any,
+    Callable,
+    Generic,
+    Literal,
+    TypedDict,
+    cast,
+)
+
+from . import licenses, requirements, specifiers, utils
+from . import version as version_module
+from .licenses import NormalizedLicenseExpression
+
+T = typing.TypeVar("T")
+
+
+if sys.version_info >= (3, 11):  # pragma: no cover
+    ExceptionGroup = ExceptionGroup
+else:  # pragma: no cover
+
+    class ExceptionGroup(Exception):
+        """A minimal implementation of :external:exc:`ExceptionGroup` from Python 3.11.
+
+        If :external:exc:`ExceptionGroup` is already defined by Python itself,
+        that version is used instead.
+        """
+
+        message: str
+        exceptions: list[Exception]
+
+        def __init__(self, message: str, exceptions: list[Exception]) -> None:
+            self.message = message
+            self.exceptions = exceptions
+
+        def __repr__(self) -> str:
+            return f"{self.__class__.__name__}({self.message!r}, {self.exceptions!r})"
+
+
+class InvalidMetadata(ValueError):
+    """A metadata field contains invalid data."""
+
+    field: str
+    """The name of the field that contains invalid data."""
+
+    def __init__(self, field: str, message: str) -> None:
+        self.field = field
+        super().__init__(message)
+
+
+# The RawMetadata class attempts to make as few assumptions about the underlying
+# serialization formats as possible. The idea is that as long as a serialization
+# formats offer some very basic primitives in *some* way then we can support
+# serializing to and from that format.
+class RawMetadata(TypedDict, total=False):
+    """A dictionary of raw core metadata.
+
+    Each field in core metadata maps to a key of this dictionary (when data is
+    provided). The key is lower-case and underscores are used instead of dashes
+    compared to the equivalent core metadata field. Any core metadata field that
+    can be specified multiple times or can hold multiple values in a single
+    field have a key with a plural name. See :class:`Metadata` whose attributes
+    match the keys of this dictionary.
+
+    Core metadata fields that can be specified multiple times are stored as a
+    list or dict depending on which is appropriate for the field. Any fields
+    which hold multiple values in a single field are stored as a list.
+
+    """
+
+    # Metadata 1.0 - PEP 241
+    metadata_version: str
+    name: str
+    version: str
+    platforms: list[str]
+    summary: str
+    description: str
+    keywords: list[str]
+    home_page: str
+    author: str
+    author_email: str
+    license: str
+
+    # Metadata 1.1 - PEP 314
+    supported_platforms: list[str]
+    download_url: str
+    classifiers: list[str]
+    requires: list[str]
+    provides: list[str]
+    obsoletes: list[str]
+
+    # Metadata 1.2 - PEP 345
+    maintainer: str
+    maintainer_email: str
+    requires_dist: list[str]
+    provides_dist: list[str]
+    obsoletes_dist: list[str]
+    requires_python: str
+    requires_external: list[str]
+    project_urls: dict[str, str]
+
+    # Metadata 2.0
+    # PEP 426 attempted to completely revamp the metadata format
+    # but got stuck without ever being able to build consensus on
+    # it and ultimately ended up withdrawn.
+    #
+    # However, a number of tools had started emitting METADATA with
+    # `2.0` Metadata-Version, so for historical reasons, this version
+    # was skipped.
+
+    # Metadata 2.1 - PEP 566
+    description_content_type: str
+    provides_extra: list[str]
+
+    # Metadata 2.2 - PEP 643
+    dynamic: list[str]
+
+    # Metadata 2.3 - PEP 685
+    # No new fields were added in PEP 685, just some edge case were
+    # tightened up to provide better interoptability.
+
+    # Metadata 2.4 - PEP 639
+    license_expression: str
+    license_files: list[str]
+
+
+_STRING_FIELDS = {
+    "author",
+    "author_email",
+    "description",
+    "description_content_type",
+    "download_url",
+    "home_page",
+    "license",
+    "license_expression",
+    "maintainer",
+    "maintainer_email",
+    "metadata_version",
+    "name",
+    "requires_python",
+    "summary",
+    "version",
+}
+
+_LIST_FIELDS = {
+    "classifiers",
+    "dynamic",
+    "license_files",
+    "obsoletes",
+    "obsoletes_dist",
+    "platforms",
+    "provides",
+    "provides_dist",
+    "provides_extra",
+    "requires",
+    "requires_dist",
+    "requires_external",
+    "supported_platforms",
+}
+
+_DICT_FIELDS = {
+    "project_urls",
+}
+
+
+def _parse_keywords(data: str) -> list[str]:
+    """Split a string of comma-separated keywords into a list of keywords."""
+    return [k.strip() for k in data.split(",")]
+
+
+def _parse_project_urls(data: list[str]) -> dict[str, str]:
+    """Parse a list of label/URL string pairings separated by a comma."""
+    urls = {}
+    for pair in data:
+        # Our logic is slightly tricky here as we want to try and do
+        # *something* reasonable with malformed data.
+        #
+        # The main thing that we have to worry about, is data that does
+        # not have a ',' at all to split the label from the Value. There
+        # isn't a singular right answer here, and we will fail validation
+        # later on (if the caller is validating) so it doesn't *really*
+        # matter, but since the missing value has to be an empty str
+        # and our return value is dict[str, str], if we let the key
+        # be the missing value, then they'd have multiple '' values that
+        # overwrite each other in a accumulating dict.
+        #
+        # The other potentional issue is that it's possible to have the
+        # same label multiple times in the metadata, with no solid "right"
+        # answer with what to do in that case. As such, we'll do the only
+        # thing we can, which is treat the field as unparseable and add it
+        # to our list of unparsed fields.
+        parts = [p.strip() for p in pair.split(",", 1)]
+        parts.extend([""] * (max(0, 2 - len(parts))))  # Ensure 2 items
+
+        # TODO: The spec doesn't say anything about if the keys should be
+        #       considered case sensitive or not... logically they should
+        #       be case-preserving and case-insensitive, but doing that
+        #       would open up more cases where we might have duplicate
+        #       entries.
+        label, url = parts
+        if label in urls:
+            # The label already exists in our set of urls, so this field
+            # is unparseable, and we can just add the whole thing to our
+            # unparseable data and stop processing it.
+            raise KeyError("duplicate labels in project urls")
+        urls[label] = url
+
+    return urls
+
+
+def _get_payload(msg: email.message.Message, source: bytes | str) -> str:
+    """Get the body of the message."""
+    # If our source is a str, then our caller has managed encodings for us,
+    # and we don't need to deal with it.
+    if isinstance(source, str):
+        payload = msg.get_payload()
+        assert isinstance(payload, str)
+        return payload
+    # If our source is a bytes, then we're managing the encoding and we need
+    # to deal with it.
+    else:
+        bpayload = msg.get_payload(decode=True)
+        assert isinstance(bpayload, bytes)
+        try:
+            return bpayload.decode("utf8", "strict")
+        except UnicodeDecodeError as exc:
+            raise ValueError("payload in an invalid encoding") from exc
+
+
+# The various parse_FORMAT functions here are intended to be as lenient as
+# possible in their parsing, while still returning a correctly typed
+# RawMetadata.
+#
+# To aid in this, we also generally want to do as little touching of the
+# data as possible, except where there are possibly some historic holdovers
+# that make valid data awkward to work with.
+#
+# While this is a lower level, intermediate format than our ``Metadata``
+# class, some light touch ups can make a massive difference in usability.
+
+# Map METADATA fields to RawMetadata.
+_EMAIL_TO_RAW_MAPPING = {
+    "author": "author",
+    "author-email": "author_email",
+    "classifier": "classifiers",
+    "description": "description",
+    "description-content-type": "description_content_type",
+    "download-url": "download_url",
+    "dynamic": "dynamic",
+    "home-page": "home_page",
+    "keywords": "keywords",
+    "license": "license",
+    "license-expression": "license_expression",
+    "license-file": "license_files",
+    "maintainer": "maintainer",
+    "maintainer-email": "maintainer_email",
+    "metadata-version": "metadata_version",
+    "name": "name",
+    "obsoletes": "obsoletes",
+    "obsoletes-dist": "obsoletes_dist",
+    "platform": "platforms",
+    "project-url": "project_urls",
+    "provides": "provides",
+    "provides-dist": "provides_dist",
+    "provides-extra": "provides_extra",
+    "requires": "requires",
+    "requires-dist": "requires_dist",
+    "requires-external": "requires_external",
+    "requires-python": "requires_python",
+    "summary": "summary",
+    "supported-platform": "supported_platforms",
+    "version": "version",
+}
+_RAW_TO_EMAIL_MAPPING = {raw: email for email, raw in _EMAIL_TO_RAW_MAPPING.items()}
+
+
+def parse_email(data: bytes | str) -> tuple[RawMetadata, dict[str, list[str]]]:
+    """Parse a distribution's metadata stored as email headers (e.g. from ``METADATA``).
+
+    This function returns a two-item tuple of dicts. The first dict is of
+    recognized fields from the core metadata specification. Fields that can be
+    parsed and translated into Python's built-in types are converted
+    appropriately. All other fields are left as-is. Fields that are allowed to
+    appear multiple times are stored as lists.
+
+    The second dict contains all other fields from the metadata. This includes
+    any unrecognized fields. It also includes any fields which are expected to
+    be parsed into a built-in type but were not formatted appropriately. Finally,
+    any fields that are expected to appear only once but are repeated are
+    included in this dict.
+
+    """
+    raw: dict[str, str | list[str] | dict[str, str]] = {}
+    unparsed: dict[str, list[str]] = {}
+
+    if isinstance(data, str):
+        parsed = email.parser.Parser(policy=email.policy.compat32).parsestr(data)
+    else:
+        parsed = email.parser.BytesParser(policy=email.policy.compat32).parsebytes(data)
+
+    # We have to wrap parsed.keys() in a set, because in the case of multiple
+    # values for a key (a list), the key will appear multiple times in the
+    # list of keys, but we're avoiding that by using get_all().
+    for name in frozenset(parsed.keys()):
+        # Header names in RFC are case insensitive, so we'll normalize to all
+        # lower case to make comparisons easier.
+        name = name.lower()
+
+        # We use get_all() here, even for fields that aren't multiple use,
+        # because otherwise someone could have e.g. two Name fields, and we
+        # would just silently ignore it rather than doing something about it.
+        headers = parsed.get_all(name) or []
+
+        # The way the email module works when parsing bytes is that it
+        # unconditionally decodes the bytes as ascii using the surrogateescape
+        # handler. When you pull that data back out (such as with get_all() ),
+        # it looks to see if the str has any surrogate escapes, and if it does
+        # it wraps it in a Header object instead of returning the string.
+        #
+        # As such, we'll look for those Header objects, and fix up the encoding.
+        value = []
+        # Flag if we have run into any issues processing the headers, thus
+        # signalling that the data belongs in 'unparsed'.
+        valid_encoding = True
+        for h in headers:
+            # It's unclear if this can return more types than just a Header or
+            # a str, so we'll just assert here to make sure.
+            assert isinstance(h, (email.header.Header, str))
+
+            # If it's a header object, we need to do our little dance to get
+            # the real data out of it. In cases where there is invalid data
+            # we're going to end up with mojibake, but there's no obvious, good
+            # way around that without reimplementing parts of the Header object
+            # ourselves.
+            #
+            # That should be fine since, if mojibacked happens, this key is
+            # going into the unparsed dict anyways.
+            if isinstance(h, email.header.Header):
+                # The Header object stores it's data as chunks, and each chunk
+                # can be independently encoded, so we'll need to check each
+                # of them.
+                chunks: list[tuple[bytes, str | None]] = []
+                for bin, encoding in email.header.decode_header(h):
+                    try:
+                        bin.decode("utf8", "strict")
+                    except UnicodeDecodeError:
+                        # Enable mojibake.
+                        encoding = "latin1"
+                        valid_encoding = False
+                    else:
+                        encoding = "utf8"
+                    chunks.append((bin, encoding))
+
+                # Turn our chunks back into a Header object, then let that
+                # Header object do the right thing to turn them into a
+                # string for us.
+                value.append(str(email.header.make_header(chunks)))
+            # This is already a string, so just add it.
+            else:
+                value.append(h)
+
+        # We've processed all of our values to get them into a list of str,
+        # but we may have mojibake data, in which case this is an unparsed
+        # field.
+        if not valid_encoding:
+            unparsed[name] = value
+            continue
+
+        raw_name = _EMAIL_TO_RAW_MAPPING.get(name)
+        if raw_name is None:
+            # This is a bit of a weird situation, we've encountered a key that
+            # we don't know what it means, so we don't know whether it's meant
+            # to be a list or not.
+            #
+            # Since we can't really tell one way or another, we'll just leave it
+            # as a list, even though it may be a single item list, because that's
+            # what makes the most sense for email headers.
+            unparsed[name] = value
+            continue
+
+        # If this is one of our string fields, then we'll check to see if our
+        # value is a list of a single item. If it is then we'll assume that
+        # it was emitted as a single string, and unwrap the str from inside
+        # the list.
+        #
+        # If it's any other kind of data, then we haven't the faintest clue
+        # what we should parse it as, and we have to just add it to our list
+        # of unparsed stuff.
+        if raw_name in _STRING_FIELDS and len(value) == 1:
+            raw[raw_name] = value[0]
+        # If this is one of our list of string fields, then we can just assign
+        # the value, since email *only* has strings, and our get_all() call
+        # above ensures that this is a list.
+        elif raw_name in _LIST_FIELDS:
+            raw[raw_name] = value
+        # Special Case: Keywords
+        # The keywords field is implemented in the metadata spec as a str,
+        # but it conceptually is a list of strings, and is serialized using
+        # ", ".join(keywords), so we'll do some light data massaging to turn
+        # this into what it logically is.
+        elif raw_name == "keywords" and len(value) == 1:
+            raw[raw_name] = _parse_keywords(value[0])
+        # Special Case: Project-URL
+        # The project urls is implemented in the metadata spec as a list of
+        # specially-formatted strings that represent a key and a value, which
+        # is fundamentally a mapping, however the email format doesn't support
+        # mappings in a sane way, so it was crammed into a list of strings
+        # instead.
+        #
+        # We will do a little light data massaging to turn this into a map as
+        # it logically should be.
+        elif raw_name == "project_urls":
+            try:
+                raw[raw_name] = _parse_project_urls(value)
+            except KeyError:
+                unparsed[name] = value
+        # Nothing that we've done has managed to parse this, so it'll just
+        # throw it in our unparseable data and move on.
+        else:
+            unparsed[name] = value
+
+    # We need to support getting the Description from the message payload in
+    # addition to getting it from the the headers. This does mean, though, there
+    # is the possibility of it being set both ways, in which case we put both
+    # in 'unparsed' since we don't know which is right.
+    try:
+        payload = _get_payload(parsed, data)
+    except ValueError:
+        unparsed.setdefault("description", []).append(
+            parsed.get_payload(decode=isinstance(data, bytes))  # type: ignore[call-overload]
+        )
+    else:
+        if payload:
+            # Check to see if we've already got a description, if so then both
+            # it, and this body move to unparseable.
+            if "description" in raw:
+                description_header = cast(str, raw.pop("description"))
+                unparsed.setdefault("description", []).extend(
+                    [description_header, payload]
+                )
+            elif "description" in unparsed:
+                unparsed["description"].append(payload)
+            else:
+                raw["description"] = payload
+
+    # We need to cast our `raw` to a metadata, because a TypedDict only support
+    # literal key names, but we're computing our key names on purpose, but the
+    # way this function is implemented, our `TypedDict` can only have valid key
+    # names.
+    return cast(RawMetadata, raw), unparsed
+
+
+_NOT_FOUND = object()
+
+
+# Keep the two values in sync.
+_VALID_METADATA_VERSIONS = ["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+_MetadataVersion = Literal["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+
+_REQUIRED_ATTRS = frozenset(["metadata_version", "name", "version"])
+
+
+class _Validator(Generic[T]):
+    """Validate a metadata field.
+
+    All _process_*() methods correspond to a core metadata field. The method is
+    called with the field's raw value. If the raw value is valid it is returned
+    in its "enriched" form (e.g. ``version.Version`` for the ``Version`` field).
+    If the raw value is invalid, :exc:`InvalidMetadata` is raised (with a cause
+    as appropriate).
+    """
+
+    name: str
+    raw_name: str
+    added: _MetadataVersion
+
+    def __init__(
+        self,
+        *,
+        added: _MetadataVersion = "1.0",
+    ) -> None:
+        self.added = added
+
+    def __set_name__(self, _owner: Metadata, name: str) -> None:
+        self.name = name
+        self.raw_name = _RAW_TO_EMAIL_MAPPING[name]
+
+    def __get__(self, instance: Metadata, _owner: type[Metadata]) -> T:
+        # With Python 3.8, the caching can be replaced with functools.cached_property().
+        # No need to check the cache as attribute lookup will resolve into the
+        # instance's __dict__ before __get__ is called.
+        cache = instance.__dict__
+        value = instance._raw.get(self.name)
+
+        # To make the _process_* methods easier, we'll check if the value is None
+        # and if this field is NOT a required attribute, and if both of those
+        # things are true, we'll skip the the converter. This will mean that the
+        # converters never have to deal with the None union.
+        if self.name in _REQUIRED_ATTRS or value is not None:
+            try:
+                converter: Callable[[Any], T] = getattr(self, f"_process_{self.name}")
+            except AttributeError:
+                pass
+            else:
+                value = converter(value)
+
+        cache[self.name] = value
+        try:
+            del instance._raw[self.name]  # type: ignore[misc]
+        except KeyError:
+            pass
+
+        return cast(T, value)
+
+    def _invalid_metadata(
+        self, msg: str, cause: Exception | None = None
+    ) -> InvalidMetadata:
+        exc = InvalidMetadata(
+            self.raw_name, msg.format_map({"field": repr(self.raw_name)})
+        )
+        exc.__cause__ = cause
+        return exc
+
+    def _process_metadata_version(self, value: str) -> _MetadataVersion:
+        # Implicitly makes Metadata-Version required.
+        if value not in _VALID_METADATA_VERSIONS:
+            raise self._invalid_metadata(f"{value!r} is not a valid metadata version")
+        return cast(_MetadataVersion, value)
+
+    def _process_name(self, value: str) -> str:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        # Validate the name as a side-effect.
+        try:
+            utils.canonicalize_name(value, validate=True)
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return value
+
+    def _process_version(self, value: str) -> version_module.Version:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        try:
+            return version_module.parse(value)
+        except version_module.InvalidVersion as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_summary(self, value: str) -> str:
+        """Check the field contains no newlines."""
+        if "\n" in value:
+            raise self._invalid_metadata("{field} must be a single line")
+        return value
+
+    def _process_description_content_type(self, value: str) -> str:
+        content_types = {"text/plain", "text/x-rst", "text/markdown"}
+        message = email.message.EmailMessage()
+        message["content-type"] = value
+
+        content_type, parameters = (
+            # Defaults to `text/plain` if parsing failed.
+            message.get_content_type().lower(),
+            message["content-type"].params,
+        )
+        # Check if content-type is valid or defaulted to `text/plain` and thus was
+        # not parseable.
+        if content_type not in content_types or content_type not in value.lower():
+            raise self._invalid_metadata(
+                f"{{field}} must be one of {list(content_types)}, not {value!r}"
+            )
+
+        charset = parameters.get("charset", "UTF-8")
+        if charset != "UTF-8":
+            raise self._invalid_metadata(
+                f"{{field}} can only specify the UTF-8 charset, not {list(charset)}"
+            )
+
+        markdown_variants = {"GFM", "CommonMark"}
+        variant = parameters.get("variant", "GFM")  # Use an acceptable default.
+        if content_type == "text/markdown" and variant not in markdown_variants:
+            raise self._invalid_metadata(
+                f"valid Markdown variants for {{field}} are {list(markdown_variants)}, "
+                f"not {variant!r}",
+            )
+        return value
+
+    def _process_dynamic(self, value: list[str]) -> list[str]:
+        for dynamic_field in map(str.lower, value):
+            if dynamic_field in {"name", "version", "metadata-version"}:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not allowed as a dynamic field"
+                )
+            elif dynamic_field not in _EMAIL_TO_RAW_MAPPING:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not a valid dynamic field"
+                )
+        return list(map(str.lower, value))
+
+    def _process_provides_extra(
+        self,
+        value: list[str],
+    ) -> list[utils.NormalizedName]:
+        normalized_names = []
+        try:
+            for name in value:
+                normalized_names.append(utils.canonicalize_name(name, validate=True))
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{name!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return normalized_names
+
+    def _process_requires_python(self, value: str) -> specifiers.SpecifierSet:
+        try:
+            return specifiers.SpecifierSet(value)
+        except specifiers.InvalidSpecifier as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_requires_dist(
+        self,
+        value: list[str],
+    ) -> list[requirements.Requirement]:
+        reqs = []
+        try:
+            for req in value:
+                reqs.append(requirements.Requirement(req))
+        except requirements.InvalidRequirement as exc:
+            raise self._invalid_metadata(
+                f"{req!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return reqs
+
+    def _process_license_expression(
+        self, value: str
+    ) -> NormalizedLicenseExpression | None:
+        try:
+            return licenses.canonicalize_license_expression(value)
+        except ValueError as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_license_files(self, value: list[str]) -> list[str]:
+        paths = []
+        for path in value:
+            if ".." in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "parent directory indicators are not allowed"
+                )
+            if "*" in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be resolved"
+                )
+            if (
+                pathlib.PurePosixPath(path).is_absolute()
+                or pathlib.PureWindowsPath(path).is_absolute()
+            ):
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be relative"
+                )
+            if pathlib.PureWindowsPath(path).as_posix() != path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "paths must use '/' delimiter"
+                )
+            paths.append(path)
+        return paths
+
+
+class Metadata:
+    """Representation of distribution metadata.
+
+    Compared to :class:`RawMetadata`, this class provides objects representing
+    metadata fields instead of only using built-in types. Any invalid metadata
+    will cause :exc:`InvalidMetadata` to be raised (with a
+    :py:attr:`~BaseException.__cause__` attribute as appropriate).
+    """
+
+    _raw: RawMetadata
+
+    @classmethod
+    def from_raw(cls, data: RawMetadata, *, validate: bool = True) -> Metadata:
+        """Create an instance from :class:`RawMetadata`.
+
+        If *validate* is true, all metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        ins = cls()
+        ins._raw = data.copy()  # Mutations occur due to caching enriched values.
+
+        if validate:
+            exceptions: list[Exception] = []
+            try:
+                metadata_version = ins.metadata_version
+                metadata_age = _VALID_METADATA_VERSIONS.index(metadata_version)
+            except InvalidMetadata as metadata_version_exc:
+                exceptions.append(metadata_version_exc)
+                metadata_version = None
+
+            # Make sure to check for the fields that are present, the required
+            # fields (so their absence can be reported).
+            fields_to_check = frozenset(ins._raw) | _REQUIRED_ATTRS
+            # Remove fields that have already been checked.
+            fields_to_check -= {"metadata_version"}
+
+            for key in fields_to_check:
+                try:
+                    if metadata_version:
+                        # Can't use getattr() as that triggers descriptor protocol which
+                        # will fail due to no value for the instance argument.
+                        try:
+                            field_metadata_version = cls.__dict__[key].added
+                        except KeyError:
+                            exc = InvalidMetadata(key, f"unrecognized field: {key!r}")
+                            exceptions.append(exc)
+                            continue
+                        field_age = _VALID_METADATA_VERSIONS.index(
+                            field_metadata_version
+                        )
+                        if field_age > metadata_age:
+                            field = _RAW_TO_EMAIL_MAPPING[key]
+                            exc = InvalidMetadata(
+                                field,
+                                f"{field} introduced in metadata version "
+                                f"{field_metadata_version}, not {metadata_version}",
+                            )
+                            exceptions.append(exc)
+                            continue
+                    getattr(ins, key)
+                except InvalidMetadata as exc:
+                    exceptions.append(exc)
+
+            if exceptions:
+                raise ExceptionGroup("invalid metadata", exceptions)
+
+        return ins
+
+    @classmethod
+    def from_email(cls, data: bytes | str, *, validate: bool = True) -> Metadata:
+        """Parse metadata from email headers.
+
+        If *validate* is true, the metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        raw, unparsed = parse_email(data)
+
+        if validate:
+            exceptions: list[Exception] = []
+            for unparsed_key in unparsed:
+                if unparsed_key in _EMAIL_TO_RAW_MAPPING:
+                    message = f"{unparsed_key!r} has invalid data"
+                else:
+                    message = f"unrecognized field: {unparsed_key!r}"
+                exceptions.append(InvalidMetadata(unparsed_key, message))
+
+            if exceptions:
+                raise ExceptionGroup("unparsed", exceptions)
+
+        try:
+            return cls.from_raw(raw, validate=validate)
+        except ExceptionGroup as exc_group:
+            raise ExceptionGroup(
+                "invalid or unparsed metadata", exc_group.exceptions
+            ) from None
+
+    metadata_version: _Validator[_MetadataVersion] = _Validator()
+    """:external:ref:`core-metadata-metadata-version`
+    (required; validated to be a valid metadata version)"""
+    # `name` is not normalized/typed to NormalizedName so as to provide access to
+    # the original/raw name.
+    name: _Validator[str] = _Validator()
+    """:external:ref:`core-metadata-name`
+    (required; validated using :func:`~packaging.utils.canonicalize_name` and its
+    *validate* parameter)"""
+    version: _Validator[version_module.Version] = _Validator()
+    """:external:ref:`core-metadata-version` (required)"""
+    dynamic: _Validator[list[str] | None] = _Validator(
+        added="2.2",
+    )
+    """:external:ref:`core-metadata-dynamic`
+    (validated against core metadata field names and lowercased)"""
+    platforms: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-platform`"""
+    supported_platforms: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-supported-platform`"""
+    summary: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-summary` (validated to contain no newlines)"""
+    description: _Validator[str | None] = _Validator()  # TODO 2.1: can be in body
+    """:external:ref:`core-metadata-description`"""
+    description_content_type: _Validator[str | None] = _Validator(added="2.1")
+    """:external:ref:`core-metadata-description-content-type` (validated)"""
+    keywords: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-keywords`"""
+    home_page: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-home-page`"""
+    download_url: _Validator[str | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-download-url`"""
+    author: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author`"""
+    author_email: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author-email`"""
+    maintainer: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer`"""
+    maintainer_email: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer-email`"""
+    license: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-license`"""
+    license_expression: _Validator[NormalizedLicenseExpression | None] = _Validator(
+        added="2.4"
+    )
+    """:external:ref:`core-metadata-license-expression`"""
+    license_files: _Validator[list[str] | None] = _Validator(added="2.4")
+    """:external:ref:`core-metadata-license-file`"""
+    classifiers: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-classifier`"""
+    requires_dist: _Validator[list[requirements.Requirement] | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-dist`"""
+    requires_python: _Validator[specifiers.SpecifierSet | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-python`"""
+    # Because `Requires-External` allows for non-PEP 440 version specifiers, we
+    # don't do any processing on the values.
+    requires_external: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-requires-external`"""
+    project_urls: _Validator[dict[str, str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-project-url`"""
+    # PEP 685 lets us raise an error if an extra doesn't pass `Name` validation
+    # regardless of metadata version.
+    provides_extra: _Validator[list[utils.NormalizedName] | None] = _Validator(
+        added="2.1",
+    )
+    """:external:ref:`core-metadata-provides-extra`"""
+    provides_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-provides-dist`"""
+    obsoletes_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-obsoletes-dist`"""
+    requires: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Requires`` (deprecated)"""
+    provides: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Provides`` (deprecated)"""
+    obsoletes: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Obsoletes`` (deprecated)"""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/requirements.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/requirements.py
new file mode 100644
index 00000000..4e068c95
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/requirements.py
@@ -0,0 +1,91 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+from __future__ import annotations
+
+from typing import Any, Iterator
+
+from ._parser import parse_requirement as _parse_requirement
+from ._tokenizer import ParserSyntaxError
+from .markers import Marker, _normalize_extra_values
+from .specifiers import SpecifierSet
+from .utils import canonicalize_name
+
+
+class InvalidRequirement(ValueError):
+    """
+    An invalid requirement was found, users should refer to PEP 508.
+    """
+
+
+class Requirement:
+    """Parse a requirement.
+
+    Parse a given requirement string into its parts, such as name, specifier,
+    URL, and extras. Raises InvalidRequirement on a badly-formed requirement
+    string.
+    """
+
+    # TODO: Can we test whether something is contained within a requirement?
+    #       If so how do we do that? Do we need to test against the _name_ of
+    #       the thing as well as the version? What about the markers?
+    # TODO: Can we normalize the name and extra name?
+
+    def __init__(self, requirement_string: str) -> None:
+        try:
+            parsed = _parse_requirement(requirement_string)
+        except ParserSyntaxError as e:
+            raise InvalidRequirement(str(e)) from e
+
+        self.name: str = parsed.name
+        self.url: str | None = parsed.url or None
+        self.extras: set[str] = set(parsed.extras or [])
+        self.specifier: SpecifierSet = SpecifierSet(parsed.specifier)
+        self.marker: Marker | None = None
+        if parsed.marker is not None:
+            self.marker = Marker.__new__(Marker)
+            self.marker._markers = _normalize_extra_values(parsed.marker)
+
+    def _iter_parts(self, name: str) -> Iterator[str]:
+        yield name
+
+        if self.extras:
+            formatted_extras = ",".join(sorted(self.extras))
+            yield f"[{formatted_extras}]"
+
+        if self.specifier:
+            yield str(self.specifier)
+
+        if self.url:
+            yield f"@ {self.url}"
+            if self.marker:
+                yield " "
+
+        if self.marker:
+            yield f"; {self.marker}"
+
+    def __str__(self) -> str:
+        return "".join(self._iter_parts(self.name))
+
+    def __repr__(self) -> str:
+        return f"<Requirement('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash(
+            (
+                self.__class__.__name__,
+                *self._iter_parts(canonicalize_name(self.name)),
+            )
+        )
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Requirement):
+            return NotImplemented
+
+        return (
+            canonicalize_name(self.name) == canonicalize_name(other.name)
+            and self.extras == other.extras
+            and self.specifier == other.specifier
+            and self.url == other.url
+            and self.marker == other.marker
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/specifiers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/specifiers.py
new file mode 100644
index 00000000..f18016e1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/specifiers.py
@@ -0,0 +1,1020 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from pip._vendor.packaging.specifiers import Specifier, SpecifierSet, InvalidSpecifier
+    from pip._vendor.packaging.version import Version
+"""
+
+from __future__ import annotations
+
+import abc
+import itertools
+import re
+from typing import Callable, Iterable, Iterator, TypeVar, Union
+
+from .utils import canonicalize_version
+from .version import Version
+
+UnparsedVersion = Union[Version, str]
+UnparsedVersionVar = TypeVar("UnparsedVersionVar", bound=UnparsedVersion)
+CallableOperator = Callable[[Version, str], bool]
+
+
+def _coerce_version(version: UnparsedVersion) -> Version:
+    if not isinstance(version, Version):
+        version = Version(version)
+    return version
+
+
+class InvalidSpecifier(ValueError):
+    """
+    Raised when attempting to create a :class:`Specifier` with a specifier
+    string that is invalid.
+
+    >>> Specifier("lolwat")
+    Traceback (most recent call last):
+        ...
+    packaging.specifiers.InvalidSpecifier: Invalid specifier: 'lolwat'
+    """
+
+
+class BaseSpecifier(metaclass=abc.ABCMeta):
+    @abc.abstractmethod
+    def __str__(self) -> str:
+        """
+        Returns the str representation of this Specifier-like object. This
+        should be representative of the Specifier itself.
+        """
+
+    @abc.abstractmethod
+    def __hash__(self) -> int:
+        """
+        Returns a hash value for this Specifier-like object.
+        """
+
+    @abc.abstractmethod
+    def __eq__(self, other: object) -> bool:
+        """
+        Returns a boolean representing whether or not the two Specifier-like
+        objects are equal.
+
+        :param other: The other object to check against.
+        """
+
+    @property
+    @abc.abstractmethod
+    def prereleases(self) -> bool | None:
+        """Whether or not pre-releases as a whole are allowed.
+
+        This can be set to either ``True`` or ``False`` to explicitly enable or disable
+        prereleases or it can be set to ``None`` (the default) to use default semantics.
+        """
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        """Setter for :attr:`prereleases`.
+
+        :param value: The value to set.
+        """
+
+    @abc.abstractmethod
+    def contains(self, item: str, prereleases: bool | None = None) -> bool:
+        """
+        Determines if the given item is contained within this specifier.
+        """
+
+    @abc.abstractmethod
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """
+        Takes an iterable of items and filters them so that only items which
+        are contained within this specifier are allowed in it.
+        """
+
+
+class Specifier(BaseSpecifier):
+    """This class abstracts handling of version specifiers.
+
+    .. tip::
+
+        It is generally not required to instantiate this manually. You should instead
+        prefer to work with :class:`SpecifierSet` instead, which can parse
+        comma-separated version specifiers (which is what package metadata contains).
+    """
+
+    _operator_regex_str = r"""
+        (?P<operator>(~=|==|!=|<=|>=|<|>|===))
+        """
+    _version_regex_str = r"""
+        (?P<version>
+            (?:
+                # The identity operators allow for an escape hatch that will
+                # do an exact string match of the version you wish to install.
+                # This will not be parsed by PEP 440 and we cannot determine
+                # any semantic meaning from it. This operator is discouraged
+                # but included entirely as an escape hatch.
+                (?<====)  # Only match for the identity operator
+                \s*
+                [^\s;)]*  # The arbitrary version can be just about anything,
+                          # we match everything except for whitespace, a
+                          # semi-colon for marker support, and a closing paren
+                          # since versions can be enclosed in them.
+            )
+            |
+            (?:
+                # The (non)equality operators allow for wild card and local
+                # versions to be specified so we have to define these two
+                # operators separately to enable that.
+                (?<===|!=)            # Only match for equals and not equals
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+
+                # You cannot use a wild card and a pre-release, post-release, a dev or
+                # local version together so group them with a | and make them optional.
+                (?:
+                    \.\*  # Wild card syntax of .*
+                    |
+                    (?:                                  # pre release
+                        [-_\.]?
+                        (alpha|beta|preview|pre|a|b|c|rc)
+                        [-_\.]?
+                        [0-9]*
+                    )?
+                    (?:                                  # post release
+                        (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                    )?
+                    (?:[-_\.]?dev[-_\.]?[0-9]*)?         # dev release
+                    (?:\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*)? # local
+                )?
+            )
+            |
+            (?:
+                # The compatible operator requires at least two digits in the
+                # release segment.
+                (?<=~=)               # Only match for the compatible operator
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)+   # release  (We have a + instead of a *)
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+            |
+            (?:
+                # All other operators only allow a sub set of what the
+                # (non)equality operators do. Specifically they do not allow
+                # local versions to be specified nor do they allow the prefix
+                # matching wild cards.
+                (?<!==|!=|~=)         # We have special cases for these
+                                      # operators so we want to make sure they
+                                      # don't match here.
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+        )
+        """
+
+    _regex = re.compile(
+        r"^\s*" + _operator_regex_str + _version_regex_str + r"\s*$",
+        re.VERBOSE | re.IGNORECASE,
+    )
+
+    _operators = {
+        "~=": "compatible",
+        "==": "equal",
+        "!=": "not_equal",
+        "<=": "less_than_equal",
+        ">=": "greater_than_equal",
+        "<": "less_than",
+        ">": "greater_than",
+        "===": "arbitrary",
+    }
+
+    def __init__(self, spec: str = "", prereleases: bool | None = None) -> None:
+        """Initialize a Specifier instance.
+
+        :param spec:
+            The string representation of a specifier which will be parsed and
+            normalized before use.
+        :param prereleases:
+            This tells the specifier if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+        :raises InvalidSpecifier:
+            If the given specifier is invalid (i.e. bad syntax).
+        """
+        match = self._regex.search(spec)
+        if not match:
+            raise InvalidSpecifier(f"Invalid specifier: {spec!r}")
+
+        self._spec: tuple[str, str] = (
+            match.group("operator").strip(),
+            match.group("version").strip(),
+        )
+
+        # Store whether or not this Specifier should accept prereleases
+        self._prereleases = prereleases
+
+    # https://github.com/python/mypy/pull/13475#pullrequestreview-1079784515
+    @property  # type: ignore[override]
+    def prereleases(self) -> bool:
+        # If there is an explicit prereleases set for this, then we'll just
+        # blindly use that.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # Look at all of our specifiers and determine if they are inclusive
+        # operators, and if they are if they are including an explicit
+        # prerelease.
+        operator, version = self._spec
+        if operator in ["==", ">=", "<=", "~=", "===", ">", "<"]:
+            # The == specifier can include a trailing .*, if it does we
+            # want to remove before parsing.
+            if operator == "==" and version.endswith(".*"):
+                version = version[:-2]
+
+            # Parse the version, and if it is a pre-release than this
+            # specifier allows pre-releases.
+            if Version(version).is_prerelease:
+                return True
+
+        return False
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    @property
+    def operator(self) -> str:
+        """The operator of this specifier.
+
+        >>> Specifier("==1.2.3").operator
+        '=='
+        """
+        return self._spec[0]
+
+    @property
+    def version(self) -> str:
+        """The version of this specifier.
+
+        >>> Specifier("==1.2.3").version
+        '1.2.3'
+        """
+        return self._spec[1]
+
+    def __repr__(self) -> str:
+        """A representation of the Specifier that shows all internal state.
+
+        >>> Specifier('>=1.0.0')
+        <Specifier('>=1.0.0')>
+        >>> Specifier('>=1.0.0', prereleases=False)
+        <Specifier('>=1.0.0', prereleases=False)>
+        >>> Specifier('>=1.0.0', prereleases=True)
+        <Specifier('>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<{self.__class__.__name__}({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the Specifier that can be round-tripped.
+
+        >>> str(Specifier('>=1.0.0'))
+        '>=1.0.0'
+        >>> str(Specifier('>=1.0.0', prereleases=False))
+        '>=1.0.0'
+        """
+        return "{}{}".format(*self._spec)
+
+    @property
+    def _canonical_spec(self) -> tuple[str, str]:
+        canonical_version = canonicalize_version(
+            self._spec[1],
+            strip_trailing_zero=(self._spec[0] != "~="),
+        )
+        return self._spec[0], canonical_version
+
+    def __hash__(self) -> int:
+        return hash(self._canonical_spec)
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two Specifier-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> Specifier("==1.2.3") == Specifier("== 1.2.3.0")
+        True
+        >>> (Specifier("==1.2.3", prereleases=False) ==
+        ...  Specifier("==1.2.3", prereleases=True))
+        True
+        >>> Specifier("==1.2.3") == "==1.2.3"
+        True
+        >>> Specifier("==1.2.3") == Specifier("==1.2.4")
+        False
+        >>> Specifier("==1.2.3") == Specifier("~=1.2.3")
+        False
+        """
+        if isinstance(other, str):
+            try:
+                other = self.__class__(str(other))
+            except InvalidSpecifier:
+                return NotImplemented
+        elif not isinstance(other, self.__class__):
+            return NotImplemented
+
+        return self._canonical_spec == other._canonical_spec
+
+    def _get_operator(self, op: str) -> CallableOperator:
+        operator_callable: CallableOperator = getattr(
+            self, f"_compare_{self._operators[op]}"
+        )
+        return operator_callable
+
+    def _compare_compatible(self, prospective: Version, spec: str) -> bool:
+        # Compatible releases have an equivalent combination of >= and ==. That
+        # is that ~=2.2 is equivalent to >=2.2,==2.*. This allows us to
+        # implement this in terms of the other specifiers instead of
+        # implementing it ourselves. The only thing we need to do is construct
+        # the other specifiers.
+
+        # We want everything but the last item in the version, but we want to
+        # ignore suffix segments.
+        prefix = _version_join(
+            list(itertools.takewhile(_is_not_suffix, _version_split(spec)))[:-1]
+        )
+
+        # Add the prefix notation to the end of our string
+        prefix += ".*"
+
+        return self._get_operator(">=")(prospective, spec) and self._get_operator("==")(
+            prospective, prefix
+        )
+
+    def _compare_equal(self, prospective: Version, spec: str) -> bool:
+        # We need special logic to handle prefix matching
+        if spec.endswith(".*"):
+            # In the case of prefix matching we want to ignore local segment.
+            normalized_prospective = canonicalize_version(
+                prospective.public, strip_trailing_zero=False
+            )
+            # Get the normalized version string ignoring the trailing .*
+            normalized_spec = canonicalize_version(spec[:-2], strip_trailing_zero=False)
+            # Split the spec out by bangs and dots, and pretend that there is
+            # an implicit dot in between a release segment and a pre-release segment.
+            split_spec = _version_split(normalized_spec)
+
+            # Split the prospective version out by bangs and dots, and pretend
+            # that there is an implicit dot in between a release segment and
+            # a pre-release segment.
+            split_prospective = _version_split(normalized_prospective)
+
+            # 0-pad the prospective version before shortening it to get the correct
+            # shortened version.
+            padded_prospective, _ = _pad_version(split_prospective, split_spec)
+
+            # Shorten the prospective version to be the same length as the spec
+            # so that we can determine if the specifier is a prefix of the
+            # prospective version or not.
+            shortened_prospective = padded_prospective[: len(split_spec)]
+
+            return shortened_prospective == split_spec
+        else:
+            # Convert our spec string into a Version
+            spec_version = Version(spec)
+
+            # If the specifier does not have a local segment, then we want to
+            # act as if the prospective version also does not have a local
+            # segment.
+            if not spec_version.local:
+                prospective = Version(prospective.public)
+
+            return prospective == spec_version
+
+    def _compare_not_equal(self, prospective: Version, spec: str) -> bool:
+        return not self._compare_equal(prospective, spec)
+
+    def _compare_less_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) <= Version(spec)
+
+    def _compare_greater_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) >= Version(spec)
+
+    def _compare_less_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is less than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective < spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a pre-release version, that we do not accept pre-release
+        # versions for the version mentioned in the specifier (e.g. <3.1 should
+        # not match 3.1.dev0, but should match 3.0.dev0).
+        if not spec.is_prerelease and prospective.is_prerelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # less than the spec version *and* it's not a pre-release of the same
+        # version in the spec.
+        return True
+
+    def _compare_greater_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is greater than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective > spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a post-release version, that we do not accept
+        # post-release versions for the version mentioned in the specifier
+        # (e.g. >3.1 should not match 3.0.post0, but should match 3.2.post0).
+        if not spec.is_postrelease and prospective.is_postrelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # Ensure that we do not allow a local version of the version mentioned
+        # in the specifier, which is technically greater than, to match.
+        if prospective.local is not None:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # greater than the spec version *and* it's not a pre-release of the
+        # same version in the spec.
+        return True
+
+    def _compare_arbitrary(self, prospective: Version, spec: str) -> bool:
+        return str(prospective).lower() == str(spec).lower()
+
+    def __contains__(self, item: str | Version) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in Specifier(">=1.2.3")
+        True
+        >>> Version("1.2.3") in Specifier(">=1.2.3")
+        True
+        >>> "1.0.0" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(self, item: UnparsedVersion, prereleases: bool | None = None) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this Specifier. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> Specifier(">=1.2.3").contains("1.2.3")
+        True
+        >>> Specifier(">=1.2.3").contains(Version("1.2.3"))
+        True
+        >>> Specifier(">=1.2.3").contains("1.0.0")
+        False
+        >>> Specifier(">=1.2.3").contains("1.3.0a1")
+        False
+        >>> Specifier(">=1.2.3", prereleases=True).contains("1.3.0a1")
+        True
+        >>> Specifier(">=1.2.3").contains("1.3.0a1", prereleases=True)
+        True
+        """
+
+        # Determine if prereleases are to be allowed or not.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # Normalize item to a Version, this allows us to have a shortcut for
+        # "2.0" in Specifier(">=2")
+        normalized_item = _coerce_version(item)
+
+        # Determine if we should be supporting prereleases in this specifier
+        # or not, if we do not support prereleases than we can short circuit
+        # logic if this version is a prereleases.
+        if normalized_item.is_prerelease and not prereleases:
+            return False
+
+        # Actually do the comparison to determine if this item is contained
+        # within this Specifier or not.
+        operator_callable: CallableOperator = self._get_operator(self.operator)
+        return operator_callable(normalized_item, self.version)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifier.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(Specifier().contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.2.3", "1.3", Version("1.4")]))
+        ['1.2.3', '1.3', <Version('1.4')>]
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.5a1"]))
+        ['1.5a1']
+        >>> list(Specifier(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(Specifier(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        """
+
+        yielded = False
+        found_prereleases = []
+
+        kw = {"prereleases": prereleases if prereleases is not None else True}
+
+        # Attempt to iterate over all the values in the iterable and if any of
+        # them match, yield them.
+        for version in iterable:
+            parsed_version = _coerce_version(version)
+
+            if self.contains(parsed_version, **kw):
+                # If our version is a prerelease, and we were not set to allow
+                # prereleases, then we'll store it for later in case nothing
+                # else matches this specifier.
+                if parsed_version.is_prerelease and not (
+                    prereleases or self.prereleases
+                ):
+                    found_prereleases.append(version)
+                # Either this is not a prerelease, or we should have been
+                # accepting prereleases from the beginning.
+                else:
+                    yielded = True
+                    yield version
+
+        # Now that we've iterated over everything, determine if we've yielded
+        # any values, and if we have not and we have any prereleases stored up
+        # then we will go ahead and yield the prereleases.
+        if not yielded and found_prereleases:
+            for version in found_prereleases:
+                yield version
+
+
+_prefix_regex = re.compile(r"^([0-9]+)((?:a|b|c|rc)[0-9]+)$")
+
+
+def _version_split(version: str) -> list[str]:
+    """Split version into components.
+
+    The split components are intended for version comparison. The logic does
+    not attempt to retain the original version string, so joining the
+    components back with :func:`_version_join` may not produce the original
+    version string.
+    """
+    result: list[str] = []
+
+    epoch, _, rest = version.rpartition("!")
+    result.append(epoch or "0")
+
+    for item in rest.split("."):
+        match = _prefix_regex.search(item)
+        if match:
+            result.extend(match.groups())
+        else:
+            result.append(item)
+    return result
+
+
+def _version_join(components: list[str]) -> str:
+    """Join split version components into a version string.
+
+    This function assumes the input came from :func:`_version_split`, where the
+    first component must be the epoch (either empty or numeric), and all other
+    components numeric.
+    """
+    epoch, *rest = components
+    return f"{epoch}!{'.'.join(rest)}"
+
+
+def _is_not_suffix(segment: str) -> bool:
+    return not any(
+        segment.startswith(prefix) for prefix in ("dev", "a", "b", "rc", "post")
+    )
+
+
+def _pad_version(left: list[str], right: list[str]) -> tuple[list[str], list[str]]:
+    left_split, right_split = [], []
+
+    # Get the release segment of our versions
+    left_split.append(list(itertools.takewhile(lambda x: x.isdigit(), left)))
+    right_split.append(list(itertools.takewhile(lambda x: x.isdigit(), right)))
+
+    # Get the rest of our versions
+    left_split.append(left[len(left_split[0]) :])
+    right_split.append(right[len(right_split[0]) :])
+
+    # Insert our padding
+    left_split.insert(1, ["0"] * max(0, len(right_split[0]) - len(left_split[0])))
+    right_split.insert(1, ["0"] * max(0, len(left_split[0]) - len(right_split[0])))
+
+    return (
+        list(itertools.chain.from_iterable(left_split)),
+        list(itertools.chain.from_iterable(right_split)),
+    )
+
+
+class SpecifierSet(BaseSpecifier):
+    """This class abstracts handling of a set of version specifiers.
+
+    It can be passed a single specifier (``>=3.0``), a comma-separated list of
+    specifiers (``>=3.0,!=3.1``), or no specifier at all.
+    """
+
+    def __init__(
+        self,
+        specifiers: str | Iterable[Specifier] = "",
+        prereleases: bool | None = None,
+    ) -> None:
+        """Initialize a SpecifierSet instance.
+
+        :param specifiers:
+            The string representation of a specifier or a comma-separated list of
+            specifiers which will be parsed and normalized before use.
+            May also be an iterable of ``Specifier`` instances, which will be used
+            as is.
+        :param prereleases:
+            This tells the SpecifierSet if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+
+        :raises InvalidSpecifier:
+            If the given ``specifiers`` are not parseable than this exception will be
+            raised.
+        """
+
+        if isinstance(specifiers, str):
+            # Split on `,` to break each individual specifier into its own item, and
+            # strip each item to remove leading/trailing whitespace.
+            split_specifiers = [s.strip() for s in specifiers.split(",") if s.strip()]
+
+            # Make each individual specifier a Specifier and save in a frozen set
+            # for later.
+            self._specs = frozenset(map(Specifier, split_specifiers))
+        else:
+            # Save the supplied specifiers in a frozen set.
+            self._specs = frozenset(specifiers)
+
+        # Store our prereleases value so we can use it later to determine if
+        # we accept prereleases or not.
+        self._prereleases = prereleases
+
+    @property
+    def prereleases(self) -> bool | None:
+        # If we have been given an explicit prerelease modifier, then we'll
+        # pass that through here.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # If we don't have any specifiers, and we don't have a forced value,
+        # then we'll just return None since we don't know if this should have
+        # pre-releases or not.
+        if not self._specs:
+            return None
+
+        # Otherwise we'll see if any of the given specifiers accept
+        # prereleases, if any of them do we'll return True, otherwise False.
+        return any(s.prereleases for s in self._specs)
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    def __repr__(self) -> str:
+        """A representation of the specifier set that shows all internal state.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> SpecifierSet('>=1.0.0,!=2.0.0')
+        <SpecifierSet('!=2.0.0,>=1.0.0')>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=False)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=False)>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=True)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<SpecifierSet({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the specifier set that can be round-tripped.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1"))
+        '!=1.0.1,>=1.0.0'
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False))
+        '!=1.0.1,>=1.0.0'
+        """
+        return ",".join(sorted(str(s) for s in self._specs))
+
+    def __hash__(self) -> int:
+        return hash(self._specs)
+
+    def __and__(self, other: SpecifierSet | str) -> SpecifierSet:
+        """Return a SpecifierSet which is a combination of the two sets.
+
+        :param other: The other object to combine with.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & '<=2.0.0,!=2.0.1'
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & SpecifierSet('<=2.0.0,!=2.0.1')
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        """
+        if isinstance(other, str):
+            other = SpecifierSet(other)
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        specifier = SpecifierSet()
+        specifier._specs = frozenset(self._specs | other._specs)
+
+        if self._prereleases is None and other._prereleases is not None:
+            specifier._prereleases = other._prereleases
+        elif self._prereleases is not None and other._prereleases is None:
+            specifier._prereleases = self._prereleases
+        elif self._prereleases == other._prereleases:
+            specifier._prereleases = self._prereleases
+        else:
+            raise ValueError(
+                "Cannot combine SpecifierSets with True and False prerelease "
+                "overrides."
+            )
+
+        return specifier
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two SpecifierSet-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> (SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False) ==
+        ...  SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == ">=1.0.0,!=1.0.1"
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.2")
+        False
+        """
+        if isinstance(other, (str, Specifier)):
+            other = SpecifierSet(str(other))
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        return self._specs == other._specs
+
+    def __len__(self) -> int:
+        """Returns the number of specifiers in this specifier set."""
+        return len(self._specs)
+
+    def __iter__(self) -> Iterator[Specifier]:
+        """
+        Returns an iterator over all the underlying :class:`Specifier` instances
+        in this specifier set.
+
+        >>> sorted(SpecifierSet(">=1.0.0,!=1.0.1"), key=str)
+        [<Specifier('!=1.0.1')>, <Specifier('>=1.0.0')>]
+        """
+        return iter(self._specs)
+
+    def __contains__(self, item: UnparsedVersion) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> Version("1.2.3") in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> "1.0.1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(
+        self,
+        item: UnparsedVersion,
+        prereleases: bool | None = None,
+        installed: bool | None = None,
+    ) -> bool:
+        """Return whether or not the item is contained in this SpecifierSet.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this SpecifierSet. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.2.3")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains(Version("1.2.3"))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.0.1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True).contains("1.3.0a1")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1", prereleases=True)
+        True
+        """
+        # Ensure that our item is a Version instance.
+        if not isinstance(item, Version):
+            item = Version(item)
+
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # We can determine if we're going to allow pre-releases by looking to
+        # see if any of the underlying items supports them. If none of them do
+        # and this item is a pre-release then we do not allow it and we can
+        # short circuit that here.
+        # Note: This means that 1.0.dev1 would not be contained in something
+        #       like >=1.0.devabc however it would be in >=1.0.debabc,>0.0.dev0
+        if not prereleases and item.is_prerelease:
+            return False
+
+        if installed and item.is_prerelease:
+            item = Version(item.base_version)
+
+        # We simply dispatch to the underlying specs here to make sure that the
+        # given version is contained within all of them.
+        # Note: This use of all() here means that an empty set of specifiers
+        #       will always return True, this is an explicit design decision.
+        return all(s.contains(item, prereleases=prereleases) for s in self._specs)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifiers in this set.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(SpecifierSet(...).contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", Version("1.4")]))
+        ['1.3', <Version('1.4')>]
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.5a1"]))
+        []
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+
+        An "empty" SpecifierSet will filter items based on the presence of prerelease
+        versions in the set.
+
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet("").filter(["1.5a1"]))
+        ['1.5a1']
+        >>> list(SpecifierSet("", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        """
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # If we have any specifiers, then we want to wrap our iterable in the
+        # filter method for each one, this will act as a logical AND amongst
+        # each specifier.
+        if self._specs:
+            for spec in self._specs:
+                iterable = spec.filter(iterable, prereleases=bool(prereleases))
+            return iter(iterable)
+        # If we do not have any specifiers, then we need to have a rough filter
+        # which will filter out any pre-releases, unless there are no final
+        # releases.
+        else:
+            filtered: list[UnparsedVersionVar] = []
+            found_prereleases: list[UnparsedVersionVar] = []
+
+            for item in iterable:
+                parsed_version = _coerce_version(item)
+
+                # Store any item which is a pre-release for later unless we've
+                # already found a final version or we are accepting prereleases
+                if parsed_version.is_prerelease and not prereleases:
+                    if not filtered:
+                        found_prereleases.append(item)
+                else:
+                    filtered.append(item)
+
+            # If we've found no items except for pre-releases, then we'll go
+            # ahead and use the pre-releases
+            if not filtered and found_prereleases and prereleases is None:
+                return iter(found_prereleases)
+
+            return iter(filtered)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/tags.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/tags.py
new file mode 100644
index 00000000..f5903402
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/tags.py
@@ -0,0 +1,617 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import logging
+import platform
+import re
+import struct
+import subprocess
+import sys
+import sysconfig
+from importlib.machinery import EXTENSION_SUFFIXES
+from typing import (
+    Iterable,
+    Iterator,
+    Sequence,
+    Tuple,
+    cast,
+)
+
+from . import _manylinux, _musllinux
+
+logger = logging.getLogger(__name__)
+
+PythonVersion = Sequence[int]
+AppleVersion = Tuple[int, int]
+
+INTERPRETER_SHORT_NAMES: dict[str, str] = {
+    "python": "py",  # Generic.
+    "cpython": "cp",
+    "pypy": "pp",
+    "ironpython": "ip",
+    "jython": "jy",
+}
+
+
+_32_BIT_INTERPRETER = struct.calcsize("P") == 4
+
+
+class Tag:
+    """
+    A representation of the tag triple for a wheel.
+
+    Instances are considered immutable and thus are hashable. Equality checking
+    is also supported.
+    """
+
+    __slots__ = ["_abi", "_hash", "_interpreter", "_platform"]
+
+    def __init__(self, interpreter: str, abi: str, platform: str) -> None:
+        self._interpreter = interpreter.lower()
+        self._abi = abi.lower()
+        self._platform = platform.lower()
+        # The __hash__ of every single element in a Set[Tag] will be evaluated each time
+        # that a set calls its `.disjoint()` method, which may be called hundreds of
+        # times when scanning a page of links for packages with tags matching that
+        # Set[Tag]. Pre-computing the value here produces significant speedups for
+        # downstream consumers.
+        self._hash = hash((self._interpreter, self._abi, self._platform))
+
+    @property
+    def interpreter(self) -> str:
+        return self._interpreter
+
+    @property
+    def abi(self) -> str:
+        return self._abi
+
+    @property
+    def platform(self) -> str:
+        return self._platform
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Tag):
+            return NotImplemented
+
+        return (
+            (self._hash == other._hash)  # Short-circuit ASAP for perf reasons.
+            and (self._platform == other._platform)
+            and (self._abi == other._abi)
+            and (self._interpreter == other._interpreter)
+        )
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __str__(self) -> str:
+        return f"{self._interpreter}-{self._abi}-{self._platform}"
+
+    def __repr__(self) -> str:
+        return f"<{self} @ {id(self)}>"
+
+
+def parse_tag(tag: str) -> frozenset[Tag]:
+    """
+    Parses the provided tag (e.g. `py3-none-any`) into a frozenset of Tag instances.
+
+    Returning a set is required due to the possibility that the tag is a
+    compressed tag set.
+    """
+    tags = set()
+    interpreters, abis, platforms = tag.split("-")
+    for interpreter in interpreters.split("."):
+        for abi in abis.split("."):
+            for platform_ in platforms.split("."):
+                tags.add(Tag(interpreter, abi, platform_))
+    return frozenset(tags)
+
+
+def _get_config_var(name: str, warn: bool = False) -> int | str | None:
+    value: int | str | None = sysconfig.get_config_var(name)
+    if value is None and warn:
+        logger.debug(
+            "Config variable '%s' is unset, Python ABI tag may be incorrect", name
+        )
+    return value
+
+
+def _normalize_string(string: str) -> str:
+    return string.replace(".", "_").replace("-", "_").replace(" ", "_")
+
+
+def _is_threaded_cpython(abis: list[str]) -> bool:
+    """
+    Determine if the ABI corresponds to a threaded (`--disable-gil`) build.
+
+    The threaded builds are indicated by a "t" in the abiflags.
+    """
+    if len(abis) == 0:
+        return False
+    # expect e.g., cp313
+    m = re.match(r"cp\d+(.*)", abis[0])
+    if not m:
+        return False
+    abiflags = m.group(1)
+    return "t" in abiflags
+
+
+def _abi3_applies(python_version: PythonVersion, threading: bool) -> bool:
+    """
+    Determine if the Python version supports abi3.
+
+    PEP 384 was first implemented in Python 3.2. The threaded (`--disable-gil`)
+    builds do not support abi3.
+    """
+    return len(python_version) > 1 and tuple(python_version) >= (3, 2) and not threading
+
+
+def _cpython_abis(py_version: PythonVersion, warn: bool = False) -> list[str]:
+    py_version = tuple(py_version)  # To allow for version comparison.
+    abis = []
+    version = _version_nodot(py_version[:2])
+    threading = debug = pymalloc = ucs4 = ""
+    with_debug = _get_config_var("Py_DEBUG", warn)
+    has_refcount = hasattr(sys, "gettotalrefcount")
+    # Windows doesn't set Py_DEBUG, so checking for support of debug-compiled
+    # extension modules is the best option.
+    # https://github.com/pypa/pip/issues/3383#issuecomment-173267692
+    has_ext = "_d.pyd" in EXTENSION_SUFFIXES
+    if with_debug or (with_debug is None and (has_refcount or has_ext)):
+        debug = "d"
+    if py_version >= (3, 13) and _get_config_var("Py_GIL_DISABLED", warn):
+        threading = "t"
+    if py_version < (3, 8):
+        with_pymalloc = _get_config_var("WITH_PYMALLOC", warn)
+        if with_pymalloc or with_pymalloc is None:
+            pymalloc = "m"
+        if py_version < (3, 3):
+            unicode_size = _get_config_var("Py_UNICODE_SIZE", warn)
+            if unicode_size == 4 or (
+                unicode_size is None and sys.maxunicode == 0x10FFFF
+            ):
+                ucs4 = "u"
+    elif debug:
+        # Debug builds can also load "normal" extension modules.
+        # We can also assume no UCS-4 or pymalloc requirement.
+        abis.append(f"cp{version}{threading}")
+    abis.insert(0, f"cp{version}{threading}{debug}{pymalloc}{ucs4}")
+    return abis
+
+
+def cpython_tags(
+    python_version: PythonVersion | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a CPython interpreter.
+
+    The tags consist of:
+    - cp<python_version>-<abi>-<platform>
+    - cp<python_version>-abi3-<platform>
+    - cp<python_version>-none-<platform>
+    - cp<less than python_version>-abi3-<platform>  # Older Python versions down to 3.2.
+
+    If python_version only specifies a major version then user-provided ABIs and
+    the 'none' ABItag will be used.
+
+    If 'abi3' or 'none' are specified in 'abis' then they will be yielded at
+    their normal position and not at the beginning.
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+
+    interpreter = f"cp{_version_nodot(python_version[:2])}"
+
+    if abis is None:
+        if len(python_version) > 1:
+            abis = _cpython_abis(python_version, warn)
+        else:
+            abis = []
+    abis = list(abis)
+    # 'abi3' and 'none' are explicitly handled later.
+    for explicit_abi in ("abi3", "none"):
+        try:
+            abis.remove(explicit_abi)
+        except ValueError:
+            pass
+
+    platforms = list(platforms or platform_tags())
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+    threading = _is_threaded_cpython(abis)
+    use_abi3 = _abi3_applies(python_version, threading)
+    if use_abi3:
+        yield from (Tag(interpreter, "abi3", platform_) for platform_ in platforms)
+    yield from (Tag(interpreter, "none", platform_) for platform_ in platforms)
+
+    if use_abi3:
+        for minor_version in range(python_version[1] - 1, 1, -1):
+            for platform_ in platforms:
+                version = _version_nodot((python_version[0], minor_version))
+                interpreter = f"cp{version}"
+                yield Tag(interpreter, "abi3", platform_)
+
+
+def _generic_abi() -> list[str]:
+    """
+    Return the ABI tag based on EXT_SUFFIX.
+    """
+    # The following are examples of `EXT_SUFFIX`.
+    # We want to keep the parts which are related to the ABI and remove the
+    # parts which are related to the platform:
+    # - linux:   '.cpython-310-x86_64-linux-gnu.so' => cp310
+    # - mac:     '.cpython-310-darwin.so'           => cp310
+    # - win:     '.cp310-win_amd64.pyd'             => cp310
+    # - win:     '.pyd'                             => cp37 (uses _cpython_abis())
+    # - pypy:    '.pypy38-pp73-x86_64-linux-gnu.so' => pypy38_pp73
+    # - graalpy: '.graalpy-38-native-x86_64-darwin.dylib'
+    #                                               => graalpy_38_native
+
+    ext_suffix = _get_config_var("EXT_SUFFIX", warn=True)
+    if not isinstance(ext_suffix, str) or ext_suffix[0] != ".":
+        raise SystemError("invalid sysconfig.get_config_var('EXT_SUFFIX')")
+    parts = ext_suffix.split(".")
+    if len(parts) < 3:
+        # CPython3.7 and earlier uses ".pyd" on Windows.
+        return _cpython_abis(sys.version_info[:2])
+    soabi = parts[1]
+    if soabi.startswith("cpython"):
+        # non-windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi.startswith("cp"):
+        # windows
+        abi = soabi.split("-")[0]
+    elif soabi.startswith("pypy"):
+        abi = "-".join(soabi.split("-")[:2])
+    elif soabi.startswith("graalpy"):
+        abi = "-".join(soabi.split("-")[:3])
+    elif soabi:
+        # pyston, ironpython, others?
+        abi = soabi
+    else:
+        return []
+    return [_normalize_string(abi)]
+
+
+def generic_tags(
+    interpreter: str | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a generic interpreter.
+
+    The tags consist of:
+    - <interpreter>-<abi>-<platform>
+
+    The "none" ABI will be added if it was not explicitly provided.
+    """
+    if not interpreter:
+        interp_name = interpreter_name()
+        interp_version = interpreter_version(warn=warn)
+        interpreter = "".join([interp_name, interp_version])
+    if abis is None:
+        abis = _generic_abi()
+    else:
+        abis = list(abis)
+    platforms = list(platforms or platform_tags())
+    if "none" not in abis:
+        abis.append("none")
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+
+def _py_interpreter_range(py_version: PythonVersion) -> Iterator[str]:
+    """
+    Yields Python versions in descending order.
+
+    After the latest version, the major-only version will be yielded, and then
+    all previous versions of that major version.
+    """
+    if len(py_version) > 1:
+        yield f"py{_version_nodot(py_version[:2])}"
+    yield f"py{py_version[0]}"
+    if len(py_version) > 1:
+        for minor in range(py_version[1] - 1, -1, -1):
+            yield f"py{_version_nodot((py_version[0], minor))}"
+
+
+def compatible_tags(
+    python_version: PythonVersion | None = None,
+    interpreter: str | None = None,
+    platforms: Iterable[str] | None = None,
+) -> Iterator[Tag]:
+    """
+    Yields the sequence of tags that are compatible with a specific version of Python.
+
+    The tags consist of:
+    - py*-none-<platform>
+    - <interpreter>-none-any  # ... if `interpreter` is provided.
+    - py*-none-any
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+    platforms = list(platforms or platform_tags())
+    for version in _py_interpreter_range(python_version):
+        for platform_ in platforms:
+            yield Tag(version, "none", platform_)
+    if interpreter:
+        yield Tag(interpreter, "none", "any")
+    for version in _py_interpreter_range(python_version):
+        yield Tag(version, "none", "any")
+
+
+def _mac_arch(arch: str, is_32bit: bool = _32_BIT_INTERPRETER) -> str:
+    if not is_32bit:
+        return arch
+
+    if arch.startswith("ppc"):
+        return "ppc"
+
+    return "i386"
+
+
+def _mac_binary_formats(version: AppleVersion, cpu_arch: str) -> list[str]:
+    formats = [cpu_arch]
+    if cpu_arch == "x86_64":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat64", "fat32"])
+
+    elif cpu_arch == "i386":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat32", "fat"])
+
+    elif cpu_arch == "ppc64":
+        # TODO: Need to care about 32-bit PPC for ppc64 through 10.2?
+        if version > (10, 5) or version < (10, 4):
+            return []
+        formats.append("fat64")
+
+    elif cpu_arch == "ppc":
+        if version > (10, 6):
+            return []
+        formats.extend(["fat32", "fat"])
+
+    if cpu_arch in {"arm64", "x86_64"}:
+        formats.append("universal2")
+
+    if cpu_arch in {"x86_64", "i386", "ppc64", "ppc", "intel"}:
+        formats.append("universal")
+
+    return formats
+
+
+def mac_platforms(
+    version: AppleVersion | None = None, arch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for a macOS system.
+
+    The `version` parameter is a two-item tuple specifying the macOS version to
+    generate platform tags for. The `arch` parameter is the CPU architecture to
+    generate platform tags for. Both parameters default to the appropriate value
+    for the current system.
+    """
+    version_str, _, cpu_arch = platform.mac_ver()
+    if version is None:
+        version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+        if version == (10, 16):
+            # When built against an older macOS SDK, Python will report macOS 10.16
+            # instead of the real version.
+            version_str = subprocess.run(
+                [
+                    sys.executable,
+                    "-sS",
+                    "-c",
+                    "import platform; print(platform.mac_ver()[0])",
+                ],
+                check=True,
+                env={"SYSTEM_VERSION_COMPAT": "0"},
+                stdout=subprocess.PIPE,
+                text=True,
+            ).stdout
+            version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+    else:
+        version = version
+    if arch is None:
+        arch = _mac_arch(cpu_arch)
+    else:
+        arch = arch
+
+    if (10, 0) <= version and version < (11, 0):
+        # Prior to Mac OS 11, each yearly release of Mac OS bumped the
+        # "minor" version number.  The major version was always 10.
+        major_version = 10
+        for minor_version in range(version[1], -1, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Starting with Mac OS 11, each yearly release bumps the major version
+        # number.   The minor versions are now the midyear updates.
+        minor_version = 0
+        for major_version in range(version[0], 10, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Mac OS 11 on x86_64 is compatible with binaries from previous releases.
+        # Arm64 support was introduced in 11.0, so no Arm binaries from previous
+        # releases exist.
+        #
+        # However, the "universal2" binary format can have a
+        # macOS version earlier than 11.0 when the x86_64 part of the binary supports
+        # that version of macOS.
+        major_version = 10
+        if arch == "x86_64":
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_formats = _mac_binary_formats(compat_version, arch)
+                for binary_format in binary_formats:
+                    yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+        else:
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_format = "universal2"
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+
+def ios_platforms(
+    version: AppleVersion | None = None, multiarch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for an iOS system.
+
+    :param version: A two-item tuple specifying the iOS version to generate
+        platform tags for. Defaults to the current iOS version.
+    :param multiarch: The CPU architecture+ABI to generate platform tags for -
+        (the value used by `sys.implementation._multiarch` e.g.,
+        `arm64_iphoneos` or `x84_64_iphonesimulator`). Defaults to the current
+        multiarch value.
+    """
+    if version is None:
+        # if iOS is the current platform, ios_ver *must* be defined. However,
+        # it won't exist for CPython versions before 3.13, which causes a mypy
+        # error.
+        _, release, _, _ = platform.ios_ver()  # type: ignore[attr-defined, unused-ignore]
+        version = cast("AppleVersion", tuple(map(int, release.split(".")[:2])))
+
+    if multiarch is None:
+        multiarch = sys.implementation._multiarch
+    multiarch = multiarch.replace("-", "_")
+
+    ios_platform_template = "ios_{major}_{minor}_{multiarch}"
+
+    # Consider any iOS major.minor version from the version requested, down to
+    # 12.0. 12.0 is the first iOS version that is known to have enough features
+    # to support CPython. Consider every possible minor release up to X.9. There
+    # highest the minor has ever gone is 8 (14.8 and 15.8) but having some extra
+    # candidates that won't ever match doesn't really hurt, and it saves us from
+    # having to keep an explicit list of known iOS versions in the code. Return
+    # the results descending order of version number.
+
+    # If the requested major version is less than 12, there won't be any matches.
+    if version[0] < 12:
+        return
+
+    # Consider the actual X.Y version that was requested.
+    yield ios_platform_template.format(
+        major=version[0], minor=version[1], multiarch=multiarch
+    )
+
+    # Consider every minor version from X.0 to the minor version prior to the
+    # version requested by the platform.
+    for minor in range(version[1] - 1, -1, -1):
+        yield ios_platform_template.format(
+            major=version[0], minor=minor, multiarch=multiarch
+        )
+
+    for major in range(version[0] - 1, 11, -1):
+        for minor in range(9, -1, -1):
+            yield ios_platform_template.format(
+                major=major, minor=minor, multiarch=multiarch
+            )
+
+
+def _linux_platforms(is_32bit: bool = _32_BIT_INTERPRETER) -> Iterator[str]:
+    linux = _normalize_string(sysconfig.get_platform())
+    if not linux.startswith("linux_"):
+        # we should never be here, just yield the sysconfig one and return
+        yield linux
+        return
+    if is_32bit:
+        if linux == "linux_x86_64":
+            linux = "linux_i686"
+        elif linux == "linux_aarch64":
+            linux = "linux_armv8l"
+    _, arch = linux.split("_", 1)
+    archs = {"armv8l": ["armv8l", "armv7l"]}.get(arch, [arch])
+    yield from _manylinux.platform_tags(archs)
+    yield from _musllinux.platform_tags(archs)
+    for arch in archs:
+        yield f"linux_{arch}"
+
+
+def _generic_platforms() -> Iterator[str]:
+    yield _normalize_string(sysconfig.get_platform())
+
+
+def platform_tags() -> Iterator[str]:
+    """
+    Provides the platform tags for this installation.
+    """
+    if platform.system() == "Darwin":
+        return mac_platforms()
+    elif platform.system() == "iOS":
+        return ios_platforms()
+    elif platform.system() == "Linux":
+        return _linux_platforms()
+    else:
+        return _generic_platforms()
+
+
+def interpreter_name() -> str:
+    """
+    Returns the name of the running interpreter.
+
+    Some implementations have a reserved, two-letter abbreviation which will
+    be returned when appropriate.
+    """
+    name = sys.implementation.name
+    return INTERPRETER_SHORT_NAMES.get(name) or name
+
+
+def interpreter_version(*, warn: bool = False) -> str:
+    """
+    Returns the version of the running interpreter.
+    """
+    version = _get_config_var("py_version_nodot", warn=warn)
+    if version:
+        version = str(version)
+    else:
+        version = _version_nodot(sys.version_info[:2])
+    return version
+
+
+def _version_nodot(version: PythonVersion) -> str:
+    return "".join(map(str, version))
+
+
+def sys_tags(*, warn: bool = False) -> Iterator[Tag]:
+    """
+    Returns the sequence of tag triples for the running interpreter.
+
+    The order of the sequence corresponds to priority order for the
+    interpreter, from most to least important.
+    """
+
+    interp_name = interpreter_name()
+    if interp_name == "cp":
+        yield from cpython_tags(warn=warn)
+    else:
+        yield from generic_tags()
+
+    if interp_name == "pp":
+        interp = "pp3"
+    elif interp_name == "cp":
+        interp = "cp" + interpreter_version(warn=warn)
+    else:
+        interp = None
+    yield from compatible_tags(interpreter=interp)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/utils.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/utils.py
new file mode 100644
index 00000000..23450953
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/utils.py
@@ -0,0 +1,163 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import functools
+import re
+from typing import NewType, Tuple, Union, cast
+
+from .tags import Tag, parse_tag
+from .version import InvalidVersion, Version, _TrimmedRelease
+
+BuildTag = Union[Tuple[()], Tuple[int, str]]
+NormalizedName = NewType("NormalizedName", str)
+
+
+class InvalidName(ValueError):
+    """
+    An invalid distribution name; users should refer to the packaging user guide.
+    """
+
+
+class InvalidWheelFilename(ValueError):
+    """
+    An invalid wheel filename was found, users should refer to PEP 427.
+    """
+
+
+class InvalidSdistFilename(ValueError):
+    """
+    An invalid sdist filename was found, users should refer to the packaging user guide.
+    """
+
+
+# Core metadata spec for `Name`
+_validate_regex = re.compile(
+    r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.IGNORECASE
+)
+_canonicalize_regex = re.compile(r"[-_.]+")
+_normalized_regex = re.compile(r"^([a-z0-9]|[a-z0-9]([a-z0-9-](?!--))*[a-z0-9])$")
+# PEP 427: The build number must start with a digit.
+_build_tag_regex = re.compile(r"(\d+)(.*)")
+
+
+def canonicalize_name(name: str, *, validate: bool = False) -> NormalizedName:
+    if validate and not _validate_regex.match(name):
+        raise InvalidName(f"name is invalid: {name!r}")
+    # This is taken from PEP 503.
+    value = _canonicalize_regex.sub("-", name).lower()
+    return cast(NormalizedName, value)
+
+
+def is_normalized_name(name: str) -> bool:
+    return _normalized_regex.match(name) is not None
+
+
+@functools.singledispatch
+def canonicalize_version(
+    version: Version | str, *, strip_trailing_zero: bool = True
+) -> str:
+    """
+    Return a canonical form of a version as a string.
+
+    >>> canonicalize_version('1.0.1')
+    '1.0.1'
+
+    Per PEP 625, versions may have multiple canonical forms, differing
+    only by trailing zeros.
+
+    >>> canonicalize_version('1.0.0')
+    '1'
+    >>> canonicalize_version('1.0.0', strip_trailing_zero=False)
+    '1.0.0'
+
+    Invalid versions are returned unaltered.
+
+    >>> canonicalize_version('foo bar baz')
+    'foo bar baz'
+    """
+    return str(_TrimmedRelease(str(version)) if strip_trailing_zero else version)
+
+
+@canonicalize_version.register
+def _(version: str, *, strip_trailing_zero: bool = True) -> str:
+    try:
+        parsed = Version(version)
+    except InvalidVersion:
+        # Legacy versions cannot be normalized
+        return version
+    return canonicalize_version(parsed, strip_trailing_zero=strip_trailing_zero)
+
+
+def parse_wheel_filename(
+    filename: str,
+) -> tuple[NormalizedName, Version, BuildTag, frozenset[Tag]]:
+    if not filename.endswith(".whl"):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (extension must be '.whl'): {filename!r}"
+        )
+
+    filename = filename[:-4]
+    dashes = filename.count("-")
+    if dashes not in (4, 5):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (wrong number of parts): {filename!r}"
+        )
+
+    parts = filename.split("-", dashes - 2)
+    name_part = parts[0]
+    # See PEP 427 for the rules on escaping the project name.
+    if "__" in name_part or re.match(r"^[\w\d._]*$", name_part, re.UNICODE) is None:
+        raise InvalidWheelFilename(f"Invalid project name: {filename!r}")
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(parts[1])
+    except InvalidVersion as e:
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (invalid version): {filename!r}"
+        ) from e
+
+    if dashes == 5:
+        build_part = parts[2]
+        build_match = _build_tag_regex.match(build_part)
+        if build_match is None:
+            raise InvalidWheelFilename(
+                f"Invalid build number: {build_part} in {filename!r}"
+            )
+        build = cast(BuildTag, (int(build_match.group(1)), build_match.group(2)))
+    else:
+        build = ()
+    tags = parse_tag(parts[-1])
+    return (name, version, build, tags)
+
+
+def parse_sdist_filename(filename: str) -> tuple[NormalizedName, Version]:
+    if filename.endswith(".tar.gz"):
+        file_stem = filename[: -len(".tar.gz")]
+    elif filename.endswith(".zip"):
+        file_stem = filename[: -len(".zip")]
+    else:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (extension must be '.tar.gz' or '.zip'):"
+            f" {filename!r}"
+        )
+
+    # We are requiring a PEP 440 version, which cannot contain dashes,
+    # so we split on the last dash.
+    name_part, sep, version_part = file_stem.rpartition("-")
+    if not sep:
+        raise InvalidSdistFilename(f"Invalid sdist filename: {filename!r}")
+
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(version_part)
+    except InvalidVersion as e:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (invalid version): {filename!r}"
+        ) from e
+
+    return (name, version)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/version.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/version.py
new file mode 100644
index 00000000..21f44ca0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/packaging/version.py
@@ -0,0 +1,582 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from pip._vendor.packaging.version import parse, Version
+"""
+
+from __future__ import annotations
+
+import itertools
+import re
+from typing import Any, Callable, NamedTuple, SupportsInt, Tuple, Union
+
+from ._structures import Infinity, InfinityType, NegativeInfinity, NegativeInfinityType
+
+__all__ = ["VERSION_PATTERN", "InvalidVersion", "Version", "parse"]
+
+LocalType = Tuple[Union[int, str], ...]
+
+CmpPrePostDevType = Union[InfinityType, NegativeInfinityType, Tuple[str, int]]
+CmpLocalType = Union[
+    NegativeInfinityType,
+    Tuple[Union[Tuple[int, str], Tuple[NegativeInfinityType, Union[int, str]]], ...],
+]
+CmpKey = Tuple[
+    int,
+    Tuple[int, ...],
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpLocalType,
+]
+VersionComparisonMethod = Callable[[CmpKey, CmpKey], bool]
+
+
+class _Version(NamedTuple):
+    epoch: int
+    release: tuple[int, ...]
+    dev: tuple[str, int] | None
+    pre: tuple[str, int] | None
+    post: tuple[str, int] | None
+    local: LocalType | None
+
+
+def parse(version: str) -> Version:
+    """Parse the given version string.
+
+    >>> parse('1.0.dev1')
+    <Version('1.0.dev1')>
+
+    :param version: The version string to parse.
+    :raises InvalidVersion: When the version string is not a valid version.
+    """
+    return Version(version)
+
+
+class InvalidVersion(ValueError):
+    """Raised when a version string is not a valid version.
+
+    >>> Version("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.version.InvalidVersion: Invalid version: 'invalid'
+    """
+
+
+class _BaseVersion:
+    _key: tuple[Any, ...]
+
+    def __hash__(self) -> int:
+        return hash(self._key)
+
+    # Please keep the duplicated `isinstance` check
+    # in the six comparisons hereunder
+    # unless you find a way to avoid adding overhead function calls.
+    def __lt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key < other._key
+
+    def __le__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key <= other._key
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key == other._key
+
+    def __ge__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key >= other._key
+
+    def __gt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key > other._key
+
+    def __ne__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key != other._key
+
+
+# Deliberately not anchored to the start and end of the string, to make it
+# easier for 3rd party code to reuse
+_VERSION_PATTERN = r"""
+    v?
+    (?:
+        (?:(?P<epoch>[0-9]+)!)?                           # epoch
+        (?P<release>[0-9]+(?:\.[0-9]+)*)                  # release segment
+        (?P<pre>                                          # pre-release
+            [-_\.]?
+            (?P<pre_l>alpha|a|beta|b|preview|pre|c|rc)
+            [-_\.]?
+            (?P<pre_n>[0-9]+)?
+        )?
+        (?P<post>                                         # post release
+            (?:-(?P<post_n1>[0-9]+))
+            |
+            (?:
+                [-_\.]?
+                (?P<post_l>post|rev|r)
+                [-_\.]?
+                (?P<post_n2>[0-9]+)?
+            )
+        )?
+        (?P<dev>                                          # dev release
+            [-_\.]?
+            (?P<dev_l>dev)
+            [-_\.]?
+            (?P<dev_n>[0-9]+)?
+        )?
+    )
+    (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
+"""
+
+VERSION_PATTERN = _VERSION_PATTERN
+"""
+A string containing the regular expression used to match a valid version.
+
+The pattern is not anchored at either end, and is intended for embedding in larger
+expressions (for example, matching a version number as part of a file name). The
+regular expression should be compiled with the ``re.VERBOSE`` and ``re.IGNORECASE``
+flags set.
+
+:meta hide-value:
+"""
+
+
+class Version(_BaseVersion):
+    """This class abstracts handling of a project's versions.
+
+    A :class:`Version` instance is comparison aware and can be compared and
+    sorted using the standard Python interfaces.
+
+    >>> v1 = Version("1.0a5")
+    >>> v2 = Version("1.0")
+    >>> v1
+    <Version('1.0a5')>
+    >>> v2
+    <Version('1.0')>
+    >>> v1 < v2
+    True
+    >>> v1 == v2
+    False
+    >>> v1 > v2
+    False
+    >>> v1 >= v2
+    False
+    >>> v1 <= v2
+    True
+    """
+
+    _regex = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.VERBOSE | re.IGNORECASE)
+    _key: CmpKey
+
+    def __init__(self, version: str) -> None:
+        """Initialize a Version object.
+
+        :param version:
+            The string representation of a version which will be parsed and normalized
+            before use.
+        :raises InvalidVersion:
+            If the ``version`` does not conform to PEP 440 in any way then this
+            exception will be raised.
+        """
+
+        # Validate the version and parse it into pieces
+        match = self._regex.search(version)
+        if not match:
+            raise InvalidVersion(f"Invalid version: {version!r}")
+
+        # Store the parsed out pieces of the version
+        self._version = _Version(
+            epoch=int(match.group("epoch")) if match.group("epoch") else 0,
+            release=tuple(int(i) for i in match.group("release").split(".")),
+            pre=_parse_letter_version(match.group("pre_l"), match.group("pre_n")),
+            post=_parse_letter_version(
+                match.group("post_l"), match.group("post_n1") or match.group("post_n2")
+            ),
+            dev=_parse_letter_version(match.group("dev_l"), match.group("dev_n")),
+            local=_parse_local_version(match.group("local")),
+        )
+
+        # Generate a key which will be used for sorting
+        self._key = _cmpkey(
+            self._version.epoch,
+            self._version.release,
+            self._version.pre,
+            self._version.post,
+            self._version.dev,
+            self._version.local,
+        )
+
+    def __repr__(self) -> str:
+        """A representation of the Version that shows all internal state.
+
+        >>> Version('1.0.0')
+        <Version('1.0.0')>
+        """
+        return f"<Version('{self}')>"
+
+    def __str__(self) -> str:
+        """A string representation of the version that can be round-tripped.
+
+        >>> str(Version("1.0a5"))
+        '1.0a5'
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        # Pre-release
+        if self.pre is not None:
+            parts.append("".join(str(x) for x in self.pre))
+
+        # Post-release
+        if self.post is not None:
+            parts.append(f".post{self.post}")
+
+        # Development release
+        if self.dev is not None:
+            parts.append(f".dev{self.dev}")
+
+        # Local version segment
+        if self.local is not None:
+            parts.append(f"+{self.local}")
+
+        return "".join(parts)
+
+    @property
+    def epoch(self) -> int:
+        """The epoch of the version.
+
+        >>> Version("2.0.0").epoch
+        0
+        >>> Version("1!2.0.0").epoch
+        1
+        """
+        return self._version.epoch
+
+    @property
+    def release(self) -> tuple[int, ...]:
+        """The components of the "release" segment of the version.
+
+        >>> Version("1.2.3").release
+        (1, 2, 3)
+        >>> Version("2.0.0").release
+        (2, 0, 0)
+        >>> Version("1!2.0.0.post0").release
+        (2, 0, 0)
+
+        Includes trailing zeroes but not the epoch or any pre-release / development /
+        post-release suffixes.
+        """
+        return self._version.release
+
+    @property
+    def pre(self) -> tuple[str, int] | None:
+        """The pre-release segment of the version.
+
+        >>> print(Version("1.2.3").pre)
+        None
+        >>> Version("1.2.3a1").pre
+        ('a', 1)
+        >>> Version("1.2.3b1").pre
+        ('b', 1)
+        >>> Version("1.2.3rc1").pre
+        ('rc', 1)
+        """
+        return self._version.pre
+
+    @property
+    def post(self) -> int | None:
+        """The post-release number of the version.
+
+        >>> print(Version("1.2.3").post)
+        None
+        >>> Version("1.2.3.post1").post
+        1
+        """
+        return self._version.post[1] if self._version.post else None
+
+    @property
+    def dev(self) -> int | None:
+        """The development number of the version.
+
+        >>> print(Version("1.2.3").dev)
+        None
+        >>> Version("1.2.3.dev1").dev
+        1
+        """
+        return self._version.dev[1] if self._version.dev else None
+
+    @property
+    def local(self) -> str | None:
+        """The local version segment of the version.
+
+        >>> print(Version("1.2.3").local)
+        None
+        >>> Version("1.2.3+abc").local
+        'abc'
+        """
+        if self._version.local:
+            return ".".join(str(x) for x in self._version.local)
+        else:
+            return None
+
+    @property
+    def public(self) -> str:
+        """The public portion of the version.
+
+        >>> Version("1.2.3").public
+        '1.2.3'
+        >>> Version("1.2.3+abc").public
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").public
+        '1!1.2.3.dev1'
+        """
+        return str(self).split("+", 1)[0]
+
+    @property
+    def base_version(self) -> str:
+        """The "base version" of the version.
+
+        >>> Version("1.2.3").base_version
+        '1.2.3'
+        >>> Version("1.2.3+abc").base_version
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").base_version
+        '1!1.2.3'
+
+        The "base version" is the public version of the project without any pre or post
+        release markers.
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        return "".join(parts)
+
+    @property
+    def is_prerelease(self) -> bool:
+        """Whether this version is a pre-release.
+
+        >>> Version("1.2.3").is_prerelease
+        False
+        >>> Version("1.2.3a1").is_prerelease
+        True
+        >>> Version("1.2.3b1").is_prerelease
+        True
+        >>> Version("1.2.3rc1").is_prerelease
+        True
+        >>> Version("1.2.3dev1").is_prerelease
+        True
+        """
+        return self.dev is not None or self.pre is not None
+
+    @property
+    def is_postrelease(self) -> bool:
+        """Whether this version is a post-release.
+
+        >>> Version("1.2.3").is_postrelease
+        False
+        >>> Version("1.2.3.post1").is_postrelease
+        True
+        """
+        return self.post is not None
+
+    @property
+    def is_devrelease(self) -> bool:
+        """Whether this version is a development release.
+
+        >>> Version("1.2.3").is_devrelease
+        False
+        >>> Version("1.2.3.dev1").is_devrelease
+        True
+        """
+        return self.dev is not None
+
+    @property
+    def major(self) -> int:
+        """The first item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").major
+        1
+        """
+        return self.release[0] if len(self.release) >= 1 else 0
+
+    @property
+    def minor(self) -> int:
+        """The second item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").minor
+        2
+        >>> Version("1").minor
+        0
+        """
+        return self.release[1] if len(self.release) >= 2 else 0
+
+    @property
+    def micro(self) -> int:
+        """The third item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").micro
+        3
+        >>> Version("1").micro
+        0
+        """
+        return self.release[2] if len(self.release) >= 3 else 0
+
+
+class _TrimmedRelease(Version):
+    @property
+    def release(self) -> tuple[int, ...]:
+        """
+        Release segment without any trailing zeros.
+
+        >>> _TrimmedRelease('1.0.0').release
+        (1,)
+        >>> _TrimmedRelease('0.0').release
+        (0,)
+        """
+        rel = super().release
+        nonzeros = (index for index, val in enumerate(rel) if val)
+        last_nonzero = max(nonzeros, default=0)
+        return rel[: last_nonzero + 1]
+
+
+def _parse_letter_version(
+    letter: str | None, number: str | bytes | SupportsInt | None
+) -> tuple[str, int] | None:
+    if letter:
+        # We consider there to be an implicit 0 in a pre-release if there is
+        # not a numeral associated with it.
+        if number is None:
+            number = 0
+
+        # We normalize any letters to their lower case form
+        letter = letter.lower()
+
+        # We consider some words to be alternate spellings of other words and
+        # in those cases we want to normalize the spellings to our preferred
+        # spelling.
+        if letter == "alpha":
+            letter = "a"
+        elif letter == "beta":
+            letter = "b"
+        elif letter in ["c", "pre", "preview"]:
+            letter = "rc"
+        elif letter in ["rev", "r"]:
+            letter = "post"
+
+        return letter, int(number)
+
+    assert not letter
+    if number:
+        # We assume if we are given a number, but we are not given a letter
+        # then this is using the implicit post release syntax (e.g. 1.0-1)
+        letter = "post"
+
+        return letter, int(number)
+
+    return None
+
+
+_local_version_separators = re.compile(r"[\._-]")
+
+
+def _parse_local_version(local: str | None) -> LocalType | None:
+    """
+    Takes a string like abc.1.twelve and turns it into ("abc", 1, "twelve").
+    """
+    if local is not None:
+        return tuple(
+            part.lower() if not part.isdigit() else int(part)
+            for part in _local_version_separators.split(local)
+        )
+    return None
+
+
+def _cmpkey(
+    epoch: int,
+    release: tuple[int, ...],
+    pre: tuple[str, int] | None,
+    post: tuple[str, int] | None,
+    dev: tuple[str, int] | None,
+    local: LocalType | None,
+) -> CmpKey:
+    # When we compare a release version, we want to compare it with all of the
+    # trailing zeros removed. So we'll use a reverse the list, drop all the now
+    # leading zeros until we come to something non zero, then take the rest
+    # re-reverse it back into the correct order and make it a tuple and use
+    # that for our sorting key.
+    _release = tuple(
+        reversed(list(itertools.dropwhile(lambda x: x == 0, reversed(release))))
+    )
+
+    # We need to "trick" the sorting algorithm to put 1.0.dev0 before 1.0a0.
+    # We'll do this by abusing the pre segment, but we _only_ want to do this
+    # if there is not a pre or a post segment. If we have one of those then
+    # the normal sorting rules will handle this case correctly.
+    if pre is None and post is None and dev is not None:
+        _pre: CmpPrePostDevType = NegativeInfinity
+    # Versions without a pre-release (except as noted above) should sort after
+    # those with one.
+    elif pre is None:
+        _pre = Infinity
+    else:
+        _pre = pre
+
+    # Versions without a post segment should sort before those with one.
+    if post is None:
+        _post: CmpPrePostDevType = NegativeInfinity
+
+    else:
+        _post = post
+
+    # Versions without a development segment should sort after those with one.
+    if dev is None:
+        _dev: CmpPrePostDevType = Infinity
+
+    else:
+        _dev = dev
+
+    if local is None:
+        # Versions without a local segment should sort before those with one.
+        _local: CmpLocalType = NegativeInfinity
+    else:
+        # Versions with a local segment need that segment parsed to implement
+        # the sorting rules in PEP440.
+        # - Alpha numeric segments sort before numeric segments
+        # - Alpha numeric segments sort lexicographically
+        # - Numeric segments sort numerically
+        # - Shorter versions sort before longer versions when the prefixes
+        #   match exactly
+        _local = tuple(
+            (i, "") if isinstance(i, int) else (NegativeInfinity, i) for i in local
+        )
+
+    return epoch, _release, _pre, _post, _dev, _local
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__init__.py
new file mode 100644
index 00000000..57ce7f10
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__init__.py
@@ -0,0 +1,3676 @@
+# TODO: Add Generic type annotations to initialized collections.
+# For now we'd simply use implicit Any/Unknown which would add redundant annotations
+# mypy: disable-error-code="var-annotated"
+"""
+Package resource API
+--------------------
+
+A resource is a logical file contained within a package, or a logical
+subdirectory thereof.  The package resource API expects resource names
+to have their path parts separated with ``/``, *not* whatever the local
+path separator is.  Do not use os.path operations to manipulate resource
+names being passed into the API.
+
+The package resource API is designed to work with normal filesystem packages,
+.egg files, and unpacked .egg files.  It can also work in a limited way with
+.zip files and with custom PEP 302 loaders that support the ``get_data()``
+method.
+
+This module is deprecated. Users are directed to :mod:`importlib.resources`,
+:mod:`importlib.metadata` and :pypi:`packaging` instead.
+"""
+
+from __future__ import annotations
+
+import sys
+
+if sys.version_info < (3, 8):  # noqa: UP036 # Check for unsupported versions
+    raise RuntimeError("Python 3.8 or later is required")
+
+import os
+import io
+import time
+import re
+import types
+from typing import (
+    Any,
+    Literal,
+    Dict,
+    Iterator,
+    Mapping,
+    MutableSequence,
+    NamedTuple,
+    NoReturn,
+    Tuple,
+    Union,
+    TYPE_CHECKING,
+    Protocol,
+    Callable,
+    Iterable,
+    TypeVar,
+    overload,
+)
+import zipfile
+import zipimport
+import warnings
+import stat
+import functools
+import pkgutil
+import operator
+import platform
+import collections
+import plistlib
+import email.parser
+import errno
+import tempfile
+import textwrap
+import inspect
+import ntpath
+import posixpath
+import importlib
+import importlib.abc
+import importlib.machinery
+from pkgutil import get_importer
+
+import _imp
+
+# capture these to bypass sandboxing
+from os import utime
+from os import open as os_open
+from os.path import isdir, split
+
+try:
+    from os import mkdir, rename, unlink
+
+    WRITE_SUPPORT = True
+except ImportError:
+    # no write support, probably under GAE
+    WRITE_SUPPORT = False
+
+from pip._internal.utils._jaraco_text import (
+    yield_lines,
+    drop_comment,
+    join_continuation,
+)
+from pip._vendor.packaging import markers as _packaging_markers
+from pip._vendor.packaging import requirements as _packaging_requirements
+from pip._vendor.packaging import utils as _packaging_utils
+from pip._vendor.packaging import version as _packaging_version
+from pip._vendor.platformdirs import user_cache_dir as _user_cache_dir
+
+if TYPE_CHECKING:
+    from _typeshed import BytesPath, StrPath, StrOrBytesPath
+    from pip._vendor.typing_extensions import Self
+
+
+# Patch: Remove deprecation warning from vendored pkg_resources.
+# Setting PYTHONWARNINGS=error to verify builds produce no warnings
+# causes immediate exceptions.
+# See https://github.com/pypa/pip/issues/12243
+
+
+_T = TypeVar("_T")
+_DistributionT = TypeVar("_DistributionT", bound="Distribution")
+# Type aliases
+_NestedStr = Union[str, Iterable[Union[str, Iterable["_NestedStr"]]]]
+_InstallerTypeT = Callable[["Requirement"], "_DistributionT"]
+_InstallerType = Callable[["Requirement"], Union["Distribution", None]]
+_PkgReqType = Union[str, "Requirement"]
+_EPDistType = Union["Distribution", _PkgReqType]
+_MetadataType = Union["IResourceProvider", None]
+_ResolvedEntryPoint = Any  # Can be any attribute in the module
+_ResourceStream = Any  # TODO / Incomplete: A readable file-like object
+# Any object works, but let's indicate we expect something like a module (optionally has __loader__ or __file__)
+_ModuleLike = Union[object, types.ModuleType]
+# Any: Should be _ModuleLike but we end up with issues where _ModuleLike doesn't have _ZipLoaderModule's __loader__
+_ProviderFactoryType = Callable[[Any], "IResourceProvider"]
+_DistFinderType = Callable[[_T, str, bool], Iterable["Distribution"]]
+_NSHandlerType = Callable[[_T, str, str, types.ModuleType], Union[str, None]]
+_AdapterT = TypeVar(
+    "_AdapterT", _DistFinderType[Any], _ProviderFactoryType, _NSHandlerType[Any]
+)
+
+
+# Use _typeshed.importlib.LoaderProtocol once available https://github.com/python/typeshed/pull/11890
+class _LoaderProtocol(Protocol):
+    def load_module(self, fullname: str, /) -> types.ModuleType: ...
+
+
+class _ZipLoaderModule(Protocol):
+    __loader__: zipimport.zipimporter
+
+
+_PEP440_FALLBACK = re.compile(r"^v?(?P<safe>(?:[0-9]+!)?[0-9]+(?:\.[0-9]+)*)", re.I)
+
+
+class PEP440Warning(RuntimeWarning):
+    """
+    Used when there is an issue with a version or specifier not complying with
+    PEP 440.
+    """
+
+
+parse_version = _packaging_version.Version
+
+
+_state_vars: dict[str, str] = {}
+
+
+def _declare_state(vartype: str, varname: str, initial_value: _T) -> _T:
+    _state_vars[varname] = vartype
+    return initial_value
+
+
+def __getstate__() -> dict[str, Any]:
+    state = {}
+    g = globals()
+    for k, v in _state_vars.items():
+        state[k] = g['_sget_' + v](g[k])
+    return state
+
+
+def __setstate__(state: dict[str, Any]) -> dict[str, Any]:
+    g = globals()
+    for k, v in state.items():
+        g['_sset_' + _state_vars[k]](k, g[k], v)
+    return state
+
+
+def _sget_dict(val):
+    return val.copy()
+
+
+def _sset_dict(key, ob, state):
+    ob.clear()
+    ob.update(state)
+
+
+def _sget_object(val):
+    return val.__getstate__()
+
+
+def _sset_object(key, ob, state):
+    ob.__setstate__(state)
+
+
+_sget_none = _sset_none = lambda *args: None
+
+
+def get_supported_platform():
+    """Return this platform's maximum compatible version.
+
+    distutils.util.get_platform() normally reports the minimum version
+    of macOS that would be required to *use* extensions produced by
+    distutils.  But what we want when checking compatibility is to know the
+    version of macOS that we are *running*.  To allow usage of packages that
+    explicitly require a newer version of macOS, we must also know the
+    current version of the OS.
+
+    If this condition occurs for any other platform with a version in its
+    platform strings, this function should be extended accordingly.
+    """
+    plat = get_build_platform()
+    m = macosVersionString.match(plat)
+    if m is not None and sys.platform == "darwin":
+        try:
+            plat = 'macosx-%s-%s' % ('.'.join(_macos_vers()[:2]), m.group(3))
+        except ValueError:
+            # not macOS
+            pass
+    return plat
+
+
+__all__ = [
+    # Basic resource access and distribution/entry point discovery
+    'require',
+    'run_script',
+    'get_provider',
+    'get_distribution',
+    'load_entry_point',
+    'get_entry_map',
+    'get_entry_info',
+    'iter_entry_points',
+    'resource_string',
+    'resource_stream',
+    'resource_filename',
+    'resource_listdir',
+    'resource_exists',
+    'resource_isdir',
+    # Environmental control
+    'declare_namespace',
+    'working_set',
+    'add_activation_listener',
+    'find_distributions',
+    'set_extraction_path',
+    'cleanup_resources',
+    'get_default_cache',
+    # Primary implementation classes
+    'Environment',
+    'WorkingSet',
+    'ResourceManager',
+    'Distribution',
+    'Requirement',
+    'EntryPoint',
+    # Exceptions
+    'ResolutionError',
+    'VersionConflict',
+    'DistributionNotFound',
+    'UnknownExtra',
+    'ExtractionError',
+    # Warnings
+    'PEP440Warning',
+    # Parsing functions and string utilities
+    'parse_requirements',
+    'parse_version',
+    'safe_name',
+    'safe_version',
+    'get_platform',
+    'compatible_platforms',
+    'yield_lines',
+    'split_sections',
+    'safe_extra',
+    'to_filename',
+    'invalid_marker',
+    'evaluate_marker',
+    # filesystem utilities
+    'ensure_directory',
+    'normalize_path',
+    # Distribution "precedence" constants
+    'EGG_DIST',
+    'BINARY_DIST',
+    'SOURCE_DIST',
+    'CHECKOUT_DIST',
+    'DEVELOP_DIST',
+    # "Provider" interfaces, implementations, and registration/lookup APIs
+    'IMetadataProvider',
+    'IResourceProvider',
+    'FileMetadata',
+    'PathMetadata',
+    'EggMetadata',
+    'EmptyProvider',
+    'empty_provider',
+    'NullProvider',
+    'EggProvider',
+    'DefaultProvider',
+    'ZipProvider',
+    'register_finder',
+    'register_namespace_handler',
+    'register_loader_type',
+    'fixup_namespace_packages',
+    'get_importer',
+    # Warnings
+    'PkgResourcesDeprecationWarning',
+    # Deprecated/backward compatibility only
+    'run_main',
+    'AvailableDistributions',
+]
+
+
+class ResolutionError(Exception):
+    """Abstract base for dependency resolution errors"""
+
+    def __repr__(self):
+        return self.__class__.__name__ + repr(self.args)
+
+
+class VersionConflict(ResolutionError):
+    """
+    An already-installed version conflicts with the requested version.
+
+    Should be initialized with the installed Distribution and the requested
+    Requirement.
+    """
+
+    _template = "{self.dist} is installed but {self.req} is required"
+
+    @property
+    def dist(self) -> Distribution:
+        return self.args[0]
+
+    @property
+    def req(self) -> Requirement:
+        return self.args[1]
+
+    def report(self):
+        return self._template.format(**locals())
+
+    def with_context(self, required_by: set[Distribution | str]):
+        """
+        If required_by is non-empty, return a version of self that is a
+        ContextualVersionConflict.
+        """
+        if not required_by:
+            return self
+        args = self.args + (required_by,)
+        return ContextualVersionConflict(*args)
+
+
+class ContextualVersionConflict(VersionConflict):
+    """
+    A VersionConflict that accepts a third parameter, the set of the
+    requirements that required the installed Distribution.
+    """
+
+    _template = VersionConflict._template + ' by {self.required_by}'
+
+    @property
+    def required_by(self) -> set[str]:
+        return self.args[2]
+
+
+class DistributionNotFound(ResolutionError):
+    """A requested distribution was not found"""
+
+    _template = (
+        "The '{self.req}' distribution was not found "
+        "and is required by {self.requirers_str}"
+    )
+
+    @property
+    def req(self) -> Requirement:
+        return self.args[0]
+
+    @property
+    def requirers(self) -> set[str] | None:
+        return self.args[1]
+
+    @property
+    def requirers_str(self):
+        if not self.requirers:
+            return 'the application'
+        return ', '.join(self.requirers)
+
+    def report(self):
+        return self._template.format(**locals())
+
+    def __str__(self):
+        return self.report()
+
+
+class UnknownExtra(ResolutionError):
+    """Distribution doesn't have an "extra feature" of the given name"""
+
+
+_provider_factories: dict[type[_ModuleLike], _ProviderFactoryType] = {}
+
+PY_MAJOR = '{}.{}'.format(*sys.version_info)
+EGG_DIST = 3
+BINARY_DIST = 2
+SOURCE_DIST = 1
+CHECKOUT_DIST = 0
+DEVELOP_DIST = -1
+
+
+def register_loader_type(
+    loader_type: type[_ModuleLike], provider_factory: _ProviderFactoryType
+):
+    """Register `provider_factory` to make providers for `loader_type`
+
+    `loader_type` is the type or class of a PEP 302 ``module.__loader__``,
+    and `provider_factory` is a function that, passed a *module* object,
+    returns an ``IResourceProvider`` for that module.
+    """
+    _provider_factories[loader_type] = provider_factory
+
+
+@overload
+def get_provider(moduleOrReq: str) -> IResourceProvider: ...
+@overload
+def get_provider(moduleOrReq: Requirement) -> Distribution: ...
+def get_provider(moduleOrReq: str | Requirement) -> IResourceProvider | Distribution:
+    """Return an IResourceProvider for the named module or requirement"""
+    if isinstance(moduleOrReq, Requirement):
+        return working_set.find(moduleOrReq) or require(str(moduleOrReq))[0]
+    try:
+        module = sys.modules[moduleOrReq]
+    except KeyError:
+        __import__(moduleOrReq)
+        module = sys.modules[moduleOrReq]
+    loader = getattr(module, '__loader__', None)
+    return _find_adapter(_provider_factories, loader)(module)
+
+
+@functools.lru_cache(maxsize=None)
+def _macos_vers():
+    version = platform.mac_ver()[0]
+    # fallback for MacPorts
+    if version == '':
+        plist = '/System/Library/CoreServices/SystemVersion.plist'
+        if os.path.exists(plist):
+            with open(plist, 'rb') as fh:
+                plist_content = plistlib.load(fh)
+            if 'ProductVersion' in plist_content:
+                version = plist_content['ProductVersion']
+    return version.split('.')
+
+
+def _macos_arch(machine):
+    return {'PowerPC': 'ppc', 'Power_Macintosh': 'ppc'}.get(machine, machine)
+
+
+def get_build_platform():
+    """Return this platform's string for platform-specific distributions
+
+    XXX Currently this is the same as ``distutils.util.get_platform()``, but it
+    needs some hacks for Linux and macOS.
+    """
+    from sysconfig import get_platform
+
+    plat = get_platform()
+    if sys.platform == "darwin" and not plat.startswith('macosx-'):
+        try:
+            version = _macos_vers()
+            machine = os.uname()[4].replace(" ", "_")
+            return "macosx-%d.%d-%s" % (
+                int(version[0]),
+                int(version[1]),
+                _macos_arch(machine),
+            )
+        except ValueError:
+            # if someone is running a non-Mac darwin system, this will fall
+            # through to the default implementation
+            pass
+    return plat
+
+
+macosVersionString = re.compile(r"macosx-(\d+)\.(\d+)-(.*)")
+darwinVersionString = re.compile(r"darwin-(\d+)\.(\d+)\.(\d+)-(.*)")
+# XXX backward compat
+get_platform = get_build_platform
+
+
+def compatible_platforms(provided: str | None, required: str | None):
+    """Can code for the `provided` platform run on the `required` platform?
+
+    Returns true if either platform is ``None``, or the platforms are equal.
+
+    XXX Needs compatibility checks for Linux and other unixy OSes.
+    """
+    if provided is None or required is None or provided == required:
+        # easy case
+        return True
+
+    # macOS special cases
+    reqMac = macosVersionString.match(required)
+    if reqMac:
+        provMac = macosVersionString.match(provided)
+
+        # is this a Mac package?
+        if not provMac:
+            # this is backwards compatibility for packages built before
+            # setuptools 0.6. All packages built after this point will
+            # use the new macOS designation.
+            provDarwin = darwinVersionString.match(provided)
+            if provDarwin:
+                dversion = int(provDarwin.group(1))
+                macosversion = "%s.%s" % (reqMac.group(1), reqMac.group(2))
+                if (
+                    dversion == 7
+                    and macosversion >= "10.3"
+                    or dversion == 8
+                    and macosversion >= "10.4"
+                ):
+                    return True
+            # egg isn't macOS or legacy darwin
+            return False
+
+        # are they the same major version and machine type?
+        if provMac.group(1) != reqMac.group(1) or provMac.group(3) != reqMac.group(3):
+            return False
+
+        # is the required OS major update >= the provided one?
+        if int(provMac.group(2)) > int(reqMac.group(2)):
+            return False
+
+        return True
+
+    # XXX Linux and other platforms' special cases should go here
+    return False
+
+
+@overload
+def get_distribution(dist: _DistributionT) -> _DistributionT: ...
+@overload
+def get_distribution(dist: _PkgReqType) -> Distribution: ...
+def get_distribution(dist: Distribution | _PkgReqType) -> Distribution:
+    """Return a current distribution object for a Requirement or string"""
+    if isinstance(dist, str):
+        dist = Requirement.parse(dist)
+    if isinstance(dist, Requirement):
+        # Bad type narrowing, dist has to be a Requirement here, so get_provider has to return Distribution
+        dist = get_provider(dist)  # type: ignore[assignment]
+    if not isinstance(dist, Distribution):
+        raise TypeError("Expected str, Requirement, or Distribution", dist)
+    return dist
+
+
+def load_entry_point(dist: _EPDistType, group: str, name: str) -> _ResolvedEntryPoint:
+    """Return `name` entry point of `group` for `dist` or raise ImportError"""
+    return get_distribution(dist).load_entry_point(group, name)
+
+
+@overload
+def get_entry_map(
+    dist: _EPDistType, group: None = None
+) -> dict[str, dict[str, EntryPoint]]: ...
+@overload
+def get_entry_map(dist: _EPDistType, group: str) -> dict[str, EntryPoint]: ...
+def get_entry_map(dist: _EPDistType, group: str | None = None):
+    """Return the entry point map for `group`, or the full entry map"""
+    return get_distribution(dist).get_entry_map(group)
+
+
+def get_entry_info(dist: _EPDistType, group: str, name: str):
+    """Return the EntryPoint object for `group`+`name`, or ``None``"""
+    return get_distribution(dist).get_entry_info(group, name)
+
+
+class IMetadataProvider(Protocol):
+    def has_metadata(self, name: str) -> bool:
+        """Does the package's distribution contain the named metadata?"""
+
+    def get_metadata(self, name: str) -> str:
+        """The named metadata resource as a string"""
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        """Yield named metadata resource as list of non-blank non-comment lines
+
+        Leading and trailing whitespace is stripped from each line, and lines
+        with ``#`` as the first non-blank character are omitted."""
+
+    def metadata_isdir(self, name: str) -> bool:
+        """Is the named metadata a directory?  (like ``os.path.isdir()``)"""
+
+    def metadata_listdir(self, name: str) -> list[str]:
+        """List of metadata names in the directory (like ``os.listdir()``)"""
+
+    def run_script(self, script_name: str, namespace: dict[str, Any]) -> None:
+        """Execute the named script in the supplied namespace dictionary"""
+
+
+class IResourceProvider(IMetadataProvider, Protocol):
+    """An object that provides access to package resources"""
+
+    def get_resource_filename(
+        self, manager: ResourceManager, resource_name: str
+    ) -> str:
+        """Return a true filesystem path for `resource_name`
+
+        `manager` must be a ``ResourceManager``"""
+
+    def get_resource_stream(
+        self, manager: ResourceManager, resource_name: str
+    ) -> _ResourceStream:
+        """Return a readable file-like object for `resource_name`
+
+        `manager` must be a ``ResourceManager``"""
+
+    def get_resource_string(
+        self, manager: ResourceManager, resource_name: str
+    ) -> bytes:
+        """Return the contents of `resource_name` as :obj:`bytes`
+
+        `manager` must be a ``ResourceManager``"""
+
+    def has_resource(self, resource_name: str) -> bool:
+        """Does the package contain the named resource?"""
+
+    def resource_isdir(self, resource_name: str) -> bool:
+        """Is the named resource a directory?  (like ``os.path.isdir()``)"""
+
+    def resource_listdir(self, resource_name: str) -> list[str]:
+        """List of resource names in the directory (like ``os.listdir()``)"""
+
+
+class WorkingSet:
+    """A collection of active distributions on sys.path (or a similar list)"""
+
+    def __init__(self, entries: Iterable[str] | None = None):
+        """Create working set from list of path entries (default=sys.path)"""
+        self.entries: list[str] = []
+        self.entry_keys = {}
+        self.by_key = {}
+        self.normalized_to_canonical_keys = {}
+        self.callbacks = []
+
+        if entries is None:
+            entries = sys.path
+
+        for entry in entries:
+            self.add_entry(entry)
+
+    @classmethod
+    def _build_master(cls):
+        """
+        Prepare the master working set.
+        """
+        ws = cls()
+        try:
+            from __main__ import __requires__
+        except ImportError:
+            # The main program does not list any requirements
+            return ws
+
+        # ensure the requirements are met
+        try:
+            ws.require(__requires__)
+        except VersionConflict:
+            return cls._build_from_requirements(__requires__)
+
+        return ws
+
+    @classmethod
+    def _build_from_requirements(cls, req_spec):
+        """
+        Build a working set from a requirement spec. Rewrites sys.path.
+        """
+        # try it without defaults already on sys.path
+        # by starting with an empty path
+        ws = cls([])
+        reqs = parse_requirements(req_spec)
+        dists = ws.resolve(reqs, Environment())
+        for dist in dists:
+            ws.add(dist)
+
+        # add any missing entries from sys.path
+        for entry in sys.path:
+            if entry not in ws.entries:
+                ws.add_entry(entry)
+
+        # then copy back to sys.path
+        sys.path[:] = ws.entries
+        return ws
+
+    def add_entry(self, entry: str):
+        """Add a path item to ``.entries``, finding any distributions on it
+
+        ``find_distributions(entry, True)`` is used to find distributions
+        corresponding to the path entry, and they are added.  `entry` is
+        always appended to ``.entries``, even if it is already present.
+        (This is because ``sys.path`` can contain the same value more than
+        once, and the ``.entries`` of the ``sys.path`` WorkingSet should always
+        equal ``sys.path``.)
+        """
+        self.entry_keys.setdefault(entry, [])
+        self.entries.append(entry)
+        for dist in find_distributions(entry, True):
+            self.add(dist, entry, False)
+
+    def __contains__(self, dist: Distribution) -> bool:
+        """True if `dist` is the active distribution for its project"""
+        return self.by_key.get(dist.key) == dist
+
+    def find(self, req: Requirement) -> Distribution | None:
+        """Find a distribution matching requirement `req`
+
+        If there is an active distribution for the requested project, this
+        returns it as long as it meets the version requirement specified by
+        `req`.  But, if there is an active distribution for the project and it
+        does *not* meet the `req` requirement, ``VersionConflict`` is raised.
+        If there is no active distribution for the requested project, ``None``
+        is returned.
+        """
+        dist = self.by_key.get(req.key)
+
+        if dist is None:
+            canonical_key = self.normalized_to_canonical_keys.get(req.key)
+
+            if canonical_key is not None:
+                req.key = canonical_key
+                dist = self.by_key.get(canonical_key)
+
+        if dist is not None and dist not in req:
+            # XXX add more info
+            raise VersionConflict(dist, req)
+        return dist
+
+    def iter_entry_points(self, group: str, name: str | None = None):
+        """Yield entry point objects from `group` matching `name`
+
+        If `name` is None, yields all entry points in `group` from all
+        distributions in the working set, otherwise only ones matching
+        both `group` and `name` are yielded (in distribution order).
+        """
+        return (
+            entry
+            for dist in self
+            for entry in dist.get_entry_map(group).values()
+            if name is None or name == entry.name
+        )
+
+    def run_script(self, requires: str, script_name: str):
+        """Locate distribution for `requires` and run `script_name` script"""
+        ns = sys._getframe(1).f_globals
+        name = ns['__name__']
+        ns.clear()
+        ns['__name__'] = name
+        self.require(requires)[0].run_script(script_name, ns)
+
+    def __iter__(self) -> Iterator[Distribution]:
+        """Yield distributions for non-duplicate projects in the working set
+
+        The yield order is the order in which the items' path entries were
+        added to the working set.
+        """
+        seen = set()
+        for item in self.entries:
+            if item not in self.entry_keys:
+                # workaround a cache issue
+                continue
+
+            for key in self.entry_keys[item]:
+                if key not in seen:
+                    seen.add(key)
+                    yield self.by_key[key]
+
+    def add(
+        self,
+        dist: Distribution,
+        entry: str | None = None,
+        insert: bool = True,
+        replace: bool = False,
+    ):
+        """Add `dist` to working set, associated with `entry`
+
+        If `entry` is unspecified, it defaults to the ``.location`` of `dist`.
+        On exit from this routine, `entry` is added to the end of the working
+        set's ``.entries`` (if it wasn't already present).
+
+        `dist` is only added to the working set if it's for a project that
+        doesn't already have a distribution in the set, unless `replace=True`.
+        If it's added, any callbacks registered with the ``subscribe()`` method
+        will be called.
+        """
+        if insert:
+            dist.insert_on(self.entries, entry, replace=replace)
+
+        if entry is None:
+            entry = dist.location
+        keys = self.entry_keys.setdefault(entry, [])
+        keys2 = self.entry_keys.setdefault(dist.location, [])
+        if not replace and dist.key in self.by_key:
+            # ignore hidden distros
+            return
+
+        self.by_key[dist.key] = dist
+        normalized_name = _packaging_utils.canonicalize_name(dist.key)
+        self.normalized_to_canonical_keys[normalized_name] = dist.key
+        if dist.key not in keys:
+            keys.append(dist.key)
+        if dist.key not in keys2:
+            keys2.append(dist.key)
+        self._added_new(dist)
+
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None,
+        installer: _InstallerTypeT[_DistributionT],
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[_DistributionT]: ...
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        *,
+        installer: _InstallerTypeT[_DistributionT],
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[_DistributionT]: ...
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[Distribution]: ...
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        installer: _InstallerType | None | _InstallerTypeT[_DistributionT] = None,
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[Distribution] | list[_DistributionT]:
+        """List all distributions needed to (recursively) meet `requirements`
+
+        `requirements` must be a sequence of ``Requirement`` objects.  `env`,
+        if supplied, should be an ``Environment`` instance.  If
+        not supplied, it defaults to all distributions available within any
+        entry or distribution in the working set.  `installer`, if supplied,
+        will be invoked with each requirement that cannot be met by an
+        already-installed distribution; it should return a ``Distribution`` or
+        ``None``.
+
+        Unless `replace_conflicting=True`, raises a VersionConflict exception
+        if
+        any requirements are found on the path that have the correct name but
+        the wrong version.  Otherwise, if an `installer` is supplied it will be
+        invoked to obtain the correct version of the requirement and activate
+        it.
+
+        `extras` is a list of the extras to be used with these requirements.
+        This is important because extra requirements may look like `my_req;
+        extra = "my_extra"`, which would otherwise be interpreted as a purely
+        optional requirement.  Instead, we want to be able to assert that these
+        requirements are truly required.
+        """
+
+        # set up the stack
+        requirements = list(requirements)[::-1]
+        # set of processed requirements
+        processed = set()
+        # key -> dist
+        best = {}
+        to_activate = []
+
+        req_extras = _ReqExtras()
+
+        # Mapping of requirement to set of distributions that required it;
+        # useful for reporting info about conflicts.
+        required_by = collections.defaultdict(set)
+
+        while requirements:
+            # process dependencies breadth-first
+            req = requirements.pop(0)
+            if req in processed:
+                # Ignore cyclic or redundant dependencies
+                continue
+
+            if not req_extras.markers_pass(req, extras):
+                continue
+
+            dist = self._resolve_dist(
+                req, best, replace_conflicting, env, installer, required_by, to_activate
+            )
+
+            # push the new requirements onto the stack
+            new_requirements = dist.requires(req.extras)[::-1]
+            requirements.extend(new_requirements)
+
+            # Register the new requirements needed by req
+            for new_requirement in new_requirements:
+                required_by[new_requirement].add(req.project_name)
+                req_extras[new_requirement] = req.extras
+
+            processed.add(req)
+
+        # return list of distros to activate
+        return to_activate
+
+    def _resolve_dist(
+        self, req, best, replace_conflicting, env, installer, required_by, to_activate
+    ) -> Distribution:
+        dist = best.get(req.key)
+        if dist is None:
+            # Find the best distribution and add it to the map
+            dist = self.by_key.get(req.key)
+            if dist is None or (dist not in req and replace_conflicting):
+                ws = self
+                if env is None:
+                    if dist is None:
+                        env = Environment(self.entries)
+                    else:
+                        # Use an empty environment and workingset to avoid
+                        # any further conflicts with the conflicting
+                        # distribution
+                        env = Environment([])
+                        ws = WorkingSet([])
+                dist = best[req.key] = env.best_match(
+                    req, ws, installer, replace_conflicting=replace_conflicting
+                )
+                if dist is None:
+                    requirers = required_by.get(req, None)
+                    raise DistributionNotFound(req, requirers)
+            to_activate.append(dist)
+        if dist not in req:
+            # Oops, the "best" so far conflicts with a dependency
+            dependent_req = required_by[req]
+            raise VersionConflict(dist, req).with_context(dependent_req)
+        return dist
+
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None,
+        installer: _InstallerTypeT[_DistributionT],
+        fallback: bool = True,
+    ) -> tuple[list[_DistributionT], dict[Distribution, Exception]]: ...
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        *,
+        installer: _InstallerTypeT[_DistributionT],
+        fallback: bool = True,
+    ) -> tuple[list[_DistributionT], dict[Distribution, Exception]]: ...
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        installer: _InstallerType | None = None,
+        fallback: bool = True,
+    ) -> tuple[list[Distribution], dict[Distribution, Exception]]: ...
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        installer: _InstallerType | None | _InstallerTypeT[_DistributionT] = None,
+        fallback: bool = True,
+    ) -> tuple[
+        list[Distribution] | list[_DistributionT],
+        dict[Distribution, Exception],
+    ]:
+        """Find all activatable distributions in `plugin_env`
+
+        Example usage::
+
+            distributions, errors = working_set.find_plugins(
+                Environment(plugin_dirlist)
+            )
+            # add plugins+libs to sys.path
+            map(working_set.add, distributions)
+            # display errors
+            print('Could not load', errors)
+
+        The `plugin_env` should be an ``Environment`` instance that contains
+        only distributions that are in the project's "plugin directory" or
+        directories. The `full_env`, if supplied, should be an ``Environment``
+        contains all currently-available distributions.  If `full_env` is not
+        supplied, one is created automatically from the ``WorkingSet`` this
+        method is called on, which will typically mean that every directory on
+        ``sys.path`` will be scanned for distributions.
+
+        `installer` is a standard installer callback as used by the
+        ``resolve()`` method. The `fallback` flag indicates whether we should
+        attempt to resolve older versions of a plugin if the newest version
+        cannot be resolved.
+
+        This method returns a 2-tuple: (`distributions`, `error_info`), where
+        `distributions` is a list of the distributions found in `plugin_env`
+        that were loadable, along with any other distributions that are needed
+        to resolve their dependencies.  `error_info` is a dictionary mapping
+        unloadable plugin distributions to an exception instance describing the
+        error that occurred. Usually this will be a ``DistributionNotFound`` or
+        ``VersionConflict`` instance.
+        """
+
+        plugin_projects = list(plugin_env)
+        # scan project names in alphabetic order
+        plugin_projects.sort()
+
+        error_info: dict[Distribution, Exception] = {}
+        distributions: dict[Distribution, Exception | None] = {}
+
+        if full_env is None:
+            env = Environment(self.entries)
+            env += plugin_env
+        else:
+            env = full_env + plugin_env
+
+        shadow_set = self.__class__([])
+        # put all our entries in shadow_set
+        list(map(shadow_set.add, self))
+
+        for project_name in plugin_projects:
+            for dist in plugin_env[project_name]:
+                req = [dist.as_requirement()]
+
+                try:
+                    resolvees = shadow_set.resolve(req, env, installer)
+
+                except ResolutionError as v:
+                    # save error info
+                    error_info[dist] = v
+                    if fallback:
+                        # try the next older version of project
+                        continue
+                    else:
+                        # give up on this project, keep going
+                        break
+
+                else:
+                    list(map(shadow_set.add, resolvees))
+                    distributions.update(dict.fromkeys(resolvees))
+
+                    # success, no need to try any more versions of this project
+                    break
+
+        sorted_distributions = list(distributions)
+        sorted_distributions.sort()
+
+        return sorted_distributions, error_info
+
+    def require(self, *requirements: _NestedStr):
+        """Ensure that distributions matching `requirements` are activated
+
+        `requirements` must be a string or a (possibly-nested) sequence
+        thereof, specifying the distributions and versions required.  The
+        return value is a sequence of the distributions that needed to be
+        activated to fulfill the requirements; all relevant distributions are
+        included, even if they were already activated in this working set.
+        """
+        needed = self.resolve(parse_requirements(requirements))
+
+        for dist in needed:
+            self.add(dist)
+
+        return needed
+
+    def subscribe(
+        self, callback: Callable[[Distribution], object], existing: bool = True
+    ):
+        """Invoke `callback` for all distributions
+
+        If `existing=True` (default),
+        call on all existing ones, as well.
+        """
+        if callback in self.callbacks:
+            return
+        self.callbacks.append(callback)
+        if not existing:
+            return
+        for dist in self:
+            callback(dist)
+
+    def _added_new(self, dist):
+        for callback in self.callbacks:
+            callback(dist)
+
+    def __getstate__(self):
+        return (
+            self.entries[:],
+            self.entry_keys.copy(),
+            self.by_key.copy(),
+            self.normalized_to_canonical_keys.copy(),
+            self.callbacks[:],
+        )
+
+    def __setstate__(self, e_k_b_n_c):
+        entries, keys, by_key, normalized_to_canonical_keys, callbacks = e_k_b_n_c
+        self.entries = entries[:]
+        self.entry_keys = keys.copy()
+        self.by_key = by_key.copy()
+        self.normalized_to_canonical_keys = normalized_to_canonical_keys.copy()
+        self.callbacks = callbacks[:]
+
+
+class _ReqExtras(Dict["Requirement", Tuple[str, ...]]):
+    """
+    Map each requirement to the extras that demanded it.
+    """
+
+    def markers_pass(self, req: Requirement, extras: tuple[str, ...] | None = None):
+        """
+        Evaluate markers for req against each extra that
+        demanded it.
+
+        Return False if the req has a marker and fails
+        evaluation. Otherwise, return True.
+        """
+        extra_evals = (
+            req.marker.evaluate({'extra': extra})
+            for extra in self.get(req, ()) + (extras or (None,))
+        )
+        return not req.marker or any(extra_evals)
+
+
+class Environment:
+    """Searchable snapshot of distributions on a search path"""
+
+    def __init__(
+        self,
+        search_path: Iterable[str] | None = None,
+        platform: str | None = get_supported_platform(),
+        python: str | None = PY_MAJOR,
+    ):
+        """Snapshot distributions available on a search path
+
+        Any distributions found on `search_path` are added to the environment.
+        `search_path` should be a sequence of ``sys.path`` items.  If not
+        supplied, ``sys.path`` is used.
+
+        `platform` is an optional string specifying the name of the platform
+        that platform-specific distributions must be compatible with.  If
+        unspecified, it defaults to the current platform.  `python` is an
+        optional string naming the desired version of Python (e.g. ``'3.6'``);
+        it defaults to the current version.
+
+        You may explicitly set `platform` (and/or `python`) to ``None`` if you
+        wish to map *all* distributions, not just those compatible with the
+        running platform or Python version.
+        """
+        self._distmap = {}
+        self.platform = platform
+        self.python = python
+        self.scan(search_path)
+
+    def can_add(self, dist: Distribution):
+        """Is distribution `dist` acceptable for this environment?
+
+        The distribution must match the platform and python version
+        requirements specified when this environment was created, or False
+        is returned.
+        """
+        py_compat = (
+            self.python is None
+            or dist.py_version is None
+            or dist.py_version == self.python
+        )
+        return py_compat and compatible_platforms(dist.platform, self.platform)
+
+    def remove(self, dist: Distribution):
+        """Remove `dist` from the environment"""
+        self._distmap[dist.key].remove(dist)
+
+    def scan(self, search_path: Iterable[str] | None = None):
+        """Scan `search_path` for distributions usable in this environment
+
+        Any distributions found are added to the environment.
+        `search_path` should be a sequence of ``sys.path`` items.  If not
+        supplied, ``sys.path`` is used.  Only distributions conforming to
+        the platform/python version defined at initialization are added.
+        """
+        if search_path is None:
+            search_path = sys.path
+
+        for item in search_path:
+            for dist in find_distributions(item):
+                self.add(dist)
+
+    def __getitem__(self, project_name: str) -> list[Distribution]:
+        """Return a newest-to-oldest list of distributions for `project_name`
+
+        Uses case-insensitive `project_name` comparison, assuming all the
+        project's distributions use their project's name converted to all
+        lowercase as their key.
+
+        """
+        distribution_key = project_name.lower()
+        return self._distmap.get(distribution_key, [])
+
+    def add(self, dist: Distribution):
+        """Add `dist` if we ``can_add()`` it and it has not already been added"""
+        if self.can_add(dist) and dist.has_version():
+            dists = self._distmap.setdefault(dist.key, [])
+            if dist not in dists:
+                dists.append(dist)
+                dists.sort(key=operator.attrgetter('hashcmp'), reverse=True)
+
+    @overload
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _InstallerTypeT[_DistributionT],
+        replace_conflicting: bool = False,
+    ) -> _DistributionT: ...
+    @overload
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _InstallerType | None = None,
+        replace_conflicting: bool = False,
+    ) -> Distribution | None: ...
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _InstallerType | None | _InstallerTypeT[_DistributionT] = None,
+        replace_conflicting: bool = False,
+    ) -> Distribution | None:
+        """Find distribution best matching `req` and usable on `working_set`
+
+        This calls the ``find(req)`` method of the `working_set` to see if a
+        suitable distribution is already active.  (This may raise
+        ``VersionConflict`` if an unsuitable version of the project is already
+        active in the specified `working_set`.)  If a suitable distribution
+        isn't active, this method returns the newest distribution in the
+        environment that meets the ``Requirement`` in `req`.  If no suitable
+        distribution is found, and `installer` is supplied, then the result of
+        calling the environment's ``obtain(req, installer)`` method will be
+        returned.
+        """
+        try:
+            dist = working_set.find(req)
+        except VersionConflict:
+            if not replace_conflicting:
+                raise
+            dist = None
+        if dist is not None:
+            return dist
+        for dist in self[req.key]:
+            if dist in req:
+                return dist
+        # try to download/install
+        return self.obtain(req, installer)
+
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: _InstallerTypeT[_DistributionT],
+    ) -> _DistributionT: ...
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: Callable[[Requirement], None] | None = None,
+    ) -> None: ...
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: _InstallerType | None = None,
+    ) -> Distribution | None: ...
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: Callable[[Requirement], None]
+        | _InstallerType
+        | None
+        | _InstallerTypeT[_DistributionT] = None,
+    ) -> Distribution | None:
+        """Obtain a distribution matching `requirement` (e.g. via download)
+
+        Obtain a distro that matches requirement (e.g. via download).  In the
+        base ``Environment`` class, this routine just returns
+        ``installer(requirement)``, unless `installer` is None, in which case
+        None is returned instead.  This method is a hook that allows subclasses
+        to attempt other ways of obtaining a distribution before falling back
+        to the `installer` argument."""
+        return installer(requirement) if installer else None
+
+    def __iter__(self) -> Iterator[str]:
+        """Yield the unique project names of the available distributions"""
+        for key in self._distmap.keys():
+            if self[key]:
+                yield key
+
+    def __iadd__(self, other: Distribution | Environment):
+        """In-place addition of a distribution or environment"""
+        if isinstance(other, Distribution):
+            self.add(other)
+        elif isinstance(other, Environment):
+            for project in other:
+                for dist in other[project]:
+                    self.add(dist)
+        else:
+            raise TypeError("Can't add %r to environment" % (other,))
+        return self
+
+    def __add__(self, other: Distribution | Environment):
+        """Add an environment or distribution to an environment"""
+        new = self.__class__([], platform=None, python=None)
+        for env in self, other:
+            new += env
+        return new
+
+
+# XXX backward compatibility
+AvailableDistributions = Environment
+
+
+class ExtractionError(RuntimeError):
+    """An error occurred extracting a resource
+
+    The following attributes are available from instances of this exception:
+
+    manager
+        The resource manager that raised this exception
+
+    cache_path
+        The base directory for resource extraction
+
+    original_error
+        The exception instance that caused extraction to fail
+    """
+
+    manager: ResourceManager
+    cache_path: str
+    original_error: BaseException | None
+
+
+class ResourceManager:
+    """Manage resource extraction and packages"""
+
+    extraction_path: str | None = None
+
+    def __init__(self):
+        self.cached_files = {}
+
+    def resource_exists(self, package_or_requirement: _PkgReqType, resource_name: str):
+        """Does the named resource exist?"""
+        return get_provider(package_or_requirement).has_resource(resource_name)
+
+    def resource_isdir(self, package_or_requirement: _PkgReqType, resource_name: str):
+        """Is the named resource an existing directory?"""
+        return get_provider(package_or_requirement).resource_isdir(resource_name)
+
+    def resource_filename(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ):
+        """Return a true filesystem path for specified resource"""
+        return get_provider(package_or_requirement).get_resource_filename(
+            self, resource_name
+        )
+
+    def resource_stream(self, package_or_requirement: _PkgReqType, resource_name: str):
+        """Return a readable file-like object for specified resource"""
+        return get_provider(package_or_requirement).get_resource_stream(
+            self, resource_name
+        )
+
+    def resource_string(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> bytes:
+        """Return specified resource as :obj:`bytes`"""
+        return get_provider(package_or_requirement).get_resource_string(
+            self, resource_name
+        )
+
+    def resource_listdir(self, package_or_requirement: _PkgReqType, resource_name: str):
+        """List the contents of the named resource directory"""
+        return get_provider(package_or_requirement).resource_listdir(resource_name)
+
+    def extraction_error(self) -> NoReturn:
+        """Give an error message for problems extracting file(s)"""
+
+        old_exc = sys.exc_info()[1]
+        cache_path = self.extraction_path or get_default_cache()
+
+        tmpl = textwrap.dedent(
+            """
+            Can't extract file(s) to egg cache
+
+            The following error occurred while trying to extract file(s)
+            to the Python egg cache:
+
+              {old_exc}
+
+            The Python egg cache directory is currently set to:
+
+              {cache_path}
+
+            Perhaps your account does not have write access to this directory?
+            You can change the cache directory by setting the PYTHON_EGG_CACHE
+            environment variable to point to an accessible directory.
+            """
+        ).lstrip()
+        err = ExtractionError(tmpl.format(**locals()))
+        err.manager = self
+        err.cache_path = cache_path
+        err.original_error = old_exc
+        raise err
+
+    def get_cache_path(self, archive_name: str, names: Iterable[StrPath] = ()):
+        """Return absolute location in cache for `archive_name` and `names`
+
+        The parent directory of the resulting path will be created if it does
+        not already exist.  `archive_name` should be the base filename of the
+        enclosing egg (which may not be the name of the enclosing zipfile!),
+        including its ".egg" extension.  `names`, if provided, should be a
+        sequence of path name parts "under" the egg's extraction location.
+
+        This method should only be called by resource providers that need to
+        obtain an extraction location, and only for names they intend to
+        extract, as it tracks the generated names for possible cleanup later.
+        """
+        extract_path = self.extraction_path or get_default_cache()
+        target_path = os.path.join(extract_path, archive_name + '-tmp', *names)
+        try:
+            _bypass_ensure_directory(target_path)
+        except Exception:
+            self.extraction_error()
+
+        self._warn_unsafe_extraction_path(extract_path)
+
+        self.cached_files[target_path] = True
+        return target_path
+
+    @staticmethod
+    def _warn_unsafe_extraction_path(path):
+        """
+        If the default extraction path is overridden and set to an insecure
+        location, such as /tmp, it opens up an opportunity for an attacker to
+        replace an extracted file with an unauthorized payload. Warn the user
+        if a known insecure location is used.
+
+        See Distribute #375 for more details.
+        """
+        if os.name == 'nt' and not path.startswith(os.environ['windir']):
+            # On Windows, permissions are generally restrictive by default
+            #  and temp directories are not writable by other users, so
+            #  bypass the warning.
+            return
+        mode = os.stat(path).st_mode
+        if mode & stat.S_IWOTH or mode & stat.S_IWGRP:
+            msg = (
+                "Extraction path is writable by group/others "
+                "and vulnerable to attack when "
+                "used with get_resource_filename ({path}). "
+                "Consider a more secure "
+                "location (set with .set_extraction_path or the "
+                "PYTHON_EGG_CACHE environment variable)."
+            ).format(**locals())
+            warnings.warn(msg, UserWarning)
+
+    def postprocess(self, tempname: StrOrBytesPath, filename: StrOrBytesPath):
+        """Perform any platform-specific postprocessing of `tempname`
+
+        This is where Mac header rewrites should be done; other platforms don't
+        have anything special they should do.
+
+        Resource providers should call this method ONLY after successfully
+        extracting a compressed resource.  They must NOT call it on resources
+        that are already in the filesystem.
+
+        `tempname` is the current (temporary) name of the file, and `filename`
+        is the name it will be renamed to by the caller after this routine
+        returns.
+        """
+
+        if os.name == 'posix':
+            # Make the resource executable
+            mode = ((os.stat(tempname).st_mode) | 0o555) & 0o7777
+            os.chmod(tempname, mode)
+
+    def set_extraction_path(self, path: str):
+        """Set the base path where resources will be extracted to, if needed.
+
+        If you do not call this routine before any extractions take place, the
+        path defaults to the return value of ``get_default_cache()``.  (Which
+        is based on the ``PYTHON_EGG_CACHE`` environment variable, with various
+        platform-specific fallbacks.  See that routine's documentation for more
+        details.)
+
+        Resources are extracted to subdirectories of this path based upon
+        information given by the ``IResourceProvider``.  You may set this to a
+        temporary directory, but then you must call ``cleanup_resources()`` to
+        delete the extracted files when done.  There is no guarantee that
+        ``cleanup_resources()`` will be able to remove all extracted files.
+
+        (Note: you may not change the extraction path for a given resource
+        manager once resources have been extracted, unless you first call
+        ``cleanup_resources()``.)
+        """
+        if self.cached_files:
+            raise ValueError("Can't change extraction path, files already extracted")
+
+        self.extraction_path = path
+
+    def cleanup_resources(self, force: bool = False) -> list[str]:
+        """
+        Delete all extracted resource files and directories, returning a list
+        of the file and directory names that could not be successfully removed.
+        This function does not have any concurrency protection, so it should
+        generally only be called when the extraction path is a temporary
+        directory exclusive to a single process.  This method is not
+        automatically called; you must call it explicitly or register it as an
+        ``atexit`` function if you wish to ensure cleanup of a temporary
+        directory used for extractions.
+        """
+        # XXX
+        return []
+
+
+def get_default_cache() -> str:
+    """
+    Return the ``PYTHON_EGG_CACHE`` environment variable
+    or a platform-relevant user cache dir for an app
+    named "Python-Eggs".
+    """
+    return os.environ.get('PYTHON_EGG_CACHE') or _user_cache_dir(appname='Python-Eggs')
+
+
+def safe_name(name: str):
+    """Convert an arbitrary string to a standard distribution name
+
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub('[^A-Za-z0-9.]+', '-', name)
+
+
+def safe_version(version: str):
+    """
+    Convert an arbitrary string to a standard version string
+    """
+    try:
+        # normalize the version
+        return str(_packaging_version.Version(version))
+    except _packaging_version.InvalidVersion:
+        version = version.replace(' ', '.')
+        return re.sub('[^A-Za-z0-9.]+', '-', version)
+
+
+def _forgiving_version(version):
+    """Fallback when ``safe_version`` is not safe enough
+    >>> parse_version(_forgiving_version('0.23ubuntu1'))
+    <Version('0.23.dev0+sanitized.ubuntu1')>
+    >>> parse_version(_forgiving_version('0.23-'))
+    <Version('0.23.dev0+sanitized')>
+    >>> parse_version(_forgiving_version('0.-_'))
+    <Version('0.dev0+sanitized')>
+    >>> parse_version(_forgiving_version('42.+?1'))
+    <Version('42.dev0+sanitized.1')>
+    >>> parse_version(_forgiving_version('hello world'))
+    <Version('0.dev0+sanitized.hello.world')>
+    """
+    version = version.replace(' ', '.')
+    match = _PEP440_FALLBACK.search(version)
+    if match:
+        safe = match["safe"]
+        rest = version[len(safe) :]
+    else:
+        safe = "0"
+        rest = version
+    local = f"sanitized.{_safe_segment(rest)}".strip(".")
+    return f"{safe}.dev0+{local}"
+
+
+def _safe_segment(segment):
+    """Convert an arbitrary string into a safe segment"""
+    segment = re.sub('[^A-Za-z0-9.]+', '-', segment)
+    segment = re.sub('-[^A-Za-z0-9]+', '-', segment)
+    return re.sub(r'\.[^A-Za-z0-9]+', '.', segment).strip(".-")
+
+
+def safe_extra(extra: str):
+    """Convert an arbitrary string to a standard 'extra' name
+
+    Any runs of non-alphanumeric characters are replaced with a single '_',
+    and the result is always lowercased.
+    """
+    return re.sub('[^A-Za-z0-9.-]+', '_', extra).lower()
+
+
+def to_filename(name: str):
+    """Convert a project or version name to its filename-escaped form
+
+    Any '-' characters are currently replaced with '_'.
+    """
+    return name.replace('-', '_')
+
+
+def invalid_marker(text: str):
+    """
+    Validate text as a PEP 508 environment marker; return an exception
+    if invalid or False otherwise.
+    """
+    try:
+        evaluate_marker(text)
+    except SyntaxError as e:
+        e.filename = None
+        e.lineno = None
+        return e
+    return False
+
+
+def evaluate_marker(text: str, extra: str | None = None) -> bool:
+    """
+    Evaluate a PEP 508 environment marker.
+    Return a boolean indicating the marker result in this environment.
+    Raise SyntaxError if marker is invalid.
+
+    This implementation uses the 'pyparsing' module.
+    """
+    try:
+        marker = _packaging_markers.Marker(text)
+        return marker.evaluate()
+    except _packaging_markers.InvalidMarker as e:
+        raise SyntaxError(e) from e
+
+
+class NullProvider:
+    """Try to implement resources and metadata for arbitrary PEP 302 loaders"""
+
+    egg_name: str | None = None
+    egg_info: str | None = None
+    loader: _LoaderProtocol | None = None
+
+    def __init__(self, module: _ModuleLike):
+        self.loader = getattr(module, '__loader__', None)
+        self.module_path = os.path.dirname(getattr(module, '__file__', ''))
+
+    def get_resource_filename(self, manager: ResourceManager, resource_name: str):
+        return self._fn(self.module_path, resource_name)
+
+    def get_resource_stream(self, manager: ResourceManager, resource_name: str):
+        return io.BytesIO(self.get_resource_string(manager, resource_name))
+
+    def get_resource_string(
+        self, manager: ResourceManager, resource_name: str
+    ) -> bytes:
+        return self._get(self._fn(self.module_path, resource_name))
+
+    def has_resource(self, resource_name: str):
+        return self._has(self._fn(self.module_path, resource_name))
+
+    def _get_metadata_path(self, name):
+        return self._fn(self.egg_info, name)
+
+    def has_metadata(self, name: str) -> bool:
+        if not self.egg_info:
+            return False
+
+        path = self._get_metadata_path(name)
+        return self._has(path)
+
+    def get_metadata(self, name: str):
+        if not self.egg_info:
+            return ""
+        path = self._get_metadata_path(name)
+        value = self._get(path)
+        try:
+            return value.decode('utf-8')
+        except UnicodeDecodeError as exc:
+            # Include the path in the error message to simplify
+            # troubleshooting, and without changing the exception type.
+            exc.reason += ' in {} file at path: {}'.format(name, path)
+            raise
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        return yield_lines(self.get_metadata(name))
+
+    def resource_isdir(self, resource_name: str):
+        return self._isdir(self._fn(self.module_path, resource_name))
+
+    def metadata_isdir(self, name: str) -> bool:
+        return bool(self.egg_info and self._isdir(self._fn(self.egg_info, name)))
+
+    def resource_listdir(self, resource_name: str):
+        return self._listdir(self._fn(self.module_path, resource_name))
+
+    def metadata_listdir(self, name: str) -> list[str]:
+        if self.egg_info:
+            return self._listdir(self._fn(self.egg_info, name))
+        return []
+
+    def run_script(self, script_name: str, namespace: dict[str, Any]):
+        script = 'scripts/' + script_name
+        if not self.has_metadata(script):
+            raise ResolutionError(
+                "Script {script!r} not found in metadata at {self.egg_info!r}".format(
+                    **locals()
+                ),
+            )
+
+        script_text = self.get_metadata(script).replace('\r\n', '\n')
+        script_text = script_text.replace('\r', '\n')
+        script_filename = self._fn(self.egg_info, script)
+        namespace['__file__'] = script_filename
+        if os.path.exists(script_filename):
+            source = _read_utf8_with_fallback(script_filename)
+            code = compile(source, script_filename, 'exec')
+            exec(code, namespace, namespace)
+        else:
+            from linecache import cache
+
+            cache[script_filename] = (
+                len(script_text),
+                0,
+                script_text.split('\n'),
+                script_filename,
+            )
+            script_code = compile(script_text, script_filename, 'exec')
+            exec(script_code, namespace, namespace)
+
+    def _has(self, path) -> bool:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _isdir(self, path) -> bool:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _listdir(self, path) -> list[str]:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _fn(self, base: str | None, resource_name: str):
+        if base is None:
+            raise TypeError(
+                "`base` parameter in `_fn` is `None`. Either override this method or check the parameter first."
+            )
+        self._validate_resource_path(resource_name)
+        if resource_name:
+            return os.path.join(base, *resource_name.split('/'))
+        return base
+
+    @staticmethod
+    def _validate_resource_path(path):
+        """
+        Validate the resource paths according to the docs.
+        https://setuptools.pypa.io/en/latest/pkg_resources.html#basic-resource-access
+
+        >>> warned = getfixture('recwarn')
+        >>> warnings.simplefilter('always')
+        >>> vrp = NullProvider._validate_resource_path
+        >>> vrp('foo/bar.txt')
+        >>> bool(warned)
+        False
+        >>> vrp('../foo/bar.txt')
+        >>> bool(warned)
+        True
+        >>> warned.clear()
+        >>> vrp('/foo/bar.txt')
+        >>> bool(warned)
+        True
+        >>> vrp('foo/../../bar.txt')
+        >>> bool(warned)
+        True
+        >>> warned.clear()
+        >>> vrp('foo/f../bar.txt')
+        >>> bool(warned)
+        False
+
+        Windows path separators are straight-up disallowed.
+        >>> vrp(r'\\foo/bar.txt')
+        Traceback (most recent call last):
+        ...
+        ValueError: Use of .. or absolute path in a resource path \
+is not allowed.
+
+        >>> vrp(r'C:\\foo/bar.txt')
+        Traceback (most recent call last):
+        ...
+        ValueError: Use of .. or absolute path in a resource path \
+is not allowed.
+
+        Blank values are allowed
+
+        >>> vrp('')
+        >>> bool(warned)
+        False
+
+        Non-string values are not.
+
+        >>> vrp(None)
+        Traceback (most recent call last):
+        ...
+        AttributeError: ...
+        """
+        invalid = (
+            os.path.pardir in path.split(posixpath.sep)
+            or posixpath.isabs(path)
+            or ntpath.isabs(path)
+            or path.startswith("\\")
+        )
+        if not invalid:
+            return
+
+        msg = "Use of .. or absolute path in a resource path is not allowed."
+
+        # Aggressively disallow Windows absolute paths
+        if (path.startswith("\\") or ntpath.isabs(path)) and not posixpath.isabs(path):
+            raise ValueError(msg)
+
+        # for compatibility, warn; in future
+        # raise ValueError(msg)
+        issue_warning(
+            msg[:-1] + " and will raise exceptions in a future release.",
+            DeprecationWarning,
+        )
+
+    def _get(self, path) -> bytes:
+        if hasattr(self.loader, 'get_data') and self.loader:
+            # Already checked get_data exists
+            return self.loader.get_data(path)  # type: ignore[attr-defined]
+        raise NotImplementedError(
+            "Can't perform this operation for loaders without 'get_data()'"
+        )
+
+
+register_loader_type(object, NullProvider)
+
+
+def _parents(path):
+    """
+    yield all parents of path including path
+    """
+    last = None
+    while path != last:
+        yield path
+        last = path
+        path, _ = os.path.split(path)
+
+
+class EggProvider(NullProvider):
+    """Provider based on a virtual filesystem"""
+
+    def __init__(self, module: _ModuleLike):
+        super().__init__(module)
+        self._setup_prefix()
+
+    def _setup_prefix(self):
+        # Assume that metadata may be nested inside a "basket"
+        # of multiple eggs and use module_path instead of .archive.
+        eggs = filter(_is_egg_path, _parents(self.module_path))
+        egg = next(eggs, None)
+        egg and self._set_egg(egg)
+
+    def _set_egg(self, path: str):
+        self.egg_name = os.path.basename(path)
+        self.egg_info = os.path.join(path, 'EGG-INFO')
+        self.egg_root = path
+
+
+class DefaultProvider(EggProvider):
+    """Provides access to package resources in the filesystem"""
+
+    def _has(self, path) -> bool:
+        return os.path.exists(path)
+
+    def _isdir(self, path) -> bool:
+        return os.path.isdir(path)
+
+    def _listdir(self, path):
+        return os.listdir(path)
+
+    def get_resource_stream(self, manager: object, resource_name: str):
+        return open(self._fn(self.module_path, resource_name), 'rb')
+
+    def _get(self, path) -> bytes:
+        with open(path, 'rb') as stream:
+            return stream.read()
+
+    @classmethod
+    def _register(cls):
+        loader_names = (
+            'SourceFileLoader',
+            'SourcelessFileLoader',
+        )
+        for name in loader_names:
+            loader_cls = getattr(importlib.machinery, name, type(None))
+            register_loader_type(loader_cls, cls)
+
+
+DefaultProvider._register()
+
+
+class EmptyProvider(NullProvider):
+    """Provider that returns nothing for all requests"""
+
+    # A special case, we don't want all Providers inheriting from NullProvider to have a potentially None module_path
+    module_path: str | None = None  # type: ignore[assignment]
+
+    _isdir = _has = lambda self, path: False
+
+    def _get(self, path) -> bytes:
+        return b''
+
+    def _listdir(self, path):
+        return []
+
+    def __init__(self):
+        pass
+
+
+empty_provider = EmptyProvider()
+
+
+class ZipManifests(Dict[str, "MemoizedZipManifests.manifest_mod"]):
+    """
+    zip manifest builder
+    """
+
+    # `path` could be `StrPath | IO[bytes]` but that violates the LSP for `MemoizedZipManifests.load`
+    @classmethod
+    def build(cls, path: str):
+        """
+        Build a dictionary similar to the zipimport directory
+        caches, except instead of tuples, store ZipInfo objects.
+
+        Use a platform-specific path separator (os.sep) for the path keys
+        for compatibility with pypy on Windows.
+        """
+        with zipfile.ZipFile(path) as zfile:
+            items = (
+                (
+                    name.replace('/', os.sep),
+                    zfile.getinfo(name),
+                )
+                for name in zfile.namelist()
+            )
+            return dict(items)
+
+    load = build
+
+
+class MemoizedZipManifests(ZipManifests):
+    """
+    Memoized zipfile manifests.
+    """
+
+    class manifest_mod(NamedTuple):
+        manifest: dict[str, zipfile.ZipInfo]
+        mtime: float
+
+    def load(self, path: str) -> dict[str, zipfile.ZipInfo]:  # type: ignore[override] # ZipManifests.load is a classmethod
+        """
+        Load a manifest at path or return a suitable manifest already loaded.
+        """
+        path = os.path.normpath(path)
+        mtime = os.stat(path).st_mtime
+
+        if path not in self or self[path].mtime != mtime:
+            manifest = self.build(path)
+            self[path] = self.manifest_mod(manifest, mtime)
+
+        return self[path].manifest
+
+
+class ZipProvider(EggProvider):
+    """Resource support for zips and eggs"""
+
+    eagers: list[str] | None = None
+    _zip_manifests = MemoizedZipManifests()
+    # ZipProvider's loader should always be a zipimporter or equivalent
+    loader: zipimport.zipimporter
+
+    def __init__(self, module: _ZipLoaderModule):
+        super().__init__(module)
+        self.zip_pre = self.loader.archive + os.sep
+
+    def _zipinfo_name(self, fspath):
+        # Convert a virtual filename (full path to file) into a zipfile subpath
+        # usable with the zipimport directory cache for our target archive
+        fspath = fspath.rstrip(os.sep)
+        if fspath == self.loader.archive:
+            return ''
+        if fspath.startswith(self.zip_pre):
+            return fspath[len(self.zip_pre) :]
+        raise AssertionError("%s is not a subpath of %s" % (fspath, self.zip_pre))
+
+    def _parts(self, zip_path):
+        # Convert a zipfile subpath into an egg-relative path part list.
+        # pseudo-fs path
+        fspath = self.zip_pre + zip_path
+        if fspath.startswith(self.egg_root + os.sep):
+            return fspath[len(self.egg_root) + 1 :].split(os.sep)
+        raise AssertionError("%s is not a subpath of %s" % (fspath, self.egg_root))
+
+    @property
+    def zipinfo(self):
+        return self._zip_manifests.load(self.loader.archive)
+
+    def get_resource_filename(self, manager: ResourceManager, resource_name: str):
+        if not self.egg_name:
+            raise NotImplementedError(
+                "resource_filename() only supported for .egg, not .zip"
+            )
+        # no need to lock for extraction, since we use temp names
+        zip_path = self._resource_to_zip(resource_name)
+        eagers = self._get_eager_resources()
+        if '/'.join(self._parts(zip_path)) in eagers:
+            for name in eagers:
+                self._extract_resource(manager, self._eager_to_zip(name))
+        return self._extract_resource(manager, zip_path)
+
+    @staticmethod
+    def _get_date_and_size(zip_stat):
+        size = zip_stat.file_size
+        # ymdhms+wday, yday, dst
+        date_time = zip_stat.date_time + (0, 0, -1)
+        # 1980 offset already done
+        timestamp = time.mktime(date_time)
+        return timestamp, size
+
+    # FIXME: 'ZipProvider._extract_resource' is too complex (12)
+    def _extract_resource(self, manager: ResourceManager, zip_path) -> str:  # noqa: C901
+        if zip_path in self._index():
+            for name in self._index()[zip_path]:
+                last = self._extract_resource(manager, os.path.join(zip_path, name))
+            # return the extracted directory name
+            return os.path.dirname(last)
+
+        timestamp, size = self._get_date_and_size(self.zipinfo[zip_path])
+
+        if not WRITE_SUPPORT:
+            raise OSError(
+                '"os.rename" and "os.unlink" are not supported on this platform'
+            )
+        try:
+            if not self.egg_name:
+                raise OSError(
+                    '"egg_name" is empty. This likely means no egg could be found from the "module_path".'
+                )
+            real_path = manager.get_cache_path(self.egg_name, self._parts(zip_path))
+
+            if self._is_current(real_path, zip_path):
+                return real_path
+
+            outf, tmpnam = _mkstemp(
+                ".$extract",
+                dir=os.path.dirname(real_path),
+            )
+            os.write(outf, self.loader.get_data(zip_path))
+            os.close(outf)
+            utime(tmpnam, (timestamp, timestamp))
+            manager.postprocess(tmpnam, real_path)
+
+            try:
+                rename(tmpnam, real_path)
+
+            except OSError:
+                if os.path.isfile(real_path):
+                    if self._is_current(real_path, zip_path):
+                        # the file became current since it was checked above,
+                        #  so proceed.
+                        return real_path
+                    # Windows, del old file and retry
+                    elif os.name == 'nt':
+                        unlink(real_path)
+                        rename(tmpnam, real_path)
+                        return real_path
+                raise
+
+        except OSError:
+            # report a user-friendly error
+            manager.extraction_error()
+
+        return real_path
+
+    def _is_current(self, file_path, zip_path):
+        """
+        Return True if the file_path is current for this zip_path
+        """
+        timestamp, size = self._get_date_and_size(self.zipinfo[zip_path])
+        if not os.path.isfile(file_path):
+            return False
+        stat = os.stat(file_path)
+        if stat.st_size != size or stat.st_mtime != timestamp:
+            return False
+        # check that the contents match
+        zip_contents = self.loader.get_data(zip_path)
+        with open(file_path, 'rb') as f:
+            file_contents = f.read()
+        return zip_contents == file_contents
+
+    def _get_eager_resources(self):
+        if self.eagers is None:
+            eagers = []
+            for name in ('native_libs.txt', 'eager_resources.txt'):
+                if self.has_metadata(name):
+                    eagers.extend(self.get_metadata_lines(name))
+            self.eagers = eagers
+        return self.eagers
+
+    def _index(self):
+        try:
+            return self._dirindex
+        except AttributeError:
+            ind = {}
+            for path in self.zipinfo:
+                parts = path.split(os.sep)
+                while parts:
+                    parent = os.sep.join(parts[:-1])
+                    if parent in ind:
+                        ind[parent].append(parts[-1])
+                        break
+                    else:
+                        ind[parent] = [parts.pop()]
+            self._dirindex = ind
+            return ind
+
+    def _has(self, fspath) -> bool:
+        zip_path = self._zipinfo_name(fspath)
+        return zip_path in self.zipinfo or zip_path in self._index()
+
+    def _isdir(self, fspath) -> bool:
+        return self._zipinfo_name(fspath) in self._index()
+
+    def _listdir(self, fspath):
+        return list(self._index().get(self._zipinfo_name(fspath), ()))
+
+    def _eager_to_zip(self, resource_name: str):
+        return self._zipinfo_name(self._fn(self.egg_root, resource_name))
+
+    def _resource_to_zip(self, resource_name: str):
+        return self._zipinfo_name(self._fn(self.module_path, resource_name))
+
+
+register_loader_type(zipimport.zipimporter, ZipProvider)
+
+
+class FileMetadata(EmptyProvider):
+    """Metadata handler for standalone PKG-INFO files
+
+    Usage::
+
+        metadata = FileMetadata("/path/to/PKG-INFO")
+
+    This provider rejects all data and metadata requests except for PKG-INFO,
+    which is treated as existing, and will be the contents of the file at
+    the provided location.
+    """
+
+    def __init__(self, path: StrPath):
+        self.path = path
+
+    def _get_metadata_path(self, name):
+        return self.path
+
+    def has_metadata(self, name: str) -> bool:
+        return name == 'PKG-INFO' and os.path.isfile(self.path)
+
+    def get_metadata(self, name: str):
+        if name != 'PKG-INFO':
+            raise KeyError("No metadata except PKG-INFO is available")
+
+        with open(self.path, encoding='utf-8', errors="replace") as f:
+            metadata = f.read()
+        self._warn_on_replacement(metadata)
+        return metadata
+
+    def _warn_on_replacement(self, metadata):
+        replacement_char = '�'
+        if replacement_char in metadata:
+            tmpl = "{self.path} could not be properly decoded in UTF-8"
+            msg = tmpl.format(**locals())
+            warnings.warn(msg)
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        return yield_lines(self.get_metadata(name))
+
+
+class PathMetadata(DefaultProvider):
+    """Metadata provider for egg directories
+
+    Usage::
+
+        # Development eggs:
+
+        egg_info = "/path/to/PackageName.egg-info"
+        base_dir = os.path.dirname(egg_info)
+        metadata = PathMetadata(base_dir, egg_info)
+        dist_name = os.path.splitext(os.path.basename(egg_info))[0]
+        dist = Distribution(basedir, project_name=dist_name, metadata=metadata)
+
+        # Unpacked egg directories:
+
+        egg_path = "/path/to/PackageName-ver-pyver-etc.egg"
+        metadata = PathMetadata(egg_path, os.path.join(egg_path,'EGG-INFO'))
+        dist = Distribution.from_filename(egg_path, metadata=metadata)
+    """
+
+    def __init__(self, path: str, egg_info: str):
+        self.module_path = path
+        self.egg_info = egg_info
+
+
+class EggMetadata(ZipProvider):
+    """Metadata provider for .egg files"""
+
+    def __init__(self, importer: zipimport.zipimporter):
+        """Create a metadata provider from a zipimporter"""
+
+        self.zip_pre = importer.archive + os.sep
+        self.loader = importer
+        if importer.prefix:
+            self.module_path = os.path.join(importer.archive, importer.prefix)
+        else:
+            self.module_path = importer.archive
+        self._setup_prefix()
+
+
+_distribution_finders: dict[type, _DistFinderType[Any]] = _declare_state(
+    'dict', '_distribution_finders', {}
+)
+
+
+def register_finder(importer_type: type[_T], distribution_finder: _DistFinderType[_T]):
+    """Register `distribution_finder` to find distributions in sys.path items
+
+    `importer_type` is the type or class of a PEP 302 "Importer" (sys.path item
+    handler), and `distribution_finder` is a callable that, passed a path
+    item and the importer instance, yields ``Distribution`` instances found on
+    that path item.  See ``pkg_resources.find_on_path`` for an example."""
+    _distribution_finders[importer_type] = distribution_finder
+
+
+def find_distributions(path_item: str, only: bool = False):
+    """Yield distributions accessible via `path_item`"""
+    importer = get_importer(path_item)
+    finder = _find_adapter(_distribution_finders, importer)
+    return finder(importer, path_item, only)
+
+
+def find_eggs_in_zip(
+    importer: zipimport.zipimporter, path_item: str, only: bool = False
+) -> Iterator[Distribution]:
+    """
+    Find eggs in zip files; possibly multiple nested eggs.
+    """
+    if importer.archive.endswith('.whl'):
+        # wheels are not supported with this finder
+        # they don't have PKG-INFO metadata, and won't ever contain eggs
+        return
+    metadata = EggMetadata(importer)
+    if metadata.has_metadata('PKG-INFO'):
+        yield Distribution.from_filename(path_item, metadata=metadata)
+    if only:
+        # don't yield nested distros
+        return
+    for subitem in metadata.resource_listdir(''):
+        if _is_egg_path(subitem):
+            subpath = os.path.join(path_item, subitem)
+            dists = find_eggs_in_zip(zipimport.zipimporter(subpath), subpath)
+            yield from dists
+        elif subitem.lower().endswith(('.dist-info', '.egg-info')):
+            subpath = os.path.join(path_item, subitem)
+            submeta = EggMetadata(zipimport.zipimporter(subpath))
+            submeta.egg_info = subpath
+            yield Distribution.from_location(path_item, subitem, submeta)
+
+
+register_finder(zipimport.zipimporter, find_eggs_in_zip)
+
+
+def find_nothing(
+    importer: object | None, path_item: str | None, only: bool | None = False
+):
+    return ()
+
+
+register_finder(object, find_nothing)
+
+
+def find_on_path(importer: object | None, path_item, only=False):
+    """Yield distributions accessible on a sys.path directory"""
+    path_item = _normalize_cached(path_item)
+
+    if _is_unpacked_egg(path_item):
+        yield Distribution.from_filename(
+            path_item,
+            metadata=PathMetadata(path_item, os.path.join(path_item, 'EGG-INFO')),
+        )
+        return
+
+    entries = (os.path.join(path_item, child) for child in safe_listdir(path_item))
+
+    # scan for .egg and .egg-info in directory
+    for entry in sorted(entries):
+        fullpath = os.path.join(path_item, entry)
+        factory = dist_factory(path_item, entry, only)
+        yield from factory(fullpath)
+
+
+def dist_factory(path_item, entry, only):
+    """Return a dist_factory for the given entry."""
+    lower = entry.lower()
+    is_egg_info = lower.endswith('.egg-info')
+    is_dist_info = lower.endswith('.dist-info') and os.path.isdir(
+        os.path.join(path_item, entry)
+    )
+    is_meta = is_egg_info or is_dist_info
+    return (
+        distributions_from_metadata
+        if is_meta
+        else find_distributions
+        if not only and _is_egg_path(entry)
+        else resolve_egg_link
+        if not only and lower.endswith('.egg-link')
+        else NoDists()
+    )
+
+
+class NoDists:
+    """
+    >>> bool(NoDists())
+    False
+
+    >>> list(NoDists()('anything'))
+    []
+    """
+
+    def __bool__(self):
+        return False
+
+    def __call__(self, fullpath):
+        return iter(())
+
+
+def safe_listdir(path: StrOrBytesPath):
+    """
+    Attempt to list contents of path, but suppress some exceptions.
+    """
+    try:
+        return os.listdir(path)
+    except (PermissionError, NotADirectoryError):
+        pass
+    except OSError as e:
+        # Ignore the directory if does not exist, not a directory or
+        # permission denied
+        if e.errno not in (errno.ENOTDIR, errno.EACCES, errno.ENOENT):
+            raise
+    return ()
+
+
+def distributions_from_metadata(path: str):
+    root = os.path.dirname(path)
+    if os.path.isdir(path):
+        if len(os.listdir(path)) == 0:
+            # empty metadata dir; skip
+            return
+        metadata: _MetadataType = PathMetadata(root, path)
+    else:
+        metadata = FileMetadata(path)
+    entry = os.path.basename(path)
+    yield Distribution.from_location(
+        root,
+        entry,
+        metadata,
+        precedence=DEVELOP_DIST,
+    )
+
+
+def non_empty_lines(path):
+    """
+    Yield non-empty lines from file at path
+    """
+    for line in _read_utf8_with_fallback(path).splitlines():
+        line = line.strip()
+        if line:
+            yield line
+
+
+def resolve_egg_link(path):
+    """
+    Given a path to an .egg-link, resolve distributions
+    present in the referenced path.
+    """
+    referenced_paths = non_empty_lines(path)
+    resolved_paths = (
+        os.path.join(os.path.dirname(path), ref) for ref in referenced_paths
+    )
+    dist_groups = map(find_distributions, resolved_paths)
+    return next(dist_groups, ())
+
+
+if hasattr(pkgutil, 'ImpImporter'):
+    register_finder(pkgutil.ImpImporter, find_on_path)
+
+register_finder(importlib.machinery.FileFinder, find_on_path)
+
+_namespace_handlers: dict[type, _NSHandlerType[Any]] = _declare_state(
+    'dict', '_namespace_handlers', {}
+)
+_namespace_packages: dict[str | None, list[str]] = _declare_state(
+    'dict', '_namespace_packages', {}
+)
+
+
+def register_namespace_handler(
+    importer_type: type[_T], namespace_handler: _NSHandlerType[_T]
+):
+    """Register `namespace_handler` to declare namespace packages
+
+    `importer_type` is the type or class of a PEP 302 "Importer" (sys.path item
+    handler), and `namespace_handler` is a callable like this::
+
+        def namespace_handler(importer, path_entry, moduleName, module):
+            # return a path_entry to use for child packages
+
+    Namespace handlers are only called if the importer object has already
+    agreed that it can handle the relevant path item, and they should only
+    return a subpath if the module __path__ does not already contain an
+    equivalent subpath.  For an example namespace handler, see
+    ``pkg_resources.file_ns_handler``.
+    """
+    _namespace_handlers[importer_type] = namespace_handler
+
+
+def _handle_ns(packageName, path_item):
+    """Ensure that named package includes a subpath of path_item (if needed)"""
+
+    importer = get_importer(path_item)
+    if importer is None:
+        return None
+
+    # use find_spec (PEP 451) and fall-back to find_module (PEP 302)
+    try:
+        spec = importer.find_spec(packageName)
+    except AttributeError:
+        # capture warnings due to #1111
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            loader = importer.find_module(packageName)
+    else:
+        loader = spec.loader if spec else None
+
+    if loader is None:
+        return None
+    module = sys.modules.get(packageName)
+    if module is None:
+        module = sys.modules[packageName] = types.ModuleType(packageName)
+        module.__path__ = []
+        _set_parent_ns(packageName)
+    elif not hasattr(module, '__path__'):
+        raise TypeError("Not a package:", packageName)
+    handler = _find_adapter(_namespace_handlers, importer)
+    subpath = handler(importer, path_item, packageName, module)
+    if subpath is not None:
+        path = module.__path__
+        path.append(subpath)
+        importlib.import_module(packageName)
+        _rebuild_mod_path(path, packageName, module)
+    return subpath
+
+
+def _rebuild_mod_path(orig_path, package_name, module: types.ModuleType):
+    """
+    Rebuild module.__path__ ensuring that all entries are ordered
+    corresponding to their sys.path order
+    """
+    sys_path = [_normalize_cached(p) for p in sys.path]
+
+    def safe_sys_path_index(entry):
+        """
+        Workaround for #520 and #513.
+        """
+        try:
+            return sys_path.index(entry)
+        except ValueError:
+            return float('inf')
+
+    def position_in_sys_path(path):
+        """
+        Return the ordinal of the path based on its position in sys.path
+        """
+        path_parts = path.split(os.sep)
+        module_parts = package_name.count('.') + 1
+        parts = path_parts[:-module_parts]
+        return safe_sys_path_index(_normalize_cached(os.sep.join(parts)))
+
+    new_path = sorted(orig_path, key=position_in_sys_path)
+    new_path = [_normalize_cached(p) for p in new_path]
+
+    if isinstance(module.__path__, list):
+        module.__path__[:] = new_path
+    else:
+        module.__path__ = new_path
+
+
+def declare_namespace(packageName: str):
+    """Declare that package 'packageName' is a namespace package"""
+
+    msg = (
+        f"Deprecated call to `pkg_resources.declare_namespace({packageName!r})`.\n"
+        "Implementing implicit namespace packages (as specified in PEP 420) "
+        "is preferred to `pkg_resources.declare_namespace`. "
+        "See https://setuptools.pypa.io/en/latest/references/"
+        "keywords.html#keyword-namespace-packages"
+    )
+    warnings.warn(msg, DeprecationWarning, stacklevel=2)
+
+    _imp.acquire_lock()
+    try:
+        if packageName in _namespace_packages:
+            return
+
+        path: MutableSequence[str] = sys.path
+        parent, _, _ = packageName.rpartition('.')
+
+        if parent:
+            declare_namespace(parent)
+            if parent not in _namespace_packages:
+                __import__(parent)
+            try:
+                path = sys.modules[parent].__path__
+            except AttributeError as e:
+                raise TypeError("Not a package:", parent) from e
+
+        # Track what packages are namespaces, so when new path items are added,
+        # they can be updated
+        _namespace_packages.setdefault(parent or None, []).append(packageName)
+        _namespace_packages.setdefault(packageName, [])
+
+        for path_item in path:
+            # Ensure all the parent's path items are reflected in the child,
+            # if they apply
+            _handle_ns(packageName, path_item)
+
+    finally:
+        _imp.release_lock()
+
+
+def fixup_namespace_packages(path_item: str, parent: str | None = None):
+    """Ensure that previously-declared namespace packages include path_item"""
+    _imp.acquire_lock()
+    try:
+        for package in _namespace_packages.get(parent, ()):
+            subpath = _handle_ns(package, path_item)
+            if subpath:
+                fixup_namespace_packages(subpath, package)
+    finally:
+        _imp.release_lock()
+
+
+def file_ns_handler(
+    importer: object,
+    path_item: StrPath,
+    packageName: str,
+    module: types.ModuleType,
+):
+    """Compute an ns-package subpath for a filesystem or zipfile importer"""
+
+    subpath = os.path.join(path_item, packageName.split('.')[-1])
+    normalized = _normalize_cached(subpath)
+    for item in module.__path__:
+        if _normalize_cached(item) == normalized:
+            break
+    else:
+        # Only return the path if it's not already there
+        return subpath
+
+
+if hasattr(pkgutil, 'ImpImporter'):
+    register_namespace_handler(pkgutil.ImpImporter, file_ns_handler)
+
+register_namespace_handler(zipimport.zipimporter, file_ns_handler)
+register_namespace_handler(importlib.machinery.FileFinder, file_ns_handler)
+
+
+def null_ns_handler(
+    importer: object,
+    path_item: str | None,
+    packageName: str | None,
+    module: _ModuleLike | None,
+):
+    return None
+
+
+register_namespace_handler(object, null_ns_handler)
+
+
+@overload
+def normalize_path(filename: StrPath) -> str: ...
+@overload
+def normalize_path(filename: BytesPath) -> bytes: ...
+def normalize_path(filename: StrOrBytesPath):
+    """Normalize a file/dir name for comparison purposes"""
+    return os.path.normcase(os.path.realpath(os.path.normpath(_cygwin_patch(filename))))
+
+
+def _cygwin_patch(filename: StrOrBytesPath):  # pragma: nocover
+    """
+    Contrary to POSIX 2008, on Cygwin, getcwd (3) contains
+    symlink components. Using
+    os.path.abspath() works around this limitation. A fix in os.getcwd()
+    would probably better, in Cygwin even more so, except
+    that this seems to be by design...
+    """
+    return os.path.abspath(filename) if sys.platform == 'cygwin' else filename
+
+
+if TYPE_CHECKING:
+    # https://github.com/python/mypy/issues/16261
+    # https://github.com/python/typeshed/issues/6347
+    @overload
+    def _normalize_cached(filename: StrPath) -> str: ...
+    @overload
+    def _normalize_cached(filename: BytesPath) -> bytes: ...
+    def _normalize_cached(filename: StrOrBytesPath) -> str | bytes: ...
+else:
+
+    @functools.lru_cache(maxsize=None)
+    def _normalize_cached(filename):
+        return normalize_path(filename)
+
+
+def _is_egg_path(path):
+    """
+    Determine if given path appears to be an egg.
+    """
+    return _is_zip_egg(path) or _is_unpacked_egg(path)
+
+
+def _is_zip_egg(path):
+    return (
+        path.lower().endswith('.egg')
+        and os.path.isfile(path)
+        and zipfile.is_zipfile(path)
+    )
+
+
+def _is_unpacked_egg(path):
+    """
+    Determine if given path appears to be an unpacked egg.
+    """
+    return path.lower().endswith('.egg') and os.path.isfile(
+        os.path.join(path, 'EGG-INFO', 'PKG-INFO')
+    )
+
+
+def _set_parent_ns(packageName):
+    parts = packageName.split('.')
+    name = parts.pop()
+    if parts:
+        parent = '.'.join(parts)
+        setattr(sys.modules[parent], name, sys.modules[packageName])
+
+
+MODULE = re.compile(r"\w+(\.\w+)*$").match
+EGG_NAME = re.compile(
+    r"""
+    (?P<name>[^-]+) (
+        -(?P<ver>[^-]+) (
+            -py(?P<pyver>[^-]+) (
+                -(?P<plat>.+)
+            )?
+        )?
+    )?
+    """,
+    re.VERBOSE | re.IGNORECASE,
+).match
+
+
+class EntryPoint:
+    """Object representing an advertised importable object"""
+
+    def __init__(
+        self,
+        name: str,
+        module_name: str,
+        attrs: Iterable[str] = (),
+        extras: Iterable[str] = (),
+        dist: Distribution | None = None,
+    ):
+        if not MODULE(module_name):
+            raise ValueError("Invalid module name", module_name)
+        self.name = name
+        self.module_name = module_name
+        self.attrs = tuple(attrs)
+        self.extras = tuple(extras)
+        self.dist = dist
+
+    def __str__(self):
+        s = "%s = %s" % (self.name, self.module_name)
+        if self.attrs:
+            s += ':' + '.'.join(self.attrs)
+        if self.extras:
+            s += ' [%s]' % ','.join(self.extras)
+        return s
+
+    def __repr__(self):
+        return "EntryPoint.parse(%r)" % str(self)
+
+    @overload
+    def load(
+        self,
+        require: Literal[True] = True,
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+    ) -> _ResolvedEntryPoint: ...
+    @overload
+    def load(
+        self,
+        require: Literal[False],
+        *args: Any,
+        **kwargs: Any,
+    ) -> _ResolvedEntryPoint: ...
+    def load(
+        self,
+        require: bool = True,
+        *args: Environment | _InstallerType | None,
+        **kwargs: Environment | _InstallerType | None,
+    ) -> _ResolvedEntryPoint:
+        """
+        Require packages for this EntryPoint, then resolve it.
+        """
+        if not require or args or kwargs:
+            warnings.warn(
+                "Parameters to load are deprecated.  Call .resolve and "
+                ".require separately.",
+                PkgResourcesDeprecationWarning,
+                stacklevel=2,
+            )
+        if require:
+            # We could pass `env` and `installer` directly,
+            # but keeping `*args` and `**kwargs` for backwards compatibility
+            self.require(*args, **kwargs)  # type: ignore
+        return self.resolve()
+
+    def resolve(self) -> _ResolvedEntryPoint:
+        """
+        Resolve the entry point from its module and attrs.
+        """
+        module = __import__(self.module_name, fromlist=['__name__'], level=0)
+        try:
+            return functools.reduce(getattr, self.attrs, module)
+        except AttributeError as exc:
+            raise ImportError(str(exc)) from exc
+
+    def require(
+        self,
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+    ):
+        if not self.dist:
+            error_cls = UnknownExtra if self.extras else AttributeError
+            raise error_cls("Can't require() without a distribution", self)
+
+        # Get the requirements for this entry point with all its extras and
+        # then resolve them. We have to pass `extras` along when resolving so
+        # that the working set knows what extras we want. Otherwise, for
+        # dist-info distributions, the working set will assume that the
+        # requirements for that extra are purely optional and skip over them.
+        reqs = self.dist.requires(self.extras)
+        items = working_set.resolve(reqs, env, installer, extras=self.extras)
+        list(map(working_set.add, items))
+
+    pattern = re.compile(
+        r'\s*'
+        r'(?P<name>.+?)\s*'
+        r'=\s*'
+        r'(?P<module>[\w.]+)\s*'
+        r'(:\s*(?P<attr>[\w.]+))?\s*'
+        r'(?P<extras>\[.*\])?\s*$'
+    )
+
+    @classmethod
+    def parse(cls, src: str, dist: Distribution | None = None):
+        """Parse a single entry point from string `src`
+
+        Entry point syntax follows the form::
+
+            name = some.module:some.attr [extra1, extra2]
+
+        The entry name and module name are required, but the ``:attrs`` and
+        ``[extras]`` parts are optional
+        """
+        m = cls.pattern.match(src)
+        if not m:
+            msg = "EntryPoint must be in 'name=module:attrs [extras]' format"
+            raise ValueError(msg, src)
+        res = m.groupdict()
+        extras = cls._parse_extras(res['extras'])
+        attrs = res['attr'].split('.') if res['attr'] else ()
+        return cls(res['name'], res['module'], attrs, extras, dist)
+
+    @classmethod
+    def _parse_extras(cls, extras_spec):
+        if not extras_spec:
+            return ()
+        req = Requirement.parse('x' + extras_spec)
+        if req.specs:
+            raise ValueError
+        return req.extras
+
+    @classmethod
+    def parse_group(
+        cls,
+        group: str,
+        lines: _NestedStr,
+        dist: Distribution | None = None,
+    ):
+        """Parse an entry point group"""
+        if not MODULE(group):
+            raise ValueError("Invalid group name", group)
+        this: dict[str, Self] = {}
+        for line in yield_lines(lines):
+            ep = cls.parse(line, dist)
+            if ep.name in this:
+                raise ValueError("Duplicate entry point", group, ep.name)
+            this[ep.name] = ep
+        return this
+
+    @classmethod
+    def parse_map(
+        cls,
+        data: str | Iterable[str] | dict[str, str | Iterable[str]],
+        dist: Distribution | None = None,
+    ):
+        """Parse a map of entry point groups"""
+        _data: Iterable[tuple[str | None, str | Iterable[str]]]
+        if isinstance(data, dict):
+            _data = data.items()
+        else:
+            _data = split_sections(data)
+        maps: dict[str, dict[str, Self]] = {}
+        for group, lines in _data:
+            if group is None:
+                if not lines:
+                    continue
+                raise ValueError("Entry points must be listed in groups")
+            group = group.strip()
+            if group in maps:
+                raise ValueError("Duplicate group name", group)
+            maps[group] = cls.parse_group(group, lines, dist)
+        return maps
+
+
+def _version_from_file(lines):
+    """
+    Given an iterable of lines from a Metadata file, return
+    the value of the Version field, if present, or None otherwise.
+    """
+
+    def is_version_line(line):
+        return line.lower().startswith('version:')
+
+    version_lines = filter(is_version_line, lines)
+    line = next(iter(version_lines), '')
+    _, _, value = line.partition(':')
+    return safe_version(value.strip()) or None
+
+
+class Distribution:
+    """Wrap an actual or potential sys.path entry w/metadata"""
+
+    PKG_INFO = 'PKG-INFO'
+
+    def __init__(
+        self,
+        location: str | None = None,
+        metadata: _MetadataType = None,
+        project_name: str | None = None,
+        version: str | None = None,
+        py_version: str | None = PY_MAJOR,
+        platform: str | None = None,
+        precedence: int = EGG_DIST,
+    ):
+        self.project_name = safe_name(project_name or 'Unknown')
+        if version is not None:
+            self._version = safe_version(version)
+        self.py_version = py_version
+        self.platform = platform
+        self.location = location
+        self.precedence = precedence
+        self._provider = metadata or empty_provider
+
+    @classmethod
+    def from_location(
+        cls,
+        location: str,
+        basename: StrPath,
+        metadata: _MetadataType = None,
+        **kw: int,  # We could set `precedence` explicitly, but keeping this as `**kw` for full backwards and subclassing compatibility
+    ) -> Distribution:
+        project_name, version, py_version, platform = [None] * 4
+        basename, ext = os.path.splitext(basename)
+        if ext.lower() in _distributionImpl:
+            cls = _distributionImpl[ext.lower()]
+
+            match = EGG_NAME(basename)
+            if match:
+                project_name, version, py_version, platform = match.group(
+                    'name', 'ver', 'pyver', 'plat'
+                )
+        return cls(
+            location,
+            metadata,
+            project_name=project_name,
+            version=version,
+            py_version=py_version,
+            platform=platform,
+            **kw,
+        )._reload_version()
+
+    def _reload_version(self):
+        return self
+
+    @property
+    def hashcmp(self):
+        return (
+            self._forgiving_parsed_version,
+            self.precedence,
+            self.key,
+            self.location,
+            self.py_version or '',
+            self.platform or '',
+        )
+
+    def __hash__(self):
+        return hash(self.hashcmp)
+
+    def __lt__(self, other: Distribution):
+        return self.hashcmp < other.hashcmp
+
+    def __le__(self, other: Distribution):
+        return self.hashcmp <= other.hashcmp
+
+    def __gt__(self, other: Distribution):
+        return self.hashcmp > other.hashcmp
+
+    def __ge__(self, other: Distribution):
+        return self.hashcmp >= other.hashcmp
+
+    def __eq__(self, other: object):
+        if not isinstance(other, self.__class__):
+            # It's not a Distribution, so they are not equal
+            return False
+        return self.hashcmp == other.hashcmp
+
+    def __ne__(self, other: object):
+        return not self == other
+
+    # These properties have to be lazy so that we don't have to load any
+    # metadata until/unless it's actually needed.  (i.e., some distributions
+    # may not know their name or version without loading PKG-INFO)
+
+    @property
+    def key(self):
+        try:
+            return self._key
+        except AttributeError:
+            self._key = key = self.project_name.lower()
+            return key
+
+    @property
+    def parsed_version(self):
+        if not hasattr(self, "_parsed_version"):
+            try:
+                self._parsed_version = parse_version(self.version)
+            except _packaging_version.InvalidVersion as ex:
+                info = f"(package: {self.project_name})"
+                if hasattr(ex, "add_note"):
+                    ex.add_note(info)  # PEP 678
+                    raise
+                raise _packaging_version.InvalidVersion(f"{str(ex)} {info}") from None
+
+        return self._parsed_version
+
+    @property
+    def _forgiving_parsed_version(self):
+        try:
+            return self.parsed_version
+        except _packaging_version.InvalidVersion as ex:
+            self._parsed_version = parse_version(_forgiving_version(self.version))
+
+            notes = "\n".join(getattr(ex, "__notes__", []))  # PEP 678
+            msg = f"""!!\n\n
+            *************************************************************************
+            {str(ex)}\n{notes}
+
+            This is a long overdue deprecation.
+            For the time being, `pkg_resources` will use `{self._parsed_version}`
+            as a replacement to avoid breaking existing environments,
+            but no future compatibility is guaranteed.
+
+            If you maintain package {self.project_name} you should implement
+            the relevant changes to adequate the project to PEP 440 immediately.
+            *************************************************************************
+            \n\n!!
+            """
+            warnings.warn(msg, DeprecationWarning)
+
+            return self._parsed_version
+
+    @property
+    def version(self):
+        try:
+            return self._version
+        except AttributeError as e:
+            version = self._get_version()
+            if version is None:
+                path = self._get_metadata_path_for_display(self.PKG_INFO)
+                msg = ("Missing 'Version:' header and/or {} file at path: {}").format(
+                    self.PKG_INFO, path
+                )
+                raise ValueError(msg, self) from e
+
+            return version
+
+    @property
+    def _dep_map(self):
+        """
+        A map of extra to its list of (direct) requirements
+        for this distribution, including the null extra.
+        """
+        try:
+            return self.__dep_map
+        except AttributeError:
+            self.__dep_map = self._filter_extras(self._build_dep_map())
+        return self.__dep_map
+
+    @staticmethod
+    def _filter_extras(dm: dict[str | None, list[Requirement]]):
+        """
+        Given a mapping of extras to dependencies, strip off
+        environment markers and filter out any dependencies
+        not matching the markers.
+        """
+        for extra in list(filter(None, dm)):
+            new_extra: str | None = extra
+            reqs = dm.pop(extra)
+            new_extra, _, marker = extra.partition(':')
+            fails_marker = marker and (
+                invalid_marker(marker) or not evaluate_marker(marker)
+            )
+            if fails_marker:
+                reqs = []
+            new_extra = safe_extra(new_extra) or None
+
+            dm.setdefault(new_extra, []).extend(reqs)
+        return dm
+
+    def _build_dep_map(self):
+        dm = {}
+        for name in 'requires.txt', 'depends.txt':
+            for extra, reqs in split_sections(self._get_metadata(name)):
+                dm.setdefault(extra, []).extend(parse_requirements(reqs))
+        return dm
+
+    def requires(self, extras: Iterable[str] = ()):
+        """List of Requirements needed for this distro if `extras` are used"""
+        dm = self._dep_map
+        deps: list[Requirement] = []
+        deps.extend(dm.get(None, ()))
+        for ext in extras:
+            try:
+                deps.extend(dm[safe_extra(ext)])
+            except KeyError as e:
+                raise UnknownExtra(
+                    "%s has no such extra feature %r" % (self, ext)
+                ) from e
+        return deps
+
+    def _get_metadata_path_for_display(self, name):
+        """
+        Return the path to the given metadata file, if available.
+        """
+        try:
+            # We need to access _get_metadata_path() on the provider object
+            # directly rather than through this class's __getattr__()
+            # since _get_metadata_path() is marked private.
+            path = self._provider._get_metadata_path(name)
+
+        # Handle exceptions e.g. in case the distribution's metadata
+        # provider doesn't support _get_metadata_path().
+        except Exception:
+            return '[could not detect]'
+
+        return path
+
+    def _get_metadata(self, name):
+        if self.has_metadata(name):
+            yield from self.get_metadata_lines(name)
+
+    def _get_version(self):
+        lines = self._get_metadata(self.PKG_INFO)
+        return _version_from_file(lines)
+
+    def activate(self, path: list[str] | None = None, replace: bool = False):
+        """Ensure distribution is importable on `path` (default=sys.path)"""
+        if path is None:
+            path = sys.path
+        self.insert_on(path, replace=replace)
+        if path is sys.path and self.location is not None:
+            fixup_namespace_packages(self.location)
+            for pkg in self._get_metadata('namespace_packages.txt'):
+                if pkg in sys.modules:
+                    declare_namespace(pkg)
+
+    def egg_name(self):
+        """Return what this distribution's standard .egg filename should be"""
+        filename = "%s-%s-py%s" % (
+            to_filename(self.project_name),
+            to_filename(self.version),
+            self.py_version or PY_MAJOR,
+        )
+
+        if self.platform:
+            filename += '-' + self.platform
+        return filename
+
+    def __repr__(self):
+        if self.location:
+            return "%s (%s)" % (self, self.location)
+        else:
+            return str(self)
+
+    def __str__(self):
+        try:
+            version = getattr(self, 'version', None)
+        except ValueError:
+            version = None
+        version = version or "[unknown version]"
+        return "%s %s" % (self.project_name, version)
+
+    def __getattr__(self, attr):
+        """Delegate all unrecognized public attributes to .metadata provider"""
+        if attr.startswith('_'):
+            raise AttributeError(attr)
+        return getattr(self._provider, attr)
+
+    def __dir__(self):
+        return list(
+            set(super().__dir__())
+            | set(attr for attr in self._provider.__dir__() if not attr.startswith('_'))
+        )
+
+    @classmethod
+    def from_filename(
+        cls,
+        filename: StrPath,
+        metadata: _MetadataType = None,
+        **kw: int,  # We could set `precedence` explicitly, but keeping this as `**kw` for full backwards and subclassing compatibility
+    ):
+        return cls.from_location(
+            _normalize_cached(filename), os.path.basename(filename), metadata, **kw
+        )
+
+    def as_requirement(self):
+        """Return a ``Requirement`` that matches this distribution exactly"""
+        if isinstance(self.parsed_version, _packaging_version.Version):
+            spec = "%s==%s" % (self.project_name, self.parsed_version)
+        else:
+            spec = "%s===%s" % (self.project_name, self.parsed_version)
+
+        return Requirement.parse(spec)
+
+    def load_entry_point(self, group: str, name: str) -> _ResolvedEntryPoint:
+        """Return the `name` entry point of `group` or raise ImportError"""
+        ep = self.get_entry_info(group, name)
+        if ep is None:
+            raise ImportError("Entry point %r not found" % ((group, name),))
+        return ep.load()
+
+    @overload
+    def get_entry_map(self, group: None = None) -> dict[str, dict[str, EntryPoint]]: ...
+    @overload
+    def get_entry_map(self, group: str) -> dict[str, EntryPoint]: ...
+    def get_entry_map(self, group: str | None = None):
+        """Return the entry point map for `group`, or the full entry map"""
+        if not hasattr(self, "_ep_map"):
+            self._ep_map = EntryPoint.parse_map(
+                self._get_metadata('entry_points.txt'), self
+            )
+        if group is not None:
+            return self._ep_map.get(group, {})
+        return self._ep_map
+
+    def get_entry_info(self, group: str, name: str):
+        """Return the EntryPoint object for `group`+`name`, or ``None``"""
+        return self.get_entry_map(group).get(name)
+
+    # FIXME: 'Distribution.insert_on' is too complex (13)
+    def insert_on(  # noqa: C901
+        self,
+        path: list[str],
+        loc=None,
+        replace: bool = False,
+    ):
+        """Ensure self.location is on path
+
+        If replace=False (default):
+            - If location is already in path anywhere, do nothing.
+            - Else:
+              - If it's an egg and its parent directory is on path,
+                insert just ahead of the parent.
+              - Else: add to the end of path.
+        If replace=True:
+            - If location is already on path anywhere (not eggs)
+              or higher priority than its parent (eggs)
+              do nothing.
+            - Else:
+              - If it's an egg and its parent directory is on path,
+                insert just ahead of the parent,
+                removing any lower-priority entries.
+              - Else: add it to the front of path.
+        """
+
+        loc = loc or self.location
+        if not loc:
+            return
+
+        nloc = _normalize_cached(loc)
+        bdir = os.path.dirname(nloc)
+        npath = [(p and _normalize_cached(p) or p) for p in path]
+
+        for p, item in enumerate(npath):
+            if item == nloc:
+                if replace:
+                    break
+                else:
+                    # don't modify path (even removing duplicates) if
+                    # found and not replace
+                    return
+            elif item == bdir and self.precedence == EGG_DIST:
+                # if it's an .egg, give it precedence over its directory
+                # UNLESS it's already been added to sys.path and replace=False
+                if (not replace) and nloc in npath[p:]:
+                    return
+                if path is sys.path:
+                    self.check_version_conflict()
+                path.insert(p, loc)
+                npath.insert(p, nloc)
+                break
+        else:
+            if path is sys.path:
+                self.check_version_conflict()
+            if replace:
+                path.insert(0, loc)
+            else:
+                path.append(loc)
+            return
+
+        # p is the spot where we found or inserted loc; now remove duplicates
+        while True:
+            try:
+                np = npath.index(nloc, p + 1)
+            except ValueError:
+                break
+            else:
+                del npath[np], path[np]
+                # ha!
+                p = np
+
+        return
+
+    def check_version_conflict(self):
+        if self.key == 'setuptools':
+            # ignore the inevitable setuptools self-conflicts  :(
+            return
+
+        nsp = dict.fromkeys(self._get_metadata('namespace_packages.txt'))
+        loc = normalize_path(self.location)
+        for modname in self._get_metadata('top_level.txt'):
+            if (
+                modname not in sys.modules
+                or modname in nsp
+                or modname in _namespace_packages
+            ):
+                continue
+            if modname in ('pkg_resources', 'setuptools', 'site'):
+                continue
+            fn = getattr(sys.modules[modname], '__file__', None)
+            if fn and (
+                normalize_path(fn).startswith(loc) or fn.startswith(self.location)
+            ):
+                continue
+            issue_warning(
+                "Module %s was already imported from %s, but %s is being added"
+                " to sys.path" % (modname, fn, self.location),
+            )
+
+    def has_version(self):
+        try:
+            self.version
+        except ValueError:
+            issue_warning("Unbuilt egg for " + repr(self))
+            return False
+        except SystemError:
+            # TODO: remove this except clause when python/cpython#103632 is fixed.
+            return False
+        return True
+
+    def clone(self, **kw: str | int | IResourceProvider | None):
+        """Copy this distribution, substituting in any changed keyword args"""
+        names = 'project_name version py_version platform location precedence'
+        for attr in names.split():
+            kw.setdefault(attr, getattr(self, attr, None))
+        kw.setdefault('metadata', self._provider)
+        # Unsafely unpacking. But keeping **kw for backwards and subclassing compatibility
+        return self.__class__(**kw)  # type:ignore[arg-type]
+
+    @property
+    def extras(self):
+        return [dep for dep in self._dep_map if dep]
+
+
+class EggInfoDistribution(Distribution):
+    def _reload_version(self):
+        """
+        Packages installed by distutils (e.g. numpy or scipy),
+        which uses an old safe_version, and so
+        their version numbers can get mangled when
+        converted to filenames (e.g., 1.11.0.dev0+2329eae to
+        1.11.0.dev0_2329eae). These distributions will not be
+        parsed properly
+        downstream by Distribution and safe_version, so
+        take an extra step and try to get the version number from
+        the metadata file itself instead of the filename.
+        """
+        md_version = self._get_version()
+        if md_version:
+            self._version = md_version
+        return self
+
+
+class DistInfoDistribution(Distribution):
+    """
+    Wrap an actual or potential sys.path entry
+    w/metadata, .dist-info style.
+    """
+
+    PKG_INFO = 'METADATA'
+    EQEQ = re.compile(r"([\(,])\s*(\d.*?)\s*([,\)])")
+
+    @property
+    def _parsed_pkg_info(self):
+        """Parse and cache metadata"""
+        try:
+            return self._pkg_info
+        except AttributeError:
+            metadata = self.get_metadata(self.PKG_INFO)
+            self._pkg_info = email.parser.Parser().parsestr(metadata)
+            return self._pkg_info
+
+    @property
+    def _dep_map(self):
+        try:
+            return self.__dep_map
+        except AttributeError:
+            self.__dep_map = self._compute_dependencies()
+            return self.__dep_map
+
+    def _compute_dependencies(self) -> dict[str | None, list[Requirement]]:
+        """Recompute this distribution's dependencies."""
+        self.__dep_map: dict[str | None, list[Requirement]] = {None: []}
+
+        reqs: list[Requirement] = []
+        # Including any condition expressions
+        for req in self._parsed_pkg_info.get_all('Requires-Dist') or []:
+            reqs.extend(parse_requirements(req))
+
+        def reqs_for_extra(extra):
+            for req in reqs:
+                if not req.marker or req.marker.evaluate({'extra': extra}):
+                    yield req
+
+        common = types.MappingProxyType(dict.fromkeys(reqs_for_extra(None)))
+        self.__dep_map[None].extend(common)
+
+        for extra in self._parsed_pkg_info.get_all('Provides-Extra') or []:
+            s_extra = safe_extra(extra.strip())
+            self.__dep_map[s_extra] = [
+                r for r in reqs_for_extra(extra) if r not in common
+            ]
+
+        return self.__dep_map
+
+
+_distributionImpl = {
+    '.egg': Distribution,
+    '.egg-info': EggInfoDistribution,
+    '.dist-info': DistInfoDistribution,
+}
+
+
+def issue_warning(*args, **kw):
+    level = 1
+    g = globals()
+    try:
+        # find the first stack frame that is *not* code in
+        # the pkg_resources module, to use for the warning
+        while sys._getframe(level).f_globals is g:
+            level += 1
+    except ValueError:
+        pass
+    warnings.warn(stacklevel=level + 1, *args, **kw)
+
+
+def parse_requirements(strs: _NestedStr):
+    """
+    Yield ``Requirement`` objects for each specification in `strs`.
+
+    `strs` must be a string, or a (possibly-nested) iterable thereof.
+    """
+    return map(Requirement, join_continuation(map(drop_comment, yield_lines(strs))))
+
+
+class RequirementParseError(_packaging_requirements.InvalidRequirement):
+    "Compatibility wrapper for InvalidRequirement"
+
+
+class Requirement(_packaging_requirements.Requirement):
+    def __init__(self, requirement_string: str):
+        """DO NOT CALL THIS UNDOCUMENTED METHOD; use Requirement.parse()!"""
+        super().__init__(requirement_string)
+        self.unsafe_name = self.name
+        project_name = safe_name(self.name)
+        self.project_name, self.key = project_name, project_name.lower()
+        self.specs = [(spec.operator, spec.version) for spec in self.specifier]
+        # packaging.requirements.Requirement uses a set for its extras. We use a variable-length tuple
+        self.extras: tuple[str] = tuple(map(safe_extra, self.extras))
+        self.hashCmp = (
+            self.key,
+            self.url,
+            self.specifier,
+            frozenset(self.extras),
+            str(self.marker) if self.marker else None,
+        )
+        self.__hash = hash(self.hashCmp)
+
+    def __eq__(self, other: object):
+        return isinstance(other, Requirement) and self.hashCmp == other.hashCmp
+
+    def __ne__(self, other):
+        return not self == other
+
+    def __contains__(self, item: Distribution | str | tuple[str, ...]) -> bool:
+        if isinstance(item, Distribution):
+            if item.key != self.key:
+                return False
+
+            item = item.version
+
+        # Allow prereleases always in order to match the previous behavior of
+        # this method. In the future this should be smarter and follow PEP 440
+        # more accurately.
+        return self.specifier.contains(item, prereleases=True)
+
+    def __hash__(self):
+        return self.__hash
+
+    def __repr__(self):
+        return "Requirement.parse(%r)" % str(self)
+
+    @staticmethod
+    def parse(s: str | Iterable[str]):
+        (req,) = parse_requirements(s)
+        return req
+
+
+def _always_object(classes):
+    """
+    Ensure object appears in the mro even
+    for old-style classes.
+    """
+    if object not in classes:
+        return classes + (object,)
+    return classes
+
+
+def _find_adapter(registry: Mapping[type, _AdapterT], ob: object) -> _AdapterT:
+    """Return an adapter factory for `ob` from `registry`"""
+    types = _always_object(inspect.getmro(getattr(ob, '__class__', type(ob))))
+    for t in types:
+        if t in registry:
+            return registry[t]
+    # _find_adapter would previously return None, and immediately be called.
+    # So we're raising a TypeError to keep backward compatibility if anyone depended on that behaviour.
+    raise TypeError(f"Could not find adapter for {registry} and {ob}")
+
+
+def ensure_directory(path: StrOrBytesPath):
+    """Ensure that the parent directory of `path` exists"""
+    dirname = os.path.dirname(path)
+    os.makedirs(dirname, exist_ok=True)
+
+
+def _bypass_ensure_directory(path):
+    """Sandbox-bypassing version of ensure_directory()"""
+    if not WRITE_SUPPORT:
+        raise OSError('"os.mkdir" not supported on this platform.')
+    dirname, filename = split(path)
+    if dirname and filename and not isdir(dirname):
+        _bypass_ensure_directory(dirname)
+        try:
+            mkdir(dirname, 0o755)
+        except FileExistsError:
+            pass
+
+
+def split_sections(s: _NestedStr) -> Iterator[tuple[str | None, list[str]]]:
+    """Split a string or iterable thereof into (section, content) pairs
+
+    Each ``section`` is a stripped version of the section header ("[section]")
+    and each ``content`` is a list of stripped lines excluding blank lines and
+    comment-only lines.  If there are any such lines before the first section
+    header, they're returned in a first ``section`` of ``None``.
+    """
+    section = None
+    content = []
+    for line in yield_lines(s):
+        if line.startswith("["):
+            if line.endswith("]"):
+                if section or content:
+                    yield section, content
+                section = line[1:-1].strip()
+                content = []
+            else:
+                raise ValueError("Invalid section heading", line)
+        else:
+            content.append(line)
+
+    # wrap up last segment
+    yield section, content
+
+
+def _mkstemp(*args, **kw):
+    old_open = os.open
+    try:
+        # temporarily bypass sandboxing
+        os.open = os_open
+        return tempfile.mkstemp(*args, **kw)
+    finally:
+        # and then put it back
+        os.open = old_open
+
+
+# Silence the PEP440Warning by default, so that end users don't get hit by it
+# randomly just because they use pkg_resources. We want to append the rule
+# because we want earlier uses of filterwarnings to take precedence over this
+# one.
+warnings.filterwarnings("ignore", category=PEP440Warning, append=True)
+
+
+class PkgResourcesDeprecationWarning(Warning):
+    """
+    Base class for warning about deprecations in ``pkg_resources``
+
+    This class is not derived from ``DeprecationWarning``, and as such is
+    visible by default.
+    """
+
+
+# Ported from ``setuptools`` to avoid introducing an import inter-dependency:
+_LOCALE_ENCODING = "locale" if sys.version_info >= (3, 10) else None
+
+
+def _read_utf8_with_fallback(file: str, fallback_encoding=_LOCALE_ENCODING) -> str:
+    """See setuptools.unicode_utils._read_utf8_with_fallback"""
+    try:
+        with open(file, "r", encoding="utf-8") as f:
+            return f.read()
+    except UnicodeDecodeError:  # pragma: no cover
+        msg = f"""\
+        ********************************************************************************
+        `encoding="utf-8"` fails with {file!r}, trying `encoding={fallback_encoding!r}`.
+
+        This fallback behaviour is considered **deprecated** and future versions of
+        `setuptools/pkg_resources` may not implement it.
+
+        Please encode {file!r} with "utf-8" to ensure future builds will succeed.
+
+        If this file was produced by `setuptools` itself, cleaning up the cached files
+        and re-building/re-installing the package with a newer version of `setuptools`
+        (e.g. by updating `build-system.requires` in its `pyproject.toml`)
+        might solve the problem.
+        ********************************************************************************
+        """
+        # TODO: Add a deadline?
+        #       See comment in setuptools.unicode_utils._Utf8EncodingNeeded
+        warnings.warn(msg, PkgResourcesDeprecationWarning, stacklevel=2)
+        with open(file, "r", encoding=fallback_encoding) as f:
+            return f.read()
+
+
+# from jaraco.functools 1.3
+def _call_aside(f, *args, **kwargs):
+    f(*args, **kwargs)
+    return f
+
+
+@_call_aside
+def _initialize(g=globals()):
+    "Set up global resource manager (deliberately not state-saved)"
+    manager = ResourceManager()
+    g['_manager'] = manager
+    g.update(
+        (name, getattr(manager, name))
+        for name in dir(manager)
+        if not name.startswith('_')
+    )
+
+
+@_call_aside
+def _initialize_master_working_set():
+    """
+    Prepare the master working set and make the ``require()``
+    API available.
+
+    This function has explicit effects on the global state
+    of pkg_resources. It is intended to be invoked once at
+    the initialization of this module.
+
+    Invocation by other packages is unsupported and done
+    at their own risk.
+    """
+    working_set = _declare_state('object', 'working_set', WorkingSet._build_master())
+
+    require = working_set.require
+    iter_entry_points = working_set.iter_entry_points
+    add_activation_listener = working_set.subscribe
+    run_script = working_set.run_script
+    # backward compatibility
+    run_main = run_script
+    # Activate all distributions already on sys.path with replace=False and
+    # ensure that all distributions added to the working set in the future
+    # (e.g. by calling ``require()``) will get activated as well,
+    # with higher priority (replace=True).
+    tuple(dist.activate(replace=False) for dist in working_set)
+    add_activation_listener(
+        lambda dist: dist.activate(replace=True),
+        existing=False,
+    )
+    working_set.entries = []
+    # match order
+    list(map(working_set.add_entry, sys.path))
+    globals().update(locals())
+
+
+if TYPE_CHECKING:
+    # All of these are set by the @_call_aside methods above
+    __resource_manager = ResourceManager()  # Won't exist at runtime
+    resource_exists = __resource_manager.resource_exists
+    resource_isdir = __resource_manager.resource_isdir
+    resource_filename = __resource_manager.resource_filename
+    resource_stream = __resource_manager.resource_stream
+    resource_string = __resource_manager.resource_string
+    resource_listdir = __resource_manager.resource_listdir
+    set_extraction_path = __resource_manager.set_extraction_path
+    cleanup_resources = __resource_manager.cleanup_resources
+
+    working_set = WorkingSet()
+    require = working_set.require
+    iter_entry_points = working_set.iter_entry_points
+    add_activation_listener = working_set.subscribe
+    run_script = working_set.run_script
+    run_main = run_script
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..86024eec
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pkg_resources/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__init__.py
new file mode 100644
index 00000000..edc21fad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__init__.py
@@ -0,0 +1,631 @@
+"""
+Utilities for determining application-specific dirs.
+
+See <https://github.com/platformdirs/platformdirs> for details and usage.
+
+"""
+
+from __future__ import annotations
+
+import os
+import sys
+from typing import TYPE_CHECKING
+
+from .api import PlatformDirsABC
+from .version import __version__
+from .version import __version_tuple__ as __version_info__
+
+if TYPE_CHECKING:
+    from pathlib import Path
+    from typing import Literal
+
+if sys.platform == "win32":
+    from pip._vendor.platformdirs.windows import Windows as _Result
+elif sys.platform == "darwin":
+    from pip._vendor.platformdirs.macos import MacOS as _Result
+else:
+    from pip._vendor.platformdirs.unix import Unix as _Result
+
+
+def _set_platform_dir_class() -> type[PlatformDirsABC]:
+    if os.getenv("ANDROID_DATA") == "/data" and os.getenv("ANDROID_ROOT") == "/system":
+        if os.getenv("SHELL") or os.getenv("PREFIX"):
+            return _Result
+
+        from pip._vendor.platformdirs.android import _android_folder  # noqa: PLC0415
+
+        if _android_folder() is not None:
+            from pip._vendor.platformdirs.android import Android  # noqa: PLC0415
+
+            return Android  # return to avoid redefinition of a result
+
+    return _Result
+
+
+if TYPE_CHECKING:
+    # Work around mypy issue: https://github.com/python/mypy/issues/10962
+    PlatformDirs = _Result
+else:
+    PlatformDirs = _set_platform_dir_class()  #: Currently active platform
+AppDirs = PlatformDirs  #: Backwards compatibility with appdirs
+
+
+def user_data_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_data_dir
+
+
+def site_data_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data directory shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_data_dir
+
+
+def user_config_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_config_dir
+
+
+def site_config_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config directory shared by the users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_config_dir
+
+
+def user_cache_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_cache_dir
+
+
+def site_cache_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_cache_dir
+
+
+def user_state_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: state directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_state_dir
+
+
+def user_log_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: log directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_log_dir
+
+
+def user_documents_dir() -> str:
+    """:returns: documents directory tied to the user"""
+    return PlatformDirs().user_documents_dir
+
+
+def user_downloads_dir() -> str:
+    """:returns: downloads directory tied to the user"""
+    return PlatformDirs().user_downloads_dir
+
+
+def user_pictures_dir() -> str:
+    """:returns: pictures directory tied to the user"""
+    return PlatformDirs().user_pictures_dir
+
+
+def user_videos_dir() -> str:
+    """:returns: videos directory tied to the user"""
+    return PlatformDirs().user_videos_dir
+
+
+def user_music_dir() -> str:
+    """:returns: music directory tied to the user"""
+    return PlatformDirs().user_music_dir
+
+
+def user_desktop_dir() -> str:
+    """:returns: desktop directory tied to the user"""
+    return PlatformDirs().user_desktop_dir
+
+
+def user_runtime_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_runtime_dir
+
+
+def site_runtime_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime directory shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_runtime_dir
+
+
+def user_data_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_data_path
+
+
+def site_data_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `multipath <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data path shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_data_path
+
+
+def user_config_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_config_path
+
+
+def site_config_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config path shared by the users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_config_path
+
+
+def site_cache_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_cache_path
+
+
+def user_cache_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_cache_path
+
+
+def user_state_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: state path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_state_path
+
+
+def user_log_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: log path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_log_path
+
+
+def user_documents_path() -> Path:
+    """:returns: documents a path tied to the user"""
+    return PlatformDirs().user_documents_path
+
+
+def user_downloads_path() -> Path:
+    """:returns: downloads path tied to the user"""
+    return PlatformDirs().user_downloads_path
+
+
+def user_pictures_path() -> Path:
+    """:returns: pictures path tied to the user"""
+    return PlatformDirs().user_pictures_path
+
+
+def user_videos_path() -> Path:
+    """:returns: videos path tied to the user"""
+    return PlatformDirs().user_videos_path
+
+
+def user_music_path() -> Path:
+    """:returns: music path tied to the user"""
+    return PlatformDirs().user_music_path
+
+
+def user_desktop_path() -> Path:
+    """:returns: desktop path tied to the user"""
+    return PlatformDirs().user_desktop_path
+
+
+def user_runtime_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_runtime_path
+
+
+def site_runtime_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime path shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_runtime_path
+
+
+__all__ = [
+    "AppDirs",
+    "PlatformDirs",
+    "PlatformDirsABC",
+    "__version__",
+    "__version_info__",
+    "site_cache_dir",
+    "site_cache_path",
+    "site_config_dir",
+    "site_config_path",
+    "site_data_dir",
+    "site_data_path",
+    "site_runtime_dir",
+    "site_runtime_path",
+    "user_cache_dir",
+    "user_cache_path",
+    "user_config_dir",
+    "user_config_path",
+    "user_data_dir",
+    "user_data_path",
+    "user_desktop_dir",
+    "user_desktop_path",
+    "user_documents_dir",
+    "user_documents_path",
+    "user_downloads_dir",
+    "user_downloads_path",
+    "user_log_dir",
+    "user_log_path",
+    "user_music_dir",
+    "user_music_path",
+    "user_pictures_dir",
+    "user_pictures_path",
+    "user_runtime_dir",
+    "user_runtime_path",
+    "user_state_dir",
+    "user_state_path",
+    "user_videos_dir",
+    "user_videos_path",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__main__.py
new file mode 100644
index 00000000..fa8a677a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__main__.py
@@ -0,0 +1,55 @@
+"""Main entry point."""
+
+from __future__ import annotations
+
+from pip._vendor.platformdirs import PlatformDirs, __version__
+
+PROPS = (
+    "user_data_dir",
+    "user_config_dir",
+    "user_cache_dir",
+    "user_state_dir",
+    "user_log_dir",
+    "user_documents_dir",
+    "user_downloads_dir",
+    "user_pictures_dir",
+    "user_videos_dir",
+    "user_music_dir",
+    "user_runtime_dir",
+    "site_data_dir",
+    "site_config_dir",
+    "site_cache_dir",
+    "site_runtime_dir",
+)
+
+
+def main() -> None:
+    """Run the main entry point."""
+    app_name = "MyApp"
+    app_author = "MyCompany"
+
+    print(f"-- platformdirs {__version__} --")  # noqa: T201
+
+    print("-- app dirs (with optional 'version')")  # noqa: T201
+    dirs = PlatformDirs(app_name, app_author, version="1.0")
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (without optional 'version')")  # noqa: T201
+    dirs = PlatformDirs(app_name, app_author)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (without optional 'appauthor')")  # noqa: T201
+    dirs = PlatformDirs(app_name)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (with disabled 'appauthor')")  # noqa: T201
+    dirs = PlatformDirs(app_name, appauthor=False)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..5f229eb1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..5bfa4953
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/android.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/android.cpython-312.pyc
new file mode 100644
index 00000000..ac4e79a7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/android.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..babb953a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc
new file mode 100644
index 00000000..c6bf6d27
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc
new file mode 100644
index 00000000..a468d5f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..ffaf3871
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc
new file mode 100644
index 00000000..2be8f9c1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/android.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/android.py
new file mode 100644
index 00000000..7004a852
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/android.py
@@ -0,0 +1,249 @@
+"""Android."""
+
+from __future__ import annotations
+
+import os
+import re
+import sys
+from functools import lru_cache
+from typing import TYPE_CHECKING, cast
+
+from .api import PlatformDirsABC
+
+
+class Android(PlatformDirsABC):
+    """
+    Follows the guidance `from here <https://android.stackexchange.com/a/216132>`_.
+
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`, `version
+    <platformdirs.api.PlatformDirsABC.version>`, `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user, e.g. ``/data/user/<userid>/<packagename>/files/<AppName>``"""
+        return self._append_app_name_and_version(cast(str, _android_folder()), "files")
+
+    @property
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_config_dir(self) -> str:
+        """
+        :return: config directory tied to the user, e.g. \
+        ``/data/user/<userid>/<packagename>/shared_prefs/<AppName>``
+        """
+        return self._append_app_name_and_version(cast(str, _android_folder()), "shared_prefs")
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `user_config_dir`"""
+        return self.user_config_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user, e.g.,``/data/user/<userid>/<packagename>/cache/<AppName>``"""
+        return self._append_app_name_and_version(cast(str, _android_folder()), "cache")
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, same as `user_cache_dir`"""
+        return self.user_cache_dir
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """
+        :return: log directory tied to the user, same as `user_cache_dir` if not opinionated else ``log`` in it,
+          e.g. ``/data/user/<userid>/<packagename>/cache/<AppName>/log``
+        """
+        path = self.user_cache_dir
+        if self.opinion:
+            path = os.path.join(path, "log")  # noqa: PTH118
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user e.g. ``/storage/emulated/0/Documents``"""
+        return _android_documents_folder()
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user e.g. ``/storage/emulated/0/Downloads``"""
+        return _android_downloads_folder()
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user e.g. ``/storage/emulated/0/Pictures``"""
+        return _android_pictures_folder()
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user e.g. ``/storage/emulated/0/DCIM/Camera``"""
+        return _android_videos_folder()
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user e.g. ``/storage/emulated/0/Music``"""
+        return _android_music_folder()
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user e.g. ``/storage/emulated/0/Desktop``"""
+        return "/storage/emulated/0/Desktop"
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, same as `user_cache_dir` if not opinionated else ``tmp`` in it,
+          e.g. ``/data/user/<userid>/<packagename>/cache/<AppName>/tmp``
+        """
+        path = self.user_cache_dir
+        if self.opinion:
+            path = os.path.join(path, "tmp")  # noqa: PTH118
+        return path
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+@lru_cache(maxsize=1)
+def _android_folder() -> str | None:  # noqa: C901
+    """:return: base folder for the Android OS or None if it cannot be found"""
+    result: str | None = None
+    # type checker isn't happy with our "import android", just don't do this when type checking see
+    # https://stackoverflow.com/a/61394121
+    if not TYPE_CHECKING:
+        try:
+            # First try to get a path to android app using python4android (if available)...
+            from android import mActivity  # noqa: PLC0415
+
+            context = cast("android.content.Context", mActivity.getApplicationContext())  # noqa: F821
+            result = context.getFilesDir().getParentFile().getAbsolutePath()
+        except Exception:  # noqa: BLE001
+            result = None
+    if result is None:
+        try:
+            # ...and fall back to using plain pyjnius, if python4android isn't available or doesn't deliver any useful
+            # result...
+            from jnius import autoclass  # noqa: PLC0415
+
+            context = autoclass("android.content.Context")
+            result = context.getFilesDir().getParentFile().getAbsolutePath()
+        except Exception:  # noqa: BLE001
+            result = None
+    if result is None:
+        # and if that fails, too, find an android folder looking at path on the sys.path
+        # warning: only works for apps installed under /data, not adopted storage etc.
+        pattern = re.compile(r"/data/(data|user/\d+)/(.+)/files")
+        for path in sys.path:
+            if pattern.match(path):
+                result = path.split("/files")[0]
+                break
+        else:
+            result = None
+    if result is None:
+        # one last try: find an android folder looking at path on the sys.path taking adopted storage paths into
+        # account
+        pattern = re.compile(r"/mnt/expand/[a-fA-F0-9-]{36}/(data|user/\d+)/(.+)/files")
+        for path in sys.path:
+            if pattern.match(path):
+                result = path.split("/files")[0]
+                break
+        else:
+            result = None
+    return result
+
+
+@lru_cache(maxsize=1)
+def _android_documents_folder() -> str:
+    """:return: documents folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        documents_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DOCUMENTS).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        documents_dir = "/storage/emulated/0/Documents"
+
+    return documents_dir
+
+
+@lru_cache(maxsize=1)
+def _android_downloads_folder() -> str:
+    """:return: downloads folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        downloads_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DOWNLOADS).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        downloads_dir = "/storage/emulated/0/Downloads"
+
+    return downloads_dir
+
+
+@lru_cache(maxsize=1)
+def _android_pictures_folder() -> str:
+    """:return: pictures folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        pictures_dir: str = context.getExternalFilesDir(environment.DIRECTORY_PICTURES).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        pictures_dir = "/storage/emulated/0/Pictures"
+
+    return pictures_dir
+
+
+@lru_cache(maxsize=1)
+def _android_videos_folder() -> str:
+    """:return: videos folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        videos_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DCIM).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        videos_dir = "/storage/emulated/0/DCIM/Camera"
+
+    return videos_dir
+
+
+@lru_cache(maxsize=1)
+def _android_music_folder() -> str:
+    """:return: music folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        music_dir: str = context.getExternalFilesDir(environment.DIRECTORY_MUSIC).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        music_dir = "/storage/emulated/0/Music"
+
+    return music_dir
+
+
+__all__ = [
+    "Android",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/api.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/api.py
new file mode 100644
index 00000000..18d660e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/api.py
@@ -0,0 +1,298 @@
+"""Base API."""
+
+from __future__ import annotations
+
+import os
+from abc import ABC, abstractmethod
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing import Iterator, Literal
+
+
+class PlatformDirsABC(ABC):  # noqa: PLR0904
+    """Abstract base class for platform directories."""
+
+    def __init__(  # noqa: PLR0913, PLR0917
+        self,
+        appname: str | None = None,
+        appauthor: str | None | Literal[False] = None,
+        version: str | None = None,
+        roaming: bool = False,  # noqa: FBT001, FBT002
+        multipath: bool = False,  # noqa: FBT001, FBT002
+        opinion: bool = True,  # noqa: FBT001, FBT002
+        ensure_exists: bool = False,  # noqa: FBT001, FBT002
+    ) -> None:
+        """
+        Create a new platform directory.
+
+        :param appname: See `appname`.
+        :param appauthor: See `appauthor`.
+        :param version: See `version`.
+        :param roaming: See `roaming`.
+        :param multipath: See `multipath`.
+        :param opinion: See `opinion`.
+        :param ensure_exists: See `ensure_exists`.
+
+        """
+        self.appname = appname  #: The name of application.
+        self.appauthor = appauthor
+        """
+        The name of the app author or distributing body for this application.
+
+        Typically, it is the owning company name. Defaults to `appname`. You may pass ``False`` to disable it.
+
+        """
+        self.version = version
+        """
+        An optional version path element to append to the path.
+
+        You might want to use this if you want multiple versions of your app to be able to run independently. If used,
+        this would typically be ``<major>.<minor>``.
+
+        """
+        self.roaming = roaming
+        """
+        Whether to use the roaming appdata directory on Windows.
+
+        That means that for users on a Windows network setup for roaming profiles, this user data will be synced on
+        login (see
+        `here <https://technet.microsoft.com/en-us/library/cc766489(WS.10).aspx>`_).
+
+        """
+        self.multipath = multipath
+        """
+        An optional parameter which indicates that the entire list of data dirs should be returned.
+
+        By default, the first item would only be returned.
+
+        """
+        self.opinion = opinion  #: A flag to indicating to use opinionated values.
+        self.ensure_exists = ensure_exists
+        """
+        Optionally create the directory (and any missing parents) upon access if it does not exist.
+
+        By default, no directories are created.
+
+        """
+
+    def _append_app_name_and_version(self, *base: str) -> str:
+        params = list(base[1:])
+        if self.appname:
+            params.append(self.appname)
+            if self.version:
+                params.append(self.version)
+        path = os.path.join(base[0], *params)  # noqa: PTH118
+        self._optionally_create_directory(path)
+        return path
+
+    def _optionally_create_directory(self, path: str) -> None:
+        if self.ensure_exists:
+            Path(path).mkdir(parents=True, exist_ok=True)
+
+    def _first_item_as_path_if_multipath(self, directory: str) -> Path:
+        if self.multipath:
+            # If multipath is True, the first path is returned.
+            directory = directory.split(os.pathsep)[0]
+        return Path(directory)
+
+    @property
+    @abstractmethod
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users"""
+
+    @property
+    @abstractmethod
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users"""
+
+    @property
+    @abstractmethod
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users"""
+
+    @property
+    @abstractmethod
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_runtime_dir(self) -> str:
+        """:return: runtime directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users"""
+
+    @property
+    def user_data_path(self) -> Path:
+        """:return: data path tied to the user"""
+        return Path(self.user_data_dir)
+
+    @property
+    def site_data_path(self) -> Path:
+        """:return: data path shared by users"""
+        return Path(self.site_data_dir)
+
+    @property
+    def user_config_path(self) -> Path:
+        """:return: config path tied to the user"""
+        return Path(self.user_config_dir)
+
+    @property
+    def site_config_path(self) -> Path:
+        """:return: config path shared by the users"""
+        return Path(self.site_config_dir)
+
+    @property
+    def user_cache_path(self) -> Path:
+        """:return: cache path tied to the user"""
+        return Path(self.user_cache_dir)
+
+    @property
+    def site_cache_path(self) -> Path:
+        """:return: cache path shared by users"""
+        return Path(self.site_cache_dir)
+
+    @property
+    def user_state_path(self) -> Path:
+        """:return: state path tied to the user"""
+        return Path(self.user_state_dir)
+
+    @property
+    def user_log_path(self) -> Path:
+        """:return: log path tied to the user"""
+        return Path(self.user_log_dir)
+
+    @property
+    def user_documents_path(self) -> Path:
+        """:return: documents a path tied to the user"""
+        return Path(self.user_documents_dir)
+
+    @property
+    def user_downloads_path(self) -> Path:
+        """:return: downloads path tied to the user"""
+        return Path(self.user_downloads_dir)
+
+    @property
+    def user_pictures_path(self) -> Path:
+        """:return: pictures path tied to the user"""
+        return Path(self.user_pictures_dir)
+
+    @property
+    def user_videos_path(self) -> Path:
+        """:return: videos path tied to the user"""
+        return Path(self.user_videos_dir)
+
+    @property
+    def user_music_path(self) -> Path:
+        """:return: music path tied to the user"""
+        return Path(self.user_music_dir)
+
+    @property
+    def user_desktop_path(self) -> Path:
+        """:return: desktop path tied to the user"""
+        return Path(self.user_desktop_dir)
+
+    @property
+    def user_runtime_path(self) -> Path:
+        """:return: runtime path tied to the user"""
+        return Path(self.user_runtime_dir)
+
+    @property
+    def site_runtime_path(self) -> Path:
+        """:return: runtime path shared by users"""
+        return Path(self.site_runtime_dir)
+
+    def iter_config_dirs(self) -> Iterator[str]:
+        """:yield: all user and site configuration directories."""
+        yield self.user_config_dir
+        yield self.site_config_dir
+
+    def iter_data_dirs(self) -> Iterator[str]:
+        """:yield: all user and site data directories."""
+        yield self.user_data_dir
+        yield self.site_data_dir
+
+    def iter_cache_dirs(self) -> Iterator[str]:
+        """:yield: all user and site cache directories."""
+        yield self.user_cache_dir
+        yield self.site_cache_dir
+
+    def iter_runtime_dirs(self) -> Iterator[str]:
+        """:yield: all user and site runtime directories."""
+        yield self.user_runtime_dir
+        yield self.site_runtime_dir
+
+    def iter_config_paths(self) -> Iterator[Path]:
+        """:yield: all user and site configuration paths."""
+        for path in self.iter_config_dirs():
+            yield Path(path)
+
+    def iter_data_paths(self) -> Iterator[Path]:
+        """:yield: all user and site data paths."""
+        for path in self.iter_data_dirs():
+            yield Path(path)
+
+    def iter_cache_paths(self) -> Iterator[Path]:
+        """:yield: all user and site cache paths."""
+        for path in self.iter_cache_dirs():
+            yield Path(path)
+
+    def iter_runtime_paths(self) -> Iterator[Path]:
+        """:yield: all user and site runtime paths."""
+        for path in self.iter_runtime_dirs():
+            yield Path(path)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/macos.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/macos.py
new file mode 100644
index 00000000..e4b0391a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/macos.py
@@ -0,0 +1,144 @@
+"""macOS."""
+
+from __future__ import annotations
+
+import os.path
+import sys
+from typing import TYPE_CHECKING
+
+from .api import PlatformDirsABC
+
+if TYPE_CHECKING:
+    from pathlib import Path
+
+
+class MacOS(PlatformDirsABC):
+    """
+    Platform directories for the macOS operating system.
+
+    Follows the guidance from
+    `Apple documentation <https://developer.apple.com/library/archive/documentation/FileManagement/Conceptual/FileSystemProgrammingGuide/MacOSXDirectories/MacOSXDirectories.html>`_.
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`,
+    `version <platformdirs.api.PlatformDirsABC.version>`,
+    `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user, e.g. ``~/Library/Application Support/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Application Support"))  # noqa: PTH111
+
+    @property
+    def site_data_dir(self) -> str:
+        """
+        :return: data directory shared by users, e.g. ``/Library/Application Support/$appname/$version``.
+          If we're using a Python binary managed by `Homebrew <https://brew.sh>`_, the directory
+          will be under the Homebrew prefix, e.g. ``/opt/homebrew/share/$appname/$version``.
+          If `multipath <platformdirs.api.PlatformDirsABC.multipath>` is enabled, and we're in Homebrew,
+          the response is a multi-path string separated by ":", e.g.
+          ``/opt/homebrew/share/$appname/$version:/Library/Application Support/$appname/$version``
+        """
+        is_homebrew = sys.prefix.startswith("/opt/homebrew")
+        path_list = [self._append_app_name_and_version("/opt/homebrew/share")] if is_homebrew else []
+        path_list.append(self._append_app_name_and_version("/Library/Application Support"))
+        if self.multipath:
+            return os.pathsep.join(path_list)
+        return path_list[0]
+
+    @property
+    def site_data_path(self) -> Path:
+        """:return: data path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_data_dir)
+
+    @property
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `site_data_dir`"""
+        return self.site_data_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user, e.g. ``~/Library/Caches/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Caches"))  # noqa: PTH111
+
+    @property
+    def site_cache_dir(self) -> str:
+        """
+        :return: cache directory shared by users, e.g. ``/Library/Caches/$appname/$version``.
+          If we're using a Python binary managed by `Homebrew <https://brew.sh>`_, the directory
+          will be under the Homebrew prefix, e.g. ``/opt/homebrew/var/cache/$appname/$version``.
+          If `multipath <platformdirs.api.PlatformDirsABC.multipath>` is enabled, and we're in Homebrew,
+          the response is a multi-path string separated by ":", e.g.
+          ``/opt/homebrew/var/cache/$appname/$version:/Library/Caches/$appname/$version``
+        """
+        is_homebrew = sys.prefix.startswith("/opt/homebrew")
+        path_list = [self._append_app_name_and_version("/opt/homebrew/var/cache")] if is_homebrew else []
+        path_list.append(self._append_app_name_and_version("/Library/Caches"))
+        if self.multipath:
+            return os.pathsep.join(path_list)
+        return path_list[0]
+
+    @property
+    def site_cache_path(self) -> Path:
+        """:return: cache path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_cache_dir)
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, e.g. ``~/Library/Logs/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Logs"))  # noqa: PTH111
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user, e.g. ``~/Documents``"""
+        return os.path.expanduser("~/Documents")  # noqa: PTH111
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user, e.g. ``~/Downloads``"""
+        return os.path.expanduser("~/Downloads")  # noqa: PTH111
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user, e.g. ``~/Pictures``"""
+        return os.path.expanduser("~/Pictures")  # noqa: PTH111
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user, e.g. ``~/Movies``"""
+        return os.path.expanduser("~/Movies")  # noqa: PTH111
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user, e.g. ``~/Music``"""
+        return os.path.expanduser("~/Music")  # noqa: PTH111
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``~/Desktop``"""
+        return os.path.expanduser("~/Desktop")  # noqa: PTH111
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """:return: runtime directory tied to the user, e.g. ``~/Library/Caches/TemporaryItems/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Caches/TemporaryItems"))  # noqa: PTH111
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+__all__ = [
+    "MacOS",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/unix.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/unix.py
new file mode 100644
index 00000000..f1942e92
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/unix.py
@@ -0,0 +1,269 @@
+"""Unix."""
+
+from __future__ import annotations
+
+import os
+import sys
+from configparser import ConfigParser
+from pathlib import Path
+from typing import Iterator, NoReturn
+
+from .api import PlatformDirsABC
+
+if sys.platform == "win32":
+
+    def getuid() -> NoReturn:
+        msg = "should only be used on Unix"
+        raise RuntimeError(msg)
+
+else:
+    from os import getuid
+
+
+class Unix(PlatformDirsABC):  # noqa: PLR0904
+    """
+    On Unix/Linux, we follow the `XDG Basedir Spec <https://specifications.freedesktop.org/basedir-spec/basedir-spec-
+    latest.html>`_.
+
+    The spec allows overriding directories with environment variables. The examples shown are the default values,
+    alongside the name of the environment variable that overrides them. Makes use of the `appname
+    <platformdirs.api.PlatformDirsABC.appname>`, `version <platformdirs.api.PlatformDirsABC.version>`, `multipath
+    <platformdirs.api.PlatformDirsABC.multipath>`, `opinion <platformdirs.api.PlatformDirsABC.opinion>`, `ensure_exists
+    <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """
+        :return: data directory tied to the user, e.g. ``~/.local/share/$appname/$version`` or
+         ``$XDG_DATA_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_DATA_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.local/share")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def _site_data_dirs(self) -> list[str]:
+        path = os.environ.get("XDG_DATA_DIRS", "")
+        if not path.strip():
+            path = f"/usr/local/share{os.pathsep}/usr/share"
+        return [self._append_app_name_and_version(p) for p in path.split(os.pathsep)]
+
+    @property
+    def site_data_dir(self) -> str:
+        """
+        :return: data directories shared by users (if `multipath <platformdirs.api.PlatformDirsABC.multipath>` is
+         enabled and ``XDG_DATA_DIRS`` is set and a multi path the response is also a multi path separated by the
+         OS path separator), e.g. ``/usr/local/share/$appname/$version`` or ``/usr/share/$appname/$version``
+        """
+        # XDG default for $XDG_DATA_DIRS; only first, if multipath is False
+        dirs = self._site_data_dirs
+        if not self.multipath:
+            return dirs[0]
+        return os.pathsep.join(dirs)
+
+    @property
+    def user_config_dir(self) -> str:
+        """
+        :return: config directory tied to the user, e.g. ``~/.config/$appname/$version`` or
+         ``$XDG_CONFIG_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_CONFIG_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.config")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def _site_config_dirs(self) -> list[str]:
+        path = os.environ.get("XDG_CONFIG_DIRS", "")
+        if not path.strip():
+            path = "/etc/xdg"
+        return [self._append_app_name_and_version(p) for p in path.split(os.pathsep)]
+
+    @property
+    def site_config_dir(self) -> str:
+        """
+        :return: config directories shared by users (if `multipath <platformdirs.api.PlatformDirsABC.multipath>`
+         is enabled and ``XDG_CONFIG_DIRS`` is set and a multi path the response is also a multi path separated by
+         the OS path separator), e.g. ``/etc/xdg/$appname/$version``
+        """
+        # XDG default for $XDG_CONFIG_DIRS only first, if multipath is False
+        dirs = self._site_config_dirs
+        if not self.multipath:
+            return dirs[0]
+        return os.pathsep.join(dirs)
+
+    @property
+    def user_cache_dir(self) -> str:
+        """
+        :return: cache directory tied to the user, e.g. ``~/.cache/$appname/$version`` or
+         ``~/$XDG_CACHE_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_CACHE_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.cache")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, e.g. ``/var/cache/$appname/$version``"""
+        return self._append_app_name_and_version("/var/cache")
+
+    @property
+    def user_state_dir(self) -> str:
+        """
+        :return: state directory tied to the user, e.g. ``~/.local/state/$appname/$version`` or
+         ``$XDG_STATE_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_STATE_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.local/state")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, same as `user_state_dir` if not opinionated else ``log`` in it"""
+        path = self.user_state_dir
+        if self.opinion:
+            path = os.path.join(path, "log")  # noqa: PTH118
+            self._optionally_create_directory(path)
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user, e.g. ``~/Documents``"""
+        return _get_user_media_dir("XDG_DOCUMENTS_DIR", "~/Documents")
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user, e.g. ``~/Downloads``"""
+        return _get_user_media_dir("XDG_DOWNLOAD_DIR", "~/Downloads")
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user, e.g. ``~/Pictures``"""
+        return _get_user_media_dir("XDG_PICTURES_DIR", "~/Pictures")
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user, e.g. ``~/Videos``"""
+        return _get_user_media_dir("XDG_VIDEOS_DIR", "~/Videos")
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user, e.g. ``~/Music``"""
+        return _get_user_media_dir("XDG_MUSIC_DIR", "~/Music")
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``~/Desktop``"""
+        return _get_user_media_dir("XDG_DESKTOP_DIR", "~/Desktop")
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, e.g. ``/run/user/$(id -u)/$appname/$version`` or
+         ``$XDG_RUNTIME_DIR/$appname/$version``.
+
+         For FreeBSD/OpenBSD/NetBSD, it would return ``/var/run/user/$(id -u)/$appname/$version`` if
+         exists, otherwise ``/tmp/runtime-$(id -u)/$appname/$version``, if``$XDG_RUNTIME_DIR``
+         is not set.
+        """
+        path = os.environ.get("XDG_RUNTIME_DIR", "")
+        if not path.strip():
+            if sys.platform.startswith(("freebsd", "openbsd", "netbsd")):
+                path = f"/var/run/user/{getuid()}"
+                if not Path(path).exists():
+                    path = f"/tmp/runtime-{getuid()}"  # noqa: S108
+            else:
+                path = f"/run/user/{getuid()}"
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """
+        :return: runtime directory shared by users, e.g. ``/run/$appname/$version`` or \
+        ``$XDG_RUNTIME_DIR/$appname/$version``.
+
+        Note that this behaves almost exactly like `user_runtime_dir` if ``$XDG_RUNTIME_DIR`` is set, but will
+        fall back to paths associated to the root user instead of a regular logged-in user if it's not set.
+
+        If you wish to ensure that a logged-in root user path is returned e.g. ``/run/user/0``, use `user_runtime_dir`
+        instead.
+
+        For FreeBSD/OpenBSD/NetBSD, it would return ``/var/run/$appname/$version`` if ``$XDG_RUNTIME_DIR`` is not set.
+        """
+        path = os.environ.get("XDG_RUNTIME_DIR", "")
+        if not path.strip():
+            if sys.platform.startswith(("freebsd", "openbsd", "netbsd")):
+                path = "/var/run"
+            else:
+                path = "/run"
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_data_path(self) -> Path:
+        """:return: data path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_data_dir)
+
+    @property
+    def site_config_path(self) -> Path:
+        """:return: config path shared by the users, returns the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_config_dir)
+
+    @property
+    def site_cache_path(self) -> Path:
+        """:return: cache path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_cache_dir)
+
+    def iter_config_dirs(self) -> Iterator[str]:
+        """:yield: all user and site configuration directories."""
+        yield self.user_config_dir
+        yield from self._site_config_dirs
+
+    def iter_data_dirs(self) -> Iterator[str]:
+        """:yield: all user and site data directories."""
+        yield self.user_data_dir
+        yield from self._site_data_dirs
+
+
+def _get_user_media_dir(env_var: str, fallback_tilde_path: str) -> str:
+    media_dir = _get_user_dirs_folder(env_var)
+    if media_dir is None:
+        media_dir = os.environ.get(env_var, "").strip()
+        if not media_dir:
+            media_dir = os.path.expanduser(fallback_tilde_path)  # noqa: PTH111
+
+    return media_dir
+
+
+def _get_user_dirs_folder(key: str) -> str | None:
+    """
+    Return directory from user-dirs.dirs config file.
+
+    See https://freedesktop.org/wiki/Software/xdg-user-dirs/.
+
+    """
+    user_dirs_config_path = Path(Unix().user_config_dir) / "user-dirs.dirs"
+    if user_dirs_config_path.exists():
+        parser = ConfigParser()
+
+        with user_dirs_config_path.open() as stream:
+            # Add fake section header, so ConfigParser doesn't complain
+            parser.read_string(f"[top]\n{stream.read()}")
+
+        if key not in parser["top"]:
+            return None
+
+        path = parser["top"][key].strip('"')
+        # Handle relative home paths
+        return path.replace("$HOME", os.path.expanduser("~"))  # noqa: PTH111
+
+    return None
+
+
+__all__ = [
+    "Unix",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/version.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/version.py
new file mode 100644
index 00000000..afb49243
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '4.3.6'
+__version_tuple__ = version_tuple = (4, 3, 6)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/windows.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/windows.py
new file mode 100644
index 00000000..d7bc9609
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/platformdirs/windows.py
@@ -0,0 +1,272 @@
+"""Windows."""
+
+from __future__ import annotations
+
+import os
+import sys
+from functools import lru_cache
+from typing import TYPE_CHECKING
+
+from .api import PlatformDirsABC
+
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+
+class Windows(PlatformDirsABC):
+    """
+    `MSDN on where to store app data files <https://learn.microsoft.com/en-us/windows/win32/shell/knownfolderid>`_.
+
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`, `appauthor
+    <platformdirs.api.PlatformDirsABC.appauthor>`, `version <platformdirs.api.PlatformDirsABC.version>`, `roaming
+    <platformdirs.api.PlatformDirsABC.roaming>`, `opinion <platformdirs.api.PlatformDirsABC.opinion>`, `ensure_exists
+    <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """
+        :return: data directory tied to the user, e.g.
+         ``%USERPROFILE%\\AppData\\Local\\$appauthor\\$appname`` (not roaming) or
+         ``%USERPROFILE%\\AppData\\Roaming\\$appauthor\\$appname`` (roaming)
+        """
+        const = "CSIDL_APPDATA" if self.roaming else "CSIDL_LOCAL_APPDATA"
+        path = os.path.normpath(get_win_folder(const))
+        return self._append_parts(path)
+
+    def _append_parts(self, path: str, *, opinion_value: str | None = None) -> str:
+        params = []
+        if self.appname:
+            if self.appauthor is not False:
+                author = self.appauthor or self.appname
+                params.append(author)
+            params.append(self.appname)
+            if opinion_value is not None and self.opinion:
+                params.append(opinion_value)
+            if self.version:
+                params.append(self.version)
+        path = os.path.join(path, *params)  # noqa: PTH118
+        self._optionally_create_directory(path)
+        return path
+
+    @property
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users, e.g. ``C:\\ProgramData\\$appauthor\\$appname``"""
+        path = os.path.normpath(get_win_folder("CSIDL_COMMON_APPDATA"))
+        return self._append_parts(path)
+
+    @property
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `site_data_dir`"""
+        return self.site_data_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """
+        :return: cache directory tied to the user (if opinionated with ``Cache`` folder within ``$appname``) e.g.
+         ``%USERPROFILE%\\AppData\\Local\\$appauthor\\$appname\\Cache\\$version``
+        """
+        path = os.path.normpath(get_win_folder("CSIDL_LOCAL_APPDATA"))
+        return self._append_parts(path, opinion_value="Cache")
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, e.g. ``C:\\ProgramData\\$appauthor\\$appname\\Cache\\$version``"""
+        path = os.path.normpath(get_win_folder("CSIDL_COMMON_APPDATA"))
+        return self._append_parts(path, opinion_value="Cache")
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, same as `user_data_dir` if not opinionated else ``Logs`` in it"""
+        path = self.user_data_dir
+        if self.opinion:
+            path = os.path.join(path, "Logs")  # noqa: PTH118
+            self._optionally_create_directory(path)
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user e.g. ``%USERPROFILE%\\Documents``"""
+        return os.path.normpath(get_win_folder("CSIDL_PERSONAL"))
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user e.g. ``%USERPROFILE%\\Downloads``"""
+        return os.path.normpath(get_win_folder("CSIDL_DOWNLOADS"))
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user e.g. ``%USERPROFILE%\\Pictures``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYPICTURES"))
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user e.g. ``%USERPROFILE%\\Videos``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYVIDEO"))
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user e.g. ``%USERPROFILE%\\Music``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYMUSIC"))
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``%USERPROFILE%\\Desktop``"""
+        return os.path.normpath(get_win_folder("CSIDL_DESKTOPDIRECTORY"))
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, e.g.
+         ``%USERPROFILE%\\AppData\\Local\\Temp\\$appauthor\\$appname``
+        """
+        path = os.path.normpath(os.path.join(get_win_folder("CSIDL_LOCAL_APPDATA"), "Temp"))  # noqa: PTH118
+        return self._append_parts(path)
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+def get_win_folder_from_env_vars(csidl_name: str) -> str:
+    """Get folder from environment variables."""
+    result = get_win_folder_if_csidl_name_not_env_var(csidl_name)
+    if result is not None:
+        return result
+
+    env_var_name = {
+        "CSIDL_APPDATA": "APPDATA",
+        "CSIDL_COMMON_APPDATA": "ALLUSERSPROFILE",
+        "CSIDL_LOCAL_APPDATA": "LOCALAPPDATA",
+    }.get(csidl_name)
+    if env_var_name is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+    result = os.environ.get(env_var_name)
+    if result is None:
+        msg = f"Unset environment variable: {env_var_name}"
+        raise ValueError(msg)
+    return result
+
+
+def get_win_folder_if_csidl_name_not_env_var(csidl_name: str) -> str | None:
+    """Get a folder for a CSIDL name that does not exist as an environment variable."""
+    if csidl_name == "CSIDL_PERSONAL":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Documents")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_DOWNLOADS":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Downloads")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYPICTURES":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Pictures")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYVIDEO":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Videos")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYMUSIC":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Music")  # noqa: PTH118
+    return None
+
+
+def get_win_folder_from_registry(csidl_name: str) -> str:
+    """
+    Get folder from the registry.
+
+    This is a fallback technique at best. I'm not sure if using the registry for these guarantees us the correct answer
+    for all CSIDL_* names.
+
+    """
+    shell_folder_name = {
+        "CSIDL_APPDATA": "AppData",
+        "CSIDL_COMMON_APPDATA": "Common AppData",
+        "CSIDL_LOCAL_APPDATA": "Local AppData",
+        "CSIDL_PERSONAL": "Personal",
+        "CSIDL_DOWNLOADS": "{374DE290-123F-4565-9164-39C4925E467B}",
+        "CSIDL_MYPICTURES": "My Pictures",
+        "CSIDL_MYVIDEO": "My Video",
+        "CSIDL_MYMUSIC": "My Music",
+    }.get(csidl_name)
+    if shell_folder_name is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+    if sys.platform != "win32":  # only needed for mypy type checker to know that this code runs only on Windows
+        raise NotImplementedError
+    import winreg  # noqa: PLC0415
+
+    key = winreg.OpenKey(winreg.HKEY_CURRENT_USER, r"Software\Microsoft\Windows\CurrentVersion\Explorer\Shell Folders")
+    directory, _ = winreg.QueryValueEx(key, shell_folder_name)
+    return str(directory)
+
+
+def get_win_folder_via_ctypes(csidl_name: str) -> str:
+    """Get folder with ctypes."""
+    # There is no 'CSIDL_DOWNLOADS'.
+    # Use 'CSIDL_PROFILE' (40) and append the default folder 'Downloads' instead.
+    # https://learn.microsoft.com/en-us/windows/win32/shell/knownfolderid
+
+    import ctypes  # noqa: PLC0415
+
+    csidl_const = {
+        "CSIDL_APPDATA": 26,
+        "CSIDL_COMMON_APPDATA": 35,
+        "CSIDL_LOCAL_APPDATA": 28,
+        "CSIDL_PERSONAL": 5,
+        "CSIDL_MYPICTURES": 39,
+        "CSIDL_MYVIDEO": 14,
+        "CSIDL_MYMUSIC": 13,
+        "CSIDL_DOWNLOADS": 40,
+        "CSIDL_DESKTOPDIRECTORY": 16,
+    }.get(csidl_name)
+    if csidl_const is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+
+    buf = ctypes.create_unicode_buffer(1024)
+    windll = getattr(ctypes, "windll")  # noqa: B009 # using getattr to avoid false positive with mypy type checker
+    windll.shell32.SHGetFolderPathW(None, csidl_const, None, 0, buf)
+
+    # Downgrade to short path name if it has high-bit chars.
+    if any(ord(c) > 255 for c in buf):  # noqa: PLR2004
+        buf2 = ctypes.create_unicode_buffer(1024)
+        if windll.kernel32.GetShortPathNameW(buf.value, buf2, 1024):
+            buf = buf2
+
+    if csidl_name == "CSIDL_DOWNLOADS":
+        return os.path.join(buf.value, "Downloads")  # noqa: PTH118
+
+    return buf.value
+
+
+def _pick_get_win_folder() -> Callable[[str], str]:
+    try:
+        import ctypes  # noqa: PLC0415
+    except ImportError:
+        pass
+    else:
+        if hasattr(ctypes, "windll"):
+            return get_win_folder_via_ctypes
+    try:
+        import winreg  # noqa: PLC0415, F401
+    except ImportError:
+        return get_win_folder_from_env_vars
+    else:
+        return get_win_folder_from_registry
+
+
+get_win_folder = lru_cache(maxsize=None)(_pick_get_win_folder())
+
+__all__ = [
+    "Windows",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__init__.py
new file mode 100644
index 00000000..60ae9bb8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__init__.py
@@ -0,0 +1,82 @@
+"""
+    Pygments
+    ~~~~~~~~
+
+    Pygments is a syntax highlighting package written in Python.
+
+    It is a generic syntax highlighter for general use in all kinds of software
+    such as forum systems, wikis or other applications that need to prettify
+    source code. Highlights are:
+
+    * a wide range of common languages and markup formats is supported
+    * special attention is paid to details, increasing quality by a fair amount
+    * support for new languages and formats are added easily
+    * a number of output formats, presently HTML, LaTeX, RTF, SVG, all image
+      formats that PIL supports, and ANSI sequences
+    * it is usable as a command-line tool and as a library
+    * ... and it highlights even Brainfuck!
+
+    The `Pygments master branch`_ is installable with ``easy_install Pygments==dev``.
+
+    .. _Pygments master branch:
+       https://github.com/pygments/pygments/archive/master.zip#egg=Pygments-dev
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+from io import StringIO, BytesIO
+
+__version__ = '2.18.0'
+__docformat__ = 'restructuredtext'
+
+__all__ = ['lex', 'format', 'highlight']
+
+
+def lex(code, lexer):
+    """
+    Lex `code` with the `lexer` (must be a `Lexer` instance)
+    and return an iterable of tokens. Currently, this only calls
+    `lexer.get_tokens()`.
+    """
+    try:
+        return lexer.get_tokens(code)
+    except TypeError:
+        # Heuristic to catch a common mistake.
+        from pip._vendor.pygments.lexer import RegexLexer
+        if isinstance(lexer, type) and issubclass(lexer, RegexLexer):
+            raise TypeError('lex() argument must be a lexer instance, '
+                            'not a class')
+        raise
+
+
+def format(tokens, formatter, outfile=None):  # pylint: disable=redefined-builtin
+    """
+    Format ``tokens`` (an iterable of tokens) with the formatter ``formatter``
+    (a `Formatter` instance).
+
+    If ``outfile`` is given and a valid file object (an object with a
+    ``write`` method), the result will be written to it, otherwise it
+    is returned as a string.
+    """
+    try:
+        if not outfile:
+            realoutfile = getattr(formatter, 'encoding', None) and BytesIO() or StringIO()
+            formatter.format(tokens, realoutfile)
+            return realoutfile.getvalue()
+        else:
+            formatter.format(tokens, outfile)
+    except TypeError:
+        # Heuristic to catch a common mistake.
+        from pip._vendor.pygments.formatter import Formatter
+        if isinstance(formatter, type) and issubclass(formatter, Formatter):
+            raise TypeError('format() argument must be a formatter instance, '
+                            'not a class')
+        raise
+
+
+def highlight(code, lexer, formatter, outfile=None):
+    """
+    This is the most high-level highlighting function. It combines `lex` and
+    `format` in one function.
+    """
+    return format(lex(code, lexer), formatter, outfile)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__main__.py
new file mode 100644
index 00000000..dcc6e5ad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__main__.py
@@ -0,0 +1,17 @@
+"""
+    pygments.__main__
+    ~~~~~~~~~~~~~~~~~
+
+    Main entry point for ``python -m pygments``.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import sys
+from pip._vendor.pygments.cmdline import main
+
+try:
+    sys.exit(main(sys.argv))
+except KeyboardInterrupt:
+    sys.exit(1)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..830cb3a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..8c08d45a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/cmdline.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/cmdline.cpython-312.pyc
new file mode 100644
index 00000000..140c2804
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/cmdline.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/console.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/console.cpython-312.pyc
new file mode 100644
index 00000000..25cbeae1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/console.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/filter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/filter.cpython-312.pyc
new file mode 100644
index 00000000..19675a9b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/filter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/formatter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/formatter.cpython-312.pyc
new file mode 100644
index 00000000..6d7a95c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/formatter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/lexer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/lexer.cpython-312.pyc
new file mode 100644
index 00000000..d55c196d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/lexer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/modeline.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/modeline.cpython-312.pyc
new file mode 100644
index 00000000..34b3e1db
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/modeline.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/plugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/plugin.cpython-312.pyc
new file mode 100644
index 00000000..ddccd158
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/plugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/regexopt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/regexopt.cpython-312.pyc
new file mode 100644
index 00000000..e4c7f194
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/regexopt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/scanner.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/scanner.cpython-312.pyc
new file mode 100644
index 00000000..867d4f41
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/scanner.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/sphinxext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/sphinxext.cpython-312.pyc
new file mode 100644
index 00000000..f90e8d93
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/sphinxext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/style.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/style.cpython-312.pyc
new file mode 100644
index 00000000..89dc801a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/style.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/token.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/token.cpython-312.pyc
new file mode 100644
index 00000000..0cb3c6ca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/token.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/unistring.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/unistring.cpython-312.pyc
new file mode 100644
index 00000000..1cfa1de8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/unistring.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..32e88681
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/cmdline.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/cmdline.py
new file mode 100644
index 00000000..0a7072ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/cmdline.py
@@ -0,0 +1,668 @@
+"""
+    pygments.cmdline
+    ~~~~~~~~~~~~~~~~
+
+    Command line interface.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import os
+import sys
+import shutil
+import argparse
+from textwrap import dedent
+
+from pip._vendor.pygments import __version__, highlight
+from pip._vendor.pygments.util import ClassNotFound, OptionError, docstring_headline, \
+    guess_decode, guess_decode_from_terminal, terminal_encoding, \
+    UnclosingTextIOWrapper
+from pip._vendor.pygments.lexers import get_all_lexers, get_lexer_by_name, guess_lexer, \
+    load_lexer_from_file, get_lexer_for_filename, find_lexer_class_for_filename
+from pip._vendor.pygments.lexers.special import TextLexer
+from pip._vendor.pygments.formatters.latex import LatexEmbeddedLexer, LatexFormatter
+from pip._vendor.pygments.formatters import get_all_formatters, get_formatter_by_name, \
+    load_formatter_from_file, get_formatter_for_filename, find_formatter_class
+from pip._vendor.pygments.formatters.terminal import TerminalFormatter
+from pip._vendor.pygments.formatters.terminal256 import Terminal256Formatter, TerminalTrueColorFormatter
+from pip._vendor.pygments.filters import get_all_filters, find_filter_class
+from pip._vendor.pygments.styles import get_all_styles, get_style_by_name
+
+
+def _parse_options(o_strs):
+    opts = {}
+    if not o_strs:
+        return opts
+    for o_str in o_strs:
+        if not o_str.strip():
+            continue
+        o_args = o_str.split(',')
+        for o_arg in o_args:
+            o_arg = o_arg.strip()
+            try:
+                o_key, o_val = o_arg.split('=', 1)
+                o_key = o_key.strip()
+                o_val = o_val.strip()
+            except ValueError:
+                opts[o_arg] = True
+            else:
+                opts[o_key] = o_val
+    return opts
+
+
+def _parse_filters(f_strs):
+    filters = []
+    if not f_strs:
+        return filters
+    for f_str in f_strs:
+        if ':' in f_str:
+            fname, fopts = f_str.split(':', 1)
+            filters.append((fname, _parse_options([fopts])))
+        else:
+            filters.append((f_str, {}))
+    return filters
+
+
+def _print_help(what, name):
+    try:
+        if what == 'lexer':
+            cls = get_lexer_by_name(name)
+            print(f"Help on the {cls.name} lexer:")
+            print(dedent(cls.__doc__))
+        elif what == 'formatter':
+            cls = find_formatter_class(name)
+            print(f"Help on the {cls.name} formatter:")
+            print(dedent(cls.__doc__))
+        elif what == 'filter':
+            cls = find_filter_class(name)
+            print(f"Help on the {name} filter:")
+            print(dedent(cls.__doc__))
+        return 0
+    except (AttributeError, ValueError):
+        print(f"{what} not found!", file=sys.stderr)
+        return 1
+
+
+def _print_list(what):
+    if what == 'lexer':
+        print()
+        print("Lexers:")
+        print("~~~~~~~")
+
+        info = []
+        for fullname, names, exts, _ in get_all_lexers():
+            tup = (', '.join(names)+':', fullname,
+                   exts and '(filenames ' + ', '.join(exts) + ')' or '')
+            info.append(tup)
+        info.sort()
+        for i in info:
+            print(('* {}\n    {} {}').format(*i))
+
+    elif what == 'formatter':
+        print()
+        print("Formatters:")
+        print("~~~~~~~~~~~")
+
+        info = []
+        for cls in get_all_formatters():
+            doc = docstring_headline(cls)
+            tup = (', '.join(cls.aliases) + ':', doc, cls.filenames and
+                   '(filenames ' + ', '.join(cls.filenames) + ')' or '')
+            info.append(tup)
+        info.sort()
+        for i in info:
+            print(('* {}\n    {} {}').format(*i))
+
+    elif what == 'filter':
+        print()
+        print("Filters:")
+        print("~~~~~~~~")
+
+        for name in get_all_filters():
+            cls = find_filter_class(name)
+            print("* " + name + ':')
+            print(f"    {docstring_headline(cls)}")
+
+    elif what == 'style':
+        print()
+        print("Styles:")
+        print("~~~~~~~")
+
+        for name in get_all_styles():
+            cls = get_style_by_name(name)
+            print("* " + name + ':')
+            print(f"    {docstring_headline(cls)}")
+
+
+def _print_list_as_json(requested_items):
+    import json
+    result = {}
+    if 'lexer' in requested_items:
+        info = {}
+        for fullname, names, filenames, mimetypes in get_all_lexers():
+            info[fullname] = {
+                'aliases': names,
+                'filenames': filenames,
+                'mimetypes': mimetypes
+            }
+        result['lexers'] = info
+
+    if 'formatter' in requested_items:
+        info = {}
+        for cls in get_all_formatters():
+            doc = docstring_headline(cls)
+            info[cls.name] = {
+                'aliases': cls.aliases,
+                'filenames': cls.filenames,
+                'doc': doc
+            }
+        result['formatters'] = info
+
+    if 'filter' in requested_items:
+        info = {}
+        for name in get_all_filters():
+            cls = find_filter_class(name)
+            info[name] = {
+                'doc': docstring_headline(cls)
+            }
+        result['filters'] = info
+
+    if 'style' in requested_items:
+        info = {}
+        for name in get_all_styles():
+            cls = get_style_by_name(name)
+            info[name] = {
+                'doc': docstring_headline(cls)
+            }
+        result['styles'] = info
+
+    json.dump(result, sys.stdout)
+
+def main_inner(parser, argns):
+    if argns.help:
+        parser.print_help()
+        return 0
+
+    if argns.V:
+        print(f'Pygments version {__version__}, (c) 2006-2024 by Georg Brandl, Matthäus '
+              'Chajdas and contributors.')
+        return 0
+
+    def is_only_option(opt):
+        return not any(v for (k, v) in vars(argns).items() if k != opt)
+
+    # handle ``pygmentize -L``
+    if argns.L is not None:
+        arg_set = set()
+        for k, v in vars(argns).items():
+            if v:
+                arg_set.add(k)
+
+        arg_set.discard('L')
+        arg_set.discard('json')
+
+        if arg_set:
+            parser.print_help(sys.stderr)
+            return 2
+
+        # print version
+        if not argns.json:
+            main(['', '-V'])
+        allowed_types = {'lexer', 'formatter', 'filter', 'style'}
+        largs = [arg.rstrip('s') for arg in argns.L]
+        if any(arg not in allowed_types for arg in largs):
+            parser.print_help(sys.stderr)
+            return 0
+        if not largs:
+            largs = allowed_types
+        if not argns.json:
+            for arg in largs:
+                _print_list(arg)
+        else:
+            _print_list_as_json(largs)
+        return 0
+
+    # handle ``pygmentize -H``
+    if argns.H:
+        if not is_only_option('H'):
+            parser.print_help(sys.stderr)
+            return 2
+        what, name = argns.H
+        if what not in ('lexer', 'formatter', 'filter'):
+            parser.print_help(sys.stderr)
+            return 2
+        return _print_help(what, name)
+
+    # parse -O options
+    parsed_opts = _parse_options(argns.O or [])
+
+    # parse -P options
+    for p_opt in argns.P or []:
+        try:
+            name, value = p_opt.split('=', 1)
+        except ValueError:
+            parsed_opts[p_opt] = True
+        else:
+            parsed_opts[name] = value
+
+    # encodings
+    inencoding = parsed_opts.get('inencoding', parsed_opts.get('encoding'))
+    outencoding = parsed_opts.get('outencoding', parsed_opts.get('encoding'))
+
+    # handle ``pygmentize -N``
+    if argns.N:
+        lexer = find_lexer_class_for_filename(argns.N)
+        if lexer is None:
+            lexer = TextLexer
+
+        print(lexer.aliases[0])
+        return 0
+
+    # handle ``pygmentize -C``
+    if argns.C:
+        inp = sys.stdin.buffer.read()
+        try:
+            lexer = guess_lexer(inp, inencoding=inencoding)
+        except ClassNotFound:
+            lexer = TextLexer
+
+        print(lexer.aliases[0])
+        return 0
+
+    # handle ``pygmentize -S``
+    S_opt = argns.S
+    a_opt = argns.a
+    if S_opt is not None:
+        f_opt = argns.f
+        if not f_opt:
+            parser.print_help(sys.stderr)
+            return 2
+        if argns.l or argns.INPUTFILE:
+            parser.print_help(sys.stderr)
+            return 2
+
+        try:
+            parsed_opts['style'] = S_opt
+            fmter = get_formatter_by_name(f_opt, **parsed_opts)
+        except ClassNotFound as err:
+            print(err, file=sys.stderr)
+            return 1
+
+        print(fmter.get_style_defs(a_opt or ''))
+        return 0
+
+    # if no -S is given, -a is not allowed
+    if argns.a is not None:
+        parser.print_help(sys.stderr)
+        return 2
+
+    # parse -F options
+    F_opts = _parse_filters(argns.F or [])
+
+    # -x: allow custom (eXternal) lexers and formatters
+    allow_custom_lexer_formatter = bool(argns.x)
+
+    # select lexer
+    lexer = None
+
+    # given by name?
+    lexername = argns.l
+    if lexername:
+        # custom lexer, located relative to user's cwd
+        if allow_custom_lexer_formatter and '.py' in lexername:
+            try:
+                filename = None
+                name = None
+                if ':' in lexername:
+                    filename, name = lexername.rsplit(':', 1)
+
+                    if '.py' in name:
+                        # This can happen on Windows: If the lexername is
+                        # C:\lexer.py -- return to normal load path in that case
+                        name = None
+
+                if filename and name:
+                    lexer = load_lexer_from_file(filename, name,
+                                                 **parsed_opts)
+                else:
+                    lexer = load_lexer_from_file(lexername, **parsed_opts)
+            except ClassNotFound as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+        else:
+            try:
+                lexer = get_lexer_by_name(lexername, **parsed_opts)
+            except (OptionError, ClassNotFound) as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+
+    # read input code
+    code = None
+
+    if argns.INPUTFILE:
+        if argns.s:
+            print('Error: -s option not usable when input file specified',
+                  file=sys.stderr)
+            return 2
+
+        infn = argns.INPUTFILE
+        try:
+            with open(infn, 'rb') as infp:
+                code = infp.read()
+        except Exception as err:
+            print('Error: cannot read infile:', err, file=sys.stderr)
+            return 1
+        if not inencoding:
+            code, inencoding = guess_decode(code)
+
+        # do we have to guess the lexer?
+        if not lexer:
+            try:
+                lexer = get_lexer_for_filename(infn, code, **parsed_opts)
+            except ClassNotFound as err:
+                if argns.g:
+                    try:
+                        lexer = guess_lexer(code, **parsed_opts)
+                    except ClassNotFound:
+                        lexer = TextLexer(**parsed_opts)
+                else:
+                    print('Error:', err, file=sys.stderr)
+                    return 1
+            except OptionError as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+
+    elif not argns.s:  # treat stdin as full file (-s support is later)
+        # read code from terminal, always in binary mode since we want to
+        # decode ourselves and be tolerant with it
+        code = sys.stdin.buffer.read()  # use .buffer to get a binary stream
+        if not inencoding:
+            code, inencoding = guess_decode_from_terminal(code, sys.stdin)
+            # else the lexer will do the decoding
+        if not lexer:
+            try:
+                lexer = guess_lexer(code, **parsed_opts)
+            except ClassNotFound:
+                lexer = TextLexer(**parsed_opts)
+
+    else:  # -s option needs a lexer with -l
+        if not lexer:
+            print('Error: when using -s a lexer has to be selected with -l',
+                  file=sys.stderr)
+            return 2
+
+    # process filters
+    for fname, fopts in F_opts:
+        try:
+            lexer.add_filter(fname, **fopts)
+        except ClassNotFound as err:
+            print('Error:', err, file=sys.stderr)
+            return 1
+
+    # select formatter
+    outfn = argns.o
+    fmter = argns.f
+    if fmter:
+        # custom formatter, located relative to user's cwd
+        if allow_custom_lexer_formatter and '.py' in fmter:
+            try:
+                filename = None
+                name = None
+                if ':' in fmter:
+                    # Same logic as above for custom lexer
+                    filename, name = fmter.rsplit(':', 1)
+
+                    if '.py' in name:
+                        name = None
+
+                if filename and name:
+                    fmter = load_formatter_from_file(filename, name,
+                                                     **parsed_opts)
+                else:
+                    fmter = load_formatter_from_file(fmter, **parsed_opts)
+            except ClassNotFound as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+        else:
+            try:
+                fmter = get_formatter_by_name(fmter, **parsed_opts)
+            except (OptionError, ClassNotFound) as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+
+    if outfn:
+        if not fmter:
+            try:
+                fmter = get_formatter_for_filename(outfn, **parsed_opts)
+            except (OptionError, ClassNotFound) as err:
+                print('Error:', err, file=sys.stderr)
+                return 1
+        try:
+            outfile = open(outfn, 'wb')
+        except Exception as err:
+            print('Error: cannot open outfile:', err, file=sys.stderr)
+            return 1
+    else:
+        if not fmter:
+            if os.environ.get('COLORTERM','') in ('truecolor', '24bit'):
+                fmter = TerminalTrueColorFormatter(**parsed_opts)
+            elif '256' in os.environ.get('TERM', ''):
+                fmter = Terminal256Formatter(**parsed_opts)
+            else:
+                fmter = TerminalFormatter(**parsed_opts)
+        outfile = sys.stdout.buffer
+
+    # determine output encoding if not explicitly selected
+    if not outencoding:
+        if outfn:
+            # output file? use lexer encoding for now (can still be None)
+            fmter.encoding = inencoding
+        else:
+            # else use terminal encoding
+            fmter.encoding = terminal_encoding(sys.stdout)
+
+    # provide coloring under Windows, if possible
+    if not outfn and sys.platform in ('win32', 'cygwin') and \
+       fmter.name in ('Terminal', 'Terminal256'):  # pragma: no cover
+        # unfortunately colorama doesn't support binary streams on Py3
+        outfile = UnclosingTextIOWrapper(outfile, encoding=fmter.encoding)
+        fmter.encoding = None
+        try:
+            import colorama.initialise
+        except ImportError:
+            pass
+        else:
+            outfile = colorama.initialise.wrap_stream(
+                outfile, convert=None, strip=None, autoreset=False, wrap=True)
+
+    # When using the LaTeX formatter and the option `escapeinside` is
+    # specified, we need a special lexer which collects escaped text
+    # before running the chosen language lexer.
+    escapeinside = parsed_opts.get('escapeinside', '')
+    if len(escapeinside) == 2 and isinstance(fmter, LatexFormatter):
+        left = escapeinside[0]
+        right = escapeinside[1]
+        lexer = LatexEmbeddedLexer(left, right, lexer)
+
+    # ... and do it!
+    if not argns.s:
+        # process whole input as per normal...
+        try:
+            highlight(code, lexer, fmter, outfile)
+        finally:
+            if outfn:
+                outfile.close()
+        return 0
+    else:
+        # line by line processing of stdin (eg: for 'tail -f')...
+        try:
+            while 1:
+                line = sys.stdin.buffer.readline()
+                if not line:
+                    break
+                if not inencoding:
+                    line = guess_decode_from_terminal(line, sys.stdin)[0]
+                highlight(line, lexer, fmter, outfile)
+                if hasattr(outfile, 'flush'):
+                    outfile.flush()
+            return 0
+        except KeyboardInterrupt:  # pragma: no cover
+            return 0
+        finally:
+            if outfn:
+                outfile.close()
+
+
+class HelpFormatter(argparse.HelpFormatter):
+    def __init__(self, prog, indent_increment=2, max_help_position=16, width=None):
+        if width is None:
+            try:
+                width = shutil.get_terminal_size().columns - 2
+            except Exception:
+                pass
+        argparse.HelpFormatter.__init__(self, prog, indent_increment,
+                                        max_help_position, width)
+
+
+def main(args=sys.argv):
+    """
+    Main command line entry point.
+    """
+    desc = "Highlight an input file and write the result to an output file."
+    parser = argparse.ArgumentParser(description=desc, add_help=False,
+                                     formatter_class=HelpFormatter)
+
+    operation = parser.add_argument_group('Main operation')
+    lexersel = operation.add_mutually_exclusive_group()
+    lexersel.add_argument(
+        '-l', metavar='LEXER',
+        help='Specify the lexer to use.  (Query names with -L.)  If not '
+        'given and -g is not present, the lexer is guessed from the filename.')
+    lexersel.add_argument(
+        '-g', action='store_true',
+        help='Guess the lexer from the file contents, or pass through '
+        'as plain text if nothing can be guessed.')
+    operation.add_argument(
+        '-F', metavar='FILTER[:options]', action='append',
+        help='Add a filter to the token stream.  (Query names with -L.) '
+        'Filter options are given after a colon if necessary.')
+    operation.add_argument(
+        '-f', metavar='FORMATTER',
+        help='Specify the formatter to use.  (Query names with -L.) '
+        'If not given, the formatter is guessed from the output filename, '
+        'and defaults to the terminal formatter if the output is to the '
+        'terminal or an unknown file extension.')
+    operation.add_argument(
+        '-O', metavar='OPTION=value[,OPTION=value,...]', action='append',
+        help='Give options to the lexer and formatter as a comma-separated '
+        'list of key-value pairs. '
+        'Example: `-O bg=light,python=cool`.')
+    operation.add_argument(
+        '-P', metavar='OPTION=value', action='append',
+        help='Give a single option to the lexer and formatter - with this '
+        'you can pass options whose value contains commas and equal signs. '
+        'Example: `-P "heading=Pygments, the Python highlighter"`.')
+    operation.add_argument(
+        '-o', metavar='OUTPUTFILE',
+        help='Where to write the output.  Defaults to standard output.')
+
+    operation.add_argument(
+        'INPUTFILE', nargs='?',
+        help='Where to read the input.  Defaults to standard input.')
+
+    flags = parser.add_argument_group('Operation flags')
+    flags.add_argument(
+        '-v', action='store_true',
+        help='Print a detailed traceback on unhandled exceptions, which '
+        'is useful for debugging and bug reports.')
+    flags.add_argument(
+        '-s', action='store_true',
+        help='Process lines one at a time until EOF, rather than waiting to '
+        'process the entire file.  This only works for stdin, only for lexers '
+        'with no line-spanning constructs, and is intended for streaming '
+        'input such as you get from `tail -f`. '
+        'Example usage: `tail -f sql.log | pygmentize -s -l sql`.')
+    flags.add_argument(
+        '-x', action='store_true',
+        help='Allow custom lexers and formatters to be loaded from a .py file '
+        'relative to the current working directory. For example, '
+        '`-l ./customlexer.py -x`. By default, this option expects a file '
+        'with a class named CustomLexer or CustomFormatter; you can also '
+        'specify your own class name with a colon (`-l ./lexer.py:MyLexer`). '
+        'Users should be very careful not to use this option with untrusted '
+        'files, because it will import and run them.')
+    flags.add_argument('--json', help='Output as JSON. This can '
+        'be only used in conjunction with -L.',
+        default=False,
+        action='store_true')
+
+    special_modes_group = parser.add_argument_group(
+        'Special modes - do not do any highlighting')
+    special_modes = special_modes_group.add_mutually_exclusive_group()
+    special_modes.add_argument(
+        '-S', metavar='STYLE -f formatter',
+        help='Print style definitions for STYLE for a formatter '
+        'given with -f. The argument given by -a is formatter '
+        'dependent.')
+    special_modes.add_argument(
+        '-L', nargs='*', metavar='WHAT',
+        help='List lexers, formatters, styles or filters -- '
+        'give additional arguments for the thing(s) you want to list '
+        '(e.g. "styles"), or omit them to list everything.')
+    special_modes.add_argument(
+        '-N', metavar='FILENAME',
+        help='Guess and print out a lexer name based solely on the given '
+        'filename. Does not take input or highlight anything. If no specific '
+        'lexer can be determined, "text" is printed.')
+    special_modes.add_argument(
+        '-C', action='store_true',
+        help='Like -N, but print out a lexer name based solely on '
+        'a given content from standard input.')
+    special_modes.add_argument(
+        '-H', action='store', nargs=2, metavar=('NAME', 'TYPE'),
+        help='Print detailed help for the object <name> of type <type>, '
+        'where <type> is one of "lexer", "formatter" or "filter".')
+    special_modes.add_argument(
+        '-V', action='store_true',
+        help='Print the package version.')
+    special_modes.add_argument(
+        '-h', '--help', action='store_true',
+        help='Print this help.')
+    special_modes_group.add_argument(
+        '-a', metavar='ARG',
+        help='Formatter-specific additional argument for the -S (print '
+        'style sheet) mode.')
+
+    argns = parser.parse_args(args[1:])
+
+    try:
+        return main_inner(parser, argns)
+    except BrokenPipeError:
+        # someone closed our stdout, e.g. by quitting a pager.
+        return 0
+    except Exception:
+        if argns.v:
+            print(file=sys.stderr)
+            print('*' * 65, file=sys.stderr)
+            print('An unhandled exception occurred while highlighting.',
+                  file=sys.stderr)
+            print('Please report the whole traceback to the issue tracker at',
+                  file=sys.stderr)
+            print('<https://github.com/pygments/pygments/issues>.',
+                  file=sys.stderr)
+            print('*' * 65, file=sys.stderr)
+            print(file=sys.stderr)
+            raise
+        import traceback
+        info = traceback.format_exception(*sys.exc_info())
+        msg = info[-1].strip()
+        if len(info) >= 3:
+            # extract relevant file and position info
+            msg += '\n   (f{})'.format(info[-2].split('\n')[0].strip()[1:])
+        print(file=sys.stderr)
+        print('*** Error while highlighting:', file=sys.stderr)
+        print(msg, file=sys.stderr)
+        print('*** If this is a bug you want to report, please rerun with -v.',
+              file=sys.stderr)
+        return 1
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/console.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/console.py
new file mode 100644
index 00000000..4c1a0621
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/console.py
@@ -0,0 +1,70 @@
+"""
+    pygments.console
+    ~~~~~~~~~~~~~~~~
+
+    Format colored console output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+esc = "\x1b["
+
+codes = {}
+codes[""] = ""
+codes["reset"] = esc + "39;49;00m"
+
+codes["bold"] = esc + "01m"
+codes["faint"] = esc + "02m"
+codes["standout"] = esc + "03m"
+codes["underline"] = esc + "04m"
+codes["blink"] = esc + "05m"
+codes["overline"] = esc + "06m"
+
+dark_colors = ["black", "red", "green", "yellow", "blue",
+               "magenta", "cyan", "gray"]
+light_colors = ["brightblack", "brightred", "brightgreen", "brightyellow", "brightblue",
+                "brightmagenta", "brightcyan", "white"]
+
+x = 30
+for dark, light in zip(dark_colors, light_colors):
+    codes[dark] = esc + "%im" % x
+    codes[light] = esc + "%im" % (60 + x)
+    x += 1
+
+del dark, light, x
+
+codes["white"] = codes["bold"]
+
+
+def reset_color():
+    return codes["reset"]
+
+
+def colorize(color_key, text):
+    return codes[color_key] + text + codes["reset"]
+
+
+def ansiformat(attr, text):
+    """
+    Format ``text`` with a color and/or some attributes::
+
+        color       normal color
+        *color*     bold color
+        _color_     underlined color
+        +color+     blinking color
+    """
+    result = []
+    if attr[:1] == attr[-1:] == '+':
+        result.append(codes['blink'])
+        attr = attr[1:-1]
+    if attr[:1] == attr[-1:] == '*':
+        result.append(codes['bold'])
+        attr = attr[1:-1]
+    if attr[:1] == attr[-1:] == '_':
+        result.append(codes['underline'])
+        attr = attr[1:-1]
+    result.append(codes[attr])
+    result.append(text)
+    result.append(codes['reset'])
+    return ''.join(result)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filter.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filter.py
new file mode 100644
index 00000000..aa6f7604
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filter.py
@@ -0,0 +1,70 @@
+"""
+    pygments.filter
+    ~~~~~~~~~~~~~~~
+
+    Module that implements the default filter.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+
+def apply_filters(stream, filters, lexer=None):
+    """
+    Use this method to apply an iterable of filters to
+    a stream. If lexer is given it's forwarded to the
+    filter, otherwise the filter receives `None`.
+    """
+    def _apply(filter_, stream):
+        yield from filter_.filter(lexer, stream)
+    for filter_ in filters:
+        stream = _apply(filter_, stream)
+    return stream
+
+
+def simplefilter(f):
+    """
+    Decorator that converts a function into a filter::
+
+        @simplefilter
+        def lowercase(self, lexer, stream, options):
+            for ttype, value in stream:
+                yield ttype, value.lower()
+    """
+    return type(f.__name__, (FunctionFilter,), {
+        '__module__': getattr(f, '__module__'),
+        '__doc__': f.__doc__,
+        'function': f,
+    })
+
+
+class Filter:
+    """
+    Default filter. Subclass this class or use the `simplefilter`
+    decorator to create own filters.
+    """
+
+    def __init__(self, **options):
+        self.options = options
+
+    def filter(self, lexer, stream):
+        raise NotImplementedError()
+
+
+class FunctionFilter(Filter):
+    """
+    Abstract class used by `simplefilter` to create simple
+    function filters on the fly. The `simplefilter` decorator
+    automatically creates subclasses of this class for
+    functions passed to it.
+    """
+    function = None
+
+    def __init__(self, **options):
+        if not hasattr(self, 'function'):
+            raise TypeError(f'{self.__class__.__name__!r} used without bound function')
+        Filter.__init__(self, **options)
+
+    def filter(self, lexer, stream):
+        # pylint: disable=not-callable
+        yield from self.function(lexer, stream, self.options)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__init__.py
new file mode 100644
index 00000000..9255ca22
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__init__.py
@@ -0,0 +1,940 @@
+"""
+    pygments.filters
+    ~~~~~~~~~~~~~~~~
+
+    Module containing filter lookup functions and default
+    filters.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+
+from pip._vendor.pygments.token import String, Comment, Keyword, Name, Error, Whitespace, \
+    string_to_tokentype
+from pip._vendor.pygments.filter import Filter
+from pip._vendor.pygments.util import get_list_opt, get_int_opt, get_bool_opt, \
+    get_choice_opt, ClassNotFound, OptionError
+from pip._vendor.pygments.plugin import find_plugin_filters
+
+
+def find_filter_class(filtername):
+    """Lookup a filter by name. Return None if not found."""
+    if filtername in FILTERS:
+        return FILTERS[filtername]
+    for name, cls in find_plugin_filters():
+        if name == filtername:
+            return cls
+    return None
+
+
+def get_filter_by_name(filtername, **options):
+    """Return an instantiated filter.
+
+    Options are passed to the filter initializer if wanted.
+    Raise a ClassNotFound if not found.
+    """
+    cls = find_filter_class(filtername)
+    if cls:
+        return cls(**options)
+    else:
+        raise ClassNotFound(f'filter {filtername!r} not found')
+
+
+def get_all_filters():
+    """Return a generator of all filter names."""
+    yield from FILTERS
+    for name, _ in find_plugin_filters():
+        yield name
+
+
+def _replace_special(ttype, value, regex, specialttype,
+                     replacefunc=lambda x: x):
+    last = 0
+    for match in regex.finditer(value):
+        start, end = match.start(), match.end()
+        if start != last:
+            yield ttype, value[last:start]
+        yield specialttype, replacefunc(value[start:end])
+        last = end
+    if last != len(value):
+        yield ttype, value[last:]
+
+
+class CodeTagFilter(Filter):
+    """Highlight special code tags in comments and docstrings.
+
+    Options accepted:
+
+    `codetags` : list of strings
+       A list of strings that are flagged as code tags.  The default is to
+       highlight ``XXX``, ``TODO``, ``FIXME``, ``BUG`` and ``NOTE``.
+
+    .. versionchanged:: 2.13
+       Now recognizes ``FIXME`` by default.
+    """
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        tags = get_list_opt(options, 'codetags',
+                            ['XXX', 'TODO', 'FIXME', 'BUG', 'NOTE'])
+        self.tag_re = re.compile(r'\b({})\b'.format('|'.join([
+            re.escape(tag) for tag in tags if tag
+        ])))
+
+    def filter(self, lexer, stream):
+        regex = self.tag_re
+        for ttype, value in stream:
+            if ttype in String.Doc or \
+               ttype in Comment and \
+               ttype not in Comment.Preproc:
+                yield from _replace_special(ttype, value, regex, Comment.Special)
+            else:
+                yield ttype, value
+
+
+class SymbolFilter(Filter):
+    """Convert mathematical symbols such as \\<longrightarrow> in Isabelle
+    or \\longrightarrow in LaTeX into Unicode characters.
+
+    This is mostly useful for HTML or console output when you want to
+    approximate the source rendering you'd see in an IDE.
+
+    Options accepted:
+
+    `lang` : string
+       The symbol language. Must be one of ``'isabelle'`` or
+       ``'latex'``.  The default is ``'isabelle'``.
+    """
+
+    latex_symbols = {
+        '\\alpha'                : '\U000003b1',
+        '\\beta'                 : '\U000003b2',
+        '\\gamma'                : '\U000003b3',
+        '\\delta'                : '\U000003b4',
+        '\\varepsilon'           : '\U000003b5',
+        '\\zeta'                 : '\U000003b6',
+        '\\eta'                  : '\U000003b7',
+        '\\vartheta'             : '\U000003b8',
+        '\\iota'                 : '\U000003b9',
+        '\\kappa'                : '\U000003ba',
+        '\\lambda'               : '\U000003bb',
+        '\\mu'                   : '\U000003bc',
+        '\\nu'                   : '\U000003bd',
+        '\\xi'                   : '\U000003be',
+        '\\pi'                   : '\U000003c0',
+        '\\varrho'               : '\U000003c1',
+        '\\sigma'                : '\U000003c3',
+        '\\tau'                  : '\U000003c4',
+        '\\upsilon'              : '\U000003c5',
+        '\\varphi'               : '\U000003c6',
+        '\\chi'                  : '\U000003c7',
+        '\\psi'                  : '\U000003c8',
+        '\\omega'                : '\U000003c9',
+        '\\Gamma'                : '\U00000393',
+        '\\Delta'                : '\U00000394',
+        '\\Theta'                : '\U00000398',
+        '\\Lambda'               : '\U0000039b',
+        '\\Xi'                   : '\U0000039e',
+        '\\Pi'                   : '\U000003a0',
+        '\\Sigma'                : '\U000003a3',
+        '\\Upsilon'              : '\U000003a5',
+        '\\Phi'                  : '\U000003a6',
+        '\\Psi'                  : '\U000003a8',
+        '\\Omega'                : '\U000003a9',
+        '\\leftarrow'            : '\U00002190',
+        '\\longleftarrow'        : '\U000027f5',
+        '\\rightarrow'           : '\U00002192',
+        '\\longrightarrow'       : '\U000027f6',
+        '\\Leftarrow'            : '\U000021d0',
+        '\\Longleftarrow'        : '\U000027f8',
+        '\\Rightarrow'           : '\U000021d2',
+        '\\Longrightarrow'       : '\U000027f9',
+        '\\leftrightarrow'       : '\U00002194',
+        '\\longleftrightarrow'   : '\U000027f7',
+        '\\Leftrightarrow'       : '\U000021d4',
+        '\\Longleftrightarrow'   : '\U000027fa',
+        '\\mapsto'               : '\U000021a6',
+        '\\longmapsto'           : '\U000027fc',
+        '\\relbar'               : '\U00002500',
+        '\\Relbar'               : '\U00002550',
+        '\\hookleftarrow'        : '\U000021a9',
+        '\\hookrightarrow'       : '\U000021aa',
+        '\\leftharpoondown'      : '\U000021bd',
+        '\\rightharpoondown'     : '\U000021c1',
+        '\\leftharpoonup'        : '\U000021bc',
+        '\\rightharpoonup'       : '\U000021c0',
+        '\\rightleftharpoons'    : '\U000021cc',
+        '\\leadsto'              : '\U0000219d',
+        '\\downharpoonleft'      : '\U000021c3',
+        '\\downharpoonright'     : '\U000021c2',
+        '\\upharpoonleft'        : '\U000021bf',
+        '\\upharpoonright'       : '\U000021be',
+        '\\restriction'          : '\U000021be',
+        '\\uparrow'              : '\U00002191',
+        '\\Uparrow'              : '\U000021d1',
+        '\\downarrow'            : '\U00002193',
+        '\\Downarrow'            : '\U000021d3',
+        '\\updownarrow'          : '\U00002195',
+        '\\Updownarrow'          : '\U000021d5',
+        '\\langle'               : '\U000027e8',
+        '\\rangle'               : '\U000027e9',
+        '\\lceil'                : '\U00002308',
+        '\\rceil'                : '\U00002309',
+        '\\lfloor'               : '\U0000230a',
+        '\\rfloor'               : '\U0000230b',
+        '\\flqq'                 : '\U000000ab',
+        '\\frqq'                 : '\U000000bb',
+        '\\bot'                  : '\U000022a5',
+        '\\top'                  : '\U000022a4',
+        '\\wedge'                : '\U00002227',
+        '\\bigwedge'             : '\U000022c0',
+        '\\vee'                  : '\U00002228',
+        '\\bigvee'               : '\U000022c1',
+        '\\forall'               : '\U00002200',
+        '\\exists'               : '\U00002203',
+        '\\nexists'              : '\U00002204',
+        '\\neg'                  : '\U000000ac',
+        '\\Box'                  : '\U000025a1',
+        '\\Diamond'              : '\U000025c7',
+        '\\vdash'                : '\U000022a2',
+        '\\models'               : '\U000022a8',
+        '\\dashv'                : '\U000022a3',
+        '\\surd'                 : '\U0000221a',
+        '\\le'                   : '\U00002264',
+        '\\ge'                   : '\U00002265',
+        '\\ll'                   : '\U0000226a',
+        '\\gg'                   : '\U0000226b',
+        '\\lesssim'              : '\U00002272',
+        '\\gtrsim'               : '\U00002273',
+        '\\lessapprox'           : '\U00002a85',
+        '\\gtrapprox'            : '\U00002a86',
+        '\\in'                   : '\U00002208',
+        '\\notin'                : '\U00002209',
+        '\\subset'               : '\U00002282',
+        '\\supset'               : '\U00002283',
+        '\\subseteq'             : '\U00002286',
+        '\\supseteq'             : '\U00002287',
+        '\\sqsubset'             : '\U0000228f',
+        '\\sqsupset'             : '\U00002290',
+        '\\sqsubseteq'           : '\U00002291',
+        '\\sqsupseteq'           : '\U00002292',
+        '\\cap'                  : '\U00002229',
+        '\\bigcap'               : '\U000022c2',
+        '\\cup'                  : '\U0000222a',
+        '\\bigcup'               : '\U000022c3',
+        '\\sqcup'                : '\U00002294',
+        '\\bigsqcup'             : '\U00002a06',
+        '\\sqcap'                : '\U00002293',
+        '\\Bigsqcap'             : '\U00002a05',
+        '\\setminus'             : '\U00002216',
+        '\\propto'               : '\U0000221d',
+        '\\uplus'                : '\U0000228e',
+        '\\bigplus'              : '\U00002a04',
+        '\\sim'                  : '\U0000223c',
+        '\\doteq'                : '\U00002250',
+        '\\simeq'                : '\U00002243',
+        '\\approx'               : '\U00002248',
+        '\\asymp'                : '\U0000224d',
+        '\\cong'                 : '\U00002245',
+        '\\equiv'                : '\U00002261',
+        '\\Join'                 : '\U000022c8',
+        '\\bowtie'               : '\U00002a1d',
+        '\\prec'                 : '\U0000227a',
+        '\\succ'                 : '\U0000227b',
+        '\\preceq'               : '\U0000227c',
+        '\\succeq'               : '\U0000227d',
+        '\\parallel'             : '\U00002225',
+        '\\mid'                  : '\U000000a6',
+        '\\pm'                   : '\U000000b1',
+        '\\mp'                   : '\U00002213',
+        '\\times'                : '\U000000d7',
+        '\\div'                  : '\U000000f7',
+        '\\cdot'                 : '\U000022c5',
+        '\\star'                 : '\U000022c6',
+        '\\circ'                 : '\U00002218',
+        '\\dagger'               : '\U00002020',
+        '\\ddagger'              : '\U00002021',
+        '\\lhd'                  : '\U000022b2',
+        '\\rhd'                  : '\U000022b3',
+        '\\unlhd'                : '\U000022b4',
+        '\\unrhd'                : '\U000022b5',
+        '\\triangleleft'         : '\U000025c3',
+        '\\triangleright'        : '\U000025b9',
+        '\\triangle'             : '\U000025b3',
+        '\\triangleq'            : '\U0000225c',
+        '\\oplus'                : '\U00002295',
+        '\\bigoplus'             : '\U00002a01',
+        '\\otimes'               : '\U00002297',
+        '\\bigotimes'            : '\U00002a02',
+        '\\odot'                 : '\U00002299',
+        '\\bigodot'              : '\U00002a00',
+        '\\ominus'               : '\U00002296',
+        '\\oslash'               : '\U00002298',
+        '\\dots'                 : '\U00002026',
+        '\\cdots'                : '\U000022ef',
+        '\\sum'                  : '\U00002211',
+        '\\prod'                 : '\U0000220f',
+        '\\coprod'               : '\U00002210',
+        '\\infty'                : '\U0000221e',
+        '\\int'                  : '\U0000222b',
+        '\\oint'                 : '\U0000222e',
+        '\\clubsuit'             : '\U00002663',
+        '\\diamondsuit'          : '\U00002662',
+        '\\heartsuit'            : '\U00002661',
+        '\\spadesuit'            : '\U00002660',
+        '\\aleph'                : '\U00002135',
+        '\\emptyset'             : '\U00002205',
+        '\\nabla'                : '\U00002207',
+        '\\partial'              : '\U00002202',
+        '\\flat'                 : '\U0000266d',
+        '\\natural'              : '\U0000266e',
+        '\\sharp'                : '\U0000266f',
+        '\\angle'                : '\U00002220',
+        '\\copyright'            : '\U000000a9',
+        '\\textregistered'       : '\U000000ae',
+        '\\textonequarter'       : '\U000000bc',
+        '\\textonehalf'          : '\U000000bd',
+        '\\textthreequarters'    : '\U000000be',
+        '\\textordfeminine'      : '\U000000aa',
+        '\\textordmasculine'     : '\U000000ba',
+        '\\euro'                 : '\U000020ac',
+        '\\pounds'               : '\U000000a3',
+        '\\yen'                  : '\U000000a5',
+        '\\textcent'             : '\U000000a2',
+        '\\textcurrency'         : '\U000000a4',
+        '\\textdegree'           : '\U000000b0',
+    }
+
+    isabelle_symbols = {
+        '\\<zero>'                 : '\U0001d7ec',
+        '\\<one>'                  : '\U0001d7ed',
+        '\\<two>'                  : '\U0001d7ee',
+        '\\<three>'                : '\U0001d7ef',
+        '\\<four>'                 : '\U0001d7f0',
+        '\\<five>'                 : '\U0001d7f1',
+        '\\<six>'                  : '\U0001d7f2',
+        '\\<seven>'                : '\U0001d7f3',
+        '\\<eight>'                : '\U0001d7f4',
+        '\\<nine>'                 : '\U0001d7f5',
+        '\\<A>'                    : '\U0001d49c',
+        '\\<B>'                    : '\U0000212c',
+        '\\<C>'                    : '\U0001d49e',
+        '\\<D>'                    : '\U0001d49f',
+        '\\<E>'                    : '\U00002130',
+        '\\<F>'                    : '\U00002131',
+        '\\<G>'                    : '\U0001d4a2',
+        '\\<H>'                    : '\U0000210b',
+        '\\<I>'                    : '\U00002110',
+        '\\<J>'                    : '\U0001d4a5',
+        '\\<K>'                    : '\U0001d4a6',
+        '\\<L>'                    : '\U00002112',
+        '\\<M>'                    : '\U00002133',
+        '\\<N>'                    : '\U0001d4a9',
+        '\\<O>'                    : '\U0001d4aa',
+        '\\<P>'                    : '\U0001d4ab',
+        '\\<Q>'                    : '\U0001d4ac',
+        '\\<R>'                    : '\U0000211b',
+        '\\<S>'                    : '\U0001d4ae',
+        '\\<T>'                    : '\U0001d4af',
+        '\\<U>'                    : '\U0001d4b0',
+        '\\<V>'                    : '\U0001d4b1',
+        '\\<W>'                    : '\U0001d4b2',
+        '\\<X>'                    : '\U0001d4b3',
+        '\\<Y>'                    : '\U0001d4b4',
+        '\\<Z>'                    : '\U0001d4b5',
+        '\\<a>'                    : '\U0001d5ba',
+        '\\<b>'                    : '\U0001d5bb',
+        '\\<c>'                    : '\U0001d5bc',
+        '\\<d>'                    : '\U0001d5bd',
+        '\\<e>'                    : '\U0001d5be',
+        '\\<f>'                    : '\U0001d5bf',
+        '\\<g>'                    : '\U0001d5c0',
+        '\\<h>'                    : '\U0001d5c1',
+        '\\<i>'                    : '\U0001d5c2',
+        '\\<j>'                    : '\U0001d5c3',
+        '\\<k>'                    : '\U0001d5c4',
+        '\\<l>'                    : '\U0001d5c5',
+        '\\<m>'                    : '\U0001d5c6',
+        '\\<n>'                    : '\U0001d5c7',
+        '\\<o>'                    : '\U0001d5c8',
+        '\\<p>'                    : '\U0001d5c9',
+        '\\<q>'                    : '\U0001d5ca',
+        '\\<r>'                    : '\U0001d5cb',
+        '\\<s>'                    : '\U0001d5cc',
+        '\\<t>'                    : '\U0001d5cd',
+        '\\<u>'                    : '\U0001d5ce',
+        '\\<v>'                    : '\U0001d5cf',
+        '\\<w>'                    : '\U0001d5d0',
+        '\\<x>'                    : '\U0001d5d1',
+        '\\<y>'                    : '\U0001d5d2',
+        '\\<z>'                    : '\U0001d5d3',
+        '\\<AA>'                   : '\U0001d504',
+        '\\<BB>'                   : '\U0001d505',
+        '\\<CC>'                   : '\U0000212d',
+        '\\<DD>'                   : '\U0001d507',
+        '\\<EE>'                   : '\U0001d508',
+        '\\<FF>'                   : '\U0001d509',
+        '\\<GG>'                   : '\U0001d50a',
+        '\\<HH>'                   : '\U0000210c',
+        '\\<II>'                   : '\U00002111',
+        '\\<JJ>'                   : '\U0001d50d',
+        '\\<KK>'                   : '\U0001d50e',
+        '\\<LL>'                   : '\U0001d50f',
+        '\\<MM>'                   : '\U0001d510',
+        '\\<NN>'                   : '\U0001d511',
+        '\\<OO>'                   : '\U0001d512',
+        '\\<PP>'                   : '\U0001d513',
+        '\\<QQ>'                   : '\U0001d514',
+        '\\<RR>'                   : '\U0000211c',
+        '\\<SS>'                   : '\U0001d516',
+        '\\<TT>'                   : '\U0001d517',
+        '\\<UU>'                   : '\U0001d518',
+        '\\<VV>'                   : '\U0001d519',
+        '\\<WW>'                   : '\U0001d51a',
+        '\\<XX>'                   : '\U0001d51b',
+        '\\<YY>'                   : '\U0001d51c',
+        '\\<ZZ>'                   : '\U00002128',
+        '\\<aa>'                   : '\U0001d51e',
+        '\\<bb>'                   : '\U0001d51f',
+        '\\<cc>'                   : '\U0001d520',
+        '\\<dd>'                   : '\U0001d521',
+        '\\<ee>'                   : '\U0001d522',
+        '\\<ff>'                   : '\U0001d523',
+        '\\<gg>'                   : '\U0001d524',
+        '\\<hh>'                   : '\U0001d525',
+        '\\<ii>'                   : '\U0001d526',
+        '\\<jj>'                   : '\U0001d527',
+        '\\<kk>'                   : '\U0001d528',
+        '\\<ll>'                   : '\U0001d529',
+        '\\<mm>'                   : '\U0001d52a',
+        '\\<nn>'                   : '\U0001d52b',
+        '\\<oo>'                   : '\U0001d52c',
+        '\\<pp>'                   : '\U0001d52d',
+        '\\<qq>'                   : '\U0001d52e',
+        '\\<rr>'                   : '\U0001d52f',
+        '\\<ss>'                   : '\U0001d530',
+        '\\<tt>'                   : '\U0001d531',
+        '\\<uu>'                   : '\U0001d532',
+        '\\<vv>'                   : '\U0001d533',
+        '\\<ww>'                   : '\U0001d534',
+        '\\<xx>'                   : '\U0001d535',
+        '\\<yy>'                   : '\U0001d536',
+        '\\<zz>'                   : '\U0001d537',
+        '\\<alpha>'                : '\U000003b1',
+        '\\<beta>'                 : '\U000003b2',
+        '\\<gamma>'                : '\U000003b3',
+        '\\<delta>'                : '\U000003b4',
+        '\\<epsilon>'              : '\U000003b5',
+        '\\<zeta>'                 : '\U000003b6',
+        '\\<eta>'                  : '\U000003b7',
+        '\\<theta>'                : '\U000003b8',
+        '\\<iota>'                 : '\U000003b9',
+        '\\<kappa>'                : '\U000003ba',
+        '\\<lambda>'               : '\U000003bb',
+        '\\<mu>'                   : '\U000003bc',
+        '\\<nu>'                   : '\U000003bd',
+        '\\<xi>'                   : '\U000003be',
+        '\\<pi>'                   : '\U000003c0',
+        '\\<rho>'                  : '\U000003c1',
+        '\\<sigma>'                : '\U000003c3',
+        '\\<tau>'                  : '\U000003c4',
+        '\\<upsilon>'              : '\U000003c5',
+        '\\<phi>'                  : '\U000003c6',
+        '\\<chi>'                  : '\U000003c7',
+        '\\<psi>'                  : '\U000003c8',
+        '\\<omega>'                : '\U000003c9',
+        '\\<Gamma>'                : '\U00000393',
+        '\\<Delta>'                : '\U00000394',
+        '\\<Theta>'                : '\U00000398',
+        '\\<Lambda>'               : '\U0000039b',
+        '\\<Xi>'                   : '\U0000039e',
+        '\\<Pi>'                   : '\U000003a0',
+        '\\<Sigma>'                : '\U000003a3',
+        '\\<Upsilon>'              : '\U000003a5',
+        '\\<Phi>'                  : '\U000003a6',
+        '\\<Psi>'                  : '\U000003a8',
+        '\\<Omega>'                : '\U000003a9',
+        '\\<bool>'                 : '\U0001d539',
+        '\\<complex>'              : '\U00002102',
+        '\\<nat>'                  : '\U00002115',
+        '\\<rat>'                  : '\U0000211a',
+        '\\<real>'                 : '\U0000211d',
+        '\\<int>'                  : '\U00002124',
+        '\\<leftarrow>'            : '\U00002190',
+        '\\<longleftarrow>'        : '\U000027f5',
+        '\\<rightarrow>'           : '\U00002192',
+        '\\<longrightarrow>'       : '\U000027f6',
+        '\\<Leftarrow>'            : '\U000021d0',
+        '\\<Longleftarrow>'        : '\U000027f8',
+        '\\<Rightarrow>'           : '\U000021d2',
+        '\\<Longrightarrow>'       : '\U000027f9',
+        '\\<leftrightarrow>'       : '\U00002194',
+        '\\<longleftrightarrow>'   : '\U000027f7',
+        '\\<Leftrightarrow>'       : '\U000021d4',
+        '\\<Longleftrightarrow>'   : '\U000027fa',
+        '\\<mapsto>'               : '\U000021a6',
+        '\\<longmapsto>'           : '\U000027fc',
+        '\\<midarrow>'             : '\U00002500',
+        '\\<Midarrow>'             : '\U00002550',
+        '\\<hookleftarrow>'        : '\U000021a9',
+        '\\<hookrightarrow>'       : '\U000021aa',
+        '\\<leftharpoondown>'      : '\U000021bd',
+        '\\<rightharpoondown>'     : '\U000021c1',
+        '\\<leftharpoonup>'        : '\U000021bc',
+        '\\<rightharpoonup>'       : '\U000021c0',
+        '\\<rightleftharpoons>'    : '\U000021cc',
+        '\\<leadsto>'              : '\U0000219d',
+        '\\<downharpoonleft>'      : '\U000021c3',
+        '\\<downharpoonright>'     : '\U000021c2',
+        '\\<upharpoonleft>'        : '\U000021bf',
+        '\\<upharpoonright>'       : '\U000021be',
+        '\\<restriction>'          : '\U000021be',
+        '\\<Colon>'                : '\U00002237',
+        '\\<up>'                   : '\U00002191',
+        '\\<Up>'                   : '\U000021d1',
+        '\\<down>'                 : '\U00002193',
+        '\\<Down>'                 : '\U000021d3',
+        '\\<updown>'               : '\U00002195',
+        '\\<Updown>'               : '\U000021d5',
+        '\\<langle>'               : '\U000027e8',
+        '\\<rangle>'               : '\U000027e9',
+        '\\<lceil>'                : '\U00002308',
+        '\\<rceil>'                : '\U00002309',
+        '\\<lfloor>'               : '\U0000230a',
+        '\\<rfloor>'               : '\U0000230b',
+        '\\<lparr>'                : '\U00002987',
+        '\\<rparr>'                : '\U00002988',
+        '\\<lbrakk>'               : '\U000027e6',
+        '\\<rbrakk>'               : '\U000027e7',
+        '\\<lbrace>'               : '\U00002983',
+        '\\<rbrace>'               : '\U00002984',
+        '\\<guillemotleft>'        : '\U000000ab',
+        '\\<guillemotright>'       : '\U000000bb',
+        '\\<bottom>'               : '\U000022a5',
+        '\\<top>'                  : '\U000022a4',
+        '\\<and>'                  : '\U00002227',
+        '\\<And>'                  : '\U000022c0',
+        '\\<or>'                   : '\U00002228',
+        '\\<Or>'                   : '\U000022c1',
+        '\\<forall>'               : '\U00002200',
+        '\\<exists>'               : '\U00002203',
+        '\\<nexists>'              : '\U00002204',
+        '\\<not>'                  : '\U000000ac',
+        '\\<box>'                  : '\U000025a1',
+        '\\<diamond>'              : '\U000025c7',
+        '\\<turnstile>'            : '\U000022a2',
+        '\\<Turnstile>'            : '\U000022a8',
+        '\\<tturnstile>'           : '\U000022a9',
+        '\\<TTurnstile>'           : '\U000022ab',
+        '\\<stileturn>'            : '\U000022a3',
+        '\\<surd>'                 : '\U0000221a',
+        '\\<le>'                   : '\U00002264',
+        '\\<ge>'                   : '\U00002265',
+        '\\<lless>'                : '\U0000226a',
+        '\\<ggreater>'             : '\U0000226b',
+        '\\<lesssim>'              : '\U00002272',
+        '\\<greatersim>'           : '\U00002273',
+        '\\<lessapprox>'           : '\U00002a85',
+        '\\<greaterapprox>'        : '\U00002a86',
+        '\\<in>'                   : '\U00002208',
+        '\\<notin>'                : '\U00002209',
+        '\\<subset>'               : '\U00002282',
+        '\\<supset>'               : '\U00002283',
+        '\\<subseteq>'             : '\U00002286',
+        '\\<supseteq>'             : '\U00002287',
+        '\\<sqsubset>'             : '\U0000228f',
+        '\\<sqsupset>'             : '\U00002290',
+        '\\<sqsubseteq>'           : '\U00002291',
+        '\\<sqsupseteq>'           : '\U00002292',
+        '\\<inter>'                : '\U00002229',
+        '\\<Inter>'                : '\U000022c2',
+        '\\<union>'                : '\U0000222a',
+        '\\<Union>'                : '\U000022c3',
+        '\\<squnion>'              : '\U00002294',
+        '\\<Squnion>'              : '\U00002a06',
+        '\\<sqinter>'              : '\U00002293',
+        '\\<Sqinter>'              : '\U00002a05',
+        '\\<setminus>'             : '\U00002216',
+        '\\<propto>'               : '\U0000221d',
+        '\\<uplus>'                : '\U0000228e',
+        '\\<Uplus>'                : '\U00002a04',
+        '\\<noteq>'                : '\U00002260',
+        '\\<sim>'                  : '\U0000223c',
+        '\\<doteq>'                : '\U00002250',
+        '\\<simeq>'                : '\U00002243',
+        '\\<approx>'               : '\U00002248',
+        '\\<asymp>'                : '\U0000224d',
+        '\\<cong>'                 : '\U00002245',
+        '\\<smile>'                : '\U00002323',
+        '\\<equiv>'                : '\U00002261',
+        '\\<frown>'                : '\U00002322',
+        '\\<Join>'                 : '\U000022c8',
+        '\\<bowtie>'               : '\U00002a1d',
+        '\\<prec>'                 : '\U0000227a',
+        '\\<succ>'                 : '\U0000227b',
+        '\\<preceq>'               : '\U0000227c',
+        '\\<succeq>'               : '\U0000227d',
+        '\\<parallel>'             : '\U00002225',
+        '\\<bar>'                  : '\U000000a6',
+        '\\<plusminus>'            : '\U000000b1',
+        '\\<minusplus>'            : '\U00002213',
+        '\\<times>'                : '\U000000d7',
+        '\\<div>'                  : '\U000000f7',
+        '\\<cdot>'                 : '\U000022c5',
+        '\\<star>'                 : '\U000022c6',
+        '\\<bullet>'               : '\U00002219',
+        '\\<circ>'                 : '\U00002218',
+        '\\<dagger>'               : '\U00002020',
+        '\\<ddagger>'              : '\U00002021',
+        '\\<lhd>'                  : '\U000022b2',
+        '\\<rhd>'                  : '\U000022b3',
+        '\\<unlhd>'                : '\U000022b4',
+        '\\<unrhd>'                : '\U000022b5',
+        '\\<triangleleft>'         : '\U000025c3',
+        '\\<triangleright>'        : '\U000025b9',
+        '\\<triangle>'             : '\U000025b3',
+        '\\<triangleq>'            : '\U0000225c',
+        '\\<oplus>'                : '\U00002295',
+        '\\<Oplus>'                : '\U00002a01',
+        '\\<otimes>'               : '\U00002297',
+        '\\<Otimes>'               : '\U00002a02',
+        '\\<odot>'                 : '\U00002299',
+        '\\<Odot>'                 : '\U00002a00',
+        '\\<ominus>'               : '\U00002296',
+        '\\<oslash>'               : '\U00002298',
+        '\\<dots>'                 : '\U00002026',
+        '\\<cdots>'                : '\U000022ef',
+        '\\<Sum>'                  : '\U00002211',
+        '\\<Prod>'                 : '\U0000220f',
+        '\\<Coprod>'               : '\U00002210',
+        '\\<infinity>'             : '\U0000221e',
+        '\\<integral>'             : '\U0000222b',
+        '\\<ointegral>'            : '\U0000222e',
+        '\\<clubsuit>'             : '\U00002663',
+        '\\<diamondsuit>'          : '\U00002662',
+        '\\<heartsuit>'            : '\U00002661',
+        '\\<spadesuit>'            : '\U00002660',
+        '\\<aleph>'                : '\U00002135',
+        '\\<emptyset>'             : '\U00002205',
+        '\\<nabla>'                : '\U00002207',
+        '\\<partial>'              : '\U00002202',
+        '\\<flat>'                 : '\U0000266d',
+        '\\<natural>'              : '\U0000266e',
+        '\\<sharp>'                : '\U0000266f',
+        '\\<angle>'                : '\U00002220',
+        '\\<copyright>'            : '\U000000a9',
+        '\\<registered>'           : '\U000000ae',
+        '\\<hyphen>'               : '\U000000ad',
+        '\\<inverse>'              : '\U000000af',
+        '\\<onequarter>'           : '\U000000bc',
+        '\\<onehalf>'              : '\U000000bd',
+        '\\<threequarters>'        : '\U000000be',
+        '\\<ordfeminine>'          : '\U000000aa',
+        '\\<ordmasculine>'         : '\U000000ba',
+        '\\<section>'              : '\U000000a7',
+        '\\<paragraph>'            : '\U000000b6',
+        '\\<exclamdown>'           : '\U000000a1',
+        '\\<questiondown>'         : '\U000000bf',
+        '\\<euro>'                 : '\U000020ac',
+        '\\<pounds>'               : '\U000000a3',
+        '\\<yen>'                  : '\U000000a5',
+        '\\<cent>'                 : '\U000000a2',
+        '\\<currency>'             : '\U000000a4',
+        '\\<degree>'               : '\U000000b0',
+        '\\<amalg>'                : '\U00002a3f',
+        '\\<mho>'                  : '\U00002127',
+        '\\<lozenge>'              : '\U000025ca',
+        '\\<wp>'                   : '\U00002118',
+        '\\<wrong>'                : '\U00002240',
+        '\\<struct>'               : '\U000022c4',
+        '\\<acute>'                : '\U000000b4',
+        '\\<index>'                : '\U00000131',
+        '\\<dieresis>'             : '\U000000a8',
+        '\\<cedilla>'              : '\U000000b8',
+        '\\<hungarumlaut>'         : '\U000002dd',
+        '\\<some>'                 : '\U000003f5',
+        '\\<newline>'              : '\U000023ce',
+        '\\<open>'                 : '\U00002039',
+        '\\<close>'                : '\U0000203a',
+        '\\<here>'                 : '\U00002302',
+        '\\<^sub>'                 : '\U000021e9',
+        '\\<^sup>'                 : '\U000021e7',
+        '\\<^bold>'                : '\U00002759',
+        '\\<^bsub>'                : '\U000021d8',
+        '\\<^esub>'                : '\U000021d9',
+        '\\<^bsup>'                : '\U000021d7',
+        '\\<^esup>'                : '\U000021d6',
+    }
+
+    lang_map = {'isabelle' : isabelle_symbols, 'latex' : latex_symbols}
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        lang = get_choice_opt(options, 'lang',
+                              ['isabelle', 'latex'], 'isabelle')
+        self.symbols = self.lang_map[lang]
+
+    def filter(self, lexer, stream):
+        for ttype, value in stream:
+            if value in self.symbols:
+                yield ttype, self.symbols[value]
+            else:
+                yield ttype, value
+
+
+class KeywordCaseFilter(Filter):
+    """Convert keywords to lowercase or uppercase or capitalize them, which
+    means first letter uppercase, rest lowercase.
+
+    This can be useful e.g. if you highlight Pascal code and want to adapt the
+    code to your styleguide.
+
+    Options accepted:
+
+    `case` : string
+       The casing to convert keywords to. Must be one of ``'lower'``,
+       ``'upper'`` or ``'capitalize'``.  The default is ``'lower'``.
+    """
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        case = get_choice_opt(options, 'case',
+                              ['lower', 'upper', 'capitalize'], 'lower')
+        self.convert = getattr(str, case)
+
+    def filter(self, lexer, stream):
+        for ttype, value in stream:
+            if ttype in Keyword:
+                yield ttype, self.convert(value)
+            else:
+                yield ttype, value
+
+
+class NameHighlightFilter(Filter):
+    """Highlight a normal Name (and Name.*) token with a different token type.
+
+    Example::
+
+        filter = NameHighlightFilter(
+            names=['foo', 'bar', 'baz'],
+            tokentype=Name.Function,
+        )
+
+    This would highlight the names "foo", "bar" and "baz"
+    as functions. `Name.Function` is the default token type.
+
+    Options accepted:
+
+    `names` : list of strings
+      A list of names that should be given the different token type.
+      There is no default.
+    `tokentype` : TokenType or string
+      A token type or a string containing a token type name that is
+      used for highlighting the strings in `names`.  The default is
+      `Name.Function`.
+    """
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        self.names = set(get_list_opt(options, 'names', []))
+        tokentype = options.get('tokentype')
+        if tokentype:
+            self.tokentype = string_to_tokentype(tokentype)
+        else:
+            self.tokentype = Name.Function
+
+    def filter(self, lexer, stream):
+        for ttype, value in stream:
+            if ttype in Name and value in self.names:
+                yield self.tokentype, value
+            else:
+                yield ttype, value
+
+
+class ErrorToken(Exception):
+    pass
+
+
+class RaiseOnErrorTokenFilter(Filter):
+    """Raise an exception when the lexer generates an error token.
+
+    Options accepted:
+
+    `excclass` : Exception class
+      The exception class to raise.
+      The default is `pygments.filters.ErrorToken`.
+
+    .. versionadded:: 0.8
+    """
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        self.exception = options.get('excclass', ErrorToken)
+        try:
+            # issubclass() will raise TypeError if first argument is not a class
+            if not issubclass(self.exception, Exception):
+                raise TypeError
+        except TypeError:
+            raise OptionError('excclass option is not an exception class')
+
+    def filter(self, lexer, stream):
+        for ttype, value in stream:
+            if ttype is Error:
+                raise self.exception(value)
+            yield ttype, value
+
+
+class VisibleWhitespaceFilter(Filter):
+    """Convert tabs, newlines and/or spaces to visible characters.
+
+    Options accepted:
+
+    `spaces` : string or bool
+      If this is a one-character string, spaces will be replaces by this string.
+      If it is another true value, spaces will be replaced by ``·`` (unicode
+      MIDDLE DOT).  If it is a false value, spaces will not be replaced.  The
+      default is ``False``.
+    `tabs` : string or bool
+      The same as for `spaces`, but the default replacement character is ``»``
+      (unicode RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK).  The default value
+      is ``False``.  Note: this will not work if the `tabsize` option for the
+      lexer is nonzero, as tabs will already have been expanded then.
+    `tabsize` : int
+      If tabs are to be replaced by this filter (see the `tabs` option), this
+      is the total number of characters that a tab should be expanded to.
+      The default is ``8``.
+    `newlines` : string or bool
+      The same as for `spaces`, but the default replacement character is ``¶``
+      (unicode PILCROW SIGN).  The default value is ``False``.
+    `wstokentype` : bool
+      If true, give whitespace the special `Whitespace` token type.  This allows
+      styling the visible whitespace differently (e.g. greyed out), but it can
+      disrupt background colors.  The default is ``True``.
+
+    .. versionadded:: 0.8
+    """
+
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        for name, default in [('spaces',   '·'),
+                              ('tabs',     '»'),
+                              ('newlines', '¶')]:
+            opt = options.get(name, False)
+            if isinstance(opt, str) and len(opt) == 1:
+                setattr(self, name, opt)
+            else:
+                setattr(self, name, (opt and default or ''))
+        tabsize = get_int_opt(options, 'tabsize', 8)
+        if self.tabs:
+            self.tabs += ' ' * (tabsize - 1)
+        if self.newlines:
+            self.newlines += '\n'
+        self.wstt = get_bool_opt(options, 'wstokentype', True)
+
+    def filter(self, lexer, stream):
+        if self.wstt:
+            spaces = self.spaces or ' '
+            tabs = self.tabs or '\t'
+            newlines = self.newlines or '\n'
+            regex = re.compile(r'\s')
+
+            def replacefunc(wschar):
+                if wschar == ' ':
+                    return spaces
+                elif wschar == '\t':
+                    return tabs
+                elif wschar == '\n':
+                    return newlines
+                return wschar
+
+            for ttype, value in stream:
+                yield from _replace_special(ttype, value, regex, Whitespace,
+                                            replacefunc)
+        else:
+            spaces, tabs, newlines = self.spaces, self.tabs, self.newlines
+            # simpler processing
+            for ttype, value in stream:
+                if spaces:
+                    value = value.replace(' ', spaces)
+                if tabs:
+                    value = value.replace('\t', tabs)
+                if newlines:
+                    value = value.replace('\n', newlines)
+                yield ttype, value
+
+
+class GobbleFilter(Filter):
+    """Gobbles source code lines (eats initial characters).
+
+    This filter drops the first ``n`` characters off every line of code.  This
+    may be useful when the source code fed to the lexer is indented by a fixed
+    amount of space that isn't desired in the output.
+
+    Options accepted:
+
+    `n` : int
+       The number of characters to gobble.
+
+    .. versionadded:: 1.2
+    """
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+        self.n = get_int_opt(options, 'n', 0)
+
+    def gobble(self, value, left):
+        if left < len(value):
+            return value[left:], 0
+        else:
+            return '', left - len(value)
+
+    def filter(self, lexer, stream):
+        n = self.n
+        left = n  # How many characters left to gobble.
+        for ttype, value in stream:
+            # Remove ``left`` tokens from first line, ``n`` from all others.
+            parts = value.split('\n')
+            (parts[0], left) = self.gobble(parts[0], left)
+            for i in range(1, len(parts)):
+                (parts[i], left) = self.gobble(parts[i], n)
+            value = '\n'.join(parts)
+
+            if value != '':
+                yield ttype, value
+
+
+class TokenMergeFilter(Filter):
+    """Merges consecutive tokens with the same token type in the output
+    stream of a lexer.
+
+    .. versionadded:: 1.2
+    """
+    def __init__(self, **options):
+        Filter.__init__(self, **options)
+
+    def filter(self, lexer, stream):
+        current_type = None
+        current_value = None
+        for ttype, value in stream:
+            if ttype is current_type:
+                current_value += value
+            else:
+                if current_type is not None:
+                    yield current_type, current_value
+                current_type = ttype
+                current_value = value
+        if current_type is not None:
+            yield current_type, current_value
+
+
+FILTERS = {
+    'codetagify':     CodeTagFilter,
+    'keywordcase':    KeywordCaseFilter,
+    'highlight':      NameHighlightFilter,
+    'raiseonerror':   RaiseOnErrorTokenFilter,
+    'whitespace':     VisibleWhitespaceFilter,
+    'gobble':         GobbleFilter,
+    'tokenmerge':     TokenMergeFilter,
+    'symbols':        SymbolFilter,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b830e3f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/filters/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatter.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatter.py
new file mode 100644
index 00000000..d2666037
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatter.py
@@ -0,0 +1,129 @@
+"""
+    pygments.formatter
+    ~~~~~~~~~~~~~~~~~~
+
+    Base formatter class.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import codecs
+
+from pip._vendor.pygments.util import get_bool_opt
+from pip._vendor.pygments.styles import get_style_by_name
+
+__all__ = ['Formatter']
+
+
+def _lookup_style(style):
+    if isinstance(style, str):
+        return get_style_by_name(style)
+    return style
+
+
+class Formatter:
+    """
+    Converts a token stream to text.
+
+    Formatters should have attributes to help selecting them. These
+    are similar to the corresponding :class:`~pygments.lexer.Lexer`
+    attributes.
+
+    .. autoattribute:: name
+       :no-value:
+
+    .. autoattribute:: aliases
+       :no-value:
+
+    .. autoattribute:: filenames
+       :no-value:
+
+    You can pass options as keyword arguments to the constructor.
+    All formatters accept these basic options:
+
+    ``style``
+        The style to use, can be a string or a Style subclass
+        (default: "default"). Not used by e.g. the
+        TerminalFormatter.
+    ``full``
+        Tells the formatter to output a "full" document, i.e.
+        a complete self-contained document. This doesn't have
+        any effect for some formatters (default: false).
+    ``title``
+        If ``full`` is true, the title that should be used to
+        caption the document (default: '').
+    ``encoding``
+        If given, must be an encoding name. This will be used to
+        convert the Unicode token strings to byte strings in the
+        output. If it is "" or None, Unicode strings will be written
+        to the output file, which most file-like objects do not
+        support (default: None).
+    ``outencoding``
+        Overrides ``encoding`` if given.
+
+    """
+
+    #: Full name for the formatter, in human-readable form.
+    name = None
+
+    #: A list of short, unique identifiers that can be used to lookup
+    #: the formatter from a list, e.g. using :func:`.get_formatter_by_name()`.
+    aliases = []
+
+    #: A list of fnmatch patterns that match filenames for which this
+    #: formatter can produce output. The patterns in this list should be unique
+    #: among all formatters.
+    filenames = []
+
+    #: If True, this formatter outputs Unicode strings when no encoding
+    #: option is given.
+    unicodeoutput = True
+
+    def __init__(self, **options):
+        """
+        As with lexers, this constructor takes arbitrary optional arguments,
+        and if you override it, you should first process your own options, then
+        call the base class implementation.
+        """
+        self.style = _lookup_style(options.get('style', 'default'))
+        self.full = get_bool_opt(options, 'full', False)
+        self.title = options.get('title', '')
+        self.encoding = options.get('encoding', None) or None
+        if self.encoding in ('guess', 'chardet'):
+            # can happen for e.g. pygmentize -O encoding=guess
+            self.encoding = 'utf-8'
+        self.encoding = options.get('outencoding') or self.encoding
+        self.options = options
+
+    def get_style_defs(self, arg=''):
+        """
+        This method must return statements or declarations suitable to define
+        the current style for subsequent highlighted text (e.g. CSS classes
+        in the `HTMLFormatter`).
+
+        The optional argument `arg` can be used to modify the generation and
+        is formatter dependent (it is standardized because it can be given on
+        the command line).
+
+        This method is called by the ``-S`` :doc:`command-line option <cmdline>`,
+        the `arg` is then given by the ``-a`` option.
+        """
+        return ''
+
+    def format(self, tokensource, outfile):
+        """
+        This method must format the tokens from the `tokensource` iterable and
+        write the formatted version to the file object `outfile`.
+
+        Formatter options can control how exactly the tokens are converted.
+        """
+        if self.encoding:
+            # wrap the outfile in a StreamWriter
+            outfile = codecs.lookup(self.encoding)[3](outfile)
+        return self.format_unencoded(tokensource, outfile)
+
+    # Allow writing Formatter[str] or Formatter[bytes]. That's equivalent to
+    # Formatter. This helps when using third-party type stubs from typeshed.
+    def __class_getitem__(cls, name):
+        return cls
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__init__.py
new file mode 100644
index 00000000..f19e9931
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__init__.py
@@ -0,0 +1,157 @@
+"""
+    pygments.formatters
+    ~~~~~~~~~~~~~~~~~~~
+
+    Pygments formatters.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+import sys
+import types
+import fnmatch
+from os.path import basename
+
+from pip._vendor.pygments.formatters._mapping import FORMATTERS
+from pip._vendor.pygments.plugin import find_plugin_formatters
+from pip._vendor.pygments.util import ClassNotFound
+
+__all__ = ['get_formatter_by_name', 'get_formatter_for_filename',
+           'get_all_formatters', 'load_formatter_from_file'] + list(FORMATTERS)
+
+_formatter_cache = {}  # classes by name
+_pattern_cache = {}
+
+
+def _fn_matches(fn, glob):
+    """Return whether the supplied file name fn matches pattern filename."""
+    if glob not in _pattern_cache:
+        pattern = _pattern_cache[glob] = re.compile(fnmatch.translate(glob))
+        return pattern.match(fn)
+    return _pattern_cache[glob].match(fn)
+
+
+def _load_formatters(module_name):
+    """Load a formatter (and all others in the module too)."""
+    mod = __import__(module_name, None, None, ['__all__'])
+    for formatter_name in mod.__all__:
+        cls = getattr(mod, formatter_name)
+        _formatter_cache[cls.name] = cls
+
+
+def get_all_formatters():
+    """Return a generator for all formatter classes."""
+    # NB: this returns formatter classes, not info like get_all_lexers().
+    for info in FORMATTERS.values():
+        if info[1] not in _formatter_cache:
+            _load_formatters(info[0])
+        yield _formatter_cache[info[1]]
+    for _, formatter in find_plugin_formatters():
+        yield formatter
+
+
+def find_formatter_class(alias):
+    """Lookup a formatter by alias.
+
+    Returns None if not found.
+    """
+    for module_name, name, aliases, _, _ in FORMATTERS.values():
+        if alias in aliases:
+            if name not in _formatter_cache:
+                _load_formatters(module_name)
+            return _formatter_cache[name]
+    for _, cls in find_plugin_formatters():
+        if alias in cls.aliases:
+            return cls
+
+
+def get_formatter_by_name(_alias, **options):
+    """
+    Return an instance of a :class:`.Formatter` subclass that has `alias` in its
+    aliases list. The formatter is given the `options` at its instantiation.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if no formatter with that
+    alias is found.
+    """
+    cls = find_formatter_class(_alias)
+    if cls is None:
+        raise ClassNotFound(f"no formatter found for name {_alias!r}")
+    return cls(**options)
+
+
+def load_formatter_from_file(filename, formattername="CustomFormatter", **options):
+    """
+    Return a `Formatter` subclass instance loaded from the provided file, relative
+    to the current directory.
+
+    The file is expected to contain a Formatter class named ``formattername``
+    (by default, CustomFormatter). Users should be very careful with the input, because
+    this method is equivalent to running ``eval()`` on the input file. The formatter is
+    given the `options` at its instantiation.
+
+    :exc:`pygments.util.ClassNotFound` is raised if there are any errors loading
+    the formatter.
+
+    .. versionadded:: 2.2
+    """
+    try:
+        # This empty dict will contain the namespace for the exec'd file
+        custom_namespace = {}
+        with open(filename, 'rb') as f:
+            exec(f.read(), custom_namespace)
+        # Retrieve the class `formattername` from that namespace
+        if formattername not in custom_namespace:
+            raise ClassNotFound(f'no valid {formattername} class found in {filename}')
+        formatter_class = custom_namespace[formattername]
+        # And finally instantiate it with the options
+        return formatter_class(**options)
+    except OSError as err:
+        raise ClassNotFound(f'cannot read {filename}: {err}')
+    except ClassNotFound:
+        raise
+    except Exception as err:
+        raise ClassNotFound(f'error when loading custom formatter: {err}')
+
+
+def get_formatter_for_filename(fn, **options):
+    """
+    Return a :class:`.Formatter` subclass instance that has a filename pattern
+    matching `fn`. The formatter is given the `options` at its instantiation.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if no formatter for that filename
+    is found.
+    """
+    fn = basename(fn)
+    for modname, name, _, filenames, _ in FORMATTERS.values():
+        for filename in filenames:
+            if _fn_matches(fn, filename):
+                if name not in _formatter_cache:
+                    _load_formatters(modname)
+                return _formatter_cache[name](**options)
+    for _name, cls in find_plugin_formatters():
+        for filename in cls.filenames:
+            if _fn_matches(fn, filename):
+                return cls(**options)
+    raise ClassNotFound(f"no formatter found for file name {fn!r}")
+
+
+class _automodule(types.ModuleType):
+    """Automatically import formatters."""
+
+    def __getattr__(self, name):
+        info = FORMATTERS.get(name)
+        if info:
+            _load_formatters(info[0])
+            cls = _formatter_cache[info[1]]
+            setattr(self, name, cls)
+            return cls
+        raise AttributeError(name)
+
+
+oldmod = sys.modules[__name__]
+newmod = _automodule(__name__)
+newmod.__dict__.update(oldmod.__dict__)
+sys.modules[__name__] = newmod
+del newmod.newmod, newmod.oldmod, newmod.sys, newmod.types
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9032f913
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/_mapping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/_mapping.cpython-312.pyc
new file mode 100644
index 00000000..9e3af21e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/_mapping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/bbcode.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/bbcode.cpython-312.pyc
new file mode 100644
index 00000000..d4b99c88
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/bbcode.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/groff.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/groff.cpython-312.pyc
new file mode 100644
index 00000000..b81218ff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/groff.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/html.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/html.cpython-312.pyc
new file mode 100644
index 00000000..3de511ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/html.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/img.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/img.cpython-312.pyc
new file mode 100644
index 00000000..851be4e5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/img.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/irc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/irc.cpython-312.pyc
new file mode 100644
index 00000000..e5ce7872
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/irc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/latex.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/latex.cpython-312.pyc
new file mode 100644
index 00000000..33230f2d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/latex.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/other.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/other.cpython-312.pyc
new file mode 100644
index 00000000..36d3f2a8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/other.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/pangomarkup.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/pangomarkup.cpython-312.pyc
new file mode 100644
index 00000000..a4695e7a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/pangomarkup.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/rtf.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/rtf.cpython-312.pyc
new file mode 100644
index 00000000..b6925a28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/rtf.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/svg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/svg.cpython-312.pyc
new file mode 100644
index 00000000..b5f859d9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/svg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal.cpython-312.pyc
new file mode 100644
index 00000000..dbc3b451
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal256.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal256.cpython-312.pyc
new file mode 100644
index 00000000..36f07013
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/__pycache__/terminal256.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/_mapping.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/_mapping.py
new file mode 100644
index 00000000..72ca8404
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/_mapping.py
@@ -0,0 +1,23 @@
+# Automatically generated by scripts/gen_mapfiles.py.
+# DO NOT EDIT BY HAND; run `tox -e mapfiles` instead.
+
+FORMATTERS = {
+    'BBCodeFormatter': ('pygments.formatters.bbcode', 'BBCode', ('bbcode', 'bb'), (), 'Format tokens with BBcodes. These formatting codes are used by many bulletin boards, so you can highlight your sourcecode with pygments before posting it there.'),
+    'BmpImageFormatter': ('pygments.formatters.img', 'img_bmp', ('bmp', 'bitmap'), ('*.bmp',), 'Create a bitmap image from source code. This uses the Python Imaging Library to generate a pixmap from the source code.'),
+    'GifImageFormatter': ('pygments.formatters.img', 'img_gif', ('gif',), ('*.gif',), 'Create a GIF image from source code. This uses the Python Imaging Library to generate a pixmap from the source code.'),
+    'GroffFormatter': ('pygments.formatters.groff', 'groff', ('groff', 'troff', 'roff'), (), 'Format tokens with groff escapes to change their color and font style.'),
+    'HtmlFormatter': ('pygments.formatters.html', 'HTML', ('html',), ('*.html', '*.htm'), "Format tokens as HTML 4 ``<span>`` tags. By default, the content is enclosed in a ``<pre>`` tag, itself wrapped in a ``<div>`` tag (but see the `nowrap` option). The ``<div>``'s CSS class can be set by the `cssclass` option."),
+    'IRCFormatter': ('pygments.formatters.irc', 'IRC', ('irc', 'IRC'), (), 'Format tokens with IRC color sequences'),
+    'ImageFormatter': ('pygments.formatters.img', 'img', ('img', 'IMG', 'png'), ('*.png',), 'Create a PNG image from source code. This uses the Python Imaging Library to generate a pixmap from the source code.'),
+    'JpgImageFormatter': ('pygments.formatters.img', 'img_jpg', ('jpg', 'jpeg'), ('*.jpg',), 'Create a JPEG image from source code. This uses the Python Imaging Library to generate a pixmap from the source code.'),
+    'LatexFormatter': ('pygments.formatters.latex', 'LaTeX', ('latex', 'tex'), ('*.tex',), 'Format tokens as LaTeX code. This needs the `fancyvrb` and `color` standard packages.'),
+    'NullFormatter': ('pygments.formatters.other', 'Text only', ('text', 'null'), ('*.txt',), 'Output the text unchanged without any formatting.'),
+    'PangoMarkupFormatter': ('pygments.formatters.pangomarkup', 'Pango Markup', ('pango', 'pangomarkup'), (), 'Format tokens as Pango Markup code. It can then be rendered to an SVG.'),
+    'RawTokenFormatter': ('pygments.formatters.other', 'Raw tokens', ('raw', 'tokens'), ('*.raw',), 'Format tokens as a raw representation for storing token streams.'),
+    'RtfFormatter': ('pygments.formatters.rtf', 'RTF', ('rtf',), ('*.rtf',), 'Format tokens as RTF markup. This formatter automatically outputs full RTF documents with color information and other useful stuff. Perfect for Copy and Paste into Microsoft(R) Word(R) documents.'),
+    'SvgFormatter': ('pygments.formatters.svg', 'SVG', ('svg',), ('*.svg',), 'Format tokens as an SVG graphics file.  This formatter is still experimental. Each line of code is a ``<text>`` element with explicit ``x`` and ``y`` coordinates containing ``<tspan>`` elements with the individual token styles.'),
+    'Terminal256Formatter': ('pygments.formatters.terminal256', 'Terminal256', ('terminal256', 'console256', '256'), (), 'Format tokens with ANSI color sequences, for output in a 256-color terminal or console.  Like in `TerminalFormatter` color sequences are terminated at newlines, so that paging the output works correctly.'),
+    'TerminalFormatter': ('pygments.formatters.terminal', 'Terminal', ('terminal', 'console'), (), 'Format tokens with ANSI color sequences, for output in a text console. Color sequences are terminated at newlines, so that paging the output works correctly.'),
+    'TerminalTrueColorFormatter': ('pygments.formatters.terminal256', 'TerminalTrueColor', ('terminal16m', 'console16m', '16m'), (), 'Format tokens with ANSI color sequences, for output in a true-color terminal or console.  Like in `TerminalFormatter` color sequences are terminated at newlines, so that paging the output works correctly.'),
+    'TestcaseFormatter': ('pygments.formatters.other', 'Testcase', ('testcase',), (), 'Format tokens as appropriate for a new testcase.'),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/bbcode.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/bbcode.py
new file mode 100644
index 00000000..5a05bd96
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/bbcode.py
@@ -0,0 +1,108 @@
+"""
+    pygments.formatters.bbcode
+    ~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    BBcode formatter.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.util import get_bool_opt
+
+__all__ = ['BBCodeFormatter']
+
+
+class BBCodeFormatter(Formatter):
+    """
+    Format tokens with BBcodes. These formatting codes are used by many
+    bulletin boards, so you can highlight your sourcecode with pygments before
+    posting it there.
+
+    This formatter has no support for background colors and borders, as there
+    are no common BBcode tags for that.
+
+    Some board systems (e.g. phpBB) don't support colors in their [code] tag,
+    so you can't use the highlighting together with that tag.
+    Text in a [code] tag usually is shown with a monospace font (which this
+    formatter can do with the ``monofont`` option) and no spaces (which you
+    need for indentation) are removed.
+
+    Additional options accepted:
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+
+    `codetag`
+        If set to true, put the output into ``[code]`` tags (default:
+        ``false``)
+
+    `monofont`
+        If set to true, add a tag to show the code with a monospace font
+        (default: ``false``).
+    """
+    name = 'BBCode'
+    aliases = ['bbcode', 'bb']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self._code = get_bool_opt(options, 'codetag', False)
+        self._mono = get_bool_opt(options, 'monofont', False)
+
+        self.styles = {}
+        self._make_styles()
+
+    def _make_styles(self):
+        for ttype, ndef in self.style:
+            start = end = ''
+            if ndef['color']:
+                start += '[color=#{}]'.format(ndef['color'])
+                end = '[/color]' + end
+            if ndef['bold']:
+                start += '[b]'
+                end = '[/b]' + end
+            if ndef['italic']:
+                start += '[i]'
+                end = '[/i]' + end
+            if ndef['underline']:
+                start += '[u]'
+                end = '[/u]' + end
+            # there are no common BBcodes for background-color and border
+
+            self.styles[ttype] = start, end
+
+    def format_unencoded(self, tokensource, outfile):
+        if self._code:
+            outfile.write('[code]')
+        if self._mono:
+            outfile.write('[font=monospace]')
+
+        lastval = ''
+        lasttype = None
+
+        for ttype, value in tokensource:
+            while ttype not in self.styles:
+                ttype = ttype.parent
+            if ttype == lasttype:
+                lastval += value
+            else:
+                if lastval:
+                    start, end = self.styles[lasttype]
+                    outfile.write(''.join((start, lastval, end)))
+                lastval = value
+                lasttype = ttype
+
+        if lastval:
+            start, end = self.styles[lasttype]
+            outfile.write(''.join((start, lastval, end)))
+
+        if self._mono:
+            outfile.write('[/font]')
+        if self._code:
+            outfile.write('[/code]')
+        if self._code or self._mono:
+            outfile.write('\n')
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/groff.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/groff.py
new file mode 100644
index 00000000..5c8a958f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/groff.py
@@ -0,0 +1,170 @@
+"""
+    pygments.formatters.groff
+    ~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for groff output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import math
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt
+
+__all__ = ['GroffFormatter']
+
+
+class GroffFormatter(Formatter):
+    """
+    Format tokens with groff escapes to change their color and font style.
+
+    .. versionadded:: 2.11
+
+    Additional options accepted:
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+
+    `monospaced`
+        If set to true, monospace font will be used (default: ``true``).
+
+    `linenos`
+        If set to true, print the line numbers (default: ``false``).
+
+    `wrap`
+        Wrap lines to the specified number of characters. Disabled if set to 0
+        (default: ``0``).
+    """
+
+    name = 'groff'
+    aliases = ['groff','troff','roff']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+
+        self.monospaced = get_bool_opt(options, 'monospaced', True)
+        self.linenos = get_bool_opt(options, 'linenos', False)
+        self._lineno = 0
+        self.wrap = get_int_opt(options, 'wrap', 0)
+        self._linelen = 0
+
+        self.styles = {}
+        self._make_styles()
+
+
+    def _make_styles(self):
+        regular = '\\f[CR]' if self.monospaced else '\\f[R]'
+        bold = '\\f[CB]' if self.monospaced else '\\f[B]'
+        italic = '\\f[CI]' if self.monospaced else '\\f[I]'
+
+        for ttype, ndef in self.style:
+            start = end = ''
+            if ndef['color']:
+                start += '\\m[{}]'.format(ndef['color'])
+                end = '\\m[]' + end
+            if ndef['bold']:
+                start += bold
+                end = regular + end
+            if ndef['italic']:
+                start += italic
+                end = regular + end
+            if ndef['bgcolor']:
+                start += '\\M[{}]'.format(ndef['bgcolor'])
+                end = '\\M[]' + end
+
+            self.styles[ttype] = start, end
+
+
+    def _define_colors(self, outfile):
+        colors = set()
+        for _, ndef in self.style:
+            if ndef['color'] is not None:
+                colors.add(ndef['color'])
+
+        for color in sorted(colors):
+            outfile.write('.defcolor ' + color + ' rgb #' + color + '\n')
+
+
+    def _write_lineno(self, outfile):
+        self._lineno += 1
+        outfile.write("%s% 4d " % (self._lineno != 1 and '\n' or '', self._lineno))
+
+
+    def _wrap_line(self, line):
+        length = len(line.rstrip('\n'))
+        space = '     ' if self.linenos else ''
+        newline = ''
+
+        if length > self.wrap:
+            for i in range(0, math.floor(length / self.wrap)):
+                chunk = line[i*self.wrap:i*self.wrap+self.wrap]
+                newline += (chunk + '\n' + space)
+            remainder = length % self.wrap
+            if remainder > 0:
+                newline += line[-remainder-1:]
+                self._linelen = remainder
+        elif self._linelen + length > self.wrap:
+            newline = ('\n' + space) + line
+            self._linelen = length
+        else:
+            newline = line
+            self._linelen += length
+
+        return newline
+
+
+    def _escape_chars(self, text):
+        text = text.replace('\\', '\\[u005C]'). \
+                    replace('.', '\\[char46]'). \
+                    replace('\'', '\\[u0027]'). \
+                    replace('`', '\\[u0060]'). \
+                    replace('~', '\\[u007E]')
+        copy = text
+
+        for char in copy:
+            if len(char) != len(char.encode()):
+                uni = char.encode('unicode_escape') \
+                    .decode()[1:] \
+                    .replace('x', 'u00') \
+                    .upper()
+                text = text.replace(char, '\\[u' + uni[1:] + ']')
+
+        return text
+
+
+    def format_unencoded(self, tokensource, outfile):
+        self._define_colors(outfile)
+
+        outfile.write('.nf\n\\f[CR]\n')
+
+        if self.linenos:
+            self._write_lineno(outfile)
+
+        for ttype, value in tokensource:
+            while ttype not in self.styles:
+                ttype = ttype.parent
+            start, end = self.styles[ttype]
+
+            for line in value.splitlines(True):
+                if self.wrap > 0:
+                    line = self._wrap_line(line)
+
+                if start and end:
+                    text = self._escape_chars(line.rstrip('\n'))
+                    if text != '':
+                        outfile.write(''.join((start, text, end)))
+                else:
+                    outfile.write(self._escape_chars(line.rstrip('\n')))
+
+                if line.endswith('\n'):
+                    if self.linenos:
+                        self._write_lineno(outfile)
+                        self._linelen = 0
+                    else:
+                        outfile.write('\n')
+                        self._linelen = 0
+
+        outfile.write('\n.fi')
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/html.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/html.py
new file mode 100644
index 00000000..7aa938f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/html.py
@@ -0,0 +1,987 @@
+"""
+    pygments.formatters.html
+    ~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for HTML output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import functools
+import os
+import sys
+import os.path
+from io import StringIO
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.token import Token, Text, STANDARD_TYPES
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt, get_list_opt
+
+try:
+    import ctags
+except ImportError:
+    ctags = None
+
+__all__ = ['HtmlFormatter']
+
+
+_escape_html_table = {
+    ord('&'): '&amp;',
+    ord('<'): '&lt;',
+    ord('>'): '&gt;',
+    ord('"'): '&quot;',
+    ord("'"): '&#39;',
+}
+
+
+def escape_html(text, table=_escape_html_table):
+    """Escape &, <, > as well as single and double quotes for HTML."""
+    return text.translate(table)
+
+
+def webify(color):
+    if color.startswith('calc') or color.startswith('var'):
+        return color
+    else:
+        return '#' + color
+
+
+def _get_ttype_class(ttype):
+    fname = STANDARD_TYPES.get(ttype)
+    if fname:
+        return fname
+    aname = ''
+    while fname is None:
+        aname = '-' + ttype[-1] + aname
+        ttype = ttype.parent
+        fname = STANDARD_TYPES.get(ttype)
+    return fname + aname
+
+
+CSSFILE_TEMPLATE = '''\
+/*
+generated by Pygments <https://pygments.org/>
+Copyright 2006-2024 by the Pygments team.
+Licensed under the BSD license, see LICENSE for details.
+*/
+%(styledefs)s
+'''
+
+DOC_HEADER = '''\
+<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"
+   "http://www.w3.org/TR/html4/strict.dtd">
+<!--
+generated by Pygments <https://pygments.org/>
+Copyright 2006-2024 by the Pygments team.
+Licensed under the BSD license, see LICENSE for details.
+-->
+<html>
+<head>
+  <title>%(title)s</title>
+  <meta http-equiv="content-type" content="text/html; charset=%(encoding)s">
+  <style type="text/css">
+''' + CSSFILE_TEMPLATE + '''
+  </style>
+</head>
+<body>
+<h2>%(title)s</h2>
+
+'''
+
+DOC_HEADER_EXTERNALCSS = '''\
+<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"
+   "http://www.w3.org/TR/html4/strict.dtd">
+
+<html>
+<head>
+  <title>%(title)s</title>
+  <meta http-equiv="content-type" content="text/html; charset=%(encoding)s">
+  <link rel="stylesheet" href="%(cssfile)s" type="text/css">
+</head>
+<body>
+<h2>%(title)s</h2>
+
+'''
+
+DOC_FOOTER = '''\
+</body>
+</html>
+'''
+
+
+class HtmlFormatter(Formatter):
+    r"""
+    Format tokens as HTML 4 ``<span>`` tags. By default, the content is enclosed
+    in a ``<pre>`` tag, itself wrapped in a ``<div>`` tag (but see the `nowrap` option).
+    The ``<div>``'s CSS class can be set by the `cssclass` option.
+
+    If the `linenos` option is set to ``"table"``, the ``<pre>`` is
+    additionally wrapped inside a ``<table>`` which has one row and two
+    cells: one containing the line numbers and one containing the code.
+    Example:
+
+    .. sourcecode:: html
+
+        <div class="highlight" >
+        <table><tr>
+          <td class="linenos" title="click to toggle"
+            onclick="with (this.firstChild.style)
+                     { display = (display == '') ? 'none' : '' }">
+            <pre>1
+            2</pre>
+          </td>
+          <td class="code">
+            <pre><span class="Ke">def </span><span class="NaFu">foo</span>(bar):
+              <span class="Ke">pass</span>
+            </pre>
+          </td>
+        </tr></table></div>
+
+    (whitespace added to improve clarity).
+
+    A list of lines can be specified using the `hl_lines` option to make these
+    lines highlighted (as of Pygments 0.11).
+
+    With the `full` option, a complete HTML 4 document is output, including
+    the style definitions inside a ``<style>`` tag, or in a separate file if
+    the `cssfile` option is given.
+
+    When `tagsfile` is set to the path of a ctags index file, it is used to
+    generate hyperlinks from names to their definition.  You must enable
+    `lineanchors` and run ctags with the `-n` option for this to work.  The
+    `python-ctags` module from PyPI must be installed to use this feature;
+    otherwise a `RuntimeError` will be raised.
+
+    The `get_style_defs(arg='')` method of a `HtmlFormatter` returns a string
+    containing CSS rules for the CSS classes used by the formatter. The
+    argument `arg` can be used to specify additional CSS selectors that
+    are prepended to the classes. A call `fmter.get_style_defs('td .code')`
+    would result in the following CSS classes:
+
+    .. sourcecode:: css
+
+        td .code .kw { font-weight: bold; color: #00FF00 }
+        td .code .cm { color: #999999 }
+        ...
+
+    If you have Pygments 0.6 or higher, you can also pass a list or tuple to the
+    `get_style_defs()` method to request multiple prefixes for the tokens:
+
+    .. sourcecode:: python
+
+        formatter.get_style_defs(['div.syntax pre', 'pre.syntax'])
+
+    The output would then look like this:
+
+    .. sourcecode:: css
+
+        div.syntax pre .kw,
+        pre.syntax .kw { font-weight: bold; color: #00FF00 }
+        div.syntax pre .cm,
+        pre.syntax .cm { color: #999999 }
+        ...
+
+    Additional options accepted:
+
+    `nowrap`
+        If set to ``True``, don't add a ``<pre>`` and a ``<div>`` tag
+        around the tokens. This disables most other options (default: ``False``).
+
+    `full`
+        Tells the formatter to output a "full" document, i.e. a complete
+        self-contained document (default: ``False``).
+
+    `title`
+        If `full` is true, the title that should be used to caption the
+        document (default: ``''``).
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``). This option has no effect if the `cssfile`
+        and `noclobber_cssfile` option are given and the file specified in
+        `cssfile` exists.
+
+    `noclasses`
+        If set to true, token ``<span>`` tags (as well as line number elements)
+        will not use CSS classes, but inline styles. This is not recommended
+        for larger pieces of code since it increases output size by quite a bit
+        (default: ``False``).
+
+    `classprefix`
+        Since the token types use relatively short class names, they may clash
+        with some of your own class names. In this case you can use the
+        `classprefix` option to give a string to prepend to all Pygments-generated
+        CSS class names for token types.
+        Note that this option also affects the output of `get_style_defs()`.
+
+    `cssclass`
+        CSS class for the wrapping ``<div>`` tag (default: ``'highlight'``).
+        If you set this option, the default selector for `get_style_defs()`
+        will be this class.
+
+        .. versionadded:: 0.9
+           If you select the ``'table'`` line numbers, the wrapping table will
+           have a CSS class of this string plus ``'table'``, the default is
+           accordingly ``'highlighttable'``.
+
+    `cssstyles`
+        Inline CSS styles for the wrapping ``<div>`` tag (default: ``''``).
+
+    `prestyles`
+        Inline CSS styles for the ``<pre>`` tag (default: ``''``).
+
+        .. versionadded:: 0.11
+
+    `cssfile`
+        If the `full` option is true and this option is given, it must be the
+        name of an external file. If the filename does not include an absolute
+        path, the file's path will be assumed to be relative to the main output
+        file's path, if the latter can be found. The stylesheet is then written
+        to this file instead of the HTML file.
+
+        .. versionadded:: 0.6
+
+    `noclobber_cssfile`
+        If `cssfile` is given and the specified file exists, the css file will
+        not be overwritten. This allows the use of the `full` option in
+        combination with a user specified css file. Default is ``False``.
+
+        .. versionadded:: 1.1
+
+    `linenos`
+        If set to ``'table'``, output line numbers as a table with two cells,
+        one containing the line numbers, the other the whole code.  This is
+        copy-and-paste-friendly, but may cause alignment problems with some
+        browsers or fonts.  If set to ``'inline'``, the line numbers will be
+        integrated in the ``<pre>`` tag that contains the code (that setting
+        is *new in Pygments 0.8*).
+
+        For compatibility with Pygments 0.7 and earlier, every true value
+        except ``'inline'`` means the same as ``'table'`` (in particular, that
+        means also ``True``).
+
+        The default value is ``False``, which means no line numbers at all.
+
+        **Note:** with the default ("table") line number mechanism, the line
+        numbers and code can have different line heights in Internet Explorer
+        unless you give the enclosing ``<pre>`` tags an explicit ``line-height``
+        CSS property (you get the default line spacing with ``line-height:
+        125%``).
+
+    `hl_lines`
+        Specify a list of lines to be highlighted. The line numbers are always
+        relative to the input (i.e. the first line is line 1) and are
+        independent of `linenostart`.
+
+        .. versionadded:: 0.11
+
+    `linenostart`
+        The line number for the first line (default: ``1``).
+
+    `linenostep`
+        If set to a number n > 1, only every nth line number is printed.
+
+    `linenospecial`
+        If set to a number n > 0, every nth line number is given the CSS
+        class ``"special"`` (default: ``0``).
+
+    `nobackground`
+        If set to ``True``, the formatter won't output the background color
+        for the wrapping element (this automatically defaults to ``False``
+        when there is no wrapping element [eg: no argument for the
+        `get_syntax_defs` method given]) (default: ``False``).
+
+        .. versionadded:: 0.6
+
+    `lineseparator`
+        This string is output between lines of code. It defaults to ``"\n"``,
+        which is enough to break a line inside ``<pre>`` tags, but you can
+        e.g. set it to ``"<br>"`` to get HTML line breaks.
+
+        .. versionadded:: 0.7
+
+    `lineanchors`
+        If set to a nonempty string, e.g. ``foo``, the formatter will wrap each
+        output line in an anchor tag with an ``id`` (and `name`) of ``foo-linenumber``.
+        This allows easy linking to certain lines.
+
+        .. versionadded:: 0.9
+
+    `linespans`
+        If set to a nonempty string, e.g. ``foo``, the formatter will wrap each
+        output line in a span tag with an ``id`` of ``foo-linenumber``.
+        This allows easy access to lines via javascript.
+
+        .. versionadded:: 1.6
+
+    `anchorlinenos`
+        If set to `True`, will wrap line numbers in <a> tags. Used in
+        combination with `linenos` and `lineanchors`.
+
+    `tagsfile`
+        If set to the path of a ctags file, wrap names in anchor tags that
+        link to their definitions. `lineanchors` should be used, and the
+        tags file should specify line numbers (see the `-n` option to ctags).
+        The tags file is assumed to be encoded in UTF-8.
+
+        .. versionadded:: 1.6
+
+    `tagurlformat`
+        A string formatting pattern used to generate links to ctags definitions.
+        Available variables are `%(path)s`, `%(fname)s` and `%(fext)s`.
+        Defaults to an empty string, resulting in just `#prefix-number` links.
+
+        .. versionadded:: 1.6
+
+    `filename`
+        A string used to generate a filename when rendering ``<pre>`` blocks,
+        for example if displaying source code. If `linenos` is set to
+        ``'table'`` then the filename will be rendered in an initial row
+        containing a single `<th>` which spans both columns.
+
+        .. versionadded:: 2.1
+
+    `wrapcode`
+        Wrap the code inside ``<pre>`` blocks using ``<code>``, as recommended
+        by the HTML5 specification.
+
+        .. versionadded:: 2.4
+
+    `debug_token_types`
+        Add ``title`` attributes to all token ``<span>`` tags that show the
+        name of the token.
+
+        .. versionadded:: 2.10
+
+
+    **Subclassing the HTML formatter**
+
+    .. versionadded:: 0.7
+
+    The HTML formatter is now built in a way that allows easy subclassing, thus
+    customizing the output HTML code. The `format()` method calls
+    `self._format_lines()` which returns a generator that yields tuples of ``(1,
+    line)``, where the ``1`` indicates that the ``line`` is a line of the
+    formatted source code.
+
+    If the `nowrap` option is set, the generator is the iterated over and the
+    resulting HTML is output.
+
+    Otherwise, `format()` calls `self.wrap()`, which wraps the generator with
+    other generators. These may add some HTML code to the one generated by
+    `_format_lines()`, either by modifying the lines generated by the latter,
+    then yielding them again with ``(1, line)``, and/or by yielding other HTML
+    code before or after the lines, with ``(0, html)``. The distinction between
+    source lines and other code makes it possible to wrap the generator multiple
+    times.
+
+    The default `wrap()` implementation adds a ``<div>`` and a ``<pre>`` tag.
+
+    A custom `HtmlFormatter` subclass could look like this:
+
+    .. sourcecode:: python
+
+        class CodeHtmlFormatter(HtmlFormatter):
+
+            def wrap(self, source, *, include_div):
+                return self._wrap_code(source)
+
+            def _wrap_code(self, source):
+                yield 0, '<code>'
+                for i, t in source:
+                    if i == 1:
+                        # it's a line of formatted code
+                        t += '<br>'
+                    yield i, t
+                yield 0, '</code>'
+
+    This results in wrapping the formatted lines with a ``<code>`` tag, where the
+    source lines are broken using ``<br>`` tags.
+
+    After calling `wrap()`, the `format()` method also adds the "line numbers"
+    and/or "full document" wrappers if the respective options are set. Then, all
+    HTML yielded by the wrapped generator is output.
+    """
+
+    name = 'HTML'
+    aliases = ['html']
+    filenames = ['*.html', '*.htm']
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self.title = self._decodeifneeded(self.title)
+        self.nowrap = get_bool_opt(options, 'nowrap', False)
+        self.noclasses = get_bool_opt(options, 'noclasses', False)
+        self.classprefix = options.get('classprefix', '')
+        self.cssclass = self._decodeifneeded(options.get('cssclass', 'highlight'))
+        self.cssstyles = self._decodeifneeded(options.get('cssstyles', ''))
+        self.prestyles = self._decodeifneeded(options.get('prestyles', ''))
+        self.cssfile = self._decodeifneeded(options.get('cssfile', ''))
+        self.noclobber_cssfile = get_bool_opt(options, 'noclobber_cssfile', False)
+        self.tagsfile = self._decodeifneeded(options.get('tagsfile', ''))
+        self.tagurlformat = self._decodeifneeded(options.get('tagurlformat', ''))
+        self.filename = self._decodeifneeded(options.get('filename', ''))
+        self.wrapcode = get_bool_opt(options, 'wrapcode', False)
+        self.span_element_openers = {}
+        self.debug_token_types = get_bool_opt(options, 'debug_token_types', False)
+
+        if self.tagsfile:
+            if not ctags:
+                raise RuntimeError('The "ctags" package must to be installed '
+                                   'to be able to use the "tagsfile" feature.')
+            self._ctags = ctags.CTags(self.tagsfile)
+
+        linenos = options.get('linenos', False)
+        if linenos == 'inline':
+            self.linenos = 2
+        elif linenos:
+            # compatibility with <= 0.7
+            self.linenos = 1
+        else:
+            self.linenos = 0
+        self.linenostart = abs(get_int_opt(options, 'linenostart', 1))
+        self.linenostep = abs(get_int_opt(options, 'linenostep', 1))
+        self.linenospecial = abs(get_int_opt(options, 'linenospecial', 0))
+        self.nobackground = get_bool_opt(options, 'nobackground', False)
+        self.lineseparator = options.get('lineseparator', '\n')
+        self.lineanchors = options.get('lineanchors', '')
+        self.linespans = options.get('linespans', '')
+        self.anchorlinenos = get_bool_opt(options, 'anchorlinenos', False)
+        self.hl_lines = set()
+        for lineno in get_list_opt(options, 'hl_lines', []):
+            try:
+                self.hl_lines.add(int(lineno))
+            except ValueError:
+                pass
+
+        self._create_stylesheet()
+
+    def _get_css_class(self, ttype):
+        """Return the css class of this token type prefixed with
+        the classprefix option."""
+        ttypeclass = _get_ttype_class(ttype)
+        if ttypeclass:
+            return self.classprefix + ttypeclass
+        return ''
+
+    def _get_css_classes(self, ttype):
+        """Return the CSS classes of this token type prefixed with the classprefix option."""
+        cls = self._get_css_class(ttype)
+        while ttype not in STANDARD_TYPES:
+            ttype = ttype.parent
+            cls = self._get_css_class(ttype) + ' ' + cls
+        return cls or ''
+
+    def _get_css_inline_styles(self, ttype):
+        """Return the inline CSS styles for this token type."""
+        cclass = self.ttype2class.get(ttype)
+        while cclass is None:
+            ttype = ttype.parent
+            cclass = self.ttype2class.get(ttype)
+        return cclass or ''
+
+    def _create_stylesheet(self):
+        t2c = self.ttype2class = {Token: ''}
+        c2s = self.class2style = {}
+        for ttype, ndef in self.style:
+            name = self._get_css_class(ttype)
+            style = ''
+            if ndef['color']:
+                style += 'color: {}; '.format(webify(ndef['color']))
+            if ndef['bold']:
+                style += 'font-weight: bold; '
+            if ndef['italic']:
+                style += 'font-style: italic; '
+            if ndef['underline']:
+                style += 'text-decoration: underline; '
+            if ndef['bgcolor']:
+                style += 'background-color: {}; '.format(webify(ndef['bgcolor']))
+            if ndef['border']:
+                style += 'border: 1px solid {}; '.format(webify(ndef['border']))
+            if style:
+                t2c[ttype] = name
+                # save len(ttype) to enable ordering the styles by
+                # hierarchy (necessary for CSS cascading rules!)
+                c2s[name] = (style[:-2], ttype, len(ttype))
+
+    def get_style_defs(self, arg=None):
+        """
+        Return CSS style definitions for the classes produced by the current
+        highlighting style. ``arg`` can be a string or list of selectors to
+        insert before the token type classes.
+        """
+        style_lines = []
+
+        style_lines.extend(self.get_linenos_style_defs())
+        style_lines.extend(self.get_background_style_defs(arg))
+        style_lines.extend(self.get_token_style_defs(arg))
+
+        return '\n'.join(style_lines)
+
+    def get_token_style_defs(self, arg=None):
+        prefix = self.get_css_prefix(arg)
+
+        styles = [
+            (level, ttype, cls, style)
+            for cls, (style, ttype, level) in self.class2style.items()
+            if cls and style
+        ]
+        styles.sort()
+
+        lines = [
+            f'{prefix(cls)} {{ {style} }} /* {repr(ttype)[6:]} */'
+            for (level, ttype, cls, style) in styles
+        ]
+
+        return lines
+
+    def get_background_style_defs(self, arg=None):
+        prefix = self.get_css_prefix(arg)
+        bg_color = self.style.background_color
+        hl_color = self.style.highlight_color
+
+        lines = []
+
+        if arg and not self.nobackground and bg_color is not None:
+            text_style = ''
+            if Text in self.ttype2class:
+                text_style = ' ' + self.class2style[self.ttype2class[Text]][0]
+            lines.insert(
+                0, '{}{{ background: {};{} }}'.format(
+                    prefix(''), bg_color, text_style
+                )
+            )
+        if hl_color is not None:
+            lines.insert(
+                0, '{} {{ background-color: {} }}'.format(prefix('hll'), hl_color)
+            )
+
+        return lines
+
+    def get_linenos_style_defs(self):
+        lines = [
+            f'pre {{ {self._pre_style} }}',
+            f'td.linenos .normal {{ {self._linenos_style} }}',
+            f'span.linenos {{ {self._linenos_style} }}',
+            f'td.linenos .special {{ {self._linenos_special_style} }}',
+            f'span.linenos.special {{ {self._linenos_special_style} }}',
+        ]
+
+        return lines
+
+    def get_css_prefix(self, arg):
+        if arg is None:
+            arg = ('cssclass' in self.options and '.'+self.cssclass or '')
+        if isinstance(arg, str):
+            args = [arg]
+        else:
+            args = list(arg)
+
+        def prefix(cls):
+            if cls:
+                cls = '.' + cls
+            tmp = []
+            for arg in args:
+                tmp.append((arg and arg + ' ' or '') + cls)
+            return ', '.join(tmp)
+
+        return prefix
+
+    @property
+    def _pre_style(self):
+        return 'line-height: 125%;'
+
+    @property
+    def _linenos_style(self):
+        color = self.style.line_number_color
+        background_color = self.style.line_number_background_color
+        return f'color: {color}; background-color: {background_color}; padding-left: 5px; padding-right: 5px;'
+
+    @property
+    def _linenos_special_style(self):
+        color = self.style.line_number_special_color
+        background_color = self.style.line_number_special_background_color
+        return f'color: {color}; background-color: {background_color}; padding-left: 5px; padding-right: 5px;'
+
+    def _decodeifneeded(self, value):
+        if isinstance(value, bytes):
+            if self.encoding:
+                return value.decode(self.encoding)
+            return value.decode()
+        return value
+
+    def _wrap_full(self, inner, outfile):
+        if self.cssfile:
+            if os.path.isabs(self.cssfile):
+                # it's an absolute filename
+                cssfilename = self.cssfile
+            else:
+                try:
+                    filename = outfile.name
+                    if not filename or filename[0] == '<':
+                        # pseudo files, e.g. name == '<fdopen>'
+                        raise AttributeError
+                    cssfilename = os.path.join(os.path.dirname(filename),
+                                               self.cssfile)
+                except AttributeError:
+                    print('Note: Cannot determine output file name, '
+                          'using current directory as base for the CSS file name',
+                          file=sys.stderr)
+                    cssfilename = self.cssfile
+            # write CSS file only if noclobber_cssfile isn't given as an option.
+            try:
+                if not os.path.exists(cssfilename) or not self.noclobber_cssfile:
+                    with open(cssfilename, "w", encoding="utf-8") as cf:
+                        cf.write(CSSFILE_TEMPLATE %
+                                 {'styledefs': self.get_style_defs('body')})
+            except OSError as err:
+                err.strerror = 'Error writing CSS file: ' + err.strerror
+                raise
+
+            yield 0, (DOC_HEADER_EXTERNALCSS %
+                      dict(title=self.title,
+                           cssfile=self.cssfile,
+                           encoding=self.encoding))
+        else:
+            yield 0, (DOC_HEADER %
+                      dict(title=self.title,
+                           styledefs=self.get_style_defs('body'),
+                           encoding=self.encoding))
+
+        yield from inner
+        yield 0, DOC_FOOTER
+
+    def _wrap_tablelinenos(self, inner):
+        dummyoutfile = StringIO()
+        lncount = 0
+        for t, line in inner:
+            if t:
+                lncount += 1
+            dummyoutfile.write(line)
+
+        fl = self.linenostart
+        mw = len(str(lncount + fl - 1))
+        sp = self.linenospecial
+        st = self.linenostep
+        anchor_name = self.lineanchors or self.linespans
+        aln = self.anchorlinenos
+        nocls = self.noclasses
+
+        lines = []
+
+        for i in range(fl, fl+lncount):
+            print_line = i % st == 0
+            special_line = sp and i % sp == 0
+
+            if print_line:
+                line = '%*d' % (mw, i)
+                if aln:
+                    line = '<a href="#%s-%d">%s</a>' % (anchor_name, i, line)
+            else:
+                line = ' ' * mw
+
+            if nocls:
+                if special_line:
+                    style = f' style="{self._linenos_special_style}"'
+                else:
+                    style = f' style="{self._linenos_style}"'
+            else:
+                if special_line:
+                    style = ' class="special"'
+                else:
+                    style = ' class="normal"'
+
+            if style:
+                line = f'<span{style}>{line}</span>'
+
+            lines.append(line)
+
+        ls = '\n'.join(lines)
+
+        # If a filename was specified, we can't put it into the code table as it
+        # would misalign the line numbers. Hence we emit a separate row for it.
+        filename_tr = ""
+        if self.filename:
+            filename_tr = (
+                '<tr><th colspan="2" class="filename">'
+                '<span class="filename">' + self.filename + '</span>'
+                '</th></tr>')
+
+        # in case you wonder about the seemingly redundant <div> here: since the
+        # content in the other cell also is wrapped in a div, some browsers in
+        # some configurations seem to mess up the formatting...
+        yield 0, (f'<table class="{self.cssclass}table">' + filename_tr +
+            '<tr><td class="linenos"><div class="linenodiv"><pre>' +
+            ls + '</pre></div></td><td class="code">')
+        yield 0, '<div>'
+        yield 0, dummyoutfile.getvalue()
+        yield 0, '</div>'
+        yield 0, '</td></tr></table>'
+
+
+    def _wrap_inlinelinenos(self, inner):
+        # need a list of lines since we need the width of a single number :(
+        inner_lines = list(inner)
+        sp = self.linenospecial
+        st = self.linenostep
+        num = self.linenostart
+        mw = len(str(len(inner_lines) + num - 1))
+        anchor_name = self.lineanchors or self.linespans
+        aln = self.anchorlinenos
+        nocls = self.noclasses
+
+        for _, inner_line in inner_lines:
+            print_line = num % st == 0
+            special_line = sp and num % sp == 0
+
+            if print_line:
+                line = '%*d' % (mw, num)
+            else:
+                line = ' ' * mw
+
+            if nocls:
+                if special_line:
+                    style = f' style="{self._linenos_special_style}"'
+                else:
+                    style = f' style="{self._linenos_style}"'
+            else:
+                if special_line:
+                    style = ' class="linenos special"'
+                else:
+                    style = ' class="linenos"'
+
+            if style:
+                linenos = f'<span{style}>{line}</span>'
+            else:
+                linenos = line
+
+            if aln:
+                yield 1, ('<a href="#%s-%d">%s</a>' % (anchor_name, num, linenos) +
+                          inner_line)
+            else:
+                yield 1, linenos + inner_line
+            num += 1
+
+    def _wrap_lineanchors(self, inner):
+        s = self.lineanchors
+        # subtract 1 since we have to increment i *before* yielding
+        i = self.linenostart - 1
+        for t, line in inner:
+            if t:
+                i += 1
+                href = "" if self.linenos else ' href="#%s-%d"' % (s, i)
+                yield 1, '<a id="%s-%d" name="%s-%d"%s></a>' % (s, i, s, i, href) + line
+            else:
+                yield 0, line
+
+    def _wrap_linespans(self, inner):
+        s = self.linespans
+        i = self.linenostart - 1
+        for t, line in inner:
+            if t:
+                i += 1
+                yield 1, '<span id="%s-%d">%s</span>' % (s, i, line)
+            else:
+                yield 0, line
+
+    def _wrap_div(self, inner):
+        style = []
+        if (self.noclasses and not self.nobackground and
+                self.style.background_color is not None):
+            style.append(f'background: {self.style.background_color}')
+        if self.cssstyles:
+            style.append(self.cssstyles)
+        style = '; '.join(style)
+
+        yield 0, ('<div' + (self.cssclass and f' class="{self.cssclass}"') +
+                  (style and (f' style="{style}"')) + '>')
+        yield from inner
+        yield 0, '</div>\n'
+
+    def _wrap_pre(self, inner):
+        style = []
+        if self.prestyles:
+            style.append(self.prestyles)
+        if self.noclasses:
+            style.append(self._pre_style)
+        style = '; '.join(style)
+
+        if self.filename and self.linenos != 1:
+            yield 0, ('<span class="filename">' + self.filename + '</span>')
+
+        # the empty span here is to keep leading empty lines from being
+        # ignored by HTML parsers
+        yield 0, ('<pre' + (style and f' style="{style}"') + '><span></span>')
+        yield from inner
+        yield 0, '</pre>'
+
+    def _wrap_code(self, inner):
+        yield 0, '<code>'
+        yield from inner
+        yield 0, '</code>'
+
+    @functools.lru_cache(maxsize=100)
+    def _translate_parts(self, value):
+        """HTML-escape a value and split it by newlines."""
+        return value.translate(_escape_html_table).split('\n')
+
+    def _format_lines(self, tokensource):
+        """
+        Just format the tokens, without any wrapping tags.
+        Yield individual lines.
+        """
+        nocls = self.noclasses
+        lsep = self.lineseparator
+        tagsfile = self.tagsfile
+
+        lspan = ''
+        line = []
+        for ttype, value in tokensource:
+            try:
+                cspan = self.span_element_openers[ttype]
+            except KeyError:
+                title = ' title="{}"'.format('.'.join(ttype)) if self.debug_token_types else ''
+                if nocls:
+                    css_style = self._get_css_inline_styles(ttype)
+                    if css_style:
+                        css_style = self.class2style[css_style][0]
+                        cspan = f'<span style="{css_style}"{title}>'
+                    else:
+                        cspan = ''
+                else:
+                    css_class = self._get_css_classes(ttype)
+                    if css_class:
+                        cspan = f'<span class="{css_class}"{title}>'
+                    else:
+                        cspan = ''
+                self.span_element_openers[ttype] = cspan
+
+            parts = self._translate_parts(value)
+
+            if tagsfile and ttype in Token.Name:
+                filename, linenumber = self._lookup_ctag(value)
+                if linenumber:
+                    base, filename = os.path.split(filename)
+                    if base:
+                        base += '/'
+                    filename, extension = os.path.splitext(filename)
+                    url = self.tagurlformat % {'path': base, 'fname': filename,
+                                               'fext': extension}
+                    parts[0] = "<a href=\"%s#%s-%d\">%s" % \
+                        (url, self.lineanchors, linenumber, parts[0])
+                    parts[-1] = parts[-1] + "</a>"
+
+            # for all but the last line
+            for part in parts[:-1]:
+                if line:
+                    # Also check for part being non-empty, so we avoid creating
+                    # empty <span> tags
+                    if lspan != cspan and part:
+                        line.extend(((lspan and '</span>'), cspan, part,
+                                     (cspan and '</span>'), lsep))
+                    else:  # both are the same, or the current part was empty
+                        line.extend((part, (lspan and '</span>'), lsep))
+                    yield 1, ''.join(line)
+                    line = []
+                elif part:
+                    yield 1, ''.join((cspan, part, (cspan and '</span>'), lsep))
+                else:
+                    yield 1, lsep
+            # for the last line
+            if line and parts[-1]:
+                if lspan != cspan:
+                    line.extend(((lspan and '</span>'), cspan, parts[-1]))
+                    lspan = cspan
+                else:
+                    line.append(parts[-1])
+            elif parts[-1]:
+                line = [cspan, parts[-1]]
+                lspan = cspan
+            # else we neither have to open a new span nor set lspan
+
+        if line:
+            line.extend(((lspan and '</span>'), lsep))
+            yield 1, ''.join(line)
+
+    def _lookup_ctag(self, token):
+        entry = ctags.TagEntry()
+        if self._ctags.find(entry, token.encode(), 0):
+            return entry['file'].decode(), entry['lineNumber']
+        else:
+            return None, None
+
+    def _highlight_lines(self, tokensource):
+        """
+        Highlighted the lines specified in the `hl_lines` option by
+        post-processing the token stream coming from `_format_lines`.
+        """
+        hls = self.hl_lines
+
+        for i, (t, value) in enumerate(tokensource):
+            if t != 1:
+                yield t, value
+            if i + 1 in hls:  # i + 1 because Python indexes start at 0
+                if self.noclasses:
+                    style = ''
+                    if self.style.highlight_color is not None:
+                        style = (f' style="background-color: {self.style.highlight_color}"')
+                    yield 1, f'<span{style}>{value}</span>'
+                else:
+                    yield 1, f'<span class="hll">{value}</span>'
+            else:
+                yield 1, value
+
+    def wrap(self, source):
+        """
+        Wrap the ``source``, which is a generator yielding
+        individual lines, in custom generators. See docstring
+        for `format`. Can be overridden.
+        """
+
+        output = source
+        if self.wrapcode:
+            output = self._wrap_code(output)
+
+        output = self._wrap_pre(output)
+
+        return output
+
+    def format_unencoded(self, tokensource, outfile):
+        """
+        The formatting process uses several nested generators; which of
+        them are used is determined by the user's options.
+
+        Each generator should take at least one argument, ``inner``,
+        and wrap the pieces of text generated by this.
+
+        Always yield 2-tuples: (code, text). If "code" is 1, the text
+        is part of the original tokensource being highlighted, if it's
+        0, the text is some piece of wrapping. This makes it possible to
+        use several different wrappers that process the original source
+        linewise, e.g. line number generators.
+        """
+        source = self._format_lines(tokensource)
+
+        # As a special case, we wrap line numbers before line highlighting
+        # so the line numbers get wrapped in the highlighting tag.
+        if not self.nowrap and self.linenos == 2:
+            source = self._wrap_inlinelinenos(source)
+
+        if self.hl_lines:
+            source = self._highlight_lines(source)
+
+        if not self.nowrap:
+            if self.lineanchors:
+                source = self._wrap_lineanchors(source)
+            if self.linespans:
+                source = self._wrap_linespans(source)
+            source = self.wrap(source)
+            if self.linenos == 1:
+                source = self._wrap_tablelinenos(source)
+            source = self._wrap_div(source)
+            if self.full:
+                source = self._wrap_full(source, outfile)
+
+        for t, piece in source:
+            outfile.write(piece)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/img.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/img.py
new file mode 100644
index 00000000..7542cfad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/img.py
@@ -0,0 +1,685 @@
+"""
+    pygments.formatters.img
+    ~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for Pixmap output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+import os
+import sys
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt, get_list_opt, \
+    get_choice_opt
+
+import subprocess
+
+# Import this carefully
+try:
+    from PIL import Image, ImageDraw, ImageFont
+    pil_available = True
+except ImportError:
+    pil_available = False
+
+try:
+    import _winreg
+except ImportError:
+    try:
+        import winreg as _winreg
+    except ImportError:
+        _winreg = None
+
+__all__ = ['ImageFormatter', 'GifImageFormatter', 'JpgImageFormatter',
+           'BmpImageFormatter']
+
+
+# For some unknown reason every font calls it something different
+STYLES = {
+    'NORMAL':     ['', 'Roman', 'Book', 'Normal', 'Regular', 'Medium'],
+    'ITALIC':     ['Oblique', 'Italic'],
+    'BOLD':       ['Bold'],
+    'BOLDITALIC': ['Bold Oblique', 'Bold Italic'],
+}
+
+# A sane default for modern systems
+DEFAULT_FONT_NAME_NIX = 'DejaVu Sans Mono'
+DEFAULT_FONT_NAME_WIN = 'Courier New'
+DEFAULT_FONT_NAME_MAC = 'Menlo'
+
+
+class PilNotAvailable(ImportError):
+    """When Python imaging library is not available"""
+
+
+class FontNotFound(Exception):
+    """When there are no usable fonts specified"""
+
+
+class FontManager:
+    """
+    Manages a set of fonts: normal, italic, bold, etc...
+    """
+
+    def __init__(self, font_name, font_size=14):
+        self.font_name = font_name
+        self.font_size = font_size
+        self.fonts = {}
+        self.encoding = None
+        self.variable = False
+        if hasattr(font_name, 'read') or os.path.isfile(font_name):
+            font = ImageFont.truetype(font_name, self.font_size)
+            self.variable = True
+            for style in STYLES:
+                self.fonts[style] = font
+
+            return
+
+        if sys.platform.startswith('win'):
+            if not font_name:
+                self.font_name = DEFAULT_FONT_NAME_WIN
+            self._create_win()
+        elif sys.platform.startswith('darwin'):
+            if not font_name:
+                self.font_name = DEFAULT_FONT_NAME_MAC
+            self._create_mac()
+        else:
+            if not font_name:
+                self.font_name = DEFAULT_FONT_NAME_NIX
+            self._create_nix()
+
+    def _get_nix_font_path(self, name, style):
+        proc = subprocess.Popen(['fc-list', f"{name}:style={style}", 'file'],
+                                stdout=subprocess.PIPE, stderr=None)
+        stdout, _ = proc.communicate()
+        if proc.returncode == 0:
+            lines = stdout.splitlines()
+            for line in lines:
+                if line.startswith(b'Fontconfig warning:'):
+                    continue
+                path = line.decode().strip().strip(':')
+                if path:
+                    return path
+            return None
+
+    def _create_nix(self):
+        for name in STYLES['NORMAL']:
+            path = self._get_nix_font_path(self.font_name, name)
+            if path is not None:
+                self.fonts['NORMAL'] = ImageFont.truetype(path, self.font_size)
+                break
+        else:
+            raise FontNotFound(f'No usable fonts named: "{self.font_name}"')
+        for style in ('ITALIC', 'BOLD', 'BOLDITALIC'):
+            for stylename in STYLES[style]:
+                path = self._get_nix_font_path(self.font_name, stylename)
+                if path is not None:
+                    self.fonts[style] = ImageFont.truetype(path, self.font_size)
+                    break
+            else:
+                if style == 'BOLDITALIC':
+                    self.fonts[style] = self.fonts['BOLD']
+                else:
+                    self.fonts[style] = self.fonts['NORMAL']
+
+    def _get_mac_font_path(self, font_map, name, style):
+        return font_map.get((name + ' ' + style).strip().lower())
+
+    def _create_mac(self):
+        font_map = {}
+        for font_dir in (os.path.join(os.getenv("HOME"), 'Library/Fonts/'),
+                         '/Library/Fonts/', '/System/Library/Fonts/'):
+            font_map.update(
+                (os.path.splitext(f)[0].lower(), os.path.join(font_dir, f))
+                for f in os.listdir(font_dir)
+                if f.lower().endswith(('ttf', 'ttc')))
+
+        for name in STYLES['NORMAL']:
+            path = self._get_mac_font_path(font_map, self.font_name, name)
+            if path is not None:
+                self.fonts['NORMAL'] = ImageFont.truetype(path, self.font_size)
+                break
+        else:
+            raise FontNotFound(f'No usable fonts named: "{self.font_name}"')
+        for style in ('ITALIC', 'BOLD', 'BOLDITALIC'):
+            for stylename in STYLES[style]:
+                path = self._get_mac_font_path(font_map, self.font_name, stylename)
+                if path is not None:
+                    self.fonts[style] = ImageFont.truetype(path, self.font_size)
+                    break
+            else:
+                if style == 'BOLDITALIC':
+                    self.fonts[style] = self.fonts['BOLD']
+                else:
+                    self.fonts[style] = self.fonts['NORMAL']
+
+    def _lookup_win(self, key, basename, styles, fail=False):
+        for suffix in ('', ' (TrueType)'):
+            for style in styles:
+                try:
+                    valname = '{}{}{}'.format(basename, style and ' '+style, suffix)
+                    val, _ = _winreg.QueryValueEx(key, valname)
+                    return val
+                except OSError:
+                    continue
+        else:
+            if fail:
+                raise FontNotFound(f'Font {basename} ({styles[0]}) not found in registry')
+            return None
+
+    def _create_win(self):
+        lookuperror = None
+        keynames = [ (_winreg.HKEY_CURRENT_USER, r'Software\Microsoft\Windows NT\CurrentVersion\Fonts'),
+                     (_winreg.HKEY_CURRENT_USER, r'Software\Microsoft\Windows\CurrentVersion\Fonts'),
+                     (_winreg.HKEY_LOCAL_MACHINE, r'Software\Microsoft\Windows NT\CurrentVersion\Fonts'),
+                     (_winreg.HKEY_LOCAL_MACHINE, r'Software\Microsoft\Windows\CurrentVersion\Fonts') ]
+        for keyname in keynames:
+            try:
+                key = _winreg.OpenKey(*keyname)
+                try:
+                    path = self._lookup_win(key, self.font_name, STYLES['NORMAL'], True)
+                    self.fonts['NORMAL'] = ImageFont.truetype(path, self.font_size)
+                    for style in ('ITALIC', 'BOLD', 'BOLDITALIC'):
+                        path = self._lookup_win(key, self.font_name, STYLES[style])
+                        if path:
+                            self.fonts[style] = ImageFont.truetype(path, self.font_size)
+                        else:
+                            if style == 'BOLDITALIC':
+                                self.fonts[style] = self.fonts['BOLD']
+                            else:
+                                self.fonts[style] = self.fonts['NORMAL']
+                    return
+                except FontNotFound as err:
+                    lookuperror = err
+                finally:
+                    _winreg.CloseKey(key)
+            except OSError:
+                pass
+        else:
+            # If we get here, we checked all registry keys and had no luck
+            # We can be in one of two situations now:
+            # * All key lookups failed. In this case lookuperror is None and we
+            #   will raise a generic error
+            # * At least one lookup failed with a FontNotFound error. In this
+            #   case, we will raise that as a more specific error
+            if lookuperror:
+                raise lookuperror
+            raise FontNotFound('Can\'t open Windows font registry key')
+
+    def get_char_size(self):
+        """
+        Get the character size.
+        """
+        return self.get_text_size('M')
+
+    def get_text_size(self, text):
+        """
+        Get the text size (width, height).
+        """
+        font = self.fonts['NORMAL']
+        if hasattr(font, 'getbbox'):  # Pillow >= 9.2.0
+            return font.getbbox(text)[2:4]
+        else:
+            return font.getsize(text)
+
+    def get_font(self, bold, oblique):
+        """
+        Get the font based on bold and italic flags.
+        """
+        if bold and oblique:
+            if self.variable:
+                return self.get_style('BOLDITALIC')
+
+            return self.fonts['BOLDITALIC']
+        elif bold:
+            if self.variable:
+                return self.get_style('BOLD')
+
+            return self.fonts['BOLD']
+        elif oblique:
+            if self.variable:
+                return self.get_style('ITALIC')
+
+            return self.fonts['ITALIC']
+        else:
+            if self.variable:
+                return self.get_style('NORMAL')
+
+            return self.fonts['NORMAL']
+
+    def get_style(self, style):
+        """
+        Get the specified style of the font if it is a variable font.
+        If not found, return the normal font.
+        """
+        font = self.fonts[style]
+        for style_name in STYLES[style]:
+            try:
+                font.set_variation_by_name(style_name)
+                return font
+            except ValueError:
+                pass
+            except OSError:
+                return font
+
+        return font
+
+
+class ImageFormatter(Formatter):
+    """
+    Create a PNG image from source code. This uses the Python Imaging Library to
+    generate a pixmap from the source code.
+
+    .. versionadded:: 0.10
+
+    Additional options accepted:
+
+    `image_format`
+        An image format to output to that is recognised by PIL, these include:
+
+        * "PNG" (default)
+        * "JPEG"
+        * "BMP"
+        * "GIF"
+
+    `line_pad`
+        The extra spacing (in pixels) between each line of text.
+
+        Default: 2
+
+    `font_name`
+        The font name to be used as the base font from which others, such as
+        bold and italic fonts will be generated.  This really should be a
+        monospace font to look sane.
+        If a filename or a file-like object is specified, the user must
+        provide different styles of the font.
+
+        Default: "Courier New" on Windows, "Menlo" on Mac OS, and
+                 "DejaVu Sans Mono" on \\*nix
+
+    `font_size`
+        The font size in points to be used.
+
+        Default: 14
+
+    `image_pad`
+        The padding, in pixels to be used at each edge of the resulting image.
+
+        Default: 10
+
+    `line_numbers`
+        Whether line numbers should be shown: True/False
+
+        Default: True
+
+    `line_number_start`
+        The line number of the first line.
+
+        Default: 1
+
+    `line_number_step`
+        The step used when printing line numbers.
+
+        Default: 1
+
+    `line_number_bg`
+        The background colour (in "#123456" format) of the line number bar, or
+        None to use the style background color.
+
+        Default: "#eed"
+
+    `line_number_fg`
+        The text color of the line numbers (in "#123456"-like format).
+
+        Default: "#886"
+
+    `line_number_chars`
+        The number of columns of line numbers allowable in the line number
+        margin.
+
+        Default: 2
+
+    `line_number_bold`
+        Whether line numbers will be bold: True/False
+
+        Default: False
+
+    `line_number_italic`
+        Whether line numbers will be italicized: True/False
+
+        Default: False
+
+    `line_number_separator`
+        Whether a line will be drawn between the line number area and the
+        source code area: True/False
+
+        Default: True
+
+    `line_number_pad`
+        The horizontal padding (in pixels) between the line number margin, and
+        the source code area.
+
+        Default: 6
+
+    `hl_lines`
+        Specify a list of lines to be highlighted.
+
+        .. versionadded:: 1.2
+
+        Default: empty list
+
+    `hl_color`
+        Specify the color for highlighting lines.
+
+        .. versionadded:: 1.2
+
+        Default: highlight color of the selected style
+    """
+
+    # Required by the pygments mapper
+    name = 'img'
+    aliases = ['img', 'IMG', 'png']
+    filenames = ['*.png']
+
+    unicodeoutput = False
+
+    default_image_format = 'png'
+
+    def __init__(self, **options):
+        """
+        See the class docstring for explanation of options.
+        """
+        if not pil_available:
+            raise PilNotAvailable(
+                'Python Imaging Library is required for this formatter')
+        Formatter.__init__(self, **options)
+        self.encoding = 'latin1'  # let pygments.format() do the right thing
+        # Read the style
+        self.styles = dict(self.style)
+        if self.style.background_color is None:
+            self.background_color = '#fff'
+        else:
+            self.background_color = self.style.background_color
+        # Image options
+        self.image_format = get_choice_opt(
+            options, 'image_format', ['png', 'jpeg', 'gif', 'bmp'],
+            self.default_image_format, normcase=True)
+        self.image_pad = get_int_opt(options, 'image_pad', 10)
+        self.line_pad = get_int_opt(options, 'line_pad', 2)
+        # The fonts
+        fontsize = get_int_opt(options, 'font_size', 14)
+        self.fonts = FontManager(options.get('font_name', ''), fontsize)
+        self.fontw, self.fonth = self.fonts.get_char_size()
+        # Line number options
+        self.line_number_fg = options.get('line_number_fg', '#886')
+        self.line_number_bg = options.get('line_number_bg', '#eed')
+        self.line_number_chars = get_int_opt(options,
+                                             'line_number_chars', 2)
+        self.line_number_bold = get_bool_opt(options,
+                                             'line_number_bold', False)
+        self.line_number_italic = get_bool_opt(options,
+                                               'line_number_italic', False)
+        self.line_number_pad = get_int_opt(options, 'line_number_pad', 6)
+        self.line_numbers = get_bool_opt(options, 'line_numbers', True)
+        self.line_number_separator = get_bool_opt(options,
+                                                  'line_number_separator', True)
+        self.line_number_step = get_int_opt(options, 'line_number_step', 1)
+        self.line_number_start = get_int_opt(options, 'line_number_start', 1)
+        if self.line_numbers:
+            self.line_number_width = (self.fontw * self.line_number_chars +
+                                      self.line_number_pad * 2)
+        else:
+            self.line_number_width = 0
+        self.hl_lines = []
+        hl_lines_str = get_list_opt(options, 'hl_lines', [])
+        for line in hl_lines_str:
+            try:
+                self.hl_lines.append(int(line))
+            except ValueError:
+                pass
+        self.hl_color = options.get('hl_color',
+                                    self.style.highlight_color) or '#f90'
+        self.drawables = []
+
+    def get_style_defs(self, arg=''):
+        raise NotImplementedError('The -S option is meaningless for the image '
+                                  'formatter. Use -O style=<stylename> instead.')
+
+    def _get_line_height(self):
+        """
+        Get the height of a line.
+        """
+        return self.fonth + self.line_pad
+
+    def _get_line_y(self, lineno):
+        """
+        Get the Y coordinate of a line number.
+        """
+        return lineno * self._get_line_height() + self.image_pad
+
+    def _get_char_width(self):
+        """
+        Get the width of a character.
+        """
+        return self.fontw
+
+    def _get_char_x(self, linelength):
+        """
+        Get the X coordinate of a character position.
+        """
+        return linelength + self.image_pad + self.line_number_width
+
+    def _get_text_pos(self, linelength, lineno):
+        """
+        Get the actual position for a character and line position.
+        """
+        return self._get_char_x(linelength), self._get_line_y(lineno)
+
+    def _get_linenumber_pos(self, lineno):
+        """
+        Get the actual position for the start of a line number.
+        """
+        return (self.image_pad, self._get_line_y(lineno))
+
+    def _get_text_color(self, style):
+        """
+        Get the correct color for the token from the style.
+        """
+        if style['color'] is not None:
+            fill = '#' + style['color']
+        else:
+            fill = '#000'
+        return fill
+
+    def _get_text_bg_color(self, style):
+        """
+        Get the correct background color for the token from the style.
+        """
+        if style['bgcolor'] is not None:
+            bg_color = '#' + style['bgcolor']
+        else:
+            bg_color = None
+        return bg_color
+
+    def _get_style_font(self, style):
+        """
+        Get the correct font for the style.
+        """
+        return self.fonts.get_font(style['bold'], style['italic'])
+
+    def _get_image_size(self, maxlinelength, maxlineno):
+        """
+        Get the required image size.
+        """
+        return (self._get_char_x(maxlinelength) + self.image_pad,
+                self._get_line_y(maxlineno + 0) + self.image_pad)
+
+    def _draw_linenumber(self, posno, lineno):
+        """
+        Remember a line number drawable to paint later.
+        """
+        self._draw_text(
+            self._get_linenumber_pos(posno),
+            str(lineno).rjust(self.line_number_chars),
+            font=self.fonts.get_font(self.line_number_bold,
+                                     self.line_number_italic),
+            text_fg=self.line_number_fg,
+            text_bg=None,
+        )
+
+    def _draw_text(self, pos, text, font, text_fg, text_bg):
+        """
+        Remember a single drawable tuple to paint later.
+        """
+        self.drawables.append((pos, text, font, text_fg, text_bg))
+
+    def _create_drawables(self, tokensource):
+        """
+        Create drawables for the token content.
+        """
+        lineno = charno = maxcharno = 0
+        maxlinelength = linelength = 0
+        for ttype, value in tokensource:
+            while ttype not in self.styles:
+                ttype = ttype.parent
+            style = self.styles[ttype]
+            # TODO: make sure tab expansion happens earlier in the chain.  It
+            # really ought to be done on the input, as to do it right here is
+            # quite complex.
+            value = value.expandtabs(4)
+            lines = value.splitlines(True)
+            # print lines
+            for i, line in enumerate(lines):
+                temp = line.rstrip('\n')
+                if temp:
+                    self._draw_text(
+                        self._get_text_pos(linelength, lineno),
+                        temp,
+                        font = self._get_style_font(style),
+                        text_fg = self._get_text_color(style),
+                        text_bg = self._get_text_bg_color(style),
+                    )
+                    temp_width, _ = self.fonts.get_text_size(temp)
+                    linelength += temp_width
+                    maxlinelength = max(maxlinelength, linelength)
+                    charno += len(temp)
+                    maxcharno = max(maxcharno, charno)
+                if line.endswith('\n'):
+                    # add a line for each extra line in the value
+                    linelength = 0
+                    charno = 0
+                    lineno += 1
+        self.maxlinelength = maxlinelength
+        self.maxcharno = maxcharno
+        self.maxlineno = lineno
+
+    def _draw_line_numbers(self):
+        """
+        Create drawables for the line numbers.
+        """
+        if not self.line_numbers:
+            return
+        for p in range(self.maxlineno):
+            n = p + self.line_number_start
+            if (n % self.line_number_step) == 0:
+                self._draw_linenumber(p, n)
+
+    def _paint_line_number_bg(self, im):
+        """
+        Paint the line number background on the image.
+        """
+        if not self.line_numbers:
+            return
+        if self.line_number_fg is None:
+            return
+        draw = ImageDraw.Draw(im)
+        recth = im.size[-1]
+        rectw = self.image_pad + self.line_number_width - self.line_number_pad
+        draw.rectangle([(0, 0), (rectw, recth)],
+                       fill=self.line_number_bg)
+        if self.line_number_separator:
+            draw.line([(rectw, 0), (rectw, recth)], fill=self.line_number_fg)
+        del draw
+
+    def format(self, tokensource, outfile):
+        """
+        Format ``tokensource``, an iterable of ``(tokentype, tokenstring)``
+        tuples and write it into ``outfile``.
+
+        This implementation calculates where it should draw each token on the
+        pixmap, then calculates the required pixmap size and draws the items.
+        """
+        self._create_drawables(tokensource)
+        self._draw_line_numbers()
+        im = Image.new(
+            'RGB',
+            self._get_image_size(self.maxlinelength, self.maxlineno),
+            self.background_color
+        )
+        self._paint_line_number_bg(im)
+        draw = ImageDraw.Draw(im)
+        # Highlight
+        if self.hl_lines:
+            x = self.image_pad + self.line_number_width - self.line_number_pad + 1
+            recth = self._get_line_height()
+            rectw = im.size[0] - x
+            for linenumber in self.hl_lines:
+                y = self._get_line_y(linenumber - 1)
+                draw.rectangle([(x, y), (x + rectw, y + recth)],
+                               fill=self.hl_color)
+        for pos, value, font, text_fg, text_bg in self.drawables:
+            if text_bg:
+                # see deprecations https://pillow.readthedocs.io/en/stable/releasenotes/9.2.0.html#font-size-and-offset-methods
+                if hasattr(draw, 'textsize'):
+                    text_size = draw.textsize(text=value, font=font)
+                else:
+                    text_size = font.getbbox(value)[2:]
+                draw.rectangle([pos[0], pos[1], pos[0] + text_size[0], pos[1] + text_size[1]], fill=text_bg)
+            draw.text(pos, value, font=font, fill=text_fg)
+        im.save(outfile, self.image_format.upper())
+
+
+# Add one formatter per format, so that the "-f gif" option gives the correct result
+# when used in pygmentize.
+
+class GifImageFormatter(ImageFormatter):
+    """
+    Create a GIF image from source code. This uses the Python Imaging Library to
+    generate a pixmap from the source code.
+
+    .. versionadded:: 1.0
+    """
+
+    name = 'img_gif'
+    aliases = ['gif']
+    filenames = ['*.gif']
+    default_image_format = 'gif'
+
+
+class JpgImageFormatter(ImageFormatter):
+    """
+    Create a JPEG image from source code. This uses the Python Imaging Library to
+    generate a pixmap from the source code.
+
+    .. versionadded:: 1.0
+    """
+
+    name = 'img_jpg'
+    aliases = ['jpg', 'jpeg']
+    filenames = ['*.jpg']
+    default_image_format = 'jpeg'
+
+
+class BmpImageFormatter(ImageFormatter):
+    """
+    Create a bitmap image from source code. This uses the Python Imaging Library to
+    generate a pixmap from the source code.
+
+    .. versionadded:: 1.0
+    """
+
+    name = 'img_bmp'
+    aliases = ['bmp', 'bitmap']
+    filenames = ['*.bmp']
+    default_image_format = 'bmp'
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/irc.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/irc.py
new file mode 100644
index 00000000..468c2876
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/irc.py
@@ -0,0 +1,154 @@
+"""
+    pygments.formatters.irc
+    ~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for IRC output
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.token import Keyword, Name, Comment, String, Error, \
+    Number, Operator, Generic, Token, Whitespace
+from pip._vendor.pygments.util import get_choice_opt
+
+
+__all__ = ['IRCFormatter']
+
+
+#: Map token types to a tuple of color values for light and dark
+#: backgrounds.
+IRC_COLORS = {
+    Token:              ('',            ''),
+
+    Whitespace:         ('gray',   'brightblack'),
+    Comment:            ('gray',   'brightblack'),
+    Comment.Preproc:    ('cyan',        'brightcyan'),
+    Keyword:            ('blue',    'brightblue'),
+    Keyword.Type:       ('cyan',        'brightcyan'),
+    Operator.Word:      ('magenta',      'brightcyan'),
+    Name.Builtin:       ('cyan',        'brightcyan'),
+    Name.Function:      ('green',   'brightgreen'),
+    Name.Namespace:     ('_cyan_',      '_brightcyan_'),
+    Name.Class:         ('_green_', '_brightgreen_'),
+    Name.Exception:     ('cyan',        'brightcyan'),
+    Name.Decorator:     ('brightblack',    'gray'),
+    Name.Variable:      ('red',     'brightred'),
+    Name.Constant:      ('red',     'brightred'),
+    Name.Attribute:     ('cyan',        'brightcyan'),
+    Name.Tag:           ('brightblue',        'brightblue'),
+    String:             ('yellow',       'yellow'),
+    Number:             ('blue',    'brightblue'),
+
+    Generic.Deleted:    ('brightred',        'brightred'),
+    Generic.Inserted:   ('green',  'brightgreen'),
+    Generic.Heading:    ('**',         '**'),
+    Generic.Subheading: ('*magenta*',   '*brightmagenta*'),
+    Generic.Error:      ('brightred',        'brightred'),
+
+    Error:              ('_brightred_',      '_brightred_'),
+}
+
+
+IRC_COLOR_MAP = {
+    'white': 0,
+    'black': 1,
+    'blue': 2,
+    'brightgreen': 3,
+    'brightred': 4,
+    'yellow': 5,
+    'magenta': 6,
+    'orange': 7,
+    'green': 7, #compat w/ ansi
+    'brightyellow': 8,
+    'lightgreen': 9,
+    'brightcyan': 9, # compat w/ ansi
+    'cyan': 10,
+    'lightblue': 11,
+    'red': 11, # compat w/ ansi
+    'brightblue': 12,
+    'brightmagenta': 13,
+    'brightblack': 14,
+    'gray': 15,
+}
+
+def ircformat(color, text):
+    if len(color) < 1:
+        return text
+    add = sub = ''
+    if '_' in color: # italic
+        add += '\x1D'
+        sub = '\x1D' + sub
+        color = color.strip('_')
+    if '*' in color: # bold
+        add += '\x02'
+        sub = '\x02' + sub
+        color = color.strip('*')
+    # underline (\x1F) not supported
+    # backgrounds (\x03FF,BB) not supported
+    if len(color) > 0: # actual color - may have issues with ircformat("red", "blah")+"10" type stuff
+        add += '\x03' + str(IRC_COLOR_MAP[color]).zfill(2)
+        sub = '\x03' + sub
+    return add + text + sub
+    return '<'+add+'>'+text+'</'+sub+'>'
+
+
+class IRCFormatter(Formatter):
+    r"""
+    Format tokens with IRC color sequences
+
+    The `get_style_defs()` method doesn't do anything special since there is
+    no support for common styles.
+
+    Options accepted:
+
+    `bg`
+        Set to ``"light"`` or ``"dark"`` depending on the terminal's background
+        (default: ``"light"``).
+
+    `colorscheme`
+        A dictionary mapping token types to (lightbg, darkbg) color names or
+        ``None`` (default: ``None`` = use builtin colorscheme).
+
+    `linenos`
+        Set to ``True`` to have line numbers in the output as well
+        (default: ``False`` = no line numbers).
+    """
+    name = 'IRC'
+    aliases = ['irc', 'IRC']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self.darkbg = get_choice_opt(options, 'bg',
+                                     ['light', 'dark'], 'light') == 'dark'
+        self.colorscheme = options.get('colorscheme', None) or IRC_COLORS
+        self.linenos = options.get('linenos', False)
+        self._lineno = 0
+
+    def _write_lineno(self, outfile):
+        if self.linenos:
+            self._lineno += 1
+            outfile.write("%04d: " % self._lineno)
+
+    def format_unencoded(self, tokensource, outfile):
+        self._write_lineno(outfile)
+
+        for ttype, value in tokensource:
+            color = self.colorscheme.get(ttype)
+            while color is None:
+                ttype = ttype[:-1]
+                color = self.colorscheme.get(ttype)
+            if color:
+                color = color[self.darkbg]
+                spl = value.split('\n')
+                for line in spl[:-1]:
+                    if line:
+                        outfile.write(ircformat(color, line))
+                    outfile.write('\n')
+                    self._write_lineno(outfile)
+                if spl[-1]:
+                    outfile.write(ircformat(color, spl[-1]))
+            else:
+                outfile.write(value)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/latex.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/latex.py
new file mode 100644
index 00000000..0ec9089b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/latex.py
@@ -0,0 +1,518 @@
+"""
+    pygments.formatters.latex
+    ~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for LaTeX fancyvrb output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from io import StringIO
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.lexer import Lexer, do_insertions
+from pip._vendor.pygments.token import Token, STANDARD_TYPES
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt
+
+
+__all__ = ['LatexFormatter']
+
+
+def escape_tex(text, commandprefix):
+    return text.replace('\\', '\x00'). \
+                replace('{', '\x01'). \
+                replace('}', '\x02'). \
+                replace('\x00', rf'\{commandprefix}Zbs{{}}'). \
+                replace('\x01', rf'\{commandprefix}Zob{{}}'). \
+                replace('\x02', rf'\{commandprefix}Zcb{{}}'). \
+                replace('^', rf'\{commandprefix}Zca{{}}'). \
+                replace('_', rf'\{commandprefix}Zus{{}}'). \
+                replace('&', rf'\{commandprefix}Zam{{}}'). \
+                replace('<', rf'\{commandprefix}Zlt{{}}'). \
+                replace('>', rf'\{commandprefix}Zgt{{}}'). \
+                replace('#', rf'\{commandprefix}Zsh{{}}'). \
+                replace('%', rf'\{commandprefix}Zpc{{}}'). \
+                replace('$', rf'\{commandprefix}Zdl{{}}'). \
+                replace('-', rf'\{commandprefix}Zhy{{}}'). \
+                replace("'", rf'\{commandprefix}Zsq{{}}'). \
+                replace('"', rf'\{commandprefix}Zdq{{}}'). \
+                replace('~', rf'\{commandprefix}Zti{{}}')
+
+
+DOC_TEMPLATE = r'''
+\documentclass{%(docclass)s}
+\usepackage{fancyvrb}
+\usepackage{color}
+\usepackage[%(encoding)s]{inputenc}
+%(preamble)s
+
+%(styledefs)s
+
+\begin{document}
+
+\section*{%(title)s}
+
+%(code)s
+\end{document}
+'''
+
+## Small explanation of the mess below :)
+#
+# The previous version of the LaTeX formatter just assigned a command to
+# each token type defined in the current style.  That obviously is
+# problematic if the highlighted code is produced for a different style
+# than the style commands themselves.
+#
+# This version works much like the HTML formatter which assigns multiple
+# CSS classes to each <span> tag, from the most specific to the least
+# specific token type, thus falling back to the parent token type if one
+# is not defined.  Here, the classes are there too and use the same short
+# forms given in token.STANDARD_TYPES.
+#
+# Highlighted code now only uses one custom command, which by default is
+# \PY and selectable by the commandprefix option (and in addition the
+# escapes \PYZat, \PYZlb and \PYZrb which haven't been renamed for
+# backwards compatibility purposes).
+#
+# \PY has two arguments: the classes, separated by +, and the text to
+# render in that style.  The classes are resolved into the respective
+# style commands by magic, which serves to ignore unknown classes.
+#
+# The magic macros are:
+# * \PY@it, \PY@bf, etc. are unconditionally wrapped around the text
+#   to render in \PY@do.  Their definition determines the style.
+# * \PY@reset resets \PY@it etc. to do nothing.
+# * \PY@toks parses the list of classes, using magic inspired by the
+#   keyval package (but modified to use plusses instead of commas
+#   because fancyvrb redefines commas inside its environments).
+# * \PY@tok processes one class, calling the \PY@tok@classname command
+#   if it exists.
+# * \PY@tok@classname sets the \PY@it etc. to reflect the chosen style
+#   for its class.
+# * \PY resets the style, parses the classnames and then calls \PY@do.
+#
+# Tip: to read this code, print it out in substituted form using e.g.
+# >>> print STYLE_TEMPLATE % {'cp': 'PY'}
+
+STYLE_TEMPLATE = r'''
+\makeatletter
+\def\%(cp)s@reset{\let\%(cp)s@it=\relax \let\%(cp)s@bf=\relax%%
+    \let\%(cp)s@ul=\relax \let\%(cp)s@tc=\relax%%
+    \let\%(cp)s@bc=\relax \let\%(cp)s@ff=\relax}
+\def\%(cp)s@tok#1{\csname %(cp)s@tok@#1\endcsname}
+\def\%(cp)s@toks#1+{\ifx\relax#1\empty\else%%
+    \%(cp)s@tok{#1}\expandafter\%(cp)s@toks\fi}
+\def\%(cp)s@do#1{\%(cp)s@bc{\%(cp)s@tc{\%(cp)s@ul{%%
+    \%(cp)s@it{\%(cp)s@bf{\%(cp)s@ff{#1}}}}}}}
+\def\%(cp)s#1#2{\%(cp)s@reset\%(cp)s@toks#1+\relax+\%(cp)s@do{#2}}
+
+%(styles)s
+
+\def\%(cp)sZbs{\char`\\}
+\def\%(cp)sZus{\char`\_}
+\def\%(cp)sZob{\char`\{}
+\def\%(cp)sZcb{\char`\}}
+\def\%(cp)sZca{\char`\^}
+\def\%(cp)sZam{\char`\&}
+\def\%(cp)sZlt{\char`\<}
+\def\%(cp)sZgt{\char`\>}
+\def\%(cp)sZsh{\char`\#}
+\def\%(cp)sZpc{\char`\%%}
+\def\%(cp)sZdl{\char`\$}
+\def\%(cp)sZhy{\char`\-}
+\def\%(cp)sZsq{\char`\'}
+\def\%(cp)sZdq{\char`\"}
+\def\%(cp)sZti{\char`\~}
+%% for compatibility with earlier versions
+\def\%(cp)sZat{@}
+\def\%(cp)sZlb{[}
+\def\%(cp)sZrb{]}
+\makeatother
+'''
+
+
+def _get_ttype_name(ttype):
+    fname = STANDARD_TYPES.get(ttype)
+    if fname:
+        return fname
+    aname = ''
+    while fname is None:
+        aname = ttype[-1] + aname
+        ttype = ttype.parent
+        fname = STANDARD_TYPES.get(ttype)
+    return fname + aname
+
+
+class LatexFormatter(Formatter):
+    r"""
+    Format tokens as LaTeX code. This needs the `fancyvrb` and `color`
+    standard packages.
+
+    Without the `full` option, code is formatted as one ``Verbatim``
+    environment, like this:
+
+    .. sourcecode:: latex
+
+        \begin{Verbatim}[commandchars=\\\{\}]
+        \PY{k}{def }\PY{n+nf}{foo}(\PY{n}{bar}):
+            \PY{k}{pass}
+        \end{Verbatim}
+
+    Wrapping can be disabled using the `nowrap` option.
+
+    The special command used here (``\PY``) and all the other macros it needs
+    are output by the `get_style_defs` method.
+
+    With the `full` option, a complete LaTeX document is output, including
+    the command definitions in the preamble.
+
+    The `get_style_defs()` method of a `LatexFormatter` returns a string
+    containing ``\def`` commands defining the macros needed inside the
+    ``Verbatim`` environments.
+
+    Additional options accepted:
+
+    `nowrap`
+        If set to ``True``, don't wrap the tokens at all, not even inside a
+        ``\begin{Verbatim}`` environment. This disables most other options
+        (default: ``False``).
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+
+    `full`
+        Tells the formatter to output a "full" document, i.e. a complete
+        self-contained document (default: ``False``).
+
+    `title`
+        If `full` is true, the title that should be used to caption the
+        document (default: ``''``).
+
+    `docclass`
+        If the `full` option is enabled, this is the document class to use
+        (default: ``'article'``).
+
+    `preamble`
+        If the `full` option is enabled, this can be further preamble commands,
+        e.g. ``\usepackage`` (default: ``''``).
+
+    `linenos`
+        If set to ``True``, output line numbers (default: ``False``).
+
+    `linenostart`
+        The line number for the first line (default: ``1``).
+
+    `linenostep`
+        If set to a number n > 1, only every nth line number is printed.
+
+    `verboptions`
+        Additional options given to the Verbatim environment (see the *fancyvrb*
+        docs for possible values) (default: ``''``).
+
+    `commandprefix`
+        The LaTeX commands used to produce colored output are constructed
+        using this prefix and some letters (default: ``'PY'``).
+
+        .. versionadded:: 0.7
+        .. versionchanged:: 0.10
+           The default is now ``'PY'`` instead of ``'C'``.
+
+    `texcomments`
+        If set to ``True``, enables LaTeX comment lines.  That is, LaTex markup
+        in comment tokens is not escaped so that LaTeX can render it (default:
+        ``False``).
+
+        .. versionadded:: 1.2
+
+    `mathescape`
+        If set to ``True``, enables LaTeX math mode escape in comments. That
+        is, ``'$...$'`` inside a comment will trigger math mode (default:
+        ``False``).
+
+        .. versionadded:: 1.2
+
+    `escapeinside`
+        If set to a string of length 2, enables escaping to LaTeX. Text
+        delimited by these 2 characters is read as LaTeX code and
+        typeset accordingly. It has no effect in string literals. It has
+        no effect in comments if `texcomments` or `mathescape` is
+        set. (default: ``''``).
+
+        .. versionadded:: 2.0
+
+    `envname`
+        Allows you to pick an alternative environment name replacing Verbatim.
+        The alternate environment still has to support Verbatim's option syntax.
+        (default: ``'Verbatim'``).
+
+        .. versionadded:: 2.0
+    """
+    name = 'LaTeX'
+    aliases = ['latex', 'tex']
+    filenames = ['*.tex']
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self.nowrap = get_bool_opt(options, 'nowrap', False)
+        self.docclass = options.get('docclass', 'article')
+        self.preamble = options.get('preamble', '')
+        self.linenos = get_bool_opt(options, 'linenos', False)
+        self.linenostart = abs(get_int_opt(options, 'linenostart', 1))
+        self.linenostep = abs(get_int_opt(options, 'linenostep', 1))
+        self.verboptions = options.get('verboptions', '')
+        self.nobackground = get_bool_opt(options, 'nobackground', False)
+        self.commandprefix = options.get('commandprefix', 'PY')
+        self.texcomments = get_bool_opt(options, 'texcomments', False)
+        self.mathescape = get_bool_opt(options, 'mathescape', False)
+        self.escapeinside = options.get('escapeinside', '')
+        if len(self.escapeinside) == 2:
+            self.left = self.escapeinside[0]
+            self.right = self.escapeinside[1]
+        else:
+            self.escapeinside = ''
+        self.envname = options.get('envname', 'Verbatim')
+
+        self._create_stylesheet()
+
+    def _create_stylesheet(self):
+        t2n = self.ttype2name = {Token: ''}
+        c2d = self.cmd2def = {}
+        cp = self.commandprefix
+
+        def rgbcolor(col):
+            if col:
+                return ','.join(['%.2f' % (int(col[i] + col[i + 1], 16) / 255.0)
+                                 for i in (0, 2, 4)])
+            else:
+                return '1,1,1'
+
+        for ttype, ndef in self.style:
+            name = _get_ttype_name(ttype)
+            cmndef = ''
+            if ndef['bold']:
+                cmndef += r'\let\$$@bf=\textbf'
+            if ndef['italic']:
+                cmndef += r'\let\$$@it=\textit'
+            if ndef['underline']:
+                cmndef += r'\let\$$@ul=\underline'
+            if ndef['roman']:
+                cmndef += r'\let\$$@ff=\textrm'
+            if ndef['sans']:
+                cmndef += r'\let\$$@ff=\textsf'
+            if ndef['mono']:
+                cmndef += r'\let\$$@ff=\textsf'
+            if ndef['color']:
+                cmndef += (r'\def\$$@tc##1{{\textcolor[rgb]{{{}}}{{##1}}}}'.format(rgbcolor(ndef['color'])))
+            if ndef['border']:
+                cmndef += (r'\def\$$@bc##1{{{{\setlength{{\fboxsep}}{{\string -\fboxrule}}'
+                           r'\fcolorbox[rgb]{{{}}}{{{}}}{{\strut ##1}}}}}}'.format(rgbcolor(ndef['border']),
+                            rgbcolor(ndef['bgcolor'])))
+            elif ndef['bgcolor']:
+                cmndef += (r'\def\$$@bc##1{{{{\setlength{{\fboxsep}}{{0pt}}'
+                           r'\colorbox[rgb]{{{}}}{{\strut ##1}}}}}}'.format(rgbcolor(ndef['bgcolor'])))
+            if cmndef == '':
+                continue
+            cmndef = cmndef.replace('$$', cp)
+            t2n[ttype] = name
+            c2d[name] = cmndef
+
+    def get_style_defs(self, arg=''):
+        """
+        Return the command sequences needed to define the commands
+        used to format text in the verbatim environment. ``arg`` is ignored.
+        """
+        cp = self.commandprefix
+        styles = []
+        for name, definition in self.cmd2def.items():
+            styles.append(rf'\@namedef{{{cp}@tok@{name}}}{{{definition}}}')
+        return STYLE_TEMPLATE % {'cp': self.commandprefix,
+                                 'styles': '\n'.join(styles)}
+
+    def format_unencoded(self, tokensource, outfile):
+        # TODO: add support for background colors
+        t2n = self.ttype2name
+        cp = self.commandprefix
+
+        if self.full:
+            realoutfile = outfile
+            outfile = StringIO()
+
+        if not self.nowrap:
+            outfile.write('\\begin{' + self.envname + '}[commandchars=\\\\\\{\\}')
+            if self.linenos:
+                start, step = self.linenostart, self.linenostep
+                outfile.write(',numbers=left' +
+                              (start and ',firstnumber=%d' % start or '') +
+                              (step and ',stepnumber=%d' % step or ''))
+            if self.mathescape or self.texcomments or self.escapeinside:
+                outfile.write(',codes={\\catcode`\\$=3\\catcode`\\^=7'
+                              '\\catcode`\\_=8\\relax}')
+            if self.verboptions:
+                outfile.write(',' + self.verboptions)
+            outfile.write(']\n')
+
+        for ttype, value in tokensource:
+            if ttype in Token.Comment:
+                if self.texcomments:
+                    # Try to guess comment starting lexeme and escape it ...
+                    start = value[0:1]
+                    for i in range(1, len(value)):
+                        if start[0] != value[i]:
+                            break
+                        start += value[i]
+
+                    value = value[len(start):]
+                    start = escape_tex(start, cp)
+
+                    # ... but do not escape inside comment.
+                    value = start + value
+                elif self.mathescape:
+                    # Only escape parts not inside a math environment.
+                    parts = value.split('$')
+                    in_math = False
+                    for i, part in enumerate(parts):
+                        if not in_math:
+                            parts[i] = escape_tex(part, cp)
+                        in_math = not in_math
+                    value = '$'.join(parts)
+                elif self.escapeinside:
+                    text = value
+                    value = ''
+                    while text:
+                        a, sep1, text = text.partition(self.left)
+                        if sep1:
+                            b, sep2, text = text.partition(self.right)
+                            if sep2:
+                                value += escape_tex(a, cp) + b
+                            else:
+                                value += escape_tex(a + sep1 + b, cp)
+                        else:
+                            value += escape_tex(a, cp)
+                else:
+                    value = escape_tex(value, cp)
+            elif ttype not in Token.Escape:
+                value = escape_tex(value, cp)
+            styles = []
+            while ttype is not Token:
+                try:
+                    styles.append(t2n[ttype])
+                except KeyError:
+                    # not in current style
+                    styles.append(_get_ttype_name(ttype))
+                ttype = ttype.parent
+            styleval = '+'.join(reversed(styles))
+            if styleval:
+                spl = value.split('\n')
+                for line in spl[:-1]:
+                    if line:
+                        outfile.write(f"\\{cp}{{{styleval}}}{{{line}}}")
+                    outfile.write('\n')
+                if spl[-1]:
+                    outfile.write(f"\\{cp}{{{styleval}}}{{{spl[-1]}}}")
+            else:
+                outfile.write(value)
+
+        if not self.nowrap:
+            outfile.write('\\end{' + self.envname + '}\n')
+
+        if self.full:
+            encoding = self.encoding or 'utf8'
+            # map known existings encodings from LaTeX distribution
+            encoding = {
+                'utf_8': 'utf8',
+                'latin_1': 'latin1',
+                'iso_8859_1': 'latin1',
+            }.get(encoding.replace('-', '_'), encoding)
+            realoutfile.write(DOC_TEMPLATE %
+                dict(docclass  = self.docclass,
+                     preamble  = self.preamble,
+                     title     = self.title,
+                     encoding  = encoding,
+                     styledefs = self.get_style_defs(),
+                     code      = outfile.getvalue()))
+
+
+class LatexEmbeddedLexer(Lexer):
+    """
+    This lexer takes one lexer as argument, the lexer for the language
+    being formatted, and the left and right delimiters for escaped text.
+
+    First everything is scanned using the language lexer to obtain
+    strings and comments. All other consecutive tokens are merged and
+    the resulting text is scanned for escaped segments, which are given
+    the Token.Escape type. Finally text that is not escaped is scanned
+    again with the language lexer.
+    """
+    def __init__(self, left, right, lang, **options):
+        self.left = left
+        self.right = right
+        self.lang = lang
+        Lexer.__init__(self, **options)
+
+    def get_tokens_unprocessed(self, text):
+        # find and remove all the escape tokens (replace with an empty string)
+        # this is very similar to DelegatingLexer.get_tokens_unprocessed.
+        buffered = ''
+        insertions = []
+        insertion_buf = []
+        for i, t, v in self._find_safe_escape_tokens(text):
+            if t is None:
+                if insertion_buf:
+                    insertions.append((len(buffered), insertion_buf))
+                    insertion_buf = []
+                buffered += v
+            else:
+                insertion_buf.append((i, t, v))
+        if insertion_buf:
+            insertions.append((len(buffered), insertion_buf))
+        return do_insertions(insertions,
+                             self.lang.get_tokens_unprocessed(buffered))
+
+    def _find_safe_escape_tokens(self, text):
+        """ find escape tokens that are not in strings or comments """
+        for i, t, v in self._filter_to(
+            self.lang.get_tokens_unprocessed(text),
+            lambda t: t in Token.Comment or t in Token.String
+        ):
+            if t is None:
+                for i2, t2, v2 in self._find_escape_tokens(v):
+                    yield i + i2, t2, v2
+            else:
+                yield i, None, v
+
+    def _filter_to(self, it, pred):
+        """ Keep only the tokens that match `pred`, merge the others together """
+        buf = ''
+        idx = 0
+        for i, t, v in it:
+            if pred(t):
+                if buf:
+                    yield idx, None, buf
+                    buf = ''
+                yield i, t, v
+            else:
+                if not buf:
+                    idx = i
+                buf += v
+        if buf:
+            yield idx, None, buf
+
+    def _find_escape_tokens(self, text):
+        """ Find escape tokens within text, give token=None otherwise """
+        index = 0
+        while text:
+            a, sep1, text = text.partition(self.left)
+            if a:
+                yield index, None, a
+                index += len(a)
+            if sep1:
+                b, sep2, text = text.partition(self.right)
+                if sep2:
+                    yield index + len(sep1), Token.Escape, b
+                    index += len(sep1) + len(b) + len(sep2)
+                else:
+                    yield index, Token.Error, sep1
+                    index += len(sep1)
+                    text = b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/other.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/other.py
new file mode 100644
index 00000000..de8d9dcf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/other.py
@@ -0,0 +1,160 @@
+"""
+    pygments.formatters.other
+    ~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Other formatters: NullFormatter, RawTokenFormatter.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.util import get_choice_opt
+from pip._vendor.pygments.token import Token
+from pip._vendor.pygments.console import colorize
+
+__all__ = ['NullFormatter', 'RawTokenFormatter', 'TestcaseFormatter']
+
+
+class NullFormatter(Formatter):
+    """
+    Output the text unchanged without any formatting.
+    """
+    name = 'Text only'
+    aliases = ['text', 'null']
+    filenames = ['*.txt']
+
+    def format(self, tokensource, outfile):
+        enc = self.encoding
+        for ttype, value in tokensource:
+            if enc:
+                outfile.write(value.encode(enc))
+            else:
+                outfile.write(value)
+
+
+class RawTokenFormatter(Formatter):
+    r"""
+    Format tokens as a raw representation for storing token streams.
+
+    The format is ``tokentype<TAB>repr(tokenstring)\n``. The output can later
+    be converted to a token stream with the `RawTokenLexer`, described in the
+    :doc:`lexer list <lexers>`.
+
+    Only two options are accepted:
+
+    `compress`
+        If set to ``'gz'`` or ``'bz2'``, compress the output with the given
+        compression algorithm after encoding (default: ``''``).
+    `error_color`
+        If set to a color name, highlight error tokens using that color.  If
+        set but with no value, defaults to ``'red'``.
+
+        .. versionadded:: 0.11
+
+    """
+    name = 'Raw tokens'
+    aliases = ['raw', 'tokens']
+    filenames = ['*.raw']
+
+    unicodeoutput = False
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        # We ignore self.encoding if it is set, since it gets set for lexer
+        # and formatter if given with -Oencoding on the command line.
+        # The RawTokenFormatter outputs only ASCII. Override here.
+        self.encoding = 'ascii'  # let pygments.format() do the right thing
+        self.compress = get_choice_opt(options, 'compress',
+                                       ['', 'none', 'gz', 'bz2'], '')
+        self.error_color = options.get('error_color', None)
+        if self.error_color is True:
+            self.error_color = 'red'
+        if self.error_color is not None:
+            try:
+                colorize(self.error_color, '')
+            except KeyError:
+                raise ValueError(f"Invalid color {self.error_color!r} specified")
+
+    def format(self, tokensource, outfile):
+        try:
+            outfile.write(b'')
+        except TypeError:
+            raise TypeError('The raw tokens formatter needs a binary '
+                            'output file')
+        if self.compress == 'gz':
+            import gzip
+            outfile = gzip.GzipFile('', 'wb', 9, outfile)
+
+            write = outfile.write
+            flush = outfile.close
+        elif self.compress == 'bz2':
+            import bz2
+            compressor = bz2.BZ2Compressor(9)
+
+            def write(text):
+                outfile.write(compressor.compress(text))
+
+            def flush():
+                outfile.write(compressor.flush())
+                outfile.flush()
+        else:
+            write = outfile.write
+            flush = outfile.flush
+
+        if self.error_color:
+            for ttype, value in tokensource:
+                line = b"%r\t%r\n" % (ttype, value)
+                if ttype is Token.Error:
+                    write(colorize(self.error_color, line))
+                else:
+                    write(line)
+        else:
+            for ttype, value in tokensource:
+                write(b"%r\t%r\n" % (ttype, value))
+        flush()
+
+
+TESTCASE_BEFORE = '''\
+    def testNeedsName(lexer):
+        fragment = %r
+        tokens = [
+'''
+TESTCASE_AFTER = '''\
+        ]
+        assert list(lexer.get_tokens(fragment)) == tokens
+'''
+
+
+class TestcaseFormatter(Formatter):
+    """
+    Format tokens as appropriate for a new testcase.
+
+    .. versionadded:: 2.0
+    """
+    name = 'Testcase'
+    aliases = ['testcase']
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        if self.encoding is not None and self.encoding != 'utf-8':
+            raise ValueError("Only None and utf-8 are allowed encodings.")
+
+    def format(self, tokensource, outfile):
+        indentation = ' ' * 12
+        rawbuf = []
+        outbuf = []
+        for ttype, value in tokensource:
+            rawbuf.append(value)
+            outbuf.append(f'{indentation}({ttype}, {value!r}),\n')
+
+        before = TESTCASE_BEFORE % (''.join(rawbuf),)
+        during = ''.join(outbuf)
+        after = TESTCASE_AFTER
+        if self.encoding is None:
+            outfile.write(before + during + after)
+        else:
+            outfile.write(before.encode('utf-8'))
+            outfile.write(during.encode('utf-8'))
+            outfile.write(after.encode('utf-8'))
+        outfile.flush()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/pangomarkup.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/pangomarkup.py
new file mode 100644
index 00000000..dfed53ab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/pangomarkup.py
@@ -0,0 +1,83 @@
+"""
+    pygments.formatters.pangomarkup
+    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for Pango markup output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.formatter import Formatter
+
+
+__all__ = ['PangoMarkupFormatter']
+
+
+_escape_table = {
+    ord('&'): '&amp;',
+    ord('<'): '&lt;',
+}
+
+
+def escape_special_chars(text, table=_escape_table):
+    """Escape & and < for Pango Markup."""
+    return text.translate(table)
+
+
+class PangoMarkupFormatter(Formatter):
+    """
+    Format tokens as Pango Markup code. It can then be rendered to an SVG.
+
+    .. versionadded:: 2.9
+    """
+
+    name = 'Pango Markup'
+    aliases = ['pango', 'pangomarkup']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+
+        self.styles = {}
+
+        for token, style in self.style:
+            start = ''
+            end = ''
+            if style['color']:
+                start += '<span fgcolor="#{}">'.format(style['color'])
+                end = '</span>' + end
+            if style['bold']:
+                start += '<b>'
+                end = '</b>' + end
+            if style['italic']:
+                start += '<i>'
+                end = '</i>' + end
+            if style['underline']:
+                start += '<u>'
+                end = '</u>' + end
+            self.styles[token] = (start, end)
+
+    def format_unencoded(self, tokensource, outfile):
+        lastval = ''
+        lasttype = None
+
+        outfile.write('<tt>')
+
+        for ttype, value in tokensource:
+            while ttype not in self.styles:
+                ttype = ttype.parent
+            if ttype == lasttype:
+                lastval += escape_special_chars(value)
+            else:
+                if lastval:
+                    stylebegin, styleend = self.styles[lasttype]
+                    outfile.write(stylebegin + lastval + styleend)
+                lastval = escape_special_chars(value)
+                lasttype = ttype
+
+        if lastval:
+            stylebegin, styleend = self.styles[lasttype]
+            outfile.write(stylebegin + lastval + styleend)
+
+        outfile.write('</tt>')
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/rtf.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/rtf.py
new file mode 100644
index 00000000..eca2a41a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/rtf.py
@@ -0,0 +1,349 @@
+"""
+    pygments.formatters.rtf
+    ~~~~~~~~~~~~~~~~~~~~~~~
+
+    A formatter that generates RTF files.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from collections import OrderedDict
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.style import _ansimap
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt, get_list_opt, surrogatepair
+
+
+__all__ = ['RtfFormatter']
+
+
+class RtfFormatter(Formatter):
+    """
+    Format tokens as RTF markup. This formatter automatically outputs full RTF
+    documents with color information and other useful stuff. Perfect for Copy and
+    Paste into Microsoft(R) Word(R) documents.
+
+    Please note that ``encoding`` and ``outencoding`` options are ignored.
+    The RTF format is ASCII natively, but handles unicode characters correctly
+    thanks to escape sequences.
+
+    .. versionadded:: 0.6
+
+    Additional options accepted:
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+
+    `fontface`
+        The used font family, for example ``Bitstream Vera Sans``. Defaults to
+        some generic font which is supposed to have fixed width.
+
+    `fontsize`
+        Size of the font used. Size is specified in half points. The
+        default is 24 half-points, giving a size 12 font.
+
+        .. versionadded:: 2.0
+
+    `linenos`
+        Turn on line numbering (default: ``False``).
+
+        .. versionadded:: 2.18
+
+    `lineno_fontsize`
+        Font size for line numbers. Size is specified in half points
+        (default: `fontsize`). 
+
+        .. versionadded:: 2.18
+
+    `lineno_padding`
+        Number of spaces between the (inline) line numbers and the
+        source code (default: ``2``).
+
+        .. versionadded:: 2.18
+
+    `linenostart`
+        The line number for the first line (default: ``1``).
+
+        .. versionadded:: 2.18
+
+    `linenostep`
+        If set to a number n > 1, only every nth line number is printed.
+
+        .. versionadded:: 2.18
+
+    `lineno_color`
+        Color for line numbers specified as a hex triplet, e.g. ``'5e5e5e'``. 
+        Defaults to the style's line number color if it is a hex triplet, 
+        otherwise ansi bright black.
+
+        .. versionadded:: 2.18
+
+    `hl_lines`
+        Specify a list of lines to be highlighted, as line numbers separated by
+        spaces, e.g. ``'3 7 8'``. The line numbers are relative to the input 
+        (i.e. the first line is line 1) unless `hl_linenostart` is set.
+
+        .. versionadded:: 2.18
+
+    `hl_color`
+        Color for highlighting the lines specified in `hl_lines`, specified as 
+        a hex triplet (default: style's `highlight_color`).
+
+        .. versionadded:: 2.18
+
+    `hl_linenostart`
+        If set to ``True`` line numbers in `hl_lines` are specified
+        relative to `linenostart` (default ``False``).
+
+        .. versionadded:: 2.18
+    """
+    name = 'RTF'
+    aliases = ['rtf']
+    filenames = ['*.rtf']
+
+    def __init__(self, **options):
+        r"""
+        Additional options accepted:
+
+        ``fontface``
+            Name of the font used. Could for example be ``'Courier New'``
+            to further specify the default which is ``'\fmodern'``. The RTF
+            specification claims that ``\fmodern`` are "Fixed-pitch serif
+            and sans serif fonts". Hope every RTF implementation thinks
+            the same about modern...
+
+        """
+        Formatter.__init__(self, **options)
+        self.fontface = options.get('fontface') or ''
+        self.fontsize = get_int_opt(options, 'fontsize', 0)
+        self.linenos = get_bool_opt(options, 'linenos', False)
+        self.lineno_fontsize = get_int_opt(options, 'lineno_fontsize',
+                                           self.fontsize)
+        self.lineno_padding = get_int_opt(options, 'lineno_padding', 2)
+        self.linenostart = abs(get_int_opt(options, 'linenostart', 1))
+        self.linenostep = abs(get_int_opt(options, 'linenostep', 1))
+        self.hl_linenostart = get_bool_opt(options, 'hl_linenostart', False)
+
+        self.hl_color = options.get('hl_color', '')
+        if not self.hl_color:
+            self.hl_color = self.style.highlight_color
+
+        self.hl_lines = []
+        for lineno in get_list_opt(options, 'hl_lines', []):
+            try:
+                lineno = int(lineno)
+                if self.hl_linenostart:
+                    lineno = lineno - self.linenostart + 1
+                self.hl_lines.append(lineno)
+            except ValueError:
+                pass
+
+        self.lineno_color = options.get('lineno_color', '')
+        if not self.lineno_color:
+            if  self.style.line_number_color == 'inherit':
+                # style color is the css value 'inherit'
+                # default to ansi bright-black
+                self.lineno_color = _ansimap['ansibrightblack']
+            else:
+                # style color is assumed to be a hex triplet as other
+                # colors in pygments/style.py
+                self.lineno_color = self.style.line_number_color
+
+        self.color_mapping = self._create_color_mapping()
+
+    def _escape(self, text):
+        return text.replace('\\', '\\\\') \
+                   .replace('{', '\\{') \
+                   .replace('}', '\\}')
+
+    def _escape_text(self, text):
+        # empty strings, should give a small performance improvement
+        if not text:
+            return ''
+
+        # escape text
+        text = self._escape(text)
+
+        buf = []
+        for c in text:
+            cn = ord(c)
+            if cn < (2**7):
+                # ASCII character
+                buf.append(str(c))
+            elif (2**7) <= cn < (2**16):
+                # single unicode escape sequence
+                buf.append('{\\u%d}' % cn)
+            elif (2**16) <= cn:
+                # RTF limits unicode to 16 bits.
+                # Force surrogate pairs
+                buf.append('{\\u%d}{\\u%d}' % surrogatepair(cn))
+
+        return ''.join(buf).replace('\n', '\\par')
+
+    @staticmethod
+    def hex_to_rtf_color(hex_color):
+        if hex_color[0] == "#":
+            hex_color = hex_color[1:]
+
+        return '\\red%d\\green%d\\blue%d;' % (
+                        int(hex_color[0:2], 16),
+                        int(hex_color[2:4], 16),
+                        int(hex_color[4:6], 16)
+                    )
+
+    def _split_tokens_on_newlines(self, tokensource):
+        """
+        Split tokens containing newline characters into multiple token
+        each representing a line of the input file. Needed for numbering
+        lines of e.g. multiline comments.
+        """
+        for ttype, value in tokensource:
+            if value == '\n':
+                yield (ttype, value)
+            elif "\n" in value:
+                lines = value.split("\n")
+                for line in lines[:-1]:
+                    yield (ttype, line+"\n")
+                if lines[-1]:
+                    yield (ttype, lines[-1])
+            else:
+                yield (ttype, value)
+
+    def _create_color_mapping(self):
+        """
+        Create a mapping of style hex colors to index/offset in
+        the RTF color table.
+        """
+        color_mapping = OrderedDict()
+        offset = 1
+
+        if self.linenos:
+            color_mapping[self.lineno_color] = offset
+            offset += 1
+
+        if self.hl_lines:
+            color_mapping[self.hl_color] = offset
+            offset += 1
+
+        for _, style in self.style:
+            for color in style['color'], style['bgcolor'], style['border']:
+                if color and color not in color_mapping:
+                    color_mapping[color] = offset
+                    offset += 1
+
+        return color_mapping
+
+    @property
+    def _lineno_template(self):
+        if self.lineno_fontsize != self.fontsize:
+            return '{{\\fs{} \\cf{} %s{}}}'.format(self.lineno_fontsize,
+                          self.color_mapping[self.lineno_color],
+                          " " * self.lineno_padding)
+
+        return '{{\\cf{} %s{}}}'.format(self.color_mapping[self.lineno_color],
+                      " " * self.lineno_padding)
+
+    @property
+    def _hl_open_str(self):
+        return rf'{{\highlight{self.color_mapping[self.hl_color]} '
+
+    @property
+    def _rtf_header(self):
+        lines = []
+        # rtf 1.8 header
+        lines.append('{\\rtf1\\ansi\\uc0\\deff0'
+                     '{\\fonttbl{\\f0\\fmodern\\fprq1\\fcharset0%s;}}'
+                     % (self.fontface and ' '
+                        + self._escape(self.fontface) or ''))
+
+        # color table
+        lines.append('{\\colortbl;')
+        for color, _ in self.color_mapping.items():
+            lines.append(self.hex_to_rtf_color(color))
+        lines.append('}')
+
+        # font and fontsize
+        lines.append('\\f0\\sa0')
+        if self.fontsize:
+            lines.append('\\fs%d' % self.fontsize)
+
+        # ensure Libre Office Writer imports and renders consecutive
+        # space characters the same width, needed for line numbering.
+        # https://bugs.documentfoundation.org/show_bug.cgi?id=144050
+        lines.append('\\dntblnsbdb')
+
+        return lines
+
+    def format_unencoded(self, tokensource, outfile):
+        for line in self._rtf_header:
+            outfile.write(line + "\n")
+
+        tokensource = self._split_tokens_on_newlines(tokensource)
+
+        # first pass of tokens to count lines, needed for line numbering
+        if self.linenos:
+            line_count = 0
+            tokens = [] # for copying the token source generator
+            for ttype, value in tokensource:
+                tokens.append((ttype, value))
+                if value.endswith("\n"):
+                    line_count += 1
+
+            # width of line number strings (for padding with spaces)
+            linenos_width = len(str(line_count+self.linenostart-1))
+
+            tokensource = tokens
+
+        # highlight stream
+        lineno = 1
+        start_new_line = True
+        for ttype, value in tokensource:
+            if start_new_line and lineno in self.hl_lines:
+                outfile.write(self._hl_open_str)
+
+            if start_new_line and self.linenos:
+                if (lineno-self.linenostart+1)%self.linenostep == 0:
+                    current_lineno = lineno + self.linenostart - 1
+                    lineno_str = str(current_lineno).rjust(linenos_width)
+                else:
+                    lineno_str = "".rjust(linenos_width)
+                outfile.write(self._lineno_template % lineno_str)
+
+            while not self.style.styles_token(ttype) and ttype.parent:
+                ttype = ttype.parent
+            style = self.style.style_for_token(ttype)
+            buf = []
+            if style['bgcolor']:
+                buf.append('\\cb%d' % self.color_mapping[style['bgcolor']])
+            if style['color']:
+                buf.append('\\cf%d' % self.color_mapping[style['color']])
+            if style['bold']:
+                buf.append('\\b')
+            if style['italic']:
+                buf.append('\\i')
+            if style['underline']:
+                buf.append('\\ul')
+            if style['border']:
+                buf.append('\\chbrdr\\chcfpat%d' %
+                           self.color_mapping[style['border']])
+            start = ''.join(buf)
+            if start:
+                outfile.write(f'{{{start} ')
+            outfile.write(self._escape_text(value))
+            if start:
+                outfile.write('}')
+            start_new_line = False
+
+            # complete line of input
+            if value.endswith("\n"):
+                # close line highlighting
+                if lineno in self.hl_lines:
+                    outfile.write('}')
+                # newline in RTF file after closing }
+                outfile.write("\n")
+
+                start_new_line = True
+                lineno += 1
+
+        outfile.write('}\n')
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/svg.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/svg.py
new file mode 100644
index 00000000..d3e018ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/svg.py
@@ -0,0 +1,185 @@
+"""
+    pygments.formatters.svg
+    ~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for SVG output.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.token import Comment
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt
+
+__all__ = ['SvgFormatter']
+
+
+def escape_html(text):
+    """Escape &, <, > as well as single and double quotes for HTML."""
+    return text.replace('&', '&amp;').  \
+                replace('<', '&lt;').   \
+                replace('>', '&gt;').   \
+                replace('"', '&quot;'). \
+                replace("'", '&#39;')
+
+
+class2style = {}
+
+class SvgFormatter(Formatter):
+    """
+    Format tokens as an SVG graphics file.  This formatter is still experimental.
+    Each line of code is a ``<text>`` element with explicit ``x`` and ``y``
+    coordinates containing ``<tspan>`` elements with the individual token styles.
+
+    By default, this formatter outputs a full SVG document including doctype
+    declaration and the ``<svg>`` root element.
+
+    .. versionadded:: 0.9
+
+    Additional options accepted:
+
+    `nowrap`
+        Don't wrap the SVG ``<text>`` elements in ``<svg><g>`` elements and
+        don't add a XML declaration and a doctype.  If true, the `fontfamily`
+        and `fontsize` options are ignored.  Defaults to ``False``.
+
+    `fontfamily`
+        The value to give the wrapping ``<g>`` element's ``font-family``
+        attribute, defaults to ``"monospace"``.
+
+    `fontsize`
+        The value to give the wrapping ``<g>`` element's ``font-size``
+        attribute, defaults to ``"14px"``.
+
+    `linenos`
+        If ``True``, add line numbers (default: ``False``).
+
+    `linenostart`
+        The line number for the first line (default: ``1``).
+
+    `linenostep`
+        If set to a number n > 1, only every nth line number is printed.
+
+    `linenowidth`
+        Maximum width devoted to line numbers (default: ``3*ystep``, sufficient
+        for up to 4-digit line numbers. Increase width for longer code blocks).
+
+    `xoffset`
+        Starting offset in X direction, defaults to ``0``.
+
+    `yoffset`
+        Starting offset in Y direction, defaults to the font size if it is given
+        in pixels, or ``20`` else.  (This is necessary since text coordinates
+        refer to the text baseline, not the top edge.)
+
+    `ystep`
+        Offset to add to the Y coordinate for each subsequent line.  This should
+        roughly be the text size plus 5.  It defaults to that value if the text
+        size is given in pixels, or ``25`` else.
+
+    `spacehack`
+        Convert spaces in the source to ``&#160;``, which are non-breaking
+        spaces.  SVG provides the ``xml:space`` attribute to control how
+        whitespace inside tags is handled, in theory, the ``preserve`` value
+        could be used to keep all whitespace as-is.  However, many current SVG
+        viewers don't obey that rule, so this option is provided as a workaround
+        and defaults to ``True``.
+    """
+    name = 'SVG'
+    aliases = ['svg']
+    filenames = ['*.svg']
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self.nowrap = get_bool_opt(options, 'nowrap', False)
+        self.fontfamily = options.get('fontfamily', 'monospace')
+        self.fontsize = options.get('fontsize', '14px')
+        self.xoffset = get_int_opt(options, 'xoffset', 0)
+        fs = self.fontsize.strip()
+        if fs.endswith('px'):
+            fs = fs[:-2].strip()
+        try:
+            int_fs = int(fs)
+        except ValueError:
+            int_fs = 20
+        self.yoffset = get_int_opt(options, 'yoffset', int_fs)
+        self.ystep = get_int_opt(options, 'ystep', int_fs + 5)
+        self.spacehack = get_bool_opt(options, 'spacehack', True)
+        self.linenos = get_bool_opt(options,'linenos',False)
+        self.linenostart = get_int_opt(options,'linenostart',1)
+        self.linenostep = get_int_opt(options,'linenostep',1)
+        self.linenowidth = get_int_opt(options,'linenowidth', 3*self.ystep)
+        self._stylecache = {}
+
+    def format_unencoded(self, tokensource, outfile):
+        """
+        Format ``tokensource``, an iterable of ``(tokentype, tokenstring)``
+        tuples and write it into ``outfile``.
+
+        For our implementation we put all lines in their own 'line group'.
+        """
+        x = self.xoffset
+        y = self.yoffset
+        if not self.nowrap:
+            if self.encoding:
+                outfile.write(f'<?xml version="1.0" encoding="{self.encoding}"?>\n')
+            else:
+                outfile.write('<?xml version="1.0"?>\n')
+            outfile.write('<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.0//EN" '
+                          '"http://www.w3.org/TR/2001/REC-SVG-20010904/DTD/'
+                          'svg10.dtd">\n')
+            outfile.write('<svg xmlns="http://www.w3.org/2000/svg">\n')
+            outfile.write(f'<g font-family="{self.fontfamily}" font-size="{self.fontsize}">\n')
+
+        counter = self.linenostart
+        counter_step = self.linenostep
+        counter_style = self._get_style(Comment)
+        line_x = x
+
+        if self.linenos:
+            if counter % counter_step == 0:
+                outfile.write(f'<text x="{x+self.linenowidth}" y="{y}" {counter_style} text-anchor="end">{counter}</text>')
+            line_x += self.linenowidth + self.ystep
+            counter += 1
+
+        outfile.write(f'<text x="{line_x}" y="{y}" xml:space="preserve">')
+        for ttype, value in tokensource:
+            style = self._get_style(ttype)
+            tspan = style and '<tspan' + style + '>' or ''
+            tspanend = tspan and '</tspan>' or ''
+            value = escape_html(value)
+            if self.spacehack:
+                value = value.expandtabs().replace(' ', '&#160;')
+            parts = value.split('\n')
+            for part in parts[:-1]:
+                outfile.write(tspan + part + tspanend)
+                y += self.ystep
+                outfile.write('</text>\n')
+                if self.linenos and counter % counter_step == 0:
+                    outfile.write(f'<text x="{x+self.linenowidth}" y="{y}" text-anchor="end" {counter_style}>{counter}</text>')
+
+                counter += 1
+                outfile.write(f'<text x="{line_x}" y="{y}" ' 'xml:space="preserve">')
+            outfile.write(tspan + parts[-1] + tspanend)
+        outfile.write('</text>')
+
+        if not self.nowrap:
+            outfile.write('</g></svg>\n')
+
+    def _get_style(self, tokentype):
+        if tokentype in self._stylecache:
+            return self._stylecache[tokentype]
+        otokentype = tokentype
+        while not self.style.styles_token(tokentype):
+            tokentype = tokentype.parent
+        value = self.style.style_for_token(tokentype)
+        result = ''
+        if value['color']:
+            result = ' fill="#' + value['color'] + '"'
+        if value['bold']:
+            result += ' font-weight="bold"'
+        if value['italic']:
+            result += ' font-style="italic"'
+        self._stylecache[otokentype] = result
+        return result
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal.py
new file mode 100644
index 00000000..51b902d3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal.py
@@ -0,0 +1,127 @@
+"""
+    pygments.formatters.terminal
+    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for terminal output with ANSI sequences.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.token import Keyword, Name, Comment, String, Error, \
+    Number, Operator, Generic, Token, Whitespace
+from pip._vendor.pygments.console import ansiformat
+from pip._vendor.pygments.util import get_choice_opt
+
+
+__all__ = ['TerminalFormatter']
+
+
+#: Map token types to a tuple of color values for light and dark
+#: backgrounds.
+TERMINAL_COLORS = {
+    Token:              ('',            ''),
+
+    Whitespace:         ('gray',   'brightblack'),
+    Comment:            ('gray',   'brightblack'),
+    Comment.Preproc:    ('cyan',        'brightcyan'),
+    Keyword:            ('blue',    'brightblue'),
+    Keyword.Type:       ('cyan',        'brightcyan'),
+    Operator.Word:      ('magenta',      'brightmagenta'),
+    Name.Builtin:       ('cyan',        'brightcyan'),
+    Name.Function:      ('green',   'brightgreen'),
+    Name.Namespace:     ('_cyan_',      '_brightcyan_'),
+    Name.Class:         ('_green_', '_brightgreen_'),
+    Name.Exception:     ('cyan',        'brightcyan'),
+    Name.Decorator:     ('brightblack',    'gray'),
+    Name.Variable:      ('red',     'brightred'),
+    Name.Constant:      ('red',     'brightred'),
+    Name.Attribute:     ('cyan',        'brightcyan'),
+    Name.Tag:           ('brightblue',        'brightblue'),
+    String:             ('yellow',       'yellow'),
+    Number:             ('blue',    'brightblue'),
+
+    Generic.Deleted:    ('brightred',        'brightred'),
+    Generic.Inserted:   ('green',  'brightgreen'),
+    Generic.Heading:    ('**',         '**'),
+    Generic.Subheading: ('*magenta*',   '*brightmagenta*'),
+    Generic.Prompt:     ('**',         '**'),
+    Generic.Error:      ('brightred',        'brightred'),
+
+    Error:              ('_brightred_',      '_brightred_'),
+}
+
+
+class TerminalFormatter(Formatter):
+    r"""
+    Format tokens with ANSI color sequences, for output in a text console.
+    Color sequences are terminated at newlines, so that paging the output
+    works correctly.
+
+    The `get_style_defs()` method doesn't do anything special since there is
+    no support for common styles.
+
+    Options accepted:
+
+    `bg`
+        Set to ``"light"`` or ``"dark"`` depending on the terminal's background
+        (default: ``"light"``).
+
+    `colorscheme`
+        A dictionary mapping token types to (lightbg, darkbg) color names or
+        ``None`` (default: ``None`` = use builtin colorscheme).
+
+    `linenos`
+        Set to ``True`` to have line numbers on the terminal output as well
+        (default: ``False`` = no line numbers).
+    """
+    name = 'Terminal'
+    aliases = ['terminal', 'console']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+        self.darkbg = get_choice_opt(options, 'bg',
+                                     ['light', 'dark'], 'light') == 'dark'
+        self.colorscheme = options.get('colorscheme', None) or TERMINAL_COLORS
+        self.linenos = options.get('linenos', False)
+        self._lineno = 0
+
+    def format(self, tokensource, outfile):
+        return Formatter.format(self, tokensource, outfile)
+
+    def _write_lineno(self, outfile):
+        self._lineno += 1
+        outfile.write("%s%04d: " % (self._lineno != 1 and '\n' or '', self._lineno))
+
+    def _get_color(self, ttype):
+        # self.colorscheme is a dict containing usually generic types, so we
+        # have to walk the tree of dots.  The base Token type must be a key,
+        # even if it's empty string, as in the default above.
+        colors = self.colorscheme.get(ttype)
+        while colors is None:
+            ttype = ttype.parent
+            colors = self.colorscheme.get(ttype)
+        return colors[self.darkbg]
+
+    def format_unencoded(self, tokensource, outfile):
+        if self.linenos:
+            self._write_lineno(outfile)
+
+        for ttype, value in tokensource:
+            color = self._get_color(ttype)
+
+            for line in value.splitlines(True):
+                if color:
+                    outfile.write(ansiformat(color, line.rstrip('\n')))
+                else:
+                    outfile.write(line.rstrip('\n'))
+                if line.endswith('\n'):
+                    if self.linenos:
+                        self._write_lineno(outfile)
+                    else:
+                        outfile.write('\n')
+
+        if self.linenos:
+            outfile.write("\n")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal256.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal256.py
new file mode 100644
index 00000000..5f254051
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/formatters/terminal256.py
@@ -0,0 +1,338 @@
+"""
+    pygments.formatters.terminal256
+    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+    Formatter for 256-color terminal output with ANSI sequences.
+
+    RGB-to-XTERM color conversion routines adapted from xterm256-conv
+    tool (http://frexx.de/xterm-256-notes/data/xterm256-conv2.tar.bz2)
+    by Wolfgang Frisch.
+
+    Formatter version 1.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+# TODO:
+#  - Options to map style's bold/underline/italic/border attributes
+#    to some ANSI attrbutes (something like 'italic=underline')
+#  - An option to output "style RGB to xterm RGB/index" conversion table
+#  - An option to indicate that we are running in "reverse background"
+#    xterm. This means that default colors are white-on-black, not
+#    black-on-while, so colors like "white background" need to be converted
+#    to "white background, black foreground", etc...
+
+from pip._vendor.pygments.formatter import Formatter
+from pip._vendor.pygments.console import codes
+from pip._vendor.pygments.style import ansicolors
+
+
+__all__ = ['Terminal256Formatter', 'TerminalTrueColorFormatter']
+
+
+class EscapeSequence:
+    def __init__(self, fg=None, bg=None, bold=False, underline=False, italic=False):
+        self.fg = fg
+        self.bg = bg
+        self.bold = bold
+        self.underline = underline
+        self.italic = italic
+
+    def escape(self, attrs):
+        if len(attrs):
+            return "\x1b[" + ";".join(attrs) + "m"
+        return ""
+
+    def color_string(self):
+        attrs = []
+        if self.fg is not None:
+            if self.fg in ansicolors:
+                esc = codes[self.fg.replace('ansi','')]
+                if ';01m' in esc:
+                    self.bold = True
+                # extract fg color code.
+                attrs.append(esc[2:4])
+            else:
+                attrs.extend(("38", "5", "%i" % self.fg))
+        if self.bg is not None:
+            if self.bg in ansicolors:
+                esc = codes[self.bg.replace('ansi','')]
+                # extract fg color code, add 10 for bg.
+                attrs.append(str(int(esc[2:4])+10))
+            else:
+                attrs.extend(("48", "5", "%i" % self.bg))
+        if self.bold:
+            attrs.append("01")
+        if self.underline:
+            attrs.append("04")
+        if self.italic:
+            attrs.append("03")
+        return self.escape(attrs)
+
+    def true_color_string(self):
+        attrs = []
+        if self.fg:
+            attrs.extend(("38", "2", str(self.fg[0]), str(self.fg[1]), str(self.fg[2])))
+        if self.bg:
+            attrs.extend(("48", "2", str(self.bg[0]), str(self.bg[1]), str(self.bg[2])))
+        if self.bold:
+            attrs.append("01")
+        if self.underline:
+            attrs.append("04")
+        if self.italic:
+            attrs.append("03")
+        return self.escape(attrs)
+
+    def reset_string(self):
+        attrs = []
+        if self.fg is not None:
+            attrs.append("39")
+        if self.bg is not None:
+            attrs.append("49")
+        if self.bold or self.underline or self.italic:
+            attrs.append("00")
+        return self.escape(attrs)
+
+
+class Terminal256Formatter(Formatter):
+    """
+    Format tokens with ANSI color sequences, for output in a 256-color
+    terminal or console.  Like in `TerminalFormatter` color sequences
+    are terminated at newlines, so that paging the output works correctly.
+
+    The formatter takes colors from a style defined by the `style` option
+    and converts them to nearest ANSI 256-color escape sequences. Bold and
+    underline attributes from the style are preserved (and displayed).
+
+    .. versionadded:: 0.9
+
+    .. versionchanged:: 2.2
+       If the used style defines foreground colors in the form ``#ansi*``, then
+       `Terminal256Formatter` will map these to non extended foreground color.
+       See :ref:`AnsiTerminalStyle` for more information.
+
+    .. versionchanged:: 2.4
+       The ANSI color names have been updated with names that are easier to
+       understand and align with colornames of other projects and terminals.
+       See :ref:`this table <new-ansi-color-names>` for more information.
+
+
+    Options accepted:
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+
+    `linenos`
+        Set to ``True`` to have line numbers on the terminal output as well
+        (default: ``False`` = no line numbers).
+    """
+    name = 'Terminal256'
+    aliases = ['terminal256', 'console256', '256']
+    filenames = []
+
+    def __init__(self, **options):
+        Formatter.__init__(self, **options)
+
+        self.xterm_colors = []
+        self.best_match = {}
+        self.style_string = {}
+
+        self.usebold = 'nobold' not in options
+        self.useunderline = 'nounderline' not in options
+        self.useitalic = 'noitalic' not in options
+
+        self._build_color_table()  # build an RGB-to-256 color conversion table
+        self._setup_styles()  # convert selected style's colors to term. colors
+
+        self.linenos = options.get('linenos', False)
+        self._lineno = 0
+
+    def _build_color_table(self):
+        # colors 0..15: 16 basic colors
+
+        self.xterm_colors.append((0x00, 0x00, 0x00))  # 0
+        self.xterm_colors.append((0xcd, 0x00, 0x00))  # 1
+        self.xterm_colors.append((0x00, 0xcd, 0x00))  # 2
+        self.xterm_colors.append((0xcd, 0xcd, 0x00))  # 3
+        self.xterm_colors.append((0x00, 0x00, 0xee))  # 4
+        self.xterm_colors.append((0xcd, 0x00, 0xcd))  # 5
+        self.xterm_colors.append((0x00, 0xcd, 0xcd))  # 6
+        self.xterm_colors.append((0xe5, 0xe5, 0xe5))  # 7
+        self.xterm_colors.append((0x7f, 0x7f, 0x7f))  # 8
+        self.xterm_colors.append((0xff, 0x00, 0x00))  # 9
+        self.xterm_colors.append((0x00, 0xff, 0x00))  # 10
+        self.xterm_colors.append((0xff, 0xff, 0x00))  # 11
+        self.xterm_colors.append((0x5c, 0x5c, 0xff))  # 12
+        self.xterm_colors.append((0xff, 0x00, 0xff))  # 13
+        self.xterm_colors.append((0x00, 0xff, 0xff))  # 14
+        self.xterm_colors.append((0xff, 0xff, 0xff))  # 15
+
+        # colors 16..232: the 6x6x6 color cube
+
+        valuerange = (0x00, 0x5f, 0x87, 0xaf, 0xd7, 0xff)
+
+        for i in range(217):
+            r = valuerange[(i // 36) % 6]
+            g = valuerange[(i // 6) % 6]
+            b = valuerange[i % 6]
+            self.xterm_colors.append((r, g, b))
+
+        # colors 233..253: grayscale
+
+        for i in range(1, 22):
+            v = 8 + i * 10
+            self.xterm_colors.append((v, v, v))
+
+    def _closest_color(self, r, g, b):
+        distance = 257*257*3  # "infinity" (>distance from #000000 to #ffffff)
+        match = 0
+
+        for i in range(0, 254):
+            values = self.xterm_colors[i]
+
+            rd = r - values[0]
+            gd = g - values[1]
+            bd = b - values[2]
+            d = rd*rd + gd*gd + bd*bd
+
+            if d < distance:
+                match = i
+                distance = d
+        return match
+
+    def _color_index(self, color):
+        index = self.best_match.get(color, None)
+        if color in ansicolors:
+            # strip the `ansi/#ansi` part and look up code
+            index = color
+            self.best_match[color] = index
+        if index is None:
+            try:
+                rgb = int(str(color), 16)
+            except ValueError:
+                rgb = 0
+
+            r = (rgb >> 16) & 0xff
+            g = (rgb >> 8) & 0xff
+            b = rgb & 0xff
+            index = self._closest_color(r, g, b)
+            self.best_match[color] = index
+        return index
+
+    def _setup_styles(self):
+        for ttype, ndef in self.style:
+            escape = EscapeSequence()
+            # get foreground from ansicolor if set
+            if ndef['ansicolor']:
+                escape.fg = self._color_index(ndef['ansicolor'])
+            elif ndef['color']:
+                escape.fg = self._color_index(ndef['color'])
+            if ndef['bgansicolor']:
+                escape.bg = self._color_index(ndef['bgansicolor'])
+            elif ndef['bgcolor']:
+                escape.bg = self._color_index(ndef['bgcolor'])
+            if self.usebold and ndef['bold']:
+                escape.bold = True
+            if self.useunderline and ndef['underline']:
+                escape.underline = True
+            if self.useitalic and ndef['italic']:
+                escape.italic = True
+            self.style_string[str(ttype)] = (escape.color_string(),
+                                             escape.reset_string())
+
+    def _write_lineno(self, outfile):
+        self._lineno += 1
+        outfile.write("%s%04d: " % (self._lineno != 1 and '\n' or '', self._lineno))
+
+    def format(self, tokensource, outfile):
+        return Formatter.format(self, tokensource, outfile)
+
+    def format_unencoded(self, tokensource, outfile):
+        if self.linenos:
+            self._write_lineno(outfile)
+
+        for ttype, value in tokensource:
+            not_found = True
+            while ttype and not_found:
+                try:
+                    # outfile.write( "<" + str(ttype) + ">" )
+                    on, off = self.style_string[str(ttype)]
+
+                    # Like TerminalFormatter, add "reset colors" escape sequence
+                    # on newline.
+                    spl = value.split('\n')
+                    for line in spl[:-1]:
+                        if line:
+                            outfile.write(on + line + off)
+                        if self.linenos:
+                            self._write_lineno(outfile)
+                        else:
+                            outfile.write('\n')
+
+                    if spl[-1]:
+                        outfile.write(on + spl[-1] + off)
+
+                    not_found = False
+                    # outfile.write( '#' + str(ttype) + '#' )
+
+                except KeyError:
+                    # ottype = ttype
+                    ttype = ttype.parent
+                    # outfile.write( '!' + str(ottype) + '->' + str(ttype) + '!' )
+
+            if not_found:
+                outfile.write(value)
+
+        if self.linenos:
+            outfile.write("\n")
+
+
+
+class TerminalTrueColorFormatter(Terminal256Formatter):
+    r"""
+    Format tokens with ANSI color sequences, for output in a true-color
+    terminal or console.  Like in `TerminalFormatter` color sequences
+    are terminated at newlines, so that paging the output works correctly.
+
+    .. versionadded:: 2.1
+
+    Options accepted:
+
+    `style`
+        The style to use, can be a string or a Style subclass (default:
+        ``'default'``).
+    """
+    name = 'TerminalTrueColor'
+    aliases = ['terminal16m', 'console16m', '16m']
+    filenames = []
+
+    def _build_color_table(self):
+        pass
+
+    def _color_tuple(self, color):
+        try:
+            rgb = int(str(color), 16)
+        except ValueError:
+            return None
+        r = (rgb >> 16) & 0xff
+        g = (rgb >> 8) & 0xff
+        b = rgb & 0xff
+        return (r, g, b)
+
+    def _setup_styles(self):
+        for ttype, ndef in self.style:
+            escape = EscapeSequence()
+            if ndef['color']:
+                escape.fg = self._color_tuple(ndef['color'])
+            if ndef['bgcolor']:
+                escape.bg = self._color_tuple(ndef['bgcolor'])
+            if self.usebold and ndef['bold']:
+                escape.bold = True
+            if self.useunderline and ndef['underline']:
+                escape.underline = True
+            if self.useitalic and ndef['italic']:
+                escape.italic = True
+            self.style_string[str(ttype)] = (escape.true_color_string(),
+                                             escape.reset_string())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexer.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexer.py
new file mode 100644
index 00000000..1348be58
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexer.py
@@ -0,0 +1,963 @@
+"""
+    pygments.lexer
+    ~~~~~~~~~~~~~~
+
+    Base lexer classes.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+import sys
+import time
+
+from pip._vendor.pygments.filter import apply_filters, Filter
+from pip._vendor.pygments.filters import get_filter_by_name
+from pip._vendor.pygments.token import Error, Text, Other, Whitespace, _TokenType
+from pip._vendor.pygments.util import get_bool_opt, get_int_opt, get_list_opt, \
+    make_analysator, Future, guess_decode
+from pip._vendor.pygments.regexopt import regex_opt
+
+__all__ = ['Lexer', 'RegexLexer', 'ExtendedRegexLexer', 'DelegatingLexer',
+           'LexerContext', 'include', 'inherit', 'bygroups', 'using', 'this',
+           'default', 'words', 'line_re']
+
+line_re = re.compile('.*?\n')
+
+_encoding_map = [(b'\xef\xbb\xbf', 'utf-8'),
+                 (b'\xff\xfe\0\0', 'utf-32'),
+                 (b'\0\0\xfe\xff', 'utf-32be'),
+                 (b'\xff\xfe', 'utf-16'),
+                 (b'\xfe\xff', 'utf-16be')]
+
+_default_analyse = staticmethod(lambda x: 0.0)
+
+
+class LexerMeta(type):
+    """
+    This metaclass automagically converts ``analyse_text`` methods into
+    static methods which always return float values.
+    """
+
+    def __new__(mcs, name, bases, d):
+        if 'analyse_text' in d:
+            d['analyse_text'] = make_analysator(d['analyse_text'])
+        return type.__new__(mcs, name, bases, d)
+
+
+class Lexer(metaclass=LexerMeta):
+    """
+    Lexer for a specific language.
+
+    See also :doc:`lexerdevelopment`, a high-level guide to writing
+    lexers.
+
+    Lexer classes have attributes used for choosing the most appropriate
+    lexer based on various criteria.
+
+    .. autoattribute:: name
+       :no-value:
+    .. autoattribute:: aliases
+       :no-value:
+    .. autoattribute:: filenames
+       :no-value:
+    .. autoattribute:: alias_filenames
+    .. autoattribute:: mimetypes
+       :no-value:
+    .. autoattribute:: priority
+
+    Lexers included in Pygments should have two additional attributes:
+
+    .. autoattribute:: url
+       :no-value:
+    .. autoattribute:: version_added
+       :no-value:
+
+    Lexers included in Pygments may have additional attributes:
+
+    .. autoattribute:: _example
+       :no-value:
+
+    You can pass options to the constructor. The basic options recognized
+    by all lexers and processed by the base `Lexer` class are:
+
+    ``stripnl``
+        Strip leading and trailing newlines from the input (default: True).
+    ``stripall``
+        Strip all leading and trailing whitespace from the input
+        (default: False).
+    ``ensurenl``
+        Make sure that the input ends with a newline (default: True).  This
+        is required for some lexers that consume input linewise.
+
+        .. versionadded:: 1.3
+
+    ``tabsize``
+        If given and greater than 0, expand tabs in the input (default: 0).
+    ``encoding``
+        If given, must be an encoding name. This encoding will be used to
+        convert the input string to Unicode, if it is not already a Unicode
+        string (default: ``'guess'``, which uses a simple UTF-8 / Locale /
+        Latin1 detection.  Can also be ``'chardet'`` to use the chardet
+        library, if it is installed.
+    ``inencoding``
+        Overrides the ``encoding`` if given.
+    """
+
+    #: Full name of the lexer, in human-readable form
+    name = None
+
+    #: A list of short, unique identifiers that can be used to look
+    #: up the lexer from a list, e.g., using `get_lexer_by_name()`.
+    aliases = []
+
+    #: A list of `fnmatch` patterns that match filenames which contain
+    #: content for this lexer. The patterns in this list should be unique among
+    #: all lexers.
+    filenames = []
+
+    #: A list of `fnmatch` patterns that match filenames which may or may not
+    #: contain content for this lexer. This list is used by the
+    #: :func:`.guess_lexer_for_filename()` function, to determine which lexers
+    #: are then included in guessing the correct one. That means that
+    #: e.g. every lexer for HTML and a template language should include
+    #: ``\*.html`` in this list.
+    alias_filenames = []
+
+    #: A list of MIME types for content that can be lexed with this lexer.
+    mimetypes = []
+
+    #: Priority, should multiple lexers match and no content is provided
+    priority = 0
+
+    #: URL of the language specification/definition. Used in the Pygments
+    #: documentation. Set to an empty string to disable.
+    url = None
+
+    #: Version of Pygments in which the lexer was added.
+    version_added = None
+
+    #: Example file name. Relative to the ``tests/examplefiles`` directory.
+    #: This is used by the documentation generator to show an example.
+    _example = None
+
+    def __init__(self, **options):
+        """
+        This constructor takes arbitrary options as keyword arguments.
+        Every subclass must first process its own options and then call
+        the `Lexer` constructor, since it processes the basic
+        options like `stripnl`.
+
+        An example looks like this:
+
+        .. sourcecode:: python
+
+           def __init__(self, **options):
+               self.compress = options.get('compress', '')
+               Lexer.__init__(self, **options)
+
+        As these options must all be specifiable as strings (due to the
+        command line usage), there are various utility functions
+        available to help with that, see `Utilities`_.
+        """
+        self.options = options
+        self.stripnl = get_bool_opt(options, 'stripnl', True)
+        self.stripall = get_bool_opt(options, 'stripall', False)
+        self.ensurenl = get_bool_opt(options, 'ensurenl', True)
+        self.tabsize = get_int_opt(options, 'tabsize', 0)
+        self.encoding = options.get('encoding', 'guess')
+        self.encoding = options.get('inencoding') or self.encoding
+        self.filters = []
+        for filter_ in get_list_opt(options, 'filters', ()):
+            self.add_filter(filter_)
+
+    def __repr__(self):
+        if self.options:
+            return f'<pygments.lexers.{self.__class__.__name__} with {self.options!r}>'
+        else:
+            return f'<pygments.lexers.{self.__class__.__name__}>'
+
+    def add_filter(self, filter_, **options):
+        """
+        Add a new stream filter to this lexer.
+        """
+        if not isinstance(filter_, Filter):
+            filter_ = get_filter_by_name(filter_, **options)
+        self.filters.append(filter_)
+
+    def analyse_text(text):
+        """
+        A static method which is called for lexer guessing.
+
+        It should analyse the text and return a float in the range
+        from ``0.0`` to ``1.0``.  If it returns ``0.0``, the lexer
+        will not be selected as the most probable one, if it returns
+        ``1.0``, it will be selected immediately.  This is used by
+        `guess_lexer`.
+
+        The `LexerMeta` metaclass automatically wraps this function so
+        that it works like a static method (no ``self`` or ``cls``
+        parameter) and the return value is automatically converted to
+        `float`. If the return value is an object that is boolean `False`
+        it's the same as if the return values was ``0.0``.
+        """
+
+    def _preprocess_lexer_input(self, text):
+        """Apply preprocessing such as decoding the input, removing BOM and normalizing newlines."""
+
+        if not isinstance(text, str):
+            if self.encoding == 'guess':
+                text, _ = guess_decode(text)
+            elif self.encoding == 'chardet':
+                try:
+                    # pip vendoring note: this code is not reachable by pip,
+                    # removed import of chardet to make it clear.
+                    raise ImportError('chardet is not vendored by pip')
+                except ImportError as e:
+                    raise ImportError('To enable chardet encoding guessing, '
+                                      'please install the chardet library '
+                                      'from http://chardet.feedparser.org/') from e
+                # check for BOM first
+                decoded = None
+                for bom, encoding in _encoding_map:
+                    if text.startswith(bom):
+                        decoded = text[len(bom):].decode(encoding, 'replace')
+                        break
+                # no BOM found, so use chardet
+                if decoded is None:
+                    enc = chardet.detect(text[:1024])  # Guess using first 1KB
+                    decoded = text.decode(enc.get('encoding') or 'utf-8',
+                                          'replace')
+                text = decoded
+            else:
+                text = text.decode(self.encoding)
+                if text.startswith('\ufeff'):
+                    text = text[len('\ufeff'):]
+        else:
+            if text.startswith('\ufeff'):
+                text = text[len('\ufeff'):]
+
+        # text now *is* a unicode string
+        text = text.replace('\r\n', '\n')
+        text = text.replace('\r', '\n')
+        if self.stripall:
+            text = text.strip()
+        elif self.stripnl:
+            text = text.strip('\n')
+        if self.tabsize > 0:
+            text = text.expandtabs(self.tabsize)
+        if self.ensurenl and not text.endswith('\n'):
+            text += '\n'
+
+        return text
+
+    def get_tokens(self, text, unfiltered=False):
+        """
+        This method is the basic interface of a lexer. It is called by
+        the `highlight()` function. It must process the text and return an
+        iterable of ``(tokentype, value)`` pairs from `text`.
+
+        Normally, you don't need to override this method. The default
+        implementation processes the options recognized by all lexers
+        (`stripnl`, `stripall` and so on), and then yields all tokens
+        from `get_tokens_unprocessed()`, with the ``index`` dropped.
+
+        If `unfiltered` is set to `True`, the filtering mechanism is
+        bypassed even if filters are defined.
+        """
+        text = self._preprocess_lexer_input(text)
+
+        def streamer():
+            for _, t, v in self.get_tokens_unprocessed(text):
+                yield t, v
+        stream = streamer()
+        if not unfiltered:
+            stream = apply_filters(stream, self.filters, self)
+        return stream
+
+    def get_tokens_unprocessed(self, text):
+        """
+        This method should process the text and return an iterable of
+        ``(index, tokentype, value)`` tuples where ``index`` is the starting
+        position of the token within the input text.
+
+        It must be overridden by subclasses. It is recommended to
+        implement it as a generator to maximize effectiveness.
+        """
+        raise NotImplementedError
+
+
+class DelegatingLexer(Lexer):
+    """
+    This lexer takes two lexer as arguments. A root lexer and
+    a language lexer. First everything is scanned using the language
+    lexer, afterwards all ``Other`` tokens are lexed using the root
+    lexer.
+
+    The lexers from the ``template`` lexer package use this base lexer.
+    """
+
+    def __init__(self, _root_lexer, _language_lexer, _needle=Other, **options):
+        self.root_lexer = _root_lexer(**options)
+        self.language_lexer = _language_lexer(**options)
+        self.needle = _needle
+        Lexer.__init__(self, **options)
+
+    def get_tokens_unprocessed(self, text):
+        buffered = ''
+        insertions = []
+        lng_buffer = []
+        for i, t, v in self.language_lexer.get_tokens_unprocessed(text):
+            if t is self.needle:
+                if lng_buffer:
+                    insertions.append((len(buffered), lng_buffer))
+                    lng_buffer = []
+                buffered += v
+            else:
+                lng_buffer.append((i, t, v))
+        if lng_buffer:
+            insertions.append((len(buffered), lng_buffer))
+        return do_insertions(insertions,
+                             self.root_lexer.get_tokens_unprocessed(buffered))
+
+
+# ------------------------------------------------------------------------------
+# RegexLexer and ExtendedRegexLexer
+#
+
+
+class include(str):  # pylint: disable=invalid-name
+    """
+    Indicates that a state should include rules from another state.
+    """
+    pass
+
+
+class _inherit:
+    """
+    Indicates the a state should inherit from its superclass.
+    """
+    def __repr__(self):
+        return 'inherit'
+
+inherit = _inherit()  # pylint: disable=invalid-name
+
+
+class combined(tuple):  # pylint: disable=invalid-name
+    """
+    Indicates a state combined from multiple states.
+    """
+
+    def __new__(cls, *args):
+        return tuple.__new__(cls, args)
+
+    def __init__(self, *args):
+        # tuple.__init__ doesn't do anything
+        pass
+
+
+class _PseudoMatch:
+    """
+    A pseudo match object constructed from a string.
+    """
+
+    def __init__(self, start, text):
+        self._text = text
+        self._start = start
+
+    def start(self, arg=None):
+        return self._start
+
+    def end(self, arg=None):
+        return self._start + len(self._text)
+
+    def group(self, arg=None):
+        if arg:
+            raise IndexError('No such group')
+        return self._text
+
+    def groups(self):
+        return (self._text,)
+
+    def groupdict(self):
+        return {}
+
+
+def bygroups(*args):
+    """
+    Callback that yields multiple actions for each group in the match.
+    """
+    def callback(lexer, match, ctx=None):
+        for i, action in enumerate(args):
+            if action is None:
+                continue
+            elif type(action) is _TokenType:
+                data = match.group(i + 1)
+                if data:
+                    yield match.start(i + 1), action, data
+            else:
+                data = match.group(i + 1)
+                if data is not None:
+                    if ctx:
+                        ctx.pos = match.start(i + 1)
+                    for item in action(lexer,
+                                       _PseudoMatch(match.start(i + 1), data), ctx):
+                        if item:
+                            yield item
+        if ctx:
+            ctx.pos = match.end()
+    return callback
+
+
+class _This:
+    """
+    Special singleton used for indicating the caller class.
+    Used by ``using``.
+    """
+
+this = _This()
+
+
+def using(_other, **kwargs):
+    """
+    Callback that processes the match with a different lexer.
+
+    The keyword arguments are forwarded to the lexer, except `state` which
+    is handled separately.
+
+    `state` specifies the state that the new lexer will start in, and can
+    be an enumerable such as ('root', 'inline', 'string') or a simple
+    string which is assumed to be on top of the root state.
+
+    Note: For that to work, `_other` must not be an `ExtendedRegexLexer`.
+    """
+    gt_kwargs = {}
+    if 'state' in kwargs:
+        s = kwargs.pop('state')
+        if isinstance(s, (list, tuple)):
+            gt_kwargs['stack'] = s
+        else:
+            gt_kwargs['stack'] = ('root', s)
+
+    if _other is this:
+        def callback(lexer, match, ctx=None):
+            # if keyword arguments are given the callback
+            # function has to create a new lexer instance
+            if kwargs:
+                # XXX: cache that somehow
+                kwargs.update(lexer.options)
+                lx = lexer.__class__(**kwargs)
+            else:
+                lx = lexer
+            s = match.start()
+            for i, t, v in lx.get_tokens_unprocessed(match.group(), **gt_kwargs):
+                yield i + s, t, v
+            if ctx:
+                ctx.pos = match.end()
+    else:
+        def callback(lexer, match, ctx=None):
+            # XXX: cache that somehow
+            kwargs.update(lexer.options)
+            lx = _other(**kwargs)
+
+            s = match.start()
+            for i, t, v in lx.get_tokens_unprocessed(match.group(), **gt_kwargs):
+                yield i + s, t, v
+            if ctx:
+                ctx.pos = match.end()
+    return callback
+
+
+class default:
+    """
+    Indicates a state or state action (e.g. #pop) to apply.
+    For example default('#pop') is equivalent to ('', Token, '#pop')
+    Note that state tuples may be used as well.
+
+    .. versionadded:: 2.0
+    """
+    def __init__(self, state):
+        self.state = state
+
+
+class words(Future):
+    """
+    Indicates a list of literal words that is transformed into an optimized
+    regex that matches any of the words.
+
+    .. versionadded:: 2.0
+    """
+    def __init__(self, words, prefix='', suffix=''):
+        self.words = words
+        self.prefix = prefix
+        self.suffix = suffix
+
+    def get(self):
+        return regex_opt(self.words, prefix=self.prefix, suffix=self.suffix)
+
+
+class RegexLexerMeta(LexerMeta):
+    """
+    Metaclass for RegexLexer, creates the self._tokens attribute from
+    self.tokens on the first instantiation.
+    """
+
+    def _process_regex(cls, regex, rflags, state):
+        """Preprocess the regular expression component of a token definition."""
+        if isinstance(regex, Future):
+            regex = regex.get()
+        return re.compile(regex, rflags).match
+
+    def _process_token(cls, token):
+        """Preprocess the token component of a token definition."""
+        assert type(token) is _TokenType or callable(token), \
+            f'token type must be simple type or callable, not {token!r}'
+        return token
+
+    def _process_new_state(cls, new_state, unprocessed, processed):
+        """Preprocess the state transition action of a token definition."""
+        if isinstance(new_state, str):
+            # an existing state
+            if new_state == '#pop':
+                return -1
+            elif new_state in unprocessed:
+                return (new_state,)
+            elif new_state == '#push':
+                return new_state
+            elif new_state[:5] == '#pop:':
+                return -int(new_state[5:])
+            else:
+                assert False, f'unknown new state {new_state!r}'
+        elif isinstance(new_state, combined):
+            # combine a new state from existing ones
+            tmp_state = '_tmp_%d' % cls._tmpname
+            cls._tmpname += 1
+            itokens = []
+            for istate in new_state:
+                assert istate != new_state, f'circular state ref {istate!r}'
+                itokens.extend(cls._process_state(unprocessed,
+                                                  processed, istate))
+            processed[tmp_state] = itokens
+            return (tmp_state,)
+        elif isinstance(new_state, tuple):
+            # push more than one state
+            for istate in new_state:
+                assert (istate in unprocessed or
+                        istate in ('#pop', '#push')), \
+                    'unknown new state ' + istate
+            return new_state
+        else:
+            assert False, f'unknown new state def {new_state!r}'
+
+    def _process_state(cls, unprocessed, processed, state):
+        """Preprocess a single state definition."""
+        assert isinstance(state, str), f"wrong state name {state!r}"
+        assert state[0] != '#', f"invalid state name {state!r}"
+        if state in processed:
+            return processed[state]
+        tokens = processed[state] = []
+        rflags = cls.flags
+        for tdef in unprocessed[state]:
+            if isinstance(tdef, include):
+                # it's a state reference
+                assert tdef != state, f"circular state reference {state!r}"
+                tokens.extend(cls._process_state(unprocessed, processed,
+                                                 str(tdef)))
+                continue
+            if isinstance(tdef, _inherit):
+                # should be processed already, but may not in the case of:
+                # 1. the state has no counterpart in any parent
+                # 2. the state includes more than one 'inherit'
+                continue
+            if isinstance(tdef, default):
+                new_state = cls._process_new_state(tdef.state, unprocessed, processed)
+                tokens.append((re.compile('').match, None, new_state))
+                continue
+
+            assert type(tdef) is tuple, f"wrong rule def {tdef!r}"
+
+            try:
+                rex = cls._process_regex(tdef[0], rflags, state)
+            except Exception as err:
+                raise ValueError(f"uncompilable regex {tdef[0]!r} in state {state!r} of {cls!r}: {err}") from err
+
+            token = cls._process_token(tdef[1])
+
+            if len(tdef) == 2:
+                new_state = None
+            else:
+                new_state = cls._process_new_state(tdef[2],
+                                                   unprocessed, processed)
+
+            tokens.append((rex, token, new_state))
+        return tokens
+
+    def process_tokendef(cls, name, tokendefs=None):
+        """Preprocess a dictionary of token definitions."""
+        processed = cls._all_tokens[name] = {}
+        tokendefs = tokendefs or cls.tokens[name]
+        for state in list(tokendefs):
+            cls._process_state(tokendefs, processed, state)
+        return processed
+
+    def get_tokendefs(cls):
+        """
+        Merge tokens from superclasses in MRO order, returning a single tokendef
+        dictionary.
+
+        Any state that is not defined by a subclass will be inherited
+        automatically.  States that *are* defined by subclasses will, by
+        default, override that state in the superclass.  If a subclass wishes to
+        inherit definitions from a superclass, it can use the special value
+        "inherit", which will cause the superclass' state definition to be
+        included at that point in the state.
+        """
+        tokens = {}
+        inheritable = {}
+        for c in cls.__mro__:
+            toks = c.__dict__.get('tokens', {})
+
+            for state, items in toks.items():
+                curitems = tokens.get(state)
+                if curitems is None:
+                    # N.b. because this is assigned by reference, sufficiently
+                    # deep hierarchies are processed incrementally (e.g. for
+                    # A(B), B(C), C(RegexLexer), B will be premodified so X(B)
+                    # will not see any inherits in B).
+                    tokens[state] = items
+                    try:
+                        inherit_ndx = items.index(inherit)
+                    except ValueError:
+                        continue
+                    inheritable[state] = inherit_ndx
+                    continue
+
+                inherit_ndx = inheritable.pop(state, None)
+                if inherit_ndx is None:
+                    continue
+
+                # Replace the "inherit" value with the items
+                curitems[inherit_ndx:inherit_ndx+1] = items
+                try:
+                    # N.b. this is the index in items (that is, the superclass
+                    # copy), so offset required when storing below.
+                    new_inh_ndx = items.index(inherit)
+                except ValueError:
+                    pass
+                else:
+                    inheritable[state] = inherit_ndx + new_inh_ndx
+
+        return tokens
+
+    def __call__(cls, *args, **kwds):
+        """Instantiate cls after preprocessing its token definitions."""
+        if '_tokens' not in cls.__dict__:
+            cls._all_tokens = {}
+            cls._tmpname = 0
+            if hasattr(cls, 'token_variants') and cls.token_variants:
+                # don't process yet
+                pass
+            else:
+                cls._tokens = cls.process_tokendef('', cls.get_tokendefs())
+
+        return type.__call__(cls, *args, **kwds)
+
+
+class RegexLexer(Lexer, metaclass=RegexLexerMeta):
+    """
+    Base for simple stateful regular expression-based lexers.
+    Simplifies the lexing process so that you need only
+    provide a list of states and regular expressions.
+    """
+
+    #: Flags for compiling the regular expressions.
+    #: Defaults to MULTILINE.
+    flags = re.MULTILINE
+
+    #: At all time there is a stack of states. Initially, the stack contains
+    #: a single state 'root'. The top of the stack is called "the current state".
+    #:
+    #: Dict of ``{'state': [(regex, tokentype, new_state), ...], ...}``
+    #:
+    #: ``new_state`` can be omitted to signify no state transition.
+    #: If ``new_state`` is a string, it is pushed on the stack. This ensure
+    #: the new current state is ``new_state``.
+    #: If ``new_state`` is a tuple of strings, all of those strings are pushed
+    #: on the stack and the current state will be the last element of the list.
+    #: ``new_state`` can also be ``combined('state1', 'state2', ...)``
+    #: to signify a new, anonymous state combined from the rules of two
+    #: or more existing ones.
+    #: Furthermore, it can be '#pop' to signify going back one step in
+    #: the state stack, or '#push' to push the current state on the stack
+    #: again. Note that if you push while in a combined state, the combined
+    #: state itself is pushed, and not only the state in which the rule is
+    #: defined.
+    #:
+    #: The tuple can also be replaced with ``include('state')``, in which
+    #: case the rules from the state named by the string are included in the
+    #: current one.
+    tokens = {}
+
+    def get_tokens_unprocessed(self, text, stack=('root',)):
+        """
+        Split ``text`` into (tokentype, text) pairs.
+
+        ``stack`` is the initial stack (default: ``['root']``)
+        """
+        pos = 0
+        tokendefs = self._tokens
+        statestack = list(stack)
+        statetokens = tokendefs[statestack[-1]]
+        while 1:
+            for rexmatch, action, new_state in statetokens:
+                m = rexmatch(text, pos)
+                if m:
+                    if action is not None:
+                        if type(action) is _TokenType:
+                            yield pos, action, m.group()
+                        else:
+                            yield from action(self, m)
+                    pos = m.end()
+                    if new_state is not None:
+                        # state transition
+                        if isinstance(new_state, tuple):
+                            for state in new_state:
+                                if state == '#pop':
+                                    if len(statestack) > 1:
+                                        statestack.pop()
+                                elif state == '#push':
+                                    statestack.append(statestack[-1])
+                                else:
+                                    statestack.append(state)
+                        elif isinstance(new_state, int):
+                            # pop, but keep at least one state on the stack
+                            # (random code leading to unexpected pops should
+                            # not allow exceptions)
+                            if abs(new_state) >= len(statestack):
+                                del statestack[1:]
+                            else:
+                                del statestack[new_state:]
+                        elif new_state == '#push':
+                            statestack.append(statestack[-1])
+                        else:
+                            assert False, f"wrong state def: {new_state!r}"
+                        statetokens = tokendefs[statestack[-1]]
+                    break
+            else:
+                # We are here only if all state tokens have been considered
+                # and there was not a match on any of them.
+                try:
+                    if text[pos] == '\n':
+                        # at EOL, reset state to "root"
+                        statestack = ['root']
+                        statetokens = tokendefs['root']
+                        yield pos, Whitespace, '\n'
+                        pos += 1
+                        continue
+                    yield pos, Error, text[pos]
+                    pos += 1
+                except IndexError:
+                    break
+
+
+class LexerContext:
+    """
+    A helper object that holds lexer position data.
+    """
+
+    def __init__(self, text, pos, stack=None, end=None):
+        self.text = text
+        self.pos = pos
+        self.end = end or len(text)  # end=0 not supported ;-)
+        self.stack = stack or ['root']
+
+    def __repr__(self):
+        return f'LexerContext({self.text!r}, {self.pos!r}, {self.stack!r})'
+
+
+class ExtendedRegexLexer(RegexLexer):
+    """
+    A RegexLexer that uses a context object to store its state.
+    """
+
+    def get_tokens_unprocessed(self, text=None, context=None):
+        """
+        Split ``text`` into (tokentype, text) pairs.
+        If ``context`` is given, use this lexer context instead.
+        """
+        tokendefs = self._tokens
+        if not context:
+            ctx = LexerContext(text, 0)
+            statetokens = tokendefs['root']
+        else:
+            ctx = context
+            statetokens = tokendefs[ctx.stack[-1]]
+            text = ctx.text
+        while 1:
+            for rexmatch, action, new_state in statetokens:
+                m = rexmatch(text, ctx.pos, ctx.end)
+                if m:
+                    if action is not None:
+                        if type(action) is _TokenType:
+                            yield ctx.pos, action, m.group()
+                            ctx.pos = m.end()
+                        else:
+                            yield from action(self, m, ctx)
+                            if not new_state:
+                                # altered the state stack?
+                                statetokens = tokendefs[ctx.stack[-1]]
+                    # CAUTION: callback must set ctx.pos!
+                    if new_state is not None:
+                        # state transition
+                        if isinstance(new_state, tuple):
+                            for state in new_state:
+                                if state == '#pop':
+                                    if len(ctx.stack) > 1:
+                                        ctx.stack.pop()
+                                elif state == '#push':
+                                    ctx.stack.append(ctx.stack[-1])
+                                else:
+                                    ctx.stack.append(state)
+                        elif isinstance(new_state, int):
+                            # see RegexLexer for why this check is made
+                            if abs(new_state) >= len(ctx.stack):
+                                del ctx.stack[1:]
+                            else:
+                                del ctx.stack[new_state:]
+                        elif new_state == '#push':
+                            ctx.stack.append(ctx.stack[-1])
+                        else:
+                            assert False, f"wrong state def: {new_state!r}"
+                        statetokens = tokendefs[ctx.stack[-1]]
+                    break
+            else:
+                try:
+                    if ctx.pos >= ctx.end:
+                        break
+                    if text[ctx.pos] == '\n':
+                        # at EOL, reset state to "root"
+                        ctx.stack = ['root']
+                        statetokens = tokendefs['root']
+                        yield ctx.pos, Text, '\n'
+                        ctx.pos += 1
+                        continue
+                    yield ctx.pos, Error, text[ctx.pos]
+                    ctx.pos += 1
+                except IndexError:
+                    break
+
+
+def do_insertions(insertions, tokens):
+    """
+    Helper for lexers which must combine the results of several
+    sublexers.
+
+    ``insertions`` is a list of ``(index, itokens)`` pairs.
+    Each ``itokens`` iterable should be inserted at position
+    ``index`` into the token stream given by the ``tokens``
+    argument.
+
+    The result is a combined token stream.
+
+    TODO: clean up the code here.
+    """
+    insertions = iter(insertions)
+    try:
+        index, itokens = next(insertions)
+    except StopIteration:
+        # no insertions
+        yield from tokens
+        return
+
+    realpos = None
+    insleft = True
+
+    # iterate over the token stream where we want to insert
+    # the tokens from the insertion list.
+    for i, t, v in tokens:
+        # first iteration. store the position of first item
+        if realpos is None:
+            realpos = i
+        oldi = 0
+        while insleft and i + len(v) >= index:
+            tmpval = v[oldi:index - i]
+            if tmpval:
+                yield realpos, t, tmpval
+                realpos += len(tmpval)
+            for it_index, it_token, it_value in itokens:
+                yield realpos, it_token, it_value
+                realpos += len(it_value)
+            oldi = index - i
+            try:
+                index, itokens = next(insertions)
+            except StopIteration:
+                insleft = False
+                break  # not strictly necessary
+        if oldi < len(v):
+            yield realpos, t, v[oldi:]
+            realpos += len(v) - oldi
+
+    # leftover tokens
+    while insleft:
+        # no normal tokens, set realpos to zero
+        realpos = realpos or 0
+        for p, t, v in itokens:
+            yield realpos, t, v
+            realpos += len(v)
+        try:
+            index, itokens = next(insertions)
+        except StopIteration:
+            insleft = False
+            break  # not strictly necessary
+
+
+class ProfilingRegexLexerMeta(RegexLexerMeta):
+    """Metaclass for ProfilingRegexLexer, collects regex timing info."""
+
+    def _process_regex(cls, regex, rflags, state):
+        if isinstance(regex, words):
+            rex = regex_opt(regex.words, prefix=regex.prefix,
+                            suffix=regex.suffix)
+        else:
+            rex = regex
+        compiled = re.compile(rex, rflags)
+
+        def match_func(text, pos, endpos=sys.maxsize):
+            info = cls._prof_data[-1].setdefault((state, rex), [0, 0.0])
+            t0 = time.time()
+            res = compiled.match(text, pos, endpos)
+            t1 = time.time()
+            info[0] += 1
+            info[1] += t1 - t0
+            return res
+        return match_func
+
+
+class ProfilingRegexLexer(RegexLexer, metaclass=ProfilingRegexLexerMeta):
+    """Drop-in replacement for RegexLexer that does profiling of its regexes."""
+
+    _prof_data = []
+    _prof_sort_index = 4  # defaults to time per call
+
+    def get_tokens_unprocessed(self, text, stack=('root',)):
+        # this needs to be a stack, since using(this) will produce nested calls
+        self.__class__._prof_data.append({})
+        yield from RegexLexer.get_tokens_unprocessed(self, text, stack)
+        rawdata = self.__class__._prof_data.pop()
+        data = sorted(((s, repr(r).strip('u\'').replace('\\\\', '\\')[:65],
+                        n, 1000 * t, 1000 * t / n)
+                       for ((s, r), (n, t)) in rawdata.items()),
+                      key=lambda x: x[self._prof_sort_index],
+                      reverse=True)
+        sum_total = sum(x[3] for x in data)
+
+        print()
+        print('Profiling result for %s lexing %d chars in %.3f ms' %
+              (self.__class__.__name__, len(text), sum_total))
+        print('=' * 110)
+        print('%-20s %-64s ncalls  tottime  percall' % ('state', 'regex'))
+        print('-' * 110)
+        for d in data:
+            print('%-20s %-65s %5d %8.4f %8.4f' % d)
+        print('=' * 110)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__init__.py
new file mode 100644
index 00000000..ac88645a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__init__.py
@@ -0,0 +1,362 @@
+"""
+    pygments.lexers
+    ~~~~~~~~~~~~~~~
+
+    Pygments lexers.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+import sys
+import types
+import fnmatch
+from os.path import basename
+
+from pip._vendor.pygments.lexers._mapping import LEXERS
+from pip._vendor.pygments.modeline import get_filetype_from_buffer
+from pip._vendor.pygments.plugin import find_plugin_lexers
+from pip._vendor.pygments.util import ClassNotFound, guess_decode
+
+COMPAT = {
+    'Python3Lexer': 'PythonLexer',
+    'Python3TracebackLexer': 'PythonTracebackLexer',
+    'LeanLexer': 'Lean3Lexer',
+}
+
+__all__ = ['get_lexer_by_name', 'get_lexer_for_filename', 'find_lexer_class',
+           'guess_lexer', 'load_lexer_from_file'] + list(LEXERS) + list(COMPAT)
+
+_lexer_cache = {}
+_pattern_cache = {}
+
+
+def _fn_matches(fn, glob):
+    """Return whether the supplied file name fn matches pattern filename."""
+    if glob not in _pattern_cache:
+        pattern = _pattern_cache[glob] = re.compile(fnmatch.translate(glob))
+        return pattern.match(fn)
+    return _pattern_cache[glob].match(fn)
+
+
+def _load_lexers(module_name):
+    """Load a lexer (and all others in the module too)."""
+    mod = __import__(module_name, None, None, ['__all__'])
+    for lexer_name in mod.__all__:
+        cls = getattr(mod, lexer_name)
+        _lexer_cache[cls.name] = cls
+
+
+def get_all_lexers(plugins=True):
+    """Return a generator of tuples in the form ``(name, aliases,
+    filenames, mimetypes)`` of all know lexers.
+
+    If *plugins* is true (the default), plugin lexers supplied by entrypoints
+    are also returned.  Otherwise, only builtin ones are considered.
+    """
+    for item in LEXERS.values():
+        yield item[1:]
+    if plugins:
+        for lexer in find_plugin_lexers():
+            yield lexer.name, lexer.aliases, lexer.filenames, lexer.mimetypes
+
+
+def find_lexer_class(name):
+    """
+    Return the `Lexer` subclass that with the *name* attribute as given by
+    the *name* argument.
+    """
+    if name in _lexer_cache:
+        return _lexer_cache[name]
+    # lookup builtin lexers
+    for module_name, lname, aliases, _, _ in LEXERS.values():
+        if name == lname:
+            _load_lexers(module_name)
+            return _lexer_cache[name]
+    # continue with lexers from setuptools entrypoints
+    for cls in find_plugin_lexers():
+        if cls.name == name:
+            return cls
+
+
+def find_lexer_class_by_name(_alias):
+    """
+    Return the `Lexer` subclass that has `alias` in its aliases list, without
+    instantiating it.
+
+    Like `get_lexer_by_name`, but does not instantiate the class.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if no lexer with that alias is
+    found.
+
+    .. versionadded:: 2.2
+    """
+    if not _alias:
+        raise ClassNotFound(f'no lexer for alias {_alias!r} found')
+    # lookup builtin lexers
+    for module_name, name, aliases, _, _ in LEXERS.values():
+        if _alias.lower() in aliases:
+            if name not in _lexer_cache:
+                _load_lexers(module_name)
+            return _lexer_cache[name]
+    # continue with lexers from setuptools entrypoints
+    for cls in find_plugin_lexers():
+        if _alias.lower() in cls.aliases:
+            return cls
+    raise ClassNotFound(f'no lexer for alias {_alias!r} found')
+
+
+def get_lexer_by_name(_alias, **options):
+    """
+    Return an instance of a `Lexer` subclass that has `alias` in its
+    aliases list. The lexer is given the `options` at its
+    instantiation.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if no lexer with that alias is
+    found.
+    """
+    if not _alias:
+        raise ClassNotFound(f'no lexer for alias {_alias!r} found')
+
+    # lookup builtin lexers
+    for module_name, name, aliases, _, _ in LEXERS.values():
+        if _alias.lower() in aliases:
+            if name not in _lexer_cache:
+                _load_lexers(module_name)
+            return _lexer_cache[name](**options)
+    # continue with lexers from setuptools entrypoints
+    for cls in find_plugin_lexers():
+        if _alias.lower() in cls.aliases:
+            return cls(**options)
+    raise ClassNotFound(f'no lexer for alias {_alias!r} found')
+
+
+def load_lexer_from_file(filename, lexername="CustomLexer", **options):
+    """Load a lexer from a file.
+
+    This method expects a file located relative to the current working
+    directory, which contains a Lexer class. By default, it expects the
+    Lexer to be name CustomLexer; you can specify your own class name
+    as the second argument to this function.
+
+    Users should be very careful with the input, because this method
+    is equivalent to running eval on the input file.
+
+    Raises ClassNotFound if there are any problems importing the Lexer.
+
+    .. versionadded:: 2.2
+    """
+    try:
+        # This empty dict will contain the namespace for the exec'd file
+        custom_namespace = {}
+        with open(filename, 'rb') as f:
+            exec(f.read(), custom_namespace)
+        # Retrieve the class `lexername` from that namespace
+        if lexername not in custom_namespace:
+            raise ClassNotFound(f'no valid {lexername} class found in {filename}')
+        lexer_class = custom_namespace[lexername]
+        # And finally instantiate it with the options
+        return lexer_class(**options)
+    except OSError as err:
+        raise ClassNotFound(f'cannot read {filename}: {err}')
+    except ClassNotFound:
+        raise
+    except Exception as err:
+        raise ClassNotFound(f'error when loading custom lexer: {err}')
+
+
+def find_lexer_class_for_filename(_fn, code=None):
+    """Get a lexer for a filename.
+
+    If multiple lexers match the filename pattern, use ``analyse_text()`` to
+    figure out which one is more appropriate.
+
+    Returns None if not found.
+    """
+    matches = []
+    fn = basename(_fn)
+    for modname, name, _, filenames, _ in LEXERS.values():
+        for filename in filenames:
+            if _fn_matches(fn, filename):
+                if name not in _lexer_cache:
+                    _load_lexers(modname)
+                matches.append((_lexer_cache[name], filename))
+    for cls in find_plugin_lexers():
+        for filename in cls.filenames:
+            if _fn_matches(fn, filename):
+                matches.append((cls, filename))
+
+    if isinstance(code, bytes):
+        # decode it, since all analyse_text functions expect unicode
+        code = guess_decode(code)
+
+    def get_rating(info):
+        cls, filename = info
+        # explicit patterns get a bonus
+        bonus = '*' not in filename and 0.5 or 0
+        # The class _always_ defines analyse_text because it's included in
+        # the Lexer class.  The default implementation returns None which
+        # gets turned into 0.0.  Run scripts/detect_missing_analyse_text.py
+        # to find lexers which need it overridden.
+        if code:
+            return cls.analyse_text(code) + bonus, cls.__name__
+        return cls.priority + bonus, cls.__name__
+
+    if matches:
+        matches.sort(key=get_rating)
+        # print "Possible lexers, after sort:", matches
+        return matches[-1][0]
+
+
+def get_lexer_for_filename(_fn, code=None, **options):
+    """Get a lexer for a filename.
+
+    Return a `Lexer` subclass instance that has a filename pattern
+    matching `fn`. The lexer is given the `options` at its
+    instantiation.
+
+    Raise :exc:`pygments.util.ClassNotFound` if no lexer for that filename
+    is found.
+
+    If multiple lexers match the filename pattern, use their ``analyse_text()``
+    methods to figure out which one is more appropriate.
+    """
+    res = find_lexer_class_for_filename(_fn, code)
+    if not res:
+        raise ClassNotFound(f'no lexer for filename {_fn!r} found')
+    return res(**options)
+
+
+def get_lexer_for_mimetype(_mime, **options):
+    """
+    Return a `Lexer` subclass instance that has `mime` in its mimetype
+    list. The lexer is given the `options` at its instantiation.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if not lexer for that mimetype
+    is found.
+    """
+    for modname, name, _, _, mimetypes in LEXERS.values():
+        if _mime in mimetypes:
+            if name not in _lexer_cache:
+                _load_lexers(modname)
+            return _lexer_cache[name](**options)
+    for cls in find_plugin_lexers():
+        if _mime in cls.mimetypes:
+            return cls(**options)
+    raise ClassNotFound(f'no lexer for mimetype {_mime!r} found')
+
+
+def _iter_lexerclasses(plugins=True):
+    """Return an iterator over all lexer classes."""
+    for key in sorted(LEXERS):
+        module_name, name = LEXERS[key][:2]
+        if name not in _lexer_cache:
+            _load_lexers(module_name)
+        yield _lexer_cache[name]
+    if plugins:
+        yield from find_plugin_lexers()
+
+
+def guess_lexer_for_filename(_fn, _text, **options):
+    """
+    As :func:`guess_lexer()`, but only lexers which have a pattern in `filenames`
+    or `alias_filenames` that matches `filename` are taken into consideration.
+
+    :exc:`pygments.util.ClassNotFound` is raised if no lexer thinks it can
+    handle the content.
+    """
+    fn = basename(_fn)
+    primary = {}
+    matching_lexers = set()
+    for lexer in _iter_lexerclasses():
+        for filename in lexer.filenames:
+            if _fn_matches(fn, filename):
+                matching_lexers.add(lexer)
+                primary[lexer] = True
+        for filename in lexer.alias_filenames:
+            if _fn_matches(fn, filename):
+                matching_lexers.add(lexer)
+                primary[lexer] = False
+    if not matching_lexers:
+        raise ClassNotFound(f'no lexer for filename {fn!r} found')
+    if len(matching_lexers) == 1:
+        return matching_lexers.pop()(**options)
+    result = []
+    for lexer in matching_lexers:
+        rv = lexer.analyse_text(_text)
+        if rv == 1.0:
+            return lexer(**options)
+        result.append((rv, lexer))
+
+    def type_sort(t):
+        # sort by:
+        # - analyse score
+        # - is primary filename pattern?
+        # - priority
+        # - last resort: class name
+        return (t[0], primary[t[1]], t[1].priority, t[1].__name__)
+    result.sort(key=type_sort)
+
+    return result[-1][1](**options)
+
+
+def guess_lexer(_text, **options):
+    """
+    Return a `Lexer` subclass instance that's guessed from the text in
+    `text`. For that, the :meth:`.analyse_text()` method of every known lexer
+    class is called with the text as argument, and the lexer which returned the
+    highest value will be instantiated and returned.
+
+    :exc:`pygments.util.ClassNotFound` is raised if no lexer thinks it can
+    handle the content.
+    """
+
+    if not isinstance(_text, str):
+        inencoding = options.get('inencoding', options.get('encoding'))
+        if inencoding:
+            _text = _text.decode(inencoding or 'utf8')
+        else:
+            _text, _ = guess_decode(_text)
+
+    # try to get a vim modeline first
+    ft = get_filetype_from_buffer(_text)
+
+    if ft is not None:
+        try:
+            return get_lexer_by_name(ft, **options)
+        except ClassNotFound:
+            pass
+
+    best_lexer = [0.0, None]
+    for lexer in _iter_lexerclasses():
+        rv = lexer.analyse_text(_text)
+        if rv == 1.0:
+            return lexer(**options)
+        if rv > best_lexer[0]:
+            best_lexer[:] = (rv, lexer)
+    if not best_lexer[0] or best_lexer[1] is None:
+        raise ClassNotFound('no lexer matching the text found')
+    return best_lexer[1](**options)
+
+
+class _automodule(types.ModuleType):
+    """Automatically import lexers."""
+
+    def __getattr__(self, name):
+        info = LEXERS.get(name)
+        if info:
+            _load_lexers(info[0])
+            cls = _lexer_cache[info[1]]
+            setattr(self, name, cls)
+            return cls
+        if name in COMPAT:
+            return getattr(self, COMPAT[name])
+        raise AttributeError(name)
+
+
+oldmod = sys.modules[__name__]
+newmod = _automodule(__name__)
+newmod.__dict__.update(oldmod.__dict__)
+sys.modules[__name__] = newmod
+del newmod.newmod, newmod.oldmod, newmod.sys, newmod.types
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3d6a861f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/_mapping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/_mapping.cpython-312.pyc
new file mode 100644
index 00000000..578cd2f2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/_mapping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/python.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/python.cpython-312.pyc
new file mode 100644
index 00000000..e1274e42
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/__pycache__/python.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/_mapping.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/_mapping.py
new file mode 100644
index 00000000..f3e5c460
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/_mapping.py
@@ -0,0 +1,589 @@
+# Automatically generated by scripts/gen_mapfiles.py.
+# DO NOT EDIT BY HAND; run `tox -e mapfiles` instead.
+
+LEXERS = {
+    'ABAPLexer': ('pip._vendor.pygments.lexers.business', 'ABAP', ('abap',), ('*.abap', '*.ABAP'), ('text/x-abap',)),
+    'AMDGPULexer': ('pip._vendor.pygments.lexers.amdgpu', 'AMDGPU', ('amdgpu',), ('*.isa',), ()),
+    'APLLexer': ('pip._vendor.pygments.lexers.apl', 'APL', ('apl',), ('*.apl', '*.aplf', '*.aplo', '*.apln', '*.aplc', '*.apli', '*.dyalog'), ()),
+    'AbnfLexer': ('pip._vendor.pygments.lexers.grammar_notation', 'ABNF', ('abnf',), ('*.abnf',), ('text/x-abnf',)),
+    'ActionScript3Lexer': ('pip._vendor.pygments.lexers.actionscript', 'ActionScript 3', ('actionscript3', 'as3'), ('*.as',), ('application/x-actionscript3', 'text/x-actionscript3', 'text/actionscript3')),
+    'ActionScriptLexer': ('pip._vendor.pygments.lexers.actionscript', 'ActionScript', ('actionscript', 'as'), ('*.as',), ('application/x-actionscript', 'text/x-actionscript', 'text/actionscript')),
+    'AdaLexer': ('pip._vendor.pygments.lexers.ada', 'Ada', ('ada', 'ada95', 'ada2005'), ('*.adb', '*.ads', '*.ada'), ('text/x-ada',)),
+    'AdlLexer': ('pip._vendor.pygments.lexers.archetype', 'ADL', ('adl',), ('*.adl', '*.adls', '*.adlf', '*.adlx'), ()),
+    'AgdaLexer': ('pip._vendor.pygments.lexers.haskell', 'Agda', ('agda',), ('*.agda',), ('text/x-agda',)),
+    'AheuiLexer': ('pip._vendor.pygments.lexers.esoteric', 'Aheui', ('aheui',), ('*.aheui',), ()),
+    'AlloyLexer': ('pip._vendor.pygments.lexers.dsls', 'Alloy', ('alloy',), ('*.als',), ('text/x-alloy',)),
+    'AmbientTalkLexer': ('pip._vendor.pygments.lexers.ambient', 'AmbientTalk', ('ambienttalk', 'ambienttalk/2', 'at'), ('*.at',), ('text/x-ambienttalk',)),
+    'AmplLexer': ('pip._vendor.pygments.lexers.ampl', 'Ampl', ('ampl',), ('*.run',), ()),
+    'Angular2HtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML + Angular2', ('html+ng2',), ('*.ng2',), ()),
+    'Angular2Lexer': ('pip._vendor.pygments.lexers.templates', 'Angular2', ('ng2',), (), ()),
+    'AntlrActionScriptLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With ActionScript Target', ('antlr-actionscript', 'antlr-as'), ('*.G', '*.g'), ()),
+    'AntlrCSharpLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With C# Target', ('antlr-csharp', 'antlr-c#'), ('*.G', '*.g'), ()),
+    'AntlrCppLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With CPP Target', ('antlr-cpp',), ('*.G', '*.g'), ()),
+    'AntlrJavaLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With Java Target', ('antlr-java',), ('*.G', '*.g'), ()),
+    'AntlrLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR', ('antlr',), (), ()),
+    'AntlrObjectiveCLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With ObjectiveC Target', ('antlr-objc',), ('*.G', '*.g'), ()),
+    'AntlrPerlLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With Perl Target', ('antlr-perl',), ('*.G', '*.g'), ()),
+    'AntlrPythonLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With Python Target', ('antlr-python',), ('*.G', '*.g'), ()),
+    'AntlrRubyLexer': ('pip._vendor.pygments.lexers.parsers', 'ANTLR With Ruby Target', ('antlr-ruby', 'antlr-rb'), ('*.G', '*.g'), ()),
+    'ApacheConfLexer': ('pip._vendor.pygments.lexers.configs', 'ApacheConf', ('apacheconf', 'aconf', 'apache'), ('.htaccess', 'apache.conf', 'apache2.conf'), ('text/x-apacheconf',)),
+    'AppleScriptLexer': ('pip._vendor.pygments.lexers.scripting', 'AppleScript', ('applescript',), ('*.applescript',), ()),
+    'ArduinoLexer': ('pip._vendor.pygments.lexers.c_like', 'Arduino', ('arduino',), ('*.ino',), ('text/x-arduino',)),
+    'ArrowLexer': ('pip._vendor.pygments.lexers.arrow', 'Arrow', ('arrow',), ('*.arw',), ()),
+    'ArturoLexer': ('pip._vendor.pygments.lexers.arturo', 'Arturo', ('arturo', 'art'), ('*.art',), ()),
+    'AscLexer': ('pip._vendor.pygments.lexers.asc', 'ASCII armored', ('asc', 'pem'), ('*.asc', '*.pem', 'id_dsa', 'id_ecdsa', 'id_ecdsa_sk', 'id_ed25519', 'id_ed25519_sk', 'id_rsa'), ('application/pgp-keys', 'application/pgp-encrypted', 'application/pgp-signature', 'application/pem-certificate-chain')),
+    'Asn1Lexer': ('pip._vendor.pygments.lexers.asn1', 'ASN.1', ('asn1',), ('*.asn1',), ()),
+    'AspectJLexer': ('pip._vendor.pygments.lexers.jvm', 'AspectJ', ('aspectj',), ('*.aj',), ('text/x-aspectj',)),
+    'AsymptoteLexer': ('pip._vendor.pygments.lexers.graphics', 'Asymptote', ('asymptote', 'asy'), ('*.asy',), ('text/x-asymptote',)),
+    'AugeasLexer': ('pip._vendor.pygments.lexers.configs', 'Augeas', ('augeas',), ('*.aug',), ()),
+    'AutoItLexer': ('pip._vendor.pygments.lexers.automation', 'AutoIt', ('autoit',), ('*.au3',), ('text/x-autoit',)),
+    'AutohotkeyLexer': ('pip._vendor.pygments.lexers.automation', 'autohotkey', ('autohotkey', 'ahk'), ('*.ahk', '*.ahkl'), ('text/x-autohotkey',)),
+    'AwkLexer': ('pip._vendor.pygments.lexers.textedit', 'Awk', ('awk', 'gawk', 'mawk', 'nawk'), ('*.awk',), ('application/x-awk',)),
+    'BBCBasicLexer': ('pip._vendor.pygments.lexers.basic', 'BBC Basic', ('bbcbasic',), ('*.bbc',), ()),
+    'BBCodeLexer': ('pip._vendor.pygments.lexers.markup', 'BBCode', ('bbcode',), (), ('text/x-bbcode',)),
+    'BCLexer': ('pip._vendor.pygments.lexers.algebra', 'BC', ('bc',), ('*.bc',), ()),
+    'BQNLexer': ('pip._vendor.pygments.lexers.bqn', 'BQN', ('bqn',), ('*.bqn',), ()),
+    'BSTLexer': ('pip._vendor.pygments.lexers.bibtex', 'BST', ('bst', 'bst-pybtex'), ('*.bst',), ()),
+    'BareLexer': ('pip._vendor.pygments.lexers.bare', 'BARE', ('bare',), ('*.bare',), ()),
+    'BaseMakefileLexer': ('pip._vendor.pygments.lexers.make', 'Base Makefile', ('basemake',), (), ()),
+    'BashLexer': ('pip._vendor.pygments.lexers.shell', 'Bash', ('bash', 'sh', 'ksh', 'zsh', 'shell', 'openrc'), ('*.sh', '*.ksh', '*.bash', '*.ebuild', '*.eclass', '*.exheres-0', '*.exlib', '*.zsh', '.bashrc', 'bashrc', '.bash_*', 'bash_*', 'zshrc', '.zshrc', '.kshrc', 'kshrc', 'PKGBUILD'), ('application/x-sh', 'application/x-shellscript', 'text/x-shellscript')),
+    'BashSessionLexer': ('pip._vendor.pygments.lexers.shell', 'Bash Session', ('console', 'shell-session'), ('*.sh-session', '*.shell-session'), ('application/x-shell-session', 'application/x-sh-session')),
+    'BatchLexer': ('pip._vendor.pygments.lexers.shell', 'Batchfile', ('batch', 'bat', 'dosbatch', 'winbatch'), ('*.bat', '*.cmd'), ('application/x-dos-batch',)),
+    'BddLexer': ('pip._vendor.pygments.lexers.bdd', 'Bdd', ('bdd',), ('*.feature',), ('text/x-bdd',)),
+    'BefungeLexer': ('pip._vendor.pygments.lexers.esoteric', 'Befunge', ('befunge',), ('*.befunge',), ('application/x-befunge',)),
+    'BerryLexer': ('pip._vendor.pygments.lexers.berry', 'Berry', ('berry', 'be'), ('*.be',), ('text/x-berry', 'application/x-berry')),
+    'BibTeXLexer': ('pip._vendor.pygments.lexers.bibtex', 'BibTeX', ('bibtex', 'bib'), ('*.bib',), ('text/x-bibtex',)),
+    'BlitzBasicLexer': ('pip._vendor.pygments.lexers.basic', 'BlitzBasic', ('blitzbasic', 'b3d', 'bplus'), ('*.bb', '*.decls'), ('text/x-bb',)),
+    'BlitzMaxLexer': ('pip._vendor.pygments.lexers.basic', 'BlitzMax', ('blitzmax', 'bmax'), ('*.bmx',), ('text/x-bmx',)),
+    'BlueprintLexer': ('pip._vendor.pygments.lexers.blueprint', 'Blueprint', ('blueprint',), ('*.blp',), ('text/x-blueprint',)),
+    'BnfLexer': ('pip._vendor.pygments.lexers.grammar_notation', 'BNF', ('bnf',), ('*.bnf',), ('text/x-bnf',)),
+    'BoaLexer': ('pip._vendor.pygments.lexers.boa', 'Boa', ('boa',), ('*.boa',), ()),
+    'BooLexer': ('pip._vendor.pygments.lexers.dotnet', 'Boo', ('boo',), ('*.boo',), ('text/x-boo',)),
+    'BoogieLexer': ('pip._vendor.pygments.lexers.verification', 'Boogie', ('boogie',), ('*.bpl',), ()),
+    'BrainfuckLexer': ('pip._vendor.pygments.lexers.esoteric', 'Brainfuck', ('brainfuck', 'bf'), ('*.bf', '*.b'), ('application/x-brainfuck',)),
+    'BugsLexer': ('pip._vendor.pygments.lexers.modeling', 'BUGS', ('bugs', 'winbugs', 'openbugs'), ('*.bug',), ()),
+    'CAmkESLexer': ('pip._vendor.pygments.lexers.esoteric', 'CAmkES', ('camkes', 'idl4'), ('*.camkes', '*.idl4'), ()),
+    'CLexer': ('pip._vendor.pygments.lexers.c_cpp', 'C', ('c',), ('*.c', '*.h', '*.idc', '*.x[bp]m'), ('text/x-chdr', 'text/x-csrc', 'image/x-xbitmap', 'image/x-xpixmap')),
+    'CMakeLexer': ('pip._vendor.pygments.lexers.make', 'CMake', ('cmake',), ('*.cmake', 'CMakeLists.txt'), ('text/x-cmake',)),
+    'CObjdumpLexer': ('pip._vendor.pygments.lexers.asm', 'c-objdump', ('c-objdump',), ('*.c-objdump',), ('text/x-c-objdump',)),
+    'CPSALexer': ('pip._vendor.pygments.lexers.lisp', 'CPSA', ('cpsa',), ('*.cpsa',), ()),
+    'CSSUL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'CSS+UL4', ('css+ul4',), ('*.cssul4',), ()),
+    'CSharpAspxLexer': ('pip._vendor.pygments.lexers.dotnet', 'aspx-cs', ('aspx-cs',), ('*.aspx', '*.asax', '*.ascx', '*.ashx', '*.asmx', '*.axd'), ()),
+    'CSharpLexer': ('pip._vendor.pygments.lexers.dotnet', 'C#', ('csharp', 'c#', 'cs'), ('*.cs',), ('text/x-csharp',)),
+    'Ca65Lexer': ('pip._vendor.pygments.lexers.asm', 'ca65 assembler', ('ca65',), ('*.s',), ()),
+    'CadlLexer': ('pip._vendor.pygments.lexers.archetype', 'cADL', ('cadl',), ('*.cadl',), ()),
+    'CapDLLexer': ('pip._vendor.pygments.lexers.esoteric', 'CapDL', ('capdl',), ('*.cdl',), ()),
+    'CapnProtoLexer': ('pip._vendor.pygments.lexers.capnproto', "Cap'n Proto", ('capnp',), ('*.capnp',), ()),
+    'CarbonLexer': ('pip._vendor.pygments.lexers.carbon', 'Carbon', ('carbon',), ('*.carbon',), ('text/x-carbon',)),
+    'CbmBasicV2Lexer': ('pip._vendor.pygments.lexers.basic', 'CBM BASIC V2', ('cbmbas',), ('*.bas',), ()),
+    'CddlLexer': ('pip._vendor.pygments.lexers.cddl', 'CDDL', ('cddl',), ('*.cddl',), ('text/x-cddl',)),
+    'CeylonLexer': ('pip._vendor.pygments.lexers.jvm', 'Ceylon', ('ceylon',), ('*.ceylon',), ('text/x-ceylon',)),
+    'Cfengine3Lexer': ('pip._vendor.pygments.lexers.configs', 'CFEngine3', ('cfengine3', 'cf3'), ('*.cf',), ()),
+    'ChaiscriptLexer': ('pip._vendor.pygments.lexers.scripting', 'ChaiScript', ('chaiscript', 'chai'), ('*.chai',), ('text/x-chaiscript', 'application/x-chaiscript')),
+    'ChapelLexer': ('pip._vendor.pygments.lexers.chapel', 'Chapel', ('chapel', 'chpl'), ('*.chpl',), ()),
+    'CharmciLexer': ('pip._vendor.pygments.lexers.c_like', 'Charmci', ('charmci',), ('*.ci',), ()),
+    'CheetahHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Cheetah', ('html+cheetah', 'html+spitfire', 'htmlcheetah'), (), ('text/html+cheetah', 'text/html+spitfire')),
+    'CheetahJavascriptLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Cheetah', ('javascript+cheetah', 'js+cheetah', 'javascript+spitfire', 'js+spitfire'), (), ('application/x-javascript+cheetah', 'text/x-javascript+cheetah', 'text/javascript+cheetah', 'application/x-javascript+spitfire', 'text/x-javascript+spitfire', 'text/javascript+spitfire')),
+    'CheetahLexer': ('pip._vendor.pygments.lexers.templates', 'Cheetah', ('cheetah', 'spitfire'), ('*.tmpl', '*.spt'), ('application/x-cheetah', 'application/x-spitfire')),
+    'CheetahXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Cheetah', ('xml+cheetah', 'xml+spitfire'), (), ('application/xml+cheetah', 'application/xml+spitfire')),
+    'CirruLexer': ('pip._vendor.pygments.lexers.webmisc', 'Cirru', ('cirru',), ('*.cirru',), ('text/x-cirru',)),
+    'ClayLexer': ('pip._vendor.pygments.lexers.c_like', 'Clay', ('clay',), ('*.clay',), ('text/x-clay',)),
+    'CleanLexer': ('pip._vendor.pygments.lexers.clean', 'Clean', ('clean',), ('*.icl', '*.dcl'), ()),
+    'ClojureLexer': ('pip._vendor.pygments.lexers.jvm', 'Clojure', ('clojure', 'clj'), ('*.clj', '*.cljc'), ('text/x-clojure', 'application/x-clojure')),
+    'ClojureScriptLexer': ('pip._vendor.pygments.lexers.jvm', 'ClojureScript', ('clojurescript', 'cljs'), ('*.cljs',), ('text/x-clojurescript', 'application/x-clojurescript')),
+    'CobolFreeformatLexer': ('pip._vendor.pygments.lexers.business', 'COBOLFree', ('cobolfree',), ('*.cbl', '*.CBL'), ()),
+    'CobolLexer': ('pip._vendor.pygments.lexers.business', 'COBOL', ('cobol',), ('*.cob', '*.COB', '*.cpy', '*.CPY'), ('text/x-cobol',)),
+    'CoffeeScriptLexer': ('pip._vendor.pygments.lexers.javascript', 'CoffeeScript', ('coffeescript', 'coffee-script', 'coffee'), ('*.coffee',), ('text/coffeescript',)),
+    'ColdfusionCFCLexer': ('pip._vendor.pygments.lexers.templates', 'Coldfusion CFC', ('cfc',), ('*.cfc',), ()),
+    'ColdfusionHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'Coldfusion HTML', ('cfm',), ('*.cfm', '*.cfml'), ('application/x-coldfusion',)),
+    'ColdfusionLexer': ('pip._vendor.pygments.lexers.templates', 'cfstatement', ('cfs',), (), ()),
+    'Comal80Lexer': ('pip._vendor.pygments.lexers.comal', 'COMAL-80', ('comal', 'comal80'), ('*.cml', '*.comal'), ()),
+    'CommonLispLexer': ('pip._vendor.pygments.lexers.lisp', 'Common Lisp', ('common-lisp', 'cl', 'lisp'), ('*.cl', '*.lisp'), ('text/x-common-lisp',)),
+    'ComponentPascalLexer': ('pip._vendor.pygments.lexers.oberon', 'Component Pascal', ('componentpascal', 'cp'), ('*.cp', '*.cps'), ('text/x-component-pascal',)),
+    'CoqLexer': ('pip._vendor.pygments.lexers.theorem', 'Coq', ('coq',), ('*.v',), ('text/x-coq',)),
+    'CplintLexer': ('pip._vendor.pygments.lexers.cplint', 'cplint', ('cplint',), ('*.ecl', '*.prolog', '*.pro', '*.pl', '*.P', '*.lpad', '*.cpl'), ('text/x-cplint',)),
+    'CppLexer': ('pip._vendor.pygments.lexers.c_cpp', 'C++', ('cpp', 'c++'), ('*.cpp', '*.hpp', '*.c++', '*.h++', '*.cc', '*.hh', '*.cxx', '*.hxx', '*.C', '*.H', '*.cp', '*.CPP', '*.tpp'), ('text/x-c++hdr', 'text/x-c++src')),
+    'CppObjdumpLexer': ('pip._vendor.pygments.lexers.asm', 'cpp-objdump', ('cpp-objdump', 'c++-objdumb', 'cxx-objdump'), ('*.cpp-objdump', '*.c++-objdump', '*.cxx-objdump'), ('text/x-cpp-objdump',)),
+    'CrmshLexer': ('pip._vendor.pygments.lexers.dsls', 'Crmsh', ('crmsh', 'pcmk'), ('*.crmsh', '*.pcmk'), ()),
+    'CrocLexer': ('pip._vendor.pygments.lexers.d', 'Croc', ('croc',), ('*.croc',), ('text/x-crocsrc',)),
+    'CryptolLexer': ('pip._vendor.pygments.lexers.haskell', 'Cryptol', ('cryptol', 'cry'), ('*.cry',), ('text/x-cryptol',)),
+    'CrystalLexer': ('pip._vendor.pygments.lexers.crystal', 'Crystal', ('cr', 'crystal'), ('*.cr',), ('text/x-crystal',)),
+    'CsoundDocumentLexer': ('pip._vendor.pygments.lexers.csound', 'Csound Document', ('csound-document', 'csound-csd'), ('*.csd',), ()),
+    'CsoundOrchestraLexer': ('pip._vendor.pygments.lexers.csound', 'Csound Orchestra', ('csound', 'csound-orc'), ('*.orc', '*.udo'), ()),
+    'CsoundScoreLexer': ('pip._vendor.pygments.lexers.csound', 'Csound Score', ('csound-score', 'csound-sco'), ('*.sco',), ()),
+    'CssDjangoLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Django/Jinja', ('css+django', 'css+jinja'), ('*.css.j2', '*.css.jinja2'), ('text/css+django', 'text/css+jinja')),
+    'CssErbLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Ruby', ('css+ruby', 'css+erb'), (), ('text/css+ruby',)),
+    'CssGenshiLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Genshi Text', ('css+genshitext', 'css+genshi'), (), ('text/css+genshi',)),
+    'CssLexer': ('pip._vendor.pygments.lexers.css', 'CSS', ('css',), ('*.css',), ('text/css',)),
+    'CssPhpLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+PHP', ('css+php',), (), ('text/css+php',)),
+    'CssSmartyLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Smarty', ('css+smarty',), (), ('text/css+smarty',)),
+    'CudaLexer': ('pip._vendor.pygments.lexers.c_like', 'CUDA', ('cuda', 'cu'), ('*.cu', '*.cuh'), ('text/x-cuda',)),
+    'CypherLexer': ('pip._vendor.pygments.lexers.graph', 'Cypher', ('cypher',), ('*.cyp', '*.cypher'), ()),
+    'CythonLexer': ('pip._vendor.pygments.lexers.python', 'Cython', ('cython', 'pyx', 'pyrex'), ('*.pyx', '*.pxd', '*.pxi'), ('text/x-cython', 'application/x-cython')),
+    'DLexer': ('pip._vendor.pygments.lexers.d', 'D', ('d',), ('*.d', '*.di'), ('text/x-dsrc',)),
+    'DObjdumpLexer': ('pip._vendor.pygments.lexers.asm', 'd-objdump', ('d-objdump',), ('*.d-objdump',), ('text/x-d-objdump',)),
+    'DarcsPatchLexer': ('pip._vendor.pygments.lexers.diff', 'Darcs Patch', ('dpatch',), ('*.dpatch', '*.darcspatch'), ()),
+    'DartLexer': ('pip._vendor.pygments.lexers.javascript', 'Dart', ('dart',), ('*.dart',), ('text/x-dart',)),
+    'Dasm16Lexer': ('pip._vendor.pygments.lexers.asm', 'DASM16', ('dasm16',), ('*.dasm16', '*.dasm'), ('text/x-dasm16',)),
+    'DaxLexer': ('pip._vendor.pygments.lexers.dax', 'Dax', ('dax',), ('*.dax',), ()),
+    'DebianControlLexer': ('pip._vendor.pygments.lexers.installers', 'Debian Control file', ('debcontrol', 'control'), ('control',), ()),
+    'DelphiLexer': ('pip._vendor.pygments.lexers.pascal', 'Delphi', ('delphi', 'pas', 'pascal', 'objectpascal'), ('*.pas', '*.dpr'), ('text/x-pascal',)),
+    'DesktopLexer': ('pip._vendor.pygments.lexers.configs', 'Desktop file', ('desktop',), ('*.desktop',), ('application/x-desktop',)),
+    'DevicetreeLexer': ('pip._vendor.pygments.lexers.devicetree', 'Devicetree', ('devicetree', 'dts'), ('*.dts', '*.dtsi'), ('text/x-c',)),
+    'DgLexer': ('pip._vendor.pygments.lexers.python', 'dg', ('dg',), ('*.dg',), ('text/x-dg',)),
+    'DiffLexer': ('pip._vendor.pygments.lexers.diff', 'Diff', ('diff', 'udiff'), ('*.diff', '*.patch'), ('text/x-diff', 'text/x-patch')),
+    'DjangoLexer': ('pip._vendor.pygments.lexers.templates', 'Django/Jinja', ('django', 'jinja'), (), ('application/x-django-templating', 'application/x-jinja')),
+    'DnsZoneLexer': ('pip._vendor.pygments.lexers.dns', 'Zone', ('zone',), ('*.zone',), ('text/dns',)),
+    'DockerLexer': ('pip._vendor.pygments.lexers.configs', 'Docker', ('docker', 'dockerfile'), ('Dockerfile', '*.docker'), ('text/x-dockerfile-config',)),
+    'DtdLexer': ('pip._vendor.pygments.lexers.html', 'DTD', ('dtd',), ('*.dtd',), ('application/xml-dtd',)),
+    'DuelLexer': ('pip._vendor.pygments.lexers.webmisc', 'Duel', ('duel', 'jbst', 'jsonml+bst'), ('*.duel', '*.jbst'), ('text/x-duel', 'text/x-jbst')),
+    'DylanConsoleLexer': ('pip._vendor.pygments.lexers.dylan', 'Dylan session', ('dylan-console', 'dylan-repl'), ('*.dylan-console',), ('text/x-dylan-console',)),
+    'DylanLexer': ('pip._vendor.pygments.lexers.dylan', 'Dylan', ('dylan',), ('*.dylan', '*.dyl', '*.intr'), ('text/x-dylan',)),
+    'DylanLidLexer': ('pip._vendor.pygments.lexers.dylan', 'DylanLID', ('dylan-lid', 'lid'), ('*.lid', '*.hdp'), ('text/x-dylan-lid',)),
+    'ECLLexer': ('pip._vendor.pygments.lexers.ecl', 'ECL', ('ecl',), ('*.ecl',), ('application/x-ecl',)),
+    'ECLexer': ('pip._vendor.pygments.lexers.c_like', 'eC', ('ec',), ('*.ec', '*.eh'), ('text/x-echdr', 'text/x-ecsrc')),
+    'EarlGreyLexer': ('pip._vendor.pygments.lexers.javascript', 'Earl Grey', ('earl-grey', 'earlgrey', 'eg'), ('*.eg',), ('text/x-earl-grey',)),
+    'EasytrieveLexer': ('pip._vendor.pygments.lexers.scripting', 'Easytrieve', ('easytrieve',), ('*.ezt', '*.mac'), ('text/x-easytrieve',)),
+    'EbnfLexer': ('pip._vendor.pygments.lexers.parsers', 'EBNF', ('ebnf',), ('*.ebnf',), ('text/x-ebnf',)),
+    'EiffelLexer': ('pip._vendor.pygments.lexers.eiffel', 'Eiffel', ('eiffel',), ('*.e',), ('text/x-eiffel',)),
+    'ElixirConsoleLexer': ('pip._vendor.pygments.lexers.erlang', 'Elixir iex session', ('iex',), (), ('text/x-elixir-shellsession',)),
+    'ElixirLexer': ('pip._vendor.pygments.lexers.erlang', 'Elixir', ('elixir', 'ex', 'exs'), ('*.ex', '*.eex', '*.exs', '*.leex'), ('text/x-elixir',)),
+    'ElmLexer': ('pip._vendor.pygments.lexers.elm', 'Elm', ('elm',), ('*.elm',), ('text/x-elm',)),
+    'ElpiLexer': ('pip._vendor.pygments.lexers.elpi', 'Elpi', ('elpi',), ('*.elpi',), ('text/x-elpi',)),
+    'EmacsLispLexer': ('pip._vendor.pygments.lexers.lisp', 'EmacsLisp', ('emacs-lisp', 'elisp', 'emacs'), ('*.el',), ('text/x-elisp', 'application/x-elisp')),
+    'EmailLexer': ('pip._vendor.pygments.lexers.email', 'E-mail', ('email', 'eml'), ('*.eml',), ('message/rfc822',)),
+    'ErbLexer': ('pip._vendor.pygments.lexers.templates', 'ERB', ('erb',), (), ('application/x-ruby-templating',)),
+    'ErlangLexer': ('pip._vendor.pygments.lexers.erlang', 'Erlang', ('erlang',), ('*.erl', '*.hrl', '*.es', '*.escript'), ('text/x-erlang',)),
+    'ErlangShellLexer': ('pip._vendor.pygments.lexers.erlang', 'Erlang erl session', ('erl',), ('*.erl-sh',), ('text/x-erl-shellsession',)),
+    'EvoqueHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Evoque', ('html+evoque',), ('*.html',), ('text/html+evoque',)),
+    'EvoqueLexer': ('pip._vendor.pygments.lexers.templates', 'Evoque', ('evoque',), ('*.evoque',), ('application/x-evoque',)),
+    'EvoqueXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Evoque', ('xml+evoque',), ('*.xml',), ('application/xml+evoque',)),
+    'ExeclineLexer': ('pip._vendor.pygments.lexers.shell', 'execline', ('execline',), ('*.exec',), ()),
+    'EzhilLexer': ('pip._vendor.pygments.lexers.ezhil', 'Ezhil', ('ezhil',), ('*.n',), ('text/x-ezhil',)),
+    'FSharpLexer': ('pip._vendor.pygments.lexers.dotnet', 'F#', ('fsharp', 'f#'), ('*.fs', '*.fsi', '*.fsx'), ('text/x-fsharp',)),
+    'FStarLexer': ('pip._vendor.pygments.lexers.ml', 'FStar', ('fstar',), ('*.fst', '*.fsti'), ('text/x-fstar',)),
+    'FactorLexer': ('pip._vendor.pygments.lexers.factor', 'Factor', ('factor',), ('*.factor',), ('text/x-factor',)),
+    'FancyLexer': ('pip._vendor.pygments.lexers.ruby', 'Fancy', ('fancy', 'fy'), ('*.fy', '*.fancypack'), ('text/x-fancysrc',)),
+    'FantomLexer': ('pip._vendor.pygments.lexers.fantom', 'Fantom', ('fan',), ('*.fan',), ('application/x-fantom',)),
+    'FelixLexer': ('pip._vendor.pygments.lexers.felix', 'Felix', ('felix', 'flx'), ('*.flx', '*.flxh'), ('text/x-felix',)),
+    'FennelLexer': ('pip._vendor.pygments.lexers.lisp', 'Fennel', ('fennel', 'fnl'), ('*.fnl',), ()),
+    'FiftLexer': ('pip._vendor.pygments.lexers.fift', 'Fift', ('fift', 'fif'), ('*.fif',), ()),
+    'FishShellLexer': ('pip._vendor.pygments.lexers.shell', 'Fish', ('fish', 'fishshell'), ('*.fish', '*.load'), ('application/x-fish',)),
+    'FlatlineLexer': ('pip._vendor.pygments.lexers.dsls', 'Flatline', ('flatline',), (), ('text/x-flatline',)),
+    'FloScriptLexer': ('pip._vendor.pygments.lexers.floscript', 'FloScript', ('floscript', 'flo'), ('*.flo',), ()),
+    'ForthLexer': ('pip._vendor.pygments.lexers.forth', 'Forth', ('forth',), ('*.frt', '*.fs'), ('application/x-forth',)),
+    'FortranFixedLexer': ('pip._vendor.pygments.lexers.fortran', 'FortranFixed', ('fortranfixed',), ('*.f', '*.F'), ()),
+    'FortranLexer': ('pip._vendor.pygments.lexers.fortran', 'Fortran', ('fortran', 'f90'), ('*.f03', '*.f90', '*.F03', '*.F90'), ('text/x-fortran',)),
+    'FoxProLexer': ('pip._vendor.pygments.lexers.foxpro', 'FoxPro', ('foxpro', 'vfp', 'clipper', 'xbase'), ('*.PRG', '*.prg'), ()),
+    'FreeFemLexer': ('pip._vendor.pygments.lexers.freefem', 'Freefem', ('freefem',), ('*.edp',), ('text/x-freefem',)),
+    'FuncLexer': ('pip._vendor.pygments.lexers.func', 'FunC', ('func', 'fc'), ('*.fc', '*.func'), ()),
+    'FutharkLexer': ('pip._vendor.pygments.lexers.futhark', 'Futhark', ('futhark',), ('*.fut',), ('text/x-futhark',)),
+    'GAPConsoleLexer': ('pip._vendor.pygments.lexers.algebra', 'GAP session', ('gap-console', 'gap-repl'), ('*.tst',), ()),
+    'GAPLexer': ('pip._vendor.pygments.lexers.algebra', 'GAP', ('gap',), ('*.g', '*.gd', '*.gi', '*.gap'), ()),
+    'GDScriptLexer': ('pip._vendor.pygments.lexers.gdscript', 'GDScript', ('gdscript', 'gd'), ('*.gd',), ('text/x-gdscript', 'application/x-gdscript')),
+    'GLShaderLexer': ('pip._vendor.pygments.lexers.graphics', 'GLSL', ('glsl',), ('*.vert', '*.frag', '*.geo'), ('text/x-glslsrc',)),
+    'GSQLLexer': ('pip._vendor.pygments.lexers.gsql', 'GSQL', ('gsql',), ('*.gsql',), ()),
+    'GasLexer': ('pip._vendor.pygments.lexers.asm', 'GAS', ('gas', 'asm'), ('*.s', '*.S'), ('text/x-gas',)),
+    'GcodeLexer': ('pip._vendor.pygments.lexers.gcodelexer', 'g-code', ('gcode',), ('*.gcode',), ()),
+    'GenshiLexer': ('pip._vendor.pygments.lexers.templates', 'Genshi', ('genshi', 'kid', 'xml+genshi', 'xml+kid'), ('*.kid',), ('application/x-genshi', 'application/x-kid')),
+    'GenshiTextLexer': ('pip._vendor.pygments.lexers.templates', 'Genshi Text', ('genshitext',), (), ('application/x-genshi-text', 'text/x-genshi')),
+    'GettextLexer': ('pip._vendor.pygments.lexers.textfmts', 'Gettext Catalog', ('pot', 'po'), ('*.pot', '*.po'), ('application/x-gettext', 'text/x-gettext', 'text/gettext')),
+    'GherkinLexer': ('pip._vendor.pygments.lexers.testing', 'Gherkin', ('gherkin', 'cucumber'), ('*.feature',), ('text/x-gherkin',)),
+    'GnuplotLexer': ('pip._vendor.pygments.lexers.graphics', 'Gnuplot', ('gnuplot',), ('*.plot', '*.plt'), ('text/x-gnuplot',)),
+    'GoLexer': ('pip._vendor.pygments.lexers.go', 'Go', ('go', 'golang'), ('*.go',), ('text/x-gosrc',)),
+    'GoloLexer': ('pip._vendor.pygments.lexers.jvm', 'Golo', ('golo',), ('*.golo',), ()),
+    'GoodDataCLLexer': ('pip._vendor.pygments.lexers.business', 'GoodData-CL', ('gooddata-cl',), ('*.gdc',), ('text/x-gooddata-cl',)),
+    'GosuLexer': ('pip._vendor.pygments.lexers.jvm', 'Gosu', ('gosu',), ('*.gs', '*.gsx', '*.gsp', '*.vark'), ('text/x-gosu',)),
+    'GosuTemplateLexer': ('pip._vendor.pygments.lexers.jvm', 'Gosu Template', ('gst',), ('*.gst',), ('text/x-gosu-template',)),
+    'GraphQLLexer': ('pip._vendor.pygments.lexers.graphql', 'GraphQL', ('graphql',), ('*.graphql',), ()),
+    'GraphvizLexer': ('pip._vendor.pygments.lexers.graphviz', 'Graphviz', ('graphviz', 'dot'), ('*.gv', '*.dot'), ('text/x-graphviz', 'text/vnd.graphviz')),
+    'GroffLexer': ('pip._vendor.pygments.lexers.markup', 'Groff', ('groff', 'nroff', 'man'), ('*.[1-9]', '*.man', '*.1p', '*.3pm'), ('application/x-troff', 'text/troff')),
+    'GroovyLexer': ('pip._vendor.pygments.lexers.jvm', 'Groovy', ('groovy',), ('*.groovy', '*.gradle'), ('text/x-groovy',)),
+    'HLSLShaderLexer': ('pip._vendor.pygments.lexers.graphics', 'HLSL', ('hlsl',), ('*.hlsl', '*.hlsli'), ('text/x-hlsl',)),
+    'HTMLUL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'HTML+UL4', ('html+ul4',), ('*.htmlul4',), ()),
+    'HamlLexer': ('pip._vendor.pygments.lexers.html', 'Haml', ('haml',), ('*.haml',), ('text/x-haml',)),
+    'HandlebarsHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Handlebars', ('html+handlebars',), ('*.handlebars', '*.hbs'), ('text/html+handlebars', 'text/x-handlebars-template')),
+    'HandlebarsLexer': ('pip._vendor.pygments.lexers.templates', 'Handlebars', ('handlebars',), (), ()),
+    'HaskellLexer': ('pip._vendor.pygments.lexers.haskell', 'Haskell', ('haskell', 'hs'), ('*.hs',), ('text/x-haskell',)),
+    'HaxeLexer': ('pip._vendor.pygments.lexers.haxe', 'Haxe', ('haxe', 'hxsl', 'hx'), ('*.hx', '*.hxsl'), ('text/haxe', 'text/x-haxe', 'text/x-hx')),
+    'HexdumpLexer': ('pip._vendor.pygments.lexers.hexdump', 'Hexdump', ('hexdump',), (), ()),
+    'HsailLexer': ('pip._vendor.pygments.lexers.asm', 'HSAIL', ('hsail', 'hsa'), ('*.hsail',), ('text/x-hsail',)),
+    'HspecLexer': ('pip._vendor.pygments.lexers.haskell', 'Hspec', ('hspec',), ('*Spec.hs',), ()),
+    'HtmlDjangoLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Django/Jinja', ('html+django', 'html+jinja', 'htmldjango'), ('*.html.j2', '*.htm.j2', '*.xhtml.j2', '*.html.jinja2', '*.htm.jinja2', '*.xhtml.jinja2'), ('text/html+django', 'text/html+jinja')),
+    'HtmlGenshiLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Genshi', ('html+genshi', 'html+kid'), (), ('text/html+genshi',)),
+    'HtmlLexer': ('pip._vendor.pygments.lexers.html', 'HTML', ('html',), ('*.html', '*.htm', '*.xhtml', '*.xslt'), ('text/html', 'application/xhtml+xml')),
+    'HtmlPhpLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+PHP', ('html+php',), ('*.phtml',), ('application/x-php', 'application/x-httpd-php', 'application/x-httpd-php3', 'application/x-httpd-php4', 'application/x-httpd-php5')),
+    'HtmlSmartyLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Smarty', ('html+smarty',), (), ('text/html+smarty',)),
+    'HttpLexer': ('pip._vendor.pygments.lexers.textfmts', 'HTTP', ('http',), (), ()),
+    'HxmlLexer': ('pip._vendor.pygments.lexers.haxe', 'Hxml', ('haxeml', 'hxml'), ('*.hxml',), ()),
+    'HyLexer': ('pip._vendor.pygments.lexers.lisp', 'Hy', ('hylang', 'hy'), ('*.hy',), ('text/x-hy', 'application/x-hy')),
+    'HybrisLexer': ('pip._vendor.pygments.lexers.scripting', 'Hybris', ('hybris',), ('*.hyb',), ('text/x-hybris', 'application/x-hybris')),
+    'IDLLexer': ('pip._vendor.pygments.lexers.idl', 'IDL', ('idl',), ('*.pro',), ('text/idl',)),
+    'IconLexer': ('pip._vendor.pygments.lexers.unicon', 'Icon', ('icon',), ('*.icon', '*.ICON'), ()),
+    'IdrisLexer': ('pip._vendor.pygments.lexers.haskell', 'Idris', ('idris', 'idr'), ('*.idr',), ('text/x-idris',)),
+    'IgorLexer': ('pip._vendor.pygments.lexers.igor', 'Igor', ('igor', 'igorpro'), ('*.ipf',), ('text/ipf',)),
+    'Inform6Lexer': ('pip._vendor.pygments.lexers.int_fiction', 'Inform 6', ('inform6', 'i6'), ('*.inf',), ()),
+    'Inform6TemplateLexer': ('pip._vendor.pygments.lexers.int_fiction', 'Inform 6 template', ('i6t',), ('*.i6t',), ()),
+    'Inform7Lexer': ('pip._vendor.pygments.lexers.int_fiction', 'Inform 7', ('inform7', 'i7'), ('*.ni', '*.i7x'), ()),
+    'IniLexer': ('pip._vendor.pygments.lexers.configs', 'INI', ('ini', 'cfg', 'dosini'), ('*.ini', '*.cfg', '*.inf', '.editorconfig'), ('text/x-ini', 'text/inf')),
+    'IoLexer': ('pip._vendor.pygments.lexers.iolang', 'Io', ('io',), ('*.io',), ('text/x-iosrc',)),
+    'IokeLexer': ('pip._vendor.pygments.lexers.jvm', 'Ioke', ('ioke', 'ik'), ('*.ik',), ('text/x-iokesrc',)),
+    'IrcLogsLexer': ('pip._vendor.pygments.lexers.textfmts', 'IRC logs', ('irc',), ('*.weechatlog',), ('text/x-irclog',)),
+    'IsabelleLexer': ('pip._vendor.pygments.lexers.theorem', 'Isabelle', ('isabelle',), ('*.thy',), ('text/x-isabelle',)),
+    'JLexer': ('pip._vendor.pygments.lexers.j', 'J', ('j',), ('*.ijs',), ('text/x-j',)),
+    'JMESPathLexer': ('pip._vendor.pygments.lexers.jmespath', 'JMESPath', ('jmespath', 'jp'), ('*.jp',), ()),
+    'JSLTLexer': ('pip._vendor.pygments.lexers.jslt', 'JSLT', ('jslt',), ('*.jslt',), ('text/x-jslt',)),
+    'JagsLexer': ('pip._vendor.pygments.lexers.modeling', 'JAGS', ('jags',), ('*.jag', '*.bug'), ()),
+    'JanetLexer': ('pip._vendor.pygments.lexers.lisp', 'Janet', ('janet',), ('*.janet', '*.jdn'), ('text/x-janet', 'application/x-janet')),
+    'JasminLexer': ('pip._vendor.pygments.lexers.jvm', 'Jasmin', ('jasmin', 'jasminxt'), ('*.j',), ()),
+    'JavaLexer': ('pip._vendor.pygments.lexers.jvm', 'Java', ('java',), ('*.java',), ('text/x-java',)),
+    'JavascriptDjangoLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Django/Jinja', ('javascript+django', 'js+django', 'javascript+jinja', 'js+jinja'), ('*.js.j2', '*.js.jinja2'), ('application/x-javascript+django', 'application/x-javascript+jinja', 'text/x-javascript+django', 'text/x-javascript+jinja', 'text/javascript+django', 'text/javascript+jinja')),
+    'JavascriptErbLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Ruby', ('javascript+ruby', 'js+ruby', 'javascript+erb', 'js+erb'), (), ('application/x-javascript+ruby', 'text/x-javascript+ruby', 'text/javascript+ruby')),
+    'JavascriptGenshiLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Genshi Text', ('js+genshitext', 'js+genshi', 'javascript+genshitext', 'javascript+genshi'), (), ('application/x-javascript+genshi', 'text/x-javascript+genshi', 'text/javascript+genshi')),
+    'JavascriptLexer': ('pip._vendor.pygments.lexers.javascript', 'JavaScript', ('javascript', 'js'), ('*.js', '*.jsm', '*.mjs', '*.cjs'), ('application/javascript', 'application/x-javascript', 'text/x-javascript', 'text/javascript')),
+    'JavascriptPhpLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+PHP', ('javascript+php', 'js+php'), (), ('application/x-javascript+php', 'text/x-javascript+php', 'text/javascript+php')),
+    'JavascriptSmartyLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Smarty', ('javascript+smarty', 'js+smarty'), (), ('application/x-javascript+smarty', 'text/x-javascript+smarty', 'text/javascript+smarty')),
+    'JavascriptUL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'Javascript+UL4', ('js+ul4',), ('*.jsul4',), ()),
+    'JclLexer': ('pip._vendor.pygments.lexers.scripting', 'JCL', ('jcl',), ('*.jcl',), ('text/x-jcl',)),
+    'JsgfLexer': ('pip._vendor.pygments.lexers.grammar_notation', 'JSGF', ('jsgf',), ('*.jsgf',), ('application/jsgf', 'application/x-jsgf', 'text/jsgf')),
+    'JsonBareObjectLexer': ('pip._vendor.pygments.lexers.data', 'JSONBareObject', (), (), ()),
+    'JsonLdLexer': ('pip._vendor.pygments.lexers.data', 'JSON-LD', ('jsonld', 'json-ld'), ('*.jsonld',), ('application/ld+json',)),
+    'JsonLexer': ('pip._vendor.pygments.lexers.data', 'JSON', ('json', 'json-object'), ('*.json', '*.jsonl', '*.ndjson', 'Pipfile.lock'), ('application/json', 'application/json-object', 'application/x-ndjson', 'application/jsonl', 'application/json-seq')),
+    'JsonnetLexer': ('pip._vendor.pygments.lexers.jsonnet', 'Jsonnet', ('jsonnet',), ('*.jsonnet', '*.libsonnet'), ()),
+    'JspLexer': ('pip._vendor.pygments.lexers.templates', 'Java Server Page', ('jsp',), ('*.jsp',), ('application/x-jsp',)),
+    'JsxLexer': ('pip._vendor.pygments.lexers.jsx', 'JSX', ('jsx', 'react'), ('*.jsx', '*.react'), ('text/jsx', 'text/typescript-jsx')),
+    'JuliaConsoleLexer': ('pip._vendor.pygments.lexers.julia', 'Julia console', ('jlcon', 'julia-repl'), (), ()),
+    'JuliaLexer': ('pip._vendor.pygments.lexers.julia', 'Julia', ('julia', 'jl'), ('*.jl',), ('text/x-julia', 'application/x-julia')),
+    'JuttleLexer': ('pip._vendor.pygments.lexers.javascript', 'Juttle', ('juttle',), ('*.juttle',), ('application/juttle', 'application/x-juttle', 'text/x-juttle', 'text/juttle')),
+    'KLexer': ('pip._vendor.pygments.lexers.q', 'K', ('k',), ('*.k',), ()),
+    'KalLexer': ('pip._vendor.pygments.lexers.javascript', 'Kal', ('kal',), ('*.kal',), ('text/kal', 'application/kal')),
+    'KconfigLexer': ('pip._vendor.pygments.lexers.configs', 'Kconfig', ('kconfig', 'menuconfig', 'linux-config', 'kernel-config'), ('Kconfig*', '*Config.in*', 'external.in*', 'standard-modules.in'), ('text/x-kconfig',)),
+    'KernelLogLexer': ('pip._vendor.pygments.lexers.textfmts', 'Kernel log', ('kmsg', 'dmesg'), ('*.kmsg', '*.dmesg'), ()),
+    'KokaLexer': ('pip._vendor.pygments.lexers.haskell', 'Koka', ('koka',), ('*.kk', '*.kki'), ('text/x-koka',)),
+    'KotlinLexer': ('pip._vendor.pygments.lexers.jvm', 'Kotlin', ('kotlin',), ('*.kt', '*.kts'), ('text/x-kotlin',)),
+    'KuinLexer': ('pip._vendor.pygments.lexers.kuin', 'Kuin', ('kuin',), ('*.kn',), ()),
+    'KustoLexer': ('pip._vendor.pygments.lexers.kusto', 'Kusto', ('kql', 'kusto'), ('*.kql', '*.kusto', '.csl'), ()),
+    'LSLLexer': ('pip._vendor.pygments.lexers.scripting', 'LSL', ('lsl',), ('*.lsl',), ('text/x-lsl',)),
+    'LassoCssLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Lasso', ('css+lasso',), (), ('text/css+lasso',)),
+    'LassoHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Lasso', ('html+lasso',), (), ('text/html+lasso', 'application/x-httpd-lasso', 'application/x-httpd-lasso[89]')),
+    'LassoJavascriptLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Lasso', ('javascript+lasso', 'js+lasso'), (), ('application/x-javascript+lasso', 'text/x-javascript+lasso', 'text/javascript+lasso')),
+    'LassoLexer': ('pip._vendor.pygments.lexers.javascript', 'Lasso', ('lasso', 'lassoscript'), ('*.lasso', '*.lasso[89]'), ('text/x-lasso',)),
+    'LassoXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Lasso', ('xml+lasso',), (), ('application/xml+lasso',)),
+    'LdaprcLexer': ('pip._vendor.pygments.lexers.ldap', 'LDAP configuration file', ('ldapconf', 'ldaprc'), ('.ldaprc', 'ldaprc', 'ldap.conf'), ('text/x-ldapconf',)),
+    'LdifLexer': ('pip._vendor.pygments.lexers.ldap', 'LDIF', ('ldif',), ('*.ldif',), ('text/x-ldif',)),
+    'Lean3Lexer': ('pip._vendor.pygments.lexers.lean', 'Lean', ('lean', 'lean3'), ('*.lean',), ('text/x-lean', 'text/x-lean3')),
+    'Lean4Lexer': ('pip._vendor.pygments.lexers.lean', 'Lean4', ('lean4',), ('*.lean',), ('text/x-lean4',)),
+    'LessCssLexer': ('pip._vendor.pygments.lexers.css', 'LessCss', ('less',), ('*.less',), ('text/x-less-css',)),
+    'LighttpdConfLexer': ('pip._vendor.pygments.lexers.configs', 'Lighttpd configuration file', ('lighttpd', 'lighty'), ('lighttpd.conf',), ('text/x-lighttpd-conf',)),
+    'LilyPondLexer': ('pip._vendor.pygments.lexers.lilypond', 'LilyPond', ('lilypond',), ('*.ly',), ()),
+    'LimboLexer': ('pip._vendor.pygments.lexers.inferno', 'Limbo', ('limbo',), ('*.b',), ('text/limbo',)),
+    'LiquidLexer': ('pip._vendor.pygments.lexers.templates', 'liquid', ('liquid',), ('*.liquid',), ()),
+    'LiterateAgdaLexer': ('pip._vendor.pygments.lexers.haskell', 'Literate Agda', ('literate-agda', 'lagda'), ('*.lagda',), ('text/x-literate-agda',)),
+    'LiterateCryptolLexer': ('pip._vendor.pygments.lexers.haskell', 'Literate Cryptol', ('literate-cryptol', 'lcryptol', 'lcry'), ('*.lcry',), ('text/x-literate-cryptol',)),
+    'LiterateHaskellLexer': ('pip._vendor.pygments.lexers.haskell', 'Literate Haskell', ('literate-haskell', 'lhaskell', 'lhs'), ('*.lhs',), ('text/x-literate-haskell',)),
+    'LiterateIdrisLexer': ('pip._vendor.pygments.lexers.haskell', 'Literate Idris', ('literate-idris', 'lidris', 'lidr'), ('*.lidr',), ('text/x-literate-idris',)),
+    'LiveScriptLexer': ('pip._vendor.pygments.lexers.javascript', 'LiveScript', ('livescript', 'live-script'), ('*.ls',), ('text/livescript',)),
+    'LlvmLexer': ('pip._vendor.pygments.lexers.asm', 'LLVM', ('llvm',), ('*.ll',), ('text/x-llvm',)),
+    'LlvmMirBodyLexer': ('pip._vendor.pygments.lexers.asm', 'LLVM-MIR Body', ('llvm-mir-body',), (), ()),
+    'LlvmMirLexer': ('pip._vendor.pygments.lexers.asm', 'LLVM-MIR', ('llvm-mir',), ('*.mir',), ()),
+    'LogosLexer': ('pip._vendor.pygments.lexers.objective', 'Logos', ('logos',), ('*.x', '*.xi', '*.xm', '*.xmi'), ('text/x-logos',)),
+    'LogtalkLexer': ('pip._vendor.pygments.lexers.prolog', 'Logtalk', ('logtalk',), ('*.lgt', '*.logtalk'), ('text/x-logtalk',)),
+    'LuaLexer': ('pip._vendor.pygments.lexers.scripting', 'Lua', ('lua',), ('*.lua', '*.wlua'), ('text/x-lua', 'application/x-lua')),
+    'LuauLexer': ('pip._vendor.pygments.lexers.scripting', 'Luau', ('luau',), ('*.luau',), ()),
+    'MCFunctionLexer': ('pip._vendor.pygments.lexers.minecraft', 'MCFunction', ('mcfunction', 'mcf'), ('*.mcfunction',), ('text/mcfunction',)),
+    'MCSchemaLexer': ('pip._vendor.pygments.lexers.minecraft', 'MCSchema', ('mcschema',), ('*.mcschema',), ('text/mcschema',)),
+    'MIMELexer': ('pip._vendor.pygments.lexers.mime', 'MIME', ('mime',), (), ('multipart/mixed', 'multipart/related', 'multipart/alternative')),
+    'MIPSLexer': ('pip._vendor.pygments.lexers.mips', 'MIPS', ('mips',), ('*.mips', '*.MIPS'), ()),
+    'MOOCodeLexer': ('pip._vendor.pygments.lexers.scripting', 'MOOCode', ('moocode', 'moo'), ('*.moo',), ('text/x-moocode',)),
+    'MSDOSSessionLexer': ('pip._vendor.pygments.lexers.shell', 'MSDOS Session', ('doscon',), (), ()),
+    'Macaulay2Lexer': ('pip._vendor.pygments.lexers.macaulay2', 'Macaulay2', ('macaulay2',), ('*.m2',), ()),
+    'MakefileLexer': ('pip._vendor.pygments.lexers.make', 'Makefile', ('make', 'makefile', 'mf', 'bsdmake'), ('*.mak', '*.mk', 'Makefile', 'makefile', 'Makefile.*', 'GNUmakefile'), ('text/x-makefile',)),
+    'MakoCssLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Mako', ('css+mako',), (), ('text/css+mako',)),
+    'MakoHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Mako', ('html+mako',), (), ('text/html+mako',)),
+    'MakoJavascriptLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Mako', ('javascript+mako', 'js+mako'), (), ('application/x-javascript+mako', 'text/x-javascript+mako', 'text/javascript+mako')),
+    'MakoLexer': ('pip._vendor.pygments.lexers.templates', 'Mako', ('mako',), ('*.mao',), ('application/x-mako',)),
+    'MakoXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Mako', ('xml+mako',), (), ('application/xml+mako',)),
+    'MaqlLexer': ('pip._vendor.pygments.lexers.business', 'MAQL', ('maql',), ('*.maql',), ('text/x-gooddata-maql', 'application/x-gooddata-maql')),
+    'MarkdownLexer': ('pip._vendor.pygments.lexers.markup', 'Markdown', ('markdown', 'md'), ('*.md', '*.markdown'), ('text/x-markdown',)),
+    'MaskLexer': ('pip._vendor.pygments.lexers.javascript', 'Mask', ('mask',), ('*.mask',), ('text/x-mask',)),
+    'MasonLexer': ('pip._vendor.pygments.lexers.templates', 'Mason', ('mason',), ('*.m', '*.mhtml', '*.mc', '*.mi', 'autohandler', 'dhandler'), ('application/x-mason',)),
+    'MathematicaLexer': ('pip._vendor.pygments.lexers.algebra', 'Mathematica', ('mathematica', 'mma', 'nb'), ('*.nb', '*.cdf', '*.nbp', '*.ma'), ('application/mathematica', 'application/vnd.wolfram.mathematica', 'application/vnd.wolfram.mathematica.package', 'application/vnd.wolfram.cdf')),
+    'MatlabLexer': ('pip._vendor.pygments.lexers.matlab', 'Matlab', ('matlab',), ('*.m',), ('text/matlab',)),
+    'MatlabSessionLexer': ('pip._vendor.pygments.lexers.matlab', 'Matlab session', ('matlabsession',), (), ()),
+    'MaximaLexer': ('pip._vendor.pygments.lexers.maxima', 'Maxima', ('maxima', 'macsyma'), ('*.mac', '*.max'), ()),
+    'MesonLexer': ('pip._vendor.pygments.lexers.meson', 'Meson', ('meson', 'meson.build'), ('meson.build', 'meson_options.txt'), ('text/x-meson',)),
+    'MiniDLexer': ('pip._vendor.pygments.lexers.d', 'MiniD', ('minid',), (), ('text/x-minidsrc',)),
+    'MiniScriptLexer': ('pip._vendor.pygments.lexers.scripting', 'MiniScript', ('miniscript', 'ms'), ('*.ms',), ('text/x-minicript', 'application/x-miniscript')),
+    'ModelicaLexer': ('pip._vendor.pygments.lexers.modeling', 'Modelica', ('modelica',), ('*.mo',), ('text/x-modelica',)),
+    'Modula2Lexer': ('pip._vendor.pygments.lexers.modula2', 'Modula-2', ('modula2', 'm2'), ('*.def', '*.mod'), ('text/x-modula2',)),
+    'MoinWikiLexer': ('pip._vendor.pygments.lexers.markup', 'MoinMoin/Trac Wiki markup', ('trac-wiki', 'moin'), (), ('text/x-trac-wiki',)),
+    'MojoLexer': ('pip._vendor.pygments.lexers.mojo', 'Mojo', ('mojo', '🔥'), ('*.mojo', '*.🔥'), ('text/x-mojo', 'application/x-mojo')),
+    'MonkeyLexer': ('pip._vendor.pygments.lexers.basic', 'Monkey', ('monkey',), ('*.monkey',), ('text/x-monkey',)),
+    'MonteLexer': ('pip._vendor.pygments.lexers.monte', 'Monte', ('monte',), ('*.mt',), ()),
+    'MoonScriptLexer': ('pip._vendor.pygments.lexers.scripting', 'MoonScript', ('moonscript', 'moon'), ('*.moon',), ('text/x-moonscript', 'application/x-moonscript')),
+    'MoselLexer': ('pip._vendor.pygments.lexers.mosel', 'Mosel', ('mosel',), ('*.mos',), ()),
+    'MozPreprocCssLexer': ('pip._vendor.pygments.lexers.markup', 'CSS+mozpreproc', ('css+mozpreproc',), ('*.css.in',), ()),
+    'MozPreprocHashLexer': ('pip._vendor.pygments.lexers.markup', 'mozhashpreproc', ('mozhashpreproc',), (), ()),
+    'MozPreprocJavascriptLexer': ('pip._vendor.pygments.lexers.markup', 'Javascript+mozpreproc', ('javascript+mozpreproc',), ('*.js.in',), ()),
+    'MozPreprocPercentLexer': ('pip._vendor.pygments.lexers.markup', 'mozpercentpreproc', ('mozpercentpreproc',), (), ()),
+    'MozPreprocXulLexer': ('pip._vendor.pygments.lexers.markup', 'XUL+mozpreproc', ('xul+mozpreproc',), ('*.xul.in',), ()),
+    'MqlLexer': ('pip._vendor.pygments.lexers.c_like', 'MQL', ('mql', 'mq4', 'mq5', 'mql4', 'mql5'), ('*.mq4', '*.mq5', '*.mqh'), ('text/x-mql',)),
+    'MscgenLexer': ('pip._vendor.pygments.lexers.dsls', 'Mscgen', ('mscgen', 'msc'), ('*.msc',), ()),
+    'MuPADLexer': ('pip._vendor.pygments.lexers.algebra', 'MuPAD', ('mupad',), ('*.mu',), ()),
+    'MxmlLexer': ('pip._vendor.pygments.lexers.actionscript', 'MXML', ('mxml',), ('*.mxml',), ()),
+    'MySqlLexer': ('pip._vendor.pygments.lexers.sql', 'MySQL', ('mysql',), (), ('text/x-mysql',)),
+    'MyghtyCssLexer': ('pip._vendor.pygments.lexers.templates', 'CSS+Myghty', ('css+myghty',), (), ('text/css+myghty',)),
+    'MyghtyHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Myghty', ('html+myghty',), (), ('text/html+myghty',)),
+    'MyghtyJavascriptLexer': ('pip._vendor.pygments.lexers.templates', 'JavaScript+Myghty', ('javascript+myghty', 'js+myghty'), (), ('application/x-javascript+myghty', 'text/x-javascript+myghty', 'text/javascript+mygthy')),
+    'MyghtyLexer': ('pip._vendor.pygments.lexers.templates', 'Myghty', ('myghty',), ('*.myt', 'autodelegate'), ('application/x-myghty',)),
+    'MyghtyXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Myghty', ('xml+myghty',), (), ('application/xml+myghty',)),
+    'NCLLexer': ('pip._vendor.pygments.lexers.ncl', 'NCL', ('ncl',), ('*.ncl',), ('text/ncl',)),
+    'NSISLexer': ('pip._vendor.pygments.lexers.installers', 'NSIS', ('nsis', 'nsi', 'nsh'), ('*.nsi', '*.nsh'), ('text/x-nsis',)),
+    'NasmLexer': ('pip._vendor.pygments.lexers.asm', 'NASM', ('nasm',), ('*.asm', '*.ASM', '*.nasm'), ('text/x-nasm',)),
+    'NasmObjdumpLexer': ('pip._vendor.pygments.lexers.asm', 'objdump-nasm', ('objdump-nasm',), ('*.objdump-intel',), ('text/x-nasm-objdump',)),
+    'NemerleLexer': ('pip._vendor.pygments.lexers.dotnet', 'Nemerle', ('nemerle',), ('*.n',), ('text/x-nemerle',)),
+    'NesCLexer': ('pip._vendor.pygments.lexers.c_like', 'nesC', ('nesc',), ('*.nc',), ('text/x-nescsrc',)),
+    'NestedTextLexer': ('pip._vendor.pygments.lexers.configs', 'NestedText', ('nestedtext', 'nt'), ('*.nt',), ()),
+    'NewLispLexer': ('pip._vendor.pygments.lexers.lisp', 'NewLisp', ('newlisp',), ('*.lsp', '*.nl', '*.kif'), ('text/x-newlisp', 'application/x-newlisp')),
+    'NewspeakLexer': ('pip._vendor.pygments.lexers.smalltalk', 'Newspeak', ('newspeak',), ('*.ns2',), ('text/x-newspeak',)),
+    'NginxConfLexer': ('pip._vendor.pygments.lexers.configs', 'Nginx configuration file', ('nginx',), ('nginx.conf',), ('text/x-nginx-conf',)),
+    'NimrodLexer': ('pip._vendor.pygments.lexers.nimrod', 'Nimrod', ('nimrod', 'nim'), ('*.nim', '*.nimrod'), ('text/x-nim',)),
+    'NitLexer': ('pip._vendor.pygments.lexers.nit', 'Nit', ('nit',), ('*.nit',), ()),
+    'NixLexer': ('pip._vendor.pygments.lexers.nix', 'Nix', ('nixos', 'nix'), ('*.nix',), ('text/x-nix',)),
+    'NodeConsoleLexer': ('pip._vendor.pygments.lexers.javascript', 'Node.js REPL console session', ('nodejsrepl',), (), ('text/x-nodejsrepl',)),
+    'NotmuchLexer': ('pip._vendor.pygments.lexers.textfmts', 'Notmuch', ('notmuch',), (), ()),
+    'NuSMVLexer': ('pip._vendor.pygments.lexers.smv', 'NuSMV', ('nusmv',), ('*.smv',), ()),
+    'NumPyLexer': ('pip._vendor.pygments.lexers.python', 'NumPy', ('numpy',), (), ()),
+    'ObjdumpLexer': ('pip._vendor.pygments.lexers.asm', 'objdump', ('objdump',), ('*.objdump',), ('text/x-objdump',)),
+    'ObjectiveCLexer': ('pip._vendor.pygments.lexers.objective', 'Objective-C', ('objective-c', 'objectivec', 'obj-c', 'objc'), ('*.m', '*.h'), ('text/x-objective-c',)),
+    'ObjectiveCppLexer': ('pip._vendor.pygments.lexers.objective', 'Objective-C++', ('objective-c++', 'objectivec++', 'obj-c++', 'objc++'), ('*.mm', '*.hh'), ('text/x-objective-c++',)),
+    'ObjectiveJLexer': ('pip._vendor.pygments.lexers.javascript', 'Objective-J', ('objective-j', 'objectivej', 'obj-j', 'objj'), ('*.j',), ('text/x-objective-j',)),
+    'OcamlLexer': ('pip._vendor.pygments.lexers.ml', 'OCaml', ('ocaml',), ('*.ml', '*.mli', '*.mll', '*.mly'), ('text/x-ocaml',)),
+    'OctaveLexer': ('pip._vendor.pygments.lexers.matlab', 'Octave', ('octave',), ('*.m',), ('text/octave',)),
+    'OdinLexer': ('pip._vendor.pygments.lexers.archetype', 'ODIN', ('odin',), ('*.odin',), ('text/odin',)),
+    'OmgIdlLexer': ('pip._vendor.pygments.lexers.c_like', 'OMG Interface Definition Language', ('omg-idl',), ('*.idl', '*.pidl'), ()),
+    'OocLexer': ('pip._vendor.pygments.lexers.ooc', 'Ooc', ('ooc',), ('*.ooc',), ('text/x-ooc',)),
+    'OpaLexer': ('pip._vendor.pygments.lexers.ml', 'Opa', ('opa',), ('*.opa',), ('text/x-opa',)),
+    'OpenEdgeLexer': ('pip._vendor.pygments.lexers.business', 'OpenEdge ABL', ('openedge', 'abl', 'progress'), ('*.p', '*.cls'), ('text/x-openedge', 'application/x-openedge')),
+    'OpenScadLexer': ('pip._vendor.pygments.lexers.openscad', 'OpenSCAD', ('openscad',), ('*.scad',), ('application/x-openscad',)),
+    'OrgLexer': ('pip._vendor.pygments.lexers.markup', 'Org Mode', ('org', 'orgmode', 'org-mode'), ('*.org',), ('text/org',)),
+    'OutputLexer': ('pip._vendor.pygments.lexers.special', 'Text output', ('output',), (), ()),
+    'PacmanConfLexer': ('pip._vendor.pygments.lexers.configs', 'PacmanConf', ('pacmanconf',), ('pacman.conf',), ()),
+    'PanLexer': ('pip._vendor.pygments.lexers.dsls', 'Pan', ('pan',), ('*.pan',), ()),
+    'ParaSailLexer': ('pip._vendor.pygments.lexers.parasail', 'ParaSail', ('parasail',), ('*.psi', '*.psl'), ('text/x-parasail',)),
+    'PawnLexer': ('pip._vendor.pygments.lexers.pawn', 'Pawn', ('pawn',), ('*.p', '*.pwn', '*.inc'), ('text/x-pawn',)),
+    'PegLexer': ('pip._vendor.pygments.lexers.grammar_notation', 'PEG', ('peg',), ('*.peg',), ('text/x-peg',)),
+    'Perl6Lexer': ('pip._vendor.pygments.lexers.perl', 'Perl6', ('perl6', 'pl6', 'raku'), ('*.pl', '*.pm', '*.nqp', '*.p6', '*.6pl', '*.p6l', '*.pl6', '*.6pm', '*.p6m', '*.pm6', '*.t', '*.raku', '*.rakumod', '*.rakutest', '*.rakudoc'), ('text/x-perl6', 'application/x-perl6')),
+    'PerlLexer': ('pip._vendor.pygments.lexers.perl', 'Perl', ('perl', 'pl'), ('*.pl', '*.pm', '*.t', '*.perl'), ('text/x-perl', 'application/x-perl')),
+    'PhixLexer': ('pip._vendor.pygments.lexers.phix', 'Phix', ('phix',), ('*.exw',), ('text/x-phix',)),
+    'PhpLexer': ('pip._vendor.pygments.lexers.php', 'PHP', ('php', 'php3', 'php4', 'php5'), ('*.php', '*.php[345]', '*.inc'), ('text/x-php',)),
+    'PigLexer': ('pip._vendor.pygments.lexers.jvm', 'Pig', ('pig',), ('*.pig',), ('text/x-pig',)),
+    'PikeLexer': ('pip._vendor.pygments.lexers.c_like', 'Pike', ('pike',), ('*.pike', '*.pmod'), ('text/x-pike',)),
+    'PkgConfigLexer': ('pip._vendor.pygments.lexers.configs', 'PkgConfig', ('pkgconfig',), ('*.pc',), ()),
+    'PlPgsqlLexer': ('pip._vendor.pygments.lexers.sql', 'PL/pgSQL', ('plpgsql',), (), ('text/x-plpgsql',)),
+    'PointlessLexer': ('pip._vendor.pygments.lexers.pointless', 'Pointless', ('pointless',), ('*.ptls',), ()),
+    'PonyLexer': ('pip._vendor.pygments.lexers.pony', 'Pony', ('pony',), ('*.pony',), ()),
+    'PortugolLexer': ('pip._vendor.pygments.lexers.pascal', 'Portugol', ('portugol',), ('*.alg', '*.portugol'), ()),
+    'PostScriptLexer': ('pip._vendor.pygments.lexers.graphics', 'PostScript', ('postscript', 'postscr'), ('*.ps', '*.eps'), ('application/postscript',)),
+    'PostgresConsoleLexer': ('pip._vendor.pygments.lexers.sql', 'PostgreSQL console (psql)', ('psql', 'postgresql-console', 'postgres-console'), (), ('text/x-postgresql-psql',)),
+    'PostgresExplainLexer': ('pip._vendor.pygments.lexers.sql', 'PostgreSQL EXPLAIN dialect', ('postgres-explain',), ('*.explain',), ('text/x-postgresql-explain',)),
+    'PostgresLexer': ('pip._vendor.pygments.lexers.sql', 'PostgreSQL SQL dialect', ('postgresql', 'postgres'), (), ('text/x-postgresql',)),
+    'PovrayLexer': ('pip._vendor.pygments.lexers.graphics', 'POVRay', ('pov',), ('*.pov', '*.inc'), ('text/x-povray',)),
+    'PowerShellLexer': ('pip._vendor.pygments.lexers.shell', 'PowerShell', ('powershell', 'pwsh', 'posh', 'ps1', 'psm1'), ('*.ps1', '*.psm1'), ('text/x-powershell',)),
+    'PowerShellSessionLexer': ('pip._vendor.pygments.lexers.shell', 'PowerShell Session', ('pwsh-session', 'ps1con'), (), ()),
+    'PraatLexer': ('pip._vendor.pygments.lexers.praat', 'Praat', ('praat',), ('*.praat', '*.proc', '*.psc'), ()),
+    'ProcfileLexer': ('pip._vendor.pygments.lexers.procfile', 'Procfile', ('procfile',), ('Procfile',), ()),
+    'PrologLexer': ('pip._vendor.pygments.lexers.prolog', 'Prolog', ('prolog',), ('*.ecl', '*.prolog', '*.pro', '*.pl'), ('text/x-prolog',)),
+    'PromQLLexer': ('pip._vendor.pygments.lexers.promql', 'PromQL', ('promql',), ('*.promql',), ()),
+    'PromelaLexer': ('pip._vendor.pygments.lexers.c_like', 'Promela', ('promela',), ('*.pml', '*.prom', '*.prm', '*.promela', '*.pr', '*.pm'), ('text/x-promela',)),
+    'PropertiesLexer': ('pip._vendor.pygments.lexers.configs', 'Properties', ('properties', 'jproperties'), ('*.properties',), ('text/x-java-properties',)),
+    'ProtoBufLexer': ('pip._vendor.pygments.lexers.dsls', 'Protocol Buffer', ('protobuf', 'proto'), ('*.proto',), ()),
+    'PrqlLexer': ('pip._vendor.pygments.lexers.prql', 'PRQL', ('prql',), ('*.prql',), ('application/prql', 'application/x-prql')),
+    'PsyshConsoleLexer': ('pip._vendor.pygments.lexers.php', 'PsySH console session for PHP', ('psysh',), (), ()),
+    'PtxLexer': ('pip._vendor.pygments.lexers.ptx', 'PTX', ('ptx',), ('*.ptx',), ('text/x-ptx',)),
+    'PugLexer': ('pip._vendor.pygments.lexers.html', 'Pug', ('pug', 'jade'), ('*.pug', '*.jade'), ('text/x-pug', 'text/x-jade')),
+    'PuppetLexer': ('pip._vendor.pygments.lexers.dsls', 'Puppet', ('puppet',), ('*.pp',), ()),
+    'PyPyLogLexer': ('pip._vendor.pygments.lexers.console', 'PyPy Log', ('pypylog', 'pypy'), ('*.pypylog',), ('application/x-pypylog',)),
+    'Python2Lexer': ('pip._vendor.pygments.lexers.python', 'Python 2.x', ('python2', 'py2'), (), ('text/x-python2', 'application/x-python2')),
+    'Python2TracebackLexer': ('pip._vendor.pygments.lexers.python', 'Python 2.x Traceback', ('py2tb',), ('*.py2tb',), ('text/x-python2-traceback',)),
+    'PythonConsoleLexer': ('pip._vendor.pygments.lexers.python', 'Python console session', ('pycon', 'python-console'), (), ('text/x-python-doctest',)),
+    'PythonLexer': ('pip._vendor.pygments.lexers.python', 'Python', ('python', 'py', 'sage', 'python3', 'py3', 'bazel', 'starlark'), ('*.py', '*.pyw', '*.pyi', '*.jy', '*.sage', '*.sc', 'SConstruct', 'SConscript', '*.bzl', 'BUCK', 'BUILD', 'BUILD.bazel', 'WORKSPACE', '*.tac'), ('text/x-python', 'application/x-python', 'text/x-python3', 'application/x-python3')),
+    'PythonTracebackLexer': ('pip._vendor.pygments.lexers.python', 'Python Traceback', ('pytb', 'py3tb'), ('*.pytb', '*.py3tb'), ('text/x-python-traceback', 'text/x-python3-traceback')),
+    'PythonUL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'Python+UL4', ('py+ul4',), ('*.pyul4',), ()),
+    'QBasicLexer': ('pip._vendor.pygments.lexers.basic', 'QBasic', ('qbasic', 'basic'), ('*.BAS', '*.bas'), ('text/basic',)),
+    'QLexer': ('pip._vendor.pygments.lexers.q', 'Q', ('q',), ('*.q',), ()),
+    'QVToLexer': ('pip._vendor.pygments.lexers.qvt', 'QVTO', ('qvto', 'qvt'), ('*.qvto',), ()),
+    'QlikLexer': ('pip._vendor.pygments.lexers.qlik', 'Qlik', ('qlik', 'qlikview', 'qliksense', 'qlikscript'), ('*.qvs', '*.qvw'), ()),
+    'QmlLexer': ('pip._vendor.pygments.lexers.webmisc', 'QML', ('qml', 'qbs'), ('*.qml', '*.qbs'), ('application/x-qml', 'application/x-qt.qbs+qml')),
+    'RConsoleLexer': ('pip._vendor.pygments.lexers.r', 'RConsole', ('rconsole', 'rout'), ('*.Rout',), ()),
+    'RNCCompactLexer': ('pip._vendor.pygments.lexers.rnc', 'Relax-NG Compact', ('rng-compact', 'rnc'), ('*.rnc',), ()),
+    'RPMSpecLexer': ('pip._vendor.pygments.lexers.installers', 'RPMSpec', ('spec',), ('*.spec',), ('text/x-rpm-spec',)),
+    'RacketLexer': ('pip._vendor.pygments.lexers.lisp', 'Racket', ('racket', 'rkt'), ('*.rkt', '*.rktd', '*.rktl'), ('text/x-racket', 'application/x-racket')),
+    'RagelCLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in C Host', ('ragel-c',), ('*.rl',), ()),
+    'RagelCppLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in CPP Host', ('ragel-cpp',), ('*.rl',), ()),
+    'RagelDLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in D Host', ('ragel-d',), ('*.rl',), ()),
+    'RagelEmbeddedLexer': ('pip._vendor.pygments.lexers.parsers', 'Embedded Ragel', ('ragel-em',), ('*.rl',), ()),
+    'RagelJavaLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in Java Host', ('ragel-java',), ('*.rl',), ()),
+    'RagelLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel', ('ragel',), (), ()),
+    'RagelObjectiveCLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in Objective C Host', ('ragel-objc',), ('*.rl',), ()),
+    'RagelRubyLexer': ('pip._vendor.pygments.lexers.parsers', 'Ragel in Ruby Host', ('ragel-ruby', 'ragel-rb'), ('*.rl',), ()),
+    'RawTokenLexer': ('pip._vendor.pygments.lexers.special', 'Raw token data', (), (), ('application/x-pygments-tokens',)),
+    'RdLexer': ('pip._vendor.pygments.lexers.r', 'Rd', ('rd',), ('*.Rd',), ('text/x-r-doc',)),
+    'ReasonLexer': ('pip._vendor.pygments.lexers.ml', 'ReasonML', ('reasonml', 'reason'), ('*.re', '*.rei'), ('text/x-reasonml',)),
+    'RebolLexer': ('pip._vendor.pygments.lexers.rebol', 'REBOL', ('rebol',), ('*.r', '*.r3', '*.reb'), ('text/x-rebol',)),
+    'RedLexer': ('pip._vendor.pygments.lexers.rebol', 'Red', ('red', 'red/system'), ('*.red', '*.reds'), ('text/x-red', 'text/x-red-system')),
+    'RedcodeLexer': ('pip._vendor.pygments.lexers.esoteric', 'Redcode', ('redcode',), ('*.cw',), ()),
+    'RegeditLexer': ('pip._vendor.pygments.lexers.configs', 'reg', ('registry',), ('*.reg',), ('text/x-windows-registry',)),
+    'ResourceLexer': ('pip._vendor.pygments.lexers.resource', 'ResourceBundle', ('resourcebundle', 'resource'), (), ()),
+    'RexxLexer': ('pip._vendor.pygments.lexers.scripting', 'Rexx', ('rexx', 'arexx'), ('*.rexx', '*.rex', '*.rx', '*.arexx'), ('text/x-rexx',)),
+    'RhtmlLexer': ('pip._vendor.pygments.lexers.templates', 'RHTML', ('rhtml', 'html+erb', 'html+ruby'), ('*.rhtml',), ('text/html+ruby',)),
+    'RideLexer': ('pip._vendor.pygments.lexers.ride', 'Ride', ('ride',), ('*.ride',), ('text/x-ride',)),
+    'RitaLexer': ('pip._vendor.pygments.lexers.rita', 'Rita', ('rita',), ('*.rita',), ('text/rita',)),
+    'RoboconfGraphLexer': ('pip._vendor.pygments.lexers.roboconf', 'Roboconf Graph', ('roboconf-graph',), ('*.graph',), ()),
+    'RoboconfInstancesLexer': ('pip._vendor.pygments.lexers.roboconf', 'Roboconf Instances', ('roboconf-instances',), ('*.instances',), ()),
+    'RobotFrameworkLexer': ('pip._vendor.pygments.lexers.robotframework', 'RobotFramework', ('robotframework',), ('*.robot', '*.resource'), ('text/x-robotframework',)),
+    'RqlLexer': ('pip._vendor.pygments.lexers.sql', 'RQL', ('rql',), ('*.rql',), ('text/x-rql',)),
+    'RslLexer': ('pip._vendor.pygments.lexers.dsls', 'RSL', ('rsl',), ('*.rsl',), ('text/rsl',)),
+    'RstLexer': ('pip._vendor.pygments.lexers.markup', 'reStructuredText', ('restructuredtext', 'rst', 'rest'), ('*.rst', '*.rest'), ('text/x-rst', 'text/prs.fallenstein.rst')),
+    'RtsLexer': ('pip._vendor.pygments.lexers.trafficscript', 'TrafficScript', ('trafficscript', 'rts'), ('*.rts',), ()),
+    'RubyConsoleLexer': ('pip._vendor.pygments.lexers.ruby', 'Ruby irb session', ('rbcon', 'irb'), (), ('text/x-ruby-shellsession',)),
+    'RubyLexer': ('pip._vendor.pygments.lexers.ruby', 'Ruby', ('ruby', 'rb', 'duby'), ('*.rb', '*.rbw', 'Rakefile', '*.rake', '*.gemspec', '*.rbx', '*.duby', 'Gemfile', 'Vagrantfile'), ('text/x-ruby', 'application/x-ruby')),
+    'RustLexer': ('pip._vendor.pygments.lexers.rust', 'Rust', ('rust', 'rs'), ('*.rs', '*.rs.in'), ('text/rust', 'text/x-rust')),
+    'SASLexer': ('pip._vendor.pygments.lexers.sas', 'SAS', ('sas',), ('*.SAS', '*.sas'), ('text/x-sas', 'text/sas', 'application/x-sas')),
+    'SLexer': ('pip._vendor.pygments.lexers.r', 'S', ('splus', 's', 'r'), ('*.S', '*.R', '.Rhistory', '.Rprofile', '.Renviron'), ('text/S-plus', 'text/S', 'text/x-r-source', 'text/x-r', 'text/x-R', 'text/x-r-history', 'text/x-r-profile')),
+    'SMLLexer': ('pip._vendor.pygments.lexers.ml', 'Standard ML', ('sml',), ('*.sml', '*.sig', '*.fun'), ('text/x-standardml', 'application/x-standardml')),
+    'SNBTLexer': ('pip._vendor.pygments.lexers.minecraft', 'SNBT', ('snbt',), ('*.snbt',), ('text/snbt',)),
+    'SarlLexer': ('pip._vendor.pygments.lexers.jvm', 'SARL', ('sarl',), ('*.sarl',), ('text/x-sarl',)),
+    'SassLexer': ('pip._vendor.pygments.lexers.css', 'Sass', ('sass',), ('*.sass',), ('text/x-sass',)),
+    'SaviLexer': ('pip._vendor.pygments.lexers.savi', 'Savi', ('savi',), ('*.savi',), ()),
+    'ScalaLexer': ('pip._vendor.pygments.lexers.jvm', 'Scala', ('scala',), ('*.scala',), ('text/x-scala',)),
+    'ScamlLexer': ('pip._vendor.pygments.lexers.html', 'Scaml', ('scaml',), ('*.scaml',), ('text/x-scaml',)),
+    'ScdocLexer': ('pip._vendor.pygments.lexers.scdoc', 'scdoc', ('scdoc', 'scd'), ('*.scd', '*.scdoc'), ()),
+    'SchemeLexer': ('pip._vendor.pygments.lexers.lisp', 'Scheme', ('scheme', 'scm'), ('*.scm', '*.ss'), ('text/x-scheme', 'application/x-scheme')),
+    'ScilabLexer': ('pip._vendor.pygments.lexers.matlab', 'Scilab', ('scilab',), ('*.sci', '*.sce', '*.tst'), ('text/scilab',)),
+    'ScssLexer': ('pip._vendor.pygments.lexers.css', 'SCSS', ('scss',), ('*.scss',), ('text/x-scss',)),
+    'SedLexer': ('pip._vendor.pygments.lexers.textedit', 'Sed', ('sed', 'gsed', 'ssed'), ('*.sed', '*.[gs]sed'), ('text/x-sed',)),
+    'ShExCLexer': ('pip._vendor.pygments.lexers.rdf', 'ShExC', ('shexc', 'shex'), ('*.shex',), ('text/shex',)),
+    'ShenLexer': ('pip._vendor.pygments.lexers.lisp', 'Shen', ('shen',), ('*.shen',), ('text/x-shen', 'application/x-shen')),
+    'SieveLexer': ('pip._vendor.pygments.lexers.sieve', 'Sieve', ('sieve',), ('*.siv', '*.sieve'), ()),
+    'SilverLexer': ('pip._vendor.pygments.lexers.verification', 'Silver', ('silver',), ('*.sil', '*.vpr'), ()),
+    'SingularityLexer': ('pip._vendor.pygments.lexers.configs', 'Singularity', ('singularity',), ('*.def', 'Singularity'), ()),
+    'SlashLexer': ('pip._vendor.pygments.lexers.slash', 'Slash', ('slash',), ('*.sla',), ()),
+    'SlimLexer': ('pip._vendor.pygments.lexers.webmisc', 'Slim', ('slim',), ('*.slim',), ('text/x-slim',)),
+    'SlurmBashLexer': ('pip._vendor.pygments.lexers.shell', 'Slurm', ('slurm', 'sbatch'), ('*.sl',), ()),
+    'SmaliLexer': ('pip._vendor.pygments.lexers.dalvik', 'Smali', ('smali',), ('*.smali',), ('text/smali',)),
+    'SmalltalkLexer': ('pip._vendor.pygments.lexers.smalltalk', 'Smalltalk', ('smalltalk', 'squeak', 'st'), ('*.st',), ('text/x-smalltalk',)),
+    'SmartGameFormatLexer': ('pip._vendor.pygments.lexers.sgf', 'SmartGameFormat', ('sgf',), ('*.sgf',), ()),
+    'SmartyLexer': ('pip._vendor.pygments.lexers.templates', 'Smarty', ('smarty',), ('*.tpl',), ('application/x-smarty',)),
+    'SmithyLexer': ('pip._vendor.pygments.lexers.smithy', 'Smithy', ('smithy',), ('*.smithy',), ()),
+    'SnobolLexer': ('pip._vendor.pygments.lexers.snobol', 'Snobol', ('snobol',), ('*.snobol',), ('text/x-snobol',)),
+    'SnowballLexer': ('pip._vendor.pygments.lexers.dsls', 'Snowball', ('snowball',), ('*.sbl',), ()),
+    'SolidityLexer': ('pip._vendor.pygments.lexers.solidity', 'Solidity', ('solidity',), ('*.sol',), ()),
+    'SoongLexer': ('pip._vendor.pygments.lexers.soong', 'Soong', ('androidbp', 'bp', 'soong'), ('Android.bp',), ()),
+    'SophiaLexer': ('pip._vendor.pygments.lexers.sophia', 'Sophia', ('sophia',), ('*.aes',), ()),
+    'SourcePawnLexer': ('pip._vendor.pygments.lexers.pawn', 'SourcePawn', ('sp',), ('*.sp',), ('text/x-sourcepawn',)),
+    'SourcesListLexer': ('pip._vendor.pygments.lexers.installers', 'Debian Sourcelist', ('debsources', 'sourceslist', 'sources.list'), ('sources.list',), ()),
+    'SparqlLexer': ('pip._vendor.pygments.lexers.rdf', 'SPARQL', ('sparql',), ('*.rq', '*.sparql'), ('application/sparql-query',)),
+    'SpiceLexer': ('pip._vendor.pygments.lexers.spice', 'Spice', ('spice', 'spicelang'), ('*.spice',), ('text/x-spice',)),
+    'SqlJinjaLexer': ('pip._vendor.pygments.lexers.templates', 'SQL+Jinja', ('sql+jinja',), ('*.sql', '*.sql.j2', '*.sql.jinja2'), ()),
+    'SqlLexer': ('pip._vendor.pygments.lexers.sql', 'SQL', ('sql',), ('*.sql',), ('text/x-sql',)),
+    'SqliteConsoleLexer': ('pip._vendor.pygments.lexers.sql', 'sqlite3con', ('sqlite3',), ('*.sqlite3-console',), ('text/x-sqlite3-console',)),
+    'SquidConfLexer': ('pip._vendor.pygments.lexers.configs', 'SquidConf', ('squidconf', 'squid.conf', 'squid'), ('squid.conf',), ('text/x-squidconf',)),
+    'SrcinfoLexer': ('pip._vendor.pygments.lexers.srcinfo', 'Srcinfo', ('srcinfo',), ('.SRCINFO',), ()),
+    'SspLexer': ('pip._vendor.pygments.lexers.templates', 'Scalate Server Page', ('ssp',), ('*.ssp',), ('application/x-ssp',)),
+    'StanLexer': ('pip._vendor.pygments.lexers.modeling', 'Stan', ('stan',), ('*.stan',), ()),
+    'StataLexer': ('pip._vendor.pygments.lexers.stata', 'Stata', ('stata', 'do'), ('*.do', '*.ado'), ('text/x-stata', 'text/stata', 'application/x-stata')),
+    'SuperColliderLexer': ('pip._vendor.pygments.lexers.supercollider', 'SuperCollider', ('supercollider', 'sc'), ('*.sc', '*.scd'), ('application/supercollider', 'text/supercollider')),
+    'SwiftLexer': ('pip._vendor.pygments.lexers.objective', 'Swift', ('swift',), ('*.swift',), ('text/x-swift',)),
+    'SwigLexer': ('pip._vendor.pygments.lexers.c_like', 'SWIG', ('swig',), ('*.swg', '*.i'), ('text/swig',)),
+    'SystemVerilogLexer': ('pip._vendor.pygments.lexers.hdl', 'systemverilog', ('systemverilog', 'sv'), ('*.sv', '*.svh'), ('text/x-systemverilog',)),
+    'SystemdLexer': ('pip._vendor.pygments.lexers.configs', 'Systemd', ('systemd',), ('*.service', '*.socket', '*.device', '*.mount', '*.automount', '*.swap', '*.target', '*.path', '*.timer', '*.slice', '*.scope'), ()),
+    'TAPLexer': ('pip._vendor.pygments.lexers.testing', 'TAP', ('tap',), ('*.tap',), ()),
+    'TNTLexer': ('pip._vendor.pygments.lexers.tnt', 'Typographic Number Theory', ('tnt',), ('*.tnt',), ()),
+    'TOMLLexer': ('pip._vendor.pygments.lexers.configs', 'TOML', ('toml',), ('*.toml', 'Pipfile', 'poetry.lock'), ('application/toml',)),
+    'TactLexer': ('pip._vendor.pygments.lexers.tact', 'Tact', ('tact',), ('*.tact',), ()),
+    'Tads3Lexer': ('pip._vendor.pygments.lexers.int_fiction', 'TADS 3', ('tads3',), ('*.t',), ()),
+    'TalLexer': ('pip._vendor.pygments.lexers.tal', 'Tal', ('tal', 'uxntal'), ('*.tal',), ('text/x-uxntal',)),
+    'TasmLexer': ('pip._vendor.pygments.lexers.asm', 'TASM', ('tasm',), ('*.asm', '*.ASM', '*.tasm'), ('text/x-tasm',)),
+    'TclLexer': ('pip._vendor.pygments.lexers.tcl', 'Tcl', ('tcl',), ('*.tcl', '*.rvt'), ('text/x-tcl', 'text/x-script.tcl', 'application/x-tcl')),
+    'TcshLexer': ('pip._vendor.pygments.lexers.shell', 'Tcsh', ('tcsh', 'csh'), ('*.tcsh', '*.csh'), ('application/x-csh',)),
+    'TcshSessionLexer': ('pip._vendor.pygments.lexers.shell', 'Tcsh Session', ('tcshcon',), (), ()),
+    'TeaTemplateLexer': ('pip._vendor.pygments.lexers.templates', 'Tea', ('tea',), ('*.tea',), ('text/x-tea',)),
+    'TealLexer': ('pip._vendor.pygments.lexers.teal', 'teal', ('teal',), ('*.teal',), ()),
+    'TeraTermLexer': ('pip._vendor.pygments.lexers.teraterm', 'Tera Term macro', ('teratermmacro', 'teraterm', 'ttl'), ('*.ttl',), ('text/x-teratermmacro',)),
+    'TermcapLexer': ('pip._vendor.pygments.lexers.configs', 'Termcap', ('termcap',), ('termcap', 'termcap.src'), ()),
+    'TerminfoLexer': ('pip._vendor.pygments.lexers.configs', 'Terminfo', ('terminfo',), ('terminfo', 'terminfo.src'), ()),
+    'TerraformLexer': ('pip._vendor.pygments.lexers.configs', 'Terraform', ('terraform', 'tf', 'hcl'), ('*.tf', '*.hcl'), ('application/x-tf', 'application/x-terraform')),
+    'TexLexer': ('pip._vendor.pygments.lexers.markup', 'TeX', ('tex', 'latex'), ('*.tex', '*.aux', '*.toc'), ('text/x-tex', 'text/x-latex')),
+    'TextLexer': ('pip._vendor.pygments.lexers.special', 'Text only', ('text',), ('*.txt',), ('text/plain',)),
+    'ThingsDBLexer': ('pip._vendor.pygments.lexers.thingsdb', 'ThingsDB', ('ti', 'thingsdb'), ('*.ti',), ()),
+    'ThriftLexer': ('pip._vendor.pygments.lexers.dsls', 'Thrift', ('thrift',), ('*.thrift',), ('application/x-thrift',)),
+    'TiddlyWiki5Lexer': ('pip._vendor.pygments.lexers.markup', 'tiddler', ('tid',), ('*.tid',), ('text/vnd.tiddlywiki',)),
+    'TlbLexer': ('pip._vendor.pygments.lexers.tlb', 'Tl-b', ('tlb',), ('*.tlb',), ()),
+    'TlsLexer': ('pip._vendor.pygments.lexers.tls', 'TLS Presentation Language', ('tls',), (), ()),
+    'TodotxtLexer': ('pip._vendor.pygments.lexers.textfmts', 'Todotxt', ('todotxt',), ('todo.txt', '*.todotxt'), ('text/x-todo',)),
+    'TransactSqlLexer': ('pip._vendor.pygments.lexers.sql', 'Transact-SQL', ('tsql', 't-sql'), ('*.sql',), ('text/x-tsql',)),
+    'TreetopLexer': ('pip._vendor.pygments.lexers.parsers', 'Treetop', ('treetop',), ('*.treetop', '*.tt'), ()),
+    'TurtleLexer': ('pip._vendor.pygments.lexers.rdf', 'Turtle', ('turtle',), ('*.ttl',), ('text/turtle', 'application/x-turtle')),
+    'TwigHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Twig', ('html+twig',), ('*.twig',), ('text/html+twig',)),
+    'TwigLexer': ('pip._vendor.pygments.lexers.templates', 'Twig', ('twig',), (), ('application/x-twig',)),
+    'TypeScriptLexer': ('pip._vendor.pygments.lexers.javascript', 'TypeScript', ('typescript', 'ts'), ('*.ts',), ('application/x-typescript', 'text/x-typescript')),
+    'TypoScriptCssDataLexer': ('pip._vendor.pygments.lexers.typoscript', 'TypoScriptCssData', ('typoscriptcssdata',), (), ()),
+    'TypoScriptHtmlDataLexer': ('pip._vendor.pygments.lexers.typoscript', 'TypoScriptHtmlData', ('typoscripthtmldata',), (), ()),
+    'TypoScriptLexer': ('pip._vendor.pygments.lexers.typoscript', 'TypoScript', ('typoscript',), ('*.typoscript',), ('text/x-typoscript',)),
+    'TypstLexer': ('pip._vendor.pygments.lexers.typst', 'Typst', ('typst',), ('*.typ',), ('text/x-typst',)),
+    'UL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'UL4', ('ul4',), ('*.ul4',), ()),
+    'UcodeLexer': ('pip._vendor.pygments.lexers.unicon', 'ucode', ('ucode',), ('*.u', '*.u1', '*.u2'), ()),
+    'UniconLexer': ('pip._vendor.pygments.lexers.unicon', 'Unicon', ('unicon',), ('*.icn',), ('text/unicon',)),
+    'UnixConfigLexer': ('pip._vendor.pygments.lexers.configs', 'Unix/Linux config files', ('unixconfig', 'linuxconfig'), (), ()),
+    'UrbiscriptLexer': ('pip._vendor.pygments.lexers.urbi', 'UrbiScript', ('urbiscript',), ('*.u',), ('application/x-urbiscript',)),
+    'UrlEncodedLexer': ('pip._vendor.pygments.lexers.html', 'urlencoded', ('urlencoded',), (), ('application/x-www-form-urlencoded',)),
+    'UsdLexer': ('pip._vendor.pygments.lexers.usd', 'USD', ('usd', 'usda'), ('*.usd', '*.usda'), ()),
+    'VBScriptLexer': ('pip._vendor.pygments.lexers.basic', 'VBScript', ('vbscript',), ('*.vbs', '*.VBS'), ()),
+    'VCLLexer': ('pip._vendor.pygments.lexers.varnish', 'VCL', ('vcl',), ('*.vcl',), ('text/x-vclsrc',)),
+    'VCLSnippetLexer': ('pip._vendor.pygments.lexers.varnish', 'VCLSnippets', ('vclsnippets', 'vclsnippet'), (), ('text/x-vclsnippet',)),
+    'VCTreeStatusLexer': ('pip._vendor.pygments.lexers.console', 'VCTreeStatus', ('vctreestatus',), (), ()),
+    'VGLLexer': ('pip._vendor.pygments.lexers.dsls', 'VGL', ('vgl',), ('*.rpf',), ()),
+    'ValaLexer': ('pip._vendor.pygments.lexers.c_like', 'Vala', ('vala', 'vapi'), ('*.vala', '*.vapi'), ('text/x-vala',)),
+    'VbNetAspxLexer': ('pip._vendor.pygments.lexers.dotnet', 'aspx-vb', ('aspx-vb',), ('*.aspx', '*.asax', '*.ascx', '*.ashx', '*.asmx', '*.axd'), ()),
+    'VbNetLexer': ('pip._vendor.pygments.lexers.dotnet', 'VB.net', ('vb.net', 'vbnet', 'lobas', 'oobas', 'sobas', 'visual-basic', 'visualbasic'), ('*.vb', '*.bas'), ('text/x-vbnet', 'text/x-vba')),
+    'VelocityHtmlLexer': ('pip._vendor.pygments.lexers.templates', 'HTML+Velocity', ('html+velocity',), (), ('text/html+velocity',)),
+    'VelocityLexer': ('pip._vendor.pygments.lexers.templates', 'Velocity', ('velocity',), ('*.vm', '*.fhtml'), ()),
+    'VelocityXmlLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Velocity', ('xml+velocity',), (), ('application/xml+velocity',)),
+    'VerifpalLexer': ('pip._vendor.pygments.lexers.verifpal', 'Verifpal', ('verifpal',), ('*.vp',), ('text/x-verifpal',)),
+    'VerilogLexer': ('pip._vendor.pygments.lexers.hdl', 'verilog', ('verilog', 'v'), ('*.v',), ('text/x-verilog',)),
+    'VhdlLexer': ('pip._vendor.pygments.lexers.hdl', 'vhdl', ('vhdl',), ('*.vhdl', '*.vhd'), ('text/x-vhdl',)),
+    'VimLexer': ('pip._vendor.pygments.lexers.textedit', 'VimL', ('vim',), ('*.vim', '.vimrc', '.exrc', '.gvimrc', '_vimrc', '_exrc', '_gvimrc', 'vimrc', 'gvimrc'), ('text/x-vim',)),
+    'VisualPrologGrammarLexer': ('pip._vendor.pygments.lexers.vip', 'Visual Prolog Grammar', ('visualprologgrammar',), ('*.vipgrm',), ()),
+    'VisualPrologLexer': ('pip._vendor.pygments.lexers.vip', 'Visual Prolog', ('visualprolog',), ('*.pro', '*.cl', '*.i', '*.pack', '*.ph'), ()),
+    'VyperLexer': ('pip._vendor.pygments.lexers.vyper', 'Vyper', ('vyper',), ('*.vy',), ()),
+    'WDiffLexer': ('pip._vendor.pygments.lexers.diff', 'WDiff', ('wdiff',), ('*.wdiff',), ()),
+    'WatLexer': ('pip._vendor.pygments.lexers.webassembly', 'WebAssembly', ('wast', 'wat'), ('*.wat', '*.wast'), ()),
+    'WebIDLLexer': ('pip._vendor.pygments.lexers.webidl', 'Web IDL', ('webidl',), ('*.webidl',), ()),
+    'WgslLexer': ('pip._vendor.pygments.lexers.wgsl', 'WebGPU Shading Language', ('wgsl',), ('*.wgsl',), ('text/wgsl',)),
+    'WhileyLexer': ('pip._vendor.pygments.lexers.whiley', 'Whiley', ('whiley',), ('*.whiley',), ('text/x-whiley',)),
+    'WikitextLexer': ('pip._vendor.pygments.lexers.markup', 'Wikitext', ('wikitext', 'mediawiki'), (), ('text/x-wiki',)),
+    'WoWTocLexer': ('pip._vendor.pygments.lexers.wowtoc', 'World of Warcraft TOC', ('wowtoc',), ('*.toc',), ()),
+    'WrenLexer': ('pip._vendor.pygments.lexers.wren', 'Wren', ('wren',), ('*.wren',), ()),
+    'X10Lexer': ('pip._vendor.pygments.lexers.x10', 'X10', ('x10', 'xten'), ('*.x10',), ('text/x-x10',)),
+    'XMLUL4Lexer': ('pip._vendor.pygments.lexers.ul4', 'XML+UL4', ('xml+ul4',), ('*.xmlul4',), ()),
+    'XQueryLexer': ('pip._vendor.pygments.lexers.webmisc', 'XQuery', ('xquery', 'xqy', 'xq', 'xql', 'xqm'), ('*.xqy', '*.xquery', '*.xq', '*.xql', '*.xqm'), ('text/xquery', 'application/xquery')),
+    'XmlDjangoLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Django/Jinja', ('xml+django', 'xml+jinja'), ('*.xml.j2', '*.xml.jinja2'), ('application/xml+django', 'application/xml+jinja')),
+    'XmlErbLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Ruby', ('xml+ruby', 'xml+erb'), (), ('application/xml+ruby',)),
+    'XmlLexer': ('pip._vendor.pygments.lexers.html', 'XML', ('xml',), ('*.xml', '*.xsl', '*.rss', '*.xslt', '*.xsd', '*.wsdl', '*.wsf'), ('text/xml', 'application/xml', 'image/svg+xml', 'application/rss+xml', 'application/atom+xml')),
+    'XmlPhpLexer': ('pip._vendor.pygments.lexers.templates', 'XML+PHP', ('xml+php',), (), ('application/xml+php',)),
+    'XmlSmartyLexer': ('pip._vendor.pygments.lexers.templates', 'XML+Smarty', ('xml+smarty',), (), ('application/xml+smarty',)),
+    'XorgLexer': ('pip._vendor.pygments.lexers.xorg', 'Xorg', ('xorg.conf',), ('xorg.conf',), ()),
+    'XppLexer': ('pip._vendor.pygments.lexers.dotnet', 'X++', ('xpp', 'x++'), ('*.xpp',), ()),
+    'XsltLexer': ('pip._vendor.pygments.lexers.html', 'XSLT', ('xslt',), ('*.xsl', '*.xslt', '*.xpl'), ('application/xsl+xml', 'application/xslt+xml')),
+    'XtendLexer': ('pip._vendor.pygments.lexers.jvm', 'Xtend', ('xtend',), ('*.xtend',), ('text/x-xtend',)),
+    'XtlangLexer': ('pip._vendor.pygments.lexers.lisp', 'xtlang', ('extempore',), ('*.xtm',), ()),
+    'YamlJinjaLexer': ('pip._vendor.pygments.lexers.templates', 'YAML+Jinja', ('yaml+jinja', 'salt', 'sls'), ('*.sls', '*.yaml.j2', '*.yml.j2', '*.yaml.jinja2', '*.yml.jinja2'), ('text/x-yaml+jinja', 'text/x-sls')),
+    'YamlLexer': ('pip._vendor.pygments.lexers.data', 'YAML', ('yaml',), ('*.yaml', '*.yml'), ('text/x-yaml',)),
+    'YangLexer': ('pip._vendor.pygments.lexers.yang', 'YANG', ('yang',), ('*.yang',), ('application/yang',)),
+    'YaraLexer': ('pip._vendor.pygments.lexers.yara', 'YARA', ('yara', 'yar'), ('*.yar',), ('text/x-yara',)),
+    'ZeekLexer': ('pip._vendor.pygments.lexers.dsls', 'Zeek', ('zeek', 'bro'), ('*.zeek', '*.bro'), ()),
+    'ZephirLexer': ('pip._vendor.pygments.lexers.php', 'Zephir', ('zephir',), ('*.zep',), ()),
+    'ZigLexer': ('pip._vendor.pygments.lexers.zig', 'Zig', ('zig',), ('*.zig',), ('text/zig',)),
+    'apdlexer': ('pip._vendor.pygments.lexers.apdlexer', 'ANSYS parametric design language', ('ansys', 'apdl'), ('*.ans',), ()),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/python.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/python.py
new file mode 100644
index 00000000..b2d07f20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/lexers/python.py
@@ -0,0 +1,1198 @@
+"""
+    pygments.lexers.python
+    ~~~~~~~~~~~~~~~~~~~~~~
+
+    Lexers for Python and related languages.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import keyword
+
+from pip._vendor.pygments.lexer import DelegatingLexer, RegexLexer, include, \
+    bygroups, using, default, words, combined, this
+from pip._vendor.pygments.util import get_bool_opt, shebang_matches
+from pip._vendor.pygments.token import Text, Comment, Operator, Keyword, Name, String, \
+    Number, Punctuation, Generic, Other, Error, Whitespace
+from pip._vendor.pygments import unistring as uni
+
+__all__ = ['PythonLexer', 'PythonConsoleLexer', 'PythonTracebackLexer',
+           'Python2Lexer', 'Python2TracebackLexer',
+           'CythonLexer', 'DgLexer', 'NumPyLexer']
+
+
+class PythonLexer(RegexLexer):
+    """
+    For Python source code (version 3.x).
+
+    .. versionchanged:: 2.5
+       This is now the default ``PythonLexer``.  It is still available as the
+       alias ``Python3Lexer``.
+    """
+
+    name = 'Python'
+    url = 'https://www.python.org'
+    aliases = ['python', 'py', 'sage', 'python3', 'py3', 'bazel', 'starlark']
+    filenames = [
+        '*.py',
+        '*.pyw',
+        # Type stubs
+        '*.pyi',
+        # Jython
+        '*.jy',
+        # Sage
+        '*.sage',
+        # SCons
+        '*.sc',
+        'SConstruct',
+        'SConscript',
+        # Skylark/Starlark (used by Bazel, Buck, and Pants)
+        '*.bzl',
+        'BUCK',
+        'BUILD',
+        'BUILD.bazel',
+        'WORKSPACE',
+        # Twisted Application infrastructure
+        '*.tac',
+    ]
+    mimetypes = ['text/x-python', 'application/x-python',
+                 'text/x-python3', 'application/x-python3']
+    version_added = '0.10'
+
+    uni_name = f"[{uni.xid_start}][{uni.xid_continue}]*"
+
+    def innerstring_rules(ttype):
+        return [
+            # the old style '%s' % (...) string formatting (still valid in Py3)
+            (r'%(\(\w+\))?[-#0 +]*([0-9]+|[*])?(\.([0-9]+|[*]))?'
+             '[hlL]?[E-GXc-giorsaux%]', String.Interpol),
+            # the new style '{}'.format(...) string formatting
+            (r'\{'
+             r'((\w+)((\.\w+)|(\[[^\]]+\]))*)?'  # field name
+             r'(\![sra])?'                       # conversion
+             r'(\:(.?[<>=\^])?[-+ ]?#?0?(\d+)?,?(\.\d+)?[E-GXb-gnosx%]?)?'
+             r'\}', String.Interpol),
+
+            # backslashes, quotes and formatting signs must be parsed one at a time
+            (r'[^\\\'"%{\n]+', ttype),
+            (r'[\'"\\]', ttype),
+            # unhandled string formatting sign
+            (r'%|(\{{1,2})', ttype)
+            # newlines are an error (use "nl" state)
+        ]
+
+    def fstring_rules(ttype):
+        return [
+            # Assuming that a '}' is the closing brace after format specifier.
+            # Sadly, this means that we won't detect syntax error. But it's
+            # more important to parse correct syntax correctly, than to
+            # highlight invalid syntax.
+            (r'\}', String.Interpol),
+            (r'\{', String.Interpol, 'expr-inside-fstring'),
+            # backslashes, quotes and formatting signs must be parsed one at a time
+            (r'[^\\\'"{}\n]+', ttype),
+            (r'[\'"\\]', ttype),
+            # newlines are an error (use "nl" state)
+        ]
+
+    tokens = {
+        'root': [
+            (r'\n', Whitespace),
+            (r'^(\s*)([rRuUbB]{,2})("""(?:.|\n)*?""")',
+             bygroups(Whitespace, String.Affix, String.Doc)),
+            (r"^(\s*)([rRuUbB]{,2})('''(?:.|\n)*?''')",
+             bygroups(Whitespace, String.Affix, String.Doc)),
+            (r'\A#!.+$', Comment.Hashbang),
+            (r'#.*$', Comment.Single),
+            (r'\\\n', Text),
+            (r'\\', Text),
+            include('keywords'),
+            include('soft-keywords'),
+            (r'(def)((?:\s|\\\s)+)', bygroups(Keyword, Text), 'funcname'),
+            (r'(class)((?:\s|\\\s)+)', bygroups(Keyword, Text), 'classname'),
+            (r'(from)((?:\s|\\\s)+)', bygroups(Keyword.Namespace, Text),
+             'fromimport'),
+            (r'(import)((?:\s|\\\s)+)', bygroups(Keyword.Namespace, Text),
+             'import'),
+            include('expr'),
+        ],
+        'expr': [
+            # raw f-strings
+            ('(?i)(rf|fr)(""")',
+             bygroups(String.Affix, String.Double),
+             combined('rfstringescape', 'tdqf')),
+            ("(?i)(rf|fr)(''')",
+             bygroups(String.Affix, String.Single),
+             combined('rfstringescape', 'tsqf')),
+            ('(?i)(rf|fr)(")',
+             bygroups(String.Affix, String.Double),
+             combined('rfstringescape', 'dqf')),
+            ("(?i)(rf|fr)(')",
+             bygroups(String.Affix, String.Single),
+             combined('rfstringescape', 'sqf')),
+            # non-raw f-strings
+            ('([fF])(""")', bygroups(String.Affix, String.Double),
+             combined('fstringescape', 'tdqf')),
+            ("([fF])(''')", bygroups(String.Affix, String.Single),
+             combined('fstringescape', 'tsqf')),
+            ('([fF])(")', bygroups(String.Affix, String.Double),
+             combined('fstringescape', 'dqf')),
+            ("([fF])(')", bygroups(String.Affix, String.Single),
+             combined('fstringescape', 'sqf')),
+            # raw bytes and strings
+            ('(?i)(rb|br|r)(""")',
+             bygroups(String.Affix, String.Double), 'tdqs'),
+            ("(?i)(rb|br|r)(''')",
+             bygroups(String.Affix, String.Single), 'tsqs'),
+            ('(?i)(rb|br|r)(")',
+             bygroups(String.Affix, String.Double), 'dqs'),
+            ("(?i)(rb|br|r)(')",
+             bygroups(String.Affix, String.Single), 'sqs'),
+            # non-raw strings
+            ('([uU]?)(""")', bygroups(String.Affix, String.Double),
+             combined('stringescape', 'tdqs')),
+            ("([uU]?)(''')", bygroups(String.Affix, String.Single),
+             combined('stringescape', 'tsqs')),
+            ('([uU]?)(")', bygroups(String.Affix, String.Double),
+             combined('stringescape', 'dqs')),
+            ("([uU]?)(')", bygroups(String.Affix, String.Single),
+             combined('stringescape', 'sqs')),
+            # non-raw bytes
+            ('([bB])(""")', bygroups(String.Affix, String.Double),
+             combined('bytesescape', 'tdqs')),
+            ("([bB])(''')", bygroups(String.Affix, String.Single),
+             combined('bytesescape', 'tsqs')),
+            ('([bB])(")', bygroups(String.Affix, String.Double),
+             combined('bytesescape', 'dqs')),
+            ("([bB])(')", bygroups(String.Affix, String.Single),
+             combined('bytesescape', 'sqs')),
+
+            (r'[^\S\n]+', Text),
+            include('numbers'),
+            (r'!=|==|<<|>>|:=|[-~+/*%=<>&^|.]', Operator),
+            (r'[]{}:(),;[]', Punctuation),
+            (r'(in|is|and|or|not)\b', Operator.Word),
+            include('expr-keywords'),
+            include('builtins'),
+            include('magicfuncs'),
+            include('magicvars'),
+            include('name'),
+        ],
+        'expr-inside-fstring': [
+            (r'[{([]', Punctuation, 'expr-inside-fstring-inner'),
+            # without format specifier
+            (r'(=\s*)?'         # debug (https://bugs.python.org/issue36817)
+             r'(\![sraf])?'     # conversion
+             r'\}', String.Interpol, '#pop'),
+            # with format specifier
+            # we'll catch the remaining '}' in the outer scope
+            (r'(=\s*)?'         # debug (https://bugs.python.org/issue36817)
+             r'(\![sraf])?'     # conversion
+             r':', String.Interpol, '#pop'),
+            (r'\s+', Whitespace),  # allow new lines
+            include('expr'),
+        ],
+        'expr-inside-fstring-inner': [
+            (r'[{([]', Punctuation, 'expr-inside-fstring-inner'),
+            (r'[])}]', Punctuation, '#pop'),
+            (r'\s+', Whitespace),  # allow new lines
+            include('expr'),
+        ],
+        'expr-keywords': [
+            # Based on https://docs.python.org/3/reference/expressions.html
+            (words((
+                'async for', 'await', 'else', 'for', 'if', 'lambda',
+                'yield', 'yield from'), suffix=r'\b'),
+             Keyword),
+            (words(('True', 'False', 'None'), suffix=r'\b'), Keyword.Constant),
+        ],
+        'keywords': [
+            (words((
+                'assert', 'async', 'await', 'break', 'continue', 'del', 'elif',
+                'else', 'except', 'finally', 'for', 'global', 'if', 'lambda',
+                'pass', 'raise', 'nonlocal', 'return', 'try', 'while', 'yield',
+                'yield from', 'as', 'with'), suffix=r'\b'),
+             Keyword),
+            (words(('True', 'False', 'None'), suffix=r'\b'), Keyword.Constant),
+        ],
+        'soft-keywords': [
+            # `match`, `case` and `_` soft keywords
+            (r'(^[ \t]*)'              # at beginning of line + possible indentation
+             r'(match|case)\b'         # a possible keyword
+             r'(?![ \t]*(?:'           # not followed by...
+             r'[:,;=^&|@~)\]}]|(?:' +  # characters and keywords that mean this isn't
+                                       # pattern matching (but None/True/False is ok)
+             r'|'.join(k for k in keyword.kwlist if k[0].islower()) + r')\b))',
+             bygroups(Text, Keyword), 'soft-keywords-inner'),
+        ],
+        'soft-keywords-inner': [
+            # optional `_` keyword
+            (r'(\s+)([^\n_]*)(_\b)', bygroups(Whitespace, using(this), Keyword)),
+            default('#pop')
+        ],
+        'builtins': [
+            (words((
+                '__import__', 'abs', 'aiter', 'all', 'any', 'bin', 'bool', 'bytearray',
+                'breakpoint', 'bytes', 'callable', 'chr', 'classmethod', 'compile',
+                'complex', 'delattr', 'dict', 'dir', 'divmod', 'enumerate', 'eval',
+                'filter', 'float', 'format', 'frozenset', 'getattr', 'globals',
+                'hasattr', 'hash', 'hex', 'id', 'input', 'int', 'isinstance',
+                'issubclass', 'iter', 'len', 'list', 'locals', 'map', 'max',
+                'memoryview', 'min', 'next', 'object', 'oct', 'open', 'ord', 'pow',
+                'print', 'property', 'range', 'repr', 'reversed', 'round', 'set',
+                'setattr', 'slice', 'sorted', 'staticmethod', 'str', 'sum', 'super',
+                'tuple', 'type', 'vars', 'zip'), prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Builtin),
+            (r'(?<!\.)(self|Ellipsis|NotImplemented|cls)\b', Name.Builtin.Pseudo),
+            (words((
+                'ArithmeticError', 'AssertionError', 'AttributeError',
+                'BaseException', 'BufferError', 'BytesWarning', 'DeprecationWarning',
+                'EOFError', 'EnvironmentError', 'Exception', 'FloatingPointError',
+                'FutureWarning', 'GeneratorExit', 'IOError', 'ImportError',
+                'ImportWarning', 'IndentationError', 'IndexError', 'KeyError',
+                'KeyboardInterrupt', 'LookupError', 'MemoryError', 'NameError',
+                'NotImplementedError', 'OSError', 'OverflowError',
+                'PendingDeprecationWarning', 'ReferenceError', 'ResourceWarning',
+                'RuntimeError', 'RuntimeWarning', 'StopIteration',
+                'SyntaxError', 'SyntaxWarning', 'SystemError', 'SystemExit',
+                'TabError', 'TypeError', 'UnboundLocalError', 'UnicodeDecodeError',
+                'UnicodeEncodeError', 'UnicodeError', 'UnicodeTranslateError',
+                'UnicodeWarning', 'UserWarning', 'ValueError', 'VMSError',
+                'Warning', 'WindowsError', 'ZeroDivisionError',
+                # new builtin exceptions from PEP 3151
+                'BlockingIOError', 'ChildProcessError', 'ConnectionError',
+                'BrokenPipeError', 'ConnectionAbortedError', 'ConnectionRefusedError',
+                'ConnectionResetError', 'FileExistsError', 'FileNotFoundError',
+                'InterruptedError', 'IsADirectoryError', 'NotADirectoryError',
+                'PermissionError', 'ProcessLookupError', 'TimeoutError',
+                # others new in Python 3
+                'StopAsyncIteration', 'ModuleNotFoundError', 'RecursionError',
+                'EncodingWarning'),
+                prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Exception),
+        ],
+        'magicfuncs': [
+            (words((
+                '__abs__', '__add__', '__aenter__', '__aexit__', '__aiter__',
+                '__and__', '__anext__', '__await__', '__bool__', '__bytes__',
+                '__call__', '__complex__', '__contains__', '__del__', '__delattr__',
+                '__delete__', '__delitem__', '__dir__', '__divmod__', '__enter__',
+                '__eq__', '__exit__', '__float__', '__floordiv__', '__format__',
+                '__ge__', '__get__', '__getattr__', '__getattribute__',
+                '__getitem__', '__gt__', '__hash__', '__iadd__', '__iand__',
+                '__ifloordiv__', '__ilshift__', '__imatmul__', '__imod__',
+                '__imul__', '__index__', '__init__', '__instancecheck__',
+                '__int__', '__invert__', '__ior__', '__ipow__', '__irshift__',
+                '__isub__', '__iter__', '__itruediv__', '__ixor__', '__le__',
+                '__len__', '__length_hint__', '__lshift__', '__lt__', '__matmul__',
+                '__missing__', '__mod__', '__mul__', '__ne__', '__neg__',
+                '__new__', '__next__', '__or__', '__pos__', '__pow__',
+                '__prepare__', '__radd__', '__rand__', '__rdivmod__', '__repr__',
+                '__reversed__', '__rfloordiv__', '__rlshift__', '__rmatmul__',
+                '__rmod__', '__rmul__', '__ror__', '__round__', '__rpow__',
+                '__rrshift__', '__rshift__', '__rsub__', '__rtruediv__',
+                '__rxor__', '__set__', '__setattr__', '__setitem__', '__str__',
+                '__sub__', '__subclasscheck__', '__truediv__',
+                '__xor__'), suffix=r'\b'),
+             Name.Function.Magic),
+        ],
+        'magicvars': [
+            (words((
+                '__annotations__', '__bases__', '__class__', '__closure__',
+                '__code__', '__defaults__', '__dict__', '__doc__', '__file__',
+                '__func__', '__globals__', '__kwdefaults__', '__module__',
+                '__mro__', '__name__', '__objclass__', '__qualname__',
+                '__self__', '__slots__', '__weakref__'), suffix=r'\b'),
+             Name.Variable.Magic),
+        ],
+        'numbers': [
+            (r'(\d(?:_?\d)*\.(?:\d(?:_?\d)*)?|(?:\d(?:_?\d)*)?\.\d(?:_?\d)*)'
+             r'([eE][+-]?\d(?:_?\d)*)?', Number.Float),
+            (r'\d(?:_?\d)*[eE][+-]?\d(?:_?\d)*j?', Number.Float),
+            (r'0[oO](?:_?[0-7])+', Number.Oct),
+            (r'0[bB](?:_?[01])+', Number.Bin),
+            (r'0[xX](?:_?[a-fA-F0-9])+', Number.Hex),
+            (r'\d(?:_?\d)*', Number.Integer),
+        ],
+        'name': [
+            (r'@' + uni_name, Name.Decorator),
+            (r'@', Operator),  # new matrix multiplication operator
+            (uni_name, Name),
+        ],
+        'funcname': [
+            include('magicfuncs'),
+            (uni_name, Name.Function, '#pop'),
+            default('#pop'),
+        ],
+        'classname': [
+            (uni_name, Name.Class, '#pop'),
+        ],
+        'import': [
+            (r'(\s+)(as)(\s+)', bygroups(Text, Keyword, Text)),
+            (r'\.', Name.Namespace),
+            (uni_name, Name.Namespace),
+            (r'(\s*)(,)(\s*)', bygroups(Text, Operator, Text)),
+            default('#pop')  # all else: go back
+        ],
+        'fromimport': [
+            (r'(\s+)(import)\b', bygroups(Text, Keyword.Namespace), '#pop'),
+            (r'\.', Name.Namespace),
+            # if None occurs here, it's "raise x from None", since None can
+            # never be a module name
+            (r'None\b', Keyword.Constant, '#pop'),
+            (uni_name, Name.Namespace),
+            default('#pop'),
+        ],
+        'rfstringescape': [
+            (r'\{\{', String.Escape),
+            (r'\}\}', String.Escape),
+        ],
+        'fstringescape': [
+            include('rfstringescape'),
+            include('stringescape'),
+        ],
+        'bytesescape': [
+            (r'\\([\\abfnrtv"\']|\n|x[a-fA-F0-9]{2}|[0-7]{1,3})', String.Escape)
+        ],
+        'stringescape': [
+            (r'\\(N\{.*?\}|u[a-fA-F0-9]{4}|U[a-fA-F0-9]{8})', String.Escape),
+            include('bytesescape')
+        ],
+        'fstrings-single': fstring_rules(String.Single),
+        'fstrings-double': fstring_rules(String.Double),
+        'strings-single': innerstring_rules(String.Single),
+        'strings-double': innerstring_rules(String.Double),
+        'dqf': [
+            (r'"', String.Double, '#pop'),
+            (r'\\\\|\\"|\\\n', String.Escape),  # included here for raw strings
+            include('fstrings-double')
+        ],
+        'sqf': [
+            (r"'", String.Single, '#pop'),
+            (r"\\\\|\\'|\\\n", String.Escape),  # included here for raw strings
+            include('fstrings-single')
+        ],
+        'dqs': [
+            (r'"', String.Double, '#pop'),
+            (r'\\\\|\\"|\\\n', String.Escape),  # included here for raw strings
+            include('strings-double')
+        ],
+        'sqs': [
+            (r"'", String.Single, '#pop'),
+            (r"\\\\|\\'|\\\n", String.Escape),  # included here for raw strings
+            include('strings-single')
+        ],
+        'tdqf': [
+            (r'"""', String.Double, '#pop'),
+            include('fstrings-double'),
+            (r'\n', String.Double)
+        ],
+        'tsqf': [
+            (r"'''", String.Single, '#pop'),
+            include('fstrings-single'),
+            (r'\n', String.Single)
+        ],
+        'tdqs': [
+            (r'"""', String.Double, '#pop'),
+            include('strings-double'),
+            (r'\n', String.Double)
+        ],
+        'tsqs': [
+            (r"'''", String.Single, '#pop'),
+            include('strings-single'),
+            (r'\n', String.Single)
+        ],
+    }
+
+    def analyse_text(text):
+        return shebang_matches(text, r'pythonw?(3(\.\d)?)?') or \
+            'import ' in text[:1000]
+
+
+Python3Lexer = PythonLexer
+
+
+class Python2Lexer(RegexLexer):
+    """
+    For Python 2.x source code.
+
+    .. versionchanged:: 2.5
+       This class has been renamed from ``PythonLexer``.  ``PythonLexer`` now
+       refers to the Python 3 variant.  File name patterns like ``*.py`` have
+       been moved to Python 3 as well.
+    """
+
+    name = 'Python 2.x'
+    url = 'https://www.python.org'
+    aliases = ['python2', 'py2']
+    filenames = []  # now taken over by PythonLexer (3.x)
+    mimetypes = ['text/x-python2', 'application/x-python2']
+    version_added = ''
+
+    def innerstring_rules(ttype):
+        return [
+            # the old style '%s' % (...) string formatting
+            (r'%(\(\w+\))?[-#0 +]*([0-9]+|[*])?(\.([0-9]+|[*]))?'
+             '[hlL]?[E-GXc-giorsux%]', String.Interpol),
+            # backslashes, quotes and formatting signs must be parsed one at a time
+            (r'[^\\\'"%\n]+', ttype),
+            (r'[\'"\\]', ttype),
+            # unhandled string formatting sign
+            (r'%', ttype),
+            # newlines are an error (use "nl" state)
+        ]
+
+    tokens = {
+        'root': [
+            (r'\n', Whitespace),
+            (r'^(\s*)([rRuUbB]{,2})("""(?:.|\n)*?""")',
+             bygroups(Whitespace, String.Affix, String.Doc)),
+            (r"^(\s*)([rRuUbB]{,2})('''(?:.|\n)*?''')",
+             bygroups(Whitespace, String.Affix, String.Doc)),
+            (r'[^\S\n]+', Text),
+            (r'\A#!.+$', Comment.Hashbang),
+            (r'#.*$', Comment.Single),
+            (r'[]{}:(),;[]', Punctuation),
+            (r'\\\n', Text),
+            (r'\\', Text),
+            (r'(in|is|and|or|not)\b', Operator.Word),
+            (r'!=|==|<<|>>|[-~+/*%=<>&^|.]', Operator),
+            include('keywords'),
+            (r'(def)((?:\s|\\\s)+)', bygroups(Keyword, Text), 'funcname'),
+            (r'(class)((?:\s|\\\s)+)', bygroups(Keyword, Text), 'classname'),
+            (r'(from)((?:\s|\\\s)+)', bygroups(Keyword.Namespace, Text),
+             'fromimport'),
+            (r'(import)((?:\s|\\\s)+)', bygroups(Keyword.Namespace, Text),
+             'import'),
+            include('builtins'),
+            include('magicfuncs'),
+            include('magicvars'),
+            include('backtick'),
+            ('([rR]|[uUbB][rR]|[rR][uUbB])(""")',
+             bygroups(String.Affix, String.Double), 'tdqs'),
+            ("([rR]|[uUbB][rR]|[rR][uUbB])(''')",
+             bygroups(String.Affix, String.Single), 'tsqs'),
+            ('([rR]|[uUbB][rR]|[rR][uUbB])(")',
+             bygroups(String.Affix, String.Double), 'dqs'),
+            ("([rR]|[uUbB][rR]|[rR][uUbB])(')",
+             bygroups(String.Affix, String.Single), 'sqs'),
+            ('([uUbB]?)(""")', bygroups(String.Affix, String.Double),
+             combined('stringescape', 'tdqs')),
+            ("([uUbB]?)(''')", bygroups(String.Affix, String.Single),
+             combined('stringescape', 'tsqs')),
+            ('([uUbB]?)(")', bygroups(String.Affix, String.Double),
+             combined('stringescape', 'dqs')),
+            ("([uUbB]?)(')", bygroups(String.Affix, String.Single),
+             combined('stringescape', 'sqs')),
+            include('name'),
+            include('numbers'),
+        ],
+        'keywords': [
+            (words((
+                'assert', 'break', 'continue', 'del', 'elif', 'else', 'except',
+                'exec', 'finally', 'for', 'global', 'if', 'lambda', 'pass',
+                'print', 'raise', 'return', 'try', 'while', 'yield',
+                'yield from', 'as', 'with'), suffix=r'\b'),
+             Keyword),
+        ],
+        'builtins': [
+            (words((
+                '__import__', 'abs', 'all', 'any', 'apply', 'basestring', 'bin',
+                'bool', 'buffer', 'bytearray', 'bytes', 'callable', 'chr', 'classmethod',
+                'cmp', 'coerce', 'compile', 'complex', 'delattr', 'dict', 'dir', 'divmod',
+                'enumerate', 'eval', 'execfile', 'exit', 'file', 'filter', 'float',
+                'frozenset', 'getattr', 'globals', 'hasattr', 'hash', 'hex', 'id',
+                'input', 'int', 'intern', 'isinstance', 'issubclass', 'iter', 'len',
+                'list', 'locals', 'long', 'map', 'max', 'min', 'next', 'object',
+                'oct', 'open', 'ord', 'pow', 'property', 'range', 'raw_input', 'reduce',
+                'reload', 'repr', 'reversed', 'round', 'set', 'setattr', 'slice',
+                'sorted', 'staticmethod', 'str', 'sum', 'super', 'tuple', 'type',
+                'unichr', 'unicode', 'vars', 'xrange', 'zip'),
+                prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Builtin),
+            (r'(?<!\.)(self|None|Ellipsis|NotImplemented|False|True|cls'
+             r')\b', Name.Builtin.Pseudo),
+            (words((
+                'ArithmeticError', 'AssertionError', 'AttributeError',
+                'BaseException', 'DeprecationWarning', 'EOFError', 'EnvironmentError',
+                'Exception', 'FloatingPointError', 'FutureWarning', 'GeneratorExit',
+                'IOError', 'ImportError', 'ImportWarning', 'IndentationError',
+                'IndexError', 'KeyError', 'KeyboardInterrupt', 'LookupError',
+                'MemoryError', 'NameError',
+                'NotImplementedError', 'OSError', 'OverflowError', 'OverflowWarning',
+                'PendingDeprecationWarning', 'ReferenceError',
+                'RuntimeError', 'RuntimeWarning', 'StandardError', 'StopIteration',
+                'SyntaxError', 'SyntaxWarning', 'SystemError', 'SystemExit',
+                'TabError', 'TypeError', 'UnboundLocalError', 'UnicodeDecodeError',
+                'UnicodeEncodeError', 'UnicodeError', 'UnicodeTranslateError',
+                'UnicodeWarning', 'UserWarning', 'ValueError', 'VMSError', 'Warning',
+                'WindowsError', 'ZeroDivisionError'), prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Exception),
+        ],
+        'magicfuncs': [
+            (words((
+                '__abs__', '__add__', '__and__', '__call__', '__cmp__', '__coerce__',
+                '__complex__', '__contains__', '__del__', '__delattr__', '__delete__',
+                '__delitem__', '__delslice__', '__div__', '__divmod__', '__enter__',
+                '__eq__', '__exit__', '__float__', '__floordiv__', '__ge__', '__get__',
+                '__getattr__', '__getattribute__', '__getitem__', '__getslice__', '__gt__',
+                '__hash__', '__hex__', '__iadd__', '__iand__', '__idiv__', '__ifloordiv__',
+                '__ilshift__', '__imod__', '__imul__', '__index__', '__init__',
+                '__instancecheck__', '__int__', '__invert__', '__iop__', '__ior__',
+                '__ipow__', '__irshift__', '__isub__', '__iter__', '__itruediv__',
+                '__ixor__', '__le__', '__len__', '__long__', '__lshift__', '__lt__',
+                '__missing__', '__mod__', '__mul__', '__ne__', '__neg__', '__new__',
+                '__nonzero__', '__oct__', '__op__', '__or__', '__pos__', '__pow__',
+                '__radd__', '__rand__', '__rcmp__', '__rdiv__', '__rdivmod__', '__repr__',
+                '__reversed__', '__rfloordiv__', '__rlshift__', '__rmod__', '__rmul__',
+                '__rop__', '__ror__', '__rpow__', '__rrshift__', '__rshift__', '__rsub__',
+                '__rtruediv__', '__rxor__', '__set__', '__setattr__', '__setitem__',
+                '__setslice__', '__str__', '__sub__', '__subclasscheck__', '__truediv__',
+                '__unicode__', '__xor__'), suffix=r'\b'),
+             Name.Function.Magic),
+        ],
+        'magicvars': [
+            (words((
+                '__bases__', '__class__', '__closure__', '__code__', '__defaults__',
+                '__dict__', '__doc__', '__file__', '__func__', '__globals__',
+                '__metaclass__', '__module__', '__mro__', '__name__', '__self__',
+                '__slots__', '__weakref__'),
+                suffix=r'\b'),
+             Name.Variable.Magic),
+        ],
+        'numbers': [
+            (r'(\d+\.\d*|\d*\.\d+)([eE][+-]?[0-9]+)?j?', Number.Float),
+            (r'\d+[eE][+-]?[0-9]+j?', Number.Float),
+            (r'0[0-7]+j?', Number.Oct),
+            (r'0[bB][01]+', Number.Bin),
+            (r'0[xX][a-fA-F0-9]+', Number.Hex),
+            (r'\d+L', Number.Integer.Long),
+            (r'\d+j?', Number.Integer)
+        ],
+        'backtick': [
+            ('`.*?`', String.Backtick),
+        ],
+        'name': [
+            (r'@[\w.]+', Name.Decorator),
+            (r'[a-zA-Z_]\w*', Name),
+        ],
+        'funcname': [
+            include('magicfuncs'),
+            (r'[a-zA-Z_]\w*', Name.Function, '#pop'),
+            default('#pop'),
+        ],
+        'classname': [
+            (r'[a-zA-Z_]\w*', Name.Class, '#pop')
+        ],
+        'import': [
+            (r'(?:[ \t]|\\\n)+', Text),
+            (r'as\b', Keyword.Namespace),
+            (r',', Operator),
+            (r'[a-zA-Z_][\w.]*', Name.Namespace),
+            default('#pop')  # all else: go back
+        ],
+        'fromimport': [
+            (r'(?:[ \t]|\\\n)+', Text),
+            (r'import\b', Keyword.Namespace, '#pop'),
+            # if None occurs here, it's "raise x from None", since None can
+            # never be a module name
+            (r'None\b', Name.Builtin.Pseudo, '#pop'),
+            # sadly, in "raise x from y" y will be highlighted as namespace too
+            (r'[a-zA-Z_.][\w.]*', Name.Namespace),
+            # anything else here also means "raise x from y" and is therefore
+            # not an error
+            default('#pop'),
+        ],
+        'stringescape': [
+            (r'\\([\\abfnrtv"\']|\n|N\{.*?\}|u[a-fA-F0-9]{4}|'
+             r'U[a-fA-F0-9]{8}|x[a-fA-F0-9]{2}|[0-7]{1,3})', String.Escape)
+        ],
+        'strings-single': innerstring_rules(String.Single),
+        'strings-double': innerstring_rules(String.Double),
+        'dqs': [
+            (r'"', String.Double, '#pop'),
+            (r'\\\\|\\"|\\\n', String.Escape),  # included here for raw strings
+            include('strings-double')
+        ],
+        'sqs': [
+            (r"'", String.Single, '#pop'),
+            (r"\\\\|\\'|\\\n", String.Escape),  # included here for raw strings
+            include('strings-single')
+        ],
+        'tdqs': [
+            (r'"""', String.Double, '#pop'),
+            include('strings-double'),
+            (r'\n', String.Double)
+        ],
+        'tsqs': [
+            (r"'''", String.Single, '#pop'),
+            include('strings-single'),
+            (r'\n', String.Single)
+        ],
+    }
+
+    def analyse_text(text):
+        return shebang_matches(text, r'pythonw?2(\.\d)?')
+
+class _PythonConsoleLexerBase(RegexLexer):
+    name = 'Python console session'
+    aliases = ['pycon', 'python-console']
+    mimetypes = ['text/x-python-doctest']
+
+    """Auxiliary lexer for `PythonConsoleLexer`.
+
+    Code tokens are output as ``Token.Other.Code``, traceback tokens as
+    ``Token.Other.Traceback``.
+    """
+    tokens = {
+        'root': [
+            (r'(>>> )(.*\n)', bygroups(Generic.Prompt, Other.Code), 'continuations'),
+            # This happens, e.g., when tracebacks are embedded in documentation;
+            # trailing whitespaces are often stripped in such contexts.
+            (r'(>>>)(\n)', bygroups(Generic.Prompt, Whitespace)),
+            (r'(\^C)?Traceback \(most recent call last\):\n', Other.Traceback, 'traceback'),
+            # SyntaxError starts with this
+            (r'  File "[^"]+", line \d+', Other.Traceback, 'traceback'),
+            (r'.*\n', Generic.Output),
+        ],
+        'continuations': [
+            (r'(\.\.\. )(.*\n)', bygroups(Generic.Prompt, Other.Code)),
+            # See above.
+            (r'(\.\.\.)(\n)', bygroups(Generic.Prompt, Whitespace)),
+            default('#pop'),
+        ],
+        'traceback': [
+            # As soon as we see a traceback, consume everything until the next
+            # >>> prompt.
+            (r'(?=>>>( |$))', Text, '#pop'),
+            (r'(KeyboardInterrupt)(\n)', bygroups(Name.Class, Whitespace)),
+            (r'.*\n', Other.Traceback),
+        ],
+    }
+
+class PythonConsoleLexer(DelegatingLexer):
+    """
+    For Python console output or doctests, such as:
+
+    .. sourcecode:: pycon
+
+        >>> a = 'foo'
+        >>> print(a)
+        foo
+        >>> 1 / 0
+        Traceback (most recent call last):
+          File "<stdin>", line 1, in <module>
+        ZeroDivisionError: integer division or modulo by zero
+
+    Additional options:
+
+    `python3`
+        Use Python 3 lexer for code.  Default is ``True``.
+
+        .. versionadded:: 1.0
+        .. versionchanged:: 2.5
+           Now defaults to ``True``.
+    """
+
+    name = 'Python console session'
+    aliases = ['pycon', 'python-console']
+    mimetypes = ['text/x-python-doctest']
+    url = 'https://python.org'
+    version_added = ''
+
+    def __init__(self, **options):
+        python3 = get_bool_opt(options, 'python3', True)
+        if python3:
+            pylexer = PythonLexer
+            tblexer = PythonTracebackLexer
+        else:
+            pylexer = Python2Lexer
+            tblexer = Python2TracebackLexer
+        # We have two auxiliary lexers. Use DelegatingLexer twice with
+        # different tokens.  TODO: DelegatingLexer should support this
+        # directly, by accepting a tuplet of auxiliary lexers and a tuple of
+        # distinguishing tokens. Then we wouldn't need this intermediary
+        # class.
+        class _ReplaceInnerCode(DelegatingLexer):
+            def __init__(self, **options):
+                super().__init__(pylexer, _PythonConsoleLexerBase, Other.Code, **options)
+        super().__init__(tblexer, _ReplaceInnerCode, Other.Traceback, **options)
+
+class PythonTracebackLexer(RegexLexer):
+    """
+    For Python 3.x tracebacks, with support for chained exceptions.
+
+    .. versionchanged:: 2.5
+       This is now the default ``PythonTracebackLexer``.  It is still available
+       as the alias ``Python3TracebackLexer``.
+    """
+
+    name = 'Python Traceback'
+    aliases = ['pytb', 'py3tb']
+    filenames = ['*.pytb', '*.py3tb']
+    mimetypes = ['text/x-python-traceback', 'text/x-python3-traceback']
+    url = 'https://python.org'
+    version_added = '1.0'
+
+    tokens = {
+        'root': [
+            (r'\n', Whitespace),
+            (r'^(\^C)?Traceback \(most recent call last\):\n', Generic.Traceback, 'intb'),
+            (r'^During handling of the above exception, another '
+             r'exception occurred:\n\n', Generic.Traceback),
+            (r'^The above exception was the direct cause of the '
+             r'following exception:\n\n', Generic.Traceback),
+            (r'^(?=  File "[^"]+", line \d+)', Generic.Traceback, 'intb'),
+            (r'^.*\n', Other),
+        ],
+        'intb': [
+            (r'^(  File )("[^"]+")(, line )(\d+)(, in )(.+)(\n)',
+             bygroups(Text, Name.Builtin, Text, Number, Text, Name, Whitespace)),
+            (r'^(  File )("[^"]+")(, line )(\d+)(\n)',
+             bygroups(Text, Name.Builtin, Text, Number, Whitespace)),
+            (r'^(    )(.+)(\n)',
+             bygroups(Whitespace, using(PythonLexer), Whitespace), 'markers'),
+            (r'^([ \t]*)(\.\.\.)(\n)',
+             bygroups(Whitespace, Comment, Whitespace)),  # for doctests...
+            (r'^([^:]+)(: )(.+)(\n)',
+             bygroups(Generic.Error, Text, Name, Whitespace), '#pop'),
+            (r'^([a-zA-Z_][\w.]*)(:?\n)',
+             bygroups(Generic.Error, Whitespace), '#pop'),
+            default('#pop'),
+        ],
+        'markers': [
+            # Either `PEP 657 <https://www.python.org/dev/peps/pep-0657/>`
+            # error locations in Python 3.11+, or single-caret markers
+            # for syntax errors before that.
+            (r'^( {4,})([~^]+)(\n)',
+             bygroups(Whitespace, Punctuation.Marker, Whitespace),
+             '#pop'),
+            default('#pop'),
+        ],
+    }
+
+
+Python3TracebackLexer = PythonTracebackLexer
+
+
+class Python2TracebackLexer(RegexLexer):
+    """
+    For Python tracebacks.
+
+    .. versionchanged:: 2.5
+       This class has been renamed from ``PythonTracebackLexer``.
+       ``PythonTracebackLexer`` now refers to the Python 3 variant.
+    """
+
+    name = 'Python 2.x Traceback'
+    aliases = ['py2tb']
+    filenames = ['*.py2tb']
+    mimetypes = ['text/x-python2-traceback']
+    url = 'https://python.org'
+    version_added = '0.7'
+
+    tokens = {
+        'root': [
+            # Cover both (most recent call last) and (innermost last)
+            # The optional ^C allows us to catch keyboard interrupt signals.
+            (r'^(\^C)?(Traceback.*\n)',
+             bygroups(Text, Generic.Traceback), 'intb'),
+            # SyntaxError starts with this.
+            (r'^(?=  File "[^"]+", line \d+)', Generic.Traceback, 'intb'),
+            (r'^.*\n', Other),
+        ],
+        'intb': [
+            (r'^(  File )("[^"]+")(, line )(\d+)(, in )(.+)(\n)',
+             bygroups(Text, Name.Builtin, Text, Number, Text, Name, Whitespace)),
+            (r'^(  File )("[^"]+")(, line )(\d+)(\n)',
+             bygroups(Text, Name.Builtin, Text, Number, Whitespace)),
+            (r'^(    )(.+)(\n)',
+             bygroups(Text, using(Python2Lexer), Whitespace), 'marker'),
+            (r'^([ \t]*)(\.\.\.)(\n)',
+             bygroups(Text, Comment, Whitespace)),  # for doctests...
+            (r'^([^:]+)(: )(.+)(\n)',
+             bygroups(Generic.Error, Text, Name, Whitespace), '#pop'),
+            (r'^([a-zA-Z_]\w*)(:?\n)',
+             bygroups(Generic.Error, Whitespace), '#pop')
+        ],
+        'marker': [
+            # For syntax errors.
+            (r'( {4,})(\^)', bygroups(Text, Punctuation.Marker), '#pop'),
+            default('#pop'),
+        ],
+    }
+
+
+class CythonLexer(RegexLexer):
+    """
+    For Pyrex and Cython source code.
+    """
+
+    name = 'Cython'
+    url = 'https://cython.org'
+    aliases = ['cython', 'pyx', 'pyrex']
+    filenames = ['*.pyx', '*.pxd', '*.pxi']
+    mimetypes = ['text/x-cython', 'application/x-cython']
+    version_added = '1.1'
+
+    tokens = {
+        'root': [
+            (r'\n', Whitespace),
+            (r'^(\s*)("""(?:.|\n)*?""")', bygroups(Whitespace, String.Doc)),
+            (r"^(\s*)('''(?:.|\n)*?''')", bygroups(Whitespace, String.Doc)),
+            (r'[^\S\n]+', Text),
+            (r'#.*$', Comment),
+            (r'[]{}:(),;[]', Punctuation),
+            (r'\\\n', Whitespace),
+            (r'\\', Text),
+            (r'(in|is|and|or|not)\b', Operator.Word),
+            (r'(<)([a-zA-Z0-9.?]+)(>)',
+             bygroups(Punctuation, Keyword.Type, Punctuation)),
+            (r'!=|==|<<|>>|[-~+/*%=<>&^|.?]', Operator),
+            (r'(from)(\d+)(<=)(\s+)(<)(\d+)(:)',
+             bygroups(Keyword, Number.Integer, Operator, Name, Operator,
+                      Name, Punctuation)),
+            include('keywords'),
+            (r'(def|property)(\s+)', bygroups(Keyword, Text), 'funcname'),
+            (r'(cp?def)(\s+)', bygroups(Keyword, Text), 'cdef'),
+            # (should actually start a block with only cdefs)
+            (r'(cdef)(:)', bygroups(Keyword, Punctuation)),
+            (r'(class|struct)(\s+)', bygroups(Keyword, Text), 'classname'),
+            (r'(from)(\s+)', bygroups(Keyword, Text), 'fromimport'),
+            (r'(c?import)(\s+)', bygroups(Keyword, Text), 'import'),
+            include('builtins'),
+            include('backtick'),
+            ('(?:[rR]|[uU][rR]|[rR][uU])"""', String, 'tdqs'),
+            ("(?:[rR]|[uU][rR]|[rR][uU])'''", String, 'tsqs'),
+            ('(?:[rR]|[uU][rR]|[rR][uU])"', String, 'dqs'),
+            ("(?:[rR]|[uU][rR]|[rR][uU])'", String, 'sqs'),
+            ('[uU]?"""', String, combined('stringescape', 'tdqs')),
+            ("[uU]?'''", String, combined('stringescape', 'tsqs')),
+            ('[uU]?"', String, combined('stringescape', 'dqs')),
+            ("[uU]?'", String, combined('stringescape', 'sqs')),
+            include('name'),
+            include('numbers'),
+        ],
+        'keywords': [
+            (words((
+                'assert', 'async', 'await', 'break', 'by', 'continue', 'ctypedef', 'del', 'elif',
+                'else', 'except', 'except?', 'exec', 'finally', 'for', 'fused', 'gil',
+                'global', 'if', 'include', 'lambda', 'nogil', 'pass', 'print',
+                'raise', 'return', 'try', 'while', 'yield', 'as', 'with'), suffix=r'\b'),
+             Keyword),
+            (r'(DEF|IF|ELIF|ELSE)\b', Comment.Preproc),
+        ],
+        'builtins': [
+            (words((
+                '__import__', 'abs', 'all', 'any', 'apply', 'basestring', 'bin', 'bint',
+                'bool', 'buffer', 'bytearray', 'bytes', 'callable', 'chr',
+                'classmethod', 'cmp', 'coerce', 'compile', 'complex', 'delattr',
+                'dict', 'dir', 'divmod', 'enumerate', 'eval', 'execfile', 'exit',
+                'file', 'filter', 'float', 'frozenset', 'getattr', 'globals',
+                'hasattr', 'hash', 'hex', 'id', 'input', 'int', 'intern', 'isinstance',
+                'issubclass', 'iter', 'len', 'list', 'locals', 'long', 'map', 'max',
+                'min', 'next', 'object', 'oct', 'open', 'ord', 'pow', 'property', 'Py_ssize_t',
+                'range', 'raw_input', 'reduce', 'reload', 'repr', 'reversed',
+                'round', 'set', 'setattr', 'slice', 'sorted', 'staticmethod',
+                'str', 'sum', 'super', 'tuple', 'type', 'unichr', 'unicode', 'unsigned',
+                'vars', 'xrange', 'zip'), prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Builtin),
+            (r'(?<!\.)(self|None|Ellipsis|NotImplemented|False|True|NULL'
+             r')\b', Name.Builtin.Pseudo),
+            (words((
+                'ArithmeticError', 'AssertionError', 'AttributeError',
+                'BaseException', 'DeprecationWarning', 'EOFError', 'EnvironmentError',
+                'Exception', 'FloatingPointError', 'FutureWarning', 'GeneratorExit',
+                'IOError', 'ImportError', 'ImportWarning', 'IndentationError',
+                'IndexError', 'KeyError', 'KeyboardInterrupt', 'LookupError',
+                'MemoryError', 'NameError', 'NotImplemented', 'NotImplementedError',
+                'OSError', 'OverflowError', 'OverflowWarning',
+                'PendingDeprecationWarning', 'ReferenceError', 'RuntimeError',
+                'RuntimeWarning', 'StandardError', 'StopIteration', 'SyntaxError',
+                'SyntaxWarning', 'SystemError', 'SystemExit', 'TabError',
+                'TypeError', 'UnboundLocalError', 'UnicodeDecodeError',
+                'UnicodeEncodeError', 'UnicodeError', 'UnicodeTranslateError',
+                'UnicodeWarning', 'UserWarning', 'ValueError', 'Warning',
+                'ZeroDivisionError'), prefix=r'(?<!\.)', suffix=r'\b'),
+             Name.Exception),
+        ],
+        'numbers': [
+            (r'(\d+\.?\d*|\d*\.\d+)([eE][+-]?[0-9]+)?', Number.Float),
+            (r'0\d+', Number.Oct),
+            (r'0[xX][a-fA-F0-9]+', Number.Hex),
+            (r'\d+L', Number.Integer.Long),
+            (r'\d+', Number.Integer)
+        ],
+        'backtick': [
+            ('`.*?`', String.Backtick),
+        ],
+        'name': [
+            (r'@\w+', Name.Decorator),
+            (r'[a-zA-Z_]\w*', Name),
+        ],
+        'funcname': [
+            (r'[a-zA-Z_]\w*', Name.Function, '#pop')
+        ],
+        'cdef': [
+            (r'(public|readonly|extern|api|inline)\b', Keyword.Reserved),
+            (r'(struct|enum|union|class)\b', Keyword),
+            (r'([a-zA-Z_]\w*)(\s*)(?=[(:#=]|$)',
+             bygroups(Name.Function, Text), '#pop'),
+            (r'([a-zA-Z_]\w*)(\s*)(,)',
+             bygroups(Name.Function, Text, Punctuation)),
+            (r'from\b', Keyword, '#pop'),
+            (r'as\b', Keyword),
+            (r':', Punctuation, '#pop'),
+            (r'(?=["\'])', Text, '#pop'),
+            (r'[a-zA-Z_]\w*', Keyword.Type),
+            (r'.', Text),
+        ],
+        'classname': [
+            (r'[a-zA-Z_]\w*', Name.Class, '#pop')
+        ],
+        'import': [
+            (r'(\s+)(as)(\s+)', bygroups(Text, Keyword, Text)),
+            (r'[a-zA-Z_][\w.]*', Name.Namespace),
+            (r'(\s*)(,)(\s*)', bygroups(Text, Operator, Text)),
+            default('#pop')  # all else: go back
+        ],
+        'fromimport': [
+            (r'(\s+)(c?import)\b', bygroups(Text, Keyword), '#pop'),
+            (r'[a-zA-Z_.][\w.]*', Name.Namespace),
+            # ``cdef foo from "header"``, or ``for foo from 0 < i < 10``
+            default('#pop'),
+        ],
+        'stringescape': [
+            (r'\\([\\abfnrtv"\']|\n|N\{.*?\}|u[a-fA-F0-9]{4}|'
+             r'U[a-fA-F0-9]{8}|x[a-fA-F0-9]{2}|[0-7]{1,3})', String.Escape)
+        ],
+        'strings': [
+            (r'%(\([a-zA-Z0-9]+\))?[-#0 +]*([0-9]+|[*])?(\.([0-9]+|[*]))?'
+             '[hlL]?[E-GXc-giorsux%]', String.Interpol),
+            (r'[^\\\'"%\n]+', String),
+            # quotes, percents and backslashes must be parsed one at a time
+            (r'[\'"\\]', String),
+            # unhandled string formatting sign
+            (r'%', String)
+            # newlines are an error (use "nl" state)
+        ],
+        'nl': [
+            (r'\n', String)
+        ],
+        'dqs': [
+            (r'"', String, '#pop'),
+            (r'\\\\|\\"|\\\n', String.Escape),  # included here again for raw strings
+            include('strings')
+        ],
+        'sqs': [
+            (r"'", String, '#pop'),
+            (r"\\\\|\\'|\\\n", String.Escape),  # included here again for raw strings
+            include('strings')
+        ],
+        'tdqs': [
+            (r'"""', String, '#pop'),
+            include('strings'),
+            include('nl')
+        ],
+        'tsqs': [
+            (r"'''", String, '#pop'),
+            include('strings'),
+            include('nl')
+        ],
+    }
+
+
+class DgLexer(RegexLexer):
+    """
+    Lexer for dg,
+    a functional and object-oriented programming language
+    running on the CPython 3 VM.
+    """
+    name = 'dg'
+    aliases = ['dg']
+    filenames = ['*.dg']
+    mimetypes = ['text/x-dg']
+    url = 'http://pyos.github.io/dg'
+    version_added = '1.6'
+
+    tokens = {
+        'root': [
+            (r'\s+', Text),
+            (r'#.*?$', Comment.Single),
+
+            (r'(?i)0b[01]+', Number.Bin),
+            (r'(?i)0o[0-7]+', Number.Oct),
+            (r'(?i)0x[0-9a-f]+', Number.Hex),
+            (r'(?i)[+-]?[0-9]+\.[0-9]+(e[+-]?[0-9]+)?j?', Number.Float),
+            (r'(?i)[+-]?[0-9]+e[+-]?\d+j?', Number.Float),
+            (r'(?i)[+-]?[0-9]+j?', Number.Integer),
+
+            (r"(?i)(br|r?b?)'''", String, combined('stringescape', 'tsqs', 'string')),
+            (r'(?i)(br|r?b?)"""', String, combined('stringescape', 'tdqs', 'string')),
+            (r"(?i)(br|r?b?)'", String, combined('stringescape', 'sqs', 'string')),
+            (r'(?i)(br|r?b?)"', String, combined('stringescape', 'dqs', 'string')),
+
+            (r"`\w+'*`", Operator),
+            (r'\b(and|in|is|or|where)\b', Operator.Word),
+            (r'[!$%&*+\-./:<-@\\^|~;,]+', Operator),
+
+            (words((
+                'bool', 'bytearray', 'bytes', 'classmethod', 'complex', 'dict', 'dict\'',
+                'float', 'frozenset', 'int', 'list', 'list\'', 'memoryview', 'object',
+                'property', 'range', 'set', 'set\'', 'slice', 'staticmethod', 'str',
+                'super', 'tuple', 'tuple\'', 'type'),
+                   prefix=r'(?<!\.)', suffix=r'(?![\'\w])'),
+             Name.Builtin),
+            (words((
+                '__import__', 'abs', 'all', 'any', 'bin', 'bind', 'chr', 'cmp', 'compile',
+                'complex', 'delattr', 'dir', 'divmod', 'drop', 'dropwhile', 'enumerate',
+                'eval', 'exhaust', 'filter', 'flip', 'foldl1?', 'format', 'fst',
+                'getattr', 'globals', 'hasattr', 'hash', 'head', 'hex', 'id', 'init',
+                'input', 'isinstance', 'issubclass', 'iter', 'iterate', 'last', 'len',
+                'locals', 'map', 'max', 'min', 'next', 'oct', 'open', 'ord', 'pow',
+                'print', 'repr', 'reversed', 'round', 'setattr', 'scanl1?', 'snd',
+                'sorted', 'sum', 'tail', 'take', 'takewhile', 'vars', 'zip'),
+                   prefix=r'(?<!\.)', suffix=r'(?![\'\w])'),
+             Name.Builtin),
+            (r"(?<!\.)(self|Ellipsis|NotImplemented|None|True|False)(?!['\w])",
+             Name.Builtin.Pseudo),
+
+            (r"(?<!\.)[A-Z]\w*(Error|Exception|Warning)'*(?!['\w])",
+             Name.Exception),
+            (r"(?<!\.)(Exception|GeneratorExit|KeyboardInterrupt|StopIteration|"
+             r"SystemExit)(?!['\w])", Name.Exception),
+
+            (r"(?<![\w.])(except|finally|for|if|import|not|otherwise|raise|"
+             r"subclass|while|with|yield)(?!['\w])", Keyword.Reserved),
+
+            (r"[A-Z_]+'*(?!['\w])", Name),
+            (r"[A-Z]\w+'*(?!['\w])", Keyword.Type),
+            (r"\w+'*", Name),
+
+            (r'[()]', Punctuation),
+            (r'.', Error),
+        ],
+        'stringescape': [
+            (r'\\([\\abfnrtv"\']|\n|N\{.*?\}|u[a-fA-F0-9]{4}|'
+             r'U[a-fA-F0-9]{8}|x[a-fA-F0-9]{2}|[0-7]{1,3})', String.Escape)
+        ],
+        'string': [
+            (r'%(\(\w+\))?[-#0 +]*([0-9]+|[*])?(\.([0-9]+|[*]))?'
+             '[hlL]?[E-GXc-giorsux%]', String.Interpol),
+            (r'[^\\\'"%\n]+', String),
+            # quotes, percents and backslashes must be parsed one at a time
+            (r'[\'"\\]', String),
+            # unhandled string formatting sign
+            (r'%', String),
+            (r'\n', String)
+        ],
+        'dqs': [
+            (r'"', String, '#pop')
+        ],
+        'sqs': [
+            (r"'", String, '#pop')
+        ],
+        'tdqs': [
+            (r'"""', String, '#pop')
+        ],
+        'tsqs': [
+            (r"'''", String, '#pop')
+        ],
+    }
+
+
+class NumPyLexer(PythonLexer):
+    """
+    A Python lexer recognizing Numerical Python builtins.
+    """
+
+    name = 'NumPy'
+    url = 'https://numpy.org/'
+    aliases = ['numpy']
+    version_added = '0.10'
+
+    # override the mimetypes to not inherit them from python
+    mimetypes = []
+    filenames = []
+
+    EXTRA_KEYWORDS = {
+        'abs', 'absolute', 'accumulate', 'add', 'alen', 'all', 'allclose',
+        'alltrue', 'alterdot', 'amax', 'amin', 'angle', 'any', 'append',
+        'apply_along_axis', 'apply_over_axes', 'arange', 'arccos', 'arccosh',
+        'arcsin', 'arcsinh', 'arctan', 'arctan2', 'arctanh', 'argmax', 'argmin',
+        'argsort', 'argwhere', 'around', 'array', 'array2string', 'array_equal',
+        'array_equiv', 'array_repr', 'array_split', 'array_str', 'arrayrange',
+        'asanyarray', 'asarray', 'asarray_chkfinite', 'ascontiguousarray',
+        'asfarray', 'asfortranarray', 'asmatrix', 'asscalar', 'astype',
+        'atleast_1d', 'atleast_2d', 'atleast_3d', 'average', 'bartlett',
+        'base_repr', 'beta', 'binary_repr', 'bincount', 'binomial',
+        'bitwise_and', 'bitwise_not', 'bitwise_or', 'bitwise_xor', 'blackman',
+        'bmat', 'broadcast', 'byte_bounds', 'bytes', 'byteswap', 'c_',
+        'can_cast', 'ceil', 'choose', 'clip', 'column_stack', 'common_type',
+        'compare_chararrays', 'compress', 'concatenate', 'conj', 'conjugate',
+        'convolve', 'copy', 'corrcoef', 'correlate', 'cos', 'cosh', 'cov',
+        'cross', 'cumprod', 'cumproduct', 'cumsum', 'delete', 'deprecate',
+        'diag', 'diagflat', 'diagonal', 'diff', 'digitize', 'disp', 'divide',
+        'dot', 'dsplit', 'dstack', 'dtype', 'dump', 'dumps', 'ediff1d', 'empty',
+        'empty_like', 'equal', 'exp', 'expand_dims', 'expm1', 'extract', 'eye',
+        'fabs', 'fastCopyAndTranspose', 'fft', 'fftfreq', 'fftshift', 'fill',
+        'finfo', 'fix', 'flat', 'flatnonzero', 'flatten', 'fliplr', 'flipud',
+        'floor', 'floor_divide', 'fmod', 'frexp', 'fromarrays', 'frombuffer',
+        'fromfile', 'fromfunction', 'fromiter', 'frompyfunc', 'fromstring',
+        'generic', 'get_array_wrap', 'get_include', 'get_numarray_include',
+        'get_numpy_include', 'get_printoptions', 'getbuffer', 'getbufsize',
+        'geterr', 'geterrcall', 'geterrobj', 'getfield', 'gradient', 'greater',
+        'greater_equal', 'gumbel', 'hamming', 'hanning', 'histogram',
+        'histogram2d', 'histogramdd', 'hsplit', 'hstack', 'hypot', 'i0',
+        'identity', 'ifft', 'imag', 'index_exp', 'indices', 'inf', 'info',
+        'inner', 'insert', 'int_asbuffer', 'interp', 'intersect1d',
+        'intersect1d_nu', 'inv', 'invert', 'iscomplex', 'iscomplexobj',
+        'isfinite', 'isfortran', 'isinf', 'isnan', 'isneginf', 'isposinf',
+        'isreal', 'isrealobj', 'isscalar', 'issctype', 'issubclass_',
+        'issubdtype', 'issubsctype', 'item', 'itemset', 'iterable', 'ix_',
+        'kaiser', 'kron', 'ldexp', 'left_shift', 'less', 'less_equal', 'lexsort',
+        'linspace', 'load', 'loads', 'loadtxt', 'log', 'log10', 'log1p', 'log2',
+        'logical_and', 'logical_not', 'logical_or', 'logical_xor', 'logspace',
+        'lstsq', 'mat', 'matrix', 'max', 'maximum', 'maximum_sctype',
+        'may_share_memory', 'mean', 'median', 'meshgrid', 'mgrid', 'min',
+        'minimum', 'mintypecode', 'mod', 'modf', 'msort', 'multiply', 'nan',
+        'nan_to_num', 'nanargmax', 'nanargmin', 'nanmax', 'nanmin', 'nansum',
+        'ndenumerate', 'ndim', 'ndindex', 'negative', 'newaxis', 'newbuffer',
+        'newbyteorder', 'nonzero', 'not_equal', 'obj2sctype', 'ogrid', 'ones',
+        'ones_like', 'outer', 'permutation', 'piecewise', 'pinv', 'pkgload',
+        'place', 'poisson', 'poly', 'poly1d', 'polyadd', 'polyder', 'polydiv',
+        'polyfit', 'polyint', 'polymul', 'polysub', 'polyval', 'power', 'prod',
+        'product', 'ptp', 'put', 'putmask', 'r_', 'randint', 'random_integers',
+        'random_sample', 'ranf', 'rank', 'ravel', 'real', 'real_if_close',
+        'recarray', 'reciprocal', 'reduce', 'remainder', 'repeat', 'require',
+        'reshape', 'resize', 'restoredot', 'right_shift', 'rint', 'roll',
+        'rollaxis', 'roots', 'rot90', 'round', 'round_', 'row_stack', 's_',
+        'sample', 'savetxt', 'sctype2char', 'searchsorted', 'seed', 'select',
+        'set_numeric_ops', 'set_printoptions', 'set_string_function',
+        'setbufsize', 'setdiff1d', 'seterr', 'seterrcall', 'seterrobj',
+        'setfield', 'setflags', 'setmember1d', 'setxor1d', 'shape',
+        'show_config', 'shuffle', 'sign', 'signbit', 'sin', 'sinc', 'sinh',
+        'size', 'slice', 'solve', 'sometrue', 'sort', 'sort_complex', 'source',
+        'split', 'sqrt', 'square', 'squeeze', 'standard_normal', 'std',
+        'subtract', 'sum', 'svd', 'swapaxes', 'take', 'tan', 'tanh', 'tensordot',
+        'test', 'tile', 'tofile', 'tolist', 'tostring', 'trace', 'transpose',
+        'trapz', 'tri', 'tril', 'trim_zeros', 'triu', 'true_divide', 'typeDict',
+        'typename', 'uniform', 'union1d', 'unique', 'unique1d', 'unravel_index',
+        'unwrap', 'vander', 'var', 'vdot', 'vectorize', 'view', 'vonmises',
+        'vsplit', 'vstack', 'weibull', 'where', 'who', 'zeros', 'zeros_like'
+    }
+
+    def get_tokens_unprocessed(self, text):
+        for index, token, value in \
+                PythonLexer.get_tokens_unprocessed(self, text):
+            if token is Name and value in self.EXTRA_KEYWORDS:
+                yield index, Keyword.Pseudo, value
+            else:
+                yield index, token, value
+
+    def analyse_text(text):
+        ltext = text[:1000]
+        return (shebang_matches(text, r'pythonw?(3(\.\d)?)?') or
+                'import ' in ltext) \
+            and ('import numpy' in ltext or 'from numpy import' in ltext)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/modeline.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/modeline.py
new file mode 100644
index 00000000..e4d9fe16
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/modeline.py
@@ -0,0 +1,43 @@
+"""
+    pygments.modeline
+    ~~~~~~~~~~~~~~~~~
+
+    A simple modeline parser (based on pymodeline).
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+
+__all__ = ['get_filetype_from_buffer']
+
+
+modeline_re = re.compile(r'''
+    (?: vi | vim | ex ) (?: [<=>]? \d* )? :
+    .* (?: ft | filetype | syn | syntax ) = ( [^:\s]+ )
+''', re.VERBOSE)
+
+
+def get_filetype_from_line(l): # noqa: E741
+    m = modeline_re.search(l)
+    if m:
+        return m.group(1)
+
+
+def get_filetype_from_buffer(buf, max_lines=5):
+    """
+    Scan the buffer for modelines and return filetype if one is found.
+    """
+    lines = buf.splitlines()
+    for line in lines[-1:-max_lines-1:-1]:
+        ret = get_filetype_from_line(line)
+        if ret:
+            return ret
+    for i in range(max_lines, -1, -1):
+        if i < len(lines):
+            ret = get_filetype_from_line(lines[i])
+            if ret:
+                return ret
+
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/plugin.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/plugin.py
new file mode 100644
index 00000000..2e462f2c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/plugin.py
@@ -0,0 +1,72 @@
+"""
+    pygments.plugin
+    ~~~~~~~~~~~~~~~
+
+    Pygments plugin interface.
+
+    lexer plugins::
+
+        [pygments.lexers]
+        yourlexer = yourmodule:YourLexer
+
+    formatter plugins::
+
+        [pygments.formatters]
+        yourformatter = yourformatter:YourFormatter
+        /.ext = yourformatter:YourFormatter
+
+    As you can see, you can define extensions for the formatter
+    with a leading slash.
+
+    syntax plugins::
+
+        [pygments.styles]
+        yourstyle = yourstyle:YourStyle
+
+    filter plugin::
+
+        [pygments.filter]
+        yourfilter = yourfilter:YourFilter
+
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+from importlib.metadata import entry_points
+
+LEXER_ENTRY_POINT = 'pygments.lexers'
+FORMATTER_ENTRY_POINT = 'pygments.formatters'
+STYLE_ENTRY_POINT = 'pygments.styles'
+FILTER_ENTRY_POINT = 'pygments.filters'
+
+
+def iter_entry_points(group_name):
+    groups = entry_points()
+    if hasattr(groups, 'select'):
+        # New interface in Python 3.10 and newer versions of the
+        # importlib_metadata backport.
+        return groups.select(group=group_name)
+    else:
+        # Older interface, deprecated in Python 3.10 and recent
+        # importlib_metadata, but we need it in Python 3.8 and 3.9.
+        return groups.get(group_name, [])
+
+
+def find_plugin_lexers():
+    for entrypoint in iter_entry_points(LEXER_ENTRY_POINT):
+        yield entrypoint.load()
+
+
+def find_plugin_formatters():
+    for entrypoint in iter_entry_points(FORMATTER_ENTRY_POINT):
+        yield entrypoint.name, entrypoint.load()
+
+
+def find_plugin_styles():
+    for entrypoint in iter_entry_points(STYLE_ENTRY_POINT):
+        yield entrypoint.name, entrypoint.load()
+
+
+def find_plugin_filters():
+    for entrypoint in iter_entry_points(FILTER_ENTRY_POINT):
+        yield entrypoint.name, entrypoint.load()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/regexopt.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/regexopt.py
new file mode 100644
index 00000000..c44eedbf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/regexopt.py
@@ -0,0 +1,91 @@
+"""
+    pygments.regexopt
+    ~~~~~~~~~~~~~~~~~
+
+    An algorithm that generates optimized regexes for matching long lists of
+    literal strings.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+from re import escape
+from os.path import commonprefix
+from itertools import groupby
+from operator import itemgetter
+
+CS_ESCAPE = re.compile(r'[\[\^\\\-\]]')
+FIRST_ELEMENT = itemgetter(0)
+
+
+def make_charset(letters):
+    return '[' + CS_ESCAPE.sub(lambda m: '\\' + m.group(), ''.join(letters)) + ']'
+
+
+def regex_opt_inner(strings, open_paren):
+    """Return a regex that matches any string in the sorted list of strings."""
+    close_paren = open_paren and ')' or ''
+    # print strings, repr(open_paren)
+    if not strings:
+        # print '-> nothing left'
+        return ''
+    first = strings[0]
+    if len(strings) == 1:
+        # print '-> only 1 string'
+        return open_paren + escape(first) + close_paren
+    if not first:
+        # print '-> first string empty'
+        return open_paren + regex_opt_inner(strings[1:], '(?:') \
+            + '?' + close_paren
+    if len(first) == 1:
+        # multiple one-char strings? make a charset
+        oneletter = []
+        rest = []
+        for s in strings:
+            if len(s) == 1:
+                oneletter.append(s)
+            else:
+                rest.append(s)
+        if len(oneletter) > 1:  # do we have more than one oneletter string?
+            if rest:
+                # print '-> 1-character + rest'
+                return open_paren + regex_opt_inner(rest, '') + '|' \
+                    + make_charset(oneletter) + close_paren
+            # print '-> only 1-character'
+            return open_paren + make_charset(oneletter) + close_paren
+    prefix = commonprefix(strings)
+    if prefix:
+        plen = len(prefix)
+        # we have a prefix for all strings
+        # print '-> prefix:', prefix
+        return open_paren + escape(prefix) \
+            + regex_opt_inner([s[plen:] for s in strings], '(?:') \
+            + close_paren
+    # is there a suffix?
+    strings_rev = [s[::-1] for s in strings]
+    suffix = commonprefix(strings_rev)
+    if suffix:
+        slen = len(suffix)
+        # print '-> suffix:', suffix[::-1]
+        return open_paren \
+            + regex_opt_inner(sorted(s[:-slen] for s in strings), '(?:') \
+            + escape(suffix[::-1]) + close_paren
+    # recurse on common 1-string prefixes
+    # print '-> last resort'
+    return open_paren + \
+        '|'.join(regex_opt_inner(list(group[1]), '')
+                 for group in groupby(strings, lambda s: s[0] == first[0])) \
+        + close_paren
+
+
+def regex_opt(strings, prefix='', suffix=''):
+    """Return a compiled regex that matches any string in the given list.
+
+    The strings to match must be literal strings, not regexes.  They will be
+    regex-escaped.
+
+    *prefix* and *suffix* are pre- and appended to the final regex.
+    """
+    strings = sorted(strings)
+    return prefix + regex_opt_inner(strings, '(') + suffix
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/scanner.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/scanner.py
new file mode 100644
index 00000000..112da349
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/scanner.py
@@ -0,0 +1,104 @@
+"""
+    pygments.scanner
+    ~~~~~~~~~~~~~~~~
+
+    This library implements a regex based scanner. Some languages
+    like Pascal are easy to parse but have some keywords that
+    depend on the context. Because of this it's impossible to lex
+    that just by using a regular expression lexer like the
+    `RegexLexer`.
+
+    Have a look at the `DelphiLexer` to get an idea of how to use
+    this scanner.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+import re
+
+
+class EndOfText(RuntimeError):
+    """
+    Raise if end of text is reached and the user
+    tried to call a match function.
+    """
+
+
+class Scanner:
+    """
+    Simple scanner
+
+    All method patterns are regular expression strings (not
+    compiled expressions!)
+    """
+
+    def __init__(self, text, flags=0):
+        """
+        :param text:    The text which should be scanned
+        :param flags:   default regular expression flags
+        """
+        self.data = text
+        self.data_length = len(text)
+        self.start_pos = 0
+        self.pos = 0
+        self.flags = flags
+        self.last = None
+        self.match = None
+        self._re_cache = {}
+
+    def eos(self):
+        """`True` if the scanner reached the end of text."""
+        return self.pos >= self.data_length
+    eos = property(eos, eos.__doc__)
+
+    def check(self, pattern):
+        """
+        Apply `pattern` on the current position and return
+        the match object. (Doesn't touch pos). Use this for
+        lookahead.
+        """
+        if self.eos:
+            raise EndOfText()
+        if pattern not in self._re_cache:
+            self._re_cache[pattern] = re.compile(pattern, self.flags)
+        return self._re_cache[pattern].match(self.data, self.pos)
+
+    def test(self, pattern):
+        """Apply a pattern on the current position and check
+        if it patches. Doesn't touch pos.
+        """
+        return self.check(pattern) is not None
+
+    def scan(self, pattern):
+        """
+        Scan the text for the given pattern and update pos/match
+        and related fields. The return value is a boolean that
+        indicates if the pattern matched. The matched value is
+        stored on the instance as ``match``, the last value is
+        stored as ``last``. ``start_pos`` is the position of the
+        pointer before the pattern was matched, ``pos`` is the
+        end position.
+        """
+        if self.eos:
+            raise EndOfText()
+        if pattern not in self._re_cache:
+            self._re_cache[pattern] = re.compile(pattern, self.flags)
+        self.last = self.match
+        m = self._re_cache[pattern].match(self.data, self.pos)
+        if m is None:
+            return False
+        self.start_pos = m.start()
+        self.pos = m.end()
+        self.match = m.group()
+        return True
+
+    def get_char(self):
+        """Scan exactly one char."""
+        self.scan('.')
+
+    def __repr__(self):
+        return '<%s %d/%d>' % (
+            self.__class__.__name__,
+            self.pos,
+            self.data_length
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/sphinxext.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/sphinxext.py
new file mode 100644
index 00000000..34077a2a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/sphinxext.py
@@ -0,0 +1,247 @@
+"""
+    pygments.sphinxext
+    ~~~~~~~~~~~~~~~~~~
+
+    Sphinx extension to generate automatic documentation of lexers,
+    formatters and filters.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import sys
+
+from docutils import nodes
+from docutils.statemachine import ViewList
+from docutils.parsers.rst import Directive
+from sphinx.util.nodes import nested_parse_with_titles
+
+
+MODULEDOC = '''
+.. module:: %s
+
+%s
+%s
+'''
+
+LEXERDOC = '''
+.. class:: %s
+
+    :Short names: %s
+    :Filenames:   %s
+    :MIME types:  %s
+
+    %s
+
+    %s
+
+'''
+
+FMTERDOC = '''
+.. class:: %s
+
+    :Short names: %s
+    :Filenames: %s
+
+    %s
+
+'''
+
+FILTERDOC = '''
+.. class:: %s
+
+    :Name: %s
+
+    %s
+
+'''
+
+
+class PygmentsDoc(Directive):
+    """
+    A directive to collect all lexers/formatters/filters and generate
+    autoclass directives for them.
+    """
+    has_content = False
+    required_arguments = 1
+    optional_arguments = 0
+    final_argument_whitespace = False
+    option_spec = {}
+
+    def run(self):
+        self.filenames = set()
+        if self.arguments[0] == 'lexers':
+            out = self.document_lexers()
+        elif self.arguments[0] == 'formatters':
+            out = self.document_formatters()
+        elif self.arguments[0] == 'filters':
+            out = self.document_filters()
+        elif self.arguments[0] == 'lexers_overview':
+            out = self.document_lexers_overview()
+        else:
+            raise Exception('invalid argument for "pygmentsdoc" directive')
+        node = nodes.compound()
+        vl = ViewList(out.split('\n'), source='')
+        nested_parse_with_titles(self.state, vl, node)
+        for fn in self.filenames:
+            self.state.document.settings.record_dependencies.add(fn)
+        return node.children
+
+    def document_lexers_overview(self):
+        """Generate a tabular overview of all lexers.
+
+        The columns are the lexer name, the extensions handled by this lexer
+        (or "None"), the aliases and a link to the lexer class."""
+        from pip._vendor.pygments.lexers._mapping import LEXERS
+        from pip._vendor.pygments.lexers import find_lexer_class
+        out = []
+
+        table = []
+
+        def format_link(name, url):
+            if url:
+                return f'`{name} <{url}>`_'
+            return name
+
+        for classname, data in sorted(LEXERS.items(), key=lambda x: x[1][1].lower()):
+            lexer_cls = find_lexer_class(data[1])
+            extensions = lexer_cls.filenames + lexer_cls.alias_filenames
+
+            table.append({
+                'name': format_link(data[1], lexer_cls.url),
+                'extensions': ', '.join(extensions).replace('*', '\\*').replace('_', '\\') or 'None',
+                'aliases': ', '.join(data[2]),
+                'class': f'{data[0]}.{classname}'
+            })
+
+        column_names = ['name', 'extensions', 'aliases', 'class']
+        column_lengths = [max([len(row[column]) for row in table if row[column]])
+                          for column in column_names]
+
+        def write_row(*columns):
+            """Format a table row"""
+            out = []
+            for length, col in zip(column_lengths, columns):
+                if col:
+                    out.append(col.ljust(length))
+                else:
+                    out.append(' '*length)
+
+            return ' '.join(out)
+
+        def write_seperator():
+            """Write a table separator row"""
+            sep = ['='*c for c in column_lengths]
+            return write_row(*sep)
+
+        out.append(write_seperator())
+        out.append(write_row('Name', 'Extension(s)', 'Short name(s)', 'Lexer class'))
+        out.append(write_seperator())
+        for row in table:
+            out.append(write_row(
+                row['name'],
+                row['extensions'],
+                row['aliases'],
+                f':class:`~{row["class"]}`'))
+        out.append(write_seperator())
+
+        return '\n'.join(out)
+
+    def document_lexers(self):
+        from pip._vendor.pygments.lexers._mapping import LEXERS
+        from pip._vendor import pygments
+        import inspect
+        import pathlib
+
+        out = []
+        modules = {}
+        moduledocstrings = {}
+        for classname, data in sorted(LEXERS.items(), key=lambda x: x[0]):
+            module = data[0]
+            mod = __import__(module, None, None, [classname])
+            self.filenames.add(mod.__file__)
+            cls = getattr(mod, classname)
+            if not cls.__doc__:
+                print(f"Warning: {classname} does not have a docstring.")
+            docstring = cls.__doc__
+            if isinstance(docstring, bytes):
+                docstring = docstring.decode('utf8')
+
+            example_file = getattr(cls, '_example', None)
+            if example_file:
+                p = pathlib.Path(inspect.getabsfile(pygments)).parent.parent /\
+                    'tests' / 'examplefiles' / example_file
+                content = p.read_text(encoding='utf-8')
+                if not content:
+                    raise Exception(
+                        f"Empty example file '{example_file}' for lexer "
+                        f"{classname}")
+
+                if data[2]:
+                    lexer_name = data[2][0]
+                    docstring += '\n\n    .. admonition:: Example\n'
+                    docstring += f'\n      .. code-block:: {lexer_name}\n\n'
+                    for line in content.splitlines():
+                        docstring += f'          {line}\n'
+
+            if cls.version_added:
+                version_line = f'.. versionadded:: {cls.version_added}'
+            else:
+                version_line = ''
+
+            modules.setdefault(module, []).append((
+                classname,
+                ', '.join(data[2]) or 'None',
+                ', '.join(data[3]).replace('*', '\\*').replace('_', '\\') or 'None',
+                ', '.join(data[4]) or 'None',
+                docstring,
+                version_line))
+            if module not in moduledocstrings:
+                moddoc = mod.__doc__
+                if isinstance(moddoc, bytes):
+                    moddoc = moddoc.decode('utf8')
+                moduledocstrings[module] = moddoc
+
+        for module, lexers in sorted(modules.items(), key=lambda x: x[0]):
+            if moduledocstrings[module] is None:
+                raise Exception(f"Missing docstring for {module}")
+            heading = moduledocstrings[module].splitlines()[4].strip().rstrip('.')
+            out.append(MODULEDOC % (module, heading, '-'*len(heading)))
+            for data in lexers:
+                out.append(LEXERDOC % data)
+
+        return ''.join(out)
+
+    def document_formatters(self):
+        from pip._vendor.pygments.formatters import FORMATTERS
+
+        out = []
+        for classname, data in sorted(FORMATTERS.items(), key=lambda x: x[0]):
+            module = data[0]
+            mod = __import__(module, None, None, [classname])
+            self.filenames.add(mod.__file__)
+            cls = getattr(mod, classname)
+            docstring = cls.__doc__
+            if isinstance(docstring, bytes):
+                docstring = docstring.decode('utf8')
+            heading = cls.__name__
+            out.append(FMTERDOC % (heading, ', '.join(data[2]) or 'None',
+                                   ', '.join(data[3]).replace('*', '\\*') or 'None',
+                                   docstring))
+        return ''.join(out)
+
+    def document_filters(self):
+        from pip._vendor.pygments.filters import FILTERS
+
+        out = []
+        for name, cls in FILTERS.items():
+            self.filenames.add(sys.modules[cls.__module__].__file__)
+            docstring = cls.__doc__
+            if isinstance(docstring, bytes):
+                docstring = docstring.decode('utf8')
+            out.append(FILTERDOC % (cls.__name__, name, docstring))
+        return ''.join(out)
+
+
+def setup(app):
+    app.add_directive('pygmentsdoc', PygmentsDoc)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/style.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/style.py
new file mode 100644
index 00000000..076e63f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/style.py
@@ -0,0 +1,203 @@
+"""
+    pygments.style
+    ~~~~~~~~~~~~~~
+
+    Basic style object.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.token import Token, STANDARD_TYPES
+
+# Default mapping of ansixxx to RGB colors.
+_ansimap = {
+    # dark
+    'ansiblack': '000000',
+    'ansired': '7f0000',
+    'ansigreen': '007f00',
+    'ansiyellow': '7f7fe0',
+    'ansiblue': '00007f',
+    'ansimagenta': '7f007f',
+    'ansicyan': '007f7f',
+    'ansigray': 'e5e5e5',
+    # normal
+    'ansibrightblack': '555555',
+    'ansibrightred': 'ff0000',
+    'ansibrightgreen': '00ff00',
+    'ansibrightyellow': 'ffff00',
+    'ansibrightblue': '0000ff',
+    'ansibrightmagenta': 'ff00ff',
+    'ansibrightcyan': '00ffff',
+    'ansiwhite': 'ffffff',
+}
+# mapping of deprecated #ansixxx colors to new color names
+_deprecated_ansicolors = {
+    # dark
+    '#ansiblack': 'ansiblack',
+    '#ansidarkred': 'ansired',
+    '#ansidarkgreen': 'ansigreen',
+    '#ansibrown': 'ansiyellow',
+    '#ansidarkblue': 'ansiblue',
+    '#ansipurple': 'ansimagenta',
+    '#ansiteal': 'ansicyan',
+    '#ansilightgray': 'ansigray',
+    # normal
+    '#ansidarkgray': 'ansibrightblack',
+    '#ansired': 'ansibrightred',
+    '#ansigreen': 'ansibrightgreen',
+    '#ansiyellow': 'ansibrightyellow',
+    '#ansiblue': 'ansibrightblue',
+    '#ansifuchsia': 'ansibrightmagenta',
+    '#ansiturquoise': 'ansibrightcyan',
+    '#ansiwhite': 'ansiwhite',
+}
+ansicolors = set(_ansimap)
+
+
+class StyleMeta(type):
+
+    def __new__(mcs, name, bases, dct):
+        obj = type.__new__(mcs, name, bases, dct)
+        for token in STANDARD_TYPES:
+            if token not in obj.styles:
+                obj.styles[token] = ''
+
+        def colorformat(text):
+            if text in ansicolors:
+                return text
+            if text[0:1] == '#':
+                col = text[1:]
+                if len(col) == 6:
+                    return col
+                elif len(col) == 3:
+                    return col[0] * 2 + col[1] * 2 + col[2] * 2
+            elif text == '':
+                return ''
+            elif text.startswith('var') or text.startswith('calc'):
+                return text
+            assert False, f"wrong color format {text!r}"
+
+        _styles = obj._styles = {}
+
+        for ttype in obj.styles:
+            for token in ttype.split():
+                if token in _styles:
+                    continue
+                ndef = _styles.get(token.parent, None)
+                styledefs = obj.styles.get(token, '').split()
+                if not ndef or token is None:
+                    ndef = ['', 0, 0, 0, '', '', 0, 0, 0]
+                elif 'noinherit' in styledefs and token is not Token:
+                    ndef = _styles[Token][:]
+                else:
+                    ndef = ndef[:]
+                _styles[token] = ndef
+                for styledef in obj.styles.get(token, '').split():
+                    if styledef == 'noinherit':
+                        pass
+                    elif styledef == 'bold':
+                        ndef[1] = 1
+                    elif styledef == 'nobold':
+                        ndef[1] = 0
+                    elif styledef == 'italic':
+                        ndef[2] = 1
+                    elif styledef == 'noitalic':
+                        ndef[2] = 0
+                    elif styledef == 'underline':
+                        ndef[3] = 1
+                    elif styledef == 'nounderline':
+                        ndef[3] = 0
+                    elif styledef[:3] == 'bg:':
+                        ndef[4] = colorformat(styledef[3:])
+                    elif styledef[:7] == 'border:':
+                        ndef[5] = colorformat(styledef[7:])
+                    elif styledef == 'roman':
+                        ndef[6] = 1
+                    elif styledef == 'sans':
+                        ndef[7] = 1
+                    elif styledef == 'mono':
+                        ndef[8] = 1
+                    else:
+                        ndef[0] = colorformat(styledef)
+
+        return obj
+
+    def style_for_token(cls, token):
+        t = cls._styles[token]
+        ansicolor = bgansicolor = None
+        color = t[0]
+        if color in _deprecated_ansicolors:
+            color = _deprecated_ansicolors[color]
+        if color in ansicolors:
+            ansicolor = color
+            color = _ansimap[color]
+        bgcolor = t[4]
+        if bgcolor in _deprecated_ansicolors:
+            bgcolor = _deprecated_ansicolors[bgcolor]
+        if bgcolor in ansicolors:
+            bgansicolor = bgcolor
+            bgcolor = _ansimap[bgcolor]
+
+        return {
+            'color':        color or None,
+            'bold':         bool(t[1]),
+            'italic':       bool(t[2]),
+            'underline':    bool(t[3]),
+            'bgcolor':      bgcolor or None,
+            'border':       t[5] or None,
+            'roman':        bool(t[6]) or None,
+            'sans':         bool(t[7]) or None,
+            'mono':         bool(t[8]) or None,
+            'ansicolor':    ansicolor,
+            'bgansicolor':  bgansicolor,
+        }
+
+    def list_styles(cls):
+        return list(cls)
+
+    def styles_token(cls, ttype):
+        return ttype in cls._styles
+
+    def __iter__(cls):
+        for token in cls._styles:
+            yield token, cls.style_for_token(token)
+
+    def __len__(cls):
+        return len(cls._styles)
+
+
+class Style(metaclass=StyleMeta):
+
+    #: overall background color (``None`` means transparent)
+    background_color = '#ffffff'
+
+    #: highlight background color
+    highlight_color = '#ffffcc'
+
+    #: line number font color
+    line_number_color = 'inherit'
+
+    #: line number background color
+    line_number_background_color = 'transparent'
+
+    #: special line number font color
+    line_number_special_color = '#000000'
+
+    #: special line number background color
+    line_number_special_background_color = '#ffffc0'
+
+    #: Style definitions for individual token types.
+    styles = {}
+
+    #: user-friendly style name (used when selecting the style, so this
+    # should be all-lowercase, no spaces, hyphens)
+    name = 'unnamed'
+
+    aliases = []
+
+    # Attribute for lexers defined within Pygments. If set
+    # to True, the style is not shown in the style gallery
+    # on the website. This is intended for language-specific
+    # styles.
+    web_style_gallery_exclude = False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__init__.py
new file mode 100644
index 00000000..712f6e69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__init__.py
@@ -0,0 +1,61 @@
+"""
+    pygments.styles
+    ~~~~~~~~~~~~~~~
+
+    Contains built-in styles.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+from pip._vendor.pygments.plugin import find_plugin_styles
+from pip._vendor.pygments.util import ClassNotFound
+from pip._vendor.pygments.styles._mapping import STYLES
+
+#: A dictionary of built-in styles, mapping style names to
+#: ``'submodule::classname'`` strings.
+#: This list is deprecated. Use `pygments.styles.STYLES` instead
+STYLE_MAP = {v[1]: v[0].split('.')[-1] + '::' + k for k, v in STYLES.items()}
+
+#: Internal reverse mapping to make `get_style_by_name` more efficient
+_STYLE_NAME_TO_MODULE_MAP = {v[1]: (v[0], k) for k, v in STYLES.items()}
+
+
+def get_style_by_name(name):
+    """
+    Return a style class by its short name. The names of the builtin styles
+    are listed in :data:`pygments.styles.STYLE_MAP`.
+
+    Will raise :exc:`pygments.util.ClassNotFound` if no style of that name is
+    found.
+    """
+    if name in _STYLE_NAME_TO_MODULE_MAP:
+        mod, cls = _STYLE_NAME_TO_MODULE_MAP[name]
+        builtin = "yes"
+    else:
+        for found_name, style in find_plugin_styles():
+            if name == found_name:
+                return style
+        # perhaps it got dropped into our styles package
+        builtin = ""
+        mod = 'pygments.styles.' + name
+        cls = name.title() + "Style"
+
+    try:
+        mod = __import__(mod, None, None, [cls])
+    except ImportError:
+        raise ClassNotFound(f"Could not find style module {mod!r}" +
+                            (builtin and ", though it should be builtin")
+                            + ".")
+    try:
+        return getattr(mod, cls)
+    except AttributeError:
+        raise ClassNotFound(f"Could not find style class {cls!r} in style module.")
+
+
+def get_all_styles():
+    """Return a generator for all styles by name, both builtin and plugin."""
+    for v in STYLES.values():
+        yield v[1]
+    for name, _ in find_plugin_styles():
+        yield name
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9f2cc4e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/_mapping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/_mapping.cpython-312.pyc
new file mode 100644
index 00000000..7efae012
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/__pycache__/_mapping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/_mapping.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/_mapping.py
new file mode 100644
index 00000000..49a7fae9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/styles/_mapping.py
@@ -0,0 +1,54 @@
+# Automatically generated by scripts/gen_mapfiles.py.
+# DO NOT EDIT BY HAND; run `tox -e mapfiles` instead.
+
+STYLES = {
+    'AbapStyle': ('pygments.styles.abap', 'abap', ()),
+    'AlgolStyle': ('pygments.styles.algol', 'algol', ()),
+    'Algol_NuStyle': ('pygments.styles.algol_nu', 'algol_nu', ()),
+    'ArduinoStyle': ('pygments.styles.arduino', 'arduino', ()),
+    'AutumnStyle': ('pygments.styles.autumn', 'autumn', ()),
+    'BlackWhiteStyle': ('pygments.styles.bw', 'bw', ()),
+    'BorlandStyle': ('pygments.styles.borland', 'borland', ()),
+    'CoffeeStyle': ('pygments.styles.coffee', 'coffee', ()),
+    'ColorfulStyle': ('pygments.styles.colorful', 'colorful', ()),
+    'DefaultStyle': ('pygments.styles.default', 'default', ()),
+    'DraculaStyle': ('pygments.styles.dracula', 'dracula', ()),
+    'EmacsStyle': ('pygments.styles.emacs', 'emacs', ()),
+    'FriendlyGrayscaleStyle': ('pygments.styles.friendly_grayscale', 'friendly_grayscale', ()),
+    'FriendlyStyle': ('pygments.styles.friendly', 'friendly', ()),
+    'FruityStyle': ('pygments.styles.fruity', 'fruity', ()),
+    'GhDarkStyle': ('pygments.styles.gh_dark', 'github-dark', ()),
+    'GruvboxDarkStyle': ('pygments.styles.gruvbox', 'gruvbox-dark', ()),
+    'GruvboxLightStyle': ('pygments.styles.gruvbox', 'gruvbox-light', ()),
+    'IgorStyle': ('pygments.styles.igor', 'igor', ()),
+    'InkPotStyle': ('pygments.styles.inkpot', 'inkpot', ()),
+    'LightbulbStyle': ('pygments.styles.lightbulb', 'lightbulb', ()),
+    'LilyPondStyle': ('pygments.styles.lilypond', 'lilypond', ()),
+    'LovelaceStyle': ('pygments.styles.lovelace', 'lovelace', ()),
+    'ManniStyle': ('pygments.styles.manni', 'manni', ()),
+    'MaterialStyle': ('pygments.styles.material', 'material', ()),
+    'MonokaiStyle': ('pygments.styles.monokai', 'monokai', ()),
+    'MurphyStyle': ('pygments.styles.murphy', 'murphy', ()),
+    'NativeStyle': ('pygments.styles.native', 'native', ()),
+    'NordDarkerStyle': ('pygments.styles.nord', 'nord-darker', ()),
+    'NordStyle': ('pygments.styles.nord', 'nord', ()),
+    'OneDarkStyle': ('pygments.styles.onedark', 'one-dark', ()),
+    'ParaisoDarkStyle': ('pygments.styles.paraiso_dark', 'paraiso-dark', ()),
+    'ParaisoLightStyle': ('pygments.styles.paraiso_light', 'paraiso-light', ()),
+    'PastieStyle': ('pygments.styles.pastie', 'pastie', ()),
+    'PerldocStyle': ('pygments.styles.perldoc', 'perldoc', ()),
+    'RainbowDashStyle': ('pygments.styles.rainbow_dash', 'rainbow_dash', ()),
+    'RrtStyle': ('pygments.styles.rrt', 'rrt', ()),
+    'SasStyle': ('pygments.styles.sas', 'sas', ()),
+    'SolarizedDarkStyle': ('pygments.styles.solarized', 'solarized-dark', ()),
+    'SolarizedLightStyle': ('pygments.styles.solarized', 'solarized-light', ()),
+    'StarofficeStyle': ('pygments.styles.staroffice', 'staroffice', ()),
+    'StataDarkStyle': ('pygments.styles.stata_dark', 'stata-dark', ()),
+    'StataLightStyle': ('pygments.styles.stata_light', 'stata-light', ()),
+    'TangoStyle': ('pygments.styles.tango', 'tango', ()),
+    'TracStyle': ('pygments.styles.trac', 'trac', ()),
+    'VimStyle': ('pygments.styles.vim', 'vim', ()),
+    'VisualStudioStyle': ('pygments.styles.vs', 'vs', ()),
+    'XcodeStyle': ('pygments.styles.xcode', 'xcode', ()),
+    'ZenburnStyle': ('pygments.styles.zenburn', 'zenburn', ()),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/token.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/token.py
new file mode 100644
index 00000000..f78018a7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/token.py
@@ -0,0 +1,214 @@
+"""
+    pygments.token
+    ~~~~~~~~~~~~~~
+
+    Basic token types and the standard tokens.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+
+class _TokenType(tuple):
+    parent = None
+
+    def split(self):
+        buf = []
+        node = self
+        while node is not None:
+            buf.append(node)
+            node = node.parent
+        buf.reverse()
+        return buf
+
+    def __init__(self, *args):
+        # no need to call super.__init__
+        self.subtypes = set()
+
+    def __contains__(self, val):
+        return self is val or (
+            type(val) is self.__class__ and
+            val[:len(self)] == self
+        )
+
+    def __getattr__(self, val):
+        if not val or not val[0].isupper():
+            return tuple.__getattribute__(self, val)
+        new = _TokenType(self + (val,))
+        setattr(self, val, new)
+        self.subtypes.add(new)
+        new.parent = self
+        return new
+
+    def __repr__(self):
+        return 'Token' + (self and '.' or '') + '.'.join(self)
+
+    def __copy__(self):
+        # These instances are supposed to be singletons
+        return self
+
+    def __deepcopy__(self, memo):
+        # These instances are supposed to be singletons
+        return self
+
+
+Token = _TokenType()
+
+# Special token types
+Text = Token.Text
+Whitespace = Text.Whitespace
+Escape = Token.Escape
+Error = Token.Error
+# Text that doesn't belong to this lexer (e.g. HTML in PHP)
+Other = Token.Other
+
+# Common token types for source code
+Keyword = Token.Keyword
+Name = Token.Name
+Literal = Token.Literal
+String = Literal.String
+Number = Literal.Number
+Punctuation = Token.Punctuation
+Operator = Token.Operator
+Comment = Token.Comment
+
+# Generic types for non-source code
+Generic = Token.Generic
+
+# String and some others are not direct children of Token.
+# alias them:
+Token.Token = Token
+Token.String = String
+Token.Number = Number
+
+
+def is_token_subtype(ttype, other):
+    """
+    Return True if ``ttype`` is a subtype of ``other``.
+
+    exists for backwards compatibility. use ``ttype in other`` now.
+    """
+    return ttype in other
+
+
+def string_to_tokentype(s):
+    """
+    Convert a string into a token type::
+
+        >>> string_to_token('String.Double')
+        Token.Literal.String.Double
+        >>> string_to_token('Token.Literal.Number')
+        Token.Literal.Number
+        >>> string_to_token('')
+        Token
+
+    Tokens that are already tokens are returned unchanged:
+
+        >>> string_to_token(String)
+        Token.Literal.String
+    """
+    if isinstance(s, _TokenType):
+        return s
+    if not s:
+        return Token
+    node = Token
+    for item in s.split('.'):
+        node = getattr(node, item)
+    return node
+
+
+# Map standard token types to short names, used in CSS class naming.
+# If you add a new item, please be sure to run this file to perform
+# a consistency check for duplicate values.
+STANDARD_TYPES = {
+    Token:                         '',
+
+    Text:                          '',
+    Whitespace:                    'w',
+    Escape:                        'esc',
+    Error:                         'err',
+    Other:                         'x',
+
+    Keyword:                       'k',
+    Keyword.Constant:              'kc',
+    Keyword.Declaration:           'kd',
+    Keyword.Namespace:             'kn',
+    Keyword.Pseudo:                'kp',
+    Keyword.Reserved:              'kr',
+    Keyword.Type:                  'kt',
+
+    Name:                          'n',
+    Name.Attribute:                'na',
+    Name.Builtin:                  'nb',
+    Name.Builtin.Pseudo:           'bp',
+    Name.Class:                    'nc',
+    Name.Constant:                 'no',
+    Name.Decorator:                'nd',
+    Name.Entity:                   'ni',
+    Name.Exception:                'ne',
+    Name.Function:                 'nf',
+    Name.Function.Magic:           'fm',
+    Name.Property:                 'py',
+    Name.Label:                    'nl',
+    Name.Namespace:                'nn',
+    Name.Other:                    'nx',
+    Name.Tag:                      'nt',
+    Name.Variable:                 'nv',
+    Name.Variable.Class:           'vc',
+    Name.Variable.Global:          'vg',
+    Name.Variable.Instance:        'vi',
+    Name.Variable.Magic:           'vm',
+
+    Literal:                       'l',
+    Literal.Date:                  'ld',
+
+    String:                        's',
+    String.Affix:                  'sa',
+    String.Backtick:               'sb',
+    String.Char:                   'sc',
+    String.Delimiter:              'dl',
+    String.Doc:                    'sd',
+    String.Double:                 's2',
+    String.Escape:                 'se',
+    String.Heredoc:                'sh',
+    String.Interpol:               'si',
+    String.Other:                  'sx',
+    String.Regex:                  'sr',
+    String.Single:                 's1',
+    String.Symbol:                 'ss',
+
+    Number:                        'm',
+    Number.Bin:                    'mb',
+    Number.Float:                  'mf',
+    Number.Hex:                    'mh',
+    Number.Integer:                'mi',
+    Number.Integer.Long:           'il',
+    Number.Oct:                    'mo',
+
+    Operator:                      'o',
+    Operator.Word:                 'ow',
+
+    Punctuation:                   'p',
+    Punctuation.Marker:            'pm',
+
+    Comment:                       'c',
+    Comment.Hashbang:              'ch',
+    Comment.Multiline:             'cm',
+    Comment.Preproc:               'cp',
+    Comment.PreprocFile:           'cpf',
+    Comment.Single:                'c1',
+    Comment.Special:               'cs',
+
+    Generic:                       'g',
+    Generic.Deleted:               'gd',
+    Generic.Emph:                  'ge',
+    Generic.Error:                 'gr',
+    Generic.Heading:               'gh',
+    Generic.Inserted:              'gi',
+    Generic.Output:                'go',
+    Generic.Prompt:                'gp',
+    Generic.Strong:                'gs',
+    Generic.Subheading:            'gu',
+    Generic.EmphStrong:            'ges',
+    Generic.Traceback:             'gt',
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/unistring.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/unistring.py
new file mode 100644
index 00000000..e2c3523e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/unistring.py
@@ -0,0 +1,153 @@
+"""
+    pygments.unistring
+    ~~~~~~~~~~~~~~~~~~
+
+    Strings of all Unicode characters of a certain category.
+    Used for matching in Unicode-aware languages. Run to regenerate.
+
+    Inspired by chartypes_create.py from the MoinMoin project.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+Cc = '\x00-\x1f\x7f-\x9f'
+
+Cf = '\xad\u0600-\u0605\u061c\u06dd\u070f\u08e2\u180e\u200b-\u200f\u202a-\u202e\u2060-\u2064\u2066-\u206f\ufeff\ufff9-\ufffb\U000110bd\U000110cd\U0001bca0-\U0001bca3\U0001d173-\U0001d17a\U000e0001\U000e0020-\U000e007f'
+
+Cn = '\u0378-\u0379\u0380-\u0383\u038b\u038d\u03a2\u0530\u0557-\u0558\u058b-\u058c\u0590\u05c8-\u05cf\u05eb-\u05ee\u05f5-\u05ff\u061d\u070e\u074b-\u074c\u07b2-\u07bf\u07fb-\u07fc\u082e-\u082f\u083f\u085c-\u085d\u085f\u086b-\u089f\u08b5\u08be-\u08d2\u0984\u098d-\u098e\u0991-\u0992\u09a9\u09b1\u09b3-\u09b5\u09ba-\u09bb\u09c5-\u09c6\u09c9-\u09ca\u09cf-\u09d6\u09d8-\u09db\u09de\u09e4-\u09e5\u09ff-\u0a00\u0a04\u0a0b-\u0a0e\u0a11-\u0a12\u0a29\u0a31\u0a34\u0a37\u0a3a-\u0a3b\u0a3d\u0a43-\u0a46\u0a49-\u0a4a\u0a4e-\u0a50\u0a52-\u0a58\u0a5d\u0a5f-\u0a65\u0a77-\u0a80\u0a84\u0a8e\u0a92\u0aa9\u0ab1\u0ab4\u0aba-\u0abb\u0ac6\u0aca\u0ace-\u0acf\u0ad1-\u0adf\u0ae4-\u0ae5\u0af2-\u0af8\u0b00\u0b04\u0b0d-\u0b0e\u0b11-\u0b12\u0b29\u0b31\u0b34\u0b3a-\u0b3b\u0b45-\u0b46\u0b49-\u0b4a\u0b4e-\u0b55\u0b58-\u0b5b\u0b5e\u0b64-\u0b65\u0b78-\u0b81\u0b84\u0b8b-\u0b8d\u0b91\u0b96-\u0b98\u0b9b\u0b9d\u0ba0-\u0ba2\u0ba5-\u0ba7\u0bab-\u0bad\u0bba-\u0bbd\u0bc3-\u0bc5\u0bc9\u0bce-\u0bcf\u0bd1-\u0bd6\u0bd8-\u0be5\u0bfb-\u0bff\u0c0d\u0c11\u0c29\u0c3a-\u0c3c\u0c45\u0c49\u0c4e-\u0c54\u0c57\u0c5b-\u0c5f\u0c64-\u0c65\u0c70-\u0c77\u0c8d\u0c91\u0ca9\u0cb4\u0cba-\u0cbb\u0cc5\u0cc9\u0cce-\u0cd4\u0cd7-\u0cdd\u0cdf\u0ce4-\u0ce5\u0cf0\u0cf3-\u0cff\u0d04\u0d0d\u0d11\u0d45\u0d49\u0d50-\u0d53\u0d64-\u0d65\u0d80-\u0d81\u0d84\u0d97-\u0d99\u0db2\u0dbc\u0dbe-\u0dbf\u0dc7-\u0dc9\u0dcb-\u0dce\u0dd5\u0dd7\u0de0-\u0de5\u0df0-\u0df1\u0df5-\u0e00\u0e3b-\u0e3e\u0e5c-\u0e80\u0e83\u0e85-\u0e86\u0e89\u0e8b-\u0e8c\u0e8e-\u0e93\u0e98\u0ea0\u0ea4\u0ea6\u0ea8-\u0ea9\u0eac\u0eba\u0ebe-\u0ebf\u0ec5\u0ec7\u0ece-\u0ecf\u0eda-\u0edb\u0ee0-\u0eff\u0f48\u0f6d-\u0f70\u0f98\u0fbd\u0fcd\u0fdb-\u0fff\u10c6\u10c8-\u10cc\u10ce-\u10cf\u1249\u124e-\u124f\u1257\u1259\u125e-\u125f\u1289\u128e-\u128f\u12b1\u12b6-\u12b7\u12bf\u12c1\u12c6-\u12c7\u12d7\u1311\u1316-\u1317\u135b-\u135c\u137d-\u137f\u139a-\u139f\u13f6-\u13f7\u13fe-\u13ff\u169d-\u169f\u16f9-\u16ff\u170d\u1715-\u171f\u1737-\u173f\u1754-\u175f\u176d\u1771\u1774-\u177f\u17de-\u17df\u17ea-\u17ef\u17fa-\u17ff\u180f\u181a-\u181f\u1879-\u187f\u18ab-\u18af\u18f6-\u18ff\u191f\u192c-\u192f\u193c-\u193f\u1941-\u1943\u196e-\u196f\u1975-\u197f\u19ac-\u19af\u19ca-\u19cf\u19db-\u19dd\u1a1c-\u1a1d\u1a5f\u1a7d-\u1a7e\u1a8a-\u1a8f\u1a9a-\u1a9f\u1aae-\u1aaf\u1abf-\u1aff\u1b4c-\u1b4f\u1b7d-\u1b7f\u1bf4-\u1bfb\u1c38-\u1c3a\u1c4a-\u1c4c\u1c89-\u1c8f\u1cbb-\u1cbc\u1cc8-\u1ccf\u1cfa-\u1cff\u1dfa\u1f16-\u1f17\u1f1e-\u1f1f\u1f46-\u1f47\u1f4e-\u1f4f\u1f58\u1f5a\u1f5c\u1f5e\u1f7e-\u1f7f\u1fb5\u1fc5\u1fd4-\u1fd5\u1fdc\u1ff0-\u1ff1\u1ff5\u1fff\u2065\u2072-\u2073\u208f\u209d-\u209f\u20c0-\u20cf\u20f1-\u20ff\u218c-\u218f\u2427-\u243f\u244b-\u245f\u2b74-\u2b75\u2b96-\u2b97\u2bc9\u2bff\u2c2f\u2c5f\u2cf4-\u2cf8\u2d26\u2d28-\u2d2c\u2d2e-\u2d2f\u2d68-\u2d6e\u2d71-\u2d7e\u2d97-\u2d9f\u2da7\u2daf\u2db7\u2dbf\u2dc7\u2dcf\u2dd7\u2ddf\u2e4f-\u2e7f\u2e9a\u2ef4-\u2eff\u2fd6-\u2fef\u2ffc-\u2fff\u3040\u3097-\u3098\u3100-\u3104\u3130\u318f\u31bb-\u31bf\u31e4-\u31ef\u321f\u32ff\u4db6-\u4dbf\u9ff0-\u9fff\ua48d-\ua48f\ua4c7-\ua4cf\ua62c-\ua63f\ua6f8-\ua6ff\ua7ba-\ua7f6\ua82c-\ua82f\ua83a-\ua83f\ua878-\ua87f\ua8c6-\ua8cd\ua8da-\ua8df\ua954-\ua95e\ua97d-\ua97f\ua9ce\ua9da-\ua9dd\ua9ff\uaa37-\uaa3f\uaa4e-\uaa4f\uaa5a-\uaa5b\uaac3-\uaada\uaaf7-\uab00\uab07-\uab08\uab0f-\uab10\uab17-\uab1f\uab27\uab2f\uab66-\uab6f\uabee-\uabef\uabfa-\uabff\ud7a4-\ud7af\ud7c7-\ud7ca\ud7fc-\ud7ff\ufa6e-\ufa6f\ufada-\ufaff\ufb07-\ufb12\ufb18-\ufb1c\ufb37\ufb3d\ufb3f\ufb42\ufb45\ufbc2-\ufbd2\ufd40-\ufd4f\ufd90-\ufd91\ufdc8-\ufdef\ufdfe-\ufdff\ufe1a-\ufe1f\ufe53\ufe67\ufe6c-\ufe6f\ufe75\ufefd-\ufefe\uff00\uffbf-\uffc1\uffc8-\uffc9\uffd0-\uffd1\uffd8-\uffd9\uffdd-\uffdf\uffe7\uffef-\ufff8\ufffe-\uffff\U0001000c\U00010027\U0001003b\U0001003e\U0001004e-\U0001004f\U0001005e-\U0001007f\U000100fb-\U000100ff\U00010103-\U00010106\U00010134-\U00010136\U0001018f\U0001019c-\U0001019f\U000101a1-\U000101cf\U000101fe-\U0001027f\U0001029d-\U0001029f\U000102d1-\U000102df\U000102fc-\U000102ff\U00010324-\U0001032c\U0001034b-\U0001034f\U0001037b-\U0001037f\U0001039e\U000103c4-\U000103c7\U000103d6-\U000103ff\U0001049e-\U0001049f\U000104aa-\U000104af\U000104d4-\U000104d7\U000104fc-\U000104ff\U00010528-\U0001052f\U00010564-\U0001056e\U00010570-\U000105ff\U00010737-\U0001073f\U00010756-\U0001075f\U00010768-\U000107ff\U00010806-\U00010807\U00010809\U00010836\U00010839-\U0001083b\U0001083d-\U0001083e\U00010856\U0001089f-\U000108a6\U000108b0-\U000108df\U000108f3\U000108f6-\U000108fa\U0001091c-\U0001091e\U0001093a-\U0001093e\U00010940-\U0001097f\U000109b8-\U000109bb\U000109d0-\U000109d1\U00010a04\U00010a07-\U00010a0b\U00010a14\U00010a18\U00010a36-\U00010a37\U00010a3b-\U00010a3e\U00010a49-\U00010a4f\U00010a59-\U00010a5f\U00010aa0-\U00010abf\U00010ae7-\U00010aea\U00010af7-\U00010aff\U00010b36-\U00010b38\U00010b56-\U00010b57\U00010b73-\U00010b77\U00010b92-\U00010b98\U00010b9d-\U00010ba8\U00010bb0-\U00010bff\U00010c49-\U00010c7f\U00010cb3-\U00010cbf\U00010cf3-\U00010cf9\U00010d28-\U00010d2f\U00010d3a-\U00010e5f\U00010e7f-\U00010eff\U00010f28-\U00010f2f\U00010f5a-\U00010fff\U0001104e-\U00011051\U00011070-\U0001107e\U000110c2-\U000110cc\U000110ce-\U000110cf\U000110e9-\U000110ef\U000110fa-\U000110ff\U00011135\U00011147-\U0001114f\U00011177-\U0001117f\U000111ce-\U000111cf\U000111e0\U000111f5-\U000111ff\U00011212\U0001123f-\U0001127f\U00011287\U00011289\U0001128e\U0001129e\U000112aa-\U000112af\U000112eb-\U000112ef\U000112fa-\U000112ff\U00011304\U0001130d-\U0001130e\U00011311-\U00011312\U00011329\U00011331\U00011334\U0001133a\U00011345-\U00011346\U00011349-\U0001134a\U0001134e-\U0001134f\U00011351-\U00011356\U00011358-\U0001135c\U00011364-\U00011365\U0001136d-\U0001136f\U00011375-\U000113ff\U0001145a\U0001145c\U0001145f-\U0001147f\U000114c8-\U000114cf\U000114da-\U0001157f\U000115b6-\U000115b7\U000115de-\U000115ff\U00011645-\U0001164f\U0001165a-\U0001165f\U0001166d-\U0001167f\U000116b8-\U000116bf\U000116ca-\U000116ff\U0001171b-\U0001171c\U0001172c-\U0001172f\U00011740-\U000117ff\U0001183c-\U0001189f\U000118f3-\U000118fe\U00011900-\U000119ff\U00011a48-\U00011a4f\U00011a84-\U00011a85\U00011aa3-\U00011abf\U00011af9-\U00011bff\U00011c09\U00011c37\U00011c46-\U00011c4f\U00011c6d-\U00011c6f\U00011c90-\U00011c91\U00011ca8\U00011cb7-\U00011cff\U00011d07\U00011d0a\U00011d37-\U00011d39\U00011d3b\U00011d3e\U00011d48-\U00011d4f\U00011d5a-\U00011d5f\U00011d66\U00011d69\U00011d8f\U00011d92\U00011d99-\U00011d9f\U00011daa-\U00011edf\U00011ef9-\U00011fff\U0001239a-\U000123ff\U0001246f\U00012475-\U0001247f\U00012544-\U00012fff\U0001342f-\U000143ff\U00014647-\U000167ff\U00016a39-\U00016a3f\U00016a5f\U00016a6a-\U00016a6d\U00016a70-\U00016acf\U00016aee-\U00016aef\U00016af6-\U00016aff\U00016b46-\U00016b4f\U00016b5a\U00016b62\U00016b78-\U00016b7c\U00016b90-\U00016e3f\U00016e9b-\U00016eff\U00016f45-\U00016f4f\U00016f7f-\U00016f8e\U00016fa0-\U00016fdf\U00016fe2-\U00016fff\U000187f2-\U000187ff\U00018af3-\U0001afff\U0001b11f-\U0001b16f\U0001b2fc-\U0001bbff\U0001bc6b-\U0001bc6f\U0001bc7d-\U0001bc7f\U0001bc89-\U0001bc8f\U0001bc9a-\U0001bc9b\U0001bca4-\U0001cfff\U0001d0f6-\U0001d0ff\U0001d127-\U0001d128\U0001d1e9-\U0001d1ff\U0001d246-\U0001d2df\U0001d2f4-\U0001d2ff\U0001d357-\U0001d35f\U0001d379-\U0001d3ff\U0001d455\U0001d49d\U0001d4a0-\U0001d4a1\U0001d4a3-\U0001d4a4\U0001d4a7-\U0001d4a8\U0001d4ad\U0001d4ba\U0001d4bc\U0001d4c4\U0001d506\U0001d50b-\U0001d50c\U0001d515\U0001d51d\U0001d53a\U0001d53f\U0001d545\U0001d547-\U0001d549\U0001d551\U0001d6a6-\U0001d6a7\U0001d7cc-\U0001d7cd\U0001da8c-\U0001da9a\U0001daa0\U0001dab0-\U0001dfff\U0001e007\U0001e019-\U0001e01a\U0001e022\U0001e025\U0001e02b-\U0001e7ff\U0001e8c5-\U0001e8c6\U0001e8d7-\U0001e8ff\U0001e94b-\U0001e94f\U0001e95a-\U0001e95d\U0001e960-\U0001ec70\U0001ecb5-\U0001edff\U0001ee04\U0001ee20\U0001ee23\U0001ee25-\U0001ee26\U0001ee28\U0001ee33\U0001ee38\U0001ee3a\U0001ee3c-\U0001ee41\U0001ee43-\U0001ee46\U0001ee48\U0001ee4a\U0001ee4c\U0001ee50\U0001ee53\U0001ee55-\U0001ee56\U0001ee58\U0001ee5a\U0001ee5c\U0001ee5e\U0001ee60\U0001ee63\U0001ee65-\U0001ee66\U0001ee6b\U0001ee73\U0001ee78\U0001ee7d\U0001ee7f\U0001ee8a\U0001ee9c-\U0001eea0\U0001eea4\U0001eeaa\U0001eebc-\U0001eeef\U0001eef2-\U0001efff\U0001f02c-\U0001f02f\U0001f094-\U0001f09f\U0001f0af-\U0001f0b0\U0001f0c0\U0001f0d0\U0001f0f6-\U0001f0ff\U0001f10d-\U0001f10f\U0001f16c-\U0001f16f\U0001f1ad-\U0001f1e5\U0001f203-\U0001f20f\U0001f23c-\U0001f23f\U0001f249-\U0001f24f\U0001f252-\U0001f25f\U0001f266-\U0001f2ff\U0001f6d5-\U0001f6df\U0001f6ed-\U0001f6ef\U0001f6fa-\U0001f6ff\U0001f774-\U0001f77f\U0001f7d9-\U0001f7ff\U0001f80c-\U0001f80f\U0001f848-\U0001f84f\U0001f85a-\U0001f85f\U0001f888-\U0001f88f\U0001f8ae-\U0001f8ff\U0001f90c-\U0001f90f\U0001f93f\U0001f971-\U0001f972\U0001f977-\U0001f979\U0001f97b\U0001f9a3-\U0001f9af\U0001f9ba-\U0001f9bf\U0001f9c3-\U0001f9cf\U0001fa00-\U0001fa5f\U0001fa6e-\U0001ffff\U0002a6d7-\U0002a6ff\U0002b735-\U0002b73f\U0002b81e-\U0002b81f\U0002cea2-\U0002ceaf\U0002ebe1-\U0002f7ff\U0002fa1e-\U000e0000\U000e0002-\U000e001f\U000e0080-\U000e00ff\U000e01f0-\U000effff\U000ffffe-\U000fffff\U0010fffe-\U0010ffff'
+
+Co = '\ue000-\uf8ff\U000f0000-\U000ffffd\U00100000-\U0010fffd'
+
+Cs = '\ud800-\udbff\\\udc00\udc01-\udfff'
+
+Ll = 'a-z\xb5\xdf-\xf6\xf8-\xff\u0101\u0103\u0105\u0107\u0109\u010b\u010d\u010f\u0111\u0113\u0115\u0117\u0119\u011b\u011d\u011f\u0121\u0123\u0125\u0127\u0129\u012b\u012d\u012f\u0131\u0133\u0135\u0137-\u0138\u013a\u013c\u013e\u0140\u0142\u0144\u0146\u0148-\u0149\u014b\u014d\u014f\u0151\u0153\u0155\u0157\u0159\u015b\u015d\u015f\u0161\u0163\u0165\u0167\u0169\u016b\u016d\u016f\u0171\u0173\u0175\u0177\u017a\u017c\u017e-\u0180\u0183\u0185\u0188\u018c-\u018d\u0192\u0195\u0199-\u019b\u019e\u01a1\u01a3\u01a5\u01a8\u01aa-\u01ab\u01ad\u01b0\u01b4\u01b6\u01b9-\u01ba\u01bd-\u01bf\u01c6\u01c9\u01cc\u01ce\u01d0\u01d2\u01d4\u01d6\u01d8\u01da\u01dc-\u01dd\u01df\u01e1\u01e3\u01e5\u01e7\u01e9\u01eb\u01ed\u01ef-\u01f0\u01f3\u01f5\u01f9\u01fb\u01fd\u01ff\u0201\u0203\u0205\u0207\u0209\u020b\u020d\u020f\u0211\u0213\u0215\u0217\u0219\u021b\u021d\u021f\u0221\u0223\u0225\u0227\u0229\u022b\u022d\u022f\u0231\u0233-\u0239\u023c\u023f-\u0240\u0242\u0247\u0249\u024b\u024d\u024f-\u0293\u0295-\u02af\u0371\u0373\u0377\u037b-\u037d\u0390\u03ac-\u03ce\u03d0-\u03d1\u03d5-\u03d7\u03d9\u03db\u03dd\u03df\u03e1\u03e3\u03e5\u03e7\u03e9\u03eb\u03ed\u03ef-\u03f3\u03f5\u03f8\u03fb-\u03fc\u0430-\u045f\u0461\u0463\u0465\u0467\u0469\u046b\u046d\u046f\u0471\u0473\u0475\u0477\u0479\u047b\u047d\u047f\u0481\u048b\u048d\u048f\u0491\u0493\u0495\u0497\u0499\u049b\u049d\u049f\u04a1\u04a3\u04a5\u04a7\u04a9\u04ab\u04ad\u04af\u04b1\u04b3\u04b5\u04b7\u04b9\u04bb\u04bd\u04bf\u04c2\u04c4\u04c6\u04c8\u04ca\u04cc\u04ce-\u04cf\u04d1\u04d3\u04d5\u04d7\u04d9\u04db\u04dd\u04df\u04e1\u04e3\u04e5\u04e7\u04e9\u04eb\u04ed\u04ef\u04f1\u04f3\u04f5\u04f7\u04f9\u04fb\u04fd\u04ff\u0501\u0503\u0505\u0507\u0509\u050b\u050d\u050f\u0511\u0513\u0515\u0517\u0519\u051b\u051d\u051f\u0521\u0523\u0525\u0527\u0529\u052b\u052d\u052f\u0560-\u0588\u10d0-\u10fa\u10fd-\u10ff\u13f8-\u13fd\u1c80-\u1c88\u1d00-\u1d2b\u1d6b-\u1d77\u1d79-\u1d9a\u1e01\u1e03\u1e05\u1e07\u1e09\u1e0b\u1e0d\u1e0f\u1e11\u1e13\u1e15\u1e17\u1e19\u1e1b\u1e1d\u1e1f\u1e21\u1e23\u1e25\u1e27\u1e29\u1e2b\u1e2d\u1e2f\u1e31\u1e33\u1e35\u1e37\u1e39\u1e3b\u1e3d\u1e3f\u1e41\u1e43\u1e45\u1e47\u1e49\u1e4b\u1e4d\u1e4f\u1e51\u1e53\u1e55\u1e57\u1e59\u1e5b\u1e5d\u1e5f\u1e61\u1e63\u1e65\u1e67\u1e69\u1e6b\u1e6d\u1e6f\u1e71\u1e73\u1e75\u1e77\u1e79\u1e7b\u1e7d\u1e7f\u1e81\u1e83\u1e85\u1e87\u1e89\u1e8b\u1e8d\u1e8f\u1e91\u1e93\u1e95-\u1e9d\u1e9f\u1ea1\u1ea3\u1ea5\u1ea7\u1ea9\u1eab\u1ead\u1eaf\u1eb1\u1eb3\u1eb5\u1eb7\u1eb9\u1ebb\u1ebd\u1ebf\u1ec1\u1ec3\u1ec5\u1ec7\u1ec9\u1ecb\u1ecd\u1ecf\u1ed1\u1ed3\u1ed5\u1ed7\u1ed9\u1edb\u1edd\u1edf\u1ee1\u1ee3\u1ee5\u1ee7\u1ee9\u1eeb\u1eed\u1eef\u1ef1\u1ef3\u1ef5\u1ef7\u1ef9\u1efb\u1efd\u1eff-\u1f07\u1f10-\u1f15\u1f20-\u1f27\u1f30-\u1f37\u1f40-\u1f45\u1f50-\u1f57\u1f60-\u1f67\u1f70-\u1f7d\u1f80-\u1f87\u1f90-\u1f97\u1fa0-\u1fa7\u1fb0-\u1fb4\u1fb6-\u1fb7\u1fbe\u1fc2-\u1fc4\u1fc6-\u1fc7\u1fd0-\u1fd3\u1fd6-\u1fd7\u1fe0-\u1fe7\u1ff2-\u1ff4\u1ff6-\u1ff7\u210a\u210e-\u210f\u2113\u212f\u2134\u2139\u213c-\u213d\u2146-\u2149\u214e\u2184\u2c30-\u2c5e\u2c61\u2c65-\u2c66\u2c68\u2c6a\u2c6c\u2c71\u2c73-\u2c74\u2c76-\u2c7b\u2c81\u2c83\u2c85\u2c87\u2c89\u2c8b\u2c8d\u2c8f\u2c91\u2c93\u2c95\u2c97\u2c99\u2c9b\u2c9d\u2c9f\u2ca1\u2ca3\u2ca5\u2ca7\u2ca9\u2cab\u2cad\u2caf\u2cb1\u2cb3\u2cb5\u2cb7\u2cb9\u2cbb\u2cbd\u2cbf\u2cc1\u2cc3\u2cc5\u2cc7\u2cc9\u2ccb\u2ccd\u2ccf\u2cd1\u2cd3\u2cd5\u2cd7\u2cd9\u2cdb\u2cdd\u2cdf\u2ce1\u2ce3-\u2ce4\u2cec\u2cee\u2cf3\u2d00-\u2d25\u2d27\u2d2d\ua641\ua643\ua645\ua647\ua649\ua64b\ua64d\ua64f\ua651\ua653\ua655\ua657\ua659\ua65b\ua65d\ua65f\ua661\ua663\ua665\ua667\ua669\ua66b\ua66d\ua681\ua683\ua685\ua687\ua689\ua68b\ua68d\ua68f\ua691\ua693\ua695\ua697\ua699\ua69b\ua723\ua725\ua727\ua729\ua72b\ua72d\ua72f-\ua731\ua733\ua735\ua737\ua739\ua73b\ua73d\ua73f\ua741\ua743\ua745\ua747\ua749\ua74b\ua74d\ua74f\ua751\ua753\ua755\ua757\ua759\ua75b\ua75d\ua75f\ua761\ua763\ua765\ua767\ua769\ua76b\ua76d\ua76f\ua771-\ua778\ua77a\ua77c\ua77f\ua781\ua783\ua785\ua787\ua78c\ua78e\ua791\ua793-\ua795\ua797\ua799\ua79b\ua79d\ua79f\ua7a1\ua7a3\ua7a5\ua7a7\ua7a9\ua7af\ua7b5\ua7b7\ua7b9\ua7fa\uab30-\uab5a\uab60-\uab65\uab70-\uabbf\ufb00-\ufb06\ufb13-\ufb17\uff41-\uff5a\U00010428-\U0001044f\U000104d8-\U000104fb\U00010cc0-\U00010cf2\U000118c0-\U000118df\U00016e60-\U00016e7f\U0001d41a-\U0001d433\U0001d44e-\U0001d454\U0001d456-\U0001d467\U0001d482-\U0001d49b\U0001d4b6-\U0001d4b9\U0001d4bb\U0001d4bd-\U0001d4c3\U0001d4c5-\U0001d4cf\U0001d4ea-\U0001d503\U0001d51e-\U0001d537\U0001d552-\U0001d56b\U0001d586-\U0001d59f\U0001d5ba-\U0001d5d3\U0001d5ee-\U0001d607\U0001d622-\U0001d63b\U0001d656-\U0001d66f\U0001d68a-\U0001d6a5\U0001d6c2-\U0001d6da\U0001d6dc-\U0001d6e1\U0001d6fc-\U0001d714\U0001d716-\U0001d71b\U0001d736-\U0001d74e\U0001d750-\U0001d755\U0001d770-\U0001d788\U0001d78a-\U0001d78f\U0001d7aa-\U0001d7c2\U0001d7c4-\U0001d7c9\U0001d7cb\U0001e922-\U0001e943'
+
+Lm = '\u02b0-\u02c1\u02c6-\u02d1\u02e0-\u02e4\u02ec\u02ee\u0374\u037a\u0559\u0640\u06e5-\u06e6\u07f4-\u07f5\u07fa\u081a\u0824\u0828\u0971\u0e46\u0ec6\u10fc\u17d7\u1843\u1aa7\u1c78-\u1c7d\u1d2c-\u1d6a\u1d78\u1d9b-\u1dbf\u2071\u207f\u2090-\u209c\u2c7c-\u2c7d\u2d6f\u2e2f\u3005\u3031-\u3035\u303b\u309d-\u309e\u30fc-\u30fe\ua015\ua4f8-\ua4fd\ua60c\ua67f\ua69c-\ua69d\ua717-\ua71f\ua770\ua788\ua7f8-\ua7f9\ua9cf\ua9e6\uaa70\uaadd\uaaf3-\uaaf4\uab5c-\uab5f\uff70\uff9e-\uff9f\U00016b40-\U00016b43\U00016f93-\U00016f9f\U00016fe0-\U00016fe1'
+
+Lo = '\xaa\xba\u01bb\u01c0-\u01c3\u0294\u05d0-\u05ea\u05ef-\u05f2\u0620-\u063f\u0641-\u064a\u066e-\u066f\u0671-\u06d3\u06d5\u06ee-\u06ef\u06fa-\u06fc\u06ff\u0710\u0712-\u072f\u074d-\u07a5\u07b1\u07ca-\u07ea\u0800-\u0815\u0840-\u0858\u0860-\u086a\u08a0-\u08b4\u08b6-\u08bd\u0904-\u0939\u093d\u0950\u0958-\u0961\u0972-\u0980\u0985-\u098c\u098f-\u0990\u0993-\u09a8\u09aa-\u09b0\u09b2\u09b6-\u09b9\u09bd\u09ce\u09dc-\u09dd\u09df-\u09e1\u09f0-\u09f1\u09fc\u0a05-\u0a0a\u0a0f-\u0a10\u0a13-\u0a28\u0a2a-\u0a30\u0a32-\u0a33\u0a35-\u0a36\u0a38-\u0a39\u0a59-\u0a5c\u0a5e\u0a72-\u0a74\u0a85-\u0a8d\u0a8f-\u0a91\u0a93-\u0aa8\u0aaa-\u0ab0\u0ab2-\u0ab3\u0ab5-\u0ab9\u0abd\u0ad0\u0ae0-\u0ae1\u0af9\u0b05-\u0b0c\u0b0f-\u0b10\u0b13-\u0b28\u0b2a-\u0b30\u0b32-\u0b33\u0b35-\u0b39\u0b3d\u0b5c-\u0b5d\u0b5f-\u0b61\u0b71\u0b83\u0b85-\u0b8a\u0b8e-\u0b90\u0b92-\u0b95\u0b99-\u0b9a\u0b9c\u0b9e-\u0b9f\u0ba3-\u0ba4\u0ba8-\u0baa\u0bae-\u0bb9\u0bd0\u0c05-\u0c0c\u0c0e-\u0c10\u0c12-\u0c28\u0c2a-\u0c39\u0c3d\u0c58-\u0c5a\u0c60-\u0c61\u0c80\u0c85-\u0c8c\u0c8e-\u0c90\u0c92-\u0ca8\u0caa-\u0cb3\u0cb5-\u0cb9\u0cbd\u0cde\u0ce0-\u0ce1\u0cf1-\u0cf2\u0d05-\u0d0c\u0d0e-\u0d10\u0d12-\u0d3a\u0d3d\u0d4e\u0d54-\u0d56\u0d5f-\u0d61\u0d7a-\u0d7f\u0d85-\u0d96\u0d9a-\u0db1\u0db3-\u0dbb\u0dbd\u0dc0-\u0dc6\u0e01-\u0e30\u0e32-\u0e33\u0e40-\u0e45\u0e81-\u0e82\u0e84\u0e87-\u0e88\u0e8a\u0e8d\u0e94-\u0e97\u0e99-\u0e9f\u0ea1-\u0ea3\u0ea5\u0ea7\u0eaa-\u0eab\u0ead-\u0eb0\u0eb2-\u0eb3\u0ebd\u0ec0-\u0ec4\u0edc-\u0edf\u0f00\u0f40-\u0f47\u0f49-\u0f6c\u0f88-\u0f8c\u1000-\u102a\u103f\u1050-\u1055\u105a-\u105d\u1061\u1065-\u1066\u106e-\u1070\u1075-\u1081\u108e\u1100-\u1248\u124a-\u124d\u1250-\u1256\u1258\u125a-\u125d\u1260-\u1288\u128a-\u128d\u1290-\u12b0\u12b2-\u12b5\u12b8-\u12be\u12c0\u12c2-\u12c5\u12c8-\u12d6\u12d8-\u1310\u1312-\u1315\u1318-\u135a\u1380-\u138f\u1401-\u166c\u166f-\u167f\u1681-\u169a\u16a0-\u16ea\u16f1-\u16f8\u1700-\u170c\u170e-\u1711\u1720-\u1731\u1740-\u1751\u1760-\u176c\u176e-\u1770\u1780-\u17b3\u17dc\u1820-\u1842\u1844-\u1878\u1880-\u1884\u1887-\u18a8\u18aa\u18b0-\u18f5\u1900-\u191e\u1950-\u196d\u1970-\u1974\u1980-\u19ab\u19b0-\u19c9\u1a00-\u1a16\u1a20-\u1a54\u1b05-\u1b33\u1b45-\u1b4b\u1b83-\u1ba0\u1bae-\u1baf\u1bba-\u1be5\u1c00-\u1c23\u1c4d-\u1c4f\u1c5a-\u1c77\u1ce9-\u1cec\u1cee-\u1cf1\u1cf5-\u1cf6\u2135-\u2138\u2d30-\u2d67\u2d80-\u2d96\u2da0-\u2da6\u2da8-\u2dae\u2db0-\u2db6\u2db8-\u2dbe\u2dc0-\u2dc6\u2dc8-\u2dce\u2dd0-\u2dd6\u2dd8-\u2dde\u3006\u303c\u3041-\u3096\u309f\u30a1-\u30fa\u30ff\u3105-\u312f\u3131-\u318e\u31a0-\u31ba\u31f0-\u31ff\u3400-\u4db5\u4e00-\u9fef\ua000-\ua014\ua016-\ua48c\ua4d0-\ua4f7\ua500-\ua60b\ua610-\ua61f\ua62a-\ua62b\ua66e\ua6a0-\ua6e5\ua78f\ua7f7\ua7fb-\ua801\ua803-\ua805\ua807-\ua80a\ua80c-\ua822\ua840-\ua873\ua882-\ua8b3\ua8f2-\ua8f7\ua8fb\ua8fd-\ua8fe\ua90a-\ua925\ua930-\ua946\ua960-\ua97c\ua984-\ua9b2\ua9e0-\ua9e4\ua9e7-\ua9ef\ua9fa-\ua9fe\uaa00-\uaa28\uaa40-\uaa42\uaa44-\uaa4b\uaa60-\uaa6f\uaa71-\uaa76\uaa7a\uaa7e-\uaaaf\uaab1\uaab5-\uaab6\uaab9-\uaabd\uaac0\uaac2\uaadb-\uaadc\uaae0-\uaaea\uaaf2\uab01-\uab06\uab09-\uab0e\uab11-\uab16\uab20-\uab26\uab28-\uab2e\uabc0-\uabe2\uac00-\ud7a3\ud7b0-\ud7c6\ud7cb-\ud7fb\uf900-\ufa6d\ufa70-\ufad9\ufb1d\ufb1f-\ufb28\ufb2a-\ufb36\ufb38-\ufb3c\ufb3e\ufb40-\ufb41\ufb43-\ufb44\ufb46-\ufbb1\ufbd3-\ufd3d\ufd50-\ufd8f\ufd92-\ufdc7\ufdf0-\ufdfb\ufe70-\ufe74\ufe76-\ufefc\uff66-\uff6f\uff71-\uff9d\uffa0-\uffbe\uffc2-\uffc7\uffca-\uffcf\uffd2-\uffd7\uffda-\uffdc\U00010000-\U0001000b\U0001000d-\U00010026\U00010028-\U0001003a\U0001003c-\U0001003d\U0001003f-\U0001004d\U00010050-\U0001005d\U00010080-\U000100fa\U00010280-\U0001029c\U000102a0-\U000102d0\U00010300-\U0001031f\U0001032d-\U00010340\U00010342-\U00010349\U00010350-\U00010375\U00010380-\U0001039d\U000103a0-\U000103c3\U000103c8-\U000103cf\U00010450-\U0001049d\U00010500-\U00010527\U00010530-\U00010563\U00010600-\U00010736\U00010740-\U00010755\U00010760-\U00010767\U00010800-\U00010805\U00010808\U0001080a-\U00010835\U00010837-\U00010838\U0001083c\U0001083f-\U00010855\U00010860-\U00010876\U00010880-\U0001089e\U000108e0-\U000108f2\U000108f4-\U000108f5\U00010900-\U00010915\U00010920-\U00010939\U00010980-\U000109b7\U000109be-\U000109bf\U00010a00\U00010a10-\U00010a13\U00010a15-\U00010a17\U00010a19-\U00010a35\U00010a60-\U00010a7c\U00010a80-\U00010a9c\U00010ac0-\U00010ac7\U00010ac9-\U00010ae4\U00010b00-\U00010b35\U00010b40-\U00010b55\U00010b60-\U00010b72\U00010b80-\U00010b91\U00010c00-\U00010c48\U00010d00-\U00010d23\U00010f00-\U00010f1c\U00010f27\U00010f30-\U00010f45\U00011003-\U00011037\U00011083-\U000110af\U000110d0-\U000110e8\U00011103-\U00011126\U00011144\U00011150-\U00011172\U00011176\U00011183-\U000111b2\U000111c1-\U000111c4\U000111da\U000111dc\U00011200-\U00011211\U00011213-\U0001122b\U00011280-\U00011286\U00011288\U0001128a-\U0001128d\U0001128f-\U0001129d\U0001129f-\U000112a8\U000112b0-\U000112de\U00011305-\U0001130c\U0001130f-\U00011310\U00011313-\U00011328\U0001132a-\U00011330\U00011332-\U00011333\U00011335-\U00011339\U0001133d\U00011350\U0001135d-\U00011361\U00011400-\U00011434\U00011447-\U0001144a\U00011480-\U000114af\U000114c4-\U000114c5\U000114c7\U00011580-\U000115ae\U000115d8-\U000115db\U00011600-\U0001162f\U00011644\U00011680-\U000116aa\U00011700-\U0001171a\U00011800-\U0001182b\U000118ff\U00011a00\U00011a0b-\U00011a32\U00011a3a\U00011a50\U00011a5c-\U00011a83\U00011a86-\U00011a89\U00011a9d\U00011ac0-\U00011af8\U00011c00-\U00011c08\U00011c0a-\U00011c2e\U00011c40\U00011c72-\U00011c8f\U00011d00-\U00011d06\U00011d08-\U00011d09\U00011d0b-\U00011d30\U00011d46\U00011d60-\U00011d65\U00011d67-\U00011d68\U00011d6a-\U00011d89\U00011d98\U00011ee0-\U00011ef2\U00012000-\U00012399\U00012480-\U00012543\U00013000-\U0001342e\U00014400-\U00014646\U00016800-\U00016a38\U00016a40-\U00016a5e\U00016ad0-\U00016aed\U00016b00-\U00016b2f\U00016b63-\U00016b77\U00016b7d-\U00016b8f\U00016f00-\U00016f44\U00016f50\U00017000-\U000187f1\U00018800-\U00018af2\U0001b000-\U0001b11e\U0001b170-\U0001b2fb\U0001bc00-\U0001bc6a\U0001bc70-\U0001bc7c\U0001bc80-\U0001bc88\U0001bc90-\U0001bc99\U0001e800-\U0001e8c4\U0001ee00-\U0001ee03\U0001ee05-\U0001ee1f\U0001ee21-\U0001ee22\U0001ee24\U0001ee27\U0001ee29-\U0001ee32\U0001ee34-\U0001ee37\U0001ee39\U0001ee3b\U0001ee42\U0001ee47\U0001ee49\U0001ee4b\U0001ee4d-\U0001ee4f\U0001ee51-\U0001ee52\U0001ee54\U0001ee57\U0001ee59\U0001ee5b\U0001ee5d\U0001ee5f\U0001ee61-\U0001ee62\U0001ee64\U0001ee67-\U0001ee6a\U0001ee6c-\U0001ee72\U0001ee74-\U0001ee77\U0001ee79-\U0001ee7c\U0001ee7e\U0001ee80-\U0001ee89\U0001ee8b-\U0001ee9b\U0001eea1-\U0001eea3\U0001eea5-\U0001eea9\U0001eeab-\U0001eebb\U00020000-\U0002a6d6\U0002a700-\U0002b734\U0002b740-\U0002b81d\U0002b820-\U0002cea1\U0002ceb0-\U0002ebe0\U0002f800-\U0002fa1d'
+
+Lt = '\u01c5\u01c8\u01cb\u01f2\u1f88-\u1f8f\u1f98-\u1f9f\u1fa8-\u1faf\u1fbc\u1fcc\u1ffc'
+
+Lu = 'A-Z\xc0-\xd6\xd8-\xde\u0100\u0102\u0104\u0106\u0108\u010a\u010c\u010e\u0110\u0112\u0114\u0116\u0118\u011a\u011c\u011e\u0120\u0122\u0124\u0126\u0128\u012a\u012c\u012e\u0130\u0132\u0134\u0136\u0139\u013b\u013d\u013f\u0141\u0143\u0145\u0147\u014a\u014c\u014e\u0150\u0152\u0154\u0156\u0158\u015a\u015c\u015e\u0160\u0162\u0164\u0166\u0168\u016a\u016c\u016e\u0170\u0172\u0174\u0176\u0178-\u0179\u017b\u017d\u0181-\u0182\u0184\u0186-\u0187\u0189-\u018b\u018e-\u0191\u0193-\u0194\u0196-\u0198\u019c-\u019d\u019f-\u01a0\u01a2\u01a4\u01a6-\u01a7\u01a9\u01ac\u01ae-\u01af\u01b1-\u01b3\u01b5\u01b7-\u01b8\u01bc\u01c4\u01c7\u01ca\u01cd\u01cf\u01d1\u01d3\u01d5\u01d7\u01d9\u01db\u01de\u01e0\u01e2\u01e4\u01e6\u01e8\u01ea\u01ec\u01ee\u01f1\u01f4\u01f6-\u01f8\u01fa\u01fc\u01fe\u0200\u0202\u0204\u0206\u0208\u020a\u020c\u020e\u0210\u0212\u0214\u0216\u0218\u021a\u021c\u021e\u0220\u0222\u0224\u0226\u0228\u022a\u022c\u022e\u0230\u0232\u023a-\u023b\u023d-\u023e\u0241\u0243-\u0246\u0248\u024a\u024c\u024e\u0370\u0372\u0376\u037f\u0386\u0388-\u038a\u038c\u038e-\u038f\u0391-\u03a1\u03a3-\u03ab\u03cf\u03d2-\u03d4\u03d8\u03da\u03dc\u03de\u03e0\u03e2\u03e4\u03e6\u03e8\u03ea\u03ec\u03ee\u03f4\u03f7\u03f9-\u03fa\u03fd-\u042f\u0460\u0462\u0464\u0466\u0468\u046a\u046c\u046e\u0470\u0472\u0474\u0476\u0478\u047a\u047c\u047e\u0480\u048a\u048c\u048e\u0490\u0492\u0494\u0496\u0498\u049a\u049c\u049e\u04a0\u04a2\u04a4\u04a6\u04a8\u04aa\u04ac\u04ae\u04b0\u04b2\u04b4\u04b6\u04b8\u04ba\u04bc\u04be\u04c0-\u04c1\u04c3\u04c5\u04c7\u04c9\u04cb\u04cd\u04d0\u04d2\u04d4\u04d6\u04d8\u04da\u04dc\u04de\u04e0\u04e2\u04e4\u04e6\u04e8\u04ea\u04ec\u04ee\u04f0\u04f2\u04f4\u04f6\u04f8\u04fa\u04fc\u04fe\u0500\u0502\u0504\u0506\u0508\u050a\u050c\u050e\u0510\u0512\u0514\u0516\u0518\u051a\u051c\u051e\u0520\u0522\u0524\u0526\u0528\u052a\u052c\u052e\u0531-\u0556\u10a0-\u10c5\u10c7\u10cd\u13a0-\u13f5\u1c90-\u1cba\u1cbd-\u1cbf\u1e00\u1e02\u1e04\u1e06\u1e08\u1e0a\u1e0c\u1e0e\u1e10\u1e12\u1e14\u1e16\u1e18\u1e1a\u1e1c\u1e1e\u1e20\u1e22\u1e24\u1e26\u1e28\u1e2a\u1e2c\u1e2e\u1e30\u1e32\u1e34\u1e36\u1e38\u1e3a\u1e3c\u1e3e\u1e40\u1e42\u1e44\u1e46\u1e48\u1e4a\u1e4c\u1e4e\u1e50\u1e52\u1e54\u1e56\u1e58\u1e5a\u1e5c\u1e5e\u1e60\u1e62\u1e64\u1e66\u1e68\u1e6a\u1e6c\u1e6e\u1e70\u1e72\u1e74\u1e76\u1e78\u1e7a\u1e7c\u1e7e\u1e80\u1e82\u1e84\u1e86\u1e88\u1e8a\u1e8c\u1e8e\u1e90\u1e92\u1e94\u1e9e\u1ea0\u1ea2\u1ea4\u1ea6\u1ea8\u1eaa\u1eac\u1eae\u1eb0\u1eb2\u1eb4\u1eb6\u1eb8\u1eba\u1ebc\u1ebe\u1ec0\u1ec2\u1ec4\u1ec6\u1ec8\u1eca\u1ecc\u1ece\u1ed0\u1ed2\u1ed4\u1ed6\u1ed8\u1eda\u1edc\u1ede\u1ee0\u1ee2\u1ee4\u1ee6\u1ee8\u1eea\u1eec\u1eee\u1ef0\u1ef2\u1ef4\u1ef6\u1ef8\u1efa\u1efc\u1efe\u1f08-\u1f0f\u1f18-\u1f1d\u1f28-\u1f2f\u1f38-\u1f3f\u1f48-\u1f4d\u1f59\u1f5b\u1f5d\u1f5f\u1f68-\u1f6f\u1fb8-\u1fbb\u1fc8-\u1fcb\u1fd8-\u1fdb\u1fe8-\u1fec\u1ff8-\u1ffb\u2102\u2107\u210b-\u210d\u2110-\u2112\u2115\u2119-\u211d\u2124\u2126\u2128\u212a-\u212d\u2130-\u2133\u213e-\u213f\u2145\u2183\u2c00-\u2c2e\u2c60\u2c62-\u2c64\u2c67\u2c69\u2c6b\u2c6d-\u2c70\u2c72\u2c75\u2c7e-\u2c80\u2c82\u2c84\u2c86\u2c88\u2c8a\u2c8c\u2c8e\u2c90\u2c92\u2c94\u2c96\u2c98\u2c9a\u2c9c\u2c9e\u2ca0\u2ca2\u2ca4\u2ca6\u2ca8\u2caa\u2cac\u2cae\u2cb0\u2cb2\u2cb4\u2cb6\u2cb8\u2cba\u2cbc\u2cbe\u2cc0\u2cc2\u2cc4\u2cc6\u2cc8\u2cca\u2ccc\u2cce\u2cd0\u2cd2\u2cd4\u2cd6\u2cd8\u2cda\u2cdc\u2cde\u2ce0\u2ce2\u2ceb\u2ced\u2cf2\ua640\ua642\ua644\ua646\ua648\ua64a\ua64c\ua64e\ua650\ua652\ua654\ua656\ua658\ua65a\ua65c\ua65e\ua660\ua662\ua664\ua666\ua668\ua66a\ua66c\ua680\ua682\ua684\ua686\ua688\ua68a\ua68c\ua68e\ua690\ua692\ua694\ua696\ua698\ua69a\ua722\ua724\ua726\ua728\ua72a\ua72c\ua72e\ua732\ua734\ua736\ua738\ua73a\ua73c\ua73e\ua740\ua742\ua744\ua746\ua748\ua74a\ua74c\ua74e\ua750\ua752\ua754\ua756\ua758\ua75a\ua75c\ua75e\ua760\ua762\ua764\ua766\ua768\ua76a\ua76c\ua76e\ua779\ua77b\ua77d-\ua77e\ua780\ua782\ua784\ua786\ua78b\ua78d\ua790\ua792\ua796\ua798\ua79a\ua79c\ua79e\ua7a0\ua7a2\ua7a4\ua7a6\ua7a8\ua7aa-\ua7ae\ua7b0-\ua7b4\ua7b6\ua7b8\uff21-\uff3a\U00010400-\U00010427\U000104b0-\U000104d3\U00010c80-\U00010cb2\U000118a0-\U000118bf\U00016e40-\U00016e5f\U0001d400-\U0001d419\U0001d434-\U0001d44d\U0001d468-\U0001d481\U0001d49c\U0001d49e-\U0001d49f\U0001d4a2\U0001d4a5-\U0001d4a6\U0001d4a9-\U0001d4ac\U0001d4ae-\U0001d4b5\U0001d4d0-\U0001d4e9\U0001d504-\U0001d505\U0001d507-\U0001d50a\U0001d50d-\U0001d514\U0001d516-\U0001d51c\U0001d538-\U0001d539\U0001d53b-\U0001d53e\U0001d540-\U0001d544\U0001d546\U0001d54a-\U0001d550\U0001d56c-\U0001d585\U0001d5a0-\U0001d5b9\U0001d5d4-\U0001d5ed\U0001d608-\U0001d621\U0001d63c-\U0001d655\U0001d670-\U0001d689\U0001d6a8-\U0001d6c0\U0001d6e2-\U0001d6fa\U0001d71c-\U0001d734\U0001d756-\U0001d76e\U0001d790-\U0001d7a8\U0001d7ca\U0001e900-\U0001e921'
+
+Mc = '\u0903\u093b\u093e-\u0940\u0949-\u094c\u094e-\u094f\u0982-\u0983\u09be-\u09c0\u09c7-\u09c8\u09cb-\u09cc\u09d7\u0a03\u0a3e-\u0a40\u0a83\u0abe-\u0ac0\u0ac9\u0acb-\u0acc\u0b02-\u0b03\u0b3e\u0b40\u0b47-\u0b48\u0b4b-\u0b4c\u0b57\u0bbe-\u0bbf\u0bc1-\u0bc2\u0bc6-\u0bc8\u0bca-\u0bcc\u0bd7\u0c01-\u0c03\u0c41-\u0c44\u0c82-\u0c83\u0cbe\u0cc0-\u0cc4\u0cc7-\u0cc8\u0cca-\u0ccb\u0cd5-\u0cd6\u0d02-\u0d03\u0d3e-\u0d40\u0d46-\u0d48\u0d4a-\u0d4c\u0d57\u0d82-\u0d83\u0dcf-\u0dd1\u0dd8-\u0ddf\u0df2-\u0df3\u0f3e-\u0f3f\u0f7f\u102b-\u102c\u1031\u1038\u103b-\u103c\u1056-\u1057\u1062-\u1064\u1067-\u106d\u1083-\u1084\u1087-\u108c\u108f\u109a-\u109c\u17b6\u17be-\u17c5\u17c7-\u17c8\u1923-\u1926\u1929-\u192b\u1930-\u1931\u1933-\u1938\u1a19-\u1a1a\u1a55\u1a57\u1a61\u1a63-\u1a64\u1a6d-\u1a72\u1b04\u1b35\u1b3b\u1b3d-\u1b41\u1b43-\u1b44\u1b82\u1ba1\u1ba6-\u1ba7\u1baa\u1be7\u1bea-\u1bec\u1bee\u1bf2-\u1bf3\u1c24-\u1c2b\u1c34-\u1c35\u1ce1\u1cf2-\u1cf3\u1cf7\u302e-\u302f\ua823-\ua824\ua827\ua880-\ua881\ua8b4-\ua8c3\ua952-\ua953\ua983\ua9b4-\ua9b5\ua9ba-\ua9bb\ua9bd-\ua9c0\uaa2f-\uaa30\uaa33-\uaa34\uaa4d\uaa7b\uaa7d\uaaeb\uaaee-\uaaef\uaaf5\uabe3-\uabe4\uabe6-\uabe7\uabe9-\uabea\uabec\U00011000\U00011002\U00011082\U000110b0-\U000110b2\U000110b7-\U000110b8\U0001112c\U00011145-\U00011146\U00011182\U000111b3-\U000111b5\U000111bf-\U000111c0\U0001122c-\U0001122e\U00011232-\U00011233\U00011235\U000112e0-\U000112e2\U00011302-\U00011303\U0001133e-\U0001133f\U00011341-\U00011344\U00011347-\U00011348\U0001134b-\U0001134d\U00011357\U00011362-\U00011363\U00011435-\U00011437\U00011440-\U00011441\U00011445\U000114b0-\U000114b2\U000114b9\U000114bb-\U000114be\U000114c1\U000115af-\U000115b1\U000115b8-\U000115bb\U000115be\U00011630-\U00011632\U0001163b-\U0001163c\U0001163e\U000116ac\U000116ae-\U000116af\U000116b6\U00011720-\U00011721\U00011726\U0001182c-\U0001182e\U00011838\U00011a39\U00011a57-\U00011a58\U00011a97\U00011c2f\U00011c3e\U00011ca9\U00011cb1\U00011cb4\U00011d8a-\U00011d8e\U00011d93-\U00011d94\U00011d96\U00011ef5-\U00011ef6\U00016f51-\U00016f7e\U0001d165-\U0001d166\U0001d16d-\U0001d172'
+
+Me = '\u0488-\u0489\u1abe\u20dd-\u20e0\u20e2-\u20e4\ua670-\ua672'
+
+Mn = '\u0300-\u036f\u0483-\u0487\u0591-\u05bd\u05bf\u05c1-\u05c2\u05c4-\u05c5\u05c7\u0610-\u061a\u064b-\u065f\u0670\u06d6-\u06dc\u06df-\u06e4\u06e7-\u06e8\u06ea-\u06ed\u0711\u0730-\u074a\u07a6-\u07b0\u07eb-\u07f3\u07fd\u0816-\u0819\u081b-\u0823\u0825-\u0827\u0829-\u082d\u0859-\u085b\u08d3-\u08e1\u08e3-\u0902\u093a\u093c\u0941-\u0948\u094d\u0951-\u0957\u0962-\u0963\u0981\u09bc\u09c1-\u09c4\u09cd\u09e2-\u09e3\u09fe\u0a01-\u0a02\u0a3c\u0a41-\u0a42\u0a47-\u0a48\u0a4b-\u0a4d\u0a51\u0a70-\u0a71\u0a75\u0a81-\u0a82\u0abc\u0ac1-\u0ac5\u0ac7-\u0ac8\u0acd\u0ae2-\u0ae3\u0afa-\u0aff\u0b01\u0b3c\u0b3f\u0b41-\u0b44\u0b4d\u0b56\u0b62-\u0b63\u0b82\u0bc0\u0bcd\u0c00\u0c04\u0c3e-\u0c40\u0c46-\u0c48\u0c4a-\u0c4d\u0c55-\u0c56\u0c62-\u0c63\u0c81\u0cbc\u0cbf\u0cc6\u0ccc-\u0ccd\u0ce2-\u0ce3\u0d00-\u0d01\u0d3b-\u0d3c\u0d41-\u0d44\u0d4d\u0d62-\u0d63\u0dca\u0dd2-\u0dd4\u0dd6\u0e31\u0e34-\u0e3a\u0e47-\u0e4e\u0eb1\u0eb4-\u0eb9\u0ebb-\u0ebc\u0ec8-\u0ecd\u0f18-\u0f19\u0f35\u0f37\u0f39\u0f71-\u0f7e\u0f80-\u0f84\u0f86-\u0f87\u0f8d-\u0f97\u0f99-\u0fbc\u0fc6\u102d-\u1030\u1032-\u1037\u1039-\u103a\u103d-\u103e\u1058-\u1059\u105e-\u1060\u1071-\u1074\u1082\u1085-\u1086\u108d\u109d\u135d-\u135f\u1712-\u1714\u1732-\u1734\u1752-\u1753\u1772-\u1773\u17b4-\u17b5\u17b7-\u17bd\u17c6\u17c9-\u17d3\u17dd\u180b-\u180d\u1885-\u1886\u18a9\u1920-\u1922\u1927-\u1928\u1932\u1939-\u193b\u1a17-\u1a18\u1a1b\u1a56\u1a58-\u1a5e\u1a60\u1a62\u1a65-\u1a6c\u1a73-\u1a7c\u1a7f\u1ab0-\u1abd\u1b00-\u1b03\u1b34\u1b36-\u1b3a\u1b3c\u1b42\u1b6b-\u1b73\u1b80-\u1b81\u1ba2-\u1ba5\u1ba8-\u1ba9\u1bab-\u1bad\u1be6\u1be8-\u1be9\u1bed\u1bef-\u1bf1\u1c2c-\u1c33\u1c36-\u1c37\u1cd0-\u1cd2\u1cd4-\u1ce0\u1ce2-\u1ce8\u1ced\u1cf4\u1cf8-\u1cf9\u1dc0-\u1df9\u1dfb-\u1dff\u20d0-\u20dc\u20e1\u20e5-\u20f0\u2cef-\u2cf1\u2d7f\u2de0-\u2dff\u302a-\u302d\u3099-\u309a\ua66f\ua674-\ua67d\ua69e-\ua69f\ua6f0-\ua6f1\ua802\ua806\ua80b\ua825-\ua826\ua8c4-\ua8c5\ua8e0-\ua8f1\ua8ff\ua926-\ua92d\ua947-\ua951\ua980-\ua982\ua9b3\ua9b6-\ua9b9\ua9bc\ua9e5\uaa29-\uaa2e\uaa31-\uaa32\uaa35-\uaa36\uaa43\uaa4c\uaa7c\uaab0\uaab2-\uaab4\uaab7-\uaab8\uaabe-\uaabf\uaac1\uaaec-\uaaed\uaaf6\uabe5\uabe8\uabed\ufb1e\ufe00-\ufe0f\ufe20-\ufe2f\U000101fd\U000102e0\U00010376-\U0001037a\U00010a01-\U00010a03\U00010a05-\U00010a06\U00010a0c-\U00010a0f\U00010a38-\U00010a3a\U00010a3f\U00010ae5-\U00010ae6\U00010d24-\U00010d27\U00010f46-\U00010f50\U00011001\U00011038-\U00011046\U0001107f-\U00011081\U000110b3-\U000110b6\U000110b9-\U000110ba\U00011100-\U00011102\U00011127-\U0001112b\U0001112d-\U00011134\U00011173\U00011180-\U00011181\U000111b6-\U000111be\U000111c9-\U000111cc\U0001122f-\U00011231\U00011234\U00011236-\U00011237\U0001123e\U000112df\U000112e3-\U000112ea\U00011300-\U00011301\U0001133b-\U0001133c\U00011340\U00011366-\U0001136c\U00011370-\U00011374\U00011438-\U0001143f\U00011442-\U00011444\U00011446\U0001145e\U000114b3-\U000114b8\U000114ba\U000114bf-\U000114c0\U000114c2-\U000114c3\U000115b2-\U000115b5\U000115bc-\U000115bd\U000115bf-\U000115c0\U000115dc-\U000115dd\U00011633-\U0001163a\U0001163d\U0001163f-\U00011640\U000116ab\U000116ad\U000116b0-\U000116b5\U000116b7\U0001171d-\U0001171f\U00011722-\U00011725\U00011727-\U0001172b\U0001182f-\U00011837\U00011839-\U0001183a\U00011a01-\U00011a0a\U00011a33-\U00011a38\U00011a3b-\U00011a3e\U00011a47\U00011a51-\U00011a56\U00011a59-\U00011a5b\U00011a8a-\U00011a96\U00011a98-\U00011a99\U00011c30-\U00011c36\U00011c38-\U00011c3d\U00011c3f\U00011c92-\U00011ca7\U00011caa-\U00011cb0\U00011cb2-\U00011cb3\U00011cb5-\U00011cb6\U00011d31-\U00011d36\U00011d3a\U00011d3c-\U00011d3d\U00011d3f-\U00011d45\U00011d47\U00011d90-\U00011d91\U00011d95\U00011d97\U00011ef3-\U00011ef4\U00016af0-\U00016af4\U00016b30-\U00016b36\U00016f8f-\U00016f92\U0001bc9d-\U0001bc9e\U0001d167-\U0001d169\U0001d17b-\U0001d182\U0001d185-\U0001d18b\U0001d1aa-\U0001d1ad\U0001d242-\U0001d244\U0001da00-\U0001da36\U0001da3b-\U0001da6c\U0001da75\U0001da84\U0001da9b-\U0001da9f\U0001daa1-\U0001daaf\U0001e000-\U0001e006\U0001e008-\U0001e018\U0001e01b-\U0001e021\U0001e023-\U0001e024\U0001e026-\U0001e02a\U0001e8d0-\U0001e8d6\U0001e944-\U0001e94a\U000e0100-\U000e01ef'
+
+Nd = '0-9\u0660-\u0669\u06f0-\u06f9\u07c0-\u07c9\u0966-\u096f\u09e6-\u09ef\u0a66-\u0a6f\u0ae6-\u0aef\u0b66-\u0b6f\u0be6-\u0bef\u0c66-\u0c6f\u0ce6-\u0cef\u0d66-\u0d6f\u0de6-\u0def\u0e50-\u0e59\u0ed0-\u0ed9\u0f20-\u0f29\u1040-\u1049\u1090-\u1099\u17e0-\u17e9\u1810-\u1819\u1946-\u194f\u19d0-\u19d9\u1a80-\u1a89\u1a90-\u1a99\u1b50-\u1b59\u1bb0-\u1bb9\u1c40-\u1c49\u1c50-\u1c59\ua620-\ua629\ua8d0-\ua8d9\ua900-\ua909\ua9d0-\ua9d9\ua9f0-\ua9f9\uaa50-\uaa59\uabf0-\uabf9\uff10-\uff19\U000104a0-\U000104a9\U00010d30-\U00010d39\U00011066-\U0001106f\U000110f0-\U000110f9\U00011136-\U0001113f\U000111d0-\U000111d9\U000112f0-\U000112f9\U00011450-\U00011459\U000114d0-\U000114d9\U00011650-\U00011659\U000116c0-\U000116c9\U00011730-\U00011739\U000118e0-\U000118e9\U00011c50-\U00011c59\U00011d50-\U00011d59\U00011da0-\U00011da9\U00016a60-\U00016a69\U00016b50-\U00016b59\U0001d7ce-\U0001d7ff\U0001e950-\U0001e959'
+
+Nl = '\u16ee-\u16f0\u2160-\u2182\u2185-\u2188\u3007\u3021-\u3029\u3038-\u303a\ua6e6-\ua6ef\U00010140-\U00010174\U00010341\U0001034a\U000103d1-\U000103d5\U00012400-\U0001246e'
+
+No = '\xb2-\xb3\xb9\xbc-\xbe\u09f4-\u09f9\u0b72-\u0b77\u0bf0-\u0bf2\u0c78-\u0c7e\u0d58-\u0d5e\u0d70-\u0d78\u0f2a-\u0f33\u1369-\u137c\u17f0-\u17f9\u19da\u2070\u2074-\u2079\u2080-\u2089\u2150-\u215f\u2189\u2460-\u249b\u24ea-\u24ff\u2776-\u2793\u2cfd\u3192-\u3195\u3220-\u3229\u3248-\u324f\u3251-\u325f\u3280-\u3289\u32b1-\u32bf\ua830-\ua835\U00010107-\U00010133\U00010175-\U00010178\U0001018a-\U0001018b\U000102e1-\U000102fb\U00010320-\U00010323\U00010858-\U0001085f\U00010879-\U0001087f\U000108a7-\U000108af\U000108fb-\U000108ff\U00010916-\U0001091b\U000109bc-\U000109bd\U000109c0-\U000109cf\U000109d2-\U000109ff\U00010a40-\U00010a48\U00010a7d-\U00010a7e\U00010a9d-\U00010a9f\U00010aeb-\U00010aef\U00010b58-\U00010b5f\U00010b78-\U00010b7f\U00010ba9-\U00010baf\U00010cfa-\U00010cff\U00010e60-\U00010e7e\U00010f1d-\U00010f26\U00010f51-\U00010f54\U00011052-\U00011065\U000111e1-\U000111f4\U0001173a-\U0001173b\U000118ea-\U000118f2\U00011c5a-\U00011c6c\U00016b5b-\U00016b61\U00016e80-\U00016e96\U0001d2e0-\U0001d2f3\U0001d360-\U0001d378\U0001e8c7-\U0001e8cf\U0001ec71-\U0001ecab\U0001ecad-\U0001ecaf\U0001ecb1-\U0001ecb4\U0001f100-\U0001f10c'
+
+Pc = '_\u203f-\u2040\u2054\ufe33-\ufe34\ufe4d-\ufe4f\uff3f'
+
+Pd = '\\-\u058a\u05be\u1400\u1806\u2010-\u2015\u2e17\u2e1a\u2e3a-\u2e3b\u2e40\u301c\u3030\u30a0\ufe31-\ufe32\ufe58\ufe63\uff0d'
+
+Pe = ')\\]}\u0f3b\u0f3d\u169c\u2046\u207e\u208e\u2309\u230b\u232a\u2769\u276b\u276d\u276f\u2771\u2773\u2775\u27c6\u27e7\u27e9\u27eb\u27ed\u27ef\u2984\u2986\u2988\u298a\u298c\u298e\u2990\u2992\u2994\u2996\u2998\u29d9\u29db\u29fd\u2e23\u2e25\u2e27\u2e29\u3009\u300b\u300d\u300f\u3011\u3015\u3017\u3019\u301b\u301e-\u301f\ufd3e\ufe18\ufe36\ufe38\ufe3a\ufe3c\ufe3e\ufe40\ufe42\ufe44\ufe48\ufe5a\ufe5c\ufe5e\uff09\uff3d\uff5d\uff60\uff63'
+
+Pf = '\xbb\u2019\u201d\u203a\u2e03\u2e05\u2e0a\u2e0d\u2e1d\u2e21'
+
+Pi = '\xab\u2018\u201b-\u201c\u201f\u2039\u2e02\u2e04\u2e09\u2e0c\u2e1c\u2e20'
+
+Po = "!-#%-'*,.-/:-;?-@\\\\\xa1\xa7\xb6-\xb7\xbf\u037e\u0387\u055a-\u055f\u0589\u05c0\u05c3\u05c6\u05f3-\u05f4\u0609-\u060a\u060c-\u060d\u061b\u061e-\u061f\u066a-\u066d\u06d4\u0700-\u070d\u07f7-\u07f9\u0830-\u083e\u085e\u0964-\u0965\u0970\u09fd\u0a76\u0af0\u0c84\u0df4\u0e4f\u0e5a-\u0e5b\u0f04-\u0f12\u0f14\u0f85\u0fd0-\u0fd4\u0fd9-\u0fda\u104a-\u104f\u10fb\u1360-\u1368\u166d-\u166e\u16eb-\u16ed\u1735-\u1736\u17d4-\u17d6\u17d8-\u17da\u1800-\u1805\u1807-\u180a\u1944-\u1945\u1a1e-\u1a1f\u1aa0-\u1aa6\u1aa8-\u1aad\u1b5a-\u1b60\u1bfc-\u1bff\u1c3b-\u1c3f\u1c7e-\u1c7f\u1cc0-\u1cc7\u1cd3\u2016-\u2017\u2020-\u2027\u2030-\u2038\u203b-\u203e\u2041-\u2043\u2047-\u2051\u2053\u2055-\u205e\u2cf9-\u2cfc\u2cfe-\u2cff\u2d70\u2e00-\u2e01\u2e06-\u2e08\u2e0b\u2e0e-\u2e16\u2e18-\u2e19\u2e1b\u2e1e-\u2e1f\u2e2a-\u2e2e\u2e30-\u2e39\u2e3c-\u2e3f\u2e41\u2e43-\u2e4e\u3001-\u3003\u303d\u30fb\ua4fe-\ua4ff\ua60d-\ua60f\ua673\ua67e\ua6f2-\ua6f7\ua874-\ua877\ua8ce-\ua8cf\ua8f8-\ua8fa\ua8fc\ua92e-\ua92f\ua95f\ua9c1-\ua9cd\ua9de-\ua9df\uaa5c-\uaa5f\uaade-\uaadf\uaaf0-\uaaf1\uabeb\ufe10-\ufe16\ufe19\ufe30\ufe45-\ufe46\ufe49-\ufe4c\ufe50-\ufe52\ufe54-\ufe57\ufe5f-\ufe61\ufe68\ufe6a-\ufe6b\uff01-\uff03\uff05-\uff07\uff0a\uff0c\uff0e-\uff0f\uff1a-\uff1b\uff1f-\uff20\uff3c\uff61\uff64-\uff65\U00010100-\U00010102\U0001039f\U000103d0\U0001056f\U00010857\U0001091f\U0001093f\U00010a50-\U00010a58\U00010a7f\U00010af0-\U00010af6\U00010b39-\U00010b3f\U00010b99-\U00010b9c\U00010f55-\U00010f59\U00011047-\U0001104d\U000110bb-\U000110bc\U000110be-\U000110c1\U00011140-\U00011143\U00011174-\U00011175\U000111c5-\U000111c8\U000111cd\U000111db\U000111dd-\U000111df\U00011238-\U0001123d\U000112a9\U0001144b-\U0001144f\U0001145b\U0001145d\U000114c6\U000115c1-\U000115d7\U00011641-\U00011643\U00011660-\U0001166c\U0001173c-\U0001173e\U0001183b\U00011a3f-\U00011a46\U00011a9a-\U00011a9c\U00011a9e-\U00011aa2\U00011c41-\U00011c45\U00011c70-\U00011c71\U00011ef7-\U00011ef8\U00012470-\U00012474\U00016a6e-\U00016a6f\U00016af5\U00016b37-\U00016b3b\U00016b44\U00016e97-\U00016e9a\U0001bc9f\U0001da87-\U0001da8b\U0001e95e-\U0001e95f"
+
+Ps = '(\\[{\u0f3a\u0f3c\u169b\u201a\u201e\u2045\u207d\u208d\u2308\u230a\u2329\u2768\u276a\u276c\u276e\u2770\u2772\u2774\u27c5\u27e6\u27e8\u27ea\u27ec\u27ee\u2983\u2985\u2987\u2989\u298b\u298d\u298f\u2991\u2993\u2995\u2997\u29d8\u29da\u29fc\u2e22\u2e24\u2e26\u2e28\u2e42\u3008\u300a\u300c\u300e\u3010\u3014\u3016\u3018\u301a\u301d\ufd3f\ufe17\ufe35\ufe37\ufe39\ufe3b\ufe3d\ufe3f\ufe41\ufe43\ufe47\ufe59\ufe5b\ufe5d\uff08\uff3b\uff5b\uff5f\uff62'
+
+Sc = '$\xa2-\xa5\u058f\u060b\u07fe-\u07ff\u09f2-\u09f3\u09fb\u0af1\u0bf9\u0e3f\u17db\u20a0-\u20bf\ua838\ufdfc\ufe69\uff04\uffe0-\uffe1\uffe5-\uffe6\U0001ecb0'
+
+Sk = '\\^`\xa8\xaf\xb4\xb8\u02c2-\u02c5\u02d2-\u02df\u02e5-\u02eb\u02ed\u02ef-\u02ff\u0375\u0384-\u0385\u1fbd\u1fbf-\u1fc1\u1fcd-\u1fcf\u1fdd-\u1fdf\u1fed-\u1fef\u1ffd-\u1ffe\u309b-\u309c\ua700-\ua716\ua720-\ua721\ua789-\ua78a\uab5b\ufbb2-\ufbc1\uff3e\uff40\uffe3\U0001f3fb-\U0001f3ff'
+
+Sm = '+<->|~\xac\xb1\xd7\xf7\u03f6\u0606-\u0608\u2044\u2052\u207a-\u207c\u208a-\u208c\u2118\u2140-\u2144\u214b\u2190-\u2194\u219a-\u219b\u21a0\u21a3\u21a6\u21ae\u21ce-\u21cf\u21d2\u21d4\u21f4-\u22ff\u2320-\u2321\u237c\u239b-\u23b3\u23dc-\u23e1\u25b7\u25c1\u25f8-\u25ff\u266f\u27c0-\u27c4\u27c7-\u27e5\u27f0-\u27ff\u2900-\u2982\u2999-\u29d7\u29dc-\u29fb\u29fe-\u2aff\u2b30-\u2b44\u2b47-\u2b4c\ufb29\ufe62\ufe64-\ufe66\uff0b\uff1c-\uff1e\uff5c\uff5e\uffe2\uffe9-\uffec\U0001d6c1\U0001d6db\U0001d6fb\U0001d715\U0001d735\U0001d74f\U0001d76f\U0001d789\U0001d7a9\U0001d7c3\U0001eef0-\U0001eef1'
+
+So = '\xa6\xa9\xae\xb0\u0482\u058d-\u058e\u060e-\u060f\u06de\u06e9\u06fd-\u06fe\u07f6\u09fa\u0b70\u0bf3-\u0bf8\u0bfa\u0c7f\u0d4f\u0d79\u0f01-\u0f03\u0f13\u0f15-\u0f17\u0f1a-\u0f1f\u0f34\u0f36\u0f38\u0fbe-\u0fc5\u0fc7-\u0fcc\u0fce-\u0fcf\u0fd5-\u0fd8\u109e-\u109f\u1390-\u1399\u1940\u19de-\u19ff\u1b61-\u1b6a\u1b74-\u1b7c\u2100-\u2101\u2103-\u2106\u2108-\u2109\u2114\u2116-\u2117\u211e-\u2123\u2125\u2127\u2129\u212e\u213a-\u213b\u214a\u214c-\u214d\u214f\u218a-\u218b\u2195-\u2199\u219c-\u219f\u21a1-\u21a2\u21a4-\u21a5\u21a7-\u21ad\u21af-\u21cd\u21d0-\u21d1\u21d3\u21d5-\u21f3\u2300-\u2307\u230c-\u231f\u2322-\u2328\u232b-\u237b\u237d-\u239a\u23b4-\u23db\u23e2-\u2426\u2440-\u244a\u249c-\u24e9\u2500-\u25b6\u25b8-\u25c0\u25c2-\u25f7\u2600-\u266e\u2670-\u2767\u2794-\u27bf\u2800-\u28ff\u2b00-\u2b2f\u2b45-\u2b46\u2b4d-\u2b73\u2b76-\u2b95\u2b98-\u2bc8\u2bca-\u2bfe\u2ce5-\u2cea\u2e80-\u2e99\u2e9b-\u2ef3\u2f00-\u2fd5\u2ff0-\u2ffb\u3004\u3012-\u3013\u3020\u3036-\u3037\u303e-\u303f\u3190-\u3191\u3196-\u319f\u31c0-\u31e3\u3200-\u321e\u322a-\u3247\u3250\u3260-\u327f\u328a-\u32b0\u32c0-\u32fe\u3300-\u33ff\u4dc0-\u4dff\ua490-\ua4c6\ua828-\ua82b\ua836-\ua837\ua839\uaa77-\uaa79\ufdfd\uffe4\uffe8\uffed-\uffee\ufffc-\ufffd\U00010137-\U0001013f\U00010179-\U00010189\U0001018c-\U0001018e\U00010190-\U0001019b\U000101a0\U000101d0-\U000101fc\U00010877-\U00010878\U00010ac8\U0001173f\U00016b3c-\U00016b3f\U00016b45\U0001bc9c\U0001d000-\U0001d0f5\U0001d100-\U0001d126\U0001d129-\U0001d164\U0001d16a-\U0001d16c\U0001d183-\U0001d184\U0001d18c-\U0001d1a9\U0001d1ae-\U0001d1e8\U0001d200-\U0001d241\U0001d245\U0001d300-\U0001d356\U0001d800-\U0001d9ff\U0001da37-\U0001da3a\U0001da6d-\U0001da74\U0001da76-\U0001da83\U0001da85-\U0001da86\U0001ecac\U0001f000-\U0001f02b\U0001f030-\U0001f093\U0001f0a0-\U0001f0ae\U0001f0b1-\U0001f0bf\U0001f0c1-\U0001f0cf\U0001f0d1-\U0001f0f5\U0001f110-\U0001f16b\U0001f170-\U0001f1ac\U0001f1e6-\U0001f202\U0001f210-\U0001f23b\U0001f240-\U0001f248\U0001f250-\U0001f251\U0001f260-\U0001f265\U0001f300-\U0001f3fa\U0001f400-\U0001f6d4\U0001f6e0-\U0001f6ec\U0001f6f0-\U0001f6f9\U0001f700-\U0001f773\U0001f780-\U0001f7d8\U0001f800-\U0001f80b\U0001f810-\U0001f847\U0001f850-\U0001f859\U0001f860-\U0001f887\U0001f890-\U0001f8ad\U0001f900-\U0001f90b\U0001f910-\U0001f93e\U0001f940-\U0001f970\U0001f973-\U0001f976\U0001f97a\U0001f97c-\U0001f9a2\U0001f9b0-\U0001f9b9\U0001f9c0-\U0001f9c2\U0001f9d0-\U0001f9ff\U0001fa60-\U0001fa6d'
+
+Zl = '\u2028'
+
+Zp = '\u2029'
+
+Zs = ' \xa0\u1680\u2000-\u200a\u202f\u205f\u3000'
+
+xid_continue = '0-9A-Z_a-z\xaa\xb5\xb7\xba\xc0-\xd6\xd8-\xf6\xf8-\u02c1\u02c6-\u02d1\u02e0-\u02e4\u02ec\u02ee\u0300-\u0374\u0376-\u0377\u037b-\u037d\u037f\u0386-\u038a\u038c\u038e-\u03a1\u03a3-\u03f5\u03f7-\u0481\u0483-\u0487\u048a-\u052f\u0531-\u0556\u0559\u0560-\u0588\u0591-\u05bd\u05bf\u05c1-\u05c2\u05c4-\u05c5\u05c7\u05d0-\u05ea\u05ef-\u05f2\u0610-\u061a\u0620-\u0669\u066e-\u06d3\u06d5-\u06dc\u06df-\u06e8\u06ea-\u06fc\u06ff\u0710-\u074a\u074d-\u07b1\u07c0-\u07f5\u07fa\u07fd\u0800-\u082d\u0840-\u085b\u0860-\u086a\u08a0-\u08b4\u08b6-\u08bd\u08d3-\u08e1\u08e3-\u0963\u0966-\u096f\u0971-\u0983\u0985-\u098c\u098f-\u0990\u0993-\u09a8\u09aa-\u09b0\u09b2\u09b6-\u09b9\u09bc-\u09c4\u09c7-\u09c8\u09cb-\u09ce\u09d7\u09dc-\u09dd\u09df-\u09e3\u09e6-\u09f1\u09fc\u09fe\u0a01-\u0a03\u0a05-\u0a0a\u0a0f-\u0a10\u0a13-\u0a28\u0a2a-\u0a30\u0a32-\u0a33\u0a35-\u0a36\u0a38-\u0a39\u0a3c\u0a3e-\u0a42\u0a47-\u0a48\u0a4b-\u0a4d\u0a51\u0a59-\u0a5c\u0a5e\u0a66-\u0a75\u0a81-\u0a83\u0a85-\u0a8d\u0a8f-\u0a91\u0a93-\u0aa8\u0aaa-\u0ab0\u0ab2-\u0ab3\u0ab5-\u0ab9\u0abc-\u0ac5\u0ac7-\u0ac9\u0acb-\u0acd\u0ad0\u0ae0-\u0ae3\u0ae6-\u0aef\u0af9-\u0aff\u0b01-\u0b03\u0b05-\u0b0c\u0b0f-\u0b10\u0b13-\u0b28\u0b2a-\u0b30\u0b32-\u0b33\u0b35-\u0b39\u0b3c-\u0b44\u0b47-\u0b48\u0b4b-\u0b4d\u0b56-\u0b57\u0b5c-\u0b5d\u0b5f-\u0b63\u0b66-\u0b6f\u0b71\u0b82-\u0b83\u0b85-\u0b8a\u0b8e-\u0b90\u0b92-\u0b95\u0b99-\u0b9a\u0b9c\u0b9e-\u0b9f\u0ba3-\u0ba4\u0ba8-\u0baa\u0bae-\u0bb9\u0bbe-\u0bc2\u0bc6-\u0bc8\u0bca-\u0bcd\u0bd0\u0bd7\u0be6-\u0bef\u0c00-\u0c0c\u0c0e-\u0c10\u0c12-\u0c28\u0c2a-\u0c39\u0c3d-\u0c44\u0c46-\u0c48\u0c4a-\u0c4d\u0c55-\u0c56\u0c58-\u0c5a\u0c60-\u0c63\u0c66-\u0c6f\u0c80-\u0c83\u0c85-\u0c8c\u0c8e-\u0c90\u0c92-\u0ca8\u0caa-\u0cb3\u0cb5-\u0cb9\u0cbc-\u0cc4\u0cc6-\u0cc8\u0cca-\u0ccd\u0cd5-\u0cd6\u0cde\u0ce0-\u0ce3\u0ce6-\u0cef\u0cf1-\u0cf2\u0d00-\u0d03\u0d05-\u0d0c\u0d0e-\u0d10\u0d12-\u0d44\u0d46-\u0d48\u0d4a-\u0d4e\u0d54-\u0d57\u0d5f-\u0d63\u0d66-\u0d6f\u0d7a-\u0d7f\u0d82-\u0d83\u0d85-\u0d96\u0d9a-\u0db1\u0db3-\u0dbb\u0dbd\u0dc0-\u0dc6\u0dca\u0dcf-\u0dd4\u0dd6\u0dd8-\u0ddf\u0de6-\u0def\u0df2-\u0df3\u0e01-\u0e3a\u0e40-\u0e4e\u0e50-\u0e59\u0e81-\u0e82\u0e84\u0e87-\u0e88\u0e8a\u0e8d\u0e94-\u0e97\u0e99-\u0e9f\u0ea1-\u0ea3\u0ea5\u0ea7\u0eaa-\u0eab\u0ead-\u0eb9\u0ebb-\u0ebd\u0ec0-\u0ec4\u0ec6\u0ec8-\u0ecd\u0ed0-\u0ed9\u0edc-\u0edf\u0f00\u0f18-\u0f19\u0f20-\u0f29\u0f35\u0f37\u0f39\u0f3e-\u0f47\u0f49-\u0f6c\u0f71-\u0f84\u0f86-\u0f97\u0f99-\u0fbc\u0fc6\u1000-\u1049\u1050-\u109d\u10a0-\u10c5\u10c7\u10cd\u10d0-\u10fa\u10fc-\u1248\u124a-\u124d\u1250-\u1256\u1258\u125a-\u125d\u1260-\u1288\u128a-\u128d\u1290-\u12b0\u12b2-\u12b5\u12b8-\u12be\u12c0\u12c2-\u12c5\u12c8-\u12d6\u12d8-\u1310\u1312-\u1315\u1318-\u135a\u135d-\u135f\u1369-\u1371\u1380-\u138f\u13a0-\u13f5\u13f8-\u13fd\u1401-\u166c\u166f-\u167f\u1681-\u169a\u16a0-\u16ea\u16ee-\u16f8\u1700-\u170c\u170e-\u1714\u1720-\u1734\u1740-\u1753\u1760-\u176c\u176e-\u1770\u1772-\u1773\u1780-\u17d3\u17d7\u17dc-\u17dd\u17e0-\u17e9\u180b-\u180d\u1810-\u1819\u1820-\u1878\u1880-\u18aa\u18b0-\u18f5\u1900-\u191e\u1920-\u192b\u1930-\u193b\u1946-\u196d\u1970-\u1974\u1980-\u19ab\u19b0-\u19c9\u19d0-\u19da\u1a00-\u1a1b\u1a20-\u1a5e\u1a60-\u1a7c\u1a7f-\u1a89\u1a90-\u1a99\u1aa7\u1ab0-\u1abd\u1b00-\u1b4b\u1b50-\u1b59\u1b6b-\u1b73\u1b80-\u1bf3\u1c00-\u1c37\u1c40-\u1c49\u1c4d-\u1c7d\u1c80-\u1c88\u1c90-\u1cba\u1cbd-\u1cbf\u1cd0-\u1cd2\u1cd4-\u1cf9\u1d00-\u1df9\u1dfb-\u1f15\u1f18-\u1f1d\u1f20-\u1f45\u1f48-\u1f4d\u1f50-\u1f57\u1f59\u1f5b\u1f5d\u1f5f-\u1f7d\u1f80-\u1fb4\u1fb6-\u1fbc\u1fbe\u1fc2-\u1fc4\u1fc6-\u1fcc\u1fd0-\u1fd3\u1fd6-\u1fdb\u1fe0-\u1fec\u1ff2-\u1ff4\u1ff6-\u1ffc\u203f-\u2040\u2054\u2071\u207f\u2090-\u209c\u20d0-\u20dc\u20e1\u20e5-\u20f0\u2102\u2107\u210a-\u2113\u2115\u2118-\u211d\u2124\u2126\u2128\u212a-\u2139\u213c-\u213f\u2145-\u2149\u214e\u2160-\u2188\u2c00-\u2c2e\u2c30-\u2c5e\u2c60-\u2ce4\u2ceb-\u2cf3\u2d00-\u2d25\u2d27\u2d2d\u2d30-\u2d67\u2d6f\u2d7f-\u2d96\u2da0-\u2da6\u2da8-\u2dae\u2db0-\u2db6\u2db8-\u2dbe\u2dc0-\u2dc6\u2dc8-\u2dce\u2dd0-\u2dd6\u2dd8-\u2dde\u2de0-\u2dff\u3005-\u3007\u3021-\u302f\u3031-\u3035\u3038-\u303c\u3041-\u3096\u3099-\u309a\u309d-\u309f\u30a1-\u30fa\u30fc-\u30ff\u3105-\u312f\u3131-\u318e\u31a0-\u31ba\u31f0-\u31ff\u3400-\u4db5\u4e00-\u9fef\ua000-\ua48c\ua4d0-\ua4fd\ua500-\ua60c\ua610-\ua62b\ua640-\ua66f\ua674-\ua67d\ua67f-\ua6f1\ua717-\ua71f\ua722-\ua788\ua78b-\ua7b9\ua7f7-\ua827\ua840-\ua873\ua880-\ua8c5\ua8d0-\ua8d9\ua8e0-\ua8f7\ua8fb\ua8fd-\ua92d\ua930-\ua953\ua960-\ua97c\ua980-\ua9c0\ua9cf-\ua9d9\ua9e0-\ua9fe\uaa00-\uaa36\uaa40-\uaa4d\uaa50-\uaa59\uaa60-\uaa76\uaa7a-\uaac2\uaadb-\uaadd\uaae0-\uaaef\uaaf2-\uaaf6\uab01-\uab06\uab09-\uab0e\uab11-\uab16\uab20-\uab26\uab28-\uab2e\uab30-\uab5a\uab5c-\uab65\uab70-\uabea\uabec-\uabed\uabf0-\uabf9\uac00-\ud7a3\ud7b0-\ud7c6\ud7cb-\ud7fb\uf900-\ufa6d\ufa70-\ufad9\ufb00-\ufb06\ufb13-\ufb17\ufb1d-\ufb28\ufb2a-\ufb36\ufb38-\ufb3c\ufb3e\ufb40-\ufb41\ufb43-\ufb44\ufb46-\ufbb1\ufbd3-\ufc5d\ufc64-\ufd3d\ufd50-\ufd8f\ufd92-\ufdc7\ufdf0-\ufdf9\ufe00-\ufe0f\ufe20-\ufe2f\ufe33-\ufe34\ufe4d-\ufe4f\ufe71\ufe73\ufe77\ufe79\ufe7b\ufe7d\ufe7f-\ufefc\uff10-\uff19\uff21-\uff3a\uff3f\uff41-\uff5a\uff66-\uffbe\uffc2-\uffc7\uffca-\uffcf\uffd2-\uffd7\uffda-\uffdc\U00010000-\U0001000b\U0001000d-\U00010026\U00010028-\U0001003a\U0001003c-\U0001003d\U0001003f-\U0001004d\U00010050-\U0001005d\U00010080-\U000100fa\U00010140-\U00010174\U000101fd\U00010280-\U0001029c\U000102a0-\U000102d0\U000102e0\U00010300-\U0001031f\U0001032d-\U0001034a\U00010350-\U0001037a\U00010380-\U0001039d\U000103a0-\U000103c3\U000103c8-\U000103cf\U000103d1-\U000103d5\U00010400-\U0001049d\U000104a0-\U000104a9\U000104b0-\U000104d3\U000104d8-\U000104fb\U00010500-\U00010527\U00010530-\U00010563\U00010600-\U00010736\U00010740-\U00010755\U00010760-\U00010767\U00010800-\U00010805\U00010808\U0001080a-\U00010835\U00010837-\U00010838\U0001083c\U0001083f-\U00010855\U00010860-\U00010876\U00010880-\U0001089e\U000108e0-\U000108f2\U000108f4-\U000108f5\U00010900-\U00010915\U00010920-\U00010939\U00010980-\U000109b7\U000109be-\U000109bf\U00010a00-\U00010a03\U00010a05-\U00010a06\U00010a0c-\U00010a13\U00010a15-\U00010a17\U00010a19-\U00010a35\U00010a38-\U00010a3a\U00010a3f\U00010a60-\U00010a7c\U00010a80-\U00010a9c\U00010ac0-\U00010ac7\U00010ac9-\U00010ae6\U00010b00-\U00010b35\U00010b40-\U00010b55\U00010b60-\U00010b72\U00010b80-\U00010b91\U00010c00-\U00010c48\U00010c80-\U00010cb2\U00010cc0-\U00010cf2\U00010d00-\U00010d27\U00010d30-\U00010d39\U00010f00-\U00010f1c\U00010f27\U00010f30-\U00010f50\U00011000-\U00011046\U00011066-\U0001106f\U0001107f-\U000110ba\U000110d0-\U000110e8\U000110f0-\U000110f9\U00011100-\U00011134\U00011136-\U0001113f\U00011144-\U00011146\U00011150-\U00011173\U00011176\U00011180-\U000111c4\U000111c9-\U000111cc\U000111d0-\U000111da\U000111dc\U00011200-\U00011211\U00011213-\U00011237\U0001123e\U00011280-\U00011286\U00011288\U0001128a-\U0001128d\U0001128f-\U0001129d\U0001129f-\U000112a8\U000112b0-\U000112ea\U000112f0-\U000112f9\U00011300-\U00011303\U00011305-\U0001130c\U0001130f-\U00011310\U00011313-\U00011328\U0001132a-\U00011330\U00011332-\U00011333\U00011335-\U00011339\U0001133b-\U00011344\U00011347-\U00011348\U0001134b-\U0001134d\U00011350\U00011357\U0001135d-\U00011363\U00011366-\U0001136c\U00011370-\U00011374\U00011400-\U0001144a\U00011450-\U00011459\U0001145e\U00011480-\U000114c5\U000114c7\U000114d0-\U000114d9\U00011580-\U000115b5\U000115b8-\U000115c0\U000115d8-\U000115dd\U00011600-\U00011640\U00011644\U00011650-\U00011659\U00011680-\U000116b7\U000116c0-\U000116c9\U00011700-\U0001171a\U0001171d-\U0001172b\U00011730-\U00011739\U00011800-\U0001183a\U000118a0-\U000118e9\U000118ff\U00011a00-\U00011a3e\U00011a47\U00011a50-\U00011a83\U00011a86-\U00011a99\U00011a9d\U00011ac0-\U00011af8\U00011c00-\U00011c08\U00011c0a-\U00011c36\U00011c38-\U00011c40\U00011c50-\U00011c59\U00011c72-\U00011c8f\U00011c92-\U00011ca7\U00011ca9-\U00011cb6\U00011d00-\U00011d06\U00011d08-\U00011d09\U00011d0b-\U00011d36\U00011d3a\U00011d3c-\U00011d3d\U00011d3f-\U00011d47\U00011d50-\U00011d59\U00011d60-\U00011d65\U00011d67-\U00011d68\U00011d6a-\U00011d8e\U00011d90-\U00011d91\U00011d93-\U00011d98\U00011da0-\U00011da9\U00011ee0-\U00011ef6\U00012000-\U00012399\U00012400-\U0001246e\U00012480-\U00012543\U00013000-\U0001342e\U00014400-\U00014646\U00016800-\U00016a38\U00016a40-\U00016a5e\U00016a60-\U00016a69\U00016ad0-\U00016aed\U00016af0-\U00016af4\U00016b00-\U00016b36\U00016b40-\U00016b43\U00016b50-\U00016b59\U00016b63-\U00016b77\U00016b7d-\U00016b8f\U00016e40-\U00016e7f\U00016f00-\U00016f44\U00016f50-\U00016f7e\U00016f8f-\U00016f9f\U00016fe0-\U00016fe1\U00017000-\U000187f1\U00018800-\U00018af2\U0001b000-\U0001b11e\U0001b170-\U0001b2fb\U0001bc00-\U0001bc6a\U0001bc70-\U0001bc7c\U0001bc80-\U0001bc88\U0001bc90-\U0001bc99\U0001bc9d-\U0001bc9e\U0001d165-\U0001d169\U0001d16d-\U0001d172\U0001d17b-\U0001d182\U0001d185-\U0001d18b\U0001d1aa-\U0001d1ad\U0001d242-\U0001d244\U0001d400-\U0001d454\U0001d456-\U0001d49c\U0001d49e-\U0001d49f\U0001d4a2\U0001d4a5-\U0001d4a6\U0001d4a9-\U0001d4ac\U0001d4ae-\U0001d4b9\U0001d4bb\U0001d4bd-\U0001d4c3\U0001d4c5-\U0001d505\U0001d507-\U0001d50a\U0001d50d-\U0001d514\U0001d516-\U0001d51c\U0001d51e-\U0001d539\U0001d53b-\U0001d53e\U0001d540-\U0001d544\U0001d546\U0001d54a-\U0001d550\U0001d552-\U0001d6a5\U0001d6a8-\U0001d6c0\U0001d6c2-\U0001d6da\U0001d6dc-\U0001d6fa\U0001d6fc-\U0001d714\U0001d716-\U0001d734\U0001d736-\U0001d74e\U0001d750-\U0001d76e\U0001d770-\U0001d788\U0001d78a-\U0001d7a8\U0001d7aa-\U0001d7c2\U0001d7c4-\U0001d7cb\U0001d7ce-\U0001d7ff\U0001da00-\U0001da36\U0001da3b-\U0001da6c\U0001da75\U0001da84\U0001da9b-\U0001da9f\U0001daa1-\U0001daaf\U0001e000-\U0001e006\U0001e008-\U0001e018\U0001e01b-\U0001e021\U0001e023-\U0001e024\U0001e026-\U0001e02a\U0001e800-\U0001e8c4\U0001e8d0-\U0001e8d6\U0001e900-\U0001e94a\U0001e950-\U0001e959\U0001ee00-\U0001ee03\U0001ee05-\U0001ee1f\U0001ee21-\U0001ee22\U0001ee24\U0001ee27\U0001ee29-\U0001ee32\U0001ee34-\U0001ee37\U0001ee39\U0001ee3b\U0001ee42\U0001ee47\U0001ee49\U0001ee4b\U0001ee4d-\U0001ee4f\U0001ee51-\U0001ee52\U0001ee54\U0001ee57\U0001ee59\U0001ee5b\U0001ee5d\U0001ee5f\U0001ee61-\U0001ee62\U0001ee64\U0001ee67-\U0001ee6a\U0001ee6c-\U0001ee72\U0001ee74-\U0001ee77\U0001ee79-\U0001ee7c\U0001ee7e\U0001ee80-\U0001ee89\U0001ee8b-\U0001ee9b\U0001eea1-\U0001eea3\U0001eea5-\U0001eea9\U0001eeab-\U0001eebb\U00020000-\U0002a6d6\U0002a700-\U0002b734\U0002b740-\U0002b81d\U0002b820-\U0002cea1\U0002ceb0-\U0002ebe0\U0002f800-\U0002fa1d\U000e0100-\U000e01ef'
+
+xid_start = 'A-Z_a-z\xaa\xb5\xba\xc0-\xd6\xd8-\xf6\xf8-\u02c1\u02c6-\u02d1\u02e0-\u02e4\u02ec\u02ee\u0370-\u0374\u0376-\u0377\u037b-\u037d\u037f\u0386\u0388-\u038a\u038c\u038e-\u03a1\u03a3-\u03f5\u03f7-\u0481\u048a-\u052f\u0531-\u0556\u0559\u0560-\u0588\u05d0-\u05ea\u05ef-\u05f2\u0620-\u064a\u066e-\u066f\u0671-\u06d3\u06d5\u06e5-\u06e6\u06ee-\u06ef\u06fa-\u06fc\u06ff\u0710\u0712-\u072f\u074d-\u07a5\u07b1\u07ca-\u07ea\u07f4-\u07f5\u07fa\u0800-\u0815\u081a\u0824\u0828\u0840-\u0858\u0860-\u086a\u08a0-\u08b4\u08b6-\u08bd\u0904-\u0939\u093d\u0950\u0958-\u0961\u0971-\u0980\u0985-\u098c\u098f-\u0990\u0993-\u09a8\u09aa-\u09b0\u09b2\u09b6-\u09b9\u09bd\u09ce\u09dc-\u09dd\u09df-\u09e1\u09f0-\u09f1\u09fc\u0a05-\u0a0a\u0a0f-\u0a10\u0a13-\u0a28\u0a2a-\u0a30\u0a32-\u0a33\u0a35-\u0a36\u0a38-\u0a39\u0a59-\u0a5c\u0a5e\u0a72-\u0a74\u0a85-\u0a8d\u0a8f-\u0a91\u0a93-\u0aa8\u0aaa-\u0ab0\u0ab2-\u0ab3\u0ab5-\u0ab9\u0abd\u0ad0\u0ae0-\u0ae1\u0af9\u0b05-\u0b0c\u0b0f-\u0b10\u0b13-\u0b28\u0b2a-\u0b30\u0b32-\u0b33\u0b35-\u0b39\u0b3d\u0b5c-\u0b5d\u0b5f-\u0b61\u0b71\u0b83\u0b85-\u0b8a\u0b8e-\u0b90\u0b92-\u0b95\u0b99-\u0b9a\u0b9c\u0b9e-\u0b9f\u0ba3-\u0ba4\u0ba8-\u0baa\u0bae-\u0bb9\u0bd0\u0c05-\u0c0c\u0c0e-\u0c10\u0c12-\u0c28\u0c2a-\u0c39\u0c3d\u0c58-\u0c5a\u0c60-\u0c61\u0c80\u0c85-\u0c8c\u0c8e-\u0c90\u0c92-\u0ca8\u0caa-\u0cb3\u0cb5-\u0cb9\u0cbd\u0cde\u0ce0-\u0ce1\u0cf1-\u0cf2\u0d05-\u0d0c\u0d0e-\u0d10\u0d12-\u0d3a\u0d3d\u0d4e\u0d54-\u0d56\u0d5f-\u0d61\u0d7a-\u0d7f\u0d85-\u0d96\u0d9a-\u0db1\u0db3-\u0dbb\u0dbd\u0dc0-\u0dc6\u0e01-\u0e30\u0e32\u0e40-\u0e46\u0e81-\u0e82\u0e84\u0e87-\u0e88\u0e8a\u0e8d\u0e94-\u0e97\u0e99-\u0e9f\u0ea1-\u0ea3\u0ea5\u0ea7\u0eaa-\u0eab\u0ead-\u0eb0\u0eb2\u0ebd\u0ec0-\u0ec4\u0ec6\u0edc-\u0edf\u0f00\u0f40-\u0f47\u0f49-\u0f6c\u0f88-\u0f8c\u1000-\u102a\u103f\u1050-\u1055\u105a-\u105d\u1061\u1065-\u1066\u106e-\u1070\u1075-\u1081\u108e\u10a0-\u10c5\u10c7\u10cd\u10d0-\u10fa\u10fc-\u1248\u124a-\u124d\u1250-\u1256\u1258\u125a-\u125d\u1260-\u1288\u128a-\u128d\u1290-\u12b0\u12b2-\u12b5\u12b8-\u12be\u12c0\u12c2-\u12c5\u12c8-\u12d6\u12d8-\u1310\u1312-\u1315\u1318-\u135a\u1380-\u138f\u13a0-\u13f5\u13f8-\u13fd\u1401-\u166c\u166f-\u167f\u1681-\u169a\u16a0-\u16ea\u16ee-\u16f8\u1700-\u170c\u170e-\u1711\u1720-\u1731\u1740-\u1751\u1760-\u176c\u176e-\u1770\u1780-\u17b3\u17d7\u17dc\u1820-\u1878\u1880-\u18a8\u18aa\u18b0-\u18f5\u1900-\u191e\u1950-\u196d\u1970-\u1974\u1980-\u19ab\u19b0-\u19c9\u1a00-\u1a16\u1a20-\u1a54\u1aa7\u1b05-\u1b33\u1b45-\u1b4b\u1b83-\u1ba0\u1bae-\u1baf\u1bba-\u1be5\u1c00-\u1c23\u1c4d-\u1c4f\u1c5a-\u1c7d\u1c80-\u1c88\u1c90-\u1cba\u1cbd-\u1cbf\u1ce9-\u1cec\u1cee-\u1cf1\u1cf5-\u1cf6\u1d00-\u1dbf\u1e00-\u1f15\u1f18-\u1f1d\u1f20-\u1f45\u1f48-\u1f4d\u1f50-\u1f57\u1f59\u1f5b\u1f5d\u1f5f-\u1f7d\u1f80-\u1fb4\u1fb6-\u1fbc\u1fbe\u1fc2-\u1fc4\u1fc6-\u1fcc\u1fd0-\u1fd3\u1fd6-\u1fdb\u1fe0-\u1fec\u1ff2-\u1ff4\u1ff6-\u1ffc\u2071\u207f\u2090-\u209c\u2102\u2107\u210a-\u2113\u2115\u2118-\u211d\u2124\u2126\u2128\u212a-\u2139\u213c-\u213f\u2145-\u2149\u214e\u2160-\u2188\u2c00-\u2c2e\u2c30-\u2c5e\u2c60-\u2ce4\u2ceb-\u2cee\u2cf2-\u2cf3\u2d00-\u2d25\u2d27\u2d2d\u2d30-\u2d67\u2d6f\u2d80-\u2d96\u2da0-\u2da6\u2da8-\u2dae\u2db0-\u2db6\u2db8-\u2dbe\u2dc0-\u2dc6\u2dc8-\u2dce\u2dd0-\u2dd6\u2dd8-\u2dde\u3005-\u3007\u3021-\u3029\u3031-\u3035\u3038-\u303c\u3041-\u3096\u309d-\u309f\u30a1-\u30fa\u30fc-\u30ff\u3105-\u312f\u3131-\u318e\u31a0-\u31ba\u31f0-\u31ff\u3400-\u4db5\u4e00-\u9fef\ua000-\ua48c\ua4d0-\ua4fd\ua500-\ua60c\ua610-\ua61f\ua62a-\ua62b\ua640-\ua66e\ua67f-\ua69d\ua6a0-\ua6ef\ua717-\ua71f\ua722-\ua788\ua78b-\ua7b9\ua7f7-\ua801\ua803-\ua805\ua807-\ua80a\ua80c-\ua822\ua840-\ua873\ua882-\ua8b3\ua8f2-\ua8f7\ua8fb\ua8fd-\ua8fe\ua90a-\ua925\ua930-\ua946\ua960-\ua97c\ua984-\ua9b2\ua9cf\ua9e0-\ua9e4\ua9e6-\ua9ef\ua9fa-\ua9fe\uaa00-\uaa28\uaa40-\uaa42\uaa44-\uaa4b\uaa60-\uaa76\uaa7a\uaa7e-\uaaaf\uaab1\uaab5-\uaab6\uaab9-\uaabd\uaac0\uaac2\uaadb-\uaadd\uaae0-\uaaea\uaaf2-\uaaf4\uab01-\uab06\uab09-\uab0e\uab11-\uab16\uab20-\uab26\uab28-\uab2e\uab30-\uab5a\uab5c-\uab65\uab70-\uabe2\uac00-\ud7a3\ud7b0-\ud7c6\ud7cb-\ud7fb\uf900-\ufa6d\ufa70-\ufad9\ufb00-\ufb06\ufb13-\ufb17\ufb1d\ufb1f-\ufb28\ufb2a-\ufb36\ufb38-\ufb3c\ufb3e\ufb40-\ufb41\ufb43-\ufb44\ufb46-\ufbb1\ufbd3-\ufc5d\ufc64-\ufd3d\ufd50-\ufd8f\ufd92-\ufdc7\ufdf0-\ufdf9\ufe71\ufe73\ufe77\ufe79\ufe7b\ufe7d\ufe7f-\ufefc\uff21-\uff3a\uff41-\uff5a\uff66-\uff9d\uffa0-\uffbe\uffc2-\uffc7\uffca-\uffcf\uffd2-\uffd7\uffda-\uffdc\U00010000-\U0001000b\U0001000d-\U00010026\U00010028-\U0001003a\U0001003c-\U0001003d\U0001003f-\U0001004d\U00010050-\U0001005d\U00010080-\U000100fa\U00010140-\U00010174\U00010280-\U0001029c\U000102a0-\U000102d0\U00010300-\U0001031f\U0001032d-\U0001034a\U00010350-\U00010375\U00010380-\U0001039d\U000103a0-\U000103c3\U000103c8-\U000103cf\U000103d1-\U000103d5\U00010400-\U0001049d\U000104b0-\U000104d3\U000104d8-\U000104fb\U00010500-\U00010527\U00010530-\U00010563\U00010600-\U00010736\U00010740-\U00010755\U00010760-\U00010767\U00010800-\U00010805\U00010808\U0001080a-\U00010835\U00010837-\U00010838\U0001083c\U0001083f-\U00010855\U00010860-\U00010876\U00010880-\U0001089e\U000108e0-\U000108f2\U000108f4-\U000108f5\U00010900-\U00010915\U00010920-\U00010939\U00010980-\U000109b7\U000109be-\U000109bf\U00010a00\U00010a10-\U00010a13\U00010a15-\U00010a17\U00010a19-\U00010a35\U00010a60-\U00010a7c\U00010a80-\U00010a9c\U00010ac0-\U00010ac7\U00010ac9-\U00010ae4\U00010b00-\U00010b35\U00010b40-\U00010b55\U00010b60-\U00010b72\U00010b80-\U00010b91\U00010c00-\U00010c48\U00010c80-\U00010cb2\U00010cc0-\U00010cf2\U00010d00-\U00010d23\U00010f00-\U00010f1c\U00010f27\U00010f30-\U00010f45\U00011003-\U00011037\U00011083-\U000110af\U000110d0-\U000110e8\U00011103-\U00011126\U00011144\U00011150-\U00011172\U00011176\U00011183-\U000111b2\U000111c1-\U000111c4\U000111da\U000111dc\U00011200-\U00011211\U00011213-\U0001122b\U00011280-\U00011286\U00011288\U0001128a-\U0001128d\U0001128f-\U0001129d\U0001129f-\U000112a8\U000112b0-\U000112de\U00011305-\U0001130c\U0001130f-\U00011310\U00011313-\U00011328\U0001132a-\U00011330\U00011332-\U00011333\U00011335-\U00011339\U0001133d\U00011350\U0001135d-\U00011361\U00011400-\U00011434\U00011447-\U0001144a\U00011480-\U000114af\U000114c4-\U000114c5\U000114c7\U00011580-\U000115ae\U000115d8-\U000115db\U00011600-\U0001162f\U00011644\U00011680-\U000116aa\U00011700-\U0001171a\U00011800-\U0001182b\U000118a0-\U000118df\U000118ff\U00011a00\U00011a0b-\U00011a32\U00011a3a\U00011a50\U00011a5c-\U00011a83\U00011a86-\U00011a89\U00011a9d\U00011ac0-\U00011af8\U00011c00-\U00011c08\U00011c0a-\U00011c2e\U00011c40\U00011c72-\U00011c8f\U00011d00-\U00011d06\U00011d08-\U00011d09\U00011d0b-\U00011d30\U00011d46\U00011d60-\U00011d65\U00011d67-\U00011d68\U00011d6a-\U00011d89\U00011d98\U00011ee0-\U00011ef2\U00012000-\U00012399\U00012400-\U0001246e\U00012480-\U00012543\U00013000-\U0001342e\U00014400-\U00014646\U00016800-\U00016a38\U00016a40-\U00016a5e\U00016ad0-\U00016aed\U00016b00-\U00016b2f\U00016b40-\U00016b43\U00016b63-\U00016b77\U00016b7d-\U00016b8f\U00016e40-\U00016e7f\U00016f00-\U00016f44\U00016f50\U00016f93-\U00016f9f\U00016fe0-\U00016fe1\U00017000-\U000187f1\U00018800-\U00018af2\U0001b000-\U0001b11e\U0001b170-\U0001b2fb\U0001bc00-\U0001bc6a\U0001bc70-\U0001bc7c\U0001bc80-\U0001bc88\U0001bc90-\U0001bc99\U0001d400-\U0001d454\U0001d456-\U0001d49c\U0001d49e-\U0001d49f\U0001d4a2\U0001d4a5-\U0001d4a6\U0001d4a9-\U0001d4ac\U0001d4ae-\U0001d4b9\U0001d4bb\U0001d4bd-\U0001d4c3\U0001d4c5-\U0001d505\U0001d507-\U0001d50a\U0001d50d-\U0001d514\U0001d516-\U0001d51c\U0001d51e-\U0001d539\U0001d53b-\U0001d53e\U0001d540-\U0001d544\U0001d546\U0001d54a-\U0001d550\U0001d552-\U0001d6a5\U0001d6a8-\U0001d6c0\U0001d6c2-\U0001d6da\U0001d6dc-\U0001d6fa\U0001d6fc-\U0001d714\U0001d716-\U0001d734\U0001d736-\U0001d74e\U0001d750-\U0001d76e\U0001d770-\U0001d788\U0001d78a-\U0001d7a8\U0001d7aa-\U0001d7c2\U0001d7c4-\U0001d7cb\U0001e800-\U0001e8c4\U0001e900-\U0001e943\U0001ee00-\U0001ee03\U0001ee05-\U0001ee1f\U0001ee21-\U0001ee22\U0001ee24\U0001ee27\U0001ee29-\U0001ee32\U0001ee34-\U0001ee37\U0001ee39\U0001ee3b\U0001ee42\U0001ee47\U0001ee49\U0001ee4b\U0001ee4d-\U0001ee4f\U0001ee51-\U0001ee52\U0001ee54\U0001ee57\U0001ee59\U0001ee5b\U0001ee5d\U0001ee5f\U0001ee61-\U0001ee62\U0001ee64\U0001ee67-\U0001ee6a\U0001ee6c-\U0001ee72\U0001ee74-\U0001ee77\U0001ee79-\U0001ee7c\U0001ee7e\U0001ee80-\U0001ee89\U0001ee8b-\U0001ee9b\U0001eea1-\U0001eea3\U0001eea5-\U0001eea9\U0001eeab-\U0001eebb\U00020000-\U0002a6d6\U0002a700-\U0002b734\U0002b740-\U0002b81d\U0002b820-\U0002cea1\U0002ceb0-\U0002ebe0\U0002f800-\U0002fa1d'
+
+cats = ['Cc', 'Cf', 'Cn', 'Co', 'Cs', 'Ll', 'Lm', 'Lo', 'Lt', 'Lu', 'Mc', 'Me', 'Mn', 'Nd', 'Nl', 'No', 'Pc', 'Pd', 'Pe', 'Pf', 'Pi', 'Po', 'Ps', 'Sc', 'Sk', 'Sm', 'So', 'Zl', 'Zp', 'Zs']
+
+# Generated from unidata 11.0.0
+
+def combine(*args):
+    return ''.join(globals()[cat] for cat in args)
+
+
+def allexcept(*args):
+    newcats = cats[:]
+    for arg in args:
+        newcats.remove(arg)
+    return ''.join(globals()[cat] for cat in newcats)
+
+
+def _handle_runs(char_list):  # pragma: no cover
+    buf = []
+    for c in char_list:
+        if len(c) == 1:
+            if buf and buf[-1][1] == chr(ord(c)-1):
+                buf[-1] = (buf[-1][0], c)
+            else:
+                buf.append((c, c))
+        else:
+            buf.append((c, c))
+    for a, b in buf:
+        if a == b:
+            yield a
+        else:
+            yield f'{a}-{b}'
+
+
+if __name__ == '__main__':  # pragma: no cover
+    import unicodedata
+
+    categories = {'xid_start': [], 'xid_continue': []}
+
+    with open(__file__, encoding='utf-8') as fp:
+        content = fp.read()
+
+    header = content[:content.find('Cc =')]
+    footer = content[content.find("def combine("):]
+
+    for code in range(0x110000):
+        c = chr(code)
+        cat = unicodedata.category(c)
+        if ord(c) == 0xdc00:
+            # Hack to avoid combining this combining with the preceding high
+            # surrogate, 0xdbff, when doing a repr.
+            c = '\\' + c
+        elif ord(c) in (0x2d, 0x5b, 0x5c, 0x5d, 0x5e):
+            # Escape regex metachars.
+            c = '\\' + c
+        categories.setdefault(cat, []).append(c)
+        # XID_START and XID_CONTINUE are special categories used for matching
+        # identifiers in Python 3.
+        if c.isidentifier():
+            categories['xid_start'].append(c)
+        if ('a' + c).isidentifier():
+            categories['xid_continue'].append(c)
+
+    with open(__file__, 'w', encoding='utf-8') as fp:
+        fp.write(header)
+
+        for cat in sorted(categories):
+            val = ''.join(_handle_runs(categories[cat]))
+            fp.write(f'{cat} = {val!a}\n\n')
+
+        cats = sorted(categories)
+        cats.remove('xid_start')
+        cats.remove('xid_continue')
+        fp.write(f'cats = {cats!r}\n\n')
+
+        fp.write(f'# Generated from unidata {unicodedata.unidata_version}\n\n')
+
+        fp.write(footer)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/util.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/util.py
new file mode 100644
index 00000000..83cf1049
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pygments/util.py
@@ -0,0 +1,324 @@
+"""
+    pygments.util
+    ~~~~~~~~~~~~~
+
+    Utility functions.
+
+    :copyright: Copyright 2006-2024 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+from io import TextIOWrapper
+
+
+split_path_re = re.compile(r'[/\\ ]')
+doctype_lookup_re = re.compile(r'''
+    <!DOCTYPE\s+(
+     [a-zA-Z_][a-zA-Z0-9]*
+     (?: \s+      # optional in HTML5
+     [a-zA-Z_][a-zA-Z0-9]*\s+
+     "[^"]*")?
+     )
+     [^>]*>
+''', re.DOTALL | re.MULTILINE | re.VERBOSE)
+tag_re = re.compile(r'<(.+?)(\s.*?)?>.*?</.+?>',
+                    re.IGNORECASE | re.DOTALL | re.MULTILINE)
+xml_decl_re = re.compile(r'\s*<\?xml[^>]*\?>', re.I)
+
+
+class ClassNotFound(ValueError):
+    """Raised if one of the lookup functions didn't find a matching class."""
+
+
+class OptionError(Exception):
+    """
+    This exception will be raised by all option processing functions if
+    the type or value of the argument is not correct.
+    """
+
+def get_choice_opt(options, optname, allowed, default=None, normcase=False):
+    """
+    If the key `optname` from the dictionary is not in the sequence
+    `allowed`, raise an error, otherwise return it.
+    """
+    string = options.get(optname, default)
+    if normcase:
+        string = string.lower()
+    if string not in allowed:
+        raise OptionError('Value for option {} must be one of {}'.format(optname, ', '.join(map(str, allowed))))
+    return string
+
+
+def get_bool_opt(options, optname, default=None):
+    """
+    Intuitively, this is `options.get(optname, default)`, but restricted to
+    Boolean value. The Booleans can be represented as string, in order to accept
+    Boolean value from the command line arguments. If the key `optname` is
+    present in the dictionary `options` and is not associated with a Boolean,
+    raise an `OptionError`. If it is absent, `default` is returned instead.
+
+    The valid string values for ``True`` are ``1``, ``yes``, ``true`` and
+    ``on``, the ones for ``False`` are ``0``, ``no``, ``false`` and ``off``
+    (matched case-insensitively).
+    """
+    string = options.get(optname, default)
+    if isinstance(string, bool):
+        return string
+    elif isinstance(string, int):
+        return bool(string)
+    elif not isinstance(string, str):
+        raise OptionError(f'Invalid type {string!r} for option {optname}; use '
+                          '1/0, yes/no, true/false, on/off')
+    elif string.lower() in ('1', 'yes', 'true', 'on'):
+        return True
+    elif string.lower() in ('0', 'no', 'false', 'off'):
+        return False
+    else:
+        raise OptionError(f'Invalid value {string!r} for option {optname}; use '
+                          '1/0, yes/no, true/false, on/off')
+
+
+def get_int_opt(options, optname, default=None):
+    """As :func:`get_bool_opt`, but interpret the value as an integer."""
+    string = options.get(optname, default)
+    try:
+        return int(string)
+    except TypeError:
+        raise OptionError(f'Invalid type {string!r} for option {optname}; you '
+                          'must give an integer value')
+    except ValueError:
+        raise OptionError(f'Invalid value {string!r} for option {optname}; you '
+                          'must give an integer value')
+
+def get_list_opt(options, optname, default=None):
+    """
+    If the key `optname` from the dictionary `options` is a string,
+    split it at whitespace and return it. If it is already a list
+    or a tuple, it is returned as a list.
+    """
+    val = options.get(optname, default)
+    if isinstance(val, str):
+        return val.split()
+    elif isinstance(val, (list, tuple)):
+        return list(val)
+    else:
+        raise OptionError(f'Invalid type {val!r} for option {optname}; you '
+                          'must give a list value')
+
+
+def docstring_headline(obj):
+    if not obj.__doc__:
+        return ''
+    res = []
+    for line in obj.__doc__.strip().splitlines():
+        if line.strip():
+            res.append(" " + line.strip())
+        else:
+            break
+    return ''.join(res).lstrip()
+
+
+def make_analysator(f):
+    """Return a static text analyser function that returns float values."""
+    def text_analyse(text):
+        try:
+            rv = f(text)
+        except Exception:
+            return 0.0
+        if not rv:
+            return 0.0
+        try:
+            return min(1.0, max(0.0, float(rv)))
+        except (ValueError, TypeError):
+            return 0.0
+    text_analyse.__doc__ = f.__doc__
+    return staticmethod(text_analyse)
+
+
+def shebang_matches(text, regex):
+    r"""Check if the given regular expression matches the last part of the
+    shebang if one exists.
+
+        >>> from pygments.util import shebang_matches
+        >>> shebang_matches('#!/usr/bin/env python', r'python(2\.\d)?')
+        True
+        >>> shebang_matches('#!/usr/bin/python2.4', r'python(2\.\d)?')
+        True
+        >>> shebang_matches('#!/usr/bin/python-ruby', r'python(2\.\d)?')
+        False
+        >>> shebang_matches('#!/usr/bin/python/ruby', r'python(2\.\d)?')
+        False
+        >>> shebang_matches('#!/usr/bin/startsomethingwith python',
+        ...                 r'python(2\.\d)?')
+        True
+
+    It also checks for common windows executable file extensions::
+
+        >>> shebang_matches('#!C:\\Python2.4\\Python.exe', r'python(2\.\d)?')
+        True
+
+    Parameters (``'-f'`` or ``'--foo'`` are ignored so ``'perl'`` does
+    the same as ``'perl -e'``)
+
+    Note that this method automatically searches the whole string (eg:
+    the regular expression is wrapped in ``'^$'``)
+    """
+    index = text.find('\n')
+    if index >= 0:
+        first_line = text[:index].lower()
+    else:
+        first_line = text.lower()
+    if first_line.startswith('#!'):
+        try:
+            found = [x for x in split_path_re.split(first_line[2:].strip())
+                     if x and not x.startswith('-')][-1]
+        except IndexError:
+            return False
+        regex = re.compile(rf'^{regex}(\.(exe|cmd|bat|bin))?$', re.IGNORECASE)
+        if regex.search(found) is not None:
+            return True
+    return False
+
+
+def doctype_matches(text, regex):
+    """Check if the doctype matches a regular expression (if present).
+
+    Note that this method only checks the first part of a DOCTYPE.
+    eg: 'html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"'
+    """
+    m = doctype_lookup_re.search(text)
+    if m is None:
+        return False
+    doctype = m.group(1)
+    return re.compile(regex, re.I).match(doctype.strip()) is not None
+
+
+def html_doctype_matches(text):
+    """Check if the file looks like it has a html doctype."""
+    return doctype_matches(text, r'html')
+
+
+_looks_like_xml_cache = {}
+
+
+def looks_like_xml(text):
+    """Check if a doctype exists or if we have some tags."""
+    if xml_decl_re.match(text):
+        return True
+    key = hash(text)
+    try:
+        return _looks_like_xml_cache[key]
+    except KeyError:
+        m = doctype_lookup_re.search(text)
+        if m is not None:
+            return True
+        rv = tag_re.search(text[:1000]) is not None
+        _looks_like_xml_cache[key] = rv
+        return rv
+
+
+def surrogatepair(c):
+    """Given a unicode character code with length greater than 16 bits,
+    return the two 16 bit surrogate pair.
+    """
+    # From example D28 of:
+    # http://www.unicode.org/book/ch03.pdf
+    return (0xd7c0 + (c >> 10), (0xdc00 + (c & 0x3ff)))
+
+
+def format_lines(var_name, seq, raw=False, indent_level=0):
+    """Formats a sequence of strings for output."""
+    lines = []
+    base_indent = ' ' * indent_level * 4
+    inner_indent = ' ' * (indent_level + 1) * 4
+    lines.append(base_indent + var_name + ' = (')
+    if raw:
+        # These should be preformatted reprs of, say, tuples.
+        for i in seq:
+            lines.append(inner_indent + i + ',')
+    else:
+        for i in seq:
+            # Force use of single quotes
+            r = repr(i + '"')
+            lines.append(inner_indent + r[:-2] + r[-1] + ',')
+    lines.append(base_indent + ')')
+    return '\n'.join(lines)
+
+
+def duplicates_removed(it, already_seen=()):
+    """
+    Returns a list with duplicates removed from the iterable `it`.
+
+    Order is preserved.
+    """
+    lst = []
+    seen = set()
+    for i in it:
+        if i in seen or i in already_seen:
+            continue
+        lst.append(i)
+        seen.add(i)
+    return lst
+
+
+class Future:
+    """Generic class to defer some work.
+
+    Handled specially in RegexLexerMeta, to support regex string construction at
+    first use.
+    """
+    def get(self):
+        raise NotImplementedError
+
+
+def guess_decode(text):
+    """Decode *text* with guessed encoding.
+
+    First try UTF-8; this should fail for non-UTF-8 encodings.
+    Then try the preferred locale encoding.
+    Fall back to latin-1, which always works.
+    """
+    try:
+        text = text.decode('utf-8')
+        return text, 'utf-8'
+    except UnicodeDecodeError:
+        try:
+            import locale
+            prefencoding = locale.getpreferredencoding()
+            text = text.decode()
+            return text, prefencoding
+        except (UnicodeDecodeError, LookupError):
+            text = text.decode('latin1')
+            return text, 'latin1'
+
+
+def guess_decode_from_terminal(text, term):
+    """Decode *text* coming from terminal *term*.
+
+    First try the terminal encoding, if given.
+    Then try UTF-8.  Then try the preferred locale encoding.
+    Fall back to latin-1, which always works.
+    """
+    if getattr(term, 'encoding', None):
+        try:
+            text = text.decode(term.encoding)
+        except UnicodeDecodeError:
+            pass
+        else:
+            return text, term.encoding
+    return guess_decode(text)
+
+
+def terminal_encoding(term):
+    """Return our best guess of encoding for the given *term*."""
+    if getattr(term, 'encoding', None):
+        return term.encoding
+    import locale
+    return locale.getpreferredencoding()
+
+
+class UnclosingTextIOWrapper(TextIOWrapper):
+    # Don't close underlying buffer on destruction.
+    def close(self):
+        self.flush()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__init__.py
new file mode 100644
index 00000000..746b89f7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__init__.py
@@ -0,0 +1,31 @@
+"""Wrappers to call pyproject.toml-based build backend hooks.
+"""
+
+from typing import TYPE_CHECKING
+
+from ._impl import (
+    BackendUnavailable,
+    BuildBackendHookCaller,
+    HookMissing,
+    UnsupportedOperation,
+    default_subprocess_runner,
+    quiet_subprocess_runner,
+)
+
+__version__ = "1.2.0"
+__all__ = [
+    "BackendUnavailable",
+    "BackendInvalid",
+    "HookMissing",
+    "UnsupportedOperation",
+    "default_subprocess_runner",
+    "quiet_subprocess_runner",
+    "BuildBackendHookCaller",
+]
+
+BackendInvalid = BackendUnavailable  # Deprecated alias, previously a separate exception
+
+if TYPE_CHECKING:
+    from ._impl import SubprocessRunner
+
+    __all__ += ["SubprocessRunner"]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..70855d6a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/_impl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/_impl.cpython-312.pyc
new file mode 100644
index 00000000..b6ee021a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/__pycache__/_impl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_impl.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_impl.py
new file mode 100644
index 00000000..d1e9d7bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_impl.py
@@ -0,0 +1,410 @@
+import json
+import os
+import sys
+import tempfile
+from contextlib import contextmanager
+from os.path import abspath
+from os.path import join as pjoin
+from subprocess import STDOUT, check_call, check_output
+from typing import TYPE_CHECKING, Any, Iterator, Mapping, Optional, Sequence
+
+from ._in_process import _in_proc_script_path
+
+if TYPE_CHECKING:
+    from typing import Protocol
+
+    class SubprocessRunner(Protocol):
+        """A protocol for the subprocess runner."""
+
+        def __call__(
+            self,
+            cmd: Sequence[str],
+            cwd: Optional[str] = None,
+            extra_environ: Optional[Mapping[str, str]] = None,
+        ) -> None:
+            ...
+
+
+def write_json(obj: Mapping[str, Any], path: str, **kwargs) -> None:
+    with open(path, "w", encoding="utf-8") as f:
+        json.dump(obj, f, **kwargs)
+
+
+def read_json(path: str) -> Mapping[str, Any]:
+    with open(path, encoding="utf-8") as f:
+        return json.load(f)
+
+
+class BackendUnavailable(Exception):
+    """Will be raised if the backend cannot be imported in the hook process."""
+
+    def __init__(
+        self,
+        traceback: str,
+        message: Optional[str] = None,
+        backend_name: Optional[str] = None,
+        backend_path: Optional[Sequence[str]] = None,
+    ) -> None:
+        # Preserving arg order for the sake of API backward compatibility.
+        self.backend_name = backend_name
+        self.backend_path = backend_path
+        self.traceback = traceback
+        super().__init__(message or "Error while importing backend")
+
+
+class HookMissing(Exception):
+    """Will be raised on missing hooks (if a fallback can't be used)."""
+
+    def __init__(self, hook_name: str) -> None:
+        super().__init__(hook_name)
+        self.hook_name = hook_name
+
+
+class UnsupportedOperation(Exception):
+    """May be raised by build_sdist if the backend indicates that it can't."""
+
+    def __init__(self, traceback: str) -> None:
+        self.traceback = traceback
+
+
+def default_subprocess_runner(
+    cmd: Sequence[str],
+    cwd: Optional[str] = None,
+    extra_environ: Optional[Mapping[str, str]] = None,
+) -> None:
+    """The default method of calling the wrapper subprocess.
+
+    This uses :func:`subprocess.check_call` under the hood.
+    """
+    env = os.environ.copy()
+    if extra_environ:
+        env.update(extra_environ)
+
+    check_call(cmd, cwd=cwd, env=env)
+
+
+def quiet_subprocess_runner(
+    cmd: Sequence[str],
+    cwd: Optional[str] = None,
+    extra_environ: Optional[Mapping[str, str]] = None,
+) -> None:
+    """Call the subprocess while suppressing output.
+
+    This uses :func:`subprocess.check_output` under the hood.
+    """
+    env = os.environ.copy()
+    if extra_environ:
+        env.update(extra_environ)
+
+    check_output(cmd, cwd=cwd, env=env, stderr=STDOUT)
+
+
+def norm_and_check(source_tree: str, requested: str) -> str:
+    """Normalise and check a backend path.
+
+    Ensure that the requested backend path is specified as a relative path,
+    and resolves to a location under the given source tree.
+
+    Return an absolute version of the requested path.
+    """
+    if os.path.isabs(requested):
+        raise ValueError("paths must be relative")
+
+    abs_source = os.path.abspath(source_tree)
+    abs_requested = os.path.normpath(os.path.join(abs_source, requested))
+    # We have to use commonprefix for Python 2.7 compatibility. So we
+    # normalise case to avoid problems because commonprefix is a character
+    # based comparison :-(
+    norm_source = os.path.normcase(abs_source)
+    norm_requested = os.path.normcase(abs_requested)
+    if os.path.commonprefix([norm_source, norm_requested]) != norm_source:
+        raise ValueError("paths must be inside source tree")
+
+    return abs_requested
+
+
+class BuildBackendHookCaller:
+    """A wrapper to call the build backend hooks for a source directory."""
+
+    def __init__(
+        self,
+        source_dir: str,
+        build_backend: str,
+        backend_path: Optional[Sequence[str]] = None,
+        runner: Optional["SubprocessRunner"] = None,
+        python_executable: Optional[str] = None,
+    ) -> None:
+        """
+        :param source_dir: The source directory to invoke the build backend for
+        :param build_backend: The build backend spec
+        :param backend_path: Additional path entries for the build backend spec
+        :param runner: The :ref:`subprocess runner <Subprocess Runners>` to use
+        :param python_executable:
+            The Python executable used to invoke the build backend
+        """
+        if runner is None:
+            runner = default_subprocess_runner
+
+        self.source_dir = abspath(source_dir)
+        self.build_backend = build_backend
+        if backend_path:
+            backend_path = [norm_and_check(self.source_dir, p) for p in backend_path]
+        self.backend_path = backend_path
+        self._subprocess_runner = runner
+        if not python_executable:
+            python_executable = sys.executable
+        self.python_executable = python_executable
+
+    @contextmanager
+    def subprocess_runner(self, runner: "SubprocessRunner") -> Iterator[None]:
+        """A context manager for temporarily overriding the default
+        :ref:`subprocess runner <Subprocess Runners>`.
+
+        :param runner: The new subprocess runner to use within the context.
+
+        .. code-block:: python
+
+            hook_caller = BuildBackendHookCaller(...)
+            with hook_caller.subprocess_runner(quiet_subprocess_runner):
+                ...
+        """
+        prev = self._subprocess_runner
+        self._subprocess_runner = runner
+        try:
+            yield
+        finally:
+            self._subprocess_runner = prev
+
+    def _supported_features(self) -> Sequence[str]:
+        """Return the list of optional features supported by the backend."""
+        return self._call_hook("_supported_features", {})
+
+    def get_requires_for_build_wheel(
+        self,
+        config_settings: Optional[Mapping[str, Any]] = None,
+    ) -> Sequence[str]:
+        """Get additional dependencies required for building a wheel.
+
+        :param config_settings: The configuration settings for the build backend
+        :returns: A list of :pep:`dependency specifiers <508>`.
+
+        .. admonition:: Fallback
+
+            If the build backend does not defined a hook with this name, an
+            empty list will be returned.
+        """
+        return self._call_hook(
+            "get_requires_for_build_wheel", {"config_settings": config_settings}
+        )
+
+    def prepare_metadata_for_build_wheel(
+        self,
+        metadata_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        _allow_fallback: bool = True,
+    ) -> str:
+        """Prepare a ``*.dist-info`` folder with metadata for this project.
+
+        :param metadata_directory: The directory to write the metadata to
+        :param config_settings: The configuration settings for the build backend
+        :param _allow_fallback:
+            Whether to allow the fallback to building a wheel and extracting
+            the metadata from it. Should be passed as a keyword argument only.
+
+        :returns: Name of the newly created subfolder within
+                  ``metadata_directory``, containing the metadata.
+
+        .. admonition:: Fallback
+
+            If the build backend does not define a hook with this name and
+            ``_allow_fallback`` is truthy, the backend will be asked to build a
+            wheel via the ``build_wheel`` hook and the dist-info extracted from
+            that will be returned.
+        """
+        return self._call_hook(
+            "prepare_metadata_for_build_wheel",
+            {
+                "metadata_directory": abspath(metadata_directory),
+                "config_settings": config_settings,
+                "_allow_fallback": _allow_fallback,
+            },
+        )
+
+    def build_wheel(
+        self,
+        wheel_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        metadata_directory: Optional[str] = None,
+    ) -> str:
+        """Build a wheel from this project.
+
+        :param wheel_directory: The directory to write the wheel to
+        :param config_settings: The configuration settings for the build backend
+        :param metadata_directory: The directory to reuse existing metadata from
+        :returns:
+            The name of the newly created wheel within ``wheel_directory``.
+
+        .. admonition:: Interaction with fallback
+
+            If the ``build_wheel`` hook was called in the fallback for
+            :meth:`prepare_metadata_for_build_wheel`, the build backend would
+            not be invoked. Instead, the previously built wheel will be copied
+            to ``wheel_directory`` and the name of that file will be returned.
+        """
+        if metadata_directory is not None:
+            metadata_directory = abspath(metadata_directory)
+        return self._call_hook(
+            "build_wheel",
+            {
+                "wheel_directory": abspath(wheel_directory),
+                "config_settings": config_settings,
+                "metadata_directory": metadata_directory,
+            },
+        )
+
+    def get_requires_for_build_editable(
+        self,
+        config_settings: Optional[Mapping[str, Any]] = None,
+    ) -> Sequence[str]:
+        """Get additional dependencies required for building an editable wheel.
+
+        :param config_settings: The configuration settings for the build backend
+        :returns: A list of :pep:`dependency specifiers <508>`.
+
+        .. admonition:: Fallback
+
+            If the build backend does not defined a hook with this name, an
+            empty list will be returned.
+        """
+        return self._call_hook(
+            "get_requires_for_build_editable", {"config_settings": config_settings}
+        )
+
+    def prepare_metadata_for_build_editable(
+        self,
+        metadata_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        _allow_fallback: bool = True,
+    ) -> Optional[str]:
+        """Prepare a ``*.dist-info`` folder with metadata for this project.
+
+        :param metadata_directory: The directory to write the metadata to
+        :param config_settings: The configuration settings for the build backend
+        :param _allow_fallback:
+            Whether to allow the fallback to building a wheel and extracting
+            the metadata from it. Should be passed as a keyword argument only.
+        :returns: Name of the newly created subfolder within
+                  ``metadata_directory``, containing the metadata.
+
+        .. admonition:: Fallback
+
+            If the build backend does not define a hook with this name and
+            ``_allow_fallback`` is truthy, the backend will be asked to build a
+            wheel via the ``build_editable`` hook and the dist-info
+            extracted from that will be returned.
+        """
+        return self._call_hook(
+            "prepare_metadata_for_build_editable",
+            {
+                "metadata_directory": abspath(metadata_directory),
+                "config_settings": config_settings,
+                "_allow_fallback": _allow_fallback,
+            },
+        )
+
+    def build_editable(
+        self,
+        wheel_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+        metadata_directory: Optional[str] = None,
+    ) -> str:
+        """Build an editable wheel from this project.
+
+        :param wheel_directory: The directory to write the wheel to
+        :param config_settings: The configuration settings for the build backend
+        :param metadata_directory: The directory to reuse existing metadata from
+        :returns:
+            The name of the newly created wheel within ``wheel_directory``.
+
+        .. admonition:: Interaction with fallback
+
+            If the ``build_editable`` hook was called in the fallback for
+            :meth:`prepare_metadata_for_build_editable`, the build backend
+            would not be invoked. Instead, the previously built wheel will be
+            copied to ``wheel_directory`` and the name of that file will be
+            returned.
+        """
+        if metadata_directory is not None:
+            metadata_directory = abspath(metadata_directory)
+        return self._call_hook(
+            "build_editable",
+            {
+                "wheel_directory": abspath(wheel_directory),
+                "config_settings": config_settings,
+                "metadata_directory": metadata_directory,
+            },
+        )
+
+    def get_requires_for_build_sdist(
+        self,
+        config_settings: Optional[Mapping[str, Any]] = None,
+    ) -> Sequence[str]:
+        """Get additional dependencies required for building an sdist.
+
+        :returns: A list of :pep:`dependency specifiers <508>`.
+        """
+        return self._call_hook(
+            "get_requires_for_build_sdist", {"config_settings": config_settings}
+        )
+
+    def build_sdist(
+        self,
+        sdist_directory: str,
+        config_settings: Optional[Mapping[str, Any]] = None,
+    ) -> str:
+        """Build an sdist from this project.
+
+        :returns:
+            The name of the newly created sdist within ``wheel_directory``.
+        """
+        return self._call_hook(
+            "build_sdist",
+            {
+                "sdist_directory": abspath(sdist_directory),
+                "config_settings": config_settings,
+            },
+        )
+
+    def _call_hook(self, hook_name: str, kwargs: Mapping[str, Any]) -> Any:
+        extra_environ = {"_PYPROJECT_HOOKS_BUILD_BACKEND": self.build_backend}
+
+        if self.backend_path:
+            backend_path = os.pathsep.join(self.backend_path)
+            extra_environ["_PYPROJECT_HOOKS_BACKEND_PATH"] = backend_path
+
+        with tempfile.TemporaryDirectory() as td:
+            hook_input = {"kwargs": kwargs}
+            write_json(hook_input, pjoin(td, "input.json"), indent=2)
+
+            # Run the hook in a subprocess
+            with _in_proc_script_path() as script:
+                python = self.python_executable
+                self._subprocess_runner(
+                    [python, abspath(str(script)), hook_name, td],
+                    cwd=self.source_dir,
+                    extra_environ=extra_environ,
+                )
+
+            data = read_json(pjoin(td, "output.json"))
+            if data.get("unsupported"):
+                raise UnsupportedOperation(data.get("traceback", ""))
+            if data.get("no_backend"):
+                raise BackendUnavailable(
+                    data.get("traceback", ""),
+                    message=data.get("backend_error", ""),
+                    backend_name=self.build_backend,
+                    backend_path=self.backend_path,
+                )
+            if data.get("hook_missing"):
+                raise HookMissing(data.get("missing_hook_name") or hook_name)
+            return data["return_val"]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__init__.py
new file mode 100644
index 00000000..906d0ba2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__init__.py
@@ -0,0 +1,21 @@
+"""This is a subpackage because the directory is on sys.path for _in_process.py
+
+The subpackage should stay as empty as possible to avoid shadowing modules that
+the backend might import.
+"""
+
+import importlib.resources as resources
+
+try:
+    resources.files
+except AttributeError:
+    # Python 3.8 compatibility
+    def _in_proc_script_path():
+        return resources.path(__package__, "_in_process.py")
+
+else:
+
+    def _in_proc_script_path():
+        return resources.as_file(
+            resources.files(__package__).joinpath("_in_process.py")
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..8c8f170c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/_in_process.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/_in_process.cpython-312.pyc
new file mode 100644
index 00000000..102407da
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/__pycache__/_in_process.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py
new file mode 100644
index 00000000..d689bab7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py
@@ -0,0 +1,389 @@
+"""This is invoked in a subprocess to call the build backend hooks.
+
+It expects:
+- Command line args: hook_name, control_dir
+- Environment variables:
+      _PYPROJECT_HOOKS_BUILD_BACKEND=entry.point:spec
+      _PYPROJECT_HOOKS_BACKEND_PATH=paths (separated with os.pathsep)
+- control_dir/input.json:
+  - {"kwargs": {...}}
+
+Results:
+- control_dir/output.json
+  - {"return_val": ...}
+"""
+import json
+import os
+import os.path
+import re
+import shutil
+import sys
+import traceback
+from glob import glob
+from importlib import import_module
+from importlib.machinery import PathFinder
+from os.path import join as pjoin
+
+# This file is run as a script, and `import wrappers` is not zip-safe, so we
+# include write_json() and read_json() from wrappers.py.
+
+
+def write_json(obj, path, **kwargs):
+    with open(path, "w", encoding="utf-8") as f:
+        json.dump(obj, f, **kwargs)
+
+
+def read_json(path):
+    with open(path, encoding="utf-8") as f:
+        return json.load(f)
+
+
+class BackendUnavailable(Exception):
+    """Raised if we cannot import the backend"""
+
+    def __init__(self, message, traceback=None):
+        super().__init__(message)
+        self.message = message
+        self.traceback = traceback
+
+
+class HookMissing(Exception):
+    """Raised if a hook is missing and we are not executing the fallback"""
+
+    def __init__(self, hook_name=None):
+        super().__init__(hook_name)
+        self.hook_name = hook_name
+
+
+def _build_backend():
+    """Find and load the build backend"""
+    backend_path = os.environ.get("_PYPROJECT_HOOKS_BACKEND_PATH")
+    ep = os.environ["_PYPROJECT_HOOKS_BUILD_BACKEND"]
+    mod_path, _, obj_path = ep.partition(":")
+
+    if backend_path:
+        # Ensure in-tree backend directories have the highest priority when importing.
+        extra_pathitems = backend_path.split(os.pathsep)
+        sys.meta_path.insert(0, _BackendPathFinder(extra_pathitems, mod_path))
+
+    try:
+        obj = import_module(mod_path)
+    except ImportError:
+        msg = f"Cannot import {mod_path!r}"
+        raise BackendUnavailable(msg, traceback.format_exc())
+
+    if obj_path:
+        for path_part in obj_path.split("."):
+            obj = getattr(obj, path_part)
+    return obj
+
+
+class _BackendPathFinder:
+    """Implements the MetaPathFinder interface to locate modules in ``backend-path``.
+
+    Since the environment provided by the frontend can contain all sorts of
+    MetaPathFinders, the only way to ensure the backend is loaded from the
+    right place is to prepend our own.
+    """
+
+    def __init__(self, backend_path, backend_module):
+        self.backend_path = backend_path
+        self.backend_module = backend_module
+        self.backend_parent, _, _ = backend_module.partition(".")
+
+    def find_spec(self, fullname, _path, _target=None):
+        if "." in fullname:
+            # Rely on importlib to find nested modules based on parent's path
+            return None
+
+        # Ignore other items in _path or sys.path and use backend_path instead:
+        spec = PathFinder.find_spec(fullname, path=self.backend_path)
+        if spec is None and fullname == self.backend_parent:
+            # According to the spec, the backend MUST be loaded from backend-path.
+            # Therefore, we can halt the import machinery and raise a clean error.
+            msg = f"Cannot find module {self.backend_module!r} in {self.backend_path!r}"
+            raise BackendUnavailable(msg)
+
+        return spec
+
+    if sys.version_info >= (3, 8):
+
+        def find_distributions(self, context=None):
+            # Delayed import: Python 3.7 does not contain importlib.metadata
+            from importlib.metadata import DistributionFinder, MetadataPathFinder
+
+            context = DistributionFinder.Context(path=self.backend_path)
+            return MetadataPathFinder.find_distributions(context=context)
+
+
+def _supported_features():
+    """Return the list of options features supported by the backend.
+
+    Returns a list of strings.
+    The only possible value is 'build_editable'.
+    """
+    backend = _build_backend()
+    features = []
+    if hasattr(backend, "build_editable"):
+        features.append("build_editable")
+    return features
+
+
+def get_requires_for_build_wheel(config_settings):
+    """Invoke the optional get_requires_for_build_wheel hook
+
+    Returns [] if the hook is not defined.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.get_requires_for_build_wheel
+    except AttributeError:
+        return []
+    else:
+        return hook(config_settings)
+
+
+def get_requires_for_build_editable(config_settings):
+    """Invoke the optional get_requires_for_build_editable hook
+
+    Returns [] if the hook is not defined.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.get_requires_for_build_editable
+    except AttributeError:
+        return []
+    else:
+        return hook(config_settings)
+
+
+def prepare_metadata_for_build_wheel(
+    metadata_directory, config_settings, _allow_fallback
+):
+    """Invoke optional prepare_metadata_for_build_wheel
+
+    Implements a fallback by building a wheel if the hook isn't defined,
+    unless _allow_fallback is False in which case HookMissing is raised.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.prepare_metadata_for_build_wheel
+    except AttributeError:
+        if not _allow_fallback:
+            raise HookMissing()
+    else:
+        return hook(metadata_directory, config_settings)
+    # fallback to build_wheel outside the try block to avoid exception chaining
+    # which can be confusing to users and is not relevant
+    whl_basename = backend.build_wheel(metadata_directory, config_settings)
+    return _get_wheel_metadata_from_wheel(
+        whl_basename, metadata_directory, config_settings
+    )
+
+
+def prepare_metadata_for_build_editable(
+    metadata_directory, config_settings, _allow_fallback
+):
+    """Invoke optional prepare_metadata_for_build_editable
+
+    Implements a fallback by building an editable wheel if the hook isn't
+    defined, unless _allow_fallback is False in which case HookMissing is
+    raised.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.prepare_metadata_for_build_editable
+    except AttributeError:
+        if not _allow_fallback:
+            raise HookMissing()
+        try:
+            build_hook = backend.build_editable
+        except AttributeError:
+            raise HookMissing(hook_name="build_editable")
+        else:
+            whl_basename = build_hook(metadata_directory, config_settings)
+            return _get_wheel_metadata_from_wheel(
+                whl_basename, metadata_directory, config_settings
+            )
+    else:
+        return hook(metadata_directory, config_settings)
+
+
+WHEEL_BUILT_MARKER = "PYPROJECT_HOOKS_ALREADY_BUILT_WHEEL"
+
+
+def _dist_info_files(whl_zip):
+    """Identify the .dist-info folder inside a wheel ZipFile."""
+    res = []
+    for path in whl_zip.namelist():
+        m = re.match(r"[^/\\]+-[^/\\]+\.dist-info/", path)
+        if m:
+            res.append(path)
+    if res:
+        return res
+    raise Exception("No .dist-info folder found in wheel")
+
+
+def _get_wheel_metadata_from_wheel(whl_basename, metadata_directory, config_settings):
+    """Extract the metadata from a wheel.
+
+    Fallback for when the build backend does not
+    define the 'get_wheel_metadata' hook.
+    """
+    from zipfile import ZipFile
+
+    with open(os.path.join(metadata_directory, WHEEL_BUILT_MARKER), "wb"):
+        pass  # Touch marker file
+
+    whl_file = os.path.join(metadata_directory, whl_basename)
+    with ZipFile(whl_file) as zipf:
+        dist_info = _dist_info_files(zipf)
+        zipf.extractall(path=metadata_directory, members=dist_info)
+    return dist_info[0].split("/")[0]
+
+
+def _find_already_built_wheel(metadata_directory):
+    """Check for a wheel already built during the get_wheel_metadata hook."""
+    if not metadata_directory:
+        return None
+    metadata_parent = os.path.dirname(metadata_directory)
+    if not os.path.isfile(pjoin(metadata_parent, WHEEL_BUILT_MARKER)):
+        return None
+
+    whl_files = glob(os.path.join(metadata_parent, "*.whl"))
+    if not whl_files:
+        print("Found wheel built marker, but no .whl files")
+        return None
+    if len(whl_files) > 1:
+        print(
+            "Found multiple .whl files; unspecified behaviour. "
+            "Will call build_wheel."
+        )
+        return None
+
+    # Exactly one .whl file
+    return whl_files[0]
+
+
+def build_wheel(wheel_directory, config_settings, metadata_directory=None):
+    """Invoke the mandatory build_wheel hook.
+
+    If a wheel was already built in the
+    prepare_metadata_for_build_wheel fallback, this
+    will copy it rather than rebuilding the wheel.
+    """
+    prebuilt_whl = _find_already_built_wheel(metadata_directory)
+    if prebuilt_whl:
+        shutil.copy2(prebuilt_whl, wheel_directory)
+        return os.path.basename(prebuilt_whl)
+
+    return _build_backend().build_wheel(
+        wheel_directory, config_settings, metadata_directory
+    )
+
+
+def build_editable(wheel_directory, config_settings, metadata_directory=None):
+    """Invoke the optional build_editable hook.
+
+    If a wheel was already built in the
+    prepare_metadata_for_build_editable fallback, this
+    will copy it rather than rebuilding the wheel.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.build_editable
+    except AttributeError:
+        raise HookMissing()
+    else:
+        prebuilt_whl = _find_already_built_wheel(metadata_directory)
+        if prebuilt_whl:
+            shutil.copy2(prebuilt_whl, wheel_directory)
+            return os.path.basename(prebuilt_whl)
+
+        return hook(wheel_directory, config_settings, metadata_directory)
+
+
+def get_requires_for_build_sdist(config_settings):
+    """Invoke the optional get_requires_for_build_wheel hook
+
+    Returns [] if the hook is not defined.
+    """
+    backend = _build_backend()
+    try:
+        hook = backend.get_requires_for_build_sdist
+    except AttributeError:
+        return []
+    else:
+        return hook(config_settings)
+
+
+class _DummyException(Exception):
+    """Nothing should ever raise this exception"""
+
+
+class GotUnsupportedOperation(Exception):
+    """For internal use when backend raises UnsupportedOperation"""
+
+    def __init__(self, traceback):
+        self.traceback = traceback
+
+
+def build_sdist(sdist_directory, config_settings):
+    """Invoke the mandatory build_sdist hook."""
+    backend = _build_backend()
+    try:
+        return backend.build_sdist(sdist_directory, config_settings)
+    except getattr(backend, "UnsupportedOperation", _DummyException):
+        raise GotUnsupportedOperation(traceback.format_exc())
+
+
+HOOK_NAMES = {
+    "get_requires_for_build_wheel",
+    "prepare_metadata_for_build_wheel",
+    "build_wheel",
+    "get_requires_for_build_editable",
+    "prepare_metadata_for_build_editable",
+    "build_editable",
+    "get_requires_for_build_sdist",
+    "build_sdist",
+    "_supported_features",
+}
+
+
+def main():
+    if len(sys.argv) < 3:
+        sys.exit("Needs args: hook_name, control_dir")
+    hook_name = sys.argv[1]
+    control_dir = sys.argv[2]
+    if hook_name not in HOOK_NAMES:
+        sys.exit("Unknown hook: %s" % hook_name)
+
+    # Remove the parent directory from sys.path to avoid polluting the backend
+    # import namespace with this directory.
+    here = os.path.dirname(__file__)
+    if here in sys.path:
+        sys.path.remove(here)
+
+    hook = globals()[hook_name]
+
+    hook_input = read_json(pjoin(control_dir, "input.json"))
+
+    json_out = {"unsupported": False, "return_val": None}
+    try:
+        json_out["return_val"] = hook(**hook_input["kwargs"])
+    except BackendUnavailable as e:
+        json_out["no_backend"] = True
+        json_out["traceback"] = e.traceback
+        json_out["backend_error"] = e.message
+    except GotUnsupportedOperation as e:
+        json_out["unsupported"] = True
+        json_out["traceback"] = e.traceback
+    except HookMissing as e:
+        json_out["hook_missing"] = True
+        json_out["missing_hook_name"] = e.hook_name or hook_name
+
+    write_json(json_out, pjoin(control_dir, "output.json"), indent=2)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__init__.py
new file mode 100644
index 00000000..04230fc8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__init__.py
@@ -0,0 +1,179 @@
+#   __
+#  /__)  _  _     _   _ _/   _
+# / (   (- (/ (/ (- _)  /  _)
+#          /
+
+"""
+Requests HTTP Library
+~~~~~~~~~~~~~~~~~~~~~
+
+Requests is an HTTP library, written in Python, for human beings.
+Basic GET usage:
+
+   >>> import requests
+   >>> r = requests.get('https://www.python.org')
+   >>> r.status_code
+   200
+   >>> b'Python is a programming language' in r.content
+   True
+
+... or POST:
+
+   >>> payload = dict(key1='value1', key2='value2')
+   >>> r = requests.post('https://httpbin.org/post', data=payload)
+   >>> print(r.text)
+   {
+     ...
+     "form": {
+       "key1": "value1",
+       "key2": "value2"
+     },
+     ...
+   }
+
+The other HTTP methods are supported - see `requests.api`. Full documentation
+is at <https://requests.readthedocs.io>.
+
+:copyright: (c) 2017 by Kenneth Reitz.
+:license: Apache 2.0, see LICENSE for more details.
+"""
+
+import warnings
+
+from pip._vendor import urllib3
+
+from .exceptions import RequestsDependencyWarning
+
+charset_normalizer_version = None
+chardet_version = None
+
+
+def check_compatibility(urllib3_version, chardet_version, charset_normalizer_version):
+    urllib3_version = urllib3_version.split(".")
+    assert urllib3_version != ["dev"]  # Verify urllib3 isn't installed from git.
+
+    # Sometimes, urllib3 only reports its version as 16.1.
+    if len(urllib3_version) == 2:
+        urllib3_version.append("0")
+
+    # Check urllib3 for compatibility.
+    major, minor, patch = urllib3_version  # noqa: F811
+    major, minor, patch = int(major), int(minor), int(patch)
+    # urllib3 >= 1.21.1
+    assert major >= 1
+    if major == 1:
+        assert minor >= 21
+
+    # Check charset_normalizer for compatibility.
+    if chardet_version:
+        major, minor, patch = chardet_version.split(".")[:3]
+        major, minor, patch = int(major), int(minor), int(patch)
+        # chardet_version >= 3.0.2, < 6.0.0
+        assert (3, 0, 2) <= (major, minor, patch) < (6, 0, 0)
+    elif charset_normalizer_version:
+        major, minor, patch = charset_normalizer_version.split(".")[:3]
+        major, minor, patch = int(major), int(minor), int(patch)
+        # charset_normalizer >= 2.0.0 < 4.0.0
+        assert (2, 0, 0) <= (major, minor, patch) < (4, 0, 0)
+    else:
+        # pip does not need or use character detection
+        pass
+
+
+def _check_cryptography(cryptography_version):
+    # cryptography < 1.3.4
+    try:
+        cryptography_version = list(map(int, cryptography_version.split(".")))
+    except ValueError:
+        return
+
+    if cryptography_version < [1, 3, 4]:
+        warning = "Old version of cryptography ({}) may cause slowdown.".format(
+            cryptography_version
+        )
+        warnings.warn(warning, RequestsDependencyWarning)
+
+
+# Check imported dependencies for compatibility.
+try:
+    check_compatibility(
+        urllib3.__version__, chardet_version, charset_normalizer_version
+    )
+except (AssertionError, ValueError):
+    warnings.warn(
+        "urllib3 ({}) or chardet ({})/charset_normalizer ({}) doesn't match a supported "
+        "version!".format(
+            urllib3.__version__, chardet_version, charset_normalizer_version
+        ),
+        RequestsDependencyWarning,
+    )
+
+# Attempt to enable urllib3's fallback for SNI support
+# if the standard library doesn't support SNI or the
+# 'ssl' library isn't available.
+try:
+    # Note: This logic prevents upgrading cryptography on Windows, if imported
+    #       as part of pip.
+    from pip._internal.utils.compat import WINDOWS
+    if not WINDOWS:
+        raise ImportError("pip internals: don't import cryptography on Windows")
+    try:
+        import ssl
+    except ImportError:
+        ssl = None
+
+    if not getattr(ssl, "HAS_SNI", False):
+        from pip._vendor.urllib3.contrib import pyopenssl
+
+        pyopenssl.inject_into_urllib3()
+
+        # Check cryptography version
+        from cryptography import __version__ as cryptography_version
+
+        _check_cryptography(cryptography_version)
+except ImportError:
+    pass
+
+# urllib3's DependencyWarnings should be silenced.
+from pip._vendor.urllib3.exceptions import DependencyWarning
+
+warnings.simplefilter("ignore", DependencyWarning)
+
+# Set default logging handler to avoid "No handler found" warnings.
+import logging
+from logging import NullHandler
+
+from . import packages, utils
+from .__version__ import (
+    __author__,
+    __author_email__,
+    __build__,
+    __cake__,
+    __copyright__,
+    __description__,
+    __license__,
+    __title__,
+    __url__,
+    __version__,
+)
+from .api import delete, get, head, options, patch, post, put, request
+from .exceptions import (
+    ConnectionError,
+    ConnectTimeout,
+    FileModeWarning,
+    HTTPError,
+    JSONDecodeError,
+    ReadTimeout,
+    RequestException,
+    Timeout,
+    TooManyRedirects,
+    URLRequired,
+)
+from .models import PreparedRequest, Request, Response
+from .sessions import Session, session
+from .status_codes import codes
+
+logging.getLogger(__name__).addHandler(NullHandler())
+
+# FileModeWarnings go off per the default.
+warnings.simplefilter("default", FileModeWarning, append=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..60417262
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__version__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__version__.cpython-312.pyc
new file mode 100644
index 00000000..2875841e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/__version__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/_internal_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/_internal_utils.cpython-312.pyc
new file mode 100644
index 00000000..bc0645e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/_internal_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/adapters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/adapters.cpython-312.pyc
new file mode 100644
index 00000000..be73f60c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/adapters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..161ed427
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/auth.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/auth.cpython-312.pyc
new file mode 100644
index 00000000..47eadcd8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/auth.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/certs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/certs.cpython-312.pyc
new file mode 100644
index 00000000..147244f0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/certs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..707ba907
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/cookies.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/cookies.cpython-312.pyc
new file mode 100644
index 00000000..cb00d8a6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/cookies.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..e4fd15d9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/help.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/help.cpython-312.pyc
new file mode 100644
index 00000000..c7124a9c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/help.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/hooks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/hooks.cpython-312.pyc
new file mode 100644
index 00000000..ca575c62
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/hooks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/models.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/models.cpython-312.pyc
new file mode 100644
index 00000000..3279f74b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/models.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/packages.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/packages.cpython-312.pyc
new file mode 100644
index 00000000..c016d04d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/packages.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/sessions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/sessions.cpython-312.pyc
new file mode 100644
index 00000000..50014ad1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/sessions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/status_codes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/status_codes.cpython-312.pyc
new file mode 100644
index 00000000..2ea7dbb4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/status_codes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/structures.cpython-312.pyc
new file mode 100644
index 00000000..16eb25e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..53b114ed
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__version__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__version__.py
new file mode 100644
index 00000000..2c105aca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/__version__.py
@@ -0,0 +1,14 @@
+# .-. .-. .-. . . .-. .-. .-. .-.
+# |(  |-  |.| | | |-  `-.  |  `-.
+# ' ' `-' `-`.`-' `-' `-'  '  `-'
+
+__title__ = "requests"
+__description__ = "Python HTTP for Humans."
+__url__ = "https://requests.readthedocs.io"
+__version__ = "2.32.3"
+__build__ = 0x023203
+__author__ = "Kenneth Reitz"
+__author_email__ = "me@kennethreitz.org"
+__license__ = "Apache-2.0"
+__copyright__ = "Copyright Kenneth Reitz"
+__cake__ = "\u2728 \U0001f370 \u2728"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/_internal_utils.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/_internal_utils.py
new file mode 100644
index 00000000..f2cf635e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/_internal_utils.py
@@ -0,0 +1,50 @@
+"""
+requests._internal_utils
+~~~~~~~~~~~~~~
+
+Provides utility functions that are consumed internally by Requests
+which depend on extremely few external helpers (such as compat)
+"""
+import re
+
+from .compat import builtin_str
+
+_VALID_HEADER_NAME_RE_BYTE = re.compile(rb"^[^:\s][^:\r\n]*$")
+_VALID_HEADER_NAME_RE_STR = re.compile(r"^[^:\s][^:\r\n]*$")
+_VALID_HEADER_VALUE_RE_BYTE = re.compile(rb"^\S[^\r\n]*$|^$")
+_VALID_HEADER_VALUE_RE_STR = re.compile(r"^\S[^\r\n]*$|^$")
+
+_HEADER_VALIDATORS_STR = (_VALID_HEADER_NAME_RE_STR, _VALID_HEADER_VALUE_RE_STR)
+_HEADER_VALIDATORS_BYTE = (_VALID_HEADER_NAME_RE_BYTE, _VALID_HEADER_VALUE_RE_BYTE)
+HEADER_VALIDATORS = {
+    bytes: _HEADER_VALIDATORS_BYTE,
+    str: _HEADER_VALIDATORS_STR,
+}
+
+
+def to_native_string(string, encoding="ascii"):
+    """Given a string object, regardless of type, returns a representation of
+    that string in the native string type, encoding and decoding where
+    necessary. This assumes ASCII unless told otherwise.
+    """
+    if isinstance(string, builtin_str):
+        out = string
+    else:
+        out = string.decode(encoding)
+
+    return out
+
+
+def unicode_is_ascii(u_string):
+    """Determine if unicode string only contains ASCII characters.
+
+    :param str u_string: unicode string to check. Must be unicode
+        and not Python 2 `str`.
+    :rtype: bool
+    """
+    assert isinstance(u_string, str)
+    try:
+        u_string.encode("ascii")
+        return True
+    except UnicodeEncodeError:
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/adapters.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/adapters.py
new file mode 100644
index 00000000..70307774
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/adapters.py
@@ -0,0 +1,719 @@
+"""
+requests.adapters
+~~~~~~~~~~~~~~~~~
+
+This module contains the transport adapters that Requests uses to define
+and maintain connections.
+"""
+
+import os.path
+import socket  # noqa: F401
+import typing
+import warnings
+
+from pip._vendor.urllib3.exceptions import ClosedPoolError, ConnectTimeoutError
+from pip._vendor.urllib3.exceptions import HTTPError as _HTTPError
+from pip._vendor.urllib3.exceptions import InvalidHeader as _InvalidHeader
+from pip._vendor.urllib3.exceptions import (
+    LocationValueError,
+    MaxRetryError,
+    NewConnectionError,
+    ProtocolError,
+)
+from pip._vendor.urllib3.exceptions import ProxyError as _ProxyError
+from pip._vendor.urllib3.exceptions import ReadTimeoutError, ResponseError
+from pip._vendor.urllib3.exceptions import SSLError as _SSLError
+from pip._vendor.urllib3.poolmanager import PoolManager, proxy_from_url
+from pip._vendor.urllib3.util import Timeout as TimeoutSauce
+from pip._vendor.urllib3.util import parse_url
+from pip._vendor.urllib3.util.retry import Retry
+from pip._vendor.urllib3.util.ssl_ import create_urllib3_context
+
+from .auth import _basic_auth_str
+from .compat import basestring, urlparse
+from .cookies import extract_cookies_to_jar
+from .exceptions import (
+    ConnectionError,
+    ConnectTimeout,
+    InvalidHeader,
+    InvalidProxyURL,
+    InvalidSchema,
+    InvalidURL,
+    ProxyError,
+    ReadTimeout,
+    RetryError,
+    SSLError,
+)
+from .models import Response
+from .structures import CaseInsensitiveDict
+from .utils import (
+    DEFAULT_CA_BUNDLE_PATH,
+    extract_zipped_paths,
+    get_auth_from_url,
+    get_encoding_from_headers,
+    prepend_scheme_if_needed,
+    select_proxy,
+    urldefragauth,
+)
+
+try:
+    from pip._vendor.urllib3.contrib.socks import SOCKSProxyManager
+except ImportError:
+
+    def SOCKSProxyManager(*args, **kwargs):
+        raise InvalidSchema("Missing dependencies for SOCKS support.")
+
+
+if typing.TYPE_CHECKING:
+    from .models import PreparedRequest
+
+
+DEFAULT_POOLBLOCK = False
+DEFAULT_POOLSIZE = 10
+DEFAULT_RETRIES = 0
+DEFAULT_POOL_TIMEOUT = None
+
+
+try:
+    import ssl  # noqa: F401
+
+    _preloaded_ssl_context = create_urllib3_context()
+    _preloaded_ssl_context.load_verify_locations(
+        extract_zipped_paths(DEFAULT_CA_BUNDLE_PATH)
+    )
+except ImportError:
+    # Bypass default SSLContext creation when Python
+    # interpreter isn't built with the ssl module.
+    _preloaded_ssl_context = None
+
+
+def _urllib3_request_context(
+    request: "PreparedRequest",
+    verify: "bool | str | None",
+    client_cert: "typing.Tuple[str, str] | str | None",
+    poolmanager: "PoolManager",
+) -> "(typing.Dict[str, typing.Any], typing.Dict[str, typing.Any])":
+    host_params = {}
+    pool_kwargs = {}
+    parsed_request_url = urlparse(request.url)
+    scheme = parsed_request_url.scheme.lower()
+    port = parsed_request_url.port
+
+    # Determine if we have and should use our default SSLContext
+    # to optimize performance on standard requests.
+    poolmanager_kwargs = getattr(poolmanager, "connection_pool_kw", {})
+    has_poolmanager_ssl_context = poolmanager_kwargs.get("ssl_context")
+    should_use_default_ssl_context = (
+        _preloaded_ssl_context is not None and not has_poolmanager_ssl_context
+    )
+
+    cert_reqs = "CERT_REQUIRED"
+    if verify is False:
+        cert_reqs = "CERT_NONE"
+    elif verify is True and should_use_default_ssl_context:
+        pool_kwargs["ssl_context"] = _preloaded_ssl_context
+    elif isinstance(verify, str):
+        if not os.path.isdir(verify):
+            pool_kwargs["ca_certs"] = verify
+        else:
+            pool_kwargs["ca_cert_dir"] = verify
+    pool_kwargs["cert_reqs"] = cert_reqs
+    if client_cert is not None:
+        if isinstance(client_cert, tuple) and len(client_cert) == 2:
+            pool_kwargs["cert_file"] = client_cert[0]
+            pool_kwargs["key_file"] = client_cert[1]
+        else:
+            # According to our docs, we allow users to specify just the client
+            # cert path
+            pool_kwargs["cert_file"] = client_cert
+    host_params = {
+        "scheme": scheme,
+        "host": parsed_request_url.hostname,
+        "port": port,
+    }
+    return host_params, pool_kwargs
+
+
+class BaseAdapter:
+    """The Base Transport Adapter"""
+
+    def __init__(self):
+        super().__init__()
+
+    def send(
+        self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None
+    ):
+        """Sends PreparedRequest object. Returns Response object.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param stream: (optional) Whether to stream the request content.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple
+        :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use
+        :param cert: (optional) Any user-provided SSL certificate to be trusted.
+        :param proxies: (optional) The proxies dictionary to apply to the request.
+        """
+        raise NotImplementedError
+
+    def close(self):
+        """Cleans up adapter specific items."""
+        raise NotImplementedError
+
+
+class HTTPAdapter(BaseAdapter):
+    """The built-in HTTP Adapter for urllib3.
+
+    Provides a general-case interface for Requests sessions to contact HTTP and
+    HTTPS urls by implementing the Transport Adapter interface. This class will
+    usually be created by the :class:`Session <Session>` class under the
+    covers.
+
+    :param pool_connections: The number of urllib3 connection pools to cache.
+    :param pool_maxsize: The maximum number of connections to save in the pool.
+    :param max_retries: The maximum number of retries each connection
+        should attempt. Note, this applies only to failed DNS lookups, socket
+        connections and connection timeouts, never to requests where data has
+        made it to the server. By default, Requests does not retry failed
+        connections. If you need granular control over the conditions under
+        which we retry a request, import urllib3's ``Retry`` class and pass
+        that instead.
+    :param pool_block: Whether the connection pool should block for connections.
+
+    Usage::
+
+      >>> import requests
+      >>> s = requests.Session()
+      >>> a = requests.adapters.HTTPAdapter(max_retries=3)
+      >>> s.mount('http://', a)
+    """
+
+    __attrs__ = [
+        "max_retries",
+        "config",
+        "_pool_connections",
+        "_pool_maxsize",
+        "_pool_block",
+    ]
+
+    def __init__(
+        self,
+        pool_connections=DEFAULT_POOLSIZE,
+        pool_maxsize=DEFAULT_POOLSIZE,
+        max_retries=DEFAULT_RETRIES,
+        pool_block=DEFAULT_POOLBLOCK,
+    ):
+        if max_retries == DEFAULT_RETRIES:
+            self.max_retries = Retry(0, read=False)
+        else:
+            self.max_retries = Retry.from_int(max_retries)
+        self.config = {}
+        self.proxy_manager = {}
+
+        super().__init__()
+
+        self._pool_connections = pool_connections
+        self._pool_maxsize = pool_maxsize
+        self._pool_block = pool_block
+
+        self.init_poolmanager(pool_connections, pool_maxsize, block=pool_block)
+
+    def __getstate__(self):
+        return {attr: getattr(self, attr, None) for attr in self.__attrs__}
+
+    def __setstate__(self, state):
+        # Can't handle by adding 'proxy_manager' to self.__attrs__ because
+        # self.poolmanager uses a lambda function, which isn't pickleable.
+        self.proxy_manager = {}
+        self.config = {}
+
+        for attr, value in state.items():
+            setattr(self, attr, value)
+
+        self.init_poolmanager(
+            self._pool_connections, self._pool_maxsize, block=self._pool_block
+        )
+
+    def init_poolmanager(
+        self, connections, maxsize, block=DEFAULT_POOLBLOCK, **pool_kwargs
+    ):
+        """Initializes a urllib3 PoolManager.
+
+        This method should not be called from user code, and is only
+        exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param connections: The number of urllib3 connection pools to cache.
+        :param maxsize: The maximum number of connections to save in the pool.
+        :param block: Block when no free connections are available.
+        :param pool_kwargs: Extra keyword arguments used to initialize the Pool Manager.
+        """
+        # save these values for pickling
+        self._pool_connections = connections
+        self._pool_maxsize = maxsize
+        self._pool_block = block
+
+        self.poolmanager = PoolManager(
+            num_pools=connections,
+            maxsize=maxsize,
+            block=block,
+            **pool_kwargs,
+        )
+
+    def proxy_manager_for(self, proxy, **proxy_kwargs):
+        """Return urllib3 ProxyManager for the given proxy.
+
+        This method should not be called from user code, and is only
+        exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param proxy: The proxy to return a urllib3 ProxyManager for.
+        :param proxy_kwargs: Extra keyword arguments used to configure the Proxy Manager.
+        :returns: ProxyManager
+        :rtype: urllib3.ProxyManager
+        """
+        if proxy in self.proxy_manager:
+            manager = self.proxy_manager[proxy]
+        elif proxy.lower().startswith("socks"):
+            username, password = get_auth_from_url(proxy)
+            manager = self.proxy_manager[proxy] = SOCKSProxyManager(
+                proxy,
+                username=username,
+                password=password,
+                num_pools=self._pool_connections,
+                maxsize=self._pool_maxsize,
+                block=self._pool_block,
+                **proxy_kwargs,
+            )
+        else:
+            proxy_headers = self.proxy_headers(proxy)
+            manager = self.proxy_manager[proxy] = proxy_from_url(
+                proxy,
+                proxy_headers=proxy_headers,
+                num_pools=self._pool_connections,
+                maxsize=self._pool_maxsize,
+                block=self._pool_block,
+                **proxy_kwargs,
+            )
+
+        return manager
+
+    def cert_verify(self, conn, url, verify, cert):
+        """Verify a SSL certificate. This method should not be called from user
+        code, and is only exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param conn: The urllib3 connection object associated with the cert.
+        :param url: The requested URL.
+        :param verify: Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use
+        :param cert: The SSL certificate to verify.
+        """
+        if url.lower().startswith("https") and verify:
+            conn.cert_reqs = "CERT_REQUIRED"
+
+            # Only load the CA certificates if 'verify' is a string indicating the CA bundle to use.
+            # Otherwise, if verify is a boolean, we don't load anything since
+            # the connection will be using a context with the default certificates already loaded,
+            # and this avoids a call to the slow load_verify_locations()
+            if verify is not True:
+                # `verify` must be a str with a path then
+                cert_loc = verify
+
+                if not os.path.exists(cert_loc):
+                    raise OSError(
+                        f"Could not find a suitable TLS CA certificate bundle, "
+                        f"invalid path: {cert_loc}"
+                    )
+
+                if not os.path.isdir(cert_loc):
+                    conn.ca_certs = cert_loc
+                else:
+                    conn.ca_cert_dir = cert_loc
+        else:
+            conn.cert_reqs = "CERT_NONE"
+            conn.ca_certs = None
+            conn.ca_cert_dir = None
+
+        if cert:
+            if not isinstance(cert, basestring):
+                conn.cert_file = cert[0]
+                conn.key_file = cert[1]
+            else:
+                conn.cert_file = cert
+                conn.key_file = None
+            if conn.cert_file and not os.path.exists(conn.cert_file):
+                raise OSError(
+                    f"Could not find the TLS certificate file, "
+                    f"invalid path: {conn.cert_file}"
+                )
+            if conn.key_file and not os.path.exists(conn.key_file):
+                raise OSError(
+                    f"Could not find the TLS key file, invalid path: {conn.key_file}"
+                )
+
+    def build_response(self, req, resp):
+        """Builds a :class:`Response <requests.Response>` object from a urllib3
+        response. This should not be called from user code, and is only exposed
+        for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`
+
+        :param req: The :class:`PreparedRequest <PreparedRequest>` used to generate the response.
+        :param resp: The urllib3 response object.
+        :rtype: requests.Response
+        """
+        response = Response()
+
+        # Fallback to None if there's no status_code, for whatever reason.
+        response.status_code = getattr(resp, "status", None)
+
+        # Make headers case-insensitive.
+        response.headers = CaseInsensitiveDict(getattr(resp, "headers", {}))
+
+        # Set encoding.
+        response.encoding = get_encoding_from_headers(response.headers)
+        response.raw = resp
+        response.reason = response.raw.reason
+
+        if isinstance(req.url, bytes):
+            response.url = req.url.decode("utf-8")
+        else:
+            response.url = req.url
+
+        # Add new cookies from the server.
+        extract_cookies_to_jar(response.cookies, req, resp)
+
+        # Give the Response some context.
+        response.request = req
+        response.connection = self
+
+        return response
+
+    def build_connection_pool_key_attributes(self, request, verify, cert=None):
+        """Build the PoolKey attributes used by urllib3 to return a connection.
+
+        This looks at the PreparedRequest, the user-specified verify value,
+        and the value of the cert parameter to determine what PoolKey values
+        to use to select a connection from a given urllib3 Connection Pool.
+
+        The SSL related pool key arguments are not consistently set. As of
+        this writing, use the following to determine what keys may be in that
+        dictionary:
+
+        * If ``verify`` is ``True``, ``"ssl_context"`` will be set and will be the
+          default Requests SSL Context
+        * If ``verify`` is ``False``, ``"ssl_context"`` will not be set but
+          ``"cert_reqs"`` will be set
+        * If ``verify`` is a string, (i.e., it is a user-specified trust bundle)
+          ``"ca_certs"`` will be set if the string is not a directory recognized
+          by :py:func:`os.path.isdir`, otherwise ``"ca_certs_dir"`` will be
+          set.
+        * If ``"cert"`` is specified, ``"cert_file"`` will always be set. If
+          ``"cert"`` is a tuple with a second item, ``"key_file"`` will also
+          be present
+
+        To override these settings, one may subclass this class, call this
+        method and use the above logic to change parameters as desired. For
+        example, if one wishes to use a custom :py:class:`ssl.SSLContext` one
+        must both set ``"ssl_context"`` and based on what else they require,
+        alter the other keys to ensure the desired behaviour.
+
+        :param request:
+            The PreparedReqest being sent over the connection.
+        :type request:
+            :class:`~requests.models.PreparedRequest`
+        :param verify:
+            Either a boolean, in which case it controls whether
+            we verify the server's TLS certificate, or a string, in which case it
+            must be a path to a CA bundle to use.
+        :param cert:
+            (optional) Any user-provided SSL certificate for client
+            authentication (a.k.a., mTLS). This may be a string (i.e., just
+            the path to a file which holds both certificate and key) or a
+            tuple of length 2 with the certificate file path and key file
+            path.
+        :returns:
+            A tuple of two dictionaries. The first is the "host parameters"
+            portion of the Pool Key including scheme, hostname, and port. The
+            second is a dictionary of SSLContext related parameters.
+        """
+        return _urllib3_request_context(request, verify, cert, self.poolmanager)
+
+    def get_connection_with_tls_context(self, request, verify, proxies=None, cert=None):
+        """Returns a urllib3 connection for the given request and TLS settings.
+        This should not be called from user code, and is only exposed for use
+        when subclassing the :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request:
+            The :class:`PreparedRequest <PreparedRequest>` object to be sent
+            over the connection.
+        :param verify:
+            Either a boolean, in which case it controls whether we verify the
+            server's TLS certificate, or a string, in which case it must be a
+            path to a CA bundle to use.
+        :param proxies:
+            (optional) The proxies dictionary to apply to the request.
+        :param cert:
+            (optional) Any user-provided SSL certificate to be used for client
+            authentication (a.k.a., mTLS).
+        :rtype:
+            urllib3.ConnectionPool
+        """
+        proxy = select_proxy(request.url, proxies)
+        try:
+            host_params, pool_kwargs = self.build_connection_pool_key_attributes(
+                request,
+                verify,
+                cert,
+            )
+        except ValueError as e:
+            raise InvalidURL(e, request=request)
+        if proxy:
+            proxy = prepend_scheme_if_needed(proxy, "http")
+            proxy_url = parse_url(proxy)
+            if not proxy_url.host:
+                raise InvalidProxyURL(
+                    "Please check proxy URL. It is malformed "
+                    "and could be missing the host."
+                )
+            proxy_manager = self.proxy_manager_for(proxy)
+            conn = proxy_manager.connection_from_host(
+                **host_params, pool_kwargs=pool_kwargs
+            )
+        else:
+            # Only scheme should be lower case
+            conn = self.poolmanager.connection_from_host(
+                **host_params, pool_kwargs=pool_kwargs
+            )
+
+        return conn
+
+    def get_connection(self, url, proxies=None):
+        """DEPRECATED: Users should move to `get_connection_with_tls_context`
+        for all subclasses of HTTPAdapter using Requests>=2.32.2.
+
+        Returns a urllib3 connection for the given URL. This should not be
+        called from user code, and is only exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param url: The URL to connect to.
+        :param proxies: (optional) A Requests-style dictionary of proxies used on this request.
+        :rtype: urllib3.ConnectionPool
+        """
+        warnings.warn(
+            (
+                "`get_connection` has been deprecated in favor of "
+                "`get_connection_with_tls_context`. Custom HTTPAdapter subclasses "
+                "will need to migrate for Requests>=2.32.2. Please see "
+                "https://github.com/psf/requests/pull/6710 for more details."
+            ),
+            DeprecationWarning,
+        )
+        proxy = select_proxy(url, proxies)
+
+        if proxy:
+            proxy = prepend_scheme_if_needed(proxy, "http")
+            proxy_url = parse_url(proxy)
+            if not proxy_url.host:
+                raise InvalidProxyURL(
+                    "Please check proxy URL. It is malformed "
+                    "and could be missing the host."
+                )
+            proxy_manager = self.proxy_manager_for(proxy)
+            conn = proxy_manager.connection_from_url(url)
+        else:
+            # Only scheme should be lower case
+            parsed = urlparse(url)
+            url = parsed.geturl()
+            conn = self.poolmanager.connection_from_url(url)
+
+        return conn
+
+    def close(self):
+        """Disposes of any internal state.
+
+        Currently, this closes the PoolManager and any active ProxyManager,
+        which closes any pooled connections.
+        """
+        self.poolmanager.clear()
+        for proxy in self.proxy_manager.values():
+            proxy.clear()
+
+    def request_url(self, request, proxies):
+        """Obtain the url to use when making the final request.
+
+        If the message is being sent through a HTTP proxy, the full URL has to
+        be used. Otherwise, we should only use the path portion of the URL.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs.
+        :rtype: str
+        """
+        proxy = select_proxy(request.url, proxies)
+        scheme = urlparse(request.url).scheme
+
+        is_proxied_http_request = proxy and scheme != "https"
+        using_socks_proxy = False
+        if proxy:
+            proxy_scheme = urlparse(proxy).scheme.lower()
+            using_socks_proxy = proxy_scheme.startswith("socks")
+
+        url = request.path_url
+        if url.startswith("//"):  # Don't confuse urllib3
+            url = f"/{url.lstrip('/')}"
+
+        if is_proxied_http_request and not using_socks_proxy:
+            url = urldefragauth(request.url)
+
+        return url
+
+    def add_headers(self, request, **kwargs):
+        """Add any headers needed by the connection. As of v2.0 this does
+        nothing by default, but is left for overriding by users that subclass
+        the :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` to add headers to.
+        :param kwargs: The keyword arguments from the call to send().
+        """
+        pass
+
+    def proxy_headers(self, proxy):
+        """Returns a dictionary of the headers to add to any request sent
+        through a proxy. This works with urllib3 magic to ensure that they are
+        correctly sent to the proxy, rather than in a tunnelled request if
+        CONNECT is being used.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param proxy: The url of the proxy being used for this request.
+        :rtype: dict
+        """
+        headers = {}
+        username, password = get_auth_from_url(proxy)
+
+        if username:
+            headers["Proxy-Authorization"] = _basic_auth_str(username, password)
+
+        return headers
+
+    def send(
+        self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None
+    ):
+        """Sends PreparedRequest object. Returns Response object.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param stream: (optional) Whether to stream the request content.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple or urllib3 Timeout object
+        :param verify: (optional) Either a boolean, in which case it controls whether
+            we verify the server's TLS certificate, or a string, in which case it
+            must be a path to a CA bundle to use
+        :param cert: (optional) Any user-provided SSL certificate to be trusted.
+        :param proxies: (optional) The proxies dictionary to apply to the request.
+        :rtype: requests.Response
+        """
+
+        try:
+            conn = self.get_connection_with_tls_context(
+                request, verify, proxies=proxies, cert=cert
+            )
+        except LocationValueError as e:
+            raise InvalidURL(e, request=request)
+
+        self.cert_verify(conn, request.url, verify, cert)
+        url = self.request_url(request, proxies)
+        self.add_headers(
+            request,
+            stream=stream,
+            timeout=timeout,
+            verify=verify,
+            cert=cert,
+            proxies=proxies,
+        )
+
+        chunked = not (request.body is None or "Content-Length" in request.headers)
+
+        if isinstance(timeout, tuple):
+            try:
+                connect, read = timeout
+                timeout = TimeoutSauce(connect=connect, read=read)
+            except ValueError:
+                raise ValueError(
+                    f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, "
+                    f"or a single float to set both timeouts to the same value."
+                )
+        elif isinstance(timeout, TimeoutSauce):
+            pass
+        else:
+            timeout = TimeoutSauce(connect=timeout, read=timeout)
+
+        try:
+            resp = conn.urlopen(
+                method=request.method,
+                url=url,
+                body=request.body,
+                headers=request.headers,
+                redirect=False,
+                assert_same_host=False,
+                preload_content=False,
+                decode_content=False,
+                retries=self.max_retries,
+                timeout=timeout,
+                chunked=chunked,
+            )
+
+        except (ProtocolError, OSError) as err:
+            raise ConnectionError(err, request=request)
+
+        except MaxRetryError as e:
+            if isinstance(e.reason, ConnectTimeoutError):
+                # TODO: Remove this in 3.0.0: see #2811
+                if not isinstance(e.reason, NewConnectionError):
+                    raise ConnectTimeout(e, request=request)
+
+            if isinstance(e.reason, ResponseError):
+                raise RetryError(e, request=request)
+
+            if isinstance(e.reason, _ProxyError):
+                raise ProxyError(e, request=request)
+
+            if isinstance(e.reason, _SSLError):
+                # This branch is for urllib3 v1.22 and later.
+                raise SSLError(e, request=request)
+
+            raise ConnectionError(e, request=request)
+
+        except ClosedPoolError as e:
+            raise ConnectionError(e, request=request)
+
+        except _ProxyError as e:
+            raise ProxyError(e)
+
+        except (_SSLError, _HTTPError) as e:
+            if isinstance(e, _SSLError):
+                # This branch is for urllib3 versions earlier than v1.22
+                raise SSLError(e, request=request)
+            elif isinstance(e, ReadTimeoutError):
+                raise ReadTimeout(e, request=request)
+            elif isinstance(e, _InvalidHeader):
+                raise InvalidHeader(e, request=request)
+            else:
+                raise
+
+        return self.build_response(request, resp)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/api.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/api.py
new file mode 100644
index 00000000..59607445
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/api.py
@@ -0,0 +1,157 @@
+"""
+requests.api
+~~~~~~~~~~~~
+
+This module implements the Requests API.
+
+:copyright: (c) 2012 by Kenneth Reitz.
+:license: Apache2, see LICENSE for more details.
+"""
+
+from . import sessions
+
+
+def request(method, url, **kwargs):
+    """Constructs and sends a :class:`Request <Request>`.
+
+    :param method: method for the new :class:`Request` object: ``GET``, ``OPTIONS``, ``HEAD``, ``POST``, ``PUT``, ``PATCH``, or ``DELETE``.
+    :param url: URL for the new :class:`Request` object.
+    :param params: (optional) Dictionary, list of tuples or bytes to send
+        in the query string for the :class:`Request`.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param headers: (optional) Dictionary of HTTP Headers to send with the :class:`Request`.
+    :param cookies: (optional) Dict or CookieJar object to send with the :class:`Request`.
+    :param files: (optional) Dictionary of ``'name': file-like-objects`` (or ``{'name': file-tuple}``) for multipart encoding upload.
+        ``file-tuple`` can be a 2-tuple ``('filename', fileobj)``, 3-tuple ``('filename', fileobj, 'content_type')``
+        or a 4-tuple ``('filename', fileobj, 'content_type', custom_headers)``, where ``'content_type'`` is a string
+        defining the content type of the given file and ``custom_headers`` a dict-like object containing additional headers
+        to add for the file.
+    :param auth: (optional) Auth tuple to enable Basic/Digest/Custom HTTP Auth.
+    :param timeout: (optional) How many seconds to wait for the server to send data
+        before giving up, as a float, or a :ref:`(connect timeout, read
+        timeout) <timeouts>` tuple.
+    :type timeout: float or tuple
+    :param allow_redirects: (optional) Boolean. Enable/disable GET/OPTIONS/POST/PUT/PATCH/DELETE/HEAD redirection. Defaults to ``True``.
+    :type allow_redirects: bool
+    :param proxies: (optional) Dictionary mapping protocol to the URL of the proxy.
+    :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use. Defaults to ``True``.
+    :param stream: (optional) if ``False``, the response content will be immediately downloaded.
+    :param cert: (optional) if String, path to ssl client cert file (.pem). If Tuple, ('cert', 'key') pair.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.request('GET', 'https://httpbin.org/get')
+      >>> req
+      <Response [200]>
+    """
+
+    # By using the 'with' statement we are sure the session is closed, thus we
+    # avoid leaving sockets open which can trigger a ResourceWarning in some
+    # cases, and look like a memory leak in others.
+    with sessions.Session() as session:
+        return session.request(method=method, url=url, **kwargs)
+
+
+def get(url, params=None, **kwargs):
+    r"""Sends a GET request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param params: (optional) Dictionary, list of tuples or bytes to send
+        in the query string for the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("get", url, params=params, **kwargs)
+
+
+def options(url, **kwargs):
+    r"""Sends an OPTIONS request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("options", url, **kwargs)
+
+
+def head(url, **kwargs):
+    r"""Sends a HEAD request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes. If
+        `allow_redirects` is not provided, it will be set to `False` (as
+        opposed to the default :meth:`request` behavior).
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    kwargs.setdefault("allow_redirects", False)
+    return request("head", url, **kwargs)
+
+
+def post(url, data=None, json=None, **kwargs):
+    r"""Sends a POST request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("post", url, data=data, json=json, **kwargs)
+
+
+def put(url, data=None, **kwargs):
+    r"""Sends a PUT request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("put", url, data=data, **kwargs)
+
+
+def patch(url, data=None, **kwargs):
+    r"""Sends a PATCH request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("patch", url, data=data, **kwargs)
+
+
+def delete(url, **kwargs):
+    r"""Sends a DELETE request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("delete", url, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/auth.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/auth.py
new file mode 100644
index 00000000..4a7ce6dc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/auth.py
@@ -0,0 +1,314 @@
+"""
+requests.auth
+~~~~~~~~~~~~~
+
+This module contains the authentication handlers for Requests.
+"""
+
+import hashlib
+import os
+import re
+import threading
+import time
+import warnings
+from base64 import b64encode
+
+from ._internal_utils import to_native_string
+from .compat import basestring, str, urlparse
+from .cookies import extract_cookies_to_jar
+from .utils import parse_dict_header
+
+CONTENT_TYPE_FORM_URLENCODED = "application/x-www-form-urlencoded"
+CONTENT_TYPE_MULTI_PART = "multipart/form-data"
+
+
+def _basic_auth_str(username, password):
+    """Returns a Basic Auth string."""
+
+    # "I want us to put a big-ol' comment on top of it that
+    # says that this behaviour is dumb but we need to preserve
+    # it because people are relying on it."
+    #    - Lukasa
+    #
+    # These are here solely to maintain backwards compatibility
+    # for things like ints. This will be removed in 3.0.0.
+    if not isinstance(username, basestring):
+        warnings.warn(
+            "Non-string usernames will no longer be supported in Requests "
+            "3.0.0. Please convert the object you've passed in ({!r}) to "
+            "a string or bytes object in the near future to avoid "
+            "problems.".format(username),
+            category=DeprecationWarning,
+        )
+        username = str(username)
+
+    if not isinstance(password, basestring):
+        warnings.warn(
+            "Non-string passwords will no longer be supported in Requests "
+            "3.0.0. Please convert the object you've passed in ({!r}) to "
+            "a string or bytes object in the near future to avoid "
+            "problems.".format(type(password)),
+            category=DeprecationWarning,
+        )
+        password = str(password)
+    # -- End Removal --
+
+    if isinstance(username, str):
+        username = username.encode("latin1")
+
+    if isinstance(password, str):
+        password = password.encode("latin1")
+
+    authstr = "Basic " + to_native_string(
+        b64encode(b":".join((username, password))).strip()
+    )
+
+    return authstr
+
+
+class AuthBase:
+    """Base class that all auth implementations derive from"""
+
+    def __call__(self, r):
+        raise NotImplementedError("Auth hooks must be callable.")
+
+
+class HTTPBasicAuth(AuthBase):
+    """Attaches HTTP Basic Authentication to the given Request object."""
+
+    def __init__(self, username, password):
+        self.username = username
+        self.password = password
+
+    def __eq__(self, other):
+        return all(
+            [
+                self.username == getattr(other, "username", None),
+                self.password == getattr(other, "password", None),
+            ]
+        )
+
+    def __ne__(self, other):
+        return not self == other
+
+    def __call__(self, r):
+        r.headers["Authorization"] = _basic_auth_str(self.username, self.password)
+        return r
+
+
+class HTTPProxyAuth(HTTPBasicAuth):
+    """Attaches HTTP Proxy Authentication to a given Request object."""
+
+    def __call__(self, r):
+        r.headers["Proxy-Authorization"] = _basic_auth_str(self.username, self.password)
+        return r
+
+
+class HTTPDigestAuth(AuthBase):
+    """Attaches HTTP Digest Authentication to the given Request object."""
+
+    def __init__(self, username, password):
+        self.username = username
+        self.password = password
+        # Keep state in per-thread local storage
+        self._thread_local = threading.local()
+
+    def init_per_thread_state(self):
+        # Ensure state is initialized just once per-thread
+        if not hasattr(self._thread_local, "init"):
+            self._thread_local.init = True
+            self._thread_local.last_nonce = ""
+            self._thread_local.nonce_count = 0
+            self._thread_local.chal = {}
+            self._thread_local.pos = None
+            self._thread_local.num_401_calls = None
+
+    def build_digest_header(self, method, url):
+        """
+        :rtype: str
+        """
+
+        realm = self._thread_local.chal["realm"]
+        nonce = self._thread_local.chal["nonce"]
+        qop = self._thread_local.chal.get("qop")
+        algorithm = self._thread_local.chal.get("algorithm")
+        opaque = self._thread_local.chal.get("opaque")
+        hash_utf8 = None
+
+        if algorithm is None:
+            _algorithm = "MD5"
+        else:
+            _algorithm = algorithm.upper()
+        # lambdas assume digest modules are imported at the top level
+        if _algorithm == "MD5" or _algorithm == "MD5-SESS":
+
+            def md5_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.md5(x).hexdigest()
+
+            hash_utf8 = md5_utf8
+        elif _algorithm == "SHA":
+
+            def sha_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha1(x).hexdigest()
+
+            hash_utf8 = sha_utf8
+        elif _algorithm == "SHA-256":
+
+            def sha256_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha256(x).hexdigest()
+
+            hash_utf8 = sha256_utf8
+        elif _algorithm == "SHA-512":
+
+            def sha512_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha512(x).hexdigest()
+
+            hash_utf8 = sha512_utf8
+
+        KD = lambda s, d: hash_utf8(f"{s}:{d}")  # noqa:E731
+
+        if hash_utf8 is None:
+            return None
+
+        # XXX not implemented yet
+        entdig = None
+        p_parsed = urlparse(url)
+        #: path is request-uri defined in RFC 2616 which should not be empty
+        path = p_parsed.path or "/"
+        if p_parsed.query:
+            path += f"?{p_parsed.query}"
+
+        A1 = f"{self.username}:{realm}:{self.password}"
+        A2 = f"{method}:{path}"
+
+        HA1 = hash_utf8(A1)
+        HA2 = hash_utf8(A2)
+
+        if nonce == self._thread_local.last_nonce:
+            self._thread_local.nonce_count += 1
+        else:
+            self._thread_local.nonce_count = 1
+        ncvalue = f"{self._thread_local.nonce_count:08x}"
+        s = str(self._thread_local.nonce_count).encode("utf-8")
+        s += nonce.encode("utf-8")
+        s += time.ctime().encode("utf-8")
+        s += os.urandom(8)
+
+        cnonce = hashlib.sha1(s).hexdigest()[:16]
+        if _algorithm == "MD5-SESS":
+            HA1 = hash_utf8(f"{HA1}:{nonce}:{cnonce}")
+
+        if not qop:
+            respdig = KD(HA1, f"{nonce}:{HA2}")
+        elif qop == "auth" or "auth" in qop.split(","):
+            noncebit = f"{nonce}:{ncvalue}:{cnonce}:auth:{HA2}"
+            respdig = KD(HA1, noncebit)
+        else:
+            # XXX handle auth-int.
+            return None
+
+        self._thread_local.last_nonce = nonce
+
+        # XXX should the partial digests be encoded too?
+        base = (
+            f'username="{self.username}", realm="{realm}", nonce="{nonce}", '
+            f'uri="{path}", response="{respdig}"'
+        )
+        if opaque:
+            base += f', opaque="{opaque}"'
+        if algorithm:
+            base += f', algorithm="{algorithm}"'
+        if entdig:
+            base += f', digest="{entdig}"'
+        if qop:
+            base += f', qop="auth", nc={ncvalue}, cnonce="{cnonce}"'
+
+        return f"Digest {base}"
+
+    def handle_redirect(self, r, **kwargs):
+        """Reset num_401_calls counter on redirects."""
+        if r.is_redirect:
+            self._thread_local.num_401_calls = 1
+
+    def handle_401(self, r, **kwargs):
+        """
+        Takes the given response and tries digest-auth, if needed.
+
+        :rtype: requests.Response
+        """
+
+        # If response is not 4xx, do not auth
+        # See https://github.com/psf/requests/issues/3772
+        if not 400 <= r.status_code < 500:
+            self._thread_local.num_401_calls = 1
+            return r
+
+        if self._thread_local.pos is not None:
+            # Rewind the file position indicator of the body to where
+            # it was to resend the request.
+            r.request.body.seek(self._thread_local.pos)
+        s_auth = r.headers.get("www-authenticate", "")
+
+        if "digest" in s_auth.lower() and self._thread_local.num_401_calls < 2:
+            self._thread_local.num_401_calls += 1
+            pat = re.compile(r"digest ", flags=re.IGNORECASE)
+            self._thread_local.chal = parse_dict_header(pat.sub("", s_auth, count=1))
+
+            # Consume content and release the original connection
+            # to allow our new request to reuse the same one.
+            r.content
+            r.close()
+            prep = r.request.copy()
+            extract_cookies_to_jar(prep._cookies, r.request, r.raw)
+            prep.prepare_cookies(prep._cookies)
+
+            prep.headers["Authorization"] = self.build_digest_header(
+                prep.method, prep.url
+            )
+            _r = r.connection.send(prep, **kwargs)
+            _r.history.append(r)
+            _r.request = prep
+
+            return _r
+
+        self._thread_local.num_401_calls = 1
+        return r
+
+    def __call__(self, r):
+        # Initialize per-thread state, if needed
+        self.init_per_thread_state()
+        # If we have a saved nonce, skip the 401
+        if self._thread_local.last_nonce:
+            r.headers["Authorization"] = self.build_digest_header(r.method, r.url)
+        try:
+            self._thread_local.pos = r.body.tell()
+        except AttributeError:
+            # In the case of HTTPDigestAuth being reused and the body of
+            # the previous request was a file-like object, pos has the
+            # file position of the previous body. Ensure it's set to
+            # None.
+            self._thread_local.pos = None
+        r.register_hook("response", self.handle_401)
+        r.register_hook("response", self.handle_redirect)
+        self._thread_local.num_401_calls = 1
+
+        return r
+
+    def __eq__(self, other):
+        return all(
+            [
+                self.username == getattr(other, "username", None),
+                self.password == getattr(other, "password", None),
+            ]
+        )
+
+    def __ne__(self, other):
+        return not self == other
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/certs.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/certs.py
new file mode 100644
index 00000000..2743144b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/certs.py
@@ -0,0 +1,17 @@
+#!/usr/bin/env python
+
+"""
+requests.certs
+~~~~~~~~~~~~~~
+
+This module returns the preferred default CA certificate bundle. There is
+only one — the one from the certifi package.
+
+If you are packaging Requests, e.g., for a Linux distribution or a managed
+environment, you can change the definition of where() to return a separately
+packaged CA bundle.
+"""
+from pip._vendor.certifi import where
+
+if __name__ == "__main__":
+    print(where())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/compat.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/compat.py
new file mode 100644
index 00000000..7081da75
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/compat.py
@@ -0,0 +1,78 @@
+"""
+requests.compat
+~~~~~~~~~~~~~~~
+
+This module previously handled import compatibility issues
+between Python 2 and Python 3. It remains for backwards
+compatibility until the next major version.
+"""
+
+import sys
+
+# -------------------
+# Character Detection
+# -------------------
+
+
+def _resolve_char_detection():
+    """Find supported character detection libraries."""
+    chardet = None
+    return chardet
+
+
+chardet = _resolve_char_detection()
+
+# -------
+# Pythons
+# -------
+
+# Syntax sugar.
+_ver = sys.version_info
+
+#: Python 2.x?
+is_py2 = _ver[0] == 2
+
+#: Python 3.x?
+is_py3 = _ver[0] == 3
+
+# Note: We've patched out simplejson support in pip because it prevents
+#       upgrading simplejson on Windows.
+import json
+from json import JSONDecodeError
+
+# Keep OrderedDict for backwards compatibility.
+from collections import OrderedDict
+from collections.abc import Callable, Mapping, MutableMapping
+from http import cookiejar as cookielib
+from http.cookies import Morsel
+from io import StringIO
+
+# --------------
+# Legacy Imports
+# --------------
+from urllib.parse import (
+    quote,
+    quote_plus,
+    unquote,
+    unquote_plus,
+    urldefrag,
+    urlencode,
+    urljoin,
+    urlparse,
+    urlsplit,
+    urlunparse,
+)
+from urllib.request import (
+    getproxies,
+    getproxies_environment,
+    parse_http_list,
+    proxy_bypass,
+    proxy_bypass_environment,
+)
+
+builtin_str = str
+str = str
+bytes = bytes
+basestring = (str, bytes)
+numeric_types = (int, float)
+integer_types = (int,)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/cookies.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/cookies.py
new file mode 100644
index 00000000..f69d0cda
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/cookies.py
@@ -0,0 +1,561 @@
+"""
+requests.cookies
+~~~~~~~~~~~~~~~~
+
+Compatibility code to be able to use `http.cookiejar.CookieJar` with requests.
+
+requests.utils imports from here, so be careful with imports.
+"""
+
+import calendar
+import copy
+import time
+
+from ._internal_utils import to_native_string
+from .compat import Morsel, MutableMapping, cookielib, urlparse, urlunparse
+
+try:
+    import threading
+except ImportError:
+    import dummy_threading as threading
+
+
+class MockRequest:
+    """Wraps a `requests.Request` to mimic a `urllib2.Request`.
+
+    The code in `http.cookiejar.CookieJar` expects this interface in order to correctly
+    manage cookie policies, i.e., determine whether a cookie can be set, given the
+    domains of the request and the cookie.
+
+    The original request object is read-only. The client is responsible for collecting
+    the new headers via `get_new_headers()` and interpreting them appropriately. You
+    probably want `get_cookie_header`, defined below.
+    """
+
+    def __init__(self, request):
+        self._r = request
+        self._new_headers = {}
+        self.type = urlparse(self._r.url).scheme
+
+    def get_type(self):
+        return self.type
+
+    def get_host(self):
+        return urlparse(self._r.url).netloc
+
+    def get_origin_req_host(self):
+        return self.get_host()
+
+    def get_full_url(self):
+        # Only return the response's URL if the user hadn't set the Host
+        # header
+        if not self._r.headers.get("Host"):
+            return self._r.url
+        # If they did set it, retrieve it and reconstruct the expected domain
+        host = to_native_string(self._r.headers["Host"], encoding="utf-8")
+        parsed = urlparse(self._r.url)
+        # Reconstruct the URL as we expect it
+        return urlunparse(
+            [
+                parsed.scheme,
+                host,
+                parsed.path,
+                parsed.params,
+                parsed.query,
+                parsed.fragment,
+            ]
+        )
+
+    def is_unverifiable(self):
+        return True
+
+    def has_header(self, name):
+        return name in self._r.headers or name in self._new_headers
+
+    def get_header(self, name, default=None):
+        return self._r.headers.get(name, self._new_headers.get(name, default))
+
+    def add_header(self, key, val):
+        """cookiejar has no legitimate use for this method; add it back if you find one."""
+        raise NotImplementedError(
+            "Cookie headers should be added with add_unredirected_header()"
+        )
+
+    def add_unredirected_header(self, name, value):
+        self._new_headers[name] = value
+
+    def get_new_headers(self):
+        return self._new_headers
+
+    @property
+    def unverifiable(self):
+        return self.is_unverifiable()
+
+    @property
+    def origin_req_host(self):
+        return self.get_origin_req_host()
+
+    @property
+    def host(self):
+        return self.get_host()
+
+
+class MockResponse:
+    """Wraps a `httplib.HTTPMessage` to mimic a `urllib.addinfourl`.
+
+    ...what? Basically, expose the parsed HTTP headers from the server response
+    the way `http.cookiejar` expects to see them.
+    """
+
+    def __init__(self, headers):
+        """Make a MockResponse for `cookiejar` to read.
+
+        :param headers: a httplib.HTTPMessage or analogous carrying the headers
+        """
+        self._headers = headers
+
+    def info(self):
+        return self._headers
+
+    def getheaders(self, name):
+        self._headers.getheaders(name)
+
+
+def extract_cookies_to_jar(jar, request, response):
+    """Extract the cookies from the response into a CookieJar.
+
+    :param jar: http.cookiejar.CookieJar (not necessarily a RequestsCookieJar)
+    :param request: our own requests.Request object
+    :param response: urllib3.HTTPResponse object
+    """
+    if not (hasattr(response, "_original_response") and response._original_response):
+        return
+    # the _original_response field is the wrapped httplib.HTTPResponse object,
+    req = MockRequest(request)
+    # pull out the HTTPMessage with the headers and put it in the mock:
+    res = MockResponse(response._original_response.msg)
+    jar.extract_cookies(res, req)
+
+
+def get_cookie_header(jar, request):
+    """
+    Produce an appropriate Cookie header string to be sent with `request`, or None.
+
+    :rtype: str
+    """
+    r = MockRequest(request)
+    jar.add_cookie_header(r)
+    return r.get_new_headers().get("Cookie")
+
+
+def remove_cookie_by_name(cookiejar, name, domain=None, path=None):
+    """Unsets a cookie by name, by default over all domains and paths.
+
+    Wraps CookieJar.clear(), is O(n).
+    """
+    clearables = []
+    for cookie in cookiejar:
+        if cookie.name != name:
+            continue
+        if domain is not None and domain != cookie.domain:
+            continue
+        if path is not None and path != cookie.path:
+            continue
+        clearables.append((cookie.domain, cookie.path, cookie.name))
+
+    for domain, path, name in clearables:
+        cookiejar.clear(domain, path, name)
+
+
+class CookieConflictError(RuntimeError):
+    """There are two cookies that meet the criteria specified in the cookie jar.
+    Use .get and .set and include domain and path args in order to be more specific.
+    """
+
+
+class RequestsCookieJar(cookielib.CookieJar, MutableMapping):
+    """Compatibility class; is a http.cookiejar.CookieJar, but exposes a dict
+    interface.
+
+    This is the CookieJar we create by default for requests and sessions that
+    don't specify one, since some clients may expect response.cookies and
+    session.cookies to support dict operations.
+
+    Requests does not use the dict interface internally; it's just for
+    compatibility with external client code. All requests code should work
+    out of the box with externally provided instances of ``CookieJar``, e.g.
+    ``LWPCookieJar`` and ``FileCookieJar``.
+
+    Unlike a regular CookieJar, this class is pickleable.
+
+    .. warning:: dictionary operations that are normally O(1) may be O(n).
+    """
+
+    def get(self, name, default=None, domain=None, path=None):
+        """Dict-like get() that also supports optional domain and path args in
+        order to resolve naming collisions from using one cookie jar over
+        multiple domains.
+
+        .. warning:: operation is O(n), not O(1).
+        """
+        try:
+            return self._find_no_duplicates(name, domain, path)
+        except KeyError:
+            return default
+
+    def set(self, name, value, **kwargs):
+        """Dict-like set() that also supports optional domain and path args in
+        order to resolve naming collisions from using one cookie jar over
+        multiple domains.
+        """
+        # support client code that unsets cookies by assignment of a None value:
+        if value is None:
+            remove_cookie_by_name(
+                self, name, domain=kwargs.get("domain"), path=kwargs.get("path")
+            )
+            return
+
+        if isinstance(value, Morsel):
+            c = morsel_to_cookie(value)
+        else:
+            c = create_cookie(name, value, **kwargs)
+        self.set_cookie(c)
+        return c
+
+    def iterkeys(self):
+        """Dict-like iterkeys() that returns an iterator of names of cookies
+        from the jar.
+
+        .. seealso:: itervalues() and iteritems().
+        """
+        for cookie in iter(self):
+            yield cookie.name
+
+    def keys(self):
+        """Dict-like keys() that returns a list of names of cookies from the
+        jar.
+
+        .. seealso:: values() and items().
+        """
+        return list(self.iterkeys())
+
+    def itervalues(self):
+        """Dict-like itervalues() that returns an iterator of values of cookies
+        from the jar.
+
+        .. seealso:: iterkeys() and iteritems().
+        """
+        for cookie in iter(self):
+            yield cookie.value
+
+    def values(self):
+        """Dict-like values() that returns a list of values of cookies from the
+        jar.
+
+        .. seealso:: keys() and items().
+        """
+        return list(self.itervalues())
+
+    def iteritems(self):
+        """Dict-like iteritems() that returns an iterator of name-value tuples
+        from the jar.
+
+        .. seealso:: iterkeys() and itervalues().
+        """
+        for cookie in iter(self):
+            yield cookie.name, cookie.value
+
+    def items(self):
+        """Dict-like items() that returns a list of name-value tuples from the
+        jar. Allows client-code to call ``dict(RequestsCookieJar)`` and get a
+        vanilla python dict of key value pairs.
+
+        .. seealso:: keys() and values().
+        """
+        return list(self.iteritems())
+
+    def list_domains(self):
+        """Utility method to list all the domains in the jar."""
+        domains = []
+        for cookie in iter(self):
+            if cookie.domain not in domains:
+                domains.append(cookie.domain)
+        return domains
+
+    def list_paths(self):
+        """Utility method to list all the paths in the jar."""
+        paths = []
+        for cookie in iter(self):
+            if cookie.path not in paths:
+                paths.append(cookie.path)
+        return paths
+
+    def multiple_domains(self):
+        """Returns True if there are multiple domains in the jar.
+        Returns False otherwise.
+
+        :rtype: bool
+        """
+        domains = []
+        for cookie in iter(self):
+            if cookie.domain is not None and cookie.domain in domains:
+                return True
+            domains.append(cookie.domain)
+        return False  # there is only one domain in jar
+
+    def get_dict(self, domain=None, path=None):
+        """Takes as an argument an optional domain and path and returns a plain
+        old Python dict of name-value pairs of cookies that meet the
+        requirements.
+
+        :rtype: dict
+        """
+        dictionary = {}
+        for cookie in iter(self):
+            if (domain is None or cookie.domain == domain) and (
+                path is None or cookie.path == path
+            ):
+                dictionary[cookie.name] = cookie.value
+        return dictionary
+
+    def __contains__(self, name):
+        try:
+            return super().__contains__(name)
+        except CookieConflictError:
+            return True
+
+    def __getitem__(self, name):
+        """Dict-like __getitem__() for compatibility with client code. Throws
+        exception if there are more than one cookie with name. In that case,
+        use the more explicit get() method instead.
+
+        .. warning:: operation is O(n), not O(1).
+        """
+        return self._find_no_duplicates(name)
+
+    def __setitem__(self, name, value):
+        """Dict-like __setitem__ for compatibility with client code. Throws
+        exception if there is already a cookie of that name in the jar. In that
+        case, use the more explicit set() method instead.
+        """
+        self.set(name, value)
+
+    def __delitem__(self, name):
+        """Deletes a cookie given a name. Wraps ``http.cookiejar.CookieJar``'s
+        ``remove_cookie_by_name()``.
+        """
+        remove_cookie_by_name(self, name)
+
+    def set_cookie(self, cookie, *args, **kwargs):
+        if (
+            hasattr(cookie.value, "startswith")
+            and cookie.value.startswith('"')
+            and cookie.value.endswith('"')
+        ):
+            cookie.value = cookie.value.replace('\\"', "")
+        return super().set_cookie(cookie, *args, **kwargs)
+
+    def update(self, other):
+        """Updates this jar with cookies from another CookieJar or dict-like"""
+        if isinstance(other, cookielib.CookieJar):
+            for cookie in other:
+                self.set_cookie(copy.copy(cookie))
+        else:
+            super().update(other)
+
+    def _find(self, name, domain=None, path=None):
+        """Requests uses this method internally to get cookie values.
+
+        If there are conflicting cookies, _find arbitrarily chooses one.
+        See _find_no_duplicates if you want an exception thrown if there are
+        conflicting cookies.
+
+        :param name: a string containing name of cookie
+        :param domain: (optional) string containing domain of cookie
+        :param path: (optional) string containing path of cookie
+        :return: cookie.value
+        """
+        for cookie in iter(self):
+            if cookie.name == name:
+                if domain is None or cookie.domain == domain:
+                    if path is None or cookie.path == path:
+                        return cookie.value
+
+        raise KeyError(f"name={name!r}, domain={domain!r}, path={path!r}")
+
+    def _find_no_duplicates(self, name, domain=None, path=None):
+        """Both ``__get_item__`` and ``get`` call this function: it's never
+        used elsewhere in Requests.
+
+        :param name: a string containing name of cookie
+        :param domain: (optional) string containing domain of cookie
+        :param path: (optional) string containing path of cookie
+        :raises KeyError: if cookie is not found
+        :raises CookieConflictError: if there are multiple cookies
+            that match name and optionally domain and path
+        :return: cookie.value
+        """
+        toReturn = None
+        for cookie in iter(self):
+            if cookie.name == name:
+                if domain is None or cookie.domain == domain:
+                    if path is None or cookie.path == path:
+                        if toReturn is not None:
+                            # if there are multiple cookies that meet passed in criteria
+                            raise CookieConflictError(
+                                f"There are multiple cookies with name, {name!r}"
+                            )
+                        # we will eventually return this as long as no cookie conflict
+                        toReturn = cookie.value
+
+        if toReturn:
+            return toReturn
+        raise KeyError(f"name={name!r}, domain={domain!r}, path={path!r}")
+
+    def __getstate__(self):
+        """Unlike a normal CookieJar, this class is pickleable."""
+        state = self.__dict__.copy()
+        # remove the unpickleable RLock object
+        state.pop("_cookies_lock")
+        return state
+
+    def __setstate__(self, state):
+        """Unlike a normal CookieJar, this class is pickleable."""
+        self.__dict__.update(state)
+        if "_cookies_lock" not in self.__dict__:
+            self._cookies_lock = threading.RLock()
+
+    def copy(self):
+        """Return a copy of this RequestsCookieJar."""
+        new_cj = RequestsCookieJar()
+        new_cj.set_policy(self.get_policy())
+        new_cj.update(self)
+        return new_cj
+
+    def get_policy(self):
+        """Return the CookiePolicy instance used."""
+        return self._policy
+
+
+def _copy_cookie_jar(jar):
+    if jar is None:
+        return None
+
+    if hasattr(jar, "copy"):
+        # We're dealing with an instance of RequestsCookieJar
+        return jar.copy()
+    # We're dealing with a generic CookieJar instance
+    new_jar = copy.copy(jar)
+    new_jar.clear()
+    for cookie in jar:
+        new_jar.set_cookie(copy.copy(cookie))
+    return new_jar
+
+
+def create_cookie(name, value, **kwargs):
+    """Make a cookie from underspecified parameters.
+
+    By default, the pair of `name` and `value` will be set for the domain ''
+    and sent on every request (this is sometimes called a "supercookie").
+    """
+    result = {
+        "version": 0,
+        "name": name,
+        "value": value,
+        "port": None,
+        "domain": "",
+        "path": "/",
+        "secure": False,
+        "expires": None,
+        "discard": True,
+        "comment": None,
+        "comment_url": None,
+        "rest": {"HttpOnly": None},
+        "rfc2109": False,
+    }
+
+    badargs = set(kwargs) - set(result)
+    if badargs:
+        raise TypeError(
+            f"create_cookie() got unexpected keyword arguments: {list(badargs)}"
+        )
+
+    result.update(kwargs)
+    result["port_specified"] = bool(result["port"])
+    result["domain_specified"] = bool(result["domain"])
+    result["domain_initial_dot"] = result["domain"].startswith(".")
+    result["path_specified"] = bool(result["path"])
+
+    return cookielib.Cookie(**result)
+
+
+def morsel_to_cookie(morsel):
+    """Convert a Morsel object into a Cookie containing the one k/v pair."""
+
+    expires = None
+    if morsel["max-age"]:
+        try:
+            expires = int(time.time() + int(morsel["max-age"]))
+        except ValueError:
+            raise TypeError(f"max-age: {morsel['max-age']} must be integer")
+    elif morsel["expires"]:
+        time_template = "%a, %d-%b-%Y %H:%M:%S GMT"
+        expires = calendar.timegm(time.strptime(morsel["expires"], time_template))
+    return create_cookie(
+        comment=morsel["comment"],
+        comment_url=bool(morsel["comment"]),
+        discard=False,
+        domain=morsel["domain"],
+        expires=expires,
+        name=morsel.key,
+        path=morsel["path"],
+        port=None,
+        rest={"HttpOnly": morsel["httponly"]},
+        rfc2109=False,
+        secure=bool(morsel["secure"]),
+        value=morsel.value,
+        version=morsel["version"] or 0,
+    )
+
+
+def cookiejar_from_dict(cookie_dict, cookiejar=None, overwrite=True):
+    """Returns a CookieJar from a key/value dictionary.
+
+    :param cookie_dict: Dict of key/values to insert into CookieJar.
+    :param cookiejar: (optional) A cookiejar to add the cookies to.
+    :param overwrite: (optional) If False, will not replace cookies
+        already in the jar with new ones.
+    :rtype: CookieJar
+    """
+    if cookiejar is None:
+        cookiejar = RequestsCookieJar()
+
+    if cookie_dict is not None:
+        names_from_jar = [cookie.name for cookie in cookiejar]
+        for name in cookie_dict:
+            if overwrite or (name not in names_from_jar):
+                cookiejar.set_cookie(create_cookie(name, cookie_dict[name]))
+
+    return cookiejar
+
+
+def merge_cookies(cookiejar, cookies):
+    """Add cookies to cookiejar and returns a merged CookieJar.
+
+    :param cookiejar: CookieJar object to add the cookies to.
+    :param cookies: Dictionary or CookieJar object to be added.
+    :rtype: CookieJar
+    """
+    if not isinstance(cookiejar, cookielib.CookieJar):
+        raise ValueError("You can only merge into CookieJar")
+
+    if isinstance(cookies, dict):
+        cookiejar = cookiejar_from_dict(cookies, cookiejar=cookiejar, overwrite=False)
+    elif isinstance(cookies, cookielib.CookieJar):
+        try:
+            cookiejar.update(cookies)
+        except AttributeError:
+            for cookie_in_jar in cookies:
+                cookiejar.set_cookie(cookie_in_jar)
+
+    return cookiejar
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/exceptions.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/exceptions.py
new file mode 100644
index 00000000..7f3660f0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/exceptions.py
@@ -0,0 +1,151 @@
+"""
+requests.exceptions
+~~~~~~~~~~~~~~~~~~~
+
+This module contains the set of Requests' exceptions.
+"""
+from pip._vendor.urllib3.exceptions import HTTPError as BaseHTTPError
+
+from .compat import JSONDecodeError as CompatJSONDecodeError
+
+
+class RequestException(IOError):
+    """There was an ambiguous exception that occurred while handling your
+    request.
+    """
+
+    def __init__(self, *args, **kwargs):
+        """Initialize RequestException with `request` and `response` objects."""
+        response = kwargs.pop("response", None)
+        self.response = response
+        self.request = kwargs.pop("request", None)
+        if response is not None and not self.request and hasattr(response, "request"):
+            self.request = self.response.request
+        super().__init__(*args, **kwargs)
+
+
+class InvalidJSONError(RequestException):
+    """A JSON error occurred."""
+
+
+class JSONDecodeError(InvalidJSONError, CompatJSONDecodeError):
+    """Couldn't decode the text into json"""
+
+    def __init__(self, *args, **kwargs):
+        """
+        Construct the JSONDecodeError instance first with all
+        args. Then use it's args to construct the IOError so that
+        the json specific args aren't used as IOError specific args
+        and the error message from JSONDecodeError is preserved.
+        """
+        CompatJSONDecodeError.__init__(self, *args)
+        InvalidJSONError.__init__(self, *self.args, **kwargs)
+
+    def __reduce__(self):
+        """
+        The __reduce__ method called when pickling the object must
+        be the one from the JSONDecodeError (be it json/simplejson)
+        as it expects all the arguments for instantiation, not just
+        one like the IOError, and the MRO would by default call the
+        __reduce__ method from the IOError due to the inheritance order.
+        """
+        return CompatJSONDecodeError.__reduce__(self)
+
+
+class HTTPError(RequestException):
+    """An HTTP error occurred."""
+
+
+class ConnectionError(RequestException):
+    """A Connection error occurred."""
+
+
+class ProxyError(ConnectionError):
+    """A proxy error occurred."""
+
+
+class SSLError(ConnectionError):
+    """An SSL error occurred."""
+
+
+class Timeout(RequestException):
+    """The request timed out.
+
+    Catching this error will catch both
+    :exc:`~requests.exceptions.ConnectTimeout` and
+    :exc:`~requests.exceptions.ReadTimeout` errors.
+    """
+
+
+class ConnectTimeout(ConnectionError, Timeout):
+    """The request timed out while trying to connect to the remote server.
+
+    Requests that produced this error are safe to retry.
+    """
+
+
+class ReadTimeout(Timeout):
+    """The server did not send any data in the allotted amount of time."""
+
+
+class URLRequired(RequestException):
+    """A valid URL is required to make a request."""
+
+
+class TooManyRedirects(RequestException):
+    """Too many redirects."""
+
+
+class MissingSchema(RequestException, ValueError):
+    """The URL scheme (e.g. http or https) is missing."""
+
+
+class InvalidSchema(RequestException, ValueError):
+    """The URL scheme provided is either invalid or unsupported."""
+
+
+class InvalidURL(RequestException, ValueError):
+    """The URL provided was somehow invalid."""
+
+
+class InvalidHeader(RequestException, ValueError):
+    """The header value provided was somehow invalid."""
+
+
+class InvalidProxyURL(InvalidURL):
+    """The proxy URL provided is invalid."""
+
+
+class ChunkedEncodingError(RequestException):
+    """The server declared chunked encoding but sent an invalid chunk."""
+
+
+class ContentDecodingError(RequestException, BaseHTTPError):
+    """Failed to decode response content."""
+
+
+class StreamConsumedError(RequestException, TypeError):
+    """The content for this response was already consumed."""
+
+
+class RetryError(RequestException):
+    """Custom retries logic failed"""
+
+
+class UnrewindableBodyError(RequestException):
+    """Requests encountered an error when trying to rewind a body."""
+
+
+# Warnings
+
+
+class RequestsWarning(Warning):
+    """Base warning for Requests."""
+
+
+class FileModeWarning(RequestsWarning, DeprecationWarning):
+    """A file was opened in text mode, but Requests determined its binary length."""
+
+
+class RequestsDependencyWarning(RequestsWarning):
+    """An imported dependency doesn't match the expected version range."""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/help.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/help.py
new file mode 100644
index 00000000..ddbb6150
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/help.py
@@ -0,0 +1,127 @@
+"""Module containing bug report helper(s)."""
+
+import json
+import platform
+import ssl
+import sys
+
+from pip._vendor import idna
+from pip._vendor import urllib3
+
+from . import __version__ as requests_version
+
+charset_normalizer = None
+chardet = None
+
+try:
+    from pip._vendor.urllib3.contrib import pyopenssl
+except ImportError:
+    pyopenssl = None
+    OpenSSL = None
+    cryptography = None
+else:
+    import cryptography
+    import OpenSSL
+
+
+def _implementation():
+    """Return a dict with the Python implementation and version.
+
+    Provide both the name and the version of the Python implementation
+    currently running. For example, on CPython 3.10.3 it will return
+    {'name': 'CPython', 'version': '3.10.3'}.
+
+    This function works best on CPython and PyPy: in particular, it probably
+    doesn't work for Jython or IronPython. Future investigation should be done
+    to work out the correct shape of the code for those platforms.
+    """
+    implementation = platform.python_implementation()
+
+    if implementation == "CPython":
+        implementation_version = platform.python_version()
+    elif implementation == "PyPy":
+        implementation_version = "{}.{}.{}".format(
+            sys.pypy_version_info.major,
+            sys.pypy_version_info.minor,
+            sys.pypy_version_info.micro,
+        )
+        if sys.pypy_version_info.releaselevel != "final":
+            implementation_version = "".join(
+                [implementation_version, sys.pypy_version_info.releaselevel]
+            )
+    elif implementation == "Jython":
+        implementation_version = platform.python_version()  # Complete Guess
+    elif implementation == "IronPython":
+        implementation_version = platform.python_version()  # Complete Guess
+    else:
+        implementation_version = "Unknown"
+
+    return {"name": implementation, "version": implementation_version}
+
+
+def info():
+    """Generate information for a bug report."""
+    try:
+        platform_info = {
+            "system": platform.system(),
+            "release": platform.release(),
+        }
+    except OSError:
+        platform_info = {
+            "system": "Unknown",
+            "release": "Unknown",
+        }
+
+    implementation_info = _implementation()
+    urllib3_info = {"version": urllib3.__version__}
+    charset_normalizer_info = {"version": None}
+    chardet_info = {"version": None}
+    if charset_normalizer:
+        charset_normalizer_info = {"version": charset_normalizer.__version__}
+    if chardet:
+        chardet_info = {"version": chardet.__version__}
+
+    pyopenssl_info = {
+        "version": None,
+        "openssl_version": "",
+    }
+    if OpenSSL:
+        pyopenssl_info = {
+            "version": OpenSSL.__version__,
+            "openssl_version": f"{OpenSSL.SSL.OPENSSL_VERSION_NUMBER:x}",
+        }
+    cryptography_info = {
+        "version": getattr(cryptography, "__version__", ""),
+    }
+    idna_info = {
+        "version": getattr(idna, "__version__", ""),
+    }
+
+    system_ssl = ssl.OPENSSL_VERSION_NUMBER
+    system_ssl_info = {"version": f"{system_ssl:x}" if system_ssl is not None else ""}
+
+    return {
+        "platform": platform_info,
+        "implementation": implementation_info,
+        "system_ssl": system_ssl_info,
+        "using_pyopenssl": pyopenssl is not None,
+        "using_charset_normalizer": chardet is None,
+        "pyOpenSSL": pyopenssl_info,
+        "urllib3": urllib3_info,
+        "chardet": chardet_info,
+        "charset_normalizer": charset_normalizer_info,
+        "cryptography": cryptography_info,
+        "idna": idna_info,
+        "requests": {
+            "version": requests_version,
+        },
+    }
+
+
+def main():
+    """Pretty-print the bug information as JSON."""
+    print(json.dumps(info(), sort_keys=True, indent=2))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/hooks.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/hooks.py
new file mode 100644
index 00000000..d181ba2e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/hooks.py
@@ -0,0 +1,33 @@
+"""
+requests.hooks
+~~~~~~~~~~~~~~
+
+This module provides the capabilities for the Requests hooks system.
+
+Available hooks:
+
+``response``:
+    The response generated from a Request.
+"""
+HOOKS = ["response"]
+
+
+def default_hooks():
+    return {event: [] for event in HOOKS}
+
+
+# TODO: response is the only one
+
+
+def dispatch_hook(key, hooks, hook_data, **kwargs):
+    """Dispatches a hook dictionary on a given piece of data."""
+    hooks = hooks or {}
+    hooks = hooks.get(key)
+    if hooks:
+        if hasattr(hooks, "__call__"):
+            hooks = [hooks]
+        for hook in hooks:
+            _hook_data = hook(hook_data, **kwargs)
+            if _hook_data is not None:
+                hook_data = _hook_data
+    return hook_data
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/models.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/models.py
new file mode 100644
index 00000000..85a008cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/models.py
@@ -0,0 +1,1037 @@
+"""
+requests.models
+~~~~~~~~~~~~~~~
+
+This module contains the primary objects that power Requests.
+"""
+
+import datetime
+
+# Import encoding now, to avoid implicit import later.
+# Implicit import within threads may cause LookupError when standard library is in a ZIP,
+# such as in Embedded Python. See https://github.com/psf/requests/issues/3578.
+import encodings.idna  # noqa: F401
+from io import UnsupportedOperation
+
+from pip._vendor.urllib3.exceptions import (
+    DecodeError,
+    LocationParseError,
+    ProtocolError,
+    ReadTimeoutError,
+    SSLError,
+)
+from pip._vendor.urllib3.fields import RequestField
+from pip._vendor.urllib3.filepost import encode_multipart_formdata
+from pip._vendor.urllib3.util import parse_url
+
+from ._internal_utils import to_native_string, unicode_is_ascii
+from .auth import HTTPBasicAuth
+from .compat import (
+    Callable,
+    JSONDecodeError,
+    Mapping,
+    basestring,
+    builtin_str,
+    chardet,
+    cookielib,
+)
+from .compat import json as complexjson
+from .compat import urlencode, urlsplit, urlunparse
+from .cookies import _copy_cookie_jar, cookiejar_from_dict, get_cookie_header
+from .exceptions import (
+    ChunkedEncodingError,
+    ConnectionError,
+    ContentDecodingError,
+    HTTPError,
+    InvalidJSONError,
+    InvalidURL,
+)
+from .exceptions import JSONDecodeError as RequestsJSONDecodeError
+from .exceptions import MissingSchema
+from .exceptions import SSLError as RequestsSSLError
+from .exceptions import StreamConsumedError
+from .hooks import default_hooks
+from .status_codes import codes
+from .structures import CaseInsensitiveDict
+from .utils import (
+    check_header_validity,
+    get_auth_from_url,
+    guess_filename,
+    guess_json_utf,
+    iter_slices,
+    parse_header_links,
+    requote_uri,
+    stream_decode_response_unicode,
+    super_len,
+    to_key_val_list,
+)
+
+#: The set of HTTP status codes that indicate an automatically
+#: processable redirect.
+REDIRECT_STATI = (
+    codes.moved,  # 301
+    codes.found,  # 302
+    codes.other,  # 303
+    codes.temporary_redirect,  # 307
+    codes.permanent_redirect,  # 308
+)
+
+DEFAULT_REDIRECT_LIMIT = 30
+CONTENT_CHUNK_SIZE = 10 * 1024
+ITER_CHUNK_SIZE = 512
+
+
+class RequestEncodingMixin:
+    @property
+    def path_url(self):
+        """Build the path URL to use."""
+
+        url = []
+
+        p = urlsplit(self.url)
+
+        path = p.path
+        if not path:
+            path = "/"
+
+        url.append(path)
+
+        query = p.query
+        if query:
+            url.append("?")
+            url.append(query)
+
+        return "".join(url)
+
+    @staticmethod
+    def _encode_params(data):
+        """Encode parameters in a piece of data.
+
+        Will successfully encode parameters when passed as a dict or a list of
+        2-tuples. Order is retained if data is a list of 2-tuples but arbitrary
+        if parameters are supplied as a dict.
+        """
+
+        if isinstance(data, (str, bytes)):
+            return data
+        elif hasattr(data, "read"):
+            return data
+        elif hasattr(data, "__iter__"):
+            result = []
+            for k, vs in to_key_val_list(data):
+                if isinstance(vs, basestring) or not hasattr(vs, "__iter__"):
+                    vs = [vs]
+                for v in vs:
+                    if v is not None:
+                        result.append(
+                            (
+                                k.encode("utf-8") if isinstance(k, str) else k,
+                                v.encode("utf-8") if isinstance(v, str) else v,
+                            )
+                        )
+            return urlencode(result, doseq=True)
+        else:
+            return data
+
+    @staticmethod
+    def _encode_files(files, data):
+        """Build the body for a multipart/form-data request.
+
+        Will successfully encode files when passed as a dict or a list of
+        tuples. Order is retained if data is a list of tuples but arbitrary
+        if parameters are supplied as a dict.
+        The tuples may be 2-tuples (filename, fileobj), 3-tuples (filename, fileobj, contentype)
+        or 4-tuples (filename, fileobj, contentype, custom_headers).
+        """
+        if not files:
+            raise ValueError("Files must be provided.")
+        elif isinstance(data, basestring):
+            raise ValueError("Data must not be a string.")
+
+        new_fields = []
+        fields = to_key_val_list(data or {})
+        files = to_key_val_list(files or {})
+
+        for field, val in fields:
+            if isinstance(val, basestring) or not hasattr(val, "__iter__"):
+                val = [val]
+            for v in val:
+                if v is not None:
+                    # Don't call str() on bytestrings: in Py3 it all goes wrong.
+                    if not isinstance(v, bytes):
+                        v = str(v)
+
+                    new_fields.append(
+                        (
+                            field.decode("utf-8")
+                            if isinstance(field, bytes)
+                            else field,
+                            v.encode("utf-8") if isinstance(v, str) else v,
+                        )
+                    )
+
+        for k, v in files:
+            # support for explicit filename
+            ft = None
+            fh = None
+            if isinstance(v, (tuple, list)):
+                if len(v) == 2:
+                    fn, fp = v
+                elif len(v) == 3:
+                    fn, fp, ft = v
+                else:
+                    fn, fp, ft, fh = v
+            else:
+                fn = guess_filename(v) or k
+                fp = v
+
+            if isinstance(fp, (str, bytes, bytearray)):
+                fdata = fp
+            elif hasattr(fp, "read"):
+                fdata = fp.read()
+            elif fp is None:
+                continue
+            else:
+                fdata = fp
+
+            rf = RequestField(name=k, data=fdata, filename=fn, headers=fh)
+            rf.make_multipart(content_type=ft)
+            new_fields.append(rf)
+
+        body, content_type = encode_multipart_formdata(new_fields)
+
+        return body, content_type
+
+
+class RequestHooksMixin:
+    def register_hook(self, event, hook):
+        """Properly register a hook."""
+
+        if event not in self.hooks:
+            raise ValueError(f'Unsupported event specified, with event name "{event}"')
+
+        if isinstance(hook, Callable):
+            self.hooks[event].append(hook)
+        elif hasattr(hook, "__iter__"):
+            self.hooks[event].extend(h for h in hook if isinstance(h, Callable))
+
+    def deregister_hook(self, event, hook):
+        """Deregister a previously registered hook.
+        Returns True if the hook existed, False if not.
+        """
+
+        try:
+            self.hooks[event].remove(hook)
+            return True
+        except ValueError:
+            return False
+
+
+class Request(RequestHooksMixin):
+    """A user-created :class:`Request <Request>` object.
+
+    Used to prepare a :class:`PreparedRequest <PreparedRequest>`, which is sent to the server.
+
+    :param method: HTTP method to use.
+    :param url: URL to send.
+    :param headers: dictionary of headers to send.
+    :param files: dictionary of {filename: fileobject} files to multipart upload.
+    :param data: the body to attach to the request. If a dictionary or
+        list of tuples ``[(key, value)]`` is provided, form-encoding will
+        take place.
+    :param json: json for the body to attach to the request (if files or data is not specified).
+    :param params: URL parameters to append to the URL. If a dictionary or
+        list of tuples ``[(key, value)]`` is provided, form-encoding will
+        take place.
+    :param auth: Auth handler or (user, pass) tuple.
+    :param cookies: dictionary or CookieJar of cookies to attach to this request.
+    :param hooks: dictionary of callback hooks, for internal usage.
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.Request('GET', 'https://httpbin.org/get')
+      >>> req.prepare()
+      <PreparedRequest [GET]>
+    """
+
+    def __init__(
+        self,
+        method=None,
+        url=None,
+        headers=None,
+        files=None,
+        data=None,
+        params=None,
+        auth=None,
+        cookies=None,
+        hooks=None,
+        json=None,
+    ):
+        # Default empty dicts for dict params.
+        data = [] if data is None else data
+        files = [] if files is None else files
+        headers = {} if headers is None else headers
+        params = {} if params is None else params
+        hooks = {} if hooks is None else hooks
+
+        self.hooks = default_hooks()
+        for k, v in list(hooks.items()):
+            self.register_hook(event=k, hook=v)
+
+        self.method = method
+        self.url = url
+        self.headers = headers
+        self.files = files
+        self.data = data
+        self.json = json
+        self.params = params
+        self.auth = auth
+        self.cookies = cookies
+
+    def __repr__(self):
+        return f"<Request [{self.method}]>"
+
+    def prepare(self):
+        """Constructs a :class:`PreparedRequest <PreparedRequest>` for transmission and returns it."""
+        p = PreparedRequest()
+        p.prepare(
+            method=self.method,
+            url=self.url,
+            headers=self.headers,
+            files=self.files,
+            data=self.data,
+            json=self.json,
+            params=self.params,
+            auth=self.auth,
+            cookies=self.cookies,
+            hooks=self.hooks,
+        )
+        return p
+
+
+class PreparedRequest(RequestEncodingMixin, RequestHooksMixin):
+    """The fully mutable :class:`PreparedRequest <PreparedRequest>` object,
+    containing the exact bytes that will be sent to the server.
+
+    Instances are generated from a :class:`Request <Request>` object, and
+    should not be instantiated manually; doing so may produce undesirable
+    effects.
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.Request('GET', 'https://httpbin.org/get')
+      >>> r = req.prepare()
+      >>> r
+      <PreparedRequest [GET]>
+
+      >>> s = requests.Session()
+      >>> s.send(r)
+      <Response [200]>
+    """
+
+    def __init__(self):
+        #: HTTP verb to send to the server.
+        self.method = None
+        #: HTTP URL to send the request to.
+        self.url = None
+        #: dictionary of HTTP headers.
+        self.headers = None
+        # The `CookieJar` used to create the Cookie header will be stored here
+        # after prepare_cookies is called
+        self._cookies = None
+        #: request body to send to the server.
+        self.body = None
+        #: dictionary of callback hooks, for internal usage.
+        self.hooks = default_hooks()
+        #: integer denoting starting position of a readable file-like body.
+        self._body_position = None
+
+    def prepare(
+        self,
+        method=None,
+        url=None,
+        headers=None,
+        files=None,
+        data=None,
+        params=None,
+        auth=None,
+        cookies=None,
+        hooks=None,
+        json=None,
+    ):
+        """Prepares the entire request with the given parameters."""
+
+        self.prepare_method(method)
+        self.prepare_url(url, params)
+        self.prepare_headers(headers)
+        self.prepare_cookies(cookies)
+        self.prepare_body(data, files, json)
+        self.prepare_auth(auth, url)
+
+        # Note that prepare_auth must be last to enable authentication schemes
+        # such as OAuth to work on a fully prepared request.
+
+        # This MUST go after prepare_auth. Authenticators could add a hook
+        self.prepare_hooks(hooks)
+
+    def __repr__(self):
+        return f"<PreparedRequest [{self.method}]>"
+
+    def copy(self):
+        p = PreparedRequest()
+        p.method = self.method
+        p.url = self.url
+        p.headers = self.headers.copy() if self.headers is not None else None
+        p._cookies = _copy_cookie_jar(self._cookies)
+        p.body = self.body
+        p.hooks = self.hooks
+        p._body_position = self._body_position
+        return p
+
+    def prepare_method(self, method):
+        """Prepares the given HTTP method."""
+        self.method = method
+        if self.method is not None:
+            self.method = to_native_string(self.method.upper())
+
+    @staticmethod
+    def _get_idna_encoded_host(host):
+        from pip._vendor import idna
+
+        try:
+            host = idna.encode(host, uts46=True).decode("utf-8")
+        except idna.IDNAError:
+            raise UnicodeError
+        return host
+
+    def prepare_url(self, url, params):
+        """Prepares the given HTTP URL."""
+        #: Accept objects that have string representations.
+        #: We're unable to blindly call unicode/str functions
+        #: as this will include the bytestring indicator (b'')
+        #: on python 3.x.
+        #: https://github.com/psf/requests/pull/2238
+        if isinstance(url, bytes):
+            url = url.decode("utf8")
+        else:
+            url = str(url)
+
+        # Remove leading whitespaces from url
+        url = url.lstrip()
+
+        # Don't do any URL preparation for non-HTTP schemes like `mailto`,
+        # `data` etc to work around exceptions from `url_parse`, which
+        # handles RFC 3986 only.
+        if ":" in url and not url.lower().startswith("http"):
+            self.url = url
+            return
+
+        # Support for unicode domain names and paths.
+        try:
+            scheme, auth, host, port, path, query, fragment = parse_url(url)
+        except LocationParseError as e:
+            raise InvalidURL(*e.args)
+
+        if not scheme:
+            raise MissingSchema(
+                f"Invalid URL {url!r}: No scheme supplied. "
+                f"Perhaps you meant https://{url}?"
+            )
+
+        if not host:
+            raise InvalidURL(f"Invalid URL {url!r}: No host supplied")
+
+        # In general, we want to try IDNA encoding the hostname if the string contains
+        # non-ASCII characters. This allows users to automatically get the correct IDNA
+        # behaviour. For strings containing only ASCII characters, we need to also verify
+        # it doesn't start with a wildcard (*), before allowing the unencoded hostname.
+        if not unicode_is_ascii(host):
+            try:
+                host = self._get_idna_encoded_host(host)
+            except UnicodeError:
+                raise InvalidURL("URL has an invalid label.")
+        elif host.startswith(("*", ".")):
+            raise InvalidURL("URL has an invalid label.")
+
+        # Carefully reconstruct the network location
+        netloc = auth or ""
+        if netloc:
+            netloc += "@"
+        netloc += host
+        if port:
+            netloc += f":{port}"
+
+        # Bare domains aren't valid URLs.
+        if not path:
+            path = "/"
+
+        if isinstance(params, (str, bytes)):
+            params = to_native_string(params)
+
+        enc_params = self._encode_params(params)
+        if enc_params:
+            if query:
+                query = f"{query}&{enc_params}"
+            else:
+                query = enc_params
+
+        url = requote_uri(urlunparse([scheme, netloc, path, None, query, fragment]))
+        self.url = url
+
+    def prepare_headers(self, headers):
+        """Prepares the given HTTP headers."""
+
+        self.headers = CaseInsensitiveDict()
+        if headers:
+            for header in headers.items():
+                # Raise exception on invalid header value.
+                check_header_validity(header)
+                name, value = header
+                self.headers[to_native_string(name)] = value
+
+    def prepare_body(self, data, files, json=None):
+        """Prepares the given HTTP body data."""
+
+        # Check if file, fo, generator, iterator.
+        # If not, run through normal process.
+
+        # Nottin' on you.
+        body = None
+        content_type = None
+
+        if not data and json is not None:
+            # urllib3 requires a bytes-like body. Python 2's json.dumps
+            # provides this natively, but Python 3 gives a Unicode string.
+            content_type = "application/json"
+
+            try:
+                body = complexjson.dumps(json, allow_nan=False)
+            except ValueError as ve:
+                raise InvalidJSONError(ve, request=self)
+
+            if not isinstance(body, bytes):
+                body = body.encode("utf-8")
+
+        is_stream = all(
+            [
+                hasattr(data, "__iter__"),
+                not isinstance(data, (basestring, list, tuple, Mapping)),
+            ]
+        )
+
+        if is_stream:
+            try:
+                length = super_len(data)
+            except (TypeError, AttributeError, UnsupportedOperation):
+                length = None
+
+            body = data
+
+            if getattr(body, "tell", None) is not None:
+                # Record the current file position before reading.
+                # This will allow us to rewind a file in the event
+                # of a redirect.
+                try:
+                    self._body_position = body.tell()
+                except OSError:
+                    # This differentiates from None, allowing us to catch
+                    # a failed `tell()` later when trying to rewind the body
+                    self._body_position = object()
+
+            if files:
+                raise NotImplementedError(
+                    "Streamed bodies and files are mutually exclusive."
+                )
+
+            if length:
+                self.headers["Content-Length"] = builtin_str(length)
+            else:
+                self.headers["Transfer-Encoding"] = "chunked"
+        else:
+            # Multi-part file uploads.
+            if files:
+                (body, content_type) = self._encode_files(files, data)
+            else:
+                if data:
+                    body = self._encode_params(data)
+                    if isinstance(data, basestring) or hasattr(data, "read"):
+                        content_type = None
+                    else:
+                        content_type = "application/x-www-form-urlencoded"
+
+            self.prepare_content_length(body)
+
+            # Add content-type if it wasn't explicitly provided.
+            if content_type and ("content-type" not in self.headers):
+                self.headers["Content-Type"] = content_type
+
+        self.body = body
+
+    def prepare_content_length(self, body):
+        """Prepare Content-Length header based on request method and body"""
+        if body is not None:
+            length = super_len(body)
+            if length:
+                # If length exists, set it. Otherwise, we fallback
+                # to Transfer-Encoding: chunked.
+                self.headers["Content-Length"] = builtin_str(length)
+        elif (
+            self.method not in ("GET", "HEAD")
+            and self.headers.get("Content-Length") is None
+        ):
+            # Set Content-Length to 0 for methods that can have a body
+            # but don't provide one. (i.e. not GET or HEAD)
+            self.headers["Content-Length"] = "0"
+
+    def prepare_auth(self, auth, url=""):
+        """Prepares the given HTTP auth data."""
+
+        # If no Auth is explicitly provided, extract it from the URL first.
+        if auth is None:
+            url_auth = get_auth_from_url(self.url)
+            auth = url_auth if any(url_auth) else None
+
+        if auth:
+            if isinstance(auth, tuple) and len(auth) == 2:
+                # special-case basic HTTP auth
+                auth = HTTPBasicAuth(*auth)
+
+            # Allow auth to make its changes.
+            r = auth(self)
+
+            # Update self to reflect the auth changes.
+            self.__dict__.update(r.__dict__)
+
+            # Recompute Content-Length
+            self.prepare_content_length(self.body)
+
+    def prepare_cookies(self, cookies):
+        """Prepares the given HTTP cookie data.
+
+        This function eventually generates a ``Cookie`` header from the
+        given cookies using cookielib. Due to cookielib's design, the header
+        will not be regenerated if it already exists, meaning this function
+        can only be called once for the life of the
+        :class:`PreparedRequest <PreparedRequest>` object. Any subsequent calls
+        to ``prepare_cookies`` will have no actual effect, unless the "Cookie"
+        header is removed beforehand.
+        """
+        if isinstance(cookies, cookielib.CookieJar):
+            self._cookies = cookies
+        else:
+            self._cookies = cookiejar_from_dict(cookies)
+
+        cookie_header = get_cookie_header(self._cookies, self)
+        if cookie_header is not None:
+            self.headers["Cookie"] = cookie_header
+
+    def prepare_hooks(self, hooks):
+        """Prepares the given hooks."""
+        # hooks can be passed as None to the prepare method and to this
+        # method. To prevent iterating over None, simply use an empty list
+        # if hooks is False-y
+        hooks = hooks or []
+        for event in hooks:
+            self.register_hook(event, hooks[event])
+
+
+class Response:
+    """The :class:`Response <Response>` object, which contains a
+    server's response to an HTTP request.
+    """
+
+    __attrs__ = [
+        "_content",
+        "status_code",
+        "headers",
+        "url",
+        "history",
+        "encoding",
+        "reason",
+        "cookies",
+        "elapsed",
+        "request",
+    ]
+
+    def __init__(self):
+        self._content = False
+        self._content_consumed = False
+        self._next = None
+
+        #: Integer Code of responded HTTP Status, e.g. 404 or 200.
+        self.status_code = None
+
+        #: Case-insensitive Dictionary of Response Headers.
+        #: For example, ``headers['content-encoding']`` will return the
+        #: value of a ``'Content-Encoding'`` response header.
+        self.headers = CaseInsensitiveDict()
+
+        #: File-like object representation of response (for advanced usage).
+        #: Use of ``raw`` requires that ``stream=True`` be set on the request.
+        #: This requirement does not apply for use internally to Requests.
+        self.raw = None
+
+        #: Final URL location of Response.
+        self.url = None
+
+        #: Encoding to decode with when accessing r.text.
+        self.encoding = None
+
+        #: A list of :class:`Response <Response>` objects from
+        #: the history of the Request. Any redirect responses will end
+        #: up here. The list is sorted from the oldest to the most recent request.
+        self.history = []
+
+        #: Textual reason of responded HTTP Status, e.g. "Not Found" or "OK".
+        self.reason = None
+
+        #: A CookieJar of Cookies the server sent back.
+        self.cookies = cookiejar_from_dict({})
+
+        #: The amount of time elapsed between sending the request
+        #: and the arrival of the response (as a timedelta).
+        #: This property specifically measures the time taken between sending
+        #: the first byte of the request and finishing parsing the headers. It
+        #: is therefore unaffected by consuming the response content or the
+        #: value of the ``stream`` keyword argument.
+        self.elapsed = datetime.timedelta(0)
+
+        #: The :class:`PreparedRequest <PreparedRequest>` object to which this
+        #: is a response.
+        self.request = None
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *args):
+        self.close()
+
+    def __getstate__(self):
+        # Consume everything; accessing the content attribute makes
+        # sure the content has been fully read.
+        if not self._content_consumed:
+            self.content
+
+        return {attr: getattr(self, attr, None) for attr in self.__attrs__}
+
+    def __setstate__(self, state):
+        for name, value in state.items():
+            setattr(self, name, value)
+
+        # pickled objects do not have .raw
+        setattr(self, "_content_consumed", True)
+        setattr(self, "raw", None)
+
+    def __repr__(self):
+        return f"<Response [{self.status_code}]>"
+
+    def __bool__(self):
+        """Returns True if :attr:`status_code` is less than 400.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code, is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        return self.ok
+
+    def __nonzero__(self):
+        """Returns True if :attr:`status_code` is less than 400.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code, is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        return self.ok
+
+    def __iter__(self):
+        """Allows you to use a response as an iterator."""
+        return self.iter_content(128)
+
+    @property
+    def ok(self):
+        """Returns True if :attr:`status_code` is less than 400, False if not.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        try:
+            self.raise_for_status()
+        except HTTPError:
+            return False
+        return True
+
+    @property
+    def is_redirect(self):
+        """True if this Response is a well-formed HTTP redirect that could have
+        been processed automatically (by :meth:`Session.resolve_redirects`).
+        """
+        return "location" in self.headers and self.status_code in REDIRECT_STATI
+
+    @property
+    def is_permanent_redirect(self):
+        """True if this Response one of the permanent versions of redirect."""
+        return "location" in self.headers and self.status_code in (
+            codes.moved_permanently,
+            codes.permanent_redirect,
+        )
+
+    @property
+    def next(self):
+        """Returns a PreparedRequest for the next request in a redirect chain, if there is one."""
+        return self._next
+
+    @property
+    def apparent_encoding(self):
+        """The apparent encoding, provided by the charset_normalizer or chardet libraries."""
+        if chardet is not None:
+            return chardet.detect(self.content)["encoding"]
+        else:
+            # If no character detection library is available, we'll fall back
+            # to a standard Python utf-8 str.
+            return "utf-8"
+
+    def iter_content(self, chunk_size=1, decode_unicode=False):
+        """Iterates over the response data.  When stream=True is set on the
+        request, this avoids reading the content at once into memory for
+        large responses.  The chunk size is the number of bytes it should
+        read into memory.  This is not necessarily the length of each item
+        returned as decoding can take place.
+
+        chunk_size must be of type int or None. A value of None will
+        function differently depending on the value of `stream`.
+        stream=True will read data as it arrives in whatever size the
+        chunks are received. If stream=False, data is returned as
+        a single chunk.
+
+        If decode_unicode is True, content will be decoded using the best
+        available encoding based on the response.
+        """
+
+        def generate():
+            # Special case for urllib3.
+            if hasattr(self.raw, "stream"):
+                try:
+                    yield from self.raw.stream(chunk_size, decode_content=True)
+                except ProtocolError as e:
+                    raise ChunkedEncodingError(e)
+                except DecodeError as e:
+                    raise ContentDecodingError(e)
+                except ReadTimeoutError as e:
+                    raise ConnectionError(e)
+                except SSLError as e:
+                    raise RequestsSSLError(e)
+            else:
+                # Standard file-like object.
+                while True:
+                    chunk = self.raw.read(chunk_size)
+                    if not chunk:
+                        break
+                    yield chunk
+
+            self._content_consumed = True
+
+        if self._content_consumed and isinstance(self._content, bool):
+            raise StreamConsumedError()
+        elif chunk_size is not None and not isinstance(chunk_size, int):
+            raise TypeError(
+                f"chunk_size must be an int, it is instead a {type(chunk_size)}."
+            )
+        # simulate reading small chunks of the content
+        reused_chunks = iter_slices(self._content, chunk_size)
+
+        stream_chunks = generate()
+
+        chunks = reused_chunks if self._content_consumed else stream_chunks
+
+        if decode_unicode:
+            chunks = stream_decode_response_unicode(chunks, self)
+
+        return chunks
+
+    def iter_lines(
+        self, chunk_size=ITER_CHUNK_SIZE, decode_unicode=False, delimiter=None
+    ):
+        """Iterates over the response data, one line at a time.  When
+        stream=True is set on the request, this avoids reading the
+        content at once into memory for large responses.
+
+        .. note:: This method is not reentrant safe.
+        """
+
+        pending = None
+
+        for chunk in self.iter_content(
+            chunk_size=chunk_size, decode_unicode=decode_unicode
+        ):
+            if pending is not None:
+                chunk = pending + chunk
+
+            if delimiter:
+                lines = chunk.split(delimiter)
+            else:
+                lines = chunk.splitlines()
+
+            if lines and lines[-1] and chunk and lines[-1][-1] == chunk[-1]:
+                pending = lines.pop()
+            else:
+                pending = None
+
+            yield from lines
+
+        if pending is not None:
+            yield pending
+
+    @property
+    def content(self):
+        """Content of the response, in bytes."""
+
+        if self._content is False:
+            # Read the contents.
+            if self._content_consumed:
+                raise RuntimeError("The content for this response was already consumed")
+
+            if self.status_code == 0 or self.raw is None:
+                self._content = None
+            else:
+                self._content = b"".join(self.iter_content(CONTENT_CHUNK_SIZE)) or b""
+
+        self._content_consumed = True
+        # don't need to release the connection; that's been handled by urllib3
+        # since we exhausted the data.
+        return self._content
+
+    @property
+    def text(self):
+        """Content of the response, in unicode.
+
+        If Response.encoding is None, encoding will be guessed using
+        ``charset_normalizer`` or ``chardet``.
+
+        The encoding of the response content is determined based solely on HTTP
+        headers, following RFC 2616 to the letter. If you can take advantage of
+        non-HTTP knowledge to make a better guess at the encoding, you should
+        set ``r.encoding`` appropriately before accessing this property.
+        """
+
+        # Try charset from content-type
+        content = None
+        encoding = self.encoding
+
+        if not self.content:
+            return ""
+
+        # Fallback to auto-detected encoding.
+        if self.encoding is None:
+            encoding = self.apparent_encoding
+
+        # Decode unicode from given encoding.
+        try:
+            content = str(self.content, encoding, errors="replace")
+        except (LookupError, TypeError):
+            # A LookupError is raised if the encoding was not found which could
+            # indicate a misspelling or similar mistake.
+            #
+            # A TypeError can be raised if encoding is None
+            #
+            # So we try blindly encoding.
+            content = str(self.content, errors="replace")
+
+        return content
+
+    def json(self, **kwargs):
+        r"""Returns the json-encoded content of a response, if any.
+
+        :param \*\*kwargs: Optional arguments that ``json.loads`` takes.
+        :raises requests.exceptions.JSONDecodeError: If the response body does not
+            contain valid json.
+        """
+
+        if not self.encoding and self.content and len(self.content) > 3:
+            # No encoding set. JSON RFC 4627 section 3 states we should expect
+            # UTF-8, -16 or -32. Detect which one to use; If the detection or
+            # decoding fails, fall back to `self.text` (using charset_normalizer to make
+            # a best guess).
+            encoding = guess_json_utf(self.content)
+            if encoding is not None:
+                try:
+                    return complexjson.loads(self.content.decode(encoding), **kwargs)
+                except UnicodeDecodeError:
+                    # Wrong UTF codec detected; usually because it's not UTF-8
+                    # but some other 8-bit codec.  This is an RFC violation,
+                    # and the server didn't bother to tell us what codec *was*
+                    # used.
+                    pass
+                except JSONDecodeError as e:
+                    raise RequestsJSONDecodeError(e.msg, e.doc, e.pos)
+
+        try:
+            return complexjson.loads(self.text, **kwargs)
+        except JSONDecodeError as e:
+            # Catch JSON-related errors and raise as requests.JSONDecodeError
+            # This aliases json.JSONDecodeError and simplejson.JSONDecodeError
+            raise RequestsJSONDecodeError(e.msg, e.doc, e.pos)
+
+    @property
+    def links(self):
+        """Returns the parsed header links of the response, if any."""
+
+        header = self.headers.get("link")
+
+        resolved_links = {}
+
+        if header:
+            links = parse_header_links(header)
+
+            for link in links:
+                key = link.get("rel") or link.get("url")
+                resolved_links[key] = link
+
+        return resolved_links
+
+    def raise_for_status(self):
+        """Raises :class:`HTTPError`, if one occurred."""
+
+        http_error_msg = ""
+        if isinstance(self.reason, bytes):
+            # We attempt to decode utf-8 first because some servers
+            # choose to localize their reason strings. If the string
+            # isn't utf-8, we fall back to iso-8859-1 for all other
+            # encodings. (See PR #3538)
+            try:
+                reason = self.reason.decode("utf-8")
+            except UnicodeDecodeError:
+                reason = self.reason.decode("iso-8859-1")
+        else:
+            reason = self.reason
+
+        if 400 <= self.status_code < 500:
+            http_error_msg = (
+                f"{self.status_code} Client Error: {reason} for url: {self.url}"
+            )
+
+        elif 500 <= self.status_code < 600:
+            http_error_msg = (
+                f"{self.status_code} Server Error: {reason} for url: {self.url}"
+            )
+
+        if http_error_msg:
+            raise HTTPError(http_error_msg, response=self)
+
+    def close(self):
+        """Releases the connection back to the pool. Once this method has been
+        called the underlying ``raw`` object must not be accessed again.
+
+        *Note: Should not normally need to be called explicitly.*
+        """
+        if not self._content_consumed:
+            self.raw.close()
+
+        release_conn = getattr(self.raw, "release_conn", None)
+        if release_conn is not None:
+            release_conn()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/packages.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/packages.py
new file mode 100644
index 00000000..200c3828
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/packages.py
@@ -0,0 +1,25 @@
+import sys
+
+from .compat import chardet
+
+# This code exists for backwards compatibility reasons.
+# I don't like it either. Just look the other way. :)
+
+for package in ("urllib3", "idna"):
+    vendored_package = "pip._vendor." + package
+    locals()[package] = __import__(vendored_package)
+    # This traversal is apparently necessary such that the identities are
+    # preserved (requests.packages.urllib3.* is urllib3.*)
+    for mod in list(sys.modules):
+        if mod == vendored_package or mod.startswith(vendored_package + '.'):
+            unprefixed_mod = mod[len("pip._vendor."):]
+            sys.modules['pip._vendor.requests.packages.' + unprefixed_mod] = sys.modules[mod]
+
+if chardet is not None:
+    target = chardet.__name__
+    for mod in list(sys.modules):
+        if mod == target or mod.startswith(f"{target}."):
+            imported_mod = sys.modules[mod]
+            sys.modules[f"requests.packages.{mod}"] = imported_mod
+            mod = mod.replace(target, "chardet")
+            sys.modules[f"requests.packages.{mod}"] = imported_mod
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/sessions.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/sessions.py
new file mode 100644
index 00000000..b387bc36
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/sessions.py
@@ -0,0 +1,831 @@
+"""
+requests.sessions
+~~~~~~~~~~~~~~~~~
+
+This module provides a Session object to manage and persist settings across
+requests (cookies, auth, proxies).
+"""
+import os
+import sys
+import time
+from collections import OrderedDict
+from datetime import timedelta
+
+from ._internal_utils import to_native_string
+from .adapters import HTTPAdapter
+from .auth import _basic_auth_str
+from .compat import Mapping, cookielib, urljoin, urlparse
+from .cookies import (
+    RequestsCookieJar,
+    cookiejar_from_dict,
+    extract_cookies_to_jar,
+    merge_cookies,
+)
+from .exceptions import (
+    ChunkedEncodingError,
+    ContentDecodingError,
+    InvalidSchema,
+    TooManyRedirects,
+)
+from .hooks import default_hooks, dispatch_hook
+
+# formerly defined here, reexposed here for backward compatibility
+from .models import (  # noqa: F401
+    DEFAULT_REDIRECT_LIMIT,
+    REDIRECT_STATI,
+    PreparedRequest,
+    Request,
+)
+from .status_codes import codes
+from .structures import CaseInsensitiveDict
+from .utils import (  # noqa: F401
+    DEFAULT_PORTS,
+    default_headers,
+    get_auth_from_url,
+    get_environ_proxies,
+    get_netrc_auth,
+    requote_uri,
+    resolve_proxies,
+    rewind_body,
+    should_bypass_proxies,
+    to_key_val_list,
+)
+
+# Preferred clock, based on which one is more accurate on a given system.
+if sys.platform == "win32":
+    preferred_clock = time.perf_counter
+else:
+    preferred_clock = time.time
+
+
+def merge_setting(request_setting, session_setting, dict_class=OrderedDict):
+    """Determines appropriate setting for a given request, taking into account
+    the explicit setting on that request, and the setting in the session. If a
+    setting is a dictionary, they will be merged together using `dict_class`
+    """
+
+    if session_setting is None:
+        return request_setting
+
+    if request_setting is None:
+        return session_setting
+
+    # Bypass if not a dictionary (e.g. verify)
+    if not (
+        isinstance(session_setting, Mapping) and isinstance(request_setting, Mapping)
+    ):
+        return request_setting
+
+    merged_setting = dict_class(to_key_val_list(session_setting))
+    merged_setting.update(to_key_val_list(request_setting))
+
+    # Remove keys that are set to None. Extract keys first to avoid altering
+    # the dictionary during iteration.
+    none_keys = [k for (k, v) in merged_setting.items() if v is None]
+    for key in none_keys:
+        del merged_setting[key]
+
+    return merged_setting
+
+
+def merge_hooks(request_hooks, session_hooks, dict_class=OrderedDict):
+    """Properly merges both requests and session hooks.
+
+    This is necessary because when request_hooks == {'response': []}, the
+    merge breaks Session hooks entirely.
+    """
+    if session_hooks is None or session_hooks.get("response") == []:
+        return request_hooks
+
+    if request_hooks is None or request_hooks.get("response") == []:
+        return session_hooks
+
+    return merge_setting(request_hooks, session_hooks, dict_class)
+
+
+class SessionRedirectMixin:
+    def get_redirect_target(self, resp):
+        """Receives a Response. Returns a redirect URI or ``None``"""
+        # Due to the nature of how requests processes redirects this method will
+        # be called at least once upon the original response and at least twice
+        # on each subsequent redirect response (if any).
+        # If a custom mixin is used to handle this logic, it may be advantageous
+        # to cache the redirect location onto the response object as a private
+        # attribute.
+        if resp.is_redirect:
+            location = resp.headers["location"]
+            # Currently the underlying http module on py3 decode headers
+            # in latin1, but empirical evidence suggests that latin1 is very
+            # rarely used with non-ASCII characters in HTTP headers.
+            # It is more likely to get UTF8 header rather than latin1.
+            # This causes incorrect handling of UTF8 encoded location headers.
+            # To solve this, we re-encode the location in latin1.
+            location = location.encode("latin1")
+            return to_native_string(location, "utf8")
+        return None
+
+    def should_strip_auth(self, old_url, new_url):
+        """Decide whether Authorization header should be removed when redirecting"""
+        old_parsed = urlparse(old_url)
+        new_parsed = urlparse(new_url)
+        if old_parsed.hostname != new_parsed.hostname:
+            return True
+        # Special case: allow http -> https redirect when using the standard
+        # ports. This isn't specified by RFC 7235, but is kept to avoid
+        # breaking backwards compatibility with older versions of requests
+        # that allowed any redirects on the same host.
+        if (
+            old_parsed.scheme == "http"
+            and old_parsed.port in (80, None)
+            and new_parsed.scheme == "https"
+            and new_parsed.port in (443, None)
+        ):
+            return False
+
+        # Handle default port usage corresponding to scheme.
+        changed_port = old_parsed.port != new_parsed.port
+        changed_scheme = old_parsed.scheme != new_parsed.scheme
+        default_port = (DEFAULT_PORTS.get(old_parsed.scheme, None), None)
+        if (
+            not changed_scheme
+            and old_parsed.port in default_port
+            and new_parsed.port in default_port
+        ):
+            return False
+
+        # Standard case: root URI must match
+        return changed_port or changed_scheme
+
+    def resolve_redirects(
+        self,
+        resp,
+        req,
+        stream=False,
+        timeout=None,
+        verify=True,
+        cert=None,
+        proxies=None,
+        yield_requests=False,
+        **adapter_kwargs,
+    ):
+        """Receives a Response. Returns a generator of Responses or Requests."""
+
+        hist = []  # keep track of history
+
+        url = self.get_redirect_target(resp)
+        previous_fragment = urlparse(req.url).fragment
+        while url:
+            prepared_request = req.copy()
+
+            # Update history and keep track of redirects.
+            # resp.history must ignore the original request in this loop
+            hist.append(resp)
+            resp.history = hist[1:]
+
+            try:
+                resp.content  # Consume socket so it can be released
+            except (ChunkedEncodingError, ContentDecodingError, RuntimeError):
+                resp.raw.read(decode_content=False)
+
+            if len(resp.history) >= self.max_redirects:
+                raise TooManyRedirects(
+                    f"Exceeded {self.max_redirects} redirects.", response=resp
+                )
+
+            # Release the connection back into the pool.
+            resp.close()
+
+            # Handle redirection without scheme (see: RFC 1808 Section 4)
+            if url.startswith("//"):
+                parsed_rurl = urlparse(resp.url)
+                url = ":".join([to_native_string(parsed_rurl.scheme), url])
+
+            # Normalize url case and attach previous fragment if needed (RFC 7231 7.1.2)
+            parsed = urlparse(url)
+            if parsed.fragment == "" and previous_fragment:
+                parsed = parsed._replace(fragment=previous_fragment)
+            elif parsed.fragment:
+                previous_fragment = parsed.fragment
+            url = parsed.geturl()
+
+            # Facilitate relative 'location' headers, as allowed by RFC 7231.
+            # (e.g. '/path/to/resource' instead of 'http://domain.tld/path/to/resource')
+            # Compliant with RFC3986, we percent encode the url.
+            if not parsed.netloc:
+                url = urljoin(resp.url, requote_uri(url))
+            else:
+                url = requote_uri(url)
+
+            prepared_request.url = to_native_string(url)
+
+            self.rebuild_method(prepared_request, resp)
+
+            # https://github.com/psf/requests/issues/1084
+            if resp.status_code not in (
+                codes.temporary_redirect,
+                codes.permanent_redirect,
+            ):
+                # https://github.com/psf/requests/issues/3490
+                purged_headers = ("Content-Length", "Content-Type", "Transfer-Encoding")
+                for header in purged_headers:
+                    prepared_request.headers.pop(header, None)
+                prepared_request.body = None
+
+            headers = prepared_request.headers
+            headers.pop("Cookie", None)
+
+            # Extract any cookies sent on the response to the cookiejar
+            # in the new request. Because we've mutated our copied prepared
+            # request, use the old one that we haven't yet touched.
+            extract_cookies_to_jar(prepared_request._cookies, req, resp.raw)
+            merge_cookies(prepared_request._cookies, self.cookies)
+            prepared_request.prepare_cookies(prepared_request._cookies)
+
+            # Rebuild auth and proxy information.
+            proxies = self.rebuild_proxies(prepared_request, proxies)
+            self.rebuild_auth(prepared_request, resp)
+
+            # A failed tell() sets `_body_position` to `object()`. This non-None
+            # value ensures `rewindable` will be True, allowing us to raise an
+            # UnrewindableBodyError, instead of hanging the connection.
+            rewindable = prepared_request._body_position is not None and (
+                "Content-Length" in headers or "Transfer-Encoding" in headers
+            )
+
+            # Attempt to rewind consumed file-like object.
+            if rewindable:
+                rewind_body(prepared_request)
+
+            # Override the original request.
+            req = prepared_request
+
+            if yield_requests:
+                yield req
+            else:
+                resp = self.send(
+                    req,
+                    stream=stream,
+                    timeout=timeout,
+                    verify=verify,
+                    cert=cert,
+                    proxies=proxies,
+                    allow_redirects=False,
+                    **adapter_kwargs,
+                )
+
+                extract_cookies_to_jar(self.cookies, prepared_request, resp.raw)
+
+                # extract redirect url, if any, for the next loop
+                url = self.get_redirect_target(resp)
+                yield resp
+
+    def rebuild_auth(self, prepared_request, response):
+        """When being redirected we may want to strip authentication from the
+        request to avoid leaking credentials. This method intelligently removes
+        and reapplies authentication where possible to avoid credential loss.
+        """
+        headers = prepared_request.headers
+        url = prepared_request.url
+
+        if "Authorization" in headers and self.should_strip_auth(
+            response.request.url, url
+        ):
+            # If we get redirected to a new host, we should strip out any
+            # authentication headers.
+            del headers["Authorization"]
+
+        # .netrc might have more auth for us on our new host.
+        new_auth = get_netrc_auth(url) if self.trust_env else None
+        if new_auth is not None:
+            prepared_request.prepare_auth(new_auth)
+
+    def rebuild_proxies(self, prepared_request, proxies):
+        """This method re-evaluates the proxy configuration by considering the
+        environment variables. If we are redirected to a URL covered by
+        NO_PROXY, we strip the proxy configuration. Otherwise, we set missing
+        proxy keys for this URL (in case they were stripped by a previous
+        redirect).
+
+        This method also replaces the Proxy-Authorization header where
+        necessary.
+
+        :rtype: dict
+        """
+        headers = prepared_request.headers
+        scheme = urlparse(prepared_request.url).scheme
+        new_proxies = resolve_proxies(prepared_request, proxies, self.trust_env)
+
+        if "Proxy-Authorization" in headers:
+            del headers["Proxy-Authorization"]
+
+        try:
+            username, password = get_auth_from_url(new_proxies[scheme])
+        except KeyError:
+            username, password = None, None
+
+        # urllib3 handles proxy authorization for us in the standard adapter.
+        # Avoid appending this to TLS tunneled requests where it may be leaked.
+        if not scheme.startswith("https") and username and password:
+            headers["Proxy-Authorization"] = _basic_auth_str(username, password)
+
+        return new_proxies
+
+    def rebuild_method(self, prepared_request, response):
+        """When being redirected we may want to change the method of the request
+        based on certain specs or browser behavior.
+        """
+        method = prepared_request.method
+
+        # https://tools.ietf.org/html/rfc7231#section-6.4.4
+        if response.status_code == codes.see_other and method != "HEAD":
+            method = "GET"
+
+        # Do what the browsers do, despite standards...
+        # First, turn 302s into GETs.
+        if response.status_code == codes.found and method != "HEAD":
+            method = "GET"
+
+        # Second, if a POST is responded to with a 301, turn it into a GET.
+        # This bizarre behaviour is explained in Issue 1704.
+        if response.status_code == codes.moved and method == "POST":
+            method = "GET"
+
+        prepared_request.method = method
+
+
+class Session(SessionRedirectMixin):
+    """A Requests session.
+
+    Provides cookie persistence, connection-pooling, and configuration.
+
+    Basic Usage::
+
+      >>> import requests
+      >>> s = requests.Session()
+      >>> s.get('https://httpbin.org/get')
+      <Response [200]>
+
+    Or as a context manager::
+
+      >>> with requests.Session() as s:
+      ...     s.get('https://httpbin.org/get')
+      <Response [200]>
+    """
+
+    __attrs__ = [
+        "headers",
+        "cookies",
+        "auth",
+        "proxies",
+        "hooks",
+        "params",
+        "verify",
+        "cert",
+        "adapters",
+        "stream",
+        "trust_env",
+        "max_redirects",
+    ]
+
+    def __init__(self):
+        #: A case-insensitive dictionary of headers to be sent on each
+        #: :class:`Request <Request>` sent from this
+        #: :class:`Session <Session>`.
+        self.headers = default_headers()
+
+        #: Default Authentication tuple or object to attach to
+        #: :class:`Request <Request>`.
+        self.auth = None
+
+        #: Dictionary mapping protocol or protocol and host to the URL of the proxy
+        #: (e.g. {'http': 'foo.bar:3128', 'http://host.name': 'foo.bar:4012'}) to
+        #: be used on each :class:`Request <Request>`.
+        self.proxies = {}
+
+        #: Event-handling hooks.
+        self.hooks = default_hooks()
+
+        #: Dictionary of querystring data to attach to each
+        #: :class:`Request <Request>`. The dictionary values may be lists for
+        #: representing multivalued query parameters.
+        self.params = {}
+
+        #: Stream response content default.
+        self.stream = False
+
+        #: SSL Verification default.
+        #: Defaults to `True`, requiring requests to verify the TLS certificate at the
+        #: remote end.
+        #: If verify is set to `False`, requests will accept any TLS certificate
+        #: presented by the server, and will ignore hostname mismatches and/or
+        #: expired certificates, which will make your application vulnerable to
+        #: man-in-the-middle (MitM) attacks.
+        #: Only set this to `False` for testing.
+        self.verify = True
+
+        #: SSL client certificate default, if String, path to ssl client
+        #: cert file (.pem). If Tuple, ('cert', 'key') pair.
+        self.cert = None
+
+        #: Maximum number of redirects allowed. If the request exceeds this
+        #: limit, a :class:`TooManyRedirects` exception is raised.
+        #: This defaults to requests.models.DEFAULT_REDIRECT_LIMIT, which is
+        #: 30.
+        self.max_redirects = DEFAULT_REDIRECT_LIMIT
+
+        #: Trust environment settings for proxy configuration, default
+        #: authentication and similar.
+        self.trust_env = True
+
+        #: A CookieJar containing all currently outstanding cookies set on this
+        #: session. By default it is a
+        #: :class:`RequestsCookieJar <requests.cookies.RequestsCookieJar>`, but
+        #: may be any other ``cookielib.CookieJar`` compatible object.
+        self.cookies = cookiejar_from_dict({})
+
+        # Default connection adapters.
+        self.adapters = OrderedDict()
+        self.mount("https://", HTTPAdapter())
+        self.mount("http://", HTTPAdapter())
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *args):
+        self.close()
+
+    def prepare_request(self, request):
+        """Constructs a :class:`PreparedRequest <PreparedRequest>` for
+        transmission and returns it. The :class:`PreparedRequest` has settings
+        merged from the :class:`Request <Request>` instance and those of the
+        :class:`Session`.
+
+        :param request: :class:`Request` instance to prepare with this
+            session's settings.
+        :rtype: requests.PreparedRequest
+        """
+        cookies = request.cookies or {}
+
+        # Bootstrap CookieJar.
+        if not isinstance(cookies, cookielib.CookieJar):
+            cookies = cookiejar_from_dict(cookies)
+
+        # Merge with session cookies
+        merged_cookies = merge_cookies(
+            merge_cookies(RequestsCookieJar(), self.cookies), cookies
+        )
+
+        # Set environment's basic authentication if not explicitly set.
+        auth = request.auth
+        if self.trust_env and not auth and not self.auth:
+            auth = get_netrc_auth(request.url)
+
+        p = PreparedRequest()
+        p.prepare(
+            method=request.method.upper(),
+            url=request.url,
+            files=request.files,
+            data=request.data,
+            json=request.json,
+            headers=merge_setting(
+                request.headers, self.headers, dict_class=CaseInsensitiveDict
+            ),
+            params=merge_setting(request.params, self.params),
+            auth=merge_setting(auth, self.auth),
+            cookies=merged_cookies,
+            hooks=merge_hooks(request.hooks, self.hooks),
+        )
+        return p
+
+    def request(
+        self,
+        method,
+        url,
+        params=None,
+        data=None,
+        headers=None,
+        cookies=None,
+        files=None,
+        auth=None,
+        timeout=None,
+        allow_redirects=True,
+        proxies=None,
+        hooks=None,
+        stream=None,
+        verify=None,
+        cert=None,
+        json=None,
+    ):
+        """Constructs a :class:`Request <Request>`, prepares it and sends it.
+        Returns :class:`Response <Response>` object.
+
+        :param method: method for the new :class:`Request` object.
+        :param url: URL for the new :class:`Request` object.
+        :param params: (optional) Dictionary or bytes to be sent in the query
+            string for the :class:`Request`.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param json: (optional) json to send in the body of the
+            :class:`Request`.
+        :param headers: (optional) Dictionary of HTTP Headers to send with the
+            :class:`Request`.
+        :param cookies: (optional) Dict or CookieJar object to send with the
+            :class:`Request`.
+        :param files: (optional) Dictionary of ``'filename': file-like-objects``
+            for multipart encoding upload.
+        :param auth: (optional) Auth tuple or callable to enable
+            Basic/Digest/Custom HTTP Auth.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple
+        :param allow_redirects: (optional) Set to True by default.
+        :type allow_redirects: bool
+        :param proxies: (optional) Dictionary mapping protocol or protocol and
+            hostname to the URL of the proxy.
+        :param hooks: (optional) Dictionary mapping hook name to one event or
+            list of events, event must be callable.
+        :param stream: (optional) whether to immediately download the response
+            content. Defaults to ``False``.
+        :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use. Defaults to ``True``. When set to
+            ``False``, requests will accept any TLS certificate presented by
+            the server, and will ignore hostname mismatches and/or expired
+            certificates, which will make your application vulnerable to
+            man-in-the-middle (MitM) attacks. Setting verify to ``False``
+            may be useful during local development or testing.
+        :param cert: (optional) if String, path to ssl client cert file (.pem).
+            If Tuple, ('cert', 'key') pair.
+        :rtype: requests.Response
+        """
+        # Create the Request.
+        req = Request(
+            method=method.upper(),
+            url=url,
+            headers=headers,
+            files=files,
+            data=data or {},
+            json=json,
+            params=params or {},
+            auth=auth,
+            cookies=cookies,
+            hooks=hooks,
+        )
+        prep = self.prepare_request(req)
+
+        proxies = proxies or {}
+
+        settings = self.merge_environment_settings(
+            prep.url, proxies, stream, verify, cert
+        )
+
+        # Send the request.
+        send_kwargs = {
+            "timeout": timeout,
+            "allow_redirects": allow_redirects,
+        }
+        send_kwargs.update(settings)
+        resp = self.send(prep, **send_kwargs)
+
+        return resp
+
+    def get(self, url, **kwargs):
+        r"""Sends a GET request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", True)
+        return self.request("GET", url, **kwargs)
+
+    def options(self, url, **kwargs):
+        r"""Sends a OPTIONS request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", True)
+        return self.request("OPTIONS", url, **kwargs)
+
+    def head(self, url, **kwargs):
+        r"""Sends a HEAD request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", False)
+        return self.request("HEAD", url, **kwargs)
+
+    def post(self, url, data=None, json=None, **kwargs):
+        r"""Sends a POST request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param json: (optional) json to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("POST", url, data=data, json=json, **kwargs)
+
+    def put(self, url, data=None, **kwargs):
+        r"""Sends a PUT request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("PUT", url, data=data, **kwargs)
+
+    def patch(self, url, data=None, **kwargs):
+        r"""Sends a PATCH request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("PATCH", url, data=data, **kwargs)
+
+    def delete(self, url, **kwargs):
+        r"""Sends a DELETE request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("DELETE", url, **kwargs)
+
+    def send(self, request, **kwargs):
+        """Send a given PreparedRequest.
+
+        :rtype: requests.Response
+        """
+        # Set defaults that the hooks can utilize to ensure they always have
+        # the correct parameters to reproduce the previous request.
+        kwargs.setdefault("stream", self.stream)
+        kwargs.setdefault("verify", self.verify)
+        kwargs.setdefault("cert", self.cert)
+        if "proxies" not in kwargs:
+            kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env)
+
+        # It's possible that users might accidentally send a Request object.
+        # Guard against that specific failure case.
+        if isinstance(request, Request):
+            raise ValueError("You can only send PreparedRequests.")
+
+        # Set up variables needed for resolve_redirects and dispatching of hooks
+        allow_redirects = kwargs.pop("allow_redirects", True)
+        stream = kwargs.get("stream")
+        hooks = request.hooks
+
+        # Get the appropriate adapter to use
+        adapter = self.get_adapter(url=request.url)
+
+        # Start time (approximately) of the request
+        start = preferred_clock()
+
+        # Send the request
+        r = adapter.send(request, **kwargs)
+
+        # Total elapsed time of the request (approximately)
+        elapsed = preferred_clock() - start
+        r.elapsed = timedelta(seconds=elapsed)
+
+        # Response manipulation hooks
+        r = dispatch_hook("response", hooks, r, **kwargs)
+
+        # Persist cookies
+        if r.history:
+            # If the hooks create history then we want those cookies too
+            for resp in r.history:
+                extract_cookies_to_jar(self.cookies, resp.request, resp.raw)
+
+        extract_cookies_to_jar(self.cookies, request, r.raw)
+
+        # Resolve redirects if allowed.
+        if allow_redirects:
+            # Redirect resolving generator.
+            gen = self.resolve_redirects(r, request, **kwargs)
+            history = [resp for resp in gen]
+        else:
+            history = []
+
+        # Shuffle things around if there's history.
+        if history:
+            # Insert the first (original) request at the start
+            history.insert(0, r)
+            # Get the last request made
+            r = history.pop()
+            r.history = history
+
+        # If redirects aren't being followed, store the response on the Request for Response.next().
+        if not allow_redirects:
+            try:
+                r._next = next(
+                    self.resolve_redirects(r, request, yield_requests=True, **kwargs)
+                )
+            except StopIteration:
+                pass
+
+        if not stream:
+            r.content
+
+        return r
+
+    def merge_environment_settings(self, url, proxies, stream, verify, cert):
+        """
+        Check the environment and merge it with some settings.
+
+        :rtype: dict
+        """
+        # Gather clues from the surrounding environment.
+        if self.trust_env:
+            # Set environment's proxies.
+            no_proxy = proxies.get("no_proxy") if proxies is not None else None
+            env_proxies = get_environ_proxies(url, no_proxy=no_proxy)
+            for k, v in env_proxies.items():
+                proxies.setdefault(k, v)
+
+            # Look for requests environment configuration
+            # and be compatible with cURL.
+            if verify is True or verify is None:
+                verify = (
+                    os.environ.get("REQUESTS_CA_BUNDLE")
+                    or os.environ.get("CURL_CA_BUNDLE")
+                    or verify
+                )
+
+        # Merge all the kwargs.
+        proxies = merge_setting(proxies, self.proxies)
+        stream = merge_setting(stream, self.stream)
+        verify = merge_setting(verify, self.verify)
+        cert = merge_setting(cert, self.cert)
+
+        return {"proxies": proxies, "stream": stream, "verify": verify, "cert": cert}
+
+    def get_adapter(self, url):
+        """
+        Returns the appropriate connection adapter for the given URL.
+
+        :rtype: requests.adapters.BaseAdapter
+        """
+        for prefix, adapter in self.adapters.items():
+            if url.lower().startswith(prefix.lower()):
+                return adapter
+
+        # Nothing matches :-/
+        raise InvalidSchema(f"No connection adapters were found for {url!r}")
+
+    def close(self):
+        """Closes all adapters and as such the session"""
+        for v in self.adapters.values():
+            v.close()
+
+    def mount(self, prefix, adapter):
+        """Registers a connection adapter to a prefix.
+
+        Adapters are sorted in descending order by prefix length.
+        """
+        self.adapters[prefix] = adapter
+        keys_to_move = [k for k in self.adapters if len(k) < len(prefix)]
+
+        for key in keys_to_move:
+            self.adapters[key] = self.adapters.pop(key)
+
+    def __getstate__(self):
+        state = {attr: getattr(self, attr, None) for attr in self.__attrs__}
+        return state
+
+    def __setstate__(self, state):
+        for attr, value in state.items():
+            setattr(self, attr, value)
+
+
+def session():
+    """
+    Returns a :class:`Session` for context-management.
+
+    .. deprecated:: 1.0.0
+
+        This method has been deprecated since version 1.0.0 and is only kept for
+        backwards compatibility. New code should use :class:`~requests.sessions.Session`
+        to create a session. This may be removed at a future date.
+
+    :rtype: Session
+    """
+    return Session()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/status_codes.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/status_codes.py
new file mode 100644
index 00000000..c7945a2f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/status_codes.py
@@ -0,0 +1,128 @@
+r"""
+The ``codes`` object defines a mapping from common names for HTTP statuses
+to their numerical codes, accessible either as attributes or as dictionary
+items.
+
+Example::
+
+    >>> import requests
+    >>> requests.codes['temporary_redirect']
+    307
+    >>> requests.codes.teapot
+    418
+    >>> requests.codes['\o/']
+    200
+
+Some codes have multiple names, and both upper- and lower-case versions of
+the names are allowed. For example, ``codes.ok``, ``codes.OK``, and
+``codes.okay`` all correspond to the HTTP status code 200.
+"""
+
+from .structures import LookupDict
+
+_codes = {
+    # Informational.
+    100: ("continue",),
+    101: ("switching_protocols",),
+    102: ("processing", "early-hints"),
+    103: ("checkpoint",),
+    122: ("uri_too_long", "request_uri_too_long"),
+    200: ("ok", "okay", "all_ok", "all_okay", "all_good", "\\o/", "✓"),
+    201: ("created",),
+    202: ("accepted",),
+    203: ("non_authoritative_info", "non_authoritative_information"),
+    204: ("no_content",),
+    205: ("reset_content", "reset"),
+    206: ("partial_content", "partial"),
+    207: ("multi_status", "multiple_status", "multi_stati", "multiple_stati"),
+    208: ("already_reported",),
+    226: ("im_used",),
+    # Redirection.
+    300: ("multiple_choices",),
+    301: ("moved_permanently", "moved", "\\o-"),
+    302: ("found",),
+    303: ("see_other", "other"),
+    304: ("not_modified",),
+    305: ("use_proxy",),
+    306: ("switch_proxy",),
+    307: ("temporary_redirect", "temporary_moved", "temporary"),
+    308: (
+        "permanent_redirect",
+        "resume_incomplete",
+        "resume",
+    ),  # "resume" and "resume_incomplete" to be removed in 3.0
+    # Client Error.
+    400: ("bad_request", "bad"),
+    401: ("unauthorized",),
+    402: ("payment_required", "payment"),
+    403: ("forbidden",),
+    404: ("not_found", "-o-"),
+    405: ("method_not_allowed", "not_allowed"),
+    406: ("not_acceptable",),
+    407: ("proxy_authentication_required", "proxy_auth", "proxy_authentication"),
+    408: ("request_timeout", "timeout"),
+    409: ("conflict",),
+    410: ("gone",),
+    411: ("length_required",),
+    412: ("precondition_failed", "precondition"),
+    413: ("request_entity_too_large", "content_too_large"),
+    414: ("request_uri_too_large", "uri_too_long"),
+    415: ("unsupported_media_type", "unsupported_media", "media_type"),
+    416: (
+        "requested_range_not_satisfiable",
+        "requested_range",
+        "range_not_satisfiable",
+    ),
+    417: ("expectation_failed",),
+    418: ("im_a_teapot", "teapot", "i_am_a_teapot"),
+    421: ("misdirected_request",),
+    422: ("unprocessable_entity", "unprocessable", "unprocessable_content"),
+    423: ("locked",),
+    424: ("failed_dependency", "dependency"),
+    425: ("unordered_collection", "unordered", "too_early"),
+    426: ("upgrade_required", "upgrade"),
+    428: ("precondition_required", "precondition"),
+    429: ("too_many_requests", "too_many"),
+    431: ("header_fields_too_large", "fields_too_large"),
+    444: ("no_response", "none"),
+    449: ("retry_with", "retry"),
+    450: ("blocked_by_windows_parental_controls", "parental_controls"),
+    451: ("unavailable_for_legal_reasons", "legal_reasons"),
+    499: ("client_closed_request",),
+    # Server Error.
+    500: ("internal_server_error", "server_error", "/o\\", "✗"),
+    501: ("not_implemented",),
+    502: ("bad_gateway",),
+    503: ("service_unavailable", "unavailable"),
+    504: ("gateway_timeout",),
+    505: ("http_version_not_supported", "http_version"),
+    506: ("variant_also_negotiates",),
+    507: ("insufficient_storage",),
+    509: ("bandwidth_limit_exceeded", "bandwidth"),
+    510: ("not_extended",),
+    511: ("network_authentication_required", "network_auth", "network_authentication"),
+}
+
+codes = LookupDict(name="status_codes")
+
+
+def _init():
+    for code, titles in _codes.items():
+        for title in titles:
+            setattr(codes, title, code)
+            if not title.startswith(("\\", "/")):
+                setattr(codes, title.upper(), code)
+
+    def doc(code):
+        names = ", ".join(f"``{n}``" for n in _codes[code])
+        return "* %d: %s" % (code, names)
+
+    global __doc__
+    __doc__ = (
+        __doc__ + "\n" + "\n".join(doc(code) for code in sorted(_codes))
+        if __doc__ is not None
+        else None
+    )
+
+
+_init()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/structures.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/structures.py
new file mode 100644
index 00000000..188e13e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/structures.py
@@ -0,0 +1,99 @@
+"""
+requests.structures
+~~~~~~~~~~~~~~~~~~~
+
+Data structures that power Requests.
+"""
+
+from collections import OrderedDict
+
+from .compat import Mapping, MutableMapping
+
+
+class CaseInsensitiveDict(MutableMapping):
+    """A case-insensitive ``dict``-like object.
+
+    Implements all methods and operations of
+    ``MutableMapping`` as well as dict's ``copy``. Also
+    provides ``lower_items``.
+
+    All keys are expected to be strings. The structure remembers the
+    case of the last key to be set, and ``iter(instance)``,
+    ``keys()``, ``items()``, ``iterkeys()``, and ``iteritems()``
+    will contain case-sensitive keys. However, querying and contains
+    testing is case insensitive::
+
+        cid = CaseInsensitiveDict()
+        cid['Accept'] = 'application/json'
+        cid['aCCEPT'] == 'application/json'  # True
+        list(cid) == ['Accept']  # True
+
+    For example, ``headers['content-encoding']`` will return the
+    value of a ``'Content-Encoding'`` response header, regardless
+    of how the header name was originally stored.
+
+    If the constructor, ``.update``, or equality comparison
+    operations are given keys that have equal ``.lower()``s, the
+    behavior is undefined.
+    """
+
+    def __init__(self, data=None, **kwargs):
+        self._store = OrderedDict()
+        if data is None:
+            data = {}
+        self.update(data, **kwargs)
+
+    def __setitem__(self, key, value):
+        # Use the lowercased key for lookups, but store the actual
+        # key alongside the value.
+        self._store[key.lower()] = (key, value)
+
+    def __getitem__(self, key):
+        return self._store[key.lower()][1]
+
+    def __delitem__(self, key):
+        del self._store[key.lower()]
+
+    def __iter__(self):
+        return (casedkey for casedkey, mappedvalue in self._store.values())
+
+    def __len__(self):
+        return len(self._store)
+
+    def lower_items(self):
+        """Like iteritems(), but with all lowercase keys."""
+        return ((lowerkey, keyval[1]) for (lowerkey, keyval) in self._store.items())
+
+    def __eq__(self, other):
+        if isinstance(other, Mapping):
+            other = CaseInsensitiveDict(other)
+        else:
+            return NotImplemented
+        # Compare insensitively
+        return dict(self.lower_items()) == dict(other.lower_items())
+
+    # Copy is required
+    def copy(self):
+        return CaseInsensitiveDict(self._store.values())
+
+    def __repr__(self):
+        return str(dict(self.items()))
+
+
+class LookupDict(dict):
+    """Dictionary lookup object."""
+
+    def __init__(self, name=None):
+        self.name = name
+        super().__init__()
+
+    def __repr__(self):
+        return f"<lookup '{self.name}'>"
+
+    def __getitem__(self, key):
+        # We allow fall-through here, so values default to None
+
+        return self.__dict__.get(key, None)
+
+    def get(self, key, default=None):
+        return self.__dict__.get(key, default)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/utils.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/utils.py
new file mode 100644
index 00000000..a35ce478
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/requests/utils.py
@@ -0,0 +1,1096 @@
+"""
+requests.utils
+~~~~~~~~~~~~~~
+
+This module provides utility functions that are used within Requests
+that are also useful for external consumption.
+"""
+
+import codecs
+import contextlib
+import io
+import os
+import re
+import socket
+import struct
+import sys
+import tempfile
+import warnings
+import zipfile
+from collections import OrderedDict
+
+from pip._vendor.urllib3.util import make_headers, parse_url
+
+from . import certs
+from .__version__ import __version__
+
+# to_native_string is unused here, but imported here for backwards compatibility
+from ._internal_utils import (  # noqa: F401
+    _HEADER_VALIDATORS_BYTE,
+    _HEADER_VALIDATORS_STR,
+    HEADER_VALIDATORS,
+    to_native_string,
+)
+from .compat import (
+    Mapping,
+    basestring,
+    bytes,
+    getproxies,
+    getproxies_environment,
+    integer_types,
+)
+from .compat import parse_http_list as _parse_list_header
+from .compat import (
+    proxy_bypass,
+    proxy_bypass_environment,
+    quote,
+    str,
+    unquote,
+    urlparse,
+    urlunparse,
+)
+from .cookies import cookiejar_from_dict
+from .exceptions import (
+    FileModeWarning,
+    InvalidHeader,
+    InvalidURL,
+    UnrewindableBodyError,
+)
+from .structures import CaseInsensitiveDict
+
+NETRC_FILES = (".netrc", "_netrc")
+
+DEFAULT_CA_BUNDLE_PATH = certs.where()
+
+DEFAULT_PORTS = {"http": 80, "https": 443}
+
+# Ensure that ', ' is used to preserve previous delimiter behavior.
+DEFAULT_ACCEPT_ENCODING = ", ".join(
+    re.split(r",\s*", make_headers(accept_encoding=True)["accept-encoding"])
+)
+
+
+if sys.platform == "win32":
+    # provide a proxy_bypass version on Windows without DNS lookups
+
+    def proxy_bypass_registry(host):
+        try:
+            import winreg
+        except ImportError:
+            return False
+
+        try:
+            internetSettings = winreg.OpenKey(
+                winreg.HKEY_CURRENT_USER,
+                r"Software\Microsoft\Windows\CurrentVersion\Internet Settings",
+            )
+            # ProxyEnable could be REG_SZ or REG_DWORD, normalizing it
+            proxyEnable = int(winreg.QueryValueEx(internetSettings, "ProxyEnable")[0])
+            # ProxyOverride is almost always a string
+            proxyOverride = winreg.QueryValueEx(internetSettings, "ProxyOverride")[0]
+        except (OSError, ValueError):
+            return False
+        if not proxyEnable or not proxyOverride:
+            return False
+
+        # make a check value list from the registry entry: replace the
+        # '<local>' string by the localhost entry and the corresponding
+        # canonical entry.
+        proxyOverride = proxyOverride.split(";")
+        # filter out empty strings to avoid re.match return true in the following code.
+        proxyOverride = filter(None, proxyOverride)
+        # now check if we match one of the registry values.
+        for test in proxyOverride:
+            if test == "<local>":
+                if "." not in host:
+                    return True
+            test = test.replace(".", r"\.")  # mask dots
+            test = test.replace("*", r".*")  # change glob sequence
+            test = test.replace("?", r".")  # change glob char
+            if re.match(test, host, re.I):
+                return True
+        return False
+
+    def proxy_bypass(host):  # noqa
+        """Return True, if the host should be bypassed.
+
+        Checks proxy settings gathered from the environment, if specified,
+        or the registry.
+        """
+        if getproxies_environment():
+            return proxy_bypass_environment(host)
+        else:
+            return proxy_bypass_registry(host)
+
+
+def dict_to_sequence(d):
+    """Returns an internal sequence dictionary update."""
+
+    if hasattr(d, "items"):
+        d = d.items()
+
+    return d
+
+
+def super_len(o):
+    total_length = None
+    current_position = 0
+
+    if isinstance(o, str):
+        o = o.encode("utf-8")
+
+    if hasattr(o, "__len__"):
+        total_length = len(o)
+
+    elif hasattr(o, "len"):
+        total_length = o.len
+
+    elif hasattr(o, "fileno"):
+        try:
+            fileno = o.fileno()
+        except (io.UnsupportedOperation, AttributeError):
+            # AttributeError is a surprising exception, seeing as how we've just checked
+            # that `hasattr(o, 'fileno')`.  It happens for objects obtained via
+            # `Tarfile.extractfile()`, per issue 5229.
+            pass
+        else:
+            total_length = os.fstat(fileno).st_size
+
+            # Having used fstat to determine the file length, we need to
+            # confirm that this file was opened up in binary mode.
+            if "b" not in o.mode:
+                warnings.warn(
+                    (
+                        "Requests has determined the content-length for this "
+                        "request using the binary size of the file: however, the "
+                        "file has been opened in text mode (i.e. without the 'b' "
+                        "flag in the mode). This may lead to an incorrect "
+                        "content-length. In Requests 3.0, support will be removed "
+                        "for files in text mode."
+                    ),
+                    FileModeWarning,
+                )
+
+    if hasattr(o, "tell"):
+        try:
+            current_position = o.tell()
+        except OSError:
+            # This can happen in some weird situations, such as when the file
+            # is actually a special file descriptor like stdin. In this
+            # instance, we don't know what the length is, so set it to zero and
+            # let requests chunk it instead.
+            if total_length is not None:
+                current_position = total_length
+        else:
+            if hasattr(o, "seek") and total_length is None:
+                # StringIO and BytesIO have seek but no usable fileno
+                try:
+                    # seek to end of file
+                    o.seek(0, 2)
+                    total_length = o.tell()
+
+                    # seek back to current position to support
+                    # partially read file-like objects
+                    o.seek(current_position or 0)
+                except OSError:
+                    total_length = 0
+
+    if total_length is None:
+        total_length = 0
+
+    return max(0, total_length - current_position)
+
+
+def get_netrc_auth(url, raise_errors=False):
+    """Returns the Requests tuple auth for a given url from netrc."""
+
+    netrc_file = os.environ.get("NETRC")
+    if netrc_file is not None:
+        netrc_locations = (netrc_file,)
+    else:
+        netrc_locations = (f"~/{f}" for f in NETRC_FILES)
+
+    try:
+        from netrc import NetrcParseError, netrc
+
+        netrc_path = None
+
+        for f in netrc_locations:
+            try:
+                loc = os.path.expanduser(f)
+            except KeyError:
+                # os.path.expanduser can fail when $HOME is undefined and
+                # getpwuid fails. See https://bugs.python.org/issue20164 &
+                # https://github.com/psf/requests/issues/1846
+                return
+
+            if os.path.exists(loc):
+                netrc_path = loc
+                break
+
+        # Abort early if there isn't one.
+        if netrc_path is None:
+            return
+
+        ri = urlparse(url)
+
+        # Strip port numbers from netloc. This weird `if...encode`` dance is
+        # used for Python 3.2, which doesn't support unicode literals.
+        splitstr = b":"
+        if isinstance(url, str):
+            splitstr = splitstr.decode("ascii")
+        host = ri.netloc.split(splitstr)[0]
+
+        try:
+            _netrc = netrc(netrc_path).authenticators(host)
+            if _netrc:
+                # Return with login / password
+                login_i = 0 if _netrc[0] else 1
+                return (_netrc[login_i], _netrc[2])
+        except (NetrcParseError, OSError):
+            # If there was a parsing error or a permissions issue reading the file,
+            # we'll just skip netrc auth unless explicitly asked to raise errors.
+            if raise_errors:
+                raise
+
+    # App Engine hackiness.
+    except (ImportError, AttributeError):
+        pass
+
+
+def guess_filename(obj):
+    """Tries to guess the filename of the given object."""
+    name = getattr(obj, "name", None)
+    if name and isinstance(name, basestring) and name[0] != "<" and name[-1] != ">":
+        return os.path.basename(name)
+
+
+def extract_zipped_paths(path):
+    """Replace nonexistent paths that look like they refer to a member of a zip
+    archive with the location of an extracted copy of the target, or else
+    just return the provided path unchanged.
+    """
+    if os.path.exists(path):
+        # this is already a valid path, no need to do anything further
+        return path
+
+    # find the first valid part of the provided path and treat that as a zip archive
+    # assume the rest of the path is the name of a member in the archive
+    archive, member = os.path.split(path)
+    while archive and not os.path.exists(archive):
+        archive, prefix = os.path.split(archive)
+        if not prefix:
+            # If we don't check for an empty prefix after the split (in other words, archive remains unchanged after the split),
+            # we _can_ end up in an infinite loop on a rare corner case affecting a small number of users
+            break
+        member = "/".join([prefix, member])
+
+    if not zipfile.is_zipfile(archive):
+        return path
+
+    zip_file = zipfile.ZipFile(archive)
+    if member not in zip_file.namelist():
+        return path
+
+    # we have a valid zip archive and a valid member of that archive
+    tmp = tempfile.gettempdir()
+    extracted_path = os.path.join(tmp, member.split("/")[-1])
+    if not os.path.exists(extracted_path):
+        # use read + write to avoid the creating nested folders, we only want the file, avoids mkdir racing condition
+        with atomic_open(extracted_path) as file_handler:
+            file_handler.write(zip_file.read(member))
+    return extracted_path
+
+
+@contextlib.contextmanager
+def atomic_open(filename):
+    """Write a file to the disk in an atomic fashion"""
+    tmp_descriptor, tmp_name = tempfile.mkstemp(dir=os.path.dirname(filename))
+    try:
+        with os.fdopen(tmp_descriptor, "wb") as tmp_handler:
+            yield tmp_handler
+        os.replace(tmp_name, filename)
+    except BaseException:
+        os.remove(tmp_name)
+        raise
+
+
+def from_key_val_list(value):
+    """Take an object and test to see if it can be represented as a
+    dictionary. Unless it can not be represented as such, return an
+    OrderedDict, e.g.,
+
+    ::
+
+        >>> from_key_val_list([('key', 'val')])
+        OrderedDict([('key', 'val')])
+        >>> from_key_val_list('string')
+        Traceback (most recent call last):
+        ...
+        ValueError: cannot encode objects that are not 2-tuples
+        >>> from_key_val_list({'key': 'val'})
+        OrderedDict([('key', 'val')])
+
+    :rtype: OrderedDict
+    """
+    if value is None:
+        return None
+
+    if isinstance(value, (str, bytes, bool, int)):
+        raise ValueError("cannot encode objects that are not 2-tuples")
+
+    return OrderedDict(value)
+
+
+def to_key_val_list(value):
+    """Take an object and test to see if it can be represented as a
+    dictionary. If it can be, return a list of tuples, e.g.,
+
+    ::
+
+        >>> to_key_val_list([('key', 'val')])
+        [('key', 'val')]
+        >>> to_key_val_list({'key': 'val'})
+        [('key', 'val')]
+        >>> to_key_val_list('string')
+        Traceback (most recent call last):
+        ...
+        ValueError: cannot encode objects that are not 2-tuples
+
+    :rtype: list
+    """
+    if value is None:
+        return None
+
+    if isinstance(value, (str, bytes, bool, int)):
+        raise ValueError("cannot encode objects that are not 2-tuples")
+
+    if isinstance(value, Mapping):
+        value = value.items()
+
+    return list(value)
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def parse_list_header(value):
+    """Parse lists as described by RFC 2068 Section 2.
+
+    In particular, parse comma-separated lists where the elements of
+    the list may include quoted-strings.  A quoted-string could
+    contain a comma.  A non-quoted string could have quotes in the
+    middle.  Quotes are removed automatically after parsing.
+
+    It basically works like :func:`parse_set_header` just that items
+    may appear multiple times and case sensitivity is preserved.
+
+    The return value is a standard :class:`list`:
+
+    >>> parse_list_header('token, "quoted value"')
+    ['token', 'quoted value']
+
+    To create a header from the :class:`list` again, use the
+    :func:`dump_header` function.
+
+    :param value: a string with a list header.
+    :return: :class:`list`
+    :rtype: list
+    """
+    result = []
+    for item in _parse_list_header(value):
+        if item[:1] == item[-1:] == '"':
+            item = unquote_header_value(item[1:-1])
+        result.append(item)
+    return result
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def parse_dict_header(value):
+    """Parse lists of key, value pairs as described by RFC 2068 Section 2 and
+    convert them into a python dict:
+
+    >>> d = parse_dict_header('foo="is a fish", bar="as well"')
+    >>> type(d) is dict
+    True
+    >>> sorted(d.items())
+    [('bar', 'as well'), ('foo', 'is a fish')]
+
+    If there is no value for a key it will be `None`:
+
+    >>> parse_dict_header('key_without_value')
+    {'key_without_value': None}
+
+    To create a header from the :class:`dict` again, use the
+    :func:`dump_header` function.
+
+    :param value: a string with a dict header.
+    :return: :class:`dict`
+    :rtype: dict
+    """
+    result = {}
+    for item in _parse_list_header(value):
+        if "=" not in item:
+            result[item] = None
+            continue
+        name, value = item.split("=", 1)
+        if value[:1] == value[-1:] == '"':
+            value = unquote_header_value(value[1:-1])
+        result[name] = value
+    return result
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def unquote_header_value(value, is_filename=False):
+    r"""Unquotes a header value.  (Reversal of :func:`quote_header_value`).
+    This does not use the real unquoting but what browsers are actually
+    using for quoting.
+
+    :param value: the header value to unquote.
+    :rtype: str
+    """
+    if value and value[0] == value[-1] == '"':
+        # this is not the real unquoting, but fixing this so that the
+        # RFC is met will result in bugs with internet explorer and
+        # probably some other browsers as well.  IE for example is
+        # uploading files with "C:\foo\bar.txt" as filename
+        value = value[1:-1]
+
+        # if this is a filename and the starting characters look like
+        # a UNC path, then just return the value without quotes.  Using the
+        # replace sequence below on a UNC path has the effect of turning
+        # the leading double slash into a single slash and then
+        # _fix_ie_filename() doesn't work correctly.  See #458.
+        if not is_filename or value[:2] != "\\\\":
+            return value.replace("\\\\", "\\").replace('\\"', '"')
+    return value
+
+
+def dict_from_cookiejar(cj):
+    """Returns a key/value dictionary from a CookieJar.
+
+    :param cj: CookieJar object to extract cookies from.
+    :rtype: dict
+    """
+
+    cookie_dict = {cookie.name: cookie.value for cookie in cj}
+    return cookie_dict
+
+
+def add_dict_to_cookiejar(cj, cookie_dict):
+    """Returns a CookieJar from a key/value dictionary.
+
+    :param cj: CookieJar to insert cookies into.
+    :param cookie_dict: Dict of key/values to insert into CookieJar.
+    :rtype: CookieJar
+    """
+
+    return cookiejar_from_dict(cookie_dict, cj)
+
+
+def get_encodings_from_content(content):
+    """Returns encodings from given content string.
+
+    :param content: bytestring to extract encodings from.
+    """
+    warnings.warn(
+        (
+            "In requests 3.0, get_encodings_from_content will be removed. For "
+            "more information, please see the discussion on issue #2266. (This"
+            " warning should only appear once.)"
+        ),
+        DeprecationWarning,
+    )
+
+    charset_re = re.compile(r'<meta.*?charset=["\']*(.+?)["\'>]', flags=re.I)
+    pragma_re = re.compile(r'<meta.*?content=["\']*;?charset=(.+?)["\'>]', flags=re.I)
+    xml_re = re.compile(r'^<\?xml.*?encoding=["\']*(.+?)["\'>]')
+
+    return (
+        charset_re.findall(content)
+        + pragma_re.findall(content)
+        + xml_re.findall(content)
+    )
+
+
+def _parse_content_type_header(header):
+    """Returns content type and parameters from given header
+
+    :param header: string
+    :return: tuple containing content type and dictionary of
+         parameters
+    """
+
+    tokens = header.split(";")
+    content_type, params = tokens[0].strip(), tokens[1:]
+    params_dict = {}
+    items_to_strip = "\"' "
+
+    for param in params:
+        param = param.strip()
+        if param:
+            key, value = param, True
+            index_of_equals = param.find("=")
+            if index_of_equals != -1:
+                key = param[:index_of_equals].strip(items_to_strip)
+                value = param[index_of_equals + 1 :].strip(items_to_strip)
+            params_dict[key.lower()] = value
+    return content_type, params_dict
+
+
+def get_encoding_from_headers(headers):
+    """Returns encodings from given HTTP Header Dict.
+
+    :param headers: dictionary to extract encoding from.
+    :rtype: str
+    """
+
+    content_type = headers.get("content-type")
+
+    if not content_type:
+        return None
+
+    content_type, params = _parse_content_type_header(content_type)
+
+    if "charset" in params:
+        return params["charset"].strip("'\"")
+
+    if "text" in content_type:
+        return "ISO-8859-1"
+
+    if "application/json" in content_type:
+        # Assume UTF-8 based on RFC 4627: https://www.ietf.org/rfc/rfc4627.txt since the charset was unset
+        return "utf-8"
+
+
+def stream_decode_response_unicode(iterator, r):
+    """Stream decodes an iterator."""
+
+    if r.encoding is None:
+        yield from iterator
+        return
+
+    decoder = codecs.getincrementaldecoder(r.encoding)(errors="replace")
+    for chunk in iterator:
+        rv = decoder.decode(chunk)
+        if rv:
+            yield rv
+    rv = decoder.decode(b"", final=True)
+    if rv:
+        yield rv
+
+
+def iter_slices(string, slice_length):
+    """Iterate over slices of a string."""
+    pos = 0
+    if slice_length is None or slice_length <= 0:
+        slice_length = len(string)
+    while pos < len(string):
+        yield string[pos : pos + slice_length]
+        pos += slice_length
+
+
+def get_unicode_from_response(r):
+    """Returns the requested content back in unicode.
+
+    :param r: Response object to get unicode content from.
+
+    Tried:
+
+    1. charset from content-type
+    2. fall back and replace all unicode characters
+
+    :rtype: str
+    """
+    warnings.warn(
+        (
+            "In requests 3.0, get_unicode_from_response will be removed. For "
+            "more information, please see the discussion on issue #2266. (This"
+            " warning should only appear once.)"
+        ),
+        DeprecationWarning,
+    )
+
+    tried_encodings = []
+
+    # Try charset from content-type
+    encoding = get_encoding_from_headers(r.headers)
+
+    if encoding:
+        try:
+            return str(r.content, encoding)
+        except UnicodeError:
+            tried_encodings.append(encoding)
+
+    # Fall back:
+    try:
+        return str(r.content, encoding, errors="replace")
+    except TypeError:
+        return r.content
+
+
+# The unreserved URI characters (RFC 3986)
+UNRESERVED_SET = frozenset(
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz" + "0123456789-._~"
+)
+
+
+def unquote_unreserved(uri):
+    """Un-escape any percent-escape sequences in a URI that are unreserved
+    characters. This leaves all reserved, illegal and non-ASCII bytes encoded.
+
+    :rtype: str
+    """
+    parts = uri.split("%")
+    for i in range(1, len(parts)):
+        h = parts[i][0:2]
+        if len(h) == 2 and h.isalnum():
+            try:
+                c = chr(int(h, 16))
+            except ValueError:
+                raise InvalidURL(f"Invalid percent-escape sequence: '{h}'")
+
+            if c in UNRESERVED_SET:
+                parts[i] = c + parts[i][2:]
+            else:
+                parts[i] = f"%{parts[i]}"
+        else:
+            parts[i] = f"%{parts[i]}"
+    return "".join(parts)
+
+
+def requote_uri(uri):
+    """Re-quote the given URI.
+
+    This function passes the given URI through an unquote/quote cycle to
+    ensure that it is fully and consistently quoted.
+
+    :rtype: str
+    """
+    safe_with_percent = "!#$%&'()*+,/:;=?@[]~"
+    safe_without_percent = "!#$&'()*+,/:;=?@[]~"
+    try:
+        # Unquote only the unreserved characters
+        # Then quote only illegal characters (do not quote reserved,
+        # unreserved, or '%')
+        return quote(unquote_unreserved(uri), safe=safe_with_percent)
+    except InvalidURL:
+        # We couldn't unquote the given URI, so let's try quoting it, but
+        # there may be unquoted '%'s in the URI. We need to make sure they're
+        # properly quoted so they do not cause issues elsewhere.
+        return quote(uri, safe=safe_without_percent)
+
+
+def address_in_network(ip, net):
+    """This function allows you to check if an IP belongs to a network subnet
+
+    Example: returns True if ip = 192.168.1.1 and net = 192.168.1.0/24
+             returns False if ip = 192.168.1.1 and net = 192.168.100.0/24
+
+    :rtype: bool
+    """
+    ipaddr = struct.unpack("=L", socket.inet_aton(ip))[0]
+    netaddr, bits = net.split("/")
+    netmask = struct.unpack("=L", socket.inet_aton(dotted_netmask(int(bits))))[0]
+    network = struct.unpack("=L", socket.inet_aton(netaddr))[0] & netmask
+    return (ipaddr & netmask) == (network & netmask)
+
+
+def dotted_netmask(mask):
+    """Converts mask from /xx format to xxx.xxx.xxx.xxx
+
+    Example: if mask is 24 function returns 255.255.255.0
+
+    :rtype: str
+    """
+    bits = 0xFFFFFFFF ^ (1 << 32 - mask) - 1
+    return socket.inet_ntoa(struct.pack(">I", bits))
+
+
+def is_ipv4_address(string_ip):
+    """
+    :rtype: bool
+    """
+    try:
+        socket.inet_aton(string_ip)
+    except OSError:
+        return False
+    return True
+
+
+def is_valid_cidr(string_network):
+    """
+    Very simple check of the cidr format in no_proxy variable.
+
+    :rtype: bool
+    """
+    if string_network.count("/") == 1:
+        try:
+            mask = int(string_network.split("/")[1])
+        except ValueError:
+            return False
+
+        if mask < 1 or mask > 32:
+            return False
+
+        try:
+            socket.inet_aton(string_network.split("/")[0])
+        except OSError:
+            return False
+    else:
+        return False
+    return True
+
+
+@contextlib.contextmanager
+def set_environ(env_name, value):
+    """Set the environment variable 'env_name' to 'value'
+
+    Save previous value, yield, and then restore the previous value stored in
+    the environment variable 'env_name'.
+
+    If 'value' is None, do nothing"""
+    value_changed = value is not None
+    if value_changed:
+        old_value = os.environ.get(env_name)
+        os.environ[env_name] = value
+    try:
+        yield
+    finally:
+        if value_changed:
+            if old_value is None:
+                del os.environ[env_name]
+            else:
+                os.environ[env_name] = old_value
+
+
+def should_bypass_proxies(url, no_proxy):
+    """
+    Returns whether we should bypass proxies or not.
+
+    :rtype: bool
+    """
+
+    # Prioritize lowercase environment variables over uppercase
+    # to keep a consistent behaviour with other http projects (curl, wget).
+    def get_proxy(key):
+        return os.environ.get(key) or os.environ.get(key.upper())
+
+    # First check whether no_proxy is defined. If it is, check that the URL
+    # we're getting isn't in the no_proxy list.
+    no_proxy_arg = no_proxy
+    if no_proxy is None:
+        no_proxy = get_proxy("no_proxy")
+    parsed = urlparse(url)
+
+    if parsed.hostname is None:
+        # URLs don't always have hostnames, e.g. file:/// urls.
+        return True
+
+    if no_proxy:
+        # We need to check whether we match here. We need to see if we match
+        # the end of the hostname, both with and without the port.
+        no_proxy = (host for host in no_proxy.replace(" ", "").split(",") if host)
+
+        if is_ipv4_address(parsed.hostname):
+            for proxy_ip in no_proxy:
+                if is_valid_cidr(proxy_ip):
+                    if address_in_network(parsed.hostname, proxy_ip):
+                        return True
+                elif parsed.hostname == proxy_ip:
+                    # If no_proxy ip was defined in plain IP notation instead of cidr notation &
+                    # matches the IP of the index
+                    return True
+        else:
+            host_with_port = parsed.hostname
+            if parsed.port:
+                host_with_port += f":{parsed.port}"
+
+            for host in no_proxy:
+                if parsed.hostname.endswith(host) or host_with_port.endswith(host):
+                    # The URL does match something in no_proxy, so we don't want
+                    # to apply the proxies on this URL.
+                    return True
+
+    with set_environ("no_proxy", no_proxy_arg):
+        # parsed.hostname can be `None` in cases such as a file URI.
+        try:
+            bypass = proxy_bypass(parsed.hostname)
+        except (TypeError, socket.gaierror):
+            bypass = False
+
+    if bypass:
+        return True
+
+    return False
+
+
+def get_environ_proxies(url, no_proxy=None):
+    """
+    Return a dict of environment proxies.
+
+    :rtype: dict
+    """
+    if should_bypass_proxies(url, no_proxy=no_proxy):
+        return {}
+    else:
+        return getproxies()
+
+
+def select_proxy(url, proxies):
+    """Select a proxy for the url, if applicable.
+
+    :param url: The url being for the request
+    :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs
+    """
+    proxies = proxies or {}
+    urlparts = urlparse(url)
+    if urlparts.hostname is None:
+        return proxies.get(urlparts.scheme, proxies.get("all"))
+
+    proxy_keys = [
+        urlparts.scheme + "://" + urlparts.hostname,
+        urlparts.scheme,
+        "all://" + urlparts.hostname,
+        "all",
+    ]
+    proxy = None
+    for proxy_key in proxy_keys:
+        if proxy_key in proxies:
+            proxy = proxies[proxy_key]
+            break
+
+    return proxy
+
+
+def resolve_proxies(request, proxies, trust_env=True):
+    """This method takes proxy information from a request and configuration
+    input to resolve a mapping of target proxies. This will consider settings
+    such as NO_PROXY to strip proxy configurations.
+
+    :param request: Request or PreparedRequest
+    :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs
+    :param trust_env: Boolean declaring whether to trust environment configs
+
+    :rtype: dict
+    """
+    proxies = proxies if proxies is not None else {}
+    url = request.url
+    scheme = urlparse(url).scheme
+    no_proxy = proxies.get("no_proxy")
+    new_proxies = proxies.copy()
+
+    if trust_env and not should_bypass_proxies(url, no_proxy=no_proxy):
+        environ_proxies = get_environ_proxies(url, no_proxy=no_proxy)
+
+        proxy = environ_proxies.get(scheme, environ_proxies.get("all"))
+
+        if proxy:
+            new_proxies.setdefault(scheme, proxy)
+    return new_proxies
+
+
+def default_user_agent(name="python-requests"):
+    """
+    Return a string representing the default user agent.
+
+    :rtype: str
+    """
+    return f"{name}/{__version__}"
+
+
+def default_headers():
+    """
+    :rtype: requests.structures.CaseInsensitiveDict
+    """
+    return CaseInsensitiveDict(
+        {
+            "User-Agent": default_user_agent(),
+            "Accept-Encoding": DEFAULT_ACCEPT_ENCODING,
+            "Accept": "*/*",
+            "Connection": "keep-alive",
+        }
+    )
+
+
+def parse_header_links(value):
+    """Return a list of parsed link headers proxies.
+
+    i.e. Link: <http:/.../front.jpeg>; rel=front; type="image/jpeg",<http://.../back.jpeg>; rel=back;type="image/jpeg"
+
+    :rtype: list
+    """
+
+    links = []
+
+    replace_chars = " '\""
+
+    value = value.strip(replace_chars)
+    if not value:
+        return links
+
+    for val in re.split(", *<", value):
+        try:
+            url, params = val.split(";", 1)
+        except ValueError:
+            url, params = val, ""
+
+        link = {"url": url.strip("<> '\"")}
+
+        for param in params.split(";"):
+            try:
+                key, value = param.split("=")
+            except ValueError:
+                break
+
+            link[key.strip(replace_chars)] = value.strip(replace_chars)
+
+        links.append(link)
+
+    return links
+
+
+# Null bytes; no need to recreate these on each call to guess_json_utf
+_null = "\x00".encode("ascii")  # encoding to ASCII for Python 3
+_null2 = _null * 2
+_null3 = _null * 3
+
+
+def guess_json_utf(data):
+    """
+    :rtype: str
+    """
+    # JSON always starts with two ASCII characters, so detection is as
+    # easy as counting the nulls and from their location and count
+    # determine the encoding. Also detect a BOM, if present.
+    sample = data[:4]
+    if sample in (codecs.BOM_UTF32_LE, codecs.BOM_UTF32_BE):
+        return "utf-32"  # BOM included
+    if sample[:3] == codecs.BOM_UTF8:
+        return "utf-8-sig"  # BOM included, MS style (discouraged)
+    if sample[:2] in (codecs.BOM_UTF16_LE, codecs.BOM_UTF16_BE):
+        return "utf-16"  # BOM included
+    nullcount = sample.count(_null)
+    if nullcount == 0:
+        return "utf-8"
+    if nullcount == 2:
+        if sample[::2] == _null2:  # 1st and 3rd are null
+            return "utf-16-be"
+        if sample[1::2] == _null2:  # 2nd and 4th are null
+            return "utf-16-le"
+        # Did not detect 2 valid UTF-16 ascii-range characters
+    if nullcount == 3:
+        if sample[:3] == _null3:
+            return "utf-32-be"
+        if sample[1:] == _null3:
+            return "utf-32-le"
+        # Did not detect a valid UTF-32 ascii-range character
+    return None
+
+
+def prepend_scheme_if_needed(url, new_scheme):
+    """Given a URL that may or may not have a scheme, prepend the given scheme.
+    Does not replace a present scheme with the one provided as an argument.
+
+    :rtype: str
+    """
+    parsed = parse_url(url)
+    scheme, auth, host, port, path, query, fragment = parsed
+
+    # A defect in urlparse determines that there isn't a netloc present in some
+    # urls. We previously assumed parsing was overly cautious, and swapped the
+    # netloc and path. Due to a lack of tests on the original defect, this is
+    # maintained with parse_url for backwards compatibility.
+    netloc = parsed.netloc
+    if not netloc:
+        netloc, path = path, netloc
+
+    if auth:
+        # parse_url doesn't provide the netloc with auth
+        # so we'll add it ourselves.
+        netloc = "@".join([auth, netloc])
+    if scheme is None:
+        scheme = new_scheme
+    if path is None:
+        path = ""
+
+    return urlunparse((scheme, netloc, path, "", query, fragment))
+
+
+def get_auth_from_url(url):
+    """Given a url with authentication components, extract them into a tuple of
+    username,password.
+
+    :rtype: (str,str)
+    """
+    parsed = urlparse(url)
+
+    try:
+        auth = (unquote(parsed.username), unquote(parsed.password))
+    except (AttributeError, TypeError):
+        auth = ("", "")
+
+    return auth
+
+
+def check_header_validity(header):
+    """Verifies that header parts don't contain leading whitespace
+    reserved characters, or return characters.
+
+    :param header: tuple, in the format (name, value).
+    """
+    name, value = header
+    _validate_header_part(header, name, 0)
+    _validate_header_part(header, value, 1)
+
+
+def _validate_header_part(header, header_part, header_validator_index):
+    if isinstance(header_part, str):
+        validator = _HEADER_VALIDATORS_STR[header_validator_index]
+    elif isinstance(header_part, bytes):
+        validator = _HEADER_VALIDATORS_BYTE[header_validator_index]
+    else:
+        raise InvalidHeader(
+            f"Header part ({header_part!r}) from {header} "
+            f"must be of type str or bytes, not {type(header_part)}"
+        )
+
+    if not validator.match(header_part):
+        header_kind = "name" if header_validator_index == 0 else "value"
+        raise InvalidHeader(
+            f"Invalid leading whitespace, reserved character(s), or return "
+            f"character(s) in header {header_kind}: {header_part!r}"
+        )
+
+
+def urldefragauth(url):
+    """
+    Given a url remove the fragment and the authentication part.
+
+    :rtype: str
+    """
+    scheme, netloc, path, params, query, fragment = urlparse(url)
+
+    # see func:`prepend_scheme_if_needed`
+    if not netloc:
+        netloc, path = path, netloc
+
+    netloc = netloc.rsplit("@", 1)[-1]
+
+    return urlunparse((scheme, netloc, path, params, query, ""))
+
+
+def rewind_body(prepared_request):
+    """Move file pointer back to its recorded starting position
+    so it can be read again on redirect.
+    """
+    body_seek = getattr(prepared_request.body, "seek", None)
+    if body_seek is not None and isinstance(
+        prepared_request._body_position, integer_types
+    ):
+        try:
+            body_seek(prepared_request._body_position)
+        except OSError:
+            raise UnrewindableBodyError(
+                "An error occurred when rewinding request body for redirect."
+            )
+    else:
+        raise UnrewindableBodyError("Unable to rewind request body for redirect.")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__init__.py
new file mode 100644
index 00000000..d92acc7b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__init__.py
@@ -0,0 +1,26 @@
+__all__ = [
+    "__version__",
+    "AbstractProvider",
+    "AbstractResolver",
+    "BaseReporter",
+    "InconsistentCandidate",
+    "Resolver",
+    "RequirementsConflicted",
+    "ResolutionError",
+    "ResolutionImpossible",
+    "ResolutionTooDeep",
+]
+
+__version__ = "1.0.1"
+
+
+from .providers import AbstractProvider, AbstractResolver
+from .reporters import BaseReporter
+from .resolvers import (
+    InconsistentCandidate,
+    RequirementsConflicted,
+    ResolutionError,
+    ResolutionImpossible,
+    ResolutionTooDeep,
+    Resolver,
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..f9903f66
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/providers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/providers.cpython-312.pyc
new file mode 100644
index 00000000..b8be4ad8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/providers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/reporters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/reporters.cpython-312.pyc
new file mode 100644
index 00000000..2647ee03
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/reporters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/resolvers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/resolvers.cpython-312.pyc
new file mode 100644
index 00000000..d33413fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/resolvers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/structs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/structs.cpython-312.pyc
new file mode 100644
index 00000000..39a8b0e5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/__pycache__/structs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..30fcc8e7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/collections_abc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/collections_abc.cpython-312.pyc
new file mode 100644
index 00000000..602a72de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/__pycache__/collections_abc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/collections_abc.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/collections_abc.py
new file mode 100644
index 00000000..1becc509
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/compat/collections_abc.py
@@ -0,0 +1,6 @@
+__all__ = ["Mapping", "Sequence"]
+
+try:
+    from collections.abc import Mapping, Sequence
+except ImportError:
+    from collections import Mapping, Sequence
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/providers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/providers.py
new file mode 100644
index 00000000..e99d87ee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/providers.py
@@ -0,0 +1,133 @@
+class AbstractProvider(object):
+    """Delegate class to provide the required interface for the resolver."""
+
+    def identify(self, requirement_or_candidate):
+        """Given a requirement, return an identifier for it.
+
+        This is used to identify a requirement, e.g. whether two requirements
+        should have their specifier parts merged.
+        """
+        raise NotImplementedError
+
+    def get_preference(
+        self,
+        identifier,
+        resolutions,
+        candidates,
+        information,
+        backtrack_causes,
+    ):
+        """Produce a sort key for given requirement based on preference.
+
+        The preference is defined as "I think this requirement should be
+        resolved first". The lower the return value is, the more preferred
+        this group of arguments is.
+
+        :param identifier: An identifier as returned by ``identify()``. This
+            identifies the dependency matches which should be returned.
+        :param resolutions: Mapping of candidates currently pinned by the
+            resolver. Each key is an identifier, and the value is a candidate.
+            The candidate may conflict with requirements from ``information``.
+        :param candidates: Mapping of each dependency's possible candidates.
+            Each value is an iterator of candidates.
+        :param information: Mapping of requirement information of each package.
+            Each value is an iterator of *requirement information*.
+        :param backtrack_causes: Sequence of requirement information that were
+            the requirements that caused the resolver to most recently backtrack.
+
+        A *requirement information* instance is a named tuple with two members:
+
+        * ``requirement`` specifies a requirement contributing to the current
+          list of candidates.
+        * ``parent`` specifies the candidate that provides (depended on) the
+          requirement, or ``None`` to indicate a root requirement.
+
+        The preference could depend on various issues, including (not
+        necessarily in this order):
+
+        * Is this package pinned in the current resolution result?
+        * How relaxed is the requirement? Stricter ones should probably be
+          worked on first? (I don't know, actually.)
+        * How many possibilities are there to satisfy this requirement? Those
+          with few left should likely be worked on first, I guess?
+        * Are there any known conflicts for this requirement? We should
+          probably work on those with the most known conflicts.
+
+        A sortable value should be returned (this will be used as the ``key``
+        parameter of the built-in sorting function). The smaller the value is,
+        the more preferred this requirement is (i.e. the sorting function
+        is called with ``reverse=False``).
+        """
+        raise NotImplementedError
+
+    def find_matches(self, identifier, requirements, incompatibilities):
+        """Find all possible candidates that satisfy the given constraints.
+
+        :param identifier: An identifier as returned by ``identify()``. This
+            identifies the dependency matches of which should be returned.
+        :param requirements: A mapping of requirements that all returned
+            candidates must satisfy. Each key is an identifier, and the value
+            an iterator of requirements for that dependency.
+        :param incompatibilities: A mapping of known incompatibilities of
+            each dependency. Each key is an identifier, and the value an
+            iterator of incompatibilities known to the resolver. All
+            incompatibilities *must* be excluded from the return value.
+
+        This should try to get candidates based on the requirements' types.
+        For VCS, local, and archive requirements, the one-and-only match is
+        returned, and for a "named" requirement, the index(es) should be
+        consulted to find concrete candidates for this requirement.
+
+        The return value should produce candidates ordered by preference; the
+        most preferred candidate should come first. The return type may be one
+        of the following:
+
+        * A callable that returns an iterator that yields candidates.
+        * An collection of candidates.
+        * An iterable of candidates. This will be consumed immediately into a
+          list of candidates.
+        """
+        raise NotImplementedError
+
+    def is_satisfied_by(self, requirement, candidate):
+        """Whether the given requirement can be satisfied by a candidate.
+
+        The candidate is guaranteed to have been generated from the
+        requirement.
+
+        A boolean should be returned to indicate whether ``candidate`` is a
+        viable solution to the requirement.
+        """
+        raise NotImplementedError
+
+    def get_dependencies(self, candidate):
+        """Get dependencies of a candidate.
+
+        This should return a collection of requirements that `candidate`
+        specifies as its dependencies.
+        """
+        raise NotImplementedError
+
+
+class AbstractResolver(object):
+    """The thing that performs the actual resolution work."""
+
+    base_exception = Exception
+
+    def __init__(self, provider, reporter):
+        self.provider = provider
+        self.reporter = reporter
+
+    def resolve(self, requirements, **kwargs):
+        """Take a collection of constraints, spit out the resolution result.
+
+        This returns a representation of the final resolution state, with one
+        guarenteed attribute ``mapping`` that contains resolved candidates as
+        values. The keys are their respective identifiers.
+
+        :param requirements: A collection of constraints.
+        :param kwargs: Additional keyword arguments that subclasses may accept.
+
+        :raises: ``self.base_exception`` or its subclass.
+        """
+        raise NotImplementedError
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/reporters.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/reporters.py
new file mode 100644
index 00000000..688b5e10
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/reporters.py
@@ -0,0 +1,43 @@
+class BaseReporter(object):
+    """Delegate class to provider progress reporting for the resolver."""
+
+    def starting(self):
+        """Called before the resolution actually starts."""
+
+    def starting_round(self, index):
+        """Called before each round of resolution starts.
+
+        The index is zero-based.
+        """
+
+    def ending_round(self, index, state):
+        """Called before each round of resolution ends.
+
+        This is NOT called if the resolution ends at this round. Use `ending`
+        if you want to report finalization. The index is zero-based.
+        """
+
+    def ending(self, state):
+        """Called before the resolution ends successfully."""
+
+    def adding_requirement(self, requirement, parent):
+        """Called when adding a new requirement into the resolve criteria.
+
+        :param requirement: The additional requirement to be applied to filter
+            the available candidaites.
+        :param parent: The candidate that requires ``requirement`` as a
+            dependency, or None if ``requirement`` is one of the root
+            requirements passed in from ``Resolver.resolve()``.
+        """
+
+    def resolving_conflicts(self, causes):
+        """Called when starting to attempt requirement conflict resolution.
+
+        :param causes: The information on the collision that caused the backtracking.
+        """
+
+    def rejecting_candidate(self, criterion, candidate):
+        """Called when rejecting a candidate during backtracking."""
+
+    def pinning(self, candidate):
+        """Called when adding a candidate to the potential solution."""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/resolvers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/resolvers.py
new file mode 100644
index 00000000..2c3d0e30
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/resolvers.py
@@ -0,0 +1,547 @@
+import collections
+import itertools
+import operator
+
+from .providers import AbstractResolver
+from .structs import DirectedGraph, IteratorMapping, build_iter_view
+
+RequirementInformation = collections.namedtuple(
+    "RequirementInformation", ["requirement", "parent"]
+)
+
+
+class ResolverException(Exception):
+    """A base class for all exceptions raised by this module.
+
+    Exceptions derived by this class should all be handled in this module. Any
+    bubbling pass the resolver should be treated as a bug.
+    """
+
+
+class RequirementsConflicted(ResolverException):
+    def __init__(self, criterion):
+        super(RequirementsConflicted, self).__init__(criterion)
+        self.criterion = criterion
+
+    def __str__(self):
+        return "Requirements conflict: {}".format(
+            ", ".join(repr(r) for r in self.criterion.iter_requirement()),
+        )
+
+
+class InconsistentCandidate(ResolverException):
+    def __init__(self, candidate, criterion):
+        super(InconsistentCandidate, self).__init__(candidate, criterion)
+        self.candidate = candidate
+        self.criterion = criterion
+
+    def __str__(self):
+        return "Provided candidate {!r} does not satisfy {}".format(
+            self.candidate,
+            ", ".join(repr(r) for r in self.criterion.iter_requirement()),
+        )
+
+
+class Criterion(object):
+    """Representation of possible resolution results of a package.
+
+    This holds three attributes:
+
+    * `information` is a collection of `RequirementInformation` pairs.
+      Each pair is a requirement contributing to this criterion, and the
+      candidate that provides the requirement.
+    * `incompatibilities` is a collection of all known not-to-work candidates
+      to exclude from consideration.
+    * `candidates` is a collection containing all possible candidates deducted
+      from the union of contributing requirements and known incompatibilities.
+      It should never be empty, except when the criterion is an attribute of a
+      raised `RequirementsConflicted` (in which case it is always empty).
+
+    .. note::
+        This class is intended to be externally immutable. **Do not** mutate
+        any of its attribute containers.
+    """
+
+    def __init__(self, candidates, information, incompatibilities):
+        self.candidates = candidates
+        self.information = information
+        self.incompatibilities = incompatibilities
+
+    def __repr__(self):
+        requirements = ", ".join(
+            "({!r}, via={!r})".format(req, parent)
+            for req, parent in self.information
+        )
+        return "Criterion({})".format(requirements)
+
+    def iter_requirement(self):
+        return (i.requirement for i in self.information)
+
+    def iter_parent(self):
+        return (i.parent for i in self.information)
+
+
+class ResolutionError(ResolverException):
+    pass
+
+
+class ResolutionImpossible(ResolutionError):
+    def __init__(self, causes):
+        super(ResolutionImpossible, self).__init__(causes)
+        # causes is a list of RequirementInformation objects
+        self.causes = causes
+
+
+class ResolutionTooDeep(ResolutionError):
+    def __init__(self, round_count):
+        super(ResolutionTooDeep, self).__init__(round_count)
+        self.round_count = round_count
+
+
+# Resolution state in a round.
+State = collections.namedtuple("State", "mapping criteria backtrack_causes")
+
+
+class Resolution(object):
+    """Stateful resolution object.
+
+    This is designed as a one-off object that holds information to kick start
+    the resolution process, and holds the results afterwards.
+    """
+
+    def __init__(self, provider, reporter):
+        self._p = provider
+        self._r = reporter
+        self._states = []
+
+    @property
+    def state(self):
+        try:
+            return self._states[-1]
+        except IndexError:
+            raise AttributeError("state")
+
+    def _push_new_state(self):
+        """Push a new state into history.
+
+        This new state will be used to hold resolution results of the next
+        coming round.
+        """
+        base = self._states[-1]
+        state = State(
+            mapping=base.mapping.copy(),
+            criteria=base.criteria.copy(),
+            backtrack_causes=base.backtrack_causes[:],
+        )
+        self._states.append(state)
+
+    def _add_to_criteria(self, criteria, requirement, parent):
+        self._r.adding_requirement(requirement=requirement, parent=parent)
+
+        identifier = self._p.identify(requirement_or_candidate=requirement)
+        criterion = criteria.get(identifier)
+        if criterion:
+            incompatibilities = list(criterion.incompatibilities)
+        else:
+            incompatibilities = []
+
+        matches = self._p.find_matches(
+            identifier=identifier,
+            requirements=IteratorMapping(
+                criteria,
+                operator.methodcaller("iter_requirement"),
+                {identifier: [requirement]},
+            ),
+            incompatibilities=IteratorMapping(
+                criteria,
+                operator.attrgetter("incompatibilities"),
+                {identifier: incompatibilities},
+            ),
+        )
+
+        if criterion:
+            information = list(criterion.information)
+            information.append(RequirementInformation(requirement, parent))
+        else:
+            information = [RequirementInformation(requirement, parent)]
+
+        criterion = Criterion(
+            candidates=build_iter_view(matches),
+            information=information,
+            incompatibilities=incompatibilities,
+        )
+        if not criterion.candidates:
+            raise RequirementsConflicted(criterion)
+        criteria[identifier] = criterion
+
+    def _remove_information_from_criteria(self, criteria, parents):
+        """Remove information from parents of criteria.
+
+        Concretely, removes all values from each criterion's ``information``
+        field that have one of ``parents`` as provider of the requirement.
+
+        :param criteria: The criteria to update.
+        :param parents: Identifiers for which to remove information from all criteria.
+        """
+        if not parents:
+            return
+        for key, criterion in criteria.items():
+            criteria[key] = Criterion(
+                criterion.candidates,
+                [
+                    information
+                    for information in criterion.information
+                    if (
+                        information.parent is None
+                        or self._p.identify(information.parent) not in parents
+                    )
+                ],
+                criterion.incompatibilities,
+            )
+
+    def _get_preference(self, name):
+        return self._p.get_preference(
+            identifier=name,
+            resolutions=self.state.mapping,
+            candidates=IteratorMapping(
+                self.state.criteria,
+                operator.attrgetter("candidates"),
+            ),
+            information=IteratorMapping(
+                self.state.criteria,
+                operator.attrgetter("information"),
+            ),
+            backtrack_causes=self.state.backtrack_causes,
+        )
+
+    def _is_current_pin_satisfying(self, name, criterion):
+        try:
+            current_pin = self.state.mapping[name]
+        except KeyError:
+            return False
+        return all(
+            self._p.is_satisfied_by(requirement=r, candidate=current_pin)
+            for r in criterion.iter_requirement()
+        )
+
+    def _get_updated_criteria(self, candidate):
+        criteria = self.state.criteria.copy()
+        for requirement in self._p.get_dependencies(candidate=candidate):
+            self._add_to_criteria(criteria, requirement, parent=candidate)
+        return criteria
+
+    def _attempt_to_pin_criterion(self, name):
+        criterion = self.state.criteria[name]
+
+        causes = []
+        for candidate in criterion.candidates:
+            try:
+                criteria = self._get_updated_criteria(candidate)
+            except RequirementsConflicted as e:
+                self._r.rejecting_candidate(e.criterion, candidate)
+                causes.append(e.criterion)
+                continue
+
+            # Check the newly-pinned candidate actually works. This should
+            # always pass under normal circumstances, but in the case of a
+            # faulty provider, we will raise an error to notify the implementer
+            # to fix find_matches() and/or is_satisfied_by().
+            satisfied = all(
+                self._p.is_satisfied_by(requirement=r, candidate=candidate)
+                for r in criterion.iter_requirement()
+            )
+            if not satisfied:
+                raise InconsistentCandidate(candidate, criterion)
+
+            self._r.pinning(candidate=candidate)
+            self.state.criteria.update(criteria)
+
+            # Put newly-pinned candidate at the end. This is essential because
+            # backtracking looks at this mapping to get the last pin.
+            self.state.mapping.pop(name, None)
+            self.state.mapping[name] = candidate
+
+            return []
+
+        # All candidates tried, nothing works. This criterion is a dead
+        # end, signal for backtracking.
+        return causes
+
+    def _backjump(self, causes):
+        """Perform backjumping.
+
+        When we enter here, the stack is like this::
+
+            [ state Z ]
+            [ state Y ]
+            [ state X ]
+            .... earlier states are irrelevant.
+
+        1. No pins worked for Z, so it does not have a pin.
+        2. We want to reset state Y to unpinned, and pin another candidate.
+        3. State X holds what state Y was before the pin, but does not
+           have the incompatibility information gathered in state Y.
+
+        Each iteration of the loop will:
+
+        1.  Identify Z. The incompatibility is not always caused by the latest
+            state. For example, given three requirements A, B and C, with
+            dependencies A1, B1 and C1, where A1 and B1 are incompatible: the
+            last state might be related to C, so we want to discard the
+            previous state.
+        2.  Discard Z.
+        3.  Discard Y but remember its incompatibility information gathered
+            previously, and the failure we're dealing with right now.
+        4.  Push a new state Y' based on X, and apply the incompatibility
+            information from Y to Y'.
+        5a. If this causes Y' to conflict, we need to backtrack again. Make Y'
+            the new Z and go back to step 2.
+        5b. If the incompatibilities apply cleanly, end backtracking.
+        """
+        incompatible_reqs = itertools.chain(
+            (c.parent for c in causes if c.parent is not None),
+            (c.requirement for c in causes),
+        )
+        incompatible_deps = {self._p.identify(r) for r in incompatible_reqs}
+        while len(self._states) >= 3:
+            # Remove the state that triggered backtracking.
+            del self._states[-1]
+
+            # Ensure to backtrack to a state that caused the incompatibility
+            incompatible_state = False
+            while not incompatible_state:
+                # Retrieve the last candidate pin and known incompatibilities.
+                try:
+                    broken_state = self._states.pop()
+                    name, candidate = broken_state.mapping.popitem()
+                except (IndexError, KeyError):
+                    raise ResolutionImpossible(causes)
+                current_dependencies = {
+                    self._p.identify(d)
+                    for d in self._p.get_dependencies(candidate)
+                }
+                incompatible_state = not current_dependencies.isdisjoint(
+                    incompatible_deps
+                )
+
+            incompatibilities_from_broken = [
+                (k, list(v.incompatibilities))
+                for k, v in broken_state.criteria.items()
+            ]
+
+            # Also mark the newly known incompatibility.
+            incompatibilities_from_broken.append((name, [candidate]))
+
+            # Create a new state from the last known-to-work one, and apply
+            # the previously gathered incompatibility information.
+            def _patch_criteria():
+                for k, incompatibilities in incompatibilities_from_broken:
+                    if not incompatibilities:
+                        continue
+                    try:
+                        criterion = self.state.criteria[k]
+                    except KeyError:
+                        continue
+                    matches = self._p.find_matches(
+                        identifier=k,
+                        requirements=IteratorMapping(
+                            self.state.criteria,
+                            operator.methodcaller("iter_requirement"),
+                        ),
+                        incompatibilities=IteratorMapping(
+                            self.state.criteria,
+                            operator.attrgetter("incompatibilities"),
+                            {k: incompatibilities},
+                        ),
+                    )
+                    candidates = build_iter_view(matches)
+                    if not candidates:
+                        return False
+                    incompatibilities.extend(criterion.incompatibilities)
+                    self.state.criteria[k] = Criterion(
+                        candidates=candidates,
+                        information=list(criterion.information),
+                        incompatibilities=incompatibilities,
+                    )
+                return True
+
+            self._push_new_state()
+            success = _patch_criteria()
+
+            # It works! Let's work on this new state.
+            if success:
+                return True
+
+            # State does not work after applying known incompatibilities.
+            # Try the still previous state.
+
+        # No way to backtrack anymore.
+        return False
+
+    def resolve(self, requirements, max_rounds):
+        if self._states:
+            raise RuntimeError("already resolved")
+
+        self._r.starting()
+
+        # Initialize the root state.
+        self._states = [
+            State(
+                mapping=collections.OrderedDict(),
+                criteria={},
+                backtrack_causes=[],
+            )
+        ]
+        for r in requirements:
+            try:
+                self._add_to_criteria(self.state.criteria, r, parent=None)
+            except RequirementsConflicted as e:
+                raise ResolutionImpossible(e.criterion.information)
+
+        # The root state is saved as a sentinel so the first ever pin can have
+        # something to backtrack to if it fails. The root state is basically
+        # pinning the virtual "root" package in the graph.
+        self._push_new_state()
+
+        for round_index in range(max_rounds):
+            self._r.starting_round(index=round_index)
+
+            unsatisfied_names = [
+                key
+                for key, criterion in self.state.criteria.items()
+                if not self._is_current_pin_satisfying(key, criterion)
+            ]
+
+            # All criteria are accounted for. Nothing more to pin, we are done!
+            if not unsatisfied_names:
+                self._r.ending(state=self.state)
+                return self.state
+
+            # keep track of satisfied names to calculate diff after pinning
+            satisfied_names = set(self.state.criteria.keys()) - set(
+                unsatisfied_names
+            )
+
+            # Choose the most preferred unpinned criterion to try.
+            name = min(unsatisfied_names, key=self._get_preference)
+            failure_causes = self._attempt_to_pin_criterion(name)
+
+            if failure_causes:
+                causes = [i for c in failure_causes for i in c.information]
+                # Backjump if pinning fails. The backjump process puts us in
+                # an unpinned state, so we can work on it in the next round.
+                self._r.resolving_conflicts(causes=causes)
+                success = self._backjump(causes)
+                self.state.backtrack_causes[:] = causes
+
+                # Dead ends everywhere. Give up.
+                if not success:
+                    raise ResolutionImpossible(self.state.backtrack_causes)
+            else:
+                # discard as information sources any invalidated names
+                # (unsatisfied names that were previously satisfied)
+                newly_unsatisfied_names = {
+                    key
+                    for key, criterion in self.state.criteria.items()
+                    if key in satisfied_names
+                    and not self._is_current_pin_satisfying(key, criterion)
+                }
+                self._remove_information_from_criteria(
+                    self.state.criteria, newly_unsatisfied_names
+                )
+                # Pinning was successful. Push a new state to do another pin.
+                self._push_new_state()
+
+            self._r.ending_round(index=round_index, state=self.state)
+
+        raise ResolutionTooDeep(max_rounds)
+
+
+def _has_route_to_root(criteria, key, all_keys, connected):
+    if key in connected:
+        return True
+    if key not in criteria:
+        return False
+    for p in criteria[key].iter_parent():
+        try:
+            pkey = all_keys[id(p)]
+        except KeyError:
+            continue
+        if pkey in connected:
+            connected.add(key)
+            return True
+        if _has_route_to_root(criteria, pkey, all_keys, connected):
+            connected.add(key)
+            return True
+    return False
+
+
+Result = collections.namedtuple("Result", "mapping graph criteria")
+
+
+def _build_result(state):
+    mapping = state.mapping
+    all_keys = {id(v): k for k, v in mapping.items()}
+    all_keys[id(None)] = None
+
+    graph = DirectedGraph()
+    graph.add(None)  # Sentinel as root dependencies' parent.
+
+    connected = {None}
+    for key, criterion in state.criteria.items():
+        if not _has_route_to_root(state.criteria, key, all_keys, connected):
+            continue
+        if key not in graph:
+            graph.add(key)
+        for p in criterion.iter_parent():
+            try:
+                pkey = all_keys[id(p)]
+            except KeyError:
+                continue
+            if pkey not in graph:
+                graph.add(pkey)
+            graph.connect(pkey, key)
+
+    return Result(
+        mapping={k: v for k, v in mapping.items() if k in connected},
+        graph=graph,
+        criteria=state.criteria,
+    )
+
+
+class Resolver(AbstractResolver):
+    """The thing that performs the actual resolution work."""
+
+    base_exception = ResolverException
+
+    def resolve(self, requirements, max_rounds=100):
+        """Take a collection of constraints, spit out the resolution result.
+
+        The return value is a representation to the final resolution result. It
+        is a tuple subclass with three public members:
+
+        * `mapping`: A dict of resolved candidates. Each key is an identifier
+            of a requirement (as returned by the provider's `identify` method),
+            and the value is the resolved candidate.
+        * `graph`: A `DirectedGraph` instance representing the dependency tree.
+            The vertices are keys of `mapping`, and each edge represents *why*
+            a particular package is included. A special vertex `None` is
+            included to represent parents of user-supplied requirements.
+        * `criteria`: A dict of "criteria" that hold detailed information on
+            how edges in the graph are derived. Each key is an identifier of a
+            requirement, and the value is a `Criterion` instance.
+
+        The following exceptions may be raised if a resolution cannot be found:
+
+        * `ResolutionImpossible`: A resolution cannot be found for the given
+            combination of requirements. The `causes` attribute of the
+            exception is a list of (requirement, parent), giving the
+            requirements that could not be satisfied.
+        * `ResolutionTooDeep`: The dependency tree is too deeply nested and
+            the resolver gave up. This is usually caused by a circular
+            dependency, but you can try to resolve this by increasing the
+            `max_rounds` argument.
+        """
+        resolution = Resolution(self.provider, self.reporter)
+        state = resolution.resolve(requirements, max_rounds=max_rounds)
+        return _build_result(state)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/structs.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/structs.py
new file mode 100644
index 00000000..359a34f6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/resolvelib/structs.py
@@ -0,0 +1,170 @@
+import itertools
+
+from .compat import collections_abc
+
+
+class DirectedGraph(object):
+    """A graph structure with directed edges."""
+
+    def __init__(self):
+        self._vertices = set()
+        self._forwards = {}  # <key> -> Set[<key>]
+        self._backwards = {}  # <key> -> Set[<key>]
+
+    def __iter__(self):
+        return iter(self._vertices)
+
+    def __len__(self):
+        return len(self._vertices)
+
+    def __contains__(self, key):
+        return key in self._vertices
+
+    def copy(self):
+        """Return a shallow copy of this graph."""
+        other = DirectedGraph()
+        other._vertices = set(self._vertices)
+        other._forwards = {k: set(v) for k, v in self._forwards.items()}
+        other._backwards = {k: set(v) for k, v in self._backwards.items()}
+        return other
+
+    def add(self, key):
+        """Add a new vertex to the graph."""
+        if key in self._vertices:
+            raise ValueError("vertex exists")
+        self._vertices.add(key)
+        self._forwards[key] = set()
+        self._backwards[key] = set()
+
+    def remove(self, key):
+        """Remove a vertex from the graph, disconnecting all edges from/to it."""
+        self._vertices.remove(key)
+        for f in self._forwards.pop(key):
+            self._backwards[f].remove(key)
+        for t in self._backwards.pop(key):
+            self._forwards[t].remove(key)
+
+    def connected(self, f, t):
+        return f in self._backwards[t] and t in self._forwards[f]
+
+    def connect(self, f, t):
+        """Connect two existing vertices.
+
+        Nothing happens if the vertices are already connected.
+        """
+        if t not in self._vertices:
+            raise KeyError(t)
+        self._forwards[f].add(t)
+        self._backwards[t].add(f)
+
+    def iter_edges(self):
+        for f, children in self._forwards.items():
+            for t in children:
+                yield f, t
+
+    def iter_children(self, key):
+        return iter(self._forwards[key])
+
+    def iter_parents(self, key):
+        return iter(self._backwards[key])
+
+
+class IteratorMapping(collections_abc.Mapping):
+    def __init__(self, mapping, accessor, appends=None):
+        self._mapping = mapping
+        self._accessor = accessor
+        self._appends = appends or {}
+
+    def __repr__(self):
+        return "IteratorMapping({!r}, {!r}, {!r})".format(
+            self._mapping,
+            self._accessor,
+            self._appends,
+        )
+
+    def __bool__(self):
+        return bool(self._mapping or self._appends)
+
+    __nonzero__ = __bool__  # XXX: Python 2.
+
+    def __contains__(self, key):
+        return key in self._mapping or key in self._appends
+
+    def __getitem__(self, k):
+        try:
+            v = self._mapping[k]
+        except KeyError:
+            return iter(self._appends[k])
+        return itertools.chain(self._accessor(v), self._appends.get(k, ()))
+
+    def __iter__(self):
+        more = (k for k in self._appends if k not in self._mapping)
+        return itertools.chain(self._mapping, more)
+
+    def __len__(self):
+        more = sum(1 for k in self._appends if k not in self._mapping)
+        return len(self._mapping) + more
+
+
+class _FactoryIterableView(object):
+    """Wrap an iterator factory returned by `find_matches()`.
+
+    Calling `iter()` on this class would invoke the underlying iterator
+    factory, making it a "collection with ordering" that can be iterated
+    through multiple times, but lacks random access methods presented in
+    built-in Python sequence types.
+    """
+
+    def __init__(self, factory):
+        self._factory = factory
+        self._iterable = None
+
+    def __repr__(self):
+        return "{}({})".format(type(self).__name__, list(self))
+
+    def __bool__(self):
+        try:
+            next(iter(self))
+        except StopIteration:
+            return False
+        return True
+
+    __nonzero__ = __bool__  # XXX: Python 2.
+
+    def __iter__(self):
+        iterable = (
+            self._factory() if self._iterable is None else self._iterable
+        )
+        self._iterable, current = itertools.tee(iterable)
+        return current
+
+
+class _SequenceIterableView(object):
+    """Wrap an iterable returned by find_matches().
+
+    This is essentially just a proxy to the underlying sequence that provides
+    the same interface as `_FactoryIterableView`.
+    """
+
+    def __init__(self, sequence):
+        self._sequence = sequence
+
+    def __repr__(self):
+        return "{}({})".format(type(self).__name__, self._sequence)
+
+    def __bool__(self):
+        return bool(self._sequence)
+
+    __nonzero__ = __bool__  # XXX: Python 2.
+
+    def __iter__(self):
+        return iter(self._sequence)
+
+
+def build_iter_view(matches):
+    """Build an iterable view from the value returned by `find_matches()`."""
+    if callable(matches):
+        return _FactoryIterableView(matches)
+    if not isinstance(matches, collections_abc.Sequence):
+        matches = list(matches)
+    return _SequenceIterableView(matches)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__init__.py
new file mode 100644
index 00000000..73f58d77
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__init__.py
@@ -0,0 +1,177 @@
+"""Rich text and beautiful formatting in the terminal."""
+
+import os
+from typing import IO, TYPE_CHECKING, Any, Callable, Optional, Union
+
+from ._extension import load_ipython_extension  # noqa: F401
+
+__all__ = ["get_console", "reconfigure", "print", "inspect", "print_json"]
+
+if TYPE_CHECKING:
+    from .console import Console
+
+# Global console used by alternative print
+_console: Optional["Console"] = None
+
+try:
+    _IMPORT_CWD = os.path.abspath(os.getcwd())
+except FileNotFoundError:
+    # Can happen if the cwd has been deleted
+    _IMPORT_CWD = ""
+
+
+def get_console() -> "Console":
+    """Get a global :class:`~rich.console.Console` instance. This function is used when Rich requires a Console,
+    and hasn't been explicitly given one.
+
+    Returns:
+        Console: A console instance.
+    """
+    global _console
+    if _console is None:
+        from .console import Console
+
+        _console = Console()
+
+    return _console
+
+
+def reconfigure(*args: Any, **kwargs: Any) -> None:
+    """Reconfigures the global console by replacing it with another.
+
+    Args:
+        *args (Any): Positional arguments for the replacement :class:`~rich.console.Console`.
+        **kwargs (Any): Keyword arguments for the replacement :class:`~rich.console.Console`.
+    """
+    from pip._vendor.rich.console import Console
+
+    new_console = Console(*args, **kwargs)
+    _console = get_console()
+    _console.__dict__ = new_console.__dict__
+
+
+def print(
+    *objects: Any,
+    sep: str = " ",
+    end: str = "\n",
+    file: Optional[IO[str]] = None,
+    flush: bool = False,
+) -> None:
+    r"""Print object(s) supplied via positional arguments.
+    This function has an identical signature to the built-in print.
+    For more advanced features, see the :class:`~rich.console.Console` class.
+
+    Args:
+        sep (str, optional): Separator between printed objects. Defaults to " ".
+        end (str, optional): Character to write at end of output. Defaults to "\\n".
+        file (IO[str], optional): File to write to, or None for stdout. Defaults to None.
+        flush (bool, optional): Has no effect as Rich always flushes output. Defaults to False.
+
+    """
+    from .console import Console
+
+    write_console = get_console() if file is None else Console(file=file)
+    return write_console.print(*objects, sep=sep, end=end)
+
+
+def print_json(
+    json: Optional[str] = None,
+    *,
+    data: Any = None,
+    indent: Union[None, int, str] = 2,
+    highlight: bool = True,
+    skip_keys: bool = False,
+    ensure_ascii: bool = False,
+    check_circular: bool = True,
+    allow_nan: bool = True,
+    default: Optional[Callable[[Any], Any]] = None,
+    sort_keys: bool = False,
+) -> None:
+    """Pretty prints JSON. Output will be valid JSON.
+
+    Args:
+        json (str): A string containing JSON.
+        data (Any): If json is not supplied, then encode this data.
+        indent (int, optional): Number of spaces to indent. Defaults to 2.
+        highlight (bool, optional): Enable highlighting of output: Defaults to True.
+        skip_keys (bool, optional): Skip keys not of a basic type. Defaults to False.
+        ensure_ascii (bool, optional): Escape all non-ascii characters. Defaults to False.
+        check_circular (bool, optional): Check for circular references. Defaults to True.
+        allow_nan (bool, optional): Allow NaN and Infinity values. Defaults to True.
+        default (Callable, optional): A callable that converts values that can not be encoded
+            in to something that can be JSON encoded. Defaults to None.
+        sort_keys (bool, optional): Sort dictionary keys. Defaults to False.
+    """
+
+    get_console().print_json(
+        json,
+        data=data,
+        indent=indent,
+        highlight=highlight,
+        skip_keys=skip_keys,
+        ensure_ascii=ensure_ascii,
+        check_circular=check_circular,
+        allow_nan=allow_nan,
+        default=default,
+        sort_keys=sort_keys,
+    )
+
+
+def inspect(
+    obj: Any,
+    *,
+    console: Optional["Console"] = None,
+    title: Optional[str] = None,
+    help: bool = False,
+    methods: bool = False,
+    docs: bool = True,
+    private: bool = False,
+    dunder: bool = False,
+    sort: bool = True,
+    all: bool = False,
+    value: bool = True,
+) -> None:
+    """Inspect any Python object.
+
+    * inspect(<OBJECT>) to see summarized info.
+    * inspect(<OBJECT>, methods=True) to see methods.
+    * inspect(<OBJECT>, help=True) to see full (non-abbreviated) help.
+    * inspect(<OBJECT>, private=True) to see private attributes (single underscore).
+    * inspect(<OBJECT>, dunder=True) to see attributes beginning with double underscore.
+    * inspect(<OBJECT>, all=True) to see all attributes.
+
+    Args:
+        obj (Any): An object to inspect.
+        title (str, optional): Title to display over inspect result, or None use type. Defaults to None.
+        help (bool, optional): Show full help text rather than just first paragraph. Defaults to False.
+        methods (bool, optional): Enable inspection of callables. Defaults to False.
+        docs (bool, optional): Also render doc strings. Defaults to True.
+        private (bool, optional): Show private attributes (beginning with underscore). Defaults to False.
+        dunder (bool, optional): Show attributes starting with double underscore. Defaults to False.
+        sort (bool, optional): Sort attributes alphabetically. Defaults to True.
+        all (bool, optional): Show all attributes. Defaults to False.
+        value (bool, optional): Pretty print value. Defaults to True.
+    """
+    _console = console or get_console()
+    from pip._vendor.rich._inspect import Inspect
+
+    # Special case for inspect(inspect)
+    is_inspect = obj is inspect
+
+    _inspect = Inspect(
+        obj,
+        title=title,
+        help=is_inspect or help,
+        methods=is_inspect or methods,
+        docs=is_inspect or docs,
+        private=private,
+        dunder=dunder,
+        sort=sort,
+        all=all,
+        value=value,
+    )
+    _console.print(_inspect)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    print("Hello, **World**")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__main__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__main__.py
new file mode 100644
index 00000000..efb7fb79
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__main__.py
@@ -0,0 +1,273 @@
+import colorsys
+import io
+from time import process_time
+
+from pip._vendor.rich import box
+from pip._vendor.rich.color import Color
+from pip._vendor.rich.console import Console, ConsoleOptions, Group, RenderableType, RenderResult
+from pip._vendor.rich.markdown import Markdown
+from pip._vendor.rich.measure import Measurement
+from pip._vendor.rich.pretty import Pretty
+from pip._vendor.rich.segment import Segment
+from pip._vendor.rich.style import Style
+from pip._vendor.rich.syntax import Syntax
+from pip._vendor.rich.table import Table
+from pip._vendor.rich.text import Text
+
+
+class ColorBox:
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        for y in range(0, 5):
+            for x in range(options.max_width):
+                h = x / options.max_width
+                l = 0.1 + ((y / 5) * 0.7)
+                r1, g1, b1 = colorsys.hls_to_rgb(h, l, 1.0)
+                r2, g2, b2 = colorsys.hls_to_rgb(h, l + 0.7 / 10, 1.0)
+                bgcolor = Color.from_rgb(r1 * 255, g1 * 255, b1 * 255)
+                color = Color.from_rgb(r2 * 255, g2 * 255, b2 * 255)
+                yield Segment("▄", Style(color=color, bgcolor=bgcolor))
+            yield Segment.line()
+
+    def __rich_measure__(
+        self, console: "Console", options: ConsoleOptions
+    ) -> Measurement:
+        return Measurement(1, options.max_width)
+
+
+def make_test_card() -> Table:
+    """Get a renderable that demonstrates a number of features."""
+    table = Table.grid(padding=1, pad_edge=True)
+    table.title = "Rich features"
+    table.add_column("Feature", no_wrap=True, justify="center", style="bold red")
+    table.add_column("Demonstration")
+
+    color_table = Table(
+        box=None,
+        expand=False,
+        show_header=False,
+        show_edge=False,
+        pad_edge=False,
+    )
+    color_table.add_row(
+        (
+            "✓ [bold green]4-bit color[/]\n"
+            "✓ [bold blue]8-bit color[/]\n"
+            "✓ [bold magenta]Truecolor (16.7 million)[/]\n"
+            "✓ [bold yellow]Dumb terminals[/]\n"
+            "✓ [bold cyan]Automatic color conversion"
+        ),
+        ColorBox(),
+    )
+
+    table.add_row("Colors", color_table)
+
+    table.add_row(
+        "Styles",
+        "All ansi styles: [bold]bold[/], [dim]dim[/], [italic]italic[/italic], [underline]underline[/], [strike]strikethrough[/], [reverse]reverse[/], and even [blink]blink[/].",
+    )
+
+    lorem = "Lorem ipsum dolor sit amet, consectetur adipiscing elit. Quisque in metus sed sapien ultricies pretium a at justo. Maecenas luctus velit et auctor maximus."
+    lorem_table = Table.grid(padding=1, collapse_padding=True)
+    lorem_table.pad_edge = False
+    lorem_table.add_row(
+        Text(lorem, justify="left", style="green"),
+        Text(lorem, justify="center", style="yellow"),
+        Text(lorem, justify="right", style="blue"),
+        Text(lorem, justify="full", style="red"),
+    )
+    table.add_row(
+        "Text",
+        Group(
+            Text.from_markup(
+                """Word wrap text. Justify [green]left[/], [yellow]center[/], [blue]right[/] or [red]full[/].\n"""
+            ),
+            lorem_table,
+        ),
+    )
+
+    def comparison(renderable1: RenderableType, renderable2: RenderableType) -> Table:
+        table = Table(show_header=False, pad_edge=False, box=None, expand=True)
+        table.add_column("1", ratio=1)
+        table.add_column("2", ratio=1)
+        table.add_row(renderable1, renderable2)
+        return table
+
+    table.add_row(
+        "Asian\nlanguage\nsupport",
+        ":flag_for_china:  该库支持中文，日文和韩文文本！\n:flag_for_japan:  ライブラリは中国語、日本語、韓国語のテキストをサポートしています\n:flag_for_south_korea:  이 라이브러리는 중국어, 일본어 및 한국어 텍스트를 지원합니다",
+    )
+
+    markup_example = (
+        "[bold magenta]Rich[/] supports a simple [i]bbcode[/i]-like [b]markup[/b] for [yellow]color[/], [underline]style[/], and emoji! "
+        ":+1: :apple: :ant: :bear: :baguette_bread: :bus: "
+    )
+    table.add_row("Markup", markup_example)
+
+    example_table = Table(
+        show_edge=False,
+        show_header=True,
+        expand=False,
+        row_styles=["none", "dim"],
+        box=box.SIMPLE,
+    )
+    example_table.add_column("[green]Date", style="green", no_wrap=True)
+    example_table.add_column("[blue]Title", style="blue")
+    example_table.add_column(
+        "[cyan]Production Budget",
+        style="cyan",
+        justify="right",
+        no_wrap=True,
+    )
+    example_table.add_column(
+        "[magenta]Box Office",
+        style="magenta",
+        justify="right",
+        no_wrap=True,
+    )
+    example_table.add_row(
+        "Dec 20, 2019",
+        "Star Wars: The Rise of Skywalker",
+        "$275,000,000",
+        "$375,126,118",
+    )
+    example_table.add_row(
+        "May 25, 2018",
+        "[b]Solo[/]: A Star Wars Story",
+        "$275,000,000",
+        "$393,151,347",
+    )
+    example_table.add_row(
+        "Dec 15, 2017",
+        "Star Wars Ep. VIII: The Last Jedi",
+        "$262,000,000",
+        "[bold]$1,332,539,889[/bold]",
+    )
+    example_table.add_row(
+        "May 19, 1999",
+        "Star Wars Ep. [b]I[/b]: [i]The phantom Menace",
+        "$115,000,000",
+        "$1,027,044,677",
+    )
+
+    table.add_row("Tables", example_table)
+
+    code = '''\
+def iter_last(values: Iterable[T]) -> Iterable[Tuple[bool, T]]:
+    """Iterate and generate a tuple with a flag for last value."""
+    iter_values = iter(values)
+    try:
+        previous_value = next(iter_values)
+    except StopIteration:
+        return
+    for value in iter_values:
+        yield False, previous_value
+        previous_value = value
+    yield True, previous_value'''
+
+    pretty_data = {
+        "foo": [
+            3.1427,
+            (
+                "Paul Atreides",
+                "Vladimir Harkonnen",
+                "Thufir Hawat",
+            ),
+        ],
+        "atomic": (False, True, None),
+    }
+    table.add_row(
+        "Syntax\nhighlighting\n&\npretty\nprinting",
+        comparison(
+            Syntax(code, "python3", line_numbers=True, indent_guides=True),
+            Pretty(pretty_data, indent_guides=True),
+        ),
+    )
+
+    markdown_example = """\
+# Markdown
+
+Supports much of the *markdown* __syntax__!
+
+- Headers
+- Basic formatting: **bold**, *italic*, `code`
+- Block quotes
+- Lists, and more...
+    """
+    table.add_row(
+        "Markdown", comparison("[cyan]" + markdown_example, Markdown(markdown_example))
+    )
+
+    table.add_row(
+        "+more!",
+        """Progress bars, columns, styled logging handler, tracebacks, etc...""",
+    )
+    return table
+
+
+if __name__ == "__main__":  # pragma: no cover
+    console = Console(
+        file=io.StringIO(),
+        force_terminal=True,
+    )
+    test_card = make_test_card()
+
+    # Print once to warm cache
+    start = process_time()
+    console.print(test_card)
+    pre_cache_taken = round((process_time() - start) * 1000.0, 1)
+
+    console.file = io.StringIO()
+
+    start = process_time()
+    console.print(test_card)
+    taken = round((process_time() - start) * 1000.0, 1)
+
+    c = Console(record=True)
+    c.print(test_card)
+
+    print(f"rendered in {pre_cache_taken}ms (cold cache)")
+    print(f"rendered in {taken}ms (warm cache)")
+
+    from pip._vendor.rich.panel import Panel
+
+    console = Console()
+
+    sponsor_message = Table.grid(padding=1)
+    sponsor_message.add_column(style="green", justify="right")
+    sponsor_message.add_column(no_wrap=True)
+
+    sponsor_message.add_row(
+        "Textualize",
+        "[u blue link=https://github.com/textualize]https://github.com/textualize",
+    )
+    sponsor_message.add_row(
+        "Twitter",
+        "[u blue link=https://twitter.com/willmcgugan]https://twitter.com/willmcgugan",
+    )
+
+    intro_message = Text.from_markup(
+        """\
+We hope you enjoy using Rich!
+
+Rich is maintained with [red]:heart:[/] by [link=https://www.textualize.io]Textualize.io[/]
+
+- Will McGugan"""
+    )
+
+    message = Table.grid(padding=2)
+    message.add_column()
+    message.add_column(no_wrap=True)
+    message.add_row(intro_message, sponsor_message)
+
+    console.print(
+        Panel.fit(
+            message,
+            box=box.ROUNDED,
+            padding=(1, 2),
+            title="[b red]Thanks for trying out Rich!",
+            border_style="bright_blue",
+        ),
+        justify="center",
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..004c8aa5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..5c2dbeb0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_cell_widths.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_cell_widths.cpython-312.pyc
new file mode 100644
index 00000000..c5a09598
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_cell_widths.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_codes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_codes.cpython-312.pyc
new file mode 100644
index 00000000..54eef923
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_codes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_replace.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_replace.cpython-312.pyc
new file mode 100644
index 00000000..ec92fde7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_emoji_replace.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_export_format.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_export_format.cpython-312.pyc
new file mode 100644
index 00000000..f2681180
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_export_format.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_extension.cpython-312.pyc
new file mode 100644
index 00000000..d14c7840
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_fileno.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_fileno.cpython-312.pyc
new file mode 100644
index 00000000..174ec9f1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_fileno.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_inspect.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_inspect.cpython-312.pyc
new file mode 100644
index 00000000..b64e428f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_inspect.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_log_render.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_log_render.cpython-312.pyc
new file mode 100644
index 00000000..926313f7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_log_render.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_loop.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_loop.cpython-312.pyc
new file mode 100644
index 00000000..2f9f7b65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_loop.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_null_file.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_null_file.cpython-312.pyc
new file mode 100644
index 00000000..009d4449
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_null_file.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_palettes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_palettes.cpython-312.pyc
new file mode 100644
index 00000000..8c039df9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_palettes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_pick.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_pick.cpython-312.pyc
new file mode 100644
index 00000000..6a4c85af
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_pick.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_ratio.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_ratio.cpython-312.pyc
new file mode 100644
index 00000000..99a840a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_ratio.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_spinners.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_spinners.cpython-312.pyc
new file mode 100644
index 00000000..6f70ef01
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_spinners.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_stack.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_stack.cpython-312.pyc
new file mode 100644
index 00000000..33ca9678
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_stack.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_timer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_timer.cpython-312.pyc
new file mode 100644
index 00000000..380b9adf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_timer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_win32_console.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_win32_console.cpython-312.pyc
new file mode 100644
index 00000000..15414529
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_win32_console.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows.cpython-312.pyc
new file mode 100644
index 00000000..98f6812a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows_renderer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows_renderer.cpython-312.pyc
new file mode 100644
index 00000000..92e2635a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_windows_renderer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_wrap.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_wrap.cpython-312.pyc
new file mode 100644
index 00000000..7184edd5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/_wrap.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/abc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/abc.cpython-312.pyc
new file mode 100644
index 00000000..c6aec332
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/abc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/align.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/align.cpython-312.pyc
new file mode 100644
index 00000000..d0d4785b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/align.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/ansi.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/ansi.cpython-312.pyc
new file mode 100644
index 00000000..147772e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/ansi.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/bar.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/bar.cpython-312.pyc
new file mode 100644
index 00000000..2ce7f7db
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/bar.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/box.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/box.cpython-312.pyc
new file mode 100644
index 00000000..0770c144
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/box.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/cells.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/cells.cpython-312.pyc
new file mode 100644
index 00000000..b312a917
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/cells.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color.cpython-312.pyc
new file mode 100644
index 00000000..96fba285
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color_triplet.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color_triplet.cpython-312.pyc
new file mode 100644
index 00000000..d75f0606
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/color_triplet.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/columns.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/columns.cpython-312.pyc
new file mode 100644
index 00000000..94ba11be
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/columns.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/console.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/console.cpython-312.pyc
new file mode 100644
index 00000000..4fee0a1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/console.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/constrain.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/constrain.cpython-312.pyc
new file mode 100644
index 00000000..00e79e2e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/constrain.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/containers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/containers.cpython-312.pyc
new file mode 100644
index 00000000..6fd4db73
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/containers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/control.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/control.cpython-312.pyc
new file mode 100644
index 00000000..ec6d39a8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/control.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/default_styles.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/default_styles.cpython-312.pyc
new file mode 100644
index 00000000..bec2fb3a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/default_styles.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/diagnose.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/diagnose.cpython-312.pyc
new file mode 100644
index 00000000..3740a624
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/diagnose.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/emoji.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/emoji.cpython-312.pyc
new file mode 100644
index 00000000..f1549eae
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/emoji.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/errors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/errors.cpython-312.pyc
new file mode 100644
index 00000000..eb44bde9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/errors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/file_proxy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/file_proxy.cpython-312.pyc
new file mode 100644
index 00000000..881a0551
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/file_proxy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/filesize.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/filesize.cpython-312.pyc
new file mode 100644
index 00000000..76655f7f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/filesize.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/highlighter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/highlighter.cpython-312.pyc
new file mode 100644
index 00000000..45d94d02
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/highlighter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/json.cpython-312.pyc
new file mode 100644
index 00000000..7f99ec87
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/jupyter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/jupyter.cpython-312.pyc
new file mode 100644
index 00000000..36941371
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/jupyter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/layout.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/layout.cpython-312.pyc
new file mode 100644
index 00000000..4605c1c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/layout.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live.cpython-312.pyc
new file mode 100644
index 00000000..ed7b92f1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live_render.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live_render.cpython-312.pyc
new file mode 100644
index 00000000..61471d4a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/live_render.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/logging.cpython-312.pyc
new file mode 100644
index 00000000..8b13e558
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/markup.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/markup.cpython-312.pyc
new file mode 100644
index 00000000..3832d51f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/markup.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/measure.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/measure.cpython-312.pyc
new file mode 100644
index 00000000..ca3d5e1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/measure.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/padding.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/padding.cpython-312.pyc
new file mode 100644
index 00000000..2f01aede
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/padding.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pager.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pager.cpython-312.pyc
new file mode 100644
index 00000000..a14d67f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pager.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/palette.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/palette.cpython-312.pyc
new file mode 100644
index 00000000..c74e6ec7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/palette.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/panel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/panel.cpython-312.pyc
new file mode 100644
index 00000000..7100f74c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/panel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pretty.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pretty.cpython-312.pyc
new file mode 100644
index 00000000..792421a2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/pretty.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress.cpython-312.pyc
new file mode 100644
index 00000000..4301d7c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress_bar.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress_bar.cpython-312.pyc
new file mode 100644
index 00000000..22a98cbf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/progress_bar.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/prompt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/prompt.cpython-312.pyc
new file mode 100644
index 00000000..8c39935d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/prompt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/protocol.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/protocol.cpython-312.pyc
new file mode 100644
index 00000000..60328a99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/protocol.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/region.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/region.cpython-312.pyc
new file mode 100644
index 00000000..509d9b4c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/region.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/repr.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/repr.cpython-312.pyc
new file mode 100644
index 00000000..fd06c5ba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/repr.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/rule.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/rule.cpython-312.pyc
new file mode 100644
index 00000000..5a938f47
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/rule.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/scope.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/scope.cpython-312.pyc
new file mode 100644
index 00000000..e5329699
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/scope.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/screen.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/screen.cpython-312.pyc
new file mode 100644
index 00000000..39e05e6e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/screen.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/segment.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/segment.cpython-312.pyc
new file mode 100644
index 00000000..67de30aa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/segment.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/spinner.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/spinner.cpython-312.pyc
new file mode 100644
index 00000000..edc6b848
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/spinner.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/status.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/status.cpython-312.pyc
new file mode 100644
index 00000000..209047b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/status.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/style.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/style.cpython-312.pyc
new file mode 100644
index 00000000..400a9b1a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/style.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/styled.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/styled.cpython-312.pyc
new file mode 100644
index 00000000..a256b7b3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/styled.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/syntax.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/syntax.cpython-312.pyc
new file mode 100644
index 00000000..94c4c313
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/syntax.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/table.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/table.cpython-312.pyc
new file mode 100644
index 00000000..8b9f7ab5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/table.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/terminal_theme.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/terminal_theme.cpython-312.pyc
new file mode 100644
index 00000000..54500d16
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/terminal_theme.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/text.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/text.cpython-312.pyc
new file mode 100644
index 00000000..64db8fd7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/text.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/theme.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/theme.cpython-312.pyc
new file mode 100644
index 00000000..ff3f4c29
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/theme.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/themes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/themes.cpython-312.pyc
new file mode 100644
index 00000000..bf04ca80
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/themes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/traceback.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/traceback.cpython-312.pyc
new file mode 100644
index 00000000..4e77db3f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/traceback.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/tree.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/tree.cpython-312.pyc
new file mode 100644
index 00000000..48206b6b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/__pycache__/tree.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_cell_widths.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_cell_widths.py
new file mode 100644
index 00000000..608ae3a7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_cell_widths.py
@@ -0,0 +1,454 @@
+# Auto generated by make_terminal_widths.py
+
+CELL_WIDTHS = [
+    (0, 0, 0),
+    (1, 31, -1),
+    (127, 159, -1),
+    (173, 173, 0),
+    (768, 879, 0),
+    (1155, 1161, 0),
+    (1425, 1469, 0),
+    (1471, 1471, 0),
+    (1473, 1474, 0),
+    (1476, 1477, 0),
+    (1479, 1479, 0),
+    (1536, 1541, 0),
+    (1552, 1562, 0),
+    (1564, 1564, 0),
+    (1611, 1631, 0),
+    (1648, 1648, 0),
+    (1750, 1757, 0),
+    (1759, 1764, 0),
+    (1767, 1768, 0),
+    (1770, 1773, 0),
+    (1807, 1807, 0),
+    (1809, 1809, 0),
+    (1840, 1866, 0),
+    (1958, 1968, 0),
+    (2027, 2035, 0),
+    (2045, 2045, 0),
+    (2070, 2073, 0),
+    (2075, 2083, 0),
+    (2085, 2087, 0),
+    (2089, 2093, 0),
+    (2137, 2139, 0),
+    (2192, 2193, 0),
+    (2200, 2207, 0),
+    (2250, 2307, 0),
+    (2362, 2364, 0),
+    (2366, 2383, 0),
+    (2385, 2391, 0),
+    (2402, 2403, 0),
+    (2433, 2435, 0),
+    (2492, 2492, 0),
+    (2494, 2500, 0),
+    (2503, 2504, 0),
+    (2507, 2509, 0),
+    (2519, 2519, 0),
+    (2530, 2531, 0),
+    (2558, 2558, 0),
+    (2561, 2563, 0),
+    (2620, 2620, 0),
+    (2622, 2626, 0),
+    (2631, 2632, 0),
+    (2635, 2637, 0),
+    (2641, 2641, 0),
+    (2672, 2673, 0),
+    (2677, 2677, 0),
+    (2689, 2691, 0),
+    (2748, 2748, 0),
+    (2750, 2757, 0),
+    (2759, 2761, 0),
+    (2763, 2765, 0),
+    (2786, 2787, 0),
+    (2810, 2815, 0),
+    (2817, 2819, 0),
+    (2876, 2876, 0),
+    (2878, 2884, 0),
+    (2887, 2888, 0),
+    (2891, 2893, 0),
+    (2901, 2903, 0),
+    (2914, 2915, 0),
+    (2946, 2946, 0),
+    (3006, 3010, 0),
+    (3014, 3016, 0),
+    (3018, 3021, 0),
+    (3031, 3031, 0),
+    (3072, 3076, 0),
+    (3132, 3132, 0),
+    (3134, 3140, 0),
+    (3142, 3144, 0),
+    (3146, 3149, 0),
+    (3157, 3158, 0),
+    (3170, 3171, 0),
+    (3201, 3203, 0),
+    (3260, 3260, 0),
+    (3262, 3268, 0),
+    (3270, 3272, 0),
+    (3274, 3277, 0),
+    (3285, 3286, 0),
+    (3298, 3299, 0),
+    (3315, 3315, 0),
+    (3328, 3331, 0),
+    (3387, 3388, 0),
+    (3390, 3396, 0),
+    (3398, 3400, 0),
+    (3402, 3405, 0),
+    (3415, 3415, 0),
+    (3426, 3427, 0),
+    (3457, 3459, 0),
+    (3530, 3530, 0),
+    (3535, 3540, 0),
+    (3542, 3542, 0),
+    (3544, 3551, 0),
+    (3570, 3571, 0),
+    (3633, 3633, 0),
+    (3636, 3642, 0),
+    (3655, 3662, 0),
+    (3761, 3761, 0),
+    (3764, 3772, 0),
+    (3784, 3790, 0),
+    (3864, 3865, 0),
+    (3893, 3893, 0),
+    (3895, 3895, 0),
+    (3897, 3897, 0),
+    (3902, 3903, 0),
+    (3953, 3972, 0),
+    (3974, 3975, 0),
+    (3981, 3991, 0),
+    (3993, 4028, 0),
+    (4038, 4038, 0),
+    (4139, 4158, 0),
+    (4182, 4185, 0),
+    (4190, 4192, 0),
+    (4194, 4196, 0),
+    (4199, 4205, 0),
+    (4209, 4212, 0),
+    (4226, 4237, 0),
+    (4239, 4239, 0),
+    (4250, 4253, 0),
+    (4352, 4447, 2),
+    (4448, 4607, 0),
+    (4957, 4959, 0),
+    (5906, 5909, 0),
+    (5938, 5940, 0),
+    (5970, 5971, 0),
+    (6002, 6003, 0),
+    (6068, 6099, 0),
+    (6109, 6109, 0),
+    (6155, 6159, 0),
+    (6277, 6278, 0),
+    (6313, 6313, 0),
+    (6432, 6443, 0),
+    (6448, 6459, 0),
+    (6679, 6683, 0),
+    (6741, 6750, 0),
+    (6752, 6780, 0),
+    (6783, 6783, 0),
+    (6832, 6862, 0),
+    (6912, 6916, 0),
+    (6964, 6980, 0),
+    (7019, 7027, 0),
+    (7040, 7042, 0),
+    (7073, 7085, 0),
+    (7142, 7155, 0),
+    (7204, 7223, 0),
+    (7376, 7378, 0),
+    (7380, 7400, 0),
+    (7405, 7405, 0),
+    (7412, 7412, 0),
+    (7415, 7417, 0),
+    (7616, 7679, 0),
+    (8203, 8207, 0),
+    (8232, 8238, 0),
+    (8288, 8292, 0),
+    (8294, 8303, 0),
+    (8400, 8432, 0),
+    (8986, 8987, 2),
+    (9001, 9002, 2),
+    (9193, 9196, 2),
+    (9200, 9200, 2),
+    (9203, 9203, 2),
+    (9725, 9726, 2),
+    (9748, 9749, 2),
+    (9800, 9811, 2),
+    (9855, 9855, 2),
+    (9875, 9875, 2),
+    (9889, 9889, 2),
+    (9898, 9899, 2),
+    (9917, 9918, 2),
+    (9924, 9925, 2),
+    (9934, 9934, 2),
+    (9940, 9940, 2),
+    (9962, 9962, 2),
+    (9970, 9971, 2),
+    (9973, 9973, 2),
+    (9978, 9978, 2),
+    (9981, 9981, 2),
+    (9989, 9989, 2),
+    (9994, 9995, 2),
+    (10024, 10024, 2),
+    (10060, 10060, 2),
+    (10062, 10062, 2),
+    (10067, 10069, 2),
+    (10071, 10071, 2),
+    (10133, 10135, 2),
+    (10160, 10160, 2),
+    (10175, 10175, 2),
+    (11035, 11036, 2),
+    (11088, 11088, 2),
+    (11093, 11093, 2),
+    (11503, 11505, 0),
+    (11647, 11647, 0),
+    (11744, 11775, 0),
+    (11904, 11929, 2),
+    (11931, 12019, 2),
+    (12032, 12245, 2),
+    (12272, 12329, 2),
+    (12330, 12335, 0),
+    (12336, 12350, 2),
+    (12353, 12438, 2),
+    (12441, 12442, 0),
+    (12443, 12543, 2),
+    (12549, 12591, 2),
+    (12593, 12686, 2),
+    (12688, 12771, 2),
+    (12783, 12830, 2),
+    (12832, 12871, 2),
+    (12880, 19903, 2),
+    (19968, 42124, 2),
+    (42128, 42182, 2),
+    (42607, 42610, 0),
+    (42612, 42621, 0),
+    (42654, 42655, 0),
+    (42736, 42737, 0),
+    (43010, 43010, 0),
+    (43014, 43014, 0),
+    (43019, 43019, 0),
+    (43043, 43047, 0),
+    (43052, 43052, 0),
+    (43136, 43137, 0),
+    (43188, 43205, 0),
+    (43232, 43249, 0),
+    (43263, 43263, 0),
+    (43302, 43309, 0),
+    (43335, 43347, 0),
+    (43360, 43388, 2),
+    (43392, 43395, 0),
+    (43443, 43456, 0),
+    (43493, 43493, 0),
+    (43561, 43574, 0),
+    (43587, 43587, 0),
+    (43596, 43597, 0),
+    (43643, 43645, 0),
+    (43696, 43696, 0),
+    (43698, 43700, 0),
+    (43703, 43704, 0),
+    (43710, 43711, 0),
+    (43713, 43713, 0),
+    (43755, 43759, 0),
+    (43765, 43766, 0),
+    (44003, 44010, 0),
+    (44012, 44013, 0),
+    (44032, 55203, 2),
+    (55216, 55295, 0),
+    (63744, 64255, 2),
+    (64286, 64286, 0),
+    (65024, 65039, 0),
+    (65040, 65049, 2),
+    (65056, 65071, 0),
+    (65072, 65106, 2),
+    (65108, 65126, 2),
+    (65128, 65131, 2),
+    (65279, 65279, 0),
+    (65281, 65376, 2),
+    (65504, 65510, 2),
+    (65529, 65531, 0),
+    (66045, 66045, 0),
+    (66272, 66272, 0),
+    (66422, 66426, 0),
+    (68097, 68099, 0),
+    (68101, 68102, 0),
+    (68108, 68111, 0),
+    (68152, 68154, 0),
+    (68159, 68159, 0),
+    (68325, 68326, 0),
+    (68900, 68903, 0),
+    (69291, 69292, 0),
+    (69373, 69375, 0),
+    (69446, 69456, 0),
+    (69506, 69509, 0),
+    (69632, 69634, 0),
+    (69688, 69702, 0),
+    (69744, 69744, 0),
+    (69747, 69748, 0),
+    (69759, 69762, 0),
+    (69808, 69818, 0),
+    (69821, 69821, 0),
+    (69826, 69826, 0),
+    (69837, 69837, 0),
+    (69888, 69890, 0),
+    (69927, 69940, 0),
+    (69957, 69958, 0),
+    (70003, 70003, 0),
+    (70016, 70018, 0),
+    (70067, 70080, 0),
+    (70089, 70092, 0),
+    (70094, 70095, 0),
+    (70188, 70199, 0),
+    (70206, 70206, 0),
+    (70209, 70209, 0),
+    (70367, 70378, 0),
+    (70400, 70403, 0),
+    (70459, 70460, 0),
+    (70462, 70468, 0),
+    (70471, 70472, 0),
+    (70475, 70477, 0),
+    (70487, 70487, 0),
+    (70498, 70499, 0),
+    (70502, 70508, 0),
+    (70512, 70516, 0),
+    (70709, 70726, 0),
+    (70750, 70750, 0),
+    (70832, 70851, 0),
+    (71087, 71093, 0),
+    (71096, 71104, 0),
+    (71132, 71133, 0),
+    (71216, 71232, 0),
+    (71339, 71351, 0),
+    (71453, 71467, 0),
+    (71724, 71738, 0),
+    (71984, 71989, 0),
+    (71991, 71992, 0),
+    (71995, 71998, 0),
+    (72000, 72000, 0),
+    (72002, 72003, 0),
+    (72145, 72151, 0),
+    (72154, 72160, 0),
+    (72164, 72164, 0),
+    (72193, 72202, 0),
+    (72243, 72249, 0),
+    (72251, 72254, 0),
+    (72263, 72263, 0),
+    (72273, 72283, 0),
+    (72330, 72345, 0),
+    (72751, 72758, 0),
+    (72760, 72767, 0),
+    (72850, 72871, 0),
+    (72873, 72886, 0),
+    (73009, 73014, 0),
+    (73018, 73018, 0),
+    (73020, 73021, 0),
+    (73023, 73029, 0),
+    (73031, 73031, 0),
+    (73098, 73102, 0),
+    (73104, 73105, 0),
+    (73107, 73111, 0),
+    (73459, 73462, 0),
+    (73472, 73473, 0),
+    (73475, 73475, 0),
+    (73524, 73530, 0),
+    (73534, 73538, 0),
+    (78896, 78912, 0),
+    (78919, 78933, 0),
+    (92912, 92916, 0),
+    (92976, 92982, 0),
+    (94031, 94031, 0),
+    (94033, 94087, 0),
+    (94095, 94098, 0),
+    (94176, 94179, 2),
+    (94180, 94180, 0),
+    (94192, 94193, 0),
+    (94208, 100343, 2),
+    (100352, 101589, 2),
+    (101632, 101640, 2),
+    (110576, 110579, 2),
+    (110581, 110587, 2),
+    (110589, 110590, 2),
+    (110592, 110882, 2),
+    (110898, 110898, 2),
+    (110928, 110930, 2),
+    (110933, 110933, 2),
+    (110948, 110951, 2),
+    (110960, 111355, 2),
+    (113821, 113822, 0),
+    (113824, 113827, 0),
+    (118528, 118573, 0),
+    (118576, 118598, 0),
+    (119141, 119145, 0),
+    (119149, 119170, 0),
+    (119173, 119179, 0),
+    (119210, 119213, 0),
+    (119362, 119364, 0),
+    (121344, 121398, 0),
+    (121403, 121452, 0),
+    (121461, 121461, 0),
+    (121476, 121476, 0),
+    (121499, 121503, 0),
+    (121505, 121519, 0),
+    (122880, 122886, 0),
+    (122888, 122904, 0),
+    (122907, 122913, 0),
+    (122915, 122916, 0),
+    (122918, 122922, 0),
+    (123023, 123023, 0),
+    (123184, 123190, 0),
+    (123566, 123566, 0),
+    (123628, 123631, 0),
+    (124140, 124143, 0),
+    (125136, 125142, 0),
+    (125252, 125258, 0),
+    (126980, 126980, 2),
+    (127183, 127183, 2),
+    (127374, 127374, 2),
+    (127377, 127386, 2),
+    (127488, 127490, 2),
+    (127504, 127547, 2),
+    (127552, 127560, 2),
+    (127568, 127569, 2),
+    (127584, 127589, 2),
+    (127744, 127776, 2),
+    (127789, 127797, 2),
+    (127799, 127868, 2),
+    (127870, 127891, 2),
+    (127904, 127946, 2),
+    (127951, 127955, 2),
+    (127968, 127984, 2),
+    (127988, 127988, 2),
+    (127992, 127994, 2),
+    (127995, 127999, 0),
+    (128000, 128062, 2),
+    (128064, 128064, 2),
+    (128066, 128252, 2),
+    (128255, 128317, 2),
+    (128331, 128334, 2),
+    (128336, 128359, 2),
+    (128378, 128378, 2),
+    (128405, 128406, 2),
+    (128420, 128420, 2),
+    (128507, 128591, 2),
+    (128640, 128709, 2),
+    (128716, 128716, 2),
+    (128720, 128722, 2),
+    (128725, 128727, 2),
+    (128732, 128735, 2),
+    (128747, 128748, 2),
+    (128756, 128764, 2),
+    (128992, 129003, 2),
+    (129008, 129008, 2),
+    (129292, 129338, 2),
+    (129340, 129349, 2),
+    (129351, 129535, 2),
+    (129648, 129660, 2),
+    (129664, 129672, 2),
+    (129680, 129725, 2),
+    (129727, 129733, 2),
+    (129742, 129755, 2),
+    (129760, 129768, 2),
+    (129776, 129784, 2),
+    (131072, 196605, 2),
+    (196608, 262141, 2),
+    (917505, 917505, 0),
+    (917536, 917631, 0),
+    (917760, 917999, 0),
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_codes.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_codes.py
new file mode 100644
index 00000000..1f2877bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_codes.py
@@ -0,0 +1,3610 @@
+EMOJI = {
+    "1st_place_medal": "🥇",
+    "2nd_place_medal": "🥈",
+    "3rd_place_medal": "🥉",
+    "ab_button_(blood_type)": "🆎",
+    "atm_sign": "🏧",
+    "a_button_(blood_type)": "🅰",
+    "afghanistan": "🇦🇫",
+    "albania": "🇦🇱",
+    "algeria": "🇩🇿",
+    "american_samoa": "🇦🇸",
+    "andorra": "🇦🇩",
+    "angola": "🇦🇴",
+    "anguilla": "🇦🇮",
+    "antarctica": "🇦🇶",
+    "antigua_&_barbuda": "🇦🇬",
+    "aquarius": "♒",
+    "argentina": "🇦🇷",
+    "aries": "♈",
+    "armenia": "🇦🇲",
+    "aruba": "🇦🇼",
+    "ascension_island": "🇦🇨",
+    "australia": "🇦🇺",
+    "austria": "🇦🇹",
+    "azerbaijan": "🇦🇿",
+    "back_arrow": "🔙",
+    "b_button_(blood_type)": "🅱",
+    "bahamas": "🇧🇸",
+    "bahrain": "🇧🇭",
+    "bangladesh": "🇧🇩",
+    "barbados": "🇧🇧",
+    "belarus": "🇧🇾",
+    "belgium": "🇧🇪",
+    "belize": "🇧🇿",
+    "benin": "🇧🇯",
+    "bermuda": "🇧🇲",
+    "bhutan": "🇧🇹",
+    "bolivia": "🇧🇴",
+    "bosnia_&_herzegovina": "🇧🇦",
+    "botswana": "🇧🇼",
+    "bouvet_island": "🇧🇻",
+    "brazil": "🇧🇷",
+    "british_indian_ocean_territory": "🇮🇴",
+    "british_virgin_islands": "🇻🇬",
+    "brunei": "🇧🇳",
+    "bulgaria": "🇧🇬",
+    "burkina_faso": "🇧🇫",
+    "burundi": "🇧🇮",
+    "cl_button": "🆑",
+    "cool_button": "🆒",
+    "cambodia": "🇰🇭",
+    "cameroon": "🇨🇲",
+    "canada": "🇨🇦",
+    "canary_islands": "🇮🇨",
+    "cancer": "♋",
+    "cape_verde": "🇨🇻",
+    "capricorn": "♑",
+    "caribbean_netherlands": "🇧🇶",
+    "cayman_islands": "🇰🇾",
+    "central_african_republic": "🇨🇫",
+    "ceuta_&_melilla": "🇪🇦",
+    "chad": "🇹🇩",
+    "chile": "🇨🇱",
+    "china": "🇨🇳",
+    "christmas_island": "🇨🇽",
+    "christmas_tree": "🎄",
+    "clipperton_island": "🇨🇵",
+    "cocos_(keeling)_islands": "🇨🇨",
+    "colombia": "🇨🇴",
+    "comoros": "🇰🇲",
+    "congo_-_brazzaville": "🇨🇬",
+    "congo_-_kinshasa": "🇨🇩",
+    "cook_islands": "🇨🇰",
+    "costa_rica": "🇨🇷",
+    "croatia": "🇭🇷",
+    "cuba": "🇨🇺",
+    "curaçao": "🇨🇼",
+    "cyprus": "🇨🇾",
+    "czechia": "🇨🇿",
+    "côte_d’ivoire": "🇨🇮",
+    "denmark": "🇩🇰",
+    "diego_garcia": "🇩🇬",
+    "djibouti": "🇩🇯",
+    "dominica": "🇩🇲",
+    "dominican_republic": "🇩🇴",
+    "end_arrow": "🔚",
+    "ecuador": "🇪🇨",
+    "egypt": "🇪🇬",
+    "el_salvador": "🇸🇻",
+    "england": "🏴\U000e0067\U000e0062\U000e0065\U000e006e\U000e0067\U000e007f",
+    "equatorial_guinea": "🇬🇶",
+    "eritrea": "🇪🇷",
+    "estonia": "🇪🇪",
+    "ethiopia": "🇪🇹",
+    "european_union": "🇪🇺",
+    "free_button": "🆓",
+    "falkland_islands": "🇫🇰",
+    "faroe_islands": "🇫🇴",
+    "fiji": "🇫🇯",
+    "finland": "🇫🇮",
+    "france": "🇫🇷",
+    "french_guiana": "🇬🇫",
+    "french_polynesia": "🇵🇫",
+    "french_southern_territories": "🇹🇫",
+    "gabon": "🇬🇦",
+    "gambia": "🇬🇲",
+    "gemini": "♊",
+    "georgia": "🇬🇪",
+    "germany": "🇩🇪",
+    "ghana": "🇬🇭",
+    "gibraltar": "🇬🇮",
+    "greece": "🇬🇷",
+    "greenland": "🇬🇱",
+    "grenada": "🇬🇩",
+    "guadeloupe": "🇬🇵",
+    "guam": "🇬🇺",
+    "guatemala": "🇬🇹",
+    "guernsey": "🇬🇬",
+    "guinea": "🇬🇳",
+    "guinea-bissau": "🇬🇼",
+    "guyana": "🇬🇾",
+    "haiti": "🇭🇹",
+    "heard_&_mcdonald_islands": "🇭🇲",
+    "honduras": "🇭🇳",
+    "hong_kong_sar_china": "🇭🇰",
+    "hungary": "🇭🇺",
+    "id_button": "🆔",
+    "iceland": "🇮🇸",
+    "india": "🇮🇳",
+    "indonesia": "🇮🇩",
+    "iran": "🇮🇷",
+    "iraq": "🇮🇶",
+    "ireland": "🇮🇪",
+    "isle_of_man": "🇮🇲",
+    "israel": "🇮🇱",
+    "italy": "🇮🇹",
+    "jamaica": "🇯🇲",
+    "japan": "🗾",
+    "japanese_acceptable_button": "🉑",
+    "japanese_application_button": "🈸",
+    "japanese_bargain_button": "🉐",
+    "japanese_castle": "🏯",
+    "japanese_congratulations_button": "㊗",
+    "japanese_discount_button": "🈹",
+    "japanese_dolls": "🎎",
+    "japanese_free_of_charge_button": "🈚",
+    "japanese_here_button": "🈁",
+    "japanese_monthly_amount_button": "🈷",
+    "japanese_no_vacancy_button": "🈵",
+    "japanese_not_free_of_charge_button": "🈶",
+    "japanese_open_for_business_button": "🈺",
+    "japanese_passing_grade_button": "🈴",
+    "japanese_post_office": "🏣",
+    "japanese_prohibited_button": "🈲",
+    "japanese_reserved_button": "🈯",
+    "japanese_secret_button": "㊙",
+    "japanese_service_charge_button": "🈂",
+    "japanese_symbol_for_beginner": "🔰",
+    "japanese_vacancy_button": "🈳",
+    "jersey": "🇯🇪",
+    "jordan": "🇯🇴",
+    "kazakhstan": "🇰🇿",
+    "kenya": "🇰🇪",
+    "kiribati": "🇰🇮",
+    "kosovo": "🇽🇰",
+    "kuwait": "🇰🇼",
+    "kyrgyzstan": "🇰🇬",
+    "laos": "🇱🇦",
+    "latvia": "🇱🇻",
+    "lebanon": "🇱🇧",
+    "leo": "♌",
+    "lesotho": "🇱🇸",
+    "liberia": "🇱🇷",
+    "libra": "♎",
+    "libya": "🇱🇾",
+    "liechtenstein": "🇱🇮",
+    "lithuania": "🇱🇹",
+    "luxembourg": "🇱🇺",
+    "macau_sar_china": "🇲🇴",
+    "macedonia": "🇲🇰",
+    "madagascar": "🇲🇬",
+    "malawi": "🇲🇼",
+    "malaysia": "🇲🇾",
+    "maldives": "🇲🇻",
+    "mali": "🇲🇱",
+    "malta": "🇲🇹",
+    "marshall_islands": "🇲🇭",
+    "martinique": "🇲🇶",
+    "mauritania": "🇲🇷",
+    "mauritius": "🇲🇺",
+    "mayotte": "🇾🇹",
+    "mexico": "🇲🇽",
+    "micronesia": "🇫🇲",
+    "moldova": "🇲🇩",
+    "monaco": "🇲🇨",
+    "mongolia": "🇲🇳",
+    "montenegro": "🇲🇪",
+    "montserrat": "🇲🇸",
+    "morocco": "🇲🇦",
+    "mozambique": "🇲🇿",
+    "mrs._claus": "🤶",
+    "mrs._claus_dark_skin_tone": "🤶🏿",
+    "mrs._claus_light_skin_tone": "🤶🏻",
+    "mrs._claus_medium-dark_skin_tone": "🤶🏾",
+    "mrs._claus_medium-light_skin_tone": "🤶🏼",
+    "mrs._claus_medium_skin_tone": "🤶🏽",
+    "myanmar_(burma)": "🇲🇲",
+    "new_button": "🆕",
+    "ng_button": "🆖",
+    "namibia": "🇳🇦",
+    "nauru": "🇳🇷",
+    "nepal": "🇳🇵",
+    "netherlands": "🇳🇱",
+    "new_caledonia": "🇳🇨",
+    "new_zealand": "🇳🇿",
+    "nicaragua": "🇳🇮",
+    "niger": "🇳🇪",
+    "nigeria": "🇳🇬",
+    "niue": "🇳🇺",
+    "norfolk_island": "🇳🇫",
+    "north_korea": "🇰🇵",
+    "northern_mariana_islands": "🇲🇵",
+    "norway": "🇳🇴",
+    "ok_button": "🆗",
+    "ok_hand": "👌",
+    "ok_hand_dark_skin_tone": "👌🏿",
+    "ok_hand_light_skin_tone": "👌🏻",
+    "ok_hand_medium-dark_skin_tone": "👌🏾",
+    "ok_hand_medium-light_skin_tone": "👌🏼",
+    "ok_hand_medium_skin_tone": "👌🏽",
+    "on!_arrow": "🔛",
+    "o_button_(blood_type)": "🅾",
+    "oman": "🇴🇲",
+    "ophiuchus": "⛎",
+    "p_button": "🅿",
+    "pakistan": "🇵🇰",
+    "palau": "🇵🇼",
+    "palestinian_territories": "🇵🇸",
+    "panama": "🇵🇦",
+    "papua_new_guinea": "🇵🇬",
+    "paraguay": "🇵🇾",
+    "peru": "🇵🇪",
+    "philippines": "🇵🇭",
+    "pisces": "♓",
+    "pitcairn_islands": "🇵🇳",
+    "poland": "🇵🇱",
+    "portugal": "🇵🇹",
+    "puerto_rico": "🇵🇷",
+    "qatar": "🇶🇦",
+    "romania": "🇷🇴",
+    "russia": "🇷🇺",
+    "rwanda": "🇷🇼",
+    "réunion": "🇷🇪",
+    "soon_arrow": "🔜",
+    "sos_button": "🆘",
+    "sagittarius": "♐",
+    "samoa": "🇼🇸",
+    "san_marino": "🇸🇲",
+    "santa_claus": "🎅",
+    "santa_claus_dark_skin_tone": "🎅🏿",
+    "santa_claus_light_skin_tone": "🎅🏻",
+    "santa_claus_medium-dark_skin_tone": "🎅🏾",
+    "santa_claus_medium-light_skin_tone": "🎅🏼",
+    "santa_claus_medium_skin_tone": "🎅🏽",
+    "saudi_arabia": "🇸🇦",
+    "scorpio": "♏",
+    "scotland": "🏴\U000e0067\U000e0062\U000e0073\U000e0063\U000e0074\U000e007f",
+    "senegal": "🇸🇳",
+    "serbia": "🇷🇸",
+    "seychelles": "🇸🇨",
+    "sierra_leone": "🇸🇱",
+    "singapore": "🇸🇬",
+    "sint_maarten": "🇸🇽",
+    "slovakia": "🇸🇰",
+    "slovenia": "🇸🇮",
+    "solomon_islands": "🇸🇧",
+    "somalia": "🇸🇴",
+    "south_africa": "🇿🇦",
+    "south_georgia_&_south_sandwich_islands": "🇬🇸",
+    "south_korea": "🇰🇷",
+    "south_sudan": "🇸🇸",
+    "spain": "🇪🇸",
+    "sri_lanka": "🇱🇰",
+    "st._barthélemy": "🇧🇱",
+    "st._helena": "🇸🇭",
+    "st._kitts_&_nevis": "🇰🇳",
+    "st._lucia": "🇱🇨",
+    "st._martin": "🇲🇫",
+    "st._pierre_&_miquelon": "🇵🇲",
+    "st._vincent_&_grenadines": "🇻🇨",
+    "statue_of_liberty": "🗽",
+    "sudan": "🇸🇩",
+    "suriname": "🇸🇷",
+    "svalbard_&_jan_mayen": "🇸🇯",
+    "swaziland": "🇸🇿",
+    "sweden": "🇸🇪",
+    "switzerland": "🇨🇭",
+    "syria": "🇸🇾",
+    "são_tomé_&_príncipe": "🇸🇹",
+    "t-rex": "🦖",
+    "top_arrow": "🔝",
+    "taiwan": "🇹🇼",
+    "tajikistan": "🇹🇯",
+    "tanzania": "🇹🇿",
+    "taurus": "♉",
+    "thailand": "🇹🇭",
+    "timor-leste": "🇹🇱",
+    "togo": "🇹🇬",
+    "tokelau": "🇹🇰",
+    "tokyo_tower": "🗼",
+    "tonga": "🇹🇴",
+    "trinidad_&_tobago": "🇹🇹",
+    "tristan_da_cunha": "🇹🇦",
+    "tunisia": "🇹🇳",
+    "turkey": "🦃",
+    "turkmenistan": "🇹🇲",
+    "turks_&_caicos_islands": "🇹🇨",
+    "tuvalu": "🇹🇻",
+    "u.s._outlying_islands": "🇺🇲",
+    "u.s._virgin_islands": "🇻🇮",
+    "up!_button": "🆙",
+    "uganda": "🇺🇬",
+    "ukraine": "🇺🇦",
+    "united_arab_emirates": "🇦🇪",
+    "united_kingdom": "🇬🇧",
+    "united_nations": "🇺🇳",
+    "united_states": "🇺🇸",
+    "uruguay": "🇺🇾",
+    "uzbekistan": "🇺🇿",
+    "vs_button": "🆚",
+    "vanuatu": "🇻🇺",
+    "vatican_city": "🇻🇦",
+    "venezuela": "🇻🇪",
+    "vietnam": "🇻🇳",
+    "virgo": "♍",
+    "wales": "🏴\U000e0067\U000e0062\U000e0077\U000e006c\U000e0073\U000e007f",
+    "wallis_&_futuna": "🇼🇫",
+    "western_sahara": "🇪🇭",
+    "yemen": "🇾🇪",
+    "zambia": "🇿🇲",
+    "zimbabwe": "🇿🇼",
+    "abacus": "🧮",
+    "adhesive_bandage": "🩹",
+    "admission_tickets": "🎟",
+    "adult": "🧑",
+    "adult_dark_skin_tone": "🧑🏿",
+    "adult_light_skin_tone": "🧑🏻",
+    "adult_medium-dark_skin_tone": "🧑🏾",
+    "adult_medium-light_skin_tone": "🧑🏼",
+    "adult_medium_skin_tone": "🧑🏽",
+    "aerial_tramway": "🚡",
+    "airplane": "✈",
+    "airplane_arrival": "🛬",
+    "airplane_departure": "🛫",
+    "alarm_clock": "⏰",
+    "alembic": "⚗",
+    "alien": "👽",
+    "alien_monster": "👾",
+    "ambulance": "🚑",
+    "american_football": "🏈",
+    "amphora": "🏺",
+    "anchor": "⚓",
+    "anger_symbol": "💢",
+    "angry_face": "😠",
+    "angry_face_with_horns": "👿",
+    "anguished_face": "😧",
+    "ant": "🐜",
+    "antenna_bars": "📶",
+    "anxious_face_with_sweat": "😰",
+    "articulated_lorry": "🚛",
+    "artist_palette": "🎨",
+    "astonished_face": "😲",
+    "atom_symbol": "⚛",
+    "auto_rickshaw": "🛺",
+    "automobile": "🚗",
+    "avocado": "🥑",
+    "axe": "🪓",
+    "baby": "👶",
+    "baby_angel": "👼",
+    "baby_angel_dark_skin_tone": "👼🏿",
+    "baby_angel_light_skin_tone": "👼🏻",
+    "baby_angel_medium-dark_skin_tone": "👼🏾",
+    "baby_angel_medium-light_skin_tone": "👼🏼",
+    "baby_angel_medium_skin_tone": "👼🏽",
+    "baby_bottle": "🍼",
+    "baby_chick": "🐤",
+    "baby_dark_skin_tone": "👶🏿",
+    "baby_light_skin_tone": "👶🏻",
+    "baby_medium-dark_skin_tone": "👶🏾",
+    "baby_medium-light_skin_tone": "👶🏼",
+    "baby_medium_skin_tone": "👶🏽",
+    "baby_symbol": "🚼",
+    "backhand_index_pointing_down": "👇",
+    "backhand_index_pointing_down_dark_skin_tone": "👇🏿",
+    "backhand_index_pointing_down_light_skin_tone": "👇🏻",
+    "backhand_index_pointing_down_medium-dark_skin_tone": "👇🏾",
+    "backhand_index_pointing_down_medium-light_skin_tone": "👇🏼",
+    "backhand_index_pointing_down_medium_skin_tone": "👇🏽",
+    "backhand_index_pointing_left": "👈",
+    "backhand_index_pointing_left_dark_skin_tone": "👈🏿",
+    "backhand_index_pointing_left_light_skin_tone": "👈🏻",
+    "backhand_index_pointing_left_medium-dark_skin_tone": "👈🏾",
+    "backhand_index_pointing_left_medium-light_skin_tone": "👈🏼",
+    "backhand_index_pointing_left_medium_skin_tone": "👈🏽",
+    "backhand_index_pointing_right": "👉",
+    "backhand_index_pointing_right_dark_skin_tone": "👉🏿",
+    "backhand_index_pointing_right_light_skin_tone": "👉🏻",
+    "backhand_index_pointing_right_medium-dark_skin_tone": "👉🏾",
+    "backhand_index_pointing_right_medium-light_skin_tone": "👉🏼",
+    "backhand_index_pointing_right_medium_skin_tone": "👉🏽",
+    "backhand_index_pointing_up": "👆",
+    "backhand_index_pointing_up_dark_skin_tone": "👆🏿",
+    "backhand_index_pointing_up_light_skin_tone": "👆🏻",
+    "backhand_index_pointing_up_medium-dark_skin_tone": "👆🏾",
+    "backhand_index_pointing_up_medium-light_skin_tone": "👆🏼",
+    "backhand_index_pointing_up_medium_skin_tone": "👆🏽",
+    "bacon": "🥓",
+    "badger": "🦡",
+    "badminton": "🏸",
+    "bagel": "🥯",
+    "baggage_claim": "🛄",
+    "baguette_bread": "🥖",
+    "balance_scale": "⚖",
+    "bald": "🦲",
+    "bald_man": "👨\u200d🦲",
+    "bald_woman": "👩\u200d🦲",
+    "ballet_shoes": "🩰",
+    "balloon": "🎈",
+    "ballot_box_with_ballot": "🗳",
+    "ballot_box_with_check": "☑",
+    "banana": "🍌",
+    "banjo": "🪕",
+    "bank": "🏦",
+    "bar_chart": "📊",
+    "barber_pole": "💈",
+    "baseball": "⚾",
+    "basket": "🧺",
+    "basketball": "🏀",
+    "bat": "🦇",
+    "bathtub": "🛁",
+    "battery": "🔋",
+    "beach_with_umbrella": "🏖",
+    "beaming_face_with_smiling_eyes": "😁",
+    "bear_face": "🐻",
+    "bearded_person": "🧔",
+    "bearded_person_dark_skin_tone": "🧔🏿",
+    "bearded_person_light_skin_tone": "🧔🏻",
+    "bearded_person_medium-dark_skin_tone": "🧔🏾",
+    "bearded_person_medium-light_skin_tone": "🧔🏼",
+    "bearded_person_medium_skin_tone": "🧔🏽",
+    "beating_heart": "💓",
+    "bed": "🛏",
+    "beer_mug": "🍺",
+    "bell": "🔔",
+    "bell_with_slash": "🔕",
+    "bellhop_bell": "🛎",
+    "bento_box": "🍱",
+    "beverage_box": "🧃",
+    "bicycle": "🚲",
+    "bikini": "👙",
+    "billed_cap": "🧢",
+    "biohazard": "☣",
+    "bird": "🐦",
+    "birthday_cake": "🎂",
+    "black_circle": "⚫",
+    "black_flag": "🏴",
+    "black_heart": "🖤",
+    "black_large_square": "⬛",
+    "black_medium-small_square": "◾",
+    "black_medium_square": "◼",
+    "black_nib": "✒",
+    "black_small_square": "▪",
+    "black_square_button": "🔲",
+    "blond-haired_man": "👱\u200d♂️",
+    "blond-haired_man_dark_skin_tone": "👱🏿\u200d♂️",
+    "blond-haired_man_light_skin_tone": "👱🏻\u200d♂️",
+    "blond-haired_man_medium-dark_skin_tone": "👱🏾\u200d♂️",
+    "blond-haired_man_medium-light_skin_tone": "👱🏼\u200d♂️",
+    "blond-haired_man_medium_skin_tone": "👱🏽\u200d♂️",
+    "blond-haired_person": "👱",
+    "blond-haired_person_dark_skin_tone": "👱🏿",
+    "blond-haired_person_light_skin_tone": "👱🏻",
+    "blond-haired_person_medium-dark_skin_tone": "👱🏾",
+    "blond-haired_person_medium-light_skin_tone": "👱🏼",
+    "blond-haired_person_medium_skin_tone": "👱🏽",
+    "blond-haired_woman": "👱\u200d♀️",
+    "blond-haired_woman_dark_skin_tone": "👱🏿\u200d♀️",
+    "blond-haired_woman_light_skin_tone": "👱🏻\u200d♀️",
+    "blond-haired_woman_medium-dark_skin_tone": "👱🏾\u200d♀️",
+    "blond-haired_woman_medium-light_skin_tone": "👱🏼\u200d♀️",
+    "blond-haired_woman_medium_skin_tone": "👱🏽\u200d♀️",
+    "blossom": "🌼",
+    "blowfish": "🐡",
+    "blue_book": "📘",
+    "blue_circle": "🔵",
+    "blue_heart": "💙",
+    "blue_square": "🟦",
+    "boar": "🐗",
+    "bomb": "💣",
+    "bone": "🦴",
+    "bookmark": "🔖",
+    "bookmark_tabs": "📑",
+    "books": "📚",
+    "bottle_with_popping_cork": "🍾",
+    "bouquet": "💐",
+    "bow_and_arrow": "🏹",
+    "bowl_with_spoon": "🥣",
+    "bowling": "🎳",
+    "boxing_glove": "🥊",
+    "boy": "👦",
+    "boy_dark_skin_tone": "👦🏿",
+    "boy_light_skin_tone": "👦🏻",
+    "boy_medium-dark_skin_tone": "👦🏾",
+    "boy_medium-light_skin_tone": "👦🏼",
+    "boy_medium_skin_tone": "👦🏽",
+    "brain": "🧠",
+    "bread": "🍞",
+    "breast-feeding": "🤱",
+    "breast-feeding_dark_skin_tone": "🤱🏿",
+    "breast-feeding_light_skin_tone": "🤱🏻",
+    "breast-feeding_medium-dark_skin_tone": "🤱🏾",
+    "breast-feeding_medium-light_skin_tone": "🤱🏼",
+    "breast-feeding_medium_skin_tone": "🤱🏽",
+    "brick": "🧱",
+    "bride_with_veil": "👰",
+    "bride_with_veil_dark_skin_tone": "👰🏿",
+    "bride_with_veil_light_skin_tone": "👰🏻",
+    "bride_with_veil_medium-dark_skin_tone": "👰🏾",
+    "bride_with_veil_medium-light_skin_tone": "👰🏼",
+    "bride_with_veil_medium_skin_tone": "👰🏽",
+    "bridge_at_night": "🌉",
+    "briefcase": "💼",
+    "briefs": "🩲",
+    "bright_button": "🔆",
+    "broccoli": "🥦",
+    "broken_heart": "💔",
+    "broom": "🧹",
+    "brown_circle": "🟤",
+    "brown_heart": "🤎",
+    "brown_square": "🟫",
+    "bug": "🐛",
+    "building_construction": "🏗",
+    "bullet_train": "🚅",
+    "burrito": "🌯",
+    "bus": "🚌",
+    "bus_stop": "🚏",
+    "bust_in_silhouette": "👤",
+    "busts_in_silhouette": "👥",
+    "butter": "🧈",
+    "butterfly": "🦋",
+    "cactus": "🌵",
+    "calendar": "📆",
+    "call_me_hand": "🤙",
+    "call_me_hand_dark_skin_tone": "🤙🏿",
+    "call_me_hand_light_skin_tone": "🤙🏻",
+    "call_me_hand_medium-dark_skin_tone": "🤙🏾",
+    "call_me_hand_medium-light_skin_tone": "🤙🏼",
+    "call_me_hand_medium_skin_tone": "🤙🏽",
+    "camel": "🐫",
+    "camera": "📷",
+    "camera_with_flash": "📸",
+    "camping": "🏕",
+    "candle": "🕯",
+    "candy": "🍬",
+    "canned_food": "🥫",
+    "canoe": "🛶",
+    "card_file_box": "🗃",
+    "card_index": "📇",
+    "card_index_dividers": "🗂",
+    "carousel_horse": "🎠",
+    "carp_streamer": "🎏",
+    "carrot": "🥕",
+    "castle": "🏰",
+    "cat": "🐱",
+    "cat_face": "🐱",
+    "cat_face_with_tears_of_joy": "😹",
+    "cat_face_with_wry_smile": "😼",
+    "chains": "⛓",
+    "chair": "🪑",
+    "chart_decreasing": "📉",
+    "chart_increasing": "📈",
+    "chart_increasing_with_yen": "💹",
+    "cheese_wedge": "🧀",
+    "chequered_flag": "🏁",
+    "cherries": "🍒",
+    "cherry_blossom": "🌸",
+    "chess_pawn": "♟",
+    "chestnut": "🌰",
+    "chicken": "🐔",
+    "child": "🧒",
+    "child_dark_skin_tone": "🧒🏿",
+    "child_light_skin_tone": "🧒🏻",
+    "child_medium-dark_skin_tone": "🧒🏾",
+    "child_medium-light_skin_tone": "🧒🏼",
+    "child_medium_skin_tone": "🧒🏽",
+    "children_crossing": "🚸",
+    "chipmunk": "🐿",
+    "chocolate_bar": "🍫",
+    "chopsticks": "🥢",
+    "church": "⛪",
+    "cigarette": "🚬",
+    "cinema": "🎦",
+    "circled_m": "Ⓜ",
+    "circus_tent": "🎪",
+    "cityscape": "🏙",
+    "cityscape_at_dusk": "🌆",
+    "clamp": "🗜",
+    "clapper_board": "🎬",
+    "clapping_hands": "👏",
+    "clapping_hands_dark_skin_tone": "👏🏿",
+    "clapping_hands_light_skin_tone": "👏🏻",
+    "clapping_hands_medium-dark_skin_tone": "👏🏾",
+    "clapping_hands_medium-light_skin_tone": "👏🏼",
+    "clapping_hands_medium_skin_tone": "👏🏽",
+    "classical_building": "🏛",
+    "clinking_beer_mugs": "🍻",
+    "clinking_glasses": "🥂",
+    "clipboard": "📋",
+    "clockwise_vertical_arrows": "🔃",
+    "closed_book": "📕",
+    "closed_mailbox_with_lowered_flag": "📪",
+    "closed_mailbox_with_raised_flag": "📫",
+    "closed_umbrella": "🌂",
+    "cloud": "☁",
+    "cloud_with_lightning": "🌩",
+    "cloud_with_lightning_and_rain": "⛈",
+    "cloud_with_rain": "🌧",
+    "cloud_with_snow": "🌨",
+    "clown_face": "🤡",
+    "club_suit": "♣",
+    "clutch_bag": "👝",
+    "coat": "🧥",
+    "cocktail_glass": "🍸",
+    "coconut": "🥥",
+    "coffin": "⚰",
+    "cold_face": "🥶",
+    "collision": "💥",
+    "comet": "☄",
+    "compass": "🧭",
+    "computer_disk": "💽",
+    "computer_mouse": "🖱",
+    "confetti_ball": "🎊",
+    "confounded_face": "😖",
+    "confused_face": "😕",
+    "construction": "🚧",
+    "construction_worker": "👷",
+    "construction_worker_dark_skin_tone": "👷🏿",
+    "construction_worker_light_skin_tone": "👷🏻",
+    "construction_worker_medium-dark_skin_tone": "👷🏾",
+    "construction_worker_medium-light_skin_tone": "👷🏼",
+    "construction_worker_medium_skin_tone": "👷🏽",
+    "control_knobs": "🎛",
+    "convenience_store": "🏪",
+    "cooked_rice": "🍚",
+    "cookie": "🍪",
+    "cooking": "🍳",
+    "copyright": "©",
+    "couch_and_lamp": "🛋",
+    "counterclockwise_arrows_button": "🔄",
+    "couple_with_heart": "💑",
+    "couple_with_heart_man_man": "👨\u200d❤️\u200d👨",
+    "couple_with_heart_woman_man": "👩\u200d❤️\u200d👨",
+    "couple_with_heart_woman_woman": "👩\u200d❤️\u200d👩",
+    "cow": "🐮",
+    "cow_face": "🐮",
+    "cowboy_hat_face": "🤠",
+    "crab": "🦀",
+    "crayon": "🖍",
+    "credit_card": "💳",
+    "crescent_moon": "🌙",
+    "cricket": "🦗",
+    "cricket_game": "🏏",
+    "crocodile": "🐊",
+    "croissant": "🥐",
+    "cross_mark": "❌",
+    "cross_mark_button": "❎",
+    "crossed_fingers": "🤞",
+    "crossed_fingers_dark_skin_tone": "🤞🏿",
+    "crossed_fingers_light_skin_tone": "🤞🏻",
+    "crossed_fingers_medium-dark_skin_tone": "🤞🏾",
+    "crossed_fingers_medium-light_skin_tone": "🤞🏼",
+    "crossed_fingers_medium_skin_tone": "🤞🏽",
+    "crossed_flags": "🎌",
+    "crossed_swords": "⚔",
+    "crown": "👑",
+    "crying_cat_face": "😿",
+    "crying_face": "😢",
+    "crystal_ball": "🔮",
+    "cucumber": "🥒",
+    "cupcake": "🧁",
+    "cup_with_straw": "🥤",
+    "curling_stone": "🥌",
+    "curly_hair": "🦱",
+    "curly-haired_man": "👨\u200d🦱",
+    "curly-haired_woman": "👩\u200d🦱",
+    "curly_loop": "➰",
+    "currency_exchange": "💱",
+    "curry_rice": "🍛",
+    "custard": "🍮",
+    "customs": "🛃",
+    "cut_of_meat": "🥩",
+    "cyclone": "🌀",
+    "dagger": "🗡",
+    "dango": "🍡",
+    "dashing_away": "💨",
+    "deaf_person": "🧏",
+    "deciduous_tree": "🌳",
+    "deer": "🦌",
+    "delivery_truck": "🚚",
+    "department_store": "🏬",
+    "derelict_house": "🏚",
+    "desert": "🏜",
+    "desert_island": "🏝",
+    "desktop_computer": "🖥",
+    "detective": "🕵",
+    "detective_dark_skin_tone": "🕵🏿",
+    "detective_light_skin_tone": "🕵🏻",
+    "detective_medium-dark_skin_tone": "🕵🏾",
+    "detective_medium-light_skin_tone": "🕵🏼",
+    "detective_medium_skin_tone": "🕵🏽",
+    "diamond_suit": "♦",
+    "diamond_with_a_dot": "💠",
+    "dim_button": "🔅",
+    "direct_hit": "🎯",
+    "disappointed_face": "😞",
+    "diving_mask": "🤿",
+    "diya_lamp": "🪔",
+    "dizzy": "💫",
+    "dizzy_face": "😵",
+    "dna": "🧬",
+    "dog": "🐶",
+    "dog_face": "🐶",
+    "dollar_banknote": "💵",
+    "dolphin": "🐬",
+    "door": "🚪",
+    "dotted_six-pointed_star": "🔯",
+    "double_curly_loop": "➿",
+    "double_exclamation_mark": "‼",
+    "doughnut": "🍩",
+    "dove": "🕊",
+    "down-left_arrow": "↙",
+    "down-right_arrow": "↘",
+    "down_arrow": "⬇",
+    "downcast_face_with_sweat": "😓",
+    "downwards_button": "🔽",
+    "dragon": "🐉",
+    "dragon_face": "🐲",
+    "dress": "👗",
+    "drooling_face": "🤤",
+    "drop_of_blood": "🩸",
+    "droplet": "💧",
+    "drum": "🥁",
+    "duck": "🦆",
+    "dumpling": "🥟",
+    "dvd": "📀",
+    "e-mail": "📧",
+    "eagle": "🦅",
+    "ear": "👂",
+    "ear_dark_skin_tone": "👂🏿",
+    "ear_light_skin_tone": "👂🏻",
+    "ear_medium-dark_skin_tone": "👂🏾",
+    "ear_medium-light_skin_tone": "👂🏼",
+    "ear_medium_skin_tone": "👂🏽",
+    "ear_of_corn": "🌽",
+    "ear_with_hearing_aid": "🦻",
+    "egg": "🍳",
+    "eggplant": "🍆",
+    "eight-pointed_star": "✴",
+    "eight-spoked_asterisk": "✳",
+    "eight-thirty": "🕣",
+    "eight_o’clock": "🕗",
+    "eject_button": "⏏",
+    "electric_plug": "🔌",
+    "elephant": "🐘",
+    "eleven-thirty": "🕦",
+    "eleven_o’clock": "🕚",
+    "elf": "🧝",
+    "elf_dark_skin_tone": "🧝🏿",
+    "elf_light_skin_tone": "🧝🏻",
+    "elf_medium-dark_skin_tone": "🧝🏾",
+    "elf_medium-light_skin_tone": "🧝🏼",
+    "elf_medium_skin_tone": "🧝🏽",
+    "envelope": "✉",
+    "envelope_with_arrow": "📩",
+    "euro_banknote": "💶",
+    "evergreen_tree": "🌲",
+    "ewe": "🐑",
+    "exclamation_mark": "❗",
+    "exclamation_question_mark": "⁉",
+    "exploding_head": "🤯",
+    "expressionless_face": "😑",
+    "eye": "👁",
+    "eye_in_speech_bubble": "👁️\u200d🗨️",
+    "eyes": "👀",
+    "face_blowing_a_kiss": "😘",
+    "face_savoring_food": "😋",
+    "face_screaming_in_fear": "😱",
+    "face_vomiting": "🤮",
+    "face_with_hand_over_mouth": "🤭",
+    "face_with_head-bandage": "🤕",
+    "face_with_medical_mask": "😷",
+    "face_with_monocle": "🧐",
+    "face_with_open_mouth": "😮",
+    "face_with_raised_eyebrow": "🤨",
+    "face_with_rolling_eyes": "🙄",
+    "face_with_steam_from_nose": "😤",
+    "face_with_symbols_on_mouth": "🤬",
+    "face_with_tears_of_joy": "😂",
+    "face_with_thermometer": "🤒",
+    "face_with_tongue": "😛",
+    "face_without_mouth": "😶",
+    "factory": "🏭",
+    "fairy": "🧚",
+    "fairy_dark_skin_tone": "🧚🏿",
+    "fairy_light_skin_tone": "🧚🏻",
+    "fairy_medium-dark_skin_tone": "🧚🏾",
+    "fairy_medium-light_skin_tone": "🧚🏼",
+    "fairy_medium_skin_tone": "🧚🏽",
+    "falafel": "🧆",
+    "fallen_leaf": "🍂",
+    "family": "👪",
+    "family_man_boy": "👨\u200d👦",
+    "family_man_boy_boy": "👨\u200d👦\u200d👦",
+    "family_man_girl": "👨\u200d👧",
+    "family_man_girl_boy": "👨\u200d👧\u200d👦",
+    "family_man_girl_girl": "👨\u200d👧\u200d👧",
+    "family_man_man_boy": "👨\u200d👨\u200d👦",
+    "family_man_man_boy_boy": "👨\u200d👨\u200d👦\u200d👦",
+    "family_man_man_girl": "👨\u200d👨\u200d👧",
+    "family_man_man_girl_boy": "👨\u200d👨\u200d👧\u200d👦",
+    "family_man_man_girl_girl": "👨\u200d👨\u200d👧\u200d👧",
+    "family_man_woman_boy": "👨\u200d👩\u200d👦",
+    "family_man_woman_boy_boy": "👨\u200d👩\u200d👦\u200d👦",
+    "family_man_woman_girl": "👨\u200d👩\u200d👧",
+    "family_man_woman_girl_boy": "👨\u200d👩\u200d👧\u200d👦",
+    "family_man_woman_girl_girl": "👨\u200d👩\u200d👧\u200d👧",
+    "family_woman_boy": "👩\u200d👦",
+    "family_woman_boy_boy": "👩\u200d👦\u200d👦",
+    "family_woman_girl": "👩\u200d👧",
+    "family_woman_girl_boy": "👩\u200d👧\u200d👦",
+    "family_woman_girl_girl": "👩\u200d👧\u200d👧",
+    "family_woman_woman_boy": "👩\u200d👩\u200d👦",
+    "family_woman_woman_boy_boy": "👩\u200d👩\u200d👦\u200d👦",
+    "family_woman_woman_girl": "👩\u200d👩\u200d👧",
+    "family_woman_woman_girl_boy": "👩\u200d👩\u200d👧\u200d👦",
+    "family_woman_woman_girl_girl": "👩\u200d👩\u200d👧\u200d👧",
+    "fast-forward_button": "⏩",
+    "fast_down_button": "⏬",
+    "fast_reverse_button": "⏪",
+    "fast_up_button": "⏫",
+    "fax_machine": "📠",
+    "fearful_face": "😨",
+    "female_sign": "♀",
+    "ferris_wheel": "🎡",
+    "ferry": "⛴",
+    "field_hockey": "🏑",
+    "file_cabinet": "🗄",
+    "file_folder": "📁",
+    "film_frames": "🎞",
+    "film_projector": "📽",
+    "fire": "🔥",
+    "fire_extinguisher": "🧯",
+    "firecracker": "🧨",
+    "fire_engine": "🚒",
+    "fireworks": "🎆",
+    "first_quarter_moon": "🌓",
+    "first_quarter_moon_face": "🌛",
+    "fish": "🐟",
+    "fish_cake_with_swirl": "🍥",
+    "fishing_pole": "🎣",
+    "five-thirty": "🕠",
+    "five_o’clock": "🕔",
+    "flag_in_hole": "⛳",
+    "flamingo": "🦩",
+    "flashlight": "🔦",
+    "flat_shoe": "🥿",
+    "fleur-de-lis": "⚜",
+    "flexed_biceps": "💪",
+    "flexed_biceps_dark_skin_tone": "💪🏿",
+    "flexed_biceps_light_skin_tone": "💪🏻",
+    "flexed_biceps_medium-dark_skin_tone": "💪🏾",
+    "flexed_biceps_medium-light_skin_tone": "💪🏼",
+    "flexed_biceps_medium_skin_tone": "💪🏽",
+    "floppy_disk": "💾",
+    "flower_playing_cards": "🎴",
+    "flushed_face": "😳",
+    "flying_disc": "🥏",
+    "flying_saucer": "🛸",
+    "fog": "🌫",
+    "foggy": "🌁",
+    "folded_hands": "🙏",
+    "folded_hands_dark_skin_tone": "🙏🏿",
+    "folded_hands_light_skin_tone": "🙏🏻",
+    "folded_hands_medium-dark_skin_tone": "🙏🏾",
+    "folded_hands_medium-light_skin_tone": "🙏🏼",
+    "folded_hands_medium_skin_tone": "🙏🏽",
+    "foot": "🦶",
+    "footprints": "👣",
+    "fork_and_knife": "🍴",
+    "fork_and_knife_with_plate": "🍽",
+    "fortune_cookie": "🥠",
+    "fountain": "⛲",
+    "fountain_pen": "🖋",
+    "four-thirty": "🕟",
+    "four_leaf_clover": "🍀",
+    "four_o’clock": "🕓",
+    "fox_face": "🦊",
+    "framed_picture": "🖼",
+    "french_fries": "🍟",
+    "fried_shrimp": "🍤",
+    "frog_face": "🐸",
+    "front-facing_baby_chick": "🐥",
+    "frowning_face": "☹",
+    "frowning_face_with_open_mouth": "😦",
+    "fuel_pump": "⛽",
+    "full_moon": "🌕",
+    "full_moon_face": "🌝",
+    "funeral_urn": "⚱",
+    "game_die": "🎲",
+    "garlic": "🧄",
+    "gear": "⚙",
+    "gem_stone": "💎",
+    "genie": "🧞",
+    "ghost": "👻",
+    "giraffe": "🦒",
+    "girl": "👧",
+    "girl_dark_skin_tone": "👧🏿",
+    "girl_light_skin_tone": "👧🏻",
+    "girl_medium-dark_skin_tone": "👧🏾",
+    "girl_medium-light_skin_tone": "👧🏼",
+    "girl_medium_skin_tone": "👧🏽",
+    "glass_of_milk": "🥛",
+    "glasses": "👓",
+    "globe_showing_americas": "🌎",
+    "globe_showing_asia-australia": "🌏",
+    "globe_showing_europe-africa": "🌍",
+    "globe_with_meridians": "🌐",
+    "gloves": "🧤",
+    "glowing_star": "🌟",
+    "goal_net": "🥅",
+    "goat": "🐐",
+    "goblin": "👺",
+    "goggles": "🥽",
+    "gorilla": "🦍",
+    "graduation_cap": "🎓",
+    "grapes": "🍇",
+    "green_apple": "🍏",
+    "green_book": "📗",
+    "green_circle": "🟢",
+    "green_heart": "💚",
+    "green_salad": "🥗",
+    "green_square": "🟩",
+    "grimacing_face": "😬",
+    "grinning_cat_face": "😺",
+    "grinning_cat_face_with_smiling_eyes": "😸",
+    "grinning_face": "😀",
+    "grinning_face_with_big_eyes": "😃",
+    "grinning_face_with_smiling_eyes": "😄",
+    "grinning_face_with_sweat": "😅",
+    "grinning_squinting_face": "😆",
+    "growing_heart": "💗",
+    "guard": "💂",
+    "guard_dark_skin_tone": "💂🏿",
+    "guard_light_skin_tone": "💂🏻",
+    "guard_medium-dark_skin_tone": "💂🏾",
+    "guard_medium-light_skin_tone": "💂🏼",
+    "guard_medium_skin_tone": "💂🏽",
+    "guide_dog": "🦮",
+    "guitar": "🎸",
+    "hamburger": "🍔",
+    "hammer": "🔨",
+    "hammer_and_pick": "⚒",
+    "hammer_and_wrench": "🛠",
+    "hamster_face": "🐹",
+    "hand_with_fingers_splayed": "🖐",
+    "hand_with_fingers_splayed_dark_skin_tone": "🖐🏿",
+    "hand_with_fingers_splayed_light_skin_tone": "🖐🏻",
+    "hand_with_fingers_splayed_medium-dark_skin_tone": "🖐🏾",
+    "hand_with_fingers_splayed_medium-light_skin_tone": "🖐🏼",
+    "hand_with_fingers_splayed_medium_skin_tone": "🖐🏽",
+    "handbag": "👜",
+    "handshake": "🤝",
+    "hatching_chick": "🐣",
+    "headphone": "🎧",
+    "hear-no-evil_monkey": "🙉",
+    "heart_decoration": "💟",
+    "heart_suit": "♥",
+    "heart_with_arrow": "💘",
+    "heart_with_ribbon": "💝",
+    "heavy_check_mark": "✔",
+    "heavy_division_sign": "➗",
+    "heavy_dollar_sign": "💲",
+    "heavy_heart_exclamation": "❣",
+    "heavy_large_circle": "⭕",
+    "heavy_minus_sign": "➖",
+    "heavy_multiplication_x": "✖",
+    "heavy_plus_sign": "➕",
+    "hedgehog": "🦔",
+    "helicopter": "🚁",
+    "herb": "🌿",
+    "hibiscus": "🌺",
+    "high-heeled_shoe": "👠",
+    "high-speed_train": "🚄",
+    "high_voltage": "⚡",
+    "hiking_boot": "🥾",
+    "hindu_temple": "🛕",
+    "hippopotamus": "🦛",
+    "hole": "🕳",
+    "honey_pot": "🍯",
+    "honeybee": "🐝",
+    "horizontal_traffic_light": "🚥",
+    "horse": "🐴",
+    "horse_face": "🐴",
+    "horse_racing": "🏇",
+    "horse_racing_dark_skin_tone": "🏇🏿",
+    "horse_racing_light_skin_tone": "🏇🏻",
+    "horse_racing_medium-dark_skin_tone": "🏇🏾",
+    "horse_racing_medium-light_skin_tone": "🏇🏼",
+    "horse_racing_medium_skin_tone": "🏇🏽",
+    "hospital": "🏥",
+    "hot_beverage": "☕",
+    "hot_dog": "🌭",
+    "hot_face": "🥵",
+    "hot_pepper": "🌶",
+    "hot_springs": "♨",
+    "hotel": "🏨",
+    "hourglass_done": "⌛",
+    "hourglass_not_done": "⏳",
+    "house": "🏠",
+    "house_with_garden": "🏡",
+    "houses": "🏘",
+    "hugging_face": "🤗",
+    "hundred_points": "💯",
+    "hushed_face": "😯",
+    "ice": "🧊",
+    "ice_cream": "🍨",
+    "ice_hockey": "🏒",
+    "ice_skate": "⛸",
+    "inbox_tray": "📥",
+    "incoming_envelope": "📨",
+    "index_pointing_up": "☝",
+    "index_pointing_up_dark_skin_tone": "☝🏿",
+    "index_pointing_up_light_skin_tone": "☝🏻",
+    "index_pointing_up_medium-dark_skin_tone": "☝🏾",
+    "index_pointing_up_medium-light_skin_tone": "☝🏼",
+    "index_pointing_up_medium_skin_tone": "☝🏽",
+    "infinity": "♾",
+    "information": "ℹ",
+    "input_latin_letters": "🔤",
+    "input_latin_lowercase": "🔡",
+    "input_latin_uppercase": "🔠",
+    "input_numbers": "🔢",
+    "input_symbols": "🔣",
+    "jack-o-lantern": "🎃",
+    "jeans": "👖",
+    "jigsaw": "🧩",
+    "joker": "🃏",
+    "joystick": "🕹",
+    "kaaba": "🕋",
+    "kangaroo": "🦘",
+    "key": "🔑",
+    "keyboard": "⌨",
+    "keycap_#": "#️⃣",
+    "keycap_*": "*️⃣",
+    "keycap_0": "0️⃣",
+    "keycap_1": "1️⃣",
+    "keycap_10": "🔟",
+    "keycap_2": "2️⃣",
+    "keycap_3": "3️⃣",
+    "keycap_4": "4️⃣",
+    "keycap_5": "5️⃣",
+    "keycap_6": "6️⃣",
+    "keycap_7": "7️⃣",
+    "keycap_8": "8️⃣",
+    "keycap_9": "9️⃣",
+    "kick_scooter": "🛴",
+    "kimono": "👘",
+    "kiss": "💋",
+    "kiss_man_man": "👨\u200d❤️\u200d💋\u200d👨",
+    "kiss_mark": "💋",
+    "kiss_woman_man": "👩\u200d❤️\u200d💋\u200d👨",
+    "kiss_woman_woman": "👩\u200d❤️\u200d💋\u200d👩",
+    "kissing_cat_face": "😽",
+    "kissing_face": "😗",
+    "kissing_face_with_closed_eyes": "😚",
+    "kissing_face_with_smiling_eyes": "😙",
+    "kitchen_knife": "🔪",
+    "kite": "🪁",
+    "kiwi_fruit": "🥝",
+    "koala": "🐨",
+    "lab_coat": "🥼",
+    "label": "🏷",
+    "lacrosse": "🥍",
+    "lady_beetle": "🐞",
+    "laptop_computer": "💻",
+    "large_blue_diamond": "🔷",
+    "large_orange_diamond": "🔶",
+    "last_quarter_moon": "🌗",
+    "last_quarter_moon_face": "🌜",
+    "last_track_button": "⏮",
+    "latin_cross": "✝",
+    "leaf_fluttering_in_wind": "🍃",
+    "leafy_green": "🥬",
+    "ledger": "📒",
+    "left-facing_fist": "🤛",
+    "left-facing_fist_dark_skin_tone": "🤛🏿",
+    "left-facing_fist_light_skin_tone": "🤛🏻",
+    "left-facing_fist_medium-dark_skin_tone": "🤛🏾",
+    "left-facing_fist_medium-light_skin_tone": "🤛🏼",
+    "left-facing_fist_medium_skin_tone": "🤛🏽",
+    "left-right_arrow": "↔",
+    "left_arrow": "⬅",
+    "left_arrow_curving_right": "↪",
+    "left_luggage": "🛅",
+    "left_speech_bubble": "🗨",
+    "leg": "🦵",
+    "lemon": "🍋",
+    "leopard": "🐆",
+    "level_slider": "🎚",
+    "light_bulb": "💡",
+    "light_rail": "🚈",
+    "link": "🔗",
+    "linked_paperclips": "🖇",
+    "lion_face": "🦁",
+    "lipstick": "💄",
+    "litter_in_bin_sign": "🚮",
+    "lizard": "🦎",
+    "llama": "🦙",
+    "lobster": "🦞",
+    "locked": "🔒",
+    "locked_with_key": "🔐",
+    "locked_with_pen": "🔏",
+    "locomotive": "🚂",
+    "lollipop": "🍭",
+    "lotion_bottle": "🧴",
+    "loudly_crying_face": "😭",
+    "loudspeaker": "📢",
+    "love-you_gesture": "🤟",
+    "love-you_gesture_dark_skin_tone": "🤟🏿",
+    "love-you_gesture_light_skin_tone": "🤟🏻",
+    "love-you_gesture_medium-dark_skin_tone": "🤟🏾",
+    "love-you_gesture_medium-light_skin_tone": "🤟🏼",
+    "love-you_gesture_medium_skin_tone": "🤟🏽",
+    "love_hotel": "🏩",
+    "love_letter": "💌",
+    "luggage": "🧳",
+    "lying_face": "🤥",
+    "mage": "🧙",
+    "mage_dark_skin_tone": "🧙🏿",
+    "mage_light_skin_tone": "🧙🏻",
+    "mage_medium-dark_skin_tone": "🧙🏾",
+    "mage_medium-light_skin_tone": "🧙🏼",
+    "mage_medium_skin_tone": "🧙🏽",
+    "magnet": "🧲",
+    "magnifying_glass_tilted_left": "🔍",
+    "magnifying_glass_tilted_right": "🔎",
+    "mahjong_red_dragon": "🀄",
+    "male_sign": "♂",
+    "man": "👨",
+    "man_and_woman_holding_hands": "👫",
+    "man_artist": "👨\u200d🎨",
+    "man_artist_dark_skin_tone": "👨🏿\u200d🎨",
+    "man_artist_light_skin_tone": "👨🏻\u200d🎨",
+    "man_artist_medium-dark_skin_tone": "👨🏾\u200d🎨",
+    "man_artist_medium-light_skin_tone": "👨🏼\u200d🎨",
+    "man_artist_medium_skin_tone": "👨🏽\u200d🎨",
+    "man_astronaut": "👨\u200d🚀",
+    "man_astronaut_dark_skin_tone": "👨🏿\u200d🚀",
+    "man_astronaut_light_skin_tone": "👨🏻\u200d🚀",
+    "man_astronaut_medium-dark_skin_tone": "👨🏾\u200d🚀",
+    "man_astronaut_medium-light_skin_tone": "👨🏼\u200d🚀",
+    "man_astronaut_medium_skin_tone": "👨🏽\u200d🚀",
+    "man_biking": "🚴\u200d♂️",
+    "man_biking_dark_skin_tone": "🚴🏿\u200d♂️",
+    "man_biking_light_skin_tone": "🚴🏻\u200d♂️",
+    "man_biking_medium-dark_skin_tone": "🚴🏾\u200d♂️",
+    "man_biking_medium-light_skin_tone": "🚴🏼\u200d♂️",
+    "man_biking_medium_skin_tone": "🚴🏽\u200d♂️",
+    "man_bouncing_ball": "⛹️\u200d♂️",
+    "man_bouncing_ball_dark_skin_tone": "⛹🏿\u200d♂️",
+    "man_bouncing_ball_light_skin_tone": "⛹🏻\u200d♂️",
+    "man_bouncing_ball_medium-dark_skin_tone": "⛹🏾\u200d♂️",
+    "man_bouncing_ball_medium-light_skin_tone": "⛹🏼\u200d♂️",
+    "man_bouncing_ball_medium_skin_tone": "⛹🏽\u200d♂️",
+    "man_bowing": "🙇\u200d♂️",
+    "man_bowing_dark_skin_tone": "🙇🏿\u200d♂️",
+    "man_bowing_light_skin_tone": "🙇🏻\u200d♂️",
+    "man_bowing_medium-dark_skin_tone": "🙇🏾\u200d♂️",
+    "man_bowing_medium-light_skin_tone": "🙇🏼\u200d♂️",
+    "man_bowing_medium_skin_tone": "🙇🏽\u200d♂️",
+    "man_cartwheeling": "🤸\u200d♂️",
+    "man_cartwheeling_dark_skin_tone": "🤸🏿\u200d♂️",
+    "man_cartwheeling_light_skin_tone": "🤸🏻\u200d♂️",
+    "man_cartwheeling_medium-dark_skin_tone": "🤸🏾\u200d♂️",
+    "man_cartwheeling_medium-light_skin_tone": "🤸🏼\u200d♂️",
+    "man_cartwheeling_medium_skin_tone": "🤸🏽\u200d♂️",
+    "man_climbing": "🧗\u200d♂️",
+    "man_climbing_dark_skin_tone": "🧗🏿\u200d♂️",
+    "man_climbing_light_skin_tone": "🧗🏻\u200d♂️",
+    "man_climbing_medium-dark_skin_tone": "🧗🏾\u200d♂️",
+    "man_climbing_medium-light_skin_tone": "🧗🏼\u200d♂️",
+    "man_climbing_medium_skin_tone": "🧗🏽\u200d♂️",
+    "man_construction_worker": "👷\u200d♂️",
+    "man_construction_worker_dark_skin_tone": "👷🏿\u200d♂️",
+    "man_construction_worker_light_skin_tone": "👷🏻\u200d♂️",
+    "man_construction_worker_medium-dark_skin_tone": "👷🏾\u200d♂️",
+    "man_construction_worker_medium-light_skin_tone": "👷🏼\u200d♂️",
+    "man_construction_worker_medium_skin_tone": "👷🏽\u200d♂️",
+    "man_cook": "👨\u200d🍳",
+    "man_cook_dark_skin_tone": "👨🏿\u200d🍳",
+    "man_cook_light_skin_tone": "👨🏻\u200d🍳",
+    "man_cook_medium-dark_skin_tone": "👨🏾\u200d🍳",
+    "man_cook_medium-light_skin_tone": "👨🏼\u200d🍳",
+    "man_cook_medium_skin_tone": "👨🏽\u200d🍳",
+    "man_dancing": "🕺",
+    "man_dancing_dark_skin_tone": "🕺🏿",
+    "man_dancing_light_skin_tone": "🕺🏻",
+    "man_dancing_medium-dark_skin_tone": "🕺🏾",
+    "man_dancing_medium-light_skin_tone": "🕺🏼",
+    "man_dancing_medium_skin_tone": "🕺🏽",
+    "man_dark_skin_tone": "👨🏿",
+    "man_detective": "🕵️\u200d♂️",
+    "man_detective_dark_skin_tone": "🕵🏿\u200d♂️",
+    "man_detective_light_skin_tone": "🕵🏻\u200d♂️",
+    "man_detective_medium-dark_skin_tone": "🕵🏾\u200d♂️",
+    "man_detective_medium-light_skin_tone": "🕵🏼\u200d♂️",
+    "man_detective_medium_skin_tone": "🕵🏽\u200d♂️",
+    "man_elf": "🧝\u200d♂️",
+    "man_elf_dark_skin_tone": "🧝🏿\u200d♂️",
+    "man_elf_light_skin_tone": "🧝🏻\u200d♂️",
+    "man_elf_medium-dark_skin_tone": "🧝🏾\u200d♂️",
+    "man_elf_medium-light_skin_tone": "🧝🏼\u200d♂️",
+    "man_elf_medium_skin_tone": "🧝🏽\u200d♂️",
+    "man_facepalming": "🤦\u200d♂️",
+    "man_facepalming_dark_skin_tone": "🤦🏿\u200d♂️",
+    "man_facepalming_light_skin_tone": "🤦🏻\u200d♂️",
+    "man_facepalming_medium-dark_skin_tone": "🤦🏾\u200d♂️",
+    "man_facepalming_medium-light_skin_tone": "🤦🏼\u200d♂️",
+    "man_facepalming_medium_skin_tone": "🤦🏽\u200d♂️",
+    "man_factory_worker": "👨\u200d🏭",
+    "man_factory_worker_dark_skin_tone": "👨🏿\u200d🏭",
+    "man_factory_worker_light_skin_tone": "👨🏻\u200d🏭",
+    "man_factory_worker_medium-dark_skin_tone": "👨🏾\u200d🏭",
+    "man_factory_worker_medium-light_skin_tone": "👨🏼\u200d🏭",
+    "man_factory_worker_medium_skin_tone": "👨🏽\u200d🏭",
+    "man_fairy": "🧚\u200d♂️",
+    "man_fairy_dark_skin_tone": "🧚🏿\u200d♂️",
+    "man_fairy_light_skin_tone": "🧚🏻\u200d♂️",
+    "man_fairy_medium-dark_skin_tone": "🧚🏾\u200d♂️",
+    "man_fairy_medium-light_skin_tone": "🧚🏼\u200d♂️",
+    "man_fairy_medium_skin_tone": "🧚🏽\u200d♂️",
+    "man_farmer": "👨\u200d🌾",
+    "man_farmer_dark_skin_tone": "👨🏿\u200d🌾",
+    "man_farmer_light_skin_tone": "👨🏻\u200d🌾",
+    "man_farmer_medium-dark_skin_tone": "👨🏾\u200d🌾",
+    "man_farmer_medium-light_skin_tone": "👨🏼\u200d🌾",
+    "man_farmer_medium_skin_tone": "👨🏽\u200d🌾",
+    "man_firefighter": "👨\u200d🚒",
+    "man_firefighter_dark_skin_tone": "👨🏿\u200d🚒",
+    "man_firefighter_light_skin_tone": "👨🏻\u200d🚒",
+    "man_firefighter_medium-dark_skin_tone": "👨🏾\u200d🚒",
+    "man_firefighter_medium-light_skin_tone": "👨🏼\u200d🚒",
+    "man_firefighter_medium_skin_tone": "👨🏽\u200d🚒",
+    "man_frowning": "🙍\u200d♂️",
+    "man_frowning_dark_skin_tone": "🙍🏿\u200d♂️",
+    "man_frowning_light_skin_tone": "🙍🏻\u200d♂️",
+    "man_frowning_medium-dark_skin_tone": "🙍🏾\u200d♂️",
+    "man_frowning_medium-light_skin_tone": "🙍🏼\u200d♂️",
+    "man_frowning_medium_skin_tone": "🙍🏽\u200d♂️",
+    "man_genie": "🧞\u200d♂️",
+    "man_gesturing_no": "🙅\u200d♂️",
+    "man_gesturing_no_dark_skin_tone": "🙅🏿\u200d♂️",
+    "man_gesturing_no_light_skin_tone": "🙅🏻\u200d♂️",
+    "man_gesturing_no_medium-dark_skin_tone": "🙅🏾\u200d♂️",
+    "man_gesturing_no_medium-light_skin_tone": "🙅🏼\u200d♂️",
+    "man_gesturing_no_medium_skin_tone": "🙅🏽\u200d♂️",
+    "man_gesturing_ok": "🙆\u200d♂️",
+    "man_gesturing_ok_dark_skin_tone": "🙆🏿\u200d♂️",
+    "man_gesturing_ok_light_skin_tone": "🙆🏻\u200d♂️",
+    "man_gesturing_ok_medium-dark_skin_tone": "🙆🏾\u200d♂️",
+    "man_gesturing_ok_medium-light_skin_tone": "🙆🏼\u200d♂️",
+    "man_gesturing_ok_medium_skin_tone": "🙆🏽\u200d♂️",
+    "man_getting_haircut": "💇\u200d♂️",
+    "man_getting_haircut_dark_skin_tone": "💇🏿\u200d♂️",
+    "man_getting_haircut_light_skin_tone": "💇🏻\u200d♂️",
+    "man_getting_haircut_medium-dark_skin_tone": "💇🏾\u200d♂️",
+    "man_getting_haircut_medium-light_skin_tone": "💇🏼\u200d♂️",
+    "man_getting_haircut_medium_skin_tone": "💇🏽\u200d♂️",
+    "man_getting_massage": "💆\u200d♂️",
+    "man_getting_massage_dark_skin_tone": "💆🏿\u200d♂️",
+    "man_getting_massage_light_skin_tone": "💆🏻\u200d♂️",
+    "man_getting_massage_medium-dark_skin_tone": "💆🏾\u200d♂️",
+    "man_getting_massage_medium-light_skin_tone": "💆🏼\u200d♂️",
+    "man_getting_massage_medium_skin_tone": "💆🏽\u200d♂️",
+    "man_golfing": "🏌️\u200d♂️",
+    "man_golfing_dark_skin_tone": "🏌🏿\u200d♂️",
+    "man_golfing_light_skin_tone": "🏌🏻\u200d♂️",
+    "man_golfing_medium-dark_skin_tone": "🏌🏾\u200d♂️",
+    "man_golfing_medium-light_skin_tone": "🏌🏼\u200d♂️",
+    "man_golfing_medium_skin_tone": "🏌🏽\u200d♂️",
+    "man_guard": "💂\u200d♂️",
+    "man_guard_dark_skin_tone": "💂🏿\u200d♂️",
+    "man_guard_light_skin_tone": "💂🏻\u200d♂️",
+    "man_guard_medium-dark_skin_tone": "💂🏾\u200d♂️",
+    "man_guard_medium-light_skin_tone": "💂🏼\u200d♂️",
+    "man_guard_medium_skin_tone": "💂🏽\u200d♂️",
+    "man_health_worker": "👨\u200d⚕️",
+    "man_health_worker_dark_skin_tone": "👨🏿\u200d⚕️",
+    "man_health_worker_light_skin_tone": "👨🏻\u200d⚕️",
+    "man_health_worker_medium-dark_skin_tone": "👨🏾\u200d⚕️",
+    "man_health_worker_medium-light_skin_tone": "👨🏼\u200d⚕️",
+    "man_health_worker_medium_skin_tone": "👨🏽\u200d⚕️",
+    "man_in_lotus_position": "🧘\u200d♂️",
+    "man_in_lotus_position_dark_skin_tone": "🧘🏿\u200d♂️",
+    "man_in_lotus_position_light_skin_tone": "🧘🏻\u200d♂️",
+    "man_in_lotus_position_medium-dark_skin_tone": "🧘🏾\u200d♂️",
+    "man_in_lotus_position_medium-light_skin_tone": "🧘🏼\u200d♂️",
+    "man_in_lotus_position_medium_skin_tone": "🧘🏽\u200d♂️",
+    "man_in_manual_wheelchair": "👨\u200d🦽",
+    "man_in_motorized_wheelchair": "👨\u200d🦼",
+    "man_in_steamy_room": "🧖\u200d♂️",
+    "man_in_steamy_room_dark_skin_tone": "🧖🏿\u200d♂️",
+    "man_in_steamy_room_light_skin_tone": "🧖🏻\u200d♂️",
+    "man_in_steamy_room_medium-dark_skin_tone": "🧖🏾\u200d♂️",
+    "man_in_steamy_room_medium-light_skin_tone": "🧖🏼\u200d♂️",
+    "man_in_steamy_room_medium_skin_tone": "🧖🏽\u200d♂️",
+    "man_in_suit_levitating": "🕴",
+    "man_in_suit_levitating_dark_skin_tone": "🕴🏿",
+    "man_in_suit_levitating_light_skin_tone": "🕴🏻",
+    "man_in_suit_levitating_medium-dark_skin_tone": "🕴🏾",
+    "man_in_suit_levitating_medium-light_skin_tone": "🕴🏼",
+    "man_in_suit_levitating_medium_skin_tone": "🕴🏽",
+    "man_in_tuxedo": "🤵",
+    "man_in_tuxedo_dark_skin_tone": "🤵🏿",
+    "man_in_tuxedo_light_skin_tone": "🤵🏻",
+    "man_in_tuxedo_medium-dark_skin_tone": "🤵🏾",
+    "man_in_tuxedo_medium-light_skin_tone": "🤵🏼",
+    "man_in_tuxedo_medium_skin_tone": "🤵🏽",
+    "man_judge": "👨\u200d⚖️",
+    "man_judge_dark_skin_tone": "👨🏿\u200d⚖️",
+    "man_judge_light_skin_tone": "👨🏻\u200d⚖️",
+    "man_judge_medium-dark_skin_tone": "👨🏾\u200d⚖️",
+    "man_judge_medium-light_skin_tone": "👨🏼\u200d⚖️",
+    "man_judge_medium_skin_tone": "👨🏽\u200d⚖️",
+    "man_juggling": "🤹\u200d♂️",
+    "man_juggling_dark_skin_tone": "🤹🏿\u200d♂️",
+    "man_juggling_light_skin_tone": "🤹🏻\u200d♂️",
+    "man_juggling_medium-dark_skin_tone": "🤹🏾\u200d♂️",
+    "man_juggling_medium-light_skin_tone": "🤹🏼\u200d♂️",
+    "man_juggling_medium_skin_tone": "🤹🏽\u200d♂️",
+    "man_lifting_weights": "🏋️\u200d♂️",
+    "man_lifting_weights_dark_skin_tone": "🏋🏿\u200d♂️",
+    "man_lifting_weights_light_skin_tone": "🏋🏻\u200d♂️",
+    "man_lifting_weights_medium-dark_skin_tone": "🏋🏾\u200d♂️",
+    "man_lifting_weights_medium-light_skin_tone": "🏋🏼\u200d♂️",
+    "man_lifting_weights_medium_skin_tone": "🏋🏽\u200d♂️",
+    "man_light_skin_tone": "👨🏻",
+    "man_mage": "🧙\u200d♂️",
+    "man_mage_dark_skin_tone": "🧙🏿\u200d♂️",
+    "man_mage_light_skin_tone": "🧙🏻\u200d♂️",
+    "man_mage_medium-dark_skin_tone": "🧙🏾\u200d♂️",
+    "man_mage_medium-light_skin_tone": "🧙🏼\u200d♂️",
+    "man_mage_medium_skin_tone": "🧙🏽\u200d♂️",
+    "man_mechanic": "👨\u200d🔧",
+    "man_mechanic_dark_skin_tone": "👨🏿\u200d🔧",
+    "man_mechanic_light_skin_tone": "👨🏻\u200d🔧",
+    "man_mechanic_medium-dark_skin_tone": "👨🏾\u200d🔧",
+    "man_mechanic_medium-light_skin_tone": "👨🏼\u200d🔧",
+    "man_mechanic_medium_skin_tone": "👨🏽\u200d🔧",
+    "man_medium-dark_skin_tone": "👨🏾",
+    "man_medium-light_skin_tone": "👨🏼",
+    "man_medium_skin_tone": "👨🏽",
+    "man_mountain_biking": "🚵\u200d♂️",
+    "man_mountain_biking_dark_skin_tone": "🚵🏿\u200d♂️",
+    "man_mountain_biking_light_skin_tone": "🚵🏻\u200d♂️",
+    "man_mountain_biking_medium-dark_skin_tone": "🚵🏾\u200d♂️",
+    "man_mountain_biking_medium-light_skin_tone": "🚵🏼\u200d♂️",
+    "man_mountain_biking_medium_skin_tone": "🚵🏽\u200d♂️",
+    "man_office_worker": "👨\u200d💼",
+    "man_office_worker_dark_skin_tone": "👨🏿\u200d💼",
+    "man_office_worker_light_skin_tone": "👨🏻\u200d💼",
+    "man_office_worker_medium-dark_skin_tone": "👨🏾\u200d💼",
+    "man_office_worker_medium-light_skin_tone": "👨🏼\u200d💼",
+    "man_office_worker_medium_skin_tone": "👨🏽\u200d💼",
+    "man_pilot": "👨\u200d✈️",
+    "man_pilot_dark_skin_tone": "👨🏿\u200d✈️",
+    "man_pilot_light_skin_tone": "👨🏻\u200d✈️",
+    "man_pilot_medium-dark_skin_tone": "👨🏾\u200d✈️",
+    "man_pilot_medium-light_skin_tone": "👨🏼\u200d✈️",
+    "man_pilot_medium_skin_tone": "👨🏽\u200d✈️",
+    "man_playing_handball": "🤾\u200d♂️",
+    "man_playing_handball_dark_skin_tone": "🤾🏿\u200d♂️",
+    "man_playing_handball_light_skin_tone": "🤾🏻\u200d♂️",
+    "man_playing_handball_medium-dark_skin_tone": "🤾🏾\u200d♂️",
+    "man_playing_handball_medium-light_skin_tone": "🤾🏼\u200d♂️",
+    "man_playing_handball_medium_skin_tone": "🤾🏽\u200d♂️",
+    "man_playing_water_polo": "🤽\u200d♂️",
+    "man_playing_water_polo_dark_skin_tone": "🤽🏿\u200d♂️",
+    "man_playing_water_polo_light_skin_tone": "🤽🏻\u200d♂️",
+    "man_playing_water_polo_medium-dark_skin_tone": "🤽🏾\u200d♂️",
+    "man_playing_water_polo_medium-light_skin_tone": "🤽🏼\u200d♂️",
+    "man_playing_water_polo_medium_skin_tone": "🤽🏽\u200d♂️",
+    "man_police_officer": "👮\u200d♂️",
+    "man_police_officer_dark_skin_tone": "👮🏿\u200d♂️",
+    "man_police_officer_light_skin_tone": "👮🏻\u200d♂️",
+    "man_police_officer_medium-dark_skin_tone": "👮🏾\u200d♂️",
+    "man_police_officer_medium-light_skin_tone": "👮🏼\u200d♂️",
+    "man_police_officer_medium_skin_tone": "👮🏽\u200d♂️",
+    "man_pouting": "🙎\u200d♂️",
+    "man_pouting_dark_skin_tone": "🙎🏿\u200d♂️",
+    "man_pouting_light_skin_tone": "🙎🏻\u200d♂️",
+    "man_pouting_medium-dark_skin_tone": "🙎🏾\u200d♂️",
+    "man_pouting_medium-light_skin_tone": "🙎🏼\u200d♂️",
+    "man_pouting_medium_skin_tone": "🙎🏽\u200d♂️",
+    "man_raising_hand": "🙋\u200d♂️",
+    "man_raising_hand_dark_skin_tone": "🙋🏿\u200d♂️",
+    "man_raising_hand_light_skin_tone": "🙋🏻\u200d♂️",
+    "man_raising_hand_medium-dark_skin_tone": "🙋🏾\u200d♂️",
+    "man_raising_hand_medium-light_skin_tone": "🙋🏼\u200d♂️",
+    "man_raising_hand_medium_skin_tone": "🙋🏽\u200d♂️",
+    "man_rowing_boat": "🚣\u200d♂️",
+    "man_rowing_boat_dark_skin_tone": "🚣🏿\u200d♂️",
+    "man_rowing_boat_light_skin_tone": "🚣🏻\u200d♂️",
+    "man_rowing_boat_medium-dark_skin_tone": "🚣🏾\u200d♂️",
+    "man_rowing_boat_medium-light_skin_tone": "🚣🏼\u200d♂️",
+    "man_rowing_boat_medium_skin_tone": "🚣🏽\u200d♂️",
+    "man_running": "🏃\u200d♂️",
+    "man_running_dark_skin_tone": "🏃🏿\u200d♂️",
+    "man_running_light_skin_tone": "🏃🏻\u200d♂️",
+    "man_running_medium-dark_skin_tone": "🏃🏾\u200d♂️",
+    "man_running_medium-light_skin_tone": "🏃🏼\u200d♂️",
+    "man_running_medium_skin_tone": "🏃🏽\u200d♂️",
+    "man_scientist": "👨\u200d🔬",
+    "man_scientist_dark_skin_tone": "👨🏿\u200d🔬",
+    "man_scientist_light_skin_tone": "👨🏻\u200d🔬",
+    "man_scientist_medium-dark_skin_tone": "👨🏾\u200d🔬",
+    "man_scientist_medium-light_skin_tone": "👨🏼\u200d🔬",
+    "man_scientist_medium_skin_tone": "👨🏽\u200d🔬",
+    "man_shrugging": "🤷\u200d♂️",
+    "man_shrugging_dark_skin_tone": "🤷🏿\u200d♂️",
+    "man_shrugging_light_skin_tone": "🤷🏻\u200d♂️",
+    "man_shrugging_medium-dark_skin_tone": "🤷🏾\u200d♂️",
+    "man_shrugging_medium-light_skin_tone": "🤷🏼\u200d♂️",
+    "man_shrugging_medium_skin_tone": "🤷🏽\u200d♂️",
+    "man_singer": "👨\u200d🎤",
+    "man_singer_dark_skin_tone": "👨🏿\u200d🎤",
+    "man_singer_light_skin_tone": "👨🏻\u200d🎤",
+    "man_singer_medium-dark_skin_tone": "👨🏾\u200d🎤",
+    "man_singer_medium-light_skin_tone": "👨🏼\u200d🎤",
+    "man_singer_medium_skin_tone": "👨🏽\u200d🎤",
+    "man_student": "👨\u200d🎓",
+    "man_student_dark_skin_tone": "👨🏿\u200d🎓",
+    "man_student_light_skin_tone": "👨🏻\u200d🎓",
+    "man_student_medium-dark_skin_tone": "👨🏾\u200d🎓",
+    "man_student_medium-light_skin_tone": "👨🏼\u200d🎓",
+    "man_student_medium_skin_tone": "👨🏽\u200d🎓",
+    "man_surfing": "🏄\u200d♂️",
+    "man_surfing_dark_skin_tone": "🏄🏿\u200d♂️",
+    "man_surfing_light_skin_tone": "🏄🏻\u200d♂️",
+    "man_surfing_medium-dark_skin_tone": "🏄🏾\u200d♂️",
+    "man_surfing_medium-light_skin_tone": "🏄🏼\u200d♂️",
+    "man_surfing_medium_skin_tone": "🏄🏽\u200d♂️",
+    "man_swimming": "🏊\u200d♂️",
+    "man_swimming_dark_skin_tone": "🏊🏿\u200d♂️",
+    "man_swimming_light_skin_tone": "🏊🏻\u200d♂️",
+    "man_swimming_medium-dark_skin_tone": "🏊🏾\u200d♂️",
+    "man_swimming_medium-light_skin_tone": "🏊🏼\u200d♂️",
+    "man_swimming_medium_skin_tone": "🏊🏽\u200d♂️",
+    "man_teacher": "👨\u200d🏫",
+    "man_teacher_dark_skin_tone": "👨🏿\u200d🏫",
+    "man_teacher_light_skin_tone": "👨🏻\u200d🏫",
+    "man_teacher_medium-dark_skin_tone": "👨🏾\u200d🏫",
+    "man_teacher_medium-light_skin_tone": "👨🏼\u200d🏫",
+    "man_teacher_medium_skin_tone": "👨🏽\u200d🏫",
+    "man_technologist": "👨\u200d💻",
+    "man_technologist_dark_skin_tone": "👨🏿\u200d💻",
+    "man_technologist_light_skin_tone": "👨🏻\u200d💻",
+    "man_technologist_medium-dark_skin_tone": "👨🏾\u200d💻",
+    "man_technologist_medium-light_skin_tone": "👨🏼\u200d💻",
+    "man_technologist_medium_skin_tone": "👨🏽\u200d💻",
+    "man_tipping_hand": "💁\u200d♂️",
+    "man_tipping_hand_dark_skin_tone": "💁🏿\u200d♂️",
+    "man_tipping_hand_light_skin_tone": "💁🏻\u200d♂️",
+    "man_tipping_hand_medium-dark_skin_tone": "💁🏾\u200d♂️",
+    "man_tipping_hand_medium-light_skin_tone": "💁🏼\u200d♂️",
+    "man_tipping_hand_medium_skin_tone": "💁🏽\u200d♂️",
+    "man_vampire": "🧛\u200d♂️",
+    "man_vampire_dark_skin_tone": "🧛🏿\u200d♂️",
+    "man_vampire_light_skin_tone": "🧛🏻\u200d♂️",
+    "man_vampire_medium-dark_skin_tone": "🧛🏾\u200d♂️",
+    "man_vampire_medium-light_skin_tone": "🧛🏼\u200d♂️",
+    "man_vampire_medium_skin_tone": "🧛🏽\u200d♂️",
+    "man_walking": "🚶\u200d♂️",
+    "man_walking_dark_skin_tone": "🚶🏿\u200d♂️",
+    "man_walking_light_skin_tone": "🚶🏻\u200d♂️",
+    "man_walking_medium-dark_skin_tone": "🚶🏾\u200d♂️",
+    "man_walking_medium-light_skin_tone": "🚶🏼\u200d♂️",
+    "man_walking_medium_skin_tone": "🚶🏽\u200d♂️",
+    "man_wearing_turban": "👳\u200d♂️",
+    "man_wearing_turban_dark_skin_tone": "👳🏿\u200d♂️",
+    "man_wearing_turban_light_skin_tone": "👳🏻\u200d♂️",
+    "man_wearing_turban_medium-dark_skin_tone": "👳🏾\u200d♂️",
+    "man_wearing_turban_medium-light_skin_tone": "👳🏼\u200d♂️",
+    "man_wearing_turban_medium_skin_tone": "👳🏽\u200d♂️",
+    "man_with_probing_cane": "👨\u200d🦯",
+    "man_with_chinese_cap": "👲",
+    "man_with_chinese_cap_dark_skin_tone": "👲🏿",
+    "man_with_chinese_cap_light_skin_tone": "👲🏻",
+    "man_with_chinese_cap_medium-dark_skin_tone": "👲🏾",
+    "man_with_chinese_cap_medium-light_skin_tone": "👲🏼",
+    "man_with_chinese_cap_medium_skin_tone": "👲🏽",
+    "man_zombie": "🧟\u200d♂️",
+    "mango": "🥭",
+    "mantelpiece_clock": "🕰",
+    "manual_wheelchair": "🦽",
+    "man’s_shoe": "👞",
+    "map_of_japan": "🗾",
+    "maple_leaf": "🍁",
+    "martial_arts_uniform": "🥋",
+    "mate": "🧉",
+    "meat_on_bone": "🍖",
+    "mechanical_arm": "🦾",
+    "mechanical_leg": "🦿",
+    "medical_symbol": "⚕",
+    "megaphone": "📣",
+    "melon": "🍈",
+    "memo": "📝",
+    "men_with_bunny_ears": "👯\u200d♂️",
+    "men_wrestling": "🤼\u200d♂️",
+    "menorah": "🕎",
+    "men’s_room": "🚹",
+    "mermaid": "🧜\u200d♀️",
+    "mermaid_dark_skin_tone": "🧜🏿\u200d♀️",
+    "mermaid_light_skin_tone": "🧜🏻\u200d♀️",
+    "mermaid_medium-dark_skin_tone": "🧜🏾\u200d♀️",
+    "mermaid_medium-light_skin_tone": "🧜🏼\u200d♀️",
+    "mermaid_medium_skin_tone": "🧜🏽\u200d♀️",
+    "merman": "🧜\u200d♂️",
+    "merman_dark_skin_tone": "🧜🏿\u200d♂️",
+    "merman_light_skin_tone": "🧜🏻\u200d♂️",
+    "merman_medium-dark_skin_tone": "🧜🏾\u200d♂️",
+    "merman_medium-light_skin_tone": "🧜🏼\u200d♂️",
+    "merman_medium_skin_tone": "🧜🏽\u200d♂️",
+    "merperson": "🧜",
+    "merperson_dark_skin_tone": "🧜🏿",
+    "merperson_light_skin_tone": "🧜🏻",
+    "merperson_medium-dark_skin_tone": "🧜🏾",
+    "merperson_medium-light_skin_tone": "🧜🏼",
+    "merperson_medium_skin_tone": "🧜🏽",
+    "metro": "🚇",
+    "microbe": "🦠",
+    "microphone": "🎤",
+    "microscope": "🔬",
+    "middle_finger": "🖕",
+    "middle_finger_dark_skin_tone": "🖕🏿",
+    "middle_finger_light_skin_tone": "🖕🏻",
+    "middle_finger_medium-dark_skin_tone": "🖕🏾",
+    "middle_finger_medium-light_skin_tone": "🖕🏼",
+    "middle_finger_medium_skin_tone": "🖕🏽",
+    "military_medal": "🎖",
+    "milky_way": "🌌",
+    "minibus": "🚐",
+    "moai": "🗿",
+    "mobile_phone": "📱",
+    "mobile_phone_off": "📴",
+    "mobile_phone_with_arrow": "📲",
+    "money-mouth_face": "🤑",
+    "money_bag": "💰",
+    "money_with_wings": "💸",
+    "monkey": "🐒",
+    "monkey_face": "🐵",
+    "monorail": "🚝",
+    "moon_cake": "🥮",
+    "moon_viewing_ceremony": "🎑",
+    "mosque": "🕌",
+    "mosquito": "🦟",
+    "motor_boat": "🛥",
+    "motor_scooter": "🛵",
+    "motorcycle": "🏍",
+    "motorized_wheelchair": "🦼",
+    "motorway": "🛣",
+    "mount_fuji": "🗻",
+    "mountain": "⛰",
+    "mountain_cableway": "🚠",
+    "mountain_railway": "🚞",
+    "mouse": "🐭",
+    "mouse_face": "🐭",
+    "mouth": "👄",
+    "movie_camera": "🎥",
+    "mushroom": "🍄",
+    "musical_keyboard": "🎹",
+    "musical_note": "🎵",
+    "musical_notes": "🎶",
+    "musical_score": "🎼",
+    "muted_speaker": "🔇",
+    "nail_polish": "💅",
+    "nail_polish_dark_skin_tone": "💅🏿",
+    "nail_polish_light_skin_tone": "💅🏻",
+    "nail_polish_medium-dark_skin_tone": "💅🏾",
+    "nail_polish_medium-light_skin_tone": "💅🏼",
+    "nail_polish_medium_skin_tone": "💅🏽",
+    "name_badge": "📛",
+    "national_park": "🏞",
+    "nauseated_face": "🤢",
+    "nazar_amulet": "🧿",
+    "necktie": "👔",
+    "nerd_face": "🤓",
+    "neutral_face": "😐",
+    "new_moon": "🌑",
+    "new_moon_face": "🌚",
+    "newspaper": "📰",
+    "next_track_button": "⏭",
+    "night_with_stars": "🌃",
+    "nine-thirty": "🕤",
+    "nine_o’clock": "🕘",
+    "no_bicycles": "🚳",
+    "no_entry": "⛔",
+    "no_littering": "🚯",
+    "no_mobile_phones": "📵",
+    "no_one_under_eighteen": "🔞",
+    "no_pedestrians": "🚷",
+    "no_smoking": "🚭",
+    "non-potable_water": "🚱",
+    "nose": "👃",
+    "nose_dark_skin_tone": "👃🏿",
+    "nose_light_skin_tone": "👃🏻",
+    "nose_medium-dark_skin_tone": "👃🏾",
+    "nose_medium-light_skin_tone": "👃🏼",
+    "nose_medium_skin_tone": "👃🏽",
+    "notebook": "📓",
+    "notebook_with_decorative_cover": "📔",
+    "nut_and_bolt": "🔩",
+    "octopus": "🐙",
+    "oden": "🍢",
+    "office_building": "🏢",
+    "ogre": "👹",
+    "oil_drum": "🛢",
+    "old_key": "🗝",
+    "old_man": "👴",
+    "old_man_dark_skin_tone": "👴🏿",
+    "old_man_light_skin_tone": "👴🏻",
+    "old_man_medium-dark_skin_tone": "👴🏾",
+    "old_man_medium-light_skin_tone": "👴🏼",
+    "old_man_medium_skin_tone": "👴🏽",
+    "old_woman": "👵",
+    "old_woman_dark_skin_tone": "👵🏿",
+    "old_woman_light_skin_tone": "👵🏻",
+    "old_woman_medium-dark_skin_tone": "👵🏾",
+    "old_woman_medium-light_skin_tone": "👵🏼",
+    "old_woman_medium_skin_tone": "👵🏽",
+    "older_adult": "🧓",
+    "older_adult_dark_skin_tone": "🧓🏿",
+    "older_adult_light_skin_tone": "🧓🏻",
+    "older_adult_medium-dark_skin_tone": "🧓🏾",
+    "older_adult_medium-light_skin_tone": "🧓🏼",
+    "older_adult_medium_skin_tone": "🧓🏽",
+    "om": "🕉",
+    "oncoming_automobile": "🚘",
+    "oncoming_bus": "🚍",
+    "oncoming_fist": "👊",
+    "oncoming_fist_dark_skin_tone": "👊🏿",
+    "oncoming_fist_light_skin_tone": "👊🏻",
+    "oncoming_fist_medium-dark_skin_tone": "👊🏾",
+    "oncoming_fist_medium-light_skin_tone": "👊🏼",
+    "oncoming_fist_medium_skin_tone": "👊🏽",
+    "oncoming_police_car": "🚔",
+    "oncoming_taxi": "🚖",
+    "one-piece_swimsuit": "🩱",
+    "one-thirty": "🕜",
+    "one_o’clock": "🕐",
+    "onion": "🧅",
+    "open_book": "📖",
+    "open_file_folder": "📂",
+    "open_hands": "👐",
+    "open_hands_dark_skin_tone": "👐🏿",
+    "open_hands_light_skin_tone": "👐🏻",
+    "open_hands_medium-dark_skin_tone": "👐🏾",
+    "open_hands_medium-light_skin_tone": "👐🏼",
+    "open_hands_medium_skin_tone": "👐🏽",
+    "open_mailbox_with_lowered_flag": "📭",
+    "open_mailbox_with_raised_flag": "📬",
+    "optical_disk": "💿",
+    "orange_book": "📙",
+    "orange_circle": "🟠",
+    "orange_heart": "🧡",
+    "orange_square": "🟧",
+    "orangutan": "🦧",
+    "orthodox_cross": "☦",
+    "otter": "🦦",
+    "outbox_tray": "📤",
+    "owl": "🦉",
+    "ox": "🐂",
+    "oyster": "🦪",
+    "package": "📦",
+    "page_facing_up": "📄",
+    "page_with_curl": "📃",
+    "pager": "📟",
+    "paintbrush": "🖌",
+    "palm_tree": "🌴",
+    "palms_up_together": "🤲",
+    "palms_up_together_dark_skin_tone": "🤲🏿",
+    "palms_up_together_light_skin_tone": "🤲🏻",
+    "palms_up_together_medium-dark_skin_tone": "🤲🏾",
+    "palms_up_together_medium-light_skin_tone": "🤲🏼",
+    "palms_up_together_medium_skin_tone": "🤲🏽",
+    "pancakes": "🥞",
+    "panda_face": "🐼",
+    "paperclip": "📎",
+    "parrot": "🦜",
+    "part_alternation_mark": "〽",
+    "party_popper": "🎉",
+    "partying_face": "🥳",
+    "passenger_ship": "🛳",
+    "passport_control": "🛂",
+    "pause_button": "⏸",
+    "paw_prints": "🐾",
+    "peace_symbol": "☮",
+    "peach": "🍑",
+    "peacock": "🦚",
+    "peanuts": "🥜",
+    "pear": "🍐",
+    "pen": "🖊",
+    "pencil": "📝",
+    "penguin": "🐧",
+    "pensive_face": "😔",
+    "people_holding_hands": "🧑\u200d🤝\u200d🧑",
+    "people_with_bunny_ears": "👯",
+    "people_wrestling": "🤼",
+    "performing_arts": "🎭",
+    "persevering_face": "😣",
+    "person_biking": "🚴",
+    "person_biking_dark_skin_tone": "🚴🏿",
+    "person_biking_light_skin_tone": "🚴🏻",
+    "person_biking_medium-dark_skin_tone": "🚴🏾",
+    "person_biking_medium-light_skin_tone": "🚴🏼",
+    "person_biking_medium_skin_tone": "🚴🏽",
+    "person_bouncing_ball": "⛹",
+    "person_bouncing_ball_dark_skin_tone": "⛹🏿",
+    "person_bouncing_ball_light_skin_tone": "⛹🏻",
+    "person_bouncing_ball_medium-dark_skin_tone": "⛹🏾",
+    "person_bouncing_ball_medium-light_skin_tone": "⛹🏼",
+    "person_bouncing_ball_medium_skin_tone": "⛹🏽",
+    "person_bowing": "🙇",
+    "person_bowing_dark_skin_tone": "🙇🏿",
+    "person_bowing_light_skin_tone": "🙇🏻",
+    "person_bowing_medium-dark_skin_tone": "🙇🏾",
+    "person_bowing_medium-light_skin_tone": "🙇🏼",
+    "person_bowing_medium_skin_tone": "🙇🏽",
+    "person_cartwheeling": "🤸",
+    "person_cartwheeling_dark_skin_tone": "🤸🏿",
+    "person_cartwheeling_light_skin_tone": "🤸🏻",
+    "person_cartwheeling_medium-dark_skin_tone": "🤸🏾",
+    "person_cartwheeling_medium-light_skin_tone": "🤸🏼",
+    "person_cartwheeling_medium_skin_tone": "🤸🏽",
+    "person_climbing": "🧗",
+    "person_climbing_dark_skin_tone": "🧗🏿",
+    "person_climbing_light_skin_tone": "🧗🏻",
+    "person_climbing_medium-dark_skin_tone": "🧗🏾",
+    "person_climbing_medium-light_skin_tone": "🧗🏼",
+    "person_climbing_medium_skin_tone": "🧗🏽",
+    "person_facepalming": "🤦",
+    "person_facepalming_dark_skin_tone": "🤦🏿",
+    "person_facepalming_light_skin_tone": "🤦🏻",
+    "person_facepalming_medium-dark_skin_tone": "🤦🏾",
+    "person_facepalming_medium-light_skin_tone": "🤦🏼",
+    "person_facepalming_medium_skin_tone": "🤦🏽",
+    "person_fencing": "🤺",
+    "person_frowning": "🙍",
+    "person_frowning_dark_skin_tone": "🙍🏿",
+    "person_frowning_light_skin_tone": "🙍🏻",
+    "person_frowning_medium-dark_skin_tone": "🙍🏾",
+    "person_frowning_medium-light_skin_tone": "🙍🏼",
+    "person_frowning_medium_skin_tone": "🙍🏽",
+    "person_gesturing_no": "🙅",
+    "person_gesturing_no_dark_skin_tone": "🙅🏿",
+    "person_gesturing_no_light_skin_tone": "🙅🏻",
+    "person_gesturing_no_medium-dark_skin_tone": "🙅🏾",
+    "person_gesturing_no_medium-light_skin_tone": "🙅🏼",
+    "person_gesturing_no_medium_skin_tone": "🙅🏽",
+    "person_gesturing_ok": "🙆",
+    "person_gesturing_ok_dark_skin_tone": "🙆🏿",
+    "person_gesturing_ok_light_skin_tone": "🙆🏻",
+    "person_gesturing_ok_medium-dark_skin_tone": "🙆🏾",
+    "person_gesturing_ok_medium-light_skin_tone": "🙆🏼",
+    "person_gesturing_ok_medium_skin_tone": "🙆🏽",
+    "person_getting_haircut": "💇",
+    "person_getting_haircut_dark_skin_tone": "💇🏿",
+    "person_getting_haircut_light_skin_tone": "💇🏻",
+    "person_getting_haircut_medium-dark_skin_tone": "💇🏾",
+    "person_getting_haircut_medium-light_skin_tone": "💇🏼",
+    "person_getting_haircut_medium_skin_tone": "💇🏽",
+    "person_getting_massage": "💆",
+    "person_getting_massage_dark_skin_tone": "💆🏿",
+    "person_getting_massage_light_skin_tone": "💆🏻",
+    "person_getting_massage_medium-dark_skin_tone": "💆🏾",
+    "person_getting_massage_medium-light_skin_tone": "💆🏼",
+    "person_getting_massage_medium_skin_tone": "💆🏽",
+    "person_golfing": "🏌",
+    "person_golfing_dark_skin_tone": "🏌🏿",
+    "person_golfing_light_skin_tone": "🏌🏻",
+    "person_golfing_medium-dark_skin_tone": "🏌🏾",
+    "person_golfing_medium-light_skin_tone": "🏌🏼",
+    "person_golfing_medium_skin_tone": "🏌🏽",
+    "person_in_bed": "🛌",
+    "person_in_bed_dark_skin_tone": "🛌🏿",
+    "person_in_bed_light_skin_tone": "🛌🏻",
+    "person_in_bed_medium-dark_skin_tone": "🛌🏾",
+    "person_in_bed_medium-light_skin_tone": "🛌🏼",
+    "person_in_bed_medium_skin_tone": "🛌🏽",
+    "person_in_lotus_position": "🧘",
+    "person_in_lotus_position_dark_skin_tone": "🧘🏿",
+    "person_in_lotus_position_light_skin_tone": "🧘🏻",
+    "person_in_lotus_position_medium-dark_skin_tone": "🧘🏾",
+    "person_in_lotus_position_medium-light_skin_tone": "🧘🏼",
+    "person_in_lotus_position_medium_skin_tone": "🧘🏽",
+    "person_in_steamy_room": "🧖",
+    "person_in_steamy_room_dark_skin_tone": "🧖🏿",
+    "person_in_steamy_room_light_skin_tone": "🧖🏻",
+    "person_in_steamy_room_medium-dark_skin_tone": "🧖🏾",
+    "person_in_steamy_room_medium-light_skin_tone": "🧖🏼",
+    "person_in_steamy_room_medium_skin_tone": "🧖🏽",
+    "person_juggling": "🤹",
+    "person_juggling_dark_skin_tone": "🤹🏿",
+    "person_juggling_light_skin_tone": "🤹🏻",
+    "person_juggling_medium-dark_skin_tone": "🤹🏾",
+    "person_juggling_medium-light_skin_tone": "🤹🏼",
+    "person_juggling_medium_skin_tone": "🤹🏽",
+    "person_kneeling": "🧎",
+    "person_lifting_weights": "🏋",
+    "person_lifting_weights_dark_skin_tone": "🏋🏿",
+    "person_lifting_weights_light_skin_tone": "🏋🏻",
+    "person_lifting_weights_medium-dark_skin_tone": "🏋🏾",
+    "person_lifting_weights_medium-light_skin_tone": "🏋🏼",
+    "person_lifting_weights_medium_skin_tone": "🏋🏽",
+    "person_mountain_biking": "🚵",
+    "person_mountain_biking_dark_skin_tone": "🚵🏿",
+    "person_mountain_biking_light_skin_tone": "🚵🏻",
+    "person_mountain_biking_medium-dark_skin_tone": "🚵🏾",
+    "person_mountain_biking_medium-light_skin_tone": "🚵🏼",
+    "person_mountain_biking_medium_skin_tone": "🚵🏽",
+    "person_playing_handball": "🤾",
+    "person_playing_handball_dark_skin_tone": "🤾🏿",
+    "person_playing_handball_light_skin_tone": "🤾🏻",
+    "person_playing_handball_medium-dark_skin_tone": "🤾🏾",
+    "person_playing_handball_medium-light_skin_tone": "🤾🏼",
+    "person_playing_handball_medium_skin_tone": "🤾🏽",
+    "person_playing_water_polo": "🤽",
+    "person_playing_water_polo_dark_skin_tone": "🤽🏿",
+    "person_playing_water_polo_light_skin_tone": "🤽🏻",
+    "person_playing_water_polo_medium-dark_skin_tone": "🤽🏾",
+    "person_playing_water_polo_medium-light_skin_tone": "🤽🏼",
+    "person_playing_water_polo_medium_skin_tone": "🤽🏽",
+    "person_pouting": "🙎",
+    "person_pouting_dark_skin_tone": "🙎🏿",
+    "person_pouting_light_skin_tone": "🙎🏻",
+    "person_pouting_medium-dark_skin_tone": "🙎🏾",
+    "person_pouting_medium-light_skin_tone": "🙎🏼",
+    "person_pouting_medium_skin_tone": "🙎🏽",
+    "person_raising_hand": "🙋",
+    "person_raising_hand_dark_skin_tone": "🙋🏿",
+    "person_raising_hand_light_skin_tone": "🙋🏻",
+    "person_raising_hand_medium-dark_skin_tone": "🙋🏾",
+    "person_raising_hand_medium-light_skin_tone": "🙋🏼",
+    "person_raising_hand_medium_skin_tone": "🙋🏽",
+    "person_rowing_boat": "🚣",
+    "person_rowing_boat_dark_skin_tone": "🚣🏿",
+    "person_rowing_boat_light_skin_tone": "🚣🏻",
+    "person_rowing_boat_medium-dark_skin_tone": "🚣🏾",
+    "person_rowing_boat_medium-light_skin_tone": "🚣🏼",
+    "person_rowing_boat_medium_skin_tone": "🚣🏽",
+    "person_running": "🏃",
+    "person_running_dark_skin_tone": "🏃🏿",
+    "person_running_light_skin_tone": "🏃🏻",
+    "person_running_medium-dark_skin_tone": "🏃🏾",
+    "person_running_medium-light_skin_tone": "🏃🏼",
+    "person_running_medium_skin_tone": "🏃🏽",
+    "person_shrugging": "🤷",
+    "person_shrugging_dark_skin_tone": "🤷🏿",
+    "person_shrugging_light_skin_tone": "🤷🏻",
+    "person_shrugging_medium-dark_skin_tone": "🤷🏾",
+    "person_shrugging_medium-light_skin_tone": "🤷🏼",
+    "person_shrugging_medium_skin_tone": "🤷🏽",
+    "person_standing": "🧍",
+    "person_surfing": "🏄",
+    "person_surfing_dark_skin_tone": "🏄🏿",
+    "person_surfing_light_skin_tone": "🏄🏻",
+    "person_surfing_medium-dark_skin_tone": "🏄🏾",
+    "person_surfing_medium-light_skin_tone": "🏄🏼",
+    "person_surfing_medium_skin_tone": "🏄🏽",
+    "person_swimming": "🏊",
+    "person_swimming_dark_skin_tone": "🏊🏿",
+    "person_swimming_light_skin_tone": "🏊🏻",
+    "person_swimming_medium-dark_skin_tone": "🏊🏾",
+    "person_swimming_medium-light_skin_tone": "🏊🏼",
+    "person_swimming_medium_skin_tone": "🏊🏽",
+    "person_taking_bath": "🛀",
+    "person_taking_bath_dark_skin_tone": "🛀🏿",
+    "person_taking_bath_light_skin_tone": "🛀🏻",
+    "person_taking_bath_medium-dark_skin_tone": "🛀🏾",
+    "person_taking_bath_medium-light_skin_tone": "🛀🏼",
+    "person_taking_bath_medium_skin_tone": "🛀🏽",
+    "person_tipping_hand": "💁",
+    "person_tipping_hand_dark_skin_tone": "💁🏿",
+    "person_tipping_hand_light_skin_tone": "💁🏻",
+    "person_tipping_hand_medium-dark_skin_tone": "💁🏾",
+    "person_tipping_hand_medium-light_skin_tone": "💁🏼",
+    "person_tipping_hand_medium_skin_tone": "💁🏽",
+    "person_walking": "🚶",
+    "person_walking_dark_skin_tone": "🚶🏿",
+    "person_walking_light_skin_tone": "🚶🏻",
+    "person_walking_medium-dark_skin_tone": "🚶🏾",
+    "person_walking_medium-light_skin_tone": "🚶🏼",
+    "person_walking_medium_skin_tone": "🚶🏽",
+    "person_wearing_turban": "👳",
+    "person_wearing_turban_dark_skin_tone": "👳🏿",
+    "person_wearing_turban_light_skin_tone": "👳🏻",
+    "person_wearing_turban_medium-dark_skin_tone": "👳🏾",
+    "person_wearing_turban_medium-light_skin_tone": "👳🏼",
+    "person_wearing_turban_medium_skin_tone": "👳🏽",
+    "petri_dish": "🧫",
+    "pick": "⛏",
+    "pie": "🥧",
+    "pig": "🐷",
+    "pig_face": "🐷",
+    "pig_nose": "🐽",
+    "pile_of_poo": "💩",
+    "pill": "💊",
+    "pinching_hand": "🤏",
+    "pine_decoration": "🎍",
+    "pineapple": "🍍",
+    "ping_pong": "🏓",
+    "pirate_flag": "🏴\u200d☠️",
+    "pistol": "🔫",
+    "pizza": "🍕",
+    "place_of_worship": "🛐",
+    "play_button": "▶",
+    "play_or_pause_button": "⏯",
+    "pleading_face": "🥺",
+    "police_car": "🚓",
+    "police_car_light": "🚨",
+    "police_officer": "👮",
+    "police_officer_dark_skin_tone": "👮🏿",
+    "police_officer_light_skin_tone": "👮🏻",
+    "police_officer_medium-dark_skin_tone": "👮🏾",
+    "police_officer_medium-light_skin_tone": "👮🏼",
+    "police_officer_medium_skin_tone": "👮🏽",
+    "poodle": "🐩",
+    "pool_8_ball": "🎱",
+    "popcorn": "🍿",
+    "post_office": "🏣",
+    "postal_horn": "📯",
+    "postbox": "📮",
+    "pot_of_food": "🍲",
+    "potable_water": "🚰",
+    "potato": "🥔",
+    "poultry_leg": "🍗",
+    "pound_banknote": "💷",
+    "pouting_cat_face": "😾",
+    "pouting_face": "😡",
+    "prayer_beads": "📿",
+    "pregnant_woman": "🤰",
+    "pregnant_woman_dark_skin_tone": "🤰🏿",
+    "pregnant_woman_light_skin_tone": "🤰🏻",
+    "pregnant_woman_medium-dark_skin_tone": "🤰🏾",
+    "pregnant_woman_medium-light_skin_tone": "🤰🏼",
+    "pregnant_woman_medium_skin_tone": "🤰🏽",
+    "pretzel": "🥨",
+    "probing_cane": "🦯",
+    "prince": "🤴",
+    "prince_dark_skin_tone": "🤴🏿",
+    "prince_light_skin_tone": "🤴🏻",
+    "prince_medium-dark_skin_tone": "🤴🏾",
+    "prince_medium-light_skin_tone": "🤴🏼",
+    "prince_medium_skin_tone": "🤴🏽",
+    "princess": "👸",
+    "princess_dark_skin_tone": "👸🏿",
+    "princess_light_skin_tone": "👸🏻",
+    "princess_medium-dark_skin_tone": "👸🏾",
+    "princess_medium-light_skin_tone": "👸🏼",
+    "princess_medium_skin_tone": "👸🏽",
+    "printer": "🖨",
+    "prohibited": "🚫",
+    "purple_circle": "🟣",
+    "purple_heart": "💜",
+    "purple_square": "🟪",
+    "purse": "👛",
+    "pushpin": "📌",
+    "question_mark": "❓",
+    "rabbit": "🐰",
+    "rabbit_face": "🐰",
+    "raccoon": "🦝",
+    "racing_car": "🏎",
+    "radio": "📻",
+    "radio_button": "🔘",
+    "radioactive": "☢",
+    "railway_car": "🚃",
+    "railway_track": "🛤",
+    "rainbow": "🌈",
+    "rainbow_flag": "🏳️\u200d🌈",
+    "raised_back_of_hand": "🤚",
+    "raised_back_of_hand_dark_skin_tone": "🤚🏿",
+    "raised_back_of_hand_light_skin_tone": "🤚🏻",
+    "raised_back_of_hand_medium-dark_skin_tone": "🤚🏾",
+    "raised_back_of_hand_medium-light_skin_tone": "🤚🏼",
+    "raised_back_of_hand_medium_skin_tone": "🤚🏽",
+    "raised_fist": "✊",
+    "raised_fist_dark_skin_tone": "✊🏿",
+    "raised_fist_light_skin_tone": "✊🏻",
+    "raised_fist_medium-dark_skin_tone": "✊🏾",
+    "raised_fist_medium-light_skin_tone": "✊🏼",
+    "raised_fist_medium_skin_tone": "✊🏽",
+    "raised_hand": "✋",
+    "raised_hand_dark_skin_tone": "✋🏿",
+    "raised_hand_light_skin_tone": "✋🏻",
+    "raised_hand_medium-dark_skin_tone": "✋🏾",
+    "raised_hand_medium-light_skin_tone": "✋🏼",
+    "raised_hand_medium_skin_tone": "✋🏽",
+    "raising_hands": "🙌",
+    "raising_hands_dark_skin_tone": "🙌🏿",
+    "raising_hands_light_skin_tone": "🙌🏻",
+    "raising_hands_medium-dark_skin_tone": "🙌🏾",
+    "raising_hands_medium-light_skin_tone": "🙌🏼",
+    "raising_hands_medium_skin_tone": "🙌🏽",
+    "ram": "🐏",
+    "rat": "🐀",
+    "razor": "🪒",
+    "ringed_planet": "🪐",
+    "receipt": "🧾",
+    "record_button": "⏺",
+    "recycling_symbol": "♻",
+    "red_apple": "🍎",
+    "red_circle": "🔴",
+    "red_envelope": "🧧",
+    "red_hair": "🦰",
+    "red-haired_man": "👨\u200d🦰",
+    "red-haired_woman": "👩\u200d🦰",
+    "red_heart": "❤",
+    "red_paper_lantern": "🏮",
+    "red_square": "🟥",
+    "red_triangle_pointed_down": "🔻",
+    "red_triangle_pointed_up": "🔺",
+    "registered": "®",
+    "relieved_face": "😌",
+    "reminder_ribbon": "🎗",
+    "repeat_button": "🔁",
+    "repeat_single_button": "🔂",
+    "rescue_worker’s_helmet": "⛑",
+    "restroom": "🚻",
+    "reverse_button": "◀",
+    "revolving_hearts": "💞",
+    "rhinoceros": "🦏",
+    "ribbon": "🎀",
+    "rice_ball": "🍙",
+    "rice_cracker": "🍘",
+    "right-facing_fist": "🤜",
+    "right-facing_fist_dark_skin_tone": "🤜🏿",
+    "right-facing_fist_light_skin_tone": "🤜🏻",
+    "right-facing_fist_medium-dark_skin_tone": "🤜🏾",
+    "right-facing_fist_medium-light_skin_tone": "🤜🏼",
+    "right-facing_fist_medium_skin_tone": "🤜🏽",
+    "right_anger_bubble": "🗯",
+    "right_arrow": "➡",
+    "right_arrow_curving_down": "⤵",
+    "right_arrow_curving_left": "↩",
+    "right_arrow_curving_up": "⤴",
+    "ring": "💍",
+    "roasted_sweet_potato": "🍠",
+    "robot_face": "🤖",
+    "rocket": "🚀",
+    "roll_of_paper": "🧻",
+    "rolled-up_newspaper": "🗞",
+    "roller_coaster": "🎢",
+    "rolling_on_the_floor_laughing": "🤣",
+    "rooster": "🐓",
+    "rose": "🌹",
+    "rosette": "🏵",
+    "round_pushpin": "📍",
+    "rugby_football": "🏉",
+    "running_shirt": "🎽",
+    "running_shoe": "👟",
+    "sad_but_relieved_face": "😥",
+    "safety_pin": "🧷",
+    "safety_vest": "🦺",
+    "salt": "🧂",
+    "sailboat": "⛵",
+    "sake": "🍶",
+    "sandwich": "🥪",
+    "sari": "🥻",
+    "satellite": "📡",
+    "satellite_antenna": "📡",
+    "sauropod": "🦕",
+    "saxophone": "🎷",
+    "scarf": "🧣",
+    "school": "🏫",
+    "school_backpack": "🎒",
+    "scissors": "✂",
+    "scorpion": "🦂",
+    "scroll": "📜",
+    "seat": "💺",
+    "see-no-evil_monkey": "🙈",
+    "seedling": "🌱",
+    "selfie": "🤳",
+    "selfie_dark_skin_tone": "🤳🏿",
+    "selfie_light_skin_tone": "🤳🏻",
+    "selfie_medium-dark_skin_tone": "🤳🏾",
+    "selfie_medium-light_skin_tone": "🤳🏼",
+    "selfie_medium_skin_tone": "🤳🏽",
+    "service_dog": "🐕\u200d🦺",
+    "seven-thirty": "🕢",
+    "seven_o’clock": "🕖",
+    "shallow_pan_of_food": "🥘",
+    "shamrock": "☘",
+    "shark": "🦈",
+    "shaved_ice": "🍧",
+    "sheaf_of_rice": "🌾",
+    "shield": "🛡",
+    "shinto_shrine": "⛩",
+    "ship": "🚢",
+    "shooting_star": "🌠",
+    "shopping_bags": "🛍",
+    "shopping_cart": "🛒",
+    "shortcake": "🍰",
+    "shorts": "🩳",
+    "shower": "🚿",
+    "shrimp": "🦐",
+    "shuffle_tracks_button": "🔀",
+    "shushing_face": "🤫",
+    "sign_of_the_horns": "🤘",
+    "sign_of_the_horns_dark_skin_tone": "🤘🏿",
+    "sign_of_the_horns_light_skin_tone": "🤘🏻",
+    "sign_of_the_horns_medium-dark_skin_tone": "🤘🏾",
+    "sign_of_the_horns_medium-light_skin_tone": "🤘🏼",
+    "sign_of_the_horns_medium_skin_tone": "🤘🏽",
+    "six-thirty": "🕡",
+    "six_o’clock": "🕕",
+    "skateboard": "🛹",
+    "skier": "⛷",
+    "skis": "🎿",
+    "skull": "💀",
+    "skull_and_crossbones": "☠",
+    "skunk": "🦨",
+    "sled": "🛷",
+    "sleeping_face": "😴",
+    "sleepy_face": "😪",
+    "slightly_frowning_face": "🙁",
+    "slightly_smiling_face": "🙂",
+    "slot_machine": "🎰",
+    "sloth": "🦥",
+    "small_airplane": "🛩",
+    "small_blue_diamond": "🔹",
+    "small_orange_diamond": "🔸",
+    "smiling_cat_face_with_heart-eyes": "😻",
+    "smiling_face": "☺",
+    "smiling_face_with_halo": "😇",
+    "smiling_face_with_3_hearts": "🥰",
+    "smiling_face_with_heart-eyes": "😍",
+    "smiling_face_with_horns": "😈",
+    "smiling_face_with_smiling_eyes": "😊",
+    "smiling_face_with_sunglasses": "😎",
+    "smirking_face": "😏",
+    "snail": "🐌",
+    "snake": "🐍",
+    "sneezing_face": "🤧",
+    "snow-capped_mountain": "🏔",
+    "snowboarder": "🏂",
+    "snowboarder_dark_skin_tone": "🏂🏿",
+    "snowboarder_light_skin_tone": "🏂🏻",
+    "snowboarder_medium-dark_skin_tone": "🏂🏾",
+    "snowboarder_medium-light_skin_tone": "🏂🏼",
+    "snowboarder_medium_skin_tone": "🏂🏽",
+    "snowflake": "❄",
+    "snowman": "☃",
+    "snowman_without_snow": "⛄",
+    "soap": "🧼",
+    "soccer_ball": "⚽",
+    "socks": "🧦",
+    "softball": "🥎",
+    "soft_ice_cream": "🍦",
+    "spade_suit": "♠",
+    "spaghetti": "🍝",
+    "sparkle": "❇",
+    "sparkler": "🎇",
+    "sparkles": "✨",
+    "sparkling_heart": "💖",
+    "speak-no-evil_monkey": "🙊",
+    "speaker_high_volume": "🔊",
+    "speaker_low_volume": "🔈",
+    "speaker_medium_volume": "🔉",
+    "speaking_head": "🗣",
+    "speech_balloon": "💬",
+    "speedboat": "🚤",
+    "spider": "🕷",
+    "spider_web": "🕸",
+    "spiral_calendar": "🗓",
+    "spiral_notepad": "🗒",
+    "spiral_shell": "🐚",
+    "spoon": "🥄",
+    "sponge": "🧽",
+    "sport_utility_vehicle": "🚙",
+    "sports_medal": "🏅",
+    "spouting_whale": "🐳",
+    "squid": "🦑",
+    "squinting_face_with_tongue": "😝",
+    "stadium": "🏟",
+    "star-struck": "🤩",
+    "star_and_crescent": "☪",
+    "star_of_david": "✡",
+    "station": "🚉",
+    "steaming_bowl": "🍜",
+    "stethoscope": "🩺",
+    "stop_button": "⏹",
+    "stop_sign": "🛑",
+    "stopwatch": "⏱",
+    "straight_ruler": "📏",
+    "strawberry": "🍓",
+    "studio_microphone": "🎙",
+    "stuffed_flatbread": "🥙",
+    "sun": "☀",
+    "sun_behind_cloud": "⛅",
+    "sun_behind_large_cloud": "🌥",
+    "sun_behind_rain_cloud": "🌦",
+    "sun_behind_small_cloud": "🌤",
+    "sun_with_face": "🌞",
+    "sunflower": "🌻",
+    "sunglasses": "😎",
+    "sunrise": "🌅",
+    "sunrise_over_mountains": "🌄",
+    "sunset": "🌇",
+    "superhero": "🦸",
+    "supervillain": "🦹",
+    "sushi": "🍣",
+    "suspension_railway": "🚟",
+    "swan": "🦢",
+    "sweat_droplets": "💦",
+    "synagogue": "🕍",
+    "syringe": "💉",
+    "t-shirt": "👕",
+    "taco": "🌮",
+    "takeout_box": "🥡",
+    "tanabata_tree": "🎋",
+    "tangerine": "🍊",
+    "taxi": "🚕",
+    "teacup_without_handle": "🍵",
+    "tear-off_calendar": "📆",
+    "teddy_bear": "🧸",
+    "telephone": "☎",
+    "telephone_receiver": "📞",
+    "telescope": "🔭",
+    "television": "📺",
+    "ten-thirty": "🕥",
+    "ten_o’clock": "🕙",
+    "tennis": "🎾",
+    "tent": "⛺",
+    "test_tube": "🧪",
+    "thermometer": "🌡",
+    "thinking_face": "🤔",
+    "thought_balloon": "💭",
+    "thread": "🧵",
+    "three-thirty": "🕞",
+    "three_o’clock": "🕒",
+    "thumbs_down": "👎",
+    "thumbs_down_dark_skin_tone": "👎🏿",
+    "thumbs_down_light_skin_tone": "👎🏻",
+    "thumbs_down_medium-dark_skin_tone": "👎🏾",
+    "thumbs_down_medium-light_skin_tone": "👎🏼",
+    "thumbs_down_medium_skin_tone": "👎🏽",
+    "thumbs_up": "👍",
+    "thumbs_up_dark_skin_tone": "👍🏿",
+    "thumbs_up_light_skin_tone": "👍🏻",
+    "thumbs_up_medium-dark_skin_tone": "👍🏾",
+    "thumbs_up_medium-light_skin_tone": "👍🏼",
+    "thumbs_up_medium_skin_tone": "👍🏽",
+    "ticket": "🎫",
+    "tiger": "🐯",
+    "tiger_face": "🐯",
+    "timer_clock": "⏲",
+    "tired_face": "😫",
+    "toolbox": "🧰",
+    "toilet": "🚽",
+    "tomato": "🍅",
+    "tongue": "👅",
+    "tooth": "🦷",
+    "top_hat": "🎩",
+    "tornado": "🌪",
+    "trackball": "🖲",
+    "tractor": "🚜",
+    "trade_mark": "™",
+    "train": "🚋",
+    "tram": "🚊",
+    "tram_car": "🚋",
+    "triangular_flag": "🚩",
+    "triangular_ruler": "📐",
+    "trident_emblem": "🔱",
+    "trolleybus": "🚎",
+    "trophy": "🏆",
+    "tropical_drink": "🍹",
+    "tropical_fish": "🐠",
+    "trumpet": "🎺",
+    "tulip": "🌷",
+    "tumbler_glass": "🥃",
+    "turtle": "🐢",
+    "twelve-thirty": "🕧",
+    "twelve_o’clock": "🕛",
+    "two-hump_camel": "🐫",
+    "two-thirty": "🕝",
+    "two_hearts": "💕",
+    "two_men_holding_hands": "👬",
+    "two_o’clock": "🕑",
+    "two_women_holding_hands": "👭",
+    "umbrella": "☂",
+    "umbrella_on_ground": "⛱",
+    "umbrella_with_rain_drops": "☔",
+    "unamused_face": "😒",
+    "unicorn_face": "🦄",
+    "unlocked": "🔓",
+    "up-down_arrow": "↕",
+    "up-left_arrow": "↖",
+    "up-right_arrow": "↗",
+    "up_arrow": "⬆",
+    "upside-down_face": "🙃",
+    "upwards_button": "🔼",
+    "vampire": "🧛",
+    "vampire_dark_skin_tone": "🧛🏿",
+    "vampire_light_skin_tone": "🧛🏻",
+    "vampire_medium-dark_skin_tone": "🧛🏾",
+    "vampire_medium-light_skin_tone": "🧛🏼",
+    "vampire_medium_skin_tone": "🧛🏽",
+    "vertical_traffic_light": "🚦",
+    "vibration_mode": "📳",
+    "victory_hand": "✌",
+    "victory_hand_dark_skin_tone": "✌🏿",
+    "victory_hand_light_skin_tone": "✌🏻",
+    "victory_hand_medium-dark_skin_tone": "✌🏾",
+    "victory_hand_medium-light_skin_tone": "✌🏼",
+    "victory_hand_medium_skin_tone": "✌🏽",
+    "video_camera": "📹",
+    "video_game": "🎮",
+    "videocassette": "📼",
+    "violin": "🎻",
+    "volcano": "🌋",
+    "volleyball": "🏐",
+    "vulcan_salute": "🖖",
+    "vulcan_salute_dark_skin_tone": "🖖🏿",
+    "vulcan_salute_light_skin_tone": "🖖🏻",
+    "vulcan_salute_medium-dark_skin_tone": "🖖🏾",
+    "vulcan_salute_medium-light_skin_tone": "🖖🏼",
+    "vulcan_salute_medium_skin_tone": "🖖🏽",
+    "waffle": "🧇",
+    "waning_crescent_moon": "🌘",
+    "waning_gibbous_moon": "🌖",
+    "warning": "⚠",
+    "wastebasket": "🗑",
+    "watch": "⌚",
+    "water_buffalo": "🐃",
+    "water_closet": "🚾",
+    "water_wave": "🌊",
+    "watermelon": "🍉",
+    "waving_hand": "👋",
+    "waving_hand_dark_skin_tone": "👋🏿",
+    "waving_hand_light_skin_tone": "👋🏻",
+    "waving_hand_medium-dark_skin_tone": "👋🏾",
+    "waving_hand_medium-light_skin_tone": "👋🏼",
+    "waving_hand_medium_skin_tone": "👋🏽",
+    "wavy_dash": "〰",
+    "waxing_crescent_moon": "🌒",
+    "waxing_gibbous_moon": "🌔",
+    "weary_cat_face": "🙀",
+    "weary_face": "😩",
+    "wedding": "💒",
+    "whale": "🐳",
+    "wheel_of_dharma": "☸",
+    "wheelchair_symbol": "♿",
+    "white_circle": "⚪",
+    "white_exclamation_mark": "❕",
+    "white_flag": "🏳",
+    "white_flower": "💮",
+    "white_hair": "🦳",
+    "white-haired_man": "👨\u200d🦳",
+    "white-haired_woman": "👩\u200d🦳",
+    "white_heart": "🤍",
+    "white_heavy_check_mark": "✅",
+    "white_large_square": "⬜",
+    "white_medium-small_square": "◽",
+    "white_medium_square": "◻",
+    "white_medium_star": "⭐",
+    "white_question_mark": "❔",
+    "white_small_square": "▫",
+    "white_square_button": "🔳",
+    "wilted_flower": "🥀",
+    "wind_chime": "🎐",
+    "wind_face": "🌬",
+    "wine_glass": "🍷",
+    "winking_face": "😉",
+    "winking_face_with_tongue": "😜",
+    "wolf_face": "🐺",
+    "woman": "👩",
+    "woman_artist": "👩\u200d🎨",
+    "woman_artist_dark_skin_tone": "👩🏿\u200d🎨",
+    "woman_artist_light_skin_tone": "👩🏻\u200d🎨",
+    "woman_artist_medium-dark_skin_tone": "👩🏾\u200d🎨",
+    "woman_artist_medium-light_skin_tone": "👩🏼\u200d🎨",
+    "woman_artist_medium_skin_tone": "👩🏽\u200d🎨",
+    "woman_astronaut": "👩\u200d🚀",
+    "woman_astronaut_dark_skin_tone": "👩🏿\u200d🚀",
+    "woman_astronaut_light_skin_tone": "👩🏻\u200d🚀",
+    "woman_astronaut_medium-dark_skin_tone": "👩🏾\u200d🚀",
+    "woman_astronaut_medium-light_skin_tone": "👩🏼\u200d🚀",
+    "woman_astronaut_medium_skin_tone": "👩🏽\u200d🚀",
+    "woman_biking": "🚴\u200d♀️",
+    "woman_biking_dark_skin_tone": "🚴🏿\u200d♀️",
+    "woman_biking_light_skin_tone": "🚴🏻\u200d♀️",
+    "woman_biking_medium-dark_skin_tone": "🚴🏾\u200d♀️",
+    "woman_biking_medium-light_skin_tone": "🚴🏼\u200d♀️",
+    "woman_biking_medium_skin_tone": "🚴🏽\u200d♀️",
+    "woman_bouncing_ball": "⛹️\u200d♀️",
+    "woman_bouncing_ball_dark_skin_tone": "⛹🏿\u200d♀️",
+    "woman_bouncing_ball_light_skin_tone": "⛹🏻\u200d♀️",
+    "woman_bouncing_ball_medium-dark_skin_tone": "⛹🏾\u200d♀️",
+    "woman_bouncing_ball_medium-light_skin_tone": "⛹🏼\u200d♀️",
+    "woman_bouncing_ball_medium_skin_tone": "⛹🏽\u200d♀️",
+    "woman_bowing": "🙇\u200d♀️",
+    "woman_bowing_dark_skin_tone": "🙇🏿\u200d♀️",
+    "woman_bowing_light_skin_tone": "🙇🏻\u200d♀️",
+    "woman_bowing_medium-dark_skin_tone": "🙇🏾\u200d♀️",
+    "woman_bowing_medium-light_skin_tone": "🙇🏼\u200d♀️",
+    "woman_bowing_medium_skin_tone": "🙇🏽\u200d♀️",
+    "woman_cartwheeling": "🤸\u200d♀️",
+    "woman_cartwheeling_dark_skin_tone": "🤸🏿\u200d♀️",
+    "woman_cartwheeling_light_skin_tone": "🤸🏻\u200d♀️",
+    "woman_cartwheeling_medium-dark_skin_tone": "🤸🏾\u200d♀️",
+    "woman_cartwheeling_medium-light_skin_tone": "🤸🏼\u200d♀️",
+    "woman_cartwheeling_medium_skin_tone": "🤸🏽\u200d♀️",
+    "woman_climbing": "🧗\u200d♀️",
+    "woman_climbing_dark_skin_tone": "🧗🏿\u200d♀️",
+    "woman_climbing_light_skin_tone": "🧗🏻\u200d♀️",
+    "woman_climbing_medium-dark_skin_tone": "🧗🏾\u200d♀️",
+    "woman_climbing_medium-light_skin_tone": "🧗🏼\u200d♀️",
+    "woman_climbing_medium_skin_tone": "🧗🏽\u200d♀️",
+    "woman_construction_worker": "👷\u200d♀️",
+    "woman_construction_worker_dark_skin_tone": "👷🏿\u200d♀️",
+    "woman_construction_worker_light_skin_tone": "👷🏻\u200d♀️",
+    "woman_construction_worker_medium-dark_skin_tone": "👷🏾\u200d♀️",
+    "woman_construction_worker_medium-light_skin_tone": "👷🏼\u200d♀️",
+    "woman_construction_worker_medium_skin_tone": "👷🏽\u200d♀️",
+    "woman_cook": "👩\u200d🍳",
+    "woman_cook_dark_skin_tone": "👩🏿\u200d🍳",
+    "woman_cook_light_skin_tone": "👩🏻\u200d🍳",
+    "woman_cook_medium-dark_skin_tone": "👩🏾\u200d🍳",
+    "woman_cook_medium-light_skin_tone": "👩🏼\u200d🍳",
+    "woman_cook_medium_skin_tone": "👩🏽\u200d🍳",
+    "woman_dancing": "💃",
+    "woman_dancing_dark_skin_tone": "💃🏿",
+    "woman_dancing_light_skin_tone": "💃🏻",
+    "woman_dancing_medium-dark_skin_tone": "💃🏾",
+    "woman_dancing_medium-light_skin_tone": "💃🏼",
+    "woman_dancing_medium_skin_tone": "💃🏽",
+    "woman_dark_skin_tone": "👩🏿",
+    "woman_detective": "🕵️\u200d♀️",
+    "woman_detective_dark_skin_tone": "🕵🏿\u200d♀️",
+    "woman_detective_light_skin_tone": "🕵🏻\u200d♀️",
+    "woman_detective_medium-dark_skin_tone": "🕵🏾\u200d♀️",
+    "woman_detective_medium-light_skin_tone": "🕵🏼\u200d♀️",
+    "woman_detective_medium_skin_tone": "🕵🏽\u200d♀️",
+    "woman_elf": "🧝\u200d♀️",
+    "woman_elf_dark_skin_tone": "🧝🏿\u200d♀️",
+    "woman_elf_light_skin_tone": "🧝🏻\u200d♀️",
+    "woman_elf_medium-dark_skin_tone": "🧝🏾\u200d♀️",
+    "woman_elf_medium-light_skin_tone": "🧝🏼\u200d♀️",
+    "woman_elf_medium_skin_tone": "🧝🏽\u200d♀️",
+    "woman_facepalming": "🤦\u200d♀️",
+    "woman_facepalming_dark_skin_tone": "🤦🏿\u200d♀️",
+    "woman_facepalming_light_skin_tone": "🤦🏻\u200d♀️",
+    "woman_facepalming_medium-dark_skin_tone": "🤦🏾\u200d♀️",
+    "woman_facepalming_medium-light_skin_tone": "🤦🏼\u200d♀️",
+    "woman_facepalming_medium_skin_tone": "🤦🏽\u200d♀️",
+    "woman_factory_worker": "👩\u200d🏭",
+    "woman_factory_worker_dark_skin_tone": "👩🏿\u200d🏭",
+    "woman_factory_worker_light_skin_tone": "👩🏻\u200d🏭",
+    "woman_factory_worker_medium-dark_skin_tone": "👩🏾\u200d🏭",
+    "woman_factory_worker_medium-light_skin_tone": "👩🏼\u200d🏭",
+    "woman_factory_worker_medium_skin_tone": "👩🏽\u200d🏭",
+    "woman_fairy": "🧚\u200d♀️",
+    "woman_fairy_dark_skin_tone": "🧚🏿\u200d♀️",
+    "woman_fairy_light_skin_tone": "🧚🏻\u200d♀️",
+    "woman_fairy_medium-dark_skin_tone": "🧚🏾\u200d♀️",
+    "woman_fairy_medium-light_skin_tone": "🧚🏼\u200d♀️",
+    "woman_fairy_medium_skin_tone": "🧚🏽\u200d♀️",
+    "woman_farmer": "👩\u200d🌾",
+    "woman_farmer_dark_skin_tone": "👩🏿\u200d🌾",
+    "woman_farmer_light_skin_tone": "👩🏻\u200d🌾",
+    "woman_farmer_medium-dark_skin_tone": "👩🏾\u200d🌾",
+    "woman_farmer_medium-light_skin_tone": "👩🏼\u200d🌾",
+    "woman_farmer_medium_skin_tone": "👩🏽\u200d🌾",
+    "woman_firefighter": "👩\u200d🚒",
+    "woman_firefighter_dark_skin_tone": "👩🏿\u200d🚒",
+    "woman_firefighter_light_skin_tone": "👩🏻\u200d🚒",
+    "woman_firefighter_medium-dark_skin_tone": "👩🏾\u200d🚒",
+    "woman_firefighter_medium-light_skin_tone": "👩🏼\u200d🚒",
+    "woman_firefighter_medium_skin_tone": "👩🏽\u200d🚒",
+    "woman_frowning": "🙍\u200d♀️",
+    "woman_frowning_dark_skin_tone": "🙍🏿\u200d♀️",
+    "woman_frowning_light_skin_tone": "🙍🏻\u200d♀️",
+    "woman_frowning_medium-dark_skin_tone": "🙍🏾\u200d♀️",
+    "woman_frowning_medium-light_skin_tone": "🙍🏼\u200d♀️",
+    "woman_frowning_medium_skin_tone": "🙍🏽\u200d♀️",
+    "woman_genie": "🧞\u200d♀️",
+    "woman_gesturing_no": "🙅\u200d♀️",
+    "woman_gesturing_no_dark_skin_tone": "🙅🏿\u200d♀️",
+    "woman_gesturing_no_light_skin_tone": "🙅🏻\u200d♀️",
+    "woman_gesturing_no_medium-dark_skin_tone": "🙅🏾\u200d♀️",
+    "woman_gesturing_no_medium-light_skin_tone": "🙅🏼\u200d♀️",
+    "woman_gesturing_no_medium_skin_tone": "🙅🏽\u200d♀️",
+    "woman_gesturing_ok": "🙆\u200d♀️",
+    "woman_gesturing_ok_dark_skin_tone": "🙆🏿\u200d♀️",
+    "woman_gesturing_ok_light_skin_tone": "🙆🏻\u200d♀️",
+    "woman_gesturing_ok_medium-dark_skin_tone": "🙆🏾\u200d♀️",
+    "woman_gesturing_ok_medium-light_skin_tone": "🙆🏼\u200d♀️",
+    "woman_gesturing_ok_medium_skin_tone": "🙆🏽\u200d♀️",
+    "woman_getting_haircut": "💇\u200d♀️",
+    "woman_getting_haircut_dark_skin_tone": "💇🏿\u200d♀️",
+    "woman_getting_haircut_light_skin_tone": "💇🏻\u200d♀️",
+    "woman_getting_haircut_medium-dark_skin_tone": "💇🏾\u200d♀️",
+    "woman_getting_haircut_medium-light_skin_tone": "💇🏼\u200d♀️",
+    "woman_getting_haircut_medium_skin_tone": "💇🏽\u200d♀️",
+    "woman_getting_massage": "💆\u200d♀️",
+    "woman_getting_massage_dark_skin_tone": "💆🏿\u200d♀️",
+    "woman_getting_massage_light_skin_tone": "💆🏻\u200d♀️",
+    "woman_getting_massage_medium-dark_skin_tone": "💆🏾\u200d♀️",
+    "woman_getting_massage_medium-light_skin_tone": "💆🏼\u200d♀️",
+    "woman_getting_massage_medium_skin_tone": "💆🏽\u200d♀️",
+    "woman_golfing": "🏌️\u200d♀️",
+    "woman_golfing_dark_skin_tone": "🏌🏿\u200d♀️",
+    "woman_golfing_light_skin_tone": "🏌🏻\u200d♀️",
+    "woman_golfing_medium-dark_skin_tone": "🏌🏾\u200d♀️",
+    "woman_golfing_medium-light_skin_tone": "🏌🏼\u200d♀️",
+    "woman_golfing_medium_skin_tone": "🏌🏽\u200d♀️",
+    "woman_guard": "💂\u200d♀️",
+    "woman_guard_dark_skin_tone": "💂🏿\u200d♀️",
+    "woman_guard_light_skin_tone": "💂🏻\u200d♀️",
+    "woman_guard_medium-dark_skin_tone": "💂🏾\u200d♀️",
+    "woman_guard_medium-light_skin_tone": "💂🏼\u200d♀️",
+    "woman_guard_medium_skin_tone": "💂🏽\u200d♀️",
+    "woman_health_worker": "👩\u200d⚕️",
+    "woman_health_worker_dark_skin_tone": "👩🏿\u200d⚕️",
+    "woman_health_worker_light_skin_tone": "👩🏻\u200d⚕️",
+    "woman_health_worker_medium-dark_skin_tone": "👩🏾\u200d⚕️",
+    "woman_health_worker_medium-light_skin_tone": "👩🏼\u200d⚕️",
+    "woman_health_worker_medium_skin_tone": "👩🏽\u200d⚕️",
+    "woman_in_lotus_position": "🧘\u200d♀️",
+    "woman_in_lotus_position_dark_skin_tone": "🧘🏿\u200d♀️",
+    "woman_in_lotus_position_light_skin_tone": "🧘🏻\u200d♀️",
+    "woman_in_lotus_position_medium-dark_skin_tone": "🧘🏾\u200d♀️",
+    "woman_in_lotus_position_medium-light_skin_tone": "🧘🏼\u200d♀️",
+    "woman_in_lotus_position_medium_skin_tone": "🧘🏽\u200d♀️",
+    "woman_in_manual_wheelchair": "👩\u200d🦽",
+    "woman_in_motorized_wheelchair": "👩\u200d🦼",
+    "woman_in_steamy_room": "🧖\u200d♀️",
+    "woman_in_steamy_room_dark_skin_tone": "🧖🏿\u200d♀️",
+    "woman_in_steamy_room_light_skin_tone": "🧖🏻\u200d♀️",
+    "woman_in_steamy_room_medium-dark_skin_tone": "🧖🏾\u200d♀️",
+    "woman_in_steamy_room_medium-light_skin_tone": "🧖🏼\u200d♀️",
+    "woman_in_steamy_room_medium_skin_tone": "🧖🏽\u200d♀️",
+    "woman_judge": "👩\u200d⚖️",
+    "woman_judge_dark_skin_tone": "👩🏿\u200d⚖️",
+    "woman_judge_light_skin_tone": "👩🏻\u200d⚖️",
+    "woman_judge_medium-dark_skin_tone": "👩🏾\u200d⚖️",
+    "woman_judge_medium-light_skin_tone": "👩🏼\u200d⚖️",
+    "woman_judge_medium_skin_tone": "👩🏽\u200d⚖️",
+    "woman_juggling": "🤹\u200d♀️",
+    "woman_juggling_dark_skin_tone": "🤹🏿\u200d♀️",
+    "woman_juggling_light_skin_tone": "🤹🏻\u200d♀️",
+    "woman_juggling_medium-dark_skin_tone": "🤹🏾\u200d♀️",
+    "woman_juggling_medium-light_skin_tone": "🤹🏼\u200d♀️",
+    "woman_juggling_medium_skin_tone": "🤹🏽\u200d♀️",
+    "woman_lifting_weights": "🏋️\u200d♀️",
+    "woman_lifting_weights_dark_skin_tone": "🏋🏿\u200d♀️",
+    "woman_lifting_weights_light_skin_tone": "🏋🏻\u200d♀️",
+    "woman_lifting_weights_medium-dark_skin_tone": "🏋🏾\u200d♀️",
+    "woman_lifting_weights_medium-light_skin_tone": "🏋🏼\u200d♀️",
+    "woman_lifting_weights_medium_skin_tone": "🏋🏽\u200d♀️",
+    "woman_light_skin_tone": "👩🏻",
+    "woman_mage": "🧙\u200d♀️",
+    "woman_mage_dark_skin_tone": "🧙🏿\u200d♀️",
+    "woman_mage_light_skin_tone": "🧙🏻\u200d♀️",
+    "woman_mage_medium-dark_skin_tone": "🧙🏾\u200d♀️",
+    "woman_mage_medium-light_skin_tone": "🧙🏼\u200d♀️",
+    "woman_mage_medium_skin_tone": "🧙🏽\u200d♀️",
+    "woman_mechanic": "👩\u200d🔧",
+    "woman_mechanic_dark_skin_tone": "👩🏿\u200d🔧",
+    "woman_mechanic_light_skin_tone": "👩🏻\u200d🔧",
+    "woman_mechanic_medium-dark_skin_tone": "👩🏾\u200d🔧",
+    "woman_mechanic_medium-light_skin_tone": "👩🏼\u200d🔧",
+    "woman_mechanic_medium_skin_tone": "👩🏽\u200d🔧",
+    "woman_medium-dark_skin_tone": "👩🏾",
+    "woman_medium-light_skin_tone": "👩🏼",
+    "woman_medium_skin_tone": "👩🏽",
+    "woman_mountain_biking": "🚵\u200d♀️",
+    "woman_mountain_biking_dark_skin_tone": "🚵🏿\u200d♀️",
+    "woman_mountain_biking_light_skin_tone": "🚵🏻\u200d♀️",
+    "woman_mountain_biking_medium-dark_skin_tone": "🚵🏾\u200d♀️",
+    "woman_mountain_biking_medium-light_skin_tone": "🚵🏼\u200d♀️",
+    "woman_mountain_biking_medium_skin_tone": "🚵🏽\u200d♀️",
+    "woman_office_worker": "👩\u200d💼",
+    "woman_office_worker_dark_skin_tone": "👩🏿\u200d💼",
+    "woman_office_worker_light_skin_tone": "👩🏻\u200d💼",
+    "woman_office_worker_medium-dark_skin_tone": "👩🏾\u200d💼",
+    "woman_office_worker_medium-light_skin_tone": "👩🏼\u200d💼",
+    "woman_office_worker_medium_skin_tone": "👩🏽\u200d💼",
+    "woman_pilot": "👩\u200d✈️",
+    "woman_pilot_dark_skin_tone": "👩🏿\u200d✈️",
+    "woman_pilot_light_skin_tone": "👩🏻\u200d✈️",
+    "woman_pilot_medium-dark_skin_tone": "👩🏾\u200d✈️",
+    "woman_pilot_medium-light_skin_tone": "👩🏼\u200d✈️",
+    "woman_pilot_medium_skin_tone": "👩🏽\u200d✈️",
+    "woman_playing_handball": "🤾\u200d♀️",
+    "woman_playing_handball_dark_skin_tone": "🤾🏿\u200d♀️",
+    "woman_playing_handball_light_skin_tone": "🤾🏻\u200d♀️",
+    "woman_playing_handball_medium-dark_skin_tone": "🤾🏾\u200d♀️",
+    "woman_playing_handball_medium-light_skin_tone": "🤾🏼\u200d♀️",
+    "woman_playing_handball_medium_skin_tone": "🤾🏽\u200d♀️",
+    "woman_playing_water_polo": "🤽\u200d♀️",
+    "woman_playing_water_polo_dark_skin_tone": "🤽🏿\u200d♀️",
+    "woman_playing_water_polo_light_skin_tone": "🤽🏻\u200d♀️",
+    "woman_playing_water_polo_medium-dark_skin_tone": "🤽🏾\u200d♀️",
+    "woman_playing_water_polo_medium-light_skin_tone": "🤽🏼\u200d♀️",
+    "woman_playing_water_polo_medium_skin_tone": "🤽🏽\u200d♀️",
+    "woman_police_officer": "👮\u200d♀️",
+    "woman_police_officer_dark_skin_tone": "👮🏿\u200d♀️",
+    "woman_police_officer_light_skin_tone": "👮🏻\u200d♀️",
+    "woman_police_officer_medium-dark_skin_tone": "👮🏾\u200d♀️",
+    "woman_police_officer_medium-light_skin_tone": "👮🏼\u200d♀️",
+    "woman_police_officer_medium_skin_tone": "👮🏽\u200d♀️",
+    "woman_pouting": "🙎\u200d♀️",
+    "woman_pouting_dark_skin_tone": "🙎🏿\u200d♀️",
+    "woman_pouting_light_skin_tone": "🙎🏻\u200d♀️",
+    "woman_pouting_medium-dark_skin_tone": "🙎🏾\u200d♀️",
+    "woman_pouting_medium-light_skin_tone": "🙎🏼\u200d♀️",
+    "woman_pouting_medium_skin_tone": "🙎🏽\u200d♀️",
+    "woman_raising_hand": "🙋\u200d♀️",
+    "woman_raising_hand_dark_skin_tone": "🙋🏿\u200d♀️",
+    "woman_raising_hand_light_skin_tone": "🙋🏻\u200d♀️",
+    "woman_raising_hand_medium-dark_skin_tone": "🙋🏾\u200d♀️",
+    "woman_raising_hand_medium-light_skin_tone": "🙋🏼\u200d♀️",
+    "woman_raising_hand_medium_skin_tone": "🙋🏽\u200d♀️",
+    "woman_rowing_boat": "🚣\u200d♀️",
+    "woman_rowing_boat_dark_skin_tone": "🚣🏿\u200d♀️",
+    "woman_rowing_boat_light_skin_tone": "🚣🏻\u200d♀️",
+    "woman_rowing_boat_medium-dark_skin_tone": "🚣🏾\u200d♀️",
+    "woman_rowing_boat_medium-light_skin_tone": "🚣🏼\u200d♀️",
+    "woman_rowing_boat_medium_skin_tone": "🚣🏽\u200d♀️",
+    "woman_running": "🏃\u200d♀️",
+    "woman_running_dark_skin_tone": "🏃🏿\u200d♀️",
+    "woman_running_light_skin_tone": "🏃🏻\u200d♀️",
+    "woman_running_medium-dark_skin_tone": "🏃🏾\u200d♀️",
+    "woman_running_medium-light_skin_tone": "🏃🏼\u200d♀️",
+    "woman_running_medium_skin_tone": "🏃🏽\u200d♀️",
+    "woman_scientist": "👩\u200d🔬",
+    "woman_scientist_dark_skin_tone": "👩🏿\u200d🔬",
+    "woman_scientist_light_skin_tone": "👩🏻\u200d🔬",
+    "woman_scientist_medium-dark_skin_tone": "👩🏾\u200d🔬",
+    "woman_scientist_medium-light_skin_tone": "👩🏼\u200d🔬",
+    "woman_scientist_medium_skin_tone": "👩🏽\u200d🔬",
+    "woman_shrugging": "🤷\u200d♀️",
+    "woman_shrugging_dark_skin_tone": "🤷🏿\u200d♀️",
+    "woman_shrugging_light_skin_tone": "🤷🏻\u200d♀️",
+    "woman_shrugging_medium-dark_skin_tone": "🤷🏾\u200d♀️",
+    "woman_shrugging_medium-light_skin_tone": "🤷🏼\u200d♀️",
+    "woman_shrugging_medium_skin_tone": "🤷🏽\u200d♀️",
+    "woman_singer": "👩\u200d🎤",
+    "woman_singer_dark_skin_tone": "👩🏿\u200d🎤",
+    "woman_singer_light_skin_tone": "👩🏻\u200d🎤",
+    "woman_singer_medium-dark_skin_tone": "👩🏾\u200d🎤",
+    "woman_singer_medium-light_skin_tone": "👩🏼\u200d🎤",
+    "woman_singer_medium_skin_tone": "👩🏽\u200d🎤",
+    "woman_student": "👩\u200d🎓",
+    "woman_student_dark_skin_tone": "👩🏿\u200d🎓",
+    "woman_student_light_skin_tone": "👩🏻\u200d🎓",
+    "woman_student_medium-dark_skin_tone": "👩🏾\u200d🎓",
+    "woman_student_medium-light_skin_tone": "👩🏼\u200d🎓",
+    "woman_student_medium_skin_tone": "👩🏽\u200d🎓",
+    "woman_surfing": "🏄\u200d♀️",
+    "woman_surfing_dark_skin_tone": "🏄🏿\u200d♀️",
+    "woman_surfing_light_skin_tone": "🏄🏻\u200d♀️",
+    "woman_surfing_medium-dark_skin_tone": "🏄🏾\u200d♀️",
+    "woman_surfing_medium-light_skin_tone": "🏄🏼\u200d♀️",
+    "woman_surfing_medium_skin_tone": "🏄🏽\u200d♀️",
+    "woman_swimming": "🏊\u200d♀️",
+    "woman_swimming_dark_skin_tone": "🏊🏿\u200d♀️",
+    "woman_swimming_light_skin_tone": "🏊🏻\u200d♀️",
+    "woman_swimming_medium-dark_skin_tone": "🏊🏾\u200d♀️",
+    "woman_swimming_medium-light_skin_tone": "🏊🏼\u200d♀️",
+    "woman_swimming_medium_skin_tone": "🏊🏽\u200d♀️",
+    "woman_teacher": "👩\u200d🏫",
+    "woman_teacher_dark_skin_tone": "👩🏿\u200d🏫",
+    "woman_teacher_light_skin_tone": "👩🏻\u200d🏫",
+    "woman_teacher_medium-dark_skin_tone": "👩🏾\u200d🏫",
+    "woman_teacher_medium-light_skin_tone": "👩🏼\u200d🏫",
+    "woman_teacher_medium_skin_tone": "👩🏽\u200d🏫",
+    "woman_technologist": "👩\u200d💻",
+    "woman_technologist_dark_skin_tone": "👩🏿\u200d💻",
+    "woman_technologist_light_skin_tone": "👩🏻\u200d💻",
+    "woman_technologist_medium-dark_skin_tone": "👩🏾\u200d💻",
+    "woman_technologist_medium-light_skin_tone": "👩🏼\u200d💻",
+    "woman_technologist_medium_skin_tone": "👩🏽\u200d💻",
+    "woman_tipping_hand": "💁\u200d♀️",
+    "woman_tipping_hand_dark_skin_tone": "💁🏿\u200d♀️",
+    "woman_tipping_hand_light_skin_tone": "💁🏻\u200d♀️",
+    "woman_tipping_hand_medium-dark_skin_tone": "💁🏾\u200d♀️",
+    "woman_tipping_hand_medium-light_skin_tone": "💁🏼\u200d♀️",
+    "woman_tipping_hand_medium_skin_tone": "💁🏽\u200d♀️",
+    "woman_vampire": "🧛\u200d♀️",
+    "woman_vampire_dark_skin_tone": "🧛🏿\u200d♀️",
+    "woman_vampire_light_skin_tone": "🧛🏻\u200d♀️",
+    "woman_vampire_medium-dark_skin_tone": "🧛🏾\u200d♀️",
+    "woman_vampire_medium-light_skin_tone": "🧛🏼\u200d♀️",
+    "woman_vampire_medium_skin_tone": "🧛🏽\u200d♀️",
+    "woman_walking": "🚶\u200d♀️",
+    "woman_walking_dark_skin_tone": "🚶🏿\u200d♀️",
+    "woman_walking_light_skin_tone": "🚶🏻\u200d♀️",
+    "woman_walking_medium-dark_skin_tone": "🚶🏾\u200d♀️",
+    "woman_walking_medium-light_skin_tone": "🚶🏼\u200d♀️",
+    "woman_walking_medium_skin_tone": "🚶🏽\u200d♀️",
+    "woman_wearing_turban": "👳\u200d♀️",
+    "woman_wearing_turban_dark_skin_tone": "👳🏿\u200d♀️",
+    "woman_wearing_turban_light_skin_tone": "👳🏻\u200d♀️",
+    "woman_wearing_turban_medium-dark_skin_tone": "👳🏾\u200d♀️",
+    "woman_wearing_turban_medium-light_skin_tone": "👳🏼\u200d♀️",
+    "woman_wearing_turban_medium_skin_tone": "👳🏽\u200d♀️",
+    "woman_with_headscarf": "🧕",
+    "woman_with_headscarf_dark_skin_tone": "🧕🏿",
+    "woman_with_headscarf_light_skin_tone": "🧕🏻",
+    "woman_with_headscarf_medium-dark_skin_tone": "🧕🏾",
+    "woman_with_headscarf_medium-light_skin_tone": "🧕🏼",
+    "woman_with_headscarf_medium_skin_tone": "🧕🏽",
+    "woman_with_probing_cane": "👩\u200d🦯",
+    "woman_zombie": "🧟\u200d♀️",
+    "woman’s_boot": "👢",
+    "woman’s_clothes": "👚",
+    "woman’s_hat": "👒",
+    "woman’s_sandal": "👡",
+    "women_with_bunny_ears": "👯\u200d♀️",
+    "women_wrestling": "🤼\u200d♀️",
+    "women’s_room": "🚺",
+    "woozy_face": "🥴",
+    "world_map": "🗺",
+    "worried_face": "😟",
+    "wrapped_gift": "🎁",
+    "wrench": "🔧",
+    "writing_hand": "✍",
+    "writing_hand_dark_skin_tone": "✍🏿",
+    "writing_hand_light_skin_tone": "✍🏻",
+    "writing_hand_medium-dark_skin_tone": "✍🏾",
+    "writing_hand_medium-light_skin_tone": "✍🏼",
+    "writing_hand_medium_skin_tone": "✍🏽",
+    "yarn": "🧶",
+    "yawning_face": "🥱",
+    "yellow_circle": "🟡",
+    "yellow_heart": "💛",
+    "yellow_square": "🟨",
+    "yen_banknote": "💴",
+    "yo-yo": "🪀",
+    "yin_yang": "☯",
+    "zany_face": "🤪",
+    "zebra": "🦓",
+    "zipper-mouth_face": "🤐",
+    "zombie": "🧟",
+    "zzz": "💤",
+    "åland_islands": "🇦🇽",
+    "keycap_asterisk": "*⃣",
+    "keycap_digit_eight": "8⃣",
+    "keycap_digit_five": "5⃣",
+    "keycap_digit_four": "4⃣",
+    "keycap_digit_nine": "9⃣",
+    "keycap_digit_one": "1⃣",
+    "keycap_digit_seven": "7⃣",
+    "keycap_digit_six": "6⃣",
+    "keycap_digit_three": "3⃣",
+    "keycap_digit_two": "2⃣",
+    "keycap_digit_zero": "0⃣",
+    "keycap_number_sign": "#⃣",
+    "light_skin_tone": "🏻",
+    "medium_light_skin_tone": "🏼",
+    "medium_skin_tone": "🏽",
+    "medium_dark_skin_tone": "🏾",
+    "dark_skin_tone": "🏿",
+    "regional_indicator_symbol_letter_a": "🇦",
+    "regional_indicator_symbol_letter_b": "🇧",
+    "regional_indicator_symbol_letter_c": "🇨",
+    "regional_indicator_symbol_letter_d": "🇩",
+    "regional_indicator_symbol_letter_e": "🇪",
+    "regional_indicator_symbol_letter_f": "🇫",
+    "regional_indicator_symbol_letter_g": "🇬",
+    "regional_indicator_symbol_letter_h": "🇭",
+    "regional_indicator_symbol_letter_i": "🇮",
+    "regional_indicator_symbol_letter_j": "🇯",
+    "regional_indicator_symbol_letter_k": "🇰",
+    "regional_indicator_symbol_letter_l": "🇱",
+    "regional_indicator_symbol_letter_m": "🇲",
+    "regional_indicator_symbol_letter_n": "🇳",
+    "regional_indicator_symbol_letter_o": "🇴",
+    "regional_indicator_symbol_letter_p": "🇵",
+    "regional_indicator_symbol_letter_q": "🇶",
+    "regional_indicator_symbol_letter_r": "🇷",
+    "regional_indicator_symbol_letter_s": "🇸",
+    "regional_indicator_symbol_letter_t": "🇹",
+    "regional_indicator_symbol_letter_u": "🇺",
+    "regional_indicator_symbol_letter_v": "🇻",
+    "regional_indicator_symbol_letter_w": "🇼",
+    "regional_indicator_symbol_letter_x": "🇽",
+    "regional_indicator_symbol_letter_y": "🇾",
+    "regional_indicator_symbol_letter_z": "🇿",
+    "airplane_arriving": "🛬",
+    "space_invader": "👾",
+    "football": "🏈",
+    "anger": "💢",
+    "angry": "😠",
+    "anguished": "😧",
+    "signal_strength": "📶",
+    "arrows_counterclockwise": "🔄",
+    "arrow_heading_down": "⤵",
+    "arrow_heading_up": "⤴",
+    "art": "🎨",
+    "astonished": "😲",
+    "athletic_shoe": "👟",
+    "atm": "🏧",
+    "car": "🚗",
+    "red_car": "🚗",
+    "angel": "👼",
+    "back": "🔙",
+    "badminton_racquet_and_shuttlecock": "🏸",
+    "dollar": "💵",
+    "euro": "💶",
+    "pound": "💷",
+    "yen": "💴",
+    "barber": "💈",
+    "bath": "🛀",
+    "bear": "🐻",
+    "heartbeat": "💓",
+    "beer": "🍺",
+    "no_bell": "🔕",
+    "bento": "🍱",
+    "bike": "🚲",
+    "bicyclist": "🚴",
+    "8ball": "🎱",
+    "biohazard_sign": "☣",
+    "birthday": "🎂",
+    "black_circle_for_record": "⏺",
+    "clubs": "♣",
+    "diamonds": "♦",
+    "arrow_double_down": "⏬",
+    "hearts": "♥",
+    "rewind": "⏪",
+    "black_left__pointing_double_triangle_with_vertical_bar": "⏮",
+    "arrow_backward": "◀",
+    "black_medium_small_square": "◾",
+    "question": "❓",
+    "fast_forward": "⏩",
+    "black_right__pointing_double_triangle_with_vertical_bar": "⏭",
+    "arrow_forward": "▶",
+    "black_right__pointing_triangle_with_double_vertical_bar": "⏯",
+    "arrow_right": "➡",
+    "spades": "♠",
+    "black_square_for_stop": "⏹",
+    "sunny": "☀",
+    "phone": "☎",
+    "recycle": "♻",
+    "arrow_double_up": "⏫",
+    "busstop": "🚏",
+    "date": "📅",
+    "flags": "🎏",
+    "cat2": "🐈",
+    "joy_cat": "😹",
+    "smirk_cat": "😼",
+    "chart_with_downwards_trend": "📉",
+    "chart_with_upwards_trend": "📈",
+    "chart": "💹",
+    "mega": "📣",
+    "checkered_flag": "🏁",
+    "accept": "🉑",
+    "ideograph_advantage": "🉐",
+    "congratulations": "㊗",
+    "secret": "㊙",
+    "m": "Ⓜ",
+    "city_sunset": "🌆",
+    "clapper": "🎬",
+    "clap": "👏",
+    "beers": "🍻",
+    "clock830": "🕣",
+    "clock8": "🕗",
+    "clock1130": "🕦",
+    "clock11": "🕚",
+    "clock530": "🕠",
+    "clock5": "🕔",
+    "clock430": "🕟",
+    "clock4": "🕓",
+    "clock930": "🕤",
+    "clock9": "🕘",
+    "clock130": "🕜",
+    "clock1": "🕐",
+    "clock730": "🕢",
+    "clock7": "🕖",
+    "clock630": "🕡",
+    "clock6": "🕕",
+    "clock1030": "🕥",
+    "clock10": "🕙",
+    "clock330": "🕞",
+    "clock3": "🕒",
+    "clock1230": "🕧",
+    "clock12": "🕛",
+    "clock230": "🕝",
+    "clock2": "🕑",
+    "arrows_clockwise": "🔃",
+    "repeat": "🔁",
+    "repeat_one": "🔂",
+    "closed_lock_with_key": "🔐",
+    "mailbox_closed": "📪",
+    "mailbox": "📫",
+    "cloud_with_tornado": "🌪",
+    "cocktail": "🍸",
+    "boom": "💥",
+    "compression": "🗜",
+    "confounded": "😖",
+    "confused": "😕",
+    "rice": "🍚",
+    "cow2": "🐄",
+    "cricket_bat_and_ball": "🏏",
+    "x": "❌",
+    "cry": "😢",
+    "curry": "🍛",
+    "dagger_knife": "🗡",
+    "dancer": "💃",
+    "dark_sunglasses": "🕶",
+    "dash": "💨",
+    "truck": "🚚",
+    "derelict_house_building": "🏚",
+    "diamond_shape_with_a_dot_inside": "💠",
+    "dart": "🎯",
+    "disappointed_relieved": "😥",
+    "disappointed": "😞",
+    "do_not_litter": "🚯",
+    "dog2": "🐕",
+    "flipper": "🐬",
+    "loop": "➿",
+    "bangbang": "‼",
+    "double_vertical_bar": "⏸",
+    "dove_of_peace": "🕊",
+    "small_red_triangle_down": "🔻",
+    "arrow_down_small": "🔽",
+    "arrow_down": "⬇",
+    "dromedary_camel": "🐪",
+    "e__mail": "📧",
+    "corn": "🌽",
+    "ear_of_rice": "🌾",
+    "earth_americas": "🌎",
+    "earth_asia": "🌏",
+    "earth_africa": "🌍",
+    "eight_pointed_black_star": "✴",
+    "eight_spoked_asterisk": "✳",
+    "eject_symbol": "⏏",
+    "bulb": "💡",
+    "emoji_modifier_fitzpatrick_type__1__2": "🏻",
+    "emoji_modifier_fitzpatrick_type__3": "🏼",
+    "emoji_modifier_fitzpatrick_type__4": "🏽",
+    "emoji_modifier_fitzpatrick_type__5": "🏾",
+    "emoji_modifier_fitzpatrick_type__6": "🏿",
+    "end": "🔚",
+    "email": "✉",
+    "european_castle": "🏰",
+    "european_post_office": "🏤",
+    "interrobang": "⁉",
+    "expressionless": "😑",
+    "eyeglasses": "👓",
+    "massage": "💆",
+    "yum": "😋",
+    "scream": "😱",
+    "kissing_heart": "😘",
+    "sweat": "😓",
+    "face_with_head__bandage": "🤕",
+    "triumph": "😤",
+    "mask": "😷",
+    "no_good": "🙅",
+    "ok_woman": "🙆",
+    "open_mouth": "😮",
+    "cold_sweat": "😰",
+    "stuck_out_tongue": "😛",
+    "stuck_out_tongue_closed_eyes": "😝",
+    "stuck_out_tongue_winking_eye": "😜",
+    "joy": "😂",
+    "no_mouth": "😶",
+    "santa": "🎅",
+    "fax": "📠",
+    "fearful": "😨",
+    "field_hockey_stick_and_ball": "🏑",
+    "first_quarter_moon_with_face": "🌛",
+    "fish_cake": "🍥",
+    "fishing_pole_and_fish": "🎣",
+    "facepunch": "👊",
+    "punch": "👊",
+    "flag_for_afghanistan": "🇦🇫",
+    "flag_for_albania": "🇦🇱",
+    "flag_for_algeria": "🇩🇿",
+    "flag_for_american_samoa": "🇦🇸",
+    "flag_for_andorra": "🇦🇩",
+    "flag_for_angola": "🇦🇴",
+    "flag_for_anguilla": "🇦🇮",
+    "flag_for_antarctica": "🇦🇶",
+    "flag_for_antigua_&_barbuda": "🇦🇬",
+    "flag_for_argentina": "🇦🇷",
+    "flag_for_armenia": "🇦🇲",
+    "flag_for_aruba": "🇦🇼",
+    "flag_for_ascension_island": "🇦🇨",
+    "flag_for_australia": "🇦🇺",
+    "flag_for_austria": "🇦🇹",
+    "flag_for_azerbaijan": "🇦🇿",
+    "flag_for_bahamas": "🇧🇸",
+    "flag_for_bahrain": "🇧🇭",
+    "flag_for_bangladesh": "🇧🇩",
+    "flag_for_barbados": "🇧🇧",
+    "flag_for_belarus": "🇧🇾",
+    "flag_for_belgium": "🇧🇪",
+    "flag_for_belize": "🇧🇿",
+    "flag_for_benin": "🇧🇯",
+    "flag_for_bermuda": "🇧🇲",
+    "flag_for_bhutan": "🇧🇹",
+    "flag_for_bolivia": "🇧🇴",
+    "flag_for_bosnia_&_herzegovina": "🇧🇦",
+    "flag_for_botswana": "🇧🇼",
+    "flag_for_bouvet_island": "🇧🇻",
+    "flag_for_brazil": "🇧🇷",
+    "flag_for_british_indian_ocean_territory": "🇮🇴",
+    "flag_for_british_virgin_islands": "🇻🇬",
+    "flag_for_brunei": "🇧🇳",
+    "flag_for_bulgaria": "🇧🇬",
+    "flag_for_burkina_faso": "🇧🇫",
+    "flag_for_burundi": "🇧🇮",
+    "flag_for_cambodia": "🇰🇭",
+    "flag_for_cameroon": "🇨🇲",
+    "flag_for_canada": "🇨🇦",
+    "flag_for_canary_islands": "🇮🇨",
+    "flag_for_cape_verde": "🇨🇻",
+    "flag_for_caribbean_netherlands": "🇧🇶",
+    "flag_for_cayman_islands": "🇰🇾",
+    "flag_for_central_african_republic": "🇨🇫",
+    "flag_for_ceuta_&_melilla": "🇪🇦",
+    "flag_for_chad": "🇹🇩",
+    "flag_for_chile": "🇨🇱",
+    "flag_for_china": "🇨🇳",
+    "flag_for_christmas_island": "🇨🇽",
+    "flag_for_clipperton_island": "🇨🇵",
+    "flag_for_cocos__islands": "🇨🇨",
+    "flag_for_colombia": "🇨🇴",
+    "flag_for_comoros": "🇰🇲",
+    "flag_for_congo____brazzaville": "🇨🇬",
+    "flag_for_congo____kinshasa": "🇨🇩",
+    "flag_for_cook_islands": "🇨🇰",
+    "flag_for_costa_rica": "🇨🇷",
+    "flag_for_croatia": "🇭🇷",
+    "flag_for_cuba": "🇨🇺",
+    "flag_for_curaçao": "🇨🇼",
+    "flag_for_cyprus": "🇨🇾",
+    "flag_for_czech_republic": "🇨🇿",
+    "flag_for_côte_d’ivoire": "🇨🇮",
+    "flag_for_denmark": "🇩🇰",
+    "flag_for_diego_garcia": "🇩🇬",
+    "flag_for_djibouti": "🇩🇯",
+    "flag_for_dominica": "🇩🇲",
+    "flag_for_dominican_republic": "🇩🇴",
+    "flag_for_ecuador": "🇪🇨",
+    "flag_for_egypt": "🇪🇬",
+    "flag_for_el_salvador": "🇸🇻",
+    "flag_for_equatorial_guinea": "🇬🇶",
+    "flag_for_eritrea": "🇪🇷",
+    "flag_for_estonia": "🇪🇪",
+    "flag_for_ethiopia": "🇪🇹",
+    "flag_for_european_union": "🇪🇺",
+    "flag_for_falkland_islands": "🇫🇰",
+    "flag_for_faroe_islands": "🇫🇴",
+    "flag_for_fiji": "🇫🇯",
+    "flag_for_finland": "🇫🇮",
+    "flag_for_france": "🇫🇷",
+    "flag_for_french_guiana": "🇬🇫",
+    "flag_for_french_polynesia": "🇵🇫",
+    "flag_for_french_southern_territories": "🇹🇫",
+    "flag_for_gabon": "🇬🇦",
+    "flag_for_gambia": "🇬🇲",
+    "flag_for_georgia": "🇬🇪",
+    "flag_for_germany": "🇩🇪",
+    "flag_for_ghana": "🇬🇭",
+    "flag_for_gibraltar": "🇬🇮",
+    "flag_for_greece": "🇬🇷",
+    "flag_for_greenland": "🇬🇱",
+    "flag_for_grenada": "🇬🇩",
+    "flag_for_guadeloupe": "🇬🇵",
+    "flag_for_guam": "🇬🇺",
+    "flag_for_guatemala": "🇬🇹",
+    "flag_for_guernsey": "🇬🇬",
+    "flag_for_guinea": "🇬🇳",
+    "flag_for_guinea__bissau": "🇬🇼",
+    "flag_for_guyana": "🇬🇾",
+    "flag_for_haiti": "🇭🇹",
+    "flag_for_heard_&_mcdonald_islands": "🇭🇲",
+    "flag_for_honduras": "🇭🇳",
+    "flag_for_hong_kong": "🇭🇰",
+    "flag_for_hungary": "🇭🇺",
+    "flag_for_iceland": "🇮🇸",
+    "flag_for_india": "🇮🇳",
+    "flag_for_indonesia": "🇮🇩",
+    "flag_for_iran": "🇮🇷",
+    "flag_for_iraq": "🇮🇶",
+    "flag_for_ireland": "🇮🇪",
+    "flag_for_isle_of_man": "🇮🇲",
+    "flag_for_israel": "🇮🇱",
+    "flag_for_italy": "🇮🇹",
+    "flag_for_jamaica": "🇯🇲",
+    "flag_for_japan": "🇯🇵",
+    "flag_for_jersey": "🇯🇪",
+    "flag_for_jordan": "🇯🇴",
+    "flag_for_kazakhstan": "🇰🇿",
+    "flag_for_kenya": "🇰🇪",
+    "flag_for_kiribati": "🇰🇮",
+    "flag_for_kosovo": "🇽🇰",
+    "flag_for_kuwait": "🇰🇼",
+    "flag_for_kyrgyzstan": "🇰🇬",
+    "flag_for_laos": "🇱🇦",
+    "flag_for_latvia": "🇱🇻",
+    "flag_for_lebanon": "🇱🇧",
+    "flag_for_lesotho": "🇱🇸",
+    "flag_for_liberia": "🇱🇷",
+    "flag_for_libya": "🇱🇾",
+    "flag_for_liechtenstein": "🇱🇮",
+    "flag_for_lithuania": "🇱🇹",
+    "flag_for_luxembourg": "🇱🇺",
+    "flag_for_macau": "🇲🇴",
+    "flag_for_macedonia": "🇲🇰",
+    "flag_for_madagascar": "🇲🇬",
+    "flag_for_malawi": "🇲🇼",
+    "flag_for_malaysia": "🇲🇾",
+    "flag_for_maldives": "🇲🇻",
+    "flag_for_mali": "🇲🇱",
+    "flag_for_malta": "🇲🇹",
+    "flag_for_marshall_islands": "🇲🇭",
+    "flag_for_martinique": "🇲🇶",
+    "flag_for_mauritania": "🇲🇷",
+    "flag_for_mauritius": "🇲🇺",
+    "flag_for_mayotte": "🇾🇹",
+    "flag_for_mexico": "🇲🇽",
+    "flag_for_micronesia": "🇫🇲",
+    "flag_for_moldova": "🇲🇩",
+    "flag_for_monaco": "🇲🇨",
+    "flag_for_mongolia": "🇲🇳",
+    "flag_for_montenegro": "🇲🇪",
+    "flag_for_montserrat": "🇲🇸",
+    "flag_for_morocco": "🇲🇦",
+    "flag_for_mozambique": "🇲🇿",
+    "flag_for_myanmar": "🇲🇲",
+    "flag_for_namibia": "🇳🇦",
+    "flag_for_nauru": "🇳🇷",
+    "flag_for_nepal": "🇳🇵",
+    "flag_for_netherlands": "🇳🇱",
+    "flag_for_new_caledonia": "🇳🇨",
+    "flag_for_new_zealand": "🇳🇿",
+    "flag_for_nicaragua": "🇳🇮",
+    "flag_for_niger": "🇳🇪",
+    "flag_for_nigeria": "🇳🇬",
+    "flag_for_niue": "🇳🇺",
+    "flag_for_norfolk_island": "🇳🇫",
+    "flag_for_north_korea": "🇰🇵",
+    "flag_for_northern_mariana_islands": "🇲🇵",
+    "flag_for_norway": "🇳🇴",
+    "flag_for_oman": "🇴🇲",
+    "flag_for_pakistan": "🇵🇰",
+    "flag_for_palau": "🇵🇼",
+    "flag_for_palestinian_territories": "🇵🇸",
+    "flag_for_panama": "🇵🇦",
+    "flag_for_papua_new_guinea": "🇵🇬",
+    "flag_for_paraguay": "🇵🇾",
+    "flag_for_peru": "🇵🇪",
+    "flag_for_philippines": "🇵🇭",
+    "flag_for_pitcairn_islands": "🇵🇳",
+    "flag_for_poland": "🇵🇱",
+    "flag_for_portugal": "🇵🇹",
+    "flag_for_puerto_rico": "🇵🇷",
+    "flag_for_qatar": "🇶🇦",
+    "flag_for_romania": "🇷🇴",
+    "flag_for_russia": "🇷🇺",
+    "flag_for_rwanda": "🇷🇼",
+    "flag_for_réunion": "🇷🇪",
+    "flag_for_samoa": "🇼🇸",
+    "flag_for_san_marino": "🇸🇲",
+    "flag_for_saudi_arabia": "🇸🇦",
+    "flag_for_senegal": "🇸🇳",
+    "flag_for_serbia": "🇷🇸",
+    "flag_for_seychelles": "🇸🇨",
+    "flag_for_sierra_leone": "🇸🇱",
+    "flag_for_singapore": "🇸🇬",
+    "flag_for_sint_maarten": "🇸🇽",
+    "flag_for_slovakia": "🇸🇰",
+    "flag_for_slovenia": "🇸🇮",
+    "flag_for_solomon_islands": "🇸🇧",
+    "flag_for_somalia": "🇸🇴",
+    "flag_for_south_africa": "🇿🇦",
+    "flag_for_south_georgia_&_south_sandwich_islands": "🇬🇸",
+    "flag_for_south_korea": "🇰🇷",
+    "flag_for_south_sudan": "🇸🇸",
+    "flag_for_spain": "🇪🇸",
+    "flag_for_sri_lanka": "🇱🇰",
+    "flag_for_st._barthélemy": "🇧🇱",
+    "flag_for_st._helena": "🇸🇭",
+    "flag_for_st._kitts_&_nevis": "🇰🇳",
+    "flag_for_st._lucia": "🇱🇨",
+    "flag_for_st._martin": "🇲🇫",
+    "flag_for_st._pierre_&_miquelon": "🇵🇲",
+    "flag_for_st._vincent_&_grenadines": "🇻🇨",
+    "flag_for_sudan": "🇸🇩",
+    "flag_for_suriname": "🇸🇷",
+    "flag_for_svalbard_&_jan_mayen": "🇸🇯",
+    "flag_for_swaziland": "🇸🇿",
+    "flag_for_sweden": "🇸🇪",
+    "flag_for_switzerland": "🇨🇭",
+    "flag_for_syria": "🇸🇾",
+    "flag_for_são_tomé_&_príncipe": "🇸🇹",
+    "flag_for_taiwan": "🇹🇼",
+    "flag_for_tajikistan": "🇹🇯",
+    "flag_for_tanzania": "🇹🇿",
+    "flag_for_thailand": "🇹🇭",
+    "flag_for_timor__leste": "🇹🇱",
+    "flag_for_togo": "🇹🇬",
+    "flag_for_tokelau": "🇹🇰",
+    "flag_for_tonga": "🇹🇴",
+    "flag_for_trinidad_&_tobago": "🇹🇹",
+    "flag_for_tristan_da_cunha": "🇹🇦",
+    "flag_for_tunisia": "🇹🇳",
+    "flag_for_turkey": "🇹🇷",
+    "flag_for_turkmenistan": "🇹🇲",
+    "flag_for_turks_&_caicos_islands": "🇹🇨",
+    "flag_for_tuvalu": "🇹🇻",
+    "flag_for_u.s._outlying_islands": "🇺🇲",
+    "flag_for_u.s._virgin_islands": "🇻🇮",
+    "flag_for_uganda": "🇺🇬",
+    "flag_for_ukraine": "🇺🇦",
+    "flag_for_united_arab_emirates": "🇦🇪",
+    "flag_for_united_kingdom": "🇬🇧",
+    "flag_for_united_states": "🇺🇸",
+    "flag_for_uruguay": "🇺🇾",
+    "flag_for_uzbekistan": "🇺🇿",
+    "flag_for_vanuatu": "🇻🇺",
+    "flag_for_vatican_city": "🇻🇦",
+    "flag_for_venezuela": "🇻🇪",
+    "flag_for_vietnam": "🇻🇳",
+    "flag_for_wallis_&_futuna": "🇼🇫",
+    "flag_for_western_sahara": "🇪🇭",
+    "flag_for_yemen": "🇾🇪",
+    "flag_for_zambia": "🇿🇲",
+    "flag_for_zimbabwe": "🇿🇼",
+    "flag_for_åland_islands": "🇦🇽",
+    "golf": "⛳",
+    "fleur__de__lis": "⚜",
+    "muscle": "💪",
+    "flushed": "😳",
+    "frame_with_picture": "🖼",
+    "fries": "🍟",
+    "frog": "🐸",
+    "hatched_chick": "🐥",
+    "frowning": "😦",
+    "fuelpump": "⛽",
+    "full_moon_with_face": "🌝",
+    "gem": "💎",
+    "star2": "🌟",
+    "golfer": "🏌",
+    "mortar_board": "🎓",
+    "grimacing": "😬",
+    "smile_cat": "😸",
+    "grinning": "😀",
+    "grin": "😁",
+    "heartpulse": "💗",
+    "guardsman": "💂",
+    "haircut": "💇",
+    "hamster": "🐹",
+    "raising_hand": "🙋",
+    "headphones": "🎧",
+    "hear_no_evil": "🙉",
+    "cupid": "💘",
+    "gift_heart": "💝",
+    "heart": "❤",
+    "exclamation": "❗",
+    "heavy_exclamation_mark": "❗",
+    "heavy_heart_exclamation_mark_ornament": "❣",
+    "o": "⭕",
+    "helm_symbol": "⎈",
+    "helmet_with_white_cross": "⛑",
+    "high_heel": "👠",
+    "bullettrain_side": "🚄",
+    "bullettrain_front": "🚅",
+    "high_brightness": "🔆",
+    "zap": "⚡",
+    "hocho": "🔪",
+    "knife": "🔪",
+    "bee": "🐝",
+    "traffic_light": "🚥",
+    "racehorse": "🐎",
+    "coffee": "☕",
+    "hotsprings": "♨",
+    "hourglass": "⌛",
+    "hourglass_flowing_sand": "⏳",
+    "house_buildings": "🏘",
+    "100": "💯",
+    "hushed": "😯",
+    "ice_hockey_stick_and_puck": "🏒",
+    "imp": "👿",
+    "information_desk_person": "💁",
+    "information_source": "ℹ",
+    "capital_abcd": "🔠",
+    "abc": "🔤",
+    "abcd": "🔡",
+    "1234": "🔢",
+    "symbols": "🔣",
+    "izakaya_lantern": "🏮",
+    "lantern": "🏮",
+    "jack_o_lantern": "🎃",
+    "dolls": "🎎",
+    "japanese_goblin": "👺",
+    "japanese_ogre": "👹",
+    "beginner": "🔰",
+    "zero": "0️⃣",
+    "one": "1️⃣",
+    "ten": "🔟",
+    "two": "2️⃣",
+    "three": "3️⃣",
+    "four": "4️⃣",
+    "five": "5️⃣",
+    "six": "6️⃣",
+    "seven": "7️⃣",
+    "eight": "8️⃣",
+    "nine": "9️⃣",
+    "couplekiss": "💏",
+    "kissing_cat": "😽",
+    "kissing": "😗",
+    "kissing_closed_eyes": "😚",
+    "kissing_smiling_eyes": "😙",
+    "beetle": "🐞",
+    "large_blue_circle": "🔵",
+    "last_quarter_moon_with_face": "🌜",
+    "leaves": "🍃",
+    "mag": "🔍",
+    "left_right_arrow": "↔",
+    "leftwards_arrow_with_hook": "↩",
+    "arrow_left": "⬅",
+    "lock": "🔒",
+    "lock_with_ink_pen": "🔏",
+    "sob": "😭",
+    "low_brightness": "🔅",
+    "lower_left_ballpoint_pen": "🖊",
+    "lower_left_crayon": "🖍",
+    "lower_left_fountain_pen": "🖋",
+    "lower_left_paintbrush": "🖌",
+    "mahjong": "🀄",
+    "couple": "👫",
+    "man_in_business_suit_levitating": "🕴",
+    "man_with_gua_pi_mao": "👲",
+    "man_with_turban": "👳",
+    "mans_shoe": "👞",
+    "shoe": "👞",
+    "menorah_with_nine_branches": "🕎",
+    "mens": "🚹",
+    "minidisc": "💽",
+    "iphone": "📱",
+    "calling": "📲",
+    "money__mouth_face": "🤑",
+    "moneybag": "💰",
+    "rice_scene": "🎑",
+    "mountain_bicyclist": "🚵",
+    "mouse2": "🐁",
+    "lips": "👄",
+    "moyai": "🗿",
+    "notes": "🎶",
+    "nail_care": "💅",
+    "ab": "🆎",
+    "negative_squared_cross_mark": "❎",
+    "a": "🅰",
+    "b": "🅱",
+    "o2": "🅾",
+    "parking": "🅿",
+    "new_moon_with_face": "🌚",
+    "no_entry_sign": "🚫",
+    "underage": "🔞",
+    "non__potable_water": "🚱",
+    "arrow_upper_right": "↗",
+    "arrow_upper_left": "↖",
+    "office": "🏢",
+    "older_man": "👴",
+    "older_woman": "👵",
+    "om_symbol": "🕉",
+    "on": "🔛",
+    "book": "📖",
+    "unlock": "🔓",
+    "mailbox_with_no_mail": "📭",
+    "mailbox_with_mail": "📬",
+    "cd": "💿",
+    "tada": "🎉",
+    "feet": "🐾",
+    "walking": "🚶",
+    "pencil2": "✏",
+    "pensive": "😔",
+    "persevere": "😣",
+    "bow": "🙇",
+    "raised_hands": "🙌",
+    "person_with_ball": "⛹",
+    "person_with_blond_hair": "👱",
+    "pray": "🙏",
+    "person_with_pouting_face": "🙎",
+    "computer": "💻",
+    "pig2": "🐖",
+    "hankey": "💩",
+    "poop": "💩",
+    "shit": "💩",
+    "bamboo": "🎍",
+    "gun": "🔫",
+    "black_joker": "🃏",
+    "rotating_light": "🚨",
+    "cop": "👮",
+    "stew": "🍲",
+    "pouch": "👝",
+    "pouting_cat": "😾",
+    "rage": "😡",
+    "put_litter_in_its_place": "🚮",
+    "rabbit2": "🐇",
+    "racing_motorcycle": "🏍",
+    "radioactive_sign": "☢",
+    "fist": "✊",
+    "hand": "✋",
+    "raised_hand_with_fingers_splayed": "🖐",
+    "raised_hand_with_part_between_middle_and_ring_fingers": "🖖",
+    "blue_car": "🚙",
+    "apple": "🍎",
+    "relieved": "😌",
+    "reversed_hand_with_middle_finger_extended": "🖕",
+    "mag_right": "🔎",
+    "arrow_right_hook": "↪",
+    "sweet_potato": "🍠",
+    "robot": "🤖",
+    "rolled__up_newspaper": "🗞",
+    "rowboat": "🚣",
+    "runner": "🏃",
+    "running": "🏃",
+    "running_shirt_with_sash": "🎽",
+    "boat": "⛵",
+    "scales": "⚖",
+    "school_satchel": "🎒",
+    "scorpius": "♏",
+    "see_no_evil": "🙈",
+    "sheep": "🐑",
+    "stars": "🌠",
+    "cake": "🍰",
+    "six_pointed_star": "🔯",
+    "ski": "🎿",
+    "sleeping_accommodation": "🛌",
+    "sleeping": "😴",
+    "sleepy": "😪",
+    "sleuth_or_spy": "🕵",
+    "heart_eyes_cat": "😻",
+    "smiley_cat": "😺",
+    "innocent": "😇",
+    "heart_eyes": "😍",
+    "smiling_imp": "😈",
+    "smiley": "😃",
+    "sweat_smile": "😅",
+    "smile": "😄",
+    "laughing": "😆",
+    "satisfied": "😆",
+    "blush": "😊",
+    "smirk": "😏",
+    "smoking": "🚬",
+    "snow_capped_mountain": "🏔",
+    "soccer": "⚽",
+    "icecream": "🍦",
+    "soon": "🔜",
+    "arrow_lower_right": "↘",
+    "arrow_lower_left": "↙",
+    "speak_no_evil": "🙊",
+    "speaker": "🔈",
+    "mute": "🔇",
+    "sound": "🔉",
+    "loud_sound": "🔊",
+    "speaking_head_in_silhouette": "🗣",
+    "spiral_calendar_pad": "🗓",
+    "spiral_note_pad": "🗒",
+    "shell": "🐚",
+    "sweat_drops": "💦",
+    "u5272": "🈹",
+    "u5408": "🈴",
+    "u55b6": "🈺",
+    "u6307": "🈯",
+    "u6708": "🈷",
+    "u6709": "🈶",
+    "u6e80": "🈵",
+    "u7121": "🈚",
+    "u7533": "🈸",
+    "u7981": "🈲",
+    "u7a7a": "🈳",
+    "cl": "🆑",
+    "cool": "🆒",
+    "free": "🆓",
+    "id": "🆔",
+    "koko": "🈁",
+    "sa": "🈂",
+    "new": "🆕",
+    "ng": "🆖",
+    "ok": "🆗",
+    "sos": "🆘",
+    "up": "🆙",
+    "vs": "🆚",
+    "steam_locomotive": "🚂",
+    "ramen": "🍜",
+    "partly_sunny": "⛅",
+    "city_sunrise": "🌇",
+    "surfer": "🏄",
+    "swimmer": "🏊",
+    "shirt": "👕",
+    "tshirt": "👕",
+    "table_tennis_paddle_and_ball": "🏓",
+    "tea": "🍵",
+    "tv": "📺",
+    "three_button_mouse": "🖱",
+    "+1": "👍",
+    "thumbsup": "👍",
+    "__1": "👎",
+    "-1": "👎",
+    "thumbsdown": "👎",
+    "thunder_cloud_and_rain": "⛈",
+    "tiger2": "🐅",
+    "tophat": "🎩",
+    "top": "🔝",
+    "tm": "™",
+    "train2": "🚆",
+    "triangular_flag_on_post": "🚩",
+    "trident": "🔱",
+    "twisted_rightwards_arrows": "🔀",
+    "unamused": "😒",
+    "small_red_triangle": "🔺",
+    "arrow_up_small": "🔼",
+    "arrow_up_down": "↕",
+    "upside__down_face": "🙃",
+    "arrow_up": "⬆",
+    "v": "✌",
+    "vhs": "📼",
+    "wc": "🚾",
+    "ocean": "🌊",
+    "waving_black_flag": "🏴",
+    "wave": "👋",
+    "waving_white_flag": "🏳",
+    "moon": "🌔",
+    "scream_cat": "🙀",
+    "weary": "😩",
+    "weight_lifter": "🏋",
+    "whale2": "🐋",
+    "wheelchair": "♿",
+    "point_down": "👇",
+    "grey_exclamation": "❕",
+    "white_frowning_face": "☹",
+    "white_check_mark": "✅",
+    "point_left": "👈",
+    "white_medium_small_square": "◽",
+    "star": "⭐",
+    "grey_question": "❔",
+    "point_right": "👉",
+    "relaxed": "☺",
+    "white_sun_behind_cloud": "🌥",
+    "white_sun_behind_cloud_with_rain": "🌦",
+    "white_sun_with_small_cloud": "🌤",
+    "point_up_2": "👆",
+    "point_up": "☝",
+    "wind_blowing_face": "🌬",
+    "wink": "😉",
+    "wolf": "🐺",
+    "dancers": "👯",
+    "boot": "👢",
+    "womans_clothes": "👚",
+    "womans_hat": "👒",
+    "sandal": "👡",
+    "womens": "🚺",
+    "worried": "😟",
+    "gift": "🎁",
+    "zipper__mouth_face": "🤐",
+    "regional_indicator_a": "🇦",
+    "regional_indicator_b": "🇧",
+    "regional_indicator_c": "🇨",
+    "regional_indicator_d": "🇩",
+    "regional_indicator_e": "🇪",
+    "regional_indicator_f": "🇫",
+    "regional_indicator_g": "🇬",
+    "regional_indicator_h": "🇭",
+    "regional_indicator_i": "🇮",
+    "regional_indicator_j": "🇯",
+    "regional_indicator_k": "🇰",
+    "regional_indicator_l": "🇱",
+    "regional_indicator_m": "🇲",
+    "regional_indicator_n": "🇳",
+    "regional_indicator_o": "🇴",
+    "regional_indicator_p": "🇵",
+    "regional_indicator_q": "🇶",
+    "regional_indicator_r": "🇷",
+    "regional_indicator_s": "🇸",
+    "regional_indicator_t": "🇹",
+    "regional_indicator_u": "🇺",
+    "regional_indicator_v": "🇻",
+    "regional_indicator_w": "🇼",
+    "regional_indicator_x": "🇽",
+    "regional_indicator_y": "🇾",
+    "regional_indicator_z": "🇿",
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_replace.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_replace.py
new file mode 100644
index 00000000..bb2cafa1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_emoji_replace.py
@@ -0,0 +1,32 @@
+from typing import Callable, Match, Optional
+import re
+
+from ._emoji_codes import EMOJI
+
+
+_ReStringMatch = Match[str]  # regex match object
+_ReSubCallable = Callable[[_ReStringMatch], str]  # Callable invoked by re.sub
+_EmojiSubMethod = Callable[[_ReSubCallable, str], str]  # Sub method of a compiled re
+
+
+def _emoji_replace(
+    text: str,
+    default_variant: Optional[str] = None,
+    _emoji_sub: _EmojiSubMethod = re.compile(r"(:(\S*?)(?:(?:\-)(emoji|text))?:)").sub,
+) -> str:
+    """Replace emoji code in text."""
+    get_emoji = EMOJI.__getitem__
+    variants = {"text": "\uFE0E", "emoji": "\uFE0F"}
+    get_variant = variants.get
+    default_variant_code = variants.get(default_variant, "") if default_variant else ""
+
+    def do_replace(match: Match[str]) -> str:
+        emoji_code, emoji_name, variant = match.groups()
+        try:
+            return get_emoji(emoji_name.lower()) + get_variant(
+                variant, default_variant_code
+            )
+        except KeyError:
+            return emoji_code
+
+    return _emoji_sub(do_replace, text)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_export_format.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_export_format.py
new file mode 100644
index 00000000..e7527e52
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_export_format.py
@@ -0,0 +1,76 @@
+CONSOLE_HTML_FORMAT = """\
+<!DOCTYPE html>
+<html>
+<head>
+<meta charset="UTF-8">
+<style>
+{stylesheet}
+body {{
+    color: {foreground};
+    background-color: {background};
+}}
+</style>
+</head>
+<body>
+    <pre style="font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"><code style="font-family:inherit">{code}</code></pre>
+</body>
+</html>
+"""
+
+CONSOLE_SVG_FORMAT = """\
+<svg class="rich-terminal" viewBox="0 0 {width} {height}" xmlns="http://www.w3.org/2000/svg">
+    <!-- Generated with Rich https://www.textualize.io -->
+    <style>
+
+    @font-face {{
+        font-family: "Fira Code";
+        src: local("FiraCode-Regular"),
+                url("https://cdnjs.cloudflare.com/ajax/libs/firacode/6.2.0/woff2/FiraCode-Regular.woff2") format("woff2"),
+                url("https://cdnjs.cloudflare.com/ajax/libs/firacode/6.2.0/woff/FiraCode-Regular.woff") format("woff");
+        font-style: normal;
+        font-weight: 400;
+    }}
+    @font-face {{
+        font-family: "Fira Code";
+        src: local("FiraCode-Bold"),
+                url("https://cdnjs.cloudflare.com/ajax/libs/firacode/6.2.0/woff2/FiraCode-Bold.woff2") format("woff2"),
+                url("https://cdnjs.cloudflare.com/ajax/libs/firacode/6.2.0/woff/FiraCode-Bold.woff") format("woff");
+        font-style: bold;
+        font-weight: 700;
+    }}
+
+    .{unique_id}-matrix {{
+        font-family: Fira Code, monospace;
+        font-size: {char_height}px;
+        line-height: {line_height}px;
+        font-variant-east-asian: full-width;
+    }}
+
+    .{unique_id}-title {{
+        font-size: 18px;
+        font-weight: bold;
+        font-family: arial;
+    }}
+
+    {styles}
+    </style>
+
+    <defs>
+    <clipPath id="{unique_id}-clip-terminal">
+      <rect x="0" y="0" width="{terminal_width}" height="{terminal_height}" />
+    </clipPath>
+    {lines}
+    </defs>
+
+    {chrome}
+    <g transform="translate({terminal_x}, {terminal_y})" clip-path="url(#{unique_id}-clip-terminal)">
+    {backgrounds}
+    <g class="{unique_id}-matrix">
+    {matrix}
+    </g>
+    </g>
+</svg>
+"""
+
+_SVG_FONT_FAMILY = "Rich Fira Code"
+_SVG_CLASSES_PREFIX = "rich-svg"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_extension.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_extension.py
new file mode 100644
index 00000000..cbd6da9b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_extension.py
@@ -0,0 +1,10 @@
+from typing import Any
+
+
+def load_ipython_extension(ip: Any) -> None:  # pragma: no cover
+    # prevent circular import
+    from pip._vendor.rich.pretty import install
+    from pip._vendor.rich.traceback import install as tr_install
+
+    install()
+    tr_install()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_fileno.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_fileno.py
new file mode 100644
index 00000000..b17ee651
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_fileno.py
@@ -0,0 +1,24 @@
+from __future__ import annotations
+
+from typing import IO, Callable
+
+
+def get_fileno(file_like: IO[str]) -> int | None:
+    """Get fileno() from a file, accounting for poorly implemented file-like objects.
+
+    Args:
+        file_like (IO): A file-like object.
+
+    Returns:
+        int | None: The result of fileno if available, or None if operation failed.
+    """
+    fileno: Callable[[], int] | None = getattr(file_like, "fileno", None)
+    if fileno is not None:
+        try:
+            return fileno()
+        except Exception:
+            # `fileno` is documented as potentially raising a OSError
+            # Alas, from the issues, there are so many poorly implemented file-like objects,
+            # that `fileno()` can raise just about anything.
+            return None
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_inspect.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_inspect.py
new file mode 100644
index 00000000..e87698d1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_inspect.py
@@ -0,0 +1,268 @@
+import inspect
+from inspect import cleandoc, getdoc, getfile, isclass, ismodule, signature
+from typing import Any, Collection, Iterable, Optional, Tuple, Type, Union
+
+from .console import Group, RenderableType
+from .control import escape_control_codes
+from .highlighter import ReprHighlighter
+from .jupyter import JupyterMixin
+from .panel import Panel
+from .pretty import Pretty
+from .table import Table
+from .text import Text, TextType
+
+
+def _first_paragraph(doc: str) -> str:
+    """Get the first paragraph from a docstring."""
+    paragraph, _, _ = doc.partition("\n\n")
+    return paragraph
+
+
+class Inspect(JupyterMixin):
+    """A renderable to inspect any Python Object.
+
+    Args:
+        obj (Any): An object to inspect.
+        title (str, optional): Title to display over inspect result, or None use type. Defaults to None.
+        help (bool, optional): Show full help text rather than just first paragraph. Defaults to False.
+        methods (bool, optional): Enable inspection of callables. Defaults to False.
+        docs (bool, optional): Also render doc strings. Defaults to True.
+        private (bool, optional): Show private attributes (beginning with underscore). Defaults to False.
+        dunder (bool, optional): Show attributes starting with double underscore. Defaults to False.
+        sort (bool, optional): Sort attributes alphabetically. Defaults to True.
+        all (bool, optional): Show all attributes. Defaults to False.
+        value (bool, optional): Pretty print value of object. Defaults to True.
+    """
+
+    def __init__(
+        self,
+        obj: Any,
+        *,
+        title: Optional[TextType] = None,
+        help: bool = False,
+        methods: bool = False,
+        docs: bool = True,
+        private: bool = False,
+        dunder: bool = False,
+        sort: bool = True,
+        all: bool = True,
+        value: bool = True,
+    ) -> None:
+        self.highlighter = ReprHighlighter()
+        self.obj = obj
+        self.title = title or self._make_title(obj)
+        if all:
+            methods = private = dunder = True
+        self.help = help
+        self.methods = methods
+        self.docs = docs or help
+        self.private = private or dunder
+        self.dunder = dunder
+        self.sort = sort
+        self.value = value
+
+    def _make_title(self, obj: Any) -> Text:
+        """Make a default title."""
+        title_str = (
+            str(obj)
+            if (isclass(obj) or callable(obj) or ismodule(obj))
+            else str(type(obj))
+        )
+        title_text = self.highlighter(title_str)
+        return title_text
+
+    def __rich__(self) -> Panel:
+        return Panel.fit(
+            Group(*self._render()),
+            title=self.title,
+            border_style="scope.border",
+            padding=(0, 1),
+        )
+
+    def _get_signature(self, name: str, obj: Any) -> Optional[Text]:
+        """Get a signature for a callable."""
+        try:
+            _signature = str(signature(obj)) + ":"
+        except ValueError:
+            _signature = "(...)"
+        except TypeError:
+            return None
+
+        source_filename: Optional[str] = None
+        try:
+            source_filename = getfile(obj)
+        except (OSError, TypeError):
+            # OSError is raised if obj has no source file, e.g. when defined in REPL.
+            pass
+
+        callable_name = Text(name, style="inspect.callable")
+        if source_filename:
+            callable_name.stylize(f"link file://{source_filename}")
+        signature_text = self.highlighter(_signature)
+
+        qualname = name or getattr(obj, "__qualname__", name)
+
+        # If obj is a module, there may be classes (which are callable) to display
+        if inspect.isclass(obj):
+            prefix = "class"
+        elif inspect.iscoroutinefunction(obj):
+            prefix = "async def"
+        else:
+            prefix = "def"
+
+        qual_signature = Text.assemble(
+            (f"{prefix} ", f"inspect.{prefix.replace(' ', '_')}"),
+            (qualname, "inspect.callable"),
+            signature_text,
+        )
+
+        return qual_signature
+
+    def _render(self) -> Iterable[RenderableType]:
+        """Render object."""
+
+        def sort_items(item: Tuple[str, Any]) -> Tuple[bool, str]:
+            key, (_error, value) = item
+            return (callable(value), key.strip("_").lower())
+
+        def safe_getattr(attr_name: str) -> Tuple[Any, Any]:
+            """Get attribute or any exception."""
+            try:
+                return (None, getattr(obj, attr_name))
+            except Exception as error:
+                return (error, None)
+
+        obj = self.obj
+        keys = dir(obj)
+        total_items = len(keys)
+        if not self.dunder:
+            keys = [key for key in keys if not key.startswith("__")]
+        if not self.private:
+            keys = [key for key in keys if not key.startswith("_")]
+        not_shown_count = total_items - len(keys)
+        items = [(key, safe_getattr(key)) for key in keys]
+        if self.sort:
+            items.sort(key=sort_items)
+
+        items_table = Table.grid(padding=(0, 1), expand=False)
+        items_table.add_column(justify="right")
+        add_row = items_table.add_row
+        highlighter = self.highlighter
+
+        if callable(obj):
+            signature = self._get_signature("", obj)
+            if signature is not None:
+                yield signature
+                yield ""
+
+        if self.docs:
+            _doc = self._get_formatted_doc(obj)
+            if _doc is not None:
+                doc_text = Text(_doc, style="inspect.help")
+                doc_text = highlighter(doc_text)
+                yield doc_text
+                yield ""
+
+        if self.value and not (isclass(obj) or callable(obj) or ismodule(obj)):
+            yield Panel(
+                Pretty(obj, indent_guides=True, max_length=10, max_string=60),
+                border_style="inspect.value.border",
+            )
+            yield ""
+
+        for key, (error, value) in items:
+            key_text = Text.assemble(
+                (
+                    key,
+                    "inspect.attr.dunder" if key.startswith("__") else "inspect.attr",
+                ),
+                (" =", "inspect.equals"),
+            )
+            if error is not None:
+                warning = key_text.copy()
+                warning.stylize("inspect.error")
+                add_row(warning, highlighter(repr(error)))
+                continue
+
+            if callable(value):
+                if not self.methods:
+                    continue
+
+                _signature_text = self._get_signature(key, value)
+                if _signature_text is None:
+                    add_row(key_text, Pretty(value, highlighter=highlighter))
+                else:
+                    if self.docs:
+                        docs = self._get_formatted_doc(value)
+                        if docs is not None:
+                            _signature_text.append("\n" if "\n" in docs else " ")
+                            doc = highlighter(docs)
+                            doc.stylize("inspect.doc")
+                            _signature_text.append(doc)
+
+                    add_row(key_text, _signature_text)
+            else:
+                add_row(key_text, Pretty(value, highlighter=highlighter))
+        if items_table.row_count:
+            yield items_table
+        elif not_shown_count:
+            yield Text.from_markup(
+                f"[b cyan]{not_shown_count}[/][i] attribute(s) not shown.[/i] "
+                f"Run [b][magenta]inspect[/]([not b]inspect[/])[/b] for options."
+            )
+
+    def _get_formatted_doc(self, object_: Any) -> Optional[str]:
+        """
+        Extract the docstring of an object, process it and returns it.
+        The processing consists in cleaning up the doctring's indentation,
+        taking only its 1st paragraph if `self.help` is not True,
+        and escape its control codes.
+
+        Args:
+            object_ (Any): the object to get the docstring from.
+
+        Returns:
+            Optional[str]: the processed docstring, or None if no docstring was found.
+        """
+        docs = getdoc(object_)
+        if docs is None:
+            return None
+        docs = cleandoc(docs).strip()
+        if not self.help:
+            docs = _first_paragraph(docs)
+        return escape_control_codes(docs)
+
+
+def get_object_types_mro(obj: Union[object, Type[Any]]) -> Tuple[type, ...]:
+    """Returns the MRO of an object's class, or of the object itself if it's a class."""
+    if not hasattr(obj, "__mro__"):
+        # N.B. we cannot use `if type(obj) is type` here because it doesn't work with
+        # some types of classes, such as the ones that use abc.ABCMeta.
+        obj = type(obj)
+    return getattr(obj, "__mro__", ())
+
+
+def get_object_types_mro_as_strings(obj: object) -> Collection[str]:
+    """
+    Returns the MRO of an object's class as full qualified names, or of the object itself if it's a class.
+
+    Examples:
+        `object_types_mro_as_strings(JSONDecoder)` will return `['json.decoder.JSONDecoder', 'builtins.object']`
+    """
+    return [
+        f'{getattr(type_, "__module__", "")}.{getattr(type_, "__qualname__", "")}'
+        for type_ in get_object_types_mro(obj)
+    ]
+
+
+def is_object_one_of_types(
+    obj: object, fully_qualified_types_names: Collection[str]
+) -> bool:
+    """
+    Returns `True` if the given object's class (or the object itself, if it's a class) has one of the
+    fully qualified names in its MRO.
+    """
+    for type_name in get_object_types_mro_as_strings(obj):
+        if type_name in fully_qualified_types_names:
+            return True
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_log_render.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_log_render.py
new file mode 100644
index 00000000..fc16c844
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_log_render.py
@@ -0,0 +1,94 @@
+from datetime import datetime
+from typing import Iterable, List, Optional, TYPE_CHECKING, Union, Callable
+
+
+from .text import Text, TextType
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleRenderable, RenderableType
+    from .table import Table
+
+FormatTimeCallable = Callable[[datetime], Text]
+
+
+class LogRender:
+    def __init__(
+        self,
+        show_time: bool = True,
+        show_level: bool = False,
+        show_path: bool = True,
+        time_format: Union[str, FormatTimeCallable] = "[%x %X]",
+        omit_repeated_times: bool = True,
+        level_width: Optional[int] = 8,
+    ) -> None:
+        self.show_time = show_time
+        self.show_level = show_level
+        self.show_path = show_path
+        self.time_format = time_format
+        self.omit_repeated_times = omit_repeated_times
+        self.level_width = level_width
+        self._last_time: Optional[Text] = None
+
+    def __call__(
+        self,
+        console: "Console",
+        renderables: Iterable["ConsoleRenderable"],
+        log_time: Optional[datetime] = None,
+        time_format: Optional[Union[str, FormatTimeCallable]] = None,
+        level: TextType = "",
+        path: Optional[str] = None,
+        line_no: Optional[int] = None,
+        link_path: Optional[str] = None,
+    ) -> "Table":
+        from .containers import Renderables
+        from .table import Table
+
+        output = Table.grid(padding=(0, 1))
+        output.expand = True
+        if self.show_time:
+            output.add_column(style="log.time")
+        if self.show_level:
+            output.add_column(style="log.level", width=self.level_width)
+        output.add_column(ratio=1, style="log.message", overflow="fold")
+        if self.show_path and path:
+            output.add_column(style="log.path")
+        row: List["RenderableType"] = []
+        if self.show_time:
+            log_time = log_time or console.get_datetime()
+            time_format = time_format or self.time_format
+            if callable(time_format):
+                log_time_display = time_format(log_time)
+            else:
+                log_time_display = Text(log_time.strftime(time_format))
+            if log_time_display == self._last_time and self.omit_repeated_times:
+                row.append(Text(" " * len(log_time_display)))
+            else:
+                row.append(log_time_display)
+                self._last_time = log_time_display
+        if self.show_level:
+            row.append(level)
+
+        row.append(Renderables(renderables))
+        if self.show_path and path:
+            path_text = Text()
+            path_text.append(
+                path, style=f"link file://{link_path}" if link_path else ""
+            )
+            if line_no:
+                path_text.append(":")
+                path_text.append(
+                    f"{line_no}",
+                    style=f"link file://{link_path}#{line_no}" if link_path else "",
+                )
+            row.append(path_text)
+
+        output.add_row(*row)
+        return output
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console
+
+    c = Console()
+    c.print("[on blue]Hello", justify="right")
+    c.log("[on blue]hello", justify="right")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_loop.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_loop.py
new file mode 100644
index 00000000..01c6cafb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_loop.py
@@ -0,0 +1,43 @@
+from typing import Iterable, Tuple, TypeVar
+
+T = TypeVar("T")
+
+
+def loop_first(values: Iterable[T]) -> Iterable[Tuple[bool, T]]:
+    """Iterate and generate a tuple with a flag for first value."""
+    iter_values = iter(values)
+    try:
+        value = next(iter_values)
+    except StopIteration:
+        return
+    yield True, value
+    for value in iter_values:
+        yield False, value
+
+
+def loop_last(values: Iterable[T]) -> Iterable[Tuple[bool, T]]:
+    """Iterate and generate a tuple with a flag for last value."""
+    iter_values = iter(values)
+    try:
+        previous_value = next(iter_values)
+    except StopIteration:
+        return
+    for value in iter_values:
+        yield False, previous_value
+        previous_value = value
+    yield True, previous_value
+
+
+def loop_first_last(values: Iterable[T]) -> Iterable[Tuple[bool, bool, T]]:
+    """Iterate and generate a tuple with a flag for first and last value."""
+    iter_values = iter(values)
+    try:
+        previous_value = next(iter_values)
+    except StopIteration:
+        return
+    first = True
+    for value in iter_values:
+        yield first, False, previous_value
+        first = False
+        previous_value = value
+    yield first, True, previous_value
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_null_file.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_null_file.py
new file mode 100644
index 00000000..6ae05d3e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_null_file.py
@@ -0,0 +1,69 @@
+from types import TracebackType
+from typing import IO, Iterable, Iterator, List, Optional, Type
+
+
+class NullFile(IO[str]):
+    def close(self) -> None:
+        pass
+
+    def isatty(self) -> bool:
+        return False
+
+    def read(self, __n: int = 1) -> str:
+        return ""
+
+    def readable(self) -> bool:
+        return False
+
+    def readline(self, __limit: int = 1) -> str:
+        return ""
+
+    def readlines(self, __hint: int = 1) -> List[str]:
+        return []
+
+    def seek(self, __offset: int, __whence: int = 1) -> int:
+        return 0
+
+    def seekable(self) -> bool:
+        return False
+
+    def tell(self) -> int:
+        return 0
+
+    def truncate(self, __size: Optional[int] = 1) -> int:
+        return 0
+
+    def writable(self) -> bool:
+        return False
+
+    def writelines(self, __lines: Iterable[str]) -> None:
+        pass
+
+    def __next__(self) -> str:
+        return ""
+
+    def __iter__(self) -> Iterator[str]:
+        return iter([""])
+
+    def __enter__(self) -> IO[str]:
+        return self
+
+    def __exit__(
+        self,
+        __t: Optional[Type[BaseException]],
+        __value: Optional[BaseException],
+        __traceback: Optional[TracebackType],
+    ) -> None:
+        pass
+
+    def write(self, text: str) -> int:
+        return 0
+
+    def flush(self) -> None:
+        pass
+
+    def fileno(self) -> int:
+        return -1
+
+
+NULL_FILE = NullFile()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_palettes.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_palettes.py
new file mode 100644
index 00000000..3c748d33
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_palettes.py
@@ -0,0 +1,309 @@
+from .palette import Palette
+
+
+# Taken from https://en.wikipedia.org/wiki/ANSI_escape_code (Windows 10 column)
+WINDOWS_PALETTE = Palette(
+    [
+        (12, 12, 12),
+        (197, 15, 31),
+        (19, 161, 14),
+        (193, 156, 0),
+        (0, 55, 218),
+        (136, 23, 152),
+        (58, 150, 221),
+        (204, 204, 204),
+        (118, 118, 118),
+        (231, 72, 86),
+        (22, 198, 12),
+        (249, 241, 165),
+        (59, 120, 255),
+        (180, 0, 158),
+        (97, 214, 214),
+        (242, 242, 242),
+    ]
+)
+
+# # The standard ansi colors (including bright variants)
+STANDARD_PALETTE = Palette(
+    [
+        (0, 0, 0),
+        (170, 0, 0),
+        (0, 170, 0),
+        (170, 85, 0),
+        (0, 0, 170),
+        (170, 0, 170),
+        (0, 170, 170),
+        (170, 170, 170),
+        (85, 85, 85),
+        (255, 85, 85),
+        (85, 255, 85),
+        (255, 255, 85),
+        (85, 85, 255),
+        (255, 85, 255),
+        (85, 255, 255),
+        (255, 255, 255),
+    ]
+)
+
+
+# The 256 color palette
+EIGHT_BIT_PALETTE = Palette(
+    [
+        (0, 0, 0),
+        (128, 0, 0),
+        (0, 128, 0),
+        (128, 128, 0),
+        (0, 0, 128),
+        (128, 0, 128),
+        (0, 128, 128),
+        (192, 192, 192),
+        (128, 128, 128),
+        (255, 0, 0),
+        (0, 255, 0),
+        (255, 255, 0),
+        (0, 0, 255),
+        (255, 0, 255),
+        (0, 255, 255),
+        (255, 255, 255),
+        (0, 0, 0),
+        (0, 0, 95),
+        (0, 0, 135),
+        (0, 0, 175),
+        (0, 0, 215),
+        (0, 0, 255),
+        (0, 95, 0),
+        (0, 95, 95),
+        (0, 95, 135),
+        (0, 95, 175),
+        (0, 95, 215),
+        (0, 95, 255),
+        (0, 135, 0),
+        (0, 135, 95),
+        (0, 135, 135),
+        (0, 135, 175),
+        (0, 135, 215),
+        (0, 135, 255),
+        (0, 175, 0),
+        (0, 175, 95),
+        (0, 175, 135),
+        (0, 175, 175),
+        (0, 175, 215),
+        (0, 175, 255),
+        (0, 215, 0),
+        (0, 215, 95),
+        (0, 215, 135),
+        (0, 215, 175),
+        (0, 215, 215),
+        (0, 215, 255),
+        (0, 255, 0),
+        (0, 255, 95),
+        (0, 255, 135),
+        (0, 255, 175),
+        (0, 255, 215),
+        (0, 255, 255),
+        (95, 0, 0),
+        (95, 0, 95),
+        (95, 0, 135),
+        (95, 0, 175),
+        (95, 0, 215),
+        (95, 0, 255),
+        (95, 95, 0),
+        (95, 95, 95),
+        (95, 95, 135),
+        (95, 95, 175),
+        (95, 95, 215),
+        (95, 95, 255),
+        (95, 135, 0),
+        (95, 135, 95),
+        (95, 135, 135),
+        (95, 135, 175),
+        (95, 135, 215),
+        (95, 135, 255),
+        (95, 175, 0),
+        (95, 175, 95),
+        (95, 175, 135),
+        (95, 175, 175),
+        (95, 175, 215),
+        (95, 175, 255),
+        (95, 215, 0),
+        (95, 215, 95),
+        (95, 215, 135),
+        (95, 215, 175),
+        (95, 215, 215),
+        (95, 215, 255),
+        (95, 255, 0),
+        (95, 255, 95),
+        (95, 255, 135),
+        (95, 255, 175),
+        (95, 255, 215),
+        (95, 255, 255),
+        (135, 0, 0),
+        (135, 0, 95),
+        (135, 0, 135),
+        (135, 0, 175),
+        (135, 0, 215),
+        (135, 0, 255),
+        (135, 95, 0),
+        (135, 95, 95),
+        (135, 95, 135),
+        (135, 95, 175),
+        (135, 95, 215),
+        (135, 95, 255),
+        (135, 135, 0),
+        (135, 135, 95),
+        (135, 135, 135),
+        (135, 135, 175),
+        (135, 135, 215),
+        (135, 135, 255),
+        (135, 175, 0),
+        (135, 175, 95),
+        (135, 175, 135),
+        (135, 175, 175),
+        (135, 175, 215),
+        (135, 175, 255),
+        (135, 215, 0),
+        (135, 215, 95),
+        (135, 215, 135),
+        (135, 215, 175),
+        (135, 215, 215),
+        (135, 215, 255),
+        (135, 255, 0),
+        (135, 255, 95),
+        (135, 255, 135),
+        (135, 255, 175),
+        (135, 255, 215),
+        (135, 255, 255),
+        (175, 0, 0),
+        (175, 0, 95),
+        (175, 0, 135),
+        (175, 0, 175),
+        (175, 0, 215),
+        (175, 0, 255),
+        (175, 95, 0),
+        (175, 95, 95),
+        (175, 95, 135),
+        (175, 95, 175),
+        (175, 95, 215),
+        (175, 95, 255),
+        (175, 135, 0),
+        (175, 135, 95),
+        (175, 135, 135),
+        (175, 135, 175),
+        (175, 135, 215),
+        (175, 135, 255),
+        (175, 175, 0),
+        (175, 175, 95),
+        (175, 175, 135),
+        (175, 175, 175),
+        (175, 175, 215),
+        (175, 175, 255),
+        (175, 215, 0),
+        (175, 215, 95),
+        (175, 215, 135),
+        (175, 215, 175),
+        (175, 215, 215),
+        (175, 215, 255),
+        (175, 255, 0),
+        (175, 255, 95),
+        (175, 255, 135),
+        (175, 255, 175),
+        (175, 255, 215),
+        (175, 255, 255),
+        (215, 0, 0),
+        (215, 0, 95),
+        (215, 0, 135),
+        (215, 0, 175),
+        (215, 0, 215),
+        (215, 0, 255),
+        (215, 95, 0),
+        (215, 95, 95),
+        (215, 95, 135),
+        (215, 95, 175),
+        (215, 95, 215),
+        (215, 95, 255),
+        (215, 135, 0),
+        (215, 135, 95),
+        (215, 135, 135),
+        (215, 135, 175),
+        (215, 135, 215),
+        (215, 135, 255),
+        (215, 175, 0),
+        (215, 175, 95),
+        (215, 175, 135),
+        (215, 175, 175),
+        (215, 175, 215),
+        (215, 175, 255),
+        (215, 215, 0),
+        (215, 215, 95),
+        (215, 215, 135),
+        (215, 215, 175),
+        (215, 215, 215),
+        (215, 215, 255),
+        (215, 255, 0),
+        (215, 255, 95),
+        (215, 255, 135),
+        (215, 255, 175),
+        (215, 255, 215),
+        (215, 255, 255),
+        (255, 0, 0),
+        (255, 0, 95),
+        (255, 0, 135),
+        (255, 0, 175),
+        (255, 0, 215),
+        (255, 0, 255),
+        (255, 95, 0),
+        (255, 95, 95),
+        (255, 95, 135),
+        (255, 95, 175),
+        (255, 95, 215),
+        (255, 95, 255),
+        (255, 135, 0),
+        (255, 135, 95),
+        (255, 135, 135),
+        (255, 135, 175),
+        (255, 135, 215),
+        (255, 135, 255),
+        (255, 175, 0),
+        (255, 175, 95),
+        (255, 175, 135),
+        (255, 175, 175),
+        (255, 175, 215),
+        (255, 175, 255),
+        (255, 215, 0),
+        (255, 215, 95),
+        (255, 215, 135),
+        (255, 215, 175),
+        (255, 215, 215),
+        (255, 215, 255),
+        (255, 255, 0),
+        (255, 255, 95),
+        (255, 255, 135),
+        (255, 255, 175),
+        (255, 255, 215),
+        (255, 255, 255),
+        (8, 8, 8),
+        (18, 18, 18),
+        (28, 28, 28),
+        (38, 38, 38),
+        (48, 48, 48),
+        (58, 58, 58),
+        (68, 68, 68),
+        (78, 78, 78),
+        (88, 88, 88),
+        (98, 98, 98),
+        (108, 108, 108),
+        (118, 118, 118),
+        (128, 128, 128),
+        (138, 138, 138),
+        (148, 148, 148),
+        (158, 158, 158),
+        (168, 168, 168),
+        (178, 178, 178),
+        (188, 188, 188),
+        (198, 198, 198),
+        (208, 208, 208),
+        (218, 218, 218),
+        (228, 228, 228),
+        (238, 238, 238),
+    ]
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_pick.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_pick.py
new file mode 100644
index 00000000..4f6d8b2d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_pick.py
@@ -0,0 +1,17 @@
+from typing import Optional
+
+
+def pick_bool(*values: Optional[bool]) -> bool:
+    """Pick the first non-none bool or return the last value.
+
+    Args:
+        *values (bool): Any number of boolean or None values.
+
+    Returns:
+        bool: First non-none boolean.
+    """
+    assert values, "1 or more values required"
+    for value in values:
+        if value is not None:
+            return value
+    return bool(value)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_ratio.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_ratio.py
new file mode 100644
index 00000000..95267b0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_ratio.py
@@ -0,0 +1,159 @@
+import sys
+from fractions import Fraction
+from math import ceil
+from typing import cast, List, Optional, Sequence
+
+if sys.version_info >= (3, 8):
+    from typing import Protocol
+else:
+    from pip._vendor.typing_extensions import Protocol  # pragma: no cover
+
+
+class Edge(Protocol):
+    """Any object that defines an edge (such as Layout)."""
+
+    size: Optional[int] = None
+    ratio: int = 1
+    minimum_size: int = 1
+
+
+def ratio_resolve(total: int, edges: Sequence[Edge]) -> List[int]:
+    """Divide total space to satisfy size, ratio, and minimum_size, constraints.
+
+    The returned list of integers should add up to total in most cases, unless it is
+    impossible to satisfy all the constraints. For instance, if there are two edges
+    with a minimum size of 20 each and `total` is 30 then the returned list will be
+    greater than total. In practice, this would mean that a Layout object would
+    clip the rows that would overflow the screen height.
+
+    Args:
+        total (int): Total number of characters.
+        edges (List[Edge]): Edges within total space.
+
+    Returns:
+        List[int]: Number of characters for each edge.
+    """
+    # Size of edge or None for yet to be determined
+    sizes = [(edge.size or None) for edge in edges]
+
+    _Fraction = Fraction
+
+    # While any edges haven't been calculated
+    while None in sizes:
+        # Get flexible edges and index to map these back on to sizes list
+        flexible_edges = [
+            (index, edge)
+            for index, (size, edge) in enumerate(zip(sizes, edges))
+            if size is None
+        ]
+        # Remaining space in total
+        remaining = total - sum(size or 0 for size in sizes)
+        if remaining <= 0:
+            # No room for flexible edges
+            return [
+                ((edge.minimum_size or 1) if size is None else size)
+                for size, edge in zip(sizes, edges)
+            ]
+        # Calculate number of characters in a ratio portion
+        portion = _Fraction(
+            remaining, sum((edge.ratio or 1) for _, edge in flexible_edges)
+        )
+
+        # If any edges will be less than their minimum, replace size with the minimum
+        for index, edge in flexible_edges:
+            if portion * edge.ratio <= edge.minimum_size:
+                sizes[index] = edge.minimum_size
+                # New fixed size will invalidate calculations, so we need to repeat the process
+                break
+        else:
+            # Distribute flexible space and compensate for rounding error
+            # Since edge sizes can only be integers we need to add the remainder
+            # to the following line
+            remainder = _Fraction(0)
+            for index, edge in flexible_edges:
+                size, remainder = divmod(portion * edge.ratio + remainder, 1)
+                sizes[index] = size
+            break
+    # Sizes now contains integers only
+    return cast(List[int], sizes)
+
+
+def ratio_reduce(
+    total: int, ratios: List[int], maximums: List[int], values: List[int]
+) -> List[int]:
+    """Divide an integer total in to parts based on ratios.
+
+    Args:
+        total (int): The total to divide.
+        ratios (List[int]): A list of integer ratios.
+        maximums (List[int]): List of maximums values for each slot.
+        values (List[int]): List of values
+
+    Returns:
+        List[int]: A list of integers guaranteed to sum to total.
+    """
+    ratios = [ratio if _max else 0 for ratio, _max in zip(ratios, maximums)]
+    total_ratio = sum(ratios)
+    if not total_ratio:
+        return values[:]
+    total_remaining = total
+    result: List[int] = []
+    append = result.append
+    for ratio, maximum, value in zip(ratios, maximums, values):
+        if ratio and total_ratio > 0:
+            distributed = min(maximum, round(ratio * total_remaining / total_ratio))
+            append(value - distributed)
+            total_remaining -= distributed
+            total_ratio -= ratio
+        else:
+            append(value)
+    return result
+
+
+def ratio_distribute(
+    total: int, ratios: List[int], minimums: Optional[List[int]] = None
+) -> List[int]:
+    """Distribute an integer total in to parts based on ratios.
+
+    Args:
+        total (int): The total to divide.
+        ratios (List[int]): A list of integer ratios.
+        minimums (List[int]): List of minimum values for each slot.
+
+    Returns:
+        List[int]: A list of integers guaranteed to sum to total.
+    """
+    if minimums:
+        ratios = [ratio if _min else 0 for ratio, _min in zip(ratios, minimums)]
+    total_ratio = sum(ratios)
+    assert total_ratio > 0, "Sum of ratios must be > 0"
+
+    total_remaining = total
+    distributed_total: List[int] = []
+    append = distributed_total.append
+    if minimums is None:
+        _minimums = [0] * len(ratios)
+    else:
+        _minimums = minimums
+    for ratio, minimum in zip(ratios, _minimums):
+        if total_ratio > 0:
+            distributed = max(minimum, ceil(ratio * total_remaining / total_ratio))
+        else:
+            distributed = total_remaining
+        append(distributed)
+        total_ratio -= ratio
+        total_remaining -= distributed
+    return distributed_total
+
+
+if __name__ == "__main__":
+    from dataclasses import dataclass
+
+    @dataclass
+    class E:
+        size: Optional[int] = None
+        ratio: int = 1
+        minimum_size: int = 1
+
+    resolved = ratio_resolve(110, [E(None, 1, 1), E(None, 1, 1), E(None, 1, 1)])
+    print(sum(resolved))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_spinners.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_spinners.py
new file mode 100644
index 00000000..d0bb1fe7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_spinners.py
@@ -0,0 +1,482 @@
+"""
+Spinners are from:
+* cli-spinners:
+    MIT License
+    Copyright (c) Sindre Sorhus <sindresorhus@gmail.com> (sindresorhus.com)
+    Permission is hereby granted, free of charge, to any person obtaining a copy
+    of this software and associated documentation files (the "Software"), to deal
+    in the Software without restriction, including without limitation the rights to
+    use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
+    the Software, and to permit persons to whom the Software is furnished to do so,
+    subject to the following conditions:
+    The above copyright notice and this permission notice shall be included
+    in all copies or substantial portions of the Software.
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+    INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+    PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE
+    FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+    ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+    IN THE SOFTWARE.
+"""
+
+SPINNERS = {
+    "dots": {
+        "interval": 80,
+        "frames": "⠋⠙⠹⠸⠼⠴⠦⠧⠇⠏",
+    },
+    "dots2": {"interval": 80, "frames": "⣾⣽⣻⢿⡿⣟⣯⣷"},
+    "dots3": {
+        "interval": 80,
+        "frames": "⠋⠙⠚⠞⠖⠦⠴⠲⠳⠓",
+    },
+    "dots4": {
+        "interval": 80,
+        "frames": "⠄⠆⠇⠋⠙⠸⠰⠠⠰⠸⠙⠋⠇⠆",
+    },
+    "dots5": {
+        "interval": 80,
+        "frames": "⠋⠙⠚⠒⠂⠂⠒⠲⠴⠦⠖⠒⠐⠐⠒⠓⠋",
+    },
+    "dots6": {
+        "interval": 80,
+        "frames": "⠁⠉⠙⠚⠒⠂⠂⠒⠲⠴⠤⠄⠄⠤⠴⠲⠒⠂⠂⠒⠚⠙⠉⠁",
+    },
+    "dots7": {
+        "interval": 80,
+        "frames": "⠈⠉⠋⠓⠒⠐⠐⠒⠖⠦⠤⠠⠠⠤⠦⠖⠒⠐⠐⠒⠓⠋⠉⠈",
+    },
+    "dots8": {
+        "interval": 80,
+        "frames": "⠁⠁⠉⠙⠚⠒⠂⠂⠒⠲⠴⠤⠄⠄⠤⠠⠠⠤⠦⠖⠒⠐⠐⠒⠓⠋⠉⠈⠈",
+    },
+    "dots9": {"interval": 80, "frames": "⢹⢺⢼⣸⣇⡧⡗⡏"},
+    "dots10": {"interval": 80, "frames": "⢄⢂⢁⡁⡈⡐⡠"},
+    "dots11": {"interval": 100, "frames": "⠁⠂⠄⡀⢀⠠⠐⠈"},
+    "dots12": {
+        "interval": 80,
+        "frames": [
+            "⢀⠀",
+            "⡀⠀",
+            "⠄⠀",
+            "⢂⠀",
+            "⡂⠀",
+            "⠅⠀",
+            "⢃⠀",
+            "⡃⠀",
+            "⠍⠀",
+            "⢋⠀",
+            "⡋⠀",
+            "⠍⠁",
+            "⢋⠁",
+            "⡋⠁",
+            "⠍⠉",
+            "⠋⠉",
+            "⠋⠉",
+            "⠉⠙",
+            "⠉⠙",
+            "⠉⠩",
+            "⠈⢙",
+            "⠈⡙",
+            "⢈⠩",
+            "⡀⢙",
+            "⠄⡙",
+            "⢂⠩",
+            "⡂⢘",
+            "⠅⡘",
+            "⢃⠨",
+            "⡃⢐",
+            "⠍⡐",
+            "⢋⠠",
+            "⡋⢀",
+            "⠍⡁",
+            "⢋⠁",
+            "⡋⠁",
+            "⠍⠉",
+            "⠋⠉",
+            "⠋⠉",
+            "⠉⠙",
+            "⠉⠙",
+            "⠉⠩",
+            "⠈⢙",
+            "⠈⡙",
+            "⠈⠩",
+            "⠀⢙",
+            "⠀⡙",
+            "⠀⠩",
+            "⠀⢘",
+            "⠀⡘",
+            "⠀⠨",
+            "⠀⢐",
+            "⠀⡐",
+            "⠀⠠",
+            "⠀⢀",
+            "⠀⡀",
+        ],
+    },
+    "dots8Bit": {
+        "interval": 80,
+        "frames": "⠀⠁⠂⠃⠄⠅⠆⠇⡀⡁⡂⡃⡄⡅⡆⡇⠈⠉⠊⠋⠌⠍⠎⠏⡈⡉⡊⡋⡌⡍⡎⡏⠐⠑⠒⠓⠔⠕⠖⠗⡐⡑⡒⡓⡔⡕⡖⡗⠘⠙⠚⠛⠜⠝⠞⠟⡘⡙"
+        "⡚⡛⡜⡝⡞⡟⠠⠡⠢⠣⠤⠥⠦⠧⡠⡡⡢⡣⡤⡥⡦⡧⠨⠩⠪⠫⠬⠭⠮⠯⡨⡩⡪⡫⡬⡭⡮⡯⠰⠱⠲⠳⠴⠵⠶⠷⡰⡱⡲⡳⡴⡵⡶⡷⠸⠹⠺⠻"
+        "⠼⠽⠾⠿⡸⡹⡺⡻⡼⡽⡾⡿⢀⢁⢂⢃⢄⢅⢆⢇⣀⣁⣂⣃⣄⣅⣆⣇⢈⢉⢊⢋⢌⢍⢎⢏⣈⣉⣊⣋⣌⣍⣎⣏⢐⢑⢒⢓⢔⢕⢖⢗⣐⣑⣒⣓⣔⣕"
+        "⣖⣗⢘⢙⢚⢛⢜⢝⢞⢟⣘⣙⣚⣛⣜⣝⣞⣟⢠⢡⢢⢣⢤⢥⢦⢧⣠⣡⣢⣣⣤⣥⣦⣧⢨⢩⢪⢫⢬⢭⢮⢯⣨⣩⣪⣫⣬⣭⣮⣯⢰⢱⢲⢳⢴⢵⢶⢷"
+        "⣰⣱⣲⣳⣴⣵⣶⣷⢸⢹⢺⢻⢼⢽⢾⢿⣸⣹⣺⣻⣼⣽⣾⣿",
+    },
+    "line": {"interval": 130, "frames": ["-", "\\", "|", "/"]},
+    "line2": {"interval": 100, "frames": "⠂-–—–-"},
+    "pipe": {"interval": 100, "frames": "┤┘┴└├┌┬┐"},
+    "simpleDots": {"interval": 400, "frames": [".  ", ".. ", "...", "   "]},
+    "simpleDotsScrolling": {
+        "interval": 200,
+        "frames": [".  ", ".. ", "...", " ..", "  .", "   "],
+    },
+    "star": {"interval": 70, "frames": "✶✸✹✺✹✷"},
+    "star2": {"interval": 80, "frames": "+x*"},
+    "flip": {
+        "interval": 70,
+        "frames": "___-``'´-___",
+    },
+    "hamburger": {"interval": 100, "frames": "☱☲☴"},
+    "growVertical": {
+        "interval": 120,
+        "frames": "▁▃▄▅▆▇▆▅▄▃",
+    },
+    "growHorizontal": {
+        "interval": 120,
+        "frames": "▏▎▍▌▋▊▉▊▋▌▍▎",
+    },
+    "balloon": {"interval": 140, "frames": " .oO@* "},
+    "balloon2": {"interval": 120, "frames": ".oO°Oo."},
+    "noise": {"interval": 100, "frames": "▓▒░"},
+    "bounce": {"interval": 120, "frames": "⠁⠂⠄⠂"},
+    "boxBounce": {"interval": 120, "frames": "▖▘▝▗"},
+    "boxBounce2": {"interval": 100, "frames": "▌▀▐▄"},
+    "triangle": {"interval": 50, "frames": "◢◣◤◥"},
+    "arc": {"interval": 100, "frames": "◜◠◝◞◡◟"},
+    "circle": {"interval": 120, "frames": "◡⊙◠"},
+    "squareCorners": {"interval": 180, "frames": "◰◳◲◱"},
+    "circleQuarters": {"interval": 120, "frames": "◴◷◶◵"},
+    "circleHalves": {"interval": 50, "frames": "◐◓◑◒"},
+    "squish": {"interval": 100, "frames": "╫╪"},
+    "toggle": {"interval": 250, "frames": "⊶⊷"},
+    "toggle2": {"interval": 80, "frames": "▫▪"},
+    "toggle3": {"interval": 120, "frames": "□■"},
+    "toggle4": {"interval": 100, "frames": "■□▪▫"},
+    "toggle5": {"interval": 100, "frames": "▮▯"},
+    "toggle6": {"interval": 300, "frames": "ဝ၀"},
+    "toggle7": {"interval": 80, "frames": "⦾⦿"},
+    "toggle8": {"interval": 100, "frames": "◍◌"},
+    "toggle9": {"interval": 100, "frames": "◉◎"},
+    "toggle10": {"interval": 100, "frames": "㊂㊀㊁"},
+    "toggle11": {"interval": 50, "frames": "⧇⧆"},
+    "toggle12": {"interval": 120, "frames": "☗☖"},
+    "toggle13": {"interval": 80, "frames": "=*-"},
+    "arrow": {"interval": 100, "frames": "←↖↑↗→↘↓↙"},
+    "arrow2": {
+        "interval": 80,
+        "frames": ["⬆️ ", "↗️ ", "➡️ ", "↘️ ", "⬇️ ", "↙️ ", "⬅️ ", "↖️ "],
+    },
+    "arrow3": {
+        "interval": 120,
+        "frames": ["▹▹▹▹▹", "▸▹▹▹▹", "▹▸▹▹▹", "▹▹▸▹▹", "▹▹▹▸▹", "▹▹▹▹▸"],
+    },
+    "bouncingBar": {
+        "interval": 80,
+        "frames": [
+            "[    ]",
+            "[=   ]",
+            "[==  ]",
+            "[=== ]",
+            "[ ===]",
+            "[  ==]",
+            "[   =]",
+            "[    ]",
+            "[   =]",
+            "[  ==]",
+            "[ ===]",
+            "[====]",
+            "[=== ]",
+            "[==  ]",
+            "[=   ]",
+        ],
+    },
+    "bouncingBall": {
+        "interval": 80,
+        "frames": [
+            "( ●    )",
+            "(  ●   )",
+            "(   ●  )",
+            "(    ● )",
+            "(     ●)",
+            "(    ● )",
+            "(   ●  )",
+            "(  ●   )",
+            "( ●    )",
+            "(●     )",
+        ],
+    },
+    "smiley": {"interval": 200, "frames": ["😄 ", "😝 "]},
+    "monkey": {"interval": 300, "frames": ["🙈 ", "🙈 ", "🙉 ", "🙊 "]},
+    "hearts": {"interval": 100, "frames": ["💛 ", "💙 ", "💜 ", "💚 ", "❤️ "]},
+    "clock": {
+        "interval": 100,
+        "frames": [
+            "🕛 ",
+            "🕐 ",
+            "🕑 ",
+            "🕒 ",
+            "🕓 ",
+            "🕔 ",
+            "🕕 ",
+            "🕖 ",
+            "🕗 ",
+            "🕘 ",
+            "🕙 ",
+            "🕚 ",
+        ],
+    },
+    "earth": {"interval": 180, "frames": ["🌍 ", "🌎 ", "🌏 "]},
+    "material": {
+        "interval": 17,
+        "frames": [
+            "█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "██▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "███▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "████▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "██████▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "██████▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "███████▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "████████▁▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "██████████▁▁▁▁▁▁▁▁▁▁",
+            "███████████▁▁▁▁▁▁▁▁▁",
+            "█████████████▁▁▁▁▁▁▁",
+            "██████████████▁▁▁▁▁▁",
+            "██████████████▁▁▁▁▁▁",
+            "▁██████████████▁▁▁▁▁",
+            "▁██████████████▁▁▁▁▁",
+            "▁██████████████▁▁▁▁▁",
+            "▁▁██████████████▁▁▁▁",
+            "▁▁▁██████████████▁▁▁",
+            "▁▁▁▁█████████████▁▁▁",
+            "▁▁▁▁██████████████▁▁",
+            "▁▁▁▁██████████████▁▁",
+            "▁▁▁▁▁██████████████▁",
+            "▁▁▁▁▁██████████████▁",
+            "▁▁▁▁▁██████████████▁",
+            "▁▁▁▁▁▁██████████████",
+            "▁▁▁▁▁▁██████████████",
+            "▁▁▁▁▁▁▁█████████████",
+            "▁▁▁▁▁▁▁█████████████",
+            "▁▁▁▁▁▁▁▁████████████",
+            "▁▁▁▁▁▁▁▁████████████",
+            "▁▁▁▁▁▁▁▁▁███████████",
+            "▁▁▁▁▁▁▁▁▁███████████",
+            "▁▁▁▁▁▁▁▁▁▁██████████",
+            "▁▁▁▁▁▁▁▁▁▁██████████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁████████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁███████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁██████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█████",
+            "█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁████",
+            "██▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁███",
+            "██▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁███",
+            "███▁▁▁▁▁▁▁▁▁▁▁▁▁▁███",
+            "████▁▁▁▁▁▁▁▁▁▁▁▁▁▁██",
+            "█████▁▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "█████▁▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "██████▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "████████▁▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "█████████▁▁▁▁▁▁▁▁▁▁▁",
+            "███████████▁▁▁▁▁▁▁▁▁",
+            "████████████▁▁▁▁▁▁▁▁",
+            "████████████▁▁▁▁▁▁▁▁",
+            "██████████████▁▁▁▁▁▁",
+            "██████████████▁▁▁▁▁▁",
+            "▁██████████████▁▁▁▁▁",
+            "▁██████████████▁▁▁▁▁",
+            "▁▁▁█████████████▁▁▁▁",
+            "▁▁▁▁▁████████████▁▁▁",
+            "▁▁▁▁▁████████████▁▁▁",
+            "▁▁▁▁▁▁███████████▁▁▁",
+            "▁▁▁▁▁▁▁▁█████████▁▁▁",
+            "▁▁▁▁▁▁▁▁█████████▁▁▁",
+            "▁▁▁▁▁▁▁▁▁█████████▁▁",
+            "▁▁▁▁▁▁▁▁▁█████████▁▁",
+            "▁▁▁▁▁▁▁▁▁▁█████████▁",
+            "▁▁▁▁▁▁▁▁▁▁▁████████▁",
+            "▁▁▁▁▁▁▁▁▁▁▁████████▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁███████▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁███████▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁███████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁███████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁████",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁███",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁███",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁██",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁██",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁██",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+            "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+        ],
+    },
+    "moon": {
+        "interval": 80,
+        "frames": ["🌑 ", "🌒 ", "🌓 ", "🌔 ", "🌕 ", "🌖 ", "🌗 ", "🌘 "],
+    },
+    "runner": {"interval": 140, "frames": ["🚶 ", "🏃 "]},
+    "pong": {
+        "interval": 80,
+        "frames": [
+            "▐⠂       ▌",
+            "▐⠈       ▌",
+            "▐ ⠂      ▌",
+            "▐ ⠠      ▌",
+            "▐  ⡀     ▌",
+            "▐  ⠠     ▌",
+            "▐   ⠂    ▌",
+            "▐   ⠈    ▌",
+            "▐    ⠂   ▌",
+            "▐    ⠠   ▌",
+            "▐     ⡀  ▌",
+            "▐     ⠠  ▌",
+            "▐      ⠂ ▌",
+            "▐      ⠈ ▌",
+            "▐       ⠂▌",
+            "▐       ⠠▌",
+            "▐       ⡀▌",
+            "▐      ⠠ ▌",
+            "▐      ⠂ ▌",
+            "▐     ⠈  ▌",
+            "▐     ⠂  ▌",
+            "▐    ⠠   ▌",
+            "▐    ⡀   ▌",
+            "▐   ⠠    ▌",
+            "▐   ⠂    ▌",
+            "▐  ⠈     ▌",
+            "▐  ⠂     ▌",
+            "▐ ⠠      ▌",
+            "▐ ⡀      ▌",
+            "▐⠠       ▌",
+        ],
+    },
+    "shark": {
+        "interval": 120,
+        "frames": [
+            "▐|\\____________▌",
+            "▐_|\\___________▌",
+            "▐__|\\__________▌",
+            "▐___|\\_________▌",
+            "▐____|\\________▌",
+            "▐_____|\\_______▌",
+            "▐______|\\______▌",
+            "▐_______|\\_____▌",
+            "▐________|\\____▌",
+            "▐_________|\\___▌",
+            "▐__________|\\__▌",
+            "▐___________|\\_▌",
+            "▐____________|\\▌",
+            "▐____________/|▌",
+            "▐___________/|_▌",
+            "▐__________/|__▌",
+            "▐_________/|___▌",
+            "▐________/|____▌",
+            "▐_______/|_____▌",
+            "▐______/|______▌",
+            "▐_____/|_______▌",
+            "▐____/|________▌",
+            "▐___/|_________▌",
+            "▐__/|__________▌",
+            "▐_/|___________▌",
+            "▐/|____________▌",
+        ],
+    },
+    "dqpb": {"interval": 100, "frames": "dqpb"},
+    "weather": {
+        "interval": 100,
+        "frames": [
+            "☀️ ",
+            "☀️ ",
+            "☀️ ",
+            "🌤 ",
+            "⛅️ ",
+            "🌥 ",
+            "☁️ ",
+            "🌧 ",
+            "🌨 ",
+            "🌧 ",
+            "🌨 ",
+            "🌧 ",
+            "🌨 ",
+            "⛈ ",
+            "🌨 ",
+            "🌧 ",
+            "🌨 ",
+            "☁️ ",
+            "🌥 ",
+            "⛅️ ",
+            "🌤 ",
+            "☀️ ",
+            "☀️ ",
+        ],
+    },
+    "christmas": {"interval": 400, "frames": "🌲🎄"},
+    "grenade": {
+        "interval": 80,
+        "frames": [
+            "،   ",
+            "′   ",
+            " ´ ",
+            " ‾ ",
+            "  ⸌",
+            "  ⸊",
+            "  |",
+            "  ⁎",
+            "  ⁕",
+            " ෴ ",
+            "  ⁓",
+            "   ",
+            "   ",
+            "   ",
+        ],
+    },
+    "point": {"interval": 125, "frames": ["∙∙∙", "●∙∙", "∙●∙", "∙∙●", "∙∙∙"]},
+    "layer": {"interval": 150, "frames": "-=≡"},
+    "betaWave": {
+        "interval": 80,
+        "frames": [
+            "ρββββββ",
+            "βρβββββ",
+            "ββρββββ",
+            "βββρβββ",
+            "ββββρββ",
+            "βββββρβ",
+            "ββββββρ",
+        ],
+    },
+    "aesthetic": {
+        "interval": 80,
+        "frames": [
+            "▰▱▱▱▱▱▱",
+            "▰▰▱▱▱▱▱",
+            "▰▰▰▱▱▱▱",
+            "▰▰▰▰▱▱▱",
+            "▰▰▰▰▰▱▱",
+            "▰▰▰▰▰▰▱",
+            "▰▰▰▰▰▰▰",
+            "▰▱▱▱▱▱▱",
+        ],
+    },
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_stack.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_stack.py
new file mode 100644
index 00000000..194564e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_stack.py
@@ -0,0 +1,16 @@
+from typing import List, TypeVar
+
+T = TypeVar("T")
+
+
+class Stack(List[T]):
+    """A small shim over builtin list."""
+
+    @property
+    def top(self) -> T:
+        """Get top of stack."""
+        return self[-1]
+
+    def push(self, item: T) -> None:
+        """Push an item on to the stack (append in stack nomenclature)."""
+        self.append(item)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_timer.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_timer.py
new file mode 100644
index 00000000..a2ca6be0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_timer.py
@@ -0,0 +1,19 @@
+"""
+Timer context manager, only used in debug.
+
+"""
+
+from time import time
+
+import contextlib
+from typing import Generator
+
+
+@contextlib.contextmanager
+def timer(subject: str = "time") -> Generator[None, None, None]:
+    """print the elapsed time. (only used in debugging)"""
+    start = time()
+    yield
+    elapsed = time() - start
+    elapsed_ms = elapsed * 1000
+    print(f"{subject} elapsed {elapsed_ms:.1f}ms")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_win32_console.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_win32_console.py
new file mode 100644
index 00000000..2eba1b9b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_win32_console.py
@@ -0,0 +1,661 @@
+"""Light wrapper around the Win32 Console API - this module should only be imported on Windows
+
+The API that this module wraps is documented at https://docs.microsoft.com/en-us/windows/console/console-functions
+"""
+
+import ctypes
+import sys
+from typing import Any
+
+windll: Any = None
+if sys.platform == "win32":
+    windll = ctypes.LibraryLoader(ctypes.WinDLL)
+else:
+    raise ImportError(f"{__name__} can only be imported on Windows")
+
+import time
+from ctypes import Structure, byref, wintypes
+from typing import IO, NamedTuple, Type, cast
+
+from pip._vendor.rich.color import ColorSystem
+from pip._vendor.rich.style import Style
+
+STDOUT = -11
+ENABLE_VIRTUAL_TERMINAL_PROCESSING = 4
+
+COORD = wintypes._COORD
+
+
+class LegacyWindowsError(Exception):
+    pass
+
+
+class WindowsCoordinates(NamedTuple):
+    """Coordinates in the Windows Console API are (y, x), not (x, y).
+    This class is intended to prevent that confusion.
+    Rows and columns are indexed from 0.
+    This class can be used in place of wintypes._COORD in arguments and argtypes.
+    """
+
+    row: int
+    col: int
+
+    @classmethod
+    def from_param(cls, value: "WindowsCoordinates") -> COORD:
+        """Converts a WindowsCoordinates into a wintypes _COORD structure.
+        This classmethod is internally called by ctypes to perform the conversion.
+
+        Args:
+            value (WindowsCoordinates): The input coordinates to convert.
+
+        Returns:
+            wintypes._COORD: The converted coordinates struct.
+        """
+        return COORD(value.col, value.row)
+
+
+class CONSOLE_SCREEN_BUFFER_INFO(Structure):
+    _fields_ = [
+        ("dwSize", COORD),
+        ("dwCursorPosition", COORD),
+        ("wAttributes", wintypes.WORD),
+        ("srWindow", wintypes.SMALL_RECT),
+        ("dwMaximumWindowSize", COORD),
+    ]
+
+
+class CONSOLE_CURSOR_INFO(ctypes.Structure):
+    _fields_ = [("dwSize", wintypes.DWORD), ("bVisible", wintypes.BOOL)]
+
+
+_GetStdHandle = windll.kernel32.GetStdHandle
+_GetStdHandle.argtypes = [
+    wintypes.DWORD,
+]
+_GetStdHandle.restype = wintypes.HANDLE
+
+
+def GetStdHandle(handle: int = STDOUT) -> wintypes.HANDLE:
+    """Retrieves a handle to the specified standard device (standard input, standard output, or standard error).
+
+    Args:
+        handle (int): Integer identifier for the handle. Defaults to -11 (stdout).
+
+    Returns:
+        wintypes.HANDLE: The handle
+    """
+    return cast(wintypes.HANDLE, _GetStdHandle(handle))
+
+
+_GetConsoleMode = windll.kernel32.GetConsoleMode
+_GetConsoleMode.argtypes = [wintypes.HANDLE, wintypes.LPDWORD]
+_GetConsoleMode.restype = wintypes.BOOL
+
+
+def GetConsoleMode(std_handle: wintypes.HANDLE) -> int:
+    """Retrieves the current input mode of a console's input buffer
+    or the current output mode of a console screen buffer.
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+
+    Raises:
+        LegacyWindowsError: If any error occurs while calling the Windows console API.
+
+    Returns:
+        int: Value representing the current console mode as documented at
+            https://docs.microsoft.com/en-us/windows/console/getconsolemode#parameters
+    """
+
+    console_mode = wintypes.DWORD()
+    success = bool(_GetConsoleMode(std_handle, console_mode))
+    if not success:
+        raise LegacyWindowsError("Unable to get legacy Windows Console Mode")
+    return console_mode.value
+
+
+_FillConsoleOutputCharacterW = windll.kernel32.FillConsoleOutputCharacterW
+_FillConsoleOutputCharacterW.argtypes = [
+    wintypes.HANDLE,
+    ctypes.c_char,
+    wintypes.DWORD,
+    cast(Type[COORD], WindowsCoordinates),
+    ctypes.POINTER(wintypes.DWORD),
+]
+_FillConsoleOutputCharacterW.restype = wintypes.BOOL
+
+
+def FillConsoleOutputCharacter(
+    std_handle: wintypes.HANDLE,
+    char: str,
+    length: int,
+    start: WindowsCoordinates,
+) -> int:
+    """Writes a character to the console screen buffer a specified number of times, beginning at the specified coordinates.
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        char (str): The character to write. Must be a string of length 1.
+        length (int): The number of times to write the character.
+        start (WindowsCoordinates): The coordinates to start writing at.
+
+    Returns:
+        int: The number of characters written.
+    """
+    character = ctypes.c_char(char.encode())
+    num_characters = wintypes.DWORD(length)
+    num_written = wintypes.DWORD(0)
+    _FillConsoleOutputCharacterW(
+        std_handle,
+        character,
+        num_characters,
+        start,
+        byref(num_written),
+    )
+    return num_written.value
+
+
+_FillConsoleOutputAttribute = windll.kernel32.FillConsoleOutputAttribute
+_FillConsoleOutputAttribute.argtypes = [
+    wintypes.HANDLE,
+    wintypes.WORD,
+    wintypes.DWORD,
+    cast(Type[COORD], WindowsCoordinates),
+    ctypes.POINTER(wintypes.DWORD),
+]
+_FillConsoleOutputAttribute.restype = wintypes.BOOL
+
+
+def FillConsoleOutputAttribute(
+    std_handle: wintypes.HANDLE,
+    attributes: int,
+    length: int,
+    start: WindowsCoordinates,
+) -> int:
+    """Sets the character attributes for a specified number of character cells,
+    beginning at the specified coordinates in a screen buffer.
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        attributes (int): Integer value representing the foreground and background colours of the cells.
+        length (int): The number of cells to set the output attribute of.
+        start (WindowsCoordinates): The coordinates of the first cell whose attributes are to be set.
+
+    Returns:
+        int: The number of cells whose attributes were actually set.
+    """
+    num_cells = wintypes.DWORD(length)
+    style_attrs = wintypes.WORD(attributes)
+    num_written = wintypes.DWORD(0)
+    _FillConsoleOutputAttribute(
+        std_handle, style_attrs, num_cells, start, byref(num_written)
+    )
+    return num_written.value
+
+
+_SetConsoleTextAttribute = windll.kernel32.SetConsoleTextAttribute
+_SetConsoleTextAttribute.argtypes = [
+    wintypes.HANDLE,
+    wintypes.WORD,
+]
+_SetConsoleTextAttribute.restype = wintypes.BOOL
+
+
+def SetConsoleTextAttribute(
+    std_handle: wintypes.HANDLE, attributes: wintypes.WORD
+) -> bool:
+    """Set the colour attributes for all text written after this function is called.
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        attributes (int): Integer value representing the foreground and background colours.
+
+
+    Returns:
+        bool: True if the attribute was set successfully, otherwise False.
+    """
+    return bool(_SetConsoleTextAttribute(std_handle, attributes))
+
+
+_GetConsoleScreenBufferInfo = windll.kernel32.GetConsoleScreenBufferInfo
+_GetConsoleScreenBufferInfo.argtypes = [
+    wintypes.HANDLE,
+    ctypes.POINTER(CONSOLE_SCREEN_BUFFER_INFO),
+]
+_GetConsoleScreenBufferInfo.restype = wintypes.BOOL
+
+
+def GetConsoleScreenBufferInfo(
+    std_handle: wintypes.HANDLE,
+) -> CONSOLE_SCREEN_BUFFER_INFO:
+    """Retrieves information about the specified console screen buffer.
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+
+    Returns:
+        CONSOLE_SCREEN_BUFFER_INFO: A CONSOLE_SCREEN_BUFFER_INFO ctype struct contain information about
+            screen size, cursor position, colour attributes, and more."""
+    console_screen_buffer_info = CONSOLE_SCREEN_BUFFER_INFO()
+    _GetConsoleScreenBufferInfo(std_handle, byref(console_screen_buffer_info))
+    return console_screen_buffer_info
+
+
+_SetConsoleCursorPosition = windll.kernel32.SetConsoleCursorPosition
+_SetConsoleCursorPosition.argtypes = [
+    wintypes.HANDLE,
+    cast(Type[COORD], WindowsCoordinates),
+]
+_SetConsoleCursorPosition.restype = wintypes.BOOL
+
+
+def SetConsoleCursorPosition(
+    std_handle: wintypes.HANDLE, coords: WindowsCoordinates
+) -> bool:
+    """Set the position of the cursor in the console screen
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        coords (WindowsCoordinates): The coordinates to move the cursor to.
+
+    Returns:
+        bool: True if the function succeeds, otherwise False.
+    """
+    return bool(_SetConsoleCursorPosition(std_handle, coords))
+
+
+_GetConsoleCursorInfo = windll.kernel32.GetConsoleCursorInfo
+_GetConsoleCursorInfo.argtypes = [
+    wintypes.HANDLE,
+    ctypes.POINTER(CONSOLE_CURSOR_INFO),
+]
+_GetConsoleCursorInfo.restype = wintypes.BOOL
+
+
+def GetConsoleCursorInfo(
+    std_handle: wintypes.HANDLE, cursor_info: CONSOLE_CURSOR_INFO
+) -> bool:
+    """Get the cursor info - used to get cursor visibility and width
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        cursor_info (CONSOLE_CURSOR_INFO): CONSOLE_CURSOR_INFO ctype struct that receives information
+            about the console's cursor.
+
+    Returns:
+          bool: True if the function succeeds, otherwise False.
+    """
+    return bool(_GetConsoleCursorInfo(std_handle, byref(cursor_info)))
+
+
+_SetConsoleCursorInfo = windll.kernel32.SetConsoleCursorInfo
+_SetConsoleCursorInfo.argtypes = [
+    wintypes.HANDLE,
+    ctypes.POINTER(CONSOLE_CURSOR_INFO),
+]
+_SetConsoleCursorInfo.restype = wintypes.BOOL
+
+
+def SetConsoleCursorInfo(
+    std_handle: wintypes.HANDLE, cursor_info: CONSOLE_CURSOR_INFO
+) -> bool:
+    """Set the cursor info - used for adjusting cursor visibility and width
+
+    Args:
+        std_handle (wintypes.HANDLE): A handle to the console input buffer or the console screen buffer.
+        cursor_info (CONSOLE_CURSOR_INFO): CONSOLE_CURSOR_INFO ctype struct containing the new cursor info.
+
+    Returns:
+          bool: True if the function succeeds, otherwise False.
+    """
+    return bool(_SetConsoleCursorInfo(std_handle, byref(cursor_info)))
+
+
+_SetConsoleTitle = windll.kernel32.SetConsoleTitleW
+_SetConsoleTitle.argtypes = [wintypes.LPCWSTR]
+_SetConsoleTitle.restype = wintypes.BOOL
+
+
+def SetConsoleTitle(title: str) -> bool:
+    """Sets the title of the current console window
+
+    Args:
+        title (str): The new title of the console window.
+
+    Returns:
+        bool: True if the function succeeds, otherwise False.
+    """
+    return bool(_SetConsoleTitle(title))
+
+
+class LegacyWindowsTerm:
+    """This class allows interaction with the legacy Windows Console API. It should only be used in the context
+    of environments where virtual terminal processing is not available. However, if it is used in a Windows environment,
+    the entire API should work.
+
+    Args:
+        file (IO[str]): The file which the Windows Console API HANDLE is retrieved from, defaults to sys.stdout.
+    """
+
+    BRIGHT_BIT = 8
+
+    # Indices are ANSI color numbers, values are the corresponding Windows Console API color numbers
+    ANSI_TO_WINDOWS = [
+        0,  # black                      The Windows colours are defined in wincon.h as follows:
+        4,  # red                         define FOREGROUND_BLUE            0x0001 -- 0000 0001
+        2,  # green                       define FOREGROUND_GREEN           0x0002 -- 0000 0010
+        6,  # yellow                      define FOREGROUND_RED             0x0004 -- 0000 0100
+        1,  # blue                        define FOREGROUND_INTENSITY       0x0008 -- 0000 1000
+        5,  # magenta                     define BACKGROUND_BLUE            0x0010 -- 0001 0000
+        3,  # cyan                        define BACKGROUND_GREEN           0x0020 -- 0010 0000
+        7,  # white                       define BACKGROUND_RED             0x0040 -- 0100 0000
+        8,  # bright black (grey)         define BACKGROUND_INTENSITY       0x0080 -- 1000 0000
+        12,  # bright red
+        10,  # bright green
+        14,  # bright yellow
+        9,  # bright blue
+        13,  # bright magenta
+        11,  # bright cyan
+        15,  # bright white
+    ]
+
+    def __init__(self, file: "IO[str]") -> None:
+        handle = GetStdHandle(STDOUT)
+        self._handle = handle
+        default_text = GetConsoleScreenBufferInfo(handle).wAttributes
+        self._default_text = default_text
+
+        self._default_fore = default_text & 7
+        self._default_back = (default_text >> 4) & 7
+        self._default_attrs = self._default_fore | (self._default_back << 4)
+
+        self._file = file
+        self.write = file.write
+        self.flush = file.flush
+
+    @property
+    def cursor_position(self) -> WindowsCoordinates:
+        """Returns the current position of the cursor (0-based)
+
+        Returns:
+            WindowsCoordinates: The current cursor position.
+        """
+        coord: COORD = GetConsoleScreenBufferInfo(self._handle).dwCursorPosition
+        return WindowsCoordinates(row=coord.Y, col=coord.X)
+
+    @property
+    def screen_size(self) -> WindowsCoordinates:
+        """Returns the current size of the console screen buffer, in character columns and rows
+
+        Returns:
+            WindowsCoordinates: The width and height of the screen as WindowsCoordinates.
+        """
+        screen_size: COORD = GetConsoleScreenBufferInfo(self._handle).dwSize
+        return WindowsCoordinates(row=screen_size.Y, col=screen_size.X)
+
+    def write_text(self, text: str) -> None:
+        """Write text directly to the terminal without any modification of styles
+
+        Args:
+            text (str): The text to write to the console
+        """
+        self.write(text)
+        self.flush()
+
+    def write_styled(self, text: str, style: Style) -> None:
+        """Write styled text to the terminal.
+
+        Args:
+            text (str): The text to write
+            style (Style): The style of the text
+        """
+        color = style.color
+        bgcolor = style.bgcolor
+        if style.reverse:
+            color, bgcolor = bgcolor, color
+
+        if color:
+            fore = color.downgrade(ColorSystem.WINDOWS).number
+            fore = fore if fore is not None else 7  # Default to ANSI 7: White
+            if style.bold:
+                fore = fore | self.BRIGHT_BIT
+            if style.dim:
+                fore = fore & ~self.BRIGHT_BIT
+            fore = self.ANSI_TO_WINDOWS[fore]
+        else:
+            fore = self._default_fore
+
+        if bgcolor:
+            back = bgcolor.downgrade(ColorSystem.WINDOWS).number
+            back = back if back is not None else 0  # Default to ANSI 0: Black
+            back = self.ANSI_TO_WINDOWS[back]
+        else:
+            back = self._default_back
+
+        assert fore is not None
+        assert back is not None
+
+        SetConsoleTextAttribute(
+            self._handle, attributes=ctypes.c_ushort(fore | (back << 4))
+        )
+        self.write_text(text)
+        SetConsoleTextAttribute(self._handle, attributes=self._default_text)
+
+    def move_cursor_to(self, new_position: WindowsCoordinates) -> None:
+        """Set the position of the cursor
+
+        Args:
+            new_position (WindowsCoordinates): The WindowsCoordinates representing the new position of the cursor.
+        """
+        if new_position.col < 0 or new_position.row < 0:
+            return
+        SetConsoleCursorPosition(self._handle, coords=new_position)
+
+    def erase_line(self) -> None:
+        """Erase all content on the line the cursor is currently located at"""
+        screen_size = self.screen_size
+        cursor_position = self.cursor_position
+        cells_to_erase = screen_size.col
+        start_coordinates = WindowsCoordinates(row=cursor_position.row, col=0)
+        FillConsoleOutputCharacter(
+            self._handle, " ", length=cells_to_erase, start=start_coordinates
+        )
+        FillConsoleOutputAttribute(
+            self._handle,
+            self._default_attrs,
+            length=cells_to_erase,
+            start=start_coordinates,
+        )
+
+    def erase_end_of_line(self) -> None:
+        """Erase all content from the cursor position to the end of that line"""
+        cursor_position = self.cursor_position
+        cells_to_erase = self.screen_size.col - cursor_position.col
+        FillConsoleOutputCharacter(
+            self._handle, " ", length=cells_to_erase, start=cursor_position
+        )
+        FillConsoleOutputAttribute(
+            self._handle,
+            self._default_attrs,
+            length=cells_to_erase,
+            start=cursor_position,
+        )
+
+    def erase_start_of_line(self) -> None:
+        """Erase all content from the cursor position to the start of that line"""
+        row, col = self.cursor_position
+        start = WindowsCoordinates(row, 0)
+        FillConsoleOutputCharacter(self._handle, " ", length=col, start=start)
+        FillConsoleOutputAttribute(
+            self._handle, self._default_attrs, length=col, start=start
+        )
+
+    def move_cursor_up(self) -> None:
+        """Move the cursor up a single cell"""
+        cursor_position = self.cursor_position
+        SetConsoleCursorPosition(
+            self._handle,
+            coords=WindowsCoordinates(
+                row=cursor_position.row - 1, col=cursor_position.col
+            ),
+        )
+
+    def move_cursor_down(self) -> None:
+        """Move the cursor down a single cell"""
+        cursor_position = self.cursor_position
+        SetConsoleCursorPosition(
+            self._handle,
+            coords=WindowsCoordinates(
+                row=cursor_position.row + 1,
+                col=cursor_position.col,
+            ),
+        )
+
+    def move_cursor_forward(self) -> None:
+        """Move the cursor forward a single cell. Wrap to the next line if required."""
+        row, col = self.cursor_position
+        if col == self.screen_size.col - 1:
+            row += 1
+            col = 0
+        else:
+            col += 1
+        SetConsoleCursorPosition(
+            self._handle, coords=WindowsCoordinates(row=row, col=col)
+        )
+
+    def move_cursor_to_column(self, column: int) -> None:
+        """Move cursor to the column specified by the zero-based column index, staying on the same row
+
+        Args:
+            column (int): The zero-based column index to move the cursor to.
+        """
+        row, _ = self.cursor_position
+        SetConsoleCursorPosition(self._handle, coords=WindowsCoordinates(row, column))
+
+    def move_cursor_backward(self) -> None:
+        """Move the cursor backward a single cell. Wrap to the previous line if required."""
+        row, col = self.cursor_position
+        if col == 0:
+            row -= 1
+            col = self.screen_size.col - 1
+        else:
+            col -= 1
+        SetConsoleCursorPosition(
+            self._handle, coords=WindowsCoordinates(row=row, col=col)
+        )
+
+    def hide_cursor(self) -> None:
+        """Hide the cursor"""
+        current_cursor_size = self._get_cursor_size()
+        invisible_cursor = CONSOLE_CURSOR_INFO(dwSize=current_cursor_size, bVisible=0)
+        SetConsoleCursorInfo(self._handle, cursor_info=invisible_cursor)
+
+    def show_cursor(self) -> None:
+        """Show the cursor"""
+        current_cursor_size = self._get_cursor_size()
+        visible_cursor = CONSOLE_CURSOR_INFO(dwSize=current_cursor_size, bVisible=1)
+        SetConsoleCursorInfo(self._handle, cursor_info=visible_cursor)
+
+    def set_title(self, title: str) -> None:
+        """Set the title of the terminal window
+
+        Args:
+            title (str): The new title of the console window
+        """
+        assert len(title) < 255, "Console title must be less than 255 characters"
+        SetConsoleTitle(title)
+
+    def _get_cursor_size(self) -> int:
+        """Get the percentage of the character cell that is filled by the cursor"""
+        cursor_info = CONSOLE_CURSOR_INFO()
+        GetConsoleCursorInfo(self._handle, cursor_info=cursor_info)
+        return int(cursor_info.dwSize)
+
+
+if __name__ == "__main__":
+    handle = GetStdHandle()
+
+    from pip._vendor.rich.console import Console
+
+    console = Console()
+
+    term = LegacyWindowsTerm(sys.stdout)
+    term.set_title("Win32 Console Examples")
+
+    style = Style(color="black", bgcolor="red")
+
+    heading = Style.parse("black on green")
+
+    # Check colour output
+    console.rule("Checking colour output")
+    console.print("[on red]on red!")
+    console.print("[blue]blue!")
+    console.print("[yellow]yellow!")
+    console.print("[bold yellow]bold yellow!")
+    console.print("[bright_yellow]bright_yellow!")
+    console.print("[dim bright_yellow]dim bright_yellow!")
+    console.print("[italic cyan]italic cyan!")
+    console.print("[bold white on blue]bold white on blue!")
+    console.print("[reverse bold white on blue]reverse bold white on blue!")
+    console.print("[bold black on cyan]bold black on cyan!")
+    console.print("[black on green]black on green!")
+    console.print("[blue on green]blue on green!")
+    console.print("[white on black]white on black!")
+    console.print("[black on white]black on white!")
+    console.print("[#1BB152 on #DA812D]#1BB152 on #DA812D!")
+
+    # Check cursor movement
+    console.rule("Checking cursor movement")
+    console.print()
+    term.move_cursor_backward()
+    term.move_cursor_backward()
+    term.write_text("went back and wrapped to prev line")
+    time.sleep(1)
+    term.move_cursor_up()
+    term.write_text("we go up")
+    time.sleep(1)
+    term.move_cursor_down()
+    term.write_text("and down")
+    time.sleep(1)
+    term.move_cursor_up()
+    term.move_cursor_backward()
+    term.move_cursor_backward()
+    term.write_text("we went up and back 2")
+    time.sleep(1)
+    term.move_cursor_down()
+    term.move_cursor_backward()
+    term.move_cursor_backward()
+    term.write_text("we went down and back 2")
+    time.sleep(1)
+
+    # Check erasing of lines
+    term.hide_cursor()
+    console.print()
+    console.rule("Checking line erasing")
+    console.print("\n...Deleting to the start of the line...")
+    term.write_text("The red arrow shows the cursor location, and direction of erase")
+    time.sleep(1)
+    term.move_cursor_to_column(16)
+    term.write_styled("<", Style.parse("black on red"))
+    term.move_cursor_backward()
+    time.sleep(1)
+    term.erase_start_of_line()
+    time.sleep(1)
+
+    console.print("\n\n...And to the end of the line...")
+    term.write_text("The red arrow shows the cursor location, and direction of erase")
+    time.sleep(1)
+
+    term.move_cursor_to_column(16)
+    term.write_styled(">", Style.parse("black on red"))
+    time.sleep(1)
+    term.erase_end_of_line()
+    time.sleep(1)
+
+    console.print("\n\n...Now the whole line will be erased...")
+    term.write_styled("I'm going to disappear!", style=Style.parse("black on cyan"))
+    time.sleep(1)
+    term.erase_line()
+
+    term.show_cursor()
+    print("\n")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows.py
new file mode 100644
index 00000000..7520a9f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows.py
@@ -0,0 +1,71 @@
+import sys
+from dataclasses import dataclass
+
+
+@dataclass
+class WindowsConsoleFeatures:
+    """Windows features available."""
+
+    vt: bool = False
+    """The console supports VT codes."""
+    truecolor: bool = False
+    """The console supports truecolor."""
+
+
+try:
+    import ctypes
+    from ctypes import LibraryLoader
+
+    if sys.platform == "win32":
+        windll = LibraryLoader(ctypes.WinDLL)
+    else:
+        windll = None
+        raise ImportError("Not windows")
+
+    from pip._vendor.rich._win32_console import (
+        ENABLE_VIRTUAL_TERMINAL_PROCESSING,
+        GetConsoleMode,
+        GetStdHandle,
+        LegacyWindowsError,
+    )
+
+except (AttributeError, ImportError, ValueError):
+    # Fallback if we can't load the Windows DLL
+    def get_windows_console_features() -> WindowsConsoleFeatures:
+        features = WindowsConsoleFeatures()
+        return features
+
+else:
+
+    def get_windows_console_features() -> WindowsConsoleFeatures:
+        """Get windows console features.
+
+        Returns:
+            WindowsConsoleFeatures: An instance of WindowsConsoleFeatures.
+        """
+        handle = GetStdHandle()
+        try:
+            console_mode = GetConsoleMode(handle)
+            success = True
+        except LegacyWindowsError:
+            console_mode = 0
+            success = False
+        vt = bool(success and console_mode & ENABLE_VIRTUAL_TERMINAL_PROCESSING)
+        truecolor = False
+        if vt:
+            win_version = sys.getwindowsversion()
+            truecolor = win_version.major > 10 or (
+                win_version.major == 10 and win_version.build >= 15063
+            )
+        features = WindowsConsoleFeatures(vt=vt, truecolor=truecolor)
+        return features
+
+
+if __name__ == "__main__":
+    import platform
+
+    features = get_windows_console_features()
+    from pip._vendor.rich import print
+
+    print(f'platform="{platform.system()}"')
+    print(repr(features))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows_renderer.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows_renderer.py
new file mode 100644
index 00000000..5ece0564
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_windows_renderer.py
@@ -0,0 +1,56 @@
+from typing import Iterable, Sequence, Tuple, cast
+
+from pip._vendor.rich._win32_console import LegacyWindowsTerm, WindowsCoordinates
+from pip._vendor.rich.segment import ControlCode, ControlType, Segment
+
+
+def legacy_windows_render(buffer: Iterable[Segment], term: LegacyWindowsTerm) -> None:
+    """Makes appropriate Windows Console API calls based on the segments in the buffer.
+
+    Args:
+        buffer (Iterable[Segment]): Iterable of Segments to convert to Win32 API calls.
+        term (LegacyWindowsTerm): Used to call the Windows Console API.
+    """
+    for text, style, control in buffer:
+        if not control:
+            if style:
+                term.write_styled(text, style)
+            else:
+                term.write_text(text)
+        else:
+            control_codes: Sequence[ControlCode] = control
+            for control_code in control_codes:
+                control_type = control_code[0]
+                if control_type == ControlType.CURSOR_MOVE_TO:
+                    _, x, y = cast(Tuple[ControlType, int, int], control_code)
+                    term.move_cursor_to(WindowsCoordinates(row=y - 1, col=x - 1))
+                elif control_type == ControlType.CARRIAGE_RETURN:
+                    term.write_text("\r")
+                elif control_type == ControlType.HOME:
+                    term.move_cursor_to(WindowsCoordinates(0, 0))
+                elif control_type == ControlType.CURSOR_UP:
+                    term.move_cursor_up()
+                elif control_type == ControlType.CURSOR_DOWN:
+                    term.move_cursor_down()
+                elif control_type == ControlType.CURSOR_FORWARD:
+                    term.move_cursor_forward()
+                elif control_type == ControlType.CURSOR_BACKWARD:
+                    term.move_cursor_backward()
+                elif control_type == ControlType.CURSOR_MOVE_TO_COLUMN:
+                    _, column = cast(Tuple[ControlType, int], control_code)
+                    term.move_cursor_to_column(column - 1)
+                elif control_type == ControlType.HIDE_CURSOR:
+                    term.hide_cursor()
+                elif control_type == ControlType.SHOW_CURSOR:
+                    term.show_cursor()
+                elif control_type == ControlType.ERASE_IN_LINE:
+                    _, mode = cast(Tuple[ControlType, int], control_code)
+                    if mode == 0:
+                        term.erase_end_of_line()
+                    elif mode == 1:
+                        term.erase_start_of_line()
+                    elif mode == 2:
+                        term.erase_line()
+                elif control_type == ControlType.SET_WINDOW_TITLE:
+                    _, title = cast(Tuple[ControlType, str], control_code)
+                    term.set_title(title)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_wrap.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_wrap.py
new file mode 100644
index 00000000..2e94ff6f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/_wrap.py
@@ -0,0 +1,93 @@
+from __future__ import annotations
+
+import re
+from typing import Iterable
+
+from ._loop import loop_last
+from .cells import cell_len, chop_cells
+
+re_word = re.compile(r"\s*\S+\s*")
+
+
+def words(text: str) -> Iterable[tuple[int, int, str]]:
+    """Yields each word from the text as a tuple
+    containing (start_index, end_index, word). A "word" in this context may
+    include the actual word and any whitespace to the right.
+    """
+    position = 0
+    word_match = re_word.match(text, position)
+    while word_match is not None:
+        start, end = word_match.span()
+        word = word_match.group(0)
+        yield start, end, word
+        word_match = re_word.match(text, end)
+
+
+def divide_line(text: str, width: int, fold: bool = True) -> list[int]:
+    """Given a string of text, and a width (measured in cells), return a list
+    of cell offsets which the string should be split at in order for it to fit
+    within the given width.
+
+    Args:
+        text: The text to examine.
+        width: The available cell width.
+        fold: If True, words longer than `width` will be folded onto a new line.
+
+    Returns:
+        A list of indices to break the line at.
+    """
+    break_positions: list[int] = []  # offsets to insert the breaks at
+    append = break_positions.append
+    cell_offset = 0
+    _cell_len = cell_len
+
+    for start, _end, word in words(text):
+        word_length = _cell_len(word.rstrip())
+        remaining_space = width - cell_offset
+        word_fits_remaining_space = remaining_space >= word_length
+
+        if word_fits_remaining_space:
+            # Simplest case - the word fits within the remaining width for this line.
+            cell_offset += _cell_len(word)
+        else:
+            # Not enough space remaining for this word on the current line.
+            if word_length > width:
+                # The word doesn't fit on any line, so we can't simply
+                # place it on the next line...
+                if fold:
+                    # Fold the word across multiple lines.
+                    folded_word = chop_cells(word, width=width)
+                    for last, line in loop_last(folded_word):
+                        if start:
+                            append(start)
+                        if last:
+                            cell_offset = _cell_len(line)
+                        else:
+                            start += len(line)
+                else:
+                    # Folding isn't allowed, so crop the word.
+                    if start:
+                        append(start)
+                    cell_offset = _cell_len(word)
+            elif cell_offset and start:
+                # The word doesn't fit within the remaining space on the current
+                # line, but it *can* fit on to the next (empty) line.
+                append(start)
+                cell_offset = _cell_len(word)
+
+    return break_positions
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from .console import Console
+
+    console = Console(width=10)
+    console.print("12345 abcdefghijklmnopqrstuvwyxzABCDEFGHIJKLMNOPQRSTUVWXYZ 12345")
+    print(chop_cells("abcdefghijklmnopqrstuvwxyz", 10))
+
+    console = Console(width=20)
+    console.rule()
+    console.print("TextualはPythonの高速アプリケーション開発フレームワークです")
+
+    console.rule()
+    console.print("アプリケーションは1670万色を使用でき")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/abc.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/abc.py
new file mode 100644
index 00000000..e6e498ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/abc.py
@@ -0,0 +1,33 @@
+from abc import ABC
+
+
+class RichRenderable(ABC):
+    """An abstract base class for Rich renderables.
+
+    Note that there is no need to extend this class, the intended use is to check if an
+    object supports the Rich renderable protocol. For example::
+
+        if isinstance(my_object, RichRenderable):
+            console.print(my_object)
+
+    """
+
+    @classmethod
+    def __subclasshook__(cls, other: type) -> bool:
+        """Check if this class supports the rich render protocol."""
+        return hasattr(other, "__rich_console__") or hasattr(other, "__rich__")
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.text import Text
+
+    t = Text()
+    print(isinstance(Text, RichRenderable))
+    print(isinstance(t, RichRenderable))
+
+    class Foo:
+        pass
+
+    f = Foo()
+    print(isinstance(f, RichRenderable))
+    print(isinstance("", RichRenderable))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/align.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/align.py
new file mode 100644
index 00000000..330dcc51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/align.py
@@ -0,0 +1,312 @@
+import sys
+from itertools import chain
+from typing import TYPE_CHECKING, Iterable, Optional
+
+if sys.version_info >= (3, 8):
+    from typing import Literal
+else:
+    from pip._vendor.typing_extensions import Literal  # pragma: no cover
+
+from .constrain import Constrain
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import StyleType
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderableType, RenderResult
+
+AlignMethod = Literal["left", "center", "right"]
+VerticalAlignMethod = Literal["top", "middle", "bottom"]
+
+
+class Align(JupyterMixin):
+    """Align a renderable by adding spaces if necessary.
+
+    Args:
+        renderable (RenderableType): A console renderable.
+        align (AlignMethod): One of "left", "center", or "right""
+        style (StyleType, optional): An optional style to apply to the background.
+        vertical (Optional[VerticalAlignMethod], optional): Optional vertical align, one of "top", "middle", or "bottom". Defaults to None.
+        pad (bool, optional): Pad the right with spaces. Defaults to True.
+        width (int, optional): Restrict contents to given width, or None to use default width. Defaults to None.
+        height (int, optional): Set height of align renderable, or None to fit to contents. Defaults to None.
+
+    Raises:
+        ValueError: if ``align`` is not one of the expected values.
+    """
+
+    def __init__(
+        self,
+        renderable: "RenderableType",
+        align: AlignMethod = "left",
+        style: Optional[StyleType] = None,
+        *,
+        vertical: Optional[VerticalAlignMethod] = None,
+        pad: bool = True,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+    ) -> None:
+        if align not in ("left", "center", "right"):
+            raise ValueError(
+                f'invalid value for align, expected "left", "center", or "right" (not {align!r})'
+            )
+        if vertical is not None and vertical not in ("top", "middle", "bottom"):
+            raise ValueError(
+                f'invalid value for vertical, expected "top", "middle", or "bottom" (not {vertical!r})'
+            )
+        self.renderable = renderable
+        self.align = align
+        self.style = style
+        self.vertical = vertical
+        self.pad = pad
+        self.width = width
+        self.height = height
+
+    def __repr__(self) -> str:
+        return f"Align({self.renderable!r}, {self.align!r})"
+
+    @classmethod
+    def left(
+        cls,
+        renderable: "RenderableType",
+        style: Optional[StyleType] = None,
+        *,
+        vertical: Optional[VerticalAlignMethod] = None,
+        pad: bool = True,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+    ) -> "Align":
+        """Align a renderable to the left."""
+        return cls(
+            renderable,
+            "left",
+            style=style,
+            vertical=vertical,
+            pad=pad,
+            width=width,
+            height=height,
+        )
+
+    @classmethod
+    def center(
+        cls,
+        renderable: "RenderableType",
+        style: Optional[StyleType] = None,
+        *,
+        vertical: Optional[VerticalAlignMethod] = None,
+        pad: bool = True,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+    ) -> "Align":
+        """Align a renderable to the center."""
+        return cls(
+            renderable,
+            "center",
+            style=style,
+            vertical=vertical,
+            pad=pad,
+            width=width,
+            height=height,
+        )
+
+    @classmethod
+    def right(
+        cls,
+        renderable: "RenderableType",
+        style: Optional[StyleType] = None,
+        *,
+        vertical: Optional[VerticalAlignMethod] = None,
+        pad: bool = True,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+    ) -> "Align":
+        """Align a renderable to the right."""
+        return cls(
+            renderable,
+            "right",
+            style=style,
+            vertical=vertical,
+            pad=pad,
+            width=width,
+            height=height,
+        )
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        align = self.align
+        width = console.measure(self.renderable, options=options).maximum
+        rendered = console.render(
+            Constrain(
+                self.renderable, width if self.width is None else min(width, self.width)
+            ),
+            options.update(height=None),
+        )
+        lines = list(Segment.split_lines(rendered))
+        width, height = Segment.get_shape(lines)
+        lines = Segment.set_shape(lines, width, height)
+        new_line = Segment.line()
+        excess_space = options.max_width - width
+        style = console.get_style(self.style) if self.style is not None else None
+
+        def generate_segments() -> Iterable[Segment]:
+            if excess_space <= 0:
+                # Exact fit
+                for line in lines:
+                    yield from line
+                    yield new_line
+
+            elif align == "left":
+                # Pad on the right
+                pad = Segment(" " * excess_space, style) if self.pad else None
+                for line in lines:
+                    yield from line
+                    if pad:
+                        yield pad
+                    yield new_line
+
+            elif align == "center":
+                # Pad left and right
+                left = excess_space // 2
+                pad = Segment(" " * left, style)
+                pad_right = (
+                    Segment(" " * (excess_space - left), style) if self.pad else None
+                )
+                for line in lines:
+                    if left:
+                        yield pad
+                    yield from line
+                    if pad_right:
+                        yield pad_right
+                    yield new_line
+
+            elif align == "right":
+                # Padding on left
+                pad = Segment(" " * excess_space, style)
+                for line in lines:
+                    yield pad
+                    yield from line
+                    yield new_line
+
+        blank_line = (
+            Segment(f"{' ' * (self.width or options.max_width)}\n", style)
+            if self.pad
+            else Segment("\n")
+        )
+
+        def blank_lines(count: int) -> Iterable[Segment]:
+            if count > 0:
+                for _ in range(count):
+                    yield blank_line
+
+        vertical_height = self.height or options.height
+        iter_segments: Iterable[Segment]
+        if self.vertical and vertical_height is not None:
+            if self.vertical == "top":
+                bottom_space = vertical_height - height
+                iter_segments = chain(generate_segments(), blank_lines(bottom_space))
+            elif self.vertical == "middle":
+                top_space = (vertical_height - height) // 2
+                bottom_space = vertical_height - top_space - height
+                iter_segments = chain(
+                    blank_lines(top_space),
+                    generate_segments(),
+                    blank_lines(bottom_space),
+                )
+            else:  #  self.vertical == "bottom":
+                top_space = vertical_height - height
+                iter_segments = chain(blank_lines(top_space), generate_segments())
+        else:
+            iter_segments = generate_segments()
+        if self.style:
+            style = console.get_style(self.style)
+            iter_segments = Segment.apply_style(iter_segments, style)
+        yield from iter_segments
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        measurement = Measurement.get(console, options, self.renderable)
+        return measurement
+
+
+class VerticalCenter(JupyterMixin):
+    """Vertically aligns a renderable.
+
+    Warn:
+        This class is deprecated and may be removed in a future version. Use Align class with
+        `vertical="middle"`.
+
+    Args:
+        renderable (RenderableType): A renderable object.
+        style (StyleType, optional): An optional style to apply to the background. Defaults to None.
+    """
+
+    def __init__(
+        self,
+        renderable: "RenderableType",
+        style: Optional[StyleType] = None,
+    ) -> None:
+        self.renderable = renderable
+        self.style = style
+
+    def __repr__(self) -> str:
+        return f"VerticalCenter({self.renderable!r})"
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        style = console.get_style(self.style) if self.style is not None else None
+        lines = console.render_lines(
+            self.renderable, options.update(height=None), pad=False
+        )
+        width, _height = Segment.get_shape(lines)
+        new_line = Segment.line()
+        height = options.height or options.size.height
+        top_space = (height - len(lines)) // 2
+        bottom_space = height - top_space - len(lines)
+        blank_line = Segment(f"{' ' * width}", style)
+
+        def blank_lines(count: int) -> Iterable[Segment]:
+            for _ in range(count):
+                yield blank_line
+                yield new_line
+
+        if top_space > 0:
+            yield from blank_lines(top_space)
+        for line in lines:
+            yield from line
+            yield new_line
+        if bottom_space > 0:
+            yield from blank_lines(bottom_space)
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        measurement = Measurement.get(console, options, self.renderable)
+        return measurement
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console, Group
+    from pip._vendor.rich.highlighter import ReprHighlighter
+    from pip._vendor.rich.panel import Panel
+
+    highlighter = ReprHighlighter()
+    console = Console()
+
+    panel = Panel(
+        Group(
+            Align.left(highlighter("align='left'")),
+            Align.center(highlighter("align='center'")),
+            Align.right(highlighter("align='right'")),
+        ),
+        width=60,
+        style="on dark_blue",
+        title="Align",
+    )
+
+    console.print(
+        Align.center(panel, vertical="middle", style="on red", height=console.height)
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/ansi.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/ansi.py
new file mode 100644
index 00000000..7de86ce5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/ansi.py
@@ -0,0 +1,241 @@
+import re
+import sys
+from contextlib import suppress
+from typing import Iterable, NamedTuple, Optional
+
+from .color import Color
+from .style import Style
+from .text import Text
+
+re_ansi = re.compile(
+    r"""
+(?:\x1b[0-?])|
+(?:\x1b\](.*?)\x1b\\)|
+(?:\x1b([(@-Z\\-_]|\[[0-?]*[ -/]*[@-~]))
+""",
+    re.VERBOSE,
+)
+
+
+class _AnsiToken(NamedTuple):
+    """Result of ansi tokenized string."""
+
+    plain: str = ""
+    sgr: Optional[str] = ""
+    osc: Optional[str] = ""
+
+
+def _ansi_tokenize(ansi_text: str) -> Iterable[_AnsiToken]:
+    """Tokenize a string in to plain text and ANSI codes.
+
+    Args:
+        ansi_text (str): A String containing ANSI codes.
+
+    Yields:
+        AnsiToken: A named tuple of (plain, sgr, osc)
+    """
+
+    position = 0
+    sgr: Optional[str]
+    osc: Optional[str]
+    for match in re_ansi.finditer(ansi_text):
+        start, end = match.span(0)
+        osc, sgr = match.groups()
+        if start > position:
+            yield _AnsiToken(ansi_text[position:start])
+        if sgr:
+            if sgr == "(":
+                position = end + 1
+                continue
+            if sgr.endswith("m"):
+                yield _AnsiToken("", sgr[1:-1], osc)
+        else:
+            yield _AnsiToken("", sgr, osc)
+        position = end
+    if position < len(ansi_text):
+        yield _AnsiToken(ansi_text[position:])
+
+
+SGR_STYLE_MAP = {
+    1: "bold",
+    2: "dim",
+    3: "italic",
+    4: "underline",
+    5: "blink",
+    6: "blink2",
+    7: "reverse",
+    8: "conceal",
+    9: "strike",
+    21: "underline2",
+    22: "not dim not bold",
+    23: "not italic",
+    24: "not underline",
+    25: "not blink",
+    26: "not blink2",
+    27: "not reverse",
+    28: "not conceal",
+    29: "not strike",
+    30: "color(0)",
+    31: "color(1)",
+    32: "color(2)",
+    33: "color(3)",
+    34: "color(4)",
+    35: "color(5)",
+    36: "color(6)",
+    37: "color(7)",
+    39: "default",
+    40: "on color(0)",
+    41: "on color(1)",
+    42: "on color(2)",
+    43: "on color(3)",
+    44: "on color(4)",
+    45: "on color(5)",
+    46: "on color(6)",
+    47: "on color(7)",
+    49: "on default",
+    51: "frame",
+    52: "encircle",
+    53: "overline",
+    54: "not frame not encircle",
+    55: "not overline",
+    90: "color(8)",
+    91: "color(9)",
+    92: "color(10)",
+    93: "color(11)",
+    94: "color(12)",
+    95: "color(13)",
+    96: "color(14)",
+    97: "color(15)",
+    100: "on color(8)",
+    101: "on color(9)",
+    102: "on color(10)",
+    103: "on color(11)",
+    104: "on color(12)",
+    105: "on color(13)",
+    106: "on color(14)",
+    107: "on color(15)",
+}
+
+
+class AnsiDecoder:
+    """Translate ANSI code in to styled Text."""
+
+    def __init__(self) -> None:
+        self.style = Style.null()
+
+    def decode(self, terminal_text: str) -> Iterable[Text]:
+        """Decode ANSI codes in an iterable of lines.
+
+        Args:
+            lines (Iterable[str]): An iterable of lines of terminal output.
+
+        Yields:
+            Text: Marked up Text.
+        """
+        for line in terminal_text.splitlines():
+            yield self.decode_line(line)
+
+    def decode_line(self, line: str) -> Text:
+        """Decode a line containing ansi codes.
+
+        Args:
+            line (str): A line of terminal output.
+
+        Returns:
+            Text: A Text instance marked up according to ansi codes.
+        """
+        from_ansi = Color.from_ansi
+        from_rgb = Color.from_rgb
+        _Style = Style
+        text = Text()
+        append = text.append
+        line = line.rsplit("\r", 1)[-1]
+        for plain_text, sgr, osc in _ansi_tokenize(line):
+            if plain_text:
+                append(plain_text, self.style or None)
+            elif osc is not None:
+                if osc.startswith("8;"):
+                    _params, semicolon, link = osc[2:].partition(";")
+                    if semicolon:
+                        self.style = self.style.update_link(link or None)
+            elif sgr is not None:
+                # Translate in to semi-colon separated codes
+                # Ignore invalid codes, because we want to be lenient
+                codes = [
+                    min(255, int(_code) if _code else 0)
+                    for _code in sgr.split(";")
+                    if _code.isdigit() or _code == ""
+                ]
+                iter_codes = iter(codes)
+                for code in iter_codes:
+                    if code == 0:
+                        # reset
+                        self.style = _Style.null()
+                    elif code in SGR_STYLE_MAP:
+                        # styles
+                        self.style += _Style.parse(SGR_STYLE_MAP[code])
+                    elif code == 38:
+                        #  Foreground
+                        with suppress(StopIteration):
+                            color_type = next(iter_codes)
+                            if color_type == 5:
+                                self.style += _Style.from_color(
+                                    from_ansi(next(iter_codes))
+                                )
+                            elif color_type == 2:
+                                self.style += _Style.from_color(
+                                    from_rgb(
+                                        next(iter_codes),
+                                        next(iter_codes),
+                                        next(iter_codes),
+                                    )
+                                )
+                    elif code == 48:
+                        # Background
+                        with suppress(StopIteration):
+                            color_type = next(iter_codes)
+                            if color_type == 5:
+                                self.style += _Style.from_color(
+                                    None, from_ansi(next(iter_codes))
+                                )
+                            elif color_type == 2:
+                                self.style += _Style.from_color(
+                                    None,
+                                    from_rgb(
+                                        next(iter_codes),
+                                        next(iter_codes),
+                                        next(iter_codes),
+                                    ),
+                                )
+
+        return text
+
+
+if sys.platform != "win32" and __name__ == "__main__":  # pragma: no cover
+    import io
+    import os
+    import pty
+    import sys
+
+    decoder = AnsiDecoder()
+
+    stdout = io.BytesIO()
+
+    def read(fd: int) -> bytes:
+        data = os.read(fd, 1024)
+        stdout.write(data)
+        return data
+
+    pty.spawn(sys.argv[1:], read)
+
+    from .console import Console
+
+    console = Console(record=True)
+
+    stdout_result = stdout.getvalue().decode("utf-8")
+    print(stdout_result)
+
+    for line in decoder.decode(stdout_result):
+        console.print(line)
+
+    console.save_html("stdout.html")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/bar.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/bar.py
new file mode 100644
index 00000000..022284b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/bar.py
@@ -0,0 +1,93 @@
+from typing import Optional, Union
+
+from .color import Color
+from .console import Console, ConsoleOptions, RenderResult
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import Style
+
+# There are left-aligned characters for 1/8 to 7/8, but
+# the right-aligned characters exist only for 1/8 and 4/8.
+BEGIN_BLOCK_ELEMENTS = ["█", "█", "█", "▐", "▐", "▐", "▕", "▕"]
+END_BLOCK_ELEMENTS = [" ", "▏", "▎", "▍", "▌", "▋", "▊", "▉"]
+FULL_BLOCK = "█"
+
+
+class Bar(JupyterMixin):
+    """Renders a solid block bar.
+
+    Args:
+        size (float): Value for the end of the bar.
+        begin (float): Begin point (between 0 and size, inclusive).
+        end (float): End point (between 0 and size, inclusive).
+        width (int, optional): Width of the bar, or ``None`` for maximum width. Defaults to None.
+        color (Union[Color, str], optional): Color of the bar. Defaults to "default".
+        bgcolor (Union[Color, str], optional): Color of bar background. Defaults to "default".
+    """
+
+    def __init__(
+        self,
+        size: float,
+        begin: float,
+        end: float,
+        *,
+        width: Optional[int] = None,
+        color: Union[Color, str] = "default",
+        bgcolor: Union[Color, str] = "default",
+    ):
+        self.size = size
+        self.begin = max(begin, 0)
+        self.end = min(end, size)
+        self.width = width
+        self.style = Style(color=color, bgcolor=bgcolor)
+
+    def __repr__(self) -> str:
+        return f"Bar({self.size}, {self.begin}, {self.end})"
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        width = min(
+            self.width if self.width is not None else options.max_width,
+            options.max_width,
+        )
+
+        if self.begin >= self.end:
+            yield Segment(" " * width, self.style)
+            yield Segment.line()
+            return
+
+        prefix_complete_eights = int(width * 8 * self.begin / self.size)
+        prefix_bar_count = prefix_complete_eights // 8
+        prefix_eights_count = prefix_complete_eights % 8
+
+        body_complete_eights = int(width * 8 * self.end / self.size)
+        body_bar_count = body_complete_eights // 8
+        body_eights_count = body_complete_eights % 8
+
+        # When start and end fall into the same cell, we ideally should render
+        # a symbol that's "center-aligned", but there is no good symbol in Unicode.
+        # In this case, we fall back to right-aligned block symbol for simplicity.
+
+        prefix = " " * prefix_bar_count
+        if prefix_eights_count:
+            prefix += BEGIN_BLOCK_ELEMENTS[prefix_eights_count]
+
+        body = FULL_BLOCK * body_bar_count
+        if body_eights_count:
+            body += END_BLOCK_ELEMENTS[body_eights_count]
+
+        suffix = " " * (width - len(body))
+
+        yield Segment(prefix + body[len(prefix) :] + suffix, self.style)
+        yield Segment.line()
+
+    def __rich_measure__(
+        self, console: Console, options: ConsoleOptions
+    ) -> Measurement:
+        return (
+            Measurement(self.width, self.width)
+            if self.width is not None
+            else Measurement(4, options.max_width)
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/box.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/box.py
new file mode 100644
index 00000000..0511a9e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/box.py
@@ -0,0 +1,480 @@
+import sys
+from typing import TYPE_CHECKING, Iterable, List
+
+if sys.version_info >= (3, 8):
+    from typing import Literal
+else:
+    from pip._vendor.typing_extensions import Literal  # pragma: no cover
+
+
+from ._loop import loop_last
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.console import ConsoleOptions
+
+
+class Box:
+    """Defines characters to render boxes.
+
+    ┌─┬┐ top
+    │ ││ head
+    ├─┼┤ head_row
+    │ ││ mid
+    ├─┼┤ row
+    ├─┼┤ foot_row
+    │ ││ foot
+    └─┴┘ bottom
+
+    Args:
+        box (str): Characters making up box.
+        ascii (bool, optional): True if this box uses ascii characters only. Default is False.
+    """
+
+    def __init__(self, box: str, *, ascii: bool = False) -> None:
+        self._box = box
+        self.ascii = ascii
+        line1, line2, line3, line4, line5, line6, line7, line8 = box.splitlines()
+        # top
+        self.top_left, self.top, self.top_divider, self.top_right = iter(line1)
+        # head
+        self.head_left, _, self.head_vertical, self.head_right = iter(line2)
+        # head_row
+        (
+            self.head_row_left,
+            self.head_row_horizontal,
+            self.head_row_cross,
+            self.head_row_right,
+        ) = iter(line3)
+
+        # mid
+        self.mid_left, _, self.mid_vertical, self.mid_right = iter(line4)
+        # row
+        self.row_left, self.row_horizontal, self.row_cross, self.row_right = iter(line5)
+        # foot_row
+        (
+            self.foot_row_left,
+            self.foot_row_horizontal,
+            self.foot_row_cross,
+            self.foot_row_right,
+        ) = iter(line6)
+        # foot
+        self.foot_left, _, self.foot_vertical, self.foot_right = iter(line7)
+        # bottom
+        self.bottom_left, self.bottom, self.bottom_divider, self.bottom_right = iter(
+            line8
+        )
+
+    def __repr__(self) -> str:
+        return "Box(...)"
+
+    def __str__(self) -> str:
+        return self._box
+
+    def substitute(self, options: "ConsoleOptions", safe: bool = True) -> "Box":
+        """Substitute this box for another if it won't render due to platform issues.
+
+        Args:
+            options (ConsoleOptions): Console options used in rendering.
+            safe (bool, optional): Substitute this for another Box if there are known problems
+                displaying on the platform (currently only relevant on Windows). Default is True.
+
+        Returns:
+            Box: A different Box or the same Box.
+        """
+        box = self
+        if options.legacy_windows and safe:
+            box = LEGACY_WINDOWS_SUBSTITUTIONS.get(box, box)
+        if options.ascii_only and not box.ascii:
+            box = ASCII
+        return box
+
+    def get_plain_headed_box(self) -> "Box":
+        """If this box uses special characters for the borders of the header, then
+        return the equivalent box that does not.
+
+        Returns:
+            Box: The most similar Box that doesn't use header-specific box characters.
+                If the current Box already satisfies this criterion, then it's returned.
+        """
+        return PLAIN_HEADED_SUBSTITUTIONS.get(self, self)
+
+    def get_top(self, widths: Iterable[int]) -> str:
+        """Get the top of a simple box.
+
+        Args:
+            widths (List[int]): Widths of columns.
+
+        Returns:
+            str: A string of box characters.
+        """
+
+        parts: List[str] = []
+        append = parts.append
+        append(self.top_left)
+        for last, width in loop_last(widths):
+            append(self.top * width)
+            if not last:
+                append(self.top_divider)
+        append(self.top_right)
+        return "".join(parts)
+
+    def get_row(
+        self,
+        widths: Iterable[int],
+        level: Literal["head", "row", "foot", "mid"] = "row",
+        edge: bool = True,
+    ) -> str:
+        """Get the top of a simple box.
+
+        Args:
+            width (List[int]): Widths of columns.
+
+        Returns:
+            str: A string of box characters.
+        """
+        if level == "head":
+            left = self.head_row_left
+            horizontal = self.head_row_horizontal
+            cross = self.head_row_cross
+            right = self.head_row_right
+        elif level == "row":
+            left = self.row_left
+            horizontal = self.row_horizontal
+            cross = self.row_cross
+            right = self.row_right
+        elif level == "mid":
+            left = self.mid_left
+            horizontal = " "
+            cross = self.mid_vertical
+            right = self.mid_right
+        elif level == "foot":
+            left = self.foot_row_left
+            horizontal = self.foot_row_horizontal
+            cross = self.foot_row_cross
+            right = self.foot_row_right
+        else:
+            raise ValueError("level must be 'head', 'row' or 'foot'")
+
+        parts: List[str] = []
+        append = parts.append
+        if edge:
+            append(left)
+        for last, width in loop_last(widths):
+            append(horizontal * width)
+            if not last:
+                append(cross)
+        if edge:
+            append(right)
+        return "".join(parts)
+
+    def get_bottom(self, widths: Iterable[int]) -> str:
+        """Get the bottom of a simple box.
+
+        Args:
+            widths (List[int]): Widths of columns.
+
+        Returns:
+            str: A string of box characters.
+        """
+
+        parts: List[str] = []
+        append = parts.append
+        append(self.bottom_left)
+        for last, width in loop_last(widths):
+            append(self.bottom * width)
+            if not last:
+                append(self.bottom_divider)
+        append(self.bottom_right)
+        return "".join(parts)
+
+
+# fmt: off
+ASCII: Box = Box(
+    "+--+\n"
+    "| ||\n"
+    "|-+|\n"
+    "| ||\n"
+    "|-+|\n"
+    "|-+|\n"
+    "| ||\n"
+    "+--+\n",
+    ascii=True,
+)
+
+ASCII2: Box = Box(
+    "+-++\n"
+    "| ||\n"
+    "+-++\n"
+    "| ||\n"
+    "+-++\n"
+    "+-++\n"
+    "| ||\n"
+    "+-++\n",
+    ascii=True,
+)
+
+ASCII_DOUBLE_HEAD: Box = Box(
+    "+-++\n"
+    "| ||\n"
+    "+=++\n"
+    "| ||\n"
+    "+-++\n"
+    "+-++\n"
+    "| ||\n"
+    "+-++\n",
+    ascii=True,
+)
+
+SQUARE: Box = Box(
+    "┌─┬┐\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "└─┴┘\n"
+)
+
+SQUARE_DOUBLE_HEAD: Box = Box(
+    "┌─┬┐\n"
+    "│ ││\n"
+    "╞═╪╡\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "└─┴┘\n"
+)
+
+MINIMAL: Box = Box(
+    "  ╷ \n"
+    "  │ \n"
+    "╶─┼╴\n"
+    "  │ \n"
+    "╶─┼╴\n"
+    "╶─┼╴\n"
+    "  │ \n"
+    "  ╵ \n"
+)
+
+
+MINIMAL_HEAVY_HEAD: Box = Box(
+    "  ╷ \n"
+    "  │ \n"
+    "╺━┿╸\n"
+    "  │ \n"
+    "╶─┼╴\n"
+    "╶─┼╴\n"
+    "  │ \n"
+    "  ╵ \n"
+)
+
+MINIMAL_DOUBLE_HEAD: Box = Box(
+    "  ╷ \n"
+    "  │ \n"
+    " ═╪ \n"
+    "  │ \n"
+    " ─┼ \n"
+    " ─┼ \n"
+    "  │ \n"
+    "  ╵ \n"
+)
+
+
+SIMPLE: Box = Box(
+    "    \n"
+    "    \n"
+    " ── \n"
+    "    \n"
+    "    \n"
+    " ── \n"
+    "    \n"
+    "    \n"
+)
+
+SIMPLE_HEAD: Box = Box(
+    "    \n"
+    "    \n"
+    " ── \n"
+    "    \n"
+    "    \n"
+    "    \n"
+    "    \n"
+    "    \n"
+)
+
+
+SIMPLE_HEAVY: Box = Box(
+    "    \n"
+    "    \n"
+    " ━━ \n"
+    "    \n"
+    "    \n"
+    " ━━ \n"
+    "    \n"
+    "    \n"
+)
+
+
+HORIZONTALS: Box = Box(
+    " ── \n"
+    "    \n"
+    " ── \n"
+    "    \n"
+    " ── \n"
+    " ── \n"
+    "    \n"
+    " ── \n"
+)
+
+ROUNDED: Box = Box(
+    "╭─┬╮\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "╰─┴╯\n"
+)
+
+HEAVY: Box = Box(
+    "┏━┳┓\n"
+    "┃ ┃┃\n"
+    "┣━╋┫\n"
+    "┃ ┃┃\n"
+    "┣━╋┫\n"
+    "┣━╋┫\n"
+    "┃ ┃┃\n"
+    "┗━┻┛\n"
+)
+
+HEAVY_EDGE: Box = Box(
+    "┏━┯┓\n"
+    "┃ │┃\n"
+    "┠─┼┨\n"
+    "┃ │┃\n"
+    "┠─┼┨\n"
+    "┠─┼┨\n"
+    "┃ │┃\n"
+    "┗━┷┛\n"
+)
+
+HEAVY_HEAD: Box = Box(
+    "┏━┳┓\n"
+    "┃ ┃┃\n"
+    "┡━╇┩\n"
+    "│ ││\n"
+    "├─┼┤\n"
+    "├─┼┤\n"
+    "│ ││\n"
+    "└─┴┘\n"
+)
+
+DOUBLE: Box = Box(
+    "╔═╦╗\n"
+    "║ ║║\n"
+    "╠═╬╣\n"
+    "║ ║║\n"
+    "╠═╬╣\n"
+    "╠═╬╣\n"
+    "║ ║║\n"
+    "╚═╩╝\n"
+)
+
+DOUBLE_EDGE: Box = Box(
+    "╔═╤╗\n"
+    "║ │║\n"
+    "╟─┼╢\n"
+    "║ │║\n"
+    "╟─┼╢\n"
+    "╟─┼╢\n"
+    "║ │║\n"
+    "╚═╧╝\n"
+)
+
+MARKDOWN: Box = Box(
+    "    \n"
+    "| ||\n"
+    "|-||\n"
+    "| ||\n"
+    "|-||\n"
+    "|-||\n"
+    "| ||\n"
+    "    \n",
+    ascii=True,
+)
+# fmt: on
+
+# Map Boxes that don't render with raster fonts on to equivalent that do
+LEGACY_WINDOWS_SUBSTITUTIONS = {
+    ROUNDED: SQUARE,
+    MINIMAL_HEAVY_HEAD: MINIMAL,
+    SIMPLE_HEAVY: SIMPLE,
+    HEAVY: SQUARE,
+    HEAVY_EDGE: SQUARE,
+    HEAVY_HEAD: SQUARE,
+}
+
+# Map headed boxes to their headerless equivalents
+PLAIN_HEADED_SUBSTITUTIONS = {
+    HEAVY_HEAD: SQUARE,
+    SQUARE_DOUBLE_HEAD: SQUARE,
+    MINIMAL_DOUBLE_HEAD: MINIMAL,
+    MINIMAL_HEAVY_HEAD: MINIMAL,
+    ASCII_DOUBLE_HEAD: ASCII2,
+}
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.columns import Columns
+    from pip._vendor.rich.panel import Panel
+
+    from . import box as box
+    from .console import Console
+    from .table import Table
+    from .text import Text
+
+    console = Console(record=True)
+
+    BOXES = [
+        "ASCII",
+        "ASCII2",
+        "ASCII_DOUBLE_HEAD",
+        "SQUARE",
+        "SQUARE_DOUBLE_HEAD",
+        "MINIMAL",
+        "MINIMAL_HEAVY_HEAD",
+        "MINIMAL_DOUBLE_HEAD",
+        "SIMPLE",
+        "SIMPLE_HEAD",
+        "SIMPLE_HEAVY",
+        "HORIZONTALS",
+        "ROUNDED",
+        "HEAVY",
+        "HEAVY_EDGE",
+        "HEAVY_HEAD",
+        "DOUBLE",
+        "DOUBLE_EDGE",
+        "MARKDOWN",
+    ]
+
+    console.print(Panel("[bold green]Box Constants", style="green"), justify="center")
+    console.print()
+
+    columns = Columns(expand=True, padding=2)
+    for box_name in sorted(BOXES):
+        table = Table(
+            show_footer=True, style="dim", border_style="not dim", expand=True
+        )
+        table.add_column("Header 1", "Footer 1")
+        table.add_column("Header 2", "Footer 2")
+        table.add_row("Cell", "Cell")
+        table.add_row("Cell", "Cell")
+        table.box = getattr(box, box_name)
+        table.title = Text(f"box.{box_name}", style="magenta")
+        columns.add_renderable(table)
+    console.print(columns)
+
+    # console.save_svg("box.svg")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/cells.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/cells.py
new file mode 100644
index 00000000..a8546227
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/cells.py
@@ -0,0 +1,174 @@
+from __future__ import annotations
+
+from functools import lru_cache
+from typing import Callable
+
+from ._cell_widths import CELL_WIDTHS
+
+# Ranges of unicode ordinals that produce a 1-cell wide character
+# This is non-exhaustive, but covers most common Western characters
+_SINGLE_CELL_UNICODE_RANGES: list[tuple[int, int]] = [
+    (0x20, 0x7E),  # Latin (excluding non-printable)
+    (0xA0, 0xAC),
+    (0xAE, 0x002FF),
+    (0x00370, 0x00482),  # Greek / Cyrillic
+    (0x02500, 0x025FC),  # Box drawing, box elements, geometric shapes
+    (0x02800, 0x028FF),  # Braille
+]
+
+# A set of characters that are a single cell wide
+_SINGLE_CELLS = frozenset(
+    [
+        character
+        for _start, _end in _SINGLE_CELL_UNICODE_RANGES
+        for character in map(chr, range(_start, _end + 1))
+    ]
+)
+
+# When called with a string this will return True if all
+# characters are single-cell, otherwise False
+_is_single_cell_widths: Callable[[str], bool] = _SINGLE_CELLS.issuperset
+
+
+@lru_cache(4096)
+def cached_cell_len(text: str) -> int:
+    """Get the number of cells required to display text.
+
+    This method always caches, which may use up a lot of memory. It is recommended to use
+    `cell_len` over this method.
+
+    Args:
+        text (str): Text to display.
+
+    Returns:
+        int: Get the number of cells required to display text.
+    """
+    if _is_single_cell_widths(text):
+        return len(text)
+    return sum(map(get_character_cell_size, text))
+
+
+def cell_len(text: str, _cell_len: Callable[[str], int] = cached_cell_len) -> int:
+    """Get the number of cells required to display text.
+
+    Args:
+        text (str): Text to display.
+
+    Returns:
+        int: Get the number of cells required to display text.
+    """
+    if len(text) < 512:
+        return _cell_len(text)
+    if _is_single_cell_widths(text):
+        return len(text)
+    return sum(map(get_character_cell_size, text))
+
+
+@lru_cache(maxsize=4096)
+def get_character_cell_size(character: str) -> int:
+    """Get the cell size of a character.
+
+    Args:
+        character (str): A single character.
+
+    Returns:
+        int: Number of cells (0, 1 or 2) occupied by that character.
+    """
+    codepoint = ord(character)
+    _table = CELL_WIDTHS
+    lower_bound = 0
+    upper_bound = len(_table) - 1
+    index = (lower_bound + upper_bound) // 2
+    while True:
+        start, end, width = _table[index]
+        if codepoint < start:
+            upper_bound = index - 1
+        elif codepoint > end:
+            lower_bound = index + 1
+        else:
+            return 0 if width == -1 else width
+        if upper_bound < lower_bound:
+            break
+        index = (lower_bound + upper_bound) // 2
+    return 1
+
+
+def set_cell_size(text: str, total: int) -> str:
+    """Set the length of a string to fit within given number of cells."""
+
+    if _is_single_cell_widths(text):
+        size = len(text)
+        if size < total:
+            return text + " " * (total - size)
+        return text[:total]
+
+    if total <= 0:
+        return ""
+    cell_size = cell_len(text)
+    if cell_size == total:
+        return text
+    if cell_size < total:
+        return text + " " * (total - cell_size)
+
+    start = 0
+    end = len(text)
+
+    # Binary search until we find the right size
+    while True:
+        pos = (start + end) // 2
+        before = text[: pos + 1]
+        before_len = cell_len(before)
+        if before_len == total + 1 and cell_len(before[-1]) == 2:
+            return before[:-1] + " "
+        if before_len == total:
+            return before
+        if before_len > total:
+            end = pos
+        else:
+            start = pos
+
+
+def chop_cells(
+    text: str,
+    width: int,
+) -> list[str]:
+    """Split text into lines such that each line fits within the available (cell) width.
+
+    Args:
+        text: The text to fold such that it fits in the given width.
+        width: The width available (number of cells).
+
+    Returns:
+        A list of strings such that each string in the list has cell width
+        less than or equal to the available width.
+    """
+    _get_character_cell_size = get_character_cell_size
+    lines: list[list[str]] = [[]]
+
+    append_new_line = lines.append
+    append_to_last_line = lines[-1].append
+
+    total_width = 0
+
+    for character in text:
+        cell_width = _get_character_cell_size(character)
+        char_doesnt_fit = total_width + cell_width > width
+
+        if char_doesnt_fit:
+            append_new_line([character])
+            append_to_last_line = lines[-1].append
+            total_width = cell_width
+        else:
+            append_to_last_line(character)
+            total_width += cell_width
+
+    return ["".join(line) for line in lines]
+
+
+if __name__ == "__main__":  # pragma: no cover
+    print(get_character_cell_size("😽"))
+    for line in chop_cells("""这是对亚洲语言支持的测试。面对模棱两可的想法，拒绝猜测的诱惑。""", 8):
+        print(line)
+    for n in range(80, 1, -1):
+        print(set_cell_size("""这是对亚洲语言支持的测试。面对模棱两可的想法，拒绝猜测的诱惑。""", n) + "|")
+        print("x" * n)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color.py
new file mode 100644
index 00000000..e2c23a6a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color.py
@@ -0,0 +1,621 @@
+import re
+import sys
+from colorsys import rgb_to_hls
+from enum import IntEnum
+from functools import lru_cache
+from typing import TYPE_CHECKING, NamedTuple, Optional, Tuple
+
+from ._palettes import EIGHT_BIT_PALETTE, STANDARD_PALETTE, WINDOWS_PALETTE
+from .color_triplet import ColorTriplet
+from .repr import Result, rich_repr
+from .terminal_theme import DEFAULT_TERMINAL_THEME
+
+if TYPE_CHECKING:  # pragma: no cover
+    from .terminal_theme import TerminalTheme
+    from .text import Text
+
+
+WINDOWS = sys.platform == "win32"
+
+
+class ColorSystem(IntEnum):
+    """One of the 3 color system supported by terminals."""
+
+    STANDARD = 1
+    EIGHT_BIT = 2
+    TRUECOLOR = 3
+    WINDOWS = 4
+
+    def __repr__(self) -> str:
+        return f"ColorSystem.{self.name}"
+
+    def __str__(self) -> str:
+        return repr(self)
+
+
+class ColorType(IntEnum):
+    """Type of color stored in Color class."""
+
+    DEFAULT = 0
+    STANDARD = 1
+    EIGHT_BIT = 2
+    TRUECOLOR = 3
+    WINDOWS = 4
+
+    def __repr__(self) -> str:
+        return f"ColorType.{self.name}"
+
+
+ANSI_COLOR_NAMES = {
+    "black": 0,
+    "red": 1,
+    "green": 2,
+    "yellow": 3,
+    "blue": 4,
+    "magenta": 5,
+    "cyan": 6,
+    "white": 7,
+    "bright_black": 8,
+    "bright_red": 9,
+    "bright_green": 10,
+    "bright_yellow": 11,
+    "bright_blue": 12,
+    "bright_magenta": 13,
+    "bright_cyan": 14,
+    "bright_white": 15,
+    "grey0": 16,
+    "gray0": 16,
+    "navy_blue": 17,
+    "dark_blue": 18,
+    "blue3": 20,
+    "blue1": 21,
+    "dark_green": 22,
+    "deep_sky_blue4": 25,
+    "dodger_blue3": 26,
+    "dodger_blue2": 27,
+    "green4": 28,
+    "spring_green4": 29,
+    "turquoise4": 30,
+    "deep_sky_blue3": 32,
+    "dodger_blue1": 33,
+    "green3": 40,
+    "spring_green3": 41,
+    "dark_cyan": 36,
+    "light_sea_green": 37,
+    "deep_sky_blue2": 38,
+    "deep_sky_blue1": 39,
+    "spring_green2": 47,
+    "cyan3": 43,
+    "dark_turquoise": 44,
+    "turquoise2": 45,
+    "green1": 46,
+    "spring_green1": 48,
+    "medium_spring_green": 49,
+    "cyan2": 50,
+    "cyan1": 51,
+    "dark_red": 88,
+    "deep_pink4": 125,
+    "purple4": 55,
+    "purple3": 56,
+    "blue_violet": 57,
+    "orange4": 94,
+    "grey37": 59,
+    "gray37": 59,
+    "medium_purple4": 60,
+    "slate_blue3": 62,
+    "royal_blue1": 63,
+    "chartreuse4": 64,
+    "dark_sea_green4": 71,
+    "pale_turquoise4": 66,
+    "steel_blue": 67,
+    "steel_blue3": 68,
+    "cornflower_blue": 69,
+    "chartreuse3": 76,
+    "cadet_blue": 73,
+    "sky_blue3": 74,
+    "steel_blue1": 81,
+    "pale_green3": 114,
+    "sea_green3": 78,
+    "aquamarine3": 79,
+    "medium_turquoise": 80,
+    "chartreuse2": 112,
+    "sea_green2": 83,
+    "sea_green1": 85,
+    "aquamarine1": 122,
+    "dark_slate_gray2": 87,
+    "dark_magenta": 91,
+    "dark_violet": 128,
+    "purple": 129,
+    "light_pink4": 95,
+    "plum4": 96,
+    "medium_purple3": 98,
+    "slate_blue1": 99,
+    "yellow4": 106,
+    "wheat4": 101,
+    "grey53": 102,
+    "gray53": 102,
+    "light_slate_grey": 103,
+    "light_slate_gray": 103,
+    "medium_purple": 104,
+    "light_slate_blue": 105,
+    "dark_olive_green3": 149,
+    "dark_sea_green": 108,
+    "light_sky_blue3": 110,
+    "sky_blue2": 111,
+    "dark_sea_green3": 150,
+    "dark_slate_gray3": 116,
+    "sky_blue1": 117,
+    "chartreuse1": 118,
+    "light_green": 120,
+    "pale_green1": 156,
+    "dark_slate_gray1": 123,
+    "red3": 160,
+    "medium_violet_red": 126,
+    "magenta3": 164,
+    "dark_orange3": 166,
+    "indian_red": 167,
+    "hot_pink3": 168,
+    "medium_orchid3": 133,
+    "medium_orchid": 134,
+    "medium_purple2": 140,
+    "dark_goldenrod": 136,
+    "light_salmon3": 173,
+    "rosy_brown": 138,
+    "grey63": 139,
+    "gray63": 139,
+    "medium_purple1": 141,
+    "gold3": 178,
+    "dark_khaki": 143,
+    "navajo_white3": 144,
+    "grey69": 145,
+    "gray69": 145,
+    "light_steel_blue3": 146,
+    "light_steel_blue": 147,
+    "yellow3": 184,
+    "dark_sea_green2": 157,
+    "light_cyan3": 152,
+    "light_sky_blue1": 153,
+    "green_yellow": 154,
+    "dark_olive_green2": 155,
+    "dark_sea_green1": 193,
+    "pale_turquoise1": 159,
+    "deep_pink3": 162,
+    "magenta2": 200,
+    "hot_pink2": 169,
+    "orchid": 170,
+    "medium_orchid1": 207,
+    "orange3": 172,
+    "light_pink3": 174,
+    "pink3": 175,
+    "plum3": 176,
+    "violet": 177,
+    "light_goldenrod3": 179,
+    "tan": 180,
+    "misty_rose3": 181,
+    "thistle3": 182,
+    "plum2": 183,
+    "khaki3": 185,
+    "light_goldenrod2": 222,
+    "light_yellow3": 187,
+    "grey84": 188,
+    "gray84": 188,
+    "light_steel_blue1": 189,
+    "yellow2": 190,
+    "dark_olive_green1": 192,
+    "honeydew2": 194,
+    "light_cyan1": 195,
+    "red1": 196,
+    "deep_pink2": 197,
+    "deep_pink1": 199,
+    "magenta1": 201,
+    "orange_red1": 202,
+    "indian_red1": 204,
+    "hot_pink": 206,
+    "dark_orange": 208,
+    "salmon1": 209,
+    "light_coral": 210,
+    "pale_violet_red1": 211,
+    "orchid2": 212,
+    "orchid1": 213,
+    "orange1": 214,
+    "sandy_brown": 215,
+    "light_salmon1": 216,
+    "light_pink1": 217,
+    "pink1": 218,
+    "plum1": 219,
+    "gold1": 220,
+    "navajo_white1": 223,
+    "misty_rose1": 224,
+    "thistle1": 225,
+    "yellow1": 226,
+    "light_goldenrod1": 227,
+    "khaki1": 228,
+    "wheat1": 229,
+    "cornsilk1": 230,
+    "grey100": 231,
+    "gray100": 231,
+    "grey3": 232,
+    "gray3": 232,
+    "grey7": 233,
+    "gray7": 233,
+    "grey11": 234,
+    "gray11": 234,
+    "grey15": 235,
+    "gray15": 235,
+    "grey19": 236,
+    "gray19": 236,
+    "grey23": 237,
+    "gray23": 237,
+    "grey27": 238,
+    "gray27": 238,
+    "grey30": 239,
+    "gray30": 239,
+    "grey35": 240,
+    "gray35": 240,
+    "grey39": 241,
+    "gray39": 241,
+    "grey42": 242,
+    "gray42": 242,
+    "grey46": 243,
+    "gray46": 243,
+    "grey50": 244,
+    "gray50": 244,
+    "grey54": 245,
+    "gray54": 245,
+    "grey58": 246,
+    "gray58": 246,
+    "grey62": 247,
+    "gray62": 247,
+    "grey66": 248,
+    "gray66": 248,
+    "grey70": 249,
+    "gray70": 249,
+    "grey74": 250,
+    "gray74": 250,
+    "grey78": 251,
+    "gray78": 251,
+    "grey82": 252,
+    "gray82": 252,
+    "grey85": 253,
+    "gray85": 253,
+    "grey89": 254,
+    "gray89": 254,
+    "grey93": 255,
+    "gray93": 255,
+}
+
+
+class ColorParseError(Exception):
+    """The color could not be parsed."""
+
+
+RE_COLOR = re.compile(
+    r"""^
+\#([0-9a-f]{6})$|
+color\(([0-9]{1,3})\)$|
+rgb\(([\d\s,]+)\)$
+""",
+    re.VERBOSE,
+)
+
+
+@rich_repr
+class Color(NamedTuple):
+    """Terminal color definition."""
+
+    name: str
+    """The name of the color (typically the input to Color.parse)."""
+    type: ColorType
+    """The type of the color."""
+    number: Optional[int] = None
+    """The color number, if a standard color, or None."""
+    triplet: Optional[ColorTriplet] = None
+    """A triplet of color components, if an RGB color."""
+
+    def __rich__(self) -> "Text":
+        """Displays the actual color if Rich printed."""
+        from .style import Style
+        from .text import Text
+
+        return Text.assemble(
+            f"<color {self.name!r} ({self.type.name.lower()})",
+            ("⬤", Style(color=self)),
+            " >",
+        )
+
+    def __rich_repr__(self) -> Result:
+        yield self.name
+        yield self.type
+        yield "number", self.number, None
+        yield "triplet", self.triplet, None
+
+    @property
+    def system(self) -> ColorSystem:
+        """Get the native color system for this color."""
+        if self.type == ColorType.DEFAULT:
+            return ColorSystem.STANDARD
+        return ColorSystem(int(self.type))
+
+    @property
+    def is_system_defined(self) -> bool:
+        """Check if the color is ultimately defined by the system."""
+        return self.system not in (ColorSystem.EIGHT_BIT, ColorSystem.TRUECOLOR)
+
+    @property
+    def is_default(self) -> bool:
+        """Check if the color is a default color."""
+        return self.type == ColorType.DEFAULT
+
+    def get_truecolor(
+        self, theme: Optional["TerminalTheme"] = None, foreground: bool = True
+    ) -> ColorTriplet:
+        """Get an equivalent color triplet for this color.
+
+        Args:
+            theme (TerminalTheme, optional): Optional terminal theme, or None to use default. Defaults to None.
+            foreground (bool, optional): True for a foreground color, or False for background. Defaults to True.
+
+        Returns:
+            ColorTriplet: A color triplet containing RGB components.
+        """
+
+        if theme is None:
+            theme = DEFAULT_TERMINAL_THEME
+        if self.type == ColorType.TRUECOLOR:
+            assert self.triplet is not None
+            return self.triplet
+        elif self.type == ColorType.EIGHT_BIT:
+            assert self.number is not None
+            return EIGHT_BIT_PALETTE[self.number]
+        elif self.type == ColorType.STANDARD:
+            assert self.number is not None
+            return theme.ansi_colors[self.number]
+        elif self.type == ColorType.WINDOWS:
+            assert self.number is not None
+            return WINDOWS_PALETTE[self.number]
+        else:  # self.type == ColorType.DEFAULT:
+            assert self.number is None
+            return theme.foreground_color if foreground else theme.background_color
+
+    @classmethod
+    def from_ansi(cls, number: int) -> "Color":
+        """Create a Color number from it's 8-bit ansi number.
+
+        Args:
+            number (int): A number between 0-255 inclusive.
+
+        Returns:
+            Color: A new Color instance.
+        """
+        return cls(
+            name=f"color({number})",
+            type=(ColorType.STANDARD if number < 16 else ColorType.EIGHT_BIT),
+            number=number,
+        )
+
+    @classmethod
+    def from_triplet(cls, triplet: "ColorTriplet") -> "Color":
+        """Create a truecolor RGB color from a triplet of values.
+
+        Args:
+            triplet (ColorTriplet): A color triplet containing red, green and blue components.
+
+        Returns:
+            Color: A new color object.
+        """
+        return cls(name=triplet.hex, type=ColorType.TRUECOLOR, triplet=triplet)
+
+    @classmethod
+    def from_rgb(cls, red: float, green: float, blue: float) -> "Color":
+        """Create a truecolor from three color components in the range(0->255).
+
+        Args:
+            red (float): Red component in range 0-255.
+            green (float): Green component in range 0-255.
+            blue (float): Blue component in range 0-255.
+
+        Returns:
+            Color: A new color object.
+        """
+        return cls.from_triplet(ColorTriplet(int(red), int(green), int(blue)))
+
+    @classmethod
+    def default(cls) -> "Color":
+        """Get a Color instance representing the default color.
+
+        Returns:
+            Color: Default color.
+        """
+        return cls(name="default", type=ColorType.DEFAULT)
+
+    @classmethod
+    @lru_cache(maxsize=1024)
+    def parse(cls, color: str) -> "Color":
+        """Parse a color definition."""
+        original_color = color
+        color = color.lower().strip()
+
+        if color == "default":
+            return cls(color, type=ColorType.DEFAULT)
+
+        color_number = ANSI_COLOR_NAMES.get(color)
+        if color_number is not None:
+            return cls(
+                color,
+                type=(ColorType.STANDARD if color_number < 16 else ColorType.EIGHT_BIT),
+                number=color_number,
+            )
+
+        color_match = RE_COLOR.match(color)
+        if color_match is None:
+            raise ColorParseError(f"{original_color!r} is not a valid color")
+
+        color_24, color_8, color_rgb = color_match.groups()
+        if color_24:
+            triplet = ColorTriplet(
+                int(color_24[0:2], 16), int(color_24[2:4], 16), int(color_24[4:6], 16)
+            )
+            return cls(color, ColorType.TRUECOLOR, triplet=triplet)
+
+        elif color_8:
+            number = int(color_8)
+            if number > 255:
+                raise ColorParseError(f"color number must be <= 255 in {color!r}")
+            return cls(
+                color,
+                type=(ColorType.STANDARD if number < 16 else ColorType.EIGHT_BIT),
+                number=number,
+            )
+
+        else:  #  color_rgb:
+            components = color_rgb.split(",")
+            if len(components) != 3:
+                raise ColorParseError(
+                    f"expected three components in {original_color!r}"
+                )
+            red, green, blue = components
+            triplet = ColorTriplet(int(red), int(green), int(blue))
+            if not all(component <= 255 for component in triplet):
+                raise ColorParseError(
+                    f"color components must be <= 255 in {original_color!r}"
+                )
+            return cls(color, ColorType.TRUECOLOR, triplet=triplet)
+
+    @lru_cache(maxsize=1024)
+    def get_ansi_codes(self, foreground: bool = True) -> Tuple[str, ...]:
+        """Get the ANSI escape codes for this color."""
+        _type = self.type
+        if _type == ColorType.DEFAULT:
+            return ("39" if foreground else "49",)
+
+        elif _type == ColorType.WINDOWS:
+            number = self.number
+            assert number is not None
+            fore, back = (30, 40) if number < 8 else (82, 92)
+            return (str(fore + number if foreground else back + number),)
+
+        elif _type == ColorType.STANDARD:
+            number = self.number
+            assert number is not None
+            fore, back = (30, 40) if number < 8 else (82, 92)
+            return (str(fore + number if foreground else back + number),)
+
+        elif _type == ColorType.EIGHT_BIT:
+            assert self.number is not None
+            return ("38" if foreground else "48", "5", str(self.number))
+
+        else:  # self.standard == ColorStandard.TRUECOLOR:
+            assert self.triplet is not None
+            red, green, blue = self.triplet
+            return ("38" if foreground else "48", "2", str(red), str(green), str(blue))
+
+    @lru_cache(maxsize=1024)
+    def downgrade(self, system: ColorSystem) -> "Color":
+        """Downgrade a color system to a system with fewer colors."""
+
+        if self.type in (ColorType.DEFAULT, system):
+            return self
+        # Convert to 8-bit color from truecolor color
+        if system == ColorSystem.EIGHT_BIT and self.system == ColorSystem.TRUECOLOR:
+            assert self.triplet is not None
+            _h, l, s = rgb_to_hls(*self.triplet.normalized)
+            # If saturation is under 15% assume it is grayscale
+            if s < 0.15:
+                gray = round(l * 25.0)
+                if gray == 0:
+                    color_number = 16
+                elif gray == 25:
+                    color_number = 231
+                else:
+                    color_number = 231 + gray
+                return Color(self.name, ColorType.EIGHT_BIT, number=color_number)
+
+            red, green, blue = self.triplet
+            six_red = red / 95 if red < 95 else 1 + (red - 95) / 40
+            six_green = green / 95 if green < 95 else 1 + (green - 95) / 40
+            six_blue = blue / 95 if blue < 95 else 1 + (blue - 95) / 40
+
+            color_number = (
+                16 + 36 * round(six_red) + 6 * round(six_green) + round(six_blue)
+            )
+            return Color(self.name, ColorType.EIGHT_BIT, number=color_number)
+
+        # Convert to standard from truecolor or 8-bit
+        elif system == ColorSystem.STANDARD:
+            if self.system == ColorSystem.TRUECOLOR:
+                assert self.triplet is not None
+                triplet = self.triplet
+            else:  # self.system == ColorSystem.EIGHT_BIT
+                assert self.number is not None
+                triplet = ColorTriplet(*EIGHT_BIT_PALETTE[self.number])
+
+            color_number = STANDARD_PALETTE.match(triplet)
+            return Color(self.name, ColorType.STANDARD, number=color_number)
+
+        elif system == ColorSystem.WINDOWS:
+            if self.system == ColorSystem.TRUECOLOR:
+                assert self.triplet is not None
+                triplet = self.triplet
+            else:  # self.system == ColorSystem.EIGHT_BIT
+                assert self.number is not None
+                if self.number < 16:
+                    return Color(self.name, ColorType.WINDOWS, number=self.number)
+                triplet = ColorTriplet(*EIGHT_BIT_PALETTE[self.number])
+
+            color_number = WINDOWS_PALETTE.match(triplet)
+            return Color(self.name, ColorType.WINDOWS, number=color_number)
+
+        return self
+
+
+def parse_rgb_hex(hex_color: str) -> ColorTriplet:
+    """Parse six hex characters in to RGB triplet."""
+    assert len(hex_color) == 6, "must be 6 characters"
+    color = ColorTriplet(
+        int(hex_color[0:2], 16), int(hex_color[2:4], 16), int(hex_color[4:6], 16)
+    )
+    return color
+
+
+def blend_rgb(
+    color1: ColorTriplet, color2: ColorTriplet, cross_fade: float = 0.5
+) -> ColorTriplet:
+    """Blend one RGB color in to another."""
+    r1, g1, b1 = color1
+    r2, g2, b2 = color2
+    new_color = ColorTriplet(
+        int(r1 + (r2 - r1) * cross_fade),
+        int(g1 + (g2 - g1) * cross_fade),
+        int(b1 + (b2 - b1) * cross_fade),
+    )
+    return new_color
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from .console import Console
+    from .table import Table
+    from .text import Text
+
+    console = Console()
+
+    table = Table(show_footer=False, show_edge=True)
+    table.add_column("Color", width=10, overflow="ellipsis")
+    table.add_column("Number", justify="right", style="yellow")
+    table.add_column("Name", style="green")
+    table.add_column("Hex", style="blue")
+    table.add_column("RGB", style="magenta")
+
+    colors = sorted((v, k) for k, v in ANSI_COLOR_NAMES.items())
+    for color_number, name in colors:
+        if "grey" in name:
+            continue
+        color_cell = Text(" " * 10, style=f"on {name}")
+        if color_number < 16:
+            table.add_row(color_cell, f"{color_number}", Text(f'"{name}"'))
+        else:
+            color = EIGHT_BIT_PALETTE[color_number]  # type: ignore[has-type]
+            table.add_row(
+                color_cell, str(color_number), Text(f'"{name}"'), color.hex, color.rgb
+            )
+
+    console.print(table)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color_triplet.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color_triplet.py
new file mode 100644
index 00000000..02cab328
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/color_triplet.py
@@ -0,0 +1,38 @@
+from typing import NamedTuple, Tuple
+
+
+class ColorTriplet(NamedTuple):
+    """The red, green, and blue components of a color."""
+
+    red: int
+    """Red component in 0 to 255 range."""
+    green: int
+    """Green component in 0 to 255 range."""
+    blue: int
+    """Blue component in 0 to 255 range."""
+
+    @property
+    def hex(self) -> str:
+        """get the color triplet in CSS style."""
+        red, green, blue = self
+        return f"#{red:02x}{green:02x}{blue:02x}"
+
+    @property
+    def rgb(self) -> str:
+        """The color in RGB format.
+
+        Returns:
+            str: An rgb color, e.g. ``"rgb(100,23,255)"``.
+        """
+        red, green, blue = self
+        return f"rgb({red},{green},{blue})"
+
+    @property
+    def normalized(self) -> Tuple[float, float, float]:
+        """Convert components into floats between 0 and 1.
+
+        Returns:
+            Tuple[float, float, float]: A tuple of three normalized colour components.
+        """
+        red, green, blue = self
+        return red / 255.0, green / 255.0, blue / 255.0
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/columns.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/columns.py
new file mode 100644
index 00000000..669a3a70
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/columns.py
@@ -0,0 +1,187 @@
+from collections import defaultdict
+from itertools import chain
+from operator import itemgetter
+from typing import Dict, Iterable, List, Optional, Tuple
+
+from .align import Align, AlignMethod
+from .console import Console, ConsoleOptions, RenderableType, RenderResult
+from .constrain import Constrain
+from .measure import Measurement
+from .padding import Padding, PaddingDimensions
+from .table import Table
+from .text import TextType
+from .jupyter import JupyterMixin
+
+
+class Columns(JupyterMixin):
+    """Display renderables in neat columns.
+
+    Args:
+        renderables (Iterable[RenderableType]): Any number of Rich renderables (including str).
+        width (int, optional): The desired width of the columns, or None to auto detect. Defaults to None.
+        padding (PaddingDimensions, optional): Optional padding around cells. Defaults to (0, 1).
+        expand (bool, optional): Expand columns to full width. Defaults to False.
+        equal (bool, optional): Arrange in to equal sized columns. Defaults to False.
+        column_first (bool, optional): Align items from top to bottom (rather than left to right). Defaults to False.
+        right_to_left (bool, optional): Start column from right hand side. Defaults to False.
+        align (str, optional): Align value ("left", "right", or "center") or None for default. Defaults to None.
+        title (TextType, optional): Optional title for Columns.
+    """
+
+    def __init__(
+        self,
+        renderables: Optional[Iterable[RenderableType]] = None,
+        padding: PaddingDimensions = (0, 1),
+        *,
+        width: Optional[int] = None,
+        expand: bool = False,
+        equal: bool = False,
+        column_first: bool = False,
+        right_to_left: bool = False,
+        align: Optional[AlignMethod] = None,
+        title: Optional[TextType] = None,
+    ) -> None:
+        self.renderables = list(renderables or [])
+        self.width = width
+        self.padding = padding
+        self.expand = expand
+        self.equal = equal
+        self.column_first = column_first
+        self.right_to_left = right_to_left
+        self.align: Optional[AlignMethod] = align
+        self.title = title
+
+    def add_renderable(self, renderable: RenderableType) -> None:
+        """Add a renderable to the columns.
+
+        Args:
+            renderable (RenderableType): Any renderable object.
+        """
+        self.renderables.append(renderable)
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        render_str = console.render_str
+        renderables = [
+            render_str(renderable) if isinstance(renderable, str) else renderable
+            for renderable in self.renderables
+        ]
+        if not renderables:
+            return
+        _top, right, _bottom, left = Padding.unpack(self.padding)
+        width_padding = max(left, right)
+        max_width = options.max_width
+        widths: Dict[int, int] = defaultdict(int)
+        column_count = len(renderables)
+
+        get_measurement = Measurement.get
+        renderable_widths = [
+            get_measurement(console, options, renderable).maximum
+            for renderable in renderables
+        ]
+        if self.equal:
+            renderable_widths = [max(renderable_widths)] * len(renderable_widths)
+
+        def iter_renderables(
+            column_count: int,
+        ) -> Iterable[Tuple[int, Optional[RenderableType]]]:
+            item_count = len(renderables)
+            if self.column_first:
+                width_renderables = list(zip(renderable_widths, renderables))
+
+                column_lengths: List[int] = [item_count // column_count] * column_count
+                for col_no in range(item_count % column_count):
+                    column_lengths[col_no] += 1
+
+                row_count = (item_count + column_count - 1) // column_count
+                cells = [[-1] * column_count for _ in range(row_count)]
+                row = col = 0
+                for index in range(item_count):
+                    cells[row][col] = index
+                    column_lengths[col] -= 1
+                    if column_lengths[col]:
+                        row += 1
+                    else:
+                        col += 1
+                        row = 0
+                for index in chain.from_iterable(cells):
+                    if index == -1:
+                        break
+                    yield width_renderables[index]
+            else:
+                yield from zip(renderable_widths, renderables)
+            # Pad odd elements with spaces
+            if item_count % column_count:
+                for _ in range(column_count - (item_count % column_count)):
+                    yield 0, None
+
+        table = Table.grid(padding=self.padding, collapse_padding=True, pad_edge=False)
+        table.expand = self.expand
+        table.title = self.title
+
+        if self.width is not None:
+            column_count = (max_width) // (self.width + width_padding)
+            for _ in range(column_count):
+                table.add_column(width=self.width)
+        else:
+            while column_count > 1:
+                widths.clear()
+                column_no = 0
+                for renderable_width, _ in iter_renderables(column_count):
+                    widths[column_no] = max(widths[column_no], renderable_width)
+                    total_width = sum(widths.values()) + width_padding * (
+                        len(widths) - 1
+                    )
+                    if total_width > max_width:
+                        column_count = len(widths) - 1
+                        break
+                    else:
+                        column_no = (column_no + 1) % column_count
+                else:
+                    break
+
+        get_renderable = itemgetter(1)
+        _renderables = [
+            get_renderable(_renderable)
+            for _renderable in iter_renderables(column_count)
+        ]
+        if self.equal:
+            _renderables = [
+                None
+                if renderable is None
+                else Constrain(renderable, renderable_widths[0])
+                for renderable in _renderables
+            ]
+        if self.align:
+            align = self.align
+            _Align = Align
+            _renderables = [
+                None if renderable is None else _Align(renderable, align)
+                for renderable in _renderables
+            ]
+
+        right_to_left = self.right_to_left
+        add_row = table.add_row
+        for start in range(0, len(_renderables), column_count):
+            row = _renderables[start : start + column_count]
+            if right_to_left:
+                row = row[::-1]
+            add_row(*row)
+        yield table
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import os
+
+    console = Console()
+
+    files = [f"{i} {s}" for i, s in enumerate(sorted(os.listdir()))]
+    columns = Columns(files, padding=(0, 1), expand=False, equal=False)
+    console.print(columns)
+    console.rule()
+    columns.column_first = True
+    console.print(columns)
+    columns.right_to_left = True
+    console.rule()
+    console.print(columns)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/console.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/console.py
new file mode 100644
index 00000000..57288454
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/console.py
@@ -0,0 +1,2661 @@
+import inspect
+import os
+import sys
+import threading
+import zlib
+from abc import ABC, abstractmethod
+from dataclasses import dataclass, field
+from datetime import datetime
+from functools import wraps
+from getpass import getpass
+from html import escape
+from inspect import isclass
+from itertools import islice
+from math import ceil
+from time import monotonic
+from types import FrameType, ModuleType, TracebackType
+from typing import (
+    IO,
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    Mapping,
+    NamedTuple,
+    Optional,
+    TextIO,
+    Tuple,
+    Type,
+    Union,
+    cast,
+)
+
+from pip._vendor.rich._null_file import NULL_FILE
+
+if sys.version_info >= (3, 8):
+    from typing import Literal, Protocol, runtime_checkable
+else:
+    from pip._vendor.typing_extensions import (
+        Literal,
+        Protocol,
+        runtime_checkable,
+    )  # pragma: no cover
+
+from . import errors, themes
+from ._emoji_replace import _emoji_replace
+from ._export_format import CONSOLE_HTML_FORMAT, CONSOLE_SVG_FORMAT
+from ._fileno import get_fileno
+from ._log_render import FormatTimeCallable, LogRender
+from .align import Align, AlignMethod
+from .color import ColorSystem, blend_rgb
+from .control import Control
+from .emoji import EmojiVariant
+from .highlighter import NullHighlighter, ReprHighlighter
+from .markup import render as render_markup
+from .measure import Measurement, measure_renderables
+from .pager import Pager, SystemPager
+from .pretty import Pretty, is_expandable
+from .protocol import rich_cast
+from .region import Region
+from .scope import render_scope
+from .screen import Screen
+from .segment import Segment
+from .style import Style, StyleType
+from .styled import Styled
+from .terminal_theme import DEFAULT_TERMINAL_THEME, SVG_EXPORT_THEME, TerminalTheme
+from .text import Text, TextType
+from .theme import Theme, ThemeStack
+
+if TYPE_CHECKING:
+    from ._windows import WindowsConsoleFeatures
+    from .live import Live
+    from .status import Status
+
+JUPYTER_DEFAULT_COLUMNS = 115
+JUPYTER_DEFAULT_LINES = 100
+WINDOWS = sys.platform == "win32"
+
+HighlighterType = Callable[[Union[str, "Text"]], "Text"]
+JustifyMethod = Literal["default", "left", "center", "right", "full"]
+OverflowMethod = Literal["fold", "crop", "ellipsis", "ignore"]
+
+
+class NoChange:
+    pass
+
+
+NO_CHANGE = NoChange()
+
+try:
+    _STDIN_FILENO = sys.__stdin__.fileno()  # type: ignore[union-attr]
+except Exception:
+    _STDIN_FILENO = 0
+try:
+    _STDOUT_FILENO = sys.__stdout__.fileno()  # type: ignore[union-attr]
+except Exception:
+    _STDOUT_FILENO = 1
+try:
+    _STDERR_FILENO = sys.__stderr__.fileno()  # type: ignore[union-attr]
+except Exception:
+    _STDERR_FILENO = 2
+
+_STD_STREAMS = (_STDIN_FILENO, _STDOUT_FILENO, _STDERR_FILENO)
+_STD_STREAMS_OUTPUT = (_STDOUT_FILENO, _STDERR_FILENO)
+
+
+_TERM_COLORS = {
+    "kitty": ColorSystem.EIGHT_BIT,
+    "256color": ColorSystem.EIGHT_BIT,
+    "16color": ColorSystem.STANDARD,
+}
+
+
+class ConsoleDimensions(NamedTuple):
+    """Size of the terminal."""
+
+    width: int
+    """The width of the console in 'cells'."""
+    height: int
+    """The height of the console in lines."""
+
+
+@dataclass
+class ConsoleOptions:
+    """Options for __rich_console__ method."""
+
+    size: ConsoleDimensions
+    """Size of console."""
+    legacy_windows: bool
+    """legacy_windows: flag for legacy windows."""
+    min_width: int
+    """Minimum width of renderable."""
+    max_width: int
+    """Maximum width of renderable."""
+    is_terminal: bool
+    """True if the target is a terminal, otherwise False."""
+    encoding: str
+    """Encoding of terminal."""
+    max_height: int
+    """Height of container (starts as terminal)"""
+    justify: Optional[JustifyMethod] = None
+    """Justify value override for renderable."""
+    overflow: Optional[OverflowMethod] = None
+    """Overflow value override for renderable."""
+    no_wrap: Optional[bool] = False
+    """Disable wrapping for text."""
+    highlight: Optional[bool] = None
+    """Highlight override for render_str."""
+    markup: Optional[bool] = None
+    """Enable markup when rendering strings."""
+    height: Optional[int] = None
+
+    @property
+    def ascii_only(self) -> bool:
+        """Check if renderables should use ascii only."""
+        return not self.encoding.startswith("utf")
+
+    def copy(self) -> "ConsoleOptions":
+        """Return a copy of the options.
+
+        Returns:
+            ConsoleOptions: a copy of self.
+        """
+        options: ConsoleOptions = ConsoleOptions.__new__(ConsoleOptions)
+        options.__dict__ = self.__dict__.copy()
+        return options
+
+    def update(
+        self,
+        *,
+        width: Union[int, NoChange] = NO_CHANGE,
+        min_width: Union[int, NoChange] = NO_CHANGE,
+        max_width: Union[int, NoChange] = NO_CHANGE,
+        justify: Union[Optional[JustifyMethod], NoChange] = NO_CHANGE,
+        overflow: Union[Optional[OverflowMethod], NoChange] = NO_CHANGE,
+        no_wrap: Union[Optional[bool], NoChange] = NO_CHANGE,
+        highlight: Union[Optional[bool], NoChange] = NO_CHANGE,
+        markup: Union[Optional[bool], NoChange] = NO_CHANGE,
+        height: Union[Optional[int], NoChange] = NO_CHANGE,
+    ) -> "ConsoleOptions":
+        """Update values, return a copy."""
+        options = self.copy()
+        if not isinstance(width, NoChange):
+            options.min_width = options.max_width = max(0, width)
+        if not isinstance(min_width, NoChange):
+            options.min_width = min_width
+        if not isinstance(max_width, NoChange):
+            options.max_width = max_width
+        if not isinstance(justify, NoChange):
+            options.justify = justify
+        if not isinstance(overflow, NoChange):
+            options.overflow = overflow
+        if not isinstance(no_wrap, NoChange):
+            options.no_wrap = no_wrap
+        if not isinstance(highlight, NoChange):
+            options.highlight = highlight
+        if not isinstance(markup, NoChange):
+            options.markup = markup
+        if not isinstance(height, NoChange):
+            if height is not None:
+                options.max_height = height
+            options.height = None if height is None else max(0, height)
+        return options
+
+    def update_width(self, width: int) -> "ConsoleOptions":
+        """Update just the width, return a copy.
+
+        Args:
+            width (int): New width (sets both min_width and max_width)
+
+        Returns:
+            ~ConsoleOptions: New console options instance.
+        """
+        options = self.copy()
+        options.min_width = options.max_width = max(0, width)
+        return options
+
+    def update_height(self, height: int) -> "ConsoleOptions":
+        """Update the height, and return a copy.
+
+        Args:
+            height (int): New height
+
+        Returns:
+            ~ConsoleOptions: New Console options instance.
+        """
+        options = self.copy()
+        options.max_height = options.height = height
+        return options
+
+    def reset_height(self) -> "ConsoleOptions":
+        """Return a copy of the options with height set to ``None``.
+
+        Returns:
+            ~ConsoleOptions: New console options instance.
+        """
+        options = self.copy()
+        options.height = None
+        return options
+
+    def update_dimensions(self, width: int, height: int) -> "ConsoleOptions":
+        """Update the width and height, and return a copy.
+
+        Args:
+            width (int): New width (sets both min_width and max_width).
+            height (int): New height.
+
+        Returns:
+            ~ConsoleOptions: New console options instance.
+        """
+        options = self.copy()
+        options.min_width = options.max_width = max(0, width)
+        options.height = options.max_height = height
+        return options
+
+
+@runtime_checkable
+class RichCast(Protocol):
+    """An object that may be 'cast' to a console renderable."""
+
+    def __rich__(
+        self,
+    ) -> Union["ConsoleRenderable", "RichCast", str]:  # pragma: no cover
+        ...
+
+
+@runtime_checkable
+class ConsoleRenderable(Protocol):
+    """An object that supports the console protocol."""
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":  # pragma: no cover
+        ...
+
+
+# A type that may be rendered by Console.
+RenderableType = Union[ConsoleRenderable, RichCast, str]
+"""A string or any object that may be rendered by Rich."""
+
+# The result of calling a __rich_console__ method.
+RenderResult = Iterable[Union[RenderableType, Segment]]
+
+_null_highlighter = NullHighlighter()
+
+
+class CaptureError(Exception):
+    """An error in the Capture context manager."""
+
+
+class NewLine:
+    """A renderable to generate new line(s)"""
+
+    def __init__(self, count: int = 1) -> None:
+        self.count = count
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Iterable[Segment]:
+        yield Segment("\n" * self.count)
+
+
+class ScreenUpdate:
+    """Render a list of lines at a given offset."""
+
+    def __init__(self, lines: List[List[Segment]], x: int, y: int) -> None:
+        self._lines = lines
+        self.x = x
+        self.y = y
+
+    def __rich_console__(
+        self, console: "Console", options: ConsoleOptions
+    ) -> RenderResult:
+        x = self.x
+        move_to = Control.move_to
+        for offset, line in enumerate(self._lines, self.y):
+            yield move_to(x, offset)
+            yield from line
+
+
+class Capture:
+    """Context manager to capture the result of printing to the console.
+    See :meth:`~rich.console.Console.capture` for how to use.
+
+    Args:
+        console (Console): A console instance to capture output.
+    """
+
+    def __init__(self, console: "Console") -> None:
+        self._console = console
+        self._result: Optional[str] = None
+
+    def __enter__(self) -> "Capture":
+        self._console.begin_capture()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self._result = self._console.end_capture()
+
+    def get(self) -> str:
+        """Get the result of the capture."""
+        if self._result is None:
+            raise CaptureError(
+                "Capture result is not available until context manager exits."
+            )
+        return self._result
+
+
+class ThemeContext:
+    """A context manager to use a temporary theme. See :meth:`~rich.console.Console.use_theme` for usage."""
+
+    def __init__(self, console: "Console", theme: Theme, inherit: bool = True) -> None:
+        self.console = console
+        self.theme = theme
+        self.inherit = inherit
+
+    def __enter__(self) -> "ThemeContext":
+        self.console.push_theme(self.theme)
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.console.pop_theme()
+
+
+class PagerContext:
+    """A context manager that 'pages' content. See :meth:`~rich.console.Console.pager` for usage."""
+
+    def __init__(
+        self,
+        console: "Console",
+        pager: Optional[Pager] = None,
+        styles: bool = False,
+        links: bool = False,
+    ) -> None:
+        self._console = console
+        self.pager = SystemPager() if pager is None else pager
+        self.styles = styles
+        self.links = links
+
+    def __enter__(self) -> "PagerContext":
+        self._console._enter_buffer()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        if exc_type is None:
+            with self._console._lock:
+                buffer: List[Segment] = self._console._buffer[:]
+                del self._console._buffer[:]
+                segments: Iterable[Segment] = buffer
+                if not self.styles:
+                    segments = Segment.strip_styles(segments)
+                elif not self.links:
+                    segments = Segment.strip_links(segments)
+                content = self._console._render_buffer(segments)
+            self.pager.show(content)
+        self._console._exit_buffer()
+
+
+class ScreenContext:
+    """A context manager that enables an alternative screen. See :meth:`~rich.console.Console.screen` for usage."""
+
+    def __init__(
+        self, console: "Console", hide_cursor: bool, style: StyleType = ""
+    ) -> None:
+        self.console = console
+        self.hide_cursor = hide_cursor
+        self.screen = Screen(style=style)
+        self._changed = False
+
+    def update(
+        self, *renderables: RenderableType, style: Optional[StyleType] = None
+    ) -> None:
+        """Update the screen.
+
+        Args:
+            renderable (RenderableType, optional): Optional renderable to replace current renderable,
+                or None for no change. Defaults to None.
+            style: (Style, optional): Replacement style, or None for no change. Defaults to None.
+        """
+        if renderables:
+            self.screen.renderable = (
+                Group(*renderables) if len(renderables) > 1 else renderables[0]
+            )
+        if style is not None:
+            self.screen.style = style
+        self.console.print(self.screen, end="")
+
+    def __enter__(self) -> "ScreenContext":
+        self._changed = self.console.set_alt_screen(True)
+        if self._changed and self.hide_cursor:
+            self.console.show_cursor(False)
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        if self._changed:
+            self.console.set_alt_screen(False)
+            if self.hide_cursor:
+                self.console.show_cursor(True)
+
+
+class Group:
+    """Takes a group of renderables and returns a renderable object that renders the group.
+
+    Args:
+        renderables (Iterable[RenderableType]): An iterable of renderable objects.
+        fit (bool, optional): Fit dimension of group to contents, or fill available space. Defaults to True.
+    """
+
+    def __init__(self, *renderables: "RenderableType", fit: bool = True) -> None:
+        self._renderables = renderables
+        self.fit = fit
+        self._render: Optional[List[RenderableType]] = None
+
+    @property
+    def renderables(self) -> List["RenderableType"]:
+        if self._render is None:
+            self._render = list(self._renderables)
+        return self._render
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        if self.fit:
+            return measure_renderables(console, options, self.renderables)
+        else:
+            return Measurement(options.max_width, options.max_width)
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> RenderResult:
+        yield from self.renderables
+
+
+def group(fit: bool = True) -> Callable[..., Callable[..., Group]]:
+    """A decorator that turns an iterable of renderables in to a group.
+
+    Args:
+        fit (bool, optional): Fit dimension of group to contents, or fill available space. Defaults to True.
+    """
+
+    def decorator(
+        method: Callable[..., Iterable[RenderableType]]
+    ) -> Callable[..., Group]:
+        """Convert a method that returns an iterable of renderables in to a Group."""
+
+        @wraps(method)
+        def _replace(*args: Any, **kwargs: Any) -> Group:
+            renderables = method(*args, **kwargs)
+            return Group(*renderables, fit=fit)
+
+        return _replace
+
+    return decorator
+
+
+def _is_jupyter() -> bool:  # pragma: no cover
+    """Check if we're running in a Jupyter notebook."""
+    try:
+        get_ipython  # type: ignore[name-defined]
+    except NameError:
+        return False
+    ipython = get_ipython()  # type: ignore[name-defined]
+    shell = ipython.__class__.__name__
+    if (
+        "google.colab" in str(ipython.__class__)
+        or os.getenv("DATABRICKS_RUNTIME_VERSION")
+        or shell == "ZMQInteractiveShell"
+    ):
+        return True  # Jupyter notebook or qtconsole
+    elif shell == "TerminalInteractiveShell":
+        return False  # Terminal running IPython
+    else:
+        return False  # Other type (?)
+
+
+COLOR_SYSTEMS = {
+    "standard": ColorSystem.STANDARD,
+    "256": ColorSystem.EIGHT_BIT,
+    "truecolor": ColorSystem.TRUECOLOR,
+    "windows": ColorSystem.WINDOWS,
+}
+
+_COLOR_SYSTEMS_NAMES = {system: name for name, system in COLOR_SYSTEMS.items()}
+
+
+@dataclass
+class ConsoleThreadLocals(threading.local):
+    """Thread local values for Console context."""
+
+    theme_stack: ThemeStack
+    buffer: List[Segment] = field(default_factory=list)
+    buffer_index: int = 0
+
+
+class RenderHook(ABC):
+    """Provides hooks in to the render process."""
+
+    @abstractmethod
+    def process_renderables(
+        self, renderables: List[ConsoleRenderable]
+    ) -> List[ConsoleRenderable]:
+        """Called with a list of objects to render.
+
+        This method can return a new list of renderables, or modify and return the same list.
+
+        Args:
+            renderables (List[ConsoleRenderable]): A number of renderable objects.
+
+        Returns:
+            List[ConsoleRenderable]: A replacement list of renderables.
+        """
+
+
+_windows_console_features: Optional["WindowsConsoleFeatures"] = None
+
+
+def get_windows_console_features() -> "WindowsConsoleFeatures":  # pragma: no cover
+    global _windows_console_features
+    if _windows_console_features is not None:
+        return _windows_console_features
+    from ._windows import get_windows_console_features
+
+    _windows_console_features = get_windows_console_features()
+    return _windows_console_features
+
+
+def detect_legacy_windows() -> bool:
+    """Detect legacy Windows."""
+    return WINDOWS and not get_windows_console_features().vt
+
+
+class Console:
+    """A high level console interface.
+
+    Args:
+        color_system (str, optional): The color system supported by your terminal,
+            either ``"standard"``, ``"256"`` or ``"truecolor"``. Leave as ``"auto"`` to autodetect.
+        force_terminal (Optional[bool], optional): Enable/disable terminal control codes, or None to auto-detect terminal. Defaults to None.
+        force_jupyter (Optional[bool], optional): Enable/disable Jupyter rendering, or None to auto-detect Jupyter. Defaults to None.
+        force_interactive (Optional[bool], optional): Enable/disable interactive mode, or None to auto detect. Defaults to None.
+        soft_wrap (Optional[bool], optional): Set soft wrap default on print method. Defaults to False.
+        theme (Theme, optional): An optional style theme object, or ``None`` for default theme.
+        stderr (bool, optional): Use stderr rather than stdout if ``file`` is not specified. Defaults to False.
+        file (IO, optional): A file object where the console should write to. Defaults to stdout.
+        quiet (bool, Optional): Boolean to suppress all output. Defaults to False.
+        width (int, optional): The width of the terminal. Leave as default to auto-detect width.
+        height (int, optional): The height of the terminal. Leave as default to auto-detect height.
+        style (StyleType, optional): Style to apply to all output, or None for no style. Defaults to None.
+        no_color (Optional[bool], optional): Enabled no color mode, or None to auto detect. Defaults to None.
+        tab_size (int, optional): Number of spaces used to replace a tab character. Defaults to 8.
+        record (bool, optional): Boolean to enable recording of terminal output,
+            required to call :meth:`export_html`, :meth:`export_svg`, and :meth:`export_text`. Defaults to False.
+        markup (bool, optional): Boolean to enable :ref:`console_markup`. Defaults to True.
+        emoji (bool, optional): Enable emoji code. Defaults to True.
+        emoji_variant (str, optional): Optional emoji variant, either "text" or "emoji". Defaults to None.
+        highlight (bool, optional): Enable automatic highlighting. Defaults to True.
+        log_time (bool, optional): Boolean to enable logging of time by :meth:`log` methods. Defaults to True.
+        log_path (bool, optional): Boolean to enable the logging of the caller by :meth:`log`. Defaults to True.
+        log_time_format (Union[str, TimeFormatterCallable], optional): If ``log_time`` is enabled, either string for strftime or callable that formats the time. Defaults to "[%X] ".
+        highlighter (HighlighterType, optional): Default highlighter.
+        legacy_windows (bool, optional): Enable legacy Windows mode, or ``None`` to auto detect. Defaults to ``None``.
+        safe_box (bool, optional): Restrict box options that don't render on legacy Windows.
+        get_datetime (Callable[[], datetime], optional): Callable that gets the current time as a datetime.datetime object (used by Console.log),
+            or None for datetime.now.
+        get_time (Callable[[], time], optional): Callable that gets the current time in seconds, default uses time.monotonic.
+    """
+
+    _environ: Mapping[str, str] = os.environ
+
+    def __init__(
+        self,
+        *,
+        color_system: Optional[
+            Literal["auto", "standard", "256", "truecolor", "windows"]
+        ] = "auto",
+        force_terminal: Optional[bool] = None,
+        force_jupyter: Optional[bool] = None,
+        force_interactive: Optional[bool] = None,
+        soft_wrap: bool = False,
+        theme: Optional[Theme] = None,
+        stderr: bool = False,
+        file: Optional[IO[str]] = None,
+        quiet: bool = False,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+        style: Optional[StyleType] = None,
+        no_color: Optional[bool] = None,
+        tab_size: int = 8,
+        record: bool = False,
+        markup: bool = True,
+        emoji: bool = True,
+        emoji_variant: Optional[EmojiVariant] = None,
+        highlight: bool = True,
+        log_time: bool = True,
+        log_path: bool = True,
+        log_time_format: Union[str, FormatTimeCallable] = "[%X]",
+        highlighter: Optional["HighlighterType"] = ReprHighlighter(),
+        legacy_windows: Optional[bool] = None,
+        safe_box: bool = True,
+        get_datetime: Optional[Callable[[], datetime]] = None,
+        get_time: Optional[Callable[[], float]] = None,
+        _environ: Optional[Mapping[str, str]] = None,
+    ):
+        # Copy of os.environ allows us to replace it for testing
+        if _environ is not None:
+            self._environ = _environ
+
+        self.is_jupyter = _is_jupyter() if force_jupyter is None else force_jupyter
+        if self.is_jupyter:
+            if width is None:
+                jupyter_columns = self._environ.get("JUPYTER_COLUMNS")
+                if jupyter_columns is not None and jupyter_columns.isdigit():
+                    width = int(jupyter_columns)
+                else:
+                    width = JUPYTER_DEFAULT_COLUMNS
+            if height is None:
+                jupyter_lines = self._environ.get("JUPYTER_LINES")
+                if jupyter_lines is not None and jupyter_lines.isdigit():
+                    height = int(jupyter_lines)
+                else:
+                    height = JUPYTER_DEFAULT_LINES
+
+        self.tab_size = tab_size
+        self.record = record
+        self._markup = markup
+        self._emoji = emoji
+        self._emoji_variant: Optional[EmojiVariant] = emoji_variant
+        self._highlight = highlight
+        self.legacy_windows: bool = (
+            (detect_legacy_windows() and not self.is_jupyter)
+            if legacy_windows is None
+            else legacy_windows
+        )
+
+        if width is None:
+            columns = self._environ.get("COLUMNS")
+            if columns is not None and columns.isdigit():
+                width = int(columns) - self.legacy_windows
+        if height is None:
+            lines = self._environ.get("LINES")
+            if lines is not None and lines.isdigit():
+                height = int(lines)
+
+        self.soft_wrap = soft_wrap
+        self._width = width
+        self._height = height
+
+        self._color_system: Optional[ColorSystem]
+
+        self._force_terminal = None
+        if force_terminal is not None:
+            self._force_terminal = force_terminal
+
+        self._file = file
+        self.quiet = quiet
+        self.stderr = stderr
+
+        if color_system is None:
+            self._color_system = None
+        elif color_system == "auto":
+            self._color_system = self._detect_color_system()
+        else:
+            self._color_system = COLOR_SYSTEMS[color_system]
+
+        self._lock = threading.RLock()
+        self._log_render = LogRender(
+            show_time=log_time,
+            show_path=log_path,
+            time_format=log_time_format,
+        )
+        self.highlighter: HighlighterType = highlighter or _null_highlighter
+        self.safe_box = safe_box
+        self.get_datetime = get_datetime or datetime.now
+        self.get_time = get_time or monotonic
+        self.style = style
+        self.no_color = (
+            no_color if no_color is not None else "NO_COLOR" in self._environ
+        )
+        self.is_interactive = (
+            (self.is_terminal and not self.is_dumb_terminal)
+            if force_interactive is None
+            else force_interactive
+        )
+
+        self._record_buffer_lock = threading.RLock()
+        self._thread_locals = ConsoleThreadLocals(
+            theme_stack=ThemeStack(themes.DEFAULT if theme is None else theme)
+        )
+        self._record_buffer: List[Segment] = []
+        self._render_hooks: List[RenderHook] = []
+        self._live: Optional["Live"] = None
+        self._is_alt_screen = False
+
+    def __repr__(self) -> str:
+        return f"<console width={self.width} {self._color_system!s}>"
+
+    @property
+    def file(self) -> IO[str]:
+        """Get the file object to write to."""
+        file = self._file or (sys.stderr if self.stderr else sys.stdout)
+        file = getattr(file, "rich_proxied_file", file)
+        if file is None:
+            file = NULL_FILE
+        return file
+
+    @file.setter
+    def file(self, new_file: IO[str]) -> None:
+        """Set a new file object."""
+        self._file = new_file
+
+    @property
+    def _buffer(self) -> List[Segment]:
+        """Get a thread local buffer."""
+        return self._thread_locals.buffer
+
+    @property
+    def _buffer_index(self) -> int:
+        """Get a thread local buffer."""
+        return self._thread_locals.buffer_index
+
+    @_buffer_index.setter
+    def _buffer_index(self, value: int) -> None:
+        self._thread_locals.buffer_index = value
+
+    @property
+    def _theme_stack(self) -> ThemeStack:
+        """Get the thread local theme stack."""
+        return self._thread_locals.theme_stack
+
+    def _detect_color_system(self) -> Optional[ColorSystem]:
+        """Detect color system from env vars."""
+        if self.is_jupyter:
+            return ColorSystem.TRUECOLOR
+        if not self.is_terminal or self.is_dumb_terminal:
+            return None
+        if WINDOWS:  # pragma: no cover
+            if self.legacy_windows:  # pragma: no cover
+                return ColorSystem.WINDOWS
+            windows_console_features = get_windows_console_features()
+            return (
+                ColorSystem.TRUECOLOR
+                if windows_console_features.truecolor
+                else ColorSystem.EIGHT_BIT
+            )
+        else:
+            color_term = self._environ.get("COLORTERM", "").strip().lower()
+            if color_term in ("truecolor", "24bit"):
+                return ColorSystem.TRUECOLOR
+            term = self._environ.get("TERM", "").strip().lower()
+            _term_name, _hyphen, colors = term.rpartition("-")
+            color_system = _TERM_COLORS.get(colors, ColorSystem.STANDARD)
+            return color_system
+
+    def _enter_buffer(self) -> None:
+        """Enter in to a buffer context, and buffer all output."""
+        self._buffer_index += 1
+
+    def _exit_buffer(self) -> None:
+        """Leave buffer context, and render content if required."""
+        self._buffer_index -= 1
+        self._check_buffer()
+
+    def set_live(self, live: "Live") -> None:
+        """Set Live instance. Used by Live context manager.
+
+        Args:
+            live (Live): Live instance using this Console.
+
+        Raises:
+            errors.LiveError: If this Console has a Live context currently active.
+        """
+        with self._lock:
+            if self._live is not None:
+                raise errors.LiveError("Only one live display may be active at once")
+            self._live = live
+
+    def clear_live(self) -> None:
+        """Clear the Live instance."""
+        with self._lock:
+            self._live = None
+
+    def push_render_hook(self, hook: RenderHook) -> None:
+        """Add a new render hook to the stack.
+
+        Args:
+            hook (RenderHook): Render hook instance.
+        """
+        with self._lock:
+            self._render_hooks.append(hook)
+
+    def pop_render_hook(self) -> None:
+        """Pop the last renderhook from the stack."""
+        with self._lock:
+            self._render_hooks.pop()
+
+    def __enter__(self) -> "Console":
+        """Own context manager to enter buffer context."""
+        self._enter_buffer()
+        return self
+
+    def __exit__(self, exc_type: Any, exc_value: Any, traceback: Any) -> None:
+        """Exit buffer context."""
+        self._exit_buffer()
+
+    def begin_capture(self) -> None:
+        """Begin capturing console output. Call :meth:`end_capture` to exit capture mode and return output."""
+        self._enter_buffer()
+
+    def end_capture(self) -> str:
+        """End capture mode and return captured string.
+
+        Returns:
+            str: Console output.
+        """
+        render_result = self._render_buffer(self._buffer)
+        del self._buffer[:]
+        self._exit_buffer()
+        return render_result
+
+    def push_theme(self, theme: Theme, *, inherit: bool = True) -> None:
+        """Push a new theme on to the top of the stack, replacing the styles from the previous theme.
+        Generally speaking, you should call :meth:`~rich.console.Console.use_theme` to get a context manager, rather
+        than calling this method directly.
+
+        Args:
+            theme (Theme): A theme instance.
+            inherit (bool, optional): Inherit existing styles. Defaults to True.
+        """
+        self._theme_stack.push_theme(theme, inherit=inherit)
+
+    def pop_theme(self) -> None:
+        """Remove theme from top of stack, restoring previous theme."""
+        self._theme_stack.pop_theme()
+
+    def use_theme(self, theme: Theme, *, inherit: bool = True) -> ThemeContext:
+        """Use a different theme for the duration of the context manager.
+
+        Args:
+            theme (Theme): Theme instance to user.
+            inherit (bool, optional): Inherit existing console styles. Defaults to True.
+
+        Returns:
+            ThemeContext: [description]
+        """
+        return ThemeContext(self, theme, inherit)
+
+    @property
+    def color_system(self) -> Optional[str]:
+        """Get color system string.
+
+        Returns:
+            Optional[str]: "standard", "256" or "truecolor".
+        """
+
+        if self._color_system is not None:
+            return _COLOR_SYSTEMS_NAMES[self._color_system]
+        else:
+            return None
+
+    @property
+    def encoding(self) -> str:
+        """Get the encoding of the console file, e.g. ``"utf-8"``.
+
+        Returns:
+            str: A standard encoding string.
+        """
+        return (getattr(self.file, "encoding", "utf-8") or "utf-8").lower()
+
+    @property
+    def is_terminal(self) -> bool:
+        """Check if the console is writing to a terminal.
+
+        Returns:
+            bool: True if the console writing to a device capable of
+            understanding terminal codes, otherwise False.
+        """
+        if self._force_terminal is not None:
+            return self._force_terminal
+
+        if hasattr(sys.stdin, "__module__") and sys.stdin.__module__.startswith(
+            "idlelib"
+        ):
+            # Return False for Idle which claims to be a tty but can't handle ansi codes
+            return False
+
+        if self.is_jupyter:
+            # return False for Jupyter, which may have FORCE_COLOR set
+            return False
+
+        # If FORCE_COLOR env var has any value at all, we assume a terminal.
+        force_color = self._environ.get("FORCE_COLOR")
+        if force_color is not None:
+            self._force_terminal = True
+            return True
+
+        isatty: Optional[Callable[[], bool]] = getattr(self.file, "isatty", None)
+        try:
+            return False if isatty is None else isatty()
+        except ValueError:
+            # in some situation (at the end of a pytest run for example) isatty() can raise
+            # ValueError: I/O operation on closed file
+            # return False because we aren't in a terminal anymore
+            return False
+
+    @property
+    def is_dumb_terminal(self) -> bool:
+        """Detect dumb terminal.
+
+        Returns:
+            bool: True if writing to a dumb terminal, otherwise False.
+
+        """
+        _term = self._environ.get("TERM", "")
+        is_dumb = _term.lower() in ("dumb", "unknown")
+        return self.is_terminal and is_dumb
+
+    @property
+    def options(self) -> ConsoleOptions:
+        """Get default console options."""
+        return ConsoleOptions(
+            max_height=self.size.height,
+            size=self.size,
+            legacy_windows=self.legacy_windows,
+            min_width=1,
+            max_width=self.width,
+            encoding=self.encoding,
+            is_terminal=self.is_terminal,
+        )
+
+    @property
+    def size(self) -> ConsoleDimensions:
+        """Get the size of the console.
+
+        Returns:
+            ConsoleDimensions: A named tuple containing the dimensions.
+        """
+
+        if self._width is not None and self._height is not None:
+            return ConsoleDimensions(self._width - self.legacy_windows, self._height)
+
+        if self.is_dumb_terminal:
+            return ConsoleDimensions(80, 25)
+
+        width: Optional[int] = None
+        height: Optional[int] = None
+
+        streams = _STD_STREAMS_OUTPUT if WINDOWS else _STD_STREAMS
+        for file_descriptor in streams:
+            try:
+                width, height = os.get_terminal_size(file_descriptor)
+            except (AttributeError, ValueError, OSError):  # Probably not a terminal
+                pass
+            else:
+                break
+
+        columns = self._environ.get("COLUMNS")
+        if columns is not None and columns.isdigit():
+            width = int(columns)
+        lines = self._environ.get("LINES")
+        if lines is not None and lines.isdigit():
+            height = int(lines)
+
+        # get_terminal_size can report 0, 0 if run from pseudo-terminal
+        width = width or 80
+        height = height or 25
+        return ConsoleDimensions(
+            width - self.legacy_windows if self._width is None else self._width,
+            height if self._height is None else self._height,
+        )
+
+    @size.setter
+    def size(self, new_size: Tuple[int, int]) -> None:
+        """Set a new size for the terminal.
+
+        Args:
+            new_size (Tuple[int, int]): New width and height.
+        """
+        width, height = new_size
+        self._width = width
+        self._height = height
+
+    @property
+    def width(self) -> int:
+        """Get the width of the console.
+
+        Returns:
+            int: The width (in characters) of the console.
+        """
+        return self.size.width
+
+    @width.setter
+    def width(self, width: int) -> None:
+        """Set width.
+
+        Args:
+            width (int): New width.
+        """
+        self._width = width
+
+    @property
+    def height(self) -> int:
+        """Get the height of the console.
+
+        Returns:
+            int: The height (in lines) of the console.
+        """
+        return self.size.height
+
+    @height.setter
+    def height(self, height: int) -> None:
+        """Set height.
+
+        Args:
+            height (int): new height.
+        """
+        self._height = height
+
+    def bell(self) -> None:
+        """Play a 'bell' sound (if supported by the terminal)."""
+        self.control(Control.bell())
+
+    def capture(self) -> Capture:
+        """A context manager to *capture* the result of print() or log() in a string,
+        rather than writing it to the console.
+
+        Example:
+            >>> from rich.console import Console
+            >>> console = Console()
+            >>> with console.capture() as capture:
+            ...     console.print("[bold magenta]Hello World[/]")
+            >>> print(capture.get())
+
+        Returns:
+            Capture: Context manager with disables writing to the terminal.
+        """
+        capture = Capture(self)
+        return capture
+
+    def pager(
+        self, pager: Optional[Pager] = None, styles: bool = False, links: bool = False
+    ) -> PagerContext:
+        """A context manager to display anything printed within a "pager". The pager application
+        is defined by the system and will typically support at least pressing a key to scroll.
+
+        Args:
+            pager (Pager, optional): A pager object, or None to use :class:`~rich.pager.SystemPager`. Defaults to None.
+            styles (bool, optional): Show styles in pager. Defaults to False.
+            links (bool, optional): Show links in pager. Defaults to False.
+
+        Example:
+            >>> from rich.console import Console
+            >>> from rich.__main__ import make_test_card
+            >>> console = Console()
+            >>> with console.pager():
+                    console.print(make_test_card())
+
+        Returns:
+            PagerContext: A context manager.
+        """
+        return PagerContext(self, pager=pager, styles=styles, links=links)
+
+    def line(self, count: int = 1) -> None:
+        """Write new line(s).
+
+        Args:
+            count (int, optional): Number of new lines. Defaults to 1.
+        """
+
+        assert count >= 0, "count must be >= 0"
+        self.print(NewLine(count))
+
+    def clear(self, home: bool = True) -> None:
+        """Clear the screen.
+
+        Args:
+            home (bool, optional): Also move the cursor to 'home' position. Defaults to True.
+        """
+        if home:
+            self.control(Control.clear(), Control.home())
+        else:
+            self.control(Control.clear())
+
+    def status(
+        self,
+        status: RenderableType,
+        *,
+        spinner: str = "dots",
+        spinner_style: StyleType = "status.spinner",
+        speed: float = 1.0,
+        refresh_per_second: float = 12.5,
+    ) -> "Status":
+        """Display a status and spinner.
+
+        Args:
+            status (RenderableType): A status renderable (str or Text typically).
+            spinner (str, optional): Name of spinner animation (see python -m rich.spinner). Defaults to "dots".
+            spinner_style (StyleType, optional): Style of spinner. Defaults to "status.spinner".
+            speed (float, optional): Speed factor for spinner animation. Defaults to 1.0.
+            refresh_per_second (float, optional): Number of refreshes per second. Defaults to 12.5.
+
+        Returns:
+            Status: A Status object that may be used as a context manager.
+        """
+        from .status import Status
+
+        status_renderable = Status(
+            status,
+            console=self,
+            spinner=spinner,
+            spinner_style=spinner_style,
+            speed=speed,
+            refresh_per_second=refresh_per_second,
+        )
+        return status_renderable
+
+    def show_cursor(self, show: bool = True) -> bool:
+        """Show or hide the cursor.
+
+        Args:
+            show (bool, optional): Set visibility of the cursor.
+        """
+        if self.is_terminal:
+            self.control(Control.show_cursor(show))
+            return True
+        return False
+
+    def set_alt_screen(self, enable: bool = True) -> bool:
+        """Enables alternative screen mode.
+
+        Note, if you enable this mode, you should ensure that is disabled before
+        the application exits. See :meth:`~rich.Console.screen` for a context manager
+        that handles this for you.
+
+        Args:
+            enable (bool, optional): Enable (True) or disable (False) alternate screen. Defaults to True.
+
+        Returns:
+            bool: True if the control codes were written.
+
+        """
+        changed = False
+        if self.is_terminal and not self.legacy_windows:
+            self.control(Control.alt_screen(enable))
+            changed = True
+            self._is_alt_screen = enable
+        return changed
+
+    @property
+    def is_alt_screen(self) -> bool:
+        """Check if the alt screen was enabled.
+
+        Returns:
+            bool: True if the alt screen was enabled, otherwise False.
+        """
+        return self._is_alt_screen
+
+    def set_window_title(self, title: str) -> bool:
+        """Set the title of the console terminal window.
+
+        Warning: There is no means within Rich of "resetting" the window title to its
+        previous value, meaning the title you set will persist even after your application
+        exits.
+
+        ``fish`` shell resets the window title before and after each command by default,
+        negating this issue. Windows Terminal and command prompt will also reset the title for you.
+        Most other shells and terminals, however, do not do this.
+
+        Some terminals may require configuration changes before you can set the title.
+        Some terminals may not support setting the title at all.
+
+        Other software (including the terminal itself, the shell, custom prompts, plugins, etc.)
+        may also set the terminal window title. This could result in whatever value you write
+        using this method being overwritten.
+
+        Args:
+            title (str): The new title of the terminal window.
+
+        Returns:
+            bool: True if the control code to change the terminal title was
+                written, otherwise False. Note that a return value of True
+                does not guarantee that the window title has actually changed,
+                since the feature may be unsupported/disabled in some terminals.
+        """
+        if self.is_terminal:
+            self.control(Control.title(title))
+            return True
+        return False
+
+    def screen(
+        self, hide_cursor: bool = True, style: Optional[StyleType] = None
+    ) -> "ScreenContext":
+        """Context manager to enable and disable 'alternative screen' mode.
+
+        Args:
+            hide_cursor (bool, optional): Also hide the cursor. Defaults to False.
+            style (Style, optional): Optional style for screen. Defaults to None.
+
+        Returns:
+            ~ScreenContext: Context which enables alternate screen on enter, and disables it on exit.
+        """
+        return ScreenContext(self, hide_cursor=hide_cursor, style=style or "")
+
+    def measure(
+        self, renderable: RenderableType, *, options: Optional[ConsoleOptions] = None
+    ) -> Measurement:
+        """Measure a renderable. Returns a :class:`~rich.measure.Measurement` object which contains
+        information regarding the number of characters required to print the renderable.
+
+        Args:
+            renderable (RenderableType): Any renderable or string.
+            options (Optional[ConsoleOptions], optional): Options to use when measuring, or None
+                to use default options. Defaults to None.
+
+        Returns:
+            Measurement: A measurement of the renderable.
+        """
+        measurement = Measurement.get(self, options or self.options, renderable)
+        return measurement
+
+    def render(
+        self, renderable: RenderableType, options: Optional[ConsoleOptions] = None
+    ) -> Iterable[Segment]:
+        """Render an object in to an iterable of `Segment` instances.
+
+        This method contains the logic for rendering objects with the console protocol.
+        You are unlikely to need to use it directly, unless you are extending the library.
+
+        Args:
+            renderable (RenderableType): An object supporting the console protocol, or
+                an object that may be converted to a string.
+            options (ConsoleOptions, optional): An options object, or None to use self.options. Defaults to None.
+
+        Returns:
+            Iterable[Segment]: An iterable of segments that may be rendered.
+        """
+
+        _options = options or self.options
+        if _options.max_width < 1:
+            # No space to render anything. This prevents potential recursion errors.
+            return
+        render_iterable: RenderResult
+
+        renderable = rich_cast(renderable)
+        if hasattr(renderable, "__rich_console__") and not isclass(renderable):
+            render_iterable = renderable.__rich_console__(self, _options)
+        elif isinstance(renderable, str):
+            text_renderable = self.render_str(
+                renderable, highlight=_options.highlight, markup=_options.markup
+            )
+            render_iterable = text_renderable.__rich_console__(self, _options)
+        else:
+            raise errors.NotRenderableError(
+                f"Unable to render {renderable!r}; "
+                "A str, Segment or object with __rich_console__ method is required"
+            )
+
+        try:
+            iter_render = iter(render_iterable)
+        except TypeError:
+            raise errors.NotRenderableError(
+                f"object {render_iterable!r} is not renderable"
+            )
+        _Segment = Segment
+        _options = _options.reset_height()
+        for render_output in iter_render:
+            if isinstance(render_output, _Segment):
+                yield render_output
+            else:
+                yield from self.render(render_output, _options)
+
+    def render_lines(
+        self,
+        renderable: RenderableType,
+        options: Optional[ConsoleOptions] = None,
+        *,
+        style: Optional[Style] = None,
+        pad: bool = True,
+        new_lines: bool = False,
+    ) -> List[List[Segment]]:
+        """Render objects in to a list of lines.
+
+        The output of render_lines is useful when further formatting of rendered console text
+        is required, such as the Panel class which draws a border around any renderable object.
+
+        Args:
+            renderable (RenderableType): Any object renderable in the console.
+            options (Optional[ConsoleOptions], optional): Console options, or None to use self.options. Default to ``None``.
+            style (Style, optional): Optional style to apply to renderables. Defaults to ``None``.
+            pad (bool, optional): Pad lines shorter than render width. Defaults to ``True``.
+            new_lines (bool, optional): Include "\n" characters at end of lines.
+
+        Returns:
+            List[List[Segment]]: A list of lines, where a line is a list of Segment objects.
+        """
+        with self._lock:
+            render_options = options or self.options
+            _rendered = self.render(renderable, render_options)
+            if style:
+                _rendered = Segment.apply_style(_rendered, style)
+
+            render_height = render_options.height
+            if render_height is not None:
+                render_height = max(0, render_height)
+
+            lines = list(
+                islice(
+                    Segment.split_and_crop_lines(
+                        _rendered,
+                        render_options.max_width,
+                        include_new_lines=new_lines,
+                        pad=pad,
+                        style=style,
+                    ),
+                    None,
+                    render_height,
+                )
+            )
+            if render_options.height is not None:
+                extra_lines = render_options.height - len(lines)
+                if extra_lines > 0:
+                    pad_line = [
+                        (
+                            [
+                                Segment(" " * render_options.max_width, style),
+                                Segment("\n"),
+                            ]
+                            if new_lines
+                            else [Segment(" " * render_options.max_width, style)]
+                        )
+                    ]
+                    lines.extend(pad_line * extra_lines)
+
+            return lines
+
+    def render_str(
+        self,
+        text: str,
+        *,
+        style: Union[str, Style] = "",
+        justify: Optional[JustifyMethod] = None,
+        overflow: Optional[OverflowMethod] = None,
+        emoji: Optional[bool] = None,
+        markup: Optional[bool] = None,
+        highlight: Optional[bool] = None,
+        highlighter: Optional[HighlighterType] = None,
+    ) -> "Text":
+        """Convert a string to a Text instance. This is called automatically if
+        you print or log a string.
+
+        Args:
+            text (str): Text to render.
+            style (Union[str, Style], optional): Style to apply to rendered text.
+            justify (str, optional): Justify method: "default", "left", "center", "full", or "right". Defaults to ``None``.
+            overflow (str, optional): Overflow method: "crop", "fold", or "ellipsis". Defaults to ``None``.
+            emoji (Optional[bool], optional): Enable emoji, or ``None`` to use Console default.
+            markup (Optional[bool], optional): Enable markup, or ``None`` to use Console default.
+            highlight (Optional[bool], optional): Enable highlighting, or ``None`` to use Console default.
+            highlighter (HighlighterType, optional): Optional highlighter to apply.
+        Returns:
+            ConsoleRenderable: Renderable object.
+
+        """
+        emoji_enabled = emoji or (emoji is None and self._emoji)
+        markup_enabled = markup or (markup is None and self._markup)
+        highlight_enabled = highlight or (highlight is None and self._highlight)
+
+        if markup_enabled:
+            rich_text = render_markup(
+                text,
+                style=style,
+                emoji=emoji_enabled,
+                emoji_variant=self._emoji_variant,
+            )
+            rich_text.justify = justify
+            rich_text.overflow = overflow
+        else:
+            rich_text = Text(
+                (
+                    _emoji_replace(text, default_variant=self._emoji_variant)
+                    if emoji_enabled
+                    else text
+                ),
+                justify=justify,
+                overflow=overflow,
+                style=style,
+            )
+
+        _highlighter = (highlighter or self.highlighter) if highlight_enabled else None
+        if _highlighter is not None:
+            highlight_text = _highlighter(str(rich_text))
+            highlight_text.copy_styles(rich_text)
+            return highlight_text
+
+        return rich_text
+
+    def get_style(
+        self, name: Union[str, Style], *, default: Optional[Union[Style, str]] = None
+    ) -> Style:
+        """Get a Style instance by its theme name or parse a definition.
+
+        Args:
+            name (str): The name of a style or a style definition.
+
+        Returns:
+            Style: A Style object.
+
+        Raises:
+            MissingStyle: If no style could be parsed from name.
+
+        """
+        if isinstance(name, Style):
+            return name
+
+        try:
+            style = self._theme_stack.get(name)
+            if style is None:
+                style = Style.parse(name)
+            return style.copy() if style.link else style
+        except errors.StyleSyntaxError as error:
+            if default is not None:
+                return self.get_style(default)
+            raise errors.MissingStyle(
+                f"Failed to get style {name!r}; {error}"
+            ) from None
+
+    def _collect_renderables(
+        self,
+        objects: Iterable[Any],
+        sep: str,
+        end: str,
+        *,
+        justify: Optional[JustifyMethod] = None,
+        emoji: Optional[bool] = None,
+        markup: Optional[bool] = None,
+        highlight: Optional[bool] = None,
+    ) -> List[ConsoleRenderable]:
+        """Combine a number of renderables and text into one renderable.
+
+        Args:
+            objects (Iterable[Any]): Anything that Rich can render.
+            sep (str): String to write between print data.
+            end (str): String to write at end of print data.
+            justify (str, optional): One of "left", "right", "center", or "full". Defaults to ``None``.
+            emoji (Optional[bool], optional): Enable emoji code, or ``None`` to use console default.
+            markup (Optional[bool], optional): Enable markup, or ``None`` to use console default.
+            highlight (Optional[bool], optional): Enable automatic highlighting, or ``None`` to use console default.
+
+        Returns:
+            List[ConsoleRenderable]: A list of things to render.
+        """
+        renderables: List[ConsoleRenderable] = []
+        _append = renderables.append
+        text: List[Text] = []
+        append_text = text.append
+
+        append = _append
+        if justify in ("left", "center", "right"):
+
+            def align_append(renderable: RenderableType) -> None:
+                _append(Align(renderable, cast(AlignMethod, justify)))
+
+            append = align_append
+
+        _highlighter: HighlighterType = _null_highlighter
+        if highlight or (highlight is None and self._highlight):
+            _highlighter = self.highlighter
+
+        def check_text() -> None:
+            if text:
+                sep_text = Text(sep, justify=justify, end=end)
+                append(sep_text.join(text))
+                text.clear()
+
+        for renderable in objects:
+            renderable = rich_cast(renderable)
+            if isinstance(renderable, str):
+                append_text(
+                    self.render_str(
+                        renderable,
+                        emoji=emoji,
+                        markup=markup,
+                        highlight=highlight,
+                        highlighter=_highlighter,
+                    )
+                )
+            elif isinstance(renderable, Text):
+                append_text(renderable)
+            elif isinstance(renderable, ConsoleRenderable):
+                check_text()
+                append(renderable)
+            elif is_expandable(renderable):
+                check_text()
+                append(Pretty(renderable, highlighter=_highlighter))
+            else:
+                append_text(_highlighter(str(renderable)))
+
+        check_text()
+
+        if self.style is not None:
+            style = self.get_style(self.style)
+            renderables = [Styled(renderable, style) for renderable in renderables]
+
+        return renderables
+
+    def rule(
+        self,
+        title: TextType = "",
+        *,
+        characters: str = "─",
+        style: Union[str, Style] = "rule.line",
+        align: AlignMethod = "center",
+    ) -> None:
+        """Draw a line with optional centered title.
+
+        Args:
+            title (str, optional): Text to render over the rule. Defaults to "".
+            characters (str, optional): Character(s) to form the line. Defaults to "─".
+            style (str, optional): Style of line. Defaults to "rule.line".
+            align (str, optional): How to align the title, one of "left", "center", or "right". Defaults to "center".
+        """
+        from .rule import Rule
+
+        rule = Rule(title=title, characters=characters, style=style, align=align)
+        self.print(rule)
+
+    def control(self, *control: Control) -> None:
+        """Insert non-printing control codes.
+
+        Args:
+            control_codes (str): Control codes, such as those that may move the cursor.
+        """
+        if not self.is_dumb_terminal:
+            with self:
+                self._buffer.extend(_control.segment for _control in control)
+
+    def out(
+        self,
+        *objects: Any,
+        sep: str = " ",
+        end: str = "\n",
+        style: Optional[Union[str, Style]] = None,
+        highlight: Optional[bool] = None,
+    ) -> None:
+        """Output to the terminal. This is a low-level way of writing to the terminal which unlike
+        :meth:`~rich.console.Console.print` won't pretty print, wrap text, or apply markup, but will
+        optionally apply highlighting and a basic style.
+
+        Args:
+            sep (str, optional): String to write between print data. Defaults to " ".
+            end (str, optional): String to write at end of print data. Defaults to "\\\\n".
+            style (Union[str, Style], optional): A style to apply to output. Defaults to None.
+            highlight (Optional[bool], optional): Enable automatic highlighting, or ``None`` to use
+                console default. Defaults to ``None``.
+        """
+        raw_output: str = sep.join(str(_object) for _object in objects)
+        self.print(
+            raw_output,
+            style=style,
+            highlight=highlight,
+            emoji=False,
+            markup=False,
+            no_wrap=True,
+            overflow="ignore",
+            crop=False,
+            end=end,
+        )
+
+    def print(
+        self,
+        *objects: Any,
+        sep: str = " ",
+        end: str = "\n",
+        style: Optional[Union[str, Style]] = None,
+        justify: Optional[JustifyMethod] = None,
+        overflow: Optional[OverflowMethod] = None,
+        no_wrap: Optional[bool] = None,
+        emoji: Optional[bool] = None,
+        markup: Optional[bool] = None,
+        highlight: Optional[bool] = None,
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+        crop: bool = True,
+        soft_wrap: Optional[bool] = None,
+        new_line_start: bool = False,
+    ) -> None:
+        """Print to the console.
+
+        Args:
+            objects (positional args): Objects to log to the terminal.
+            sep (str, optional): String to write between print data. Defaults to " ".
+            end (str, optional): String to write at end of print data. Defaults to "\\\\n".
+            style (Union[str, Style], optional): A style to apply to output. Defaults to None.
+            justify (str, optional): Justify method: "default", "left", "right", "center", or "full". Defaults to ``None``.
+            overflow (str, optional): Overflow method: "ignore", "crop", "fold", or "ellipsis". Defaults to None.
+            no_wrap (Optional[bool], optional): Disable word wrapping. Defaults to None.
+            emoji (Optional[bool], optional): Enable emoji code, or ``None`` to use console default. Defaults to ``None``.
+            markup (Optional[bool], optional): Enable markup, or ``None`` to use console default. Defaults to ``None``.
+            highlight (Optional[bool], optional): Enable automatic highlighting, or ``None`` to use console default. Defaults to ``None``.
+            width (Optional[int], optional): Width of output, or ``None`` to auto-detect. Defaults to ``None``.
+            crop (Optional[bool], optional): Crop output to width of terminal. Defaults to True.
+            soft_wrap (bool, optional): Enable soft wrap mode which disables word wrapping and cropping of text or ``None`` for
+                Console default. Defaults to ``None``.
+            new_line_start (bool, False): Insert a new line at the start if the output contains more than one line. Defaults to ``False``.
+        """
+        if not objects:
+            objects = (NewLine(),)
+
+        if soft_wrap is None:
+            soft_wrap = self.soft_wrap
+        if soft_wrap:
+            if no_wrap is None:
+                no_wrap = True
+            if overflow is None:
+                overflow = "ignore"
+            crop = False
+        render_hooks = self._render_hooks[:]
+        with self:
+            renderables = self._collect_renderables(
+                objects,
+                sep,
+                end,
+                justify=justify,
+                emoji=emoji,
+                markup=markup,
+                highlight=highlight,
+            )
+            for hook in render_hooks:
+                renderables = hook.process_renderables(renderables)
+            render_options = self.options.update(
+                justify=justify,
+                overflow=overflow,
+                width=min(width, self.width) if width is not None else NO_CHANGE,
+                height=height,
+                no_wrap=no_wrap,
+                markup=markup,
+                highlight=highlight,
+            )
+
+            new_segments: List[Segment] = []
+            extend = new_segments.extend
+            render = self.render
+            if style is None:
+                for renderable in renderables:
+                    extend(render(renderable, render_options))
+            else:
+                for renderable in renderables:
+                    extend(
+                        Segment.apply_style(
+                            render(renderable, render_options), self.get_style(style)
+                        )
+                    )
+            if new_line_start:
+                if (
+                    len("".join(segment.text for segment in new_segments).splitlines())
+                    > 1
+                ):
+                    new_segments.insert(0, Segment.line())
+            if crop:
+                buffer_extend = self._buffer.extend
+                for line in Segment.split_and_crop_lines(
+                    new_segments, self.width, pad=False
+                ):
+                    buffer_extend(line)
+            else:
+                self._buffer.extend(new_segments)
+
+    def print_json(
+        self,
+        json: Optional[str] = None,
+        *,
+        data: Any = None,
+        indent: Union[None, int, str] = 2,
+        highlight: bool = True,
+        skip_keys: bool = False,
+        ensure_ascii: bool = False,
+        check_circular: bool = True,
+        allow_nan: bool = True,
+        default: Optional[Callable[[Any], Any]] = None,
+        sort_keys: bool = False,
+    ) -> None:
+        """Pretty prints JSON. Output will be valid JSON.
+
+        Args:
+            json (Optional[str]): A string containing JSON.
+            data (Any): If json is not supplied, then encode this data.
+            indent (Union[None, int, str], optional): Number of spaces to indent. Defaults to 2.
+            highlight (bool, optional): Enable highlighting of output: Defaults to True.
+            skip_keys (bool, optional): Skip keys not of a basic type. Defaults to False.
+            ensure_ascii (bool, optional): Escape all non-ascii characters. Defaults to False.
+            check_circular (bool, optional): Check for circular references. Defaults to True.
+            allow_nan (bool, optional): Allow NaN and Infinity values. Defaults to True.
+            default (Callable, optional): A callable that converts values that can not be encoded
+                in to something that can be JSON encoded. Defaults to None.
+            sort_keys (bool, optional): Sort dictionary keys. Defaults to False.
+        """
+        from pip._vendor.rich.json import JSON
+
+        if json is None:
+            json_renderable = JSON.from_data(
+                data,
+                indent=indent,
+                highlight=highlight,
+                skip_keys=skip_keys,
+                ensure_ascii=ensure_ascii,
+                check_circular=check_circular,
+                allow_nan=allow_nan,
+                default=default,
+                sort_keys=sort_keys,
+            )
+        else:
+            if not isinstance(json, str):
+                raise TypeError(
+                    f"json must be str. Did you mean print_json(data={json!r}) ?"
+                )
+            json_renderable = JSON(
+                json,
+                indent=indent,
+                highlight=highlight,
+                skip_keys=skip_keys,
+                ensure_ascii=ensure_ascii,
+                check_circular=check_circular,
+                allow_nan=allow_nan,
+                default=default,
+                sort_keys=sort_keys,
+            )
+        self.print(json_renderable, soft_wrap=True)
+
+    def update_screen(
+        self,
+        renderable: RenderableType,
+        *,
+        region: Optional[Region] = None,
+        options: Optional[ConsoleOptions] = None,
+    ) -> None:
+        """Update the screen at a given offset.
+
+        Args:
+            renderable (RenderableType): A Rich renderable.
+            region (Region, optional): Region of screen to update, or None for entire screen. Defaults to None.
+            x (int, optional): x offset. Defaults to 0.
+            y (int, optional): y offset. Defaults to 0.
+
+        Raises:
+            errors.NoAltScreen: If the Console isn't in alt screen mode.
+
+        """
+        if not self.is_alt_screen:
+            raise errors.NoAltScreen("Alt screen must be enabled to call update_screen")
+        render_options = options or self.options
+        if region is None:
+            x = y = 0
+            render_options = render_options.update_dimensions(
+                render_options.max_width, render_options.height or self.height
+            )
+        else:
+            x, y, width, height = region
+            render_options = render_options.update_dimensions(width, height)
+
+        lines = self.render_lines(renderable, options=render_options)
+        self.update_screen_lines(lines, x, y)
+
+    def update_screen_lines(
+        self, lines: List[List[Segment]], x: int = 0, y: int = 0
+    ) -> None:
+        """Update lines of the screen at a given offset.
+
+        Args:
+            lines (List[List[Segment]]): Rendered lines (as produced by :meth:`~rich.Console.render_lines`).
+            x (int, optional): x offset (column no). Defaults to 0.
+            y (int, optional): y offset (column no). Defaults to 0.
+
+        Raises:
+            errors.NoAltScreen: If the Console isn't in alt screen mode.
+        """
+        if not self.is_alt_screen:
+            raise errors.NoAltScreen("Alt screen must be enabled to call update_screen")
+        screen_update = ScreenUpdate(lines, x, y)
+        segments = self.render(screen_update)
+        self._buffer.extend(segments)
+        self._check_buffer()
+
+    def print_exception(
+        self,
+        *,
+        width: Optional[int] = 100,
+        extra_lines: int = 3,
+        theme: Optional[str] = None,
+        word_wrap: bool = False,
+        show_locals: bool = False,
+        suppress: Iterable[Union[str, ModuleType]] = (),
+        max_frames: int = 100,
+    ) -> None:
+        """Prints a rich render of the last exception and traceback.
+
+        Args:
+            width (Optional[int], optional): Number of characters used to render code. Defaults to 100.
+            extra_lines (int, optional): Additional lines of code to render. Defaults to 3.
+            theme (str, optional): Override pygments theme used in traceback
+            word_wrap (bool, optional): Enable word wrapping of long lines. Defaults to False.
+            show_locals (bool, optional): Enable display of local variables. Defaults to False.
+            suppress (Iterable[Union[str, ModuleType]]): Optional sequence of modules or paths to exclude from traceback.
+            max_frames (int): Maximum number of frames to show in a traceback, 0 for no maximum. Defaults to 100.
+        """
+        from .traceback import Traceback
+
+        traceback = Traceback(
+            width=width,
+            extra_lines=extra_lines,
+            theme=theme,
+            word_wrap=word_wrap,
+            show_locals=show_locals,
+            suppress=suppress,
+            max_frames=max_frames,
+        )
+        self.print(traceback)
+
+    @staticmethod
+    def _caller_frame_info(
+        offset: int,
+        currentframe: Callable[[], Optional[FrameType]] = inspect.currentframe,
+    ) -> Tuple[str, int, Dict[str, Any]]:
+        """Get caller frame information.
+
+        Args:
+            offset (int): the caller offset within the current frame stack.
+            currentframe (Callable[[], Optional[FrameType]], optional): the callable to use to
+                retrieve the current frame. Defaults to ``inspect.currentframe``.
+
+        Returns:
+            Tuple[str, int, Dict[str, Any]]: A tuple containing the filename, the line number and
+                the dictionary of local variables associated with the caller frame.
+
+        Raises:
+            RuntimeError: If the stack offset is invalid.
+        """
+        # Ignore the frame of this local helper
+        offset += 1
+
+        frame = currentframe()
+        if frame is not None:
+            # Use the faster currentframe where implemented
+            while offset and frame is not None:
+                frame = frame.f_back
+                offset -= 1
+            assert frame is not None
+            return frame.f_code.co_filename, frame.f_lineno, frame.f_locals
+        else:
+            # Fallback to the slower stack
+            frame_info = inspect.stack()[offset]
+            return frame_info.filename, frame_info.lineno, frame_info.frame.f_locals
+
+    def log(
+        self,
+        *objects: Any,
+        sep: str = " ",
+        end: str = "\n",
+        style: Optional[Union[str, Style]] = None,
+        justify: Optional[JustifyMethod] = None,
+        emoji: Optional[bool] = None,
+        markup: Optional[bool] = None,
+        highlight: Optional[bool] = None,
+        log_locals: bool = False,
+        _stack_offset: int = 1,
+    ) -> None:
+        """Log rich content to the terminal.
+
+        Args:
+            objects (positional args): Objects to log to the terminal.
+            sep (str, optional): String to write between print data. Defaults to " ".
+            end (str, optional): String to write at end of print data. Defaults to "\\\\n".
+            style (Union[str, Style], optional): A style to apply to output. Defaults to None.
+            justify (str, optional): One of "left", "right", "center", or "full". Defaults to ``None``.
+            emoji (Optional[bool], optional): Enable emoji code, or ``None`` to use console default. Defaults to None.
+            markup (Optional[bool], optional): Enable markup, or ``None`` to use console default. Defaults to None.
+            highlight (Optional[bool], optional): Enable automatic highlighting, or ``None`` to use console default. Defaults to None.
+            log_locals (bool, optional): Boolean to enable logging of locals where ``log()``
+                was called. Defaults to False.
+            _stack_offset (int, optional): Offset of caller from end of call stack. Defaults to 1.
+        """
+        if not objects:
+            objects = (NewLine(),)
+
+        render_hooks = self._render_hooks[:]
+
+        with self:
+            renderables = self._collect_renderables(
+                objects,
+                sep,
+                end,
+                justify=justify,
+                emoji=emoji,
+                markup=markup,
+                highlight=highlight,
+            )
+            if style is not None:
+                renderables = [Styled(renderable, style) for renderable in renderables]
+
+            filename, line_no, locals = self._caller_frame_info(_stack_offset)
+            link_path = None if filename.startswith("<") else os.path.abspath(filename)
+            path = filename.rpartition(os.sep)[-1]
+            if log_locals:
+                locals_map = {
+                    key: value
+                    for key, value in locals.items()
+                    if not key.startswith("__")
+                }
+                renderables.append(render_scope(locals_map, title="[i]locals"))
+
+            renderables = [
+                self._log_render(
+                    self,
+                    renderables,
+                    log_time=self.get_datetime(),
+                    path=path,
+                    line_no=line_no,
+                    link_path=link_path,
+                )
+            ]
+            for hook in render_hooks:
+                renderables = hook.process_renderables(renderables)
+            new_segments: List[Segment] = []
+            extend = new_segments.extend
+            render = self.render
+            render_options = self.options
+            for renderable in renderables:
+                extend(render(renderable, render_options))
+            buffer_extend = self._buffer.extend
+            for line in Segment.split_and_crop_lines(
+                new_segments, self.width, pad=False
+            ):
+                buffer_extend(line)
+
+    def on_broken_pipe(self) -> None:
+        """This function is called when a `BrokenPipeError` is raised.
+
+        This can occur when piping Textual output in Linux and macOS.
+        The default implementation is to exit the app, but you could implement
+        this method in a subclass to change the behavior.
+
+        See https://docs.python.org/3/library/signal.html#note-on-sigpipe for details.
+        """
+        self.quiet = True
+        devnull = os.open(os.devnull, os.O_WRONLY)
+        os.dup2(devnull, sys.stdout.fileno())
+        raise SystemExit(1)
+
+    def _check_buffer(self) -> None:
+        """Check if the buffer may be rendered. Render it if it can (e.g. Console.quiet is False)
+        Rendering is supported on Windows, Unix and Jupyter environments. For
+        legacy Windows consoles, the win32 API is called directly.
+        This method will also record what it renders if recording is enabled via Console.record.
+        """
+        if self.quiet:
+            del self._buffer[:]
+            return
+
+        try:
+            self._write_buffer()
+        except BrokenPipeError:
+            self.on_broken_pipe()
+
+    def _write_buffer(self) -> None:
+        """Write the buffer to the output file."""
+
+        with self._lock:
+            if self.record and not self._buffer_index:
+                with self._record_buffer_lock:
+                    self._record_buffer.extend(self._buffer[:])
+
+            if self._buffer_index == 0:
+                if self.is_jupyter:  # pragma: no cover
+                    from .jupyter import display
+
+                    display(self._buffer, self._render_buffer(self._buffer[:]))
+                    del self._buffer[:]
+                else:
+                    if WINDOWS:
+                        use_legacy_windows_render = False
+                        if self.legacy_windows:
+                            fileno = get_fileno(self.file)
+                            if fileno is not None:
+                                use_legacy_windows_render = (
+                                    fileno in _STD_STREAMS_OUTPUT
+                                )
+
+                        if use_legacy_windows_render:
+                            from pip._vendor.rich._win32_console import LegacyWindowsTerm
+                            from pip._vendor.rich._windows_renderer import legacy_windows_render
+
+                            buffer = self._buffer[:]
+                            if self.no_color and self._color_system:
+                                buffer = list(Segment.remove_color(buffer))
+
+                            legacy_windows_render(buffer, LegacyWindowsTerm(self.file))
+                        else:
+                            # Either a non-std stream on legacy Windows, or modern Windows.
+                            text = self._render_buffer(self._buffer[:])
+                            # https://bugs.python.org/issue37871
+                            # https://github.com/python/cpython/issues/82052
+                            # We need to avoid writing more than 32Kb in a single write, due to the above bug
+                            write = self.file.write
+                            # Worse case scenario, every character is 4 bytes of utf-8
+                            MAX_WRITE = 32 * 1024 // 4
+                            try:
+                                if len(text) <= MAX_WRITE:
+                                    write(text)
+                                else:
+                                    batch: List[str] = []
+                                    batch_append = batch.append
+                                    size = 0
+                                    for line in text.splitlines(True):
+                                        if size + len(line) > MAX_WRITE and batch:
+                                            write("".join(batch))
+                                            batch.clear()
+                                            size = 0
+                                        batch_append(line)
+                                        size += len(line)
+                                    if batch:
+                                        write("".join(batch))
+                                        batch.clear()
+                            except UnicodeEncodeError as error:
+                                error.reason = f"{error.reason}\n*** You may need to add PYTHONIOENCODING=utf-8 to your environment ***"
+                                raise
+                    else:
+                        text = self._render_buffer(self._buffer[:])
+                        try:
+                            self.file.write(text)
+                        except UnicodeEncodeError as error:
+                            error.reason = f"{error.reason}\n*** You may need to add PYTHONIOENCODING=utf-8 to your environment ***"
+                            raise
+
+                    self.file.flush()
+                    del self._buffer[:]
+
+    def _render_buffer(self, buffer: Iterable[Segment]) -> str:
+        """Render buffered output, and clear buffer."""
+        output: List[str] = []
+        append = output.append
+        color_system = self._color_system
+        legacy_windows = self.legacy_windows
+        not_terminal = not self.is_terminal
+        if self.no_color and color_system:
+            buffer = Segment.remove_color(buffer)
+        for text, style, control in buffer:
+            if style:
+                append(
+                    style.render(
+                        text,
+                        color_system=color_system,
+                        legacy_windows=legacy_windows,
+                    )
+                )
+            elif not (not_terminal and control):
+                append(text)
+
+        rendered = "".join(output)
+        return rendered
+
+    def input(
+        self,
+        prompt: TextType = "",
+        *,
+        markup: bool = True,
+        emoji: bool = True,
+        password: bool = False,
+        stream: Optional[TextIO] = None,
+    ) -> str:
+        """Displays a prompt and waits for input from the user. The prompt may contain color / style.
+
+        It works in the same way as Python's builtin :func:`input` function and provides elaborate line editing and history features if Python's builtin :mod:`readline` module is previously loaded.
+
+        Args:
+            prompt (Union[str, Text]): Text to render in the prompt.
+            markup (bool, optional): Enable console markup (requires a str prompt). Defaults to True.
+            emoji (bool, optional): Enable emoji (requires a str prompt). Defaults to True.
+            password: (bool, optional): Hide typed text. Defaults to False.
+            stream: (TextIO, optional): Optional file to read input from (rather than stdin). Defaults to None.
+
+        Returns:
+            str: Text read from stdin.
+        """
+        if prompt:
+            self.print(prompt, markup=markup, emoji=emoji, end="")
+        if password:
+            result = getpass("", stream=stream)
+        else:
+            if stream:
+                result = stream.readline()
+            else:
+                result = input()
+        return result
+
+    def export_text(self, *, clear: bool = True, styles: bool = False) -> str:
+        """Generate text from console contents (requires record=True argument in constructor).
+
+        Args:
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``.
+            styles (bool, optional): If ``True``, ansi escape codes will be included. ``False`` for plain text.
+                Defaults to ``False``.
+
+        Returns:
+            str: String containing console contents.
+
+        """
+        assert (
+            self.record
+        ), "To export console contents set record=True in the constructor or instance"
+
+        with self._record_buffer_lock:
+            if styles:
+                text = "".join(
+                    (style.render(text) if style else text)
+                    for text, style, _ in self._record_buffer
+                )
+            else:
+                text = "".join(
+                    segment.text
+                    for segment in self._record_buffer
+                    if not segment.control
+                )
+            if clear:
+                del self._record_buffer[:]
+        return text
+
+    def save_text(self, path: str, *, clear: bool = True, styles: bool = False) -> None:
+        """Generate text from console and save to a given location (requires record=True argument in constructor).
+
+        Args:
+            path (str): Path to write text files.
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``.
+            styles (bool, optional): If ``True``, ansi style codes will be included. ``False`` for plain text.
+                Defaults to ``False``.
+
+        """
+        text = self.export_text(clear=clear, styles=styles)
+        with open(path, "w", encoding="utf-8") as write_file:
+            write_file.write(text)
+
+    def export_html(
+        self,
+        *,
+        theme: Optional[TerminalTheme] = None,
+        clear: bool = True,
+        code_format: Optional[str] = None,
+        inline_styles: bool = False,
+    ) -> str:
+        """Generate HTML from console contents (requires record=True argument in constructor).
+
+        Args:
+            theme (TerminalTheme, optional): TerminalTheme object containing console colors.
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``.
+            code_format (str, optional): Format string to render HTML. In addition to '{foreground}',
+                '{background}', and '{code}', should contain '{stylesheet}' if inline_styles is ``False``.
+            inline_styles (bool, optional): If ``True`` styles will be inlined in to spans, which makes files
+                larger but easier to cut and paste markup. If ``False``, styles will be embedded in a style tag.
+                Defaults to False.
+
+        Returns:
+            str: String containing console contents as HTML.
+        """
+        assert (
+            self.record
+        ), "To export console contents set record=True in the constructor or instance"
+        fragments: List[str] = []
+        append = fragments.append
+        _theme = theme or DEFAULT_TERMINAL_THEME
+        stylesheet = ""
+
+        render_code_format = CONSOLE_HTML_FORMAT if code_format is None else code_format
+
+        with self._record_buffer_lock:
+            if inline_styles:
+                for text, style, _ in Segment.filter_control(
+                    Segment.simplify(self._record_buffer)
+                ):
+                    text = escape(text)
+                    if style:
+                        rule = style.get_html_style(_theme)
+                        if style.link:
+                            text = f'<a href="{style.link}">{text}</a>'
+                        text = f'<span style="{rule}">{text}</span>' if rule else text
+                    append(text)
+            else:
+                styles: Dict[str, int] = {}
+                for text, style, _ in Segment.filter_control(
+                    Segment.simplify(self._record_buffer)
+                ):
+                    text = escape(text)
+                    if style:
+                        rule = style.get_html_style(_theme)
+                        style_number = styles.setdefault(rule, len(styles) + 1)
+                        if style.link:
+                            text = f'<a class="r{style_number}" href="{style.link}">{text}</a>'
+                        else:
+                            text = f'<span class="r{style_number}">{text}</span>'
+                    append(text)
+                stylesheet_rules: List[str] = []
+                stylesheet_append = stylesheet_rules.append
+                for style_rule, style_number in styles.items():
+                    if style_rule:
+                        stylesheet_append(f".r{style_number} {{{style_rule}}}")
+                stylesheet = "\n".join(stylesheet_rules)
+
+            rendered_code = render_code_format.format(
+                code="".join(fragments),
+                stylesheet=stylesheet,
+                foreground=_theme.foreground_color.hex,
+                background=_theme.background_color.hex,
+            )
+            if clear:
+                del self._record_buffer[:]
+        return rendered_code
+
+    def save_html(
+        self,
+        path: str,
+        *,
+        theme: Optional[TerminalTheme] = None,
+        clear: bool = True,
+        code_format: str = CONSOLE_HTML_FORMAT,
+        inline_styles: bool = False,
+    ) -> None:
+        """Generate HTML from console contents and write to a file (requires record=True argument in constructor).
+
+        Args:
+            path (str): Path to write html file.
+            theme (TerminalTheme, optional): TerminalTheme object containing console colors.
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``.
+            code_format (str, optional): Format string to render HTML. In addition to '{foreground}',
+                '{background}', and '{code}', should contain '{stylesheet}' if inline_styles is ``False``.
+            inline_styles (bool, optional): If ``True`` styles will be inlined in to spans, which makes files
+                larger but easier to cut and paste markup. If ``False``, styles will be embedded in a style tag.
+                Defaults to False.
+
+        """
+        html = self.export_html(
+            theme=theme,
+            clear=clear,
+            code_format=code_format,
+            inline_styles=inline_styles,
+        )
+        with open(path, "w", encoding="utf-8") as write_file:
+            write_file.write(html)
+
+    def export_svg(
+        self,
+        *,
+        title: str = "Rich",
+        theme: Optional[TerminalTheme] = None,
+        clear: bool = True,
+        code_format: str = CONSOLE_SVG_FORMAT,
+        font_aspect_ratio: float = 0.61,
+        unique_id: Optional[str] = None,
+    ) -> str:
+        """
+        Generate an SVG from the console contents (requires record=True in Console constructor).
+
+        Args:
+            title (str, optional): The title of the tab in the output image
+            theme (TerminalTheme, optional): The ``TerminalTheme`` object to use to style the terminal
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``
+            code_format (str, optional): Format string used to generate the SVG. Rich will inject a number of variables
+                into the string in order to form the final SVG output. The default template used and the variables
+                injected by Rich can be found by inspecting the ``console.CONSOLE_SVG_FORMAT`` variable.
+            font_aspect_ratio (float, optional): The width to height ratio of the font used in the ``code_format``
+                string. Defaults to 0.61, which is the width to height ratio of Fira Code (the default font).
+                If you aren't specifying a different font inside ``code_format``, you probably don't need this.
+            unique_id (str, optional): unique id that is used as the prefix for various elements (CSS styles, node
+                ids). If not set, this defaults to a computed value based on the recorded content.
+        """
+
+        from pip._vendor.rich.cells import cell_len
+
+        style_cache: Dict[Style, str] = {}
+
+        def get_svg_style(style: Style) -> str:
+            """Convert a Style to CSS rules for SVG."""
+            if style in style_cache:
+                return style_cache[style]
+            css_rules = []
+            color = (
+                _theme.foreground_color
+                if (style.color is None or style.color.is_default)
+                else style.color.get_truecolor(_theme)
+            )
+            bgcolor = (
+                _theme.background_color
+                if (style.bgcolor is None or style.bgcolor.is_default)
+                else style.bgcolor.get_truecolor(_theme)
+            )
+            if style.reverse:
+                color, bgcolor = bgcolor, color
+            if style.dim:
+                color = blend_rgb(color, bgcolor, 0.4)
+            css_rules.append(f"fill: {color.hex}")
+            if style.bold:
+                css_rules.append("font-weight: bold")
+            if style.italic:
+                css_rules.append("font-style: italic;")
+            if style.underline:
+                css_rules.append("text-decoration: underline;")
+            if style.strike:
+                css_rules.append("text-decoration: line-through;")
+
+            css = ";".join(css_rules)
+            style_cache[style] = css
+            return css
+
+        _theme = theme or SVG_EXPORT_THEME
+
+        width = self.width
+        char_height = 20
+        char_width = char_height * font_aspect_ratio
+        line_height = char_height * 1.22
+
+        margin_top = 1
+        margin_right = 1
+        margin_bottom = 1
+        margin_left = 1
+
+        padding_top = 40
+        padding_right = 8
+        padding_bottom = 8
+        padding_left = 8
+
+        padding_width = padding_left + padding_right
+        padding_height = padding_top + padding_bottom
+        margin_width = margin_left + margin_right
+        margin_height = margin_top + margin_bottom
+
+        text_backgrounds: List[str] = []
+        text_group: List[str] = []
+        classes: Dict[str, int] = {}
+        style_no = 1
+
+        def escape_text(text: str) -> str:
+            """HTML escape text and replace spaces with nbsp."""
+            return escape(text).replace(" ", "&#160;")
+
+        def make_tag(
+            name: str, content: Optional[str] = None, **attribs: object
+        ) -> str:
+            """Make a tag from name, content, and attributes."""
+
+            def stringify(value: object) -> str:
+                if isinstance(value, (float)):
+                    return format(value, "g")
+                return str(value)
+
+            tag_attribs = " ".join(
+                f'{k.lstrip("_").replace("_", "-")}="{stringify(v)}"'
+                for k, v in attribs.items()
+            )
+            return (
+                f"<{name} {tag_attribs}>{content}</{name}>"
+                if content
+                else f"<{name} {tag_attribs}/>"
+            )
+
+        with self._record_buffer_lock:
+            segments = list(Segment.filter_control(self._record_buffer))
+            if clear:
+                self._record_buffer.clear()
+
+        if unique_id is None:
+            unique_id = "terminal-" + str(
+                zlib.adler32(
+                    ("".join(repr(segment) for segment in segments)).encode(
+                        "utf-8",
+                        "ignore",
+                    )
+                    + title.encode("utf-8", "ignore")
+                )
+            )
+        y = 0
+        for y, line in enumerate(Segment.split_and_crop_lines(segments, length=width)):
+            x = 0
+            for text, style, _control in line:
+                style = style or Style()
+                rules = get_svg_style(style)
+                if rules not in classes:
+                    classes[rules] = style_no
+                    style_no += 1
+                class_name = f"r{classes[rules]}"
+
+                if style.reverse:
+                    has_background = True
+                    background = (
+                        _theme.foreground_color.hex
+                        if style.color is None
+                        else style.color.get_truecolor(_theme).hex
+                    )
+                else:
+                    bgcolor = style.bgcolor
+                    has_background = bgcolor is not None and not bgcolor.is_default
+                    background = (
+                        _theme.background_color.hex
+                        if style.bgcolor is None
+                        else style.bgcolor.get_truecolor(_theme).hex
+                    )
+
+                text_length = cell_len(text)
+                if has_background:
+                    text_backgrounds.append(
+                        make_tag(
+                            "rect",
+                            fill=background,
+                            x=x * char_width,
+                            y=y * line_height + 1.5,
+                            width=char_width * text_length,
+                            height=line_height + 0.25,
+                            shape_rendering="crispEdges",
+                        )
+                    )
+
+                if text != " " * len(text):
+                    text_group.append(
+                        make_tag(
+                            "text",
+                            escape_text(text),
+                            _class=f"{unique_id}-{class_name}",
+                            x=x * char_width,
+                            y=y * line_height + char_height,
+                            textLength=char_width * len(text),
+                            clip_path=f"url(#{unique_id}-line-{y})",
+                        )
+                    )
+                x += cell_len(text)
+
+        line_offsets = [line_no * line_height + 1.5 for line_no in range(y)]
+        lines = "\n".join(
+            f"""<clipPath id="{unique_id}-line-{line_no}">
+    {make_tag("rect", x=0, y=offset, width=char_width * width, height=line_height + 0.25)}
+            </clipPath>"""
+            for line_no, offset in enumerate(line_offsets)
+        )
+
+        styles = "\n".join(
+            f".{unique_id}-r{rule_no} {{ {css} }}" for css, rule_no in classes.items()
+        )
+        backgrounds = "".join(text_backgrounds)
+        matrix = "".join(text_group)
+
+        terminal_width = ceil(width * char_width + padding_width)
+        terminal_height = (y + 1) * line_height + padding_height
+        chrome = make_tag(
+            "rect",
+            fill=_theme.background_color.hex,
+            stroke="rgba(255,255,255,0.35)",
+            stroke_width="1",
+            x=margin_left,
+            y=margin_top,
+            width=terminal_width,
+            height=terminal_height,
+            rx=8,
+        )
+
+        title_color = _theme.foreground_color.hex
+        if title:
+            chrome += make_tag(
+                "text",
+                escape_text(title),
+                _class=f"{unique_id}-title",
+                fill=title_color,
+                text_anchor="middle",
+                x=terminal_width // 2,
+                y=margin_top + char_height + 6,
+            )
+        chrome += f"""
+            <g transform="translate(26,22)">
+            <circle cx="0" cy="0" r="7" fill="#ff5f57"/>
+            <circle cx="22" cy="0" r="7" fill="#febc2e"/>
+            <circle cx="44" cy="0" r="7" fill="#28c840"/>
+            </g>
+        """
+
+        svg = code_format.format(
+            unique_id=unique_id,
+            char_width=char_width,
+            char_height=char_height,
+            line_height=line_height,
+            terminal_width=char_width * width - 1,
+            terminal_height=(y + 1) * line_height - 1,
+            width=terminal_width + margin_width,
+            height=terminal_height + margin_height,
+            terminal_x=margin_left + padding_left,
+            terminal_y=margin_top + padding_top,
+            styles=styles,
+            chrome=chrome,
+            backgrounds=backgrounds,
+            matrix=matrix,
+            lines=lines,
+        )
+        return svg
+
+    def save_svg(
+        self,
+        path: str,
+        *,
+        title: str = "Rich",
+        theme: Optional[TerminalTheme] = None,
+        clear: bool = True,
+        code_format: str = CONSOLE_SVG_FORMAT,
+        font_aspect_ratio: float = 0.61,
+        unique_id: Optional[str] = None,
+    ) -> None:
+        """Generate an SVG file from the console contents (requires record=True in Console constructor).
+
+        Args:
+            path (str): The path to write the SVG to.
+            title (str, optional): The title of the tab in the output image
+            theme (TerminalTheme, optional): The ``TerminalTheme`` object to use to style the terminal
+            clear (bool, optional): Clear record buffer after exporting. Defaults to ``True``
+            code_format (str, optional): Format string used to generate the SVG. Rich will inject a number of variables
+                into the string in order to form the final SVG output. The default template used and the variables
+                injected by Rich can be found by inspecting the ``console.CONSOLE_SVG_FORMAT`` variable.
+            font_aspect_ratio (float, optional): The width to height ratio of the font used in the ``code_format``
+                string. Defaults to 0.61, which is the width to height ratio of Fira Code (the default font).
+                If you aren't specifying a different font inside ``code_format``, you probably don't need this.
+            unique_id (str, optional): unique id that is used as the prefix for various elements (CSS styles, node
+                ids). If not set, this defaults to a computed value based on the recorded content.
+        """
+        svg = self.export_svg(
+            title=title,
+            theme=theme,
+            clear=clear,
+            code_format=code_format,
+            font_aspect_ratio=font_aspect_ratio,
+            unique_id=unique_id,
+        )
+        with open(path, "w", encoding="utf-8") as write_file:
+            write_file.write(svg)
+
+
+def _svg_hash(svg_main_code: str) -> str:
+    """Returns a unique hash for the given SVG main code.
+
+    Args:
+        svg_main_code (str): The content we're going to inject in the SVG envelope.
+
+    Returns:
+        str: a hash of the given content
+    """
+    return str(zlib.adler32(svg_main_code.encode()))
+
+
+if __name__ == "__main__":  # pragma: no cover
+    console = Console(record=True)
+
+    console.log(
+        "JSONRPC [i]request[/i]",
+        5,
+        1.3,
+        True,
+        False,
+        None,
+        {
+            "jsonrpc": "2.0",
+            "method": "subtract",
+            "params": {"minuend": 42, "subtrahend": 23},
+            "id": 3,
+        },
+    )
+
+    console.log("Hello, World!", "{'a': 1}", repr(console))
+
+    console.print(
+        {
+            "name": None,
+            "empty": [],
+            "quiz": {
+                "sport": {
+                    "answered": True,
+                    "q1": {
+                        "question": "Which one is correct team name in NBA?",
+                        "options": [
+                            "New York Bulls",
+                            "Los Angeles Kings",
+                            "Golden State Warriors",
+                            "Huston Rocket",
+                        ],
+                        "answer": "Huston Rocket",
+                    },
+                },
+                "maths": {
+                    "answered": False,
+                    "q1": {
+                        "question": "5 + 7 = ?",
+                        "options": [10, 11, 12, 13],
+                        "answer": 12,
+                    },
+                    "q2": {
+                        "question": "12 - 8 = ?",
+                        "options": [1, 2, 3, 4],
+                        "answer": 4,
+                    },
+                },
+            },
+        }
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/constrain.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/constrain.py
new file mode 100644
index 00000000..65fdf563
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/constrain.py
@@ -0,0 +1,37 @@
+from typing import Optional, TYPE_CHECKING
+
+from .jupyter import JupyterMixin
+from .measure import Measurement
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderableType, RenderResult
+
+
+class Constrain(JupyterMixin):
+    """Constrain the width of a renderable to a given number of characters.
+
+    Args:
+        renderable (RenderableType): A renderable object.
+        width (int, optional): The maximum width (in characters) to render. Defaults to 80.
+    """
+
+    def __init__(self, renderable: "RenderableType", width: Optional[int] = 80) -> None:
+        self.renderable = renderable
+        self.width = width
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        if self.width is None:
+            yield self.renderable
+        else:
+            child_options = options.update_width(min(self.width, options.max_width))
+            yield from console.render(self.renderable, child_options)
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        if self.width is not None:
+            options = options.update_width(self.width)
+        measurement = Measurement.get(console, options, self.renderable)
+        return measurement
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/containers.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/containers.py
new file mode 100644
index 00000000..901ff8ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/containers.py
@@ -0,0 +1,167 @@
+from itertools import zip_longest
+from typing import (
+    TYPE_CHECKING,
+    Iterable,
+    Iterator,
+    List,
+    Optional,
+    TypeVar,
+    Union,
+    overload,
+)
+
+if TYPE_CHECKING:
+    from .console import (
+        Console,
+        ConsoleOptions,
+        JustifyMethod,
+        OverflowMethod,
+        RenderResult,
+        RenderableType,
+    )
+    from .text import Text
+
+from .cells import cell_len
+from .measure import Measurement
+
+T = TypeVar("T")
+
+
+class Renderables:
+    """A list subclass which renders its contents to the console."""
+
+    def __init__(
+        self, renderables: Optional[Iterable["RenderableType"]] = None
+    ) -> None:
+        self._renderables: List["RenderableType"] = (
+            list(renderables) if renderables is not None else []
+        )
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        """Console render method to insert line-breaks."""
+        yield from self._renderables
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        dimensions = [
+            Measurement.get(console, options, renderable)
+            for renderable in self._renderables
+        ]
+        if not dimensions:
+            return Measurement(1, 1)
+        _min = max(dimension.minimum for dimension in dimensions)
+        _max = max(dimension.maximum for dimension in dimensions)
+        return Measurement(_min, _max)
+
+    def append(self, renderable: "RenderableType") -> None:
+        self._renderables.append(renderable)
+
+    def __iter__(self) -> Iterable["RenderableType"]:
+        return iter(self._renderables)
+
+
+class Lines:
+    """A list subclass which can render to the console."""
+
+    def __init__(self, lines: Iterable["Text"] = ()) -> None:
+        self._lines: List["Text"] = list(lines)
+
+    def __repr__(self) -> str:
+        return f"Lines({self._lines!r})"
+
+    def __iter__(self) -> Iterator["Text"]:
+        return iter(self._lines)
+
+    @overload
+    def __getitem__(self, index: int) -> "Text":
+        ...
+
+    @overload
+    def __getitem__(self, index: slice) -> List["Text"]:
+        ...
+
+    def __getitem__(self, index: Union[slice, int]) -> Union["Text", List["Text"]]:
+        return self._lines[index]
+
+    def __setitem__(self, index: int, value: "Text") -> "Lines":
+        self._lines[index] = value
+        return self
+
+    def __len__(self) -> int:
+        return self._lines.__len__()
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        """Console render method to insert line-breaks."""
+        yield from self._lines
+
+    def append(self, line: "Text") -> None:
+        self._lines.append(line)
+
+    def extend(self, lines: Iterable["Text"]) -> None:
+        self._lines.extend(lines)
+
+    def pop(self, index: int = -1) -> "Text":
+        return self._lines.pop(index)
+
+    def justify(
+        self,
+        console: "Console",
+        width: int,
+        justify: "JustifyMethod" = "left",
+        overflow: "OverflowMethod" = "fold",
+    ) -> None:
+        """Justify and overflow text to a given width.
+
+        Args:
+            console (Console): Console instance.
+            width (int): Number of cells available per line.
+            justify (str, optional): Default justify method for text: "left", "center", "full" or "right". Defaults to "left".
+            overflow (str, optional): Default overflow for text: "crop", "fold", or "ellipsis". Defaults to "fold".
+
+        """
+        from .text import Text
+
+        if justify == "left":
+            for line in self._lines:
+                line.truncate(width, overflow=overflow, pad=True)
+        elif justify == "center":
+            for line in self._lines:
+                line.rstrip()
+                line.truncate(width, overflow=overflow)
+                line.pad_left((width - cell_len(line.plain)) // 2)
+                line.pad_right(width - cell_len(line.plain))
+        elif justify == "right":
+            for line in self._lines:
+                line.rstrip()
+                line.truncate(width, overflow=overflow)
+                line.pad_left(width - cell_len(line.plain))
+        elif justify == "full":
+            for line_index, line in enumerate(self._lines):
+                if line_index == len(self._lines) - 1:
+                    break
+                words = line.split(" ")
+                words_size = sum(cell_len(word.plain) for word in words)
+                num_spaces = len(words) - 1
+                spaces = [1 for _ in range(num_spaces)]
+                index = 0
+                if spaces:
+                    while words_size + num_spaces < width:
+                        spaces[len(spaces) - index - 1] += 1
+                        num_spaces += 1
+                        index = (index + 1) % len(spaces)
+                tokens: List[Text] = []
+                for index, (word, next_word) in enumerate(
+                    zip_longest(words, words[1:])
+                ):
+                    tokens.append(word)
+                    if index < len(spaces):
+                        style = word.get_style_at_offset(console, -1)
+                        next_style = next_word.get_style_at_offset(console, 0)
+                        space_style = style if style == next_style else line.style
+                        tokens.append(Text(" " * spaces[index], style=space_style))
+                self[line_index] = Text("").join(tokens)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/control.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/control.py
new file mode 100644
index 00000000..88fcb929
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/control.py
@@ -0,0 +1,225 @@
+import sys
+import time
+from typing import TYPE_CHECKING, Callable, Dict, Iterable, List, Union
+
+if sys.version_info >= (3, 8):
+    from typing import Final
+else:
+    from pip._vendor.typing_extensions import Final  # pragma: no cover
+
+from .segment import ControlCode, ControlType, Segment
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderResult
+
+STRIP_CONTROL_CODES: Final = [
+    7,  # Bell
+    8,  # Backspace
+    11,  # Vertical tab
+    12,  # Form feed
+    13,  # Carriage return
+]
+_CONTROL_STRIP_TRANSLATE: Final = {
+    _codepoint: None for _codepoint in STRIP_CONTROL_CODES
+}
+
+CONTROL_ESCAPE: Final = {
+    7: "\\a",
+    8: "\\b",
+    11: "\\v",
+    12: "\\f",
+    13: "\\r",
+}
+
+CONTROL_CODES_FORMAT: Dict[int, Callable[..., str]] = {
+    ControlType.BELL: lambda: "\x07",
+    ControlType.CARRIAGE_RETURN: lambda: "\r",
+    ControlType.HOME: lambda: "\x1b[H",
+    ControlType.CLEAR: lambda: "\x1b[2J",
+    ControlType.ENABLE_ALT_SCREEN: lambda: "\x1b[?1049h",
+    ControlType.DISABLE_ALT_SCREEN: lambda: "\x1b[?1049l",
+    ControlType.SHOW_CURSOR: lambda: "\x1b[?25h",
+    ControlType.HIDE_CURSOR: lambda: "\x1b[?25l",
+    ControlType.CURSOR_UP: lambda param: f"\x1b[{param}A",
+    ControlType.CURSOR_DOWN: lambda param: f"\x1b[{param}B",
+    ControlType.CURSOR_FORWARD: lambda param: f"\x1b[{param}C",
+    ControlType.CURSOR_BACKWARD: lambda param: f"\x1b[{param}D",
+    ControlType.CURSOR_MOVE_TO_COLUMN: lambda param: f"\x1b[{param+1}G",
+    ControlType.ERASE_IN_LINE: lambda param: f"\x1b[{param}K",
+    ControlType.CURSOR_MOVE_TO: lambda x, y: f"\x1b[{y+1};{x+1}H",
+    ControlType.SET_WINDOW_TITLE: lambda title: f"\x1b]0;{title}\x07",
+}
+
+
+class Control:
+    """A renderable that inserts a control code (non printable but may move cursor).
+
+    Args:
+        *codes (str): Positional arguments are either a :class:`~rich.segment.ControlType` enum or a
+            tuple of ControlType and an integer parameter
+    """
+
+    __slots__ = ["segment"]
+
+    def __init__(self, *codes: Union[ControlType, ControlCode]) -> None:
+        control_codes: List[ControlCode] = [
+            (code,) if isinstance(code, ControlType) else code for code in codes
+        ]
+        _format_map = CONTROL_CODES_FORMAT
+        rendered_codes = "".join(
+            _format_map[code](*parameters) for code, *parameters in control_codes
+        )
+        self.segment = Segment(rendered_codes, None, control_codes)
+
+    @classmethod
+    def bell(cls) -> "Control":
+        """Ring the 'bell'."""
+        return cls(ControlType.BELL)
+
+    @classmethod
+    def home(cls) -> "Control":
+        """Move cursor to 'home' position."""
+        return cls(ControlType.HOME)
+
+    @classmethod
+    def move(cls, x: int = 0, y: int = 0) -> "Control":
+        """Move cursor relative to current position.
+
+        Args:
+            x (int): X offset.
+            y (int): Y offset.
+
+        Returns:
+            ~Control: Control object.
+
+        """
+
+        def get_codes() -> Iterable[ControlCode]:
+            control = ControlType
+            if x:
+                yield (
+                    control.CURSOR_FORWARD if x > 0 else control.CURSOR_BACKWARD,
+                    abs(x),
+                )
+            if y:
+                yield (
+                    control.CURSOR_DOWN if y > 0 else control.CURSOR_UP,
+                    abs(y),
+                )
+
+        control = cls(*get_codes())
+        return control
+
+    @classmethod
+    def move_to_column(cls, x: int, y: int = 0) -> "Control":
+        """Move to the given column, optionally add offset to row.
+
+        Returns:
+            x (int): absolute x (column)
+            y (int): optional y offset (row)
+
+        Returns:
+            ~Control: Control object.
+        """
+
+        return (
+            cls(
+                (ControlType.CURSOR_MOVE_TO_COLUMN, x),
+                (
+                    ControlType.CURSOR_DOWN if y > 0 else ControlType.CURSOR_UP,
+                    abs(y),
+                ),
+            )
+            if y
+            else cls((ControlType.CURSOR_MOVE_TO_COLUMN, x))
+        )
+
+    @classmethod
+    def move_to(cls, x: int, y: int) -> "Control":
+        """Move cursor to absolute position.
+
+        Args:
+            x (int): x offset (column)
+            y (int): y offset (row)
+
+        Returns:
+            ~Control: Control object.
+        """
+        return cls((ControlType.CURSOR_MOVE_TO, x, y))
+
+    @classmethod
+    def clear(cls) -> "Control":
+        """Clear the screen."""
+        return cls(ControlType.CLEAR)
+
+    @classmethod
+    def show_cursor(cls, show: bool) -> "Control":
+        """Show or hide the cursor."""
+        return cls(ControlType.SHOW_CURSOR if show else ControlType.HIDE_CURSOR)
+
+    @classmethod
+    def alt_screen(cls, enable: bool) -> "Control":
+        """Enable or disable alt screen."""
+        if enable:
+            return cls(ControlType.ENABLE_ALT_SCREEN, ControlType.HOME)
+        else:
+            return cls(ControlType.DISABLE_ALT_SCREEN)
+
+    @classmethod
+    def title(cls, title: str) -> "Control":
+        """Set the terminal window title
+
+        Args:
+            title (str): The new terminal window title
+        """
+        return cls((ControlType.SET_WINDOW_TITLE, title))
+
+    def __str__(self) -> str:
+        return self.segment.text
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        if self.segment.text:
+            yield self.segment
+
+
+def strip_control_codes(
+    text: str, _translate_table: Dict[int, None] = _CONTROL_STRIP_TRANSLATE
+) -> str:
+    """Remove control codes from text.
+
+    Args:
+        text (str): A string possibly contain control codes.
+
+    Returns:
+        str: String with control codes removed.
+    """
+    return text.translate(_translate_table)
+
+
+def escape_control_codes(
+    text: str,
+    _translate_table: Dict[int, str] = CONTROL_ESCAPE,
+) -> str:
+    """Replace control codes with their "escaped" equivalent in the given text.
+    (e.g. "\b" becomes "\\b")
+
+    Args:
+        text (str): A string possibly containing control codes.
+
+    Returns:
+        str: String with control codes replaced with their escaped version.
+    """
+    return text.translate(_translate_table)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console
+
+    console = Console()
+    console.print("Look at the title of your terminal window ^")
+    # console.print(Control((ControlType.SET_WINDOW_TITLE, "Hello, world!")))
+    for i in range(10):
+        console.set_window_title("🚀 Loading" + "." * i)
+        time.sleep(0.5)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/default_styles.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/default_styles.py
new file mode 100644
index 00000000..6c0d7323
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/default_styles.py
@@ -0,0 +1,191 @@
+from typing import Dict
+
+from .style import Style
+
+DEFAULT_STYLES: Dict[str, Style] = {
+    "none": Style.null(),
+    "reset": Style(
+        color="default",
+        bgcolor="default",
+        dim=False,
+        bold=False,
+        italic=False,
+        underline=False,
+        blink=False,
+        blink2=False,
+        reverse=False,
+        conceal=False,
+        strike=False,
+    ),
+    "dim": Style(dim=True),
+    "bright": Style(dim=False),
+    "bold": Style(bold=True),
+    "strong": Style(bold=True),
+    "code": Style(reverse=True, bold=True),
+    "italic": Style(italic=True),
+    "emphasize": Style(italic=True),
+    "underline": Style(underline=True),
+    "blink": Style(blink=True),
+    "blink2": Style(blink2=True),
+    "reverse": Style(reverse=True),
+    "strike": Style(strike=True),
+    "black": Style(color="black"),
+    "red": Style(color="red"),
+    "green": Style(color="green"),
+    "yellow": Style(color="yellow"),
+    "magenta": Style(color="magenta"),
+    "cyan": Style(color="cyan"),
+    "white": Style(color="white"),
+    "inspect.attr": Style(color="yellow", italic=True),
+    "inspect.attr.dunder": Style(color="yellow", italic=True, dim=True),
+    "inspect.callable": Style(bold=True, color="red"),
+    "inspect.async_def": Style(italic=True, color="bright_cyan"),
+    "inspect.def": Style(italic=True, color="bright_cyan"),
+    "inspect.class": Style(italic=True, color="bright_cyan"),
+    "inspect.error": Style(bold=True, color="red"),
+    "inspect.equals": Style(),
+    "inspect.help": Style(color="cyan"),
+    "inspect.doc": Style(dim=True),
+    "inspect.value.border": Style(color="green"),
+    "live.ellipsis": Style(bold=True, color="red"),
+    "layout.tree.row": Style(dim=False, color="red"),
+    "layout.tree.column": Style(dim=False, color="blue"),
+    "logging.keyword": Style(bold=True, color="yellow"),
+    "logging.level.notset": Style(dim=True),
+    "logging.level.debug": Style(color="green"),
+    "logging.level.info": Style(color="blue"),
+    "logging.level.warning": Style(color="yellow"),
+    "logging.level.error": Style(color="red", bold=True),
+    "logging.level.critical": Style(color="red", bold=True, reverse=True),
+    "log.level": Style.null(),
+    "log.time": Style(color="cyan", dim=True),
+    "log.message": Style.null(),
+    "log.path": Style(dim=True),
+    "repr.ellipsis": Style(color="yellow"),
+    "repr.indent": Style(color="green", dim=True),
+    "repr.error": Style(color="red", bold=True),
+    "repr.str": Style(color="green", italic=False, bold=False),
+    "repr.brace": Style(bold=True),
+    "repr.comma": Style(bold=True),
+    "repr.ipv4": Style(bold=True, color="bright_green"),
+    "repr.ipv6": Style(bold=True, color="bright_green"),
+    "repr.eui48": Style(bold=True, color="bright_green"),
+    "repr.eui64": Style(bold=True, color="bright_green"),
+    "repr.tag_start": Style(bold=True),
+    "repr.tag_name": Style(color="bright_magenta", bold=True),
+    "repr.tag_contents": Style(color="default"),
+    "repr.tag_end": Style(bold=True),
+    "repr.attrib_name": Style(color="yellow", italic=False),
+    "repr.attrib_equal": Style(bold=True),
+    "repr.attrib_value": Style(color="magenta", italic=False),
+    "repr.number": Style(color="cyan", bold=True, italic=False),
+    "repr.number_complex": Style(color="cyan", bold=True, italic=False),  # same
+    "repr.bool_true": Style(color="bright_green", italic=True),
+    "repr.bool_false": Style(color="bright_red", italic=True),
+    "repr.none": Style(color="magenta", italic=True),
+    "repr.url": Style(underline=True, color="bright_blue", italic=False, bold=False),
+    "repr.uuid": Style(color="bright_yellow", bold=False),
+    "repr.call": Style(color="magenta", bold=True),
+    "repr.path": Style(color="magenta"),
+    "repr.filename": Style(color="bright_magenta"),
+    "rule.line": Style(color="bright_green"),
+    "rule.text": Style.null(),
+    "json.brace": Style(bold=True),
+    "json.bool_true": Style(color="bright_green", italic=True),
+    "json.bool_false": Style(color="bright_red", italic=True),
+    "json.null": Style(color="magenta", italic=True),
+    "json.number": Style(color="cyan", bold=True, italic=False),
+    "json.str": Style(color="green", italic=False, bold=False),
+    "json.key": Style(color="blue", bold=True),
+    "prompt": Style.null(),
+    "prompt.choices": Style(color="magenta", bold=True),
+    "prompt.default": Style(color="cyan", bold=True),
+    "prompt.invalid": Style(color="red"),
+    "prompt.invalid.choice": Style(color="red"),
+    "pretty": Style.null(),
+    "scope.border": Style(color="blue"),
+    "scope.key": Style(color="yellow", italic=True),
+    "scope.key.special": Style(color="yellow", italic=True, dim=True),
+    "scope.equals": Style(color="red"),
+    "table.header": Style(bold=True),
+    "table.footer": Style(bold=True),
+    "table.cell": Style.null(),
+    "table.title": Style(italic=True),
+    "table.caption": Style(italic=True, dim=True),
+    "traceback.error": Style(color="red", italic=True),
+    "traceback.border.syntax_error": Style(color="bright_red"),
+    "traceback.border": Style(color="red"),
+    "traceback.text": Style.null(),
+    "traceback.title": Style(color="red", bold=True),
+    "traceback.exc_type": Style(color="bright_red", bold=True),
+    "traceback.exc_value": Style.null(),
+    "traceback.offset": Style(color="bright_red", bold=True),
+    "traceback.error_range": Style(underline=True, bold=True, dim=False),
+    "bar.back": Style(color="grey23"),
+    "bar.complete": Style(color="rgb(249,38,114)"),
+    "bar.finished": Style(color="rgb(114,156,31)"),
+    "bar.pulse": Style(color="rgb(249,38,114)"),
+    "progress.description": Style.null(),
+    "progress.filesize": Style(color="green"),
+    "progress.filesize.total": Style(color="green"),
+    "progress.download": Style(color="green"),
+    "progress.elapsed": Style(color="yellow"),
+    "progress.percentage": Style(color="magenta"),
+    "progress.remaining": Style(color="cyan"),
+    "progress.data.speed": Style(color="red"),
+    "progress.spinner": Style(color="green"),
+    "status.spinner": Style(color="green"),
+    "tree": Style(),
+    "tree.line": Style(),
+    "markdown.paragraph": Style(),
+    "markdown.text": Style(),
+    "markdown.em": Style(italic=True),
+    "markdown.emph": Style(italic=True),  # For commonmark backwards compatibility
+    "markdown.strong": Style(bold=True),
+    "markdown.code": Style(bold=True, color="cyan", bgcolor="black"),
+    "markdown.code_block": Style(color="cyan", bgcolor="black"),
+    "markdown.block_quote": Style(color="magenta"),
+    "markdown.list": Style(color="cyan"),
+    "markdown.item": Style(),
+    "markdown.item.bullet": Style(color="yellow", bold=True),
+    "markdown.item.number": Style(color="yellow", bold=True),
+    "markdown.hr": Style(color="yellow"),
+    "markdown.h1.border": Style(),
+    "markdown.h1": Style(bold=True),
+    "markdown.h2": Style(bold=True, underline=True),
+    "markdown.h3": Style(bold=True),
+    "markdown.h4": Style(bold=True, dim=True),
+    "markdown.h5": Style(underline=True),
+    "markdown.h6": Style(italic=True),
+    "markdown.h7": Style(italic=True, dim=True),
+    "markdown.link": Style(color="bright_blue"),
+    "markdown.link_url": Style(color="blue", underline=True),
+    "markdown.s": Style(strike=True),
+    "iso8601.date": Style(color="blue"),
+    "iso8601.time": Style(color="magenta"),
+    "iso8601.timezone": Style(color="yellow"),
+}
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import argparse
+    import io
+
+    from pip._vendor.rich.console import Console
+    from pip._vendor.rich.table import Table
+    from pip._vendor.rich.text import Text
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--html", action="store_true", help="Export as HTML table")
+    args = parser.parse_args()
+    html: bool = args.html
+    console = Console(record=True, width=70, file=io.StringIO()) if html else Console()
+
+    table = Table("Name", "Styling")
+
+    for style_name, style in DEFAULT_STYLES.items():
+        table.add_row(Text(style_name, style=style), str(style))
+
+    console.print(table)
+    if html:
+        print(console.export_html(inline_styles=True))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/diagnose.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/diagnose.py
new file mode 100644
index 00000000..ad361838
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/diagnose.py
@@ -0,0 +1,37 @@
+import os
+import platform
+
+from pip._vendor.rich import inspect
+from pip._vendor.rich.console import Console, get_windows_console_features
+from pip._vendor.rich.panel import Panel
+from pip._vendor.rich.pretty import Pretty
+
+
+def report() -> None:  # pragma: no cover
+    """Print a report to the terminal with debugging information"""
+    console = Console()
+    inspect(console)
+    features = get_windows_console_features()
+    inspect(features)
+
+    env_names = (
+        "TERM",
+        "COLORTERM",
+        "CLICOLOR",
+        "NO_COLOR",
+        "TERM_PROGRAM",
+        "COLUMNS",
+        "LINES",
+        "JUPYTER_COLUMNS",
+        "JUPYTER_LINES",
+        "JPY_PARENT_PID",
+        "VSCODE_VERBOSE_LOGGING",
+    )
+    env = {name: os.getenv(name) for name in env_names}
+    console.print(Panel.fit((Pretty(env)), title="[b]Environment Variables"))
+
+    console.print(f'platform="{platform.system()}"')
+
+
+if __name__ == "__main__":  # pragma: no cover
+    report()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/emoji.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/emoji.py
new file mode 100644
index 00000000..791f0465
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/emoji.py
@@ -0,0 +1,96 @@
+import sys
+from typing import TYPE_CHECKING, Optional, Union
+
+from .jupyter import JupyterMixin
+from .segment import Segment
+from .style import Style
+from ._emoji_codes import EMOJI
+from ._emoji_replace import _emoji_replace
+
+if sys.version_info >= (3, 8):
+    from typing import Literal
+else:
+    from pip._vendor.typing_extensions import Literal  # pragma: no cover
+
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderResult
+
+
+EmojiVariant = Literal["emoji", "text"]
+
+
+class NoEmoji(Exception):
+    """No emoji by that name."""
+
+
+class Emoji(JupyterMixin):
+    __slots__ = ["name", "style", "_char", "variant"]
+
+    VARIANTS = {"text": "\uFE0E", "emoji": "\uFE0F"}
+
+    def __init__(
+        self,
+        name: str,
+        style: Union[str, Style] = "none",
+        variant: Optional[EmojiVariant] = None,
+    ) -> None:
+        """A single emoji character.
+
+        Args:
+            name (str): Name of emoji.
+            style (Union[str, Style], optional): Optional style. Defaults to None.
+
+        Raises:
+            NoEmoji: If the emoji doesn't exist.
+        """
+        self.name = name
+        self.style = style
+        self.variant = variant
+        try:
+            self._char = EMOJI[name]
+        except KeyError:
+            raise NoEmoji(f"No emoji called {name!r}")
+        if variant is not None:
+            self._char += self.VARIANTS.get(variant, "")
+
+    @classmethod
+    def replace(cls, text: str) -> str:
+        """Replace emoji markup with corresponding unicode characters.
+
+        Args:
+            text (str): A string with emojis codes, e.g. "Hello :smiley:!"
+
+        Returns:
+            str: A string with emoji codes replaces with actual emoji.
+        """
+        return _emoji_replace(text)
+
+    def __repr__(self) -> str:
+        return f"<emoji {self.name!r}>"
+
+    def __str__(self) -> str:
+        return self._char
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        yield Segment(self._char, console.get_style(self.style))
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sys
+
+    from pip._vendor.rich.columns import Columns
+    from pip._vendor.rich.console import Console
+
+    console = Console(record=True)
+
+    columns = Columns(
+        (f":{name}: {name}" for name in sorted(EMOJI.keys()) if "\u200D" not in name),
+        column_first=True,
+    )
+
+    console.print(columns)
+    if len(sys.argv) > 1:
+        console.save_html(sys.argv[1])
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/errors.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/errors.py
new file mode 100644
index 00000000..0bcbe53e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/errors.py
@@ -0,0 +1,34 @@
+class ConsoleError(Exception):
+    """An error in console operation."""
+
+
+class StyleError(Exception):
+    """An error in styles."""
+
+
+class StyleSyntaxError(ConsoleError):
+    """Style was badly formatted."""
+
+
+class MissingStyle(StyleError):
+    """No such style."""
+
+
+class StyleStackError(ConsoleError):
+    """Style stack is invalid."""
+
+
+class NotRenderableError(ConsoleError):
+    """Object is not renderable."""
+
+
+class MarkupError(ConsoleError):
+    """Markup was badly formatted."""
+
+
+class LiveError(ConsoleError):
+    """Error related to Live display."""
+
+
+class NoAltScreen(ConsoleError):
+    """Alt screen mode was required."""
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/file_proxy.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/file_proxy.py
new file mode 100644
index 00000000..4b0b0da6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/file_proxy.py
@@ -0,0 +1,57 @@
+import io
+from typing import IO, TYPE_CHECKING, Any, List
+
+from .ansi import AnsiDecoder
+from .text import Text
+
+if TYPE_CHECKING:
+    from .console import Console
+
+
+class FileProxy(io.TextIOBase):
+    """Wraps a file (e.g. sys.stdout) and redirects writes to a console."""
+
+    def __init__(self, console: "Console", file: IO[str]) -> None:
+        self.__console = console
+        self.__file = file
+        self.__buffer: List[str] = []
+        self.__ansi_decoder = AnsiDecoder()
+
+    @property
+    def rich_proxied_file(self) -> IO[str]:
+        """Get proxied file."""
+        return self.__file
+
+    def __getattr__(self, name: str) -> Any:
+        return getattr(self.__file, name)
+
+    def write(self, text: str) -> int:
+        if not isinstance(text, str):
+            raise TypeError(f"write() argument must be str, not {type(text).__name__}")
+        buffer = self.__buffer
+        lines: List[str] = []
+        while text:
+            line, new_line, text = text.partition("\n")
+            if new_line:
+                lines.append("".join(buffer) + line)
+                buffer.clear()
+            else:
+                buffer.append(line)
+                break
+        if lines:
+            console = self.__console
+            with console:
+                output = Text("\n").join(
+                    self.__ansi_decoder.decode_line(line) for line in lines
+                )
+                console.print(output)
+        return len(text)
+
+    def flush(self) -> None:
+        output = "".join(self.__buffer)
+        if output:
+            self.__console.print(output)
+        del self.__buffer[:]
+
+    def fileno(self) -> int:
+        return self.__file.fileno()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/filesize.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/filesize.py
new file mode 100644
index 00000000..83bc9118
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/filesize.py
@@ -0,0 +1,88 @@
+"""Functions for reporting filesizes. Borrowed from https://github.com/PyFilesystem/pyfilesystem2
+
+The functions declared in this module should cover the different
+use cases needed to generate a string representation of a file size
+using several different units. Since there are many standards regarding
+file size units, three different functions have been implemented.
+
+See Also:
+    * `Wikipedia: Binary prefix <https://en.wikipedia.org/wiki/Binary_prefix>`_
+
+"""
+
+__all__ = ["decimal"]
+
+from typing import Iterable, List, Optional, Tuple
+
+
+def _to_str(
+    size: int,
+    suffixes: Iterable[str],
+    base: int,
+    *,
+    precision: Optional[int] = 1,
+    separator: Optional[str] = " ",
+) -> str:
+    if size == 1:
+        return "1 byte"
+    elif size < base:
+        return f"{size:,} bytes"
+
+    for i, suffix in enumerate(suffixes, 2):  # noqa: B007
+        unit = base**i
+        if size < unit:
+            break
+    return "{:,.{precision}f}{separator}{}".format(
+        (base * size / unit),
+        suffix,
+        precision=precision,
+        separator=separator,
+    )
+
+
+def pick_unit_and_suffix(size: int, suffixes: List[str], base: int) -> Tuple[int, str]:
+    """Pick a suffix and base for the given size."""
+    for i, suffix in enumerate(suffixes):
+        unit = base**i
+        if size < unit * base:
+            break
+    return unit, suffix
+
+
+def decimal(
+    size: int,
+    *,
+    precision: Optional[int] = 1,
+    separator: Optional[str] = " ",
+) -> str:
+    """Convert a filesize in to a string (powers of 1000, SI prefixes).
+
+    In this convention, ``1000 B = 1 kB``.
+
+    This is typically the format used to advertise the storage
+    capacity of USB flash drives and the like (*256 MB* meaning
+    actually a storage capacity of more than *256 000 000 B*),
+    or used by **Mac OS X** since v10.6 to report file sizes.
+
+    Arguments:
+        int (size): A file size.
+        int (precision): The number of decimal places to include (default = 1).
+        str (separator): The string to separate the value from the units (default = " ").
+
+    Returns:
+        `str`: A string containing a abbreviated file size and units.
+
+    Example:
+        >>> filesize.decimal(30000)
+        '30.0 kB'
+        >>> filesize.decimal(30000, precision=2, separator="")
+        '30.00kB'
+
+    """
+    return _to_str(
+        size,
+        ("kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"),
+        1000,
+        precision=precision,
+        separator=separator,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/highlighter.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/highlighter.py
new file mode 100644
index 00000000..e4c462e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/highlighter.py
@@ -0,0 +1,232 @@
+import re
+from abc import ABC, abstractmethod
+from typing import List, Union
+
+from .text import Span, Text
+
+
+def _combine_regex(*regexes: str) -> str:
+    """Combine a number of regexes in to a single regex.
+
+    Returns:
+        str: New regex with all regexes ORed together.
+    """
+    return "|".join(regexes)
+
+
+class Highlighter(ABC):
+    """Abstract base class for highlighters."""
+
+    def __call__(self, text: Union[str, Text]) -> Text:
+        """Highlight a str or Text instance.
+
+        Args:
+            text (Union[str, ~Text]): Text to highlight.
+
+        Raises:
+            TypeError: If not called with text or str.
+
+        Returns:
+            Text: A test instance with highlighting applied.
+        """
+        if isinstance(text, str):
+            highlight_text = Text(text)
+        elif isinstance(text, Text):
+            highlight_text = text.copy()
+        else:
+            raise TypeError(f"str or Text instance required, not {text!r}")
+        self.highlight(highlight_text)
+        return highlight_text
+
+    @abstractmethod
+    def highlight(self, text: Text) -> None:
+        """Apply highlighting in place to text.
+
+        Args:
+            text (~Text): A text object highlight.
+        """
+
+
+class NullHighlighter(Highlighter):
+    """A highlighter object that doesn't highlight.
+
+    May be used to disable highlighting entirely.
+
+    """
+
+    def highlight(self, text: Text) -> None:
+        """Nothing to do"""
+
+
+class RegexHighlighter(Highlighter):
+    """Applies highlighting from a list of regular expressions."""
+
+    highlights: List[str] = []
+    base_style: str = ""
+
+    def highlight(self, text: Text) -> None:
+        """Highlight :class:`rich.text.Text` using regular expressions.
+
+        Args:
+            text (~Text): Text to highlighted.
+
+        """
+
+        highlight_regex = text.highlight_regex
+        for re_highlight in self.highlights:
+            highlight_regex(re_highlight, style_prefix=self.base_style)
+
+
+class ReprHighlighter(RegexHighlighter):
+    """Highlights the text typically produced from ``__repr__`` methods."""
+
+    base_style = "repr."
+    highlights = [
+        r"(?P<tag_start><)(?P<tag_name>[-\w.:|]*)(?P<tag_contents>[\w\W]*)(?P<tag_end>>)",
+        r'(?P<attrib_name>[\w_]{1,50})=(?P<attrib_value>"?[\w_]+"?)?',
+        r"(?P<brace>[][{}()])",
+        _combine_regex(
+            r"(?P<ipv4>[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3})",
+            r"(?P<ipv6>([A-Fa-f0-9]{1,4}::?){1,7}[A-Fa-f0-9]{1,4})",
+            r"(?P<eui64>(?:[0-9A-Fa-f]{1,2}-){7}[0-9A-Fa-f]{1,2}|(?:[0-9A-Fa-f]{1,2}:){7}[0-9A-Fa-f]{1,2}|(?:[0-9A-Fa-f]{4}\.){3}[0-9A-Fa-f]{4})",
+            r"(?P<eui48>(?:[0-9A-Fa-f]{1,2}-){5}[0-9A-Fa-f]{1,2}|(?:[0-9A-Fa-f]{1,2}:){5}[0-9A-Fa-f]{1,2}|(?:[0-9A-Fa-f]{4}\.){2}[0-9A-Fa-f]{4})",
+            r"(?P<uuid>[a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12})",
+            r"(?P<call>[\w.]*?)\(",
+            r"\b(?P<bool_true>True)\b|\b(?P<bool_false>False)\b|\b(?P<none>None)\b",
+            r"(?P<ellipsis>\.\.\.)",
+            r"(?P<number_complex>(?<!\w)(?:\-?[0-9]+\.?[0-9]*(?:e[-+]?\d+?)?)(?:[-+](?:[0-9]+\.?[0-9]*(?:e[-+]?\d+)?))?j)",
+            r"(?P<number>(?<!\w)\-?[0-9]+\.?[0-9]*(e[-+]?\d+?)?\b|0x[0-9a-fA-F]*)",
+            r"(?P<path>\B(/[-\w._+]+)*\/)(?P<filename>[-\w._+]*)?",
+            r"(?<![\\\w])(?P<str>b?'''.*?(?<!\\)'''|b?'.*?(?<!\\)'|b?\"\"\".*?(?<!\\)\"\"\"|b?\".*?(?<!\\)\")",
+            r"(?P<url>(file|https|http|ws|wss)://[-0-9a-zA-Z$_+!`(),.?/;:&=%#~@]*)",
+        ),
+    ]
+
+
+class JSONHighlighter(RegexHighlighter):
+    """Highlights JSON"""
+
+    # Captures the start and end of JSON strings, handling escaped quotes
+    JSON_STR = r"(?<![\\\w])(?P<str>b?\".*?(?<!\\)\")"
+    JSON_WHITESPACE = {" ", "\n", "\r", "\t"}
+
+    base_style = "json."
+    highlights = [
+        _combine_regex(
+            r"(?P<brace>[\{\[\(\)\]\}])",
+            r"\b(?P<bool_true>true)\b|\b(?P<bool_false>false)\b|\b(?P<null>null)\b",
+            r"(?P<number>(?<!\w)\-?[0-9]+\.?[0-9]*(e[\-\+]?\d+?)?\b|0x[0-9a-fA-F]*)",
+            JSON_STR,
+        ),
+    ]
+
+    def highlight(self, text: Text) -> None:
+        super().highlight(text)
+
+        # Additional work to handle highlighting JSON keys
+        plain = text.plain
+        append = text.spans.append
+        whitespace = self.JSON_WHITESPACE
+        for match in re.finditer(self.JSON_STR, plain):
+            start, end = match.span()
+            cursor = end
+            while cursor < len(plain):
+                char = plain[cursor]
+                cursor += 1
+                if char == ":":
+                    append(Span(start, end, "json.key"))
+                elif char in whitespace:
+                    continue
+                break
+
+
+class ISO8601Highlighter(RegexHighlighter):
+    """Highlights the ISO8601 date time strings.
+    Regex reference: https://www.oreilly.com/library/view/regular-expressions-cookbook/9781449327453/ch04s07.html
+    """
+
+    base_style = "iso8601."
+    highlights = [
+        #
+        # Dates
+        #
+        # Calendar month (e.g. 2008-08). The hyphen is required
+        r"^(?P<year>[0-9]{4})-(?P<month>1[0-2]|0[1-9])$",
+        # Calendar date w/o hyphens (e.g. 20080830)
+        r"^(?P<date>(?P<year>[0-9]{4})(?P<month>1[0-2]|0[1-9])(?P<day>3[01]|0[1-9]|[12][0-9]))$",
+        # Ordinal date (e.g. 2008-243). The hyphen is optional
+        r"^(?P<date>(?P<year>[0-9]{4})-?(?P<day>36[0-6]|3[0-5][0-9]|[12][0-9]{2}|0[1-9][0-9]|00[1-9]))$",
+        #
+        # Weeks
+        #
+        # Week of the year (e.g., 2008-W35). The hyphen is optional
+        r"^(?P<date>(?P<year>[0-9]{4})-?W(?P<week>5[0-3]|[1-4][0-9]|0[1-9]))$",
+        # Week date (e.g., 2008-W35-6). The hyphens are optional
+        r"^(?P<date>(?P<year>[0-9]{4})-?W(?P<week>5[0-3]|[1-4][0-9]|0[1-9])-?(?P<day>[1-7]))$",
+        #
+        # Times
+        #
+        # Hours and minutes (e.g., 17:21). The colon is optional
+        r"^(?P<time>(?P<hour>2[0-3]|[01][0-9]):?(?P<minute>[0-5][0-9]))$",
+        # Hours, minutes, and seconds w/o colons (e.g., 172159)
+        r"^(?P<time>(?P<hour>2[0-3]|[01][0-9])(?P<minute>[0-5][0-9])(?P<second>[0-5][0-9]))$",
+        # Time zone designator (e.g., Z, +07 or +07:00). The colons and the minutes are optional
+        r"^(?P<timezone>(Z|[+-](?:2[0-3]|[01][0-9])(?::?(?:[0-5][0-9]))?))$",
+        # Hours, minutes, and seconds with time zone designator (e.g., 17:21:59+07:00).
+        # All the colons are optional. The minutes in the time zone designator are also optional
+        r"^(?P<time>(?P<hour>2[0-3]|[01][0-9])(?P<minute>[0-5][0-9])(?P<second>[0-5][0-9]))(?P<timezone>Z|[+-](?:2[0-3]|[01][0-9])(?::?(?:[0-5][0-9]))?)$",
+        #
+        # Date and Time
+        #
+        # Calendar date with hours, minutes, and seconds (e.g., 2008-08-30 17:21:59 or 20080830 172159).
+        # A space is required between the date and the time. The hyphens and colons are optional.
+        # This regex matches dates and times that specify some hyphens or colons but omit others.
+        # This does not follow ISO 8601
+        r"^(?P<date>(?P<year>[0-9]{4})(?P<hyphen>-)?(?P<month>1[0-2]|0[1-9])(?(hyphen)-)(?P<day>3[01]|0[1-9]|[12][0-9])) (?P<time>(?P<hour>2[0-3]|[01][0-9])(?(hyphen):)(?P<minute>[0-5][0-9])(?(hyphen):)(?P<second>[0-5][0-9]))$",
+        #
+        # XML Schema dates and times
+        #
+        # Date, with optional time zone (e.g., 2008-08-30 or 2008-08-30+07:00).
+        # Hyphens are required. This is the XML Schema 'date' type
+        r"^(?P<date>(?P<year>-?(?:[1-9][0-9]*)?[0-9]{4})-(?P<month>1[0-2]|0[1-9])-(?P<day>3[01]|0[1-9]|[12][0-9]))(?P<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$",
+        # Time, with optional fractional seconds and time zone (e.g., 01:45:36 or 01:45:36.123+07:00).
+        # There is no limit on the number of digits for the fractional seconds. This is the XML Schema 'time' type
+        r"^(?P<time>(?P<hour>2[0-3]|[01][0-9]):(?P<minute>[0-5][0-9]):(?P<second>[0-5][0-9])(?P<frac>\.[0-9]+)?)(?P<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$",
+        # Date and time, with optional fractional seconds and time zone (e.g., 2008-08-30T01:45:36 or 2008-08-30T01:45:36.123Z).
+        # This is the XML Schema 'dateTime' type
+        r"^(?P<date>(?P<year>-?(?:[1-9][0-9]*)?[0-9]{4})-(?P<month>1[0-2]|0[1-9])-(?P<day>3[01]|0[1-9]|[12][0-9]))T(?P<time>(?P<hour>2[0-3]|[01][0-9]):(?P<minute>[0-5][0-9]):(?P<second>[0-5][0-9])(?P<ms>\.[0-9]+)?)(?P<timezone>Z|[+-](?:2[0-3]|[01][0-9]):[0-5][0-9])?$",
+    ]
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from .console import Console
+
+    console = Console()
+    console.print("[bold green]hello world![/bold green]")
+    console.print("'[bold green]hello world![/bold green]'")
+
+    console.print(" /foo")
+    console.print("/foo/")
+    console.print("/foo/bar")
+    console.print("foo/bar/baz")
+
+    console.print("/foo/bar/baz?foo=bar+egg&egg=baz")
+    console.print("/foo/bar/baz/")
+    console.print("/foo/bar/baz/egg")
+    console.print("/foo/bar/baz/egg.py")
+    console.print("/foo/bar/baz/egg.py word")
+    console.print(" /foo/bar/baz/egg.py word")
+    console.print("foo /foo/bar/baz/egg.py word")
+    console.print("foo /foo/bar/ba._++z/egg+.py word")
+    console.print("https://example.org?foo=bar#header")
+
+    console.print(1234567.34)
+    console.print(1 / 2)
+    console.print(-1 / 123123123123)
+
+    console.print(
+        "127.0.1.1 bar 192.168.1.4 2001:0db8:85a3:0000:0000:8a2e:0370:7334 foo"
+    )
+    import json
+
+    console.print_json(json.dumps(obj={"name": "apple", "count": 1}), indent=None)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/json.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/json.py
new file mode 100644
index 00000000..4087c79b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/json.py
@@ -0,0 +1,139 @@
+from pathlib import Path
+from json import loads, dumps
+from typing import Any, Callable, Optional, Union
+
+from .text import Text
+from .highlighter import JSONHighlighter, NullHighlighter
+
+
+class JSON:
+    """A renderable which pretty prints JSON.
+
+    Args:
+        json (str): JSON encoded data.
+        indent (Union[None, int, str], optional): Number of characters to indent by. Defaults to 2.
+        highlight (bool, optional): Enable highlighting. Defaults to True.
+        skip_keys (bool, optional): Skip keys not of a basic type. Defaults to False.
+        ensure_ascii (bool, optional): Escape all non-ascii characters. Defaults to False.
+        check_circular (bool, optional): Check for circular references. Defaults to True.
+        allow_nan (bool, optional): Allow NaN and Infinity values. Defaults to True.
+        default (Callable, optional): A callable that converts values that can not be encoded
+            in to something that can be JSON encoded. Defaults to None.
+        sort_keys (bool, optional): Sort dictionary keys. Defaults to False.
+    """
+
+    def __init__(
+        self,
+        json: str,
+        indent: Union[None, int, str] = 2,
+        highlight: bool = True,
+        skip_keys: bool = False,
+        ensure_ascii: bool = False,
+        check_circular: bool = True,
+        allow_nan: bool = True,
+        default: Optional[Callable[[Any], Any]] = None,
+        sort_keys: bool = False,
+    ) -> None:
+        data = loads(json)
+        json = dumps(
+            data,
+            indent=indent,
+            skipkeys=skip_keys,
+            ensure_ascii=ensure_ascii,
+            check_circular=check_circular,
+            allow_nan=allow_nan,
+            default=default,
+            sort_keys=sort_keys,
+        )
+        highlighter = JSONHighlighter() if highlight else NullHighlighter()
+        self.text = highlighter(json)
+        self.text.no_wrap = True
+        self.text.overflow = None
+
+    @classmethod
+    def from_data(
+        cls,
+        data: Any,
+        indent: Union[None, int, str] = 2,
+        highlight: bool = True,
+        skip_keys: bool = False,
+        ensure_ascii: bool = False,
+        check_circular: bool = True,
+        allow_nan: bool = True,
+        default: Optional[Callable[[Any], Any]] = None,
+        sort_keys: bool = False,
+    ) -> "JSON":
+        """Encodes a JSON object from arbitrary data.
+
+        Args:
+            data (Any): An object that may be encoded in to JSON
+            indent (Union[None, int, str], optional): Number of characters to indent by. Defaults to 2.
+            highlight (bool, optional): Enable highlighting. Defaults to True.
+            default (Callable, optional): Optional callable which will be called for objects that cannot be serialized. Defaults to None.
+            skip_keys (bool, optional): Skip keys not of a basic type. Defaults to False.
+            ensure_ascii (bool, optional): Escape all non-ascii characters. Defaults to False.
+            check_circular (bool, optional): Check for circular references. Defaults to True.
+            allow_nan (bool, optional): Allow NaN and Infinity values. Defaults to True.
+            default (Callable, optional): A callable that converts values that can not be encoded
+                in to something that can be JSON encoded. Defaults to None.
+            sort_keys (bool, optional): Sort dictionary keys. Defaults to False.
+
+        Returns:
+            JSON: New JSON object from the given data.
+        """
+        json_instance: "JSON" = cls.__new__(cls)
+        json = dumps(
+            data,
+            indent=indent,
+            skipkeys=skip_keys,
+            ensure_ascii=ensure_ascii,
+            check_circular=check_circular,
+            allow_nan=allow_nan,
+            default=default,
+            sort_keys=sort_keys,
+        )
+        highlighter = JSONHighlighter() if highlight else NullHighlighter()
+        json_instance.text = highlighter(json)
+        json_instance.text.no_wrap = True
+        json_instance.text.overflow = None
+        return json_instance
+
+    def __rich__(self) -> Text:
+        return self.text
+
+
+if __name__ == "__main__":
+    import argparse
+    import sys
+
+    parser = argparse.ArgumentParser(description="Pretty print json")
+    parser.add_argument(
+        "path",
+        metavar="PATH",
+        help="path to file, or - for stdin",
+    )
+    parser.add_argument(
+        "-i",
+        "--indent",
+        metavar="SPACES",
+        type=int,
+        help="Number of spaces in an indent",
+        default=2,
+    )
+    args = parser.parse_args()
+
+    from pip._vendor.rich.console import Console
+
+    console = Console()
+    error_console = Console(stderr=True)
+
+    try:
+        if args.path == "-":
+            json_data = sys.stdin.read()
+        else:
+            json_data = Path(args.path).read_text()
+    except Exception as error:
+        error_console.print(f"Unable to read {args.path!r}; {error}")
+        sys.exit(-1)
+
+    console.print(JSON(json_data, indent=args.indent), soft_wrap=True)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/jupyter.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/jupyter.py
new file mode 100644
index 00000000..22f4d716
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/jupyter.py
@@ -0,0 +1,101 @@
+from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Sequence
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.console import ConsoleRenderable
+
+from . import get_console
+from .segment import Segment
+from .terminal_theme import DEFAULT_TERMINAL_THEME
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.console import ConsoleRenderable
+
+JUPYTER_HTML_FORMAT = """\
+<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">{code}</pre>
+"""
+
+
+class JupyterRenderable:
+    """A shim to write html to Jupyter notebook."""
+
+    def __init__(self, html: str, text: str) -> None:
+        self.html = html
+        self.text = text
+
+    def _repr_mimebundle_(
+        self, include: Sequence[str], exclude: Sequence[str], **kwargs: Any
+    ) -> Dict[str, str]:
+        data = {"text/plain": self.text, "text/html": self.html}
+        if include:
+            data = {k: v for (k, v) in data.items() if k in include}
+        if exclude:
+            data = {k: v for (k, v) in data.items() if k not in exclude}
+        return data
+
+
+class JupyterMixin:
+    """Add to an Rich renderable to make it render in Jupyter notebook."""
+
+    __slots__ = ()
+
+    def _repr_mimebundle_(
+        self: "ConsoleRenderable",
+        include: Sequence[str],
+        exclude: Sequence[str],
+        **kwargs: Any,
+    ) -> Dict[str, str]:
+        console = get_console()
+        segments = list(console.render(self, console.options))
+        html = _render_segments(segments)
+        text = console._render_buffer(segments)
+        data = {"text/plain": text, "text/html": html}
+        if include:
+            data = {k: v for (k, v) in data.items() if k in include}
+        if exclude:
+            data = {k: v for (k, v) in data.items() if k not in exclude}
+        return data
+
+
+def _render_segments(segments: Iterable[Segment]) -> str:
+    def escape(text: str) -> str:
+        """Escape html."""
+        return text.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
+
+    fragments: List[str] = []
+    append_fragment = fragments.append
+    theme = DEFAULT_TERMINAL_THEME
+    for text, style, control in Segment.simplify(segments):
+        if control:
+            continue
+        text = escape(text)
+        if style:
+            rule = style.get_html_style(theme)
+            text = f'<span style="{rule}">{text}</span>' if rule else text
+            if style.link:
+                text = f'<a href="{style.link}" target="_blank">{text}</a>'
+        append_fragment(text)
+
+    code = "".join(fragments)
+    html = JUPYTER_HTML_FORMAT.format(code=code)
+
+    return html
+
+
+def display(segments: Iterable[Segment], text: str) -> None:
+    """Render segments to Jupyter."""
+    html = _render_segments(segments)
+    jupyter_renderable = JupyterRenderable(html, text)
+    try:
+        from IPython.display import display as ipython_display
+
+        ipython_display(jupyter_renderable)
+    except ModuleNotFoundError:
+        # Handle the case where the Console has force_jupyter=True,
+        # but IPython is not installed.
+        pass
+
+
+def print(*args: Any, **kwargs: Any) -> None:
+    """Proxy for Console print."""
+    console = get_console()
+    return console.print(*args, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/layout.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/layout.py
new file mode 100644
index 00000000..a6f1a31b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/layout.py
@@ -0,0 +1,442 @@
+from abc import ABC, abstractmethod
+from itertools import islice
+from operator import itemgetter
+from threading import RLock
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Sequence,
+    Tuple,
+    Union,
+)
+
+from ._ratio import ratio_resolve
+from .align import Align
+from .console import Console, ConsoleOptions, RenderableType, RenderResult
+from .highlighter import ReprHighlighter
+from .panel import Panel
+from .pretty import Pretty
+from .region import Region
+from .repr import Result, rich_repr
+from .segment import Segment
+from .style import StyleType
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.tree import Tree
+
+
+class LayoutRender(NamedTuple):
+    """An individual layout render."""
+
+    region: Region
+    render: List[List[Segment]]
+
+
+RegionMap = Dict["Layout", Region]
+RenderMap = Dict["Layout", LayoutRender]
+
+
+class LayoutError(Exception):
+    """Layout related error."""
+
+
+class NoSplitter(LayoutError):
+    """Requested splitter does not exist."""
+
+
+class _Placeholder:
+    """An internal renderable used as a Layout placeholder."""
+
+    highlighter = ReprHighlighter()
+
+    def __init__(self, layout: "Layout", style: StyleType = "") -> None:
+        self.layout = layout
+        self.style = style
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        width = options.max_width
+        height = options.height or options.size.height
+        layout = self.layout
+        title = (
+            f"{layout.name!r} ({width} x {height})"
+            if layout.name
+            else f"({width} x {height})"
+        )
+        yield Panel(
+            Align.center(Pretty(layout), vertical="middle"),
+            style=self.style,
+            title=self.highlighter(title),
+            border_style="blue",
+            height=height,
+        )
+
+
+class Splitter(ABC):
+    """Base class for a splitter."""
+
+    name: str = ""
+
+    @abstractmethod
+    def get_tree_icon(self) -> str:
+        """Get the icon (emoji) used in layout.tree"""
+
+    @abstractmethod
+    def divide(
+        self, children: Sequence["Layout"], region: Region
+    ) -> Iterable[Tuple["Layout", Region]]:
+        """Divide a region amongst several child layouts.
+
+        Args:
+            children (Sequence(Layout)): A number of child layouts.
+            region (Region): A rectangular region to divide.
+        """
+
+
+class RowSplitter(Splitter):
+    """Split a layout region in to rows."""
+
+    name = "row"
+
+    def get_tree_icon(self) -> str:
+        return "[layout.tree.row]⬌"
+
+    def divide(
+        self, children: Sequence["Layout"], region: Region
+    ) -> Iterable[Tuple["Layout", Region]]:
+        x, y, width, height = region
+        render_widths = ratio_resolve(width, children)
+        offset = 0
+        _Region = Region
+        for child, child_width in zip(children, render_widths):
+            yield child, _Region(x + offset, y, child_width, height)
+            offset += child_width
+
+
+class ColumnSplitter(Splitter):
+    """Split a layout region in to columns."""
+
+    name = "column"
+
+    def get_tree_icon(self) -> str:
+        return "[layout.tree.column]⬍"
+
+    def divide(
+        self, children: Sequence["Layout"], region: Region
+    ) -> Iterable[Tuple["Layout", Region]]:
+        x, y, width, height = region
+        render_heights = ratio_resolve(height, children)
+        offset = 0
+        _Region = Region
+        for child, child_height in zip(children, render_heights):
+            yield child, _Region(x, y + offset, width, child_height)
+            offset += child_height
+
+
+@rich_repr
+class Layout:
+    """A renderable to divide a fixed height in to rows or columns.
+
+    Args:
+        renderable (RenderableType, optional): Renderable content, or None for placeholder. Defaults to None.
+        name (str, optional): Optional identifier for Layout. Defaults to None.
+        size (int, optional): Optional fixed size of layout. Defaults to None.
+        minimum_size (int, optional): Minimum size of layout. Defaults to 1.
+        ratio (int, optional): Optional ratio for flexible layout. Defaults to 1.
+        visible (bool, optional): Visibility of layout. Defaults to True.
+    """
+
+    splitters = {"row": RowSplitter, "column": ColumnSplitter}
+
+    def __init__(
+        self,
+        renderable: Optional[RenderableType] = None,
+        *,
+        name: Optional[str] = None,
+        size: Optional[int] = None,
+        minimum_size: int = 1,
+        ratio: int = 1,
+        visible: bool = True,
+    ) -> None:
+        self._renderable = renderable or _Placeholder(self)
+        self.size = size
+        self.minimum_size = minimum_size
+        self.ratio = ratio
+        self.name = name
+        self.visible = visible
+        self.splitter: Splitter = self.splitters["column"]()
+        self._children: List[Layout] = []
+        self._render_map: RenderMap = {}
+        self._lock = RLock()
+
+    def __rich_repr__(self) -> Result:
+        yield "name", self.name, None
+        yield "size", self.size, None
+        yield "minimum_size", self.minimum_size, 1
+        yield "ratio", self.ratio, 1
+
+    @property
+    def renderable(self) -> RenderableType:
+        """Layout renderable."""
+        return self if self._children else self._renderable
+
+    @property
+    def children(self) -> List["Layout"]:
+        """Gets (visible) layout children."""
+        return [child for child in self._children if child.visible]
+
+    @property
+    def map(self) -> RenderMap:
+        """Get a map of the last render."""
+        return self._render_map
+
+    def get(self, name: str) -> Optional["Layout"]:
+        """Get a named layout, or None if it doesn't exist.
+
+        Args:
+            name (str): Name of layout.
+
+        Returns:
+            Optional[Layout]: Layout instance or None if no layout was found.
+        """
+        if self.name == name:
+            return self
+        else:
+            for child in self._children:
+                named_layout = child.get(name)
+                if named_layout is not None:
+                    return named_layout
+        return None
+
+    def __getitem__(self, name: str) -> "Layout":
+        layout = self.get(name)
+        if layout is None:
+            raise KeyError(f"No layout with name {name!r}")
+        return layout
+
+    @property
+    def tree(self) -> "Tree":
+        """Get a tree renderable to show layout structure."""
+        from pip._vendor.rich.styled import Styled
+        from pip._vendor.rich.table import Table
+        from pip._vendor.rich.tree import Tree
+
+        def summary(layout: "Layout") -> Table:
+            icon = layout.splitter.get_tree_icon()
+
+            table = Table.grid(padding=(0, 1, 0, 0))
+
+            text: RenderableType = (
+                Pretty(layout) if layout.visible else Styled(Pretty(layout), "dim")
+            )
+            table.add_row(icon, text)
+            _summary = table
+            return _summary
+
+        layout = self
+        tree = Tree(
+            summary(layout),
+            guide_style=f"layout.tree.{layout.splitter.name}",
+            highlight=True,
+        )
+
+        def recurse(tree: "Tree", layout: "Layout") -> None:
+            for child in layout._children:
+                recurse(
+                    tree.add(
+                        summary(child),
+                        guide_style=f"layout.tree.{child.splitter.name}",
+                    ),
+                    child,
+                )
+
+        recurse(tree, self)
+        return tree
+
+    def split(
+        self,
+        *layouts: Union["Layout", RenderableType],
+        splitter: Union[Splitter, str] = "column",
+    ) -> None:
+        """Split the layout in to multiple sub-layouts.
+
+        Args:
+            *layouts (Layout): Positional arguments should be (sub) Layout instances.
+            splitter (Union[Splitter, str]): Splitter instance or name of splitter.
+        """
+        _layouts = [
+            layout if isinstance(layout, Layout) else Layout(layout)
+            for layout in layouts
+        ]
+        try:
+            self.splitter = (
+                splitter
+                if isinstance(splitter, Splitter)
+                else self.splitters[splitter]()
+            )
+        except KeyError:
+            raise NoSplitter(f"No splitter called {splitter!r}")
+        self._children[:] = _layouts
+
+    def add_split(self, *layouts: Union["Layout", RenderableType]) -> None:
+        """Add a new layout(s) to existing split.
+
+        Args:
+            *layouts (Union[Layout, RenderableType]): Positional arguments should be renderables or (sub) Layout instances.
+
+        """
+        _layouts = (
+            layout if isinstance(layout, Layout) else Layout(layout)
+            for layout in layouts
+        )
+        self._children.extend(_layouts)
+
+    def split_row(self, *layouts: Union["Layout", RenderableType]) -> None:
+        """Split the layout in to a row (layouts side by side).
+
+        Args:
+            *layouts (Layout): Positional arguments should be (sub) Layout instances.
+        """
+        self.split(*layouts, splitter="row")
+
+    def split_column(self, *layouts: Union["Layout", RenderableType]) -> None:
+        """Split the layout in to a column (layouts stacked on top of each other).
+
+        Args:
+            *layouts (Layout): Positional arguments should be (sub) Layout instances.
+        """
+        self.split(*layouts, splitter="column")
+
+    def unsplit(self) -> None:
+        """Reset splits to initial state."""
+        del self._children[:]
+
+    def update(self, renderable: RenderableType) -> None:
+        """Update renderable.
+
+        Args:
+            renderable (RenderableType): New renderable object.
+        """
+        with self._lock:
+            self._renderable = renderable
+
+    def refresh_screen(self, console: "Console", layout_name: str) -> None:
+        """Refresh a sub-layout.
+
+        Args:
+            console (Console): Console instance where Layout is to be rendered.
+            layout_name (str): Name of layout.
+        """
+        with self._lock:
+            layout = self[layout_name]
+            region, _lines = self._render_map[layout]
+            (x, y, width, height) = region
+            lines = console.render_lines(
+                layout, console.options.update_dimensions(width, height)
+            )
+            self._render_map[layout] = LayoutRender(region, lines)
+            console.update_screen_lines(lines, x, y)
+
+    def _make_region_map(self, width: int, height: int) -> RegionMap:
+        """Create a dict that maps layout on to Region."""
+        stack: List[Tuple[Layout, Region]] = [(self, Region(0, 0, width, height))]
+        push = stack.append
+        pop = stack.pop
+        layout_regions: List[Tuple[Layout, Region]] = []
+        append_layout_region = layout_regions.append
+        while stack:
+            append_layout_region(pop())
+            layout, region = layout_regions[-1]
+            children = layout.children
+            if children:
+                for child_and_region in layout.splitter.divide(children, region):
+                    push(child_and_region)
+
+        region_map = {
+            layout: region
+            for layout, region in sorted(layout_regions, key=itemgetter(1))
+        }
+        return region_map
+
+    def render(self, console: Console, options: ConsoleOptions) -> RenderMap:
+        """Render the sub_layouts.
+
+        Args:
+            console (Console): Console instance.
+            options (ConsoleOptions): Console options.
+
+        Returns:
+            RenderMap: A dict that maps Layout on to a tuple of Region, lines
+        """
+        render_width = options.max_width
+        render_height = options.height or console.height
+        region_map = self._make_region_map(render_width, render_height)
+        layout_regions = [
+            (layout, region)
+            for layout, region in region_map.items()
+            if not layout.children
+        ]
+        render_map: Dict["Layout", "LayoutRender"] = {}
+        render_lines = console.render_lines
+        update_dimensions = options.update_dimensions
+
+        for layout, region in layout_regions:
+            lines = render_lines(
+                layout.renderable, update_dimensions(region.width, region.height)
+            )
+            render_map[layout] = LayoutRender(region, lines)
+        return render_map
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        with self._lock:
+            width = options.max_width or console.width
+            height = options.height or console.height
+            render_map = self.render(console, options.update_dimensions(width, height))
+            self._render_map = render_map
+            layout_lines: List[List[Segment]] = [[] for _ in range(height)]
+            _islice = islice
+            for region, lines in render_map.values():
+                _x, y, _layout_width, layout_height = region
+                for row, line in zip(
+                    _islice(layout_lines, y, y + layout_height), lines
+                ):
+                    row.extend(line)
+
+            new_line = Segment.line()
+            for layout_row in layout_lines:
+                yield from layout_row
+                yield new_line
+
+
+if __name__ == "__main__":
+    from pip._vendor.rich.console import Console
+
+    console = Console()
+    layout = Layout()
+
+    layout.split_column(
+        Layout(name="header", size=3),
+        Layout(ratio=1, name="main"),
+        Layout(size=10, name="footer"),
+    )
+
+    layout["main"].split_row(Layout(name="side"), Layout(name="body", ratio=2))
+
+    layout["body"].split_row(Layout(name="content", ratio=2), Layout(name="s2"))
+
+    layout["s2"].split_column(
+        Layout(name="top"), Layout(name="middle"), Layout(name="bottom")
+    )
+
+    layout["side"].split_column(Layout(layout.tree, name="left1"), Layout(name="left2"))
+
+    layout["content"].update("foo")
+
+    console.print(layout)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live.py
new file mode 100644
index 00000000..8738cf09
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live.py
@@ -0,0 +1,375 @@
+import sys
+from threading import Event, RLock, Thread
+from types import TracebackType
+from typing import IO, Any, Callable, List, Optional, TextIO, Type, cast
+
+from . import get_console
+from .console import Console, ConsoleRenderable, RenderableType, RenderHook
+from .control import Control
+from .file_proxy import FileProxy
+from .jupyter import JupyterMixin
+from .live_render import LiveRender, VerticalOverflowMethod
+from .screen import Screen
+from .text import Text
+
+
+class _RefreshThread(Thread):
+    """A thread that calls refresh() at regular intervals."""
+
+    def __init__(self, live: "Live", refresh_per_second: float) -> None:
+        self.live = live
+        self.refresh_per_second = refresh_per_second
+        self.done = Event()
+        super().__init__(daemon=True)
+
+    def stop(self) -> None:
+        self.done.set()
+
+    def run(self) -> None:
+        while not self.done.wait(1 / self.refresh_per_second):
+            with self.live._lock:
+                if not self.done.is_set():
+                    self.live.refresh()
+
+
+class Live(JupyterMixin, RenderHook):
+    """Renders an auto-updating live display of any given renderable.
+
+    Args:
+        renderable (RenderableType, optional): The renderable to live display. Defaults to displaying nothing.
+        console (Console, optional): Optional Console instance. Defaults to an internal Console instance writing to stdout.
+        screen (bool, optional): Enable alternate screen mode. Defaults to False.
+        auto_refresh (bool, optional): Enable auto refresh. If disabled, you will need to call `refresh()` or `update()` with refresh flag. Defaults to True
+        refresh_per_second (float, optional): Number of times per second to refresh the live display. Defaults to 4.
+        transient (bool, optional): Clear the renderable on exit (has no effect when screen=True). Defaults to False.
+        redirect_stdout (bool, optional): Enable redirection of stdout, so ``print`` may be used. Defaults to True.
+        redirect_stderr (bool, optional): Enable redirection of stderr. Defaults to True.
+        vertical_overflow (VerticalOverflowMethod, optional): How to handle renderable when it is too tall for the console. Defaults to "ellipsis".
+        get_renderable (Callable[[], RenderableType], optional): Optional callable to get renderable. Defaults to None.
+    """
+
+    def __init__(
+        self,
+        renderable: Optional[RenderableType] = None,
+        *,
+        console: Optional[Console] = None,
+        screen: bool = False,
+        auto_refresh: bool = True,
+        refresh_per_second: float = 4,
+        transient: bool = False,
+        redirect_stdout: bool = True,
+        redirect_stderr: bool = True,
+        vertical_overflow: VerticalOverflowMethod = "ellipsis",
+        get_renderable: Optional[Callable[[], RenderableType]] = None,
+    ) -> None:
+        assert refresh_per_second > 0, "refresh_per_second must be > 0"
+        self._renderable = renderable
+        self.console = console if console is not None else get_console()
+        self._screen = screen
+        self._alt_screen = False
+
+        self._redirect_stdout = redirect_stdout
+        self._redirect_stderr = redirect_stderr
+        self._restore_stdout: Optional[IO[str]] = None
+        self._restore_stderr: Optional[IO[str]] = None
+
+        self._lock = RLock()
+        self.ipy_widget: Optional[Any] = None
+        self.auto_refresh = auto_refresh
+        self._started: bool = False
+        self.transient = True if screen else transient
+
+        self._refresh_thread: Optional[_RefreshThread] = None
+        self.refresh_per_second = refresh_per_second
+
+        self.vertical_overflow = vertical_overflow
+        self._get_renderable = get_renderable
+        self._live_render = LiveRender(
+            self.get_renderable(), vertical_overflow=vertical_overflow
+        )
+
+    @property
+    def is_started(self) -> bool:
+        """Check if live display has been started."""
+        return self._started
+
+    def get_renderable(self) -> RenderableType:
+        renderable = (
+            self._get_renderable()
+            if self._get_renderable is not None
+            else self._renderable
+        )
+        return renderable or ""
+
+    def start(self, refresh: bool = False) -> None:
+        """Start live rendering display.
+
+        Args:
+            refresh (bool, optional): Also refresh. Defaults to False.
+        """
+        with self._lock:
+            if self._started:
+                return
+            self.console.set_live(self)
+            self._started = True
+            if self._screen:
+                self._alt_screen = self.console.set_alt_screen(True)
+            self.console.show_cursor(False)
+            self._enable_redirect_io()
+            self.console.push_render_hook(self)
+            if refresh:
+                try:
+                    self.refresh()
+                except Exception:
+                    # If refresh fails, we want to stop the redirection of sys.stderr,
+                    # so the error stacktrace is properly displayed in the terminal.
+                    # (or, if the code that calls Rich captures the exception and wants to display something,
+                    # let this be displayed in the terminal).
+                    self.stop()
+                    raise
+            if self.auto_refresh:
+                self._refresh_thread = _RefreshThread(self, self.refresh_per_second)
+                self._refresh_thread.start()
+
+    def stop(self) -> None:
+        """Stop live rendering display."""
+        with self._lock:
+            if not self._started:
+                return
+            self.console.clear_live()
+            self._started = False
+
+            if self.auto_refresh and self._refresh_thread is not None:
+                self._refresh_thread.stop()
+                self._refresh_thread = None
+            # allow it to fully render on the last even if overflow
+            self.vertical_overflow = "visible"
+            with self.console:
+                try:
+                    if not self._alt_screen and not self.console.is_jupyter:
+                        self.refresh()
+                finally:
+                    self._disable_redirect_io()
+                    self.console.pop_render_hook()
+                    if not self._alt_screen and self.console.is_terminal:
+                        self.console.line()
+                    self.console.show_cursor(True)
+                    if self._alt_screen:
+                        self.console.set_alt_screen(False)
+
+                    if self.transient and not self._alt_screen:
+                        self.console.control(self._live_render.restore_cursor())
+                    if self.ipy_widget is not None and self.transient:
+                        self.ipy_widget.close()  # pragma: no cover
+
+    def __enter__(self) -> "Live":
+        self.start(refresh=self._renderable is not None)
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.stop()
+
+    def _enable_redirect_io(self) -> None:
+        """Enable redirecting of stdout / stderr."""
+        if self.console.is_terminal or self.console.is_jupyter:
+            if self._redirect_stdout and not isinstance(sys.stdout, FileProxy):
+                self._restore_stdout = sys.stdout
+                sys.stdout = cast("TextIO", FileProxy(self.console, sys.stdout))
+            if self._redirect_stderr and not isinstance(sys.stderr, FileProxy):
+                self._restore_stderr = sys.stderr
+                sys.stderr = cast("TextIO", FileProxy(self.console, sys.stderr))
+
+    def _disable_redirect_io(self) -> None:
+        """Disable redirecting of stdout / stderr."""
+        if self._restore_stdout:
+            sys.stdout = cast("TextIO", self._restore_stdout)
+            self._restore_stdout = None
+        if self._restore_stderr:
+            sys.stderr = cast("TextIO", self._restore_stderr)
+            self._restore_stderr = None
+
+    @property
+    def renderable(self) -> RenderableType:
+        """Get the renderable that is being displayed
+
+        Returns:
+            RenderableType: Displayed renderable.
+        """
+        renderable = self.get_renderable()
+        return Screen(renderable) if self._alt_screen else renderable
+
+    def update(self, renderable: RenderableType, *, refresh: bool = False) -> None:
+        """Update the renderable that is being displayed
+
+        Args:
+            renderable (RenderableType): New renderable to use.
+            refresh (bool, optional): Refresh the display. Defaults to False.
+        """
+        if isinstance(renderable, str):
+            renderable = self.console.render_str(renderable)
+        with self._lock:
+            self._renderable = renderable
+            if refresh:
+                self.refresh()
+
+    def refresh(self) -> None:
+        """Update the display of the Live Render."""
+        with self._lock:
+            self._live_render.set_renderable(self.renderable)
+            if self.console.is_jupyter:  # pragma: no cover
+                try:
+                    from IPython.display import display
+                    from ipywidgets import Output
+                except ImportError:
+                    import warnings
+
+                    warnings.warn('install "ipywidgets" for Jupyter support')
+                else:
+                    if self.ipy_widget is None:
+                        self.ipy_widget = Output()
+                        display(self.ipy_widget)
+
+                    with self.ipy_widget:
+                        self.ipy_widget.clear_output(wait=True)
+                        self.console.print(self._live_render.renderable)
+            elif self.console.is_terminal and not self.console.is_dumb_terminal:
+                with self.console:
+                    self.console.print(Control())
+            elif (
+                not self._started and not self.transient
+            ):  # if it is finished allow files or dumb-terminals to see final result
+                with self.console:
+                    self.console.print(Control())
+
+    def process_renderables(
+        self, renderables: List[ConsoleRenderable]
+    ) -> List[ConsoleRenderable]:
+        """Process renderables to restore cursor and display progress."""
+        self._live_render.vertical_overflow = self.vertical_overflow
+        if self.console.is_interactive:
+            # lock needs acquiring as user can modify live_render renderable at any time unlike in Progress.
+            with self._lock:
+                reset = (
+                    Control.home()
+                    if self._alt_screen
+                    else self._live_render.position_cursor()
+                )
+                renderables = [reset, *renderables, self._live_render]
+        elif (
+            not self._started and not self.transient
+        ):  # if it is finished render the final output for files or dumb_terminals
+            renderables = [*renderables, self._live_render]
+
+        return renderables
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import random
+    import time
+    from itertools import cycle
+    from typing import Dict, List, Tuple
+
+    from .align import Align
+    from .console import Console
+    from .live import Live as Live
+    from .panel import Panel
+    from .rule import Rule
+    from .syntax import Syntax
+    from .table import Table
+
+    console = Console()
+
+    syntax = Syntax(
+        '''def loop_last(values: Iterable[T]) -> Iterable[Tuple[bool, T]]:
+    """Iterate and generate a tuple with a flag for last value."""
+    iter_values = iter(values)
+    try:
+        previous_value = next(iter_values)
+    except StopIteration:
+        return
+    for value in iter_values:
+        yield False, previous_value
+        previous_value = value
+    yield True, previous_value''',
+        "python",
+        line_numbers=True,
+    )
+
+    table = Table("foo", "bar", "baz")
+    table.add_row("1", "2", "3")
+
+    progress_renderables = [
+        "You can make the terminal shorter and taller to see the live table hide"
+        "Text may be printed while the progress bars are rendering.",
+        Panel("In fact, [i]any[/i] renderable will work"),
+        "Such as [magenta]tables[/]...",
+        table,
+        "Pretty printed structures...",
+        {"type": "example", "text": "Pretty printed"},
+        "Syntax...",
+        syntax,
+        Rule("Give it a try!"),
+    ]
+
+    examples = cycle(progress_renderables)
+
+    exchanges = [
+        "SGD",
+        "MYR",
+        "EUR",
+        "USD",
+        "AUD",
+        "JPY",
+        "CNH",
+        "HKD",
+        "CAD",
+        "INR",
+        "DKK",
+        "GBP",
+        "RUB",
+        "NZD",
+        "MXN",
+        "IDR",
+        "TWD",
+        "THB",
+        "VND",
+    ]
+    with Live(console=console) as live_table:
+        exchange_rate_dict: Dict[Tuple[str, str], float] = {}
+
+        for index in range(100):
+            select_exchange = exchanges[index % len(exchanges)]
+
+            for exchange in exchanges:
+                if exchange == select_exchange:
+                    continue
+                time.sleep(0.4)
+                if random.randint(0, 10) < 1:
+                    console.log(next(examples))
+                exchange_rate_dict[(select_exchange, exchange)] = 200 / (
+                    (random.random() * 320) + 1
+                )
+                if len(exchange_rate_dict) > len(exchanges) - 1:
+                    exchange_rate_dict.pop(list(exchange_rate_dict.keys())[0])
+                table = Table(title="Exchange Rates")
+
+                table.add_column("Source Currency")
+                table.add_column("Destination Currency")
+                table.add_column("Exchange Rate")
+
+                for (source, dest), exchange_rate in exchange_rate_dict.items():
+                    table.add_row(
+                        source,
+                        dest,
+                        Text(
+                            f"{exchange_rate:.4f}",
+                            style="red" if exchange_rate < 1.0 else "green",
+                        ),
+                    )
+
+                live_table.update(Align.center(table))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live_render.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live_render.py
new file mode 100644
index 00000000..e20745df
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/live_render.py
@@ -0,0 +1,112 @@
+import sys
+from typing import Optional, Tuple
+
+if sys.version_info >= (3, 8):
+    from typing import Literal
+else:
+    from pip._vendor.typing_extensions import Literal  # pragma: no cover
+
+
+from ._loop import loop_last
+from .console import Console, ConsoleOptions, RenderableType, RenderResult
+from .control import Control
+from .segment import ControlType, Segment
+from .style import StyleType
+from .text import Text
+
+VerticalOverflowMethod = Literal["crop", "ellipsis", "visible"]
+
+
+class LiveRender:
+    """Creates a renderable that may be updated.
+
+    Args:
+        renderable (RenderableType): Any renderable object.
+        style (StyleType, optional): An optional style to apply to the renderable. Defaults to "".
+    """
+
+    def __init__(
+        self,
+        renderable: RenderableType,
+        style: StyleType = "",
+        vertical_overflow: VerticalOverflowMethod = "ellipsis",
+    ) -> None:
+        self.renderable = renderable
+        self.style = style
+        self.vertical_overflow = vertical_overflow
+        self._shape: Optional[Tuple[int, int]] = None
+
+    def set_renderable(self, renderable: RenderableType) -> None:
+        """Set a new renderable.
+
+        Args:
+            renderable (RenderableType): Any renderable object, including str.
+        """
+        self.renderable = renderable
+
+    def position_cursor(self) -> Control:
+        """Get control codes to move cursor to beginning of live render.
+
+        Returns:
+            Control: A control instance that may be printed.
+        """
+        if self._shape is not None:
+            _, height = self._shape
+            return Control(
+                ControlType.CARRIAGE_RETURN,
+                (ControlType.ERASE_IN_LINE, 2),
+                *(
+                    (
+                        (ControlType.CURSOR_UP, 1),
+                        (ControlType.ERASE_IN_LINE, 2),
+                    )
+                    * (height - 1)
+                )
+            )
+        return Control()
+
+    def restore_cursor(self) -> Control:
+        """Get control codes to clear the render and restore the cursor to its previous position.
+
+        Returns:
+            Control: A Control instance that may be printed.
+        """
+        if self._shape is not None:
+            _, height = self._shape
+            return Control(
+                ControlType.CARRIAGE_RETURN,
+                *((ControlType.CURSOR_UP, 1), (ControlType.ERASE_IN_LINE, 2)) * height
+            )
+        return Control()
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        renderable = self.renderable
+        style = console.get_style(self.style)
+        lines = console.render_lines(renderable, options, style=style, pad=False)
+        shape = Segment.get_shape(lines)
+
+        _, height = shape
+        if height > options.size.height:
+            if self.vertical_overflow == "crop":
+                lines = lines[: options.size.height]
+                shape = Segment.get_shape(lines)
+            elif self.vertical_overflow == "ellipsis":
+                lines = lines[: (options.size.height - 1)]
+                overflow_text = Text(
+                    "...",
+                    overflow="crop",
+                    justify="center",
+                    end="",
+                    style="live.ellipsis",
+                )
+                lines.append(list(console.render(overflow_text)))
+                shape = Segment.get_shape(lines)
+        self._shape = shape
+
+        new_line = Segment.line()
+        for last, line in loop_last(lines):
+            yield from line
+            if not last:
+                yield new_line
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/logging.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/logging.py
new file mode 100644
index 00000000..ff8d5d95
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/logging.py
@@ -0,0 +1,297 @@
+import logging
+from datetime import datetime
+from logging import Handler, LogRecord
+from pathlib import Path
+from types import ModuleType
+from typing import ClassVar, Iterable, List, Optional, Type, Union
+
+from pip._vendor.rich._null_file import NullFile
+
+from . import get_console
+from ._log_render import FormatTimeCallable, LogRender
+from .console import Console, ConsoleRenderable
+from .highlighter import Highlighter, ReprHighlighter
+from .text import Text
+from .traceback import Traceback
+
+
+class RichHandler(Handler):
+    """A logging handler that renders output with Rich. The time / level / message and file are displayed in columns.
+    The level is color coded, and the message is syntax highlighted.
+
+    Note:
+        Be careful when enabling console markup in log messages if you have configured logging for libraries not
+        under your control. If a dependency writes messages containing square brackets, it may not produce the intended output.
+
+    Args:
+        level (Union[int, str], optional): Log level. Defaults to logging.NOTSET.
+        console (:class:`~rich.console.Console`, optional): Optional console instance to write logs.
+            Default will use a global console instance writing to stdout.
+        show_time (bool, optional): Show a column for the time. Defaults to True.
+        omit_repeated_times (bool, optional): Omit repetition of the same time. Defaults to True.
+        show_level (bool, optional): Show a column for the level. Defaults to True.
+        show_path (bool, optional): Show the path to the original log call. Defaults to True.
+        enable_link_path (bool, optional): Enable terminal link of path column to file. Defaults to True.
+        highlighter (Highlighter, optional): Highlighter to style log messages, or None to use ReprHighlighter. Defaults to None.
+        markup (bool, optional): Enable console markup in log messages. Defaults to False.
+        rich_tracebacks (bool, optional): Enable rich tracebacks with syntax highlighting and formatting. Defaults to False.
+        tracebacks_width (Optional[int], optional): Number of characters used to render tracebacks, or None for full width. Defaults to None.
+        tracebacks_code_width (int, optional): Number of code characters used to render tracebacks, or None for full width. Defaults to 88.
+        tracebacks_extra_lines (int, optional): Additional lines of code to render tracebacks, or None for full width. Defaults to None.
+        tracebacks_theme (str, optional): Override pygments theme used in traceback.
+        tracebacks_word_wrap (bool, optional): Enable word wrapping of long tracebacks lines. Defaults to True.
+        tracebacks_show_locals (bool, optional): Enable display of locals in tracebacks. Defaults to False.
+        tracebacks_suppress (Sequence[Union[str, ModuleType]]): Optional sequence of modules or paths to exclude from traceback.
+        tracebacks_max_frames (int, optional): Optional maximum number of frames returned by traceback.
+        locals_max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to 10.
+        locals_max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to 80.
+        log_time_format (Union[str, TimeFormatterCallable], optional): If ``log_time`` is enabled, either string for strftime or callable that formats the time. Defaults to "[%x %X] ".
+        keywords (List[str], optional): List of words to highlight instead of ``RichHandler.KEYWORDS``.
+    """
+
+    KEYWORDS: ClassVar[Optional[List[str]]] = [
+        "GET",
+        "POST",
+        "HEAD",
+        "PUT",
+        "DELETE",
+        "OPTIONS",
+        "TRACE",
+        "PATCH",
+    ]
+    HIGHLIGHTER_CLASS: ClassVar[Type[Highlighter]] = ReprHighlighter
+
+    def __init__(
+        self,
+        level: Union[int, str] = logging.NOTSET,
+        console: Optional[Console] = None,
+        *,
+        show_time: bool = True,
+        omit_repeated_times: bool = True,
+        show_level: bool = True,
+        show_path: bool = True,
+        enable_link_path: bool = True,
+        highlighter: Optional[Highlighter] = None,
+        markup: bool = False,
+        rich_tracebacks: bool = False,
+        tracebacks_width: Optional[int] = None,
+        tracebacks_code_width: int = 88,
+        tracebacks_extra_lines: int = 3,
+        tracebacks_theme: Optional[str] = None,
+        tracebacks_word_wrap: bool = True,
+        tracebacks_show_locals: bool = False,
+        tracebacks_suppress: Iterable[Union[str, ModuleType]] = (),
+        tracebacks_max_frames: int = 100,
+        locals_max_length: int = 10,
+        locals_max_string: int = 80,
+        log_time_format: Union[str, FormatTimeCallable] = "[%x %X]",
+        keywords: Optional[List[str]] = None,
+    ) -> None:
+        super().__init__(level=level)
+        self.console = console or get_console()
+        self.highlighter = highlighter or self.HIGHLIGHTER_CLASS()
+        self._log_render = LogRender(
+            show_time=show_time,
+            show_level=show_level,
+            show_path=show_path,
+            time_format=log_time_format,
+            omit_repeated_times=omit_repeated_times,
+            level_width=None,
+        )
+        self.enable_link_path = enable_link_path
+        self.markup = markup
+        self.rich_tracebacks = rich_tracebacks
+        self.tracebacks_width = tracebacks_width
+        self.tracebacks_extra_lines = tracebacks_extra_lines
+        self.tracebacks_theme = tracebacks_theme
+        self.tracebacks_word_wrap = tracebacks_word_wrap
+        self.tracebacks_show_locals = tracebacks_show_locals
+        self.tracebacks_suppress = tracebacks_suppress
+        self.tracebacks_max_frames = tracebacks_max_frames
+        self.tracebacks_code_width = tracebacks_code_width
+        self.locals_max_length = locals_max_length
+        self.locals_max_string = locals_max_string
+        self.keywords = keywords
+
+    def get_level_text(self, record: LogRecord) -> Text:
+        """Get the level name from the record.
+
+        Args:
+            record (LogRecord): LogRecord instance.
+
+        Returns:
+            Text: A tuple of the style and level name.
+        """
+        level_name = record.levelname
+        level_text = Text.styled(
+            level_name.ljust(8), f"logging.level.{level_name.lower()}"
+        )
+        return level_text
+
+    def emit(self, record: LogRecord) -> None:
+        """Invoked by logging."""
+        message = self.format(record)
+        traceback = None
+        if (
+            self.rich_tracebacks
+            and record.exc_info
+            and record.exc_info != (None, None, None)
+        ):
+            exc_type, exc_value, exc_traceback = record.exc_info
+            assert exc_type is not None
+            assert exc_value is not None
+            traceback = Traceback.from_exception(
+                exc_type,
+                exc_value,
+                exc_traceback,
+                width=self.tracebacks_width,
+                code_width=self.tracebacks_code_width,
+                extra_lines=self.tracebacks_extra_lines,
+                theme=self.tracebacks_theme,
+                word_wrap=self.tracebacks_word_wrap,
+                show_locals=self.tracebacks_show_locals,
+                locals_max_length=self.locals_max_length,
+                locals_max_string=self.locals_max_string,
+                suppress=self.tracebacks_suppress,
+                max_frames=self.tracebacks_max_frames,
+            )
+            message = record.getMessage()
+            if self.formatter:
+                record.message = record.getMessage()
+                formatter = self.formatter
+                if hasattr(formatter, "usesTime") and formatter.usesTime():
+                    record.asctime = formatter.formatTime(record, formatter.datefmt)
+                message = formatter.formatMessage(record)
+
+        message_renderable = self.render_message(record, message)
+        log_renderable = self.render(
+            record=record, traceback=traceback, message_renderable=message_renderable
+        )
+        if isinstance(self.console.file, NullFile):
+            # Handles pythonw, where stdout/stderr are null, and we return NullFile
+            # instance from Console.file. In this case, we still want to make a log record
+            # even though we won't be writing anything to a file.
+            self.handleError(record)
+        else:
+            try:
+                self.console.print(log_renderable)
+            except Exception:
+                self.handleError(record)
+
+    def render_message(self, record: LogRecord, message: str) -> "ConsoleRenderable":
+        """Render message text in to Text.
+
+        Args:
+            record (LogRecord): logging Record.
+            message (str): String containing log message.
+
+        Returns:
+            ConsoleRenderable: Renderable to display log message.
+        """
+        use_markup = getattr(record, "markup", self.markup)
+        message_text = Text.from_markup(message) if use_markup else Text(message)
+
+        highlighter = getattr(record, "highlighter", self.highlighter)
+        if highlighter:
+            message_text = highlighter(message_text)
+
+        if self.keywords is None:
+            self.keywords = self.KEYWORDS
+
+        if self.keywords:
+            message_text.highlight_words(self.keywords, "logging.keyword")
+
+        return message_text
+
+    def render(
+        self,
+        *,
+        record: LogRecord,
+        traceback: Optional[Traceback],
+        message_renderable: "ConsoleRenderable",
+    ) -> "ConsoleRenderable":
+        """Render log for display.
+
+        Args:
+            record (LogRecord): logging Record.
+            traceback (Optional[Traceback]): Traceback instance or None for no Traceback.
+            message_renderable (ConsoleRenderable): Renderable (typically Text) containing log message contents.
+
+        Returns:
+            ConsoleRenderable: Renderable to display log.
+        """
+        path = Path(record.pathname).name
+        level = self.get_level_text(record)
+        time_format = None if self.formatter is None else self.formatter.datefmt
+        log_time = datetime.fromtimestamp(record.created)
+
+        log_renderable = self._log_render(
+            self.console,
+            [message_renderable] if not traceback else [message_renderable, traceback],
+            log_time=log_time,
+            time_format=time_format,
+            level=level,
+            path=path,
+            line_no=record.lineno,
+            link_path=record.pathname if self.enable_link_path else None,
+        )
+        return log_renderable
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from time import sleep
+
+    FORMAT = "%(message)s"
+    # FORMAT = "%(asctime)-15s - %(levelname)s - %(message)s"
+    logging.basicConfig(
+        level="NOTSET",
+        format=FORMAT,
+        datefmt="[%X]",
+        handlers=[RichHandler(rich_tracebacks=True, tracebacks_show_locals=True)],
+    )
+    log = logging.getLogger("rich")
+
+    log.info("Server starting...")
+    log.info("Listening on http://127.0.0.1:8080")
+    sleep(1)
+
+    log.info("GET /index.html 200 1298")
+    log.info("GET /imgs/backgrounds/back1.jpg 200 54386")
+    log.info("GET /css/styles.css 200 54386")
+    log.warning("GET /favicon.ico 404 242")
+    sleep(1)
+
+    log.debug(
+        "JSONRPC request\n--> %r\n<-- %r",
+        {
+            "version": "1.1",
+            "method": "confirmFruitPurchase",
+            "params": [["apple", "orange", "mangoes", "pomelo"], 1.123],
+            "id": "194521489",
+        },
+        {"version": "1.1", "result": True, "error": None, "id": "194521489"},
+    )
+    log.debug(
+        "Loading configuration file /adasd/asdasd/qeqwe/qwrqwrqwr/sdgsdgsdg/werwerwer/dfgerert/ertertert/ertetert/werwerwer"
+    )
+    log.error("Unable to find 'pomelo' in database!")
+    log.info("POST /jsonrpc/ 200 65532")
+    log.info("POST /admin/ 401 42234")
+    log.warning("password was rejected for admin site.")
+
+    def divide() -> None:
+        number = 1
+        divisor = 0
+        foos = ["foo"] * 100
+        log.debug("in divide")
+        try:
+            number / divisor
+        except:
+            log.exception("An error of some kind occurred!")
+
+    divide()
+    sleep(1)
+    log.critical("Out of memory!")
+    log.info("Server exited with code=-1")
+    log.info("[bold]EXITING...[/bold]", extra=dict(markup=True))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/markup.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/markup.py
new file mode 100644
index 00000000..f6171878
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/markup.py
@@ -0,0 +1,251 @@
+import re
+from ast import literal_eval
+from operator import attrgetter
+from typing import Callable, Iterable, List, Match, NamedTuple, Optional, Tuple, Union
+
+from ._emoji_replace import _emoji_replace
+from .emoji import EmojiVariant
+from .errors import MarkupError
+from .style import Style
+from .text import Span, Text
+
+RE_TAGS = re.compile(
+    r"""((\\*)\[([a-z#/@][^[]*?)])""",
+    re.VERBOSE,
+)
+
+RE_HANDLER = re.compile(r"^([\w.]*?)(\(.*?\))?$")
+
+
+class Tag(NamedTuple):
+    """A tag in console markup."""
+
+    name: str
+    """The tag name. e.g. 'bold'."""
+    parameters: Optional[str]
+    """Any additional parameters after the name."""
+
+    def __str__(self) -> str:
+        return (
+            self.name if self.parameters is None else f"{self.name} {self.parameters}"
+        )
+
+    @property
+    def markup(self) -> str:
+        """Get the string representation of this tag."""
+        return (
+            f"[{self.name}]"
+            if self.parameters is None
+            else f"[{self.name}={self.parameters}]"
+        )
+
+
+_ReStringMatch = Match[str]  # regex match object
+_ReSubCallable = Callable[[_ReStringMatch], str]  # Callable invoked by re.sub
+_EscapeSubMethod = Callable[[_ReSubCallable, str], str]  # Sub method of a compiled re
+
+
+def escape(
+    markup: str,
+    _escape: _EscapeSubMethod = re.compile(r"(\\*)(\[[a-z#/@][^[]*?])").sub,
+) -> str:
+    """Escapes text so that it won't be interpreted as markup.
+
+    Args:
+        markup (str): Content to be inserted in to markup.
+
+    Returns:
+        str: Markup with square brackets escaped.
+    """
+
+    def escape_backslashes(match: Match[str]) -> str:
+        """Called by re.sub replace matches."""
+        backslashes, text = match.groups()
+        return f"{backslashes}{backslashes}\\{text}"
+
+    markup = _escape(escape_backslashes, markup)
+    if markup.endswith("\\") and not markup.endswith("\\\\"):
+        return markup + "\\"
+
+    return markup
+
+
+def _parse(markup: str) -> Iterable[Tuple[int, Optional[str], Optional[Tag]]]:
+    """Parse markup in to an iterable of tuples of (position, text, tag).
+
+    Args:
+        markup (str): A string containing console markup
+
+    """
+    position = 0
+    _divmod = divmod
+    _Tag = Tag
+    for match in RE_TAGS.finditer(markup):
+        full_text, escapes, tag_text = match.groups()
+        start, end = match.span()
+        if start > position:
+            yield start, markup[position:start], None
+        if escapes:
+            backslashes, escaped = _divmod(len(escapes), 2)
+            if backslashes:
+                # Literal backslashes
+                yield start, "\\" * backslashes, None
+                start += backslashes * 2
+            if escaped:
+                # Escape of tag
+                yield start, full_text[len(escapes) :], None
+                position = end
+                continue
+        text, equals, parameters = tag_text.partition("=")
+        yield start, None, _Tag(text, parameters if equals else None)
+        position = end
+    if position < len(markup):
+        yield position, markup[position:], None
+
+
+def render(
+    markup: str,
+    style: Union[str, Style] = "",
+    emoji: bool = True,
+    emoji_variant: Optional[EmojiVariant] = None,
+) -> Text:
+    """Render console markup in to a Text instance.
+
+    Args:
+        markup (str): A string containing console markup.
+        style: (Union[str, Style]): The style to use.
+        emoji (bool, optional): Also render emoji code. Defaults to True.
+        emoji_variant (str, optional): Optional emoji variant, either "text" or "emoji". Defaults to None.
+
+
+    Raises:
+        MarkupError: If there is a syntax error in the markup.
+
+    Returns:
+        Text: A test instance.
+    """
+    emoji_replace = _emoji_replace
+    if "[" not in markup:
+        return Text(
+            emoji_replace(markup, default_variant=emoji_variant) if emoji else markup,
+            style=style,
+        )
+    text = Text(style=style)
+    append = text.append
+    normalize = Style.normalize
+
+    style_stack: List[Tuple[int, Tag]] = []
+    pop = style_stack.pop
+
+    spans: List[Span] = []
+    append_span = spans.append
+
+    _Span = Span
+    _Tag = Tag
+
+    def pop_style(style_name: str) -> Tuple[int, Tag]:
+        """Pop tag matching given style name."""
+        for index, (_, tag) in enumerate(reversed(style_stack), 1):
+            if tag.name == style_name:
+                return pop(-index)
+        raise KeyError(style_name)
+
+    for position, plain_text, tag in _parse(markup):
+        if plain_text is not None:
+            # Handle open brace escapes, where the brace is not part of a tag.
+            plain_text = plain_text.replace("\\[", "[")
+            append(emoji_replace(plain_text) if emoji else plain_text)
+        elif tag is not None:
+            if tag.name.startswith("/"):  # Closing tag
+                style_name = tag.name[1:].strip()
+
+                if style_name:  # explicit close
+                    style_name = normalize(style_name)
+                    try:
+                        start, open_tag = pop_style(style_name)
+                    except KeyError:
+                        raise MarkupError(
+                            f"closing tag '{tag.markup}' at position {position} doesn't match any open tag"
+                        ) from None
+                else:  # implicit close
+                    try:
+                        start, open_tag = pop()
+                    except IndexError:
+                        raise MarkupError(
+                            f"closing tag '[/]' at position {position} has nothing to close"
+                        ) from None
+
+                if open_tag.name.startswith("@"):
+                    if open_tag.parameters:
+                        handler_name = ""
+                        parameters = open_tag.parameters.strip()
+                        handler_match = RE_HANDLER.match(parameters)
+                        if handler_match is not None:
+                            handler_name, match_parameters = handler_match.groups()
+                            parameters = (
+                                "()" if match_parameters is None else match_parameters
+                            )
+
+                        try:
+                            meta_params = literal_eval(parameters)
+                        except SyntaxError as error:
+                            raise MarkupError(
+                                f"error parsing {parameters!r} in {open_tag.parameters!r}; {error.msg}"
+                            )
+                        except Exception as error:
+                            raise MarkupError(
+                                f"error parsing {open_tag.parameters!r}; {error}"
+                            ) from None
+
+                        if handler_name:
+                            meta_params = (
+                                handler_name,
+                                meta_params
+                                if isinstance(meta_params, tuple)
+                                else (meta_params,),
+                            )
+
+                    else:
+                        meta_params = ()
+
+                    append_span(
+                        _Span(
+                            start, len(text), Style(meta={open_tag.name: meta_params})
+                        )
+                    )
+                else:
+                    append_span(_Span(start, len(text), str(open_tag)))
+
+            else:  # Opening tag
+                normalized_tag = _Tag(normalize(tag.name), tag.parameters)
+                style_stack.append((len(text), normalized_tag))
+
+    text_length = len(text)
+    while style_stack:
+        start, tag = style_stack.pop()
+        style = str(tag)
+        if style:
+            append_span(_Span(start, text_length, style))
+
+    text.spans = sorted(spans[::-1], key=attrgetter("start"))
+    return text
+
+
+if __name__ == "__main__":  # pragma: no cover
+    MARKUP = [
+        "[red]Hello World[/red]",
+        "[magenta]Hello [b]World[/b]",
+        "[bold]Bold[italic] bold and italic [/bold]italic[/italic]",
+        "Click [link=https://www.willmcgugan.com]here[/link] to visit my Blog",
+        ":warning-emoji: [bold red blink] DANGER![/]",
+    ]
+
+    from pip._vendor.rich import print
+    from pip._vendor.rich.table import Table
+
+    grid = Table("Markup", "Result", padding=(0, 1))
+
+    for markup in MARKUP:
+        grid.add_row(Text(markup), markup)
+
+    print(grid)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/measure.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/measure.py
new file mode 100644
index 00000000..a508ffa8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/measure.py
@@ -0,0 +1,151 @@
+from operator import itemgetter
+from typing import TYPE_CHECKING, Callable, NamedTuple, Optional, Sequence
+
+from . import errors
+from .protocol import is_renderable, rich_cast
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderableType
+
+
+class Measurement(NamedTuple):
+    """Stores the minimum and maximum widths (in characters) required to render an object."""
+
+    minimum: int
+    """Minimum number of cells required to render."""
+    maximum: int
+    """Maximum number of cells required to render."""
+
+    @property
+    def span(self) -> int:
+        """Get difference between maximum and minimum."""
+        return self.maximum - self.minimum
+
+    def normalize(self) -> "Measurement":
+        """Get measurement that ensures that minimum <= maximum and minimum >= 0
+
+        Returns:
+            Measurement: A normalized measurement.
+        """
+        minimum, maximum = self
+        minimum = min(max(0, minimum), maximum)
+        return Measurement(max(0, minimum), max(0, max(minimum, maximum)))
+
+    def with_maximum(self, width: int) -> "Measurement":
+        """Get a RenderableWith where the widths are <= width.
+
+        Args:
+            width (int): Maximum desired width.
+
+        Returns:
+            Measurement: New Measurement object.
+        """
+        minimum, maximum = self
+        return Measurement(min(minimum, width), min(maximum, width))
+
+    def with_minimum(self, width: int) -> "Measurement":
+        """Get a RenderableWith where the widths are >= width.
+
+        Args:
+            width (int): Minimum desired width.
+
+        Returns:
+            Measurement: New Measurement object.
+        """
+        minimum, maximum = self
+        width = max(0, width)
+        return Measurement(max(minimum, width), max(maximum, width))
+
+    def clamp(
+        self, min_width: Optional[int] = None, max_width: Optional[int] = None
+    ) -> "Measurement":
+        """Clamp a measurement within the specified range.
+
+        Args:
+            min_width (int): Minimum desired width, or ``None`` for no minimum. Defaults to None.
+            max_width (int): Maximum desired width, or ``None`` for no maximum. Defaults to None.
+
+        Returns:
+            Measurement: New Measurement object.
+        """
+        measurement = self
+        if min_width is not None:
+            measurement = measurement.with_minimum(min_width)
+        if max_width is not None:
+            measurement = measurement.with_maximum(max_width)
+        return measurement
+
+    @classmethod
+    def get(
+        cls, console: "Console", options: "ConsoleOptions", renderable: "RenderableType"
+    ) -> "Measurement":
+        """Get a measurement for a renderable.
+
+        Args:
+            console (~rich.console.Console): Console instance.
+            options (~rich.console.ConsoleOptions): Console options.
+            renderable (RenderableType): An object that may be rendered with Rich.
+
+        Raises:
+            errors.NotRenderableError: If the object is not renderable.
+
+        Returns:
+            Measurement: Measurement object containing range of character widths required to render the object.
+        """
+        _max_width = options.max_width
+        if _max_width < 1:
+            return Measurement(0, 0)
+        if isinstance(renderable, str):
+            renderable = console.render_str(
+                renderable, markup=options.markup, highlight=False
+            )
+        renderable = rich_cast(renderable)
+        if is_renderable(renderable):
+            get_console_width: Optional[
+                Callable[["Console", "ConsoleOptions"], "Measurement"]
+            ] = getattr(renderable, "__rich_measure__", None)
+            if get_console_width is not None:
+                render_width = (
+                    get_console_width(console, options)
+                    .normalize()
+                    .with_maximum(_max_width)
+                )
+                if render_width.maximum < 1:
+                    return Measurement(0, 0)
+                return render_width.normalize()
+            else:
+                return Measurement(0, _max_width)
+        else:
+            raise errors.NotRenderableError(
+                f"Unable to get render width for {renderable!r}; "
+                "a str, Segment, or object with __rich_console__ method is required"
+            )
+
+
+def measure_renderables(
+    console: "Console",
+    options: "ConsoleOptions",
+    renderables: Sequence["RenderableType"],
+) -> "Measurement":
+    """Get a measurement that would fit a number of renderables.
+
+    Args:
+        console (~rich.console.Console): Console instance.
+        options (~rich.console.ConsoleOptions): Console options.
+        renderables (Iterable[RenderableType]): One or more renderable objects.
+
+    Returns:
+        Measurement: Measurement object containing range of character widths required to
+            contain all given renderables.
+    """
+    if not renderables:
+        return Measurement(0, 0)
+    get_measurement = Measurement.get
+    measurements = [
+        get_measurement(console, options, renderable) for renderable in renderables
+    ]
+    measured_width = Measurement(
+        max(measurements, key=itemgetter(0)).minimum,
+        max(measurements, key=itemgetter(1)).maximum,
+    )
+    return measured_width
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/padding.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/padding.py
new file mode 100644
index 00000000..0161cd18
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/padding.py
@@ -0,0 +1,141 @@
+from typing import TYPE_CHECKING, List, Optional, Tuple, Union
+
+if TYPE_CHECKING:
+    from .console import (
+        Console,
+        ConsoleOptions,
+        RenderableType,
+        RenderResult,
+    )
+
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import Style
+
+PaddingDimensions = Union[int, Tuple[int], Tuple[int, int], Tuple[int, int, int, int]]
+
+
+class Padding(JupyterMixin):
+    """Draw space around content.
+
+    Example:
+        >>> print(Padding("Hello", (2, 4), style="on blue"))
+
+    Args:
+        renderable (RenderableType): String or other renderable.
+        pad (Union[int, Tuple[int]]): Padding for top, right, bottom, and left borders.
+            May be specified with 1, 2, or 4 integers (CSS style).
+        style (Union[str, Style], optional): Style for padding characters. Defaults to "none".
+        expand (bool, optional): Expand padding to fit available width. Defaults to True.
+    """
+
+    def __init__(
+        self,
+        renderable: "RenderableType",
+        pad: "PaddingDimensions" = (0, 0, 0, 0),
+        *,
+        style: Union[str, Style] = "none",
+        expand: bool = True,
+    ):
+        self.renderable = renderable
+        self.top, self.right, self.bottom, self.left = self.unpack(pad)
+        self.style = style
+        self.expand = expand
+
+    @classmethod
+    def indent(cls, renderable: "RenderableType", level: int) -> "Padding":
+        """Make padding instance to render an indent.
+
+        Args:
+            renderable (RenderableType): String or other renderable.
+            level (int): Number of characters to indent.
+
+        Returns:
+            Padding: A Padding instance.
+        """
+
+        return Padding(renderable, pad=(0, 0, 0, level), expand=False)
+
+    @staticmethod
+    def unpack(pad: "PaddingDimensions") -> Tuple[int, int, int, int]:
+        """Unpack padding specified in CSS style."""
+        if isinstance(pad, int):
+            return (pad, pad, pad, pad)
+        if len(pad) == 1:
+            _pad = pad[0]
+            return (_pad, _pad, _pad, _pad)
+        if len(pad) == 2:
+            pad_top, pad_right = pad
+            return (pad_top, pad_right, pad_top, pad_right)
+        if len(pad) == 4:
+            top, right, bottom, left = pad
+            return (top, right, bottom, left)
+        raise ValueError(f"1, 2 or 4 integers required for padding; {len(pad)} given")
+
+    def __repr__(self) -> str:
+        return f"Padding({self.renderable!r}, ({self.top},{self.right},{self.bottom},{self.left}))"
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        style = console.get_style(self.style)
+        if self.expand:
+            width = options.max_width
+        else:
+            width = min(
+                Measurement.get(console, options, self.renderable).maximum
+                + self.left
+                + self.right,
+                options.max_width,
+            )
+        render_options = options.update_width(width - self.left - self.right)
+        if render_options.height is not None:
+            render_options = render_options.update_height(
+                height=render_options.height - self.top - self.bottom
+            )
+        lines = console.render_lines(
+            self.renderable, render_options, style=style, pad=True
+        )
+        _Segment = Segment
+
+        left = _Segment(" " * self.left, style) if self.left else None
+        right = (
+            [_Segment(f'{" " * self.right}', style), _Segment.line()]
+            if self.right
+            else [_Segment.line()]
+        )
+        blank_line: Optional[List[Segment]] = None
+        if self.top:
+            blank_line = [_Segment(f'{" " * width}\n', style)]
+            yield from blank_line * self.top
+        if left:
+            for line in lines:
+                yield left
+                yield from line
+                yield from right
+        else:
+            for line in lines:
+                yield from line
+                yield from right
+        if self.bottom:
+            blank_line = blank_line or [_Segment(f'{" " * width}\n', style)]
+            yield from blank_line * self.bottom
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        max_width = options.max_width
+        extra_width = self.left + self.right
+        if max_width - extra_width < 1:
+            return Measurement(max_width, max_width)
+        measure_min, measure_max = Measurement.get(console, options, self.renderable)
+        measurement = Measurement(measure_min + extra_width, measure_max + extra_width)
+        measurement = measurement.with_maximum(max_width)
+        return measurement
+
+
+if __name__ == "__main__":  #  pragma: no cover
+    from pip._vendor.rich import print
+
+    print(Padding("Hello, World", (2, 4), style="on blue"))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pager.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pager.py
new file mode 100644
index 00000000..a3f7aa62
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pager.py
@@ -0,0 +1,34 @@
+from abc import ABC, abstractmethod
+from typing import Any
+
+
+class Pager(ABC):
+    """Base class for a pager."""
+
+    @abstractmethod
+    def show(self, content: str) -> None:
+        """Show content in pager.
+
+        Args:
+            content (str): Content to be displayed.
+        """
+
+
+class SystemPager(Pager):
+    """Uses the pager installed on the system."""
+
+    def _pager(self, content: str) -> Any:  #  pragma: no cover
+        return __import__("pydoc").pager(content)
+
+    def show(self, content: str) -> None:
+        """Use the same pager used by pydoc."""
+        self._pager(content)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from .__main__ import make_test_card
+    from .console import Console
+
+    console = Console()
+    with console.pager(styles=True):
+        console.print(make_test_card())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/palette.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/palette.py
new file mode 100644
index 00000000..fa0c4dd4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/palette.py
@@ -0,0 +1,100 @@
+from math import sqrt
+from functools import lru_cache
+from typing import Sequence, Tuple, TYPE_CHECKING
+
+from .color_triplet import ColorTriplet
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.table import Table
+
+
+class Palette:
+    """A palette of available colors."""
+
+    def __init__(self, colors: Sequence[Tuple[int, int, int]]):
+        self._colors = colors
+
+    def __getitem__(self, number: int) -> ColorTriplet:
+        return ColorTriplet(*self._colors[number])
+
+    def __rich__(self) -> "Table":
+        from pip._vendor.rich.color import Color
+        from pip._vendor.rich.style import Style
+        from pip._vendor.rich.text import Text
+        from pip._vendor.rich.table import Table
+
+        table = Table(
+            "index",
+            "RGB",
+            "Color",
+            title="Palette",
+            caption=f"{len(self._colors)} colors",
+            highlight=True,
+            caption_justify="right",
+        )
+        for index, color in enumerate(self._colors):
+            table.add_row(
+                str(index),
+                repr(color),
+                Text(" " * 16, style=Style(bgcolor=Color.from_rgb(*color))),
+            )
+        return table
+
+    # This is somewhat inefficient and needs caching
+    @lru_cache(maxsize=1024)
+    def match(self, color: Tuple[int, int, int]) -> int:
+        """Find a color from a palette that most closely matches a given color.
+
+        Args:
+            color (Tuple[int, int, int]): RGB components in range 0 > 255.
+
+        Returns:
+            int: Index of closes matching color.
+        """
+        red1, green1, blue1 = color
+        _sqrt = sqrt
+        get_color = self._colors.__getitem__
+
+        def get_color_distance(index: int) -> float:
+            """Get the distance to a color."""
+            red2, green2, blue2 = get_color(index)
+            red_mean = (red1 + red2) // 2
+            red = red1 - red2
+            green = green1 - green2
+            blue = blue1 - blue2
+            return _sqrt(
+                (((512 + red_mean) * red * red) >> 8)
+                + 4 * green * green
+                + (((767 - red_mean) * blue * blue) >> 8)
+            )
+
+        min_index = min(range(len(self._colors)), key=get_color_distance)
+        return min_index
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import colorsys
+    from typing import Iterable
+    from pip._vendor.rich.color import Color
+    from pip._vendor.rich.console import Console, ConsoleOptions
+    from pip._vendor.rich.segment import Segment
+    from pip._vendor.rich.style import Style
+
+    class ColorBox:
+        def __rich_console__(
+            self, console: Console, options: ConsoleOptions
+        ) -> Iterable[Segment]:
+            height = console.size.height - 3
+            for y in range(0, height):
+                for x in range(options.max_width):
+                    h = x / options.max_width
+                    l = y / (height + 1)
+                    r1, g1, b1 = colorsys.hls_to_rgb(h, l, 1.0)
+                    r2, g2, b2 = colorsys.hls_to_rgb(h, l + (1 / height / 2), 1.0)
+                    bgcolor = Color.from_rgb(r1 * 255, g1 * 255, b1 * 255)
+                    color = Color.from_rgb(r2 * 255, g2 * 255, b2 * 255)
+                    yield Segment("▄", Style(color=color, bgcolor=bgcolor))
+                yield Segment.line()
+
+    console = Console()
+    console.print(ColorBox())
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/panel.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/panel.py
new file mode 100644
index 00000000..8cfa6f4a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/panel.py
@@ -0,0 +1,318 @@
+from typing import TYPE_CHECKING, Optional
+
+from .align import AlignMethod
+from .box import ROUNDED, Box
+from .cells import cell_len
+from .jupyter import JupyterMixin
+from .measure import Measurement, measure_renderables
+from .padding import Padding, PaddingDimensions
+from .segment import Segment
+from .style import Style, StyleType
+from .text import Text, TextType
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderableType, RenderResult
+
+
+class Panel(JupyterMixin):
+    """A console renderable that draws a border around its contents.
+
+    Example:
+        >>> console.print(Panel("Hello, World!"))
+
+    Args:
+        renderable (RenderableType): A console renderable object.
+        box (Box, optional): A Box instance that defines the look of the border (see :ref:`appendix_box`. Defaults to box.ROUNDED.
+        title (Optional[TextType], optional): Optional title displayed in panel header. Defaults to None.
+        title_align (AlignMethod, optional): Alignment of title. Defaults to "center".
+        subtitle (Optional[TextType], optional): Optional subtitle displayed in panel footer. Defaults to None.
+        subtitle_align (AlignMethod, optional): Alignment of subtitle. Defaults to "center".
+        safe_box (bool, optional): Disable box characters that don't display on windows legacy terminal with *raster* fonts. Defaults to True.
+        expand (bool, optional): If True the panel will stretch to fill the console width, otherwise it will be sized to fit the contents. Defaults to True.
+        style (str, optional): The style of the panel (border and contents). Defaults to "none".
+        border_style (str, optional): The style of the border. Defaults to "none".
+        width (Optional[int], optional): Optional width of panel. Defaults to None to auto-detect.
+        height (Optional[int], optional): Optional height of panel. Defaults to None to auto-detect.
+        padding (Optional[PaddingDimensions]): Optional padding around renderable. Defaults to 0.
+        highlight (bool, optional): Enable automatic highlighting of panel title (if str). Defaults to False.
+    """
+
+    def __init__(
+        self,
+        renderable: "RenderableType",
+        box: Box = ROUNDED,
+        *,
+        title: Optional[TextType] = None,
+        title_align: AlignMethod = "center",
+        subtitle: Optional[TextType] = None,
+        subtitle_align: AlignMethod = "center",
+        safe_box: Optional[bool] = None,
+        expand: bool = True,
+        style: StyleType = "none",
+        border_style: StyleType = "none",
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+        padding: PaddingDimensions = (0, 1),
+        highlight: bool = False,
+    ) -> None:
+        self.renderable = renderable
+        self.box = box
+        self.title = title
+        self.title_align: AlignMethod = title_align
+        self.subtitle = subtitle
+        self.subtitle_align = subtitle_align
+        self.safe_box = safe_box
+        self.expand = expand
+        self.style = style
+        self.border_style = border_style
+        self.width = width
+        self.height = height
+        self.padding = padding
+        self.highlight = highlight
+
+    @classmethod
+    def fit(
+        cls,
+        renderable: "RenderableType",
+        box: Box = ROUNDED,
+        *,
+        title: Optional[TextType] = None,
+        title_align: AlignMethod = "center",
+        subtitle: Optional[TextType] = None,
+        subtitle_align: AlignMethod = "center",
+        safe_box: Optional[bool] = None,
+        style: StyleType = "none",
+        border_style: StyleType = "none",
+        width: Optional[int] = None,
+        height: Optional[int] = None,
+        padding: PaddingDimensions = (0, 1),
+        highlight: bool = False,
+    ) -> "Panel":
+        """An alternative constructor that sets expand=False."""
+        return cls(
+            renderable,
+            box,
+            title=title,
+            title_align=title_align,
+            subtitle=subtitle,
+            subtitle_align=subtitle_align,
+            safe_box=safe_box,
+            style=style,
+            border_style=border_style,
+            width=width,
+            height=height,
+            padding=padding,
+            highlight=highlight,
+            expand=False,
+        )
+
+    @property
+    def _title(self) -> Optional[Text]:
+        if self.title:
+            title_text = (
+                Text.from_markup(self.title)
+                if isinstance(self.title, str)
+                else self.title.copy()
+            )
+            title_text.end = ""
+            title_text.plain = title_text.plain.replace("\n", " ")
+            title_text.no_wrap = True
+            title_text.expand_tabs()
+            title_text.pad(1)
+            return title_text
+        return None
+
+    @property
+    def _subtitle(self) -> Optional[Text]:
+        if self.subtitle:
+            subtitle_text = (
+                Text.from_markup(self.subtitle)
+                if isinstance(self.subtitle, str)
+                else self.subtitle.copy()
+            )
+            subtitle_text.end = ""
+            subtitle_text.plain = subtitle_text.plain.replace("\n", " ")
+            subtitle_text.no_wrap = True
+            subtitle_text.expand_tabs()
+            subtitle_text.pad(1)
+            return subtitle_text
+        return None
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        _padding = Padding.unpack(self.padding)
+        renderable = (
+            Padding(self.renderable, _padding) if any(_padding) else self.renderable
+        )
+        style = console.get_style(self.style)
+        partial_border_style = console.get_style(self.border_style)
+        border_style = style + partial_border_style
+        width = (
+            options.max_width
+            if self.width is None
+            else min(options.max_width, self.width)
+        )
+
+        safe_box: bool = console.safe_box if self.safe_box is None else self.safe_box
+        box = self.box.substitute(options, safe=safe_box)
+
+        def align_text(
+            text: Text, width: int, align: str, character: str, style: Style
+        ) -> Text:
+            """Gets new aligned text.
+
+            Args:
+                text (Text): Title or subtitle text.
+                width (int): Desired width.
+                align (str): Alignment.
+                character (str): Character for alignment.
+                style (Style): Border style
+
+            Returns:
+                Text: New text instance
+            """
+            text = text.copy()
+            text.truncate(width)
+            excess_space = width - cell_len(text.plain)
+            if text.style:
+                text.stylize(console.get_style(text.style))
+
+            if excess_space:
+                if align == "left":
+                    return Text.assemble(
+                        text,
+                        (character * excess_space, style),
+                        no_wrap=True,
+                        end="",
+                    )
+                elif align == "center":
+                    left = excess_space // 2
+                    return Text.assemble(
+                        (character * left, style),
+                        text,
+                        (character * (excess_space - left), style),
+                        no_wrap=True,
+                        end="",
+                    )
+                else:
+                    return Text.assemble(
+                        (character * excess_space, style),
+                        text,
+                        no_wrap=True,
+                        end="",
+                    )
+            return text
+
+        title_text = self._title
+        if title_text is not None:
+            title_text.stylize_before(partial_border_style)
+
+        child_width = (
+            width - 2
+            if self.expand
+            else console.measure(
+                renderable, options=options.update_width(width - 2)
+            ).maximum
+        )
+        child_height = self.height or options.height or None
+        if child_height:
+            child_height -= 2
+        if title_text is not None:
+            child_width = min(
+                options.max_width - 2, max(child_width, title_text.cell_len + 2)
+            )
+
+        width = child_width + 2
+        child_options = options.update(
+            width=child_width, height=child_height, highlight=self.highlight
+        )
+        lines = console.render_lines(renderable, child_options, style=style)
+
+        line_start = Segment(box.mid_left, border_style)
+        line_end = Segment(f"{box.mid_right}", border_style)
+        new_line = Segment.line()
+        if title_text is None or width <= 4:
+            yield Segment(box.get_top([width - 2]), border_style)
+        else:
+            title_text = align_text(
+                title_text,
+                width - 4,
+                self.title_align,
+                box.top,
+                border_style,
+            )
+            yield Segment(box.top_left + box.top, border_style)
+            yield from console.render(title_text, child_options.update_width(width - 4))
+            yield Segment(box.top + box.top_right, border_style)
+
+        yield new_line
+        for line in lines:
+            yield line_start
+            yield from line
+            yield line_end
+            yield new_line
+
+        subtitle_text = self._subtitle
+        if subtitle_text is not None:
+            subtitle_text.stylize_before(partial_border_style)
+
+        if subtitle_text is None or width <= 4:
+            yield Segment(box.get_bottom([width - 2]), border_style)
+        else:
+            subtitle_text = align_text(
+                subtitle_text,
+                width - 4,
+                self.subtitle_align,
+                box.bottom,
+                border_style,
+            )
+            yield Segment(box.bottom_left + box.bottom, border_style)
+            yield from console.render(
+                subtitle_text, child_options.update_width(width - 4)
+            )
+            yield Segment(box.bottom + box.bottom_right, border_style)
+
+        yield new_line
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        _title = self._title
+        _, right, _, left = Padding.unpack(self.padding)
+        padding = left + right
+        renderables = [self.renderable, _title] if _title else [self.renderable]
+
+        if self.width is None:
+            width = (
+                measure_renderables(
+                    console,
+                    options.update_width(options.max_width - padding - 2),
+                    renderables,
+                ).maximum
+                + padding
+                + 2
+            )
+        else:
+            width = self.width
+        return Measurement(width, width)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from .console import Console
+
+    c = Console()
+
+    from .box import DOUBLE, ROUNDED
+    from .padding import Padding
+
+    p = Panel(
+        "Hello, World!",
+        title="rich.Panel",
+        style="white on blue",
+        box=DOUBLE,
+        padding=1,
+    )
+
+    c.print()
+    c.print(p)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pretty.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pretty.py
new file mode 100644
index 00000000..c4a274f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/pretty.py
@@ -0,0 +1,1016 @@
+import builtins
+import collections
+import dataclasses
+import inspect
+import os
+import reprlib
+import sys
+from array import array
+from collections import Counter, UserDict, UserList, defaultdict, deque
+from dataclasses import dataclass, fields, is_dataclass
+from inspect import isclass
+from itertools import islice
+from types import MappingProxyType
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    DefaultDict,
+    Deque,
+    Dict,
+    Iterable,
+    List,
+    Optional,
+    Sequence,
+    Set,
+    Tuple,
+    Union,
+)
+
+from pip._vendor.rich.repr import RichReprResult
+
+try:
+    import attr as _attr_module
+
+    _has_attrs = hasattr(_attr_module, "ib")
+except ImportError:  # pragma: no cover
+    _has_attrs = False
+
+from . import get_console
+from ._loop import loop_last
+from ._pick import pick_bool
+from .abc import RichRenderable
+from .cells import cell_len
+from .highlighter import ReprHighlighter
+from .jupyter import JupyterMixin, JupyterRenderable
+from .measure import Measurement
+from .text import Text
+
+if TYPE_CHECKING:
+    from .console import (
+        Console,
+        ConsoleOptions,
+        HighlighterType,
+        JustifyMethod,
+        OverflowMethod,
+        RenderResult,
+    )
+
+
+def _is_attr_object(obj: Any) -> bool:
+    """Check if an object was created with attrs module."""
+    return _has_attrs and _attr_module.has(type(obj))
+
+
+def _get_attr_fields(obj: Any) -> Sequence["_attr_module.Attribute[Any]"]:
+    """Get fields for an attrs object."""
+    return _attr_module.fields(type(obj)) if _has_attrs else []
+
+
+def _is_dataclass_repr(obj: object) -> bool:
+    """Check if an instance of a dataclass contains the default repr.
+
+    Args:
+        obj (object): A dataclass instance.
+
+    Returns:
+        bool: True if the default repr is used, False if there is a custom repr.
+    """
+    # Digging in to a lot of internals here
+    # Catching all exceptions in case something is missing on a non CPython implementation
+    try:
+        return obj.__repr__.__code__.co_filename in (
+            dataclasses.__file__,
+            reprlib.__file__,
+        )
+    except Exception:  # pragma: no coverage
+        return False
+
+
+_dummy_namedtuple = collections.namedtuple("_dummy_namedtuple", [])
+
+
+def _has_default_namedtuple_repr(obj: object) -> bool:
+    """Check if an instance of namedtuple contains the default repr
+
+    Args:
+        obj (object): A namedtuple
+
+    Returns:
+        bool: True if the default repr is used, False if there's a custom repr.
+    """
+    obj_file = None
+    try:
+        obj_file = inspect.getfile(obj.__repr__)
+    except (OSError, TypeError):
+        # OSError handles case where object is defined in __main__ scope, e.g. REPL - no filename available.
+        # TypeError trapped defensively, in case of object without filename slips through.
+        pass
+    default_repr_file = inspect.getfile(_dummy_namedtuple.__repr__)
+    return obj_file == default_repr_file
+
+
+def _ipy_display_hook(
+    value: Any,
+    console: Optional["Console"] = None,
+    overflow: "OverflowMethod" = "ignore",
+    crop: bool = False,
+    indent_guides: bool = False,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+    max_depth: Optional[int] = None,
+    expand_all: bool = False,
+) -> Union[str, None]:
+    # needed here to prevent circular import:
+    from .console import ConsoleRenderable
+
+    # always skip rich generated jupyter renderables or None values
+    if _safe_isinstance(value, JupyterRenderable) or value is None:
+        return None
+
+    console = console or get_console()
+
+    with console.capture() as capture:
+        # certain renderables should start on a new line
+        if _safe_isinstance(value, ConsoleRenderable):
+            console.line()
+        console.print(
+            (
+                value
+                if _safe_isinstance(value, RichRenderable)
+                else Pretty(
+                    value,
+                    overflow=overflow,
+                    indent_guides=indent_guides,
+                    max_length=max_length,
+                    max_string=max_string,
+                    max_depth=max_depth,
+                    expand_all=expand_all,
+                    margin=12,
+                )
+            ),
+            crop=crop,
+            new_line_start=True,
+            end="",
+        )
+    # strip trailing newline, not usually part of a text repr
+    # I'm not sure if this should be prevented at a lower level
+    return capture.get().rstrip("\n")
+
+
+def _safe_isinstance(
+    obj: object, class_or_tuple: Union[type, Tuple[type, ...]]
+) -> bool:
+    """isinstance can fail in rare cases, for example types with no __class__"""
+    try:
+        return isinstance(obj, class_or_tuple)
+    except Exception:
+        return False
+
+
+def install(
+    console: Optional["Console"] = None,
+    overflow: "OverflowMethod" = "ignore",
+    crop: bool = False,
+    indent_guides: bool = False,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+    max_depth: Optional[int] = None,
+    expand_all: bool = False,
+) -> None:
+    """Install automatic pretty printing in the Python REPL.
+
+    Args:
+        console (Console, optional): Console instance or ``None`` to use global console. Defaults to None.
+        overflow (Optional[OverflowMethod], optional): Overflow method. Defaults to "ignore".
+        crop (Optional[bool], optional): Enable cropping of long lines. Defaults to False.
+        indent_guides (bool, optional): Enable indentation guides. Defaults to False.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to None.
+        max_depth (int, optional): Maximum depth of nested data structures, or None for no maximum. Defaults to None.
+        expand_all (bool, optional): Expand all containers. Defaults to False.
+        max_frames (int): Maximum number of frames to show in a traceback, 0 for no maximum. Defaults to 100.
+    """
+    from pip._vendor.rich import get_console
+
+    console = console or get_console()
+    assert console is not None
+
+    def display_hook(value: Any) -> None:
+        """Replacement sys.displayhook which prettifies objects with Rich."""
+        if value is not None:
+            assert console is not None
+            builtins._ = None  # type: ignore[attr-defined]
+            console.print(
+                (
+                    value
+                    if _safe_isinstance(value, RichRenderable)
+                    else Pretty(
+                        value,
+                        overflow=overflow,
+                        indent_guides=indent_guides,
+                        max_length=max_length,
+                        max_string=max_string,
+                        max_depth=max_depth,
+                        expand_all=expand_all,
+                    )
+                ),
+                crop=crop,
+            )
+            builtins._ = value  # type: ignore[attr-defined]
+
+    try:
+        ip = get_ipython()  # type: ignore[name-defined]
+    except NameError:
+        sys.displayhook = display_hook
+    else:
+        from IPython.core.formatters import BaseFormatter
+
+        class RichFormatter(BaseFormatter):  # type: ignore[misc]
+            pprint: bool = True
+
+            def __call__(self, value: Any) -> Any:
+                if self.pprint:
+                    return _ipy_display_hook(
+                        value,
+                        console=get_console(),
+                        overflow=overflow,
+                        indent_guides=indent_guides,
+                        max_length=max_length,
+                        max_string=max_string,
+                        max_depth=max_depth,
+                        expand_all=expand_all,
+                    )
+                else:
+                    return repr(value)
+
+        # replace plain text formatter with rich formatter
+        rich_formatter = RichFormatter()
+        ip.display_formatter.formatters["text/plain"] = rich_formatter
+
+
+class Pretty(JupyterMixin):
+    """A rich renderable that pretty prints an object.
+
+    Args:
+        _object (Any): An object to pretty print.
+        highlighter (HighlighterType, optional): Highlighter object to apply to result, or None for ReprHighlighter. Defaults to None.
+        indent_size (int, optional): Number of spaces in indent. Defaults to 4.
+        justify (JustifyMethod, optional): Justify method, or None for default. Defaults to None.
+        overflow (OverflowMethod, optional): Overflow method, or None for default. Defaults to None.
+        no_wrap (Optional[bool], optional): Disable word wrapping. Defaults to False.
+        indent_guides (bool, optional): Enable indentation guides. Defaults to False.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to None.
+        max_depth (int, optional): Maximum depth of nested data structures, or None for no maximum. Defaults to None.
+        expand_all (bool, optional): Expand all containers. Defaults to False.
+        margin (int, optional): Subtrace a margin from width to force containers to expand earlier. Defaults to 0.
+        insert_line (bool, optional): Insert a new line if the output has multiple new lines. Defaults to False.
+    """
+
+    def __init__(
+        self,
+        _object: Any,
+        highlighter: Optional["HighlighterType"] = None,
+        *,
+        indent_size: int = 4,
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        no_wrap: Optional[bool] = False,
+        indent_guides: bool = False,
+        max_length: Optional[int] = None,
+        max_string: Optional[int] = None,
+        max_depth: Optional[int] = None,
+        expand_all: bool = False,
+        margin: int = 0,
+        insert_line: bool = False,
+    ) -> None:
+        self._object = _object
+        self.highlighter = highlighter or ReprHighlighter()
+        self.indent_size = indent_size
+        self.justify: Optional["JustifyMethod"] = justify
+        self.overflow: Optional["OverflowMethod"] = overflow
+        self.no_wrap = no_wrap
+        self.indent_guides = indent_guides
+        self.max_length = max_length
+        self.max_string = max_string
+        self.max_depth = max_depth
+        self.expand_all = expand_all
+        self.margin = margin
+        self.insert_line = insert_line
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        pretty_str = pretty_repr(
+            self._object,
+            max_width=options.max_width - self.margin,
+            indent_size=self.indent_size,
+            max_length=self.max_length,
+            max_string=self.max_string,
+            max_depth=self.max_depth,
+            expand_all=self.expand_all,
+        )
+        pretty_text = Text.from_ansi(
+            pretty_str,
+            justify=self.justify or options.justify,
+            overflow=self.overflow or options.overflow,
+            no_wrap=pick_bool(self.no_wrap, options.no_wrap),
+            style="pretty",
+        )
+        pretty_text = (
+            self.highlighter(pretty_text)
+            if pretty_text
+            else Text(
+                f"{type(self._object)}.__repr__ returned empty string",
+                style="dim italic",
+            )
+        )
+        if self.indent_guides and not options.ascii_only:
+            pretty_text = pretty_text.with_indent_guides(
+                self.indent_size, style="repr.indent"
+            )
+        if self.insert_line and "\n" in pretty_text:
+            yield ""
+        yield pretty_text
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        pretty_str = pretty_repr(
+            self._object,
+            max_width=options.max_width,
+            indent_size=self.indent_size,
+            max_length=self.max_length,
+            max_string=self.max_string,
+            max_depth=self.max_depth,
+            expand_all=self.expand_all,
+        )
+        text_width = (
+            max(cell_len(line) for line in pretty_str.splitlines()) if pretty_str else 0
+        )
+        return Measurement(text_width, text_width)
+
+
+def _get_braces_for_defaultdict(_object: DefaultDict[Any, Any]) -> Tuple[str, str, str]:
+    return (
+        f"defaultdict({_object.default_factory!r}, {{",
+        "})",
+        f"defaultdict({_object.default_factory!r}, {{}})",
+    )
+
+
+def _get_braces_for_deque(_object: Deque[Any]) -> Tuple[str, str, str]:
+    if _object.maxlen is None:
+        return ("deque([", "])", "deque()")
+    return (
+        "deque([",
+        f"], maxlen={_object.maxlen})",
+        f"deque(maxlen={_object.maxlen})",
+    )
+
+
+def _get_braces_for_array(_object: "array[Any]") -> Tuple[str, str, str]:
+    return (f"array({_object.typecode!r}, [", "])", f"array({_object.typecode!r})")
+
+
+_BRACES: Dict[type, Callable[[Any], Tuple[str, str, str]]] = {
+    os._Environ: lambda _object: ("environ({", "})", "environ({})"),
+    array: _get_braces_for_array,
+    defaultdict: _get_braces_for_defaultdict,
+    Counter: lambda _object: ("Counter({", "})", "Counter()"),
+    deque: _get_braces_for_deque,
+    dict: lambda _object: ("{", "}", "{}"),
+    UserDict: lambda _object: ("{", "}", "{}"),
+    frozenset: lambda _object: ("frozenset({", "})", "frozenset()"),
+    list: lambda _object: ("[", "]", "[]"),
+    UserList: lambda _object: ("[", "]", "[]"),
+    set: lambda _object: ("{", "}", "set()"),
+    tuple: lambda _object: ("(", ")", "()"),
+    MappingProxyType: lambda _object: ("mappingproxy({", "})", "mappingproxy({})"),
+}
+_CONTAINERS = tuple(_BRACES.keys())
+_MAPPING_CONTAINERS = (dict, os._Environ, MappingProxyType, UserDict)
+
+
+def is_expandable(obj: Any) -> bool:
+    """Check if an object may be expanded by pretty print."""
+    return (
+        _safe_isinstance(obj, _CONTAINERS)
+        or (is_dataclass(obj))
+        or (hasattr(obj, "__rich_repr__"))
+        or _is_attr_object(obj)
+    ) and not isclass(obj)
+
+
+@dataclass
+class Node:
+    """A node in a repr tree. May be atomic or a container."""
+
+    key_repr: str = ""
+    value_repr: str = ""
+    open_brace: str = ""
+    close_brace: str = ""
+    empty: str = ""
+    last: bool = False
+    is_tuple: bool = False
+    is_namedtuple: bool = False
+    children: Optional[List["Node"]] = None
+    key_separator: str = ": "
+    separator: str = ", "
+
+    def iter_tokens(self) -> Iterable[str]:
+        """Generate tokens for this node."""
+        if self.key_repr:
+            yield self.key_repr
+            yield self.key_separator
+        if self.value_repr:
+            yield self.value_repr
+        elif self.children is not None:
+            if self.children:
+                yield self.open_brace
+                if self.is_tuple and not self.is_namedtuple and len(self.children) == 1:
+                    yield from self.children[0].iter_tokens()
+                    yield ","
+                else:
+                    for child in self.children:
+                        yield from child.iter_tokens()
+                        if not child.last:
+                            yield self.separator
+                yield self.close_brace
+            else:
+                yield self.empty
+
+    def check_length(self, start_length: int, max_length: int) -> bool:
+        """Check the length fits within a limit.
+
+        Args:
+            start_length (int): Starting length of the line (indent, prefix, suffix).
+            max_length (int): Maximum length.
+
+        Returns:
+            bool: True if the node can be rendered within max length, otherwise False.
+        """
+        total_length = start_length
+        for token in self.iter_tokens():
+            total_length += cell_len(token)
+            if total_length > max_length:
+                return False
+        return True
+
+    def __str__(self) -> str:
+        repr_text = "".join(self.iter_tokens())
+        return repr_text
+
+    def render(
+        self, max_width: int = 80, indent_size: int = 4, expand_all: bool = False
+    ) -> str:
+        """Render the node to a pretty repr.
+
+        Args:
+            max_width (int, optional): Maximum width of the repr. Defaults to 80.
+            indent_size (int, optional): Size of indents. Defaults to 4.
+            expand_all (bool, optional): Expand all levels. Defaults to False.
+
+        Returns:
+            str: A repr string of the original object.
+        """
+        lines = [_Line(node=self, is_root=True)]
+        line_no = 0
+        while line_no < len(lines):
+            line = lines[line_no]
+            if line.expandable and not line.expanded:
+                if expand_all or not line.check_length(max_width):
+                    lines[line_no : line_no + 1] = line.expand(indent_size)
+            line_no += 1
+
+        repr_str = "\n".join(str(line) for line in lines)
+        return repr_str
+
+
+@dataclass
+class _Line:
+    """A line in repr output."""
+
+    parent: Optional["_Line"] = None
+    is_root: bool = False
+    node: Optional[Node] = None
+    text: str = ""
+    suffix: str = ""
+    whitespace: str = ""
+    expanded: bool = False
+    last: bool = False
+
+    @property
+    def expandable(self) -> bool:
+        """Check if the line may be expanded."""
+        return bool(self.node is not None and self.node.children)
+
+    def check_length(self, max_length: int) -> bool:
+        """Check this line fits within a given number of cells."""
+        start_length = (
+            len(self.whitespace) + cell_len(self.text) + cell_len(self.suffix)
+        )
+        assert self.node is not None
+        return self.node.check_length(start_length, max_length)
+
+    def expand(self, indent_size: int) -> Iterable["_Line"]:
+        """Expand this line by adding children on their own line."""
+        node = self.node
+        assert node is not None
+        whitespace = self.whitespace
+        assert node.children
+        if node.key_repr:
+            new_line = yield _Line(
+                text=f"{node.key_repr}{node.key_separator}{node.open_brace}",
+                whitespace=whitespace,
+            )
+        else:
+            new_line = yield _Line(text=node.open_brace, whitespace=whitespace)
+        child_whitespace = self.whitespace + " " * indent_size
+        tuple_of_one = node.is_tuple and len(node.children) == 1
+        for last, child in loop_last(node.children):
+            separator = "," if tuple_of_one else node.separator
+            line = _Line(
+                parent=new_line,
+                node=child,
+                whitespace=child_whitespace,
+                suffix=separator,
+                last=last and not tuple_of_one,
+            )
+            yield line
+
+        yield _Line(
+            text=node.close_brace,
+            whitespace=whitespace,
+            suffix=self.suffix,
+            last=self.last,
+        )
+
+    def __str__(self) -> str:
+        if self.last:
+            return f"{self.whitespace}{self.text}{self.node or ''}"
+        else:
+            return (
+                f"{self.whitespace}{self.text}{self.node or ''}{self.suffix.rstrip()}"
+            )
+
+
+def _is_namedtuple(obj: Any) -> bool:
+    """Checks if an object is most likely a namedtuple. It is possible
+    to craft an object that passes this check and isn't a namedtuple, but
+    there is only a minuscule chance of this happening unintentionally.
+
+    Args:
+        obj (Any): The object to test
+
+    Returns:
+        bool: True if the object is a namedtuple. False otherwise.
+    """
+    try:
+        fields = getattr(obj, "_fields", None)
+    except Exception:
+        # Being very defensive - if we cannot get the attr then its not a namedtuple
+        return False
+    return isinstance(obj, tuple) and isinstance(fields, tuple)
+
+
+def traverse(
+    _object: Any,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+    max_depth: Optional[int] = None,
+) -> Node:
+    """Traverse object and generate a tree.
+
+    Args:
+        _object (Any): Object to be traversed.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of string before truncating, or None to disable truncating.
+            Defaults to None.
+        max_depth (int, optional): Maximum depth of data structures, or None for no maximum.
+            Defaults to None.
+
+    Returns:
+        Node: The root of a tree structure which can be used to render a pretty repr.
+    """
+
+    def to_repr(obj: Any) -> str:
+        """Get repr string for an object, but catch errors."""
+        if (
+            max_string is not None
+            and _safe_isinstance(obj, (bytes, str))
+            and len(obj) > max_string
+        ):
+            truncated = len(obj) - max_string
+            obj_repr = f"{obj[:max_string]!r}+{truncated}"
+        else:
+            try:
+                obj_repr = repr(obj)
+            except Exception as error:
+                obj_repr = f"<repr-error {str(error)!r}>"
+        return obj_repr
+
+    visited_ids: Set[int] = set()
+    push_visited = visited_ids.add
+    pop_visited = visited_ids.remove
+
+    def _traverse(obj: Any, root: bool = False, depth: int = 0) -> Node:
+        """Walk the object depth first."""
+
+        obj_id = id(obj)
+        if obj_id in visited_ids:
+            # Recursion detected
+            return Node(value_repr="...")
+
+        obj_type = type(obj)
+        children: List[Node]
+        reached_max_depth = max_depth is not None and depth >= max_depth
+
+        def iter_rich_args(rich_args: Any) -> Iterable[Union[Any, Tuple[str, Any]]]:
+            for arg in rich_args:
+                if _safe_isinstance(arg, tuple):
+                    if len(arg) == 3:
+                        key, child, default = arg
+                        if default == child:
+                            continue
+                        yield key, child
+                    elif len(arg) == 2:
+                        key, child = arg
+                        yield key, child
+                    elif len(arg) == 1:
+                        yield arg[0]
+                else:
+                    yield arg
+
+        try:
+            fake_attributes = hasattr(
+                obj, "awehoi234_wdfjwljet234_234wdfoijsdfmmnxpi492"
+            )
+        except Exception:
+            fake_attributes = False
+
+        rich_repr_result: Optional[RichReprResult] = None
+        if not fake_attributes:
+            try:
+                if hasattr(obj, "__rich_repr__") and not isclass(obj):
+                    rich_repr_result = obj.__rich_repr__()
+            except Exception:
+                pass
+
+        if rich_repr_result is not None:
+            push_visited(obj_id)
+            angular = getattr(obj.__rich_repr__, "angular", False)
+            args = list(iter_rich_args(rich_repr_result))
+            class_name = obj.__class__.__name__
+
+            if args:
+                children = []
+                append = children.append
+
+                if reached_max_depth:
+                    if angular:
+                        node = Node(value_repr=f"<{class_name}...>")
+                    else:
+                        node = Node(value_repr=f"{class_name}(...)")
+                else:
+                    if angular:
+                        node = Node(
+                            open_brace=f"<{class_name} ",
+                            close_brace=">",
+                            children=children,
+                            last=root,
+                            separator=" ",
+                        )
+                    else:
+                        node = Node(
+                            open_brace=f"{class_name}(",
+                            close_brace=")",
+                            children=children,
+                            last=root,
+                        )
+                    for last, arg in loop_last(args):
+                        if _safe_isinstance(arg, tuple):
+                            key, child = arg
+                            child_node = _traverse(child, depth=depth + 1)
+                            child_node.last = last
+                            child_node.key_repr = key
+                            child_node.key_separator = "="
+                            append(child_node)
+                        else:
+                            child_node = _traverse(arg, depth=depth + 1)
+                            child_node.last = last
+                            append(child_node)
+            else:
+                node = Node(
+                    value_repr=f"<{class_name}>" if angular else f"{class_name}()",
+                    children=[],
+                    last=root,
+                )
+            pop_visited(obj_id)
+        elif _is_attr_object(obj) and not fake_attributes:
+            push_visited(obj_id)
+            children = []
+            append = children.append
+
+            attr_fields = _get_attr_fields(obj)
+            if attr_fields:
+                if reached_max_depth:
+                    node = Node(value_repr=f"{obj.__class__.__name__}(...)")
+                else:
+                    node = Node(
+                        open_brace=f"{obj.__class__.__name__}(",
+                        close_brace=")",
+                        children=children,
+                        last=root,
+                    )
+
+                    def iter_attrs() -> (
+                        Iterable[Tuple[str, Any, Optional[Callable[[Any], str]]]]
+                    ):
+                        """Iterate over attr fields and values."""
+                        for attr in attr_fields:
+                            if attr.repr:
+                                try:
+                                    value = getattr(obj, attr.name)
+                                except Exception as error:
+                                    # Can happen, albeit rarely
+                                    yield (attr.name, error, None)
+                                else:
+                                    yield (
+                                        attr.name,
+                                        value,
+                                        attr.repr if callable(attr.repr) else None,
+                                    )
+
+                    for last, (name, value, repr_callable) in loop_last(iter_attrs()):
+                        if repr_callable:
+                            child_node = Node(value_repr=str(repr_callable(value)))
+                        else:
+                            child_node = _traverse(value, depth=depth + 1)
+                        child_node.last = last
+                        child_node.key_repr = name
+                        child_node.key_separator = "="
+                        append(child_node)
+            else:
+                node = Node(
+                    value_repr=f"{obj.__class__.__name__}()", children=[], last=root
+                )
+            pop_visited(obj_id)
+        elif (
+            is_dataclass(obj)
+            and not _safe_isinstance(obj, type)
+            and not fake_attributes
+            and _is_dataclass_repr(obj)
+        ):
+            push_visited(obj_id)
+            children = []
+            append = children.append
+            if reached_max_depth:
+                node = Node(value_repr=f"{obj.__class__.__name__}(...)")
+            else:
+                node = Node(
+                    open_brace=f"{obj.__class__.__name__}(",
+                    close_brace=")",
+                    children=children,
+                    last=root,
+                    empty=f"{obj.__class__.__name__}()",
+                )
+
+                for last, field in loop_last(
+                    field
+                    for field in fields(obj)
+                    if field.repr and hasattr(obj, field.name)
+                ):
+                    child_node = _traverse(getattr(obj, field.name), depth=depth + 1)
+                    child_node.key_repr = field.name
+                    child_node.last = last
+                    child_node.key_separator = "="
+                    append(child_node)
+
+            pop_visited(obj_id)
+        elif _is_namedtuple(obj) and _has_default_namedtuple_repr(obj):
+            push_visited(obj_id)
+            class_name = obj.__class__.__name__
+            if reached_max_depth:
+                # If we've reached the max depth, we still show the class name, but not its contents
+                node = Node(
+                    value_repr=f"{class_name}(...)",
+                )
+            else:
+                children = []
+                append = children.append
+                node = Node(
+                    open_brace=f"{class_name}(",
+                    close_brace=")",
+                    children=children,
+                    empty=f"{class_name}()",
+                )
+                for last, (key, value) in loop_last(obj._asdict().items()):
+                    child_node = _traverse(value, depth=depth + 1)
+                    child_node.key_repr = key
+                    child_node.last = last
+                    child_node.key_separator = "="
+                    append(child_node)
+            pop_visited(obj_id)
+        elif _safe_isinstance(obj, _CONTAINERS):
+            for container_type in _CONTAINERS:
+                if _safe_isinstance(obj, container_type):
+                    obj_type = container_type
+                    break
+
+            push_visited(obj_id)
+
+            open_brace, close_brace, empty = _BRACES[obj_type](obj)
+
+            if reached_max_depth:
+                node = Node(value_repr=f"{open_brace}...{close_brace}")
+            elif obj_type.__repr__ != type(obj).__repr__:
+                node = Node(value_repr=to_repr(obj), last=root)
+            elif obj:
+                children = []
+                node = Node(
+                    open_brace=open_brace,
+                    close_brace=close_brace,
+                    children=children,
+                    last=root,
+                )
+                append = children.append
+                num_items = len(obj)
+                last_item_index = num_items - 1
+
+                if _safe_isinstance(obj, _MAPPING_CONTAINERS):
+                    iter_items = iter(obj.items())
+                    if max_length is not None:
+                        iter_items = islice(iter_items, max_length)
+                    for index, (key, child) in enumerate(iter_items):
+                        child_node = _traverse(child, depth=depth + 1)
+                        child_node.key_repr = to_repr(key)
+                        child_node.last = index == last_item_index
+                        append(child_node)
+                else:
+                    iter_values = iter(obj)
+                    if max_length is not None:
+                        iter_values = islice(iter_values, max_length)
+                    for index, child in enumerate(iter_values):
+                        child_node = _traverse(child, depth=depth + 1)
+                        child_node.last = index == last_item_index
+                        append(child_node)
+                if max_length is not None and num_items > max_length:
+                    append(Node(value_repr=f"... +{num_items - max_length}", last=True))
+            else:
+                node = Node(empty=empty, children=[], last=root)
+
+            pop_visited(obj_id)
+        else:
+            node = Node(value_repr=to_repr(obj), last=root)
+        node.is_tuple = type(obj) == tuple
+        node.is_namedtuple = _is_namedtuple(obj)
+        return node
+
+    node = _traverse(_object, root=True)
+    return node
+
+
+def pretty_repr(
+    _object: Any,
+    *,
+    max_width: int = 80,
+    indent_size: int = 4,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+    max_depth: Optional[int] = None,
+    expand_all: bool = False,
+) -> str:
+    """Prettify repr string by expanding on to new lines to fit within a given width.
+
+    Args:
+        _object (Any): Object to repr.
+        max_width (int, optional): Desired maximum width of repr string. Defaults to 80.
+        indent_size (int, optional): Number of spaces to indent. Defaults to 4.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of string before truncating, or None to disable truncating.
+            Defaults to None.
+        max_depth (int, optional): Maximum depth of nested data structure, or None for no depth.
+            Defaults to None.
+        expand_all (bool, optional): Expand all containers regardless of available width. Defaults to False.
+
+    Returns:
+        str: A possibly multi-line representation of the object.
+    """
+
+    if _safe_isinstance(_object, Node):
+        node = _object
+    else:
+        node = traverse(
+            _object, max_length=max_length, max_string=max_string, max_depth=max_depth
+        )
+    repr_str: str = node.render(
+        max_width=max_width, indent_size=indent_size, expand_all=expand_all
+    )
+    return repr_str
+
+
+def pprint(
+    _object: Any,
+    *,
+    console: Optional["Console"] = None,
+    indent_guides: bool = True,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+    max_depth: Optional[int] = None,
+    expand_all: bool = False,
+) -> None:
+    """A convenience function for pretty printing.
+
+    Args:
+        _object (Any): Object to pretty print.
+        console (Console, optional): Console instance, or None to use default. Defaults to None.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of strings before truncating, or None to disable. Defaults to None.
+        max_depth (int, optional): Maximum depth for nested data structures, or None for unlimited depth. Defaults to None.
+        indent_guides (bool, optional): Enable indentation guides. Defaults to True.
+        expand_all (bool, optional): Expand all containers. Defaults to False.
+    """
+    _console = get_console() if console is None else console
+    _console.print(
+        Pretty(
+            _object,
+            max_length=max_length,
+            max_string=max_string,
+            max_depth=max_depth,
+            indent_guides=indent_guides,
+            expand_all=expand_all,
+            overflow="ignore",
+        ),
+        soft_wrap=True,
+    )
+
+
+if __name__ == "__main__":  # pragma: no cover
+
+    class BrokenRepr:
+        def __repr__(self) -> str:
+            1 / 0
+            return "this will fail"
+
+    from typing import NamedTuple
+
+    class StockKeepingUnit(NamedTuple):
+        name: str
+        description: str
+        price: float
+        category: str
+        reviews: List[str]
+
+    d = defaultdict(int)
+    d["foo"] = 5
+    data = {
+        "foo": [
+            1,
+            "Hello World!",
+            100.123,
+            323.232,
+            432324.0,
+            {5, 6, 7, (1, 2, 3, 4), 8},
+        ],
+        "bar": frozenset({1, 2, 3}),
+        "defaultdict": defaultdict(
+            list, {"crumble": ["apple", "rhubarb", "butter", "sugar", "flour"]}
+        ),
+        "counter": Counter(
+            [
+                "apple",
+                "orange",
+                "pear",
+                "kumquat",
+                "kumquat",
+                "durian" * 100,
+            ]
+        ),
+        "atomic": (False, True, None),
+        "namedtuple": StockKeepingUnit(
+            "Sparkling British Spring Water",
+            "Carbonated spring water",
+            0.9,
+            "water",
+            ["its amazing!", "its terrible!"],
+        ),
+        "Broken": BrokenRepr(),
+    }
+    data["foo"].append(data)  # type: ignore[attr-defined]
+
+    from pip._vendor.rich import print
+
+    print(Pretty(data, indent_guides=True, max_string=20))
+
+    class Thing:
+        def __repr__(self) -> str:
+            return "Hello\x1b[38;5;239m World!"
+
+    print(Pretty(Thing()))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress.py
new file mode 100644
index 00000000..ec086d98
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress.py
@@ -0,0 +1,1715 @@
+import io
+import sys
+import typing
+import warnings
+from abc import ABC, abstractmethod
+from collections import deque
+from dataclasses import dataclass, field
+from datetime import timedelta
+from io import RawIOBase, UnsupportedOperation
+from math import ceil
+from mmap import mmap
+from operator import length_hint
+from os import PathLike, stat
+from threading import Event, RLock, Thread
+from types import TracebackType
+from typing import (
+    Any,
+    BinaryIO,
+    Callable,
+    ContextManager,
+    Deque,
+    Dict,
+    Generic,
+    Iterable,
+    List,
+    NamedTuple,
+    NewType,
+    Optional,
+    Sequence,
+    TextIO,
+    Tuple,
+    Type,
+    TypeVar,
+    Union,
+)
+
+if sys.version_info >= (3, 8):
+    from typing import Literal
+else:
+    from pip._vendor.typing_extensions import Literal  # pragma: no cover
+
+if sys.version_info >= (3, 11):
+    from typing import Self
+else:
+    from pip._vendor.typing_extensions import Self  # pragma: no cover
+
+from . import filesize, get_console
+from .console import Console, Group, JustifyMethod, RenderableType
+from .highlighter import Highlighter
+from .jupyter import JupyterMixin
+from .live import Live
+from .progress_bar import ProgressBar
+from .spinner import Spinner
+from .style import StyleType
+from .table import Column, Table
+from .text import Text, TextType
+
+TaskID = NewType("TaskID", int)
+
+ProgressType = TypeVar("ProgressType")
+
+GetTimeCallable = Callable[[], float]
+
+
+_I = typing.TypeVar("_I", TextIO, BinaryIO)
+
+
+class _TrackThread(Thread):
+    """A thread to periodically update progress."""
+
+    def __init__(self, progress: "Progress", task_id: "TaskID", update_period: float):
+        self.progress = progress
+        self.task_id = task_id
+        self.update_period = update_period
+        self.done = Event()
+
+        self.completed = 0
+        super().__init__(daemon=True)
+
+    def run(self) -> None:
+        task_id = self.task_id
+        advance = self.progress.advance
+        update_period = self.update_period
+        last_completed = 0
+        wait = self.done.wait
+        while not wait(update_period) and self.progress.live.is_started:
+            completed = self.completed
+            if last_completed != completed:
+                advance(task_id, completed - last_completed)
+                last_completed = completed
+
+        self.progress.update(self.task_id, completed=self.completed, refresh=True)
+
+    def __enter__(self) -> "_TrackThread":
+        self.start()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.done.set()
+        self.join()
+
+
+def track(
+    sequence: Union[Sequence[ProgressType], Iterable[ProgressType]],
+    description: str = "Working...",
+    total: Optional[float] = None,
+    completed: int = 0,
+    auto_refresh: bool = True,
+    console: Optional[Console] = None,
+    transient: bool = False,
+    get_time: Optional[Callable[[], float]] = None,
+    refresh_per_second: float = 10,
+    style: StyleType = "bar.back",
+    complete_style: StyleType = "bar.complete",
+    finished_style: StyleType = "bar.finished",
+    pulse_style: StyleType = "bar.pulse",
+    update_period: float = 0.1,
+    disable: bool = False,
+    show_speed: bool = True,
+) -> Iterable[ProgressType]:
+    """Track progress by iterating over a sequence.
+
+    Args:
+        sequence (Iterable[ProgressType]): A sequence (must support "len") you wish to iterate over.
+        description (str, optional): Description of task show next to progress bar. Defaults to "Working".
+        total: (float, optional): Total number of steps. Default is len(sequence).
+        completed (int, optional): Number of steps completed so far. Defaults to 0.
+        auto_refresh (bool, optional): Automatic refresh, disable to force a refresh after each iteration. Default is True.
+        transient: (bool, optional): Clear the progress on exit. Defaults to False.
+        console (Console, optional): Console to write to. Default creates internal Console instance.
+        refresh_per_second (float): Number of times per second to refresh the progress information. Defaults to 10.
+        style (StyleType, optional): Style for the bar background. Defaults to "bar.back".
+        complete_style (StyleType, optional): Style for the completed bar. Defaults to "bar.complete".
+        finished_style (StyleType, optional): Style for a finished bar. Defaults to "bar.finished".
+        pulse_style (StyleType, optional): Style for pulsing bars. Defaults to "bar.pulse".
+        update_period (float, optional): Minimum time (in seconds) between calls to update(). Defaults to 0.1.
+        disable (bool, optional): Disable display of progress.
+        show_speed (bool, optional): Show speed if total isn't known. Defaults to True.
+    Returns:
+        Iterable[ProgressType]: An iterable of the values in the sequence.
+
+    """
+
+    columns: List["ProgressColumn"] = (
+        [TextColumn("[progress.description]{task.description}")] if description else []
+    )
+    columns.extend(
+        (
+            BarColumn(
+                style=style,
+                complete_style=complete_style,
+                finished_style=finished_style,
+                pulse_style=pulse_style,
+            ),
+            TaskProgressColumn(show_speed=show_speed),
+            TimeRemainingColumn(elapsed_when_finished=True),
+        )
+    )
+    progress = Progress(
+        *columns,
+        auto_refresh=auto_refresh,
+        console=console,
+        transient=transient,
+        get_time=get_time,
+        refresh_per_second=refresh_per_second or 10,
+        disable=disable,
+    )
+
+    with progress:
+        yield from progress.track(
+            sequence,
+            total=total,
+            completed=completed,
+            description=description,
+            update_period=update_period,
+        )
+
+
+class _Reader(RawIOBase, BinaryIO):
+    """A reader that tracks progress while it's being read from."""
+
+    def __init__(
+        self,
+        handle: BinaryIO,
+        progress: "Progress",
+        task: TaskID,
+        close_handle: bool = True,
+    ) -> None:
+        self.handle = handle
+        self.progress = progress
+        self.task = task
+        self.close_handle = close_handle
+        self._closed = False
+
+    def __enter__(self) -> "_Reader":
+        self.handle.__enter__()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.close()
+
+    def __iter__(self) -> BinaryIO:
+        return self
+
+    def __next__(self) -> bytes:
+        line = next(self.handle)
+        self.progress.advance(self.task, advance=len(line))
+        return line
+
+    @property
+    def closed(self) -> bool:
+        return self._closed
+
+    def fileno(self) -> int:
+        return self.handle.fileno()
+
+    def isatty(self) -> bool:
+        return self.handle.isatty()
+
+    @property
+    def mode(self) -> str:
+        return self.handle.mode
+
+    @property
+    def name(self) -> str:
+        return self.handle.name
+
+    def readable(self) -> bool:
+        return self.handle.readable()
+
+    def seekable(self) -> bool:
+        return self.handle.seekable()
+
+    def writable(self) -> bool:
+        return False
+
+    def read(self, size: int = -1) -> bytes:
+        block = self.handle.read(size)
+        self.progress.advance(self.task, advance=len(block))
+        return block
+
+    def readinto(self, b: Union[bytearray, memoryview, mmap]):  # type: ignore[no-untyped-def, override]
+        n = self.handle.readinto(b)  # type: ignore[attr-defined]
+        self.progress.advance(self.task, advance=n)
+        return n
+
+    def readline(self, size: int = -1) -> bytes:  # type: ignore[override]
+        line = self.handle.readline(size)
+        self.progress.advance(self.task, advance=len(line))
+        return line
+
+    def readlines(self, hint: int = -1) -> List[bytes]:
+        lines = self.handle.readlines(hint)
+        self.progress.advance(self.task, advance=sum(map(len, lines)))
+        return lines
+
+    def close(self) -> None:
+        if self.close_handle:
+            self.handle.close()
+        self._closed = True
+
+    def seek(self, offset: int, whence: int = 0) -> int:
+        pos = self.handle.seek(offset, whence)
+        self.progress.update(self.task, completed=pos)
+        return pos
+
+    def tell(self) -> int:
+        return self.handle.tell()
+
+    def write(self, s: Any) -> int:
+        raise UnsupportedOperation("write")
+
+    def writelines(self, lines: Iterable[Any]) -> None:
+        raise UnsupportedOperation("writelines")
+
+
+class _ReadContext(ContextManager[_I], Generic[_I]):
+    """A utility class to handle a context for both a reader and a progress."""
+
+    def __init__(self, progress: "Progress", reader: _I) -> None:
+        self.progress = progress
+        self.reader: _I = reader
+
+    def __enter__(self) -> _I:
+        self.progress.start()
+        return self.reader.__enter__()
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.progress.stop()
+        self.reader.__exit__(exc_type, exc_val, exc_tb)
+
+
+def wrap_file(
+    file: BinaryIO,
+    total: int,
+    *,
+    description: str = "Reading...",
+    auto_refresh: bool = True,
+    console: Optional[Console] = None,
+    transient: bool = False,
+    get_time: Optional[Callable[[], float]] = None,
+    refresh_per_second: float = 10,
+    style: StyleType = "bar.back",
+    complete_style: StyleType = "bar.complete",
+    finished_style: StyleType = "bar.finished",
+    pulse_style: StyleType = "bar.pulse",
+    disable: bool = False,
+) -> ContextManager[BinaryIO]:
+    """Read bytes from a file while tracking progress.
+
+    Args:
+        file (Union[str, PathLike[str], BinaryIO]): The path to the file to read, or a file-like object in binary mode.
+        total (int): Total number of bytes to read.
+        description (str, optional): Description of task show next to progress bar. Defaults to "Reading".
+        auto_refresh (bool, optional): Automatic refresh, disable to force a refresh after each iteration. Default is True.
+        transient: (bool, optional): Clear the progress on exit. Defaults to False.
+        console (Console, optional): Console to write to. Default creates internal Console instance.
+        refresh_per_second (float): Number of times per second to refresh the progress information. Defaults to 10.
+        style (StyleType, optional): Style for the bar background. Defaults to "bar.back".
+        complete_style (StyleType, optional): Style for the completed bar. Defaults to "bar.complete".
+        finished_style (StyleType, optional): Style for a finished bar. Defaults to "bar.finished".
+        pulse_style (StyleType, optional): Style for pulsing bars. Defaults to "bar.pulse".
+        disable (bool, optional): Disable display of progress.
+    Returns:
+        ContextManager[BinaryIO]: A context manager yielding a progress reader.
+
+    """
+
+    columns: List["ProgressColumn"] = (
+        [TextColumn("[progress.description]{task.description}")] if description else []
+    )
+    columns.extend(
+        (
+            BarColumn(
+                style=style,
+                complete_style=complete_style,
+                finished_style=finished_style,
+                pulse_style=pulse_style,
+            ),
+            DownloadColumn(),
+            TimeRemainingColumn(),
+        )
+    )
+    progress = Progress(
+        *columns,
+        auto_refresh=auto_refresh,
+        console=console,
+        transient=transient,
+        get_time=get_time,
+        refresh_per_second=refresh_per_second or 10,
+        disable=disable,
+    )
+
+    reader = progress.wrap_file(file, total=total, description=description)
+    return _ReadContext(progress, reader)
+
+
+@typing.overload
+def open(
+    file: Union[str, "PathLike[str]", bytes],
+    mode: Union[Literal["rt"], Literal["r"]],
+    buffering: int = -1,
+    encoding: Optional[str] = None,
+    errors: Optional[str] = None,
+    newline: Optional[str] = None,
+    *,
+    total: Optional[int] = None,
+    description: str = "Reading...",
+    auto_refresh: bool = True,
+    console: Optional[Console] = None,
+    transient: bool = False,
+    get_time: Optional[Callable[[], float]] = None,
+    refresh_per_second: float = 10,
+    style: StyleType = "bar.back",
+    complete_style: StyleType = "bar.complete",
+    finished_style: StyleType = "bar.finished",
+    pulse_style: StyleType = "bar.pulse",
+    disable: bool = False,
+) -> ContextManager[TextIO]:
+    pass
+
+
+@typing.overload
+def open(
+    file: Union[str, "PathLike[str]", bytes],
+    mode: Literal["rb"],
+    buffering: int = -1,
+    encoding: Optional[str] = None,
+    errors: Optional[str] = None,
+    newline: Optional[str] = None,
+    *,
+    total: Optional[int] = None,
+    description: str = "Reading...",
+    auto_refresh: bool = True,
+    console: Optional[Console] = None,
+    transient: bool = False,
+    get_time: Optional[Callable[[], float]] = None,
+    refresh_per_second: float = 10,
+    style: StyleType = "bar.back",
+    complete_style: StyleType = "bar.complete",
+    finished_style: StyleType = "bar.finished",
+    pulse_style: StyleType = "bar.pulse",
+    disable: bool = False,
+) -> ContextManager[BinaryIO]:
+    pass
+
+
+def open(
+    file: Union[str, "PathLike[str]", bytes],
+    mode: Union[Literal["rb"], Literal["rt"], Literal["r"]] = "r",
+    buffering: int = -1,
+    encoding: Optional[str] = None,
+    errors: Optional[str] = None,
+    newline: Optional[str] = None,
+    *,
+    total: Optional[int] = None,
+    description: str = "Reading...",
+    auto_refresh: bool = True,
+    console: Optional[Console] = None,
+    transient: bool = False,
+    get_time: Optional[Callable[[], float]] = None,
+    refresh_per_second: float = 10,
+    style: StyleType = "bar.back",
+    complete_style: StyleType = "bar.complete",
+    finished_style: StyleType = "bar.finished",
+    pulse_style: StyleType = "bar.pulse",
+    disable: bool = False,
+) -> Union[ContextManager[BinaryIO], ContextManager[TextIO]]:
+    """Read bytes from a file while tracking progress.
+
+    Args:
+        path (Union[str, PathLike[str], BinaryIO]): The path to the file to read, or a file-like object in binary mode.
+        mode (str): The mode to use to open the file. Only supports "r", "rb" or "rt".
+        buffering (int): The buffering strategy to use, see :func:`io.open`.
+        encoding (str, optional): The encoding to use when reading in text mode, see :func:`io.open`.
+        errors (str, optional): The error handling strategy for decoding errors, see :func:`io.open`.
+        newline (str, optional): The strategy for handling newlines in text mode, see :func:`io.open`
+        total: (int, optional): Total number of bytes to read. Must be provided if reading from a file handle. Default for a path is os.stat(file).st_size.
+        description (str, optional): Description of task show next to progress bar. Defaults to "Reading".
+        auto_refresh (bool, optional): Automatic refresh, disable to force a refresh after each iteration. Default is True.
+        transient: (bool, optional): Clear the progress on exit. Defaults to False.
+        console (Console, optional): Console to write to. Default creates internal Console instance.
+        refresh_per_second (float): Number of times per second to refresh the progress information. Defaults to 10.
+        style (StyleType, optional): Style for the bar background. Defaults to "bar.back".
+        complete_style (StyleType, optional): Style for the completed bar. Defaults to "bar.complete".
+        finished_style (StyleType, optional): Style for a finished bar. Defaults to "bar.finished".
+        pulse_style (StyleType, optional): Style for pulsing bars. Defaults to "bar.pulse".
+        disable (bool, optional): Disable display of progress.
+        encoding (str, optional): The encoding to use when reading in text mode.
+
+    Returns:
+        ContextManager[BinaryIO]: A context manager yielding a progress reader.
+
+    """
+
+    columns: List["ProgressColumn"] = (
+        [TextColumn("[progress.description]{task.description}")] if description else []
+    )
+    columns.extend(
+        (
+            BarColumn(
+                style=style,
+                complete_style=complete_style,
+                finished_style=finished_style,
+                pulse_style=pulse_style,
+            ),
+            DownloadColumn(),
+            TimeRemainingColumn(),
+        )
+    )
+    progress = Progress(
+        *columns,
+        auto_refresh=auto_refresh,
+        console=console,
+        transient=transient,
+        get_time=get_time,
+        refresh_per_second=refresh_per_second or 10,
+        disable=disable,
+    )
+
+    reader = progress.open(
+        file,
+        mode=mode,
+        buffering=buffering,
+        encoding=encoding,
+        errors=errors,
+        newline=newline,
+        total=total,
+        description=description,
+    )
+    return _ReadContext(progress, reader)  # type: ignore[return-value, type-var]
+
+
+class ProgressColumn(ABC):
+    """Base class for a widget to use in progress display."""
+
+    max_refresh: Optional[float] = None
+
+    def __init__(self, table_column: Optional[Column] = None) -> None:
+        self._table_column = table_column
+        self._renderable_cache: Dict[TaskID, Tuple[float, RenderableType]] = {}
+        self._update_time: Optional[float] = None
+
+    def get_table_column(self) -> Column:
+        """Get a table column, used to build tasks table."""
+        return self._table_column or Column()
+
+    def __call__(self, task: "Task") -> RenderableType:
+        """Called by the Progress object to return a renderable for the given task.
+
+        Args:
+            task (Task): An object containing information regarding the task.
+
+        Returns:
+            RenderableType: Anything renderable (including str).
+        """
+        current_time = task.get_time()
+        if self.max_refresh is not None and not task.completed:
+            try:
+                timestamp, renderable = self._renderable_cache[task.id]
+            except KeyError:
+                pass
+            else:
+                if timestamp + self.max_refresh > current_time:
+                    return renderable
+
+        renderable = self.render(task)
+        self._renderable_cache[task.id] = (current_time, renderable)
+        return renderable
+
+    @abstractmethod
+    def render(self, task: "Task") -> RenderableType:
+        """Should return a renderable object."""
+
+
+class RenderableColumn(ProgressColumn):
+    """A column to insert an arbitrary column.
+
+    Args:
+        renderable (RenderableType, optional): Any renderable. Defaults to empty string.
+    """
+
+    def __init__(
+        self, renderable: RenderableType = "", *, table_column: Optional[Column] = None
+    ):
+        self.renderable = renderable
+        super().__init__(table_column=table_column)
+
+    def render(self, task: "Task") -> RenderableType:
+        return self.renderable
+
+
+class SpinnerColumn(ProgressColumn):
+    """A column with a 'spinner' animation.
+
+    Args:
+        spinner_name (str, optional): Name of spinner animation. Defaults to "dots".
+        style (StyleType, optional): Style of spinner. Defaults to "progress.spinner".
+        speed (float, optional): Speed factor of spinner. Defaults to 1.0.
+        finished_text (TextType, optional): Text used when task is finished. Defaults to " ".
+    """
+
+    def __init__(
+        self,
+        spinner_name: str = "dots",
+        style: Optional[StyleType] = "progress.spinner",
+        speed: float = 1.0,
+        finished_text: TextType = " ",
+        table_column: Optional[Column] = None,
+    ):
+        self.spinner = Spinner(spinner_name, style=style, speed=speed)
+        self.finished_text = (
+            Text.from_markup(finished_text)
+            if isinstance(finished_text, str)
+            else finished_text
+        )
+        super().__init__(table_column=table_column)
+
+    def set_spinner(
+        self,
+        spinner_name: str,
+        spinner_style: Optional[StyleType] = "progress.spinner",
+        speed: float = 1.0,
+    ) -> None:
+        """Set a new spinner.
+
+        Args:
+            spinner_name (str): Spinner name, see python -m rich.spinner.
+            spinner_style (Optional[StyleType], optional): Spinner style. Defaults to "progress.spinner".
+            speed (float, optional): Speed factor of spinner. Defaults to 1.0.
+        """
+        self.spinner = Spinner(spinner_name, style=spinner_style, speed=speed)
+
+    def render(self, task: "Task") -> RenderableType:
+        text = (
+            self.finished_text
+            if task.finished
+            else self.spinner.render(task.get_time())
+        )
+        return text
+
+
+class TextColumn(ProgressColumn):
+    """A column containing text."""
+
+    def __init__(
+        self,
+        text_format: str,
+        style: StyleType = "none",
+        justify: JustifyMethod = "left",
+        markup: bool = True,
+        highlighter: Optional[Highlighter] = None,
+        table_column: Optional[Column] = None,
+    ) -> None:
+        self.text_format = text_format
+        self.justify: JustifyMethod = justify
+        self.style = style
+        self.markup = markup
+        self.highlighter = highlighter
+        super().__init__(table_column=table_column or Column(no_wrap=True))
+
+    def render(self, task: "Task") -> Text:
+        _text = self.text_format.format(task=task)
+        if self.markup:
+            text = Text.from_markup(_text, style=self.style, justify=self.justify)
+        else:
+            text = Text(_text, style=self.style, justify=self.justify)
+        if self.highlighter:
+            self.highlighter.highlight(text)
+        return text
+
+
+class BarColumn(ProgressColumn):
+    """Renders a visual progress bar.
+
+    Args:
+        bar_width (Optional[int], optional): Width of bar or None for full width. Defaults to 40.
+        style (StyleType, optional): Style for the bar background. Defaults to "bar.back".
+        complete_style (StyleType, optional): Style for the completed bar. Defaults to "bar.complete".
+        finished_style (StyleType, optional): Style for a finished bar. Defaults to "bar.finished".
+        pulse_style (StyleType, optional): Style for pulsing bars. Defaults to "bar.pulse".
+    """
+
+    def __init__(
+        self,
+        bar_width: Optional[int] = 40,
+        style: StyleType = "bar.back",
+        complete_style: StyleType = "bar.complete",
+        finished_style: StyleType = "bar.finished",
+        pulse_style: StyleType = "bar.pulse",
+        table_column: Optional[Column] = None,
+    ) -> None:
+        self.bar_width = bar_width
+        self.style = style
+        self.complete_style = complete_style
+        self.finished_style = finished_style
+        self.pulse_style = pulse_style
+        super().__init__(table_column=table_column)
+
+    def render(self, task: "Task") -> ProgressBar:
+        """Gets a progress bar widget for a task."""
+        return ProgressBar(
+            total=max(0, task.total) if task.total is not None else None,
+            completed=max(0, task.completed),
+            width=None if self.bar_width is None else max(1, self.bar_width),
+            pulse=not task.started,
+            animation_time=task.get_time(),
+            style=self.style,
+            complete_style=self.complete_style,
+            finished_style=self.finished_style,
+            pulse_style=self.pulse_style,
+        )
+
+
+class TimeElapsedColumn(ProgressColumn):
+    """Renders time elapsed."""
+
+    def render(self, task: "Task") -> Text:
+        """Show time elapsed."""
+        elapsed = task.finished_time if task.finished else task.elapsed
+        if elapsed is None:
+            return Text("-:--:--", style="progress.elapsed")
+        delta = timedelta(seconds=max(0, int(elapsed)))
+        return Text(str(delta), style="progress.elapsed")
+
+
+class TaskProgressColumn(TextColumn):
+    """Show task progress as a percentage.
+
+    Args:
+        text_format (str, optional): Format for percentage display. Defaults to "[progress.percentage]{task.percentage:>3.0f}%".
+        text_format_no_percentage (str, optional): Format if percentage is unknown. Defaults to "".
+        style (StyleType, optional): Style of output. Defaults to "none".
+        justify (JustifyMethod, optional): Text justification. Defaults to "left".
+        markup (bool, optional): Enable markup. Defaults to True.
+        highlighter (Optional[Highlighter], optional): Highlighter to apply to output. Defaults to None.
+        table_column (Optional[Column], optional): Table Column to use. Defaults to None.
+        show_speed (bool, optional): Show speed if total is unknown. Defaults to False.
+    """
+
+    def __init__(
+        self,
+        text_format: str = "[progress.percentage]{task.percentage:>3.0f}%",
+        text_format_no_percentage: str = "",
+        style: StyleType = "none",
+        justify: JustifyMethod = "left",
+        markup: bool = True,
+        highlighter: Optional[Highlighter] = None,
+        table_column: Optional[Column] = None,
+        show_speed: bool = False,
+    ) -> None:
+        self.text_format_no_percentage = text_format_no_percentage
+        self.show_speed = show_speed
+        super().__init__(
+            text_format=text_format,
+            style=style,
+            justify=justify,
+            markup=markup,
+            highlighter=highlighter,
+            table_column=table_column,
+        )
+
+    @classmethod
+    def render_speed(cls, speed: Optional[float]) -> Text:
+        """Render the speed in iterations per second.
+
+        Args:
+            task (Task): A Task object.
+
+        Returns:
+            Text: Text object containing the task speed.
+        """
+        if speed is None:
+            return Text("", style="progress.percentage")
+        unit, suffix = filesize.pick_unit_and_suffix(
+            int(speed),
+            ["", "×10³", "×10⁶", "×10⁹", "×10¹²"],
+            1000,
+        )
+        data_speed = speed / unit
+        return Text(f"{data_speed:.1f}{suffix} it/s", style="progress.percentage")
+
+    def render(self, task: "Task") -> Text:
+        if task.total is None and self.show_speed:
+            return self.render_speed(task.finished_speed or task.speed)
+        text_format = (
+            self.text_format_no_percentage if task.total is None else self.text_format
+        )
+        _text = text_format.format(task=task)
+        if self.markup:
+            text = Text.from_markup(_text, style=self.style, justify=self.justify)
+        else:
+            text = Text(_text, style=self.style, justify=self.justify)
+        if self.highlighter:
+            self.highlighter.highlight(text)
+        return text
+
+
+class TimeRemainingColumn(ProgressColumn):
+    """Renders estimated time remaining.
+
+    Args:
+        compact (bool, optional): Render MM:SS when time remaining is less than an hour. Defaults to False.
+        elapsed_when_finished (bool, optional): Render time elapsed when the task is finished. Defaults to False.
+    """
+
+    # Only refresh twice a second to prevent jitter
+    max_refresh = 0.5
+
+    def __init__(
+        self,
+        compact: bool = False,
+        elapsed_when_finished: bool = False,
+        table_column: Optional[Column] = None,
+    ):
+        self.compact = compact
+        self.elapsed_when_finished = elapsed_when_finished
+        super().__init__(table_column=table_column)
+
+    def render(self, task: "Task") -> Text:
+        """Show time remaining."""
+        if self.elapsed_when_finished and task.finished:
+            task_time = task.finished_time
+            style = "progress.elapsed"
+        else:
+            task_time = task.time_remaining
+            style = "progress.remaining"
+
+        if task.total is None:
+            return Text("", style=style)
+
+        if task_time is None:
+            return Text("--:--" if self.compact else "-:--:--", style=style)
+
+        # Based on https://github.com/tqdm/tqdm/blob/master/tqdm/std.py
+        minutes, seconds = divmod(int(task_time), 60)
+        hours, minutes = divmod(minutes, 60)
+
+        if self.compact and not hours:
+            formatted = f"{minutes:02d}:{seconds:02d}"
+        else:
+            formatted = f"{hours:d}:{minutes:02d}:{seconds:02d}"
+
+        return Text(formatted, style=style)
+
+
+class FileSizeColumn(ProgressColumn):
+    """Renders completed filesize."""
+
+    def render(self, task: "Task") -> Text:
+        """Show data completed."""
+        data_size = filesize.decimal(int(task.completed))
+        return Text(data_size, style="progress.filesize")
+
+
+class TotalFileSizeColumn(ProgressColumn):
+    """Renders total filesize."""
+
+    def render(self, task: "Task") -> Text:
+        """Show data completed."""
+        data_size = filesize.decimal(int(task.total)) if task.total is not None else ""
+        return Text(data_size, style="progress.filesize.total")
+
+
+class MofNCompleteColumn(ProgressColumn):
+    """Renders completed count/total, e.g. '  10/1000'.
+
+    Best for bounded tasks with int quantities.
+
+    Space pads the completed count so that progress length does not change as task progresses
+    past powers of 10.
+
+    Args:
+        separator (str, optional): Text to separate completed and total values. Defaults to "/".
+    """
+
+    def __init__(self, separator: str = "/", table_column: Optional[Column] = None):
+        self.separator = separator
+        super().__init__(table_column=table_column)
+
+    def render(self, task: "Task") -> Text:
+        """Show completed/total."""
+        completed = int(task.completed)
+        total = int(task.total) if task.total is not None else "?"
+        total_width = len(str(total))
+        return Text(
+            f"{completed:{total_width}d}{self.separator}{total}",
+            style="progress.download",
+        )
+
+
+class DownloadColumn(ProgressColumn):
+    """Renders file size downloaded and total, e.g. '0.5/2.3 GB'.
+
+    Args:
+        binary_units (bool, optional): Use binary units, KiB, MiB etc. Defaults to False.
+    """
+
+    def __init__(
+        self, binary_units: bool = False, table_column: Optional[Column] = None
+    ) -> None:
+        self.binary_units = binary_units
+        super().__init__(table_column=table_column)
+
+    def render(self, task: "Task") -> Text:
+        """Calculate common unit for completed and total."""
+        completed = int(task.completed)
+
+        unit_and_suffix_calculation_base = (
+            int(task.total) if task.total is not None else completed
+        )
+        if self.binary_units:
+            unit, suffix = filesize.pick_unit_and_suffix(
+                unit_and_suffix_calculation_base,
+                ["bytes", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB"],
+                1024,
+            )
+        else:
+            unit, suffix = filesize.pick_unit_and_suffix(
+                unit_and_suffix_calculation_base,
+                ["bytes", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"],
+                1000,
+            )
+        precision = 0 if unit == 1 else 1
+
+        completed_ratio = completed / unit
+        completed_str = f"{completed_ratio:,.{precision}f}"
+
+        if task.total is not None:
+            total = int(task.total)
+            total_ratio = total / unit
+            total_str = f"{total_ratio:,.{precision}f}"
+        else:
+            total_str = "?"
+
+        download_status = f"{completed_str}/{total_str} {suffix}"
+        download_text = Text(download_status, style="progress.download")
+        return download_text
+
+
+class TransferSpeedColumn(ProgressColumn):
+    """Renders human readable transfer speed."""
+
+    def render(self, task: "Task") -> Text:
+        """Show data transfer speed."""
+        speed = task.finished_speed or task.speed
+        if speed is None:
+            return Text("?", style="progress.data.speed")
+        data_speed = filesize.decimal(int(speed))
+        return Text(f"{data_speed}/s", style="progress.data.speed")
+
+
+class ProgressSample(NamedTuple):
+    """Sample of progress for a given time."""
+
+    timestamp: float
+    """Timestamp of sample."""
+    completed: float
+    """Number of steps completed."""
+
+
+@dataclass
+class Task:
+    """Information regarding a progress task.
+
+    This object should be considered read-only outside of the :class:`~Progress` class.
+
+    """
+
+    id: TaskID
+    """Task ID associated with this task (used in Progress methods)."""
+
+    description: str
+    """str: Description of the task."""
+
+    total: Optional[float]
+    """Optional[float]: Total number of steps in this task."""
+
+    completed: float
+    """float: Number of steps completed"""
+
+    _get_time: GetTimeCallable
+    """Callable to get the current time."""
+
+    finished_time: Optional[float] = None
+    """float: Time task was finished."""
+
+    visible: bool = True
+    """bool: Indicates if this task is visible in the progress display."""
+
+    fields: Dict[str, Any] = field(default_factory=dict)
+    """dict: Arbitrary fields passed in via Progress.update."""
+
+    start_time: Optional[float] = field(default=None, init=False, repr=False)
+    """Optional[float]: Time this task was started, or None if not started."""
+
+    stop_time: Optional[float] = field(default=None, init=False, repr=False)
+    """Optional[float]: Time this task was stopped, or None if not stopped."""
+
+    finished_speed: Optional[float] = None
+    """Optional[float]: The last speed for a finished task."""
+
+    _progress: Deque[ProgressSample] = field(
+        default_factory=lambda: deque(maxlen=1000), init=False, repr=False
+    )
+
+    _lock: RLock = field(repr=False, default_factory=RLock)
+    """Thread lock."""
+
+    def get_time(self) -> float:
+        """float: Get the current time, in seconds."""
+        return self._get_time()
+
+    @property
+    def started(self) -> bool:
+        """bool: Check if the task as started."""
+        return self.start_time is not None
+
+    @property
+    def remaining(self) -> Optional[float]:
+        """Optional[float]: Get the number of steps remaining, if a non-None total was set."""
+        if self.total is None:
+            return None
+        return self.total - self.completed
+
+    @property
+    def elapsed(self) -> Optional[float]:
+        """Optional[float]: Time elapsed since task was started, or ``None`` if the task hasn't started."""
+        if self.start_time is None:
+            return None
+        if self.stop_time is not None:
+            return self.stop_time - self.start_time
+        return self.get_time() - self.start_time
+
+    @property
+    def finished(self) -> bool:
+        """Check if the task has finished."""
+        return self.finished_time is not None
+
+    @property
+    def percentage(self) -> float:
+        """float: Get progress of task as a percentage. If a None total was set, returns 0"""
+        if not self.total:
+            return 0.0
+        completed = (self.completed / self.total) * 100.0
+        completed = min(100.0, max(0.0, completed))
+        return completed
+
+    @property
+    def speed(self) -> Optional[float]:
+        """Optional[float]: Get the estimated speed in steps per second."""
+        if self.start_time is None:
+            return None
+        with self._lock:
+            progress = self._progress
+            if not progress:
+                return None
+            total_time = progress[-1].timestamp - progress[0].timestamp
+            if total_time == 0:
+                return None
+            iter_progress = iter(progress)
+            next(iter_progress)
+            total_completed = sum(sample.completed for sample in iter_progress)
+            speed = total_completed / total_time
+            return speed
+
+    @property
+    def time_remaining(self) -> Optional[float]:
+        """Optional[float]: Get estimated time to completion, or ``None`` if no data."""
+        if self.finished:
+            return 0.0
+        speed = self.speed
+        if not speed:
+            return None
+        remaining = self.remaining
+        if remaining is None:
+            return None
+        estimate = ceil(remaining / speed)
+        return estimate
+
+    def _reset(self) -> None:
+        """Reset progress."""
+        self._progress.clear()
+        self.finished_time = None
+        self.finished_speed = None
+
+
+class Progress(JupyterMixin):
+    """Renders an auto-updating progress bar(s).
+
+    Args:
+        console (Console, optional): Optional Console instance. Defaults to an internal Console instance writing to stdout.
+        auto_refresh (bool, optional): Enable auto refresh. If disabled, you will need to call `refresh()`.
+        refresh_per_second (Optional[float], optional): Number of times per second to refresh the progress information or None to use default (10). Defaults to None.
+        speed_estimate_period: (float, optional): Period (in seconds) used to calculate the speed estimate. Defaults to 30.
+        transient: (bool, optional): Clear the progress on exit. Defaults to False.
+        redirect_stdout: (bool, optional): Enable redirection of stdout, so ``print`` may be used. Defaults to True.
+        redirect_stderr: (bool, optional): Enable redirection of stderr. Defaults to True.
+        get_time: (Callable, optional): A callable that gets the current time, or None to use Console.get_time. Defaults to None.
+        disable (bool, optional): Disable progress display. Defaults to False
+        expand (bool, optional): Expand tasks table to fit width. Defaults to False.
+    """
+
+    def __init__(
+        self,
+        *columns: Union[str, ProgressColumn],
+        console: Optional[Console] = None,
+        auto_refresh: bool = True,
+        refresh_per_second: float = 10,
+        speed_estimate_period: float = 30.0,
+        transient: bool = False,
+        redirect_stdout: bool = True,
+        redirect_stderr: bool = True,
+        get_time: Optional[GetTimeCallable] = None,
+        disable: bool = False,
+        expand: bool = False,
+    ) -> None:
+        assert refresh_per_second > 0, "refresh_per_second must be > 0"
+        self._lock = RLock()
+        self.columns = columns or self.get_default_columns()
+        self.speed_estimate_period = speed_estimate_period
+
+        self.disable = disable
+        self.expand = expand
+        self._tasks: Dict[TaskID, Task] = {}
+        self._task_index: TaskID = TaskID(0)
+        self.live = Live(
+            console=console or get_console(),
+            auto_refresh=auto_refresh,
+            refresh_per_second=refresh_per_second,
+            transient=transient,
+            redirect_stdout=redirect_stdout,
+            redirect_stderr=redirect_stderr,
+            get_renderable=self.get_renderable,
+        )
+        self.get_time = get_time or self.console.get_time
+        self.print = self.console.print
+        self.log = self.console.log
+
+    @classmethod
+    def get_default_columns(cls) -> Tuple[ProgressColumn, ...]:
+        """Get the default columns used for a new Progress instance:
+           - a text column for the description (TextColumn)
+           - the bar itself (BarColumn)
+           - a text column showing completion percentage (TextColumn)
+           - an estimated-time-remaining column (TimeRemainingColumn)
+        If the Progress instance is created without passing a columns argument,
+        the default columns defined here will be used.
+
+        You can also create a Progress instance using custom columns before
+        and/or after the defaults, as in this example:
+
+            progress = Progress(
+                SpinnerColumn(),
+                *Progress.get_default_columns(),
+                "Elapsed:",
+                TimeElapsedColumn(),
+            )
+
+        This code shows the creation of a Progress display, containing
+        a spinner to the left, the default columns, and a labeled elapsed
+        time column.
+        """
+        return (
+            TextColumn("[progress.description]{task.description}"),
+            BarColumn(),
+            TaskProgressColumn(),
+            TimeRemainingColumn(),
+        )
+
+    @property
+    def console(self) -> Console:
+        return self.live.console
+
+    @property
+    def tasks(self) -> List[Task]:
+        """Get a list of Task instances."""
+        with self._lock:
+            return list(self._tasks.values())
+
+    @property
+    def task_ids(self) -> List[TaskID]:
+        """A list of task IDs."""
+        with self._lock:
+            return list(self._tasks.keys())
+
+    @property
+    def finished(self) -> bool:
+        """Check if all tasks have been completed."""
+        with self._lock:
+            if not self._tasks:
+                return True
+            return all(task.finished for task in self._tasks.values())
+
+    def start(self) -> None:
+        """Start the progress display."""
+        if not self.disable:
+            self.live.start(refresh=True)
+
+    def stop(self) -> None:
+        """Stop the progress display."""
+        self.live.stop()
+        if not self.console.is_interactive and not self.console.is_jupyter:
+            self.console.print()
+
+    def __enter__(self) -> Self:
+        self.start()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.stop()
+
+    def track(
+        self,
+        sequence: Union[Iterable[ProgressType], Sequence[ProgressType]],
+        total: Optional[float] = None,
+        completed: int = 0,
+        task_id: Optional[TaskID] = None,
+        description: str = "Working...",
+        update_period: float = 0.1,
+    ) -> Iterable[ProgressType]:
+        """Track progress by iterating over a sequence.
+
+        Args:
+            sequence (Sequence[ProgressType]): A sequence of values you want to iterate over and track progress.
+            total: (float, optional): Total number of steps. Default is len(sequence).
+            completed (int, optional): Number of steps completed so far. Defaults to 0.
+            task_id: (TaskID): Task to track. Default is new task.
+            description: (str, optional): Description of task, if new task is created.
+            update_period (float, optional): Minimum time (in seconds) between calls to update(). Defaults to 0.1.
+
+        Returns:
+            Iterable[ProgressType]: An iterable of values taken from the provided sequence.
+        """
+        if total is None:
+            total = float(length_hint(sequence)) or None
+
+        if task_id is None:
+            task_id = self.add_task(description, total=total, completed=completed)
+        else:
+            self.update(task_id, total=total, completed=completed)
+
+        if self.live.auto_refresh:
+            with _TrackThread(self, task_id, update_period) as track_thread:
+                for value in sequence:
+                    yield value
+                    track_thread.completed += 1
+        else:
+            advance = self.advance
+            refresh = self.refresh
+            for value in sequence:
+                yield value
+                advance(task_id, 1)
+                refresh()
+
+    def wrap_file(
+        self,
+        file: BinaryIO,
+        total: Optional[int] = None,
+        *,
+        task_id: Optional[TaskID] = None,
+        description: str = "Reading...",
+    ) -> BinaryIO:
+        """Track progress file reading from a binary file.
+
+        Args:
+            file (BinaryIO): A file-like object opened in binary mode.
+            total (int, optional): Total number of bytes to read. This must be provided unless a task with a total is also given.
+            task_id (TaskID): Task to track. Default is new task.
+            description (str, optional): Description of task, if new task is created.
+
+        Returns:
+            BinaryIO: A readable file-like object in binary mode.
+
+        Raises:
+            ValueError: When no total value can be extracted from the arguments or the task.
+        """
+        # attempt to recover the total from the task
+        total_bytes: Optional[float] = None
+        if total is not None:
+            total_bytes = total
+        elif task_id is not None:
+            with self._lock:
+                total_bytes = self._tasks[task_id].total
+        if total_bytes is None:
+            raise ValueError(
+                f"unable to get the total number of bytes, please specify 'total'"
+            )
+
+        # update total of task or create new task
+        if task_id is None:
+            task_id = self.add_task(description, total=total_bytes)
+        else:
+            self.update(task_id, total=total_bytes)
+
+        return _Reader(file, self, task_id, close_handle=False)
+
+    @typing.overload
+    def open(
+        self,
+        file: Union[str, "PathLike[str]", bytes],
+        mode: Literal["rb"],
+        buffering: int = -1,
+        encoding: Optional[str] = None,
+        errors: Optional[str] = None,
+        newline: Optional[str] = None,
+        *,
+        total: Optional[int] = None,
+        task_id: Optional[TaskID] = None,
+        description: str = "Reading...",
+    ) -> BinaryIO:
+        pass
+
+    @typing.overload
+    def open(
+        self,
+        file: Union[str, "PathLike[str]", bytes],
+        mode: Union[Literal["r"], Literal["rt"]],
+        buffering: int = -1,
+        encoding: Optional[str] = None,
+        errors: Optional[str] = None,
+        newline: Optional[str] = None,
+        *,
+        total: Optional[int] = None,
+        task_id: Optional[TaskID] = None,
+        description: str = "Reading...",
+    ) -> TextIO:
+        pass
+
+    def open(
+        self,
+        file: Union[str, "PathLike[str]", bytes],
+        mode: Union[Literal["rb"], Literal["rt"], Literal["r"]] = "r",
+        buffering: int = -1,
+        encoding: Optional[str] = None,
+        errors: Optional[str] = None,
+        newline: Optional[str] = None,
+        *,
+        total: Optional[int] = None,
+        task_id: Optional[TaskID] = None,
+        description: str = "Reading...",
+    ) -> Union[BinaryIO, TextIO]:
+        """Track progress while reading from a binary file.
+
+        Args:
+            path (Union[str, PathLike[str]]): The path to the file to read.
+            mode (str): The mode to use to open the file. Only supports "r", "rb" or "rt".
+            buffering (int): The buffering strategy to use, see :func:`io.open`.
+            encoding (str, optional): The encoding to use when reading in text mode, see :func:`io.open`.
+            errors (str, optional): The error handling strategy for decoding errors, see :func:`io.open`.
+            newline (str, optional): The strategy for handling newlines in text mode, see :func:`io.open`.
+            total (int, optional): Total number of bytes to read. If none given, os.stat(path).st_size is used.
+            task_id (TaskID): Task to track. Default is new task.
+            description (str, optional): Description of task, if new task is created.
+
+        Returns:
+            BinaryIO: A readable file-like object in binary mode.
+
+        Raises:
+            ValueError: When an invalid mode is given.
+        """
+        # normalize the mode (always rb, rt)
+        _mode = "".join(sorted(mode, reverse=False))
+        if _mode not in ("br", "rt", "r"):
+            raise ValueError(f"invalid mode {mode!r}")
+
+        # patch buffering to provide the same behaviour as the builtin `open`
+        line_buffering = buffering == 1
+        if _mode == "br" and buffering == 1:
+            warnings.warn(
+                "line buffering (buffering=1) isn't supported in binary mode, the default buffer size will be used",
+                RuntimeWarning,
+            )
+            buffering = -1
+        elif _mode in ("rt", "r"):
+            if buffering == 0:
+                raise ValueError("can't have unbuffered text I/O")
+            elif buffering == 1:
+                buffering = -1
+
+        # attempt to get the total with `os.stat`
+        if total is None:
+            total = stat(file).st_size
+
+        # update total of task or create new task
+        if task_id is None:
+            task_id = self.add_task(description, total=total)
+        else:
+            self.update(task_id, total=total)
+
+        # open the file in binary mode,
+        handle = io.open(file, "rb", buffering=buffering)
+        reader = _Reader(handle, self, task_id, close_handle=True)
+
+        # wrap the reader in a `TextIOWrapper` if text mode
+        if mode in ("r", "rt"):
+            return io.TextIOWrapper(
+                reader,
+                encoding=encoding,
+                errors=errors,
+                newline=newline,
+                line_buffering=line_buffering,
+            )
+
+        return reader
+
+    def start_task(self, task_id: TaskID) -> None:
+        """Start a task.
+
+        Starts a task (used when calculating elapsed time). You may need to call this manually,
+        if you called ``add_task`` with ``start=False``.
+
+        Args:
+            task_id (TaskID): ID of task.
+        """
+        with self._lock:
+            task = self._tasks[task_id]
+            if task.start_time is None:
+                task.start_time = self.get_time()
+
+    def stop_task(self, task_id: TaskID) -> None:
+        """Stop a task.
+
+        This will freeze the elapsed time on the task.
+
+        Args:
+            task_id (TaskID): ID of task.
+        """
+        with self._lock:
+            task = self._tasks[task_id]
+            current_time = self.get_time()
+            if task.start_time is None:
+                task.start_time = current_time
+            task.stop_time = current_time
+
+    def update(
+        self,
+        task_id: TaskID,
+        *,
+        total: Optional[float] = None,
+        completed: Optional[float] = None,
+        advance: Optional[float] = None,
+        description: Optional[str] = None,
+        visible: Optional[bool] = None,
+        refresh: bool = False,
+        **fields: Any,
+    ) -> None:
+        """Update information associated with a task.
+
+        Args:
+            task_id (TaskID): Task id (returned by add_task).
+            total (float, optional): Updates task.total if not None.
+            completed (float, optional): Updates task.completed if not None.
+            advance (float, optional): Add a value to task.completed if not None.
+            description (str, optional): Change task description if not None.
+            visible (bool, optional): Set visible flag if not None.
+            refresh (bool): Force a refresh of progress information. Default is False.
+            **fields (Any): Additional data fields required for rendering.
+        """
+        with self._lock:
+            task = self._tasks[task_id]
+            completed_start = task.completed
+
+            if total is not None and total != task.total:
+                task.total = total
+                task._reset()
+            if advance is not None:
+                task.completed += advance
+            if completed is not None:
+                task.completed = completed
+            if description is not None:
+                task.description = description
+            if visible is not None:
+                task.visible = visible
+            task.fields.update(fields)
+            update_completed = task.completed - completed_start
+
+            current_time = self.get_time()
+            old_sample_time = current_time - self.speed_estimate_period
+            _progress = task._progress
+
+            popleft = _progress.popleft
+            while _progress and _progress[0].timestamp < old_sample_time:
+                popleft()
+            if update_completed > 0:
+                _progress.append(ProgressSample(current_time, update_completed))
+            if (
+                task.total is not None
+                and task.completed >= task.total
+                and task.finished_time is None
+            ):
+                task.finished_time = task.elapsed
+
+        if refresh:
+            self.refresh()
+
+    def reset(
+        self,
+        task_id: TaskID,
+        *,
+        start: bool = True,
+        total: Optional[float] = None,
+        completed: int = 0,
+        visible: Optional[bool] = None,
+        description: Optional[str] = None,
+        **fields: Any,
+    ) -> None:
+        """Reset a task so completed is 0 and the clock is reset.
+
+        Args:
+            task_id (TaskID): ID of task.
+            start (bool, optional): Start the task after reset. Defaults to True.
+            total (float, optional): New total steps in task, or None to use current total. Defaults to None.
+            completed (int, optional): Number of steps completed. Defaults to 0.
+            visible (bool, optional): Enable display of the task. Defaults to True.
+            description (str, optional): Change task description if not None. Defaults to None.
+            **fields (str): Additional data fields required for rendering.
+        """
+        current_time = self.get_time()
+        with self._lock:
+            task = self._tasks[task_id]
+            task._reset()
+            task.start_time = current_time if start else None
+            if total is not None:
+                task.total = total
+            task.completed = completed
+            if visible is not None:
+                task.visible = visible
+            if fields:
+                task.fields = fields
+            if description is not None:
+                task.description = description
+            task.finished_time = None
+        self.refresh()
+
+    def advance(self, task_id: TaskID, advance: float = 1) -> None:
+        """Advance task by a number of steps.
+
+        Args:
+            task_id (TaskID): ID of task.
+            advance (float): Number of steps to advance. Default is 1.
+        """
+        current_time = self.get_time()
+        with self._lock:
+            task = self._tasks[task_id]
+            completed_start = task.completed
+            task.completed += advance
+            update_completed = task.completed - completed_start
+            old_sample_time = current_time - self.speed_estimate_period
+            _progress = task._progress
+
+            popleft = _progress.popleft
+            while _progress and _progress[0].timestamp < old_sample_time:
+                popleft()
+            while len(_progress) > 1000:
+                popleft()
+            _progress.append(ProgressSample(current_time, update_completed))
+            if (
+                task.total is not None
+                and task.completed >= task.total
+                and task.finished_time is None
+            ):
+                task.finished_time = task.elapsed
+                task.finished_speed = task.speed
+
+    def refresh(self) -> None:
+        """Refresh (render) the progress information."""
+        if not self.disable and self.live.is_started:
+            self.live.refresh()
+
+    def get_renderable(self) -> RenderableType:
+        """Get a renderable for the progress display."""
+        renderable = Group(*self.get_renderables())
+        return renderable
+
+    def get_renderables(self) -> Iterable[RenderableType]:
+        """Get a number of renderables for the progress display."""
+        table = self.make_tasks_table(self.tasks)
+        yield table
+
+    def make_tasks_table(self, tasks: Iterable[Task]) -> Table:
+        """Get a table to render the Progress display.
+
+        Args:
+            tasks (Iterable[Task]): An iterable of Task instances, one per row of the table.
+
+        Returns:
+            Table: A table instance.
+        """
+        table_columns = (
+            (
+                Column(no_wrap=True)
+                if isinstance(_column, str)
+                else _column.get_table_column().copy()
+            )
+            for _column in self.columns
+        )
+        table = Table.grid(*table_columns, padding=(0, 1), expand=self.expand)
+
+        for task in tasks:
+            if task.visible:
+                table.add_row(
+                    *(
+                        (
+                            column.format(task=task)
+                            if isinstance(column, str)
+                            else column(task)
+                        )
+                        for column in self.columns
+                    )
+                )
+        return table
+
+    def __rich__(self) -> RenderableType:
+        """Makes the Progress class itself renderable."""
+        with self._lock:
+            return self.get_renderable()
+
+    def add_task(
+        self,
+        description: str,
+        start: bool = True,
+        total: Optional[float] = 100.0,
+        completed: int = 0,
+        visible: bool = True,
+        **fields: Any,
+    ) -> TaskID:
+        """Add a new 'task' to the Progress display.
+
+        Args:
+            description (str): A description of the task.
+            start (bool, optional): Start the task immediately (to calculate elapsed time). If set to False,
+                you will need to call `start` manually. Defaults to True.
+            total (float, optional): Number of total steps in the progress if known.
+                Set to None to render a pulsing animation. Defaults to 100.
+            completed (int, optional): Number of steps completed so far. Defaults to 0.
+            visible (bool, optional): Enable display of the task. Defaults to True.
+            **fields (str): Additional data fields required for rendering.
+
+        Returns:
+            TaskID: An ID you can use when calling `update`.
+        """
+        with self._lock:
+            task = Task(
+                self._task_index,
+                description,
+                total,
+                completed,
+                visible=visible,
+                fields=fields,
+                _get_time=self.get_time,
+                _lock=self._lock,
+            )
+            self._tasks[self._task_index] = task
+            if start:
+                self.start_task(self._task_index)
+            new_task_index = self._task_index
+            self._task_index = TaskID(int(self._task_index) + 1)
+        self.refresh()
+        return new_task_index
+
+    def remove_task(self, task_id: TaskID) -> None:
+        """Delete a task if it exists.
+
+        Args:
+            task_id (TaskID): A task ID.
+
+        """
+        with self._lock:
+            del self._tasks[task_id]
+
+
+if __name__ == "__main__":  # pragma: no coverage
+    import random
+    import time
+
+    from .panel import Panel
+    from .rule import Rule
+    from .syntax import Syntax
+    from .table import Table
+
+    syntax = Syntax(
+        '''def loop_last(values: Iterable[T]) -> Iterable[Tuple[bool, T]]:
+    """Iterate and generate a tuple with a flag for last value."""
+    iter_values = iter(values)
+    try:
+        previous_value = next(iter_values)
+    except StopIteration:
+        return
+    for value in iter_values:
+        yield False, previous_value
+        previous_value = value
+    yield True, previous_value''',
+        "python",
+        line_numbers=True,
+    )
+
+    table = Table("foo", "bar", "baz")
+    table.add_row("1", "2", "3")
+
+    progress_renderables = [
+        "Text may be printed while the progress bars are rendering.",
+        Panel("In fact, [i]any[/i] renderable will work"),
+        "Such as [magenta]tables[/]...",
+        table,
+        "Pretty printed structures...",
+        {"type": "example", "text": "Pretty printed"},
+        "Syntax...",
+        syntax,
+        Rule("Give it a try!"),
+    ]
+
+    from itertools import cycle
+
+    examples = cycle(progress_renderables)
+
+    console = Console(record=True)
+
+    with Progress(
+        SpinnerColumn(),
+        *Progress.get_default_columns(),
+        TimeElapsedColumn(),
+        console=console,
+        transient=False,
+    ) as progress:
+        task1 = progress.add_task("[red]Downloading", total=1000)
+        task2 = progress.add_task("[green]Processing", total=1000)
+        task3 = progress.add_task("[yellow]Thinking", total=None)
+
+        while not progress.finished:
+            progress.update(task1, advance=0.5)
+            progress.update(task2, advance=0.3)
+            time.sleep(0.01)
+            if random.randint(0, 100) < 1:
+                progress.log(next(examples))
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress_bar.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress_bar.py
new file mode 100644
index 00000000..41794f76
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/progress_bar.py
@@ -0,0 +1,223 @@
+import math
+from functools import lru_cache
+from time import monotonic
+from typing import Iterable, List, Optional
+
+from .color import Color, blend_rgb
+from .color_triplet import ColorTriplet
+from .console import Console, ConsoleOptions, RenderResult
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import Style, StyleType
+
+# Number of characters before 'pulse' animation repeats
+PULSE_SIZE = 20
+
+
+class ProgressBar(JupyterMixin):
+    """Renders a (progress) bar. Used by rich.progress.
+
+    Args:
+        total (float, optional): Number of steps in the bar. Defaults to 100. Set to None to render a pulsing animation.
+        completed (float, optional): Number of steps completed. Defaults to 0.
+        width (int, optional): Width of the bar, or ``None`` for maximum width. Defaults to None.
+        pulse (bool, optional): Enable pulse effect. Defaults to False. Will pulse if a None total was passed.
+        style (StyleType, optional): Style for the bar background. Defaults to "bar.back".
+        complete_style (StyleType, optional): Style for the completed bar. Defaults to "bar.complete".
+        finished_style (StyleType, optional): Style for a finished bar. Defaults to "bar.finished".
+        pulse_style (StyleType, optional): Style for pulsing bars. Defaults to "bar.pulse".
+        animation_time (Optional[float], optional): Time in seconds to use for animation, or None to use system time.
+    """
+
+    def __init__(
+        self,
+        total: Optional[float] = 100.0,
+        completed: float = 0,
+        width: Optional[int] = None,
+        pulse: bool = False,
+        style: StyleType = "bar.back",
+        complete_style: StyleType = "bar.complete",
+        finished_style: StyleType = "bar.finished",
+        pulse_style: StyleType = "bar.pulse",
+        animation_time: Optional[float] = None,
+    ):
+        self.total = total
+        self.completed = completed
+        self.width = width
+        self.pulse = pulse
+        self.style = style
+        self.complete_style = complete_style
+        self.finished_style = finished_style
+        self.pulse_style = pulse_style
+        self.animation_time = animation_time
+
+        self._pulse_segments: Optional[List[Segment]] = None
+
+    def __repr__(self) -> str:
+        return f"<Bar {self.completed!r} of {self.total!r}>"
+
+    @property
+    def percentage_completed(self) -> Optional[float]:
+        """Calculate percentage complete."""
+        if self.total is None:
+            return None
+        completed = (self.completed / self.total) * 100.0
+        completed = min(100, max(0.0, completed))
+        return completed
+
+    @lru_cache(maxsize=16)
+    def _get_pulse_segments(
+        self,
+        fore_style: Style,
+        back_style: Style,
+        color_system: str,
+        no_color: bool,
+        ascii: bool = False,
+    ) -> List[Segment]:
+        """Get a list of segments to render a pulse animation.
+
+        Returns:
+            List[Segment]: A list of segments, one segment per character.
+        """
+        bar = "-" if ascii else "━"
+        segments: List[Segment] = []
+        if color_system not in ("standard", "eight_bit", "truecolor") or no_color:
+            segments += [Segment(bar, fore_style)] * (PULSE_SIZE // 2)
+            segments += [Segment(" " if no_color else bar, back_style)] * (
+                PULSE_SIZE - (PULSE_SIZE // 2)
+            )
+            return segments
+
+        append = segments.append
+        fore_color = (
+            fore_style.color.get_truecolor()
+            if fore_style.color
+            else ColorTriplet(255, 0, 255)
+        )
+        back_color = (
+            back_style.color.get_truecolor()
+            if back_style.color
+            else ColorTriplet(0, 0, 0)
+        )
+        cos = math.cos
+        pi = math.pi
+        _Segment = Segment
+        _Style = Style
+        from_triplet = Color.from_triplet
+
+        for index in range(PULSE_SIZE):
+            position = index / PULSE_SIZE
+            fade = 0.5 + cos(position * pi * 2) / 2.0
+            color = blend_rgb(fore_color, back_color, cross_fade=fade)
+            append(_Segment(bar, _Style(color=from_triplet(color))))
+        return segments
+
+    def update(self, completed: float, total: Optional[float] = None) -> None:
+        """Update progress with new values.
+
+        Args:
+            completed (float): Number of steps completed.
+            total (float, optional): Total number of steps, or ``None`` to not change. Defaults to None.
+        """
+        self.completed = completed
+        self.total = total if total is not None else self.total
+
+    def _render_pulse(
+        self, console: Console, width: int, ascii: bool = False
+    ) -> Iterable[Segment]:
+        """Renders the pulse animation.
+
+        Args:
+            console (Console): Console instance.
+            width (int): Width in characters of pulse animation.
+
+        Returns:
+            RenderResult: [description]
+
+        Yields:
+            Iterator[Segment]: Segments to render pulse
+        """
+        fore_style = console.get_style(self.pulse_style, default="white")
+        back_style = console.get_style(self.style, default="black")
+
+        pulse_segments = self._get_pulse_segments(
+            fore_style, back_style, console.color_system, console.no_color, ascii=ascii
+        )
+        segment_count = len(pulse_segments)
+        current_time = (
+            monotonic() if self.animation_time is None else self.animation_time
+        )
+        segments = pulse_segments * (int(width / segment_count) + 2)
+        offset = int(-current_time * 15) % segment_count
+        segments = segments[offset : offset + width]
+        yield from segments
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        width = min(self.width or options.max_width, options.max_width)
+        ascii = options.legacy_windows or options.ascii_only
+        should_pulse = self.pulse or self.total is None
+        if should_pulse:
+            yield from self._render_pulse(console, width, ascii=ascii)
+            return
+
+        completed: Optional[float] = (
+            min(self.total, max(0, self.completed)) if self.total is not None else None
+        )
+
+        bar = "-" if ascii else "━"
+        half_bar_right = " " if ascii else "╸"
+        half_bar_left = " " if ascii else "╺"
+        complete_halves = (
+            int(width * 2 * completed / self.total)
+            if self.total and completed is not None
+            else width * 2
+        )
+        bar_count = complete_halves // 2
+        half_bar_count = complete_halves % 2
+        style = console.get_style(self.style)
+        is_finished = self.total is None or self.completed >= self.total
+        complete_style = console.get_style(
+            self.finished_style if is_finished else self.complete_style
+        )
+        _Segment = Segment
+        if bar_count:
+            yield _Segment(bar * bar_count, complete_style)
+        if half_bar_count:
+            yield _Segment(half_bar_right * half_bar_count, complete_style)
+
+        if not console.no_color:
+            remaining_bars = width - bar_count - half_bar_count
+            if remaining_bars and console.color_system is not None:
+                if not half_bar_count and bar_count:
+                    yield _Segment(half_bar_left, style)
+                    remaining_bars -= 1
+                if remaining_bars:
+                    yield _Segment(bar * remaining_bars, style)
+
+    def __rich_measure__(
+        self, console: Console, options: ConsoleOptions
+    ) -> Measurement:
+        return (
+            Measurement(self.width, self.width)
+            if self.width is not None
+            else Measurement(4, options.max_width)
+        )
+
+
+if __name__ == "__main__":  # pragma: no cover
+    console = Console()
+    bar = ProgressBar(width=50, total=100)
+
+    import time
+
+    console.show_cursor(False)
+    for n in range(0, 101, 1):
+        bar.update(n)
+        console.print(bar)
+        console.file.write("\r")
+        time.sleep(0.05)
+    console.show_cursor(True)
+    console.print()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/prompt.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/prompt.py
new file mode 100644
index 00000000..fccb70db
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/prompt.py
@@ -0,0 +1,400 @@
+from typing import Any, Generic, List, Optional, TextIO, TypeVar, Union, overload
+
+from . import get_console
+from .console import Console
+from .text import Text, TextType
+
+PromptType = TypeVar("PromptType")
+DefaultType = TypeVar("DefaultType")
+
+
+class PromptError(Exception):
+    """Exception base class for prompt related errors."""
+
+
+class InvalidResponse(PromptError):
+    """Exception to indicate a response was invalid. Raise this within process_response() to indicate an error
+    and provide an error message.
+
+    Args:
+        message (Union[str, Text]): Error message.
+    """
+
+    def __init__(self, message: TextType) -> None:
+        self.message = message
+
+    def __rich__(self) -> TextType:
+        return self.message
+
+
+class PromptBase(Generic[PromptType]):
+    """Ask the user for input until a valid response is received. This is the base class, see one of
+    the concrete classes for examples.
+
+    Args:
+        prompt (TextType, optional): Prompt text. Defaults to "".
+        console (Console, optional): A Console instance or None to use global console. Defaults to None.
+        password (bool, optional): Enable password input. Defaults to False.
+        choices (List[str], optional): A list of valid choices. Defaults to None.
+        case_sensitive (bool, optional): Matching of choices should be case-sensitive. Defaults to True.
+        show_default (bool, optional): Show default in prompt. Defaults to True.
+        show_choices (bool, optional): Show choices in prompt. Defaults to True.
+    """
+
+    response_type: type = str
+
+    validate_error_message = "[prompt.invalid]Please enter a valid value"
+    illegal_choice_message = (
+        "[prompt.invalid.choice]Please select one of the available options"
+    )
+    prompt_suffix = ": "
+
+    choices: Optional[List[str]] = None
+
+    def __init__(
+        self,
+        prompt: TextType = "",
+        *,
+        console: Optional[Console] = None,
+        password: bool = False,
+        choices: Optional[List[str]] = None,
+        case_sensitive: bool = True,
+        show_default: bool = True,
+        show_choices: bool = True,
+    ) -> None:
+        self.console = console or get_console()
+        self.prompt = (
+            Text.from_markup(prompt, style="prompt")
+            if isinstance(prompt, str)
+            else prompt
+        )
+        self.password = password
+        if choices is not None:
+            self.choices = choices
+        self.case_sensitive = case_sensitive
+        self.show_default = show_default
+        self.show_choices = show_choices
+
+    @classmethod
+    @overload
+    def ask(
+        cls,
+        prompt: TextType = "",
+        *,
+        console: Optional[Console] = None,
+        password: bool = False,
+        choices: Optional[List[str]] = None,
+        case_sensitive: bool = True,
+        show_default: bool = True,
+        show_choices: bool = True,
+        default: DefaultType,
+        stream: Optional[TextIO] = None,
+    ) -> Union[DefaultType, PromptType]:
+        ...
+
+    @classmethod
+    @overload
+    def ask(
+        cls,
+        prompt: TextType = "",
+        *,
+        console: Optional[Console] = None,
+        password: bool = False,
+        choices: Optional[List[str]] = None,
+        case_sensitive: bool = True,
+        show_default: bool = True,
+        show_choices: bool = True,
+        stream: Optional[TextIO] = None,
+    ) -> PromptType:
+        ...
+
+    @classmethod
+    def ask(
+        cls,
+        prompt: TextType = "",
+        *,
+        console: Optional[Console] = None,
+        password: bool = False,
+        choices: Optional[List[str]] = None,
+        case_sensitive: bool = True,
+        show_default: bool = True,
+        show_choices: bool = True,
+        default: Any = ...,
+        stream: Optional[TextIO] = None,
+    ) -> Any:
+        """Shortcut to construct and run a prompt loop and return the result.
+
+        Example:
+            >>> filename = Prompt.ask("Enter a filename")
+
+        Args:
+            prompt (TextType, optional): Prompt text. Defaults to "".
+            console (Console, optional): A Console instance or None to use global console. Defaults to None.
+            password (bool, optional): Enable password input. Defaults to False.
+            choices (List[str], optional): A list of valid choices. Defaults to None.
+            case_sensitive (bool, optional): Matching of choices should be case-sensitive. Defaults to True.
+            show_default (bool, optional): Show default in prompt. Defaults to True.
+            show_choices (bool, optional): Show choices in prompt. Defaults to True.
+            stream (TextIO, optional): Optional text file open for reading to get input. Defaults to None.
+        """
+        _prompt = cls(
+            prompt,
+            console=console,
+            password=password,
+            choices=choices,
+            case_sensitive=case_sensitive,
+            show_default=show_default,
+            show_choices=show_choices,
+        )
+        return _prompt(default=default, stream=stream)
+
+    def render_default(self, default: DefaultType) -> Text:
+        """Turn the supplied default in to a Text instance.
+
+        Args:
+            default (DefaultType): Default value.
+
+        Returns:
+            Text: Text containing rendering of default value.
+        """
+        return Text(f"({default})", "prompt.default")
+
+    def make_prompt(self, default: DefaultType) -> Text:
+        """Make prompt text.
+
+        Args:
+            default (DefaultType): Default value.
+
+        Returns:
+            Text: Text to display in prompt.
+        """
+        prompt = self.prompt.copy()
+        prompt.end = ""
+
+        if self.show_choices and self.choices:
+            _choices = "/".join(self.choices)
+            choices = f"[{_choices}]"
+            prompt.append(" ")
+            prompt.append(choices, "prompt.choices")
+
+        if (
+            default != ...
+            and self.show_default
+            and isinstance(default, (str, self.response_type))
+        ):
+            prompt.append(" ")
+            _default = self.render_default(default)
+            prompt.append(_default)
+
+        prompt.append(self.prompt_suffix)
+
+        return prompt
+
+    @classmethod
+    def get_input(
+        cls,
+        console: Console,
+        prompt: TextType,
+        password: bool,
+        stream: Optional[TextIO] = None,
+    ) -> str:
+        """Get input from user.
+
+        Args:
+            console (Console): Console instance.
+            prompt (TextType): Prompt text.
+            password (bool): Enable password entry.
+
+        Returns:
+            str: String from user.
+        """
+        return console.input(prompt, password=password, stream=stream)
+
+    def check_choice(self, value: str) -> bool:
+        """Check value is in the list of valid choices.
+
+        Args:
+            value (str): Value entered by user.
+
+        Returns:
+            bool: True if choice was valid, otherwise False.
+        """
+        assert self.choices is not None
+        if self.case_sensitive:
+            return value.strip() in self.choices
+        return value.strip().lower() in [choice.lower() for choice in self.choices]
+
+    def process_response(self, value: str) -> PromptType:
+        """Process response from user, convert to prompt type.
+
+        Args:
+            value (str): String typed by user.
+
+        Raises:
+            InvalidResponse: If ``value`` is invalid.
+
+        Returns:
+            PromptType: The value to be returned from ask method.
+        """
+        value = value.strip()
+        try:
+            return_value: PromptType = self.response_type(value)
+        except ValueError:
+            raise InvalidResponse(self.validate_error_message)
+
+        if self.choices is not None:
+            if not self.check_choice(value):
+                raise InvalidResponse(self.illegal_choice_message)
+
+            if not self.case_sensitive:
+                # return the original choice, not the lower case version
+                return_value = self.response_type(
+                    self.choices[
+                        [choice.lower() for choice in self.choices].index(value.lower())
+                    ]
+                )
+        return return_value
+
+    def on_validate_error(self, value: str, error: InvalidResponse) -> None:
+        """Called to handle validation error.
+
+        Args:
+            value (str): String entered by user.
+            error (InvalidResponse): Exception instance the initiated the error.
+        """
+        self.console.print(error)
+
+    def pre_prompt(self) -> None:
+        """Hook to display something before the prompt."""
+
+    @overload
+    def __call__(self, *, stream: Optional[TextIO] = None) -> PromptType:
+        ...
+
+    @overload
+    def __call__(
+        self, *, default: DefaultType, stream: Optional[TextIO] = None
+    ) -> Union[PromptType, DefaultType]:
+        ...
+
+    def __call__(self, *, default: Any = ..., stream: Optional[TextIO] = None) -> Any:
+        """Run the prompt loop.
+
+        Args:
+            default (Any, optional): Optional default value.
+
+        Returns:
+            PromptType: Processed value.
+        """
+        while True:
+            self.pre_prompt()
+            prompt = self.make_prompt(default)
+            value = self.get_input(self.console, prompt, self.password, stream=stream)
+            if value == "" and default != ...:
+                return default
+            try:
+                return_value = self.process_response(value)
+            except InvalidResponse as error:
+                self.on_validate_error(value, error)
+                continue
+            else:
+                return return_value
+
+
+class Prompt(PromptBase[str]):
+    """A prompt that returns a str.
+
+    Example:
+        >>> name = Prompt.ask("Enter your name")
+
+
+    """
+
+    response_type = str
+
+
+class IntPrompt(PromptBase[int]):
+    """A prompt that returns an integer.
+
+    Example:
+        >>> burrito_count = IntPrompt.ask("How many burritos do you want to order")
+
+    """
+
+    response_type = int
+    validate_error_message = "[prompt.invalid]Please enter a valid integer number"
+
+
+class FloatPrompt(PromptBase[float]):
+    """A prompt that returns a float.
+
+    Example:
+        >>> temperature = FloatPrompt.ask("Enter desired temperature")
+
+    """
+
+    response_type = float
+    validate_error_message = "[prompt.invalid]Please enter a number"
+
+
+class Confirm(PromptBase[bool]):
+    """A yes / no confirmation prompt.
+
+    Example:
+        >>> if Confirm.ask("Continue"):
+                run_job()
+
+    """
+
+    response_type = bool
+    validate_error_message = "[prompt.invalid]Please enter Y or N"
+    choices: List[str] = ["y", "n"]
+
+    def render_default(self, default: DefaultType) -> Text:
+        """Render the default as (y) or (n) rather than True/False."""
+        yes, no = self.choices
+        return Text(f"({yes})" if default else f"({no})", style="prompt.default")
+
+    def process_response(self, value: str) -> bool:
+        """Convert choices to a bool."""
+        value = value.strip().lower()
+        if value not in self.choices:
+            raise InvalidResponse(self.validate_error_message)
+        return value == self.choices[0]
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich import print
+
+    if Confirm.ask("Run [i]prompt[/i] tests?", default=True):
+        while True:
+            result = IntPrompt.ask(
+                ":rocket: Enter a number between [b]1[/b] and [b]10[/b]", default=5
+            )
+            if result >= 1 and result <= 10:
+                break
+            print(":pile_of_poo: [prompt.invalid]Number must be between 1 and 10")
+        print(f"number={result}")
+
+        while True:
+            password = Prompt.ask(
+                "Please enter a password [cyan](must be at least 5 characters)",
+                password=True,
+            )
+            if len(password) >= 5:
+                break
+            print("[prompt.invalid]password too short")
+        print(f"password={password!r}")
+
+        fruit = Prompt.ask("Enter a fruit", choices=["apple", "orange", "pear"])
+        print(f"fruit={fruit!r}")
+
+        doggie = Prompt.ask(
+            "What's the best Dog? (Case INSENSITIVE)",
+            choices=["Border Terrier", "Collie", "Labradoodle"],
+            case_sensitive=False,
+        )
+        print(f"doggie={doggie!r}")
+
+    else:
+        print("[b]OK :loudly_crying_face:")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/protocol.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/protocol.py
new file mode 100644
index 00000000..12ab2371
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/protocol.py
@@ -0,0 +1,42 @@
+from typing import Any, cast, Set, TYPE_CHECKING
+from inspect import isclass
+
+if TYPE_CHECKING:
+    from pip._vendor.rich.console import RenderableType
+
+_GIBBERISH = """aihwerij235234ljsdnp34ksodfipwoe234234jlskjdf"""
+
+
+def is_renderable(check_object: Any) -> bool:
+    """Check if an object may be rendered by Rich."""
+    return (
+        isinstance(check_object, str)
+        or hasattr(check_object, "__rich__")
+        or hasattr(check_object, "__rich_console__")
+    )
+
+
+def rich_cast(renderable: object) -> "RenderableType":
+    """Cast an object to a renderable by calling __rich__ if present.
+
+    Args:
+        renderable (object): A potentially renderable object
+
+    Returns:
+        object: The result of recursively calling __rich__.
+    """
+    from pip._vendor.rich.console import RenderableType
+
+    rich_visited_set: Set[type] = set()  # Prevent potential infinite loop
+    while hasattr(renderable, "__rich__") and not isclass(renderable):
+        # Detect object which claim to have all the attributes
+        if hasattr(renderable, _GIBBERISH):
+            return repr(renderable)
+        cast_method = getattr(renderable, "__rich__")
+        renderable = cast_method()
+        renderable_type = type(renderable)
+        if renderable_type in rich_visited_set:
+            break
+        rich_visited_set.add(renderable_type)
+
+    return cast(RenderableType, renderable)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/region.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/region.py
new file mode 100644
index 00000000..75b3631c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/region.py
@@ -0,0 +1,10 @@
+from typing import NamedTuple
+
+
+class Region(NamedTuple):
+    """Defines a rectangular region of the screen."""
+
+    x: int
+    y: int
+    width: int
+    height: int
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/repr.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/repr.py
new file mode 100644
index 00000000..10efc427
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/repr.py
@@ -0,0 +1,149 @@
+import inspect
+from functools import partial
+from typing import (
+    Any,
+    Callable,
+    Iterable,
+    List,
+    Optional,
+    Tuple,
+    Type,
+    TypeVar,
+    Union,
+    overload,
+)
+
+T = TypeVar("T")
+
+
+Result = Iterable[Union[Any, Tuple[Any], Tuple[str, Any], Tuple[str, Any, Any]]]
+RichReprResult = Result
+
+
+class ReprError(Exception):
+    """An error occurred when attempting to build a repr."""
+
+
+@overload
+def auto(cls: Optional[Type[T]]) -> Type[T]:
+    ...
+
+
+@overload
+def auto(*, angular: bool = False) -> Callable[[Type[T]], Type[T]]:
+    ...
+
+
+def auto(
+    cls: Optional[Type[T]] = None, *, angular: Optional[bool] = None
+) -> Union[Type[T], Callable[[Type[T]], Type[T]]]:
+    """Class decorator to create __repr__ from __rich_repr__"""
+
+    def do_replace(cls: Type[T], angular: Optional[bool] = None) -> Type[T]:
+        def auto_repr(self: T) -> str:
+            """Create repr string from __rich_repr__"""
+            repr_str: List[str] = []
+            append = repr_str.append
+
+            angular: bool = getattr(self.__rich_repr__, "angular", False)  # type: ignore[attr-defined]
+            for arg in self.__rich_repr__():  # type: ignore[attr-defined]
+                if isinstance(arg, tuple):
+                    if len(arg) == 1:
+                        append(repr(arg[0]))
+                    else:
+                        key, value, *default = arg
+                        if key is None:
+                            append(repr(value))
+                        else:
+                            if default and default[0] == value:
+                                continue
+                            append(f"{key}={value!r}")
+                else:
+                    append(repr(arg))
+            if angular:
+                return f"<{self.__class__.__name__} {' '.join(repr_str)}>"
+            else:
+                return f"{self.__class__.__name__}({', '.join(repr_str)})"
+
+        def auto_rich_repr(self: Type[T]) -> Result:
+            """Auto generate __rich_rep__ from signature of __init__"""
+            try:
+                signature = inspect.signature(self.__init__)
+                for name, param in signature.parameters.items():
+                    if param.kind == param.POSITIONAL_ONLY:
+                        yield getattr(self, name)
+                    elif param.kind in (
+                        param.POSITIONAL_OR_KEYWORD,
+                        param.KEYWORD_ONLY,
+                    ):
+                        if param.default is param.empty:
+                            yield getattr(self, param.name)
+                        else:
+                            yield param.name, getattr(self, param.name), param.default
+            except Exception as error:
+                raise ReprError(
+                    f"Failed to auto generate __rich_repr__; {error}"
+                ) from None
+
+        if not hasattr(cls, "__rich_repr__"):
+            auto_rich_repr.__doc__ = "Build a rich repr"
+            cls.__rich_repr__ = auto_rich_repr  # type: ignore[attr-defined]
+
+        auto_repr.__doc__ = "Return repr(self)"
+        cls.__repr__ = auto_repr  # type: ignore[assignment]
+        if angular is not None:
+            cls.__rich_repr__.angular = angular  # type: ignore[attr-defined]
+        return cls
+
+    if cls is None:
+        return partial(do_replace, angular=angular)
+    else:
+        return do_replace(cls, angular=angular)
+
+
+@overload
+def rich_repr(cls: Optional[Type[T]]) -> Type[T]:
+    ...
+
+
+@overload
+def rich_repr(*, angular: bool = False) -> Callable[[Type[T]], Type[T]]:
+    ...
+
+
+def rich_repr(
+    cls: Optional[Type[T]] = None, *, angular: bool = False
+) -> Union[Type[T], Callable[[Type[T]], Type[T]]]:
+    if cls is None:
+        return auto(angular=angular)
+    else:
+        return auto(cls)
+
+
+if __name__ == "__main__":
+
+    @auto
+    class Foo:
+        def __rich_repr__(self) -> Result:
+            yield "foo"
+            yield "bar", {"shopping": ["eggs", "ham", "pineapple"]}
+            yield "buy", "hand sanitizer"
+
+    foo = Foo()
+    from pip._vendor.rich.console import Console
+
+    console = Console()
+
+    console.rule("Standard repr")
+    console.print(foo)
+
+    console.print(foo, width=60)
+    console.print(foo, width=30)
+
+    console.rule("Angular repr")
+    Foo.__rich_repr__.angular = True  # type: ignore[attr-defined]
+
+    console.print(foo)
+
+    console.print(foo, width=60)
+    console.print(foo, width=30)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/rule.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/rule.py
new file mode 100644
index 00000000..fd00ce6e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/rule.py
@@ -0,0 +1,130 @@
+from typing import Union
+
+from .align import AlignMethod
+from .cells import cell_len, set_cell_size
+from .console import Console, ConsoleOptions, RenderResult
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .style import Style
+from .text import Text
+
+
+class Rule(JupyterMixin):
+    """A console renderable to draw a horizontal rule (line).
+
+    Args:
+        title (Union[str, Text], optional): Text to render in the rule. Defaults to "".
+        characters (str, optional): Character(s) used to draw the line. Defaults to "─".
+        style (StyleType, optional): Style of Rule. Defaults to "rule.line".
+        end (str, optional): Character at end of Rule. defaults to "\\\\n"
+        align (str, optional): How to align the title, one of "left", "center", or "right". Defaults to "center".
+    """
+
+    def __init__(
+        self,
+        title: Union[str, Text] = "",
+        *,
+        characters: str = "─",
+        style: Union[str, Style] = "rule.line",
+        end: str = "\n",
+        align: AlignMethod = "center",
+    ) -> None:
+        if cell_len(characters) < 1:
+            raise ValueError(
+                "'characters' argument must have a cell width of at least 1"
+            )
+        if align not in ("left", "center", "right"):
+            raise ValueError(
+                f'invalid value for align, expected "left", "center", "right" (not {align!r})'
+            )
+        self.title = title
+        self.characters = characters
+        self.style = style
+        self.end = end
+        self.align = align
+
+    def __repr__(self) -> str:
+        return f"Rule({self.title!r}, {self.characters!r})"
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        width = options.max_width
+
+        characters = (
+            "-"
+            if (options.ascii_only and not self.characters.isascii())
+            else self.characters
+        )
+
+        chars_len = cell_len(characters)
+        if not self.title:
+            yield self._rule_line(chars_len, width)
+            return
+
+        if isinstance(self.title, Text):
+            title_text = self.title
+        else:
+            title_text = console.render_str(self.title, style="rule.text")
+
+        title_text.plain = title_text.plain.replace("\n", " ")
+        title_text.expand_tabs()
+
+        required_space = 4 if self.align == "center" else 2
+        truncate_width = max(0, width - required_space)
+        if not truncate_width:
+            yield self._rule_line(chars_len, width)
+            return
+
+        rule_text = Text(end=self.end)
+        if self.align == "center":
+            title_text.truncate(truncate_width, overflow="ellipsis")
+            side_width = (width - cell_len(title_text.plain)) // 2
+            left = Text(characters * (side_width // chars_len + 1))
+            left.truncate(side_width - 1)
+            right_length = width - cell_len(left.plain) - cell_len(title_text.plain)
+            right = Text(characters * (side_width // chars_len + 1))
+            right.truncate(right_length)
+            rule_text.append(left.plain + " ", self.style)
+            rule_text.append(title_text)
+            rule_text.append(" " + right.plain, self.style)
+        elif self.align == "left":
+            title_text.truncate(truncate_width, overflow="ellipsis")
+            rule_text.append(title_text)
+            rule_text.append(" ")
+            rule_text.append(characters * (width - rule_text.cell_len), self.style)
+        elif self.align == "right":
+            title_text.truncate(truncate_width, overflow="ellipsis")
+            rule_text.append(characters * (width - title_text.cell_len - 1), self.style)
+            rule_text.append(" ")
+            rule_text.append(title_text)
+
+        rule_text.plain = set_cell_size(rule_text.plain, width)
+        yield rule_text
+
+    def _rule_line(self, chars_len: int, width: int) -> Text:
+        rule_text = Text(self.characters * ((width // chars_len) + 1), self.style)
+        rule_text.truncate(width)
+        rule_text.plain = set_cell_size(rule_text.plain, width)
+        return rule_text
+
+    def __rich_measure__(
+        self, console: Console, options: ConsoleOptions
+    ) -> Measurement:
+        return Measurement(1, 1)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sys
+
+    from pip._vendor.rich.console import Console
+
+    try:
+        text = sys.argv[1]
+    except IndexError:
+        text = "Hello, World"
+    console = Console()
+    console.print(Rule(title=text))
+
+    console = Console()
+    console.print(Rule("foo"), width=4)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/scope.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/scope.py
new file mode 100644
index 00000000..c9d134cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/scope.py
@@ -0,0 +1,86 @@
+from collections.abc import Mapping
+from typing import TYPE_CHECKING, Any, Optional, Tuple
+
+from .highlighter import ReprHighlighter
+from .panel import Panel
+from .pretty import Pretty
+from .table import Table
+from .text import Text, TextType
+
+if TYPE_CHECKING:
+    from .console import ConsoleRenderable
+
+
+def render_scope(
+    scope: "Mapping[str, Any]",
+    *,
+    title: Optional[TextType] = None,
+    sort_keys: bool = True,
+    indent_guides: bool = False,
+    max_length: Optional[int] = None,
+    max_string: Optional[int] = None,
+) -> "ConsoleRenderable":
+    """Render python variables in a given scope.
+
+    Args:
+        scope (Mapping): A mapping containing variable names and values.
+        title (str, optional): Optional title. Defaults to None.
+        sort_keys (bool, optional): Enable sorting of items. Defaults to True.
+        indent_guides (bool, optional): Enable indentation guides. Defaults to False.
+        max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to None.
+        max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to None.
+
+    Returns:
+        ConsoleRenderable: A renderable object.
+    """
+    highlighter = ReprHighlighter()
+    items_table = Table.grid(padding=(0, 1), expand=False)
+    items_table.add_column(justify="right")
+
+    def sort_items(item: Tuple[str, Any]) -> Tuple[bool, str]:
+        """Sort special variables first, then alphabetically."""
+        key, _ = item
+        return (not key.startswith("__"), key.lower())
+
+    items = sorted(scope.items(), key=sort_items) if sort_keys else scope.items()
+    for key, value in items:
+        key_text = Text.assemble(
+            (key, "scope.key.special" if key.startswith("__") else "scope.key"),
+            (" =", "scope.equals"),
+        )
+        items_table.add_row(
+            key_text,
+            Pretty(
+                value,
+                highlighter=highlighter,
+                indent_guides=indent_guides,
+                max_length=max_length,
+                max_string=max_string,
+            ),
+        )
+    return Panel.fit(
+        items_table,
+        title=title,
+        border_style="scope.border",
+        padding=(0, 1),
+    )
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich import print
+
+    print()
+
+    def test(foo: float, bar: float) -> None:
+        list_of_things = [1, 2, 3, None, 4, True, False, "Hello World"]
+        dict_of_things = {
+            "version": "1.1",
+            "method": "confirmFruitPurchase",
+            "params": [["apple", "orange", "mangoes", "pomelo"], 1.123],
+            "id": "194521489",
+        }
+        print(render_scope(locals(), title="[i]locals", sort_keys=False))
+
+    test(20.3423, 3.1427)
+    print()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/screen.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/screen.py
new file mode 100644
index 00000000..7f416e1e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/screen.py
@@ -0,0 +1,54 @@
+from typing import Optional, TYPE_CHECKING
+
+from .segment import Segment
+from .style import StyleType
+from ._loop import loop_last
+
+
+if TYPE_CHECKING:
+    from .console import (
+        Console,
+        ConsoleOptions,
+        RenderResult,
+        RenderableType,
+        Group,
+    )
+
+
+class Screen:
+    """A renderable that fills the terminal screen and crops excess.
+
+    Args:
+        renderable (RenderableType): Child renderable.
+        style (StyleType, optional): Optional background style. Defaults to None.
+    """
+
+    renderable: "RenderableType"
+
+    def __init__(
+        self,
+        *renderables: "RenderableType",
+        style: Optional[StyleType] = None,
+        application_mode: bool = False,
+    ) -> None:
+        from pip._vendor.rich.console import Group
+
+        self.renderable = Group(*renderables)
+        self.style = style
+        self.application_mode = application_mode
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        width, height = options.size
+        style = console.get_style(self.style) if self.style else None
+        render_options = options.update(width=width, height=height)
+        lines = console.render_lines(
+            self.renderable or "", render_options, style=style, pad=True
+        )
+        lines = Segment.set_shape(lines, width, height, style=style)
+        new_line = Segment("\n\r") if self.application_mode else Segment.line()
+        for last, line in loop_last(lines):
+            yield from line
+            if not last:
+                yield new_line
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/segment.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/segment.py
new file mode 100644
index 00000000..4b5f9979
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/segment.py
@@ -0,0 +1,752 @@
+from enum import IntEnum
+from functools import lru_cache
+from itertools import filterfalse
+from logging import getLogger
+from operator import attrgetter
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Sequence,
+    Tuple,
+    Type,
+    Union,
+)
+
+from .cells import (
+    _is_single_cell_widths,
+    cached_cell_len,
+    cell_len,
+    get_character_cell_size,
+    set_cell_size,
+)
+from .repr import Result, rich_repr
+from .style import Style
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderResult
+
+log = getLogger("rich")
+
+
+class ControlType(IntEnum):
+    """Non-printable control codes which typically translate to ANSI codes."""
+
+    BELL = 1
+    CARRIAGE_RETURN = 2
+    HOME = 3
+    CLEAR = 4
+    SHOW_CURSOR = 5
+    HIDE_CURSOR = 6
+    ENABLE_ALT_SCREEN = 7
+    DISABLE_ALT_SCREEN = 8
+    CURSOR_UP = 9
+    CURSOR_DOWN = 10
+    CURSOR_FORWARD = 11
+    CURSOR_BACKWARD = 12
+    CURSOR_MOVE_TO_COLUMN = 13
+    CURSOR_MOVE_TO = 14
+    ERASE_IN_LINE = 15
+    SET_WINDOW_TITLE = 16
+
+
+ControlCode = Union[
+    Tuple[ControlType],
+    Tuple[ControlType, Union[int, str]],
+    Tuple[ControlType, int, int],
+]
+
+
+@rich_repr()
+class Segment(NamedTuple):
+    """A piece of text with associated style. Segments are produced by the Console render process and
+    are ultimately converted in to strings to be written to the terminal.
+
+    Args:
+        text (str): A piece of text.
+        style (:class:`~rich.style.Style`, optional): An optional style to apply to the text.
+        control (Tuple[ControlCode], optional): Optional sequence of control codes.
+
+    Attributes:
+        cell_length (int): The cell length of this Segment.
+    """
+
+    text: str
+    style: Optional[Style] = None
+    control: Optional[Sequence[ControlCode]] = None
+
+    @property
+    def cell_length(self) -> int:
+        """The number of terminal cells required to display self.text.
+
+        Returns:
+            int: A number of cells.
+        """
+        text, _style, control = self
+        return 0 if control else cell_len(text)
+
+    def __rich_repr__(self) -> Result:
+        yield self.text
+        if self.control is None:
+            if self.style is not None:
+                yield self.style
+        else:
+            yield self.style
+            yield self.control
+
+    def __bool__(self) -> bool:
+        """Check if the segment contains text."""
+        return bool(self.text)
+
+    @property
+    def is_control(self) -> bool:
+        """Check if the segment contains control codes."""
+        return self.control is not None
+
+    @classmethod
+    @lru_cache(1024 * 16)
+    def _split_cells(cls, segment: "Segment", cut: int) -> Tuple["Segment", "Segment"]:
+        """Split a segment in to two at a given cell position.
+
+        Note that splitting a double-width character, may result in that character turning
+        into two spaces.
+
+        Args:
+            segment (Segment): A segment to split.
+            cut (int): A cell position to cut on.
+
+        Returns:
+            A tuple of two segments.
+        """
+        text, style, control = segment
+        _Segment = Segment
+        cell_length = segment.cell_length
+        if cut >= cell_length:
+            return segment, _Segment("", style, control)
+
+        cell_size = get_character_cell_size
+
+        pos = int((cut / cell_length) * len(text))
+
+        while True:
+            before = text[:pos]
+            cell_pos = cell_len(before)
+            out_by = cell_pos - cut
+            if not out_by:
+                return (
+                    _Segment(before, style, control),
+                    _Segment(text[pos:], style, control),
+                )
+            if out_by == -1 and cell_size(text[pos]) == 2:
+                return (
+                    _Segment(text[:pos] + " ", style, control),
+                    _Segment(" " + text[pos + 1 :], style, control),
+                )
+            if out_by == +1 and cell_size(text[pos - 1]) == 2:
+                return (
+                    _Segment(text[: pos - 1] + " ", style, control),
+                    _Segment(" " + text[pos:], style, control),
+                )
+            if cell_pos < cut:
+                pos += 1
+            else:
+                pos -= 1
+
+    def split_cells(self, cut: int) -> Tuple["Segment", "Segment"]:
+        """Split segment in to two segments at the specified column.
+
+        If the cut point falls in the middle of a 2-cell wide character then it is replaced
+        by two spaces, to preserve the display width of the parent segment.
+
+        Args:
+            cut (int): Offset within the segment to cut.
+
+        Returns:
+            Tuple[Segment, Segment]: Two segments.
+        """
+        text, style, control = self
+        assert cut >= 0
+
+        if _is_single_cell_widths(text):
+            # Fast path with all 1 cell characters
+            if cut >= len(text):
+                return self, Segment("", style, control)
+            return (
+                Segment(text[:cut], style, control),
+                Segment(text[cut:], style, control),
+            )
+
+        return self._split_cells(self, cut)
+
+    @classmethod
+    def line(cls) -> "Segment":
+        """Make a new line segment."""
+        return cls("\n")
+
+    @classmethod
+    def apply_style(
+        cls,
+        segments: Iterable["Segment"],
+        style: Optional[Style] = None,
+        post_style: Optional[Style] = None,
+    ) -> Iterable["Segment"]:
+        """Apply style(s) to an iterable of segments.
+
+        Returns an iterable of segments where the style is replaced by ``style + segment.style + post_style``.
+
+        Args:
+            segments (Iterable[Segment]): Segments to process.
+            style (Style, optional): Base style. Defaults to None.
+            post_style (Style, optional): Style to apply on top of segment style. Defaults to None.
+
+        Returns:
+            Iterable[Segments]: A new iterable of segments (possibly the same iterable).
+        """
+        result_segments = segments
+        if style:
+            apply = style.__add__
+            result_segments = (
+                cls(text, None if control else apply(_style), control)
+                for text, _style, control in result_segments
+            )
+        if post_style:
+            result_segments = (
+                cls(
+                    text,
+                    (
+                        None
+                        if control
+                        else (_style + post_style if _style else post_style)
+                    ),
+                    control,
+                )
+                for text, _style, control in result_segments
+            )
+        return result_segments
+
+    @classmethod
+    def filter_control(
+        cls, segments: Iterable["Segment"], is_control: bool = False
+    ) -> Iterable["Segment"]:
+        """Filter segments by ``is_control`` attribute.
+
+        Args:
+            segments (Iterable[Segment]): An iterable of Segment instances.
+            is_control (bool, optional): is_control flag to match in search.
+
+        Returns:
+            Iterable[Segment]: And iterable of Segment instances.
+
+        """
+        if is_control:
+            return filter(attrgetter("control"), segments)
+        else:
+            return filterfalse(attrgetter("control"), segments)
+
+    @classmethod
+    def split_lines(cls, segments: Iterable["Segment"]) -> Iterable[List["Segment"]]:
+        """Split a sequence of segments in to a list of lines.
+
+        Args:
+            segments (Iterable[Segment]): Segments potentially containing line feeds.
+
+        Yields:
+            Iterable[List[Segment]]: Iterable of segment lists, one per line.
+        """
+        line: List[Segment] = []
+        append = line.append
+
+        for segment in segments:
+            if "\n" in segment.text and not segment.control:
+                text, style, _ = segment
+                while text:
+                    _text, new_line, text = text.partition("\n")
+                    if _text:
+                        append(cls(_text, style))
+                    if new_line:
+                        yield line
+                        line = []
+                        append = line.append
+            else:
+                append(segment)
+        if line:
+            yield line
+
+    @classmethod
+    def split_and_crop_lines(
+        cls,
+        segments: Iterable["Segment"],
+        length: int,
+        style: Optional[Style] = None,
+        pad: bool = True,
+        include_new_lines: bool = True,
+    ) -> Iterable[List["Segment"]]:
+        """Split segments in to lines, and crop lines greater than a given length.
+
+        Args:
+            segments (Iterable[Segment]): An iterable of segments, probably
+                generated from console.render.
+            length (int): Desired line length.
+            style (Style, optional): Style to use for any padding.
+            pad (bool): Enable padding of lines that are less than `length`.
+
+        Returns:
+            Iterable[List[Segment]]: An iterable of lines of segments.
+        """
+        line: List[Segment] = []
+        append = line.append
+
+        adjust_line_length = cls.adjust_line_length
+        new_line_segment = cls("\n")
+
+        for segment in segments:
+            if "\n" in segment.text and not segment.control:
+                text, segment_style, _ = segment
+                while text:
+                    _text, new_line, text = text.partition("\n")
+                    if _text:
+                        append(cls(_text, segment_style))
+                    if new_line:
+                        cropped_line = adjust_line_length(
+                            line, length, style=style, pad=pad
+                        )
+                        if include_new_lines:
+                            cropped_line.append(new_line_segment)
+                        yield cropped_line
+                        line.clear()
+            else:
+                append(segment)
+        if line:
+            yield adjust_line_length(line, length, style=style, pad=pad)
+
+    @classmethod
+    def adjust_line_length(
+        cls,
+        line: List["Segment"],
+        length: int,
+        style: Optional[Style] = None,
+        pad: bool = True,
+    ) -> List["Segment"]:
+        """Adjust a line to a given width (cropping or padding as required).
+
+        Args:
+            segments (Iterable[Segment]): A list of segments in a single line.
+            length (int): The desired width of the line.
+            style (Style, optional): The style of padding if used (space on the end). Defaults to None.
+            pad (bool, optional): Pad lines with spaces if they are shorter than `length`. Defaults to True.
+
+        Returns:
+            List[Segment]: A line of segments with the desired length.
+        """
+        line_length = sum(segment.cell_length for segment in line)
+        new_line: List[Segment]
+
+        if line_length < length:
+            if pad:
+                new_line = line + [cls(" " * (length - line_length), style)]
+            else:
+                new_line = line[:]
+        elif line_length > length:
+            new_line = []
+            append = new_line.append
+            line_length = 0
+            for segment in line:
+                segment_length = segment.cell_length
+                if line_length + segment_length < length or segment.control:
+                    append(segment)
+                    line_length += segment_length
+                else:
+                    text, segment_style, _ = segment
+                    text = set_cell_size(text, length - line_length)
+                    append(cls(text, segment_style))
+                    break
+        else:
+            new_line = line[:]
+        return new_line
+
+    @classmethod
+    def get_line_length(cls, line: List["Segment"]) -> int:
+        """Get the length of list of segments.
+
+        Args:
+            line (List[Segment]): A line encoded as a list of Segments (assumes no '\\\\n' characters),
+
+        Returns:
+            int: The length of the line.
+        """
+        _cell_len = cell_len
+        return sum(_cell_len(text) for text, style, control in line if not control)
+
+    @classmethod
+    def get_shape(cls, lines: List[List["Segment"]]) -> Tuple[int, int]:
+        """Get the shape (enclosing rectangle) of a list of lines.
+
+        Args:
+            lines (List[List[Segment]]): A list of lines (no '\\\\n' characters).
+
+        Returns:
+            Tuple[int, int]: Width and height in characters.
+        """
+        get_line_length = cls.get_line_length
+        max_width = max(get_line_length(line) for line in lines) if lines else 0
+        return (max_width, len(lines))
+
+    @classmethod
+    def set_shape(
+        cls,
+        lines: List[List["Segment"]],
+        width: int,
+        height: Optional[int] = None,
+        style: Optional[Style] = None,
+        new_lines: bool = False,
+    ) -> List[List["Segment"]]:
+        """Set the shape of a list of lines (enclosing rectangle).
+
+        Args:
+            lines (List[List[Segment]]): A list of lines.
+            width (int): Desired width.
+            height (int, optional): Desired height or None for no change.
+            style (Style, optional): Style of any padding added.
+            new_lines (bool, optional): Padded lines should include "\n". Defaults to False.
+
+        Returns:
+            List[List[Segment]]: New list of lines.
+        """
+        _height = height or len(lines)
+
+        blank = (
+            [cls(" " * width + "\n", style)] if new_lines else [cls(" " * width, style)]
+        )
+
+        adjust_line_length = cls.adjust_line_length
+        shaped_lines = lines[:_height]
+        shaped_lines[:] = [
+            adjust_line_length(line, width, style=style) for line in lines
+        ]
+        if len(shaped_lines) < _height:
+            shaped_lines.extend([blank] * (_height - len(shaped_lines)))
+        return shaped_lines
+
+    @classmethod
+    def align_top(
+        cls: Type["Segment"],
+        lines: List[List["Segment"]],
+        width: int,
+        height: int,
+        style: Style,
+        new_lines: bool = False,
+    ) -> List[List["Segment"]]:
+        """Aligns lines to top (adds extra lines to bottom as required).
+
+        Args:
+            lines (List[List[Segment]]): A list of lines.
+            width (int): Desired width.
+            height (int, optional): Desired height or None for no change.
+            style (Style): Style of any padding added.
+            new_lines (bool, optional): Padded lines should include "\n". Defaults to False.
+
+        Returns:
+            List[List[Segment]]: New list of lines.
+        """
+        extra_lines = height - len(lines)
+        if not extra_lines:
+            return lines[:]
+        lines = lines[:height]
+        blank = cls(" " * width + "\n", style) if new_lines else cls(" " * width, style)
+        lines = lines + [[blank]] * extra_lines
+        return lines
+
+    @classmethod
+    def align_bottom(
+        cls: Type["Segment"],
+        lines: List[List["Segment"]],
+        width: int,
+        height: int,
+        style: Style,
+        new_lines: bool = False,
+    ) -> List[List["Segment"]]:
+        """Aligns render to bottom (adds extra lines above as required).
+
+        Args:
+            lines (List[List[Segment]]): A list of lines.
+            width (int): Desired width.
+            height (int, optional): Desired height or None for no change.
+            style (Style): Style of any padding added. Defaults to None.
+            new_lines (bool, optional): Padded lines should include "\n". Defaults to False.
+
+        Returns:
+            List[List[Segment]]: New list of lines.
+        """
+        extra_lines = height - len(lines)
+        if not extra_lines:
+            return lines[:]
+        lines = lines[:height]
+        blank = cls(" " * width + "\n", style) if new_lines else cls(" " * width, style)
+        lines = [[blank]] * extra_lines + lines
+        return lines
+
+    @classmethod
+    def align_middle(
+        cls: Type["Segment"],
+        lines: List[List["Segment"]],
+        width: int,
+        height: int,
+        style: Style,
+        new_lines: bool = False,
+    ) -> List[List["Segment"]]:
+        """Aligns lines to middle (adds extra lines to above and below as required).
+
+        Args:
+            lines (List[List[Segment]]): A list of lines.
+            width (int): Desired width.
+            height (int, optional): Desired height or None for no change.
+            style (Style): Style of any padding added.
+            new_lines (bool, optional): Padded lines should include "\n". Defaults to False.
+
+        Returns:
+            List[List[Segment]]: New list of lines.
+        """
+        extra_lines = height - len(lines)
+        if not extra_lines:
+            return lines[:]
+        lines = lines[:height]
+        blank = cls(" " * width + "\n", style) if new_lines else cls(" " * width, style)
+        top_lines = extra_lines // 2
+        bottom_lines = extra_lines - top_lines
+        lines = [[blank]] * top_lines + lines + [[blank]] * bottom_lines
+        return lines
+
+    @classmethod
+    def simplify(cls, segments: Iterable["Segment"]) -> Iterable["Segment"]:
+        """Simplify an iterable of segments by combining contiguous segments with the same style.
+
+        Args:
+            segments (Iterable[Segment]): An iterable of segments.
+
+        Returns:
+            Iterable[Segment]: A possibly smaller iterable of segments that will render the same way.
+        """
+        iter_segments = iter(segments)
+        try:
+            last_segment = next(iter_segments)
+        except StopIteration:
+            return
+
+        _Segment = Segment
+        for segment in iter_segments:
+            if last_segment.style == segment.style and not segment.control:
+                last_segment = _Segment(
+                    last_segment.text + segment.text, last_segment.style
+                )
+            else:
+                yield last_segment
+                last_segment = segment
+        yield last_segment
+
+    @classmethod
+    def strip_links(cls, segments: Iterable["Segment"]) -> Iterable["Segment"]:
+        """Remove all links from an iterable of styles.
+
+        Args:
+            segments (Iterable[Segment]): An iterable segments.
+
+        Yields:
+            Segment: Segments with link removed.
+        """
+        for segment in segments:
+            if segment.control or segment.style is None:
+                yield segment
+            else:
+                text, style, _control = segment
+                yield cls(text, style.update_link(None) if style else None)
+
+    @classmethod
+    def strip_styles(cls, segments: Iterable["Segment"]) -> Iterable["Segment"]:
+        """Remove all styles from an iterable of segments.
+
+        Args:
+            segments (Iterable[Segment]): An iterable segments.
+
+        Yields:
+            Segment: Segments with styles replace with None
+        """
+        for text, _style, control in segments:
+            yield cls(text, None, control)
+
+    @classmethod
+    def remove_color(cls, segments: Iterable["Segment"]) -> Iterable["Segment"]:
+        """Remove all color from an iterable of segments.
+
+        Args:
+            segments (Iterable[Segment]): An iterable segments.
+
+        Yields:
+            Segment: Segments with colorless style.
+        """
+
+        cache: Dict[Style, Style] = {}
+        for text, style, control in segments:
+            if style:
+                colorless_style = cache.get(style)
+                if colorless_style is None:
+                    colorless_style = style.without_color
+                    cache[style] = colorless_style
+                yield cls(text, colorless_style, control)
+            else:
+                yield cls(text, None, control)
+
+    @classmethod
+    def divide(
+        cls, segments: Iterable["Segment"], cuts: Iterable[int]
+    ) -> Iterable[List["Segment"]]:
+        """Divides an iterable of segments in to portions.
+
+        Args:
+            cuts (Iterable[int]): Cell positions where to divide.
+
+        Yields:
+            [Iterable[List[Segment]]]: An iterable of Segments in List.
+        """
+        split_segments: List["Segment"] = []
+        add_segment = split_segments.append
+
+        iter_cuts = iter(cuts)
+
+        while True:
+            cut = next(iter_cuts, -1)
+            if cut == -1:
+                return
+            if cut != 0:
+                break
+            yield []
+        pos = 0
+
+        segments_clear = split_segments.clear
+        segments_copy = split_segments.copy
+
+        _cell_len = cached_cell_len
+        for segment in segments:
+            text, _style, control = segment
+            while text:
+                end_pos = pos if control else pos + _cell_len(text)
+                if end_pos < cut:
+                    add_segment(segment)
+                    pos = end_pos
+                    break
+
+                if end_pos == cut:
+                    add_segment(segment)
+                    yield segments_copy()
+                    segments_clear()
+                    pos = end_pos
+
+                    cut = next(iter_cuts, -1)
+                    if cut == -1:
+                        if split_segments:
+                            yield segments_copy()
+                        return
+
+                    break
+
+                else:
+                    before, segment = segment.split_cells(cut - pos)
+                    text, _style, control = segment
+                    add_segment(before)
+                    yield segments_copy()
+                    segments_clear()
+                    pos = cut
+
+                cut = next(iter_cuts, -1)
+                if cut == -1:
+                    if split_segments:
+                        yield segments_copy()
+                    return
+
+        yield segments_copy()
+
+
+class Segments:
+    """A simple renderable to render an iterable of segments. This class may be useful if
+    you want to print segments outside of a __rich_console__ method.
+
+    Args:
+        segments (Iterable[Segment]): An iterable of segments.
+        new_lines (bool, optional): Add new lines between segments. Defaults to False.
+    """
+
+    def __init__(self, segments: Iterable[Segment], new_lines: bool = False) -> None:
+        self.segments = list(segments)
+        self.new_lines = new_lines
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        if self.new_lines:
+            line = Segment.line()
+            for segment in self.segments:
+                yield segment
+                yield line
+        else:
+            yield from self.segments
+
+
+class SegmentLines:
+    def __init__(self, lines: Iterable[List[Segment]], new_lines: bool = False) -> None:
+        """A simple renderable containing a number of lines of segments. May be used as an intermediate
+        in rendering process.
+
+        Args:
+            lines (Iterable[List[Segment]]): Lists of segments forming lines.
+            new_lines (bool, optional): Insert new lines after each line. Defaults to False.
+        """
+        self.lines = list(lines)
+        self.new_lines = new_lines
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        if self.new_lines:
+            new_line = Segment.line()
+            for line in self.lines:
+                yield from line
+                yield new_line
+        else:
+            for line in self.lines:
+                yield from line
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console
+    from pip._vendor.rich.syntax import Syntax
+    from pip._vendor.rich.text import Text
+
+    code = """from rich.console import Console
+console = Console()
+text = Text.from_markup("Hello, [bold magenta]World[/]!")
+console.print(text)"""
+
+    text = Text.from_markup("Hello, [bold magenta]World[/]!")
+
+    console = Console()
+
+    console.rule("rich.Segment")
+    console.print(
+        "A Segment is the last step in the Rich render process before generating text with ANSI codes."
+    )
+    console.print("\nConsider the following code:\n")
+    console.print(Syntax(code, "python", line_numbers=True))
+    console.print()
+    console.print(
+        "When you call [b]print()[/b], Rich [i]renders[/i] the object in to the following:\n"
+    )
+    fragments = list(console.render(text))
+    console.print(fragments)
+    console.print()
+    console.print("The Segments are then processed to produce the following output:\n")
+    console.print(text)
+    console.print(
+        "\nYou will only need to know this if you are implementing your own Rich renderables."
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/spinner.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/spinner.py
new file mode 100644
index 00000000..70570b6b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/spinner.py
@@ -0,0 +1,138 @@
+from typing import cast, List, Optional, TYPE_CHECKING, Union
+
+from ._spinners import SPINNERS
+from .measure import Measurement
+from .table import Table
+from .text import Text
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderResult, RenderableType
+    from .style import StyleType
+
+
+class Spinner:
+    """A spinner animation.
+
+    Args:
+        name (str): Name of spinner (run python -m rich.spinner).
+        text (RenderableType, optional): A renderable to display at the right of the spinner (str or Text typically). Defaults to "".
+        style (StyleType, optional): Style for spinner animation. Defaults to None.
+        speed (float, optional): Speed factor for animation. Defaults to 1.0.
+
+    Raises:
+        KeyError: If name isn't one of the supported spinner animations.
+    """
+
+    def __init__(
+        self,
+        name: str,
+        text: "RenderableType" = "",
+        *,
+        style: Optional["StyleType"] = None,
+        speed: float = 1.0,
+    ) -> None:
+        try:
+            spinner = SPINNERS[name]
+        except KeyError:
+            raise KeyError(f"no spinner called {name!r}")
+        self.text: "Union[RenderableType, Text]" = (
+            Text.from_markup(text) if isinstance(text, str) else text
+        )
+        self.name = name
+        self.frames = cast(List[str], spinner["frames"])[:]
+        self.interval = cast(float, spinner["interval"])
+        self.start_time: Optional[float] = None
+        self.style = style
+        self.speed = speed
+        self.frame_no_offset: float = 0.0
+        self._update_speed = 0.0
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        yield self.render(console.get_time())
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        text = self.render(0)
+        return Measurement.get(console, options, text)
+
+    def render(self, time: float) -> "RenderableType":
+        """Render the spinner for a given time.
+
+        Args:
+            time (float): Time in seconds.
+
+        Returns:
+            RenderableType: A renderable containing animation frame.
+        """
+        if self.start_time is None:
+            self.start_time = time
+
+        frame_no = ((time - self.start_time) * self.speed) / (
+            self.interval / 1000.0
+        ) + self.frame_no_offset
+        frame = Text(
+            self.frames[int(frame_no) % len(self.frames)], style=self.style or ""
+        )
+
+        if self._update_speed:
+            self.frame_no_offset = frame_no
+            self.start_time = time
+            self.speed = self._update_speed
+            self._update_speed = 0.0
+
+        if not self.text:
+            return frame
+        elif isinstance(self.text, (str, Text)):
+            return Text.assemble(frame, " ", self.text)
+        else:
+            table = Table.grid(padding=1)
+            table.add_row(frame, self.text)
+            return table
+
+    def update(
+        self,
+        *,
+        text: "RenderableType" = "",
+        style: Optional["StyleType"] = None,
+        speed: Optional[float] = None,
+    ) -> None:
+        """Updates attributes of a spinner after it has been started.
+
+        Args:
+            text (RenderableType, optional): A renderable to display at the right of the spinner (str or Text typically). Defaults to "".
+            style (StyleType, optional): Style for spinner animation. Defaults to None.
+            speed (float, optional): Speed factor for animation. Defaults to None.
+        """
+        if text:
+            self.text = Text.from_markup(text) if isinstance(text, str) else text
+        if style:
+            self.style = style
+        if speed:
+            self._update_speed = speed
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from time import sleep
+
+    from .columns import Columns
+    from .panel import Panel
+    from .live import Live
+
+    all_spinners = Columns(
+        [
+            Spinner(spinner_name, text=Text(repr(spinner_name), style="green"))
+            for spinner_name in sorted(SPINNERS.keys())
+        ],
+        column_first=True,
+        expand=True,
+    )
+
+    with Live(
+        Panel(all_spinners, title="Spinners", border_style="blue"),
+        refresh_per_second=20,
+    ) as live:
+        while True:
+            sleep(0.1)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/status.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/status.py
new file mode 100644
index 00000000..65744838
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/status.py
@@ -0,0 +1,131 @@
+from types import TracebackType
+from typing import Optional, Type
+
+from .console import Console, RenderableType
+from .jupyter import JupyterMixin
+from .live import Live
+from .spinner import Spinner
+from .style import StyleType
+
+
+class Status(JupyterMixin):
+    """Displays a status indicator with a 'spinner' animation.
+
+    Args:
+        status (RenderableType): A status renderable (str or Text typically).
+        console (Console, optional): Console instance to use, or None for global console. Defaults to None.
+        spinner (str, optional): Name of spinner animation (see python -m rich.spinner). Defaults to "dots".
+        spinner_style (StyleType, optional): Style of spinner. Defaults to "status.spinner".
+        speed (float, optional): Speed factor for spinner animation. Defaults to 1.0.
+        refresh_per_second (float, optional): Number of refreshes per second. Defaults to 12.5.
+    """
+
+    def __init__(
+        self,
+        status: RenderableType,
+        *,
+        console: Optional[Console] = None,
+        spinner: str = "dots",
+        spinner_style: StyleType = "status.spinner",
+        speed: float = 1.0,
+        refresh_per_second: float = 12.5,
+    ):
+        self.status = status
+        self.spinner_style = spinner_style
+        self.speed = speed
+        self._spinner = Spinner(spinner, text=status, style=spinner_style, speed=speed)
+        self._live = Live(
+            self.renderable,
+            console=console,
+            refresh_per_second=refresh_per_second,
+            transient=True,
+        )
+
+    @property
+    def renderable(self) -> Spinner:
+        return self._spinner
+
+    @property
+    def console(self) -> "Console":
+        """Get the Console used by the Status objects."""
+        return self._live.console
+
+    def update(
+        self,
+        status: Optional[RenderableType] = None,
+        *,
+        spinner: Optional[str] = None,
+        spinner_style: Optional[StyleType] = None,
+        speed: Optional[float] = None,
+    ) -> None:
+        """Update status.
+
+        Args:
+            status (Optional[RenderableType], optional): New status renderable or None for no change. Defaults to None.
+            spinner (Optional[str], optional): New spinner or None for no change. Defaults to None.
+            spinner_style (Optional[StyleType], optional): New spinner style or None for no change. Defaults to None.
+            speed (Optional[float], optional): Speed factor for spinner animation or None for no change. Defaults to None.
+        """
+        if status is not None:
+            self.status = status
+        if spinner_style is not None:
+            self.spinner_style = spinner_style
+        if speed is not None:
+            self.speed = speed
+        if spinner is not None:
+            self._spinner = Spinner(
+                spinner, text=self.status, style=self.spinner_style, speed=self.speed
+            )
+            self._live.update(self.renderable, refresh=True)
+        else:
+            self._spinner.update(
+                text=self.status, style=self.spinner_style, speed=self.speed
+            )
+
+    def start(self) -> None:
+        """Start the status animation."""
+        self._live.start()
+
+    def stop(self) -> None:
+        """Stop the spinner animation."""
+        self._live.stop()
+
+    def __rich__(self) -> RenderableType:
+        return self.renderable
+
+    def __enter__(self) -> "Status":
+        self.start()
+        return self
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> None:
+        self.stop()
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from time import sleep
+
+    from .console import Console
+
+    console = Console()
+    with console.status("[magenta]Covid detector booting up") as status:
+        sleep(3)
+        console.log("Importing advanced AI")
+        sleep(3)
+        console.log("Advanced Covid AI Ready")
+        sleep(3)
+        status.update(status="[bold blue] Scanning for Covid", spinner="earth")
+        sleep(3)
+        console.log("Found 10,000,000,000 copies of Covid32.exe")
+        sleep(3)
+        status.update(
+            status="[bold red]Moving Covid32.exe to Trash",
+            spinner="bouncingBall",
+            spinner_style="yellow",
+        )
+        sleep(5)
+    console.print("[bold green]Covid deleted successfully")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/style.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/style.py
new file mode 100644
index 00000000..262fd6ec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/style.py
@@ -0,0 +1,796 @@
+import sys
+from functools import lru_cache
+from marshal import dumps, loads
+from random import randint
+from typing import Any, Dict, Iterable, List, Optional, Type, Union, cast
+
+from . import errors
+from .color import Color, ColorParseError, ColorSystem, blend_rgb
+from .repr import Result, rich_repr
+from .terminal_theme import DEFAULT_TERMINAL_THEME, TerminalTheme
+
+# Style instances and style definitions are often interchangeable
+StyleType = Union[str, "Style"]
+
+
+class _Bit:
+    """A descriptor to get/set a style attribute bit."""
+
+    __slots__ = ["bit"]
+
+    def __init__(self, bit_no: int) -> None:
+        self.bit = 1 << bit_no
+
+    def __get__(self, obj: "Style", objtype: Type["Style"]) -> Optional[bool]:
+        if obj._set_attributes & self.bit:
+            return obj._attributes & self.bit != 0
+        return None
+
+
+@rich_repr
+class Style:
+    """A terminal style.
+
+    A terminal style consists of a color (`color`), a background color (`bgcolor`), and a number of attributes, such
+    as bold, italic etc. The attributes have 3 states: they can either be on
+    (``True``), off (``False``), or not set (``None``).
+
+    Args:
+        color (Union[Color, str], optional): Color of terminal text. Defaults to None.
+        bgcolor (Union[Color, str], optional): Color of terminal background. Defaults to None.
+        bold (bool, optional): Enable bold text. Defaults to None.
+        dim (bool, optional): Enable dim text. Defaults to None.
+        italic (bool, optional): Enable italic text. Defaults to None.
+        underline (bool, optional): Enable underlined text. Defaults to None.
+        blink (bool, optional): Enabled blinking text. Defaults to None.
+        blink2 (bool, optional): Enable fast blinking text. Defaults to None.
+        reverse (bool, optional): Enabled reverse text. Defaults to None.
+        conceal (bool, optional): Enable concealed text. Defaults to None.
+        strike (bool, optional): Enable strikethrough text. Defaults to None.
+        underline2 (bool, optional): Enable doubly underlined text. Defaults to None.
+        frame (bool, optional): Enable framed text. Defaults to None.
+        encircle (bool, optional): Enable encircled text. Defaults to None.
+        overline (bool, optional): Enable overlined text. Defaults to None.
+        link (str, link): Link URL. Defaults to None.
+
+    """
+
+    _color: Optional[Color]
+    _bgcolor: Optional[Color]
+    _attributes: int
+    _set_attributes: int
+    _hash: Optional[int]
+    _null: bool
+    _meta: Optional[bytes]
+
+    __slots__ = [
+        "_color",
+        "_bgcolor",
+        "_attributes",
+        "_set_attributes",
+        "_link",
+        "_link_id",
+        "_ansi",
+        "_style_definition",
+        "_hash",
+        "_null",
+        "_meta",
+    ]
+
+    # maps bits on to SGR parameter
+    _style_map = {
+        0: "1",
+        1: "2",
+        2: "3",
+        3: "4",
+        4: "5",
+        5: "6",
+        6: "7",
+        7: "8",
+        8: "9",
+        9: "21",
+        10: "51",
+        11: "52",
+        12: "53",
+    }
+
+    STYLE_ATTRIBUTES = {
+        "dim": "dim",
+        "d": "dim",
+        "bold": "bold",
+        "b": "bold",
+        "italic": "italic",
+        "i": "italic",
+        "underline": "underline",
+        "u": "underline",
+        "blink": "blink",
+        "blink2": "blink2",
+        "reverse": "reverse",
+        "r": "reverse",
+        "conceal": "conceal",
+        "c": "conceal",
+        "strike": "strike",
+        "s": "strike",
+        "underline2": "underline2",
+        "uu": "underline2",
+        "frame": "frame",
+        "encircle": "encircle",
+        "overline": "overline",
+        "o": "overline",
+    }
+
+    def __init__(
+        self,
+        *,
+        color: Optional[Union[Color, str]] = None,
+        bgcolor: Optional[Union[Color, str]] = None,
+        bold: Optional[bool] = None,
+        dim: Optional[bool] = None,
+        italic: Optional[bool] = None,
+        underline: Optional[bool] = None,
+        blink: Optional[bool] = None,
+        blink2: Optional[bool] = None,
+        reverse: Optional[bool] = None,
+        conceal: Optional[bool] = None,
+        strike: Optional[bool] = None,
+        underline2: Optional[bool] = None,
+        frame: Optional[bool] = None,
+        encircle: Optional[bool] = None,
+        overline: Optional[bool] = None,
+        link: Optional[str] = None,
+        meta: Optional[Dict[str, Any]] = None,
+    ):
+        self._ansi: Optional[str] = None
+        self._style_definition: Optional[str] = None
+
+        def _make_color(color: Union[Color, str]) -> Color:
+            return color if isinstance(color, Color) else Color.parse(color)
+
+        self._color = None if color is None else _make_color(color)
+        self._bgcolor = None if bgcolor is None else _make_color(bgcolor)
+        self._set_attributes = sum(
+            (
+                bold is not None,
+                dim is not None and 2,
+                italic is not None and 4,
+                underline is not None and 8,
+                blink is not None and 16,
+                blink2 is not None and 32,
+                reverse is not None and 64,
+                conceal is not None and 128,
+                strike is not None and 256,
+                underline2 is not None and 512,
+                frame is not None and 1024,
+                encircle is not None and 2048,
+                overline is not None and 4096,
+            )
+        )
+        self._attributes = (
+            sum(
+                (
+                    bold and 1 or 0,
+                    dim and 2 or 0,
+                    italic and 4 or 0,
+                    underline and 8 or 0,
+                    blink and 16 or 0,
+                    blink2 and 32 or 0,
+                    reverse and 64 or 0,
+                    conceal and 128 or 0,
+                    strike and 256 or 0,
+                    underline2 and 512 or 0,
+                    frame and 1024 or 0,
+                    encircle and 2048 or 0,
+                    overline and 4096 or 0,
+                )
+            )
+            if self._set_attributes
+            else 0
+        )
+
+        self._link = link
+        self._meta = None if meta is None else dumps(meta)
+        self._link_id = (
+            f"{randint(0, 999999)}{hash(self._meta)}" if (link or meta) else ""
+        )
+        self._hash: Optional[int] = None
+        self._null = not (self._set_attributes or color or bgcolor or link or meta)
+
+    @classmethod
+    def null(cls) -> "Style":
+        """Create an 'null' style, equivalent to Style(), but more performant."""
+        return NULL_STYLE
+
+    @classmethod
+    def from_color(
+        cls, color: Optional[Color] = None, bgcolor: Optional[Color] = None
+    ) -> "Style":
+        """Create a new style with colors and no attributes.
+
+        Returns:
+            color (Optional[Color]): A (foreground) color, or None for no color. Defaults to None.
+            bgcolor (Optional[Color]): A (background) color, or None for no color. Defaults to None.
+        """
+        style: Style = cls.__new__(Style)
+        style._ansi = None
+        style._style_definition = None
+        style._color = color
+        style._bgcolor = bgcolor
+        style._set_attributes = 0
+        style._attributes = 0
+        style._link = None
+        style._link_id = ""
+        style._meta = None
+        style._null = not (color or bgcolor)
+        style._hash = None
+        return style
+
+    @classmethod
+    def from_meta(cls, meta: Optional[Dict[str, Any]]) -> "Style":
+        """Create a new style with meta data.
+
+        Returns:
+            meta (Optional[Dict[str, Any]]): A dictionary of meta data. Defaults to None.
+        """
+        style: Style = cls.__new__(Style)
+        style._ansi = None
+        style._style_definition = None
+        style._color = None
+        style._bgcolor = None
+        style._set_attributes = 0
+        style._attributes = 0
+        style._link = None
+        style._meta = dumps(meta)
+        style._link_id = f"{randint(0, 999999)}{hash(style._meta)}"
+        style._hash = None
+        style._null = not (meta)
+        return style
+
+    @classmethod
+    def on(cls, meta: Optional[Dict[str, Any]] = None, **handlers: Any) -> "Style":
+        """Create a blank style with meta information.
+
+        Example:
+            style = Style.on(click=self.on_click)
+
+        Args:
+            meta (Optional[Dict[str, Any]], optional): An optional dict of meta information.
+            **handlers (Any): Keyword arguments are translated in to handlers.
+
+        Returns:
+            Style: A Style with meta information attached.
+        """
+        meta = {} if meta is None else meta
+        meta.update({f"@{key}": value for key, value in handlers.items()})
+        return cls.from_meta(meta)
+
+    bold = _Bit(0)
+    dim = _Bit(1)
+    italic = _Bit(2)
+    underline = _Bit(3)
+    blink = _Bit(4)
+    blink2 = _Bit(5)
+    reverse = _Bit(6)
+    conceal = _Bit(7)
+    strike = _Bit(8)
+    underline2 = _Bit(9)
+    frame = _Bit(10)
+    encircle = _Bit(11)
+    overline = _Bit(12)
+
+    @property
+    def link_id(self) -> str:
+        """Get a link id, used in ansi code for links."""
+        return self._link_id
+
+    def __str__(self) -> str:
+        """Re-generate style definition from attributes."""
+        if self._style_definition is None:
+            attributes: List[str] = []
+            append = attributes.append
+            bits = self._set_attributes
+            if bits & 0b0000000001111:
+                if bits & 1:
+                    append("bold" if self.bold else "not bold")
+                if bits & (1 << 1):
+                    append("dim" if self.dim else "not dim")
+                if bits & (1 << 2):
+                    append("italic" if self.italic else "not italic")
+                if bits & (1 << 3):
+                    append("underline" if self.underline else "not underline")
+            if bits & 0b0000111110000:
+                if bits & (1 << 4):
+                    append("blink" if self.blink else "not blink")
+                if bits & (1 << 5):
+                    append("blink2" if self.blink2 else "not blink2")
+                if bits & (1 << 6):
+                    append("reverse" if self.reverse else "not reverse")
+                if bits & (1 << 7):
+                    append("conceal" if self.conceal else "not conceal")
+                if bits & (1 << 8):
+                    append("strike" if self.strike else "not strike")
+            if bits & 0b1111000000000:
+                if bits & (1 << 9):
+                    append("underline2" if self.underline2 else "not underline2")
+                if bits & (1 << 10):
+                    append("frame" if self.frame else "not frame")
+                if bits & (1 << 11):
+                    append("encircle" if self.encircle else "not encircle")
+                if bits & (1 << 12):
+                    append("overline" if self.overline else "not overline")
+            if self._color is not None:
+                append(self._color.name)
+            if self._bgcolor is not None:
+                append("on")
+                append(self._bgcolor.name)
+            if self._link:
+                append("link")
+                append(self._link)
+            self._style_definition = " ".join(attributes) or "none"
+        return self._style_definition
+
+    def __bool__(self) -> bool:
+        """A Style is false if it has no attributes, colors, or links."""
+        return not self._null
+
+    def _make_ansi_codes(self, color_system: ColorSystem) -> str:
+        """Generate ANSI codes for this style.
+
+        Args:
+            color_system (ColorSystem): Color system.
+
+        Returns:
+            str: String containing codes.
+        """
+
+        if self._ansi is None:
+            sgr: List[str] = []
+            append = sgr.append
+            _style_map = self._style_map
+            attributes = self._attributes & self._set_attributes
+            if attributes:
+                if attributes & 1:
+                    append(_style_map[0])
+                if attributes & 2:
+                    append(_style_map[1])
+                if attributes & 4:
+                    append(_style_map[2])
+                if attributes & 8:
+                    append(_style_map[3])
+                if attributes & 0b0000111110000:
+                    for bit in range(4, 9):
+                        if attributes & (1 << bit):
+                            append(_style_map[bit])
+                if attributes & 0b1111000000000:
+                    for bit in range(9, 13):
+                        if attributes & (1 << bit):
+                            append(_style_map[bit])
+            if self._color is not None:
+                sgr.extend(self._color.downgrade(color_system).get_ansi_codes())
+            if self._bgcolor is not None:
+                sgr.extend(
+                    self._bgcolor.downgrade(color_system).get_ansi_codes(
+                        foreground=False
+                    )
+                )
+            self._ansi = ";".join(sgr)
+        return self._ansi
+
+    @classmethod
+    @lru_cache(maxsize=1024)
+    def normalize(cls, style: str) -> str:
+        """Normalize a style definition so that styles with the same effect have the same string
+        representation.
+
+        Args:
+            style (str): A style definition.
+
+        Returns:
+            str: Normal form of style definition.
+        """
+        try:
+            return str(cls.parse(style))
+        except errors.StyleSyntaxError:
+            return style.strip().lower()
+
+    @classmethod
+    def pick_first(cls, *values: Optional[StyleType]) -> StyleType:
+        """Pick first non-None style."""
+        for value in values:
+            if value is not None:
+                return value
+        raise ValueError("expected at least one non-None style")
+
+    def __rich_repr__(self) -> Result:
+        yield "color", self.color, None
+        yield "bgcolor", self.bgcolor, None
+        yield "bold", self.bold, None,
+        yield "dim", self.dim, None,
+        yield "italic", self.italic, None
+        yield "underline", self.underline, None,
+        yield "blink", self.blink, None
+        yield "blink2", self.blink2, None
+        yield "reverse", self.reverse, None
+        yield "conceal", self.conceal, None
+        yield "strike", self.strike, None
+        yield "underline2", self.underline2, None
+        yield "frame", self.frame, None
+        yield "encircle", self.encircle, None
+        yield "link", self.link, None
+        if self._meta:
+            yield "meta", self.meta
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Style):
+            return NotImplemented
+        return self.__hash__() == other.__hash__()
+
+    def __ne__(self, other: Any) -> bool:
+        if not isinstance(other, Style):
+            return NotImplemented
+        return self.__hash__() != other.__hash__()
+
+    def __hash__(self) -> int:
+        if self._hash is not None:
+            return self._hash
+        self._hash = hash(
+            (
+                self._color,
+                self._bgcolor,
+                self._attributes,
+                self._set_attributes,
+                self._link,
+                self._meta,
+            )
+        )
+        return self._hash
+
+    @property
+    def color(self) -> Optional[Color]:
+        """The foreground color or None if it is not set."""
+        return self._color
+
+    @property
+    def bgcolor(self) -> Optional[Color]:
+        """The background color or None if it is not set."""
+        return self._bgcolor
+
+    @property
+    def link(self) -> Optional[str]:
+        """Link text, if set."""
+        return self._link
+
+    @property
+    def transparent_background(self) -> bool:
+        """Check if the style specified a transparent background."""
+        return self.bgcolor is None or self.bgcolor.is_default
+
+    @property
+    def background_style(self) -> "Style":
+        """A Style with background only."""
+        return Style(bgcolor=self.bgcolor)
+
+    @property
+    def meta(self) -> Dict[str, Any]:
+        """Get meta information (can not be changed after construction)."""
+        return {} if self._meta is None else cast(Dict[str, Any], loads(self._meta))
+
+    @property
+    def without_color(self) -> "Style":
+        """Get a copy of the style with color removed."""
+        if self._null:
+            return NULL_STYLE
+        style: Style = self.__new__(Style)
+        style._ansi = None
+        style._style_definition = None
+        style._color = None
+        style._bgcolor = None
+        style._attributes = self._attributes
+        style._set_attributes = self._set_attributes
+        style._link = self._link
+        style._link_id = f"{randint(0, 999999)}" if self._link else ""
+        style._null = False
+        style._meta = None
+        style._hash = None
+        return style
+
+    @classmethod
+    @lru_cache(maxsize=4096)
+    def parse(cls, style_definition: str) -> "Style":
+        """Parse a style definition.
+
+        Args:
+            style_definition (str): A string containing a style.
+
+        Raises:
+            errors.StyleSyntaxError: If the style definition syntax is invalid.
+
+        Returns:
+            `Style`: A Style instance.
+        """
+        if style_definition.strip() == "none" or not style_definition:
+            return cls.null()
+
+        STYLE_ATTRIBUTES = cls.STYLE_ATTRIBUTES
+        color: Optional[str] = None
+        bgcolor: Optional[str] = None
+        attributes: Dict[str, Optional[Any]] = {}
+        link: Optional[str] = None
+
+        words = iter(style_definition.split())
+        for original_word in words:
+            word = original_word.lower()
+            if word == "on":
+                word = next(words, "")
+                if not word:
+                    raise errors.StyleSyntaxError("color expected after 'on'")
+                try:
+                    Color.parse(word) is None
+                except ColorParseError as error:
+                    raise errors.StyleSyntaxError(
+                        f"unable to parse {word!r} as background color; {error}"
+                    ) from None
+                bgcolor = word
+
+            elif word == "not":
+                word = next(words, "")
+                attribute = STYLE_ATTRIBUTES.get(word)
+                if attribute is None:
+                    raise errors.StyleSyntaxError(
+                        f"expected style attribute after 'not', found {word!r}"
+                    )
+                attributes[attribute] = False
+
+            elif word == "link":
+                word = next(words, "")
+                if not word:
+                    raise errors.StyleSyntaxError("URL expected after 'link'")
+                link = word
+
+            elif word in STYLE_ATTRIBUTES:
+                attributes[STYLE_ATTRIBUTES[word]] = True
+
+            else:
+                try:
+                    Color.parse(word)
+                except ColorParseError as error:
+                    raise errors.StyleSyntaxError(
+                        f"unable to parse {word!r} as color; {error}"
+                    ) from None
+                color = word
+        style = Style(color=color, bgcolor=bgcolor, link=link, **attributes)
+        return style
+
+    @lru_cache(maxsize=1024)
+    def get_html_style(self, theme: Optional[TerminalTheme] = None) -> str:
+        """Get a CSS style rule."""
+        theme = theme or DEFAULT_TERMINAL_THEME
+        css: List[str] = []
+        append = css.append
+
+        color = self.color
+        bgcolor = self.bgcolor
+        if self.reverse:
+            color, bgcolor = bgcolor, color
+        if self.dim:
+            foreground_color = (
+                theme.foreground_color if color is None else color.get_truecolor(theme)
+            )
+            color = Color.from_triplet(
+                blend_rgb(foreground_color, theme.background_color, 0.5)
+            )
+        if color is not None:
+            theme_color = color.get_truecolor(theme)
+            append(f"color: {theme_color.hex}")
+            append(f"text-decoration-color: {theme_color.hex}")
+        if bgcolor is not None:
+            theme_color = bgcolor.get_truecolor(theme, foreground=False)
+            append(f"background-color: {theme_color.hex}")
+        if self.bold:
+            append("font-weight: bold")
+        if self.italic:
+            append("font-style: italic")
+        if self.underline:
+            append("text-decoration: underline")
+        if self.strike:
+            append("text-decoration: line-through")
+        if self.overline:
+            append("text-decoration: overline")
+        return "; ".join(css)
+
+    @classmethod
+    def combine(cls, styles: Iterable["Style"]) -> "Style":
+        """Combine styles and get result.
+
+        Args:
+            styles (Iterable[Style]): Styles to combine.
+
+        Returns:
+            Style: A new style instance.
+        """
+        iter_styles = iter(styles)
+        return sum(iter_styles, next(iter_styles))
+
+    @classmethod
+    def chain(cls, *styles: "Style") -> "Style":
+        """Combine styles from positional argument in to a single style.
+
+        Args:
+            *styles (Iterable[Style]): Styles to combine.
+
+        Returns:
+            Style: A new style instance.
+        """
+        iter_styles = iter(styles)
+        return sum(iter_styles, next(iter_styles))
+
+    def copy(self) -> "Style":
+        """Get a copy of this style.
+
+        Returns:
+            Style: A new Style instance with identical attributes.
+        """
+        if self._null:
+            return NULL_STYLE
+        style: Style = self.__new__(Style)
+        style._ansi = self._ansi
+        style._style_definition = self._style_definition
+        style._color = self._color
+        style._bgcolor = self._bgcolor
+        style._attributes = self._attributes
+        style._set_attributes = self._set_attributes
+        style._link = self._link
+        style._link_id = f"{randint(0, 999999)}" if self._link else ""
+        style._hash = self._hash
+        style._null = False
+        style._meta = self._meta
+        return style
+
+    @lru_cache(maxsize=128)
+    def clear_meta_and_links(self) -> "Style":
+        """Get a copy of this style with link and meta information removed.
+
+        Returns:
+            Style: New style object.
+        """
+        if self._null:
+            return NULL_STYLE
+        style: Style = self.__new__(Style)
+        style._ansi = self._ansi
+        style._style_definition = self._style_definition
+        style._color = self._color
+        style._bgcolor = self._bgcolor
+        style._attributes = self._attributes
+        style._set_attributes = self._set_attributes
+        style._link = None
+        style._link_id = ""
+        style._hash = None
+        style._null = False
+        style._meta = None
+        return style
+
+    def update_link(self, link: Optional[str] = None) -> "Style":
+        """Get a copy with a different value for link.
+
+        Args:
+            link (str, optional): New value for link. Defaults to None.
+
+        Returns:
+            Style: A new Style instance.
+        """
+        style: Style = self.__new__(Style)
+        style._ansi = self._ansi
+        style._style_definition = self._style_definition
+        style._color = self._color
+        style._bgcolor = self._bgcolor
+        style._attributes = self._attributes
+        style._set_attributes = self._set_attributes
+        style._link = link
+        style._link_id = f"{randint(0, 999999)}" if link else ""
+        style._hash = None
+        style._null = False
+        style._meta = self._meta
+        return style
+
+    def render(
+        self,
+        text: str = "",
+        *,
+        color_system: Optional[ColorSystem] = ColorSystem.TRUECOLOR,
+        legacy_windows: bool = False,
+    ) -> str:
+        """Render the ANSI codes for the style.
+
+        Args:
+            text (str, optional): A string to style. Defaults to "".
+            color_system (Optional[ColorSystem], optional): Color system to render to. Defaults to ColorSystem.TRUECOLOR.
+
+        Returns:
+            str: A string containing ANSI style codes.
+        """
+        if not text or color_system is None:
+            return text
+        attrs = self._ansi or self._make_ansi_codes(color_system)
+        rendered = f"\x1b[{attrs}m{text}\x1b[0m" if attrs else text
+        if self._link and not legacy_windows:
+            rendered = (
+                f"\x1b]8;id={self._link_id};{self._link}\x1b\\{rendered}\x1b]8;;\x1b\\"
+            )
+        return rendered
+
+    def test(self, text: Optional[str] = None) -> None:
+        """Write text with style directly to terminal.
+
+        This method is for testing purposes only.
+
+        Args:
+            text (Optional[str], optional): Text to style or None for style name.
+
+        """
+        text = text or str(self)
+        sys.stdout.write(f"{self.render(text)}\n")
+
+    @lru_cache(maxsize=1024)
+    def _add(self, style: Optional["Style"]) -> "Style":
+        if style is None or style._null:
+            return self
+        if self._null:
+            return style
+        new_style: Style = self.__new__(Style)
+        new_style._ansi = None
+        new_style._style_definition = None
+        new_style._color = style._color or self._color
+        new_style._bgcolor = style._bgcolor or self._bgcolor
+        new_style._attributes = (self._attributes & ~style._set_attributes) | (
+            style._attributes & style._set_attributes
+        )
+        new_style._set_attributes = self._set_attributes | style._set_attributes
+        new_style._link = style._link or self._link
+        new_style._link_id = style._link_id or self._link_id
+        new_style._null = style._null
+        if self._meta and style._meta:
+            new_style._meta = dumps({**self.meta, **style.meta})
+        else:
+            new_style._meta = self._meta or style._meta
+        new_style._hash = None
+        return new_style
+
+    def __add__(self, style: Optional["Style"]) -> "Style":
+        combined_style = self._add(style)
+        return combined_style.copy() if combined_style.link else combined_style
+
+
+NULL_STYLE = Style()
+
+
+class StyleStack:
+    """A stack of styles."""
+
+    __slots__ = ["_stack"]
+
+    def __init__(self, default_style: "Style") -> None:
+        self._stack: List[Style] = [default_style]
+
+    def __repr__(self) -> str:
+        return f"<stylestack {self._stack!r}>"
+
+    @property
+    def current(self) -> Style:
+        """Get the Style at the top of the stack."""
+        return self._stack[-1]
+
+    def push(self, style: Style) -> None:
+        """Push a new style on to the stack.
+
+        Args:
+            style (Style): New style to combine with current style.
+        """
+        self._stack.append(self._stack[-1] + style)
+
+    def pop(self) -> Style:
+        """Pop last style and discard.
+
+        Returns:
+            Style: New current style (also available as stack.current)
+        """
+        self._stack.pop()
+        return self._stack[-1]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/styled.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/styled.py
new file mode 100644
index 00000000..91cd0db3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/styled.py
@@ -0,0 +1,42 @@
+from typing import TYPE_CHECKING
+
+from .measure import Measurement
+from .segment import Segment
+from .style import StyleType
+
+if TYPE_CHECKING:
+    from .console import Console, ConsoleOptions, RenderResult, RenderableType
+
+
+class Styled:
+    """Apply a style to a renderable.
+
+    Args:
+        renderable (RenderableType): Any renderable.
+        style (StyleType): A style to apply across the entire renderable.
+    """
+
+    def __init__(self, renderable: "RenderableType", style: "StyleType") -> None:
+        self.renderable = renderable
+        self.style = style
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        style = console.get_style(self.style)
+        rendered_segments = console.render(self.renderable, options)
+        segments = Segment.apply_style(rendered_segments, style)
+        return segments
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        return Measurement.get(console, options, self.renderable)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich import print
+    from pip._vendor.rich.panel import Panel
+
+    panel = Styled(Panel("hello"), "on blue")
+    print(panel)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/syntax.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/syntax.py
new file mode 100644
index 00000000..f3d483c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/syntax.py
@@ -0,0 +1,966 @@
+import os.path
+import re
+import sys
+import textwrap
+from abc import ABC, abstractmethod
+from pathlib import Path
+from typing import (
+    Any,
+    Dict,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Sequence,
+    Set,
+    Tuple,
+    Type,
+    Union,
+)
+
+from pip._vendor.pygments.lexer import Lexer
+from pip._vendor.pygments.lexers import get_lexer_by_name, guess_lexer_for_filename
+from pip._vendor.pygments.style import Style as PygmentsStyle
+from pip._vendor.pygments.styles import get_style_by_name
+from pip._vendor.pygments.token import (
+    Comment,
+    Error,
+    Generic,
+    Keyword,
+    Name,
+    Number,
+    Operator,
+    String,
+    Token,
+    Whitespace,
+)
+from pip._vendor.pygments.util import ClassNotFound
+
+from pip._vendor.rich.containers import Lines
+from pip._vendor.rich.padding import Padding, PaddingDimensions
+
+from ._loop import loop_first
+from .cells import cell_len
+from .color import Color, blend_rgb
+from .console import Console, ConsoleOptions, JustifyMethod, RenderResult
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment, Segments
+from .style import Style, StyleType
+from .text import Text
+
+TokenType = Tuple[str, ...]
+
+WINDOWS = sys.platform == "win32"
+DEFAULT_THEME = "monokai"
+
+# The following styles are based on https://github.com/pygments/pygments/blob/master/pygments/formatters/terminal.py
+# A few modifications were made
+
+ANSI_LIGHT: Dict[TokenType, Style] = {
+    Token: Style(),
+    Whitespace: Style(color="white"),
+    Comment: Style(dim=True),
+    Comment.Preproc: Style(color="cyan"),
+    Keyword: Style(color="blue"),
+    Keyword.Type: Style(color="cyan"),
+    Operator.Word: Style(color="magenta"),
+    Name.Builtin: Style(color="cyan"),
+    Name.Function: Style(color="green"),
+    Name.Namespace: Style(color="cyan", underline=True),
+    Name.Class: Style(color="green", underline=True),
+    Name.Exception: Style(color="cyan"),
+    Name.Decorator: Style(color="magenta", bold=True),
+    Name.Variable: Style(color="red"),
+    Name.Constant: Style(color="red"),
+    Name.Attribute: Style(color="cyan"),
+    Name.Tag: Style(color="bright_blue"),
+    String: Style(color="yellow"),
+    Number: Style(color="blue"),
+    Generic.Deleted: Style(color="bright_red"),
+    Generic.Inserted: Style(color="green"),
+    Generic.Heading: Style(bold=True),
+    Generic.Subheading: Style(color="magenta", bold=True),
+    Generic.Prompt: Style(bold=True),
+    Generic.Error: Style(color="bright_red"),
+    Error: Style(color="red", underline=True),
+}
+
+ANSI_DARK: Dict[TokenType, Style] = {
+    Token: Style(),
+    Whitespace: Style(color="bright_black"),
+    Comment: Style(dim=True),
+    Comment.Preproc: Style(color="bright_cyan"),
+    Keyword: Style(color="bright_blue"),
+    Keyword.Type: Style(color="bright_cyan"),
+    Operator.Word: Style(color="bright_magenta"),
+    Name.Builtin: Style(color="bright_cyan"),
+    Name.Function: Style(color="bright_green"),
+    Name.Namespace: Style(color="bright_cyan", underline=True),
+    Name.Class: Style(color="bright_green", underline=True),
+    Name.Exception: Style(color="bright_cyan"),
+    Name.Decorator: Style(color="bright_magenta", bold=True),
+    Name.Variable: Style(color="bright_red"),
+    Name.Constant: Style(color="bright_red"),
+    Name.Attribute: Style(color="bright_cyan"),
+    Name.Tag: Style(color="bright_blue"),
+    String: Style(color="yellow"),
+    Number: Style(color="bright_blue"),
+    Generic.Deleted: Style(color="bright_red"),
+    Generic.Inserted: Style(color="bright_green"),
+    Generic.Heading: Style(bold=True),
+    Generic.Subheading: Style(color="bright_magenta", bold=True),
+    Generic.Prompt: Style(bold=True),
+    Generic.Error: Style(color="bright_red"),
+    Error: Style(color="red", underline=True),
+}
+
+RICH_SYNTAX_THEMES = {"ansi_light": ANSI_LIGHT, "ansi_dark": ANSI_DARK}
+NUMBERS_COLUMN_DEFAULT_PADDING = 2
+
+
+class SyntaxTheme(ABC):
+    """Base class for a syntax theme."""
+
+    @abstractmethod
+    def get_style_for_token(self, token_type: TokenType) -> Style:
+        """Get a style for a given Pygments token."""
+        raise NotImplementedError  # pragma: no cover
+
+    @abstractmethod
+    def get_background_style(self) -> Style:
+        """Get the background color."""
+        raise NotImplementedError  # pragma: no cover
+
+
+class PygmentsSyntaxTheme(SyntaxTheme):
+    """Syntax theme that delegates to Pygments theme."""
+
+    def __init__(self, theme: Union[str, Type[PygmentsStyle]]) -> None:
+        self._style_cache: Dict[TokenType, Style] = {}
+        if isinstance(theme, str):
+            try:
+                self._pygments_style_class = get_style_by_name(theme)
+            except ClassNotFound:
+                self._pygments_style_class = get_style_by_name("default")
+        else:
+            self._pygments_style_class = theme
+
+        self._background_color = self._pygments_style_class.background_color
+        self._background_style = Style(bgcolor=self._background_color)
+
+    def get_style_for_token(self, token_type: TokenType) -> Style:
+        """Get a style from a Pygments class."""
+        try:
+            return self._style_cache[token_type]
+        except KeyError:
+            try:
+                pygments_style = self._pygments_style_class.style_for_token(token_type)
+            except KeyError:
+                style = Style.null()
+            else:
+                color = pygments_style["color"]
+                bgcolor = pygments_style["bgcolor"]
+                style = Style(
+                    color="#" + color if color else "#000000",
+                    bgcolor="#" + bgcolor if bgcolor else self._background_color,
+                    bold=pygments_style["bold"],
+                    italic=pygments_style["italic"],
+                    underline=pygments_style["underline"],
+                )
+            self._style_cache[token_type] = style
+        return style
+
+    def get_background_style(self) -> Style:
+        return self._background_style
+
+
+class ANSISyntaxTheme(SyntaxTheme):
+    """Syntax theme to use standard colors."""
+
+    def __init__(self, style_map: Dict[TokenType, Style]) -> None:
+        self.style_map = style_map
+        self._missing_style = Style.null()
+        self._background_style = Style.null()
+        self._style_cache: Dict[TokenType, Style] = {}
+
+    def get_style_for_token(self, token_type: TokenType) -> Style:
+        """Look up style in the style map."""
+        try:
+            return self._style_cache[token_type]
+        except KeyError:
+            # Styles form a hierarchy
+            # We need to go from most to least specific
+            # e.g. ("foo", "bar", "baz") to ("foo", "bar")  to ("foo",)
+            get_style = self.style_map.get
+            token = tuple(token_type)
+            style = self._missing_style
+            while token:
+                _style = get_style(token)
+                if _style is not None:
+                    style = _style
+                    break
+                token = token[:-1]
+            self._style_cache[token_type] = style
+            return style
+
+    def get_background_style(self) -> Style:
+        return self._background_style
+
+
+SyntaxPosition = Tuple[int, int]
+
+
+class _SyntaxHighlightRange(NamedTuple):
+    """
+    A range to highlight in a Syntax object.
+    `start` and `end` are 2-integers tuples, where the first integer is the line number
+    (starting from 1) and the second integer is the column index (starting from 0).
+    """
+
+    style: StyleType
+    start: SyntaxPosition
+    end: SyntaxPosition
+    style_before: bool = False
+
+
+class Syntax(JupyterMixin):
+    """Construct a Syntax object to render syntax highlighted code.
+
+    Args:
+        code (str): Code to highlight.
+        lexer (Lexer | str): Lexer to use (see https://pygments.org/docs/lexers/)
+        theme (str, optional): Color theme, aka Pygments style (see https://pygments.org/docs/styles/#getting-a-list-of-available-styles). Defaults to "monokai".
+        dedent (bool, optional): Enable stripping of initial whitespace. Defaults to False.
+        line_numbers (bool, optional): Enable rendering of line numbers. Defaults to False.
+        start_line (int, optional): Starting number for line numbers. Defaults to 1.
+        line_range (Tuple[int | None, int | None], optional): If given should be a tuple of the start and end line to render.
+            A value of None in the tuple indicates the range is open in that direction.
+        highlight_lines (Set[int]): A set of line numbers to highlight.
+        code_width: Width of code to render (not including line numbers), or ``None`` to use all available width.
+        tab_size (int, optional): Size of tabs. Defaults to 4.
+        word_wrap (bool, optional): Enable word wrapping.
+        background_color (str, optional): Optional background color, or None to use theme color. Defaults to None.
+        indent_guides (bool, optional): Show indent guides. Defaults to False.
+        padding (PaddingDimensions): Padding to apply around the syntax. Defaults to 0 (no padding).
+    """
+
+    _pygments_style_class: Type[PygmentsStyle]
+    _theme: SyntaxTheme
+
+    @classmethod
+    def get_theme(cls, name: Union[str, SyntaxTheme]) -> SyntaxTheme:
+        """Get a syntax theme instance."""
+        if isinstance(name, SyntaxTheme):
+            return name
+        theme: SyntaxTheme
+        if name in RICH_SYNTAX_THEMES:
+            theme = ANSISyntaxTheme(RICH_SYNTAX_THEMES[name])
+        else:
+            theme = PygmentsSyntaxTheme(name)
+        return theme
+
+    def __init__(
+        self,
+        code: str,
+        lexer: Union[Lexer, str],
+        *,
+        theme: Union[str, SyntaxTheme] = DEFAULT_THEME,
+        dedent: bool = False,
+        line_numbers: bool = False,
+        start_line: int = 1,
+        line_range: Optional[Tuple[Optional[int], Optional[int]]] = None,
+        highlight_lines: Optional[Set[int]] = None,
+        code_width: Optional[int] = None,
+        tab_size: int = 4,
+        word_wrap: bool = False,
+        background_color: Optional[str] = None,
+        indent_guides: bool = False,
+        padding: PaddingDimensions = 0,
+    ) -> None:
+        self.code = code
+        self._lexer = lexer
+        self.dedent = dedent
+        self.line_numbers = line_numbers
+        self.start_line = start_line
+        self.line_range = line_range
+        self.highlight_lines = highlight_lines or set()
+        self.code_width = code_width
+        self.tab_size = tab_size
+        self.word_wrap = word_wrap
+        self.background_color = background_color
+        self.background_style = (
+            Style(bgcolor=background_color) if background_color else Style()
+        )
+        self.indent_guides = indent_guides
+        self.padding = padding
+
+        self._theme = self.get_theme(theme)
+        self._stylized_ranges: List[_SyntaxHighlightRange] = []
+
+    @classmethod
+    def from_path(
+        cls,
+        path: str,
+        encoding: str = "utf-8",
+        lexer: Optional[Union[Lexer, str]] = None,
+        theme: Union[str, SyntaxTheme] = DEFAULT_THEME,
+        dedent: bool = False,
+        line_numbers: bool = False,
+        line_range: Optional[Tuple[int, int]] = None,
+        start_line: int = 1,
+        highlight_lines: Optional[Set[int]] = None,
+        code_width: Optional[int] = None,
+        tab_size: int = 4,
+        word_wrap: bool = False,
+        background_color: Optional[str] = None,
+        indent_guides: bool = False,
+        padding: PaddingDimensions = 0,
+    ) -> "Syntax":
+        """Construct a Syntax object from a file.
+
+        Args:
+            path (str): Path to file to highlight.
+            encoding (str): Encoding of file.
+            lexer (str | Lexer, optional): Lexer to use. If None, lexer will be auto-detected from path/file content.
+            theme (str, optional): Color theme, aka Pygments style (see https://pygments.org/docs/styles/#getting-a-list-of-available-styles). Defaults to "emacs".
+            dedent (bool, optional): Enable stripping of initial whitespace. Defaults to True.
+            line_numbers (bool, optional): Enable rendering of line numbers. Defaults to False.
+            start_line (int, optional): Starting number for line numbers. Defaults to 1.
+            line_range (Tuple[int, int], optional): If given should be a tuple of the start and end line to render.
+            highlight_lines (Set[int]): A set of line numbers to highlight.
+            code_width: Width of code to render (not including line numbers), or ``None`` to use all available width.
+            tab_size (int, optional): Size of tabs. Defaults to 4.
+            word_wrap (bool, optional): Enable word wrapping of code.
+            background_color (str, optional): Optional background color, or None to use theme color. Defaults to None.
+            indent_guides (bool, optional): Show indent guides. Defaults to False.
+            padding (PaddingDimensions): Padding to apply around the syntax. Defaults to 0 (no padding).
+
+        Returns:
+            [Syntax]: A Syntax object that may be printed to the console
+        """
+        code = Path(path).read_text(encoding=encoding)
+
+        if not lexer:
+            lexer = cls.guess_lexer(path, code=code)
+
+        return cls(
+            code,
+            lexer,
+            theme=theme,
+            dedent=dedent,
+            line_numbers=line_numbers,
+            line_range=line_range,
+            start_line=start_line,
+            highlight_lines=highlight_lines,
+            code_width=code_width,
+            tab_size=tab_size,
+            word_wrap=word_wrap,
+            background_color=background_color,
+            indent_guides=indent_guides,
+            padding=padding,
+        )
+
+    @classmethod
+    def guess_lexer(cls, path: str, code: Optional[str] = None) -> str:
+        """Guess the alias of the Pygments lexer to use based on a path and an optional string of code.
+        If code is supplied, it will use a combination of the code and the filename to determine the
+        best lexer to use. For example, if the file is ``index.html`` and the file contains Django
+        templating syntax, then "html+django" will be returned. If the file is ``index.html``, and no
+        templating language is used, the "html" lexer will be used. If no string of code
+        is supplied, the lexer will be chosen based on the file extension..
+
+        Args:
+             path (AnyStr): The path to the file containing the code you wish to know the lexer for.
+             code (str, optional): Optional string of code that will be used as a fallback if no lexer
+                is found for the supplied path.
+
+        Returns:
+            str: The name of the Pygments lexer that best matches the supplied path/code.
+        """
+        lexer: Optional[Lexer] = None
+        lexer_name = "default"
+        if code:
+            try:
+                lexer = guess_lexer_for_filename(path, code)
+            except ClassNotFound:
+                pass
+
+        if not lexer:
+            try:
+                _, ext = os.path.splitext(path)
+                if ext:
+                    extension = ext.lstrip(".").lower()
+                    lexer = get_lexer_by_name(extension)
+            except ClassNotFound:
+                pass
+
+        if lexer:
+            if lexer.aliases:
+                lexer_name = lexer.aliases[0]
+            else:
+                lexer_name = lexer.name
+
+        return lexer_name
+
+    def _get_base_style(self) -> Style:
+        """Get the base style."""
+        default_style = self._theme.get_background_style() + self.background_style
+        return default_style
+
+    def _get_token_color(self, token_type: TokenType) -> Optional[Color]:
+        """Get a color (if any) for the given token.
+
+        Args:
+            token_type (TokenType): A token type tuple from Pygments.
+
+        Returns:
+            Optional[Color]: Color from theme, or None for no color.
+        """
+        style = self._theme.get_style_for_token(token_type)
+        return style.color
+
+    @property
+    def lexer(self) -> Optional[Lexer]:
+        """The lexer for this syntax, or None if no lexer was found.
+
+        Tries to find the lexer by name if a string was passed to the constructor.
+        """
+
+        if isinstance(self._lexer, Lexer):
+            return self._lexer
+        try:
+            return get_lexer_by_name(
+                self._lexer,
+                stripnl=False,
+                ensurenl=True,
+                tabsize=self.tab_size,
+            )
+        except ClassNotFound:
+            return None
+
+    @property
+    def default_lexer(self) -> Lexer:
+        """A Pygments Lexer to use if one is not specified or invalid."""
+        return get_lexer_by_name(
+            "text",
+            stripnl=False,
+            ensurenl=True,
+            tabsize=self.tab_size,
+        )
+
+    def highlight(
+        self,
+        code: str,
+        line_range: Optional[Tuple[Optional[int], Optional[int]]] = None,
+    ) -> Text:
+        """Highlight code and return a Text instance.
+
+        Args:
+            code (str): Code to highlight.
+            line_range(Tuple[int, int], optional): Optional line range to highlight.
+
+        Returns:
+            Text: A text instance containing highlighted syntax.
+        """
+
+        base_style = self._get_base_style()
+        justify: JustifyMethod = (
+            "default" if base_style.transparent_background else "left"
+        )
+
+        text = Text(
+            justify=justify,
+            style=base_style,
+            tab_size=self.tab_size,
+            no_wrap=not self.word_wrap,
+        )
+        _get_theme_style = self._theme.get_style_for_token
+
+        lexer = self.lexer or self.default_lexer
+
+        if lexer is None:
+            text.append(code)
+        else:
+            if line_range:
+                # More complicated path to only stylize a portion of the code
+                # This speeds up further operations as there are less spans to process
+                line_start, line_end = line_range
+
+                def line_tokenize() -> Iterable[Tuple[Any, str]]:
+                    """Split tokens to one per line."""
+                    assert lexer  # required to make MyPy happy - we know lexer is not None at this point
+
+                    for token_type, token in lexer.get_tokens(code):
+                        while token:
+                            line_token, new_line, token = token.partition("\n")
+                            yield token_type, line_token + new_line
+
+                def tokens_to_spans() -> Iterable[Tuple[str, Optional[Style]]]:
+                    """Convert tokens to spans."""
+                    tokens = iter(line_tokenize())
+                    line_no = 0
+                    _line_start = line_start - 1 if line_start else 0
+
+                    # Skip over tokens until line start
+                    while line_no < _line_start:
+                        try:
+                            _token_type, token = next(tokens)
+                        except StopIteration:
+                            break
+                        yield (token, None)
+                        if token.endswith("\n"):
+                            line_no += 1
+                    # Generate spans until line end
+                    for token_type, token in tokens:
+                        yield (token, _get_theme_style(token_type))
+                        if token.endswith("\n"):
+                            line_no += 1
+                            if line_end and line_no >= line_end:
+                                break
+
+                text.append_tokens(tokens_to_spans())
+
+            else:
+                text.append_tokens(
+                    (token, _get_theme_style(token_type))
+                    for token_type, token in lexer.get_tokens(code)
+                )
+            if self.background_color is not None:
+                text.stylize(f"on {self.background_color}")
+
+        if self._stylized_ranges:
+            self._apply_stylized_ranges(text)
+
+        return text
+
+    def stylize_range(
+        self,
+        style: StyleType,
+        start: SyntaxPosition,
+        end: SyntaxPosition,
+        style_before: bool = False,
+    ) -> None:
+        """
+        Adds a custom style on a part of the code, that will be applied to the syntax display when it's rendered.
+        Line numbers are 1-based, while column indexes are 0-based.
+
+        Args:
+            style (StyleType): The style to apply.
+            start (Tuple[int, int]): The start of the range, in the form `[line number, column index]`.
+            end (Tuple[int, int]): The end of the range, in the form `[line number, column index]`.
+            style_before (bool): Apply the style before any existing styles.
+        """
+        self._stylized_ranges.append(
+            _SyntaxHighlightRange(style, start, end, style_before)
+        )
+
+    def _get_line_numbers_color(self, blend: float = 0.3) -> Color:
+        background_style = self._theme.get_background_style() + self.background_style
+        background_color = background_style.bgcolor
+        if background_color is None or background_color.is_system_defined:
+            return Color.default()
+        foreground_color = self._get_token_color(Token.Text)
+        if foreground_color is None or foreground_color.is_system_defined:
+            return foreground_color or Color.default()
+        new_color = blend_rgb(
+            background_color.get_truecolor(),
+            foreground_color.get_truecolor(),
+            cross_fade=blend,
+        )
+        return Color.from_triplet(new_color)
+
+    @property
+    def _numbers_column_width(self) -> int:
+        """Get the number of characters used to render the numbers column."""
+        column_width = 0
+        if self.line_numbers:
+            column_width = (
+                len(str(self.start_line + self.code.count("\n")))
+                + NUMBERS_COLUMN_DEFAULT_PADDING
+            )
+        return column_width
+
+    def _get_number_styles(self, console: Console) -> Tuple[Style, Style, Style]:
+        """Get background, number, and highlight styles for line numbers."""
+        background_style = self._get_base_style()
+        if background_style.transparent_background:
+            return Style.null(), Style(dim=True), Style.null()
+        if console.color_system in ("256", "truecolor"):
+            number_style = Style.chain(
+                background_style,
+                self._theme.get_style_for_token(Token.Text),
+                Style(color=self._get_line_numbers_color()),
+                self.background_style,
+            )
+            highlight_number_style = Style.chain(
+                background_style,
+                self._theme.get_style_for_token(Token.Text),
+                Style(bold=True, color=self._get_line_numbers_color(0.9)),
+                self.background_style,
+            )
+        else:
+            number_style = background_style + Style(dim=True)
+            highlight_number_style = background_style + Style(dim=False)
+        return background_style, number_style, highlight_number_style
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        _, right, _, left = Padding.unpack(self.padding)
+        padding = left + right
+        if self.code_width is not None:
+            width = self.code_width + self._numbers_column_width + padding + 1
+            return Measurement(self._numbers_column_width, width)
+        lines = self.code.splitlines()
+        width = (
+            self._numbers_column_width
+            + padding
+            + (max(cell_len(line) for line in lines) if lines else 0)
+        )
+        if self.line_numbers:
+            width += 1
+        return Measurement(self._numbers_column_width, width)
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        segments = Segments(self._get_syntax(console, options))
+        if self.padding:
+            yield Padding(segments, style=self._get_base_style(), pad=self.padding)
+        else:
+            yield segments
+
+    def _get_syntax(
+        self,
+        console: Console,
+        options: ConsoleOptions,
+    ) -> Iterable[Segment]:
+        """
+        Get the Segments for the Syntax object, excluding any vertical/horizontal padding
+        """
+        transparent_background = self._get_base_style().transparent_background
+        code_width = (
+            (
+                (options.max_width - self._numbers_column_width - 1)
+                if self.line_numbers
+                else options.max_width
+            )
+            if self.code_width is None
+            else self.code_width
+        )
+
+        ends_on_nl, processed_code = self._process_code(self.code)
+        text = self.highlight(processed_code, self.line_range)
+
+        if not self.line_numbers and not self.word_wrap and not self.line_range:
+            if not ends_on_nl:
+                text.remove_suffix("\n")
+            # Simple case of just rendering text
+            style = (
+                self._get_base_style()
+                + self._theme.get_style_for_token(Comment)
+                + Style(dim=True)
+                + self.background_style
+            )
+            if self.indent_guides and not options.ascii_only:
+                text = text.with_indent_guides(self.tab_size, style=style)
+                text.overflow = "crop"
+            if style.transparent_background:
+                yield from console.render(
+                    text, options=options.update(width=code_width)
+                )
+            else:
+                syntax_lines = console.render_lines(
+                    text,
+                    options.update(width=code_width, height=None, justify="left"),
+                    style=self.background_style,
+                    pad=True,
+                    new_lines=True,
+                )
+                for syntax_line in syntax_lines:
+                    yield from syntax_line
+            return
+
+        start_line, end_line = self.line_range or (None, None)
+        line_offset = 0
+        if start_line:
+            line_offset = max(0, start_line - 1)
+        lines: Union[List[Text], Lines] = text.split("\n", allow_blank=ends_on_nl)
+        if self.line_range:
+            if line_offset > len(lines):
+                return
+            lines = lines[line_offset:end_line]
+
+        if self.indent_guides and not options.ascii_only:
+            style = (
+                self._get_base_style()
+                + self._theme.get_style_for_token(Comment)
+                + Style(dim=True)
+                + self.background_style
+            )
+            lines = (
+                Text("\n")
+                .join(lines)
+                .with_indent_guides(self.tab_size, style=style + Style(italic=False))
+                .split("\n", allow_blank=True)
+            )
+
+        numbers_column_width = self._numbers_column_width
+        render_options = options.update(width=code_width)
+
+        highlight_line = self.highlight_lines.__contains__
+        _Segment = Segment
+        new_line = _Segment("\n")
+
+        line_pointer = "> " if options.legacy_windows else "❱ "
+
+        (
+            background_style,
+            number_style,
+            highlight_number_style,
+        ) = self._get_number_styles(console)
+
+        for line_no, line in enumerate(lines, self.start_line + line_offset):
+            if self.word_wrap:
+                wrapped_lines = console.render_lines(
+                    line,
+                    render_options.update(height=None, justify="left"),
+                    style=background_style,
+                    pad=not transparent_background,
+                )
+            else:
+                segments = list(line.render(console, end=""))
+                if options.no_wrap:
+                    wrapped_lines = [segments]
+                else:
+                    wrapped_lines = [
+                        _Segment.adjust_line_length(
+                            segments,
+                            render_options.max_width,
+                            style=background_style,
+                            pad=not transparent_background,
+                        )
+                    ]
+
+            if self.line_numbers:
+                wrapped_line_left_pad = _Segment(
+                    " " * numbers_column_width + " ", background_style
+                )
+                for first, wrapped_line in loop_first(wrapped_lines):
+                    if first:
+                        line_column = str(line_no).rjust(numbers_column_width - 2) + " "
+                        if highlight_line(line_no):
+                            yield _Segment(line_pointer, Style(color="red"))
+                            yield _Segment(line_column, highlight_number_style)
+                        else:
+                            yield _Segment("  ", highlight_number_style)
+                            yield _Segment(line_column, number_style)
+                    else:
+                        yield wrapped_line_left_pad
+                    yield from wrapped_line
+                    yield new_line
+            else:
+                for wrapped_line in wrapped_lines:
+                    yield from wrapped_line
+                    yield new_line
+
+    def _apply_stylized_ranges(self, text: Text) -> None:
+        """
+        Apply stylized ranges to a text instance,
+        using the given code to determine the right portion to apply the style to.
+
+        Args:
+            text (Text): Text instance to apply the style to.
+        """
+        code = text.plain
+        newlines_offsets = [
+            # Let's add outer boundaries at each side of the list:
+            0,
+            # N.B. using "\n" here is much faster than using metacharacters such as "^" or "\Z":
+            *[
+                match.start() + 1
+                for match in re.finditer("\n", code, flags=re.MULTILINE)
+            ],
+            len(code) + 1,
+        ]
+
+        for stylized_range in self._stylized_ranges:
+            start = _get_code_index_for_syntax_position(
+                newlines_offsets, stylized_range.start
+            )
+            end = _get_code_index_for_syntax_position(
+                newlines_offsets, stylized_range.end
+            )
+            if start is not None and end is not None:
+                if stylized_range.style_before:
+                    text.stylize_before(stylized_range.style, start, end)
+                else:
+                    text.stylize(stylized_range.style, start, end)
+
+    def _process_code(self, code: str) -> Tuple[bool, str]:
+        """
+        Applies various processing to a raw code string
+        (normalises it so it always ends with a line return, dedents it if necessary, etc.)
+
+        Args:
+            code (str): The raw code string to process
+
+        Returns:
+            Tuple[bool, str]: the boolean indicates whether the raw code ends with a line return,
+                while the string is the processed code.
+        """
+        ends_on_nl = code.endswith("\n")
+        processed_code = code if ends_on_nl else code + "\n"
+        processed_code = (
+            textwrap.dedent(processed_code) if self.dedent else processed_code
+        )
+        processed_code = processed_code.expandtabs(self.tab_size)
+        return ends_on_nl, processed_code
+
+
+def _get_code_index_for_syntax_position(
+    newlines_offsets: Sequence[int], position: SyntaxPosition
+) -> Optional[int]:
+    """
+    Returns the index of the code string for the given positions.
+
+    Args:
+        newlines_offsets (Sequence[int]): The offset of each newline character found in the code snippet.
+        position (SyntaxPosition): The position to search for.
+
+    Returns:
+        Optional[int]: The index of the code string for this position, or `None`
+            if the given position's line number is out of range (if it's the column that is out of range
+            we silently clamp its value so that it reaches the end of the line)
+    """
+    lines_count = len(newlines_offsets)
+
+    line_number, column_index = position
+    if line_number > lines_count or len(newlines_offsets) < (line_number + 1):
+        return None  # `line_number` is out of range
+    line_index = line_number - 1
+    line_length = newlines_offsets[line_index + 1] - newlines_offsets[line_index] - 1
+    # If `column_index` is out of range: let's silently clamp it:
+    column_index = min(line_length, column_index)
+    return newlines_offsets[line_index] + column_index
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import argparse
+    import sys
+
+    parser = argparse.ArgumentParser(
+        description="Render syntax to the console with Rich"
+    )
+    parser.add_argument(
+        "path",
+        metavar="PATH",
+        help="path to file, or - for stdin",
+    )
+    parser.add_argument(
+        "-c",
+        "--force-color",
+        dest="force_color",
+        action="store_true",
+        default=None,
+        help="force color for non-terminals",
+    )
+    parser.add_argument(
+        "-i",
+        "--indent-guides",
+        dest="indent_guides",
+        action="store_true",
+        default=False,
+        help="display indent guides",
+    )
+    parser.add_argument(
+        "-l",
+        "--line-numbers",
+        dest="line_numbers",
+        action="store_true",
+        help="render line numbers",
+    )
+    parser.add_argument(
+        "-w",
+        "--width",
+        type=int,
+        dest="width",
+        default=None,
+        help="width of output (default will auto-detect)",
+    )
+    parser.add_argument(
+        "-r",
+        "--wrap",
+        dest="word_wrap",
+        action="store_true",
+        default=False,
+        help="word wrap long lines",
+    )
+    parser.add_argument(
+        "-s",
+        "--soft-wrap",
+        action="store_true",
+        dest="soft_wrap",
+        default=False,
+        help="enable soft wrapping mode",
+    )
+    parser.add_argument(
+        "-t", "--theme", dest="theme", default="monokai", help="pygments theme"
+    )
+    parser.add_argument(
+        "-b",
+        "--background-color",
+        dest="background_color",
+        default=None,
+        help="Override background color",
+    )
+    parser.add_argument(
+        "-x",
+        "--lexer",
+        default=None,
+        dest="lexer_name",
+        help="Lexer name",
+    )
+    parser.add_argument(
+        "-p", "--padding", type=int, default=0, dest="padding", help="Padding"
+    )
+    parser.add_argument(
+        "--highlight-line",
+        type=int,
+        default=None,
+        dest="highlight_line",
+        help="The line number (not index!) to highlight",
+    )
+    args = parser.parse_args()
+
+    from pip._vendor.rich.console import Console
+
+    console = Console(force_terminal=args.force_color, width=args.width)
+
+    if args.path == "-":
+        code = sys.stdin.read()
+        syntax = Syntax(
+            code=code,
+            lexer=args.lexer_name,
+            line_numbers=args.line_numbers,
+            word_wrap=args.word_wrap,
+            theme=args.theme,
+            background_color=args.background_color,
+            indent_guides=args.indent_guides,
+            padding=args.padding,
+            highlight_lines={args.highlight_line},
+        )
+    else:
+        syntax = Syntax.from_path(
+            args.path,
+            lexer=args.lexer_name,
+            line_numbers=args.line_numbers,
+            word_wrap=args.word_wrap,
+            theme=args.theme,
+            background_color=args.background_color,
+            indent_guides=args.indent_guides,
+            padding=args.padding,
+            highlight_lines={args.highlight_line},
+        )
+    console.print(syntax, soft_wrap=args.soft_wrap)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/table.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/table.py
new file mode 100644
index 00000000..654c8555
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/table.py
@@ -0,0 +1,1007 @@
+from dataclasses import dataclass, field, replace
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Sequence,
+    Tuple,
+    Union,
+)
+
+from . import box, errors
+from ._loop import loop_first_last, loop_last
+from ._pick import pick_bool
+from ._ratio import ratio_distribute, ratio_reduce
+from .align import VerticalAlignMethod
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .padding import Padding, PaddingDimensions
+from .protocol import is_renderable
+from .segment import Segment
+from .style import Style, StyleType
+from .text import Text, TextType
+
+if TYPE_CHECKING:
+    from .console import (
+        Console,
+        ConsoleOptions,
+        JustifyMethod,
+        OverflowMethod,
+        RenderableType,
+        RenderResult,
+    )
+
+
+@dataclass
+class Column:
+    """Defines a column within a ~Table.
+
+    Args:
+        title (Union[str, Text], optional): The title of the table rendered at the top. Defaults to None.
+        caption (Union[str, Text], optional): The table caption rendered below. Defaults to None.
+        width (int, optional): The width in characters of the table, or ``None`` to automatically fit. Defaults to None.
+        min_width (Optional[int], optional): The minimum width of the table, or ``None`` for no minimum. Defaults to None.
+        box (box.Box, optional): One of the constants in box.py used to draw the edges (see :ref:`appendix_box`), or ``None`` for no box lines. Defaults to box.HEAVY_HEAD.
+        safe_box (Optional[bool], optional): Disable box characters that don't display on windows legacy terminal with *raster* fonts. Defaults to True.
+        padding (PaddingDimensions, optional): Padding for cells (top, right, bottom, left). Defaults to (0, 1).
+        collapse_padding (bool, optional): Enable collapsing of padding around cells. Defaults to False.
+        pad_edge (bool, optional): Enable padding of edge cells. Defaults to True.
+        expand (bool, optional): Expand the table to fit the available space if ``True``, otherwise the table width will be auto-calculated. Defaults to False.
+        show_header (bool, optional): Show a header row. Defaults to True.
+        show_footer (bool, optional): Show a footer row. Defaults to False.
+        show_edge (bool, optional): Draw a box around the outside of the table. Defaults to True.
+        show_lines (bool, optional): Draw lines between every row. Defaults to False.
+        leading (int, optional): Number of blank lines between rows (precludes ``show_lines``). Defaults to 0.
+        style (Union[str, Style], optional): Default style for the table. Defaults to "none".
+        row_styles (List[Union, str], optional): Optional list of row styles, if more than one style is given then the styles will alternate. Defaults to None.
+        header_style (Union[str, Style], optional): Style of the header. Defaults to "table.header".
+        footer_style (Union[str, Style], optional): Style of the footer. Defaults to "table.footer".
+        border_style (Union[str, Style], optional): Style of the border. Defaults to None.
+        title_style (Union[str, Style], optional): Style of the title. Defaults to None.
+        caption_style (Union[str, Style], optional): Style of the caption. Defaults to None.
+        title_justify (str, optional): Justify method for title. Defaults to "center".
+        caption_justify (str, optional): Justify method for caption. Defaults to "center".
+        highlight (bool, optional): Highlight cell contents (if str). Defaults to False.
+    """
+
+    header: "RenderableType" = ""
+    """RenderableType: Renderable for the header (typically a string)"""
+
+    footer: "RenderableType" = ""
+    """RenderableType: Renderable for the footer (typically a string)"""
+
+    header_style: StyleType = ""
+    """StyleType: The style of the header."""
+
+    footer_style: StyleType = ""
+    """StyleType: The style of the footer."""
+
+    style: StyleType = ""
+    """StyleType: The style of the column."""
+
+    justify: "JustifyMethod" = "left"
+    """str: How to justify text within the column ("left", "center", "right", or "full")"""
+
+    vertical: "VerticalAlignMethod" = "top"
+    """str: How to vertically align content ("top", "middle", or "bottom")"""
+
+    overflow: "OverflowMethod" = "ellipsis"
+    """str: Overflow method."""
+
+    width: Optional[int] = None
+    """Optional[int]: Width of the column, or ``None`` (default) to auto calculate width."""
+
+    min_width: Optional[int] = None
+    """Optional[int]: Minimum width of column, or ``None`` for no minimum. Defaults to None."""
+
+    max_width: Optional[int] = None
+    """Optional[int]: Maximum width of column, or ``None`` for no maximum. Defaults to None."""
+
+    ratio: Optional[int] = None
+    """Optional[int]: Ratio to use when calculating column width, or ``None`` (default) to adapt to column contents."""
+
+    no_wrap: bool = False
+    """bool: Prevent wrapping of text within the column. Defaults to ``False``."""
+
+    highlight: bool = False
+    """bool: Apply highlighter to column. Defaults to ``False``."""
+
+    _index: int = 0
+    """Index of column."""
+
+    _cells: List["RenderableType"] = field(default_factory=list)
+
+    def copy(self) -> "Column":
+        """Return a copy of this Column."""
+        return replace(self, _cells=[])
+
+    @property
+    def cells(self) -> Iterable["RenderableType"]:
+        """Get all cells in the column, not including header."""
+        yield from self._cells
+
+    @property
+    def flexible(self) -> bool:
+        """Check if this column is flexible."""
+        return self.ratio is not None
+
+
+@dataclass
+class Row:
+    """Information regarding a row."""
+
+    style: Optional[StyleType] = None
+    """Style to apply to row."""
+
+    end_section: bool = False
+    """Indicated end of section, which will force a line beneath the row."""
+
+
+class _Cell(NamedTuple):
+    """A single cell in a table."""
+
+    style: StyleType
+    """Style to apply to cell."""
+    renderable: "RenderableType"
+    """Cell renderable."""
+    vertical: VerticalAlignMethod
+    """Cell vertical alignment."""
+
+
+class Table(JupyterMixin):
+    """A console renderable to draw a table.
+
+    Args:
+        *headers (Union[Column, str]): Column headers, either as a string, or :class:`~rich.table.Column` instance.
+        title (Union[str, Text], optional): The title of the table rendered at the top. Defaults to None.
+        caption (Union[str, Text], optional): The table caption rendered below. Defaults to None.
+        width (int, optional): The width in characters of the table, or ``None`` to automatically fit. Defaults to None.
+        min_width (Optional[int], optional): The minimum width of the table, or ``None`` for no minimum. Defaults to None.
+        box (box.Box, optional): One of the constants in box.py used to draw the edges (see :ref:`appendix_box`), or ``None`` for no box lines. Defaults to box.HEAVY_HEAD.
+        safe_box (Optional[bool], optional): Disable box characters that don't display on windows legacy terminal with *raster* fonts. Defaults to True.
+        padding (PaddingDimensions, optional): Padding for cells (top, right, bottom, left). Defaults to (0, 1).
+        collapse_padding (bool, optional): Enable collapsing of padding around cells. Defaults to False.
+        pad_edge (bool, optional): Enable padding of edge cells. Defaults to True.
+        expand (bool, optional): Expand the table to fit the available space if ``True``, otherwise the table width will be auto-calculated. Defaults to False.
+        show_header (bool, optional): Show a header row. Defaults to True.
+        show_footer (bool, optional): Show a footer row. Defaults to False.
+        show_edge (bool, optional): Draw a box around the outside of the table. Defaults to True.
+        show_lines (bool, optional): Draw lines between every row. Defaults to False.
+        leading (int, optional): Number of blank lines between rows (precludes ``show_lines``). Defaults to 0.
+        style (Union[str, Style], optional): Default style for the table. Defaults to "none".
+        row_styles (List[Union, str], optional): Optional list of row styles, if more than one style is given then the styles will alternate. Defaults to None.
+        header_style (Union[str, Style], optional): Style of the header. Defaults to "table.header".
+        footer_style (Union[str, Style], optional): Style of the footer. Defaults to "table.footer".
+        border_style (Union[str, Style], optional): Style of the border. Defaults to None.
+        title_style (Union[str, Style], optional): Style of the title. Defaults to None.
+        caption_style (Union[str, Style], optional): Style of the caption. Defaults to None.
+        title_justify (str, optional): Justify method for title. Defaults to "center".
+        caption_justify (str, optional): Justify method for caption. Defaults to "center".
+        highlight (bool, optional): Highlight cell contents (if str). Defaults to False.
+    """
+
+    columns: List[Column]
+    rows: List[Row]
+
+    def __init__(
+        self,
+        *headers: Union[Column, str],
+        title: Optional[TextType] = None,
+        caption: Optional[TextType] = None,
+        width: Optional[int] = None,
+        min_width: Optional[int] = None,
+        box: Optional[box.Box] = box.HEAVY_HEAD,
+        safe_box: Optional[bool] = None,
+        padding: PaddingDimensions = (0, 1),
+        collapse_padding: bool = False,
+        pad_edge: bool = True,
+        expand: bool = False,
+        show_header: bool = True,
+        show_footer: bool = False,
+        show_edge: bool = True,
+        show_lines: bool = False,
+        leading: int = 0,
+        style: StyleType = "none",
+        row_styles: Optional[Iterable[StyleType]] = None,
+        header_style: Optional[StyleType] = "table.header",
+        footer_style: Optional[StyleType] = "table.footer",
+        border_style: Optional[StyleType] = None,
+        title_style: Optional[StyleType] = None,
+        caption_style: Optional[StyleType] = None,
+        title_justify: "JustifyMethod" = "center",
+        caption_justify: "JustifyMethod" = "center",
+        highlight: bool = False,
+    ) -> None:
+        self.columns: List[Column] = []
+        self.rows: List[Row] = []
+        self.title = title
+        self.caption = caption
+        self.width = width
+        self.min_width = min_width
+        self.box = box
+        self.safe_box = safe_box
+        self._padding = Padding.unpack(padding)
+        self.pad_edge = pad_edge
+        self._expand = expand
+        self.show_header = show_header
+        self.show_footer = show_footer
+        self.show_edge = show_edge
+        self.show_lines = show_lines
+        self.leading = leading
+        self.collapse_padding = collapse_padding
+        self.style = style
+        self.header_style = header_style or ""
+        self.footer_style = footer_style or ""
+        self.border_style = border_style
+        self.title_style = title_style
+        self.caption_style = caption_style
+        self.title_justify: "JustifyMethod" = title_justify
+        self.caption_justify: "JustifyMethod" = caption_justify
+        self.highlight = highlight
+        self.row_styles: Sequence[StyleType] = list(row_styles or [])
+        append_column = self.columns.append
+        for header in headers:
+            if isinstance(header, str):
+                self.add_column(header=header)
+            else:
+                header._index = len(self.columns)
+                append_column(header)
+
+    @classmethod
+    def grid(
+        cls,
+        *headers: Union[Column, str],
+        padding: PaddingDimensions = 0,
+        collapse_padding: bool = True,
+        pad_edge: bool = False,
+        expand: bool = False,
+    ) -> "Table":
+        """Get a table with no lines, headers, or footer.
+
+        Args:
+            *headers (Union[Column, str]): Column headers, either as a string, or :class:`~rich.table.Column` instance.
+            padding (PaddingDimensions, optional): Get padding around cells. Defaults to 0.
+            collapse_padding (bool, optional): Enable collapsing of padding around cells. Defaults to True.
+            pad_edge (bool, optional): Enable padding around edges of table. Defaults to False.
+            expand (bool, optional): Expand the table to fit the available space if ``True``, otherwise the table width will be auto-calculated. Defaults to False.
+
+        Returns:
+            Table: A table instance.
+        """
+        return cls(
+            *headers,
+            box=None,
+            padding=padding,
+            collapse_padding=collapse_padding,
+            show_header=False,
+            show_footer=False,
+            show_edge=False,
+            pad_edge=pad_edge,
+            expand=expand,
+        )
+
+    @property
+    def expand(self) -> bool:
+        """Setting a non-None self.width implies expand."""
+        return self._expand or self.width is not None
+
+    @expand.setter
+    def expand(self, expand: bool) -> None:
+        """Set expand."""
+        self._expand = expand
+
+    @property
+    def _extra_width(self) -> int:
+        """Get extra width to add to cell content."""
+        width = 0
+        if self.box and self.show_edge:
+            width += 2
+        if self.box:
+            width += len(self.columns) - 1
+        return width
+
+    @property
+    def row_count(self) -> int:
+        """Get the current number of rows."""
+        return len(self.rows)
+
+    def get_row_style(self, console: "Console", index: int) -> StyleType:
+        """Get the current row style."""
+        style = Style.null()
+        if self.row_styles:
+            style += console.get_style(self.row_styles[index % len(self.row_styles)])
+        row_style = self.rows[index].style
+        if row_style is not None:
+            style += console.get_style(row_style)
+        return style
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        max_width = options.max_width
+        if self.width is not None:
+            max_width = self.width
+        if max_width < 0:
+            return Measurement(0, 0)
+
+        extra_width = self._extra_width
+        max_width = sum(
+            self._calculate_column_widths(
+                console, options.update_width(max_width - extra_width)
+            )
+        )
+        _measure_column = self._measure_column
+
+        measurements = [
+            _measure_column(console, options.update_width(max_width), column)
+            for column in self.columns
+        ]
+        minimum_width = (
+            sum(measurement.minimum for measurement in measurements) + extra_width
+        )
+        maximum_width = (
+            sum(measurement.maximum for measurement in measurements) + extra_width
+            if (self.width is None)
+            else self.width
+        )
+        measurement = Measurement(minimum_width, maximum_width)
+        measurement = measurement.clamp(self.min_width)
+        return measurement
+
+    @property
+    def padding(self) -> Tuple[int, int, int, int]:
+        """Get cell padding."""
+        return self._padding
+
+    @padding.setter
+    def padding(self, padding: PaddingDimensions) -> "Table":
+        """Set cell padding."""
+        self._padding = Padding.unpack(padding)
+        return self
+
+    def add_column(
+        self,
+        header: "RenderableType" = "",
+        footer: "RenderableType" = "",
+        *,
+        header_style: Optional[StyleType] = None,
+        highlight: Optional[bool] = None,
+        footer_style: Optional[StyleType] = None,
+        style: Optional[StyleType] = None,
+        justify: "JustifyMethod" = "left",
+        vertical: "VerticalAlignMethod" = "top",
+        overflow: "OverflowMethod" = "ellipsis",
+        width: Optional[int] = None,
+        min_width: Optional[int] = None,
+        max_width: Optional[int] = None,
+        ratio: Optional[int] = None,
+        no_wrap: bool = False,
+    ) -> None:
+        """Add a column to the table.
+
+        Args:
+            header (RenderableType, optional): Text or renderable for the header.
+                Defaults to "".
+            footer (RenderableType, optional): Text or renderable for the footer.
+                Defaults to "".
+            header_style (Union[str, Style], optional): Style for the header, or None for default. Defaults to None.
+            highlight (bool, optional): Whether to highlight the text. The default of None uses the value of the table (self) object.
+            footer_style (Union[str, Style], optional): Style for the footer, or None for default. Defaults to None.
+            style (Union[str, Style], optional): Style for the column cells, or None for default. Defaults to None.
+            justify (JustifyMethod, optional): Alignment for cells. Defaults to "left".
+            vertical (VerticalAlignMethod, optional): Vertical alignment, one of "top", "middle", or "bottom". Defaults to "top".
+            overflow (OverflowMethod): Overflow method: "crop", "fold", "ellipsis". Defaults to "ellipsis".
+            width (int, optional): Desired width of column in characters, or None to fit to contents. Defaults to None.
+            min_width (Optional[int], optional): Minimum width of column, or ``None`` for no minimum. Defaults to None.
+            max_width (Optional[int], optional): Maximum width of column, or ``None`` for no maximum. Defaults to None.
+            ratio (int, optional): Flexible ratio for the column (requires ``Table.expand`` or ``Table.width``). Defaults to None.
+            no_wrap (bool, optional): Set to ``True`` to disable wrapping of this column.
+        """
+
+        column = Column(
+            _index=len(self.columns),
+            header=header,
+            footer=footer,
+            header_style=header_style or "",
+            highlight=highlight if highlight is not None else self.highlight,
+            footer_style=footer_style or "",
+            style=style or "",
+            justify=justify,
+            vertical=vertical,
+            overflow=overflow,
+            width=width,
+            min_width=min_width,
+            max_width=max_width,
+            ratio=ratio,
+            no_wrap=no_wrap,
+        )
+        self.columns.append(column)
+
+    def add_row(
+        self,
+        *renderables: Optional["RenderableType"],
+        style: Optional[StyleType] = None,
+        end_section: bool = False,
+    ) -> None:
+        """Add a row of renderables.
+
+        Args:
+            *renderables (None or renderable): Each cell in a row must be a renderable object (including str),
+                or ``None`` for a blank cell.
+            style (StyleType, optional): An optional style to apply to the entire row. Defaults to None.
+            end_section (bool, optional): End a section and draw a line. Defaults to False.
+
+        Raises:
+            errors.NotRenderableError: If you add something that can't be rendered.
+        """
+
+        def add_cell(column: Column, renderable: "RenderableType") -> None:
+            column._cells.append(renderable)
+
+        cell_renderables: List[Optional["RenderableType"]] = list(renderables)
+
+        columns = self.columns
+        if len(cell_renderables) < len(columns):
+            cell_renderables = [
+                *cell_renderables,
+                *[None] * (len(columns) - len(cell_renderables)),
+            ]
+        for index, renderable in enumerate(cell_renderables):
+            if index == len(columns):
+                column = Column(_index=index, highlight=self.highlight)
+                for _ in self.rows:
+                    add_cell(column, Text(""))
+                self.columns.append(column)
+            else:
+                column = columns[index]
+            if renderable is None:
+                add_cell(column, "")
+            elif is_renderable(renderable):
+                add_cell(column, renderable)
+            else:
+                raise errors.NotRenderableError(
+                    f"unable to render {type(renderable).__name__}; a string or other renderable object is required"
+                )
+        self.rows.append(Row(style=style, end_section=end_section))
+
+    def add_section(self) -> None:
+        """Add a new section (draw a line after current row)."""
+
+        if self.rows:
+            self.rows[-1].end_section = True
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        if not self.columns:
+            yield Segment("\n")
+            return
+
+        max_width = options.max_width
+        if self.width is not None:
+            max_width = self.width
+
+        extra_width = self._extra_width
+        widths = self._calculate_column_widths(
+            console, options.update_width(max_width - extra_width)
+        )
+        table_width = sum(widths) + extra_width
+
+        render_options = options.update(
+            width=table_width, highlight=self.highlight, height=None
+        )
+
+        def render_annotation(
+            text: TextType, style: StyleType, justify: "JustifyMethod" = "center"
+        ) -> "RenderResult":
+            render_text = (
+                console.render_str(text, style=style, highlight=False)
+                if isinstance(text, str)
+                else text
+            )
+            return console.render(
+                render_text, options=render_options.update(justify=justify)
+            )
+
+        if self.title:
+            yield from render_annotation(
+                self.title,
+                style=Style.pick_first(self.title_style, "table.title"),
+                justify=self.title_justify,
+            )
+        yield from self._render(console, render_options, widths)
+        if self.caption:
+            yield from render_annotation(
+                self.caption,
+                style=Style.pick_first(self.caption_style, "table.caption"),
+                justify=self.caption_justify,
+            )
+
+    def _calculate_column_widths(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> List[int]:
+        """Calculate the widths of each column, including padding, not including borders."""
+        max_width = options.max_width
+        columns = self.columns
+        width_ranges = [
+            self._measure_column(console, options, column) for column in columns
+        ]
+        widths = [_range.maximum or 1 for _range in width_ranges]
+        get_padding_width = self._get_padding_width
+        extra_width = self._extra_width
+        if self.expand:
+            ratios = [col.ratio or 0 for col in columns if col.flexible]
+            if any(ratios):
+                fixed_widths = [
+                    0 if column.flexible else _range.maximum
+                    for _range, column in zip(width_ranges, columns)
+                ]
+                flex_minimum = [
+                    (column.width or 1) + get_padding_width(column._index)
+                    for column in columns
+                    if column.flexible
+                ]
+                flexible_width = max_width - sum(fixed_widths)
+                flex_widths = ratio_distribute(flexible_width, ratios, flex_minimum)
+                iter_flex_widths = iter(flex_widths)
+                for index, column in enumerate(columns):
+                    if column.flexible:
+                        widths[index] = fixed_widths[index] + next(iter_flex_widths)
+        table_width = sum(widths)
+
+        if table_width > max_width:
+            widths = self._collapse_widths(
+                widths,
+                [(column.width is None and not column.no_wrap) for column in columns],
+                max_width,
+            )
+            table_width = sum(widths)
+            # last resort, reduce columns evenly
+            if table_width > max_width:
+                excess_width = table_width - max_width
+                widths = ratio_reduce(excess_width, [1] * len(widths), widths, widths)
+                table_width = sum(widths)
+
+            width_ranges = [
+                self._measure_column(console, options.update_width(width), column)
+                for width, column in zip(widths, columns)
+            ]
+            widths = [_range.maximum or 0 for _range in width_ranges]
+
+        if (table_width < max_width and self.expand) or (
+            self.min_width is not None and table_width < (self.min_width - extra_width)
+        ):
+            _max_width = (
+                max_width
+                if self.min_width is None
+                else min(self.min_width - extra_width, max_width)
+            )
+            pad_widths = ratio_distribute(_max_width - table_width, widths)
+            widths = [_width + pad for _width, pad in zip(widths, pad_widths)]
+
+        return widths
+
+    @classmethod
+    def _collapse_widths(
+        cls, widths: List[int], wrapable: List[bool], max_width: int
+    ) -> List[int]:
+        """Reduce widths so that the total is under max_width.
+
+        Args:
+            widths (List[int]): List of widths.
+            wrapable (List[bool]): List of booleans that indicate if a column may shrink.
+            max_width (int): Maximum width to reduce to.
+
+        Returns:
+            List[int]: A new list of widths.
+        """
+        total_width = sum(widths)
+        excess_width = total_width - max_width
+        if any(wrapable):
+            while total_width and excess_width > 0:
+                max_column = max(
+                    width for width, allow_wrap in zip(widths, wrapable) if allow_wrap
+                )
+                second_max_column = max(
+                    width if allow_wrap and width != max_column else 0
+                    for width, allow_wrap in zip(widths, wrapable)
+                )
+                column_difference = max_column - second_max_column
+                ratios = [
+                    (1 if (width == max_column and allow_wrap) else 0)
+                    for width, allow_wrap in zip(widths, wrapable)
+                ]
+                if not any(ratios) or not column_difference:
+                    break
+                max_reduce = [min(excess_width, column_difference)] * len(widths)
+                widths = ratio_reduce(excess_width, ratios, max_reduce, widths)
+
+                total_width = sum(widths)
+                excess_width = total_width - max_width
+        return widths
+
+    def _get_cells(
+        self, console: "Console", column_index: int, column: Column
+    ) -> Iterable[_Cell]:
+        """Get all the cells with padding and optional header."""
+
+        collapse_padding = self.collapse_padding
+        pad_edge = self.pad_edge
+        padding = self.padding
+        any_padding = any(padding)
+
+        first_column = column_index == 0
+        last_column = column_index == len(self.columns) - 1
+
+        _padding_cache: Dict[Tuple[bool, bool], Tuple[int, int, int, int]] = {}
+
+        def get_padding(first_row: bool, last_row: bool) -> Tuple[int, int, int, int]:
+            cached = _padding_cache.get((first_row, last_row))
+            if cached:
+                return cached
+            top, right, bottom, left = padding
+
+            if collapse_padding:
+                if not first_column:
+                    left = max(0, left - right)
+                if not last_row:
+                    bottom = max(0, top - bottom)
+
+            if not pad_edge:
+                if first_column:
+                    left = 0
+                if last_column:
+                    right = 0
+                if first_row:
+                    top = 0
+                if last_row:
+                    bottom = 0
+            _padding = (top, right, bottom, left)
+            _padding_cache[(first_row, last_row)] = _padding
+            return _padding
+
+        raw_cells: List[Tuple[StyleType, "RenderableType"]] = []
+        _append = raw_cells.append
+        get_style = console.get_style
+        if self.show_header:
+            header_style = get_style(self.header_style or "") + get_style(
+                column.header_style
+            )
+            _append((header_style, column.header))
+        cell_style = get_style(column.style or "")
+        for cell in column.cells:
+            _append((cell_style, cell))
+        if self.show_footer:
+            footer_style = get_style(self.footer_style or "") + get_style(
+                column.footer_style
+            )
+            _append((footer_style, column.footer))
+
+        if any_padding:
+            _Padding = Padding
+            for first, last, (style, renderable) in loop_first_last(raw_cells):
+                yield _Cell(
+                    style,
+                    _Padding(renderable, get_padding(first, last)),
+                    getattr(renderable, "vertical", None) or column.vertical,
+                )
+        else:
+            for style, renderable in raw_cells:
+                yield _Cell(
+                    style,
+                    renderable,
+                    getattr(renderable, "vertical", None) or column.vertical,
+                )
+
+    def _get_padding_width(self, column_index: int) -> int:
+        """Get extra width from padding."""
+        _, pad_right, _, pad_left = self.padding
+        if self.collapse_padding:
+            if column_index > 0:
+                pad_left = max(0, pad_left - pad_right)
+        return pad_left + pad_right
+
+    def _measure_column(
+        self,
+        console: "Console",
+        options: "ConsoleOptions",
+        column: Column,
+    ) -> Measurement:
+        """Get the minimum and maximum width of the column."""
+
+        max_width = options.max_width
+        if max_width < 1:
+            return Measurement(0, 0)
+
+        padding_width = self._get_padding_width(column._index)
+
+        if column.width is not None:
+            # Fixed width column
+            return Measurement(
+                column.width + padding_width, column.width + padding_width
+            ).with_maximum(max_width)
+        # Flexible column, we need to measure contents
+        min_widths: List[int] = []
+        max_widths: List[int] = []
+        append_min = min_widths.append
+        append_max = max_widths.append
+        get_render_width = Measurement.get
+        for cell in self._get_cells(console, column._index, column):
+            _min, _max = get_render_width(console, options, cell.renderable)
+            append_min(_min)
+            append_max(_max)
+
+        measurement = Measurement(
+            max(min_widths) if min_widths else 1,
+            max(max_widths) if max_widths else max_width,
+        ).with_maximum(max_width)
+        measurement = measurement.clamp(
+            None if column.min_width is None else column.min_width + padding_width,
+            None if column.max_width is None else column.max_width + padding_width,
+        )
+        return measurement
+
+    def _render(
+        self, console: "Console", options: "ConsoleOptions", widths: List[int]
+    ) -> "RenderResult":
+        table_style = console.get_style(self.style or "")
+
+        border_style = table_style + console.get_style(self.border_style or "")
+        _column_cells = (
+            self._get_cells(console, column_index, column)
+            for column_index, column in enumerate(self.columns)
+        )
+        row_cells: List[Tuple[_Cell, ...]] = list(zip(*_column_cells))
+        _box = (
+            self.box.substitute(
+                options, safe=pick_bool(self.safe_box, console.safe_box)
+            )
+            if self.box
+            else None
+        )
+        _box = _box.get_plain_headed_box() if _box and not self.show_header else _box
+
+        new_line = Segment.line()
+
+        columns = self.columns
+        show_header = self.show_header
+        show_footer = self.show_footer
+        show_edge = self.show_edge
+        show_lines = self.show_lines
+        leading = self.leading
+
+        _Segment = Segment
+        if _box:
+            box_segments = [
+                (
+                    _Segment(_box.head_left, border_style),
+                    _Segment(_box.head_right, border_style),
+                    _Segment(_box.head_vertical, border_style),
+                ),
+                (
+                    _Segment(_box.mid_left, border_style),
+                    _Segment(_box.mid_right, border_style),
+                    _Segment(_box.mid_vertical, border_style),
+                ),
+                (
+                    _Segment(_box.foot_left, border_style),
+                    _Segment(_box.foot_right, border_style),
+                    _Segment(_box.foot_vertical, border_style),
+                ),
+            ]
+            if show_edge:
+                yield _Segment(_box.get_top(widths), border_style)
+                yield new_line
+        else:
+            box_segments = []
+
+        get_row_style = self.get_row_style
+        get_style = console.get_style
+
+        for index, (first, last, row_cell) in enumerate(loop_first_last(row_cells)):
+            header_row = first and show_header
+            footer_row = last and show_footer
+            row = (
+                self.rows[index - show_header]
+                if (not header_row and not footer_row)
+                else None
+            )
+            max_height = 1
+            cells: List[List[List[Segment]]] = []
+            if header_row or footer_row:
+                row_style = Style.null()
+            else:
+                row_style = get_style(
+                    get_row_style(console, index - 1 if show_header else index)
+                )
+            for width, cell, column in zip(widths, row_cell, columns):
+                render_options = options.update(
+                    width=width,
+                    justify=column.justify,
+                    no_wrap=column.no_wrap,
+                    overflow=column.overflow,
+                    height=None,
+                    highlight=column.highlight,
+                )
+                lines = console.render_lines(
+                    cell.renderable,
+                    render_options,
+                    style=get_style(cell.style) + row_style,
+                )
+                max_height = max(max_height, len(lines))
+                cells.append(lines)
+
+            row_height = max(len(cell) for cell in cells)
+
+            def align_cell(
+                cell: List[List[Segment]],
+                vertical: "VerticalAlignMethod",
+                width: int,
+                style: Style,
+            ) -> List[List[Segment]]:
+                if header_row:
+                    vertical = "bottom"
+                elif footer_row:
+                    vertical = "top"
+
+                if vertical == "top":
+                    return _Segment.align_top(cell, width, row_height, style)
+                elif vertical == "middle":
+                    return _Segment.align_middle(cell, width, row_height, style)
+                return _Segment.align_bottom(cell, width, row_height, style)
+
+            cells[:] = [
+                _Segment.set_shape(
+                    align_cell(
+                        cell,
+                        _cell.vertical,
+                        width,
+                        get_style(_cell.style) + row_style,
+                    ),
+                    width,
+                    max_height,
+                )
+                for width, _cell, cell, column in zip(widths, row_cell, cells, columns)
+            ]
+
+            if _box:
+                if last and show_footer:
+                    yield _Segment(
+                        _box.get_row(widths, "foot", edge=show_edge), border_style
+                    )
+                    yield new_line
+                left, right, _divider = box_segments[0 if first else (2 if last else 1)]
+
+                # If the column divider is whitespace also style it with the row background
+                divider = (
+                    _divider
+                    if _divider.text.strip()
+                    else _Segment(
+                        _divider.text, row_style.background_style + _divider.style
+                    )
+                )
+                for line_no in range(max_height):
+                    if show_edge:
+                        yield left
+                    for last_cell, rendered_cell in loop_last(cells):
+                        yield from rendered_cell[line_no]
+                        if not last_cell:
+                            yield divider
+                    if show_edge:
+                        yield right
+                    yield new_line
+            else:
+                for line_no in range(max_height):
+                    for rendered_cell in cells:
+                        yield from rendered_cell[line_no]
+                    yield new_line
+            if _box and first and show_header:
+                yield _Segment(
+                    _box.get_row(widths, "head", edge=show_edge), border_style
+                )
+                yield new_line
+            end_section = row and row.end_section
+            if _box and (show_lines or leading or end_section):
+                if (
+                    not last
+                    and not (show_footer and index >= len(row_cells) - 2)
+                    and not (show_header and header_row)
+                ):
+                    if leading:
+                        yield _Segment(
+                            _box.get_row(widths, "mid", edge=show_edge) * leading,
+                            border_style,
+                        )
+                    else:
+                        yield _Segment(
+                            _box.get_row(widths, "row", edge=show_edge), border_style
+                        )
+                    yield new_line
+
+        if _box and show_edge:
+            yield _Segment(_box.get_bottom(widths), border_style)
+            yield new_line
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console
+    from pip._vendor.rich.highlighter import ReprHighlighter
+    from pip._vendor.rich.table import Table as Table
+
+    from ._timer import timer
+
+    with timer("Table render"):
+        table = Table(
+            title="Star Wars Movies",
+            caption="Rich example table",
+            caption_justify="right",
+        )
+
+        table.add_column(
+            "Released", header_style="bright_cyan", style="cyan", no_wrap=True
+        )
+        table.add_column("Title", style="magenta")
+        table.add_column("Box Office", justify="right", style="green")
+
+        table.add_row(
+            "Dec 20, 2019",
+            "Star Wars: The Rise of Skywalker",
+            "$952,110,690",
+        )
+        table.add_row("May 25, 2018", "Solo: A Star Wars Story", "$393,151,347")
+        table.add_row(
+            "Dec 15, 2017",
+            "Star Wars Ep. V111: The Last Jedi",
+            "$1,332,539,889",
+            style="on black",
+            end_section=True,
+        )
+        table.add_row(
+            "Dec 16, 2016",
+            "Rogue One: A Star Wars Story",
+            "$1,332,439,889",
+        )
+
+        def header(text: str) -> None:
+            console.print()
+            console.rule(highlight(text))
+            console.print()
+
+        console = Console()
+        highlight = ReprHighlighter()
+        header("Example Table")
+        console.print(table, justify="center")
+
+        table.expand = True
+        header("expand=True")
+        console.print(table)
+
+        table.width = 50
+        header("width=50")
+
+        console.print(table, justify="center")
+
+        table.width = None
+        table.expand = False
+        table.row_styles = ["dim", "none"]
+        header("row_styles=['dim', 'none']")
+
+        console.print(table, justify="center")
+
+        table.width = None
+        table.expand = False
+        table.row_styles = ["dim", "none"]
+        table.leading = 1
+        header("leading=1, row_styles=['dim', 'none']")
+        console.print(table, justify="center")
+
+        table.width = None
+        table.expand = False
+        table.row_styles = ["dim", "none"]
+        table.show_lines = True
+        table.leading = 0
+        header("show_lines=True, row_styles=['dim', 'none']")
+        console.print(table, justify="center")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/terminal_theme.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/terminal_theme.py
new file mode 100644
index 00000000..565e9d96
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/terminal_theme.py
@@ -0,0 +1,153 @@
+from typing import List, Optional, Tuple
+
+from .color_triplet import ColorTriplet
+from .palette import Palette
+
+_ColorTuple = Tuple[int, int, int]
+
+
+class TerminalTheme:
+    """A color theme used when exporting console content.
+
+    Args:
+        background (Tuple[int, int, int]): The background color.
+        foreground (Tuple[int, int, int]): The foreground (text) color.
+        normal (List[Tuple[int, int, int]]): A list of 8 normal intensity colors.
+        bright (List[Tuple[int, int, int]], optional): A list of 8 bright colors, or None
+            to repeat normal intensity. Defaults to None.
+    """
+
+    def __init__(
+        self,
+        background: _ColorTuple,
+        foreground: _ColorTuple,
+        normal: List[_ColorTuple],
+        bright: Optional[List[_ColorTuple]] = None,
+    ) -> None:
+        self.background_color = ColorTriplet(*background)
+        self.foreground_color = ColorTriplet(*foreground)
+        self.ansi_colors = Palette(normal + (bright or normal))
+
+
+DEFAULT_TERMINAL_THEME = TerminalTheme(
+    (255, 255, 255),
+    (0, 0, 0),
+    [
+        (0, 0, 0),
+        (128, 0, 0),
+        (0, 128, 0),
+        (128, 128, 0),
+        (0, 0, 128),
+        (128, 0, 128),
+        (0, 128, 128),
+        (192, 192, 192),
+    ],
+    [
+        (128, 128, 128),
+        (255, 0, 0),
+        (0, 255, 0),
+        (255, 255, 0),
+        (0, 0, 255),
+        (255, 0, 255),
+        (0, 255, 255),
+        (255, 255, 255),
+    ],
+)
+
+MONOKAI = TerminalTheme(
+    (12, 12, 12),
+    (217, 217, 217),
+    [
+        (26, 26, 26),
+        (244, 0, 95),
+        (152, 224, 36),
+        (253, 151, 31),
+        (157, 101, 255),
+        (244, 0, 95),
+        (88, 209, 235),
+        (196, 197, 181),
+        (98, 94, 76),
+    ],
+    [
+        (244, 0, 95),
+        (152, 224, 36),
+        (224, 213, 97),
+        (157, 101, 255),
+        (244, 0, 95),
+        (88, 209, 235),
+        (246, 246, 239),
+    ],
+)
+DIMMED_MONOKAI = TerminalTheme(
+    (25, 25, 25),
+    (185, 188, 186),
+    [
+        (58, 61, 67),
+        (190, 63, 72),
+        (135, 154, 59),
+        (197, 166, 53),
+        (79, 118, 161),
+        (133, 92, 141),
+        (87, 143, 164),
+        (185, 188, 186),
+        (136, 137, 135),
+    ],
+    [
+        (251, 0, 31),
+        (15, 114, 47),
+        (196, 112, 51),
+        (24, 109, 227),
+        (251, 0, 103),
+        (46, 112, 109),
+        (253, 255, 185),
+    ],
+)
+NIGHT_OWLISH = TerminalTheme(
+    (255, 255, 255),
+    (64, 63, 83),
+    [
+        (1, 22, 39),
+        (211, 66, 62),
+        (42, 162, 152),
+        (218, 170, 1),
+        (72, 118, 214),
+        (64, 63, 83),
+        (8, 145, 106),
+        (122, 129, 129),
+        (122, 129, 129),
+    ],
+    [
+        (247, 110, 110),
+        (73, 208, 197),
+        (218, 194, 107),
+        (92, 167, 228),
+        (105, 112, 152),
+        (0, 201, 144),
+        (152, 159, 177),
+    ],
+)
+
+SVG_EXPORT_THEME = TerminalTheme(
+    (41, 41, 41),
+    (197, 200, 198),
+    [
+        (75, 78, 85),
+        (204, 85, 90),
+        (152, 168, 75),
+        (208, 179, 68),
+        (96, 138, 177),
+        (152, 114, 159),
+        (104, 160, 179),
+        (197, 200, 198),
+        (154, 155, 153),
+    ],
+    [
+        (255, 38, 39),
+        (0, 130, 61),
+        (208, 132, 66),
+        (25, 132, 233),
+        (255, 44, 122),
+        (57, 130, 128),
+        (253, 253, 197),
+    ],
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/text.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/text.py
new file mode 100644
index 00000000..5a0c6b14
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/text.py
@@ -0,0 +1,1361 @@
+import re
+from functools import partial, reduce
+from math import gcd
+from operator import itemgetter
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    NamedTuple,
+    Optional,
+    Pattern,
+    Tuple,
+    Union,
+)
+
+from ._loop import loop_last
+from ._pick import pick_bool
+from ._wrap import divide_line
+from .align import AlignMethod
+from .cells import cell_len, set_cell_size
+from .containers import Lines
+from .control import strip_control_codes
+from .emoji import EmojiVariant
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import Style, StyleType
+
+if TYPE_CHECKING:  # pragma: no cover
+    from .console import Console, ConsoleOptions, JustifyMethod, OverflowMethod
+
+DEFAULT_JUSTIFY: "JustifyMethod" = "default"
+DEFAULT_OVERFLOW: "OverflowMethod" = "fold"
+
+
+_re_whitespace = re.compile(r"\s+$")
+
+TextType = Union[str, "Text"]
+"""A plain string or a :class:`Text` instance."""
+
+GetStyleCallable = Callable[[str], Optional[StyleType]]
+
+
+class Span(NamedTuple):
+    """A marked up region in some text."""
+
+    start: int
+    """Span start index."""
+    end: int
+    """Span end index."""
+    style: Union[str, Style]
+    """Style associated with the span."""
+
+    def __repr__(self) -> str:
+        return f"Span({self.start}, {self.end}, {self.style!r})"
+
+    def __bool__(self) -> bool:
+        return self.end > self.start
+
+    def split(self, offset: int) -> Tuple["Span", Optional["Span"]]:
+        """Split a span in to 2 from a given offset."""
+
+        if offset < self.start:
+            return self, None
+        if offset >= self.end:
+            return self, None
+
+        start, end, style = self
+        span1 = Span(start, min(end, offset), style)
+        span2 = Span(span1.end, end, style)
+        return span1, span2
+
+    def move(self, offset: int) -> "Span":
+        """Move start and end by a given offset.
+
+        Args:
+            offset (int): Number of characters to add to start and end.
+
+        Returns:
+            TextSpan: A new TextSpan with adjusted position.
+        """
+        start, end, style = self
+        return Span(start + offset, end + offset, style)
+
+    def right_crop(self, offset: int) -> "Span":
+        """Crop the span at the given offset.
+
+        Args:
+            offset (int): A value between start and end.
+
+        Returns:
+            Span: A new (possibly smaller) span.
+        """
+        start, end, style = self
+        if offset >= end:
+            return self
+        return Span(start, min(offset, end), style)
+
+    def extend(self, cells: int) -> "Span":
+        """Extend the span by the given number of cells.
+
+        Args:
+            cells (int): Additional space to add to end of span.
+
+        Returns:
+            Span: A span.
+        """
+        if cells:
+            start, end, style = self
+            return Span(start, end + cells, style)
+        else:
+            return self
+
+
+class Text(JupyterMixin):
+    """Text with color / style.
+
+    Args:
+        text (str, optional): Default unstyled text. Defaults to "".
+        style (Union[str, Style], optional): Base style for text. Defaults to "".
+        justify (str, optional): Justify method: "left", "center", "full", "right". Defaults to None.
+        overflow (str, optional): Overflow method: "crop", "fold", "ellipsis". Defaults to None.
+        no_wrap (bool, optional): Disable text wrapping, or None for default. Defaults to None.
+        end (str, optional): Character to end text with. Defaults to "\\\\n".
+        tab_size (int): Number of spaces per tab, or ``None`` to use ``console.tab_size``. Defaults to None.
+        spans (List[Span], optional). A list of predefined style spans. Defaults to None.
+    """
+
+    __slots__ = [
+        "_text",
+        "style",
+        "justify",
+        "overflow",
+        "no_wrap",
+        "end",
+        "tab_size",
+        "_spans",
+        "_length",
+    ]
+
+    def __init__(
+        self,
+        text: str = "",
+        style: Union[str, Style] = "",
+        *,
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        no_wrap: Optional[bool] = None,
+        end: str = "\n",
+        tab_size: Optional[int] = None,
+        spans: Optional[List[Span]] = None,
+    ) -> None:
+        sanitized_text = strip_control_codes(text)
+        self._text = [sanitized_text]
+        self.style = style
+        self.justify: Optional["JustifyMethod"] = justify
+        self.overflow: Optional["OverflowMethod"] = overflow
+        self.no_wrap = no_wrap
+        self.end = end
+        self.tab_size = tab_size
+        self._spans: List[Span] = spans or []
+        self._length: int = len(sanitized_text)
+
+    def __len__(self) -> int:
+        return self._length
+
+    def __bool__(self) -> bool:
+        return bool(self._length)
+
+    def __str__(self) -> str:
+        return self.plain
+
+    def __repr__(self) -> str:
+        return f"<text {self.plain!r} {self._spans!r} {self.style!r}>"
+
+    def __add__(self, other: Any) -> "Text":
+        if isinstance(other, (str, Text)):
+            result = self.copy()
+            result.append(other)
+            return result
+        return NotImplemented
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Text):
+            return NotImplemented
+        return self.plain == other.plain and self._spans == other._spans
+
+    def __contains__(self, other: object) -> bool:
+        if isinstance(other, str):
+            return other in self.plain
+        elif isinstance(other, Text):
+            return other.plain in self.plain
+        return False
+
+    def __getitem__(self, slice: Union[int, slice]) -> "Text":
+        def get_text_at(offset: int) -> "Text":
+            _Span = Span
+            text = Text(
+                self.plain[offset],
+                spans=[
+                    _Span(0, 1, style)
+                    for start, end, style in self._spans
+                    if end > offset >= start
+                ],
+                end="",
+            )
+            return text
+
+        if isinstance(slice, int):
+            return get_text_at(slice)
+        else:
+            start, stop, step = slice.indices(len(self.plain))
+            if step == 1:
+                lines = self.divide([start, stop])
+                return lines[1]
+            else:
+                # This would be a bit of work to implement efficiently
+                # For now, its not required
+                raise TypeError("slices with step!=1 are not supported")
+
+    @property
+    def cell_len(self) -> int:
+        """Get the number of cells required to render this text."""
+        return cell_len(self.plain)
+
+    @property
+    def markup(self) -> str:
+        """Get console markup to render this Text.
+
+        Returns:
+            str: A string potentially creating markup tags.
+        """
+        from .markup import escape
+
+        output: List[str] = []
+
+        plain = self.plain
+        markup_spans = [
+            (0, False, self.style),
+            *((span.start, False, span.style) for span in self._spans),
+            *((span.end, True, span.style) for span in self._spans),
+            (len(plain), True, self.style),
+        ]
+        markup_spans.sort(key=itemgetter(0, 1))
+        position = 0
+        append = output.append
+        for offset, closing, style in markup_spans:
+            if offset > position:
+                append(escape(plain[position:offset]))
+                position = offset
+            if style:
+                append(f"[/{style}]" if closing else f"[{style}]")
+        markup = "".join(output)
+        return markup
+
+    @classmethod
+    def from_markup(
+        cls,
+        text: str,
+        *,
+        style: Union[str, Style] = "",
+        emoji: bool = True,
+        emoji_variant: Optional[EmojiVariant] = None,
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        end: str = "\n",
+    ) -> "Text":
+        """Create Text instance from markup.
+
+        Args:
+            text (str): A string containing console markup.
+            style (Union[str, Style], optional): Base style for text. Defaults to "".
+            emoji (bool, optional): Also render emoji code. Defaults to True.
+            emoji_variant (str, optional): Optional emoji variant, either "text" or "emoji". Defaults to None.
+            justify (str, optional): Justify method: "left", "center", "full", "right". Defaults to None.
+            overflow (str, optional): Overflow method: "crop", "fold", "ellipsis". Defaults to None.
+            end (str, optional): Character to end text with. Defaults to "\\\\n".
+
+        Returns:
+            Text: A Text instance with markup rendered.
+        """
+        from .markup import render
+
+        rendered_text = render(text, style, emoji=emoji, emoji_variant=emoji_variant)
+        rendered_text.justify = justify
+        rendered_text.overflow = overflow
+        rendered_text.end = end
+        return rendered_text
+
+    @classmethod
+    def from_ansi(
+        cls,
+        text: str,
+        *,
+        style: Union[str, Style] = "",
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        no_wrap: Optional[bool] = None,
+        end: str = "\n",
+        tab_size: Optional[int] = 8,
+    ) -> "Text":
+        """Create a Text object from a string containing ANSI escape codes.
+
+        Args:
+            text (str): A string containing escape codes.
+            style (Union[str, Style], optional): Base style for text. Defaults to "".
+            justify (str, optional): Justify method: "left", "center", "full", "right". Defaults to None.
+            overflow (str, optional): Overflow method: "crop", "fold", "ellipsis". Defaults to None.
+            no_wrap (bool, optional): Disable text wrapping, or None for default. Defaults to None.
+            end (str, optional): Character to end text with. Defaults to "\\\\n".
+            tab_size (int): Number of spaces per tab, or ``None`` to use ``console.tab_size``. Defaults to None.
+        """
+        from .ansi import AnsiDecoder
+
+        joiner = Text(
+            "\n",
+            justify=justify,
+            overflow=overflow,
+            no_wrap=no_wrap,
+            end=end,
+            tab_size=tab_size,
+            style=style,
+        )
+        decoder = AnsiDecoder()
+        result = joiner.join(line for line in decoder.decode(text))
+        return result
+
+    @classmethod
+    def styled(
+        cls,
+        text: str,
+        style: StyleType = "",
+        *,
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+    ) -> "Text":
+        """Construct a Text instance with a pre-applied styled. A style applied in this way won't be used
+        to pad the text when it is justified.
+
+        Args:
+            text (str): A string containing console markup.
+            style (Union[str, Style]): Style to apply to the text. Defaults to "".
+            justify (str, optional): Justify method: "left", "center", "full", "right". Defaults to None.
+            overflow (str, optional): Overflow method: "crop", "fold", "ellipsis". Defaults to None.
+
+        Returns:
+            Text: A text instance with a style applied to the entire string.
+        """
+        styled_text = cls(text, justify=justify, overflow=overflow)
+        styled_text.stylize(style)
+        return styled_text
+
+    @classmethod
+    def assemble(
+        cls,
+        *parts: Union[str, "Text", Tuple[str, StyleType]],
+        style: Union[str, Style] = "",
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        no_wrap: Optional[bool] = None,
+        end: str = "\n",
+        tab_size: int = 8,
+        meta: Optional[Dict[str, Any]] = None,
+    ) -> "Text":
+        """Construct a text instance by combining a sequence of strings with optional styles.
+        The positional arguments should be either strings, or a tuple of string + style.
+
+        Args:
+            style (Union[str, Style], optional): Base style for text. Defaults to "".
+            justify (str, optional): Justify method: "left", "center", "full", "right". Defaults to None.
+            overflow (str, optional): Overflow method: "crop", "fold", "ellipsis". Defaults to None.
+            no_wrap (bool, optional): Disable text wrapping, or None for default. Defaults to None.
+            end (str, optional): Character to end text with. Defaults to "\\\\n".
+            tab_size (int): Number of spaces per tab, or ``None`` to use ``console.tab_size``. Defaults to None.
+            meta (Dict[str, Any], optional). Meta data to apply to text, or None for no meta data. Default to None
+
+        Returns:
+            Text: A new text instance.
+        """
+        text = cls(
+            style=style,
+            justify=justify,
+            overflow=overflow,
+            no_wrap=no_wrap,
+            end=end,
+            tab_size=tab_size,
+        )
+        append = text.append
+        _Text = Text
+        for part in parts:
+            if isinstance(part, (_Text, str)):
+                append(part)
+            else:
+                append(*part)
+        if meta:
+            text.apply_meta(meta)
+        return text
+
+    @property
+    def plain(self) -> str:
+        """Get the text as a single string."""
+        if len(self._text) != 1:
+            self._text[:] = ["".join(self._text)]
+        return self._text[0]
+
+    @plain.setter
+    def plain(self, new_text: str) -> None:
+        """Set the text to a new value."""
+        if new_text != self.plain:
+            sanitized_text = strip_control_codes(new_text)
+            self._text[:] = [sanitized_text]
+            old_length = self._length
+            self._length = len(sanitized_text)
+            if old_length > self._length:
+                self._trim_spans()
+
+    @property
+    def spans(self) -> List[Span]:
+        """Get a reference to the internal list of spans."""
+        return self._spans
+
+    @spans.setter
+    def spans(self, spans: List[Span]) -> None:
+        """Set spans."""
+        self._spans = spans[:]
+
+    def blank_copy(self, plain: str = "") -> "Text":
+        """Return a new Text instance with copied metadata (but not the string or spans)."""
+        copy_self = Text(
+            plain,
+            style=self.style,
+            justify=self.justify,
+            overflow=self.overflow,
+            no_wrap=self.no_wrap,
+            end=self.end,
+            tab_size=self.tab_size,
+        )
+        return copy_self
+
+    def copy(self) -> "Text":
+        """Return a copy of this instance."""
+        copy_self = Text(
+            self.plain,
+            style=self.style,
+            justify=self.justify,
+            overflow=self.overflow,
+            no_wrap=self.no_wrap,
+            end=self.end,
+            tab_size=self.tab_size,
+        )
+        copy_self._spans[:] = self._spans
+        return copy_self
+
+    def stylize(
+        self,
+        style: Union[str, Style],
+        start: int = 0,
+        end: Optional[int] = None,
+    ) -> None:
+        """Apply a style to the text, or a portion of the text.
+
+        Args:
+            style (Union[str, Style]): Style instance or style definition to apply.
+            start (int): Start offset (negative indexing is supported). Defaults to 0.
+            end (Optional[int], optional): End offset (negative indexing is supported), or None for end of text. Defaults to None.
+        """
+        if style:
+            length = len(self)
+            if start < 0:
+                start = length + start
+            if end is None:
+                end = length
+            if end < 0:
+                end = length + end
+            if start >= length or end <= start:
+                # Span not in text or not valid
+                return
+            self._spans.append(Span(start, min(length, end), style))
+
+    def stylize_before(
+        self,
+        style: Union[str, Style],
+        start: int = 0,
+        end: Optional[int] = None,
+    ) -> None:
+        """Apply a style to the text, or a portion of the text. Styles will be applied before other styles already present.
+
+        Args:
+            style (Union[str, Style]): Style instance or style definition to apply.
+            start (int): Start offset (negative indexing is supported). Defaults to 0.
+            end (Optional[int], optional): End offset (negative indexing is supported), or None for end of text. Defaults to None.
+        """
+        if style:
+            length = len(self)
+            if start < 0:
+                start = length + start
+            if end is None:
+                end = length
+            if end < 0:
+                end = length + end
+            if start >= length or end <= start:
+                # Span not in text or not valid
+                return
+            self._spans.insert(0, Span(start, min(length, end), style))
+
+    def apply_meta(
+        self, meta: Dict[str, Any], start: int = 0, end: Optional[int] = None
+    ) -> None:
+        """Apply metadata to the text, or a portion of the text.
+
+        Args:
+            meta (Dict[str, Any]): A dict of meta information.
+            start (int): Start offset (negative indexing is supported). Defaults to 0.
+            end (Optional[int], optional): End offset (negative indexing is supported), or None for end of text. Defaults to None.
+
+        """
+        style = Style.from_meta(meta)
+        self.stylize(style, start=start, end=end)
+
+    def on(self, meta: Optional[Dict[str, Any]] = None, **handlers: Any) -> "Text":
+        """Apply event handlers (used by Textual project).
+
+        Example:
+            >>> from rich.text import Text
+            >>> text = Text("hello world")
+            >>> text.on(click="view.toggle('world')")
+
+        Args:
+            meta (Dict[str, Any]): Mapping of meta information.
+            **handlers: Keyword args are prefixed with "@" to defined handlers.
+
+        Returns:
+            Text: Self is returned to method may be chained.
+        """
+        meta = {} if meta is None else meta
+        meta.update({f"@{key}": value for key, value in handlers.items()})
+        self.stylize(Style.from_meta(meta))
+        return self
+
+    def remove_suffix(self, suffix: str) -> None:
+        """Remove a suffix if it exists.
+
+        Args:
+            suffix (str): Suffix to remove.
+        """
+        if self.plain.endswith(suffix):
+            self.right_crop(len(suffix))
+
+    def get_style_at_offset(self, console: "Console", offset: int) -> Style:
+        """Get the style of a character at give offset.
+
+        Args:
+            console (~Console): Console where text will be rendered.
+            offset (int): Offset in to text (negative indexing supported)
+
+        Returns:
+            Style: A Style instance.
+        """
+        # TODO: This is a little inefficient, it is only used by full justify
+        if offset < 0:
+            offset = len(self) + offset
+        get_style = console.get_style
+        style = get_style(self.style).copy()
+        for start, end, span_style in self._spans:
+            if end > offset >= start:
+                style += get_style(span_style, default="")
+        return style
+
+    def extend_style(self, spaces: int) -> None:
+        """Extend the Text given number of spaces where the spaces have the same style as the last character.
+
+        Args:
+            spaces (int): Number of spaces to add to the Text.
+        """
+        if spaces <= 0:
+            return
+        spans = self.spans
+        new_spaces = " " * spaces
+        if spans:
+            end_offset = len(self)
+            self._spans[:] = [
+                span.extend(spaces) if span.end >= end_offset else span
+                for span in spans
+            ]
+            self._text.append(new_spaces)
+            self._length += spaces
+        else:
+            self.plain += new_spaces
+
+    def highlight_regex(
+        self,
+        re_highlight: Union[Pattern[str], str],
+        style: Optional[Union[GetStyleCallable, StyleType]] = None,
+        *,
+        style_prefix: str = "",
+    ) -> int:
+        """Highlight text with a regular expression, where group names are
+        translated to styles.
+
+        Args:
+            re_highlight (Union[re.Pattern, str]): A regular expression object or string.
+            style (Union[GetStyleCallable, StyleType]): Optional style to apply to whole match, or a callable
+                which accepts the matched text and returns a style. Defaults to None.
+            style_prefix (str, optional): Optional prefix to add to style group names.
+
+        Returns:
+            int: Number of regex matches
+        """
+        count = 0
+        append_span = self._spans.append
+        _Span = Span
+        plain = self.plain
+        if isinstance(re_highlight, str):
+            re_highlight = re.compile(re_highlight)
+        for match in re_highlight.finditer(plain):
+            get_span = match.span
+            if style:
+                start, end = get_span()
+                match_style = style(plain[start:end]) if callable(style) else style
+                if match_style is not None and end > start:
+                    append_span(_Span(start, end, match_style))
+
+            count += 1
+            for name in match.groupdict().keys():
+                start, end = get_span(name)
+                if start != -1 and end > start:
+                    append_span(_Span(start, end, f"{style_prefix}{name}"))
+        return count
+
+    def highlight_words(
+        self,
+        words: Iterable[str],
+        style: Union[str, Style],
+        *,
+        case_sensitive: bool = True,
+    ) -> int:
+        """Highlight words with a style.
+
+        Args:
+            words (Iterable[str]): Words to highlight.
+            style (Union[str, Style]): Style to apply.
+            case_sensitive (bool, optional): Enable case sensitive matching. Defaults to True.
+
+        Returns:
+            int: Number of words highlighted.
+        """
+        re_words = "|".join(re.escape(word) for word in words)
+        add_span = self._spans.append
+        count = 0
+        _Span = Span
+        for match in re.finditer(
+            re_words, self.plain, flags=0 if case_sensitive else re.IGNORECASE
+        ):
+            start, end = match.span(0)
+            add_span(_Span(start, end, style))
+            count += 1
+        return count
+
+    def rstrip(self) -> None:
+        """Strip whitespace from end of text."""
+        self.plain = self.plain.rstrip()
+
+    def rstrip_end(self, size: int) -> None:
+        """Remove whitespace beyond a certain width at the end of the text.
+
+        Args:
+            size (int): The desired size of the text.
+        """
+        text_length = len(self)
+        if text_length > size:
+            excess = text_length - size
+            whitespace_match = _re_whitespace.search(self.plain)
+            if whitespace_match is not None:
+                whitespace_count = len(whitespace_match.group(0))
+                self.right_crop(min(whitespace_count, excess))
+
+    def set_length(self, new_length: int) -> None:
+        """Set new length of the text, clipping or padding is required."""
+        length = len(self)
+        if length != new_length:
+            if length < new_length:
+                self.pad_right(new_length - length)
+            else:
+                self.right_crop(length - new_length)
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Iterable[Segment]:
+        tab_size: int = console.tab_size if self.tab_size is None else self.tab_size
+        justify = self.justify or options.justify or DEFAULT_JUSTIFY
+
+        overflow = self.overflow or options.overflow or DEFAULT_OVERFLOW
+
+        lines = self.wrap(
+            console,
+            options.max_width,
+            justify=justify,
+            overflow=overflow,
+            tab_size=tab_size or 8,
+            no_wrap=pick_bool(self.no_wrap, options.no_wrap, False),
+        )
+        all_lines = Text("\n").join(lines)
+        yield from all_lines.render(console, end=self.end)
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> Measurement:
+        text = self.plain
+        lines = text.splitlines()
+        max_text_width = max(cell_len(line) for line in lines) if lines else 0
+        words = text.split()
+        min_text_width = (
+            max(cell_len(word) for word in words) if words else max_text_width
+        )
+        return Measurement(min_text_width, max_text_width)
+
+    def render(self, console: "Console", end: str = "") -> Iterable["Segment"]:
+        """Render the text as Segments.
+
+        Args:
+            console (Console): Console instance.
+            end (Optional[str], optional): Optional end character.
+
+        Returns:
+            Iterable[Segment]: Result of render that may be written to the console.
+        """
+        _Segment = Segment
+        text = self.plain
+        if not self._spans:
+            yield Segment(text)
+            if end:
+                yield _Segment(end)
+            return
+        get_style = partial(console.get_style, default=Style.null())
+
+        enumerated_spans = list(enumerate(self._spans, 1))
+        style_map = {index: get_style(span.style) for index, span in enumerated_spans}
+        style_map[0] = get_style(self.style)
+
+        spans = [
+            (0, False, 0),
+            *((span.start, False, index) for index, span in enumerated_spans),
+            *((span.end, True, index) for index, span in enumerated_spans),
+            (len(text), True, 0),
+        ]
+        spans.sort(key=itemgetter(0, 1))
+
+        stack: List[int] = []
+        stack_append = stack.append
+        stack_pop = stack.remove
+
+        style_cache: Dict[Tuple[Style, ...], Style] = {}
+        style_cache_get = style_cache.get
+        combine = Style.combine
+
+        def get_current_style() -> Style:
+            """Construct current style from stack."""
+            styles = tuple(style_map[_style_id] for _style_id in sorted(stack))
+            cached_style = style_cache_get(styles)
+            if cached_style is not None:
+                return cached_style
+            current_style = combine(styles)
+            style_cache[styles] = current_style
+            return current_style
+
+        for (offset, leaving, style_id), (next_offset, _, _) in zip(spans, spans[1:]):
+            if leaving:
+                stack_pop(style_id)
+            else:
+                stack_append(style_id)
+            if next_offset > offset:
+                yield _Segment(text[offset:next_offset], get_current_style())
+        if end:
+            yield _Segment(end)
+
+    def join(self, lines: Iterable["Text"]) -> "Text":
+        """Join text together with this instance as the separator.
+
+        Args:
+            lines (Iterable[Text]): An iterable of Text instances to join.
+
+        Returns:
+            Text: A new text instance containing join text.
+        """
+
+        new_text = self.blank_copy()
+
+        def iter_text() -> Iterable["Text"]:
+            if self.plain:
+                for last, line in loop_last(lines):
+                    yield line
+                    if not last:
+                        yield self
+            else:
+                yield from lines
+
+        extend_text = new_text._text.extend
+        append_span = new_text._spans.append
+        extend_spans = new_text._spans.extend
+        offset = 0
+        _Span = Span
+
+        for text in iter_text():
+            extend_text(text._text)
+            if text.style:
+                append_span(_Span(offset, offset + len(text), text.style))
+            extend_spans(
+                _Span(offset + start, offset + end, style)
+                for start, end, style in text._spans
+            )
+            offset += len(text)
+        new_text._length = offset
+        return new_text
+
+    def expand_tabs(self, tab_size: Optional[int] = None) -> None:
+        """Converts tabs to spaces.
+
+        Args:
+            tab_size (int, optional): Size of tabs. Defaults to 8.
+
+        """
+        if "\t" not in self.plain:
+            return
+        if tab_size is None:
+            tab_size = self.tab_size
+        if tab_size is None:
+            tab_size = 8
+
+        new_text: List[Text] = []
+        append = new_text.append
+
+        for line in self.split("\n", include_separator=True):
+            if "\t" not in line.plain:
+                append(line)
+            else:
+                cell_position = 0
+                parts = line.split("\t", include_separator=True)
+                for part in parts:
+                    if part.plain.endswith("\t"):
+                        part._text[-1] = part._text[-1][:-1] + " "
+                        cell_position += part.cell_len
+                        tab_remainder = cell_position % tab_size
+                        if tab_remainder:
+                            spaces = tab_size - tab_remainder
+                            part.extend_style(spaces)
+                            cell_position += spaces
+                    else:
+                        cell_position += part.cell_len
+                    append(part)
+
+        result = Text("").join(new_text)
+
+        self._text = [result.plain]
+        self._length = len(self.plain)
+        self._spans[:] = result._spans
+
+    def truncate(
+        self,
+        max_width: int,
+        *,
+        overflow: Optional["OverflowMethod"] = None,
+        pad: bool = False,
+    ) -> None:
+        """Truncate text if it is longer that a given width.
+
+        Args:
+            max_width (int): Maximum number of characters in text.
+            overflow (str, optional): Overflow method: "crop", "fold", or "ellipsis". Defaults to None, to use self.overflow.
+            pad (bool, optional): Pad with spaces if the length is less than max_width. Defaults to False.
+        """
+        _overflow = overflow or self.overflow or DEFAULT_OVERFLOW
+        if _overflow != "ignore":
+            length = cell_len(self.plain)
+            if length > max_width:
+                if _overflow == "ellipsis":
+                    self.plain = set_cell_size(self.plain, max_width - 1) + "…"
+                else:
+                    self.plain = set_cell_size(self.plain, max_width)
+            if pad and length < max_width:
+                spaces = max_width - length
+                self._text = [f"{self.plain}{' ' * spaces}"]
+                self._length = len(self.plain)
+
+    def _trim_spans(self) -> None:
+        """Remove or modify any spans that are over the end of the text."""
+        max_offset = len(self.plain)
+        _Span = Span
+        self._spans[:] = [
+            (
+                span
+                if span.end < max_offset
+                else _Span(span.start, min(max_offset, span.end), span.style)
+            )
+            for span in self._spans
+            if span.start < max_offset
+        ]
+
+    def pad(self, count: int, character: str = " ") -> None:
+        """Pad left and right with a given number of characters.
+
+        Args:
+            count (int): Width of padding.
+            character (str): The character to pad with. Must be a string of length 1.
+        """
+        assert len(character) == 1, "Character must be a string of length 1"
+        if count:
+            pad_characters = character * count
+            self.plain = f"{pad_characters}{self.plain}{pad_characters}"
+            _Span = Span
+            self._spans[:] = [
+                _Span(start + count, end + count, style)
+                for start, end, style in self._spans
+            ]
+
+    def pad_left(self, count: int, character: str = " ") -> None:
+        """Pad the left with a given character.
+
+        Args:
+            count (int): Number of characters to pad.
+            character (str, optional): Character to pad with. Defaults to " ".
+        """
+        assert len(character) == 1, "Character must be a string of length 1"
+        if count:
+            self.plain = f"{character * count}{self.plain}"
+            _Span = Span
+            self._spans[:] = [
+                _Span(start + count, end + count, style)
+                for start, end, style in self._spans
+            ]
+
+    def pad_right(self, count: int, character: str = " ") -> None:
+        """Pad the right with a given character.
+
+        Args:
+            count (int): Number of characters to pad.
+            character (str, optional): Character to pad with. Defaults to " ".
+        """
+        assert len(character) == 1, "Character must be a string of length 1"
+        if count:
+            self.plain = f"{self.plain}{character * count}"
+
+    def align(self, align: AlignMethod, width: int, character: str = " ") -> None:
+        """Align text to a given width.
+
+        Args:
+            align (AlignMethod): One of "left", "center", or "right".
+            width (int): Desired width.
+            character (str, optional): Character to pad with. Defaults to " ".
+        """
+        self.truncate(width)
+        excess_space = width - cell_len(self.plain)
+        if excess_space:
+            if align == "left":
+                self.pad_right(excess_space, character)
+            elif align == "center":
+                left = excess_space // 2
+                self.pad_left(left, character)
+                self.pad_right(excess_space - left, character)
+            else:
+                self.pad_left(excess_space, character)
+
+    def append(
+        self, text: Union["Text", str], style: Optional[Union[str, "Style"]] = None
+    ) -> "Text":
+        """Add text with an optional style.
+
+        Args:
+            text (Union[Text, str]): A str or Text to append.
+            style (str, optional): A style name. Defaults to None.
+
+        Returns:
+            Text: Returns self for chaining.
+        """
+
+        if not isinstance(text, (str, Text)):
+            raise TypeError("Only str or Text can be appended to Text")
+
+        if len(text):
+            if isinstance(text, str):
+                sanitized_text = strip_control_codes(text)
+                self._text.append(sanitized_text)
+                offset = len(self)
+                text_length = len(sanitized_text)
+                if style:
+                    self._spans.append(Span(offset, offset + text_length, style))
+                self._length += text_length
+            elif isinstance(text, Text):
+                _Span = Span
+                if style is not None:
+                    raise ValueError(
+                        "style must not be set when appending Text instance"
+                    )
+                text_length = self._length
+                if text.style:
+                    self._spans.append(
+                        _Span(text_length, text_length + len(text), text.style)
+                    )
+                self._text.append(text.plain)
+                self._spans.extend(
+                    _Span(start + text_length, end + text_length, style)
+                    for start, end, style in text._spans.copy()
+                )
+                self._length += len(text)
+        return self
+
+    def append_text(self, text: "Text") -> "Text":
+        """Append another Text instance. This method is more performant that Text.append, but
+        only works for Text.
+
+        Args:
+            text (Text): The Text instance to append to this instance.
+
+        Returns:
+            Text: Returns self for chaining.
+        """
+        _Span = Span
+        text_length = self._length
+        if text.style:
+            self._spans.append(_Span(text_length, text_length + len(text), text.style))
+        self._text.append(text.plain)
+        self._spans.extend(
+            _Span(start + text_length, end + text_length, style)
+            for start, end, style in text._spans.copy()
+        )
+        self._length += len(text)
+        return self
+
+    def append_tokens(
+        self, tokens: Iterable[Tuple[str, Optional[StyleType]]]
+    ) -> "Text":
+        """Append iterable of str and style. Style may be a Style instance or a str style definition.
+
+        Args:
+            tokens (Iterable[Tuple[str, Optional[StyleType]]]): An iterable of tuples containing str content and style.
+
+        Returns:
+            Text: Returns self for chaining.
+        """
+        append_text = self._text.append
+        append_span = self._spans.append
+        _Span = Span
+        offset = len(self)
+        for content, style in tokens:
+            content = strip_control_codes(content)
+            append_text(content)
+            if style:
+                append_span(_Span(offset, offset + len(content), style))
+            offset += len(content)
+        self._length = offset
+        return self
+
+    def copy_styles(self, text: "Text") -> None:
+        """Copy styles from another Text instance.
+
+        Args:
+            text (Text): A Text instance to copy styles from, must be the same length.
+        """
+        self._spans.extend(text._spans)
+
+    def split(
+        self,
+        separator: str = "\n",
+        *,
+        include_separator: bool = False,
+        allow_blank: bool = False,
+    ) -> Lines:
+        """Split rich text in to lines, preserving styles.
+
+        Args:
+            separator (str, optional): String to split on. Defaults to "\\\\n".
+            include_separator (bool, optional): Include the separator in the lines. Defaults to False.
+            allow_blank (bool, optional): Return a blank line if the text ends with a separator. Defaults to False.
+
+        Returns:
+            List[RichText]: A list of rich text, one per line of the original.
+        """
+        assert separator, "separator must not be empty"
+
+        text = self.plain
+        if separator not in text:
+            return Lines([self.copy()])
+
+        if include_separator:
+            lines = self.divide(
+                match.end() for match in re.finditer(re.escape(separator), text)
+            )
+        else:
+
+            def flatten_spans() -> Iterable[int]:
+                for match in re.finditer(re.escape(separator), text):
+                    start, end = match.span()
+                    yield start
+                    yield end
+
+            lines = Lines(
+                line for line in self.divide(flatten_spans()) if line.plain != separator
+            )
+
+        if not allow_blank and text.endswith(separator):
+            lines.pop()
+
+        return lines
+
+    def divide(self, offsets: Iterable[int]) -> Lines:
+        """Divide text in to a number of lines at given offsets.
+
+        Args:
+            offsets (Iterable[int]): Offsets used to divide text.
+
+        Returns:
+            Lines: New RichText instances between offsets.
+        """
+        _offsets = list(offsets)
+
+        if not _offsets:
+            return Lines([self.copy()])
+
+        text = self.plain
+        text_length = len(text)
+        divide_offsets = [0, *_offsets, text_length]
+        line_ranges = list(zip(divide_offsets, divide_offsets[1:]))
+
+        style = self.style
+        justify = self.justify
+        overflow = self.overflow
+        _Text = Text
+        new_lines = Lines(
+            _Text(
+                text[start:end],
+                style=style,
+                justify=justify,
+                overflow=overflow,
+            )
+            for start, end in line_ranges
+        )
+        if not self._spans:
+            return new_lines
+
+        _line_appends = [line._spans.append for line in new_lines._lines]
+        line_count = len(line_ranges)
+        _Span = Span
+
+        for span_start, span_end, style in self._spans:
+            lower_bound = 0
+            upper_bound = line_count
+            start_line_no = (lower_bound + upper_bound) // 2
+
+            while True:
+                line_start, line_end = line_ranges[start_line_no]
+                if span_start < line_start:
+                    upper_bound = start_line_no - 1
+                elif span_start > line_end:
+                    lower_bound = start_line_no + 1
+                else:
+                    break
+                start_line_no = (lower_bound + upper_bound) // 2
+
+            if span_end < line_end:
+                end_line_no = start_line_no
+            else:
+                end_line_no = lower_bound = start_line_no
+                upper_bound = line_count
+
+                while True:
+                    line_start, line_end = line_ranges[end_line_no]
+                    if span_end < line_start:
+                        upper_bound = end_line_no - 1
+                    elif span_end > line_end:
+                        lower_bound = end_line_no + 1
+                    else:
+                        break
+                    end_line_no = (lower_bound + upper_bound) // 2
+
+            for line_no in range(start_line_no, end_line_no + 1):
+                line_start, line_end = line_ranges[line_no]
+                new_start = max(0, span_start - line_start)
+                new_end = min(span_end - line_start, line_end - line_start)
+                if new_end > new_start:
+                    _line_appends[line_no](_Span(new_start, new_end, style))
+
+        return new_lines
+
+    def right_crop(self, amount: int = 1) -> None:
+        """Remove a number of characters from the end of the text."""
+        max_offset = len(self.plain) - amount
+        _Span = Span
+        self._spans[:] = [
+            (
+                span
+                if span.end < max_offset
+                else _Span(span.start, min(max_offset, span.end), span.style)
+            )
+            for span in self._spans
+            if span.start < max_offset
+        ]
+        self._text = [self.plain[:-amount]]
+        self._length -= amount
+
+    def wrap(
+        self,
+        console: "Console",
+        width: int,
+        *,
+        justify: Optional["JustifyMethod"] = None,
+        overflow: Optional["OverflowMethod"] = None,
+        tab_size: int = 8,
+        no_wrap: Optional[bool] = None,
+    ) -> Lines:
+        """Word wrap the text.
+
+        Args:
+            console (Console): Console instance.
+            width (int): Number of cells available per line.
+            justify (str, optional): Justify method: "default", "left", "center", "full", "right". Defaults to "default".
+            overflow (str, optional): Overflow method: "crop", "fold", or "ellipsis". Defaults to None.
+            tab_size (int, optional): Default tab size. Defaults to 8.
+            no_wrap (bool, optional): Disable wrapping, Defaults to False.
+
+        Returns:
+            Lines: Number of lines.
+        """
+        wrap_justify = justify or self.justify or DEFAULT_JUSTIFY
+        wrap_overflow = overflow or self.overflow or DEFAULT_OVERFLOW
+
+        no_wrap = pick_bool(no_wrap, self.no_wrap, False) or overflow == "ignore"
+
+        lines = Lines()
+        for line in self.split(allow_blank=True):
+            if "\t" in line:
+                line.expand_tabs(tab_size)
+            if no_wrap:
+                new_lines = Lines([line])
+            else:
+                offsets = divide_line(str(line), width, fold=wrap_overflow == "fold")
+                new_lines = line.divide(offsets)
+            for line in new_lines:
+                line.rstrip_end(width)
+            if wrap_justify:
+                new_lines.justify(
+                    console, width, justify=wrap_justify, overflow=wrap_overflow
+                )
+            for line in new_lines:
+                line.truncate(width, overflow=wrap_overflow)
+            lines.extend(new_lines)
+        return lines
+
+    def fit(self, width: int) -> Lines:
+        """Fit the text in to given width by chopping in to lines.
+
+        Args:
+            width (int): Maximum characters in a line.
+
+        Returns:
+            Lines: Lines container.
+        """
+        lines: Lines = Lines()
+        append = lines.append
+        for line in self.split():
+            line.set_length(width)
+            append(line)
+        return lines
+
+    def detect_indentation(self) -> int:
+        """Auto-detect indentation of code.
+
+        Returns:
+            int: Number of spaces used to indent code.
+        """
+
+        _indentations = {
+            len(match.group(1))
+            for match in re.finditer(r"^( *)(.*)$", self.plain, flags=re.MULTILINE)
+        }
+
+        try:
+            indentation = (
+                reduce(gcd, [indent for indent in _indentations if not indent % 2]) or 1
+            )
+        except TypeError:
+            indentation = 1
+
+        return indentation
+
+    def with_indent_guides(
+        self,
+        indent_size: Optional[int] = None,
+        *,
+        character: str = "│",
+        style: StyleType = "dim green",
+    ) -> "Text":
+        """Adds indent guide lines to text.
+
+        Args:
+            indent_size (Optional[int]): Size of indentation, or None to auto detect. Defaults to None.
+            character (str, optional): Character to use for indentation. Defaults to "│".
+            style (Union[Style, str], optional): Style of indent guides.
+
+        Returns:
+            Text: New text with indentation guides.
+        """
+
+        _indent_size = self.detect_indentation() if indent_size is None else indent_size
+
+        text = self.copy()
+        text.expand_tabs()
+        indent_line = f"{character}{' ' * (_indent_size - 1)}"
+
+        re_indent = re.compile(r"^( *)(.*)$")
+        new_lines: List[Text] = []
+        add_line = new_lines.append
+        blank_lines = 0
+        for line in text.split(allow_blank=True):
+            match = re_indent.match(line.plain)
+            if not match or not match.group(2):
+                blank_lines += 1
+                continue
+            indent = match.group(1)
+            full_indents, remaining_space = divmod(len(indent), _indent_size)
+            new_indent = f"{indent_line * full_indents}{' ' * remaining_space}"
+            line.plain = new_indent + line.plain[len(new_indent) :]
+            line.stylize(style, 0, len(new_indent))
+            if blank_lines:
+                new_lines.extend([Text(new_indent, style=style)] * blank_lines)
+                blank_lines = 0
+            add_line(line)
+        if blank_lines:
+            new_lines.extend([Text("", style=style)] * blank_lines)
+
+        new_text = text.blank_copy("\n").join(new_lines)
+        return new_text
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Console
+
+    text = Text(
+        """\nLorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.\n"""
+    )
+    text.highlight_words(["Lorem"], "bold")
+    text.highlight_words(["ipsum"], "italic")
+
+    console = Console()
+
+    console.rule("justify='left'")
+    console.print(text, style="red")
+    console.print()
+
+    console.rule("justify='center'")
+    console.print(text, style="green", justify="center")
+    console.print()
+
+    console.rule("justify='right'")
+    console.print(text, style="blue", justify="right")
+    console.print()
+
+    console.rule("justify='full'")
+    console.print(text, style="magenta", justify="full")
+    console.print()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/theme.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/theme.py
new file mode 100644
index 00000000..227f1d86
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/theme.py
@@ -0,0 +1,115 @@
+import configparser
+from typing import IO, Dict, List, Mapping, Optional
+
+from .default_styles import DEFAULT_STYLES
+from .style import Style, StyleType
+
+
+class Theme:
+    """A container for style information, used by :class:`~rich.console.Console`.
+
+    Args:
+        styles (Dict[str, Style], optional): A mapping of style names on to styles. Defaults to None for a theme with no styles.
+        inherit (bool, optional): Inherit default styles. Defaults to True.
+    """
+
+    styles: Dict[str, Style]
+
+    def __init__(
+        self, styles: Optional[Mapping[str, StyleType]] = None, inherit: bool = True
+    ):
+        self.styles = DEFAULT_STYLES.copy() if inherit else {}
+        if styles is not None:
+            self.styles.update(
+                {
+                    name: style if isinstance(style, Style) else Style.parse(style)
+                    for name, style in styles.items()
+                }
+            )
+
+    @property
+    def config(self) -> str:
+        """Get contents of a config file for this theme."""
+        config = "[styles]\n" + "\n".join(
+            f"{name} = {style}" for name, style in sorted(self.styles.items())
+        )
+        return config
+
+    @classmethod
+    def from_file(
+        cls, config_file: IO[str], source: Optional[str] = None, inherit: bool = True
+    ) -> "Theme":
+        """Load a theme from a text mode file.
+
+        Args:
+            config_file (IO[str]): An open conf file.
+            source (str, optional): The filename of the open file. Defaults to None.
+            inherit (bool, optional): Inherit default styles. Defaults to True.
+
+        Returns:
+            Theme: A New theme instance.
+        """
+        config = configparser.ConfigParser()
+        config.read_file(config_file, source=source)
+        styles = {name: Style.parse(value) for name, value in config.items("styles")}
+        theme = Theme(styles, inherit=inherit)
+        return theme
+
+    @classmethod
+    def read(
+        cls, path: str, inherit: bool = True, encoding: Optional[str] = None
+    ) -> "Theme":
+        """Read a theme from a path.
+
+        Args:
+            path (str): Path to a config file readable by Python configparser module.
+            inherit (bool, optional): Inherit default styles. Defaults to True.
+            encoding (str, optional): Encoding of the config file. Defaults to None.
+
+        Returns:
+            Theme: A new theme instance.
+        """
+        with open(path, encoding=encoding) as config_file:
+            return cls.from_file(config_file, source=path, inherit=inherit)
+
+
+class ThemeStackError(Exception):
+    """Base exception for errors related to the theme stack."""
+
+
+class ThemeStack:
+    """A stack of themes.
+
+    Args:
+        theme (Theme): A theme instance
+    """
+
+    def __init__(self, theme: Theme) -> None:
+        self._entries: List[Dict[str, Style]] = [theme.styles]
+        self.get = self._entries[-1].get
+
+    def push_theme(self, theme: Theme, inherit: bool = True) -> None:
+        """Push a theme on the top of the stack.
+
+        Args:
+            theme (Theme): A Theme instance.
+            inherit (boolean, optional): Inherit styles from current top of stack.
+        """
+        styles: Dict[str, Style]
+        styles = (
+            {**self._entries[-1], **theme.styles} if inherit else theme.styles.copy()
+        )
+        self._entries.append(styles)
+        self.get = self._entries[-1].get
+
+    def pop_theme(self) -> None:
+        """Pop (and discard) the top-most theme."""
+        if len(self._entries) == 1:
+            raise ThemeStackError("Unable to pop base theme")
+        self._entries.pop()
+        self.get = self._entries[-1].get
+
+
+if __name__ == "__main__":  # pragma: no cover
+    theme = Theme()
+    print(theme.config)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/themes.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/themes.py
new file mode 100644
index 00000000..bf6db104
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/themes.py
@@ -0,0 +1,5 @@
+from .default_styles import DEFAULT_STYLES
+from .theme import Theme
+
+
+DEFAULT = Theme(DEFAULT_STYLES)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/traceback.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/traceback.py
new file mode 100644
index 00000000..28d742b4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/traceback.py
@@ -0,0 +1,797 @@
+import inspect
+import linecache
+import os
+import sys
+from dataclasses import dataclass, field
+from itertools import islice
+from traceback import walk_tb
+from types import ModuleType, TracebackType
+from typing import (
+    Any,
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    Optional,
+    Sequence,
+    Tuple,
+    Type,
+    Union,
+)
+
+from pip._vendor.pygments.lexers import guess_lexer_for_filename
+from pip._vendor.pygments.token import Comment, Keyword, Name, Number, Operator, String
+from pip._vendor.pygments.token import Text as TextToken
+from pip._vendor.pygments.token import Token
+from pip._vendor.pygments.util import ClassNotFound
+
+from . import pretty
+from ._loop import loop_last
+from .columns import Columns
+from .console import Console, ConsoleOptions, ConsoleRenderable, RenderResult, group
+from .constrain import Constrain
+from .highlighter import RegexHighlighter, ReprHighlighter
+from .panel import Panel
+from .scope import render_scope
+from .style import Style
+from .syntax import Syntax
+from .text import Text
+from .theme import Theme
+
+WINDOWS = sys.platform == "win32"
+
+LOCALS_MAX_LENGTH = 10
+LOCALS_MAX_STRING = 80
+
+
+def install(
+    *,
+    console: Optional[Console] = None,
+    width: Optional[int] = 100,
+    code_width: Optional[int] = 88,
+    extra_lines: int = 3,
+    theme: Optional[str] = None,
+    word_wrap: bool = False,
+    show_locals: bool = False,
+    locals_max_length: int = LOCALS_MAX_LENGTH,
+    locals_max_string: int = LOCALS_MAX_STRING,
+    locals_hide_dunder: bool = True,
+    locals_hide_sunder: Optional[bool] = None,
+    indent_guides: bool = True,
+    suppress: Iterable[Union[str, ModuleType]] = (),
+    max_frames: int = 100,
+) -> Callable[[Type[BaseException], BaseException, Optional[TracebackType]], Any]:
+    """Install a rich traceback handler.
+
+    Once installed, any tracebacks will be printed with syntax highlighting and rich formatting.
+
+
+    Args:
+        console (Optional[Console], optional): Console to write exception to. Default uses internal Console instance.
+        width (Optional[int], optional): Width (in characters) of traceback. Defaults to 100.
+        code_width (Optional[int], optional): Code width (in characters) of traceback. Defaults to 88.
+        extra_lines (int, optional): Extra lines of code. Defaults to 3.
+        theme (Optional[str], optional): Pygments theme to use in traceback. Defaults to ``None`` which will pick
+            a theme appropriate for the platform.
+        word_wrap (bool, optional): Enable word wrapping of long lines. Defaults to False.
+        show_locals (bool, optional): Enable display of local variables. Defaults to False.
+        locals_max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to 10.
+        locals_max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to 80.
+        locals_hide_dunder (bool, optional): Hide locals prefixed with double underscore. Defaults to True.
+        locals_hide_sunder (bool, optional): Hide locals prefixed with single underscore. Defaults to False.
+        indent_guides (bool, optional): Enable indent guides in code and locals. Defaults to True.
+        suppress (Sequence[Union[str, ModuleType]]): Optional sequence of modules or paths to exclude from traceback.
+
+    Returns:
+        Callable: The previous exception handler that was replaced.
+
+    """
+    traceback_console = Console(stderr=True) if console is None else console
+
+    locals_hide_sunder = (
+        True
+        if (traceback_console.is_jupyter and locals_hide_sunder is None)
+        else locals_hide_sunder
+    )
+
+    def excepthook(
+        type_: Type[BaseException],
+        value: BaseException,
+        traceback: Optional[TracebackType],
+    ) -> None:
+        traceback_console.print(
+            Traceback.from_exception(
+                type_,
+                value,
+                traceback,
+                width=width,
+                code_width=code_width,
+                extra_lines=extra_lines,
+                theme=theme,
+                word_wrap=word_wrap,
+                show_locals=show_locals,
+                locals_max_length=locals_max_length,
+                locals_max_string=locals_max_string,
+                locals_hide_dunder=locals_hide_dunder,
+                locals_hide_sunder=bool(locals_hide_sunder),
+                indent_guides=indent_guides,
+                suppress=suppress,
+                max_frames=max_frames,
+            )
+        )
+
+    def ipy_excepthook_closure(ip: Any) -> None:  # pragma: no cover
+        tb_data = {}  # store information about showtraceback call
+        default_showtraceback = ip.showtraceback  # keep reference of default traceback
+
+        def ipy_show_traceback(*args: Any, **kwargs: Any) -> None:
+            """wrap the default ip.showtraceback to store info for ip._showtraceback"""
+            nonlocal tb_data
+            tb_data = kwargs
+            default_showtraceback(*args, **kwargs)
+
+        def ipy_display_traceback(
+            *args: Any, is_syntax: bool = False, **kwargs: Any
+        ) -> None:
+            """Internally called traceback from ip._showtraceback"""
+            nonlocal tb_data
+            exc_tuple = ip._get_exc_info()
+
+            # do not display trace on syntax error
+            tb: Optional[TracebackType] = None if is_syntax else exc_tuple[2]
+
+            # determine correct tb_offset
+            compiled = tb_data.get("running_compiled_code", False)
+            tb_offset = tb_data.get("tb_offset", 1 if compiled else 0)
+            # remove ipython internal frames from trace with tb_offset
+            for _ in range(tb_offset):
+                if tb is None:
+                    break
+                tb = tb.tb_next
+
+            excepthook(exc_tuple[0], exc_tuple[1], tb)
+            tb_data = {}  # clear data upon usage
+
+        # replace _showtraceback instead of showtraceback to allow ipython features such as debugging to work
+        # this is also what the ipython docs recommends to modify when subclassing InteractiveShell
+        ip._showtraceback = ipy_display_traceback
+        # add wrapper to capture tb_data
+        ip.showtraceback = ipy_show_traceback
+        ip.showsyntaxerror = lambda *args, **kwargs: ipy_display_traceback(
+            *args, is_syntax=True, **kwargs
+        )
+
+    try:  # pragma: no cover
+        # if within ipython, use customized traceback
+        ip = get_ipython()  # type: ignore[name-defined]
+        ipy_excepthook_closure(ip)
+        return sys.excepthook
+    except Exception:
+        # otherwise use default system hook
+        old_excepthook = sys.excepthook
+        sys.excepthook = excepthook
+        return old_excepthook
+
+
+@dataclass
+class Frame:
+    filename: str
+    lineno: int
+    name: str
+    line: str = ""
+    locals: Optional[Dict[str, pretty.Node]] = None
+    last_instruction: Optional[Tuple[Tuple[int, int], Tuple[int, int]]] = None
+
+
+@dataclass
+class _SyntaxError:
+    offset: int
+    filename: str
+    line: str
+    lineno: int
+    msg: str
+
+
+@dataclass
+class Stack:
+    exc_type: str
+    exc_value: str
+    syntax_error: Optional[_SyntaxError] = None
+    is_cause: bool = False
+    frames: List[Frame] = field(default_factory=list)
+
+
+@dataclass
+class Trace:
+    stacks: List[Stack]
+
+
+class PathHighlighter(RegexHighlighter):
+    highlights = [r"(?P<dim>.*/)(?P<bold>.+)"]
+
+
+class Traceback:
+    """A Console renderable that renders a traceback.
+
+    Args:
+        trace (Trace, optional): A `Trace` object produced from `extract`. Defaults to None, which uses
+            the last exception.
+        width (Optional[int], optional): Number of characters used to traceback. Defaults to 100.
+        code_width (Optional[int], optional): Number of code characters used to traceback. Defaults to 88.
+        extra_lines (int, optional): Additional lines of code to render. Defaults to 3.
+        theme (str, optional): Override pygments theme used in traceback.
+        word_wrap (bool, optional): Enable word wrapping of long lines. Defaults to False.
+        show_locals (bool, optional): Enable display of local variables. Defaults to False.
+        indent_guides (bool, optional): Enable indent guides in code and locals. Defaults to True.
+        locals_max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+            Defaults to 10.
+        locals_max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to 80.
+        locals_hide_dunder (bool, optional): Hide locals prefixed with double underscore. Defaults to True.
+        locals_hide_sunder (bool, optional): Hide locals prefixed with single underscore. Defaults to False.
+        suppress (Sequence[Union[str, ModuleType]]): Optional sequence of modules or paths to exclude from traceback.
+        max_frames (int): Maximum number of frames to show in a traceback, 0 for no maximum. Defaults to 100.
+
+    """
+
+    LEXERS = {
+        "": "text",
+        ".py": "python",
+        ".pxd": "cython",
+        ".pyx": "cython",
+        ".pxi": "pyrex",
+    }
+
+    def __init__(
+        self,
+        trace: Optional[Trace] = None,
+        *,
+        width: Optional[int] = 100,
+        code_width: Optional[int] = 88,
+        extra_lines: int = 3,
+        theme: Optional[str] = None,
+        word_wrap: bool = False,
+        show_locals: bool = False,
+        locals_max_length: int = LOCALS_MAX_LENGTH,
+        locals_max_string: int = LOCALS_MAX_STRING,
+        locals_hide_dunder: bool = True,
+        locals_hide_sunder: bool = False,
+        indent_guides: bool = True,
+        suppress: Iterable[Union[str, ModuleType]] = (),
+        max_frames: int = 100,
+    ):
+        if trace is None:
+            exc_type, exc_value, traceback = sys.exc_info()
+            if exc_type is None or exc_value is None or traceback is None:
+                raise ValueError(
+                    "Value for 'trace' required if not called in except: block"
+                )
+            trace = self.extract(
+                exc_type, exc_value, traceback, show_locals=show_locals
+            )
+        self.trace = trace
+        self.width = width
+        self.code_width = code_width
+        self.extra_lines = extra_lines
+        self.theme = Syntax.get_theme(theme or "ansi_dark")
+        self.word_wrap = word_wrap
+        self.show_locals = show_locals
+        self.indent_guides = indent_guides
+        self.locals_max_length = locals_max_length
+        self.locals_max_string = locals_max_string
+        self.locals_hide_dunder = locals_hide_dunder
+        self.locals_hide_sunder = locals_hide_sunder
+
+        self.suppress: Sequence[str] = []
+        for suppress_entity in suppress:
+            if not isinstance(suppress_entity, str):
+                assert (
+                    suppress_entity.__file__ is not None
+                ), f"{suppress_entity!r} must be a module with '__file__' attribute"
+                path = os.path.dirname(suppress_entity.__file__)
+            else:
+                path = suppress_entity
+            path = os.path.normpath(os.path.abspath(path))
+            self.suppress.append(path)
+        self.max_frames = max(4, max_frames) if max_frames > 0 else 0
+
+    @classmethod
+    def from_exception(
+        cls,
+        exc_type: Type[Any],
+        exc_value: BaseException,
+        traceback: Optional[TracebackType],
+        *,
+        width: Optional[int] = 100,
+        code_width: Optional[int] = 88,
+        extra_lines: int = 3,
+        theme: Optional[str] = None,
+        word_wrap: bool = False,
+        show_locals: bool = False,
+        locals_max_length: int = LOCALS_MAX_LENGTH,
+        locals_max_string: int = LOCALS_MAX_STRING,
+        locals_hide_dunder: bool = True,
+        locals_hide_sunder: bool = False,
+        indent_guides: bool = True,
+        suppress: Iterable[Union[str, ModuleType]] = (),
+        max_frames: int = 100,
+    ) -> "Traceback":
+        """Create a traceback from exception info
+
+        Args:
+            exc_type (Type[BaseException]): Exception type.
+            exc_value (BaseException): Exception value.
+            traceback (TracebackType): Python Traceback object.
+            width (Optional[int], optional): Number of characters used to traceback. Defaults to 100.
+            code_width (Optional[int], optional): Number of code characters used to traceback. Defaults to 88.
+            extra_lines (int, optional): Additional lines of code to render. Defaults to 3.
+            theme (str, optional): Override pygments theme used in traceback.
+            word_wrap (bool, optional): Enable word wrapping of long lines. Defaults to False.
+            show_locals (bool, optional): Enable display of local variables. Defaults to False.
+            indent_guides (bool, optional): Enable indent guides in code and locals. Defaults to True.
+            locals_max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+                Defaults to 10.
+            locals_max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to 80.
+            locals_hide_dunder (bool, optional): Hide locals prefixed with double underscore. Defaults to True.
+            locals_hide_sunder (bool, optional): Hide locals prefixed with single underscore. Defaults to False.
+            suppress (Iterable[Union[str, ModuleType]]): Optional sequence of modules or paths to exclude from traceback.
+            max_frames (int): Maximum number of frames to show in a traceback, 0 for no maximum. Defaults to 100.
+
+        Returns:
+            Traceback: A Traceback instance that may be printed.
+        """
+        rich_traceback = cls.extract(
+            exc_type,
+            exc_value,
+            traceback,
+            show_locals=show_locals,
+            locals_max_length=locals_max_length,
+            locals_max_string=locals_max_string,
+            locals_hide_dunder=locals_hide_dunder,
+            locals_hide_sunder=locals_hide_sunder,
+        )
+
+        return cls(
+            rich_traceback,
+            width=width,
+            code_width=code_width,
+            extra_lines=extra_lines,
+            theme=theme,
+            word_wrap=word_wrap,
+            show_locals=show_locals,
+            indent_guides=indent_guides,
+            locals_max_length=locals_max_length,
+            locals_max_string=locals_max_string,
+            locals_hide_dunder=locals_hide_dunder,
+            locals_hide_sunder=locals_hide_sunder,
+            suppress=suppress,
+            max_frames=max_frames,
+        )
+
+    @classmethod
+    def extract(
+        cls,
+        exc_type: Type[BaseException],
+        exc_value: BaseException,
+        traceback: Optional[TracebackType],
+        *,
+        show_locals: bool = False,
+        locals_max_length: int = LOCALS_MAX_LENGTH,
+        locals_max_string: int = LOCALS_MAX_STRING,
+        locals_hide_dunder: bool = True,
+        locals_hide_sunder: bool = False,
+    ) -> Trace:
+        """Extract traceback information.
+
+        Args:
+            exc_type (Type[BaseException]): Exception type.
+            exc_value (BaseException): Exception value.
+            traceback (TracebackType): Python Traceback object.
+            show_locals (bool, optional): Enable display of local variables. Defaults to False.
+            locals_max_length (int, optional): Maximum length of containers before abbreviating, or None for no abbreviation.
+                Defaults to 10.
+            locals_max_string (int, optional): Maximum length of string before truncating, or None to disable. Defaults to 80.
+            locals_hide_dunder (bool, optional): Hide locals prefixed with double underscore. Defaults to True.
+            locals_hide_sunder (bool, optional): Hide locals prefixed with single underscore. Defaults to False.
+
+        Returns:
+            Trace: A Trace instance which you can use to construct a `Traceback`.
+        """
+
+        stacks: List[Stack] = []
+        is_cause = False
+
+        from pip._vendor.rich import _IMPORT_CWD
+
+        def safe_str(_object: Any) -> str:
+            """Don't allow exceptions from __str__ to propagate."""
+            try:
+                return str(_object)
+            except Exception:
+                return "<exception str() failed>"
+
+        while True:
+            stack = Stack(
+                exc_type=safe_str(exc_type.__name__),
+                exc_value=safe_str(exc_value),
+                is_cause=is_cause,
+            )
+
+            if isinstance(exc_value, SyntaxError):
+                stack.syntax_error = _SyntaxError(
+                    offset=exc_value.offset or 0,
+                    filename=exc_value.filename or "?",
+                    lineno=exc_value.lineno or 0,
+                    line=exc_value.text or "",
+                    msg=exc_value.msg,
+                )
+
+            stacks.append(stack)
+            append = stack.frames.append
+
+            def get_locals(
+                iter_locals: Iterable[Tuple[str, object]]
+            ) -> Iterable[Tuple[str, object]]:
+                """Extract locals from an iterator of key pairs."""
+                if not (locals_hide_dunder or locals_hide_sunder):
+                    yield from iter_locals
+                    return
+                for key, value in iter_locals:
+                    if locals_hide_dunder and key.startswith("__"):
+                        continue
+                    if locals_hide_sunder and key.startswith("_"):
+                        continue
+                    yield key, value
+
+            for frame_summary, line_no in walk_tb(traceback):
+                filename = frame_summary.f_code.co_filename
+
+                last_instruction: Optional[Tuple[Tuple[int, int], Tuple[int, int]]]
+                last_instruction = None
+                if sys.version_info >= (3, 11):
+                    instruction_index = frame_summary.f_lasti // 2
+                    instruction_position = next(
+                        islice(
+                            frame_summary.f_code.co_positions(),
+                            instruction_index,
+                            instruction_index + 1,
+                        )
+                    )
+                    (
+                        start_line,
+                        end_line,
+                        start_column,
+                        end_column,
+                    ) = instruction_position
+                    if (
+                        start_line is not None
+                        and end_line is not None
+                        and start_column is not None
+                        and end_column is not None
+                    ):
+                        last_instruction = (
+                            (start_line, start_column),
+                            (end_line, end_column),
+                        )
+
+                if filename and not filename.startswith("<"):
+                    if not os.path.isabs(filename):
+                        filename = os.path.join(_IMPORT_CWD, filename)
+                if frame_summary.f_locals.get("_rich_traceback_omit", False):
+                    continue
+
+                frame = Frame(
+                    filename=filename or "?",
+                    lineno=line_no,
+                    name=frame_summary.f_code.co_name,
+                    locals=(
+                        {
+                            key: pretty.traverse(
+                                value,
+                                max_length=locals_max_length,
+                                max_string=locals_max_string,
+                            )
+                            for key, value in get_locals(frame_summary.f_locals.items())
+                            if not (inspect.isfunction(value) or inspect.isclass(value))
+                        }
+                        if show_locals
+                        else None
+                    ),
+                    last_instruction=last_instruction,
+                )
+                append(frame)
+                if frame_summary.f_locals.get("_rich_traceback_guard", False):
+                    del stack.frames[:]
+
+            cause = getattr(exc_value, "__cause__", None)
+            if cause:
+                exc_type = cause.__class__
+                exc_value = cause
+                # __traceback__ can be None, e.g. for exceptions raised by the
+                # 'multiprocessing' module
+                traceback = cause.__traceback__
+                is_cause = True
+                continue
+
+            cause = exc_value.__context__
+            if cause and not getattr(exc_value, "__suppress_context__", False):
+                exc_type = cause.__class__
+                exc_value = cause
+                traceback = cause.__traceback__
+                is_cause = False
+                continue
+            # No cover, code is reached but coverage doesn't recognize it.
+            break  # pragma: no cover
+
+        trace = Trace(stacks=stacks)
+        return trace
+
+    def __rich_console__(
+        self, console: Console, options: ConsoleOptions
+    ) -> RenderResult:
+        theme = self.theme
+        background_style = theme.get_background_style()
+        token_style = theme.get_style_for_token
+
+        traceback_theme = Theme(
+            {
+                "pretty": token_style(TextToken),
+                "pygments.text": token_style(Token),
+                "pygments.string": token_style(String),
+                "pygments.function": token_style(Name.Function),
+                "pygments.number": token_style(Number),
+                "repr.indent": token_style(Comment) + Style(dim=True),
+                "repr.str": token_style(String),
+                "repr.brace": token_style(TextToken) + Style(bold=True),
+                "repr.number": token_style(Number),
+                "repr.bool_true": token_style(Keyword.Constant),
+                "repr.bool_false": token_style(Keyword.Constant),
+                "repr.none": token_style(Keyword.Constant),
+                "scope.border": token_style(String.Delimiter),
+                "scope.equals": token_style(Operator),
+                "scope.key": token_style(Name),
+                "scope.key.special": token_style(Name.Constant) + Style(dim=True),
+            },
+            inherit=False,
+        )
+
+        highlighter = ReprHighlighter()
+        for last, stack in loop_last(reversed(self.trace.stacks)):
+            if stack.frames:
+                stack_renderable: ConsoleRenderable = Panel(
+                    self._render_stack(stack),
+                    title="[traceback.title]Traceback [dim](most recent call last)",
+                    style=background_style,
+                    border_style="traceback.border",
+                    expand=True,
+                    padding=(0, 1),
+                )
+                stack_renderable = Constrain(stack_renderable, self.width)
+                with console.use_theme(traceback_theme):
+                    yield stack_renderable
+            if stack.syntax_error is not None:
+                with console.use_theme(traceback_theme):
+                    yield Constrain(
+                        Panel(
+                            self._render_syntax_error(stack.syntax_error),
+                            style=background_style,
+                            border_style="traceback.border.syntax_error",
+                            expand=True,
+                            padding=(0, 1),
+                            width=self.width,
+                        ),
+                        self.width,
+                    )
+                yield Text.assemble(
+                    (f"{stack.exc_type}: ", "traceback.exc_type"),
+                    highlighter(stack.syntax_error.msg),
+                )
+            elif stack.exc_value:
+                yield Text.assemble(
+                    (f"{stack.exc_type}: ", "traceback.exc_type"),
+                    highlighter(stack.exc_value),
+                )
+            else:
+                yield Text.assemble((f"{stack.exc_type}", "traceback.exc_type"))
+
+            if not last:
+                if stack.is_cause:
+                    yield Text.from_markup(
+                        "\n[i]The above exception was the direct cause of the following exception:\n",
+                    )
+                else:
+                    yield Text.from_markup(
+                        "\n[i]During handling of the above exception, another exception occurred:\n",
+                    )
+
+    @group()
+    def _render_syntax_error(self, syntax_error: _SyntaxError) -> RenderResult:
+        highlighter = ReprHighlighter()
+        path_highlighter = PathHighlighter()
+        if syntax_error.filename != "<stdin>":
+            if os.path.exists(syntax_error.filename):
+                text = Text.assemble(
+                    (f" {syntax_error.filename}", "pygments.string"),
+                    (":", "pygments.text"),
+                    (str(syntax_error.lineno), "pygments.number"),
+                    style="pygments.text",
+                )
+                yield path_highlighter(text)
+        syntax_error_text = highlighter(syntax_error.line.rstrip())
+        syntax_error_text.no_wrap = True
+        offset = min(syntax_error.offset - 1, len(syntax_error_text))
+        syntax_error_text.stylize("bold underline", offset, offset)
+        syntax_error_text += Text.from_markup(
+            "\n" + " " * offset + "[traceback.offset]▲[/]",
+            style="pygments.text",
+        )
+        yield syntax_error_text
+
+    @classmethod
+    def _guess_lexer(cls, filename: str, code: str) -> str:
+        ext = os.path.splitext(filename)[-1]
+        if not ext:
+            # No extension, look at first line to see if it is a hashbang
+            # Note, this is an educated guess and not a guarantee
+            # If it fails, the only downside is that the code is highlighted strangely
+            new_line_index = code.index("\n")
+            first_line = code[:new_line_index] if new_line_index != -1 else code
+            if first_line.startswith("#!") and "python" in first_line.lower():
+                return "python"
+        try:
+            return cls.LEXERS.get(ext) or guess_lexer_for_filename(filename, code).name
+        except ClassNotFound:
+            return "text"
+
+    @group()
+    def _render_stack(self, stack: Stack) -> RenderResult:
+        path_highlighter = PathHighlighter()
+        theme = self.theme
+
+        def read_code(filename: str) -> str:
+            """Read files, and cache results on filename.
+
+            Args:
+                filename (str): Filename to read
+
+            Returns:
+                str: Contents of file
+            """
+            return "".join(linecache.getlines(filename))
+
+        def render_locals(frame: Frame) -> Iterable[ConsoleRenderable]:
+            if frame.locals:
+                yield render_scope(
+                    frame.locals,
+                    title="locals",
+                    indent_guides=self.indent_guides,
+                    max_length=self.locals_max_length,
+                    max_string=self.locals_max_string,
+                )
+
+        exclude_frames: Optional[range] = None
+        if self.max_frames != 0:
+            exclude_frames = range(
+                self.max_frames // 2,
+                len(stack.frames) - self.max_frames // 2,
+            )
+
+        excluded = False
+        for frame_index, frame in enumerate(stack.frames):
+            if exclude_frames and frame_index in exclude_frames:
+                excluded = True
+                continue
+
+            if excluded:
+                assert exclude_frames is not None
+                yield Text(
+                    f"\n... {len(exclude_frames)} frames hidden ...",
+                    justify="center",
+                    style="traceback.error",
+                )
+                excluded = False
+
+            first = frame_index == 0
+            frame_filename = frame.filename
+            suppressed = any(frame_filename.startswith(path) for path in self.suppress)
+
+            if os.path.exists(frame.filename):
+                text = Text.assemble(
+                    path_highlighter(Text(frame.filename, style="pygments.string")),
+                    (":", "pygments.text"),
+                    (str(frame.lineno), "pygments.number"),
+                    " in ",
+                    (frame.name, "pygments.function"),
+                    style="pygments.text",
+                )
+            else:
+                text = Text.assemble(
+                    "in ",
+                    (frame.name, "pygments.function"),
+                    (":", "pygments.text"),
+                    (str(frame.lineno), "pygments.number"),
+                    style="pygments.text",
+                )
+            if not frame.filename.startswith("<") and not first:
+                yield ""
+            yield text
+            if frame.filename.startswith("<"):
+                yield from render_locals(frame)
+                continue
+            if not suppressed:
+                try:
+                    code = read_code(frame.filename)
+                    if not code:
+                        # code may be an empty string if the file doesn't exist, OR
+                        # if the traceback filename is generated dynamically
+                        continue
+                    lexer_name = self._guess_lexer(frame.filename, code)
+                    syntax = Syntax(
+                        code,
+                        lexer_name,
+                        theme=theme,
+                        line_numbers=True,
+                        line_range=(
+                            frame.lineno - self.extra_lines,
+                            frame.lineno + self.extra_lines,
+                        ),
+                        highlight_lines={frame.lineno},
+                        word_wrap=self.word_wrap,
+                        code_width=self.code_width,
+                        indent_guides=self.indent_guides,
+                        dedent=False,
+                    )
+                    yield ""
+                except Exception as error:
+                    yield Text.assemble(
+                        (f"\n{error}", "traceback.error"),
+                    )
+                else:
+                    if frame.last_instruction is not None:
+                        start, end = frame.last_instruction
+                        syntax.stylize_range(
+                            style="traceback.error_range",
+                            start=start,
+                            end=end,
+                            style_before=True,
+                        )
+                    yield (
+                        Columns(
+                            [
+                                syntax,
+                                *render_locals(frame),
+                            ],
+                            padding=1,
+                        )
+                        if frame.locals
+                        else syntax
+                    )
+
+
+if __name__ == "__main__":  # pragma: no cover
+    install(show_locals=True)
+    import sys
+
+    def bar(
+        a: Any,
+    ) -> None:  # 这是对亚洲语言支持的测试。面对模棱两可的想法，拒绝猜测的诱惑
+        one = 1
+        print(one / a)
+
+    def foo(a: Any) -> None:
+        _rich_traceback_guard = True
+        zed = {
+            "characters": {
+                "Paul Atreides",
+                "Vladimir Harkonnen",
+                "Thufir Hawat",
+                "Duncan Idaho",
+            },
+            "atomic_types": (None, False, True),
+        }
+        bar(a)
+
+    def error() -> None:
+        foo(0)
+
+    error()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/tree.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/tree.py
new file mode 100644
index 00000000..27c5cf7b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/rich/tree.py
@@ -0,0 +1,257 @@
+from typing import Iterator, List, Optional, Tuple
+
+from ._loop import loop_first, loop_last
+from .console import Console, ConsoleOptions, RenderableType, RenderResult
+from .jupyter import JupyterMixin
+from .measure import Measurement
+from .segment import Segment
+from .style import Style, StyleStack, StyleType
+from .styled import Styled
+
+GuideType = Tuple[str, str, str, str]
+
+
+class Tree(JupyterMixin):
+    """A renderable for a tree structure.
+
+    Attributes:
+        ASCII_GUIDES (GuideType): Guide lines used when Console.ascii_only is True.
+        TREE_GUIDES (List[GuideType, GuideType, GuideType]): Default guide lines.
+
+    Args:
+        label (RenderableType): The renderable or str for the tree label.
+        style (StyleType, optional): Style of this tree. Defaults to "tree".
+        guide_style (StyleType, optional): Style of the guide lines. Defaults to "tree.line".
+        expanded (bool, optional): Also display children. Defaults to True.
+        highlight (bool, optional): Highlight renderable (if str). Defaults to False.
+        hide_root (bool, optional): Hide the root node. Defaults to False.
+    """
+
+    ASCII_GUIDES = ("    ", "|   ", "+-- ", "`-- ")
+    TREE_GUIDES = [
+        ("    ", "│   ", "├── ", "└── "),
+        ("    ", "┃   ", "┣━━ ", "┗━━ "),
+        ("    ", "║   ", "╠══ ", "╚══ "),
+    ]
+
+    def __init__(
+        self,
+        label: RenderableType,
+        *,
+        style: StyleType = "tree",
+        guide_style: StyleType = "tree.line",
+        expanded: bool = True,
+        highlight: bool = False,
+        hide_root: bool = False,
+    ) -> None:
+        self.label = label
+        self.style = style
+        self.guide_style = guide_style
+        self.children: List[Tree] = []
+        self.expanded = expanded
+        self.highlight = highlight
+        self.hide_root = hide_root
+
+    def add(
+        self,
+        label: RenderableType,
+        *,
+        style: Optional[StyleType] = None,
+        guide_style: Optional[StyleType] = None,
+        expanded: bool = True,
+        highlight: Optional[bool] = False,
+    ) -> "Tree":
+        """Add a child tree.
+
+        Args:
+            label (RenderableType): The renderable or str for the tree label.
+            style (StyleType, optional): Style of this tree. Defaults to "tree".
+            guide_style (StyleType, optional): Style of the guide lines. Defaults to "tree.line".
+            expanded (bool, optional): Also display children. Defaults to True.
+            highlight (Optional[bool], optional): Highlight renderable (if str). Defaults to False.
+
+        Returns:
+            Tree: A new child Tree, which may be further modified.
+        """
+        node = Tree(
+            label,
+            style=self.style if style is None else style,
+            guide_style=self.guide_style if guide_style is None else guide_style,
+            expanded=expanded,
+            highlight=self.highlight if highlight is None else highlight,
+        )
+        self.children.append(node)
+        return node
+
+    def __rich_console__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "RenderResult":
+        stack: List[Iterator[Tuple[bool, Tree]]] = []
+        pop = stack.pop
+        push = stack.append
+        new_line = Segment.line()
+
+        get_style = console.get_style
+        null_style = Style.null()
+        guide_style = get_style(self.guide_style, default="") or null_style
+        SPACE, CONTINUE, FORK, END = range(4)
+
+        _Segment = Segment
+
+        def make_guide(index: int, style: Style) -> Segment:
+            """Make a Segment for a level of the guide lines."""
+            if options.ascii_only:
+                line = self.ASCII_GUIDES[index]
+            else:
+                guide = 1 if style.bold else (2 if style.underline2 else 0)
+                line = self.TREE_GUIDES[0 if options.legacy_windows else guide][index]
+            return _Segment(line, style)
+
+        levels: List[Segment] = [make_guide(CONTINUE, guide_style)]
+        push(iter(loop_last([self])))
+
+        guide_style_stack = StyleStack(get_style(self.guide_style))
+        style_stack = StyleStack(get_style(self.style))
+        remove_guide_styles = Style(bold=False, underline2=False)
+
+        depth = 0
+
+        while stack:
+            stack_node = pop()
+            try:
+                last, node = next(stack_node)
+            except StopIteration:
+                levels.pop()
+                if levels:
+                    guide_style = levels[-1].style or null_style
+                    levels[-1] = make_guide(FORK, guide_style)
+                    guide_style_stack.pop()
+                    style_stack.pop()
+                continue
+            push(stack_node)
+            if last:
+                levels[-1] = make_guide(END, levels[-1].style or null_style)
+
+            guide_style = guide_style_stack.current + get_style(node.guide_style)
+            style = style_stack.current + get_style(node.style)
+            prefix = levels[(2 if self.hide_root else 1) :]
+            renderable_lines = console.render_lines(
+                Styled(node.label, style),
+                options.update(
+                    width=options.max_width
+                    - sum(level.cell_length for level in prefix),
+                    highlight=self.highlight,
+                    height=None,
+                ),
+                pad=options.justify is not None,
+            )
+
+            if not (depth == 0 and self.hide_root):
+                for first, line in loop_first(renderable_lines):
+                    if prefix:
+                        yield from _Segment.apply_style(
+                            prefix,
+                            style.background_style,
+                            post_style=remove_guide_styles,
+                        )
+                    yield from line
+                    yield new_line
+                    if first and prefix:
+                        prefix[-1] = make_guide(
+                            SPACE if last else CONTINUE, prefix[-1].style or null_style
+                        )
+
+            if node.expanded and node.children:
+                levels[-1] = make_guide(
+                    SPACE if last else CONTINUE, levels[-1].style or null_style
+                )
+                levels.append(
+                    make_guide(END if len(node.children) == 1 else FORK, guide_style)
+                )
+                style_stack.push(get_style(node.style))
+                guide_style_stack.push(get_style(node.guide_style))
+                push(iter(loop_last(node.children)))
+                depth += 1
+
+    def __rich_measure__(
+        self, console: "Console", options: "ConsoleOptions"
+    ) -> "Measurement":
+        stack: List[Iterator[Tree]] = [iter([self])]
+        pop = stack.pop
+        push = stack.append
+        minimum = 0
+        maximum = 0
+        measure = Measurement.get
+        level = 0
+        while stack:
+            iter_tree = pop()
+            try:
+                tree = next(iter_tree)
+            except StopIteration:
+                level -= 1
+                continue
+            push(iter_tree)
+            min_measure, max_measure = measure(console, options, tree.label)
+            indent = level * 4
+            minimum = max(min_measure + indent, minimum)
+            maximum = max(max_measure + indent, maximum)
+            if tree.expanded and tree.children:
+                push(iter(tree.children))
+                level += 1
+        return Measurement(minimum, maximum)
+
+
+if __name__ == "__main__":  # pragma: no cover
+    from pip._vendor.rich.console import Group
+    from pip._vendor.rich.markdown import Markdown
+    from pip._vendor.rich.panel import Panel
+    from pip._vendor.rich.syntax import Syntax
+    from pip._vendor.rich.table import Table
+
+    table = Table(row_styles=["", "dim"])
+
+    table.add_column("Released", style="cyan", no_wrap=True)
+    table.add_column("Title", style="magenta")
+    table.add_column("Box Office", justify="right", style="green")
+
+    table.add_row("Dec 20, 2019", "Star Wars: The Rise of Skywalker", "$952,110,690")
+    table.add_row("May 25, 2018", "Solo: A Star Wars Story", "$393,151,347")
+    table.add_row("Dec 15, 2017", "Star Wars Ep. V111: The Last Jedi", "$1,332,539,889")
+    table.add_row("Dec 16, 2016", "Rogue One: A Star Wars Story", "$1,332,439,889")
+
+    code = """\
+class Segment(NamedTuple):
+    text: str = ""
+    style: Optional[Style] = None
+    is_control: bool = False
+"""
+    syntax = Syntax(code, "python", theme="monokai", line_numbers=True)
+
+    markdown = Markdown(
+        """\
+### example.md
+> Hello, World!
+>
+> Markdown _all_ the things
+"""
+    )
+
+    root = Tree("🌲 [b green]Rich Tree", highlight=True, hide_root=True)
+
+    node = root.add(":file_folder: Renderables", guide_style="red")
+    simple_node = node.add(":file_folder: [bold yellow]Atomic", guide_style="uu green")
+    simple_node.add(Group("📄 Syntax", syntax))
+    simple_node.add(Group("📄 Markdown", Panel(markdown, border_style="green")))
+
+    containers_node = node.add(
+        ":file_folder: [bold magenta]Containers", guide_style="bold magenta"
+    )
+    containers_node.expanded = True
+    panel = Panel.fit("Just a panel", border_style="red")
+    containers_node.add(Group("📄 Panels", panel))
+
+    containers_node.add(Group("📄 [b magenta]Table", table))
+
+    console = Console()
+
+    console.print(root)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__init__.py
new file mode 100644
index 00000000..2b08d6e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__init__.py
@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+__all__ = ("loads", "load", "TOMLDecodeError")
+__version__ = "2.2.1"  # DO NOT EDIT THIS LINE MANUALLY. LET bump2version UTILITY DO IT
+
+from ._parser import TOMLDecodeError, load, loads
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7cf3b656
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..425a5842
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_re.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_re.cpython-312.pyc
new file mode 100644
index 00000000..a399f3d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_re.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_types.cpython-312.pyc
new file mode 100644
index 00000000..dd678aea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/__pycache__/_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_parser.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_parser.py
new file mode 100644
index 00000000..b548e8b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_parser.py
@@ -0,0 +1,770 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from __future__ import annotations
+
+from collections.abc import Iterable
+import string
+import sys
+from types import MappingProxyType
+from typing import IO, Any, Final, NamedTuple
+import warnings
+
+from ._re import (
+    RE_DATETIME,
+    RE_LOCALTIME,
+    RE_NUMBER,
+    match_to_datetime,
+    match_to_localtime,
+    match_to_number,
+)
+from ._types import Key, ParseFloat, Pos
+
+# Inline tables/arrays are implemented using recursion. Pathologically
+# nested documents cause pure Python to raise RecursionError (which is OK),
+# but mypyc binary wheels will crash unrecoverably (not OK). According to
+# mypyc docs this will be fixed in the future:
+# https://mypyc.readthedocs.io/en/latest/differences_from_python.html#stack-overflows
+# Before mypyc's fix is in, recursion needs to be limited by this library.
+# Choosing `sys.getrecursionlimit()` as maximum inline table/array nesting
+# level, as it allows more nesting than pure Python, but still seems a far
+# lower number than where mypyc binaries crash.
+MAX_INLINE_NESTING: Final = sys.getrecursionlimit()
+
+ASCII_CTRL: Final = frozenset(chr(i) for i in range(32)) | frozenset(chr(127))
+
+# Neither of these sets include quotation mark or backslash. They are
+# currently handled as separate cases in the parser functions.
+ILLEGAL_BASIC_STR_CHARS: Final = ASCII_CTRL - frozenset("\t")
+ILLEGAL_MULTILINE_BASIC_STR_CHARS: Final = ASCII_CTRL - frozenset("\t\n")
+
+ILLEGAL_LITERAL_STR_CHARS: Final = ILLEGAL_BASIC_STR_CHARS
+ILLEGAL_MULTILINE_LITERAL_STR_CHARS: Final = ILLEGAL_MULTILINE_BASIC_STR_CHARS
+
+ILLEGAL_COMMENT_CHARS: Final = ILLEGAL_BASIC_STR_CHARS
+
+TOML_WS: Final = frozenset(" \t")
+TOML_WS_AND_NEWLINE: Final = TOML_WS | frozenset("\n")
+BARE_KEY_CHARS: Final = frozenset(string.ascii_letters + string.digits + "-_")
+KEY_INITIAL_CHARS: Final = BARE_KEY_CHARS | frozenset("\"'")
+HEXDIGIT_CHARS: Final = frozenset(string.hexdigits)
+
+BASIC_STR_ESCAPE_REPLACEMENTS: Final = MappingProxyType(
+    {
+        "\\b": "\u0008",  # backspace
+        "\\t": "\u0009",  # tab
+        "\\n": "\u000A",  # linefeed
+        "\\f": "\u000C",  # form feed
+        "\\r": "\u000D",  # carriage return
+        '\\"': "\u0022",  # quote
+        "\\\\": "\u005C",  # backslash
+    }
+)
+
+
+class DEPRECATED_DEFAULT:
+    """Sentinel to be used as default arg during deprecation
+    period of TOMLDecodeError's free-form arguments."""
+
+
+class TOMLDecodeError(ValueError):
+    """An error raised if a document is not valid TOML.
+
+    Adds the following attributes to ValueError:
+    msg: The unformatted error message
+    doc: The TOML document being parsed
+    pos: The index of doc where parsing failed
+    lineno: The line corresponding to pos
+    colno: The column corresponding to pos
+    """
+
+    def __init__(
+        self,
+        msg: str | type[DEPRECATED_DEFAULT] = DEPRECATED_DEFAULT,
+        doc: str | type[DEPRECATED_DEFAULT] = DEPRECATED_DEFAULT,
+        pos: Pos | type[DEPRECATED_DEFAULT] = DEPRECATED_DEFAULT,
+        *args: Any,
+    ):
+        if (
+            args
+            or not isinstance(msg, str)
+            or not isinstance(doc, str)
+            or not isinstance(pos, int)
+        ):
+            warnings.warn(
+                "Free-form arguments for TOMLDecodeError are deprecated. "
+                "Please set 'msg' (str), 'doc' (str) and 'pos' (int) arguments only.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            if pos is not DEPRECATED_DEFAULT:
+                args = pos, *args
+            if doc is not DEPRECATED_DEFAULT:
+                args = doc, *args
+            if msg is not DEPRECATED_DEFAULT:
+                args = msg, *args
+            ValueError.__init__(self, *args)
+            return
+
+        lineno = doc.count("\n", 0, pos) + 1
+        if lineno == 1:
+            colno = pos + 1
+        else:
+            colno = pos - doc.rindex("\n", 0, pos)
+
+        if pos >= len(doc):
+            coord_repr = "end of document"
+        else:
+            coord_repr = f"line {lineno}, column {colno}"
+        errmsg = f"{msg} (at {coord_repr})"
+        ValueError.__init__(self, errmsg)
+
+        self.msg = msg
+        self.doc = doc
+        self.pos = pos
+        self.lineno = lineno
+        self.colno = colno
+
+
+def load(__fp: IO[bytes], *, parse_float: ParseFloat = float) -> dict[str, Any]:
+    """Parse TOML from a binary file object."""
+    b = __fp.read()
+    try:
+        s = b.decode()
+    except AttributeError:
+        raise TypeError(
+            "File must be opened in binary mode, e.g. use `open('foo.toml', 'rb')`"
+        ) from None
+    return loads(s, parse_float=parse_float)
+
+
+def loads(__s: str, *, parse_float: ParseFloat = float) -> dict[str, Any]:  # noqa: C901
+    """Parse TOML from a string."""
+
+    # The spec allows converting "\r\n" to "\n", even in string
+    # literals. Let's do so to simplify parsing.
+    try:
+        src = __s.replace("\r\n", "\n")
+    except (AttributeError, TypeError):
+        raise TypeError(
+            f"Expected str object, not '{type(__s).__qualname__}'"
+        ) from None
+    pos = 0
+    out = Output(NestedDict(), Flags())
+    header: Key = ()
+    parse_float = make_safe_parse_float(parse_float)
+
+    # Parse one statement at a time
+    # (typically means one line in TOML source)
+    while True:
+        # 1. Skip line leading whitespace
+        pos = skip_chars(src, pos, TOML_WS)
+
+        # 2. Parse rules. Expect one of the following:
+        #    - end of file
+        #    - end of line
+        #    - comment
+        #    - key/value pair
+        #    - append dict to list (and move to its namespace)
+        #    - create dict (and move to its namespace)
+        # Skip trailing whitespace when applicable.
+        try:
+            char = src[pos]
+        except IndexError:
+            break
+        if char == "\n":
+            pos += 1
+            continue
+        if char in KEY_INITIAL_CHARS:
+            pos = key_value_rule(src, pos, out, header, parse_float)
+            pos = skip_chars(src, pos, TOML_WS)
+        elif char == "[":
+            try:
+                second_char: str | None = src[pos + 1]
+            except IndexError:
+                second_char = None
+            out.flags.finalize_pending()
+            if second_char == "[":
+                pos, header = create_list_rule(src, pos, out)
+            else:
+                pos, header = create_dict_rule(src, pos, out)
+            pos = skip_chars(src, pos, TOML_WS)
+        elif char != "#":
+            raise TOMLDecodeError("Invalid statement", src, pos)
+
+        # 3. Skip comment
+        pos = skip_comment(src, pos)
+
+        # 4. Expect end of line or end of file
+        try:
+            char = src[pos]
+        except IndexError:
+            break
+        if char != "\n":
+            raise TOMLDecodeError(
+                "Expected newline or end of document after a statement", src, pos
+            )
+        pos += 1
+
+    return out.data.dict
+
+
+class Flags:
+    """Flags that map to parsed keys/namespaces."""
+
+    # Marks an immutable namespace (inline array or inline table).
+    FROZEN: Final = 0
+    # Marks a nest that has been explicitly created and can no longer
+    # be opened using the "[table]" syntax.
+    EXPLICIT_NEST: Final = 1
+
+    def __init__(self) -> None:
+        self._flags: dict[str, dict] = {}
+        self._pending_flags: set[tuple[Key, int]] = set()
+
+    def add_pending(self, key: Key, flag: int) -> None:
+        self._pending_flags.add((key, flag))
+
+    def finalize_pending(self) -> None:
+        for key, flag in self._pending_flags:
+            self.set(key, flag, recursive=False)
+        self._pending_flags.clear()
+
+    def unset_all(self, key: Key) -> None:
+        cont = self._flags
+        for k in key[:-1]:
+            if k not in cont:
+                return
+            cont = cont[k]["nested"]
+        cont.pop(key[-1], None)
+
+    def set(self, key: Key, flag: int, *, recursive: bool) -> None:  # noqa: A003
+        cont = self._flags
+        key_parent, key_stem = key[:-1], key[-1]
+        for k in key_parent:
+            if k not in cont:
+                cont[k] = {"flags": set(), "recursive_flags": set(), "nested": {}}
+            cont = cont[k]["nested"]
+        if key_stem not in cont:
+            cont[key_stem] = {"flags": set(), "recursive_flags": set(), "nested": {}}
+        cont[key_stem]["recursive_flags" if recursive else "flags"].add(flag)
+
+    def is_(self, key: Key, flag: int) -> bool:
+        if not key:
+            return False  # document root has no flags
+        cont = self._flags
+        for k in key[:-1]:
+            if k not in cont:
+                return False
+            inner_cont = cont[k]
+            if flag in inner_cont["recursive_flags"]:
+                return True
+            cont = inner_cont["nested"]
+        key_stem = key[-1]
+        if key_stem in cont:
+            inner_cont = cont[key_stem]
+            return flag in inner_cont["flags"] or flag in inner_cont["recursive_flags"]
+        return False
+
+
+class NestedDict:
+    def __init__(self) -> None:
+        # The parsed content of the TOML document
+        self.dict: dict[str, Any] = {}
+
+    def get_or_create_nest(
+        self,
+        key: Key,
+        *,
+        access_lists: bool = True,
+    ) -> dict:
+        cont: Any = self.dict
+        for k in key:
+            if k not in cont:
+                cont[k] = {}
+            cont = cont[k]
+            if access_lists and isinstance(cont, list):
+                cont = cont[-1]
+            if not isinstance(cont, dict):
+                raise KeyError("There is no nest behind this key")
+        return cont
+
+    def append_nest_to_list(self, key: Key) -> None:
+        cont = self.get_or_create_nest(key[:-1])
+        last_key = key[-1]
+        if last_key in cont:
+            list_ = cont[last_key]
+            if not isinstance(list_, list):
+                raise KeyError("An object other than list found behind this key")
+            list_.append({})
+        else:
+            cont[last_key] = [{}]
+
+
+class Output(NamedTuple):
+    data: NestedDict
+    flags: Flags
+
+
+def skip_chars(src: str, pos: Pos, chars: Iterable[str]) -> Pos:
+    try:
+        while src[pos] in chars:
+            pos += 1
+    except IndexError:
+        pass
+    return pos
+
+
+def skip_until(
+    src: str,
+    pos: Pos,
+    expect: str,
+    *,
+    error_on: frozenset[str],
+    error_on_eof: bool,
+) -> Pos:
+    try:
+        new_pos = src.index(expect, pos)
+    except ValueError:
+        new_pos = len(src)
+        if error_on_eof:
+            raise TOMLDecodeError(f"Expected {expect!r}", src, new_pos) from None
+
+    if not error_on.isdisjoint(src[pos:new_pos]):
+        while src[pos] not in error_on:
+            pos += 1
+        raise TOMLDecodeError(f"Found invalid character {src[pos]!r}", src, pos)
+    return new_pos
+
+
+def skip_comment(src: str, pos: Pos) -> Pos:
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char == "#":
+        return skip_until(
+            src, pos + 1, "\n", error_on=ILLEGAL_COMMENT_CHARS, error_on_eof=False
+        )
+    return pos
+
+
+def skip_comments_and_array_ws(src: str, pos: Pos) -> Pos:
+    while True:
+        pos_before_skip = pos
+        pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
+        pos = skip_comment(src, pos)
+        if pos == pos_before_skip:
+            return pos
+
+
+def create_dict_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
+    pos += 1  # Skip "["
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, key = parse_key(src, pos)
+
+    if out.flags.is_(key, Flags.EXPLICIT_NEST) or out.flags.is_(key, Flags.FROZEN):
+        raise TOMLDecodeError(f"Cannot declare {key} twice", src, pos)
+    out.flags.set(key, Flags.EXPLICIT_NEST, recursive=False)
+    try:
+        out.data.get_or_create_nest(key)
+    except KeyError:
+        raise TOMLDecodeError("Cannot overwrite a value", src, pos) from None
+
+    if not src.startswith("]", pos):
+        raise TOMLDecodeError(
+            "Expected ']' at the end of a table declaration", src, pos
+        )
+    return pos + 1, key
+
+
+def create_list_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
+    pos += 2  # Skip "[["
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, key = parse_key(src, pos)
+
+    if out.flags.is_(key, Flags.FROZEN):
+        raise TOMLDecodeError(f"Cannot mutate immutable namespace {key}", src, pos)
+    # Free the namespace now that it points to another empty list item...
+    out.flags.unset_all(key)
+    # ...but this key precisely is still prohibited from table declaration
+    out.flags.set(key, Flags.EXPLICIT_NEST, recursive=False)
+    try:
+        out.data.append_nest_to_list(key)
+    except KeyError:
+        raise TOMLDecodeError("Cannot overwrite a value", src, pos) from None
+
+    if not src.startswith("]]", pos):
+        raise TOMLDecodeError(
+            "Expected ']]' at the end of an array declaration", src, pos
+        )
+    return pos + 2, key
+
+
+def key_value_rule(
+    src: str, pos: Pos, out: Output, header: Key, parse_float: ParseFloat
+) -> Pos:
+    pos, key, value = parse_key_value_pair(src, pos, parse_float, nest_lvl=0)
+    key_parent, key_stem = key[:-1], key[-1]
+    abs_key_parent = header + key_parent
+
+    relative_path_cont_keys = (header + key[:i] for i in range(1, len(key)))
+    for cont_key in relative_path_cont_keys:
+        # Check that dotted key syntax does not redefine an existing table
+        if out.flags.is_(cont_key, Flags.EXPLICIT_NEST):
+            raise TOMLDecodeError(f"Cannot redefine namespace {cont_key}", src, pos)
+        # Containers in the relative path can't be opened with the table syntax or
+        # dotted key/value syntax in following table sections.
+        out.flags.add_pending(cont_key, Flags.EXPLICIT_NEST)
+
+    if out.flags.is_(abs_key_parent, Flags.FROZEN):
+        raise TOMLDecodeError(
+            f"Cannot mutate immutable namespace {abs_key_parent}", src, pos
+        )
+
+    try:
+        nest = out.data.get_or_create_nest(abs_key_parent)
+    except KeyError:
+        raise TOMLDecodeError("Cannot overwrite a value", src, pos) from None
+    if key_stem in nest:
+        raise TOMLDecodeError("Cannot overwrite a value", src, pos)
+    # Mark inline table and array namespaces recursively immutable
+    if isinstance(value, (dict, list)):
+        out.flags.set(header + key, Flags.FROZEN, recursive=True)
+    nest[key_stem] = value
+    return pos
+
+
+def parse_key_value_pair(
+    src: str, pos: Pos, parse_float: ParseFloat, nest_lvl: int
+) -> tuple[Pos, Key, Any]:
+    pos, key = parse_key(src, pos)
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char != "=":
+        raise TOMLDecodeError("Expected '=' after a key in a key/value pair", src, pos)
+    pos += 1
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, value = parse_value(src, pos, parse_float, nest_lvl)
+    return pos, key, value
+
+
+def parse_key(src: str, pos: Pos) -> tuple[Pos, Key]:
+    pos, key_part = parse_key_part(src, pos)
+    key: Key = (key_part,)
+    pos = skip_chars(src, pos, TOML_WS)
+    while True:
+        try:
+            char: str | None = src[pos]
+        except IndexError:
+            char = None
+        if char != ".":
+            return pos, key
+        pos += 1
+        pos = skip_chars(src, pos, TOML_WS)
+        pos, key_part = parse_key_part(src, pos)
+        key += (key_part,)
+        pos = skip_chars(src, pos, TOML_WS)
+
+
+def parse_key_part(src: str, pos: Pos) -> tuple[Pos, str]:
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char in BARE_KEY_CHARS:
+        start_pos = pos
+        pos = skip_chars(src, pos, BARE_KEY_CHARS)
+        return pos, src[start_pos:pos]
+    if char == "'":
+        return parse_literal_str(src, pos)
+    if char == '"':
+        return parse_one_line_basic_str(src, pos)
+    raise TOMLDecodeError("Invalid initial character for a key part", src, pos)
+
+
+def parse_one_line_basic_str(src: str, pos: Pos) -> tuple[Pos, str]:
+    pos += 1
+    return parse_basic_str(src, pos, multiline=False)
+
+
+def parse_array(
+    src: str, pos: Pos, parse_float: ParseFloat, nest_lvl: int
+) -> tuple[Pos, list]:
+    pos += 1
+    array: list = []
+
+    pos = skip_comments_and_array_ws(src, pos)
+    if src.startswith("]", pos):
+        return pos + 1, array
+    while True:
+        pos, val = parse_value(src, pos, parse_float, nest_lvl)
+        array.append(val)
+        pos = skip_comments_and_array_ws(src, pos)
+
+        c = src[pos : pos + 1]
+        if c == "]":
+            return pos + 1, array
+        if c != ",":
+            raise TOMLDecodeError("Unclosed array", src, pos)
+        pos += 1
+
+        pos = skip_comments_and_array_ws(src, pos)
+        if src.startswith("]", pos):
+            return pos + 1, array
+
+
+def parse_inline_table(
+    src: str, pos: Pos, parse_float: ParseFloat, nest_lvl: int
+) -> tuple[Pos, dict]:
+    pos += 1
+    nested_dict = NestedDict()
+    flags = Flags()
+
+    pos = skip_chars(src, pos, TOML_WS)
+    if src.startswith("}", pos):
+        return pos + 1, nested_dict.dict
+    while True:
+        pos, key, value = parse_key_value_pair(src, pos, parse_float, nest_lvl)
+        key_parent, key_stem = key[:-1], key[-1]
+        if flags.is_(key, Flags.FROZEN):
+            raise TOMLDecodeError(f"Cannot mutate immutable namespace {key}", src, pos)
+        try:
+            nest = nested_dict.get_or_create_nest(key_parent, access_lists=False)
+        except KeyError:
+            raise TOMLDecodeError("Cannot overwrite a value", src, pos) from None
+        if key_stem in nest:
+            raise TOMLDecodeError(f"Duplicate inline table key {key_stem!r}", src, pos)
+        nest[key_stem] = value
+        pos = skip_chars(src, pos, TOML_WS)
+        c = src[pos : pos + 1]
+        if c == "}":
+            return pos + 1, nested_dict.dict
+        if c != ",":
+            raise TOMLDecodeError("Unclosed inline table", src, pos)
+        if isinstance(value, (dict, list)):
+            flags.set(key, Flags.FROZEN, recursive=True)
+        pos += 1
+        pos = skip_chars(src, pos, TOML_WS)
+
+
+def parse_basic_str_escape(
+    src: str, pos: Pos, *, multiline: bool = False
+) -> tuple[Pos, str]:
+    escape_id = src[pos : pos + 2]
+    pos += 2
+    if multiline and escape_id in {"\\ ", "\\\t", "\\\n"}:
+        # Skip whitespace until next non-whitespace character or end of
+        # the doc. Error if non-whitespace is found before newline.
+        if escape_id != "\\\n":
+            pos = skip_chars(src, pos, TOML_WS)
+            try:
+                char = src[pos]
+            except IndexError:
+                return pos, ""
+            if char != "\n":
+                raise TOMLDecodeError("Unescaped '\\' in a string", src, pos)
+            pos += 1
+        pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
+        return pos, ""
+    if escape_id == "\\u":
+        return parse_hex_char(src, pos, 4)
+    if escape_id == "\\U":
+        return parse_hex_char(src, pos, 8)
+    try:
+        return pos, BASIC_STR_ESCAPE_REPLACEMENTS[escape_id]
+    except KeyError:
+        raise TOMLDecodeError("Unescaped '\\' in a string", src, pos) from None
+
+
+def parse_basic_str_escape_multiline(src: str, pos: Pos) -> tuple[Pos, str]:
+    return parse_basic_str_escape(src, pos, multiline=True)
+
+
+def parse_hex_char(src: str, pos: Pos, hex_len: int) -> tuple[Pos, str]:
+    hex_str = src[pos : pos + hex_len]
+    if len(hex_str) != hex_len or not HEXDIGIT_CHARS.issuperset(hex_str):
+        raise TOMLDecodeError("Invalid hex value", src, pos)
+    pos += hex_len
+    hex_int = int(hex_str, 16)
+    if not is_unicode_scalar_value(hex_int):
+        raise TOMLDecodeError(
+            "Escaped character is not a Unicode scalar value", src, pos
+        )
+    return pos, chr(hex_int)
+
+
+def parse_literal_str(src: str, pos: Pos) -> tuple[Pos, str]:
+    pos += 1  # Skip starting apostrophe
+    start_pos = pos
+    pos = skip_until(
+        src, pos, "'", error_on=ILLEGAL_LITERAL_STR_CHARS, error_on_eof=True
+    )
+    return pos + 1, src[start_pos:pos]  # Skip ending apostrophe
+
+
+def parse_multiline_str(src: str, pos: Pos, *, literal: bool) -> tuple[Pos, str]:
+    pos += 3
+    if src.startswith("\n", pos):
+        pos += 1
+
+    if literal:
+        delim = "'"
+        end_pos = skip_until(
+            src,
+            pos,
+            "'''",
+            error_on=ILLEGAL_MULTILINE_LITERAL_STR_CHARS,
+            error_on_eof=True,
+        )
+        result = src[pos:end_pos]
+        pos = end_pos + 3
+    else:
+        delim = '"'
+        pos, result = parse_basic_str(src, pos, multiline=True)
+
+    # Add at maximum two extra apostrophes/quotes if the end sequence
+    # is 4 or 5 chars long instead of just 3.
+    if not src.startswith(delim, pos):
+        return pos, result
+    pos += 1
+    if not src.startswith(delim, pos):
+        return pos, result + delim
+    pos += 1
+    return pos, result + (delim * 2)
+
+
+def parse_basic_str(src: str, pos: Pos, *, multiline: bool) -> tuple[Pos, str]:
+    if multiline:
+        error_on = ILLEGAL_MULTILINE_BASIC_STR_CHARS
+        parse_escapes = parse_basic_str_escape_multiline
+    else:
+        error_on = ILLEGAL_BASIC_STR_CHARS
+        parse_escapes = parse_basic_str_escape
+    result = ""
+    start_pos = pos
+    while True:
+        try:
+            char = src[pos]
+        except IndexError:
+            raise TOMLDecodeError("Unterminated string", src, pos) from None
+        if char == '"':
+            if not multiline:
+                return pos + 1, result + src[start_pos:pos]
+            if src.startswith('"""', pos):
+                return pos + 3, result + src[start_pos:pos]
+            pos += 1
+            continue
+        if char == "\\":
+            result += src[start_pos:pos]
+            pos, parsed_escape = parse_escapes(src, pos)
+            result += parsed_escape
+            start_pos = pos
+            continue
+        if char in error_on:
+            raise TOMLDecodeError(f"Illegal character {char!r}", src, pos)
+        pos += 1
+
+
+def parse_value(  # noqa: C901
+    src: str, pos: Pos, parse_float: ParseFloat, nest_lvl: int
+) -> tuple[Pos, Any]:
+    if nest_lvl > MAX_INLINE_NESTING:
+        # Pure Python should have raised RecursionError already.
+        # This ensures mypyc binaries eventually do the same.
+        raise RecursionError(  # pragma: no cover
+            "TOML inline arrays/tables are nested more than the allowed"
+            f" {MAX_INLINE_NESTING} levels"
+        )
+
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+
+    # IMPORTANT: order conditions based on speed of checking and likelihood
+
+    # Basic strings
+    if char == '"':
+        if src.startswith('"""', pos):
+            return parse_multiline_str(src, pos, literal=False)
+        return parse_one_line_basic_str(src, pos)
+
+    # Literal strings
+    if char == "'":
+        if src.startswith("'''", pos):
+            return parse_multiline_str(src, pos, literal=True)
+        return parse_literal_str(src, pos)
+
+    # Booleans
+    if char == "t":
+        if src.startswith("true", pos):
+            return pos + 4, True
+    if char == "f":
+        if src.startswith("false", pos):
+            return pos + 5, False
+
+    # Arrays
+    if char == "[":
+        return parse_array(src, pos, parse_float, nest_lvl + 1)
+
+    # Inline tables
+    if char == "{":
+        return parse_inline_table(src, pos, parse_float, nest_lvl + 1)
+
+    # Dates and times
+    datetime_match = RE_DATETIME.match(src, pos)
+    if datetime_match:
+        try:
+            datetime_obj = match_to_datetime(datetime_match)
+        except ValueError as e:
+            raise TOMLDecodeError("Invalid date or datetime", src, pos) from e
+        return datetime_match.end(), datetime_obj
+    localtime_match = RE_LOCALTIME.match(src, pos)
+    if localtime_match:
+        return localtime_match.end(), match_to_localtime(localtime_match)
+
+    # Integers and "normal" floats.
+    # The regex will greedily match any type starting with a decimal
+    # char, so needs to be located after handling of dates and times.
+    number_match = RE_NUMBER.match(src, pos)
+    if number_match:
+        return number_match.end(), match_to_number(number_match, parse_float)
+
+    # Special floats
+    first_three = src[pos : pos + 3]
+    if first_three in {"inf", "nan"}:
+        return pos + 3, parse_float(first_three)
+    first_four = src[pos : pos + 4]
+    if first_four in {"-inf", "+inf", "-nan", "+nan"}:
+        return pos + 4, parse_float(first_four)
+
+    raise TOMLDecodeError("Invalid value", src, pos)
+
+
+def is_unicode_scalar_value(codepoint: int) -> bool:
+    return (0 <= codepoint <= 55295) or (57344 <= codepoint <= 1114111)
+
+
+def make_safe_parse_float(parse_float: ParseFloat) -> ParseFloat:
+    """A decorator to make `parse_float` safe.
+
+    `parse_float` must not return dicts or lists, because these types
+    would be mixed with parsed TOML tables and arrays, thus confusing
+    the parser. The returned decorated callable raises `ValueError`
+    instead of returning illegal types.
+    """
+    # The default `float` callable never returns illegal types. Optimize it.
+    if parse_float is float:
+        return float
+
+    def safe_parse_float(float_str: str) -> Any:
+        float_value = parse_float(float_str)
+        if isinstance(float_value, (dict, list)):
+            raise ValueError("parse_float must not return dicts or lists")
+        return float_value
+
+    return safe_parse_float
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_re.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_re.py
new file mode 100644
index 00000000..51348661
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_re.py
@@ -0,0 +1,112 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from __future__ import annotations
+
+from datetime import date, datetime, time, timedelta, timezone, tzinfo
+from functools import lru_cache
+import re
+from typing import Any, Final
+
+from ._types import ParseFloat
+
+# E.g.
+# - 00:32:00.999999
+# - 00:32:00
+_TIME_RE_STR: Final = (
+    r"([01][0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(?:\.([0-9]{1,6})[0-9]*)?"
+)
+
+RE_NUMBER: Final = re.compile(
+    r"""
+0
+(?:
+    x[0-9A-Fa-f](?:_?[0-9A-Fa-f])*   # hex
+    |
+    b[01](?:_?[01])*                 # bin
+    |
+    o[0-7](?:_?[0-7])*               # oct
+)
+|
+[+-]?(?:0|[1-9](?:_?[0-9])*)         # dec, integer part
+(?P<floatpart>
+    (?:\.[0-9](?:_?[0-9])*)?         # optional fractional part
+    (?:[eE][+-]?[0-9](?:_?[0-9])*)?  # optional exponent part
+)
+""",
+    flags=re.VERBOSE,
+)
+RE_LOCALTIME: Final = re.compile(_TIME_RE_STR)
+RE_DATETIME: Final = re.compile(
+    rf"""
+([0-9]{{4}})-(0[1-9]|1[0-2])-(0[1-9]|[12][0-9]|3[01])  # date, e.g. 1988-10-27
+(?:
+    [Tt ]
+    {_TIME_RE_STR}
+    (?:([Zz])|([+-])([01][0-9]|2[0-3]):([0-5][0-9]))?  # optional time offset
+)?
+""",
+    flags=re.VERBOSE,
+)
+
+
+def match_to_datetime(match: re.Match) -> datetime | date:
+    """Convert a `RE_DATETIME` match to `datetime.datetime` or `datetime.date`.
+
+    Raises ValueError if the match does not correspond to a valid date
+    or datetime.
+    """
+    (
+        year_str,
+        month_str,
+        day_str,
+        hour_str,
+        minute_str,
+        sec_str,
+        micros_str,
+        zulu_time,
+        offset_sign_str,
+        offset_hour_str,
+        offset_minute_str,
+    ) = match.groups()
+    year, month, day = int(year_str), int(month_str), int(day_str)
+    if hour_str is None:
+        return date(year, month, day)
+    hour, minute, sec = int(hour_str), int(minute_str), int(sec_str)
+    micros = int(micros_str.ljust(6, "0")) if micros_str else 0
+    if offset_sign_str:
+        tz: tzinfo | None = cached_tz(
+            offset_hour_str, offset_minute_str, offset_sign_str
+        )
+    elif zulu_time:
+        tz = timezone.utc
+    else:  # local date-time
+        tz = None
+    return datetime(year, month, day, hour, minute, sec, micros, tzinfo=tz)
+
+
+# No need to limit cache size. This is only ever called on input
+# that matched RE_DATETIME, so there is an implicit bound of
+# 24 (hours) * 60 (minutes) * 2 (offset direction) = 2880.
+@lru_cache(maxsize=None)
+def cached_tz(hour_str: str, minute_str: str, sign_str: str) -> timezone:
+    sign = 1 if sign_str == "+" else -1
+    return timezone(
+        timedelta(
+            hours=sign * int(hour_str),
+            minutes=sign * int(minute_str),
+        )
+    )
+
+
+def match_to_localtime(match: re.Match) -> time:
+    hour_str, minute_str, sec_str, micros_str = match.groups()
+    micros = int(micros_str.ljust(6, "0")) if micros_str else 0
+    return time(int(hour_str), int(minute_str), int(sec_str), micros)
+
+
+def match_to_number(match: re.Match, parse_float: ParseFloat) -> Any:
+    if match.group("floatpart"):
+        return parse_float(match.group())
+    return int(match.group(), 0)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_types.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_types.py
new file mode 100644
index 00000000..d949412e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/_types.py
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from typing import Any, Callable, Tuple
+
+# Type annotations
+ParseFloat = Callable[[str], Any]
+Key = Tuple[str, ...]
+Pos = int
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/py.typed
new file mode 100644
index 00000000..7632ecf7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/tomli/py.typed
@@ -0,0 +1 @@
+# Marker file for PEP 561
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__init__.py
new file mode 100644
index 00000000..e468bf8c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__init__.py
@@ -0,0 +1,36 @@
+"""Verify certificates using native system trust stores"""
+
+import sys as _sys
+
+if _sys.version_info < (3, 10):
+    raise ImportError("truststore requires Python 3.10 or later")
+
+# Detect Python runtimes which don't implement SSLObject.get_unverified_chain() API
+# This API only became public in Python 3.13 but was available in CPython and PyPy since 3.10.
+if _sys.version_info < (3, 13):
+    try:
+        import ssl as _ssl
+    except ImportError:
+        raise ImportError("truststore requires the 'ssl' module")
+    else:
+        _sslmem = _ssl.MemoryBIO()
+        _sslobj = _ssl.create_default_context().wrap_bio(
+            _sslmem,
+            _sslmem,
+        )
+        try:
+            while not hasattr(_sslobj, "get_unverified_chain"):
+                _sslobj = _sslobj._sslobj  # type: ignore[attr-defined]
+        except AttributeError:
+            raise ImportError(
+                "truststore requires peer certificate chain APIs to be available"
+            ) from None
+
+        del _ssl, _sslobj, _sslmem  # noqa: F821
+
+from ._api import SSLContext, extract_from_ssl, inject_into_ssl  # noqa: E402
+
+del _api, _sys  # type: ignore[name-defined] # noqa: F821
+
+__all__ = ["SSLContext", "inject_into_ssl", "extract_from_ssl"]
+__version__ = "0.10.0"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9a52b496
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_api.cpython-312.pyc
new file mode 100644
index 00000000..bff73913
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_macos.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_macos.cpython-312.pyc
new file mode 100644
index 00000000..96bb09b2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_macos.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_openssl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_openssl.cpython-312.pyc
new file mode 100644
index 00000000..29748231
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_openssl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_ssl_constants.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_ssl_constants.cpython-312.pyc
new file mode 100644
index 00000000..cd01ded6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_ssl_constants.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_windows.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_windows.cpython-312.pyc
new file mode 100644
index 00000000..ecb166e7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/__pycache__/_windows.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_api.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_api.py
new file mode 100644
index 00000000..aeb023af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_api.py
@@ -0,0 +1,316 @@
+import os
+import platform
+import socket
+import ssl
+import sys
+import typing
+
+import _ssl  # type: ignore[import-not-found]
+
+from ._ssl_constants import (
+    _original_SSLContext,
+    _original_super_SSLContext,
+    _truststore_SSLContext_dunder_class,
+    _truststore_SSLContext_super_class,
+)
+
+if platform.system() == "Windows":
+    from ._windows import _configure_context, _verify_peercerts_impl
+elif platform.system() == "Darwin":
+    from ._macos import _configure_context, _verify_peercerts_impl
+else:
+    from ._openssl import _configure_context, _verify_peercerts_impl
+
+if typing.TYPE_CHECKING:
+    from pip._vendor.typing_extensions import Buffer
+
+# From typeshed/stdlib/ssl.pyi
+_StrOrBytesPath: typing.TypeAlias = str | bytes | os.PathLike[str] | os.PathLike[bytes]
+_PasswordType: typing.TypeAlias = str | bytes | typing.Callable[[], str | bytes]
+
+
+def inject_into_ssl() -> None:
+    """Injects the :class:`truststore.SSLContext` into the ``ssl``
+    module by replacing :class:`ssl.SSLContext`.
+    """
+    setattr(ssl, "SSLContext", SSLContext)
+    # urllib3 holds on to its own reference of ssl.SSLContext
+    # so we need to replace that reference too.
+    try:
+        import pip._vendor.urllib3.util.ssl_ as urllib3_ssl
+
+        setattr(urllib3_ssl, "SSLContext", SSLContext)
+    except ImportError:
+        pass
+
+
+def extract_from_ssl() -> None:
+    """Restores the :class:`ssl.SSLContext` class to its original state"""
+    setattr(ssl, "SSLContext", _original_SSLContext)
+    try:
+        import pip._vendor.urllib3.util.ssl_ as urllib3_ssl
+
+        urllib3_ssl.SSLContext = _original_SSLContext  # type: ignore[assignment]
+    except ImportError:
+        pass
+
+
+class SSLContext(_truststore_SSLContext_super_class):  # type: ignore[misc]
+    """SSLContext API that uses system certificates on all platforms"""
+
+    @property  # type: ignore[misc]
+    def __class__(self) -> type:
+        # Dirty hack to get around isinstance() checks
+        # for ssl.SSLContext instances in aiohttp/trustme
+        # when using non-CPython implementations.
+        return _truststore_SSLContext_dunder_class or SSLContext
+
+    def __init__(self, protocol: int = None) -> None:  # type: ignore[assignment]
+        self._ctx = _original_SSLContext(protocol)
+
+        class TruststoreSSLObject(ssl.SSLObject):
+            # This object exists because wrap_bio() doesn't
+            # immediately do the handshake so we need to do
+            # certificate verifications after SSLObject.do_handshake()
+
+            def do_handshake(self) -> None:
+                ret = super().do_handshake()
+                _verify_peercerts(self, server_hostname=self.server_hostname)
+                return ret
+
+        self._ctx.sslobject_class = TruststoreSSLObject
+
+    def wrap_socket(
+        self,
+        sock: socket.socket,
+        server_side: bool = False,
+        do_handshake_on_connect: bool = True,
+        suppress_ragged_eofs: bool = True,
+        server_hostname: str | None = None,
+        session: ssl.SSLSession | None = None,
+    ) -> ssl.SSLSocket:
+        # Use a context manager here because the
+        # inner SSLContext holds on to our state
+        # but also does the actual handshake.
+        with _configure_context(self._ctx):
+            ssl_sock = self._ctx.wrap_socket(
+                sock,
+                server_side=server_side,
+                server_hostname=server_hostname,
+                do_handshake_on_connect=do_handshake_on_connect,
+                suppress_ragged_eofs=suppress_ragged_eofs,
+                session=session,
+            )
+        try:
+            _verify_peercerts(ssl_sock, server_hostname=server_hostname)
+        except Exception:
+            ssl_sock.close()
+            raise
+        return ssl_sock
+
+    def wrap_bio(
+        self,
+        incoming: ssl.MemoryBIO,
+        outgoing: ssl.MemoryBIO,
+        server_side: bool = False,
+        server_hostname: str | None = None,
+        session: ssl.SSLSession | None = None,
+    ) -> ssl.SSLObject:
+        with _configure_context(self._ctx):
+            ssl_obj = self._ctx.wrap_bio(
+                incoming,
+                outgoing,
+                server_hostname=server_hostname,
+                server_side=server_side,
+                session=session,
+            )
+        return ssl_obj
+
+    def load_verify_locations(
+        self,
+        cafile: str | bytes | os.PathLike[str] | os.PathLike[bytes] | None = None,
+        capath: str | bytes | os.PathLike[str] | os.PathLike[bytes] | None = None,
+        cadata: typing.Union[str, "Buffer", None] = None,
+    ) -> None:
+        return self._ctx.load_verify_locations(
+            cafile=cafile, capath=capath, cadata=cadata
+        )
+
+    def load_cert_chain(
+        self,
+        certfile: _StrOrBytesPath,
+        keyfile: _StrOrBytesPath | None = None,
+        password: _PasswordType | None = None,
+    ) -> None:
+        return self._ctx.load_cert_chain(
+            certfile=certfile, keyfile=keyfile, password=password
+        )
+
+    def load_default_certs(
+        self, purpose: ssl.Purpose = ssl.Purpose.SERVER_AUTH
+    ) -> None:
+        return self._ctx.load_default_certs(purpose)
+
+    def set_alpn_protocols(self, alpn_protocols: typing.Iterable[str]) -> None:
+        return self._ctx.set_alpn_protocols(alpn_protocols)
+
+    def set_npn_protocols(self, npn_protocols: typing.Iterable[str]) -> None:
+        return self._ctx.set_npn_protocols(npn_protocols)
+
+    def set_ciphers(self, __cipherlist: str) -> None:
+        return self._ctx.set_ciphers(__cipherlist)
+
+    def get_ciphers(self) -> typing.Any:
+        return self._ctx.get_ciphers()
+
+    def session_stats(self) -> dict[str, int]:
+        return self._ctx.session_stats()
+
+    def cert_store_stats(self) -> dict[str, int]:
+        raise NotImplementedError()
+
+    def set_default_verify_paths(self) -> None:
+        self._ctx.set_default_verify_paths()
+
+    @typing.overload
+    def get_ca_certs(
+        self, binary_form: typing.Literal[False] = ...
+    ) -> list[typing.Any]: ...
+
+    @typing.overload
+    def get_ca_certs(self, binary_form: typing.Literal[True] = ...) -> list[bytes]: ...
+
+    @typing.overload
+    def get_ca_certs(self, binary_form: bool = ...) -> typing.Any: ...
+
+    def get_ca_certs(self, binary_form: bool = False) -> list[typing.Any] | list[bytes]:
+        raise NotImplementedError()
+
+    @property
+    def check_hostname(self) -> bool:
+        return self._ctx.check_hostname
+
+    @check_hostname.setter
+    def check_hostname(self, value: bool) -> None:
+        self._ctx.check_hostname = value
+
+    @property
+    def hostname_checks_common_name(self) -> bool:
+        return self._ctx.hostname_checks_common_name
+
+    @hostname_checks_common_name.setter
+    def hostname_checks_common_name(self, value: bool) -> None:
+        self._ctx.hostname_checks_common_name = value
+
+    @property
+    def keylog_filename(self) -> str:
+        return self._ctx.keylog_filename
+
+    @keylog_filename.setter
+    def keylog_filename(self, value: str) -> None:
+        self._ctx.keylog_filename = value
+
+    @property
+    def maximum_version(self) -> ssl.TLSVersion:
+        return self._ctx.maximum_version
+
+    @maximum_version.setter
+    def maximum_version(self, value: ssl.TLSVersion) -> None:
+        _original_super_SSLContext.maximum_version.__set__(  # type: ignore[attr-defined]
+            self._ctx, value
+        )
+
+    @property
+    def minimum_version(self) -> ssl.TLSVersion:
+        return self._ctx.minimum_version
+
+    @minimum_version.setter
+    def minimum_version(self, value: ssl.TLSVersion) -> None:
+        _original_super_SSLContext.minimum_version.__set__(  # type: ignore[attr-defined]
+            self._ctx, value
+        )
+
+    @property
+    def options(self) -> ssl.Options:
+        return self._ctx.options
+
+    @options.setter
+    def options(self, value: ssl.Options) -> None:
+        _original_super_SSLContext.options.__set__(  # type: ignore[attr-defined]
+            self._ctx, value
+        )
+
+    @property
+    def post_handshake_auth(self) -> bool:
+        return self._ctx.post_handshake_auth
+
+    @post_handshake_auth.setter
+    def post_handshake_auth(self, value: bool) -> None:
+        self._ctx.post_handshake_auth = value
+
+    @property
+    def protocol(self) -> ssl._SSLMethod:
+        return self._ctx.protocol
+
+    @property
+    def security_level(self) -> int:
+        return self._ctx.security_level
+
+    @property
+    def verify_flags(self) -> ssl.VerifyFlags:
+        return self._ctx.verify_flags
+
+    @verify_flags.setter
+    def verify_flags(self, value: ssl.VerifyFlags) -> None:
+        _original_super_SSLContext.verify_flags.__set__(  # type: ignore[attr-defined]
+            self._ctx, value
+        )
+
+    @property
+    def verify_mode(self) -> ssl.VerifyMode:
+        return self._ctx.verify_mode
+
+    @verify_mode.setter
+    def verify_mode(self, value: ssl.VerifyMode) -> None:
+        _original_super_SSLContext.verify_mode.__set__(  # type: ignore[attr-defined]
+            self._ctx, value
+        )
+
+
+# Python 3.13+ makes get_unverified_chain() a public API that only returns DER
+# encoded certificates. We detect whether we need to call public_bytes() for 3.10->3.12
+# Pre-3.13 returned None instead of an empty list from get_unverified_chain()
+if sys.version_info >= (3, 13):
+
+    def _get_unverified_chain_bytes(sslobj: ssl.SSLObject) -> list[bytes]:
+        unverified_chain = sslobj.get_unverified_chain() or ()  # type: ignore[attr-defined]
+        return [
+            cert if isinstance(cert, bytes) else cert.public_bytes(_ssl.ENCODING_DER)
+            for cert in unverified_chain
+        ]
+
+else:
+
+    def _get_unverified_chain_bytes(sslobj: ssl.SSLObject) -> list[bytes]:
+        unverified_chain = sslobj.get_unverified_chain() or ()  # type: ignore[attr-defined]
+        return [cert.public_bytes(_ssl.ENCODING_DER) for cert in unverified_chain]
+
+
+def _verify_peercerts(
+    sock_or_sslobj: ssl.SSLSocket | ssl.SSLObject, server_hostname: str | None
+) -> None:
+    """
+    Verifies the peer certificates from an SSLSocket or SSLObject
+    against the certificates in the OS trust store.
+    """
+    sslobj: ssl.SSLObject = sock_or_sslobj  # type: ignore[assignment]
+    try:
+        while not hasattr(sslobj, "get_unverified_chain"):
+            sslobj = sslobj._sslobj  # type: ignore[attr-defined]
+    except AttributeError:
+        pass
+
+    cert_bytes = _get_unverified_chain_bytes(sslobj)
+    _verify_peercerts_impl(
+        sock_or_sslobj.context, cert_bytes, server_hostname=server_hostname
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_macos.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_macos.py
new file mode 100644
index 00000000..34503077
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_macos.py
@@ -0,0 +1,571 @@
+import contextlib
+import ctypes
+import platform
+import ssl
+import typing
+from ctypes import (
+    CDLL,
+    POINTER,
+    c_bool,
+    c_char_p,
+    c_int32,
+    c_long,
+    c_uint32,
+    c_ulong,
+    c_void_p,
+)
+from ctypes.util import find_library
+
+from ._ssl_constants import _set_ssl_context_verify_mode
+
+_mac_version = platform.mac_ver()[0]
+_mac_version_info = tuple(map(int, _mac_version.split(".")))
+if _mac_version_info < (10, 8):
+    raise ImportError(
+        f"Only OS X 10.8 and newer are supported, not {_mac_version_info[0]}.{_mac_version_info[1]}"
+    )
+
+_is_macos_version_10_14_or_later = _mac_version_info >= (10, 14)
+
+
+def _load_cdll(name: str, macos10_16_path: str) -> CDLL:
+    """Loads a CDLL by name, falling back to known path on 10.16+"""
+    try:
+        # Big Sur is technically 11 but we use 10.16 due to the Big Sur
+        # beta being labeled as 10.16.
+        path: str | None
+        if _mac_version_info >= (10, 16):
+            path = macos10_16_path
+        else:
+            path = find_library(name)
+        if not path:
+            raise OSError  # Caught and reraised as 'ImportError'
+        return CDLL(path, use_errno=True)
+    except OSError:
+        raise ImportError(f"The library {name} failed to load") from None
+
+
+Security = _load_cdll(
+    "Security", "/System/Library/Frameworks/Security.framework/Security"
+)
+CoreFoundation = _load_cdll(
+    "CoreFoundation",
+    "/System/Library/Frameworks/CoreFoundation.framework/CoreFoundation",
+)
+
+Boolean = c_bool
+CFIndex = c_long
+CFStringEncoding = c_uint32
+CFData = c_void_p
+CFString = c_void_p
+CFArray = c_void_p
+CFMutableArray = c_void_p
+CFError = c_void_p
+CFType = c_void_p
+CFTypeID = c_ulong
+CFTypeRef = POINTER(CFType)
+CFAllocatorRef = c_void_p
+
+OSStatus = c_int32
+
+CFErrorRef = POINTER(CFError)
+CFDataRef = POINTER(CFData)
+CFStringRef = POINTER(CFString)
+CFArrayRef = POINTER(CFArray)
+CFMutableArrayRef = POINTER(CFMutableArray)
+CFArrayCallBacks = c_void_p
+CFOptionFlags = c_uint32
+
+SecCertificateRef = POINTER(c_void_p)
+SecPolicyRef = POINTER(c_void_p)
+SecTrustRef = POINTER(c_void_p)
+SecTrustResultType = c_uint32
+SecTrustOptionFlags = c_uint32
+
+try:
+    Security.SecCertificateCreateWithData.argtypes = [CFAllocatorRef, CFDataRef]
+    Security.SecCertificateCreateWithData.restype = SecCertificateRef
+
+    Security.SecCertificateCopyData.argtypes = [SecCertificateRef]
+    Security.SecCertificateCopyData.restype = CFDataRef
+
+    Security.SecCopyErrorMessageString.argtypes = [OSStatus, c_void_p]
+    Security.SecCopyErrorMessageString.restype = CFStringRef
+
+    Security.SecTrustSetAnchorCertificates.argtypes = [SecTrustRef, CFArrayRef]
+    Security.SecTrustSetAnchorCertificates.restype = OSStatus
+
+    Security.SecTrustSetAnchorCertificatesOnly.argtypes = [SecTrustRef, Boolean]
+    Security.SecTrustSetAnchorCertificatesOnly.restype = OSStatus
+
+    Security.SecPolicyCreateRevocation.argtypes = [CFOptionFlags]
+    Security.SecPolicyCreateRevocation.restype = SecPolicyRef
+
+    Security.SecPolicyCreateSSL.argtypes = [Boolean, CFStringRef]
+    Security.SecPolicyCreateSSL.restype = SecPolicyRef
+
+    Security.SecTrustCreateWithCertificates.argtypes = [
+        CFTypeRef,
+        CFTypeRef,
+        POINTER(SecTrustRef),
+    ]
+    Security.SecTrustCreateWithCertificates.restype = OSStatus
+
+    Security.SecTrustGetTrustResult.argtypes = [
+        SecTrustRef,
+        POINTER(SecTrustResultType),
+    ]
+    Security.SecTrustGetTrustResult.restype = OSStatus
+
+    Security.SecTrustEvaluate.argtypes = [
+        SecTrustRef,
+        POINTER(SecTrustResultType),
+    ]
+    Security.SecTrustEvaluate.restype = OSStatus
+
+    Security.SecTrustRef = SecTrustRef  # type: ignore[attr-defined]
+    Security.SecTrustResultType = SecTrustResultType  # type: ignore[attr-defined]
+    Security.OSStatus = OSStatus  # type: ignore[attr-defined]
+
+    kSecRevocationUseAnyAvailableMethod = 3
+    kSecRevocationRequirePositiveResponse = 8
+
+    CoreFoundation.CFRelease.argtypes = [CFTypeRef]
+    CoreFoundation.CFRelease.restype = None
+
+    CoreFoundation.CFGetTypeID.argtypes = [CFTypeRef]
+    CoreFoundation.CFGetTypeID.restype = CFTypeID
+
+    CoreFoundation.CFStringCreateWithCString.argtypes = [
+        CFAllocatorRef,
+        c_char_p,
+        CFStringEncoding,
+    ]
+    CoreFoundation.CFStringCreateWithCString.restype = CFStringRef
+
+    CoreFoundation.CFStringGetCStringPtr.argtypes = [CFStringRef, CFStringEncoding]
+    CoreFoundation.CFStringGetCStringPtr.restype = c_char_p
+
+    CoreFoundation.CFStringGetCString.argtypes = [
+        CFStringRef,
+        c_char_p,
+        CFIndex,
+        CFStringEncoding,
+    ]
+    CoreFoundation.CFStringGetCString.restype = c_bool
+
+    CoreFoundation.CFDataCreate.argtypes = [CFAllocatorRef, c_char_p, CFIndex]
+    CoreFoundation.CFDataCreate.restype = CFDataRef
+
+    CoreFoundation.CFDataGetLength.argtypes = [CFDataRef]
+    CoreFoundation.CFDataGetLength.restype = CFIndex
+
+    CoreFoundation.CFDataGetBytePtr.argtypes = [CFDataRef]
+    CoreFoundation.CFDataGetBytePtr.restype = c_void_p
+
+    CoreFoundation.CFArrayCreate.argtypes = [
+        CFAllocatorRef,
+        POINTER(CFTypeRef),
+        CFIndex,
+        CFArrayCallBacks,
+    ]
+    CoreFoundation.CFArrayCreate.restype = CFArrayRef
+
+    CoreFoundation.CFArrayCreateMutable.argtypes = [
+        CFAllocatorRef,
+        CFIndex,
+        CFArrayCallBacks,
+    ]
+    CoreFoundation.CFArrayCreateMutable.restype = CFMutableArrayRef
+
+    CoreFoundation.CFArrayAppendValue.argtypes = [CFMutableArrayRef, c_void_p]
+    CoreFoundation.CFArrayAppendValue.restype = None
+
+    CoreFoundation.CFArrayGetCount.argtypes = [CFArrayRef]
+    CoreFoundation.CFArrayGetCount.restype = CFIndex
+
+    CoreFoundation.CFArrayGetValueAtIndex.argtypes = [CFArrayRef, CFIndex]
+    CoreFoundation.CFArrayGetValueAtIndex.restype = c_void_p
+
+    CoreFoundation.CFErrorGetCode.argtypes = [CFErrorRef]
+    CoreFoundation.CFErrorGetCode.restype = CFIndex
+
+    CoreFoundation.CFErrorCopyDescription.argtypes = [CFErrorRef]
+    CoreFoundation.CFErrorCopyDescription.restype = CFStringRef
+
+    CoreFoundation.kCFAllocatorDefault = CFAllocatorRef.in_dll(  # type: ignore[attr-defined]
+        CoreFoundation, "kCFAllocatorDefault"
+    )
+    CoreFoundation.kCFTypeArrayCallBacks = c_void_p.in_dll(  # type: ignore[attr-defined]
+        CoreFoundation, "kCFTypeArrayCallBacks"
+    )
+
+    CoreFoundation.CFTypeRef = CFTypeRef  # type: ignore[attr-defined]
+    CoreFoundation.CFArrayRef = CFArrayRef  # type: ignore[attr-defined]
+    CoreFoundation.CFStringRef = CFStringRef  # type: ignore[attr-defined]
+    CoreFoundation.CFErrorRef = CFErrorRef  # type: ignore[attr-defined]
+
+except AttributeError as e:
+    raise ImportError(f"Error initializing ctypes: {e}") from None
+
+# SecTrustEvaluateWithError is macOS 10.14+
+if _is_macos_version_10_14_or_later:
+    try:
+        Security.SecTrustEvaluateWithError.argtypes = [
+            SecTrustRef,
+            POINTER(CFErrorRef),
+        ]
+        Security.SecTrustEvaluateWithError.restype = c_bool
+    except AttributeError as e:
+        raise ImportError(f"Error initializing ctypes: {e}") from None
+
+
+def _handle_osstatus(result: OSStatus, _: typing.Any, args: typing.Any) -> typing.Any:
+    """
+    Raises an error if the OSStatus value is non-zero.
+    """
+    if int(result) == 0:
+        return args
+
+    # Returns a CFString which we need to transform
+    # into a UTF-8 Python string.
+    error_message_cfstring = None
+    try:
+        error_message_cfstring = Security.SecCopyErrorMessageString(result, None)
+
+        # First step is convert the CFString into a C string pointer.
+        # We try the fast no-copy way first.
+        error_message_cfstring_c_void_p = ctypes.cast(
+            error_message_cfstring, ctypes.POINTER(ctypes.c_void_p)
+        )
+        message = CoreFoundation.CFStringGetCStringPtr(
+            error_message_cfstring_c_void_p, CFConst.kCFStringEncodingUTF8
+        )
+
+        # Quoting the Apple dev docs:
+        #
+        # "A pointer to a C string or NULL if the internal
+        # storage of theString does not allow this to be
+        # returned efficiently."
+        #
+        # So we need to get our hands dirty.
+        if message is None:
+            buffer = ctypes.create_string_buffer(1024)
+            result = CoreFoundation.CFStringGetCString(
+                error_message_cfstring_c_void_p,
+                buffer,
+                1024,
+                CFConst.kCFStringEncodingUTF8,
+            )
+            if not result:
+                raise OSError("Error copying C string from CFStringRef")
+            message = buffer.value
+
+    finally:
+        if error_message_cfstring is not None:
+            CoreFoundation.CFRelease(error_message_cfstring)
+
+    # If no message can be found for this status we come
+    # up with a generic one that forwards the status code.
+    if message is None or message == "":
+        message = f"SecureTransport operation returned a non-zero OSStatus: {result}"
+
+    raise ssl.SSLError(message)
+
+
+Security.SecTrustCreateWithCertificates.errcheck = _handle_osstatus  # type: ignore[assignment]
+Security.SecTrustSetAnchorCertificates.errcheck = _handle_osstatus  # type: ignore[assignment]
+Security.SecTrustSetAnchorCertificatesOnly.errcheck = _handle_osstatus  # type: ignore[assignment]
+Security.SecTrustGetTrustResult.errcheck = _handle_osstatus  # type: ignore[assignment]
+Security.SecTrustEvaluate.errcheck = _handle_osstatus  # type: ignore[assignment]
+
+
+class CFConst:
+    """CoreFoundation constants"""
+
+    kCFStringEncodingUTF8 = CFStringEncoding(0x08000100)
+
+    errSecIncompleteCertRevocationCheck = -67635
+    errSecHostNameMismatch = -67602
+    errSecCertificateExpired = -67818
+    errSecNotTrusted = -67843
+
+
+def _bytes_to_cf_data_ref(value: bytes) -> CFDataRef:  # type: ignore[valid-type]
+    return CoreFoundation.CFDataCreate(  # type: ignore[no-any-return]
+        CoreFoundation.kCFAllocatorDefault, value, len(value)
+    )
+
+
+def _bytes_to_cf_string(value: bytes) -> CFString:
+    """
+    Given a Python binary data, create a CFString.
+    The string must be CFReleased by the caller.
+    """
+    c_str = ctypes.c_char_p(value)
+    cf_str = CoreFoundation.CFStringCreateWithCString(
+        CoreFoundation.kCFAllocatorDefault,
+        c_str,
+        CFConst.kCFStringEncodingUTF8,
+    )
+    return cf_str  # type: ignore[no-any-return]
+
+
+def _cf_string_ref_to_str(cf_string_ref: CFStringRef) -> str | None:  # type: ignore[valid-type]
+    """
+    Creates a Unicode string from a CFString object. Used entirely for error
+    reporting.
+    Yes, it annoys me quite a lot that this function is this complex.
+    """
+
+    string = CoreFoundation.CFStringGetCStringPtr(
+        cf_string_ref, CFConst.kCFStringEncodingUTF8
+    )
+    if string is None:
+        buffer = ctypes.create_string_buffer(1024)
+        result = CoreFoundation.CFStringGetCString(
+            cf_string_ref, buffer, 1024, CFConst.kCFStringEncodingUTF8
+        )
+        if not result:
+            raise OSError("Error copying C string from CFStringRef")
+        string = buffer.value
+    if string is not None:
+        string = string.decode("utf-8")
+    return string  # type: ignore[no-any-return]
+
+
+def _der_certs_to_cf_cert_array(certs: list[bytes]) -> CFMutableArrayRef:  # type: ignore[valid-type]
+    """Builds a CFArray of SecCertificateRefs from a list of DER-encoded certificates.
+    Responsibility of the caller to call CoreFoundation.CFRelease on the CFArray.
+    """
+    cf_array = CoreFoundation.CFArrayCreateMutable(
+        CoreFoundation.kCFAllocatorDefault,
+        0,
+        ctypes.byref(CoreFoundation.kCFTypeArrayCallBacks),
+    )
+    if not cf_array:
+        raise MemoryError("Unable to allocate memory!")
+
+    for cert_data in certs:
+        cf_data = None
+        sec_cert_ref = None
+        try:
+            cf_data = _bytes_to_cf_data_ref(cert_data)
+            sec_cert_ref = Security.SecCertificateCreateWithData(
+                CoreFoundation.kCFAllocatorDefault, cf_data
+            )
+            CoreFoundation.CFArrayAppendValue(cf_array, sec_cert_ref)
+        finally:
+            if cf_data:
+                CoreFoundation.CFRelease(cf_data)
+            if sec_cert_ref:
+                CoreFoundation.CFRelease(sec_cert_ref)
+
+    return cf_array  # type: ignore[no-any-return]
+
+
+@contextlib.contextmanager
+def _configure_context(ctx: ssl.SSLContext) -> typing.Iterator[None]:
+    check_hostname = ctx.check_hostname
+    verify_mode = ctx.verify_mode
+    ctx.check_hostname = False
+    _set_ssl_context_verify_mode(ctx, ssl.CERT_NONE)
+    try:
+        yield
+    finally:
+        ctx.check_hostname = check_hostname
+        _set_ssl_context_verify_mode(ctx, verify_mode)
+
+
+def _verify_peercerts_impl(
+    ssl_context: ssl.SSLContext,
+    cert_chain: list[bytes],
+    server_hostname: str | None = None,
+) -> None:
+    certs = None
+    policies = None
+    trust = None
+    try:
+        # Only set a hostname on the policy if we're verifying the hostname
+        # on the leaf certificate.
+        if server_hostname is not None and ssl_context.check_hostname:
+            cf_str_hostname = None
+            try:
+                cf_str_hostname = _bytes_to_cf_string(server_hostname.encode("ascii"))
+                ssl_policy = Security.SecPolicyCreateSSL(True, cf_str_hostname)
+            finally:
+                if cf_str_hostname:
+                    CoreFoundation.CFRelease(cf_str_hostname)
+        else:
+            ssl_policy = Security.SecPolicyCreateSSL(True, None)
+
+        policies = ssl_policy
+        if ssl_context.verify_flags & ssl.VERIFY_CRL_CHECK_CHAIN:
+            # Add explicit policy requiring positive revocation checks
+            policies = CoreFoundation.CFArrayCreateMutable(
+                CoreFoundation.kCFAllocatorDefault,
+                0,
+                ctypes.byref(CoreFoundation.kCFTypeArrayCallBacks),
+            )
+            CoreFoundation.CFArrayAppendValue(policies, ssl_policy)
+            CoreFoundation.CFRelease(ssl_policy)
+            revocation_policy = Security.SecPolicyCreateRevocation(
+                kSecRevocationUseAnyAvailableMethod
+                | kSecRevocationRequirePositiveResponse
+            )
+            CoreFoundation.CFArrayAppendValue(policies, revocation_policy)
+            CoreFoundation.CFRelease(revocation_policy)
+        elif ssl_context.verify_flags & ssl.VERIFY_CRL_CHECK_LEAF:
+            raise NotImplementedError("VERIFY_CRL_CHECK_LEAF not implemented for macOS")
+
+        certs = None
+        try:
+            certs = _der_certs_to_cf_cert_array(cert_chain)
+
+            # Now that we have certificates loaded and a SecPolicy
+            # we can finally create a SecTrust object!
+            trust = Security.SecTrustRef()
+            Security.SecTrustCreateWithCertificates(
+                certs, policies, ctypes.byref(trust)
+            )
+
+        finally:
+            # The certs are now being held by SecTrust so we can
+            # release our handles for the array.
+            if certs:
+                CoreFoundation.CFRelease(certs)
+
+        # If there are additional trust anchors to load we need to transform
+        # the list of DER-encoded certificates into a CFArray.
+        ctx_ca_certs_der: list[bytes] | None = ssl_context.get_ca_certs(
+            binary_form=True
+        )
+        if ctx_ca_certs_der:
+            ctx_ca_certs = None
+            try:
+                ctx_ca_certs = _der_certs_to_cf_cert_array(ctx_ca_certs_der)
+                Security.SecTrustSetAnchorCertificates(trust, ctx_ca_certs)
+            finally:
+                if ctx_ca_certs:
+                    CoreFoundation.CFRelease(ctx_ca_certs)
+
+        # We always want system certificates.
+        Security.SecTrustSetAnchorCertificatesOnly(trust, False)
+
+        # macOS 10.13 and earlier don't support SecTrustEvaluateWithError()
+        # so we use SecTrustEvaluate() which means we need to construct error
+        # messages ourselves.
+        if _is_macos_version_10_14_or_later:
+            _verify_peercerts_impl_macos_10_14(ssl_context, trust)
+        else:
+            _verify_peercerts_impl_macos_10_13(ssl_context, trust)
+    finally:
+        if policies:
+            CoreFoundation.CFRelease(policies)
+        if trust:
+            CoreFoundation.CFRelease(trust)
+
+
+def _verify_peercerts_impl_macos_10_13(
+    ssl_context: ssl.SSLContext, sec_trust_ref: typing.Any
+) -> None:
+    """Verify using 'SecTrustEvaluate' API for macOS 10.13 and earlier.
+    macOS 10.14 added the 'SecTrustEvaluateWithError' API.
+    """
+    sec_trust_result_type = Security.SecTrustResultType()
+    Security.SecTrustEvaluate(sec_trust_ref, ctypes.byref(sec_trust_result_type))
+
+    try:
+        sec_trust_result_type_as_int = int(sec_trust_result_type.value)
+    except (ValueError, TypeError):
+        sec_trust_result_type_as_int = -1
+
+    # Apple doesn't document these values in their own API docs.
+    # See: https://github.com/xybp888/iOS-SDKs/blob/master/iPhoneOS13.0.sdk/System/Library/Frameworks/Security.framework/Headers/SecTrust.h#L84
+    if (
+        ssl_context.verify_mode == ssl.CERT_REQUIRED
+        and sec_trust_result_type_as_int not in (1, 4)
+    ):
+        # Note that we're not able to ignore only hostname errors
+        # for macOS 10.13 and earlier, so check_hostname=False will
+        # still return an error.
+        sec_trust_result_type_to_message = {
+            0: "Invalid trust result type",
+            # 1: "Trust evaluation succeeded",
+            2: "User confirmation required",
+            3: "User specified that certificate is not trusted",
+            # 4: "Trust result is unspecified",
+            5: "Recoverable trust failure occurred",
+            6: "Fatal trust failure occurred",
+            7: "Other error occurred, certificate may be revoked",
+        }
+        error_message = sec_trust_result_type_to_message.get(
+            sec_trust_result_type_as_int,
+            f"Unknown trust result: {sec_trust_result_type_as_int}",
+        )
+
+        err = ssl.SSLCertVerificationError(error_message)
+        err.verify_message = error_message
+        err.verify_code = sec_trust_result_type_as_int
+        raise err
+
+
+def _verify_peercerts_impl_macos_10_14(
+    ssl_context: ssl.SSLContext, sec_trust_ref: typing.Any
+) -> None:
+    """Verify using 'SecTrustEvaluateWithError' API for macOS 10.14+."""
+    cf_error = CoreFoundation.CFErrorRef()
+    sec_trust_eval_result = Security.SecTrustEvaluateWithError(
+        sec_trust_ref, ctypes.byref(cf_error)
+    )
+    # sec_trust_eval_result is a bool (0 or 1)
+    # where 1 means that the certs are trusted.
+    if sec_trust_eval_result == 1:
+        is_trusted = True
+    elif sec_trust_eval_result == 0:
+        is_trusted = False
+    else:
+        raise ssl.SSLError(
+            f"Unknown result from Security.SecTrustEvaluateWithError: {sec_trust_eval_result!r}"
+        )
+
+    cf_error_code = 0
+    if not is_trusted:
+        cf_error_code = CoreFoundation.CFErrorGetCode(cf_error)
+
+        # If the error is a known failure that we're
+        # explicitly okay with from SSLContext configuration
+        # we can set is_trusted accordingly.
+        if ssl_context.verify_mode != ssl.CERT_REQUIRED and (
+            cf_error_code == CFConst.errSecNotTrusted
+            or cf_error_code == CFConst.errSecCertificateExpired
+        ):
+            is_trusted = True
+
+    # If we're still not trusted then we start to
+    # construct and raise the SSLCertVerificationError.
+    if not is_trusted:
+        cf_error_string_ref = None
+        try:
+            cf_error_string_ref = CoreFoundation.CFErrorCopyDescription(cf_error)
+
+            # Can this ever return 'None' if there's a CFError?
+            cf_error_message = (
+                _cf_string_ref_to_str(cf_error_string_ref)
+                or "Certificate verification failed"
+            )
+
+            # TODO: Not sure if we need the SecTrustResultType for anything?
+            # We only care whether or not it's a success or failure for now.
+            sec_trust_result_type = Security.SecTrustResultType()
+            Security.SecTrustGetTrustResult(
+                sec_trust_ref, ctypes.byref(sec_trust_result_type)
+            )
+
+            err = ssl.SSLCertVerificationError(cf_error_message)
+            err.verify_message = cf_error_message
+            err.verify_code = cf_error_code
+            raise err
+        finally:
+            if cf_error_string_ref:
+                CoreFoundation.CFRelease(cf_error_string_ref)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_openssl.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_openssl.py
new file mode 100644
index 00000000..9951cf75
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_openssl.py
@@ -0,0 +1,66 @@
+import contextlib
+import os
+import re
+import ssl
+import typing
+
+# candidates based on https://github.com/tiran/certifi-system-store by Christian Heimes
+_CA_FILE_CANDIDATES = [
+    # Alpine, Arch, Fedora 34+, OpenWRT, RHEL 9+, BSD
+    "/etc/ssl/cert.pem",
+    # Fedora <= 34, RHEL <= 9, CentOS <= 9
+    "/etc/pki/tls/cert.pem",
+    # Debian, Ubuntu (requires ca-certificates)
+    "/etc/ssl/certs/ca-certificates.crt",
+    # SUSE
+    "/etc/ssl/ca-bundle.pem",
+]
+
+_HASHED_CERT_FILENAME_RE = re.compile(r"^[0-9a-fA-F]{8}\.[0-9]$")
+
+
+@contextlib.contextmanager
+def _configure_context(ctx: ssl.SSLContext) -> typing.Iterator[None]:
+    # First, check whether the default locations from OpenSSL
+    # seem like they will give us a usable set of CA certs.
+    # ssl.get_default_verify_paths already takes care of:
+    # - getting cafile from either the SSL_CERT_FILE env var
+    #   or the path configured when OpenSSL was compiled,
+    #   and verifying that that path exists
+    # - getting capath from either the SSL_CERT_DIR env var
+    #   or the path configured when OpenSSL was compiled,
+    #   and verifying that that path exists
+    # In addition we'll check whether capath appears to contain certs.
+    defaults = ssl.get_default_verify_paths()
+    if defaults.cafile or (defaults.capath and _capath_contains_certs(defaults.capath)):
+        ctx.set_default_verify_paths()
+    else:
+        # cafile from OpenSSL doesn't exist
+        # and capath from OpenSSL doesn't contain certs.
+        # Let's search other common locations instead.
+        for cafile in _CA_FILE_CANDIDATES:
+            if os.path.isfile(cafile):
+                ctx.load_verify_locations(cafile=cafile)
+                break
+
+    yield
+
+
+def _capath_contains_certs(capath: str) -> bool:
+    """Check whether capath exists and contains certs in the expected format."""
+    if not os.path.isdir(capath):
+        return False
+    for name in os.listdir(capath):
+        if _HASHED_CERT_FILENAME_RE.match(name):
+            return True
+    return False
+
+
+def _verify_peercerts_impl(
+    ssl_context: ssl.SSLContext,
+    cert_chain: list[bytes],
+    server_hostname: str | None = None,
+) -> None:
+    # This is a no-op because we've enabled SSLContext's built-in
+    # verification via verify_mode=CERT_REQUIRED, and don't need to repeat it.
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_ssl_constants.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_ssl_constants.py
new file mode 100644
index 00000000..b1ee7a3c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_ssl_constants.py
@@ -0,0 +1,31 @@
+import ssl
+import sys
+import typing
+
+# Hold on to the original class so we can create it consistently
+# even if we inject our own SSLContext into the ssl module.
+_original_SSLContext = ssl.SSLContext
+_original_super_SSLContext = super(_original_SSLContext, _original_SSLContext)
+
+# CPython is known to be good, but non-CPython implementations
+# may implement SSLContext differently so to be safe we don't
+# subclass the SSLContext.
+
+# This is returned by truststore.SSLContext.__class__()
+_truststore_SSLContext_dunder_class: typing.Optional[type]
+
+# This value is the superclass of truststore.SSLContext.
+_truststore_SSLContext_super_class: type
+
+if sys.implementation.name == "cpython":
+    _truststore_SSLContext_super_class = _original_SSLContext
+    _truststore_SSLContext_dunder_class = None
+else:
+    _truststore_SSLContext_super_class = object
+    _truststore_SSLContext_dunder_class = _original_SSLContext
+
+
+def _set_ssl_context_verify_mode(
+    ssl_context: ssl.SSLContext, verify_mode: ssl.VerifyMode
+) -> None:
+    _original_super_SSLContext.verify_mode.__set__(ssl_context, verify_mode)  # type: ignore[attr-defined]
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_windows.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_windows.py
new file mode 100644
index 00000000..a9bf9abd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/_windows.py
@@ -0,0 +1,567 @@
+import contextlib
+import ssl
+import typing
+from ctypes import WinDLL  # type: ignore
+from ctypes import WinError  # type: ignore
+from ctypes import (
+    POINTER,
+    Structure,
+    c_char_p,
+    c_ulong,
+    c_void_p,
+    c_wchar_p,
+    cast,
+    create_unicode_buffer,
+    pointer,
+    sizeof,
+)
+from ctypes.wintypes import (
+    BOOL,
+    DWORD,
+    HANDLE,
+    LONG,
+    LPCSTR,
+    LPCVOID,
+    LPCWSTR,
+    LPFILETIME,
+    LPSTR,
+    LPWSTR,
+)
+from typing import TYPE_CHECKING, Any
+
+from ._ssl_constants import _set_ssl_context_verify_mode
+
+HCERTCHAINENGINE = HANDLE
+HCERTSTORE = HANDLE
+HCRYPTPROV_LEGACY = HANDLE
+
+
+class CERT_CONTEXT(Structure):
+    _fields_ = (
+        ("dwCertEncodingType", DWORD),
+        ("pbCertEncoded", c_void_p),
+        ("cbCertEncoded", DWORD),
+        ("pCertInfo", c_void_p),
+        ("hCertStore", HCERTSTORE),
+    )
+
+
+PCERT_CONTEXT = POINTER(CERT_CONTEXT)
+PCCERT_CONTEXT = POINTER(PCERT_CONTEXT)
+
+
+class CERT_ENHKEY_USAGE(Structure):
+    _fields_ = (
+        ("cUsageIdentifier", DWORD),
+        ("rgpszUsageIdentifier", POINTER(LPSTR)),
+    )
+
+
+PCERT_ENHKEY_USAGE = POINTER(CERT_ENHKEY_USAGE)
+
+
+class CERT_USAGE_MATCH(Structure):
+    _fields_ = (
+        ("dwType", DWORD),
+        ("Usage", CERT_ENHKEY_USAGE),
+    )
+
+
+class CERT_CHAIN_PARA(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("RequestedUsage", CERT_USAGE_MATCH),
+        ("RequestedIssuancePolicy", CERT_USAGE_MATCH),
+        ("dwUrlRetrievalTimeout", DWORD),
+        ("fCheckRevocationFreshnessTime", BOOL),
+        ("dwRevocationFreshnessTime", DWORD),
+        ("pftCacheResync", LPFILETIME),
+        ("pStrongSignPara", c_void_p),
+        ("dwStrongSignFlags", DWORD),
+    )
+
+
+if TYPE_CHECKING:
+    PCERT_CHAIN_PARA = pointer[CERT_CHAIN_PARA]  # type: ignore[misc]
+else:
+    PCERT_CHAIN_PARA = POINTER(CERT_CHAIN_PARA)
+
+
+class CERT_TRUST_STATUS(Structure):
+    _fields_ = (
+        ("dwErrorStatus", DWORD),
+        ("dwInfoStatus", DWORD),
+    )
+
+
+class CERT_CHAIN_ELEMENT(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("pCertContext", PCERT_CONTEXT),
+        ("TrustStatus", CERT_TRUST_STATUS),
+        ("pRevocationInfo", c_void_p),
+        ("pIssuanceUsage", PCERT_ENHKEY_USAGE),
+        ("pApplicationUsage", PCERT_ENHKEY_USAGE),
+        ("pwszExtendedErrorInfo", LPCWSTR),
+    )
+
+
+PCERT_CHAIN_ELEMENT = POINTER(CERT_CHAIN_ELEMENT)
+
+
+class CERT_SIMPLE_CHAIN(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("TrustStatus", CERT_TRUST_STATUS),
+        ("cElement", DWORD),
+        ("rgpElement", POINTER(PCERT_CHAIN_ELEMENT)),
+        ("pTrustListInfo", c_void_p),
+        ("fHasRevocationFreshnessTime", BOOL),
+        ("dwRevocationFreshnessTime", DWORD),
+    )
+
+
+PCERT_SIMPLE_CHAIN = POINTER(CERT_SIMPLE_CHAIN)
+
+
+class CERT_CHAIN_CONTEXT(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("TrustStatus", CERT_TRUST_STATUS),
+        ("cChain", DWORD),
+        ("rgpChain", POINTER(PCERT_SIMPLE_CHAIN)),
+        ("cLowerQualityChainContext", DWORD),
+        ("rgpLowerQualityChainContext", c_void_p),
+        ("fHasRevocationFreshnessTime", BOOL),
+        ("dwRevocationFreshnessTime", DWORD),
+    )
+
+
+PCERT_CHAIN_CONTEXT = POINTER(CERT_CHAIN_CONTEXT)
+PCCERT_CHAIN_CONTEXT = POINTER(PCERT_CHAIN_CONTEXT)
+
+
+class SSL_EXTRA_CERT_CHAIN_POLICY_PARA(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("dwAuthType", DWORD),
+        ("fdwChecks", DWORD),
+        ("pwszServerName", LPCWSTR),
+    )
+
+
+class CERT_CHAIN_POLICY_PARA(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("dwFlags", DWORD),
+        ("pvExtraPolicyPara", c_void_p),
+    )
+
+
+PCERT_CHAIN_POLICY_PARA = POINTER(CERT_CHAIN_POLICY_PARA)
+
+
+class CERT_CHAIN_POLICY_STATUS(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("dwError", DWORD),
+        ("lChainIndex", LONG),
+        ("lElementIndex", LONG),
+        ("pvExtraPolicyStatus", c_void_p),
+    )
+
+
+PCERT_CHAIN_POLICY_STATUS = POINTER(CERT_CHAIN_POLICY_STATUS)
+
+
+class CERT_CHAIN_ENGINE_CONFIG(Structure):
+    _fields_ = (
+        ("cbSize", DWORD),
+        ("hRestrictedRoot", HCERTSTORE),
+        ("hRestrictedTrust", HCERTSTORE),
+        ("hRestrictedOther", HCERTSTORE),
+        ("cAdditionalStore", DWORD),
+        ("rghAdditionalStore", c_void_p),
+        ("dwFlags", DWORD),
+        ("dwUrlRetrievalTimeout", DWORD),
+        ("MaximumCachedCertificates", DWORD),
+        ("CycleDetectionModulus", DWORD),
+        ("hExclusiveRoot", HCERTSTORE),
+        ("hExclusiveTrustedPeople", HCERTSTORE),
+        ("dwExclusiveFlags", DWORD),
+    )
+
+
+PCERT_CHAIN_ENGINE_CONFIG = POINTER(CERT_CHAIN_ENGINE_CONFIG)
+PHCERTCHAINENGINE = POINTER(HCERTCHAINENGINE)
+
+X509_ASN_ENCODING = 0x00000001
+PKCS_7_ASN_ENCODING = 0x00010000
+CERT_STORE_PROV_MEMORY = b"Memory"
+CERT_STORE_ADD_USE_EXISTING = 2
+USAGE_MATCH_TYPE_OR = 1
+OID_PKIX_KP_SERVER_AUTH = c_char_p(b"1.3.6.1.5.5.7.3.1")
+CERT_CHAIN_REVOCATION_CHECK_END_CERT = 0x10000000
+CERT_CHAIN_REVOCATION_CHECK_CHAIN = 0x20000000
+CERT_CHAIN_POLICY_IGNORE_ALL_NOT_TIME_VALID_FLAGS = 0x00000007
+CERT_CHAIN_POLICY_IGNORE_INVALID_BASIC_CONSTRAINTS_FLAG = 0x00000008
+CERT_CHAIN_POLICY_ALLOW_UNKNOWN_CA_FLAG = 0x00000010
+CERT_CHAIN_POLICY_IGNORE_INVALID_NAME_FLAG = 0x00000040
+CERT_CHAIN_POLICY_IGNORE_WRONG_USAGE_FLAG = 0x00000020
+CERT_CHAIN_POLICY_IGNORE_INVALID_POLICY_FLAG = 0x00000080
+CERT_CHAIN_POLICY_IGNORE_ALL_REV_UNKNOWN_FLAGS = 0x00000F00
+CERT_CHAIN_POLICY_ALLOW_TESTROOT_FLAG = 0x00008000
+CERT_CHAIN_POLICY_TRUST_TESTROOT_FLAG = 0x00004000
+SECURITY_FLAG_IGNORE_CERT_CN_INVALID = 0x00001000
+AUTHTYPE_SERVER = 2
+CERT_CHAIN_POLICY_SSL = 4
+FORMAT_MESSAGE_FROM_SYSTEM = 0x00001000
+FORMAT_MESSAGE_IGNORE_INSERTS = 0x00000200
+
+# Flags to set for SSLContext.verify_mode=CERT_NONE
+CERT_CHAIN_POLICY_VERIFY_MODE_NONE_FLAGS = (
+    CERT_CHAIN_POLICY_IGNORE_ALL_NOT_TIME_VALID_FLAGS
+    | CERT_CHAIN_POLICY_IGNORE_INVALID_BASIC_CONSTRAINTS_FLAG
+    | CERT_CHAIN_POLICY_ALLOW_UNKNOWN_CA_FLAG
+    | CERT_CHAIN_POLICY_IGNORE_INVALID_NAME_FLAG
+    | CERT_CHAIN_POLICY_IGNORE_WRONG_USAGE_FLAG
+    | CERT_CHAIN_POLICY_IGNORE_INVALID_POLICY_FLAG
+    | CERT_CHAIN_POLICY_IGNORE_ALL_REV_UNKNOWN_FLAGS
+    | CERT_CHAIN_POLICY_ALLOW_TESTROOT_FLAG
+    | CERT_CHAIN_POLICY_TRUST_TESTROOT_FLAG
+)
+
+wincrypt = WinDLL("crypt32.dll")
+kernel32 = WinDLL("kernel32.dll")
+
+
+def _handle_win_error(result: bool, _: Any, args: Any) -> Any:
+    if not result:
+        # Note, actually raises OSError after calling GetLastError and FormatMessage
+        raise WinError()
+    return args
+
+
+CertCreateCertificateChainEngine = wincrypt.CertCreateCertificateChainEngine
+CertCreateCertificateChainEngine.argtypes = (
+    PCERT_CHAIN_ENGINE_CONFIG,
+    PHCERTCHAINENGINE,
+)
+CertCreateCertificateChainEngine.errcheck = _handle_win_error
+
+CertOpenStore = wincrypt.CertOpenStore
+CertOpenStore.argtypes = (LPCSTR, DWORD, HCRYPTPROV_LEGACY, DWORD, c_void_p)
+CertOpenStore.restype = HCERTSTORE
+CertOpenStore.errcheck = _handle_win_error
+
+CertAddEncodedCertificateToStore = wincrypt.CertAddEncodedCertificateToStore
+CertAddEncodedCertificateToStore.argtypes = (
+    HCERTSTORE,
+    DWORD,
+    c_char_p,
+    DWORD,
+    DWORD,
+    PCCERT_CONTEXT,
+)
+CertAddEncodedCertificateToStore.restype = BOOL
+
+CertCreateCertificateContext = wincrypt.CertCreateCertificateContext
+CertCreateCertificateContext.argtypes = (DWORD, c_char_p, DWORD)
+CertCreateCertificateContext.restype = PCERT_CONTEXT
+CertCreateCertificateContext.errcheck = _handle_win_error
+
+CertGetCertificateChain = wincrypt.CertGetCertificateChain
+CertGetCertificateChain.argtypes = (
+    HCERTCHAINENGINE,
+    PCERT_CONTEXT,
+    LPFILETIME,
+    HCERTSTORE,
+    PCERT_CHAIN_PARA,
+    DWORD,
+    c_void_p,
+    PCCERT_CHAIN_CONTEXT,
+)
+CertGetCertificateChain.restype = BOOL
+CertGetCertificateChain.errcheck = _handle_win_error
+
+CertVerifyCertificateChainPolicy = wincrypt.CertVerifyCertificateChainPolicy
+CertVerifyCertificateChainPolicy.argtypes = (
+    c_ulong,
+    PCERT_CHAIN_CONTEXT,
+    PCERT_CHAIN_POLICY_PARA,
+    PCERT_CHAIN_POLICY_STATUS,
+)
+CertVerifyCertificateChainPolicy.restype = BOOL
+
+CertCloseStore = wincrypt.CertCloseStore
+CertCloseStore.argtypes = (HCERTSTORE, DWORD)
+CertCloseStore.restype = BOOL
+CertCloseStore.errcheck = _handle_win_error
+
+CertFreeCertificateChain = wincrypt.CertFreeCertificateChain
+CertFreeCertificateChain.argtypes = (PCERT_CHAIN_CONTEXT,)
+
+CertFreeCertificateContext = wincrypt.CertFreeCertificateContext
+CertFreeCertificateContext.argtypes = (PCERT_CONTEXT,)
+
+CertFreeCertificateChainEngine = wincrypt.CertFreeCertificateChainEngine
+CertFreeCertificateChainEngine.argtypes = (HCERTCHAINENGINE,)
+
+FormatMessageW = kernel32.FormatMessageW
+FormatMessageW.argtypes = (
+    DWORD,
+    LPCVOID,
+    DWORD,
+    DWORD,
+    LPWSTR,
+    DWORD,
+    c_void_p,
+)
+FormatMessageW.restype = DWORD
+
+
+def _verify_peercerts_impl(
+    ssl_context: ssl.SSLContext,
+    cert_chain: list[bytes],
+    server_hostname: str | None = None,
+) -> None:
+    """Verify the cert_chain from the server using Windows APIs."""
+
+    # If the peer didn't send any certificates then
+    # we can't do verification. Raise an error.
+    if not cert_chain:
+        raise ssl.SSLCertVerificationError("Peer sent no certificates to verify")
+
+    pCertContext = None
+    hIntermediateCertStore = CertOpenStore(CERT_STORE_PROV_MEMORY, 0, None, 0, None)
+    try:
+        # Add intermediate certs to an in-memory cert store
+        for cert_bytes in cert_chain[1:]:
+            CertAddEncodedCertificateToStore(
+                hIntermediateCertStore,
+                X509_ASN_ENCODING | PKCS_7_ASN_ENCODING,
+                cert_bytes,
+                len(cert_bytes),
+                CERT_STORE_ADD_USE_EXISTING,
+                None,
+            )
+
+        # Cert context for leaf cert
+        leaf_cert = cert_chain[0]
+        pCertContext = CertCreateCertificateContext(
+            X509_ASN_ENCODING | PKCS_7_ASN_ENCODING, leaf_cert, len(leaf_cert)
+        )
+
+        # Chain params to match certs for serverAuth extended usage
+        cert_enhkey_usage = CERT_ENHKEY_USAGE()
+        cert_enhkey_usage.cUsageIdentifier = 1
+        cert_enhkey_usage.rgpszUsageIdentifier = (c_char_p * 1)(OID_PKIX_KP_SERVER_AUTH)
+        cert_usage_match = CERT_USAGE_MATCH()
+        cert_usage_match.Usage = cert_enhkey_usage
+        chain_params = CERT_CHAIN_PARA()
+        chain_params.RequestedUsage = cert_usage_match
+        chain_params.cbSize = sizeof(chain_params)
+        pChainPara = pointer(chain_params)
+
+        if ssl_context.verify_flags & ssl.VERIFY_CRL_CHECK_CHAIN:
+            chain_flags = CERT_CHAIN_REVOCATION_CHECK_CHAIN
+        elif ssl_context.verify_flags & ssl.VERIFY_CRL_CHECK_LEAF:
+            chain_flags = CERT_CHAIN_REVOCATION_CHECK_END_CERT
+        else:
+            chain_flags = 0
+
+        try:
+            # First attempt to verify using the default Windows system trust roots
+            # (default chain engine).
+            _get_and_verify_cert_chain(
+                ssl_context,
+                None,
+                hIntermediateCertStore,
+                pCertContext,
+                pChainPara,
+                server_hostname,
+                chain_flags=chain_flags,
+            )
+        except ssl.SSLCertVerificationError as e:
+            # If that fails but custom CA certs have been added
+            # to the SSLContext using load_verify_locations,
+            # try verifying using a custom chain engine
+            # that trusts the custom CA certs.
+            custom_ca_certs: list[bytes] | None = ssl_context.get_ca_certs(
+                binary_form=True
+            )
+            if custom_ca_certs:
+                try:
+                    _verify_using_custom_ca_certs(
+                        ssl_context,
+                        custom_ca_certs,
+                        hIntermediateCertStore,
+                        pCertContext,
+                        pChainPara,
+                        server_hostname,
+                        chain_flags=chain_flags,
+                    )
+                # Raise the original error, not the new error.
+                except ssl.SSLCertVerificationError:
+                    raise e from None
+            else:
+                raise
+    finally:
+        CertCloseStore(hIntermediateCertStore, 0)
+        if pCertContext:
+            CertFreeCertificateContext(pCertContext)
+
+
+def _get_and_verify_cert_chain(
+    ssl_context: ssl.SSLContext,
+    hChainEngine: HCERTCHAINENGINE | None,
+    hIntermediateCertStore: HCERTSTORE,
+    pPeerCertContext: c_void_p,
+    pChainPara: PCERT_CHAIN_PARA,  # type: ignore[valid-type]
+    server_hostname: str | None,
+    chain_flags: int,
+) -> None:
+    ppChainContext = None
+    try:
+        # Get cert chain
+        ppChainContext = pointer(PCERT_CHAIN_CONTEXT())
+        CertGetCertificateChain(
+            hChainEngine,  # chain engine
+            pPeerCertContext,  # leaf cert context
+            None,  # current system time
+            hIntermediateCertStore,  # additional in-memory cert store
+            pChainPara,  # chain-building parameters
+            chain_flags,
+            None,  # reserved
+            ppChainContext,  # the resulting chain context
+        )
+        pChainContext = ppChainContext.contents
+
+        # Verify cert chain
+        ssl_extra_cert_chain_policy_para = SSL_EXTRA_CERT_CHAIN_POLICY_PARA()
+        ssl_extra_cert_chain_policy_para.cbSize = sizeof(
+            ssl_extra_cert_chain_policy_para
+        )
+        ssl_extra_cert_chain_policy_para.dwAuthType = AUTHTYPE_SERVER
+        ssl_extra_cert_chain_policy_para.fdwChecks = 0
+        if ssl_context.check_hostname is False:
+            ssl_extra_cert_chain_policy_para.fdwChecks = (
+                SECURITY_FLAG_IGNORE_CERT_CN_INVALID
+            )
+        if server_hostname:
+            ssl_extra_cert_chain_policy_para.pwszServerName = c_wchar_p(server_hostname)
+
+        chain_policy = CERT_CHAIN_POLICY_PARA()
+        chain_policy.pvExtraPolicyPara = cast(
+            pointer(ssl_extra_cert_chain_policy_para), c_void_p
+        )
+        if ssl_context.verify_mode == ssl.CERT_NONE:
+            chain_policy.dwFlags |= CERT_CHAIN_POLICY_VERIFY_MODE_NONE_FLAGS
+        chain_policy.cbSize = sizeof(chain_policy)
+
+        pPolicyPara = pointer(chain_policy)
+        policy_status = CERT_CHAIN_POLICY_STATUS()
+        policy_status.cbSize = sizeof(policy_status)
+        pPolicyStatus = pointer(policy_status)
+        CertVerifyCertificateChainPolicy(
+            CERT_CHAIN_POLICY_SSL,
+            pChainContext,
+            pPolicyPara,
+            pPolicyStatus,
+        )
+
+        # Check status
+        error_code = policy_status.dwError
+        if error_code:
+            # Try getting a human readable message for an error code.
+            error_message_buf = create_unicode_buffer(1024)
+            error_message_chars = FormatMessageW(
+                FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS,
+                None,
+                error_code,
+                0,
+                error_message_buf,
+                sizeof(error_message_buf),
+                None,
+            )
+
+            # See if we received a message for the error,
+            # otherwise we use a generic error with the
+            # error code and hope that it's search-able.
+            if error_message_chars <= 0:
+                error_message = f"Certificate chain policy error {error_code:#x} [{policy_status.lElementIndex}]"
+            else:
+                error_message = error_message_buf.value.strip()
+
+            err = ssl.SSLCertVerificationError(error_message)
+            err.verify_message = error_message
+            err.verify_code = error_code
+            raise err from None
+    finally:
+        if ppChainContext:
+            CertFreeCertificateChain(ppChainContext.contents)
+
+
+def _verify_using_custom_ca_certs(
+    ssl_context: ssl.SSLContext,
+    custom_ca_certs: list[bytes],
+    hIntermediateCertStore: HCERTSTORE,
+    pPeerCertContext: c_void_p,
+    pChainPara: PCERT_CHAIN_PARA,  # type: ignore[valid-type]
+    server_hostname: str | None,
+    chain_flags: int,
+) -> None:
+    hChainEngine = None
+    hRootCertStore = CertOpenStore(CERT_STORE_PROV_MEMORY, 0, None, 0, None)
+    try:
+        # Add custom CA certs to an in-memory cert store
+        for cert_bytes in custom_ca_certs:
+            CertAddEncodedCertificateToStore(
+                hRootCertStore,
+                X509_ASN_ENCODING | PKCS_7_ASN_ENCODING,
+                cert_bytes,
+                len(cert_bytes),
+                CERT_STORE_ADD_USE_EXISTING,
+                None,
+            )
+
+        # Create a custom cert chain engine which exclusively trusts
+        # certs from our hRootCertStore
+        cert_chain_engine_config = CERT_CHAIN_ENGINE_CONFIG()
+        cert_chain_engine_config.cbSize = sizeof(cert_chain_engine_config)
+        cert_chain_engine_config.hExclusiveRoot = hRootCertStore
+        pConfig = pointer(cert_chain_engine_config)
+        phChainEngine = pointer(HCERTCHAINENGINE())
+        CertCreateCertificateChainEngine(
+            pConfig,
+            phChainEngine,
+        )
+        hChainEngine = phChainEngine.contents
+
+        # Get and verify a cert chain using the custom chain engine
+        _get_and_verify_cert_chain(
+            ssl_context,
+            hChainEngine,
+            hIntermediateCertStore,
+            pPeerCertContext,
+            pChainPara,
+            server_hostname,
+            chain_flags,
+        )
+    finally:
+        if hChainEngine:
+            CertFreeCertificateChainEngine(hChainEngine)
+        CertCloseStore(hRootCertStore, 0)
+
+
+@contextlib.contextmanager
+def _configure_context(ctx: ssl.SSLContext) -> typing.Iterator[None]:
+    check_hostname = ctx.check_hostname
+    verify_mode = ctx.verify_mode
+    ctx.check_hostname = False
+    _set_ssl_context_verify_mode(ctx, ssl.CERT_NONE)
+    try:
+        yield
+    finally:
+        ctx.check_hostname = check_hostname
+        _set_ssl_context_verify_mode(ctx, verify_mode)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/py.typed b/backend/venv/lib/python3.12/site-packages/pip/_vendor/truststore/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/typing_extensions.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/typing_extensions.py
new file mode 100644
index 00000000..e429384e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/typing_extensions.py
@@ -0,0 +1,3641 @@
+import abc
+import collections
+import collections.abc
+import contextlib
+import functools
+import inspect
+import operator
+import sys
+import types as _types
+import typing
+import warnings
+
+__all__ = [
+    # Super-special typing primitives.
+    'Any',
+    'ClassVar',
+    'Concatenate',
+    'Final',
+    'LiteralString',
+    'ParamSpec',
+    'ParamSpecArgs',
+    'ParamSpecKwargs',
+    'Self',
+    'Type',
+    'TypeVar',
+    'TypeVarTuple',
+    'Unpack',
+
+    # ABCs (from collections.abc).
+    'Awaitable',
+    'AsyncIterator',
+    'AsyncIterable',
+    'Coroutine',
+    'AsyncGenerator',
+    'AsyncContextManager',
+    'Buffer',
+    'ChainMap',
+
+    # Concrete collection types.
+    'ContextManager',
+    'Counter',
+    'Deque',
+    'DefaultDict',
+    'NamedTuple',
+    'OrderedDict',
+    'TypedDict',
+
+    # Structural checks, a.k.a. protocols.
+    'SupportsAbs',
+    'SupportsBytes',
+    'SupportsComplex',
+    'SupportsFloat',
+    'SupportsIndex',
+    'SupportsInt',
+    'SupportsRound',
+
+    # One-off things.
+    'Annotated',
+    'assert_never',
+    'assert_type',
+    'clear_overloads',
+    'dataclass_transform',
+    'deprecated',
+    'Doc',
+    'get_overloads',
+    'final',
+    'get_args',
+    'get_origin',
+    'get_original_bases',
+    'get_protocol_members',
+    'get_type_hints',
+    'IntVar',
+    'is_protocol',
+    'is_typeddict',
+    'Literal',
+    'NewType',
+    'overload',
+    'override',
+    'Protocol',
+    'reveal_type',
+    'runtime',
+    'runtime_checkable',
+    'Text',
+    'TypeAlias',
+    'TypeAliasType',
+    'TypeGuard',
+    'TypeIs',
+    'TYPE_CHECKING',
+    'Never',
+    'NoReturn',
+    'ReadOnly',
+    'Required',
+    'NotRequired',
+
+    # Pure aliases, have always been in typing
+    'AbstractSet',
+    'AnyStr',
+    'BinaryIO',
+    'Callable',
+    'Collection',
+    'Container',
+    'Dict',
+    'ForwardRef',
+    'FrozenSet',
+    'Generator',
+    'Generic',
+    'Hashable',
+    'IO',
+    'ItemsView',
+    'Iterable',
+    'Iterator',
+    'KeysView',
+    'List',
+    'Mapping',
+    'MappingView',
+    'Match',
+    'MutableMapping',
+    'MutableSequence',
+    'MutableSet',
+    'NoDefault',
+    'Optional',
+    'Pattern',
+    'Reversible',
+    'Sequence',
+    'Set',
+    'Sized',
+    'TextIO',
+    'Tuple',
+    'Union',
+    'ValuesView',
+    'cast',
+    'no_type_check',
+    'no_type_check_decorator',
+]
+
+# for backward compatibility
+PEP_560 = True
+GenericMeta = type
+_PEP_696_IMPLEMENTED = sys.version_info >= (3, 13, 0, "beta")
+
+# The functions below are modified copies of typing internal helpers.
+# They are needed by _ProtocolMeta and they provide support for PEP 646.
+
+
+class _Sentinel:
+    def __repr__(self):
+        return "<sentinel>"
+
+
+_marker = _Sentinel()
+
+
+if sys.version_info >= (3, 10):
+    def _should_collect_from_parameters(t):
+        return isinstance(
+            t, (typing._GenericAlias, _types.GenericAlias, _types.UnionType)
+        )
+elif sys.version_info >= (3, 9):
+    def _should_collect_from_parameters(t):
+        return isinstance(t, (typing._GenericAlias, _types.GenericAlias))
+else:
+    def _should_collect_from_parameters(t):
+        return isinstance(t, typing._GenericAlias) and not t._special
+
+
+NoReturn = typing.NoReturn
+
+# Some unconstrained type variables.  These are used by the container types.
+# (These are not for export.)
+T = typing.TypeVar('T')  # Any type.
+KT = typing.TypeVar('KT')  # Key type.
+VT = typing.TypeVar('VT')  # Value type.
+T_co = typing.TypeVar('T_co', covariant=True)  # Any type covariant containers.
+T_contra = typing.TypeVar('T_contra', contravariant=True)  # Ditto contravariant.
+
+
+if sys.version_info >= (3, 11):
+    from typing import Any
+else:
+
+    class _AnyMeta(type):
+        def __instancecheck__(self, obj):
+            if self is Any:
+                raise TypeError("typing_extensions.Any cannot be used with isinstance()")
+            return super().__instancecheck__(obj)
+
+        def __repr__(self):
+            if self is Any:
+                return "typing_extensions.Any"
+            return super().__repr__()
+
+    class Any(metaclass=_AnyMeta):
+        """Special type indicating an unconstrained type.
+        - Any is compatible with every type.
+        - Any assumed to have all methods.
+        - All values assumed to be instances of Any.
+        Note that all the above statements are true from the point of view of
+        static type checkers. At runtime, Any should not be used with instance
+        checks.
+        """
+        def __new__(cls, *args, **kwargs):
+            if cls is Any:
+                raise TypeError("Any cannot be instantiated")
+            return super().__new__(cls, *args, **kwargs)
+
+
+ClassVar = typing.ClassVar
+
+
+class _ExtensionsSpecialForm(typing._SpecialForm, _root=True):
+    def __repr__(self):
+        return 'typing_extensions.' + self._name
+
+
+Final = typing.Final
+
+if sys.version_info >= (3, 11):
+    final = typing.final
+else:
+    # @final exists in 3.8+, but we backport it for all versions
+    # before 3.11 to keep support for the __final__ attribute.
+    # See https://bugs.python.org/issue46342
+    def final(f):
+        """This decorator can be used to indicate to type checkers that
+        the decorated method cannot be overridden, and decorated class
+        cannot be subclassed. For example:
+
+            class Base:
+                @final
+                def done(self) -> None:
+                    ...
+            class Sub(Base):
+                def done(self) -> None:  # Error reported by type checker
+                    ...
+            @final
+            class Leaf:
+                ...
+            class Other(Leaf):  # Error reported by type checker
+                ...
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__final__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+        """
+        try:
+            f.__final__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return f
+
+
+def IntVar(name):
+    return typing.TypeVar(name)
+
+
+# A Literal bug was fixed in 3.11.0, 3.10.1 and 3.9.8
+if sys.version_info >= (3, 10, 1):
+    Literal = typing.Literal
+else:
+    def _flatten_literal_params(parameters):
+        """An internal helper for Literal creation: flatten Literals among parameters"""
+        params = []
+        for p in parameters:
+            if isinstance(p, _LiteralGenericAlias):
+                params.extend(p.__args__)
+            else:
+                params.append(p)
+        return tuple(params)
+
+    def _value_and_type_iter(params):
+        for p in params:
+            yield p, type(p)
+
+    class _LiteralGenericAlias(typing._GenericAlias, _root=True):
+        def __eq__(self, other):
+            if not isinstance(other, _LiteralGenericAlias):
+                return NotImplemented
+            these_args_deduped = set(_value_and_type_iter(self.__args__))
+            other_args_deduped = set(_value_and_type_iter(other.__args__))
+            return these_args_deduped == other_args_deduped
+
+        def __hash__(self):
+            return hash(frozenset(_value_and_type_iter(self.__args__)))
+
+    class _LiteralForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, doc: str):
+            self._name = 'Literal'
+            self._doc = self.__doc__ = doc
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+
+            parameters = _flatten_literal_params(parameters)
+
+            val_type_pairs = list(_value_and_type_iter(parameters))
+            try:
+                deduped_pairs = set(val_type_pairs)
+            except TypeError:
+                # unhashable parameters
+                pass
+            else:
+                # similar logic to typing._deduplicate on Python 3.9+
+                if len(deduped_pairs) < len(val_type_pairs):
+                    new_parameters = []
+                    for pair in val_type_pairs:
+                        if pair in deduped_pairs:
+                            new_parameters.append(pair[0])
+                            deduped_pairs.remove(pair)
+                    assert not deduped_pairs, deduped_pairs
+                    parameters = tuple(new_parameters)
+
+            return _LiteralGenericAlias(self, parameters)
+
+    Literal = _LiteralForm(doc="""\
+                           A type that can be used to indicate to type checkers
+                           that the corresponding value has a value literally equivalent
+                           to the provided parameter. For example:
+
+                               var: Literal[4] = 4
+
+                           The type checker understands that 'var' is literally equal to
+                           the value 4 and no other value.
+
+                           Literal[...] cannot be subclassed. There is no runtime
+                           checking verifying that the parameter is actually a value
+                           instead of a type.""")
+
+
+_overload_dummy = typing._overload_dummy
+
+
+if hasattr(typing, "get_overloads"):  # 3.11+
+    overload = typing.overload
+    get_overloads = typing.get_overloads
+    clear_overloads = typing.clear_overloads
+else:
+    # {module: {qualname: {firstlineno: func}}}
+    _overload_registry = collections.defaultdict(
+        functools.partial(collections.defaultdict, dict)
+    )
+
+    def overload(func):
+        """Decorator for overloaded functions/methods.
+
+        In a stub file, place two or more stub definitions for the same
+        function in a row, each decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+
+        In a non-stub file (i.e. a regular .py file), do the same but
+        follow it with an implementation.  The implementation should *not*
+        be decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+        def utf8(value):
+            # implementation goes here
+
+        The overloads for a function can be retrieved at runtime using the
+        get_overloads() function.
+        """
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        try:
+            _overload_registry[f.__module__][f.__qualname__][
+                f.__code__.co_firstlineno
+            ] = func
+        except AttributeError:
+            # Not a normal function; ignore.
+            pass
+        return _overload_dummy
+
+    def get_overloads(func):
+        """Return all defined overloads for *func* as a sequence."""
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        if f.__module__ not in _overload_registry:
+            return []
+        mod_dict = _overload_registry[f.__module__]
+        if f.__qualname__ not in mod_dict:
+            return []
+        return list(mod_dict[f.__qualname__].values())
+
+    def clear_overloads():
+        """Clear all overloads in the registry."""
+        _overload_registry.clear()
+
+
+# This is not a real generic class.  Don't use outside annotations.
+Type = typing.Type
+
+# Various ABCs mimicking those in collections.abc.
+# A few are simply re-exported for completeness.
+Awaitable = typing.Awaitable
+Coroutine = typing.Coroutine
+AsyncIterable = typing.AsyncIterable
+AsyncIterator = typing.AsyncIterator
+Deque = typing.Deque
+DefaultDict = typing.DefaultDict
+OrderedDict = typing.OrderedDict
+Counter = typing.Counter
+ChainMap = typing.ChainMap
+Text = typing.Text
+TYPE_CHECKING = typing.TYPE_CHECKING
+
+
+if sys.version_info >= (3, 13, 0, "beta"):
+    from typing import AsyncContextManager, AsyncGenerator, ContextManager, Generator
+else:
+    def _is_dunder(attr):
+        return attr.startswith('__') and attr.endswith('__')
+
+    # Python <3.9 doesn't have typing._SpecialGenericAlias
+    _special_generic_alias_base = getattr(
+        typing, "_SpecialGenericAlias", typing._GenericAlias
+    )
+
+    class _SpecialGenericAlias(_special_generic_alias_base, _root=True):
+        def __init__(self, origin, nparams, *, inst=True, name=None, defaults=()):
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                self.__origin__ = origin
+                self._nparams = nparams
+                super().__init__(origin, nparams, special=True, inst=inst, name=name)
+            else:
+                # Python >= 3.9
+                super().__init__(origin, nparams, inst=inst, name=name)
+            self._defaults = defaults
+
+        def __setattr__(self, attr, val):
+            allowed_attrs = {'_name', '_inst', '_nparams', '_defaults'}
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                allowed_attrs.add("__origin__")
+            if _is_dunder(attr) or attr in allowed_attrs:
+                object.__setattr__(self, attr, val)
+            else:
+                setattr(self.__origin__, attr, val)
+
+        @typing._tp_cache
+        def __getitem__(self, params):
+            if not isinstance(params, tuple):
+                params = (params,)
+            msg = "Parameters to generic types must be types."
+            params = tuple(typing._type_check(p, msg) for p in params)
+            if (
+                self._defaults
+                and len(params) < self._nparams
+                and len(params) + len(self._defaults) >= self._nparams
+            ):
+                params = (*params, *self._defaults[len(params) - self._nparams:])
+            actual_len = len(params)
+
+            if actual_len != self._nparams:
+                if self._defaults:
+                    expected = f"at least {self._nparams - len(self._defaults)}"
+                else:
+                    expected = str(self._nparams)
+                if not self._nparams:
+                    raise TypeError(f"{self} is not a generic class")
+                raise TypeError(
+                    f"Too {'many' if actual_len > self._nparams else 'few'}"
+                    f" arguments for {self};"
+                    f" actual {actual_len}, expected {expected}"
+                )
+            return self.copy_with(params)
+
+    _NoneType = type(None)
+    Generator = _SpecialGenericAlias(
+        collections.abc.Generator, 3, defaults=(_NoneType, _NoneType)
+    )
+    AsyncGenerator = _SpecialGenericAlias(
+        collections.abc.AsyncGenerator, 2, defaults=(_NoneType,)
+    )
+    ContextManager = _SpecialGenericAlias(
+        contextlib.AbstractContextManager,
+        2,
+        name="ContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+    AsyncContextManager = _SpecialGenericAlias(
+        contextlib.AbstractAsyncContextManager,
+        2,
+        name="AsyncContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+
+
+_PROTO_ALLOWLIST = {
+    'collections.abc': [
+        'Callable', 'Awaitable', 'Iterable', 'Iterator', 'AsyncIterable',
+        'Hashable', 'Sized', 'Container', 'Collection', 'Reversible', 'Buffer',
+    ],
+    'contextlib': ['AbstractContextManager', 'AbstractAsyncContextManager'],
+    'typing_extensions': ['Buffer'],
+}
+
+
+_EXCLUDED_ATTRS = frozenset(typing.EXCLUDED_ATTRIBUTES) | {
+    "__match_args__", "__protocol_attrs__", "__non_callable_proto_members__",
+    "__final__",
+}
+
+
+def _get_protocol_attrs(cls):
+    attrs = set()
+    for base in cls.__mro__[:-1]:  # without object
+        if base.__name__ in {'Protocol', 'Generic'}:
+            continue
+        annotations = getattr(base, '__annotations__', {})
+        for attr in (*base.__dict__, *annotations):
+            if (not attr.startswith('_abc_') and attr not in _EXCLUDED_ATTRS):
+                attrs.add(attr)
+    return attrs
+
+
+def _caller(depth=2):
+    try:
+        return sys._getframe(depth).f_globals.get('__name__', '__main__')
+    except (AttributeError, ValueError):  # For platforms without _getframe()
+        return None
+
+
+# `__match_args__` attribute was removed from protocol members in 3.13,
+# we want to backport this change to older Python versions.
+if sys.version_info >= (3, 13):
+    Protocol = typing.Protocol
+else:
+    def _allow_reckless_class_checks(depth=3):
+        """Allow instance and class checks for special stdlib modules.
+        The abc and functools modules indiscriminately call isinstance() and
+        issubclass() on the whole MRO of a user class, which may contain protocols.
+        """
+        return _caller(depth) in {'abc', 'functools', None}
+
+    def _no_init(self, *args, **kwargs):
+        if type(self)._is_protocol:
+            raise TypeError('Protocols cannot be instantiated')
+
+    def _type_check_issubclass_arg_1(arg):
+        """Raise TypeError if `arg` is not an instance of `type`
+        in `issubclass(arg, <protocol>)`.
+
+        In most cases, this is verified by type.__subclasscheck__.
+        Checking it again unnecessarily would slow down issubclass() checks,
+        so, we don't perform this check unless we absolutely have to.
+
+        For various error paths, however,
+        we want to ensure that *this* error message is shown to the user
+        where relevant, rather than a typing.py-specific error message.
+        """
+        if not isinstance(arg, type):
+            # Same error message as for issubclass(1, int).
+            raise TypeError('issubclass() arg 1 must be a class')
+
+    # Inheriting from typing._ProtocolMeta isn't actually desirable,
+    # but is necessary to allow typing.Protocol and typing_extensions.Protocol
+    # to mix without getting TypeErrors about "metaclass conflict"
+    class _ProtocolMeta(type(typing.Protocol)):
+        # This metaclass is somewhat unfortunate,
+        # but is necessary for several reasons...
+        #
+        # NOTE: DO NOT call super() in any methods in this class
+        # That would call the methods on typing._ProtocolMeta on Python 3.8-3.11
+        # and those are slow
+        def __new__(mcls, name, bases, namespace, **kwargs):
+            if name == "Protocol" and len(bases) < 2:
+                pass
+            elif {Protocol, typing.Protocol} & set(bases):
+                for base in bases:
+                    if not (
+                        base in {object, typing.Generic, Protocol, typing.Protocol}
+                        or base.__name__ in _PROTO_ALLOWLIST.get(base.__module__, [])
+                        or is_protocol(base)
+                    ):
+                        raise TypeError(
+                            f"Protocols can only inherit from other protocols, "
+                            f"got {base!r}"
+                        )
+            return abc.ABCMeta.__new__(mcls, name, bases, namespace, **kwargs)
+
+        def __init__(cls, *args, **kwargs):
+            abc.ABCMeta.__init__(cls, *args, **kwargs)
+            if getattr(cls, "_is_protocol", False):
+                cls.__protocol_attrs__ = _get_protocol_attrs(cls)
+
+        def __subclasscheck__(cls, other):
+            if cls is Protocol:
+                return type.__subclasscheck__(cls, other)
+            if (
+                getattr(cls, '_is_protocol', False)
+                and not _allow_reckless_class_checks()
+            ):
+                if not getattr(cls, '_is_runtime_protocol', False):
+                    _type_check_issubclass_arg_1(other)
+                    raise TypeError(
+                        "Instance and class checks can only be used with "
+                        "@runtime_checkable protocols"
+                    )
+                if (
+                    # this attribute is set by @runtime_checkable:
+                    cls.__non_callable_proto_members__
+                    and cls.__dict__.get("__subclasshook__") is _proto_hook
+                ):
+                    _type_check_issubclass_arg_1(other)
+                    non_method_attrs = sorted(cls.__non_callable_proto_members__)
+                    raise TypeError(
+                        "Protocols with non-method members don't support issubclass()."
+                        f" Non-method members: {str(non_method_attrs)[1:-1]}."
+                    )
+            return abc.ABCMeta.__subclasscheck__(cls, other)
+
+        def __instancecheck__(cls, instance):
+            # We need this method for situations where attributes are
+            # assigned in __init__.
+            if cls is Protocol:
+                return type.__instancecheck__(cls, instance)
+            if not getattr(cls, "_is_protocol", False):
+                # i.e., it's a concrete subclass of a protocol
+                return abc.ABCMeta.__instancecheck__(cls, instance)
+
+            if (
+                not getattr(cls, '_is_runtime_protocol', False) and
+                not _allow_reckless_class_checks()
+            ):
+                raise TypeError("Instance and class checks can only be used with"
+                                " @runtime_checkable protocols")
+
+            if abc.ABCMeta.__instancecheck__(cls, instance):
+                return True
+
+            for attr in cls.__protocol_attrs__:
+                try:
+                    val = inspect.getattr_static(instance, attr)
+                except AttributeError:
+                    break
+                # this attribute is set by @runtime_checkable:
+                if val is None and attr not in cls.__non_callable_proto_members__:
+                    break
+            else:
+                return True
+
+            return False
+
+        def __eq__(cls, other):
+            # Hack so that typing.Generic.__class_getitem__
+            # treats typing_extensions.Protocol
+            # as equivalent to typing.Protocol
+            if abc.ABCMeta.__eq__(cls, other) is True:
+                return True
+            return cls is Protocol and other is typing.Protocol
+
+        # This has to be defined, or the abc-module cache
+        # complains about classes with this metaclass being unhashable,
+        # if we define only __eq__!
+        def __hash__(cls) -> int:
+            return type.__hash__(cls)
+
+    @classmethod
+    def _proto_hook(cls, other):
+        if not cls.__dict__.get('_is_protocol', False):
+            return NotImplemented
+
+        for attr in cls.__protocol_attrs__:
+            for base in other.__mro__:
+                # Check if the members appears in the class dictionary...
+                if attr in base.__dict__:
+                    if base.__dict__[attr] is None:
+                        return NotImplemented
+                    break
+
+                # ...or in annotations, if it is a sub-protocol.
+                annotations = getattr(base, '__annotations__', {})
+                if (
+                    isinstance(annotations, collections.abc.Mapping)
+                    and attr in annotations
+                    and is_protocol(other)
+                ):
+                    break
+            else:
+                return NotImplemented
+        return True
+
+    class Protocol(typing.Generic, metaclass=_ProtocolMeta):
+        __doc__ = typing.Protocol.__doc__
+        __slots__ = ()
+        _is_protocol = True
+        _is_runtime_protocol = False
+
+        def __init_subclass__(cls, *args, **kwargs):
+            super().__init_subclass__(*args, **kwargs)
+
+            # Determine if this is a protocol or a concrete subclass.
+            if not cls.__dict__.get('_is_protocol', False):
+                cls._is_protocol = any(b is Protocol for b in cls.__bases__)
+
+            # Set (or override) the protocol subclass hook.
+            if '__subclasshook__' not in cls.__dict__:
+                cls.__subclasshook__ = _proto_hook
+
+            # Prohibit instantiation for protocol classes
+            if cls._is_protocol and cls.__init__ is Protocol.__init__:
+                cls.__init__ = _no_init
+
+
+if sys.version_info >= (3, 13):
+    runtime_checkable = typing.runtime_checkable
+else:
+    def runtime_checkable(cls):
+        """Mark a protocol class as a runtime protocol.
+
+        Such protocol can be used with isinstance() and issubclass().
+        Raise TypeError if applied to a non-protocol class.
+        This allows a simple-minded structural check very similar to
+        one trick ponies in collections.abc such as Iterable.
+
+        For example::
+
+            @runtime_checkable
+            class Closable(Protocol):
+                def close(self): ...
+
+            assert isinstance(open('/some/file'), Closable)
+
+        Warning: this will check only the presence of the required methods,
+        not their type signatures!
+        """
+        if not issubclass(cls, typing.Generic) or not getattr(cls, '_is_protocol', False):
+            raise TypeError(f'@runtime_checkable can be only applied to protocol classes,'
+                            f' got {cls!r}')
+        cls._is_runtime_protocol = True
+
+        # typing.Protocol classes on <=3.11 break if we execute this block,
+        # because typing.Protocol classes on <=3.11 don't have a
+        # `__protocol_attrs__` attribute, and this block relies on the
+        # `__protocol_attrs__` attribute. Meanwhile, typing.Protocol classes on 3.12.2+
+        # break if we *don't* execute this block, because *they* assume that all
+        # protocol classes have a `__non_callable_proto_members__` attribute
+        # (which this block sets)
+        if isinstance(cls, _ProtocolMeta) or sys.version_info >= (3, 12, 2):
+            # PEP 544 prohibits using issubclass()
+            # with protocols that have non-method members.
+            # See gh-113320 for why we compute this attribute here,
+            # rather than in `_ProtocolMeta.__init__`
+            cls.__non_callable_proto_members__ = set()
+            for attr in cls.__protocol_attrs__:
+                try:
+                    is_callable = callable(getattr(cls, attr, None))
+                except Exception as e:
+                    raise TypeError(
+                        f"Failed to determine whether protocol member {attr!r} "
+                        "is a method member"
+                    ) from e
+                else:
+                    if not is_callable:
+                        cls.__non_callable_proto_members__.add(attr)
+
+        return cls
+
+
+# The "runtime" alias exists for backwards compatibility.
+runtime = runtime_checkable
+
+
+# Our version of runtime-checkable protocols is faster on Python 3.8-3.11
+if sys.version_info >= (3, 12):
+    SupportsInt = typing.SupportsInt
+    SupportsFloat = typing.SupportsFloat
+    SupportsComplex = typing.SupportsComplex
+    SupportsBytes = typing.SupportsBytes
+    SupportsIndex = typing.SupportsIndex
+    SupportsAbs = typing.SupportsAbs
+    SupportsRound = typing.SupportsRound
+else:
+    @runtime_checkable
+    class SupportsInt(Protocol):
+        """An ABC with one abstract method __int__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __int__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsFloat(Protocol):
+        """An ABC with one abstract method __float__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __float__(self) -> float:
+            pass
+
+    @runtime_checkable
+    class SupportsComplex(Protocol):
+        """An ABC with one abstract method __complex__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __complex__(self) -> complex:
+            pass
+
+    @runtime_checkable
+    class SupportsBytes(Protocol):
+        """An ABC with one abstract method __bytes__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __bytes__(self) -> bytes:
+            pass
+
+    @runtime_checkable
+    class SupportsIndex(Protocol):
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __index__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsAbs(Protocol[T_co]):
+        """
+        An ABC with one abstract method __abs__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __abs__(self) -> T_co:
+            pass
+
+    @runtime_checkable
+    class SupportsRound(Protocol[T_co]):
+        """
+        An ABC with one abstract method __round__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __round__(self, ndigits: int = 0) -> T_co:
+            pass
+
+
+def _ensure_subclassable(mro_entries):
+    def inner(func):
+        if sys.implementation.name == "pypy" and sys.version_info < (3, 9):
+            cls_dict = {
+                "__call__": staticmethod(func),
+                "__mro_entries__": staticmethod(mro_entries)
+            }
+            t = type(func.__name__, (), cls_dict)
+            return functools.update_wrapper(t(), func)
+        else:
+            func.__mro_entries__ = mro_entries
+            return func
+    return inner
+
+
+# Update this to something like >=3.13.0b1 if and when
+# PEP 728 is implemented in CPython
+_PEP_728_IMPLEMENTED = False
+
+if _PEP_728_IMPLEMENTED:
+    # The standard library TypedDict in Python 3.8 does not store runtime information
+    # about which (if any) keys are optional.  See https://bugs.python.org/issue38834
+    # The standard library TypedDict in Python 3.9.0/1 does not honour the "total"
+    # keyword with old-style TypedDict().  See https://bugs.python.org/issue42059
+    # The standard library TypedDict below Python 3.11 does not store runtime
+    # information about optional and required keys when using Required or NotRequired.
+    # Generic TypedDicts are also impossible using typing.TypedDict on Python <3.11.
+    # Aaaand on 3.12 we add __orig_bases__ to TypedDict
+    # to enable better runtime introspection.
+    # On 3.13 we deprecate some odd ways of creating TypedDicts.
+    # Also on 3.13, PEP 705 adds the ReadOnly[] qualifier.
+    # PEP 728 (still pending) makes more changes.
+    TypedDict = typing.TypedDict
+    _TypedDictMeta = typing._TypedDictMeta
+    is_typeddict = typing.is_typeddict
+else:
+    # 3.10.0 and later
+    _TAKES_MODULE = "module" in inspect.signature(typing._type_check).parameters
+
+    def _get_typeddict_qualifiers(annotation_type):
+        while True:
+            annotation_origin = get_origin(annotation_type)
+            if annotation_origin is Annotated:
+                annotation_args = get_args(annotation_type)
+                if annotation_args:
+                    annotation_type = annotation_args[0]
+                else:
+                    break
+            elif annotation_origin is Required:
+                yield Required
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is NotRequired:
+                yield NotRequired
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is ReadOnly:
+                yield ReadOnly
+                annotation_type, = get_args(annotation_type)
+            else:
+                break
+
+    class _TypedDictMeta(type):
+        def __new__(cls, name, bases, ns, *, total=True, closed=False):
+            """Create new typed dict class object.
+
+            This method is called when TypedDict is subclassed,
+            or when TypedDict is instantiated. This way
+            TypedDict supports all three syntax forms described in its docstring.
+            Subclasses and instances of TypedDict return actual dictionaries.
+            """
+            for base in bases:
+                if type(base) is not _TypedDictMeta and base is not typing.Generic:
+                    raise TypeError('cannot inherit from both a TypedDict type '
+                                    'and a non-TypedDict base class')
+
+            if any(issubclass(b, typing.Generic) for b in bases):
+                generic_base = (typing.Generic,)
+            else:
+                generic_base = ()
+
+            # typing.py generally doesn't let you inherit from plain Generic, unless
+            # the name of the class happens to be "Protocol"
+            tp_dict = type.__new__(_TypedDictMeta, "Protocol", (*generic_base, dict), ns)
+            tp_dict.__name__ = name
+            if tp_dict.__qualname__ == "Protocol":
+                tp_dict.__qualname__ = name
+
+            if not hasattr(tp_dict, '__orig_bases__'):
+                tp_dict.__orig_bases__ = bases
+
+            annotations = {}
+            if "__annotations__" in ns:
+                own_annotations = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                own_annotations = ns["__annotate__"](1)
+            else:
+                own_annotations = {}
+            msg = "TypedDict('Name', {f0: t0, f1: t1, ...}); each t must be a type"
+            if _TAKES_MODULE:
+                own_annotations = {
+                    n: typing._type_check(tp, msg, module=tp_dict.__module__)
+                    for n, tp in own_annotations.items()
+                }
+            else:
+                own_annotations = {
+                    n: typing._type_check(tp, msg)
+                    for n, tp in own_annotations.items()
+                }
+            required_keys = set()
+            optional_keys = set()
+            readonly_keys = set()
+            mutable_keys = set()
+            extra_items_type = None
+
+            for base in bases:
+                base_dict = base.__dict__
+
+                annotations.update(base_dict.get('__annotations__', {}))
+                required_keys.update(base_dict.get('__required_keys__', ()))
+                optional_keys.update(base_dict.get('__optional_keys__', ()))
+                readonly_keys.update(base_dict.get('__readonly_keys__', ()))
+                mutable_keys.update(base_dict.get('__mutable_keys__', ()))
+                base_extra_items_type = base_dict.get('__extra_items__', None)
+                if base_extra_items_type is not None:
+                    extra_items_type = base_extra_items_type
+
+            if closed and extra_items_type is None:
+                extra_items_type = Never
+            if closed and "__extra_items__" in own_annotations:
+                annotation_type = own_annotations.pop("__extra_items__")
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+                if Required in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "Required"
+                    )
+                if NotRequired in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "NotRequired"
+                    )
+                extra_items_type = annotation_type
+
+            annotations.update(own_annotations)
+            for annotation_key, annotation_type in own_annotations.items():
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+
+                if Required in qualifiers:
+                    required_keys.add(annotation_key)
+                elif NotRequired in qualifiers:
+                    optional_keys.add(annotation_key)
+                elif total:
+                    required_keys.add(annotation_key)
+                else:
+                    optional_keys.add(annotation_key)
+                if ReadOnly in qualifiers:
+                    mutable_keys.discard(annotation_key)
+                    readonly_keys.add(annotation_key)
+                else:
+                    mutable_keys.add(annotation_key)
+                    readonly_keys.discard(annotation_key)
+
+            tp_dict.__annotations__ = annotations
+            tp_dict.__required_keys__ = frozenset(required_keys)
+            tp_dict.__optional_keys__ = frozenset(optional_keys)
+            tp_dict.__readonly_keys__ = frozenset(readonly_keys)
+            tp_dict.__mutable_keys__ = frozenset(mutable_keys)
+            if not hasattr(tp_dict, '__total__'):
+                tp_dict.__total__ = total
+            tp_dict.__closed__ = closed
+            tp_dict.__extra_items__ = extra_items_type
+            return tp_dict
+
+        __call__ = dict  # static method
+
+        def __subclasscheck__(cls, other):
+            # Typed dicts are only for static structural subtyping.
+            raise TypeError('TypedDict does not support instance and class checks')
+
+        __instancecheck__ = __subclasscheck__
+
+    _TypedDict = type.__new__(_TypedDictMeta, 'TypedDict', (), {})
+
+    @_ensure_subclassable(lambda bases: (_TypedDict,))
+    def TypedDict(typename, fields=_marker, /, *, total=True, closed=False, **kwargs):
+        """A simple typed namespace. At runtime it is equivalent to a plain dict.
+
+        TypedDict creates a dictionary type such that a type checker will expect all
+        instances to have a certain set of keys, where each key is
+        associated with a value of a consistent type. This expectation
+        is not checked at runtime.
+
+        Usage::
+
+            class Point2D(TypedDict):
+                x: int
+                y: int
+                label: str
+
+            a: Point2D = {'x': 1, 'y': 2, 'label': 'good'}  # OK
+            b: Point2D = {'z': 3, 'label': 'bad'}           # Fails type check
+
+            assert Point2D(x=1, y=2, label='first') == dict(x=1, y=2, label='first')
+
+        The type info can be accessed via the Point2D.__annotations__ dict, and
+        the Point2D.__required_keys__ and Point2D.__optional_keys__ frozensets.
+        TypedDict supports an additional equivalent form::
+
+            Point2D = TypedDict('Point2D', {'x': int, 'y': int, 'label': str})
+
+        By default, all keys must be present in a TypedDict. It is possible
+        to override this by specifying totality::
+
+            class Point2D(TypedDict, total=False):
+                x: int
+                y: int
+
+        This means that a Point2D TypedDict can have any of the keys omitted. A type
+        checker is only expected to support a literal False or True as the value of
+        the total argument. True is the default, and makes all items defined in the
+        class body be required.
+
+        The Required and NotRequired special forms can also be used to mark
+        individual keys as being required or not required::
+
+            class Point2D(TypedDict):
+                x: int  # the "x" key must always be present (Required is the default)
+                y: NotRequired[int]  # the "y" key can be omitted
+
+        See PEP 655 for more details on Required and NotRequired.
+        """
+        if fields is _marker or fields is None:
+            if fields is _marker:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+
+            example = f"`{typename} = TypedDict({typename!r}, {{}})`"
+            deprecation_msg = (
+                f"{deprecated_thing} is deprecated and will be disallowed in "
+                "Python 3.15. To create a TypedDict class with 0 fields "
+                "using the functional syntax, pass an empty dictionary, e.g. "
+            ) + example + "."
+            warnings.warn(deprecation_msg, DeprecationWarning, stacklevel=2)
+            if closed is not False and closed is not True:
+                kwargs["closed"] = closed
+                closed = False
+            fields = kwargs
+        elif kwargs:
+            raise TypeError("TypedDict takes either a dict or keyword arguments,"
+                            " but not both")
+        if kwargs:
+            if sys.version_info >= (3, 13):
+                raise TypeError("TypedDict takes no keyword arguments")
+            warnings.warn(
+                "The kwargs-based syntax for TypedDict definitions is deprecated "
+                "in Python 3.11, will be removed in Python 3.13, and may not be "
+                "understood by third-party type checkers.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+
+        ns = {'__annotations__': dict(fields)}
+        module = _caller()
+        if module is not None:
+            # Setting correct module is necessary to make typed dict classes pickleable.
+            ns['__module__'] = module
+
+        td = _TypedDictMeta(typename, (), ns, total=total, closed=closed)
+        td.__orig_bases__ = (TypedDict,)
+        return td
+
+    if hasattr(typing, "_TypedDictMeta"):
+        _TYPEDDICT_TYPES = (typing._TypedDictMeta, _TypedDictMeta)
+    else:
+        _TYPEDDICT_TYPES = (_TypedDictMeta,)
+
+    def is_typeddict(tp):
+        """Check if an annotation is a TypedDict class
+
+        For example::
+            class Film(TypedDict):
+                title: str
+                year: int
+
+            is_typeddict(Film)  # => True
+            is_typeddict(Union[list, str])  # => False
+        """
+        # On 3.8, this would otherwise return True
+        if hasattr(typing, "TypedDict") and tp is typing.TypedDict:
+            return False
+        return isinstance(tp, _TYPEDDICT_TYPES)
+
+
+if hasattr(typing, "assert_type"):
+    assert_type = typing.assert_type
+
+else:
+    def assert_type(val, typ, /):
+        """Assert (to the type checker) that the value is of the given type.
+
+        When the type checker encounters a call to assert_type(), it
+        emits an error if the value is not of the specified type::
+
+            def greet(name: str) -> None:
+                assert_type(name, str)  # ok
+                assert_type(name, int)  # type checker error
+
+        At runtime this returns the first argument unchanged and otherwise
+        does nothing.
+        """
+        return val
+
+
+if hasattr(typing, "ReadOnly"):  # 3.13+
+    get_type_hints = typing.get_type_hints
+else:  # <=3.13
+    # replaces _strip_annotations()
+    def _strip_extras(t):
+        """Strips Annotated, Required and NotRequired from a given type."""
+        if isinstance(t, _AnnotatedAlias):
+            return _strip_extras(t.__origin__)
+        if hasattr(t, "__origin__") and t.__origin__ in (Required, NotRequired, ReadOnly):
+            return _strip_extras(t.__args__[0])
+        if isinstance(t, typing._GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return t.copy_with(stripped_args)
+        if hasattr(_types, "GenericAlias") and isinstance(t, _types.GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return _types.GenericAlias(t.__origin__, stripped_args)
+        if hasattr(_types, "UnionType") and isinstance(t, _types.UnionType):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return functools.reduce(operator.or_, stripped_args)
+
+        return t
+
+    def get_type_hints(obj, globalns=None, localns=None, include_extras=False):
+        """Return type hints for an object.
+
+        This is often the same as obj.__annotations__, but it handles
+        forward references encoded as string literals, adds Optional[t] if a
+        default value equal to None is set and recursively replaces all
+        'Annotated[T, ...]', 'Required[T]' or 'NotRequired[T]' with 'T'
+        (unless 'include_extras=True').
+
+        The argument may be a module, class, method, or function. The annotations
+        are returned as a dictionary. For classes, annotations include also
+        inherited members.
+
+        TypeError is raised if the argument is not of a type that can contain
+        annotations, and an empty dictionary is returned if no annotations are
+        present.
+
+        BEWARE -- the behavior of globalns and localns is counterintuitive
+        (unless you are familiar with how eval() and exec() work).  The
+        search order is locals first, then globals.
+
+        - If no dict arguments are passed, an attempt is made to use the
+          globals from obj (or the respective module's globals for classes),
+          and these are also used as the locals.  If the object does not appear
+          to have globals, an empty dictionary is used.
+
+        - If one dict argument is passed, it is used for both globals and
+          locals.
+
+        - If two dict arguments are passed, they specify globals and
+          locals, respectively.
+        """
+        if hasattr(typing, "Annotated"):  # 3.9+
+            hint = typing.get_type_hints(
+                obj, globalns=globalns, localns=localns, include_extras=True
+            )
+        else:  # 3.8
+            hint = typing.get_type_hints(obj, globalns=globalns, localns=localns)
+        if include_extras:
+            return hint
+        return {k: _strip_extras(t) for k, t in hint.items()}
+
+
+# Python 3.9+ has PEP 593 (Annotated)
+if hasattr(typing, 'Annotated'):
+    Annotated = typing.Annotated
+    # Not exported and not a public API, but needed for get_origin() and get_args()
+    # to work.
+    _AnnotatedAlias = typing._AnnotatedAlias
+# 3.8
+else:
+    class _AnnotatedAlias(typing._GenericAlias, _root=True):
+        """Runtime representation of an annotated type.
+
+        At its core 'Annotated[t, dec1, dec2, ...]' is an alias for the type 't'
+        with extra annotations. The alias behaves like a normal typing alias,
+        instantiating is the same as instantiating the underlying type, binding
+        it to types is also the same.
+        """
+        def __init__(self, origin, metadata):
+            if isinstance(origin, _AnnotatedAlias):
+                metadata = origin.__metadata__ + metadata
+                origin = origin.__origin__
+            super().__init__(origin, origin)
+            self.__metadata__ = metadata
+
+        def copy_with(self, params):
+            assert len(params) == 1
+            new_type = params[0]
+            return _AnnotatedAlias(new_type, self.__metadata__)
+
+        def __repr__(self):
+            return (f"typing_extensions.Annotated[{typing._type_repr(self.__origin__)}, "
+                    f"{', '.join(repr(a) for a in self.__metadata__)}]")
+
+        def __reduce__(self):
+            return operator.getitem, (
+                Annotated, (self.__origin__, *self.__metadata__)
+            )
+
+        def __eq__(self, other):
+            if not isinstance(other, _AnnotatedAlias):
+                return NotImplemented
+            if self.__origin__ != other.__origin__:
+                return False
+            return self.__metadata__ == other.__metadata__
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__metadata__))
+
+    class Annotated:
+        """Add context specific metadata to a type.
+
+        Example: Annotated[int, runtime_check.Unsigned] indicates to the
+        hypothetical runtime_check module that this type is an unsigned int.
+        Every other consumer of this type can ignore this metadata and treat
+        this type as int.
+
+        The first argument to Annotated must be a valid type (and will be in
+        the __origin__ field), the remaining arguments are kept as a tuple in
+        the __extra__ field.
+
+        Details:
+
+        - It's an error to call `Annotated` with less than two arguments.
+        - Nested Annotated are flattened::
+
+            Annotated[Annotated[T, Ann1, Ann2], Ann3] == Annotated[T, Ann1, Ann2, Ann3]
+
+        - Instantiating an annotated type is equivalent to instantiating the
+        underlying type::
+
+            Annotated[C, Ann1](5) == C(5)
+
+        - Annotated can be used as a generic type alias::
+
+            Optimized = Annotated[T, runtime.Optimize()]
+            Optimized[int] == Annotated[int, runtime.Optimize()]
+
+            OptimizedList = Annotated[List[T], runtime.Optimize()]
+            OptimizedList[int] == Annotated[List[int], runtime.Optimize()]
+        """
+
+        __slots__ = ()
+
+        def __new__(cls, *args, **kwargs):
+            raise TypeError("Type Annotated cannot be instantiated.")
+
+        @typing._tp_cache
+        def __class_getitem__(cls, params):
+            if not isinstance(params, tuple) or len(params) < 2:
+                raise TypeError("Annotated[...] should be used "
+                                "with at least two arguments (a type and an "
+                                "annotation).")
+            allowed_special_forms = (ClassVar, Final)
+            if get_origin(params[0]) in allowed_special_forms:
+                origin = params[0]
+            else:
+                msg = "Annotated[t, ...]: t must be a type."
+                origin = typing._type_check(params[0], msg)
+            metadata = tuple(params[1:])
+            return _AnnotatedAlias(origin, metadata)
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                f"Cannot subclass {cls.__module__}.Annotated"
+            )
+
+# Python 3.8 has get_origin() and get_args() but those implementations aren't
+# Annotated-aware, so we can't use those. Python 3.9's versions don't support
+# ParamSpecArgs and ParamSpecKwargs, so only Python 3.10's versions will do.
+if sys.version_info[:2] >= (3, 10):
+    get_origin = typing.get_origin
+    get_args = typing.get_args
+# 3.8-3.9
+else:
+    try:
+        # 3.9+
+        from typing import _BaseGenericAlias
+    except ImportError:
+        _BaseGenericAlias = typing._GenericAlias
+    try:
+        # 3.9+
+        from typing import GenericAlias as _typing_GenericAlias
+    except ImportError:
+        _typing_GenericAlias = typing._GenericAlias
+
+    def get_origin(tp):
+        """Get the unsubscripted version of a type.
+
+        This supports generic types, Callable, Tuple, Union, Literal, Final, ClassVar
+        and Annotated. Return None for unsupported types. Examples::
+
+            get_origin(Literal[42]) is Literal
+            get_origin(int) is None
+            get_origin(ClassVar[int]) is ClassVar
+            get_origin(Generic) is Generic
+            get_origin(Generic[T]) is Generic
+            get_origin(Union[T, int]) is Union
+            get_origin(List[Tuple[T, T]][int]) == list
+            get_origin(P.args) is P
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return Annotated
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias, _BaseGenericAlias,
+                           ParamSpecArgs, ParamSpecKwargs)):
+            return tp.__origin__
+        if tp is typing.Generic:
+            return typing.Generic
+        return None
+
+    def get_args(tp):
+        """Get type arguments with all substitutions performed.
+
+        For unions, basic simplifications used by Union constructor are performed.
+        Examples::
+            get_args(Dict[str, int]) == (str, int)
+            get_args(int) == ()
+            get_args(Union[int, Union[T, int], str][int]) == (int, str)
+            get_args(Union[int, Tuple[T, int]][str]) == (int, Tuple[str, int])
+            get_args(Callable[[], T][int]) == ([], int)
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return (tp.__origin__, *tp.__metadata__)
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias)):
+            if getattr(tp, "_special", False):
+                return ()
+            res = tp.__args__
+            if get_origin(tp) is collections.abc.Callable and res[0] is not Ellipsis:
+                res = (list(res[:-1]), res[-1])
+            return res
+        return ()
+
+
+# 3.10+
+if hasattr(typing, 'TypeAlias'):
+    TypeAlias = typing.TypeAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeAlias(self, parameters):
+        """Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example above.
+        """
+        raise TypeError(f"{self} is not subscriptable")
+# 3.8
+else:
+    TypeAlias = _ExtensionsSpecialForm(
+        'TypeAlias',
+        doc="""Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example
+        above."""
+    )
+
+
+if hasattr(typing, "NoDefault"):
+    NoDefault = typing.NoDefault
+else:
+    class NoDefaultTypeMeta(type):
+        def __setattr__(cls, attr, value):
+            # TypeError is consistent with the behavior of NoneType
+            raise TypeError(
+                f"cannot set {attr!r} attribute of immutable type {cls.__name__!r}"
+            )
+
+    class NoDefaultType(metaclass=NoDefaultTypeMeta):
+        """The type of the NoDefault singleton."""
+
+        __slots__ = ()
+
+        def __new__(cls):
+            return globals().get("NoDefault") or object.__new__(cls)
+
+        def __repr__(self):
+            return "typing_extensions.NoDefault"
+
+        def __reduce__(self):
+            return "NoDefault"
+
+    NoDefault = NoDefaultType()
+    del NoDefaultType, NoDefaultTypeMeta
+
+
+def _set_default(type_param, default):
+    type_param.has_default = lambda: default is not NoDefault
+    type_param.__default__ = default
+
+
+def _set_module(typevarlike):
+    # for pickling:
+    def_mod = _caller(depth=3)
+    if def_mod != 'typing_extensions':
+        typevarlike.__module__ = def_mod
+
+
+class _DefaultMixin:
+    """Mixin for TypeVarLike defaults."""
+
+    __slots__ = ()
+    __init__ = _set_default
+
+
+# Classes using this metaclass must provide a _backported_typevarlike ClassVar
+class _TypeVarLikeMeta(type):
+    def __instancecheck__(cls, __instance: Any) -> bool:
+        return isinstance(__instance, cls._backported_typevarlike)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVar
+else:
+    # Add default and infer_variance parameters from PEP 696 and 695
+    class TypeVar(metaclass=_TypeVarLikeMeta):
+        """Type variable."""
+
+        _backported_typevarlike = typing.TypeVar
+
+        def __new__(cls, name, *constraints, bound=None,
+                    covariant=False, contravariant=False,
+                    default=NoDefault, infer_variance=False):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented (3.12+), can pass infer_variance to typing.TypeVar
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant,
+                                         infer_variance=infer_variance)
+            else:
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant)
+                if infer_variance and (covariant or contravariant):
+                    raise ValueError("Variance cannot be specified with infer_variance.")
+                typevar.__infer_variance__ = infer_variance
+
+            _set_default(typevar, default)
+            _set_module(typevar)
+
+            def _tvar_prepare_subst(alias, args):
+                if (
+                    typevar.has_default()
+                    and alias.__parameters__.index(typevar) == len(args)
+                ):
+                    args += (typevar.__default__,)
+                return args
+
+            typevar.__typing_prepare_subst__ = _tvar_prepare_subst
+            return typevar
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.TypeVar' is not an acceptable base type")
+
+
+# Python 3.10+ has PEP 612
+if hasattr(typing, 'ParamSpecArgs'):
+    ParamSpecArgs = typing.ParamSpecArgs
+    ParamSpecKwargs = typing.ParamSpecKwargs
+# 3.8-3.9
+else:
+    class _Immutable:
+        """Mixin to indicate that object should not be copied."""
+        __slots__ = ()
+
+        def __copy__(self):
+            return self
+
+        def __deepcopy__(self, memo):
+            return self
+
+    class ParamSpecArgs(_Immutable):
+        """The args for a ParamSpec object.
+
+        Given a ParamSpec object P, P.args is an instance of ParamSpecArgs.
+
+        ParamSpecArgs objects have a reference back to their ParamSpec:
+
+        P.args.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.args"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecArgs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+    class ParamSpecKwargs(_Immutable):
+        """The kwargs for a ParamSpec object.
+
+        Given a ParamSpec object P, P.kwargs is an instance of ParamSpecKwargs.
+
+        ParamSpecKwargs objects have a reference back to their ParamSpec:
+
+        P.kwargs.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.kwargs"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecKwargs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import ParamSpec
+
+# 3.10+
+elif hasattr(typing, 'ParamSpec'):
+
+    # Add default parameter - PEP 696
+    class ParamSpec(metaclass=_TypeVarLikeMeta):
+        """Parameter specification."""
+
+        _backported_typevarlike = typing.ParamSpec
+
+        def __new__(cls, name, *, bound=None,
+                    covariant=False, contravariant=False,
+                    infer_variance=False, default=NoDefault):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented, can pass infer_variance to typing.TypeVar
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant,
+                                             infer_variance=infer_variance)
+            else:
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant)
+                paramspec.__infer_variance__ = infer_variance
+
+            _set_default(paramspec, default)
+            _set_module(paramspec)
+
+            def _paramspec_prepare_subst(alias, args):
+                params = alias.__parameters__
+                i = params.index(paramspec)
+                if i == len(args) and paramspec.has_default():
+                    args = [*args, paramspec.__default__]
+                if i >= len(args):
+                    raise TypeError(f"Too few arguments for {alias}")
+                # Special case where Z[[int, str, bool]] == Z[int, str, bool] in PEP 612.
+                if len(params) == 1 and not typing._is_param_expr(args[0]):
+                    assert i == 0
+                    args = (args,)
+                # Convert lists to tuples to help other libraries cache the results.
+                elif isinstance(args[i], list):
+                    args = (*args[:i], tuple(args[i]), *args[i + 1:])
+                return args
+
+            paramspec.__typing_prepare_subst__ = _paramspec_prepare_subst
+            return paramspec
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.ParamSpec' is not an acceptable base type")
+
+# 3.8-3.9
+else:
+
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class ParamSpec(list, _DefaultMixin):
+        """Parameter specification variable.
+
+        Usage::
+
+           P = ParamSpec('P')
+
+        Parameter specification variables exist primarily for the benefit of static
+        type checkers.  They are used to forward the parameter types of one
+        callable to another callable, a pattern commonly found in higher order
+        functions and decorators.  They are only valid when used in ``Concatenate``,
+        or s the first argument to ``Callable``. In Python 3.10 and higher,
+        they are also supported in user-defined Generics at runtime.
+        See class Generic for more information on generic types.  An
+        example for annotating a decorator::
+
+           T = TypeVar('T')
+           P = ParamSpec('P')
+
+           def add_logging(f: Callable[P, T]) -> Callable[P, T]:
+               '''A type-safe decorator to add logging to a function.'''
+               def inner(*args: P.args, **kwargs: P.kwargs) -> T:
+                   logging.info(f'{f.__name__} was called')
+                   return f(*args, **kwargs)
+               return inner
+
+           @add_logging
+           def add_two(x: float, y: float) -> float:
+               '''Add two numbers together.'''
+               return x + y
+
+        Parameter specification variables defined with covariant=True or
+        contravariant=True can be used to declare covariant or contravariant
+        generic types.  These keyword arguments are valid, but their actual semantics
+        are yet to be decided.  See PEP 612 for details.
+
+        Parameter specification variables can be introspected. e.g.:
+
+           P.__name__ == 'T'
+           P.__bound__ == None
+           P.__covariant__ == False
+           P.__contravariant__ == False
+
+        Note that only parameter specification variables defined in global scope can
+        be pickled.
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        @property
+        def args(self):
+            return ParamSpecArgs(self)
+
+        @property
+        def kwargs(self):
+            return ParamSpecKwargs(self)
+
+        def __init__(self, name, *, bound=None, covariant=False, contravariant=False,
+                     infer_variance=False, default=NoDefault):
+            list.__init__(self, [self])
+            self.__name__ = name
+            self.__covariant__ = bool(covariant)
+            self.__contravariant__ = bool(contravariant)
+            self.__infer_variance__ = bool(infer_variance)
+            if bound:
+                self.__bound__ = typing._type_check(bound, 'Bound must be a type.')
+            else:
+                self.__bound__ = None
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __repr__(self):
+            if self.__infer_variance__:
+                prefix = ''
+            elif self.__covariant__:
+                prefix = '+'
+            elif self.__contravariant__:
+                prefix = '-'
+            else:
+                prefix = '~'
+            return prefix + self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        # Hack to get typing._type_check to pass.
+        def __call__(self, *args, **kwargs):
+            pass
+
+
+# 3.8-3.9
+if not hasattr(typing, 'Concatenate'):
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class _ConcatenateGenericAlias(list):
+
+        # Trick Generic into looking into this for __parameters__.
+        __class__ = typing._GenericAlias
+
+        # Flag in 3.8.
+        _special = False
+
+        def __init__(self, origin, args):
+            super().__init__(args)
+            self.__origin__ = origin
+            self.__args__ = args
+
+        def __repr__(self):
+            _type_repr = typing._type_repr
+            return (f'{_type_repr(self.__origin__)}'
+                    f'[{", ".join(_type_repr(arg) for arg in self.__args__)}]')
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__args__))
+
+        # Hack to get typing._type_check to pass in Generic.
+        def __call__(self, *args, **kwargs):
+            pass
+
+        @property
+        def __parameters__(self):
+            return tuple(
+                tp for tp in self.__args__ if isinstance(tp, (typing.TypeVar, ParamSpec))
+            )
+
+
+# 3.8-3.9
+@typing._tp_cache
+def _concatenate_getitem(self, parameters):
+    if parameters == ():
+        raise TypeError("Cannot take a Concatenate of no types.")
+    if not isinstance(parameters, tuple):
+        parameters = (parameters,)
+    if not isinstance(parameters[-1], ParamSpec):
+        raise TypeError("The last parameter to Concatenate should be a "
+                        "ParamSpec variable.")
+    msg = "Concatenate[arg, ...]: each arg must be a type."
+    parameters = tuple(typing._type_check(p, msg) for p in parameters)
+    return _ConcatenateGenericAlias(self, parameters)
+
+
+# 3.10+
+if hasattr(typing, 'Concatenate'):
+    Concatenate = typing.Concatenate
+    _ConcatenateGenericAlias = typing._ConcatenateGenericAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def Concatenate(self, parameters):
+        """Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """
+        return _concatenate_getitem(self, parameters)
+# 3.8
+else:
+    class _ConcatenateForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            return _concatenate_getitem(self, parameters)
+
+    Concatenate = _ConcatenateForm(
+        'Concatenate',
+        doc="""Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """)
+
+# 3.10+
+if hasattr(typing, 'TypeGuard'):
+    TypeGuard = typing.TypeGuard
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeGuard(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeGuardForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeGuard = _TypeGuardForm(
+        'TypeGuard',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """)
+
+# 3.13+
+if hasattr(typing, 'TypeIs'):
+    TypeIs = typing.TypeIs
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeIs(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeIsForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeIs = _TypeIsForm(
+        'TypeIs',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """)
+
+
+# Vendored from cpython typing._SpecialFrom
+class _SpecialForm(typing._Final, _root=True):
+    __slots__ = ('_name', '__doc__', '_getitem')
+
+    def __init__(self, getitem):
+        self._getitem = getitem
+        self._name = getitem.__name__
+        self.__doc__ = getitem.__doc__
+
+    def __getattr__(self, item):
+        if item in {'__name__', '__qualname__'}:
+            return self._name
+
+        raise AttributeError(item)
+
+    def __mro_entries__(self, bases):
+        raise TypeError(f"Cannot subclass {self!r}")
+
+    def __repr__(self):
+        return f'typing_extensions.{self._name}'
+
+    def __reduce__(self):
+        return self._name
+
+    def __call__(self, *args, **kwds):
+        raise TypeError(f"Cannot instantiate {self!r}")
+
+    def __or__(self, other):
+        return typing.Union[self, other]
+
+    def __ror__(self, other):
+        return typing.Union[other, self]
+
+    def __instancecheck__(self, obj):
+        raise TypeError(f"{self} cannot be used with isinstance()")
+
+    def __subclasscheck__(self, cls):
+        raise TypeError(f"{self} cannot be used with issubclass()")
+
+    @typing._tp_cache
+    def __getitem__(self, parameters):
+        return self._getitem(self, parameters)
+
+
+if hasattr(typing, "LiteralString"):  # 3.11+
+    LiteralString = typing.LiteralString
+else:
+    @_SpecialForm
+    def LiteralString(self, params):
+        """Represents an arbitrary literal string.
+
+        Example::
+
+          from pip._vendor.typing_extensions import LiteralString
+
+          def query(sql: LiteralString) -> ...:
+              ...
+
+          query("SELECT * FROM table")  # ok
+          query(f"SELECT * FROM {input()}")  # not ok
+
+        See PEP 675 for details.
+
+        """
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Self"):  # 3.11+
+    Self = typing.Self
+else:
+    @_SpecialForm
+    def Self(self, params):
+        """Used to spell the type of "self" in classes.
+
+        Example::
+
+          from typing import Self
+
+          class ReturnsSelf:
+              def parse(self, data: bytes) -> Self:
+                  ...
+                  return self
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Never"):  # 3.11+
+    Never = typing.Never
+else:
+    @_SpecialForm
+    def Never(self, params):
+        """The bottom type, a type that has no members.
+
+        This can be used to define a function that should never be
+        called, or a function that never returns::
+
+            from pip._vendor.typing_extensions import Never
+
+            def never_call_me(arg: Never) -> None:
+                pass
+
+            def int_or_str(arg: int | str) -> None:
+                never_call_me(arg)  # type checker error
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        never_call_me(arg)  # ok, arg is of type Never
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, 'Required'):  # 3.11+
+    Required = typing.Required
+    NotRequired = typing.NotRequired
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.10
+    @_ExtensionsSpecialForm
+    def Required(self, parameters):
+        """A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+    @_ExtensionsSpecialForm
+    def NotRequired(self, parameters):
+        """A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _RequiredForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    Required = _RequiredForm(
+        'Required',
+        doc="""A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """)
+    NotRequired = _RequiredForm(
+        'NotRequired',
+        doc="""A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """)
+
+
+if hasattr(typing, 'ReadOnly'):
+    ReadOnly = typing.ReadOnly
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.12
+    @_ExtensionsSpecialForm
+    def ReadOnly(self, parameters):
+        """A special typing construct to mark an item of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this property.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _ReadOnlyForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    ReadOnly = _ReadOnlyForm(
+        'ReadOnly',
+        doc="""A special typing construct to mark a key of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this propery.
+        """)
+
+
+_UNPACK_DOC = """\
+Type unpack operator.
+
+The type unpack operator takes the child types from some container type,
+such as `tuple[int, str]` or a `TypeVarTuple`, and 'pulls them out'. For
+example:
+
+  # For some generic class `Foo`:
+  Foo[Unpack[tuple[int, str]]]  # Equivalent to Foo[int, str]
+
+  Ts = TypeVarTuple('Ts')
+  # Specifies that `Bar` is generic in an arbitrary number of types.
+  # (Think of `Ts` as a tuple of an arbitrary number of individual
+  #  `TypeVar`s, which the `Unpack` is 'pulling out' directly into the
+  #  `Generic[]`.)
+  class Bar(Generic[Unpack[Ts]]): ...
+  Bar[int]  # Valid
+  Bar[int, str]  # Also valid
+
+From Python 3.11, this can also be done using the `*` operator:
+
+    Foo[*tuple[int, str]]
+    class Bar(Generic[*Ts]): ...
+
+The operator can also be used along with a `TypedDict` to annotate
+`**kwargs` in a function signature. For instance:
+
+  class Movie(TypedDict):
+    name: str
+    year: int
+
+  # This function expects two keyword arguments - *name* of type `str` and
+  # *year* of type `int`.
+  def foo(**kwargs: Unpack[Movie]): ...
+
+Note that there is only some runtime checking of this operator. Not
+everything the runtime allows may be accepted by static type checkers.
+
+For more information, see PEP 646 and PEP 692.
+"""
+
+
+if sys.version_info >= (3, 12):  # PEP 692 changed the repr of Unpack[]
+    Unpack = typing.Unpack
+
+    def _is_unpack(obj):
+        return get_origin(obj) is Unpack
+
+elif sys.version_info[:2] >= (3, 9):  # 3.9+
+    class _UnpackSpecialForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, getitem):
+            super().__init__(getitem)
+            self.__doc__ = _UNPACK_DOC
+
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+        @property
+        def __typing_unpacked_tuple_args__(self):
+            assert self.__origin__ is Unpack
+            assert len(self.__args__) == 1
+            arg, = self.__args__
+            if isinstance(arg, (typing._GenericAlias, _types.GenericAlias)):
+                if arg.__origin__ is not tuple:
+                    raise TypeError("Unpack[...] must be used with a tuple type")
+                return arg.__args__
+            return None
+
+    @_UnpackSpecialForm
+    def Unpack(self, parameters):
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return _UnpackAlias(self, (item,))
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+else:  # 3.8
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+    class _UnpackForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return _UnpackAlias(self, (item,))
+
+    Unpack = _UnpackForm('Unpack', doc=_UNPACK_DOC)
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVarTuple
+
+elif hasattr(typing, "TypeVarTuple"):  # 3.11+
+
+    def _unpack_args(*args):
+        newargs = []
+        for arg in args:
+            subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+            if subargs is not None and not (subargs and subargs[-1] is ...):
+                newargs.extend(subargs)
+            else:
+                newargs.append(arg)
+        return newargs
+
+    # Add default parameter - PEP 696
+    class TypeVarTuple(metaclass=_TypeVarLikeMeta):
+        """Type variable tuple."""
+
+        _backported_typevarlike = typing.TypeVarTuple
+
+        def __new__(cls, name, *, default=NoDefault):
+            tvt = typing.TypeVarTuple(name)
+            _set_default(tvt, default)
+            _set_module(tvt)
+
+            def _typevartuple_prepare_subst(alias, args):
+                params = alias.__parameters__
+                typevartuple_index = params.index(tvt)
+                for param in params[typevartuple_index + 1:]:
+                    if isinstance(param, TypeVarTuple):
+                        raise TypeError(
+                            f"More than one TypeVarTuple parameter in {alias}"
+                        )
+
+                alen = len(args)
+                plen = len(params)
+                left = typevartuple_index
+                right = plen - typevartuple_index - 1
+                var_tuple_index = None
+                fillarg = None
+                for k, arg in enumerate(args):
+                    if not isinstance(arg, type):
+                        subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+                        if subargs and len(subargs) == 2 and subargs[-1] is ...:
+                            if var_tuple_index is not None:
+                                raise TypeError(
+                                    "More than one unpacked "
+                                    "arbitrary-length tuple argument"
+                                )
+                            var_tuple_index = k
+                            fillarg = subargs[0]
+                if var_tuple_index is not None:
+                    left = min(left, var_tuple_index)
+                    right = min(right, alen - var_tuple_index - 1)
+                elif left + right > alen:
+                    raise TypeError(f"Too few arguments for {alias};"
+                                    f" actual {alen}, expected at least {plen - 1}")
+                if left == alen - right and tvt.has_default():
+                    replacement = _unpack_args(tvt.__default__)
+                else:
+                    replacement = args[left: alen - right]
+
+                return (
+                    *args[:left],
+                    *([fillarg] * (typevartuple_index - left)),
+                    replacement,
+                    *([fillarg] * (plen - right - left - typevartuple_index - 1)),
+                    *args[alen - right:],
+                )
+
+            tvt.__typing_prepare_subst__ = _typevartuple_prepare_subst
+            return tvt
+
+        def __init_subclass__(self, *args, **kwds):
+            raise TypeError("Cannot subclass special typing classes")
+
+else:  # <=3.10
+    class TypeVarTuple(_DefaultMixin):
+        """Type variable tuple.
+
+        Usage::
+
+            Ts = TypeVarTuple('Ts')
+
+        In the same way that a normal type variable is a stand-in for a single
+        type such as ``int``, a type variable *tuple* is a stand-in for a *tuple*
+        type such as ``Tuple[int, str]``.
+
+        Type variable tuples can be used in ``Generic`` declarations.
+        Consider the following example::
+
+            class Array(Generic[*Ts]): ...
+
+        The ``Ts`` type variable tuple here behaves like ``tuple[T1, T2]``,
+        where ``T1`` and ``T2`` are type variables. To use these type variables
+        as type parameters of ``Array``, we must *unpack* the type variable tuple using
+        the star operator: ``*Ts``. The signature of ``Array`` then behaves
+        as if we had simply written ``class Array(Generic[T1, T2]): ...``.
+        In contrast to ``Generic[T1, T2]``, however, ``Generic[*Shape]`` allows
+        us to parameterise the class with an *arbitrary* number of type parameters.
+
+        Type variable tuples can be used anywhere a normal ``TypeVar`` can.
+        This includes class definitions, as shown above, as well as function
+        signatures and variable annotations::
+
+            class Array(Generic[*Ts]):
+
+                def __init__(self, shape: Tuple[*Ts]):
+                    self._shape: Tuple[*Ts] = shape
+
+                def get_shape(self) -> Tuple[*Ts]:
+                    return self._shape
+
+            shape = (Height(480), Width(640))
+            x: Array[Height, Width] = Array(shape)
+            y = abs(x)  # Inferred type is Array[Height, Width]
+            z = x + x   #        ...    is Array[Height, Width]
+            x.get_shape()  #     ...    is tuple[Height, Width]
+
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        def __iter__(self):
+            yield self.__unpacked__
+
+        def __init__(self, name, *, default=NoDefault):
+            self.__name__ = name
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+            self.__unpacked__ = Unpack[self]
+
+        def __repr__(self):
+            return self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(self, *args, **kwds):
+            if '_root' not in kwds:
+                raise TypeError("Cannot subclass special typing classes")
+
+
+if hasattr(typing, "reveal_type"):  # 3.11+
+    reveal_type = typing.reveal_type
+else:  # <=3.10
+    def reveal_type(obj: T, /) -> T:
+        """Reveal the inferred type of a variable.
+
+        When a static type checker encounters a call to ``reveal_type()``,
+        it will emit the inferred type of the argument::
+
+            x: int = 1
+            reveal_type(x)
+
+        Running a static type checker (e.g., ``mypy``) on this example
+        will produce output similar to 'Revealed type is "builtins.int"'.
+
+        At runtime, the function prints the runtime type of the
+        argument and returns it unchanged.
+
+        """
+        print(f"Runtime type is {type(obj).__name__!r}", file=sys.stderr)
+        return obj
+
+
+if hasattr(typing, "_ASSERT_NEVER_REPR_MAX_LENGTH"):  # 3.11+
+    _ASSERT_NEVER_REPR_MAX_LENGTH = typing._ASSERT_NEVER_REPR_MAX_LENGTH
+else:  # <=3.10
+    _ASSERT_NEVER_REPR_MAX_LENGTH = 100
+
+
+if hasattr(typing, "assert_never"):  # 3.11+
+    assert_never = typing.assert_never
+else:  # <=3.10
+    def assert_never(arg: Never, /) -> Never:
+        """Assert to the type checker that a line of code is unreachable.
+
+        Example::
+
+            def int_or_str(arg: int | str) -> None:
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        assert_never(arg)
+
+        If a type checker finds that a call to assert_never() is
+        reachable, it will emit an error.
+
+        At runtime, this throws an exception when called.
+
+        """
+        value = repr(arg)
+        if len(value) > _ASSERT_NEVER_REPR_MAX_LENGTH:
+            value = value[:_ASSERT_NEVER_REPR_MAX_LENGTH] + '...'
+        raise AssertionError(f"Expected code to be unreachable, but got: {value}")
+
+
+if sys.version_info >= (3, 12):  # 3.12+
+    # dataclass_transform exists in 3.11 but lacks the frozen_default parameter
+    dataclass_transform = typing.dataclass_transform
+else:  # <=3.11
+    def dataclass_transform(
+        *,
+        eq_default: bool = True,
+        order_default: bool = False,
+        kw_only_default: bool = False,
+        frozen_default: bool = False,
+        field_specifiers: typing.Tuple[
+            typing.Union[typing.Type[typing.Any], typing.Callable[..., typing.Any]],
+            ...
+        ] = (),
+        **kwargs: typing.Any,
+    ) -> typing.Callable[[T], T]:
+        """Decorator that marks a function, class, or metaclass as providing
+        dataclass-like behavior.
+
+        Example:
+
+            from pip._vendor.typing_extensions import dataclass_transform
+
+            _T = TypeVar("_T")
+
+            # Used on a decorator function
+            @dataclass_transform()
+            def create_model(cls: type[_T]) -> type[_T]:
+                ...
+                return cls
+
+            @create_model
+            class CustomerModel:
+                id: int
+                name: str
+
+            # Used on a base class
+            @dataclass_transform()
+            class ModelBase: ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+            # Used on a metaclass
+            @dataclass_transform()
+            class ModelMeta(type): ...
+
+            class ModelBase(metaclass=ModelMeta): ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+        Each of the ``CustomerModel`` classes defined in this example will now
+        behave similarly to a dataclass created with the ``@dataclasses.dataclass``
+        decorator. For example, the type checker will synthesize an ``__init__``
+        method.
+
+        The arguments to this decorator can be used to customize this behavior:
+        - ``eq_default`` indicates whether the ``eq`` parameter is assumed to be
+          True or False if it is omitted by the caller.
+        - ``order_default`` indicates whether the ``order`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``kw_only_default`` indicates whether the ``kw_only`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``frozen_default`` indicates whether the ``frozen`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``field_specifiers`` specifies a static list of supported classes
+          or functions that describe fields, similar to ``dataclasses.field()``.
+
+        At runtime, this decorator records its arguments in the
+        ``__dataclass_transform__`` attribute on the decorated object.
+
+        See PEP 681 for details.
+
+        """
+        def decorator(cls_or_fn):
+            cls_or_fn.__dataclass_transform__ = {
+                "eq_default": eq_default,
+                "order_default": order_default,
+                "kw_only_default": kw_only_default,
+                "frozen_default": frozen_default,
+                "field_specifiers": field_specifiers,
+                "kwargs": kwargs,
+            }
+            return cls_or_fn
+        return decorator
+
+
+if hasattr(typing, "override"):  # 3.12+
+    override = typing.override
+else:  # <=3.11
+    _F = typing.TypeVar("_F", bound=typing.Callable[..., typing.Any])
+
+    def override(arg: _F, /) -> _F:
+        """Indicate that a method is intended to override a method in a base class.
+
+        Usage:
+
+            class Base:
+                def method(self) -> None:
+                    pass
+
+            class Child(Base):
+                @override
+                def method(self) -> None:
+                    super().method()
+
+        When this decorator is applied to a method, the type checker will
+        validate that it overrides a method with the same name on a base class.
+        This helps prevent bugs that may occur when a base class is changed
+        without an equivalent change to a child class.
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__override__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+
+        See PEP 698 for details.
+
+        """
+        try:
+            arg.__override__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return arg
+
+
+if hasattr(warnings, "deprecated"):
+    deprecated = warnings.deprecated
+else:
+    _T = typing.TypeVar("_T")
+
+    class deprecated:
+        """Indicate that a class, function or overload is deprecated.
+
+        When this decorator is applied to an object, the type checker
+        will generate a diagnostic on usage of the deprecated object.
+
+        Usage:
+
+            @deprecated("Use B instead")
+            class A:
+                pass
+
+            @deprecated("Use g instead")
+            def f():
+                pass
+
+            @overload
+            @deprecated("int support is deprecated")
+            def g(x: int) -> int: ...
+            @overload
+            def g(x: str) -> int: ...
+
+        The warning specified by *category* will be emitted at runtime
+        on use of deprecated objects. For functions, that happens on calls;
+        for classes, on instantiation and on creation of subclasses.
+        If the *category* is ``None``, no warning is emitted at runtime.
+        The *stacklevel* determines where the
+        warning is emitted. If it is ``1`` (the default), the warning
+        is emitted at the direct caller of the deprecated object; if it
+        is higher, it is emitted further up the stack.
+        Static type checker behavior is not affected by the *category*
+        and *stacklevel* arguments.
+
+        The deprecation message passed to the decorator is saved in the
+        ``__deprecated__`` attribute on the decorated object.
+        If applied to an overload, the decorator
+        must be after the ``@overload`` decorator for the attribute to
+        exist on the overload as returned by ``get_overloads()``.
+
+        See PEP 702 for details.
+
+        """
+        def __init__(
+            self,
+            message: str,
+            /,
+            *,
+            category: typing.Optional[typing.Type[Warning]] = DeprecationWarning,
+            stacklevel: int = 1,
+        ) -> None:
+            if not isinstance(message, str):
+                raise TypeError(
+                    "Expected an object of type str for 'message', not "
+                    f"{type(message).__name__!r}"
+                )
+            self.message = message
+            self.category = category
+            self.stacklevel = stacklevel
+
+        def __call__(self, arg: _T, /) -> _T:
+            # Make sure the inner functions created below don't
+            # retain a reference to self.
+            msg = self.message
+            category = self.category
+            stacklevel = self.stacklevel
+            if category is None:
+                arg.__deprecated__ = msg
+                return arg
+            elif isinstance(arg, type):
+                import functools
+                from types import MethodType
+
+                original_new = arg.__new__
+
+                @functools.wraps(original_new)
+                def __new__(cls, *args, **kwargs):
+                    if cls is arg:
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    if original_new is not object.__new__:
+                        return original_new(cls, *args, **kwargs)
+                    # Mirrors a similar check in object.__new__.
+                    elif cls.__init__ is object.__init__ and (args or kwargs):
+                        raise TypeError(f"{cls.__name__}() takes no arguments")
+                    else:
+                        return original_new(cls)
+
+                arg.__new__ = staticmethod(__new__)
+
+                original_init_subclass = arg.__init_subclass__
+                # We need slightly different behavior if __init_subclass__
+                # is a bound method (likely if it was implemented in Python)
+                if isinstance(original_init_subclass, MethodType):
+                    original_init_subclass = original_init_subclass.__func__
+
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = classmethod(__init_subclass__)
+                # Or otherwise, which likely means it's a builtin such as
+                # object's implementation of __init_subclass__.
+                else:
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = __init_subclass__
+
+                arg.__deprecated__ = __new__.__deprecated__ = msg
+                __init_subclass__.__deprecated__ = msg
+                return arg
+            elif callable(arg):
+                import functools
+
+                @functools.wraps(arg)
+                def wrapper(*args, **kwargs):
+                    warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    return arg(*args, **kwargs)
+
+                arg.__deprecated__ = wrapper.__deprecated__ = msg
+                return wrapper
+            else:
+                raise TypeError(
+                    "@deprecated decorator with non-None category must be applied to "
+                    f"a class or callable, not {arg!r}"
+                )
+
+
+# We have to do some monkey patching to deal with the dual nature of
+# Unpack/TypeVarTuple:
+# - We want Unpack to be a kind of TypeVar so it gets accepted in
+#   Generic[Unpack[Ts]]
+# - We want it to *not* be treated as a TypeVar for the purposes of
+#   counting generic parameters, so that when we subscript a generic,
+#   the runtime doesn't try to substitute the Unpack with the subscripted type.
+if not hasattr(typing, "TypeVarTuple"):
+    def _check_generic(cls, parameters, elen=_marker):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        if elen is _marker:
+            if not hasattr(cls, "__parameters__") or not cls.__parameters__:
+                raise TypeError(f"{cls} is not a generic class")
+            elen = len(cls.__parameters__)
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+                num_tv_tuples = sum(isinstance(p, TypeVarTuple) for p in parameters)
+                if (num_tv_tuples > 0) and (alen >= elen - num_tv_tuples):
+                    return
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            things = "arguments" if sys.version_info >= (3, 10) else "parameters"
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} {things}"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+else:
+    # Python 3.11+
+
+    def _check_generic(cls, parameters, elen):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} arguments"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+
+if not _PEP_696_IMPLEMENTED:
+    typing._check_generic = _check_generic
+
+
+def _has_generic_or_protocol_as_origin() -> bool:
+    try:
+        frame = sys._getframe(2)
+    # - Catch AttributeError: not all Python implementations have sys._getframe()
+    # - Catch ValueError: maybe we're called from an unexpected module
+    #   and the call stack isn't deep enough
+    except (AttributeError, ValueError):
+        return False  # err on the side of leniency
+    else:
+        # If we somehow get invoked from outside typing.py,
+        # also err on the side of leniency
+        if frame.f_globals.get("__name__") != "typing":
+            return False
+        origin = frame.f_locals.get("origin")
+        # Cannot use "in" because origin may be an object with a buggy __eq__ that
+        # throws an error.
+        return origin is typing.Generic or origin is Protocol or origin is typing.Protocol
+
+
+_TYPEVARTUPLE_TYPES = {TypeVarTuple, getattr(typing, "TypeVarTuple", None)}
+
+
+def _is_unpacked_typevartuple(x) -> bool:
+    if get_origin(x) is not Unpack:
+        return False
+    args = get_args(x)
+    return (
+        bool(args)
+        and len(args) == 1
+        and type(args[0]) in _TYPEVARTUPLE_TYPES
+    )
+
+
+# Python 3.11+ _collect_type_vars was renamed to _collect_parameters
+if hasattr(typing, '_collect_type_vars'):
+    def _collect_type_vars(types, typevar_types=None):
+        """Collect all type variable contained in types in order of
+        first appearance (lexicographic order). For example::
+
+            _collect_type_vars((T, List[S, T])) == (T, S)
+        """
+        if typevar_types is None:
+            typevar_types = typing.TypeVar
+        tvars = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with a default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in types:
+            if _is_unpacked_typevartuple(t):
+                type_var_tuple_encountered = True
+            elif isinstance(t, typevar_types) and t not in tvars:
+                if enforce_default_ordering:
+                    has_default = getattr(t, '__default__', NoDefault) is not NoDefault
+                    if has_default:
+                        if type_var_tuple_encountered:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+                        default_encountered = True
+                    elif default_encountered:
+                        raise TypeError(f'Type parameter {t!r} without a default'
+                                        ' follows type parameter with a default')
+
+                tvars.append(t)
+            if _should_collect_from_parameters(t):
+                tvars.extend([t for t in t.__parameters__ if t not in tvars])
+        return tuple(tvars)
+
+    typing._collect_type_vars = _collect_type_vars
+else:
+    def _collect_parameters(args):
+        """Collect all type variables and parameter specifications in args
+        in order of first appearance (lexicographic order).
+
+        For example::
+
+            assert _collect_parameters((T, Callable[P, T])) == (T, P)
+        """
+        parameters = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in args:
+            if isinstance(t, type):
+                # We don't want __parameters__ descriptor of a bare Python class.
+                pass
+            elif isinstance(t, tuple):
+                # `t` might be a tuple, when `ParamSpec` is substituted with
+                # `[T, int]`, or `[int, *Ts]`, etc.
+                for x in t:
+                    for collected in _collect_parameters([x]):
+                        if collected not in parameters:
+                            parameters.append(collected)
+            elif hasattr(t, '__typing_subst__'):
+                if t not in parameters:
+                    if enforce_default_ordering:
+                        has_default = (
+                            getattr(t, '__default__', NoDefault) is not NoDefault
+                        )
+
+                        if type_var_tuple_encountered and has_default:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+
+                        if has_default:
+                            default_encountered = True
+                        elif default_encountered:
+                            raise TypeError(f'Type parameter {t!r} without a default'
+                                            ' follows type parameter with a default')
+
+                    parameters.append(t)
+            else:
+                if _is_unpacked_typevartuple(t):
+                    type_var_tuple_encountered = True
+                for x in getattr(t, '__parameters__', ()):
+                    if x not in parameters:
+                        parameters.append(x)
+
+        return tuple(parameters)
+
+    if not _PEP_696_IMPLEMENTED:
+        typing._collect_parameters = _collect_parameters
+
+# Backport typing.NamedTuple as it exists in Python 3.13.
+# In 3.11, the ability to define generic `NamedTuple`s was supported.
+# This was explicitly disallowed in 3.9-3.10, and only half-worked in <=3.8.
+# On 3.12, we added __orig_bases__ to call-based NamedTuples
+# On 3.13, we deprecated kwargs-based NamedTuples
+if sys.version_info >= (3, 13):
+    NamedTuple = typing.NamedTuple
+else:
+    def _make_nmtuple(name, types, module, defaults=()):
+        fields = [n for n, t in types]
+        annotations = {n: typing._type_check(t, f"field {n} annotation must be a type")
+                       for n, t in types}
+        nm_tpl = collections.namedtuple(name, fields,
+                                        defaults=defaults, module=module)
+        nm_tpl.__annotations__ = nm_tpl.__new__.__annotations__ = annotations
+        # The `_field_types` attribute was removed in 3.9;
+        # in earlier versions, it is the same as the `__annotations__` attribute
+        if sys.version_info < (3, 9):
+            nm_tpl._field_types = annotations
+        return nm_tpl
+
+    _prohibited_namedtuple_fields = typing._prohibited
+    _special_namedtuple_fields = frozenset({'__module__', '__name__', '__annotations__'})
+
+    class _NamedTupleMeta(type):
+        def __new__(cls, typename, bases, ns):
+            assert _NamedTuple in bases
+            for base in bases:
+                if base is not _NamedTuple and base is not typing.Generic:
+                    raise TypeError(
+                        'can only inherit from a NamedTuple type and Generic')
+            bases = tuple(tuple if base is _NamedTuple else base for base in bases)
+            if "__annotations__" in ns:
+                types = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                types = ns["__annotate__"](1)
+            else:
+                types = {}
+            default_names = []
+            for field_name in types:
+                if field_name in ns:
+                    default_names.append(field_name)
+                elif default_names:
+                    raise TypeError(f"Non-default namedtuple field {field_name} "
+                                    f"cannot follow default field"
+                                    f"{'s' if len(default_names) > 1 else ''} "
+                                    f"{', '.join(default_names)}")
+            nm_tpl = _make_nmtuple(
+                typename, types.items(),
+                defaults=[ns[n] for n in default_names],
+                module=ns['__module__']
+            )
+            nm_tpl.__bases__ = bases
+            if typing.Generic in bases:
+                if hasattr(typing, '_generic_class_getitem'):  # 3.12+
+                    nm_tpl.__class_getitem__ = classmethod(typing._generic_class_getitem)
+                else:
+                    class_getitem = typing.Generic.__class_getitem__.__func__
+                    nm_tpl.__class_getitem__ = classmethod(class_getitem)
+            # update from user namespace without overriding special namedtuple attributes
+            for key, val in ns.items():
+                if key in _prohibited_namedtuple_fields:
+                    raise AttributeError("Cannot overwrite NamedTuple attribute " + key)
+                elif key not in _special_namedtuple_fields:
+                    if key not in nm_tpl._fields:
+                        setattr(nm_tpl, key, ns[key])
+                    try:
+                        set_name = type(val).__set_name__
+                    except AttributeError:
+                        pass
+                    else:
+                        try:
+                            set_name(val, nm_tpl, key)
+                        except BaseException as e:
+                            msg = (
+                                f"Error calling __set_name__ on {type(val).__name__!r} "
+                                f"instance {key!r} in {typename!r}"
+                            )
+                            # BaseException.add_note() existed on py311,
+                            # but the __set_name__ machinery didn't start
+                            # using add_note() until py312.
+                            # Making sure exceptions are raised in the same way
+                            # as in "normal" classes seems most important here.
+                            if sys.version_info >= (3, 12):
+                                e.add_note(msg)
+                                raise
+                            else:
+                                raise RuntimeError(msg) from e
+
+            if typing.Generic in bases:
+                nm_tpl.__init_subclass__()
+            return nm_tpl
+
+    _NamedTuple = type.__new__(_NamedTupleMeta, 'NamedTuple', (), {})
+
+    def _namedtuple_mro_entries(bases):
+        assert NamedTuple in bases
+        return (_NamedTuple,)
+
+    @_ensure_subclassable(_namedtuple_mro_entries)
+    def NamedTuple(typename, fields=_marker, /, **kwargs):
+        """Typed version of namedtuple.
+
+        Usage::
+
+            class Employee(NamedTuple):
+                name: str
+                id: int
+
+        This is equivalent to::
+
+            Employee = collections.namedtuple('Employee', ['name', 'id'])
+
+        The resulting class has an extra __annotations__ attribute, giving a
+        dict that maps field names to types.  (The field names are also in
+        the _fields attribute, which is part of the namedtuple API.)
+        An alternative equivalent functional syntax is also accepted::
+
+            Employee = NamedTuple('Employee', [('name', str), ('id', int)])
+        """
+        if fields is _marker:
+            if kwargs:
+                deprecated_thing = "Creating NamedTuple classes using keyword arguments"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "Use the class-based or functional syntax instead."
+                )
+            else:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif fields is None:
+            if kwargs:
+                raise TypeError(
+                    "Cannot pass `None` as the 'fields' parameter "
+                    "and also specify fields using keyword arguments"
+                )
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif kwargs:
+            raise TypeError("Either list of fields or keywords"
+                            " can be provided to NamedTuple, not both")
+        if fields is _marker or fields is None:
+            warnings.warn(
+                deprecation_msg.format(name=deprecated_thing, remove="3.15"),
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            fields = kwargs.items()
+        nt = _make_nmtuple(typename, fields, module=_caller())
+        nt.__orig_bases__ = (NamedTuple,)
+        return nt
+
+
+if hasattr(collections.abc, "Buffer"):
+    Buffer = collections.abc.Buffer
+else:
+    class Buffer(abc.ABC):  # noqa: B024
+        """Base class for classes that implement the buffer protocol.
+
+        The buffer protocol allows Python objects to expose a low-level
+        memory buffer interface. Before Python 3.12, it is not possible
+        to implement the buffer protocol in pure Python code, or even
+        to check whether a class implements the buffer protocol. In
+        Python 3.12 and higher, the ``__buffer__`` method allows access
+        to the buffer protocol from Python code, and the
+        ``collections.abc.Buffer`` ABC allows checking whether a class
+        implements the buffer protocol.
+
+        To indicate support for the buffer protocol in earlier versions,
+        inherit from this ABC, either in a stub file or at runtime,
+        or use ABC registration. This ABC provides no methods, because
+        there is no Python-accessible methods shared by pre-3.12 buffer
+        classes. It is useful primarily for static checks.
+
+        """
+
+    # As a courtesy, register the most common stdlib buffer classes.
+    Buffer.register(memoryview)
+    Buffer.register(bytearray)
+    Buffer.register(bytes)
+
+
+# Backport of types.get_original_bases, available on 3.12+ in CPython
+if hasattr(_types, "get_original_bases"):
+    get_original_bases = _types.get_original_bases
+else:
+    def get_original_bases(cls, /):
+        """Return the class's "original" bases prior to modification by `__mro_entries__`.
+
+        Examples::
+
+            from typing import TypeVar, Generic
+            from pip._vendor.typing_extensions import NamedTuple, TypedDict
+
+            T = TypeVar("T")
+            class Foo(Generic[T]): ...
+            class Bar(Foo[int], float): ...
+            class Baz(list[str]): ...
+            Eggs = NamedTuple("Eggs", [("a", int), ("b", str)])
+            Spam = TypedDict("Spam", {"a": int, "b": str})
+
+            assert get_original_bases(Bar) == (Foo[int], float)
+            assert get_original_bases(Baz) == (list[str],)
+            assert get_original_bases(Eggs) == (NamedTuple,)
+            assert get_original_bases(Spam) == (TypedDict,)
+            assert get_original_bases(int) == (object,)
+        """
+        try:
+            return cls.__dict__.get("__orig_bases__", cls.__bases__)
+        except AttributeError:
+            raise TypeError(
+                f'Expected an instance of type, not {type(cls).__name__!r}'
+            ) from None
+
+
+# NewType is a class on Python 3.10+, making it pickleable
+# The error message for subclassing instances of NewType was improved on 3.11+
+if sys.version_info >= (3, 11):
+    NewType = typing.NewType
+else:
+    class NewType:
+        """NewType creates simple unique types with almost zero
+        runtime overhead. NewType(name, tp) is considered a subtype of tp
+        by static type checkers. At runtime, NewType(name, tp) returns
+        a dummy callable that simply returns its argument. Usage::
+            UserId = NewType('UserId', int)
+            def name_by_id(user_id: UserId) -> str:
+                ...
+            UserId('user')          # Fails type check
+            name_by_id(42)          # Fails type check
+            name_by_id(UserId(42))  # OK
+            num = UserId(5) + 1     # type: int
+        """
+
+        def __call__(self, obj, /):
+            return obj
+
+        def __init__(self, name, tp):
+            self.__qualname__ = name
+            if '.' in name:
+                name = name.rpartition('.')[-1]
+            self.__name__ = name
+            self.__supertype__ = tp
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __mro_entries__(self, bases):
+            # We defined __mro_entries__ to get a better error message
+            # if a user attempts to subclass a NewType instance. bpo-46170
+            supercls_name = self.__name__
+
+            class Dummy:
+                def __init_subclass__(cls):
+                    subcls_name = cls.__name__
+                    raise TypeError(
+                        f"Cannot subclass an instance of NewType. "
+                        f"Perhaps you were looking for: "
+                        f"`{subcls_name} = NewType({subcls_name!r}, {supercls_name})`"
+                    )
+
+            return (Dummy,)
+
+        def __repr__(self):
+            return f'{self.__module__}.{self.__qualname__}'
+
+        def __reduce__(self):
+            return self.__qualname__
+
+        if sys.version_info >= (3, 10):
+            # PEP 604 methods
+            # It doesn't make sense to have these methods on Python <3.10
+
+            def __or__(self, other):
+                return typing.Union[self, other]
+
+            def __ror__(self, other):
+                return typing.Union[other, self]
+
+
+if hasattr(typing, "TypeAliasType"):
+    TypeAliasType = typing.TypeAliasType
+else:
+    def _is_unionable(obj):
+        """Corresponds to is_unionable() in unionobject.c in CPython."""
+        return obj is None or isinstance(obj, (
+            type,
+            _types.GenericAlias,
+            _types.UnionType,
+            TypeAliasType,
+        ))
+
+    class TypeAliasType:
+        """Create named, parameterized type aliases.
+
+        This provides a backport of the new `type` statement in Python 3.12:
+
+            type ListOrSet[T] = list[T] | set[T]
+
+        is equivalent to:
+
+            T = TypeVar("T")
+            ListOrSet = TypeAliasType("ListOrSet", list[T] | set[T], type_params=(T,))
+
+        The name ListOrSet can then be used as an alias for the type it refers to.
+
+        The type_params argument should contain all the type parameters used
+        in the value of the type alias. If the alias is not generic, this
+        argument is omitted.
+
+        Static type checkers should only support type aliases declared using
+        TypeAliasType that follow these rules:
+
+        - The first argument (the name) must be a string literal.
+        - The TypeAliasType instance must be immediately assigned to a variable
+          of the same name. (For example, 'X = TypeAliasType("Y", int)' is invalid,
+          as is 'X, Y = TypeAliasType("X", int), TypeAliasType("Y", int)').
+
+        """
+
+        def __init__(self, name: str, value, *, type_params=()):
+            if not isinstance(name, str):
+                raise TypeError("TypeAliasType name must be a string")
+            self.__value__ = value
+            self.__type_params__ = type_params
+
+            parameters = []
+            for type_param in type_params:
+                if isinstance(type_param, TypeVarTuple):
+                    parameters.extend(type_param)
+                else:
+                    parameters.append(type_param)
+            self.__parameters__ = tuple(parameters)
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+            # Setting this attribute closes the TypeAliasType from further modification
+            self.__name__ = name
+
+        def __setattr__(self, name: str, value: object, /) -> None:
+            if hasattr(self, "__name__"):
+                self._raise_attribute_error(name)
+            super().__setattr__(name, value)
+
+        def __delattr__(self, name: str, /) -> Never:
+            self._raise_attribute_error(name)
+
+        def _raise_attribute_error(self, name: str) -> Never:
+            # Match the Python 3.12 error messages exactly
+            if name == "__name__":
+                raise AttributeError("readonly attribute")
+            elif name in {"__value__", "__type_params__", "__parameters__", "__module__"}:
+                raise AttributeError(
+                    f"attribute '{name}' of 'typing.TypeAliasType' objects "
+                    "is not writable"
+                )
+            else:
+                raise AttributeError(
+                    f"'typing.TypeAliasType' object has no attribute '{name}'"
+                )
+
+        def __repr__(self) -> str:
+            return self.__name__
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+            parameters = [
+                typing._type_check(
+                    item, f'Subscripting {self.__name__} requires a type.'
+                )
+                for item in parameters
+            ]
+            return typing._GenericAlias(self, tuple(parameters))
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                "type 'typing_extensions.TypeAliasType' is not an acceptable base type"
+            )
+
+        # The presence of this method convinces typing._type_check
+        # that TypeAliasTypes are types.
+        def __call__(self):
+            raise TypeError("Type alias is not callable")
+
+        if sys.version_info >= (3, 10):
+            def __or__(self, right):
+                # For forward compatibility with 3.12, reject Unions
+                # that are not accepted by the built-in Union.
+                if not _is_unionable(right):
+                    return NotImplemented
+                return typing.Union[self, right]
+
+            def __ror__(self, left):
+                if not _is_unionable(left):
+                    return NotImplemented
+                return typing.Union[left, self]
+
+
+if hasattr(typing, "is_protocol"):
+    is_protocol = typing.is_protocol
+    get_protocol_members = typing.get_protocol_members
+else:
+    def is_protocol(tp: type, /) -> bool:
+        """Return True if the given type is a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, is_protocol
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> is_protocol(P)
+            True
+            >>> is_protocol(int)
+            False
+        """
+        return (
+            isinstance(tp, type)
+            and getattr(tp, '_is_protocol', False)
+            and tp is not Protocol
+            and tp is not typing.Protocol
+        )
+
+    def get_protocol_members(tp: type, /) -> typing.FrozenSet[str]:
+        """Return the set of members defined in a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, get_protocol_members
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> get_protocol_members(P)
+            frozenset({'a', 'b'})
+
+        Raise a TypeError for arguments that are not Protocols.
+        """
+        if not is_protocol(tp):
+            raise TypeError(f'{tp!r} is not a Protocol')
+        if hasattr(tp, '__protocol_attrs__'):
+            return frozenset(tp.__protocol_attrs__)
+        return frozenset(_get_protocol_attrs(tp))
+
+
+if hasattr(typing, "Doc"):
+    Doc = typing.Doc
+else:
+    class Doc:
+        """Define the documentation of a type annotation using ``Annotated``, to be
+         used in class attributes, function and method parameters, return values,
+         and variables.
+
+        The value should be a positional-only string literal to allow static tools
+        like editors and documentation generators to use it.
+
+        This complements docstrings.
+
+        The string value passed is available in the attribute ``documentation``.
+
+        Example::
+
+            >>> from typing_extensions import Annotated, Doc
+            >>> def hi(to: Annotated[str, Doc("Who to say hi to")]) -> None: ...
+        """
+        def __init__(self, documentation: str, /) -> None:
+            self.documentation = documentation
+
+        def __repr__(self) -> str:
+            return f"Doc({self.documentation!r})"
+
+        def __hash__(self) -> int:
+            return hash(self.documentation)
+
+        def __eq__(self, other: object) -> bool:
+            if not isinstance(other, Doc):
+                return NotImplemented
+            return self.documentation == other.documentation
+
+
+_CapsuleType = getattr(_types, "CapsuleType", None)
+
+if _CapsuleType is None:
+    try:
+        import _socket
+    except ImportError:
+        pass
+    else:
+        _CAPI = getattr(_socket, "CAPI", None)
+        if _CAPI is not None:
+            _CapsuleType = type(_CAPI)
+
+if _CapsuleType is not None:
+    CapsuleType = _CapsuleType
+    __all__.append("CapsuleType")
+
+
+# Aliases for items that have always been in typing.
+# Explicitly assign these (rather than using `from typing import *` at the top),
+# so that we get a CI error if one of these is deleted from typing.py
+# in a future version of Python
+AbstractSet = typing.AbstractSet
+AnyStr = typing.AnyStr
+BinaryIO = typing.BinaryIO
+Callable = typing.Callable
+Collection = typing.Collection
+Container = typing.Container
+Dict = typing.Dict
+ForwardRef = typing.ForwardRef
+FrozenSet = typing.FrozenSet
+Generic = typing.Generic
+Hashable = typing.Hashable
+IO = typing.IO
+ItemsView = typing.ItemsView
+Iterable = typing.Iterable
+Iterator = typing.Iterator
+KeysView = typing.KeysView
+List = typing.List
+Mapping = typing.Mapping
+MappingView = typing.MappingView
+Match = typing.Match
+MutableMapping = typing.MutableMapping
+MutableSequence = typing.MutableSequence
+MutableSet = typing.MutableSet
+Optional = typing.Optional
+Pattern = typing.Pattern
+Reversible = typing.Reversible
+Sequence = typing.Sequence
+Set = typing.Set
+Sized = typing.Sized
+TextIO = typing.TextIO
+Tuple = typing.Tuple
+Union = typing.Union
+ValuesView = typing.ValuesView
+cast = typing.cast
+no_type_check = typing.no_type_check
+no_type_check_decorator = typing.no_type_check_decorator
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__init__.py
new file mode 100644
index 00000000..c6fa3821
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__init__.py
@@ -0,0 +1,102 @@
+"""
+Python HTTP library with thread-safe connection pooling, file post support, user friendly, and more
+"""
+from __future__ import absolute_import
+
+# Set default logging handler to avoid "No handler found" warnings.
+import logging
+import warnings
+from logging import NullHandler
+
+from . import exceptions
+from ._version import __version__
+from .connectionpool import HTTPConnectionPool, HTTPSConnectionPool, connection_from_url
+from .filepost import encode_multipart_formdata
+from .poolmanager import PoolManager, ProxyManager, proxy_from_url
+from .response import HTTPResponse
+from .util.request import make_headers
+from .util.retry import Retry
+from .util.timeout import Timeout
+from .util.url import get_host
+
+# === NOTE TO REPACKAGERS AND VENDORS ===
+# Please delete this block, this logic is only
+# for urllib3 being distributed via PyPI.
+# See: https://github.com/urllib3/urllib3/issues/2680
+try:
+    import urllib3_secure_extra  # type: ignore # noqa: F401
+except ImportError:
+    pass
+else:
+    warnings.warn(
+        "'urllib3[secure]' extra is deprecated and will be removed "
+        "in a future release of urllib3 2.x. Read more in this issue: "
+        "https://github.com/urllib3/urllib3/issues/2680",
+        category=DeprecationWarning,
+        stacklevel=2,
+    )
+
+__author__ = "Andrey Petrov (andrey.petrov@shazow.net)"
+__license__ = "MIT"
+__version__ = __version__
+
+__all__ = (
+    "HTTPConnectionPool",
+    "HTTPSConnectionPool",
+    "PoolManager",
+    "ProxyManager",
+    "HTTPResponse",
+    "Retry",
+    "Timeout",
+    "add_stderr_logger",
+    "connection_from_url",
+    "disable_warnings",
+    "encode_multipart_formdata",
+    "get_host",
+    "make_headers",
+    "proxy_from_url",
+)
+
+logging.getLogger(__name__).addHandler(NullHandler())
+
+
+def add_stderr_logger(level=logging.DEBUG):
+    """
+    Helper for quickly adding a StreamHandler to the logger. Useful for
+    debugging.
+
+    Returns the handler after adding it.
+    """
+    # This method needs to be in this __init__.py to get the __name__ correct
+    # even if urllib3 is vendored within another package.
+    logger = logging.getLogger(__name__)
+    handler = logging.StreamHandler()
+    handler.setFormatter(logging.Formatter("%(asctime)s %(levelname)s %(message)s"))
+    logger.addHandler(handler)
+    logger.setLevel(level)
+    logger.debug("Added a stderr logging handler to logger: %s", __name__)
+    return handler
+
+
+# ... Clean up.
+del NullHandler
+
+
+# All warning filters *must* be appended unless you're really certain that they
+# shouldn't be: otherwise, it's very hard for users to use most Python
+# mechanisms to silence them.
+# SecurityWarning's always go off by default.
+warnings.simplefilter("always", exceptions.SecurityWarning, append=True)
+# SubjectAltNameWarning's should go off once per host
+warnings.simplefilter("default", exceptions.SubjectAltNameWarning, append=True)
+# InsecurePlatformWarning's don't vary between requests, so we keep it default.
+warnings.simplefilter("default", exceptions.InsecurePlatformWarning, append=True)
+# SNIMissingWarnings should go off only once.
+warnings.simplefilter("default", exceptions.SNIMissingWarning, append=True)
+
+
+def disable_warnings(category=exceptions.HTTPWarning):
+    """
+    Helper for quickly disabling all urllib3 warnings.
+    """
+    warnings.simplefilter("ignore", category)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..45abccc3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_collections.cpython-312.pyc
new file mode 100644
index 00000000..ccd8c738
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..f50cc248
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..9f281f5e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connectionpool.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connectionpool.cpython-312.pyc
new file mode 100644
index 00000000..8dfa01f7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/connectionpool.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..eac2ebf6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/fields.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/fields.cpython-312.pyc
new file mode 100644
index 00000000..b6fb550f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/fields.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/filepost.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/filepost.cpython-312.pyc
new file mode 100644
index 00000000..9b15bccd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/filepost.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/poolmanager.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/poolmanager.cpython-312.pyc
new file mode 100644
index 00000000..819470d4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/poolmanager.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..c003be6a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..387b5358
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_collections.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_collections.py
new file mode 100644
index 00000000..bceb8451
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_collections.py
@@ -0,0 +1,355 @@
+from __future__ import absolute_import
+
+try:
+    from collections.abc import Mapping, MutableMapping
+except ImportError:
+    from collections import Mapping, MutableMapping
+try:
+    from threading import RLock
+except ImportError:  # Platform-specific: No threads available
+
+    class RLock:
+        def __enter__(self):
+            pass
+
+        def __exit__(self, exc_type, exc_value, traceback):
+            pass
+
+
+from collections import OrderedDict
+
+from .exceptions import InvalidHeader
+from .packages import six
+from .packages.six import iterkeys, itervalues
+
+__all__ = ["RecentlyUsedContainer", "HTTPHeaderDict"]
+
+
+_Null = object()
+
+
+class RecentlyUsedContainer(MutableMapping):
+    """
+    Provides a thread-safe dict-like container which maintains up to
+    ``maxsize`` keys while throwing away the least-recently-used keys beyond
+    ``maxsize``.
+
+    :param maxsize:
+        Maximum number of recent elements to retain.
+
+    :param dispose_func:
+        Every time an item is evicted from the container,
+        ``dispose_func(value)`` is called.  Callback which will get called
+    """
+
+    ContainerCls = OrderedDict
+
+    def __init__(self, maxsize=10, dispose_func=None):
+        self._maxsize = maxsize
+        self.dispose_func = dispose_func
+
+        self._container = self.ContainerCls()
+        self.lock = RLock()
+
+    def __getitem__(self, key):
+        # Re-insert the item, moving it to the end of the eviction line.
+        with self.lock:
+            item = self._container.pop(key)
+            self._container[key] = item
+            return item
+
+    def __setitem__(self, key, value):
+        evicted_value = _Null
+        with self.lock:
+            # Possibly evict the existing value of 'key'
+            evicted_value = self._container.get(key, _Null)
+            self._container[key] = value
+
+            # If we didn't evict an existing value, we might have to evict the
+            # least recently used item from the beginning of the container.
+            if len(self._container) > self._maxsize:
+                _key, evicted_value = self._container.popitem(last=False)
+
+        if self.dispose_func and evicted_value is not _Null:
+            self.dispose_func(evicted_value)
+
+    def __delitem__(self, key):
+        with self.lock:
+            value = self._container.pop(key)
+
+        if self.dispose_func:
+            self.dispose_func(value)
+
+    def __len__(self):
+        with self.lock:
+            return len(self._container)
+
+    def __iter__(self):
+        raise NotImplementedError(
+            "Iteration over this class is unlikely to be threadsafe."
+        )
+
+    def clear(self):
+        with self.lock:
+            # Copy pointers to all values, then wipe the mapping
+            values = list(itervalues(self._container))
+            self._container.clear()
+
+        if self.dispose_func:
+            for value in values:
+                self.dispose_func(value)
+
+    def keys(self):
+        with self.lock:
+            return list(iterkeys(self._container))
+
+
+class HTTPHeaderDict(MutableMapping):
+    """
+    :param headers:
+        An iterable of field-value pairs. Must not contain multiple field names
+        when compared case-insensitively.
+
+    :param kwargs:
+        Additional field-value pairs to pass in to ``dict.update``.
+
+    A ``dict`` like container for storing HTTP Headers.
+
+    Field names are stored and compared case-insensitively in compliance with
+    RFC 7230. Iteration provides the first case-sensitive key seen for each
+    case-insensitive pair.
+
+    Using ``__setitem__`` syntax overwrites fields that compare equal
+    case-insensitively in order to maintain ``dict``'s api. For fields that
+    compare equal, instead create a new ``HTTPHeaderDict`` and use ``.add``
+    in a loop.
+
+    If multiple fields that are equal case-insensitively are passed to the
+    constructor or ``.update``, the behavior is undefined and some will be
+    lost.
+
+    >>> headers = HTTPHeaderDict()
+    >>> headers.add('Set-Cookie', 'foo=bar')
+    >>> headers.add('set-cookie', 'baz=quxx')
+    >>> headers['content-length'] = '7'
+    >>> headers['SET-cookie']
+    'foo=bar, baz=quxx'
+    >>> headers['Content-Length']
+    '7'
+    """
+
+    def __init__(self, headers=None, **kwargs):
+        super(HTTPHeaderDict, self).__init__()
+        self._container = OrderedDict()
+        if headers is not None:
+            if isinstance(headers, HTTPHeaderDict):
+                self._copy_from(headers)
+            else:
+                self.extend(headers)
+        if kwargs:
+            self.extend(kwargs)
+
+    def __setitem__(self, key, val):
+        self._container[key.lower()] = [key, val]
+        return self._container[key.lower()]
+
+    def __getitem__(self, key):
+        val = self._container[key.lower()]
+        return ", ".join(val[1:])
+
+    def __delitem__(self, key):
+        del self._container[key.lower()]
+
+    def __contains__(self, key):
+        return key.lower() in self._container
+
+    def __eq__(self, other):
+        if not isinstance(other, Mapping) and not hasattr(other, "keys"):
+            return False
+        if not isinstance(other, type(self)):
+            other = type(self)(other)
+        return dict((k.lower(), v) for k, v in self.itermerged()) == dict(
+            (k.lower(), v) for k, v in other.itermerged()
+        )
+
+    def __ne__(self, other):
+        return not self.__eq__(other)
+
+    if six.PY2:  # Python 2
+        iterkeys = MutableMapping.iterkeys
+        itervalues = MutableMapping.itervalues
+
+    __marker = object()
+
+    def __len__(self):
+        return len(self._container)
+
+    def __iter__(self):
+        # Only provide the originally cased names
+        for vals in self._container.values():
+            yield vals[0]
+
+    def pop(self, key, default=__marker):
+        """D.pop(k[,d]) -> v, remove specified key and return the corresponding value.
+        If key is not found, d is returned if given, otherwise KeyError is raised.
+        """
+        # Using the MutableMapping function directly fails due to the private marker.
+        # Using ordinary dict.pop would expose the internal structures.
+        # So let's reinvent the wheel.
+        try:
+            value = self[key]
+        except KeyError:
+            if default is self.__marker:
+                raise
+            return default
+        else:
+            del self[key]
+            return value
+
+    def discard(self, key):
+        try:
+            del self[key]
+        except KeyError:
+            pass
+
+    def add(self, key, val):
+        """Adds a (name, value) pair, doesn't overwrite the value if it already
+        exists.
+
+        >>> headers = HTTPHeaderDict(foo='bar')
+        >>> headers.add('Foo', 'baz')
+        >>> headers['foo']
+        'bar, baz'
+        """
+        key_lower = key.lower()
+        new_vals = [key, val]
+        # Keep the common case aka no item present as fast as possible
+        vals = self._container.setdefault(key_lower, new_vals)
+        if new_vals is not vals:
+            vals.append(val)
+
+    def extend(self, *args, **kwargs):
+        """Generic import function for any type of header-like object.
+        Adapted version of MutableMapping.update in order to insert items
+        with self.add instead of self.__setitem__
+        """
+        if len(args) > 1:
+            raise TypeError(
+                "extend() takes at most 1 positional "
+                "arguments ({0} given)".format(len(args))
+            )
+        other = args[0] if len(args) >= 1 else ()
+
+        if isinstance(other, HTTPHeaderDict):
+            for key, val in other.iteritems():
+                self.add(key, val)
+        elif isinstance(other, Mapping):
+            for key in other:
+                self.add(key, other[key])
+        elif hasattr(other, "keys"):
+            for key in other.keys():
+                self.add(key, other[key])
+        else:
+            for key, value in other:
+                self.add(key, value)
+
+        for key, value in kwargs.items():
+            self.add(key, value)
+
+    def getlist(self, key, default=__marker):
+        """Returns a list of all the values for the named field. Returns an
+        empty list if the key doesn't exist."""
+        try:
+            vals = self._container[key.lower()]
+        except KeyError:
+            if default is self.__marker:
+                return []
+            return default
+        else:
+            return vals[1:]
+
+    def _prepare_for_method_change(self):
+        """
+        Remove content-specific header fields before changing the request
+        method to GET or HEAD according to RFC 9110, Section 15.4.
+        """
+        content_specific_headers = [
+            "Content-Encoding",
+            "Content-Language",
+            "Content-Location",
+            "Content-Type",
+            "Content-Length",
+            "Digest",
+            "Last-Modified",
+        ]
+        for header in content_specific_headers:
+            self.discard(header)
+        return self
+
+    # Backwards compatibility for httplib
+    getheaders = getlist
+    getallmatchingheaders = getlist
+    iget = getlist
+
+    # Backwards compatibility for http.cookiejar
+    get_all = getlist
+
+    def __repr__(self):
+        return "%s(%s)" % (type(self).__name__, dict(self.itermerged()))
+
+    def _copy_from(self, other):
+        for key in other:
+            val = other.getlist(key)
+            if isinstance(val, list):
+                # Don't need to convert tuples
+                val = list(val)
+            self._container[key.lower()] = [key] + val
+
+    def copy(self):
+        clone = type(self)()
+        clone._copy_from(self)
+        return clone
+
+    def iteritems(self):
+        """Iterate over all header lines, including duplicate ones."""
+        for key in self:
+            vals = self._container[key.lower()]
+            for val in vals[1:]:
+                yield vals[0], val
+
+    def itermerged(self):
+        """Iterate over all headers, merging duplicate ones together."""
+        for key in self:
+            val = self._container[key.lower()]
+            yield val[0], ", ".join(val[1:])
+
+    def items(self):
+        return list(self.iteritems())
+
+    @classmethod
+    def from_httplib(cls, message):  # Python 2
+        """Read headers from a Python 2 httplib message object."""
+        # python2.7 does not expose a proper API for exporting multiheaders
+        # efficiently. This function re-reads raw lines from the message
+        # object and extracts the multiheaders properly.
+        obs_fold_continued_leaders = (" ", "\t")
+        headers = []
+
+        for line in message.headers:
+            if line.startswith(obs_fold_continued_leaders):
+                if not headers:
+                    # We received a header line that starts with OWS as described
+                    # in RFC-7230 S3.2.4. This indicates a multiline header, but
+                    # there exists no previous header to which we can attach it.
+                    raise InvalidHeader(
+                        "Header continuation with no previous header: %s" % line
+                    )
+                else:
+                    key, value = headers[-1]
+                    headers[-1] = (key, value + " " + line.strip())
+                    continue
+
+            key, value = line.split(":", 1)
+            headers.append((key, value.strip()))
+
+        return cls(headers)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_version.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_version.py
new file mode 100644
index 00000000..d49df2a0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/_version.py
@@ -0,0 +1,2 @@
+# This file is protected via CODEOWNERS
+__version__ = "1.26.20"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connection.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connection.py
new file mode 100644
index 00000000..de35b63d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connection.py
@@ -0,0 +1,572 @@
+from __future__ import absolute_import
+
+import datetime
+import logging
+import os
+import re
+import socket
+import warnings
+from socket import error as SocketError
+from socket import timeout as SocketTimeout
+
+from .packages import six
+from .packages.six.moves.http_client import HTTPConnection as _HTTPConnection
+from .packages.six.moves.http_client import HTTPException  # noqa: F401
+from .util.proxy import create_proxy_ssl_context
+
+try:  # Compiled with SSL?
+    import ssl
+
+    BaseSSLError = ssl.SSLError
+except (ImportError, AttributeError):  # Platform-specific: No SSL.
+    ssl = None
+
+    class BaseSSLError(BaseException):
+        pass
+
+
+try:
+    # Python 3: not a no-op, we're adding this to the namespace so it can be imported.
+    ConnectionError = ConnectionError
+except NameError:
+    # Python 2
+    class ConnectionError(Exception):
+        pass
+
+
+try:  # Python 3:
+    # Not a no-op, we're adding this to the namespace so it can be imported.
+    BrokenPipeError = BrokenPipeError
+except NameError:  # Python 2:
+
+    class BrokenPipeError(Exception):
+        pass
+
+
+from ._collections import HTTPHeaderDict  # noqa (historical, removed in v2)
+from ._version import __version__
+from .exceptions import (
+    ConnectTimeoutError,
+    NewConnectionError,
+    SubjectAltNameWarning,
+    SystemTimeWarning,
+)
+from .util import SKIP_HEADER, SKIPPABLE_HEADERS, connection
+from .util.ssl_ import (
+    assert_fingerprint,
+    create_urllib3_context,
+    is_ipaddress,
+    resolve_cert_reqs,
+    resolve_ssl_version,
+    ssl_wrap_socket,
+)
+from .util.ssl_match_hostname import CertificateError, match_hostname
+
+log = logging.getLogger(__name__)
+
+port_by_scheme = {"http": 80, "https": 443}
+
+# When it comes time to update this value as a part of regular maintenance
+# (ie test_recent_date is failing) update it to ~6 months before the current date.
+RECENT_DATE = datetime.date(2024, 1, 1)
+
+_CONTAINS_CONTROL_CHAR_RE = re.compile(r"[^-!#$%&'*+.^_`|~0-9a-zA-Z]")
+
+
+class HTTPConnection(_HTTPConnection, object):
+    """
+    Based on :class:`http.client.HTTPConnection` but provides an extra constructor
+    backwards-compatibility layer between older and newer Pythons.
+
+    Additional keyword parameters are used to configure attributes of the connection.
+    Accepted parameters include:
+
+    - ``strict``: See the documentation on :class:`urllib3.connectionpool.HTTPConnectionPool`
+    - ``source_address``: Set the source address for the current connection.
+    - ``socket_options``: Set specific options on the underlying socket. If not specified, then
+      defaults are loaded from ``HTTPConnection.default_socket_options`` which includes disabling
+      Nagle's algorithm (sets TCP_NODELAY to 1) unless the connection is behind a proxy.
+
+      For example, if you wish to enable TCP Keep Alive in addition to the defaults,
+      you might pass:
+
+      .. code-block:: python
+
+         HTTPConnection.default_socket_options + [
+             (socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1),
+         ]
+
+      Or you may want to disable the defaults by passing an empty list (e.g., ``[]``).
+    """
+
+    default_port = port_by_scheme["http"]
+
+    #: Disable Nagle's algorithm by default.
+    #: ``[(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)]``
+    default_socket_options = [(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)]
+
+    #: Whether this connection verifies the host's certificate.
+    is_verified = False
+
+    #: Whether this proxy connection (if used) verifies the proxy host's
+    #: certificate.
+    proxy_is_verified = None
+
+    def __init__(self, *args, **kw):
+        if not six.PY2:
+            kw.pop("strict", None)
+
+        # Pre-set source_address.
+        self.source_address = kw.get("source_address")
+
+        #: The socket options provided by the user. If no options are
+        #: provided, we use the default options.
+        self.socket_options = kw.pop("socket_options", self.default_socket_options)
+
+        # Proxy options provided by the user.
+        self.proxy = kw.pop("proxy", None)
+        self.proxy_config = kw.pop("proxy_config", None)
+
+        _HTTPConnection.__init__(self, *args, **kw)
+
+    @property
+    def host(self):
+        """
+        Getter method to remove any trailing dots that indicate the hostname is an FQDN.
+
+        In general, SSL certificates don't include the trailing dot indicating a
+        fully-qualified domain name, and thus, they don't validate properly when
+        checked against a domain name that includes the dot. In addition, some
+        servers may not expect to receive the trailing dot when provided.
+
+        However, the hostname with trailing dot is critical to DNS resolution; doing a
+        lookup with the trailing dot will properly only resolve the appropriate FQDN,
+        whereas a lookup without a trailing dot will search the system's search domain
+        list. Thus, it's important to keep the original host around for use only in
+        those cases where it's appropriate (i.e., when doing DNS lookup to establish the
+        actual TCP connection across which we're going to send HTTP requests).
+        """
+        return self._dns_host.rstrip(".")
+
+    @host.setter
+    def host(self, value):
+        """
+        Setter for the `host` property.
+
+        We assume that only urllib3 uses the _dns_host attribute; httplib itself
+        only uses `host`, and it seems reasonable that other libraries follow suit.
+        """
+        self._dns_host = value
+
+    def _new_conn(self):
+        """Establish a socket connection and set nodelay settings on it.
+
+        :return: New socket connection.
+        """
+        extra_kw = {}
+        if self.source_address:
+            extra_kw["source_address"] = self.source_address
+
+        if self.socket_options:
+            extra_kw["socket_options"] = self.socket_options
+
+        try:
+            conn = connection.create_connection(
+                (self._dns_host, self.port), self.timeout, **extra_kw
+            )
+
+        except SocketTimeout:
+            raise ConnectTimeoutError(
+                self,
+                "Connection to %s timed out. (connect timeout=%s)"
+                % (self.host, self.timeout),
+            )
+
+        except SocketError as e:
+            raise NewConnectionError(
+                self, "Failed to establish a new connection: %s" % e
+            )
+
+        return conn
+
+    def _is_using_tunnel(self):
+        # Google App Engine's httplib does not define _tunnel_host
+        return getattr(self, "_tunnel_host", None)
+
+    def _prepare_conn(self, conn):
+        self.sock = conn
+        if self._is_using_tunnel():
+            # TODO: Fix tunnel so it doesn't depend on self.sock state.
+            self._tunnel()
+            # Mark this connection as not reusable
+            self.auto_open = 0
+
+    def connect(self):
+        conn = self._new_conn()
+        self._prepare_conn(conn)
+
+    def putrequest(self, method, url, *args, **kwargs):
+        """ """
+        # Empty docstring because the indentation of CPython's implementation
+        # is broken but we don't want this method in our documentation.
+        match = _CONTAINS_CONTROL_CHAR_RE.search(method)
+        if match:
+            raise ValueError(
+                "Method cannot contain non-token characters %r (found at least %r)"
+                % (method, match.group())
+            )
+
+        return _HTTPConnection.putrequest(self, method, url, *args, **kwargs)
+
+    def putheader(self, header, *values):
+        """ """
+        if not any(isinstance(v, str) and v == SKIP_HEADER for v in values):
+            _HTTPConnection.putheader(self, header, *values)
+        elif six.ensure_str(header.lower()) not in SKIPPABLE_HEADERS:
+            raise ValueError(
+                "urllib3.util.SKIP_HEADER only supports '%s'"
+                % ("', '".join(map(str.title, sorted(SKIPPABLE_HEADERS))),)
+            )
+
+    def request(self, method, url, body=None, headers=None):
+        # Update the inner socket's timeout value to send the request.
+        # This only triggers if the connection is re-used.
+        if getattr(self, "sock", None) is not None:
+            self.sock.settimeout(self.timeout)
+
+        if headers is None:
+            headers = {}
+        else:
+            # Avoid modifying the headers passed into .request()
+            headers = headers.copy()
+        if "user-agent" not in (six.ensure_str(k.lower()) for k in headers):
+            headers["User-Agent"] = _get_default_user_agent()
+        super(HTTPConnection, self).request(method, url, body=body, headers=headers)
+
+    def request_chunked(self, method, url, body=None, headers=None):
+        """
+        Alternative to the common request method, which sends the
+        body with chunked encoding and not as one block
+        """
+        headers = headers or {}
+        header_keys = set([six.ensure_str(k.lower()) for k in headers])
+        skip_accept_encoding = "accept-encoding" in header_keys
+        skip_host = "host" in header_keys
+        self.putrequest(
+            method, url, skip_accept_encoding=skip_accept_encoding, skip_host=skip_host
+        )
+        if "user-agent" not in header_keys:
+            self.putheader("User-Agent", _get_default_user_agent())
+        for header, value in headers.items():
+            self.putheader(header, value)
+        if "transfer-encoding" not in header_keys:
+            self.putheader("Transfer-Encoding", "chunked")
+        self.endheaders()
+
+        if body is not None:
+            stringish_types = six.string_types + (bytes,)
+            if isinstance(body, stringish_types):
+                body = (body,)
+            for chunk in body:
+                if not chunk:
+                    continue
+                if not isinstance(chunk, bytes):
+                    chunk = chunk.encode("utf8")
+                len_str = hex(len(chunk))[2:]
+                to_send = bytearray(len_str.encode())
+                to_send += b"\r\n"
+                to_send += chunk
+                to_send += b"\r\n"
+                self.send(to_send)
+
+        # After the if clause, to always have a closed body
+        self.send(b"0\r\n\r\n")
+
+
+class HTTPSConnection(HTTPConnection):
+    """
+    Many of the parameters to this constructor are passed to the underlying SSL
+    socket by means of :py:func:`urllib3.util.ssl_wrap_socket`.
+    """
+
+    default_port = port_by_scheme["https"]
+
+    cert_reqs = None
+    ca_certs = None
+    ca_cert_dir = None
+    ca_cert_data = None
+    ssl_version = None
+    assert_fingerprint = None
+    tls_in_tls_required = False
+
+    def __init__(
+        self,
+        host,
+        port=None,
+        key_file=None,
+        cert_file=None,
+        key_password=None,
+        strict=None,
+        timeout=socket._GLOBAL_DEFAULT_TIMEOUT,
+        ssl_context=None,
+        server_hostname=None,
+        **kw
+    ):
+
+        HTTPConnection.__init__(self, host, port, strict=strict, timeout=timeout, **kw)
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.key_password = key_password
+        self.ssl_context = ssl_context
+        self.server_hostname = server_hostname
+
+        # Required property for Google AppEngine 1.9.0 which otherwise causes
+        # HTTPS requests to go out as HTTP. (See Issue #356)
+        self._protocol = "https"
+
+    def set_cert(
+        self,
+        key_file=None,
+        cert_file=None,
+        cert_reqs=None,
+        key_password=None,
+        ca_certs=None,
+        assert_hostname=None,
+        assert_fingerprint=None,
+        ca_cert_dir=None,
+        ca_cert_data=None,
+    ):
+        """
+        This method should only be called once, before the connection is used.
+        """
+        # If cert_reqs is not provided we'll assume CERT_REQUIRED unless we also
+        # have an SSLContext object in which case we'll use its verify_mode.
+        if cert_reqs is None:
+            if self.ssl_context is not None:
+                cert_reqs = self.ssl_context.verify_mode
+            else:
+                cert_reqs = resolve_cert_reqs(None)
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.cert_reqs = cert_reqs
+        self.key_password = key_password
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+        self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
+        self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
+        self.ca_cert_data = ca_cert_data
+
+    def connect(self):
+        # Add certificate verification
+        self.sock = conn = self._new_conn()
+        hostname = self.host
+        tls_in_tls = False
+
+        if self._is_using_tunnel():
+            if self.tls_in_tls_required:
+                self.sock = conn = self._connect_tls_proxy(hostname, conn)
+                tls_in_tls = True
+
+            # Calls self._set_hostport(), so self.host is
+            # self._tunnel_host below.
+            self._tunnel()
+            # Mark this connection as not reusable
+            self.auto_open = 0
+
+            # Override the host with the one we're requesting data from.
+            hostname = self._tunnel_host
+
+        server_hostname = hostname
+        if self.server_hostname is not None:
+            server_hostname = self.server_hostname
+
+        is_time_off = datetime.date.today() < RECENT_DATE
+        if is_time_off:
+            warnings.warn(
+                (
+                    "System time is way off (before {0}). This will probably "
+                    "lead to SSL verification errors"
+                ).format(RECENT_DATE),
+                SystemTimeWarning,
+            )
+
+        # Wrap socket using verification with the root certs in
+        # trusted_root_certs
+        default_ssl_context = False
+        if self.ssl_context is None:
+            default_ssl_context = True
+            self.ssl_context = create_urllib3_context(
+                ssl_version=resolve_ssl_version(self.ssl_version),
+                cert_reqs=resolve_cert_reqs(self.cert_reqs),
+            )
+
+        context = self.ssl_context
+        context.verify_mode = resolve_cert_reqs(self.cert_reqs)
+
+        # Try to load OS default certs if none are given.
+        # Works well on Windows (requires Python3.4+)
+        if (
+            not self.ca_certs
+            and not self.ca_cert_dir
+            and not self.ca_cert_data
+            and default_ssl_context
+            and hasattr(context, "load_default_certs")
+        ):
+            context.load_default_certs()
+
+        self.sock = ssl_wrap_socket(
+            sock=conn,
+            keyfile=self.key_file,
+            certfile=self.cert_file,
+            key_password=self.key_password,
+            ca_certs=self.ca_certs,
+            ca_cert_dir=self.ca_cert_dir,
+            ca_cert_data=self.ca_cert_data,
+            server_hostname=server_hostname,
+            ssl_context=context,
+            tls_in_tls=tls_in_tls,
+        )
+
+        # If we're using all defaults and the connection
+        # is TLSv1 or TLSv1.1 we throw a DeprecationWarning
+        # for the host.
+        if (
+            default_ssl_context
+            and self.ssl_version is None
+            and hasattr(self.sock, "version")
+            and self.sock.version() in {"TLSv1", "TLSv1.1"}
+        ):  # Defensive:
+            warnings.warn(
+                "Negotiating TLSv1/TLSv1.1 by default is deprecated "
+                "and will be disabled in urllib3 v2.0.0. Connecting to "
+                "'%s' with '%s' can be enabled by explicitly opting-in "
+                "with 'ssl_version'" % (self.host, self.sock.version()),
+                DeprecationWarning,
+            )
+
+        if self.assert_fingerprint:
+            assert_fingerprint(
+                self.sock.getpeercert(binary_form=True), self.assert_fingerprint
+            )
+        elif (
+            context.verify_mode != ssl.CERT_NONE
+            and not getattr(context, "check_hostname", False)
+            and self.assert_hostname is not False
+        ):
+            # While urllib3 attempts to always turn off hostname matching from
+            # the TLS library, this cannot always be done. So we check whether
+            # the TLS Library still thinks it's matching hostnames.
+            cert = self.sock.getpeercert()
+            if not cert.get("subjectAltName", ()):
+                warnings.warn(
+                    (
+                        "Certificate for {0} has no `subjectAltName`, falling back to check for a "
+                        "`commonName` for now. This feature is being removed by major browsers and "
+                        "deprecated by RFC 2818. (See https://github.com/urllib3/urllib3/issues/497 "
+                        "for details.)".format(hostname)
+                    ),
+                    SubjectAltNameWarning,
+                )
+            _match_hostname(cert, self.assert_hostname or server_hostname)
+
+        self.is_verified = (
+            context.verify_mode == ssl.CERT_REQUIRED
+            or self.assert_fingerprint is not None
+        )
+
+    def _connect_tls_proxy(self, hostname, conn):
+        """
+        Establish a TLS connection to the proxy using the provided SSL context.
+        """
+        proxy_config = self.proxy_config
+        ssl_context = proxy_config.ssl_context
+        if ssl_context:
+            # If the user provided a proxy context, we assume CA and client
+            # certificates have already been set
+            return ssl_wrap_socket(
+                sock=conn,
+                server_hostname=hostname,
+                ssl_context=ssl_context,
+            )
+
+        ssl_context = create_proxy_ssl_context(
+            self.ssl_version,
+            self.cert_reqs,
+            self.ca_certs,
+            self.ca_cert_dir,
+            self.ca_cert_data,
+        )
+
+        # If no cert was provided, use only the default options for server
+        # certificate validation
+        socket = ssl_wrap_socket(
+            sock=conn,
+            ca_certs=self.ca_certs,
+            ca_cert_dir=self.ca_cert_dir,
+            ca_cert_data=self.ca_cert_data,
+            server_hostname=hostname,
+            ssl_context=ssl_context,
+        )
+
+        if ssl_context.verify_mode != ssl.CERT_NONE and not getattr(
+            ssl_context, "check_hostname", False
+        ):
+            # While urllib3 attempts to always turn off hostname matching from
+            # the TLS library, this cannot always be done. So we check whether
+            # the TLS Library still thinks it's matching hostnames.
+            cert = socket.getpeercert()
+            if not cert.get("subjectAltName", ()):
+                warnings.warn(
+                    (
+                        "Certificate for {0} has no `subjectAltName`, falling back to check for a "
+                        "`commonName` for now. This feature is being removed by major browsers and "
+                        "deprecated by RFC 2818. (See https://github.com/urllib3/urllib3/issues/497 "
+                        "for details.)".format(hostname)
+                    ),
+                    SubjectAltNameWarning,
+                )
+            _match_hostname(cert, hostname)
+
+        self.proxy_is_verified = ssl_context.verify_mode == ssl.CERT_REQUIRED
+        return socket
+
+
+def _match_hostname(cert, asserted_hostname):
+    # Our upstream implementation of ssl.match_hostname()
+    # only applies this normalization to IP addresses so it doesn't
+    # match DNS SANs so we do the same thing!
+    stripped_hostname = asserted_hostname.strip("u[]")
+    if is_ipaddress(stripped_hostname):
+        asserted_hostname = stripped_hostname
+
+    try:
+        match_hostname(cert, asserted_hostname)
+    except CertificateError as e:
+        log.warning(
+            "Certificate did not match expected hostname: %s. Certificate: %s",
+            asserted_hostname,
+            cert,
+        )
+        # Add cert to exception and reraise so client code can inspect
+        # the cert when catching the exception, if they want to
+        e._peer_cert = cert
+        raise
+
+
+def _get_default_user_agent():
+    return "python-urllib3/%s" % __version__
+
+
+class DummyConnection(object):
+    """Used to detect a failed ConnectionCls import."""
+
+    pass
+
+
+if not ssl:
+    HTTPSConnection = DummyConnection  # noqa: F811
+
+
+VerifiedHTTPSConnection = HTTPSConnection
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connectionpool.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connectionpool.py
new file mode 100644
index 00000000..0872ed77
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/connectionpool.py
@@ -0,0 +1,1140 @@
+from __future__ import absolute_import
+
+import errno
+import logging
+import re
+import socket
+import sys
+import warnings
+from socket import error as SocketError
+from socket import timeout as SocketTimeout
+
+from ._collections import HTTPHeaderDict
+from .connection import (
+    BaseSSLError,
+    BrokenPipeError,
+    DummyConnection,
+    HTTPConnection,
+    HTTPException,
+    HTTPSConnection,
+    VerifiedHTTPSConnection,
+    port_by_scheme,
+)
+from .exceptions import (
+    ClosedPoolError,
+    EmptyPoolError,
+    HeaderParsingError,
+    HostChangedError,
+    InsecureRequestWarning,
+    LocationValueError,
+    MaxRetryError,
+    NewConnectionError,
+    ProtocolError,
+    ProxyError,
+    ReadTimeoutError,
+    SSLError,
+    TimeoutError,
+)
+from .packages import six
+from .packages.six.moves import queue
+from .request import RequestMethods
+from .response import HTTPResponse
+from .util.connection import is_connection_dropped
+from .util.proxy import connection_requires_http_tunnel
+from .util.queue import LifoQueue
+from .util.request import set_file_position
+from .util.response import assert_header_parsing
+from .util.retry import Retry
+from .util.ssl_match_hostname import CertificateError
+from .util.timeout import Timeout
+from .util.url import Url, _encode_target
+from .util.url import _normalize_host as normalize_host
+from .util.url import get_host, parse_url
+
+try:  # Platform-specific: Python 3
+    import weakref
+
+    weakref_finalize = weakref.finalize
+except AttributeError:  # Platform-specific: Python 2
+    from .packages.backports.weakref_finalize import weakref_finalize
+
+xrange = six.moves.xrange
+
+log = logging.getLogger(__name__)
+
+_Default = object()
+
+
+# Pool objects
+class ConnectionPool(object):
+    """
+    Base class for all connection pools, such as
+    :class:`.HTTPConnectionPool` and :class:`.HTTPSConnectionPool`.
+
+    .. note::
+       ConnectionPool.urlopen() does not normalize or percent-encode target URIs
+       which is useful if your target server doesn't support percent-encoded
+       target URIs.
+    """
+
+    scheme = None
+    QueueCls = LifoQueue
+
+    def __init__(self, host, port=None):
+        if not host:
+            raise LocationValueError("No host specified.")
+
+        self.host = _normalize_host(host, scheme=self.scheme)
+        self._proxy_host = host.lower()
+        self.port = port
+
+    def __str__(self):
+        return "%s(host=%r, port=%r)" % (type(self).__name__, self.host, self.port)
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.close()
+        # Return False to re-raise any potential exceptions
+        return False
+
+    def close(self):
+        """
+        Close all pooled connections and disable the pool.
+        """
+        pass
+
+
+# This is taken from http://hg.python.org/cpython/file/7aaba721ebc0/Lib/socket.py#l252
+_blocking_errnos = {errno.EAGAIN, errno.EWOULDBLOCK}
+
+
+class HTTPConnectionPool(ConnectionPool, RequestMethods):
+    """
+    Thread-safe connection pool for one host.
+
+    :param host:
+        Host used for this HTTP Connection (e.g. "localhost"), passed into
+        :class:`http.client.HTTPConnection`.
+
+    :param port:
+        Port used for this HTTP Connection (None is equivalent to 80), passed
+        into :class:`http.client.HTTPConnection`.
+
+    :param strict:
+        Causes BadStatusLine to be raised if the status line can't be parsed
+        as a valid HTTP/1.0 or 1.1 status line, passed into
+        :class:`http.client.HTTPConnection`.
+
+        .. note::
+           Only works in Python 2. This parameter is ignored in Python 3.
+
+    :param timeout:
+        Socket timeout in seconds for each individual connection. This can
+        be a float or integer, which sets the timeout for the HTTP request,
+        or an instance of :class:`urllib3.util.Timeout` which gives you more
+        fine-grained control over request timeouts. After the constructor has
+        been parsed, this is always a `urllib3.util.Timeout` object.
+
+    :param maxsize:
+        Number of connections to save that can be reused. More than 1 is useful
+        in multithreaded situations. If ``block`` is set to False, more
+        connections will be created but they will not be saved once they've
+        been used.
+
+    :param block:
+        If set to True, no more than ``maxsize`` connections will be used at
+        a time. When no free connections are available, the call will block
+        until a connection has been released. This is a useful side effect for
+        particular multithreaded situations where one does not want to use more
+        than maxsize connections per host to prevent flooding.
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+
+    :param retries:
+        Retry configuration to use by default with requests in this pool.
+
+    :param _proxy:
+        Parsed proxy URL, should not be used directly, instead, see
+        :class:`urllib3.ProxyManager`
+
+    :param _proxy_headers:
+        A dictionary with proxy headers, should not be used directly,
+        instead, see :class:`urllib3.ProxyManager`
+
+    :param \\**conn_kw:
+        Additional parameters are used to create fresh :class:`urllib3.connection.HTTPConnection`,
+        :class:`urllib3.connection.HTTPSConnection` instances.
+    """
+
+    scheme = "http"
+    ConnectionCls = HTTPConnection
+    ResponseCls = HTTPResponse
+
+    def __init__(
+        self,
+        host,
+        port=None,
+        strict=False,
+        timeout=Timeout.DEFAULT_TIMEOUT,
+        maxsize=1,
+        block=False,
+        headers=None,
+        retries=None,
+        _proxy=None,
+        _proxy_headers=None,
+        _proxy_config=None,
+        **conn_kw
+    ):
+        ConnectionPool.__init__(self, host, port)
+        RequestMethods.__init__(self, headers)
+
+        self.strict = strict
+
+        if not isinstance(timeout, Timeout):
+            timeout = Timeout.from_float(timeout)
+
+        if retries is None:
+            retries = Retry.DEFAULT
+
+        self.timeout = timeout
+        self.retries = retries
+
+        self.pool = self.QueueCls(maxsize)
+        self.block = block
+
+        self.proxy = _proxy
+        self.proxy_headers = _proxy_headers or {}
+        self.proxy_config = _proxy_config
+
+        # Fill the queue up so that doing get() on it will block properly
+        for _ in xrange(maxsize):
+            self.pool.put(None)
+
+        # These are mostly for testing and debugging purposes.
+        self.num_connections = 0
+        self.num_requests = 0
+        self.conn_kw = conn_kw
+
+        if self.proxy:
+            # Enable Nagle's algorithm for proxies, to avoid packet fragmentation.
+            # We cannot know if the user has added default socket options, so we cannot replace the
+            # list.
+            self.conn_kw.setdefault("socket_options", [])
+
+            self.conn_kw["proxy"] = self.proxy
+            self.conn_kw["proxy_config"] = self.proxy_config
+
+        # Do not pass 'self' as callback to 'finalize'.
+        # Then the 'finalize' would keep an endless living (leak) to self.
+        # By just passing a reference to the pool allows the garbage collector
+        # to free self if nobody else has a reference to it.
+        pool = self.pool
+
+        # Close all the HTTPConnections in the pool before the
+        # HTTPConnectionPool object is garbage collected.
+        weakref_finalize(self, _close_pool_connections, pool)
+
+    def _new_conn(self):
+        """
+        Return a fresh :class:`HTTPConnection`.
+        """
+        self.num_connections += 1
+        log.debug(
+            "Starting new HTTP connection (%d): %s:%s",
+            self.num_connections,
+            self.host,
+            self.port or "80",
+        )
+
+        conn = self.ConnectionCls(
+            host=self.host,
+            port=self.port,
+            timeout=self.timeout.connect_timeout,
+            strict=self.strict,
+            **self.conn_kw
+        )
+        return conn
+
+    def _get_conn(self, timeout=None):
+        """
+        Get a connection. Will return a pooled connection if one is available.
+
+        If no connections are available and :prop:`.block` is ``False``, then a
+        fresh connection is returned.
+
+        :param timeout:
+            Seconds to wait before giving up and raising
+            :class:`urllib3.exceptions.EmptyPoolError` if the pool is empty and
+            :prop:`.block` is ``True``.
+        """
+        conn = None
+        try:
+            conn = self.pool.get(block=self.block, timeout=timeout)
+
+        except AttributeError:  # self.pool is None
+            raise ClosedPoolError(self, "Pool is closed.")
+
+        except queue.Empty:
+            if self.block:
+                raise EmptyPoolError(
+                    self,
+                    "Pool reached maximum size and no more connections are allowed.",
+                )
+            pass  # Oh well, we'll create a new connection then
+
+        # If this is a persistent connection, check if it got disconnected
+        if conn and is_connection_dropped(conn):
+            log.debug("Resetting dropped connection: %s", self.host)
+            conn.close()
+            if getattr(conn, "auto_open", 1) == 0:
+                # This is a proxied connection that has been mutated by
+                # http.client._tunnel() and cannot be reused (since it would
+                # attempt to bypass the proxy)
+                conn = None
+
+        return conn or self._new_conn()
+
+    def _put_conn(self, conn):
+        """
+        Put a connection back into the pool.
+
+        :param conn:
+            Connection object for the current host and port as returned by
+            :meth:`._new_conn` or :meth:`._get_conn`.
+
+        If the pool is already full, the connection is closed and discarded
+        because we exceeded maxsize. If connections are discarded frequently,
+        then maxsize should be increased.
+
+        If the pool is closed, then the connection will be closed and discarded.
+        """
+        try:
+            self.pool.put(conn, block=False)
+            return  # Everything is dandy, done.
+        except AttributeError:
+            # self.pool is None.
+            pass
+        except queue.Full:
+            # This should never happen if self.block == True
+            log.warning(
+                "Connection pool is full, discarding connection: %s. Connection pool size: %s",
+                self.host,
+                self.pool.qsize(),
+            )
+        # Connection never got put back into the pool, close it.
+        if conn:
+            conn.close()
+
+    def _validate_conn(self, conn):
+        """
+        Called right before a request is made, after the socket is created.
+        """
+        pass
+
+    def _prepare_proxy(self, conn):
+        # Nothing to do for HTTP connections.
+        pass
+
+    def _get_timeout(self, timeout):
+        """Helper that always returns a :class:`urllib3.util.Timeout`"""
+        if timeout is _Default:
+            return self.timeout.clone()
+
+        if isinstance(timeout, Timeout):
+            return timeout.clone()
+        else:
+            # User passed us an int/float. This is for backwards compatibility,
+            # can be removed later
+            return Timeout.from_float(timeout)
+
+    def _raise_timeout(self, err, url, timeout_value):
+        """Is the error actually a timeout? Will raise a ReadTimeout or pass"""
+
+        if isinstance(err, SocketTimeout):
+            raise ReadTimeoutError(
+                self, url, "Read timed out. (read timeout=%s)" % timeout_value
+            )
+
+        # See the above comment about EAGAIN in Python 3. In Python 2 we have
+        # to specifically catch it and throw the timeout error
+        if hasattr(err, "errno") and err.errno in _blocking_errnos:
+            raise ReadTimeoutError(
+                self, url, "Read timed out. (read timeout=%s)" % timeout_value
+            )
+
+        # Catch possible read timeouts thrown as SSL errors. If not the
+        # case, rethrow the original. We need to do this because of:
+        # http://bugs.python.org/issue10272
+        if "timed out" in str(err) or "did not complete (read)" in str(
+            err
+        ):  # Python < 2.7.4
+            raise ReadTimeoutError(
+                self, url, "Read timed out. (read timeout=%s)" % timeout_value
+            )
+
+    def _make_request(
+        self, conn, method, url, timeout=_Default, chunked=False, **httplib_request_kw
+    ):
+        """
+        Perform a request on a given urllib connection object taken from our
+        pool.
+
+        :param conn:
+            a connection from one of our connection pools
+
+        :param timeout:
+            Socket timeout in seconds for the request. This can be a
+            float or integer, which will set the same timeout value for
+            the socket connect and the socket read, or an instance of
+            :class:`urllib3.util.Timeout`, which gives you more fine-grained
+            control over your timeouts.
+        """
+        self.num_requests += 1
+
+        timeout_obj = self._get_timeout(timeout)
+        timeout_obj.start_connect()
+        conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout)
+
+        # Trigger any extra validation we need to do.
+        try:
+            self._validate_conn(conn)
+        except (SocketTimeout, BaseSSLError) as e:
+            # Py2 raises this as a BaseSSLError, Py3 raises it as socket timeout.
+            self._raise_timeout(err=e, url=url, timeout_value=conn.timeout)
+            raise
+
+        # conn.request() calls http.client.*.request, not the method in
+        # urllib3.request. It also calls makefile (recv) on the socket.
+        try:
+            if chunked:
+                conn.request_chunked(method, url, **httplib_request_kw)
+            else:
+                conn.request(method, url, **httplib_request_kw)
+
+        # We are swallowing BrokenPipeError (errno.EPIPE) since the server is
+        # legitimately able to close the connection after sending a valid response.
+        # With this behaviour, the received response is still readable.
+        except BrokenPipeError:
+            # Python 3
+            pass
+        except IOError as e:
+            # Python 2 and macOS/Linux
+            # EPIPE and ESHUTDOWN are BrokenPipeError on Python 2, and EPROTOTYPE/ECONNRESET are needed on macOS
+            # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/
+            if e.errno not in {
+                errno.EPIPE,
+                errno.ESHUTDOWN,
+                errno.EPROTOTYPE,
+                errno.ECONNRESET,
+            }:
+                raise
+
+        # Reset the timeout for the recv() on the socket
+        read_timeout = timeout_obj.read_timeout
+
+        # App Engine doesn't have a sock attr
+        if getattr(conn, "sock", None):
+            # In Python 3 socket.py will catch EAGAIN and return None when you
+            # try and read into the file pointer created by http.client, which
+            # instead raises a BadStatusLine exception. Instead of catching
+            # the exception and assuming all BadStatusLine exceptions are read
+            # timeouts, check for a zero timeout before making the request.
+            if read_timeout == 0:
+                raise ReadTimeoutError(
+                    self, url, "Read timed out. (read timeout=%s)" % read_timeout
+                )
+            if read_timeout is Timeout.DEFAULT_TIMEOUT:
+                conn.sock.settimeout(socket.getdefaulttimeout())
+            else:  # None or a value
+                conn.sock.settimeout(read_timeout)
+
+        # Receive the response from the server
+        try:
+            try:
+                # Python 2.7, use buffering of HTTP responses
+                httplib_response = conn.getresponse(buffering=True)
+            except TypeError:
+                # Python 3
+                try:
+                    httplib_response = conn.getresponse()
+                except BaseException as e:
+                    # Remove the TypeError from the exception chain in
+                    # Python 3 (including for exceptions like SystemExit).
+                    # Otherwise it looks like a bug in the code.
+                    six.raise_from(e, None)
+        except (SocketTimeout, BaseSSLError, SocketError) as e:
+            self._raise_timeout(err=e, url=url, timeout_value=read_timeout)
+            raise
+
+        # AppEngine doesn't have a version attr.
+        http_version = getattr(conn, "_http_vsn_str", "HTTP/?")
+        log.debug(
+            '%s://%s:%s "%s %s %s" %s %s',
+            self.scheme,
+            self.host,
+            self.port,
+            method,
+            url,
+            http_version,
+            httplib_response.status,
+            httplib_response.length,
+        )
+
+        try:
+            assert_header_parsing(httplib_response.msg)
+        except (HeaderParsingError, TypeError) as hpe:  # Platform-specific: Python 3
+            log.warning(
+                "Failed to parse headers (url=%s): %s",
+                self._absolute_url(url),
+                hpe,
+                exc_info=True,
+            )
+
+        return httplib_response
+
+    def _absolute_url(self, path):
+        return Url(scheme=self.scheme, host=self.host, port=self.port, path=path).url
+
+    def close(self):
+        """
+        Close all pooled connections and disable the pool.
+        """
+        if self.pool is None:
+            return
+        # Disable access to the pool
+        old_pool, self.pool = self.pool, None
+
+        # Close all the HTTPConnections in the pool.
+        _close_pool_connections(old_pool)
+
+    def is_same_host(self, url):
+        """
+        Check if the given ``url`` is a member of the same host as this
+        connection pool.
+        """
+        if url.startswith("/"):
+            return True
+
+        # TODO: Add optional support for socket.gethostbyname checking.
+        scheme, host, port = get_host(url)
+        if host is not None:
+            host = _normalize_host(host, scheme=scheme)
+
+        # Use explicit default port for comparison when none is given
+        if self.port and not port:
+            port = port_by_scheme.get(scheme)
+        elif not self.port and port == port_by_scheme.get(scheme):
+            port = None
+
+        return (scheme, host, port) == (self.scheme, self.host, self.port)
+
+    def urlopen(
+        self,
+        method,
+        url,
+        body=None,
+        headers=None,
+        retries=None,
+        redirect=True,
+        assert_same_host=True,
+        timeout=_Default,
+        pool_timeout=None,
+        release_conn=None,
+        chunked=False,
+        body_pos=None,
+        **response_kw
+    ):
+        """
+        Get a connection from the pool and perform an HTTP request. This is the
+        lowest level call for making a request, so you'll need to specify all
+        the raw details.
+
+        .. note::
+
+           More commonly, it's appropriate to use a convenience method provided
+           by :class:`.RequestMethods`, such as :meth:`request`.
+
+        .. note::
+
+           `release_conn` will only behave as expected if
+           `preload_content=False` because we want to make
+           `preload_content=False` the default behaviour someday soon without
+           breaking backwards compatibility.
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param body:
+            Data to send in the request body, either :class:`str`, :class:`bytes`,
+            an iterable of :class:`str`/:class:`bytes`, or a file-like object.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+
+        :param retries:
+            Configure the number of retries to allow before raising a
+            :class:`~urllib3.exceptions.MaxRetryError` exception.
+
+            Pass ``None`` to retry until you receive a response. Pass a
+            :class:`~urllib3.util.retry.Retry` object for fine-grained control
+            over different types of retries.
+            Pass an integer number to retry connection errors that many times,
+            but no other types of errors. Pass zero to never retry.
+
+            If ``False``, then retries are disabled and any exception is raised
+            immediately. Also, instead of raising a MaxRetryError on redirects,
+            the redirect response will be returned.
+
+        :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int.
+
+        :param redirect:
+            If True, automatically handle redirects (status codes 301, 302,
+            303, 307, 308). Each redirect counts as a retry. Disabling retries
+            will disable redirect, too.
+
+        :param assert_same_host:
+            If ``True``, will make sure that the host of the pool requests is
+            consistent else will raise HostChangedError. When ``False``, you can
+            use the pool on an HTTP proxy and request foreign hosts.
+
+        :param timeout:
+            If specified, overrides the default timeout for this one
+            request. It may be a float (in seconds) or an instance of
+            :class:`urllib3.util.Timeout`.
+
+        :param pool_timeout:
+            If set and the pool is set to block=True, then this method will
+            block for ``pool_timeout`` seconds and raise EmptyPoolError if no
+            connection is available within the time period.
+
+        :param release_conn:
+            If False, then the urlopen call will not release the connection
+            back into the pool once a response is received (but will release if
+            you read the entire contents of the response such as when
+            `preload_content=True`). This is useful if you're not preloading
+            the response's content immediately. You will need to call
+            ``r.release_conn()`` on the response ``r`` to return the connection
+            back into the pool. If None, it takes the value of
+            ``response_kw.get('preload_content', True)``.
+
+        :param chunked:
+            If True, urllib3 will send the body using chunked transfer
+            encoding. Otherwise, urllib3 will send the body using the standard
+            content-length form. Defaults to False.
+
+        :param int body_pos:
+            Position to seek to in file-like body in the event of a retry or
+            redirect. Typically this won't need to be set because urllib3 will
+            auto-populate the value when needed.
+
+        :param \\**response_kw:
+            Additional parameters are passed to
+            :meth:`urllib3.response.HTTPResponse.from_httplib`
+        """
+
+        parsed_url = parse_url(url)
+        destination_scheme = parsed_url.scheme
+
+        if headers is None:
+            headers = self.headers
+
+        if not isinstance(retries, Retry):
+            retries = Retry.from_int(retries, redirect=redirect, default=self.retries)
+
+        if release_conn is None:
+            release_conn = response_kw.get("preload_content", True)
+
+        # Check host
+        if assert_same_host and not self.is_same_host(url):
+            raise HostChangedError(self, url, retries)
+
+        # Ensure that the URL we're connecting to is properly encoded
+        if url.startswith("/"):
+            url = six.ensure_str(_encode_target(url))
+        else:
+            url = six.ensure_str(parsed_url.url)
+
+        conn = None
+
+        # Track whether `conn` needs to be released before
+        # returning/raising/recursing. Update this variable if necessary, and
+        # leave `release_conn` constant throughout the function. That way, if
+        # the function recurses, the original value of `release_conn` will be
+        # passed down into the recursive call, and its value will be respected.
+        #
+        # See issue #651 [1] for details.
+        #
+        # [1] <https://github.com/urllib3/urllib3/issues/651>
+        release_this_conn = release_conn
+
+        http_tunnel_required = connection_requires_http_tunnel(
+            self.proxy, self.proxy_config, destination_scheme
+        )
+
+        # Merge the proxy headers. Only done when not using HTTP CONNECT. We
+        # have to copy the headers dict so we can safely change it without those
+        # changes being reflected in anyone else's copy.
+        if not http_tunnel_required:
+            headers = headers.copy()
+            headers.update(self.proxy_headers)
+
+        # Must keep the exception bound to a separate variable or else Python 3
+        # complains about UnboundLocalError.
+        err = None
+
+        # Keep track of whether we cleanly exited the except block. This
+        # ensures we do proper cleanup in finally.
+        clean_exit = False
+
+        # Rewind body position, if needed. Record current position
+        # for future rewinds in the event of a redirect/retry.
+        body_pos = set_file_position(body, body_pos)
+
+        try:
+            # Request a connection from the queue.
+            timeout_obj = self._get_timeout(timeout)
+            conn = self._get_conn(timeout=pool_timeout)
+
+            conn.timeout = timeout_obj.connect_timeout
+
+            is_new_proxy_conn = self.proxy is not None and not getattr(
+                conn, "sock", None
+            )
+            if is_new_proxy_conn and http_tunnel_required:
+                self._prepare_proxy(conn)
+
+            # Make the request on the httplib connection object.
+            httplib_response = self._make_request(
+                conn,
+                method,
+                url,
+                timeout=timeout_obj,
+                body=body,
+                headers=headers,
+                chunked=chunked,
+            )
+
+            # If we're going to release the connection in ``finally:``, then
+            # the response doesn't need to know about the connection. Otherwise
+            # it will also try to release it and we'll have a double-release
+            # mess.
+            response_conn = conn if not release_conn else None
+
+            # Pass method to Response for length checking
+            response_kw["request_method"] = method
+
+            # Import httplib's response into our own wrapper object
+            response = self.ResponseCls.from_httplib(
+                httplib_response,
+                pool=self,
+                connection=response_conn,
+                retries=retries,
+                **response_kw
+            )
+
+            # Everything went great!
+            clean_exit = True
+
+        except EmptyPoolError:
+            # Didn't get a connection from the pool, no need to clean up
+            clean_exit = True
+            release_this_conn = False
+            raise
+
+        except (
+            TimeoutError,
+            HTTPException,
+            SocketError,
+            ProtocolError,
+            BaseSSLError,
+            SSLError,
+            CertificateError,
+        ) as e:
+            # Discard the connection for these exceptions. It will be
+            # replaced during the next _get_conn() call.
+            clean_exit = False
+
+            def _is_ssl_error_message_from_http_proxy(ssl_error):
+                # We're trying to detect the message 'WRONG_VERSION_NUMBER' but
+                # SSLErrors are kinda all over the place when it comes to the message,
+                # so we try to cover our bases here!
+                message = " ".join(re.split("[^a-z]", str(ssl_error).lower()))
+                return (
+                    "wrong version number" in message
+                    or "unknown protocol" in message
+                    or "record layer failure" in message
+                )
+
+            # Try to detect a common user error with proxies which is to
+            # set an HTTP proxy to be HTTPS when it should be 'http://'
+            # (ie {'http': 'http://proxy', 'https': 'https://proxy'})
+            # Instead we add a nice error message and point to a URL.
+            if (
+                isinstance(e, BaseSSLError)
+                and self.proxy
+                and _is_ssl_error_message_from_http_proxy(e)
+                and conn.proxy
+                and conn.proxy.scheme == "https"
+            ):
+                e = ProxyError(
+                    "Your proxy appears to only use HTTP and not HTTPS, "
+                    "try changing your proxy URL to be HTTP. See: "
+                    "https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html"
+                    "#https-proxy-error-http-proxy",
+                    SSLError(e),
+                )
+            elif isinstance(e, (BaseSSLError, CertificateError)):
+                e = SSLError(e)
+            elif isinstance(e, (SocketError, NewConnectionError)) and self.proxy:
+                e = ProxyError("Cannot connect to proxy.", e)
+            elif isinstance(e, (SocketError, HTTPException)):
+                e = ProtocolError("Connection aborted.", e)
+
+            retries = retries.increment(
+                method, url, error=e, _pool=self, _stacktrace=sys.exc_info()[2]
+            )
+            retries.sleep()
+
+            # Keep track of the error for the retry warning.
+            err = e
+
+        finally:
+            if not clean_exit:
+                # We hit some kind of exception, handled or otherwise. We need
+                # to throw the connection away unless explicitly told not to.
+                # Close the connection, set the variable to None, and make sure
+                # we put the None back in the pool to avoid leaking it.
+                conn = conn and conn.close()
+                release_this_conn = True
+
+            if release_this_conn:
+                # Put the connection back to be reused. If the connection is
+                # expired then it will be None, which will get replaced with a
+                # fresh connection during _get_conn.
+                self._put_conn(conn)
+
+        if not conn:
+            # Try again
+            log.warning(
+                "Retrying (%r) after connection broken by '%r': %s", retries, err, url
+            )
+            return self.urlopen(
+                method,
+                url,
+                body,
+                headers,
+                retries,
+                redirect,
+                assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                **response_kw
+            )
+
+        # Handle redirect?
+        redirect_location = redirect and response.get_redirect_location()
+        if redirect_location:
+            if response.status == 303:
+                # Change the method according to RFC 9110, Section 15.4.4.
+                method = "GET"
+                # And lose the body not to transfer anything sensitive.
+                body = None
+                headers = HTTPHeaderDict(headers)._prepare_for_method_change()
+
+            try:
+                retries = retries.increment(method, url, response=response, _pool=self)
+            except MaxRetryError:
+                if retries.raise_on_redirect:
+                    response.drain_conn()
+                    raise
+                return response
+
+            response.drain_conn()
+            retries.sleep_for_retry(response)
+            log.debug("Redirecting %s -> %s", url, redirect_location)
+            return self.urlopen(
+                method,
+                redirect_location,
+                body,
+                headers,
+                retries=retries,
+                redirect=redirect,
+                assert_same_host=assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                **response_kw
+            )
+
+        # Check if we should retry the HTTP response.
+        has_retry_after = bool(response.headers.get("Retry-After"))
+        if retries.is_retry(method, response.status, has_retry_after):
+            try:
+                retries = retries.increment(method, url, response=response, _pool=self)
+            except MaxRetryError:
+                if retries.raise_on_status:
+                    response.drain_conn()
+                    raise
+                return response
+
+            response.drain_conn()
+            retries.sleep(response)
+            log.debug("Retry: %s", url)
+            return self.urlopen(
+                method,
+                url,
+                body,
+                headers,
+                retries=retries,
+                redirect=redirect,
+                assert_same_host=assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                **response_kw
+            )
+
+        return response
+
+
+class HTTPSConnectionPool(HTTPConnectionPool):
+    """
+    Same as :class:`.HTTPConnectionPool`, but HTTPS.
+
+    :class:`.HTTPSConnection` uses one of ``assert_fingerprint``,
+    ``assert_hostname`` and ``host`` in this order to verify connections.
+    If ``assert_hostname`` is False, no verification is done.
+
+    The ``key_file``, ``cert_file``, ``cert_reqs``, ``ca_certs``,
+    ``ca_cert_dir``, ``ssl_version``, ``key_password`` are only used if :mod:`ssl`
+    is available and are fed into :meth:`urllib3.util.ssl_wrap_socket` to upgrade
+    the connection socket into an SSL socket.
+    """
+
+    scheme = "https"
+    ConnectionCls = HTTPSConnection
+
+    def __init__(
+        self,
+        host,
+        port=None,
+        strict=False,
+        timeout=Timeout.DEFAULT_TIMEOUT,
+        maxsize=1,
+        block=False,
+        headers=None,
+        retries=None,
+        _proxy=None,
+        _proxy_headers=None,
+        key_file=None,
+        cert_file=None,
+        cert_reqs=None,
+        key_password=None,
+        ca_certs=None,
+        ssl_version=None,
+        assert_hostname=None,
+        assert_fingerprint=None,
+        ca_cert_dir=None,
+        **conn_kw
+    ):
+
+        HTTPConnectionPool.__init__(
+            self,
+            host,
+            port,
+            strict,
+            timeout,
+            maxsize,
+            block,
+            headers,
+            retries,
+            _proxy,
+            _proxy_headers,
+            **conn_kw
+        )
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.cert_reqs = cert_reqs
+        self.key_password = key_password
+        self.ca_certs = ca_certs
+        self.ca_cert_dir = ca_cert_dir
+        self.ssl_version = ssl_version
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+
+    def _prepare_conn(self, conn):
+        """
+        Prepare the ``connection`` for :meth:`urllib3.util.ssl_wrap_socket`
+        and establish the tunnel if proxy is used.
+        """
+
+        if isinstance(conn, VerifiedHTTPSConnection):
+            conn.set_cert(
+                key_file=self.key_file,
+                key_password=self.key_password,
+                cert_file=self.cert_file,
+                cert_reqs=self.cert_reqs,
+                ca_certs=self.ca_certs,
+                ca_cert_dir=self.ca_cert_dir,
+                assert_hostname=self.assert_hostname,
+                assert_fingerprint=self.assert_fingerprint,
+            )
+            conn.ssl_version = self.ssl_version
+        return conn
+
+    def _prepare_proxy(self, conn):
+        """
+        Establishes a tunnel connection through HTTP CONNECT.
+
+        Tunnel connection is established early because otherwise httplib would
+        improperly set Host: header to proxy's IP:port.
+        """
+
+        conn.set_tunnel(self._proxy_host, self.port, self.proxy_headers)
+
+        if self.proxy.scheme == "https":
+            conn.tls_in_tls_required = True
+
+        conn.connect()
+
+    def _new_conn(self):
+        """
+        Return a fresh :class:`http.client.HTTPSConnection`.
+        """
+        self.num_connections += 1
+        log.debug(
+            "Starting new HTTPS connection (%d): %s:%s",
+            self.num_connections,
+            self.host,
+            self.port or "443",
+        )
+
+        if not self.ConnectionCls or self.ConnectionCls is DummyConnection:
+            raise SSLError(
+                "Can't connect to HTTPS URL because the SSL module is not available."
+            )
+
+        actual_host = self.host
+        actual_port = self.port
+        if self.proxy is not None:
+            actual_host = self.proxy.host
+            actual_port = self.proxy.port
+
+        conn = self.ConnectionCls(
+            host=actual_host,
+            port=actual_port,
+            timeout=self.timeout.connect_timeout,
+            strict=self.strict,
+            cert_file=self.cert_file,
+            key_file=self.key_file,
+            key_password=self.key_password,
+            **self.conn_kw
+        )
+
+        return self._prepare_conn(conn)
+
+    def _validate_conn(self, conn):
+        """
+        Called right before a request is made, after the socket is created.
+        """
+        super(HTTPSConnectionPool, self)._validate_conn(conn)
+
+        # Force connect early to allow us to validate the connection.
+        if not getattr(conn, "sock", None):  # AppEngine might not have  `.sock`
+            conn.connect()
+
+        if not conn.is_verified:
+            warnings.warn(
+                (
+                    "Unverified HTTPS request is being made to host '%s'. "
+                    "Adding certificate verification is strongly advised. See: "
+                    "https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html"
+                    "#ssl-warnings" % conn.host
+                ),
+                InsecureRequestWarning,
+            )
+
+        if getattr(conn, "proxy_is_verified", None) is False:
+            warnings.warn(
+                (
+                    "Unverified HTTPS connection done to an HTTPS proxy. "
+                    "Adding certificate verification is strongly advised. See: "
+                    "https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html"
+                    "#ssl-warnings"
+                ),
+                InsecureRequestWarning,
+            )
+
+
+def connection_from_url(url, **kw):
+    """
+    Given a url, return an :class:`.ConnectionPool` instance of its host.
+
+    This is a shortcut for not having to parse out the scheme, host, and port
+    of the url before creating an :class:`.ConnectionPool` instance.
+
+    :param url:
+        Absolute URL string that must include the scheme. Port is optional.
+
+    :param \\**kw:
+        Passes additional parameters to the constructor of the appropriate
+        :class:`.ConnectionPool`. Useful for specifying things like
+        timeout, maxsize, headers, etc.
+
+    Example::
+
+        >>> conn = connection_from_url('http://google.com/')
+        >>> r = conn.request('GET', '/')
+    """
+    scheme, host, port = get_host(url)
+    port = port or port_by_scheme.get(scheme, 80)
+    if scheme == "https":
+        return HTTPSConnectionPool(host, port=port, **kw)
+    else:
+        return HTTPConnectionPool(host, port=port, **kw)
+
+
+def _normalize_host(host, scheme):
+    """
+    Normalize hosts for comparisons and use with sockets.
+    """
+
+    host = normalize_host(host, scheme)
+
+    # httplib doesn't like it when we include brackets in IPv6 addresses
+    # Specifically, if we include brackets but also pass the port then
+    # httplib crazily doubles up the square brackets on the Host header.
+    # Instead, we need to make sure we never pass ``None`` as the port.
+    # However, for backward compatibility reasons we can't actually
+    # *assert* that.  See http://bugs.python.org/issue28539
+    if host.startswith("[") and host.endswith("]"):
+        host = host[1:-1]
+    return host
+
+
+def _close_pool_connections(pool):
+    """Drains a queue of connections and closes each one."""
+    try:
+        while True:
+            conn = pool.get(block=False)
+            if conn:
+                conn.close()
+    except queue.Empty:
+        pass  # Done.
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..41ffe605
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/_appengine_environ.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/_appengine_environ.cpython-312.pyc
new file mode 100644
index 00000000..3f0714b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/_appengine_environ.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/appengine.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/appengine.cpython-312.pyc
new file mode 100644
index 00000000..bbca6f28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/appengine.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/ntlmpool.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/ntlmpool.cpython-312.pyc
new file mode 100644
index 00000000..cbe4b21f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/ntlmpool.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc
new file mode 100644
index 00000000..ee7d6853
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/securetransport.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/securetransport.cpython-312.pyc
new file mode 100644
index 00000000..057f0b6e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/securetransport.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/socks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/socks.cpython-312.pyc
new file mode 100644
index 00000000..65ae7edc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/__pycache__/socks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_appengine_environ.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_appengine_environ.py
new file mode 100644
index 00000000..8765b907
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_appengine_environ.py
@@ -0,0 +1,36 @@
+"""
+This module provides means to detect the App Engine environment.
+"""
+
+import os
+
+
+def is_appengine():
+    return is_local_appengine() or is_prod_appengine()
+
+
+def is_appengine_sandbox():
+    """Reports if the app is running in the first generation sandbox.
+
+    The second generation runtimes are technically still in a sandbox, but it
+    is much less restrictive, so generally you shouldn't need to check for it.
+    see https://cloud.google.com/appengine/docs/standard/runtimes
+    """
+    return is_appengine() and os.environ["APPENGINE_RUNTIME"] == "python27"
+
+
+def is_local_appengine():
+    return "APPENGINE_RUNTIME" in os.environ and os.environ.get(
+        "SERVER_SOFTWARE", ""
+    ).startswith("Development/")
+
+
+def is_prod_appengine():
+    return "APPENGINE_RUNTIME" in os.environ and os.environ.get(
+        "SERVER_SOFTWARE", ""
+    ).startswith("Google App Engine/")
+
+
+def is_prod_appengine_mvms():
+    """Deprecated."""
+    return False
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d9f1d77b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/bindings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/bindings.cpython-312.pyc
new file mode 100644
index 00000000..e183158c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/bindings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/low_level.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/low_level.cpython-312.pyc
new file mode 100644
index 00000000..63c24080
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/__pycache__/low_level.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/bindings.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/bindings.py
new file mode 100644
index 00000000..264d564d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/bindings.py
@@ -0,0 +1,519 @@
+"""
+This module uses ctypes to bind a whole bunch of functions and constants from
+SecureTransport. The goal here is to provide the low-level API to
+SecureTransport. These are essentially the C-level functions and constants, and
+they're pretty gross to work with.
+
+This code is a bastardised version of the code found in Will Bond's oscrypto
+library. An enormous debt is owed to him for blazing this trail for us. For
+that reason, this code should be considered to be covered both by urllib3's
+license and by oscrypto's:
+
+    Copyright (c) 2015-2016 Will Bond <will@wbond.net>
+
+    Permission is hereby granted, free of charge, to any person obtaining a
+    copy of this software and associated documentation files (the "Software"),
+    to deal in the Software without restriction, including without limitation
+    the rights to use, copy, modify, merge, publish, distribute, sublicense,
+    and/or sell copies of the Software, and to permit persons to whom the
+    Software is furnished to do so, subject to the following conditions:
+
+    The above copyright notice and this permission notice shall be included in
+    all copies or substantial portions of the Software.
+
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+    FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+    DEALINGS IN THE SOFTWARE.
+"""
+from __future__ import absolute_import
+
+import platform
+from ctypes import (
+    CDLL,
+    CFUNCTYPE,
+    POINTER,
+    c_bool,
+    c_byte,
+    c_char_p,
+    c_int32,
+    c_long,
+    c_size_t,
+    c_uint32,
+    c_ulong,
+    c_void_p,
+)
+from ctypes.util import find_library
+
+from ...packages.six import raise_from
+
+if platform.system() != "Darwin":
+    raise ImportError("Only macOS is supported")
+
+version = platform.mac_ver()[0]
+version_info = tuple(map(int, version.split(".")))
+if version_info < (10, 8):
+    raise OSError(
+        "Only OS X 10.8 and newer are supported, not %s.%s"
+        % (version_info[0], version_info[1])
+    )
+
+
+def load_cdll(name, macos10_16_path):
+    """Loads a CDLL by name, falling back to known path on 10.16+"""
+    try:
+        # Big Sur is technically 11 but we use 10.16 due to the Big Sur
+        # beta being labeled as 10.16.
+        if version_info >= (10, 16):
+            path = macos10_16_path
+        else:
+            path = find_library(name)
+        if not path:
+            raise OSError  # Caught and reraised as 'ImportError'
+        return CDLL(path, use_errno=True)
+    except OSError:
+        raise_from(ImportError("The library %s failed to load" % name), None)
+
+
+Security = load_cdll(
+    "Security", "/System/Library/Frameworks/Security.framework/Security"
+)
+CoreFoundation = load_cdll(
+    "CoreFoundation",
+    "/System/Library/Frameworks/CoreFoundation.framework/CoreFoundation",
+)
+
+
+Boolean = c_bool
+CFIndex = c_long
+CFStringEncoding = c_uint32
+CFData = c_void_p
+CFString = c_void_p
+CFArray = c_void_p
+CFMutableArray = c_void_p
+CFDictionary = c_void_p
+CFError = c_void_p
+CFType = c_void_p
+CFTypeID = c_ulong
+
+CFTypeRef = POINTER(CFType)
+CFAllocatorRef = c_void_p
+
+OSStatus = c_int32
+
+CFDataRef = POINTER(CFData)
+CFStringRef = POINTER(CFString)
+CFArrayRef = POINTER(CFArray)
+CFMutableArrayRef = POINTER(CFMutableArray)
+CFDictionaryRef = POINTER(CFDictionary)
+CFArrayCallBacks = c_void_p
+CFDictionaryKeyCallBacks = c_void_p
+CFDictionaryValueCallBacks = c_void_p
+
+SecCertificateRef = POINTER(c_void_p)
+SecExternalFormat = c_uint32
+SecExternalItemType = c_uint32
+SecIdentityRef = POINTER(c_void_p)
+SecItemImportExportFlags = c_uint32
+SecItemImportExportKeyParameters = c_void_p
+SecKeychainRef = POINTER(c_void_p)
+SSLProtocol = c_uint32
+SSLCipherSuite = c_uint32
+SSLContextRef = POINTER(c_void_p)
+SecTrustRef = POINTER(c_void_p)
+SSLConnectionRef = c_uint32
+SecTrustResultType = c_uint32
+SecTrustOptionFlags = c_uint32
+SSLProtocolSide = c_uint32
+SSLConnectionType = c_uint32
+SSLSessionOption = c_uint32
+
+
+try:
+    Security.SecItemImport.argtypes = [
+        CFDataRef,
+        CFStringRef,
+        POINTER(SecExternalFormat),
+        POINTER(SecExternalItemType),
+        SecItemImportExportFlags,
+        POINTER(SecItemImportExportKeyParameters),
+        SecKeychainRef,
+        POINTER(CFArrayRef),
+    ]
+    Security.SecItemImport.restype = OSStatus
+
+    Security.SecCertificateGetTypeID.argtypes = []
+    Security.SecCertificateGetTypeID.restype = CFTypeID
+
+    Security.SecIdentityGetTypeID.argtypes = []
+    Security.SecIdentityGetTypeID.restype = CFTypeID
+
+    Security.SecKeyGetTypeID.argtypes = []
+    Security.SecKeyGetTypeID.restype = CFTypeID
+
+    Security.SecCertificateCreateWithData.argtypes = [CFAllocatorRef, CFDataRef]
+    Security.SecCertificateCreateWithData.restype = SecCertificateRef
+
+    Security.SecCertificateCopyData.argtypes = [SecCertificateRef]
+    Security.SecCertificateCopyData.restype = CFDataRef
+
+    Security.SecCopyErrorMessageString.argtypes = [OSStatus, c_void_p]
+    Security.SecCopyErrorMessageString.restype = CFStringRef
+
+    Security.SecIdentityCreateWithCertificate.argtypes = [
+        CFTypeRef,
+        SecCertificateRef,
+        POINTER(SecIdentityRef),
+    ]
+    Security.SecIdentityCreateWithCertificate.restype = OSStatus
+
+    Security.SecKeychainCreate.argtypes = [
+        c_char_p,
+        c_uint32,
+        c_void_p,
+        Boolean,
+        c_void_p,
+        POINTER(SecKeychainRef),
+    ]
+    Security.SecKeychainCreate.restype = OSStatus
+
+    Security.SecKeychainDelete.argtypes = [SecKeychainRef]
+    Security.SecKeychainDelete.restype = OSStatus
+
+    Security.SecPKCS12Import.argtypes = [
+        CFDataRef,
+        CFDictionaryRef,
+        POINTER(CFArrayRef),
+    ]
+    Security.SecPKCS12Import.restype = OSStatus
+
+    SSLReadFunc = CFUNCTYPE(OSStatus, SSLConnectionRef, c_void_p, POINTER(c_size_t))
+    SSLWriteFunc = CFUNCTYPE(
+        OSStatus, SSLConnectionRef, POINTER(c_byte), POINTER(c_size_t)
+    )
+
+    Security.SSLSetIOFuncs.argtypes = [SSLContextRef, SSLReadFunc, SSLWriteFunc]
+    Security.SSLSetIOFuncs.restype = OSStatus
+
+    Security.SSLSetPeerID.argtypes = [SSLContextRef, c_char_p, c_size_t]
+    Security.SSLSetPeerID.restype = OSStatus
+
+    Security.SSLSetCertificate.argtypes = [SSLContextRef, CFArrayRef]
+    Security.SSLSetCertificate.restype = OSStatus
+
+    Security.SSLSetCertificateAuthorities.argtypes = [SSLContextRef, CFTypeRef, Boolean]
+    Security.SSLSetCertificateAuthorities.restype = OSStatus
+
+    Security.SSLSetConnection.argtypes = [SSLContextRef, SSLConnectionRef]
+    Security.SSLSetConnection.restype = OSStatus
+
+    Security.SSLSetPeerDomainName.argtypes = [SSLContextRef, c_char_p, c_size_t]
+    Security.SSLSetPeerDomainName.restype = OSStatus
+
+    Security.SSLHandshake.argtypes = [SSLContextRef]
+    Security.SSLHandshake.restype = OSStatus
+
+    Security.SSLRead.argtypes = [SSLContextRef, c_char_p, c_size_t, POINTER(c_size_t)]
+    Security.SSLRead.restype = OSStatus
+
+    Security.SSLWrite.argtypes = [SSLContextRef, c_char_p, c_size_t, POINTER(c_size_t)]
+    Security.SSLWrite.restype = OSStatus
+
+    Security.SSLClose.argtypes = [SSLContextRef]
+    Security.SSLClose.restype = OSStatus
+
+    Security.SSLGetNumberSupportedCiphers.argtypes = [SSLContextRef, POINTER(c_size_t)]
+    Security.SSLGetNumberSupportedCiphers.restype = OSStatus
+
+    Security.SSLGetSupportedCiphers.argtypes = [
+        SSLContextRef,
+        POINTER(SSLCipherSuite),
+        POINTER(c_size_t),
+    ]
+    Security.SSLGetSupportedCiphers.restype = OSStatus
+
+    Security.SSLSetEnabledCiphers.argtypes = [
+        SSLContextRef,
+        POINTER(SSLCipherSuite),
+        c_size_t,
+    ]
+    Security.SSLSetEnabledCiphers.restype = OSStatus
+
+    Security.SSLGetNumberEnabledCiphers.argtype = [SSLContextRef, POINTER(c_size_t)]
+    Security.SSLGetNumberEnabledCiphers.restype = OSStatus
+
+    Security.SSLGetEnabledCiphers.argtypes = [
+        SSLContextRef,
+        POINTER(SSLCipherSuite),
+        POINTER(c_size_t),
+    ]
+    Security.SSLGetEnabledCiphers.restype = OSStatus
+
+    Security.SSLGetNegotiatedCipher.argtypes = [SSLContextRef, POINTER(SSLCipherSuite)]
+    Security.SSLGetNegotiatedCipher.restype = OSStatus
+
+    Security.SSLGetNegotiatedProtocolVersion.argtypes = [
+        SSLContextRef,
+        POINTER(SSLProtocol),
+    ]
+    Security.SSLGetNegotiatedProtocolVersion.restype = OSStatus
+
+    Security.SSLCopyPeerTrust.argtypes = [SSLContextRef, POINTER(SecTrustRef)]
+    Security.SSLCopyPeerTrust.restype = OSStatus
+
+    Security.SecTrustSetAnchorCertificates.argtypes = [SecTrustRef, CFArrayRef]
+    Security.SecTrustSetAnchorCertificates.restype = OSStatus
+
+    Security.SecTrustSetAnchorCertificatesOnly.argstypes = [SecTrustRef, Boolean]
+    Security.SecTrustSetAnchorCertificatesOnly.restype = OSStatus
+
+    Security.SecTrustEvaluate.argtypes = [SecTrustRef, POINTER(SecTrustResultType)]
+    Security.SecTrustEvaluate.restype = OSStatus
+
+    Security.SecTrustGetCertificateCount.argtypes = [SecTrustRef]
+    Security.SecTrustGetCertificateCount.restype = CFIndex
+
+    Security.SecTrustGetCertificateAtIndex.argtypes = [SecTrustRef, CFIndex]
+    Security.SecTrustGetCertificateAtIndex.restype = SecCertificateRef
+
+    Security.SSLCreateContext.argtypes = [
+        CFAllocatorRef,
+        SSLProtocolSide,
+        SSLConnectionType,
+    ]
+    Security.SSLCreateContext.restype = SSLContextRef
+
+    Security.SSLSetSessionOption.argtypes = [SSLContextRef, SSLSessionOption, Boolean]
+    Security.SSLSetSessionOption.restype = OSStatus
+
+    Security.SSLSetProtocolVersionMin.argtypes = [SSLContextRef, SSLProtocol]
+    Security.SSLSetProtocolVersionMin.restype = OSStatus
+
+    Security.SSLSetProtocolVersionMax.argtypes = [SSLContextRef, SSLProtocol]
+    Security.SSLSetProtocolVersionMax.restype = OSStatus
+
+    try:
+        Security.SSLSetALPNProtocols.argtypes = [SSLContextRef, CFArrayRef]
+        Security.SSLSetALPNProtocols.restype = OSStatus
+    except AttributeError:
+        # Supported only in 10.12+
+        pass
+
+    Security.SecCopyErrorMessageString.argtypes = [OSStatus, c_void_p]
+    Security.SecCopyErrorMessageString.restype = CFStringRef
+
+    Security.SSLReadFunc = SSLReadFunc
+    Security.SSLWriteFunc = SSLWriteFunc
+    Security.SSLContextRef = SSLContextRef
+    Security.SSLProtocol = SSLProtocol
+    Security.SSLCipherSuite = SSLCipherSuite
+    Security.SecIdentityRef = SecIdentityRef
+    Security.SecKeychainRef = SecKeychainRef
+    Security.SecTrustRef = SecTrustRef
+    Security.SecTrustResultType = SecTrustResultType
+    Security.SecExternalFormat = SecExternalFormat
+    Security.OSStatus = OSStatus
+
+    Security.kSecImportExportPassphrase = CFStringRef.in_dll(
+        Security, "kSecImportExportPassphrase"
+    )
+    Security.kSecImportItemIdentity = CFStringRef.in_dll(
+        Security, "kSecImportItemIdentity"
+    )
+
+    # CoreFoundation time!
+    CoreFoundation.CFRetain.argtypes = [CFTypeRef]
+    CoreFoundation.CFRetain.restype = CFTypeRef
+
+    CoreFoundation.CFRelease.argtypes = [CFTypeRef]
+    CoreFoundation.CFRelease.restype = None
+
+    CoreFoundation.CFGetTypeID.argtypes = [CFTypeRef]
+    CoreFoundation.CFGetTypeID.restype = CFTypeID
+
+    CoreFoundation.CFStringCreateWithCString.argtypes = [
+        CFAllocatorRef,
+        c_char_p,
+        CFStringEncoding,
+    ]
+    CoreFoundation.CFStringCreateWithCString.restype = CFStringRef
+
+    CoreFoundation.CFStringGetCStringPtr.argtypes = [CFStringRef, CFStringEncoding]
+    CoreFoundation.CFStringGetCStringPtr.restype = c_char_p
+
+    CoreFoundation.CFStringGetCString.argtypes = [
+        CFStringRef,
+        c_char_p,
+        CFIndex,
+        CFStringEncoding,
+    ]
+    CoreFoundation.CFStringGetCString.restype = c_bool
+
+    CoreFoundation.CFDataCreate.argtypes = [CFAllocatorRef, c_char_p, CFIndex]
+    CoreFoundation.CFDataCreate.restype = CFDataRef
+
+    CoreFoundation.CFDataGetLength.argtypes = [CFDataRef]
+    CoreFoundation.CFDataGetLength.restype = CFIndex
+
+    CoreFoundation.CFDataGetBytePtr.argtypes = [CFDataRef]
+    CoreFoundation.CFDataGetBytePtr.restype = c_void_p
+
+    CoreFoundation.CFDictionaryCreate.argtypes = [
+        CFAllocatorRef,
+        POINTER(CFTypeRef),
+        POINTER(CFTypeRef),
+        CFIndex,
+        CFDictionaryKeyCallBacks,
+        CFDictionaryValueCallBacks,
+    ]
+    CoreFoundation.CFDictionaryCreate.restype = CFDictionaryRef
+
+    CoreFoundation.CFDictionaryGetValue.argtypes = [CFDictionaryRef, CFTypeRef]
+    CoreFoundation.CFDictionaryGetValue.restype = CFTypeRef
+
+    CoreFoundation.CFArrayCreate.argtypes = [
+        CFAllocatorRef,
+        POINTER(CFTypeRef),
+        CFIndex,
+        CFArrayCallBacks,
+    ]
+    CoreFoundation.CFArrayCreate.restype = CFArrayRef
+
+    CoreFoundation.CFArrayCreateMutable.argtypes = [
+        CFAllocatorRef,
+        CFIndex,
+        CFArrayCallBacks,
+    ]
+    CoreFoundation.CFArrayCreateMutable.restype = CFMutableArrayRef
+
+    CoreFoundation.CFArrayAppendValue.argtypes = [CFMutableArrayRef, c_void_p]
+    CoreFoundation.CFArrayAppendValue.restype = None
+
+    CoreFoundation.CFArrayGetCount.argtypes = [CFArrayRef]
+    CoreFoundation.CFArrayGetCount.restype = CFIndex
+
+    CoreFoundation.CFArrayGetValueAtIndex.argtypes = [CFArrayRef, CFIndex]
+    CoreFoundation.CFArrayGetValueAtIndex.restype = c_void_p
+
+    CoreFoundation.kCFAllocatorDefault = CFAllocatorRef.in_dll(
+        CoreFoundation, "kCFAllocatorDefault"
+    )
+    CoreFoundation.kCFTypeArrayCallBacks = c_void_p.in_dll(
+        CoreFoundation, "kCFTypeArrayCallBacks"
+    )
+    CoreFoundation.kCFTypeDictionaryKeyCallBacks = c_void_p.in_dll(
+        CoreFoundation, "kCFTypeDictionaryKeyCallBacks"
+    )
+    CoreFoundation.kCFTypeDictionaryValueCallBacks = c_void_p.in_dll(
+        CoreFoundation, "kCFTypeDictionaryValueCallBacks"
+    )
+
+    CoreFoundation.CFTypeRef = CFTypeRef
+    CoreFoundation.CFArrayRef = CFArrayRef
+    CoreFoundation.CFStringRef = CFStringRef
+    CoreFoundation.CFDictionaryRef = CFDictionaryRef
+
+except (AttributeError):
+    raise ImportError("Error initializing ctypes")
+
+
+class CFConst(object):
+    """
+    A class object that acts as essentially a namespace for CoreFoundation
+    constants.
+    """
+
+    kCFStringEncodingUTF8 = CFStringEncoding(0x08000100)
+
+
+class SecurityConst(object):
+    """
+    A class object that acts as essentially a namespace for Security constants.
+    """
+
+    kSSLSessionOptionBreakOnServerAuth = 0
+
+    kSSLProtocol2 = 1
+    kSSLProtocol3 = 2
+    kTLSProtocol1 = 4
+    kTLSProtocol11 = 7
+    kTLSProtocol12 = 8
+    # SecureTransport does not support TLS 1.3 even if there's a constant for it
+    kTLSProtocol13 = 10
+    kTLSProtocolMaxSupported = 999
+
+    kSSLClientSide = 1
+    kSSLStreamType = 0
+
+    kSecFormatPEMSequence = 10
+
+    kSecTrustResultInvalid = 0
+    kSecTrustResultProceed = 1
+    # This gap is present on purpose: this was kSecTrustResultConfirm, which
+    # is deprecated.
+    kSecTrustResultDeny = 3
+    kSecTrustResultUnspecified = 4
+    kSecTrustResultRecoverableTrustFailure = 5
+    kSecTrustResultFatalTrustFailure = 6
+    kSecTrustResultOtherError = 7
+
+    errSSLProtocol = -9800
+    errSSLWouldBlock = -9803
+    errSSLClosedGraceful = -9805
+    errSSLClosedNoNotify = -9816
+    errSSLClosedAbort = -9806
+
+    errSSLXCertChainInvalid = -9807
+    errSSLCrypto = -9809
+    errSSLInternal = -9810
+    errSSLCertExpired = -9814
+    errSSLCertNotYetValid = -9815
+    errSSLUnknownRootCert = -9812
+    errSSLNoRootCert = -9813
+    errSSLHostNameMismatch = -9843
+    errSSLPeerHandshakeFail = -9824
+    errSSLPeerUserCancelled = -9839
+    errSSLWeakPeerEphemeralDHKey = -9850
+    errSSLServerAuthCompleted = -9841
+    errSSLRecordOverflow = -9847
+
+    errSecVerifyFailed = -67808
+    errSecNoTrustSettings = -25263
+    errSecItemNotFound = -25300
+    errSecInvalidTrustSettings = -25262
+
+    # Cipher suites. We only pick the ones our default cipher string allows.
+    # Source: https://developer.apple.com/documentation/security/1550981-ssl_cipher_suite_values
+    TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384 = 0xC02C
+    TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 = 0xC030
+    TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256 = 0xC02B
+    TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 = 0xC02F
+    TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256 = 0xCCA9
+    TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256 = 0xCCA8
+    TLS_DHE_RSA_WITH_AES_256_GCM_SHA384 = 0x009F
+    TLS_DHE_RSA_WITH_AES_128_GCM_SHA256 = 0x009E
+    TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384 = 0xC024
+    TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384 = 0xC028
+    TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA = 0xC00A
+    TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA = 0xC014
+    TLS_DHE_RSA_WITH_AES_256_CBC_SHA256 = 0x006B
+    TLS_DHE_RSA_WITH_AES_256_CBC_SHA = 0x0039
+    TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256 = 0xC023
+    TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256 = 0xC027
+    TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA = 0xC009
+    TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA = 0xC013
+    TLS_DHE_RSA_WITH_AES_128_CBC_SHA256 = 0x0067
+    TLS_DHE_RSA_WITH_AES_128_CBC_SHA = 0x0033
+    TLS_RSA_WITH_AES_256_GCM_SHA384 = 0x009D
+    TLS_RSA_WITH_AES_128_GCM_SHA256 = 0x009C
+    TLS_RSA_WITH_AES_256_CBC_SHA256 = 0x003D
+    TLS_RSA_WITH_AES_128_CBC_SHA256 = 0x003C
+    TLS_RSA_WITH_AES_256_CBC_SHA = 0x0035
+    TLS_RSA_WITH_AES_128_CBC_SHA = 0x002F
+    TLS_AES_128_GCM_SHA256 = 0x1301
+    TLS_AES_256_GCM_SHA384 = 0x1302
+    TLS_AES_128_CCM_8_SHA256 = 0x1305
+    TLS_AES_128_CCM_SHA256 = 0x1304
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/low_level.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/low_level.py
new file mode 100644
index 00000000..fa0b245d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/_securetransport/low_level.py
@@ -0,0 +1,397 @@
+"""
+Low-level helpers for the SecureTransport bindings.
+
+These are Python functions that are not directly related to the high-level APIs
+but are necessary to get them to work. They include a whole bunch of low-level
+CoreFoundation messing about and memory management. The concerns in this module
+are almost entirely about trying to avoid memory leaks and providing
+appropriate and useful assistance to the higher-level code.
+"""
+import base64
+import ctypes
+import itertools
+import os
+import re
+import ssl
+import struct
+import tempfile
+
+from .bindings import CFConst, CoreFoundation, Security
+
+# This regular expression is used to grab PEM data out of a PEM bundle.
+_PEM_CERTS_RE = re.compile(
+    b"-----BEGIN CERTIFICATE-----\n(.*?)\n-----END CERTIFICATE-----", re.DOTALL
+)
+
+
+def _cf_data_from_bytes(bytestring):
+    """
+    Given a bytestring, create a CFData object from it. This CFData object must
+    be CFReleased by the caller.
+    """
+    return CoreFoundation.CFDataCreate(
+        CoreFoundation.kCFAllocatorDefault, bytestring, len(bytestring)
+    )
+
+
+def _cf_dictionary_from_tuples(tuples):
+    """
+    Given a list of Python tuples, create an associated CFDictionary.
+    """
+    dictionary_size = len(tuples)
+
+    # We need to get the dictionary keys and values out in the same order.
+    keys = (t[0] for t in tuples)
+    values = (t[1] for t in tuples)
+    cf_keys = (CoreFoundation.CFTypeRef * dictionary_size)(*keys)
+    cf_values = (CoreFoundation.CFTypeRef * dictionary_size)(*values)
+
+    return CoreFoundation.CFDictionaryCreate(
+        CoreFoundation.kCFAllocatorDefault,
+        cf_keys,
+        cf_values,
+        dictionary_size,
+        CoreFoundation.kCFTypeDictionaryKeyCallBacks,
+        CoreFoundation.kCFTypeDictionaryValueCallBacks,
+    )
+
+
+def _cfstr(py_bstr):
+    """
+    Given a Python binary data, create a CFString.
+    The string must be CFReleased by the caller.
+    """
+    c_str = ctypes.c_char_p(py_bstr)
+    cf_str = CoreFoundation.CFStringCreateWithCString(
+        CoreFoundation.kCFAllocatorDefault,
+        c_str,
+        CFConst.kCFStringEncodingUTF8,
+    )
+    return cf_str
+
+
+def _create_cfstring_array(lst):
+    """
+    Given a list of Python binary data, create an associated CFMutableArray.
+    The array must be CFReleased by the caller.
+
+    Raises an ssl.SSLError on failure.
+    """
+    cf_arr = None
+    try:
+        cf_arr = CoreFoundation.CFArrayCreateMutable(
+            CoreFoundation.kCFAllocatorDefault,
+            0,
+            ctypes.byref(CoreFoundation.kCFTypeArrayCallBacks),
+        )
+        if not cf_arr:
+            raise MemoryError("Unable to allocate memory!")
+        for item in lst:
+            cf_str = _cfstr(item)
+            if not cf_str:
+                raise MemoryError("Unable to allocate memory!")
+            try:
+                CoreFoundation.CFArrayAppendValue(cf_arr, cf_str)
+            finally:
+                CoreFoundation.CFRelease(cf_str)
+    except BaseException as e:
+        if cf_arr:
+            CoreFoundation.CFRelease(cf_arr)
+        raise ssl.SSLError("Unable to allocate array: %s" % (e,))
+    return cf_arr
+
+
+def _cf_string_to_unicode(value):
+    """
+    Creates a Unicode string from a CFString object. Used entirely for error
+    reporting.
+
+    Yes, it annoys me quite a lot that this function is this complex.
+    """
+    value_as_void_p = ctypes.cast(value, ctypes.POINTER(ctypes.c_void_p))
+
+    string = CoreFoundation.CFStringGetCStringPtr(
+        value_as_void_p, CFConst.kCFStringEncodingUTF8
+    )
+    if string is None:
+        buffer = ctypes.create_string_buffer(1024)
+        result = CoreFoundation.CFStringGetCString(
+            value_as_void_p, buffer, 1024, CFConst.kCFStringEncodingUTF8
+        )
+        if not result:
+            raise OSError("Error copying C string from CFStringRef")
+        string = buffer.value
+    if string is not None:
+        string = string.decode("utf-8")
+    return string
+
+
+def _assert_no_error(error, exception_class=None):
+    """
+    Checks the return code and throws an exception if there is an error to
+    report
+    """
+    if error == 0:
+        return
+
+    cf_error_string = Security.SecCopyErrorMessageString(error, None)
+    output = _cf_string_to_unicode(cf_error_string)
+    CoreFoundation.CFRelease(cf_error_string)
+
+    if output is None or output == u"":
+        output = u"OSStatus %s" % error
+
+    if exception_class is None:
+        exception_class = ssl.SSLError
+
+    raise exception_class(output)
+
+
+def _cert_array_from_pem(pem_bundle):
+    """
+    Given a bundle of certs in PEM format, turns them into a CFArray of certs
+    that can be used to validate a cert chain.
+    """
+    # Normalize the PEM bundle's line endings.
+    pem_bundle = pem_bundle.replace(b"\r\n", b"\n")
+
+    der_certs = [
+        base64.b64decode(match.group(1)) for match in _PEM_CERTS_RE.finditer(pem_bundle)
+    ]
+    if not der_certs:
+        raise ssl.SSLError("No root certificates specified")
+
+    cert_array = CoreFoundation.CFArrayCreateMutable(
+        CoreFoundation.kCFAllocatorDefault,
+        0,
+        ctypes.byref(CoreFoundation.kCFTypeArrayCallBacks),
+    )
+    if not cert_array:
+        raise ssl.SSLError("Unable to allocate memory!")
+
+    try:
+        for der_bytes in der_certs:
+            certdata = _cf_data_from_bytes(der_bytes)
+            if not certdata:
+                raise ssl.SSLError("Unable to allocate memory!")
+            cert = Security.SecCertificateCreateWithData(
+                CoreFoundation.kCFAllocatorDefault, certdata
+            )
+            CoreFoundation.CFRelease(certdata)
+            if not cert:
+                raise ssl.SSLError("Unable to build cert object!")
+
+            CoreFoundation.CFArrayAppendValue(cert_array, cert)
+            CoreFoundation.CFRelease(cert)
+    except Exception:
+        # We need to free the array before the exception bubbles further.
+        # We only want to do that if an error occurs: otherwise, the caller
+        # should free.
+        CoreFoundation.CFRelease(cert_array)
+        raise
+
+    return cert_array
+
+
+def _is_cert(item):
+    """
+    Returns True if a given CFTypeRef is a certificate.
+    """
+    expected = Security.SecCertificateGetTypeID()
+    return CoreFoundation.CFGetTypeID(item) == expected
+
+
+def _is_identity(item):
+    """
+    Returns True if a given CFTypeRef is an identity.
+    """
+    expected = Security.SecIdentityGetTypeID()
+    return CoreFoundation.CFGetTypeID(item) == expected
+
+
+def _temporary_keychain():
+    """
+    This function creates a temporary Mac keychain that we can use to work with
+    credentials. This keychain uses a one-time password and a temporary file to
+    store the data. We expect to have one keychain per socket. The returned
+    SecKeychainRef must be freed by the caller, including calling
+    SecKeychainDelete.
+
+    Returns a tuple of the SecKeychainRef and the path to the temporary
+    directory that contains it.
+    """
+    # Unfortunately, SecKeychainCreate requires a path to a keychain. This
+    # means we cannot use mkstemp to use a generic temporary file. Instead,
+    # we're going to create a temporary directory and a filename to use there.
+    # This filename will be 8 random bytes expanded into base64. We also need
+    # some random bytes to password-protect the keychain we're creating, so we
+    # ask for 40 random bytes.
+    random_bytes = os.urandom(40)
+    filename = base64.b16encode(random_bytes[:8]).decode("utf-8")
+    password = base64.b16encode(random_bytes[8:])  # Must be valid UTF-8
+    tempdirectory = tempfile.mkdtemp()
+
+    keychain_path = os.path.join(tempdirectory, filename).encode("utf-8")
+
+    # We now want to create the keychain itself.
+    keychain = Security.SecKeychainRef()
+    status = Security.SecKeychainCreate(
+        keychain_path, len(password), password, False, None, ctypes.byref(keychain)
+    )
+    _assert_no_error(status)
+
+    # Having created the keychain, we want to pass it off to the caller.
+    return keychain, tempdirectory
+
+
+def _load_items_from_file(keychain, path):
+    """
+    Given a single file, loads all the trust objects from it into arrays and
+    the keychain.
+    Returns a tuple of lists: the first list is a list of identities, the
+    second a list of certs.
+    """
+    certificates = []
+    identities = []
+    result_array = None
+
+    with open(path, "rb") as f:
+        raw_filedata = f.read()
+
+    try:
+        filedata = CoreFoundation.CFDataCreate(
+            CoreFoundation.kCFAllocatorDefault, raw_filedata, len(raw_filedata)
+        )
+        result_array = CoreFoundation.CFArrayRef()
+        result = Security.SecItemImport(
+            filedata,  # cert data
+            None,  # Filename, leaving it out for now
+            None,  # What the type of the file is, we don't care
+            None,  # what's in the file, we don't care
+            0,  # import flags
+            None,  # key params, can include passphrase in the future
+            keychain,  # The keychain to insert into
+            ctypes.byref(result_array),  # Results
+        )
+        _assert_no_error(result)
+
+        # A CFArray is not very useful to us as an intermediary
+        # representation, so we are going to extract the objects we want
+        # and then free the array. We don't need to keep hold of keys: the
+        # keychain already has them!
+        result_count = CoreFoundation.CFArrayGetCount(result_array)
+        for index in range(result_count):
+            item = CoreFoundation.CFArrayGetValueAtIndex(result_array, index)
+            item = ctypes.cast(item, CoreFoundation.CFTypeRef)
+
+            if _is_cert(item):
+                CoreFoundation.CFRetain(item)
+                certificates.append(item)
+            elif _is_identity(item):
+                CoreFoundation.CFRetain(item)
+                identities.append(item)
+    finally:
+        if result_array:
+            CoreFoundation.CFRelease(result_array)
+
+        CoreFoundation.CFRelease(filedata)
+
+    return (identities, certificates)
+
+
+def _load_client_cert_chain(keychain, *paths):
+    """
+    Load certificates and maybe keys from a number of files. Has the end goal
+    of returning a CFArray containing one SecIdentityRef, and then zero or more
+    SecCertificateRef objects, suitable for use as a client certificate trust
+    chain.
+    """
+    # Ok, the strategy.
+    #
+    # This relies on knowing that macOS will not give you a SecIdentityRef
+    # unless you have imported a key into a keychain. This is a somewhat
+    # artificial limitation of macOS (for example, it doesn't necessarily
+    # affect iOS), but there is nothing inside Security.framework that lets you
+    # get a SecIdentityRef without having a key in a keychain.
+    #
+    # So the policy here is we take all the files and iterate them in order.
+    # Each one will use SecItemImport to have one or more objects loaded from
+    # it. We will also point at a keychain that macOS can use to work with the
+    # private key.
+    #
+    # Once we have all the objects, we'll check what we actually have. If we
+    # already have a SecIdentityRef in hand, fab: we'll use that. Otherwise,
+    # we'll take the first certificate (which we assume to be our leaf) and
+    # ask the keychain to give us a SecIdentityRef with that cert's associated
+    # key.
+    #
+    # We'll then return a CFArray containing the trust chain: one
+    # SecIdentityRef and then zero-or-more SecCertificateRef objects. The
+    # responsibility for freeing this CFArray will be with the caller. This
+    # CFArray must remain alive for the entire connection, so in practice it
+    # will be stored with a single SSLSocket, along with the reference to the
+    # keychain.
+    certificates = []
+    identities = []
+
+    # Filter out bad paths.
+    paths = (path for path in paths if path)
+
+    try:
+        for file_path in paths:
+            new_identities, new_certs = _load_items_from_file(keychain, file_path)
+            identities.extend(new_identities)
+            certificates.extend(new_certs)
+
+        # Ok, we have everything. The question is: do we have an identity? If
+        # not, we want to grab one from the first cert we have.
+        if not identities:
+            new_identity = Security.SecIdentityRef()
+            status = Security.SecIdentityCreateWithCertificate(
+                keychain, certificates[0], ctypes.byref(new_identity)
+            )
+            _assert_no_error(status)
+            identities.append(new_identity)
+
+            # We now want to release the original certificate, as we no longer
+            # need it.
+            CoreFoundation.CFRelease(certificates.pop(0))
+
+        # We now need to build a new CFArray that holds the trust chain.
+        trust_chain = CoreFoundation.CFArrayCreateMutable(
+            CoreFoundation.kCFAllocatorDefault,
+            0,
+            ctypes.byref(CoreFoundation.kCFTypeArrayCallBacks),
+        )
+        for item in itertools.chain(identities, certificates):
+            # ArrayAppendValue does a CFRetain on the item. That's fine,
+            # because the finally block will release our other refs to them.
+            CoreFoundation.CFArrayAppendValue(trust_chain, item)
+
+        return trust_chain
+    finally:
+        for obj in itertools.chain(identities, certificates):
+            CoreFoundation.CFRelease(obj)
+
+
+TLS_PROTOCOL_VERSIONS = {
+    "SSLv2": (0, 2),
+    "SSLv3": (3, 0),
+    "TLSv1": (3, 1),
+    "TLSv1.1": (3, 2),
+    "TLSv1.2": (3, 3),
+}
+
+
+def _build_tls_unknown_ca_alert(version):
+    """
+    Builds a TLS alert record for an unknown CA.
+    """
+    ver_maj, ver_min = TLS_PROTOCOL_VERSIONS[version]
+    severity_fatal = 0x02
+    description_unknown_ca = 0x30
+    msg = struct.pack(">BB", severity_fatal, description_unknown_ca)
+    msg_len = len(msg)
+    record_type_alert = 0x15
+    record = struct.pack(">BBBH", record_type_alert, ver_maj, ver_min, msg_len) + msg
+    return record
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/appengine.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/appengine.py
new file mode 100644
index 00000000..1717ee22
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/appengine.py
@@ -0,0 +1,314 @@
+"""
+This module provides a pool manager that uses Google App Engine's
+`URLFetch Service <https://cloud.google.com/appengine/docs/python/urlfetch>`_.
+
+Example usage::
+
+    from pip._vendor.urllib3 import PoolManager
+    from pip._vendor.urllib3.contrib.appengine import AppEngineManager, is_appengine_sandbox
+
+    if is_appengine_sandbox():
+        # AppEngineManager uses AppEngine's URLFetch API behind the scenes
+        http = AppEngineManager()
+    else:
+        # PoolManager uses a socket-level API behind the scenes
+        http = PoolManager()
+
+    r = http.request('GET', 'https://google.com/')
+
+There are `limitations <https://cloud.google.com/appengine/docs/python/\
+urlfetch/#Python_Quotas_and_limits>`_ to the URLFetch service and it may not be
+the best choice for your application. There are three options for using
+urllib3 on Google App Engine:
+
+1. You can use :class:`AppEngineManager` with URLFetch. URLFetch is
+   cost-effective in many circumstances as long as your usage is within the
+   limitations.
+2. You can use a normal :class:`~urllib3.PoolManager` by enabling sockets.
+   Sockets also have `limitations and restrictions
+   <https://cloud.google.com/appengine/docs/python/sockets/\
+   #limitations-and-restrictions>`_ and have a lower free quota than URLFetch.
+   To use sockets, be sure to specify the following in your ``app.yaml``::
+
+        env_variables:
+            GAE_USE_SOCKETS_HTTPLIB : 'true'
+
+3. If you are using `App Engine Flexible
+<https://cloud.google.com/appengine/docs/flexible/>`_, you can use the standard
+:class:`PoolManager` without any configuration or special environment variables.
+"""
+
+from __future__ import absolute_import
+
+import io
+import logging
+import warnings
+
+from ..exceptions import (
+    HTTPError,
+    HTTPWarning,
+    MaxRetryError,
+    ProtocolError,
+    SSLError,
+    TimeoutError,
+)
+from ..packages.six.moves.urllib.parse import urljoin
+from ..request import RequestMethods
+from ..response import HTTPResponse
+from ..util.retry import Retry
+from ..util.timeout import Timeout
+from . import _appengine_environ
+
+try:
+    from google.appengine.api import urlfetch
+except ImportError:
+    urlfetch = None
+
+
+log = logging.getLogger(__name__)
+
+
+class AppEnginePlatformWarning(HTTPWarning):
+    pass
+
+
+class AppEnginePlatformError(HTTPError):
+    pass
+
+
+class AppEngineManager(RequestMethods):
+    """
+    Connection manager for Google App Engine sandbox applications.
+
+    This manager uses the URLFetch service directly instead of using the
+    emulated httplib, and is subject to URLFetch limitations as described in
+    the App Engine documentation `here
+    <https://cloud.google.com/appengine/docs/python/urlfetch>`_.
+
+    Notably it will raise an :class:`AppEnginePlatformError` if:
+        * URLFetch is not available.
+        * If you attempt to use this on App Engine Flexible, as full socket
+          support is available.
+        * If a request size is more than 10 megabytes.
+        * If a response size is more than 32 megabytes.
+        * If you use an unsupported request method such as OPTIONS.
+
+    Beyond those cases, it will raise normal urllib3 errors.
+    """
+
+    def __init__(
+        self,
+        headers=None,
+        retries=None,
+        validate_certificate=True,
+        urlfetch_retries=True,
+    ):
+        if not urlfetch:
+            raise AppEnginePlatformError(
+                "URLFetch is not available in this environment."
+            )
+
+        warnings.warn(
+            "urllib3 is using URLFetch on Google App Engine sandbox instead "
+            "of sockets. To use sockets directly instead of URLFetch see "
+            "https://urllib3.readthedocs.io/en/1.26.x/reference/urllib3.contrib.html.",
+            AppEnginePlatformWarning,
+        )
+
+        RequestMethods.__init__(self, headers)
+        self.validate_certificate = validate_certificate
+        self.urlfetch_retries = urlfetch_retries
+
+        self.retries = retries or Retry.DEFAULT
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        # Return False to re-raise any potential exceptions
+        return False
+
+    def urlopen(
+        self,
+        method,
+        url,
+        body=None,
+        headers=None,
+        retries=None,
+        redirect=True,
+        timeout=Timeout.DEFAULT_TIMEOUT,
+        **response_kw
+    ):
+
+        retries = self._get_retries(retries, redirect)
+
+        try:
+            follow_redirects = redirect and retries.redirect != 0 and retries.total
+            response = urlfetch.fetch(
+                url,
+                payload=body,
+                method=method,
+                headers=headers or {},
+                allow_truncated=False,
+                follow_redirects=self.urlfetch_retries and follow_redirects,
+                deadline=self._get_absolute_timeout(timeout),
+                validate_certificate=self.validate_certificate,
+            )
+        except urlfetch.DeadlineExceededError as e:
+            raise TimeoutError(self, e)
+
+        except urlfetch.InvalidURLError as e:
+            if "too large" in str(e):
+                raise AppEnginePlatformError(
+                    "URLFetch request too large, URLFetch only "
+                    "supports requests up to 10mb in size.",
+                    e,
+                )
+            raise ProtocolError(e)
+
+        except urlfetch.DownloadError as e:
+            if "Too many redirects" in str(e):
+                raise MaxRetryError(self, url, reason=e)
+            raise ProtocolError(e)
+
+        except urlfetch.ResponseTooLargeError as e:
+            raise AppEnginePlatformError(
+                "URLFetch response too large, URLFetch only supports"
+                "responses up to 32mb in size.",
+                e,
+            )
+
+        except urlfetch.SSLCertificateError as e:
+            raise SSLError(e)
+
+        except urlfetch.InvalidMethodError as e:
+            raise AppEnginePlatformError(
+                "URLFetch does not support method: %s" % method, e
+            )
+
+        http_response = self._urlfetch_response_to_http_response(
+            response, retries=retries, **response_kw
+        )
+
+        # Handle redirect?
+        redirect_location = redirect and http_response.get_redirect_location()
+        if redirect_location:
+            # Check for redirect response
+            if self.urlfetch_retries and retries.raise_on_redirect:
+                raise MaxRetryError(self, url, "too many redirects")
+            else:
+                if http_response.status == 303:
+                    method = "GET"
+
+                try:
+                    retries = retries.increment(
+                        method, url, response=http_response, _pool=self
+                    )
+                except MaxRetryError:
+                    if retries.raise_on_redirect:
+                        raise MaxRetryError(self, url, "too many redirects")
+                    return http_response
+
+                retries.sleep_for_retry(http_response)
+                log.debug("Redirecting %s -> %s", url, redirect_location)
+                redirect_url = urljoin(url, redirect_location)
+                return self.urlopen(
+                    method,
+                    redirect_url,
+                    body,
+                    headers,
+                    retries=retries,
+                    redirect=redirect,
+                    timeout=timeout,
+                    **response_kw
+                )
+
+        # Check if we should retry the HTTP response.
+        has_retry_after = bool(http_response.headers.get("Retry-After"))
+        if retries.is_retry(method, http_response.status, has_retry_after):
+            retries = retries.increment(method, url, response=http_response, _pool=self)
+            log.debug("Retry: %s", url)
+            retries.sleep(http_response)
+            return self.urlopen(
+                method,
+                url,
+                body=body,
+                headers=headers,
+                retries=retries,
+                redirect=redirect,
+                timeout=timeout,
+                **response_kw
+            )
+
+        return http_response
+
+    def _urlfetch_response_to_http_response(self, urlfetch_resp, **response_kw):
+
+        if is_prod_appengine():
+            # Production GAE handles deflate encoding automatically, but does
+            # not remove the encoding header.
+            content_encoding = urlfetch_resp.headers.get("content-encoding")
+
+            if content_encoding == "deflate":
+                del urlfetch_resp.headers["content-encoding"]
+
+        transfer_encoding = urlfetch_resp.headers.get("transfer-encoding")
+        # We have a full response's content,
+        # so let's make sure we don't report ourselves as chunked data.
+        if transfer_encoding == "chunked":
+            encodings = transfer_encoding.split(",")
+            encodings.remove("chunked")
+            urlfetch_resp.headers["transfer-encoding"] = ",".join(encodings)
+
+        original_response = HTTPResponse(
+            # In order for decoding to work, we must present the content as
+            # a file-like object.
+            body=io.BytesIO(urlfetch_resp.content),
+            msg=urlfetch_resp.header_msg,
+            headers=urlfetch_resp.headers,
+            status=urlfetch_resp.status_code,
+            **response_kw
+        )
+
+        return HTTPResponse(
+            body=io.BytesIO(urlfetch_resp.content),
+            headers=urlfetch_resp.headers,
+            status=urlfetch_resp.status_code,
+            original_response=original_response,
+            **response_kw
+        )
+
+    def _get_absolute_timeout(self, timeout):
+        if timeout is Timeout.DEFAULT_TIMEOUT:
+            return None  # Defer to URLFetch's default.
+        if isinstance(timeout, Timeout):
+            if timeout._read is not None or timeout._connect is not None:
+                warnings.warn(
+                    "URLFetch does not support granular timeout settings, "
+                    "reverting to total or default URLFetch timeout.",
+                    AppEnginePlatformWarning,
+                )
+            return timeout.total
+        return timeout
+
+    def _get_retries(self, retries, redirect):
+        if not isinstance(retries, Retry):
+            retries = Retry.from_int(retries, redirect=redirect, default=self.retries)
+
+        if retries.connect or retries.read or retries.redirect:
+            warnings.warn(
+                "URLFetch only supports total retries and does not "
+                "recognize connect, read, or redirect retry parameters.",
+                AppEnginePlatformWarning,
+            )
+
+        return retries
+
+
+# Alias methods from _appengine_environ to maintain public API interface.
+
+is_appengine = _appengine_environ.is_appengine
+is_appengine_sandbox = _appengine_environ.is_appengine_sandbox
+is_local_appengine = _appengine_environ.is_local_appengine
+is_prod_appengine = _appengine_environ.is_prod_appengine
+is_prod_appengine_mvms = _appengine_environ.is_prod_appengine_mvms
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/ntlmpool.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/ntlmpool.py
new file mode 100644
index 00000000..47166575
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/ntlmpool.py
@@ -0,0 +1,130 @@
+"""
+NTLM authenticating pool, contributed by erikcederstran
+
+Issue #10, see: http://code.google.com/p/urllib3/issues/detail?id=10
+"""
+from __future__ import absolute_import
+
+import warnings
+from logging import getLogger
+
+from ntlm import ntlm
+
+from .. import HTTPSConnectionPool
+from ..packages.six.moves.http_client import HTTPSConnection
+
+warnings.warn(
+    "The 'urllib3.contrib.ntlmpool' module is deprecated and will be removed "
+    "in urllib3 v2.0 release, urllib3 is not able to support it properly due "
+    "to reasons listed in issue: https://github.com/urllib3/urllib3/issues/2282. "
+    "If you are a user of this module please comment in the mentioned issue.",
+    DeprecationWarning,
+)
+
+log = getLogger(__name__)
+
+
+class NTLMConnectionPool(HTTPSConnectionPool):
+    """
+    Implements an NTLM authentication version of an urllib3 connection pool
+    """
+
+    scheme = "https"
+
+    def __init__(self, user, pw, authurl, *args, **kwargs):
+        """
+        authurl is a random URL on the server that is protected by NTLM.
+        user is the Windows user, probably in the DOMAIN\\username format.
+        pw is the password for the user.
+        """
+        super(NTLMConnectionPool, self).__init__(*args, **kwargs)
+        self.authurl = authurl
+        self.rawuser = user
+        user_parts = user.split("\\", 1)
+        self.domain = user_parts[0].upper()
+        self.user = user_parts[1]
+        self.pw = pw
+
+    def _new_conn(self):
+        # Performs the NTLM handshake that secures the connection. The socket
+        # must be kept open while requests are performed.
+        self.num_connections += 1
+        log.debug(
+            "Starting NTLM HTTPS connection no. %d: https://%s%s",
+            self.num_connections,
+            self.host,
+            self.authurl,
+        )
+
+        headers = {"Connection": "Keep-Alive"}
+        req_header = "Authorization"
+        resp_header = "www-authenticate"
+
+        conn = HTTPSConnection(host=self.host, port=self.port)
+
+        # Send negotiation message
+        headers[req_header] = "NTLM %s" % ntlm.create_NTLM_NEGOTIATE_MESSAGE(
+            self.rawuser
+        )
+        log.debug("Request headers: %s", headers)
+        conn.request("GET", self.authurl, None, headers)
+        res = conn.getresponse()
+        reshdr = dict(res.headers)
+        log.debug("Response status: %s %s", res.status, res.reason)
+        log.debug("Response headers: %s", reshdr)
+        log.debug("Response data: %s [...]", res.read(100))
+
+        # Remove the reference to the socket, so that it can not be closed by
+        # the response object (we want to keep the socket open)
+        res.fp = None
+
+        # Server should respond with a challenge message
+        auth_header_values = reshdr[resp_header].split(", ")
+        auth_header_value = None
+        for s in auth_header_values:
+            if s[:5] == "NTLM ":
+                auth_header_value = s[5:]
+        if auth_header_value is None:
+            raise Exception(
+                "Unexpected %s response header: %s" % (resp_header, reshdr[resp_header])
+            )
+
+        # Send authentication message
+        ServerChallenge, NegotiateFlags = ntlm.parse_NTLM_CHALLENGE_MESSAGE(
+            auth_header_value
+        )
+        auth_msg = ntlm.create_NTLM_AUTHENTICATE_MESSAGE(
+            ServerChallenge, self.user, self.domain, self.pw, NegotiateFlags
+        )
+        headers[req_header] = "NTLM %s" % auth_msg
+        log.debug("Request headers: %s", headers)
+        conn.request("GET", self.authurl, None, headers)
+        res = conn.getresponse()
+        log.debug("Response status: %s %s", res.status, res.reason)
+        log.debug("Response headers: %s", dict(res.headers))
+        log.debug("Response data: %s [...]", res.read()[:100])
+        if res.status != 200:
+            if res.status == 401:
+                raise Exception("Server rejected request: wrong username or password")
+            raise Exception("Wrong server response: %s %s" % (res.status, res.reason))
+
+        res.fp = None
+        log.debug("Connection established")
+        return conn
+
+    def urlopen(
+        self,
+        method,
+        url,
+        body=None,
+        headers=None,
+        retries=3,
+        redirect=True,
+        assert_same_host=True,
+    ):
+        if headers is None:
+            headers = {}
+        headers["Connection"] = "Keep-Alive"
+        return super(NTLMConnectionPool, self).urlopen(
+            method, url, body, headers, retries, redirect, assert_same_host
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/pyopenssl.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/pyopenssl.py
new file mode 100644
index 00000000..19e4aa97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/pyopenssl.py
@@ -0,0 +1,518 @@
+"""
+TLS with SNI_-support for Python 2. Follow these instructions if you would
+like to verify TLS certificates in Python 2. Note, the default libraries do
+*not* do certificate checking; you need to do additional work to validate
+certificates yourself.
+
+This needs the following packages installed:
+
+* `pyOpenSSL`_ (tested with 16.0.0)
+* `cryptography`_ (minimum 1.3.4, from pyopenssl)
+* `idna`_ (minimum 2.0, from cryptography)
+
+However, pyopenssl depends on cryptography, which depends on idna, so while we
+use all three directly here we end up having relatively few packages required.
+
+You can install them with the following command:
+
+.. code-block:: bash
+
+    $ python -m pip install pyopenssl cryptography idna
+
+To activate certificate checking, call
+:func:`~urllib3.contrib.pyopenssl.inject_into_urllib3` from your Python code
+before you begin making HTTP requests. This can be done in a ``sitecustomize``
+module, or at any other time before your application begins using ``urllib3``,
+like this:
+
+.. code-block:: python
+
+    try:
+        import pip._vendor.urllib3.contrib.pyopenssl as pyopenssl
+        pyopenssl.inject_into_urllib3()
+    except ImportError:
+        pass
+
+Now you can use :mod:`urllib3` as you normally would, and it will support SNI
+when the required modules are installed.
+
+Activating this module also has the positive side effect of disabling SSL/TLS
+compression in Python 2 (see `CRIME attack`_).
+
+.. _sni: https://en.wikipedia.org/wiki/Server_Name_Indication
+.. _crime attack: https://en.wikipedia.org/wiki/CRIME_(security_exploit)
+.. _pyopenssl: https://www.pyopenssl.org
+.. _cryptography: https://cryptography.io
+.. _idna: https://github.com/kjd/idna
+"""
+from __future__ import absolute_import
+
+import OpenSSL.crypto
+import OpenSSL.SSL
+from cryptography import x509
+from cryptography.hazmat.backends.openssl import backend as openssl_backend
+
+try:
+    from cryptography.x509 import UnsupportedExtension
+except ImportError:
+    # UnsupportedExtension is gone in cryptography >= 2.1.0
+    class UnsupportedExtension(Exception):
+        pass
+
+
+from io import BytesIO
+from socket import error as SocketError
+from socket import timeout
+
+try:  # Platform-specific: Python 2
+    from socket import _fileobject
+except ImportError:  # Platform-specific: Python 3
+    _fileobject = None
+    from ..packages.backports.makefile import backport_makefile
+
+import logging
+import ssl
+import sys
+import warnings
+
+from .. import util
+from ..packages import six
+from ..util.ssl_ import PROTOCOL_TLS_CLIENT
+
+warnings.warn(
+    "'urllib3.contrib.pyopenssl' module is deprecated and will be removed "
+    "in a future release of urllib3 2.x. Read more in this issue: "
+    "https://github.com/urllib3/urllib3/issues/2680",
+    category=DeprecationWarning,
+    stacklevel=2,
+)
+
+__all__ = ["inject_into_urllib3", "extract_from_urllib3"]
+
+# SNI always works.
+HAS_SNI = True
+
+# Map from urllib3 to PyOpenSSL compatible parameter-values.
+_openssl_versions = {
+    util.PROTOCOL_TLS: OpenSSL.SSL.SSLv23_METHOD,
+    PROTOCOL_TLS_CLIENT: OpenSSL.SSL.SSLv23_METHOD,
+    ssl.PROTOCOL_TLSv1: OpenSSL.SSL.TLSv1_METHOD,
+}
+
+if hasattr(ssl, "PROTOCOL_SSLv3") and hasattr(OpenSSL.SSL, "SSLv3_METHOD"):
+    _openssl_versions[ssl.PROTOCOL_SSLv3] = OpenSSL.SSL.SSLv3_METHOD
+
+if hasattr(ssl, "PROTOCOL_TLSv1_1") and hasattr(OpenSSL.SSL, "TLSv1_1_METHOD"):
+    _openssl_versions[ssl.PROTOCOL_TLSv1_1] = OpenSSL.SSL.TLSv1_1_METHOD
+
+if hasattr(ssl, "PROTOCOL_TLSv1_2") and hasattr(OpenSSL.SSL, "TLSv1_2_METHOD"):
+    _openssl_versions[ssl.PROTOCOL_TLSv1_2] = OpenSSL.SSL.TLSv1_2_METHOD
+
+
+_stdlib_to_openssl_verify = {
+    ssl.CERT_NONE: OpenSSL.SSL.VERIFY_NONE,
+    ssl.CERT_OPTIONAL: OpenSSL.SSL.VERIFY_PEER,
+    ssl.CERT_REQUIRED: OpenSSL.SSL.VERIFY_PEER
+    + OpenSSL.SSL.VERIFY_FAIL_IF_NO_PEER_CERT,
+}
+_openssl_to_stdlib_verify = dict((v, k) for k, v in _stdlib_to_openssl_verify.items())
+
+# OpenSSL will only write 16K at a time
+SSL_WRITE_BLOCKSIZE = 16384
+
+orig_util_HAS_SNI = util.HAS_SNI
+orig_util_SSLContext = util.ssl_.SSLContext
+
+
+log = logging.getLogger(__name__)
+
+
+def inject_into_urllib3():
+    "Monkey-patch urllib3 with PyOpenSSL-backed SSL-support."
+
+    _validate_dependencies_met()
+
+    util.SSLContext = PyOpenSSLContext
+    util.ssl_.SSLContext = PyOpenSSLContext
+    util.HAS_SNI = HAS_SNI
+    util.ssl_.HAS_SNI = HAS_SNI
+    util.IS_PYOPENSSL = True
+    util.ssl_.IS_PYOPENSSL = True
+
+
+def extract_from_urllib3():
+    "Undo monkey-patching by :func:`inject_into_urllib3`."
+
+    util.SSLContext = orig_util_SSLContext
+    util.ssl_.SSLContext = orig_util_SSLContext
+    util.HAS_SNI = orig_util_HAS_SNI
+    util.ssl_.HAS_SNI = orig_util_HAS_SNI
+    util.IS_PYOPENSSL = False
+    util.ssl_.IS_PYOPENSSL = False
+
+
+def _validate_dependencies_met():
+    """
+    Verifies that PyOpenSSL's package-level dependencies have been met.
+    Throws `ImportError` if they are not met.
+    """
+    # Method added in `cryptography==1.1`; not available in older versions
+    from cryptography.x509.extensions import Extensions
+
+    if getattr(Extensions, "get_extension_for_class", None) is None:
+        raise ImportError(
+            "'cryptography' module missing required functionality.  "
+            "Try upgrading to v1.3.4 or newer."
+        )
+
+    # pyOpenSSL 0.14 and above use cryptography for OpenSSL bindings. The _x509
+    # attribute is only present on those versions.
+    from OpenSSL.crypto import X509
+
+    x509 = X509()
+    if getattr(x509, "_x509", None) is None:
+        raise ImportError(
+            "'pyOpenSSL' module missing required functionality. "
+            "Try upgrading to v0.14 or newer."
+        )
+
+
+def _dnsname_to_stdlib(name):
+    """
+    Converts a dNSName SubjectAlternativeName field to the form used by the
+    standard library on the given Python version.
+
+    Cryptography produces a dNSName as a unicode string that was idna-decoded
+    from ASCII bytes. We need to idna-encode that string to get it back, and
+    then on Python 3 we also need to convert to unicode via UTF-8 (the stdlib
+    uses PyUnicode_FromStringAndSize on it, which decodes via UTF-8).
+
+    If the name cannot be idna-encoded then we return None signalling that
+    the name given should be skipped.
+    """
+
+    def idna_encode(name):
+        """
+        Borrowed wholesale from the Python Cryptography Project. It turns out
+        that we can't just safely call `idna.encode`: it can explode for
+        wildcard names. This avoids that problem.
+        """
+        from pip._vendor import idna
+
+        try:
+            for prefix in [u"*.", u"."]:
+                if name.startswith(prefix):
+                    name = name[len(prefix) :]
+                    return prefix.encode("ascii") + idna.encode(name)
+            return idna.encode(name)
+        except idna.core.IDNAError:
+            return None
+
+    # Don't send IPv6 addresses through the IDNA encoder.
+    if ":" in name:
+        return name
+
+    name = idna_encode(name)
+    if name is None:
+        return None
+    elif sys.version_info >= (3, 0):
+        name = name.decode("utf-8")
+    return name
+
+
+def get_subj_alt_name(peer_cert):
+    """
+    Given an PyOpenSSL certificate, provides all the subject alternative names.
+    """
+    # Pass the cert to cryptography, which has much better APIs for this.
+    if hasattr(peer_cert, "to_cryptography"):
+        cert = peer_cert.to_cryptography()
+    else:
+        der = OpenSSL.crypto.dump_certificate(OpenSSL.crypto.FILETYPE_ASN1, peer_cert)
+        cert = x509.load_der_x509_certificate(der, openssl_backend)
+
+    # We want to find the SAN extension. Ask Cryptography to locate it (it's
+    # faster than looping in Python)
+    try:
+        ext = cert.extensions.get_extension_for_class(x509.SubjectAlternativeName).value
+    except x509.ExtensionNotFound:
+        # No such extension, return the empty list.
+        return []
+    except (
+        x509.DuplicateExtension,
+        UnsupportedExtension,
+        x509.UnsupportedGeneralNameType,
+        UnicodeError,
+    ) as e:
+        # A problem has been found with the quality of the certificate. Assume
+        # no SAN field is present.
+        log.warning(
+            "A problem was encountered with the certificate that prevented "
+            "urllib3 from finding the SubjectAlternativeName field. This can "
+            "affect certificate validation. The error was %s",
+            e,
+        )
+        return []
+
+    # We want to return dNSName and iPAddress fields. We need to cast the IPs
+    # back to strings because the match_hostname function wants them as
+    # strings.
+    # Sadly the DNS names need to be idna encoded and then, on Python 3, UTF-8
+    # decoded. This is pretty frustrating, but that's what the standard library
+    # does with certificates, and so we need to attempt to do the same.
+    # We also want to skip over names which cannot be idna encoded.
+    names = [
+        ("DNS", name)
+        for name in map(_dnsname_to_stdlib, ext.get_values_for_type(x509.DNSName))
+        if name is not None
+    ]
+    names.extend(
+        ("IP Address", str(name)) for name in ext.get_values_for_type(x509.IPAddress)
+    )
+
+    return names
+
+
+class WrappedSocket(object):
+    """API-compatibility wrapper for Python OpenSSL's Connection-class.
+
+    Note: _makefile_refs, _drop() and _reuse() are needed for the garbage
+    collector of pypy.
+    """
+
+    def __init__(self, connection, socket, suppress_ragged_eofs=True):
+        self.connection = connection
+        self.socket = socket
+        self.suppress_ragged_eofs = suppress_ragged_eofs
+        self._makefile_refs = 0
+        self._closed = False
+
+    def fileno(self):
+        return self.socket.fileno()
+
+    # Copy-pasted from Python 3.5 source code
+    def _decref_socketios(self):
+        if self._makefile_refs > 0:
+            self._makefile_refs -= 1
+        if self._closed:
+            self.close()
+
+    def recv(self, *args, **kwargs):
+        try:
+            data = self.connection.recv(*args, **kwargs)
+        except OpenSSL.SSL.SysCallError as e:
+            if self.suppress_ragged_eofs and e.args == (-1, "Unexpected EOF"):
+                return b""
+            else:
+                raise SocketError(str(e))
+        except OpenSSL.SSL.ZeroReturnError:
+            if self.connection.get_shutdown() == OpenSSL.SSL.RECEIVED_SHUTDOWN:
+                return b""
+            else:
+                raise
+        except OpenSSL.SSL.WantReadError:
+            if not util.wait_for_read(self.socket, self.socket.gettimeout()):
+                raise timeout("The read operation timed out")
+            else:
+                return self.recv(*args, **kwargs)
+
+        # TLS 1.3 post-handshake authentication
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError("read error: %r" % e)
+        else:
+            return data
+
+    def recv_into(self, *args, **kwargs):
+        try:
+            return self.connection.recv_into(*args, **kwargs)
+        except OpenSSL.SSL.SysCallError as e:
+            if self.suppress_ragged_eofs and e.args == (-1, "Unexpected EOF"):
+                return 0
+            else:
+                raise SocketError(str(e))
+        except OpenSSL.SSL.ZeroReturnError:
+            if self.connection.get_shutdown() == OpenSSL.SSL.RECEIVED_SHUTDOWN:
+                return 0
+            else:
+                raise
+        except OpenSSL.SSL.WantReadError:
+            if not util.wait_for_read(self.socket, self.socket.gettimeout()):
+                raise timeout("The read operation timed out")
+            else:
+                return self.recv_into(*args, **kwargs)
+
+        # TLS 1.3 post-handshake authentication
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError("read error: %r" % e)
+
+    def settimeout(self, timeout):
+        return self.socket.settimeout(timeout)
+
+    def _send_until_done(self, data):
+        while True:
+            try:
+                return self.connection.send(data)
+            except OpenSSL.SSL.WantWriteError:
+                if not util.wait_for_write(self.socket, self.socket.gettimeout()):
+                    raise timeout()
+                continue
+            except OpenSSL.SSL.SysCallError as e:
+                raise SocketError(str(e))
+
+    def sendall(self, data):
+        total_sent = 0
+        while total_sent < len(data):
+            sent = self._send_until_done(
+                data[total_sent : total_sent + SSL_WRITE_BLOCKSIZE]
+            )
+            total_sent += sent
+
+    def shutdown(self):
+        # FIXME rethrow compatible exceptions should we ever use this
+        self.connection.shutdown()
+
+    def close(self):
+        if self._makefile_refs < 1:
+            try:
+                self._closed = True
+                return self.connection.close()
+            except OpenSSL.SSL.Error:
+                return
+        else:
+            self._makefile_refs -= 1
+
+    def getpeercert(self, binary_form=False):
+        x509 = self.connection.get_peer_certificate()
+
+        if not x509:
+            return x509
+
+        if binary_form:
+            return OpenSSL.crypto.dump_certificate(OpenSSL.crypto.FILETYPE_ASN1, x509)
+
+        return {
+            "subject": ((("commonName", x509.get_subject().CN),),),
+            "subjectAltName": get_subj_alt_name(x509),
+        }
+
+    def version(self):
+        return self.connection.get_protocol_version_name()
+
+    def _reuse(self):
+        self._makefile_refs += 1
+
+    def _drop(self):
+        if self._makefile_refs < 1:
+            self.close()
+        else:
+            self._makefile_refs -= 1
+
+
+if _fileobject:  # Platform-specific: Python 2
+
+    def makefile(self, mode, bufsize=-1):
+        self._makefile_refs += 1
+        return _fileobject(self, mode, bufsize, close=True)
+
+else:  # Platform-specific: Python 3
+    makefile = backport_makefile
+
+WrappedSocket.makefile = makefile
+
+
+class PyOpenSSLContext(object):
+    """
+    I am a wrapper class for the PyOpenSSL ``Context`` object. I am responsible
+    for translating the interface of the standard library ``SSLContext`` object
+    to calls into PyOpenSSL.
+    """
+
+    def __init__(self, protocol):
+        self.protocol = _openssl_versions[protocol]
+        self._ctx = OpenSSL.SSL.Context(self.protocol)
+        self._options = 0
+        self.check_hostname = False
+
+    @property
+    def options(self):
+        return self._options
+
+    @options.setter
+    def options(self, value):
+        self._options = value
+        self._ctx.set_options(value)
+
+    @property
+    def verify_mode(self):
+        return _openssl_to_stdlib_verify[self._ctx.get_verify_mode()]
+
+    @verify_mode.setter
+    def verify_mode(self, value):
+        self._ctx.set_verify(_stdlib_to_openssl_verify[value], _verify_callback)
+
+    def set_default_verify_paths(self):
+        self._ctx.set_default_verify_paths()
+
+    def set_ciphers(self, ciphers):
+        if isinstance(ciphers, six.text_type):
+            ciphers = ciphers.encode("utf-8")
+        self._ctx.set_cipher_list(ciphers)
+
+    def load_verify_locations(self, cafile=None, capath=None, cadata=None):
+        if cafile is not None:
+            cafile = cafile.encode("utf-8")
+        if capath is not None:
+            capath = capath.encode("utf-8")
+        try:
+            self._ctx.load_verify_locations(cafile, capath)
+            if cadata is not None:
+                self._ctx.load_verify_locations(BytesIO(cadata))
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError("unable to load trusted certificates: %r" % e)
+
+    def load_cert_chain(self, certfile, keyfile=None, password=None):
+        self._ctx.use_certificate_chain_file(certfile)
+        if password is not None:
+            if not isinstance(password, six.binary_type):
+                password = password.encode("utf-8")
+            self._ctx.set_passwd_cb(lambda *_: password)
+        self._ctx.use_privatekey_file(keyfile or certfile)
+
+    def set_alpn_protocols(self, protocols):
+        protocols = [six.ensure_binary(p) for p in protocols]
+        return self._ctx.set_alpn_protos(protocols)
+
+    def wrap_socket(
+        self,
+        sock,
+        server_side=False,
+        do_handshake_on_connect=True,
+        suppress_ragged_eofs=True,
+        server_hostname=None,
+    ):
+        cnx = OpenSSL.SSL.Connection(self._ctx, sock)
+
+        if isinstance(server_hostname, six.text_type):  # Platform-specific: Python 3
+            server_hostname = server_hostname.encode("utf-8")
+
+        if server_hostname is not None:
+            cnx.set_tlsext_host_name(server_hostname)
+
+        cnx.set_connect_state()
+
+        while True:
+            try:
+                cnx.do_handshake()
+            except OpenSSL.SSL.WantReadError:
+                if not util.wait_for_read(sock, sock.gettimeout()):
+                    raise timeout("select timed out")
+                continue
+            except OpenSSL.SSL.Error as e:
+                raise ssl.SSLError("bad handshake: %r" % e)
+            break
+
+        return WrappedSocket(cnx, sock)
+
+
+def _verify_callback(cnx, x509, err_no, err_depth, return_code):
+    return err_no == 0
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/securetransport.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/securetransport.py
new file mode 100644
index 00000000..722ee4e1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/securetransport.py
@@ -0,0 +1,920 @@
+"""
+SecureTranport support for urllib3 via ctypes.
+
+This makes platform-native TLS available to urllib3 users on macOS without the
+use of a compiler. This is an important feature because the Python Package
+Index is moving to become a TLSv1.2-or-higher server, and the default OpenSSL
+that ships with macOS is not capable of doing TLSv1.2. The only way to resolve
+this is to give macOS users an alternative solution to the problem, and that
+solution is to use SecureTransport.
+
+We use ctypes here because this solution must not require a compiler. That's
+because pip is not allowed to require a compiler either.
+
+This is not intended to be a seriously long-term solution to this problem.
+The hope is that PEP 543 will eventually solve this issue for us, at which
+point we can retire this contrib module. But in the short term, we need to
+solve the impending tire fire that is Python on Mac without this kind of
+contrib module. So...here we are.
+
+To use this module, simply import and inject it::
+
+    import pip._vendor.urllib3.contrib.securetransport as securetransport
+    securetransport.inject_into_urllib3()
+
+Happy TLSing!
+
+This code is a bastardised version of the code found in Will Bond's oscrypto
+library. An enormous debt is owed to him for blazing this trail for us. For
+that reason, this code should be considered to be covered both by urllib3's
+license and by oscrypto's:
+
+.. code-block::
+
+    Copyright (c) 2015-2016 Will Bond <will@wbond.net>
+
+    Permission is hereby granted, free of charge, to any person obtaining a
+    copy of this software and associated documentation files (the "Software"),
+    to deal in the Software without restriction, including without limitation
+    the rights to use, copy, modify, merge, publish, distribute, sublicense,
+    and/or sell copies of the Software, and to permit persons to whom the
+    Software is furnished to do so, subject to the following conditions:
+
+    The above copyright notice and this permission notice shall be included in
+    all copies or substantial portions of the Software.
+
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+    FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+    DEALINGS IN THE SOFTWARE.
+"""
+from __future__ import absolute_import
+
+import contextlib
+import ctypes
+import errno
+import os.path
+import shutil
+import socket
+import ssl
+import struct
+import threading
+import weakref
+
+from .. import util
+from ..packages import six
+from ..util.ssl_ import PROTOCOL_TLS_CLIENT
+from ._securetransport.bindings import CoreFoundation, Security, SecurityConst
+from ._securetransport.low_level import (
+    _assert_no_error,
+    _build_tls_unknown_ca_alert,
+    _cert_array_from_pem,
+    _create_cfstring_array,
+    _load_client_cert_chain,
+    _temporary_keychain,
+)
+
+try:  # Platform-specific: Python 2
+    from socket import _fileobject
+except ImportError:  # Platform-specific: Python 3
+    _fileobject = None
+    from ..packages.backports.makefile import backport_makefile
+
+__all__ = ["inject_into_urllib3", "extract_from_urllib3"]
+
+# SNI always works
+HAS_SNI = True
+
+orig_util_HAS_SNI = util.HAS_SNI
+orig_util_SSLContext = util.ssl_.SSLContext
+
+# This dictionary is used by the read callback to obtain a handle to the
+# calling wrapped socket. This is a pretty silly approach, but for now it'll
+# do. I feel like I should be able to smuggle a handle to the wrapped socket
+# directly in the SSLConnectionRef, but for now this approach will work I
+# guess.
+#
+# We need to lock around this structure for inserts, but we don't do it for
+# reads/writes in the callbacks. The reasoning here goes as follows:
+#
+#    1. It is not possible to call into the callbacks before the dictionary is
+#       populated, so once in the callback the id must be in the dictionary.
+#    2. The callbacks don't mutate the dictionary, they only read from it, and
+#       so cannot conflict with any of the insertions.
+#
+# This is good: if we had to lock in the callbacks we'd drastically slow down
+# the performance of this code.
+_connection_refs = weakref.WeakValueDictionary()
+_connection_ref_lock = threading.Lock()
+
+# Limit writes to 16kB. This is OpenSSL's limit, but we'll cargo-cult it over
+# for no better reason than we need *a* limit, and this one is right there.
+SSL_WRITE_BLOCKSIZE = 16384
+
+# This is our equivalent of util.ssl_.DEFAULT_CIPHERS, but expanded out to
+# individual cipher suites. We need to do this because this is how
+# SecureTransport wants them.
+CIPHER_SUITES = [
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256,
+    SecurityConst.TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_256_GCM_SHA384,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_128_GCM_SHA256,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256,
+    SecurityConst.TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256,
+    SecurityConst.TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_256_CBC_SHA256,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_256_CBC_SHA,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_128_CBC_SHA256,
+    SecurityConst.TLS_DHE_RSA_WITH_AES_128_CBC_SHA,
+    SecurityConst.TLS_AES_256_GCM_SHA384,
+    SecurityConst.TLS_AES_128_GCM_SHA256,
+    SecurityConst.TLS_RSA_WITH_AES_256_GCM_SHA384,
+    SecurityConst.TLS_RSA_WITH_AES_128_GCM_SHA256,
+    SecurityConst.TLS_AES_128_CCM_8_SHA256,
+    SecurityConst.TLS_AES_128_CCM_SHA256,
+    SecurityConst.TLS_RSA_WITH_AES_256_CBC_SHA256,
+    SecurityConst.TLS_RSA_WITH_AES_128_CBC_SHA256,
+    SecurityConst.TLS_RSA_WITH_AES_256_CBC_SHA,
+    SecurityConst.TLS_RSA_WITH_AES_128_CBC_SHA,
+]
+
+# Basically this is simple: for PROTOCOL_SSLv23 we turn it into a low of
+# TLSv1 and a high of TLSv1.2. For everything else, we pin to that version.
+# TLSv1 to 1.2 are supported on macOS 10.8+
+_protocol_to_min_max = {
+    util.PROTOCOL_TLS: (SecurityConst.kTLSProtocol1, SecurityConst.kTLSProtocol12),
+    PROTOCOL_TLS_CLIENT: (SecurityConst.kTLSProtocol1, SecurityConst.kTLSProtocol12),
+}
+
+if hasattr(ssl, "PROTOCOL_SSLv2"):
+    _protocol_to_min_max[ssl.PROTOCOL_SSLv2] = (
+        SecurityConst.kSSLProtocol2,
+        SecurityConst.kSSLProtocol2,
+    )
+if hasattr(ssl, "PROTOCOL_SSLv3"):
+    _protocol_to_min_max[ssl.PROTOCOL_SSLv3] = (
+        SecurityConst.kSSLProtocol3,
+        SecurityConst.kSSLProtocol3,
+    )
+if hasattr(ssl, "PROTOCOL_TLSv1"):
+    _protocol_to_min_max[ssl.PROTOCOL_TLSv1] = (
+        SecurityConst.kTLSProtocol1,
+        SecurityConst.kTLSProtocol1,
+    )
+if hasattr(ssl, "PROTOCOL_TLSv1_1"):
+    _protocol_to_min_max[ssl.PROTOCOL_TLSv1_1] = (
+        SecurityConst.kTLSProtocol11,
+        SecurityConst.kTLSProtocol11,
+    )
+if hasattr(ssl, "PROTOCOL_TLSv1_2"):
+    _protocol_to_min_max[ssl.PROTOCOL_TLSv1_2] = (
+        SecurityConst.kTLSProtocol12,
+        SecurityConst.kTLSProtocol12,
+    )
+
+
+def inject_into_urllib3():
+    """
+    Monkey-patch urllib3 with SecureTransport-backed SSL-support.
+    """
+    util.SSLContext = SecureTransportContext
+    util.ssl_.SSLContext = SecureTransportContext
+    util.HAS_SNI = HAS_SNI
+    util.ssl_.HAS_SNI = HAS_SNI
+    util.IS_SECURETRANSPORT = True
+    util.ssl_.IS_SECURETRANSPORT = True
+
+
+def extract_from_urllib3():
+    """
+    Undo monkey-patching by :func:`inject_into_urllib3`.
+    """
+    util.SSLContext = orig_util_SSLContext
+    util.ssl_.SSLContext = orig_util_SSLContext
+    util.HAS_SNI = orig_util_HAS_SNI
+    util.ssl_.HAS_SNI = orig_util_HAS_SNI
+    util.IS_SECURETRANSPORT = False
+    util.ssl_.IS_SECURETRANSPORT = False
+
+
+def _read_callback(connection_id, data_buffer, data_length_pointer):
+    """
+    SecureTransport read callback. This is called by ST to request that data
+    be returned from the socket.
+    """
+    wrapped_socket = None
+    try:
+        wrapped_socket = _connection_refs.get(connection_id)
+        if wrapped_socket is None:
+            return SecurityConst.errSSLInternal
+        base_socket = wrapped_socket.socket
+
+        requested_length = data_length_pointer[0]
+
+        timeout = wrapped_socket.gettimeout()
+        error = None
+        read_count = 0
+
+        try:
+            while read_count < requested_length:
+                if timeout is None or timeout >= 0:
+                    if not util.wait_for_read(base_socket, timeout):
+                        raise socket.error(errno.EAGAIN, "timed out")
+
+                remaining = requested_length - read_count
+                buffer = (ctypes.c_char * remaining).from_address(
+                    data_buffer + read_count
+                )
+                chunk_size = base_socket.recv_into(buffer, remaining)
+                read_count += chunk_size
+                if not chunk_size:
+                    if not read_count:
+                        return SecurityConst.errSSLClosedGraceful
+                    break
+        except (socket.error) as e:
+            error = e.errno
+
+            if error is not None and error != errno.EAGAIN:
+                data_length_pointer[0] = read_count
+                if error == errno.ECONNRESET or error == errno.EPIPE:
+                    return SecurityConst.errSSLClosedAbort
+                raise
+
+        data_length_pointer[0] = read_count
+
+        if read_count != requested_length:
+            return SecurityConst.errSSLWouldBlock
+
+        return 0
+    except Exception as e:
+        if wrapped_socket is not None:
+            wrapped_socket._exception = e
+        return SecurityConst.errSSLInternal
+
+
+def _write_callback(connection_id, data_buffer, data_length_pointer):
+    """
+    SecureTransport write callback. This is called by ST to request that data
+    actually be sent on the network.
+    """
+    wrapped_socket = None
+    try:
+        wrapped_socket = _connection_refs.get(connection_id)
+        if wrapped_socket is None:
+            return SecurityConst.errSSLInternal
+        base_socket = wrapped_socket.socket
+
+        bytes_to_write = data_length_pointer[0]
+        data = ctypes.string_at(data_buffer, bytes_to_write)
+
+        timeout = wrapped_socket.gettimeout()
+        error = None
+        sent = 0
+
+        try:
+            while sent < bytes_to_write:
+                if timeout is None or timeout >= 0:
+                    if not util.wait_for_write(base_socket, timeout):
+                        raise socket.error(errno.EAGAIN, "timed out")
+                chunk_sent = base_socket.send(data)
+                sent += chunk_sent
+
+                # This has some needless copying here, but I'm not sure there's
+                # much value in optimising this data path.
+                data = data[chunk_sent:]
+        except (socket.error) as e:
+            error = e.errno
+
+            if error is not None and error != errno.EAGAIN:
+                data_length_pointer[0] = sent
+                if error == errno.ECONNRESET or error == errno.EPIPE:
+                    return SecurityConst.errSSLClosedAbort
+                raise
+
+        data_length_pointer[0] = sent
+
+        if sent != bytes_to_write:
+            return SecurityConst.errSSLWouldBlock
+
+        return 0
+    except Exception as e:
+        if wrapped_socket is not None:
+            wrapped_socket._exception = e
+        return SecurityConst.errSSLInternal
+
+
+# We need to keep these two objects references alive: if they get GC'd while
+# in use then SecureTransport could attempt to call a function that is in freed
+# memory. That would be...uh...bad. Yeah, that's the word. Bad.
+_read_callback_pointer = Security.SSLReadFunc(_read_callback)
+_write_callback_pointer = Security.SSLWriteFunc(_write_callback)
+
+
+class WrappedSocket(object):
+    """
+    API-compatibility wrapper for Python's OpenSSL wrapped socket object.
+
+    Note: _makefile_refs, _drop(), and _reuse() are needed for the garbage
+    collector of PyPy.
+    """
+
+    def __init__(self, socket):
+        self.socket = socket
+        self.context = None
+        self._makefile_refs = 0
+        self._closed = False
+        self._exception = None
+        self._keychain = None
+        self._keychain_dir = None
+        self._client_cert_chain = None
+
+        # We save off the previously-configured timeout and then set it to
+        # zero. This is done because we use select and friends to handle the
+        # timeouts, but if we leave the timeout set on the lower socket then
+        # Python will "kindly" call select on that socket again for us. Avoid
+        # that by forcing the timeout to zero.
+        self._timeout = self.socket.gettimeout()
+        self.socket.settimeout(0)
+
+    @contextlib.contextmanager
+    def _raise_on_error(self):
+        """
+        A context manager that can be used to wrap calls that do I/O from
+        SecureTransport. If any of the I/O callbacks hit an exception, this
+        context manager will correctly propagate the exception after the fact.
+        This avoids silently swallowing those exceptions.
+
+        It also correctly forces the socket closed.
+        """
+        self._exception = None
+
+        # We explicitly don't catch around this yield because in the unlikely
+        # event that an exception was hit in the block we don't want to swallow
+        # it.
+        yield
+        if self._exception is not None:
+            exception, self._exception = self._exception, None
+            self.close()
+            raise exception
+
+    def _set_ciphers(self):
+        """
+        Sets up the allowed ciphers. By default this matches the set in
+        util.ssl_.DEFAULT_CIPHERS, at least as supported by macOS. This is done
+        custom and doesn't allow changing at this time, mostly because parsing
+        OpenSSL cipher strings is going to be a freaking nightmare.
+        """
+        ciphers = (Security.SSLCipherSuite * len(CIPHER_SUITES))(*CIPHER_SUITES)
+        result = Security.SSLSetEnabledCiphers(
+            self.context, ciphers, len(CIPHER_SUITES)
+        )
+        _assert_no_error(result)
+
+    def _set_alpn_protocols(self, protocols):
+        """
+        Sets up the ALPN protocols on the context.
+        """
+        if not protocols:
+            return
+        protocols_arr = _create_cfstring_array(protocols)
+        try:
+            result = Security.SSLSetALPNProtocols(self.context, protocols_arr)
+            _assert_no_error(result)
+        finally:
+            CoreFoundation.CFRelease(protocols_arr)
+
+    def _custom_validate(self, verify, trust_bundle):
+        """
+        Called when we have set custom validation. We do this in two cases:
+        first, when cert validation is entirely disabled; and second, when
+        using a custom trust DB.
+        Raises an SSLError if the connection is not trusted.
+        """
+        # If we disabled cert validation, just say: cool.
+        if not verify:
+            return
+
+        successes = (
+            SecurityConst.kSecTrustResultUnspecified,
+            SecurityConst.kSecTrustResultProceed,
+        )
+        try:
+            trust_result = self._evaluate_trust(trust_bundle)
+            if trust_result in successes:
+                return
+            reason = "error code: %d" % (trust_result,)
+        except Exception as e:
+            # Do not trust on error
+            reason = "exception: %r" % (e,)
+
+        # SecureTransport does not send an alert nor shuts down the connection.
+        rec = _build_tls_unknown_ca_alert(self.version())
+        self.socket.sendall(rec)
+        # close the connection immediately
+        # l_onoff = 1, activate linger
+        # l_linger = 0, linger for 0 seoncds
+        opts = struct.pack("ii", 1, 0)
+        self.socket.setsockopt(socket.SOL_SOCKET, socket.SO_LINGER, opts)
+        self.close()
+        raise ssl.SSLError("certificate verify failed, %s" % reason)
+
+    def _evaluate_trust(self, trust_bundle):
+        # We want data in memory, so load it up.
+        if os.path.isfile(trust_bundle):
+            with open(trust_bundle, "rb") as f:
+                trust_bundle = f.read()
+
+        cert_array = None
+        trust = Security.SecTrustRef()
+
+        try:
+            # Get a CFArray that contains the certs we want.
+            cert_array = _cert_array_from_pem(trust_bundle)
+
+            # Ok, now the hard part. We want to get the SecTrustRef that ST has
+            # created for this connection, shove our CAs into it, tell ST to
+            # ignore everything else it knows, and then ask if it can build a
+            # chain. This is a buuuunch of code.
+            result = Security.SSLCopyPeerTrust(self.context, ctypes.byref(trust))
+            _assert_no_error(result)
+            if not trust:
+                raise ssl.SSLError("Failed to copy trust reference")
+
+            result = Security.SecTrustSetAnchorCertificates(trust, cert_array)
+            _assert_no_error(result)
+
+            result = Security.SecTrustSetAnchorCertificatesOnly(trust, True)
+            _assert_no_error(result)
+
+            trust_result = Security.SecTrustResultType()
+            result = Security.SecTrustEvaluate(trust, ctypes.byref(trust_result))
+            _assert_no_error(result)
+        finally:
+            if trust:
+                CoreFoundation.CFRelease(trust)
+
+            if cert_array is not None:
+                CoreFoundation.CFRelease(cert_array)
+
+        return trust_result.value
+
+    def handshake(
+        self,
+        server_hostname,
+        verify,
+        trust_bundle,
+        min_version,
+        max_version,
+        client_cert,
+        client_key,
+        client_key_passphrase,
+        alpn_protocols,
+    ):
+        """
+        Actually performs the TLS handshake. This is run automatically by
+        wrapped socket, and shouldn't be needed in user code.
+        """
+        # First, we do the initial bits of connection setup. We need to create
+        # a context, set its I/O funcs, and set the connection reference.
+        self.context = Security.SSLCreateContext(
+            None, SecurityConst.kSSLClientSide, SecurityConst.kSSLStreamType
+        )
+        result = Security.SSLSetIOFuncs(
+            self.context, _read_callback_pointer, _write_callback_pointer
+        )
+        _assert_no_error(result)
+
+        # Here we need to compute the handle to use. We do this by taking the
+        # id of self modulo 2**31 - 1. If this is already in the dictionary, we
+        # just keep incrementing by one until we find a free space.
+        with _connection_ref_lock:
+            handle = id(self) % 2147483647
+            while handle in _connection_refs:
+                handle = (handle + 1) % 2147483647
+            _connection_refs[handle] = self
+
+        result = Security.SSLSetConnection(self.context, handle)
+        _assert_no_error(result)
+
+        # If we have a server hostname, we should set that too.
+        if server_hostname:
+            if not isinstance(server_hostname, bytes):
+                server_hostname = server_hostname.encode("utf-8")
+
+            result = Security.SSLSetPeerDomainName(
+                self.context, server_hostname, len(server_hostname)
+            )
+            _assert_no_error(result)
+
+        # Setup the ciphers.
+        self._set_ciphers()
+
+        # Setup the ALPN protocols.
+        self._set_alpn_protocols(alpn_protocols)
+
+        # Set the minimum and maximum TLS versions.
+        result = Security.SSLSetProtocolVersionMin(self.context, min_version)
+        _assert_no_error(result)
+
+        result = Security.SSLSetProtocolVersionMax(self.context, max_version)
+        _assert_no_error(result)
+
+        # If there's a trust DB, we need to use it. We do that by telling
+        # SecureTransport to break on server auth. We also do that if we don't
+        # want to validate the certs at all: we just won't actually do any
+        # authing in that case.
+        if not verify or trust_bundle is not None:
+            result = Security.SSLSetSessionOption(
+                self.context, SecurityConst.kSSLSessionOptionBreakOnServerAuth, True
+            )
+            _assert_no_error(result)
+
+        # If there's a client cert, we need to use it.
+        if client_cert:
+            self._keychain, self._keychain_dir = _temporary_keychain()
+            self._client_cert_chain = _load_client_cert_chain(
+                self._keychain, client_cert, client_key
+            )
+            result = Security.SSLSetCertificate(self.context, self._client_cert_chain)
+            _assert_no_error(result)
+
+        while True:
+            with self._raise_on_error():
+                result = Security.SSLHandshake(self.context)
+
+                if result == SecurityConst.errSSLWouldBlock:
+                    raise socket.timeout("handshake timed out")
+                elif result == SecurityConst.errSSLServerAuthCompleted:
+                    self._custom_validate(verify, trust_bundle)
+                    continue
+                else:
+                    _assert_no_error(result)
+                    break
+
+    def fileno(self):
+        return self.socket.fileno()
+
+    # Copy-pasted from Python 3.5 source code
+    def _decref_socketios(self):
+        if self._makefile_refs > 0:
+            self._makefile_refs -= 1
+        if self._closed:
+            self.close()
+
+    def recv(self, bufsiz):
+        buffer = ctypes.create_string_buffer(bufsiz)
+        bytes_read = self.recv_into(buffer, bufsiz)
+        data = buffer[:bytes_read]
+        return data
+
+    def recv_into(self, buffer, nbytes=None):
+        # Read short on EOF.
+        if self._closed:
+            return 0
+
+        if nbytes is None:
+            nbytes = len(buffer)
+
+        buffer = (ctypes.c_char * nbytes).from_buffer(buffer)
+        processed_bytes = ctypes.c_size_t(0)
+
+        with self._raise_on_error():
+            result = Security.SSLRead(
+                self.context, buffer, nbytes, ctypes.byref(processed_bytes)
+            )
+
+        # There are some result codes that we want to treat as "not always
+        # errors". Specifically, those are errSSLWouldBlock,
+        # errSSLClosedGraceful, and errSSLClosedNoNotify.
+        if result == SecurityConst.errSSLWouldBlock:
+            # If we didn't process any bytes, then this was just a time out.
+            # However, we can get errSSLWouldBlock in situations when we *did*
+            # read some data, and in those cases we should just read "short"
+            # and return.
+            if processed_bytes.value == 0:
+                # Timed out, no data read.
+                raise socket.timeout("recv timed out")
+        elif result in (
+            SecurityConst.errSSLClosedGraceful,
+            SecurityConst.errSSLClosedNoNotify,
+        ):
+            # The remote peer has closed this connection. We should do so as
+            # well. Note that we don't actually return here because in
+            # principle this could actually be fired along with return data.
+            # It's unlikely though.
+            self.close()
+        else:
+            _assert_no_error(result)
+
+        # Ok, we read and probably succeeded. We should return whatever data
+        # was actually read.
+        return processed_bytes.value
+
+    def settimeout(self, timeout):
+        self._timeout = timeout
+
+    def gettimeout(self):
+        return self._timeout
+
+    def send(self, data):
+        processed_bytes = ctypes.c_size_t(0)
+
+        with self._raise_on_error():
+            result = Security.SSLWrite(
+                self.context, data, len(data), ctypes.byref(processed_bytes)
+            )
+
+        if result == SecurityConst.errSSLWouldBlock and processed_bytes.value == 0:
+            # Timed out
+            raise socket.timeout("send timed out")
+        else:
+            _assert_no_error(result)
+
+        # We sent, and probably succeeded. Tell them how much we sent.
+        return processed_bytes.value
+
+    def sendall(self, data):
+        total_sent = 0
+        while total_sent < len(data):
+            sent = self.send(data[total_sent : total_sent + SSL_WRITE_BLOCKSIZE])
+            total_sent += sent
+
+    def shutdown(self):
+        with self._raise_on_error():
+            Security.SSLClose(self.context)
+
+    def close(self):
+        # TODO: should I do clean shutdown here? Do I have to?
+        if self._makefile_refs < 1:
+            self._closed = True
+            if self.context:
+                CoreFoundation.CFRelease(self.context)
+                self.context = None
+            if self._client_cert_chain:
+                CoreFoundation.CFRelease(self._client_cert_chain)
+                self._client_cert_chain = None
+            if self._keychain:
+                Security.SecKeychainDelete(self._keychain)
+                CoreFoundation.CFRelease(self._keychain)
+                shutil.rmtree(self._keychain_dir)
+                self._keychain = self._keychain_dir = None
+            return self.socket.close()
+        else:
+            self._makefile_refs -= 1
+
+    def getpeercert(self, binary_form=False):
+        # Urgh, annoying.
+        #
+        # Here's how we do this:
+        #
+        # 1. Call SSLCopyPeerTrust to get hold of the trust object for this
+        #    connection.
+        # 2. Call SecTrustGetCertificateAtIndex for index 0 to get the leaf.
+        # 3. To get the CN, call SecCertificateCopyCommonName and process that
+        #    string so that it's of the appropriate type.
+        # 4. To get the SAN, we need to do something a bit more complex:
+        #    a. Call SecCertificateCopyValues to get the data, requesting
+        #       kSecOIDSubjectAltName.
+        #    b. Mess about with this dictionary to try to get the SANs out.
+        #
+        # This is gross. Really gross. It's going to be a few hundred LoC extra
+        # just to repeat something that SecureTransport can *already do*. So my
+        # operating assumption at this time is that what we want to do is
+        # instead to just flag to urllib3 that it shouldn't do its own hostname
+        # validation when using SecureTransport.
+        if not binary_form:
+            raise ValueError("SecureTransport only supports dumping binary certs")
+        trust = Security.SecTrustRef()
+        certdata = None
+        der_bytes = None
+
+        try:
+            # Grab the trust store.
+            result = Security.SSLCopyPeerTrust(self.context, ctypes.byref(trust))
+            _assert_no_error(result)
+            if not trust:
+                # Probably we haven't done the handshake yet. No biggie.
+                return None
+
+            cert_count = Security.SecTrustGetCertificateCount(trust)
+            if not cert_count:
+                # Also a case that might happen if we haven't handshaked.
+                # Handshook? Handshaken?
+                return None
+
+            leaf = Security.SecTrustGetCertificateAtIndex(trust, 0)
+            assert leaf
+
+            # Ok, now we want the DER bytes.
+            certdata = Security.SecCertificateCopyData(leaf)
+            assert certdata
+
+            data_length = CoreFoundation.CFDataGetLength(certdata)
+            data_buffer = CoreFoundation.CFDataGetBytePtr(certdata)
+            der_bytes = ctypes.string_at(data_buffer, data_length)
+        finally:
+            if certdata:
+                CoreFoundation.CFRelease(certdata)
+            if trust:
+                CoreFoundation.CFRelease(trust)
+
+        return der_bytes
+
+    def version(self):
+        protocol = Security.SSLProtocol()
+        result = Security.SSLGetNegotiatedProtocolVersion(
+            self.context, ctypes.byref(protocol)
+        )
+        _assert_no_error(result)
+        if protocol.value == SecurityConst.kTLSProtocol13:
+            raise ssl.SSLError("SecureTransport does not support TLS 1.3")
+        elif protocol.value == SecurityConst.kTLSProtocol12:
+            return "TLSv1.2"
+        elif protocol.value == SecurityConst.kTLSProtocol11:
+            return "TLSv1.1"
+        elif protocol.value == SecurityConst.kTLSProtocol1:
+            return "TLSv1"
+        elif protocol.value == SecurityConst.kSSLProtocol3:
+            return "SSLv3"
+        elif protocol.value == SecurityConst.kSSLProtocol2:
+            return "SSLv2"
+        else:
+            raise ssl.SSLError("Unknown TLS version: %r" % protocol)
+
+    def _reuse(self):
+        self._makefile_refs += 1
+
+    def _drop(self):
+        if self._makefile_refs < 1:
+            self.close()
+        else:
+            self._makefile_refs -= 1
+
+
+if _fileobject:  # Platform-specific: Python 2
+
+    def makefile(self, mode, bufsize=-1):
+        self._makefile_refs += 1
+        return _fileobject(self, mode, bufsize, close=True)
+
+else:  # Platform-specific: Python 3
+
+    def makefile(self, mode="r", buffering=None, *args, **kwargs):
+        # We disable buffering with SecureTransport because it conflicts with
+        # the buffering that ST does internally (see issue #1153 for more).
+        buffering = 0
+        return backport_makefile(self, mode, buffering, *args, **kwargs)
+
+
+WrappedSocket.makefile = makefile
+
+
+class SecureTransportContext(object):
+    """
+    I am a wrapper class for the SecureTransport library, to translate the
+    interface of the standard library ``SSLContext`` object to calls into
+    SecureTransport.
+    """
+
+    def __init__(self, protocol):
+        self._min_version, self._max_version = _protocol_to_min_max[protocol]
+        self._options = 0
+        self._verify = False
+        self._trust_bundle = None
+        self._client_cert = None
+        self._client_key = None
+        self._client_key_passphrase = None
+        self._alpn_protocols = None
+
+    @property
+    def check_hostname(self):
+        """
+        SecureTransport cannot have its hostname checking disabled. For more,
+        see the comment on getpeercert() in this file.
+        """
+        return True
+
+    @check_hostname.setter
+    def check_hostname(self, value):
+        """
+        SecureTransport cannot have its hostname checking disabled. For more,
+        see the comment on getpeercert() in this file.
+        """
+        pass
+
+    @property
+    def options(self):
+        # TODO: Well, crap.
+        #
+        # So this is the bit of the code that is the most likely to cause us
+        # trouble. Essentially we need to enumerate all of the SSL options that
+        # users might want to use and try to see if we can sensibly translate
+        # them, or whether we should just ignore them.
+        return self._options
+
+    @options.setter
+    def options(self, value):
+        # TODO: Update in line with above.
+        self._options = value
+
+    @property
+    def verify_mode(self):
+        return ssl.CERT_REQUIRED if self._verify else ssl.CERT_NONE
+
+    @verify_mode.setter
+    def verify_mode(self, value):
+        self._verify = True if value == ssl.CERT_REQUIRED else False
+
+    def set_default_verify_paths(self):
+        # So, this has to do something a bit weird. Specifically, what it does
+        # is nothing.
+        #
+        # This means that, if we had previously had load_verify_locations
+        # called, this does not undo that. We need to do that because it turns
+        # out that the rest of the urllib3 code will attempt to load the
+        # default verify paths if it hasn't been told about any paths, even if
+        # the context itself was sometime earlier. We resolve that by just
+        # ignoring it.
+        pass
+
+    def load_default_certs(self):
+        return self.set_default_verify_paths()
+
+    def set_ciphers(self, ciphers):
+        # For now, we just require the default cipher string.
+        if ciphers != util.ssl_.DEFAULT_CIPHERS:
+            raise ValueError("SecureTransport doesn't support custom cipher strings")
+
+    def load_verify_locations(self, cafile=None, capath=None, cadata=None):
+        # OK, we only really support cadata and cafile.
+        if capath is not None:
+            raise ValueError("SecureTransport does not support cert directories")
+
+        # Raise if cafile does not exist.
+        if cafile is not None:
+            with open(cafile):
+                pass
+
+        self._trust_bundle = cafile or cadata
+
+    def load_cert_chain(self, certfile, keyfile=None, password=None):
+        self._client_cert = certfile
+        self._client_key = keyfile
+        self._client_cert_passphrase = password
+
+    def set_alpn_protocols(self, protocols):
+        """
+        Sets the ALPN protocols that will later be set on the context.
+
+        Raises a NotImplementedError if ALPN is not supported.
+        """
+        if not hasattr(Security, "SSLSetALPNProtocols"):
+            raise NotImplementedError(
+                "SecureTransport supports ALPN only in macOS 10.12+"
+            )
+        self._alpn_protocols = [six.ensure_binary(p) for p in protocols]
+
+    def wrap_socket(
+        self,
+        sock,
+        server_side=False,
+        do_handshake_on_connect=True,
+        suppress_ragged_eofs=True,
+        server_hostname=None,
+    ):
+        # So, what do we do here? Firstly, we assert some properties. This is a
+        # stripped down shim, so there is some functionality we don't support.
+        # See PEP 543 for the real deal.
+        assert not server_side
+        assert do_handshake_on_connect
+        assert suppress_ragged_eofs
+
+        # Ok, we're good to go. Now we want to create the wrapped socket object
+        # and store it in the appropriate place.
+        wrapped_socket = WrappedSocket(sock)
+
+        # Now we can handshake
+        wrapped_socket.handshake(
+            server_hostname,
+            self._verify,
+            self._trust_bundle,
+            self._min_version,
+            self._max_version,
+            self._client_cert,
+            self._client_key,
+            self._client_key_passphrase,
+            self._alpn_protocols,
+        )
+        return wrapped_socket
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/socks.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/socks.py
new file mode 100644
index 00000000..c326e80d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/contrib/socks.py
@@ -0,0 +1,216 @@
+# -*- coding: utf-8 -*-
+"""
+This module contains provisional support for SOCKS proxies from within
+urllib3. This module supports SOCKS4, SOCKS4A (an extension of SOCKS4), and
+SOCKS5. To enable its functionality, either install PySocks or install this
+module with the ``socks`` extra.
+
+The SOCKS implementation supports the full range of urllib3 features. It also
+supports the following SOCKS features:
+
+- SOCKS4A (``proxy_url='socks4a://...``)
+- SOCKS4 (``proxy_url='socks4://...``)
+- SOCKS5 with remote DNS (``proxy_url='socks5h://...``)
+- SOCKS5 with local DNS (``proxy_url='socks5://...``)
+- Usernames and passwords for the SOCKS proxy
+
+.. note::
+   It is recommended to use ``socks5h://`` or ``socks4a://`` schemes in
+   your ``proxy_url`` to ensure that DNS resolution is done from the remote
+   server instead of client-side when connecting to a domain name.
+
+SOCKS4 supports IPv4 and domain names with the SOCKS4A extension. SOCKS5
+supports IPv4, IPv6, and domain names.
+
+When connecting to a SOCKS4 proxy the ``username`` portion of the ``proxy_url``
+will be sent as the ``userid`` section of the SOCKS request:
+
+.. code-block:: python
+
+    proxy_url="socks4a://<userid>@proxy-host"
+
+When connecting to a SOCKS5 proxy the ``username`` and ``password`` portion
+of the ``proxy_url`` will be sent as the username/password to authenticate
+with the proxy:
+
+.. code-block:: python
+
+    proxy_url="socks5h://<username>:<password>@proxy-host"
+
+"""
+from __future__ import absolute_import
+
+try:
+    import socks
+except ImportError:
+    import warnings
+
+    from ..exceptions import DependencyWarning
+
+    warnings.warn(
+        (
+            "SOCKS support in urllib3 requires the installation of optional "
+            "dependencies: specifically, PySocks.  For more information, see "
+            "https://urllib3.readthedocs.io/en/1.26.x/contrib.html#socks-proxies"
+        ),
+        DependencyWarning,
+    )
+    raise
+
+from socket import error as SocketError
+from socket import timeout as SocketTimeout
+
+from ..connection import HTTPConnection, HTTPSConnection
+from ..connectionpool import HTTPConnectionPool, HTTPSConnectionPool
+from ..exceptions import ConnectTimeoutError, NewConnectionError
+from ..poolmanager import PoolManager
+from ..util.url import parse_url
+
+try:
+    import ssl
+except ImportError:
+    ssl = None
+
+
+class SOCKSConnection(HTTPConnection):
+    """
+    A plain-text HTTP connection that connects via a SOCKS proxy.
+    """
+
+    def __init__(self, *args, **kwargs):
+        self._socks_options = kwargs.pop("_socks_options")
+        super(SOCKSConnection, self).__init__(*args, **kwargs)
+
+    def _new_conn(self):
+        """
+        Establish a new connection via the SOCKS proxy.
+        """
+        extra_kw = {}
+        if self.source_address:
+            extra_kw["source_address"] = self.source_address
+
+        if self.socket_options:
+            extra_kw["socket_options"] = self.socket_options
+
+        try:
+            conn = socks.create_connection(
+                (self.host, self.port),
+                proxy_type=self._socks_options["socks_version"],
+                proxy_addr=self._socks_options["proxy_host"],
+                proxy_port=self._socks_options["proxy_port"],
+                proxy_username=self._socks_options["username"],
+                proxy_password=self._socks_options["password"],
+                proxy_rdns=self._socks_options["rdns"],
+                timeout=self.timeout,
+                **extra_kw
+            )
+
+        except SocketTimeout:
+            raise ConnectTimeoutError(
+                self,
+                "Connection to %s timed out. (connect timeout=%s)"
+                % (self.host, self.timeout),
+            )
+
+        except socks.ProxyError as e:
+            # This is fragile as hell, but it seems to be the only way to raise
+            # useful errors here.
+            if e.socket_err:
+                error = e.socket_err
+                if isinstance(error, SocketTimeout):
+                    raise ConnectTimeoutError(
+                        self,
+                        "Connection to %s timed out. (connect timeout=%s)"
+                        % (self.host, self.timeout),
+                    )
+                else:
+                    raise NewConnectionError(
+                        self, "Failed to establish a new connection: %s" % error
+                    )
+            else:
+                raise NewConnectionError(
+                    self, "Failed to establish a new connection: %s" % e
+                )
+
+        except SocketError as e:  # Defensive: PySocks should catch all these.
+            raise NewConnectionError(
+                self, "Failed to establish a new connection: %s" % e
+            )
+
+        return conn
+
+
+# We don't need to duplicate the Verified/Unverified distinction from
+# urllib3/connection.py here because the HTTPSConnection will already have been
+# correctly set to either the Verified or Unverified form by that module. This
+# means the SOCKSHTTPSConnection will automatically be the correct type.
+class SOCKSHTTPSConnection(SOCKSConnection, HTTPSConnection):
+    pass
+
+
+class SOCKSHTTPConnectionPool(HTTPConnectionPool):
+    ConnectionCls = SOCKSConnection
+
+
+class SOCKSHTTPSConnectionPool(HTTPSConnectionPool):
+    ConnectionCls = SOCKSHTTPSConnection
+
+
+class SOCKSProxyManager(PoolManager):
+    """
+    A version of the urllib3 ProxyManager that routes connections via the
+    defined SOCKS proxy.
+    """
+
+    pool_classes_by_scheme = {
+        "http": SOCKSHTTPConnectionPool,
+        "https": SOCKSHTTPSConnectionPool,
+    }
+
+    def __init__(
+        self,
+        proxy_url,
+        username=None,
+        password=None,
+        num_pools=10,
+        headers=None,
+        **connection_pool_kw
+    ):
+        parsed = parse_url(proxy_url)
+
+        if username is None and password is None and parsed.auth is not None:
+            split = parsed.auth.split(":")
+            if len(split) == 2:
+                username, password = split
+        if parsed.scheme == "socks5":
+            socks_version = socks.PROXY_TYPE_SOCKS5
+            rdns = False
+        elif parsed.scheme == "socks5h":
+            socks_version = socks.PROXY_TYPE_SOCKS5
+            rdns = True
+        elif parsed.scheme == "socks4":
+            socks_version = socks.PROXY_TYPE_SOCKS4
+            rdns = False
+        elif parsed.scheme == "socks4a":
+            socks_version = socks.PROXY_TYPE_SOCKS4
+            rdns = True
+        else:
+            raise ValueError("Unable to determine SOCKS version from %s" % proxy_url)
+
+        self.proxy_url = proxy_url
+
+        socks_options = {
+            "socks_version": socks_version,
+            "proxy_host": parsed.host,
+            "proxy_port": parsed.port,
+            "username": username,
+            "password": password,
+            "rdns": rdns,
+        }
+        connection_pool_kw["_socks_options"] = socks_options
+
+        super(SOCKSProxyManager, self).__init__(
+            num_pools, headers, **connection_pool_kw
+        )
+
+        self.pool_classes_by_scheme = SOCKSProxyManager.pool_classes_by_scheme
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/exceptions.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/exceptions.py
new file mode 100644
index 00000000..cba6f3f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/exceptions.py
@@ -0,0 +1,323 @@
+from __future__ import absolute_import
+
+from .packages.six.moves.http_client import IncompleteRead as httplib_IncompleteRead
+
+# Base Exceptions
+
+
+class HTTPError(Exception):
+    """Base exception used by this module."""
+
+    pass
+
+
+class HTTPWarning(Warning):
+    """Base warning used by this module."""
+
+    pass
+
+
+class PoolError(HTTPError):
+    """Base exception for errors caused within a pool."""
+
+    def __init__(self, pool, message):
+        self.pool = pool
+        HTTPError.__init__(self, "%s: %s" % (pool, message))
+
+    def __reduce__(self):
+        # For pickling purposes.
+        return self.__class__, (None, None)
+
+
+class RequestError(PoolError):
+    """Base exception for PoolErrors that have associated URLs."""
+
+    def __init__(self, pool, url, message):
+        self.url = url
+        PoolError.__init__(self, pool, message)
+
+    def __reduce__(self):
+        # For pickling purposes.
+        return self.__class__, (None, self.url, None)
+
+
+class SSLError(HTTPError):
+    """Raised when SSL certificate fails in an HTTPS connection."""
+
+    pass
+
+
+class ProxyError(HTTPError):
+    """Raised when the connection to a proxy fails."""
+
+    def __init__(self, message, error, *args):
+        super(ProxyError, self).__init__(message, error, *args)
+        self.original_error = error
+
+
+class DecodeError(HTTPError):
+    """Raised when automatic decoding based on Content-Type fails."""
+
+    pass
+
+
+class ProtocolError(HTTPError):
+    """Raised when something unexpected happens mid-request/response."""
+
+    pass
+
+
+#: Renamed to ProtocolError but aliased for backwards compatibility.
+ConnectionError = ProtocolError
+
+
+# Leaf Exceptions
+
+
+class MaxRetryError(RequestError):
+    """Raised when the maximum number of retries is exceeded.
+
+    :param pool: The connection pool
+    :type pool: :class:`~urllib3.connectionpool.HTTPConnectionPool`
+    :param string url: The requested Url
+    :param exceptions.Exception reason: The underlying error
+
+    """
+
+    def __init__(self, pool, url, reason=None):
+        self.reason = reason
+
+        message = "Max retries exceeded with url: %s (Caused by %r)" % (url, reason)
+
+        RequestError.__init__(self, pool, url, message)
+
+
+class HostChangedError(RequestError):
+    """Raised when an existing pool gets a request for a foreign host."""
+
+    def __init__(self, pool, url, retries=3):
+        message = "Tried to open a foreign host with url: %s" % url
+        RequestError.__init__(self, pool, url, message)
+        self.retries = retries
+
+
+class TimeoutStateError(HTTPError):
+    """Raised when passing an invalid state to a timeout"""
+
+    pass
+
+
+class TimeoutError(HTTPError):
+    """Raised when a socket timeout error occurs.
+
+    Catching this error will catch both :exc:`ReadTimeoutErrors
+    <ReadTimeoutError>` and :exc:`ConnectTimeoutErrors <ConnectTimeoutError>`.
+    """
+
+    pass
+
+
+class ReadTimeoutError(TimeoutError, RequestError):
+    """Raised when a socket timeout occurs while receiving data from a server"""
+
+    pass
+
+
+# This timeout error does not have a URL attached and needs to inherit from the
+# base HTTPError
+class ConnectTimeoutError(TimeoutError):
+    """Raised when a socket timeout occurs while connecting to a server"""
+
+    pass
+
+
+class NewConnectionError(ConnectTimeoutError, PoolError):
+    """Raised when we fail to establish a new connection. Usually ECONNREFUSED."""
+
+    pass
+
+
+class EmptyPoolError(PoolError):
+    """Raised when a pool runs out of connections and no more are allowed."""
+
+    pass
+
+
+class ClosedPoolError(PoolError):
+    """Raised when a request enters a pool after the pool has been closed."""
+
+    pass
+
+
+class LocationValueError(ValueError, HTTPError):
+    """Raised when there is something wrong with a given URL input."""
+
+    pass
+
+
+class LocationParseError(LocationValueError):
+    """Raised when get_host or similar fails to parse the URL input."""
+
+    def __init__(self, location):
+        message = "Failed to parse: %s" % location
+        HTTPError.__init__(self, message)
+
+        self.location = location
+
+
+class URLSchemeUnknown(LocationValueError):
+    """Raised when a URL input has an unsupported scheme."""
+
+    def __init__(self, scheme):
+        message = "Not supported URL scheme %s" % scheme
+        super(URLSchemeUnknown, self).__init__(message)
+
+        self.scheme = scheme
+
+
+class ResponseError(HTTPError):
+    """Used as a container for an error reason supplied in a MaxRetryError."""
+
+    GENERIC_ERROR = "too many error responses"
+    SPECIFIC_ERROR = "too many {status_code} error responses"
+
+
+class SecurityWarning(HTTPWarning):
+    """Warned when performing security reducing actions"""
+
+    pass
+
+
+class SubjectAltNameWarning(SecurityWarning):
+    """Warned when connecting to a host with a certificate missing a SAN."""
+
+    pass
+
+
+class InsecureRequestWarning(SecurityWarning):
+    """Warned when making an unverified HTTPS request."""
+
+    pass
+
+
+class SystemTimeWarning(SecurityWarning):
+    """Warned when system time is suspected to be wrong"""
+
+    pass
+
+
+class InsecurePlatformWarning(SecurityWarning):
+    """Warned when certain TLS/SSL configuration is not available on a platform."""
+
+    pass
+
+
+class SNIMissingWarning(HTTPWarning):
+    """Warned when making a HTTPS request without SNI available."""
+
+    pass
+
+
+class DependencyWarning(HTTPWarning):
+    """
+    Warned when an attempt is made to import a module with missing optional
+    dependencies.
+    """
+
+    pass
+
+
+class ResponseNotChunked(ProtocolError, ValueError):
+    """Response needs to be chunked in order to read it as chunks."""
+
+    pass
+
+
+class BodyNotHttplibCompatible(HTTPError):
+    """
+    Body should be :class:`http.client.HTTPResponse` like
+    (have an fp attribute which returns raw chunks) for read_chunked().
+    """
+
+    pass
+
+
+class IncompleteRead(HTTPError, httplib_IncompleteRead):
+    """
+    Response length doesn't match expected Content-Length
+
+    Subclass of :class:`http.client.IncompleteRead` to allow int value
+    for ``partial`` to avoid creating large objects on streamed reads.
+    """
+
+    def __init__(self, partial, expected):
+        super(IncompleteRead, self).__init__(partial, expected)
+
+    def __repr__(self):
+        return "IncompleteRead(%i bytes read, %i more expected)" % (
+            self.partial,
+            self.expected,
+        )
+
+
+class InvalidChunkLength(HTTPError, httplib_IncompleteRead):
+    """Invalid chunk length in a chunked response."""
+
+    def __init__(self, response, length):
+        super(InvalidChunkLength, self).__init__(
+            response.tell(), response.length_remaining
+        )
+        self.response = response
+        self.length = length
+
+    def __repr__(self):
+        return "InvalidChunkLength(got length %r, %i bytes read)" % (
+            self.length,
+            self.partial,
+        )
+
+
+class InvalidHeader(HTTPError):
+    """The header provided was somehow invalid."""
+
+    pass
+
+
+class ProxySchemeUnknown(AssertionError, URLSchemeUnknown):
+    """ProxyManager does not support the supplied scheme"""
+
+    # TODO(t-8ch): Stop inheriting from AssertionError in v2.0.
+
+    def __init__(self, scheme):
+        # 'localhost' is here because our URL parser parses
+        # localhost:8080 -> scheme=localhost, remove if we fix this.
+        if scheme == "localhost":
+            scheme = None
+        if scheme is None:
+            message = "Proxy URL had no scheme, should start with http:// or https://"
+        else:
+            message = (
+                "Proxy URL had unsupported scheme %s, should use http:// or https://"
+                % scheme
+            )
+        super(ProxySchemeUnknown, self).__init__(message)
+
+
+class ProxySchemeUnsupported(ValueError):
+    """Fetching HTTPS resources through HTTPS proxies is unsupported"""
+
+    pass
+
+
+class HeaderParsingError(HTTPError):
+    """Raised by assert_header_parsing, but we convert it to a log.warning statement."""
+
+    def __init__(self, defects, unparsed_data):
+        message = "%s, unparsed data: %r" % (defects or "Unknown", unparsed_data)
+        super(HeaderParsingError, self).__init__(message)
+
+
+class UnrewindableBodyError(HTTPError):
+    """urllib3 encountered an error when trying to rewind a body"""
+
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/fields.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/fields.py
new file mode 100644
index 00000000..9d630f49
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/fields.py
@@ -0,0 +1,274 @@
+from __future__ import absolute_import
+
+import email.utils
+import mimetypes
+import re
+
+from .packages import six
+
+
+def guess_content_type(filename, default="application/octet-stream"):
+    """
+    Guess the "Content-Type" of a file.
+
+    :param filename:
+        The filename to guess the "Content-Type" of using :mod:`mimetypes`.
+    :param default:
+        If no "Content-Type" can be guessed, default to `default`.
+    """
+    if filename:
+        return mimetypes.guess_type(filename)[0] or default
+    return default
+
+
+def format_header_param_rfc2231(name, value):
+    """
+    Helper function to format and quote a single header parameter using the
+    strategy defined in RFC 2231.
+
+    Particularly useful for header parameters which might contain
+    non-ASCII values, like file names. This follows
+    `RFC 2388 Section 4.4 <https://tools.ietf.org/html/rfc2388#section-4.4>`_.
+
+    :param name:
+        The name of the parameter, a string expected to be ASCII only.
+    :param value:
+        The value of the parameter, provided as ``bytes`` or `str``.
+    :ret:
+        An RFC-2231-formatted unicode string.
+    """
+    if isinstance(value, six.binary_type):
+        value = value.decode("utf-8")
+
+    if not any(ch in value for ch in '"\\\r\n'):
+        result = u'%s="%s"' % (name, value)
+        try:
+            result.encode("ascii")
+        except (UnicodeEncodeError, UnicodeDecodeError):
+            pass
+        else:
+            return result
+
+    if six.PY2:  # Python 2:
+        value = value.encode("utf-8")
+
+    # encode_rfc2231 accepts an encoded string and returns an ascii-encoded
+    # string in Python 2 but accepts and returns unicode strings in Python 3
+    value = email.utils.encode_rfc2231(value, "utf-8")
+    value = "%s*=%s" % (name, value)
+
+    if six.PY2:  # Python 2:
+        value = value.decode("utf-8")
+
+    return value
+
+
+_HTML5_REPLACEMENTS = {
+    u"\u0022": u"%22",
+    # Replace "\" with "\\".
+    u"\u005C": u"\u005C\u005C",
+}
+
+# All control characters from 0x00 to 0x1F *except* 0x1B.
+_HTML5_REPLACEMENTS.update(
+    {
+        six.unichr(cc): u"%{:02X}".format(cc)
+        for cc in range(0x00, 0x1F + 1)
+        if cc not in (0x1B,)
+    }
+)
+
+
+def _replace_multiple(value, needles_and_replacements):
+    def replacer(match):
+        return needles_and_replacements[match.group(0)]
+
+    pattern = re.compile(
+        r"|".join([re.escape(needle) for needle in needles_and_replacements.keys()])
+    )
+
+    result = pattern.sub(replacer, value)
+
+    return result
+
+
+def format_header_param_html5(name, value):
+    """
+    Helper function to format and quote a single header parameter using the
+    HTML5 strategy.
+
+    Particularly useful for header parameters which might contain
+    non-ASCII values, like file names. This follows the `HTML5 Working Draft
+    Section 4.10.22.7`_ and matches the behavior of curl and modern browsers.
+
+    .. _HTML5 Working Draft Section 4.10.22.7:
+        https://w3c.github.io/html/sec-forms.html#multipart-form-data
+
+    :param name:
+        The name of the parameter, a string expected to be ASCII only.
+    :param value:
+        The value of the parameter, provided as ``bytes`` or `str``.
+    :ret:
+        A unicode string, stripped of troublesome characters.
+    """
+    if isinstance(value, six.binary_type):
+        value = value.decode("utf-8")
+
+    value = _replace_multiple(value, _HTML5_REPLACEMENTS)
+
+    return u'%s="%s"' % (name, value)
+
+
+# For backwards-compatibility.
+format_header_param = format_header_param_html5
+
+
+class RequestField(object):
+    """
+    A data container for request body parameters.
+
+    :param name:
+        The name of this request field. Must be unicode.
+    :param data:
+        The data/value body.
+    :param filename:
+        An optional filename of the request field. Must be unicode.
+    :param headers:
+        An optional dict-like object of headers to initially use for the field.
+    :param header_formatter:
+        An optional callable that is used to encode and format the headers. By
+        default, this is :func:`format_header_param_html5`.
+    """
+
+    def __init__(
+        self,
+        name,
+        data,
+        filename=None,
+        headers=None,
+        header_formatter=format_header_param_html5,
+    ):
+        self._name = name
+        self._filename = filename
+        self.data = data
+        self.headers = {}
+        if headers:
+            self.headers = dict(headers)
+        self.header_formatter = header_formatter
+
+    @classmethod
+    def from_tuples(cls, fieldname, value, header_formatter=format_header_param_html5):
+        """
+        A :class:`~urllib3.fields.RequestField` factory from old-style tuple parameters.
+
+        Supports constructing :class:`~urllib3.fields.RequestField` from
+        parameter of key/value strings AND key/filetuple. A filetuple is a
+        (filename, data, MIME type) tuple where the MIME type is optional.
+        For example::
+
+            'foo': 'bar',
+            'fakefile': ('foofile.txt', 'contents of foofile'),
+            'realfile': ('barfile.txt', open('realfile').read()),
+            'typedfile': ('bazfile.bin', open('bazfile').read(), 'image/jpeg'),
+            'nonamefile': 'contents of nonamefile field',
+
+        Field names and filenames must be unicode.
+        """
+        if isinstance(value, tuple):
+            if len(value) == 3:
+                filename, data, content_type = value
+            else:
+                filename, data = value
+                content_type = guess_content_type(filename)
+        else:
+            filename = None
+            content_type = None
+            data = value
+
+        request_param = cls(
+            fieldname, data, filename=filename, header_formatter=header_formatter
+        )
+        request_param.make_multipart(content_type=content_type)
+
+        return request_param
+
+    def _render_part(self, name, value):
+        """
+        Overridable helper function to format a single header parameter. By
+        default, this calls ``self.header_formatter``.
+
+        :param name:
+            The name of the parameter, a string expected to be ASCII only.
+        :param value:
+            The value of the parameter, provided as a unicode string.
+        """
+
+        return self.header_formatter(name, value)
+
+    def _render_parts(self, header_parts):
+        """
+        Helper function to format and quote a single header.
+
+        Useful for single headers that are composed of multiple items. E.g.,
+        'Content-Disposition' fields.
+
+        :param header_parts:
+            A sequence of (k, v) tuples or a :class:`dict` of (k, v) to format
+            as `k1="v1"; k2="v2"; ...`.
+        """
+        parts = []
+        iterable = header_parts
+        if isinstance(header_parts, dict):
+            iterable = header_parts.items()
+
+        for name, value in iterable:
+            if value is not None:
+                parts.append(self._render_part(name, value))
+
+        return u"; ".join(parts)
+
+    def render_headers(self):
+        """
+        Renders the headers for this request field.
+        """
+        lines = []
+
+        sort_keys = ["Content-Disposition", "Content-Type", "Content-Location"]
+        for sort_key in sort_keys:
+            if self.headers.get(sort_key, False):
+                lines.append(u"%s: %s" % (sort_key, self.headers[sort_key]))
+
+        for header_name, header_value in self.headers.items():
+            if header_name not in sort_keys:
+                if header_value:
+                    lines.append(u"%s: %s" % (header_name, header_value))
+
+        lines.append(u"\r\n")
+        return u"\r\n".join(lines)
+
+    def make_multipart(
+        self, content_disposition=None, content_type=None, content_location=None
+    ):
+        """
+        Makes this request field into a multipart request field.
+
+        This method overrides "Content-Disposition", "Content-Type" and
+        "Content-Location" headers to the request parameter.
+
+        :param content_type:
+            The 'Content-Type' of the request body.
+        :param content_location:
+            The 'Content-Location' of the request body.
+
+        """
+        self.headers["Content-Disposition"] = content_disposition or u"form-data"
+        self.headers["Content-Disposition"] += u"; ".join(
+            [
+                u"",
+                self._render_parts(
+                    ((u"name", self._name), (u"filename", self._filename))
+                ),
+            ]
+        )
+        self.headers["Content-Type"] = content_type
+        self.headers["Content-Location"] = content_location
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/filepost.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/filepost.py
new file mode 100644
index 00000000..36c9252c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/filepost.py
@@ -0,0 +1,98 @@
+from __future__ import absolute_import
+
+import binascii
+import codecs
+import os
+from io import BytesIO
+
+from .fields import RequestField
+from .packages import six
+from .packages.six import b
+
+writer = codecs.lookup("utf-8")[3]
+
+
+def choose_boundary():
+    """
+    Our embarrassingly-simple replacement for mimetools.choose_boundary.
+    """
+    boundary = binascii.hexlify(os.urandom(16))
+    if not six.PY2:
+        boundary = boundary.decode("ascii")
+    return boundary
+
+
+def iter_field_objects(fields):
+    """
+    Iterate over fields.
+
+    Supports list of (k, v) tuples and dicts, and lists of
+    :class:`~urllib3.fields.RequestField`.
+
+    """
+    if isinstance(fields, dict):
+        i = six.iteritems(fields)
+    else:
+        i = iter(fields)
+
+    for field in i:
+        if isinstance(field, RequestField):
+            yield field
+        else:
+            yield RequestField.from_tuples(*field)
+
+
+def iter_fields(fields):
+    """
+    .. deprecated:: 1.6
+
+    Iterate over fields.
+
+    The addition of :class:`~urllib3.fields.RequestField` makes this function
+    obsolete. Instead, use :func:`iter_field_objects`, which returns
+    :class:`~urllib3.fields.RequestField` objects.
+
+    Supports list of (k, v) tuples and dicts.
+    """
+    if isinstance(fields, dict):
+        return ((k, v) for k, v in six.iteritems(fields))
+
+    return ((k, v) for k, v in fields)
+
+
+def encode_multipart_formdata(fields, boundary=None):
+    """
+    Encode a dictionary of ``fields`` using the multipart/form-data MIME format.
+
+    :param fields:
+        Dictionary of fields or list of (key, :class:`~urllib3.fields.RequestField`).
+
+    :param boundary:
+        If not specified, then a random boundary will be generated using
+        :func:`urllib3.filepost.choose_boundary`.
+    """
+    body = BytesIO()
+    if boundary is None:
+        boundary = choose_boundary()
+
+    for field in iter_field_objects(fields):
+        body.write(b("--%s\r\n" % (boundary)))
+
+        writer(body).write(field.render_headers())
+        data = field.data
+
+        if isinstance(data, int):
+            data = str(data)  # Backwards compatibility
+
+        if isinstance(data, six.text_type):
+            writer(body).write(data)
+        else:
+            body.write(data)
+
+        body.write(b"\r\n")
+
+    body.write(b("--%s--\r\n" % (boundary)))
+
+    content_type = str("multipart/form-data; boundary=%s" % boundary)
+
+    return body.getvalue(), content_type
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..5c70a0d5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/six.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/six.cpython-312.pyc
new file mode 100644
index 00000000..7f7af1d7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/__pycache__/six.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ca73098a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/makefile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/makefile.cpython-312.pyc
new file mode 100644
index 00000000..c76085ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/makefile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/weakref_finalize.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/weakref_finalize.cpython-312.pyc
new file mode 100644
index 00000000..8579bb9d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/__pycache__/weakref_finalize.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/makefile.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/makefile.py
new file mode 100644
index 00000000..b8fb2154
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/makefile.py
@@ -0,0 +1,51 @@
+# -*- coding: utf-8 -*-
+"""
+backports.makefile
+~~~~~~~~~~~~~~~~~~
+
+Backports the Python 3 ``socket.makefile`` method for use with anything that
+wants to create a "fake" socket object.
+"""
+import io
+from socket import SocketIO
+
+
+def backport_makefile(
+    self, mode="r", buffering=None, encoding=None, errors=None, newline=None
+):
+    """
+    Backport of ``socket.makefile`` from Python 3.5.
+    """
+    if not set(mode) <= {"r", "w", "b"}:
+        raise ValueError("invalid mode %r (only r, w, b allowed)" % (mode,))
+    writing = "w" in mode
+    reading = "r" in mode or not writing
+    assert reading or writing
+    binary = "b" in mode
+    rawmode = ""
+    if reading:
+        rawmode += "r"
+    if writing:
+        rawmode += "w"
+    raw = SocketIO(self, rawmode)
+    self._makefile_refs += 1
+    if buffering is None:
+        buffering = -1
+    if buffering < 0:
+        buffering = io.DEFAULT_BUFFER_SIZE
+    if buffering == 0:
+        if not binary:
+            raise ValueError("unbuffered streams must be binary")
+        return raw
+    if reading and writing:
+        buffer = io.BufferedRWPair(raw, raw, buffering)
+    elif reading:
+        buffer = io.BufferedReader(raw, buffering)
+    else:
+        assert writing
+        buffer = io.BufferedWriter(raw, buffering)
+    if binary:
+        return buffer
+    text = io.TextIOWrapper(buffer, encoding, errors, newline)
+    text.mode = mode
+    return text
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/weakref_finalize.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/weakref_finalize.py
new file mode 100644
index 00000000..a2f2966e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/backports/weakref_finalize.py
@@ -0,0 +1,155 @@
+# -*- coding: utf-8 -*-
+"""
+backports.weakref_finalize
+~~~~~~~~~~~~~~~~~~
+
+Backports the Python 3 ``weakref.finalize`` method.
+"""
+from __future__ import absolute_import
+
+import itertools
+import sys
+from weakref import ref
+
+__all__ = ["weakref_finalize"]
+
+
+class weakref_finalize(object):
+    """Class for finalization of weakrefable objects
+    finalize(obj, func, *args, **kwargs) returns a callable finalizer
+    object which will be called when obj is garbage collected. The
+    first time the finalizer is called it evaluates func(*arg, **kwargs)
+    and returns the result. After this the finalizer is dead, and
+    calling it just returns None.
+    When the program exits any remaining finalizers for which the
+    atexit attribute is true will be run in reverse order of creation.
+    By default atexit is true.
+    """
+
+    # Finalizer objects don't have any state of their own.  They are
+    # just used as keys to lookup _Info objects in the registry.  This
+    # ensures that they cannot be part of a ref-cycle.
+
+    __slots__ = ()
+    _registry = {}
+    _shutdown = False
+    _index_iter = itertools.count()
+    _dirty = False
+    _registered_with_atexit = False
+
+    class _Info(object):
+        __slots__ = ("weakref", "func", "args", "kwargs", "atexit", "index")
+
+    def __init__(self, obj, func, *args, **kwargs):
+        if not self._registered_with_atexit:
+            # We may register the exit function more than once because
+            # of a thread race, but that is harmless
+            import atexit
+
+            atexit.register(self._exitfunc)
+            weakref_finalize._registered_with_atexit = True
+        info = self._Info()
+        info.weakref = ref(obj, self)
+        info.func = func
+        info.args = args
+        info.kwargs = kwargs or None
+        info.atexit = True
+        info.index = next(self._index_iter)
+        self._registry[self] = info
+        weakref_finalize._dirty = True
+
+    def __call__(self, _=None):
+        """If alive then mark as dead and return func(*args, **kwargs);
+        otherwise return None"""
+        info = self._registry.pop(self, None)
+        if info and not self._shutdown:
+            return info.func(*info.args, **(info.kwargs or {}))
+
+    def detach(self):
+        """If alive then mark as dead and return (obj, func, args, kwargs);
+        otherwise return None"""
+        info = self._registry.get(self)
+        obj = info and info.weakref()
+        if obj is not None and self._registry.pop(self, None):
+            return (obj, info.func, info.args, info.kwargs or {})
+
+    def peek(self):
+        """If alive then return (obj, func, args, kwargs);
+        otherwise return None"""
+        info = self._registry.get(self)
+        obj = info and info.weakref()
+        if obj is not None:
+            return (obj, info.func, info.args, info.kwargs or {})
+
+    @property
+    def alive(self):
+        """Whether finalizer is alive"""
+        return self in self._registry
+
+    @property
+    def atexit(self):
+        """Whether finalizer should be called at exit"""
+        info = self._registry.get(self)
+        return bool(info) and info.atexit
+
+    @atexit.setter
+    def atexit(self, value):
+        info = self._registry.get(self)
+        if info:
+            info.atexit = bool(value)
+
+    def __repr__(self):
+        info = self._registry.get(self)
+        obj = info and info.weakref()
+        if obj is None:
+            return "<%s object at %#x; dead>" % (type(self).__name__, id(self))
+        else:
+            return "<%s object at %#x; for %r at %#x>" % (
+                type(self).__name__,
+                id(self),
+                type(obj).__name__,
+                id(obj),
+            )
+
+    @classmethod
+    def _select_for_exit(cls):
+        # Return live finalizers marked for exit, oldest first
+        L = [(f, i) for (f, i) in cls._registry.items() if i.atexit]
+        L.sort(key=lambda item: item[1].index)
+        return [f for (f, i) in L]
+
+    @classmethod
+    def _exitfunc(cls):
+        # At shutdown invoke finalizers for which atexit is true.
+        # This is called once all other non-daemonic threads have been
+        # joined.
+        reenable_gc = False
+        try:
+            if cls._registry:
+                import gc
+
+                if gc.isenabled():
+                    reenable_gc = True
+                    gc.disable()
+                pending = None
+                while True:
+                    if pending is None or weakref_finalize._dirty:
+                        pending = cls._select_for_exit()
+                        weakref_finalize._dirty = False
+                    if not pending:
+                        break
+                    f = pending.pop()
+                    try:
+                        # gc is disabled, so (assuming no daemonic
+                        # threads) the following is the only line in
+                        # this function which might trigger creation
+                        # of a new finalizer
+                        f()
+                    except Exception:
+                        sys.excepthook(*sys.exc_info())
+                    assert f not in cls._registry
+        finally:
+            # prevent any more finalizers from executing during shutdown
+            weakref_finalize._shutdown = True
+            if reenable_gc:
+                gc.enable()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/six.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/six.py
new file mode 100644
index 00000000..f099a3dc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/packages/six.py
@@ -0,0 +1,1076 @@
+# Copyright (c) 2010-2020 Benjamin Peterson
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+"""Utilities for writing code that runs on Python 2 and 3"""
+
+from __future__ import absolute_import
+
+import functools
+import itertools
+import operator
+import sys
+import types
+
+__author__ = "Benjamin Peterson <benjamin@python.org>"
+__version__ = "1.16.0"
+
+
+# Useful for very coarse version differentiation.
+PY2 = sys.version_info[0] == 2
+PY3 = sys.version_info[0] == 3
+PY34 = sys.version_info[0:2] >= (3, 4)
+
+if PY3:
+    string_types = (str,)
+    integer_types = (int,)
+    class_types = (type,)
+    text_type = str
+    binary_type = bytes
+
+    MAXSIZE = sys.maxsize
+else:
+    string_types = (basestring,)
+    integer_types = (int, long)
+    class_types = (type, types.ClassType)
+    text_type = unicode
+    binary_type = str
+
+    if sys.platform.startswith("java"):
+        # Jython always uses 32 bits.
+        MAXSIZE = int((1 << 31) - 1)
+    else:
+        # It's possible to have sizeof(long) != sizeof(Py_ssize_t).
+        class X(object):
+            def __len__(self):
+                return 1 << 31
+
+        try:
+            len(X())
+        except OverflowError:
+            # 32-bit
+            MAXSIZE = int((1 << 31) - 1)
+        else:
+            # 64-bit
+            MAXSIZE = int((1 << 63) - 1)
+        del X
+
+if PY34:
+    from importlib.util import spec_from_loader
+else:
+    spec_from_loader = None
+
+
+def _add_doc(func, doc):
+    """Add documentation to a function."""
+    func.__doc__ = doc
+
+
+def _import_module(name):
+    """Import module, returning the module after the last dot."""
+    __import__(name)
+    return sys.modules[name]
+
+
+class _LazyDescr(object):
+    def __init__(self, name):
+        self.name = name
+
+    def __get__(self, obj, tp):
+        result = self._resolve()
+        setattr(obj, self.name, result)  # Invokes __set__.
+        try:
+            # This is a bit ugly, but it avoids running this again by
+            # removing this descriptor.
+            delattr(obj.__class__, self.name)
+        except AttributeError:
+            pass
+        return result
+
+
+class MovedModule(_LazyDescr):
+    def __init__(self, name, old, new=None):
+        super(MovedModule, self).__init__(name)
+        if PY3:
+            if new is None:
+                new = name
+            self.mod = new
+        else:
+            self.mod = old
+
+    def _resolve(self):
+        return _import_module(self.mod)
+
+    def __getattr__(self, attr):
+        _module = self._resolve()
+        value = getattr(_module, attr)
+        setattr(self, attr, value)
+        return value
+
+
+class _LazyModule(types.ModuleType):
+    def __init__(self, name):
+        super(_LazyModule, self).__init__(name)
+        self.__doc__ = self.__class__.__doc__
+
+    def __dir__(self):
+        attrs = ["__doc__", "__name__"]
+        attrs += [attr.name for attr in self._moved_attributes]
+        return attrs
+
+    # Subclasses should override this
+    _moved_attributes = []
+
+
+class MovedAttribute(_LazyDescr):
+    def __init__(self, name, old_mod, new_mod, old_attr=None, new_attr=None):
+        super(MovedAttribute, self).__init__(name)
+        if PY3:
+            if new_mod is None:
+                new_mod = name
+            self.mod = new_mod
+            if new_attr is None:
+                if old_attr is None:
+                    new_attr = name
+                else:
+                    new_attr = old_attr
+            self.attr = new_attr
+        else:
+            self.mod = old_mod
+            if old_attr is None:
+                old_attr = name
+            self.attr = old_attr
+
+    def _resolve(self):
+        module = _import_module(self.mod)
+        return getattr(module, self.attr)
+
+
+class _SixMetaPathImporter(object):
+
+    """
+    A meta path importer to import six.moves and its submodules.
+
+    This class implements a PEP302 finder and loader. It should be compatible
+    with Python 2.5 and all existing versions of Python3
+    """
+
+    def __init__(self, six_module_name):
+        self.name = six_module_name
+        self.known_modules = {}
+
+    def _add_module(self, mod, *fullnames):
+        for fullname in fullnames:
+            self.known_modules[self.name + "." + fullname] = mod
+
+    def _get_module(self, fullname):
+        return self.known_modules[self.name + "." + fullname]
+
+    def find_module(self, fullname, path=None):
+        if fullname in self.known_modules:
+            return self
+        return None
+
+    def find_spec(self, fullname, path, target=None):
+        if fullname in self.known_modules:
+            return spec_from_loader(fullname, self)
+        return None
+
+    def __get_module(self, fullname):
+        try:
+            return self.known_modules[fullname]
+        except KeyError:
+            raise ImportError("This loader does not know module " + fullname)
+
+    def load_module(self, fullname):
+        try:
+            # in case of a reload
+            return sys.modules[fullname]
+        except KeyError:
+            pass
+        mod = self.__get_module(fullname)
+        if isinstance(mod, MovedModule):
+            mod = mod._resolve()
+        else:
+            mod.__loader__ = self
+        sys.modules[fullname] = mod
+        return mod
+
+    def is_package(self, fullname):
+        """
+        Return true, if the named module is a package.
+
+        We need this method to get correct spec objects with
+        Python 3.4 (see PEP451)
+        """
+        return hasattr(self.__get_module(fullname), "__path__")
+
+    def get_code(self, fullname):
+        """Return None
+
+        Required, if is_package is implemented"""
+        self.__get_module(fullname)  # eventually raises ImportError
+        return None
+
+    get_source = get_code  # same as get_code
+
+    def create_module(self, spec):
+        return self.load_module(spec.name)
+
+    def exec_module(self, module):
+        pass
+
+
+_importer = _SixMetaPathImporter(__name__)
+
+
+class _MovedItems(_LazyModule):
+
+    """Lazy loading of moved objects"""
+
+    __path__ = []  # mark as package
+
+
+_moved_attributes = [
+    MovedAttribute("cStringIO", "cStringIO", "io", "StringIO"),
+    MovedAttribute("filter", "itertools", "builtins", "ifilter", "filter"),
+    MovedAttribute(
+        "filterfalse", "itertools", "itertools", "ifilterfalse", "filterfalse"
+    ),
+    MovedAttribute("input", "__builtin__", "builtins", "raw_input", "input"),
+    MovedAttribute("intern", "__builtin__", "sys"),
+    MovedAttribute("map", "itertools", "builtins", "imap", "map"),
+    MovedAttribute("getcwd", "os", "os", "getcwdu", "getcwd"),
+    MovedAttribute("getcwdb", "os", "os", "getcwd", "getcwdb"),
+    MovedAttribute("getoutput", "commands", "subprocess"),
+    MovedAttribute("range", "__builtin__", "builtins", "xrange", "range"),
+    MovedAttribute(
+        "reload_module", "__builtin__", "importlib" if PY34 else "imp", "reload"
+    ),
+    MovedAttribute("reduce", "__builtin__", "functools"),
+    MovedAttribute("shlex_quote", "pipes", "shlex", "quote"),
+    MovedAttribute("StringIO", "StringIO", "io"),
+    MovedAttribute("UserDict", "UserDict", "collections"),
+    MovedAttribute("UserList", "UserList", "collections"),
+    MovedAttribute("UserString", "UserString", "collections"),
+    MovedAttribute("xrange", "__builtin__", "builtins", "xrange", "range"),
+    MovedAttribute("zip", "itertools", "builtins", "izip", "zip"),
+    MovedAttribute(
+        "zip_longest", "itertools", "itertools", "izip_longest", "zip_longest"
+    ),
+    MovedModule("builtins", "__builtin__"),
+    MovedModule("configparser", "ConfigParser"),
+    MovedModule(
+        "collections_abc",
+        "collections",
+        "collections.abc" if sys.version_info >= (3, 3) else "collections",
+    ),
+    MovedModule("copyreg", "copy_reg"),
+    MovedModule("dbm_gnu", "gdbm", "dbm.gnu"),
+    MovedModule("dbm_ndbm", "dbm", "dbm.ndbm"),
+    MovedModule(
+        "_dummy_thread",
+        "dummy_thread",
+        "_dummy_thread" if sys.version_info < (3, 9) else "_thread",
+    ),
+    MovedModule("http_cookiejar", "cookielib", "http.cookiejar"),
+    MovedModule("http_cookies", "Cookie", "http.cookies"),
+    MovedModule("html_entities", "htmlentitydefs", "html.entities"),
+    MovedModule("html_parser", "HTMLParser", "html.parser"),
+    MovedModule("http_client", "httplib", "http.client"),
+    MovedModule("email_mime_base", "email.MIMEBase", "email.mime.base"),
+    MovedModule("email_mime_image", "email.MIMEImage", "email.mime.image"),
+    MovedModule("email_mime_multipart", "email.MIMEMultipart", "email.mime.multipart"),
+    MovedModule(
+        "email_mime_nonmultipart", "email.MIMENonMultipart", "email.mime.nonmultipart"
+    ),
+    MovedModule("email_mime_text", "email.MIMEText", "email.mime.text"),
+    MovedModule("BaseHTTPServer", "BaseHTTPServer", "http.server"),
+    MovedModule("CGIHTTPServer", "CGIHTTPServer", "http.server"),
+    MovedModule("SimpleHTTPServer", "SimpleHTTPServer", "http.server"),
+    MovedModule("cPickle", "cPickle", "pickle"),
+    MovedModule("queue", "Queue"),
+    MovedModule("reprlib", "repr"),
+    MovedModule("socketserver", "SocketServer"),
+    MovedModule("_thread", "thread", "_thread"),
+    MovedModule("tkinter", "Tkinter"),
+    MovedModule("tkinter_dialog", "Dialog", "tkinter.dialog"),
+    MovedModule("tkinter_filedialog", "FileDialog", "tkinter.filedialog"),
+    MovedModule("tkinter_scrolledtext", "ScrolledText", "tkinter.scrolledtext"),
+    MovedModule("tkinter_simpledialog", "SimpleDialog", "tkinter.simpledialog"),
+    MovedModule("tkinter_tix", "Tix", "tkinter.tix"),
+    MovedModule("tkinter_ttk", "ttk", "tkinter.ttk"),
+    MovedModule("tkinter_constants", "Tkconstants", "tkinter.constants"),
+    MovedModule("tkinter_dnd", "Tkdnd", "tkinter.dnd"),
+    MovedModule("tkinter_colorchooser", "tkColorChooser", "tkinter.colorchooser"),
+    MovedModule("tkinter_commondialog", "tkCommonDialog", "tkinter.commondialog"),
+    MovedModule("tkinter_tkfiledialog", "tkFileDialog", "tkinter.filedialog"),
+    MovedModule("tkinter_font", "tkFont", "tkinter.font"),
+    MovedModule("tkinter_messagebox", "tkMessageBox", "tkinter.messagebox"),
+    MovedModule("tkinter_tksimpledialog", "tkSimpleDialog", "tkinter.simpledialog"),
+    MovedModule("urllib_parse", __name__ + ".moves.urllib_parse", "urllib.parse"),
+    MovedModule("urllib_error", __name__ + ".moves.urllib_error", "urllib.error"),
+    MovedModule("urllib", __name__ + ".moves.urllib", __name__ + ".moves.urllib"),
+    MovedModule("urllib_robotparser", "robotparser", "urllib.robotparser"),
+    MovedModule("xmlrpc_client", "xmlrpclib", "xmlrpc.client"),
+    MovedModule("xmlrpc_server", "SimpleXMLRPCServer", "xmlrpc.server"),
+]
+# Add windows specific modules.
+if sys.platform == "win32":
+    _moved_attributes += [
+        MovedModule("winreg", "_winreg"),
+    ]
+
+for attr in _moved_attributes:
+    setattr(_MovedItems, attr.name, attr)
+    if isinstance(attr, MovedModule):
+        _importer._add_module(attr, "moves." + attr.name)
+del attr
+
+_MovedItems._moved_attributes = _moved_attributes
+
+moves = _MovedItems(__name__ + ".moves")
+_importer._add_module(moves, "moves")
+
+
+class Module_six_moves_urllib_parse(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_parse"""
+
+
+_urllib_parse_moved_attributes = [
+    MovedAttribute("ParseResult", "urlparse", "urllib.parse"),
+    MovedAttribute("SplitResult", "urlparse", "urllib.parse"),
+    MovedAttribute("parse_qs", "urlparse", "urllib.parse"),
+    MovedAttribute("parse_qsl", "urlparse", "urllib.parse"),
+    MovedAttribute("urldefrag", "urlparse", "urllib.parse"),
+    MovedAttribute("urljoin", "urlparse", "urllib.parse"),
+    MovedAttribute("urlparse", "urlparse", "urllib.parse"),
+    MovedAttribute("urlsplit", "urlparse", "urllib.parse"),
+    MovedAttribute("urlunparse", "urlparse", "urllib.parse"),
+    MovedAttribute("urlunsplit", "urlparse", "urllib.parse"),
+    MovedAttribute("quote", "urllib", "urllib.parse"),
+    MovedAttribute("quote_plus", "urllib", "urllib.parse"),
+    MovedAttribute("unquote", "urllib", "urllib.parse"),
+    MovedAttribute("unquote_plus", "urllib", "urllib.parse"),
+    MovedAttribute(
+        "unquote_to_bytes", "urllib", "urllib.parse", "unquote", "unquote_to_bytes"
+    ),
+    MovedAttribute("urlencode", "urllib", "urllib.parse"),
+    MovedAttribute("splitquery", "urllib", "urllib.parse"),
+    MovedAttribute("splittag", "urllib", "urllib.parse"),
+    MovedAttribute("splituser", "urllib", "urllib.parse"),
+    MovedAttribute("splitvalue", "urllib", "urllib.parse"),
+    MovedAttribute("uses_fragment", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_netloc", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_params", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_query", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_relative", "urlparse", "urllib.parse"),
+]
+for attr in _urllib_parse_moved_attributes:
+    setattr(Module_six_moves_urllib_parse, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_parse._moved_attributes = _urllib_parse_moved_attributes
+
+_importer._add_module(
+    Module_six_moves_urllib_parse(__name__ + ".moves.urllib_parse"),
+    "moves.urllib_parse",
+    "moves.urllib.parse",
+)
+
+
+class Module_six_moves_urllib_error(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_error"""
+
+
+_urllib_error_moved_attributes = [
+    MovedAttribute("URLError", "urllib2", "urllib.error"),
+    MovedAttribute("HTTPError", "urllib2", "urllib.error"),
+    MovedAttribute("ContentTooShortError", "urllib", "urllib.error"),
+]
+for attr in _urllib_error_moved_attributes:
+    setattr(Module_six_moves_urllib_error, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_error._moved_attributes = _urllib_error_moved_attributes
+
+_importer._add_module(
+    Module_six_moves_urllib_error(__name__ + ".moves.urllib.error"),
+    "moves.urllib_error",
+    "moves.urllib.error",
+)
+
+
+class Module_six_moves_urllib_request(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_request"""
+
+
+_urllib_request_moved_attributes = [
+    MovedAttribute("urlopen", "urllib2", "urllib.request"),
+    MovedAttribute("install_opener", "urllib2", "urllib.request"),
+    MovedAttribute("build_opener", "urllib2", "urllib.request"),
+    MovedAttribute("pathname2url", "urllib", "urllib.request"),
+    MovedAttribute("url2pathname", "urllib", "urllib.request"),
+    MovedAttribute("getproxies", "urllib", "urllib.request"),
+    MovedAttribute("Request", "urllib2", "urllib.request"),
+    MovedAttribute("OpenerDirector", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPDefaultErrorHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPRedirectHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPCookieProcessor", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyHandler", "urllib2", "urllib.request"),
+    MovedAttribute("BaseHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPPasswordMgr", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPPasswordMgrWithDefaultRealm", "urllib2", "urllib.request"),
+    MovedAttribute("AbstractBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("AbstractDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPSHandler", "urllib2", "urllib.request"),
+    MovedAttribute("FileHandler", "urllib2", "urllib.request"),
+    MovedAttribute("FTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("CacheFTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("UnknownHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPErrorProcessor", "urllib2", "urllib.request"),
+    MovedAttribute("urlretrieve", "urllib", "urllib.request"),
+    MovedAttribute("urlcleanup", "urllib", "urllib.request"),
+    MovedAttribute("URLopener", "urllib", "urllib.request"),
+    MovedAttribute("FancyURLopener", "urllib", "urllib.request"),
+    MovedAttribute("proxy_bypass", "urllib", "urllib.request"),
+    MovedAttribute("parse_http_list", "urllib2", "urllib.request"),
+    MovedAttribute("parse_keqv_list", "urllib2", "urllib.request"),
+]
+for attr in _urllib_request_moved_attributes:
+    setattr(Module_six_moves_urllib_request, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_request._moved_attributes = _urllib_request_moved_attributes
+
+_importer._add_module(
+    Module_six_moves_urllib_request(__name__ + ".moves.urllib.request"),
+    "moves.urllib_request",
+    "moves.urllib.request",
+)
+
+
+class Module_six_moves_urllib_response(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_response"""
+
+
+_urllib_response_moved_attributes = [
+    MovedAttribute("addbase", "urllib", "urllib.response"),
+    MovedAttribute("addclosehook", "urllib", "urllib.response"),
+    MovedAttribute("addinfo", "urllib", "urllib.response"),
+    MovedAttribute("addinfourl", "urllib", "urllib.response"),
+]
+for attr in _urllib_response_moved_attributes:
+    setattr(Module_six_moves_urllib_response, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_response._moved_attributes = _urllib_response_moved_attributes
+
+_importer._add_module(
+    Module_six_moves_urllib_response(__name__ + ".moves.urllib.response"),
+    "moves.urllib_response",
+    "moves.urllib.response",
+)
+
+
+class Module_six_moves_urllib_robotparser(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_robotparser"""
+
+
+_urllib_robotparser_moved_attributes = [
+    MovedAttribute("RobotFileParser", "robotparser", "urllib.robotparser"),
+]
+for attr in _urllib_robotparser_moved_attributes:
+    setattr(Module_six_moves_urllib_robotparser, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_robotparser._moved_attributes = (
+    _urllib_robotparser_moved_attributes
+)
+
+_importer._add_module(
+    Module_six_moves_urllib_robotparser(__name__ + ".moves.urllib.robotparser"),
+    "moves.urllib_robotparser",
+    "moves.urllib.robotparser",
+)
+
+
+class Module_six_moves_urllib(types.ModuleType):
+
+    """Create a six.moves.urllib namespace that resembles the Python 3 namespace"""
+
+    __path__ = []  # mark as package
+    parse = _importer._get_module("moves.urllib_parse")
+    error = _importer._get_module("moves.urllib_error")
+    request = _importer._get_module("moves.urllib_request")
+    response = _importer._get_module("moves.urllib_response")
+    robotparser = _importer._get_module("moves.urllib_robotparser")
+
+    def __dir__(self):
+        return ["parse", "error", "request", "response", "robotparser"]
+
+
+_importer._add_module(
+    Module_six_moves_urllib(__name__ + ".moves.urllib"), "moves.urllib"
+)
+
+
+def add_move(move):
+    """Add an item to six.moves."""
+    setattr(_MovedItems, move.name, move)
+
+
+def remove_move(name):
+    """Remove item from six.moves."""
+    try:
+        delattr(_MovedItems, name)
+    except AttributeError:
+        try:
+            del moves.__dict__[name]
+        except KeyError:
+            raise AttributeError("no such move, %r" % (name,))
+
+
+if PY3:
+    _meth_func = "__func__"
+    _meth_self = "__self__"
+
+    _func_closure = "__closure__"
+    _func_code = "__code__"
+    _func_defaults = "__defaults__"
+    _func_globals = "__globals__"
+else:
+    _meth_func = "im_func"
+    _meth_self = "im_self"
+
+    _func_closure = "func_closure"
+    _func_code = "func_code"
+    _func_defaults = "func_defaults"
+    _func_globals = "func_globals"
+
+
+try:
+    advance_iterator = next
+except NameError:
+
+    def advance_iterator(it):
+        return it.next()
+
+
+next = advance_iterator
+
+
+try:
+    callable = callable
+except NameError:
+
+    def callable(obj):
+        return any("__call__" in klass.__dict__ for klass in type(obj).__mro__)
+
+
+if PY3:
+
+    def get_unbound_function(unbound):
+        return unbound
+
+    create_bound_method = types.MethodType
+
+    def create_unbound_method(func, cls):
+        return func
+
+    Iterator = object
+else:
+
+    def get_unbound_function(unbound):
+        return unbound.im_func
+
+    def create_bound_method(func, obj):
+        return types.MethodType(func, obj, obj.__class__)
+
+    def create_unbound_method(func, cls):
+        return types.MethodType(func, None, cls)
+
+    class Iterator(object):
+        def next(self):
+            return type(self).__next__(self)
+
+    callable = callable
+_add_doc(
+    get_unbound_function, """Get the function out of a possibly unbound function"""
+)
+
+
+get_method_function = operator.attrgetter(_meth_func)
+get_method_self = operator.attrgetter(_meth_self)
+get_function_closure = operator.attrgetter(_func_closure)
+get_function_code = operator.attrgetter(_func_code)
+get_function_defaults = operator.attrgetter(_func_defaults)
+get_function_globals = operator.attrgetter(_func_globals)
+
+
+if PY3:
+
+    def iterkeys(d, **kw):
+        return iter(d.keys(**kw))
+
+    def itervalues(d, **kw):
+        return iter(d.values(**kw))
+
+    def iteritems(d, **kw):
+        return iter(d.items(**kw))
+
+    def iterlists(d, **kw):
+        return iter(d.lists(**kw))
+
+    viewkeys = operator.methodcaller("keys")
+
+    viewvalues = operator.methodcaller("values")
+
+    viewitems = operator.methodcaller("items")
+else:
+
+    def iterkeys(d, **kw):
+        return d.iterkeys(**kw)
+
+    def itervalues(d, **kw):
+        return d.itervalues(**kw)
+
+    def iteritems(d, **kw):
+        return d.iteritems(**kw)
+
+    def iterlists(d, **kw):
+        return d.iterlists(**kw)
+
+    viewkeys = operator.methodcaller("viewkeys")
+
+    viewvalues = operator.methodcaller("viewvalues")
+
+    viewitems = operator.methodcaller("viewitems")
+
+_add_doc(iterkeys, "Return an iterator over the keys of a dictionary.")
+_add_doc(itervalues, "Return an iterator over the values of a dictionary.")
+_add_doc(iteritems, "Return an iterator over the (key, value) pairs of a dictionary.")
+_add_doc(
+    iterlists, "Return an iterator over the (key, [values]) pairs of a dictionary."
+)
+
+
+if PY3:
+
+    def b(s):
+        return s.encode("latin-1")
+
+    def u(s):
+        return s
+
+    unichr = chr
+    import struct
+
+    int2byte = struct.Struct(">B").pack
+    del struct
+    byte2int = operator.itemgetter(0)
+    indexbytes = operator.getitem
+    iterbytes = iter
+    import io
+
+    StringIO = io.StringIO
+    BytesIO = io.BytesIO
+    del io
+    _assertCountEqual = "assertCountEqual"
+    if sys.version_info[1] <= 1:
+        _assertRaisesRegex = "assertRaisesRegexp"
+        _assertRegex = "assertRegexpMatches"
+        _assertNotRegex = "assertNotRegexpMatches"
+    else:
+        _assertRaisesRegex = "assertRaisesRegex"
+        _assertRegex = "assertRegex"
+        _assertNotRegex = "assertNotRegex"
+else:
+
+    def b(s):
+        return s
+
+    # Workaround for standalone backslash
+
+    def u(s):
+        return unicode(s.replace(r"\\", r"\\\\"), "unicode_escape")
+
+    unichr = unichr
+    int2byte = chr
+
+    def byte2int(bs):
+        return ord(bs[0])
+
+    def indexbytes(buf, i):
+        return ord(buf[i])
+
+    iterbytes = functools.partial(itertools.imap, ord)
+    import StringIO
+
+    StringIO = BytesIO = StringIO.StringIO
+    _assertCountEqual = "assertItemsEqual"
+    _assertRaisesRegex = "assertRaisesRegexp"
+    _assertRegex = "assertRegexpMatches"
+    _assertNotRegex = "assertNotRegexpMatches"
+_add_doc(b, """Byte literal""")
+_add_doc(u, """Text literal""")
+
+
+def assertCountEqual(self, *args, **kwargs):
+    return getattr(self, _assertCountEqual)(*args, **kwargs)
+
+
+def assertRaisesRegex(self, *args, **kwargs):
+    return getattr(self, _assertRaisesRegex)(*args, **kwargs)
+
+
+def assertRegex(self, *args, **kwargs):
+    return getattr(self, _assertRegex)(*args, **kwargs)
+
+
+def assertNotRegex(self, *args, **kwargs):
+    return getattr(self, _assertNotRegex)(*args, **kwargs)
+
+
+if PY3:
+    exec_ = getattr(moves.builtins, "exec")
+
+    def reraise(tp, value, tb=None):
+        try:
+            if value is None:
+                value = tp()
+            if value.__traceback__ is not tb:
+                raise value.with_traceback(tb)
+            raise value
+        finally:
+            value = None
+            tb = None
+
+else:
+
+    def exec_(_code_, _globs_=None, _locs_=None):
+        """Execute code in a namespace."""
+        if _globs_ is None:
+            frame = sys._getframe(1)
+            _globs_ = frame.f_globals
+            if _locs_ is None:
+                _locs_ = frame.f_locals
+            del frame
+        elif _locs_ is None:
+            _locs_ = _globs_
+        exec ("""exec _code_ in _globs_, _locs_""")
+
+    exec_(
+        """def reraise(tp, value, tb=None):
+    try:
+        raise tp, value, tb
+    finally:
+        tb = None
+"""
+    )
+
+
+if sys.version_info[:2] > (3,):
+    exec_(
+        """def raise_from(value, from_value):
+    try:
+        raise value from from_value
+    finally:
+        value = None
+"""
+    )
+else:
+
+    def raise_from(value, from_value):
+        raise value
+
+
+print_ = getattr(moves.builtins, "print", None)
+if print_ is None:
+
+    def print_(*args, **kwargs):
+        """The new-style print function for Python 2.4 and 2.5."""
+        fp = kwargs.pop("file", sys.stdout)
+        if fp is None:
+            return
+
+        def write(data):
+            if not isinstance(data, basestring):
+                data = str(data)
+            # If the file has an encoding, encode unicode with it.
+            if (
+                isinstance(fp, file)
+                and isinstance(data, unicode)
+                and fp.encoding is not None
+            ):
+                errors = getattr(fp, "errors", None)
+                if errors is None:
+                    errors = "strict"
+                data = data.encode(fp.encoding, errors)
+            fp.write(data)
+
+        want_unicode = False
+        sep = kwargs.pop("sep", None)
+        if sep is not None:
+            if isinstance(sep, unicode):
+                want_unicode = True
+            elif not isinstance(sep, str):
+                raise TypeError("sep must be None or a string")
+        end = kwargs.pop("end", None)
+        if end is not None:
+            if isinstance(end, unicode):
+                want_unicode = True
+            elif not isinstance(end, str):
+                raise TypeError("end must be None or a string")
+        if kwargs:
+            raise TypeError("invalid keyword arguments to print()")
+        if not want_unicode:
+            for arg in args:
+                if isinstance(arg, unicode):
+                    want_unicode = True
+                    break
+        if want_unicode:
+            newline = unicode("\n")
+            space = unicode(" ")
+        else:
+            newline = "\n"
+            space = " "
+        if sep is None:
+            sep = space
+        if end is None:
+            end = newline
+        for i, arg in enumerate(args):
+            if i:
+                write(sep)
+            write(arg)
+        write(end)
+
+
+if sys.version_info[:2] < (3, 3):
+    _print = print_
+
+    def print_(*args, **kwargs):
+        fp = kwargs.get("file", sys.stdout)
+        flush = kwargs.pop("flush", False)
+        _print(*args, **kwargs)
+        if flush and fp is not None:
+            fp.flush()
+
+
+_add_doc(reraise, """Reraise an exception.""")
+
+if sys.version_info[0:2] < (3, 4):
+    # This does exactly the same what the :func:`py3:functools.update_wrapper`
+    # function does on Python versions after 3.2. It sets the ``__wrapped__``
+    # attribute on ``wrapper`` object and it doesn't raise an error if any of
+    # the attributes mentioned in ``assigned`` and ``updated`` are missing on
+    # ``wrapped`` object.
+    def _update_wrapper(
+        wrapper,
+        wrapped,
+        assigned=functools.WRAPPER_ASSIGNMENTS,
+        updated=functools.WRAPPER_UPDATES,
+    ):
+        for attr in assigned:
+            try:
+                value = getattr(wrapped, attr)
+            except AttributeError:
+                continue
+            else:
+                setattr(wrapper, attr, value)
+        for attr in updated:
+            getattr(wrapper, attr).update(getattr(wrapped, attr, {}))
+        wrapper.__wrapped__ = wrapped
+        return wrapper
+
+    _update_wrapper.__doc__ = functools.update_wrapper.__doc__
+
+    def wraps(
+        wrapped,
+        assigned=functools.WRAPPER_ASSIGNMENTS,
+        updated=functools.WRAPPER_UPDATES,
+    ):
+        return functools.partial(
+            _update_wrapper, wrapped=wrapped, assigned=assigned, updated=updated
+        )
+
+    wraps.__doc__ = functools.wraps.__doc__
+
+else:
+    wraps = functools.wraps
+
+
+def with_metaclass(meta, *bases):
+    """Create a base class with a metaclass."""
+    # This requires a bit of explanation: the basic idea is to make a dummy
+    # metaclass for one level of class instantiation that replaces itself with
+    # the actual metaclass.
+    class metaclass(type):
+        def __new__(cls, name, this_bases, d):
+            if sys.version_info[:2] >= (3, 7):
+                # This version introduced PEP 560 that requires a bit
+                # of extra care (we mimic what is done by __build_class__).
+                resolved_bases = types.resolve_bases(bases)
+                if resolved_bases is not bases:
+                    d["__orig_bases__"] = bases
+            else:
+                resolved_bases = bases
+            return meta(name, resolved_bases, d)
+
+        @classmethod
+        def __prepare__(cls, name, this_bases):
+            return meta.__prepare__(name, bases)
+
+    return type.__new__(metaclass, "temporary_class", (), {})
+
+
+def add_metaclass(metaclass):
+    """Class decorator for creating a class with a metaclass."""
+
+    def wrapper(cls):
+        orig_vars = cls.__dict__.copy()
+        slots = orig_vars.get("__slots__")
+        if slots is not None:
+            if isinstance(slots, str):
+                slots = [slots]
+            for slots_var in slots:
+                orig_vars.pop(slots_var)
+        orig_vars.pop("__dict__", None)
+        orig_vars.pop("__weakref__", None)
+        if hasattr(cls, "__qualname__"):
+            orig_vars["__qualname__"] = cls.__qualname__
+        return metaclass(cls.__name__, cls.__bases__, orig_vars)
+
+    return wrapper
+
+
+def ensure_binary(s, encoding="utf-8", errors="strict"):
+    """Coerce **s** to six.binary_type.
+
+    For Python 2:
+      - `unicode` -> encoded to `str`
+      - `str` -> `str`
+
+    For Python 3:
+      - `str` -> encoded to `bytes`
+      - `bytes` -> `bytes`
+    """
+    if isinstance(s, binary_type):
+        return s
+    if isinstance(s, text_type):
+        return s.encode(encoding, errors)
+    raise TypeError("not expecting type '%s'" % type(s))
+
+
+def ensure_str(s, encoding="utf-8", errors="strict"):
+    """Coerce *s* to `str`.
+
+    For Python 2:
+      - `unicode` -> encoded to `str`
+      - `str` -> `str`
+
+    For Python 3:
+      - `str` -> `str`
+      - `bytes` -> decoded to `str`
+    """
+    # Optimization: Fast return for the common case.
+    if type(s) is str:
+        return s
+    if PY2 and isinstance(s, text_type):
+        return s.encode(encoding, errors)
+    elif PY3 and isinstance(s, binary_type):
+        return s.decode(encoding, errors)
+    elif not isinstance(s, (text_type, binary_type)):
+        raise TypeError("not expecting type '%s'" % type(s))
+    return s
+
+
+def ensure_text(s, encoding="utf-8", errors="strict"):
+    """Coerce *s* to six.text_type.
+
+    For Python 2:
+      - `unicode` -> `unicode`
+      - `str` -> `unicode`
+
+    For Python 3:
+      - `str` -> `str`
+      - `bytes` -> decoded to `str`
+    """
+    if isinstance(s, binary_type):
+        return s.decode(encoding, errors)
+    elif isinstance(s, text_type):
+        return s
+    else:
+        raise TypeError("not expecting type '%s'" % type(s))
+
+
+def python_2_unicode_compatible(klass):
+    """
+    A class decorator that defines __unicode__ and __str__ methods under Python 2.
+    Under Python 3 it does nothing.
+
+    To support Python 2 and 3 with a single code base, define a __str__ method
+    returning text and apply this decorator to the class.
+    """
+    if PY2:
+        if "__str__" not in klass.__dict__:
+            raise ValueError(
+                "@python_2_unicode_compatible cannot be applied "
+                "to %s because it doesn't define __str__()." % klass.__name__
+            )
+        klass.__unicode__ = klass.__str__
+        klass.__str__ = lambda self: self.__unicode__().encode("utf-8")
+    return klass
+
+
+# Complete the moves implementation.
+# This code is at the end of this module to speed up module loading.
+# Turn this module into a package.
+__path__ = []  # required for PEP 302 and PEP 451
+__package__ = __name__  # see PEP 366 @ReservedAssignment
+if globals().get("__spec__") is not None:
+    __spec__.submodule_search_locations = []  # PEP 451 @UndefinedVariable
+# Remove other six meta path importers, since they cause problems. This can
+# happen if six is removed from sys.modules and then reloaded. (Setuptools does
+# this for some reason.)
+if sys.meta_path:
+    for i, importer in enumerate(sys.meta_path):
+        # Here's some real nastiness: Another "instance" of the six module might
+        # be floating around. Therefore, we can't use isinstance() to check for
+        # the six meta path importer, since the other six instance will have
+        # inserted an importer with different class.
+        if (
+            type(importer).__name__ == "_SixMetaPathImporter"
+            and importer.name == __name__
+        ):
+            del sys.meta_path[i]
+            break
+    del i, importer
+# Finally, add the importer to the meta path import hook.
+sys.meta_path.append(_importer)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/poolmanager.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/poolmanager.py
new file mode 100644
index 00000000..fb51bf7d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/poolmanager.py
@@ -0,0 +1,540 @@
+from __future__ import absolute_import
+
+import collections
+import functools
+import logging
+
+from ._collections import HTTPHeaderDict, RecentlyUsedContainer
+from .connectionpool import HTTPConnectionPool, HTTPSConnectionPool, port_by_scheme
+from .exceptions import (
+    LocationValueError,
+    MaxRetryError,
+    ProxySchemeUnknown,
+    ProxySchemeUnsupported,
+    URLSchemeUnknown,
+)
+from .packages import six
+from .packages.six.moves.urllib.parse import urljoin
+from .request import RequestMethods
+from .util.proxy import connection_requires_http_tunnel
+from .util.retry import Retry
+from .util.url import parse_url
+
+__all__ = ["PoolManager", "ProxyManager", "proxy_from_url"]
+
+
+log = logging.getLogger(__name__)
+
+SSL_KEYWORDS = (
+    "key_file",
+    "cert_file",
+    "cert_reqs",
+    "ca_certs",
+    "ssl_version",
+    "ca_cert_dir",
+    "ssl_context",
+    "key_password",
+    "server_hostname",
+)
+
+# All known keyword arguments that could be provided to the pool manager, its
+# pools, or the underlying connections. This is used to construct a pool key.
+_key_fields = (
+    "key_scheme",  # str
+    "key_host",  # str
+    "key_port",  # int
+    "key_timeout",  # int or float or Timeout
+    "key_retries",  # int or Retry
+    "key_strict",  # bool
+    "key_block",  # bool
+    "key_source_address",  # str
+    "key_key_file",  # str
+    "key_key_password",  # str
+    "key_cert_file",  # str
+    "key_cert_reqs",  # str
+    "key_ca_certs",  # str
+    "key_ssl_version",  # str
+    "key_ca_cert_dir",  # str
+    "key_ssl_context",  # instance of ssl.SSLContext or urllib3.util.ssl_.SSLContext
+    "key_maxsize",  # int
+    "key_headers",  # dict
+    "key__proxy",  # parsed proxy url
+    "key__proxy_headers",  # dict
+    "key__proxy_config",  # class
+    "key_socket_options",  # list of (level (int), optname (int), value (int or str)) tuples
+    "key__socks_options",  # dict
+    "key_assert_hostname",  # bool or string
+    "key_assert_fingerprint",  # str
+    "key_server_hostname",  # str
+)
+
+#: The namedtuple class used to construct keys for the connection pool.
+#: All custom key schemes should include the fields in this key at a minimum.
+PoolKey = collections.namedtuple("PoolKey", _key_fields)
+
+_proxy_config_fields = ("ssl_context", "use_forwarding_for_https")
+ProxyConfig = collections.namedtuple("ProxyConfig", _proxy_config_fields)
+
+
+def _default_key_normalizer(key_class, request_context):
+    """
+    Create a pool key out of a request context dictionary.
+
+    According to RFC 3986, both the scheme and host are case-insensitive.
+    Therefore, this function normalizes both before constructing the pool
+    key for an HTTPS request. If you wish to change this behaviour, provide
+    alternate callables to ``key_fn_by_scheme``.
+
+    :param key_class:
+        The class to use when constructing the key. This should be a namedtuple
+        with the ``scheme`` and ``host`` keys at a minimum.
+    :type  key_class: namedtuple
+    :param request_context:
+        A dictionary-like object that contain the context for a request.
+    :type  request_context: dict
+
+    :return: A namedtuple that can be used as a connection pool key.
+    :rtype:  PoolKey
+    """
+    # Since we mutate the dictionary, make a copy first
+    context = request_context.copy()
+    context["scheme"] = context["scheme"].lower()
+    context["host"] = context["host"].lower()
+
+    # These are both dictionaries and need to be transformed into frozensets
+    for key in ("headers", "_proxy_headers", "_socks_options"):
+        if key in context and context[key] is not None:
+            context[key] = frozenset(context[key].items())
+
+    # The socket_options key may be a list and needs to be transformed into a
+    # tuple.
+    socket_opts = context.get("socket_options")
+    if socket_opts is not None:
+        context["socket_options"] = tuple(socket_opts)
+
+    # Map the kwargs to the names in the namedtuple - this is necessary since
+    # namedtuples can't have fields starting with '_'.
+    for key in list(context.keys()):
+        context["key_" + key] = context.pop(key)
+
+    # Default to ``None`` for keys missing from the context
+    for field in key_class._fields:
+        if field not in context:
+            context[field] = None
+
+    return key_class(**context)
+
+
+#: A dictionary that maps a scheme to a callable that creates a pool key.
+#: This can be used to alter the way pool keys are constructed, if desired.
+#: Each PoolManager makes a copy of this dictionary so they can be configured
+#: globally here, or individually on the instance.
+key_fn_by_scheme = {
+    "http": functools.partial(_default_key_normalizer, PoolKey),
+    "https": functools.partial(_default_key_normalizer, PoolKey),
+}
+
+pool_classes_by_scheme = {"http": HTTPConnectionPool, "https": HTTPSConnectionPool}
+
+
+class PoolManager(RequestMethods):
+    """
+    Allows for arbitrary requests while transparently keeping track of
+    necessary connection pools for you.
+
+    :param num_pools:
+        Number of connection pools to cache before discarding the least
+        recently used pool.
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+
+    :param \\**connection_pool_kw:
+        Additional parameters are used to create fresh
+        :class:`urllib3.connectionpool.ConnectionPool` instances.
+
+    Example::
+
+        >>> manager = PoolManager(num_pools=2)
+        >>> r = manager.request('GET', 'http://google.com/')
+        >>> r = manager.request('GET', 'http://google.com/mail')
+        >>> r = manager.request('GET', 'http://yahoo.com/')
+        >>> len(manager.pools)
+        2
+
+    """
+
+    proxy = None
+    proxy_config = None
+
+    def __init__(self, num_pools=10, headers=None, **connection_pool_kw):
+        RequestMethods.__init__(self, headers)
+        self.connection_pool_kw = connection_pool_kw
+        self.pools = RecentlyUsedContainer(num_pools)
+
+        # Locally set the pool classes and keys so other PoolManagers can
+        # override them.
+        self.pool_classes_by_scheme = pool_classes_by_scheme
+        self.key_fn_by_scheme = key_fn_by_scheme.copy()
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.clear()
+        # Return False to re-raise any potential exceptions
+        return False
+
+    def _new_pool(self, scheme, host, port, request_context=None):
+        """
+        Create a new :class:`urllib3.connectionpool.ConnectionPool` based on host, port, scheme, and
+        any additional pool keyword arguments.
+
+        If ``request_context`` is provided, it is provided as keyword arguments
+        to the pool class used. This method is used to actually create the
+        connection pools handed out by :meth:`connection_from_url` and
+        companion methods. It is intended to be overridden for customization.
+        """
+        pool_cls = self.pool_classes_by_scheme[scheme]
+        if request_context is None:
+            request_context = self.connection_pool_kw.copy()
+
+        # Although the context has everything necessary to create the pool,
+        # this function has historically only used the scheme, host, and port
+        # in the positional args. When an API change is acceptable these can
+        # be removed.
+        for key in ("scheme", "host", "port"):
+            request_context.pop(key, None)
+
+        if scheme == "http":
+            for kw in SSL_KEYWORDS:
+                request_context.pop(kw, None)
+
+        return pool_cls(host, port, **request_context)
+
+    def clear(self):
+        """
+        Empty our store of pools and direct them all to close.
+
+        This will not affect in-flight connections, but they will not be
+        re-used after completion.
+        """
+        self.pools.clear()
+
+    def connection_from_host(self, host, port=None, scheme="http", pool_kwargs=None):
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the host, port, and scheme.
+
+        If ``port`` isn't given, it will be derived from the ``scheme`` using
+        ``urllib3.connectionpool.port_by_scheme``. If ``pool_kwargs`` is
+        provided, it is merged with the instance's ``connection_pool_kw``
+        variable and used to create the new connection pool, if one is
+        needed.
+        """
+
+        if not host:
+            raise LocationValueError("No host specified.")
+
+        request_context = self._merge_pool_kwargs(pool_kwargs)
+        request_context["scheme"] = scheme or "http"
+        if not port:
+            port = port_by_scheme.get(request_context["scheme"].lower(), 80)
+        request_context["port"] = port
+        request_context["host"] = host
+
+        return self.connection_from_context(request_context)
+
+    def connection_from_context(self, request_context):
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the request context.
+
+        ``request_context`` must at least contain the ``scheme`` key and its
+        value must be a key in ``key_fn_by_scheme`` instance variable.
+        """
+        scheme = request_context["scheme"].lower()
+        pool_key_constructor = self.key_fn_by_scheme.get(scheme)
+        if not pool_key_constructor:
+            raise URLSchemeUnknown(scheme)
+        pool_key = pool_key_constructor(request_context)
+
+        return self.connection_from_pool_key(pool_key, request_context=request_context)
+
+    def connection_from_pool_key(self, pool_key, request_context=None):
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the provided pool key.
+
+        ``pool_key`` should be a namedtuple that only contains immutable
+        objects. At a minimum it must have the ``scheme``, ``host``, and
+        ``port`` fields.
+        """
+        with self.pools.lock:
+            # If the scheme, host, or port doesn't match existing open
+            # connections, open a new ConnectionPool.
+            pool = self.pools.get(pool_key)
+            if pool:
+                return pool
+
+            # Make a fresh ConnectionPool of the desired type
+            scheme = request_context["scheme"]
+            host = request_context["host"]
+            port = request_context["port"]
+            pool = self._new_pool(scheme, host, port, request_context=request_context)
+            self.pools[pool_key] = pool
+
+        return pool
+
+    def connection_from_url(self, url, pool_kwargs=None):
+        """
+        Similar to :func:`urllib3.connectionpool.connection_from_url`.
+
+        If ``pool_kwargs`` is not provided and a new pool needs to be
+        constructed, ``self.connection_pool_kw`` is used to initialize
+        the :class:`urllib3.connectionpool.ConnectionPool`. If ``pool_kwargs``
+        is provided, it is used instead. Note that if a new pool does not
+        need to be created for the request, the provided ``pool_kwargs`` are
+        not used.
+        """
+        u = parse_url(url)
+        return self.connection_from_host(
+            u.host, port=u.port, scheme=u.scheme, pool_kwargs=pool_kwargs
+        )
+
+    def _merge_pool_kwargs(self, override):
+        """
+        Merge a dictionary of override values for self.connection_pool_kw.
+
+        This does not modify self.connection_pool_kw and returns a new dict.
+        Any keys in the override dictionary with a value of ``None`` are
+        removed from the merged dictionary.
+        """
+        base_pool_kwargs = self.connection_pool_kw.copy()
+        if override:
+            for key, value in override.items():
+                if value is None:
+                    try:
+                        del base_pool_kwargs[key]
+                    except KeyError:
+                        pass
+                else:
+                    base_pool_kwargs[key] = value
+        return base_pool_kwargs
+
+    def _proxy_requires_url_absolute_form(self, parsed_url):
+        """
+        Indicates if the proxy requires the complete destination URL in the
+        request.  Normally this is only needed when not using an HTTP CONNECT
+        tunnel.
+        """
+        if self.proxy is None:
+            return False
+
+        return not connection_requires_http_tunnel(
+            self.proxy, self.proxy_config, parsed_url.scheme
+        )
+
+    def _validate_proxy_scheme_url_selection(self, url_scheme):
+        """
+        Validates that were not attempting to do TLS in TLS connections on
+        Python2 or with unsupported SSL implementations.
+        """
+        if self.proxy is None or url_scheme != "https":
+            return
+
+        if self.proxy.scheme != "https":
+            return
+
+        if six.PY2 and not self.proxy_config.use_forwarding_for_https:
+            raise ProxySchemeUnsupported(
+                "Contacting HTTPS destinations through HTTPS proxies "
+                "'via CONNECT tunnels' is not supported in Python 2"
+            )
+
+    def urlopen(self, method, url, redirect=True, **kw):
+        """
+        Same as :meth:`urllib3.HTTPConnectionPool.urlopen`
+        with custom cross-host redirect logic and only sends the request-uri
+        portion of the ``url``.
+
+        The given ``url`` parameter must be absolute, such that an appropriate
+        :class:`urllib3.connectionpool.ConnectionPool` can be chosen for it.
+        """
+        u = parse_url(url)
+        self._validate_proxy_scheme_url_selection(u.scheme)
+
+        conn = self.connection_from_host(u.host, port=u.port, scheme=u.scheme)
+
+        kw["assert_same_host"] = False
+        kw["redirect"] = False
+
+        if "headers" not in kw:
+            kw["headers"] = self.headers.copy()
+
+        if self._proxy_requires_url_absolute_form(u):
+            response = conn.urlopen(method, url, **kw)
+        else:
+            response = conn.urlopen(method, u.request_uri, **kw)
+
+        redirect_location = redirect and response.get_redirect_location()
+        if not redirect_location:
+            return response
+
+        # Support relative URLs for redirecting.
+        redirect_location = urljoin(url, redirect_location)
+
+        if response.status == 303:
+            # Change the method according to RFC 9110, Section 15.4.4.
+            method = "GET"
+            # And lose the body not to transfer anything sensitive.
+            kw["body"] = None
+            kw["headers"] = HTTPHeaderDict(kw["headers"])._prepare_for_method_change()
+
+        retries = kw.get("retries")
+        if not isinstance(retries, Retry):
+            retries = Retry.from_int(retries, redirect=redirect)
+
+        # Strip headers marked as unsafe to forward to the redirected location.
+        # Check remove_headers_on_redirect to avoid a potential network call within
+        # conn.is_same_host() which may use socket.gethostbyname() in the future.
+        if retries.remove_headers_on_redirect and not conn.is_same_host(
+            redirect_location
+        ):
+            headers = list(six.iterkeys(kw["headers"]))
+            for header in headers:
+                if header.lower() in retries.remove_headers_on_redirect:
+                    kw["headers"].pop(header, None)
+
+        try:
+            retries = retries.increment(method, url, response=response, _pool=conn)
+        except MaxRetryError:
+            if retries.raise_on_redirect:
+                response.drain_conn()
+                raise
+            return response
+
+        kw["retries"] = retries
+        kw["redirect"] = redirect
+
+        log.info("Redirecting %s -> %s", url, redirect_location)
+
+        response.drain_conn()
+        return self.urlopen(method, redirect_location, **kw)
+
+
+class ProxyManager(PoolManager):
+    """
+    Behaves just like :class:`PoolManager`, but sends all requests through
+    the defined proxy, using the CONNECT method for HTTPS URLs.
+
+    :param proxy_url:
+        The URL of the proxy to be used.
+
+    :param proxy_headers:
+        A dictionary containing headers that will be sent to the proxy. In case
+        of HTTP they are being sent with each request, while in the
+        HTTPS/CONNECT case they are sent only once. Could be used for proxy
+        authentication.
+
+    :param proxy_ssl_context:
+        The proxy SSL context is used to establish the TLS connection to the
+        proxy when using HTTPS proxies.
+
+    :param use_forwarding_for_https:
+        (Defaults to False) If set to True will forward requests to the HTTPS
+        proxy to be made on behalf of the client instead of creating a TLS
+        tunnel via the CONNECT method. **Enabling this flag means that request
+        and response headers and content will be visible from the HTTPS proxy**
+        whereas tunneling keeps request and response headers and content
+        private.  IP address, target hostname, SNI, and port are always visible
+        to an HTTPS proxy even when this flag is disabled.
+
+    Example:
+        >>> proxy = urllib3.ProxyManager('http://localhost:3128/')
+        >>> r1 = proxy.request('GET', 'http://google.com/')
+        >>> r2 = proxy.request('GET', 'http://httpbin.org/')
+        >>> len(proxy.pools)
+        1
+        >>> r3 = proxy.request('GET', 'https://httpbin.org/')
+        >>> r4 = proxy.request('GET', 'https://twitter.com/')
+        >>> len(proxy.pools)
+        3
+
+    """
+
+    def __init__(
+        self,
+        proxy_url,
+        num_pools=10,
+        headers=None,
+        proxy_headers=None,
+        proxy_ssl_context=None,
+        use_forwarding_for_https=False,
+        **connection_pool_kw
+    ):
+
+        if isinstance(proxy_url, HTTPConnectionPool):
+            proxy_url = "%s://%s:%i" % (
+                proxy_url.scheme,
+                proxy_url.host,
+                proxy_url.port,
+            )
+        proxy = parse_url(proxy_url)
+
+        if proxy.scheme not in ("http", "https"):
+            raise ProxySchemeUnknown(proxy.scheme)
+
+        if not proxy.port:
+            port = port_by_scheme.get(proxy.scheme, 80)
+            proxy = proxy._replace(port=port)
+
+        self.proxy = proxy
+        self.proxy_headers = proxy_headers or {}
+        self.proxy_ssl_context = proxy_ssl_context
+        self.proxy_config = ProxyConfig(proxy_ssl_context, use_forwarding_for_https)
+
+        connection_pool_kw["_proxy"] = self.proxy
+        connection_pool_kw["_proxy_headers"] = self.proxy_headers
+        connection_pool_kw["_proxy_config"] = self.proxy_config
+
+        super(ProxyManager, self).__init__(num_pools, headers, **connection_pool_kw)
+
+    def connection_from_host(self, host, port=None, scheme="http", pool_kwargs=None):
+        if scheme == "https":
+            return super(ProxyManager, self).connection_from_host(
+                host, port, scheme, pool_kwargs=pool_kwargs
+            )
+
+        return super(ProxyManager, self).connection_from_host(
+            self.proxy.host, self.proxy.port, self.proxy.scheme, pool_kwargs=pool_kwargs
+        )
+
+    def _set_proxy_headers(self, url, headers=None):
+        """
+        Sets headers needed by proxies: specifically, the Accept and Host
+        headers. Only sets headers not provided by the user.
+        """
+        headers_ = {"Accept": "*/*"}
+
+        netloc = parse_url(url).netloc
+        if netloc:
+            headers_["Host"] = netloc
+
+        if headers:
+            headers_.update(headers)
+        return headers_
+
+    def urlopen(self, method, url, redirect=True, **kw):
+        "Same as HTTP(S)ConnectionPool.urlopen, ``url`` must be absolute."
+        u = parse_url(url)
+        if not connection_requires_http_tunnel(self.proxy, self.proxy_config, u.scheme):
+            # For connections using HTTP CONNECT, httplib sets the necessary
+            # headers on the CONNECT to the proxy. If we're not using CONNECT,
+            # we'll definitely need to set 'Host' at the very least.
+            headers = kw.get("headers", self.headers)
+            kw["headers"] = self._set_proxy_headers(url, headers)
+
+        return super(ProxyManager, self).urlopen(method, url, redirect=redirect, **kw)
+
+
+def proxy_from_url(url, **kw):
+    return ProxyManager(proxy_url=url, **kw)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/request.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/request.py
new file mode 100644
index 00000000..3b4cf999
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/request.py
@@ -0,0 +1,191 @@
+from __future__ import absolute_import
+
+import sys
+
+from .filepost import encode_multipart_formdata
+from .packages import six
+from .packages.six.moves.urllib.parse import urlencode
+
+__all__ = ["RequestMethods"]
+
+
+class RequestMethods(object):
+    """
+    Convenience mixin for classes who implement a :meth:`urlopen` method, such
+    as :class:`urllib3.HTTPConnectionPool` and
+    :class:`urllib3.PoolManager`.
+
+    Provides behavior for making common types of HTTP request methods and
+    decides which type of request field encoding to use.
+
+    Specifically,
+
+    :meth:`.request_encode_url` is for sending requests whose fields are
+    encoded in the URL (such as GET, HEAD, DELETE).
+
+    :meth:`.request_encode_body` is for sending requests whose fields are
+    encoded in the *body* of the request using multipart or www-form-urlencoded
+    (such as for POST, PUT, PATCH).
+
+    :meth:`.request` is for making any kind of request, it will look up the
+    appropriate encoding format and use one of the above two methods to make
+    the request.
+
+    Initializer parameters:
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+    """
+
+    _encode_url_methods = {"DELETE", "GET", "HEAD", "OPTIONS"}
+
+    def __init__(self, headers=None):
+        self.headers = headers or {}
+
+    def urlopen(
+        self,
+        method,
+        url,
+        body=None,
+        headers=None,
+        encode_multipart=True,
+        multipart_boundary=None,
+        **kw
+    ):  # Abstract
+        raise NotImplementedError(
+            "Classes extending RequestMethods must implement "
+            "their own ``urlopen`` method."
+        )
+
+    def request(self, method, url, fields=None, headers=None, **urlopen_kw):
+        """
+        Make a request using :meth:`urlopen` with the appropriate encoding of
+        ``fields`` based on the ``method`` used.
+
+        This is a convenience method that requires the least amount of manual
+        effort. It can be used in most situations, while still having the
+        option to drop down to more specific methods when necessary, such as
+        :meth:`request_encode_url`, :meth:`request_encode_body`,
+        or even the lowest level :meth:`urlopen`.
+        """
+        method = method.upper()
+
+        urlopen_kw["request_url"] = url
+
+        if method in self._encode_url_methods:
+            return self.request_encode_url(
+                method, url, fields=fields, headers=headers, **urlopen_kw
+            )
+        else:
+            return self.request_encode_body(
+                method, url, fields=fields, headers=headers, **urlopen_kw
+            )
+
+    def request_encode_url(self, method, url, fields=None, headers=None, **urlopen_kw):
+        """
+        Make a request using :meth:`urlopen` with the ``fields`` encoded in
+        the url. This is useful for request methods like GET, HEAD, DELETE, etc.
+        """
+        if headers is None:
+            headers = self.headers
+
+        extra_kw = {"headers": headers}
+        extra_kw.update(urlopen_kw)
+
+        if fields:
+            url += "?" + urlencode(fields)
+
+        return self.urlopen(method, url, **extra_kw)
+
+    def request_encode_body(
+        self,
+        method,
+        url,
+        fields=None,
+        headers=None,
+        encode_multipart=True,
+        multipart_boundary=None,
+        **urlopen_kw
+    ):
+        """
+        Make a request using :meth:`urlopen` with the ``fields`` encoded in
+        the body. This is useful for request methods like POST, PUT, PATCH, etc.
+
+        When ``encode_multipart=True`` (default), then
+        :func:`urllib3.encode_multipart_formdata` is used to encode
+        the payload with the appropriate content type. Otherwise
+        :func:`urllib.parse.urlencode` is used with the
+        'application/x-www-form-urlencoded' content type.
+
+        Multipart encoding must be used when posting files, and it's reasonably
+        safe to use it in other times too. However, it may break request
+        signing, such as with OAuth.
+
+        Supports an optional ``fields`` parameter of key/value strings AND
+        key/filetuple. A filetuple is a (filename, data, MIME type) tuple where
+        the MIME type is optional. For example::
+
+            fields = {
+                'foo': 'bar',
+                'fakefile': ('foofile.txt', 'contents of foofile'),
+                'realfile': ('barfile.txt', open('realfile').read()),
+                'typedfile': ('bazfile.bin', open('bazfile').read(),
+                              'image/jpeg'),
+                'nonamefile': 'contents of nonamefile field',
+            }
+
+        When uploading a file, providing a filename (the first parameter of the
+        tuple) is optional but recommended to best mimic behavior of browsers.
+
+        Note that if ``headers`` are supplied, the 'Content-Type' header will
+        be overwritten because it depends on the dynamic random boundary string
+        which is used to compose the body of the request. The random boundary
+        string can be explicitly set with the ``multipart_boundary`` parameter.
+        """
+        if headers is None:
+            headers = self.headers
+
+        extra_kw = {"headers": {}}
+
+        if fields:
+            if "body" in urlopen_kw:
+                raise TypeError(
+                    "request got values for both 'fields' and 'body', can only specify one."
+                )
+
+            if encode_multipart:
+                body, content_type = encode_multipart_formdata(
+                    fields, boundary=multipart_boundary
+                )
+            else:
+                body, content_type = (
+                    urlencode(fields),
+                    "application/x-www-form-urlencoded",
+                )
+
+            extra_kw["body"] = body
+            extra_kw["headers"] = {"Content-Type": content_type}
+
+        extra_kw["headers"].update(headers)
+        extra_kw.update(urlopen_kw)
+
+        return self.urlopen(method, url, **extra_kw)
+
+
+if not six.PY2:
+
+    class RequestModule(sys.modules[__name__].__class__):
+        def __call__(self, *args, **kwargs):
+            """
+            If user tries to call this module directly urllib3 v2.x style raise an error to the user
+            suggesting they may need urllib3 v2
+            """
+            raise TypeError(
+                "'module' object is not callable\n"
+                "urllib3.request() method is not supported in this release, "
+                "upgrade to urllib3 v2 to use it\n"
+                "see https://urllib3.readthedocs.io/en/stable/v2-migration-guide.html"
+            )
+
+    sys.modules[__name__].__class__ = RequestModule
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/response.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/response.py
new file mode 100644
index 00000000..8909f845
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/response.py
@@ -0,0 +1,879 @@
+from __future__ import absolute_import
+
+import io
+import logging
+import sys
+import warnings
+import zlib
+from contextlib import contextmanager
+from socket import error as SocketError
+from socket import timeout as SocketTimeout
+
+brotli = None
+
+from . import util
+from ._collections import HTTPHeaderDict
+from .connection import BaseSSLError, HTTPException
+from .exceptions import (
+    BodyNotHttplibCompatible,
+    DecodeError,
+    HTTPError,
+    IncompleteRead,
+    InvalidChunkLength,
+    InvalidHeader,
+    ProtocolError,
+    ReadTimeoutError,
+    ResponseNotChunked,
+    SSLError,
+)
+from .packages import six
+from .util.response import is_fp_closed, is_response_to_head
+
+log = logging.getLogger(__name__)
+
+
+class DeflateDecoder(object):
+    def __init__(self):
+        self._first_try = True
+        self._data = b""
+        self._obj = zlib.decompressobj()
+
+    def __getattr__(self, name):
+        return getattr(self._obj, name)
+
+    def decompress(self, data):
+        if not data:
+            return data
+
+        if not self._first_try:
+            return self._obj.decompress(data)
+
+        self._data += data
+        try:
+            decompressed = self._obj.decompress(data)
+            if decompressed:
+                self._first_try = False
+                self._data = None
+            return decompressed
+        except zlib.error:
+            self._first_try = False
+            self._obj = zlib.decompressobj(-zlib.MAX_WBITS)
+            try:
+                return self.decompress(self._data)
+            finally:
+                self._data = None
+
+
+class GzipDecoderState(object):
+
+    FIRST_MEMBER = 0
+    OTHER_MEMBERS = 1
+    SWALLOW_DATA = 2
+
+
+class GzipDecoder(object):
+    def __init__(self):
+        self._obj = zlib.decompressobj(16 + zlib.MAX_WBITS)
+        self._state = GzipDecoderState.FIRST_MEMBER
+
+    def __getattr__(self, name):
+        return getattr(self._obj, name)
+
+    def decompress(self, data):
+        ret = bytearray()
+        if self._state == GzipDecoderState.SWALLOW_DATA or not data:
+            return bytes(ret)
+        while True:
+            try:
+                ret += self._obj.decompress(data)
+            except zlib.error:
+                previous_state = self._state
+                # Ignore data after the first error
+                self._state = GzipDecoderState.SWALLOW_DATA
+                if previous_state == GzipDecoderState.OTHER_MEMBERS:
+                    # Allow trailing garbage acceptable in other gzip clients
+                    return bytes(ret)
+                raise
+            data = self._obj.unused_data
+            if not data:
+                return bytes(ret)
+            self._state = GzipDecoderState.OTHER_MEMBERS
+            self._obj = zlib.decompressobj(16 + zlib.MAX_WBITS)
+
+
+if brotli is not None:
+
+    class BrotliDecoder(object):
+        # Supports both 'brotlipy' and 'Brotli' packages
+        # since they share an import name. The top branches
+        # are for 'brotlipy' and bottom branches for 'Brotli'
+        def __init__(self):
+            self._obj = brotli.Decompressor()
+            if hasattr(self._obj, "decompress"):
+                self.decompress = self._obj.decompress
+            else:
+                self.decompress = self._obj.process
+
+        def flush(self):
+            if hasattr(self._obj, "flush"):
+                return self._obj.flush()
+            return b""
+
+
+class MultiDecoder(object):
+    """
+    From RFC7231:
+        If one or more encodings have been applied to a representation, the
+        sender that applied the encodings MUST generate a Content-Encoding
+        header field that lists the content codings in the order in which
+        they were applied.
+    """
+
+    def __init__(self, modes):
+        self._decoders = [_get_decoder(m.strip()) for m in modes.split(",")]
+
+    def flush(self):
+        return self._decoders[0].flush()
+
+    def decompress(self, data):
+        for d in reversed(self._decoders):
+            data = d.decompress(data)
+        return data
+
+
+def _get_decoder(mode):
+    if "," in mode:
+        return MultiDecoder(mode)
+
+    if mode == "gzip":
+        return GzipDecoder()
+
+    if brotli is not None and mode == "br":
+        return BrotliDecoder()
+
+    return DeflateDecoder()
+
+
+class HTTPResponse(io.IOBase):
+    """
+    HTTP Response container.
+
+    Backwards-compatible with :class:`http.client.HTTPResponse` but the response ``body`` is
+    loaded and decoded on-demand when the ``data`` property is accessed.  This
+    class is also compatible with the Python standard library's :mod:`io`
+    module, and can hence be treated as a readable object in the context of that
+    framework.
+
+    Extra parameters for behaviour not present in :class:`http.client.HTTPResponse`:
+
+    :param preload_content:
+        If True, the response's body will be preloaded during construction.
+
+    :param decode_content:
+        If True, will attempt to decode the body based on the
+        'content-encoding' header.
+
+    :param original_response:
+        When this HTTPResponse wrapper is generated from an :class:`http.client.HTTPResponse`
+        object, it's convenient to include the original for debug purposes. It's
+        otherwise unused.
+
+    :param retries:
+        The retries contains the last :class:`~urllib3.util.retry.Retry` that
+        was used during the request.
+
+    :param enforce_content_length:
+        Enforce content length checking. Body returned by server must match
+        value of Content-Length header, if present. Otherwise, raise error.
+    """
+
+    CONTENT_DECODERS = ["gzip", "deflate"]
+    if brotli is not None:
+        CONTENT_DECODERS += ["br"]
+    REDIRECT_STATUSES = [301, 302, 303, 307, 308]
+
+    def __init__(
+        self,
+        body="",
+        headers=None,
+        status=0,
+        version=0,
+        reason=None,
+        strict=0,
+        preload_content=True,
+        decode_content=True,
+        original_response=None,
+        pool=None,
+        connection=None,
+        msg=None,
+        retries=None,
+        enforce_content_length=False,
+        request_method=None,
+        request_url=None,
+        auto_close=True,
+    ):
+
+        if isinstance(headers, HTTPHeaderDict):
+            self.headers = headers
+        else:
+            self.headers = HTTPHeaderDict(headers)
+        self.status = status
+        self.version = version
+        self.reason = reason
+        self.strict = strict
+        self.decode_content = decode_content
+        self.retries = retries
+        self.enforce_content_length = enforce_content_length
+        self.auto_close = auto_close
+
+        self._decoder = None
+        self._body = None
+        self._fp = None
+        self._original_response = original_response
+        self._fp_bytes_read = 0
+        self.msg = msg
+        self._request_url = request_url
+
+        if body and isinstance(body, (six.string_types, bytes)):
+            self._body = body
+
+        self._pool = pool
+        self._connection = connection
+
+        if hasattr(body, "read"):
+            self._fp = body
+
+        # Are we using the chunked-style of transfer encoding?
+        self.chunked = False
+        self.chunk_left = None
+        tr_enc = self.headers.get("transfer-encoding", "").lower()
+        # Don't incur the penalty of creating a list and then discarding it
+        encodings = (enc.strip() for enc in tr_enc.split(","))
+        if "chunked" in encodings:
+            self.chunked = True
+
+        # Determine length of response
+        self.length_remaining = self._init_length(request_method)
+
+        # If requested, preload the body.
+        if preload_content and not self._body:
+            self._body = self.read(decode_content=decode_content)
+
+    def get_redirect_location(self):
+        """
+        Should we redirect and where to?
+
+        :returns: Truthy redirect location string if we got a redirect status
+            code and valid location. ``None`` if redirect status and no
+            location. ``False`` if not a redirect status code.
+        """
+        if self.status in self.REDIRECT_STATUSES:
+            return self.headers.get("location")
+
+        return False
+
+    def release_conn(self):
+        if not self._pool or not self._connection:
+            return
+
+        self._pool._put_conn(self._connection)
+        self._connection = None
+
+    def drain_conn(self):
+        """
+        Read and discard any remaining HTTP response data in the response connection.
+
+        Unread data in the HTTPResponse connection blocks the connection from being released back to the pool.
+        """
+        try:
+            self.read()
+        except (HTTPError, SocketError, BaseSSLError, HTTPException):
+            pass
+
+    @property
+    def data(self):
+        # For backwards-compat with earlier urllib3 0.4 and earlier.
+        if self._body:
+            return self._body
+
+        if self._fp:
+            return self.read(cache_content=True)
+
+    @property
+    def connection(self):
+        return self._connection
+
+    def isclosed(self):
+        return is_fp_closed(self._fp)
+
+    def tell(self):
+        """
+        Obtain the number of bytes pulled over the wire so far. May differ from
+        the amount of content returned by :meth:``urllib3.response.HTTPResponse.read``
+        if bytes are encoded on the wire (e.g, compressed).
+        """
+        return self._fp_bytes_read
+
+    def _init_length(self, request_method):
+        """
+        Set initial length value for Response content if available.
+        """
+        length = self.headers.get("content-length")
+
+        if length is not None:
+            if self.chunked:
+                # This Response will fail with an IncompleteRead if it can't be
+                # received as chunked. This method falls back to attempt reading
+                # the response before raising an exception.
+                log.warning(
+                    "Received response with both Content-Length and "
+                    "Transfer-Encoding set. This is expressly forbidden "
+                    "by RFC 7230 sec 3.3.2. Ignoring Content-Length and "
+                    "attempting to process response as Transfer-Encoding: "
+                    "chunked."
+                )
+                return None
+
+            try:
+                # RFC 7230 section 3.3.2 specifies multiple content lengths can
+                # be sent in a single Content-Length header
+                # (e.g. Content-Length: 42, 42). This line ensures the values
+                # are all valid ints and that as long as the `set` length is 1,
+                # all values are the same. Otherwise, the header is invalid.
+                lengths = set([int(val) for val in length.split(",")])
+                if len(lengths) > 1:
+                    raise InvalidHeader(
+                        "Content-Length contained multiple "
+                        "unmatching values (%s)" % length
+                    )
+                length = lengths.pop()
+            except ValueError:
+                length = None
+            else:
+                if length < 0:
+                    length = None
+
+        # Convert status to int for comparison
+        # In some cases, httplib returns a status of "_UNKNOWN"
+        try:
+            status = int(self.status)
+        except ValueError:
+            status = 0
+
+        # Check for responses that shouldn't include a body
+        if status in (204, 304) or 100 <= status < 200 or request_method == "HEAD":
+            length = 0
+
+        return length
+
+    def _init_decoder(self):
+        """
+        Set-up the _decoder attribute if necessary.
+        """
+        # Note: content-encoding value should be case-insensitive, per RFC 7230
+        # Section 3.2
+        content_encoding = self.headers.get("content-encoding", "").lower()
+        if self._decoder is None:
+            if content_encoding in self.CONTENT_DECODERS:
+                self._decoder = _get_decoder(content_encoding)
+            elif "," in content_encoding:
+                encodings = [
+                    e.strip()
+                    for e in content_encoding.split(",")
+                    if e.strip() in self.CONTENT_DECODERS
+                ]
+                if len(encodings):
+                    self._decoder = _get_decoder(content_encoding)
+
+    DECODER_ERROR_CLASSES = (IOError, zlib.error)
+    if brotli is not None:
+        DECODER_ERROR_CLASSES += (brotli.error,)
+
+    def _decode(self, data, decode_content, flush_decoder):
+        """
+        Decode the data passed in and potentially flush the decoder.
+        """
+        if not decode_content:
+            return data
+
+        try:
+            if self._decoder:
+                data = self._decoder.decompress(data)
+        except self.DECODER_ERROR_CLASSES as e:
+            content_encoding = self.headers.get("content-encoding", "").lower()
+            raise DecodeError(
+                "Received response with content-encoding: %s, but "
+                "failed to decode it." % content_encoding,
+                e,
+            )
+        if flush_decoder:
+            data += self._flush_decoder()
+
+        return data
+
+    def _flush_decoder(self):
+        """
+        Flushes the decoder. Should only be called if the decoder is actually
+        being used.
+        """
+        if self._decoder:
+            buf = self._decoder.decompress(b"")
+            return buf + self._decoder.flush()
+
+        return b""
+
+    @contextmanager
+    def _error_catcher(self):
+        """
+        Catch low-level python exceptions, instead re-raising urllib3
+        variants, so that low-level exceptions are not leaked in the
+        high-level api.
+
+        On exit, release the connection back to the pool.
+        """
+        clean_exit = False
+
+        try:
+            try:
+                yield
+
+            except SocketTimeout:
+                # FIXME: Ideally we'd like to include the url in the ReadTimeoutError but
+                # there is yet no clean way to get at it from this context.
+                raise ReadTimeoutError(self._pool, None, "Read timed out.")
+
+            except BaseSSLError as e:
+                # FIXME: Is there a better way to differentiate between SSLErrors?
+                if "read operation timed out" not in str(e):
+                    # SSL errors related to framing/MAC get wrapped and reraised here
+                    raise SSLError(e)
+
+                raise ReadTimeoutError(self._pool, None, "Read timed out.")
+
+            except (HTTPException, SocketError) as e:
+                # This includes IncompleteRead.
+                raise ProtocolError("Connection broken: %r" % e, e)
+
+            # If no exception is thrown, we should avoid cleaning up
+            # unnecessarily.
+            clean_exit = True
+        finally:
+            # If we didn't terminate cleanly, we need to throw away our
+            # connection.
+            if not clean_exit:
+                # The response may not be closed but we're not going to use it
+                # anymore so close it now to ensure that the connection is
+                # released back to the pool.
+                if self._original_response:
+                    self._original_response.close()
+
+                # Closing the response may not actually be sufficient to close
+                # everything, so if we have a hold of the connection close that
+                # too.
+                if self._connection:
+                    self._connection.close()
+
+            # If we hold the original response but it's closed now, we should
+            # return the connection back to the pool.
+            if self._original_response and self._original_response.isclosed():
+                self.release_conn()
+
+    def _fp_read(self, amt):
+        """
+        Read a response with the thought that reading the number of bytes
+        larger than can fit in a 32-bit int at a time via SSL in some
+        known cases leads to an overflow error that has to be prevented
+        if `amt` or `self.length_remaining` indicate that a problem may
+        happen.
+
+        The known cases:
+          * 3.8 <= CPython < 3.9.7 because of a bug
+            https://github.com/urllib3/urllib3/issues/2513#issuecomment-1152559900.
+          * urllib3 injected with pyOpenSSL-backed SSL-support.
+          * CPython < 3.10 only when `amt` does not fit 32-bit int.
+        """
+        assert self._fp
+        c_int_max = 2 ** 31 - 1
+        if (
+            (
+                (amt and amt > c_int_max)
+                or (self.length_remaining and self.length_remaining > c_int_max)
+            )
+            and not util.IS_SECURETRANSPORT
+            and (util.IS_PYOPENSSL or sys.version_info < (3, 10))
+        ):
+            buffer = io.BytesIO()
+            # Besides `max_chunk_amt` being a maximum chunk size, it
+            # affects memory overhead of reading a response by this
+            # method in CPython.
+            # `c_int_max` equal to 2 GiB - 1 byte is the actual maximum
+            # chunk size that does not lead to an overflow error, but
+            # 256 MiB is a compromise.
+            max_chunk_amt = 2 ** 28
+            while amt is None or amt != 0:
+                if amt is not None:
+                    chunk_amt = min(amt, max_chunk_amt)
+                    amt -= chunk_amt
+                else:
+                    chunk_amt = max_chunk_amt
+                data = self._fp.read(chunk_amt)
+                if not data:
+                    break
+                buffer.write(data)
+                del data  # to reduce peak memory usage by `max_chunk_amt`.
+            return buffer.getvalue()
+        else:
+            # StringIO doesn't like amt=None
+            return self._fp.read(amt) if amt is not None else self._fp.read()
+
+    def read(self, amt=None, decode_content=None, cache_content=False):
+        """
+        Similar to :meth:`http.client.HTTPResponse.read`, but with two additional
+        parameters: ``decode_content`` and ``cache_content``.
+
+        :param amt:
+            How much of the content to read. If specified, caching is skipped
+            because it doesn't make sense to cache partial content as the full
+            response.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+
+        :param cache_content:
+            If True, will save the returned data such that the same result is
+            returned despite of the state of the underlying file object. This
+            is useful if you want the ``.data`` property to continue working
+            after having ``.read()`` the file object. (Overridden if ``amt`` is
+            set.)
+        """
+        self._init_decoder()
+        if decode_content is None:
+            decode_content = self.decode_content
+
+        if self._fp is None:
+            return
+
+        flush_decoder = False
+        fp_closed = getattr(self._fp, "closed", False)
+
+        with self._error_catcher():
+            data = self._fp_read(amt) if not fp_closed else b""
+            if amt is None:
+                flush_decoder = True
+            else:
+                cache_content = False
+                if (
+                    amt != 0 and not data
+                ):  # Platform-specific: Buggy versions of Python.
+                    # Close the connection when no data is returned
+                    #
+                    # This is redundant to what httplib/http.client _should_
+                    # already do.  However, versions of python released before
+                    # December 15, 2012 (http://bugs.python.org/issue16298) do
+                    # not properly close the connection in all cases. There is
+                    # no harm in redundantly calling close.
+                    self._fp.close()
+                    flush_decoder = True
+                    if self.enforce_content_length and self.length_remaining not in (
+                        0,
+                        None,
+                    ):
+                        # This is an edge case that httplib failed to cover due
+                        # to concerns of backward compatibility. We're
+                        # addressing it here to make sure IncompleteRead is
+                        # raised during streaming, so all calls with incorrect
+                        # Content-Length are caught.
+                        raise IncompleteRead(self._fp_bytes_read, self.length_remaining)
+
+        if data:
+            self._fp_bytes_read += len(data)
+            if self.length_remaining is not None:
+                self.length_remaining -= len(data)
+
+            data = self._decode(data, decode_content, flush_decoder)
+
+            if cache_content:
+                self._body = data
+
+        return data
+
+    def stream(self, amt=2 ** 16, decode_content=None):
+        """
+        A generator wrapper for the read() method. A call will block until
+        ``amt`` bytes have been read from the connection or until the
+        connection is closed.
+
+        :param amt:
+            How much of the content to read. The generator will return up to
+            much data per iteration, but may return less. This is particularly
+            likely when using compressed data. However, the empty string will
+            never be returned.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        if self.chunked and self.supports_chunked_reads():
+            for line in self.read_chunked(amt, decode_content=decode_content):
+                yield line
+        else:
+            while not is_fp_closed(self._fp):
+                data = self.read(amt=amt, decode_content=decode_content)
+
+                if data:
+                    yield data
+
+    @classmethod
+    def from_httplib(ResponseCls, r, **response_kw):
+        """
+        Given an :class:`http.client.HTTPResponse` instance ``r``, return a
+        corresponding :class:`urllib3.response.HTTPResponse` object.
+
+        Remaining parameters are passed to the HTTPResponse constructor, along
+        with ``original_response=r``.
+        """
+        headers = r.msg
+
+        if not isinstance(headers, HTTPHeaderDict):
+            if six.PY2:
+                # Python 2.7
+                headers = HTTPHeaderDict.from_httplib(headers)
+            else:
+                headers = HTTPHeaderDict(headers.items())
+
+        # HTTPResponse objects in Python 3 don't have a .strict attribute
+        strict = getattr(r, "strict", 0)
+        resp = ResponseCls(
+            body=r,
+            headers=headers,
+            status=r.status,
+            version=r.version,
+            reason=r.reason,
+            strict=strict,
+            original_response=r,
+            **response_kw
+        )
+        return resp
+
+    # Backwards-compatibility methods for http.client.HTTPResponse
+    def getheaders(self):
+        warnings.warn(
+            "HTTPResponse.getheaders() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead access HTTPResponse.headers directly.",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.headers
+
+    def getheader(self, name, default=None):
+        warnings.warn(
+            "HTTPResponse.getheader() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead use HTTPResponse.headers.get(name, default).",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.headers.get(name, default)
+
+    # Backwards compatibility for http.cookiejar
+    def info(self):
+        return self.headers
+
+    # Overrides from io.IOBase
+    def close(self):
+        if not self.closed:
+            self._fp.close()
+
+        if self._connection:
+            self._connection.close()
+
+        if not self.auto_close:
+            io.IOBase.close(self)
+
+    @property
+    def closed(self):
+        if not self.auto_close:
+            return io.IOBase.closed.__get__(self)
+        elif self._fp is None:
+            return True
+        elif hasattr(self._fp, "isclosed"):
+            return self._fp.isclosed()
+        elif hasattr(self._fp, "closed"):
+            return self._fp.closed
+        else:
+            return True
+
+    def fileno(self):
+        if self._fp is None:
+            raise IOError("HTTPResponse has no file to get a fileno from")
+        elif hasattr(self._fp, "fileno"):
+            return self._fp.fileno()
+        else:
+            raise IOError(
+                "The file-like object this HTTPResponse is wrapped "
+                "around has no file descriptor"
+            )
+
+    def flush(self):
+        if (
+            self._fp is not None
+            and hasattr(self._fp, "flush")
+            and not getattr(self._fp, "closed", False)
+        ):
+            return self._fp.flush()
+
+    def readable(self):
+        # This method is required for `io` module compatibility.
+        return True
+
+    def readinto(self, b):
+        # This method is required for `io` module compatibility.
+        temp = self.read(len(b))
+        if len(temp) == 0:
+            return 0
+        else:
+            b[: len(temp)] = temp
+            return len(temp)
+
+    def supports_chunked_reads(self):
+        """
+        Checks if the underlying file-like object looks like a
+        :class:`http.client.HTTPResponse` object. We do this by testing for
+        the fp attribute. If it is present we assume it returns raw chunks as
+        processed by read_chunked().
+        """
+        return hasattr(self._fp, "fp")
+
+    def _update_chunk_length(self):
+        # First, we'll figure out length of a chunk and then
+        # we'll try to read it from socket.
+        if self.chunk_left is not None:
+            return
+        line = self._fp.fp.readline()
+        line = line.split(b";", 1)[0]
+        try:
+            self.chunk_left = int(line, 16)
+        except ValueError:
+            # Invalid chunked protocol response, abort.
+            self.close()
+            raise InvalidChunkLength(self, line)
+
+    def _handle_chunk(self, amt):
+        returned_chunk = None
+        if amt is None:
+            chunk = self._fp._safe_read(self.chunk_left)
+            returned_chunk = chunk
+            self._fp._safe_read(2)  # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+        elif amt < self.chunk_left:
+            value = self._fp._safe_read(amt)
+            self.chunk_left = self.chunk_left - amt
+            returned_chunk = value
+        elif amt == self.chunk_left:
+            value = self._fp._safe_read(amt)
+            self._fp._safe_read(2)  # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+            returned_chunk = value
+        else:  # amt > self.chunk_left
+            returned_chunk = self._fp._safe_read(self.chunk_left)
+            self._fp._safe_read(2)  # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+        return returned_chunk
+
+    def read_chunked(self, amt=None, decode_content=None):
+        """
+        Similar to :meth:`HTTPResponse.read`, but with an additional
+        parameter: ``decode_content``.
+
+        :param amt:
+            How much of the content to read. If specified, caching is skipped
+            because it doesn't make sense to cache partial content as the full
+            response.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        self._init_decoder()
+        # FIXME: Rewrite this method and make it a class with a better structured logic.
+        if not self.chunked:
+            raise ResponseNotChunked(
+                "Response is not chunked. "
+                "Header 'transfer-encoding: chunked' is missing."
+            )
+        if not self.supports_chunked_reads():
+            raise BodyNotHttplibCompatible(
+                "Body should be http.client.HTTPResponse like. "
+                "It should have have an fp attribute which returns raw chunks."
+            )
+
+        with self._error_catcher():
+            # Don't bother reading the body of a HEAD request.
+            if self._original_response and is_response_to_head(self._original_response):
+                self._original_response.close()
+                return
+
+            # If a response is already read and closed
+            # then return immediately.
+            if self._fp.fp is None:
+                return
+
+            while True:
+                self._update_chunk_length()
+                if self.chunk_left == 0:
+                    break
+                chunk = self._handle_chunk(amt)
+                decoded = self._decode(
+                    chunk, decode_content=decode_content, flush_decoder=False
+                )
+                if decoded:
+                    yield decoded
+
+            if decode_content:
+                # On CPython and PyPy, we should never need to flush the
+                # decoder. However, on Jython we *might* need to, so
+                # lets defensively do it anyway.
+                decoded = self._flush_decoder()
+                if decoded:  # Platform-specific: Jython.
+                    yield decoded
+
+            # Chunk content ends with \r\n: discard it.
+            while True:
+                line = self._fp.fp.readline()
+                if not line:
+                    # Some sites may not end with '\r\n'.
+                    break
+                if line == b"\r\n":
+                    break
+
+            # We read everything; close the "file".
+            if self._original_response:
+                self._original_response.close()
+
+    def geturl(self):
+        """
+        Returns the URL that was the source of this response.
+        If the request that generated this response redirected, this method
+        will return the final redirect location.
+        """
+        if self.retries is not None and len(self.retries.history):
+            return self.retries.history[-1].redirect_location
+        else:
+            return self._request_url
+
+    def __iter__(self):
+        buffer = []
+        for chunk in self.stream(decode_content=True):
+            if b"\n" in chunk:
+                chunk = chunk.split(b"\n")
+                yield b"".join(buffer) + chunk[0] + b"\n"
+                for x in chunk[1:-1]:
+                    yield x + b"\n"
+                if chunk[-1]:
+                    buffer = [chunk[-1]]
+                else:
+                    buffer = []
+            else:
+                buffer.append(chunk)
+        if buffer:
+            yield b"".join(buffer)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__init__.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__init__.py
new file mode 100644
index 00000000..4547fc52
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__init__.py
@@ -0,0 +1,49 @@
+from __future__ import absolute_import
+
+# For backwards compatibility, provide imports that used to be here.
+from .connection import is_connection_dropped
+from .request import SKIP_HEADER, SKIPPABLE_HEADERS, make_headers
+from .response import is_fp_closed
+from .retry import Retry
+from .ssl_ import (
+    ALPN_PROTOCOLS,
+    HAS_SNI,
+    IS_PYOPENSSL,
+    IS_SECURETRANSPORT,
+    PROTOCOL_TLS,
+    SSLContext,
+    assert_fingerprint,
+    resolve_cert_reqs,
+    resolve_ssl_version,
+    ssl_wrap_socket,
+)
+from .timeout import Timeout, current_time
+from .url import Url, get_host, parse_url, split_first
+from .wait import wait_for_read, wait_for_write
+
+__all__ = (
+    "HAS_SNI",
+    "IS_PYOPENSSL",
+    "IS_SECURETRANSPORT",
+    "SSLContext",
+    "PROTOCOL_TLS",
+    "ALPN_PROTOCOLS",
+    "Retry",
+    "Timeout",
+    "Url",
+    "assert_fingerprint",
+    "current_time",
+    "is_connection_dropped",
+    "is_fp_closed",
+    "get_host",
+    "parse_url",
+    "make_headers",
+    "resolve_cert_reqs",
+    "resolve_ssl_version",
+    "split_first",
+    "ssl_wrap_socket",
+    "wait_for_read",
+    "wait_for_write",
+    "SKIP_HEADER",
+    "SKIPPABLE_HEADERS",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e599a005
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..4ce42de8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/proxy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/proxy.cpython-312.pyc
new file mode 100644
index 00000000..d6c3cbd4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/proxy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/queue.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/queue.cpython-312.pyc
new file mode 100644
index 00000000..9a71f2fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/queue.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..5760208e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..d30f2d14
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/retry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/retry.cpython-312.pyc
new file mode 100644
index 00000000..bc336fee
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/retry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_.cpython-312.pyc
new file mode 100644
index 00000000..7bf800f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc
new file mode 100644
index 00000000..69ac3609
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssltransport.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssltransport.cpython-312.pyc
new file mode 100644
index 00000000..eecedfc8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/ssltransport.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/timeout.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/timeout.cpython-312.pyc
new file mode 100644
index 00000000..8d0fa64b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/timeout.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/url.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/url.cpython-312.pyc
new file mode 100644
index 00000000..546781db
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/url.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/wait.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/wait.cpython-312.pyc
new file mode 100644
index 00000000..e59fda1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/__pycache__/wait.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/connection.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/connection.py
new file mode 100644
index 00000000..6af1138f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/connection.py
@@ -0,0 +1,149 @@
+from __future__ import absolute_import
+
+import socket
+
+from ..contrib import _appengine_environ
+from ..exceptions import LocationParseError
+from ..packages import six
+from .wait import NoWayToWaitForSocketError, wait_for_read
+
+
+def is_connection_dropped(conn):  # Platform-specific
+    """
+    Returns True if the connection is dropped and should be closed.
+
+    :param conn:
+        :class:`http.client.HTTPConnection` object.
+
+    Note: For platforms like AppEngine, this will always return ``False`` to
+    let the platform handle connection recycling transparently for us.
+    """
+    sock = getattr(conn, "sock", False)
+    if sock is False:  # Platform-specific: AppEngine
+        return False
+    if sock is None:  # Connection already closed (such as by httplib).
+        return True
+    try:
+        # Returns True if readable, which here means it's been dropped
+        return wait_for_read(sock, timeout=0.0)
+    except NoWayToWaitForSocketError:  # Platform-specific: AppEngine
+        return False
+
+
+# This function is copied from socket.py in the Python 2.7 standard
+# library test suite. Added to its signature is only `socket_options`.
+# One additional modification is that we avoid binding to IPv6 servers
+# discovered in DNS if the system doesn't have IPv6 functionality.
+def create_connection(
+    address,
+    timeout=socket._GLOBAL_DEFAULT_TIMEOUT,
+    source_address=None,
+    socket_options=None,
+):
+    """Connect to *address* and return the socket object.
+
+    Convenience function.  Connect to *address* (a 2-tuple ``(host,
+    port)``) and return the socket object.  Passing the optional
+    *timeout* parameter will set the timeout on the socket instance
+    before attempting to connect.  If no *timeout* is supplied, the
+    global default timeout setting returned by :func:`socket.getdefaulttimeout`
+    is used.  If *source_address* is set it must be a tuple of (host, port)
+    for the socket to bind as a source address before making the connection.
+    An host of '' or port 0 tells the OS to use the default.
+    """
+
+    host, port = address
+    if host.startswith("["):
+        host = host.strip("[]")
+    err = None
+
+    # Using the value from allowed_gai_family() in the context of getaddrinfo lets
+    # us select whether to work with IPv4 DNS records, IPv6 records, or both.
+    # The original create_connection function always returns all records.
+    family = allowed_gai_family()
+
+    try:
+        host.encode("idna")
+    except UnicodeError:
+        return six.raise_from(
+            LocationParseError(u"'%s', label empty or too long" % host), None
+        )
+
+    for res in socket.getaddrinfo(host, port, family, socket.SOCK_STREAM):
+        af, socktype, proto, canonname, sa = res
+        sock = None
+        try:
+            sock = socket.socket(af, socktype, proto)
+
+            # If provided, set socket level options before connecting.
+            _set_socket_options(sock, socket_options)
+
+            if timeout is not socket._GLOBAL_DEFAULT_TIMEOUT:
+                sock.settimeout(timeout)
+            if source_address:
+                sock.bind(source_address)
+            sock.connect(sa)
+            return sock
+
+        except socket.error as e:
+            err = e
+            if sock is not None:
+                sock.close()
+                sock = None
+
+    if err is not None:
+        raise err
+
+    raise socket.error("getaddrinfo returns an empty list")
+
+
+def _set_socket_options(sock, options):
+    if options is None:
+        return
+
+    for opt in options:
+        sock.setsockopt(*opt)
+
+
+def allowed_gai_family():
+    """This function is designed to work in the context of
+    getaddrinfo, where family=socket.AF_UNSPEC is the default and
+    will perform a DNS search for both IPv6 and IPv4 records."""
+
+    family = socket.AF_INET
+    if HAS_IPV6:
+        family = socket.AF_UNSPEC
+    return family
+
+
+def _has_ipv6(host):
+    """Returns True if the system can bind an IPv6 address."""
+    sock = None
+    has_ipv6 = False
+
+    # App Engine doesn't support IPV6 sockets and actually has a quota on the
+    # number of sockets that can be used, so just early out here instead of
+    # creating a socket needlessly.
+    # See https://github.com/urllib3/urllib3/issues/1446
+    if _appengine_environ.is_appengine_sandbox():
+        return False
+
+    if socket.has_ipv6:
+        # has_ipv6 returns true if cPython was compiled with IPv6 support.
+        # It does not tell us if the system has IPv6 support enabled. To
+        # determine that we must bind to an IPv6 address.
+        # https://github.com/urllib3/urllib3/pull/611
+        # https://bugs.python.org/issue658327
+        try:
+            sock = socket.socket(socket.AF_INET6)
+            sock.bind((host, 0))
+            has_ipv6 = True
+        except Exception:
+            pass
+
+    if sock:
+        sock.close()
+    return has_ipv6
+
+
+HAS_IPV6 = _has_ipv6("::1")
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/proxy.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/proxy.py
new file mode 100644
index 00000000..2199cc7b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/proxy.py
@@ -0,0 +1,57 @@
+from .ssl_ import create_urllib3_context, resolve_cert_reqs, resolve_ssl_version
+
+
+def connection_requires_http_tunnel(
+    proxy_url=None, proxy_config=None, destination_scheme=None
+):
+    """
+    Returns True if the connection requires an HTTP CONNECT through the proxy.
+
+    :param URL proxy_url:
+        URL of the proxy.
+    :param ProxyConfig proxy_config:
+        Proxy configuration from poolmanager.py
+    :param str destination_scheme:
+        The scheme of the destination. (i.e https, http, etc)
+    """
+    # If we're not using a proxy, no way to use a tunnel.
+    if proxy_url is None:
+        return False
+
+    # HTTP destinations never require tunneling, we always forward.
+    if destination_scheme == "http":
+        return False
+
+    # Support for forwarding with HTTPS proxies and HTTPS destinations.
+    if (
+        proxy_url.scheme == "https"
+        and proxy_config
+        and proxy_config.use_forwarding_for_https
+    ):
+        return False
+
+    # Otherwise always use a tunnel.
+    return True
+
+
+def create_proxy_ssl_context(
+    ssl_version, cert_reqs, ca_certs=None, ca_cert_dir=None, ca_cert_data=None
+):
+    """
+    Generates a default proxy ssl context if one hasn't been provided by the
+    user.
+    """
+    ssl_context = create_urllib3_context(
+        ssl_version=resolve_ssl_version(ssl_version),
+        cert_reqs=resolve_cert_reqs(cert_reqs),
+    )
+
+    if (
+        not ca_certs
+        and not ca_cert_dir
+        and not ca_cert_data
+        and hasattr(ssl_context, "load_default_certs")
+    ):
+        ssl_context.load_default_certs()
+
+    return ssl_context
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/queue.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/queue.py
new file mode 100644
index 00000000..41784104
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/queue.py
@@ -0,0 +1,22 @@
+import collections
+
+from ..packages import six
+from ..packages.six.moves import queue
+
+if six.PY2:
+    # Queue is imported for side effects on MS Windows. See issue #229.
+    import Queue as _unused_module_Queue  # noqa: F401
+
+
+class LifoQueue(queue.Queue):
+    def _init(self, _):
+        self.queue = collections.deque()
+
+    def _qsize(self, len=len):
+        return len(self.queue)
+
+    def _put(self, item):
+        self.queue.append(item)
+
+    def _get(self):
+        return self.queue.pop()
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/request.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/request.py
new file mode 100644
index 00000000..330766ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/request.py
@@ -0,0 +1,137 @@
+from __future__ import absolute_import
+
+from base64 import b64encode
+
+from ..exceptions import UnrewindableBodyError
+from ..packages.six import b, integer_types
+
+# Pass as a value within ``headers`` to skip
+# emitting some HTTP headers that are added automatically.
+# The only headers that are supported are ``Accept-Encoding``,
+# ``Host``, and ``User-Agent``.
+SKIP_HEADER = "@@@SKIP_HEADER@@@"
+SKIPPABLE_HEADERS = frozenset(["accept-encoding", "host", "user-agent"])
+
+ACCEPT_ENCODING = "gzip,deflate"
+
+_FAILEDTELL = object()
+
+
+def make_headers(
+    keep_alive=None,
+    accept_encoding=None,
+    user_agent=None,
+    basic_auth=None,
+    proxy_basic_auth=None,
+    disable_cache=None,
+):
+    """
+    Shortcuts for generating request headers.
+
+    :param keep_alive:
+        If ``True``, adds 'connection: keep-alive' header.
+
+    :param accept_encoding:
+        Can be a boolean, list, or string.
+        ``True`` translates to 'gzip,deflate'.
+        List will get joined by comma.
+        String will be used as provided.
+
+    :param user_agent:
+        String representing the user-agent you want, such as
+        "python-urllib3/0.6"
+
+    :param basic_auth:
+        Colon-separated username:password string for 'authorization: basic ...'
+        auth header.
+
+    :param proxy_basic_auth:
+        Colon-separated username:password string for 'proxy-authorization: basic ...'
+        auth header.
+
+    :param disable_cache:
+        If ``True``, adds 'cache-control: no-cache' header.
+
+    Example::
+
+        >>> make_headers(keep_alive=True, user_agent="Batman/1.0")
+        {'connection': 'keep-alive', 'user-agent': 'Batman/1.0'}
+        >>> make_headers(accept_encoding=True)
+        {'accept-encoding': 'gzip,deflate'}
+    """
+    headers = {}
+    if accept_encoding:
+        if isinstance(accept_encoding, str):
+            pass
+        elif isinstance(accept_encoding, list):
+            accept_encoding = ",".join(accept_encoding)
+        else:
+            accept_encoding = ACCEPT_ENCODING
+        headers["accept-encoding"] = accept_encoding
+
+    if user_agent:
+        headers["user-agent"] = user_agent
+
+    if keep_alive:
+        headers["connection"] = "keep-alive"
+
+    if basic_auth:
+        headers["authorization"] = "Basic " + b64encode(b(basic_auth)).decode("utf-8")
+
+    if proxy_basic_auth:
+        headers["proxy-authorization"] = "Basic " + b64encode(
+            b(proxy_basic_auth)
+        ).decode("utf-8")
+
+    if disable_cache:
+        headers["cache-control"] = "no-cache"
+
+    return headers
+
+
+def set_file_position(body, pos):
+    """
+    If a position is provided, move file to that point.
+    Otherwise, we'll attempt to record a position for future use.
+    """
+    if pos is not None:
+        rewind_body(body, pos)
+    elif getattr(body, "tell", None) is not None:
+        try:
+            pos = body.tell()
+        except (IOError, OSError):
+            # This differentiates from None, allowing us to catch
+            # a failed `tell()` later when trying to rewind the body.
+            pos = _FAILEDTELL
+
+    return pos
+
+
+def rewind_body(body, body_pos):
+    """
+    Attempt to rewind body to a certain position.
+    Primarily used for request redirects and retries.
+
+    :param body:
+        File-like object that supports seek.
+
+    :param int pos:
+        Position to seek to in file.
+    """
+    body_seek = getattr(body, "seek", None)
+    if body_seek is not None and isinstance(body_pos, integer_types):
+        try:
+            body_seek(body_pos)
+        except (IOError, OSError):
+            raise UnrewindableBodyError(
+                "An error occurred when rewinding request body for redirect/retry."
+            )
+    elif body_pos is _FAILEDTELL:
+        raise UnrewindableBodyError(
+            "Unable to record file position for rewinding "
+            "request body during a redirect/retry."
+        )
+    else:
+        raise ValueError(
+            "body_pos must be of type integer, instead it was %s." % type(body_pos)
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/response.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/response.py
new file mode 100644
index 00000000..5ea609cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/response.py
@@ -0,0 +1,107 @@
+from __future__ import absolute_import
+
+from email.errors import MultipartInvariantViolationDefect, StartBoundaryNotFoundDefect
+
+from ..exceptions import HeaderParsingError
+from ..packages.six.moves import http_client as httplib
+
+
+def is_fp_closed(obj):
+    """
+    Checks whether a given file-like object is closed.
+
+    :param obj:
+        The file-like object to check.
+    """
+
+    try:
+        # Check `isclosed()` first, in case Python3 doesn't set `closed`.
+        # GH Issue #928
+        return obj.isclosed()
+    except AttributeError:
+        pass
+
+    try:
+        # Check via the official file-like-object way.
+        return obj.closed
+    except AttributeError:
+        pass
+
+    try:
+        # Check if the object is a container for another file-like object that
+        # gets released on exhaustion (e.g. HTTPResponse).
+        return obj.fp is None
+    except AttributeError:
+        pass
+
+    raise ValueError("Unable to determine whether fp is closed.")
+
+
+def assert_header_parsing(headers):
+    """
+    Asserts whether all headers have been successfully parsed.
+    Extracts encountered errors from the result of parsing headers.
+
+    Only works on Python 3.
+
+    :param http.client.HTTPMessage headers: Headers to verify.
+
+    :raises urllib3.exceptions.HeaderParsingError:
+        If parsing errors are found.
+    """
+
+    # This will fail silently if we pass in the wrong kind of parameter.
+    # To make debugging easier add an explicit check.
+    if not isinstance(headers, httplib.HTTPMessage):
+        raise TypeError("expected httplib.Message, got {0}.".format(type(headers)))
+
+    defects = getattr(headers, "defects", None)
+    get_payload = getattr(headers, "get_payload", None)
+
+    unparsed_data = None
+    if get_payload:
+        # get_payload is actually email.message.Message.get_payload;
+        # we're only interested in the result if it's not a multipart message
+        if not headers.is_multipart():
+            payload = get_payload()
+
+            if isinstance(payload, (bytes, str)):
+                unparsed_data = payload
+    if defects:
+        # httplib is assuming a response body is available
+        # when parsing headers even when httplib only sends
+        # header data to parse_headers() This results in
+        # defects on multipart responses in particular.
+        # See: https://github.com/urllib3/urllib3/issues/800
+
+        # So we ignore the following defects:
+        # - StartBoundaryNotFoundDefect:
+        #     The claimed start boundary was never found.
+        # - MultipartInvariantViolationDefect:
+        #     A message claimed to be a multipart but no subparts were found.
+        defects = [
+            defect
+            for defect in defects
+            if not isinstance(
+                defect, (StartBoundaryNotFoundDefect, MultipartInvariantViolationDefect)
+            )
+        ]
+
+    if defects or unparsed_data:
+        raise HeaderParsingError(defects=defects, unparsed_data=unparsed_data)
+
+
+def is_response_to_head(response):
+    """
+    Checks whether the request of a response has been a HEAD-request.
+    Handles the quirks of AppEngine.
+
+    :param http.client.HTTPResponse response:
+        Response to check if the originating request
+        used 'HEAD' as a method.
+    """
+    # FIXME: Can we do this somehow without accessing private httplib _method?
+    method = response._method
+    if isinstance(method, int):  # Platform-specific: Appengine
+        return method == 3
+    return method.upper() == "HEAD"
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/retry.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/retry.py
new file mode 100644
index 00000000..9a1e90d0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/retry.py
@@ -0,0 +1,622 @@
+from __future__ import absolute_import
+
+import email
+import logging
+import re
+import time
+import warnings
+from collections import namedtuple
+from itertools import takewhile
+
+from ..exceptions import (
+    ConnectTimeoutError,
+    InvalidHeader,
+    MaxRetryError,
+    ProtocolError,
+    ProxyError,
+    ReadTimeoutError,
+    ResponseError,
+)
+from ..packages import six
+
+log = logging.getLogger(__name__)
+
+
+# Data structure for representing the metadata of requests that result in a retry.
+RequestHistory = namedtuple(
+    "RequestHistory", ["method", "url", "error", "status", "redirect_location"]
+)
+
+
+# TODO: In v2 we can remove this sentinel and metaclass with deprecated options.
+_Default = object()
+
+
+class _RetryMeta(type):
+    @property
+    def DEFAULT_METHOD_WHITELIST(cls):
+        warnings.warn(
+            "Using 'Retry.DEFAULT_METHOD_WHITELIST' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_ALLOWED_METHODS' instead",
+            DeprecationWarning,
+        )
+        return cls.DEFAULT_ALLOWED_METHODS
+
+    @DEFAULT_METHOD_WHITELIST.setter
+    def DEFAULT_METHOD_WHITELIST(cls, value):
+        warnings.warn(
+            "Using 'Retry.DEFAULT_METHOD_WHITELIST' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_ALLOWED_METHODS' instead",
+            DeprecationWarning,
+        )
+        cls.DEFAULT_ALLOWED_METHODS = value
+
+    @property
+    def DEFAULT_REDIRECT_HEADERS_BLACKLIST(cls):
+        warnings.warn(
+            "Using 'Retry.DEFAULT_REDIRECT_HEADERS_BLACKLIST' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_REMOVE_HEADERS_ON_REDIRECT' instead",
+            DeprecationWarning,
+        )
+        return cls.DEFAULT_REMOVE_HEADERS_ON_REDIRECT
+
+    @DEFAULT_REDIRECT_HEADERS_BLACKLIST.setter
+    def DEFAULT_REDIRECT_HEADERS_BLACKLIST(cls, value):
+        warnings.warn(
+            "Using 'Retry.DEFAULT_REDIRECT_HEADERS_BLACKLIST' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_REMOVE_HEADERS_ON_REDIRECT' instead",
+            DeprecationWarning,
+        )
+        cls.DEFAULT_REMOVE_HEADERS_ON_REDIRECT = value
+
+    @property
+    def BACKOFF_MAX(cls):
+        warnings.warn(
+            "Using 'Retry.BACKOFF_MAX' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_BACKOFF_MAX' instead",
+            DeprecationWarning,
+        )
+        return cls.DEFAULT_BACKOFF_MAX
+
+    @BACKOFF_MAX.setter
+    def BACKOFF_MAX(cls, value):
+        warnings.warn(
+            "Using 'Retry.BACKOFF_MAX' is deprecated and "
+            "will be removed in v2.0. Use 'Retry.DEFAULT_BACKOFF_MAX' instead",
+            DeprecationWarning,
+        )
+        cls.DEFAULT_BACKOFF_MAX = value
+
+
+@six.add_metaclass(_RetryMeta)
+class Retry(object):
+    """Retry configuration.
+
+    Each retry attempt will create a new Retry object with updated values, so
+    they can be safely reused.
+
+    Retries can be defined as a default for a pool::
+
+        retries = Retry(connect=5, read=2, redirect=5)
+        http = PoolManager(retries=retries)
+        response = http.request('GET', 'http://example.com/')
+
+    Or per-request (which overrides the default for the pool)::
+
+        response = http.request('GET', 'http://example.com/', retries=Retry(10))
+
+    Retries can be disabled by passing ``False``::
+
+        response = http.request('GET', 'http://example.com/', retries=False)
+
+    Errors will be wrapped in :class:`~urllib3.exceptions.MaxRetryError` unless
+    retries are disabled, in which case the causing exception will be raised.
+
+    :param int total:
+        Total number of retries to allow. Takes precedence over other counts.
+
+        Set to ``None`` to remove this constraint and fall back on other
+        counts.
+
+        Set to ``0`` to fail on the first retry.
+
+        Set to ``False`` to disable and imply ``raise_on_redirect=False``.
+
+    :param int connect:
+        How many connection-related errors to retry on.
+
+        These are errors raised before the request is sent to the remote server,
+        which we assume has not triggered the server to process the request.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int read:
+        How many times to retry on read errors.
+
+        These errors are raised after the request was sent to the server, so the
+        request may have side-effects.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int redirect:
+        How many redirects to perform. Limit this to avoid infinite redirect
+        loops.
+
+        A redirect is a HTTP response with a status code 301, 302, 303, 307 or
+        308.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+        Set to ``False`` to disable and imply ``raise_on_redirect=False``.
+
+    :param int status:
+        How many times to retry on bad status codes.
+
+        These are retries made on responses, where status code matches
+        ``status_forcelist``.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int other:
+        How many times to retry on other errors.
+
+        Other errors are errors that are not connect, read, redirect or status errors.
+        These errors might be raised after the request was sent to the server, so the
+        request might have side-effects.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+        If ``total`` is not set, it's a good idea to set this to 0 to account
+        for unexpected edge cases and avoid infinite retry loops.
+
+    :param iterable allowed_methods:
+        Set of uppercased HTTP method verbs that we should retry on.
+
+        By default, we only retry on methods which are considered to be
+        idempotent (multiple requests with the same parameters end with the
+        same state). See :attr:`Retry.DEFAULT_ALLOWED_METHODS`.
+
+        Set to a ``False`` value to retry on any verb.
+
+        .. warning::
+
+            Previously this parameter was named ``method_whitelist``, that
+            usage is deprecated in v1.26.0 and will be removed in v2.0.
+
+    :param iterable status_forcelist:
+        A set of integer HTTP status codes that we should force a retry on.
+        A retry is initiated if the request method is in ``allowed_methods``
+        and the response status code is in ``status_forcelist``.
+
+        By default, this is disabled with ``None``.
+
+    :param float backoff_factor:
+        A backoff factor to apply between attempts after the second try
+        (most errors are resolved immediately by a second try without a
+        delay). urllib3 will sleep for::
+
+            {backoff factor} * (2 ** ({number of total retries} - 1))
+
+        seconds. If the backoff_factor is 0.1, then :func:`.sleep` will sleep
+        for [0.0s, 0.2s, 0.4s, ...] between retries. It will never be longer
+        than :attr:`Retry.DEFAULT_BACKOFF_MAX`.
+
+        By default, backoff is disabled (set to 0).
+
+    :param bool raise_on_redirect: Whether, if the number of redirects is
+        exhausted, to raise a MaxRetryError, or to return a response with a
+        response code in the 3xx range.
+
+    :param bool raise_on_status: Similar meaning to ``raise_on_redirect``:
+        whether we should raise an exception, or return a response,
+        if status falls in ``status_forcelist`` range and retries have
+        been exhausted.
+
+    :param tuple history: The history of the request encountered during
+        each call to :meth:`~Retry.increment`. The list is in the order
+        the requests occurred. Each list item is of class :class:`RequestHistory`.
+
+    :param bool respect_retry_after_header:
+        Whether to respect Retry-After header on status codes defined as
+        :attr:`Retry.RETRY_AFTER_STATUS_CODES` or not.
+
+    :param iterable remove_headers_on_redirect:
+        Sequence of headers to remove from the request when a response
+        indicating a redirect is returned before firing off the redirected
+        request.
+    """
+
+    #: Default methods to be used for ``allowed_methods``
+    DEFAULT_ALLOWED_METHODS = frozenset(
+        ["HEAD", "GET", "PUT", "DELETE", "OPTIONS", "TRACE"]
+    )
+
+    #: Default status codes to be used for ``status_forcelist``
+    RETRY_AFTER_STATUS_CODES = frozenset([413, 429, 503])
+
+    #: Default headers to be used for ``remove_headers_on_redirect``
+    DEFAULT_REMOVE_HEADERS_ON_REDIRECT = frozenset(
+        ["Cookie", "Authorization", "Proxy-Authorization"]
+    )
+
+    #: Maximum backoff time.
+    DEFAULT_BACKOFF_MAX = 120
+
+    def __init__(
+        self,
+        total=10,
+        connect=None,
+        read=None,
+        redirect=None,
+        status=None,
+        other=None,
+        allowed_methods=_Default,
+        status_forcelist=None,
+        backoff_factor=0,
+        raise_on_redirect=True,
+        raise_on_status=True,
+        history=None,
+        respect_retry_after_header=True,
+        remove_headers_on_redirect=_Default,
+        # TODO: Deprecated, remove in v2.0
+        method_whitelist=_Default,
+    ):
+
+        if method_whitelist is not _Default:
+            if allowed_methods is not _Default:
+                raise ValueError(
+                    "Using both 'allowed_methods' and "
+                    "'method_whitelist' together is not allowed. "
+                    "Instead only use 'allowed_methods'"
+                )
+            warnings.warn(
+                "Using 'method_whitelist' with Retry is deprecated and "
+                "will be removed in v2.0. Use 'allowed_methods' instead",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            allowed_methods = method_whitelist
+        if allowed_methods is _Default:
+            allowed_methods = self.DEFAULT_ALLOWED_METHODS
+        if remove_headers_on_redirect is _Default:
+            remove_headers_on_redirect = self.DEFAULT_REMOVE_HEADERS_ON_REDIRECT
+
+        self.total = total
+        self.connect = connect
+        self.read = read
+        self.status = status
+        self.other = other
+
+        if redirect is False or total is False:
+            redirect = 0
+            raise_on_redirect = False
+
+        self.redirect = redirect
+        self.status_forcelist = status_forcelist or set()
+        self.allowed_methods = allowed_methods
+        self.backoff_factor = backoff_factor
+        self.raise_on_redirect = raise_on_redirect
+        self.raise_on_status = raise_on_status
+        self.history = history or tuple()
+        self.respect_retry_after_header = respect_retry_after_header
+        self.remove_headers_on_redirect = frozenset(
+            [h.lower() for h in remove_headers_on_redirect]
+        )
+
+    def new(self, **kw):
+        params = dict(
+            total=self.total,
+            connect=self.connect,
+            read=self.read,
+            redirect=self.redirect,
+            status=self.status,
+            other=self.other,
+            status_forcelist=self.status_forcelist,
+            backoff_factor=self.backoff_factor,
+            raise_on_redirect=self.raise_on_redirect,
+            raise_on_status=self.raise_on_status,
+            history=self.history,
+            remove_headers_on_redirect=self.remove_headers_on_redirect,
+            respect_retry_after_header=self.respect_retry_after_header,
+        )
+
+        # TODO: If already given in **kw we use what's given to us
+        # If not given we need to figure out what to pass. We decide
+        # based on whether our class has the 'method_whitelist' property
+        # and if so we pass the deprecated 'method_whitelist' otherwise
+        # we use 'allowed_methods'. Remove in v2.0
+        if "method_whitelist" not in kw and "allowed_methods" not in kw:
+            if "method_whitelist" in self.__dict__:
+                warnings.warn(
+                    "Using 'method_whitelist' with Retry is deprecated and "
+                    "will be removed in v2.0. Use 'allowed_methods' instead",
+                    DeprecationWarning,
+                )
+                params["method_whitelist"] = self.allowed_methods
+            else:
+                params["allowed_methods"] = self.allowed_methods
+
+        params.update(kw)
+        return type(self)(**params)
+
+    @classmethod
+    def from_int(cls, retries, redirect=True, default=None):
+        """Backwards-compatibility for the old retries format."""
+        if retries is None:
+            retries = default if default is not None else cls.DEFAULT
+
+        if isinstance(retries, Retry):
+            return retries
+
+        redirect = bool(redirect) and None
+        new_retries = cls(retries, redirect=redirect)
+        log.debug("Converted retries value: %r -> %r", retries, new_retries)
+        return new_retries
+
+    def get_backoff_time(self):
+        """Formula for computing the current backoff
+
+        :rtype: float
+        """
+        # We want to consider only the last consecutive errors sequence (Ignore redirects).
+        consecutive_errors_len = len(
+            list(
+                takewhile(lambda x: x.redirect_location is None, reversed(self.history))
+            )
+        )
+        if consecutive_errors_len <= 1:
+            return 0
+
+        backoff_value = self.backoff_factor * (2 ** (consecutive_errors_len - 1))
+        return min(self.DEFAULT_BACKOFF_MAX, backoff_value)
+
+    def parse_retry_after(self, retry_after):
+        # Whitespace: https://tools.ietf.org/html/rfc7230#section-3.2.4
+        if re.match(r"^\s*[0-9]+\s*$", retry_after):
+            seconds = int(retry_after)
+        else:
+            retry_date_tuple = email.utils.parsedate_tz(retry_after)
+            if retry_date_tuple is None:
+                raise InvalidHeader("Invalid Retry-After header: %s" % retry_after)
+            if retry_date_tuple[9] is None:  # Python 2
+                # Assume UTC if no timezone was specified
+                # On Python2.7, parsedate_tz returns None for a timezone offset
+                # instead of 0 if no timezone is given, where mktime_tz treats
+                # a None timezone offset as local time.
+                retry_date_tuple = retry_date_tuple[:9] + (0,) + retry_date_tuple[10:]
+
+            retry_date = email.utils.mktime_tz(retry_date_tuple)
+            seconds = retry_date - time.time()
+
+        if seconds < 0:
+            seconds = 0
+
+        return seconds
+
+    def get_retry_after(self, response):
+        """Get the value of Retry-After in seconds."""
+
+        retry_after = response.headers.get("Retry-After")
+
+        if retry_after is None:
+            return None
+
+        return self.parse_retry_after(retry_after)
+
+    def sleep_for_retry(self, response=None):
+        retry_after = self.get_retry_after(response)
+        if retry_after:
+            time.sleep(retry_after)
+            return True
+
+        return False
+
+    def _sleep_backoff(self):
+        backoff = self.get_backoff_time()
+        if backoff <= 0:
+            return
+        time.sleep(backoff)
+
+    def sleep(self, response=None):
+        """Sleep between retry attempts.
+
+        This method will respect a server's ``Retry-After`` response header
+        and sleep the duration of the time requested. If that is not present, it
+        will use an exponential backoff. By default, the backoff factor is 0 and
+        this method will return immediately.
+        """
+
+        if self.respect_retry_after_header and response:
+            slept = self.sleep_for_retry(response)
+            if slept:
+                return
+
+        self._sleep_backoff()
+
+    def _is_connection_error(self, err):
+        """Errors when we're fairly sure that the server did not receive the
+        request, so it should be safe to retry.
+        """
+        if isinstance(err, ProxyError):
+            err = err.original_error
+        return isinstance(err, ConnectTimeoutError)
+
+    def _is_read_error(self, err):
+        """Errors that occur after the request has been started, so we should
+        assume that the server began processing it.
+        """
+        return isinstance(err, (ReadTimeoutError, ProtocolError))
+
+    def _is_method_retryable(self, method):
+        """Checks if a given HTTP method should be retried upon, depending if
+        it is included in the allowed_methods
+        """
+        # TODO: For now favor if the Retry implementation sets its own method_whitelist
+        # property outside of our constructor to avoid breaking custom implementations.
+        if "method_whitelist" in self.__dict__:
+            warnings.warn(
+                "Using 'method_whitelist' with Retry is deprecated and "
+                "will be removed in v2.0. Use 'allowed_methods' instead",
+                DeprecationWarning,
+            )
+            allowed_methods = self.method_whitelist
+        else:
+            allowed_methods = self.allowed_methods
+
+        if allowed_methods and method.upper() not in allowed_methods:
+            return False
+        return True
+
+    def is_retry(self, method, status_code, has_retry_after=False):
+        """Is this method/status code retryable? (Based on allowlists and control
+        variables such as the number of total retries to allow, whether to
+        respect the Retry-After header, whether this header is present, and
+        whether the returned status code is on the list of status codes to
+        be retried upon on the presence of the aforementioned header)
+        """
+        if not self._is_method_retryable(method):
+            return False
+
+        if self.status_forcelist and status_code in self.status_forcelist:
+            return True
+
+        return (
+            self.total
+            and self.respect_retry_after_header
+            and has_retry_after
+            and (status_code in self.RETRY_AFTER_STATUS_CODES)
+        )
+
+    def is_exhausted(self):
+        """Are we out of retries?"""
+        retry_counts = (
+            self.total,
+            self.connect,
+            self.read,
+            self.redirect,
+            self.status,
+            self.other,
+        )
+        retry_counts = list(filter(None, retry_counts))
+        if not retry_counts:
+            return False
+
+        return min(retry_counts) < 0
+
+    def increment(
+        self,
+        method=None,
+        url=None,
+        response=None,
+        error=None,
+        _pool=None,
+        _stacktrace=None,
+    ):
+        """Return a new Retry object with incremented retry counters.
+
+        :param response: A response object, or None, if the server did not
+            return a response.
+        :type response: :class:`~urllib3.response.HTTPResponse`
+        :param Exception error: An error encountered during the request, or
+            None if the response was received successfully.
+
+        :return: A new ``Retry`` object.
+        """
+        if self.total is False and error:
+            # Disabled, indicate to re-raise the error.
+            raise six.reraise(type(error), error, _stacktrace)
+
+        total = self.total
+        if total is not None:
+            total -= 1
+
+        connect = self.connect
+        read = self.read
+        redirect = self.redirect
+        status_count = self.status
+        other = self.other
+        cause = "unknown"
+        status = None
+        redirect_location = None
+
+        if error and self._is_connection_error(error):
+            # Connect retry?
+            if connect is False:
+                raise six.reraise(type(error), error, _stacktrace)
+            elif connect is not None:
+                connect -= 1
+
+        elif error and self._is_read_error(error):
+            # Read retry?
+            if read is False or not self._is_method_retryable(method):
+                raise six.reraise(type(error), error, _stacktrace)
+            elif read is not None:
+                read -= 1
+
+        elif error:
+            # Other retry?
+            if other is not None:
+                other -= 1
+
+        elif response and response.get_redirect_location():
+            # Redirect retry?
+            if redirect is not None:
+                redirect -= 1
+            cause = "too many redirects"
+            redirect_location = response.get_redirect_location()
+            status = response.status
+
+        else:
+            # Incrementing because of a server error like a 500 in
+            # status_forcelist and the given method is in the allowed_methods
+            cause = ResponseError.GENERIC_ERROR
+            if response and response.status:
+                if status_count is not None:
+                    status_count -= 1
+                cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status)
+                status = response.status
+
+        history = self.history + (
+            RequestHistory(method, url, error, status, redirect_location),
+        )
+
+        new_retry = self.new(
+            total=total,
+            connect=connect,
+            read=read,
+            redirect=redirect,
+            status=status_count,
+            other=other,
+            history=history,
+        )
+
+        if new_retry.is_exhausted():
+            raise MaxRetryError(_pool, url, error or ResponseError(cause))
+
+        log.debug("Incremented Retry for (url='%s'): %r", url, new_retry)
+
+        return new_retry
+
+    def __repr__(self):
+        return (
+            "{cls.__name__}(total={self.total}, connect={self.connect}, "
+            "read={self.read}, redirect={self.redirect}, status={self.status})"
+        ).format(cls=type(self), self=self)
+
+    def __getattr__(self, item):
+        if item == "method_whitelist":
+            # TODO: Remove this deprecated alias in v2.0
+            warnings.warn(
+                "Using 'method_whitelist' with Retry is deprecated and "
+                "will be removed in v2.0. Use 'allowed_methods' instead",
+                DeprecationWarning,
+            )
+            return self.allowed_methods
+        try:
+            return getattr(super(Retry, self), item)
+        except AttributeError:
+            return getattr(Retry, item)
+
+
+# For backwards compatibility (equivalent to pre-v1.9):
+Retry.DEFAULT = Retry(3)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_.py
new file mode 100644
index 00000000..0a6a0e06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_.py
@@ -0,0 +1,504 @@
+from __future__ import absolute_import
+
+import hashlib
+import hmac
+import os
+import sys
+import warnings
+from binascii import hexlify, unhexlify
+
+from ..exceptions import (
+    InsecurePlatformWarning,
+    ProxySchemeUnsupported,
+    SNIMissingWarning,
+    SSLError,
+)
+from ..packages import six
+from .url import BRACELESS_IPV6_ADDRZ_RE, IPV4_RE
+
+SSLContext = None
+SSLTransport = None
+HAS_SNI = False
+IS_PYOPENSSL = False
+IS_SECURETRANSPORT = False
+ALPN_PROTOCOLS = ["http/1.1"]
+
+# Maps the length of a digest to a possible hash function producing this digest
+HASHFUNC_MAP = {
+    length: getattr(hashlib, algorithm, None)
+    for length, algorithm in ((32, "md5"), (40, "sha1"), (64, "sha256"))
+}
+
+
+def _const_compare_digest_backport(a, b):
+    """
+    Compare two digests of equal length in constant time.
+
+    The digests must be of type str/bytes.
+    Returns True if the digests match, and False otherwise.
+    """
+    result = abs(len(a) - len(b))
+    for left, right in zip(bytearray(a), bytearray(b)):
+        result |= left ^ right
+    return result == 0
+
+
+_const_compare_digest = getattr(hmac, "compare_digest", _const_compare_digest_backport)
+
+try:  # Test for SSL features
+    import ssl
+    from ssl import CERT_REQUIRED, wrap_socket
+except ImportError:
+    pass
+
+try:
+    from ssl import HAS_SNI  # Has SNI?
+except ImportError:
+    pass
+
+try:
+    from .ssltransport import SSLTransport
+except ImportError:
+    pass
+
+
+try:  # Platform-specific: Python 3.6
+    from ssl import PROTOCOL_TLS
+
+    PROTOCOL_SSLv23 = PROTOCOL_TLS
+except ImportError:
+    try:
+        from ssl import PROTOCOL_SSLv23 as PROTOCOL_TLS
+
+        PROTOCOL_SSLv23 = PROTOCOL_TLS
+    except ImportError:
+        PROTOCOL_SSLv23 = PROTOCOL_TLS = 2
+
+try:
+    from ssl import PROTOCOL_TLS_CLIENT
+except ImportError:
+    PROTOCOL_TLS_CLIENT = PROTOCOL_TLS
+
+
+try:
+    from ssl import OP_NO_COMPRESSION, OP_NO_SSLv2, OP_NO_SSLv3
+except ImportError:
+    OP_NO_SSLv2, OP_NO_SSLv3 = 0x1000000, 0x2000000
+    OP_NO_COMPRESSION = 0x20000
+
+
+try:  # OP_NO_TICKET was added in Python 3.6
+    from ssl import OP_NO_TICKET
+except ImportError:
+    OP_NO_TICKET = 0x4000
+
+
+# A secure default.
+# Sources for more information on TLS ciphers:
+#
+# - https://wiki.mozilla.org/Security/Server_Side_TLS
+# - https://www.ssllabs.com/projects/best-practices/index.html
+# - https://hynek.me/articles/hardening-your-web-servers-ssl-ciphers/
+#
+# The general intent is:
+# - prefer cipher suites that offer perfect forward secrecy (DHE/ECDHE),
+# - prefer ECDHE over DHE for better performance,
+# - prefer any AES-GCM and ChaCha20 over any AES-CBC for better performance and
+#   security,
+# - prefer AES-GCM over ChaCha20 because hardware-accelerated AES is common,
+# - disable NULL authentication, MD5 MACs, DSS, and other
+#   insecure ciphers for security reasons.
+# - NOTE: TLS 1.3 cipher suites are managed through a different interface
+#   not exposed by CPython (yet!) and are enabled by default if they're available.
+DEFAULT_CIPHERS = ":".join(
+    [
+        "ECDHE+AESGCM",
+        "ECDHE+CHACHA20",
+        "DHE+AESGCM",
+        "DHE+CHACHA20",
+        "ECDH+AESGCM",
+        "DH+AESGCM",
+        "ECDH+AES",
+        "DH+AES",
+        "RSA+AESGCM",
+        "RSA+AES",
+        "!aNULL",
+        "!eNULL",
+        "!MD5",
+        "!DSS",
+    ]
+)
+
+try:
+    from ssl import SSLContext  # Modern SSL?
+except ImportError:
+
+    class SSLContext(object):  # Platform-specific: Python 2
+        def __init__(self, protocol_version):
+            self.protocol = protocol_version
+            # Use default values from a real SSLContext
+            self.check_hostname = False
+            self.verify_mode = ssl.CERT_NONE
+            self.ca_certs = None
+            self.options = 0
+            self.certfile = None
+            self.keyfile = None
+            self.ciphers = None
+
+        def load_cert_chain(self, certfile, keyfile):
+            self.certfile = certfile
+            self.keyfile = keyfile
+
+        def load_verify_locations(self, cafile=None, capath=None, cadata=None):
+            self.ca_certs = cafile
+
+            if capath is not None:
+                raise SSLError("CA directories not supported in older Pythons")
+
+            if cadata is not None:
+                raise SSLError("CA data not supported in older Pythons")
+
+        def set_ciphers(self, cipher_suite):
+            self.ciphers = cipher_suite
+
+        def wrap_socket(self, socket, server_hostname=None, server_side=False):
+            warnings.warn(
+                "A true SSLContext object is not available. This prevents "
+                "urllib3 from configuring SSL appropriately and may cause "
+                "certain SSL connections to fail. You can upgrade to a newer "
+                "version of Python to solve this. For more information, see "
+                "https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html"
+                "#ssl-warnings",
+                InsecurePlatformWarning,
+            )
+            kwargs = {
+                "keyfile": self.keyfile,
+                "certfile": self.certfile,
+                "ca_certs": self.ca_certs,
+                "cert_reqs": self.verify_mode,
+                "ssl_version": self.protocol,
+                "server_side": server_side,
+            }
+            return wrap_socket(socket, ciphers=self.ciphers, **kwargs)
+
+
+def assert_fingerprint(cert, fingerprint):
+    """
+    Checks if given fingerprint matches the supplied certificate.
+
+    :param cert:
+        Certificate as bytes object.
+    :param fingerprint:
+        Fingerprint as string of hexdigits, can be interspersed by colons.
+    """
+
+    fingerprint = fingerprint.replace(":", "").lower()
+    digest_length = len(fingerprint)
+    if digest_length not in HASHFUNC_MAP:
+        raise SSLError("Fingerprint of invalid length: {0}".format(fingerprint))
+    hashfunc = HASHFUNC_MAP.get(digest_length)
+    if hashfunc is None:
+        raise SSLError(
+            "Hash function implementation unavailable for fingerprint length: {0}".format(
+                digest_length
+            )
+        )
+
+    # We need encode() here for py32; works on py2 and p33.
+    fingerprint_bytes = unhexlify(fingerprint.encode())
+
+    cert_digest = hashfunc(cert).digest()
+
+    if not _const_compare_digest(cert_digest, fingerprint_bytes):
+        raise SSLError(
+            'Fingerprints did not match. Expected "{0}", got "{1}".'.format(
+                fingerprint, hexlify(cert_digest)
+            )
+        )
+
+
+def resolve_cert_reqs(candidate):
+    """
+    Resolves the argument to a numeric constant, which can be passed to
+    the wrap_socket function/method from the ssl module.
+    Defaults to :data:`ssl.CERT_REQUIRED`.
+    If given a string it is assumed to be the name of the constant in the
+    :mod:`ssl` module or its abbreviation.
+    (So you can specify `REQUIRED` instead of `CERT_REQUIRED`.
+    If it's neither `None` nor a string we assume it is already the numeric
+    constant which can directly be passed to wrap_socket.
+    """
+    if candidate is None:
+        return CERT_REQUIRED
+
+    if isinstance(candidate, str):
+        res = getattr(ssl, candidate, None)
+        if res is None:
+            res = getattr(ssl, "CERT_" + candidate)
+        return res
+
+    return candidate
+
+
+def resolve_ssl_version(candidate):
+    """
+    like resolve_cert_reqs
+    """
+    if candidate is None:
+        return PROTOCOL_TLS
+
+    if isinstance(candidate, str):
+        res = getattr(ssl, candidate, None)
+        if res is None:
+            res = getattr(ssl, "PROTOCOL_" + candidate)
+        return res
+
+    return candidate
+
+
+def create_urllib3_context(
+    ssl_version=None, cert_reqs=None, options=None, ciphers=None
+):
+    """All arguments have the same meaning as ``ssl_wrap_socket``.
+
+    By default, this function does a lot of the same work that
+    ``ssl.create_default_context`` does on Python 3.4+. It:
+
+    - Disables SSLv2, SSLv3, and compression
+    - Sets a restricted set of server ciphers
+
+    If you wish to enable SSLv3, you can do::
+
+        from pip._vendor.urllib3.util import ssl_
+        context = ssl_.create_urllib3_context()
+        context.options &= ~ssl_.OP_NO_SSLv3
+
+    You can do the same to enable compression (substituting ``COMPRESSION``
+    for ``SSLv3`` in the last line above).
+
+    :param ssl_version:
+        The desired protocol version to use. This will default to
+        PROTOCOL_SSLv23 which will negotiate the highest protocol that both
+        the server and your installation of OpenSSL support.
+    :param cert_reqs:
+        Whether to require the certificate verification. This defaults to
+        ``ssl.CERT_REQUIRED``.
+    :param options:
+        Specific OpenSSL options. These default to ``ssl.OP_NO_SSLv2``,
+        ``ssl.OP_NO_SSLv3``, ``ssl.OP_NO_COMPRESSION``, and ``ssl.OP_NO_TICKET``.
+    :param ciphers:
+        Which cipher suites to allow the server to select.
+    :returns:
+        Constructed SSLContext object with specified options
+    :rtype: SSLContext
+    """
+    # PROTOCOL_TLS is deprecated in Python 3.10
+    if not ssl_version or ssl_version == PROTOCOL_TLS:
+        ssl_version = PROTOCOL_TLS_CLIENT
+
+    context = SSLContext(ssl_version)
+
+    context.set_ciphers(ciphers or DEFAULT_CIPHERS)
+
+    # Setting the default here, as we may have no ssl module on import
+    cert_reqs = ssl.CERT_REQUIRED if cert_reqs is None else cert_reqs
+
+    if options is None:
+        options = 0
+        # SSLv2 is easily broken and is considered harmful and dangerous
+        options |= OP_NO_SSLv2
+        # SSLv3 has several problems and is now dangerous
+        options |= OP_NO_SSLv3
+        # Disable compression to prevent CRIME attacks for OpenSSL 1.0+
+        # (issue #309)
+        options |= OP_NO_COMPRESSION
+        # TLSv1.2 only. Unless set explicitly, do not request tickets.
+        # This may save some bandwidth on wire, and although the ticket is encrypted,
+        # there is a risk associated with it being on wire,
+        # if the server is not rotating its ticketing keys properly.
+        options |= OP_NO_TICKET
+
+    context.options |= options
+
+    # Enable post-handshake authentication for TLS 1.3, see GH #1634. PHA is
+    # necessary for conditional client cert authentication with TLS 1.3.
+    # The attribute is None for OpenSSL <= 1.1.0 or does not exist in older
+    # versions of Python.  We only enable on Python 3.7.4+ or if certificate
+    # verification is enabled to work around Python issue #37428
+    # See: https://bugs.python.org/issue37428
+    if (cert_reqs == ssl.CERT_REQUIRED or sys.version_info >= (3, 7, 4)) and getattr(
+        context, "post_handshake_auth", None
+    ) is not None:
+        context.post_handshake_auth = True
+
+    def disable_check_hostname():
+        if (
+            getattr(context, "check_hostname", None) is not None
+        ):  # Platform-specific: Python 3.2
+            # We do our own verification, including fingerprints and alternative
+            # hostnames. So disable it here
+            context.check_hostname = False
+
+    # The order of the below lines setting verify_mode and check_hostname
+    # matter due to safe-guards SSLContext has to prevent an SSLContext with
+    # check_hostname=True, verify_mode=NONE/OPTIONAL. This is made even more
+    # complex because we don't know whether PROTOCOL_TLS_CLIENT will be used
+    # or not so we don't know the initial state of the freshly created SSLContext.
+    if cert_reqs == ssl.CERT_REQUIRED:
+        context.verify_mode = cert_reqs
+        disable_check_hostname()
+    else:
+        disable_check_hostname()
+        context.verify_mode = cert_reqs
+
+    # Enable logging of TLS session keys via defacto standard environment variable
+    # 'SSLKEYLOGFILE', if the feature is available (Python 3.8+). Skip empty values.
+    if hasattr(context, "keylog_filename"):
+        sslkeylogfile = os.environ.get("SSLKEYLOGFILE")
+        if sslkeylogfile:
+            context.keylog_filename = sslkeylogfile
+
+    return context
+
+
+def ssl_wrap_socket(
+    sock,
+    keyfile=None,
+    certfile=None,
+    cert_reqs=None,
+    ca_certs=None,
+    server_hostname=None,
+    ssl_version=None,
+    ciphers=None,
+    ssl_context=None,
+    ca_cert_dir=None,
+    key_password=None,
+    ca_cert_data=None,
+    tls_in_tls=False,
+):
+    """
+    All arguments except for server_hostname, ssl_context, and ca_cert_dir have
+    the same meaning as they do when using :func:`ssl.wrap_socket`.
+
+    :param server_hostname:
+        When SNI is supported, the expected hostname of the certificate
+    :param ssl_context:
+        A pre-made :class:`SSLContext` object. If none is provided, one will
+        be created using :func:`create_urllib3_context`.
+    :param ciphers:
+        A string of ciphers we wish the client to support.
+    :param ca_cert_dir:
+        A directory containing CA certificates in multiple separate files, as
+        supported by OpenSSL's -CApath flag or the capath argument to
+        SSLContext.load_verify_locations().
+    :param key_password:
+        Optional password if the keyfile is encrypted.
+    :param ca_cert_data:
+        Optional string containing CA certificates in PEM format suitable for
+        passing as the cadata parameter to SSLContext.load_verify_locations()
+    :param tls_in_tls:
+        Use SSLTransport to wrap the existing socket.
+    """
+    context = ssl_context
+    if context is None:
+        # Note: This branch of code and all the variables in it are no longer
+        # used by urllib3 itself. We should consider deprecating and removing
+        # this code.
+        context = create_urllib3_context(ssl_version, cert_reqs, ciphers=ciphers)
+
+    if ca_certs or ca_cert_dir or ca_cert_data:
+        try:
+            context.load_verify_locations(ca_certs, ca_cert_dir, ca_cert_data)
+        except (IOError, OSError) as e:
+            raise SSLError(e)
+
+    elif ssl_context is None and hasattr(context, "load_default_certs"):
+        # try to load OS default certs; works well on Windows (require Python3.4+)
+        context.load_default_certs()
+
+    # Attempt to detect if we get the goofy behavior of the
+    # keyfile being encrypted and OpenSSL asking for the
+    # passphrase via the terminal and instead error out.
+    if keyfile and key_password is None and _is_key_file_encrypted(keyfile):
+        raise SSLError("Client private key is encrypted, password is required")
+
+    if certfile:
+        if key_password is None:
+            context.load_cert_chain(certfile, keyfile)
+        else:
+            context.load_cert_chain(certfile, keyfile, key_password)
+
+    try:
+        if hasattr(context, "set_alpn_protocols"):
+            context.set_alpn_protocols(ALPN_PROTOCOLS)
+    except NotImplementedError:  # Defensive: in CI, we always have set_alpn_protocols
+        pass
+
+    # If we detect server_hostname is an IP address then the SNI
+    # extension should not be used according to RFC3546 Section 3.1
+    use_sni_hostname = server_hostname and not is_ipaddress(server_hostname)
+    # SecureTransport uses server_hostname in certificate verification.
+    send_sni = (use_sni_hostname and HAS_SNI) or (
+        IS_SECURETRANSPORT and server_hostname
+    )
+    # Do not warn the user if server_hostname is an invalid SNI hostname.
+    if not HAS_SNI and use_sni_hostname:
+        warnings.warn(
+            "An HTTPS request has been made, but the SNI (Server Name "
+            "Indication) extension to TLS is not available on this platform. "
+            "This may cause the server to present an incorrect TLS "
+            "certificate, which can cause validation failures. You can upgrade to "
+            "a newer version of Python to solve this. For more information, see "
+            "https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html"
+            "#ssl-warnings",
+            SNIMissingWarning,
+        )
+
+    if send_sni:
+        ssl_sock = _ssl_wrap_socket_impl(
+            sock, context, tls_in_tls, server_hostname=server_hostname
+        )
+    else:
+        ssl_sock = _ssl_wrap_socket_impl(sock, context, tls_in_tls)
+    return ssl_sock
+
+
+def is_ipaddress(hostname):
+    """Detects whether the hostname given is an IPv4 or IPv6 address.
+    Also detects IPv6 addresses with Zone IDs.
+
+    :param str hostname: Hostname to examine.
+    :return: True if the hostname is an IP address, False otherwise.
+    """
+    if not six.PY2 and isinstance(hostname, bytes):
+        # IDN A-label bytes are ASCII compatible.
+        hostname = hostname.decode("ascii")
+    return bool(IPV4_RE.match(hostname) or BRACELESS_IPV6_ADDRZ_RE.match(hostname))
+
+
+def _is_key_file_encrypted(key_file):
+    """Detects if a key file is encrypted or not."""
+    with open(key_file, "r") as f:
+        for line in f:
+            # Look for Proc-Type: 4,ENCRYPTED
+            if "ENCRYPTED" in line:
+                return True
+
+    return False
+
+
+def _ssl_wrap_socket_impl(sock, ssl_context, tls_in_tls, server_hostname=None):
+    if tls_in_tls:
+        if not SSLTransport:
+            # Import error, ssl is not available.
+            raise ProxySchemeUnsupported(
+                "TLS in TLS requires support for the 'ssl' module"
+            )
+
+        SSLTransport._validate_ssl_context_for_tls_in_tls(ssl_context)
+        return SSLTransport(sock, ssl_context, server_hostname)
+
+    if server_hostname:
+        return ssl_context.wrap_socket(sock, server_hostname=server_hostname)
+    else:
+        return ssl_context.wrap_socket(sock)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_match_hostname.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_match_hostname.py
new file mode 100644
index 00000000..1dd950c4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssl_match_hostname.py
@@ -0,0 +1,159 @@
+"""The match_hostname() function from Python 3.3.3, essential when using SSL."""
+
+# Note: This file is under the PSF license as the code comes from the python
+# stdlib.   http://docs.python.org/3/license.html
+
+import re
+import sys
+
+# ipaddress has been backported to 2.6+ in pypi.  If it is installed on the
+# system, use it to handle IPAddress ServerAltnames (this was added in
+# python-3.5) otherwise only do DNS matching.  This allows
+# util.ssl_match_hostname to continue to be used in Python 2.7.
+try:
+    import ipaddress
+except ImportError:
+    ipaddress = None
+
+__version__ = "3.5.0.1"
+
+
+class CertificateError(ValueError):
+    pass
+
+
+def _dnsname_match(dn, hostname, max_wildcards=1):
+    """Matching according to RFC 6125, section 6.4.3
+
+    http://tools.ietf.org/html/rfc6125#section-6.4.3
+    """
+    pats = []
+    if not dn:
+        return False
+
+    # Ported from python3-syntax:
+    # leftmost, *remainder = dn.split(r'.')
+    parts = dn.split(r".")
+    leftmost = parts[0]
+    remainder = parts[1:]
+
+    wildcards = leftmost.count("*")
+    if wildcards > max_wildcards:
+        # Issue #17980: avoid denials of service by refusing more
+        # than one wildcard per fragment.  A survey of established
+        # policy among SSL implementations showed it to be a
+        # reasonable choice.
+        raise CertificateError(
+            "too many wildcards in certificate DNS name: " + repr(dn)
+        )
+
+    # speed up common case w/o wildcards
+    if not wildcards:
+        return dn.lower() == hostname.lower()
+
+    # RFC 6125, section 6.4.3, subitem 1.
+    # The client SHOULD NOT attempt to match a presented identifier in which
+    # the wildcard character comprises a label other than the left-most label.
+    if leftmost == "*":
+        # When '*' is a fragment by itself, it matches a non-empty dotless
+        # fragment.
+        pats.append("[^.]+")
+    elif leftmost.startswith("xn--") or hostname.startswith("xn--"):
+        # RFC 6125, section 6.4.3, subitem 3.
+        # The client SHOULD NOT attempt to match a presented identifier
+        # where the wildcard character is embedded within an A-label or
+        # U-label of an internationalized domain name.
+        pats.append(re.escape(leftmost))
+    else:
+        # Otherwise, '*' matches any dotless string, e.g. www*
+        pats.append(re.escape(leftmost).replace(r"\*", "[^.]*"))
+
+    # add the remaining fragments, ignore any wildcards
+    for frag in remainder:
+        pats.append(re.escape(frag))
+
+    pat = re.compile(r"\A" + r"\.".join(pats) + r"\Z", re.IGNORECASE)
+    return pat.match(hostname)
+
+
+def _to_unicode(obj):
+    if isinstance(obj, str) and sys.version_info < (3,):
+        # ignored flake8 # F821 to support python 2.7 function
+        obj = unicode(obj, encoding="ascii", errors="strict")  # noqa: F821
+    return obj
+
+
+def _ipaddress_match(ipname, host_ip):
+    """Exact matching of IP addresses.
+
+    RFC 6125 explicitly doesn't define an algorithm for this
+    (section 1.7.2 - "Out of Scope").
+    """
+    # OpenSSL may add a trailing newline to a subjectAltName's IP address
+    # Divergence from upstream: ipaddress can't handle byte str
+    ip = ipaddress.ip_address(_to_unicode(ipname).rstrip())
+    return ip == host_ip
+
+
+def match_hostname(cert, hostname):
+    """Verify that *cert* (in decoded format as returned by
+    SSLSocket.getpeercert()) matches the *hostname*.  RFC 2818 and RFC 6125
+    rules are followed, but IP addresses are not accepted for *hostname*.
+
+    CertificateError is raised on failure. On success, the function
+    returns nothing.
+    """
+    if not cert:
+        raise ValueError(
+            "empty or no certificate, match_hostname needs a "
+            "SSL socket or SSL context with either "
+            "CERT_OPTIONAL or CERT_REQUIRED"
+        )
+    try:
+        # Divergence from upstream: ipaddress can't handle byte str
+        host_ip = ipaddress.ip_address(_to_unicode(hostname))
+    except (UnicodeError, ValueError):
+        # ValueError: Not an IP address (common case)
+        # UnicodeError: Divergence from upstream: Have to deal with ipaddress not taking
+        # byte strings.  addresses should be all ascii, so we consider it not
+        # an ipaddress in this case
+        host_ip = None
+    except AttributeError:
+        # Divergence from upstream: Make ipaddress library optional
+        if ipaddress is None:
+            host_ip = None
+        else:  # Defensive
+            raise
+    dnsnames = []
+    san = cert.get("subjectAltName", ())
+    for key, value in san:
+        if key == "DNS":
+            if host_ip is None and _dnsname_match(value, hostname):
+                return
+            dnsnames.append(value)
+        elif key == "IP Address":
+            if host_ip is not None and _ipaddress_match(value, host_ip):
+                return
+            dnsnames.append(value)
+    if not dnsnames:
+        # The subject is only checked when there is no dNSName entry
+        # in subjectAltName
+        for sub in cert.get("subject", ()):
+            for key, value in sub:
+                # XXX according to RFC 2818, the most specific Common Name
+                # must be used.
+                if key == "commonName":
+                    if _dnsname_match(value, hostname):
+                        return
+                    dnsnames.append(value)
+    if len(dnsnames) > 1:
+        raise CertificateError(
+            "hostname %r "
+            "doesn't match either of %s" % (hostname, ", ".join(map(repr, dnsnames)))
+        )
+    elif len(dnsnames) == 1:
+        raise CertificateError("hostname %r doesn't match %r" % (hostname, dnsnames[0]))
+    else:
+        raise CertificateError(
+            "no appropriate commonName or subjectAltName fields were found"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssltransport.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssltransport.py
new file mode 100644
index 00000000..4a7105d1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/ssltransport.py
@@ -0,0 +1,221 @@
+import io
+import socket
+import ssl
+
+from ..exceptions import ProxySchemeUnsupported
+from ..packages import six
+
+SSL_BLOCKSIZE = 16384
+
+
+class SSLTransport:
+    """
+    The SSLTransport wraps an existing socket and establishes an SSL connection.
+
+    Contrary to Python's implementation of SSLSocket, it allows you to chain
+    multiple TLS connections together. It's particularly useful if you need to
+    implement TLS within TLS.
+
+    The class supports most of the socket API operations.
+    """
+
+    @staticmethod
+    def _validate_ssl_context_for_tls_in_tls(ssl_context):
+        """
+        Raises a ProxySchemeUnsupported if the provided ssl_context can't be used
+        for TLS in TLS.
+
+        The only requirement is that the ssl_context provides the 'wrap_bio'
+        methods.
+        """
+
+        if not hasattr(ssl_context, "wrap_bio"):
+            if six.PY2:
+                raise ProxySchemeUnsupported(
+                    "TLS in TLS requires SSLContext.wrap_bio() which isn't "
+                    "supported on Python 2"
+                )
+            else:
+                raise ProxySchemeUnsupported(
+                    "TLS in TLS requires SSLContext.wrap_bio() which isn't "
+                    "available on non-native SSLContext"
+                )
+
+    def __init__(
+        self, socket, ssl_context, server_hostname=None, suppress_ragged_eofs=True
+    ):
+        """
+        Create an SSLTransport around socket using the provided ssl_context.
+        """
+        self.incoming = ssl.MemoryBIO()
+        self.outgoing = ssl.MemoryBIO()
+
+        self.suppress_ragged_eofs = suppress_ragged_eofs
+        self.socket = socket
+
+        self.sslobj = ssl_context.wrap_bio(
+            self.incoming, self.outgoing, server_hostname=server_hostname
+        )
+
+        # Perform initial handshake.
+        self._ssl_io_loop(self.sslobj.do_handshake)
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *_):
+        self.close()
+
+    def fileno(self):
+        return self.socket.fileno()
+
+    def read(self, len=1024, buffer=None):
+        return self._wrap_ssl_read(len, buffer)
+
+    def recv(self, len=1024, flags=0):
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to recv")
+        return self._wrap_ssl_read(len)
+
+    def recv_into(self, buffer, nbytes=None, flags=0):
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to recv_into")
+        if buffer and (nbytes is None):
+            nbytes = len(buffer)
+        elif nbytes is None:
+            nbytes = 1024
+        return self.read(nbytes, buffer)
+
+    def sendall(self, data, flags=0):
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to sendall")
+        count = 0
+        with memoryview(data) as view, view.cast("B") as byte_view:
+            amount = len(byte_view)
+            while count < amount:
+                v = self.send(byte_view[count:])
+                count += v
+
+    def send(self, data, flags=0):
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to send")
+        response = self._ssl_io_loop(self.sslobj.write, data)
+        return response
+
+    def makefile(
+        self, mode="r", buffering=None, encoding=None, errors=None, newline=None
+    ):
+        """
+        Python's httpclient uses makefile and buffered io when reading HTTP
+        messages and we need to support it.
+
+        This is unfortunately a copy and paste of socket.py makefile with small
+        changes to point to the socket directly.
+        """
+        if not set(mode) <= {"r", "w", "b"}:
+            raise ValueError("invalid mode %r (only r, w, b allowed)" % (mode,))
+
+        writing = "w" in mode
+        reading = "r" in mode or not writing
+        assert reading or writing
+        binary = "b" in mode
+        rawmode = ""
+        if reading:
+            rawmode += "r"
+        if writing:
+            rawmode += "w"
+        raw = socket.SocketIO(self, rawmode)
+        self.socket._io_refs += 1
+        if buffering is None:
+            buffering = -1
+        if buffering < 0:
+            buffering = io.DEFAULT_BUFFER_SIZE
+        if buffering == 0:
+            if not binary:
+                raise ValueError("unbuffered streams must be binary")
+            return raw
+        if reading and writing:
+            buffer = io.BufferedRWPair(raw, raw, buffering)
+        elif reading:
+            buffer = io.BufferedReader(raw, buffering)
+        else:
+            assert writing
+            buffer = io.BufferedWriter(raw, buffering)
+        if binary:
+            return buffer
+        text = io.TextIOWrapper(buffer, encoding, errors, newline)
+        text.mode = mode
+        return text
+
+    def unwrap(self):
+        self._ssl_io_loop(self.sslobj.unwrap)
+
+    def close(self):
+        self.socket.close()
+
+    def getpeercert(self, binary_form=False):
+        return self.sslobj.getpeercert(binary_form)
+
+    def version(self):
+        return self.sslobj.version()
+
+    def cipher(self):
+        return self.sslobj.cipher()
+
+    def selected_alpn_protocol(self):
+        return self.sslobj.selected_alpn_protocol()
+
+    def selected_npn_protocol(self):
+        return self.sslobj.selected_npn_protocol()
+
+    def shared_ciphers(self):
+        return self.sslobj.shared_ciphers()
+
+    def compression(self):
+        return self.sslobj.compression()
+
+    def settimeout(self, value):
+        self.socket.settimeout(value)
+
+    def gettimeout(self):
+        return self.socket.gettimeout()
+
+    def _decref_socketios(self):
+        self.socket._decref_socketios()
+
+    def _wrap_ssl_read(self, len, buffer=None):
+        try:
+            return self._ssl_io_loop(self.sslobj.read, len, buffer)
+        except ssl.SSLError as e:
+            if e.errno == ssl.SSL_ERROR_EOF and self.suppress_ragged_eofs:
+                return 0  # eof, return 0.
+            else:
+                raise
+
+    def _ssl_io_loop(self, func, *args):
+        """Performs an I/O loop between incoming/outgoing and the socket."""
+        should_loop = True
+        ret = None
+
+        while should_loop:
+            errno = None
+            try:
+                ret = func(*args)
+            except ssl.SSLError as e:
+                if e.errno not in (ssl.SSL_ERROR_WANT_READ, ssl.SSL_ERROR_WANT_WRITE):
+                    # WANT_READ, and WANT_WRITE are expected, others are not.
+                    raise e
+                errno = e.errno
+
+            buf = self.outgoing.read()
+            self.socket.sendall(buf)
+
+            if errno is None:
+                should_loop = False
+            elif errno == ssl.SSL_ERROR_WANT_READ:
+                buf = self.socket.recv(SSL_BLOCKSIZE)
+                if buf:
+                    self.incoming.write(buf)
+                else:
+                    self.incoming.write_eof()
+        return ret
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/timeout.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/timeout.py
new file mode 100644
index 00000000..78e18a62
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/timeout.py
@@ -0,0 +1,271 @@
+from __future__ import absolute_import
+
+import time
+
+# The default socket timeout, used by httplib to indicate that no timeout was; specified by the user
+from socket import _GLOBAL_DEFAULT_TIMEOUT, getdefaulttimeout
+
+from ..exceptions import TimeoutStateError
+
+# A sentinel value to indicate that no timeout was specified by the user in
+# urllib3
+_Default = object()
+
+
+# Use time.monotonic if available.
+current_time = getattr(time, "monotonic", time.time)
+
+
+class Timeout(object):
+    """Timeout configuration.
+
+    Timeouts can be defined as a default for a pool:
+
+    .. code-block:: python
+
+       timeout = Timeout(connect=2.0, read=7.0)
+       http = PoolManager(timeout=timeout)
+       response = http.request('GET', 'http://example.com/')
+
+    Or per-request (which overrides the default for the pool):
+
+    .. code-block:: python
+
+       response = http.request('GET', 'http://example.com/', timeout=Timeout(10))
+
+    Timeouts can be disabled by setting all the parameters to ``None``:
+
+    .. code-block:: python
+
+       no_timeout = Timeout(connect=None, read=None)
+       response = http.request('GET', 'http://example.com/, timeout=no_timeout)
+
+
+    :param total:
+        This combines the connect and read timeouts into one; the read timeout
+        will be set to the time leftover from the connect attempt. In the
+        event that both a connect timeout and a total are specified, or a read
+        timeout and a total are specified, the shorter timeout will be applied.
+
+        Defaults to None.
+
+    :type total: int, float, or None
+
+    :param connect:
+        The maximum amount of time (in seconds) to wait for a connection
+        attempt to a server to succeed. Omitting the parameter will default the
+        connect timeout to the system default, probably `the global default
+        timeout in socket.py
+        <http://hg.python.org/cpython/file/603b4d593758/Lib/socket.py#l535>`_.
+        None will set an infinite timeout for connection attempts.
+
+    :type connect: int, float, or None
+
+    :param read:
+        The maximum amount of time (in seconds) to wait between consecutive
+        read operations for a response from the server. Omitting the parameter
+        will default the read timeout to the system default, probably `the
+        global default timeout in socket.py
+        <http://hg.python.org/cpython/file/603b4d593758/Lib/socket.py#l535>`_.
+        None will set an infinite timeout.
+
+    :type read: int, float, or None
+
+    .. note::
+
+        Many factors can affect the total amount of time for urllib3 to return
+        an HTTP response.
+
+        For example, Python's DNS resolver does not obey the timeout specified
+        on the socket. Other factors that can affect total request time include
+        high CPU load, high swap, the program running at a low priority level,
+        or other behaviors.
+
+        In addition, the read and total timeouts only measure the time between
+        read operations on the socket connecting the client and the server,
+        not the total amount of time for the request to return a complete
+        response. For most requests, the timeout is raised because the server
+        has not sent the first byte in the specified time. This is not always
+        the case; if a server streams one byte every fifteen seconds, a timeout
+        of 20 seconds will not trigger, even though the request will take
+        several minutes to complete.
+
+        If your goal is to cut off any request after a set amount of wall clock
+        time, consider having a second "watcher" thread to cut off a slow
+        request.
+    """
+
+    #: A sentinel object representing the default timeout value
+    DEFAULT_TIMEOUT = _GLOBAL_DEFAULT_TIMEOUT
+
+    def __init__(self, total=None, connect=_Default, read=_Default):
+        self._connect = self._validate_timeout(connect, "connect")
+        self._read = self._validate_timeout(read, "read")
+        self.total = self._validate_timeout(total, "total")
+        self._start_connect = None
+
+    def __repr__(self):
+        return "%s(connect=%r, read=%r, total=%r)" % (
+            type(self).__name__,
+            self._connect,
+            self._read,
+            self.total,
+        )
+
+    # __str__ provided for backwards compatibility
+    __str__ = __repr__
+
+    @classmethod
+    def resolve_default_timeout(cls, timeout):
+        return getdefaulttimeout() if timeout is cls.DEFAULT_TIMEOUT else timeout
+
+    @classmethod
+    def _validate_timeout(cls, value, name):
+        """Check that a timeout attribute is valid.
+
+        :param value: The timeout value to validate
+        :param name: The name of the timeout attribute to validate. This is
+            used to specify in error messages.
+        :return: The validated and casted version of the given value.
+        :raises ValueError: If it is a numeric value less than or equal to
+            zero, or the type is not an integer, float, or None.
+        """
+        if value is _Default:
+            return cls.DEFAULT_TIMEOUT
+
+        if value is None or value is cls.DEFAULT_TIMEOUT:
+            return value
+
+        if isinstance(value, bool):
+            raise ValueError(
+                "Timeout cannot be a boolean value. It must "
+                "be an int, float or None."
+            )
+        try:
+            float(value)
+        except (TypeError, ValueError):
+            raise ValueError(
+                "Timeout value %s was %s, but it must be an "
+                "int, float or None." % (name, value)
+            )
+
+        try:
+            if value <= 0:
+                raise ValueError(
+                    "Attempted to set %s timeout to %s, but the "
+                    "timeout cannot be set to a value less "
+                    "than or equal to 0." % (name, value)
+                )
+        except TypeError:
+            # Python 3
+            raise ValueError(
+                "Timeout value %s was %s, but it must be an "
+                "int, float or None." % (name, value)
+            )
+
+        return value
+
+    @classmethod
+    def from_float(cls, timeout):
+        """Create a new Timeout from a legacy timeout value.
+
+        The timeout value used by httplib.py sets the same timeout on the
+        connect(), and recv() socket requests. This creates a :class:`Timeout`
+        object that sets the individual timeouts to the ``timeout`` value
+        passed to this function.
+
+        :param timeout: The legacy timeout value.
+        :type timeout: integer, float, sentinel default object, or None
+        :return: Timeout object
+        :rtype: :class:`Timeout`
+        """
+        return Timeout(read=timeout, connect=timeout)
+
+    def clone(self):
+        """Create a copy of the timeout object
+
+        Timeout properties are stored per-pool but each request needs a fresh
+        Timeout object to ensure each one has its own start/stop configured.
+
+        :return: a copy of the timeout object
+        :rtype: :class:`Timeout`
+        """
+        # We can't use copy.deepcopy because that will also create a new object
+        # for _GLOBAL_DEFAULT_TIMEOUT, which socket.py uses as a sentinel to
+        # detect the user default.
+        return Timeout(connect=self._connect, read=self._read, total=self.total)
+
+    def start_connect(self):
+        """Start the timeout clock, used during a connect() attempt
+
+        :raises urllib3.exceptions.TimeoutStateError: if you attempt
+            to start a timer that has been started already.
+        """
+        if self._start_connect is not None:
+            raise TimeoutStateError("Timeout timer has already been started.")
+        self._start_connect = current_time()
+        return self._start_connect
+
+    def get_connect_duration(self):
+        """Gets the time elapsed since the call to :meth:`start_connect`.
+
+        :return: Elapsed time in seconds.
+        :rtype: float
+        :raises urllib3.exceptions.TimeoutStateError: if you attempt
+            to get duration for a timer that hasn't been started.
+        """
+        if self._start_connect is None:
+            raise TimeoutStateError(
+                "Can't get connect duration for timer that has not started."
+            )
+        return current_time() - self._start_connect
+
+    @property
+    def connect_timeout(self):
+        """Get the value to use when setting a connection timeout.
+
+        This will be a positive float or integer, the value None
+        (never timeout), or the default system timeout.
+
+        :return: Connect timeout.
+        :rtype: int, float, :attr:`Timeout.DEFAULT_TIMEOUT` or None
+        """
+        if self.total is None:
+            return self._connect
+
+        if self._connect is None or self._connect is self.DEFAULT_TIMEOUT:
+            return self.total
+
+        return min(self._connect, self.total)
+
+    @property
+    def read_timeout(self):
+        """Get the value for the read timeout.
+
+        This assumes some time has elapsed in the connection timeout and
+        computes the read timeout appropriately.
+
+        If self.total is set, the read timeout is dependent on the amount of
+        time taken by the connect timeout. If the connection time has not been
+        established, a :exc:`~urllib3.exceptions.TimeoutStateError` will be
+        raised.
+
+        :return: Value to use for the read timeout.
+        :rtype: int, float, :attr:`Timeout.DEFAULT_TIMEOUT` or None
+        :raises urllib3.exceptions.TimeoutStateError: If :meth:`start_connect`
+            has not yet been called on this object.
+        """
+        if (
+            self.total is not None
+            and self.total is not self.DEFAULT_TIMEOUT
+            and self._read is not None
+            and self._read is not self.DEFAULT_TIMEOUT
+        ):
+            # In case the connect timeout has not yet been established.
+            if self._start_connect is None:
+                return self._read
+            return max(0, min(self.total - self.get_connect_duration(), self._read))
+        elif self.total is not None and self.total is not self.DEFAULT_TIMEOUT:
+            return max(0, self.total - self.get_connect_duration())
+        else:
+            return self._read
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/url.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/url.py
new file mode 100644
index 00000000..a960b2f3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/url.py
@@ -0,0 +1,435 @@
+from __future__ import absolute_import
+
+import re
+from collections import namedtuple
+
+from ..exceptions import LocationParseError
+from ..packages import six
+
+url_attrs = ["scheme", "auth", "host", "port", "path", "query", "fragment"]
+
+# We only want to normalize urls with an HTTP(S) scheme.
+# urllib3 infers URLs without a scheme (None) to be http.
+NORMALIZABLE_SCHEMES = ("http", "https", None)
+
+# Almost all of these patterns were derived from the
+# 'rfc3986' module: https://github.com/python-hyper/rfc3986
+PERCENT_RE = re.compile(r"%[a-fA-F0-9]{2}")
+SCHEME_RE = re.compile(r"^(?:[a-zA-Z][a-zA-Z0-9+-]*:|/)")
+URI_RE = re.compile(
+    r"^(?:([a-zA-Z][a-zA-Z0-9+.-]*):)?"
+    r"(?://([^\\/?#]*))?"
+    r"([^?#]*)"
+    r"(?:\?([^#]*))?"
+    r"(?:#(.*))?$",
+    re.UNICODE | re.DOTALL,
+)
+
+IPV4_PAT = r"(?:[0-9]{1,3}\.){3}[0-9]{1,3}"
+HEX_PAT = "[0-9A-Fa-f]{1,4}"
+LS32_PAT = "(?:{hex}:{hex}|{ipv4})".format(hex=HEX_PAT, ipv4=IPV4_PAT)
+_subs = {"hex": HEX_PAT, "ls32": LS32_PAT}
+_variations = [
+    #                            6( h16 ":" ) ls32
+    "(?:%(hex)s:){6}%(ls32)s",
+    #                       "::" 5( h16 ":" ) ls32
+    "::(?:%(hex)s:){5}%(ls32)s",
+    # [               h16 ] "::" 4( h16 ":" ) ls32
+    "(?:%(hex)s)?::(?:%(hex)s:){4}%(ls32)s",
+    # [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
+    "(?:(?:%(hex)s:)?%(hex)s)?::(?:%(hex)s:){3}%(ls32)s",
+    # [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
+    "(?:(?:%(hex)s:){0,2}%(hex)s)?::(?:%(hex)s:){2}%(ls32)s",
+    # [ *3( h16 ":" ) h16 ] "::"    h16 ":"   ls32
+    "(?:(?:%(hex)s:){0,3}%(hex)s)?::%(hex)s:%(ls32)s",
+    # [ *4( h16 ":" ) h16 ] "::"              ls32
+    "(?:(?:%(hex)s:){0,4}%(hex)s)?::%(ls32)s",
+    # [ *5( h16 ":" ) h16 ] "::"              h16
+    "(?:(?:%(hex)s:){0,5}%(hex)s)?::%(hex)s",
+    # [ *6( h16 ":" ) h16 ] "::"
+    "(?:(?:%(hex)s:){0,6}%(hex)s)?::",
+]
+
+UNRESERVED_PAT = r"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789._\-~"
+IPV6_PAT = "(?:" + "|".join([x % _subs for x in _variations]) + ")"
+ZONE_ID_PAT = "(?:%25|%)(?:[" + UNRESERVED_PAT + "]|%[a-fA-F0-9]{2})+"
+IPV6_ADDRZ_PAT = r"\[" + IPV6_PAT + r"(?:" + ZONE_ID_PAT + r")?\]"
+REG_NAME_PAT = r"(?:[^\[\]%:/?#]|%[a-fA-F0-9]{2})*"
+TARGET_RE = re.compile(r"^(/[^?#]*)(?:\?([^#]*))?(?:#.*)?$")
+
+IPV4_RE = re.compile("^" + IPV4_PAT + "$")
+IPV6_RE = re.compile("^" + IPV6_PAT + "$")
+IPV6_ADDRZ_RE = re.compile("^" + IPV6_ADDRZ_PAT + "$")
+BRACELESS_IPV6_ADDRZ_RE = re.compile("^" + IPV6_ADDRZ_PAT[2:-2] + "$")
+ZONE_ID_RE = re.compile("(" + ZONE_ID_PAT + r")\]$")
+
+_HOST_PORT_PAT = ("^(%s|%s|%s)(?::0*?(|0|[1-9][0-9]{0,4}))?$") % (
+    REG_NAME_PAT,
+    IPV4_PAT,
+    IPV6_ADDRZ_PAT,
+)
+_HOST_PORT_RE = re.compile(_HOST_PORT_PAT, re.UNICODE | re.DOTALL)
+
+UNRESERVED_CHARS = set(
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789._-~"
+)
+SUB_DELIM_CHARS = set("!$&'()*+,;=")
+USERINFO_CHARS = UNRESERVED_CHARS | SUB_DELIM_CHARS | {":"}
+PATH_CHARS = USERINFO_CHARS | {"@", "/"}
+QUERY_CHARS = FRAGMENT_CHARS = PATH_CHARS | {"?"}
+
+
+class Url(namedtuple("Url", url_attrs)):
+    """
+    Data structure for representing an HTTP URL. Used as a return value for
+    :func:`parse_url`. Both the scheme and host are normalized as they are
+    both case-insensitive according to RFC 3986.
+    """
+
+    __slots__ = ()
+
+    def __new__(
+        cls,
+        scheme=None,
+        auth=None,
+        host=None,
+        port=None,
+        path=None,
+        query=None,
+        fragment=None,
+    ):
+        if path and not path.startswith("/"):
+            path = "/" + path
+        if scheme is not None:
+            scheme = scheme.lower()
+        return super(Url, cls).__new__(
+            cls, scheme, auth, host, port, path, query, fragment
+        )
+
+    @property
+    def hostname(self):
+        """For backwards-compatibility with urlparse. We're nice like that."""
+        return self.host
+
+    @property
+    def request_uri(self):
+        """Absolute path including the query string."""
+        uri = self.path or "/"
+
+        if self.query is not None:
+            uri += "?" + self.query
+
+        return uri
+
+    @property
+    def netloc(self):
+        """Network location including host and port"""
+        if self.port:
+            return "%s:%d" % (self.host, self.port)
+        return self.host
+
+    @property
+    def url(self):
+        """
+        Convert self into a url
+
+        This function should more or less round-trip with :func:`.parse_url`. The
+        returned url may not be exactly the same as the url inputted to
+        :func:`.parse_url`, but it should be equivalent by the RFC (e.g., urls
+        with a blank port will have : removed).
+
+        Example: ::
+
+            >>> U = parse_url('http://google.com/mail/')
+            >>> U.url
+            'http://google.com/mail/'
+            >>> Url('http', 'username:password', 'host.com', 80,
+            ... '/path', 'query', 'fragment').url
+            'http://username:password@host.com:80/path?query#fragment'
+        """
+        scheme, auth, host, port, path, query, fragment = self
+        url = u""
+
+        # We use "is not None" we want things to happen with empty strings (or 0 port)
+        if scheme is not None:
+            url += scheme + u"://"
+        if auth is not None:
+            url += auth + u"@"
+        if host is not None:
+            url += host
+        if port is not None:
+            url += u":" + str(port)
+        if path is not None:
+            url += path
+        if query is not None:
+            url += u"?" + query
+        if fragment is not None:
+            url += u"#" + fragment
+
+        return url
+
+    def __str__(self):
+        return self.url
+
+
+def split_first(s, delims):
+    """
+    .. deprecated:: 1.25
+
+    Given a string and an iterable of delimiters, split on the first found
+    delimiter. Return two split parts and the matched delimiter.
+
+    If not found, then the first part is the full input string.
+
+    Example::
+
+        >>> split_first('foo/bar?baz', '?/=')
+        ('foo', 'bar?baz', '/')
+        >>> split_first('foo/bar?baz', '123')
+        ('foo/bar?baz', '', None)
+
+    Scales linearly with number of delims. Not ideal for large number of delims.
+    """
+    min_idx = None
+    min_delim = None
+    for d in delims:
+        idx = s.find(d)
+        if idx < 0:
+            continue
+
+        if min_idx is None or idx < min_idx:
+            min_idx = idx
+            min_delim = d
+
+    if min_idx is None or min_idx < 0:
+        return s, "", None
+
+    return s[:min_idx], s[min_idx + 1 :], min_delim
+
+
+def _encode_invalid_chars(component, allowed_chars, encoding="utf-8"):
+    """Percent-encodes a URI component without reapplying
+    onto an already percent-encoded component.
+    """
+    if component is None:
+        return component
+
+    component = six.ensure_text(component)
+
+    # Normalize existing percent-encoded bytes.
+    # Try to see if the component we're encoding is already percent-encoded
+    # so we can skip all '%' characters but still encode all others.
+    component, percent_encodings = PERCENT_RE.subn(
+        lambda match: match.group(0).upper(), component
+    )
+
+    uri_bytes = component.encode("utf-8", "surrogatepass")
+    is_percent_encoded = percent_encodings == uri_bytes.count(b"%")
+    encoded_component = bytearray()
+
+    for i in range(0, len(uri_bytes)):
+        # Will return a single character bytestring on both Python 2 & 3
+        byte = uri_bytes[i : i + 1]
+        byte_ord = ord(byte)
+        if (is_percent_encoded and byte == b"%") or (
+            byte_ord < 128 and byte.decode() in allowed_chars
+        ):
+            encoded_component += byte
+            continue
+        encoded_component.extend(b"%" + (hex(byte_ord)[2:].encode().zfill(2).upper()))
+
+    return encoded_component.decode(encoding)
+
+
+def _remove_path_dot_segments(path):
+    # See http://tools.ietf.org/html/rfc3986#section-5.2.4 for pseudo-code
+    segments = path.split("/")  # Turn the path into a list of segments
+    output = []  # Initialize the variable to use to store output
+
+    for segment in segments:
+        # '.' is the current directory, so ignore it, it is superfluous
+        if segment == ".":
+            continue
+        # Anything other than '..', should be appended to the output
+        elif segment != "..":
+            output.append(segment)
+        # In this case segment == '..', if we can, we should pop the last
+        # element
+        elif output:
+            output.pop()
+
+    # If the path starts with '/' and the output is empty or the first string
+    # is non-empty
+    if path.startswith("/") and (not output or output[0]):
+        output.insert(0, "")
+
+    # If the path starts with '/.' or '/..' ensure we add one more empty
+    # string to add a trailing '/'
+    if path.endswith(("/.", "/..")):
+        output.append("")
+
+    return "/".join(output)
+
+
+def _normalize_host(host, scheme):
+    if host:
+        if isinstance(host, six.binary_type):
+            host = six.ensure_str(host)
+
+        if scheme in NORMALIZABLE_SCHEMES:
+            is_ipv6 = IPV6_ADDRZ_RE.match(host)
+            if is_ipv6:
+                # IPv6 hosts of the form 'a::b%zone' are encoded in a URL as
+                # such per RFC 6874: 'a::b%25zone'. Unquote the ZoneID
+                # separator as necessary to return a valid RFC 4007 scoped IP.
+                match = ZONE_ID_RE.search(host)
+                if match:
+                    start, end = match.span(1)
+                    zone_id = host[start:end]
+
+                    if zone_id.startswith("%25") and zone_id != "%25":
+                        zone_id = zone_id[3:]
+                    else:
+                        zone_id = zone_id[1:]
+                    zone_id = "%" + _encode_invalid_chars(zone_id, UNRESERVED_CHARS)
+                    return host[:start].lower() + zone_id + host[end:]
+                else:
+                    return host.lower()
+            elif not IPV4_RE.match(host):
+                return six.ensure_str(
+                    b".".join([_idna_encode(label) for label in host.split(".")])
+                )
+    return host
+
+
+def _idna_encode(name):
+    if name and any(ord(x) >= 128 for x in name):
+        try:
+            from pip._vendor import idna
+        except ImportError:
+            six.raise_from(
+                LocationParseError("Unable to parse URL without the 'idna' module"),
+                None,
+            )
+        try:
+            return idna.encode(name.lower(), strict=True, std3_rules=True)
+        except idna.IDNAError:
+            six.raise_from(
+                LocationParseError(u"Name '%s' is not a valid IDNA label" % name), None
+            )
+    return name.lower().encode("ascii")
+
+
+def _encode_target(target):
+    """Percent-encodes a request target so that there are no invalid characters"""
+    path, query = TARGET_RE.match(target).groups()
+    target = _encode_invalid_chars(path, PATH_CHARS)
+    query = _encode_invalid_chars(query, QUERY_CHARS)
+    if query is not None:
+        target += "?" + query
+    return target
+
+
+def parse_url(url):
+    """
+    Given a url, return a parsed :class:`.Url` namedtuple. Best-effort is
+    performed to parse incomplete urls. Fields not provided will be None.
+    This parser is RFC 3986 and RFC 6874 compliant.
+
+    The parser logic and helper functions are based heavily on
+    work done in the ``rfc3986`` module.
+
+    :param str url: URL to parse into a :class:`.Url` namedtuple.
+
+    Partly backwards-compatible with :mod:`urlparse`.
+
+    Example::
+
+        >>> parse_url('http://google.com/mail/')
+        Url(scheme='http', host='google.com', port=None, path='/mail/', ...)
+        >>> parse_url('google.com:80')
+        Url(scheme=None, host='google.com', port=80, path=None, ...)
+        >>> parse_url('/foo?bar')
+        Url(scheme=None, host=None, port=None, path='/foo', query='bar', ...)
+    """
+    if not url:
+        # Empty
+        return Url()
+
+    source_url = url
+    if not SCHEME_RE.search(url):
+        url = "//" + url
+
+    try:
+        scheme, authority, path, query, fragment = URI_RE.match(url).groups()
+        normalize_uri = scheme is None or scheme.lower() in NORMALIZABLE_SCHEMES
+
+        if scheme:
+            scheme = scheme.lower()
+
+        if authority:
+            auth, _, host_port = authority.rpartition("@")
+            auth = auth or None
+            host, port = _HOST_PORT_RE.match(host_port).groups()
+            if auth and normalize_uri:
+                auth = _encode_invalid_chars(auth, USERINFO_CHARS)
+            if port == "":
+                port = None
+        else:
+            auth, host, port = None, None, None
+
+        if port is not None:
+            port = int(port)
+            if not (0 <= port <= 65535):
+                raise LocationParseError(url)
+
+        host = _normalize_host(host, scheme)
+
+        if normalize_uri and path:
+            path = _remove_path_dot_segments(path)
+            path = _encode_invalid_chars(path, PATH_CHARS)
+        if normalize_uri and query:
+            query = _encode_invalid_chars(query, QUERY_CHARS)
+        if normalize_uri and fragment:
+            fragment = _encode_invalid_chars(fragment, FRAGMENT_CHARS)
+
+    except (ValueError, AttributeError):
+        return six.raise_from(LocationParseError(source_url), None)
+
+    # For the sake of backwards compatibility we put empty
+    # string values for path if there are any defined values
+    # beyond the path in the URL.
+    # TODO: Remove this when we break backwards compatibility.
+    if not path:
+        if query is not None or fragment is not None:
+            path = ""
+        else:
+            path = None
+
+    # Ensure that each part of the URL is a `str` for
+    # backwards compatibility.
+    if isinstance(url, six.text_type):
+        ensure_func = six.ensure_text
+    else:
+        ensure_func = six.ensure_str
+
+    def ensure_type(x):
+        return x if x is None else ensure_func(x)
+
+    return Url(
+        scheme=ensure_type(scheme),
+        auth=ensure_type(auth),
+        host=ensure_type(host),
+        port=port,
+        path=ensure_type(path),
+        query=ensure_type(query),
+        fragment=ensure_type(fragment),
+    )
+
+
+def get_host(url):
+    """
+    Deprecated. Use :func:`parse_url` instead.
+    """
+    p = parse_url(url)
+    return p.scheme or "http", p.hostname, p.port
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/wait.py b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/wait.py
new file mode 100644
index 00000000..21b4590b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/urllib3/util/wait.py
@@ -0,0 +1,152 @@
+import errno
+import select
+import sys
+from functools import partial
+
+try:
+    from time import monotonic
+except ImportError:
+    from time import time as monotonic
+
+__all__ = ["NoWayToWaitForSocketError", "wait_for_read", "wait_for_write"]
+
+
+class NoWayToWaitForSocketError(Exception):
+    pass
+
+
+# How should we wait on sockets?
+#
+# There are two types of APIs you can use for waiting on sockets: the fancy
+# modern stateful APIs like epoll/kqueue, and the older stateless APIs like
+# select/poll. The stateful APIs are more efficient when you have a lots of
+# sockets to keep track of, because you can set them up once and then use them
+# lots of times. But we only ever want to wait on a single socket at a time
+# and don't want to keep track of state, so the stateless APIs are actually
+# more efficient. So we want to use select() or poll().
+#
+# Now, how do we choose between select() and poll()? On traditional Unixes,
+# select() has a strange calling convention that makes it slow, or fail
+# altogether, for high-numbered file descriptors. The point of poll() is to fix
+# that, so on Unixes, we prefer poll().
+#
+# On Windows, there is no poll() (or at least Python doesn't provide a wrapper
+# for it), but that's OK, because on Windows, select() doesn't have this
+# strange calling convention; plain select() works fine.
+#
+# So: on Windows we use select(), and everywhere else we use poll(). We also
+# fall back to select() in case poll() is somehow broken or missing.
+
+if sys.version_info >= (3, 5):
+    # Modern Python, that retries syscalls by default
+    def _retry_on_intr(fn, timeout):
+        return fn(timeout)
+
+else:
+    # Old and broken Pythons.
+    def _retry_on_intr(fn, timeout):
+        if timeout is None:
+            deadline = float("inf")
+        else:
+            deadline = monotonic() + timeout
+
+        while True:
+            try:
+                return fn(timeout)
+            # OSError for 3 <= pyver < 3.5, select.error for pyver <= 2.7
+            except (OSError, select.error) as e:
+                # 'e.args[0]' incantation works for both OSError and select.error
+                if e.args[0] != errno.EINTR:
+                    raise
+                else:
+                    timeout = deadline - monotonic()
+                    if timeout < 0:
+                        timeout = 0
+                    if timeout == float("inf"):
+                        timeout = None
+                    continue
+
+
+def select_wait_for_socket(sock, read=False, write=False, timeout=None):
+    if not read and not write:
+        raise RuntimeError("must specify at least one of read=True, write=True")
+    rcheck = []
+    wcheck = []
+    if read:
+        rcheck.append(sock)
+    if write:
+        wcheck.append(sock)
+    # When doing a non-blocking connect, most systems signal success by
+    # marking the socket writable. Windows, though, signals success by marked
+    # it as "exceptional". We paper over the difference by checking the write
+    # sockets for both conditions. (The stdlib selectors module does the same
+    # thing.)
+    fn = partial(select.select, rcheck, wcheck, wcheck)
+    rready, wready, xready = _retry_on_intr(fn, timeout)
+    return bool(rready or wready or xready)
+
+
+def poll_wait_for_socket(sock, read=False, write=False, timeout=None):
+    if not read and not write:
+        raise RuntimeError("must specify at least one of read=True, write=True")
+    mask = 0
+    if read:
+        mask |= select.POLLIN
+    if write:
+        mask |= select.POLLOUT
+    poll_obj = select.poll()
+    poll_obj.register(sock, mask)
+
+    # For some reason, poll() takes timeout in milliseconds
+    def do_poll(t):
+        if t is not None:
+            t *= 1000
+        return poll_obj.poll(t)
+
+    return bool(_retry_on_intr(do_poll, timeout))
+
+
+def null_wait_for_socket(*args, **kwargs):
+    raise NoWayToWaitForSocketError("no select-equivalent available")
+
+
+def _have_working_poll():
+    # Apparently some systems have a select.poll that fails as soon as you try
+    # to use it, either due to strange configuration or broken monkeypatching
+    # from libraries like eventlet/greenlet.
+    try:
+        poll_obj = select.poll()
+        _retry_on_intr(poll_obj.poll, 0)
+    except (AttributeError, OSError):
+        return False
+    else:
+        return True
+
+
+def wait_for_socket(*args, **kwargs):
+    # We delay choosing which implementation to use until the first time we're
+    # called. We could do it at import time, but then we might make the wrong
+    # decision if someone goes wild with monkeypatching select.poll after
+    # we're imported.
+    global wait_for_socket
+    if _have_working_poll():
+        wait_for_socket = poll_wait_for_socket
+    elif hasattr(select, "select"):
+        wait_for_socket = select_wait_for_socket
+    else:  # Platform-specific: Appengine.
+        wait_for_socket = null_wait_for_socket
+    return wait_for_socket(*args, **kwargs)
+
+
+def wait_for_read(sock, timeout=None):
+    """Waits for reading to be available on a given socket.
+    Returns True if the socket is readable, or False if the timeout expired.
+    """
+    return wait_for_socket(sock, read=True, timeout=timeout)
+
+
+def wait_for_write(sock, timeout=None):
+    """Waits for writing to be available on a given socket.
+    Returns True if the socket is readable, or False if the timeout expired.
+    """
+    return wait_for_socket(sock, write=True, timeout=timeout)
diff --git a/backend/venv/lib/python3.12/site-packages/pip/_vendor/vendor.txt b/backend/venv/lib/python3.12/site-packages/pip/_vendor/vendor.txt
new file mode 100644
index 00000000..f04a9c1e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/_vendor/vendor.txt
@@ -0,0 +1,18 @@
+CacheControl==0.14.1
+distlib==0.3.9
+distro==1.9.0
+msgpack==1.1.0
+packaging==24.2
+platformdirs==4.3.6
+pyproject-hooks==1.2.0
+requests==2.32.3
+    certifi==2024.8.30
+    idna==3.10
+    urllib3==1.26.20
+rich==13.9.4
+    pygments==2.18.0
+    typing_extensions==4.12.2
+resolvelib==1.0.1
+setuptools==70.3.0
+tomli==2.2.1
+truststore==0.10.0
diff --git a/backend/venv/lib/python3.12/site-packages/pip/py.typed b/backend/venv/lib/python3.12/site-packages/pip/py.typed
new file mode 100644
index 00000000..493b53e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pip/py.typed
@@ -0,0 +1,4 @@
+pip is a command line program. While it is implemented in Python, and so is
+available for import, you must not use pip's internal APIs in this way. Typing
+information is provided as a convenience only and is not a guarantee. Expect
+unannounced changes to the API and types in releases.
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/__init__.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/__init__.py
new file mode 100644
index 00000000..68feeb05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/__init__.py
@@ -0,0 +1,3709 @@
+"""
+Package resource API
+--------------------
+
+A resource is a logical file contained within a package, or a logical
+subdirectory thereof.  The package resource API expects resource names
+to have their path parts separated with ``/``, *not* whatever the local
+path separator is.  Do not use os.path operations to manipulate resource
+names being passed into the API.
+
+The package resource API is designed to work with normal filesystem packages,
+.egg files, and unpacked .egg files.  It can also work in a limited way with
+.zip files and with custom PEP 302 loaders that support the ``get_data()``
+method.
+
+This module is deprecated. Users are directed to :mod:`importlib.resources`,
+:mod:`importlib.metadata` and :pypi:`packaging` instead.
+"""
+
+from __future__ import annotations
+
+import sys
+
+if sys.version_info < (3, 9):  # noqa: UP036 # Check for unsupported versions
+    raise RuntimeError("Python 3.9 or later is required")
+
+import _imp
+import collections
+import email.parser
+import errno
+import functools
+import importlib
+import importlib.abc
+import importlib.machinery
+import inspect
+import io
+import ntpath
+import operator
+import os
+import pkgutil
+import platform
+import plistlib
+import posixpath
+import re
+import stat
+import tempfile
+import textwrap
+import time
+import types
+import warnings
+import zipfile
+import zipimport
+from collections.abc import Iterable, Iterator, Mapping, MutableSequence
+from pkgutil import get_importer
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    BinaryIO,
+    Callable,
+    Literal,
+    NamedTuple,
+    NoReturn,
+    Protocol,
+    TypeVar,
+    Union,
+    overload,
+)
+
+sys.path.extend(((vendor_path := os.path.join(os.path.dirname(os.path.dirname(__file__)), 'setuptools', '_vendor')) not in sys.path) * [vendor_path])  # fmt: skip
+# workaround for #4476
+sys.modules.pop('backports', None)
+
+# capture these to bypass sandboxing
+from os import open as os_open, utime  # isort: skip
+from os.path import isdir, split  # isort: skip
+
+try:
+    from os import mkdir, rename, unlink
+
+    WRITE_SUPPORT = True
+except ImportError:
+    # no write support, probably under GAE
+    WRITE_SUPPORT = False
+
+import packaging.markers
+import packaging.requirements
+import packaging.specifiers
+import packaging.utils
+import packaging.version
+from jaraco.text import drop_comment, join_continuation, yield_lines
+from platformdirs import user_cache_dir as _user_cache_dir
+
+if TYPE_CHECKING:
+    from _typeshed import BytesPath, StrOrBytesPath, StrPath
+    from _typeshed.importlib import LoaderProtocol
+    from typing_extensions import Self, TypeAlias
+
+warnings.warn(
+    "pkg_resources is deprecated as an API. "
+    "See https://setuptools.pypa.io/en/latest/pkg_resources.html",
+    DeprecationWarning,
+    stacklevel=2,
+)
+
+_T = TypeVar("_T")
+_DistributionT = TypeVar("_DistributionT", bound="Distribution")
+# Type aliases
+_NestedStr: TypeAlias = Union[str, Iterable[Union[str, Iterable["_NestedStr"]]]]
+_StrictInstallerType: TypeAlias = Callable[["Requirement"], "_DistributionT"]
+_InstallerType: TypeAlias = Callable[["Requirement"], Union["Distribution", None]]
+_PkgReqType: TypeAlias = Union[str, "Requirement"]
+_EPDistType: TypeAlias = Union["Distribution", _PkgReqType]
+_MetadataType: TypeAlias = Union["IResourceProvider", None]
+_ResolvedEntryPoint: TypeAlias = Any  # Can be any attribute in the module
+_ResourceStream: TypeAlias = Any  # TODO / Incomplete: A readable file-like object
+# Any object works, but let's indicate we expect something like a module (optionally has __loader__ or __file__)
+_ModuleLike: TypeAlias = Union[object, types.ModuleType]
+# Any: Should be _ModuleLike but we end up with issues where _ModuleLike doesn't have _ZipLoaderModule's __loader__
+_ProviderFactoryType: TypeAlias = Callable[[Any], "IResourceProvider"]
+_DistFinderType: TypeAlias = Callable[[_T, str, bool], Iterable["Distribution"]]
+_NSHandlerType: TypeAlias = Callable[[_T, str, str, types.ModuleType], Union[str, None]]
+_AdapterT = TypeVar(
+    "_AdapterT", _DistFinderType[Any], _ProviderFactoryType, _NSHandlerType[Any]
+)
+
+
+class _ZipLoaderModule(Protocol):
+    __loader__: zipimport.zipimporter
+
+
+_PEP440_FALLBACK = re.compile(r"^v?(?P<safe>(?:[0-9]+!)?[0-9]+(?:\.[0-9]+)*)", re.I)
+
+
+class PEP440Warning(RuntimeWarning):
+    """
+    Used when there is an issue with a version or specifier not complying with
+    PEP 440.
+    """
+
+
+parse_version = packaging.version.Version
+
+_state_vars: dict[str, str] = {}
+
+
+def _declare_state(vartype: str, varname: str, initial_value: _T) -> _T:
+    _state_vars[varname] = vartype
+    return initial_value
+
+
+def __getstate__() -> dict[str, Any]:
+    state = {}
+    g = globals()
+    for k, v in _state_vars.items():
+        state[k] = g['_sget_' + v](g[k])
+    return state
+
+
+def __setstate__(state: dict[str, Any]) -> dict[str, Any]:
+    g = globals()
+    for k, v in state.items():
+        g['_sset_' + _state_vars[k]](k, g[k], v)
+    return state
+
+
+def _sget_dict(val):
+    return val.copy()
+
+
+def _sset_dict(key, ob, state) -> None:
+    ob.clear()
+    ob.update(state)
+
+
+def _sget_object(val):
+    return val.__getstate__()
+
+
+def _sset_object(key, ob, state) -> None:
+    ob.__setstate__(state)
+
+
+_sget_none = _sset_none = lambda *args: None
+
+
+def get_supported_platform():
+    """Return this platform's maximum compatible version.
+
+    distutils.util.get_platform() normally reports the minimum version
+    of macOS that would be required to *use* extensions produced by
+    distutils.  But what we want when checking compatibility is to know the
+    version of macOS that we are *running*.  To allow usage of packages that
+    explicitly require a newer version of macOS, we must also know the
+    current version of the OS.
+
+    If this condition occurs for any other platform with a version in its
+    platform strings, this function should be extended accordingly.
+    """
+    plat = get_build_platform()
+    m = macosVersionString.match(plat)
+    if m is not None and sys.platform == "darwin":
+        try:
+            major_minor = '.'.join(_macos_vers()[:2])
+            build = m.group(3)
+            plat = f'macosx-{major_minor}-{build}'
+        except ValueError:
+            # not macOS
+            pass
+    return plat
+
+
+__all__ = [
+    # Basic resource access and distribution/entry point discovery
+    'require',
+    'run_script',
+    'get_provider',
+    'get_distribution',
+    'load_entry_point',
+    'get_entry_map',
+    'get_entry_info',
+    'iter_entry_points',
+    'resource_string',
+    'resource_stream',
+    'resource_filename',
+    'resource_listdir',
+    'resource_exists',
+    'resource_isdir',
+    # Environmental control
+    'declare_namespace',
+    'working_set',
+    'add_activation_listener',
+    'find_distributions',
+    'set_extraction_path',
+    'cleanup_resources',
+    'get_default_cache',
+    # Primary implementation classes
+    'Environment',
+    'WorkingSet',
+    'ResourceManager',
+    'Distribution',
+    'Requirement',
+    'EntryPoint',
+    # Exceptions
+    'ResolutionError',
+    'VersionConflict',
+    'DistributionNotFound',
+    'UnknownExtra',
+    'ExtractionError',
+    # Warnings
+    'PEP440Warning',
+    # Parsing functions and string utilities
+    'parse_requirements',
+    'parse_version',
+    'safe_name',
+    'safe_version',
+    'get_platform',
+    'compatible_platforms',
+    'yield_lines',
+    'split_sections',
+    'safe_extra',
+    'to_filename',
+    'invalid_marker',
+    'evaluate_marker',
+    # filesystem utilities
+    'ensure_directory',
+    'normalize_path',
+    # Distribution "precedence" constants
+    'EGG_DIST',
+    'BINARY_DIST',
+    'SOURCE_DIST',
+    'CHECKOUT_DIST',
+    'DEVELOP_DIST',
+    # "Provider" interfaces, implementations, and registration/lookup APIs
+    'IMetadataProvider',
+    'IResourceProvider',
+    'FileMetadata',
+    'PathMetadata',
+    'EggMetadata',
+    'EmptyProvider',
+    'empty_provider',
+    'NullProvider',
+    'EggProvider',
+    'DefaultProvider',
+    'ZipProvider',
+    'register_finder',
+    'register_namespace_handler',
+    'register_loader_type',
+    'fixup_namespace_packages',
+    'get_importer',
+    # Warnings
+    'PkgResourcesDeprecationWarning',
+    # Deprecated/backward compatibility only
+    'run_main',
+    'AvailableDistributions',
+]
+
+
+class ResolutionError(Exception):
+    """Abstract base for dependency resolution errors"""
+
+    def __repr__(self) -> str:
+        return self.__class__.__name__ + repr(self.args)
+
+
+class VersionConflict(ResolutionError):
+    """
+    An already-installed version conflicts with the requested version.
+
+    Should be initialized with the installed Distribution and the requested
+    Requirement.
+    """
+
+    _template = "{self.dist} is installed but {self.req} is required"
+
+    @property
+    def dist(self) -> Distribution:
+        return self.args[0]
+
+    @property
+    def req(self) -> Requirement:
+        return self.args[1]
+
+    def report(self):
+        return self._template.format(**locals())
+
+    def with_context(
+        self, required_by: set[Distribution | str]
+    ) -> Self | ContextualVersionConflict:
+        """
+        If required_by is non-empty, return a version of self that is a
+        ContextualVersionConflict.
+        """
+        if not required_by:
+            return self
+        args = self.args + (required_by,)
+        return ContextualVersionConflict(*args)
+
+
+class ContextualVersionConflict(VersionConflict):
+    """
+    A VersionConflict that accepts a third parameter, the set of the
+    requirements that required the installed Distribution.
+    """
+
+    _template = VersionConflict._template + ' by {self.required_by}'
+
+    @property
+    def required_by(self) -> set[str]:
+        return self.args[2]
+
+
+class DistributionNotFound(ResolutionError):
+    """A requested distribution was not found"""
+
+    _template = (
+        "The '{self.req}' distribution was not found "
+        "and is required by {self.requirers_str}"
+    )
+
+    @property
+    def req(self) -> Requirement:
+        return self.args[0]
+
+    @property
+    def requirers(self) -> set[str] | None:
+        return self.args[1]
+
+    @property
+    def requirers_str(self):
+        if not self.requirers:
+            return 'the application'
+        return ', '.join(self.requirers)
+
+    def report(self):
+        return self._template.format(**locals())
+
+    def __str__(self) -> str:
+        return self.report()
+
+
+class UnknownExtra(ResolutionError):
+    """Distribution doesn't have an "extra feature" of the given name"""
+
+
+_provider_factories: dict[type[_ModuleLike], _ProviderFactoryType] = {}
+
+PY_MAJOR = f'{sys.version_info.major}.{sys.version_info.minor}'
+EGG_DIST = 3
+BINARY_DIST = 2
+SOURCE_DIST = 1
+CHECKOUT_DIST = 0
+DEVELOP_DIST = -1
+
+
+def register_loader_type(
+    loader_type: type[_ModuleLike], provider_factory: _ProviderFactoryType
+) -> None:
+    """Register `provider_factory` to make providers for `loader_type`
+
+    `loader_type` is the type or class of a PEP 302 ``module.__loader__``,
+    and `provider_factory` is a function that, passed a *module* object,
+    returns an ``IResourceProvider`` for that module.
+    """
+    _provider_factories[loader_type] = provider_factory
+
+
+@overload
+def get_provider(moduleOrReq: str) -> IResourceProvider: ...
+@overload
+def get_provider(moduleOrReq: Requirement) -> Distribution: ...
+def get_provider(moduleOrReq: str | Requirement) -> IResourceProvider | Distribution:
+    """Return an IResourceProvider for the named module or requirement"""
+    if isinstance(moduleOrReq, Requirement):
+        return working_set.find(moduleOrReq) or require(str(moduleOrReq))[0]
+    try:
+        module = sys.modules[moduleOrReq]
+    except KeyError:
+        __import__(moduleOrReq)
+        module = sys.modules[moduleOrReq]
+    loader = getattr(module, '__loader__', None)
+    return _find_adapter(_provider_factories, loader)(module)
+
+
+@functools.cache
+def _macos_vers():
+    version = platform.mac_ver()[0]
+    # fallback for MacPorts
+    if version == '':
+        plist = '/System/Library/CoreServices/SystemVersion.plist'
+        if os.path.exists(plist):
+            with open(plist, 'rb') as fh:
+                plist_content = plistlib.load(fh)
+            if 'ProductVersion' in plist_content:
+                version = plist_content['ProductVersion']
+    return version.split('.')
+
+
+def _macos_arch(machine):
+    return {'PowerPC': 'ppc', 'Power_Macintosh': 'ppc'}.get(machine, machine)
+
+
+def get_build_platform():
+    """Return this platform's string for platform-specific distributions
+
+    XXX Currently this is the same as ``distutils.util.get_platform()``, but it
+    needs some hacks for Linux and macOS.
+    """
+    from sysconfig import get_platform
+
+    plat = get_platform()
+    if sys.platform == "darwin" and not plat.startswith('macosx-'):
+        try:
+            version = _macos_vers()
+            machine = _macos_arch(os.uname()[4].replace(" ", "_"))
+            return f"macosx-{version[0]}.{version[1]}-{machine}"
+        except ValueError:
+            # if someone is running a non-Mac darwin system, this will fall
+            # through to the default implementation
+            pass
+    return plat
+
+
+macosVersionString = re.compile(r"macosx-(\d+)\.(\d+)-(.*)")
+darwinVersionString = re.compile(r"darwin-(\d+)\.(\d+)\.(\d+)-(.*)")
+# XXX backward compat
+get_platform = get_build_platform
+
+
+def compatible_platforms(provided: str | None, required: str | None) -> bool:
+    """Can code for the `provided` platform run on the `required` platform?
+
+    Returns true if either platform is ``None``, or the platforms are equal.
+
+    XXX Needs compatibility checks for Linux and other unixy OSes.
+    """
+    if provided is None or required is None or provided == required:
+        # easy case
+        return True
+
+    # macOS special cases
+    reqMac = macosVersionString.match(required)
+    if reqMac:
+        provMac = macosVersionString.match(provided)
+
+        # is this a Mac package?
+        if not provMac:
+            # this is backwards compatibility for packages built before
+            # setuptools 0.6. All packages built after this point will
+            # use the new macOS designation.
+            provDarwin = darwinVersionString.match(provided)
+            if provDarwin:
+                dversion = int(provDarwin.group(1))
+                macosversion = f"{reqMac.group(1)}.{reqMac.group(2)}"
+                if (
+                    dversion == 7
+                    and macosversion >= "10.3"
+                    or dversion == 8
+                    and macosversion >= "10.4"
+                ):
+                    return True
+            # egg isn't macOS or legacy darwin
+            return False
+
+        # are they the same major version and machine type?
+        if provMac.group(1) != reqMac.group(1) or provMac.group(3) != reqMac.group(3):
+            return False
+
+        # is the required OS major update >= the provided one?
+        if int(provMac.group(2)) > int(reqMac.group(2)):
+            return False
+
+        return True
+
+    # XXX Linux and other platforms' special cases should go here
+    return False
+
+
+@overload
+def get_distribution(dist: _DistributionT) -> _DistributionT: ...
+@overload
+def get_distribution(dist: _PkgReqType) -> Distribution: ...
+def get_distribution(dist: Distribution | _PkgReqType) -> Distribution:
+    """Return a current distribution object for a Requirement or string"""
+    if isinstance(dist, str):
+        dist = Requirement.parse(dist)
+    if isinstance(dist, Requirement):
+        dist = get_provider(dist)
+    if not isinstance(dist, Distribution):
+        raise TypeError("Expected str, Requirement, or Distribution", dist)
+    return dist
+
+
+def load_entry_point(dist: _EPDistType, group: str, name: str) -> _ResolvedEntryPoint:
+    """Return `name` entry point of `group` for `dist` or raise ImportError"""
+    return get_distribution(dist).load_entry_point(group, name)
+
+
+@overload
+def get_entry_map(
+    dist: _EPDistType, group: None = None
+) -> dict[str, dict[str, EntryPoint]]: ...
+@overload
+def get_entry_map(dist: _EPDistType, group: str) -> dict[str, EntryPoint]: ...
+def get_entry_map(dist: _EPDistType, group: str | None = None):
+    """Return the entry point map for `group`, or the full entry map"""
+    return get_distribution(dist).get_entry_map(group)
+
+
+def get_entry_info(dist: _EPDistType, group: str, name: str) -> EntryPoint | None:
+    """Return the EntryPoint object for `group`+`name`, or ``None``"""
+    return get_distribution(dist).get_entry_info(group, name)
+
+
+class IMetadataProvider(Protocol):
+    def has_metadata(self, name: str) -> bool:
+        """Does the package's distribution contain the named metadata?"""
+        ...
+
+    def get_metadata(self, name: str) -> str:
+        """The named metadata resource as a string"""
+        ...
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        """Yield named metadata resource as list of non-blank non-comment lines
+
+        Leading and trailing whitespace is stripped from each line, and lines
+        with ``#`` as the first non-blank character are omitted."""
+        ...
+
+    def metadata_isdir(self, name: str) -> bool:
+        """Is the named metadata a directory?  (like ``os.path.isdir()``)"""
+        ...
+
+    def metadata_listdir(self, name: str) -> list[str]:
+        """List of metadata names in the directory (like ``os.listdir()``)"""
+        ...
+
+    def run_script(self, script_name: str, namespace: dict[str, Any]) -> None:
+        """Execute the named script in the supplied namespace dictionary"""
+        ...
+
+
+class IResourceProvider(IMetadataProvider, Protocol):
+    """An object that provides access to package resources"""
+
+    def get_resource_filename(
+        self, manager: ResourceManager, resource_name: str
+    ) -> str:
+        """Return a true filesystem path for `resource_name`
+
+        `manager` must be a ``ResourceManager``"""
+        ...
+
+    def get_resource_stream(
+        self, manager: ResourceManager, resource_name: str
+    ) -> _ResourceStream:
+        """Return a readable file-like object for `resource_name`
+
+        `manager` must be a ``ResourceManager``"""
+        ...
+
+    def get_resource_string(
+        self, manager: ResourceManager, resource_name: str
+    ) -> bytes:
+        """Return the contents of `resource_name` as :obj:`bytes`
+
+        `manager` must be a ``ResourceManager``"""
+        ...
+
+    def has_resource(self, resource_name: str) -> bool:
+        """Does the package contain the named resource?"""
+        ...
+
+    def resource_isdir(self, resource_name: str) -> bool:
+        """Is the named resource a directory?  (like ``os.path.isdir()``)"""
+        ...
+
+    def resource_listdir(self, resource_name: str) -> list[str]:
+        """List of resource names in the directory (like ``os.listdir()``)"""
+        ...
+
+
+class WorkingSet:
+    """A collection of active distributions on sys.path (or a similar list)"""
+
+    def __init__(self, entries: Iterable[str] | None = None) -> None:
+        """Create working set from list of path entries (default=sys.path)"""
+        self.entries: list[str] = []
+        self.entry_keys: dict[str | None, list[str]] = {}
+        self.by_key: dict[str, Distribution] = {}
+        self.normalized_to_canonical_keys: dict[str, str] = {}
+        self.callbacks: list[Callable[[Distribution], object]] = []
+
+        if entries is None:
+            entries = sys.path
+
+        for entry in entries:
+            self.add_entry(entry)
+
+    @classmethod
+    def _build_master(cls):
+        """
+        Prepare the master working set.
+        """
+        ws = cls()
+        try:
+            from __main__ import __requires__
+        except ImportError:
+            # The main program does not list any requirements
+            return ws
+
+        # ensure the requirements are met
+        try:
+            ws.require(__requires__)
+        except VersionConflict:
+            return cls._build_from_requirements(__requires__)
+
+        return ws
+
+    @classmethod
+    def _build_from_requirements(cls, req_spec):
+        """
+        Build a working set from a requirement spec. Rewrites sys.path.
+        """
+        # try it without defaults already on sys.path
+        # by starting with an empty path
+        ws = cls([])
+        reqs = parse_requirements(req_spec)
+        dists = ws.resolve(reqs, Environment())
+        for dist in dists:
+            ws.add(dist)
+
+        # add any missing entries from sys.path
+        for entry in sys.path:
+            if entry not in ws.entries:
+                ws.add_entry(entry)
+
+        # then copy back to sys.path
+        sys.path[:] = ws.entries
+        return ws
+
+    def add_entry(self, entry: str) -> None:
+        """Add a path item to ``.entries``, finding any distributions on it
+
+        ``find_distributions(entry, True)`` is used to find distributions
+        corresponding to the path entry, and they are added.  `entry` is
+        always appended to ``.entries``, even if it is already present.
+        (This is because ``sys.path`` can contain the same value more than
+        once, and the ``.entries`` of the ``sys.path`` WorkingSet should always
+        equal ``sys.path``.)
+        """
+        self.entry_keys.setdefault(entry, [])
+        self.entries.append(entry)
+        for dist in find_distributions(entry, True):
+            self.add(dist, entry, False)
+
+    def __contains__(self, dist: Distribution) -> bool:
+        """True if `dist` is the active distribution for its project"""
+        return self.by_key.get(dist.key) == dist
+
+    def find(self, req: Requirement) -> Distribution | None:
+        """Find a distribution matching requirement `req`
+
+        If there is an active distribution for the requested project, this
+        returns it as long as it meets the version requirement specified by
+        `req`.  But, if there is an active distribution for the project and it
+        does *not* meet the `req` requirement, ``VersionConflict`` is raised.
+        If there is no active distribution for the requested project, ``None``
+        is returned.
+        """
+        dist = self.by_key.get(req.key)
+
+        if dist is None:
+            canonical_key = self.normalized_to_canonical_keys.get(req.key)
+
+            if canonical_key is not None:
+                req.key = canonical_key
+                dist = self.by_key.get(canonical_key)
+
+        if dist is not None and dist not in req:
+            # XXX add more info
+            raise VersionConflict(dist, req)
+        return dist
+
+    def iter_entry_points(
+        self, group: str, name: str | None = None
+    ) -> Iterator[EntryPoint]:
+        """Yield entry point objects from `group` matching `name`
+
+        If `name` is None, yields all entry points in `group` from all
+        distributions in the working set, otherwise only ones matching
+        both `group` and `name` are yielded (in distribution order).
+        """
+        return (
+            entry
+            for dist in self
+            for entry in dist.get_entry_map(group).values()
+            if name is None or name == entry.name
+        )
+
+    def run_script(self, requires: str, script_name: str) -> None:
+        """Locate distribution for `requires` and run `script_name` script"""
+        ns = sys._getframe(1).f_globals
+        name = ns['__name__']
+        ns.clear()
+        ns['__name__'] = name
+        self.require(requires)[0].run_script(script_name, ns)
+
+    def __iter__(self) -> Iterator[Distribution]:
+        """Yield distributions for non-duplicate projects in the working set
+
+        The yield order is the order in which the items' path entries were
+        added to the working set.
+        """
+        seen = set()
+        for item in self.entries:
+            if item not in self.entry_keys:
+                # workaround a cache issue
+                continue
+
+            for key in self.entry_keys[item]:
+                if key not in seen:
+                    seen.add(key)
+                    yield self.by_key[key]
+
+    def add(
+        self,
+        dist: Distribution,
+        entry: str | None = None,
+        insert: bool = True,
+        replace: bool = False,
+    ) -> None:
+        """Add `dist` to working set, associated with `entry`
+
+        If `entry` is unspecified, it defaults to the ``.location`` of `dist`.
+        On exit from this routine, `entry` is added to the end of the working
+        set's ``.entries`` (if it wasn't already present).
+
+        `dist` is only added to the working set if it's for a project that
+        doesn't already have a distribution in the set, unless `replace=True`.
+        If it's added, any callbacks registered with the ``subscribe()`` method
+        will be called.
+        """
+        if insert:
+            dist.insert_on(self.entries, entry, replace=replace)
+
+        if entry is None:
+            entry = dist.location
+        keys = self.entry_keys.setdefault(entry, [])
+        keys2 = self.entry_keys.setdefault(dist.location, [])
+        if not replace and dist.key in self.by_key:
+            # ignore hidden distros
+            return
+
+        self.by_key[dist.key] = dist
+        normalized_name = packaging.utils.canonicalize_name(dist.key)
+        self.normalized_to_canonical_keys[normalized_name] = dist.key
+        if dist.key not in keys:
+            keys.append(dist.key)
+        if dist.key not in keys2:
+            keys2.append(dist.key)
+        self._added_new(dist)
+
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None,
+        installer: _StrictInstallerType[_DistributionT],
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[_DistributionT]: ...
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        *,
+        installer: _StrictInstallerType[_DistributionT],
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[_DistributionT]: ...
+    @overload
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[Distribution]: ...
+    def resolve(
+        self,
+        requirements: Iterable[Requirement],
+        env: Environment | None = None,
+        installer: _InstallerType | None | _StrictInstallerType[_DistributionT] = None,
+        replace_conflicting: bool = False,
+        extras: tuple[str, ...] | None = None,
+    ) -> list[Distribution] | list[_DistributionT]:
+        """List all distributions needed to (recursively) meet `requirements`
+
+        `requirements` must be a sequence of ``Requirement`` objects.  `env`,
+        if supplied, should be an ``Environment`` instance.  If
+        not supplied, it defaults to all distributions available within any
+        entry or distribution in the working set.  `installer`, if supplied,
+        will be invoked with each requirement that cannot be met by an
+        already-installed distribution; it should return a ``Distribution`` or
+        ``None``.
+
+        Unless `replace_conflicting=True`, raises a VersionConflict exception
+        if
+        any requirements are found on the path that have the correct name but
+        the wrong version.  Otherwise, if an `installer` is supplied it will be
+        invoked to obtain the correct version of the requirement and activate
+        it.
+
+        `extras` is a list of the extras to be used with these requirements.
+        This is important because extra requirements may look like `my_req;
+        extra = "my_extra"`, which would otherwise be interpreted as a purely
+        optional requirement.  Instead, we want to be able to assert that these
+        requirements are truly required.
+        """
+
+        # set up the stack
+        requirements = list(requirements)[::-1]
+        # set of processed requirements
+        processed = set()
+        # key -> dist
+        best: dict[str, Distribution] = {}
+        to_activate: list[Distribution] = []
+
+        req_extras = _ReqExtras()
+
+        # Mapping of requirement to set of distributions that required it;
+        # useful for reporting info about conflicts.
+        required_by = collections.defaultdict[Requirement, set[str]](set)
+
+        while requirements:
+            # process dependencies breadth-first
+            req = requirements.pop(0)
+            if req in processed:
+                # Ignore cyclic or redundant dependencies
+                continue
+
+            if not req_extras.markers_pass(req, extras):
+                continue
+
+            dist = self._resolve_dist(
+                req, best, replace_conflicting, env, installer, required_by, to_activate
+            )
+
+            # push the new requirements onto the stack
+            new_requirements = dist.requires(req.extras)[::-1]
+            requirements.extend(new_requirements)
+
+            # Register the new requirements needed by req
+            for new_requirement in new_requirements:
+                required_by[new_requirement].add(req.project_name)
+                req_extras[new_requirement] = req.extras
+
+            processed.add(req)
+
+        # return list of distros to activate
+        return to_activate
+
+    def _resolve_dist(
+        self, req, best, replace_conflicting, env, installer, required_by, to_activate
+    ) -> Distribution:
+        dist = best.get(req.key)
+        if dist is None:
+            # Find the best distribution and add it to the map
+            dist = self.by_key.get(req.key)
+            if dist is None or (dist not in req and replace_conflicting):
+                ws = self
+                if env is None:
+                    if dist is None:
+                        env = Environment(self.entries)
+                    else:
+                        # Use an empty environment and workingset to avoid
+                        # any further conflicts with the conflicting
+                        # distribution
+                        env = Environment([])
+                        ws = WorkingSet([])
+                dist = best[req.key] = env.best_match(
+                    req, ws, installer, replace_conflicting=replace_conflicting
+                )
+                if dist is None:
+                    requirers = required_by.get(req, None)
+                    raise DistributionNotFound(req, requirers)
+            to_activate.append(dist)
+        if dist not in req:
+            # Oops, the "best" so far conflicts with a dependency
+            dependent_req = required_by[req]
+            raise VersionConflict(dist, req).with_context(dependent_req)
+        return dist
+
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None,
+        installer: _StrictInstallerType[_DistributionT],
+        fallback: bool = True,
+    ) -> tuple[list[_DistributionT], dict[Distribution, Exception]]: ...
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        *,
+        installer: _StrictInstallerType[_DistributionT],
+        fallback: bool = True,
+    ) -> tuple[list[_DistributionT], dict[Distribution, Exception]]: ...
+    @overload
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        installer: _InstallerType | None = None,
+        fallback: bool = True,
+    ) -> tuple[list[Distribution], dict[Distribution, Exception]]: ...
+    def find_plugins(
+        self,
+        plugin_env: Environment,
+        full_env: Environment | None = None,
+        installer: _InstallerType | None | _StrictInstallerType[_DistributionT] = None,
+        fallback: bool = True,
+    ) -> tuple[
+        list[Distribution] | list[_DistributionT],
+        dict[Distribution, Exception],
+    ]:
+        """Find all activatable distributions in `plugin_env`
+
+        Example usage::
+
+            distributions, errors = working_set.find_plugins(
+                Environment(plugin_dirlist)
+            )
+            # add plugins+libs to sys.path
+            map(working_set.add, distributions)
+            # display errors
+            print('Could not load', errors)
+
+        The `plugin_env` should be an ``Environment`` instance that contains
+        only distributions that are in the project's "plugin directory" or
+        directories. The `full_env`, if supplied, should be an ``Environment``
+        contains all currently-available distributions.  If `full_env` is not
+        supplied, one is created automatically from the ``WorkingSet`` this
+        method is called on, which will typically mean that every directory on
+        ``sys.path`` will be scanned for distributions.
+
+        `installer` is a standard installer callback as used by the
+        ``resolve()`` method. The `fallback` flag indicates whether we should
+        attempt to resolve older versions of a plugin if the newest version
+        cannot be resolved.
+
+        This method returns a 2-tuple: (`distributions`, `error_info`), where
+        `distributions` is a list of the distributions found in `plugin_env`
+        that were loadable, along with any other distributions that are needed
+        to resolve their dependencies.  `error_info` is a dictionary mapping
+        unloadable plugin distributions to an exception instance describing the
+        error that occurred. Usually this will be a ``DistributionNotFound`` or
+        ``VersionConflict`` instance.
+        """
+
+        plugin_projects = list(plugin_env)
+        # scan project names in alphabetic order
+        plugin_projects.sort()
+
+        error_info: dict[Distribution, Exception] = {}
+        distributions: dict[Distribution, Exception | None] = {}
+
+        if full_env is None:
+            env = Environment(self.entries)
+            env += plugin_env
+        else:
+            env = full_env + plugin_env
+
+        shadow_set = self.__class__([])
+        # put all our entries in shadow_set
+        list(map(shadow_set.add, self))
+
+        for project_name in plugin_projects:
+            for dist in plugin_env[project_name]:
+                req = [dist.as_requirement()]
+
+                try:
+                    resolvees = shadow_set.resolve(req, env, installer)
+
+                except ResolutionError as v:
+                    # save error info
+                    error_info[dist] = v
+                    if fallback:
+                        # try the next older version of project
+                        continue
+                    else:
+                        # give up on this project, keep going
+                        break
+
+                else:
+                    list(map(shadow_set.add, resolvees))
+                    distributions.update(dict.fromkeys(resolvees))
+
+                    # success, no need to try any more versions of this project
+                    break
+
+        sorted_distributions = list(distributions)
+        sorted_distributions.sort()
+
+        return sorted_distributions, error_info
+
+    def require(self, *requirements: _NestedStr) -> list[Distribution]:
+        """Ensure that distributions matching `requirements` are activated
+
+        `requirements` must be a string or a (possibly-nested) sequence
+        thereof, specifying the distributions and versions required.  The
+        return value is a sequence of the distributions that needed to be
+        activated to fulfill the requirements; all relevant distributions are
+        included, even if they were already activated in this working set.
+        """
+        needed = self.resolve(parse_requirements(requirements))
+
+        for dist in needed:
+            self.add(dist)
+
+        return needed
+
+    def subscribe(
+        self, callback: Callable[[Distribution], object], existing: bool = True
+    ) -> None:
+        """Invoke `callback` for all distributions
+
+        If `existing=True` (default),
+        call on all existing ones, as well.
+        """
+        if callback in self.callbacks:
+            return
+        self.callbacks.append(callback)
+        if not existing:
+            return
+        for dist in self:
+            callback(dist)
+
+    def _added_new(self, dist) -> None:
+        for callback in self.callbacks:
+            callback(dist)
+
+    def __getstate__(
+        self,
+    ) -> tuple[
+        list[str],
+        dict[str | None, list[str]],
+        dict[str, Distribution],
+        dict[str, str],
+        list[Callable[[Distribution], object]],
+    ]:
+        return (
+            self.entries[:],
+            self.entry_keys.copy(),
+            self.by_key.copy(),
+            self.normalized_to_canonical_keys.copy(),
+            self.callbacks[:],
+        )
+
+    def __setstate__(self, e_k_b_n_c) -> None:
+        entries, keys, by_key, normalized_to_canonical_keys, callbacks = e_k_b_n_c
+        self.entries = entries[:]
+        self.entry_keys = keys.copy()
+        self.by_key = by_key.copy()
+        self.normalized_to_canonical_keys = normalized_to_canonical_keys.copy()
+        self.callbacks = callbacks[:]
+
+
+class _ReqExtras(dict["Requirement", tuple[str, ...]]):
+    """
+    Map each requirement to the extras that demanded it.
+    """
+
+    def markers_pass(self, req: Requirement, extras: tuple[str, ...] | None = None):
+        """
+        Evaluate markers for req against each extra that
+        demanded it.
+
+        Return False if the req has a marker and fails
+        evaluation. Otherwise, return True.
+        """
+        return not req.marker or any(
+            req.marker.evaluate({'extra': extra})
+            for extra in self.get(req, ()) + (extras or ("",))
+        )
+
+
+class Environment:
+    """Searchable snapshot of distributions on a search path"""
+
+    def __init__(
+        self,
+        search_path: Iterable[str] | None = None,
+        platform: str | None = get_supported_platform(),
+        python: str | None = PY_MAJOR,
+    ) -> None:
+        """Snapshot distributions available on a search path
+
+        Any distributions found on `search_path` are added to the environment.
+        `search_path` should be a sequence of ``sys.path`` items.  If not
+        supplied, ``sys.path`` is used.
+
+        `platform` is an optional string specifying the name of the platform
+        that platform-specific distributions must be compatible with.  If
+        unspecified, it defaults to the current platform.  `python` is an
+        optional string naming the desired version of Python (e.g. ``'3.6'``);
+        it defaults to the current version.
+
+        You may explicitly set `platform` (and/or `python`) to ``None`` if you
+        wish to map *all* distributions, not just those compatible with the
+        running platform or Python version.
+        """
+        self._distmap: dict[str, list[Distribution]] = {}
+        self.platform = platform
+        self.python = python
+        self.scan(search_path)
+
+    def can_add(self, dist: Distribution) -> bool:
+        """Is distribution `dist` acceptable for this environment?
+
+        The distribution must match the platform and python version
+        requirements specified when this environment was created, or False
+        is returned.
+        """
+        py_compat = (
+            self.python is None
+            or dist.py_version is None
+            or dist.py_version == self.python
+        )
+        return py_compat and compatible_platforms(dist.platform, self.platform)
+
+    def remove(self, dist: Distribution) -> None:
+        """Remove `dist` from the environment"""
+        self._distmap[dist.key].remove(dist)
+
+    def scan(self, search_path: Iterable[str] | None = None) -> None:
+        """Scan `search_path` for distributions usable in this environment
+
+        Any distributions found are added to the environment.
+        `search_path` should be a sequence of ``sys.path`` items.  If not
+        supplied, ``sys.path`` is used.  Only distributions conforming to
+        the platform/python version defined at initialization are added.
+        """
+        if search_path is None:
+            search_path = sys.path
+
+        for item in search_path:
+            for dist in find_distributions(item):
+                self.add(dist)
+
+    def __getitem__(self, project_name: str) -> list[Distribution]:
+        """Return a newest-to-oldest list of distributions for `project_name`
+
+        Uses case-insensitive `project_name` comparison, assuming all the
+        project's distributions use their project's name converted to all
+        lowercase as their key.
+
+        """
+        distribution_key = project_name.lower()
+        return self._distmap.get(distribution_key, [])
+
+    def add(self, dist: Distribution) -> None:
+        """Add `dist` if we ``can_add()`` it and it has not already been added"""
+        if self.can_add(dist) and dist.has_version():
+            dists = self._distmap.setdefault(dist.key, [])
+            if dist not in dists:
+                dists.append(dist)
+                dists.sort(key=operator.attrgetter('hashcmp'), reverse=True)
+
+    @overload
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _StrictInstallerType[_DistributionT],
+        replace_conflicting: bool = False,
+    ) -> _DistributionT: ...
+    @overload
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _InstallerType | None = None,
+        replace_conflicting: bool = False,
+    ) -> Distribution | None: ...
+    def best_match(
+        self,
+        req: Requirement,
+        working_set: WorkingSet,
+        installer: _InstallerType | None | _StrictInstallerType[_DistributionT] = None,
+        replace_conflicting: bool = False,
+    ) -> Distribution | None:
+        """Find distribution best matching `req` and usable on `working_set`
+
+        This calls the ``find(req)`` method of the `working_set` to see if a
+        suitable distribution is already active.  (This may raise
+        ``VersionConflict`` if an unsuitable version of the project is already
+        active in the specified `working_set`.)  If a suitable distribution
+        isn't active, this method returns the newest distribution in the
+        environment that meets the ``Requirement`` in `req`.  If no suitable
+        distribution is found, and `installer` is supplied, then the result of
+        calling the environment's ``obtain(req, installer)`` method will be
+        returned.
+        """
+        try:
+            dist = working_set.find(req)
+        except VersionConflict:
+            if not replace_conflicting:
+                raise
+            dist = None
+        if dist is not None:
+            return dist
+        for dist in self[req.key]:
+            if dist in req:
+                return dist
+        # try to download/install
+        return self.obtain(req, installer)
+
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: _StrictInstallerType[_DistributionT],
+    ) -> _DistributionT: ...
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: Callable[[Requirement], None] | None = None,
+    ) -> None: ...
+    @overload
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: _InstallerType | None = None,
+    ) -> Distribution | None: ...
+    def obtain(
+        self,
+        requirement: Requirement,
+        installer: Callable[[Requirement], None]
+        | _InstallerType
+        | None
+        | _StrictInstallerType[_DistributionT] = None,
+    ) -> Distribution | None:
+        """Obtain a distribution matching `requirement` (e.g. via download)
+
+        Obtain a distro that matches requirement (e.g. via download).  In the
+        base ``Environment`` class, this routine just returns
+        ``installer(requirement)``, unless `installer` is None, in which case
+        None is returned instead.  This method is a hook that allows subclasses
+        to attempt other ways of obtaining a distribution before falling back
+        to the `installer` argument."""
+        return installer(requirement) if installer else None
+
+    def __iter__(self) -> Iterator[str]:
+        """Yield the unique project names of the available distributions"""
+        for key in self._distmap.keys():
+            if self[key]:
+                yield key
+
+    def __iadd__(self, other: Distribution | Environment) -> Self:
+        """In-place addition of a distribution or environment"""
+        if isinstance(other, Distribution):
+            self.add(other)
+        elif isinstance(other, Environment):
+            for project in other:
+                for dist in other[project]:
+                    self.add(dist)
+        else:
+            raise TypeError(f"Can't add {other!r} to environment")
+        return self
+
+    def __add__(self, other: Distribution | Environment) -> Self:
+        """Add an environment or distribution to an environment"""
+        new = self.__class__([], platform=None, python=None)
+        for env in self, other:
+            new += env
+        return new
+
+
+# XXX backward compatibility
+AvailableDistributions = Environment
+
+
+class ExtractionError(RuntimeError):
+    """An error occurred extracting a resource
+
+    The following attributes are available from instances of this exception:
+
+    manager
+        The resource manager that raised this exception
+
+    cache_path
+        The base directory for resource extraction
+
+    original_error
+        The exception instance that caused extraction to fail
+    """
+
+    manager: ResourceManager
+    cache_path: str
+    original_error: BaseException | None
+
+
+class ResourceManager:
+    """Manage resource extraction and packages"""
+
+    extraction_path: str | None = None
+
+    def __init__(self) -> None:
+        # acts like a set
+        self.cached_files: dict[str, Literal[True]] = {}
+
+    def resource_exists(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> bool:
+        """Does the named resource exist?"""
+        return get_provider(package_or_requirement).has_resource(resource_name)
+
+    def resource_isdir(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> bool:
+        """Is the named resource an existing directory?"""
+        return get_provider(package_or_requirement).resource_isdir(resource_name)
+
+    def resource_filename(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> str:
+        """Return a true filesystem path for specified resource"""
+        return get_provider(package_or_requirement).get_resource_filename(
+            self, resource_name
+        )
+
+    def resource_stream(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> _ResourceStream:
+        """Return a readable file-like object for specified resource"""
+        return get_provider(package_or_requirement).get_resource_stream(
+            self, resource_name
+        )
+
+    def resource_string(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> bytes:
+        """Return specified resource as :obj:`bytes`"""
+        return get_provider(package_or_requirement).get_resource_string(
+            self, resource_name
+        )
+
+    def resource_listdir(
+        self, package_or_requirement: _PkgReqType, resource_name: str
+    ) -> list[str]:
+        """List the contents of the named resource directory"""
+        return get_provider(package_or_requirement).resource_listdir(resource_name)
+
+    def extraction_error(self) -> NoReturn:
+        """Give an error message for problems extracting file(s)"""
+
+        old_exc = sys.exc_info()[1]
+        cache_path = self.extraction_path or get_default_cache()
+
+        tmpl = textwrap.dedent(
+            """
+            Can't extract file(s) to egg cache
+
+            The following error occurred while trying to extract file(s)
+            to the Python egg cache:
+
+              {old_exc}
+
+            The Python egg cache directory is currently set to:
+
+              {cache_path}
+
+            Perhaps your account does not have write access to this directory?
+            You can change the cache directory by setting the PYTHON_EGG_CACHE
+            environment variable to point to an accessible directory.
+            """
+        ).lstrip()
+        err = ExtractionError(tmpl.format(**locals()))
+        err.manager = self
+        err.cache_path = cache_path
+        err.original_error = old_exc
+        raise err
+
+    def get_cache_path(self, archive_name: str, names: Iterable[StrPath] = ()) -> str:
+        """Return absolute location in cache for `archive_name` and `names`
+
+        The parent directory of the resulting path will be created if it does
+        not already exist.  `archive_name` should be the base filename of the
+        enclosing egg (which may not be the name of the enclosing zipfile!),
+        including its ".egg" extension.  `names`, if provided, should be a
+        sequence of path name parts "under" the egg's extraction location.
+
+        This method should only be called by resource providers that need to
+        obtain an extraction location, and only for names they intend to
+        extract, as it tracks the generated names for possible cleanup later.
+        """
+        extract_path = self.extraction_path or get_default_cache()
+        target_path = os.path.join(extract_path, archive_name + '-tmp', *names)
+        try:
+            _bypass_ensure_directory(target_path)
+        except Exception:
+            self.extraction_error()
+
+        self._warn_unsafe_extraction_path(extract_path)
+
+        self.cached_files[target_path] = True
+        return target_path
+
+    @staticmethod
+    def _warn_unsafe_extraction_path(path) -> None:
+        """
+        If the default extraction path is overridden and set to an insecure
+        location, such as /tmp, it opens up an opportunity for an attacker to
+        replace an extracted file with an unauthorized payload. Warn the user
+        if a known insecure location is used.
+
+        See Distribute #375 for more details.
+        """
+        if os.name == 'nt' and not path.startswith(os.environ['windir']):
+            # On Windows, permissions are generally restrictive by default
+            #  and temp directories are not writable by other users, so
+            #  bypass the warning.
+            return
+        mode = os.stat(path).st_mode
+        if mode & stat.S_IWOTH or mode & stat.S_IWGRP:
+            msg = (
+                "Extraction path is writable by group/others "
+                "and vulnerable to attack when "
+                "used with get_resource_filename ({path}). "
+                "Consider a more secure "
+                "location (set with .set_extraction_path or the "
+                "PYTHON_EGG_CACHE environment variable)."
+            ).format(**locals())
+            warnings.warn(msg, UserWarning)
+
+    def postprocess(self, tempname: StrOrBytesPath, filename: StrOrBytesPath) -> None:
+        """Perform any platform-specific postprocessing of `tempname`
+
+        This is where Mac header rewrites should be done; other platforms don't
+        have anything special they should do.
+
+        Resource providers should call this method ONLY after successfully
+        extracting a compressed resource.  They must NOT call it on resources
+        that are already in the filesystem.
+
+        `tempname` is the current (temporary) name of the file, and `filename`
+        is the name it will be renamed to by the caller after this routine
+        returns.
+        """
+
+        if os.name == 'posix':
+            # Make the resource executable
+            mode = ((os.stat(tempname).st_mode) | 0o555) & 0o7777
+            os.chmod(tempname, mode)
+
+    def set_extraction_path(self, path: str) -> None:
+        """Set the base path where resources will be extracted to, if needed.
+
+        If you do not call this routine before any extractions take place, the
+        path defaults to the return value of ``get_default_cache()``.  (Which
+        is based on the ``PYTHON_EGG_CACHE`` environment variable, with various
+        platform-specific fallbacks.  See that routine's documentation for more
+        details.)
+
+        Resources are extracted to subdirectories of this path based upon
+        information given by the ``IResourceProvider``.  You may set this to a
+        temporary directory, but then you must call ``cleanup_resources()`` to
+        delete the extracted files when done.  There is no guarantee that
+        ``cleanup_resources()`` will be able to remove all extracted files.
+
+        (Note: you may not change the extraction path for a given resource
+        manager once resources have been extracted, unless you first call
+        ``cleanup_resources()``.)
+        """
+        if self.cached_files:
+            raise ValueError("Can't change extraction path, files already extracted")
+
+        self.extraction_path = path
+
+    def cleanup_resources(self, force: bool = False) -> list[str]:
+        """
+        Delete all extracted resource files and directories, returning a list
+        of the file and directory names that could not be successfully removed.
+        This function does not have any concurrency protection, so it should
+        generally only be called when the extraction path is a temporary
+        directory exclusive to a single process.  This method is not
+        automatically called; you must call it explicitly or register it as an
+        ``atexit`` function if you wish to ensure cleanup of a temporary
+        directory used for extractions.
+        """
+        # XXX
+        return []
+
+
+def get_default_cache() -> str:
+    """
+    Return the ``PYTHON_EGG_CACHE`` environment variable
+    or a platform-relevant user cache dir for an app
+    named "Python-Eggs".
+    """
+    return os.environ.get('PYTHON_EGG_CACHE') or _user_cache_dir(appname='Python-Eggs')
+
+
+def safe_name(name: str) -> str:
+    """Convert an arbitrary string to a standard distribution name
+
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub('[^A-Za-z0-9.]+', '-', name)
+
+
+def safe_version(version: str) -> str:
+    """
+    Convert an arbitrary string to a standard version string
+    """
+    try:
+        # normalize the version
+        return str(packaging.version.Version(version))
+    except packaging.version.InvalidVersion:
+        version = version.replace(' ', '.')
+        return re.sub('[^A-Za-z0-9.]+', '-', version)
+
+
+def _forgiving_version(version) -> str:
+    """Fallback when ``safe_version`` is not safe enough
+    >>> parse_version(_forgiving_version('0.23ubuntu1'))
+    <Version('0.23.dev0+sanitized.ubuntu1')>
+    >>> parse_version(_forgiving_version('0.23-'))
+    <Version('0.23.dev0+sanitized')>
+    >>> parse_version(_forgiving_version('0.-_'))
+    <Version('0.dev0+sanitized')>
+    >>> parse_version(_forgiving_version('42.+?1'))
+    <Version('42.dev0+sanitized.1')>
+    >>> parse_version(_forgiving_version('hello world'))
+    <Version('0.dev0+sanitized.hello.world')>
+    """
+    version = version.replace(' ', '.')
+    match = _PEP440_FALLBACK.search(version)
+    if match:
+        safe = match["safe"]
+        rest = version[len(safe) :]
+    else:
+        safe = "0"
+        rest = version
+    local = f"sanitized.{_safe_segment(rest)}".strip(".")
+    return f"{safe}.dev0+{local}"
+
+
+def _safe_segment(segment):
+    """Convert an arbitrary string into a safe segment"""
+    segment = re.sub('[^A-Za-z0-9.]+', '-', segment)
+    segment = re.sub('-[^A-Za-z0-9]+', '-', segment)
+    return re.sub(r'\.[^A-Za-z0-9]+', '.', segment).strip(".-")
+
+
+def safe_extra(extra: str) -> str:
+    """Convert an arbitrary string to a standard 'extra' name
+
+    Any runs of non-alphanumeric characters are replaced with a single '_',
+    and the result is always lowercased.
+    """
+    return re.sub('[^A-Za-z0-9.-]+', '_', extra).lower()
+
+
+def to_filename(name: str) -> str:
+    """Convert a project or version name to its filename-escaped form
+
+    Any '-' characters are currently replaced with '_'.
+    """
+    return name.replace('-', '_')
+
+
+def invalid_marker(text: str) -> SyntaxError | Literal[False]:
+    """
+    Validate text as a PEP 508 environment marker; return an exception
+    if invalid or False otherwise.
+    """
+    try:
+        evaluate_marker(text)
+    except SyntaxError as e:
+        e.filename = None
+        e.lineno = None
+        return e
+    return False
+
+
+def evaluate_marker(text: str, extra: str | None = None) -> bool:
+    """
+    Evaluate a PEP 508 environment marker.
+    Return a boolean indicating the marker result in this environment.
+    Raise SyntaxError if marker is invalid.
+
+    This implementation uses the 'pyparsing' module.
+    """
+    try:
+        marker = packaging.markers.Marker(text)
+        return marker.evaluate()
+    except packaging.markers.InvalidMarker as e:
+        raise SyntaxError(e) from e
+
+
+class NullProvider:
+    """Try to implement resources and metadata for arbitrary PEP 302 loaders"""
+
+    egg_name: str | None = None
+    egg_info: str | None = None
+    loader: LoaderProtocol | None = None
+
+    def __init__(self, module: _ModuleLike) -> None:
+        self.loader = getattr(module, '__loader__', None)
+        self.module_path = os.path.dirname(getattr(module, '__file__', ''))
+
+    def get_resource_filename(
+        self, manager: ResourceManager, resource_name: str
+    ) -> str:
+        return self._fn(self.module_path, resource_name)
+
+    def get_resource_stream(
+        self, manager: ResourceManager, resource_name: str
+    ) -> BinaryIO:
+        return io.BytesIO(self.get_resource_string(manager, resource_name))
+
+    def get_resource_string(
+        self, manager: ResourceManager, resource_name: str
+    ) -> bytes:
+        return self._get(self._fn(self.module_path, resource_name))
+
+    def has_resource(self, resource_name: str) -> bool:
+        return self._has(self._fn(self.module_path, resource_name))
+
+    def _get_metadata_path(self, name):
+        return self._fn(self.egg_info, name)
+
+    def has_metadata(self, name: str) -> bool:
+        if not self.egg_info:
+            return False
+
+        path = self._get_metadata_path(name)
+        return self._has(path)
+
+    def get_metadata(self, name: str) -> str:
+        if not self.egg_info:
+            return ""
+        path = self._get_metadata_path(name)
+        value = self._get(path)
+        try:
+            return value.decode('utf-8')
+        except UnicodeDecodeError as exc:
+            # Include the path in the error message to simplify
+            # troubleshooting, and without changing the exception type.
+            exc.reason += f' in {name} file at path: {path}'
+            raise
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        return yield_lines(self.get_metadata(name))
+
+    def resource_isdir(self, resource_name: str) -> bool:
+        return self._isdir(self._fn(self.module_path, resource_name))
+
+    def metadata_isdir(self, name: str) -> bool:
+        return bool(self.egg_info and self._isdir(self._fn(self.egg_info, name)))
+
+    def resource_listdir(self, resource_name: str) -> list[str]:
+        return self._listdir(self._fn(self.module_path, resource_name))
+
+    def metadata_listdir(self, name: str) -> list[str]:
+        if self.egg_info:
+            return self._listdir(self._fn(self.egg_info, name))
+        return []
+
+    def run_script(self, script_name: str, namespace: dict[str, Any]) -> None:
+        script = 'scripts/' + script_name
+        if not self.has_metadata(script):
+            raise ResolutionError(
+                "Script {script!r} not found in metadata at {self.egg_info!r}".format(
+                    **locals()
+                ),
+            )
+
+        script_text = self.get_metadata(script).replace('\r\n', '\n')
+        script_text = script_text.replace('\r', '\n')
+        script_filename = self._fn(self.egg_info, script)
+        namespace['__file__'] = script_filename
+        if os.path.exists(script_filename):
+            source = _read_utf8_with_fallback(script_filename)
+            code = compile(source, script_filename, 'exec')
+            exec(code, namespace, namespace)
+        else:
+            from linecache import cache
+
+            cache[script_filename] = (
+                len(script_text),
+                0,
+                script_text.split('\n'),
+                script_filename,
+            )
+            script_code = compile(script_text, script_filename, 'exec')
+            exec(script_code, namespace, namespace)
+
+    def _has(self, path) -> bool:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _isdir(self, path) -> bool:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _listdir(self, path) -> list[str]:
+        raise NotImplementedError(
+            "Can't perform this operation for unregistered loader type"
+        )
+
+    def _fn(self, base: str | None, resource_name: str):
+        if base is None:
+            raise TypeError(
+                "`base` parameter in `_fn` is `None`. Either override this method or check the parameter first."
+            )
+        self._validate_resource_path(resource_name)
+        if resource_name:
+            return os.path.join(base, *resource_name.split('/'))
+        return base
+
+    @staticmethod
+    def _validate_resource_path(path) -> None:
+        """
+        Validate the resource paths according to the docs.
+        https://setuptools.pypa.io/en/latest/pkg_resources.html#basic-resource-access
+
+        >>> warned = getfixture('recwarn')
+        >>> warnings.simplefilter('always')
+        >>> vrp = NullProvider._validate_resource_path
+        >>> vrp('foo/bar.txt')
+        >>> bool(warned)
+        False
+        >>> vrp('../foo/bar.txt')
+        >>> bool(warned)
+        True
+        >>> warned.clear()
+        >>> vrp('/foo/bar.txt')
+        >>> bool(warned)
+        True
+        >>> vrp('foo/../../bar.txt')
+        >>> bool(warned)
+        True
+        >>> warned.clear()
+        >>> vrp('foo/f../bar.txt')
+        >>> bool(warned)
+        False
+
+        Windows path separators are straight-up disallowed.
+        >>> vrp(r'\\foo/bar.txt')
+        Traceback (most recent call last):
+        ...
+        ValueError: Use of .. or absolute path in a resource path \
+is not allowed.
+
+        >>> vrp(r'C:\\foo/bar.txt')
+        Traceback (most recent call last):
+        ...
+        ValueError: Use of .. or absolute path in a resource path \
+is not allowed.
+
+        Blank values are allowed
+
+        >>> vrp('')
+        >>> bool(warned)
+        False
+
+        Non-string values are not.
+
+        >>> vrp(None)
+        Traceback (most recent call last):
+        ...
+        AttributeError: ...
+        """
+        invalid = (
+            os.path.pardir in path.split(posixpath.sep)
+            or posixpath.isabs(path)
+            or ntpath.isabs(path)
+            or path.startswith("\\")
+        )
+        if not invalid:
+            return
+
+        msg = "Use of .. or absolute path in a resource path is not allowed."
+
+        # Aggressively disallow Windows absolute paths
+        if (path.startswith("\\") or ntpath.isabs(path)) and not posixpath.isabs(path):
+            raise ValueError(msg)
+
+        # for compatibility, warn; in future
+        # raise ValueError(msg)
+        issue_warning(
+            msg[:-1] + " and will raise exceptions in a future release.",
+            DeprecationWarning,
+        )
+
+    def _get(self, path) -> bytes:
+        if hasattr(self.loader, 'get_data') and self.loader:
+            # Already checked get_data exists
+            return self.loader.get_data(path)  # type: ignore[attr-defined]
+        raise NotImplementedError(
+            "Can't perform this operation for loaders without 'get_data()'"
+        )
+
+
+register_loader_type(object, NullProvider)
+
+
+def _parents(path):
+    """
+    yield all parents of path including path
+    """
+    last = None
+    while path != last:
+        yield path
+        last = path
+        path, _ = os.path.split(path)
+
+
+class EggProvider(NullProvider):
+    """Provider based on a virtual filesystem"""
+
+    def __init__(self, module: _ModuleLike) -> None:
+        super().__init__(module)
+        self._setup_prefix()
+
+    def _setup_prefix(self):
+        # Assume that metadata may be nested inside a "basket"
+        # of multiple eggs and use module_path instead of .archive.
+        eggs = filter(_is_egg_path, _parents(self.module_path))
+        egg = next(eggs, None)
+        egg and self._set_egg(egg)
+
+    def _set_egg(self, path: str) -> None:
+        self.egg_name = os.path.basename(path)
+        self.egg_info = os.path.join(path, 'EGG-INFO')
+        self.egg_root = path
+
+
+class DefaultProvider(EggProvider):
+    """Provides access to package resources in the filesystem"""
+
+    def _has(self, path) -> bool:
+        return os.path.exists(path)
+
+    def _isdir(self, path) -> bool:
+        return os.path.isdir(path)
+
+    def _listdir(self, path):
+        return os.listdir(path)
+
+    def get_resource_stream(
+        self, manager: object, resource_name: str
+    ) -> io.BufferedReader:
+        return open(self._fn(self.module_path, resource_name), 'rb')
+
+    def _get(self, path) -> bytes:
+        with open(path, 'rb') as stream:
+            return stream.read()
+
+    @classmethod
+    def _register(cls) -> None:
+        loader_names = (
+            'SourceFileLoader',
+            'SourcelessFileLoader',
+        )
+        for name in loader_names:
+            loader_cls = getattr(importlib.machinery, name, type(None))
+            register_loader_type(loader_cls, cls)
+
+
+DefaultProvider._register()
+
+
+class EmptyProvider(NullProvider):
+    """Provider that returns nothing for all requests"""
+
+    # A special case, we don't want all Providers inheriting from NullProvider to have a potentially None module_path
+    module_path: str | None = None  # type: ignore[assignment]
+
+    _isdir = _has = lambda self, path: False
+
+    def _get(self, path) -> bytes:
+        return b''
+
+    def _listdir(self, path):
+        return []
+
+    def __init__(self) -> None:
+        pass
+
+
+empty_provider = EmptyProvider()
+
+
+class ZipManifests(dict[str, "MemoizedZipManifests.manifest_mod"]):
+    """
+    zip manifest builder
+    """
+
+    # `path` could be `StrPath | IO[bytes]` but that violates the LSP for `MemoizedZipManifests.load`
+    @classmethod
+    def build(cls, path: str) -> dict[str, zipfile.ZipInfo]:
+        """
+        Build a dictionary similar to the zipimport directory
+        caches, except instead of tuples, store ZipInfo objects.
+
+        Use a platform-specific path separator (os.sep) for the path keys
+        for compatibility with pypy on Windows.
+        """
+        with zipfile.ZipFile(path) as zfile:
+            items = (
+                (
+                    name.replace('/', os.sep),
+                    zfile.getinfo(name),
+                )
+                for name in zfile.namelist()
+            )
+            return dict(items)
+
+    load = build
+
+
+class MemoizedZipManifests(ZipManifests):
+    """
+    Memoized zipfile manifests.
+    """
+
+    class manifest_mod(NamedTuple):
+        manifest: dict[str, zipfile.ZipInfo]
+        mtime: float
+
+    def load(self, path: str) -> dict[str, zipfile.ZipInfo]:  # type: ignore[override] # ZipManifests.load is a classmethod
+        """
+        Load a manifest at path or return a suitable manifest already loaded.
+        """
+        path = os.path.normpath(path)
+        mtime = os.stat(path).st_mtime
+
+        if path not in self or self[path].mtime != mtime:
+            manifest = self.build(path)
+            self[path] = self.manifest_mod(manifest, mtime)
+
+        return self[path].manifest
+
+
+class ZipProvider(EggProvider):
+    """Resource support for zips and eggs"""
+
+    eagers: list[str] | None = None
+    _zip_manifests = MemoizedZipManifests()
+    # ZipProvider's loader should always be a zipimporter or equivalent
+    loader: zipimport.zipimporter
+
+    def __init__(self, module: _ZipLoaderModule) -> None:
+        super().__init__(module)
+        self.zip_pre = self.loader.archive + os.sep
+
+    def _zipinfo_name(self, fspath):
+        # Convert a virtual filename (full path to file) into a zipfile subpath
+        # usable with the zipimport directory cache for our target archive
+        fspath = fspath.rstrip(os.sep)
+        if fspath == self.loader.archive:
+            return ''
+        if fspath.startswith(self.zip_pre):
+            return fspath[len(self.zip_pre) :]
+        raise AssertionError(f"{fspath} is not a subpath of {self.zip_pre}")
+
+    def _parts(self, zip_path):
+        # Convert a zipfile subpath into an egg-relative path part list.
+        # pseudo-fs path
+        fspath = self.zip_pre + zip_path
+        if fspath.startswith(self.egg_root + os.sep):
+            return fspath[len(self.egg_root) + 1 :].split(os.sep)
+        raise AssertionError(f"{fspath} is not a subpath of {self.egg_root}")
+
+    @property
+    def zipinfo(self):
+        return self._zip_manifests.load(self.loader.archive)
+
+    def get_resource_filename(
+        self, manager: ResourceManager, resource_name: str
+    ) -> str:
+        if not self.egg_name:
+            raise NotImplementedError(
+                "resource_filename() only supported for .egg, not .zip"
+            )
+        # no need to lock for extraction, since we use temp names
+        zip_path = self._resource_to_zip(resource_name)
+        eagers = self._get_eager_resources()
+        if '/'.join(self._parts(zip_path)) in eagers:
+            for name in eagers:
+                self._extract_resource(manager, self._eager_to_zip(name))
+        return self._extract_resource(manager, zip_path)
+
+    @staticmethod
+    def _get_date_and_size(zip_stat):
+        size = zip_stat.file_size
+        # ymdhms+wday, yday, dst
+        date_time = zip_stat.date_time + (0, 0, -1)
+        # 1980 offset already done
+        timestamp = time.mktime(date_time)
+        return timestamp, size
+
+    # FIXME: 'ZipProvider._extract_resource' is too complex (12)
+    def _extract_resource(self, manager: ResourceManager, zip_path) -> str:  # noqa: C901
+        if zip_path in self._index():
+            for name in self._index()[zip_path]:
+                last = self._extract_resource(manager, os.path.join(zip_path, name))
+            # return the extracted directory name
+            return os.path.dirname(last)
+
+        timestamp, _size = self._get_date_and_size(self.zipinfo[zip_path])
+
+        if not WRITE_SUPPORT:
+            raise OSError(
+                '"os.rename" and "os.unlink" are not supported on this platform'
+            )
+        try:
+            if not self.egg_name:
+                raise OSError(
+                    '"egg_name" is empty. This likely means no egg could be found from the "module_path".'
+                )
+            real_path = manager.get_cache_path(self.egg_name, self._parts(zip_path))
+
+            if self._is_current(real_path, zip_path):
+                return real_path
+
+            outf, tmpnam = _mkstemp(
+                ".$extract",
+                dir=os.path.dirname(real_path),
+            )
+            os.write(outf, self.loader.get_data(zip_path))
+            os.close(outf)
+            utime(tmpnam, (timestamp, timestamp))
+            manager.postprocess(tmpnam, real_path)
+
+            try:
+                rename(tmpnam, real_path)
+
+            except OSError:
+                if os.path.isfile(real_path):
+                    if self._is_current(real_path, zip_path):
+                        # the file became current since it was checked above,
+                        #  so proceed.
+                        return real_path
+                    # Windows, del old file and retry
+                    elif os.name == 'nt':
+                        unlink(real_path)
+                        rename(tmpnam, real_path)
+                        return real_path
+                raise
+
+        except OSError:
+            # report a user-friendly error
+            manager.extraction_error()
+
+        return real_path
+
+    def _is_current(self, file_path, zip_path):
+        """
+        Return True if the file_path is current for this zip_path
+        """
+        timestamp, size = self._get_date_and_size(self.zipinfo[zip_path])
+        if not os.path.isfile(file_path):
+            return False
+        stat = os.stat(file_path)
+        if stat.st_size != size or stat.st_mtime != timestamp:
+            return False
+        # check that the contents match
+        zip_contents = self.loader.get_data(zip_path)
+        with open(file_path, 'rb') as f:
+            file_contents = f.read()
+        return zip_contents == file_contents
+
+    def _get_eager_resources(self):
+        if self.eagers is None:
+            eagers = []
+            for name in ('native_libs.txt', 'eager_resources.txt'):
+                if self.has_metadata(name):
+                    eagers.extend(self.get_metadata_lines(name))
+            self.eagers = eagers
+        return self.eagers
+
+    def _index(self):
+        try:
+            return self._dirindex
+        except AttributeError:
+            ind = {}
+            for path in self.zipinfo:
+                parts = path.split(os.sep)
+                while parts:
+                    parent = os.sep.join(parts[:-1])
+                    if parent in ind:
+                        ind[parent].append(parts[-1])
+                        break
+                    else:
+                        ind[parent] = [parts.pop()]
+            self._dirindex = ind
+            return ind
+
+    def _has(self, fspath) -> bool:
+        zip_path = self._zipinfo_name(fspath)
+        return zip_path in self.zipinfo or zip_path in self._index()
+
+    def _isdir(self, fspath) -> bool:
+        return self._zipinfo_name(fspath) in self._index()
+
+    def _listdir(self, fspath):
+        return list(self._index().get(self._zipinfo_name(fspath), ()))
+
+    def _eager_to_zip(self, resource_name: str):
+        return self._zipinfo_name(self._fn(self.egg_root, resource_name))
+
+    def _resource_to_zip(self, resource_name: str):
+        return self._zipinfo_name(self._fn(self.module_path, resource_name))
+
+
+register_loader_type(zipimport.zipimporter, ZipProvider)
+
+
+class FileMetadata(EmptyProvider):
+    """Metadata handler for standalone PKG-INFO files
+
+    Usage::
+
+        metadata = FileMetadata("/path/to/PKG-INFO")
+
+    This provider rejects all data and metadata requests except for PKG-INFO,
+    which is treated as existing, and will be the contents of the file at
+    the provided location.
+    """
+
+    def __init__(self, path: StrPath) -> None:
+        self.path = path
+
+    def _get_metadata_path(self, name):
+        return self.path
+
+    def has_metadata(self, name: str) -> bool:
+        return name == 'PKG-INFO' and os.path.isfile(self.path)
+
+    def get_metadata(self, name: str) -> str:
+        if name != 'PKG-INFO':
+            raise KeyError("No metadata except PKG-INFO is available")
+
+        with open(self.path, encoding='utf-8', errors="replace") as f:
+            metadata = f.read()
+        self._warn_on_replacement(metadata)
+        return metadata
+
+    def _warn_on_replacement(self, metadata) -> None:
+        replacement_char = '�'
+        if replacement_char in metadata:
+            tmpl = "{self.path} could not be properly decoded in UTF-8"
+            msg = tmpl.format(**locals())
+            warnings.warn(msg)
+
+    def get_metadata_lines(self, name: str) -> Iterator[str]:
+        return yield_lines(self.get_metadata(name))
+
+
+class PathMetadata(DefaultProvider):
+    """Metadata provider for egg directories
+
+    Usage::
+
+        # Development eggs:
+
+        egg_info = "/path/to/PackageName.egg-info"
+        base_dir = os.path.dirname(egg_info)
+        metadata = PathMetadata(base_dir, egg_info)
+        dist_name = os.path.splitext(os.path.basename(egg_info))[0]
+        dist = Distribution(basedir, project_name=dist_name, metadata=metadata)
+
+        # Unpacked egg directories:
+
+        egg_path = "/path/to/PackageName-ver-pyver-etc.egg"
+        metadata = PathMetadata(egg_path, os.path.join(egg_path,'EGG-INFO'))
+        dist = Distribution.from_filename(egg_path, metadata=metadata)
+    """
+
+    def __init__(self, path: str, egg_info: str) -> None:
+        self.module_path = path
+        self.egg_info = egg_info
+
+
+class EggMetadata(ZipProvider):
+    """Metadata provider for .egg files"""
+
+    def __init__(self, importer: zipimport.zipimporter) -> None:
+        """Create a metadata provider from a zipimporter"""
+
+        self.zip_pre = importer.archive + os.sep
+        self.loader = importer
+        if importer.prefix:
+            self.module_path = os.path.join(importer.archive, importer.prefix)
+        else:
+            self.module_path = importer.archive
+        self._setup_prefix()
+
+
+_distribution_finders: dict[type, _DistFinderType[Any]] = _declare_state(
+    'dict', '_distribution_finders', {}
+)
+
+
+def register_finder(
+    importer_type: type[_T], distribution_finder: _DistFinderType[_T]
+) -> None:
+    """Register `distribution_finder` to find distributions in sys.path items
+
+    `importer_type` is the type or class of a PEP 302 "Importer" (sys.path item
+    handler), and `distribution_finder` is a callable that, passed a path
+    item and the importer instance, yields ``Distribution`` instances found on
+    that path item.  See ``pkg_resources.find_on_path`` for an example."""
+    _distribution_finders[importer_type] = distribution_finder
+
+
+def find_distributions(path_item: str, only: bool = False) -> Iterable[Distribution]:
+    """Yield distributions accessible via `path_item`"""
+    importer = get_importer(path_item)
+    finder = _find_adapter(_distribution_finders, importer)
+    return finder(importer, path_item, only)
+
+
+def find_eggs_in_zip(
+    importer: zipimport.zipimporter, path_item: str, only: bool = False
+) -> Iterator[Distribution]:
+    """
+    Find eggs in zip files; possibly multiple nested eggs.
+    """
+    if importer.archive.endswith('.whl'):
+        # wheels are not supported with this finder
+        # they don't have PKG-INFO metadata, and won't ever contain eggs
+        return
+    metadata = EggMetadata(importer)
+    if metadata.has_metadata('PKG-INFO'):
+        yield Distribution.from_filename(path_item, metadata=metadata)
+    if only:
+        # don't yield nested distros
+        return
+    for subitem in metadata.resource_listdir(''):
+        if _is_egg_path(subitem):
+            subpath = os.path.join(path_item, subitem)
+            dists = find_eggs_in_zip(zipimport.zipimporter(subpath), subpath)
+            yield from dists
+        elif subitem.lower().endswith(('.dist-info', '.egg-info')):
+            subpath = os.path.join(path_item, subitem)
+            submeta = EggMetadata(zipimport.zipimporter(subpath))
+            submeta.egg_info = subpath
+            yield Distribution.from_location(path_item, subitem, submeta)
+
+
+register_finder(zipimport.zipimporter, find_eggs_in_zip)
+
+
+def find_nothing(
+    importer: object | None, path_item: str | None, only: bool | None = False
+):
+    return ()
+
+
+register_finder(object, find_nothing)
+
+
+def find_on_path(importer: object | None, path_item, only=False):
+    """Yield distributions accessible on a sys.path directory"""
+    path_item = _normalize_cached(path_item)
+
+    if _is_unpacked_egg(path_item):
+        yield Distribution.from_filename(
+            path_item,
+            metadata=PathMetadata(path_item, os.path.join(path_item, 'EGG-INFO')),
+        )
+        return
+
+    entries = (os.path.join(path_item, child) for child in safe_listdir(path_item))
+
+    # scan for .egg and .egg-info in directory
+    for entry in sorted(entries):
+        fullpath = os.path.join(path_item, entry)
+        factory = dist_factory(path_item, entry, only)
+        yield from factory(fullpath)
+
+
+def dist_factory(path_item, entry, only):
+    """Return a dist_factory for the given entry."""
+    lower = entry.lower()
+    is_egg_info = lower.endswith('.egg-info')
+    is_dist_info = lower.endswith('.dist-info') and os.path.isdir(
+        os.path.join(path_item, entry)
+    )
+    is_meta = is_egg_info or is_dist_info
+    return (
+        distributions_from_metadata
+        if is_meta
+        else find_distributions
+        if not only and _is_egg_path(entry)
+        else resolve_egg_link
+        if not only and lower.endswith('.egg-link')
+        else NoDists()
+    )
+
+
+class NoDists:
+    """
+    >>> bool(NoDists())
+    False
+
+    >>> list(NoDists()('anything'))
+    []
+    """
+
+    def __bool__(self) -> Literal[False]:
+        return False
+
+    def __call__(self, fullpath: object):
+        return iter(())
+
+
+def safe_listdir(path: StrOrBytesPath):
+    """
+    Attempt to list contents of path, but suppress some exceptions.
+    """
+    try:
+        return os.listdir(path)
+    except (PermissionError, NotADirectoryError):
+        pass
+    except OSError as e:
+        # Ignore the directory if does not exist, not a directory or
+        # permission denied
+        if e.errno not in (errno.ENOTDIR, errno.EACCES, errno.ENOENT):
+            raise
+    return ()
+
+
+def distributions_from_metadata(path: str):
+    root = os.path.dirname(path)
+    if os.path.isdir(path):
+        if len(os.listdir(path)) == 0:
+            # empty metadata dir; skip
+            return
+        metadata: _MetadataType = PathMetadata(root, path)
+    else:
+        metadata = FileMetadata(path)
+    entry = os.path.basename(path)
+    yield Distribution.from_location(
+        root,
+        entry,
+        metadata,
+        precedence=DEVELOP_DIST,
+    )
+
+
+def non_empty_lines(path):
+    """
+    Yield non-empty lines from file at path
+    """
+    for line in _read_utf8_with_fallback(path).splitlines():
+        line = line.strip()
+        if line:
+            yield line
+
+
+def resolve_egg_link(path):
+    """
+    Given a path to an .egg-link, resolve distributions
+    present in the referenced path.
+    """
+    referenced_paths = non_empty_lines(path)
+    resolved_paths = (
+        os.path.join(os.path.dirname(path), ref) for ref in referenced_paths
+    )
+    dist_groups = map(find_distributions, resolved_paths)
+    return next(dist_groups, ())
+
+
+if hasattr(pkgutil, 'ImpImporter'):
+    register_finder(pkgutil.ImpImporter, find_on_path)
+
+register_finder(importlib.machinery.FileFinder, find_on_path)
+
+_namespace_handlers: dict[type, _NSHandlerType[Any]] = _declare_state(
+    'dict', '_namespace_handlers', {}
+)
+_namespace_packages: dict[str | None, list[str]] = _declare_state(
+    'dict', '_namespace_packages', {}
+)
+
+
+def register_namespace_handler(
+    importer_type: type[_T], namespace_handler: _NSHandlerType[_T]
+) -> None:
+    """Register `namespace_handler` to declare namespace packages
+
+    `importer_type` is the type or class of a PEP 302 "Importer" (sys.path item
+    handler), and `namespace_handler` is a callable like this::
+
+        def namespace_handler(importer, path_entry, moduleName, module):
+            # return a path_entry to use for child packages
+
+    Namespace handlers are only called if the importer object has already
+    agreed that it can handle the relevant path item, and they should only
+    return a subpath if the module __path__ does not already contain an
+    equivalent subpath.  For an example namespace handler, see
+    ``pkg_resources.file_ns_handler``.
+    """
+    _namespace_handlers[importer_type] = namespace_handler
+
+
+def _handle_ns(packageName, path_item):
+    """Ensure that named package includes a subpath of path_item (if needed)"""
+
+    importer = get_importer(path_item)
+    if importer is None:
+        return None
+
+    # use find_spec (PEP 451) and fall-back to find_module (PEP 302)
+    try:
+        spec = importer.find_spec(packageName)
+    except AttributeError:
+        # capture warnings due to #1111
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            loader = importer.find_module(packageName)
+    else:
+        loader = spec.loader if spec else None
+
+    if loader is None:
+        return None
+    module = sys.modules.get(packageName)
+    if module is None:
+        module = sys.modules[packageName] = types.ModuleType(packageName)
+        module.__path__ = []
+        _set_parent_ns(packageName)
+    elif not hasattr(module, '__path__'):
+        raise TypeError("Not a package:", packageName)
+    handler = _find_adapter(_namespace_handlers, importer)
+    subpath = handler(importer, path_item, packageName, module)
+    if subpath is not None:
+        path = module.__path__
+        path.append(subpath)
+        importlib.import_module(packageName)
+        _rebuild_mod_path(path, packageName, module)
+    return subpath
+
+
+def _rebuild_mod_path(orig_path, package_name, module: types.ModuleType) -> None:
+    """
+    Rebuild module.__path__ ensuring that all entries are ordered
+    corresponding to their sys.path order
+    """
+    sys_path = [_normalize_cached(p) for p in sys.path]
+
+    def safe_sys_path_index(entry):
+        """
+        Workaround for #520 and #513.
+        """
+        try:
+            return sys_path.index(entry)
+        except ValueError:
+            return float('inf')
+
+    def position_in_sys_path(path):
+        """
+        Return the ordinal of the path based on its position in sys.path
+        """
+        path_parts = path.split(os.sep)
+        module_parts = package_name.count('.') + 1
+        parts = path_parts[:-module_parts]
+        return safe_sys_path_index(_normalize_cached(os.sep.join(parts)))
+
+    new_path = sorted(orig_path, key=position_in_sys_path)
+    new_path = [_normalize_cached(p) for p in new_path]
+
+    if isinstance(module.__path__, list):
+        module.__path__[:] = new_path
+    else:
+        module.__path__ = new_path
+
+
+def declare_namespace(packageName: str) -> None:
+    """Declare that package 'packageName' is a namespace package"""
+
+    msg = (
+        f"Deprecated call to `pkg_resources.declare_namespace({packageName!r})`.\n"
+        "Implementing implicit namespace packages (as specified in PEP 420) "
+        "is preferred to `pkg_resources.declare_namespace`. "
+        "See https://setuptools.pypa.io/en/latest/references/"
+        "keywords.html#keyword-namespace-packages"
+    )
+    warnings.warn(msg, DeprecationWarning, stacklevel=2)
+
+    _imp.acquire_lock()
+    try:
+        if packageName in _namespace_packages:
+            return
+
+        path: MutableSequence[str] = sys.path
+        parent, _, _ = packageName.rpartition('.')
+
+        if parent:
+            declare_namespace(parent)
+            if parent not in _namespace_packages:
+                __import__(parent)
+            try:
+                path = sys.modules[parent].__path__
+            except AttributeError as e:
+                raise TypeError("Not a package:", parent) from e
+
+        # Track what packages are namespaces, so when new path items are added,
+        # they can be updated
+        _namespace_packages.setdefault(parent or None, []).append(packageName)
+        _namespace_packages.setdefault(packageName, [])
+
+        for path_item in path:
+            # Ensure all the parent's path items are reflected in the child,
+            # if they apply
+            _handle_ns(packageName, path_item)
+
+    finally:
+        _imp.release_lock()
+
+
+def fixup_namespace_packages(path_item: str, parent: str | None = None) -> None:
+    """Ensure that previously-declared namespace packages include path_item"""
+    _imp.acquire_lock()
+    try:
+        for package in _namespace_packages.get(parent, ()):
+            subpath = _handle_ns(package, path_item)
+            if subpath:
+                fixup_namespace_packages(subpath, package)
+    finally:
+        _imp.release_lock()
+
+
+def file_ns_handler(
+    importer: object,
+    path_item: StrPath,
+    packageName: str,
+    module: types.ModuleType,
+):
+    """Compute an ns-package subpath for a filesystem or zipfile importer"""
+
+    subpath = os.path.join(path_item, packageName.split('.')[-1])
+    normalized = _normalize_cached(subpath)
+    for item in module.__path__:
+        if _normalize_cached(item) == normalized:
+            break
+    else:
+        # Only return the path if it's not already there
+        return subpath
+
+
+if hasattr(pkgutil, 'ImpImporter'):
+    register_namespace_handler(pkgutil.ImpImporter, file_ns_handler)
+
+register_namespace_handler(zipimport.zipimporter, file_ns_handler)
+register_namespace_handler(importlib.machinery.FileFinder, file_ns_handler)
+
+
+def null_ns_handler(
+    importer: object,
+    path_item: str | None,
+    packageName: str | None,
+    module: _ModuleLike | None,
+) -> None:
+    return None
+
+
+register_namespace_handler(object, null_ns_handler)
+
+
+@overload
+def normalize_path(filename: StrPath) -> str: ...
+@overload
+def normalize_path(filename: BytesPath) -> bytes: ...
+def normalize_path(filename: StrOrBytesPath) -> str | bytes:
+    """Normalize a file/dir name for comparison purposes"""
+    return os.path.normcase(os.path.realpath(os.path.normpath(_cygwin_patch(filename))))
+
+
+def _cygwin_patch(filename: StrOrBytesPath):  # pragma: nocover
+    """
+    Contrary to POSIX 2008, on Cygwin, getcwd (3) contains
+    symlink components. Using
+    os.path.abspath() works around this limitation. A fix in os.getcwd()
+    would probably better, in Cygwin even more so, except
+    that this seems to be by design...
+    """
+    return os.path.abspath(filename) if sys.platform == 'cygwin' else filename
+
+
+if TYPE_CHECKING:
+    # https://github.com/python/mypy/issues/16261
+    # https://github.com/python/typeshed/issues/6347
+    @overload
+    def _normalize_cached(filename: StrPath) -> str: ...
+    @overload
+    def _normalize_cached(filename: BytesPath) -> bytes: ...
+    def _normalize_cached(filename: StrOrBytesPath) -> str | bytes: ...
+
+else:
+
+    @functools.cache
+    def _normalize_cached(filename):
+        return normalize_path(filename)
+
+
+def _is_egg_path(path):
+    """
+    Determine if given path appears to be an egg.
+    """
+    return _is_zip_egg(path) or _is_unpacked_egg(path)
+
+
+def _is_zip_egg(path):
+    return (
+        path.lower().endswith('.egg')
+        and os.path.isfile(path)
+        and zipfile.is_zipfile(path)
+    )
+
+
+def _is_unpacked_egg(path):
+    """
+    Determine if given path appears to be an unpacked egg.
+    """
+    return path.lower().endswith('.egg') and os.path.isfile(
+        os.path.join(path, 'EGG-INFO', 'PKG-INFO')
+    )
+
+
+def _set_parent_ns(packageName) -> None:
+    parts = packageName.split('.')
+    name = parts.pop()
+    if parts:
+        parent = '.'.join(parts)
+        setattr(sys.modules[parent], name, sys.modules[packageName])
+
+
+MODULE = re.compile(r"\w+(\.\w+)*$").match
+EGG_NAME = re.compile(
+    r"""
+    (?P<name>[^-]+) (
+        -(?P<ver>[^-]+) (
+            -py(?P<pyver>[^-]+) (
+                -(?P<plat>.+)
+            )?
+        )?
+    )?
+    """,
+    re.VERBOSE | re.IGNORECASE,
+).match
+
+
+class EntryPoint:
+    """Object representing an advertised importable object"""
+
+    def __init__(
+        self,
+        name: str,
+        module_name: str,
+        attrs: Iterable[str] = (),
+        extras: Iterable[str] = (),
+        dist: Distribution | None = None,
+    ) -> None:
+        if not MODULE(module_name):
+            raise ValueError("Invalid module name", module_name)
+        self.name = name
+        self.module_name = module_name
+        self.attrs = tuple(attrs)
+        self.extras = tuple(extras)
+        self.dist = dist
+
+    def __str__(self) -> str:
+        s = f"{self.name} = {self.module_name}"
+        if self.attrs:
+            s += ':' + '.'.join(self.attrs)
+        if self.extras:
+            extras = ','.join(self.extras)
+            s += f' [{extras}]'
+        return s
+
+    def __repr__(self) -> str:
+        return f"EntryPoint.parse({str(self)!r})"
+
+    @overload
+    def load(
+        self,
+        require: Literal[True] = True,
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+    ) -> _ResolvedEntryPoint: ...
+    @overload
+    def load(
+        self,
+        require: Literal[False],
+        *args: Any,
+        **kwargs: Any,
+    ) -> _ResolvedEntryPoint: ...
+    def load(
+        self,
+        require: bool = True,
+        *args: Environment | _InstallerType | None,
+        **kwargs: Environment | _InstallerType | None,
+    ) -> _ResolvedEntryPoint:
+        """
+        Require packages for this EntryPoint, then resolve it.
+        """
+        if not require or args or kwargs:
+            warnings.warn(
+                "Parameters to load are deprecated.  Call .resolve and "
+                ".require separately.",
+                PkgResourcesDeprecationWarning,
+                stacklevel=2,
+            )
+        if require:
+            # We could pass `env` and `installer` directly,
+            # but keeping `*args` and `**kwargs` for backwards compatibility
+            self.require(*args, **kwargs)  # type: ignore[arg-type]
+        return self.resolve()
+
+    def resolve(self) -> _ResolvedEntryPoint:
+        """
+        Resolve the entry point from its module and attrs.
+        """
+        module = __import__(self.module_name, fromlist=['__name__'], level=0)
+        try:
+            return functools.reduce(getattr, self.attrs, module)
+        except AttributeError as exc:
+            raise ImportError(str(exc)) from exc
+
+    def require(
+        self,
+        env: Environment | None = None,
+        installer: _InstallerType | None = None,
+    ) -> None:
+        if not self.dist:
+            error_cls = UnknownExtra if self.extras else AttributeError
+            raise error_cls("Can't require() without a distribution", self)
+
+        # Get the requirements for this entry point with all its extras and
+        # then resolve them. We have to pass `extras` along when resolving so
+        # that the working set knows what extras we want. Otherwise, for
+        # dist-info distributions, the working set will assume that the
+        # requirements for that extra are purely optional and skip over them.
+        reqs = self.dist.requires(self.extras)
+        items = working_set.resolve(reqs, env, installer, extras=self.extras)
+        list(map(working_set.add, items))
+
+    pattern = re.compile(
+        r'\s*'
+        r'(?P<name>.+?)\s*'
+        r'=\s*'
+        r'(?P<module>[\w.]+)\s*'
+        r'(:\s*(?P<attr>[\w.]+))?\s*'
+        r'(?P<extras>\[.*\])?\s*$'
+    )
+
+    @classmethod
+    def parse(cls, src: str, dist: Distribution | None = None) -> Self:
+        """Parse a single entry point from string `src`
+
+        Entry point syntax follows the form::
+
+            name = some.module:some.attr [extra1, extra2]
+
+        The entry name and module name are required, but the ``:attrs`` and
+        ``[extras]`` parts are optional
+        """
+        m = cls.pattern.match(src)
+        if not m:
+            msg = "EntryPoint must be in 'name=module:attrs [extras]' format"
+            raise ValueError(msg, src)
+        res = m.groupdict()
+        extras = cls._parse_extras(res['extras'])
+        attrs = res['attr'].split('.') if res['attr'] else ()
+        return cls(res['name'], res['module'], attrs, extras, dist)
+
+    @classmethod
+    def _parse_extras(cls, extras_spec):
+        if not extras_spec:
+            return ()
+        req = Requirement.parse('x' + extras_spec)
+        if req.specs:
+            raise ValueError
+        return req.extras
+
+    @classmethod
+    def parse_group(
+        cls,
+        group: str,
+        lines: _NestedStr,
+        dist: Distribution | None = None,
+    ) -> dict[str, Self]:
+        """Parse an entry point group"""
+        if not MODULE(group):
+            raise ValueError("Invalid group name", group)
+        this: dict[str, Self] = {}
+        for line in yield_lines(lines):
+            ep = cls.parse(line, dist)
+            if ep.name in this:
+                raise ValueError("Duplicate entry point", group, ep.name)
+            this[ep.name] = ep
+        return this
+
+    @classmethod
+    def parse_map(
+        cls,
+        data: str | Iterable[str] | dict[str, str | Iterable[str]],
+        dist: Distribution | None = None,
+    ) -> dict[str, dict[str, Self]]:
+        """Parse a map of entry point groups"""
+        _data: Iterable[tuple[str | None, str | Iterable[str]]]
+        if isinstance(data, dict):
+            _data = data.items()
+        else:
+            _data = split_sections(data)
+        maps: dict[str, dict[str, Self]] = {}
+        for group, lines in _data:
+            if group is None:
+                if not lines:
+                    continue
+                raise ValueError("Entry points must be listed in groups")
+            group = group.strip()
+            if group in maps:
+                raise ValueError("Duplicate group name", group)
+            maps[group] = cls.parse_group(group, lines, dist)
+        return maps
+
+
+def _version_from_file(lines):
+    """
+    Given an iterable of lines from a Metadata file, return
+    the value of the Version field, if present, or None otherwise.
+    """
+
+    def is_version_line(line):
+        return line.lower().startswith('version:')
+
+    version_lines = filter(is_version_line, lines)
+    line = next(iter(version_lines), '')
+    _, _, value = line.partition(':')
+    return safe_version(value.strip()) or None
+
+
+class Distribution:
+    """Wrap an actual or potential sys.path entry w/metadata"""
+
+    PKG_INFO = 'PKG-INFO'
+
+    def __init__(
+        self,
+        location: str | None = None,
+        metadata: _MetadataType = None,
+        project_name: str | None = None,
+        version: str | None = None,
+        py_version: str | None = PY_MAJOR,
+        platform: str | None = None,
+        precedence: int = EGG_DIST,
+    ) -> None:
+        self.project_name = safe_name(project_name or 'Unknown')
+        if version is not None:
+            self._version = safe_version(version)
+        self.py_version = py_version
+        self.platform = platform
+        self.location = location
+        self.precedence = precedence
+        self._provider = metadata or empty_provider
+
+    @classmethod
+    def from_location(
+        cls,
+        location: str,
+        basename: StrPath,
+        metadata: _MetadataType = None,
+        **kw: int,  # We could set `precedence` explicitly, but keeping this as `**kw` for full backwards and subclassing compatibility
+    ) -> Distribution:
+        project_name, version, py_version, platform = [None] * 4
+        basename, ext = os.path.splitext(basename)
+        if ext.lower() in _distributionImpl:
+            cls = _distributionImpl[ext.lower()]
+
+            match = EGG_NAME(basename)
+            if match:
+                project_name, version, py_version, platform = match.group(
+                    'name', 'ver', 'pyver', 'plat'
+                )
+        return cls(
+            location,
+            metadata,
+            project_name=project_name,
+            version=version,
+            py_version=py_version,
+            platform=platform,
+            **kw,
+        )._reload_version()
+
+    def _reload_version(self):
+        return self
+
+    @property
+    def hashcmp(self):
+        return (
+            self._forgiving_parsed_version,
+            self.precedence,
+            self.key,
+            self.location,
+            self.py_version or '',
+            self.platform or '',
+        )
+
+    def __hash__(self) -> int:
+        return hash(self.hashcmp)
+
+    def __lt__(self, other: Distribution) -> bool:
+        return self.hashcmp < other.hashcmp
+
+    def __le__(self, other: Distribution) -> bool:
+        return self.hashcmp <= other.hashcmp
+
+    def __gt__(self, other: Distribution) -> bool:
+        return self.hashcmp > other.hashcmp
+
+    def __ge__(self, other: Distribution) -> bool:
+        return self.hashcmp >= other.hashcmp
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, self.__class__):
+            # It's not a Distribution, so they are not equal
+            return False
+        return self.hashcmp == other.hashcmp
+
+    def __ne__(self, other: object) -> bool:
+        return not self == other
+
+    # These properties have to be lazy so that we don't have to load any
+    # metadata until/unless it's actually needed.  (i.e., some distributions
+    # may not know their name or version without loading PKG-INFO)
+
+    @property
+    def key(self):
+        try:
+            return self._key
+        except AttributeError:
+            self._key = key = self.project_name.lower()
+            return key
+
+    @property
+    def parsed_version(self):
+        if not hasattr(self, "_parsed_version"):
+            try:
+                self._parsed_version = parse_version(self.version)
+            except packaging.version.InvalidVersion as ex:
+                info = f"(package: {self.project_name})"
+                if hasattr(ex, "add_note"):
+                    ex.add_note(info)  # PEP 678
+                    raise
+                raise packaging.version.InvalidVersion(f"{str(ex)} {info}") from None
+
+        return self._parsed_version
+
+    @property
+    def _forgiving_parsed_version(self):
+        try:
+            return self.parsed_version
+        except packaging.version.InvalidVersion as ex:
+            self._parsed_version = parse_version(_forgiving_version(self.version))
+
+            notes = "\n".join(getattr(ex, "__notes__", []))  # PEP 678
+            msg = f"""!!\n\n
+            *************************************************************************
+            {str(ex)}\n{notes}
+
+            This is a long overdue deprecation.
+            For the time being, `pkg_resources` will use `{self._parsed_version}`
+            as a replacement to avoid breaking existing environments,
+            but no future compatibility is guaranteed.
+
+            If you maintain package {self.project_name} you should implement
+            the relevant changes to adequate the project to PEP 440 immediately.
+            *************************************************************************
+            \n\n!!
+            """
+            warnings.warn(msg, DeprecationWarning)
+
+            return self._parsed_version
+
+    @property
+    def version(self):
+        try:
+            return self._version
+        except AttributeError as e:
+            version = self._get_version()
+            if version is None:
+                path = self._get_metadata_path_for_display(self.PKG_INFO)
+                msg = f"Missing 'Version:' header and/or {self.PKG_INFO} file at path: {path}"
+                raise ValueError(msg, self) from e
+
+            return version
+
+    @property
+    def _dep_map(self):
+        """
+        A map of extra to its list of (direct) requirements
+        for this distribution, including the null extra.
+        """
+        try:
+            return self.__dep_map
+        except AttributeError:
+            self.__dep_map = self._filter_extras(self._build_dep_map())
+        return self.__dep_map
+
+    @staticmethod
+    def _filter_extras(
+        dm: dict[str | None, list[Requirement]],
+    ) -> dict[str | None, list[Requirement]]:
+        """
+        Given a mapping of extras to dependencies, strip off
+        environment markers and filter out any dependencies
+        not matching the markers.
+        """
+        for extra in list(filter(None, dm)):
+            new_extra: str | None = extra
+            reqs = dm.pop(extra)
+            new_extra, _, marker = extra.partition(':')
+            fails_marker = marker and (
+                invalid_marker(marker) or not evaluate_marker(marker)
+            )
+            if fails_marker:
+                reqs = []
+            new_extra = safe_extra(new_extra) or None
+
+            dm.setdefault(new_extra, []).extend(reqs)
+        return dm
+
+    def _build_dep_map(self):
+        dm = {}
+        for name in 'requires.txt', 'depends.txt':
+            for extra, reqs in split_sections(self._get_metadata(name)):
+                dm.setdefault(extra, []).extend(parse_requirements(reqs))
+        return dm
+
+    def requires(self, extras: Iterable[str] = ()) -> list[Requirement]:
+        """List of Requirements needed for this distro if `extras` are used"""
+        dm = self._dep_map
+        deps: list[Requirement] = []
+        deps.extend(dm.get(None, ()))
+        for ext in extras:
+            try:
+                deps.extend(dm[safe_extra(ext)])
+            except KeyError as e:
+                raise UnknownExtra(f"{self} has no such extra feature {ext!r}") from e
+        return deps
+
+    def _get_metadata_path_for_display(self, name):
+        """
+        Return the path to the given metadata file, if available.
+        """
+        try:
+            # We need to access _get_metadata_path() on the provider object
+            # directly rather than through this class's __getattr__()
+            # since _get_metadata_path() is marked private.
+            path = self._provider._get_metadata_path(name)
+
+        # Handle exceptions e.g. in case the distribution's metadata
+        # provider doesn't support _get_metadata_path().
+        except Exception:
+            return '[could not detect]'
+
+        return path
+
+    def _get_metadata(self, name):
+        if self.has_metadata(name):
+            yield from self.get_metadata_lines(name)
+
+    def _get_version(self):
+        lines = self._get_metadata(self.PKG_INFO)
+        return _version_from_file(lines)
+
+    def activate(self, path: list[str] | None = None, replace: bool = False) -> None:
+        """Ensure distribution is importable on `path` (default=sys.path)"""
+        if path is None:
+            path = sys.path
+        self.insert_on(path, replace=replace)
+        if path is sys.path and self.location is not None:
+            fixup_namespace_packages(self.location)
+            for pkg in self._get_metadata('namespace_packages.txt'):
+                if pkg in sys.modules:
+                    declare_namespace(pkg)
+
+    def egg_name(self):
+        """Return what this distribution's standard .egg filename should be"""
+        filename = f"{to_filename(self.project_name)}-{to_filename(self.version)}-py{self.py_version or PY_MAJOR}"
+
+        if self.platform:
+            filename += '-' + self.platform
+        return filename
+
+    def __repr__(self) -> str:
+        if self.location:
+            return f"{self} ({self.location})"
+        else:
+            return str(self)
+
+    def __str__(self) -> str:
+        try:
+            version = getattr(self, 'version', None)
+        except ValueError:
+            version = None
+        version = version or "[unknown version]"
+        return f"{self.project_name} {version}"
+
+    def __getattr__(self, attr: str):
+        """Delegate all unrecognized public attributes to .metadata provider"""
+        if attr.startswith('_'):
+            raise AttributeError(attr)
+        return getattr(self._provider, attr)
+
+    def __dir__(self):
+        return list(
+            set(super().__dir__())
+            | set(attr for attr in self._provider.__dir__() if not attr.startswith('_'))
+        )
+
+    @classmethod
+    def from_filename(
+        cls,
+        filename: StrPath,
+        metadata: _MetadataType = None,
+        **kw: int,  # We could set `precedence` explicitly, but keeping this as `**kw` for full backwards and subclassing compatibility
+    ) -> Distribution:
+        return cls.from_location(
+            _normalize_cached(filename), os.path.basename(filename), metadata, **kw
+        )
+
+    def as_requirement(self):
+        """Return a ``Requirement`` that matches this distribution exactly"""
+        if isinstance(self.parsed_version, packaging.version.Version):
+            spec = f"{self.project_name}=={self.parsed_version}"
+        else:
+            spec = f"{self.project_name}==={self.parsed_version}"
+
+        return Requirement.parse(spec)
+
+    def load_entry_point(self, group: str, name: str) -> _ResolvedEntryPoint:
+        """Return the `name` entry point of `group` or raise ImportError"""
+        ep = self.get_entry_info(group, name)
+        if ep is None:
+            raise ImportError(f"Entry point {(group, name)!r} not found")
+        return ep.load()
+
+    @overload
+    def get_entry_map(self, group: None = None) -> dict[str, dict[str, EntryPoint]]: ...
+    @overload
+    def get_entry_map(self, group: str) -> dict[str, EntryPoint]: ...
+    def get_entry_map(self, group: str | None = None):
+        """Return the entry point map for `group`, or the full entry map"""
+        if not hasattr(self, "_ep_map"):
+            self._ep_map = EntryPoint.parse_map(
+                self._get_metadata('entry_points.txt'), self
+            )
+        if group is not None:
+            return self._ep_map.get(group, {})
+        return self._ep_map
+
+    def get_entry_info(self, group: str, name: str) -> EntryPoint | None:
+        """Return the EntryPoint object for `group`+`name`, or ``None``"""
+        return self.get_entry_map(group).get(name)
+
+    # FIXME: 'Distribution.insert_on' is too complex (13)
+    def insert_on(  # noqa: C901
+        self,
+        path: list[str],
+        loc=None,
+        replace: bool = False,
+    ) -> None:
+        """Ensure self.location is on path
+
+        If replace=False (default):
+            - If location is already in path anywhere, do nothing.
+            - Else:
+              - If it's an egg and its parent directory is on path,
+                insert just ahead of the parent.
+              - Else: add to the end of path.
+        If replace=True:
+            - If location is already on path anywhere (not eggs)
+              or higher priority than its parent (eggs)
+              do nothing.
+            - Else:
+              - If it's an egg and its parent directory is on path,
+                insert just ahead of the parent,
+                removing any lower-priority entries.
+              - Else: add it to the front of path.
+        """
+
+        loc = loc or self.location
+        if not loc:
+            return
+
+        nloc = _normalize_cached(loc)
+        bdir = os.path.dirname(nloc)
+        npath = [(p and _normalize_cached(p) or p) for p in path]
+
+        for p, item in enumerate(npath):
+            if item == nloc:
+                if replace:
+                    break
+                else:
+                    # don't modify path (even removing duplicates) if
+                    # found and not replace
+                    return
+            elif item == bdir and self.precedence == EGG_DIST:
+                # if it's an .egg, give it precedence over its directory
+                # UNLESS it's already been added to sys.path and replace=False
+                if (not replace) and nloc in npath[p:]:
+                    return
+                if path is sys.path:
+                    self.check_version_conflict()
+                path.insert(p, loc)
+                npath.insert(p, nloc)
+                break
+        else:
+            if path is sys.path:
+                self.check_version_conflict()
+            if replace:
+                path.insert(0, loc)
+            else:
+                path.append(loc)
+            return
+
+        # p is the spot where we found or inserted loc; now remove duplicates
+        while True:
+            try:
+                np = npath.index(nloc, p + 1)
+            except ValueError:
+                break
+            else:
+                del npath[np], path[np]
+                # ha!
+                p = np
+
+        return
+
+    def check_version_conflict(self):
+        if self.key == 'setuptools':
+            # ignore the inevitable setuptools self-conflicts  :(
+            return
+
+        nsp = dict.fromkeys(self._get_metadata('namespace_packages.txt'))
+        loc = normalize_path(self.location)
+        for modname in self._get_metadata('top_level.txt'):
+            if (
+                modname not in sys.modules
+                or modname in nsp
+                or modname in _namespace_packages
+            ):
+                continue
+            if modname in ('pkg_resources', 'setuptools', 'site'):
+                continue
+            fn = getattr(sys.modules[modname], '__file__', None)
+            if fn and (
+                normalize_path(fn).startswith(loc) or fn.startswith(self.location)
+            ):
+                continue
+            issue_warning(
+                f"Module {modname} was already imported from {fn}, "
+                f"but {self.location} is being added to sys.path",
+            )
+
+    def has_version(self) -> bool:
+        try:
+            self.version
+        except ValueError:
+            issue_warning("Unbuilt egg for " + repr(self))
+            return False
+        except SystemError:
+            # TODO: remove this except clause when python/cpython#103632 is fixed.
+            return False
+        return True
+
+    def clone(self, **kw: str | int | IResourceProvider | None) -> Self:
+        """Copy this distribution, substituting in any changed keyword args"""
+        names = 'project_name version py_version platform location precedence'
+        for attr in names.split():
+            kw.setdefault(attr, getattr(self, attr, None))
+        kw.setdefault('metadata', self._provider)
+        # Unsafely unpacking. But keeping **kw for backwards and subclassing compatibility
+        return self.__class__(**kw)  # type:ignore[arg-type]
+
+    @property
+    def extras(self):
+        return [dep for dep in self._dep_map if dep]
+
+
+class EggInfoDistribution(Distribution):
+    def _reload_version(self):
+        """
+        Packages installed by distutils (e.g. numpy or scipy),
+        which uses an old safe_version, and so
+        their version numbers can get mangled when
+        converted to filenames (e.g., 1.11.0.dev0+2329eae to
+        1.11.0.dev0_2329eae). These distributions will not be
+        parsed properly
+        downstream by Distribution and safe_version, so
+        take an extra step and try to get the version number from
+        the metadata file itself instead of the filename.
+        """
+        md_version = self._get_version()
+        if md_version:
+            self._version = md_version
+        return self
+
+
+class DistInfoDistribution(Distribution):
+    """
+    Wrap an actual or potential sys.path entry
+    w/metadata, .dist-info style.
+    """
+
+    PKG_INFO = 'METADATA'
+    EQEQ = re.compile(r"([\(,])\s*(\d.*?)\s*([,\)])")
+
+    @property
+    def _parsed_pkg_info(self):
+        """Parse and cache metadata"""
+        try:
+            return self._pkg_info
+        except AttributeError:
+            metadata = self.get_metadata(self.PKG_INFO)
+            self._pkg_info = email.parser.Parser().parsestr(metadata)
+            return self._pkg_info
+
+    @property
+    def _dep_map(self):
+        try:
+            return self.__dep_map
+        except AttributeError:
+            self.__dep_map = self._compute_dependencies()
+            return self.__dep_map
+
+    def _compute_dependencies(self) -> dict[str | None, list[Requirement]]:
+        """Recompute this distribution's dependencies."""
+        self.__dep_map: dict[str | None, list[Requirement]] = {None: []}
+
+        reqs: list[Requirement] = []
+        # Including any condition expressions
+        for req in self._parsed_pkg_info.get_all('Requires-Dist') or []:
+            reqs.extend(parse_requirements(req))
+
+        def reqs_for_extra(extra):
+            for req in reqs:
+                if not req.marker or req.marker.evaluate({'extra': extra}):
+                    yield req
+
+        common = types.MappingProxyType(dict.fromkeys(reqs_for_extra(None)))
+        self.__dep_map[None].extend(common)
+
+        for extra in self._parsed_pkg_info.get_all('Provides-Extra') or []:
+            s_extra = safe_extra(extra.strip())
+            self.__dep_map[s_extra] = [
+                r for r in reqs_for_extra(extra) if r not in common
+            ]
+
+        return self.__dep_map
+
+
+_distributionImpl = {
+    '.egg': Distribution,
+    '.egg-info': EggInfoDistribution,
+    '.dist-info': DistInfoDistribution,
+}
+
+
+def issue_warning(*args, **kw):
+    level = 1
+    g = globals()
+    try:
+        # find the first stack frame that is *not* code in
+        # the pkg_resources module, to use for the warning
+        while sys._getframe(level).f_globals is g:
+            level += 1
+    except ValueError:
+        pass
+    warnings.warn(stacklevel=level + 1, *args, **kw)
+
+
+def parse_requirements(strs: _NestedStr) -> map[Requirement]:
+    """
+    Yield ``Requirement`` objects for each specification in `strs`.
+
+    `strs` must be a string, or a (possibly-nested) iterable thereof.
+    """
+    return map(Requirement, join_continuation(map(drop_comment, yield_lines(strs))))
+
+
+class RequirementParseError(packaging.requirements.InvalidRequirement):
+    "Compatibility wrapper for InvalidRequirement"
+
+
+class Requirement(packaging.requirements.Requirement):
+    # prefer variable length tuple to set (as found in
+    # packaging.requirements.Requirement)
+    extras: tuple[str, ...]  # type: ignore[assignment]
+
+    def __init__(self, requirement_string: str) -> None:
+        """DO NOT CALL THIS UNDOCUMENTED METHOD; use Requirement.parse()!"""
+        super().__init__(requirement_string)
+        self.unsafe_name = self.name
+        project_name = safe_name(self.name)
+        self.project_name, self.key = project_name, project_name.lower()
+        self.specs = [(spec.operator, spec.version) for spec in self.specifier]
+        self.extras = tuple(map(safe_extra, self.extras))
+        self.hashCmp = (
+            self.key,
+            self.url,
+            self.specifier,
+            frozenset(self.extras),
+            str(self.marker) if self.marker else None,
+        )
+        self.__hash = hash(self.hashCmp)
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, Requirement) and self.hashCmp == other.hashCmp
+
+    def __ne__(self, other: object) -> bool:
+        return not self == other
+
+    def __contains__(
+        self, item: Distribution | packaging.specifiers.UnparsedVersion
+    ) -> bool:
+        if isinstance(item, Distribution):
+            if item.key != self.key:
+                return False
+
+            version = item.version
+        else:
+            version = item
+
+        # Allow prereleases always in order to match the previous behavior of
+        # this method. In the future this should be smarter and follow PEP 440
+        # more accurately.
+        return self.specifier.contains(
+            version,
+            prereleases=True,
+        )
+
+    def __hash__(self) -> int:
+        return self.__hash
+
+    def __repr__(self) -> str:
+        return f"Requirement.parse({str(self)!r})"
+
+    @staticmethod
+    def parse(s: str | Iterable[str]) -> Requirement:
+        (req,) = parse_requirements(s)
+        return req
+
+
+def _always_object(classes):
+    """
+    Ensure object appears in the mro even
+    for old-style classes.
+    """
+    if object not in classes:
+        return classes + (object,)
+    return classes
+
+
+def _find_adapter(registry: Mapping[type, _AdapterT], ob: object) -> _AdapterT:
+    """Return an adapter factory for `ob` from `registry`"""
+    types = _always_object(inspect.getmro(getattr(ob, '__class__', type(ob))))
+    for t in types:
+        if t in registry:
+            return registry[t]
+    # _find_adapter would previously return None, and immediately be called.
+    # So we're raising a TypeError to keep backward compatibility if anyone depended on that behaviour.
+    raise TypeError(f"Could not find adapter for {registry} and {ob}")
+
+
+def ensure_directory(path: StrOrBytesPath) -> None:
+    """Ensure that the parent directory of `path` exists"""
+    dirname = os.path.dirname(path)
+    os.makedirs(dirname, exist_ok=True)
+
+
+def _bypass_ensure_directory(path) -> None:
+    """Sandbox-bypassing version of ensure_directory()"""
+    if not WRITE_SUPPORT:
+        raise OSError('"os.mkdir" not supported on this platform.')
+    dirname, filename = split(path)
+    if dirname and filename and not isdir(dirname):
+        _bypass_ensure_directory(dirname)
+        try:
+            mkdir(dirname, 0o755)
+        except FileExistsError:
+            pass
+
+
+def split_sections(s: _NestedStr) -> Iterator[tuple[str | None, list[str]]]:
+    """Split a string or iterable thereof into (section, content) pairs
+
+    Each ``section`` is a stripped version of the section header ("[section]")
+    and each ``content`` is a list of stripped lines excluding blank lines and
+    comment-only lines.  If there are any such lines before the first section
+    header, they're returned in a first ``section`` of ``None``.
+    """
+    section = None
+    content: list[str] = []
+    for line in yield_lines(s):
+        if line.startswith("["):
+            if line.endswith("]"):
+                if section or content:
+                    yield section, content
+                section = line[1:-1].strip()
+                content = []
+            else:
+                raise ValueError("Invalid section heading", line)
+        else:
+            content.append(line)
+
+    # wrap up last segment
+    yield section, content
+
+
+def _mkstemp(*args, **kw):
+    old_open = os.open
+    try:
+        # temporarily bypass sandboxing
+        os.open = os_open
+        return tempfile.mkstemp(*args, **kw)
+    finally:
+        # and then put it back
+        os.open = old_open
+
+
+# Silence the PEP440Warning by default, so that end users don't get hit by it
+# randomly just because they use pkg_resources. We want to append the rule
+# because we want earlier uses of filterwarnings to take precedence over this
+# one.
+warnings.filterwarnings("ignore", category=PEP440Warning, append=True)
+
+
+class PkgResourcesDeprecationWarning(Warning):
+    """
+    Base class for warning about deprecations in ``pkg_resources``
+
+    This class is not derived from ``DeprecationWarning``, and as such is
+    visible by default.
+    """
+
+
+# Ported from ``setuptools`` to avoid introducing an import inter-dependency:
+_LOCALE_ENCODING = "locale" if sys.version_info >= (3, 10) else None
+
+
+# This must go before calls to `_call_aside`. See https://github.com/pypa/setuptools/pull/4422
+def _read_utf8_with_fallback(file: str, fallback_encoding=_LOCALE_ENCODING) -> str:
+    """See setuptools.unicode_utils._read_utf8_with_fallback"""
+    try:
+        with open(file, "r", encoding="utf-8") as f:
+            return f.read()
+    except UnicodeDecodeError:  # pragma: no cover
+        msg = f"""\
+        ********************************************************************************
+        `encoding="utf-8"` fails with {file!r}, trying `encoding={fallback_encoding!r}`.
+
+        This fallback behaviour is considered **deprecated** and future versions of
+        `setuptools/pkg_resources` may not implement it.
+
+        Please encode {file!r} with "utf-8" to ensure future builds will succeed.
+
+        If this file was produced by `setuptools` itself, cleaning up the cached files
+        and re-building/re-installing the package with a newer version of `setuptools`
+        (e.g. by updating `build-system.requires` in its `pyproject.toml`)
+        might solve the problem.
+        ********************************************************************************
+        """
+        # TODO: Add a deadline?
+        #       See comment in setuptools.unicode_utils._Utf8EncodingNeeded
+        warnings.warn(msg, PkgResourcesDeprecationWarning, stacklevel=2)
+        with open(file, "r", encoding=fallback_encoding) as f:
+            return f.read()
+
+
+# from jaraco.functools 1.3
+def _call_aside(f, *args, **kwargs):
+    f(*args, **kwargs)
+    return f
+
+
+@_call_aside
+def _initialize(g=globals()) -> None:
+    "Set up global resource manager (deliberately not state-saved)"
+    manager = ResourceManager()
+    g['_manager'] = manager
+    g.update(
+        (name, getattr(manager, name))
+        for name in dir(manager)
+        if not name.startswith('_')
+    )
+
+
+@_call_aside
+def _initialize_master_working_set() -> None:
+    """
+    Prepare the master working set and make the ``require()``
+    API available.
+
+    This function has explicit effects on the global state
+    of pkg_resources. It is intended to be invoked once at
+    the initialization of this module.
+
+    Invocation by other packages is unsupported and done
+    at their own risk.
+    """
+    working_set = _declare_state('object', 'working_set', WorkingSet._build_master())
+
+    require = working_set.require
+    iter_entry_points = working_set.iter_entry_points
+    add_activation_listener = working_set.subscribe
+    run_script = working_set.run_script
+    # backward compatibility
+    run_main = run_script
+    # Activate all distributions already on sys.path with replace=False and
+    # ensure that all distributions added to the working set in the future
+    # (e.g. by calling ``require()``) will get activated as well,
+    # with higher priority (replace=True).
+    tuple(dist.activate(replace=False) for dist in working_set)
+    add_activation_listener(
+        lambda dist: dist.activate(replace=True),
+        existing=False,
+    )
+    working_set.entries = []
+    # match order
+    list(map(working_set.add_entry, sys.path))
+    globals().update(locals())
+
+
+if TYPE_CHECKING:
+    # All of these are set by the @_call_aside methods above
+    __resource_manager = ResourceManager()  # Won't exist at runtime
+    resource_exists = __resource_manager.resource_exists
+    resource_isdir = __resource_manager.resource_isdir
+    resource_filename = __resource_manager.resource_filename
+    resource_stream = __resource_manager.resource_stream
+    resource_string = __resource_manager.resource_string
+    resource_listdir = __resource_manager.resource_listdir
+    set_extraction_path = __resource_manager.set_extraction_path
+    cleanup_resources = __resource_manager.cleanup_resources
+
+    working_set = WorkingSet()
+    require = working_set.require
+    iter_entry_points = working_set.iter_entry_points
+    add_activation_listener = working_set.subscribe
+    run_script = working_set.run_script
+    run_main = run_script
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b44127c9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/api_tests.txt b/backend/venv/lib/python3.12/site-packages/pkg_resources/api_tests.txt
new file mode 100644
index 00000000..d72b85aa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/api_tests.txt
@@ -0,0 +1,424 @@
+Pluggable Distributions of Python Software
+==========================================
+
+Distributions
+-------------
+
+A "Distribution" is a collection of files that represent a "Release" of a
+"Project" as of a particular point in time, denoted by a
+"Version"::
+
+    >>> import sys, pkg_resources
+    >>> from pkg_resources import Distribution
+    >>> Distribution(project_name="Foo", version="1.2")
+    Foo 1.2
+
+Distributions have a location, which can be a filename, URL, or really anything
+else you care to use::
+
+    >>> dist = Distribution(
+    ...     location="http://example.com/something",
+    ...     project_name="Bar", version="0.9"
+    ... )
+
+    >>> dist
+    Bar 0.9 (http://example.com/something)
+
+
+Distributions have various introspectable attributes::
+
+    >>> dist.location
+    'http://example.com/something'
+
+    >>> dist.project_name
+    'Bar'
+
+    >>> dist.version
+    '0.9'
+
+    >>> dist.py_version == '{}.{}'.format(*sys.version_info)
+    True
+
+    >>> print(dist.platform)
+    None
+
+Including various computed attributes::
+
+    >>> from pkg_resources import parse_version
+    >>> dist.parsed_version == parse_version(dist.version)
+    True
+
+    >>> dist.key    # case-insensitive form of the project name
+    'bar'
+
+Distributions are compared (and hashed) by version first::
+
+    >>> Distribution(version='1.0') == Distribution(version='1.0')
+    True
+    >>> Distribution(version='1.0') == Distribution(version='1.1')
+    False
+    >>> Distribution(version='1.0') <  Distribution(version='1.1')
+    True
+
+but also by project name (case-insensitive), platform, Python version,
+location, etc.::
+
+    >>> Distribution(project_name="Foo",version="1.0") == \
+    ... Distribution(project_name="Foo",version="1.0")
+    True
+
+    >>> Distribution(project_name="Foo",version="1.0") == \
+    ... Distribution(project_name="foo",version="1.0")
+    True
+
+    >>> Distribution(project_name="Foo",version="1.0") == \
+    ... Distribution(project_name="Foo",version="1.1")
+    False
+
+    >>> Distribution(project_name="Foo",py_version="2.3",version="1.0") == \
+    ... Distribution(project_name="Foo",py_version="2.4",version="1.0")
+    False
+
+    >>> Distribution(location="spam",version="1.0") == \
+    ... Distribution(location="spam",version="1.0")
+    True
+
+    >>> Distribution(location="spam",version="1.0") == \
+    ... Distribution(location="baz",version="1.0")
+    False
+
+
+
+Hash and compare distribution by prio/plat
+
+Get version from metadata
+provider capabilities
+egg_name()
+as_requirement()
+from_location, from_filename (w/path normalization)
+
+Releases may have zero or more "Requirements", which indicate
+what releases of another project the release requires in order to
+function.  A Requirement names the other project, expresses some criteria
+as to what releases of that project are acceptable, and lists any "Extras"
+that the requiring release may need from that project.  (An Extra is an
+optional feature of a Release, that can only be used if its additional
+Requirements are satisfied.)
+
+
+
+The Working Set
+---------------
+
+A collection of active distributions is called a Working Set.  Note that a
+Working Set can contain any importable distribution, not just pluggable ones.
+For example, the Python standard library is an importable distribution that
+will usually be part of the Working Set, even though it is not pluggable.
+Similarly, when you are doing development work on a project, the files you are
+editing are also a Distribution.  (And, with a little attention to the
+directory names used,  and including some additional metadata, such a
+"development distribution" can be made pluggable as well.)
+
+    >>> from pkg_resources import WorkingSet
+
+A working set's entries are the sys.path entries that correspond to the active
+distributions.  By default, the working set's entries are the items on
+``sys.path``::
+
+    >>> ws = WorkingSet()
+    >>> ws.entries == sys.path
+    True
+
+But you can also create an empty working set explicitly, and add distributions
+to it::
+
+    >>> ws = WorkingSet([])
+    >>> ws.add(dist)
+    >>> ws.entries
+    ['http://example.com/something']
+    >>> dist in ws
+    True
+    >>> Distribution('foo',version="") in ws
+    False
+
+And you can iterate over its distributions::
+
+    >>> list(ws)
+    [Bar 0.9 (http://example.com/something)]
+
+Adding the same distribution more than once is a no-op::
+
+    >>> ws.add(dist)
+    >>> list(ws)
+    [Bar 0.9 (http://example.com/something)]
+
+For that matter, adding multiple distributions for the same project also does
+nothing, because a working set can only hold one active distribution per
+project -- the first one added to it::
+
+    >>> ws.add(
+    ...     Distribution(
+    ...         'http://example.com/something', project_name="Bar",
+    ...         version="7.2"
+    ...     )
+    ... )
+    >>> list(ws)
+    [Bar 0.9 (http://example.com/something)]
+
+You can append a path entry to a working set using ``add_entry()``::
+
+    >>> ws.entries
+    ['http://example.com/something']
+    >>> ws.add_entry(pkg_resources.__file__)
+    >>> ws.entries
+    ['http://example.com/something', '...pkg_resources...']
+
+Multiple additions result in multiple entries, even if the entry is already in
+the working set (because ``sys.path`` can contain the same entry more than
+once)::
+
+    >>> ws.add_entry(pkg_resources.__file__)
+    >>> ws.entries
+    ['...example.com...', '...pkg_resources...', '...pkg_resources...']
+
+And you can specify the path entry a distribution was found under, using the
+optional second parameter to ``add()``::
+
+    >>> ws = WorkingSet([])
+    >>> ws.add(dist,"foo")
+    >>> ws.entries
+    ['foo']
+
+But even if a distribution is found under multiple path entries, it still only
+shows up once when iterating the working set:
+
+    >>> ws.add_entry(ws.entries[0])
+    >>> list(ws)
+    [Bar 0.9 (http://example.com/something)]
+
+You can ask a WorkingSet to ``find()`` a distribution matching a requirement::
+
+    >>> from pkg_resources import Requirement
+    >>> print(ws.find(Requirement.parse("Foo==1.0")))   # no match, return None
+    None
+
+    >>> ws.find(Requirement.parse("Bar==0.9"))  # match, return distribution
+    Bar 0.9 (http://example.com/something)
+
+Note that asking for a conflicting version of a distribution already in a
+working set triggers a ``pkg_resources.VersionConflict`` error:
+
+    >>> try:
+    ...     ws.find(Requirement.parse("Bar==1.0"))
+    ... except pkg_resources.VersionConflict as exc:
+    ...     print(str(exc))
+    ... else:
+    ...     raise AssertionError("VersionConflict was not raised")
+    (Bar 0.9 (http://example.com/something), Requirement.parse('Bar==1.0'))
+
+You can subscribe a callback function to receive notifications whenever a new
+distribution is added to a working set.  The callback is immediately invoked
+once for each existing distribution in the working set, and then is called
+again for new distributions added thereafter::
+
+    >>> def added(dist): print("Added %s" % dist)
+    >>> ws.subscribe(added)
+    Added Bar 0.9
+    >>> foo12 = Distribution(project_name="Foo", version="1.2", location="f12")
+    >>> ws.add(foo12)
+    Added Foo 1.2
+
+Note, however, that only the first distribution added for a given project name
+will trigger a callback, even during the initial ``subscribe()`` callback::
+
+    >>> foo14 = Distribution(project_name="Foo", version="1.4", location="f14")
+    >>> ws.add(foo14)   # no callback, because Foo 1.2 is already active
+
+    >>> ws = WorkingSet([])
+    >>> ws.add(foo12)
+    >>> ws.add(foo14)
+    >>> ws.subscribe(added)
+    Added Foo 1.2
+
+And adding a callback more than once has no effect, either::
+
+    >>> ws.subscribe(added)     # no callbacks
+
+    # and no double-callbacks on subsequent additions, either
+    >>> just_a_test = Distribution(project_name="JustATest", version="0.99")
+    >>> ws.add(just_a_test)
+    Added JustATest 0.99
+
+
+Finding Plugins
+---------------
+
+``WorkingSet`` objects can be used to figure out what plugins in an
+``Environment`` can be loaded without any resolution errors::
+
+    >>> from pkg_resources import Environment
+
+    >>> plugins = Environment([])   # normally, a list of plugin directories
+    >>> plugins.add(foo12)
+    >>> plugins.add(foo14)
+    >>> plugins.add(just_a_test)
+
+In the simplest case, we just get the newest version of each distribution in
+the plugin environment::
+
+    >>> ws = WorkingSet([])
+    >>> ws.find_plugins(plugins)
+    ([JustATest 0.99, Foo 1.4 (f14)], {})
+
+But if there's a problem with a version conflict or missing requirements, the
+method falls back to older versions, and the error info dict will contain an
+exception instance for each unloadable plugin::
+
+    >>> ws.add(foo12)   # this will conflict with Foo 1.4
+    >>> ws.find_plugins(plugins)
+    ([JustATest 0.99, Foo 1.2 (f12)], {Foo 1.4 (f14): VersionConflict(...)})
+
+But if you disallow fallbacks, the failed plugin will be skipped instead of
+trying older versions::
+
+    >>> ws.find_plugins(plugins, fallback=False)
+    ([JustATest 0.99], {Foo 1.4 (f14): VersionConflict(...)})
+
+
+
+Platform Compatibility Rules
+----------------------------
+
+On the Mac, there are potential compatibility issues for modules compiled
+on newer versions of macOS than what the user is running. Additionally,
+macOS will soon have two platforms to contend with: Intel and PowerPC.
+
+Basic equality works as on other platforms::
+
+    >>> from pkg_resources import compatible_platforms as cp
+    >>> reqd = 'macosx-10.4-ppc'
+    >>> cp(reqd, reqd)
+    True
+    >>> cp("win32", reqd)
+    False
+
+Distributions made on other machine types are not compatible::
+
+    >>> cp("macosx-10.4-i386", reqd)
+    False
+
+Distributions made on earlier versions of the OS are compatible, as
+long as they are from the same top-level version. The patchlevel version
+number does not matter::
+
+    >>> cp("macosx-10.4-ppc", reqd)
+    True
+    >>> cp("macosx-10.3-ppc", reqd)
+    True
+    >>> cp("macosx-10.5-ppc", reqd)
+    False
+    >>> cp("macosx-9.5-ppc", reqd)
+    False
+
+Backwards compatibility for packages made via earlier versions of
+setuptools is provided as well::
+
+    >>> cp("darwin-8.2.0-Power_Macintosh", reqd)
+    True
+    >>> cp("darwin-7.2.0-Power_Macintosh", reqd)
+    True
+    >>> cp("darwin-8.2.0-Power_Macintosh", "macosx-10.3-ppc")
+    False
+
+
+Environment Markers
+-------------------
+
+    >>> from pkg_resources import invalid_marker as im, evaluate_marker as em
+    >>> import os
+
+    >>> print(im("sys_platform"))
+    Expected marker operator, one of <=, <, !=, ==, >=, >, ~=, ===, in, not in
+        sys_platform
+                    ^
+
+    >>> print(im("sys_platform=="))
+    Expected a marker variable or quoted string
+        sys_platform==
+                      ^
+
+    >>> print(im("sys_platform=='win32'"))
+    False
+
+    >>> print(im("sys=='x'"))
+    Expected a marker variable or quoted string
+        sys=='x'
+        ^
+
+    >>> print(im("(extra)"))
+    Expected marker operator, one of <=, <, !=, ==, >=, >, ~=, ===, in, not in
+        (extra)
+              ^
+
+    >>> print(im("(extra"))
+    Expected marker operator, one of <=, <, !=, ==, >=, >, ~=, ===, in, not in
+        (extra
+              ^
+
+    >>> print(im("os.open('foo')=='y'"))
+    Expected a marker variable or quoted string
+        os.open('foo')=='y'
+        ^
+
+    >>> print(im("'x'=='y' and os.open('foo')=='y'"))   # no short-circuit!
+    Expected a marker variable or quoted string
+        'x'=='y' and os.open('foo')=='y'
+                     ^
+
+    >>> print(im("'x'=='x' or os.open('foo')=='y'"))   # no short-circuit!
+    Expected a marker variable or quoted string
+        'x'=='x' or os.open('foo')=='y'
+                    ^
+
+    >>> print(im("r'x'=='x'"))
+    Expected a marker variable or quoted string
+        r'x'=='x'
+        ^
+
+    >>> print(im("'''x'''=='x'"))
+    Expected marker operator, one of <=, <, !=, ==, >=, >, ~=, ===, in, not in
+        '''x'''=='x'
+          ^
+
+    >>> print(im('"""x"""=="x"'))
+    Expected marker operator, one of <=, <, !=, ==, >=, >, ~=, ===, in, not in
+        """x"""=="x"
+          ^
+
+    >>> print(im(r"x\n=='x'"))
+    Expected a marker variable or quoted string
+        x\n=='x'
+        ^
+
+    >>> print(im("os.open=='y'"))
+    Expected a marker variable or quoted string
+        os.open=='y'
+        ^
+
+    >>> em("sys_platform=='win32'") == (sys.platform=='win32')
+    True
+
+    >>> em("python_version >= '2.7'")
+    True
+
+    >>> em("python_version > '2.6'")
+    True
+
+    >>> im("implementation_name=='cpython'")
+    False
+
+    >>> im("platform_python_implementation=='CPython'")
+    False
+
+    >>> im("implementation_version=='3.5.1'")
+    False
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/py.typed b/backend/venv/lib/python3.12/site-packages/pkg_resources/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__init__.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b5e120e5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_find_distributions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_find_distributions.cpython-312.pyc
new file mode 100644
index 00000000..45567e4d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_find_distributions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_integration_zope_interface.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_integration_zope_interface.cpython-312.pyc
new file mode 100644
index 00000000..cd4bf420
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_integration_zope_interface.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_markers.cpython-312.pyc
new file mode 100644
index 00000000..5df84711
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_pkg_resources.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_pkg_resources.cpython-312.pyc
new file mode 100644
index 00000000..f0750a33
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_pkg_resources.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_resources.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_resources.cpython-312.pyc
new file mode 100644
index 00000000..ef627097
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_resources.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_working_set.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_working_set.cpython-312.pyc
new file mode 100644
index 00000000..7ed306af
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/__pycache__/test_working_set.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/__pycache__/setup.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/__pycache__/setup.cpython-312.pyc
new file mode 100644
index 00000000..7e0e5763
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/__pycache__/setup.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/setup.cfg b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/setup.cfg
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/setup.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/setup.py
new file mode 100644
index 00000000..ce908064
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-source/setup.py
@@ -0,0 +1,7 @@
+import setuptools
+
+setuptools.setup(
+    name="my-test-package",
+    version="1.0",
+    zip_safe=True,
+)
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-zip/my-test-package.zip b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-zip/my-test-package.zip
new file mode 100644
index 00000000..81f9a017
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package-zip/my-test-package.zip differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/PKG-INFO b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/PKG-INFO
new file mode 100644
index 00000000..7328e3f7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/PKG-INFO
@@ -0,0 +1,10 @@
+Metadata-Version: 1.0
+Name: my-test-package
+Version: 1.0
+Summary: UNKNOWN
+Home-page: UNKNOWN
+Author: UNKNOWN
+Author-email: UNKNOWN
+License: UNKNOWN
+Description: UNKNOWN
+Platform: UNKNOWN
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/SOURCES.txt b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/SOURCES.txt
new file mode 100644
index 00000000..3c4ee167
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/SOURCES.txt
@@ -0,0 +1,7 @@
+setup.cfg
+setup.py
+my_test_package.egg-info/PKG-INFO
+my_test_package.egg-info/SOURCES.txt
+my_test_package.egg-info/dependency_links.txt
+my_test_package.egg-info/top_level.txt
+my_test_package.egg-info/zip-safe
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/dependency_links.txt b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/dependency_links.txt
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/dependency_links.txt
@@ -0,0 +1 @@
+
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/top_level.txt b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/top_level.txt
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/top_level.txt
@@ -0,0 +1 @@
+
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/zip-safe b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/zip-safe
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/zip-safe
@@ -0,0 +1 @@
+
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_zipped-egg/my_test_package-1.0-py3.7.egg b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_zipped-egg/my_test_package-1.0-py3.7.egg
new file mode 100644
index 00000000..5115b895
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/data/my-test-package_zipped-egg/my_test_package-1.0-py3.7.egg differ
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_find_distributions.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_find_distributions.py
new file mode 100644
index 00000000..301b36d6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_find_distributions.py
@@ -0,0 +1,56 @@
+import shutil
+from pathlib import Path
+
+import pytest
+
+import pkg_resources
+
+TESTS_DATA_DIR = Path(__file__).parent / 'data'
+
+
+class TestFindDistributions:
+    @pytest.fixture
+    def target_dir(self, tmpdir):
+        target_dir = tmpdir.mkdir('target')
+        # place a .egg named directory in the target that is not an egg:
+        target_dir.mkdir('not.an.egg')
+        return target_dir
+
+    def test_non_egg_dir_named_egg(self, target_dir):
+        dists = pkg_resources.find_distributions(str(target_dir))
+        assert not list(dists)
+
+    def test_standalone_egg_directory(self, target_dir):
+        shutil.copytree(
+            TESTS_DATA_DIR / 'my-test-package_unpacked-egg',
+            target_dir,
+            dirs_exist_ok=True,
+        )
+        dists = pkg_resources.find_distributions(str(target_dir))
+        assert [dist.project_name for dist in dists] == ['my-test-package']
+        dists = pkg_resources.find_distributions(str(target_dir), only=True)
+        assert not list(dists)
+
+    def test_zipped_egg(self, target_dir):
+        shutil.copytree(
+            TESTS_DATA_DIR / 'my-test-package_zipped-egg',
+            target_dir,
+            dirs_exist_ok=True,
+        )
+        dists = pkg_resources.find_distributions(str(target_dir))
+        assert [dist.project_name for dist in dists] == ['my-test-package']
+        dists = pkg_resources.find_distributions(str(target_dir), only=True)
+        assert not list(dists)
+
+    def test_zipped_sdist_one_level_removed(self, target_dir):
+        shutil.copytree(
+            TESTS_DATA_DIR / 'my-test-package-zip', target_dir, dirs_exist_ok=True
+        )
+        dists = pkg_resources.find_distributions(
+            str(target_dir / "my-test-package.zip")
+        )
+        assert [dist.project_name for dist in dists] == ['my-test-package']
+        dists = pkg_resources.find_distributions(
+            str(target_dir / "my-test-package.zip"), only=True
+        )
+        assert not list(dists)
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_integration_zope_interface.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_integration_zope_interface.py
new file mode 100644
index 00000000..4e37c340
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_integration_zope_interface.py
@@ -0,0 +1,54 @@
+import platform
+from inspect import cleandoc
+
+import jaraco.path
+import pytest
+
+pytestmark = pytest.mark.integration
+
+
+# For the sake of simplicity this test uses fixtures defined in
+# `setuptools.test.fixtures`,
+# and it also exercise conditions considered deprecated...
+# So if needed this test can be deleted.
+@pytest.mark.skipif(
+    platform.system() != "Linux",
+    reason="only demonstrated to fail on Linux in #4399",
+)
+def test_interop_pkg_resources_iter_entry_points(tmp_path, venv):
+    """
+    Importing pkg_resources.iter_entry_points on console_scripts
+    seems to cause trouble with zope-interface, when deprecates installation method
+    is used. See #4399.
+    """
+    project = {
+        "pkg": {
+            "foo.py": cleandoc(
+                """
+                from pkg_resources import iter_entry_points
+
+                def bar():
+                    print("Print me if you can")
+                """
+            ),
+            "setup.py": cleandoc(
+                """
+                from setuptools import setup, find_packages
+
+                setup(
+                    install_requires=["zope-interface==6.4.post2"],
+                    entry_points={
+                        "console_scripts": [
+                            "foo=foo:bar",
+                        ],
+                    },
+                )
+                """
+            ),
+        }
+    }
+    jaraco.path.build(project, prefix=tmp_path)
+    cmd = ["pip", "install", "-e", ".", "--no-use-pep517"]
+    venv.run(cmd, cwd=tmp_path / "pkg")  # Needs this version of pkg_resources installed
+    out = venv.run(["foo"])
+    assert "Print me if you can" in out
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_markers.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_markers.py
new file mode 100644
index 00000000..9306d5b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_markers.py
@@ -0,0 +1,8 @@
+from unittest import mock
+
+from pkg_resources import evaluate_marker
+
+
+@mock.patch('platform.python_version', return_value='2.7.10')
+def test_ordering(python_version_mock):
+    assert evaluate_marker("python_full_version > '2.7.3'") is True
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_pkg_resources.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_pkg_resources.py
new file mode 100644
index 00000000..0f696e85
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_pkg_resources.py
@@ -0,0 +1,427 @@
+from __future__ import annotations
+
+import builtins
+import datetime
+import os
+import plistlib
+import stat
+import subprocess
+import sys
+import tempfile
+import zipfile
+from unittest import mock
+
+import pytest
+
+import pkg_resources
+from pkg_resources import DistInfoDistribution, Distribution, EggInfoDistribution
+
+import distutils.command.install_egg_info
+import distutils.dist
+
+
+class EggRemover(str):
+    def __call__(self):
+        if self in sys.path:
+            sys.path.remove(self)
+        if os.path.exists(self):
+            os.remove(self)
+
+
+class TestZipProvider:
+    finalizers: list[EggRemover] = []
+
+    ref_time = datetime.datetime(2013, 5, 12, 13, 25, 0)
+    "A reference time for a file modification"
+
+    @classmethod
+    def setup_class(cls):
+        "create a zip egg and add it to sys.path"
+        egg = tempfile.NamedTemporaryFile(suffix='.egg', delete=False)
+        zip_egg = zipfile.ZipFile(egg, 'w')
+        zip_info = zipfile.ZipInfo()
+        zip_info.filename = 'mod.py'
+        zip_info.date_time = cls.ref_time.timetuple()
+        zip_egg.writestr(zip_info, 'x = 3\n')
+        zip_info = zipfile.ZipInfo()
+        zip_info.filename = 'data.dat'
+        zip_info.date_time = cls.ref_time.timetuple()
+        zip_egg.writestr(zip_info, 'hello, world!')
+        zip_info = zipfile.ZipInfo()
+        zip_info.filename = 'subdir/mod2.py'
+        zip_info.date_time = cls.ref_time.timetuple()
+        zip_egg.writestr(zip_info, 'x = 6\n')
+        zip_info = zipfile.ZipInfo()
+        zip_info.filename = 'subdir/data2.dat'
+        zip_info.date_time = cls.ref_time.timetuple()
+        zip_egg.writestr(zip_info, 'goodbye, world!')
+        zip_egg.close()
+        egg.close()
+
+        sys.path.append(egg.name)
+        subdir = os.path.join(egg.name, 'subdir')
+        sys.path.append(subdir)
+        cls.finalizers.append(EggRemover(subdir))
+        cls.finalizers.append(EggRemover(egg.name))
+
+    @classmethod
+    def teardown_class(cls):
+        for finalizer in cls.finalizers:
+            finalizer()
+
+    def test_resource_listdir(self):
+        import mod  # pyright: ignore[reportMissingImports] # Temporary package for test
+
+        zp = pkg_resources.ZipProvider(mod)
+
+        expected_root = ['data.dat', 'mod.py', 'subdir']
+        assert sorted(zp.resource_listdir('')) == expected_root
+
+        expected_subdir = ['data2.dat', 'mod2.py']
+        assert sorted(zp.resource_listdir('subdir')) == expected_subdir
+        assert sorted(zp.resource_listdir('subdir/')) == expected_subdir
+
+        assert zp.resource_listdir('nonexistent') == []
+        assert zp.resource_listdir('nonexistent/') == []
+
+        import mod2  # pyright: ignore[reportMissingImports] # Temporary package for test
+
+        zp2 = pkg_resources.ZipProvider(mod2)
+
+        assert sorted(zp2.resource_listdir('')) == expected_subdir
+
+        assert zp2.resource_listdir('subdir') == []
+        assert zp2.resource_listdir('subdir/') == []
+
+    def test_resource_filename_rewrites_on_change(self):
+        """
+        If a previous call to get_resource_filename has saved the file, but
+        the file has been subsequently mutated with different file of the
+        same size and modification time, it should not be overwritten on a
+        subsequent call to get_resource_filename.
+        """
+        import mod  # pyright: ignore[reportMissingImports] # Temporary package for test
+
+        manager = pkg_resources.ResourceManager()
+        zp = pkg_resources.ZipProvider(mod)
+        filename = zp.get_resource_filename(manager, 'data.dat')
+        actual = datetime.datetime.fromtimestamp(os.stat(filename).st_mtime)
+        assert actual == self.ref_time
+        f = open(filename, 'w', encoding="utf-8")
+        f.write('hello, world?')
+        f.close()
+        ts = self.ref_time.timestamp()
+        os.utime(filename, (ts, ts))
+        filename = zp.get_resource_filename(manager, 'data.dat')
+        with open(filename, encoding="utf-8") as f:
+            assert f.read() == 'hello, world!'
+        manager.cleanup_resources()
+
+
+class TestResourceManager:
+    def test_get_cache_path(self):
+        mgr = pkg_resources.ResourceManager()
+        path = mgr.get_cache_path('foo')
+        type_ = str(type(path))
+        message = "Unexpected type from get_cache_path: " + type_
+        assert isinstance(path, str), message
+
+    def test_get_cache_path_race(self, tmpdir):
+        # Patch to os.path.isdir to create a race condition
+        def patched_isdir(dirname, unpatched_isdir=pkg_resources.isdir):
+            patched_isdir.dirnames.append(dirname)
+
+            was_dir = unpatched_isdir(dirname)
+            if not was_dir:
+                os.makedirs(dirname)
+            return was_dir
+
+        patched_isdir.dirnames = []
+
+        # Get a cache path with a "race condition"
+        mgr = pkg_resources.ResourceManager()
+        mgr.set_extraction_path(str(tmpdir))
+
+        archive_name = os.sep.join(('foo', 'bar', 'baz'))
+        with mock.patch.object(pkg_resources, 'isdir', new=patched_isdir):
+            mgr.get_cache_path(archive_name)
+
+        # Because this test relies on the implementation details of this
+        # function, these assertions are a sentinel to ensure that the
+        # test suite will not fail silently if the implementation changes.
+        called_dirnames = patched_isdir.dirnames
+        assert len(called_dirnames) == 2
+        assert called_dirnames[0].split(os.sep)[-2:] == ['foo', 'bar']
+        assert called_dirnames[1].split(os.sep)[-1:] == ['foo']
+
+    """
+    Tests to ensure that pkg_resources runs independently from setuptools.
+    """
+
+    def test_setuptools_not_imported(self):
+        """
+        In a separate Python environment, import pkg_resources and assert
+        that action doesn't cause setuptools to be imported.
+        """
+        lines = (
+            'import pkg_resources',
+            'import sys',
+            ('assert "setuptools" not in sys.modules, "setuptools was imported"'),
+        )
+        cmd = [sys.executable, '-c', '; '.join(lines)]
+        subprocess.check_call(cmd)
+
+
+def make_test_distribution(metadata_path, metadata):
+    """
+    Make a test Distribution object, and return it.
+
+    :param metadata_path: the path to the metadata file that should be
+        created. This should be inside a distribution directory that should
+        also be created. For example, an argument value might end with
+        "<project>.dist-info/METADATA".
+    :param metadata: the desired contents of the metadata file, as bytes.
+    """
+    dist_dir = os.path.dirname(metadata_path)
+    os.mkdir(dist_dir)
+    with open(metadata_path, 'wb') as f:
+        f.write(metadata)
+    dists = list(pkg_resources.distributions_from_metadata(dist_dir))
+    (dist,) = dists
+
+    return dist
+
+
+def test_get_metadata__bad_utf8(tmpdir):
+    """
+    Test a metadata file with bytes that can't be decoded as utf-8.
+    """
+    filename = 'METADATA'
+    # Convert the tmpdir LocalPath object to a string before joining.
+    metadata_path = os.path.join(str(tmpdir), 'foo.dist-info', filename)
+    # Encode a non-ascii string with the wrong encoding (not utf-8).
+    metadata = 'née'.encode('iso-8859-1')
+    dist = make_test_distribution(metadata_path, metadata=metadata)
+
+    with pytest.raises(UnicodeDecodeError) as excinfo:
+        dist.get_metadata(filename)
+
+    exc = excinfo.value
+    actual = str(exc)
+    expected = (
+        # The error message starts with "'utf-8' codec ..." However, the
+        # spelling of "utf-8" can vary (e.g. "utf8") so we don't include it
+        "codec can't decode byte 0xe9 in position 1: "
+        'invalid continuation byte in METADATA file at path: '
+    )
+    assert expected in actual, f'actual: {actual}'
+    assert actual.endswith(metadata_path), f'actual: {actual}'
+
+
+def make_distribution_no_version(tmpdir, basename):
+    """
+    Create a distribution directory with no file containing the version.
+    """
+    dist_dir = tmpdir / basename
+    dist_dir.ensure_dir()
+    # Make the directory non-empty so distributions_from_metadata()
+    # will detect it and yield it.
+    dist_dir.join('temp.txt').ensure()
+
+    dists = list(pkg_resources.distributions_from_metadata(dist_dir))
+    assert len(dists) == 1
+    (dist,) = dists
+
+    return dist, dist_dir
+
+
+@pytest.mark.parametrize(
+    ("suffix", "expected_filename", "expected_dist_type"),
+    [
+        ('egg-info', 'PKG-INFO', EggInfoDistribution),
+        ('dist-info', 'METADATA', DistInfoDistribution),
+    ],
+)
+@pytest.mark.xfail(
+    sys.version_info[:2] == (3, 12) and sys.version_info.releaselevel != 'final',
+    reason="https://github.com/python/cpython/issues/103632",
+)
+def test_distribution_version_missing(
+    tmpdir, suffix, expected_filename, expected_dist_type
+):
+    """
+    Test Distribution.version when the "Version" header is missing.
+    """
+    basename = f'foo.{suffix}'
+    dist, dist_dir = make_distribution_no_version(tmpdir, basename)
+
+    expected_text = (
+        f"Missing 'Version:' header and/or {expected_filename} file at path: "
+    )
+    metadata_path = os.path.join(dist_dir, expected_filename)
+
+    # Now check the exception raised when the "version" attribute is accessed.
+    with pytest.raises(ValueError) as excinfo:
+        dist.version
+
+    err = str(excinfo.value)
+    # Include a string expression after the assert so the full strings
+    # will be visible for inspection on failure.
+    assert expected_text in err, str((expected_text, err))
+
+    # Also check the args passed to the ValueError.
+    msg, dist = excinfo.value.args
+    assert expected_text in msg
+    # Check that the message portion contains the path.
+    assert metadata_path in msg, str((metadata_path, msg))
+    assert type(dist) is expected_dist_type
+
+
+@pytest.mark.xfail(
+    sys.version_info[:2] == (3, 12) and sys.version_info.releaselevel != 'final',
+    reason="https://github.com/python/cpython/issues/103632",
+)
+def test_distribution_version_missing_undetected_path():
+    """
+    Test Distribution.version when the "Version" header is missing and
+    the path can't be detected.
+    """
+    # Create a Distribution object with no metadata argument, which results
+    # in an empty metadata provider.
+    dist = Distribution('/foo')
+    with pytest.raises(ValueError) as excinfo:
+        dist.version
+
+    msg, dist = excinfo.value.args
+    expected = (
+        "Missing 'Version:' header and/or PKG-INFO file at path: [could not detect]"
+    )
+    assert msg == expected
+
+
+@pytest.mark.parametrize('only', [False, True])
+def test_dist_info_is_not_dir(tmp_path, only):
+    """Test path containing a file with dist-info extension."""
+    dist_info = tmp_path / 'foobar.dist-info'
+    dist_info.touch()
+    assert not pkg_resources.dist_factory(str(tmp_path), str(dist_info), only)
+
+
+def test_macos_vers_fallback(monkeypatch, tmp_path):
+    """Regression test for pkg_resources._macos_vers"""
+    orig_open = builtins.open
+
+    # Pretend we need to use the plist file
+    monkeypatch.setattr('platform.mac_ver', mock.Mock(return_value=('', (), '')))
+
+    # Create fake content for the fake plist file
+    with open(tmp_path / 'fake.plist', 'wb') as fake_file:
+        plistlib.dump({"ProductVersion": "11.4"}, fake_file)
+
+    # Pretend the fake file exists
+    monkeypatch.setattr('os.path.exists', mock.Mock(return_value=True))
+
+    def fake_open(file, *args, **kwargs):
+        return orig_open(tmp_path / 'fake.plist', *args, **kwargs)
+
+    # Ensure that the _macos_vers works correctly
+    with mock.patch('builtins.open', mock.Mock(side_effect=fake_open)) as m:
+        pkg_resources._macos_vers.cache_clear()
+        assert pkg_resources._macos_vers() == ["11", "4"]
+        pkg_resources._macos_vers.cache_clear()
+
+    m.assert_called()
+
+
+class TestDeepVersionLookupDistutils:
+    @pytest.fixture
+    def env(self, tmpdir):
+        """
+        Create a package environment, similar to a virtualenv,
+        in which packages are installed.
+        """
+
+        class Environment(str):
+            pass
+
+        env = Environment(tmpdir)
+        tmpdir.chmod(stat.S_IRWXU)
+        subs = 'home', 'lib', 'scripts', 'data', 'egg-base'
+        env.paths = dict((dirname, str(tmpdir / dirname)) for dirname in subs)
+        list(map(os.mkdir, env.paths.values()))
+        return env
+
+    def create_foo_pkg(self, env, version):
+        """
+        Create a foo package installed (distutils-style) to env.paths['lib']
+        as version.
+        """
+        ld = "This package has unicode metadata! ❄"
+        attrs = dict(name='foo', version=version, long_description=ld)
+        dist = distutils.dist.Distribution(attrs)
+        iei_cmd = distutils.command.install_egg_info.install_egg_info(dist)
+        iei_cmd.initialize_options()
+        iei_cmd.install_dir = env.paths['lib']
+        iei_cmd.finalize_options()
+        iei_cmd.run()
+
+    def test_version_resolved_from_egg_info(self, env):
+        version = '1.11.0.dev0+2329eae'
+        self.create_foo_pkg(env, version)
+
+        # this requirement parsing will raise a VersionConflict unless the
+        # .egg-info file is parsed (see #419 on BitBucket)
+        req = pkg_resources.Requirement.parse('foo>=1.9')
+        dist = pkg_resources.WorkingSet([env.paths['lib']]).find(req)
+        assert dist.version == version
+
+    @pytest.mark.parametrize(
+        ("unnormalized", "normalized"),
+        [
+            ('foo', 'foo'),
+            ('foo/', 'foo'),
+            ('foo/bar', 'foo/bar'),
+            ('foo/bar/', 'foo/bar'),
+        ],
+    )
+    def test_normalize_path_trailing_sep(self, unnormalized, normalized):
+        """Ensure the trailing slash is cleaned for path comparison.
+
+        See pypa/setuptools#1519.
+        """
+        result_from_unnormalized = pkg_resources.normalize_path(unnormalized)
+        result_from_normalized = pkg_resources.normalize_path(normalized)
+        assert result_from_unnormalized == result_from_normalized
+
+    @pytest.mark.skipif(
+        os.path.normcase('A') != os.path.normcase('a'),
+        reason='Testing case-insensitive filesystems.',
+    )
+    @pytest.mark.parametrize(
+        ("unnormalized", "normalized"),
+        [
+            ('MiXeD/CasE', 'mixed/case'),
+        ],
+    )
+    def test_normalize_path_normcase(self, unnormalized, normalized):
+        """Ensure mixed case is normalized on case-insensitive filesystems."""
+        result_from_unnormalized = pkg_resources.normalize_path(unnormalized)
+        result_from_normalized = pkg_resources.normalize_path(normalized)
+        assert result_from_unnormalized == result_from_normalized
+
+    @pytest.mark.skipif(
+        os.path.sep != '\\',
+        reason='Testing systems using backslashes as path separators.',
+    )
+    @pytest.mark.parametrize(
+        ("unnormalized", "expected"),
+        [
+            ('forward/slash', 'forward\\slash'),
+            ('forward/slash/', 'forward\\slash'),
+            ('backward\\slash\\', 'backward\\slash'),
+        ],
+    )
+    def test_normalize_path_backslash_sep(self, unnormalized, expected):
+        """Ensure path seps are cleaned on backslash path sep systems."""
+        result = pkg_resources.normalize_path(unnormalized)
+        assert result.endswith(expected)
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_resources.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_resources.py
new file mode 100644
index 00000000..70436c08
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_resources.py
@@ -0,0 +1,869 @@
+import itertools
+import os
+import platform
+import string
+import sys
+
+import pytest
+from packaging.specifiers import SpecifierSet
+
+import pkg_resources
+from pkg_resources import (
+    Distribution,
+    EntryPoint,
+    Requirement,
+    VersionConflict,
+    WorkingSet,
+    parse_requirements,
+    parse_version,
+    safe_name,
+    safe_version,
+)
+
+
+# from Python 3.6 docs. Available from itertools on Python 3.10
+def pairwise(iterable):
+    "s -> (s0,s1), (s1,s2), (s2, s3), ..."
+    a, b = itertools.tee(iterable)
+    next(b, None)
+    return zip(a, b)
+
+
+class Metadata(pkg_resources.EmptyProvider):
+    """Mock object to return metadata as if from an on-disk distribution"""
+
+    def __init__(self, *pairs) -> None:
+        self.metadata = dict(pairs)
+
+    def has_metadata(self, name) -> bool:
+        return name in self.metadata
+
+    def get_metadata(self, name):
+        return self.metadata[name]
+
+    def get_metadata_lines(self, name):
+        return pkg_resources.yield_lines(self.get_metadata(name))
+
+
+dist_from_fn = pkg_resources.Distribution.from_filename
+
+
+class TestDistro:
+    def testCollection(self):
+        # empty path should produce no distributions
+        ad = pkg_resources.Environment([], platform=None, python=None)
+        assert list(ad) == []
+        assert ad['FooPkg'] == []
+        ad.add(dist_from_fn("FooPkg-1.3_1.egg"))
+        ad.add(dist_from_fn("FooPkg-1.4-py2.4-win32.egg"))
+        ad.add(dist_from_fn("FooPkg-1.2-py2.4.egg"))
+
+        # Name is in there now
+        assert ad['FooPkg']
+        # But only 1 package
+        assert list(ad) == ['foopkg']
+
+        # Distributions sort by version
+        expected = ['1.4', '1.3-1', '1.2']
+        assert [dist.version for dist in ad['FooPkg']] == expected
+
+        # Removing a distribution leaves sequence alone
+        ad.remove(ad['FooPkg'][1])
+        assert [dist.version for dist in ad['FooPkg']] == ['1.4', '1.2']
+
+        # And inserting adds them in order
+        ad.add(dist_from_fn("FooPkg-1.9.egg"))
+        assert [dist.version for dist in ad['FooPkg']] == ['1.9', '1.4', '1.2']
+
+        ws = WorkingSet([])
+        foo12 = dist_from_fn("FooPkg-1.2-py2.4.egg")
+        foo14 = dist_from_fn("FooPkg-1.4-py2.4-win32.egg")
+        (req,) = parse_requirements("FooPkg>=1.3")
+
+        # Nominal case: no distros on path, should yield all applicable
+        assert ad.best_match(req, ws).version == '1.9'
+        # If a matching distro is already installed, should return only that
+        ws.add(foo14)
+        assert ad.best_match(req, ws).version == '1.4'
+
+        # If the first matching distro is unsuitable, it's a version conflict
+        ws = WorkingSet([])
+        ws.add(foo12)
+        ws.add(foo14)
+        with pytest.raises(VersionConflict):
+            ad.best_match(req, ws)
+
+        # If more than one match on the path, the first one takes precedence
+        ws = WorkingSet([])
+        ws.add(foo14)
+        ws.add(foo12)
+        ws.add(foo14)
+        assert ad.best_match(req, ws).version == '1.4'
+
+    def checkFooPkg(self, d):
+        assert d.project_name == "FooPkg"
+        assert d.key == "foopkg"
+        assert d.version == "1.3.post1"
+        assert d.py_version == "2.4"
+        assert d.platform == "win32"
+        assert d.parsed_version == parse_version("1.3-1")
+
+    def testDistroBasics(self):
+        d = Distribution(
+            "/some/path",
+            project_name="FooPkg",
+            version="1.3-1",
+            py_version="2.4",
+            platform="win32",
+        )
+        self.checkFooPkg(d)
+
+        d = Distribution("/some/path")
+        assert d.py_version == f'{sys.version_info.major}.{sys.version_info.minor}'
+        assert d.platform is None
+
+    def testDistroParse(self):
+        d = dist_from_fn("FooPkg-1.3.post1-py2.4-win32.egg")
+        self.checkFooPkg(d)
+        d = dist_from_fn("FooPkg-1.3.post1-py2.4-win32.egg-info")
+        self.checkFooPkg(d)
+
+    def testDistroMetadata(self):
+        d = Distribution(
+            "/some/path",
+            project_name="FooPkg",
+            py_version="2.4",
+            platform="win32",
+            metadata=Metadata(('PKG-INFO', "Metadata-Version: 1.0\nVersion: 1.3-1\n")),
+        )
+        self.checkFooPkg(d)
+
+    def distRequires(self, txt):
+        return Distribution("/foo", metadata=Metadata(('depends.txt', txt)))
+
+    def checkRequires(self, dist, txt, extras=()):
+        assert list(dist.requires(extras)) == list(parse_requirements(txt))
+
+    def testDistroDependsSimple(self):
+        for v in "Twisted>=1.5", "Twisted>=1.5\nZConfig>=2.0":
+            self.checkRequires(self.distRequires(v), v)
+
+    needs_object_dir = pytest.mark.skipif(
+        not hasattr(object, '__dir__'),
+        reason='object.__dir__ necessary for self.__dir__ implementation',
+    )
+
+    def test_distribution_dir(self):
+        d = pkg_resources.Distribution()
+        dir(d)
+
+    @needs_object_dir
+    def test_distribution_dir_includes_provider_dir(self):
+        d = pkg_resources.Distribution()
+        before = d.__dir__()
+        assert 'test_attr' not in before
+        d._provider.test_attr = None
+        after = d.__dir__()
+        assert len(after) == len(before) + 1
+        assert 'test_attr' in after
+
+    @needs_object_dir
+    def test_distribution_dir_ignores_provider_dir_leading_underscore(self):
+        d = pkg_resources.Distribution()
+        before = d.__dir__()
+        assert '_test_attr' not in before
+        d._provider._test_attr = None
+        after = d.__dir__()
+        assert len(after) == len(before)
+        assert '_test_attr' not in after
+
+    def testResolve(self):
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        # Resolving no requirements -> nothing to install
+        assert list(ws.resolve([], ad)) == []
+        # Request something not in the collection -> DistributionNotFound
+        with pytest.raises(pkg_resources.DistributionNotFound):
+            ws.resolve(parse_requirements("Foo"), ad)
+
+        Foo = Distribution.from_filename(
+            "/foo_dir/Foo-1.2.egg",
+            metadata=Metadata(('depends.txt', "[bar]\nBaz>=2.0")),
+        )
+        ad.add(Foo)
+        ad.add(Distribution.from_filename("Foo-0.9.egg"))
+
+        # Request thing(s) that are available -> list to activate
+        for i in range(3):
+            targets = list(ws.resolve(parse_requirements("Foo"), ad))
+            assert targets == [Foo]
+            list(map(ws.add, targets))
+        with pytest.raises(VersionConflict):
+            ws.resolve(parse_requirements("Foo==0.9"), ad)
+        ws = WorkingSet([])  # reset
+
+        # Request an extra that causes an unresolved dependency for "Baz"
+        with pytest.raises(pkg_resources.DistributionNotFound):
+            ws.resolve(parse_requirements("Foo[bar]"), ad)
+        Baz = Distribution.from_filename(
+            "/foo_dir/Baz-2.1.egg", metadata=Metadata(('depends.txt', "Foo"))
+        )
+        ad.add(Baz)
+
+        # Activation list now includes resolved dependency
+        assert list(ws.resolve(parse_requirements("Foo[bar]"), ad)) == [Foo, Baz]
+        # Requests for conflicting versions produce VersionConflict
+        with pytest.raises(VersionConflict) as vc:
+            ws.resolve(parse_requirements("Foo==1.2\nFoo!=1.2"), ad)
+
+        msg = 'Foo 0.9 is installed but Foo==1.2 is required'
+        assert vc.value.report() == msg
+
+    def test_environment_marker_evaluation_negative(self):
+        """Environment markers are evaluated at resolution time."""
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        res = ws.resolve(parse_requirements("Foo;python_version<'2'"), ad)
+        assert list(res) == []
+
+    def test_environment_marker_evaluation_positive(self):
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        Foo = Distribution.from_filename("/foo_dir/Foo-1.2.dist-info")
+        ad.add(Foo)
+        res = ws.resolve(parse_requirements("Foo;python_version>='2'"), ad)
+        assert list(res) == [Foo]
+
+    def test_environment_marker_evaluation_called(self):
+        """
+        If one package foo requires bar without any extras,
+        markers should pass for bar without extras.
+        """
+        (parent_req,) = parse_requirements("foo")
+        (req,) = parse_requirements("bar;python_version>='2'")
+        req_extras = pkg_resources._ReqExtras({req: parent_req.extras})
+        assert req_extras.markers_pass(req)
+
+        (parent_req,) = parse_requirements("foo[]")
+        (req,) = parse_requirements("bar;python_version>='2'")
+        req_extras = pkg_resources._ReqExtras({req: parent_req.extras})
+        assert req_extras.markers_pass(req)
+
+    def test_marker_evaluation_with_extras(self):
+        """Extras are also evaluated as markers at resolution time."""
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        Foo = Distribution.from_filename(
+            "/foo_dir/Foo-1.2.dist-info",
+            metadata=Metadata((
+                "METADATA",
+                "Provides-Extra: baz\nRequires-Dist: quux; extra=='baz'",
+            )),
+        )
+        ad.add(Foo)
+        assert list(ws.resolve(parse_requirements("Foo"), ad)) == [Foo]
+        quux = Distribution.from_filename("/foo_dir/quux-1.0.dist-info")
+        ad.add(quux)
+        res = list(ws.resolve(parse_requirements("Foo[baz]"), ad))
+        assert res == [Foo, quux]
+
+    def test_marker_evaluation_with_extras_normlized(self):
+        """Extras are also evaluated as markers at resolution time."""
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        Foo = Distribution.from_filename(
+            "/foo_dir/Foo-1.2.dist-info",
+            metadata=Metadata((
+                "METADATA",
+                "Provides-Extra: baz-lightyear\n"
+                "Requires-Dist: quux; extra=='baz-lightyear'",
+            )),
+        )
+        ad.add(Foo)
+        assert list(ws.resolve(parse_requirements("Foo"), ad)) == [Foo]
+        quux = Distribution.from_filename("/foo_dir/quux-1.0.dist-info")
+        ad.add(quux)
+        res = list(ws.resolve(parse_requirements("Foo[baz-lightyear]"), ad))
+        assert res == [Foo, quux]
+
+    def test_marker_evaluation_with_multiple_extras(self):
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        Foo = Distribution.from_filename(
+            "/foo_dir/Foo-1.2.dist-info",
+            metadata=Metadata((
+                "METADATA",
+                "Provides-Extra: baz\n"
+                "Requires-Dist: quux; extra=='baz'\n"
+                "Provides-Extra: bar\n"
+                "Requires-Dist: fred; extra=='bar'\n",
+            )),
+        )
+        ad.add(Foo)
+        quux = Distribution.from_filename("/foo_dir/quux-1.0.dist-info")
+        ad.add(quux)
+        fred = Distribution.from_filename("/foo_dir/fred-0.1.dist-info")
+        ad.add(fred)
+        res = list(ws.resolve(parse_requirements("Foo[baz,bar]"), ad))
+        assert sorted(res) == [fred, quux, Foo]
+
+    def test_marker_evaluation_with_extras_loop(self):
+        ad = pkg_resources.Environment([])
+        ws = WorkingSet([])
+        a = Distribution.from_filename(
+            "/foo_dir/a-0.2.dist-info",
+            metadata=Metadata(("METADATA", "Requires-Dist: c[a]")),
+        )
+        b = Distribution.from_filename(
+            "/foo_dir/b-0.3.dist-info",
+            metadata=Metadata(("METADATA", "Requires-Dist: c[b]")),
+        )
+        c = Distribution.from_filename(
+            "/foo_dir/c-1.0.dist-info",
+            metadata=Metadata((
+                "METADATA",
+                "Provides-Extra: a\n"
+                "Requires-Dist: b;extra=='a'\n"
+                "Provides-Extra: b\n"
+                "Requires-Dist: foo;extra=='b'",
+            )),
+        )
+        foo = Distribution.from_filename("/foo_dir/foo-0.1.dist-info")
+        for dist in (a, b, c, foo):
+            ad.add(dist)
+        res = list(ws.resolve(parse_requirements("a"), ad))
+        assert res == [a, c, b, foo]
+
+    @pytest.mark.xfail(
+        sys.version_info[:2] == (3, 12) and sys.version_info.releaselevel != 'final',
+        reason="https://github.com/python/cpython/issues/103632",
+    )
+    def testDistroDependsOptions(self):
+        d = self.distRequires(
+            """
+            Twisted>=1.5
+            [docgen]
+            ZConfig>=2.0
+            docutils>=0.3
+            [fastcgi]
+            fcgiapp>=0.1"""
+        )
+        self.checkRequires(d, "Twisted>=1.5")
+        self.checkRequires(
+            d, "Twisted>=1.5 ZConfig>=2.0 docutils>=0.3".split(), ["docgen"]
+        )
+        self.checkRequires(d, "Twisted>=1.5 fcgiapp>=0.1".split(), ["fastcgi"])
+        self.checkRequires(
+            d,
+            "Twisted>=1.5 ZConfig>=2.0 docutils>=0.3 fcgiapp>=0.1".split(),
+            ["docgen", "fastcgi"],
+        )
+        self.checkRequires(
+            d,
+            "Twisted>=1.5 fcgiapp>=0.1 ZConfig>=2.0 docutils>=0.3".split(),
+            ["fastcgi", "docgen"],
+        )
+        with pytest.raises(pkg_resources.UnknownExtra):
+            d.requires(["foo"])
+
+
+class TestWorkingSet:
+    def test_find_conflicting(self):
+        ws = WorkingSet([])
+        Foo = Distribution.from_filename("/foo_dir/Foo-1.2.egg")
+        ws.add(Foo)
+
+        # create a requirement that conflicts with Foo 1.2
+        req = next(parse_requirements("Foo<1.2"))
+
+        with pytest.raises(VersionConflict) as vc:
+            ws.find(req)
+
+        msg = 'Foo 1.2 is installed but Foo<1.2 is required'
+        assert vc.value.report() == msg
+
+    def test_resolve_conflicts_with_prior(self):
+        """
+        A ContextualVersionConflict should be raised when a requirement
+        conflicts with a prior requirement for a different package.
+        """
+        # Create installation where Foo depends on Baz 1.0 and Bar depends on
+        # Baz 2.0.
+        ws = WorkingSet([])
+        md = Metadata(('depends.txt', "Baz==1.0"))
+        Foo = Distribution.from_filename("/foo_dir/Foo-1.0.egg", metadata=md)
+        ws.add(Foo)
+        md = Metadata(('depends.txt', "Baz==2.0"))
+        Bar = Distribution.from_filename("/foo_dir/Bar-1.0.egg", metadata=md)
+        ws.add(Bar)
+        Baz = Distribution.from_filename("/foo_dir/Baz-1.0.egg")
+        ws.add(Baz)
+        Baz = Distribution.from_filename("/foo_dir/Baz-2.0.egg")
+        ws.add(Baz)
+
+        with pytest.raises(VersionConflict) as vc:
+            ws.resolve(parse_requirements("Foo\nBar\n"))
+
+        msg = "Baz 1.0 is installed but Baz==2.0 is required by "
+        msg += repr(set(['Bar']))
+        assert vc.value.report() == msg
+
+
+class TestEntryPoints:
+    def assertfields(self, ep):
+        assert ep.name == "foo"
+        assert ep.module_name == "pkg_resources.tests.test_resources"
+        assert ep.attrs == ("TestEntryPoints",)
+        assert ep.extras == ("x",)
+        assert ep.load() is TestEntryPoints
+        expect = "foo = pkg_resources.tests.test_resources:TestEntryPoints [x]"
+        assert str(ep) == expect
+
+    def setup_method(self, method):
+        self.dist = Distribution.from_filename(
+            "FooPkg-1.2-py2.4.egg", metadata=Metadata(('requires.txt', '[x]'))
+        )
+
+    def testBasics(self):
+        ep = EntryPoint(
+            "foo",
+            "pkg_resources.tests.test_resources",
+            ["TestEntryPoints"],
+            ["x"],
+            self.dist,
+        )
+        self.assertfields(ep)
+
+    def testParse(self):
+        s = "foo = pkg_resources.tests.test_resources:TestEntryPoints [x]"
+        ep = EntryPoint.parse(s, self.dist)
+        self.assertfields(ep)
+
+        ep = EntryPoint.parse("bar baz=  spammity[PING]")
+        assert ep.name == "bar baz"
+        assert ep.module_name == "spammity"
+        assert ep.attrs == ()
+        assert ep.extras == ("ping",)
+
+        ep = EntryPoint.parse(" fizzly =  wocka:foo")
+        assert ep.name == "fizzly"
+        assert ep.module_name == "wocka"
+        assert ep.attrs == ("foo",)
+        assert ep.extras == ()
+
+        # plus in the name
+        spec = "html+mako = mako.ext.pygmentplugin:MakoHtmlLexer"
+        ep = EntryPoint.parse(spec)
+        assert ep.name == 'html+mako'
+
+    reject_specs = "foo", "x=a:b:c", "q=x/na", "fez=pish:tush-z", "x=f[a]>2"
+
+    @pytest.mark.parametrize("reject_spec", reject_specs)
+    def test_reject_spec(self, reject_spec):
+        with pytest.raises(ValueError):
+            EntryPoint.parse(reject_spec)
+
+    def test_printable_name(self):
+        """
+        Allow any printable character in the name.
+        """
+        # Create a name with all printable characters; strip the whitespace.
+        name = string.printable.strip()
+        spec = "{name} = module:attr".format(**locals())
+        ep = EntryPoint.parse(spec)
+        assert ep.name == name
+
+    def checkSubMap(self, m):
+        assert len(m) == len(self.submap_expect)
+        for key, ep in self.submap_expect.items():
+            assert m.get(key).name == ep.name
+            assert m.get(key).module_name == ep.module_name
+            assert sorted(m.get(key).attrs) == sorted(ep.attrs)
+            assert sorted(m.get(key).extras) == sorted(ep.extras)
+
+    submap_expect = dict(
+        feature1=EntryPoint('feature1', 'somemodule', ['somefunction']),
+        feature2=EntryPoint(
+            'feature2', 'another.module', ['SomeClass'], ['extra1', 'extra2']
+        ),
+        feature3=EntryPoint('feature3', 'this.module', extras=['something']),
+    )
+    submap_str = """
+            # define features for blah blah
+            feature1 = somemodule:somefunction
+            feature2 = another.module:SomeClass [extra1,extra2]
+            feature3 = this.module [something]
+    """
+
+    def testParseList(self):
+        self.checkSubMap(EntryPoint.parse_group("xyz", self.submap_str))
+        with pytest.raises(ValueError):
+            EntryPoint.parse_group("x a", "foo=bar")
+        with pytest.raises(ValueError):
+            EntryPoint.parse_group("x", ["foo=baz", "foo=bar"])
+
+    def testParseMap(self):
+        m = EntryPoint.parse_map({'xyz': self.submap_str})
+        self.checkSubMap(m['xyz'])
+        assert list(m.keys()) == ['xyz']
+        m = EntryPoint.parse_map("[xyz]\n" + self.submap_str)
+        self.checkSubMap(m['xyz'])
+        assert list(m.keys()) == ['xyz']
+        with pytest.raises(ValueError):
+            EntryPoint.parse_map(["[xyz]", "[xyz]"])
+        with pytest.raises(ValueError):
+            EntryPoint.parse_map(self.submap_str)
+
+    def testDeprecationWarnings(self):
+        ep = EntryPoint(
+            "foo", "pkg_resources.tests.test_resources", ["TestEntryPoints"], ["x"]
+        )
+        with pytest.warns(pkg_resources.PkgResourcesDeprecationWarning):
+            ep.load(require=False)
+
+
+class TestRequirements:
+    def testBasics(self):
+        r = Requirement.parse("Twisted>=1.2")
+        assert str(r) == "Twisted>=1.2"
+        assert repr(r) == "Requirement.parse('Twisted>=1.2')"
+        assert r == Requirement("Twisted>=1.2")
+        assert r == Requirement("twisTed>=1.2")
+        assert r != Requirement("Twisted>=2.0")
+        assert r != Requirement("Zope>=1.2")
+        assert r != Requirement("Zope>=3.0")
+        assert r != Requirement("Twisted[extras]>=1.2")
+
+    def testOrdering(self):
+        r1 = Requirement("Twisted==1.2c1,>=1.2")
+        r2 = Requirement("Twisted>=1.2,==1.2c1")
+        assert r1 == r2
+        assert str(r1) == str(r2)
+        assert str(r2) == "Twisted==1.2c1,>=1.2"
+        assert Requirement("Twisted") != Requirement(
+            "Twisted @ https://localhost/twisted.zip"
+        )
+
+    def testBasicContains(self):
+        r = Requirement("Twisted>=1.2")
+        foo_dist = Distribution.from_filename("FooPkg-1.3_1.egg")
+        twist11 = Distribution.from_filename("Twisted-1.1.egg")
+        twist12 = Distribution.from_filename("Twisted-1.2.egg")
+        assert parse_version('1.2') in r
+        assert parse_version('1.1') not in r
+        assert '1.2' in r
+        assert '1.1' not in r
+        assert foo_dist not in r
+        assert twist11 not in r
+        assert twist12 in r
+
+    def testOptionsAndHashing(self):
+        r1 = Requirement.parse("Twisted[foo,bar]>=1.2")
+        r2 = Requirement.parse("Twisted[bar,FOO]>=1.2")
+        assert r1 == r2
+        assert set(r1.extras) == set(("foo", "bar"))
+        assert set(r2.extras) == set(("foo", "bar"))
+        assert hash(r1) == hash(r2)
+        assert hash(r1) == hash((
+            "twisted",
+            None,
+            SpecifierSet(">=1.2"),
+            frozenset(["foo", "bar"]),
+            None,
+        ))
+        assert hash(
+            Requirement.parse("Twisted @ https://localhost/twisted.zip")
+        ) == hash((
+            "twisted",
+            "https://localhost/twisted.zip",
+            SpecifierSet(),
+            frozenset(),
+            None,
+        ))
+
+    def testVersionEquality(self):
+        r1 = Requirement.parse("foo==0.3a2")
+        r2 = Requirement.parse("foo!=0.3a4")
+        d = Distribution.from_filename
+
+        assert d("foo-0.3a4.egg") not in r1
+        assert d("foo-0.3a1.egg") not in r1
+        assert d("foo-0.3a4.egg") not in r2
+
+        assert d("foo-0.3a2.egg") in r1
+        assert d("foo-0.3a2.egg") in r2
+        assert d("foo-0.3a3.egg") in r2
+        assert d("foo-0.3a5.egg") in r2
+
+    def testSetuptoolsProjectName(self):
+        """
+        The setuptools project should implement the setuptools package.
+        """
+
+        assert Requirement.parse('setuptools').project_name == 'setuptools'
+        # setuptools 0.7 and higher means setuptools.
+        assert Requirement.parse('setuptools == 0.7').project_name == 'setuptools'
+        assert Requirement.parse('setuptools == 0.7a1').project_name == 'setuptools'
+        assert Requirement.parse('setuptools >= 0.7').project_name == 'setuptools'
+
+
+class TestParsing:
+    def testEmptyParse(self):
+        assert list(parse_requirements('')) == []
+
+    def testYielding(self):
+        for inp, out in [
+            ([], []),
+            ('x', ['x']),
+            ([[]], []),
+            (' x\n y', ['x', 'y']),
+            (['x\n\n', 'y'], ['x', 'y']),
+        ]:
+            assert list(pkg_resources.yield_lines(inp)) == out
+
+    def testSplitting(self):
+        sample = """
+                    x
+                    [Y]
+                    z
+
+                    a
+                    [b ]
+                    # foo
+                    c
+                    [ d]
+                    [q]
+                    v
+                    """
+        assert list(pkg_resources.split_sections(sample)) == [
+            (None, ["x"]),
+            ("Y", ["z", "a"]),
+            ("b", ["c"]),
+            ("d", []),
+            ("q", ["v"]),
+        ]
+        with pytest.raises(ValueError):
+            list(pkg_resources.split_sections("[foo"))
+
+    def testSafeName(self):
+        assert safe_name("adns-python") == "adns-python"
+        assert safe_name("WSGI Utils") == "WSGI-Utils"
+        assert safe_name("WSGI  Utils") == "WSGI-Utils"
+        assert safe_name("Money$$$Maker") == "Money-Maker"
+        assert safe_name("peak.web") != "peak-web"
+
+    def testSafeVersion(self):
+        assert safe_version("1.2-1") == "1.2.post1"
+        assert safe_version("1.2 alpha") == "1.2.alpha"
+        assert safe_version("2.3.4 20050521") == "2.3.4.20050521"
+        assert safe_version("Money$$$Maker") == "Money-Maker"
+        assert safe_version("peak.web") == "peak.web"
+
+    def testSimpleRequirements(self):
+        assert list(parse_requirements('Twis-Ted>=1.2-1')) == [
+            Requirement('Twis-Ted>=1.2-1')
+        ]
+        assert list(parse_requirements('Twisted >=1.2, \\ # more\n<2.0')) == [
+            Requirement('Twisted>=1.2,<2.0')
+        ]
+        assert Requirement.parse("FooBar==1.99a3") == Requirement("FooBar==1.99a3")
+        with pytest.raises(ValueError):
+            Requirement.parse(">=2.3")
+        with pytest.raises(ValueError):
+            Requirement.parse("x\\")
+        with pytest.raises(ValueError):
+            Requirement.parse("x==2 q")
+        with pytest.raises(ValueError):
+            Requirement.parse("X==1\nY==2")
+        with pytest.raises(ValueError):
+            Requirement.parse("#")
+
+    def test_requirements_with_markers(self):
+        assert Requirement.parse("foobar;os_name=='a'") == Requirement.parse(
+            "foobar;os_name=='a'"
+        )
+        assert Requirement.parse(
+            "name==1.1;python_version=='2.7'"
+        ) != Requirement.parse("name==1.1;python_version=='3.6'")
+        assert Requirement.parse(
+            "name==1.0;python_version=='2.7'"
+        ) != Requirement.parse("name==1.2;python_version=='2.7'")
+        assert Requirement.parse(
+            "name[foo]==1.0;python_version=='3.6'"
+        ) != Requirement.parse("name[foo,bar]==1.0;python_version=='3.6'")
+
+    def test_local_version(self):
+        parse_requirements('foo==1.0+org1')
+
+    def test_spaces_between_multiple_versions(self):
+        parse_requirements('foo>=1.0, <3')
+        parse_requirements('foo >= 1.0, < 3')
+
+    @pytest.mark.parametrize(
+        ("lower", "upper"),
+        [
+            ('1.2-rc1', '1.2rc1'),
+            ('0.4', '0.4.0'),
+            ('0.4.0.0', '0.4.0'),
+            ('0.4.0-0', '0.4-0'),
+            ('0post1', '0.0post1'),
+            ('0pre1', '0.0c1'),
+            ('0.0.0preview1', '0c1'),
+            ('0.0c1', '0-rc1'),
+            ('1.2a1', '1.2.a.1'),
+            ('1.2.a', '1.2a'),
+        ],
+    )
+    def testVersionEquality(self, lower, upper):
+        assert parse_version(lower) == parse_version(upper)
+
+    torture = """
+        0.80.1-3 0.80.1-2 0.80.1-1 0.79.9999+0.80.0pre4-1
+        0.79.9999+0.80.0pre2-3 0.79.9999+0.80.0pre2-2
+        0.77.2-1 0.77.1-1 0.77.0-1
+        """
+
+    @pytest.mark.parametrize(
+        ("lower", "upper"),
+        [
+            ('2.1', '2.1.1'),
+            ('2a1', '2b0'),
+            ('2a1', '2.1'),
+            ('2.3a1', '2.3'),
+            ('2.1-1', '2.1-2'),
+            ('2.1-1', '2.1.1'),
+            ('2.1', '2.1post4'),
+            ('2.1a0-20040501', '2.1'),
+            ('1.1', '02.1'),
+            ('3.2', '3.2.post0'),
+            ('3.2post1', '3.2post2'),
+            ('0.4', '4.0'),
+            ('0.0.4', '0.4.0'),
+            ('0post1', '0.4post1'),
+            ('2.1.0-rc1', '2.1.0'),
+            ('2.1dev', '2.1a0'),
+        ]
+        + list(pairwise(reversed(torture.split()))),
+    )
+    def testVersionOrdering(self, lower, upper):
+        assert parse_version(lower) < parse_version(upper)
+
+    def testVersionHashable(self):
+        """
+        Ensure that our versions stay hashable even though we've subclassed
+        them and added some shim code to them.
+        """
+        assert hash(parse_version("1.0")) == hash(parse_version("1.0"))
+
+
+class TestNamespaces:
+    ns_str = "__import__('pkg_resources').declare_namespace(__name__)\n"
+
+    @pytest.fixture
+    def symlinked_tmpdir(self, tmpdir):
+        """
+        Where available, return the tempdir as a symlink,
+        which as revealed in #231 is more fragile than
+        a natural tempdir.
+        """
+        if not hasattr(os, 'symlink'):
+            yield str(tmpdir)
+            return
+
+        link_name = str(tmpdir) + '-linked'
+        os.symlink(str(tmpdir), link_name)
+        try:
+            yield type(tmpdir)(link_name)
+        finally:
+            os.unlink(link_name)
+
+    @pytest.fixture(autouse=True)
+    def patched_path(self, tmpdir):
+        """
+        Patch sys.path to include the 'site-pkgs' dir. Also
+        restore pkg_resources._namespace_packages to its
+        former state.
+        """
+        saved_ns_pkgs = pkg_resources._namespace_packages.copy()
+        saved_sys_path = sys.path[:]
+        site_pkgs = tmpdir.mkdir('site-pkgs')
+        sys.path.append(str(site_pkgs))
+        try:
+            yield
+        finally:
+            pkg_resources._namespace_packages = saved_ns_pkgs
+            sys.path = saved_sys_path
+
+    issue591 = pytest.mark.xfail(platform.system() == 'Windows', reason="#591")
+
+    @issue591
+    def test_two_levels_deep(self, symlinked_tmpdir):
+        """
+        Test nested namespace packages
+        Create namespace packages in the following tree :
+            site-packages-1/pkg1/pkg2
+            site-packages-2/pkg1/pkg2
+        Check both are in the _namespace_packages dict and that their __path__
+        is correct
+        """
+        real_tmpdir = symlinked_tmpdir.realpath()
+        tmpdir = symlinked_tmpdir
+        sys.path.append(str(tmpdir / 'site-pkgs2'))
+        site_dirs = tmpdir / 'site-pkgs', tmpdir / 'site-pkgs2'
+        for site in site_dirs:
+            pkg1 = site / 'pkg1'
+            pkg2 = pkg1 / 'pkg2'
+            pkg2.ensure_dir()
+            (pkg1 / '__init__.py').write_text(self.ns_str, encoding='utf-8')
+            (pkg2 / '__init__.py').write_text(self.ns_str, encoding='utf-8')
+        with pytest.warns(DeprecationWarning, match="pkg_resources.declare_namespace"):
+            import pkg1  # pyright: ignore[reportMissingImports] # Temporary package for test
+        assert "pkg1" in pkg_resources._namespace_packages
+        # attempt to import pkg2 from site-pkgs2
+        with pytest.warns(DeprecationWarning, match="pkg_resources.declare_namespace"):
+            import pkg1.pkg2  # pyright: ignore[reportMissingImports] # Temporary package for test
+        # check the _namespace_packages dict
+        assert "pkg1.pkg2" in pkg_resources._namespace_packages
+        assert pkg_resources._namespace_packages["pkg1"] == ["pkg1.pkg2"]
+        # check the __path__ attribute contains both paths
+        expected = [
+            str(real_tmpdir / "site-pkgs" / "pkg1" / "pkg2"),
+            str(real_tmpdir / "site-pkgs2" / "pkg1" / "pkg2"),
+        ]
+        assert pkg1.pkg2.__path__ == expected
+
+    @issue591
+    def test_path_order(self, symlinked_tmpdir):
+        """
+        Test that if multiple versions of the same namespace package subpackage
+        are on different sys.path entries, that only the one earliest on
+        sys.path is imported, and that the namespace package's __path__ is in
+        the correct order.
+
+        Regression test for https://github.com/pypa/setuptools/issues/207
+        """
+
+        tmpdir = symlinked_tmpdir
+        site_dirs = (
+            tmpdir / "site-pkgs",
+            tmpdir / "site-pkgs2",
+            tmpdir / "site-pkgs3",
+        )
+
+        vers_str = "__version__ = %r"
+
+        for number, site in enumerate(site_dirs, 1):
+            if number > 1:
+                sys.path.append(str(site))
+            nspkg = site / 'nspkg'
+            subpkg = nspkg / 'subpkg'
+            subpkg.ensure_dir()
+            (nspkg / '__init__.py').write_text(self.ns_str, encoding='utf-8')
+            (subpkg / '__init__.py').write_text(vers_str % number, encoding='utf-8')
+
+        with pytest.warns(DeprecationWarning, match="pkg_resources.declare_namespace"):
+            import nspkg  # pyright: ignore[reportMissingImports] # Temporary package for test
+            import nspkg.subpkg  # pyright: ignore[reportMissingImports] # Temporary package for test
+        expected = [str(site.realpath() / 'nspkg') for site in site_dirs]
+        assert nspkg.__path__ == expected
+        assert nspkg.subpkg.__version__ == 1
diff --git a/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_working_set.py b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_working_set.py
new file mode 100644
index 00000000..7bb84952
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pkg_resources/tests/test_working_set.py
@@ -0,0 +1,501 @@
+import functools
+import inspect
+import re
+import textwrap
+
+import pytest
+
+import pkg_resources
+
+from .test_resources import Metadata
+
+
+def strip_comments(s):
+    return '\n'.join(
+        line
+        for line in s.split('\n')
+        if line.strip() and not line.strip().startswith('#')
+    )
+
+
+def parse_distributions(s):
+    """
+    Parse a series of distribution specs of the form:
+    {project_name}-{version}
+       [optional, indented requirements specification]
+
+    Example:
+
+        foo-0.2
+        bar-1.0
+          foo>=3.0
+          [feature]
+          baz
+
+    yield 2 distributions:
+        - project_name=foo, version=0.2
+        - project_name=bar, version=1.0,
+          requires=['foo>=3.0', 'baz; extra=="feature"']
+    """
+    s = s.strip()
+    for spec in re.split(r'\n(?=[^\s])', s):
+        if not spec:
+            continue
+        fields = spec.split('\n', 1)
+        assert 1 <= len(fields) <= 2
+        name, version = fields.pop(0).rsplit('-', 1)
+        if fields:
+            requires = textwrap.dedent(fields.pop(0))
+            metadata = Metadata(('requires.txt', requires))
+        else:
+            metadata = None
+        dist = pkg_resources.Distribution(
+            project_name=name, version=version, metadata=metadata
+        )
+        yield dist
+
+
+class FakeInstaller:
+    def __init__(self, installable_dists) -> None:
+        self._installable_dists = installable_dists
+
+    def __call__(self, req):
+        return next(
+            iter(filter(lambda dist: dist in req, self._installable_dists)), None
+        )
+
+
+def parametrize_test_working_set_resolve(*test_list):
+    idlist = []
+    argvalues = []
+    for test in test_list:
+        (
+            name,
+            installed_dists,
+            installable_dists,
+            requirements,
+            expected1,
+            expected2,
+        ) = (
+            strip_comments(s.lstrip())
+            for s in textwrap.dedent(test).lstrip().split('\n\n', 5)
+        )
+        installed_dists = list(parse_distributions(installed_dists))
+        installable_dists = list(parse_distributions(installable_dists))
+        requirements = list(pkg_resources.parse_requirements(requirements))
+        for id_, replace_conflicting, expected in (
+            (name, False, expected1),
+            (name + '_replace_conflicting', True, expected2),
+        ):
+            idlist.append(id_)
+            expected = strip_comments(expected.strip())
+            if re.match(r'\w+$', expected):
+                expected = getattr(pkg_resources, expected)
+                assert issubclass(expected, Exception)
+            else:
+                expected = list(parse_distributions(expected))
+            argvalues.append(
+                pytest.param(
+                    installed_dists,
+                    installable_dists,
+                    requirements,
+                    replace_conflicting,
+                    expected,
+                )
+            )
+    return pytest.mark.parametrize(
+        'installed_dists,installable_dists,'
+        'requirements,replace_conflicting,'
+        'resolved_dists_or_exception',
+        argvalues,
+        ids=idlist,
+    )
+
+
+@parametrize_test_working_set_resolve(
+    """
+    # id
+    noop
+
+    # installed
+
+    # installable
+
+    # wanted
+
+    # resolved
+
+    # resolved [replace conflicting]
+    """,
+    """
+    # id
+    already_installed
+
+    # installed
+    foo-3.0
+
+    # installable
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    foo-3.0
+
+    # resolved [replace conflicting]
+    foo-3.0
+    """,
+    """
+    # id
+    installable_not_installed
+
+    # installed
+
+    # installable
+    foo-3.0
+    foo-4.0
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    foo-3.0
+
+    # resolved [replace conflicting]
+    foo-3.0
+    """,
+    """
+    # id
+    not_installable
+
+    # installed
+
+    # installable
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    DistributionNotFound
+
+    # resolved [replace conflicting]
+    DistributionNotFound
+    """,
+    """
+    # id
+    no_matching_version
+
+    # installed
+
+    # installable
+    foo-3.1
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    DistributionNotFound
+
+    # resolved [replace conflicting]
+    DistributionNotFound
+    """,
+    """
+    # id
+    installable_with_installed_conflict
+
+    # installed
+    foo-3.1
+
+    # installable
+    foo-3.5
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    foo-3.5
+    """,
+    """
+    # id
+    not_installable_with_installed_conflict
+
+    # installed
+    foo-3.1
+
+    # installable
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    DistributionNotFound
+    """,
+    """
+    # id
+    installed_with_installed_require
+
+    # installed
+    foo-3.9
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # installable
+
+    # wanted
+    baz
+
+    # resolved
+    foo-3.9
+    baz-0.1
+
+    # resolved [replace conflicting]
+    foo-3.9
+    baz-0.1
+    """,
+    """
+    # id
+    installed_with_conflicting_installed_require
+
+    # installed
+    foo-5
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # installable
+
+    # wanted
+    baz
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    DistributionNotFound
+    """,
+    """
+    # id
+    installed_with_installable_conflicting_require
+
+    # installed
+    foo-5
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # installable
+    foo-2.9
+
+    # wanted
+    baz
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    baz-0.1
+    foo-2.9
+    """,
+    """
+    # id
+    installed_with_installable_require
+
+    # installed
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # installable
+    foo-3.9
+
+    # wanted
+    baz
+
+    # resolved
+    foo-3.9
+    baz-0.1
+
+    # resolved [replace conflicting]
+    foo-3.9
+    baz-0.1
+    """,
+    """
+    # id
+    installable_with_installed_require
+
+    # installed
+    foo-3.9
+
+    # installable
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # wanted
+    baz
+
+    # resolved
+    foo-3.9
+    baz-0.1
+
+    # resolved [replace conflicting]
+    foo-3.9
+    baz-0.1
+    """,
+    """
+    # id
+    installable_with_installable_require
+
+    # installed
+
+    # installable
+    foo-3.9
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # wanted
+    baz
+
+    # resolved
+    foo-3.9
+    baz-0.1
+
+    # resolved [replace conflicting]
+    foo-3.9
+    baz-0.1
+    """,
+    """
+    # id
+    installable_with_conflicting_installable_require
+
+    # installed
+    foo-5
+
+    # installable
+    foo-2.9
+    baz-0.1
+        foo>=2.1,!=3.1,<4
+
+    # wanted
+    baz
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    baz-0.1
+    foo-2.9
+    """,
+    """
+    # id
+    conflicting_installables
+
+    # installed
+
+    # installable
+    foo-2.9
+    foo-5.0
+
+    # wanted
+    foo>=2.1,!=3.1,<4
+    foo>=4
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    VersionConflict
+    """,
+    """
+    # id
+    installables_with_conflicting_requires
+
+    # installed
+
+    # installable
+    foo-2.9
+        dep==1.0
+    baz-5.0
+        dep==2.0
+    dep-1.0
+    dep-2.0
+
+    # wanted
+    foo
+    baz
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    VersionConflict
+    """,
+    """
+    # id
+    installables_with_conflicting_nested_requires
+
+    # installed
+
+    # installable
+    foo-2.9
+        dep1
+    dep1-1.0
+        subdep<1.0
+    baz-5.0
+        dep2
+    dep2-1.0
+        subdep>1.0
+    subdep-0.9
+    subdep-1.1
+
+    # wanted
+    foo
+    baz
+
+    # resolved
+    VersionConflict
+
+    # resolved [replace conflicting]
+    VersionConflict
+    """,
+    """
+    # id
+    wanted_normalized_name_installed_canonical
+
+    # installed
+    foo.bar-3.6
+
+    # installable
+
+    # wanted
+    foo-bar==3.6
+
+    # resolved
+    foo.bar-3.6
+
+    # resolved [replace conflicting]
+    foo.bar-3.6
+    """,
+)
+def test_working_set_resolve(
+    installed_dists,
+    installable_dists,
+    requirements,
+    replace_conflicting,
+    resolved_dists_or_exception,
+):
+    ws = pkg_resources.WorkingSet([])
+    list(map(ws.add, installed_dists))
+    resolve_call = functools.partial(
+        ws.resolve,
+        requirements,
+        installer=FakeInstaller(installable_dists),
+        replace_conflicting=replace_conflicting,
+    )
+    if inspect.isclass(resolved_dists_or_exception):
+        with pytest.raises(resolved_dists_or_exception):
+            resolve_call()
+    else:
+        assert sorted(resolve_call()) == sorted(resolved_dists_or_exception)
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/LICENSE
new file mode 100644
index 00000000..85f4dd63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/LICENSE
@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2015 holger krekel (rather uses bitbucket/hpk42)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/METADATA
new file mode 100644
index 00000000..2d697b0d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/METADATA
@@ -0,0 +1,155 @@
+Metadata-Version: 2.1
+Name: pluggy
+Version: 1.5.0
+Summary: plugin and hook calling mechanisms for python
+Home-page: https://github.com/pytest-dev/pluggy
+Author: Holger Krekel
+Author-email: holger@merlinux.eu
+License: MIT
+Platform: unix
+Platform: linux
+Platform: osx
+Platform: win32
+Classifier: Development Status :: 6 - Mature
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: POSIX
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Topic :: Software Development :: Testing
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Utilities
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Provides-Extra: dev
+Requires-Dist: pre-commit ; extra == 'dev'
+Requires-Dist: tox ; extra == 'dev'
+Provides-Extra: testing
+Requires-Dist: pytest ; extra == 'testing'
+Requires-Dist: pytest-benchmark ; extra == 'testing'
+
+====================================================
+pluggy - A minimalist production ready plugin system
+====================================================
+
+|pypi| |conda-forge| |versions| |github-actions| |gitter| |black| |codecov|
+
+This is the core framework used by the `pytest`_, `tox`_, and `devpi`_ projects.
+
+Please `read the docs`_ to learn more!
+
+A definitive example
+====================
+.. code-block:: python
+
+    import pluggy
+
+    hookspec = pluggy.HookspecMarker("myproject")
+    hookimpl = pluggy.HookimplMarker("myproject")
+
+
+    class MySpec:
+        """A hook specification namespace."""
+
+        @hookspec
+        def myhook(self, arg1, arg2):
+            """My special little hook that you can customize."""
+
+
+    class Plugin_1:
+        """A hook implementation namespace."""
+
+        @hookimpl
+        def myhook(self, arg1, arg2):
+            print("inside Plugin_1.myhook()")
+            return arg1 + arg2
+
+
+    class Plugin_2:
+        """A 2nd hook implementation namespace."""
+
+        @hookimpl
+        def myhook(self, arg1, arg2):
+            print("inside Plugin_2.myhook()")
+            return arg1 - arg2
+
+
+    # create a manager and add the spec
+    pm = pluggy.PluginManager("myproject")
+    pm.add_hookspecs(MySpec)
+
+    # register plugins
+    pm.register(Plugin_1())
+    pm.register(Plugin_2())
+
+    # call our ``myhook`` hook
+    results = pm.hook.myhook(arg1=1, arg2=2)
+    print(results)
+
+
+Running this directly gets us::
+
+    $ python docs/examples/toy-example.py
+    inside Plugin_2.myhook()
+    inside Plugin_1.myhook()
+    [-1, 3]
+
+
+.. badges
+
+.. |pypi| image:: https://img.shields.io/pypi/v/pluggy.svg
+    :target: https://pypi.org/pypi/pluggy
+
+.. |versions| image:: https://img.shields.io/pypi/pyversions/pluggy.svg
+    :target: https://pypi.org/pypi/pluggy
+
+.. |github-actions| image:: https://github.com/pytest-dev/pluggy/workflows/main/badge.svg
+    :target: https://github.com/pytest-dev/pluggy/actions
+
+.. |conda-forge| image:: https://img.shields.io/conda/vn/conda-forge/pluggy.svg
+    :target: https://anaconda.org/conda-forge/pytest
+
+.. |gitter| image:: https://badges.gitter.im/pytest-dev/pluggy.svg
+    :alt: Join the chat at https://gitter.im/pytest-dev/pluggy
+    :target: https://gitter.im/pytest-dev/pluggy?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge
+
+.. |black| image:: https://img.shields.io/badge/code%20style-black-000000.svg
+    :target: https://github.com/ambv/black
+
+.. |codecov| image:: https://codecov.io/gh/pytest-dev/pluggy/branch/master/graph/badge.svg
+    :target: https://codecov.io/gh/pytest-dev/pluggy
+    :alt: Code coverage Status
+
+.. links
+.. _pytest:
+    http://pytest.org
+.. _tox:
+    https://tox.readthedocs.org
+.. _devpi:
+    http://doc.devpi.net
+.. _read the docs:
+   https://pluggy.readthedocs.io/en/latest/
+
+
+Support pluggy
+--------------
+
+`Open Collective`_ is an online funding platform for open and transparent communities.
+It provides tools to raise money and share your finances in full transparency.
+
+It is the platform of choice for individuals and companies that want to make one-time or
+monthly donations directly to the project.
+
+``pluggy`` is part of the ``pytest-dev`` project, see more details in the `pytest collective`_.
+
+.. _Open Collective: https://opencollective.com
+.. _pytest collective: https://opencollective.com/pytest
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/RECORD
new file mode 100644
index 00000000..544bdfb1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/RECORD
@@ -0,0 +1,23 @@
+pluggy-1.5.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pluggy-1.5.0.dist-info/LICENSE,sha256=1rZebCE6XQtXeRHTTW5ZSbn1nXbCOMUHGi8_wWz7JgY,1110
+pluggy-1.5.0.dist-info/METADATA,sha256=6JeHn3o9P9iqwK20MgVHdoqxick1SS3SORb65Iyb-Fw,4812
+pluggy-1.5.0.dist-info/RECORD,,
+pluggy-1.5.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+pluggy-1.5.0.dist-info/top_level.txt,sha256=xKSCRhai-v9MckvMuWqNz16c1tbsmOggoMSwTgcpYHE,7
+pluggy/__init__.py,sha256=U8qtIRmmr0SRdbxAF8VJJs01jMUYgKAc9oAjYYCLgz4,980
+pluggy/__pycache__/__init__.cpython-312.pyc,,
+pluggy/__pycache__/_callers.cpython-312.pyc,,
+pluggy/__pycache__/_hooks.cpython-312.pyc,,
+pluggy/__pycache__/_manager.cpython-312.pyc,,
+pluggy/__pycache__/_result.cpython-312.pyc,,
+pluggy/__pycache__/_tracing.cpython-312.pyc,,
+pluggy/__pycache__/_version.cpython-312.pyc,,
+pluggy/__pycache__/_warnings.cpython-312.pyc,,
+pluggy/_callers.py,sha256=8k8i3GVBT_gtccCPFpN8Ww0towWSnSazrl0vbP9UXSY,7316
+pluggy/_hooks.py,sha256=m-3qVLDdn4S9y3pffLOpMQeDI4PDw8hrATK1SC8rQkU,25108
+pluggy/_manager.py,sha256=ylIDFwrUP_mMAGpdRPj9zwxukG7nWJAfY1yylXyXAMo,20265
+pluggy/_result.py,sha256=eEak-7Ie88bRkylsgbLwB6iMogogIMZheq8W3bImmcs,2849
+pluggy/_tracing.py,sha256=kSBr25F_rNklV2QhLD6h1jx6Z1kcKDRbuYvF5jv35pU,2089
+pluggy/_version.py,sha256=OYzqgMEgfFG0au4hzbEdgYI-c7Hxo3wdBtrpEjK1RoY,411
+pluggy/_warnings.py,sha256=td0AvZBpfamriCC3OqsLwxMh-SzAMjfjmc58T5vP3lw,828
+pluggy/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/top_level.txt
new file mode 100644
index 00000000..11bdb5c1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy-1.5.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+pluggy
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__init__.py b/backend/venv/lib/python3.12/site-packages/pluggy/__init__.py
new file mode 100644
index 00000000..36ce1680
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/__init__.py
@@ -0,0 +1,37 @@
+try:
+    from ._version import version as __version__
+except ImportError:
+    # broken installation, we don't even try
+    # unknown only works because we do poor mans version compare
+    __version__ = "unknown"
+
+__all__ = [
+    "__version__",
+    "PluginManager",
+    "PluginValidationError",
+    "HookCaller",
+    "HookCallError",
+    "HookspecOpts",
+    "HookimplOpts",
+    "HookImpl",
+    "HookRelay",
+    "HookspecMarker",
+    "HookimplMarker",
+    "Result",
+    "PluggyWarning",
+    "PluggyTeardownRaisedWarning",
+]
+
+from ._hooks import HookCaller
+from ._hooks import HookImpl
+from ._hooks import HookimplMarker
+from ._hooks import HookimplOpts
+from ._hooks import HookRelay
+from ._hooks import HookspecMarker
+from ._hooks import HookspecOpts
+from ._manager import PluginManager
+from ._manager import PluginValidationError
+from ._result import HookCallError
+from ._result import Result
+from ._warnings import PluggyTeardownRaisedWarning
+from ._warnings import PluggyWarning
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..26bf0cd1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_callers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_callers.cpython-312.pyc
new file mode 100644
index 00000000..b1b54df2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_callers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_hooks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_hooks.cpython-312.pyc
new file mode 100644
index 00000000..08b01baf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_hooks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_manager.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_manager.cpython-312.pyc
new file mode 100644
index 00000000..f7137898
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_manager.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_result.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_result.cpython-312.pyc
new file mode 100644
index 00000000..027e7b6c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_result.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_tracing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_tracing.cpython-312.pyc
new file mode 100644
index 00000000..1871a7d7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_tracing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..62ef7a41
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_warnings.cpython-312.pyc
new file mode 100644
index 00000000..9030f0ed
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pluggy/__pycache__/_warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_callers.py b/backend/venv/lib/python3.12/site-packages/pluggy/_callers.py
new file mode 100644
index 00000000..d01f925c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_callers.py
@@ -0,0 +1,182 @@
+"""
+Call loop machinery
+"""
+
+from __future__ import annotations
+
+from typing import cast
+from typing import Generator
+from typing import Mapping
+from typing import NoReturn
+from typing import Sequence
+from typing import Tuple
+from typing import Union
+import warnings
+
+from ._hooks import HookImpl
+from ._result import HookCallError
+from ._result import Result
+from ._warnings import PluggyTeardownRaisedWarning
+
+
+# Need to distinguish between old- and new-style hook wrappers.
+# Wrapping with a tuple is the fastest type-safe way I found to do it.
+Teardown = Union[
+    Tuple[Generator[None, Result[object], None], HookImpl],
+    Generator[None, object, object],
+]
+
+
+def _raise_wrapfail(
+    wrap_controller: (
+        Generator[None, Result[object], None] | Generator[None, object, object]
+    ),
+    msg: str,
+) -> NoReturn:
+    co = wrap_controller.gi_code
+    raise RuntimeError(
+        "wrap_controller at %r %s:%d %s"
+        % (co.co_name, co.co_filename, co.co_firstlineno, msg)
+    )
+
+
+def _warn_teardown_exception(
+    hook_name: str, hook_impl: HookImpl, e: BaseException
+) -> None:
+    msg = "A plugin raised an exception during an old-style hookwrapper teardown.\n"
+    msg += f"Plugin: {hook_impl.plugin_name}, Hook: {hook_name}\n"
+    msg += f"{type(e).__name__}: {e}\n"
+    msg += "For more information see https://pluggy.readthedocs.io/en/stable/api_reference.html#pluggy.PluggyTeardownRaisedWarning"  # noqa: E501
+    warnings.warn(PluggyTeardownRaisedWarning(msg), stacklevel=5)
+
+
+def _multicall(
+    hook_name: str,
+    hook_impls: Sequence[HookImpl],
+    caller_kwargs: Mapping[str, object],
+    firstresult: bool,
+) -> object | list[object]:
+    """Execute a call into multiple python functions/methods and return the
+    result(s).
+
+    ``caller_kwargs`` comes from HookCaller.__call__().
+    """
+    __tracebackhide__ = True
+    results: list[object] = []
+    exception = None
+    only_new_style_wrappers = True
+    try:  # run impl and wrapper setup functions in a loop
+        teardowns: list[Teardown] = []
+        try:
+            for hook_impl in reversed(hook_impls):
+                try:
+                    args = [caller_kwargs[argname] for argname in hook_impl.argnames]
+                except KeyError:
+                    for argname in hook_impl.argnames:
+                        if argname not in caller_kwargs:
+                            raise HookCallError(
+                                f"hook call must provide argument {argname!r}"
+                            )
+
+                if hook_impl.hookwrapper:
+                    only_new_style_wrappers = False
+                    try:
+                        # If this cast is not valid, a type error is raised below,
+                        # which is the desired response.
+                        res = hook_impl.function(*args)
+                        wrapper_gen = cast(Generator[None, Result[object], None], res)
+                        next(wrapper_gen)  # first yield
+                        teardowns.append((wrapper_gen, hook_impl))
+                    except StopIteration:
+                        _raise_wrapfail(wrapper_gen, "did not yield")
+                elif hook_impl.wrapper:
+                    try:
+                        # If this cast is not valid, a type error is raised below,
+                        # which is the desired response.
+                        res = hook_impl.function(*args)
+                        function_gen = cast(Generator[None, object, object], res)
+                        next(function_gen)  # first yield
+                        teardowns.append(function_gen)
+                    except StopIteration:
+                        _raise_wrapfail(function_gen, "did not yield")
+                else:
+                    res = hook_impl.function(*args)
+                    if res is not None:
+                        results.append(res)
+                        if firstresult:  # halt further impl calls
+                            break
+        except BaseException as exc:
+            exception = exc
+    finally:
+        # Fast path - only new-style wrappers, no Result.
+        if only_new_style_wrappers:
+            if firstresult:  # first result hooks return a single value
+                result = results[0] if results else None
+            else:
+                result = results
+
+            # run all wrapper post-yield blocks
+            for teardown in reversed(teardowns):
+                try:
+                    if exception is not None:
+                        teardown.throw(exception)  # type: ignore[union-attr]
+                    else:
+                        teardown.send(result)  # type: ignore[union-attr]
+                    # Following is unreachable for a well behaved hook wrapper.
+                    # Try to force finalizers otherwise postponed till GC action.
+                    # Note: close() may raise if generator handles GeneratorExit.
+                    teardown.close()  # type: ignore[union-attr]
+                except StopIteration as si:
+                    result = si.value
+                    exception = None
+                    continue
+                except BaseException as e:
+                    exception = e
+                    continue
+                _raise_wrapfail(teardown, "has second yield")  # type: ignore[arg-type]
+
+            if exception is not None:
+                raise exception.with_traceback(exception.__traceback__)
+            else:
+                return result
+
+        # Slow path - need to support old-style wrappers.
+        else:
+            if firstresult:  # first result hooks return a single value
+                outcome: Result[object | list[object]] = Result(
+                    results[0] if results else None, exception
+                )
+            else:
+                outcome = Result(results, exception)
+
+            # run all wrapper post-yield blocks
+            for teardown in reversed(teardowns):
+                if isinstance(teardown, tuple):
+                    try:
+                        teardown[0].send(outcome)
+                    except StopIteration:
+                        pass
+                    except BaseException as e:
+                        _warn_teardown_exception(hook_name, teardown[1], e)
+                        raise
+                    else:
+                        _raise_wrapfail(teardown[0], "has second yield")
+                else:
+                    try:
+                        if outcome._exception is not None:
+                            teardown.throw(outcome._exception)
+                        else:
+                            teardown.send(outcome._result)
+                        # Following is unreachable for a well behaved hook wrapper.
+                        # Try to force finalizers otherwise postponed till GC action.
+                        # Note: close() may raise if generator handles GeneratorExit.
+                        teardown.close()
+                    except StopIteration as si:
+                        outcome.force_result(si.value)
+                        continue
+                    except BaseException as e:
+                        outcome.force_exception(e)
+                        continue
+                    _raise_wrapfail(teardown, "has second yield")
+
+            return outcome.get_result()
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_hooks.py b/backend/venv/lib/python3.12/site-packages/pluggy/_hooks.py
new file mode 100644
index 00000000..362d7918
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_hooks.py
@@ -0,0 +1,715 @@
+"""
+Internal hook annotation, representation and calling machinery.
+"""
+
+from __future__ import annotations
+
+import inspect
+import sys
+from types import ModuleType
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import Final
+from typing import final
+from typing import Generator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypedDict
+from typing import TypeVar
+from typing import Union
+import warnings
+
+from ._result import Result
+
+
+_T = TypeVar("_T")
+_F = TypeVar("_F", bound=Callable[..., object])
+_Namespace = Union[ModuleType, type]
+_Plugin = object
+_HookExec = Callable[
+    [str, Sequence["HookImpl"], Mapping[str, object], bool],
+    Union[object, List[object]],
+]
+_HookImplFunction = Callable[..., Union[_T, Generator[None, Result[_T], None]]]
+
+
+class HookspecOpts(TypedDict):
+    """Options for a hook specification."""
+
+    #: Whether the hook is :ref:`first result only <firstresult>`.
+    firstresult: bool
+    #: Whether the hook is :ref:`historic <historic>`.
+    historic: bool
+    #: Whether the hook :ref:`warns when implemented <warn_on_impl>`.
+    warn_on_impl: Warning | None
+    #: Whether the hook warns when :ref:`certain arguments are requested
+    #: <warn_on_impl>`.
+    #:
+    #: .. versionadded:: 1.5
+    warn_on_impl_args: Mapping[str, Warning] | None
+
+
+class HookimplOpts(TypedDict):
+    """Options for a hook implementation."""
+
+    #: Whether the hook implementation is a :ref:`wrapper <hookwrapper>`.
+    wrapper: bool
+    #: Whether the hook implementation is an :ref:`old-style wrapper
+    #: <old_style_hookwrappers>`.
+    hookwrapper: bool
+    #: Whether validation against a hook specification is :ref:`optional
+    #: <optionalhook>`.
+    optionalhook: bool
+    #: Whether to try to order this hook implementation :ref:`first
+    #: <callorder>`.
+    tryfirst: bool
+    #: Whether to try to order this hook implementation :ref:`last
+    #: <callorder>`.
+    trylast: bool
+    #: The name of the hook specification to match, see :ref:`specname`.
+    specname: str | None
+
+
+@final
+class HookspecMarker:
+    """Decorator for marking functions as hook specifications.
+
+    Instantiate it with a project_name to get a decorator.
+    Calling :meth:`PluginManager.add_hookspecs` later will discover all marked
+    functions if the :class:`PluginManager` uses the same project name.
+    """
+
+    __slots__ = ("project_name",)
+
+    def __init__(self, project_name: str) -> None:
+        self.project_name: Final = project_name
+
+    @overload
+    def __call__(
+        self,
+        function: _F,
+        firstresult: bool = False,
+        historic: bool = False,
+        warn_on_impl: Warning | None = None,
+        warn_on_impl_args: Mapping[str, Warning] | None = None,
+    ) -> _F: ...
+
+    @overload  # noqa: F811
+    def __call__(  # noqa: F811
+        self,
+        function: None = ...,
+        firstresult: bool = ...,
+        historic: bool = ...,
+        warn_on_impl: Warning | None = ...,
+        warn_on_impl_args: Mapping[str, Warning] | None = ...,
+    ) -> Callable[[_F], _F]: ...
+
+    def __call__(  # noqa: F811
+        self,
+        function: _F | None = None,
+        firstresult: bool = False,
+        historic: bool = False,
+        warn_on_impl: Warning | None = None,
+        warn_on_impl_args: Mapping[str, Warning] | None = None,
+    ) -> _F | Callable[[_F], _F]:
+        """If passed a function, directly sets attributes on the function
+        which will make it discoverable to :meth:`PluginManager.add_hookspecs`.
+
+        If passed no function, returns a decorator which can be applied to a
+        function later using the attributes supplied.
+
+        :param firstresult:
+            If ``True``, the 1:N hook call (N being the number of registered
+            hook implementation functions) will stop at I<=N when the I'th
+            function returns a non-``None`` result. See :ref:`firstresult`.
+
+        :param historic:
+            If ``True``, every call to the hook will be memorized and replayed
+            on plugins registered after the call was made. See :ref:`historic`.
+
+        :param warn_on_impl:
+            If given, every implementation of this hook will trigger the given
+            warning. See :ref:`warn_on_impl`.
+
+        :param warn_on_impl_args:
+            If given, every implementation of this hook which requests one of
+            the arguments in the dict will trigger the corresponding warning.
+            See :ref:`warn_on_impl`.
+
+            .. versionadded:: 1.5
+        """
+
+        def setattr_hookspec_opts(func: _F) -> _F:
+            if historic and firstresult:
+                raise ValueError("cannot have a historic firstresult hook")
+            opts: HookspecOpts = {
+                "firstresult": firstresult,
+                "historic": historic,
+                "warn_on_impl": warn_on_impl,
+                "warn_on_impl_args": warn_on_impl_args,
+            }
+            setattr(func, self.project_name + "_spec", opts)
+            return func
+
+        if function is not None:
+            return setattr_hookspec_opts(function)
+        else:
+            return setattr_hookspec_opts
+
+
+@final
+class HookimplMarker:
+    """Decorator for marking functions as hook implementations.
+
+    Instantiate it with a ``project_name`` to get a decorator.
+    Calling :meth:`PluginManager.register` later will discover all marked
+    functions if the :class:`PluginManager` uses the same project name.
+    """
+
+    __slots__ = ("project_name",)
+
+    def __init__(self, project_name: str) -> None:
+        self.project_name: Final = project_name
+
+    @overload
+    def __call__(
+        self,
+        function: _F,
+        hookwrapper: bool = ...,
+        optionalhook: bool = ...,
+        tryfirst: bool = ...,
+        trylast: bool = ...,
+        specname: str | None = ...,
+        wrapper: bool = ...,
+    ) -> _F: ...
+
+    @overload  # noqa: F811
+    def __call__(  # noqa: F811
+        self,
+        function: None = ...,
+        hookwrapper: bool = ...,
+        optionalhook: bool = ...,
+        tryfirst: bool = ...,
+        trylast: bool = ...,
+        specname: str | None = ...,
+        wrapper: bool = ...,
+    ) -> Callable[[_F], _F]: ...
+
+    def __call__(  # noqa: F811
+        self,
+        function: _F | None = None,
+        hookwrapper: bool = False,
+        optionalhook: bool = False,
+        tryfirst: bool = False,
+        trylast: bool = False,
+        specname: str | None = None,
+        wrapper: bool = False,
+    ) -> _F | Callable[[_F], _F]:
+        """If passed a function, directly sets attributes on the function
+        which will make it discoverable to :meth:`PluginManager.register`.
+
+        If passed no function, returns a decorator which can be applied to a
+        function later using the attributes supplied.
+
+        :param optionalhook:
+            If ``True``, a missing matching hook specification will not result
+            in an error (by default it is an error if no matching spec is
+            found). See :ref:`optionalhook`.
+
+        :param tryfirst:
+            If ``True``, this hook implementation will run as early as possible
+            in the chain of N hook implementations for a specification. See
+            :ref:`callorder`.
+
+        :param trylast:
+            If ``True``, this hook implementation will run as late as possible
+            in the chain of N hook implementations for a specification. See
+            :ref:`callorder`.
+
+        :param wrapper:
+            If ``True`` ("new-style hook wrapper"), the hook implementation
+            needs to execute exactly one ``yield``. The code before the
+            ``yield`` is run early before any non-hook-wrapper function is run.
+            The code after the ``yield`` is run after all non-hook-wrapper
+            functions have run. The ``yield`` receives the result value of the
+            inner calls, or raises the exception of inner calls (including
+            earlier hook wrapper calls). The return value of the function
+            becomes the return value of the hook, and a raised exception becomes
+            the exception of the hook. See :ref:`hookwrapper`.
+
+        :param hookwrapper:
+            If ``True`` ("old-style hook wrapper"), the hook implementation
+            needs to execute exactly one ``yield``. The code before the
+            ``yield`` is run early before any non-hook-wrapper function is run.
+            The code after the ``yield`` is run after all non-hook-wrapper
+            function have run  The ``yield`` receives a :class:`Result` object
+            representing the exception or result outcome of the inner calls
+            (including earlier hook wrapper calls). This option is mutually
+            exclusive with ``wrapper``. See :ref:`old_style_hookwrapper`.
+
+        :param specname:
+            If provided, the given name will be used instead of the function
+            name when matching this hook implementation to a hook specification
+            during registration. See :ref:`specname`.
+
+        .. versionadded:: 1.2.0
+            The ``wrapper`` parameter.
+        """
+
+        def setattr_hookimpl_opts(func: _F) -> _F:
+            opts: HookimplOpts = {
+                "wrapper": wrapper,
+                "hookwrapper": hookwrapper,
+                "optionalhook": optionalhook,
+                "tryfirst": tryfirst,
+                "trylast": trylast,
+                "specname": specname,
+            }
+            setattr(func, self.project_name + "_impl", opts)
+            return func
+
+        if function is None:
+            return setattr_hookimpl_opts
+        else:
+            return setattr_hookimpl_opts(function)
+
+
+def normalize_hookimpl_opts(opts: HookimplOpts) -> None:
+    opts.setdefault("tryfirst", False)
+    opts.setdefault("trylast", False)
+    opts.setdefault("wrapper", False)
+    opts.setdefault("hookwrapper", False)
+    opts.setdefault("optionalhook", False)
+    opts.setdefault("specname", None)
+
+
+_PYPY = hasattr(sys, "pypy_version_info")
+
+
+def varnames(func: object) -> tuple[tuple[str, ...], tuple[str, ...]]:
+    """Return tuple of positional and keywrord argument names for a function,
+    method, class or callable.
+
+    In case of a class, its ``__init__`` method is considered.
+    For methods the ``self`` parameter is not included.
+    """
+    if inspect.isclass(func):
+        try:
+            func = func.__init__
+        except AttributeError:
+            return (), ()
+    elif not inspect.isroutine(func):  # callable object?
+        try:
+            func = getattr(func, "__call__", func)
+        except Exception:
+            return (), ()
+
+    try:
+        # func MUST be a function or method here or we won't parse any args.
+        sig = inspect.signature(
+            func.__func__ if inspect.ismethod(func) else func  # type:ignore[arg-type]
+        )
+    except TypeError:
+        return (), ()
+
+    _valid_param_kinds = (
+        inspect.Parameter.POSITIONAL_ONLY,
+        inspect.Parameter.POSITIONAL_OR_KEYWORD,
+    )
+    _valid_params = {
+        name: param
+        for name, param in sig.parameters.items()
+        if param.kind in _valid_param_kinds
+    }
+    args = tuple(_valid_params)
+    defaults = (
+        tuple(
+            param.default
+            for param in _valid_params.values()
+            if param.default is not param.empty
+        )
+        or None
+    )
+
+    if defaults:
+        index = -len(defaults)
+        args, kwargs = args[:index], tuple(args[index:])
+    else:
+        kwargs = ()
+
+    # strip any implicit instance arg
+    # pypy3 uses "obj" instead of "self" for default dunder methods
+    if not _PYPY:
+        implicit_names: tuple[str, ...] = ("self",)
+    else:
+        implicit_names = ("self", "obj")
+    if args:
+        qualname: str = getattr(func, "__qualname__", "")
+        if inspect.ismethod(func) or ("." in qualname and args[0] in implicit_names):
+            args = args[1:]
+
+    return args, kwargs
+
+
+@final
+class HookRelay:
+    """Hook holder object for performing 1:N hook calls where N is the number
+    of registered plugins."""
+
+    __slots__ = ("__dict__",)
+
+    def __init__(self) -> None:
+        """:meta private:"""
+
+    if TYPE_CHECKING:
+
+        def __getattr__(self, name: str) -> HookCaller: ...
+
+
+# Historical name (pluggy<=1.2), kept for backward compatibility.
+_HookRelay = HookRelay
+
+
+_CallHistory = List[Tuple[Mapping[str, object], Optional[Callable[[Any], None]]]]
+
+
+class HookCaller:
+    """A caller of all registered implementations of a hook specification."""
+
+    __slots__ = (
+        "name",
+        "spec",
+        "_hookexec",
+        "_hookimpls",
+        "_call_history",
+    )
+
+    def __init__(
+        self,
+        name: str,
+        hook_execute: _HookExec,
+        specmodule_or_class: _Namespace | None = None,
+        spec_opts: HookspecOpts | None = None,
+    ) -> None:
+        """:meta private:"""
+        #: Name of the hook getting called.
+        self.name: Final = name
+        self._hookexec: Final = hook_execute
+        # The hookimpls list. The caller iterates it *in reverse*. Format:
+        # 1. trylast nonwrappers
+        # 2. nonwrappers
+        # 3. tryfirst nonwrappers
+        # 4. trylast wrappers
+        # 5. wrappers
+        # 6. tryfirst wrappers
+        self._hookimpls: Final[list[HookImpl]] = []
+        self._call_history: _CallHistory | None = None
+        # TODO: Document, or make private.
+        self.spec: HookSpec | None = None
+        if specmodule_or_class is not None:
+            assert spec_opts is not None
+            self.set_specification(specmodule_or_class, spec_opts)
+
+    # TODO: Document, or make private.
+    def has_spec(self) -> bool:
+        return self.spec is not None
+
+    # TODO: Document, or make private.
+    def set_specification(
+        self,
+        specmodule_or_class: _Namespace,
+        spec_opts: HookspecOpts,
+    ) -> None:
+        if self.spec is not None:
+            raise ValueError(
+                f"Hook {self.spec.name!r} is already registered "
+                f"within namespace {self.spec.namespace}"
+            )
+        self.spec = HookSpec(specmodule_or_class, self.name, spec_opts)
+        if spec_opts.get("historic"):
+            self._call_history = []
+
+    def is_historic(self) -> bool:
+        """Whether this caller is :ref:`historic <historic>`."""
+        return self._call_history is not None
+
+    def _remove_plugin(self, plugin: _Plugin) -> None:
+        for i, method in enumerate(self._hookimpls):
+            if method.plugin == plugin:
+                del self._hookimpls[i]
+                return
+        raise ValueError(f"plugin {plugin!r} not found")
+
+    def get_hookimpls(self) -> list[HookImpl]:
+        """Get all registered hook implementations for this hook."""
+        return self._hookimpls.copy()
+
+    def _add_hookimpl(self, hookimpl: HookImpl) -> None:
+        """Add an implementation to the callback chain."""
+        for i, method in enumerate(self._hookimpls):
+            if method.hookwrapper or method.wrapper:
+                splitpoint = i
+                break
+        else:
+            splitpoint = len(self._hookimpls)
+        if hookimpl.hookwrapper or hookimpl.wrapper:
+            start, end = splitpoint, len(self._hookimpls)
+        else:
+            start, end = 0, splitpoint
+
+        if hookimpl.trylast:
+            self._hookimpls.insert(start, hookimpl)
+        elif hookimpl.tryfirst:
+            self._hookimpls.insert(end, hookimpl)
+        else:
+            # find last non-tryfirst method
+            i = end - 1
+            while i >= start and self._hookimpls[i].tryfirst:
+                i -= 1
+            self._hookimpls.insert(i + 1, hookimpl)
+
+    def __repr__(self) -> str:
+        return f"<HookCaller {self.name!r}>"
+
+    def _verify_all_args_are_provided(self, kwargs: Mapping[str, object]) -> None:
+        # This is written to avoid expensive operations when not needed.
+        if self.spec:
+            for argname in self.spec.argnames:
+                if argname not in kwargs:
+                    notincall = ", ".join(
+                        repr(argname)
+                        for argname in self.spec.argnames
+                        # Avoid self.spec.argnames - kwargs.keys() - doesn't preserve order.
+                        if argname not in kwargs.keys()
+                    )
+                    warnings.warn(
+                        "Argument(s) {} which are declared in the hookspec "
+                        "cannot be found in this hook call".format(notincall),
+                        stacklevel=2,
+                    )
+                    break
+
+    def __call__(self, **kwargs: object) -> Any:
+        """Call the hook.
+
+        Only accepts keyword arguments, which should match the hook
+        specification.
+
+        Returns the result(s) of calling all registered plugins, see
+        :ref:`calling`.
+        """
+        assert (
+            not self.is_historic()
+        ), "Cannot directly call a historic hook - use call_historic instead."
+        self._verify_all_args_are_provided(kwargs)
+        firstresult = self.spec.opts.get("firstresult", False) if self.spec else False
+        # Copy because plugins may register other plugins during iteration (#438).
+        return self._hookexec(self.name, self._hookimpls.copy(), kwargs, firstresult)
+
+    def call_historic(
+        self,
+        result_callback: Callable[[Any], None] | None = None,
+        kwargs: Mapping[str, object] | None = None,
+    ) -> None:
+        """Call the hook with given ``kwargs`` for all registered plugins and
+        for all plugins which will be registered afterwards, see
+        :ref:`historic`.
+
+        :param result_callback:
+            If provided, will be called for each non-``None`` result obtained
+            from a hook implementation.
+        """
+        assert self._call_history is not None
+        kwargs = kwargs or {}
+        self._verify_all_args_are_provided(kwargs)
+        self._call_history.append((kwargs, result_callback))
+        # Historizing hooks don't return results.
+        # Remember firstresult isn't compatible with historic.
+        # Copy because plugins may register other plugins during iteration (#438).
+        res = self._hookexec(self.name, self._hookimpls.copy(), kwargs, False)
+        if result_callback is None:
+            return
+        if isinstance(res, list):
+            for x in res:
+                result_callback(x)
+
+    def call_extra(
+        self, methods: Sequence[Callable[..., object]], kwargs: Mapping[str, object]
+    ) -> Any:
+        """Call the hook with some additional temporarily participating
+        methods using the specified ``kwargs`` as call parameters, see
+        :ref:`call_extra`."""
+        assert (
+            not self.is_historic()
+        ), "Cannot directly call a historic hook - use call_historic instead."
+        self._verify_all_args_are_provided(kwargs)
+        opts: HookimplOpts = {
+            "wrapper": False,
+            "hookwrapper": False,
+            "optionalhook": False,
+            "trylast": False,
+            "tryfirst": False,
+            "specname": None,
+        }
+        hookimpls = self._hookimpls.copy()
+        for method in methods:
+            hookimpl = HookImpl(None, "<temp>", method, opts)
+            # Find last non-tryfirst nonwrapper method.
+            i = len(hookimpls) - 1
+            while i >= 0 and (
+                # Skip wrappers.
+                (hookimpls[i].hookwrapper or hookimpls[i].wrapper)
+                # Skip tryfirst nonwrappers.
+                or hookimpls[i].tryfirst
+            ):
+                i -= 1
+            hookimpls.insert(i + 1, hookimpl)
+        firstresult = self.spec.opts.get("firstresult", False) if self.spec else False
+        return self._hookexec(self.name, hookimpls, kwargs, firstresult)
+
+    def _maybe_apply_history(self, method: HookImpl) -> None:
+        """Apply call history to a new hookimpl if it is marked as historic."""
+        if self.is_historic():
+            assert self._call_history is not None
+            for kwargs, result_callback in self._call_history:
+                res = self._hookexec(self.name, [method], kwargs, False)
+                if res and result_callback is not None:
+                    # XXX: remember firstresult isn't compat with historic
+                    assert isinstance(res, list)
+                    result_callback(res[0])
+
+
+# Historical name (pluggy<=1.2), kept for backward compatibility.
+_HookCaller = HookCaller
+
+
+class _SubsetHookCaller(HookCaller):
+    """A proxy to another HookCaller which manages calls to all registered
+    plugins except the ones from remove_plugins."""
+
+    # This class is unusual: in inhertits from `HookCaller` so all of
+    # the *code* runs in the class, but it delegates all underlying *data*
+    # to the original HookCaller.
+    # `subset_hook_caller` used to be implemented by creating a full-fledged
+    # HookCaller, copying all hookimpls from the original. This had problems
+    # with memory leaks (#346) and historic calls (#347), which make a proxy
+    # approach better.
+    # An alternative implementation is to use a `_getattr__`/`__getattribute__`
+    # proxy, however that adds more overhead and is more tricky to implement.
+
+    __slots__ = (
+        "_orig",
+        "_remove_plugins",
+    )
+
+    def __init__(self, orig: HookCaller, remove_plugins: AbstractSet[_Plugin]) -> None:
+        self._orig = orig
+        self._remove_plugins = remove_plugins
+        self.name = orig.name  # type: ignore[misc]
+        self._hookexec = orig._hookexec  # type: ignore[misc]
+
+    @property  # type: ignore[misc]
+    def _hookimpls(self) -> list[HookImpl]:
+        return [
+            impl
+            for impl in self._orig._hookimpls
+            if impl.plugin not in self._remove_plugins
+        ]
+
+    @property
+    def spec(self) -> HookSpec | None:  # type: ignore[override]
+        return self._orig.spec
+
+    @property
+    def _call_history(self) -> _CallHistory | None:  # type: ignore[override]
+        return self._orig._call_history
+
+    def __repr__(self) -> str:
+        return f"<_SubsetHookCaller {self.name!r}>"
+
+
+@final
+class HookImpl:
+    """A hook implementation in a :class:`HookCaller`."""
+
+    __slots__ = (
+        "function",
+        "argnames",
+        "kwargnames",
+        "plugin",
+        "opts",
+        "plugin_name",
+        "wrapper",
+        "hookwrapper",
+        "optionalhook",
+        "tryfirst",
+        "trylast",
+    )
+
+    def __init__(
+        self,
+        plugin: _Plugin,
+        plugin_name: str,
+        function: _HookImplFunction[object],
+        hook_impl_opts: HookimplOpts,
+    ) -> None:
+        """:meta private:"""
+        #: The hook implementation function.
+        self.function: Final = function
+        argnames, kwargnames = varnames(self.function)
+        #: The positional parameter names of ``function```.
+        self.argnames: Final = argnames
+        #: The keyword parameter names of ``function```.
+        self.kwargnames: Final = kwargnames
+        #: The plugin which defined this hook implementation.
+        self.plugin: Final = plugin
+        #: The :class:`HookimplOpts` used to configure this hook implementation.
+        self.opts: Final = hook_impl_opts
+        #: The name of the plugin which defined this hook implementation.
+        self.plugin_name: Final = plugin_name
+        #: Whether the hook implementation is a :ref:`wrapper <hookwrapper>`.
+        self.wrapper: Final = hook_impl_opts["wrapper"]
+        #: Whether the hook implementation is an :ref:`old-style wrapper
+        #: <old_style_hookwrappers>`.
+        self.hookwrapper: Final = hook_impl_opts["hookwrapper"]
+        #: Whether validation against a hook specification is :ref:`optional
+        #: <optionalhook>`.
+        self.optionalhook: Final = hook_impl_opts["optionalhook"]
+        #: Whether to try to order this hook implementation :ref:`first
+        #: <callorder>`.
+        self.tryfirst: Final = hook_impl_opts["tryfirst"]
+        #: Whether to try to order this hook implementation :ref:`last
+        #: <callorder>`.
+        self.trylast: Final = hook_impl_opts["trylast"]
+
+    def __repr__(self) -> str:
+        return f"<HookImpl plugin_name={self.plugin_name!r}, plugin={self.plugin!r}>"
+
+
+@final
+class HookSpec:
+    __slots__ = (
+        "namespace",
+        "function",
+        "name",
+        "argnames",
+        "kwargnames",
+        "opts",
+        "warn_on_impl",
+        "warn_on_impl_args",
+    )
+
+    def __init__(self, namespace: _Namespace, name: str, opts: HookspecOpts) -> None:
+        self.namespace = namespace
+        self.function: Callable[..., object] = getattr(namespace, name)
+        self.name = name
+        self.argnames, self.kwargnames = varnames(self.function)
+        self.opts = opts
+        self.warn_on_impl = opts.get("warn_on_impl")
+        self.warn_on_impl_args = opts.get("warn_on_impl_args")
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_manager.py b/backend/venv/lib/python3.12/site-packages/pluggy/_manager.py
new file mode 100644
index 00000000..9998dd81
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_manager.py
@@ -0,0 +1,528 @@
+from __future__ import annotations
+
+import inspect
+import types
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Final
+from typing import Iterable
+from typing import Mapping
+from typing import Sequence
+from typing import TYPE_CHECKING
+import warnings
+
+from . import _tracing
+from ._callers import _multicall
+from ._hooks import _HookImplFunction
+from ._hooks import _Namespace
+from ._hooks import _Plugin
+from ._hooks import _SubsetHookCaller
+from ._hooks import HookCaller
+from ._hooks import HookImpl
+from ._hooks import HookimplOpts
+from ._hooks import HookRelay
+from ._hooks import HookspecOpts
+from ._hooks import normalize_hookimpl_opts
+from ._result import Result
+
+
+if TYPE_CHECKING:
+    # importtlib.metadata import is slow, defer it.
+    import importlib.metadata
+
+
+_BeforeTrace = Callable[[str, Sequence[HookImpl], Mapping[str, Any]], None]
+_AfterTrace = Callable[[Result[Any], str, Sequence[HookImpl], Mapping[str, Any]], None]
+
+
+def _warn_for_function(warning: Warning, function: Callable[..., object]) -> None:
+    func = cast(types.FunctionType, function)
+    warnings.warn_explicit(
+        warning,
+        type(warning),
+        lineno=func.__code__.co_firstlineno,
+        filename=func.__code__.co_filename,
+    )
+
+
+class PluginValidationError(Exception):
+    """Plugin failed validation.
+
+    :param plugin: The plugin which failed validation.
+    :param message: Error message.
+    """
+
+    def __init__(self, plugin: _Plugin, message: str) -> None:
+        super().__init__(message)
+        #: The plugin which failed validation.
+        self.plugin = plugin
+
+
+class DistFacade:
+    """Emulate a pkg_resources Distribution"""
+
+    def __init__(self, dist: importlib.metadata.Distribution) -> None:
+        self._dist = dist
+
+    @property
+    def project_name(self) -> str:
+        name: str = self.metadata["name"]
+        return name
+
+    def __getattr__(self, attr: str, default=None):
+        return getattr(self._dist, attr, default)
+
+    def __dir__(self) -> list[str]:
+        return sorted(dir(self._dist) + ["_dist", "project_name"])
+
+
+class PluginManager:
+    """Core class which manages registration of plugin objects and 1:N hook
+    calling.
+
+    You can register new hooks by calling :meth:`add_hookspecs(module_or_class)
+    <PluginManager.add_hookspecs>`.
+
+    You can register plugin objects (which contain hook implementations) by
+    calling :meth:`register(plugin) <PluginManager.register>`.
+
+    For debugging purposes you can call :meth:`PluginManager.enable_tracing`
+    which will subsequently send debug information to the trace helper.
+
+    :param project_name:
+        The short project name. Prefer snake case. Make sure it's unique!
+    """
+
+    def __init__(self, project_name: str) -> None:
+        #: The project name.
+        self.project_name: Final = project_name
+        self._name2plugin: Final[dict[str, _Plugin]] = {}
+        self._plugin_distinfo: Final[list[tuple[_Plugin, DistFacade]]] = []
+        #: The "hook relay", used to call a hook on all registered plugins.
+        #: See :ref:`calling`.
+        self.hook: Final = HookRelay()
+        #: The tracing entry point. See :ref:`tracing`.
+        self.trace: Final[_tracing.TagTracerSub] = _tracing.TagTracer().get(
+            "pluginmanage"
+        )
+        self._inner_hookexec = _multicall
+
+    def _hookexec(
+        self,
+        hook_name: str,
+        methods: Sequence[HookImpl],
+        kwargs: Mapping[str, object],
+        firstresult: bool,
+    ) -> object | list[object]:
+        # called from all hookcaller instances.
+        # enable_tracing will set its own wrapping function at self._inner_hookexec
+        return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
+
+    def register(self, plugin: _Plugin, name: str | None = None) -> str | None:
+        """Register a plugin and return its name.
+
+        :param name:
+            The name under which to register the plugin. If not specified, a
+            name is generated using :func:`get_canonical_name`.
+
+        :returns:
+            The plugin name. If the name is blocked from registering, returns
+            ``None``.
+
+        If the plugin is already registered, raises a :exc:`ValueError`.
+        """
+        plugin_name = name or self.get_canonical_name(plugin)
+
+        if plugin_name in self._name2plugin:
+            if self._name2plugin.get(plugin_name, -1) is None:
+                return None  # blocked plugin, return None to indicate no registration
+            raise ValueError(
+                "Plugin name already registered: %s=%s\n%s"
+                % (plugin_name, plugin, self._name2plugin)
+            )
+
+        if plugin in self._name2plugin.values():
+            raise ValueError(
+                "Plugin already registered under a different name: %s=%s\n%s"
+                % (plugin_name, plugin, self._name2plugin)
+            )
+
+        # XXX if an error happens we should make sure no state has been
+        # changed at point of return
+        self._name2plugin[plugin_name] = plugin
+
+        # register matching hook implementations of the plugin
+        for name in dir(plugin):
+            hookimpl_opts = self.parse_hookimpl_opts(plugin, name)
+            if hookimpl_opts is not None:
+                normalize_hookimpl_opts(hookimpl_opts)
+                method: _HookImplFunction[object] = getattr(plugin, name)
+                hookimpl = HookImpl(plugin, plugin_name, method, hookimpl_opts)
+                name = hookimpl_opts.get("specname") or name
+                hook: HookCaller | None = getattr(self.hook, name, None)
+                if hook is None:
+                    hook = HookCaller(name, self._hookexec)
+                    setattr(self.hook, name, hook)
+                elif hook.has_spec():
+                    self._verify_hook(hook, hookimpl)
+                    hook._maybe_apply_history(hookimpl)
+                hook._add_hookimpl(hookimpl)
+        return plugin_name
+
+    def parse_hookimpl_opts(self, plugin: _Plugin, name: str) -> HookimplOpts | None:
+        """Try to obtain a hook implementation from an item with the given name
+        in the given plugin which is being searched for hook impls.
+
+        :returns:
+            The parsed hookimpl options, or None to skip the given item.
+
+        This method can be overridden by ``PluginManager`` subclasses to
+        customize how hook implementation are picked up. By default, returns the
+        options for items decorated with :class:`HookimplMarker`.
+        """
+        method: object = getattr(plugin, name)
+        if not inspect.isroutine(method):
+            return None
+        try:
+            res: HookimplOpts | None = getattr(
+                method, self.project_name + "_impl", None
+            )
+        except Exception:
+            res = {}  # type: ignore[assignment]
+        if res is not None and not isinstance(res, dict):
+            # false positive
+            res = None  # type:ignore[unreachable]
+        return res
+
+    def unregister(
+        self, plugin: _Plugin | None = None, name: str | None = None
+    ) -> Any | None:
+        """Unregister a plugin and all of its hook implementations.
+
+        The plugin can be specified either by the plugin object or the plugin
+        name. If both are specified, they must agree.
+
+        Returns the unregistered plugin, or ``None`` if not found.
+        """
+        if name is None:
+            assert plugin is not None, "one of name or plugin needs to be specified"
+            name = self.get_name(plugin)
+            assert name is not None, "plugin is not registered"
+
+        if plugin is None:
+            plugin = self.get_plugin(name)
+            if plugin is None:
+                return None
+
+        hookcallers = self.get_hookcallers(plugin)
+        if hookcallers:
+            for hookcaller in hookcallers:
+                hookcaller._remove_plugin(plugin)
+
+        # if self._name2plugin[name] == None registration was blocked: ignore
+        if self._name2plugin.get(name):
+            assert name is not None
+            del self._name2plugin[name]
+
+        return plugin
+
+    def set_blocked(self, name: str) -> None:
+        """Block registrations of the given name, unregister if already registered."""
+        self.unregister(name=name)
+        self._name2plugin[name] = None
+
+    def is_blocked(self, name: str) -> bool:
+        """Return whether the given plugin name is blocked."""
+        return name in self._name2plugin and self._name2plugin[name] is None
+
+    def unblock(self, name: str) -> bool:
+        """Unblocks a name.
+
+        Returns whether the name was actually blocked.
+        """
+        if self._name2plugin.get(name, -1) is None:
+            del self._name2plugin[name]
+            return True
+        return False
+
+    def add_hookspecs(self, module_or_class: _Namespace) -> None:
+        """Add new hook specifications defined in the given ``module_or_class``.
+
+        Functions are recognized as hook specifications if they have been
+        decorated with a matching :class:`HookspecMarker`.
+        """
+        names = []
+        for name in dir(module_or_class):
+            spec_opts = self.parse_hookspec_opts(module_or_class, name)
+            if spec_opts is not None:
+                hc: HookCaller | None = getattr(self.hook, name, None)
+                if hc is None:
+                    hc = HookCaller(name, self._hookexec, module_or_class, spec_opts)
+                    setattr(self.hook, name, hc)
+                else:
+                    # Plugins registered this hook without knowing the spec.
+                    hc.set_specification(module_or_class, spec_opts)
+                    for hookfunction in hc.get_hookimpls():
+                        self._verify_hook(hc, hookfunction)
+                names.append(name)
+
+        if not names:
+            raise ValueError(
+                f"did not find any {self.project_name!r} hooks in {module_or_class!r}"
+            )
+
+    def parse_hookspec_opts(
+        self, module_or_class: _Namespace, name: str
+    ) -> HookspecOpts | None:
+        """Try to obtain a hook specification from an item with the given name
+        in the given module or class which is being searched for hook specs.
+
+        :returns:
+            The parsed hookspec options for defining a hook, or None to skip the
+            given item.
+
+        This method can be overridden by ``PluginManager`` subclasses to
+        customize how hook specifications are picked up. By default, returns the
+        options for items decorated with :class:`HookspecMarker`.
+        """
+        method = getattr(module_or_class, name)
+        opts: HookspecOpts | None = getattr(method, self.project_name + "_spec", None)
+        return opts
+
+    def get_plugins(self) -> set[Any]:
+        """Return a set of all registered plugin objects."""
+        return {x for x in self._name2plugin.values() if x is not None}
+
+    def is_registered(self, plugin: _Plugin) -> bool:
+        """Return whether the plugin is already registered."""
+        return any(plugin == val for val in self._name2plugin.values())
+
+    def get_canonical_name(self, plugin: _Plugin) -> str:
+        """Return a canonical name for a plugin object.
+
+        Note that a plugin may be registered under a different name
+        specified by the caller of :meth:`register(plugin, name) <register>`.
+        To obtain the name of a registered plugin use :meth:`get_name(plugin)
+        <get_name>` instead.
+        """
+        name: str | None = getattr(plugin, "__name__", None)
+        return name or str(id(plugin))
+
+    def get_plugin(self, name: str) -> Any | None:
+        """Return the plugin registered under the given name, if any."""
+        return self._name2plugin.get(name)
+
+    def has_plugin(self, name: str) -> bool:
+        """Return whether a plugin with the given name is registered."""
+        return self.get_plugin(name) is not None
+
+    def get_name(self, plugin: _Plugin) -> str | None:
+        """Return the name the plugin is registered under, or ``None`` if
+        is isn't."""
+        for name, val in self._name2plugin.items():
+            if plugin == val:
+                return name
+        return None
+
+    def _verify_hook(self, hook: HookCaller, hookimpl: HookImpl) -> None:
+        if hook.is_historic() and (hookimpl.hookwrapper or hookimpl.wrapper):
+            raise PluginValidationError(
+                hookimpl.plugin,
+                "Plugin %r\nhook %r\nhistoric incompatible with yield/wrapper/hookwrapper"
+                % (hookimpl.plugin_name, hook.name),
+            )
+
+        assert hook.spec is not None
+        if hook.spec.warn_on_impl:
+            _warn_for_function(hook.spec.warn_on_impl, hookimpl.function)
+
+        # positional arg checking
+        notinspec = set(hookimpl.argnames) - set(hook.spec.argnames)
+        if notinspec:
+            raise PluginValidationError(
+                hookimpl.plugin,
+                "Plugin %r for hook %r\nhookimpl definition: %s\n"
+                "Argument(s) %s are declared in the hookimpl but "
+                "can not be found in the hookspec"
+                % (
+                    hookimpl.plugin_name,
+                    hook.name,
+                    _formatdef(hookimpl.function),
+                    notinspec,
+                ),
+            )
+
+        if hook.spec.warn_on_impl_args:
+            for hookimpl_argname in hookimpl.argnames:
+                argname_warning = hook.spec.warn_on_impl_args.get(hookimpl_argname)
+                if argname_warning is not None:
+                    _warn_for_function(argname_warning, hookimpl.function)
+
+        if (
+            hookimpl.wrapper or hookimpl.hookwrapper
+        ) and not inspect.isgeneratorfunction(hookimpl.function):
+            raise PluginValidationError(
+                hookimpl.plugin,
+                "Plugin %r for hook %r\nhookimpl definition: %s\n"
+                "Declared as wrapper=True or hookwrapper=True "
+                "but function is not a generator function"
+                % (hookimpl.plugin_name, hook.name, _formatdef(hookimpl.function)),
+            )
+
+        if hookimpl.wrapper and hookimpl.hookwrapper:
+            raise PluginValidationError(
+                hookimpl.plugin,
+                "Plugin %r for hook %r\nhookimpl definition: %s\n"
+                "The wrapper=True and hookwrapper=True options are mutually exclusive"
+                % (hookimpl.plugin_name, hook.name, _formatdef(hookimpl.function)),
+            )
+
+    def check_pending(self) -> None:
+        """Verify that all hooks which have not been verified against a
+        hook specification are optional, otherwise raise
+        :exc:`PluginValidationError`."""
+        for name in self.hook.__dict__:
+            if name[0] != "_":
+                hook: HookCaller = getattr(self.hook, name)
+                if not hook.has_spec():
+                    for hookimpl in hook.get_hookimpls():
+                        if not hookimpl.optionalhook:
+                            raise PluginValidationError(
+                                hookimpl.plugin,
+                                "unknown hook %r in plugin %r"
+                                % (name, hookimpl.plugin),
+                            )
+
+    def load_setuptools_entrypoints(self, group: str, name: str | None = None) -> int:
+        """Load modules from querying the specified setuptools ``group``.
+
+        :param group:
+            Entry point group to load plugins.
+        :param name:
+            If given, loads only plugins with the given ``name``.
+
+        :return:
+            The number of plugins loaded by this call.
+        """
+        import importlib.metadata
+
+        count = 0
+        for dist in list(importlib.metadata.distributions()):
+            for ep in dist.entry_points:
+                if (
+                    ep.group != group
+                    or (name is not None and ep.name != name)
+                    # already registered
+                    or self.get_plugin(ep.name)
+                    or self.is_blocked(ep.name)
+                ):
+                    continue
+                plugin = ep.load()
+                self.register(plugin, name=ep.name)
+                self._plugin_distinfo.append((plugin, DistFacade(dist)))
+                count += 1
+        return count
+
+    def list_plugin_distinfo(self) -> list[tuple[_Plugin, DistFacade]]:
+        """Return a list of (plugin, distinfo) pairs for all
+        setuptools-registered plugins."""
+        return list(self._plugin_distinfo)
+
+    def list_name_plugin(self) -> list[tuple[str, _Plugin]]:
+        """Return a list of (name, plugin) pairs for all registered plugins."""
+        return list(self._name2plugin.items())
+
+    def get_hookcallers(self, plugin: _Plugin) -> list[HookCaller] | None:
+        """Get all hook callers for the specified plugin.
+
+        :returns:
+            The hook callers, or ``None`` if ``plugin`` is not registered in
+            this plugin manager.
+        """
+        if self.get_name(plugin) is None:
+            return None
+        hookcallers = []
+        for hookcaller in self.hook.__dict__.values():
+            for hookimpl in hookcaller.get_hookimpls():
+                if hookimpl.plugin is plugin:
+                    hookcallers.append(hookcaller)
+        return hookcallers
+
+    def add_hookcall_monitoring(
+        self, before: _BeforeTrace, after: _AfterTrace
+    ) -> Callable[[], None]:
+        """Add before/after tracing functions for all hooks.
+
+        Returns an undo function which, when called, removes the added tracers.
+
+        ``before(hook_name, hook_impls, kwargs)`` will be called ahead
+        of all hook calls and receive a hookcaller instance, a list
+        of HookImpl instances and the keyword arguments for the hook call.
+
+        ``after(outcome, hook_name, hook_impls, kwargs)`` receives the
+        same arguments as ``before`` but also a :class:`~pluggy.Result` object
+        which represents the result of the overall hook call.
+        """
+        oldcall = self._inner_hookexec
+
+        def traced_hookexec(
+            hook_name: str,
+            hook_impls: Sequence[HookImpl],
+            caller_kwargs: Mapping[str, object],
+            firstresult: bool,
+        ) -> object | list[object]:
+            before(hook_name, hook_impls, caller_kwargs)
+            outcome = Result.from_call(
+                lambda: oldcall(hook_name, hook_impls, caller_kwargs, firstresult)
+            )
+            after(outcome, hook_name, hook_impls, caller_kwargs)
+            return outcome.get_result()
+
+        self._inner_hookexec = traced_hookexec
+
+        def undo() -> None:
+            self._inner_hookexec = oldcall
+
+        return undo
+
+    def enable_tracing(self) -> Callable[[], None]:
+        """Enable tracing of hook calls.
+
+        Returns an undo function which, when called, removes the added tracing.
+        """
+        hooktrace = self.trace.root.get("hook")
+
+        def before(
+            hook_name: str, methods: Sequence[HookImpl], kwargs: Mapping[str, object]
+        ) -> None:
+            hooktrace.root.indent += 1
+            hooktrace(hook_name, kwargs)
+
+        def after(
+            outcome: Result[object],
+            hook_name: str,
+            methods: Sequence[HookImpl],
+            kwargs: Mapping[str, object],
+        ) -> None:
+            if outcome.exception is None:
+                hooktrace("finish", hook_name, "-->", outcome.get_result())
+            hooktrace.root.indent -= 1
+
+        return self.add_hookcall_monitoring(before, after)
+
+    def subset_hook_caller(
+        self, name: str, remove_plugins: Iterable[_Plugin]
+    ) -> HookCaller:
+        """Return a proxy :class:`~pluggy.HookCaller` instance for the named
+        method which manages calls to all registered plugins except the ones
+        from remove_plugins."""
+        orig: HookCaller = getattr(self.hook, name)
+        plugins_to_remove = {plug for plug in remove_plugins if hasattr(plug, name)}
+        if plugins_to_remove:
+            return _SubsetHookCaller(orig, plugins_to_remove)
+        return orig
+
+
+def _formatdef(func: Callable[..., object]) -> str:
+    return f"{func.__name__}{inspect.signature(func)}"
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_result.py b/backend/venv/lib/python3.12/site-packages/pluggy/_result.py
new file mode 100644
index 00000000..f9a081c4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_result.py
@@ -0,0 +1,104 @@
+"""
+Hook wrapper "result" utilities.
+"""
+
+from __future__ import annotations
+
+from types import TracebackType
+from typing import Callable
+from typing import cast
+from typing import final
+from typing import Generic
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+
+
+_ExcInfo = Tuple[Type[BaseException], BaseException, Optional[TracebackType]]
+ResultType = TypeVar("ResultType")
+
+
+class HookCallError(Exception):
+    """Hook was called incorrectly."""
+
+
+@final
+class Result(Generic[ResultType]):
+    """An object used to inspect and set the result in a :ref:`hook wrapper
+    <hookwrappers>`."""
+
+    __slots__ = ("_result", "_exception")
+
+    def __init__(
+        self,
+        result: ResultType | None,
+        exception: BaseException | None,
+    ) -> None:
+        """:meta private:"""
+        self._result = result
+        self._exception = exception
+
+    @property
+    def excinfo(self) -> _ExcInfo | None:
+        """:meta private:"""
+        exc = self._exception
+        if exc is None:
+            return None
+        else:
+            return (type(exc), exc, exc.__traceback__)
+
+    @property
+    def exception(self) -> BaseException | None:
+        """:meta private:"""
+        return self._exception
+
+    @classmethod
+    def from_call(cls, func: Callable[[], ResultType]) -> Result[ResultType]:
+        """:meta private:"""
+        __tracebackhide__ = True
+        result = exception = None
+        try:
+            result = func()
+        except BaseException as exc:
+            exception = exc
+        return cls(result, exception)
+
+    def force_result(self, result: ResultType) -> None:
+        """Force the result(s) to ``result``.
+
+        If the hook was marked as a ``firstresult`` a single value should
+        be set, otherwise set a (modified) list of results. Any exceptions
+        found during invocation will be deleted.
+
+        This overrides any previous result or exception.
+        """
+        self._result = result
+        self._exception = None
+
+    def force_exception(self, exception: BaseException) -> None:
+        """Force the result to fail with ``exception``.
+
+        This overrides any previous result or exception.
+
+        .. versionadded:: 1.1.0
+        """
+        self._result = None
+        self._exception = exception
+
+    def get_result(self) -> ResultType:
+        """Get the result(s) for this hook call.
+
+        If the hook was marked as a ``firstresult`` only a single value
+        will be returned, otherwise a list of results.
+        """
+        __tracebackhide__ = True
+        exc = self._exception
+        if exc is None:
+            return cast(ResultType, self._result)
+        else:
+            raise exc.with_traceback(exc.__traceback__)
+
+
+# Historical name (pluggy<=1.2), kept for backward compatibility.
+_Result = Result
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_tracing.py b/backend/venv/lib/python3.12/site-packages/pluggy/_tracing.py
new file mode 100644
index 00000000..cd238ad7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_tracing.py
@@ -0,0 +1,73 @@
+"""
+Tracing utils
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Sequence
+from typing import Tuple
+
+
+_Writer = Callable[[str], object]
+_Processor = Callable[[Tuple[str, ...], Tuple[Any, ...]], object]
+
+
+class TagTracer:
+    def __init__(self) -> None:
+        self._tags2proc: dict[tuple[str, ...], _Processor] = {}
+        self._writer: _Writer | None = None
+        self.indent = 0
+
+    def get(self, name: str) -> TagTracerSub:
+        return TagTracerSub(self, (name,))
+
+    def _format_message(self, tags: Sequence[str], args: Sequence[object]) -> str:
+        if isinstance(args[-1], dict):
+            extra = args[-1]
+            args = args[:-1]
+        else:
+            extra = {}
+
+        content = " ".join(map(str, args))
+        indent = "  " * self.indent
+
+        lines = ["{}{} [{}]\n".format(indent, content, ":".join(tags))]
+
+        for name, value in extra.items():
+            lines.append(f"{indent}    {name}: {value}\n")
+
+        return "".join(lines)
+
+    def _processmessage(self, tags: tuple[str, ...], args: tuple[object, ...]) -> None:
+        if self._writer is not None and args:
+            self._writer(self._format_message(tags, args))
+        try:
+            processor = self._tags2proc[tags]
+        except KeyError:
+            pass
+        else:
+            processor(tags, args)
+
+    def setwriter(self, writer: _Writer | None) -> None:
+        self._writer = writer
+
+    def setprocessor(self, tags: str | tuple[str, ...], processor: _Processor) -> None:
+        if isinstance(tags, str):
+            tags = tuple(tags.split(":"))
+        else:
+            assert isinstance(tags, tuple)
+        self._tags2proc[tags] = processor
+
+
+class TagTracerSub:
+    def __init__(self, root: TagTracer, tags: tuple[str, ...]) -> None:
+        self.root = root
+        self.tags = tags
+
+    def __call__(self, *args: object) -> None:
+        self.root._processmessage(self.tags, args)
+
+    def get(self, name: str) -> TagTracerSub:
+        return self.__class__(self.root, self.tags + (name,))
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_version.py b/backend/venv/lib/python3.12/site-packages/pluggy/_version.py
new file mode 100644
index 00000000..c565007e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '1.5.0'
+__version_tuple__ = version_tuple = (1, 5, 0)
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/_warnings.py b/backend/venv/lib/python3.12/site-packages/pluggy/_warnings.py
new file mode 100644
index 00000000..6356c770
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pluggy/_warnings.py
@@ -0,0 +1,27 @@
+from typing import final
+
+
+class PluggyWarning(UserWarning):
+    """Base class for all warnings emitted by pluggy."""
+
+    __module__ = "pluggy"
+
+
+@final
+class PluggyTeardownRaisedWarning(PluggyWarning):
+    """A plugin raised an exception during an :ref:`old-style hookwrapper
+    <old_style_hookwrappers>` teardown.
+
+    Such exceptions are not handled by pluggy, and may cause subsequent
+    teardowns to be executed at unexpected times, or be skipped entirely.
+
+    This is an issue in the plugin implementation.
+
+    If the exception is unintended, fix the underlying cause.
+
+    If the exception is intended, switch to :ref:`new-style hook wrappers
+    <hookwrappers>`, or use :func:`result.force_exception()
+    <pluggy.Result.force_exception>` to set the exception instead of raising.
+    """
+
+    __module__ = "pluggy"
diff --git a/backend/venv/lib/python3.12/site-packages/pluggy/py.typed b/backend/venv/lib/python3.12/site-packages/pluggy/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/py.py b/backend/venv/lib/python3.12/site-packages/py.py
new file mode 100644
index 00000000..5c661e66
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/py.py
@@ -0,0 +1,15 @@
+# shim for pylib going away
+# if pylib is installed this file will get skipped
+# (`py/__init__.py` has higher precedence)
+from __future__ import annotations
+
+import sys
+
+import _pytest._py.error as error
+import _pytest._py.path as path
+
+
+sys.modules["py.error"] = error
+sys.modules["py.path"] = path
+
+__all__ = ["error", "path"]
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/AUTHORS b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/AUTHORS
new file mode 100644
index 00000000..e25156cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/AUTHORS
@@ -0,0 +1,472 @@
+Holger Krekel, holger at merlinux eu
+merlinux GmbH, Germany, office at merlinux eu
+
+Contributors include::
+
+Aaron Coleman
+Abdeali JK
+Abdelrahman Elbehery
+Abhijeet Kasurde
+Adam Johnson
+Adam Stewart
+Adam Uhlir
+Ahn Ki-Wook
+Akhilesh Ramakrishnan
+Akiomi Kamakura
+Alan Velasco
+Alessio Izzo
+Alex Jones
+Alex Lambson
+Alexander Johnson
+Alexander King
+Alexei Kozlenok
+Alice Purcell
+Allan Feldman
+Aly Sivji
+Amir Elkess
+Anatoly Bubenkoff
+Anders Hovmöller
+Andras Mitzki
+Andras Tim
+Andrea Cimatoribus
+Andreas Motl
+Andreas Zeidler
+Andrew Shapton
+Andrey Paramonov
+Andrzej Klajnert
+Andrzej Ostrowski
+Andy Freeland
+Anita Hammer
+Anthon van der Neut
+Anthony Shaw
+Anthony Sottile
+Anton Grinevich
+Anton Lodder
+Antony Lee
+Arel Cordero
+Arias Emmanuel
+Ariel Pillemer
+Armin Rigo
+Aron Coyle
+Aron Curzon
+Arthur Richard
+Ashish Kurmi
+Aviral Verma
+Aviv Palivoda
+Babak Keyvani
+Barney Gale
+Ben Brown
+Ben Gartner
+Ben Leith
+Ben Webb
+Benjamin Peterson
+Benjamin Schubert
+Bernard Pratz
+Bo Wu
+Bob Ippolito
+Brian Dorsey
+Brian Larsen
+Brian Maissy
+Brian Okken
+Brianna Laugher
+Bruno Oliveira
+Cal Jacobson
+Cal Leeming
+Carl Friedrich Bolz
+Carlos Jenkins
+Ceridwen
+Charles Cloud
+Charles Machalow
+Charnjit SiNGH (CCSJ)
+Cheuk Ting Ho
+Chris Mahoney
+Chris Lamb
+Chris NeJame
+Chris Rose
+Chris Wheeler
+Christian Boelsen
+Christian Clauss
+Christian Fetzer
+Christian Neumüller
+Christian Theunert
+Christian Tismer
+Christine Mecklenborg
+Christoph Buelter
+Christopher Dignam
+Christopher Gilling
+Claire Cecil
+Claudio Madotto
+Clément M.T. Robert
+Cornelius Riemenschneider
+CrazyMerlyn
+Cristian Vera
+Cyrus Maden
+Damian Skrzypczak
+Daniel Grana
+Daniel Hahler
+Daniel Miller
+Daniel Nuri
+Daniel Sánchez Castelló
+Daniel Valenzuela Zenteno
+Daniel Wandschneider
+Daniele Procida
+Danielle Jenkins
+Daniil Galiev
+Dave Hunt
+David Díaz-Barquero
+David Mohr
+David Paul Röthlisberger
+David Peled
+David Szotten
+David Vierra
+Daw-Ran Liou
+Debi Mishra
+Denis Kirisov
+Denivy Braiam Rück
+Dheeraj C K
+Dhiren Serai
+Diego Russo
+Dmitry Dygalo
+Dmitry Pribysh
+Dominic Mortlock
+Duncan Betts
+Edison Gustavo Muenz
+Edoardo Batini
+Edson Tadeu M. Manoel
+Eduardo Schettino
+Edward Haigh
+Eero Vaher
+Eli Boyarski
+Elizaveta Shashkova
+Éloi Rivard
+Endre Galaczi
+Eric Hunsberger
+Eric Liu
+Eric Siegerman
+Eric Yuan
+Erik Aronesty
+Erik Hasse
+Erik M. Bray
+Evan Kepner
+Evgeny Seliverstov
+Fabian Sturm
+Fabien Zarifian
+Fabio Zadrozny
+Farbod Ahmadian
+faph
+Felix Hofstätter
+Felix Nieuwenhuizen
+Feng Ma
+Florian Bruhin
+Florian Dahlitz
+Floris Bruynooghe
+Frank Hoffmann
+Fraser Stark
+Gabriel Landau
+Gabriel Reis
+Garvit Shubham
+Gene Wood
+George Kussumoto
+Georgy Dyuldin
+Gergely Kalmár
+Gleb Nikonorov
+Graeme Smecher
+Graham Horler
+Greg Price
+Gregory Lee
+Grig Gheorghiu
+Grigorii Eremeev (budulianin)
+Guido Wesdorp
+Guoqiang Zhang
+Harald Armin Massa
+Harshna
+Henk-Jaap Wagenaar
+Holger Kohr
+Hugo van Kemenade
+Hui Wang (coldnight)
+Ian Bicking
+Ian Lesperance
+Ilya Konstantinov
+Ionuț Turturică
+Isaac Virshup
+Israel Fruchter
+Itxaso Aizpurua
+Iwan Briquemont
+Jaap Broekhuizen
+Jake VanderPlas
+Jakob van Santen
+Jakub Mitoraj
+James Bourbeau
+James Frost
+Jan Balster
+Janne Vanhala
+Jason R. Coombs
+Javier Domingo Cansino
+Javier Romero
+Jeff Rackauckas
+Jeff Widman
+Jenni Rinker
+Jens Tröger
+John Eddie Ayson
+John Litborn
+John Towler
+Jon Parise
+Jon Sonesen
+Jonas Obrist
+Jordan Guymon
+Jordan Moldow
+Jordan Speicher
+Joseph Hunkeler
+Joseph Sawaya
+Josh Karpel
+Joshua Bronson
+Jurko Gospodnetić
+Justice Ndou
+Justyna Janczyszyn
+Kale Kundert
+Kamran Ahmad
+Kenny Y
+Karl O. Pinc
+Karthikeyan Singaravelan
+Katarzyna Jachim
+Katarzyna Król
+Katerina Koukiou
+Keri Volans
+Kevin C
+Kevin Cox
+Kevin Hierro Carrasco
+Kevin J. Foley
+Kian Eliasi
+Kian-Meng Ang
+Kodi B. Arfer
+Kojo Idrissa
+Kostis Anagnostopoulos
+Kristoffer Nordström
+Kyle Altendorf
+Lawrence Mitchell
+Lee Kamentsky
+Leonardus Chen
+Lev Maximov
+Levon Saldamli
+Lewis Cowles
+Llandy Riveron Del Risco
+Loic Esteve
+lovetheguitar
+Lukas Bednar
+Luke Murphy
+Maciek Fijalkowski
+Maho
+Maik Figura
+Mandeep Bhutani
+Manuel Krebber
+Marc Mueller
+Marc Schlaich
+Marcelo Duarte Trevisani
+Marcin Bachry
+Marc Bresson
+Marco Gorelli
+Mark Abramowitz
+Mark Dickinson
+Mark Vong
+Marko Pacak
+Markus Unterwaditzer
+Martijn Faassen
+Martin Altmayer
+Martin K. Scherer
+Martin Prusse
+Mathieu Clabaut
+Matt Bachmann
+Matt Duck
+Matt Williams
+Matthias Hafner
+Maxim Filipenko
+Maximilian Cosmo Sitter
+mbyt
+Michael Aquilina
+Michael Birtwell
+Michael Droettboom
+Michael Goerz
+Michael Krebs
+Michael Seifert
+Michael Vogt
+Michal Wajszczuk
+Michał Górny
+Michał Zięba
+Mickey Pashov
+Mihai Capotă
+Mihail Milushev
+Mike Hoyle (hoylemd)
+Mike Lundy
+Milan Lesnek
+Miro Hrončok
+mrbean-bremen
+Nathan Goldbaum
+Nathaniel Compton
+Nathaniel Waisbrot
+Nauman Ahmed
+Ned Batchelder
+Neil Martin
+Neven Mundar
+Nicholas Devenish
+Nicholas Murphy
+Niclas Olofsson
+Nicolas Delaby
+Nicolas Simonds
+Nico Vidal
+Nikolay Kondratyev
+Nipunn Koorapati
+Oleg Pidsadnyi
+Oleg Sushchenko
+Olga Matoula
+Oliver Bestwalter
+Omar Kohl
+Omer Hadari
+Ondřej Súkup
+Oscar Benjamin
+Parth Patel
+Patrick Hayes
+Patrick Lannigan
+Paul Müller
+Paul Reece
+Pauli Virtanen
+Pavel Karateev
+Paweł Adamczak
+Pedro Algarvio
+Petter Strandmark
+Philipp Loose
+Pierre Sassoulas
+Pieter Mulder
+Piotr Banaszkiewicz
+Piotr Helm
+Poulami Sau
+Prakhar Gurunani
+Prashant Anand
+Prashant Sharma
+Pulkit Goyal
+Punyashloka Biswal
+Quentin Pradet
+q0w
+Ralf Schmitt
+Ralph Giles
+Ram Rachum
+Ran Benita
+Raphael Castaneda
+Raphael Pierzina
+Rafal Semik
+Raquel Alegre
+Ravi Chandra
+Reagan Lee
+Rob Arrow
+Robert Holt
+Roberto Aldera
+Roberto Polli
+Roland Puntaier
+Romain Dorgueil
+Roman Bolshakov
+Ronny Pfannschmidt
+Ross Lawley
+Ruaridh Williamson
+Russel Winder
+Russell Martin
+Ryan Puddephatt
+Ryan Wooden
+Sadra Barikbin
+Saiprasad Kale
+Samuel Colvin
+Samuel Dion-Girardeau
+Samuel Jirovec
+Samuel Searles-Bryant
+Samuel Therrien (Avasam)
+Samuele Pedroni
+Sanket Duthade
+Sankt Petersbug
+Saravanan Padmanaban
+Sean Malloy
+Segev Finer
+Serhii Mozghovyi
+Seth Junot
+Shantanu Jain
+Sharad Nair
+Shubham Adep
+Simon Blanchard
+Simon Gomizelj
+Simon Holesch
+Simon Kerr
+Skylar Downes
+Srinivas Reddy Thatiparthy
+Stefaan Lippens
+Stefan Farmbauer
+Stefan Scherfke
+Stefan Zimmermann
+Stefanie Molin
+Stefano Taschini
+Steffen Allner
+Stephan Obermann
+Sven-Hendrik Haase
+Sviatoslav Sydorenko
+Sylvain Marié
+Tadek Teleżyński
+Takafumi Arakaki
+Taneli Hukkinen
+Tanvi Mehta
+Tanya Agarwal
+Tarcisio Fischer
+Tareq Alayan
+Tatiana Ovary
+Ted Xiao
+Terje Runde
+Thomas Grainger
+Thomas Hisch
+Tianyu Dongfang
+Tim Hoffmann
+Tim Strazny
+TJ Bruno
+Tobias Diez
+Tom Dalton
+Tom Viner
+Tomáš Gavenčiak
+Tomer Keren
+Tony Narlock
+Tor Colvin
+Trevor Bekolay
+Tushar Sadhwani
+Tyler Goodlet
+Tyler Smart
+Tzu-ping Chung
+Vasily Kuznetsov
+Victor Maryama
+Victor Rodriguez
+Victor Uriarte
+Vidar T. Fauske
+Vijay Arora
+Virendra Patil
+Virgil Dupras
+Vitaly Lashmanov
+Vivaan Verma
+Vlad Dragos
+Vlad Radziuk
+Vladyslav Rachek
+Volodymyr Kochetkov
+Volodymyr Piskun
+Wei Lin
+Wil Cooley
+William Lee
+Wim Glenn
+Wouter van Ackooy
+Xixi Zhao
+Xuan Luong
+Xuecong Liao
+Yannick Péroux
+Yao Xiao
+Yoav Caspi
+Yuliang Shao
+Yusuke Kadowaki
+Yutian Li
+Yuval Shimon
+Zac Hatfield-Dodds
+Zach Snicker
+Zachary Kneupper
+Zachary OBrien
+Zhouxin Qiu
+Zoltán Máté
+Zsolt Cserna
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/LICENSE
new file mode 100644
index 00000000..c3f1657f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/LICENSE
@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2004 Holger Krekel and others
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+of the Software, and to permit persons to whom the Software is furnished to do
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/METADATA
new file mode 100644
index 00000000..b8f89610
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/METADATA
@@ -0,0 +1,212 @@
+Metadata-Version: 2.1
+Name: pytest
+Version: 8.3.4
+Summary: pytest: simple powerful testing with Python
+Author: Holger Krekel, Bruno Oliveira, Ronny Pfannschmidt, Floris Bruynooghe, Brianna Laugher, Florian Bruhin, Others (See AUTHORS)
+License: MIT
+Project-URL: Changelog, https://docs.pytest.org/en/stable/changelog.html
+Project-URL: Homepage, https://docs.pytest.org/en/latest/
+Project-URL: Source, https://github.com/pytest-dev/pytest
+Project-URL: Tracker, https://github.com/pytest-dev/pytest/issues
+Project-URL: Twitter, https://twitter.com/pytestdotorg
+Keywords: test,unittest
+Classifier: Development Status :: 6 - Mature
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX
+Classifier: Operating System :: Unix
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development :: Testing
+Classifier: Topic :: Utilities
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+License-File: AUTHORS
+Requires-Dist: colorama; sys_platform == "win32"
+Requires-Dist: exceptiongroup>=1.0.0rc8; python_version < "3.11"
+Requires-Dist: iniconfig
+Requires-Dist: packaging
+Requires-Dist: pluggy<2,>=1.5
+Requires-Dist: tomli>=1; python_version < "3.11"
+Provides-Extra: dev
+Requires-Dist: argcomplete; extra == "dev"
+Requires-Dist: attrs>=19.2; extra == "dev"
+Requires-Dist: hypothesis>=3.56; extra == "dev"
+Requires-Dist: mock; extra == "dev"
+Requires-Dist: pygments>=2.7.2; extra == "dev"
+Requires-Dist: requests; extra == "dev"
+Requires-Dist: setuptools; extra == "dev"
+Requires-Dist: xmlschema; extra == "dev"
+
+.. image:: https://github.com/pytest-dev/pytest/raw/main/doc/en/img/pytest_logo_curves.svg
+   :target: https://docs.pytest.org/en/stable/
+   :align: center
+   :height: 200
+   :alt: pytest
+
+
+------
+
+.. image:: https://img.shields.io/pypi/v/pytest.svg
+    :target: https://pypi.org/project/pytest/
+
+.. image:: https://img.shields.io/conda/vn/conda-forge/pytest.svg
+    :target: https://anaconda.org/conda-forge/pytest
+
+.. image:: https://img.shields.io/pypi/pyversions/pytest.svg
+    :target: https://pypi.org/project/pytest/
+
+.. image:: https://codecov.io/gh/pytest-dev/pytest/branch/main/graph/badge.svg
+    :target: https://codecov.io/gh/pytest-dev/pytest
+    :alt: Code coverage Status
+
+.. image:: https://github.com/pytest-dev/pytest/actions/workflows/test.yml/badge.svg
+    :target: https://github.com/pytest-dev/pytest/actions?query=workflow%3Atest
+
+.. image:: https://results.pre-commit.ci/badge/github/pytest-dev/pytest/main.svg
+   :target: https://results.pre-commit.ci/latest/github/pytest-dev/pytest/main
+   :alt: pre-commit.ci status
+
+.. image:: https://www.codetriage.com/pytest-dev/pytest/badges/users.svg
+    :target: https://www.codetriage.com/pytest-dev/pytest
+
+.. image:: https://readthedocs.org/projects/pytest/badge/?version=latest
+    :target: https://pytest.readthedocs.io/en/latest/?badge=latest
+    :alt: Documentation Status
+
+.. image:: https://img.shields.io/badge/Discord-pytest--dev-blue
+    :target: https://discord.com/invite/pytest-dev
+    :alt: Discord
+
+.. image:: https://img.shields.io/badge/Libera%20chat-%23pytest-orange
+    :target: https://web.libera.chat/#pytest
+    :alt: Libera chat
+
+
+The ``pytest`` framework makes it easy to write small tests, yet
+scales to support complex functional testing for applications and libraries.
+
+An example of a simple test:
+
+.. code-block:: python
+
+    # content of test_sample.py
+    def inc(x):
+        return x + 1
+
+
+    def test_answer():
+        assert inc(3) == 5
+
+
+To execute it::
+
+    $ pytest
+    ============================= test session starts =============================
+    collected 1 items
+
+    test_sample.py F
+
+    ================================== FAILURES ===================================
+    _________________________________ test_answer _________________________________
+
+        def test_answer():
+    >       assert inc(3) == 5
+    E       assert 4 == 5
+    E        +  where 4 = inc(3)
+
+    test_sample.py:5: AssertionError
+    ========================== 1 failed in 0.04 seconds ===========================
+
+
+Due to ``pytest``'s detailed assertion introspection, only plain ``assert`` statements are used. See `getting-started <https://docs.pytest.org/en/stable/getting-started.html#our-first-test-run>`_ for more examples.
+
+
+Features
+--------
+
+- Detailed info on failing `assert statements <https://docs.pytest.org/en/stable/how-to/assert.html>`_ (no need to remember ``self.assert*`` names)
+
+- `Auto-discovery
+  <https://docs.pytest.org/en/stable/explanation/goodpractices.html#python-test-discovery>`_
+  of test modules and functions
+
+- `Modular fixtures <https://docs.pytest.org/en/stable/explanation/fixtures.html>`_ for
+  managing small or parametrized long-lived test resources
+
+- Can run `unittest <https://docs.pytest.org/en/stable/how-to/unittest.html>`_ (or trial)
+  test suites out of the box
+
+- Python 3.8+ or PyPy3
+
+- Rich plugin architecture, with over 1300+ `external plugins <https://docs.pytest.org/en/latest/reference/plugin_list.html>`_ and thriving community
+
+
+Documentation
+-------------
+
+For full documentation, including installation, tutorials and PDF documents, please see https://docs.pytest.org/en/stable/.
+
+
+Bugs/Requests
+-------------
+
+Please use the `GitHub issue tracker <https://github.com/pytest-dev/pytest/issues>`_ to submit bugs or request features.
+
+
+Changelog
+---------
+
+Consult the `Changelog <https://docs.pytest.org/en/stable/changelog.html>`__ page for fixes and enhancements of each version.
+
+
+Support pytest
+--------------
+
+`Open Collective`_ is an online funding platform for open and transparent communities.
+It provides tools to raise money and share your finances in full transparency.
+
+It is the platform of choice for individuals and companies that want to make one-time or
+monthly donations directly to the project.
+
+See more details in the `pytest collective`_.
+
+.. _Open Collective: https://opencollective.com
+.. _pytest collective: https://opencollective.com/pytest
+
+
+pytest for enterprise
+---------------------
+
+Available as part of the Tidelift Subscription.
+
+The maintainers of pytest and thousands of other packages are working with Tidelift to deliver commercial support and
+maintenance for the open source dependencies you use to build your applications.
+Save time, reduce risk, and improve code health, while paying the maintainers of the exact dependencies you use.
+
+`Learn more. <https://tidelift.com/subscription/pkg/pypi-pytest?utm_source=pypi-pytest&utm_medium=referral&utm_campaign=enterprise&utm_term=repo>`_
+
+Security
+^^^^^^^^
+
+pytest has never been associated with a security vulnerability, but in any case, to report a
+security vulnerability please use the `Tidelift security contact <https://tidelift.com/security>`_.
+Tidelift will coordinate the fix and disclosure.
+
+
+License
+-------
+
+Copyright Holger Krekel and others, 2004.
+
+Distributed under the terms of the `MIT`_ license, pytest is free and open source software.
+
+.. _`MIT`: https://github.com/pytest-dev/pytest/blob/main/LICENSE
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/RECORD
new file mode 100644
index 00000000..dca47e1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/RECORD
@@ -0,0 +1,155 @@
+../../../bin/py.test,sha256=p_cNJhPYF2W5QdsdFP_FO8q-6WKXkUBXIk-Vtrl2nKg,278
+../../../bin/pytest,sha256=p_cNJhPYF2W5QdsdFP_FO8q-6WKXkUBXIk-Vtrl2nKg,278
+__pycache__/py.cpython-312.pyc,,
+_pytest/__init__.py,sha256=4IdRJhnW5XG2KlaJkOxn5_TC9WeQ5tXDSF7tbb4vEso,391
+_pytest/__pycache__/__init__.cpython-312.pyc,,
+_pytest/__pycache__/_argcomplete.cpython-312.pyc,,
+_pytest/__pycache__/_version.cpython-312.pyc,,
+_pytest/__pycache__/cacheprovider.cpython-312.pyc,,
+_pytest/__pycache__/capture.cpython-312.pyc,,
+_pytest/__pycache__/compat.cpython-312.pyc,,
+_pytest/__pycache__/debugging.cpython-312.pyc,,
+_pytest/__pycache__/deprecated.cpython-312.pyc,,
+_pytest/__pycache__/doctest.cpython-312.pyc,,
+_pytest/__pycache__/faulthandler.cpython-312.pyc,,
+_pytest/__pycache__/fixtures.cpython-312.pyc,,
+_pytest/__pycache__/freeze_support.cpython-312.pyc,,
+_pytest/__pycache__/helpconfig.cpython-312.pyc,,
+_pytest/__pycache__/hookspec.cpython-312.pyc,,
+_pytest/__pycache__/junitxml.cpython-312.pyc,,
+_pytest/__pycache__/legacypath.cpython-312.pyc,,
+_pytest/__pycache__/logging.cpython-312.pyc,,
+_pytest/__pycache__/main.cpython-312.pyc,,
+_pytest/__pycache__/monkeypatch.cpython-312.pyc,,
+_pytest/__pycache__/nodes.cpython-312.pyc,,
+_pytest/__pycache__/outcomes.cpython-312.pyc,,
+_pytest/__pycache__/pastebin.cpython-312.pyc,,
+_pytest/__pycache__/pathlib.cpython-312.pyc,,
+_pytest/__pycache__/pytester.cpython-312.pyc,,
+_pytest/__pycache__/pytester_assertions.cpython-312.pyc,,
+_pytest/__pycache__/python.cpython-312.pyc,,
+_pytest/__pycache__/python_api.cpython-312.pyc,,
+_pytest/__pycache__/python_path.cpython-312.pyc,,
+_pytest/__pycache__/recwarn.cpython-312.pyc,,
+_pytest/__pycache__/reports.cpython-312.pyc,,
+_pytest/__pycache__/runner.cpython-312.pyc,,
+_pytest/__pycache__/scope.cpython-312.pyc,,
+_pytest/__pycache__/setuponly.cpython-312.pyc,,
+_pytest/__pycache__/setupplan.cpython-312.pyc,,
+_pytest/__pycache__/skipping.cpython-312.pyc,,
+_pytest/__pycache__/stash.cpython-312.pyc,,
+_pytest/__pycache__/stepwise.cpython-312.pyc,,
+_pytest/__pycache__/terminal.cpython-312.pyc,,
+_pytest/__pycache__/threadexception.cpython-312.pyc,,
+_pytest/__pycache__/timing.cpython-312.pyc,,
+_pytest/__pycache__/tmpdir.cpython-312.pyc,,
+_pytest/__pycache__/unittest.cpython-312.pyc,,
+_pytest/__pycache__/unraisableexception.cpython-312.pyc,,
+_pytest/__pycache__/warning_types.cpython-312.pyc,,
+_pytest/__pycache__/warnings.cpython-312.pyc,,
+_pytest/_argcomplete.py,sha256=gh0pna66p4LVb2D8ST4568WGxvdInGT43m6slYhqNqU,3776
+_pytest/_code/__init__.py,sha256=5h7R-LFINKh7p8QR1HgdjvSGo1ysVJz28MQ9h7ipHK4,521
+_pytest/_code/__pycache__/__init__.cpython-312.pyc,,
+_pytest/_code/__pycache__/code.cpython-312.pyc,,
+_pytest/_code/__pycache__/source.cpython-312.pyc,,
+_pytest/_code/code.py,sha256=umPdqLxq8UgWKAItTEvF6ZOq5dF65mzCJHFaZHzTNGY,50133
+_pytest/_code/source.py,sha256=2w9OZFOrRpiVaD_UdUS1T2XC7c2Is2GZn0iQy-lZfwk,7278
+_pytest/_io/__init__.py,sha256=pkLF29VEFr6Dlr3eOtJL8sf47RLFt1Jf4X1DZBPlYmc,190
+_pytest/_io/__pycache__/__init__.cpython-312.pyc,,
+_pytest/_io/__pycache__/pprint.cpython-312.pyc,,
+_pytest/_io/__pycache__/saferepr.cpython-312.pyc,,
+_pytest/_io/__pycache__/terminalwriter.cpython-312.pyc,,
+_pytest/_io/__pycache__/wcwidth.cpython-312.pyc,,
+_pytest/_io/pprint.py,sha256=BCe8K7Zc0drYC5_JKZBBMVrhK84ARlmPpk9vSWPYhaE,19633
+_pytest/_io/saferepr.py,sha256=Hhx5F-75iz03hdk-WO86Bmy9RBuRHsuJj-YUzozfrgo,4082
+_pytest/_io/terminalwriter.py,sha256=dQ07zJ1-vlpFqWBBu_c0cHxT0yXcGSu7o7LxDCEyB3s,9319
+_pytest/_io/wcwidth.py,sha256=cUEJ74UhweICwbKvU2q6noZcNgD0QlBEB9CfakGYaqA,1289
+_pytest/_py/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+_pytest/_py/__pycache__/__init__.cpython-312.pyc,,
+_pytest/_py/__pycache__/error.cpython-312.pyc,,
+_pytest/_py/__pycache__/path.cpython-312.pyc,,
+_pytest/_py/error.py,sha256=S1BRlfXSD248OFNzAuZ5O9w9W6fr2NUn0X8wYFGMNk0,3015
+_pytest/_py/path.py,sha256=Xl4UspvrwwKYNlLZDGbjhUrnD6fuBmFxxchgltmwGek,49211
+_pytest/_version.py,sha256=lMdMvJTmZKMLyCj9eyDEhOf6ttbh5vLbkANi6yBfmkc,411
+_pytest/assertion/__init__.py,sha256=lK1YNNAk1VqCK-Y5C5hMJMqJQyxQT11HuDH3w85a3Zc,6791
+_pytest/assertion/__pycache__/__init__.cpython-312.pyc,,
+_pytest/assertion/__pycache__/rewrite.cpython-312.pyc,,
+_pytest/assertion/__pycache__/truncate.cpython-312.pyc,,
+_pytest/assertion/__pycache__/util.cpython-312.pyc,,
+_pytest/assertion/rewrite.py,sha256=OltHwIVPXCPe0Ejgta_4hWjY8Xo3BIGbQEyLREfXZCA,48272
+_pytest/assertion/truncate.py,sha256=GYl5iqDXUuKQHgd_mthWl3ZjxBbDVQliBhks1Ty00kE,4459
+_pytest/assertion/util.py,sha256=6Vg5dZDuIXak9OLupvKyavLDroATgpU6ilkclc0OlJY,20265
+_pytest/cacheprovider.py,sha256=BFQVkna56tlO-v9qaNJVHIcF30hIVGrP9St_vMp4w98,22373
+_pytest/capture.py,sha256=iiu_k5_0ASbINS5_o7ZxMShGaWZFQD-y7jtU-NiJtMs,34680
+_pytest/compat.py,sha256=sPcVQwPd45GaqsgIZEbCTR04GKhkVmIfft6QnKj3hmo,11467
+_pytest/config/__init__.py,sha256=Ch5YizaRrCfMykEuZdHF0RaIyvtWvoSXL4v6E1Cu-FY,70645
+_pytest/config/__pycache__/__init__.cpython-312.pyc,,
+_pytest/config/__pycache__/argparsing.cpython-312.pyc,,
+_pytest/config/__pycache__/compat.cpython-312.pyc,,
+_pytest/config/__pycache__/exceptions.cpython-312.pyc,,
+_pytest/config/__pycache__/findpaths.cpython-312.pyc,,
+_pytest/config/argparsing.py,sha256=dNjEvFh2C34XMoiE_R7liJv5cryXUz2WR2VsxdnQdjo,20562
+_pytest/config/compat.py,sha256=-m8G4-LLezCd4KZO6JQufEz7cRDqUSOjIwCtiKWpJvY,2938
+_pytest/config/exceptions.py,sha256=lUKnOtpRqK-qNL6JfOP-8tRqpmHU34CVxguR5y0Qfbw,288
+_pytest/config/findpaths.py,sha256=h4zq5AbLaZGpkeEcD2Xg-rJimh9I5pE042qQOTZT7NM,8062
+_pytest/debugging.py,sha256=yRmmOexsaDeFky37IrD2e9svz8CWebB7L2fSUy4LvuE,13260
+_pytest/deprecated.py,sha256=sO9UiqEdy9Z-NCvDoYYA0QtafYogAb7lP5M9N_Hpnak,3147
+_pytest/doctest.py,sha256=7WJprJGYj7_9Lyr-L49wJ7q5ZwDVj1FBhA9_CX7JdLc,26255
+_pytest/faulthandler.py,sha256=dT0H-MLi62SXeiKjLQJ0EVPuxkTlNOxpWtNxA5uBJPs,3674
+_pytest/fixtures.py,sha256=I5t3pW2lHaVPbN1rAQ9sdX0a3QrpoW_U5VP-Vxejxmg,73550
+_pytest/freeze_support.py,sha256=1EfzuxPd2oV9Ira26K5J4r9ppFZjnGi-xKzsBXe8B4g,1291
+_pytest/helpconfig.py,sha256=ibnZNxKzToLmx-2ZrZKCP9t6jJvpAIlmqdf9a0rhOoI,8895
+_pytest/hookspec.py,sha256=G-wKdmV3pecpeeiIAmzgPUMr22kz-CsqSpWEM-uiamg,42825
+_pytest/junitxml.py,sha256=FnYwq0wAR4Cixzj-a9qhyulUSEpMyjX9ALbjza_We74,25574
+_pytest/legacypath.py,sha256=_l6v8akNMfTc5TAjvbc6M-_t157p9QE6-118WM0DRt8,16588
+_pytest/logging.py,sha256=QfaUUx-T0FiKBJBBb3bDllt8O8eTE7Mpigq7wvDepRc,35124
+_pytest/main.py,sha256=Oowez36UkOwJXkTRq4rVuJRRr18ItBnz_YDjgAmFCV8,37416
+_pytest/mark/__init__.py,sha256=bHORyCAsGnGJq7Tpm7A2sNQX31jwU1TgufM9DYcrTfQ,9307
+_pytest/mark/__pycache__/__init__.cpython-312.pyc,,
+_pytest/mark/__pycache__/expression.cpython-312.pyc,,
+_pytest/mark/__pycache__/structures.cpython-312.pyc,,
+_pytest/mark/expression.py,sha256=H6LmX0MWlxe0uBmuXIpQEntrLtyqIhEJv07YvA79eDQ,10152
+_pytest/mark/structures.py,sha256=6hiIR3d4zxy35Yiw961r9sYrNl-T5WS8_0auSmpdiB0,21039
+_pytest/monkeypatch.py,sha256=SKgteVJz1puqYQ3el6-ju5ZsNABqpoMUuRC6nn3tFpc,14598
+_pytest/nodes.py,sha256=Hqyplow99hb-Zz0KKzL0K3cQ0rCgDXK65vBp6ave3u8,26483
+_pytest/outcomes.py,sha256=SeW14rRKnGSt7K_NxY7HGnedoJawFHwQi2anAYYugk8,10532
+_pytest/pastebin.py,sha256=Ja1z3Z6cXNElobpwy97FiyR5DDexZrDEB6vufmNvE4o,3978
+_pytest/pathlib.py,sha256=onXoMMo4cbp-DR03XQuRimuIm_DrHNg3RujwP4Z7tic,36617
+_pytest/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+_pytest/pytester.py,sha256=-D_SNLfp_AQRMP7GOo6NsXlektiYod79pxBY-2RdUT0,61552
+_pytest/pytester_assertions.py,sha256=EIHeb1llN9TVRfBJqQnwvjKeG-KpzURNq8mtbK7vcyA,2244
+_pytest/python.py,sha256=mkJek4hqp7GMIyk6uPNWclI2dYlg78tTjymdcZViTJM,64851
+_pytest/python_api.py,sha256=KyIP4xCyd2BpNFlY-28a0E50XxMXy7HSQi6ZaCw3qZg,40122
+_pytest/python_path.py,sha256=fGP7iR_XMFRPijg4niILo44gWUWLlD635fYO5Abd6IM,745
+_pytest/recwarn.py,sha256=M_xZw1EMireN8CZMmlI_sCiVsun8Rcq0FlnRkPeWdYQ,13227
+_pytest/reports.py,sha256=5OM_OyQHIS09PW6T_8kAJNS67GvcpvP-lKcna2LcSZ0,21331
+_pytest/runner.py,sha256=LDWKfhiIzWNkXqr1xwex-l1yhsWkdWCJko4bYM-etQ8,19436
+_pytest/scope.py,sha256=MyzqXUuWP1-BgbbCBzJyILiS_jicZO2LNUMUjv7vhh0,2798
+_pytest/setuponly.py,sha256=HNY9Ern-wex9iWSHxJU6ODA0yzYIH65QCkgNZ_BmbuA,3306
+_pytest/setupplan.py,sha256=l-ycFNxDZPyY52wh4f7yaqhzZ7SW1ijSKnQLmqzDZWA,1184
+_pytest/skipping.py,sha256=XbZKDPek9ex8aRXEoEy5iv0_e1b0sUi0PZrWqLBapek,10217
+_pytest/stash.py,sha256=5pE3kDx4q855TW9aVvYTdrkkKlMDU6-xiX4luKpJEgI,3090
+_pytest/stepwise.py,sha256=lYFm6kg000n_WEGOEQAho0j6dRCKJvgKz1Ya2Zz-0Zc,4596
+_pytest/terminal.py,sha256=-xT17xSJs9bu90wqRBc3WckaWTNTPOmVkZlO1X16Wyo,57393
+_pytest/threadexception.py,sha256=GHLYwCYK6I13Xv6bISO2crvPj9Z5ADKgVnUD7m1Oa14,3005
+_pytest/timing.py,sha256=URwa2JENXYkIN_9LFgEmJ4ric7SW8O6a8woS_TN6jXI,413
+_pytest/tmpdir.py,sha256=bo40r_gpxS7AdB_BANpSgh_fejHiXaGWrBxHpax9wtw,11375
+_pytest/unittest.py,sha256=wew7w2q5SqgdPppFzv0evwrTLWmMCwKFQvSUyEX2C0Q,15614
+_pytest/unraisableexception.py,sha256=-L6ln8mRnqqPBskzarua49st4ioXoKgllZ3oMmRuCKU,3252
+_pytest/warning_types.py,sha256=m2_Y3zydUZNzPpu88n8wPNWqaxfaATMKEo_zAgXMqyY,4388
+_pytest/warnings.py,sha256=ExyXdM9ZsIUX4o5GCt43fR-YWhIHSuUbV6GbKEVXeiA,5211
+py.py,sha256=txZ1tdmEW6CBTp6Idn-I2sOzzA0xKNoCi9Re27Uj6HE,329
+pytest-8.3.4.dist-info/AUTHORS,sha256=XobBrGbApvzdviFsaW1chh5icAuN555aWaMBHcmuqQE,7026
+pytest-8.3.4.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pytest-8.3.4.dist-info/LICENSE,sha256=yoNqX57Mo7LzUCMPqiCkj7ixRWU7VWjXhIYt-GRwa5s,1091
+pytest-8.3.4.dist-info/METADATA,sha256=f5v2O_PCDdT8dVKotHCLiHzXKMTS9hTO2YsKQ6_P3ig,7510
+pytest-8.3.4.dist-info/RECORD,,
+pytest-8.3.4.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pytest-8.3.4.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
+pytest-8.3.4.dist-info/entry_points.txt,sha256=8IPrHPH3LNZQ7v5tNEOcNTZYk_SheNg64jsTM9erqL4,77
+pytest-8.3.4.dist-info/top_level.txt,sha256=yyhjvmXH7-JOaoQIdmNQHPuoBCxOyXS3jIths_6C8A4,18
+pytest/__init__.py,sha256=jm6h0ZECJdDXlX0i5F20mN3ypV--T7osmtMHzzzY8ug,5169
+pytest/__main__.py,sha256=oVDrGGo7N0TNyzXntUblcgTKbhHGWtivcX5TC7tEcKo,154
+pytest/__pycache__/__init__.cpython-312.pyc,,
+pytest/__pycache__/__main__.cpython-312.pyc,,
+pytest/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/WHEEL
new file mode 100644
index 00000000..ae527e7d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/entry_points.txt
new file mode 100644
index 00000000..192205df
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/entry_points.txt
@@ -0,0 +1,3 @@
+[console_scripts]
+py.test = pytest:console_main
+pytest = pytest:console_main
diff --git a/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/top_level.txt
new file mode 100644
index 00000000..3084ae51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest-8.3.4.dist-info/top_level.txt
@@ -0,0 +1,3 @@
+_pytest
+py
+pytest
diff --git a/backend/venv/lib/python3.12/site-packages/pytest/__init__.py b/backend/venv/lib/python3.12/site-packages/pytest/__init__.py
new file mode 100644
index 00000000..90abcdab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest/__init__.py
@@ -0,0 +1,172 @@
+# PYTHON_ARGCOMPLETE_OK
+"""pytest: unit and functional testing with Python."""
+
+from __future__ import annotations
+
+from _pytest import __version__
+from _pytest import version_tuple
+from _pytest._code import ExceptionInfo
+from _pytest.assertion import register_assert_rewrite
+from _pytest.cacheprovider import Cache
+from _pytest.capture import CaptureFixture
+from _pytest.config import cmdline
+from _pytest.config import Config
+from _pytest.config import console_main
+from _pytest.config import ExitCode
+from _pytest.config import hookimpl
+from _pytest.config import hookspec
+from _pytest.config import main
+from _pytest.config import PytestPluginManager
+from _pytest.config import UsageError
+from _pytest.config.argparsing import OptionGroup
+from _pytest.config.argparsing import Parser
+from _pytest.debugging import pytestPDB as __pytestPDB
+from _pytest.doctest import DoctestItem
+from _pytest.fixtures import fixture
+from _pytest.fixtures import FixtureDef
+from _pytest.fixtures import FixtureLookupError
+from _pytest.fixtures import FixtureRequest
+from _pytest.fixtures import yield_fixture
+from _pytest.freeze_support import freeze_includes
+from _pytest.legacypath import TempdirFactory
+from _pytest.legacypath import Testdir
+from _pytest.logging import LogCaptureFixture
+from _pytest.main import Dir
+from _pytest.main import Session
+from _pytest.mark import Mark
+from _pytest.mark import MARK_GEN as mark
+from _pytest.mark import MarkDecorator
+from _pytest.mark import MarkGenerator
+from _pytest.mark import param
+from _pytest.monkeypatch import MonkeyPatch
+from _pytest.nodes import Collector
+from _pytest.nodes import Directory
+from _pytest.nodes import File
+from _pytest.nodes import Item
+from _pytest.outcomes import exit
+from _pytest.outcomes import fail
+from _pytest.outcomes import importorskip
+from _pytest.outcomes import skip
+from _pytest.outcomes import xfail
+from _pytest.pytester import HookRecorder
+from _pytest.pytester import LineMatcher
+from _pytest.pytester import Pytester
+from _pytest.pytester import RecordedHookCall
+from _pytest.pytester import RunResult
+from _pytest.python import Class
+from _pytest.python import Function
+from _pytest.python import Metafunc
+from _pytest.python import Module
+from _pytest.python import Package
+from _pytest.python_api import approx
+from _pytest.python_api import raises
+from _pytest.recwarn import deprecated_call
+from _pytest.recwarn import WarningsRecorder
+from _pytest.recwarn import warns
+from _pytest.reports import CollectReport
+from _pytest.reports import TestReport
+from _pytest.runner import CallInfo
+from _pytest.stash import Stash
+from _pytest.stash import StashKey
+from _pytest.terminal import TestShortLogReport
+from _pytest.tmpdir import TempPathFactory
+from _pytest.warning_types import PytestAssertRewriteWarning
+from _pytest.warning_types import PytestCacheWarning
+from _pytest.warning_types import PytestCollectionWarning
+from _pytest.warning_types import PytestConfigWarning
+from _pytest.warning_types import PytestDeprecationWarning
+from _pytest.warning_types import PytestExperimentalApiWarning
+from _pytest.warning_types import PytestRemovedIn9Warning
+from _pytest.warning_types import PytestReturnNotNoneWarning
+from _pytest.warning_types import PytestUnhandledCoroutineWarning
+from _pytest.warning_types import PytestUnhandledThreadExceptionWarning
+from _pytest.warning_types import PytestUnknownMarkWarning
+from _pytest.warning_types import PytestUnraisableExceptionWarning
+from _pytest.warning_types import PytestWarning
+
+
+set_trace = __pytestPDB.set_trace
+
+
+__all__ = [
+    "__version__",
+    "approx",
+    "Cache",
+    "CallInfo",
+    "CaptureFixture",
+    "Class",
+    "cmdline",
+    "Collector",
+    "CollectReport",
+    "Config",
+    "console_main",
+    "deprecated_call",
+    "Dir",
+    "Directory",
+    "DoctestItem",
+    "exit",
+    "ExceptionInfo",
+    "ExitCode",
+    "fail",
+    "File",
+    "fixture",
+    "FixtureDef",
+    "FixtureLookupError",
+    "FixtureRequest",
+    "freeze_includes",
+    "Function",
+    "hookimpl",
+    "HookRecorder",
+    "hookspec",
+    "importorskip",
+    "Item",
+    "LineMatcher",
+    "LogCaptureFixture",
+    "main",
+    "mark",
+    "Mark",
+    "MarkDecorator",
+    "MarkGenerator",
+    "Metafunc",
+    "Module",
+    "MonkeyPatch",
+    "OptionGroup",
+    "Package",
+    "param",
+    "Parser",
+    "PytestAssertRewriteWarning",
+    "PytestCacheWarning",
+    "PytestCollectionWarning",
+    "PytestConfigWarning",
+    "PytestDeprecationWarning",
+    "PytestExperimentalApiWarning",
+    "PytestRemovedIn9Warning",
+    "PytestReturnNotNoneWarning",
+    "Pytester",
+    "PytestPluginManager",
+    "PytestUnhandledCoroutineWarning",
+    "PytestUnhandledThreadExceptionWarning",
+    "PytestUnknownMarkWarning",
+    "PytestUnraisableExceptionWarning",
+    "PytestWarning",
+    "raises",
+    "RecordedHookCall",
+    "register_assert_rewrite",
+    "RunResult",
+    "Session",
+    "set_trace",
+    "skip",
+    "Stash",
+    "StashKey",
+    "version_tuple",
+    "TempdirFactory",
+    "TempPathFactory",
+    "Testdir",
+    "TestReport",
+    "TestShortLogReport",
+    "UsageError",
+    "WarningsRecorder",
+    "warns",
+    "xfail",
+    "yield_fixture",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/pytest/__main__.py b/backend/venv/lib/python3.12/site-packages/pytest/__main__.py
new file mode 100644
index 00000000..cccab5d5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest/__main__.py
@@ -0,0 +1,9 @@
+"""The pytest entry point."""
+
+from __future__ import annotations
+
+import pytest
+
+
+if __name__ == "__main__":
+    raise SystemExit(pytest.console_main())
diff --git a/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..809a3fce
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..46f97853
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest/py.typed b/backend/venv/lib/python3.12/site-packages/pytest/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/AUTHORS.rst b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/AUTHORS.rst
new file mode 100644
index 00000000..22ddfc63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/AUTHORS.rst
@@ -0,0 +1,28 @@
+Authors
+=======
+
+`Oleg Pidsadnyi <oleg.pidsadnyi@gmail.com>`_
+    original idea, initial implementation and further improvements
+`Anatoly Bubenkov <bubenkoff@gmail.com>`_
+    key implementation idea and realization, many new features and improvements
+
+These people have contributed to `pytest-bdd`, in alphabetical order:
+
+* `Adam Coddington <me@adamcoddington.net>`_
+* `Albert-Jan Nijburg <albertjan@curit.com>`_
+* `Alessio Bogon <youtux>`_
+* `Andrey Makhnach <andrey.makhnach@gmail.com>`_
+* `Aron Curzon <curzona@gmail.com>`_
+* `Dmitrijs Milajevs <dimazest@gmail.com>`_
+* `Dmitry Kolyagin <pauk-slon>`_
+* `Florian Bruhin <me@the-compiler.org>`_
+* `Floris Bruynooghe <flub@devork.be>`_
+* `Harro van der Klauw <hvdklauw@gmail.com>`_
+* `Hugo van Kemenade <https://github.com/hugovk>`_
+* `Kyle Adams <https://github.com/kadams54>`_
+* `Laurence Rowe <l@lrowe.co.uk>`_
+* `Leonardo Santagada <santagada@github.com>`_
+* `Milosz Sliwinski <sliwinski-milosz>`_
+* `Michiel Holtkamp <github@elfstone.nl>`_
+* `Robin Pedersen <ropez@github.com>`_
+* `Sergey Kraynev <sergejyit@gmail.com>`_
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/LICENSE.txt
new file mode 100644
index 00000000..811392e6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/LICENSE.txt
@@ -0,0 +1,7 @@
+Copyright (C) 2013-2014 Oleg Pidsadnyi, Anatoly Bubenkov and others
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/METADATA
new file mode 100644
index 00000000..8dd27b99
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/METADATA
@@ -0,0 +1,1663 @@
+Metadata-Version: 2.1
+Name: pytest-bdd
+Version: 8.1.0
+Summary: BDD for pytest
+Home-page: https://pytest-bdd.readthedocs.io/
+License: MIT
+Author: Oleg Pidsadnyi
+Author-email: oleg.pidsadnyi@gmail.com
+Maintainer: Alessio Bogon
+Maintainer-email: 778703+youtux@users.noreply.github.com
+Requires-Python: >=3.9
+Classifier: Development Status :: 6 - Mature
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development :: Testing
+Classifier: Topic :: Utilities
+Requires-Dist: Mako
+Requires-Dist: gherkin-official (>=29.0.0,<30.0.0)
+Requires-Dist: packaging
+Requires-Dist: parse
+Requires-Dist: parse-type
+Requires-Dist: pytest (>=7.0.0)
+Requires-Dist: typing-extensions
+Project-URL: Documentation, https://pytest-bdd.readthedocs.io/
+Project-URL: Repository, https://github.com/pytest-dev/pytest-bdd
+Description-Content-Type: text/x-rst
+
+Pytest-BDD: the BDD framework for pytest
+========================================
+
+.. image:: https://img.shields.io/pypi/v/pytest-bdd.svg
+   :target: https://pypi.python.org/pypi/pytest-bdd
+.. image:: https://codecov.io/gh/pytest-dev/pytest-bdd/branch/master/graph/badge.svg
+   :target: https://codecov.io/gh/pytest-dev/pytest-bdd
+.. image:: https://github.com/pytest-dev/pytest-bdd/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/pytest-dev/pytest-bdd/actions/workflows/main.yml
+.. image:: https://readthedocs.org/projects/pytest-bdd/badge/?version=stable
+   :target: https://readthedocs.org/projects/pytest-bdd/
+   :alt: Documentation Status
+
+pytest-bdd implements a subset of the Gherkin language to enable automating project
+requirements testing and to facilitate behavioral driven development.
+
+Unlike many other BDD tools, it does not require a separate runner and benefits from
+the power and flexibility of pytest. It enables unifying unit and functional
+tests, reduces the burden of continuous integration server configuration and allows the reuse of
+test setups.
+
+Pytest fixtures written for unit tests can be reused for setup and actions
+mentioned in feature steps with dependency injection. This allows a true BDD
+just-enough specification of the requirements without maintaining any context object
+containing the side effects of Gherkin imperative declarations.
+
+.. _behave: https://pypi.python.org/pypi/behave
+.. _pytest-splinter: https://github.com/pytest-dev/pytest-splinter
+
+Install pytest-bdd
+------------------
+
+::
+
+    pip install pytest-bdd
+
+
+Example
+-------
+
+An example test for a blog hosting software could look like this.
+Note that pytest-splinter_ is used to get the browser fixture.
+
+.. code-block:: gherkin
+
+    # content of publish_article.feature
+
+    Feature: Blog
+        A site where you can publish your articles.
+
+        Scenario: Publishing the article
+            Given I'm an author user
+            And I have an article
+
+            When I go to the article page
+            And I press the publish button
+
+            Then I should not see the error message
+            And the article should be published
+
+Note that only one feature is allowed per feature file.
+
+.. code-block:: python
+
+    # content of test_publish_article.py
+
+    from pytest_bdd import scenario, given, when, then
+
+    @scenario('publish_article.feature', 'Publishing the article')
+    def test_publish():
+        pass
+
+
+    @given("I'm an author user")
+    def author_user(auth, author):
+        auth['user'] = author.user
+
+
+    @given("I have an article", target_fixture="article")
+    def article(author):
+        return create_test_article(author=author)
+
+
+    @when("I go to the article page")
+    def go_to_article(article, browser):
+        browser.visit(urljoin(browser.url, '/manage/articles/{0}/'.format(article.id)))
+
+
+    @when("I press the publish button")
+    def publish_article(browser):
+        browser.find_by_css('button[name=publish]').first.click()
+
+
+    @then("I should not see the error message")
+    def no_error_message(browser):
+        with pytest.raises(ElementDoesNotExist):
+            browser.find_by_css('.message.error').first
+
+
+    @then("the article should be published")
+    def article_is_published(article):
+        article.refresh()  # Refresh the object in the SQLAlchemy session
+        assert article.is_published
+
+
+Scenario decorator
+------------------
+
+Functions decorated with the `scenario` decorator behave like a normal test function,
+and they will be executed after all scenario steps.
+
+
+.. code-block:: python
+
+    from pytest_bdd import scenario, given, when, then
+
+    @scenario('publish_article.feature', 'Publishing the article')
+    def test_publish(browser):
+        assert article.title in browser.html
+
+
+.. NOTE:: It is however encouraged to try as much as possible to have your logic only inside the Given, When, Then steps.
+
+
+Step aliases
+------------
+
+Sometimes, one has to declare the same fixtures or steps with
+different names for better readability. In order to use the same step
+function with multiple step names simply decorate it multiple times:
+
+.. code-block:: python
+
+    @given("I have an article")
+    @given("there's an article")
+    def article(author, target_fixture="article"):
+        return create_test_article(author=author)
+
+Note that the given step aliases are independent and will be executed
+when mentioned.
+
+For example if you associate your resource to some owner or not. Admin
+user can’t be an author of the article, but articles should have a
+default author.
+
+.. code-block:: gherkin
+
+    Feature: Resource owner
+        Scenario: I'm the author
+            Given I'm an author
+            And I have an article
+
+
+        Scenario: I'm the admin
+            Given I'm the admin
+            And there's an article
+
+
+Using Asterisks in Place of Keywords
+------------------------------------
+
+To avoid redundancy or unnecessary repetition of keywords
+such as "And" or "But" in Gherkin scenarios,
+you can use an asterisk (*) as a shorthand.
+The asterisk acts as a wildcard, allowing for the same functionality
+without repeating the keyword explicitly.
+It improves readability by making the steps easier to follow,
+especially when the specific keyword does not add value to the scenario's clarity.
+
+The asterisk will work the same as other step keywords - Given, When, Then - it follows.
+
+For example:
+
+.. code-block:: gherkin
+
+    Feature: Resource owner
+        Scenario: I'm the author
+            Given I'm an author
+            * I have an article
+            * I have a pen
+
+
+.. code-block:: python
+
+    from pytest_bdd import given
+
+    @given("I'm an author")
+    def _():
+        pass
+
+    @given("I have an article")
+    def _():
+        pass
+
+    @given("I have a pen")
+    def _():
+        pass
+
+
+In the scenario above, the asterisk (*) replaces the And or Given keywords.
+This allows for cleaner scenarios while still linking related steps together in the context of the scenario.
+
+This approach is particularly useful when you have a series of steps
+that do not require explicitly stating whether they are part of the "Given", "When", or "Then" context
+but are part of the logical flow of the scenario.
+
+
+Step arguments
+--------------
+
+Often it's possible to reuse steps giving them a parameter(s).
+This allows to have single implementation and multiple use, so less code.
+Also opens the possibility to use same step twice in single scenario and with different arguments!
+And even more, there are several types of step parameter parsers at your disposal
+(idea taken from behave_ implementation):
+
+.. _pypi_parse: http://pypi.python.org/pypi/parse
+.. _pypi_parse_type: http://pypi.python.org/pypi/parse_type
+
+**string** (the default)
+    This is the default and can be considered as a `null` or `exact` parser. It parses no parameters
+    and matches the step name by equality of strings.
+**parse** (based on: pypi_parse_)
+    Provides a simple parser that replaces regular expressions for
+    step parameters with a readable syntax like ``{param:Type}``.
+    The syntax is inspired by the Python builtin ``string.format()``
+    function.
+    Step parameters must use the named fields syntax of pypi_parse_
+    in step definitions. The named fields are extracted,
+    optionally type converted and then used as step function arguments.
+    Supports type conversions by using type converters passed via `extra_types`
+**cfparse** (extends: pypi_parse_, based on: pypi_parse_type_)
+    Provides an extended parser with "Cardinality Field" (CF) support.
+    Automatically creates missing type converters for related cardinality
+    as long as a type converter for cardinality=1 is provided.
+    Supports parse expressions like:
+    * ``{values:Type+}`` (cardinality=1..N, many)
+    * ``{values:Type*}`` (cardinality=0..N, many0)
+    * ``{value:Type?}``  (cardinality=0..1, optional)
+    Supports type conversions (as above).
+**re**
+    This uses full regular expressions to parse the clause text. You will
+    need to use named groups "(?P<name>...)" to define the variables pulled
+    from the text and passed to your ``step()`` function.
+    Type conversion can only be done via `converters` step decorator argument (see example below).
+
+The default parser is `string`, so just plain one-to-one match to the keyword definition.
+Parsers except `string`, as well as their optional arguments are specified like:
+
+for `cfparse` parser
+
+.. code-block:: python
+
+    from pytest_bdd import parsers
+
+    @given(
+        parsers.cfparse("there are {start:Number} cucumbers", extra_types={"Number": int}),
+        target_fixture="cucumbers",
+    )
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+for `re` parser
+
+.. code-block:: python
+
+    from pytest_bdd import parsers
+
+    @given(
+        parsers.re(r"there are (?P<start>\d+) cucumbers"),
+        converters={"start": int},
+        target_fixture="cucumbers",
+    )
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+
+Example:
+
+.. code-block:: gherkin
+
+    Feature: Step arguments
+        Scenario: Arguments for given, when, then
+            Given there are 5 cucumbers
+
+            When I eat 3 cucumbers
+            And I eat 2 cucumbers
+
+            Then I should have 0 cucumbers
+
+
+The code will look like:
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios, given, when, then, parsers
+
+
+    scenarios("arguments.feature")
+
+
+    @given(parsers.parse("there are {start:d} cucumbers"), target_fixture="cucumbers")
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+
+    @when(parsers.parse("I eat {eat:d} cucumbers"))
+    def eat_cucumbers(cucumbers, eat):
+        cucumbers["eat"] += eat
+
+
+    @then(parsers.parse("I should have {left:d} cucumbers"))
+    def should_have_left_cucumbers(cucumbers, left):
+        assert cucumbers["start"] - cucumbers["eat"] == left
+
+Example code also shows possibility to pass argument converters which may be useful if you need to postprocess step
+arguments after the parser.
+
+You can implement your own step parser. It's interface is quite simple. The code can look like:
+
+.. code-block:: python
+
+    import re
+    from pytest_bdd import given, parsers
+
+
+    class MyParser(parsers.StepParser):
+        """Custom parser."""
+
+        def __init__(self, name, **kwargs):
+            """Compile regex."""
+            super().__init__(name)
+            self.regex = re.compile(re.sub("%(.+)%", "(?P<\1>.+)", self.name), **kwargs)
+
+        def parse_arguments(self, name):
+            """Get step arguments.
+
+            :return: `dict` of step arguments
+            """
+            return self.regex.match(name).groupdict()
+
+        def is_matching(self, name):
+            """Match given name with the step name."""
+            return bool(self.regex.match(name))
+
+
+    @given(parsers.parse("there are %start% cucumbers"), target_fixture="cucumbers")
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+
+Override fixtures via given steps
+---------------------------------
+
+Dependency injection is not a panacea if you have complex structure of your test setup data. Sometimes there's a need
+such a given step which would imperatively change the fixture only for certain test (scenario), while for other tests
+it will stay untouched. To allow this, special parameter `target_fixture` exists in the `given` decorator:
+
+.. code-block:: python
+
+    from pytest_bdd import given
+
+    @pytest.fixture
+    def foo():
+        return "foo"
+
+
+    @given("I have injecting given", target_fixture="foo")
+    def injecting_given():
+        return "injected foo"
+
+
+    @then('foo should be "injected foo"')
+    def foo_is_foo(foo):
+        assert foo == 'injected foo'
+
+
+.. code-block:: gherkin
+
+    Feature: Target fixture
+        Scenario: Test given fixture injection
+            Given I have injecting given
+            Then foo should be "injected foo"
+
+
+In this example, the existing fixture `foo` will be overridden by given step `I have injecting given` only for the scenario it's
+used in.
+
+Sometimes it is also useful to let `when` and `then` steps provide a fixture as well.
+A common use case is when we have to assert the outcome of an HTTP request:
+
+.. code-block:: python
+
+    # content of test_blog.py
+
+    from pytest_bdd import scenarios, given, when, then
+
+    from my_app.models import Article
+
+    scenarios("blog.feature")
+
+
+    @given("there is an article", target_fixture="article")
+    def there_is_an_article():
+        return Article()
+
+
+    @when("I request the deletion of the article", target_fixture="request_result")
+    def there_should_be_a_new_article(article, http_client):
+        return http_client.delete(f"/articles/{article.uid}")
+
+
+    @then("the request should be successful")
+    def article_is_published(request_result):
+        assert request_result.status_code == 200
+
+
+.. code-block:: gherkin
+
+    # content of blog.feature
+
+    Feature: Blog
+        Scenario: Deleting the article
+            Given there is an article
+
+            When I request the deletion of the article
+
+            Then the request should be successful
+
+
+Scenarios shortcut
+------------------
+
+If you have a relatively large set of feature files, it's boring to manually bind scenarios to the tests using the scenario decorator. Of course with the manual approach you get all the power to be able to additionally parametrize the test, give the test function a nice name, document it, etc, but in the majority of the cases you don't need that.
+Instead, you want to bind all the scenarios found in the ``features`` folder(s) recursively automatically, by using the ``scenarios`` helper.
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios
+
+    # assume 'features' subfolder is in this file's directory
+    scenarios('features')
+
+That's all you need to do to bind all scenarios found in the ``features`` folder!
+Note that you can pass multiple paths, and those paths can be either feature files or feature folders.
+
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios
+
+    # pass multiple paths/files
+    scenarios('features', 'other_features/some.feature', 'some_other_features')
+
+But what if you need to manually bind a certain scenario, leaving others to be automatically bound?
+Just write your scenario in a "normal" way, but ensure you do it **before** the call of ``scenarios`` helper.
+
+
+.. code-block:: python
+
+    from pytest_bdd import scenario, scenarios
+
+    @scenario('features/some.feature', 'Test something')
+    def test_something():
+        pass
+
+    # assume 'features' subfolder is in this file's directory
+    scenarios('features')
+
+In the example above, the ``test_something`` scenario binding will be kept manual, other scenarios found in the ``features`` folder will be bound automatically.
+
+
+Scenario outlines
+-----------------
+
+Scenarios can be parametrized to cover multiple cases. These are called `Scenario Outlines <https://cucumber.io/docs/gherkin/reference/#scenario-outline>`_ in Gherkin, and the variable templates are written using angular brackets (e.g. ``<var_name>``).
+
+Example:
+
+.. code-block:: gherkin
+
+    # content of scenario_outlines.feature
+
+    Feature: Scenario outlines
+        Scenario Outline: Outlined given, when, then
+            Given there are <start> cucumbers
+            When I eat <eat> cucumbers
+            Then I should have <left> cucumbers
+
+            Examples:
+            | start | eat | left |
+            |  12   |  5  |  7   |
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios, given, when, then, parsers
+
+
+    scenarios("scenario_outlines.feature")
+
+
+    @given(parsers.parse("there are {start:d} cucumbers"), target_fixture="cucumbers")
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+
+    @when(parsers.parse("I eat {eat:d} cucumbers"))
+    def eat_cucumbers(cucumbers, eat):
+        cucumbers["eat"] += eat
+
+
+    @then(parsers.parse("I should have {left:d} cucumbers"))
+    def should_have_left_cucumbers(cucumbers, left):
+        assert cucumbers["start"] - cucumbers["eat"] == left
+
+
+Example parameters from example tables can not only be used in steps, but also embedded directly within docstrings and datatables, allowing for dynamic substitution.
+This provides added flexibility for scenarios that require complex setups or validations.
+
+Example:
+
+.. code-block:: gherkin
+
+    # content of docstring_and_datatable_with_params.feature
+
+    Feature: Docstring and Datatable with example parameters
+        Scenario Outline: Using parameters in docstrings and datatables
+            Given the following configuration:
+                """
+                username: <username>
+                password: <password>
+                """
+            When the user logs in
+            Then the response should contain:
+                | field     | value      |
+                | username  | <username> |
+                | logged_in | true       |
+
+            Examples:
+            | username  | password  |
+            | user1     | pass123   |
+            | user2     | 123secure |
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios, given, when, then
+    import json
+
+    # Load scenarios from the feature file
+    scenarios("docstring_and_datatable_with_params.feature")
+
+
+    @given("the following configuration:")
+    def given_user_config(docstring):
+        print(docstring)
+
+
+    @when("the user logs in")
+    def user_logs_in(logged_in):
+        logged_in = True
+
+
+    @then("the response should contain:")
+    def response_should_contain(datatable):
+        assert datatable[1][1] in ["user1", "user2"]
+
+
+Scenario Outlines with Multiple Example Tables
+----------------------------------------------
+
+In `pytest-bdd`, you can use multiple example tables in a scenario outline to test
+different sets of input data under various conditions.
+You can define separate `Examples` blocks, each with its own table of data,
+and optionally tag them to differentiate between positive, negative, or any other conditions.
+
+Example:
+
+.. code-block:: gherkin
+
+    # content of scenario_outline.feature
+
+    Feature: Scenario outlines with multiple examples tables
+        Scenario Outline: Outlined with multiple example tables
+            Given there are <start> cucumbers
+            When I eat <eat> cucumbers
+            Then I should have <left> cucumbers
+
+            @positive
+            Examples: Positive results
+                | start | eat | left |
+                |  12   |  5  |  7   |
+                |  5    |  4  |  1   |
+
+            @negative
+            Examples: Impossible negative results
+                | start | eat | left |
+                |  3    |  9  |  -6  |
+                |  1    |  4  |  -3  |
+
+.. code-block:: python
+
+    from pytest_bdd import scenarios, given, when, then, parsers
+
+
+    scenarios("scenario_outline.feature")
+
+
+    @given(parsers.parse("there are {start:d} cucumbers"), target_fixture="cucumbers")
+    def given_cucumbers(start):
+        return {"start": start, "eat": 0}
+
+
+    @when(parsers.parse("I eat {eat:d} cucumbers"))
+    def eat_cucumbers(cucumbers, eat):
+        cucumbers["eat"] += eat
+
+
+    @then(parsers.parse("I should have {left:d} cucumbers"))
+    def should_have_left_cucumbers(cucumbers, left):
+        assert cucumbers["start"] - cucumbers["eat"] == left
+
+
+When you filter scenarios by a tag, only the examples associated with that tag will be executed.
+This allows you to run a specific subset of your test cases based on the tag.
+For example, in the following scenario outline, if you filter by the @positive tag,
+only the examples under the "Positive results" table will be executed, and the "Negative results" table will be ignored.
+
+.. code-block:: bash
+
+    pytest -k "positive"
+
+
+Handling Empty Example Cells
+----------------------------
+
+By default, empty cells in the example tables are interpreted as empty strings ("").
+However, there may be cases where it is more appropriate to handle them as ``None``.
+In such scenarios, you can use a converter with the ``parsers.re`` parser to define a custom behavior for empty values.
+
+For example, the following code demonstrates how to use a custom converter to return ``None`` when an empty cell is encountered:
+
+.. code-block:: gherkin
+
+    # content of empty_example_cells.feature
+
+    Feature: Handling empty example cells
+        Scenario Outline: Using converters for empty cells
+            Given I am starting lunch
+            Then there are <start> cucumbers
+
+            Examples:
+            | start |
+            |       |
+
+.. code-block:: python
+
+    from pytest_bdd import then, parsers
+
+
+    # Define a converter that returns None for empty strings
+    def empty_to_none(value):
+        return None if value.strip() == "" else value
+
+
+    @given("I am starting lunch")
+    def _():
+        pass
+
+
+    @then(
+        parsers.re("there are (?P<start>.*?) cucumbers"),
+        converters={"start": empty_to_none}
+    )
+    def _(start):
+        # Example assertion to demonstrate the conversion
+        assert start is None
+
+
+Here, the `start` cell in the example table is empty.
+When the ``parsers.re`` parser is combined with the ``empty_to_none`` converter,
+the empty cell will be converted to ``None`` and can be handled accordingly in the step definition.
+
+
+Rules
+-----
+
+In Gherkin, `Rules` allow you to group related scenarios or examples under a shared context.
+This is useful when you want to define different conditions or behaviours
+for multiple examples that follow a similar structure.
+You can use either ``Scenario`` or ``Example`` to define individual cases, as they are aliases and function identically.
+
+Additionally, **tags** applied to a rule will be automatically applied to all the **examples or scenarios**
+under that rule, making it easier to organize and filter tests during execution.
+
+Example:
+
+.. code-block:: gherkin
+
+    Feature: Rules and examples
+
+        @feature_tag
+        Rule: A rule for valid cases
+
+            @rule_tag
+            Example: Valid case 1
+                Given I have a valid input
+                When I process the input
+                Then the result should be successful
+
+        Rule: A rule for invalid cases
+
+            Example: Invalid case
+                Given I have an invalid input
+                When I process the input
+                Then the result should be an error
+
+
+Datatables
+----------
+
+The ``datatable`` argument allows you to utilise data tables defined in your Gherkin scenarios
+directly within your test functions. This is particularly useful for scenarios that require tabular data as input,
+enabling you to manage and manipulate this data conveniently.
+
+When you use the ``datatable`` argument in a step definition, it will return the table as a list of lists,
+where each inner list represents a row from the table.
+
+For example, the Gherkin table:
+
+.. code-block:: gherkin
+
+    | name  | email            |
+    | John  | john@example.com |
+
+Will be returned by the ``datatable`` argument as:
+
+.. code-block:: python
+
+    [
+        ["name", "email"],
+        ["John", "john@example.com"]
+    ]
+
+.. NOTE:: When using the datatable argument, it is essential to ensure that the step to which it is applied
+          actually has an associated data table. If the step does not have an associated data table,
+          attempting to use the datatable argument will raise an error.
+          Make sure that your Gherkin steps correctly reference the data table when defined.
+
+Full example:
+
+.. code-block:: gherkin
+
+    Feature: Manage user accounts
+
+      Scenario: Creating a new user with roles and permissions
+        Given the following user details:
+          | name  | email             | age |
+          | John  | john@example.com  | 30  |
+          | Alice | alice@example.com | 25  |
+
+        When each user is assigned the following roles:
+          | Admin       | Full access to the system |
+          | Contributor | Can add content           |
+
+        And the page is saved
+
+        Then the user should have the following permissions:
+          | permission     | allowed |
+          | view dashboard | true    |
+          | edit content   | true    |
+          | delete content | false   |
+
+.. code-block:: python
+
+    from pytest_bdd import given, when, then
+
+
+    @given("the following user details:", target_fixture="users")
+    def _(datatable):
+        users = []
+        for row in datatable[1:]:
+            users.append(row)
+
+        print(users)
+        return users
+
+
+    @when("each user is assigned the following roles:")
+    def _(datatable, users):
+        roles = datatable
+        for user in users:
+            for role_row in datatable:
+                assign_role(user, role_row)
+
+
+    @when("the page is saved")
+    def _():
+        save_page()
+
+
+    @then("the user should have the following permissions:")
+    def _(datatable, users):
+        expected_permissions = []
+        for row in datatable[1:]:
+            expected_permissions.append(row)
+
+        assert users_have_correct_permissions(users, expected_permissions)
+
+
+Docstrings
+----------
+
+The `docstring` argument allows you to access the Gherkin docstring defined in your steps as a multiline string.
+The content of the docstring is passed as a single string, with each line separated by `\\n`.
+Leading indentation are stripped.
+
+For example, the Gherkin docstring:
+
+
+.. code-block:: gherkin
+
+    """
+    This is a sample docstring.
+    It spans multiple lines.
+    """
+
+
+Will be returned as:
+
+.. code-block:: python
+
+    "This is a sample docstring.\nIt spans multiple lines."
+
+
+Full example:
+
+.. code-block:: gherkin
+
+    Feature: Docstring
+
+      Scenario: Step with docstrings
+        Given some steps will have docstrings
+
+        Then a step has a docstring
+        """
+        This is a docstring
+        on two lines
+        """
+
+        And a step provides a docstring with lower indentation
+        """
+    This is a docstring
+        """
+
+        And this step has no docstring
+
+        And this step has a greater indentation
+        """
+            This is a docstring
+        """
+
+        And this step has no docstring
+
+.. code-block:: python
+
+        from pytest_bdd import given, then
+
+
+        @given("some steps will have docstrings")
+        def _():
+            pass
+
+        @then("a step has a docstring")
+        def _(docstring):
+            assert docstring == "This is a docstring\non two lines"
+
+        @then("a step provides a docstring with lower indentation")
+        def _(docstring):
+            assert docstring == "This is a docstring"
+
+        @then("this step has a greater indentation")
+        def _(docstring):
+            assert docstring == "This is a docstring"
+
+        @then("this step has no docstring")
+        def _():
+            pass
+
+
+.. note::   The ``docstring`` argument can only be used for steps that have an associated docstring.
+            Otherwise, an error will be thrown.
+
+Organizing your scenarios
+-------------------------
+
+The more features and scenarios you have, the more important the question of their organization becomes.
+The things you can do (and that is also a recommended way):
+
+* organize your feature files in the folders by semantic groups:
+
+::
+
+    features
+    │
+    ├──frontend
+    │  │
+    │  └──auth
+    │     │
+    │     └──login.feature
+    └──backend
+       │
+       └──auth
+          │
+          └──login.feature
+
+This looks fine, but how do you run tests only for a certain feature?
+As pytest-bdd uses pytest, and bdd scenarios are actually normal tests. But test files
+are separate from the feature files, the mapping is up to developers, so the test files structure can look
+completely different:
+
+::
+
+    tests
+    │
+    └──functional
+       │
+       └──test_auth.py
+          │
+          └ """Authentication tests."""
+            from pytest_bdd import scenario
+
+            @scenario('frontend/auth/login.feature')
+            def test_logging_in_frontend():
+                pass
+
+            @scenario('backend/auth/login.feature')
+            def test_logging_in_backend():
+                pass
+
+
+For picking up tests to run we can use the
+`tests selection <https://pytest.org/en/7.1.x/how-to/usage.html#specifying-which-tests-to-run>`_ technique. The problem is that
+you have to know how your tests are organized, knowing only the feature files organization is not enough.
+Cucumber uses `tags <https://cucumber.io/docs/cucumber/api/#tags>`_ as a way of categorizing your features
+and scenarios, which pytest-bdd supports. For example, we could have:
+
+.. code-block:: gherkin
+
+    @login @backend
+    Feature: Login
+
+      @successful
+      Scenario: Successful login
+
+
+pytest-bdd uses `pytest markers <http://pytest.org/latest/mark.html>`_ as a `storage` of the tags for the given
+scenario test, so we can use standard test selection:
+
+.. code-block:: bash
+
+    pytest -m "backend and login and successful"
+
+The feature and scenario markers are not different from standard pytest markers, and the ``@`` symbol is stripped out automatically to allow test selector expressions. If you want to have bdd-related tags to be distinguishable from the other test markers, use a prefix like ``bdd``.
+Note that if you use pytest with the ``--strict-markers`` option, all Gherkin tags mentioned in the feature files should be also in the ``markers`` setting of the ``pytest.ini`` config. Also for tags please use names which are python-compatible variable names, i.e. start with a non-number, only underscores or alphanumeric characters, etc. That way you can safely use tags for tests filtering.
+
+You can customize how tags are converted to pytest marks by implementing the
+``pytest_bdd_apply_tag`` hook and returning ``True`` from it:
+
+.. code-block:: python
+
+   def pytest_bdd_apply_tag(tag, function):
+       if tag == 'todo':
+           marker = pytest.mark.skip(reason="Not implemented yet")
+           marker(function)
+           return True
+       else:
+           # Fall back to the default behavior of pytest-bdd
+           return None
+
+Test setup
+----------
+
+Test setup is implemented within the Given section. Even though these steps
+are executed imperatively to apply possible side-effects, pytest-bdd is trying
+to benefit of the PyTest fixtures which is based on the dependency injection
+and makes the setup more declarative style.
+
+.. code-block:: python
+
+    @given("I have a beautiful article", target_fixture="article")
+    def article():
+        return Article(is_beautiful=True)
+
+The target PyTest fixture "article" gets the return value and any other step can depend on it.
+
+.. code-block:: gherkin
+
+    Feature: The power of PyTest
+        Scenario: Symbolic name across steps
+            Given I have a beautiful article
+            When I publish this article
+
+The When step is referencing the ``article`` to publish it.
+
+.. code-block:: python
+
+    @when("I publish this article")
+    def publish_article(article):
+        article.publish()
+
+
+Many other BDD toolkits operate on a global context and put the side effects there.
+This makes it very difficult to implement the steps, because the dependencies
+appear only as the side-effects during run-time and not declared in the code.
+The "publish article" step has to trust that the article is already in the context,
+has to know the name of the attribute it is stored there, the type etc.
+
+In pytest-bdd you just declare an argument of the step function that it depends on
+and the PyTest will make sure to provide it.
+
+Still side effects can be applied in the imperative style by design of the BDD.
+
+.. code-block:: gherkin
+
+    Feature: News website
+        Scenario: Publishing an article
+            Given I have a beautiful article
+            And my article is published
+
+Functional tests can reuse your fixture libraries created for the unit-tests and upgrade
+them by applying the side effects.
+
+.. code-block:: python
+
+    @pytest.fixture
+    def article():
+        return Article(is_beautiful=True)
+
+
+    @given("I have a beautiful article")
+    def i_have_a_beautiful_article(article):
+        pass
+
+
+    @given("my article is published")
+    def published_article(article):
+        article.publish()
+        return article
+
+
+This way side-effects were applied to our article and PyTest makes sure that all
+steps that require the "article" fixture will receive the same object. The value
+of the "published_article" and the "article" fixtures is the same object.
+
+Fixtures are evaluated **only once** within the PyTest scope and their values are cached.
+
+
+Backgrounds
+-----------
+
+It's often the case that to cover certain feature, you'll need multiple scenarios. And it's logical that the
+setup for those scenarios will have some common parts (if not equal). For this, there are `backgrounds`.
+pytest-bdd implements `Gherkin backgrounds <https://cucumber.io/docs/gherkin/reference/#background>`_ for
+features.
+
+.. code-block:: gherkin
+
+    Feature: Multiple site support
+
+      Background:
+        Given a global administrator named "Greg"
+        And a blog named "Greg's anti-tax rants"
+        And a customer named "Wilson"
+        And a blog named "Expensive Therapy" owned by "Wilson"
+
+      Scenario: Wilson posts to his own blog
+        Given I am logged in as Wilson
+        When I try to post to "Expensive Therapy"
+        Then I should see "Your article was published."
+
+      Scenario: Greg posts to a client's blog
+        Given I am logged in as Greg
+        When I try to post to "Expensive Therapy"
+        Then I should see "Your article was published."
+
+In this example, all steps from the background will be executed before all the scenario's own given
+steps, adding a possibility to prepare some common setup for multiple scenarios in a single feature.
+About best practices for Background, please read Gherkin's
+`Tips for using Background <https://cucumber.io/docs/gherkin/reference/#tips-for-using-background>`_.
+
+.. NOTE:: Only "Given" steps should be used in "Background" section.
+          Steps "When" and "Then" are prohibited, because their purposes are
+          related to actions and consuming outcomes; that is in conflict with the
+          aim of "Background" - to prepare the system for tests or "put the system
+          in a known state" as "Given" does it.
+
+
+Reusing fixtures
+----------------
+
+Sometimes scenarios define new names for an existing fixture that can be
+inherited (reused). For example, if we have the pytest fixture:
+
+
+.. code-block:: python
+
+    @pytest.fixture
+    def article():
+       """Test article."""
+       return Article()
+
+
+Then this fixture can be reused with other names using given():
+
+
+.. code-block:: python
+
+    @given('I have a beautiful article')
+    def i_have_an_article(article):
+       """I have an article."""
+
+
+Reusing steps
+-------------
+
+It is possible to define some common steps in the parent ``conftest.py`` and
+simply expect them in the child test file.
+
+.. code-block:: gherkin
+
+    # content of common_steps.feature
+
+    Scenario: All steps are declared in the conftest
+        Given I have a bar
+        Then bar should have value "bar"
+
+.. code-block:: python
+
+    # content of conftest.py
+
+    from pytest_bdd import given, then
+
+
+    @given("I have a bar", target_fixture="bar")
+    def bar():
+        return "bar"
+
+
+    @then('bar should have value "bar"')
+    def bar_is_bar(bar):
+        assert bar == "bar"
+
+.. code-block:: python
+
+    # content of test_common.py
+
+    @scenario("common_steps.feature", "All steps are declared in the conftest")
+    def test_conftest():
+        pass
+
+There are no definitions of steps in the test file. They were
+collected from the parent conftest.py.
+
+
+Default steps
+-------------
+
+Here is the list of steps that are implemented inside pytest-bdd:
+
+given
+    * trace - enters the `pdb` debugger via `pytest.set_trace()`
+when
+    * trace - enters the `pdb` debugger via `pytest.set_trace()`
+then
+    * trace - enters the `pdb` debugger via `pytest.set_trace()`
+
+
+Feature file paths
+------------------
+
+By default, pytest-bdd will use the current module's path as the base path for finding feature files, but this behaviour can be changed in the pytest configuration file (i.e. `pytest.ini`, `tox.ini` or `setup.cfg`) by declaring the new base path in the `bdd_features_base_dir` key. The path is interpreted as relative to the `pytest root directory <https://docs.pytest.org/en/latest/reference/customize.html#rootdir>`__.
+You can also override the features base path on a per-scenario basis, in order to override the path for specific tests.
+
+pytest.ini:
+
+.. code-block:: ini
+
+    [pytest]
+    bdd_features_base_dir = features/
+
+tests/test_publish_article.py:
+
+.. code-block:: python
+
+    from pytest_bdd import scenario
+
+
+    @scenario("foo.feature", "Foo feature in features/foo.feature")
+    def test_foo():
+        pass
+
+
+    @scenario(
+        "foo.feature",
+        "Foo feature in tests/local-features/foo.feature",
+        features_base_dir="./local-features/",
+    )
+    def test_foo_local():
+        pass
+
+
+The `features_base_dir` parameter can also be passed to the `@scenario` decorator.
+
+
+Avoid retyping the feature file name
+------------------------------------
+
+If you want to avoid retyping the feature file name when defining your scenarios in a test file, use ``functools.partial``.
+This will make your life much easier when defining multiple scenarios in a test file. For example:
+
+.. code-block:: python
+
+    # content of test_publish_article.py
+
+    from functools import partial
+
+    import pytest_bdd
+
+
+    scenario = partial(pytest_bdd.scenario, "/path/to/publish_article.feature")
+
+
+    @scenario("Publishing the article")
+    def test_publish():
+        pass
+
+
+    @scenario("Publishing the article as unprivileged user")
+    def test_publish_unprivileged():
+        pass
+
+
+You can learn more about `functools.partial <https://docs.python.org/3/library/functools.html#functools.partial>`_
+in the Python docs.
+
+
+Programmatic step generation
+----------------------------
+Sometimes you have step definitions that would be much easier to automate rather than writing them manually over and over again.
+This is common, for example, when using libraries like `pytest-factoryboy <https://pytest-factoryboy.readthedocs.io/>`_ that automatically creates fixtures.
+Writing step definitions for every model can become a tedious task.
+
+For this reason, pytest-bdd provides a way to generate step definitions automatically.
+
+The trick is to pass the ``stacklevel`` parameter to the ``given``, ``when``, ``then``, ``step`` decorators. This will instruct them to inject the step fixtures in the appropriate module, rather than just injecting them in the caller frame.
+
+Let's look at a concrete example; let's say you have a class ``Wallet`` that has some amount of each currency:
+
+.. code-block:: python
+
+    # contents of wallet.py
+
+    import dataclass
+
+    @dataclass
+    class Wallet:
+        verified: bool
+
+        amount_eur: int
+        amount_usd: int
+        amount_gbp: int
+        amount_jpy: int
+
+
+You can use pytest-factoryboy to automatically create model fixtures for this class:
+
+.. code-block:: python
+
+    # contents of wallet_factory.py
+
+    from wallet import Wallet
+
+    import factory
+    from pytest_factoryboy import register
+
+    class WalletFactory(factory.Factory):
+        class Meta:
+            model = Wallet
+
+        amount_eur = 0
+        amount_usd = 0
+        amount_gbp = 0
+        amount_jpy = 0
+
+    register(Wallet)  # creates the "wallet" fixture
+    register(Wallet, "second_wallet")  # creates the "second_wallet" fixture
+
+
+Now we can define a function ``generate_wallet_steps(...)`` that creates the steps for any wallet fixture (in our case, it will be ``wallet`` and ``second_wallet``):
+
+.. code-block:: python
+
+    # contents of wallet_steps.py
+
+    import re
+    from dataclasses import fields
+
+    import factory
+    import pytest
+    from pytest_bdd import given, when, then, scenarios, parsers
+
+
+    def generate_wallet_steps(model_name="wallet", stacklevel=1):
+        stacklevel += 1
+
+        human_name = model_name.replace("_", " ")  # "second_wallet" -> "second wallet"
+
+        @given(f"I have a {human_name}", target_fixture=model_name, stacklevel=stacklevel)
+        def _(request):
+            return request.getfixturevalue(model_name)
+
+        # Generate steps for currency fields:
+        for field in fields(Wallet):
+            match = re.fullmatch(r"amount_(?P<currency>[a-z]{3})", field.name)
+            if not match:
+                continue
+            currency = match["currency"]
+
+            @given(
+                parsers.parse(f"I have {{value:d}} {currency.upper()} in my {human_name}"),
+                target_fixture=f"{model_name}__amount_{currency}",
+                stacklevel=stacklevel,
+            )
+            def _(value: int) -> int:
+                return value
+
+            @then(
+                parsers.parse(f"I should have {{value:d}} {currency.upper()} in my {human_name}"),
+                stacklevel=stacklevel,
+            )
+            def _(value: int, _currency=currency, _model_name=model_name) -> None:
+                wallet = request.getfixturevalue(_model_name)
+                assert getattr(wallet, f"amount_{_currency}") == value
+
+    # Inject the steps into the current module
+    generate_wallet_steps("wallet")
+    generate_wallet_steps("second_wallet")
+
+
+This last file, ``wallet_steps.py``, now contains all the step definitions for our "wallet" and "second_wallet" fixtures.
+
+We can now define a scenario like this:
+
+.. code-block:: gherkin
+
+    # contents of wallet.feature
+    Feature: A feature
+
+        Scenario: Wallet EUR amount stays constant
+            Given I have 10 EUR in my wallet
+            And I have a wallet
+            Then I should have 10 EUR in my wallet
+
+        Scenario: Second wallet JPY amount stays constant
+            Given I have 100 JPY in my second wallet
+            And I have a second wallet
+            Then I should have 100 JPY in my second wallet
+
+
+and finally a test file that puts it all together and run the scenarios:
+
+.. code-block:: python
+
+    # contents of test_wallet.py
+
+    from pytest_factoryboy import scenarios
+
+    from wallet_factory import *  # import the registered fixtures "wallet" and "second_wallet"
+    from wallet_steps import *  # import all the step definitions into this test file
+
+    scenarios("wallet.feature")
+
+
+Hooks
+-----
+
+pytest-bdd exposes several `pytest hooks <https://docs.pytest.org/en/7.1.x/reference/reference.html#hooks>`_
+which might be helpful building useful reporting, visualization, etc. on top of it:
+
+* `pytest_bdd_before_scenario(request, feature, scenario)` - Called before scenario is executed
+
+* `pytest_bdd_after_scenario(request, feature, scenario)` - Called after scenario is executed
+  (even if one of steps has failed)
+
+* `pytest_bdd_before_step(request, feature, scenario, step, step_func)` - Called before step function
+  is executed and its arguments evaluated
+
+* `pytest_bdd_before_step_call(request, feature, scenario, step, step_func, step_func_args)` - Called before step
+  function is executed with evaluated arguments
+
+* `pytest_bdd_after_step(request, feature, scenario, step, step_func, step_func_args)` - Called after step function
+  is successfully executed
+
+* `pytest_bdd_step_error(request, feature, scenario, step, step_func, step_func_args, exception)` - Called when step
+  function failed to execute
+
+* `pytest_bdd_step_func_lookup_error(request, feature, scenario, step, exception)` - Called when step lookup failed
+
+
+Browser testing
+---------------
+
+Tools recommended to use for browser testing:
+
+* pytest-splinter_ - pytest `splinter <https://splinter.readthedocs.io/>`_ integration for the real browser testing
+
+
+Reporting
+---------
+
+It's important to have nice reporting out of your bdd tests. Cucumber introduced some kind of standard for
+`json format <https://www.relishapp.com/cucumber/cucumber/docs/json-output-formatter>`_
+which can be used for, for example, by `this <https://plugins.jenkins.io/cucumber-testresult-plugin/>`_ Jenkins
+plugin.
+
+To have an output in json format:
+
+::
+
+    pytest --cucumberjson=<path to json report>
+
+This will output an expanded (meaning scenario outlines will be expanded to several scenarios) Cucumber format.
+
+To enable gherkin-formatted output on terminal, use `--gherkin-terminal-reporter` in conjunction with the `-v` or `-vv` options:
+
+::
+
+    pytest -v --gherkin-terminal-reporter
+
+
+Test code generation helpers
+----------------------------
+
+For newcomers it's sometimes hard to write all needed test code without being frustrated.
+To simplify their life, a simple code generator was implemented. It allows to create fully functional
+(but of course empty) tests and step definitions for a given feature file.
+It's done as a separate console script provided by pytest-bdd package:
+
+::
+
+    pytest-bdd generate <feature file name> .. <feature file nameN>
+
+It will print the generated code to the standard output so you can easily redirect it to the file:
+
+::
+
+    pytest-bdd generate features/some.feature > tests/functional/test_some.py
+
+
+Advanced code generation
+------------------------
+
+For more experienced users, there's a smart code generation/suggestion feature. It will only generate the
+test code which is not yet there, checking existing tests and step definitions the same way it's done during the
+test execution. The code suggestion tool is called via passing additional pytest arguments:
+
+::
+
+    pytest --generate-missing --feature features tests/functional
+
+The output will be like:
+
+::
+
+    ============================= test session starts ==============================
+    platform linux2 -- Python 2.7.6 -- py-1.4.24 -- pytest-2.6.2
+    plugins: xdist, pep8, cov, cache, bdd, bdd, bdd
+    collected 2 items
+
+    Scenario is not bound to any test: "Code is generated for scenarios which are not bound to any tests" in feature "Missing code generation" in /tmp/pytest-552/testdir/test_generate_missing0/tests/generation.feature
+    --------------------------------------------------------------------------------
+
+    Step is not defined: "I have a custom bar" in scenario: "Code is generated for scenario steps which are not yet defined(implemented)" in feature "Missing code generation" in /tmp/pytest-552/testdir/test_generate_missing0/tests/generation.feature
+    --------------------------------------------------------------------------------
+    Please place the code above to the test file(s):
+
+    @scenario('tests/generation.feature', 'Code is generated for scenarios which are not bound to any tests')
+    def test_Code_is_generated_for_scenarios_which_are_not_bound_to_any_tests():
+        """Code is generated for scenarios which are not bound to any tests."""
+
+
+    @given("I have a custom bar")
+    def I_have_a_custom_bar():
+        """I have a custom bar."""
+
+As as side effect, the tool will validate the files for format errors, also some of the logic bugs, for example the
+ordering of the types of the steps.
+
+
+.. _Migration from 5.x.x:
+
+Migration of your tests from versions 5.x.x
+-------------------------------------------
+
+The primary focus of the pytest-bdd is the compatibility with the latest gherkin developments
+e.g. multiple scenario outline example tables with tags support etc.
+
+In order to provide the best compatibility, it is best to support the features described in the official
+gherkin reference. This means deprecation of some non-standard features that were implemented in pytest-bdd.
+
+
+Removal of the feature examples
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+The example tables on the feature level are no longer supported. If you had examples on the feature level, you should copy them to each individual scenario.
+
+
+Removal of the vertical examples
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+Vertical example tables are no longer supported since the official gherkin doesn't support them.
+The example tables should have horizontal orientation.
+
+
+Step arguments are no longer fixtures
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+Step parsed arguments conflicted with the fixtures. Now they no longer define fixture.
+If the fixture has to be defined by the step, the target_fixture param should be used.
+
+
+Variable templates in steps are only parsed for Scenario Outlines
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+In previous versions of pytest, steps containing ``<variable>`` would be parsed both by ``Scenario`` and ``Scenario Outline``.
+Now they are only parsed within a ``Scenario Outline``.
+
+
+.. _Migration from 4.x.x:
+
+Migration of your tests from versions 4.x.x
+-------------------------------------------
+
+Replace usage of <parameter> inside step definitions with parsed {parameter}
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+Templated steps (e.g. ``@given("there are <start> cucumbers")``) should now the use step argument parsers in order to match the scenario outlines and get the values from the example tables. The values from the example tables are no longer passed as fixtures, although if you define your step to use a parser, the parameters will be still provided as fixtures.
+
+.. code-block:: python
+
+    # Old step definition:
+    @given("there are <start> cucumbers")
+    def given_cucumbers(start):
+        pass
+
+
+    # New step definition:
+    @given(parsers.parse("there are {start} cucumbers"))
+    def given_cucumbers(start):
+        pass
+
+
+Scenario `example_converters` are removed in favor of the converters provided on the step level:
+
+.. code-block:: python
+
+    # Old code:
+    @given("there are <start> cucumbers")
+    def given_cucumbers(start):
+        return {"start": start}
+
+    @scenario("outline.feature", "Outlined", example_converters={"start": float})
+    def test_outline():
+        pass
+
+
+    # New code:
+    @given(parsers.parse("there are {start} cucumbers"), converters={"start": float})
+    def given_cucumbers(start):
+        return {"start": start}
+
+    @scenario("outline.feature", "Outlined")
+    def test_outline():
+        pass
+
+
+Refuse combining scenario outline and pytest parametrization
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+The significant downside of combining scenario outline and pytest parametrization approach was an inability to see the
+test table from the feature file.
+
+
+.. _Migration from 3.x.x:
+
+Migration of your tests from versions 3.x.x
+-------------------------------------------
+
+
+Given steps are no longer fixtures. In case it is needed to make given step setup a fixture,
+the target_fixture parameter should be used.
+
+
+.. code-block:: python
+
+    @given("there's an article", target_fixture="article")
+    def there_is_an_article():
+        return Article()
+
+
+Given steps no longer have the `fixture` parameter. In fact the step may depend on multiple fixtures.
+Just normal step declaration with the dependency injection should be used.
+
+.. code-block:: python
+
+    @given("there's an article")
+    def there_is_an_article(article):
+        pass
+
+
+Strict gherkin option is removed, so the ``strict_gherkin`` parameter can be removed from the scenario decorators
+as well as ``bdd_strict_gherkin`` from the ini files.
+
+Step validation handlers for the hook ``pytest_bdd_step_validation_error`` should be removed.
+
+License
+-------
+
+This software is licensed under the `MIT License <https://opensource.org/licenses/MIT>`_.
+
+© 2013 Oleg Pidsadnyi, Anatoly Bubenkov and others
+
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/RECORD
new file mode 100644
index 00000000..2b86202f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/RECORD
@@ -0,0 +1,47 @@
+../../../bin/pytest-bdd,sha256=T6-R6GfOWXKqdy8SWORK5RAk7UuXabI_0FyauIdcY9I,274
+pytest_bdd-8.1.0.dist-info/AUTHORS.rst,sha256=Lw63eLUno26AwCg87bHMG190jIv3iT5eOwypTatrMCg,1077
+pytest_bdd-8.1.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+pytest_bdd-8.1.0.dist-info/LICENSE.txt,sha256=f1n6ugf-djAFtxG5hSL2UllEqhEpYa--h6NgJUkftXE,1092
+pytest_bdd-8.1.0.dist-info/METADATA,sha256=hWA-YAJ3Xc7_dol3Pcuhb1w4tcqLAM4qA_M7yT1qvZU,53006
+pytest_bdd-8.1.0.dist-info/RECORD,,
+pytest_bdd-8.1.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pytest_bdd-8.1.0.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+pytest_bdd-8.1.0.dist-info/entry_points.txt,sha256=ARYHol5wm3zqNlNNttotPBM-1Kpurc97GEurz6KJDmo,95
+pytest_bdd/__init__.py,sha256=G3508mX3FTqcO7fczQRdQrqLWteH_arhM8e0UHqQBT4,241
+pytest_bdd/__pycache__/__init__.cpython-312.pyc,,
+pytest_bdd/__pycache__/compat.cpython-312.pyc,,
+pytest_bdd/__pycache__/cucumber_json.cpython-312.pyc,,
+pytest_bdd/__pycache__/exceptions.cpython-312.pyc,,
+pytest_bdd/__pycache__/feature.cpython-312.pyc,,
+pytest_bdd/__pycache__/generation.cpython-312.pyc,,
+pytest_bdd/__pycache__/gherkin_parser.cpython-312.pyc,,
+pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312.pyc,,
+pytest_bdd/__pycache__/hooks.cpython-312.pyc,,
+pytest_bdd/__pycache__/parser.cpython-312.pyc,,
+pytest_bdd/__pycache__/parsers.cpython-312.pyc,,
+pytest_bdd/__pycache__/plugin.cpython-312.pyc,,
+pytest_bdd/__pycache__/reporting.cpython-312.pyc,,
+pytest_bdd/__pycache__/scenario.cpython-312.pyc,,
+pytest_bdd/__pycache__/scripts.cpython-312.pyc,,
+pytest_bdd/__pycache__/steps.cpython-312.pyc,,
+pytest_bdd/__pycache__/types.cpython-312.pyc,,
+pytest_bdd/__pycache__/utils.cpython-312.pyc,,
+pytest_bdd/compat.py,sha256=_VO7a_zvxZ6jZz1NXc82S_zYn6gkfDrOlbfUvys2alY,2917
+pytest_bdd/cucumber_json.py,sha256=BZVsgE3nB7A-7tXVeyxIDb6ODiSIdZbCDnGYT4lL4NI,5636
+pytest_bdd/exceptions.py,sha256=EM1xNokASuL5wZc3WYF-nkBl8teNbnFCEba0cdNl54Q,1346
+pytest_bdd/feature.py,sha256=kKQ-5uE8qFYXdLAZrJa_rev2jlRyn8U5GZGWOvz0_FU,2818
+pytest_bdd/generation.py,sha256=onwqZh1S0z8M0IL1mVoDriqqX25Ub0B3dMuU8PVwwKA,6861
+pytest_bdd/gherkin_parser.py,sha256=dRmMcoRFIYMLdO8fQs6enY3Y_MT_VTx4vfJ3JyUKIEI,10249
+pytest_bdd/gherkin_terminal_reporter.py,sha256=3Hf9wGATgNYybzLXWDt5jIqgLRG524GJxOoEav_hoe8,4655
+pytest_bdd/hooks.py,sha256=DKt4MR2EIGdORIFK5Ge14gZhTXNLH2RSZUiOldakqMA,2145
+pytest_bdd/parser.py,sha256=6ZzeCabvs1g5RlJkyb2OHHqUPolSh0BsAbKW7-tlxAQ,19712
+pytest_bdd/parsers.py,sha256=bKgY9YGXZfG-9hSHSuqtdGncBVIF08FF5hbyrfeyg9w,3109
+pytest_bdd/plugin.py,sha256=Mzw_tDdixkK_LZFilA0R3WoT3roIzASZedIlCUTHEkg,3867
+pytest_bdd/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pytest_bdd/reporting.py,sha256=DRucx3fD3iKrDv4PVrF4pH6LJV0gp0EixOT92qL-2po,5826
+pytest_bdd/scenario.py,sha256=BoFKGj4nXiE0zZQTNoruAZWYLz7AVoBcdLngyE4pttY,17598
+pytest_bdd/scripts.py,sha256=F63VtUqwckqVTMCQ-gdNAikbWXEVYjELcGG8o4CgAWY,2640
+pytest_bdd/steps.py,sha256=qExNgK71HT8lyrORRafS93nWe4vbK26-AvJ06TEsL00,6596
+pytest_bdd/templates/test.py.mak,sha256=6SfxpzMleeuSQbwGpiWJ4ZbUSq_LHXW5p0AHYU8RQCM,620
+pytest_bdd/types.py,sha256=rGQaZuBSmrhz0-HdGPv-p0MVbhDOqHqJ9oPBwYOyjRE,379
+pytest_bdd/utils.py,sha256=fqhWDODj7ihZHsIJQ44qdmf4LGmM7nqBmbcSABhiwRI,2597
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/WHEEL
new file mode 100644
index 00000000..8b9b3a1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: poetry-core 1.9.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/entry_points.txt
new file mode 100644
index 00000000..ca70180c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd-8.1.0.dist-info/entry_points.txt
@@ -0,0 +1,6 @@
+[console_scripts]
+pytest-bdd=pytest_bdd.scripts:main
+
+[pytest11]
+pytest-bdd=pytest_bdd.plugin
+
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__init__.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__init__.py
new file mode 100644
index 00000000..f5078b97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__init__.py
@@ -0,0 +1,8 @@
+"""pytest-bdd public API."""
+
+from __future__ import annotations
+
+from pytest_bdd.scenario import scenario, scenarios
+from pytest_bdd.steps import given, step, then, when
+
+__all__ = ["given", "when", "step", "then", "scenario", "scenarios"]
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..77db9145
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3430628d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..21e1ae04
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..de3b9006
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..7fa2315b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312.pyc
new file mode 100644
index 00000000..d5f864d5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/cucumber_json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..b7d8ac96
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..e42b3062
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..0264bf36
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312.pyc
new file mode 100644
index 00000000..898a6cfe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/feature.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..fbd69e02
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312.pyc
new file mode 100644
index 00000000..22fda128
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/generation.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..538b185e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312.pyc
new file mode 100644
index 00000000..174dd97b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..4c52aa92
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312.pyc
new file mode 100644
index 00000000..fe574163
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/gherkin_terminal_reporter.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..f43e46d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312.pyc
new file mode 100644
index 00000000..0abbe522
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/hooks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..37843bd3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312.pyc
new file mode 100644
index 00000000..0677449f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..6af2da81
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312.pyc
new file mode 100644
index 00000000..ad9dcea5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/parsers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..ceec2fbb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312.pyc
new file mode 100644
index 00000000..32de0e5b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/plugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..7406d6ee
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312.pyc
new file mode 100644
index 00000000..bfab326d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/reporting.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..841fc93c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312.pyc
new file mode 100644
index 00000000..c24639c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scenario.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scripts.cpython-312.pyc
new file mode 100644
index 00000000..47e65cd1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..f96bf5c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312.pyc
new file mode 100644
index 00000000..42fb553c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/steps.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..f8a15921
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..324a33be
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312-pytest-8.3.4.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312-pytest-8.3.4.pyc
new file mode 100644
index 00000000..37583b53
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312-pytest-8.3.4.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..089288a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/pytest_bdd/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/compat.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/compat.py
new file mode 100644
index 00000000..bdce0074
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/compat.py
@@ -0,0 +1,81 @@
+from __future__ import annotations
+
+from collections.abc import Sequence
+from importlib.metadata import version
+from typing import Any
+
+from _pytest.fixtures import FixtureDef, FixtureManager, FixtureRequest
+from _pytest.nodes import Node
+from packaging.version import parse as parse_version
+
+pytest_version = parse_version(version("pytest"))
+
+__all__ = ["getfixturedefs", "inject_fixture"]
+
+if pytest_version.release >= (8, 1):
+
+    def getfixturedefs(fixturemanager: FixtureManager, fixturename: str, node: Node) -> Sequence[FixtureDef] | None:
+        return fixturemanager.getfixturedefs(fixturename, node)
+
+    def inject_fixture(request: FixtureRequest, arg: str, value: Any) -> None:
+        """Inject fixture into pytest fixture request.
+
+        :param request: pytest fixture request
+        :param arg: argument name
+        :param value: argument value
+        """
+        # Ensure there's a fixture definition for the argument
+        request._fixturemanager._register_fixture(
+            name=arg,
+            func=lambda: value,
+            nodeid=request.node.nodeid,
+        )
+        # Note the fixture we just registered will have a lower priority
+        # if there was already one registered, so we need to force its value
+        # to the one we want to inject.
+        fixture_def = request._get_active_fixturedef(arg)
+        fixture_def.cached_result = (value, None, None)  # type: ignore
+
+else:
+
+    def getfixturedefs(fixturemanager: FixtureManager, fixturename: str, node: Node) -> Sequence[FixtureDef] | None:
+        return fixturemanager.getfixturedefs(fixturename, node.nodeid)  # type: ignore
+
+    def inject_fixture(request: FixtureRequest, arg: str, value: Any) -> None:
+        """Inject fixture into pytest fixture request.
+
+        :param request: pytest fixture request
+        :param arg: argument name
+        :param value: argument value
+        """
+        fd = FixtureDef(
+            fixturemanager=request._fixturemanager,  # type: ignore
+            baseid=None,
+            argname=arg,
+            func=lambda: value,
+            scope="function",
+            params=None,
+        )
+        fd.cached_result = (value, 0, None)
+
+        old_fd = request._fixture_defs.get(arg)
+        add_fixturename = arg not in request.fixturenames
+
+        def fin() -> None:
+            request._fixturemanager._arg2fixturedefs[arg].remove(fd)
+
+            if old_fd is not None:
+                request._fixture_defs[arg] = old_fd
+
+            if add_fixturename:
+                request._pyfuncitem._fixtureinfo.names_closure.remove(arg)
+
+        request.addfinalizer(fin)
+
+        # inject fixture definition
+        request._fixturemanager._arg2fixturedefs.setdefault(arg, []).append(fd)
+
+        # inject fixture value in request cache
+        request._fixture_defs[arg] = fd
+        if add_fixturename:
+            request._pyfuncitem._fixtureinfo.names_closure.append(arg)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/cucumber_json.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/cucumber_json.py
new file mode 100644
index 00000000..c9bd975a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/cucumber_json.py
@@ -0,0 +1,149 @@
+"""Cucumber json output formatter."""
+
+from __future__ import annotations
+
+import json
+import math
+import os
+import time
+import typing
+
+if typing.TYPE_CHECKING:
+    from typing import Any
+
+    from _pytest.config import Config
+    from _pytest.config.argparsing import Parser
+    from _pytest.reports import TestReport
+    from _pytest.terminal import TerminalReporter
+
+
+def add_options(parser: Parser) -> None:
+    """Add pytest-bdd options."""
+    group = parser.getgroup("bdd", "Cucumber JSON")
+    group.addoption(
+        "--cucumberjson",
+        "--cucumber-json",
+        action="store",
+        dest="cucumber_json_path",
+        metavar="path",
+        default=None,
+        help="create cucumber json style report file at given path.",
+    )
+
+
+def configure(config: Config) -> None:
+    cucumber_json_path = config.option.cucumber_json_path
+    # prevent opening json log on worker nodes (xdist)
+    if cucumber_json_path and not hasattr(config, "workerinput"):
+        config._bddcucumberjson = LogBDDCucumberJSON(cucumber_json_path)  # type: ignore[attr-defined]
+        config.pluginmanager.register(config._bddcucumberjson)  # type: ignore[attr-defined]
+
+
+def unconfigure(config: Config) -> None:
+    xml = getattr(config, "_bddcucumberjson", None)  # type: ignore[attr-defined]
+    if xml is not None:
+        del config._bddcucumberjson  # type: ignore[attr-defined]
+        config.pluginmanager.unregister(xml)
+
+
+class LogBDDCucumberJSON:
+    """Logging plugin for cucumber like json output."""
+
+    def __init__(self, logfile: str) -> None:
+        logfile = os.path.expanduser(os.path.expandvars(logfile))
+        self.logfile = os.path.normpath(os.path.abspath(logfile))
+        self.features: dict[str, dict] = {}
+
+    def _get_result(self, step: dict[str, Any], report: TestReport, error_message: bool = False) -> dict[str, Any]:
+        """Get scenario test run result.
+
+        :param step: `Step` step we get result for
+        :param report: pytest `Report` object
+        :return: `dict` in form {"status": "<passed|failed|skipped>", ["error_message": "<error_message>"]}
+        """
+        result: dict[str, Any] = {}
+        if report.passed or not step["failed"]:  # ignore setup/teardown
+            result = {"status": "passed"}
+        elif report.failed:
+            result = {"status": "failed", "error_message": str(report.longrepr) if error_message else ""}
+        elif report.skipped:
+            result = {"status": "skipped"}
+        result["duration"] = int(math.floor((10**9) * step["duration"]))  # nanosec
+        return result
+
+    def _serialize_tags(self, item: dict[str, Any]) -> list[dict[str, Any]]:
+        """Serialize item's tags.
+
+        :param item: json-serialized `Scenario` or `Feature`.
+        :return: `list` of `dict` in the form of:
+            [
+                {
+                    "name": "<tag>",
+                    "line": 2,
+                }
+            ]
+        """
+        return [{"name": tag, "line": item["line_number"] - 1} for tag in item["tags"]]
+
+    def pytest_runtest_logreport(self, report: TestReport) -> None:
+        try:
+            scenario = report.scenario
+        except AttributeError:
+            # skip reporting for non-bdd tests
+            return
+
+        if not scenario["steps"] or report.when != "call":
+            # skip if there isn't a result or scenario has no steps
+            return
+
+        def stepmap(step: dict[str, Any]) -> dict[str, Any]:
+            error_message = False
+            if step["failed"] and not scenario.setdefault("failed", False):
+                scenario["failed"] = True
+                error_message = True
+
+            step_name = step["name"]
+
+            return {
+                "keyword": step["keyword"],
+                "name": step_name,
+                "line": step["line_number"],
+                "match": {"location": ""},
+                "result": self._get_result(step, report, error_message),
+            }
+
+        if scenario["feature"]["filename"] not in self.features:
+            self.features[scenario["feature"]["filename"]] = {
+                "keyword": scenario["feature"]["keyword"],
+                "uri": scenario["feature"]["rel_filename"],
+                "name": scenario["feature"]["name"] or scenario["feature"]["rel_filename"],
+                "id": scenario["feature"]["rel_filename"].lower().replace(" ", "-"),
+                "line": scenario["feature"]["line_number"],
+                "description": scenario["feature"]["description"],
+                "language": scenario["feature"]["language"],
+                "tags": self._serialize_tags(scenario["feature"]),
+                "elements": [],
+            }
+
+        self.features[scenario["feature"]["filename"]]["elements"].append(
+            {
+                "keyword": scenario["keyword"],
+                "id": report.item["name"],
+                "name": scenario["name"],
+                "line": scenario["line_number"],
+                "description": scenario["description"],
+                "tags": self._serialize_tags(scenario),
+                "type": "scenario",
+                "steps": [stepmap(step) for step in scenario["steps"]],
+            }
+        )
+
+    def pytest_sessionstart(self) -> None:
+        self.suite_start_time = time.time()
+
+    def pytest_sessionfinish(self) -> None:
+        with open(self.logfile, "w", encoding="utf-8") as logfile:
+            logfile.write(json.dumps(list(self.features.values())))
+
+    def pytest_terminal_summary(self, terminalreporter: TerminalReporter) -> None:
+        terminalreporter.write_sep("-", f"generated json file: {self.logfile}")
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/exceptions.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/exceptions.py
new file mode 100644
index 00000000..b46a8221
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/exceptions.py
@@ -0,0 +1,65 @@
+"""pytest-bdd Exceptions."""
+
+from __future__ import annotations
+
+
+class StepImplementationError(Exception):
+    """Step implementation error."""
+
+
+class ScenarioIsDecoratorOnly(Exception):
+    """Scenario can be only used as decorator."""
+
+
+class ScenarioValidationError(Exception):
+    """Base class for scenario validation."""
+
+
+class ScenarioNotFound(ScenarioValidationError):
+    """Scenario Not Found."""
+
+
+class StepDefinitionNotFoundError(Exception):
+    """Step definition not found."""
+
+
+class NoScenariosFound(Exception):
+    """No scenarios found."""
+
+
+class GherkinParseError(Exception):
+    """Base class for all Gherkin parsing errors."""
+
+    def __init__(self, message: str, line: int, line_content: str, filename: str) -> None:
+        super().__init__(message)
+        self.message = message
+        self.line = line
+        self.line_content = line_content
+        self.filename = filename
+
+    def __str__(self) -> str:
+        return f"{self.message}\nLine number: {self.line}\nLine: {self.line_content}\nFile: {self.filename}"
+
+
+class FeatureError(GherkinParseError):
+    pass
+
+
+class BackgroundError(GherkinParseError):
+    pass
+
+
+class ScenarioError(GherkinParseError):
+    pass
+
+
+class StepError(GherkinParseError):
+    pass
+
+
+class RuleError(GherkinParseError):
+    pass
+
+
+class TokenError(GherkinParseError):
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/feature.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/feature.py
new file mode 100644
index 00000000..6a7aa55a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/feature.py
@@ -0,0 +1,80 @@
+"""Feature.
+
+The way of describing the behavior is based on Gherkin language, but a very
+limited version. It doesn't support any parameter tables.
+If the parametrization is needed to generate more test cases it can be done
+on the fixture level of the pytest.
+The <variable> syntax can be used here to make a connection between steps and
+it will also validate the parameters mentioned in the steps with ones
+provided in the pytest parametrization table.
+
+Syntax example:
+
+    Feature: Articles
+        Scenario: Publishing the article
+            Given I'm an author user
+            And I have an article
+            When I go to the article page
+            And I press the publish button
+            Then I should not see the error message
+            And the article should be published  # Note: will query the database
+
+:note: The "#" symbol is used for comments.
+:note: There are no multiline steps, the description of the step must fit in
+one line.
+"""
+
+from __future__ import annotations
+
+import glob
+import os.path
+
+from .parser import Feature, FeatureParser
+
+# Global features dictionary
+features: dict[str, Feature] = {}
+
+
+def get_feature(base_path: str, filename: str, encoding: str = "utf-8") -> Feature:
+    """Get a feature by the filename.
+
+    :param str base_path: Base feature directory.
+    :param str filename: Filename of the feature file.
+    :param str encoding: Feature file encoding.
+
+    :return: `Feature` instance from the parsed feature cache.
+
+    :note: The features are parsed on the execution of the test and
+           stored in the global variable cache to improve the performance
+           when multiple scenarios are referencing the same file.
+    """
+    __tracebackhide__ = True
+    full_name = os.path.abspath(os.path.join(base_path, filename))
+    feature = features.get(full_name)
+    if not feature:
+        feature = FeatureParser(base_path, filename, encoding).parse()
+        features[full_name] = feature
+    return feature
+
+
+def get_features(paths: list[str], encoding: str = "utf-8") -> list[Feature]:
+    """Get features for given paths.
+
+    :param list paths: `list` of paths (file or dirs)
+
+    :return: `list` of `Feature` objects.
+    """
+    seen_names = set()
+    _features = []
+    for path in paths:
+        if path not in seen_names:
+            seen_names.add(path)
+            if os.path.isdir(path):
+                file_paths = list(glob.iglob(os.path.join(path, "**", "*.feature"), recursive=True))
+                _features.extend(get_features(file_paths, encoding=encoding))
+            else:
+                base, name = os.path.split(path)
+                feature = get_feature(base, name, encoding=encoding)
+                _features.append(feature)
+    _features.sort(key=lambda _feature: _feature.name or _feature.filename)
+    return _features
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/generation.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/generation.py
new file mode 100644
index 00000000..04da0231
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/generation.py
@@ -0,0 +1,202 @@
+"""pytest-bdd missing test code generation."""
+
+from __future__ import annotations
+
+import itertools
+import os.path
+from typing import TYPE_CHECKING, cast
+
+from _pytest._io import TerminalWriter
+from mako.lookup import TemplateLookup  # type: ignore
+
+from .compat import getfixturedefs
+from .feature import get_features
+from .parser import Feature, ScenarioTemplate, Step
+from .scenario import inject_fixturedefs_for_step, make_python_docstring, make_python_name, make_string_literal
+from .steps import get_step_fixture_name
+from .types import STEP_TYPES
+
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+    from typing import Any
+
+    from _pytest.config import Config
+    from _pytest.config.argparsing import Parser
+    from _pytest.fixtures import FixtureDef, FixtureManager
+    from _pytest.main import Session
+    from _pytest.python import Function
+
+
+template_lookup = TemplateLookup(directories=[os.path.join(os.path.dirname(__file__), "templates")])
+
+
+def add_options(parser: Parser) -> None:
+    """Add pytest-bdd options."""
+    group = parser.getgroup("bdd", "Generation")
+
+    group._addoption(
+        "--generate-missing",
+        action="store_true",
+        dest="generate_missing",
+        default=False,
+        help="Generate missing bdd test code for given feature files and exit.",
+    )
+
+    group._addoption(
+        "--feature",
+        metavar="FILE_OR_DIR",
+        action="append",
+        dest="features",
+        help="Feature file or directory to generate missing code for. Multiple allowed.",
+    )
+
+
+def cmdline_main(config: Config) -> int | None:
+    """Check config option to show missing code."""
+    if config.option.generate_missing:
+        return show_missing_code(config)
+    return None  # Make mypy happy
+
+
+def generate_code(features: list[Feature], scenarios: list[ScenarioTemplate], steps: list[Step]) -> str:
+    """Generate test code for the given filenames."""
+    grouped_steps = group_steps(steps)
+    template = template_lookup.get_template("test.py.mak")
+    code = template.render(
+        features=features,
+        scenarios=scenarios,
+        steps=grouped_steps,
+        make_python_name=make_python_name,
+        make_python_docstring=make_python_docstring,
+        make_string_literal=make_string_literal,
+    )
+    return cast(str, code)
+
+
+def show_missing_code(config: Config) -> int:
+    """Wrap pytest session to show missing code."""
+    from _pytest.main import wrap_session
+
+    return wrap_session(config, _show_missing_code_main)
+
+
+def print_missing_code(scenarios: list[ScenarioTemplate], steps: list[Step]) -> None:
+    """Print missing code with TerminalWriter."""
+    tw = TerminalWriter()
+    scenario = step = None
+
+    for scenario in scenarios:
+        tw.line()
+        tw.line(
+            (
+                f'Scenario "{scenario.name}" is not bound to any test in the feature "{scenario.feature.name}" '
+                f"in the file {scenario.feature.filename}:{scenario.line_number}"
+            ),
+            red=True,
+        )
+
+    if scenario:
+        tw.sep("-", red=True)
+
+    for step in steps:
+        tw.line()
+        if step.scenario is not None:
+            tw.line(
+                (
+                    f'Step {step} is not defined in the scenario "{step.scenario.name}" '
+                    f'in the feature "{step.scenario.feature.name}" in the file '
+                    f"{step.scenario.feature.filename}:{step.line_number}"
+                ),
+                red=True,
+            )
+        elif step.background is not None:
+            message = f"Background step {step} is not defined."
+            tw.line(message, red=True)
+
+    if step:
+        tw.sep("-", red=True)
+
+    tw.line("Please place the code above to the test file(s):")
+    tw.line()
+
+    features = sorted(
+        (scenario.feature for scenario in scenarios), key=lambda feature: feature.name or feature.filename
+    )
+    code = generate_code(features, scenarios, steps)
+    tw.write(code)
+
+
+def _find_step_fixturedef(
+    fixturemanager: FixtureManager, item: Function, step: Step
+) -> Sequence[FixtureDef[Any]] | None:
+    """Find step fixturedef."""
+    with inject_fixturedefs_for_step(step=step, fixturemanager=fixturemanager, node=item):
+        bdd_name = get_step_fixture_name(step=step)
+        return getfixturedefs(fixturemanager, bdd_name, item)
+
+
+def parse_feature_files(paths: list[str], **kwargs: Any) -> tuple[list[Feature], list[ScenarioTemplate], list[Step]]:
+    """Parse feature files of given paths.
+
+    :param paths: `list` of paths (file or dirs)
+
+    :return: `list` of `tuple` in form:
+             (`list` of `Feature` objects, `list` of `Scenario` objects, `list` of `Step` objects).
+    """
+    features = get_features(paths, **kwargs)
+    scenarios = sorted(
+        itertools.chain.from_iterable(feature.scenarios.values() for feature in features),
+        key=lambda scenario: (scenario.feature.name or scenario.feature.filename, scenario.name),
+    )
+    steps = sorted((step for scenario in scenarios for step in scenario.steps), key=lambda step: step.name)
+    return features, scenarios, steps
+
+
+def group_steps(steps: list[Step]) -> list[Step]:
+    """Group steps by type."""
+    steps = sorted(steps, key=lambda step: step.type)
+    seen_steps = set()
+    grouped_steps = []
+    for step in itertools.chain.from_iterable(
+        sorted(group, key=lambda step: step.name) for _, group in itertools.groupby(steps, lambda step: step.type)
+    ):
+        if step.name not in seen_steps:
+            grouped_steps.append(step)
+            seen_steps.add(step.name)
+    grouped_steps.sort(key=lambda step: STEP_TYPES.index(step.type))
+    return grouped_steps
+
+
+def _show_missing_code_main(config: Config, session: Session) -> None:
+    """Preparing fixture duplicates for output."""
+    tw = TerminalWriter()
+    session.perform_collect()
+
+    fm = session._fixturemanager
+
+    if config.option.features is None:
+        tw.line("The --feature parameter is required.", red=True)
+        session.exitstatus = 100
+        return
+
+    features, scenarios, steps = parse_feature_files(config.option.features)
+
+    for item in session.items:
+        if scenario := getattr(item.obj, "__scenario__", None):  # type: ignore
+            if scenario in scenarios:
+                scenarios.remove(scenario)
+            for step in scenario.steps:
+                if _find_step_fixturedef(fm, item, step=step):  # type: ignore
+                    try:
+                        steps.remove(step)
+                    except ValueError:
+                        pass
+    for scenario in scenarios:
+        for step in scenario.steps:
+            if step.background is None:
+                steps.remove(step)
+    grouped_steps = group_steps(steps)
+    print_missing_code(scenarios, grouped_steps)
+
+    if scenarios or steps:
+        session.exitstatus = 100
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_parser.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_parser.py
new file mode 100644
index 00000000..737b0dd8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_parser.py
@@ -0,0 +1,345 @@
+from __future__ import annotations
+
+import linecache
+import re
+import textwrap
+import typing
+from collections.abc import Sequence
+from dataclasses import dataclass, field
+from typing import Any
+
+from gherkin.errors import CompositeParserException  # type: ignore
+from gherkin.parser import Parser  # type: ignore
+
+from . import exceptions
+
+if typing.TYPE_CHECKING:
+    from typing_extensions import Self
+
+
+ERROR_PATTERNS = [
+    (
+        re.compile(r"expected:.*got 'Feature.*'"),
+        exceptions.FeatureError,
+        "Multiple features are not allowed in a single feature file.",
+    ),
+    (
+        re.compile(r"expected:.*got '(?:Given|When|Then|And|But).*'"),
+        exceptions.FeatureError,
+        "Step definition outside of a Scenario or a Background.",
+    ),
+    (
+        re.compile(r"expected:.*got 'Background.*'"),
+        exceptions.BackgroundError,
+        "Multiple 'Background' sections detected. Only one 'Background' is allowed per feature.",
+    ),
+    (
+        re.compile(r"expected:.*got 'Scenario.*'"),
+        exceptions.ScenarioError,
+        "Misplaced or incorrect 'Scenario' keyword. Ensure it's correctly placed. There might be a missing Feature section.",
+    ),
+    (
+        re.compile(r"expected:.*got 'Given.*'"),
+        exceptions.StepError,
+        "Improper step keyword detected. Ensure correct order and indentation for steps (Given, When, Then, etc.).",
+    ),
+    (
+        re.compile(r"expected:.*got 'Rule.*'"),
+        exceptions.RuleError,
+        "Misplaced or incorrectly formatted 'Rule'. Ensure it follows the feature structure.",
+    ),
+    (
+        re.compile(r"expected:.*got '.*'"),
+        exceptions.TokenError,
+        "Unexpected token found. Check Gherkin syntax near the reported error.",
+    ),
+]
+
+
+@dataclass
+class Location:
+    column: int
+    line: int
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(column=data["column"], line=data["line"])
+
+
+@dataclass
+class Comment:
+    location: Location
+    text: str
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(location=Location.from_dict(data["location"]), text=data["text"])
+
+
+@dataclass
+class Cell:
+    location: Location
+    value: str
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(location=Location.from_dict(data["location"]), value=_to_raw_string(data["value"]))
+
+
+@dataclass
+class Row:
+    id: str
+    location: Location
+    cells: list[Cell]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            id=data["id"],
+            location=Location.from_dict(data["location"]),
+            cells=[Cell.from_dict(cell) for cell in data["cells"]],
+        )
+
+
+@dataclass
+class ExamplesTable:
+    location: Location
+    tags: list[Tag]
+    name: str | None = None
+    table_header: Row | None = None
+    table_body: list[Row] | None = field(default_factory=list)
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            location=Location.from_dict(data["location"]),
+            name=data.get("name"),
+            table_header=Row.from_dict(data["tableHeader"]) if data.get("tableHeader") else None,
+            table_body=[Row.from_dict(row) for row in data.get("tableBody", [])],
+            tags=[Tag.from_dict(tag) for tag in data["tags"]],
+        )
+
+
+@dataclass
+class DataTable:
+    location: Location
+    rows: list[Row]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            location=Location.from_dict(data["location"]), rows=[Row.from_dict(row) for row in data.get("rows", [])]
+        )
+
+    def raw(self) -> Sequence[Sequence[object]]:
+        return [[cell.value for cell in row.cells] for row in self.rows]
+
+
+@dataclass
+class DocString:
+    content: str
+    delimiter: str
+    location: Location
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            content=textwrap.dedent(data["content"]),
+            delimiter=data["delimiter"],
+            location=Location.from_dict(data["location"]),
+        )
+
+
+@dataclass
+class Step:
+    id: str
+    location: Location
+    keyword: str
+    keyword_type: str
+    text: str
+    datatable: DataTable | None = None
+    docstring: DocString | None = None
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            id=data["id"],
+            location=Location.from_dict(data["location"]),
+            keyword=data["keyword"].strip(),
+            keyword_type=data["keywordType"],
+            text=data["text"],
+            datatable=DataTable.from_dict(data["dataTable"]) if data.get("dataTable") else None,
+            docstring=DocString.from_dict(data["docString"]) if data.get("docString") else None,
+        )
+
+
+@dataclass
+class Tag:
+    id: str
+    location: Location
+    name: str
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(id=data["id"], location=Location.from_dict(data["location"]), name=data["name"])
+
+
+@dataclass
+class Scenario:
+    id: str
+    location: Location
+    keyword: str
+    name: str
+    description: str
+    steps: list[Step]
+    tags: list[Tag]
+    examples: list[ExamplesTable] = field(default_factory=list)
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            id=data["id"],
+            location=Location.from_dict(data["location"]),
+            keyword=data["keyword"],
+            name=data["name"],
+            description=data["description"],
+            steps=[Step.from_dict(step) for step in data["steps"]],
+            tags=[Tag.from_dict(tag) for tag in data["tags"]],
+            examples=[ExamplesTable.from_dict(example) for example in data["examples"]],
+        )
+
+
+@dataclass
+class Rule:
+    id: str
+    location: Location
+    keyword: str
+    name: str
+    description: str
+    tags: list[Tag]
+    children: list[Child]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            id=data["id"],
+            location=Location.from_dict(data["location"]),
+            keyword=data["keyword"],
+            name=data["name"],
+            description=data["description"],
+            tags=[Tag.from_dict(tag) for tag in data["tags"]],
+            children=[Child.from_dict(child) for child in data["children"]],
+        )
+
+
+@dataclass
+class Background:
+    id: str
+    location: Location
+    keyword: str
+    name: str
+    description: str
+    steps: list[Step]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            id=data["id"],
+            location=Location.from_dict(data["location"]),
+            keyword=data["keyword"],
+            name=data["name"],
+            description=data["description"],
+            steps=[Step.from_dict(step) for step in data["steps"]],
+        )
+
+
+@dataclass
+class Child:
+    background: Background | None = None
+    rule: Rule | None = None
+    scenario: Scenario | None = None
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            background=Background.from_dict(data["background"]) if data.get("background") else None,
+            rule=Rule.from_dict(data["rule"]) if data.get("rule") else None,
+            scenario=Scenario.from_dict(data["scenario"]) if data.get("scenario") else None,
+        )
+
+
+@dataclass
+class Feature:
+    location: Location
+    language: str
+    keyword: str
+    tags: list[Tag]
+    name: str
+    description: str
+    children: list[Child]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            location=Location.from_dict(data["location"]),
+            language=data["language"],
+            keyword=data["keyword"],
+            tags=[Tag.from_dict(tag) for tag in data["tags"]],
+            name=data["name"],
+            description=data["description"],
+            children=[Child.from_dict(child) for child in data["children"]],
+        )
+
+
+@dataclass
+class GherkinDocument:
+    feature: Feature
+    comments: list[Comment]
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> Self:
+        return cls(
+            feature=Feature.from_dict(data["feature"]),
+            comments=[Comment.from_dict(comment) for comment in data["comments"]],
+        )
+
+
+def _to_raw_string(normal_string: str) -> str:
+    return normal_string.replace("\\", "\\\\")
+
+
+def get_gherkin_document(abs_filename: str, encoding: str = "utf-8") -> GherkinDocument:
+    with open(abs_filename, encoding=encoding) as f:
+        feature_file_text = f.read()
+
+    try:
+        gherkin_data = Parser().parse(feature_file_text)
+    except CompositeParserException as e:
+        message = e.args[0]
+        line = e.errors[0].location["line"]
+        line_content = linecache.getline(abs_filename, e.errors[0].location["line"]).rstrip("\n")
+        filename = abs_filename
+        handle_gherkin_parser_error(message, line, line_content, filename, e)
+        # If no patterns matched, raise a generic GherkinParserError
+        raise exceptions.GherkinParseError(f"Unknown parsing error: {message}", line, line_content, filename) from e
+
+    # At this point, the `gherkin_data` should be valid if no exception was raised
+    return GherkinDocument.from_dict(gherkin_data)
+
+
+def handle_gherkin_parser_error(
+    raw_error: str, line: int, line_content: str, filename: str, original_exception: Exception | None = None
+) -> None:
+    """Map the error message to a specific exception type and raise it."""
+    # Split the raw_error into individual lines
+    error_lines = raw_error.splitlines()
+
+    # Check each line against all error patterns
+    for error_line in error_lines:
+        for pattern, exception_class, message in ERROR_PATTERNS:
+            if pattern.search(error_line):
+                # If a match is found, raise the corresponding exception with the formatted message
+                if original_exception:
+                    raise exception_class(message, line, line_content, filename) from original_exception
+                else:
+                    raise exception_class(message, line, line_content, filename)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_terminal_reporter.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_terminal_reporter.py
new file mode 100644
index 00000000..4aecd77f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/gherkin_terminal_reporter.py
@@ -0,0 +1,116 @@
+from __future__ import annotations
+
+import typing
+
+from _pytest.terminal import TerminalReporter
+
+if typing.TYPE_CHECKING:
+    from typing import Any
+
+    from _pytest.config import Config
+    from _pytest.config.argparsing import Parser
+    from _pytest.reports import TestReport
+
+
+def add_options(parser: Parser) -> None:
+    group = parser.getgroup("terminal reporting", "reporting", after="general")
+    group._addoption(
+        "--gherkin-terminal-reporter",
+        action="store_true",
+        dest="gherkin_terminal_reporter",
+        default=False,
+        help="enable gherkin output",
+    )
+
+
+def configure(config: Config) -> None:
+    if config.option.gherkin_terminal_reporter:
+        # Get the standard terminal reporter plugin and replace it with our
+        current_reporter = config.pluginmanager.getplugin("terminalreporter")
+        if current_reporter.__class__ != TerminalReporter:
+            raise Exception(
+                "gherkin-terminal-reporter is not compatible with any other terminal reporter."
+                "You can use only one terminal reporter."
+                "Currently '{0}' is used."
+                "Please decide to use one by deactivating {0} or gherkin-terminal-reporter.".format(
+                    current_reporter.__class__
+                )
+            )
+        gherkin_reporter = GherkinTerminalReporter(config)
+        config.pluginmanager.unregister(current_reporter)
+        config.pluginmanager.register(gherkin_reporter, "terminalreporter")
+        if config.pluginmanager.getplugin("dsession"):
+            raise Exception("gherkin-terminal-reporter is not compatible with 'xdist' plugin.")
+
+
+class GherkinTerminalReporter(TerminalReporter):  # type: ignore
+    def __init__(self, config: Config) -> None:
+        super().__init__(config)
+        self.current_rule = None
+
+    def pytest_runtest_logreport(self, report: TestReport) -> Any:
+        rep = report
+        res = self.config.hook.pytest_report_teststatus(report=rep, config=self.config)
+        cat, letter, word = res
+
+        if not letter and not word:
+            # probably passed setup/teardown
+            return None
+
+        if isinstance(word, tuple):
+            word, word_markup = word
+        elif rep.passed:
+            word_markup = {"green": True}
+        elif rep.failed:
+            word_markup = {"red": True}
+        elif rep.skipped:
+            word_markup = {"yellow": True}
+        feature_markup = {"blue": True}
+        scenario_markup = word_markup
+        rule_markup = {"purple": True}
+
+        if self.verbosity <= 0 or not hasattr(report, "scenario"):
+            return super().pytest_runtest_logreport(rep)
+
+        rule = report.scenario.get("rule")
+        indent = "    " if rule else ""
+
+        if self.verbosity == 1:
+            self.ensure_newline()
+            self._tw.write(f"{report.scenario['feature']['keyword']}: ", **feature_markup)
+            self._tw.write(report.scenario["feature"]["name"], **feature_markup)
+            self._tw.write("\n")
+
+            if rule and rule["name"] != self.current_rule:
+                self._tw.write(f"  {rule['keyword']}: ", **rule_markup)
+                self._tw.write(rule["name"], **rule_markup)
+                self._tw.write("\n")
+                self.current_rule = rule["name"]
+
+            self._tw.write(f"{indent}    {report.scenario['keyword']}: ", **scenario_markup)
+            self._tw.write(report.scenario["name"], **scenario_markup)
+            self._tw.write(" ")
+            self._tw.write(word, **word_markup)
+            self._tw.write("\n")
+        elif self.verbosity > 1:
+            self.ensure_newline()
+            self._tw.write(f"{report.scenario['feature']['keyword']}: ", **feature_markup)
+            self._tw.write(report.scenario["feature"]["name"], **feature_markup)
+            self._tw.write("\n")
+
+            if rule and rule["name"] != self.current_rule:
+                self._tw.write(f"  {rule['keyword']}: ", **rule_markup)
+                self._tw.write(rule["name"], **rule_markup)
+                self._tw.write("\n")
+                self.current_rule = rule["name"]
+
+            self._tw.write(f"{indent}    {report.scenario['keyword']}: ", **scenario_markup)
+            self._tw.write(report.scenario["name"], **scenario_markup)
+            self._tw.write("\n")
+            for step in report.scenario["steps"]:
+                self._tw.write(f"{indent}        {step['keyword']} {step['name']}\n", **scenario_markup)
+            self._tw.write(f"{indent}    {word}", **word_markup)
+            self._tw.write("\n\n")
+
+        self.stats.setdefault(cat, []).append(rep)
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/hooks.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/hooks.py
new file mode 100644
index 00000000..48e1cedb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/hooks.py
@@ -0,0 +1,74 @@
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+from _pytest.fixtures import FixtureRequest
+
+from pytest_bdd.parser import Feature, Scenario, Step
+
+"""Pytest-bdd pytest hooks."""
+
+
+def pytest_bdd_before_scenario(request: FixtureRequest, feature: Feature, scenario: Scenario) -> object:
+    """Called before scenario is executed."""
+
+
+def pytest_bdd_after_scenario(request: FixtureRequest, feature: Feature, scenario: Scenario) -> object:
+    """Called after scenario is executed."""
+
+
+def pytest_bdd_before_step(
+    request: FixtureRequest, feature: Feature, scenario: Scenario, step: Step, step_func: Callable[..., object]
+) -> object:
+    """Called before step function is set up."""
+
+
+def pytest_bdd_before_step_call(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., object],
+    step_func_args: dict[str, object],
+) -> object:
+    """Called before step function is executed."""
+
+
+def pytest_bdd_after_step(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., object],
+    step_func_args: dict[str, object],
+) -> object:
+    """Called after step function is successfully executed."""
+
+
+def pytest_bdd_step_error(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., object],
+    step_func_args: dict[str, object],
+    exception: Exception,
+) -> object:
+    """Called when step function failed to execute."""
+
+
+def pytest_bdd_step_func_lookup_error(
+    request: FixtureRequest, feature: Feature, scenario: Scenario, step: Step, exception: Exception
+) -> object:
+    """Called when step lookup failed."""
+
+
+@pytest.hookspec(firstresult=True)
+def pytest_bdd_apply_tag(tag: str, function: Callable[..., object]) -> object:
+    """Apply a tag (from a ``.feature`` file) to the given scenario.
+
+    The default implementation does the equivalent of
+    ``getattr(pytest.mark, tag)(function)``, but you can override this hook and
+    return ``True`` to do more sophisticated handling of tags.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/parser.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/parser.py
new file mode 100644
index 00000000..2ede93ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/parser.py
@@ -0,0 +1,543 @@
+from __future__ import annotations
+
+import copy
+import os.path
+import re
+import textwrap
+from collections import OrderedDict
+from collections.abc import Generator, Iterable, Mapping, Sequence
+from dataclasses import dataclass, field
+from typing import Any
+
+from .exceptions import StepError
+from .gherkin_parser import Background as GherkinBackground
+from .gherkin_parser import DataTable
+from .gherkin_parser import Feature as GherkinFeature
+from .gherkin_parser import GherkinDocument
+from .gherkin_parser import Rule as GherkinRule
+from .gherkin_parser import Scenario as GherkinScenario
+from .gherkin_parser import Step as GherkinStep
+from .gherkin_parser import Tag as GherkinTag
+from .gherkin_parser import get_gherkin_document
+from .types import STEP_TYPE_BY_PARSER_KEYWORD
+
+PARAM_RE = re.compile(r"<(.+?)>")
+
+
+def render_string(input_string: str, render_context: Mapping[str, object]) -> str:
+    """
+    Render the string with the given context,
+    but avoid replacing text inside angle brackets if context is missing.
+
+    Args:
+        input_string (str): The string for which to render/replace params.
+        render_context (Mapping[str, object]): The context for rendering the string.
+
+    Returns:
+        str: The rendered string with parameters replaced only if they exist in the context.
+    """
+
+    def replacer(m: re.Match) -> str:
+        varname = m.group(1)
+        # If the context contains the variable, replace it. Otherwise, leave it unchanged.
+        return str(render_context.get(varname, f"<{varname}>"))
+
+    return PARAM_RE.sub(replacer, input_string)
+
+
+def get_tag_names(tag_data: list[GherkinTag]) -> set[str]:
+    """Extract tag names from tag data.
+
+    Args:
+        tag_data (List[dict]): The tag data to extract names from.
+
+    Returns:
+        set[str]: A set of tag names.
+    """
+    return {tag.name.lstrip("@") for tag in tag_data}
+
+
+@dataclass(eq=False)
+class Feature:
+    """Represents a feature parsed from a feature file.
+
+    Attributes:
+        scenarios (OrderedDict[str, ScenarioTemplate]): A dictionary of scenarios in the feature.
+        filename (str): The absolute path of the feature file.
+        rel_filename (str): The relative path of the feature file.
+        name (Optional[str]): The name of the feature.
+        tags (set[str]): A set of tags associated with the feature.
+        background (Optional[Background]): The background steps for the feature, if any.
+        line_number (int): The line number where the feature starts in the file.
+        description (str): The description of the feature.
+    """
+
+    scenarios: OrderedDict[str, ScenarioTemplate]
+    filename: str
+    rel_filename: str
+    language: str
+    keyword: str
+    name: str | None
+    tags: set[str]
+    background: Background | None
+    line_number: int
+    description: str
+
+
+@dataclass(eq=False)
+class Examples:
+    """Represents examples used in scenarios for parameterization.
+
+    Attributes:
+        line_number (Optional[int]): The line number where the examples start.
+        name (Optional[str]): The name of the examples.
+        example_params (List[str]): The names of the parameters for the examples.
+        examples (List[Sequence[str]]): The list of example rows.
+    """
+
+    line_number: int | None = None
+    name: str | None = None
+    example_params: list[str] = field(default_factory=list)
+    examples: list[Sequence[str]] = field(default_factory=list)
+    tags: set[str] = field(default_factory=set)
+
+    def set_param_names(self, keys: Iterable[str]) -> None:
+        """Set the parameter names for the examples.
+
+        Args:
+            keys (Iterable[str]): The parameter names to set.
+        """
+        self.example_params = [str(key) for key in keys]
+
+    def add_example(self, values: Sequence[str]) -> None:
+        """Add a new example row.
+
+        Args:
+            values (Sequence[str]): The values for the example row.
+        """
+        self.examples.append([str(value) if value is not None else "" for value in values])
+
+    def as_contexts(self) -> Iterable[dict[str, Any]]:
+        """Generate contexts for the examples.
+
+        Yields:
+            Dict[str, Any]: A dictionary mapping parameter names to their values for each example row.
+        """
+        for row in self.examples:
+            assert len(self.example_params) == len(row)
+            yield dict(zip(self.example_params, row))
+
+    def __bool__(self) -> bool:
+        """Check if there are any examples.
+
+        Returns:
+            bool: True if there are examples, False otherwise.
+        """
+        return bool(self.examples)
+
+
+@dataclass(eq=False)
+class Rule:
+    keyword: str
+    name: str
+    description: str
+    tags: set[str]
+    background: Background | None = None
+
+
+@dataclass(eq=False)
+class ScenarioTemplate:
+    """Represents a scenario template within a feature.
+
+    Attributes:
+        feature (Feature): The feature to which this scenario belongs.
+        keyword (str): The keyword used to define the scenario.
+        name (str): The name of the scenario.
+        line_number (int): The line number where the scenario starts in the file.
+        templated (bool): Whether the scenario is templated.
+        description (Optional[str]): The description of the scenario.
+        tags (set[str]): A set of tags associated with the scenario.
+        _steps (List[Step]): The list of steps in the scenario (internal use only).
+        examples (Optional[Examples]): The examples used for parameterization in the scenario.
+        rule (Optional[Rule]): The rule to which the scenario may belong (None = no rule).
+    """
+
+    feature: Feature
+    keyword: str
+    name: str
+    line_number: int
+    templated: bool
+    description: str | None = None
+    tags: set[str] = field(default_factory=set)
+    _steps: list[Step] = field(init=False, default_factory=list)
+    examples: list[Examples] = field(default_factory=list[Examples])
+    rule: Rule | None = None
+
+    def add_step(self, step: Step) -> None:
+        """Add a step to the scenario.
+
+        Args:
+            step (Step): The step to add.
+        """
+        step.scenario = self
+        self._steps.append(step)
+
+    @property
+    def all_background_steps(self) -> list[Step]:
+        steps = []
+        # Add background steps from the feature
+        if self.feature.background:
+            steps.extend(self.feature.background.steps)
+        if self.rule is not None and self.rule.background is not None:
+            # Add background steps from the rule
+            steps.extend(self.rule.background.steps)
+        return steps
+
+    @property
+    def steps(self) -> list[Step]:
+        """Get all steps for the scenario, including background steps.
+
+        Returns:
+            List[Step]: A list of steps, including any background steps from the feature.
+        """
+        return self.all_background_steps + self._steps
+
+    def render(self, context: Mapping[str, Any]) -> Scenario:
+        """Render the scenario with the given context.
+
+        Args:
+            context (Mapping[str, Any]): The context for rendering steps.
+
+        Returns:
+            Scenario: A Scenario object with steps rendered based on the context.
+        """
+        base_steps = self.all_background_steps + self._steps
+        scenario_steps = [
+            Step(
+                name=render_string(step.name, context),
+                type=step.type,
+                indent=step.indent,
+                line_number=step.line_number,
+                keyword=step.keyword,
+                datatable=step.render_datatable(step.datatable, context) if step.datatable else None,
+                docstring=render_string(step.docstring, context) if step.docstring else None,
+            )
+            for step in base_steps
+        ]
+        return Scenario(
+            feature=self.feature,
+            keyword=self.keyword,
+            name=render_string(self.name, context),
+            line_number=self.line_number,
+            steps=scenario_steps,
+            tags=self.tags,
+            description=self.description,
+            rule=self.rule,
+        )
+
+
+@dataclass(eq=False)
+class Scenario:
+    """Represents a scenario with steps.
+
+    Attributes:
+        feature (Feature): The feature to which this scenario belongs.
+        keyword (str): The keyword used to define the scenario.
+        name (str): The name of the scenario.
+        line_number (int): The line number where the scenario starts in the file.
+        steps (List[Step]): The list of steps in the scenario.
+        description (Optional[str]): The description of the scenario.
+        tags (set[str]): A set of tags associated with the scenario.
+    """
+
+    feature: Feature
+    keyword: str
+    name: str
+    line_number: int
+    steps: list[Step]
+    description: str | None = None
+    tags: set[str] = field(default_factory=set)
+    rule: Rule | None = None
+
+
+@dataclass(eq=False)
+class Step:
+    """Represents a step within a scenario or background.
+
+    Attributes:
+        type (str): The type of step (e.g., 'given', 'when', 'then').
+        name (str): The name of the step.
+        line_number (int): The line number where the step starts in the file.
+        indent (int): The indentation level of the step.
+        keyword (str): The keyword used for the step (e.g., 'Given', 'When', 'Then').
+        failed (bool): Whether the step has failed (internal use only).
+        scenario (Optional[ScenarioTemplate]): The scenario to which this step belongs (internal use only).
+        background (Optional[Background]): The background to which this step belongs (internal use only).
+    """
+
+    type: str
+    name: str
+    line_number: int
+    indent: int
+    keyword: str
+    docstring: str | None = None
+    datatable: DataTable | None = None
+    failed: bool = field(init=False, default=False)
+    scenario: ScenarioTemplate | None = field(init=False, default=None)
+    background: Background | None = field(init=False, default=None)
+
+    def __init__(
+        self,
+        name: str,
+        type: str,
+        indent: int,
+        line_number: int,
+        keyword: str,
+        datatable: DataTable | None = None,
+        docstring: str | None = None,
+    ) -> None:
+        """Initialize a step.
+
+        Args:
+            name (str): The name of the step.
+            type (str): The type of the step (e.g., 'given', 'when', 'then').
+            indent (int): The indentation level of the step.
+            line_number (int): The line number where the step starts in the file.
+            keyword (str): The keyword used for the step (e.g., 'Given', 'When', 'Then').
+        """
+        self.name = name
+        self.type = type
+        self.indent = indent
+        self.line_number = line_number
+        self.keyword = keyword
+        self.datatable = datatable
+        self.docstring = docstring
+
+    def __str__(self) -> str:
+        """Return a string representation of the step.
+
+        Returns:
+            str: A string representation of the step.
+        """
+        return f'{self.type.capitalize()} "{self.name}"'
+
+    @staticmethod
+    def render_datatable(datatable: DataTable, context: Mapping[str, object]) -> DataTable:
+        """
+        Render the datatable with the given context,
+        but avoid replacing text inside angle brackets if context is missing.
+
+        Args:
+            datatable (DataTable): The datatable to render.
+            context (Mapping[str, Any]): The context for rendering the datatable.
+
+        Returns:
+            datatable (DataTable): The rendered datatable with parameters replaced only if they exist in the context.
+        """
+        rendered_datatable = copy.deepcopy(datatable)
+        for row in rendered_datatable.rows:
+            for cell in row.cells:
+                cell.value = render_string(cell.value, context)
+        return rendered_datatable
+
+
+@dataclass(eq=False)
+class Background:
+    """Represents the background steps for a feature.
+
+    Attributes:
+        line_number (int): The line number where the background starts in the file.
+        steps (List[Step]): The list of steps in the background.
+    """
+
+    line_number: int
+    steps: list[Step] = field(init=False, default_factory=list)
+
+    def add_step(self, step: Step) -> None:
+        """Add a step to the background.
+
+        Args:
+            step (Step): The step to add.
+        """
+        step.background = self
+        self.steps.append(step)
+
+
+class FeatureParser:
+    """Converts a feature file into a Feature object.
+
+    Args:
+        basedir (str): The basedir for locating feature files.
+        filename (str): The filename of the feature file.
+        encoding (str): File encoding of the feature file to parse.
+    """
+
+    def __init__(self, basedir: str, filename: str, encoding: str = "utf-8"):
+        self.abs_filename = os.path.abspath(os.path.join(basedir, filename))
+        self.rel_filename = os.path.join(os.path.basename(basedir), filename)
+        self.encoding = encoding
+
+    def parse_steps(self, steps_data: list[GherkinStep]) -> list[Step]:
+        """Parse a list of step data into Step objects.
+
+        Args:
+            steps_data (List[dict]): The list of step data.
+
+        Returns:
+            List[Step]: A list of Step objects.
+        """
+
+        if not steps_data:
+            return []
+
+        first_step = steps_data[0]
+        if first_step.keyword_type not in STEP_TYPE_BY_PARSER_KEYWORD:
+            raise StepError(
+                message=f"First step in a scenario or background must start with 'Given', 'When' or 'Then', but got {first_step.keyword}.",
+                line=first_step.location.line,
+                line_content=first_step.text,
+                filename=self.abs_filename,
+            )
+
+        steps = []
+        current_type = STEP_TYPE_BY_PARSER_KEYWORD[first_step.keyword_type]
+        for step in steps_data:
+            current_type = STEP_TYPE_BY_PARSER_KEYWORD.get(step.keyword_type, current_type)
+            steps.append(
+                Step(
+                    name=step.text,
+                    type=current_type,
+                    indent=step.location.column - 1,
+                    line_number=step.location.line,
+                    keyword=step.keyword.title(),
+                    datatable=step.datatable,
+                    docstring=step.docstring.content if step.docstring else None,
+                )
+            )
+        return steps
+
+    def parse_scenario(
+        self, scenario_data: GherkinScenario, feature: Feature, rule: Rule | None = None
+    ) -> ScenarioTemplate:
+        """Parse a scenario data dictionary into a ScenarioTemplate object.
+
+        Args:
+            scenario_data (dict): The dictionary containing scenario data.
+            feature (Feature): The feature to which this scenario belongs.
+            rule (Optional[Rule]): The rule to which this scenario may belong. (None = no rule)
+
+        Returns:
+            ScenarioTemplate: A ScenarioTemplate object representing the parsed scenario.
+        """
+        templated = bool(scenario_data.examples)
+        scenario = ScenarioTemplate(
+            feature=feature,
+            keyword=scenario_data.keyword,
+            name=scenario_data.name,
+            line_number=scenario_data.location.line,
+            templated=templated,
+            tags=get_tag_names(scenario_data.tags),
+            description=textwrap.dedent(scenario_data.description),
+            rule=rule,
+        )
+        for step in self.parse_steps(scenario_data.steps):
+            scenario.add_step(step)
+
+        # Loop over multiple example tables if they exist
+        for example_data in scenario_data.examples:
+            examples = Examples(
+                line_number=example_data.location.line,
+                name=example_data.name,
+                tags=get_tag_names(example_data.tags),
+            )
+            if example_data.table_header is not None:
+                param_names = [cell.value for cell in example_data.table_header.cells]
+                examples.set_param_names(param_names)
+                if example_data.table_body is not None:
+                    for row in example_data.table_body:
+                        values = [cell.value or "" for cell in row.cells]
+                        examples.add_example(values)
+                scenario.examples.append(examples)
+
+        return scenario
+
+    def parse_background(self, background_data: GherkinBackground) -> Background:
+        background = Background(
+            line_number=background_data.location.line,
+        )
+        background.steps = self.parse_steps(background_data.steps)
+        for step in background.steps:
+            step.background = background
+        return background
+
+    def _parse_feature_file(self) -> GherkinDocument:
+        """Parse a feature file into a Feature object.
+
+        Returns:
+            Dict: A Gherkin document representation of the feature file.
+        """
+        return get_gherkin_document(self.abs_filename, self.encoding)
+
+    def parse(self) -> Feature:
+        """Parse the feature file and return a Feature object with its backgrounds, rules, and scenarios."""
+        gherkin_doc: GherkinDocument = self._parse_feature_file()
+        feature_data: GherkinFeature = gherkin_doc.feature
+        feature = Feature(
+            scenarios=OrderedDict(),
+            keyword=feature_data.keyword,
+            filename=self.abs_filename,
+            rel_filename=self.rel_filename,
+            name=feature_data.name,
+            tags=get_tag_names(feature_data.tags),
+            background=None,
+            line_number=feature_data.location.line,
+            description=textwrap.dedent(feature_data.description),
+            language=feature_data.language,
+        )
+
+        for child in feature_data.children:
+            if child.background:
+                feature.background = self.parse_background(child.background)
+            elif child.rule:
+                self._parse_and_add_rule(child.rule, feature)
+            elif child.scenario:
+                self._parse_and_add_scenario(child.scenario, feature)
+
+        return feature
+
+    def _parse_and_add_rule(self, rule_data: GherkinRule, feature: Feature) -> None:
+        """Parse a rule, including its background and scenarios, and add to the feature."""
+        background = self._extract_rule_background(rule_data)
+
+        rule = Rule(
+            keyword=rule_data.keyword,
+            name=rule_data.name,
+            description=rule_data.description,
+            tags=get_tag_names(rule_data.tags),
+            background=background,
+        )
+
+        for scenario in self._extract_rule_scenarios(rule_data, feature, rule):
+            feature.scenarios[scenario.name] = scenario
+
+    def _extract_rule_background(self, rule_data: GherkinRule) -> Background | None:
+        """Extract the first background from rule children if it exists."""
+        for child in rule_data.children:
+            if child.background:
+                return self.parse_background(child.background)
+        return None
+
+    def _extract_rule_scenarios(
+        self, rule_data: GherkinRule, feature: Feature, rule: Rule
+    ) -> Generator[ScenarioTemplate]:
+        """Yield each parsed scenario under a rule."""
+        for child in rule_data.children:
+            if child.scenario:
+                yield self.parse_scenario(child.scenario, feature, rule)
+
+    def _parse_and_add_scenario(
+        self, scenario_data: GherkinScenario, feature: Feature, rule: Rule | None = None
+    ) -> None:
+        """Parse an individual scenario and add it to the feature's scenarios."""
+        scenario = self.parse_scenario(scenario_data, feature, rule)
+        feature.scenarios[scenario.name] = scenario
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/parsers.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/parsers.py
new file mode 100644
index 00000000..26b348dd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/parsers.py
@@ -0,0 +1,120 @@
+"""Step parsers."""
+
+from __future__ import annotations
+
+import abc
+import re as base_re
+from typing import Any, TypeVar, cast, overload
+
+import parse as base_parse
+from parse_type import cfparse as base_cfparse
+
+
+class StepParser(abc.ABC):
+    """Parser of the individual step."""
+
+    def __init__(self, name: str) -> None:
+        self.name = name
+
+    @abc.abstractmethod
+    def parse_arguments(self, name: str) -> dict[str, Any] | None:
+        """Get step arguments from the given step name.
+
+        :return: `dict` of step arguments
+        """
+        ...
+
+    @abc.abstractmethod
+    def is_matching(self, name: str) -> bool:
+        """Match given name with the step name."""
+        ...
+
+
+class re(StepParser):
+    """Regex step parser."""
+
+    def __init__(self, name: str, *args: Any, **kwargs: Any) -> None:
+        """Compile regex."""
+        super().__init__(name)
+        self.regex = base_re.compile(self.name, *args, **kwargs)
+
+    def parse_arguments(self, name: str) -> dict[str, str] | None:
+        """Get step arguments.
+
+        :return: `dict` of step arguments
+        """
+        match = self.regex.fullmatch(name)
+        if match is None:
+            return None
+        return match.groupdict()
+
+    def is_matching(self, name: str) -> bool:
+        """Match given name with the step name."""
+        return bool(self.regex.fullmatch(name))
+
+
+class parse(StepParser):
+    """parse step parser."""
+
+    def __init__(self, name: str, *args: Any, **kwargs: Any) -> None:
+        """Compile parse expression."""
+        super().__init__(name)
+        self.parser = base_parse.compile(self.name, *args, **kwargs)
+
+    def parse_arguments(self, name: str) -> dict[str, Any]:
+        """Get step arguments.
+
+        :return: `dict` of step arguments
+        """
+        return cast(dict[str, Any], self.parser.parse(name).named)
+
+    def is_matching(self, name: str) -> bool:
+        """Match given name with the step name."""
+        try:
+            return bool(self.parser.parse(name))
+        except ValueError:
+            return False
+
+
+class cfparse(parse):
+    """cfparse step parser."""
+
+    def __init__(self, name: str, *args: Any, **kwargs: Any) -> None:
+        """Compile parse expression."""
+        super(parse, self).__init__(name)
+        self.parser = base_cfparse.Parser(self.name, *args, **kwargs)
+
+
+class string(StepParser):
+    """Exact string step parser."""
+
+    def parse_arguments(self, name: str) -> dict:
+        """No parameters are available for simple string step.
+
+        :return: `dict` of step arguments
+        """
+        return {}
+
+    def is_matching(self, name: str) -> bool:
+        """Match given name with the step name."""
+        return self.name == name
+
+
+TStepParser = TypeVar("TStepParser", bound=StepParser)
+
+
+@overload
+def get_parser(step_name: str) -> string: ...
+
+
+@overload
+def get_parser(step_name: TStepParser) -> TStepParser: ...
+
+
+def get_parser(step_name: str | StepParser) -> StepParser:
+    """Get parser by given name."""
+
+    if isinstance(step_name, StepParser):
+        return step_name
+
+    return string(step_name)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/plugin.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/plugin.py
new file mode 100644
index 00000000..c397134b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/plugin.py
@@ -0,0 +1,139 @@
+"""Pytest plugin entry point. Used for any fixtures needed."""
+
+from __future__ import annotations
+
+from collections.abc import Generator
+from typing import TYPE_CHECKING, Any, Callable, TypeVar, cast
+
+import pytest
+from typing_extensions import ParamSpec
+
+from . import cucumber_json, generation, gherkin_terminal_reporter, given, reporting, then, when
+from .utils import CONFIG_STACK
+
+if TYPE_CHECKING:
+    from _pytest.config import Config, PytestPluginManager
+    from _pytest.config.argparsing import Parser
+    from _pytest.fixtures import FixtureRequest
+    from _pytest.nodes import Item
+    from _pytest.runner import CallInfo
+    from pluggy._result import _Result
+
+    from .parser import Feature, Scenario, Step
+
+
+P = ParamSpec("P")
+T = TypeVar("T")
+
+
+def pytest_addhooks(pluginmanager: PytestPluginManager) -> None:
+    """Register plugin hooks."""
+    from pytest_bdd import hooks
+
+    pluginmanager.add_hookspecs(hooks)
+
+
+@given("trace")
+@when("trace")
+@then("trace")
+def _() -> None:
+    """Enter pytest's pdb trace."""
+    pytest.set_trace()
+
+
+@pytest.fixture
+def _pytest_bdd_example() -> dict:
+    """The current scenario outline parametrization.
+
+    This is used internally by pytest_bdd.
+
+    If no outline is used, we just return an empty dict to render
+    the current template without any actual variable.
+    Otherwise, pytest_bdd will add all the context variables in this fixture
+    from the example definitions in the feature file.
+    """
+    return {}
+
+
+def pytest_addoption(parser: Parser) -> None:
+    """Add pytest-bdd options."""
+    add_bdd_ini(parser)
+    cucumber_json.add_options(parser)
+    generation.add_options(parser)
+    gherkin_terminal_reporter.add_options(parser)
+
+
+def add_bdd_ini(parser: Parser) -> None:
+    parser.addini("bdd_features_base_dir", "Base features directory.")
+
+
+@pytest.hookimpl(trylast=True)
+def pytest_configure(config: Config) -> None:
+    """Configure all subplugins."""
+    CONFIG_STACK.append(config)
+    cucumber_json.configure(config)
+    gherkin_terminal_reporter.configure(config)
+
+
+def pytest_unconfigure(config: Config) -> None:
+    """Unconfigure all subplugins."""
+    if CONFIG_STACK:
+        CONFIG_STACK.pop()
+    cucumber_json.unconfigure(config)
+
+
+@pytest.hookimpl(hookwrapper=True)
+def pytest_runtest_makereport(item: Item, call: CallInfo) -> Generator[None, _Result, None]:
+    outcome = yield
+    reporting.runtest_makereport(item, call, outcome.get_result())
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_bdd_before_scenario(request: FixtureRequest, feature: Feature, scenario: Scenario) -> None:
+    reporting.before_scenario(request, feature, scenario)
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_bdd_step_error(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., Any],
+    step_func_args: dict,
+    exception: Exception,
+) -> None:
+    reporting.step_error(request, feature, scenario, step, step_func, step_func_args, exception)
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_bdd_before_step(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., Any],
+) -> None:
+    reporting.before_step(request, feature, scenario, step, step_func)
+
+
+@pytest.hookimpl(tryfirst=True)
+def pytest_bdd_after_step(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., Any],
+    step_func_args: dict[str, Any],
+) -> None:
+    reporting.after_step(request, feature, scenario, step, step_func, step_func_args)
+
+
+def pytest_cmdline_main(config: Config) -> int | None:
+    return generation.cmdline_main(config)
+
+
+def pytest_bdd_apply_tag(tag: str, function: Callable[P, T]) -> Callable[P, T]:
+    mark = getattr(pytest.mark, tag)
+    marked = mark(function)
+    return cast(Callable[P, T], marked)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/py.typed b/backend/venv/lib/python3.12/site-packages/pytest_bdd/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/reporting.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/reporting.py
new file mode 100644
index 00000000..e92cc3ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/reporting.py
@@ -0,0 +1,197 @@
+"""Reporting functionality.
+
+Collection of the scenario execution statuses, timing and other information
+that enriches the pytest test reporting.
+"""
+
+from __future__ import annotations
+
+import time
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing import Any, Callable
+
+    from _pytest.fixtures import FixtureRequest
+    from _pytest.nodes import Item
+    from _pytest.reports import TestReport
+    from _pytest.runner import CallInfo
+
+    from .parser import Feature, Scenario, Step
+
+
+class StepReport:
+    """Step execution report."""
+
+    failed = False
+    stopped = None
+
+    def __init__(self, step: Step) -> None:
+        """Step report constructor.
+
+        :param pytest_bdd.parser.Step step: Step.
+        """
+        self.step = step
+        self.started = time.perf_counter()
+
+    def serialize(self) -> dict[str, Any]:
+        """Serialize the step execution report.
+
+        :return: Serialized step execution report.
+        :rtype: dict
+        """
+        return {
+            "name": self.step.name,
+            "type": self.step.type,
+            "keyword": self.step.keyword,
+            "line_number": self.step.line_number,
+            "failed": self.failed,
+            "duration": self.duration,
+        }
+
+    def finalize(self, failed: bool) -> None:
+        """Stop collecting information and finalize the report.
+
+        :param bool failed: Whether the step execution is failed.
+        """
+        self.stopped = time.perf_counter()
+        self.failed = failed
+
+    @property
+    def duration(self) -> float:
+        """Step execution duration.
+
+        :return: Step execution duration.
+        :rtype: float
+        """
+        if self.stopped is None:
+            return 0
+
+        return self.stopped - self.started
+
+
+class ScenarioReport:
+    """Scenario execution report."""
+
+    def __init__(self, scenario: Scenario) -> None:
+        """Scenario report constructor.
+
+        :param pytest_bdd.parser.Scenario scenario: Scenario.
+        """
+        self.scenario: Scenario = scenario
+        self.step_reports: list[StepReport] = []
+
+    @property
+    def current_step_report(self) -> StepReport:
+        """Get current step report.
+
+        :return: Last or current step report.
+        :rtype: pytest_bdd.reporting.StepReport
+        """
+        return self.step_reports[-1]
+
+    def add_step_report(self, step_report: StepReport) -> None:
+        """Add new step report.
+
+        :param step_report: New current step report.
+        :type step_report: pytest_bdd.reporting.StepReport
+        """
+        self.step_reports.append(step_report)
+
+    def serialize(self) -> dict[str, Any]:
+        """Serialize scenario execution report in order to transfer reporting from nodes in the distributed mode.
+
+        :return: Serialized report.
+        :rtype: dict
+        """
+        scenario = self.scenario
+        feature = scenario.feature
+
+        serialized = {
+            "steps": [step_report.serialize() for step_report in self.step_reports],
+            "keyword": scenario.keyword,
+            "name": scenario.name,
+            "line_number": scenario.line_number,
+            "tags": sorted(scenario.tags),
+            "description": scenario.description,
+            "feature": {
+                "keyword": feature.keyword,
+                "name": feature.name,
+                "filename": feature.filename,
+                "rel_filename": feature.rel_filename,
+                "language": feature.language,
+                "line_number": feature.line_number,
+                "description": feature.description,
+                "tags": sorted(feature.tags),
+            },
+        }
+
+        if scenario.rule:
+            serialized["rule"] = {
+                "keyword": scenario.rule.keyword,
+                "name": scenario.rule.name,
+                "description": scenario.rule.description,
+                "tags": scenario.rule.tags,
+            }
+
+        return serialized
+
+    def fail(self) -> None:
+        """Stop collecting information and finalize the report as failed."""
+        self.current_step_report.finalize(failed=True)
+        remaining_steps = self.scenario.steps[len(self.step_reports) :]
+
+        # Fail the rest of the steps and make reports.
+        for step in remaining_steps:
+            report = StepReport(step=step)
+            report.finalize(failed=True)
+            self.add_step_report(report)
+
+
+def runtest_makereport(item: Item, call: CallInfo, rep: TestReport) -> None:
+    """Store item in the report object."""
+    scenario_report = getattr(item, "__scenario_report__", None)
+    if scenario_report is not None:
+        rep.scenario = scenario_report.serialize()  # type: ignore
+        rep.item = {"name": item.name}  # type: ignore
+
+
+def before_scenario(request: FixtureRequest, feature: Feature, scenario: Scenario) -> None:
+    """Create scenario report for the item."""
+    request.node.__scenario_report__ = ScenarioReport(scenario=scenario)
+
+
+def step_error(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., Any],
+    step_func_args: dict,
+    exception: Exception,
+) -> None:
+    """Finalize the step report as failed."""
+    request.node.__scenario_report__.fail()
+
+
+def before_step(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable[..., Any],
+) -> None:
+    """Store step start time."""
+    request.node.__scenario_report__.add_step_report(StepReport(step=step))
+
+
+def after_step(
+    request: FixtureRequest,
+    feature: Feature,
+    scenario: Scenario,
+    step: Step,
+    step_func: Callable,
+    step_func_args: dict,
+) -> None:
+    """Finalize the step report as successful."""
+    request.node.__scenario_report__.current_step_report.finalize(failed=False)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/scenario.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/scenario.py
new file mode 100644
index 00000000..cb2a126f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/scenario.py
@@ -0,0 +1,479 @@
+"""Scenario implementation.
+
+The pytest will collect the test case and the steps will be executed
+line by line.
+
+Example:
+
+test_publish_article = scenario(
+    feature_name="publish_article.feature",
+    scenario_name="Publishing the article",
+)
+"""
+
+from __future__ import annotations
+
+import contextlib
+import logging
+import os
+import re
+from collections.abc import Iterable, Iterator
+from inspect import signature
+from typing import TYPE_CHECKING, Any, Callable, TypeVar, cast
+
+import pytest
+from _pytest.fixtures import FixtureDef, FixtureManager, FixtureRequest, call_fixture_func
+from typing_extensions import ParamSpec
+
+from . import exceptions
+from .compat import getfixturedefs, inject_fixture
+from .feature import get_feature, get_features
+from .steps import StepFunctionContext, get_step_fixture_name
+from .utils import CONFIG_STACK, get_caller_module_locals, get_caller_module_path, get_required_args, identity
+
+if TYPE_CHECKING:
+    from _pytest.mark.structures import ParameterSet
+    from _pytest.nodes import Node
+
+    from .parser import Feature, Scenario, ScenarioTemplate, Step
+
+P = ParamSpec("P")
+T = TypeVar("T")
+
+logger = logging.getLogger(__name__)
+
+PYTHON_REPLACE_REGEX = re.compile(r"\W")
+ALPHA_REGEX = re.compile(r"^\d+_*")
+
+STEP_ARGUMENT_DATATABLE = "datatable"
+STEP_ARGUMENT_DOCSTRING = "docstring"
+STEP_ARGUMENTS_RESERVED_NAMES = {STEP_ARGUMENT_DATATABLE, STEP_ARGUMENT_DOCSTRING}
+
+
+def find_fixturedefs_for_step(step: Step, fixturemanager: FixtureManager, node: Node) -> Iterable[FixtureDef[Any]]:
+    """Find the fixture defs that can parse a step."""
+    # happens to be that _arg2fixturedefs is changed during the iteration so we use a copy
+    fixture_def_by_name = list(fixturemanager._arg2fixturedefs.items())
+    for fixturename, fixturedefs in fixture_def_by_name:
+        for _, fixturedef in enumerate(fixturedefs):
+            step_func_context = getattr(fixturedef.func, "_pytest_bdd_step_context", None)
+            if step_func_context is None:
+                continue
+
+            if step_func_context.type is not None and step_func_context.type != step.type:
+                continue
+
+            match = step_func_context.parser.is_matching(step.name)
+            if not match:
+                continue
+
+            fixturedefs = cast(list[FixtureDef[Any]], getfixturedefs(fixturemanager, fixturename, node) or [])
+            if fixturedef not in fixturedefs:
+                continue
+
+            yield fixturedef
+
+
+# Function copied from pytest 8.0 (removed in later versions).
+def iterparentnodeids(nodeid: str) -> Iterator[str]:
+    """Return the parent node IDs of a given node ID, inclusive.
+
+    For the node ID
+
+        "testing/code/test_excinfo.py::TestFormattedExcinfo::test_repr_source"
+
+    the result would be
+
+        ""
+        "testing"
+        "testing/code"
+        "testing/code/test_excinfo.py"
+        "testing/code/test_excinfo.py::TestFormattedExcinfo"
+        "testing/code/test_excinfo.py::TestFormattedExcinfo::test_repr_source"
+
+    Note that / components are only considered until the first ::.
+    """
+    SEP = "/"
+    pos = 0
+    first_colons: int | None = nodeid.find("::")
+    if first_colons == -1:
+        first_colons = None
+    # The root Session node - always present.
+    yield ""
+    # Eagerly consume SEP parts until first colons.
+    while True:
+        at = nodeid.find(SEP, pos, first_colons)
+        if at == -1:
+            break
+        if at > 0:
+            yield nodeid[:at]
+        pos = at + len(SEP)
+    # Eagerly consume :: parts.
+    while True:
+        at = nodeid.find("::", pos)
+        if at == -1:
+            break
+        if at > 0:
+            yield nodeid[:at]
+        pos = at + len("::")
+    # The node ID itself.
+    if nodeid:
+        yield nodeid
+
+
+@contextlib.contextmanager
+def inject_fixturedefs_for_step(step: Step, fixturemanager: FixtureManager, node: Node) -> Iterator[None]:
+    """Inject fixture definitions that can parse a step.
+
+    We fist iterate over all the fixturedefs that can parse the step.
+
+    Then we sort them by their "path" (list of parent IDs) so that we respect the fixture scoping rules.
+
+    Finally, we inject them into the request.
+    """
+    bdd_name = get_step_fixture_name(step=step)
+
+    fixturedefs = list(find_fixturedefs_for_step(step=step, fixturemanager=fixturemanager, node=node))
+
+    # Sort the fixture definitions by their "path", so that the `bdd_name` fixture will
+    # respect the fixture scope
+
+    def get_fixture_path(fixture_def: FixtureDef) -> list[str]:
+        return list(iterparentnodeids(fixture_def.baseid))
+
+    fixturedefs.sort(key=lambda x: get_fixture_path(x))
+
+    if not fixturedefs:
+        yield
+        return
+
+    logger.debug("Adding providers for fixture %r: %r", bdd_name, fixturedefs)
+    fixturemanager._arg2fixturedefs[bdd_name] = fixturedefs
+
+    try:
+        yield
+    finally:
+        del fixturemanager._arg2fixturedefs[bdd_name]
+
+
+def get_step_function(request: FixtureRequest, step: Step) -> StepFunctionContext | None:
+    """Get the step function (context) for the given step.
+
+    We first figure out what's the step fixture name that we have to inject.
+
+    Then we let `patch_argumented_step_functions` find out what step definition fixtures can parse the current step,
+    and it will inject them for the step fixture name.
+
+    Finally, we let request.getfixturevalue(...) fetch the step definition fixture.
+    """
+    __tracebackhide__ = True
+    bdd_name = get_step_fixture_name(step=step)
+
+    with inject_fixturedefs_for_step(step=step, fixturemanager=request._fixturemanager, node=request.node):
+        try:
+            return cast(StepFunctionContext, request.getfixturevalue(bdd_name))
+        except pytest.FixtureLookupError:
+            return None
+
+
+def parse_step_arguments(step: Step, context: StepFunctionContext) -> dict[str, object]:
+    """Parse step arguments."""
+    parsed_args = context.parser.parse_arguments(step.name)
+
+    assert parsed_args is not None, (
+        f"Unexpected `NoneType` returned from " f"parse_arguments(...) in parser: {context.parser!r}"
+    )
+
+    reserved_args = set(parsed_args.keys()) & STEP_ARGUMENTS_RESERVED_NAMES
+    if reserved_args:
+        reserved_arguments_str = ", ".join(repr(arg) for arg in reserved_args)
+        raise exceptions.StepImplementationError(
+            f"Step {step.name!r} defines argument names that are reserved: {reserved_arguments_str}. "
+            "Please use different names."
+        )
+
+    converted_args = {key: (context.converters.get(key, identity)(value)) for key, value in parsed_args.items()}
+
+    return converted_args
+
+
+def _execute_step_function(
+    request: FixtureRequest, scenario: Scenario, step: Step, context: StepFunctionContext
+) -> None:
+    """Execute step function."""
+    __tracebackhide__ = True
+
+    func_sig = signature(context.step_func)
+
+    kw = {
+        "request": request,
+        "feature": scenario.feature,
+        "scenario": scenario,
+        "step": step,
+        "step_func": context.step_func,
+        "step_func_args": {},
+    }
+    request.config.hook.pytest_bdd_before_step(**kw)
+
+    try:
+        parsed_args = parse_step_arguments(step=step, context=context)
+
+        # Filter out the arguments that are not in the function signature
+        kwargs = {k: v for k, v in parsed_args.items() if k in func_sig.parameters}
+
+        if STEP_ARGUMENT_DATATABLE in func_sig.parameters and step.datatable is not None:
+            kwargs[STEP_ARGUMENT_DATATABLE] = step.datatable.raw()
+        if STEP_ARGUMENT_DOCSTRING in func_sig.parameters and step.docstring is not None:
+            kwargs[STEP_ARGUMENT_DOCSTRING] = step.docstring
+
+        # Fill the missing arguments requesting the fixture values
+        kwargs |= {
+            arg: request.getfixturevalue(arg) for arg in get_required_args(context.step_func) if arg not in kwargs
+        }
+
+        kw["step_func_args"] = kwargs
+
+        request.config.hook.pytest_bdd_before_step_call(**kw)
+
+        # Execute the step as if it was a pytest fixture using `call_fixture_func`,
+        # so that we can allow "yield" statements in it
+        return_value = call_fixture_func(fixturefunc=context.step_func, request=request, kwargs=kwargs)
+
+    except Exception as exception:
+        request.config.hook.pytest_bdd_step_error(exception=exception, **kw)
+        raise
+
+    if context.target_fixture is not None:
+        inject_fixture(request, context.target_fixture, return_value)
+
+    request.config.hook.pytest_bdd_after_step(**kw)
+
+
+def _execute_scenario(feature: Feature, scenario: Scenario, request: FixtureRequest) -> None:
+    """Execute the scenario.
+
+    :param feature: Feature.
+    :param scenario: Scenario.
+    :param request: request.
+    """
+    __tracebackhide__ = True
+    request.config.hook.pytest_bdd_before_scenario(request=request, feature=feature, scenario=scenario)
+
+    try:
+        for step in scenario.steps:
+            step_func_context = get_step_function(request=request, step=step)
+            if step_func_context is None:
+                exc = exceptions.StepDefinitionNotFoundError(
+                    f"Step definition is not found: {step}. "
+                    f'Line {step.line_number} in scenario "{scenario.name}" in the feature "{scenario.feature.filename}"'
+                )
+                request.config.hook.pytest_bdd_step_func_lookup_error(
+                    request=request, feature=feature, scenario=scenario, step=step, exception=exc
+                )
+                raise exc
+            _execute_step_function(request, scenario, step, step_func_context)
+    finally:
+        request.config.hook.pytest_bdd_after_scenario(request=request, feature=feature, scenario=scenario)
+
+
+def _get_scenario_decorator(
+    feature: Feature, feature_name: str, templated_scenario: ScenarioTemplate, scenario_name: str
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    # HACK: Ideally we would use `def decorator(fn)`, but we want to return a custom exception
+    # when the decorator is misused.
+    # Pytest inspect the signature to determine the required fixtures, and in that case it would look
+    # for a fixture called "fn" that doesn't exist (if it exists then it's even worse).
+    # It will error with a "fixture 'fn' not found" message instead.
+    # We can avoid this hack by using a pytest hook and check for misuse instead.
+    def decorator(*args: Callable[P, T]) -> Callable[P, T]:
+        if not args:
+            raise exceptions.ScenarioIsDecoratorOnly(
+                "scenario function can only be used as a decorator. Refer to the documentation."
+            )
+        [fn] = args
+        func_args = get_required_args(fn)
+
+        def scenario_wrapper(request: FixtureRequest, _pytest_bdd_example: dict[str, str]) -> Any:
+            __tracebackhide__ = True
+            scenario = templated_scenario.render(_pytest_bdd_example)
+            _execute_scenario(feature, scenario, request)
+            fixture_values = [request.getfixturevalue(arg) for arg in func_args]
+            return fn(*fixture_values)
+
+        if func_args:
+            # We need to tell pytest that the original function requires its fixtures,
+            # otherwise indirect fixtures would not work.
+            scenario_wrapper = pytest.mark.usefixtures(*func_args)(scenario_wrapper)
+
+        example_parametrizations = collect_example_parametrizations(templated_scenario)
+        if example_parametrizations is not None:
+            # Parametrize the scenario outlines
+            scenario_wrapper = pytest.mark.parametrize(
+                "_pytest_bdd_example",
+                example_parametrizations,
+            )(scenario_wrapper)
+
+        rule_tags = set() if templated_scenario.rule is None else templated_scenario.rule.tags
+        for tag in templated_scenario.tags | feature.tags | rule_tags:
+            config = CONFIG_STACK[-1]
+            config.hook.pytest_bdd_apply_tag(tag=tag, function=scenario_wrapper)
+
+        scenario_wrapper.__doc__ = f"{feature_name}: {scenario_name}"
+        scenario_wrapper.__scenario__ = templated_scenario  # type: ignore[attr-defined]
+        return cast(Callable[P, T], scenario_wrapper)
+
+    return decorator
+
+
+def collect_example_parametrizations(
+    templated_scenario: ScenarioTemplate,
+) -> list[ParameterSet] | None:
+    parametrizations = []
+
+    for examples in templated_scenario.examples:
+        tags: set = examples.tags or set()
+
+        example_marks = [getattr(pytest.mark, tag) for tag in tags]
+
+        for context in examples.as_contexts():
+            param_id = "-".join(context.values())
+            parametrizations.append(
+                pytest.param(
+                    context,
+                    id=param_id,
+                    marks=example_marks,
+                ),
+            )
+
+    return parametrizations or None
+
+
+def scenario(
+    feature_name: str,
+    scenario_name: str,
+    encoding: str = "utf-8",
+    features_base_dir: str | None = None,
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """Scenario decorator.
+
+    :param str feature_name: Feature file name. Absolute or relative to the configured feature base path.
+    :param str scenario_name: Scenario name.
+    :param str encoding: Feature file encoding.
+    :param features_base_dir: Optional base dir location for locating feature files. If not set, it will try and resolve using property set in .ini file, then the caller_module_path.
+    """
+    __tracebackhide__ = True
+    scenario_name = scenario_name
+    caller_module_path = get_caller_module_path()
+
+    # Get the feature
+    if features_base_dir is None:
+        features_base_dir = get_features_base_dir(caller_module_path)
+    feature = get_feature(features_base_dir, feature_name, encoding=encoding)
+
+    # Get the scenario
+    try:
+        scenario = feature.scenarios[scenario_name]
+    except KeyError:
+        feature_name = feature.name or "[Empty]"
+        raise exceptions.ScenarioNotFound(
+            f'Scenario "{scenario_name}" in feature "{feature_name}" in {feature.filename} is not found.'
+        )
+
+    return _get_scenario_decorator(
+        feature=feature, feature_name=feature_name, templated_scenario=scenario, scenario_name=scenario_name
+    )
+
+
+def get_features_base_dir(caller_module_path: str) -> str:
+    d = get_from_ini("bdd_features_base_dir")
+    if d is None:
+        return os.path.dirname(caller_module_path)
+    rootdir = CONFIG_STACK[-1].rootpath
+    return os.path.join(rootdir, d)
+
+
+def get_from_ini(key: str, default: str | None = None) -> str | None:
+    """Get value from ini config. Return default if value has not been set.
+
+    Use if the default value is dynamic. Otherwise, set default on addini call.
+    """
+    config = CONFIG_STACK[-1]
+    value = config.getini(key)
+    if not isinstance(value, str):
+        raise TypeError(f"Expected a string for configuration option {value!r}, got a {type(value)} instead")
+    return value if value != "" else default
+
+
+def make_python_name(string: str) -> str:
+    """Make python attribute name out of a given string."""
+    string = re.sub(PYTHON_REPLACE_REGEX, "", string.replace(" ", "_"))
+    return re.sub(ALPHA_REGEX, "", string).lower()
+
+
+def make_python_docstring(string: str) -> str:
+    """Make a python docstring literal out of a given string."""
+    return '"""{}."""'.format(string.replace('"""', '\\"\\"\\"'))
+
+
+def make_string_literal(string: str) -> str:
+    """Make python string literal out of a given string."""
+    return "'{}'".format(string.replace("'", "\\'"))
+
+
+def get_python_name_generator(name: str) -> Iterable[str]:
+    """Generate a sequence of suitable python names out of given arbitrary string name."""
+    python_name = make_python_name(name)
+    suffix = ""
+    index = 0
+
+    def get_name() -> str:
+        return f"test_{python_name}{suffix}"
+
+    while True:
+        yield get_name()
+        index += 1
+        suffix = f"_{index}"
+
+
+def scenarios(*feature_paths: str, **kwargs: Any) -> None:
+    """Parse features from the paths and put all found scenarios in the caller module.
+
+    :param *feature_paths: feature file paths to use for scenarios
+    """
+    caller_locals = get_caller_module_locals()
+    caller_path = get_caller_module_path()
+
+    features_base_dir = kwargs.get("features_base_dir")
+    if features_base_dir is None:
+        features_base_dir = get_features_base_dir(caller_path)
+
+    abs_feature_paths = []
+    for path in feature_paths:
+        if not os.path.isabs(path):
+            path = os.path.abspath(os.path.join(features_base_dir, path))
+        abs_feature_paths.append(path)
+    found = False
+
+    module_scenarios = frozenset(
+        (attr.__scenario__.feature.filename, attr.__scenario__.name)
+        for name, attr in caller_locals.items()
+        if hasattr(attr, "__scenario__")
+    )
+
+    for feature in get_features(abs_feature_paths):
+        for scenario_name, scenario_object in feature.scenarios.items():
+            # skip already bound scenarios
+            if (scenario_object.feature.filename, scenario_name) not in module_scenarios:
+
+                @scenario(feature.filename, scenario_name, **kwargs)
+                def _scenario() -> None:
+                    pass  # pragma: no cover
+
+                for test_name in get_python_name_generator(scenario_name):
+                    if test_name not in caller_locals:
+                        # found a unique test name
+                        caller_locals[test_name] = _scenario
+                        break
+            found = True
+    if not found:
+        raise exceptions.NoScenariosFound(abs_feature_paths)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/scripts.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/scripts.py
new file mode 100644
index 00000000..562575b3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/scripts.py
@@ -0,0 +1,76 @@
+"""pytest-bdd scripts."""
+
+from __future__ import annotations
+
+import argparse
+import glob
+import os.path
+import re
+
+from .generation import generate_code, parse_feature_files
+
+MIGRATE_REGEX = re.compile(r"\s?(\w+)\s=\sscenario\((.+)\)", flags=re.MULTILINE)
+
+
+def migrate_tests(args: argparse.Namespace) -> None:
+    """Migrate outdated tests to the most recent form."""
+    path = args.path
+    for file_path in glob.iglob(os.path.join(os.path.abspath(path), "**", "*.py"), recursive=True):
+        migrate_tests_in_file(file_path)
+
+
+def migrate_tests_in_file(file_path: str) -> None:
+    """Migrate all bdd-based tests in the given test file."""
+    try:
+        with open(file_path, "r+") as fd:
+            content = fd.read()
+            new_content = MIGRATE_REGEX.sub(r"\n@scenario(\2)\ndef \1():\n    pass\n", content)
+            if new_content != content:
+                # the regex above potentially causes the end of the file to
+                # have an extra newline
+                new_content = new_content.rstrip("\n") + "\n"
+                fd.seek(0)
+                fd.write(new_content)
+                print(f"migrated: {file_path}")
+            else:
+                print(f"skipped: {file_path}")
+    except OSError:
+        pass
+
+
+def check_existense(file_name: str) -> str:
+    """Check file or directory name for existence."""
+    if not os.path.exists(file_name):
+        raise argparse.ArgumentTypeError(f"{file_name} is an invalid file or directory name")
+    return file_name
+
+
+def print_generated_code(args: argparse.Namespace) -> None:
+    """Print generated test code for the given filenames."""
+    features, scenarios, steps = parse_feature_files(args.files)
+    code = generate_code(features, scenarios, steps)
+    print(code)
+
+
+def main() -> None:
+    """Main entry point."""
+    parser = argparse.ArgumentParser(prog="pytest-bdd")
+    subparsers = parser.add_subparsers(help="sub-command help", dest="command")
+    subparsers.required = True
+    parser_generate = subparsers.add_parser("generate", help="generate help")
+    parser_generate.add_argument(
+        "files",
+        metavar="FEATURE_FILE",
+        type=check_existense,
+        nargs="+",
+        help="Feature files to generate test code with",
+    )
+    parser_generate.set_defaults(func=print_generated_code)
+
+    parser_migrate = subparsers.add_parser("migrate", help="migrate help")
+    parser_migrate.add_argument("path", metavar="PATH", help="Migrate outdated tests to the most recent form")
+    parser_migrate.set_defaults(func=migrate_tests)
+
+    args = parser.parse_args()
+    if hasattr(args, "func"):
+        args.func(args)
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/steps.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/steps.py
new file mode 100644
index 00000000..0f3899ad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/steps.py
@@ -0,0 +1,207 @@
+"""Step decorators.
+
+Example:
+
+@given("I have an article", target_fixture="article")
+def _(author):
+    return create_test_article(author=author)
+
+
+@when("I go to the article page")
+def _(browser, article):
+    browser.visit(urljoin(browser.url, "/articles/{0}/".format(article.id)))
+
+
+@then("I should not see the error message")
+def _(browser):
+    with pytest.raises(ElementDoesNotExist):
+        browser.find_by_css(".message.error").first
+
+
+Multiple names for the steps:
+
+@given("I have an article")
+@given("there is an article")
+def _(author):
+    return create_test_article(author=author)
+
+
+Reusing existing fixtures for a different step name:
+
+
+@given("I have a beautiful article")
+def _(article):
+    pass
+
+"""
+
+from __future__ import annotations
+
+import enum
+from collections.abc import Iterable
+from dataclasses import dataclass, field
+from itertools import count
+from typing import Any, Callable, Literal, TypeVar
+
+import pytest
+from typing_extensions import ParamSpec
+
+from .parser import Step
+from .parsers import StepParser, get_parser
+from .types import GIVEN, THEN, WHEN
+from .utils import get_caller_module_locals
+
+P = ParamSpec("P")
+T = TypeVar("T")
+
+
+@enum.unique
+class StepNamePrefix(enum.Enum):
+    step_def = "pytestbdd_stepdef"
+    step_impl = "pytestbdd_stepimpl"
+
+
+@dataclass
+class StepFunctionContext:
+    type: Literal["given", "when", "then"] | None
+    step_func: Callable[..., Any]
+    parser: StepParser
+    converters: dict[str, Callable[[str], Any]] = field(default_factory=dict)
+    target_fixture: str | None = None
+
+
+def get_step_fixture_name(step: Step) -> str:
+    """Get step fixture name"""
+    return f"{StepNamePrefix.step_impl.value}_{step.type}_{step.name}"
+
+
+def given(
+    name: str | StepParser,
+    converters: dict[str, Callable[[str], Any]] | None = None,
+    target_fixture: str | None = None,
+    stacklevel: int = 1,
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """Given step decorator.
+
+    :param name: Step name or a parser object.
+    :param converters: Optional `dict` of the argument or parameter converters in form
+                       {<param_name>: <converter function>}.
+    :param target_fixture: Target fixture name to replace by steps definition function.
+    :param stacklevel: Stack level to find the caller frame. This is used when injecting the step definition fixture.
+
+    :return: Decorator function for the step.
+    """
+    return step(name, GIVEN, converters=converters, target_fixture=target_fixture, stacklevel=stacklevel)
+
+
+def when(
+    name: str | StepParser,
+    converters: dict[str, Callable[[str], Any]] | None = None,
+    target_fixture: str | None = None,
+    stacklevel: int = 1,
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """When step decorator.
+
+    :param name: Step name or a parser object.
+    :param converters: Optional `dict` of the argument or parameter converters in form
+                       {<param_name>: <converter function>}.
+    :param target_fixture: Target fixture name to replace by steps definition function.
+    :param stacklevel: Stack level to find the caller frame. This is used when injecting the step definition fixture.
+
+    :return: Decorator function for the step.
+    """
+    return step(name, WHEN, converters=converters, target_fixture=target_fixture, stacklevel=stacklevel)
+
+
+def then(
+    name: str | StepParser,
+    converters: dict[str, Callable[[str], Any]] | None = None,
+    target_fixture: str | None = None,
+    stacklevel: int = 1,
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """Then step decorator.
+
+    :param name: Step name or a parser object.
+    :param converters: Optional `dict` of the argument or parameter converters in form
+                       {<param_name>: <converter function>}.
+    :param target_fixture: Target fixture name to replace by steps definition function.
+    :param stacklevel: Stack level to find the caller frame. This is used when injecting the step definition fixture.
+
+    :return: Decorator function for the step.
+    """
+    return step(name, THEN, converters=converters, target_fixture=target_fixture, stacklevel=stacklevel)
+
+
+def step(
+    name: str | StepParser,
+    type_: Literal["given", "when", "then"] | None = None,
+    converters: dict[str, Callable[[str], Any]] | None = None,
+    target_fixture: str | None = None,
+    stacklevel: int = 1,
+) -> Callable[[Callable[P, T]], Callable[P, T]]:
+    """Generic step decorator.
+
+    :param name: Step name as in the feature file.
+    :param type_: Step type ("given", "when" or "then"). If None, this step will work for all the types.
+    :param converters: Optional step arguments converters mapping.
+    :param target_fixture: Optional fixture name to replace by step definition.
+    :param stacklevel: Stack level to find the caller frame. This is used when injecting the step definition fixture.
+
+    :return: Decorator function for the step.
+
+    Example:
+    >>> @step("there is a wallet", target_fixture="wallet")
+    >>> def _() -> dict[str, int]:
+    >>>     return {"eur": 0, "usd": 0}
+
+    """
+    if converters is None:
+        converters = {}
+
+    def decorator(func: Callable[P, T]) -> Callable[P, T]:
+        parser = get_parser(name)
+
+        context = StepFunctionContext(
+            type=type_,
+            step_func=func,
+            parser=parser,
+            converters=converters,
+            target_fixture=target_fixture,
+        )
+
+        def step_function_marker() -> StepFunctionContext:
+            return context
+
+        step_function_marker._pytest_bdd_step_context = context  # type: ignore
+
+        caller_locals = get_caller_module_locals(stacklevel=stacklevel)
+        fixture_step_name = find_unique_name(
+            f"{StepNamePrefix.step_def.value}_{type_ or '*'}_{parser.name}", seen=caller_locals.keys()
+        )
+        caller_locals[fixture_step_name] = pytest.fixture(name=fixture_step_name)(step_function_marker)
+        return func
+
+    return decorator
+
+
+def find_unique_name(name: str, seen: Iterable[str]) -> str:
+    """Find a unique name among a set of strings.
+
+    New names are generated by appending an increasing number at the end of the name.
+
+    Example:
+    >>> find_unique_name("foo", ["foo", "foo_1"])
+    'foo_2'
+    """
+    seen = set(seen)
+    if name not in seen:
+        return name
+
+    # Generate new names with increasing numbers
+    for i in count(1):
+        new_name = f"{name}_{i}"
+        if new_name not in seen:
+            return new_name
+
+    # This line will never be reached, but it's here to satisfy mypy
+    raise RuntimeError("Unable to find a unique name")
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/templates/test.py.mak b/backend/venv/lib/python3.12/site-packages/pytest_bdd/templates/test.py.mak
new file mode 100644
index 00000000..9f790153
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/templates/test.py.mak
@@ -0,0 +1,29 @@
+% if features:
+"""${ features[0].name or features[0].rel_filename } feature tests."""
+
+from pytest_bdd import (
+    given,
+    scenario,
+    then,
+    when,
+)
+
+
+% endif
+% for scenario in sorted(scenarios, key=lambda scenario: scenario.name):
+@scenario('${scenario.feature.rel_filename}', ${ make_string_literal(scenario.name)})
+def test_${ make_python_name(scenario.name)}():
+    ${make_python_docstring(scenario.name)}
+
+
+% endfor
+% for step in steps:
+@${step.type}(${ make_string_literal(step.name)})
+def _():
+    ${make_python_docstring(step.name)}
+    raise NotImplementedError
+% if not loop.last:
+
+
+% endif
+% endfor
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/types.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/types.py
new file mode 100644
index 00000000..58cf4965
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/types.py
@@ -0,0 +1,20 @@
+"""Common type definitions."""
+
+from __future__ import annotations
+
+import typing
+
+if typing.TYPE_CHECKING:
+    from typing_extensions import Literal
+
+GIVEN: Literal["given"] = "given"
+WHEN: Literal["when"] = "when"
+THEN: Literal["then"] = "then"
+
+STEP_TYPES = (GIVEN, WHEN, THEN)
+
+STEP_TYPE_BY_PARSER_KEYWORD = {
+    "Context": GIVEN,
+    "Action": WHEN,
+    "Outcome": THEN,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/pytest_bdd/utils.py b/backend/venv/lib/python3.12/site-packages/pytest_bdd/utils.py
new file mode 100644
index 00000000..a72c86df
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/pytest_bdd/utils.py
@@ -0,0 +1,89 @@
+"""Various utility functions."""
+
+from __future__ import annotations
+
+import base64
+import pickle
+import re
+from inspect import getframeinfo, signature
+from sys import _getframe
+from typing import TYPE_CHECKING, TypeVar, cast
+
+if TYPE_CHECKING:
+    from typing import Any, Callable
+
+    from _pytest.config import Config
+    from _pytest.pytester import RunResult
+
+T = TypeVar("T")
+
+CONFIG_STACK: list[Config] = []
+
+
+def get_required_args(func: Callable[..., Any]) -> list[str]:
+    """Get a list of argument that are required for a function.
+
+    :param func: The function to inspect.
+
+    :return: A list of argument names.
+    """
+    params = signature(func).parameters.values()
+    return [
+        param.name for param in params if param.kind == param.POSITIONAL_OR_KEYWORD and param.default is param.empty
+    ]
+
+
+def get_caller_module_locals(stacklevel: int = 1) -> dict[str, Any]:
+    """Get the caller module locals dictionary.
+
+    We use sys._getframe instead of inspect.stack(0) because the latter is way slower, since it iterates over
+    all the frames in the stack.
+    """
+    return _getframe(stacklevel + 1).f_locals
+
+
+def get_caller_module_path(depth: int = 2) -> str:
+    """Get the caller module path.
+
+    We use sys._getframe instead of inspect.stack(0) because the latter is way slower, since it iterates over
+    all the frames in the stack.
+    """
+    frame = _getframe(depth)
+    return getframeinfo(frame, context=0).filename
+
+
+_DUMP_START = "_pytest_bdd_>>>"
+_DUMP_END = "<<<_pytest_bdd_"
+
+
+def dump_obj(*objects: Any) -> None:
+    """Dump objects to stdout so that they can be inspected by the test suite."""
+    for obj in objects:
+        dump = pickle.dumps(obj, protocol=pickle.HIGHEST_PROTOCOL)
+        encoded = base64.b64encode(dump).decode("ascii")
+        print(f"{_DUMP_START}{encoded}{_DUMP_END}")
+
+
+def collect_dumped_objects(result: RunResult) -> list:
+    """Parse all the objects dumped with `dump_object` from the result.
+
+    Note: You must run the result with output to stdout enabled.
+    For example, using ``pytester.runpytest("-s")``.
+    """
+    stdout = str(result.stdout)
+    payloads = re.findall(rf"{_DUMP_START}(.*?){_DUMP_END}", stdout)
+    return [pickle.loads(base64.b64decode(payload)) for payload in payloads]
+
+
+def setdefault(obj: object, name: str, default: T) -> T:
+    """Just like dict.setdefault, but for objects."""
+    try:
+        return cast(T, getattr(obj, name))
+    except AttributeError:
+        setattr(obj, name, default)
+        return default
+
+
+def identity(x: T) -> T:
+    """Return the argument."""
+    return x
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/LICENSE
new file mode 100644
index 00000000..67db8588
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/LICENSE
@@ -0,0 +1,175 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/METADATA
new file mode 100644
index 00000000..72d9dc53
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/METADATA
@@ -0,0 +1,119 @@
+Metadata-Version: 2.1
+Name: requests
+Version: 2.32.3
+Summary: Python HTTP for Humans.
+Home-page: https://requests.readthedocs.io
+Author: Kenneth Reitz
+Author-email: me@kennethreitz.org
+License: Apache-2.0
+Project-URL: Documentation, https://requests.readthedocs.io
+Project-URL: Source, https://github.com/psf/requests
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Natural Language :: English
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Internet :: WWW/HTTP
+Classifier: Topic :: Software Development :: Libraries
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: charset-normalizer <4,>=2
+Requires-Dist: idna <4,>=2.5
+Requires-Dist: urllib3 <3,>=1.21.1
+Requires-Dist: certifi >=2017.4.17
+Provides-Extra: security
+Provides-Extra: socks
+Requires-Dist: PySocks !=1.5.7,>=1.5.6 ; extra == 'socks'
+Provides-Extra: use_chardet_on_py3
+Requires-Dist: chardet <6,>=3.0.2 ; extra == 'use_chardet_on_py3'
+
+# Requests
+
+**Requests** is a simple, yet elegant, HTTP library.
+
+```python
+>>> import requests
+>>> r = requests.get('https://httpbin.org/basic-auth/user/pass', auth=('user', 'pass'))
+>>> r.status_code
+200
+>>> r.headers['content-type']
+'application/json; charset=utf8'
+>>> r.encoding
+'utf-8'
+>>> r.text
+'{"authenticated": true, ...'
+>>> r.json()
+{'authenticated': True, ...}
+```
+
+Requests allows you to send HTTP/1.1 requests extremely easily. There’s no need to manually add query strings to your URLs, or to form-encode your `PUT` & `POST` data — but nowadays, just use the `json` method!
+
+Requests is one of the most downloaded Python packages today, pulling in around `30M downloads / week`— according to GitHub, Requests is currently [depended upon](https://github.com/psf/requests/network/dependents?package_id=UGFja2FnZS01NzA4OTExNg%3D%3D) by `1,000,000+` repositories. You may certainly put your trust in this code.
+
+[![Downloads](https://static.pepy.tech/badge/requests/month)](https://pepy.tech/project/requests)
+[![Supported Versions](https://img.shields.io/pypi/pyversions/requests.svg)](https://pypi.org/project/requests)
+[![Contributors](https://img.shields.io/github/contributors/psf/requests.svg)](https://github.com/psf/requests/graphs/contributors)
+
+## Installing Requests and Supported Versions
+
+Requests is available on PyPI:
+
+```console
+$ python -m pip install requests
+```
+
+Requests officially supports Python 3.8+.
+
+## Supported Features & Best–Practices
+
+Requests is ready for the demands of building robust and reliable HTTP–speaking applications, for the needs of today.
+
+- Keep-Alive & Connection Pooling
+- International Domains and URLs
+- Sessions with Cookie Persistence
+- Browser-style TLS/SSL Verification
+- Basic & Digest Authentication
+- Familiar `dict`–like Cookies
+- Automatic Content Decompression and Decoding
+- Multi-part File Uploads
+- SOCKS Proxy Support
+- Connection Timeouts
+- Streaming Downloads
+- Automatic honoring of `.netrc`
+- Chunked HTTP Requests
+
+## API Reference and User Guide available on [Read the Docs](https://requests.readthedocs.io)
+
+[![Read the Docs](https://raw.githubusercontent.com/psf/requests/main/ext/ss.png)](https://requests.readthedocs.io)
+
+## Cloning the repository
+
+When cloning the Requests repository, you may need to add the `-c
+fetch.fsck.badTimezone=ignore` flag to avoid an error about a bad commit (see
+[this issue](https://github.com/psf/requests/issues/2690) for more background):
+
+```shell
+git clone -c fetch.fsck.badTimezone=ignore https://github.com/psf/requests.git
+```
+
+You can also apply this setting to your global Git config:
+
+```shell
+git config --global fetch.fsck.badTimezone ignore
+```
+
+---
+
+[![Kenneth Reitz](https://raw.githubusercontent.com/psf/requests/main/ext/kr.png)](https://kennethreitz.org) [![Python Software Foundation](https://raw.githubusercontent.com/psf/requests/main/ext/psf.png)](https://www.python.org/psf)
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/RECORD
new file mode 100644
index 00000000..f34e5a1d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/RECORD
@@ -0,0 +1,43 @@
+requests-2.32.3.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+requests-2.32.3.dist-info/LICENSE,sha256=CeipvOyAZxBGUsFoaFqwkx54aPnIKEtm9a5u2uXxEws,10142
+requests-2.32.3.dist-info/METADATA,sha256=ZY7oRUweLnb7jCEnEW9hFWs7IpQbNVnAA4ncpwA4WBo,4610
+requests-2.32.3.dist-info/RECORD,,
+requests-2.32.3.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+requests-2.32.3.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+requests-2.32.3.dist-info/top_level.txt,sha256=fMSVmHfb5rbGOo6xv-O_tUX6j-WyixssE-SnwcDRxNQ,9
+requests/__init__.py,sha256=4xaAERmPDIBPsa2PsjpU9r06yooK-2mZKHTZAhWRWts,5072
+requests/__pycache__/__init__.cpython-312.pyc,,
+requests/__pycache__/__version__.cpython-312.pyc,,
+requests/__pycache__/_internal_utils.cpython-312.pyc,,
+requests/__pycache__/adapters.cpython-312.pyc,,
+requests/__pycache__/api.cpython-312.pyc,,
+requests/__pycache__/auth.cpython-312.pyc,,
+requests/__pycache__/certs.cpython-312.pyc,,
+requests/__pycache__/compat.cpython-312.pyc,,
+requests/__pycache__/cookies.cpython-312.pyc,,
+requests/__pycache__/exceptions.cpython-312.pyc,,
+requests/__pycache__/help.cpython-312.pyc,,
+requests/__pycache__/hooks.cpython-312.pyc,,
+requests/__pycache__/models.cpython-312.pyc,,
+requests/__pycache__/packages.cpython-312.pyc,,
+requests/__pycache__/sessions.cpython-312.pyc,,
+requests/__pycache__/status_codes.cpython-312.pyc,,
+requests/__pycache__/structures.cpython-312.pyc,,
+requests/__pycache__/utils.cpython-312.pyc,,
+requests/__version__.py,sha256=FVfglgZmNQnmYPXpOohDU58F5EUb_-VnSTaAesS187g,435
+requests/_internal_utils.py,sha256=nMQymr4hs32TqVo5AbCrmcJEhvPUh7xXlluyqwslLiQ,1495
+requests/adapters.py,sha256=KIcecscqam6reOCXRl4DwP4jX8Jcl8sd57ft17KR2cQ,27451
+requests/api.py,sha256=_Zb9Oa7tzVIizTKwFrPjDEY9ejtm_OnSRERnADxGsQs,6449
+requests/auth.py,sha256=kF75tqnLctZ9Mf_hm9TZIj4cQWnN5uxRz8oWsx5wmR0,10186
+requests/certs.py,sha256=Z9Sb410Anv6jUFTyss0jFFhU6xst8ctELqfy8Ev23gw,429
+requests/compat.py,sha256=C5w_DPLSurXPgcdWU78fora0APmbYkX2G89QvH5xzPA,1817
+requests/cookies.py,sha256=bNi-iqEj4NPZ00-ob-rHvzkvObzN3lEpgw3g6paS3Xw,18590
+requests/exceptions.py,sha256=jJPS1UWATs86ShVUaLorTiJb1SaGuoNEWgICJep-VkY,4260
+requests/help.py,sha256=gPX5d_H7Xd88aDABejhqGgl9B1VFRTt5BmiYvL3PzIQ,3875
+requests/hooks.py,sha256=CiuysiHA39V5UfcCBXFIx83IrDpuwfN9RcTUgv28ftQ,733
+requests/models.py,sha256=k42roXzC8u_OagAPQi9U4MkfO7i4r2FdaqvMqstPehc,35418
+requests/packages.py,sha256=_g0gZ681UyAlKHRjH6kanbaoxx2eAb6qzcXiODyTIoc,904
+requests/sessions.py,sha256=ykTI8UWGSltOfH07HKollH7kTBGw4WhiBVaQGmckTw4,30495
+requests/status_codes.py,sha256=iJUAeA25baTdw-6PfD0eF4qhpINDJRJI-yaMqxs4LEI,4322
+requests/structures.py,sha256=-IbmhVz06S-5aPSZuUthZ6-6D9XOjRuTXHOabY041XM,2912
+requests/utils.py,sha256=HiQC6Nq_Da3ktaMiFzQkh-dCk3iQHHKEsYS5kDc-8Cw,33619
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/top_level.txt
new file mode 100644
index 00000000..f2293605
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests-2.32.3.dist-info/top_level.txt
@@ -0,0 +1 @@
+requests
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__init__.py b/backend/venv/lib/python3.12/site-packages/requests/__init__.py
new file mode 100644
index 00000000..051cda13
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/__init__.py
@@ -0,0 +1,184 @@
+#   __
+#  /__)  _  _     _   _ _/   _
+# / (   (- (/ (/ (- _)  /  _)
+#          /
+
+"""
+Requests HTTP Library
+~~~~~~~~~~~~~~~~~~~~~
+
+Requests is an HTTP library, written in Python, for human beings.
+Basic GET usage:
+
+   >>> import requests
+   >>> r = requests.get('https://www.python.org')
+   >>> r.status_code
+   200
+   >>> b'Python is a programming language' in r.content
+   True
+
+... or POST:
+
+   >>> payload = dict(key1='value1', key2='value2')
+   >>> r = requests.post('https://httpbin.org/post', data=payload)
+   >>> print(r.text)
+   {
+     ...
+     "form": {
+       "key1": "value1",
+       "key2": "value2"
+     },
+     ...
+   }
+
+The other HTTP methods are supported - see `requests.api`. Full documentation
+is at <https://requests.readthedocs.io>.
+
+:copyright: (c) 2017 by Kenneth Reitz.
+:license: Apache 2.0, see LICENSE for more details.
+"""
+
+import warnings
+
+import urllib3
+
+from .exceptions import RequestsDependencyWarning
+
+try:
+    from charset_normalizer import __version__ as charset_normalizer_version
+except ImportError:
+    charset_normalizer_version = None
+
+try:
+    from chardet import __version__ as chardet_version
+except ImportError:
+    chardet_version = None
+
+
+def check_compatibility(urllib3_version, chardet_version, charset_normalizer_version):
+    urllib3_version = urllib3_version.split(".")
+    assert urllib3_version != ["dev"]  # Verify urllib3 isn't installed from git.
+
+    # Sometimes, urllib3 only reports its version as 16.1.
+    if len(urllib3_version) == 2:
+        urllib3_version.append("0")
+
+    # Check urllib3 for compatibility.
+    major, minor, patch = urllib3_version  # noqa: F811
+    major, minor, patch = int(major), int(minor), int(patch)
+    # urllib3 >= 1.21.1
+    assert major >= 1
+    if major == 1:
+        assert minor >= 21
+
+    # Check charset_normalizer for compatibility.
+    if chardet_version:
+        major, minor, patch = chardet_version.split(".")[:3]
+        major, minor, patch = int(major), int(minor), int(patch)
+        # chardet_version >= 3.0.2, < 6.0.0
+        assert (3, 0, 2) <= (major, minor, patch) < (6, 0, 0)
+    elif charset_normalizer_version:
+        major, minor, patch = charset_normalizer_version.split(".")[:3]
+        major, minor, patch = int(major), int(minor), int(patch)
+        # charset_normalizer >= 2.0.0 < 4.0.0
+        assert (2, 0, 0) <= (major, minor, patch) < (4, 0, 0)
+    else:
+        warnings.warn(
+            "Unable to find acceptable character detection dependency "
+            "(chardet or charset_normalizer).",
+            RequestsDependencyWarning,
+        )
+
+
+def _check_cryptography(cryptography_version):
+    # cryptography < 1.3.4
+    try:
+        cryptography_version = list(map(int, cryptography_version.split(".")))
+    except ValueError:
+        return
+
+    if cryptography_version < [1, 3, 4]:
+        warning = "Old version of cryptography ({}) may cause slowdown.".format(
+            cryptography_version
+        )
+        warnings.warn(warning, RequestsDependencyWarning)
+
+
+# Check imported dependencies for compatibility.
+try:
+    check_compatibility(
+        urllib3.__version__, chardet_version, charset_normalizer_version
+    )
+except (AssertionError, ValueError):
+    warnings.warn(
+        "urllib3 ({}) or chardet ({})/charset_normalizer ({}) doesn't match a supported "
+        "version!".format(
+            urllib3.__version__, chardet_version, charset_normalizer_version
+        ),
+        RequestsDependencyWarning,
+    )
+
+# Attempt to enable urllib3's fallback for SNI support
+# if the standard library doesn't support SNI or the
+# 'ssl' library isn't available.
+try:
+    try:
+        import ssl
+    except ImportError:
+        ssl = None
+
+    if not getattr(ssl, "HAS_SNI", False):
+        from urllib3.contrib import pyopenssl
+
+        pyopenssl.inject_into_urllib3()
+
+        # Check cryptography version
+        from cryptography import __version__ as cryptography_version
+
+        _check_cryptography(cryptography_version)
+except ImportError:
+    pass
+
+# urllib3's DependencyWarnings should be silenced.
+from urllib3.exceptions import DependencyWarning
+
+warnings.simplefilter("ignore", DependencyWarning)
+
+# Set default logging handler to avoid "No handler found" warnings.
+import logging
+from logging import NullHandler
+
+from . import packages, utils
+from .__version__ import (
+    __author__,
+    __author_email__,
+    __build__,
+    __cake__,
+    __copyright__,
+    __description__,
+    __license__,
+    __title__,
+    __url__,
+    __version__,
+)
+from .api import delete, get, head, options, patch, post, put, request
+from .exceptions import (
+    ConnectionError,
+    ConnectTimeout,
+    FileModeWarning,
+    HTTPError,
+    JSONDecodeError,
+    ReadTimeout,
+    RequestException,
+    Timeout,
+    TooManyRedirects,
+    URLRequired,
+)
+from .models import PreparedRequest, Request, Response
+from .sessions import Session, session
+from .status_codes import codes
+
+logging.getLogger(__name__).addHandler(NullHandler())
+
+# FileModeWarnings go off per the default.
+warnings.simplefilter("default", FileModeWarning, append=True)
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..1c2a087b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__version__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__version__.cpython-312.pyc
new file mode 100644
index 00000000..34438a6f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/__version__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/_internal_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/_internal_utils.cpython-312.pyc
new file mode 100644
index 00000000..91af81b5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/_internal_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/adapters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/adapters.cpython-312.pyc
new file mode 100644
index 00000000..607f8b91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/adapters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..3d5a6fcb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/auth.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/auth.cpython-312.pyc
new file mode 100644
index 00000000..a5a113fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/auth.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/certs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/certs.cpython-312.pyc
new file mode 100644
index 00000000..0f5fa191
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/certs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..03af2f6c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/cookies.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/cookies.cpython-312.pyc
new file mode 100644
index 00000000..1d10f2e0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/cookies.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..8d3d52d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/help.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/help.cpython-312.pyc
new file mode 100644
index 00000000..907b11c8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/help.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/hooks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/hooks.cpython-312.pyc
new file mode 100644
index 00000000..738534a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/hooks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/models.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/models.cpython-312.pyc
new file mode 100644
index 00000000..0ed55fa8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/models.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/packages.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/packages.cpython-312.pyc
new file mode 100644
index 00000000..f20bdcb3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/packages.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/sessions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/sessions.cpython-312.pyc
new file mode 100644
index 00000000..fb243c79
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/sessions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/status_codes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/status_codes.cpython-312.pyc
new file mode 100644
index 00000000..c7f5c641
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/status_codes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/structures.cpython-312.pyc
new file mode 100644
index 00000000..00f65edd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..5c130dde
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/requests/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/requests/__version__.py b/backend/venv/lib/python3.12/site-packages/requests/__version__.py
new file mode 100644
index 00000000..2c105aca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/__version__.py
@@ -0,0 +1,14 @@
+# .-. .-. .-. . . .-. .-. .-. .-.
+# |(  |-  |.| | | |-  `-.  |  `-.
+# ' ' `-' `-`.`-' `-' `-'  '  `-'
+
+__title__ = "requests"
+__description__ = "Python HTTP for Humans."
+__url__ = "https://requests.readthedocs.io"
+__version__ = "2.32.3"
+__build__ = 0x023203
+__author__ = "Kenneth Reitz"
+__author_email__ = "me@kennethreitz.org"
+__license__ = "Apache-2.0"
+__copyright__ = "Copyright Kenneth Reitz"
+__cake__ = "\u2728 \U0001f370 \u2728"
diff --git a/backend/venv/lib/python3.12/site-packages/requests/_internal_utils.py b/backend/venv/lib/python3.12/site-packages/requests/_internal_utils.py
new file mode 100644
index 00000000..f2cf635e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/_internal_utils.py
@@ -0,0 +1,50 @@
+"""
+requests._internal_utils
+~~~~~~~~~~~~~~
+
+Provides utility functions that are consumed internally by Requests
+which depend on extremely few external helpers (such as compat)
+"""
+import re
+
+from .compat import builtin_str
+
+_VALID_HEADER_NAME_RE_BYTE = re.compile(rb"^[^:\s][^:\r\n]*$")
+_VALID_HEADER_NAME_RE_STR = re.compile(r"^[^:\s][^:\r\n]*$")
+_VALID_HEADER_VALUE_RE_BYTE = re.compile(rb"^\S[^\r\n]*$|^$")
+_VALID_HEADER_VALUE_RE_STR = re.compile(r"^\S[^\r\n]*$|^$")
+
+_HEADER_VALIDATORS_STR = (_VALID_HEADER_NAME_RE_STR, _VALID_HEADER_VALUE_RE_STR)
+_HEADER_VALIDATORS_BYTE = (_VALID_HEADER_NAME_RE_BYTE, _VALID_HEADER_VALUE_RE_BYTE)
+HEADER_VALIDATORS = {
+    bytes: _HEADER_VALIDATORS_BYTE,
+    str: _HEADER_VALIDATORS_STR,
+}
+
+
+def to_native_string(string, encoding="ascii"):
+    """Given a string object, regardless of type, returns a representation of
+    that string in the native string type, encoding and decoding where
+    necessary. This assumes ASCII unless told otherwise.
+    """
+    if isinstance(string, builtin_str):
+        out = string
+    else:
+        out = string.decode(encoding)
+
+    return out
+
+
+def unicode_is_ascii(u_string):
+    """Determine if unicode string only contains ASCII characters.
+
+    :param str u_string: unicode string to check. Must be unicode
+        and not Python 2 `str`.
+    :rtype: bool
+    """
+    assert isinstance(u_string, str)
+    try:
+        u_string.encode("ascii")
+        return True
+    except UnicodeEncodeError:
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/requests/adapters.py b/backend/venv/lib/python3.12/site-packages/requests/adapters.py
new file mode 100644
index 00000000..9a58b160
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/adapters.py
@@ -0,0 +1,719 @@
+"""
+requests.adapters
+~~~~~~~~~~~~~~~~~
+
+This module contains the transport adapters that Requests uses to define
+and maintain connections.
+"""
+
+import os.path
+import socket  # noqa: F401
+import typing
+import warnings
+
+from urllib3.exceptions import ClosedPoolError, ConnectTimeoutError
+from urllib3.exceptions import HTTPError as _HTTPError
+from urllib3.exceptions import InvalidHeader as _InvalidHeader
+from urllib3.exceptions import (
+    LocationValueError,
+    MaxRetryError,
+    NewConnectionError,
+    ProtocolError,
+)
+from urllib3.exceptions import ProxyError as _ProxyError
+from urllib3.exceptions import ReadTimeoutError, ResponseError
+from urllib3.exceptions import SSLError as _SSLError
+from urllib3.poolmanager import PoolManager, proxy_from_url
+from urllib3.util import Timeout as TimeoutSauce
+from urllib3.util import parse_url
+from urllib3.util.retry import Retry
+from urllib3.util.ssl_ import create_urllib3_context
+
+from .auth import _basic_auth_str
+from .compat import basestring, urlparse
+from .cookies import extract_cookies_to_jar
+from .exceptions import (
+    ConnectionError,
+    ConnectTimeout,
+    InvalidHeader,
+    InvalidProxyURL,
+    InvalidSchema,
+    InvalidURL,
+    ProxyError,
+    ReadTimeout,
+    RetryError,
+    SSLError,
+)
+from .models import Response
+from .structures import CaseInsensitiveDict
+from .utils import (
+    DEFAULT_CA_BUNDLE_PATH,
+    extract_zipped_paths,
+    get_auth_from_url,
+    get_encoding_from_headers,
+    prepend_scheme_if_needed,
+    select_proxy,
+    urldefragauth,
+)
+
+try:
+    from urllib3.contrib.socks import SOCKSProxyManager
+except ImportError:
+
+    def SOCKSProxyManager(*args, **kwargs):
+        raise InvalidSchema("Missing dependencies for SOCKS support.")
+
+
+if typing.TYPE_CHECKING:
+    from .models import PreparedRequest
+
+
+DEFAULT_POOLBLOCK = False
+DEFAULT_POOLSIZE = 10
+DEFAULT_RETRIES = 0
+DEFAULT_POOL_TIMEOUT = None
+
+
+try:
+    import ssl  # noqa: F401
+
+    _preloaded_ssl_context = create_urllib3_context()
+    _preloaded_ssl_context.load_verify_locations(
+        extract_zipped_paths(DEFAULT_CA_BUNDLE_PATH)
+    )
+except ImportError:
+    # Bypass default SSLContext creation when Python
+    # interpreter isn't built with the ssl module.
+    _preloaded_ssl_context = None
+
+
+def _urllib3_request_context(
+    request: "PreparedRequest",
+    verify: "bool | str | None",
+    client_cert: "typing.Tuple[str, str] | str | None",
+    poolmanager: "PoolManager",
+) -> "(typing.Dict[str, typing.Any], typing.Dict[str, typing.Any])":
+    host_params = {}
+    pool_kwargs = {}
+    parsed_request_url = urlparse(request.url)
+    scheme = parsed_request_url.scheme.lower()
+    port = parsed_request_url.port
+
+    # Determine if we have and should use our default SSLContext
+    # to optimize performance on standard requests.
+    poolmanager_kwargs = getattr(poolmanager, "connection_pool_kw", {})
+    has_poolmanager_ssl_context = poolmanager_kwargs.get("ssl_context")
+    should_use_default_ssl_context = (
+        _preloaded_ssl_context is not None and not has_poolmanager_ssl_context
+    )
+
+    cert_reqs = "CERT_REQUIRED"
+    if verify is False:
+        cert_reqs = "CERT_NONE"
+    elif verify is True and should_use_default_ssl_context:
+        pool_kwargs["ssl_context"] = _preloaded_ssl_context
+    elif isinstance(verify, str):
+        if not os.path.isdir(verify):
+            pool_kwargs["ca_certs"] = verify
+        else:
+            pool_kwargs["ca_cert_dir"] = verify
+    pool_kwargs["cert_reqs"] = cert_reqs
+    if client_cert is not None:
+        if isinstance(client_cert, tuple) and len(client_cert) == 2:
+            pool_kwargs["cert_file"] = client_cert[0]
+            pool_kwargs["key_file"] = client_cert[1]
+        else:
+            # According to our docs, we allow users to specify just the client
+            # cert path
+            pool_kwargs["cert_file"] = client_cert
+    host_params = {
+        "scheme": scheme,
+        "host": parsed_request_url.hostname,
+        "port": port,
+    }
+    return host_params, pool_kwargs
+
+
+class BaseAdapter:
+    """The Base Transport Adapter"""
+
+    def __init__(self):
+        super().__init__()
+
+    def send(
+        self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None
+    ):
+        """Sends PreparedRequest object. Returns Response object.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param stream: (optional) Whether to stream the request content.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple
+        :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use
+        :param cert: (optional) Any user-provided SSL certificate to be trusted.
+        :param proxies: (optional) The proxies dictionary to apply to the request.
+        """
+        raise NotImplementedError
+
+    def close(self):
+        """Cleans up adapter specific items."""
+        raise NotImplementedError
+
+
+class HTTPAdapter(BaseAdapter):
+    """The built-in HTTP Adapter for urllib3.
+
+    Provides a general-case interface for Requests sessions to contact HTTP and
+    HTTPS urls by implementing the Transport Adapter interface. This class will
+    usually be created by the :class:`Session <Session>` class under the
+    covers.
+
+    :param pool_connections: The number of urllib3 connection pools to cache.
+    :param pool_maxsize: The maximum number of connections to save in the pool.
+    :param max_retries: The maximum number of retries each connection
+        should attempt. Note, this applies only to failed DNS lookups, socket
+        connections and connection timeouts, never to requests where data has
+        made it to the server. By default, Requests does not retry failed
+        connections. If you need granular control over the conditions under
+        which we retry a request, import urllib3's ``Retry`` class and pass
+        that instead.
+    :param pool_block: Whether the connection pool should block for connections.
+
+    Usage::
+
+      >>> import requests
+      >>> s = requests.Session()
+      >>> a = requests.adapters.HTTPAdapter(max_retries=3)
+      >>> s.mount('http://', a)
+    """
+
+    __attrs__ = [
+        "max_retries",
+        "config",
+        "_pool_connections",
+        "_pool_maxsize",
+        "_pool_block",
+    ]
+
+    def __init__(
+        self,
+        pool_connections=DEFAULT_POOLSIZE,
+        pool_maxsize=DEFAULT_POOLSIZE,
+        max_retries=DEFAULT_RETRIES,
+        pool_block=DEFAULT_POOLBLOCK,
+    ):
+        if max_retries == DEFAULT_RETRIES:
+            self.max_retries = Retry(0, read=False)
+        else:
+            self.max_retries = Retry.from_int(max_retries)
+        self.config = {}
+        self.proxy_manager = {}
+
+        super().__init__()
+
+        self._pool_connections = pool_connections
+        self._pool_maxsize = pool_maxsize
+        self._pool_block = pool_block
+
+        self.init_poolmanager(pool_connections, pool_maxsize, block=pool_block)
+
+    def __getstate__(self):
+        return {attr: getattr(self, attr, None) for attr in self.__attrs__}
+
+    def __setstate__(self, state):
+        # Can't handle by adding 'proxy_manager' to self.__attrs__ because
+        # self.poolmanager uses a lambda function, which isn't pickleable.
+        self.proxy_manager = {}
+        self.config = {}
+
+        for attr, value in state.items():
+            setattr(self, attr, value)
+
+        self.init_poolmanager(
+            self._pool_connections, self._pool_maxsize, block=self._pool_block
+        )
+
+    def init_poolmanager(
+        self, connections, maxsize, block=DEFAULT_POOLBLOCK, **pool_kwargs
+    ):
+        """Initializes a urllib3 PoolManager.
+
+        This method should not be called from user code, and is only
+        exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param connections: The number of urllib3 connection pools to cache.
+        :param maxsize: The maximum number of connections to save in the pool.
+        :param block: Block when no free connections are available.
+        :param pool_kwargs: Extra keyword arguments used to initialize the Pool Manager.
+        """
+        # save these values for pickling
+        self._pool_connections = connections
+        self._pool_maxsize = maxsize
+        self._pool_block = block
+
+        self.poolmanager = PoolManager(
+            num_pools=connections,
+            maxsize=maxsize,
+            block=block,
+            **pool_kwargs,
+        )
+
+    def proxy_manager_for(self, proxy, **proxy_kwargs):
+        """Return urllib3 ProxyManager for the given proxy.
+
+        This method should not be called from user code, and is only
+        exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param proxy: The proxy to return a urllib3 ProxyManager for.
+        :param proxy_kwargs: Extra keyword arguments used to configure the Proxy Manager.
+        :returns: ProxyManager
+        :rtype: urllib3.ProxyManager
+        """
+        if proxy in self.proxy_manager:
+            manager = self.proxy_manager[proxy]
+        elif proxy.lower().startswith("socks"):
+            username, password = get_auth_from_url(proxy)
+            manager = self.proxy_manager[proxy] = SOCKSProxyManager(
+                proxy,
+                username=username,
+                password=password,
+                num_pools=self._pool_connections,
+                maxsize=self._pool_maxsize,
+                block=self._pool_block,
+                **proxy_kwargs,
+            )
+        else:
+            proxy_headers = self.proxy_headers(proxy)
+            manager = self.proxy_manager[proxy] = proxy_from_url(
+                proxy,
+                proxy_headers=proxy_headers,
+                num_pools=self._pool_connections,
+                maxsize=self._pool_maxsize,
+                block=self._pool_block,
+                **proxy_kwargs,
+            )
+
+        return manager
+
+    def cert_verify(self, conn, url, verify, cert):
+        """Verify a SSL certificate. This method should not be called from user
+        code, and is only exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param conn: The urllib3 connection object associated with the cert.
+        :param url: The requested URL.
+        :param verify: Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use
+        :param cert: The SSL certificate to verify.
+        """
+        if url.lower().startswith("https") and verify:
+            conn.cert_reqs = "CERT_REQUIRED"
+
+            # Only load the CA certificates if 'verify' is a string indicating the CA bundle to use.
+            # Otherwise, if verify is a boolean, we don't load anything since
+            # the connection will be using a context with the default certificates already loaded,
+            # and this avoids a call to the slow load_verify_locations()
+            if verify is not True:
+                # `verify` must be a str with a path then
+                cert_loc = verify
+
+                if not os.path.exists(cert_loc):
+                    raise OSError(
+                        f"Could not find a suitable TLS CA certificate bundle, "
+                        f"invalid path: {cert_loc}"
+                    )
+
+                if not os.path.isdir(cert_loc):
+                    conn.ca_certs = cert_loc
+                else:
+                    conn.ca_cert_dir = cert_loc
+        else:
+            conn.cert_reqs = "CERT_NONE"
+            conn.ca_certs = None
+            conn.ca_cert_dir = None
+
+        if cert:
+            if not isinstance(cert, basestring):
+                conn.cert_file = cert[0]
+                conn.key_file = cert[1]
+            else:
+                conn.cert_file = cert
+                conn.key_file = None
+            if conn.cert_file and not os.path.exists(conn.cert_file):
+                raise OSError(
+                    f"Could not find the TLS certificate file, "
+                    f"invalid path: {conn.cert_file}"
+                )
+            if conn.key_file and not os.path.exists(conn.key_file):
+                raise OSError(
+                    f"Could not find the TLS key file, invalid path: {conn.key_file}"
+                )
+
+    def build_response(self, req, resp):
+        """Builds a :class:`Response <requests.Response>` object from a urllib3
+        response. This should not be called from user code, and is only exposed
+        for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`
+
+        :param req: The :class:`PreparedRequest <PreparedRequest>` used to generate the response.
+        :param resp: The urllib3 response object.
+        :rtype: requests.Response
+        """
+        response = Response()
+
+        # Fallback to None if there's no status_code, for whatever reason.
+        response.status_code = getattr(resp, "status", None)
+
+        # Make headers case-insensitive.
+        response.headers = CaseInsensitiveDict(getattr(resp, "headers", {}))
+
+        # Set encoding.
+        response.encoding = get_encoding_from_headers(response.headers)
+        response.raw = resp
+        response.reason = response.raw.reason
+
+        if isinstance(req.url, bytes):
+            response.url = req.url.decode("utf-8")
+        else:
+            response.url = req.url
+
+        # Add new cookies from the server.
+        extract_cookies_to_jar(response.cookies, req, resp)
+
+        # Give the Response some context.
+        response.request = req
+        response.connection = self
+
+        return response
+
+    def build_connection_pool_key_attributes(self, request, verify, cert=None):
+        """Build the PoolKey attributes used by urllib3 to return a connection.
+
+        This looks at the PreparedRequest, the user-specified verify value,
+        and the value of the cert parameter to determine what PoolKey values
+        to use to select a connection from a given urllib3 Connection Pool.
+
+        The SSL related pool key arguments are not consistently set. As of
+        this writing, use the following to determine what keys may be in that
+        dictionary:
+
+        * If ``verify`` is ``True``, ``"ssl_context"`` will be set and will be the
+          default Requests SSL Context
+        * If ``verify`` is ``False``, ``"ssl_context"`` will not be set but
+          ``"cert_reqs"`` will be set
+        * If ``verify`` is a string, (i.e., it is a user-specified trust bundle)
+          ``"ca_certs"`` will be set if the string is not a directory recognized
+          by :py:func:`os.path.isdir`, otherwise ``"ca_certs_dir"`` will be
+          set.
+        * If ``"cert"`` is specified, ``"cert_file"`` will always be set. If
+          ``"cert"`` is a tuple with a second item, ``"key_file"`` will also
+          be present
+
+        To override these settings, one may subclass this class, call this
+        method and use the above logic to change parameters as desired. For
+        example, if one wishes to use a custom :py:class:`ssl.SSLContext` one
+        must both set ``"ssl_context"`` and based on what else they require,
+        alter the other keys to ensure the desired behaviour.
+
+        :param request:
+            The PreparedReqest being sent over the connection.
+        :type request:
+            :class:`~requests.models.PreparedRequest`
+        :param verify:
+            Either a boolean, in which case it controls whether
+            we verify the server's TLS certificate, or a string, in which case it
+            must be a path to a CA bundle to use.
+        :param cert:
+            (optional) Any user-provided SSL certificate for client
+            authentication (a.k.a., mTLS). This may be a string (i.e., just
+            the path to a file which holds both certificate and key) or a
+            tuple of length 2 with the certificate file path and key file
+            path.
+        :returns:
+            A tuple of two dictionaries. The first is the "host parameters"
+            portion of the Pool Key including scheme, hostname, and port. The
+            second is a dictionary of SSLContext related parameters.
+        """
+        return _urllib3_request_context(request, verify, cert, self.poolmanager)
+
+    def get_connection_with_tls_context(self, request, verify, proxies=None, cert=None):
+        """Returns a urllib3 connection for the given request and TLS settings.
+        This should not be called from user code, and is only exposed for use
+        when subclassing the :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request:
+            The :class:`PreparedRequest <PreparedRequest>` object to be sent
+            over the connection.
+        :param verify:
+            Either a boolean, in which case it controls whether we verify the
+            server's TLS certificate, or a string, in which case it must be a
+            path to a CA bundle to use.
+        :param proxies:
+            (optional) The proxies dictionary to apply to the request.
+        :param cert:
+            (optional) Any user-provided SSL certificate to be used for client
+            authentication (a.k.a., mTLS).
+        :rtype:
+            urllib3.ConnectionPool
+        """
+        proxy = select_proxy(request.url, proxies)
+        try:
+            host_params, pool_kwargs = self.build_connection_pool_key_attributes(
+                request,
+                verify,
+                cert,
+            )
+        except ValueError as e:
+            raise InvalidURL(e, request=request)
+        if proxy:
+            proxy = prepend_scheme_if_needed(proxy, "http")
+            proxy_url = parse_url(proxy)
+            if not proxy_url.host:
+                raise InvalidProxyURL(
+                    "Please check proxy URL. It is malformed "
+                    "and could be missing the host."
+                )
+            proxy_manager = self.proxy_manager_for(proxy)
+            conn = proxy_manager.connection_from_host(
+                **host_params, pool_kwargs=pool_kwargs
+            )
+        else:
+            # Only scheme should be lower case
+            conn = self.poolmanager.connection_from_host(
+                **host_params, pool_kwargs=pool_kwargs
+            )
+
+        return conn
+
+    def get_connection(self, url, proxies=None):
+        """DEPRECATED: Users should move to `get_connection_with_tls_context`
+        for all subclasses of HTTPAdapter using Requests>=2.32.2.
+
+        Returns a urllib3 connection for the given URL. This should not be
+        called from user code, and is only exposed for use when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param url: The URL to connect to.
+        :param proxies: (optional) A Requests-style dictionary of proxies used on this request.
+        :rtype: urllib3.ConnectionPool
+        """
+        warnings.warn(
+            (
+                "`get_connection` has been deprecated in favor of "
+                "`get_connection_with_tls_context`. Custom HTTPAdapter subclasses "
+                "will need to migrate for Requests>=2.32.2. Please see "
+                "https://github.com/psf/requests/pull/6710 for more details."
+            ),
+            DeprecationWarning,
+        )
+        proxy = select_proxy(url, proxies)
+
+        if proxy:
+            proxy = prepend_scheme_if_needed(proxy, "http")
+            proxy_url = parse_url(proxy)
+            if not proxy_url.host:
+                raise InvalidProxyURL(
+                    "Please check proxy URL. It is malformed "
+                    "and could be missing the host."
+                )
+            proxy_manager = self.proxy_manager_for(proxy)
+            conn = proxy_manager.connection_from_url(url)
+        else:
+            # Only scheme should be lower case
+            parsed = urlparse(url)
+            url = parsed.geturl()
+            conn = self.poolmanager.connection_from_url(url)
+
+        return conn
+
+    def close(self):
+        """Disposes of any internal state.
+
+        Currently, this closes the PoolManager and any active ProxyManager,
+        which closes any pooled connections.
+        """
+        self.poolmanager.clear()
+        for proxy in self.proxy_manager.values():
+            proxy.clear()
+
+    def request_url(self, request, proxies):
+        """Obtain the url to use when making the final request.
+
+        If the message is being sent through a HTTP proxy, the full URL has to
+        be used. Otherwise, we should only use the path portion of the URL.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs.
+        :rtype: str
+        """
+        proxy = select_proxy(request.url, proxies)
+        scheme = urlparse(request.url).scheme
+
+        is_proxied_http_request = proxy and scheme != "https"
+        using_socks_proxy = False
+        if proxy:
+            proxy_scheme = urlparse(proxy).scheme.lower()
+            using_socks_proxy = proxy_scheme.startswith("socks")
+
+        url = request.path_url
+        if url.startswith("//"):  # Don't confuse urllib3
+            url = f"/{url.lstrip('/')}"
+
+        if is_proxied_http_request and not using_socks_proxy:
+            url = urldefragauth(request.url)
+
+        return url
+
+    def add_headers(self, request, **kwargs):
+        """Add any headers needed by the connection. As of v2.0 this does
+        nothing by default, but is left for overriding by users that subclass
+        the :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` to add headers to.
+        :param kwargs: The keyword arguments from the call to send().
+        """
+        pass
+
+    def proxy_headers(self, proxy):
+        """Returns a dictionary of the headers to add to any request sent
+        through a proxy. This works with urllib3 magic to ensure that they are
+        correctly sent to the proxy, rather than in a tunnelled request if
+        CONNECT is being used.
+
+        This should not be called from user code, and is only exposed for use
+        when subclassing the
+        :class:`HTTPAdapter <requests.adapters.HTTPAdapter>`.
+
+        :param proxy: The url of the proxy being used for this request.
+        :rtype: dict
+        """
+        headers = {}
+        username, password = get_auth_from_url(proxy)
+
+        if username:
+            headers["Proxy-Authorization"] = _basic_auth_str(username, password)
+
+        return headers
+
+    def send(
+        self, request, stream=False, timeout=None, verify=True, cert=None, proxies=None
+    ):
+        """Sends PreparedRequest object. Returns Response object.
+
+        :param request: The :class:`PreparedRequest <PreparedRequest>` being sent.
+        :param stream: (optional) Whether to stream the request content.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple or urllib3 Timeout object
+        :param verify: (optional) Either a boolean, in which case it controls whether
+            we verify the server's TLS certificate, or a string, in which case it
+            must be a path to a CA bundle to use
+        :param cert: (optional) Any user-provided SSL certificate to be trusted.
+        :param proxies: (optional) The proxies dictionary to apply to the request.
+        :rtype: requests.Response
+        """
+
+        try:
+            conn = self.get_connection_with_tls_context(
+                request, verify, proxies=proxies, cert=cert
+            )
+        except LocationValueError as e:
+            raise InvalidURL(e, request=request)
+
+        self.cert_verify(conn, request.url, verify, cert)
+        url = self.request_url(request, proxies)
+        self.add_headers(
+            request,
+            stream=stream,
+            timeout=timeout,
+            verify=verify,
+            cert=cert,
+            proxies=proxies,
+        )
+
+        chunked = not (request.body is None or "Content-Length" in request.headers)
+
+        if isinstance(timeout, tuple):
+            try:
+                connect, read = timeout
+                timeout = TimeoutSauce(connect=connect, read=read)
+            except ValueError:
+                raise ValueError(
+                    f"Invalid timeout {timeout}. Pass a (connect, read) timeout tuple, "
+                    f"or a single float to set both timeouts to the same value."
+                )
+        elif isinstance(timeout, TimeoutSauce):
+            pass
+        else:
+            timeout = TimeoutSauce(connect=timeout, read=timeout)
+
+        try:
+            resp = conn.urlopen(
+                method=request.method,
+                url=url,
+                body=request.body,
+                headers=request.headers,
+                redirect=False,
+                assert_same_host=False,
+                preload_content=False,
+                decode_content=False,
+                retries=self.max_retries,
+                timeout=timeout,
+                chunked=chunked,
+            )
+
+        except (ProtocolError, OSError) as err:
+            raise ConnectionError(err, request=request)
+
+        except MaxRetryError as e:
+            if isinstance(e.reason, ConnectTimeoutError):
+                # TODO: Remove this in 3.0.0: see #2811
+                if not isinstance(e.reason, NewConnectionError):
+                    raise ConnectTimeout(e, request=request)
+
+            if isinstance(e.reason, ResponseError):
+                raise RetryError(e, request=request)
+
+            if isinstance(e.reason, _ProxyError):
+                raise ProxyError(e, request=request)
+
+            if isinstance(e.reason, _SSLError):
+                # This branch is for urllib3 v1.22 and later.
+                raise SSLError(e, request=request)
+
+            raise ConnectionError(e, request=request)
+
+        except ClosedPoolError as e:
+            raise ConnectionError(e, request=request)
+
+        except _ProxyError as e:
+            raise ProxyError(e)
+
+        except (_SSLError, _HTTPError) as e:
+            if isinstance(e, _SSLError):
+                # This branch is for urllib3 versions earlier than v1.22
+                raise SSLError(e, request=request)
+            elif isinstance(e, ReadTimeoutError):
+                raise ReadTimeout(e, request=request)
+            elif isinstance(e, _InvalidHeader):
+                raise InvalidHeader(e, request=request)
+            else:
+                raise
+
+        return self.build_response(request, resp)
diff --git a/backend/venv/lib/python3.12/site-packages/requests/api.py b/backend/venv/lib/python3.12/site-packages/requests/api.py
new file mode 100644
index 00000000..59607445
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/api.py
@@ -0,0 +1,157 @@
+"""
+requests.api
+~~~~~~~~~~~~
+
+This module implements the Requests API.
+
+:copyright: (c) 2012 by Kenneth Reitz.
+:license: Apache2, see LICENSE for more details.
+"""
+
+from . import sessions
+
+
+def request(method, url, **kwargs):
+    """Constructs and sends a :class:`Request <Request>`.
+
+    :param method: method for the new :class:`Request` object: ``GET``, ``OPTIONS``, ``HEAD``, ``POST``, ``PUT``, ``PATCH``, or ``DELETE``.
+    :param url: URL for the new :class:`Request` object.
+    :param params: (optional) Dictionary, list of tuples or bytes to send
+        in the query string for the :class:`Request`.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param headers: (optional) Dictionary of HTTP Headers to send with the :class:`Request`.
+    :param cookies: (optional) Dict or CookieJar object to send with the :class:`Request`.
+    :param files: (optional) Dictionary of ``'name': file-like-objects`` (or ``{'name': file-tuple}``) for multipart encoding upload.
+        ``file-tuple`` can be a 2-tuple ``('filename', fileobj)``, 3-tuple ``('filename', fileobj, 'content_type')``
+        or a 4-tuple ``('filename', fileobj, 'content_type', custom_headers)``, where ``'content_type'`` is a string
+        defining the content type of the given file and ``custom_headers`` a dict-like object containing additional headers
+        to add for the file.
+    :param auth: (optional) Auth tuple to enable Basic/Digest/Custom HTTP Auth.
+    :param timeout: (optional) How many seconds to wait for the server to send data
+        before giving up, as a float, or a :ref:`(connect timeout, read
+        timeout) <timeouts>` tuple.
+    :type timeout: float or tuple
+    :param allow_redirects: (optional) Boolean. Enable/disable GET/OPTIONS/POST/PUT/PATCH/DELETE/HEAD redirection. Defaults to ``True``.
+    :type allow_redirects: bool
+    :param proxies: (optional) Dictionary mapping protocol to the URL of the proxy.
+    :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use. Defaults to ``True``.
+    :param stream: (optional) if ``False``, the response content will be immediately downloaded.
+    :param cert: (optional) if String, path to ssl client cert file (.pem). If Tuple, ('cert', 'key') pair.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.request('GET', 'https://httpbin.org/get')
+      >>> req
+      <Response [200]>
+    """
+
+    # By using the 'with' statement we are sure the session is closed, thus we
+    # avoid leaving sockets open which can trigger a ResourceWarning in some
+    # cases, and look like a memory leak in others.
+    with sessions.Session() as session:
+        return session.request(method=method, url=url, **kwargs)
+
+
+def get(url, params=None, **kwargs):
+    r"""Sends a GET request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param params: (optional) Dictionary, list of tuples or bytes to send
+        in the query string for the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("get", url, params=params, **kwargs)
+
+
+def options(url, **kwargs):
+    r"""Sends an OPTIONS request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("options", url, **kwargs)
+
+
+def head(url, **kwargs):
+    r"""Sends a HEAD request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes. If
+        `allow_redirects` is not provided, it will be set to `False` (as
+        opposed to the default :meth:`request` behavior).
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    kwargs.setdefault("allow_redirects", False)
+    return request("head", url, **kwargs)
+
+
+def post(url, data=None, json=None, **kwargs):
+    r"""Sends a POST request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("post", url, data=data, json=json, **kwargs)
+
+
+def put(url, data=None, **kwargs):
+    r"""Sends a PUT request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("put", url, data=data, **kwargs)
+
+
+def patch(url, data=None, **kwargs):
+    r"""Sends a PATCH request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+        object to send in the body of the :class:`Request`.
+    :param json: (optional) A JSON serializable Python object to send in the body of the :class:`Request`.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("patch", url, data=data, **kwargs)
+
+
+def delete(url, **kwargs):
+    r"""Sends a DELETE request.
+
+    :param url: URL for the new :class:`Request` object.
+    :param \*\*kwargs: Optional arguments that ``request`` takes.
+    :return: :class:`Response <Response>` object
+    :rtype: requests.Response
+    """
+
+    return request("delete", url, **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/requests/auth.py b/backend/venv/lib/python3.12/site-packages/requests/auth.py
new file mode 100644
index 00000000..4a7ce6dc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/auth.py
@@ -0,0 +1,314 @@
+"""
+requests.auth
+~~~~~~~~~~~~~
+
+This module contains the authentication handlers for Requests.
+"""
+
+import hashlib
+import os
+import re
+import threading
+import time
+import warnings
+from base64 import b64encode
+
+from ._internal_utils import to_native_string
+from .compat import basestring, str, urlparse
+from .cookies import extract_cookies_to_jar
+from .utils import parse_dict_header
+
+CONTENT_TYPE_FORM_URLENCODED = "application/x-www-form-urlencoded"
+CONTENT_TYPE_MULTI_PART = "multipart/form-data"
+
+
+def _basic_auth_str(username, password):
+    """Returns a Basic Auth string."""
+
+    # "I want us to put a big-ol' comment on top of it that
+    # says that this behaviour is dumb but we need to preserve
+    # it because people are relying on it."
+    #    - Lukasa
+    #
+    # These are here solely to maintain backwards compatibility
+    # for things like ints. This will be removed in 3.0.0.
+    if not isinstance(username, basestring):
+        warnings.warn(
+            "Non-string usernames will no longer be supported in Requests "
+            "3.0.0. Please convert the object you've passed in ({!r}) to "
+            "a string or bytes object in the near future to avoid "
+            "problems.".format(username),
+            category=DeprecationWarning,
+        )
+        username = str(username)
+
+    if not isinstance(password, basestring):
+        warnings.warn(
+            "Non-string passwords will no longer be supported in Requests "
+            "3.0.0. Please convert the object you've passed in ({!r}) to "
+            "a string or bytes object in the near future to avoid "
+            "problems.".format(type(password)),
+            category=DeprecationWarning,
+        )
+        password = str(password)
+    # -- End Removal --
+
+    if isinstance(username, str):
+        username = username.encode("latin1")
+
+    if isinstance(password, str):
+        password = password.encode("latin1")
+
+    authstr = "Basic " + to_native_string(
+        b64encode(b":".join((username, password))).strip()
+    )
+
+    return authstr
+
+
+class AuthBase:
+    """Base class that all auth implementations derive from"""
+
+    def __call__(self, r):
+        raise NotImplementedError("Auth hooks must be callable.")
+
+
+class HTTPBasicAuth(AuthBase):
+    """Attaches HTTP Basic Authentication to the given Request object."""
+
+    def __init__(self, username, password):
+        self.username = username
+        self.password = password
+
+    def __eq__(self, other):
+        return all(
+            [
+                self.username == getattr(other, "username", None),
+                self.password == getattr(other, "password", None),
+            ]
+        )
+
+    def __ne__(self, other):
+        return not self == other
+
+    def __call__(self, r):
+        r.headers["Authorization"] = _basic_auth_str(self.username, self.password)
+        return r
+
+
+class HTTPProxyAuth(HTTPBasicAuth):
+    """Attaches HTTP Proxy Authentication to a given Request object."""
+
+    def __call__(self, r):
+        r.headers["Proxy-Authorization"] = _basic_auth_str(self.username, self.password)
+        return r
+
+
+class HTTPDigestAuth(AuthBase):
+    """Attaches HTTP Digest Authentication to the given Request object."""
+
+    def __init__(self, username, password):
+        self.username = username
+        self.password = password
+        # Keep state in per-thread local storage
+        self._thread_local = threading.local()
+
+    def init_per_thread_state(self):
+        # Ensure state is initialized just once per-thread
+        if not hasattr(self._thread_local, "init"):
+            self._thread_local.init = True
+            self._thread_local.last_nonce = ""
+            self._thread_local.nonce_count = 0
+            self._thread_local.chal = {}
+            self._thread_local.pos = None
+            self._thread_local.num_401_calls = None
+
+    def build_digest_header(self, method, url):
+        """
+        :rtype: str
+        """
+
+        realm = self._thread_local.chal["realm"]
+        nonce = self._thread_local.chal["nonce"]
+        qop = self._thread_local.chal.get("qop")
+        algorithm = self._thread_local.chal.get("algorithm")
+        opaque = self._thread_local.chal.get("opaque")
+        hash_utf8 = None
+
+        if algorithm is None:
+            _algorithm = "MD5"
+        else:
+            _algorithm = algorithm.upper()
+        # lambdas assume digest modules are imported at the top level
+        if _algorithm == "MD5" or _algorithm == "MD5-SESS":
+
+            def md5_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.md5(x).hexdigest()
+
+            hash_utf8 = md5_utf8
+        elif _algorithm == "SHA":
+
+            def sha_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha1(x).hexdigest()
+
+            hash_utf8 = sha_utf8
+        elif _algorithm == "SHA-256":
+
+            def sha256_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha256(x).hexdigest()
+
+            hash_utf8 = sha256_utf8
+        elif _algorithm == "SHA-512":
+
+            def sha512_utf8(x):
+                if isinstance(x, str):
+                    x = x.encode("utf-8")
+                return hashlib.sha512(x).hexdigest()
+
+            hash_utf8 = sha512_utf8
+
+        KD = lambda s, d: hash_utf8(f"{s}:{d}")  # noqa:E731
+
+        if hash_utf8 is None:
+            return None
+
+        # XXX not implemented yet
+        entdig = None
+        p_parsed = urlparse(url)
+        #: path is request-uri defined in RFC 2616 which should not be empty
+        path = p_parsed.path or "/"
+        if p_parsed.query:
+            path += f"?{p_parsed.query}"
+
+        A1 = f"{self.username}:{realm}:{self.password}"
+        A2 = f"{method}:{path}"
+
+        HA1 = hash_utf8(A1)
+        HA2 = hash_utf8(A2)
+
+        if nonce == self._thread_local.last_nonce:
+            self._thread_local.nonce_count += 1
+        else:
+            self._thread_local.nonce_count = 1
+        ncvalue = f"{self._thread_local.nonce_count:08x}"
+        s = str(self._thread_local.nonce_count).encode("utf-8")
+        s += nonce.encode("utf-8")
+        s += time.ctime().encode("utf-8")
+        s += os.urandom(8)
+
+        cnonce = hashlib.sha1(s).hexdigest()[:16]
+        if _algorithm == "MD5-SESS":
+            HA1 = hash_utf8(f"{HA1}:{nonce}:{cnonce}")
+
+        if not qop:
+            respdig = KD(HA1, f"{nonce}:{HA2}")
+        elif qop == "auth" or "auth" in qop.split(","):
+            noncebit = f"{nonce}:{ncvalue}:{cnonce}:auth:{HA2}"
+            respdig = KD(HA1, noncebit)
+        else:
+            # XXX handle auth-int.
+            return None
+
+        self._thread_local.last_nonce = nonce
+
+        # XXX should the partial digests be encoded too?
+        base = (
+            f'username="{self.username}", realm="{realm}", nonce="{nonce}", '
+            f'uri="{path}", response="{respdig}"'
+        )
+        if opaque:
+            base += f', opaque="{opaque}"'
+        if algorithm:
+            base += f', algorithm="{algorithm}"'
+        if entdig:
+            base += f', digest="{entdig}"'
+        if qop:
+            base += f', qop="auth", nc={ncvalue}, cnonce="{cnonce}"'
+
+        return f"Digest {base}"
+
+    def handle_redirect(self, r, **kwargs):
+        """Reset num_401_calls counter on redirects."""
+        if r.is_redirect:
+            self._thread_local.num_401_calls = 1
+
+    def handle_401(self, r, **kwargs):
+        """
+        Takes the given response and tries digest-auth, if needed.
+
+        :rtype: requests.Response
+        """
+
+        # If response is not 4xx, do not auth
+        # See https://github.com/psf/requests/issues/3772
+        if not 400 <= r.status_code < 500:
+            self._thread_local.num_401_calls = 1
+            return r
+
+        if self._thread_local.pos is not None:
+            # Rewind the file position indicator of the body to where
+            # it was to resend the request.
+            r.request.body.seek(self._thread_local.pos)
+        s_auth = r.headers.get("www-authenticate", "")
+
+        if "digest" in s_auth.lower() and self._thread_local.num_401_calls < 2:
+            self._thread_local.num_401_calls += 1
+            pat = re.compile(r"digest ", flags=re.IGNORECASE)
+            self._thread_local.chal = parse_dict_header(pat.sub("", s_auth, count=1))
+
+            # Consume content and release the original connection
+            # to allow our new request to reuse the same one.
+            r.content
+            r.close()
+            prep = r.request.copy()
+            extract_cookies_to_jar(prep._cookies, r.request, r.raw)
+            prep.prepare_cookies(prep._cookies)
+
+            prep.headers["Authorization"] = self.build_digest_header(
+                prep.method, prep.url
+            )
+            _r = r.connection.send(prep, **kwargs)
+            _r.history.append(r)
+            _r.request = prep
+
+            return _r
+
+        self._thread_local.num_401_calls = 1
+        return r
+
+    def __call__(self, r):
+        # Initialize per-thread state, if needed
+        self.init_per_thread_state()
+        # If we have a saved nonce, skip the 401
+        if self._thread_local.last_nonce:
+            r.headers["Authorization"] = self.build_digest_header(r.method, r.url)
+        try:
+            self._thread_local.pos = r.body.tell()
+        except AttributeError:
+            # In the case of HTTPDigestAuth being reused and the body of
+            # the previous request was a file-like object, pos has the
+            # file position of the previous body. Ensure it's set to
+            # None.
+            self._thread_local.pos = None
+        r.register_hook("response", self.handle_401)
+        r.register_hook("response", self.handle_redirect)
+        self._thread_local.num_401_calls = 1
+
+        return r
+
+    def __eq__(self, other):
+        return all(
+            [
+                self.username == getattr(other, "username", None),
+                self.password == getattr(other, "password", None),
+            ]
+        )
+
+    def __ne__(self, other):
+        return not self == other
diff --git a/backend/venv/lib/python3.12/site-packages/requests/certs.py b/backend/venv/lib/python3.12/site-packages/requests/certs.py
new file mode 100644
index 00000000..be422c3e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/certs.py
@@ -0,0 +1,17 @@
+#!/usr/bin/env python
+
+"""
+requests.certs
+~~~~~~~~~~~~~~
+
+This module returns the preferred default CA certificate bundle. There is
+only one — the one from the certifi package.
+
+If you are packaging Requests, e.g., for a Linux distribution or a managed
+environment, you can change the definition of where() to return a separately
+packaged CA bundle.
+"""
+from certifi import where
+
+if __name__ == "__main__":
+    print(where())
diff --git a/backend/venv/lib/python3.12/site-packages/requests/compat.py b/backend/venv/lib/python3.12/site-packages/requests/compat.py
new file mode 100644
index 00000000..095de1b6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/compat.py
@@ -0,0 +1,94 @@
+"""
+requests.compat
+~~~~~~~~~~~~~~~
+
+This module previously handled import compatibility issues
+between Python 2 and Python 3. It remains for backwards
+compatibility until the next major version.
+"""
+
+import importlib
+import sys
+
+# -------------------
+# Character Detection
+# -------------------
+
+
+def _resolve_char_detection():
+    """Find supported character detection libraries."""
+    chardet = None
+    for lib in ("chardet", "charset_normalizer"):
+        if chardet is None:
+            try:
+                chardet = importlib.import_module(lib)
+            except ImportError:
+                pass
+    return chardet
+
+
+chardet = _resolve_char_detection()
+
+# -------
+# Pythons
+# -------
+
+# Syntax sugar.
+_ver = sys.version_info
+
+#: Python 2.x?
+is_py2 = _ver[0] == 2
+
+#: Python 3.x?
+is_py3 = _ver[0] == 3
+
+# json/simplejson module import resolution
+has_simplejson = False
+try:
+    import simplejson as json
+
+    has_simplejson = True
+except ImportError:
+    import json
+
+if has_simplejson:
+    from simplejson import JSONDecodeError
+else:
+    from json import JSONDecodeError
+
+# Keep OrderedDict for backwards compatibility.
+from collections import OrderedDict
+from collections.abc import Callable, Mapping, MutableMapping
+from http import cookiejar as cookielib
+from http.cookies import Morsel
+from io import StringIO
+
+# --------------
+# Legacy Imports
+# --------------
+from urllib.parse import (
+    quote,
+    quote_plus,
+    unquote,
+    unquote_plus,
+    urldefrag,
+    urlencode,
+    urljoin,
+    urlparse,
+    urlsplit,
+    urlunparse,
+)
+from urllib.request import (
+    getproxies,
+    getproxies_environment,
+    parse_http_list,
+    proxy_bypass,
+    proxy_bypass_environment,
+)
+
+builtin_str = str
+str = str
+bytes = bytes
+basestring = (str, bytes)
+numeric_types = (int, float)
+integer_types = (int,)
diff --git a/backend/venv/lib/python3.12/site-packages/requests/cookies.py b/backend/venv/lib/python3.12/site-packages/requests/cookies.py
new file mode 100644
index 00000000..f69d0cda
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/cookies.py
@@ -0,0 +1,561 @@
+"""
+requests.cookies
+~~~~~~~~~~~~~~~~
+
+Compatibility code to be able to use `http.cookiejar.CookieJar` with requests.
+
+requests.utils imports from here, so be careful with imports.
+"""
+
+import calendar
+import copy
+import time
+
+from ._internal_utils import to_native_string
+from .compat import Morsel, MutableMapping, cookielib, urlparse, urlunparse
+
+try:
+    import threading
+except ImportError:
+    import dummy_threading as threading
+
+
+class MockRequest:
+    """Wraps a `requests.Request` to mimic a `urllib2.Request`.
+
+    The code in `http.cookiejar.CookieJar` expects this interface in order to correctly
+    manage cookie policies, i.e., determine whether a cookie can be set, given the
+    domains of the request and the cookie.
+
+    The original request object is read-only. The client is responsible for collecting
+    the new headers via `get_new_headers()` and interpreting them appropriately. You
+    probably want `get_cookie_header`, defined below.
+    """
+
+    def __init__(self, request):
+        self._r = request
+        self._new_headers = {}
+        self.type = urlparse(self._r.url).scheme
+
+    def get_type(self):
+        return self.type
+
+    def get_host(self):
+        return urlparse(self._r.url).netloc
+
+    def get_origin_req_host(self):
+        return self.get_host()
+
+    def get_full_url(self):
+        # Only return the response's URL if the user hadn't set the Host
+        # header
+        if not self._r.headers.get("Host"):
+            return self._r.url
+        # If they did set it, retrieve it and reconstruct the expected domain
+        host = to_native_string(self._r.headers["Host"], encoding="utf-8")
+        parsed = urlparse(self._r.url)
+        # Reconstruct the URL as we expect it
+        return urlunparse(
+            [
+                parsed.scheme,
+                host,
+                parsed.path,
+                parsed.params,
+                parsed.query,
+                parsed.fragment,
+            ]
+        )
+
+    def is_unverifiable(self):
+        return True
+
+    def has_header(self, name):
+        return name in self._r.headers or name in self._new_headers
+
+    def get_header(self, name, default=None):
+        return self._r.headers.get(name, self._new_headers.get(name, default))
+
+    def add_header(self, key, val):
+        """cookiejar has no legitimate use for this method; add it back if you find one."""
+        raise NotImplementedError(
+            "Cookie headers should be added with add_unredirected_header()"
+        )
+
+    def add_unredirected_header(self, name, value):
+        self._new_headers[name] = value
+
+    def get_new_headers(self):
+        return self._new_headers
+
+    @property
+    def unverifiable(self):
+        return self.is_unverifiable()
+
+    @property
+    def origin_req_host(self):
+        return self.get_origin_req_host()
+
+    @property
+    def host(self):
+        return self.get_host()
+
+
+class MockResponse:
+    """Wraps a `httplib.HTTPMessage` to mimic a `urllib.addinfourl`.
+
+    ...what? Basically, expose the parsed HTTP headers from the server response
+    the way `http.cookiejar` expects to see them.
+    """
+
+    def __init__(self, headers):
+        """Make a MockResponse for `cookiejar` to read.
+
+        :param headers: a httplib.HTTPMessage or analogous carrying the headers
+        """
+        self._headers = headers
+
+    def info(self):
+        return self._headers
+
+    def getheaders(self, name):
+        self._headers.getheaders(name)
+
+
+def extract_cookies_to_jar(jar, request, response):
+    """Extract the cookies from the response into a CookieJar.
+
+    :param jar: http.cookiejar.CookieJar (not necessarily a RequestsCookieJar)
+    :param request: our own requests.Request object
+    :param response: urllib3.HTTPResponse object
+    """
+    if not (hasattr(response, "_original_response") and response._original_response):
+        return
+    # the _original_response field is the wrapped httplib.HTTPResponse object,
+    req = MockRequest(request)
+    # pull out the HTTPMessage with the headers and put it in the mock:
+    res = MockResponse(response._original_response.msg)
+    jar.extract_cookies(res, req)
+
+
+def get_cookie_header(jar, request):
+    """
+    Produce an appropriate Cookie header string to be sent with `request`, or None.
+
+    :rtype: str
+    """
+    r = MockRequest(request)
+    jar.add_cookie_header(r)
+    return r.get_new_headers().get("Cookie")
+
+
+def remove_cookie_by_name(cookiejar, name, domain=None, path=None):
+    """Unsets a cookie by name, by default over all domains and paths.
+
+    Wraps CookieJar.clear(), is O(n).
+    """
+    clearables = []
+    for cookie in cookiejar:
+        if cookie.name != name:
+            continue
+        if domain is not None and domain != cookie.domain:
+            continue
+        if path is not None and path != cookie.path:
+            continue
+        clearables.append((cookie.domain, cookie.path, cookie.name))
+
+    for domain, path, name in clearables:
+        cookiejar.clear(domain, path, name)
+
+
+class CookieConflictError(RuntimeError):
+    """There are two cookies that meet the criteria specified in the cookie jar.
+    Use .get and .set and include domain and path args in order to be more specific.
+    """
+
+
+class RequestsCookieJar(cookielib.CookieJar, MutableMapping):
+    """Compatibility class; is a http.cookiejar.CookieJar, but exposes a dict
+    interface.
+
+    This is the CookieJar we create by default for requests and sessions that
+    don't specify one, since some clients may expect response.cookies and
+    session.cookies to support dict operations.
+
+    Requests does not use the dict interface internally; it's just for
+    compatibility with external client code. All requests code should work
+    out of the box with externally provided instances of ``CookieJar``, e.g.
+    ``LWPCookieJar`` and ``FileCookieJar``.
+
+    Unlike a regular CookieJar, this class is pickleable.
+
+    .. warning:: dictionary operations that are normally O(1) may be O(n).
+    """
+
+    def get(self, name, default=None, domain=None, path=None):
+        """Dict-like get() that also supports optional domain and path args in
+        order to resolve naming collisions from using one cookie jar over
+        multiple domains.
+
+        .. warning:: operation is O(n), not O(1).
+        """
+        try:
+            return self._find_no_duplicates(name, domain, path)
+        except KeyError:
+            return default
+
+    def set(self, name, value, **kwargs):
+        """Dict-like set() that also supports optional domain and path args in
+        order to resolve naming collisions from using one cookie jar over
+        multiple domains.
+        """
+        # support client code that unsets cookies by assignment of a None value:
+        if value is None:
+            remove_cookie_by_name(
+                self, name, domain=kwargs.get("domain"), path=kwargs.get("path")
+            )
+            return
+
+        if isinstance(value, Morsel):
+            c = morsel_to_cookie(value)
+        else:
+            c = create_cookie(name, value, **kwargs)
+        self.set_cookie(c)
+        return c
+
+    def iterkeys(self):
+        """Dict-like iterkeys() that returns an iterator of names of cookies
+        from the jar.
+
+        .. seealso:: itervalues() and iteritems().
+        """
+        for cookie in iter(self):
+            yield cookie.name
+
+    def keys(self):
+        """Dict-like keys() that returns a list of names of cookies from the
+        jar.
+
+        .. seealso:: values() and items().
+        """
+        return list(self.iterkeys())
+
+    def itervalues(self):
+        """Dict-like itervalues() that returns an iterator of values of cookies
+        from the jar.
+
+        .. seealso:: iterkeys() and iteritems().
+        """
+        for cookie in iter(self):
+            yield cookie.value
+
+    def values(self):
+        """Dict-like values() that returns a list of values of cookies from the
+        jar.
+
+        .. seealso:: keys() and items().
+        """
+        return list(self.itervalues())
+
+    def iteritems(self):
+        """Dict-like iteritems() that returns an iterator of name-value tuples
+        from the jar.
+
+        .. seealso:: iterkeys() and itervalues().
+        """
+        for cookie in iter(self):
+            yield cookie.name, cookie.value
+
+    def items(self):
+        """Dict-like items() that returns a list of name-value tuples from the
+        jar. Allows client-code to call ``dict(RequestsCookieJar)`` and get a
+        vanilla python dict of key value pairs.
+
+        .. seealso:: keys() and values().
+        """
+        return list(self.iteritems())
+
+    def list_domains(self):
+        """Utility method to list all the domains in the jar."""
+        domains = []
+        for cookie in iter(self):
+            if cookie.domain not in domains:
+                domains.append(cookie.domain)
+        return domains
+
+    def list_paths(self):
+        """Utility method to list all the paths in the jar."""
+        paths = []
+        for cookie in iter(self):
+            if cookie.path not in paths:
+                paths.append(cookie.path)
+        return paths
+
+    def multiple_domains(self):
+        """Returns True if there are multiple domains in the jar.
+        Returns False otherwise.
+
+        :rtype: bool
+        """
+        domains = []
+        for cookie in iter(self):
+            if cookie.domain is not None and cookie.domain in domains:
+                return True
+            domains.append(cookie.domain)
+        return False  # there is only one domain in jar
+
+    def get_dict(self, domain=None, path=None):
+        """Takes as an argument an optional domain and path and returns a plain
+        old Python dict of name-value pairs of cookies that meet the
+        requirements.
+
+        :rtype: dict
+        """
+        dictionary = {}
+        for cookie in iter(self):
+            if (domain is None or cookie.domain == domain) and (
+                path is None or cookie.path == path
+            ):
+                dictionary[cookie.name] = cookie.value
+        return dictionary
+
+    def __contains__(self, name):
+        try:
+            return super().__contains__(name)
+        except CookieConflictError:
+            return True
+
+    def __getitem__(self, name):
+        """Dict-like __getitem__() for compatibility with client code. Throws
+        exception if there are more than one cookie with name. In that case,
+        use the more explicit get() method instead.
+
+        .. warning:: operation is O(n), not O(1).
+        """
+        return self._find_no_duplicates(name)
+
+    def __setitem__(self, name, value):
+        """Dict-like __setitem__ for compatibility with client code. Throws
+        exception if there is already a cookie of that name in the jar. In that
+        case, use the more explicit set() method instead.
+        """
+        self.set(name, value)
+
+    def __delitem__(self, name):
+        """Deletes a cookie given a name. Wraps ``http.cookiejar.CookieJar``'s
+        ``remove_cookie_by_name()``.
+        """
+        remove_cookie_by_name(self, name)
+
+    def set_cookie(self, cookie, *args, **kwargs):
+        if (
+            hasattr(cookie.value, "startswith")
+            and cookie.value.startswith('"')
+            and cookie.value.endswith('"')
+        ):
+            cookie.value = cookie.value.replace('\\"', "")
+        return super().set_cookie(cookie, *args, **kwargs)
+
+    def update(self, other):
+        """Updates this jar with cookies from another CookieJar or dict-like"""
+        if isinstance(other, cookielib.CookieJar):
+            for cookie in other:
+                self.set_cookie(copy.copy(cookie))
+        else:
+            super().update(other)
+
+    def _find(self, name, domain=None, path=None):
+        """Requests uses this method internally to get cookie values.
+
+        If there are conflicting cookies, _find arbitrarily chooses one.
+        See _find_no_duplicates if you want an exception thrown if there are
+        conflicting cookies.
+
+        :param name: a string containing name of cookie
+        :param domain: (optional) string containing domain of cookie
+        :param path: (optional) string containing path of cookie
+        :return: cookie.value
+        """
+        for cookie in iter(self):
+            if cookie.name == name:
+                if domain is None or cookie.domain == domain:
+                    if path is None or cookie.path == path:
+                        return cookie.value
+
+        raise KeyError(f"name={name!r}, domain={domain!r}, path={path!r}")
+
+    def _find_no_duplicates(self, name, domain=None, path=None):
+        """Both ``__get_item__`` and ``get`` call this function: it's never
+        used elsewhere in Requests.
+
+        :param name: a string containing name of cookie
+        :param domain: (optional) string containing domain of cookie
+        :param path: (optional) string containing path of cookie
+        :raises KeyError: if cookie is not found
+        :raises CookieConflictError: if there are multiple cookies
+            that match name and optionally domain and path
+        :return: cookie.value
+        """
+        toReturn = None
+        for cookie in iter(self):
+            if cookie.name == name:
+                if domain is None or cookie.domain == domain:
+                    if path is None or cookie.path == path:
+                        if toReturn is not None:
+                            # if there are multiple cookies that meet passed in criteria
+                            raise CookieConflictError(
+                                f"There are multiple cookies with name, {name!r}"
+                            )
+                        # we will eventually return this as long as no cookie conflict
+                        toReturn = cookie.value
+
+        if toReturn:
+            return toReturn
+        raise KeyError(f"name={name!r}, domain={domain!r}, path={path!r}")
+
+    def __getstate__(self):
+        """Unlike a normal CookieJar, this class is pickleable."""
+        state = self.__dict__.copy()
+        # remove the unpickleable RLock object
+        state.pop("_cookies_lock")
+        return state
+
+    def __setstate__(self, state):
+        """Unlike a normal CookieJar, this class is pickleable."""
+        self.__dict__.update(state)
+        if "_cookies_lock" not in self.__dict__:
+            self._cookies_lock = threading.RLock()
+
+    def copy(self):
+        """Return a copy of this RequestsCookieJar."""
+        new_cj = RequestsCookieJar()
+        new_cj.set_policy(self.get_policy())
+        new_cj.update(self)
+        return new_cj
+
+    def get_policy(self):
+        """Return the CookiePolicy instance used."""
+        return self._policy
+
+
+def _copy_cookie_jar(jar):
+    if jar is None:
+        return None
+
+    if hasattr(jar, "copy"):
+        # We're dealing with an instance of RequestsCookieJar
+        return jar.copy()
+    # We're dealing with a generic CookieJar instance
+    new_jar = copy.copy(jar)
+    new_jar.clear()
+    for cookie in jar:
+        new_jar.set_cookie(copy.copy(cookie))
+    return new_jar
+
+
+def create_cookie(name, value, **kwargs):
+    """Make a cookie from underspecified parameters.
+
+    By default, the pair of `name` and `value` will be set for the domain ''
+    and sent on every request (this is sometimes called a "supercookie").
+    """
+    result = {
+        "version": 0,
+        "name": name,
+        "value": value,
+        "port": None,
+        "domain": "",
+        "path": "/",
+        "secure": False,
+        "expires": None,
+        "discard": True,
+        "comment": None,
+        "comment_url": None,
+        "rest": {"HttpOnly": None},
+        "rfc2109": False,
+    }
+
+    badargs = set(kwargs) - set(result)
+    if badargs:
+        raise TypeError(
+            f"create_cookie() got unexpected keyword arguments: {list(badargs)}"
+        )
+
+    result.update(kwargs)
+    result["port_specified"] = bool(result["port"])
+    result["domain_specified"] = bool(result["domain"])
+    result["domain_initial_dot"] = result["domain"].startswith(".")
+    result["path_specified"] = bool(result["path"])
+
+    return cookielib.Cookie(**result)
+
+
+def morsel_to_cookie(morsel):
+    """Convert a Morsel object into a Cookie containing the one k/v pair."""
+
+    expires = None
+    if morsel["max-age"]:
+        try:
+            expires = int(time.time() + int(morsel["max-age"]))
+        except ValueError:
+            raise TypeError(f"max-age: {morsel['max-age']} must be integer")
+    elif morsel["expires"]:
+        time_template = "%a, %d-%b-%Y %H:%M:%S GMT"
+        expires = calendar.timegm(time.strptime(morsel["expires"], time_template))
+    return create_cookie(
+        comment=morsel["comment"],
+        comment_url=bool(morsel["comment"]),
+        discard=False,
+        domain=morsel["domain"],
+        expires=expires,
+        name=morsel.key,
+        path=morsel["path"],
+        port=None,
+        rest={"HttpOnly": morsel["httponly"]},
+        rfc2109=False,
+        secure=bool(morsel["secure"]),
+        value=morsel.value,
+        version=morsel["version"] or 0,
+    )
+
+
+def cookiejar_from_dict(cookie_dict, cookiejar=None, overwrite=True):
+    """Returns a CookieJar from a key/value dictionary.
+
+    :param cookie_dict: Dict of key/values to insert into CookieJar.
+    :param cookiejar: (optional) A cookiejar to add the cookies to.
+    :param overwrite: (optional) If False, will not replace cookies
+        already in the jar with new ones.
+    :rtype: CookieJar
+    """
+    if cookiejar is None:
+        cookiejar = RequestsCookieJar()
+
+    if cookie_dict is not None:
+        names_from_jar = [cookie.name for cookie in cookiejar]
+        for name in cookie_dict:
+            if overwrite or (name not in names_from_jar):
+                cookiejar.set_cookie(create_cookie(name, cookie_dict[name]))
+
+    return cookiejar
+
+
+def merge_cookies(cookiejar, cookies):
+    """Add cookies to cookiejar and returns a merged CookieJar.
+
+    :param cookiejar: CookieJar object to add the cookies to.
+    :param cookies: Dictionary or CookieJar object to be added.
+    :rtype: CookieJar
+    """
+    if not isinstance(cookiejar, cookielib.CookieJar):
+        raise ValueError("You can only merge into CookieJar")
+
+    if isinstance(cookies, dict):
+        cookiejar = cookiejar_from_dict(cookies, cookiejar=cookiejar, overwrite=False)
+    elif isinstance(cookies, cookielib.CookieJar):
+        try:
+            cookiejar.update(cookies)
+        except AttributeError:
+            for cookie_in_jar in cookies:
+                cookiejar.set_cookie(cookie_in_jar)
+
+    return cookiejar
diff --git a/backend/venv/lib/python3.12/site-packages/requests/exceptions.py b/backend/venv/lib/python3.12/site-packages/requests/exceptions.py
new file mode 100644
index 00000000..83986b48
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/exceptions.py
@@ -0,0 +1,151 @@
+"""
+requests.exceptions
+~~~~~~~~~~~~~~~~~~~
+
+This module contains the set of Requests' exceptions.
+"""
+from urllib3.exceptions import HTTPError as BaseHTTPError
+
+from .compat import JSONDecodeError as CompatJSONDecodeError
+
+
+class RequestException(IOError):
+    """There was an ambiguous exception that occurred while handling your
+    request.
+    """
+
+    def __init__(self, *args, **kwargs):
+        """Initialize RequestException with `request` and `response` objects."""
+        response = kwargs.pop("response", None)
+        self.response = response
+        self.request = kwargs.pop("request", None)
+        if response is not None and not self.request and hasattr(response, "request"):
+            self.request = self.response.request
+        super().__init__(*args, **kwargs)
+
+
+class InvalidJSONError(RequestException):
+    """A JSON error occurred."""
+
+
+class JSONDecodeError(InvalidJSONError, CompatJSONDecodeError):
+    """Couldn't decode the text into json"""
+
+    def __init__(self, *args, **kwargs):
+        """
+        Construct the JSONDecodeError instance first with all
+        args. Then use it's args to construct the IOError so that
+        the json specific args aren't used as IOError specific args
+        and the error message from JSONDecodeError is preserved.
+        """
+        CompatJSONDecodeError.__init__(self, *args)
+        InvalidJSONError.__init__(self, *self.args, **kwargs)
+
+    def __reduce__(self):
+        """
+        The __reduce__ method called when pickling the object must
+        be the one from the JSONDecodeError (be it json/simplejson)
+        as it expects all the arguments for instantiation, not just
+        one like the IOError, and the MRO would by default call the
+        __reduce__ method from the IOError due to the inheritance order.
+        """
+        return CompatJSONDecodeError.__reduce__(self)
+
+
+class HTTPError(RequestException):
+    """An HTTP error occurred."""
+
+
+class ConnectionError(RequestException):
+    """A Connection error occurred."""
+
+
+class ProxyError(ConnectionError):
+    """A proxy error occurred."""
+
+
+class SSLError(ConnectionError):
+    """An SSL error occurred."""
+
+
+class Timeout(RequestException):
+    """The request timed out.
+
+    Catching this error will catch both
+    :exc:`~requests.exceptions.ConnectTimeout` and
+    :exc:`~requests.exceptions.ReadTimeout` errors.
+    """
+
+
+class ConnectTimeout(ConnectionError, Timeout):
+    """The request timed out while trying to connect to the remote server.
+
+    Requests that produced this error are safe to retry.
+    """
+
+
+class ReadTimeout(Timeout):
+    """The server did not send any data in the allotted amount of time."""
+
+
+class URLRequired(RequestException):
+    """A valid URL is required to make a request."""
+
+
+class TooManyRedirects(RequestException):
+    """Too many redirects."""
+
+
+class MissingSchema(RequestException, ValueError):
+    """The URL scheme (e.g. http or https) is missing."""
+
+
+class InvalidSchema(RequestException, ValueError):
+    """The URL scheme provided is either invalid or unsupported."""
+
+
+class InvalidURL(RequestException, ValueError):
+    """The URL provided was somehow invalid."""
+
+
+class InvalidHeader(RequestException, ValueError):
+    """The header value provided was somehow invalid."""
+
+
+class InvalidProxyURL(InvalidURL):
+    """The proxy URL provided is invalid."""
+
+
+class ChunkedEncodingError(RequestException):
+    """The server declared chunked encoding but sent an invalid chunk."""
+
+
+class ContentDecodingError(RequestException, BaseHTTPError):
+    """Failed to decode response content."""
+
+
+class StreamConsumedError(RequestException, TypeError):
+    """The content for this response was already consumed."""
+
+
+class RetryError(RequestException):
+    """Custom retries logic failed"""
+
+
+class UnrewindableBodyError(RequestException):
+    """Requests encountered an error when trying to rewind a body."""
+
+
+# Warnings
+
+
+class RequestsWarning(Warning):
+    """Base warning for Requests."""
+
+
+class FileModeWarning(RequestsWarning, DeprecationWarning):
+    """A file was opened in text mode, but Requests determined its binary length."""
+
+
+class RequestsDependencyWarning(RequestsWarning):
+    """An imported dependency doesn't match the expected version range."""
diff --git a/backend/venv/lib/python3.12/site-packages/requests/help.py b/backend/venv/lib/python3.12/site-packages/requests/help.py
new file mode 100644
index 00000000..8fbcd656
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/help.py
@@ -0,0 +1,134 @@
+"""Module containing bug report helper(s)."""
+
+import json
+import platform
+import ssl
+import sys
+
+import idna
+import urllib3
+
+from . import __version__ as requests_version
+
+try:
+    import charset_normalizer
+except ImportError:
+    charset_normalizer = None
+
+try:
+    import chardet
+except ImportError:
+    chardet = None
+
+try:
+    from urllib3.contrib import pyopenssl
+except ImportError:
+    pyopenssl = None
+    OpenSSL = None
+    cryptography = None
+else:
+    import cryptography
+    import OpenSSL
+
+
+def _implementation():
+    """Return a dict with the Python implementation and version.
+
+    Provide both the name and the version of the Python implementation
+    currently running. For example, on CPython 3.10.3 it will return
+    {'name': 'CPython', 'version': '3.10.3'}.
+
+    This function works best on CPython and PyPy: in particular, it probably
+    doesn't work for Jython or IronPython. Future investigation should be done
+    to work out the correct shape of the code for those platforms.
+    """
+    implementation = platform.python_implementation()
+
+    if implementation == "CPython":
+        implementation_version = platform.python_version()
+    elif implementation == "PyPy":
+        implementation_version = "{}.{}.{}".format(
+            sys.pypy_version_info.major,
+            sys.pypy_version_info.minor,
+            sys.pypy_version_info.micro,
+        )
+        if sys.pypy_version_info.releaselevel != "final":
+            implementation_version = "".join(
+                [implementation_version, sys.pypy_version_info.releaselevel]
+            )
+    elif implementation == "Jython":
+        implementation_version = platform.python_version()  # Complete Guess
+    elif implementation == "IronPython":
+        implementation_version = platform.python_version()  # Complete Guess
+    else:
+        implementation_version = "Unknown"
+
+    return {"name": implementation, "version": implementation_version}
+
+
+def info():
+    """Generate information for a bug report."""
+    try:
+        platform_info = {
+            "system": platform.system(),
+            "release": platform.release(),
+        }
+    except OSError:
+        platform_info = {
+            "system": "Unknown",
+            "release": "Unknown",
+        }
+
+    implementation_info = _implementation()
+    urllib3_info = {"version": urllib3.__version__}
+    charset_normalizer_info = {"version": None}
+    chardet_info = {"version": None}
+    if charset_normalizer:
+        charset_normalizer_info = {"version": charset_normalizer.__version__}
+    if chardet:
+        chardet_info = {"version": chardet.__version__}
+
+    pyopenssl_info = {
+        "version": None,
+        "openssl_version": "",
+    }
+    if OpenSSL:
+        pyopenssl_info = {
+            "version": OpenSSL.__version__,
+            "openssl_version": f"{OpenSSL.SSL.OPENSSL_VERSION_NUMBER:x}",
+        }
+    cryptography_info = {
+        "version": getattr(cryptography, "__version__", ""),
+    }
+    idna_info = {
+        "version": getattr(idna, "__version__", ""),
+    }
+
+    system_ssl = ssl.OPENSSL_VERSION_NUMBER
+    system_ssl_info = {"version": f"{system_ssl:x}" if system_ssl is not None else ""}
+
+    return {
+        "platform": platform_info,
+        "implementation": implementation_info,
+        "system_ssl": system_ssl_info,
+        "using_pyopenssl": pyopenssl is not None,
+        "using_charset_normalizer": chardet is None,
+        "pyOpenSSL": pyopenssl_info,
+        "urllib3": urllib3_info,
+        "chardet": chardet_info,
+        "charset_normalizer": charset_normalizer_info,
+        "cryptography": cryptography_info,
+        "idna": idna_info,
+        "requests": {
+            "version": requests_version,
+        },
+    }
+
+
+def main():
+    """Pretty-print the bug information as JSON."""
+    print(json.dumps(info(), sort_keys=True, indent=2))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/requests/hooks.py b/backend/venv/lib/python3.12/site-packages/requests/hooks.py
new file mode 100644
index 00000000..d181ba2e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/hooks.py
@@ -0,0 +1,33 @@
+"""
+requests.hooks
+~~~~~~~~~~~~~~
+
+This module provides the capabilities for the Requests hooks system.
+
+Available hooks:
+
+``response``:
+    The response generated from a Request.
+"""
+HOOKS = ["response"]
+
+
+def default_hooks():
+    return {event: [] for event in HOOKS}
+
+
+# TODO: response is the only one
+
+
+def dispatch_hook(key, hooks, hook_data, **kwargs):
+    """Dispatches a hook dictionary on a given piece of data."""
+    hooks = hooks or {}
+    hooks = hooks.get(key)
+    if hooks:
+        if hasattr(hooks, "__call__"):
+            hooks = [hooks]
+        for hook in hooks:
+            _hook_data = hook(hook_data, **kwargs)
+            if _hook_data is not None:
+                hook_data = _hook_data
+    return hook_data
diff --git a/backend/venv/lib/python3.12/site-packages/requests/models.py b/backend/venv/lib/python3.12/site-packages/requests/models.py
new file mode 100644
index 00000000..8f56ca7d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/models.py
@@ -0,0 +1,1037 @@
+"""
+requests.models
+~~~~~~~~~~~~~~~
+
+This module contains the primary objects that power Requests.
+"""
+
+import datetime
+
+# Import encoding now, to avoid implicit import later.
+# Implicit import within threads may cause LookupError when standard library is in a ZIP,
+# such as in Embedded Python. See https://github.com/psf/requests/issues/3578.
+import encodings.idna  # noqa: F401
+from io import UnsupportedOperation
+
+from urllib3.exceptions import (
+    DecodeError,
+    LocationParseError,
+    ProtocolError,
+    ReadTimeoutError,
+    SSLError,
+)
+from urllib3.fields import RequestField
+from urllib3.filepost import encode_multipart_formdata
+from urllib3.util import parse_url
+
+from ._internal_utils import to_native_string, unicode_is_ascii
+from .auth import HTTPBasicAuth
+from .compat import (
+    Callable,
+    JSONDecodeError,
+    Mapping,
+    basestring,
+    builtin_str,
+    chardet,
+    cookielib,
+)
+from .compat import json as complexjson
+from .compat import urlencode, urlsplit, urlunparse
+from .cookies import _copy_cookie_jar, cookiejar_from_dict, get_cookie_header
+from .exceptions import (
+    ChunkedEncodingError,
+    ConnectionError,
+    ContentDecodingError,
+    HTTPError,
+    InvalidJSONError,
+    InvalidURL,
+)
+from .exceptions import JSONDecodeError as RequestsJSONDecodeError
+from .exceptions import MissingSchema
+from .exceptions import SSLError as RequestsSSLError
+from .exceptions import StreamConsumedError
+from .hooks import default_hooks
+from .status_codes import codes
+from .structures import CaseInsensitiveDict
+from .utils import (
+    check_header_validity,
+    get_auth_from_url,
+    guess_filename,
+    guess_json_utf,
+    iter_slices,
+    parse_header_links,
+    requote_uri,
+    stream_decode_response_unicode,
+    super_len,
+    to_key_val_list,
+)
+
+#: The set of HTTP status codes that indicate an automatically
+#: processable redirect.
+REDIRECT_STATI = (
+    codes.moved,  # 301
+    codes.found,  # 302
+    codes.other,  # 303
+    codes.temporary_redirect,  # 307
+    codes.permanent_redirect,  # 308
+)
+
+DEFAULT_REDIRECT_LIMIT = 30
+CONTENT_CHUNK_SIZE = 10 * 1024
+ITER_CHUNK_SIZE = 512
+
+
+class RequestEncodingMixin:
+    @property
+    def path_url(self):
+        """Build the path URL to use."""
+
+        url = []
+
+        p = urlsplit(self.url)
+
+        path = p.path
+        if not path:
+            path = "/"
+
+        url.append(path)
+
+        query = p.query
+        if query:
+            url.append("?")
+            url.append(query)
+
+        return "".join(url)
+
+    @staticmethod
+    def _encode_params(data):
+        """Encode parameters in a piece of data.
+
+        Will successfully encode parameters when passed as a dict or a list of
+        2-tuples. Order is retained if data is a list of 2-tuples but arbitrary
+        if parameters are supplied as a dict.
+        """
+
+        if isinstance(data, (str, bytes)):
+            return data
+        elif hasattr(data, "read"):
+            return data
+        elif hasattr(data, "__iter__"):
+            result = []
+            for k, vs in to_key_val_list(data):
+                if isinstance(vs, basestring) or not hasattr(vs, "__iter__"):
+                    vs = [vs]
+                for v in vs:
+                    if v is not None:
+                        result.append(
+                            (
+                                k.encode("utf-8") if isinstance(k, str) else k,
+                                v.encode("utf-8") if isinstance(v, str) else v,
+                            )
+                        )
+            return urlencode(result, doseq=True)
+        else:
+            return data
+
+    @staticmethod
+    def _encode_files(files, data):
+        """Build the body for a multipart/form-data request.
+
+        Will successfully encode files when passed as a dict or a list of
+        tuples. Order is retained if data is a list of tuples but arbitrary
+        if parameters are supplied as a dict.
+        The tuples may be 2-tuples (filename, fileobj), 3-tuples (filename, fileobj, contentype)
+        or 4-tuples (filename, fileobj, contentype, custom_headers).
+        """
+        if not files:
+            raise ValueError("Files must be provided.")
+        elif isinstance(data, basestring):
+            raise ValueError("Data must not be a string.")
+
+        new_fields = []
+        fields = to_key_val_list(data or {})
+        files = to_key_val_list(files or {})
+
+        for field, val in fields:
+            if isinstance(val, basestring) or not hasattr(val, "__iter__"):
+                val = [val]
+            for v in val:
+                if v is not None:
+                    # Don't call str() on bytestrings: in Py3 it all goes wrong.
+                    if not isinstance(v, bytes):
+                        v = str(v)
+
+                    new_fields.append(
+                        (
+                            field.decode("utf-8")
+                            if isinstance(field, bytes)
+                            else field,
+                            v.encode("utf-8") if isinstance(v, str) else v,
+                        )
+                    )
+
+        for k, v in files:
+            # support for explicit filename
+            ft = None
+            fh = None
+            if isinstance(v, (tuple, list)):
+                if len(v) == 2:
+                    fn, fp = v
+                elif len(v) == 3:
+                    fn, fp, ft = v
+                else:
+                    fn, fp, ft, fh = v
+            else:
+                fn = guess_filename(v) or k
+                fp = v
+
+            if isinstance(fp, (str, bytes, bytearray)):
+                fdata = fp
+            elif hasattr(fp, "read"):
+                fdata = fp.read()
+            elif fp is None:
+                continue
+            else:
+                fdata = fp
+
+            rf = RequestField(name=k, data=fdata, filename=fn, headers=fh)
+            rf.make_multipart(content_type=ft)
+            new_fields.append(rf)
+
+        body, content_type = encode_multipart_formdata(new_fields)
+
+        return body, content_type
+
+
+class RequestHooksMixin:
+    def register_hook(self, event, hook):
+        """Properly register a hook."""
+
+        if event not in self.hooks:
+            raise ValueError(f'Unsupported event specified, with event name "{event}"')
+
+        if isinstance(hook, Callable):
+            self.hooks[event].append(hook)
+        elif hasattr(hook, "__iter__"):
+            self.hooks[event].extend(h for h in hook if isinstance(h, Callable))
+
+    def deregister_hook(self, event, hook):
+        """Deregister a previously registered hook.
+        Returns True if the hook existed, False if not.
+        """
+
+        try:
+            self.hooks[event].remove(hook)
+            return True
+        except ValueError:
+            return False
+
+
+class Request(RequestHooksMixin):
+    """A user-created :class:`Request <Request>` object.
+
+    Used to prepare a :class:`PreparedRequest <PreparedRequest>`, which is sent to the server.
+
+    :param method: HTTP method to use.
+    :param url: URL to send.
+    :param headers: dictionary of headers to send.
+    :param files: dictionary of {filename: fileobject} files to multipart upload.
+    :param data: the body to attach to the request. If a dictionary or
+        list of tuples ``[(key, value)]`` is provided, form-encoding will
+        take place.
+    :param json: json for the body to attach to the request (if files or data is not specified).
+    :param params: URL parameters to append to the URL. If a dictionary or
+        list of tuples ``[(key, value)]`` is provided, form-encoding will
+        take place.
+    :param auth: Auth handler or (user, pass) tuple.
+    :param cookies: dictionary or CookieJar of cookies to attach to this request.
+    :param hooks: dictionary of callback hooks, for internal usage.
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.Request('GET', 'https://httpbin.org/get')
+      >>> req.prepare()
+      <PreparedRequest [GET]>
+    """
+
+    def __init__(
+        self,
+        method=None,
+        url=None,
+        headers=None,
+        files=None,
+        data=None,
+        params=None,
+        auth=None,
+        cookies=None,
+        hooks=None,
+        json=None,
+    ):
+        # Default empty dicts for dict params.
+        data = [] if data is None else data
+        files = [] if files is None else files
+        headers = {} if headers is None else headers
+        params = {} if params is None else params
+        hooks = {} if hooks is None else hooks
+
+        self.hooks = default_hooks()
+        for k, v in list(hooks.items()):
+            self.register_hook(event=k, hook=v)
+
+        self.method = method
+        self.url = url
+        self.headers = headers
+        self.files = files
+        self.data = data
+        self.json = json
+        self.params = params
+        self.auth = auth
+        self.cookies = cookies
+
+    def __repr__(self):
+        return f"<Request [{self.method}]>"
+
+    def prepare(self):
+        """Constructs a :class:`PreparedRequest <PreparedRequest>` for transmission and returns it."""
+        p = PreparedRequest()
+        p.prepare(
+            method=self.method,
+            url=self.url,
+            headers=self.headers,
+            files=self.files,
+            data=self.data,
+            json=self.json,
+            params=self.params,
+            auth=self.auth,
+            cookies=self.cookies,
+            hooks=self.hooks,
+        )
+        return p
+
+
+class PreparedRequest(RequestEncodingMixin, RequestHooksMixin):
+    """The fully mutable :class:`PreparedRequest <PreparedRequest>` object,
+    containing the exact bytes that will be sent to the server.
+
+    Instances are generated from a :class:`Request <Request>` object, and
+    should not be instantiated manually; doing so may produce undesirable
+    effects.
+
+    Usage::
+
+      >>> import requests
+      >>> req = requests.Request('GET', 'https://httpbin.org/get')
+      >>> r = req.prepare()
+      >>> r
+      <PreparedRequest [GET]>
+
+      >>> s = requests.Session()
+      >>> s.send(r)
+      <Response [200]>
+    """
+
+    def __init__(self):
+        #: HTTP verb to send to the server.
+        self.method = None
+        #: HTTP URL to send the request to.
+        self.url = None
+        #: dictionary of HTTP headers.
+        self.headers = None
+        # The `CookieJar` used to create the Cookie header will be stored here
+        # after prepare_cookies is called
+        self._cookies = None
+        #: request body to send to the server.
+        self.body = None
+        #: dictionary of callback hooks, for internal usage.
+        self.hooks = default_hooks()
+        #: integer denoting starting position of a readable file-like body.
+        self._body_position = None
+
+    def prepare(
+        self,
+        method=None,
+        url=None,
+        headers=None,
+        files=None,
+        data=None,
+        params=None,
+        auth=None,
+        cookies=None,
+        hooks=None,
+        json=None,
+    ):
+        """Prepares the entire request with the given parameters."""
+
+        self.prepare_method(method)
+        self.prepare_url(url, params)
+        self.prepare_headers(headers)
+        self.prepare_cookies(cookies)
+        self.prepare_body(data, files, json)
+        self.prepare_auth(auth, url)
+
+        # Note that prepare_auth must be last to enable authentication schemes
+        # such as OAuth to work on a fully prepared request.
+
+        # This MUST go after prepare_auth. Authenticators could add a hook
+        self.prepare_hooks(hooks)
+
+    def __repr__(self):
+        return f"<PreparedRequest [{self.method}]>"
+
+    def copy(self):
+        p = PreparedRequest()
+        p.method = self.method
+        p.url = self.url
+        p.headers = self.headers.copy() if self.headers is not None else None
+        p._cookies = _copy_cookie_jar(self._cookies)
+        p.body = self.body
+        p.hooks = self.hooks
+        p._body_position = self._body_position
+        return p
+
+    def prepare_method(self, method):
+        """Prepares the given HTTP method."""
+        self.method = method
+        if self.method is not None:
+            self.method = to_native_string(self.method.upper())
+
+    @staticmethod
+    def _get_idna_encoded_host(host):
+        import idna
+
+        try:
+            host = idna.encode(host, uts46=True).decode("utf-8")
+        except idna.IDNAError:
+            raise UnicodeError
+        return host
+
+    def prepare_url(self, url, params):
+        """Prepares the given HTTP URL."""
+        #: Accept objects that have string representations.
+        #: We're unable to blindly call unicode/str functions
+        #: as this will include the bytestring indicator (b'')
+        #: on python 3.x.
+        #: https://github.com/psf/requests/pull/2238
+        if isinstance(url, bytes):
+            url = url.decode("utf8")
+        else:
+            url = str(url)
+
+        # Remove leading whitespaces from url
+        url = url.lstrip()
+
+        # Don't do any URL preparation for non-HTTP schemes like `mailto`,
+        # `data` etc to work around exceptions from `url_parse`, which
+        # handles RFC 3986 only.
+        if ":" in url and not url.lower().startswith("http"):
+            self.url = url
+            return
+
+        # Support for unicode domain names and paths.
+        try:
+            scheme, auth, host, port, path, query, fragment = parse_url(url)
+        except LocationParseError as e:
+            raise InvalidURL(*e.args)
+
+        if not scheme:
+            raise MissingSchema(
+                f"Invalid URL {url!r}: No scheme supplied. "
+                f"Perhaps you meant https://{url}?"
+            )
+
+        if not host:
+            raise InvalidURL(f"Invalid URL {url!r}: No host supplied")
+
+        # In general, we want to try IDNA encoding the hostname if the string contains
+        # non-ASCII characters. This allows users to automatically get the correct IDNA
+        # behaviour. For strings containing only ASCII characters, we need to also verify
+        # it doesn't start with a wildcard (*), before allowing the unencoded hostname.
+        if not unicode_is_ascii(host):
+            try:
+                host = self._get_idna_encoded_host(host)
+            except UnicodeError:
+                raise InvalidURL("URL has an invalid label.")
+        elif host.startswith(("*", ".")):
+            raise InvalidURL("URL has an invalid label.")
+
+        # Carefully reconstruct the network location
+        netloc = auth or ""
+        if netloc:
+            netloc += "@"
+        netloc += host
+        if port:
+            netloc += f":{port}"
+
+        # Bare domains aren't valid URLs.
+        if not path:
+            path = "/"
+
+        if isinstance(params, (str, bytes)):
+            params = to_native_string(params)
+
+        enc_params = self._encode_params(params)
+        if enc_params:
+            if query:
+                query = f"{query}&{enc_params}"
+            else:
+                query = enc_params
+
+        url = requote_uri(urlunparse([scheme, netloc, path, None, query, fragment]))
+        self.url = url
+
+    def prepare_headers(self, headers):
+        """Prepares the given HTTP headers."""
+
+        self.headers = CaseInsensitiveDict()
+        if headers:
+            for header in headers.items():
+                # Raise exception on invalid header value.
+                check_header_validity(header)
+                name, value = header
+                self.headers[to_native_string(name)] = value
+
+    def prepare_body(self, data, files, json=None):
+        """Prepares the given HTTP body data."""
+
+        # Check if file, fo, generator, iterator.
+        # If not, run through normal process.
+
+        # Nottin' on you.
+        body = None
+        content_type = None
+
+        if not data and json is not None:
+            # urllib3 requires a bytes-like body. Python 2's json.dumps
+            # provides this natively, but Python 3 gives a Unicode string.
+            content_type = "application/json"
+
+            try:
+                body = complexjson.dumps(json, allow_nan=False)
+            except ValueError as ve:
+                raise InvalidJSONError(ve, request=self)
+
+            if not isinstance(body, bytes):
+                body = body.encode("utf-8")
+
+        is_stream = all(
+            [
+                hasattr(data, "__iter__"),
+                not isinstance(data, (basestring, list, tuple, Mapping)),
+            ]
+        )
+
+        if is_stream:
+            try:
+                length = super_len(data)
+            except (TypeError, AttributeError, UnsupportedOperation):
+                length = None
+
+            body = data
+
+            if getattr(body, "tell", None) is not None:
+                # Record the current file position before reading.
+                # This will allow us to rewind a file in the event
+                # of a redirect.
+                try:
+                    self._body_position = body.tell()
+                except OSError:
+                    # This differentiates from None, allowing us to catch
+                    # a failed `tell()` later when trying to rewind the body
+                    self._body_position = object()
+
+            if files:
+                raise NotImplementedError(
+                    "Streamed bodies and files are mutually exclusive."
+                )
+
+            if length:
+                self.headers["Content-Length"] = builtin_str(length)
+            else:
+                self.headers["Transfer-Encoding"] = "chunked"
+        else:
+            # Multi-part file uploads.
+            if files:
+                (body, content_type) = self._encode_files(files, data)
+            else:
+                if data:
+                    body = self._encode_params(data)
+                    if isinstance(data, basestring) or hasattr(data, "read"):
+                        content_type = None
+                    else:
+                        content_type = "application/x-www-form-urlencoded"
+
+            self.prepare_content_length(body)
+
+            # Add content-type if it wasn't explicitly provided.
+            if content_type and ("content-type" not in self.headers):
+                self.headers["Content-Type"] = content_type
+
+        self.body = body
+
+    def prepare_content_length(self, body):
+        """Prepare Content-Length header based on request method and body"""
+        if body is not None:
+            length = super_len(body)
+            if length:
+                # If length exists, set it. Otherwise, we fallback
+                # to Transfer-Encoding: chunked.
+                self.headers["Content-Length"] = builtin_str(length)
+        elif (
+            self.method not in ("GET", "HEAD")
+            and self.headers.get("Content-Length") is None
+        ):
+            # Set Content-Length to 0 for methods that can have a body
+            # but don't provide one. (i.e. not GET or HEAD)
+            self.headers["Content-Length"] = "0"
+
+    def prepare_auth(self, auth, url=""):
+        """Prepares the given HTTP auth data."""
+
+        # If no Auth is explicitly provided, extract it from the URL first.
+        if auth is None:
+            url_auth = get_auth_from_url(self.url)
+            auth = url_auth if any(url_auth) else None
+
+        if auth:
+            if isinstance(auth, tuple) and len(auth) == 2:
+                # special-case basic HTTP auth
+                auth = HTTPBasicAuth(*auth)
+
+            # Allow auth to make its changes.
+            r = auth(self)
+
+            # Update self to reflect the auth changes.
+            self.__dict__.update(r.__dict__)
+
+            # Recompute Content-Length
+            self.prepare_content_length(self.body)
+
+    def prepare_cookies(self, cookies):
+        """Prepares the given HTTP cookie data.
+
+        This function eventually generates a ``Cookie`` header from the
+        given cookies using cookielib. Due to cookielib's design, the header
+        will not be regenerated if it already exists, meaning this function
+        can only be called once for the life of the
+        :class:`PreparedRequest <PreparedRequest>` object. Any subsequent calls
+        to ``prepare_cookies`` will have no actual effect, unless the "Cookie"
+        header is removed beforehand.
+        """
+        if isinstance(cookies, cookielib.CookieJar):
+            self._cookies = cookies
+        else:
+            self._cookies = cookiejar_from_dict(cookies)
+
+        cookie_header = get_cookie_header(self._cookies, self)
+        if cookie_header is not None:
+            self.headers["Cookie"] = cookie_header
+
+    def prepare_hooks(self, hooks):
+        """Prepares the given hooks."""
+        # hooks can be passed as None to the prepare method and to this
+        # method. To prevent iterating over None, simply use an empty list
+        # if hooks is False-y
+        hooks = hooks or []
+        for event in hooks:
+            self.register_hook(event, hooks[event])
+
+
+class Response:
+    """The :class:`Response <Response>` object, which contains a
+    server's response to an HTTP request.
+    """
+
+    __attrs__ = [
+        "_content",
+        "status_code",
+        "headers",
+        "url",
+        "history",
+        "encoding",
+        "reason",
+        "cookies",
+        "elapsed",
+        "request",
+    ]
+
+    def __init__(self):
+        self._content = False
+        self._content_consumed = False
+        self._next = None
+
+        #: Integer Code of responded HTTP Status, e.g. 404 or 200.
+        self.status_code = None
+
+        #: Case-insensitive Dictionary of Response Headers.
+        #: For example, ``headers['content-encoding']`` will return the
+        #: value of a ``'Content-Encoding'`` response header.
+        self.headers = CaseInsensitiveDict()
+
+        #: File-like object representation of response (for advanced usage).
+        #: Use of ``raw`` requires that ``stream=True`` be set on the request.
+        #: This requirement does not apply for use internally to Requests.
+        self.raw = None
+
+        #: Final URL location of Response.
+        self.url = None
+
+        #: Encoding to decode with when accessing r.text.
+        self.encoding = None
+
+        #: A list of :class:`Response <Response>` objects from
+        #: the history of the Request. Any redirect responses will end
+        #: up here. The list is sorted from the oldest to the most recent request.
+        self.history = []
+
+        #: Textual reason of responded HTTP Status, e.g. "Not Found" or "OK".
+        self.reason = None
+
+        #: A CookieJar of Cookies the server sent back.
+        self.cookies = cookiejar_from_dict({})
+
+        #: The amount of time elapsed between sending the request
+        #: and the arrival of the response (as a timedelta).
+        #: This property specifically measures the time taken between sending
+        #: the first byte of the request and finishing parsing the headers. It
+        #: is therefore unaffected by consuming the response content or the
+        #: value of the ``stream`` keyword argument.
+        self.elapsed = datetime.timedelta(0)
+
+        #: The :class:`PreparedRequest <PreparedRequest>` object to which this
+        #: is a response.
+        self.request = None
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *args):
+        self.close()
+
+    def __getstate__(self):
+        # Consume everything; accessing the content attribute makes
+        # sure the content has been fully read.
+        if not self._content_consumed:
+            self.content
+
+        return {attr: getattr(self, attr, None) for attr in self.__attrs__}
+
+    def __setstate__(self, state):
+        for name, value in state.items():
+            setattr(self, name, value)
+
+        # pickled objects do not have .raw
+        setattr(self, "_content_consumed", True)
+        setattr(self, "raw", None)
+
+    def __repr__(self):
+        return f"<Response [{self.status_code}]>"
+
+    def __bool__(self):
+        """Returns True if :attr:`status_code` is less than 400.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code, is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        return self.ok
+
+    def __nonzero__(self):
+        """Returns True if :attr:`status_code` is less than 400.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code, is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        return self.ok
+
+    def __iter__(self):
+        """Allows you to use a response as an iterator."""
+        return self.iter_content(128)
+
+    @property
+    def ok(self):
+        """Returns True if :attr:`status_code` is less than 400, False if not.
+
+        This attribute checks if the status code of the response is between
+        400 and 600 to see if there was a client error or a server error. If
+        the status code is between 200 and 400, this will return True. This
+        is **not** a check to see if the response code is ``200 OK``.
+        """
+        try:
+            self.raise_for_status()
+        except HTTPError:
+            return False
+        return True
+
+    @property
+    def is_redirect(self):
+        """True if this Response is a well-formed HTTP redirect that could have
+        been processed automatically (by :meth:`Session.resolve_redirects`).
+        """
+        return "location" in self.headers and self.status_code in REDIRECT_STATI
+
+    @property
+    def is_permanent_redirect(self):
+        """True if this Response one of the permanent versions of redirect."""
+        return "location" in self.headers and self.status_code in (
+            codes.moved_permanently,
+            codes.permanent_redirect,
+        )
+
+    @property
+    def next(self):
+        """Returns a PreparedRequest for the next request in a redirect chain, if there is one."""
+        return self._next
+
+    @property
+    def apparent_encoding(self):
+        """The apparent encoding, provided by the charset_normalizer or chardet libraries."""
+        if chardet is not None:
+            return chardet.detect(self.content)["encoding"]
+        else:
+            # If no character detection library is available, we'll fall back
+            # to a standard Python utf-8 str.
+            return "utf-8"
+
+    def iter_content(self, chunk_size=1, decode_unicode=False):
+        """Iterates over the response data.  When stream=True is set on the
+        request, this avoids reading the content at once into memory for
+        large responses.  The chunk size is the number of bytes it should
+        read into memory.  This is not necessarily the length of each item
+        returned as decoding can take place.
+
+        chunk_size must be of type int or None. A value of None will
+        function differently depending on the value of `stream`.
+        stream=True will read data as it arrives in whatever size the
+        chunks are received. If stream=False, data is returned as
+        a single chunk.
+
+        If decode_unicode is True, content will be decoded using the best
+        available encoding based on the response.
+        """
+
+        def generate():
+            # Special case for urllib3.
+            if hasattr(self.raw, "stream"):
+                try:
+                    yield from self.raw.stream(chunk_size, decode_content=True)
+                except ProtocolError as e:
+                    raise ChunkedEncodingError(e)
+                except DecodeError as e:
+                    raise ContentDecodingError(e)
+                except ReadTimeoutError as e:
+                    raise ConnectionError(e)
+                except SSLError as e:
+                    raise RequestsSSLError(e)
+            else:
+                # Standard file-like object.
+                while True:
+                    chunk = self.raw.read(chunk_size)
+                    if not chunk:
+                        break
+                    yield chunk
+
+            self._content_consumed = True
+
+        if self._content_consumed and isinstance(self._content, bool):
+            raise StreamConsumedError()
+        elif chunk_size is not None and not isinstance(chunk_size, int):
+            raise TypeError(
+                f"chunk_size must be an int, it is instead a {type(chunk_size)}."
+            )
+        # simulate reading small chunks of the content
+        reused_chunks = iter_slices(self._content, chunk_size)
+
+        stream_chunks = generate()
+
+        chunks = reused_chunks if self._content_consumed else stream_chunks
+
+        if decode_unicode:
+            chunks = stream_decode_response_unicode(chunks, self)
+
+        return chunks
+
+    def iter_lines(
+        self, chunk_size=ITER_CHUNK_SIZE, decode_unicode=False, delimiter=None
+    ):
+        """Iterates over the response data, one line at a time.  When
+        stream=True is set on the request, this avoids reading the
+        content at once into memory for large responses.
+
+        .. note:: This method is not reentrant safe.
+        """
+
+        pending = None
+
+        for chunk in self.iter_content(
+            chunk_size=chunk_size, decode_unicode=decode_unicode
+        ):
+            if pending is not None:
+                chunk = pending + chunk
+
+            if delimiter:
+                lines = chunk.split(delimiter)
+            else:
+                lines = chunk.splitlines()
+
+            if lines and lines[-1] and chunk and lines[-1][-1] == chunk[-1]:
+                pending = lines.pop()
+            else:
+                pending = None
+
+            yield from lines
+
+        if pending is not None:
+            yield pending
+
+    @property
+    def content(self):
+        """Content of the response, in bytes."""
+
+        if self._content is False:
+            # Read the contents.
+            if self._content_consumed:
+                raise RuntimeError("The content for this response was already consumed")
+
+            if self.status_code == 0 or self.raw is None:
+                self._content = None
+            else:
+                self._content = b"".join(self.iter_content(CONTENT_CHUNK_SIZE)) or b""
+
+        self._content_consumed = True
+        # don't need to release the connection; that's been handled by urllib3
+        # since we exhausted the data.
+        return self._content
+
+    @property
+    def text(self):
+        """Content of the response, in unicode.
+
+        If Response.encoding is None, encoding will be guessed using
+        ``charset_normalizer`` or ``chardet``.
+
+        The encoding of the response content is determined based solely on HTTP
+        headers, following RFC 2616 to the letter. If you can take advantage of
+        non-HTTP knowledge to make a better guess at the encoding, you should
+        set ``r.encoding`` appropriately before accessing this property.
+        """
+
+        # Try charset from content-type
+        content = None
+        encoding = self.encoding
+
+        if not self.content:
+            return ""
+
+        # Fallback to auto-detected encoding.
+        if self.encoding is None:
+            encoding = self.apparent_encoding
+
+        # Decode unicode from given encoding.
+        try:
+            content = str(self.content, encoding, errors="replace")
+        except (LookupError, TypeError):
+            # A LookupError is raised if the encoding was not found which could
+            # indicate a misspelling or similar mistake.
+            #
+            # A TypeError can be raised if encoding is None
+            #
+            # So we try blindly encoding.
+            content = str(self.content, errors="replace")
+
+        return content
+
+    def json(self, **kwargs):
+        r"""Returns the json-encoded content of a response, if any.
+
+        :param \*\*kwargs: Optional arguments that ``json.loads`` takes.
+        :raises requests.exceptions.JSONDecodeError: If the response body does not
+            contain valid json.
+        """
+
+        if not self.encoding and self.content and len(self.content) > 3:
+            # No encoding set. JSON RFC 4627 section 3 states we should expect
+            # UTF-8, -16 or -32. Detect which one to use; If the detection or
+            # decoding fails, fall back to `self.text` (using charset_normalizer to make
+            # a best guess).
+            encoding = guess_json_utf(self.content)
+            if encoding is not None:
+                try:
+                    return complexjson.loads(self.content.decode(encoding), **kwargs)
+                except UnicodeDecodeError:
+                    # Wrong UTF codec detected; usually because it's not UTF-8
+                    # but some other 8-bit codec.  This is an RFC violation,
+                    # and the server didn't bother to tell us what codec *was*
+                    # used.
+                    pass
+                except JSONDecodeError as e:
+                    raise RequestsJSONDecodeError(e.msg, e.doc, e.pos)
+
+        try:
+            return complexjson.loads(self.text, **kwargs)
+        except JSONDecodeError as e:
+            # Catch JSON-related errors and raise as requests.JSONDecodeError
+            # This aliases json.JSONDecodeError and simplejson.JSONDecodeError
+            raise RequestsJSONDecodeError(e.msg, e.doc, e.pos)
+
+    @property
+    def links(self):
+        """Returns the parsed header links of the response, if any."""
+
+        header = self.headers.get("link")
+
+        resolved_links = {}
+
+        if header:
+            links = parse_header_links(header)
+
+            for link in links:
+                key = link.get("rel") or link.get("url")
+                resolved_links[key] = link
+
+        return resolved_links
+
+    def raise_for_status(self):
+        """Raises :class:`HTTPError`, if one occurred."""
+
+        http_error_msg = ""
+        if isinstance(self.reason, bytes):
+            # We attempt to decode utf-8 first because some servers
+            # choose to localize their reason strings. If the string
+            # isn't utf-8, we fall back to iso-8859-1 for all other
+            # encodings. (See PR #3538)
+            try:
+                reason = self.reason.decode("utf-8")
+            except UnicodeDecodeError:
+                reason = self.reason.decode("iso-8859-1")
+        else:
+            reason = self.reason
+
+        if 400 <= self.status_code < 500:
+            http_error_msg = (
+                f"{self.status_code} Client Error: {reason} for url: {self.url}"
+            )
+
+        elif 500 <= self.status_code < 600:
+            http_error_msg = (
+                f"{self.status_code} Server Error: {reason} for url: {self.url}"
+            )
+
+        if http_error_msg:
+            raise HTTPError(http_error_msg, response=self)
+
+    def close(self):
+        """Releases the connection back to the pool. Once this method has been
+        called the underlying ``raw`` object must not be accessed again.
+
+        *Note: Should not normally need to be called explicitly.*
+        """
+        if not self._content_consumed:
+            self.raw.close()
+
+        release_conn = getattr(self.raw, "release_conn", None)
+        if release_conn is not None:
+            release_conn()
diff --git a/backend/venv/lib/python3.12/site-packages/requests/packages.py b/backend/venv/lib/python3.12/site-packages/requests/packages.py
new file mode 100644
index 00000000..5ab3d8e2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/packages.py
@@ -0,0 +1,23 @@
+import sys
+
+from .compat import chardet
+
+# This code exists for backwards compatibility reasons.
+# I don't like it either. Just look the other way. :)
+
+for package in ("urllib3", "idna"):
+    locals()[package] = __import__(package)
+    # This traversal is apparently necessary such that the identities are
+    # preserved (requests.packages.urllib3.* is urllib3.*)
+    for mod in list(sys.modules):
+        if mod == package or mod.startswith(f"{package}."):
+            sys.modules[f"requests.packages.{mod}"] = sys.modules[mod]
+
+if chardet is not None:
+    target = chardet.__name__
+    for mod in list(sys.modules):
+        if mod == target or mod.startswith(f"{target}."):
+            imported_mod = sys.modules[mod]
+            sys.modules[f"requests.packages.{mod}"] = imported_mod
+            mod = mod.replace(target, "chardet")
+            sys.modules[f"requests.packages.{mod}"] = imported_mod
diff --git a/backend/venv/lib/python3.12/site-packages/requests/sessions.py b/backend/venv/lib/python3.12/site-packages/requests/sessions.py
new file mode 100644
index 00000000..b387bc36
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/sessions.py
@@ -0,0 +1,831 @@
+"""
+requests.sessions
+~~~~~~~~~~~~~~~~~
+
+This module provides a Session object to manage and persist settings across
+requests (cookies, auth, proxies).
+"""
+import os
+import sys
+import time
+from collections import OrderedDict
+from datetime import timedelta
+
+from ._internal_utils import to_native_string
+from .adapters import HTTPAdapter
+from .auth import _basic_auth_str
+from .compat import Mapping, cookielib, urljoin, urlparse
+from .cookies import (
+    RequestsCookieJar,
+    cookiejar_from_dict,
+    extract_cookies_to_jar,
+    merge_cookies,
+)
+from .exceptions import (
+    ChunkedEncodingError,
+    ContentDecodingError,
+    InvalidSchema,
+    TooManyRedirects,
+)
+from .hooks import default_hooks, dispatch_hook
+
+# formerly defined here, reexposed here for backward compatibility
+from .models import (  # noqa: F401
+    DEFAULT_REDIRECT_LIMIT,
+    REDIRECT_STATI,
+    PreparedRequest,
+    Request,
+)
+from .status_codes import codes
+from .structures import CaseInsensitiveDict
+from .utils import (  # noqa: F401
+    DEFAULT_PORTS,
+    default_headers,
+    get_auth_from_url,
+    get_environ_proxies,
+    get_netrc_auth,
+    requote_uri,
+    resolve_proxies,
+    rewind_body,
+    should_bypass_proxies,
+    to_key_val_list,
+)
+
+# Preferred clock, based on which one is more accurate on a given system.
+if sys.platform == "win32":
+    preferred_clock = time.perf_counter
+else:
+    preferred_clock = time.time
+
+
+def merge_setting(request_setting, session_setting, dict_class=OrderedDict):
+    """Determines appropriate setting for a given request, taking into account
+    the explicit setting on that request, and the setting in the session. If a
+    setting is a dictionary, they will be merged together using `dict_class`
+    """
+
+    if session_setting is None:
+        return request_setting
+
+    if request_setting is None:
+        return session_setting
+
+    # Bypass if not a dictionary (e.g. verify)
+    if not (
+        isinstance(session_setting, Mapping) and isinstance(request_setting, Mapping)
+    ):
+        return request_setting
+
+    merged_setting = dict_class(to_key_val_list(session_setting))
+    merged_setting.update(to_key_val_list(request_setting))
+
+    # Remove keys that are set to None. Extract keys first to avoid altering
+    # the dictionary during iteration.
+    none_keys = [k for (k, v) in merged_setting.items() if v is None]
+    for key in none_keys:
+        del merged_setting[key]
+
+    return merged_setting
+
+
+def merge_hooks(request_hooks, session_hooks, dict_class=OrderedDict):
+    """Properly merges both requests and session hooks.
+
+    This is necessary because when request_hooks == {'response': []}, the
+    merge breaks Session hooks entirely.
+    """
+    if session_hooks is None or session_hooks.get("response") == []:
+        return request_hooks
+
+    if request_hooks is None or request_hooks.get("response") == []:
+        return session_hooks
+
+    return merge_setting(request_hooks, session_hooks, dict_class)
+
+
+class SessionRedirectMixin:
+    def get_redirect_target(self, resp):
+        """Receives a Response. Returns a redirect URI or ``None``"""
+        # Due to the nature of how requests processes redirects this method will
+        # be called at least once upon the original response and at least twice
+        # on each subsequent redirect response (if any).
+        # If a custom mixin is used to handle this logic, it may be advantageous
+        # to cache the redirect location onto the response object as a private
+        # attribute.
+        if resp.is_redirect:
+            location = resp.headers["location"]
+            # Currently the underlying http module on py3 decode headers
+            # in latin1, but empirical evidence suggests that latin1 is very
+            # rarely used with non-ASCII characters in HTTP headers.
+            # It is more likely to get UTF8 header rather than latin1.
+            # This causes incorrect handling of UTF8 encoded location headers.
+            # To solve this, we re-encode the location in latin1.
+            location = location.encode("latin1")
+            return to_native_string(location, "utf8")
+        return None
+
+    def should_strip_auth(self, old_url, new_url):
+        """Decide whether Authorization header should be removed when redirecting"""
+        old_parsed = urlparse(old_url)
+        new_parsed = urlparse(new_url)
+        if old_parsed.hostname != new_parsed.hostname:
+            return True
+        # Special case: allow http -> https redirect when using the standard
+        # ports. This isn't specified by RFC 7235, but is kept to avoid
+        # breaking backwards compatibility with older versions of requests
+        # that allowed any redirects on the same host.
+        if (
+            old_parsed.scheme == "http"
+            and old_parsed.port in (80, None)
+            and new_parsed.scheme == "https"
+            and new_parsed.port in (443, None)
+        ):
+            return False
+
+        # Handle default port usage corresponding to scheme.
+        changed_port = old_parsed.port != new_parsed.port
+        changed_scheme = old_parsed.scheme != new_parsed.scheme
+        default_port = (DEFAULT_PORTS.get(old_parsed.scheme, None), None)
+        if (
+            not changed_scheme
+            and old_parsed.port in default_port
+            and new_parsed.port in default_port
+        ):
+            return False
+
+        # Standard case: root URI must match
+        return changed_port or changed_scheme
+
+    def resolve_redirects(
+        self,
+        resp,
+        req,
+        stream=False,
+        timeout=None,
+        verify=True,
+        cert=None,
+        proxies=None,
+        yield_requests=False,
+        **adapter_kwargs,
+    ):
+        """Receives a Response. Returns a generator of Responses or Requests."""
+
+        hist = []  # keep track of history
+
+        url = self.get_redirect_target(resp)
+        previous_fragment = urlparse(req.url).fragment
+        while url:
+            prepared_request = req.copy()
+
+            # Update history and keep track of redirects.
+            # resp.history must ignore the original request in this loop
+            hist.append(resp)
+            resp.history = hist[1:]
+
+            try:
+                resp.content  # Consume socket so it can be released
+            except (ChunkedEncodingError, ContentDecodingError, RuntimeError):
+                resp.raw.read(decode_content=False)
+
+            if len(resp.history) >= self.max_redirects:
+                raise TooManyRedirects(
+                    f"Exceeded {self.max_redirects} redirects.", response=resp
+                )
+
+            # Release the connection back into the pool.
+            resp.close()
+
+            # Handle redirection without scheme (see: RFC 1808 Section 4)
+            if url.startswith("//"):
+                parsed_rurl = urlparse(resp.url)
+                url = ":".join([to_native_string(parsed_rurl.scheme), url])
+
+            # Normalize url case and attach previous fragment if needed (RFC 7231 7.1.2)
+            parsed = urlparse(url)
+            if parsed.fragment == "" and previous_fragment:
+                parsed = parsed._replace(fragment=previous_fragment)
+            elif parsed.fragment:
+                previous_fragment = parsed.fragment
+            url = parsed.geturl()
+
+            # Facilitate relative 'location' headers, as allowed by RFC 7231.
+            # (e.g. '/path/to/resource' instead of 'http://domain.tld/path/to/resource')
+            # Compliant with RFC3986, we percent encode the url.
+            if not parsed.netloc:
+                url = urljoin(resp.url, requote_uri(url))
+            else:
+                url = requote_uri(url)
+
+            prepared_request.url = to_native_string(url)
+
+            self.rebuild_method(prepared_request, resp)
+
+            # https://github.com/psf/requests/issues/1084
+            if resp.status_code not in (
+                codes.temporary_redirect,
+                codes.permanent_redirect,
+            ):
+                # https://github.com/psf/requests/issues/3490
+                purged_headers = ("Content-Length", "Content-Type", "Transfer-Encoding")
+                for header in purged_headers:
+                    prepared_request.headers.pop(header, None)
+                prepared_request.body = None
+
+            headers = prepared_request.headers
+            headers.pop("Cookie", None)
+
+            # Extract any cookies sent on the response to the cookiejar
+            # in the new request. Because we've mutated our copied prepared
+            # request, use the old one that we haven't yet touched.
+            extract_cookies_to_jar(prepared_request._cookies, req, resp.raw)
+            merge_cookies(prepared_request._cookies, self.cookies)
+            prepared_request.prepare_cookies(prepared_request._cookies)
+
+            # Rebuild auth and proxy information.
+            proxies = self.rebuild_proxies(prepared_request, proxies)
+            self.rebuild_auth(prepared_request, resp)
+
+            # A failed tell() sets `_body_position` to `object()`. This non-None
+            # value ensures `rewindable` will be True, allowing us to raise an
+            # UnrewindableBodyError, instead of hanging the connection.
+            rewindable = prepared_request._body_position is not None and (
+                "Content-Length" in headers or "Transfer-Encoding" in headers
+            )
+
+            # Attempt to rewind consumed file-like object.
+            if rewindable:
+                rewind_body(prepared_request)
+
+            # Override the original request.
+            req = prepared_request
+
+            if yield_requests:
+                yield req
+            else:
+                resp = self.send(
+                    req,
+                    stream=stream,
+                    timeout=timeout,
+                    verify=verify,
+                    cert=cert,
+                    proxies=proxies,
+                    allow_redirects=False,
+                    **adapter_kwargs,
+                )
+
+                extract_cookies_to_jar(self.cookies, prepared_request, resp.raw)
+
+                # extract redirect url, if any, for the next loop
+                url = self.get_redirect_target(resp)
+                yield resp
+
+    def rebuild_auth(self, prepared_request, response):
+        """When being redirected we may want to strip authentication from the
+        request to avoid leaking credentials. This method intelligently removes
+        and reapplies authentication where possible to avoid credential loss.
+        """
+        headers = prepared_request.headers
+        url = prepared_request.url
+
+        if "Authorization" in headers and self.should_strip_auth(
+            response.request.url, url
+        ):
+            # If we get redirected to a new host, we should strip out any
+            # authentication headers.
+            del headers["Authorization"]
+
+        # .netrc might have more auth for us on our new host.
+        new_auth = get_netrc_auth(url) if self.trust_env else None
+        if new_auth is not None:
+            prepared_request.prepare_auth(new_auth)
+
+    def rebuild_proxies(self, prepared_request, proxies):
+        """This method re-evaluates the proxy configuration by considering the
+        environment variables. If we are redirected to a URL covered by
+        NO_PROXY, we strip the proxy configuration. Otherwise, we set missing
+        proxy keys for this URL (in case they were stripped by a previous
+        redirect).
+
+        This method also replaces the Proxy-Authorization header where
+        necessary.
+
+        :rtype: dict
+        """
+        headers = prepared_request.headers
+        scheme = urlparse(prepared_request.url).scheme
+        new_proxies = resolve_proxies(prepared_request, proxies, self.trust_env)
+
+        if "Proxy-Authorization" in headers:
+            del headers["Proxy-Authorization"]
+
+        try:
+            username, password = get_auth_from_url(new_proxies[scheme])
+        except KeyError:
+            username, password = None, None
+
+        # urllib3 handles proxy authorization for us in the standard adapter.
+        # Avoid appending this to TLS tunneled requests where it may be leaked.
+        if not scheme.startswith("https") and username and password:
+            headers["Proxy-Authorization"] = _basic_auth_str(username, password)
+
+        return new_proxies
+
+    def rebuild_method(self, prepared_request, response):
+        """When being redirected we may want to change the method of the request
+        based on certain specs or browser behavior.
+        """
+        method = prepared_request.method
+
+        # https://tools.ietf.org/html/rfc7231#section-6.4.4
+        if response.status_code == codes.see_other and method != "HEAD":
+            method = "GET"
+
+        # Do what the browsers do, despite standards...
+        # First, turn 302s into GETs.
+        if response.status_code == codes.found and method != "HEAD":
+            method = "GET"
+
+        # Second, if a POST is responded to with a 301, turn it into a GET.
+        # This bizarre behaviour is explained in Issue 1704.
+        if response.status_code == codes.moved and method == "POST":
+            method = "GET"
+
+        prepared_request.method = method
+
+
+class Session(SessionRedirectMixin):
+    """A Requests session.
+
+    Provides cookie persistence, connection-pooling, and configuration.
+
+    Basic Usage::
+
+      >>> import requests
+      >>> s = requests.Session()
+      >>> s.get('https://httpbin.org/get')
+      <Response [200]>
+
+    Or as a context manager::
+
+      >>> with requests.Session() as s:
+      ...     s.get('https://httpbin.org/get')
+      <Response [200]>
+    """
+
+    __attrs__ = [
+        "headers",
+        "cookies",
+        "auth",
+        "proxies",
+        "hooks",
+        "params",
+        "verify",
+        "cert",
+        "adapters",
+        "stream",
+        "trust_env",
+        "max_redirects",
+    ]
+
+    def __init__(self):
+        #: A case-insensitive dictionary of headers to be sent on each
+        #: :class:`Request <Request>` sent from this
+        #: :class:`Session <Session>`.
+        self.headers = default_headers()
+
+        #: Default Authentication tuple or object to attach to
+        #: :class:`Request <Request>`.
+        self.auth = None
+
+        #: Dictionary mapping protocol or protocol and host to the URL of the proxy
+        #: (e.g. {'http': 'foo.bar:3128', 'http://host.name': 'foo.bar:4012'}) to
+        #: be used on each :class:`Request <Request>`.
+        self.proxies = {}
+
+        #: Event-handling hooks.
+        self.hooks = default_hooks()
+
+        #: Dictionary of querystring data to attach to each
+        #: :class:`Request <Request>`. The dictionary values may be lists for
+        #: representing multivalued query parameters.
+        self.params = {}
+
+        #: Stream response content default.
+        self.stream = False
+
+        #: SSL Verification default.
+        #: Defaults to `True`, requiring requests to verify the TLS certificate at the
+        #: remote end.
+        #: If verify is set to `False`, requests will accept any TLS certificate
+        #: presented by the server, and will ignore hostname mismatches and/or
+        #: expired certificates, which will make your application vulnerable to
+        #: man-in-the-middle (MitM) attacks.
+        #: Only set this to `False` for testing.
+        self.verify = True
+
+        #: SSL client certificate default, if String, path to ssl client
+        #: cert file (.pem). If Tuple, ('cert', 'key') pair.
+        self.cert = None
+
+        #: Maximum number of redirects allowed. If the request exceeds this
+        #: limit, a :class:`TooManyRedirects` exception is raised.
+        #: This defaults to requests.models.DEFAULT_REDIRECT_LIMIT, which is
+        #: 30.
+        self.max_redirects = DEFAULT_REDIRECT_LIMIT
+
+        #: Trust environment settings for proxy configuration, default
+        #: authentication and similar.
+        self.trust_env = True
+
+        #: A CookieJar containing all currently outstanding cookies set on this
+        #: session. By default it is a
+        #: :class:`RequestsCookieJar <requests.cookies.RequestsCookieJar>`, but
+        #: may be any other ``cookielib.CookieJar`` compatible object.
+        self.cookies = cookiejar_from_dict({})
+
+        # Default connection adapters.
+        self.adapters = OrderedDict()
+        self.mount("https://", HTTPAdapter())
+        self.mount("http://", HTTPAdapter())
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *args):
+        self.close()
+
+    def prepare_request(self, request):
+        """Constructs a :class:`PreparedRequest <PreparedRequest>` for
+        transmission and returns it. The :class:`PreparedRequest` has settings
+        merged from the :class:`Request <Request>` instance and those of the
+        :class:`Session`.
+
+        :param request: :class:`Request` instance to prepare with this
+            session's settings.
+        :rtype: requests.PreparedRequest
+        """
+        cookies = request.cookies or {}
+
+        # Bootstrap CookieJar.
+        if not isinstance(cookies, cookielib.CookieJar):
+            cookies = cookiejar_from_dict(cookies)
+
+        # Merge with session cookies
+        merged_cookies = merge_cookies(
+            merge_cookies(RequestsCookieJar(), self.cookies), cookies
+        )
+
+        # Set environment's basic authentication if not explicitly set.
+        auth = request.auth
+        if self.trust_env and not auth and not self.auth:
+            auth = get_netrc_auth(request.url)
+
+        p = PreparedRequest()
+        p.prepare(
+            method=request.method.upper(),
+            url=request.url,
+            files=request.files,
+            data=request.data,
+            json=request.json,
+            headers=merge_setting(
+                request.headers, self.headers, dict_class=CaseInsensitiveDict
+            ),
+            params=merge_setting(request.params, self.params),
+            auth=merge_setting(auth, self.auth),
+            cookies=merged_cookies,
+            hooks=merge_hooks(request.hooks, self.hooks),
+        )
+        return p
+
+    def request(
+        self,
+        method,
+        url,
+        params=None,
+        data=None,
+        headers=None,
+        cookies=None,
+        files=None,
+        auth=None,
+        timeout=None,
+        allow_redirects=True,
+        proxies=None,
+        hooks=None,
+        stream=None,
+        verify=None,
+        cert=None,
+        json=None,
+    ):
+        """Constructs a :class:`Request <Request>`, prepares it and sends it.
+        Returns :class:`Response <Response>` object.
+
+        :param method: method for the new :class:`Request` object.
+        :param url: URL for the new :class:`Request` object.
+        :param params: (optional) Dictionary or bytes to be sent in the query
+            string for the :class:`Request`.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param json: (optional) json to send in the body of the
+            :class:`Request`.
+        :param headers: (optional) Dictionary of HTTP Headers to send with the
+            :class:`Request`.
+        :param cookies: (optional) Dict or CookieJar object to send with the
+            :class:`Request`.
+        :param files: (optional) Dictionary of ``'filename': file-like-objects``
+            for multipart encoding upload.
+        :param auth: (optional) Auth tuple or callable to enable
+            Basic/Digest/Custom HTTP Auth.
+        :param timeout: (optional) How long to wait for the server to send
+            data before giving up, as a float, or a :ref:`(connect timeout,
+            read timeout) <timeouts>` tuple.
+        :type timeout: float or tuple
+        :param allow_redirects: (optional) Set to True by default.
+        :type allow_redirects: bool
+        :param proxies: (optional) Dictionary mapping protocol or protocol and
+            hostname to the URL of the proxy.
+        :param hooks: (optional) Dictionary mapping hook name to one event or
+            list of events, event must be callable.
+        :param stream: (optional) whether to immediately download the response
+            content. Defaults to ``False``.
+        :param verify: (optional) Either a boolean, in which case it controls whether we verify
+            the server's TLS certificate, or a string, in which case it must be a path
+            to a CA bundle to use. Defaults to ``True``. When set to
+            ``False``, requests will accept any TLS certificate presented by
+            the server, and will ignore hostname mismatches and/or expired
+            certificates, which will make your application vulnerable to
+            man-in-the-middle (MitM) attacks. Setting verify to ``False``
+            may be useful during local development or testing.
+        :param cert: (optional) if String, path to ssl client cert file (.pem).
+            If Tuple, ('cert', 'key') pair.
+        :rtype: requests.Response
+        """
+        # Create the Request.
+        req = Request(
+            method=method.upper(),
+            url=url,
+            headers=headers,
+            files=files,
+            data=data or {},
+            json=json,
+            params=params or {},
+            auth=auth,
+            cookies=cookies,
+            hooks=hooks,
+        )
+        prep = self.prepare_request(req)
+
+        proxies = proxies or {}
+
+        settings = self.merge_environment_settings(
+            prep.url, proxies, stream, verify, cert
+        )
+
+        # Send the request.
+        send_kwargs = {
+            "timeout": timeout,
+            "allow_redirects": allow_redirects,
+        }
+        send_kwargs.update(settings)
+        resp = self.send(prep, **send_kwargs)
+
+        return resp
+
+    def get(self, url, **kwargs):
+        r"""Sends a GET request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", True)
+        return self.request("GET", url, **kwargs)
+
+    def options(self, url, **kwargs):
+        r"""Sends a OPTIONS request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", True)
+        return self.request("OPTIONS", url, **kwargs)
+
+    def head(self, url, **kwargs):
+        r"""Sends a HEAD request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        kwargs.setdefault("allow_redirects", False)
+        return self.request("HEAD", url, **kwargs)
+
+    def post(self, url, data=None, json=None, **kwargs):
+        r"""Sends a POST request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param json: (optional) json to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("POST", url, data=data, json=json, **kwargs)
+
+    def put(self, url, data=None, **kwargs):
+        r"""Sends a PUT request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("PUT", url, data=data, **kwargs)
+
+    def patch(self, url, data=None, **kwargs):
+        r"""Sends a PATCH request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param data: (optional) Dictionary, list of tuples, bytes, or file-like
+            object to send in the body of the :class:`Request`.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("PATCH", url, data=data, **kwargs)
+
+    def delete(self, url, **kwargs):
+        r"""Sends a DELETE request. Returns :class:`Response` object.
+
+        :param url: URL for the new :class:`Request` object.
+        :param \*\*kwargs: Optional arguments that ``request`` takes.
+        :rtype: requests.Response
+        """
+
+        return self.request("DELETE", url, **kwargs)
+
+    def send(self, request, **kwargs):
+        """Send a given PreparedRequest.
+
+        :rtype: requests.Response
+        """
+        # Set defaults that the hooks can utilize to ensure they always have
+        # the correct parameters to reproduce the previous request.
+        kwargs.setdefault("stream", self.stream)
+        kwargs.setdefault("verify", self.verify)
+        kwargs.setdefault("cert", self.cert)
+        if "proxies" not in kwargs:
+            kwargs["proxies"] = resolve_proxies(request, self.proxies, self.trust_env)
+
+        # It's possible that users might accidentally send a Request object.
+        # Guard against that specific failure case.
+        if isinstance(request, Request):
+            raise ValueError("You can only send PreparedRequests.")
+
+        # Set up variables needed for resolve_redirects and dispatching of hooks
+        allow_redirects = kwargs.pop("allow_redirects", True)
+        stream = kwargs.get("stream")
+        hooks = request.hooks
+
+        # Get the appropriate adapter to use
+        adapter = self.get_adapter(url=request.url)
+
+        # Start time (approximately) of the request
+        start = preferred_clock()
+
+        # Send the request
+        r = adapter.send(request, **kwargs)
+
+        # Total elapsed time of the request (approximately)
+        elapsed = preferred_clock() - start
+        r.elapsed = timedelta(seconds=elapsed)
+
+        # Response manipulation hooks
+        r = dispatch_hook("response", hooks, r, **kwargs)
+
+        # Persist cookies
+        if r.history:
+            # If the hooks create history then we want those cookies too
+            for resp in r.history:
+                extract_cookies_to_jar(self.cookies, resp.request, resp.raw)
+
+        extract_cookies_to_jar(self.cookies, request, r.raw)
+
+        # Resolve redirects if allowed.
+        if allow_redirects:
+            # Redirect resolving generator.
+            gen = self.resolve_redirects(r, request, **kwargs)
+            history = [resp for resp in gen]
+        else:
+            history = []
+
+        # Shuffle things around if there's history.
+        if history:
+            # Insert the first (original) request at the start
+            history.insert(0, r)
+            # Get the last request made
+            r = history.pop()
+            r.history = history
+
+        # If redirects aren't being followed, store the response on the Request for Response.next().
+        if not allow_redirects:
+            try:
+                r._next = next(
+                    self.resolve_redirects(r, request, yield_requests=True, **kwargs)
+                )
+            except StopIteration:
+                pass
+
+        if not stream:
+            r.content
+
+        return r
+
+    def merge_environment_settings(self, url, proxies, stream, verify, cert):
+        """
+        Check the environment and merge it with some settings.
+
+        :rtype: dict
+        """
+        # Gather clues from the surrounding environment.
+        if self.trust_env:
+            # Set environment's proxies.
+            no_proxy = proxies.get("no_proxy") if proxies is not None else None
+            env_proxies = get_environ_proxies(url, no_proxy=no_proxy)
+            for k, v in env_proxies.items():
+                proxies.setdefault(k, v)
+
+            # Look for requests environment configuration
+            # and be compatible with cURL.
+            if verify is True or verify is None:
+                verify = (
+                    os.environ.get("REQUESTS_CA_BUNDLE")
+                    or os.environ.get("CURL_CA_BUNDLE")
+                    or verify
+                )
+
+        # Merge all the kwargs.
+        proxies = merge_setting(proxies, self.proxies)
+        stream = merge_setting(stream, self.stream)
+        verify = merge_setting(verify, self.verify)
+        cert = merge_setting(cert, self.cert)
+
+        return {"proxies": proxies, "stream": stream, "verify": verify, "cert": cert}
+
+    def get_adapter(self, url):
+        """
+        Returns the appropriate connection adapter for the given URL.
+
+        :rtype: requests.adapters.BaseAdapter
+        """
+        for prefix, adapter in self.adapters.items():
+            if url.lower().startswith(prefix.lower()):
+                return adapter
+
+        # Nothing matches :-/
+        raise InvalidSchema(f"No connection adapters were found for {url!r}")
+
+    def close(self):
+        """Closes all adapters and as such the session"""
+        for v in self.adapters.values():
+            v.close()
+
+    def mount(self, prefix, adapter):
+        """Registers a connection adapter to a prefix.
+
+        Adapters are sorted in descending order by prefix length.
+        """
+        self.adapters[prefix] = adapter
+        keys_to_move = [k for k in self.adapters if len(k) < len(prefix)]
+
+        for key in keys_to_move:
+            self.adapters[key] = self.adapters.pop(key)
+
+    def __getstate__(self):
+        state = {attr: getattr(self, attr, None) for attr in self.__attrs__}
+        return state
+
+    def __setstate__(self, state):
+        for attr, value in state.items():
+            setattr(self, attr, value)
+
+
+def session():
+    """
+    Returns a :class:`Session` for context-management.
+
+    .. deprecated:: 1.0.0
+
+        This method has been deprecated since version 1.0.0 and is only kept for
+        backwards compatibility. New code should use :class:`~requests.sessions.Session`
+        to create a session. This may be removed at a future date.
+
+    :rtype: Session
+    """
+    return Session()
diff --git a/backend/venv/lib/python3.12/site-packages/requests/status_codes.py b/backend/venv/lib/python3.12/site-packages/requests/status_codes.py
new file mode 100644
index 00000000..c7945a2f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/status_codes.py
@@ -0,0 +1,128 @@
+r"""
+The ``codes`` object defines a mapping from common names for HTTP statuses
+to their numerical codes, accessible either as attributes or as dictionary
+items.
+
+Example::
+
+    >>> import requests
+    >>> requests.codes['temporary_redirect']
+    307
+    >>> requests.codes.teapot
+    418
+    >>> requests.codes['\o/']
+    200
+
+Some codes have multiple names, and both upper- and lower-case versions of
+the names are allowed. For example, ``codes.ok``, ``codes.OK``, and
+``codes.okay`` all correspond to the HTTP status code 200.
+"""
+
+from .structures import LookupDict
+
+_codes = {
+    # Informational.
+    100: ("continue",),
+    101: ("switching_protocols",),
+    102: ("processing", "early-hints"),
+    103: ("checkpoint",),
+    122: ("uri_too_long", "request_uri_too_long"),
+    200: ("ok", "okay", "all_ok", "all_okay", "all_good", "\\o/", "✓"),
+    201: ("created",),
+    202: ("accepted",),
+    203: ("non_authoritative_info", "non_authoritative_information"),
+    204: ("no_content",),
+    205: ("reset_content", "reset"),
+    206: ("partial_content", "partial"),
+    207: ("multi_status", "multiple_status", "multi_stati", "multiple_stati"),
+    208: ("already_reported",),
+    226: ("im_used",),
+    # Redirection.
+    300: ("multiple_choices",),
+    301: ("moved_permanently", "moved", "\\o-"),
+    302: ("found",),
+    303: ("see_other", "other"),
+    304: ("not_modified",),
+    305: ("use_proxy",),
+    306: ("switch_proxy",),
+    307: ("temporary_redirect", "temporary_moved", "temporary"),
+    308: (
+        "permanent_redirect",
+        "resume_incomplete",
+        "resume",
+    ),  # "resume" and "resume_incomplete" to be removed in 3.0
+    # Client Error.
+    400: ("bad_request", "bad"),
+    401: ("unauthorized",),
+    402: ("payment_required", "payment"),
+    403: ("forbidden",),
+    404: ("not_found", "-o-"),
+    405: ("method_not_allowed", "not_allowed"),
+    406: ("not_acceptable",),
+    407: ("proxy_authentication_required", "proxy_auth", "proxy_authentication"),
+    408: ("request_timeout", "timeout"),
+    409: ("conflict",),
+    410: ("gone",),
+    411: ("length_required",),
+    412: ("precondition_failed", "precondition"),
+    413: ("request_entity_too_large", "content_too_large"),
+    414: ("request_uri_too_large", "uri_too_long"),
+    415: ("unsupported_media_type", "unsupported_media", "media_type"),
+    416: (
+        "requested_range_not_satisfiable",
+        "requested_range",
+        "range_not_satisfiable",
+    ),
+    417: ("expectation_failed",),
+    418: ("im_a_teapot", "teapot", "i_am_a_teapot"),
+    421: ("misdirected_request",),
+    422: ("unprocessable_entity", "unprocessable", "unprocessable_content"),
+    423: ("locked",),
+    424: ("failed_dependency", "dependency"),
+    425: ("unordered_collection", "unordered", "too_early"),
+    426: ("upgrade_required", "upgrade"),
+    428: ("precondition_required", "precondition"),
+    429: ("too_many_requests", "too_many"),
+    431: ("header_fields_too_large", "fields_too_large"),
+    444: ("no_response", "none"),
+    449: ("retry_with", "retry"),
+    450: ("blocked_by_windows_parental_controls", "parental_controls"),
+    451: ("unavailable_for_legal_reasons", "legal_reasons"),
+    499: ("client_closed_request",),
+    # Server Error.
+    500: ("internal_server_error", "server_error", "/o\\", "✗"),
+    501: ("not_implemented",),
+    502: ("bad_gateway",),
+    503: ("service_unavailable", "unavailable"),
+    504: ("gateway_timeout",),
+    505: ("http_version_not_supported", "http_version"),
+    506: ("variant_also_negotiates",),
+    507: ("insufficient_storage",),
+    509: ("bandwidth_limit_exceeded", "bandwidth"),
+    510: ("not_extended",),
+    511: ("network_authentication_required", "network_auth", "network_authentication"),
+}
+
+codes = LookupDict(name="status_codes")
+
+
+def _init():
+    for code, titles in _codes.items():
+        for title in titles:
+            setattr(codes, title, code)
+            if not title.startswith(("\\", "/")):
+                setattr(codes, title.upper(), code)
+
+    def doc(code):
+        names = ", ".join(f"``{n}``" for n in _codes[code])
+        return "* %d: %s" % (code, names)
+
+    global __doc__
+    __doc__ = (
+        __doc__ + "\n" + "\n".join(doc(code) for code in sorted(_codes))
+        if __doc__ is not None
+        else None
+    )
+
+
+_init()
diff --git a/backend/venv/lib/python3.12/site-packages/requests/structures.py b/backend/venv/lib/python3.12/site-packages/requests/structures.py
new file mode 100644
index 00000000..188e13e4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/structures.py
@@ -0,0 +1,99 @@
+"""
+requests.structures
+~~~~~~~~~~~~~~~~~~~
+
+Data structures that power Requests.
+"""
+
+from collections import OrderedDict
+
+from .compat import Mapping, MutableMapping
+
+
+class CaseInsensitiveDict(MutableMapping):
+    """A case-insensitive ``dict``-like object.
+
+    Implements all methods and operations of
+    ``MutableMapping`` as well as dict's ``copy``. Also
+    provides ``lower_items``.
+
+    All keys are expected to be strings. The structure remembers the
+    case of the last key to be set, and ``iter(instance)``,
+    ``keys()``, ``items()``, ``iterkeys()``, and ``iteritems()``
+    will contain case-sensitive keys. However, querying and contains
+    testing is case insensitive::
+
+        cid = CaseInsensitiveDict()
+        cid['Accept'] = 'application/json'
+        cid['aCCEPT'] == 'application/json'  # True
+        list(cid) == ['Accept']  # True
+
+    For example, ``headers['content-encoding']`` will return the
+    value of a ``'Content-Encoding'`` response header, regardless
+    of how the header name was originally stored.
+
+    If the constructor, ``.update``, or equality comparison
+    operations are given keys that have equal ``.lower()``s, the
+    behavior is undefined.
+    """
+
+    def __init__(self, data=None, **kwargs):
+        self._store = OrderedDict()
+        if data is None:
+            data = {}
+        self.update(data, **kwargs)
+
+    def __setitem__(self, key, value):
+        # Use the lowercased key for lookups, but store the actual
+        # key alongside the value.
+        self._store[key.lower()] = (key, value)
+
+    def __getitem__(self, key):
+        return self._store[key.lower()][1]
+
+    def __delitem__(self, key):
+        del self._store[key.lower()]
+
+    def __iter__(self):
+        return (casedkey for casedkey, mappedvalue in self._store.values())
+
+    def __len__(self):
+        return len(self._store)
+
+    def lower_items(self):
+        """Like iteritems(), but with all lowercase keys."""
+        return ((lowerkey, keyval[1]) for (lowerkey, keyval) in self._store.items())
+
+    def __eq__(self, other):
+        if isinstance(other, Mapping):
+            other = CaseInsensitiveDict(other)
+        else:
+            return NotImplemented
+        # Compare insensitively
+        return dict(self.lower_items()) == dict(other.lower_items())
+
+    # Copy is required
+    def copy(self):
+        return CaseInsensitiveDict(self._store.values())
+
+    def __repr__(self):
+        return str(dict(self.items()))
+
+
+class LookupDict(dict):
+    """Dictionary lookup object."""
+
+    def __init__(self, name=None):
+        self.name = name
+        super().__init__()
+
+    def __repr__(self):
+        return f"<lookup '{self.name}'>"
+
+    def __getitem__(self, key):
+        # We allow fall-through here, so values default to None
+
+        return self.__dict__.get(key, None)
+
+    def get(self, key, default=None):
+        return self.__dict__.get(key, default)
diff --git a/backend/venv/lib/python3.12/site-packages/requests/utils.py b/backend/venv/lib/python3.12/site-packages/requests/utils.py
new file mode 100644
index 00000000..ae6c42f6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/requests/utils.py
@@ -0,0 +1,1096 @@
+"""
+requests.utils
+~~~~~~~~~~~~~~
+
+This module provides utility functions that are used within Requests
+that are also useful for external consumption.
+"""
+
+import codecs
+import contextlib
+import io
+import os
+import re
+import socket
+import struct
+import sys
+import tempfile
+import warnings
+import zipfile
+from collections import OrderedDict
+
+from urllib3.util import make_headers, parse_url
+
+from . import certs
+from .__version__ import __version__
+
+# to_native_string is unused here, but imported here for backwards compatibility
+from ._internal_utils import (  # noqa: F401
+    _HEADER_VALIDATORS_BYTE,
+    _HEADER_VALIDATORS_STR,
+    HEADER_VALIDATORS,
+    to_native_string,
+)
+from .compat import (
+    Mapping,
+    basestring,
+    bytes,
+    getproxies,
+    getproxies_environment,
+    integer_types,
+)
+from .compat import parse_http_list as _parse_list_header
+from .compat import (
+    proxy_bypass,
+    proxy_bypass_environment,
+    quote,
+    str,
+    unquote,
+    urlparse,
+    urlunparse,
+)
+from .cookies import cookiejar_from_dict
+from .exceptions import (
+    FileModeWarning,
+    InvalidHeader,
+    InvalidURL,
+    UnrewindableBodyError,
+)
+from .structures import CaseInsensitiveDict
+
+NETRC_FILES = (".netrc", "_netrc")
+
+DEFAULT_CA_BUNDLE_PATH = certs.where()
+
+DEFAULT_PORTS = {"http": 80, "https": 443}
+
+# Ensure that ', ' is used to preserve previous delimiter behavior.
+DEFAULT_ACCEPT_ENCODING = ", ".join(
+    re.split(r",\s*", make_headers(accept_encoding=True)["accept-encoding"])
+)
+
+
+if sys.platform == "win32":
+    # provide a proxy_bypass version on Windows without DNS lookups
+
+    def proxy_bypass_registry(host):
+        try:
+            import winreg
+        except ImportError:
+            return False
+
+        try:
+            internetSettings = winreg.OpenKey(
+                winreg.HKEY_CURRENT_USER,
+                r"Software\Microsoft\Windows\CurrentVersion\Internet Settings",
+            )
+            # ProxyEnable could be REG_SZ or REG_DWORD, normalizing it
+            proxyEnable = int(winreg.QueryValueEx(internetSettings, "ProxyEnable")[0])
+            # ProxyOverride is almost always a string
+            proxyOverride = winreg.QueryValueEx(internetSettings, "ProxyOverride")[0]
+        except (OSError, ValueError):
+            return False
+        if not proxyEnable or not proxyOverride:
+            return False
+
+        # make a check value list from the registry entry: replace the
+        # '<local>' string by the localhost entry and the corresponding
+        # canonical entry.
+        proxyOverride = proxyOverride.split(";")
+        # filter out empty strings to avoid re.match return true in the following code.
+        proxyOverride = filter(None, proxyOverride)
+        # now check if we match one of the registry values.
+        for test in proxyOverride:
+            if test == "<local>":
+                if "." not in host:
+                    return True
+            test = test.replace(".", r"\.")  # mask dots
+            test = test.replace("*", r".*")  # change glob sequence
+            test = test.replace("?", r".")  # change glob char
+            if re.match(test, host, re.I):
+                return True
+        return False
+
+    def proxy_bypass(host):  # noqa
+        """Return True, if the host should be bypassed.
+
+        Checks proxy settings gathered from the environment, if specified,
+        or the registry.
+        """
+        if getproxies_environment():
+            return proxy_bypass_environment(host)
+        else:
+            return proxy_bypass_registry(host)
+
+
+def dict_to_sequence(d):
+    """Returns an internal sequence dictionary update."""
+
+    if hasattr(d, "items"):
+        d = d.items()
+
+    return d
+
+
+def super_len(o):
+    total_length = None
+    current_position = 0
+
+    if isinstance(o, str):
+        o = o.encode("utf-8")
+
+    if hasattr(o, "__len__"):
+        total_length = len(o)
+
+    elif hasattr(o, "len"):
+        total_length = o.len
+
+    elif hasattr(o, "fileno"):
+        try:
+            fileno = o.fileno()
+        except (io.UnsupportedOperation, AttributeError):
+            # AttributeError is a surprising exception, seeing as how we've just checked
+            # that `hasattr(o, 'fileno')`.  It happens for objects obtained via
+            # `Tarfile.extractfile()`, per issue 5229.
+            pass
+        else:
+            total_length = os.fstat(fileno).st_size
+
+            # Having used fstat to determine the file length, we need to
+            # confirm that this file was opened up in binary mode.
+            if "b" not in o.mode:
+                warnings.warn(
+                    (
+                        "Requests has determined the content-length for this "
+                        "request using the binary size of the file: however, the "
+                        "file has been opened in text mode (i.e. without the 'b' "
+                        "flag in the mode). This may lead to an incorrect "
+                        "content-length. In Requests 3.0, support will be removed "
+                        "for files in text mode."
+                    ),
+                    FileModeWarning,
+                )
+
+    if hasattr(o, "tell"):
+        try:
+            current_position = o.tell()
+        except OSError:
+            # This can happen in some weird situations, such as when the file
+            # is actually a special file descriptor like stdin. In this
+            # instance, we don't know what the length is, so set it to zero and
+            # let requests chunk it instead.
+            if total_length is not None:
+                current_position = total_length
+        else:
+            if hasattr(o, "seek") and total_length is None:
+                # StringIO and BytesIO have seek but no usable fileno
+                try:
+                    # seek to end of file
+                    o.seek(0, 2)
+                    total_length = o.tell()
+
+                    # seek back to current position to support
+                    # partially read file-like objects
+                    o.seek(current_position or 0)
+                except OSError:
+                    total_length = 0
+
+    if total_length is None:
+        total_length = 0
+
+    return max(0, total_length - current_position)
+
+
+def get_netrc_auth(url, raise_errors=False):
+    """Returns the Requests tuple auth for a given url from netrc."""
+
+    netrc_file = os.environ.get("NETRC")
+    if netrc_file is not None:
+        netrc_locations = (netrc_file,)
+    else:
+        netrc_locations = (f"~/{f}" for f in NETRC_FILES)
+
+    try:
+        from netrc import NetrcParseError, netrc
+
+        netrc_path = None
+
+        for f in netrc_locations:
+            try:
+                loc = os.path.expanduser(f)
+            except KeyError:
+                # os.path.expanduser can fail when $HOME is undefined and
+                # getpwuid fails. See https://bugs.python.org/issue20164 &
+                # https://github.com/psf/requests/issues/1846
+                return
+
+            if os.path.exists(loc):
+                netrc_path = loc
+                break
+
+        # Abort early if there isn't one.
+        if netrc_path is None:
+            return
+
+        ri = urlparse(url)
+
+        # Strip port numbers from netloc. This weird `if...encode`` dance is
+        # used for Python 3.2, which doesn't support unicode literals.
+        splitstr = b":"
+        if isinstance(url, str):
+            splitstr = splitstr.decode("ascii")
+        host = ri.netloc.split(splitstr)[0]
+
+        try:
+            _netrc = netrc(netrc_path).authenticators(host)
+            if _netrc:
+                # Return with login / password
+                login_i = 0 if _netrc[0] else 1
+                return (_netrc[login_i], _netrc[2])
+        except (NetrcParseError, OSError):
+            # If there was a parsing error or a permissions issue reading the file,
+            # we'll just skip netrc auth unless explicitly asked to raise errors.
+            if raise_errors:
+                raise
+
+    # App Engine hackiness.
+    except (ImportError, AttributeError):
+        pass
+
+
+def guess_filename(obj):
+    """Tries to guess the filename of the given object."""
+    name = getattr(obj, "name", None)
+    if name and isinstance(name, basestring) and name[0] != "<" and name[-1] != ">":
+        return os.path.basename(name)
+
+
+def extract_zipped_paths(path):
+    """Replace nonexistent paths that look like they refer to a member of a zip
+    archive with the location of an extracted copy of the target, or else
+    just return the provided path unchanged.
+    """
+    if os.path.exists(path):
+        # this is already a valid path, no need to do anything further
+        return path
+
+    # find the first valid part of the provided path and treat that as a zip archive
+    # assume the rest of the path is the name of a member in the archive
+    archive, member = os.path.split(path)
+    while archive and not os.path.exists(archive):
+        archive, prefix = os.path.split(archive)
+        if not prefix:
+            # If we don't check for an empty prefix after the split (in other words, archive remains unchanged after the split),
+            # we _can_ end up in an infinite loop on a rare corner case affecting a small number of users
+            break
+        member = "/".join([prefix, member])
+
+    if not zipfile.is_zipfile(archive):
+        return path
+
+    zip_file = zipfile.ZipFile(archive)
+    if member not in zip_file.namelist():
+        return path
+
+    # we have a valid zip archive and a valid member of that archive
+    tmp = tempfile.gettempdir()
+    extracted_path = os.path.join(tmp, member.split("/")[-1])
+    if not os.path.exists(extracted_path):
+        # use read + write to avoid the creating nested folders, we only want the file, avoids mkdir racing condition
+        with atomic_open(extracted_path) as file_handler:
+            file_handler.write(zip_file.read(member))
+    return extracted_path
+
+
+@contextlib.contextmanager
+def atomic_open(filename):
+    """Write a file to the disk in an atomic fashion"""
+    tmp_descriptor, tmp_name = tempfile.mkstemp(dir=os.path.dirname(filename))
+    try:
+        with os.fdopen(tmp_descriptor, "wb") as tmp_handler:
+            yield tmp_handler
+        os.replace(tmp_name, filename)
+    except BaseException:
+        os.remove(tmp_name)
+        raise
+
+
+def from_key_val_list(value):
+    """Take an object and test to see if it can be represented as a
+    dictionary. Unless it can not be represented as such, return an
+    OrderedDict, e.g.,
+
+    ::
+
+        >>> from_key_val_list([('key', 'val')])
+        OrderedDict([('key', 'val')])
+        >>> from_key_val_list('string')
+        Traceback (most recent call last):
+        ...
+        ValueError: cannot encode objects that are not 2-tuples
+        >>> from_key_val_list({'key': 'val'})
+        OrderedDict([('key', 'val')])
+
+    :rtype: OrderedDict
+    """
+    if value is None:
+        return None
+
+    if isinstance(value, (str, bytes, bool, int)):
+        raise ValueError("cannot encode objects that are not 2-tuples")
+
+    return OrderedDict(value)
+
+
+def to_key_val_list(value):
+    """Take an object and test to see if it can be represented as a
+    dictionary. If it can be, return a list of tuples, e.g.,
+
+    ::
+
+        >>> to_key_val_list([('key', 'val')])
+        [('key', 'val')]
+        >>> to_key_val_list({'key': 'val'})
+        [('key', 'val')]
+        >>> to_key_val_list('string')
+        Traceback (most recent call last):
+        ...
+        ValueError: cannot encode objects that are not 2-tuples
+
+    :rtype: list
+    """
+    if value is None:
+        return None
+
+    if isinstance(value, (str, bytes, bool, int)):
+        raise ValueError("cannot encode objects that are not 2-tuples")
+
+    if isinstance(value, Mapping):
+        value = value.items()
+
+    return list(value)
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def parse_list_header(value):
+    """Parse lists as described by RFC 2068 Section 2.
+
+    In particular, parse comma-separated lists where the elements of
+    the list may include quoted-strings.  A quoted-string could
+    contain a comma.  A non-quoted string could have quotes in the
+    middle.  Quotes are removed automatically after parsing.
+
+    It basically works like :func:`parse_set_header` just that items
+    may appear multiple times and case sensitivity is preserved.
+
+    The return value is a standard :class:`list`:
+
+    >>> parse_list_header('token, "quoted value"')
+    ['token', 'quoted value']
+
+    To create a header from the :class:`list` again, use the
+    :func:`dump_header` function.
+
+    :param value: a string with a list header.
+    :return: :class:`list`
+    :rtype: list
+    """
+    result = []
+    for item in _parse_list_header(value):
+        if item[:1] == item[-1:] == '"':
+            item = unquote_header_value(item[1:-1])
+        result.append(item)
+    return result
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def parse_dict_header(value):
+    """Parse lists of key, value pairs as described by RFC 2068 Section 2 and
+    convert them into a python dict:
+
+    >>> d = parse_dict_header('foo="is a fish", bar="as well"')
+    >>> type(d) is dict
+    True
+    >>> sorted(d.items())
+    [('bar', 'as well'), ('foo', 'is a fish')]
+
+    If there is no value for a key it will be `None`:
+
+    >>> parse_dict_header('key_without_value')
+    {'key_without_value': None}
+
+    To create a header from the :class:`dict` again, use the
+    :func:`dump_header` function.
+
+    :param value: a string with a dict header.
+    :return: :class:`dict`
+    :rtype: dict
+    """
+    result = {}
+    for item in _parse_list_header(value):
+        if "=" not in item:
+            result[item] = None
+            continue
+        name, value = item.split("=", 1)
+        if value[:1] == value[-1:] == '"':
+            value = unquote_header_value(value[1:-1])
+        result[name] = value
+    return result
+
+
+# From mitsuhiko/werkzeug (used with permission).
+def unquote_header_value(value, is_filename=False):
+    r"""Unquotes a header value.  (Reversal of :func:`quote_header_value`).
+    This does not use the real unquoting but what browsers are actually
+    using for quoting.
+
+    :param value: the header value to unquote.
+    :rtype: str
+    """
+    if value and value[0] == value[-1] == '"':
+        # this is not the real unquoting, but fixing this so that the
+        # RFC is met will result in bugs with internet explorer and
+        # probably some other browsers as well.  IE for example is
+        # uploading files with "C:\foo\bar.txt" as filename
+        value = value[1:-1]
+
+        # if this is a filename and the starting characters look like
+        # a UNC path, then just return the value without quotes.  Using the
+        # replace sequence below on a UNC path has the effect of turning
+        # the leading double slash into a single slash and then
+        # _fix_ie_filename() doesn't work correctly.  See #458.
+        if not is_filename or value[:2] != "\\\\":
+            return value.replace("\\\\", "\\").replace('\\"', '"')
+    return value
+
+
+def dict_from_cookiejar(cj):
+    """Returns a key/value dictionary from a CookieJar.
+
+    :param cj: CookieJar object to extract cookies from.
+    :rtype: dict
+    """
+
+    cookie_dict = {cookie.name: cookie.value for cookie in cj}
+    return cookie_dict
+
+
+def add_dict_to_cookiejar(cj, cookie_dict):
+    """Returns a CookieJar from a key/value dictionary.
+
+    :param cj: CookieJar to insert cookies into.
+    :param cookie_dict: Dict of key/values to insert into CookieJar.
+    :rtype: CookieJar
+    """
+
+    return cookiejar_from_dict(cookie_dict, cj)
+
+
+def get_encodings_from_content(content):
+    """Returns encodings from given content string.
+
+    :param content: bytestring to extract encodings from.
+    """
+    warnings.warn(
+        (
+            "In requests 3.0, get_encodings_from_content will be removed. For "
+            "more information, please see the discussion on issue #2266. (This"
+            " warning should only appear once.)"
+        ),
+        DeprecationWarning,
+    )
+
+    charset_re = re.compile(r'<meta.*?charset=["\']*(.+?)["\'>]', flags=re.I)
+    pragma_re = re.compile(r'<meta.*?content=["\']*;?charset=(.+?)["\'>]', flags=re.I)
+    xml_re = re.compile(r'^<\?xml.*?encoding=["\']*(.+?)["\'>]')
+
+    return (
+        charset_re.findall(content)
+        + pragma_re.findall(content)
+        + xml_re.findall(content)
+    )
+
+
+def _parse_content_type_header(header):
+    """Returns content type and parameters from given header
+
+    :param header: string
+    :return: tuple containing content type and dictionary of
+         parameters
+    """
+
+    tokens = header.split(";")
+    content_type, params = tokens[0].strip(), tokens[1:]
+    params_dict = {}
+    items_to_strip = "\"' "
+
+    for param in params:
+        param = param.strip()
+        if param:
+            key, value = param, True
+            index_of_equals = param.find("=")
+            if index_of_equals != -1:
+                key = param[:index_of_equals].strip(items_to_strip)
+                value = param[index_of_equals + 1 :].strip(items_to_strip)
+            params_dict[key.lower()] = value
+    return content_type, params_dict
+
+
+def get_encoding_from_headers(headers):
+    """Returns encodings from given HTTP Header Dict.
+
+    :param headers: dictionary to extract encoding from.
+    :rtype: str
+    """
+
+    content_type = headers.get("content-type")
+
+    if not content_type:
+        return None
+
+    content_type, params = _parse_content_type_header(content_type)
+
+    if "charset" in params:
+        return params["charset"].strip("'\"")
+
+    if "text" in content_type:
+        return "ISO-8859-1"
+
+    if "application/json" in content_type:
+        # Assume UTF-8 based on RFC 4627: https://www.ietf.org/rfc/rfc4627.txt since the charset was unset
+        return "utf-8"
+
+
+def stream_decode_response_unicode(iterator, r):
+    """Stream decodes an iterator."""
+
+    if r.encoding is None:
+        yield from iterator
+        return
+
+    decoder = codecs.getincrementaldecoder(r.encoding)(errors="replace")
+    for chunk in iterator:
+        rv = decoder.decode(chunk)
+        if rv:
+            yield rv
+    rv = decoder.decode(b"", final=True)
+    if rv:
+        yield rv
+
+
+def iter_slices(string, slice_length):
+    """Iterate over slices of a string."""
+    pos = 0
+    if slice_length is None or slice_length <= 0:
+        slice_length = len(string)
+    while pos < len(string):
+        yield string[pos : pos + slice_length]
+        pos += slice_length
+
+
+def get_unicode_from_response(r):
+    """Returns the requested content back in unicode.
+
+    :param r: Response object to get unicode content from.
+
+    Tried:
+
+    1. charset from content-type
+    2. fall back and replace all unicode characters
+
+    :rtype: str
+    """
+    warnings.warn(
+        (
+            "In requests 3.0, get_unicode_from_response will be removed. For "
+            "more information, please see the discussion on issue #2266. (This"
+            " warning should only appear once.)"
+        ),
+        DeprecationWarning,
+    )
+
+    tried_encodings = []
+
+    # Try charset from content-type
+    encoding = get_encoding_from_headers(r.headers)
+
+    if encoding:
+        try:
+            return str(r.content, encoding)
+        except UnicodeError:
+            tried_encodings.append(encoding)
+
+    # Fall back:
+    try:
+        return str(r.content, encoding, errors="replace")
+    except TypeError:
+        return r.content
+
+
+# The unreserved URI characters (RFC 3986)
+UNRESERVED_SET = frozenset(
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz" + "0123456789-._~"
+)
+
+
+def unquote_unreserved(uri):
+    """Un-escape any percent-escape sequences in a URI that are unreserved
+    characters. This leaves all reserved, illegal and non-ASCII bytes encoded.
+
+    :rtype: str
+    """
+    parts = uri.split("%")
+    for i in range(1, len(parts)):
+        h = parts[i][0:2]
+        if len(h) == 2 and h.isalnum():
+            try:
+                c = chr(int(h, 16))
+            except ValueError:
+                raise InvalidURL(f"Invalid percent-escape sequence: '{h}'")
+
+            if c in UNRESERVED_SET:
+                parts[i] = c + parts[i][2:]
+            else:
+                parts[i] = f"%{parts[i]}"
+        else:
+            parts[i] = f"%{parts[i]}"
+    return "".join(parts)
+
+
+def requote_uri(uri):
+    """Re-quote the given URI.
+
+    This function passes the given URI through an unquote/quote cycle to
+    ensure that it is fully and consistently quoted.
+
+    :rtype: str
+    """
+    safe_with_percent = "!#$%&'()*+,/:;=?@[]~"
+    safe_without_percent = "!#$&'()*+,/:;=?@[]~"
+    try:
+        # Unquote only the unreserved characters
+        # Then quote only illegal characters (do not quote reserved,
+        # unreserved, or '%')
+        return quote(unquote_unreserved(uri), safe=safe_with_percent)
+    except InvalidURL:
+        # We couldn't unquote the given URI, so let's try quoting it, but
+        # there may be unquoted '%'s in the URI. We need to make sure they're
+        # properly quoted so they do not cause issues elsewhere.
+        return quote(uri, safe=safe_without_percent)
+
+
+def address_in_network(ip, net):
+    """This function allows you to check if an IP belongs to a network subnet
+
+    Example: returns True if ip = 192.168.1.1 and net = 192.168.1.0/24
+             returns False if ip = 192.168.1.1 and net = 192.168.100.0/24
+
+    :rtype: bool
+    """
+    ipaddr = struct.unpack("=L", socket.inet_aton(ip))[0]
+    netaddr, bits = net.split("/")
+    netmask = struct.unpack("=L", socket.inet_aton(dotted_netmask(int(bits))))[0]
+    network = struct.unpack("=L", socket.inet_aton(netaddr))[0] & netmask
+    return (ipaddr & netmask) == (network & netmask)
+
+
+def dotted_netmask(mask):
+    """Converts mask from /xx format to xxx.xxx.xxx.xxx
+
+    Example: if mask is 24 function returns 255.255.255.0
+
+    :rtype: str
+    """
+    bits = 0xFFFFFFFF ^ (1 << 32 - mask) - 1
+    return socket.inet_ntoa(struct.pack(">I", bits))
+
+
+def is_ipv4_address(string_ip):
+    """
+    :rtype: bool
+    """
+    try:
+        socket.inet_aton(string_ip)
+    except OSError:
+        return False
+    return True
+
+
+def is_valid_cidr(string_network):
+    """
+    Very simple check of the cidr format in no_proxy variable.
+
+    :rtype: bool
+    """
+    if string_network.count("/") == 1:
+        try:
+            mask = int(string_network.split("/")[1])
+        except ValueError:
+            return False
+
+        if mask < 1 or mask > 32:
+            return False
+
+        try:
+            socket.inet_aton(string_network.split("/")[0])
+        except OSError:
+            return False
+    else:
+        return False
+    return True
+
+
+@contextlib.contextmanager
+def set_environ(env_name, value):
+    """Set the environment variable 'env_name' to 'value'
+
+    Save previous value, yield, and then restore the previous value stored in
+    the environment variable 'env_name'.
+
+    If 'value' is None, do nothing"""
+    value_changed = value is not None
+    if value_changed:
+        old_value = os.environ.get(env_name)
+        os.environ[env_name] = value
+    try:
+        yield
+    finally:
+        if value_changed:
+            if old_value is None:
+                del os.environ[env_name]
+            else:
+                os.environ[env_name] = old_value
+
+
+def should_bypass_proxies(url, no_proxy):
+    """
+    Returns whether we should bypass proxies or not.
+
+    :rtype: bool
+    """
+
+    # Prioritize lowercase environment variables over uppercase
+    # to keep a consistent behaviour with other http projects (curl, wget).
+    def get_proxy(key):
+        return os.environ.get(key) or os.environ.get(key.upper())
+
+    # First check whether no_proxy is defined. If it is, check that the URL
+    # we're getting isn't in the no_proxy list.
+    no_proxy_arg = no_proxy
+    if no_proxy is None:
+        no_proxy = get_proxy("no_proxy")
+    parsed = urlparse(url)
+
+    if parsed.hostname is None:
+        # URLs don't always have hostnames, e.g. file:/// urls.
+        return True
+
+    if no_proxy:
+        # We need to check whether we match here. We need to see if we match
+        # the end of the hostname, both with and without the port.
+        no_proxy = (host for host in no_proxy.replace(" ", "").split(",") if host)
+
+        if is_ipv4_address(parsed.hostname):
+            for proxy_ip in no_proxy:
+                if is_valid_cidr(proxy_ip):
+                    if address_in_network(parsed.hostname, proxy_ip):
+                        return True
+                elif parsed.hostname == proxy_ip:
+                    # If no_proxy ip was defined in plain IP notation instead of cidr notation &
+                    # matches the IP of the index
+                    return True
+        else:
+            host_with_port = parsed.hostname
+            if parsed.port:
+                host_with_port += f":{parsed.port}"
+
+            for host in no_proxy:
+                if parsed.hostname.endswith(host) or host_with_port.endswith(host):
+                    # The URL does match something in no_proxy, so we don't want
+                    # to apply the proxies on this URL.
+                    return True
+
+    with set_environ("no_proxy", no_proxy_arg):
+        # parsed.hostname can be `None` in cases such as a file URI.
+        try:
+            bypass = proxy_bypass(parsed.hostname)
+        except (TypeError, socket.gaierror):
+            bypass = False
+
+    if bypass:
+        return True
+
+    return False
+
+
+def get_environ_proxies(url, no_proxy=None):
+    """
+    Return a dict of environment proxies.
+
+    :rtype: dict
+    """
+    if should_bypass_proxies(url, no_proxy=no_proxy):
+        return {}
+    else:
+        return getproxies()
+
+
+def select_proxy(url, proxies):
+    """Select a proxy for the url, if applicable.
+
+    :param url: The url being for the request
+    :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs
+    """
+    proxies = proxies or {}
+    urlparts = urlparse(url)
+    if urlparts.hostname is None:
+        return proxies.get(urlparts.scheme, proxies.get("all"))
+
+    proxy_keys = [
+        urlparts.scheme + "://" + urlparts.hostname,
+        urlparts.scheme,
+        "all://" + urlparts.hostname,
+        "all",
+    ]
+    proxy = None
+    for proxy_key in proxy_keys:
+        if proxy_key in proxies:
+            proxy = proxies[proxy_key]
+            break
+
+    return proxy
+
+
+def resolve_proxies(request, proxies, trust_env=True):
+    """This method takes proxy information from a request and configuration
+    input to resolve a mapping of target proxies. This will consider settings
+    such as NO_PROXY to strip proxy configurations.
+
+    :param request: Request or PreparedRequest
+    :param proxies: A dictionary of schemes or schemes and hosts to proxy URLs
+    :param trust_env: Boolean declaring whether to trust environment configs
+
+    :rtype: dict
+    """
+    proxies = proxies if proxies is not None else {}
+    url = request.url
+    scheme = urlparse(url).scheme
+    no_proxy = proxies.get("no_proxy")
+    new_proxies = proxies.copy()
+
+    if trust_env and not should_bypass_proxies(url, no_proxy=no_proxy):
+        environ_proxies = get_environ_proxies(url, no_proxy=no_proxy)
+
+        proxy = environ_proxies.get(scheme, environ_proxies.get("all"))
+
+        if proxy:
+            new_proxies.setdefault(scheme, proxy)
+    return new_proxies
+
+
+def default_user_agent(name="python-requests"):
+    """
+    Return a string representing the default user agent.
+
+    :rtype: str
+    """
+    return f"{name}/{__version__}"
+
+
+def default_headers():
+    """
+    :rtype: requests.structures.CaseInsensitiveDict
+    """
+    return CaseInsensitiveDict(
+        {
+            "User-Agent": default_user_agent(),
+            "Accept-Encoding": DEFAULT_ACCEPT_ENCODING,
+            "Accept": "*/*",
+            "Connection": "keep-alive",
+        }
+    )
+
+
+def parse_header_links(value):
+    """Return a list of parsed link headers proxies.
+
+    i.e. Link: <http:/.../front.jpeg>; rel=front; type="image/jpeg",<http://.../back.jpeg>; rel=back;type="image/jpeg"
+
+    :rtype: list
+    """
+
+    links = []
+
+    replace_chars = " '\""
+
+    value = value.strip(replace_chars)
+    if not value:
+        return links
+
+    for val in re.split(", *<", value):
+        try:
+            url, params = val.split(";", 1)
+        except ValueError:
+            url, params = val, ""
+
+        link = {"url": url.strip("<> '\"")}
+
+        for param in params.split(";"):
+            try:
+                key, value = param.split("=")
+            except ValueError:
+                break
+
+            link[key.strip(replace_chars)] = value.strip(replace_chars)
+
+        links.append(link)
+
+    return links
+
+
+# Null bytes; no need to recreate these on each call to guess_json_utf
+_null = "\x00".encode("ascii")  # encoding to ASCII for Python 3
+_null2 = _null * 2
+_null3 = _null * 3
+
+
+def guess_json_utf(data):
+    """
+    :rtype: str
+    """
+    # JSON always starts with two ASCII characters, so detection is as
+    # easy as counting the nulls and from their location and count
+    # determine the encoding. Also detect a BOM, if present.
+    sample = data[:4]
+    if sample in (codecs.BOM_UTF32_LE, codecs.BOM_UTF32_BE):
+        return "utf-32"  # BOM included
+    if sample[:3] == codecs.BOM_UTF8:
+        return "utf-8-sig"  # BOM included, MS style (discouraged)
+    if sample[:2] in (codecs.BOM_UTF16_LE, codecs.BOM_UTF16_BE):
+        return "utf-16"  # BOM included
+    nullcount = sample.count(_null)
+    if nullcount == 0:
+        return "utf-8"
+    if nullcount == 2:
+        if sample[::2] == _null2:  # 1st and 3rd are null
+            return "utf-16-be"
+        if sample[1::2] == _null2:  # 2nd and 4th are null
+            return "utf-16-le"
+        # Did not detect 2 valid UTF-16 ascii-range characters
+    if nullcount == 3:
+        if sample[:3] == _null3:
+            return "utf-32-be"
+        if sample[1:] == _null3:
+            return "utf-32-le"
+        # Did not detect a valid UTF-32 ascii-range character
+    return None
+
+
+def prepend_scheme_if_needed(url, new_scheme):
+    """Given a URL that may or may not have a scheme, prepend the given scheme.
+    Does not replace a present scheme with the one provided as an argument.
+
+    :rtype: str
+    """
+    parsed = parse_url(url)
+    scheme, auth, host, port, path, query, fragment = parsed
+
+    # A defect in urlparse determines that there isn't a netloc present in some
+    # urls. We previously assumed parsing was overly cautious, and swapped the
+    # netloc and path. Due to a lack of tests on the original defect, this is
+    # maintained with parse_url for backwards compatibility.
+    netloc = parsed.netloc
+    if not netloc:
+        netloc, path = path, netloc
+
+    if auth:
+        # parse_url doesn't provide the netloc with auth
+        # so we'll add it ourselves.
+        netloc = "@".join([auth, netloc])
+    if scheme is None:
+        scheme = new_scheme
+    if path is None:
+        path = ""
+
+    return urlunparse((scheme, netloc, path, "", query, fragment))
+
+
+def get_auth_from_url(url):
+    """Given a url with authentication components, extract them into a tuple of
+    username,password.
+
+    :rtype: (str,str)
+    """
+    parsed = urlparse(url)
+
+    try:
+        auth = (unquote(parsed.username), unquote(parsed.password))
+    except (AttributeError, TypeError):
+        auth = ("", "")
+
+    return auth
+
+
+def check_header_validity(header):
+    """Verifies that header parts don't contain leading whitespace
+    reserved characters, or return characters.
+
+    :param header: tuple, in the format (name, value).
+    """
+    name, value = header
+    _validate_header_part(header, name, 0)
+    _validate_header_part(header, value, 1)
+
+
+def _validate_header_part(header, header_part, header_validator_index):
+    if isinstance(header_part, str):
+        validator = _HEADER_VALIDATORS_STR[header_validator_index]
+    elif isinstance(header_part, bytes):
+        validator = _HEADER_VALIDATORS_BYTE[header_validator_index]
+    else:
+        raise InvalidHeader(
+            f"Header part ({header_part!r}) from {header} "
+            f"must be of type str or bytes, not {type(header_part)}"
+        )
+
+    if not validator.match(header_part):
+        header_kind = "name" if header_validator_index == 0 else "value"
+        raise InvalidHeader(
+            f"Invalid leading whitespace, reserved character(s), or return "
+            f"character(s) in header {header_kind}: {header_part!r}"
+        )
+
+
+def urldefragauth(url):
+    """
+    Given a url remove the fragment and the authentication part.
+
+    :rtype: str
+    """
+    scheme, netloc, path, params, query, fragment = urlparse(url)
+
+    # see func:`prepend_scheme_if_needed`
+    if not netloc:
+        netloc, path = path, netloc
+
+    netloc = netloc.rsplit("@", 1)[-1]
+
+    return urlunparse((scheme, netloc, path, params, query, ""))
+
+
+def rewind_body(prepared_request):
+    """Move file pointer back to its recorded starting position
+    so it can be read again on redirect.
+    """
+    body_seek = getattr(prepared_request.body, "seek", None)
+    if body_seek is not None and isinstance(
+        prepared_request._body_position, integer_types
+    ):
+        try:
+            body_seek(prepared_request._body_position)
+        except OSError:
+            raise UnrewindableBodyError(
+                "An error occurred when rewinding request body for redirect."
+            )
+    else:
+        raise UnrewindableBodyError("Unable to rewind request body for redirect.")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/METADATA
new file mode 100644
index 00000000..0c941771
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/METADATA
@@ -0,0 +1,142 @@
+Metadata-Version: 2.2
+Name: setuptools
+Version: 75.8.0
+Summary: Easily download, build, install, upgrade, and uninstall Python packages
+Author-email: Python Packaging Authority <distutils-sig@python.org>
+Project-URL: Source, https://github.com/pypa/setuptools
+Project-URL: Documentation, https://setuptools.pypa.io/
+Project-URL: Changelog, https://setuptools.pypa.io/en/stable/history.html
+Keywords: CPAN PyPI distutils eggs package management
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: System :: Archiving :: Packaging
+Classifier: Topic :: System :: Systems Administration
+Classifier: Topic :: Utilities
+Requires-Python: >=3.9
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Provides-Extra: test
+Requires-Dist: pytest!=8.1.*,>=6; extra == "test"
+Requires-Dist: virtualenv>=13.0.0; extra == "test"
+Requires-Dist: wheel>=0.44.0; extra == "test"
+Requires-Dist: pip>=19.1; extra == "test"
+Requires-Dist: packaging>=24.2; extra == "test"
+Requires-Dist: jaraco.envs>=2.2; extra == "test"
+Requires-Dist: pytest-xdist>=3; extra == "test"
+Requires-Dist: jaraco.path>=3.7.2; extra == "test"
+Requires-Dist: build[virtualenv]>=1.0.3; extra == "test"
+Requires-Dist: filelock>=3.4.0; extra == "test"
+Requires-Dist: ini2toml[lite]>=0.14; extra == "test"
+Requires-Dist: tomli-w>=1.0.0; extra == "test"
+Requires-Dist: pytest-timeout; extra == "test"
+Requires-Dist: pytest-perf; sys_platform != "cygwin" and extra == "test"
+Requires-Dist: jaraco.develop>=7.21; (python_version >= "3.9" and sys_platform != "cygwin") and extra == "test"
+Requires-Dist: pytest-home>=0.5; extra == "test"
+Requires-Dist: pytest-subprocess; extra == "test"
+Requires-Dist: pyproject-hooks!=1.1; extra == "test"
+Requires-Dist: jaraco.test>=5.5; extra == "test"
+Provides-Extra: doc
+Requires-Dist: sphinx>=3.5; extra == "doc"
+Requires-Dist: jaraco.packaging>=9.3; extra == "doc"
+Requires-Dist: rst.linker>=1.9; extra == "doc"
+Requires-Dist: furo; extra == "doc"
+Requires-Dist: sphinx-lint; extra == "doc"
+Requires-Dist: jaraco.tidelift>=1.4; extra == "doc"
+Requires-Dist: pygments-github-lexers==0.0.5; extra == "doc"
+Requires-Dist: sphinx-favicon; extra == "doc"
+Requires-Dist: sphinx-inline-tabs; extra == "doc"
+Requires-Dist: sphinx-reredirects; extra == "doc"
+Requires-Dist: sphinxcontrib-towncrier; extra == "doc"
+Requires-Dist: sphinx-notfound-page<2,>=1; extra == "doc"
+Requires-Dist: pyproject-hooks!=1.1; extra == "doc"
+Requires-Dist: towncrier<24.7; extra == "doc"
+Provides-Extra: ssl
+Provides-Extra: certs
+Provides-Extra: core
+Requires-Dist: packaging>=24.2; extra == "core"
+Requires-Dist: more_itertools>=8.8; extra == "core"
+Requires-Dist: jaraco.text>=3.7; extra == "core"
+Requires-Dist: importlib_metadata>=6; python_version < "3.10" and extra == "core"
+Requires-Dist: tomli>=2.0.1; python_version < "3.11" and extra == "core"
+Requires-Dist: wheel>=0.43.0; extra == "core"
+Requires-Dist: platformdirs>=4.2.2; extra == "core"
+Requires-Dist: jaraco.collections; extra == "core"
+Requires-Dist: jaraco.functools>=4; extra == "core"
+Requires-Dist: packaging; extra == "core"
+Requires-Dist: more_itertools; extra == "core"
+Provides-Extra: check
+Requires-Dist: pytest-checkdocs>=2.4; extra == "check"
+Requires-Dist: pytest-ruff>=0.2.1; sys_platform != "cygwin" and extra == "check"
+Requires-Dist: ruff>=0.8.0; sys_platform != "cygwin" and extra == "check"
+Provides-Extra: cover
+Requires-Dist: pytest-cov; extra == "cover"
+Provides-Extra: enabler
+Requires-Dist: pytest-enabler>=2.2; extra == "enabler"
+Provides-Extra: type
+Requires-Dist: pytest-mypy; extra == "type"
+Requires-Dist: mypy==1.14.*; extra == "type"
+Requires-Dist: importlib_metadata>=7.0.2; python_version < "3.10" and extra == "type"
+Requires-Dist: jaraco.develop>=7.21; sys_platform != "cygwin" and extra == "type"
+
+.. |pypi-version| image:: https://img.shields.io/pypi/v/setuptools.svg
+   :target: https://pypi.org/project/setuptools
+
+.. |py-version| image:: https://img.shields.io/pypi/pyversions/setuptools.svg
+
+.. |test-badge| image:: https://github.com/pypa/setuptools/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/pypa/setuptools/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. |ruff-badge| image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+   :target: https://github.com/astral-sh/ruff
+   :alt: Ruff
+
+.. |docs-badge| image:: https://img.shields.io/readthedocs/setuptools/latest.svg
+   :target: https://setuptools.pypa.io
+
+.. |skeleton-badge| image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. |codecov-badge| image:: https://img.shields.io/codecov/c/github/pypa/setuptools/master.svg?logo=codecov&logoColor=white
+   :target: https://codecov.io/gh/pypa/setuptools
+
+.. |tidelift-badge| image:: https://tidelift.com/badges/github/pypa/setuptools?style=flat
+   :target: https://tidelift.com/subscription/pkg/pypi-setuptools?utm_source=pypi-setuptools&utm_medium=readme
+
+.. |discord-badge| image:: https://img.shields.io/discord/803025117553754132
+   :target: https://discord.com/channels/803025117553754132/815945031150993468
+   :alt: Discord
+
+|pypi-version| |py-version| |test-badge| |ruff-badge| |docs-badge| |skeleton-badge| |codecov-badge| |discord-badge|
+
+See the `Quickstart <https://setuptools.pypa.io/en/latest/userguide/quickstart.html>`_
+and the `User's Guide <https://setuptools.pypa.io/en/latest/userguide/>`_ for
+instructions on how to use Setuptools.
+
+Questions and comments should be directed to `GitHub Discussions
+<https://github.com/pypa/setuptools/discussions>`_.
+Bug reports and especially tested patches may be
+submitted directly to the `bug tracker
+<https://github.com/pypa/setuptools/issues>`_.
+
+
+Code of Conduct
+===============
+
+Everyone interacting in the setuptools project's codebases, issue trackers,
+chat rooms, and fora is expected to follow the
+`PSF Code of Conduct <https://github.com/pypa/.github/blob/main/CODE_OF_CONDUCT.md>`_.
+
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+Setuptools and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-setuptools?utm_source=pypi-setuptools&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/RECORD
new file mode 100644
index 00000000..f57fbed1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/RECORD
@@ -0,0 +1,853 @@
+_distutils_hack/__init__.py,sha256=34HmvLo07j45Uvd2VR-2aRQ7lJD91sTK6zJgn5fphbQ,6755
+_distutils_hack/__pycache__/__init__.cpython-312.pyc,,
+_distutils_hack/__pycache__/override.cpython-312.pyc,,
+_distutils_hack/override.py,sha256=Eu_s-NF6VIZ4Cqd0tbbA5wtWky2IZPNd8et6GLt1mzo,44
+distutils-precedence.pth,sha256=JjjOniUA5XKl4N5_rtZmHrVp0baW_LoHsN0iPaX10iQ,151
+pkg_resources/__init__.py,sha256=fuydd-87on31-3P0wTNTZiFn0quiY18pZdP6zRXalQs,126098
+pkg_resources/__pycache__/__init__.cpython-312.pyc,,
+pkg_resources/api_tests.txt,sha256=XEdvy4igHHrq2qNHNMHnlfO6XSQKNqOyLHbl6QcpfAI,12595
+pkg_resources/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pkg_resources/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pkg_resources/tests/__pycache__/__init__.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_find_distributions.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_integration_zope_interface.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_markers.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_pkg_resources.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_resources.cpython-312.pyc,,
+pkg_resources/tests/__pycache__/test_working_set.cpython-312.pyc,,
+pkg_resources/tests/data/my-test-package-source/__pycache__/setup.cpython-312.pyc,,
+pkg_resources/tests/data/my-test-package-source/setup.cfg,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+pkg_resources/tests/data/my-test-package-source/setup.py,sha256=1VobhAZbMb7M9mfhb_NE8PwDsvukoWLs9aUAS0pYhe8,105
+pkg_resources/tests/data/my-test-package-zip/my-test-package.zip,sha256=AYRcQ39GVePPnMT8TknP1gdDHyJnXhthESmpAjnzSCI,1809
+pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/PKG-INFO,sha256=JvWv9Io2PAuYwEEw2fBW4Qc5YvdbkscpKX1kmLzsoHk,187
+pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/SOURCES.txt,sha256=4ClkH8eTovZrdVrJFsVuxdbMEF--lBVSuKonDAPE5Jc,208
+pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/dependency_links.txt,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/top_level.txt,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+pkg_resources/tests/data/my-test-package_unpacked-egg/my_test_package-1.0-py3.7.egg/EGG-INFO/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+pkg_resources/tests/data/my-test-package_zipped-egg/my_test_package-1.0-py3.7.egg,sha256=ZTlMGxjRGiKDNkiA2c75jbQH2TWIteP00irF9gvczbo,843
+pkg_resources/tests/test_find_distributions.py,sha256=U91cov5L1COAIWLNq3Xy4plU7_MnOE1WtXMu6iV2waM,1972
+pkg_resources/tests/test_integration_zope_interface.py,sha256=nzVoK557KZQN0V3DIQ1sVeaCOgt4Kpl-CODAWsO7pmc,1652
+pkg_resources/tests/test_markers.py,sha256=0orKg7UMDf7fnuNQvRMOc-EF9EAP_JTQnk4mtGgbW50,241
+pkg_resources/tests/test_pkg_resources.py,sha256=UMgFTBF8WN-yBOaWcsmKlr1wqEz7cu2B9OnPeXMmkMI,15207
+pkg_resources/tests/test_resources.py,sha256=K0LqMAUGpRQ9pUb9K0vyI7GesvtlQvTH074m-E2VQlo,31252
+pkg_resources/tests/test_working_set.py,sha256=ZUJ8Su47v1YfRDkf8B6nBPHcdChVdzgRcxppHTwJs9k,8539
+setuptools-75.8.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools-75.8.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools-75.8.0.dist-info/METADATA,sha256=ijjJeedHZjDBD3J3R9dLc6NzWwVLAMg3l_BqhSEzLkk,6670
+setuptools-75.8.0.dist-info/RECORD,,
+setuptools-75.8.0.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+setuptools-75.8.0.dist-info/entry_points.txt,sha256=zkgthpf_Fa9NVE9p6FKT3Xk9DR1faAcRU4coggsV7jA,2449
+setuptools-75.8.0.dist-info/top_level.txt,sha256=d9yL39v_W7qmKDDSH6sT4bE0j_Ls1M3P161OGgdsm4g,41
+setuptools/__init__.py,sha256=AQsMeRFWz9CQ9aBtcSkbB4Dn8t2w86-GPriklpoAjew,10406
+setuptools/__pycache__/__init__.cpython-312.pyc,,
+setuptools/__pycache__/_core_metadata.cpython-312.pyc,,
+setuptools/__pycache__/_entry_points.cpython-312.pyc,,
+setuptools/__pycache__/_imp.cpython-312.pyc,,
+setuptools/__pycache__/_importlib.cpython-312.pyc,,
+setuptools/__pycache__/_itertools.cpython-312.pyc,,
+setuptools/__pycache__/_normalization.cpython-312.pyc,,
+setuptools/__pycache__/_path.cpython-312.pyc,,
+setuptools/__pycache__/_reqs.cpython-312.pyc,,
+setuptools/__pycache__/_shutil.cpython-312.pyc,,
+setuptools/__pycache__/_static.cpython-312.pyc,,
+setuptools/__pycache__/archive_util.cpython-312.pyc,,
+setuptools/__pycache__/build_meta.cpython-312.pyc,,
+setuptools/__pycache__/depends.cpython-312.pyc,,
+setuptools/__pycache__/discovery.cpython-312.pyc,,
+setuptools/__pycache__/dist.cpython-312.pyc,,
+setuptools/__pycache__/errors.cpython-312.pyc,,
+setuptools/__pycache__/extension.cpython-312.pyc,,
+setuptools/__pycache__/glob.cpython-312.pyc,,
+setuptools/__pycache__/installer.cpython-312.pyc,,
+setuptools/__pycache__/launch.cpython-312.pyc,,
+setuptools/__pycache__/logging.cpython-312.pyc,,
+setuptools/__pycache__/modified.cpython-312.pyc,,
+setuptools/__pycache__/monkey.cpython-312.pyc,,
+setuptools/__pycache__/msvc.cpython-312.pyc,,
+setuptools/__pycache__/namespaces.cpython-312.pyc,,
+setuptools/__pycache__/package_index.cpython-312.pyc,,
+setuptools/__pycache__/sandbox.cpython-312.pyc,,
+setuptools/__pycache__/unicode_utils.cpython-312.pyc,,
+setuptools/__pycache__/version.cpython-312.pyc,,
+setuptools/__pycache__/warnings.cpython-312.pyc,,
+setuptools/__pycache__/wheel.cpython-312.pyc,,
+setuptools/__pycache__/windows_support.cpython-312.pyc,,
+setuptools/_core_metadata.py,sha256=38kBNNxBe1SQAQ-WtnH9cyo7R_YNN47AdQqd0OT0Iuo,11070
+setuptools/_distutils/__init__.py,sha256=xGYuhWwLG07J0Q49BVnEjPy6wyDcd6veJMDJX7ljlyM,359
+setuptools/_distutils/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/_log.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/_macos_compat.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/_modified.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/_msvccompiler.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/archive_util.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/ccompiler.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/cmd.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/core.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/cygwinccompiler.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/debug.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/dep_util.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/dir_util.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/dist.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/errors.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/extension.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/fancy_getopt.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/file_util.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/filelist.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/log.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/spawn.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/sysconfig.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/text_file.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/unixccompiler.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/util.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/version.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/versionpredicate.cpython-312.pyc,,
+setuptools/_distutils/__pycache__/zosccompiler.cpython-312.pyc,,
+setuptools/_distutils/_log.py,sha256=i-lNTTcXS8TmWITJ6DODGvtW5z5tMattJQ76h8rZxQU,42
+setuptools/_distutils/_macos_compat.py,sha256=JzUGhF4E5yIITHbUaPobZEWjGHdrrcNV63z86S4RjBc,239
+setuptools/_distutils/_modified.py,sha256=JZvIUKHidnO_yddOaS9oaXdSrWnyQMifataAkvpsnIU,2446
+setuptools/_distutils/_msvccompiler.py,sha256=C88msUcJV1aUqK7YBXz5ZAUbkZefg6evxhjHrTTOiOs,20798
+setuptools/_distutils/archive_util.py,sha256=15jHbLmCDcnZ75J2tFFyCmCP6yF2aWEzVz-lusaeyr4,7844
+setuptools/_distutils/ccompiler.py,sha256=OKzwMW7f05TRXAETsKg34PV1f3sRbGmAj1j1650JB-E,49190
+setuptools/_distutils/cmd.py,sha256=_LTLcweNAsCmKvUJB9FH4gCtd59_d59X0MJLHlwObuE,18668
+setuptools/_distutils/command/__init__.py,sha256=GfFAzbBqk1qxSH4BdaKioKS4hRRnD44BAmwEN85C4u8,386
+setuptools/_distutils/command/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/_framework_compat.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/bdist.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/bdist_dumb.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/bdist_rpm.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/build.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/build_clib.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/build_ext.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/build_py.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/build_scripts.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/check.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/clean.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/config.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install_data.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install_egg_info.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install_headers.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install_lib.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/install_scripts.cpython-312.pyc,,
+setuptools/_distutils/command/__pycache__/sdist.cpython-312.pyc,,
+setuptools/_distutils/command/_framework_compat.py,sha256=0iZdSJYzGRWCCvzRDKE-R0-_yaAYvFMd1ylXb2eYXug,1609
+setuptools/_distutils/command/bdist.py,sha256=utgCVN6xIUvAFArc_Z_O5TE3X4G00Cirh46QnFOaooA,5423
+setuptools/_distutils/command/bdist_dumb.py,sha256=Gi_d1Nz4l6Gz_xU4LRfRVRKBri22WoNPM7uYyX2ksdk,4582
+setuptools/_distutils/command/bdist_rpm.py,sha256=DfZgvPmm3PTAd39YzLeQ8fmbyRGaXo-nmnUzYEtccg0,21686
+setuptools/_distutils/command/build.py,sha256=ibbwCEdNbXJSlDWf01Rt6wRCsE3FmwqizGqf9kLnL5I,5768
+setuptools/_distutils/command/build_clib.py,sha256=MAJ-SZOqn0hvy0TnIbRozbeiDTiHCzLGGvCJKw21Du8,7727
+setuptools/_distutils/command/build_ext.py,sha256=LKLrELoIw2HTk2_FEBVquYC7rfqFzjMmaWSFUQENa0U,32048
+setuptools/_distutils/command/build_py.py,sha256=cNe8vwMhyPJ2gN6jot9cOY5OuUPUvT6j9fIFcC-Fcik,16552
+setuptools/_distutils/command/build_scripts.py,sha256=d1Qlv_Z3T2dEwP3nvO7xEeVjXF_VTn_JWqG68TpCa20,5600
+setuptools/_distutils/command/check.py,sha256=piX-jHyR6S1te2GKO1q-gnW3FeqM6mMyg8swnHWrNtQ,4912
+setuptools/_distutils/command/clean.py,sha256=qlKth74jWLZjKa2nxOay_2Fua6MVNTroApaQOva2dwc,2595
+setuptools/_distutils/command/config.py,sha256=FKd2vUSVOp0rpVUer4bj6D94-fyxyF8HJxitRlZFc9c,13008
+setuptools/_distutils/command/install.py,sha256=mXTwnho9e4kYICKI5-ZC4rV7JT0c6M8C8jVYAfofeF0,30122
+setuptools/_distutils/command/install_data.py,sha256=TcWFlYd5NQDefqiQVnkLPeFP0n1bexfh6IDs8k9HXTA,2810
+setuptools/_distutils/command/install_egg_info.py,sha256=ffiLoU1ivQJ8q2_WL7ZygZbUcOsgdFLKL7otEIJWWkI,2868
+setuptools/_distutils/command/install_headers.py,sha256=1QFukeAJxStck8dHiJze6p4XCSTutBlPEIMR5vzKly4,1251
+setuptools/_distutils/command/install_lib.py,sha256=pKLNE1rnqdoSo8bq9efQbQuQprg5TGswFpvKka1F3Fg,8330
+setuptools/_distutils/command/install_scripts.py,sha256=Qw8KrC24mcIeJEvcBNKISMpi75npSm6jzW6BOzA9G7g,1937
+setuptools/_distutils/command/sdist.py,sha256=FNSz7cRI4Q1v7oDHoVAxBAKdoyg6aZHdYHa25TrHm24,18837
+setuptools/_distutils/compat/__init__.py,sha256=tA0rz-pnl0LPMmeJq7yLlMwKBMvwK-FpnIwliYkK-iE,396
+setuptools/_distutils/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_distutils/compat/__pycache__/py39.cpython-312.pyc,,
+setuptools/_distutils/compat/py39.py,sha256=hOsD6lwZLqZoMnacNJ3P6nUA-LJQhEpVtYTzVH0o96M,1964
+setuptools/_distutils/core.py,sha256=GEHKaFC48T3o-_SmH4864GvKyx1IgbVC6ISIPVlx7a4,9364
+setuptools/_distutils/cygwinccompiler.py,sha256=1LQO0p-AwDSNzLJk_KPIKp62eiDpkGZ4fMMs2N3o94w,11891
+setuptools/_distutils/debug.py,sha256=N6MrTAqK6l9SVk6tWweR108PM8Ol7qNlfyV-nHcLhsY,139
+setuptools/_distutils/dep_util.py,sha256=xN75p6ZpHhMiHEc-rpL2XilJQynHnDNiafHteaZ4tjU,349
+setuptools/_distutils/dir_util.py,sha256=DXPUlfVVGsg9B-Jgg4At_j9T7vM60OgwNXkQHqTo7-I,7236
+setuptools/_distutils/dist.py,sha256=9NOaF7Tz32uA9tl1SMkwe5bb8YM6hEto01PGvuo-uDU,51529
+setuptools/_distutils/errors.py,sha256=bZ3cL1YpmYVHJYEgw8UM8vlsLMIpeAXqi6IDSV9Yqhw,3325
+setuptools/_distutils/extension.py,sha256=N6GyiMs_qxMyDjcOXuKRiEICbzFcVXbYD3DuD9_hD5k,10755
+setuptools/_distutils/fancy_getopt.py,sha256=PjdO-bWCW0imV_UN-MGEw9R2GP2OiE8pHjITgmTAY3Q,17895
+setuptools/_distutils/file_util.py,sha256=YFQL_pD3hLuER9II_H6-hDC_YIGEookdd4wedLuiTW0,7978
+setuptools/_distutils/filelist.py,sha256=DX3dWnDTAmEk5e7SkGYdX6RJHFCzZFH8qtIm7qFAabE,13567
+setuptools/_distutils/log.py,sha256=VyBs5j7z4-K6XTEEBThUc9HyMpoPLGtQpERqbz5ylww,1200
+setuptools/_distutils/spawn.py,sha256=kCtEAFJgPpSZXbozrVfuB7B4omYUhr-hLaukpuXGXYg,3634
+setuptools/_distutils/sysconfig.py,sha256=M_9pKSCgIXa_etvIzzh2AOuvBLAIWB70Wmn0ET4haKk,19266
+setuptools/_distutils/tests/__init__.py,sha256=j-IoPZEtQv3EOPuqNTwalr6GLyRjzCC-OOaNvZzmHsI,1485
+setuptools/_distutils/tests/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/support.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_archive_util.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_bdist.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_bdist_dumb.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_bdist_rpm.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_build.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_build_clib.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_build_ext.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_build_py.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_build_scripts.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_ccompiler.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_check.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_clean.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_cmd.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_config_cmd.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_core.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_cygwinccompiler.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_dir_util.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_dist.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_extension.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_file_util.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_filelist.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_install.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_install_data.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_install_headers.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_install_lib.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_install_scripts.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_log.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_mingwccompiler.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_modified.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_msvccompiler.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_sdist.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_spawn.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_sysconfig.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_text_file.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_unixccompiler.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_util.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_version.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/test_versionpredicate.cpython-312.pyc,,
+setuptools/_distutils/tests/__pycache__/unix_compat.cpython-312.pyc,,
+setuptools/_distutils/tests/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_distutils/tests/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_distutils/tests/compat/__pycache__/py39.cpython-312.pyc,,
+setuptools/_distutils/tests/compat/py39.py,sha256=t0GBTM-30jX-9zCfkwlNBFtzzabemx6065mJ0d9_VRw,1026
+setuptools/_distutils/tests/support.py,sha256=tjsYsyxvpTK4NrkCseh2ujvDIGV0Mf_b5SI5fP2T0yM,4099
+setuptools/_distutils/tests/test_archive_util.py,sha256=jozimSwPBF-JoJfN_vDaiVGZp66BNcWZGh34FlW57DQ,11787
+setuptools/_distutils/tests/test_bdist.py,sha256=xNHxUsLlHsZQRwkzLb_iSD24s-9Mk-NX2ffBWwOyPyc,1396
+setuptools/_distutils/tests/test_bdist_dumb.py,sha256=QF05MHNhPOdZyh88Xpw8KsO64s7pRFkl8KL-RoV4XK0,2247
+setuptools/_distutils/tests/test_bdist_rpm.py,sha256=Hdm-pwWgyaoGdGbEcGZa8cRhGU45y8gHK8umOanTjik,3932
+setuptools/_distutils/tests/test_build.py,sha256=JJY5XpOZco25ZY0pstxl-iI8mHsWP0ujf5o8aOtuZYY,1742
+setuptools/_distutils/tests/test_build_clib.py,sha256=Mo1ZFb4C1VXBYOGvnallwN7YCnTtr24akLDO8Zi4CsY,4331
+setuptools/_distutils/tests/test_build_ext.py,sha256=I6HUzX6MgDW25v4B39XBLqPfVT8SyGOl1W-32ehIZOM,19954
+setuptools/_distutils/tests/test_build_py.py,sha256=NsfmRrojOHBXNMqWR_mp5g4PLTgjhD7iZFUffGZFIdw,6882
+setuptools/_distutils/tests/test_build_scripts.py,sha256=cD-FRy-oX55sXRX5Ez5xQCaeHrWajyKc4Xuwv2fe48w,2880
+setuptools/_distutils/tests/test_ccompiler.py,sha256=uLvdaaE18FJ2nTH-MQ67IPknqYaG1OQ-z00ok0JjwMA,3026
+setuptools/_distutils/tests/test_check.py,sha256=hHSV07qf7YoSxGsTbbsUQ9tssZz5RRNdbrY1s2SwaFI,6226
+setuptools/_distutils/tests/test_clean.py,sha256=hPH6jfIpGFUrvWbF1txkiNVSNaAxt2wq5XjV499zO4E,1240
+setuptools/_distutils/tests/test_cmd.py,sha256=bgRB79mitoOKR1OiyZHnCogvGxt3pWkxeTqIC04lQWQ,3254
+setuptools/_distutils/tests/test_config_cmd.py,sha256=Zs6WX0IfxDvmuC19XzuVNnYCnTr9Y-hl73TAmDSBN4Y,2664
+setuptools/_distutils/tests/test_core.py,sha256=L7XKVAxa-MGoAZeANopnuK9fRKneYhkSQpgw8XQvcF8,3829
+setuptools/_distutils/tests/test_cygwinccompiler.py,sha256=iqxsDy0Z5ZTRgxM8ARzPXakitQod2V8aG5qet_J5tTg,2753
+setuptools/_distutils/tests/test_dir_util.py,sha256=E84lC-k4riVUwURyWaQ0Jqx2ui2-io-0RuJa3M7qkJs,4500
+setuptools/_distutils/tests/test_dist.py,sha256=a6wlc5fQJd5qQ6HOndzcupNhjTxvj6-_JLtpuYvaP1M,18793
+setuptools/_distutils/tests/test_extension.py,sha256=-YejLgZCuycFrOBd64pVH0JvwMc9NwhzHvQxvvjXHqk,3670
+setuptools/_distutils/tests/test_file_util.py,sha256=livjnl3FkilQlrB2rFdFQq9nvjEVZHynNya0bfzv_b4,3522
+setuptools/_distutils/tests/test_filelist.py,sha256=rJwkqCUfkGDgWlD22TozsT8ycbupMHB8DXqThzwT1T4,10766
+setuptools/_distutils/tests/test_install.py,sha256=TfCB0ykhIxydIC2Q4SuTAZzSHvteMHgrBL9whoSgK9Q,8618
+setuptools/_distutils/tests/test_install_data.py,sha256=vKq3K97k0hBAnOg38nmwEdf7cEDVr9rTVyCeJolgb4A,2464
+setuptools/_distutils/tests/test_install_headers.py,sha256=PVAYpo_tYl980Qf64DPOmmSvyefIHdU06f7VsJeZykE,936
+setuptools/_distutils/tests/test_install_lib.py,sha256=qri6Rl-maNTQrNDV8DbeXNl0hjsfRIKiI4rfZLrmWBI,3612
+setuptools/_distutils/tests/test_install_scripts.py,sha256=KE3v0cDkFW-90IOID-OmZZGM2mhy-ZkEuuW7UXS2SHw,1600
+setuptools/_distutils/tests/test_log.py,sha256=isFtOufloCyEdZaQOV7cVUr46GwtdVMj43mGBB5XH7k,323
+setuptools/_distutils/tests/test_mingwccompiler.py,sha256=mBl8W8QIO2xy4eOj6aAEVom4lobwpHM-HvFUIXu6q0c,2202
+setuptools/_distutils/tests/test_modified.py,sha256=h1--bOWmtJo1bpVV6uRhdnS9br71CBiNDM1MDwSGpug,4221
+setuptools/_distutils/tests/test_msvccompiler.py,sha256=xUrfyCwCO57DEsxcoL6s-YG3YIZRllYsKuagZbBPFJ0,4301
+setuptools/_distutils/tests/test_sdist.py,sha256=InsbU09aeVdJtj8QywREnUZuV32bvi47eRmH3iWQwkk,15058
+setuptools/_distutils/tests/test_spawn.py,sha256=eS8w9D7bTxyFLSyRahJWeuh8Kc1F8RWWiY_dSG5B5Bc,4803
+setuptools/_distutils/tests/test_sysconfig.py,sha256=lxM8LsUi1TomjDV4HoYK8u5nUoBkeNL60Uq8PY1DcwU,11986
+setuptools/_distutils/tests/test_text_file.py,sha256=WQWSB5AfdBDZaMA8BFgipJPnsJb_2SKMfL90fSkRVtw,3460
+setuptools/_distutils/tests/test_unixccompiler.py,sha256=WwCE0zPYNE2nVSDuTYJwtR-BpmE0tuTJlCNQfeV_JOI,11835
+setuptools/_distutils/tests/test_util.py,sha256=H9zlZ4z4Vh4TfjNYDBsxP7wguQLpxCfJYyOcm1yZU3c,7988
+setuptools/_distutils/tests/test_version.py,sha256=ahfg_mP8wRy1sgwY-_Px5hrjgf6_upTIpnCgpR4yWRk,2750
+setuptools/_distutils/tests/test_versionpredicate.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_distutils/tests/unix_compat.py,sha256=z-op6C2iVdX1aq5BIBR7cqOxijKE97alNwJqHNdLpoI,386
+setuptools/_distutils/text_file.py,sha256=z4dkOJBr9Bo2LG0TNqm8sD63LEEaKSSP0J0bWBrFG3c,12101
+setuptools/_distutils/unixccompiler.py,sha256=QJHNcQiMtWcOQ4WzupzIv1nQwBENo-bNkeVCSVmT4Jk,15437
+setuptools/_distutils/util.py,sha256=bXhftuz61oszGvwzXUtv9kteGNe-7YmFTEu33D6pQhc,17493
+setuptools/_distutils/version.py,sha256=vImT5-ECXkQ21oKL0XYFiTqK6NyM09cpzBNoA_34CQU,12619
+setuptools/_distutils/versionpredicate.py,sha256=qBWQ6wTj12ODytoTmIydefIY2jb4uY1sdbgbuLn-IJM,5205
+setuptools/_distutils/zosccompiler.py,sha256=bb2dQoGnsv4LmoQBfjhDsaOpt_p5R7y_28l1cltmG94,6589
+setuptools/_entry_points.py,sha256=Y3QUE9JKFW_YyquDnpffNWSs6f3jKEt1e-dnx--9-Kw,2310
+setuptools/_imp.py,sha256=YY1EjZEN-0zYci1cxO10B_adAEOr7i8eK8JoCc9Ierc,2435
+setuptools/_importlib.py,sha256=aKIjcK0HKXNz2D-XTrxaixGn_juTkONwmu3dcheMOF0,223
+setuptools/_itertools.py,sha256=jWRfsIrpC7myooz3hDURj9GtvpswZeKXg2HakmEhNjo,657
+setuptools/_normalization.py,sha256=t-SeXc0jU2weQY9BA3qGlRThzBND1oYK5Hpzg1_533g,4536
+setuptools/_path.py,sha256=cPv41v03HD7uEYqCIo-E_cGRfpPVr4lywBCiK-HSrCg,2685
+setuptools/_reqs.py,sha256=QI3C9uOBSNRccu208qPnixHx51nxCry7_nPTIJaSYxM,1438
+setuptools/_shutil.py,sha256=cAOllcoyMTXs5JLoybQi29yI5gABk82hepJyOBv2bMw,1496
+setuptools/_static.py,sha256=w10bv06NSpOFhwCNC3FQkeVNzciZvIzNZp8NX301bhs,4855
+setuptools/_vendor/__pycache__/typing_extensions.cpython-312.pyc,,
+setuptools/_vendor/autocommand-2.2.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/autocommand-2.2.2.dist-info/LICENSE,sha256=reeNBJgtaZctREqOFKlPh6IzTdOFXMgDSOqOJAqg3y0,7634
+setuptools/_vendor/autocommand-2.2.2.dist-info/METADATA,sha256=OADZuR3O6iBlpu1ieTgzYul6w4uOVrk0P0BO5TGGAJk,15006
+setuptools/_vendor/autocommand-2.2.2.dist-info/RECORD,sha256=giu6ZrQVJvpUcYa4AiH4XaUNZSvuVJPb_l0UCFES8MM,1308
+setuptools/_vendor/autocommand-2.2.2.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+setuptools/_vendor/autocommand-2.2.2.dist-info/top_level.txt,sha256=AzfhgKKS8EdAwWUTSF8mgeVQbXOY9kokHB6kSqwwqu0,12
+setuptools/_vendor/autocommand/__init__.py,sha256=zko5Rnvolvb-UXjCx_2ArPTGBWwUK5QY4LIQIKYR7As,1037
+setuptools/_vendor/autocommand/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/autocommand/__pycache__/autoasync.cpython-312.pyc,,
+setuptools/_vendor/autocommand/__pycache__/autocommand.cpython-312.pyc,,
+setuptools/_vendor/autocommand/__pycache__/automain.cpython-312.pyc,,
+setuptools/_vendor/autocommand/__pycache__/autoparse.cpython-312.pyc,,
+setuptools/_vendor/autocommand/__pycache__/errors.cpython-312.pyc,,
+setuptools/_vendor/autocommand/autoasync.py,sha256=AMdyrxNS4pqWJfP_xuoOcImOHWD-qT7x06wmKN1Vp-U,5680
+setuptools/_vendor/autocommand/autocommand.py,sha256=hmkEmQ72HtL55gnURVjDOnsfYlGd5lLXbvT4KG496Qw,2505
+setuptools/_vendor/autocommand/automain.py,sha256=A2b8i754Mxc_DjU9WFr6vqYDWlhz0cn8miu8d8EsxV8,2076
+setuptools/_vendor/autocommand/autoparse.py,sha256=WVWmZJPcbzUKXP40raQw_0HD8qPJ2V9VG1eFFmmnFxw,11642
+setuptools/_vendor/autocommand/errors.py,sha256=7aa3roh9Herd6nIKpQHNWEslWE8oq7GiHYVUuRqORnA,886
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/METADATA,sha256=ghXFTq132dxaEIolxr3HK1mZqm9iyUmaRANZQSr6WlE,2020
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/RECORD,sha256=JYofHISeEXUGmlWl1s41ev3QTjTNXeJwk-Ss7HqdLOE,1360
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/backports.tarfile-1.2.0.dist-info/top_level.txt,sha256=cGjaLMOoBR1FK0ApojtzWVmViTtJ7JGIK_HwXiEsvtU,10
+setuptools/_vendor/backports/__init__.py,sha256=iOEMwnlORWezdO8-2vxBIPSR37D7JGjluZ8f55vzxls,81
+setuptools/_vendor/backports/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/backports/tarfile/__init__.py,sha256=Pwf2qUIfB0SolJPCKcx3vz3UEu_aids4g4sAfxy94qg,108491
+setuptools/_vendor/backports/tarfile/__main__.py,sha256=Yw2oGT1afrz2eBskzdPYL8ReB_3liApmhFkN2EbDmc4,59
+setuptools/_vendor/backports/tarfile/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/backports/tarfile/__pycache__/__main__.cpython-312.pyc,,
+setuptools/_vendor/backports/tarfile/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/backports/tarfile/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/backports/tarfile/compat/__pycache__/py38.cpython-312.pyc,,
+setuptools/_vendor/backports/tarfile/compat/py38.py,sha256=iYkyt_gvWjLzGUTJD9TuTfMMjOk-ersXZmRlvQYN2qE,568
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/METADATA,sha256=anuQ7_7h4J1bSEzfcjIBakPi2cyVQ7y7jklLHsBeH1k,4648
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/RECORD,sha256=DY08buueu-hsrH1ghhVSQzwynanqUSSLYdAr4uXmQDA,2518
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/WHEEL,sha256=mguMlWGMX-VHnMpKOjjQidIo1ssRlCFu4a4mBpz1s2M,91
+setuptools/_vendor/importlib_metadata-8.0.0.dist-info/top_level.txt,sha256=CO3fD9yylANiXkrMo4qHLV_mqXL2sC5JFKgt1yWAT-A,19
+setuptools/_vendor/importlib_metadata/__init__.py,sha256=tZNB-23h8Bixi9uCrQqj9Yf0aeC--Josdy3IZRIQeB0,33798
+setuptools/_vendor/importlib_metadata/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_adapters.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_collections.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_compat.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_functools.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_itertools.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_meta.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/_text.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/__pycache__/diagnose.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/_adapters.py,sha256=rIhWTwBvYA1bV7i-5FfVX38qEXDTXFeS5cb5xJtP3ks,2317
+setuptools/_vendor/importlib_metadata/_collections.py,sha256=CJ0OTCHIjWA0ZIVS4voORAsn2R4R2cQBEtPsZEJpASY,743
+setuptools/_vendor/importlib_metadata/_compat.py,sha256=73QKrN9KNoaZzhbX5yPCCZa-FaALwXe8TPlDR72JgBU,1314
+setuptools/_vendor/importlib_metadata/_functools.py,sha256=PsY2-4rrKX4RVeRC1oGp1lB1pmC9eKN88_f-bD9uOoA,2895
+setuptools/_vendor/importlib_metadata/_itertools.py,sha256=cvr_2v8BRbxcIl5x5ldfqdHjhI8Yi8s8yk50G_nm6jQ,2068
+setuptools/_vendor/importlib_metadata/_meta.py,sha256=nxZ7C8GVlcBFAKWyVOn_dn7ot_twBcbm1NmvjIetBHI,1801
+setuptools/_vendor/importlib_metadata/_text.py,sha256=HCsFksZpJLeTP3NEk_ngrAeXVRRtTrtyh9eOABoRP4A,2166
+setuptools/_vendor/importlib_metadata/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/importlib_metadata/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/compat/__pycache__/py311.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/compat/__pycache__/py39.cpython-312.pyc,,
+setuptools/_vendor/importlib_metadata/compat/py311.py,sha256=uqm-K-uohyj1042TH4a9Er_I5o7667DvulcD-gC_fSA,608
+setuptools/_vendor/importlib_metadata/compat/py39.py,sha256=cPkMv6-0ilK-0Jw_Tkn0xYbOKJZc4WJKQHow0c2T44w,1102
+setuptools/_vendor/importlib_metadata/diagnose.py,sha256=nkSRMiowlmkhLYhKhvCg9glmt_11Cox-EmLzEbqYTa8,379
+setuptools/_vendor/importlib_metadata/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/inflect-7.3.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/inflect-7.3.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/inflect-7.3.1.dist-info/METADATA,sha256=ZgMNY0WAZRs-U8wZiV2SMfjSKqBrMngXyDMs_CAwMwg,21079
+setuptools/_vendor/inflect-7.3.1.dist-info/RECORD,sha256=XXg0rBuiYSxoAQUP3lenuYsPNqz4jDwtTzdv2JEbMJE,943
+setuptools/_vendor/inflect-7.3.1.dist-info/WHEEL,sha256=y4mX-SOX4fYIkonsAGA5N0Oy-8_gI4FXw5HNI1xqvWg,91
+setuptools/_vendor/inflect-7.3.1.dist-info/top_level.txt,sha256=m52ujdp10CqT6jh1XQxZT6kEntcnv-7Tl7UiGNTzWZA,8
+setuptools/_vendor/inflect/__init__.py,sha256=Jxy1HJXZiZ85kHeLAhkmvz6EMTdFqBe-duvt34R6IOc,103796
+setuptools/_vendor/inflect/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/inflect/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/inflect/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/inflect/compat/__pycache__/py38.cpython-312.pyc,,
+setuptools/_vendor/inflect/compat/py38.py,sha256=oObVfVnWX9_OpnOuEJn1mFbJxVhwyR5epbiTNXDDaso,160
+setuptools/_vendor/inflect/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/METADATA,sha256=IMUaliNsA5X1Ox9MXUWOagch5R4Wwb_3M7erp29dBtg,3933
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/RECORD,sha256=HptivXDkpfom6VlMu4CGD_7KPev-6Hc9rvp3TNJZygY,873
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/WHEEL,sha256=Mdi9PDNwEZptOjTlUcAth7XJDFtKrHYaQMPulZeBCiQ,91
+setuptools/_vendor/jaraco.collections-5.1.0.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/METADATA,sha256=xDtguJej0tN9iEXCUvxEJh2a7xceIRVBEakBLSr__tY,4020
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/RECORD,sha256=VRl7iKeEQyl7stgnp1uq50CzOJYlHYcoNdS0x17C9X4,641
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/jaraco.context-5.3.0.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/METADATA,sha256=i4aUaQDX-jjdEQK5wevhegyx8JyLfin2HyvaSk3FHso,2891
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/RECORD,sha256=YyqnwE98S8wBwCevW5vHb-iVj0oYEDW5V6O9MBS6JIs,843
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/jaraco.functools-4.0.1.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/METADATA,sha256=AzWdm6ViMfDOPoQMfLWn2zgBQSGJScyqeN29TcuWXVI,3658
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/RECORD,sha256=gW2UV0HcokYJk4jKPu10_AZnrLqjb3C1WbJJTDl5sfY,1500
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/jaraco.text-3.12.1.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+setuptools/_vendor/jaraco/__pycache__/context.cpython-312.pyc,,
+setuptools/_vendor/jaraco/collections/__init__.py,sha256=Pc1-SqjWm81ad1P0-GttpkwO_LWlnaY6gUq8gcKh2v0,26640
+setuptools/_vendor/jaraco/collections/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/jaraco/collections/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/jaraco/context.py,sha256=REoLIxDkO5MfEYowt_WoupNCRoxBS5v7YX2PbW8lIcs,9552
+setuptools/_vendor/jaraco/functools/__init__.py,sha256=hEAJaS2uSZRuF_JY4CxCHIYh79ZpxaPp9OiHyr9EJ1w,16642
+setuptools/_vendor/jaraco/functools/__init__.pyi,sha256=gk3dsgHzo5F_U74HzAvpNivFAPCkPJ1b2-yCd62dfnw,3878
+setuptools/_vendor/jaraco/functools/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/jaraco/functools/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/jaraco/text/Lorem ipsum.txt,sha256=N_7c_79zxOufBY9HZ3yzMgOkNv-TkOTTio4BydrSjgs,1335
+setuptools/_vendor/jaraco/text/__init__.py,sha256=Y2YUqXR_orUoDaY4SkPRe6ZZhb5HUHB_Ah9RCNsVyho,16250
+setuptools/_vendor/jaraco/text/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/__pycache__/layouts.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/__pycache__/show-newlines.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/__pycache__/strip-prefix.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/__pycache__/to-dvorak.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/__pycache__/to-qwerty.cpython-312.pyc,,
+setuptools/_vendor/jaraco/text/layouts.py,sha256=HTC8aSTLZ7uXipyOXapRMC158juecjK6RVwitfmZ9_w,643
+setuptools/_vendor/jaraco/text/show-newlines.py,sha256=WGQa65e8lyhb92LUOLqVn6KaCtoeVgVws6WtSRmLk6w,904
+setuptools/_vendor/jaraco/text/strip-prefix.py,sha256=NfVXV8JVNo6nqcuYASfMV7_y4Eo8zMQqlCOGvAnRIVw,412
+setuptools/_vendor/jaraco/text/to-dvorak.py,sha256=1SNcbSsvISpXXg-LnybIHHY-RUFOQr36zcHkY1pWFqw,119
+setuptools/_vendor/jaraco/text/to-qwerty.py,sha256=s4UMQUnPwFn_dB5uZC27BurHOQcYondBfzIpVL5pEzw,119
+setuptools/_vendor/more_itertools-10.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/more_itertools-10.3.0.dist-info/LICENSE,sha256=CfHIyelBrz5YTVlkHqm4fYPAyw_QB-te85Gn4mQ8GkY,1053
+setuptools/_vendor/more_itertools-10.3.0.dist-info/METADATA,sha256=BFO90O-fLNiVQMpj7oIS5ztzgJUUQZ3TA32P5HH3N-A,36293
+setuptools/_vendor/more_itertools-10.3.0.dist-info/RECORD,sha256=d8jnPgGNwP1-ntbICwWkQEVF9kH7CFIgzkKzaLWao9M,1259
+setuptools/_vendor/more_itertools-10.3.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/more_itertools-10.3.0.dist-info/WHEEL,sha256=rSgq_JpHF9fHR1lx53qwg_1-2LypZE_qmcuXbVUq948,81
+setuptools/_vendor/more_itertools/__init__.py,sha256=dtAbGjTDmn_ghiU5YXfhyDy0phAlXVdt5klZA5fUa-Q,149
+setuptools/_vendor/more_itertools/__init__.pyi,sha256=5B3eTzON1BBuOLob1vCflyEb2lSd6usXQQ-Cv-hXkeA,43
+setuptools/_vendor/more_itertools/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/more_itertools/__pycache__/more.cpython-312.pyc,,
+setuptools/_vendor/more_itertools/__pycache__/recipes.cpython-312.pyc,,
+setuptools/_vendor/more_itertools/more.py,sha256=1E5kzFncRKTDw0cYv1yRXMgDdunstLQd1QStcnL6U90,148370
+setuptools/_vendor/more_itertools/more.pyi,sha256=iXXeqt48Nxe8VGmIWpkVXuKpR2FYNuu2DU8nQLWCCu0,21484
+setuptools/_vendor/more_itertools/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/more_itertools/recipes.py,sha256=WedhhfhGVgr6zii8fIbGJVmRTw0ZKRiLKnYBDGJv4nY,28591
+setuptools/_vendor/more_itertools/recipes.pyi,sha256=T_mdGpcFdfrP3JSWbwzYP9JyNV-Go-7RPfpxfftAWlA,4617
+setuptools/_vendor/packaging-24.2.dist-info/INSTALLER,sha256=5hhM4Q4mYTT9z6QB6PGpUAW81PGNFrYrdXMj4oM_6ak,2
+setuptools/_vendor/packaging-24.2.dist-info/LICENSE,sha256=ytHvW9NA1z4HS6YU0m996spceUDD2MNIUuZcSQlobEg,197
+setuptools/_vendor/packaging-24.2.dist-info/LICENSE.APACHE,sha256=DVQuDIgE45qn836wDaWnYhSdxoLXgpRRKH4RuTjpRZQ,10174
+setuptools/_vendor/packaging-24.2.dist-info/LICENSE.BSD,sha256=tw5-m3QvHMb5SLNMFqo5_-zpQZY2S8iP8NIYDwAo-sU,1344
+setuptools/_vendor/packaging-24.2.dist-info/METADATA,sha256=ohH86s6k5mIfQxY2TS0LcSfADeOFa4BiCC-bxZV-pNs,3204
+setuptools/_vendor/packaging-24.2.dist-info/RECORD,sha256=Y4DrXM0KY0ArfzhbAEa1LYFPwW3WEgEeL4iCqXe-A-M,2009
+setuptools/_vendor/packaging-24.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/packaging-24.2.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+setuptools/_vendor/packaging/__init__.py,sha256=dk4Ta_vmdVJxYHDcfyhvQNw8V3PgSBomKNXqg-D2JDY,494
+setuptools/_vendor/packaging/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_parser.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_structures.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/markers.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/metadata.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/requirements.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/tags.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/utils.cpython-312.pyc,,
+setuptools/_vendor/packaging/__pycache__/version.cpython-312.pyc,,
+setuptools/_vendor/packaging/_elffile.py,sha256=cflAQAkE25tzhYmq_aCi72QfbT_tn891tPzfpbeHOwE,3306
+setuptools/_vendor/packaging/_manylinux.py,sha256=vl5OCoz4kx80H5rwXKeXWjl9WNISGmr4ZgTpTP9lU9c,9612
+setuptools/_vendor/packaging/_musllinux.py,sha256=p9ZqNYiOItGee8KcZFeHF_YcdhVwGHdK6r-8lgixvGQ,2694
+setuptools/_vendor/packaging/_parser.py,sha256=s_TvTvDNK0NrM2QB3VKThdWFM4Nc0P6JnkObkl3MjpM,10236
+setuptools/_vendor/packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+setuptools/_vendor/packaging/_tokenizer.py,sha256=J6v5H7Jzvb-g81xp_2QACKwO7LxHQA6ikryMU7zXwN8,5273
+setuptools/_vendor/packaging/licenses/__init__.py,sha256=1x5M1nEYjcgwEbLt0dXwz2ukjr18DiCzC0sraQqJ-Ww,5715
+setuptools/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc,,
+setuptools/_vendor/packaging/licenses/_spdx.py,sha256=oAm1ztPFwlsmCKe7lAAsv_OIOfS1cWDu9bNBkeu-2ns,48398
+setuptools/_vendor/packaging/markers.py,sha256=c89TNzB7ZdGYhkovm6PYmqGyHxXlYVaLW591PHUNKD8,10561
+setuptools/_vendor/packaging/metadata.py,sha256=YJibM7GYe4re8-0a3OlXmGS-XDgTEoO4tlBt2q25Bng,34762
+setuptools/_vendor/packaging/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/packaging/requirements.py,sha256=gYyRSAdbrIyKDY66ugIDUQjRMvxkH2ALioTmX3tnL6o,2947
+setuptools/_vendor/packaging/specifiers.py,sha256=GG1wPNMcL0fMJO68vF53wKMdwnfehDcaI-r9NpTfilA,40074
+setuptools/_vendor/packaging/tags.py,sha256=CFqrJzAzc2XNGexerH__T-Y5Iwq7WbsYXsiLERLWxY0,21014
+setuptools/_vendor/packaging/utils.py,sha256=0F3Hh9OFuRgrhTgGZUl5K22Fv1YP2tZl1z_2gO6kJiA,5050
+setuptools/_vendor/packaging/version.py,sha256=olfyuk_DPbflNkJ4wBWetXQ17c74x3DB501degUv7DY,16676
+setuptools/_vendor/platformdirs-4.2.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/platformdirs-4.2.2.dist-info/METADATA,sha256=zmsie01G1MtXR0wgIv5XpVeTO7idr0WWvfmxKsKWuGk,11429
+setuptools/_vendor/platformdirs-4.2.2.dist-info/RECORD,sha256=TCEddtQu1A78Os_Mhm2JEqcYr7yit-UYSUQjZtbpn-g,1642
+setuptools/_vendor/platformdirs-4.2.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/platformdirs-4.2.2.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
+setuptools/_vendor/platformdirs-4.2.2.dist-info/licenses/LICENSE,sha256=KeD9YukphQ6G6yjD_czwzv30-pSHkBHP-z0NS-1tTbY,1089
+setuptools/_vendor/platformdirs/__init__.py,sha256=EMGE8qeHRR9CzDFr8kL3tA8hdZZniYjXBVZd0UGTWK0,22225
+setuptools/_vendor/platformdirs/__main__.py,sha256=HnsUQHpiBaiTxwcmwVw-nFaPdVNZtQIdi1eWDtI-MzI,1493
+setuptools/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/android.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/api.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/version.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc,,
+setuptools/_vendor/platformdirs/android.py,sha256=xZXY9Jd46WOsxT2U6-5HsNtDZ-IQqxcEUrBLl3hYk4o,9016
+setuptools/_vendor/platformdirs/api.py,sha256=QBYdUac2eC521ek_y53uD1Dcq-lJX8IgSRVd4InC6uc,8996
+setuptools/_vendor/platformdirs/macos.py,sha256=wftsbsvq6nZ0WORXSiCrZNkRHz_WKuktl0a6mC7MFkI,5580
+setuptools/_vendor/platformdirs/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/platformdirs/unix.py,sha256=Cci9Wqt35dAMsg6HT9nRGHSBW5obb0pR3AE1JJnsCXg,10643
+setuptools/_vendor/platformdirs/version.py,sha256=r7F76tZRjgQKzrpx_I0_ZMQOMU-PS7eGnHD7zEK3KB0,411
+setuptools/_vendor/platformdirs/windows.py,sha256=IFpiohUBwxPtCzlyKwNtxyW4Jk8haa6W8o59mfrDXVo,10125
+setuptools/_vendor/tomli-2.0.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/tomli-2.0.1.dist-info/LICENSE,sha256=uAgWsNUwuKzLTCIReDeQmEpuO2GSLCte6S8zcqsnQv4,1072
+setuptools/_vendor/tomli-2.0.1.dist-info/METADATA,sha256=zPDceKmPwJGLWtZykrHixL7WVXWmJGzZ1jyRT5lCoPI,8875
+setuptools/_vendor/tomli-2.0.1.dist-info/RECORD,sha256=DLn5pFGh42WsVLTIhmLh2gy1SnLRalJY-wq_-dPhwCI,999
+setuptools/_vendor/tomli-2.0.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/tomli-2.0.1.dist-info/WHEEL,sha256=jPMR_Dzkc4X4icQtmz81lnNY_kAsfog7ry7qoRvYLXw,81
+setuptools/_vendor/tomli/__init__.py,sha256=JhUwV66DB1g4Hvt1UQCVMdfCu-IgAV8FXmvDU9onxd4,396
+setuptools/_vendor/tomli/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/tomli/__pycache__/_parser.cpython-312.pyc,,
+setuptools/_vendor/tomli/__pycache__/_re.cpython-312.pyc,,
+setuptools/_vendor/tomli/__pycache__/_types.cpython-312.pyc,,
+setuptools/_vendor/tomli/_parser.py,sha256=g9-ENaALS-B8dokYpCuzUFalWlog7T-SIYMjLZSWrtM,22633
+setuptools/_vendor/tomli/_re.py,sha256=dbjg5ChZT23Ka9z9DHOXfdtSpPwUfdgMXnj8NOoly-w,2943
+setuptools/_vendor/tomli/_types.py,sha256=-GTG2VUqkpxwMqzmVO4F7ybKddIbAnuAHXfmWQcTi3Q,254
+setuptools/_vendor/tomli/py.typed,sha256=8PjyZ1aVoQpRVvt71muvuq5qE-jTFZkK-GLHkhdebmc,26
+setuptools/_vendor/typeguard-4.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/typeguard-4.3.0.dist-info/LICENSE,sha256=YWP3mH37ONa8MgzitwsvArhivEESZRbVUu8c1DJH51g,1130
+setuptools/_vendor/typeguard-4.3.0.dist-info/METADATA,sha256=z2dcHAp0TwhYCFU5Deh8x31nazElgujUz9tbuP0pjSE,3717
+setuptools/_vendor/typeguard-4.3.0.dist-info/RECORD,sha256=SKUZWVgkeDUidUKM7s1473fXmsna55bjmi6vJUAoJVI,2402
+setuptools/_vendor/typeguard-4.3.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/typeguard-4.3.0.dist-info/entry_points.txt,sha256=qp7NQ1aLtiSgMQqo6gWlfGpy0IIXzoMJmeQTLpzqFZQ,48
+setuptools/_vendor/typeguard-4.3.0.dist-info/top_level.txt,sha256=4z28AhuDodwRS_c1J_l8H51t5QuwfTseskYzlxp6grs,10
+setuptools/_vendor/typeguard/__init__.py,sha256=Onh4w38elPCjtlcU3JY9k3h70NjsxXIkAflmQn-Z0FY,2071
+setuptools/_vendor/typeguard/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_checkers.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_config.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_decorators.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_exceptions.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_functions.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_importhook.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_memo.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_pytest_plugin.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_suppression.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_transformer.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_union_transformer.cpython-312.pyc,,
+setuptools/_vendor/typeguard/__pycache__/_utils.cpython-312.pyc,,
+setuptools/_vendor/typeguard/_checkers.py,sha256=JRrgKicdOEfIBoNEtegYCEIlhpad-a1u1Em7GCj0WCI,31360
+setuptools/_vendor/typeguard/_config.py,sha256=nIz8QwDa-oFO3L9O8_6srzlmd99pSby2wOM4Wb7F_B0,2846
+setuptools/_vendor/typeguard/_decorators.py,sha256=v6dsIeWvPhExGLP_wXF-RmDUyjZf_Ak28g7gBJ_v0-0,9033
+setuptools/_vendor/typeguard/_exceptions.py,sha256=ZIPeiV-FBd5Emw2EaWd2Fvlsrwi4ocwT2fVGBIAtHcQ,1121
+setuptools/_vendor/typeguard/_functions.py,sha256=ibgSAKa5ptIm1eR9ARG0BSozAFJPFNASZqhPVyQeqig,10393
+setuptools/_vendor/typeguard/_importhook.py,sha256=ugjCDvFcdWMU7UugqlJG91IpVNpEIxtRr-99s0h1k7M,6389
+setuptools/_vendor/typeguard/_memo.py,sha256=1juQV_vxnD2JYKbSrebiQuj4oKHz6n67v9pYA-CCISg,1303
+setuptools/_vendor/typeguard/_pytest_plugin.py,sha256=-fcSqkv54rIfIF8pDavY5YQPkj4OX8GMt_lL7CQSD4I,4416
+setuptools/_vendor/typeguard/_suppression.py,sha256=VQfzxcwIbu3if0f7VBkKM7hkYOA7tNFw9a7jMBsmMg4,2266
+setuptools/_vendor/typeguard/_transformer.py,sha256=9Ha7_QhdwoUni_6hvdY-hZbuEergowHrNL2vzHIakFY,44937
+setuptools/_vendor/typeguard/_union_transformer.py,sha256=v_42r7-6HuRX2SoFwnyJ-E5PlxXpVeUJPJR1-HU9qSo,1354
+setuptools/_vendor/typeguard/_utils.py,sha256=5HhO1rPn5f1M6ymkVAEv7Xmlz1cX-j0OnTMlyHqqrR8,5270
+setuptools/_vendor/typeguard/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/typing_extensions-4.12.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/typing_extensions-4.12.2.dist-info/LICENSE,sha256=Oy-B_iHRgcSZxZolbI4ZaEVdZonSaaqFNzv7avQdo78,13936
+setuptools/_vendor/typing_extensions-4.12.2.dist-info/METADATA,sha256=BeUQIa8cnYbrjWx-N8TOznM9UGW5Gm2DicVpDtRA8W0,3018
+setuptools/_vendor/typing_extensions-4.12.2.dist-info/RECORD,sha256=dxAALYGXHmMqpqL8M9xddKr118quIgQKZdPjFQOwXuk,571
+setuptools/_vendor/typing_extensions-4.12.2.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+setuptools/_vendor/typing_extensions.py,sha256=gwekpyG9DVG3lxWKX4ni8u7nk3We5slG98mA9F3DJQw,134451
+setuptools/_vendor/wheel-0.43.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/wheel-0.43.0.dist-info/LICENSE.txt,sha256=MMI2GGeRCPPo6h0qZYx8pBe9_IkcmO8aifpP8MmChlQ,1107
+setuptools/_vendor/wheel-0.43.0.dist-info/METADATA,sha256=WbrCKwClnT5WCKVrjPjvxDgxo2tyeS7kOJyc1GaceEE,2153
+setuptools/_vendor/wheel-0.43.0.dist-info/RECORD,sha256=eD5lR0JhGviM2fAL8BpDGiGdtTZVbmP_mBx71nMHCsk,4557
+setuptools/_vendor/wheel-0.43.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/wheel-0.43.0.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+setuptools/_vendor/wheel-0.43.0.dist-info/entry_points.txt,sha256=rTY1BbkPHhkGMm4Q3F0pIzJBzW2kMxoG1oriffvGdA0,104
+setuptools/_vendor/wheel/__init__.py,sha256=D6jhH00eMzbgrXGAeOwVfD5i-lCAMMycuG1L0useDlo,59
+setuptools/_vendor/wheel/__main__.py,sha256=NkMUnuTCGcOkgY0IBLgBCVC_BGGcWORx2K8jYGS12UE,455
+setuptools/_vendor/wheel/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/__main__.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/_setuptools_logging.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/bdist_wheel.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/macosx_libfile.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/metadata.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/util.cpython-312.pyc,,
+setuptools/_vendor/wheel/__pycache__/wheelfile.cpython-312.pyc,,
+setuptools/_vendor/wheel/_setuptools_logging.py,sha256=NoCnjJ4DFEZ45Eo-2BdXLsWJCwGkait1tp_17paleVw,746
+setuptools/_vendor/wheel/bdist_wheel.py,sha256=OKJyp9E831zJrxoRfmM9AgOjByG1CB-pzF5kXQFmaKk,20938
+setuptools/_vendor/wheel/cli/__init__.py,sha256=eBNhnPwWTtdKAJHy77lvz7gOQ5Eu3GavGugXxhSsn-U,4264
+setuptools/_vendor/wheel/cli/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/wheel/cli/__pycache__/convert.cpython-312.pyc,,
+setuptools/_vendor/wheel/cli/__pycache__/pack.cpython-312.pyc,,
+setuptools/_vendor/wheel/cli/__pycache__/tags.cpython-312.pyc,,
+setuptools/_vendor/wheel/cli/__pycache__/unpack.cpython-312.pyc,,
+setuptools/_vendor/wheel/cli/convert.py,sha256=qJcpYGKqdfw1P6BelgN1Hn_suNgM6bvyEWFlZeuSWx0,9439
+setuptools/_vendor/wheel/cli/pack.py,sha256=CAFcHdBVulvsHYJlndKVO7KMI9JqBTZz5ii0PKxxCOs,3103
+setuptools/_vendor/wheel/cli/tags.py,sha256=lHw-LaWrkS5Jy_qWcw-6pSjeNM6yAjDnqKI3E5JTTCU,4760
+setuptools/_vendor/wheel/cli/unpack.py,sha256=Y_J7ynxPSoFFTT7H0fMgbBlVErwyDGcObgme5MBuz58,1021
+setuptools/_vendor/wheel/macosx_libfile.py,sha256=HnW6OPdN993psStvwl49xtx2kw7hoVbe6nvwmf8WsKI,16103
+setuptools/_vendor/wheel/metadata.py,sha256=q-xCCqSAK7HzyZxK9A6_HAWmhqS1oB4BFw1-rHQxBiQ,5884
+setuptools/_vendor/wheel/util.py,sha256=e0jpnsbbM9QhaaMSyap-_ZgUxcxwpyLDk6RHcrduPLg,621
+setuptools/_vendor/wheel/vendored/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/wheel/vendored/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_elffile.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_manylinux.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_musllinux.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_parser.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_structures.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/_tokenizer.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/markers.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/requirements.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/specifiers.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/tags.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/utils.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/__pycache__/version.cpython-312.pyc,,
+setuptools/_vendor/wheel/vendored/packaging/_elffile.py,sha256=hbmK8OD6Z7fY6hwinHEUcD1by7czkGiNYu7ShnFEk2k,3266
+setuptools/_vendor/wheel/vendored/packaging/_manylinux.py,sha256=P7sdR5_7XBY09LVYYPhHmydMJIIwPXWsh4olk74Uuj4,9588
+setuptools/_vendor/wheel/vendored/packaging/_musllinux.py,sha256=z1s8To2hQ0vpn_d-O2i5qxGwEK8WmGlLt3d_26V7NeY,2674
+setuptools/_vendor/wheel/vendored/packaging/_parser.py,sha256=4tT4emSl2qTaU7VTQE1Xa9o1jMPCsBezsYBxyNMUN-s,10347
+setuptools/_vendor/wheel/vendored/packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+setuptools/_vendor/wheel/vendored/packaging/_tokenizer.py,sha256=alCtbwXhOFAmFGZ6BQ-wCTSFoRAJ2z-ysIf7__MTJ_k,5292
+setuptools/_vendor/wheel/vendored/packaging/markers.py,sha256=_TSPI1BhJYO7Bp9AzTmHQxIqHEVXaTjmDh9G-w8qzPA,8232
+setuptools/_vendor/wheel/vendored/packaging/requirements.py,sha256=dgoBeVprPu2YE6Q8nGfwOPTjATHbRa_ZGLyXhFEln6Q,2933
+setuptools/_vendor/wheel/vendored/packaging/specifiers.py,sha256=IWSt0SrLSP72heWhAC8UL0eGvas7XIQHjqiViVfmPKE,39778
+setuptools/_vendor/wheel/vendored/packaging/tags.py,sha256=fedHXiOHkBxNZTXotXv8uXPmMFU9ae-TKBujgYHigcA,18950
+setuptools/_vendor/wheel/vendored/packaging/utils.py,sha256=XgdmP3yx9-wQEFjO7OvMj9RjEf5JlR5HFFR69v7SQ9E,5268
+setuptools/_vendor/wheel/vendored/packaging/version.py,sha256=PFJaYZDxBgyxkfYhH3SQw4qfE9ICCWrTmitvq14y3bs,16234
+setuptools/_vendor/wheel/vendored/vendor.txt,sha256=Z2ENjB1i5prfez8CdM1Sdr3c6Zxv2rRRolMpLmBncAE,16
+setuptools/_vendor/wheel/wheelfile.py,sha256=DtJDWoZMvnBh4leNMDPGOprQU9d_dp6q-MmV0U--4xc,7694
+setuptools/_vendor/zipp-3.19.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+setuptools/_vendor/zipp-3.19.2.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+setuptools/_vendor/zipp-3.19.2.dist-info/METADATA,sha256=UIrk_kMIHGSwsKKChYizqMw0MMZpPRZ2ZiVpQAsN_bE,3575
+setuptools/_vendor/zipp-3.19.2.dist-info/RECORD,sha256=8xby4D_ZrefrvAsVRwaEjiu4_VaLkJNRCfDY484rm_4,1039
+setuptools/_vendor/zipp-3.19.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/zipp-3.19.2.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+setuptools/_vendor/zipp-3.19.2.dist-info/top_level.txt,sha256=iAbdoSHfaGqBfVb2XuR9JqSQHCoOsOtG6y9C_LSpqFw,5
+setuptools/_vendor/zipp/__init__.py,sha256=QuI1g00G4fRAcGt-HqbV0oWIkmSgedCGGYsHHYzNa8A,13412
+setuptools/_vendor/zipp/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/zipp/__pycache__/glob.cpython-312.pyc,,
+setuptools/_vendor/zipp/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/_vendor/zipp/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/_vendor/zipp/compat/__pycache__/py310.cpython-312.pyc,,
+setuptools/_vendor/zipp/compat/py310.py,sha256=eZpkW0zRtunkhEh8jjX3gCGe22emoKCBJw72Zt4RkhA,219
+setuptools/_vendor/zipp/glob.py,sha256=etWpnfEoRyfUvrUsi6sTiGmErvPwe6HzY6pT8jg_lUI,3082
+setuptools/archive_util.py,sha256=Tl_64hSTtc4y8x7xa98rFVUbG24oArpjzLAYGYP2_sI,7356
+setuptools/build_meta.py,sha256=rry-LowqvWFsxG6Qm5QWetHJGeETzRdiQ5-bs4bOkjo,20446
+setuptools/cli-32.exe,sha256=MqzBvFQxFsviz_EMuGd3LfLyVP8mNMhwrvC0bEtpb9s,11776
+setuptools/cli-64.exe,sha256=u7PeVwdinmpgoMI4zUd7KPB_AGaYL9qVP6b87DkHOko,14336
+setuptools/cli-arm64.exe,sha256=uafQjaiA36yLz1SOuksG-1m28JsX0zFIoPZhgyiSbGE,13824
+setuptools/cli.exe,sha256=MqzBvFQxFsviz_EMuGd3LfLyVP8mNMhwrvC0bEtpb9s,11776
+setuptools/command/__init__.py,sha256=wdSrlNR0P6nCz9_oFtCAiAkeFJMsZa1jPcpXT53f0SM,803
+setuptools/command/__pycache__/__init__.cpython-312.pyc,,
+setuptools/command/__pycache__/_requirestxt.cpython-312.pyc,,
+setuptools/command/__pycache__/alias.cpython-312.pyc,,
+setuptools/command/__pycache__/bdist_egg.cpython-312.pyc,,
+setuptools/command/__pycache__/bdist_rpm.cpython-312.pyc,,
+setuptools/command/__pycache__/bdist_wheel.cpython-312.pyc,,
+setuptools/command/__pycache__/build.cpython-312.pyc,,
+setuptools/command/__pycache__/build_clib.cpython-312.pyc,,
+setuptools/command/__pycache__/build_ext.cpython-312.pyc,,
+setuptools/command/__pycache__/build_py.cpython-312.pyc,,
+setuptools/command/__pycache__/develop.cpython-312.pyc,,
+setuptools/command/__pycache__/dist_info.cpython-312.pyc,,
+setuptools/command/__pycache__/easy_install.cpython-312.pyc,,
+setuptools/command/__pycache__/editable_wheel.cpython-312.pyc,,
+setuptools/command/__pycache__/egg_info.cpython-312.pyc,,
+setuptools/command/__pycache__/install.cpython-312.pyc,,
+setuptools/command/__pycache__/install_egg_info.cpython-312.pyc,,
+setuptools/command/__pycache__/install_lib.cpython-312.pyc,,
+setuptools/command/__pycache__/install_scripts.cpython-312.pyc,,
+setuptools/command/__pycache__/rotate.cpython-312.pyc,,
+setuptools/command/__pycache__/saveopts.cpython-312.pyc,,
+setuptools/command/__pycache__/sdist.cpython-312.pyc,,
+setuptools/command/__pycache__/setopt.cpython-312.pyc,,
+setuptools/command/__pycache__/test.cpython-312.pyc,,
+setuptools/command/_requirestxt.py,sha256=ItYMTJGh_i5TlQstX_nFopqEhkC4PJFadBL2Zd3V670,4228
+setuptools/command/alias.py,sha256=rDdrMt32DS6qf3K7tjZZyHD_dMKrm77AXcAtx-nBQ0I,2380
+setuptools/command/bdist_egg.py,sha256=3eDucQ4fdeYMsLO9PhBfY1JkcMLhZXgnAI_9FdFNsEE,16972
+setuptools/command/bdist_rpm.py,sha256=LyqI49w48SKk0FmuHsE9MLzX1SuXjL7YMNbZMFZqFII,1435
+setuptools/command/bdist_wheel.py,sha256=QCtkl0iSLhM5F65hACha-tAGxOdxu9dtJNBhKCa4wLI,22274
+setuptools/command/build.py,sha256=eI7STMERGGZEpzk1tvJN8p9IOjAAXMcGLzljv2mwI3M,6052
+setuptools/command/build_clib.py,sha256=AbgpPIF_3qL8fZr3JIebI-WHTMTBiMfrFkVQz8K40G4,4528
+setuptools/command/build_ext.py,sha256=bUH4M0NizaJJrv10wK-ZD3uY0TxCSZlYQDhiwwzHslM,18377
+setuptools/command/build_py.py,sha256=DCbjvB18kkL-xUK5rvlzm0C6twTeOxNhyvJDxxa7fII,15539
+setuptools/command/develop.py,sha256=zX22119sI1G1gfJ1gNCE4hkg2zbLKx0uUwvNmC5bIu8,6886
+setuptools/command/dist_info.py,sha256=HU752iLLmmYMHbsDBgz2ubRjkgJobugOp8H71LzzDys,3450
+setuptools/command/easy_install.py,sha256=0Z4kFlE78Ae2AfHXYTxZFUa2t3qlNqXCtQu4J1Nx8iA,87870
+setuptools/command/editable_wheel.py,sha256=Efoy1UnT-GO7ID1BdoHNdeHzNkvOXYrciATfEfYyLdE,35626
+setuptools/command/egg_info.py,sha256=WWUozR3DZCrWsTQhHXOygMiEUcrjLWphET0-Zsocsm4,25982
+setuptools/command/install.py,sha256=1PKfoE4F9fTDHmccflDnkb289CIcldY3aPtHyYuusHM,7045
+setuptools/command/install_egg_info.py,sha256=3I9IPCH7D59Sh-6aVYz-h6wwyxq-wkxrKwKg3nDdJqs,2075
+setuptools/command/install_lib.py,sha256=9n1_U83eHcERL_a_rv_LhHCkhXlLdqyZ4SdBow-9qcE,4319
+setuptools/command/install_scripts.py,sha256=tVOCj3e8OTIrkoL_bGbT5pOksdxZfQblH_bdI4DtVV4,2637
+setuptools/command/launcher manifest.xml,sha256=xlLbjWrB01tKC0-hlVkOKkiSPbzMml2eOPtJ_ucCnbE,628
+setuptools/command/rotate.py,sha256=XNd_BEEOWAJHW1FcLTMUWWl4QB6zAuk7b8VWQg3FHos,2187
+setuptools/command/saveopts.py,sha256=Np0PVb7SD7oTbu9Z9sosS7D-CkkIkU7x4glu5Es1tjA,692
+setuptools/command/sdist.py,sha256=JaQm2-ebXI2kvyrBjJKP8yNLPa5eMbMeis88CXBMYlk,7374
+setuptools/command/setopt.py,sha256=xZF2RCc4ABvE9eHHAzF50-fkQg3au8fcRUVVGd58k3U,5100
+setuptools/command/test.py,sha256=k7xcq7D7bEehgxarbw-dW3AtmGZORqz8HjKR6FGJ3jk,1343
+setuptools/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/compat/__pycache__/py310.cpython-312.pyc,,
+setuptools/compat/__pycache__/py311.cpython-312.pyc,,
+setuptools/compat/__pycache__/py312.cpython-312.pyc,,
+setuptools/compat/__pycache__/py39.cpython-312.pyc,,
+setuptools/compat/py310.py,sha256=8sqwWczIcrkzeAbhaim4pKVd4tXZdcqmebgdvzji0rc,141
+setuptools/compat/py311.py,sha256=e6tJAFwZEP82hmMBl10HYeSypelo_Ti2wTjKZVKLwOE,790
+setuptools/compat/py312.py,sha256=vYKVtdrdOTsO_R90dJkEXsFwfMJFuIFJflhIgHrjJ-Y,366
+setuptools/compat/py39.py,sha256=BJMtnkfcqyTfccqjYQxfoRtU2nTnWaEESBVkshTiXqY,493
+setuptools/config/NOTICE,sha256=Ld3wiBgpejuJ1D2V_2WdjahXQRCMkTbfo6TYVsBiO9g,493
+setuptools/config/__init__.py,sha256=aiPnL9BJn1O6MfmuNXyn8W2Lp8u9qizRVqwPiOdPIjY,1499
+setuptools/config/__pycache__/__init__.cpython-312.pyc,,
+setuptools/config/__pycache__/_apply_pyprojecttoml.cpython-312.pyc,,
+setuptools/config/__pycache__/expand.cpython-312.pyc,,
+setuptools/config/__pycache__/pyprojecttoml.cpython-312.pyc,,
+setuptools/config/__pycache__/setupcfg.cpython-312.pyc,,
+setuptools/config/_apply_pyprojecttoml.py,sha256=PTeYQZtIS5CRMw2L1lwnN7jptJkcCQZjsYqt4n_H03o,16988
+setuptools/config/_validate_pyproject/NOTICE,sha256=Ccm86pXKCG-Lxb7RdOQLyDWyl9QPtfhru7Vw_gpVgac,18737
+setuptools/config/_validate_pyproject/__init__.py,sha256=dnp6T7ePP1R5z4OuC7Fd2dkFlIrtIfizUfvpGJP6nz0,1042
+setuptools/config/_validate_pyproject/__pycache__/__init__.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/__pycache__/error_reporting.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/__pycache__/extra_validations.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_exceptions.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_validations.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/__pycache__/formats.cpython-312.pyc,,
+setuptools/config/_validate_pyproject/error_reporting.py,sha256=meldD7nBQdolQhvG-43r1Ue-gU1n7ORAJR86vh3Rrvk,11813
+setuptools/config/_validate_pyproject/extra_validations.py,sha256=kd0SWYrsp3IXF9KGAM8QpeaKpGyMsNgL-tjh9TPfhyY,1625
+setuptools/config/_validate_pyproject/fastjsonschema_exceptions.py,sha256=w749JgqKi8clBFcObdcbZVqsmF4oJ_QByhZ1SGbUFNw,1612
+setuptools/config/_validate_pyproject/fastjsonschema_validations.py,sha256=T8RvJSi53YBfDB4mRPuF6gsPxx8-u8KYX56y3qokp_4,335460
+setuptools/config/_validate_pyproject/formats.py,sha256=Mjz6Rjx1BMDw2XTMAfS-sM5x5Fv5aX2Zk_q5M_7rf_c,12814
+setuptools/config/distutils.schema.json,sha256=Tcp32kRnhwORGw_9p6GEi08lj2h15tQRzOYBbzGmcBU,972
+setuptools/config/expand.py,sha256=JNAktRCsyyRB-rQodbPnCucmLWqcYvzCDC8Ebn2Z7xM,16041
+setuptools/config/pyprojecttoml.py,sha256=YMu5PdbJJI5azp6kR_boM1mflf5nqOA-InF4s6LnLgw,18320
+setuptools/config/setupcfg.py,sha256=VZDkwE7DYv45SbadJD8CwKrDtiXvjgllL8PYSvoRCyg,26575
+setuptools/config/setuptools.schema.json,sha256=dZBRuSEnZkatoVlt1kVwG8ocTeRdO7BD0xvOWKH54PY,16071
+setuptools/depends.py,sha256=jKYfjmt_2ZQYVghb8L9bU7LJ6erHJ5ze-K_fKV1BMXk,5965
+setuptools/discovery.py,sha256=-42c3XhwzkfodDKKP50C2YBzr11fncAgmUzBdBRb0-Q,21258
+setuptools/dist.py,sha256=9hze5pIoeYSPFOqvhz0hituaxEPsV4wN9sY3hcf3hfg,38815
+setuptools/errors.py,sha256=gY2x2PIaIgy01yRANRC-zcCwxDCqCScgJoCOZFe0yio,3024
+setuptools/extension.py,sha256=KCnv9p3tgm0ZVqtgE451fyILsm4hCyvOiUtOu787D-4,6683
+setuptools/glob.py,sha256=AC_B33DY8g-CHELxDsJrtwFrpiucSAZsakPFdSOQzhc,6062
+setuptools/gui-32.exe,sha256=hdrh6V13hF8stZvKw9Sv50u-TJGpvMW_SnHNQxBNvnw,11776
+setuptools/gui-64.exe,sha256=NHG2FA6txkEid9u-_j_vjDRaDxpZd2CGuAo2GMOoPjs,14336
+setuptools/gui-arm64.exe,sha256=5pT0dDQFyLWSb_RX22_n8aEt7HwWqcOGR4TT9OB64Jc,13824
+setuptools/gui.exe,sha256=hdrh6V13hF8stZvKw9Sv50u-TJGpvMW_SnHNQxBNvnw,11776
+setuptools/installer.py,sha256=_4Wegx4r3L05sMo3-IlqFp-OuxnWyBqjyMZ7LWQXmh8,5110
+setuptools/launch.py,sha256=IBb5lEv69CyuZ9ewIrmKlXh154kdLmP29LKfTMkximE,820
+setuptools/logging.py,sha256=W16iHJ1HcCXYQ0RxyrEfJ83FT4175tCtoYg-E6uSpVI,1261
+setuptools/modified.py,sha256=ZwbfBfCFP88ltvbv_dJDz-t1LsQjnM-JUpgZnnQZjjM,568
+setuptools/monkey.py,sha256=FwMWl2n1v2bHbeqBy-o9g8yUNaAkYFbszCbXe9d5Za8,3717
+setuptools/msvc.py,sha256=t3d_HZ6gMqyqnhz5dPRBuMywTTPk3LylIWkzt0RCVnM,41255
+setuptools/namespaces.py,sha256=2GGqYY1BNDEhMtBc1rHTv7klgmNVRdksJeW-L1f--ys,3171
+setuptools/package_index.py,sha256=yHXj4Z_QaUeh8W6rwq08I5LxnyHrSkNf01uXHwdE8K0,39095
+setuptools/sandbox.py,sha256=fMqtcOuipHO6RKPh1YB5o7d985dLKo76Whp3vrIei2E,14906
+setuptools/script (dev).tmpl,sha256=RUzQzCQUaXtwdLtYHWYbIQmOaES5Brqq1FvUA_tu-5I,218
+setuptools/script.tmpl,sha256=WGTt5piezO27c-Dbx6l5Q4T3Ff20A5z7872hv3aAhYY,138
+setuptools/tests/__init__.py,sha256=AnBfls2iJbTDQzmMKeLRt-9lxhaOHUVOZEgXv89Uwvs,335
+setuptools/tests/__pycache__/__init__.cpython-312.pyc,,
+setuptools/tests/__pycache__/contexts.cpython-312.pyc,,
+setuptools/tests/__pycache__/environment.cpython-312.pyc,,
+setuptools/tests/__pycache__/fixtures.cpython-312.pyc,,
+setuptools/tests/__pycache__/mod_with_constant.cpython-312.pyc,,
+setuptools/tests/__pycache__/namespaces.cpython-312.pyc,,
+setuptools/tests/__pycache__/script-with-bom.cpython-312.pyc,,
+setuptools/tests/__pycache__/server.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_archive_util.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_bdist_deprecations.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_bdist_egg.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_bdist_wheel.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_build.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_build_clib.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_build_ext.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_build_meta.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_build_py.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_config_discovery.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_core_metadata.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_depends.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_develop.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_dist.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_dist_info.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_distutils_adoption.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_easy_install.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_editable_install.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_egg_info.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_extern.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_find_packages.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_find_py_modules.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_glob.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_install_scripts.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_logging.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_manifest.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_namespaces.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_packageindex.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_sandbox.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_sdist.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_setopt.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_setuptools.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_shutil_wrapper.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_unicode_utils.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_virtualenv.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_warnings.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_wheel.cpython-312.pyc,,
+setuptools/tests/__pycache__/test_windows_wrappers.cpython-312.pyc,,
+setuptools/tests/__pycache__/text.cpython-312.pyc,,
+setuptools/tests/__pycache__/textwrap.cpython-312.pyc,,
+setuptools/tests/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/tests/compat/__pycache__/__init__.cpython-312.pyc,,
+setuptools/tests/compat/__pycache__/py39.cpython-312.pyc,,
+setuptools/tests/compat/py39.py,sha256=eUy7_F-6KRTOIKl-veshUu6I0EdTSdBZMh0EV0lZ1-g,135
+setuptools/tests/config/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/tests/config/__pycache__/__init__.cpython-312.pyc,,
+setuptools/tests/config/__pycache__/test_apply_pyprojecttoml.cpython-312.pyc,,
+setuptools/tests/config/__pycache__/test_expand.cpython-312.pyc,,
+setuptools/tests/config/__pycache__/test_pyprojecttoml.cpython-312.pyc,,
+setuptools/tests/config/__pycache__/test_pyprojecttoml_dynamic_deps.cpython-312.pyc,,
+setuptools/tests/config/__pycache__/test_setupcfg.cpython-312.pyc,,
+setuptools/tests/config/downloads/__init__.py,sha256=9ixnDEdyL_arKbUzfuiJftAj9bGxKz8M9alOFZMjx9Y,1827
+setuptools/tests/config/downloads/__pycache__/__init__.cpython-312.pyc,,
+setuptools/tests/config/downloads/__pycache__/preload.cpython-312.pyc,,
+setuptools/tests/config/downloads/preload.py,sha256=sIGGZpY3cmMpMwiJYYYYHG2ifZJkvJgEotRFtiulV1I,450
+setuptools/tests/config/setupcfg_examples.txt,sha256=cAbVvCbkFZuTUL6xRRzRgqyB0rLvJTfvw3D30glo2OE,1912
+setuptools/tests/config/test_apply_pyprojecttoml.py,sha256=Vjfnmr2_7GZmCcBj8q-rU99gg2eGrJPbqJjq_dSt1uU,20286
+setuptools/tests/config/test_expand.py,sha256=S0oT6JvgA_oujR4YS4RUuf5gmOt1CTQV66RQDzV8xd4,8933
+setuptools/tests/config/test_pyprojecttoml.py,sha256=0LefSljUhA6MqtJ5AVzLhomqZcYiFKdu_1ckDeMT1LY,12406
+setuptools/tests/config/test_pyprojecttoml_dynamic_deps.py,sha256=9W73-yLhZJmvCiO4rTiQoBpZT5wNA90Xbd5n2HCshd4,3271
+setuptools/tests/config/test_setupcfg.py,sha256=r5ZQJ6mt6mybbxdfv8u6Pl4K5PtFNiNUWxcohiwz8Aw,33189
+setuptools/tests/contexts.py,sha256=TAdZKxmmodx1ExMVo01o4QpRjpIpo4X3IWKq_BnjxpU,3480
+setuptools/tests/environment.py,sha256=95_UtTaRiuvwYC9eXKEHbn02kDtZysvZq3UZJmPUj1I,3102
+setuptools/tests/fixtures.py,sha256=-V7iD6BeE2E0Rw6dVvTOCm36JG8ZTTnrXhN0GISlgrg,5197
+setuptools/tests/indexes/test_links_priority/external.html,sha256=eL9euOuE93JKZdqlXxBOlHbKwIuNuIdq7GBRpsaPMcU,92
+setuptools/tests/indexes/test_links_priority/simple/foobar/index.html,sha256=DD-TKr7UU4zAjHHz4VexYDNSAzR27levSh1c-k3ZdLE,174
+setuptools/tests/integration/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+setuptools/tests/integration/__pycache__/__init__.cpython-312.pyc,,
+setuptools/tests/integration/__pycache__/helpers.cpython-312.pyc,,
+setuptools/tests/integration/__pycache__/test_pip_install_sdist.cpython-312.pyc,,
+setuptools/tests/integration/helpers.py,sha256=3PHcS9SCA-fwVJmUP2ad5NQOttJAETI5Nnoc_xroO5k,2522
+setuptools/tests/integration/test_pip_install_sdist.py,sha256=uZW03e141p--A16Ot4YAugBS3h6yAhElPaQyTSJAnzI,8204
+setuptools/tests/mod_with_constant.py,sha256=X_Kj80M55w1tmQ4f7uZY91ZTALo4hKVT6EHxgYocUMQ,22
+setuptools/tests/namespaces.py,sha256=HPcI3nR5MCFWXpaADIJ1fwKxymcQgBkuw87Ic5PUSAQ,2774
+setuptools/tests/script-with-bom.py,sha256=hRRgIizEULGiG_ZTNoMY46HhKhxpWfy5FGcD6Qbh5fc,18
+setuptools/tests/server.py,sha256=0FDZf0cSInCP5n1haK7AxHo3jD261RK7D3-RjP-F53k,2397
+setuptools/tests/test_archive_util.py,sha256=buuKdY8XkW26Pe3IKAoBRGHG0MDumnuNoPg2WsAQzIg,845
+setuptools/tests/test_bdist_deprecations.py,sha256=75Xq3gYn79LIIyusEltbHan0bEgAt2e_CaL7KLS8-KQ,775
+setuptools/tests/test_bdist_egg.py,sha256=6PaYN1F3JDbIh1uK0urv7yJFcx98z5dn9SOJ8Mv91l8,1957
+setuptools/tests/test_bdist_wheel.py,sha256=C3Pp7p2fUCGvuH0Idq8lCC2ABhsu4G2iN5JaocyNHQw,19906
+setuptools/tests/test_build.py,sha256=wJgMz2hwHADcLFg-nXrwRVhus7hjmAeEGgrpIQwCGnA,798
+setuptools/tests/test_build_clib.py,sha256=bX51XRAf4uO7IuHFpjePnoK8mE74N2gsoeEqF-ofgws,3123
+setuptools/tests/test_build_ext.py,sha256=e4ZSxsYPB5zq1KSqGEuATZ0t0PJQzMhjjkKJ-hIjcgc,10099
+setuptools/tests/test_build_meta.py,sha256=iRoGOf5vehltTldbghc97pHj-8nDya_Nl9-Bsdyl8Q0,33574
+setuptools/tests/test_build_py.py,sha256=-1DnEEJgavQFPzoLgHc-QriFv0llaOp1YEr7LDXCSBo,14187
+setuptools/tests/test_config_discovery.py,sha256=FqV-lOtkqaI-ayzU2zocSdD5TaRAgCZnixNDilKA6FQ,22580
+setuptools/tests/test_core_metadata.py,sha256=FpnPjsuRnA-8jRDvBe2oI_3vivwjNTPuF2CswCONXA4,19102
+setuptools/tests/test_depends.py,sha256=yQBXoQbNQlJit6mbRVoz6Bb553f3sNrq02lZimNz5XY,424
+setuptools/tests/test_develop.py,sha256=CLzXZ8-b5-VFTuau4P4yXEdLx1UdyTFcOfrV0qyUIdE,5142
+setuptools/tests/test_dist.py,sha256=HbTZp7D5NzLAh3btrU8LeuZv4SfAADa9EVubeZyu3iE,8897
+setuptools/tests/test_dist_info.py,sha256=Agy3L0M3hF5dxXwSpLJwIbgJI3ODy_bsH23cEUR2OlM,7094
+setuptools/tests/test_distutils_adoption.py,sha256=_eynrOfyEqXFEmjUJhzpe8GXPyTUPvNSObs4qAAmBy8,5987
+setuptools/tests/test_easy_install.py,sha256=D8hsk-Z0hYcrKoyKM8AnKR0IdVfgXDlR2yARSoXDzTc,53308
+setuptools/tests/test_editable_install.py,sha256=7eTEtpT0k7QeVyZg64eh3kZn-SjckuB9LcokOuV37DI,43383
+setuptools/tests/test_egg_info.py,sha256=0Bi29Im-Pz_SvKtKVkq7sy2tYEdrvfvCOyexfTx3jyo,43963
+setuptools/tests/test_extern.py,sha256=rpKU6oCcksumLwf5TeKlDluFQ0TUfbPwTLQbpxcFrCU,296
+setuptools/tests/test_find_packages.py,sha256=CTLAcTzWGWBLCcd2aAsUVkvO3ibrlqexFBdDKOWPoq8,7819
+setuptools/tests/test_find_py_modules.py,sha256=zQjuhIG5TQN2SJPix9ARo4DL_w84Ln8QsHDUjjbrtAQ,2404
+setuptools/tests/test_glob.py,sha256=P3JvpH-kXQ4BZ3zvRF-zKxOgwyWzwIaQIz0WHdxS0kk,887
+setuptools/tests/test_install_scripts.py,sha256=scIrJ6a_ssKqg4vIBNaUjmAKHEYLUUZ9WKnPeKnE6gc,3433
+setuptools/tests/test_logging.py,sha256=zlE5DlldukC7Jc54FNvDV_7ux3ErAkrfrN5CSsnNOUQ,2099
+setuptools/tests/test_manifest.py,sha256=eMg65pIA52DizB6mpktSU-b8CjwaNCS5MSgL_V1LrFI,18562
+setuptools/tests/test_namespaces.py,sha256=Y6utoe5PHHqL_DlgawqB9F8XpsUDPvvw1sQMenK04e0,4515
+setuptools/tests/test_packageindex.py,sha256=qEjLHpSu2gAkegwEstzHQT-Om1uQIYjA8zeNzEX79uo,8775
+setuptools/tests/test_sandbox.py,sha256=shUWE7fLTWe7Jzdfi6NheBWauG3oUukbUV46cArD0u0,4330
+setuptools/tests/test_sdist.py,sha256=0SIflXKf3g0hNPY8ZZbqx_zV4SK1oo6X6A853sgKYh4,32428
+setuptools/tests/test_setopt.py,sha256=3VxxM4ATfP-P4AGnDjoWCnHr5-i9CSEQTFYU1-FTnvI,1365
+setuptools/tests/test_setuptools.py,sha256=_eIhqKf45-OtHqxRf20KndOZJlJdS0PuFLXBO3M-LN8,9008
+setuptools/tests/test_shutil_wrapper.py,sha256=g15E11PtZxG-InB2BWNFyH-svObXx2XcMhgMLJPuFnc,641
+setuptools/tests/test_unicode_utils.py,sha256=xWfEEl8jkQCt9othUTXJfFmdyATAFggJs2tTxjbumbw,316
+setuptools/tests/test_virtualenv.py,sha256=g-njC_9JTAs1YVx_1dGJ_Q6RlInO4qKVu9-XAgNb6TY,3730
+setuptools/tests/test_warnings.py,sha256=zwR2zcnCeCeDqILZlJOPAcuyPHoDvGu1OtOVYiLMk74,3347
+setuptools/tests/test_wheel.py,sha256=J-83W1KdXTgAjFZE3H-ytohhvDE1iqdbE5YF5jLQlGQ,19370
+setuptools/tests/test_windows_wrappers.py,sha256=aF6UTowN3yzCgdBh9nDQVvYIfSYogrTK776TEyXEBqg,7881
+setuptools/tests/text.py,sha256=a12197pMVTvB6FAWQ0ujT8fIQiLIWJlFAl1UCaDUDfg,123
+setuptools/tests/textwrap.py,sha256=FNNNq_MiaEJx88PnsbJQIRxmj1qmgcAOCXXRsODPJN4,98
+setuptools/unicode_utils.py,sha256=d9M9xmxKQIJYwNZFa4IPpIFEKTbxIAae1jNoVxu12Aw,3190
+setuptools/version.py,sha256=WJCeUuyq74Aok2TeK9-OexZOu8XrlQy7-y0BEuWNovQ,161
+setuptools/warnings.py,sha256=oY0Se5eOqje_FEyjTgonUc0XGwgsrI5cgm1kkwulz_w,3796
+setuptools/wheel.py,sha256=xkAtvgm7uPTyYV2zqVmQ0wA8kLwOyRT2Jes1zAy07Ks,8624
+setuptools/windows_support.py,sha256=wW4IYLM1Bv7Z1MaauP2xmPjyy-wkmQnXdyvXscAf9fw,726
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/WHEEL
new file mode 100644
index 00000000..505164bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.8.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/entry_points.txt
new file mode 100644
index 00000000..0db0a6c8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/entry_points.txt
@@ -0,0 +1,51 @@
+[distutils.commands]
+alias = setuptools.command.alias:alias
+bdist_egg = setuptools.command.bdist_egg:bdist_egg
+bdist_rpm = setuptools.command.bdist_rpm:bdist_rpm
+bdist_wheel = setuptools.command.bdist_wheel:bdist_wheel
+build = setuptools.command.build:build
+build_clib = setuptools.command.build_clib:build_clib
+build_ext = setuptools.command.build_ext:build_ext
+build_py = setuptools.command.build_py:build_py
+develop = setuptools.command.develop:develop
+dist_info = setuptools.command.dist_info:dist_info
+easy_install = setuptools.command.easy_install:easy_install
+editable_wheel = setuptools.command.editable_wheel:editable_wheel
+egg_info = setuptools.command.egg_info:egg_info
+install = setuptools.command.install:install
+install_egg_info = setuptools.command.install_egg_info:install_egg_info
+install_lib = setuptools.command.install_lib:install_lib
+install_scripts = setuptools.command.install_scripts:install_scripts
+rotate = setuptools.command.rotate:rotate
+saveopts = setuptools.command.saveopts:saveopts
+sdist = setuptools.command.sdist:sdist
+setopt = setuptools.command.setopt:setopt
+
+[distutils.setup_keywords]
+dependency_links = setuptools.dist:assert_string_list
+eager_resources = setuptools.dist:assert_string_list
+entry_points = setuptools.dist:check_entry_points
+exclude_package_data = setuptools.dist:check_package_data
+extras_require = setuptools.dist:check_extras
+include_package_data = setuptools.dist:assert_bool
+install_requires = setuptools.dist:check_requirements
+namespace_packages = setuptools.dist:check_nsp
+package_data = setuptools.dist:check_package_data
+packages = setuptools.dist:check_packages
+python_requires = setuptools.dist:check_specifier
+setup_requires = setuptools.dist:check_requirements
+use_2to3 = setuptools.dist:invalid_unless_false
+zip_safe = setuptools.dist:assert_bool
+
+[egg_info.writers]
+PKG-INFO = setuptools.command.egg_info:write_pkg_info
+dependency_links.txt = setuptools.command.egg_info:overwrite_arg
+eager_resources.txt = setuptools.command.egg_info:overwrite_arg
+entry_points.txt = setuptools.command.egg_info:write_entries
+namespace_packages.txt = setuptools.command.egg_info:overwrite_arg
+requires.txt = setuptools.command.egg_info:write_requirements
+top_level.txt = setuptools.command.egg_info:write_toplevel_names
+
+[setuptools.finalize_distribution_options]
+keywords = setuptools.dist:Distribution._finalize_setup_keywords
+parent_finalize = setuptools.dist:_Distribution.finalize_options
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/top_level.txt
new file mode 100644
index 00000000..b5ac1070
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools-75.8.0.dist-info/top_level.txt
@@ -0,0 +1,3 @@
+_distutils_hack
+pkg_resources
+setuptools
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/__init__.py
new file mode 100644
index 00000000..64464dfa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/__init__.py
@@ -0,0 +1,286 @@
+"""Extensions to the 'distutils' for large or complex distributions"""
+# mypy: disable_error_code=override
+# Command.reinitialize_command has an extra **kw param that distutils doesn't have
+# Can't disable on the exact line because distutils doesn't exists on Python 3.12
+# and mypy isn't aware of distutils_hack, causing distutils.core.Command to be Any,
+# and a [unused-ignore] to be raised on 3.12+
+
+from __future__ import annotations
+
+import functools
+import os
+import re
+import sys
+from abc import abstractmethod
+from collections.abc import Mapping
+from typing import TYPE_CHECKING, TypeVar, overload
+
+sys.path.extend(((vendor_path := os.path.join(os.path.dirname(os.path.dirname(__file__)), 'setuptools', '_vendor')) not in sys.path) * [vendor_path])  # fmt: skip
+# workaround for #4476
+sys.modules.pop('backports', None)
+
+import _distutils_hack.override  # noqa: F401
+
+from . import logging, monkey
+from .depends import Require
+from .discovery import PackageFinder, PEP420PackageFinder
+from .dist import Distribution
+from .extension import Extension
+from .version import __version__ as __version__
+from .warnings import SetuptoolsDeprecationWarning
+
+import distutils.core
+from distutils.errors import DistutilsOptionError
+
+__all__ = [
+    'setup',
+    'Distribution',
+    'Command',
+    'Extension',
+    'Require',
+    'SetuptoolsDeprecationWarning',
+    'find_packages',
+    'find_namespace_packages',
+]
+
+_CommandT = TypeVar("_CommandT", bound="_Command")
+
+bootstrap_install_from = None
+
+find_packages = PackageFinder.find
+find_namespace_packages = PEP420PackageFinder.find
+
+
+def _install_setup_requires(attrs):
+    # Note: do not use `setuptools.Distribution` directly, as
+    # our PEP 517 backend patch `distutils.core.Distribution`.
+    class MinimalDistribution(distutils.core.Distribution):
+        """
+        A minimal version of a distribution for supporting the
+        fetch_build_eggs interface.
+        """
+
+        def __init__(self, attrs: Mapping[str, object]) -> None:
+            _incl = 'dependency_links', 'setup_requires'
+            filtered = {k: attrs[k] for k in set(_incl) & set(attrs)}
+            super().__init__(filtered)
+            # Prevent accidentally triggering discovery with incomplete set of attrs
+            self.set_defaults._disable()
+
+        def _get_project_config_files(self, filenames=None):
+            """Ignore ``pyproject.toml``, they are not related to setup_requires"""
+            try:
+                cfg, _toml = super()._split_standard_project_metadata(filenames)
+            except Exception:
+                return filenames, ()
+            return cfg, ()
+
+        def finalize_options(self):
+            """
+            Disable finalize_options to avoid building the working set.
+            Ref #2158.
+            """
+
+    dist = MinimalDistribution(attrs)
+
+    # Honor setup.cfg's options.
+    dist.parse_config_files(ignore_option_errors=True)
+    if dist.setup_requires:
+        _fetch_build_eggs(dist)
+
+
+def _fetch_build_eggs(dist: Distribution):
+    try:
+        dist.fetch_build_eggs(dist.setup_requires)
+    except Exception as ex:
+        msg = """
+        It is possible a package already installed in your system
+        contains an version that is invalid according to PEP 440.
+        You can try `pip install --use-pep517` as a workaround for this problem,
+        or rely on a new virtual environment.
+
+        If the problem refers to a package that is not installed yet,
+        please contact that package's maintainers or distributors.
+        """
+        if "InvalidVersion" in ex.__class__.__name__:
+            if hasattr(ex, "add_note"):
+                ex.add_note(msg)  # PEP 678
+            else:
+                dist.announce(f"\n{msg}\n")
+        raise
+
+
+def setup(**attrs):
+    logging.configure()
+    # Make sure we have any requirements needed to interpret 'attrs'.
+    _install_setup_requires(attrs)
+    return distutils.core.setup(**attrs)
+
+
+setup.__doc__ = distutils.core.setup.__doc__
+
+if TYPE_CHECKING:
+    # Work around a mypy issue where type[T] can't be used as a base: https://github.com/python/mypy/issues/10962
+    from distutils.core import Command as _Command
+else:
+    _Command = monkey.get_unpatched(distutils.core.Command)
+
+
+class Command(_Command):
+    """
+    Setuptools internal actions are organized using a *command design pattern*.
+    This means that each action (or group of closely related actions) executed during
+    the build should be implemented as a ``Command`` subclass.
+
+    These commands are abstractions and do not necessarily correspond to a command that
+    can (or should) be executed via a terminal, in a CLI fashion (although historically
+    they would).
+
+    When creating a new command from scratch, custom defined classes **SHOULD** inherit
+    from ``setuptools.Command`` and implement a few mandatory methods.
+    Between these mandatory methods, are listed:
+    :meth:`initialize_options`, :meth:`finalize_options` and :meth:`run`.
+
+    A useful analogy for command classes is to think of them as subroutines with local
+    variables called "options".  The options are "declared" in :meth:`initialize_options`
+    and "defined" (given their final values, aka "finalized") in :meth:`finalize_options`,
+    both of which must be defined by every command class. The "body" of the subroutine,
+    (where it does all the work) is the :meth:`run` method.
+    Between :meth:`initialize_options` and :meth:`finalize_options`, ``setuptools`` may set
+    the values for options/attributes based on user's input (or circumstance),
+    which means that the implementation should be careful to not overwrite values in
+    :meth:`finalize_options` unless necessary.
+
+    Please note that other commands (or other parts of setuptools) may also overwrite
+    the values of the command's options/attributes multiple times during the build
+    process.
+    Therefore it is important to consistently implement :meth:`initialize_options` and
+    :meth:`finalize_options`. For example, all derived attributes (or attributes that
+    depend on the value of other attributes) **SHOULD** be recomputed in
+    :meth:`finalize_options`.
+
+    When overwriting existing commands, custom defined classes **MUST** abide by the
+    same APIs implemented by the original class. They also **SHOULD** inherit from the
+    original class.
+    """
+
+    command_consumes_arguments = False
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    def __init__(self, dist: Distribution, **kw) -> None:
+        """
+        Construct the command for dist, updating
+        vars(self) with any keyword parameters.
+        """
+        super().__init__(dist)
+        vars(self).update(kw)
+
+    def _ensure_stringlike(self, option, what, default=None):
+        val = getattr(self, option)
+        if val is None:
+            setattr(self, option, default)
+            return default
+        elif not isinstance(val, str):
+            raise DistutilsOptionError(f"'{option}' must be a {what} (got `{val}`)")
+        return val
+
+    def ensure_string_list(self, option: str) -> None:
+        r"""Ensure that 'option' is a list of strings.  If 'option' is
+        currently a string, we split it either on /,\s*/ or /\s+/, so
+        "foo bar baz", "foo,bar,baz", and "foo,   bar baz" all become
+        ["foo", "bar", "baz"].
+
+        ..
+           TODO: This method seems to be similar to the one in ``distutils.cmd``
+           Probably it is just here for backward compatibility with old Python versions?
+
+        :meta private:
+        """
+        val = getattr(self, option)
+        if val is None:
+            return
+        elif isinstance(val, str):
+            setattr(self, option, re.split(r',\s*|\s+', val))
+        else:
+            if isinstance(val, list):
+                ok = all(isinstance(v, str) for v in val)
+            else:
+                ok = False
+            if not ok:
+                raise DistutilsOptionError(
+                    f"'{option}' must be a list of strings (got {val!r})"
+                )
+
+    @overload
+    def reinitialize_command(
+        self, command: str, reinit_subcommands: bool = False, **kw
+    ) -> _Command: ...
+    @overload
+    def reinitialize_command(
+        self, command: _CommandT, reinit_subcommands: bool = False, **kw
+    ) -> _CommandT: ...
+    def reinitialize_command(
+        self, command: str | _Command, reinit_subcommands: bool = False, **kw
+    ) -> _Command:
+        cmd = _Command.reinitialize_command(self, command, reinit_subcommands)
+        vars(cmd).update(kw)
+        return cmd  # pyright: ignore[reportReturnType] # pypa/distutils#307
+
+    @abstractmethod
+    def initialize_options(self) -> None:
+        """
+        Set or (reset) all options/attributes/caches used by the command
+        to their default values. Note that these values may be overwritten during
+        the build.
+        """
+        raise NotImplementedError
+
+    @abstractmethod
+    def finalize_options(self) -> None:
+        """
+        Set final values for all options/attributes used by the command.
+        Most of the time, each option/attribute/cache should only be set if it does not
+        have any value yet (e.g. ``if self.attr is None: self.attr = val``).
+        """
+        raise NotImplementedError
+
+    @abstractmethod
+    def run(self) -> None:
+        """
+        Execute the actions intended by the command.
+        (Side effects **SHOULD** only take place when :meth:`run` is executed,
+        for example, creating new files or writing to the terminal output).
+        """
+        raise NotImplementedError
+
+
+def _find_all_simple(path):
+    """
+    Find all files under 'path'
+    """
+    results = (
+        os.path.join(base, file)
+        for base, dirs, files in os.walk(path, followlinks=True)
+        for file in files
+    )
+    return filter(os.path.isfile, results)
+
+
+def findall(dir=os.curdir):
+    """
+    Find all files under 'dir' and return the list of full filenames.
+    Unless dir is '.', return full filenames with dir prepended.
+    """
+    files = _find_all_simple(dir)
+    if dir == os.curdir:
+        make_rel = functools.partial(os.path.relpath, start=dir)
+        files = map(make_rel, files)
+    return list(files)
+
+
+class sic(str):
+    """Treat this string as-is (https://en.wikipedia.org/wiki/Sic)"""
+
+
+# Apply monkey patches
+monkey.patch_all()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b4691960
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_core_metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_core_metadata.cpython-312.pyc
new file mode 100644
index 00000000..84e8333f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_core_metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_entry_points.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_entry_points.cpython-312.pyc
new file mode 100644
index 00000000..7ff9a411
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_entry_points.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_imp.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_imp.cpython-312.pyc
new file mode 100644
index 00000000..4529fda5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_imp.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_importlib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_importlib.cpython-312.pyc
new file mode 100644
index 00000000..ae0daea2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_importlib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_itertools.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_itertools.cpython-312.pyc
new file mode 100644
index 00000000..f1e459cc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_itertools.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_normalization.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_normalization.cpython-312.pyc
new file mode 100644
index 00000000..73a9ff43
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_normalization.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_path.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_path.cpython-312.pyc
new file mode 100644
index 00000000..6dab3956
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_path.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_reqs.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_reqs.cpython-312.pyc
new file mode 100644
index 00000000..ca068041
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_reqs.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_shutil.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_shutil.cpython-312.pyc
new file mode 100644
index 00000000..6ebad107
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_shutil.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_static.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_static.cpython-312.pyc
new file mode 100644
index 00000000..151d6eb2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/_static.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/archive_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/archive_util.cpython-312.pyc
new file mode 100644
index 00000000..308af5ef
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/archive_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/build_meta.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/build_meta.cpython-312.pyc
new file mode 100644
index 00000000..c9bd3b2b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/build_meta.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/depends.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/depends.cpython-312.pyc
new file mode 100644
index 00000000..649e6943
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/depends.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/discovery.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/discovery.cpython-312.pyc
new file mode 100644
index 00000000..c364d071
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/discovery.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/dist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/dist.cpython-312.pyc
new file mode 100644
index 00000000..b305f05c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/dist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/errors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/errors.cpython-312.pyc
new file mode 100644
index 00000000..8da1537b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/errors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/extension.cpython-312.pyc
new file mode 100644
index 00000000..c2ef812d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/glob.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/glob.cpython-312.pyc
new file mode 100644
index 00000000..4fd18ad5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/glob.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/installer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/installer.cpython-312.pyc
new file mode 100644
index 00000000..9f8db909
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/installer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/launch.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/launch.cpython-312.pyc
new file mode 100644
index 00000000..b4b23250
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/launch.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/logging.cpython-312.pyc
new file mode 100644
index 00000000..9b2d9435
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/modified.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/modified.cpython-312.pyc
new file mode 100644
index 00000000..53a48092
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/modified.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/monkey.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/monkey.cpython-312.pyc
new file mode 100644
index 00000000..5f9dfb76
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/monkey.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/msvc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/msvc.cpython-312.pyc
new file mode 100644
index 00000000..32e56605
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/msvc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/namespaces.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/namespaces.cpython-312.pyc
new file mode 100644
index 00000000..f168dec7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/namespaces.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/package_index.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/package_index.cpython-312.pyc
new file mode 100644
index 00000000..1f7bbec6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/package_index.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/sandbox.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/sandbox.cpython-312.pyc
new file mode 100644
index 00000000..2cfd98f9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/sandbox.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/unicode_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/unicode_utils.cpython-312.pyc
new file mode 100644
index 00000000..ea70dd0c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/unicode_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..da83153e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/warnings.cpython-312.pyc
new file mode 100644
index 00000000..a286b3d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/wheel.cpython-312.pyc
new file mode 100644
index 00000000..6d185af0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/windows_support.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/windows_support.cpython-312.pyc
new file mode 100644
index 00000000..e5ddc4e2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/__pycache__/windows_support.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_core_metadata.py b/backend/venv/lib/python3.12/site-packages/setuptools/_core_metadata.py
new file mode 100644
index 00000000..850cc409
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_core_metadata.py
@@ -0,0 +1,321 @@
+"""
+Handling of Core Metadata for Python packages (including reading and writing).
+
+See: https://packaging.python.org/en/latest/specifications/core-metadata/
+"""
+
+from __future__ import annotations
+
+import os
+import stat
+import textwrap
+from email import message_from_file
+from email.message import Message
+from tempfile import NamedTemporaryFile
+
+from packaging.markers import Marker
+from packaging.requirements import Requirement
+from packaging.utils import canonicalize_name, canonicalize_version
+from packaging.version import Version
+
+from . import _normalization, _reqs
+from ._static import is_static
+from .warnings import SetuptoolsDeprecationWarning
+
+from distutils.util import rfc822_escape
+
+
+def get_metadata_version(self):
+    mv = getattr(self, 'metadata_version', None)
+    if mv is None:
+        mv = Version('2.2')
+        self.metadata_version = mv
+    return mv
+
+
+def rfc822_unescape(content: str) -> str:
+    """Reverse RFC-822 escaping by removing leading whitespaces from content."""
+    lines = content.splitlines()
+    if len(lines) == 1:
+        return lines[0].lstrip()
+    return '\n'.join((lines[0].lstrip(), textwrap.dedent('\n'.join(lines[1:]))))
+
+
+def _read_field_from_msg(msg: Message, field: str) -> str | None:
+    """Read Message header field."""
+    value = msg[field]
+    if value == 'UNKNOWN':
+        return None
+    return value
+
+
+def _read_field_unescaped_from_msg(msg: Message, field: str) -> str | None:
+    """Read Message header field and apply rfc822_unescape."""
+    value = _read_field_from_msg(msg, field)
+    if value is None:
+        return value
+    return rfc822_unescape(value)
+
+
+def _read_list_from_msg(msg: Message, field: str) -> list[str] | None:
+    """Read Message header field and return all results as list."""
+    values = msg.get_all(field, None)
+    if values == []:
+        return None
+    return values
+
+
+def _read_payload_from_msg(msg: Message) -> str | None:
+    value = str(msg.get_payload()).strip()
+    if value == 'UNKNOWN' or not value:
+        return None
+    return value
+
+
+def read_pkg_file(self, file):
+    """Reads the metadata values from a file object."""
+    msg = message_from_file(file)
+
+    self.metadata_version = Version(msg['metadata-version'])
+    self.name = _read_field_from_msg(msg, 'name')
+    self.version = _read_field_from_msg(msg, 'version')
+    self.description = _read_field_from_msg(msg, 'summary')
+    # we are filling author only.
+    self.author = _read_field_from_msg(msg, 'author')
+    self.maintainer = None
+    self.author_email = _read_field_from_msg(msg, 'author-email')
+    self.maintainer_email = None
+    self.url = _read_field_from_msg(msg, 'home-page')
+    self.download_url = _read_field_from_msg(msg, 'download-url')
+    self.license = _read_field_unescaped_from_msg(msg, 'license')
+
+    self.long_description = _read_field_unescaped_from_msg(msg, 'description')
+    if self.long_description is None and self.metadata_version >= Version('2.1'):
+        self.long_description = _read_payload_from_msg(msg)
+    self.description = _read_field_from_msg(msg, 'summary')
+
+    if 'keywords' in msg:
+        self.keywords = _read_field_from_msg(msg, 'keywords').split(',')
+
+    self.platforms = _read_list_from_msg(msg, 'platform')
+    self.classifiers = _read_list_from_msg(msg, 'classifier')
+
+    # PEP 314 - these fields only exist in 1.1
+    if self.metadata_version == Version('1.1'):
+        self.requires = _read_list_from_msg(msg, 'requires')
+        self.provides = _read_list_from_msg(msg, 'provides')
+        self.obsoletes = _read_list_from_msg(msg, 'obsoletes')
+    else:
+        self.requires = None
+        self.provides = None
+        self.obsoletes = None
+
+    self.license_files = _read_list_from_msg(msg, 'license-file')
+
+
+def single_line(val):
+    """
+    Quick and dirty validation for Summary pypa/setuptools#1390.
+    """
+    if '\n' in val:
+        # TODO: Replace with `raise ValueError("newlines not allowed")`
+        # after reviewing #2893.
+        msg = "newlines are not allowed in `summary` and will break in the future"
+        SetuptoolsDeprecationWarning.emit("Invalid config.", msg)
+        # due_date is undefined. Controversial change, there was a lot of push back.
+        val = val.strip().split('\n')[0]
+    return val
+
+
+def write_pkg_info(self, base_dir):
+    """Write the PKG-INFO file into the release tree."""
+    temp = ""
+    final = os.path.join(base_dir, 'PKG-INFO')
+    try:
+        # Use a temporary file while writing to avoid race conditions
+        # (e.g. `importlib.metadata` reading `.egg-info/PKG-INFO`):
+        with NamedTemporaryFile("w", encoding="utf-8", dir=base_dir, delete=False) as f:
+            temp = f.name
+            self.write_pkg_file(f)
+        permissions = stat.S_IMODE(os.lstat(temp).st_mode)
+        os.chmod(temp, permissions | stat.S_IRGRP | stat.S_IROTH)
+        os.replace(temp, final)  # atomic operation.
+    finally:
+        if temp and os.path.exists(temp):
+            os.remove(temp)
+
+
+# Based on Python 3.5 version
+def write_pkg_file(self, file):  # noqa: C901  # is too complex (14)  # FIXME
+    """Write the PKG-INFO format data to a file object."""
+    version = self.get_metadata_version()
+
+    def write_field(key, value):
+        file.write(f"{key}: {value}\n")
+
+    write_field('Metadata-Version', str(version))
+    write_field('Name', self.get_name())
+    write_field('Version', self.get_version())
+
+    summary = self.get_description()
+    if summary:
+        write_field('Summary', single_line(summary))
+
+    optional_fields = (
+        ('Home-page', 'url'),
+        ('Download-URL', 'download_url'),
+        ('Author', 'author'),
+        ('Author-email', 'author_email'),
+        ('Maintainer', 'maintainer'),
+        ('Maintainer-email', 'maintainer_email'),
+    )
+
+    for field, attr in optional_fields:
+        attr_val = getattr(self, attr, None)
+        if attr_val is not None:
+            write_field(field, attr_val)
+
+    license = self.get_license()
+    if license:
+        write_field('License', rfc822_escape(license))
+
+    for label, url in self.project_urls.items():
+        write_field('Project-URL', f'{label}, {url}')
+
+    keywords = ','.join(self.get_keywords())
+    if keywords:
+        write_field('Keywords', keywords)
+
+    platforms = self.get_platforms() or []
+    for platform in platforms:
+        write_field('Platform', platform)
+
+    self._write_list(file, 'Classifier', self.get_classifiers())
+
+    # PEP 314
+    self._write_list(file, 'Requires', self.get_requires())
+    self._write_list(file, 'Provides', self.get_provides())
+    self._write_list(file, 'Obsoletes', self.get_obsoletes())
+
+    # Setuptools specific for PEP 345
+    if hasattr(self, 'python_requires'):
+        write_field('Requires-Python', self.python_requires)
+
+    # PEP 566
+    if self.long_description_content_type:
+        write_field('Description-Content-Type', self.long_description_content_type)
+
+    self._write_list(file, 'License-File', self.license_files or [])
+    _write_requirements(self, file)
+
+    for field, attr in _POSSIBLE_DYNAMIC_FIELDS.items():
+        if (val := getattr(self, attr, None)) and not is_static(val):
+            write_field('Dynamic', field)
+
+    long_description = self.get_long_description()
+    if long_description:
+        file.write(f"\n{long_description}")
+        if not long_description.endswith("\n"):
+            file.write("\n")
+
+
+def _write_requirements(self, file):
+    for req in _reqs.parse(self.install_requires):
+        file.write(f"Requires-Dist: {req}\n")
+
+    processed_extras = {}
+    for augmented_extra, reqs in self.extras_require.items():
+        # Historically, setuptools allows "augmented extras": `<extra>:<condition>`
+        unsafe_extra, _, condition = augmented_extra.partition(":")
+        unsafe_extra = unsafe_extra.strip()
+        extra = _normalization.safe_extra(unsafe_extra)
+
+        if extra:
+            _write_provides_extra(file, processed_extras, extra, unsafe_extra)
+        for req in _reqs.parse_strings(reqs):
+            r = _include_extra(req, extra, condition.strip())
+            file.write(f"Requires-Dist: {r}\n")
+
+    return processed_extras
+
+
+def _include_extra(req: str, extra: str, condition: str) -> Requirement:
+    r = Requirement(req)  # create a fresh object that can be modified
+    parts = (
+        f"({r.marker})" if r.marker else None,
+        f"({condition})" if condition else None,
+        f"extra == {extra!r}" if extra else None,
+    )
+    r.marker = Marker(" and ".join(x for x in parts if x))
+    return r
+
+
+def _write_provides_extra(file, processed_extras, safe, unsafe):
+    previous = processed_extras.get(safe)
+    if previous == unsafe:
+        SetuptoolsDeprecationWarning.emit(
+            'Ambiguity during "extra" normalization for dependencies.',
+            f"""
+            {previous!r} and {unsafe!r} normalize to the same value:\n
+                {safe!r}\n
+            In future versions, setuptools might halt the build process.
+            """,
+            see_url="https://peps.python.org/pep-0685/",
+        )
+    else:
+        processed_extras[safe] = unsafe
+        file.write(f"Provides-Extra: {safe}\n")
+
+
+# from pypa/distutils#244; needed only until that logic is always available
+def get_fullname(self):
+    return _distribution_fullname(self.get_name(), self.get_version())
+
+
+def _distribution_fullname(name: str, version: str) -> str:
+    """
+    >>> _distribution_fullname('setup.tools', '1.0-2')
+    'setup_tools-1.0.post2'
+    >>> _distribution_fullname('setup-tools', '1.2post2')
+    'setup_tools-1.2.post2'
+    >>> _distribution_fullname('setup-tools', '1.0-r2')
+    'setup_tools-1.0.post2'
+    >>> _distribution_fullname('setup.tools', '1.0.post')
+    'setup_tools-1.0.post0'
+    >>> _distribution_fullname('setup.tools', '1.0+ubuntu-1')
+    'setup_tools-1.0+ubuntu.1'
+    """
+    return "{}-{}".format(
+        canonicalize_name(name).replace('-', '_'),
+        canonicalize_version(version, strip_trailing_zero=False),
+    )
+
+
+_POSSIBLE_DYNAMIC_FIELDS = {
+    # Core Metadata Field x related Distribution attribute
+    "author": "author",
+    "author-email": "author_email",
+    "classifier": "classifiers",
+    "description": "long_description",
+    "description-content-type": "long_description_content_type",
+    "download-url": "download_url",
+    "home-page": "url",
+    "keywords": "keywords",
+    "license": "license",
+    # "license-file": "license_files", # XXX: does PEP 639 exempt Dynamic ??
+    "maintainer": "maintainer",
+    "maintainer-email": "maintainer_email",
+    "obsoletes": "obsoletes",
+    # "obsoletes-dist": "obsoletes_dist",  # NOT USED
+    "platform": "platforms",
+    "project-url": "project_urls",
+    "provides": "provides",
+    # "provides-dist": "provides_dist",  # NOT USED
+    "provides-extra": "extras_require",
+    "requires": "requires",
+    "requires-dist": "install_requires",
+    # "requires-external": "requires_external",  # NOT USED
+    "requires-python": "python_requires",
+    "summary": "description",
+    # "supported-platform": "supported_platforms",  # NOT USED
+}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__init__.py
new file mode 100644
index 00000000..e374d5c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__init__.py
@@ -0,0 +1,14 @@
+import importlib
+import sys
+
+__version__, _, _ = sys.version.partition(' ')
+
+
+try:
+    # Allow Debian and pkgsrc (only) to customize system
+    # behavior. Ref pypa/distutils#2 and pypa/distutils#16.
+    # This hook is deprecated and no other environments
+    # should use it.
+    importlib.import_module('_distutils_system_mod')
+except ImportError:
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..69620337
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_log.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_log.cpython-312.pyc
new file mode 100644
index 00000000..a7c92e91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_log.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_macos_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_macos_compat.cpython-312.pyc
new file mode 100644
index 00000000..2f08eff4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_macos_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_modified.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_modified.cpython-312.pyc
new file mode 100644
index 00000000..89a0102c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_modified.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_msvccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_msvccompiler.cpython-312.pyc
new file mode 100644
index 00000000..d236b536
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/_msvccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/archive_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/archive_util.cpython-312.pyc
new file mode 100644
index 00000000..51fa8016
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/archive_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/ccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/ccompiler.cpython-312.pyc
new file mode 100644
index 00000000..01fbe08a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/ccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cmd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cmd.cpython-312.pyc
new file mode 100644
index 00000000..5657c528
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cmd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/core.cpython-312.pyc
new file mode 100644
index 00000000..33e171d8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cygwinccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cygwinccompiler.cpython-312.pyc
new file mode 100644
index 00000000..64c8aef1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/cygwinccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/debug.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/debug.cpython-312.pyc
new file mode 100644
index 00000000..09e0512b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/debug.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dep_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dep_util.cpython-312.pyc
new file mode 100644
index 00000000..5df2e43e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dep_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dir_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dir_util.cpython-312.pyc
new file mode 100644
index 00000000..ce0eba09
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dir_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dist.cpython-312.pyc
new file mode 100644
index 00000000..d5a03e8d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/dist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/errors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/errors.cpython-312.pyc
new file mode 100644
index 00000000..8ee6f15e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/errors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/extension.cpython-312.pyc
new file mode 100644
index 00000000..afda0081
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/fancy_getopt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/fancy_getopt.cpython-312.pyc
new file mode 100644
index 00000000..d55f4d83
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/fancy_getopt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/file_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/file_util.cpython-312.pyc
new file mode 100644
index 00000000..948b3268
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/file_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/filelist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/filelist.cpython-312.pyc
new file mode 100644
index 00000000..3c72f261
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/filelist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/log.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/log.cpython-312.pyc
new file mode 100644
index 00000000..928ddbde
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/log.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/spawn.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/spawn.cpython-312.pyc
new file mode 100644
index 00000000..6133c806
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/spawn.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/sysconfig.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/sysconfig.cpython-312.pyc
new file mode 100644
index 00000000..de58a499
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/sysconfig.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/text_file.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/text_file.cpython-312.pyc
new file mode 100644
index 00000000..2f13c11d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/text_file.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/unixccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/unixccompiler.cpython-312.pyc
new file mode 100644
index 00000000..cfb56db9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/unixccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..13b9545d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..14a280fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/versionpredicate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/versionpredicate.cpython-312.pyc
new file mode 100644
index 00000000..c9568198
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/versionpredicate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/zosccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/zosccompiler.cpython-312.pyc
new file mode 100644
index 00000000..1c526682
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/__pycache__/zosccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_log.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_log.py
new file mode 100644
index 00000000..0148f157
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_log.py
@@ -0,0 +1,3 @@
+import logging
+
+log = logging.getLogger()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_macos_compat.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_macos_compat.py
new file mode 100644
index 00000000..76ecb96a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_macos_compat.py
@@ -0,0 +1,12 @@
+import importlib
+import sys
+
+
+def bypass_compiler_fixup(cmd, args):
+    return cmd
+
+
+if sys.platform == 'darwin':
+    compiler_fixup = importlib.import_module('_osx_support').compiler_fixup
+else:
+    compiler_fixup = bypass_compiler_fixup
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_modified.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_modified.py
new file mode 100644
index 00000000..7cdca939
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_modified.py
@@ -0,0 +1,73 @@
+"""Timestamp comparison of files and groups of files."""
+
+import functools
+import os.path
+
+from jaraco.functools import splat
+
+from .compat.py39 import zip_strict
+from .errors import DistutilsFileError
+
+
+def _newer(source, target):
+    return not os.path.exists(target) or (
+        os.path.getmtime(source) > os.path.getmtime(target)
+    )
+
+
+def newer(source, target):
+    """
+    Is source modified more recently than target.
+
+    Returns True if 'source' is modified more recently than
+    'target' or if 'target' does not exist.
+
+    Raises DistutilsFileError if 'source' does not exist.
+    """
+    if not os.path.exists(source):
+        raise DistutilsFileError(f"file '{os.path.abspath(source)}' does not exist")
+
+    return _newer(source, target)
+
+
+def newer_pairwise(sources, targets, newer=newer):
+    """
+    Filter filenames where sources are newer than targets.
+
+    Walk two filename iterables in parallel, testing if each source is newer
+    than its corresponding target.  Returns a pair of lists (sources,
+    targets) where source is newer than target, according to the semantics
+    of 'newer()'.
+    """
+    newer_pairs = filter(splat(newer), zip_strict(sources, targets))
+    return tuple(map(list, zip(*newer_pairs))) or ([], [])
+
+
+def newer_group(sources, target, missing='error'):
+    """
+    Is target out-of-date with respect to any file in sources.
+
+    Return True if 'target' is out-of-date with respect to any file
+    listed in 'sources'. In other words, if 'target' exists and is newer
+    than every file in 'sources', return False; otherwise return True.
+    ``missing`` controls how to handle a missing source file:
+
+    - error (default): allow the ``stat()`` call to fail.
+    - ignore: silently disregard any missing source files.
+    - newer: treat missing source files as "target out of date". This
+      mode is handy in "dry-run" mode: it will pretend to carry out
+      commands that wouldn't work because inputs are missing, but
+      that doesn't matter because dry-run won't run the commands.
+    """
+
+    def missing_as_newer(source):
+        return missing == 'newer' and not os.path.exists(source)
+
+    ignored = os.path.exists if missing == 'ignore' else None
+    return not os.path.exists(target) or any(
+        missing_as_newer(source) or _newer(source, target)
+        for source in filter(ignored, sources)
+    )
+
+
+newer_pairwise_group = functools.partial(newer_pairwise, newer=newer_group)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_msvccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_msvccompiler.py
new file mode 100644
index 00000000..97b067c6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/_msvccompiler.py
@@ -0,0 +1,604 @@
+"""distutils._msvccompiler
+
+Contains MSVCCompiler, an implementation of the abstract CCompiler class
+for Microsoft Visual Studio 2015.
+
+This module requires VS 2015 or later.
+"""
+
+# Written by Perry Stoll
+# hacked by Robin Becker and Thomas Heller to do a better job of
+#   finding DevStudio (through the registry)
+# ported to VS 2005 and VS 2008 by Christian Heimes
+# ported to VS 2015 by Steve Dower
+
+import contextlib
+import os
+import subprocess
+import unittest.mock as mock
+import warnings
+
+with contextlib.suppress(ImportError):
+    import winreg
+
+from itertools import count
+
+from ._log import log
+from .ccompiler import CCompiler, gen_lib_options
+from .errors import (
+    CompileError,
+    DistutilsExecError,
+    DistutilsPlatformError,
+    LibError,
+    LinkError,
+)
+from .util import get_host_platform, get_platform
+
+
+def _find_vc2015():
+    try:
+        key = winreg.OpenKeyEx(
+            winreg.HKEY_LOCAL_MACHINE,
+            r"Software\Microsoft\VisualStudio\SxS\VC7",
+            access=winreg.KEY_READ | winreg.KEY_WOW64_32KEY,
+        )
+    except OSError:
+        log.debug("Visual C++ is not registered")
+        return None, None
+
+    best_version = 0
+    best_dir = None
+    with key:
+        for i in count():
+            try:
+                v, vc_dir, vt = winreg.EnumValue(key, i)
+            except OSError:
+                break
+            if v and vt == winreg.REG_SZ and os.path.isdir(vc_dir):
+                try:
+                    version = int(float(v))
+                except (ValueError, TypeError):
+                    continue
+                if version >= 14 and version > best_version:
+                    best_version, best_dir = version, vc_dir
+    return best_version, best_dir
+
+
+def _find_vc2017():
+    """Returns "15, path" based on the result of invoking vswhere.exe
+    If no install is found, returns "None, None"
+
+    The version is returned to avoid unnecessarily changing the function
+    result. It may be ignored when the path is not None.
+
+    If vswhere.exe is not available, by definition, VS 2017 is not
+    installed.
+    """
+    root = os.environ.get("ProgramFiles(x86)") or os.environ.get("ProgramFiles")
+    if not root:
+        return None, None
+
+    variant = 'arm64' if get_platform() == 'win-arm64' else 'x86.x64'
+    suitable_components = (
+        f"Microsoft.VisualStudio.Component.VC.Tools.{variant}",
+        "Microsoft.VisualStudio.Workload.WDExpress",
+    )
+
+    for component in suitable_components:
+        # Workaround for `-requiresAny` (only available on VS 2017 > 15.6)
+        with contextlib.suppress(
+            subprocess.CalledProcessError, OSError, UnicodeDecodeError
+        ):
+            path = (
+                subprocess.check_output([
+                    os.path.join(
+                        root, "Microsoft Visual Studio", "Installer", "vswhere.exe"
+                    ),
+                    "-latest",
+                    "-prerelease",
+                    "-requires",
+                    component,
+                    "-property",
+                    "installationPath",
+                    "-products",
+                    "*",
+                ])
+                .decode(encoding="mbcs", errors="strict")
+                .strip()
+            )
+
+            path = os.path.join(path, "VC", "Auxiliary", "Build")
+            if os.path.isdir(path):
+                return 15, path
+
+    return None, None  # no suitable component found
+
+
+PLAT_SPEC_TO_RUNTIME = {
+    'x86': 'x86',
+    'x86_amd64': 'x64',
+    'x86_arm': 'arm',
+    'x86_arm64': 'arm64',
+}
+
+
+def _find_vcvarsall(plat_spec):
+    # bpo-38597: Removed vcruntime return value
+    _, best_dir = _find_vc2017()
+
+    if not best_dir:
+        best_version, best_dir = _find_vc2015()
+
+    if not best_dir:
+        log.debug("No suitable Visual C++ version found")
+        return None, None
+
+    vcvarsall = os.path.join(best_dir, "vcvarsall.bat")
+    if not os.path.isfile(vcvarsall):
+        log.debug("%s cannot be found", vcvarsall)
+        return None, None
+
+    return vcvarsall, None
+
+
+def _get_vc_env(plat_spec):
+    if os.getenv("DISTUTILS_USE_SDK"):
+        return {key.lower(): value for key, value in os.environ.items()}
+
+    vcvarsall, _ = _find_vcvarsall(plat_spec)
+    if not vcvarsall:
+        raise DistutilsPlatformError(
+            'Microsoft Visual C++ 14.0 or greater is required. '
+            'Get it with "Microsoft C++ Build Tools": '
+            'https://visualstudio.microsoft.com/visual-cpp-build-tools/'
+        )
+
+    try:
+        out = subprocess.check_output(
+            f'cmd /u /c "{vcvarsall}" {plat_spec} && set',
+            stderr=subprocess.STDOUT,
+        ).decode('utf-16le', errors='replace')
+    except subprocess.CalledProcessError as exc:
+        log.error(exc.output)
+        raise DistutilsPlatformError(f"Error executing {exc.cmd}")
+
+    env = {
+        key.lower(): value
+        for key, _, value in (line.partition('=') for line in out.splitlines())
+        if key and value
+    }
+
+    return env
+
+
+def _find_exe(exe, paths=None):
+    """Return path to an MSVC executable program.
+
+    Tries to find the program in several places: first, one of the
+    MSVC program search paths from the registry; next, the directories
+    in the PATH environment variable.  If any of those work, return an
+    absolute path that is known to exist.  If none of them work, just
+    return the original program name, 'exe'.
+    """
+    if not paths:
+        paths = os.getenv('path').split(os.pathsep)
+    for p in paths:
+        fn = os.path.join(os.path.abspath(p), exe)
+        if os.path.isfile(fn):
+            return fn
+    return exe
+
+
+_vcvars_names = {
+    'win32': 'x86',
+    'win-amd64': 'amd64',
+    'win-arm32': 'arm',
+    'win-arm64': 'arm64',
+}
+
+
+def _get_vcvars_spec(host_platform, platform):
+    """
+    Given a host platform and platform, determine the spec for vcvarsall.
+
+    Uses the native MSVC host if the host platform would need expensive
+    emulation for x86.
+
+    >>> _get_vcvars_spec('win-arm64', 'win32')
+    'arm64_x86'
+    >>> _get_vcvars_spec('win-arm64', 'win-amd64')
+    'arm64_amd64'
+
+    Otherwise, always cross-compile from x86 to work with the
+    lighter-weight MSVC installs that do not include native 64-bit tools.
+
+    >>> _get_vcvars_spec('win32', 'win32')
+    'x86'
+    >>> _get_vcvars_spec('win-arm32', 'win-arm32')
+    'x86_arm'
+    >>> _get_vcvars_spec('win-amd64', 'win-arm64')
+    'x86_arm64'
+    """
+    if host_platform != 'win-arm64':
+        host_platform = 'win32'
+    vc_hp = _vcvars_names[host_platform]
+    vc_plat = _vcvars_names[platform]
+    return vc_hp if vc_hp == vc_plat else f'{vc_hp}_{vc_plat}'
+
+
+class MSVCCompiler(CCompiler):
+    """Concrete class that implements an interface to Microsoft Visual C++,
+    as defined by the CCompiler abstract class."""
+
+    compiler_type = 'msvc'
+
+    # Just set this so CCompiler's constructor doesn't barf.  We currently
+    # don't use the 'set_executables()' bureaucracy provided by CCompiler,
+    # as it really isn't necessary for this sort of single-compiler class.
+    # Would be nice to have a consistent interface with UnixCCompiler,
+    # though, so it's worth thinking about.
+    executables = {}
+
+    # Private class data (need to distinguish C from C++ source for compiler)
+    _c_extensions = ['.c']
+    _cpp_extensions = ['.cc', '.cpp', '.cxx']
+    _rc_extensions = ['.rc']
+    _mc_extensions = ['.mc']
+
+    # Needed for the filename generation methods provided by the
+    # base class, CCompiler.
+    src_extensions = _c_extensions + _cpp_extensions + _rc_extensions + _mc_extensions
+    res_extension = '.res'
+    obj_extension = '.obj'
+    static_lib_extension = '.lib'
+    shared_lib_extension = '.dll'
+    static_lib_format = shared_lib_format = '%s%s'
+    exe_extension = '.exe'
+
+    def __init__(self, verbose=False, dry_run=False, force=False):
+        super().__init__(verbose, dry_run, force)
+        # target platform (.plat_name is consistent with 'bdist')
+        self.plat_name = None
+        self.initialized = False
+
+    @classmethod
+    def _configure(cls, vc_env):
+        """
+        Set class-level include/lib dirs.
+        """
+        cls.include_dirs = cls._parse_path(vc_env.get('include', ''))
+        cls.library_dirs = cls._parse_path(vc_env.get('lib', ''))
+
+    @staticmethod
+    def _parse_path(val):
+        return [dir.rstrip(os.sep) for dir in val.split(os.pathsep) if dir]
+
+    def initialize(self, plat_name=None):
+        # multi-init means we would need to check platform same each time...
+        assert not self.initialized, "don't init multiple times"
+        if plat_name is None:
+            plat_name = get_platform()
+        # sanity check for platforms to prevent obscure errors later.
+        if plat_name not in _vcvars_names:
+            raise DistutilsPlatformError(
+                f"--plat-name must be one of {tuple(_vcvars_names)}"
+            )
+
+        plat_spec = _get_vcvars_spec(get_host_platform(), plat_name)
+
+        vc_env = _get_vc_env(plat_spec)
+        if not vc_env:
+            raise DistutilsPlatformError(
+                "Unable to find a compatible Visual Studio installation."
+            )
+        self._configure(vc_env)
+
+        self._paths = vc_env.get('path', '')
+        paths = self._paths.split(os.pathsep)
+        self.cc = _find_exe("cl.exe", paths)
+        self.linker = _find_exe("link.exe", paths)
+        self.lib = _find_exe("lib.exe", paths)
+        self.rc = _find_exe("rc.exe", paths)  # resource compiler
+        self.mc = _find_exe("mc.exe", paths)  # message compiler
+        self.mt = _find_exe("mt.exe", paths)  # message compiler
+
+        self.preprocess_options = None
+        # bpo-38597: Always compile with dynamic linking
+        # Future releases of Python 3.x will include all past
+        # versions of vcruntime*.dll for compatibility.
+        self.compile_options = ['/nologo', '/O2', '/W3', '/GL', '/DNDEBUG', '/MD']
+
+        self.compile_options_debug = [
+            '/nologo',
+            '/Od',
+            '/MDd',
+            '/Zi',
+            '/W3',
+            '/D_DEBUG',
+        ]
+
+        ldflags = ['/nologo', '/INCREMENTAL:NO', '/LTCG']
+
+        ldflags_debug = ['/nologo', '/INCREMENTAL:NO', '/LTCG', '/DEBUG:FULL']
+
+        self.ldflags_exe = [*ldflags, '/MANIFEST:EMBED,ID=1']
+        self.ldflags_exe_debug = [*ldflags_debug, '/MANIFEST:EMBED,ID=1']
+        self.ldflags_shared = [
+            *ldflags,
+            '/DLL',
+            '/MANIFEST:EMBED,ID=2',
+            '/MANIFESTUAC:NO',
+        ]
+        self.ldflags_shared_debug = [
+            *ldflags_debug,
+            '/DLL',
+            '/MANIFEST:EMBED,ID=2',
+            '/MANIFESTUAC:NO',
+        ]
+        self.ldflags_static = [*ldflags]
+        self.ldflags_static_debug = [*ldflags_debug]
+
+        self._ldflags = {
+            (CCompiler.EXECUTABLE, None): self.ldflags_exe,
+            (CCompiler.EXECUTABLE, False): self.ldflags_exe,
+            (CCompiler.EXECUTABLE, True): self.ldflags_exe_debug,
+            (CCompiler.SHARED_OBJECT, None): self.ldflags_shared,
+            (CCompiler.SHARED_OBJECT, False): self.ldflags_shared,
+            (CCompiler.SHARED_OBJECT, True): self.ldflags_shared_debug,
+            (CCompiler.SHARED_LIBRARY, None): self.ldflags_static,
+            (CCompiler.SHARED_LIBRARY, False): self.ldflags_static,
+            (CCompiler.SHARED_LIBRARY, True): self.ldflags_static_debug,
+        }
+
+        self.initialized = True
+
+    # -- Worker methods ------------------------------------------------
+
+    @property
+    def out_extensions(self):
+        return {
+            **super().out_extensions,
+            **{
+                ext: self.res_extension
+                for ext in self._rc_extensions + self._mc_extensions
+            },
+        }
+
+    def compile(  # noqa: C901
+        self,
+        sources,
+        output_dir=None,
+        macros=None,
+        include_dirs=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        depends=None,
+    ):
+        if not self.initialized:
+            self.initialize()
+        compile_info = self._setup_compile(
+            output_dir, macros, include_dirs, sources, depends, extra_postargs
+        )
+        macros, objects, extra_postargs, pp_opts, build = compile_info
+
+        compile_opts = extra_preargs or []
+        compile_opts.append('/c')
+        if debug:
+            compile_opts.extend(self.compile_options_debug)
+        else:
+            compile_opts.extend(self.compile_options)
+
+        add_cpp_opts = False
+
+        for obj in objects:
+            try:
+                src, ext = build[obj]
+            except KeyError:
+                continue
+            if debug:
+                # pass the full pathname to MSVC in debug mode,
+                # this allows the debugger to find the source file
+                # without asking the user to browse for it
+                src = os.path.abspath(src)
+
+            if ext in self._c_extensions:
+                input_opt = "/Tc" + src
+            elif ext in self._cpp_extensions:
+                input_opt = "/Tp" + src
+                add_cpp_opts = True
+            elif ext in self._rc_extensions:
+                # compile .RC to .RES file
+                input_opt = src
+                output_opt = "/fo" + obj
+                try:
+                    self.spawn([self.rc] + pp_opts + [output_opt, input_opt])
+                except DistutilsExecError as msg:
+                    raise CompileError(msg)
+                continue
+            elif ext in self._mc_extensions:
+                # Compile .MC to .RC file to .RES file.
+                #   * '-h dir' specifies the directory for the
+                #     generated include file
+                #   * '-r dir' specifies the target directory of the
+                #     generated RC file and the binary message resource
+                #     it includes
+                #
+                # For now (since there are no options to change this),
+                # we use the source-directory for the include file and
+                # the build directory for the RC file and message
+                # resources. This works at least for win32all.
+                h_dir = os.path.dirname(src)
+                rc_dir = os.path.dirname(obj)
+                try:
+                    # first compile .MC to .RC and .H file
+                    self.spawn([self.mc, '-h', h_dir, '-r', rc_dir, src])
+                    base, _ = os.path.splitext(os.path.basename(src))
+                    rc_file = os.path.join(rc_dir, base + '.rc')
+                    # then compile .RC to .RES file
+                    self.spawn([self.rc, "/fo" + obj, rc_file])
+
+                except DistutilsExecError as msg:
+                    raise CompileError(msg)
+                continue
+            else:
+                # how to handle this file?
+                raise CompileError(f"Don't know how to compile {src} to {obj}")
+
+            args = [self.cc] + compile_opts + pp_opts
+            if add_cpp_opts:
+                args.append('/EHsc')
+            args.extend((input_opt, "/Fo" + obj))
+            args.extend(extra_postargs)
+
+            try:
+                self.spawn(args)
+            except DistutilsExecError as msg:
+                raise CompileError(msg)
+
+        return objects
+
+    def create_static_lib(
+        self, objects, output_libname, output_dir=None, debug=False, target_lang=None
+    ):
+        if not self.initialized:
+            self.initialize()
+        objects, output_dir = self._fix_object_args(objects, output_dir)
+        output_filename = self.library_filename(output_libname, output_dir=output_dir)
+
+        if self._need_link(objects, output_filename):
+            lib_args = objects + ['/OUT:' + output_filename]
+            if debug:
+                pass  # XXX what goes here?
+            try:
+                log.debug('Executing "%s" %s', self.lib, ' '.join(lib_args))
+                self.spawn([self.lib] + lib_args)
+            except DistutilsExecError as msg:
+                raise LibError(msg)
+        else:
+            log.debug("skipping %s (up-to-date)", output_filename)
+
+    def link(
+        self,
+        target_desc,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        if not self.initialized:
+            self.initialize()
+        objects, output_dir = self._fix_object_args(objects, output_dir)
+        fixed_args = self._fix_lib_args(libraries, library_dirs, runtime_library_dirs)
+        libraries, library_dirs, runtime_library_dirs = fixed_args
+
+        if runtime_library_dirs:
+            self.warn(
+                "I don't know what to do with 'runtime_library_dirs': "
+                + str(runtime_library_dirs)
+            )
+
+        lib_opts = gen_lib_options(self, library_dirs, runtime_library_dirs, libraries)
+        if output_dir is not None:
+            output_filename = os.path.join(output_dir, output_filename)
+
+        if self._need_link(objects, output_filename):
+            ldflags = self._ldflags[target_desc, debug]
+
+            export_opts = ["/EXPORT:" + sym for sym in (export_symbols or [])]
+
+            ld_args = (
+                ldflags + lib_opts + export_opts + objects + ['/OUT:' + output_filename]
+            )
+
+            # The MSVC linker generates .lib and .exp files, which cannot be
+            # suppressed by any linker switches. The .lib files may even be
+            # needed! Make sure they are generated in the temporary build
+            # directory. Since they have different names for debug and release
+            # builds, they can go into the same directory.
+            build_temp = os.path.dirname(objects[0])
+            if export_symbols is not None:
+                (dll_name, dll_ext) = os.path.splitext(
+                    os.path.basename(output_filename)
+                )
+                implib_file = os.path.join(build_temp, self.library_filename(dll_name))
+                ld_args.append('/IMPLIB:' + implib_file)
+
+            if extra_preargs:
+                ld_args[:0] = extra_preargs
+            if extra_postargs:
+                ld_args.extend(extra_postargs)
+
+            output_dir = os.path.dirname(os.path.abspath(output_filename))
+            self.mkpath(output_dir)
+            try:
+                log.debug('Executing "%s" %s', self.linker, ' '.join(ld_args))
+                self.spawn([self.linker] + ld_args)
+            except DistutilsExecError as msg:
+                raise LinkError(msg)
+        else:
+            log.debug("skipping %s (up-to-date)", output_filename)
+
+    def spawn(self, cmd):
+        env = dict(os.environ, PATH=self._paths)
+        with self._fallback_spawn(cmd, env) as fallback:
+            return super().spawn(cmd, env=env)
+        return fallback.value
+
+    @contextlib.contextmanager
+    def _fallback_spawn(self, cmd, env):
+        """
+        Discovered in pypa/distutils#15, some tools monkeypatch the compiler,
+        so the 'env' kwarg causes a TypeError. Detect this condition and
+        restore the legacy, unsafe behavior.
+        """
+        bag = type('Bag', (), {})()
+        try:
+            yield bag
+        except TypeError as exc:
+            if "unexpected keyword argument 'env'" not in str(exc):
+                raise
+        else:
+            return
+        warnings.warn("Fallback spawn triggered. Please update distutils monkeypatch.")
+        with mock.patch.dict('os.environ', env):
+            bag.value = super().spawn(cmd)
+
+    # -- Miscellaneous methods -----------------------------------------
+    # These are all used by the 'gen_lib_options() function, in
+    # ccompiler.py.
+
+    def library_dir_option(self, dir):
+        return "/LIBPATH:" + dir
+
+    def runtime_library_dir_option(self, dir):
+        raise DistutilsPlatformError(
+            "don't know how to set runtime library search path for MSVC"
+        )
+
+    def library_option(self, lib):
+        return self.library_filename(lib)
+
+    def find_library_file(self, dirs, lib, debug=False):
+        # Prefer a debugging library if found (and requested), but deal
+        # with it if we don't have one.
+        if debug:
+            try_names = [lib + "_d", lib]
+        else:
+            try_names = [lib]
+        for dir in dirs:
+            for name in try_names:
+                libfile = os.path.join(dir, self.library_filename(name))
+                if os.path.isfile(libfile):
+                    return libfile
+        else:
+            # Oops, didn't find it in *any* of 'dirs'
+            return None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/archive_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/archive_util.py
new file mode 100644
index 00000000..5bb6df76
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/archive_util.py
@@ -0,0 +1,264 @@
+"""distutils.archive_util
+
+Utility functions for creating archive files (tarballs, zip files,
+that sort of thing)."""
+
+import os
+
+try:
+    import zipfile
+except ImportError:
+    zipfile = None
+
+
+from ._log import log
+from .dir_util import mkpath
+from .errors import DistutilsExecError
+from .spawn import spawn
+
+try:
+    from pwd import getpwnam
+except ImportError:
+    getpwnam = None
+
+try:
+    from grp import getgrnam
+except ImportError:
+    getgrnam = None
+
+
+def _get_gid(name):
+    """Returns a gid, given a group name."""
+    if getgrnam is None or name is None:
+        return None
+    try:
+        result = getgrnam(name)
+    except KeyError:
+        result = None
+    if result is not None:
+        return result[2]
+    return None
+
+
+def _get_uid(name):
+    """Returns an uid, given a user name."""
+    if getpwnam is None or name is None:
+        return None
+    try:
+        result = getpwnam(name)
+    except KeyError:
+        result = None
+    if result is not None:
+        return result[2]
+    return None
+
+
+def make_tarball(
+    base_name,
+    base_dir,
+    compress="gzip",
+    verbose=False,
+    dry_run=False,
+    owner=None,
+    group=None,
+):
+    """Create a (possibly compressed) tar file from all the files under
+    'base_dir'.
+
+    'compress' must be "gzip" (the default), "bzip2", "xz", or None.
+
+    'owner' and 'group' can be used to define an owner and a group for the
+    archive that is being built. If not provided, the current owner and group
+    will be used.
+
+    The output tar file will be named 'base_dir' +  ".tar", possibly plus
+    the appropriate compression extension (".gz", ".bz2", ".xz" or ".Z").
+
+    Returns the output filename.
+    """
+    tar_compression = {
+        'gzip': 'gz',
+        'bzip2': 'bz2',
+        'xz': 'xz',
+        None: '',
+    }
+    compress_ext = {'gzip': '.gz', 'bzip2': '.bz2', 'xz': '.xz'}
+
+    # flags for compression program, each element of list will be an argument
+    if compress is not None and compress not in compress_ext.keys():
+        raise ValueError(
+            "bad value for 'compress': must be None, 'gzip', 'bzip2', 'xz'"
+        )
+
+    archive_name = base_name + '.tar'
+    archive_name += compress_ext.get(compress, '')
+
+    mkpath(os.path.dirname(archive_name), dry_run=dry_run)
+
+    # creating the tarball
+    import tarfile  # late import so Python build itself doesn't break
+
+    log.info('Creating tar archive')
+
+    uid = _get_uid(owner)
+    gid = _get_gid(group)
+
+    def _set_uid_gid(tarinfo):
+        if gid is not None:
+            tarinfo.gid = gid
+            tarinfo.gname = group
+        if uid is not None:
+            tarinfo.uid = uid
+            tarinfo.uname = owner
+        return tarinfo
+
+    if not dry_run:
+        tar = tarfile.open(archive_name, f'w|{tar_compression[compress]}')
+        try:
+            tar.add(base_dir, filter=_set_uid_gid)
+        finally:
+            tar.close()
+
+    return archive_name
+
+
+def make_zipfile(base_name, base_dir, verbose=False, dry_run=False):  # noqa: C901
+    """Create a zip file from all the files under 'base_dir'.
+
+    The output zip file will be named 'base_name' + ".zip".  Uses either the
+    "zipfile" Python module (if available) or the InfoZIP "zip" utility
+    (if installed and found on the default search path).  If neither tool is
+    available, raises DistutilsExecError.  Returns the name of the output zip
+    file.
+    """
+    zip_filename = base_name + ".zip"
+    mkpath(os.path.dirname(zip_filename), dry_run=dry_run)
+
+    # If zipfile module is not available, try spawning an external
+    # 'zip' command.
+    if zipfile is None:
+        if verbose:
+            zipoptions = "-r"
+        else:
+            zipoptions = "-rq"
+
+        try:
+            spawn(["zip", zipoptions, zip_filename, base_dir], dry_run=dry_run)
+        except DistutilsExecError:
+            # XXX really should distinguish between "couldn't find
+            # external 'zip' command" and "zip failed".
+            raise DistutilsExecError(
+                f"unable to create zip file '{zip_filename}': "
+                "could neither import the 'zipfile' module nor "
+                "find a standalone zip utility"
+            )
+
+    else:
+        log.info("creating '%s' and adding '%s' to it", zip_filename, base_dir)
+
+        if not dry_run:
+            try:
+                zip = zipfile.ZipFile(
+                    zip_filename, "w", compression=zipfile.ZIP_DEFLATED
+                )
+            except RuntimeError:
+                zip = zipfile.ZipFile(zip_filename, "w", compression=zipfile.ZIP_STORED)
+
+            with zip:
+                if base_dir != os.curdir:
+                    path = os.path.normpath(os.path.join(base_dir, ''))
+                    zip.write(path, path)
+                    log.info("adding '%s'", path)
+                for dirpath, dirnames, filenames in os.walk(base_dir):
+                    for name in dirnames:
+                        path = os.path.normpath(os.path.join(dirpath, name, ''))
+                        zip.write(path, path)
+                        log.info("adding '%s'", path)
+                    for name in filenames:
+                        path = os.path.normpath(os.path.join(dirpath, name))
+                        if os.path.isfile(path):
+                            zip.write(path, path)
+                            log.info("adding '%s'", path)
+
+    return zip_filename
+
+
+ARCHIVE_FORMATS = {
+    'gztar': (make_tarball, [('compress', 'gzip')], "gzip'ed tar-file"),
+    'bztar': (make_tarball, [('compress', 'bzip2')], "bzip2'ed tar-file"),
+    'xztar': (make_tarball, [('compress', 'xz')], "xz'ed tar-file"),
+    'ztar': (make_tarball, [('compress', 'compress')], "compressed tar file"),
+    'tar': (make_tarball, [('compress', None)], "uncompressed tar file"),
+    'zip': (make_zipfile, [], "ZIP file"),
+}
+
+
+def check_archive_formats(formats):
+    """Returns the first format from the 'format' list that is unknown.
+
+    If all formats are known, returns None
+    """
+    for format in formats:
+        if format not in ARCHIVE_FORMATS:
+            return format
+    return None
+
+
+def make_archive(
+    base_name,
+    format,
+    root_dir=None,
+    base_dir=None,
+    verbose=False,
+    dry_run=False,
+    owner=None,
+    group=None,
+):
+    """Create an archive file (eg. zip or tar).
+
+    'base_name' is the name of the file to create, minus any format-specific
+    extension; 'format' is the archive format: one of "zip", "tar", "gztar",
+    "bztar", "xztar", or "ztar".
+
+    'root_dir' is a directory that will be the root directory of the
+    archive; ie. we typically chdir into 'root_dir' before creating the
+    archive.  'base_dir' is the directory where we start archiving from;
+    ie. 'base_dir' will be the common prefix of all files and
+    directories in the archive.  'root_dir' and 'base_dir' both default
+    to the current directory.  Returns the name of the archive file.
+
+    'owner' and 'group' are used when creating a tar archive. By default,
+    uses the current owner and group.
+    """
+    save_cwd = os.getcwd()
+    if root_dir is not None:
+        log.debug("changing into '%s'", root_dir)
+        base_name = os.path.abspath(base_name)
+        if not dry_run:
+            os.chdir(root_dir)
+
+    if base_dir is None:
+        base_dir = os.curdir
+
+    kwargs = {'dry_run': dry_run}
+
+    try:
+        format_info = ARCHIVE_FORMATS[format]
+    except KeyError:
+        raise ValueError(f"unknown archive format '{format}'")
+
+    func = format_info[0]
+    kwargs.update(format_info[1])
+
+    if format != 'zip':
+        kwargs['owner'] = owner
+        kwargs['group'] = group
+
+    try:
+        filename = func(base_name, base_dir, **kwargs)
+    finally:
+        if root_dir is not None:
+            log.debug("changing back to '%s'", save_cwd)
+            os.chdir(save_cwd)
+
+    return filename
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/ccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/ccompiler.py
new file mode 100644
index 00000000..714f13d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/ccompiler.py
@@ -0,0 +1,1263 @@
+"""distutils.ccompiler
+
+Contains CCompiler, an abstract base class that defines the interface
+for the Distutils compiler abstraction model."""
+
+import os
+import pathlib
+import re
+import sys
+import types
+import warnings
+
+from more_itertools import always_iterable
+
+from ._log import log
+from ._modified import newer_group
+from .dir_util import mkpath
+from .errors import (
+    CompileError,
+    DistutilsModuleError,
+    DistutilsPlatformError,
+    LinkError,
+    UnknownFileError,
+)
+from .file_util import move_file
+from .spawn import spawn
+from .util import execute, is_mingw, split_quoted
+
+
+class CCompiler:
+    """Abstract base class to define the interface that must be implemented
+    by real compiler classes.  Also has some utility methods used by
+    several compiler classes.
+
+    The basic idea behind a compiler abstraction class is that each
+    instance can be used for all the compile/link steps in building a
+    single project.  Thus, attributes common to all of those compile and
+    link steps -- include directories, macros to define, libraries to link
+    against, etc. -- are attributes of the compiler instance.  To allow for
+    variability in how individual files are treated, most of those
+    attributes may be varied on a per-compilation or per-link basis.
+    """
+
+    # 'compiler_type' is a class attribute that identifies this class.  It
+    # keeps code that wants to know what kind of compiler it's dealing with
+    # from having to import all possible compiler classes just to do an
+    # 'isinstance'.  In concrete CCompiler subclasses, 'compiler_type'
+    # should really, really be one of the keys of the 'compiler_class'
+    # dictionary (see below -- used by the 'new_compiler()' factory
+    # function) -- authors of new compiler interface classes are
+    # responsible for updating 'compiler_class'!
+    compiler_type = None
+
+    # XXX things not handled by this compiler abstraction model:
+    #   * client can't provide additional options for a compiler,
+    #     e.g. warning, optimization, debugging flags.  Perhaps this
+    #     should be the domain of concrete compiler abstraction classes
+    #     (UnixCCompiler, MSVCCompiler, etc.) -- or perhaps the base
+    #     class should have methods for the common ones.
+    #   * can't completely override the include or library searchg
+    #     path, ie. no "cc -I -Idir1 -Idir2" or "cc -L -Ldir1 -Ldir2".
+    #     I'm not sure how widely supported this is even by Unix
+    #     compilers, much less on other platforms.  And I'm even less
+    #     sure how useful it is; maybe for cross-compiling, but
+    #     support for that is a ways off.  (And anyways, cross
+    #     compilers probably have a dedicated binary with the
+    #     right paths compiled in.  I hope.)
+    #   * can't do really freaky things with the library list/library
+    #     dirs, e.g. "-Ldir1 -lfoo -Ldir2 -lfoo" to link against
+    #     different versions of libfoo.a in different locations.  I
+    #     think this is useless without the ability to null out the
+    #     library search path anyways.
+
+    # Subclasses that rely on the standard filename generation methods
+    # implemented below should override these; see the comment near
+    # those methods ('object_filenames()' et. al.) for details:
+    src_extensions = None  # list of strings
+    obj_extension = None  # string
+    static_lib_extension = None
+    shared_lib_extension = None  # string
+    static_lib_format = None  # format string
+    shared_lib_format = None  # prob. same as static_lib_format
+    exe_extension = None  # string
+
+    # Default language settings. language_map is used to detect a source
+    # file or Extension target language, checking source filenames.
+    # language_order is used to detect the language precedence, when deciding
+    # what language to use when mixing source types. For example, if some
+    # extension has two files with ".c" extension, and one with ".cpp", it
+    # is still linked as c++.
+    language_map = {
+        ".c": "c",
+        ".cc": "c++",
+        ".cpp": "c++",
+        ".cxx": "c++",
+        ".m": "objc",
+    }
+    language_order = ["c++", "objc", "c"]
+
+    include_dirs = []
+    """
+    include dirs specific to this compiler class
+    """
+
+    library_dirs = []
+    """
+    library dirs specific to this compiler class
+    """
+
+    def __init__(self, verbose=False, dry_run=False, force=False):
+        self.dry_run = dry_run
+        self.force = force
+        self.verbose = verbose
+
+        # 'output_dir': a common output directory for object, library,
+        # shared object, and shared library files
+        self.output_dir = None
+
+        # 'macros': a list of macro definitions (or undefinitions).  A
+        # macro definition is a 2-tuple (name, value), where the value is
+        # either a string or None (no explicit value).  A macro
+        # undefinition is a 1-tuple (name,).
+        self.macros = []
+
+        # 'include_dirs': a list of directories to search for include files
+        self.include_dirs = []
+
+        # 'libraries': a list of libraries to include in any link
+        # (library names, not filenames: eg. "foo" not "libfoo.a")
+        self.libraries = []
+
+        # 'library_dirs': a list of directories to search for libraries
+        self.library_dirs = []
+
+        # 'runtime_library_dirs': a list of directories to search for
+        # shared libraries/objects at runtime
+        self.runtime_library_dirs = []
+
+        # 'objects': a list of object files (or similar, such as explicitly
+        # named library files) to include on any link
+        self.objects = []
+
+        for key in self.executables.keys():
+            self.set_executable(key, self.executables[key])
+
+    def set_executables(self, **kwargs):
+        """Define the executables (and options for them) that will be run
+        to perform the various stages of compilation.  The exact set of
+        executables that may be specified here depends on the compiler
+        class (via the 'executables' class attribute), but most will have:
+          compiler      the C/C++ compiler
+          linker_so     linker used to create shared objects and libraries
+          linker_exe    linker used to create binary executables
+          archiver      static library creator
+
+        On platforms with a command-line (Unix, DOS/Windows), each of these
+        is a string that will be split into executable name and (optional)
+        list of arguments.  (Splitting the string is done similarly to how
+        Unix shells operate: words are delimited by spaces, but quotes and
+        backslashes can override this.  See
+        'distutils.util.split_quoted()'.)
+        """
+
+        # Note that some CCompiler implementation classes will define class
+        # attributes 'cpp', 'cc', etc. with hard-coded executable names;
+        # this is appropriate when a compiler class is for exactly one
+        # compiler/OS combination (eg. MSVCCompiler).  Other compiler
+        # classes (UnixCCompiler, in particular) are driven by information
+        # discovered at run-time, since there are many different ways to do
+        # basically the same things with Unix C compilers.
+
+        for key in kwargs:
+            if key not in self.executables:
+                raise ValueError(
+                    f"unknown executable '{key}' for class {self.__class__.__name__}"
+                )
+            self.set_executable(key, kwargs[key])
+
+    def set_executable(self, key, value):
+        if isinstance(value, str):
+            setattr(self, key, split_quoted(value))
+        else:
+            setattr(self, key, value)
+
+    def _find_macro(self, name):
+        i = 0
+        for defn in self.macros:
+            if defn[0] == name:
+                return i
+            i += 1
+        return None
+
+    def _check_macro_definitions(self, definitions):
+        """Ensure that every element of 'definitions' is valid."""
+        for defn in definitions:
+            self._check_macro_definition(*defn)
+
+    def _check_macro_definition(self, defn):
+        """
+        Raise a TypeError if defn is not valid.
+
+        A valid definition is either a (name, value) 2-tuple or a (name,) tuple.
+        """
+        if not isinstance(defn, tuple) or not self._is_valid_macro(*defn):
+            raise TypeError(
+                f"invalid macro definition '{defn}': "
+                "must be tuple (string,), (string, string), or (string, None)"
+            )
+
+    @staticmethod
+    def _is_valid_macro(name, value=None):
+        """
+        A valid macro is a ``name : str`` and a ``value : str | None``.
+        """
+        return isinstance(name, str) and isinstance(value, (str, types.NoneType))
+
+    # -- Bookkeeping methods -------------------------------------------
+
+    def define_macro(self, name, value=None):
+        """Define a preprocessor macro for all compilations driven by this
+        compiler object.  The optional parameter 'value' should be a
+        string; if it is not supplied, then the macro will be defined
+        without an explicit value and the exact outcome depends on the
+        compiler used (XXX true? does ANSI say anything about this?)
+        """
+        # Delete from the list of macro definitions/undefinitions if
+        # already there (so that this one will take precedence).
+        i = self._find_macro(name)
+        if i is not None:
+            del self.macros[i]
+
+        self.macros.append((name, value))
+
+    def undefine_macro(self, name):
+        """Undefine a preprocessor macro for all compilations driven by
+        this compiler object.  If the same macro is defined by
+        'define_macro()' and undefined by 'undefine_macro()' the last call
+        takes precedence (including multiple redefinitions or
+        undefinitions).  If the macro is redefined/undefined on a
+        per-compilation basis (ie. in the call to 'compile()'), then that
+        takes precedence.
+        """
+        # Delete from the list of macro definitions/undefinitions if
+        # already there (so that this one will take precedence).
+        i = self._find_macro(name)
+        if i is not None:
+            del self.macros[i]
+
+        undefn = (name,)
+        self.macros.append(undefn)
+
+    def add_include_dir(self, dir):
+        """Add 'dir' to the list of directories that will be searched for
+        header files.  The compiler is instructed to search directories in
+        the order in which they are supplied by successive calls to
+        'add_include_dir()'.
+        """
+        self.include_dirs.append(dir)
+
+    def set_include_dirs(self, dirs):
+        """Set the list of directories that will be searched to 'dirs' (a
+        list of strings).  Overrides any preceding calls to
+        'add_include_dir()'; subsequence calls to 'add_include_dir()' add
+        to the list passed to 'set_include_dirs()'.  This does not affect
+        any list of standard include directories that the compiler may
+        search by default.
+        """
+        self.include_dirs = dirs[:]
+
+    def add_library(self, libname):
+        """Add 'libname' to the list of libraries that will be included in
+        all links driven by this compiler object.  Note that 'libname'
+        should *not* be the name of a file containing a library, but the
+        name of the library itself: the actual filename will be inferred by
+        the linker, the compiler, or the compiler class (depending on the
+        platform).
+
+        The linker will be instructed to link against libraries in the
+        order they were supplied to 'add_library()' and/or
+        'set_libraries()'.  It is perfectly valid to duplicate library
+        names; the linker will be instructed to link against libraries as
+        many times as they are mentioned.
+        """
+        self.libraries.append(libname)
+
+    def set_libraries(self, libnames):
+        """Set the list of libraries to be included in all links driven by
+        this compiler object to 'libnames' (a list of strings).  This does
+        not affect any standard system libraries that the linker may
+        include by default.
+        """
+        self.libraries = libnames[:]
+
+    def add_library_dir(self, dir):
+        """Add 'dir' to the list of directories that will be searched for
+        libraries specified to 'add_library()' and 'set_libraries()'.  The
+        linker will be instructed to search for libraries in the order they
+        are supplied to 'add_library_dir()' and/or 'set_library_dirs()'.
+        """
+        self.library_dirs.append(dir)
+
+    def set_library_dirs(self, dirs):
+        """Set the list of library search directories to 'dirs' (a list of
+        strings).  This does not affect any standard library search path
+        that the linker may search by default.
+        """
+        self.library_dirs = dirs[:]
+
+    def add_runtime_library_dir(self, dir):
+        """Add 'dir' to the list of directories that will be searched for
+        shared libraries at runtime.
+        """
+        self.runtime_library_dirs.append(dir)
+
+    def set_runtime_library_dirs(self, dirs):
+        """Set the list of directories to search for shared libraries at
+        runtime to 'dirs' (a list of strings).  This does not affect any
+        standard search path that the runtime linker may search by
+        default.
+        """
+        self.runtime_library_dirs = dirs[:]
+
+    def add_link_object(self, object):
+        """Add 'object' to the list of object files (or analogues, such as
+        explicitly named library files or the output of "resource
+        compilers") to be included in every link driven by this compiler
+        object.
+        """
+        self.objects.append(object)
+
+    def set_link_objects(self, objects):
+        """Set the list of object files (or analogues) to be included in
+        every link to 'objects'.  This does not affect any standard object
+        files that the linker may include by default (such as system
+        libraries).
+        """
+        self.objects = objects[:]
+
+    # -- Private utility methods --------------------------------------
+    # (here for the convenience of subclasses)
+
+    # Helper method to prep compiler in subclass compile() methods
+
+    def _setup_compile(self, outdir, macros, incdirs, sources, depends, extra):
+        """Process arguments and decide which source files to compile."""
+        outdir, macros, incdirs = self._fix_compile_args(outdir, macros, incdirs)
+
+        if extra is None:
+            extra = []
+
+        # Get the list of expected output (object) files
+        objects = self.object_filenames(sources, strip_dir=False, output_dir=outdir)
+        assert len(objects) == len(sources)
+
+        pp_opts = gen_preprocess_options(macros, incdirs)
+
+        build = {}
+        for i in range(len(sources)):
+            src = sources[i]
+            obj = objects[i]
+            ext = os.path.splitext(src)[1]
+            self.mkpath(os.path.dirname(obj))
+            build[obj] = (src, ext)
+
+        return macros, objects, extra, pp_opts, build
+
+    def _get_cc_args(self, pp_opts, debug, before):
+        # works for unixccompiler, cygwinccompiler
+        cc_args = pp_opts + ['-c']
+        if debug:
+            cc_args[:0] = ['-g']
+        if before:
+            cc_args[:0] = before
+        return cc_args
+
+    def _fix_compile_args(self, output_dir, macros, include_dirs):
+        """Typecheck and fix-up some of the arguments to the 'compile()'
+        method, and return fixed-up values.  Specifically: if 'output_dir'
+        is None, replaces it with 'self.output_dir'; ensures that 'macros'
+        is a list, and augments it with 'self.macros'; ensures that
+        'include_dirs' is a list, and augments it with 'self.include_dirs'.
+        Guarantees that the returned values are of the correct type,
+        i.e. for 'output_dir' either string or None, and for 'macros' and
+        'include_dirs' either list or None.
+        """
+        if output_dir is None:
+            output_dir = self.output_dir
+        elif not isinstance(output_dir, str):
+            raise TypeError("'output_dir' must be a string or None")
+
+        if macros is None:
+            macros = list(self.macros)
+        elif isinstance(macros, list):
+            macros = macros + (self.macros or [])
+        else:
+            raise TypeError("'macros' (if supplied) must be a list of tuples")
+
+        if include_dirs is None:
+            include_dirs = list(self.include_dirs)
+        elif isinstance(include_dirs, (list, tuple)):
+            include_dirs = list(include_dirs) + (self.include_dirs or [])
+        else:
+            raise TypeError("'include_dirs' (if supplied) must be a list of strings")
+
+        # add include dirs for class
+        include_dirs += self.__class__.include_dirs
+
+        return output_dir, macros, include_dirs
+
+    def _prep_compile(self, sources, output_dir, depends=None):
+        """Decide which source files must be recompiled.
+
+        Determine the list of object files corresponding to 'sources',
+        and figure out which ones really need to be recompiled.
+        Return a list of all object files and a dictionary telling
+        which source files can be skipped.
+        """
+        # Get the list of expected output (object) files
+        objects = self.object_filenames(sources, output_dir=output_dir)
+        assert len(objects) == len(sources)
+
+        # Return an empty dict for the "which source files can be skipped"
+        # return value to preserve API compatibility.
+        return objects, {}
+
+    def _fix_object_args(self, objects, output_dir):
+        """Typecheck and fix up some arguments supplied to various methods.
+        Specifically: ensure that 'objects' is a list; if output_dir is
+        None, replace with self.output_dir.  Return fixed versions of
+        'objects' and 'output_dir'.
+        """
+        if not isinstance(objects, (list, tuple)):
+            raise TypeError("'objects' must be a list or tuple of strings")
+        objects = list(objects)
+
+        if output_dir is None:
+            output_dir = self.output_dir
+        elif not isinstance(output_dir, str):
+            raise TypeError("'output_dir' must be a string or None")
+
+        return (objects, output_dir)
+
+    def _fix_lib_args(self, libraries, library_dirs, runtime_library_dirs):
+        """Typecheck and fix up some of the arguments supplied to the
+        'link_*' methods.  Specifically: ensure that all arguments are
+        lists, and augment them with their permanent versions
+        (eg. 'self.libraries' augments 'libraries').  Return a tuple with
+        fixed versions of all arguments.
+        """
+        if libraries is None:
+            libraries = list(self.libraries)
+        elif isinstance(libraries, (list, tuple)):
+            libraries = list(libraries) + (self.libraries or [])
+        else:
+            raise TypeError("'libraries' (if supplied) must be a list of strings")
+
+        if library_dirs is None:
+            library_dirs = list(self.library_dirs)
+        elif isinstance(library_dirs, (list, tuple)):
+            library_dirs = list(library_dirs) + (self.library_dirs or [])
+        else:
+            raise TypeError("'library_dirs' (if supplied) must be a list of strings")
+
+        # add library dirs for class
+        library_dirs += self.__class__.library_dirs
+
+        if runtime_library_dirs is None:
+            runtime_library_dirs = list(self.runtime_library_dirs)
+        elif isinstance(runtime_library_dirs, (list, tuple)):
+            runtime_library_dirs = list(runtime_library_dirs) + (
+                self.runtime_library_dirs or []
+            )
+        else:
+            raise TypeError(
+                "'runtime_library_dirs' (if supplied) must be a list of strings"
+            )
+
+        return (libraries, library_dirs, runtime_library_dirs)
+
+    def _need_link(self, objects, output_file):
+        """Return true if we need to relink the files listed in 'objects'
+        to recreate 'output_file'.
+        """
+        if self.force:
+            return True
+        else:
+            if self.dry_run:
+                newer = newer_group(objects, output_file, missing='newer')
+            else:
+                newer = newer_group(objects, output_file)
+            return newer
+
+    def detect_language(self, sources):
+        """Detect the language of a given file, or list of files. Uses
+        language_map, and language_order to do the job.
+        """
+        if not isinstance(sources, list):
+            sources = [sources]
+        lang = None
+        index = len(self.language_order)
+        for source in sources:
+            base, ext = os.path.splitext(source)
+            extlang = self.language_map.get(ext)
+            try:
+                extindex = self.language_order.index(extlang)
+                if extindex < index:
+                    lang = extlang
+                    index = extindex
+            except ValueError:
+                pass
+        return lang
+
+    # -- Worker methods ------------------------------------------------
+    # (must be implemented by subclasses)
+
+    def preprocess(
+        self,
+        source,
+        output_file=None,
+        macros=None,
+        include_dirs=None,
+        extra_preargs=None,
+        extra_postargs=None,
+    ):
+        """Preprocess a single C/C++ source file, named in 'source'.
+        Output will be written to file named 'output_file', or stdout if
+        'output_file' not supplied.  'macros' is a list of macro
+        definitions as for 'compile()', which will augment the macros set
+        with 'define_macro()' and 'undefine_macro()'.  'include_dirs' is a
+        list of directory names that will be added to the default list.
+
+        Raises PreprocessError on failure.
+        """
+        pass
+
+    def compile(
+        self,
+        sources,
+        output_dir=None,
+        macros=None,
+        include_dirs=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        depends=None,
+    ):
+        """Compile one or more source files.
+
+        'sources' must be a list of filenames, most likely C/C++
+        files, but in reality anything that can be handled by a
+        particular compiler and compiler class (eg. MSVCCompiler can
+        handle resource files in 'sources').  Return a list of object
+        filenames, one per source filename in 'sources'.  Depending on
+        the implementation, not all source files will necessarily be
+        compiled, but all corresponding object filenames will be
+        returned.
+
+        If 'output_dir' is given, object files will be put under it, while
+        retaining their original path component.  That is, "foo/bar.c"
+        normally compiles to "foo/bar.o" (for a Unix implementation); if
+        'output_dir' is "build", then it would compile to
+        "build/foo/bar.o".
+
+        'macros', if given, must be a list of macro definitions.  A macro
+        definition is either a (name, value) 2-tuple or a (name,) 1-tuple.
+        The former defines a macro; if the value is None, the macro is
+        defined without an explicit value.  The 1-tuple case undefines a
+        macro.  Later definitions/redefinitions/ undefinitions take
+        precedence.
+
+        'include_dirs', if given, must be a list of strings, the
+        directories to add to the default include file search path for this
+        compilation only.
+
+        'debug' is a boolean; if true, the compiler will be instructed to
+        output debug symbols in (or alongside) the object file(s).
+
+        'extra_preargs' and 'extra_postargs' are implementation- dependent.
+        On platforms that have the notion of a command-line (e.g. Unix,
+        DOS/Windows), they are most likely lists of strings: extra
+        command-line arguments to prepend/append to the compiler command
+        line.  On other platforms, consult the implementation class
+        documentation.  In any event, they are intended as an escape hatch
+        for those occasions when the abstract compiler framework doesn't
+        cut the mustard.
+
+        'depends', if given, is a list of filenames that all targets
+        depend on.  If a source file is older than any file in
+        depends, then the source file will be recompiled.  This
+        supports dependency tracking, but only at a coarse
+        granularity.
+
+        Raises CompileError on failure.
+        """
+        # A concrete compiler class can either override this method
+        # entirely or implement _compile().
+        macros, objects, extra_postargs, pp_opts, build = self._setup_compile(
+            output_dir, macros, include_dirs, sources, depends, extra_postargs
+        )
+        cc_args = self._get_cc_args(pp_opts, debug, extra_preargs)
+
+        for obj in objects:
+            try:
+                src, ext = build[obj]
+            except KeyError:
+                continue
+            self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts)
+
+        # Return *all* object filenames, not just the ones we just built.
+        return objects
+
+    def _compile(self, obj, src, ext, cc_args, extra_postargs, pp_opts):
+        """Compile 'src' to product 'obj'."""
+        # A concrete compiler class that does not override compile()
+        # should implement _compile().
+        pass
+
+    def create_static_lib(
+        self, objects, output_libname, output_dir=None, debug=False, target_lang=None
+    ):
+        """Link a bunch of stuff together to create a static library file.
+        The "bunch of stuff" consists of the list of object files supplied
+        as 'objects', the extra object files supplied to
+        'add_link_object()' and/or 'set_link_objects()', the libraries
+        supplied to 'add_library()' and/or 'set_libraries()', and the
+        libraries supplied as 'libraries' (if any).
+
+        'output_libname' should be a library name, not a filename; the
+        filename will be inferred from the library name.  'output_dir' is
+        the directory where the library file will be put.
+
+        'debug' is a boolean; if true, debugging information will be
+        included in the library (note that on most platforms, it is the
+        compile step where this matters: the 'debug' flag is included here
+        just for consistency).
+
+        'target_lang' is the target language for which the given objects
+        are being compiled. This allows specific linkage time treatment of
+        certain languages.
+
+        Raises LibError on failure.
+        """
+        pass
+
+    # values for target_desc parameter in link()
+    SHARED_OBJECT = "shared_object"
+    SHARED_LIBRARY = "shared_library"
+    EXECUTABLE = "executable"
+
+    def link(
+        self,
+        target_desc,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        """Link a bunch of stuff together to create an executable or
+        shared library file.
+
+        The "bunch of stuff" consists of the list of object files supplied
+        as 'objects'.  'output_filename' should be a filename.  If
+        'output_dir' is supplied, 'output_filename' is relative to it
+        (i.e. 'output_filename' can provide directory components if
+        needed).
+
+        'libraries' is a list of libraries to link against.  These are
+        library names, not filenames, since they're translated into
+        filenames in a platform-specific way (eg. "foo" becomes "libfoo.a"
+        on Unix and "foo.lib" on DOS/Windows).  However, they can include a
+        directory component, which means the linker will look in that
+        specific directory rather than searching all the normal locations.
+
+        'library_dirs', if supplied, should be a list of directories to
+        search for libraries that were specified as bare library names
+        (ie. no directory component).  These are on top of the system
+        default and those supplied to 'add_library_dir()' and/or
+        'set_library_dirs()'.  'runtime_library_dirs' is a list of
+        directories that will be embedded into the shared library and used
+        to search for other shared libraries that *it* depends on at
+        run-time.  (This may only be relevant on Unix.)
+
+        'export_symbols' is a list of symbols that the shared library will
+        export.  (This appears to be relevant only on Windows.)
+
+        'debug' is as for 'compile()' and 'create_static_lib()', with the
+        slight distinction that it actually matters on most platforms (as
+        opposed to 'create_static_lib()', which includes a 'debug' flag
+        mostly for form's sake).
+
+        'extra_preargs' and 'extra_postargs' are as for 'compile()' (except
+        of course that they supply command-line arguments for the
+        particular linker being used).
+
+        'target_lang' is the target language for which the given objects
+        are being compiled. This allows specific linkage time treatment of
+        certain languages.
+
+        Raises LinkError on failure.
+        """
+        raise NotImplementedError
+
+    # Old 'link_*()' methods, rewritten to use the new 'link()' method.
+
+    def link_shared_lib(
+        self,
+        objects,
+        output_libname,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        self.link(
+            CCompiler.SHARED_LIBRARY,
+            objects,
+            self.library_filename(output_libname, lib_type='shared'),
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            export_symbols,
+            debug,
+            extra_preargs,
+            extra_postargs,
+            build_temp,
+            target_lang,
+        )
+
+    def link_shared_object(
+        self,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        self.link(
+            CCompiler.SHARED_OBJECT,
+            objects,
+            output_filename,
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            export_symbols,
+            debug,
+            extra_preargs,
+            extra_postargs,
+            build_temp,
+            target_lang,
+        )
+
+    def link_executable(
+        self,
+        objects,
+        output_progname,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        target_lang=None,
+    ):
+        self.link(
+            CCompiler.EXECUTABLE,
+            objects,
+            self.executable_filename(output_progname),
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            None,
+            debug,
+            extra_preargs,
+            extra_postargs,
+            None,
+            target_lang,
+        )
+
+    # -- Miscellaneous methods -----------------------------------------
+    # These are all used by the 'gen_lib_options() function; there is
+    # no appropriate default implementation so subclasses should
+    # implement all of these.
+
+    def library_dir_option(self, dir):
+        """Return the compiler option to add 'dir' to the list of
+        directories searched for libraries.
+        """
+        raise NotImplementedError
+
+    def runtime_library_dir_option(self, dir):
+        """Return the compiler option to add 'dir' to the list of
+        directories searched for runtime libraries.
+        """
+        raise NotImplementedError
+
+    def library_option(self, lib):
+        """Return the compiler option to add 'lib' to the list of libraries
+        linked into the shared library or executable.
+        """
+        raise NotImplementedError
+
+    def has_function(  # noqa: C901
+        self,
+        funcname,
+        includes=None,
+        include_dirs=None,
+        libraries=None,
+        library_dirs=None,
+    ):
+        """Return a boolean indicating whether funcname is provided as
+        a symbol on the current platform.  The optional arguments can
+        be used to augment the compilation environment.
+
+        The libraries argument is a list of flags to be passed to the
+        linker to make additional symbol definitions available for
+        linking.
+
+        The includes and include_dirs arguments are deprecated.
+        Usually, supplying include files with function declarations
+        will cause function detection to fail even in cases where the
+        symbol is available for linking.
+
+        """
+        # this can't be included at module scope because it tries to
+        # import math which might not be available at that point - maybe
+        # the necessary logic should just be inlined?
+        import tempfile
+
+        if includes is None:
+            includes = []
+        else:
+            warnings.warn("includes is deprecated", DeprecationWarning)
+        if include_dirs is None:
+            include_dirs = []
+        else:
+            warnings.warn("include_dirs is deprecated", DeprecationWarning)
+        if libraries is None:
+            libraries = []
+        if library_dirs is None:
+            library_dirs = []
+        fd, fname = tempfile.mkstemp(".c", funcname, text=True)
+        with os.fdopen(fd, "w", encoding='utf-8') as f:
+            for incl in includes:
+                f.write(f"""#include "{incl}"\n""")
+            if not includes:
+                # Use "char func(void);" as the prototype to follow
+                # what autoconf does.  This prototype does not match
+                # any well-known function the compiler might recognize
+                # as a builtin, so this ends up as a true link test.
+                # Without a fake prototype, the test would need to
+                # know the exact argument types, and the has_function
+                # interface does not provide that level of information.
+                f.write(
+                    f"""\
+#ifdef __cplusplus
+extern "C"
+#endif
+char {funcname}(void);
+"""
+                )
+            f.write(
+                f"""\
+int main (int argc, char **argv) {{
+    {funcname}();
+    return 0;
+}}
+"""
+            )
+
+        try:
+            objects = self.compile([fname], include_dirs=include_dirs)
+        except CompileError:
+            return False
+        finally:
+            os.remove(fname)
+
+        try:
+            self.link_executable(
+                objects, "a.out", libraries=libraries, library_dirs=library_dirs
+            )
+        except (LinkError, TypeError):
+            return False
+        else:
+            os.remove(
+                self.executable_filename("a.out", output_dir=self.output_dir or '')
+            )
+        finally:
+            for fn in objects:
+                os.remove(fn)
+        return True
+
+    def find_library_file(self, dirs, lib, debug=False):
+        """Search the specified list of directories for a static or shared
+        library file 'lib' and return the full path to that file.  If
+        'debug' true, look for a debugging version (if that makes sense on
+        the current platform).  Return None if 'lib' wasn't found in any of
+        the specified directories.
+        """
+        raise NotImplementedError
+
+    # -- Filename generation methods -----------------------------------
+
+    # The default implementation of the filename generating methods are
+    # prejudiced towards the Unix/DOS/Windows view of the world:
+    #   * object files are named by replacing the source file extension
+    #     (eg. .c/.cpp -> .o/.obj)
+    #   * library files (shared or static) are named by plugging the
+    #     library name and extension into a format string, eg.
+    #     "lib%s.%s" % (lib_name, ".a") for Unix static libraries
+    #   * executables are named by appending an extension (possibly
+    #     empty) to the program name: eg. progname + ".exe" for
+    #     Windows
+    #
+    # To reduce redundant code, these methods expect to find
+    # several attributes in the current object (presumably defined
+    # as class attributes):
+    #   * src_extensions -
+    #     list of C/C++ source file extensions, eg. ['.c', '.cpp']
+    #   * obj_extension -
+    #     object file extension, eg. '.o' or '.obj'
+    #   * static_lib_extension -
+    #     extension for static library files, eg. '.a' or '.lib'
+    #   * shared_lib_extension -
+    #     extension for shared library/object files, eg. '.so', '.dll'
+    #   * static_lib_format -
+    #     format string for generating static library filenames,
+    #     eg. 'lib%s.%s' or '%s.%s'
+    #   * shared_lib_format
+    #     format string for generating shared library filenames
+    #     (probably same as static_lib_format, since the extension
+    #     is one of the intended parameters to the format string)
+    #   * exe_extension -
+    #     extension for executable files, eg. '' or '.exe'
+
+    def object_filenames(self, source_filenames, strip_dir=False, output_dir=''):
+        if output_dir is None:
+            output_dir = ''
+        return list(
+            self._make_out_path(output_dir, strip_dir, src_name)
+            for src_name in source_filenames
+        )
+
+    @property
+    def out_extensions(self):
+        return dict.fromkeys(self.src_extensions, self.obj_extension)
+
+    def _make_out_path(self, output_dir, strip_dir, src_name):
+        return self._make_out_path_exts(
+            output_dir, strip_dir, src_name, self.out_extensions
+        )
+
+    @classmethod
+    def _make_out_path_exts(cls, output_dir, strip_dir, src_name, extensions):
+        r"""
+        >>> exts = {'.c': '.o'}
+        >>> CCompiler._make_out_path_exts('.', False, '/foo/bar.c', exts).replace('\\', '/')
+        './foo/bar.o'
+        >>> CCompiler._make_out_path_exts('.', True, '/foo/bar.c', exts).replace('\\', '/')
+        './bar.o'
+        """
+        src = pathlib.PurePath(src_name)
+        # Ensure base is relative to honor output_dir (python/cpython#37775).
+        base = cls._make_relative(src)
+        try:
+            new_ext = extensions[src.suffix]
+        except LookupError:
+            raise UnknownFileError(f"unknown file type '{src.suffix}' (from '{src}')")
+        if strip_dir:
+            base = pathlib.PurePath(base.name)
+        return os.path.join(output_dir, base.with_suffix(new_ext))
+
+    @staticmethod
+    def _make_relative(base: pathlib.Path):
+        return base.relative_to(base.anchor)
+
+    def shared_object_filename(self, basename, strip_dir=False, output_dir=''):
+        assert output_dir is not None
+        if strip_dir:
+            basename = os.path.basename(basename)
+        return os.path.join(output_dir, basename + self.shared_lib_extension)
+
+    def executable_filename(self, basename, strip_dir=False, output_dir=''):
+        assert output_dir is not None
+        if strip_dir:
+            basename = os.path.basename(basename)
+        return os.path.join(output_dir, basename + (self.exe_extension or ''))
+
+    def library_filename(
+        self,
+        libname,
+        lib_type='static',
+        strip_dir=False,
+        output_dir='',  # or 'shared'
+    ):
+        assert output_dir is not None
+        expected = '"static", "shared", "dylib", "xcode_stub"'
+        if lib_type not in eval(expected):
+            raise ValueError(f"'lib_type' must be {expected}")
+        fmt = getattr(self, lib_type + "_lib_format")
+        ext = getattr(self, lib_type + "_lib_extension")
+
+        dir, base = os.path.split(libname)
+        filename = fmt % (base, ext)
+        if strip_dir:
+            dir = ''
+
+        return os.path.join(output_dir, dir, filename)
+
+    # -- Utility methods -----------------------------------------------
+
+    def announce(self, msg, level=1):
+        log.debug(msg)
+
+    def debug_print(self, msg):
+        from distutils.debug import DEBUG
+
+        if DEBUG:
+            print(msg)
+
+    def warn(self, msg):
+        sys.stderr.write(f"warning: {msg}\n")
+
+    def execute(self, func, args, msg=None, level=1):
+        execute(func, args, msg, self.dry_run)
+
+    def spawn(self, cmd, **kwargs):
+        spawn(cmd, dry_run=self.dry_run, **kwargs)
+
+    def move_file(self, src, dst):
+        return move_file(src, dst, dry_run=self.dry_run)
+
+    def mkpath(self, name, mode=0o777):
+        mkpath(name, mode, dry_run=self.dry_run)
+
+
+# Map a sys.platform/os.name ('posix', 'nt') to the default compiler
+# type for that platform. Keys are interpreted as re match
+# patterns. Order is important; platform mappings are preferred over
+# OS names.
+_default_compilers = (
+    # Platform string mappings
+    # on a cygwin built python we can use gcc like an ordinary UNIXish
+    # compiler
+    ('cygwin.*', 'unix'),
+    ('zos', 'zos'),
+    # OS name mappings
+    ('posix', 'unix'),
+    ('nt', 'msvc'),
+)
+
+
+def get_default_compiler(osname=None, platform=None):
+    """Determine the default compiler to use for the given platform.
+
+    osname should be one of the standard Python OS names (i.e. the
+    ones returned by os.name) and platform the common value
+    returned by sys.platform for the platform in question.
+
+    The default values are os.name and sys.platform in case the
+    parameters are not given.
+    """
+    if osname is None:
+        osname = os.name
+    if platform is None:
+        platform = sys.platform
+    # Mingw is a special case where sys.platform is 'win32' but we
+    # want to use the 'mingw32' compiler, so check it first
+    if is_mingw():
+        return 'mingw32'
+    for pattern, compiler in _default_compilers:
+        if (
+            re.match(pattern, platform) is not None
+            or re.match(pattern, osname) is not None
+        ):
+            return compiler
+    # Default to Unix compiler
+    return 'unix'
+
+
+# Map compiler types to (module_name, class_name) pairs -- ie. where to
+# find the code that implements an interface to this compiler.  (The module
+# is assumed to be in the 'distutils' package.)
+compiler_class = {
+    'unix': ('unixccompiler', 'UnixCCompiler', "standard UNIX-style compiler"),
+    'msvc': ('_msvccompiler', 'MSVCCompiler', "Microsoft Visual C++"),
+    'cygwin': (
+        'cygwinccompiler',
+        'CygwinCCompiler',
+        "Cygwin port of GNU C Compiler for Win32",
+    ),
+    'mingw32': (
+        'cygwinccompiler',
+        'Mingw32CCompiler',
+        "Mingw32 port of GNU C Compiler for Win32",
+    ),
+    'bcpp': ('bcppcompiler', 'BCPPCompiler', "Borland C++ Compiler"),
+    'zos': ('zosccompiler', 'zOSCCompiler', 'IBM XL C/C++ Compilers'),
+}
+
+
+def show_compilers():
+    """Print list of available compilers (used by the "--help-compiler"
+    options to "build", "build_ext", "build_clib").
+    """
+    # XXX this "knows" that the compiler option it's describing is
+    # "--compiler", which just happens to be the case for the three
+    # commands that use it.
+    from distutils.fancy_getopt import FancyGetopt
+
+    compilers = sorted(
+        ("compiler=" + compiler, None, compiler_class[compiler][2])
+        for compiler in compiler_class.keys()
+    )
+    pretty_printer = FancyGetopt(compilers)
+    pretty_printer.print_help("List of available compilers:")
+
+
+def new_compiler(plat=None, compiler=None, verbose=False, dry_run=False, force=False):
+    """Generate an instance of some CCompiler subclass for the supplied
+    platform/compiler combination.  'plat' defaults to 'os.name'
+    (eg. 'posix', 'nt'), and 'compiler' defaults to the default compiler
+    for that platform.  Currently only 'posix' and 'nt' are supported, and
+    the default compilers are "traditional Unix interface" (UnixCCompiler
+    class) and Visual C++ (MSVCCompiler class).  Note that it's perfectly
+    possible to ask for a Unix compiler object under Windows, and a
+    Microsoft compiler object under Unix -- if you supply a value for
+    'compiler', 'plat' is ignored.
+    """
+    if plat is None:
+        plat = os.name
+
+    try:
+        if compiler is None:
+            compiler = get_default_compiler(plat)
+
+        (module_name, class_name, long_description) = compiler_class[compiler]
+    except KeyError:
+        msg = f"don't know how to compile C/C++ code on platform '{plat}'"
+        if compiler is not None:
+            msg = msg + f" with '{compiler}' compiler"
+        raise DistutilsPlatformError(msg)
+
+    try:
+        module_name = "distutils." + module_name
+        __import__(module_name)
+        module = sys.modules[module_name]
+        klass = vars(module)[class_name]
+    except ImportError:
+        raise DistutilsModuleError(
+            f"can't compile C/C++ code: unable to load module '{module_name}'"
+        )
+    except KeyError:
+        raise DistutilsModuleError(
+            f"can't compile C/C++ code: unable to find class '{class_name}' "
+            f"in module '{module_name}'"
+        )
+
+    # XXX The None is necessary to preserve backwards compatibility
+    # with classes that expect verbose to be the first positional
+    # argument.
+    return klass(None, dry_run, force)
+
+
+def gen_preprocess_options(macros, include_dirs):
+    """Generate C pre-processor options (-D, -U, -I) as used by at least
+    two types of compilers: the typical Unix compiler and Visual C++.
+    'macros' is the usual thing, a list of 1- or 2-tuples, where (name,)
+    means undefine (-U) macro 'name', and (name,value) means define (-D)
+    macro 'name' to 'value'.  'include_dirs' is just a list of directory
+    names to be added to the header file search path (-I).  Returns a list
+    of command-line options suitable for either Unix compilers or Visual
+    C++.
+    """
+    # XXX it would be nice (mainly aesthetic, and so we don't generate
+    # stupid-looking command lines) to go over 'macros' and eliminate
+    # redundant definitions/undefinitions (ie. ensure that only the
+    # latest mention of a particular macro winds up on the command
+    # line).  I don't think it's essential, though, since most (all?)
+    # Unix C compilers only pay attention to the latest -D or -U
+    # mention of a macro on their command line.  Similar situation for
+    # 'include_dirs'.  I'm punting on both for now.  Anyways, weeding out
+    # redundancies like this should probably be the province of
+    # CCompiler, since the data structures used are inherited from it
+    # and therefore common to all CCompiler classes.
+    pp_opts = []
+    for macro in macros:
+        if not (isinstance(macro, tuple) and 1 <= len(macro) <= 2):
+            raise TypeError(
+                f"bad macro definition '{macro}': "
+                "each element of 'macros' list must be a 1- or 2-tuple"
+            )
+
+        if len(macro) == 1:  # undefine this macro
+            pp_opts.append(f"-U{macro[0]}")
+        elif len(macro) == 2:
+            if macro[1] is None:  # define with no explicit value
+                pp_opts.append(f"-D{macro[0]}")
+            else:
+                # XXX *don't* need to be clever about quoting the
+                # macro value here, because we're going to avoid the
+                # shell at all costs when we spawn the command!
+                pp_opts.append("-D{}={}".format(*macro))
+
+    pp_opts.extend(f"-I{dir}" for dir in include_dirs)
+    return pp_opts
+
+
+def gen_lib_options(compiler, library_dirs, runtime_library_dirs, libraries):
+    """Generate linker options for searching library directories and
+    linking with specific libraries.  'libraries' and 'library_dirs' are,
+    respectively, lists of library names (not filenames!) and search
+    directories.  Returns a list of command-line options suitable for use
+    with some compiler (depending on the two format strings passed in).
+    """
+    lib_opts = [compiler.library_dir_option(dir) for dir in library_dirs]
+
+    for dir in runtime_library_dirs:
+        lib_opts.extend(always_iterable(compiler.runtime_library_dir_option(dir)))
+
+    # XXX it's important that we *not* remove redundant library mentions!
+    # sometimes you really do have to say "-lfoo -lbar -lfoo" in order to
+    # resolve all symbols.  I just hope we never have to say "-lfoo obj.o
+    # -lbar" to get things to work -- that's certainly a possibility, but a
+    # pretty nasty way to arrange your C code.
+
+    for lib in libraries:
+        (lib_dir, lib_name) = os.path.split(lib)
+        if lib_dir:
+            lib_file = compiler.find_library_file([lib_dir], lib_name)
+            if lib_file:
+                lib_opts.append(lib_file)
+            else:
+                compiler.warn(
+                    f"no library file corresponding to '{lib}' found (skipping)"
+                )
+        else:
+            lib_opts.append(compiler.library_option(lib))
+    return lib_opts
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cmd.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cmd.py
new file mode 100644
index 00000000..9c6fa656
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cmd.py
@@ -0,0 +1,462 @@
+"""distutils.cmd
+
+Provides the Command class, the base class for the command classes
+in the distutils.command package.
+"""
+
+from __future__ import annotations
+
+import logging
+import os
+import re
+import sys
+from collections.abc import Callable
+from typing import Any, ClassVar, TypeVar, overload
+
+from . import _modified, archive_util, dir_util, file_util, util
+from ._log import log
+from .errors import DistutilsOptionError
+
+_CommandT = TypeVar("_CommandT", bound="Command")
+
+
+class Command:
+    """Abstract base class for defining command classes, the "worker bees"
+    of the Distutils.  A useful analogy for command classes is to think of
+    them as subroutines with local variables called "options".  The options
+    are "declared" in 'initialize_options()' and "defined" (given their
+    final values, aka "finalized") in 'finalize_options()', both of which
+    must be defined by every command class.  The distinction between the
+    two is necessary because option values might come from the outside
+    world (command line, config file, ...), and any options dependent on
+    other options must be computed *after* these outside influences have
+    been processed -- hence 'finalize_options()'.  The "body" of the
+    subroutine, where it does all its work based on the values of its
+    options, is the 'run()' method, which must also be implemented by every
+    command class.
+    """
+
+    # 'sub_commands' formalizes the notion of a "family" of commands,
+    # eg. "install" as the parent with sub-commands "install_lib",
+    # "install_headers", etc.  The parent of a family of commands
+    # defines 'sub_commands' as a class attribute; it's a list of
+    #    (command_name : string, predicate : unbound_method | string | None)
+    # tuples, where 'predicate' is a method of the parent command that
+    # determines whether the corresponding command is applicable in the
+    # current situation.  (Eg. we "install_headers" is only applicable if
+    # we have any C header files to install.)  If 'predicate' is None,
+    # that command is always applicable.
+    #
+    # 'sub_commands' is usually defined at the *end* of a class, because
+    # predicates can be unbound methods, so they must already have been
+    # defined.  The canonical example is the "install" command.
+    sub_commands: ClassVar[  # Any to work around variance issues
+        list[tuple[str, Callable[[Any], bool] | None]]
+    ] = []
+
+    user_options: ClassVar[
+        # Specifying both because list is invariant. Avoids mypy override assignment issues
+        list[tuple[str, str, str]] | list[tuple[str, str | None, str]]
+    ] = []
+
+    # -- Creation/initialization methods -------------------------------
+
+    def __init__(self, dist):
+        """Create and initialize a new Command object.  Most importantly,
+        invokes the 'initialize_options()' method, which is the real
+        initializer and depends on the actual command being
+        instantiated.
+        """
+        # late import because of mutual dependence between these classes
+        from distutils.dist import Distribution
+
+        if not isinstance(dist, Distribution):
+            raise TypeError("dist must be a Distribution instance")
+        if self.__class__ is Command:
+            raise RuntimeError("Command is an abstract class")
+
+        self.distribution = dist
+        self.initialize_options()
+
+        # Per-command versions of the global flags, so that the user can
+        # customize Distutils' behaviour command-by-command and let some
+        # commands fall back on the Distribution's behaviour.  None means
+        # "not defined, check self.distribution's copy", while 0 or 1 mean
+        # false and true (duh).  Note that this means figuring out the real
+        # value of each flag is a touch complicated -- hence "self._dry_run"
+        # will be handled by __getattr__, below.
+        # XXX This needs to be fixed.
+        self._dry_run = None
+
+        # verbose is largely ignored, but needs to be set for
+        # backwards compatibility (I think)?
+        self.verbose = dist.verbose
+
+        # Some commands define a 'self.force' option to ignore file
+        # timestamps, but methods defined *here* assume that
+        # 'self.force' exists for all commands.  So define it here
+        # just to be safe.
+        self.force = None
+
+        # The 'help' flag is just used for command-line parsing, so
+        # none of that complicated bureaucracy is needed.
+        self.help = False
+
+        # 'finalized' records whether or not 'finalize_options()' has been
+        # called.  'finalize_options()' itself should not pay attention to
+        # this flag: it is the business of 'ensure_finalized()', which
+        # always calls 'finalize_options()', to respect/update it.
+        self.finalized = False
+
+    # XXX A more explicit way to customize dry_run would be better.
+    def __getattr__(self, attr):
+        if attr == 'dry_run':
+            myval = getattr(self, "_" + attr)
+            if myval is None:
+                return getattr(self.distribution, attr)
+            else:
+                return myval
+        else:
+            raise AttributeError(attr)
+
+    def ensure_finalized(self):
+        if not self.finalized:
+            self.finalize_options()
+        self.finalized = True
+
+    # Subclasses must define:
+    #   initialize_options()
+    #     provide default values for all options; may be customized by
+    #     setup script, by options from config file(s), or by command-line
+    #     options
+    #   finalize_options()
+    #     decide on the final values for all options; this is called
+    #     after all possible intervention from the outside world
+    #     (command-line, option file, etc.) has been processed
+    #   run()
+    #     run the command: do whatever it is we're here to do,
+    #     controlled by the command's various option values
+
+    def initialize_options(self):
+        """Set default values for all the options that this command
+        supports.  Note that these defaults may be overridden by other
+        commands, by the setup script, by config files, or by the
+        command-line.  Thus, this is not the place to code dependencies
+        between options; generally, 'initialize_options()' implementations
+        are just a bunch of "self.foo = None" assignments.
+
+        This method must be implemented by all command classes.
+        """
+        raise RuntimeError(
+            f"abstract method -- subclass {self.__class__} must override"
+        )
+
+    def finalize_options(self):
+        """Set final values for all the options that this command supports.
+        This is always called as late as possible, ie.  after any option
+        assignments from the command-line or from other commands have been
+        done.  Thus, this is the place to code option dependencies: if
+        'foo' depends on 'bar', then it is safe to set 'foo' from 'bar' as
+        long as 'foo' still has the same value it was assigned in
+        'initialize_options()'.
+
+        This method must be implemented by all command classes.
+        """
+        raise RuntimeError(
+            f"abstract method -- subclass {self.__class__} must override"
+        )
+
+    def dump_options(self, header=None, indent=""):
+        from distutils.fancy_getopt import longopt_xlate
+
+        if header is None:
+            header = f"command options for '{self.get_command_name()}':"
+        self.announce(indent + header, level=logging.INFO)
+        indent = indent + "  "
+        for option, _, _ in self.user_options:
+            option = option.translate(longopt_xlate)
+            if option[-1] == "=":
+                option = option[:-1]
+            value = getattr(self, option)
+            self.announce(indent + f"{option} = {value}", level=logging.INFO)
+
+    def run(self):
+        """A command's raison d'etre: carry out the action it exists to
+        perform, controlled by the options initialized in
+        'initialize_options()', customized by other commands, the setup
+        script, the command-line, and config files, and finalized in
+        'finalize_options()'.  All terminal output and filesystem
+        interaction should be done by 'run()'.
+
+        This method must be implemented by all command classes.
+        """
+        raise RuntimeError(
+            f"abstract method -- subclass {self.__class__} must override"
+        )
+
+    def announce(self, msg, level=logging.DEBUG):
+        log.log(level, msg)
+
+    def debug_print(self, msg):
+        """Print 'msg' to stdout if the global DEBUG (taken from the
+        DISTUTILS_DEBUG environment variable) flag is true.
+        """
+        from distutils.debug import DEBUG
+
+        if DEBUG:
+            print(msg)
+            sys.stdout.flush()
+
+    # -- Option validation methods -------------------------------------
+    # (these are very handy in writing the 'finalize_options()' method)
+    #
+    # NB. the general philosophy here is to ensure that a particular option
+    # value meets certain type and value constraints.  If not, we try to
+    # force it into conformance (eg. if we expect a list but have a string,
+    # split the string on comma and/or whitespace).  If we can't force the
+    # option into conformance, raise DistutilsOptionError.  Thus, command
+    # classes need do nothing more than (eg.)
+    #   self.ensure_string_list('foo')
+    # and they can be guaranteed that thereafter, self.foo will be
+    # a list of strings.
+
+    def _ensure_stringlike(self, option, what, default=None):
+        val = getattr(self, option)
+        if val is None:
+            setattr(self, option, default)
+            return default
+        elif not isinstance(val, str):
+            raise DistutilsOptionError(f"'{option}' must be a {what} (got `{val}`)")
+        return val
+
+    def ensure_string(self, option, default=None):
+        """Ensure that 'option' is a string; if not defined, set it to
+        'default'.
+        """
+        self._ensure_stringlike(option, "string", default)
+
+    def ensure_string_list(self, option):
+        r"""Ensure that 'option' is a list of strings.  If 'option' is
+        currently a string, we split it either on /,\s*/ or /\s+/, so
+        "foo bar baz", "foo,bar,baz", and "foo,   bar baz" all become
+        ["foo", "bar", "baz"].
+        """
+        val = getattr(self, option)
+        if val is None:
+            return
+        elif isinstance(val, str):
+            setattr(self, option, re.split(r',\s*|\s+', val))
+        else:
+            if isinstance(val, list):
+                ok = all(isinstance(v, str) for v in val)
+            else:
+                ok = False
+            if not ok:
+                raise DistutilsOptionError(
+                    f"'{option}' must be a list of strings (got {val!r})"
+                )
+
+    def _ensure_tested_string(self, option, tester, what, error_fmt, default=None):
+        val = self._ensure_stringlike(option, what, default)
+        if val is not None and not tester(val):
+            raise DistutilsOptionError(
+                ("error in '%s' option: " + error_fmt) % (option, val)
+            )
+
+    def ensure_filename(self, option):
+        """Ensure that 'option' is the name of an existing file."""
+        self._ensure_tested_string(
+            option, os.path.isfile, "filename", "'%s' does not exist or is not a file"
+        )
+
+    def ensure_dirname(self, option):
+        self._ensure_tested_string(
+            option,
+            os.path.isdir,
+            "directory name",
+            "'%s' does not exist or is not a directory",
+        )
+
+    # -- Convenience methods for commands ------------------------------
+
+    def get_command_name(self):
+        if hasattr(self, 'command_name'):
+            return self.command_name
+        else:
+            return self.__class__.__name__
+
+    def set_undefined_options(self, src_cmd, *option_pairs):
+        """Set the values of any "undefined" options from corresponding
+        option values in some other command object.  "Undefined" here means
+        "is None", which is the convention used to indicate that an option
+        has not been changed between 'initialize_options()' and
+        'finalize_options()'.  Usually called from 'finalize_options()' for
+        options that depend on some other command rather than another
+        option of the same command.  'src_cmd' is the other command from
+        which option values will be taken (a command object will be created
+        for it if necessary); the remaining arguments are
+        '(src_option,dst_option)' tuples which mean "take the value of
+        'src_option' in the 'src_cmd' command object, and copy it to
+        'dst_option' in the current command object".
+        """
+        # Option_pairs: list of (src_option, dst_option) tuples
+        src_cmd_obj = self.distribution.get_command_obj(src_cmd)
+        src_cmd_obj.ensure_finalized()
+        for src_option, dst_option in option_pairs:
+            if getattr(self, dst_option) is None:
+                setattr(self, dst_option, getattr(src_cmd_obj, src_option))
+
+    def get_finalized_command(self, command, create=True):
+        """Wrapper around Distribution's 'get_command_obj()' method: find
+        (create if necessary and 'create' is true) the command object for
+        'command', call its 'ensure_finalized()' method, and return the
+        finalized command object.
+        """
+        cmd_obj = self.distribution.get_command_obj(command, create)
+        cmd_obj.ensure_finalized()
+        return cmd_obj
+
+    # XXX rename to 'get_reinitialized_command()'? (should do the
+    # same in dist.py, if so)
+    @overload
+    def reinitialize_command(
+        self, command: str, reinit_subcommands: bool = False
+    ) -> Command: ...
+    @overload
+    def reinitialize_command(
+        self, command: _CommandT, reinit_subcommands: bool = False
+    ) -> _CommandT: ...
+    def reinitialize_command(
+        self, command: str | Command, reinit_subcommands=False
+    ) -> Command:
+        return self.distribution.reinitialize_command(command, reinit_subcommands)
+
+    def run_command(self, command):
+        """Run some other command: uses the 'run_command()' method of
+        Distribution, which creates and finalizes the command object if
+        necessary and then invokes its 'run()' method.
+        """
+        self.distribution.run_command(command)
+
+    def get_sub_commands(self):
+        """Determine the sub-commands that are relevant in the current
+        distribution (ie., that need to be run).  This is based on the
+        'sub_commands' class attribute: each tuple in that list may include
+        a method that we call to determine if the subcommand needs to be
+        run for the current distribution.  Return a list of command names.
+        """
+        commands = []
+        for cmd_name, method in self.sub_commands:
+            if method is None or method(self):
+                commands.append(cmd_name)
+        return commands
+
+    # -- External world manipulation -----------------------------------
+
+    def warn(self, msg):
+        log.warning("warning: %s: %s\n", self.get_command_name(), msg)
+
+    def execute(self, func, args, msg=None, level=1):
+        util.execute(func, args, msg, dry_run=self.dry_run)
+
+    def mkpath(self, name, mode=0o777):
+        dir_util.mkpath(name, mode, dry_run=self.dry_run)
+
+    def copy_file(
+        self,
+        infile,
+        outfile,
+        preserve_mode=True,
+        preserve_times=True,
+        link=None,
+        level=1,
+    ):
+        """Copy a file respecting verbose, dry-run and force flags.  (The
+        former two default to whatever is in the Distribution object, and
+        the latter defaults to false for commands that don't define it.)"""
+        return file_util.copy_file(
+            infile,
+            outfile,
+            preserve_mode,
+            preserve_times,
+            not self.force,
+            link,
+            dry_run=self.dry_run,
+        )
+
+    def copy_tree(
+        self,
+        infile,
+        outfile,
+        preserve_mode=True,
+        preserve_times=True,
+        preserve_symlinks=False,
+        level=1,
+    ):
+        """Copy an entire directory tree respecting verbose, dry-run,
+        and force flags.
+        """
+        return dir_util.copy_tree(
+            infile,
+            outfile,
+            preserve_mode,
+            preserve_times,
+            preserve_symlinks,
+            not self.force,
+            dry_run=self.dry_run,
+        )
+
+    def move_file(self, src, dst, level=1):
+        """Move a file respecting dry-run flag."""
+        return file_util.move_file(src, dst, dry_run=self.dry_run)
+
+    def spawn(self, cmd, search_path=True, level=1):
+        """Spawn an external command respecting dry-run flag."""
+        from distutils.spawn import spawn
+
+        spawn(cmd, search_path, dry_run=self.dry_run)
+
+    def make_archive(
+        self, base_name, format, root_dir=None, base_dir=None, owner=None, group=None
+    ):
+        return archive_util.make_archive(
+            base_name,
+            format,
+            root_dir,
+            base_dir,
+            dry_run=self.dry_run,
+            owner=owner,
+            group=group,
+        )
+
+    def make_file(
+        self, infiles, outfile, func, args, exec_msg=None, skip_msg=None, level=1
+    ):
+        """Special case of 'execute()' for operations that process one or
+        more input files and generate one output file.  Works just like
+        'execute()', except the operation is skipped and a different
+        message printed if 'outfile' already exists and is newer than all
+        files listed in 'infiles'.  If the command defined 'self.force',
+        and it is true, then the command is unconditionally run -- does no
+        timestamp checks.
+        """
+        if skip_msg is None:
+            skip_msg = f"skipping {outfile} (inputs unchanged)"
+
+        # Allow 'infiles' to be a single string
+        if isinstance(infiles, str):
+            infiles = (infiles,)
+        elif not isinstance(infiles, (list, tuple)):
+            raise TypeError("'infiles' must be a string, or a list or tuple of strings")
+
+        if exec_msg is None:
+            exec_msg = "generating {} from {}".format(outfile, ', '.join(infiles))
+
+        # If 'outfile' must be regenerated (either because it doesn't
+        # exist, is out-of-date, or the 'force' flag is true) then
+        # perform the action that presumably regenerates it
+        if self.force or _modified.newer_group(infiles, outfile):
+            self.execute(func, args, exec_msg, level)
+        # Otherwise, print the "skip" message
+        else:
+            log.debug(skip_msg)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__init__.py
new file mode 100644
index 00000000..0f8a1692
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__init__.py
@@ -0,0 +1,23 @@
+"""distutils.command
+
+Package containing implementation of all the standard Distutils
+commands."""
+
+__all__ = [
+    'build',
+    'build_py',
+    'build_ext',
+    'build_clib',
+    'build_scripts',
+    'clean',
+    'install',
+    'install_lib',
+    'install_headers',
+    'install_scripts',
+    'install_data',
+    'sdist',
+    'bdist',
+    'bdist_dumb',
+    'bdist_rpm',
+    'check',
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3d112978
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/_framework_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/_framework_compat.cpython-312.pyc
new file mode 100644
index 00000000..0af763e8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/_framework_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist.cpython-312.pyc
new file mode 100644
index 00000000..96eddc87
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_dumb.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_dumb.cpython-312.pyc
new file mode 100644
index 00000000..2e04a8e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_dumb.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_rpm.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_rpm.cpython-312.pyc
new file mode 100644
index 00000000..b7a135f1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/bdist_rpm.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build.cpython-312.pyc
new file mode 100644
index 00000000..b51069c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_clib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_clib.cpython-312.pyc
new file mode 100644
index 00000000..84b77414
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_clib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_ext.cpython-312.pyc
new file mode 100644
index 00000000..0e298db8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_py.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_py.cpython-312.pyc
new file mode 100644
index 00000000..421c1d13
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_py.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_scripts.cpython-312.pyc
new file mode 100644
index 00000000..2304ab62
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/build_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/check.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/check.cpython-312.pyc
new file mode 100644
index 00000000..15ecc645
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/check.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/clean.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/clean.cpython-312.pyc
new file mode 100644
index 00000000..44ba47fe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/clean.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..fbe4e280
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install.cpython-312.pyc
new file mode 100644
index 00000000..8f9e1f32
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_data.cpython-312.pyc
new file mode 100644
index 00000000..2ed2d438
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_egg_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_egg_info.cpython-312.pyc
new file mode 100644
index 00000000..690644a9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_egg_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_headers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_headers.cpython-312.pyc
new file mode 100644
index 00000000..605748aa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_headers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_lib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_lib.cpython-312.pyc
new file mode 100644
index 00000000..4f11289c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_lib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_scripts.cpython-312.pyc
new file mode 100644
index 00000000..2660bbd4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/install_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/sdist.cpython-312.pyc
new file mode 100644
index 00000000..8c935e40
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/__pycache__/sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/_framework_compat.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/_framework_compat.py
new file mode 100644
index 00000000..00d34bc7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/_framework_compat.py
@@ -0,0 +1,54 @@
+"""
+Backward compatibility for homebrew builds on macOS.
+"""
+
+import functools
+import os
+import subprocess
+import sys
+import sysconfig
+
+
+@functools.lru_cache
+def enabled():
+    """
+    Only enabled for Python 3.9 framework homebrew builds
+    except ensurepip and venv.
+    """
+    PY39 = (3, 9) < sys.version_info < (3, 10)
+    framework = sys.platform == 'darwin' and sys._framework
+    homebrew = "Cellar" in sysconfig.get_config_var('projectbase')
+    venv = sys.prefix != sys.base_prefix
+    ensurepip = os.environ.get("ENSUREPIP_OPTIONS")
+    return PY39 and framework and homebrew and not venv and not ensurepip
+
+
+schemes = dict(
+    osx_framework_library=dict(
+        stdlib='{installed_base}/{platlibdir}/python{py_version_short}',
+        platstdlib='{platbase}/{platlibdir}/python{py_version_short}',
+        purelib='{homebrew_prefix}/lib/python{py_version_short}/site-packages',
+        platlib='{homebrew_prefix}/{platlibdir}/python{py_version_short}/site-packages',
+        include='{installed_base}/include/python{py_version_short}{abiflags}',
+        platinclude='{installed_platbase}/include/python{py_version_short}{abiflags}',
+        scripts='{homebrew_prefix}/bin',
+        data='{homebrew_prefix}',
+    )
+)
+
+
+@functools.lru_cache
+def vars():
+    if not enabled():
+        return {}
+    homebrew_prefix = subprocess.check_output(['brew', '--prefix'], text=True).strip()
+    return locals()
+
+
+def scheme(name):
+    """
+    Override the selected scheme for posix_prefix.
+    """
+    if not enabled() or not name.endswith('_prefix'):
+        return name
+    return 'osx_framework_library'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist.py
new file mode 100644
index 00000000..1ec3c35f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist.py
@@ -0,0 +1,156 @@
+"""distutils.command.bdist
+
+Implements the Distutils 'bdist' command (create a built [binary]
+distribution)."""
+
+import os
+import warnings
+from typing import ClassVar
+
+from ..core import Command
+from ..errors import DistutilsOptionError, DistutilsPlatformError
+from ..util import get_platform
+
+
+def show_formats():
+    """Print list of available formats (arguments to "--format" option)."""
+    from ..fancy_getopt import FancyGetopt
+
+    formats = [
+        ("formats=" + format, None, bdist.format_commands[format][1])
+        for format in bdist.format_commands
+    ]
+    pretty_printer = FancyGetopt(formats)
+    pretty_printer.print_help("List of available distribution formats:")
+
+
+class ListCompat(dict[str, tuple[str, str]]):
+    # adapter to allow for Setuptools compatibility in format_commands
+    def append(self, item):
+        warnings.warn(
+            """format_commands is now a dict. append is deprecated.""",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+
+class bdist(Command):
+    description = "create a built (binary) distribution"
+
+    user_options = [
+        ('bdist-base=', 'b', "temporary directory for creating built distributions"),
+        (
+            'plat-name=',
+            'p',
+            "platform name to embed in generated filenames "
+            f"[default: {get_platform()}]",
+        ),
+        ('formats=', None, "formats for distribution (comma-separated list)"),
+        (
+            'dist-dir=',
+            'd',
+            "directory to put final built distributions in [default: dist]",
+        ),
+        ('skip-build', None, "skip rebuilding everything (for testing/debugging)"),
+        (
+            'owner=',
+            'u',
+            "Owner name used when creating a tar file [default: current user]",
+        ),
+        (
+            'group=',
+            'g',
+            "Group name used when creating a tar file [default: current group]",
+        ),
+    ]
+
+    boolean_options = ['skip-build']
+
+    help_options = [
+        ('help-formats', None, "lists available distribution formats", show_formats),
+    ]
+
+    # The following commands do not take a format option from bdist
+    no_format_option: ClassVar[tuple[str, ...]] = ('bdist_rpm',)
+
+    # This won't do in reality: will need to distinguish RPM-ish Linux,
+    # Debian-ish Linux, Solaris, FreeBSD, ..., Windows, Mac OS.
+    default_format = {'posix': 'gztar', 'nt': 'zip'}
+
+    # Define commands in preferred order for the --help-formats option
+    format_commands = ListCompat({
+        'rpm': ('bdist_rpm', "RPM distribution"),
+        'gztar': ('bdist_dumb', "gzip'ed tar file"),
+        'bztar': ('bdist_dumb', "bzip2'ed tar file"),
+        'xztar': ('bdist_dumb', "xz'ed tar file"),
+        'ztar': ('bdist_dumb', "compressed tar file"),
+        'tar': ('bdist_dumb', "tar file"),
+        'zip': ('bdist_dumb', "ZIP file"),
+    })
+
+    # for compatibility until consumers only reference format_commands
+    format_command = format_commands
+
+    def initialize_options(self):
+        self.bdist_base = None
+        self.plat_name = None
+        self.formats = None
+        self.dist_dir = None
+        self.skip_build = False
+        self.group = None
+        self.owner = None
+
+    def finalize_options(self):
+        # have to finalize 'plat_name' before 'bdist_base'
+        if self.plat_name is None:
+            if self.skip_build:
+                self.plat_name = get_platform()
+            else:
+                self.plat_name = self.get_finalized_command('build').plat_name
+
+        # 'bdist_base' -- parent of per-built-distribution-format
+        # temporary directories (eg. we'll probably have
+        # "build/bdist.<plat>/dumb", "build/bdist.<plat>/rpm", etc.)
+        if self.bdist_base is None:
+            build_base = self.get_finalized_command('build').build_base
+            self.bdist_base = os.path.join(build_base, 'bdist.' + self.plat_name)
+
+        self.ensure_string_list('formats')
+        if self.formats is None:
+            try:
+                self.formats = [self.default_format[os.name]]
+            except KeyError:
+                raise DistutilsPlatformError(
+                    "don't know how to create built distributions "
+                    f"on platform {os.name}"
+                )
+
+        if self.dist_dir is None:
+            self.dist_dir = "dist"
+
+    def run(self):
+        # Figure out which sub-commands we need to run.
+        commands = []
+        for format in self.formats:
+            try:
+                commands.append(self.format_commands[format][0])
+            except KeyError:
+                raise DistutilsOptionError(f"invalid format '{format}'")
+
+        # Reinitialize and run each command.
+        for i in range(len(self.formats)):
+            cmd_name = commands[i]
+            sub_cmd = self.reinitialize_command(cmd_name)
+            if cmd_name not in self.no_format_option:
+                sub_cmd.format = self.formats[i]
+
+            # passing the owner and group names for tar archiving
+            if cmd_name == 'bdist_dumb':
+                sub_cmd.owner = self.owner
+                sub_cmd.group = self.group
+
+            # If we're going to need to run this command again, tell it to
+            # keep its temporary files around so subsequent runs go faster.
+            if cmd_name in commands[i + 1 :]:
+                sub_cmd.keep_temp = True
+            self.run_command(cmd_name)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_dumb.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_dumb.py
new file mode 100644
index 00000000..67b0c8cc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_dumb.py
@@ -0,0 +1,140 @@
+"""distutils.command.bdist_dumb
+
+Implements the Distutils 'bdist_dumb' command (create a "dumb" built
+distribution -- i.e., just an archive to be unpacked under $prefix or
+$exec_prefix)."""
+
+import os
+from distutils._log import log
+
+from ..core import Command
+from ..dir_util import ensure_relative, remove_tree
+from ..errors import DistutilsPlatformError
+from ..sysconfig import get_python_version
+from ..util import get_platform
+
+
+class bdist_dumb(Command):
+    description = "create a \"dumb\" built distribution"
+
+    user_options = [
+        ('bdist-dir=', 'd', "temporary directory for creating the distribution"),
+        (
+            'plat-name=',
+            'p',
+            "platform name to embed in generated filenames "
+            f"[default: {get_platform()}]",
+        ),
+        (
+            'format=',
+            'f',
+            "archive format to create (tar, gztar, bztar, xztar, ztar, zip)",
+        ),
+        (
+            'keep-temp',
+            'k',
+            "keep the pseudo-installation tree around after creating the distribution archive",
+        ),
+        ('dist-dir=', 'd', "directory to put final built distributions in"),
+        ('skip-build', None, "skip rebuilding everything (for testing/debugging)"),
+        (
+            'relative',
+            None,
+            "build the archive using relative paths [default: false]",
+        ),
+        (
+            'owner=',
+            'u',
+            "Owner name used when creating a tar file [default: current user]",
+        ),
+        (
+            'group=',
+            'g',
+            "Group name used when creating a tar file [default: current group]",
+        ),
+    ]
+
+    boolean_options = ['keep-temp', 'skip-build', 'relative']
+
+    default_format = {'posix': 'gztar', 'nt': 'zip'}
+
+    def initialize_options(self):
+        self.bdist_dir = None
+        self.plat_name = None
+        self.format = None
+        self.keep_temp = False
+        self.dist_dir = None
+        self.skip_build = None
+        self.relative = False
+        self.owner = None
+        self.group = None
+
+    def finalize_options(self):
+        if self.bdist_dir is None:
+            bdist_base = self.get_finalized_command('bdist').bdist_base
+            self.bdist_dir = os.path.join(bdist_base, 'dumb')
+
+        if self.format is None:
+            try:
+                self.format = self.default_format[os.name]
+            except KeyError:
+                raise DistutilsPlatformError(
+                    "don't know how to create dumb built distributions "
+                    f"on platform {os.name}"
+                )
+
+        self.set_undefined_options(
+            'bdist',
+            ('dist_dir', 'dist_dir'),
+            ('plat_name', 'plat_name'),
+            ('skip_build', 'skip_build'),
+        )
+
+    def run(self):
+        if not self.skip_build:
+            self.run_command('build')
+
+        install = self.reinitialize_command('install', reinit_subcommands=True)
+        install.root = self.bdist_dir
+        install.skip_build = self.skip_build
+        install.warn_dir = False
+
+        log.info("installing to %s", self.bdist_dir)
+        self.run_command('install')
+
+        # And make an archive relative to the root of the
+        # pseudo-installation tree.
+        archive_basename = f"{self.distribution.get_fullname()}.{self.plat_name}"
+
+        pseudoinstall_root = os.path.join(self.dist_dir, archive_basename)
+        if not self.relative:
+            archive_root = self.bdist_dir
+        else:
+            if self.distribution.has_ext_modules() and (
+                install.install_base != install.install_platbase
+            ):
+                raise DistutilsPlatformError(
+                    "can't make a dumb built distribution where "
+                    f"base and platbase are different ({install.install_base!r}, {install.install_platbase!r})"
+                )
+            else:
+                archive_root = os.path.join(
+                    self.bdist_dir, ensure_relative(install.install_base)
+                )
+
+        # Make the archive
+        filename = self.make_archive(
+            pseudoinstall_root,
+            self.format,
+            root_dir=archive_root,
+            owner=self.owner,
+            group=self.group,
+        )
+        if self.distribution.has_ext_modules():
+            pyversion = get_python_version()
+        else:
+            pyversion = 'any'
+        self.distribution.dist_files.append(('bdist_dumb', pyversion, filename))
+
+        if not self.keep_temp:
+            remove_tree(self.bdist_dir, dry_run=self.dry_run)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_rpm.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_rpm.py
new file mode 100644
index 00000000..d443eb09
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/bdist_rpm.py
@@ -0,0 +1,597 @@
+"""distutils.command.bdist_rpm
+
+Implements the Distutils 'bdist_rpm' command (create RPM source and binary
+distributions)."""
+
+import os
+import subprocess
+import sys
+from distutils._log import log
+
+from ..core import Command
+from ..debug import DEBUG
+from ..errors import (
+    DistutilsExecError,
+    DistutilsFileError,
+    DistutilsOptionError,
+    DistutilsPlatformError,
+)
+from ..file_util import write_file
+from ..sysconfig import get_python_version
+
+
+class bdist_rpm(Command):
+    description = "create an RPM distribution"
+
+    user_options = [
+        ('bdist-base=', None, "base directory for creating built distributions"),
+        (
+            'rpm-base=',
+            None,
+            "base directory for creating RPMs (defaults to \"rpm\" under "
+            "--bdist-base; must be specified for RPM 2)",
+        ),
+        (
+            'dist-dir=',
+            'd',
+            "directory to put final RPM files in (and .spec files if --spec-only)",
+        ),
+        (
+            'python=',
+            None,
+            "path to Python interpreter to hard-code in the .spec file "
+            "[default: \"python\"]",
+        ),
+        (
+            'fix-python',
+            None,
+            "hard-code the exact path to the current Python interpreter in "
+            "the .spec file",
+        ),
+        ('spec-only', None, "only regenerate spec file"),
+        ('source-only', None, "only generate source RPM"),
+        ('binary-only', None, "only generate binary RPM"),
+        ('use-bzip2', None, "use bzip2 instead of gzip to create source distribution"),
+        # More meta-data: too RPM-specific to put in the setup script,
+        # but needs to go in the .spec file -- so we make these options
+        # to "bdist_rpm".  The idea is that packagers would put this
+        # info in setup.cfg, although they are of course free to
+        # supply it on the command line.
+        (
+            'distribution-name=',
+            None,
+            "name of the (Linux) distribution to which this "
+            "RPM applies (*not* the name of the module distribution!)",
+        ),
+        ('group=', None, "package classification [default: \"Development/Libraries\"]"),
+        ('release=', None, "RPM release number"),
+        ('serial=', None, "RPM serial number"),
+        (
+            'vendor=',
+            None,
+            "RPM \"vendor\" (eg. \"Joe Blow <joe@example.com>\") "
+            "[default: maintainer or author from setup script]",
+        ),
+        (
+            'packager=',
+            None,
+            "RPM packager (eg. \"Jane Doe <jane@example.net>\") [default: vendor]",
+        ),
+        ('doc-files=', None, "list of documentation files (space or comma-separated)"),
+        ('changelog=', None, "RPM changelog"),
+        ('icon=', None, "name of icon file"),
+        ('provides=', None, "capabilities provided by this package"),
+        ('requires=', None, "capabilities required by this package"),
+        ('conflicts=', None, "capabilities which conflict with this package"),
+        ('build-requires=', None, "capabilities required to build this package"),
+        ('obsoletes=', None, "capabilities made obsolete by this package"),
+        ('no-autoreq', None, "do not automatically calculate dependencies"),
+        # Actions to take when building RPM
+        ('keep-temp', 'k', "don't clean up RPM build directory"),
+        ('no-keep-temp', None, "clean up RPM build directory [default]"),
+        (
+            'use-rpm-opt-flags',
+            None,
+            "compile with RPM_OPT_FLAGS when building from source RPM",
+        ),
+        ('no-rpm-opt-flags', None, "do not pass any RPM CFLAGS to compiler"),
+        ('rpm3-mode', None, "RPM 3 compatibility mode (default)"),
+        ('rpm2-mode', None, "RPM 2 compatibility mode"),
+        # Add the hooks necessary for specifying custom scripts
+        ('prep-script=', None, "Specify a script for the PREP phase of RPM building"),
+        ('build-script=', None, "Specify a script for the BUILD phase of RPM building"),
+        (
+            'pre-install=',
+            None,
+            "Specify a script for the pre-INSTALL phase of RPM building",
+        ),
+        (
+            'install-script=',
+            None,
+            "Specify a script for the INSTALL phase of RPM building",
+        ),
+        (
+            'post-install=',
+            None,
+            "Specify a script for the post-INSTALL phase of RPM building",
+        ),
+        (
+            'pre-uninstall=',
+            None,
+            "Specify a script for the pre-UNINSTALL phase of RPM building",
+        ),
+        (
+            'post-uninstall=',
+            None,
+            "Specify a script for the post-UNINSTALL phase of RPM building",
+        ),
+        ('clean-script=', None, "Specify a script for the CLEAN phase of RPM building"),
+        (
+            'verify-script=',
+            None,
+            "Specify a script for the VERIFY phase of the RPM build",
+        ),
+        # Allow a packager to explicitly force an architecture
+        ('force-arch=', None, "Force an architecture onto the RPM build process"),
+        ('quiet', 'q', "Run the INSTALL phase of RPM building in quiet mode"),
+    ]
+
+    boolean_options = [
+        'keep-temp',
+        'use-rpm-opt-flags',
+        'rpm3-mode',
+        'no-autoreq',
+        'quiet',
+    ]
+
+    negative_opt = {
+        'no-keep-temp': 'keep-temp',
+        'no-rpm-opt-flags': 'use-rpm-opt-flags',
+        'rpm2-mode': 'rpm3-mode',
+    }
+
+    def initialize_options(self):
+        self.bdist_base = None
+        self.rpm_base = None
+        self.dist_dir = None
+        self.python = None
+        self.fix_python = None
+        self.spec_only = None
+        self.binary_only = None
+        self.source_only = None
+        self.use_bzip2 = None
+
+        self.distribution_name = None
+        self.group = None
+        self.release = None
+        self.serial = None
+        self.vendor = None
+        self.packager = None
+        self.doc_files = None
+        self.changelog = None
+        self.icon = None
+
+        self.prep_script = None
+        self.build_script = None
+        self.install_script = None
+        self.clean_script = None
+        self.verify_script = None
+        self.pre_install = None
+        self.post_install = None
+        self.pre_uninstall = None
+        self.post_uninstall = None
+        self.prep = None
+        self.provides = None
+        self.requires = None
+        self.conflicts = None
+        self.build_requires = None
+        self.obsoletes = None
+
+        self.keep_temp = False
+        self.use_rpm_opt_flags = True
+        self.rpm3_mode = True
+        self.no_autoreq = False
+
+        self.force_arch = None
+        self.quiet = False
+
+    def finalize_options(self):
+        self.set_undefined_options('bdist', ('bdist_base', 'bdist_base'))
+        if self.rpm_base is None:
+            if not self.rpm3_mode:
+                raise DistutilsOptionError("you must specify --rpm-base in RPM 2 mode")
+            self.rpm_base = os.path.join(self.bdist_base, "rpm")
+
+        if self.python is None:
+            if self.fix_python:
+                self.python = sys.executable
+            else:
+                self.python = "python3"
+        elif self.fix_python:
+            raise DistutilsOptionError(
+                "--python and --fix-python are mutually exclusive options"
+            )
+
+        if os.name != 'posix':
+            raise DistutilsPlatformError(
+                f"don't know how to create RPM distributions on platform {os.name}"
+            )
+        if self.binary_only and self.source_only:
+            raise DistutilsOptionError(
+                "cannot supply both '--source-only' and '--binary-only'"
+            )
+
+        # don't pass CFLAGS to pure python distributions
+        if not self.distribution.has_ext_modules():
+            self.use_rpm_opt_flags = False
+
+        self.set_undefined_options('bdist', ('dist_dir', 'dist_dir'))
+        self.finalize_package_data()
+
+    def finalize_package_data(self):
+        self.ensure_string('group', "Development/Libraries")
+        self.ensure_string(
+            'vendor',
+            f"{self.distribution.get_contact()} <{self.distribution.get_contact_email()}>",
+        )
+        self.ensure_string('packager')
+        self.ensure_string_list('doc_files')
+        if isinstance(self.doc_files, list):
+            for readme in ('README', 'README.txt'):
+                if os.path.exists(readme) and readme not in self.doc_files:
+                    self.doc_files.append(readme)
+
+        self.ensure_string('release', "1")
+        self.ensure_string('serial')  # should it be an int?
+
+        self.ensure_string('distribution_name')
+
+        self.ensure_string('changelog')
+        # Format changelog correctly
+        self.changelog = self._format_changelog(self.changelog)
+
+        self.ensure_filename('icon')
+
+        self.ensure_filename('prep_script')
+        self.ensure_filename('build_script')
+        self.ensure_filename('install_script')
+        self.ensure_filename('clean_script')
+        self.ensure_filename('verify_script')
+        self.ensure_filename('pre_install')
+        self.ensure_filename('post_install')
+        self.ensure_filename('pre_uninstall')
+        self.ensure_filename('post_uninstall')
+
+        # XXX don't forget we punted on summaries and descriptions -- they
+        # should be handled here eventually!
+
+        # Now *this* is some meta-data that belongs in the setup script...
+        self.ensure_string_list('provides')
+        self.ensure_string_list('requires')
+        self.ensure_string_list('conflicts')
+        self.ensure_string_list('build_requires')
+        self.ensure_string_list('obsoletes')
+
+        self.ensure_string('force_arch')
+
+    def run(self):  # noqa: C901
+        if DEBUG:
+            print("before _get_package_data():")
+            print("vendor =", self.vendor)
+            print("packager =", self.packager)
+            print("doc_files =", self.doc_files)
+            print("changelog =", self.changelog)
+
+        # make directories
+        if self.spec_only:
+            spec_dir = self.dist_dir
+            self.mkpath(spec_dir)
+        else:
+            rpm_dir = {}
+            for d in ('SOURCES', 'SPECS', 'BUILD', 'RPMS', 'SRPMS'):
+                rpm_dir[d] = os.path.join(self.rpm_base, d)
+                self.mkpath(rpm_dir[d])
+            spec_dir = rpm_dir['SPECS']
+
+        # Spec file goes into 'dist_dir' if '--spec-only specified',
+        # build/rpm.<plat> otherwise.
+        spec_path = os.path.join(spec_dir, f"{self.distribution.get_name()}.spec")
+        self.execute(
+            write_file, (spec_path, self._make_spec_file()), f"writing '{spec_path}'"
+        )
+
+        if self.spec_only:  # stop if requested
+            return
+
+        # Make a source distribution and copy to SOURCES directory with
+        # optional icon.
+        saved_dist_files = self.distribution.dist_files[:]
+        sdist = self.reinitialize_command('sdist')
+        if self.use_bzip2:
+            sdist.formats = ['bztar']
+        else:
+            sdist.formats = ['gztar']
+        self.run_command('sdist')
+        self.distribution.dist_files = saved_dist_files
+
+        source = sdist.get_archive_files()[0]
+        source_dir = rpm_dir['SOURCES']
+        self.copy_file(source, source_dir)
+
+        if self.icon:
+            if os.path.exists(self.icon):
+                self.copy_file(self.icon, source_dir)
+            else:
+                raise DistutilsFileError(f"icon file '{self.icon}' does not exist")
+
+        # build package
+        log.info("building RPMs")
+        rpm_cmd = ['rpmbuild']
+
+        if self.source_only:  # what kind of RPMs?
+            rpm_cmd.append('-bs')
+        elif self.binary_only:
+            rpm_cmd.append('-bb')
+        else:
+            rpm_cmd.append('-ba')
+        rpm_cmd.extend(['--define', f'__python {self.python}'])
+        if self.rpm3_mode:
+            rpm_cmd.extend(['--define', f'_topdir {os.path.abspath(self.rpm_base)}'])
+        if not self.keep_temp:
+            rpm_cmd.append('--clean')
+
+        if self.quiet:
+            rpm_cmd.append('--quiet')
+
+        rpm_cmd.append(spec_path)
+        # Determine the binary rpm names that should be built out of this spec
+        # file
+        # Note that some of these may not be really built (if the file
+        # list is empty)
+        nvr_string = "%{name}-%{version}-%{release}"
+        src_rpm = nvr_string + ".src.rpm"
+        non_src_rpm = "%{arch}/" + nvr_string + ".%{arch}.rpm"
+        q_cmd = rf"rpm -q --qf '{src_rpm} {non_src_rpm}\n' --specfile '{spec_path}'"
+
+        out = os.popen(q_cmd)
+        try:
+            binary_rpms = []
+            source_rpm = None
+            while True:
+                line = out.readline()
+                if not line:
+                    break
+                ell = line.strip().split()
+                assert len(ell) == 2
+                binary_rpms.append(ell[1])
+                # The source rpm is named after the first entry in the spec file
+                if source_rpm is None:
+                    source_rpm = ell[0]
+
+            status = out.close()
+            if status:
+                raise DistutilsExecError(f"Failed to execute: {q_cmd!r}")
+
+        finally:
+            out.close()
+
+        self.spawn(rpm_cmd)
+
+        if not self.dry_run:
+            if self.distribution.has_ext_modules():
+                pyversion = get_python_version()
+            else:
+                pyversion = 'any'
+
+            if not self.binary_only:
+                srpm = os.path.join(rpm_dir['SRPMS'], source_rpm)
+                assert os.path.exists(srpm)
+                self.move_file(srpm, self.dist_dir)
+                filename = os.path.join(self.dist_dir, source_rpm)
+                self.distribution.dist_files.append(('bdist_rpm', pyversion, filename))
+
+            if not self.source_only:
+                for rpm in binary_rpms:
+                    rpm = os.path.join(rpm_dir['RPMS'], rpm)
+                    if os.path.exists(rpm):
+                        self.move_file(rpm, self.dist_dir)
+                        filename = os.path.join(self.dist_dir, os.path.basename(rpm))
+                        self.distribution.dist_files.append((
+                            'bdist_rpm',
+                            pyversion,
+                            filename,
+                        ))
+
+    def _dist_path(self, path):
+        return os.path.join(self.dist_dir, os.path.basename(path))
+
+    def _make_spec_file(self):  # noqa: C901
+        """Generate the text of an RPM spec file and return it as a
+        list of strings (one per line).
+        """
+        # definitions and headers
+        spec_file = [
+            '%define name ' + self.distribution.get_name(),
+            '%define version ' + self.distribution.get_version().replace('-', '_'),
+            '%define unmangled_version ' + self.distribution.get_version(),
+            '%define release ' + self.release.replace('-', '_'),
+            '',
+            'Summary: ' + (self.distribution.get_description() or "UNKNOWN"),
+        ]
+
+        # Workaround for #14443 which affects some RPM based systems such as
+        # RHEL6 (and probably derivatives)
+        vendor_hook = subprocess.getoutput('rpm --eval %{__os_install_post}')
+        # Generate a potential replacement value for __os_install_post (whilst
+        # normalizing the whitespace to simplify the test for whether the
+        # invocation of brp-python-bytecompile passes in __python):
+        vendor_hook = '\n'.join([
+            f'  {line.strip()} \\' for line in vendor_hook.splitlines()
+        ])
+        problem = "brp-python-bytecompile \\\n"
+        fixed = "brp-python-bytecompile %{__python} \\\n"
+        fixed_hook = vendor_hook.replace(problem, fixed)
+        if fixed_hook != vendor_hook:
+            spec_file.append('# Workaround for https://bugs.python.org/issue14443')
+            spec_file.append('%define __os_install_post ' + fixed_hook + '\n')
+
+        # put locale summaries into spec file
+        # XXX not supported for now (hard to put a dictionary
+        # in a config file -- arg!)
+        # for locale in self.summaries.keys():
+        #    spec_file.append('Summary(%s): %s' % (locale,
+        #                                          self.summaries[locale]))
+
+        spec_file.extend([
+            'Name: %{name}',
+            'Version: %{version}',
+            'Release: %{release}',
+        ])
+
+        # XXX yuck! this filename is available from the "sdist" command,
+        # but only after it has run: and we create the spec file before
+        # running "sdist", in case of --spec-only.
+        if self.use_bzip2:
+            spec_file.append('Source0: %{name}-%{unmangled_version}.tar.bz2')
+        else:
+            spec_file.append('Source0: %{name}-%{unmangled_version}.tar.gz')
+
+        spec_file.extend([
+            'License: ' + (self.distribution.get_license() or "UNKNOWN"),
+            'Group: ' + self.group,
+            'BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-buildroot',
+            'Prefix: %{_prefix}',
+        ])
+
+        if not self.force_arch:
+            # noarch if no extension modules
+            if not self.distribution.has_ext_modules():
+                spec_file.append('BuildArch: noarch')
+        else:
+            spec_file.append(f'BuildArch: {self.force_arch}')
+
+        for field in (
+            'Vendor',
+            'Packager',
+            'Provides',
+            'Requires',
+            'Conflicts',
+            'Obsoletes',
+        ):
+            val = getattr(self, field.lower())
+            if isinstance(val, list):
+                spec_file.append('{}: {}'.format(field, ' '.join(val)))
+            elif val is not None:
+                spec_file.append(f'{field}: {val}')
+
+        if self.distribution.get_url():
+            spec_file.append('Url: ' + self.distribution.get_url())
+
+        if self.distribution_name:
+            spec_file.append('Distribution: ' + self.distribution_name)
+
+        if self.build_requires:
+            spec_file.append('BuildRequires: ' + ' '.join(self.build_requires))
+
+        if self.icon:
+            spec_file.append('Icon: ' + os.path.basename(self.icon))
+
+        if self.no_autoreq:
+            spec_file.append('AutoReq: 0')
+
+        spec_file.extend([
+            '',
+            '%description',
+            self.distribution.get_long_description() or "",
+        ])
+
+        # put locale descriptions into spec file
+        # XXX again, suppressed because config file syntax doesn't
+        # easily support this ;-(
+        # for locale in self.descriptions.keys():
+        #    spec_file.extend([
+        #        '',
+        #        '%description -l ' + locale,
+        #        self.descriptions[locale],
+        #        ])
+
+        # rpm scripts
+        # figure out default build script
+        def_setup_call = f"{self.python} {os.path.basename(sys.argv[0])}"
+        def_build = f"{def_setup_call} build"
+        if self.use_rpm_opt_flags:
+            def_build = 'env CFLAGS="$RPM_OPT_FLAGS" ' + def_build
+
+        # insert contents of files
+
+        # XXX this is kind of misleading: user-supplied options are files
+        # that we open and interpolate into the spec file, but the defaults
+        # are just text that we drop in as-is.  Hmmm.
+
+        install_cmd = f'{def_setup_call} install -O1 --root=$RPM_BUILD_ROOT --record=INSTALLED_FILES'
+
+        script_options = [
+            ('prep', 'prep_script', "%setup -n %{name}-%{unmangled_version}"),
+            ('build', 'build_script', def_build),
+            ('install', 'install_script', install_cmd),
+            ('clean', 'clean_script', "rm -rf $RPM_BUILD_ROOT"),
+            ('verifyscript', 'verify_script', None),
+            ('pre', 'pre_install', None),
+            ('post', 'post_install', None),
+            ('preun', 'pre_uninstall', None),
+            ('postun', 'post_uninstall', None),
+        ]
+
+        for rpm_opt, attr, default in script_options:
+            # Insert contents of file referred to, if no file is referred to
+            # use 'default' as contents of script
+            val = getattr(self, attr)
+            if val or default:
+                spec_file.extend([
+                    '',
+                    '%' + rpm_opt,
+                ])
+                if val:
+                    with open(val) as f:
+                        spec_file.extend(f.read().split('\n'))
+                else:
+                    spec_file.append(default)
+
+        # files section
+        spec_file.extend([
+            '',
+            '%files -f INSTALLED_FILES',
+            '%defattr(-,root,root)',
+        ])
+
+        if self.doc_files:
+            spec_file.append('%doc ' + ' '.join(self.doc_files))
+
+        if self.changelog:
+            spec_file.extend([
+                '',
+                '%changelog',
+            ])
+            spec_file.extend(self.changelog)
+
+        return spec_file
+
+    def _format_changelog(self, changelog):
+        """Format the changelog correctly and convert it to a list of strings"""
+        if not changelog:
+            return changelog
+        new_changelog = []
+        for line in changelog.strip().split('\n'):
+            line = line.strip()
+            if line[0] == '*':
+                new_changelog.extend(['', line])
+            elif line[0] == '-':
+                new_changelog.append(line)
+            else:
+                new_changelog.append('  ' + line)
+
+        # strip trailing newline inserted by first changelog entry
+        if not new_changelog[0]:
+            del new_changelog[0]
+
+        return new_changelog
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build.py
new file mode 100644
index 00000000..ccd2c706
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build.py
@@ -0,0 +1,157 @@
+"""distutils.command.build
+
+Implements the Distutils 'build' command."""
+
+import os
+import sys
+import sysconfig
+
+from ..core import Command
+from ..errors import DistutilsOptionError
+from ..util import get_platform
+
+
+def show_compilers():
+    from ..ccompiler import show_compilers
+
+    show_compilers()
+
+
+class build(Command):
+    description = "build everything needed to install"
+
+    user_options = [
+        ('build-base=', 'b', "base directory for build library"),
+        ('build-purelib=', None, "build directory for platform-neutral distributions"),
+        ('build-platlib=', None, "build directory for platform-specific distributions"),
+        (
+            'build-lib=',
+            None,
+            "build directory for all distribution (defaults to either build-purelib or build-platlib",
+        ),
+        ('build-scripts=', None, "build directory for scripts"),
+        ('build-temp=', 't', "temporary build directory"),
+        (
+            'plat-name=',
+            'p',
+            f"platform name to build for, if supported [default: {get_platform()}]",
+        ),
+        ('compiler=', 'c', "specify the compiler type"),
+        ('parallel=', 'j', "number of parallel build jobs"),
+        ('debug', 'g', "compile extensions and libraries with debugging information"),
+        ('force', 'f', "forcibly build everything (ignore file timestamps)"),
+        ('executable=', 'e', "specify final destination interpreter path (build.py)"),
+    ]
+
+    boolean_options = ['debug', 'force']
+
+    help_options = [
+        ('help-compiler', None, "list available compilers", show_compilers),
+    ]
+
+    def initialize_options(self):
+        self.build_base = 'build'
+        # these are decided only after 'build_base' has its final value
+        # (unless overridden by the user or client)
+        self.build_purelib = None
+        self.build_platlib = None
+        self.build_lib = None
+        self.build_temp = None
+        self.build_scripts = None
+        self.compiler = None
+        self.plat_name = None
+        self.debug = None
+        self.force = False
+        self.executable = None
+        self.parallel = None
+
+    def finalize_options(self):  # noqa: C901
+        if self.plat_name is None:
+            self.plat_name = get_platform()
+        else:
+            # plat-name only supported for windows (other platforms are
+            # supported via ./configure flags, if at all).  Avoid misleading
+            # other platforms.
+            if os.name != 'nt':
+                raise DistutilsOptionError(
+                    "--plat-name only supported on Windows (try "
+                    "using './configure --help' on your platform)"
+                )
+
+        plat_specifier = f".{self.plat_name}-{sys.implementation.cache_tag}"
+
+        # Python 3.13+ with --disable-gil shouldn't share build directories
+        if sysconfig.get_config_var('Py_GIL_DISABLED'):
+            plat_specifier += 't'
+
+        # Make it so Python 2.x and Python 2.x with --with-pydebug don't
+        # share the same build directories. Doing so confuses the build
+        # process for C modules
+        if hasattr(sys, 'gettotalrefcount'):
+            plat_specifier += '-pydebug'
+
+        # 'build_purelib' and 'build_platlib' just default to 'lib' and
+        # 'lib.<plat>' under the base build directory.  We only use one of
+        # them for a given distribution, though --
+        if self.build_purelib is None:
+            self.build_purelib = os.path.join(self.build_base, 'lib')
+        if self.build_platlib is None:
+            self.build_platlib = os.path.join(self.build_base, 'lib' + plat_specifier)
+
+        # 'build_lib' is the actual directory that we will use for this
+        # particular module distribution -- if user didn't supply it, pick
+        # one of 'build_purelib' or 'build_platlib'.
+        if self.build_lib is None:
+            if self.distribution.has_ext_modules():
+                self.build_lib = self.build_platlib
+            else:
+                self.build_lib = self.build_purelib
+
+        # 'build_temp' -- temporary directory for compiler turds,
+        # "build/temp.<plat>"
+        if self.build_temp is None:
+            self.build_temp = os.path.join(self.build_base, 'temp' + plat_specifier)
+        if self.build_scripts is None:
+            self.build_scripts = os.path.join(
+                self.build_base,
+                f'scripts-{sys.version_info.major}.{sys.version_info.minor}',
+            )
+
+        if self.executable is None and sys.executable:
+            self.executable = os.path.normpath(sys.executable)
+
+        if isinstance(self.parallel, str):
+            try:
+                self.parallel = int(self.parallel)
+            except ValueError:
+                raise DistutilsOptionError("parallel should be an integer")
+
+    def run(self):
+        # Run all relevant sub-commands.  This will be some subset of:
+        #  - build_py      - pure Python modules
+        #  - build_clib    - standalone C libraries
+        #  - build_ext     - Python extensions
+        #  - build_scripts - (Python) scripts
+        for cmd_name in self.get_sub_commands():
+            self.run_command(cmd_name)
+
+    # -- Predicates for the sub-command list ---------------------------
+
+    def has_pure_modules(self):
+        return self.distribution.has_pure_modules()
+
+    def has_c_libraries(self):
+        return self.distribution.has_c_libraries()
+
+    def has_ext_modules(self):
+        return self.distribution.has_ext_modules()
+
+    def has_scripts(self):
+        return self.distribution.has_scripts()
+
+    sub_commands = [
+        ('build_py', has_pure_modules),
+        ('build_clib', has_c_libraries),
+        ('build_ext', has_ext_modules),
+        ('build_scripts', has_scripts),
+    ]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_clib.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_clib.py
new file mode 100644
index 00000000..3e183276
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_clib.py
@@ -0,0 +1,208 @@
+"""distutils.command.build_clib
+
+Implements the Distutils 'build_clib' command, to build a C/C++ library
+that is included in the module distribution and needed by an extension
+module."""
+
+
+# XXX this module has *lots* of code ripped-off quite transparently from
+# build_ext.py -- not surprisingly really, as the work required to build
+# a static library from a collection of C source files is not really all
+# that different from what's required to build a shared object file from
+# a collection of C source files.  Nevertheless, I haven't done the
+# necessary refactoring to account for the overlap in code between the
+# two modules, mainly because a number of subtle details changed in the
+# cut 'n paste.  Sigh.
+
+import os
+from distutils._log import log
+from typing import ClassVar
+
+from ..core import Command
+from ..errors import DistutilsSetupError
+from ..sysconfig import customize_compiler
+
+
+def show_compilers():
+    from ..ccompiler import show_compilers
+
+    show_compilers()
+
+
+class build_clib(Command):
+    description = "build C/C++ libraries used by Python extensions"
+
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ('build-clib=', 'b', "directory to build C/C++ libraries to"),
+        ('build-temp=', 't', "directory to put temporary build by-products"),
+        ('debug', 'g', "compile with debugging information"),
+        ('force', 'f', "forcibly build everything (ignore file timestamps)"),
+        ('compiler=', 'c', "specify the compiler type"),
+    ]
+
+    boolean_options = ['debug', 'force']
+
+    help_options = [
+        ('help-compiler', None, "list available compilers", show_compilers),
+    ]
+
+    def initialize_options(self):
+        self.build_clib = None
+        self.build_temp = None
+
+        # List of libraries to build
+        self.libraries = None
+
+        # Compilation options for all libraries
+        self.include_dirs = None
+        self.define = None
+        self.undef = None
+        self.debug = None
+        self.force = False
+        self.compiler = None
+
+    def finalize_options(self):
+        # This might be confusing: both build-clib and build-temp default
+        # to build-temp as defined by the "build" command.  This is because
+        # I think that C libraries are really just temporary build
+        # by-products, at least from the point of view of building Python
+        # extensions -- but I want to keep my options open.
+        self.set_undefined_options(
+            'build',
+            ('build_temp', 'build_clib'),
+            ('build_temp', 'build_temp'),
+            ('compiler', 'compiler'),
+            ('debug', 'debug'),
+            ('force', 'force'),
+        )
+
+        self.libraries = self.distribution.libraries
+        if self.libraries:
+            self.check_library_list(self.libraries)
+
+        if self.include_dirs is None:
+            self.include_dirs = self.distribution.include_dirs or []
+        if isinstance(self.include_dirs, str):
+            self.include_dirs = self.include_dirs.split(os.pathsep)
+
+        # XXX same as for build_ext -- what about 'self.define' and
+        # 'self.undef' ?
+
+    def run(self):
+        if not self.libraries:
+            return
+
+        # Yech -- this is cut 'n pasted from build_ext.py!
+        from ..ccompiler import new_compiler
+
+        self.compiler = new_compiler(
+            compiler=self.compiler, dry_run=self.dry_run, force=self.force
+        )
+        customize_compiler(self.compiler)
+
+        if self.include_dirs is not None:
+            self.compiler.set_include_dirs(self.include_dirs)
+        if self.define is not None:
+            # 'define' option is a list of (name,value) tuples
+            for name, value in self.define:
+                self.compiler.define_macro(name, value)
+        if self.undef is not None:
+            for macro in self.undef:
+                self.compiler.undefine_macro(macro)
+
+        self.build_libraries(self.libraries)
+
+    def check_library_list(self, libraries):
+        """Ensure that the list of libraries is valid.
+
+        `library` is presumably provided as a command option 'libraries'.
+        This method checks that it is a list of 2-tuples, where the tuples
+        are (library_name, build_info_dict).
+
+        Raise DistutilsSetupError if the structure is invalid anywhere;
+        just returns otherwise.
+        """
+        if not isinstance(libraries, list):
+            raise DistutilsSetupError("'libraries' option must be a list of tuples")
+
+        for lib in libraries:
+            if not isinstance(lib, tuple) and len(lib) != 2:
+                raise DistutilsSetupError("each element of 'libraries' must a 2-tuple")
+
+            name, build_info = lib
+
+            if not isinstance(name, str):
+                raise DistutilsSetupError(
+                    "first element of each tuple in 'libraries' "
+                    "must be a string (the library name)"
+                )
+
+            if '/' in name or (os.sep != '/' and os.sep in name):
+                raise DistutilsSetupError(
+                    f"bad library name '{lib[0]}': may not contain directory separators"
+                )
+
+            if not isinstance(build_info, dict):
+                raise DistutilsSetupError(
+                    "second element of each tuple in 'libraries' "
+                    "must be a dictionary (build info)"
+                )
+
+    def get_library_names(self):
+        # Assume the library list is valid -- 'check_library_list()' is
+        # called from 'finalize_options()', so it should be!
+        if not self.libraries:
+            return None
+
+        lib_names = []
+        for lib_name, _build_info in self.libraries:
+            lib_names.append(lib_name)
+        return lib_names
+
+    def get_source_files(self):
+        self.check_library_list(self.libraries)
+        filenames = []
+        for lib_name, build_info in self.libraries:
+            sources = build_info.get('sources')
+            if sources is None or not isinstance(sources, (list, tuple)):
+                raise DistutilsSetupError(
+                    f"in 'libraries' option (library '{lib_name}'), "
+                    "'sources' must be present and must be "
+                    "a list of source filenames"
+                )
+
+            filenames.extend(sources)
+        return filenames
+
+    def build_libraries(self, libraries):
+        for lib_name, build_info in libraries:
+            sources = build_info.get('sources')
+            if sources is None or not isinstance(sources, (list, tuple)):
+                raise DistutilsSetupError(
+                    f"in 'libraries' option (library '{lib_name}'), "
+                    "'sources' must be present and must be "
+                    "a list of source filenames"
+                )
+            sources = list(sources)
+
+            log.info("building '%s' library", lib_name)
+
+            # First, compile the source code to object files in the library
+            # directory.  (This should probably change to putting object
+            # files in a temporary build directory.)
+            macros = build_info.get('macros')
+            include_dirs = build_info.get('include_dirs')
+            objects = self.compiler.compile(
+                sources,
+                output_dir=self.build_temp,
+                macros=macros,
+                include_dirs=include_dirs,
+                debug=self.debug,
+            )
+
+            # Now "link" the object files together into a static library.
+            # (On Unix at least, this isn't really linking -- it just
+            # builds an archive.  Whatever.)
+            self.compiler.create_static_lib(
+                objects, lib_name, output_dir=self.build_clib, debug=self.debug
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_ext.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_ext.py
new file mode 100644
index 00000000..cf60bd0a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_ext.py
@@ -0,0 +1,800 @@
+"""distutils.command.build_ext
+
+Implements the Distutils 'build_ext' command, for building extension
+modules (currently limited to C extensions, should accommodate C++
+extensions ASAP)."""
+
+import contextlib
+import os
+import re
+import sys
+from distutils._log import log
+from site import USER_BASE
+
+from .._modified import newer_group
+from ..core import Command
+from ..errors import (
+    CCompilerError,
+    CompileError,
+    DistutilsError,
+    DistutilsOptionError,
+    DistutilsPlatformError,
+    DistutilsSetupError,
+)
+from ..extension import Extension
+from ..sysconfig import customize_compiler, get_config_h_filename, get_python_version
+from ..util import get_platform, is_freethreaded, is_mingw
+
+# An extension name is just a dot-separated list of Python NAMEs (ie.
+# the same as a fully-qualified module name).
+extension_name_re = re.compile(r'^[a-zA-Z_][a-zA-Z_0-9]*(\.[a-zA-Z_][a-zA-Z_0-9]*)*$')
+
+
+def show_compilers():
+    from ..ccompiler import show_compilers
+
+    show_compilers()
+
+
+class build_ext(Command):
+    description = "build C/C++ extensions (compile/link to build directory)"
+
+    # XXX thoughts on how to deal with complex command-line options like
+    # these, i.e. how to make it so fancy_getopt can suck them off the
+    # command line and make it look like setup.py defined the appropriate
+    # lists of tuples of what-have-you.
+    #   - each command needs a callback to process its command-line options
+    #   - Command.__init__() needs access to its share of the whole
+    #     command line (must ultimately come from
+    #     Distribution.parse_command_line())
+    #   - it then calls the current command class' option-parsing
+    #     callback to deal with weird options like -D, which have to
+    #     parse the option text and churn out some custom data
+    #     structure
+    #   - that data structure (in this case, a list of 2-tuples)
+    #     will then be present in the command object by the time
+    #     we get to finalize_options() (i.e. the constructor
+    #     takes care of both command-line and client options
+    #     in between initialize_options() and finalize_options())
+
+    sep_by = f" (separated by '{os.pathsep}')"
+    user_options = [
+        ('build-lib=', 'b', "directory for compiled extension modules"),
+        ('build-temp=', 't', "directory for temporary files (build by-products)"),
+        (
+            'plat-name=',
+            'p',
+            "platform name to cross-compile for, if supported "
+            f"[default: {get_platform()}]",
+        ),
+        (
+            'inplace',
+            'i',
+            "ignore build-lib and put compiled extensions into the source "
+            "directory alongside your pure Python modules",
+        ),
+        (
+            'include-dirs=',
+            'I',
+            "list of directories to search for header files" + sep_by,
+        ),
+        ('define=', 'D', "C preprocessor macros to define"),
+        ('undef=', 'U', "C preprocessor macros to undefine"),
+        ('libraries=', 'l', "external C libraries to link with"),
+        (
+            'library-dirs=',
+            'L',
+            "directories to search for external C libraries" + sep_by,
+        ),
+        ('rpath=', 'R', "directories to search for shared C libraries at runtime"),
+        ('link-objects=', 'O', "extra explicit link objects to include in the link"),
+        ('debug', 'g', "compile/link with debugging information"),
+        ('force', 'f', "forcibly build everything (ignore file timestamps)"),
+        ('compiler=', 'c', "specify the compiler type"),
+        ('parallel=', 'j', "number of parallel build jobs"),
+        ('swig-cpp', None, "make SWIG create C++ files (default is C)"),
+        ('swig-opts=', None, "list of SWIG command line options"),
+        ('swig=', None, "path to the SWIG executable"),
+        ('user', None, "add user include, library and rpath"),
+    ]
+
+    boolean_options = ['inplace', 'debug', 'force', 'swig-cpp', 'user']
+
+    help_options = [
+        ('help-compiler', None, "list available compilers", show_compilers),
+    ]
+
+    def initialize_options(self):
+        self.extensions = None
+        self.build_lib = None
+        self.plat_name = None
+        self.build_temp = None
+        self.inplace = False
+        self.package = None
+
+        self.include_dirs = None
+        self.define = None
+        self.undef = None
+        self.libraries = None
+        self.library_dirs = None
+        self.rpath = None
+        self.link_objects = None
+        self.debug = None
+        self.force = None
+        self.compiler = None
+        self.swig = None
+        self.swig_cpp = None
+        self.swig_opts = None
+        self.user = None
+        self.parallel = None
+
+    @staticmethod
+    def _python_lib_dir(sysconfig):
+        """
+        Resolve Python's library directory for building extensions
+        that rely on a shared Python library.
+
+        See python/cpython#44264 and python/cpython#48686
+        """
+        if not sysconfig.get_config_var('Py_ENABLE_SHARED'):
+            return
+
+        if sysconfig.python_build:
+            yield '.'
+            return
+
+        if sys.platform == 'zos':
+            # On z/OS, a user is not required to install Python to
+            # a predetermined path, but can use Python portably
+            installed_dir = sysconfig.get_config_var('base')
+            lib_dir = sysconfig.get_config_var('platlibdir')
+            yield os.path.join(installed_dir, lib_dir)
+        else:
+            # building third party extensions
+            yield sysconfig.get_config_var('LIBDIR')
+
+    def finalize_options(self):  # noqa: C901
+        from distutils import sysconfig
+
+        self.set_undefined_options(
+            'build',
+            ('build_lib', 'build_lib'),
+            ('build_temp', 'build_temp'),
+            ('compiler', 'compiler'),
+            ('debug', 'debug'),
+            ('force', 'force'),
+            ('parallel', 'parallel'),
+            ('plat_name', 'plat_name'),
+        )
+
+        if self.package is None:
+            self.package = self.distribution.ext_package
+
+        self.extensions = self.distribution.ext_modules
+
+        # Make sure Python's include directories (for Python.h, pyconfig.h,
+        # etc.) are in the include search path.
+        py_include = sysconfig.get_python_inc()
+        plat_py_include = sysconfig.get_python_inc(plat_specific=True)
+        if self.include_dirs is None:
+            self.include_dirs = self.distribution.include_dirs or []
+        if isinstance(self.include_dirs, str):
+            self.include_dirs = self.include_dirs.split(os.pathsep)
+
+        # If in a virtualenv, add its include directory
+        # Issue 16116
+        if sys.exec_prefix != sys.base_exec_prefix:
+            self.include_dirs.append(os.path.join(sys.exec_prefix, 'include'))
+
+        # Put the Python "system" include dir at the end, so that
+        # any local include dirs take precedence.
+        self.include_dirs.extend(py_include.split(os.path.pathsep))
+        if plat_py_include != py_include:
+            self.include_dirs.extend(plat_py_include.split(os.path.pathsep))
+
+        self.ensure_string_list('libraries')
+        self.ensure_string_list('link_objects')
+
+        # Life is easier if we're not forever checking for None, so
+        # simplify these options to empty lists if unset
+        if self.libraries is None:
+            self.libraries = []
+        if self.library_dirs is None:
+            self.library_dirs = []
+        elif isinstance(self.library_dirs, str):
+            self.library_dirs = self.library_dirs.split(os.pathsep)
+
+        if self.rpath is None:
+            self.rpath = []
+        elif isinstance(self.rpath, str):
+            self.rpath = self.rpath.split(os.pathsep)
+
+        # for extensions under windows use different directories
+        # for Release and Debug builds.
+        # also Python's library directory must be appended to library_dirs
+        if os.name == 'nt' and not is_mingw():
+            # the 'libs' directory is for binary installs - we assume that
+            # must be the *native* platform.  But we don't really support
+            # cross-compiling via a binary install anyway, so we let it go.
+            self.library_dirs.append(os.path.join(sys.exec_prefix, 'libs'))
+            if sys.base_exec_prefix != sys.prefix:  # Issue 16116
+                self.library_dirs.append(os.path.join(sys.base_exec_prefix, 'libs'))
+            if self.debug:
+                self.build_temp = os.path.join(self.build_temp, "Debug")
+            else:
+                self.build_temp = os.path.join(self.build_temp, "Release")
+
+            # Append the source distribution include and library directories,
+            # this allows distutils on windows to work in the source tree
+            self.include_dirs.append(os.path.dirname(get_config_h_filename()))
+            self.library_dirs.append(sys.base_exec_prefix)
+
+            # Use the .lib files for the correct architecture
+            if self.plat_name == 'win32':
+                suffix = 'win32'
+            else:
+                # win-amd64
+                suffix = self.plat_name[4:]
+            new_lib = os.path.join(sys.exec_prefix, 'PCbuild')
+            if suffix:
+                new_lib = os.path.join(new_lib, suffix)
+            self.library_dirs.append(new_lib)
+
+        # For extensions under Cygwin, Python's library directory must be
+        # appended to library_dirs
+        if sys.platform[:6] == 'cygwin':
+            if not sysconfig.python_build:
+                # building third party extensions
+                self.library_dirs.append(
+                    os.path.join(
+                        sys.prefix, "lib", "python" + get_python_version(), "config"
+                    )
+                )
+            else:
+                # building python standard extensions
+                self.library_dirs.append('.')
+
+        self.library_dirs.extend(self._python_lib_dir(sysconfig))
+
+        # The argument parsing will result in self.define being a string, but
+        # it has to be a list of 2-tuples.  All the preprocessor symbols
+        # specified by the 'define' option will be set to '1'.  Multiple
+        # symbols can be separated with commas.
+
+        if self.define:
+            defines = self.define.split(',')
+            self.define = [(symbol, '1') for symbol in defines]
+
+        # The option for macros to undefine is also a string from the
+        # option parsing, but has to be a list.  Multiple symbols can also
+        # be separated with commas here.
+        if self.undef:
+            self.undef = self.undef.split(',')
+
+        if self.swig_opts is None:
+            self.swig_opts = []
+        else:
+            self.swig_opts = self.swig_opts.split(' ')
+
+        # Finally add the user include and library directories if requested
+        if self.user:
+            user_include = os.path.join(USER_BASE, "include")
+            user_lib = os.path.join(USER_BASE, "lib")
+            if os.path.isdir(user_include):
+                self.include_dirs.append(user_include)
+            if os.path.isdir(user_lib):
+                self.library_dirs.append(user_lib)
+                self.rpath.append(user_lib)
+
+        if isinstance(self.parallel, str):
+            try:
+                self.parallel = int(self.parallel)
+            except ValueError:
+                raise DistutilsOptionError("parallel should be an integer")
+
+    def run(self):  # noqa: C901
+        from ..ccompiler import new_compiler
+
+        # 'self.extensions', as supplied by setup.py, is a list of
+        # Extension instances.  See the documentation for Extension (in
+        # distutils.extension) for details.
+        #
+        # For backwards compatibility with Distutils 0.8.2 and earlier, we
+        # also allow the 'extensions' list to be a list of tuples:
+        #    (ext_name, build_info)
+        # where build_info is a dictionary containing everything that
+        # Extension instances do except the name, with a few things being
+        # differently named.  We convert these 2-tuples to Extension
+        # instances as needed.
+
+        if not self.extensions:
+            return
+
+        # If we were asked to build any C/C++ libraries, make sure that the
+        # directory where we put them is in the library search path for
+        # linking extensions.
+        if self.distribution.has_c_libraries():
+            build_clib = self.get_finalized_command('build_clib')
+            self.libraries.extend(build_clib.get_library_names() or [])
+            self.library_dirs.append(build_clib.build_clib)
+
+        # Setup the CCompiler object that we'll use to do all the
+        # compiling and linking
+        self.compiler = new_compiler(
+            compiler=self.compiler,
+            verbose=self.verbose,
+            dry_run=self.dry_run,
+            force=self.force,
+        )
+        customize_compiler(self.compiler)
+        # If we are cross-compiling, init the compiler now (if we are not
+        # cross-compiling, init would not hurt, but people may rely on
+        # late initialization of compiler even if they shouldn't...)
+        if os.name == 'nt' and self.plat_name != get_platform():
+            self.compiler.initialize(self.plat_name)
+
+        # The official Windows free threaded Python installer doesn't set
+        # Py_GIL_DISABLED because its pyconfig.h is shared with the
+        # default build, so define it here (pypa/setuptools#4662).
+        if os.name == 'nt' and is_freethreaded():
+            self.compiler.define_macro('Py_GIL_DISABLED', '1')
+
+        # And make sure that any compile/link-related options (which might
+        # come from the command-line or from the setup script) are set in
+        # that CCompiler object -- that way, they automatically apply to
+        # all compiling and linking done here.
+        if self.include_dirs is not None:
+            self.compiler.set_include_dirs(self.include_dirs)
+        if self.define is not None:
+            # 'define' option is a list of (name,value) tuples
+            for name, value in self.define:
+                self.compiler.define_macro(name, value)
+        if self.undef is not None:
+            for macro in self.undef:
+                self.compiler.undefine_macro(macro)
+        if self.libraries is not None:
+            self.compiler.set_libraries(self.libraries)
+        if self.library_dirs is not None:
+            self.compiler.set_library_dirs(self.library_dirs)
+        if self.rpath is not None:
+            self.compiler.set_runtime_library_dirs(self.rpath)
+        if self.link_objects is not None:
+            self.compiler.set_link_objects(self.link_objects)
+
+        # Now actually compile and link everything.
+        self.build_extensions()
+
+    def check_extensions_list(self, extensions):  # noqa: C901
+        """Ensure that the list of extensions (presumably provided as a
+        command option 'extensions') is valid, i.e. it is a list of
+        Extension objects.  We also support the old-style list of 2-tuples,
+        where the tuples are (ext_name, build_info), which are converted to
+        Extension instances here.
+
+        Raise DistutilsSetupError if the structure is invalid anywhere;
+        just returns otherwise.
+        """
+        if not isinstance(extensions, list):
+            raise DistutilsSetupError(
+                "'ext_modules' option must be a list of Extension instances"
+            )
+
+        for i, ext in enumerate(extensions):
+            if isinstance(ext, Extension):
+                continue  # OK! (assume type-checking done
+                # by Extension constructor)
+
+            if not isinstance(ext, tuple) or len(ext) != 2:
+                raise DistutilsSetupError(
+                    "each element of 'ext_modules' option must be an "
+                    "Extension instance or 2-tuple"
+                )
+
+            ext_name, build_info = ext
+
+            log.warning(
+                "old-style (ext_name, build_info) tuple found in "
+                "ext_modules for extension '%s' "
+                "-- please convert to Extension instance",
+                ext_name,
+            )
+
+            if not (isinstance(ext_name, str) and extension_name_re.match(ext_name)):
+                raise DistutilsSetupError(
+                    "first element of each tuple in 'ext_modules' "
+                    "must be the extension name (a string)"
+                )
+
+            if not isinstance(build_info, dict):
+                raise DistutilsSetupError(
+                    "second element of each tuple in 'ext_modules' "
+                    "must be a dictionary (build info)"
+                )
+
+            # OK, the (ext_name, build_info) dict is type-safe: convert it
+            # to an Extension instance.
+            ext = Extension(ext_name, build_info['sources'])
+
+            # Easy stuff: one-to-one mapping from dict elements to
+            # instance attributes.
+            for key in (
+                'include_dirs',
+                'library_dirs',
+                'libraries',
+                'extra_objects',
+                'extra_compile_args',
+                'extra_link_args',
+            ):
+                val = build_info.get(key)
+                if val is not None:
+                    setattr(ext, key, val)
+
+            # Medium-easy stuff: same syntax/semantics, different names.
+            ext.runtime_library_dirs = build_info.get('rpath')
+            if 'def_file' in build_info:
+                log.warning("'def_file' element of build info dict no longer supported")
+
+            # Non-trivial stuff: 'macros' split into 'define_macros'
+            # and 'undef_macros'.
+            macros = build_info.get('macros')
+            if macros:
+                ext.define_macros = []
+                ext.undef_macros = []
+                for macro in macros:
+                    if not (isinstance(macro, tuple) and len(macro) in (1, 2)):
+                        raise DistutilsSetupError(
+                            "'macros' element of build info dict must be 1- or 2-tuple"
+                        )
+                    if len(macro) == 1:
+                        ext.undef_macros.append(macro[0])
+                    elif len(macro) == 2:
+                        ext.define_macros.append(macro)
+
+            extensions[i] = ext
+
+    def get_source_files(self):
+        self.check_extensions_list(self.extensions)
+        filenames = []
+
+        # Wouldn't it be neat if we knew the names of header files too...
+        for ext in self.extensions:
+            filenames.extend(ext.sources)
+        return filenames
+
+    def get_outputs(self):
+        # Sanity check the 'extensions' list -- can't assume this is being
+        # done in the same run as a 'build_extensions()' call (in fact, we
+        # can probably assume that it *isn't*!).
+        self.check_extensions_list(self.extensions)
+
+        # And build the list of output (built) filenames.  Note that this
+        # ignores the 'inplace' flag, and assumes everything goes in the
+        # "build" tree.
+        return [self.get_ext_fullpath(ext.name) for ext in self.extensions]
+
+    def build_extensions(self):
+        # First, sanity-check the 'extensions' list
+        self.check_extensions_list(self.extensions)
+        if self.parallel:
+            self._build_extensions_parallel()
+        else:
+            self._build_extensions_serial()
+
+    def _build_extensions_parallel(self):
+        workers = self.parallel
+        if self.parallel is True:
+            workers = os.cpu_count()  # may return None
+        try:
+            from concurrent.futures import ThreadPoolExecutor
+        except ImportError:
+            workers = None
+
+        if workers is None:
+            self._build_extensions_serial()
+            return
+
+        with ThreadPoolExecutor(max_workers=workers) as executor:
+            futures = [
+                executor.submit(self.build_extension, ext) for ext in self.extensions
+            ]
+            for ext, fut in zip(self.extensions, futures):
+                with self._filter_build_errors(ext):
+                    fut.result()
+
+    def _build_extensions_serial(self):
+        for ext in self.extensions:
+            with self._filter_build_errors(ext):
+                self.build_extension(ext)
+
+    @contextlib.contextmanager
+    def _filter_build_errors(self, ext):
+        try:
+            yield
+        except (CCompilerError, DistutilsError, CompileError) as e:
+            if not ext.optional:
+                raise
+            self.warn(f'building extension "{ext.name}" failed: {e}')
+
+    def build_extension(self, ext):
+        sources = ext.sources
+        if sources is None or not isinstance(sources, (list, tuple)):
+            raise DistutilsSetupError(
+                f"in 'ext_modules' option (extension '{ext.name}'), "
+                "'sources' must be present and must be "
+                "a list of source filenames"
+            )
+        # sort to make the resulting .so file build reproducible
+        sources = sorted(sources)
+
+        ext_path = self.get_ext_fullpath(ext.name)
+        depends = sources + ext.depends
+        if not (self.force or newer_group(depends, ext_path, 'newer')):
+            log.debug("skipping '%s' extension (up-to-date)", ext.name)
+            return
+        else:
+            log.info("building '%s' extension", ext.name)
+
+        # First, scan the sources for SWIG definition files (.i), run
+        # SWIG on 'em to create .c files, and modify the sources list
+        # accordingly.
+        sources = self.swig_sources(sources, ext)
+
+        # Next, compile the source code to object files.
+
+        # XXX not honouring 'define_macros' or 'undef_macros' -- the
+        # CCompiler API needs to change to accommodate this, and I
+        # want to do one thing at a time!
+
+        # Two possible sources for extra compiler arguments:
+        #   - 'extra_compile_args' in Extension object
+        #   - CFLAGS environment variable (not particularly
+        #     elegant, but people seem to expect it and I
+        #     guess it's useful)
+        # The environment variable should take precedence, and
+        # any sensible compiler will give precedence to later
+        # command line args.  Hence we combine them in order:
+        extra_args = ext.extra_compile_args or []
+
+        macros = ext.define_macros[:]
+        for undef in ext.undef_macros:
+            macros.append((undef,))
+
+        objects = self.compiler.compile(
+            sources,
+            output_dir=self.build_temp,
+            macros=macros,
+            include_dirs=ext.include_dirs,
+            debug=self.debug,
+            extra_postargs=extra_args,
+            depends=ext.depends,
+        )
+
+        # XXX outdated variable, kept here in case third-part code
+        # needs it.
+        self._built_objects = objects[:]
+
+        # Now link the object files together into a "shared object" --
+        # of course, first we have to figure out all the other things
+        # that go into the mix.
+        if ext.extra_objects:
+            objects.extend(ext.extra_objects)
+        extra_args = ext.extra_link_args or []
+
+        # Detect target language, if not provided
+        language = ext.language or self.compiler.detect_language(sources)
+
+        self.compiler.link_shared_object(
+            objects,
+            ext_path,
+            libraries=self.get_libraries(ext),
+            library_dirs=ext.library_dirs,
+            runtime_library_dirs=ext.runtime_library_dirs,
+            extra_postargs=extra_args,
+            export_symbols=self.get_export_symbols(ext),
+            debug=self.debug,
+            build_temp=self.build_temp,
+            target_lang=language,
+        )
+
+    def swig_sources(self, sources, extension):
+        """Walk the list of source files in 'sources', looking for SWIG
+        interface (.i) files.  Run SWIG on all that are found, and
+        return a modified 'sources' list with SWIG source files replaced
+        by the generated C (or C++) files.
+        """
+        new_sources = []
+        swig_sources = []
+        swig_targets = {}
+
+        # XXX this drops generated C/C++ files into the source tree, which
+        # is fine for developers who want to distribute the generated
+        # source -- but there should be an option to put SWIG output in
+        # the temp dir.
+
+        if self.swig_cpp:
+            log.warning("--swig-cpp is deprecated - use --swig-opts=-c++")
+
+        if (
+            self.swig_cpp
+            or ('-c++' in self.swig_opts)
+            or ('-c++' in extension.swig_opts)
+        ):
+            target_ext = '.cpp'
+        else:
+            target_ext = '.c'
+
+        for source in sources:
+            (base, ext) = os.path.splitext(source)
+            if ext == ".i":  # SWIG interface file
+                new_sources.append(base + '_wrap' + target_ext)
+                swig_sources.append(source)
+                swig_targets[source] = new_sources[-1]
+            else:
+                new_sources.append(source)
+
+        if not swig_sources:
+            return new_sources
+
+        swig = self.swig or self.find_swig()
+        swig_cmd = [swig, "-python"]
+        swig_cmd.extend(self.swig_opts)
+        if self.swig_cpp:
+            swig_cmd.append("-c++")
+
+        # Do not override commandline arguments
+        if not self.swig_opts:
+            swig_cmd.extend(extension.swig_opts)
+
+        for source in swig_sources:
+            target = swig_targets[source]
+            log.info("swigging %s to %s", source, target)
+            self.spawn(swig_cmd + ["-o", target, source])
+
+        return new_sources
+
+    def find_swig(self):
+        """Return the name of the SWIG executable.  On Unix, this is
+        just "swig" -- it should be in the PATH.  Tries a bit harder on
+        Windows.
+        """
+        if os.name == "posix":
+            return "swig"
+        elif os.name == "nt":
+            # Look for SWIG in its standard installation directory on
+            # Windows (or so I presume!).  If we find it there, great;
+            # if not, act like Unix and assume it's in the PATH.
+            for vers in ("1.3", "1.2", "1.1"):
+                fn = os.path.join(f"c:\\swig{vers}", "swig.exe")
+                if os.path.isfile(fn):
+                    return fn
+            else:
+                return "swig.exe"
+        else:
+            raise DistutilsPlatformError(
+                f"I don't know how to find (much less run) SWIG on platform '{os.name}'"
+            )
+
+    # -- Name generators -----------------------------------------------
+    # (extension names, filenames, whatever)
+    def get_ext_fullpath(self, ext_name):
+        """Returns the path of the filename for a given extension.
+
+        The file is located in `build_lib` or directly in the package
+        (inplace option).
+        """
+        fullname = self.get_ext_fullname(ext_name)
+        modpath = fullname.split('.')
+        filename = self.get_ext_filename(modpath[-1])
+
+        if not self.inplace:
+            # no further work needed
+            # returning :
+            #   build_dir/package/path/filename
+            filename = os.path.join(*modpath[:-1] + [filename])
+            return os.path.join(self.build_lib, filename)
+
+        # the inplace option requires to find the package directory
+        # using the build_py command for that
+        package = '.'.join(modpath[0:-1])
+        build_py = self.get_finalized_command('build_py')
+        package_dir = os.path.abspath(build_py.get_package_dir(package))
+
+        # returning
+        #   package_dir/filename
+        return os.path.join(package_dir, filename)
+
+    def get_ext_fullname(self, ext_name):
+        """Returns the fullname of a given extension name.
+
+        Adds the `package.` prefix"""
+        if self.package is None:
+            return ext_name
+        else:
+            return self.package + '.' + ext_name
+
+    def get_ext_filename(self, ext_name):
+        r"""Convert the name of an extension (eg. "foo.bar") into the name
+        of the file from which it will be loaded (eg. "foo/bar.so", or
+        "foo\bar.pyd").
+        """
+        from ..sysconfig import get_config_var
+
+        ext_path = ext_name.split('.')
+        ext_suffix = get_config_var('EXT_SUFFIX')
+        return os.path.join(*ext_path) + ext_suffix
+
+    def get_export_symbols(self, ext):
+        """Return the list of symbols that a shared extension has to
+        export.  This either uses 'ext.export_symbols' or, if it's not
+        provided, "PyInit_" + module_name.  Only relevant on Windows, where
+        the .pyd file (DLL) must export the module "PyInit_" function.
+        """
+        name = ext.name.split('.')[-1]
+        try:
+            # Unicode module name support as defined in PEP-489
+            # https://peps.python.org/pep-0489/#export-hook-name
+            name.encode('ascii')
+        except UnicodeEncodeError:
+            suffix = 'U_' + name.encode('punycode').replace(b'-', b'_').decode('ascii')
+        else:
+            suffix = "_" + name
+
+        initfunc_name = "PyInit" + suffix
+        if initfunc_name not in ext.export_symbols:
+            ext.export_symbols.append(initfunc_name)
+        return ext.export_symbols
+
+    def get_libraries(self, ext):  # noqa: C901
+        """Return the list of libraries to link against when building a
+        shared extension.  On most platforms, this is just 'ext.libraries';
+        on Windows, we add the Python library (eg. python20.dll).
+        """
+        # The python library is always needed on Windows.  For MSVC, this
+        # is redundant, since the library is mentioned in a pragma in
+        # pyconfig.h that MSVC groks.  The other Windows compilers all seem
+        # to need it mentioned explicitly, though, so that's what we do.
+        # Append '_d' to the python import library on debug builds.
+        if sys.platform == "win32" and not is_mingw():
+            from .._msvccompiler import MSVCCompiler
+
+            if not isinstance(self.compiler, MSVCCompiler):
+                template = "python%d%d"
+                if self.debug:
+                    template = template + '_d'
+                pythonlib = template % (
+                    sys.hexversion >> 24,
+                    (sys.hexversion >> 16) & 0xFF,
+                )
+                # don't extend ext.libraries, it may be shared with other
+                # extensions, it is a reference to the original list
+                return ext.libraries + [pythonlib]
+        else:
+            # On Android only the main executable and LD_PRELOADs are considered
+            # to be RTLD_GLOBAL, all the dependencies of the main executable
+            # remain RTLD_LOCAL and so the shared libraries must be linked with
+            # libpython when python is built with a shared python library (issue
+            # bpo-21536).
+            # On Cygwin (and if required, other POSIX-like platforms based on
+            # Windows like MinGW) it is simply necessary that all symbols in
+            # shared libraries are resolved at link time.
+            from ..sysconfig import get_config_var
+
+            link_libpython = False
+            if get_config_var('Py_ENABLE_SHARED'):
+                # A native build on an Android device or on Cygwin
+                if hasattr(sys, 'getandroidapilevel'):
+                    link_libpython = True
+                elif sys.platform == 'cygwin' or is_mingw():
+                    link_libpython = True
+                elif '_PYTHON_HOST_PLATFORM' in os.environ:
+                    # We are cross-compiling for one of the relevant platforms
+                    if get_config_var('ANDROID_API_LEVEL') != 0:
+                        link_libpython = True
+                    elif get_config_var('MACHDEP') == 'cygwin':
+                        link_libpython = True
+
+            if link_libpython:
+                ldversion = get_config_var('LDVERSION')
+                return ext.libraries + ['python' + ldversion]
+
+        return ext.libraries
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_py.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_py.py
new file mode 100644
index 00000000..49d71034
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_py.py
@@ -0,0 +1,406 @@
+"""distutils.command.build_py
+
+Implements the Distutils 'build_py' command."""
+
+import glob
+import importlib.util
+import os
+import sys
+from distutils._log import log
+
+from ..core import Command
+from ..errors import DistutilsFileError, DistutilsOptionError
+from ..util import convert_path
+
+
+class build_py(Command):
+    description = "\"build\" pure Python modules (copy to build directory)"
+
+    user_options = [
+        ('build-lib=', 'd', "directory to \"build\" (copy) to"),
+        ('compile', 'c', "compile .py to .pyc"),
+        ('no-compile', None, "don't compile .py files [default]"),
+        (
+            'optimize=',
+            'O',
+            "also compile with optimization: -O1 for \"python -O\", "
+            "-O2 for \"python -OO\", and -O0 to disable [default: -O0]",
+        ),
+        ('force', 'f', "forcibly build everything (ignore file timestamps)"),
+    ]
+
+    boolean_options = ['compile', 'force']
+    negative_opt = {'no-compile': 'compile'}
+
+    def initialize_options(self):
+        self.build_lib = None
+        self.py_modules = None
+        self.package = None
+        self.package_data = None
+        self.package_dir = None
+        self.compile = False
+        self.optimize = 0
+        self.force = None
+
+    def finalize_options(self):
+        self.set_undefined_options(
+            'build', ('build_lib', 'build_lib'), ('force', 'force')
+        )
+
+        # Get the distribution options that are aliases for build_py
+        # options -- list of packages and list of modules.
+        self.packages = self.distribution.packages
+        self.py_modules = self.distribution.py_modules
+        self.package_data = self.distribution.package_data
+        self.package_dir = {}
+        if self.distribution.package_dir:
+            for name, path in self.distribution.package_dir.items():
+                self.package_dir[name] = convert_path(path)
+        self.data_files = self.get_data_files()
+
+        # Ick, copied straight from install_lib.py (fancy_getopt needs a
+        # type system!  Hell, *everything* needs a type system!!!)
+        if not isinstance(self.optimize, int):
+            try:
+                self.optimize = int(self.optimize)
+                assert 0 <= self.optimize <= 2
+            except (ValueError, AssertionError):
+                raise DistutilsOptionError("optimize must be 0, 1, or 2")
+
+    def run(self):
+        # XXX copy_file by default preserves atime and mtime.  IMHO this is
+        # the right thing to do, but perhaps it should be an option -- in
+        # particular, a site administrator might want installed files to
+        # reflect the time of installation rather than the last
+        # modification time before the installed release.
+
+        # XXX copy_file by default preserves mode, which appears to be the
+        # wrong thing to do: if a file is read-only in the working
+        # directory, we want it to be installed read/write so that the next
+        # installation of the same module distribution can overwrite it
+        # without problems.  (This might be a Unix-specific issue.)  Thus
+        # we turn off 'preserve_mode' when copying to the build directory,
+        # since the build directory is supposed to be exactly what the
+        # installation will look like (ie. we preserve mode when
+        # installing).
+
+        # Two options control which modules will be installed: 'packages'
+        # and 'py_modules'.  The former lets us work with whole packages, not
+        # specifying individual modules at all; the latter is for
+        # specifying modules one-at-a-time.
+
+        if self.py_modules:
+            self.build_modules()
+        if self.packages:
+            self.build_packages()
+            self.build_package_data()
+
+        self.byte_compile(self.get_outputs(include_bytecode=False))
+
+    def get_data_files(self):
+        """Generate list of '(package,src_dir,build_dir,filenames)' tuples"""
+        data = []
+        if not self.packages:
+            return data
+        for package in self.packages:
+            # Locate package source directory
+            src_dir = self.get_package_dir(package)
+
+            # Compute package build directory
+            build_dir = os.path.join(*([self.build_lib] + package.split('.')))
+
+            # Length of path to strip from found files
+            plen = 0
+            if src_dir:
+                plen = len(src_dir) + 1
+
+            # Strip directory from globbed filenames
+            filenames = [file[plen:] for file in self.find_data_files(package, src_dir)]
+            data.append((package, src_dir, build_dir, filenames))
+        return data
+
+    def find_data_files(self, package, src_dir):
+        """Return filenames for package's data files in 'src_dir'"""
+        globs = self.package_data.get('', []) + self.package_data.get(package, [])
+        files = []
+        for pattern in globs:
+            # Each pattern has to be converted to a platform-specific path
+            filelist = glob.glob(
+                os.path.join(glob.escape(src_dir), convert_path(pattern))
+            )
+            # Files that match more than one pattern are only added once
+            files.extend([
+                fn for fn in filelist if fn not in files and os.path.isfile(fn)
+            ])
+        return files
+
+    def build_package_data(self):
+        """Copy data files into build directory"""
+        for _package, src_dir, build_dir, filenames in self.data_files:
+            for filename in filenames:
+                target = os.path.join(build_dir, filename)
+                self.mkpath(os.path.dirname(target))
+                self.copy_file(
+                    os.path.join(src_dir, filename), target, preserve_mode=False
+                )
+
+    def get_package_dir(self, package):
+        """Return the directory, relative to the top of the source
+        distribution, where package 'package' should be found
+        (at least according to the 'package_dir' option, if any)."""
+        path = package.split('.')
+
+        if not self.package_dir:
+            if path:
+                return os.path.join(*path)
+            else:
+                return ''
+        else:
+            tail = []
+            while path:
+                try:
+                    pdir = self.package_dir['.'.join(path)]
+                except KeyError:
+                    tail.insert(0, path[-1])
+                    del path[-1]
+                else:
+                    tail.insert(0, pdir)
+                    return os.path.join(*tail)
+            else:
+                # Oops, got all the way through 'path' without finding a
+                # match in package_dir.  If package_dir defines a directory
+                # for the root (nameless) package, then fallback on it;
+                # otherwise, we might as well have not consulted
+                # package_dir at all, as we just use the directory implied
+                # by 'tail' (which should be the same as the original value
+                # of 'path' at this point).
+                pdir = self.package_dir.get('')
+                if pdir is not None:
+                    tail.insert(0, pdir)
+
+                if tail:
+                    return os.path.join(*tail)
+                else:
+                    return ''
+
+    def check_package(self, package, package_dir):
+        # Empty dir name means current directory, which we can probably
+        # assume exists.  Also, os.path.exists and isdir don't know about
+        # my "empty string means current dir" convention, so we have to
+        # circumvent them.
+        if package_dir != "":
+            if not os.path.exists(package_dir):
+                raise DistutilsFileError(
+                    f"package directory '{package_dir}' does not exist"
+                )
+            if not os.path.isdir(package_dir):
+                raise DistutilsFileError(
+                    f"supposed package directory '{package_dir}' exists, "
+                    "but is not a directory"
+                )
+
+        # Directories without __init__.py are namespace packages (PEP 420).
+        if package:
+            init_py = os.path.join(package_dir, "__init__.py")
+            if os.path.isfile(init_py):
+                return init_py
+
+        # Either not in a package at all (__init__.py not expected), or
+        # __init__.py doesn't exist -- so don't return the filename.
+        return None
+
+    def check_module(self, module, module_file):
+        if not os.path.isfile(module_file):
+            log.warning("file %s (for module %s) not found", module_file, module)
+            return False
+        else:
+            return True
+
+    def find_package_modules(self, package, package_dir):
+        self.check_package(package, package_dir)
+        module_files = glob.glob(os.path.join(glob.escape(package_dir), "*.py"))
+        modules = []
+        setup_script = os.path.abspath(self.distribution.script_name)
+
+        for f in module_files:
+            abs_f = os.path.abspath(f)
+            if abs_f != setup_script:
+                module = os.path.splitext(os.path.basename(f))[0]
+                modules.append((package, module, f))
+            else:
+                self.debug_print(f"excluding {setup_script}")
+        return modules
+
+    def find_modules(self):
+        """Finds individually-specified Python modules, ie. those listed by
+        module name in 'self.py_modules'.  Returns a list of tuples (package,
+        module_base, filename): 'package' is a tuple of the path through
+        package-space to the module; 'module_base' is the bare (no
+        packages, no dots) module name, and 'filename' is the path to the
+        ".py" file (relative to the distribution root) that implements the
+        module.
+        """
+        # Map package names to tuples of useful info about the package:
+        #    (package_dir, checked)
+        # package_dir - the directory where we'll find source files for
+        #   this package
+        # checked - true if we have checked that the package directory
+        #   is valid (exists, contains __init__.py, ... ?)
+        packages = {}
+
+        # List of (package, module, filename) tuples to return
+        modules = []
+
+        # We treat modules-in-packages almost the same as toplevel modules,
+        # just the "package" for a toplevel is empty (either an empty
+        # string or empty list, depending on context).  Differences:
+        #   - don't check for __init__.py in directory for empty package
+        for module in self.py_modules:
+            path = module.split('.')
+            package = '.'.join(path[0:-1])
+            module_base = path[-1]
+
+            try:
+                (package_dir, checked) = packages[package]
+            except KeyError:
+                package_dir = self.get_package_dir(package)
+                checked = False
+
+            if not checked:
+                init_py = self.check_package(package, package_dir)
+                packages[package] = (package_dir, 1)
+                if init_py:
+                    modules.append((package, "__init__", init_py))
+
+            # XXX perhaps we should also check for just .pyc files
+            # (so greedy closed-source bastards can distribute Python
+            # modules too)
+            module_file = os.path.join(package_dir, module_base + ".py")
+            if not self.check_module(module, module_file):
+                continue
+
+            modules.append((package, module_base, module_file))
+
+        return modules
+
+    def find_all_modules(self):
+        """Compute the list of all modules that will be built, whether
+        they are specified one-module-at-a-time ('self.py_modules') or
+        by whole packages ('self.packages').  Return a list of tuples
+        (package, module, module_file), just like 'find_modules()' and
+        'find_package_modules()' do."""
+        modules = []
+        if self.py_modules:
+            modules.extend(self.find_modules())
+        if self.packages:
+            for package in self.packages:
+                package_dir = self.get_package_dir(package)
+                m = self.find_package_modules(package, package_dir)
+                modules.extend(m)
+        return modules
+
+    def get_source_files(self):
+        return [module[-1] for module in self.find_all_modules()]
+
+    def get_module_outfile(self, build_dir, package, module):
+        outfile_path = [build_dir] + list(package) + [module + ".py"]
+        return os.path.join(*outfile_path)
+
+    def get_outputs(self, include_bytecode=True):
+        modules = self.find_all_modules()
+        outputs = []
+        for package, module, _module_file in modules:
+            package = package.split('.')
+            filename = self.get_module_outfile(self.build_lib, package, module)
+            outputs.append(filename)
+            if include_bytecode:
+                if self.compile:
+                    outputs.append(
+                        importlib.util.cache_from_source(filename, optimization='')
+                    )
+                if self.optimize > 0:
+                    outputs.append(
+                        importlib.util.cache_from_source(
+                            filename, optimization=self.optimize
+                        )
+                    )
+
+        outputs += [
+            os.path.join(build_dir, filename)
+            for package, src_dir, build_dir, filenames in self.data_files
+            for filename in filenames
+        ]
+
+        return outputs
+
+    def build_module(self, module, module_file, package):
+        if isinstance(package, str):
+            package = package.split('.')
+        elif not isinstance(package, (list, tuple)):
+            raise TypeError(
+                "'package' must be a string (dot-separated), list, or tuple"
+            )
+
+        # Now put the module source file into the "build" area -- this is
+        # easy, we just copy it somewhere under self.build_lib (the build
+        # directory for Python source).
+        outfile = self.get_module_outfile(self.build_lib, package, module)
+        dir = os.path.dirname(outfile)
+        self.mkpath(dir)
+        return self.copy_file(module_file, outfile, preserve_mode=False)
+
+    def build_modules(self):
+        modules = self.find_modules()
+        for package, module, module_file in modules:
+            # Now "build" the module -- ie. copy the source file to
+            # self.build_lib (the build directory for Python source).
+            # (Actually, it gets copied to the directory for this package
+            # under self.build_lib.)
+            self.build_module(module, module_file, package)
+
+    def build_packages(self):
+        for package in self.packages:
+            # Get list of (package, module, module_file) tuples based on
+            # scanning the package directory.  'package' is only included
+            # in the tuple so that 'find_modules()' and
+            # 'find_package_tuples()' have a consistent interface; it's
+            # ignored here (apart from a sanity check).  Also, 'module' is
+            # the *unqualified* module name (ie. no dots, no package -- we
+            # already know its package!), and 'module_file' is the path to
+            # the .py file, relative to the current directory
+            # (ie. including 'package_dir').
+            package_dir = self.get_package_dir(package)
+            modules = self.find_package_modules(package, package_dir)
+
+            # Now loop over the modules we found, "building" each one (just
+            # copy it to self.build_lib).
+            for package_, module, module_file in modules:
+                assert package == package_
+                self.build_module(module, module_file, package)
+
+    def byte_compile(self, files):
+        if sys.dont_write_bytecode:
+            self.warn('byte-compiling is disabled, skipping.')
+            return
+
+        from ..util import byte_compile
+
+        prefix = self.build_lib
+        if prefix[-1] != os.sep:
+            prefix = prefix + os.sep
+
+        # XXX this code is essentially the same as the 'byte_compile()
+        # method of the "install_lib" command, except for the determination
+        # of the 'prefix' string.  Hmmm.
+        if self.compile:
+            byte_compile(
+                files, optimize=0, force=self.force, prefix=prefix, dry_run=self.dry_run
+            )
+        if self.optimize > 0:
+            byte_compile(
+                files,
+                optimize=self.optimize,
+                force=self.force,
+                prefix=prefix,
+                dry_run=self.dry_run,
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_scripts.py
new file mode 100644
index 00000000..1c6fd3ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/build_scripts.py
@@ -0,0 +1,171 @@
+"""distutils.command.build_scripts
+
+Implements the Distutils 'build_scripts' command."""
+
+import os
+import re
+import tokenize
+from distutils import sysconfig
+from distutils._log import log
+from stat import ST_MODE
+from typing import ClassVar
+
+from .._modified import newer
+from ..core import Command
+from ..util import convert_path
+
+shebang_pattern = re.compile('^#!.*python[0-9.]*([ \t].*)?$')
+"""
+Pattern matching a Python interpreter indicated in first line of a script.
+"""
+
+# for Setuptools compatibility
+first_line_re = shebang_pattern
+
+
+class build_scripts(Command):
+    description = "\"build\" scripts (copy and fixup #! line)"
+
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ('build-dir=', 'd', "directory to \"build\" (copy) to"),
+        ('force', 'f', "forcibly build everything (ignore file timestamps"),
+        ('executable=', 'e', "specify final destination interpreter path"),
+    ]
+
+    boolean_options = ['force']
+
+    def initialize_options(self):
+        self.build_dir = None
+        self.scripts = None
+        self.force = None
+        self.executable = None
+
+    def finalize_options(self):
+        self.set_undefined_options(
+            'build',
+            ('build_scripts', 'build_dir'),
+            ('force', 'force'),
+            ('executable', 'executable'),
+        )
+        self.scripts = self.distribution.scripts
+
+    def get_source_files(self):
+        return self.scripts
+
+    def run(self):
+        if not self.scripts:
+            return
+        self.copy_scripts()
+
+    def copy_scripts(self):
+        """
+        Copy each script listed in ``self.scripts``.
+
+        If a script is marked as a Python script (first line matches
+        'shebang_pattern', i.e. starts with ``#!`` and contains
+        "python"), then adjust in the copy the first line to refer to
+        the current Python interpreter.
+        """
+        self.mkpath(self.build_dir)
+        outfiles = []
+        updated_files = []
+        for script in self.scripts:
+            self._copy_script(script, outfiles, updated_files)
+
+        self._change_modes(outfiles)
+
+        return outfiles, updated_files
+
+    def _copy_script(self, script, outfiles, updated_files):  # noqa: C901
+        shebang_match = None
+        script = convert_path(script)
+        outfile = os.path.join(self.build_dir, os.path.basename(script))
+        outfiles.append(outfile)
+
+        if not self.force and not newer(script, outfile):
+            log.debug("not copying %s (up-to-date)", script)
+            return
+
+        # Always open the file, but ignore failures in dry-run mode
+        # in order to attempt to copy directly.
+        try:
+            f = tokenize.open(script)
+        except OSError:
+            if not self.dry_run:
+                raise
+            f = None
+        else:
+            first_line = f.readline()
+            if not first_line:
+                self.warn(f"{script} is an empty file (skipping)")
+                return
+
+            shebang_match = shebang_pattern.match(first_line)
+
+        updated_files.append(outfile)
+        if shebang_match:
+            log.info("copying and adjusting %s -> %s", script, self.build_dir)
+            if not self.dry_run:
+                if not sysconfig.python_build:
+                    executable = self.executable
+                else:
+                    executable = os.path.join(
+                        sysconfig.get_config_var("BINDIR"),
+                        "python{}{}".format(
+                            sysconfig.get_config_var("VERSION"),
+                            sysconfig.get_config_var("EXE"),
+                        ),
+                    )
+                post_interp = shebang_match.group(1) or ''
+                shebang = "#!" + executable + post_interp + "\n"
+                self._validate_shebang(shebang, f.encoding)
+                with open(outfile, "w", encoding=f.encoding) as outf:
+                    outf.write(shebang)
+                    outf.writelines(f.readlines())
+            if f:
+                f.close()
+        else:
+            if f:
+                f.close()
+            self.copy_file(script, outfile)
+
+    def _change_modes(self, outfiles):
+        if os.name != 'posix':
+            return
+
+        for file in outfiles:
+            self._change_mode(file)
+
+    def _change_mode(self, file):
+        if self.dry_run:
+            log.info("changing mode of %s", file)
+            return
+
+        oldmode = os.stat(file)[ST_MODE] & 0o7777
+        newmode = (oldmode | 0o555) & 0o7777
+        if newmode != oldmode:
+            log.info("changing mode of %s from %o to %o", file, oldmode, newmode)
+            os.chmod(file, newmode)
+
+    @staticmethod
+    def _validate_shebang(shebang, encoding):
+        # Python parser starts to read a script using UTF-8 until
+        # it gets a #coding:xxx cookie. The shebang has to be the
+        # first line of a file, the #coding:xxx cookie cannot be
+        # written before. So the shebang has to be encodable to
+        # UTF-8.
+        try:
+            shebang.encode('utf-8')
+        except UnicodeEncodeError:
+            raise ValueError(f"The shebang ({shebang!r}) is not encodable to utf-8")
+
+        # If the script is encoded to a custom encoding (use a
+        # #coding:xxx cookie), the shebang has to be encodable to
+        # the script encoding too.
+        try:
+            shebang.encode(encoding)
+        except UnicodeEncodeError:
+            raise ValueError(
+                f"The shebang ({shebang!r}) is not encodable "
+                f"to the script encoding ({encoding})"
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/check.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/check.py
new file mode 100644
index 00000000..078c1ce8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/check.py
@@ -0,0 +1,152 @@
+"""distutils.command.check
+
+Implements the Distutils 'check' command.
+"""
+
+import contextlib
+from typing import ClassVar
+
+from ..core import Command
+from ..errors import DistutilsSetupError
+
+with contextlib.suppress(ImportError):
+    import docutils.frontend
+    import docutils.nodes
+    import docutils.parsers.rst
+    import docutils.utils
+
+    class SilentReporter(docutils.utils.Reporter):
+        def __init__(
+            self,
+            source,
+            report_level,
+            halt_level,
+            stream=None,
+            debug=False,
+            encoding='ascii',
+            error_handler='replace',
+        ):
+            self.messages = []
+            super().__init__(
+                source, report_level, halt_level, stream, debug, encoding, error_handler
+            )
+
+        def system_message(self, level, message, *children, **kwargs):
+            self.messages.append((level, message, children, kwargs))
+            return docutils.nodes.system_message(
+                message, *children, level=level, type=self.levels[level], **kwargs
+            )
+
+
+class check(Command):
+    """This command checks the meta-data of the package."""
+
+    description = "perform some checks on the package"
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ('metadata', 'm', 'Verify meta-data'),
+        (
+            'restructuredtext',
+            'r',
+            'Checks if long string meta-data syntax are reStructuredText-compliant',
+        ),
+        ('strict', 's', 'Will exit with an error if a check fails'),
+    ]
+
+    boolean_options = ['metadata', 'restructuredtext', 'strict']
+
+    def initialize_options(self):
+        """Sets default values for options."""
+        self.restructuredtext = False
+        self.metadata = 1
+        self.strict = False
+        self._warnings = 0
+
+    def finalize_options(self):
+        pass
+
+    def warn(self, msg):
+        """Counts the number of warnings that occurs."""
+        self._warnings += 1
+        return Command.warn(self, msg)
+
+    def run(self):
+        """Runs the command."""
+        # perform the various tests
+        if self.metadata:
+            self.check_metadata()
+        if self.restructuredtext:
+            if 'docutils' in globals():
+                try:
+                    self.check_restructuredtext()
+                except TypeError as exc:
+                    raise DistutilsSetupError(str(exc))
+            elif self.strict:
+                raise DistutilsSetupError('The docutils package is needed.')
+
+        # let's raise an error in strict mode, if we have at least
+        # one warning
+        if self.strict and self._warnings > 0:
+            raise DistutilsSetupError('Please correct your package.')
+
+    def check_metadata(self):
+        """Ensures that all required elements of meta-data are supplied.
+
+        Required fields:
+            name, version
+
+        Warns if any are missing.
+        """
+        metadata = self.distribution.metadata
+
+        missing = [
+            attr for attr in ('name', 'version') if not getattr(metadata, attr, None)
+        ]
+
+        if missing:
+            self.warn("missing required meta-data: {}".format(', '.join(missing)))
+
+    def check_restructuredtext(self):
+        """Checks if the long string fields are reST-compliant."""
+        data = self.distribution.get_long_description()
+        for warning in self._check_rst_data(data):
+            line = warning[-1].get('line')
+            if line is None:
+                warning = warning[1]
+            else:
+                warning = f'{warning[1]} (line {line})'
+            self.warn(warning)
+
+    def _check_rst_data(self, data):
+        """Returns warnings when the provided data doesn't compile."""
+        # the include and csv_table directives need this to be a path
+        source_path = self.distribution.script_name or 'setup.py'
+        parser = docutils.parsers.rst.Parser()
+        settings = docutils.frontend.OptionParser(
+            components=(docutils.parsers.rst.Parser,)
+        ).get_default_values()
+        settings.tab_width = 4
+        settings.pep_references = None
+        settings.rfc_references = None
+        reporter = SilentReporter(
+            source_path,
+            settings.report_level,
+            settings.halt_level,
+            stream=settings.warning_stream,
+            debug=settings.debug,
+            encoding=settings.error_encoding,
+            error_handler=settings.error_encoding_error_handler,
+        )
+
+        document = docutils.nodes.document(settings, reporter, source=source_path)
+        document.note_source(source_path, -1)
+        try:
+            parser.parse(data, document)
+        except AttributeError as e:
+            reporter.messages.append((
+                -1,
+                f'Could not finish the parsing: {e}.',
+                '',
+                {},
+            ))
+
+        return reporter.messages
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/clean.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/clean.py
new file mode 100644
index 00000000..fb54a60e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/clean.py
@@ -0,0 +1,76 @@
+"""distutils.command.clean
+
+Implements the Distutils 'clean' command."""
+
+# contributed by Bastian Kleineidam <calvin@cs.uni-sb.de>, added 2000-03-18
+
+import os
+from distutils._log import log
+
+from ..core import Command
+from ..dir_util import remove_tree
+
+
+class clean(Command):
+    description = "clean up temporary files from 'build' command"
+    user_options = [
+        ('build-base=', 'b', "base build directory [default: 'build.build-base']"),
+        (
+            'build-lib=',
+            None,
+            "build directory for all modules [default: 'build.build-lib']",
+        ),
+        ('build-temp=', 't', "temporary build directory [default: 'build.build-temp']"),
+        (
+            'build-scripts=',
+            None,
+            "build directory for scripts [default: 'build.build-scripts']",
+        ),
+        ('bdist-base=', None, "temporary directory for built distributions"),
+        ('all', 'a', "remove all build output, not just temporary by-products"),
+    ]
+
+    boolean_options = ['all']
+
+    def initialize_options(self):
+        self.build_base = None
+        self.build_lib = None
+        self.build_temp = None
+        self.build_scripts = None
+        self.bdist_base = None
+        self.all = None
+
+    def finalize_options(self):
+        self.set_undefined_options(
+            'build',
+            ('build_base', 'build_base'),
+            ('build_lib', 'build_lib'),
+            ('build_scripts', 'build_scripts'),
+            ('build_temp', 'build_temp'),
+        )
+        self.set_undefined_options('bdist', ('bdist_base', 'bdist_base'))
+
+    def run(self):
+        # remove the build/temp.<plat> directory (unless it's already
+        # gone)
+        if os.path.exists(self.build_temp):
+            remove_tree(self.build_temp, dry_run=self.dry_run)
+        else:
+            log.debug("'%s' does not exist -- can't clean it", self.build_temp)
+
+        if self.all:
+            # remove build directories
+            for directory in (self.build_lib, self.bdist_base, self.build_scripts):
+                if os.path.exists(directory):
+                    remove_tree(directory, dry_run=self.dry_run)
+                else:
+                    log.warning("'%s' does not exist -- can't clean it", directory)
+
+        # just for the heck of it, try to remove the base build directory:
+        # we might have emptied it right now, but if not we don't care
+        if not self.dry_run:
+            try:
+                os.rmdir(self.build_base)
+                log.info("removing '%s'", self.build_base)
+            except OSError:
+                pass
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/config.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/config.py
new file mode 100644
index 00000000..fe83c292
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/config.py
@@ -0,0 +1,369 @@
+"""distutils.command.config
+
+Implements the Distutils 'config' command, a (mostly) empty command class
+that exists mainly to be sub-classed by specific module distributions and
+applications.  The idea is that while every "config" command is different,
+at least they're all named the same, and users always see "config" in the
+list of standard commands.  Also, this is a good place to put common
+configure-like tasks: "try to compile this C code", or "figure out where
+this header file lives".
+"""
+
+from __future__ import annotations
+
+import os
+import pathlib
+import re
+from collections.abc import Sequence
+from distutils._log import log
+
+from ..core import Command
+from ..errors import DistutilsExecError
+from ..sysconfig import customize_compiler
+
+LANG_EXT = {"c": ".c", "c++": ".cxx"}
+
+
+class config(Command):
+    description = "prepare to build"
+
+    user_options = [
+        ('compiler=', None, "specify the compiler type"),
+        ('cc=', None, "specify the compiler executable"),
+        ('include-dirs=', 'I', "list of directories to search for header files"),
+        ('define=', 'D', "C preprocessor macros to define"),
+        ('undef=', 'U', "C preprocessor macros to undefine"),
+        ('libraries=', 'l', "external C libraries to link with"),
+        ('library-dirs=', 'L', "directories to search for external C libraries"),
+        ('noisy', None, "show every action (compile, link, run, ...) taken"),
+        (
+            'dump-source',
+            None,
+            "dump generated source files before attempting to compile them",
+        ),
+    ]
+
+    # The three standard command methods: since the "config" command
+    # does nothing by default, these are empty.
+
+    def initialize_options(self):
+        self.compiler = None
+        self.cc = None
+        self.include_dirs = None
+        self.libraries = None
+        self.library_dirs = None
+
+        # maximal output for now
+        self.noisy = 1
+        self.dump_source = 1
+
+        # list of temporary files generated along-the-way that we have
+        # to clean at some point
+        self.temp_files = []
+
+    def finalize_options(self):
+        if self.include_dirs is None:
+            self.include_dirs = self.distribution.include_dirs or []
+        elif isinstance(self.include_dirs, str):
+            self.include_dirs = self.include_dirs.split(os.pathsep)
+
+        if self.libraries is None:
+            self.libraries = []
+        elif isinstance(self.libraries, str):
+            self.libraries = [self.libraries]
+
+        if self.library_dirs is None:
+            self.library_dirs = []
+        elif isinstance(self.library_dirs, str):
+            self.library_dirs = self.library_dirs.split(os.pathsep)
+
+    def run(self):
+        pass
+
+    # Utility methods for actual "config" commands.  The interfaces are
+    # loosely based on Autoconf macros of similar names.  Sub-classes
+    # may use these freely.
+
+    def _check_compiler(self):
+        """Check that 'self.compiler' really is a CCompiler object;
+        if not, make it one.
+        """
+        # We do this late, and only on-demand, because this is an expensive
+        # import.
+        from ..ccompiler import CCompiler, new_compiler
+
+        if not isinstance(self.compiler, CCompiler):
+            self.compiler = new_compiler(
+                compiler=self.compiler, dry_run=self.dry_run, force=True
+            )
+            customize_compiler(self.compiler)
+            if self.include_dirs:
+                self.compiler.set_include_dirs(self.include_dirs)
+            if self.libraries:
+                self.compiler.set_libraries(self.libraries)
+            if self.library_dirs:
+                self.compiler.set_library_dirs(self.library_dirs)
+
+    def _gen_temp_sourcefile(self, body, headers, lang):
+        filename = "_configtest" + LANG_EXT[lang]
+        with open(filename, "w", encoding='utf-8') as file:
+            if headers:
+                for header in headers:
+                    file.write(f"#include <{header}>\n")
+                file.write("\n")
+            file.write(body)
+            if body[-1] != "\n":
+                file.write("\n")
+        return filename
+
+    def _preprocess(self, body, headers, include_dirs, lang):
+        src = self._gen_temp_sourcefile(body, headers, lang)
+        out = "_configtest.i"
+        self.temp_files.extend([src, out])
+        self.compiler.preprocess(src, out, include_dirs=include_dirs)
+        return (src, out)
+
+    def _compile(self, body, headers, include_dirs, lang):
+        src = self._gen_temp_sourcefile(body, headers, lang)
+        if self.dump_source:
+            dump_file(src, f"compiling '{src}':")
+        (obj,) = self.compiler.object_filenames([src])
+        self.temp_files.extend([src, obj])
+        self.compiler.compile([src], include_dirs=include_dirs)
+        return (src, obj)
+
+    def _link(self, body, headers, include_dirs, libraries, library_dirs, lang):
+        (src, obj) = self._compile(body, headers, include_dirs, lang)
+        prog = os.path.splitext(os.path.basename(src))[0]
+        self.compiler.link_executable(
+            [obj],
+            prog,
+            libraries=libraries,
+            library_dirs=library_dirs,
+            target_lang=lang,
+        )
+
+        if self.compiler.exe_extension is not None:
+            prog = prog + self.compiler.exe_extension
+        self.temp_files.append(prog)
+
+        return (src, obj, prog)
+
+    def _clean(self, *filenames):
+        if not filenames:
+            filenames = self.temp_files
+            self.temp_files = []
+        log.info("removing: %s", ' '.join(filenames))
+        for filename in filenames:
+            try:
+                os.remove(filename)
+            except OSError:
+                pass
+
+    # XXX these ignore the dry-run flag: what to do, what to do? even if
+    # you want a dry-run build, you still need some sort of configuration
+    # info.  My inclination is to make it up to the real config command to
+    # consult 'dry_run', and assume a default (minimal) configuration if
+    # true.  The problem with trying to do it here is that you'd have to
+    # return either true or false from all the 'try' methods, neither of
+    # which is correct.
+
+    # XXX need access to the header search path and maybe default macros.
+
+    def try_cpp(self, body=None, headers=None, include_dirs=None, lang="c"):
+        """Construct a source file from 'body' (a string containing lines
+        of C/C++ code) and 'headers' (a list of header files to include)
+        and run it through the preprocessor.  Return true if the
+        preprocessor succeeded, false if there were any errors.
+        ('body' probably isn't of much use, but what the heck.)
+        """
+        from ..ccompiler import CompileError
+
+        self._check_compiler()
+        ok = True
+        try:
+            self._preprocess(body, headers, include_dirs, lang)
+        except CompileError:
+            ok = False
+
+        self._clean()
+        return ok
+
+    def search_cpp(self, pattern, body=None, headers=None, include_dirs=None, lang="c"):
+        """Construct a source file (just like 'try_cpp()'), run it through
+        the preprocessor, and return true if any line of the output matches
+        'pattern'.  'pattern' should either be a compiled regex object or a
+        string containing a regex.  If both 'body' and 'headers' are None,
+        preprocesses an empty file -- which can be useful to determine the
+        symbols the preprocessor and compiler set by default.
+        """
+        self._check_compiler()
+        src, out = self._preprocess(body, headers, include_dirs, lang)
+
+        if isinstance(pattern, str):
+            pattern = re.compile(pattern)
+
+        with open(out, encoding='utf-8') as file:
+            match = any(pattern.search(line) for line in file)
+
+        self._clean()
+        return match
+
+    def try_compile(self, body, headers=None, include_dirs=None, lang="c"):
+        """Try to compile a source file built from 'body' and 'headers'.
+        Return true on success, false otherwise.
+        """
+        from ..ccompiler import CompileError
+
+        self._check_compiler()
+        try:
+            self._compile(body, headers, include_dirs, lang)
+            ok = True
+        except CompileError:
+            ok = False
+
+        log.info(ok and "success!" or "failure.")
+        self._clean()
+        return ok
+
+    def try_link(
+        self,
+        body,
+        headers=None,
+        include_dirs=None,
+        libraries=None,
+        library_dirs=None,
+        lang="c",
+    ):
+        """Try to compile and link a source file, built from 'body' and
+        'headers', to executable form.  Return true on success, false
+        otherwise.
+        """
+        from ..ccompiler import CompileError, LinkError
+
+        self._check_compiler()
+        try:
+            self._link(body, headers, include_dirs, libraries, library_dirs, lang)
+            ok = True
+        except (CompileError, LinkError):
+            ok = False
+
+        log.info(ok and "success!" or "failure.")
+        self._clean()
+        return ok
+
+    def try_run(
+        self,
+        body,
+        headers=None,
+        include_dirs=None,
+        libraries=None,
+        library_dirs=None,
+        lang="c",
+    ):
+        """Try to compile, link to an executable, and run a program
+        built from 'body' and 'headers'.  Return true on success, false
+        otherwise.
+        """
+        from ..ccompiler import CompileError, LinkError
+
+        self._check_compiler()
+        try:
+            src, obj, exe = self._link(
+                body, headers, include_dirs, libraries, library_dirs, lang
+            )
+            self.spawn([exe])
+            ok = True
+        except (CompileError, LinkError, DistutilsExecError):
+            ok = False
+
+        log.info(ok and "success!" or "failure.")
+        self._clean()
+        return ok
+
+    # -- High-level methods --------------------------------------------
+    # (these are the ones that are actually likely to be useful
+    # when implementing a real-world config command!)
+
+    def check_func(
+        self,
+        func,
+        headers=None,
+        include_dirs=None,
+        libraries=None,
+        library_dirs=None,
+        decl=False,
+        call=False,
+    ):
+        """Determine if function 'func' is available by constructing a
+        source file that refers to 'func', and compiles and links it.
+        If everything succeeds, returns true; otherwise returns false.
+
+        The constructed source file starts out by including the header
+        files listed in 'headers'.  If 'decl' is true, it then declares
+        'func' (as "int func()"); you probably shouldn't supply 'headers'
+        and set 'decl' true in the same call, or you might get errors about
+        a conflicting declarations for 'func'.  Finally, the constructed
+        'main()' function either references 'func' or (if 'call' is true)
+        calls it.  'libraries' and 'library_dirs' are used when
+        linking.
+        """
+        self._check_compiler()
+        body = []
+        if decl:
+            body.append(f"int {func} ();")
+        body.append("int main () {")
+        if call:
+            body.append(f"  {func}();")
+        else:
+            body.append(f"  {func};")
+        body.append("}")
+        body = "\n".join(body) + "\n"
+
+        return self.try_link(body, headers, include_dirs, libraries, library_dirs)
+
+    def check_lib(
+        self,
+        library,
+        library_dirs=None,
+        headers=None,
+        include_dirs=None,
+        other_libraries: Sequence[str] = [],
+    ):
+        """Determine if 'library' is available to be linked against,
+        without actually checking that any particular symbols are provided
+        by it.  'headers' will be used in constructing the source file to
+        be compiled, but the only effect of this is to check if all the
+        header files listed are available.  Any libraries listed in
+        'other_libraries' will be included in the link, in case 'library'
+        has symbols that depend on other libraries.
+        """
+        self._check_compiler()
+        return self.try_link(
+            "int main (void) { }",
+            headers,
+            include_dirs,
+            [library] + list(other_libraries),
+            library_dirs,
+        )
+
+    def check_header(self, header, include_dirs=None, library_dirs=None, lang="c"):
+        """Determine if the system header file named by 'header_file'
+        exists and can be found by the preprocessor; return true if so,
+        false otherwise.
+        """
+        return self.try_cpp(
+            body="/* No body */", headers=[header], include_dirs=include_dirs
+        )
+
+
+def dump_file(filename, head=None):
+    """Dumps a file content into log.info.
+
+    If head is not None, will be dumped before the file content.
+    """
+    if head is None:
+        log.info('%s', filename)
+    else:
+        log.info(head)
+    log.info(pathlib.Path(filename).read_text(encoding='utf-8'))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install.py
new file mode 100644
index 00000000..94009950
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install.py
@@ -0,0 +1,811 @@
+"""distutils.command.install
+
+Implements the Distutils 'install' command."""
+
+import contextlib
+import itertools
+import os
+import sys
+import sysconfig
+from distutils._log import log
+from site import USER_BASE, USER_SITE
+
+import jaraco.collections
+
+from ..core import Command
+from ..debug import DEBUG
+from ..errors import DistutilsOptionError, DistutilsPlatformError
+from ..file_util import write_file
+from ..sysconfig import get_config_vars
+from ..util import change_root, convert_path, get_platform, subst_vars
+from . import _framework_compat as fw
+
+HAS_USER_SITE = True
+
+WINDOWS_SCHEME = {
+    'purelib': '{base}/Lib/site-packages',
+    'platlib': '{base}/Lib/site-packages',
+    'headers': '{base}/Include/{dist_name}',
+    'scripts': '{base}/Scripts',
+    'data': '{base}',
+}
+
+INSTALL_SCHEMES = {
+    'posix_prefix': {
+        'purelib': '{base}/lib/{implementation_lower}{py_version_short}/site-packages',
+        'platlib': '{platbase}/{platlibdir}/{implementation_lower}'
+        '{py_version_short}/site-packages',
+        'headers': '{base}/include/{implementation_lower}'
+        '{py_version_short}{abiflags}/{dist_name}',
+        'scripts': '{base}/bin',
+        'data': '{base}',
+    },
+    'posix_home': {
+        'purelib': '{base}/lib/{implementation_lower}',
+        'platlib': '{base}/{platlibdir}/{implementation_lower}',
+        'headers': '{base}/include/{implementation_lower}/{dist_name}',
+        'scripts': '{base}/bin',
+        'data': '{base}',
+    },
+    'nt': WINDOWS_SCHEME,
+    'pypy': {
+        'purelib': '{base}/site-packages',
+        'platlib': '{base}/site-packages',
+        'headers': '{base}/include/{dist_name}',
+        'scripts': '{base}/bin',
+        'data': '{base}',
+    },
+    'pypy_nt': {
+        'purelib': '{base}/site-packages',
+        'platlib': '{base}/site-packages',
+        'headers': '{base}/include/{dist_name}',
+        'scripts': '{base}/Scripts',
+        'data': '{base}',
+    },
+}
+
+# user site schemes
+if HAS_USER_SITE:
+    INSTALL_SCHEMES['nt_user'] = {
+        'purelib': '{usersite}',
+        'platlib': '{usersite}',
+        'headers': '{userbase}/{implementation}{py_version_nodot_plat}'
+        '/Include/{dist_name}',
+        'scripts': '{userbase}/{implementation}{py_version_nodot_plat}/Scripts',
+        'data': '{userbase}',
+    }
+
+    INSTALL_SCHEMES['posix_user'] = {
+        'purelib': '{usersite}',
+        'platlib': '{usersite}',
+        'headers': '{userbase}/include/{implementation_lower}'
+        '{py_version_short}{abiflags}/{dist_name}',
+        'scripts': '{userbase}/bin',
+        'data': '{userbase}',
+    }
+
+
+INSTALL_SCHEMES.update(fw.schemes)
+
+
+# The keys to an installation scheme; if any new types of files are to be
+# installed, be sure to add an entry to every installation scheme above,
+# and to SCHEME_KEYS here.
+SCHEME_KEYS = ('purelib', 'platlib', 'headers', 'scripts', 'data')
+
+
+def _load_sysconfig_schemes():
+    with contextlib.suppress(AttributeError):
+        return {
+            scheme: sysconfig.get_paths(scheme, expand=False)
+            for scheme in sysconfig.get_scheme_names()
+        }
+
+
+def _load_schemes():
+    """
+    Extend default schemes with schemes from sysconfig.
+    """
+
+    sysconfig_schemes = _load_sysconfig_schemes() or {}
+
+    return {
+        scheme: {
+            **INSTALL_SCHEMES.get(scheme, {}),
+            **sysconfig_schemes.get(scheme, {}),
+        }
+        for scheme in set(itertools.chain(INSTALL_SCHEMES, sysconfig_schemes))
+    }
+
+
+def _get_implementation():
+    if hasattr(sys, 'pypy_version_info'):
+        return 'PyPy'
+    else:
+        return 'Python'
+
+
+def _select_scheme(ob, name):
+    scheme = _inject_headers(name, _load_scheme(_resolve_scheme(name)))
+    vars(ob).update(_remove_set(ob, _scheme_attrs(scheme)))
+
+
+def _remove_set(ob, attrs):
+    """
+    Include only attrs that are None in ob.
+    """
+    return {key: value for key, value in attrs.items() if getattr(ob, key) is None}
+
+
+def _resolve_scheme(name):
+    os_name, sep, key = name.partition('_')
+    try:
+        resolved = sysconfig.get_preferred_scheme(key)
+    except Exception:
+        resolved = fw.scheme(_pypy_hack(name))
+    return resolved
+
+
+def _load_scheme(name):
+    return _load_schemes()[name]
+
+
+def _inject_headers(name, scheme):
+    """
+    Given a scheme name and the resolved scheme,
+    if the scheme does not include headers, resolve
+    the fallback scheme for the name and use headers
+    from it. pypa/distutils#88
+    """
+    # Bypass the preferred scheme, which may not
+    # have defined headers.
+    fallback = _load_scheme(_pypy_hack(name))
+    scheme.setdefault('headers', fallback['headers'])
+    return scheme
+
+
+def _scheme_attrs(scheme):
+    """Resolve install directories by applying the install schemes."""
+    return {f'install_{key}': scheme[key] for key in SCHEME_KEYS}
+
+
+def _pypy_hack(name):
+    PY37 = sys.version_info < (3, 8)
+    old_pypy = hasattr(sys, 'pypy_version_info') and PY37
+    prefix = not name.endswith(('_user', '_home'))
+    pypy_name = 'pypy' + '_nt' * (os.name == 'nt')
+    return pypy_name if old_pypy and prefix else name
+
+
+class install(Command):
+    description = "install everything from build directory"
+
+    user_options = [
+        # Select installation scheme and set base director(y|ies)
+        ('prefix=', None, "installation prefix"),
+        ('exec-prefix=', None, "(Unix only) prefix for platform-specific files"),
+        ('home=', None, "(Unix only) home directory to install under"),
+        # Or, just set the base director(y|ies)
+        (
+            'install-base=',
+            None,
+            "base installation directory (instead of --prefix or --home)",
+        ),
+        (
+            'install-platbase=',
+            None,
+            "base installation directory for platform-specific files (instead of --exec-prefix or --home)",
+        ),
+        ('root=', None, "install everything relative to this alternate root directory"),
+        # Or, explicitly set the installation scheme
+        (
+            'install-purelib=',
+            None,
+            "installation directory for pure Python module distributions",
+        ),
+        (
+            'install-platlib=',
+            None,
+            "installation directory for non-pure module distributions",
+        ),
+        (
+            'install-lib=',
+            None,
+            "installation directory for all module distributions (overrides --install-purelib and --install-platlib)",
+        ),
+        ('install-headers=', None, "installation directory for C/C++ headers"),
+        ('install-scripts=', None, "installation directory for Python scripts"),
+        ('install-data=', None, "installation directory for data files"),
+        # Byte-compilation options -- see install_lib.py for details, as
+        # these are duplicated from there (but only install_lib does
+        # anything with them).
+        ('compile', 'c', "compile .py to .pyc [default]"),
+        ('no-compile', None, "don't compile .py files"),
+        (
+            'optimize=',
+            'O',
+            "also compile with optimization: -O1 for \"python -O\", "
+            "-O2 for \"python -OO\", and -O0 to disable [default: -O0]",
+        ),
+        # Miscellaneous control options
+        ('force', 'f', "force installation (overwrite any existing files)"),
+        ('skip-build', None, "skip rebuilding everything (for testing/debugging)"),
+        # Where to install documentation (eventually!)
+        # ('doc-format=', None, "format of documentation to generate"),
+        # ('install-man=', None, "directory for Unix man pages"),
+        # ('install-html=', None, "directory for HTML documentation"),
+        # ('install-info=', None, "directory for GNU info files"),
+        ('record=', None, "filename in which to record list of installed files"),
+    ]
+
+    boolean_options = ['compile', 'force', 'skip-build']
+
+    if HAS_USER_SITE:
+        user_options.append((
+            'user',
+            None,
+            f"install in user site-package '{USER_SITE}'",
+        ))
+        boolean_options.append('user')
+
+    negative_opt = {'no-compile': 'compile'}
+
+    def initialize_options(self):
+        """Initializes options."""
+        # High-level options: these select both an installation base
+        # and scheme.
+        self.prefix = None
+        self.exec_prefix = None
+        self.home = None
+        self.user = False
+
+        # These select only the installation base; it's up to the user to
+        # specify the installation scheme (currently, that means supplying
+        # the --install-{platlib,purelib,scripts,data} options).
+        self.install_base = None
+        self.install_platbase = None
+        self.root = None
+
+        # These options are the actual installation directories; if not
+        # supplied by the user, they are filled in using the installation
+        # scheme implied by prefix/exec-prefix/home and the contents of
+        # that installation scheme.
+        self.install_purelib = None  # for pure module distributions
+        self.install_platlib = None  # non-pure (dists w/ extensions)
+        self.install_headers = None  # for C/C++ headers
+        self.install_lib = None  # set to either purelib or platlib
+        self.install_scripts = None
+        self.install_data = None
+        self.install_userbase = USER_BASE
+        self.install_usersite = USER_SITE
+
+        self.compile = None
+        self.optimize = None
+
+        # Deprecated
+        # These two are for putting non-packagized distributions into their
+        # own directory and creating a .pth file if it makes sense.
+        # 'extra_path' comes from the setup file; 'install_path_file' can
+        # be turned off if it makes no sense to install a .pth file.  (But
+        # better to install it uselessly than to guess wrong and not
+        # install it when it's necessary and would be used!)  Currently,
+        # 'install_path_file' is always true unless some outsider meddles
+        # with it.
+        self.extra_path = None
+        self.install_path_file = True
+
+        # 'force' forces installation, even if target files are not
+        # out-of-date.  'skip_build' skips running the "build" command,
+        # handy if you know it's not necessary.  'warn_dir' (which is *not*
+        # a user option, it's just there so the bdist_* commands can turn
+        # it off) determines whether we warn about installing to a
+        # directory not in sys.path.
+        self.force = False
+        self.skip_build = False
+        self.warn_dir = True
+
+        # These are only here as a conduit from the 'build' command to the
+        # 'install_*' commands that do the real work.  ('build_base' isn't
+        # actually used anywhere, but it might be useful in future.)  They
+        # are not user options, because if the user told the install
+        # command where the build directory is, that wouldn't affect the
+        # build command.
+        self.build_base = None
+        self.build_lib = None
+
+        # Not defined yet because we don't know anything about
+        # documentation yet.
+        # self.install_man = None
+        # self.install_html = None
+        # self.install_info = None
+
+        self.record = None
+
+    # -- Option finalizing methods -------------------------------------
+    # (This is rather more involved than for most commands,
+    # because this is where the policy for installing third-
+    # party Python modules on various platforms given a wide
+    # array of user input is decided.  Yes, it's quite complex!)
+
+    def finalize_options(self):  # noqa: C901
+        """Finalizes options."""
+        # This method (and its helpers, like 'finalize_unix()',
+        # 'finalize_other()', and 'select_scheme()') is where the default
+        # installation directories for modules, extension modules, and
+        # anything else we care to install from a Python module
+        # distribution.  Thus, this code makes a pretty important policy
+        # statement about how third-party stuff is added to a Python
+        # installation!  Note that the actual work of installation is done
+        # by the relatively simple 'install_*' commands; they just take
+        # their orders from the installation directory options determined
+        # here.
+
+        # Check for errors/inconsistencies in the options; first, stuff
+        # that's wrong on any platform.
+
+        if (self.prefix or self.exec_prefix or self.home) and (
+            self.install_base or self.install_platbase
+        ):
+            raise DistutilsOptionError(
+                "must supply either prefix/exec-prefix/home or install-base/install-platbase -- not both"
+            )
+
+        if self.home and (self.prefix or self.exec_prefix):
+            raise DistutilsOptionError(
+                "must supply either home or prefix/exec-prefix -- not both"
+            )
+
+        if self.user and (
+            self.prefix
+            or self.exec_prefix
+            or self.home
+            or self.install_base
+            or self.install_platbase
+        ):
+            raise DistutilsOptionError(
+                "can't combine user with prefix, "
+                "exec_prefix/home, or install_(plat)base"
+            )
+
+        # Next, stuff that's wrong (or dubious) only on certain platforms.
+        if os.name != "posix":
+            if self.exec_prefix:
+                self.warn("exec-prefix option ignored on this platform")
+                self.exec_prefix = None
+
+        # Now the interesting logic -- so interesting that we farm it out
+        # to other methods.  The goal of these methods is to set the final
+        # values for the install_{lib,scripts,data,...}  options, using as
+        # input a heady brew of prefix, exec_prefix, home, install_base,
+        # install_platbase, user-supplied versions of
+        # install_{purelib,platlib,lib,scripts,data,...}, and the
+        # install schemes.  Phew!
+
+        self.dump_dirs("pre-finalize_{unix,other}")
+
+        if os.name == 'posix':
+            self.finalize_unix()
+        else:
+            self.finalize_other()
+
+        self.dump_dirs("post-finalize_{unix,other}()")
+
+        # Expand configuration variables, tilde, etc. in self.install_base
+        # and self.install_platbase -- that way, we can use $base or
+        # $platbase in the other installation directories and not worry
+        # about needing recursive variable expansion (shudder).
+
+        py_version = sys.version.split()[0]
+        (prefix, exec_prefix) = get_config_vars('prefix', 'exec_prefix')
+        try:
+            abiflags = sys.abiflags
+        except AttributeError:
+            # sys.abiflags may not be defined on all platforms.
+            abiflags = ''
+        local_vars = {
+            'dist_name': self.distribution.get_name(),
+            'dist_version': self.distribution.get_version(),
+            'dist_fullname': self.distribution.get_fullname(),
+            'py_version': py_version,
+            'py_version_short': f'{sys.version_info.major}.{sys.version_info.minor}',
+            'py_version_nodot': f'{sys.version_info.major}{sys.version_info.minor}',
+            'sys_prefix': prefix,
+            'prefix': prefix,
+            'sys_exec_prefix': exec_prefix,
+            'exec_prefix': exec_prefix,
+            'abiflags': abiflags,
+            'platlibdir': getattr(sys, 'platlibdir', 'lib'),
+            'implementation_lower': _get_implementation().lower(),
+            'implementation': _get_implementation(),
+        }
+
+        # vars for compatibility on older Pythons
+        compat_vars = dict(
+            # Python 3.9 and earlier
+            py_version_nodot_plat=getattr(sys, 'winver', '').replace('.', ''),
+        )
+
+        if HAS_USER_SITE:
+            local_vars['userbase'] = self.install_userbase
+            local_vars['usersite'] = self.install_usersite
+
+        self.config_vars = jaraco.collections.DictStack([
+            fw.vars(),
+            compat_vars,
+            sysconfig.get_config_vars(),
+            local_vars,
+        ])
+
+        self.expand_basedirs()
+
+        self.dump_dirs("post-expand_basedirs()")
+
+        # Now define config vars for the base directories so we can expand
+        # everything else.
+        local_vars['base'] = self.install_base
+        local_vars['platbase'] = self.install_platbase
+
+        if DEBUG:
+            from pprint import pprint
+
+            print("config vars:")
+            pprint(dict(self.config_vars))
+
+        # Expand "~" and configuration variables in the installation
+        # directories.
+        self.expand_dirs()
+
+        self.dump_dirs("post-expand_dirs()")
+
+        # Create directories in the home dir:
+        if self.user:
+            self.create_home_path()
+
+        # Pick the actual directory to install all modules to: either
+        # install_purelib or install_platlib, depending on whether this
+        # module distribution is pure or not.  Of course, if the user
+        # already specified install_lib, use their selection.
+        if self.install_lib is None:
+            if self.distribution.has_ext_modules():  # has extensions: non-pure
+                self.install_lib = self.install_platlib
+            else:
+                self.install_lib = self.install_purelib
+
+        # Convert directories from Unix /-separated syntax to the local
+        # convention.
+        self.convert_paths(
+            'lib',
+            'purelib',
+            'platlib',
+            'scripts',
+            'data',
+            'headers',
+            'userbase',
+            'usersite',
+        )
+
+        # Deprecated
+        # Well, we're not actually fully completely finalized yet: we still
+        # have to deal with 'extra_path', which is the hack for allowing
+        # non-packagized module distributions (hello, Numerical Python!) to
+        # get their own directories.
+        self.handle_extra_path()
+        self.install_libbase = self.install_lib  # needed for .pth file
+        self.install_lib = os.path.join(self.install_lib, self.extra_dirs)
+
+        # If a new root directory was supplied, make all the installation
+        # dirs relative to it.
+        if self.root is not None:
+            self.change_roots(
+                'libbase', 'lib', 'purelib', 'platlib', 'scripts', 'data', 'headers'
+            )
+
+        self.dump_dirs("after prepending root")
+
+        # Find out the build directories, ie. where to install from.
+        self.set_undefined_options(
+            'build', ('build_base', 'build_base'), ('build_lib', 'build_lib')
+        )
+
+        # Punt on doc directories for now -- after all, we're punting on
+        # documentation completely!
+
+    def dump_dirs(self, msg):
+        """Dumps the list of user options."""
+        if not DEBUG:
+            return
+        from ..fancy_getopt import longopt_xlate
+
+        log.debug(msg + ":")
+        for opt in self.user_options:
+            opt_name = opt[0]
+            if opt_name[-1] == "=":
+                opt_name = opt_name[0:-1]
+            if opt_name in self.negative_opt:
+                opt_name = self.negative_opt[opt_name]
+                opt_name = opt_name.translate(longopt_xlate)
+                val = not getattr(self, opt_name)
+            else:
+                opt_name = opt_name.translate(longopt_xlate)
+                val = getattr(self, opt_name)
+            log.debug("  %s: %s", opt_name, val)
+
+    def finalize_unix(self):
+        """Finalizes options for posix platforms."""
+        if self.install_base is not None or self.install_platbase is not None:
+            incomplete_scheme = (
+                (
+                    self.install_lib is None
+                    and self.install_purelib is None
+                    and self.install_platlib is None
+                )
+                or self.install_headers is None
+                or self.install_scripts is None
+                or self.install_data is None
+            )
+            if incomplete_scheme:
+                raise DistutilsOptionError(
+                    "install-base or install-platbase supplied, but "
+                    "installation scheme is incomplete"
+                )
+            return
+
+        if self.user:
+            if self.install_userbase is None:
+                raise DistutilsPlatformError("User base directory is not specified")
+            self.install_base = self.install_platbase = self.install_userbase
+            self.select_scheme("posix_user")
+        elif self.home is not None:
+            self.install_base = self.install_platbase = self.home
+            self.select_scheme("posix_home")
+        else:
+            if self.prefix is None:
+                if self.exec_prefix is not None:
+                    raise DistutilsOptionError(
+                        "must not supply exec-prefix without prefix"
+                    )
+
+                # Allow Fedora to add components to the prefix
+                _prefix_addition = getattr(sysconfig, '_prefix_addition', "")
+
+                self.prefix = os.path.normpath(sys.prefix) + _prefix_addition
+                self.exec_prefix = os.path.normpath(sys.exec_prefix) + _prefix_addition
+
+            else:
+                if self.exec_prefix is None:
+                    self.exec_prefix = self.prefix
+
+            self.install_base = self.prefix
+            self.install_platbase = self.exec_prefix
+            self.select_scheme("posix_prefix")
+
+    def finalize_other(self):
+        """Finalizes options for non-posix platforms"""
+        if self.user:
+            if self.install_userbase is None:
+                raise DistutilsPlatformError("User base directory is not specified")
+            self.install_base = self.install_platbase = self.install_userbase
+            self.select_scheme(os.name + "_user")
+        elif self.home is not None:
+            self.install_base = self.install_platbase = self.home
+            self.select_scheme("posix_home")
+        else:
+            if self.prefix is None:
+                self.prefix = os.path.normpath(sys.prefix)
+
+            self.install_base = self.install_platbase = self.prefix
+            try:
+                self.select_scheme(os.name)
+            except KeyError:
+                raise DistutilsPlatformError(
+                    f"I don't know how to install stuff on '{os.name}'"
+                )
+
+    def select_scheme(self, name):
+        _select_scheme(self, name)
+
+    def _expand_attrs(self, attrs):
+        for attr in attrs:
+            val = getattr(self, attr)
+            if val is not None:
+                if os.name in ('posix', 'nt'):
+                    val = os.path.expanduser(val)
+                val = subst_vars(val, self.config_vars)
+                setattr(self, attr, val)
+
+    def expand_basedirs(self):
+        """Calls `os.path.expanduser` on install_base, install_platbase and
+        root."""
+        self._expand_attrs(['install_base', 'install_platbase', 'root'])
+
+    def expand_dirs(self):
+        """Calls `os.path.expanduser` on install dirs."""
+        self._expand_attrs([
+            'install_purelib',
+            'install_platlib',
+            'install_lib',
+            'install_headers',
+            'install_scripts',
+            'install_data',
+        ])
+
+    def convert_paths(self, *names):
+        """Call `convert_path` over `names`."""
+        for name in names:
+            attr = "install_" + name
+            setattr(self, attr, convert_path(getattr(self, attr)))
+
+    def handle_extra_path(self):
+        """Set `path_file` and `extra_dirs` using `extra_path`."""
+        if self.extra_path is None:
+            self.extra_path = self.distribution.extra_path
+
+        if self.extra_path is not None:
+            log.warning(
+                "Distribution option extra_path is deprecated. "
+                "See issue27919 for details."
+            )
+            if isinstance(self.extra_path, str):
+                self.extra_path = self.extra_path.split(',')
+
+            if len(self.extra_path) == 1:
+                path_file = extra_dirs = self.extra_path[0]
+            elif len(self.extra_path) == 2:
+                path_file, extra_dirs = self.extra_path
+            else:
+                raise DistutilsOptionError(
+                    "'extra_path' option must be a list, tuple, or "
+                    "comma-separated string with 1 or 2 elements"
+                )
+
+            # convert to local form in case Unix notation used (as it
+            # should be in setup scripts)
+            extra_dirs = convert_path(extra_dirs)
+        else:
+            path_file = None
+            extra_dirs = ''
+
+        # XXX should we warn if path_file and not extra_dirs? (in which
+        # case the path file would be harmless but pointless)
+        self.path_file = path_file
+        self.extra_dirs = extra_dirs
+
+    def change_roots(self, *names):
+        """Change the install directories pointed by name using root."""
+        for name in names:
+            attr = "install_" + name
+            setattr(self, attr, change_root(self.root, getattr(self, attr)))
+
+    def create_home_path(self):
+        """Create directories under ~."""
+        if not self.user:
+            return
+        home = convert_path(os.path.expanduser("~"))
+        for path in self.config_vars.values():
+            if str(path).startswith(home) and not os.path.isdir(path):
+                self.debug_print(f"os.makedirs('{path}', 0o700)")
+                os.makedirs(path, 0o700)
+
+    # -- Command execution methods -------------------------------------
+
+    def run(self):
+        """Runs the command."""
+        # Obviously have to build before we can install
+        if not self.skip_build:
+            self.run_command('build')
+            # If we built for any other platform, we can't install.
+            build_plat = self.distribution.get_command_obj('build').plat_name
+            # check warn_dir - it is a clue that the 'install' is happening
+            # internally, and not to sys.path, so we don't check the platform
+            # matches what we are running.
+            if self.warn_dir and build_plat != get_platform():
+                raise DistutilsPlatformError("Can't install when cross-compiling")
+
+        # Run all sub-commands (at least those that need to be run)
+        for cmd_name in self.get_sub_commands():
+            self.run_command(cmd_name)
+
+        if self.path_file:
+            self.create_path_file()
+
+        # write list of installed files, if requested.
+        if self.record:
+            outputs = self.get_outputs()
+            if self.root:  # strip any package prefix
+                root_len = len(self.root)
+                for counter in range(len(outputs)):
+                    outputs[counter] = outputs[counter][root_len:]
+            self.execute(
+                write_file,
+                (self.record, outputs),
+                f"writing list of installed files to '{self.record}'",
+            )
+
+        sys_path = map(os.path.normpath, sys.path)
+        sys_path = map(os.path.normcase, sys_path)
+        install_lib = os.path.normcase(os.path.normpath(self.install_lib))
+        if (
+            self.warn_dir
+            and not (self.path_file and self.install_path_file)
+            and install_lib not in sys_path
+        ):
+            log.debug(
+                (
+                    "modules installed to '%s', which is not in "
+                    "Python's module search path (sys.path) -- "
+                    "you'll have to change the search path yourself"
+                ),
+                self.install_lib,
+            )
+
+    def create_path_file(self):
+        """Creates the .pth file"""
+        filename = os.path.join(self.install_libbase, self.path_file + ".pth")
+        if self.install_path_file:
+            self.execute(
+                write_file, (filename, [self.extra_dirs]), f"creating {filename}"
+            )
+        else:
+            self.warn(f"path file '{filename}' not created")
+
+    # -- Reporting methods ---------------------------------------------
+
+    def get_outputs(self):
+        """Assembles the outputs of all the sub-commands."""
+        outputs = []
+        for cmd_name in self.get_sub_commands():
+            cmd = self.get_finalized_command(cmd_name)
+            # Add the contents of cmd.get_outputs(), ensuring
+            # that outputs doesn't contain duplicate entries
+            for filename in cmd.get_outputs():
+                if filename not in outputs:
+                    outputs.append(filename)
+
+        if self.path_file and self.install_path_file:
+            outputs.append(os.path.join(self.install_libbase, self.path_file + ".pth"))
+
+        return outputs
+
+    def get_inputs(self):
+        """Returns the inputs of all the sub-commands"""
+        # XXX gee, this looks familiar ;-(
+        inputs = []
+        for cmd_name in self.get_sub_commands():
+            cmd = self.get_finalized_command(cmd_name)
+            inputs.extend(cmd.get_inputs())
+
+        return inputs
+
+    # -- Predicates for sub-command list -------------------------------
+
+    def has_lib(self):
+        """Returns true if the current distribution has any Python
+        modules to install."""
+        return (
+            self.distribution.has_pure_modules() or self.distribution.has_ext_modules()
+        )
+
+    def has_headers(self):
+        """Returns true if the current distribution has any headers to
+        install."""
+        return self.distribution.has_headers()
+
+    def has_scripts(self):
+        """Returns true if the current distribution has any scripts to.
+        install."""
+        return self.distribution.has_scripts()
+
+    def has_data(self):
+        """Returns true if the current distribution has any data to.
+        install."""
+        return self.distribution.has_data_files()
+
+    # 'sub_commands': a list of commands this command might have to run to
+    # get its work done.  See cmd.py for more info.
+    sub_commands = [
+        ('install_lib', has_lib),
+        ('install_headers', has_headers),
+        ('install_scripts', has_scripts),
+        ('install_data', has_data),
+        ('install_egg_info', lambda self: True),
+    ]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_data.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_data.py
new file mode 100644
index 00000000..36f5bcc8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_data.py
@@ -0,0 +1,93 @@
+"""distutils.command.install_data
+
+Implements the Distutils 'install_data' command, for installing
+platform-independent data files."""
+
+# contributed by Bastian Kleineidam
+
+from __future__ import annotations
+
+import functools
+import os
+from collections.abc import Iterable
+
+from ..core import Command
+from ..util import change_root, convert_path
+
+
+class install_data(Command):
+    description = "install data files"
+
+    user_options = [
+        (
+            'install-dir=',
+            'd',
+            "base directory for installing data files [default: installation base dir]",
+        ),
+        ('root=', None, "install everything relative to this alternate root directory"),
+        ('force', 'f', "force installation (overwrite existing files)"),
+    ]
+
+    boolean_options = ['force']
+
+    def initialize_options(self):
+        self.install_dir = None
+        self.outfiles = []
+        self.root = None
+        self.force = False
+        self.data_files = self.distribution.data_files
+        self.warn_dir = True
+
+    def finalize_options(self):
+        self.set_undefined_options(
+            'install',
+            ('install_data', 'install_dir'),
+            ('root', 'root'),
+            ('force', 'force'),
+        )
+
+    def run(self):
+        self.mkpath(self.install_dir)
+        for f in self.data_files:
+            self._copy(f)
+
+    @functools.singledispatchmethod
+    def _copy(self, f: tuple[str | os.PathLike, Iterable[str | os.PathLike]]):
+        # it's a tuple with path to install to and a list of files
+        dir = convert_path(f[0])
+        if not os.path.isabs(dir):
+            dir = os.path.join(self.install_dir, dir)
+        elif self.root:
+            dir = change_root(self.root, dir)
+        self.mkpath(dir)
+
+        if f[1] == []:
+            # If there are no files listed, the user must be
+            # trying to create an empty directory, so add the
+            # directory to the list of output files.
+            self.outfiles.append(dir)
+        else:
+            # Copy files, adding them to the list of output files.
+            for data in f[1]:
+                data = convert_path(data)
+                (out, _) = self.copy_file(data, dir)
+                self.outfiles.append(out)
+
+    @_copy.register(str)
+    @_copy.register(os.PathLike)
+    def _(self, f: str | os.PathLike):
+        # it's a simple file, so copy it
+        f = convert_path(f)
+        if self.warn_dir:
+            self.warn(
+                "setup script did not provide a directory for "
+                f"'{f}' -- installing right in '{self.install_dir}'"
+            )
+        (out, _) = self.copy_file(f, self.install_dir)
+        self.outfiles.append(out)
+
+    def get_inputs(self):
+        return self.data_files or []
+
+    def get_outputs(self):
+        return self.outfiles
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_egg_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_egg_info.py
new file mode 100644
index 00000000..230e94ab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_egg_info.py
@@ -0,0 +1,91 @@
+"""
+distutils.command.install_egg_info
+
+Implements the Distutils 'install_egg_info' command, for installing
+a package's PKG-INFO metadata.
+"""
+
+import os
+import re
+import sys
+from typing import ClassVar
+
+from .. import dir_util
+from .._log import log
+from ..cmd import Command
+
+
+class install_egg_info(Command):
+    """Install an .egg-info file for the package"""
+
+    description = "Install package's PKG-INFO metadata as an .egg-info file"
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ('install-dir=', 'd', "directory to install to"),
+    ]
+
+    def initialize_options(self):
+        self.install_dir = None
+
+    @property
+    def basename(self):
+        """
+        Allow basename to be overridden by child class.
+        Ref pypa/distutils#2.
+        """
+        name = to_filename(safe_name(self.distribution.get_name()))
+        version = to_filename(safe_version(self.distribution.get_version()))
+        return f"{name}-{version}-py{sys.version_info.major}.{sys.version_info.minor}.egg-info"
+
+    def finalize_options(self):
+        self.set_undefined_options('install_lib', ('install_dir', 'install_dir'))
+        self.target = os.path.join(self.install_dir, self.basename)
+        self.outputs = [self.target]
+
+    def run(self):
+        target = self.target
+        if os.path.isdir(target) and not os.path.islink(target):
+            dir_util.remove_tree(target, dry_run=self.dry_run)
+        elif os.path.exists(target):
+            self.execute(os.unlink, (self.target,), "Removing " + target)
+        elif not os.path.isdir(self.install_dir):
+            self.execute(
+                os.makedirs, (self.install_dir,), "Creating " + self.install_dir
+            )
+        log.info("Writing %s", target)
+        if not self.dry_run:
+            with open(target, 'w', encoding='UTF-8') as f:
+                self.distribution.metadata.write_pkg_file(f)
+
+    def get_outputs(self):
+        return self.outputs
+
+
+# The following routines are taken from setuptools' pkg_resources module and
+# can be replaced by importing them from pkg_resources once it is included
+# in the stdlib.
+
+
+def safe_name(name):
+    """Convert an arbitrary string to a standard distribution name
+
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub('[^A-Za-z0-9.]+', '-', name)
+
+
+def safe_version(version):
+    """Convert an arbitrary string to a standard version string
+
+    Spaces become dots, and all other non-alphanumeric characters become
+    dashes, with runs of multiple dashes condensed to a single dash.
+    """
+    version = version.replace(' ', '.')
+    return re.sub('[^A-Za-z0-9.]+', '-', version)
+
+
+def to_filename(name):
+    """Convert a project or version name to its filename-escaped form
+
+    Any '-' characters are currently replaced with '_'.
+    """
+    return name.replace('-', '_')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_headers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_headers.py
new file mode 100644
index 00000000..586121e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_headers.py
@@ -0,0 +1,46 @@
+"""distutils.command.install_headers
+
+Implements the Distutils 'install_headers' command, to install C/C++ header
+files to the Python include directory."""
+
+from typing import ClassVar
+
+from ..core import Command
+
+
+# XXX force is never used
+class install_headers(Command):
+    description = "install C/C++ header files"
+
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ('install-dir=', 'd', "directory to install header files to"),
+        ('force', 'f', "force installation (overwrite existing files)"),
+    ]
+
+    boolean_options = ['force']
+
+    def initialize_options(self):
+        self.install_dir = None
+        self.force = False
+        self.outfiles = []
+
+    def finalize_options(self):
+        self.set_undefined_options(
+            'install', ('install_headers', 'install_dir'), ('force', 'force')
+        )
+
+    def run(self):
+        headers = self.distribution.headers
+        if not headers:
+            return
+
+        self.mkpath(self.install_dir)
+        for header in headers:
+            (out, _) = self.copy_file(header, self.install_dir)
+            self.outfiles.append(out)
+
+    def get_inputs(self):
+        return self.distribution.headers or []
+
+    def get_outputs(self):
+        return self.outfiles
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_lib.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_lib.py
new file mode 100644
index 00000000..4c1230a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_lib.py
@@ -0,0 +1,234 @@
+"""distutils.command.install_lib
+
+Implements the Distutils 'install_lib' command
+(install all Python modules)."""
+
+import importlib.util
+import os
+import sys
+
+from ..core import Command
+from ..errors import DistutilsOptionError
+
+# Extension for Python source files.
+PYTHON_SOURCE_EXTENSION = ".py"
+
+
+class install_lib(Command):
+    description = "install all Python modules (extensions and pure Python)"
+
+    # The byte-compilation options are a tad confusing.  Here are the
+    # possible scenarios:
+    #   1) no compilation at all (--no-compile --no-optimize)
+    #   2) compile .pyc only (--compile --no-optimize; default)
+    #   3) compile .pyc and "opt-1" .pyc (--compile --optimize)
+    #   4) compile "opt-1" .pyc only (--no-compile --optimize)
+    #   5) compile .pyc and "opt-2" .pyc (--compile --optimize-more)
+    #   6) compile "opt-2" .pyc only (--no-compile --optimize-more)
+    #
+    # The UI for this is two options, 'compile' and 'optimize'.
+    # 'compile' is strictly boolean, and only decides whether to
+    # generate .pyc files.  'optimize' is three-way (0, 1, or 2), and
+    # decides both whether to generate .pyc files and what level of
+    # optimization to use.
+
+    user_options = [
+        ('install-dir=', 'd', "directory to install to"),
+        ('build-dir=', 'b', "build directory (where to install from)"),
+        ('force', 'f', "force installation (overwrite existing files)"),
+        ('compile', 'c', "compile .py to .pyc [default]"),
+        ('no-compile', None, "don't compile .py files"),
+        (
+            'optimize=',
+            'O',
+            "also compile with optimization: -O1 for \"python -O\", "
+            "-O2 for \"python -OO\", and -O0 to disable [default: -O0]",
+        ),
+        ('skip-build', None, "skip the build steps"),
+    ]
+
+    boolean_options = ['force', 'compile', 'skip-build']
+    negative_opt = {'no-compile': 'compile'}
+
+    def initialize_options(self):
+        # let the 'install' command dictate our installation directory
+        self.install_dir = None
+        self.build_dir = None
+        self.force = False
+        self.compile = None
+        self.optimize = None
+        self.skip_build = None
+
+    def finalize_options(self):
+        # Get all the information we need to install pure Python modules
+        # from the umbrella 'install' command -- build (source) directory,
+        # install (target) directory, and whether to compile .py files.
+        self.set_undefined_options(
+            'install',
+            ('build_lib', 'build_dir'),
+            ('install_lib', 'install_dir'),
+            ('force', 'force'),
+            ('compile', 'compile'),
+            ('optimize', 'optimize'),
+            ('skip_build', 'skip_build'),
+        )
+
+        if self.compile is None:
+            self.compile = True
+        if self.optimize is None:
+            self.optimize = False
+
+        if not isinstance(self.optimize, int):
+            try:
+                self.optimize = int(self.optimize)
+            except ValueError:
+                pass
+            if self.optimize not in (0, 1, 2):
+                raise DistutilsOptionError("optimize must be 0, 1, or 2")
+
+    def run(self):
+        # Make sure we have built everything we need first
+        self.build()
+
+        # Install everything: simply dump the entire contents of the build
+        # directory to the installation directory (that's the beauty of
+        # having a build directory!)
+        outfiles = self.install()
+
+        # (Optionally) compile .py to .pyc
+        if outfiles is not None and self.distribution.has_pure_modules():
+            self.byte_compile(outfiles)
+
+    # -- Top-level worker functions ------------------------------------
+    # (called from 'run()')
+
+    def build(self):
+        if not self.skip_build:
+            if self.distribution.has_pure_modules():
+                self.run_command('build_py')
+            if self.distribution.has_ext_modules():
+                self.run_command('build_ext')
+
+    def install(self):
+        if os.path.isdir(self.build_dir):
+            outfiles = self.copy_tree(self.build_dir, self.install_dir)
+        else:
+            self.warn(
+                f"'{self.build_dir}' does not exist -- no Python modules to install"
+            )
+            return
+        return outfiles
+
+    def byte_compile(self, files):
+        if sys.dont_write_bytecode:
+            self.warn('byte-compiling is disabled, skipping.')
+            return
+
+        from ..util import byte_compile
+
+        # Get the "--root" directory supplied to the "install" command,
+        # and use it as a prefix to strip off the purported filename
+        # encoded in bytecode files.  This is far from complete, but it
+        # should at least generate usable bytecode in RPM distributions.
+        install_root = self.get_finalized_command('install').root
+
+        if self.compile:
+            byte_compile(
+                files,
+                optimize=0,
+                force=self.force,
+                prefix=install_root,
+                dry_run=self.dry_run,
+            )
+        if self.optimize > 0:
+            byte_compile(
+                files,
+                optimize=self.optimize,
+                force=self.force,
+                prefix=install_root,
+                verbose=self.verbose,
+                dry_run=self.dry_run,
+            )
+
+    # -- Utility methods -----------------------------------------------
+
+    def _mutate_outputs(self, has_any, build_cmd, cmd_option, output_dir):
+        if not has_any:
+            return []
+
+        build_cmd = self.get_finalized_command(build_cmd)
+        build_files = build_cmd.get_outputs()
+        build_dir = getattr(build_cmd, cmd_option)
+
+        prefix_len = len(build_dir) + len(os.sep)
+        outputs = [os.path.join(output_dir, file[prefix_len:]) for file in build_files]
+
+        return outputs
+
+    def _bytecode_filenames(self, py_filenames):
+        bytecode_files = []
+        for py_file in py_filenames:
+            # Since build_py handles package data installation, the
+            # list of outputs can contain more than just .py files.
+            # Make sure we only report bytecode for the .py files.
+            ext = os.path.splitext(os.path.normcase(py_file))[1]
+            if ext != PYTHON_SOURCE_EXTENSION:
+                continue
+            if self.compile:
+                bytecode_files.append(
+                    importlib.util.cache_from_source(py_file, optimization='')
+                )
+            if self.optimize > 0:
+                bytecode_files.append(
+                    importlib.util.cache_from_source(
+                        py_file, optimization=self.optimize
+                    )
+                )
+
+        return bytecode_files
+
+    # -- External interface --------------------------------------------
+    # (called by outsiders)
+
+    def get_outputs(self):
+        """Return the list of files that would be installed if this command
+        were actually run.  Not affected by the "dry-run" flag or whether
+        modules have actually been built yet.
+        """
+        pure_outputs = self._mutate_outputs(
+            self.distribution.has_pure_modules(),
+            'build_py',
+            'build_lib',
+            self.install_dir,
+        )
+        if self.compile:
+            bytecode_outputs = self._bytecode_filenames(pure_outputs)
+        else:
+            bytecode_outputs = []
+
+        ext_outputs = self._mutate_outputs(
+            self.distribution.has_ext_modules(),
+            'build_ext',
+            'build_lib',
+            self.install_dir,
+        )
+
+        return pure_outputs + bytecode_outputs + ext_outputs
+
+    def get_inputs(self):
+        """Get the list of files that are input to this command, ie. the
+        files that get installed as they are named in the build tree.
+        The files in this list correspond one-to-one to the output
+        filenames returned by 'get_outputs()'.
+        """
+        inputs = []
+
+        if self.distribution.has_pure_modules():
+            build_py = self.get_finalized_command('build_py')
+            inputs.extend(build_py.get_outputs())
+
+        if self.distribution.has_ext_modules():
+            build_ext = self.get_finalized_command('build_ext')
+            inputs.extend(build_ext.get_outputs())
+
+        return inputs
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_scripts.py
new file mode 100644
index 00000000..bb43387f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/install_scripts.py
@@ -0,0 +1,61 @@
+"""distutils.command.install_scripts
+
+Implements the Distutils 'install_scripts' command, for installing
+Python scripts."""
+
+# contributed by Bastian Kleineidam
+
+import os
+from distutils._log import log
+from stat import ST_MODE
+
+from ..core import Command
+
+
+class install_scripts(Command):
+    description = "install scripts (Python or otherwise)"
+
+    user_options = [
+        ('install-dir=', 'd', "directory to install scripts to"),
+        ('build-dir=', 'b', "build directory (where to install from)"),
+        ('force', 'f', "force installation (overwrite existing files)"),
+        ('skip-build', None, "skip the build steps"),
+    ]
+
+    boolean_options = ['force', 'skip-build']
+
+    def initialize_options(self):
+        self.install_dir = None
+        self.force = False
+        self.build_dir = None
+        self.skip_build = None
+
+    def finalize_options(self):
+        self.set_undefined_options('build', ('build_scripts', 'build_dir'))
+        self.set_undefined_options(
+            'install',
+            ('install_scripts', 'install_dir'),
+            ('force', 'force'),
+            ('skip_build', 'skip_build'),
+        )
+
+    def run(self):
+        if not self.skip_build:
+            self.run_command('build_scripts')
+        self.outfiles = self.copy_tree(self.build_dir, self.install_dir)
+        if os.name == 'posix':
+            # Set the executable bits (owner, group, and world) on
+            # all the scripts we just installed.
+            for file in self.get_outputs():
+                if self.dry_run:
+                    log.info("changing mode of %s", file)
+                else:
+                    mode = ((os.stat(file)[ST_MODE]) | 0o555) & 0o7777
+                    log.info("changing mode of %s to %o", file, mode)
+                    os.chmod(file, mode)
+
+    def get_inputs(self):
+        return self.distribution.scripts or []
+
+    def get_outputs(self):
+        return self.outfiles or []
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/sdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/sdist.py
new file mode 100644
index 00000000..acb3a416
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/command/sdist.py
@@ -0,0 +1,515 @@
+"""distutils.command.sdist
+
+Implements the Distutils 'sdist' command (create a source distribution)."""
+
+import os
+import sys
+from distutils import archive_util, dir_util, file_util
+from distutils._log import log
+from glob import glob
+from itertools import filterfalse
+from typing import ClassVar
+
+from ..core import Command
+from ..errors import DistutilsOptionError, DistutilsTemplateError
+from ..filelist import FileList
+from ..text_file import TextFile
+from ..util import convert_path
+
+
+def show_formats():
+    """Print all possible values for the 'formats' option (used by
+    the "--help-formats" command-line option).
+    """
+    from ..archive_util import ARCHIVE_FORMATS
+    from ..fancy_getopt import FancyGetopt
+
+    formats = sorted(
+        ("formats=" + format, None, ARCHIVE_FORMATS[format][2])
+        for format in ARCHIVE_FORMATS.keys()
+    )
+    FancyGetopt(formats).print_help("List of available source distribution formats:")
+
+
+class sdist(Command):
+    description = "create a source distribution (tarball, zip file, etc.)"
+
+    def checking_metadata(self):
+        """Callable used for the check sub-command.
+
+        Placed here so user_options can view it"""
+        return self.metadata_check
+
+    user_options = [
+        ('template=', 't', "name of manifest template file [default: MANIFEST.in]"),
+        ('manifest=', 'm', "name of manifest file [default: MANIFEST]"),
+        (
+            'use-defaults',
+            None,
+            "include the default file set in the manifest "
+            "[default; disable with --no-defaults]",
+        ),
+        ('no-defaults', None, "don't include the default file set"),
+        (
+            'prune',
+            None,
+            "specifically exclude files/directories that should not be "
+            "distributed (build tree, RCS/CVS dirs, etc.) "
+            "[default; disable with --no-prune]",
+        ),
+        ('no-prune', None, "don't automatically exclude anything"),
+        (
+            'manifest-only',
+            'o',
+            "just regenerate the manifest and then stop (implies --force-manifest)",
+        ),
+        (
+            'force-manifest',
+            'f',
+            "forcibly regenerate the manifest and carry on as usual. "
+            "Deprecated: now the manifest is always regenerated.",
+        ),
+        ('formats=', None, "formats for source distribution (comma-separated list)"),
+        (
+            'keep-temp',
+            'k',
+            "keep the distribution tree around after creating " + "archive file(s)",
+        ),
+        (
+            'dist-dir=',
+            'd',
+            "directory to put the source distribution archive(s) in [default: dist]",
+        ),
+        (
+            'metadata-check',
+            None,
+            "Ensure that all required elements of meta-data "
+            "are supplied. Warn if any missing. [default]",
+        ),
+        (
+            'owner=',
+            'u',
+            "Owner name used when creating a tar file [default: current user]",
+        ),
+        (
+            'group=',
+            'g',
+            "Group name used when creating a tar file [default: current group]",
+        ),
+    ]
+
+    boolean_options = [
+        'use-defaults',
+        'prune',
+        'manifest-only',
+        'force-manifest',
+        'keep-temp',
+        'metadata-check',
+    ]
+
+    help_options = [
+        ('help-formats', None, "list available distribution formats", show_formats),
+    ]
+
+    negative_opt = {'no-defaults': 'use-defaults', 'no-prune': 'prune'}
+
+    sub_commands = [('check', checking_metadata)]
+
+    READMES: ClassVar[tuple[str, ...]] = ('README', 'README.txt', 'README.rst')
+
+    def initialize_options(self):
+        # 'template' and 'manifest' are, respectively, the names of
+        # the manifest template and manifest file.
+        self.template = None
+        self.manifest = None
+
+        # 'use_defaults': if true, we will include the default file set
+        # in the manifest
+        self.use_defaults = True
+        self.prune = True
+
+        self.manifest_only = False
+        self.force_manifest = False
+
+        self.formats = ['gztar']
+        self.keep_temp = False
+        self.dist_dir = None
+
+        self.archive_files = None
+        self.metadata_check = 1
+        self.owner = None
+        self.group = None
+
+    def finalize_options(self):
+        if self.manifest is None:
+            self.manifest = "MANIFEST"
+        if self.template is None:
+            self.template = "MANIFEST.in"
+
+        self.ensure_string_list('formats')
+
+        bad_format = archive_util.check_archive_formats(self.formats)
+        if bad_format:
+            raise DistutilsOptionError(f"unknown archive format '{bad_format}'")
+
+        if self.dist_dir is None:
+            self.dist_dir = "dist"
+
+    def run(self):
+        # 'filelist' contains the list of files that will make up the
+        # manifest
+        self.filelist = FileList()
+
+        # Run sub commands
+        for cmd_name in self.get_sub_commands():
+            self.run_command(cmd_name)
+
+        # Do whatever it takes to get the list of files to process
+        # (process the manifest template, read an existing manifest,
+        # whatever).  File list is accumulated in 'self.filelist'.
+        self.get_file_list()
+
+        # If user just wanted us to regenerate the manifest, stop now.
+        if self.manifest_only:
+            return
+
+        # Otherwise, go ahead and create the source distribution tarball,
+        # or zipfile, or whatever.
+        self.make_distribution()
+
+    def get_file_list(self):
+        """Figure out the list of files to include in the source
+        distribution, and put it in 'self.filelist'.  This might involve
+        reading the manifest template (and writing the manifest), or just
+        reading the manifest, or just using the default file set -- it all
+        depends on the user's options.
+        """
+        # new behavior when using a template:
+        # the file list is recalculated every time because
+        # even if MANIFEST.in or setup.py are not changed
+        # the user might have added some files in the tree that
+        # need to be included.
+        #
+        #  This makes --force the default and only behavior with templates.
+        template_exists = os.path.isfile(self.template)
+        if not template_exists and self._manifest_is_not_generated():
+            self.read_manifest()
+            self.filelist.sort()
+            self.filelist.remove_duplicates()
+            return
+
+        if not template_exists:
+            self.warn(
+                ("manifest template '%s' does not exist " + "(using default file list)")
+                % self.template
+            )
+        self.filelist.findall()
+
+        if self.use_defaults:
+            self.add_defaults()
+
+        if template_exists:
+            self.read_template()
+
+        if self.prune:
+            self.prune_file_list()
+
+        self.filelist.sort()
+        self.filelist.remove_duplicates()
+        self.write_manifest()
+
+    def add_defaults(self):
+        """Add all the default files to self.filelist:
+          - README or README.txt
+          - setup.py
+          - tests/test*.py and test/test*.py
+          - all pure Python modules mentioned in setup script
+          - all files pointed by package_data (build_py)
+          - all files defined in data_files.
+          - all files defined as scripts.
+          - all C sources listed as part of extensions or C libraries
+            in the setup script (doesn't catch C headers!)
+        Warns if (README or README.txt) or setup.py are missing; everything
+        else is optional.
+        """
+        self._add_defaults_standards()
+        self._add_defaults_optional()
+        self._add_defaults_python()
+        self._add_defaults_data_files()
+        self._add_defaults_ext()
+        self._add_defaults_c_libs()
+        self._add_defaults_scripts()
+
+    @staticmethod
+    def _cs_path_exists(fspath):
+        """
+        Case-sensitive path existence check
+
+        >>> sdist._cs_path_exists(__file__)
+        True
+        >>> sdist._cs_path_exists(__file__.upper())
+        False
+        """
+        if not os.path.exists(fspath):
+            return False
+        # make absolute so we always have a directory
+        abspath = os.path.abspath(fspath)
+        directory, filename = os.path.split(abspath)
+        return filename in os.listdir(directory)
+
+    def _add_defaults_standards(self):
+        standards = [self.READMES, self.distribution.script_name]
+        for fn in standards:
+            if isinstance(fn, tuple):
+                alts = fn
+                got_it = False
+                for fn in alts:
+                    if self._cs_path_exists(fn):
+                        got_it = True
+                        self.filelist.append(fn)
+                        break
+
+                if not got_it:
+                    self.warn(
+                        "standard file not found: should have one of " + ', '.join(alts)
+                    )
+            else:
+                if self._cs_path_exists(fn):
+                    self.filelist.append(fn)
+                else:
+                    self.warn(f"standard file '{fn}' not found")
+
+    def _add_defaults_optional(self):
+        optional = ['tests/test*.py', 'test/test*.py', 'setup.cfg']
+        for pattern in optional:
+            files = filter(os.path.isfile, glob(pattern))
+            self.filelist.extend(files)
+
+    def _add_defaults_python(self):
+        # build_py is used to get:
+        #  - python modules
+        #  - files defined in package_data
+        build_py = self.get_finalized_command('build_py')
+
+        # getting python files
+        if self.distribution.has_pure_modules():
+            self.filelist.extend(build_py.get_source_files())
+
+        # getting package_data files
+        # (computed in build_py.data_files by build_py.finalize_options)
+        for _pkg, src_dir, _build_dir, filenames in build_py.data_files:
+            for filename in filenames:
+                self.filelist.append(os.path.join(src_dir, filename))
+
+    def _add_defaults_data_files(self):
+        # getting distribution.data_files
+        if self.distribution.has_data_files():
+            for item in self.distribution.data_files:
+                if isinstance(item, str):
+                    # plain file
+                    item = convert_path(item)
+                    if os.path.isfile(item):
+                        self.filelist.append(item)
+                else:
+                    # a (dirname, filenames) tuple
+                    dirname, filenames = item
+                    for f in filenames:
+                        f = convert_path(f)
+                        if os.path.isfile(f):
+                            self.filelist.append(f)
+
+    def _add_defaults_ext(self):
+        if self.distribution.has_ext_modules():
+            build_ext = self.get_finalized_command('build_ext')
+            self.filelist.extend(build_ext.get_source_files())
+
+    def _add_defaults_c_libs(self):
+        if self.distribution.has_c_libraries():
+            build_clib = self.get_finalized_command('build_clib')
+            self.filelist.extend(build_clib.get_source_files())
+
+    def _add_defaults_scripts(self):
+        if self.distribution.has_scripts():
+            build_scripts = self.get_finalized_command('build_scripts')
+            self.filelist.extend(build_scripts.get_source_files())
+
+    def read_template(self):
+        """Read and parse manifest template file named by self.template.
+
+        (usually "MANIFEST.in") The parsing and processing is done by
+        'self.filelist', which updates itself accordingly.
+        """
+        log.info("reading manifest template '%s'", self.template)
+        template = TextFile(
+            self.template,
+            strip_comments=True,
+            skip_blanks=True,
+            join_lines=True,
+            lstrip_ws=True,
+            rstrip_ws=True,
+            collapse_join=True,
+        )
+
+        try:
+            while True:
+                line = template.readline()
+                if line is None:  # end of file
+                    break
+
+                try:
+                    self.filelist.process_template_line(line)
+                # the call above can raise a DistutilsTemplateError for
+                # malformed lines, or a ValueError from the lower-level
+                # convert_path function
+                except (DistutilsTemplateError, ValueError) as msg:
+                    self.warn(
+                        f"{template.filename}, line {int(template.current_line)}: {msg}"
+                    )
+        finally:
+            template.close()
+
+    def prune_file_list(self):
+        """Prune off branches that might slip into the file list as created
+        by 'read_template()', but really don't belong there:
+          * the build tree (typically "build")
+          * the release tree itself (only an issue if we ran "sdist"
+            previously with --keep-temp, or it aborted)
+          * any RCS, CVS, .svn, .hg, .git, .bzr, _darcs directories
+        """
+        build = self.get_finalized_command('build')
+        base_dir = self.distribution.get_fullname()
+
+        self.filelist.exclude_pattern(None, prefix=os.fspath(build.build_base))
+        self.filelist.exclude_pattern(None, prefix=base_dir)
+
+        if sys.platform == 'win32':
+            seps = r'/|\\'
+        else:
+            seps = '/'
+
+        vcs_dirs = ['RCS', 'CVS', r'\.svn', r'\.hg', r'\.git', r'\.bzr', '_darcs']
+        vcs_ptrn = r'(^|{})({})({}).*'.format(seps, '|'.join(vcs_dirs), seps)
+        self.filelist.exclude_pattern(vcs_ptrn, is_regex=True)
+
+    def write_manifest(self):
+        """Write the file list in 'self.filelist' (presumably as filled in
+        by 'add_defaults()' and 'read_template()') to the manifest file
+        named by 'self.manifest'.
+        """
+        if self._manifest_is_not_generated():
+            log.info(
+                f"not writing to manually maintained manifest file '{self.manifest}'"
+            )
+            return
+
+        content = self.filelist.files[:]
+        content.insert(0, '# file GENERATED by distutils, do NOT edit')
+        self.execute(
+            file_util.write_file,
+            (self.manifest, content),
+            f"writing manifest file '{self.manifest}'",
+        )
+
+    def _manifest_is_not_generated(self):
+        # check for special comment used in 3.1.3 and higher
+        if not os.path.isfile(self.manifest):
+            return False
+
+        with open(self.manifest, encoding='utf-8') as fp:
+            first_line = next(fp)
+        return first_line != '# file GENERATED by distutils, do NOT edit\n'
+
+    def read_manifest(self):
+        """Read the manifest file (named by 'self.manifest') and use it to
+        fill in 'self.filelist', the list of files to include in the source
+        distribution.
+        """
+        log.info("reading manifest file '%s'", self.manifest)
+        with open(self.manifest, encoding='utf-8') as lines:
+            self.filelist.extend(
+                # ignore comments and blank lines
+                filter(None, filterfalse(is_comment, map(str.strip, lines)))
+            )
+
+    def make_release_tree(self, base_dir, files):
+        """Create the directory tree that will become the source
+        distribution archive.  All directories implied by the filenames in
+        'files' are created under 'base_dir', and then we hard link or copy
+        (if hard linking is unavailable) those files into place.
+        Essentially, this duplicates the developer's source tree, but in a
+        directory named after the distribution, containing only the files
+        to be distributed.
+        """
+        # Create all the directories under 'base_dir' necessary to
+        # put 'files' there; the 'mkpath()' is just so we don't die
+        # if the manifest happens to be empty.
+        self.mkpath(base_dir)
+        dir_util.create_tree(base_dir, files, dry_run=self.dry_run)
+
+        # And walk over the list of files, either making a hard link (if
+        # os.link exists) to each one that doesn't already exist in its
+        # corresponding location under 'base_dir', or copying each file
+        # that's out-of-date in 'base_dir'.  (Usually, all files will be
+        # out-of-date, because by default we blow away 'base_dir' when
+        # we're done making the distribution archives.)
+
+        if hasattr(os, 'link'):  # can make hard links on this system
+            link = 'hard'
+            msg = f"making hard links in {base_dir}..."
+        else:  # nope, have to copy
+            link = None
+            msg = f"copying files to {base_dir}..."
+
+        if not files:
+            log.warning("no files to distribute -- empty manifest?")
+        else:
+            log.info(msg)
+        for file in files:
+            if not os.path.isfile(file):
+                log.warning("'%s' not a regular file -- skipping", file)
+            else:
+                dest = os.path.join(base_dir, file)
+                self.copy_file(file, dest, link=link)
+
+        self.distribution.metadata.write_pkg_info(base_dir)
+
+    def make_distribution(self):
+        """Create the source distribution(s).  First, we create the release
+        tree with 'make_release_tree()'; then, we create all required
+        archive files (according to 'self.formats') from the release tree.
+        Finally, we clean up by blowing away the release tree (unless
+        'self.keep_temp' is true).  The list of archive files created is
+        stored so it can be retrieved later by 'get_archive_files()'.
+        """
+        # Don't warn about missing meta-data here -- should be (and is!)
+        # done elsewhere.
+        base_dir = self.distribution.get_fullname()
+        base_name = os.path.join(self.dist_dir, base_dir)
+
+        self.make_release_tree(base_dir, self.filelist.files)
+        archive_files = []  # remember names of files we create
+        # tar archive must be created last to avoid overwrite and remove
+        if 'tar' in self.formats:
+            self.formats.append(self.formats.pop(self.formats.index('tar')))
+
+        for fmt in self.formats:
+            file = self.make_archive(
+                base_name, fmt, base_dir=base_dir, owner=self.owner, group=self.group
+            )
+            archive_files.append(file)
+            self.distribution.dist_files.append(('sdist', '', file))
+
+        self.archive_files = archive_files
+
+        if not self.keep_temp:
+            dir_util.remove_tree(base_dir, dry_run=self.dry_run)
+
+    def get_archive_files(self):
+        """Return the list of archive files created when the command
+        was run, or None if the command hasn't run yet.
+        """
+        return self.archive_files
+
+
+def is_comment(line):
+    return line.startswith('#')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__init__.py
new file mode 100644
index 00000000..c715ee9c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__init__.py
@@ -0,0 +1,13 @@
+from __future__ import annotations
+
+
+def consolidate_linker_args(args: list[str]) -> list[str] | str:
+    """
+    Ensure the return value is a string for backward compatibility.
+
+    Retain until at least 2025-04-31. See pypa/distutils#246
+    """
+
+    if not all(arg.startswith('-Wl,') for arg in args):
+        return args
+    return '-Wl,' + ','.join(arg.removeprefix('-Wl,') for arg in args)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..50573ac7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/py39.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/py39.cpython-312.pyc
new file mode 100644
index 00000000..a2ddc956
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/__pycache__/py39.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/py39.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/py39.py
new file mode 100644
index 00000000..1b436d76
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/compat/py39.py
@@ -0,0 +1,66 @@
+import functools
+import itertools
+import platform
+import sys
+
+
+def add_ext_suffix_39(vars):
+    """
+    Ensure vars contains 'EXT_SUFFIX'. pypa/distutils#130
+    """
+    import _imp
+
+    ext_suffix = _imp.extension_suffixes()[0]
+    vars.update(
+        EXT_SUFFIX=ext_suffix,
+        # sysconfig sets SO to match EXT_SUFFIX, so maintain
+        # that expectation.
+        # https://github.com/python/cpython/blob/785cc6770588de087d09e89a69110af2542be208/Lib/sysconfig.py#L671-L673
+        SO=ext_suffix,
+    )
+
+
+needs_ext_suffix = sys.version_info < (3, 10) and platform.system() == 'Windows'
+add_ext_suffix = add_ext_suffix_39 if needs_ext_suffix else lambda vars: None
+
+
+# from more_itertools
+class UnequalIterablesError(ValueError):
+    def __init__(self, details=None):
+        msg = 'Iterables have different lengths'
+        if details is not None:
+            msg += (': index 0 has length {}; index {} has length {}').format(*details)
+
+        super().__init__(msg)
+
+
+# from more_itertools
+def _zip_equal_generator(iterables):
+    _marker = object()
+    for combo in itertools.zip_longest(*iterables, fillvalue=_marker):
+        for val in combo:
+            if val is _marker:
+                raise UnequalIterablesError()
+        yield combo
+
+
+# from more_itertools
+def _zip_equal(*iterables):
+    # Check whether the iterables are all the same size.
+    try:
+        first_size = len(iterables[0])
+        for i, it in enumerate(iterables[1:], 1):
+            size = len(it)
+            if size != first_size:
+                raise UnequalIterablesError(details=(first_size, i, size))
+        # All sizes are equal, we can use the built-in zip.
+        return zip(*iterables)
+    # If any one of the iterables didn't have a length, start reading
+    # them until one runs out.
+    except TypeError:
+        return _zip_equal_generator(iterables)
+
+
+zip_strict = (
+    _zip_equal if sys.version_info < (3, 10) else functools.partial(zip, strict=True)
+)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/core.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/core.py
new file mode 100644
index 00000000..bd62546b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/core.py
@@ -0,0 +1,289 @@
+"""distutils.core
+
+The only module that needs to be imported to use the Distutils; provides
+the 'setup' function (which is to be called from the setup script).  Also
+indirectly provides the Distribution and Command classes, although they are
+really defined in distutils.dist and distutils.cmd.
+"""
+
+from __future__ import annotations
+
+import os
+import sys
+import tokenize
+from collections.abc import Iterable
+
+from .cmd import Command
+from .debug import DEBUG
+
+# Mainly import these so setup scripts can "from distutils.core import" them.
+from .dist import Distribution
+from .errors import (
+    CCompilerError,
+    DistutilsArgError,
+    DistutilsError,
+    DistutilsSetupError,
+)
+from .extension import Extension
+
+__all__ = ['Distribution', 'Command', 'Extension', 'setup']
+
+# This is a barebones help message generated displayed when the user
+# runs the setup script with no arguments at all.  More useful help
+# is generated with various --help options: global help, list commands,
+# and per-command help.
+USAGE = """\
+usage: %(script)s [global_opts] cmd1 [cmd1_opts] [cmd2 [cmd2_opts] ...]
+   or: %(script)s --help [cmd1 cmd2 ...]
+   or: %(script)s --help-commands
+   or: %(script)s cmd --help
+"""
+
+
+def gen_usage(script_name):
+    script = os.path.basename(script_name)
+    return USAGE % locals()
+
+
+# Some mild magic to control the behaviour of 'setup()' from 'run_setup()'.
+_setup_stop_after = None
+_setup_distribution = None
+
+# Legal keyword arguments for the setup() function
+setup_keywords = (
+    'distclass',
+    'script_name',
+    'script_args',
+    'options',
+    'name',
+    'version',
+    'author',
+    'author_email',
+    'maintainer',
+    'maintainer_email',
+    'url',
+    'license',
+    'description',
+    'long_description',
+    'keywords',
+    'platforms',
+    'classifiers',
+    'download_url',
+    'requires',
+    'provides',
+    'obsoletes',
+)
+
+# Legal keyword arguments for the Extension constructor
+extension_keywords = (
+    'name',
+    'sources',
+    'include_dirs',
+    'define_macros',
+    'undef_macros',
+    'library_dirs',
+    'libraries',
+    'runtime_library_dirs',
+    'extra_objects',
+    'extra_compile_args',
+    'extra_link_args',
+    'swig_opts',
+    'export_symbols',
+    'depends',
+    'language',
+)
+
+
+def setup(**attrs):  # noqa: C901
+    """The gateway to the Distutils: do everything your setup script needs
+    to do, in a highly flexible and user-driven way.  Briefly: create a
+    Distribution instance; find and parse config files; parse the command
+    line; run each Distutils command found there, customized by the options
+    supplied to 'setup()' (as keyword arguments), in config files, and on
+    the command line.
+
+    The Distribution instance might be an instance of a class supplied via
+    the 'distclass' keyword argument to 'setup'; if no such class is
+    supplied, then the Distribution class (in dist.py) is instantiated.
+    All other arguments to 'setup' (except for 'cmdclass') are used to set
+    attributes of the Distribution instance.
+
+    The 'cmdclass' argument, if supplied, is a dictionary mapping command
+    names to command classes.  Each command encountered on the command line
+    will be turned into a command class, which is in turn instantiated; any
+    class found in 'cmdclass' is used in place of the default, which is
+    (for command 'foo_bar') class 'foo_bar' in module
+    'distutils.command.foo_bar'.  The command class must provide a
+    'user_options' attribute which is a list of option specifiers for
+    'distutils.fancy_getopt'.  Any command-line options between the current
+    and the next command are used to set attributes of the current command
+    object.
+
+    When the entire command-line has been successfully parsed, calls the
+    'run()' method on each command object in turn.  This method will be
+    driven entirely by the Distribution object (which each command object
+    has a reference to, thanks to its constructor), and the
+    command-specific options that became attributes of each command
+    object.
+    """
+
+    global _setup_stop_after, _setup_distribution
+
+    # Determine the distribution class -- either caller-supplied or
+    # our Distribution (see below).
+    klass = attrs.get('distclass')
+    if klass:
+        attrs.pop('distclass')
+    else:
+        klass = Distribution
+
+    if 'script_name' not in attrs:
+        attrs['script_name'] = os.path.basename(sys.argv[0])
+    if 'script_args' not in attrs:
+        attrs['script_args'] = sys.argv[1:]
+
+    # Create the Distribution instance, using the remaining arguments
+    # (ie. everything except distclass) to initialize it
+    try:
+        _setup_distribution = dist = klass(attrs)
+    except DistutilsSetupError as msg:
+        if 'name' not in attrs:
+            raise SystemExit(f"error in setup command: {msg}")
+        else:
+            raise SystemExit("error in {} setup command: {}".format(attrs['name'], msg))
+
+    if _setup_stop_after == "init":
+        return dist
+
+    # Find and parse the config file(s): they will override options from
+    # the setup script, but be overridden by the command line.
+    dist.parse_config_files()
+
+    if DEBUG:
+        print("options (after parsing config files):")
+        dist.dump_option_dicts()
+
+    if _setup_stop_after == "config":
+        return dist
+
+    # Parse the command line and override config files; any
+    # command-line errors are the end user's fault, so turn them into
+    # SystemExit to suppress tracebacks.
+    try:
+        ok = dist.parse_command_line()
+    except DistutilsArgError as msg:
+        raise SystemExit(gen_usage(dist.script_name) + f"\nerror: {msg}")
+
+    if DEBUG:
+        print("options (after parsing command line):")
+        dist.dump_option_dicts()
+
+    if _setup_stop_after == "commandline":
+        return dist
+
+    # And finally, run all the commands found on the command line.
+    if ok:
+        return run_commands(dist)
+
+    return dist
+
+
+# setup ()
+
+
+def run_commands(dist):
+    """Given a Distribution object run all the commands,
+    raising ``SystemExit`` errors in the case of failure.
+
+    This function assumes that either ``sys.argv`` or ``dist.script_args``
+    is already set accordingly.
+    """
+    try:
+        dist.run_commands()
+    except KeyboardInterrupt:
+        raise SystemExit("interrupted")
+    except OSError as exc:
+        if DEBUG:
+            sys.stderr.write(f"error: {exc}\n")
+            raise
+        else:
+            raise SystemExit(f"error: {exc}")
+
+    except (DistutilsError, CCompilerError) as msg:
+        if DEBUG:
+            raise
+        else:
+            raise SystemExit("error: " + str(msg))
+
+    return dist
+
+
+def run_setup(script_name, script_args: Iterable[str] | None = None, stop_after="run"):
+    """Run a setup script in a somewhat controlled environment, and
+    return the Distribution instance that drives things.  This is useful
+    if you need to find out the distribution meta-data (passed as
+    keyword args from 'script' to 'setup()', or the contents of the
+    config files or command-line.
+
+    'script_name' is a file that will be read and run with 'exec()';
+    'sys.argv[0]' will be replaced with 'script' for the duration of the
+    call.  'script_args' is a list of strings; if supplied,
+    'sys.argv[1:]' will be replaced by 'script_args' for the duration of
+    the call.
+
+    'stop_after' tells 'setup()' when to stop processing; possible
+    values:
+      init
+        stop after the Distribution instance has been created and
+        populated with the keyword arguments to 'setup()'
+      config
+        stop after config files have been parsed (and their data
+        stored in the Distribution instance)
+      commandline
+        stop after the command-line ('sys.argv[1:]' or 'script_args')
+        have been parsed (and the data stored in the Distribution)
+      run [default]
+        stop after all commands have been run (the same as if 'setup()'
+        had been called in the usual way
+
+    Returns the Distribution instance, which provides all information
+    used to drive the Distutils.
+    """
+    if stop_after not in ('init', 'config', 'commandline', 'run'):
+        raise ValueError(f"invalid value for 'stop_after': {stop_after!r}")
+
+    global _setup_stop_after, _setup_distribution
+    _setup_stop_after = stop_after
+
+    save_argv = sys.argv.copy()
+    g = {'__file__': script_name, '__name__': '__main__'}
+    try:
+        try:
+            sys.argv[0] = script_name
+            if script_args is not None:
+                sys.argv[1:] = script_args
+            # tokenize.open supports automatic encoding detection
+            with tokenize.open(script_name) as f:
+                code = f.read().replace(r'\r\n', r'\n')
+                exec(code, g)
+        finally:
+            sys.argv = save_argv
+            _setup_stop_after = None
+    except SystemExit:
+        # Hmm, should we do something if exiting with a non-zero code
+        # (ie. error)?
+        pass
+
+    if _setup_distribution is None:
+        raise RuntimeError(
+            "'distutils.core.setup()' was never called -- "
+            f"perhaps '{script_name}' is not a Distutils setup script?"
+        )
+
+    # I wonder if the setup script's namespace -- g and l -- would be of
+    # any interest to callers?
+    # print "_setup_distribution:", _setup_distribution
+    return _setup_distribution
+
+
+# run_setup ()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cygwinccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cygwinccompiler.py
new file mode 100644
index 00000000..3c67524e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/cygwinccompiler.py
@@ -0,0 +1,339 @@
+"""distutils.cygwinccompiler
+
+Provides the CygwinCCompiler class, a subclass of UnixCCompiler that
+handles the Cygwin port of the GNU C compiler to Windows.  It also contains
+the Mingw32CCompiler class which handles the mingw32 port of GCC (same as
+cygwin in no-cygwin mode).
+"""
+
+import copy
+import os
+import pathlib
+import shlex
+import sys
+import warnings
+from subprocess import check_output
+
+from .errors import (
+    CCompilerError,
+    CompileError,
+    DistutilsExecError,
+    DistutilsPlatformError,
+)
+from .file_util import write_file
+from .sysconfig import get_config_vars
+from .unixccompiler import UnixCCompiler
+from .version import LooseVersion, suppress_known_deprecation
+
+
+def get_msvcr():
+    """No longer needed, but kept for backward compatibility."""
+    return []
+
+
+_runtime_library_dirs_msg = (
+    "Unable to set runtime library search path on Windows, "
+    "usually indicated by `runtime_library_dirs` parameter to Extension"
+)
+
+
+class CygwinCCompiler(UnixCCompiler):
+    """Handles the Cygwin port of the GNU C compiler to Windows."""
+
+    compiler_type = 'cygwin'
+    obj_extension = ".o"
+    static_lib_extension = ".a"
+    shared_lib_extension = ".dll.a"
+    dylib_lib_extension = ".dll"
+    static_lib_format = "lib%s%s"
+    shared_lib_format = "lib%s%s"
+    dylib_lib_format = "cyg%s%s"
+    exe_extension = ".exe"
+
+    def __init__(self, verbose=False, dry_run=False, force=False):
+        super().__init__(verbose, dry_run, force)
+
+        status, details = check_config_h()
+        self.debug_print(f"Python's GCC status: {status} (details: {details})")
+        if status is not CONFIG_H_OK:
+            self.warn(
+                "Python's pyconfig.h doesn't seem to support your compiler. "
+                f"Reason: {details}. "
+                "Compiling may fail because of undefined preprocessor macros."
+            )
+
+        self.cc, self.cxx = get_config_vars('CC', 'CXX')
+
+        # Override 'CC' and 'CXX' environment variables for
+        # building using MINGW compiler for MSVC python.
+        self.cc = os.environ.get('CC', self.cc or 'gcc')
+        self.cxx = os.environ.get('CXX', self.cxx or 'g++')
+
+        self.linker_dll = self.cc
+        self.linker_dll_cxx = self.cxx
+        shared_option = "-shared"
+
+        self.set_executables(
+            compiler=f'{self.cc} -mcygwin -O -Wall',
+            compiler_so=f'{self.cc} -mcygwin -mdll -O -Wall',
+            compiler_cxx=f'{self.cxx} -mcygwin -O -Wall',
+            compiler_so_cxx=f'{self.cxx} -mcygwin -mdll -O -Wall',
+            linker_exe=f'{self.cc} -mcygwin',
+            linker_so=f'{self.linker_dll} -mcygwin {shared_option}',
+            linker_exe_cxx=f'{self.cxx} -mcygwin',
+            linker_so_cxx=f'{self.linker_dll_cxx} -mcygwin {shared_option}',
+        )
+
+        self.dll_libraries = get_msvcr()
+
+    @property
+    def gcc_version(self):
+        # Older numpy depended on this existing to check for ancient
+        # gcc versions. This doesn't make much sense with clang etc so
+        # just hardcode to something recent.
+        # https://github.com/numpy/numpy/pull/20333
+        warnings.warn(
+            "gcc_version attribute of CygwinCCompiler is deprecated. "
+            "Instead of returning actual gcc version a fixed value 11.2.0 is returned.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        with suppress_known_deprecation():
+            return LooseVersion("11.2.0")
+
+    def _compile(self, obj, src, ext, cc_args, extra_postargs, pp_opts):
+        """Compiles the source by spawning GCC and windres if needed."""
+        if ext in ('.rc', '.res'):
+            # gcc needs '.res' and '.rc' compiled to object files !!!
+            try:
+                self.spawn(["windres", "-i", src, "-o", obj])
+            except DistutilsExecError as msg:
+                raise CompileError(msg)
+        else:  # for other files use the C-compiler
+            try:
+                if self.detect_language(src) == 'c++':
+                    self.spawn(
+                        self.compiler_so_cxx
+                        + cc_args
+                        + [src, '-o', obj]
+                        + extra_postargs
+                    )
+                else:
+                    self.spawn(
+                        self.compiler_so + cc_args + [src, '-o', obj] + extra_postargs
+                    )
+            except DistutilsExecError as msg:
+                raise CompileError(msg)
+
+    def link(
+        self,
+        target_desc,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        """Link the objects."""
+        # use separate copies, so we can modify the lists
+        extra_preargs = copy.copy(extra_preargs or [])
+        libraries = copy.copy(libraries or [])
+        objects = copy.copy(objects or [])
+
+        if runtime_library_dirs:
+            self.warn(_runtime_library_dirs_msg)
+
+        # Additional libraries
+        libraries.extend(self.dll_libraries)
+
+        # handle export symbols by creating a def-file
+        # with executables this only works with gcc/ld as linker
+        if (export_symbols is not None) and (
+            target_desc != self.EXECUTABLE or self.linker_dll == "gcc"
+        ):
+            # (The linker doesn't do anything if output is up-to-date.
+            # So it would probably better to check if we really need this,
+            # but for this we had to insert some unchanged parts of
+            # UnixCCompiler, and this is not what we want.)
+
+            # we want to put some files in the same directory as the
+            # object files are, build_temp doesn't help much
+            # where are the object files
+            temp_dir = os.path.dirname(objects[0])
+            # name of dll to give the helper files the same base name
+            (dll_name, dll_extension) = os.path.splitext(
+                os.path.basename(output_filename)
+            )
+
+            # generate the filenames for these files
+            def_file = os.path.join(temp_dir, dll_name + ".def")
+
+            # Generate .def file
+            contents = [f"LIBRARY {os.path.basename(output_filename)}", "EXPORTS"]
+            contents.extend(export_symbols)
+            self.execute(write_file, (def_file, contents), f"writing {def_file}")
+
+            # next add options for def-file
+
+            # for gcc/ld the def-file is specified as any object files
+            objects.append(def_file)
+
+        # end: if ((export_symbols is not None) and
+        #        (target_desc != self.EXECUTABLE or self.linker_dll == "gcc")):
+
+        # who wants symbols and a many times larger output file
+        # should explicitly switch the debug mode on
+        # otherwise we let ld strip the output file
+        # (On my machine: 10KiB < stripped_file < ??100KiB
+        #   unstripped_file = stripped_file + XXX KiB
+        #  ( XXX=254 for a typical python extension))
+        if not debug:
+            extra_preargs.append("-s")
+
+        UnixCCompiler.link(
+            self,
+            target_desc,
+            objects,
+            output_filename,
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            None,  # export_symbols, we do this in our def-file
+            debug,
+            extra_preargs,
+            extra_postargs,
+            build_temp,
+            target_lang,
+        )
+
+    def runtime_library_dir_option(self, dir):
+        # cygwin doesn't support rpath. While in theory we could error
+        # out like MSVC does, code might expect it to work like on Unix, so
+        # just warn and hope for the best.
+        self.warn(_runtime_library_dirs_msg)
+        return []
+
+    # -- Miscellaneous methods -----------------------------------------
+
+    def _make_out_path(self, output_dir, strip_dir, src_name):
+        # use normcase to make sure '.rc' is really '.rc' and not '.RC'
+        norm_src_name = os.path.normcase(src_name)
+        return super()._make_out_path(output_dir, strip_dir, norm_src_name)
+
+    @property
+    def out_extensions(self):
+        """
+        Add support for rc and res files.
+        """
+        return {
+            **super().out_extensions,
+            **{ext: ext + self.obj_extension for ext in ('.res', '.rc')},
+        }
+
+
+# the same as cygwin plus some additional parameters
+class Mingw32CCompiler(CygwinCCompiler):
+    """Handles the Mingw32 port of the GNU C compiler to Windows."""
+
+    compiler_type = 'mingw32'
+
+    def __init__(self, verbose=False, dry_run=False, force=False):
+        super().__init__(verbose, dry_run, force)
+
+        shared_option = "-shared"
+
+        if is_cygwincc(self.cc):
+            raise CCompilerError('Cygwin gcc cannot be used with --compiler=mingw32')
+
+        self.set_executables(
+            compiler=f'{self.cc} -O -Wall',
+            compiler_so=f'{self.cc} -shared -O -Wall',
+            compiler_so_cxx=f'{self.cxx} -shared -O -Wall',
+            compiler_cxx=f'{self.cxx} -O -Wall',
+            linker_exe=f'{self.cc}',
+            linker_so=f'{self.linker_dll} {shared_option}',
+            linker_exe_cxx=f'{self.cxx}',
+            linker_so_cxx=f'{self.linker_dll_cxx} {shared_option}',
+        )
+
+    def runtime_library_dir_option(self, dir):
+        raise DistutilsPlatformError(_runtime_library_dirs_msg)
+
+
+# Because these compilers aren't configured in Python's pyconfig.h file by
+# default, we should at least warn the user if he is using an unmodified
+# version.
+
+CONFIG_H_OK = "ok"
+CONFIG_H_NOTOK = "not ok"
+CONFIG_H_UNCERTAIN = "uncertain"
+
+
+def check_config_h():
+    """Check if the current Python installation appears amenable to building
+    extensions with GCC.
+
+    Returns a tuple (status, details), where 'status' is one of the following
+    constants:
+
+    - CONFIG_H_OK: all is well, go ahead and compile
+    - CONFIG_H_NOTOK: doesn't look good
+    - CONFIG_H_UNCERTAIN: not sure -- unable to read pyconfig.h
+
+    'details' is a human-readable string explaining the situation.
+
+    Note there are two ways to conclude "OK": either 'sys.version' contains
+    the string "GCC" (implying that this Python was built with GCC), or the
+    installed "pyconfig.h" contains the string "__GNUC__".
+    """
+
+    # XXX since this function also checks sys.version, it's not strictly a
+    # "pyconfig.h" check -- should probably be renamed...
+
+    from distutils import sysconfig
+
+    # if sys.version contains GCC then python was compiled with GCC, and the
+    # pyconfig.h file should be OK
+    if "GCC" in sys.version:
+        return CONFIG_H_OK, "sys.version mentions 'GCC'"
+
+    # Clang would also work
+    if "Clang" in sys.version:
+        return CONFIG_H_OK, "sys.version mentions 'Clang'"
+
+    # let's see if __GNUC__ is mentioned in python.h
+    fn = sysconfig.get_config_h_filename()
+    try:
+        config_h = pathlib.Path(fn).read_text(encoding='utf-8')
+    except OSError as exc:
+        return (CONFIG_H_UNCERTAIN, f"couldn't read '{fn}': {exc.strerror}")
+    else:
+        substring = '__GNUC__'
+        if substring in config_h:
+            code = CONFIG_H_OK
+            mention_inflected = 'mentions'
+        else:
+            code = CONFIG_H_NOTOK
+            mention_inflected = 'does not mention'
+        return code, f"{fn!r} {mention_inflected} {substring!r}"
+
+
+def is_cygwincc(cc):
+    """Try to determine if the compiler that would be used is from cygwin."""
+    out_string = check_output(shlex.split(cc) + ['-dumpmachine'])
+    return out_string.strip().endswith(b'cygwin')
+
+
+get_versions = None
+"""
+A stand-in for the previous get_versions() function to prevent failures
+when monkeypatched. See pypa/setuptools#2969.
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/debug.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/debug.py
new file mode 100644
index 00000000..daf1660f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/debug.py
@@ -0,0 +1,5 @@
+import os
+
+# If DISTUTILS_DEBUG is anything other than the empty string, we run in
+# debug mode.
+DEBUG = os.environ.get('DISTUTILS_DEBUG')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dep_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dep_util.py
new file mode 100644
index 00000000..09a8a2e1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dep_util.py
@@ -0,0 +1,14 @@
+import warnings
+
+from . import _modified
+
+
+def __getattr__(name):
+    if name not in ['newer', 'newer_group', 'newer_pairwise']:
+        raise AttributeError(name)
+    warnings.warn(
+        "dep_util is Deprecated. Use functions from setuptools instead.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    return getattr(_modified, name)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dir_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dir_util.py
new file mode 100644
index 00000000..d9782602
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dir_util.py
@@ -0,0 +1,244 @@
+"""distutils.dir_util
+
+Utility functions for manipulating directories and directory trees."""
+
+import functools
+import itertools
+import os
+import pathlib
+
+from . import file_util
+from ._log import log
+from .errors import DistutilsFileError, DistutilsInternalError
+
+
+class SkipRepeatAbsolutePaths(set):
+    """
+    Cache for mkpath.
+
+    In addition to cheapening redundant calls, eliminates redundant
+    "creating /foo/bar/baz" messages in dry-run mode.
+    """
+
+    def __init__(self):
+        SkipRepeatAbsolutePaths.instance = self
+
+    @classmethod
+    def clear(cls):
+        super(cls, cls.instance).clear()
+
+    def wrap(self, func):
+        @functools.wraps(func)
+        def wrapper(path, *args, **kwargs):
+            if path.absolute() in self:
+                return
+            result = func(path, *args, **kwargs)
+            self.add(path.absolute())
+            return result
+
+        return wrapper
+
+
+# Python 3.8 compatibility
+wrapper = SkipRepeatAbsolutePaths().wrap
+
+
+@functools.singledispatch
+@wrapper
+def mkpath(name: pathlib.Path, mode=0o777, verbose=True, dry_run=False) -> None:
+    """Create a directory and any missing ancestor directories.
+
+    If the directory already exists (or if 'name' is the empty string, which
+    means the current directory, which of course exists), then do nothing.
+    Raise DistutilsFileError if unable to create some directory along the way
+    (eg. some sub-path exists, but is a file rather than a directory).
+    If 'verbose' is true, log the directory created.
+    """
+    if verbose and not name.is_dir():
+        log.info("creating %s", name)
+
+    try:
+        dry_run or name.mkdir(mode=mode, parents=True, exist_ok=True)
+    except OSError as exc:
+        raise DistutilsFileError(f"could not create '{name}': {exc.args[-1]}")
+
+
+@mkpath.register
+def _(name: str, *args, **kwargs):
+    return mkpath(pathlib.Path(name), *args, **kwargs)
+
+
+@mkpath.register
+def _(name: None, *args, **kwargs):
+    """
+    Detect a common bug -- name is None.
+    """
+    raise DistutilsInternalError(f"mkpath: 'name' must be a string (got {name!r})")
+
+
+def create_tree(base_dir, files, mode=0o777, verbose=True, dry_run=False):
+    """Create all the empty directories under 'base_dir' needed to put 'files'
+    there.
+
+    'base_dir' is just the name of a directory which doesn't necessarily
+    exist yet; 'files' is a list of filenames to be interpreted relative to
+    'base_dir'.  'base_dir' + the directory portion of every file in 'files'
+    will be created if it doesn't already exist.  'mode', 'verbose' and
+    'dry_run' flags are as for 'mkpath()'.
+    """
+    # First get the list of directories to create
+    need_dir = set(os.path.join(base_dir, os.path.dirname(file)) for file in files)
+
+    # Now create them
+    for dir in sorted(need_dir):
+        mkpath(dir, mode, verbose=verbose, dry_run=dry_run)
+
+
+def copy_tree(
+    src,
+    dst,
+    preserve_mode=True,
+    preserve_times=True,
+    preserve_symlinks=False,
+    update=False,
+    verbose=True,
+    dry_run=False,
+):
+    """Copy an entire directory tree 'src' to a new location 'dst'.
+
+    Both 'src' and 'dst' must be directory names.  If 'src' is not a
+    directory, raise DistutilsFileError.  If 'dst' does not exist, it is
+    created with 'mkpath()'.  The end result of the copy is that every
+    file in 'src' is copied to 'dst', and directories under 'src' are
+    recursively copied to 'dst'.  Return the list of files that were
+    copied or might have been copied, using their output name.  The
+    return value is unaffected by 'update' or 'dry_run': it is simply
+    the list of all files under 'src', with the names changed to be
+    under 'dst'.
+
+    'preserve_mode' and 'preserve_times' are the same as for
+    'copy_file'; note that they only apply to regular files, not to
+    directories.  If 'preserve_symlinks' is true, symlinks will be
+    copied as symlinks (on platforms that support them!); otherwise
+    (the default), the destination of the symlink will be copied.
+    'update' and 'verbose' are the same as for 'copy_file'.
+    """
+    if not dry_run and not os.path.isdir(src):
+        raise DistutilsFileError(f"cannot copy tree '{src}': not a directory")
+    try:
+        names = os.listdir(src)
+    except OSError as e:
+        if dry_run:
+            names = []
+        else:
+            raise DistutilsFileError(f"error listing files in '{src}': {e.strerror}")
+
+    if not dry_run:
+        mkpath(dst, verbose=verbose)
+
+    copy_one = functools.partial(
+        _copy_one,
+        src=src,
+        dst=dst,
+        preserve_symlinks=preserve_symlinks,
+        verbose=verbose,
+        dry_run=dry_run,
+        preserve_mode=preserve_mode,
+        preserve_times=preserve_times,
+        update=update,
+    )
+    return list(itertools.chain.from_iterable(map(copy_one, names)))
+
+
+def _copy_one(
+    name,
+    *,
+    src,
+    dst,
+    preserve_symlinks,
+    verbose,
+    dry_run,
+    preserve_mode,
+    preserve_times,
+    update,
+):
+    src_name = os.path.join(src, name)
+    dst_name = os.path.join(dst, name)
+
+    if name.startswith('.nfs'):
+        # skip NFS rename files
+        return
+
+    if preserve_symlinks and os.path.islink(src_name):
+        link_dest = os.readlink(src_name)
+        if verbose >= 1:
+            log.info("linking %s -> %s", dst_name, link_dest)
+        if not dry_run:
+            os.symlink(link_dest, dst_name)
+        yield dst_name
+
+    elif os.path.isdir(src_name):
+        yield from copy_tree(
+            src_name,
+            dst_name,
+            preserve_mode,
+            preserve_times,
+            preserve_symlinks,
+            update,
+            verbose=verbose,
+            dry_run=dry_run,
+        )
+    else:
+        file_util.copy_file(
+            src_name,
+            dst_name,
+            preserve_mode,
+            preserve_times,
+            update,
+            verbose=verbose,
+            dry_run=dry_run,
+        )
+        yield dst_name
+
+
+def _build_cmdtuple(path, cmdtuples):
+    """Helper for remove_tree()."""
+    for f in os.listdir(path):
+        real_f = os.path.join(path, f)
+        if os.path.isdir(real_f) and not os.path.islink(real_f):
+            _build_cmdtuple(real_f, cmdtuples)
+        else:
+            cmdtuples.append((os.remove, real_f))
+    cmdtuples.append((os.rmdir, path))
+
+
+def remove_tree(directory, verbose=True, dry_run=False):
+    """Recursively remove an entire directory tree.
+
+    Any errors are ignored (apart from being reported to stdout if 'verbose'
+    is true).
+    """
+    if verbose >= 1:
+        log.info("removing '%s' (and everything under it)", directory)
+    if dry_run:
+        return
+    cmdtuples = []
+    _build_cmdtuple(directory, cmdtuples)
+    for cmd in cmdtuples:
+        try:
+            cmd[0](cmd[1])
+            # Clear the cache
+            SkipRepeatAbsolutePaths.clear()
+        except OSError as exc:
+            log.warning("error removing %s: %s", directory, exc)
+
+
+def ensure_relative(path):
+    """Take the full path 'path', and make it a relative path.
+
+    This is useful to make 'path' the second argument to os.path.join().
+    """
+    drive, path = os.path.splitdrive(path)
+    if path[0:1] == os.sep:
+        path = drive + path[1:]
+    return path
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dist.py
new file mode 100644
index 00000000..33ed8ebd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/dist.py
@@ -0,0 +1,1317 @@
+"""distutils.dist
+
+Provides the Distribution class, which represents the module distribution
+being built/installed/distributed.
+"""
+
+from __future__ import annotations
+
+import contextlib
+import logging
+import os
+import pathlib
+import re
+import sys
+import warnings
+from collections.abc import Iterable
+from email import message_from_file
+from typing import TYPE_CHECKING, Literal, TypeVar, overload
+
+from packaging.utils import canonicalize_name, canonicalize_version
+
+from ._log import log
+from .debug import DEBUG
+from .errors import (
+    DistutilsArgError,
+    DistutilsClassError,
+    DistutilsModuleError,
+    DistutilsOptionError,
+)
+from .fancy_getopt import FancyGetopt, translate_longopt
+from .util import check_environ, rfc822_escape, strtobool
+
+if TYPE_CHECKING:
+    # type-only import because of mutual dependence between these modules
+    from .cmd import Command
+
+_CommandT = TypeVar("_CommandT", bound="Command")
+
+# Regex to define acceptable Distutils command names.  This is not *quite*
+# the same as a Python NAME -- I don't allow leading underscores.  The fact
+# that they're very similar is no coincidence; the default naming scheme is
+# to look for a Python module named after the command.
+command_re = re.compile(r'^[a-zA-Z]([a-zA-Z0-9_]*)$')
+
+
+def _ensure_list(value, fieldname):
+    if isinstance(value, str):
+        # a string containing comma separated values is okay.  It will
+        # be converted to a list by Distribution.finalize_options().
+        pass
+    elif not isinstance(value, list):
+        # passing a tuple or an iterator perhaps, warn and convert
+        typename = type(value).__name__
+        msg = "Warning: '{fieldname}' should be a list, got type '{typename}'"
+        msg = msg.format(**locals())
+        log.warning(msg)
+        value = list(value)
+    return value
+
+
+class Distribution:
+    """The core of the Distutils.  Most of the work hiding behind 'setup'
+    is really done within a Distribution instance, which farms the work out
+    to the Distutils commands specified on the command line.
+
+    Setup scripts will almost never instantiate Distribution directly,
+    unless the 'setup()' function is totally inadequate to their needs.
+    However, it is conceivable that a setup script might wish to subclass
+    Distribution for some specialized purpose, and then pass the subclass
+    to 'setup()' as the 'distclass' keyword argument.  If so, it is
+    necessary to respect the expectations that 'setup' has of Distribution.
+    See the code for 'setup()', in core.py, for details.
+    """
+
+    # 'global_options' describes the command-line options that may be
+    # supplied to the setup script prior to any actual commands.
+    # Eg. "./setup.py -n" or "./setup.py --quiet" both take advantage of
+    # these global options.  This list should be kept to a bare minimum,
+    # since every global option is also valid as a command option -- and we
+    # don't want to pollute the commands with too many options that they
+    # have minimal control over.
+    # The fourth entry for verbose means that it can be repeated.
+    global_options = [
+        ('verbose', 'v', "run verbosely (default)", 1),
+        ('quiet', 'q', "run quietly (turns verbosity off)"),
+        ('dry-run', 'n', "don't actually do anything"),
+        ('help', 'h', "show detailed help message"),
+        ('no-user-cfg', None, 'ignore pydistutils.cfg in your home directory'),
+    ]
+
+    # 'common_usage' is a short (2-3 line) string describing the common
+    # usage of the setup script.
+    common_usage = """\
+Common commands: (see '--help-commands' for more)
+
+  setup.py build      will build the package underneath 'build/'
+  setup.py install    will install the package
+"""
+
+    # options that are not propagated to the commands
+    display_options = [
+        ('help-commands', None, "list all available commands"),
+        ('name', None, "print package name"),
+        ('version', 'V', "print package version"),
+        ('fullname', None, "print <package name>-<version>"),
+        ('author', None, "print the author's name"),
+        ('author-email', None, "print the author's email address"),
+        ('maintainer', None, "print the maintainer's name"),
+        ('maintainer-email', None, "print the maintainer's email address"),
+        ('contact', None, "print the maintainer's name if known, else the author's"),
+        (
+            'contact-email',
+            None,
+            "print the maintainer's email address if known, else the author's",
+        ),
+        ('url', None, "print the URL for this package"),
+        ('license', None, "print the license of the package"),
+        ('licence', None, "alias for --license"),
+        ('description', None, "print the package description"),
+        ('long-description', None, "print the long package description"),
+        ('platforms', None, "print the list of platforms"),
+        ('classifiers', None, "print the list of classifiers"),
+        ('keywords', None, "print the list of keywords"),
+        ('provides', None, "print the list of packages/modules provided"),
+        ('requires', None, "print the list of packages/modules required"),
+        ('obsoletes', None, "print the list of packages/modules made obsolete"),
+    ]
+    display_option_names = [translate_longopt(x[0]) for x in display_options]
+
+    # negative options are options that exclude other options
+    negative_opt = {'quiet': 'verbose'}
+
+    # -- Creation/initialization methods -------------------------------
+
+    def __init__(self, attrs=None):  # noqa: C901
+        """Construct a new Distribution instance: initialize all the
+        attributes of a Distribution, and then use 'attrs' (a dictionary
+        mapping attribute names to values) to assign some of those
+        attributes their "real" values.  (Any attributes not mentioned in
+        'attrs' will be assigned to some null value: 0, None, an empty list
+        or dictionary, etc.)  Most importantly, initialize the
+        'command_obj' attribute to the empty dictionary; this will be
+        filled in with real command objects by 'parse_command_line()'.
+        """
+
+        # Default values for our command-line options
+        self.verbose = True
+        self.dry_run = False
+        self.help = False
+        for attr in self.display_option_names:
+            setattr(self, attr, False)
+
+        # Store the distribution meta-data (name, version, author, and so
+        # forth) in a separate object -- we're getting to have enough
+        # information here (and enough command-line options) that it's
+        # worth it.  Also delegate 'get_XXX()' methods to the 'metadata'
+        # object in a sneaky and underhanded (but efficient!) way.
+        self.metadata = DistributionMetadata()
+        for basename in self.metadata._METHOD_BASENAMES:
+            method_name = "get_" + basename
+            setattr(self, method_name, getattr(self.metadata, method_name))
+
+        # 'cmdclass' maps command names to class objects, so we
+        # can 1) quickly figure out which class to instantiate when
+        # we need to create a new command object, and 2) have a way
+        # for the setup script to override command classes
+        self.cmdclass = {}
+
+        # 'command_packages' is a list of packages in which commands
+        # are searched for.  The factory for command 'foo' is expected
+        # to be named 'foo' in the module 'foo' in one of the packages
+        # named here.  This list is searched from the left; an error
+        # is raised if no named package provides the command being
+        # searched for.  (Always access using get_command_packages().)
+        self.command_packages = None
+
+        # 'script_name' and 'script_args' are usually set to sys.argv[0]
+        # and sys.argv[1:], but they can be overridden when the caller is
+        # not necessarily a setup script run from the command-line.
+        self.script_name = None
+        self.script_args: list[str] | None = None
+
+        # 'command_options' is where we store command options between
+        # parsing them (from config files, the command-line, etc.) and when
+        # they are actually needed -- ie. when the command in question is
+        # instantiated.  It is a dictionary of dictionaries of 2-tuples:
+        #   command_options = { command_name : { option : (source, value) } }
+        self.command_options = {}
+
+        # 'dist_files' is the list of (command, pyversion, file) that
+        # have been created by any dist commands run so far. This is
+        # filled regardless of whether the run is dry or not. pyversion
+        # gives sysconfig.get_python_version() if the dist file is
+        # specific to a Python version, 'any' if it is good for all
+        # Python versions on the target platform, and '' for a source
+        # file. pyversion should not be used to specify minimum or
+        # maximum required Python versions; use the metainfo for that
+        # instead.
+        self.dist_files = []
+
+        # These options are really the business of various commands, rather
+        # than of the Distribution itself.  We provide aliases for them in
+        # Distribution as a convenience to the developer.
+        self.packages = None
+        self.package_data = {}
+        self.package_dir = None
+        self.py_modules = None
+        self.libraries = None
+        self.headers = None
+        self.ext_modules = None
+        self.ext_package = None
+        self.include_dirs = None
+        self.extra_path = None
+        self.scripts = None
+        self.data_files = None
+        self.password = ''
+
+        # And now initialize bookkeeping stuff that can't be supplied by
+        # the caller at all.  'command_obj' maps command names to
+        # Command instances -- that's how we enforce that every command
+        # class is a singleton.
+        self.command_obj = {}
+
+        # 'have_run' maps command names to boolean values; it keeps track
+        # of whether we have actually run a particular command, to make it
+        # cheap to "run" a command whenever we think we might need to -- if
+        # it's already been done, no need for expensive filesystem
+        # operations, we just check the 'have_run' dictionary and carry on.
+        # It's only safe to query 'have_run' for a command class that has
+        # been instantiated -- a false value will be inserted when the
+        # command object is created, and replaced with a true value when
+        # the command is successfully run.  Thus it's probably best to use
+        # '.get()' rather than a straight lookup.
+        self.have_run = {}
+
+        # Now we'll use the attrs dictionary (ultimately, keyword args from
+        # the setup script) to possibly override any or all of these
+        # distribution options.
+
+        if attrs:
+            # Pull out the set of command options and work on them
+            # specifically.  Note that this order guarantees that aliased
+            # command options will override any supplied redundantly
+            # through the general options dictionary.
+            options = attrs.get('options')
+            if options is not None:
+                del attrs['options']
+                for command, cmd_options in options.items():
+                    opt_dict = self.get_option_dict(command)
+                    for opt, val in cmd_options.items():
+                        opt_dict[opt] = ("setup script", val)
+
+            if 'licence' in attrs:
+                attrs['license'] = attrs['licence']
+                del attrs['licence']
+                msg = "'licence' distribution option is deprecated; use 'license'"
+                warnings.warn(msg)
+
+            # Now work on the rest of the attributes.  Any attribute that's
+            # not already defined is invalid!
+            for key, val in attrs.items():
+                if hasattr(self.metadata, "set_" + key):
+                    getattr(self.metadata, "set_" + key)(val)
+                elif hasattr(self.metadata, key):
+                    setattr(self.metadata, key, val)
+                elif hasattr(self, key):
+                    setattr(self, key, val)
+                else:
+                    msg = f"Unknown distribution option: {key!r}"
+                    warnings.warn(msg)
+
+        # no-user-cfg is handled before other command line args
+        # because other args override the config files, and this
+        # one is needed before we can load the config files.
+        # If attrs['script_args'] wasn't passed, assume false.
+        #
+        # This also make sure we just look at the global options
+        self.want_user_cfg = True
+
+        if self.script_args is not None:
+            # Coerce any possible iterable from attrs into a list
+            self.script_args = list(self.script_args)
+            for arg in self.script_args:
+                if not arg.startswith('-'):
+                    break
+                if arg == '--no-user-cfg':
+                    self.want_user_cfg = False
+                    break
+
+        self.finalize_options()
+
+    def get_option_dict(self, command):
+        """Get the option dictionary for a given command.  If that
+        command's option dictionary hasn't been created yet, then create it
+        and return the new dictionary; otherwise, return the existing
+        option dictionary.
+        """
+        dict = self.command_options.get(command)
+        if dict is None:
+            dict = self.command_options[command] = {}
+        return dict
+
+    def dump_option_dicts(self, header=None, commands=None, indent=""):
+        from pprint import pformat
+
+        if commands is None:  # dump all command option dicts
+            commands = sorted(self.command_options.keys())
+
+        if header is not None:
+            self.announce(indent + header)
+            indent = indent + "  "
+
+        if not commands:
+            self.announce(indent + "no commands known yet")
+            return
+
+        for cmd_name in commands:
+            opt_dict = self.command_options.get(cmd_name)
+            if opt_dict is None:
+                self.announce(indent + f"no option dict for '{cmd_name}' command")
+            else:
+                self.announce(indent + f"option dict for '{cmd_name}' command:")
+                out = pformat(opt_dict)
+                for line in out.split('\n'):
+                    self.announce(indent + "  " + line)
+
+    # -- Config file finding/parsing methods ---------------------------
+
+    def find_config_files(self):
+        """Find as many configuration files as should be processed for this
+        platform, and return a list of filenames in the order in which they
+        should be parsed.  The filenames returned are guaranteed to exist
+        (modulo nasty race conditions).
+
+        There are multiple possible config files:
+        - distutils.cfg in the Distutils installation directory (i.e.
+          where the top-level Distutils __inst__.py file lives)
+        - a file in the user's home directory named .pydistutils.cfg
+          on Unix and pydistutils.cfg on Windows/Mac; may be disabled
+          with the ``--no-user-cfg`` option
+        - setup.cfg in the current directory
+        - a file named by an environment variable
+        """
+        check_environ()
+        files = [str(path) for path in self._gen_paths() if os.path.isfile(path)]
+
+        if DEBUG:
+            self.announce("using config files: {}".format(', '.join(files)))
+
+        return files
+
+    def _gen_paths(self):
+        # The system-wide Distutils config file
+        sys_dir = pathlib.Path(sys.modules['distutils'].__file__).parent
+        yield sys_dir / "distutils.cfg"
+
+        # The per-user config file
+        prefix = '.' * (os.name == 'posix')
+        filename = prefix + 'pydistutils.cfg'
+        if self.want_user_cfg:
+            with contextlib.suppress(RuntimeError):
+                yield pathlib.Path('~').expanduser() / filename
+
+        # All platforms support local setup.cfg
+        yield pathlib.Path('setup.cfg')
+
+        # Additional config indicated in the environment
+        with contextlib.suppress(TypeError):
+            yield pathlib.Path(os.getenv("DIST_EXTRA_CONFIG"))
+
+    def parse_config_files(self, filenames=None):  # noqa: C901
+        from configparser import ConfigParser
+
+        # Ignore install directory options if we have a venv
+        if sys.prefix != sys.base_prefix:
+            ignore_options = [
+                'install-base',
+                'install-platbase',
+                'install-lib',
+                'install-platlib',
+                'install-purelib',
+                'install-headers',
+                'install-scripts',
+                'install-data',
+                'prefix',
+                'exec-prefix',
+                'home',
+                'user',
+                'root',
+            ]
+        else:
+            ignore_options = []
+
+        ignore_options = frozenset(ignore_options)
+
+        if filenames is None:
+            filenames = self.find_config_files()
+
+        if DEBUG:
+            self.announce("Distribution.parse_config_files():")
+
+        parser = ConfigParser()
+        for filename in filenames:
+            if DEBUG:
+                self.announce(f"  reading {filename}")
+            parser.read(filename, encoding='utf-8')
+            for section in parser.sections():
+                options = parser.options(section)
+                opt_dict = self.get_option_dict(section)
+
+                for opt in options:
+                    if opt != '__name__' and opt not in ignore_options:
+                        val = parser.get(section, opt)
+                        opt = opt.replace('-', '_')
+                        opt_dict[opt] = (filename, val)
+
+            # Make the ConfigParser forget everything (so we retain
+            # the original filenames that options come from)
+            parser.__init__()
+
+        # If there was a "global" section in the config file, use it
+        # to set Distribution options.
+
+        if 'global' in self.command_options:
+            for opt, (_src, val) in self.command_options['global'].items():
+                alias = self.negative_opt.get(opt)
+                try:
+                    if alias:
+                        setattr(self, alias, not strtobool(val))
+                    elif opt in ('verbose', 'dry_run'):  # ugh!
+                        setattr(self, opt, strtobool(val))
+                    else:
+                        setattr(self, opt, val)
+                except ValueError as msg:
+                    raise DistutilsOptionError(msg)
+
+    # -- Command-line parsing methods ----------------------------------
+
+    def parse_command_line(self):
+        """Parse the setup script's command line, taken from the
+        'script_args' instance attribute (which defaults to 'sys.argv[1:]'
+        -- see 'setup()' in core.py).  This list is first processed for
+        "global options" -- options that set attributes of the Distribution
+        instance.  Then, it is alternately scanned for Distutils commands
+        and options for that command.  Each new command terminates the
+        options for the previous command.  The allowed options for a
+        command are determined by the 'user_options' attribute of the
+        command class -- thus, we have to be able to load command classes
+        in order to parse the command line.  Any error in that 'options'
+        attribute raises DistutilsGetoptError; any error on the
+        command-line raises DistutilsArgError.  If no Distutils commands
+        were found on the command line, raises DistutilsArgError.  Return
+        true if command-line was successfully parsed and we should carry
+        on with executing commands; false if no errors but we shouldn't
+        execute commands (currently, this only happens if user asks for
+        help).
+        """
+        #
+        # We now have enough information to show the Macintosh dialog
+        # that allows the user to interactively specify the "command line".
+        #
+        toplevel_options = self._get_toplevel_options()
+
+        # We have to parse the command line a bit at a time -- global
+        # options, then the first command, then its options, and so on --
+        # because each command will be handled by a different class, and
+        # the options that are valid for a particular class aren't known
+        # until we have loaded the command class, which doesn't happen
+        # until we know what the command is.
+
+        self.commands = []
+        parser = FancyGetopt(toplevel_options + self.display_options)
+        parser.set_negative_aliases(self.negative_opt)
+        parser.set_aliases({'licence': 'license'})
+        args = parser.getopt(args=self.script_args, object=self)
+        option_order = parser.get_option_order()
+        logging.getLogger().setLevel(logging.WARN - 10 * self.verbose)
+
+        # for display options we return immediately
+        if self.handle_display_options(option_order):
+            return
+        while args:
+            args = self._parse_command_opts(parser, args)
+            if args is None:  # user asked for help (and got it)
+                return
+
+        # Handle the cases of --help as a "global" option, ie.
+        # "setup.py --help" and "setup.py --help command ...".  For the
+        # former, we show global options (--verbose, --dry-run, etc.)
+        # and display-only options (--name, --version, etc.); for the
+        # latter, we omit the display-only options and show help for
+        # each command listed on the command line.
+        if self.help:
+            self._show_help(
+                parser, display_options=len(self.commands) == 0, commands=self.commands
+            )
+            return
+
+        # Oops, no commands found -- an end-user error
+        if not self.commands:
+            raise DistutilsArgError("no commands supplied")
+
+        # All is well: return true
+        return True
+
+    def _get_toplevel_options(self):
+        """Return the non-display options recognized at the top level.
+
+        This includes options that are recognized *only* at the top
+        level as well as options recognized for commands.
+        """
+        return self.global_options + [
+            (
+                "command-packages=",
+                None,
+                "list of packages that provide distutils commands",
+            ),
+        ]
+
+    def _parse_command_opts(self, parser, args):  # noqa: C901
+        """Parse the command-line options for a single command.
+        'parser' must be a FancyGetopt instance; 'args' must be the list
+        of arguments, starting with the current command (whose options
+        we are about to parse).  Returns a new version of 'args' with
+        the next command at the front of the list; will be the empty
+        list if there are no more commands on the command line.  Returns
+        None if the user asked for help on this command.
+        """
+        # late import because of mutual dependence between these modules
+        from distutils.cmd import Command
+
+        # Pull the current command from the head of the command line
+        command = args[0]
+        if not command_re.match(command):
+            raise SystemExit(f"invalid command name '{command}'")
+        self.commands.append(command)
+
+        # Dig up the command class that implements this command, so we
+        # 1) know that it's a valid command, and 2) know which options
+        # it takes.
+        try:
+            cmd_class = self.get_command_class(command)
+        except DistutilsModuleError as msg:
+            raise DistutilsArgError(msg)
+
+        # Require that the command class be derived from Command -- want
+        # to be sure that the basic "command" interface is implemented.
+        if not issubclass(cmd_class, Command):
+            raise DistutilsClassError(
+                f"command class {cmd_class} must subclass Command"
+            )
+
+        # Also make sure that the command object provides a list of its
+        # known options.
+        if not (
+            hasattr(cmd_class, 'user_options')
+            and isinstance(cmd_class.user_options, list)
+        ):
+            msg = (
+                "command class %s must provide "
+                "'user_options' attribute (a list of tuples)"
+            )
+            raise DistutilsClassError(msg % cmd_class)
+
+        # If the command class has a list of negative alias options,
+        # merge it in with the global negative aliases.
+        negative_opt = self.negative_opt
+        if hasattr(cmd_class, 'negative_opt'):
+            negative_opt = negative_opt.copy()
+            negative_opt.update(cmd_class.negative_opt)
+
+        # Check for help_options in command class.  They have a different
+        # format (tuple of four) so we need to preprocess them here.
+        if hasattr(cmd_class, 'help_options') and isinstance(
+            cmd_class.help_options, list
+        ):
+            help_options = fix_help_options(cmd_class.help_options)
+        else:
+            help_options = []
+
+        # All commands support the global options too, just by adding
+        # in 'global_options'.
+        parser.set_option_table(
+            self.global_options + cmd_class.user_options + help_options
+        )
+        parser.set_negative_aliases(negative_opt)
+        (args, opts) = parser.getopt(args[1:])
+        if hasattr(opts, 'help') and opts.help:
+            self._show_help(parser, display_options=False, commands=[cmd_class])
+            return
+
+        if hasattr(cmd_class, 'help_options') and isinstance(
+            cmd_class.help_options, list
+        ):
+            help_option_found = 0
+            for help_option, _short, _desc, func in cmd_class.help_options:
+                if hasattr(opts, parser.get_attr_name(help_option)):
+                    help_option_found = 1
+                    if callable(func):
+                        func()
+                    else:
+                        raise DistutilsClassError(
+                            f"invalid help function {func!r} for help option '{help_option}': "
+                            "must be a callable object (function, etc.)"
+                        )
+
+            if help_option_found:
+                return
+
+        # Put the options from the command-line into their official
+        # holding pen, the 'command_options' dictionary.
+        opt_dict = self.get_option_dict(command)
+        for name, value in vars(opts).items():
+            opt_dict[name] = ("command line", value)
+
+        return args
+
+    def finalize_options(self):
+        """Set final values for all the options on the Distribution
+        instance, analogous to the .finalize_options() method of Command
+        objects.
+        """
+        for attr in ('keywords', 'platforms'):
+            value = getattr(self.metadata, attr)
+            if value is None:
+                continue
+            if isinstance(value, str):
+                value = [elm.strip() for elm in value.split(',')]
+                setattr(self.metadata, attr, value)
+
+    def _show_help(
+        self, parser, global_options=True, display_options=True, commands: Iterable = ()
+    ):
+        """Show help for the setup script command-line in the form of
+        several lists of command-line options.  'parser' should be a
+        FancyGetopt instance; do not expect it to be returned in the
+        same state, as its option table will be reset to make it
+        generate the correct help text.
+
+        If 'global_options' is true, lists the global options:
+        --verbose, --dry-run, etc.  If 'display_options' is true, lists
+        the "display-only" options: --name, --version, etc.  Finally,
+        lists per-command help for every command name or command class
+        in 'commands'.
+        """
+        # late import because of mutual dependence between these modules
+        from distutils.cmd import Command
+        from distutils.core import gen_usage
+
+        if global_options:
+            if display_options:
+                options = self._get_toplevel_options()
+            else:
+                options = self.global_options
+            parser.set_option_table(options)
+            parser.print_help(self.common_usage + "\nGlobal options:")
+            print()
+
+        if display_options:
+            parser.set_option_table(self.display_options)
+            parser.print_help(
+                "Information display options (just display information, ignore any commands)"
+            )
+            print()
+
+        for command in commands:
+            if isinstance(command, type) and issubclass(command, Command):
+                klass = command
+            else:
+                klass = self.get_command_class(command)
+            if hasattr(klass, 'help_options') and isinstance(klass.help_options, list):
+                parser.set_option_table(
+                    klass.user_options + fix_help_options(klass.help_options)
+                )
+            else:
+                parser.set_option_table(klass.user_options)
+            parser.print_help(f"Options for '{klass.__name__}' command:")
+            print()
+
+        print(gen_usage(self.script_name))
+
+    def handle_display_options(self, option_order):
+        """If there were any non-global "display-only" options
+        (--help-commands or the metadata display options) on the command
+        line, display the requested info and return true; else return
+        false.
+        """
+        from distutils.core import gen_usage
+
+        # User just wants a list of commands -- we'll print it out and stop
+        # processing now (ie. if they ran "setup --help-commands foo bar",
+        # we ignore "foo bar").
+        if self.help_commands:
+            self.print_commands()
+            print()
+            print(gen_usage(self.script_name))
+            return 1
+
+        # If user supplied any of the "display metadata" options, then
+        # display that metadata in the order in which the user supplied the
+        # metadata options.
+        any_display_options = 0
+        is_display_option = set()
+        for option in self.display_options:
+            is_display_option.add(option[0])
+
+        for opt, val in option_order:
+            if val and opt in is_display_option:
+                opt = translate_longopt(opt)
+                value = getattr(self.metadata, "get_" + opt)()
+                if opt in ('keywords', 'platforms'):
+                    print(','.join(value))
+                elif opt in ('classifiers', 'provides', 'requires', 'obsoletes'):
+                    print('\n'.join(value))
+                else:
+                    print(value)
+                any_display_options = 1
+
+        return any_display_options
+
+    def print_command_list(self, commands, header, max_length):
+        """Print a subset of the list of all commands -- used by
+        'print_commands()'.
+        """
+        print(header + ":")
+
+        for cmd in commands:
+            klass = self.cmdclass.get(cmd)
+            if not klass:
+                klass = self.get_command_class(cmd)
+            try:
+                description = klass.description
+            except AttributeError:
+                description = "(no description available)"
+
+            print(f"  {cmd:<{max_length}}  {description}")
+
+    def print_commands(self):
+        """Print out a help message listing all available commands with a
+        description of each.  The list is divided into "standard commands"
+        (listed in distutils.command.__all__) and "extra commands"
+        (mentioned in self.cmdclass, but not a standard command).  The
+        descriptions come from the command class attribute
+        'description'.
+        """
+        import distutils.command
+
+        std_commands = distutils.command.__all__
+        is_std = set(std_commands)
+
+        extra_commands = [cmd for cmd in self.cmdclass.keys() if cmd not in is_std]
+
+        max_length = 0
+        for cmd in std_commands + extra_commands:
+            if len(cmd) > max_length:
+                max_length = len(cmd)
+
+        self.print_command_list(std_commands, "Standard commands", max_length)
+        if extra_commands:
+            print()
+            self.print_command_list(extra_commands, "Extra commands", max_length)
+
+    def get_command_list(self):
+        """Get a list of (command, description) tuples.
+        The list is divided into "standard commands" (listed in
+        distutils.command.__all__) and "extra commands" (mentioned in
+        self.cmdclass, but not a standard command).  The descriptions come
+        from the command class attribute 'description'.
+        """
+        # Currently this is only used on Mac OS, for the Mac-only GUI
+        # Distutils interface (by Jack Jansen)
+        import distutils.command
+
+        std_commands = distutils.command.__all__
+        is_std = set(std_commands)
+
+        extra_commands = [cmd for cmd in self.cmdclass.keys() if cmd not in is_std]
+
+        rv = []
+        for cmd in std_commands + extra_commands:
+            klass = self.cmdclass.get(cmd)
+            if not klass:
+                klass = self.get_command_class(cmd)
+            try:
+                description = klass.description
+            except AttributeError:
+                description = "(no description available)"
+            rv.append((cmd, description))
+        return rv
+
+    # -- Command class/object methods ----------------------------------
+
+    def get_command_packages(self):
+        """Return a list of packages from which commands are loaded."""
+        pkgs = self.command_packages
+        if not isinstance(pkgs, list):
+            if pkgs is None:
+                pkgs = ''
+            pkgs = [pkg.strip() for pkg in pkgs.split(',') if pkg != '']
+            if "distutils.command" not in pkgs:
+                pkgs.insert(0, "distutils.command")
+            self.command_packages = pkgs
+        return pkgs
+
+    def get_command_class(self, command):
+        """Return the class that implements the Distutils command named by
+        'command'.  First we check the 'cmdclass' dictionary; if the
+        command is mentioned there, we fetch the class object from the
+        dictionary and return it.  Otherwise we load the command module
+        ("distutils.command." + command) and fetch the command class from
+        the module.  The loaded class is also stored in 'cmdclass'
+        to speed future calls to 'get_command_class()'.
+
+        Raises DistutilsModuleError if the expected module could not be
+        found, or if that module does not define the expected class.
+        """
+        klass = self.cmdclass.get(command)
+        if klass:
+            return klass
+
+        for pkgname in self.get_command_packages():
+            module_name = f"{pkgname}.{command}"
+            klass_name = command
+
+            try:
+                __import__(module_name)
+                module = sys.modules[module_name]
+            except ImportError:
+                continue
+
+            try:
+                klass = getattr(module, klass_name)
+            except AttributeError:
+                raise DistutilsModuleError(
+                    f"invalid command '{command}' (no class '{klass_name}' in module '{module_name}')"
+                )
+
+            self.cmdclass[command] = klass
+            return klass
+
+        raise DistutilsModuleError(f"invalid command '{command}'")
+
+    @overload
+    def get_command_obj(
+        self, command: str, create: Literal[True] = True
+    ) -> Command: ...
+    @overload
+    def get_command_obj(
+        self, command: str, create: Literal[False]
+    ) -> Command | None: ...
+    def get_command_obj(self, command: str, create: bool = True) -> Command | None:
+        """Return the command object for 'command'.  Normally this object
+        is cached on a previous call to 'get_command_obj()'; if no command
+        object for 'command' is in the cache, then we either create and
+        return it (if 'create' is true) or return None.
+        """
+        cmd_obj = self.command_obj.get(command)
+        if not cmd_obj and create:
+            if DEBUG:
+                self.announce(
+                    "Distribution.get_command_obj(): "
+                    f"creating '{command}' command object"
+                )
+
+            klass = self.get_command_class(command)
+            cmd_obj = self.command_obj[command] = klass(self)
+            self.have_run[command] = False
+
+            # Set any options that were supplied in config files
+            # or on the command line.  (NB. support for error
+            # reporting is lame here: any errors aren't reported
+            # until 'finalize_options()' is called, which means
+            # we won't report the source of the error.)
+            options = self.command_options.get(command)
+            if options:
+                self._set_command_options(cmd_obj, options)
+
+        return cmd_obj
+
+    def _set_command_options(self, command_obj, option_dict=None):  # noqa: C901
+        """Set the options for 'command_obj' from 'option_dict'.  Basically
+        this means copying elements of a dictionary ('option_dict') to
+        attributes of an instance ('command').
+
+        'command_obj' must be a Command instance.  If 'option_dict' is not
+        supplied, uses the standard option dictionary for this command
+        (from 'self.command_options').
+        """
+        command_name = command_obj.get_command_name()
+        if option_dict is None:
+            option_dict = self.get_option_dict(command_name)
+
+        if DEBUG:
+            self.announce(f"  setting options for '{command_name}' command:")
+        for option, (source, value) in option_dict.items():
+            if DEBUG:
+                self.announce(f"    {option} = {value} (from {source})")
+            try:
+                bool_opts = [translate_longopt(o) for o in command_obj.boolean_options]
+            except AttributeError:
+                bool_opts = []
+            try:
+                neg_opt = command_obj.negative_opt
+            except AttributeError:
+                neg_opt = {}
+
+            try:
+                is_string = isinstance(value, str)
+                if option in neg_opt and is_string:
+                    setattr(command_obj, neg_opt[option], not strtobool(value))
+                elif option in bool_opts and is_string:
+                    setattr(command_obj, option, strtobool(value))
+                elif hasattr(command_obj, option):
+                    setattr(command_obj, option, value)
+                else:
+                    raise DistutilsOptionError(
+                        f"error in {source}: command '{command_name}' has no such option '{option}'"
+                    )
+            except ValueError as msg:
+                raise DistutilsOptionError(msg)
+
+    @overload
+    def reinitialize_command(
+        self, command: str, reinit_subcommands: bool = False
+    ) -> Command: ...
+    @overload
+    def reinitialize_command(
+        self, command: _CommandT, reinit_subcommands: bool = False
+    ) -> _CommandT: ...
+    def reinitialize_command(
+        self, command: str | Command, reinit_subcommands=False
+    ) -> Command:
+        """Reinitializes a command to the state it was in when first
+        returned by 'get_command_obj()': ie., initialized but not yet
+        finalized.  This provides the opportunity to sneak option
+        values in programmatically, overriding or supplementing
+        user-supplied values from the config files and command line.
+        You'll have to re-finalize the command object (by calling
+        'finalize_options()' or 'ensure_finalized()') before using it for
+        real.
+
+        'command' should be a command name (string) or command object.  If
+        'reinit_subcommands' is true, also reinitializes the command's
+        sub-commands, as declared by the 'sub_commands' class attribute (if
+        it has one).  See the "install" command for an example.  Only
+        reinitializes the sub-commands that actually matter, ie. those
+        whose test predicates return true.
+
+        Returns the reinitialized command object.
+        """
+        from distutils.cmd import Command
+
+        if not isinstance(command, Command):
+            command_name = command
+            command = self.get_command_obj(command_name)
+        else:
+            command_name = command.get_command_name()
+
+        if not command.finalized:
+            return command
+        command.initialize_options()
+        command.finalized = False
+        self.have_run[command_name] = False
+        self._set_command_options(command)
+
+        if reinit_subcommands:
+            for sub in command.get_sub_commands():
+                self.reinitialize_command(sub, reinit_subcommands)
+
+        return command
+
+    # -- Methods that operate on the Distribution ----------------------
+
+    def announce(self, msg, level=logging.INFO):
+        log.log(level, msg)
+
+    def run_commands(self):
+        """Run each command that was seen on the setup script command line.
+        Uses the list of commands found and cache of command objects
+        created by 'get_command_obj()'.
+        """
+        for cmd in self.commands:
+            self.run_command(cmd)
+
+    # -- Methods that operate on its Commands --------------------------
+
+    def run_command(self, command):
+        """Do whatever it takes to run a command (including nothing at all,
+        if the command has already been run).  Specifically: if we have
+        already created and run the command named by 'command', return
+        silently without doing anything.  If the command named by 'command'
+        doesn't even have a command object yet, create one.  Then invoke
+        'run()' on that command object (or an existing one).
+        """
+        # Already been here, done that? then return silently.
+        if self.have_run.get(command):
+            return
+
+        log.info("running %s", command)
+        cmd_obj = self.get_command_obj(command)
+        cmd_obj.ensure_finalized()
+        cmd_obj.run()
+        self.have_run[command] = True
+
+    # -- Distribution query methods ------------------------------------
+
+    def has_pure_modules(self):
+        return len(self.packages or self.py_modules or []) > 0
+
+    def has_ext_modules(self):
+        return self.ext_modules and len(self.ext_modules) > 0
+
+    def has_c_libraries(self):
+        return self.libraries and len(self.libraries) > 0
+
+    def has_modules(self):
+        return self.has_pure_modules() or self.has_ext_modules()
+
+    def has_headers(self):
+        return self.headers and len(self.headers) > 0
+
+    def has_scripts(self):
+        return self.scripts and len(self.scripts) > 0
+
+    def has_data_files(self):
+        return self.data_files and len(self.data_files) > 0
+
+    def is_pure(self):
+        return (
+            self.has_pure_modules()
+            and not self.has_ext_modules()
+            and not self.has_c_libraries()
+        )
+
+    # -- Metadata query methods ----------------------------------------
+
+    # If you're looking for 'get_name()', 'get_version()', and so forth,
+    # they are defined in a sneaky way: the constructor binds self.get_XXX
+    # to self.metadata.get_XXX.  The actual code is in the
+    # DistributionMetadata class, below.
+
+
+class DistributionMetadata:
+    """Dummy class to hold the distribution meta-data: name, version,
+    author, and so forth.
+    """
+
+    _METHOD_BASENAMES = (
+        "name",
+        "version",
+        "author",
+        "author_email",
+        "maintainer",
+        "maintainer_email",
+        "url",
+        "license",
+        "description",
+        "long_description",
+        "keywords",
+        "platforms",
+        "fullname",
+        "contact",
+        "contact_email",
+        "classifiers",
+        "download_url",
+        # PEP 314
+        "provides",
+        "requires",
+        "obsoletes",
+    )
+
+    def __init__(self, path=None):
+        if path is not None:
+            self.read_pkg_file(open(path))
+        else:
+            self.name = None
+            self.version = None
+            self.author = None
+            self.author_email = None
+            self.maintainer = None
+            self.maintainer_email = None
+            self.url = None
+            self.license = None
+            self.description = None
+            self.long_description = None
+            self.keywords = None
+            self.platforms = None
+            self.classifiers = None
+            self.download_url = None
+            # PEP 314
+            self.provides = None
+            self.requires = None
+            self.obsoletes = None
+
+    def read_pkg_file(self, file):
+        """Reads the metadata values from a file object."""
+        msg = message_from_file(file)
+
+        def _read_field(name):
+            value = msg[name]
+            if value and value != "UNKNOWN":
+                return value
+
+        def _read_list(name):
+            values = msg.get_all(name, None)
+            if values == []:
+                return None
+            return values
+
+        metadata_version = msg['metadata-version']
+        self.name = _read_field('name')
+        self.version = _read_field('version')
+        self.description = _read_field('summary')
+        # we are filling author only.
+        self.author = _read_field('author')
+        self.maintainer = None
+        self.author_email = _read_field('author-email')
+        self.maintainer_email = None
+        self.url = _read_field('home-page')
+        self.license = _read_field('license')
+
+        if 'download-url' in msg:
+            self.download_url = _read_field('download-url')
+        else:
+            self.download_url = None
+
+        self.long_description = _read_field('description')
+        self.description = _read_field('summary')
+
+        if 'keywords' in msg:
+            self.keywords = _read_field('keywords').split(',')
+
+        self.platforms = _read_list('platform')
+        self.classifiers = _read_list('classifier')
+
+        # PEP 314 - these fields only exist in 1.1
+        if metadata_version == '1.1':
+            self.requires = _read_list('requires')
+            self.provides = _read_list('provides')
+            self.obsoletes = _read_list('obsoletes')
+        else:
+            self.requires = None
+            self.provides = None
+            self.obsoletes = None
+
+    def write_pkg_info(self, base_dir):
+        """Write the PKG-INFO file into the release tree."""
+        with open(
+            os.path.join(base_dir, 'PKG-INFO'), 'w', encoding='UTF-8'
+        ) as pkg_info:
+            self.write_pkg_file(pkg_info)
+
+    def write_pkg_file(self, file):
+        """Write the PKG-INFO format data to a file object."""
+        version = '1.0'
+        if (
+            self.provides
+            or self.requires
+            or self.obsoletes
+            or self.classifiers
+            or self.download_url
+        ):
+            version = '1.1'
+
+        # required fields
+        file.write(f'Metadata-Version: {version}\n')
+        file.write(f'Name: {self.get_name()}\n')
+        file.write(f'Version: {self.get_version()}\n')
+
+        def maybe_write(header, val):
+            if val:
+                file.write(f"{header}: {val}\n")
+
+        # optional fields
+        maybe_write("Summary", self.get_description())
+        maybe_write("Home-page", self.get_url())
+        maybe_write("Author", self.get_contact())
+        maybe_write("Author-email", self.get_contact_email())
+        maybe_write("License", self.get_license())
+        maybe_write("Download-URL", self.download_url)
+        maybe_write("Description", rfc822_escape(self.get_long_description() or ""))
+        maybe_write("Keywords", ",".join(self.get_keywords()))
+
+        self._write_list(file, 'Platform', self.get_platforms())
+        self._write_list(file, 'Classifier', self.get_classifiers())
+
+        # PEP 314
+        self._write_list(file, 'Requires', self.get_requires())
+        self._write_list(file, 'Provides', self.get_provides())
+        self._write_list(file, 'Obsoletes', self.get_obsoletes())
+
+    def _write_list(self, file, name, values):
+        values = values or []
+        for value in values:
+            file.write(f'{name}: {value}\n')
+
+    # -- Metadata query methods ----------------------------------------
+
+    def get_name(self):
+        return self.name or "UNKNOWN"
+
+    def get_version(self):
+        return self.version or "0.0.0"
+
+    def get_fullname(self):
+        return self._fullname(self.get_name(), self.get_version())
+
+    @staticmethod
+    def _fullname(name: str, version: str) -> str:
+        """
+        >>> DistributionMetadata._fullname('setup.tools', '1.0-2')
+        'setup_tools-1.0.post2'
+        >>> DistributionMetadata._fullname('setup-tools', '1.2post2')
+        'setup_tools-1.2.post2'
+        >>> DistributionMetadata._fullname('setup-tools', '1.0-r2')
+        'setup_tools-1.0.post2'
+        >>> DistributionMetadata._fullname('setup.tools', '1.0.post')
+        'setup_tools-1.0.post0'
+        >>> DistributionMetadata._fullname('setup.tools', '1.0+ubuntu-1')
+        'setup_tools-1.0+ubuntu.1'
+        """
+        return "{}-{}".format(
+            canonicalize_name(name).replace('-', '_'),
+            canonicalize_version(version, strip_trailing_zero=False),
+        )
+
+    def get_author(self):
+        return self.author
+
+    def get_author_email(self):
+        return self.author_email
+
+    def get_maintainer(self):
+        return self.maintainer
+
+    def get_maintainer_email(self):
+        return self.maintainer_email
+
+    def get_contact(self):
+        return self.maintainer or self.author
+
+    def get_contact_email(self):
+        return self.maintainer_email or self.author_email
+
+    def get_url(self):
+        return self.url
+
+    def get_license(self):
+        return self.license
+
+    get_licence = get_license
+
+    def get_description(self):
+        return self.description
+
+    def get_long_description(self):
+        return self.long_description
+
+    def get_keywords(self):
+        return self.keywords or []
+
+    def set_keywords(self, value):
+        self.keywords = _ensure_list(value, 'keywords')
+
+    def get_platforms(self):
+        return self.platforms
+
+    def set_platforms(self, value):
+        self.platforms = _ensure_list(value, 'platforms')
+
+    def get_classifiers(self):
+        return self.classifiers or []
+
+    def set_classifiers(self, value):
+        self.classifiers = _ensure_list(value, 'classifiers')
+
+    def get_download_url(self):
+        return self.download_url
+
+    # PEP 314
+    def get_requires(self):
+        return self.requires or []
+
+    def set_requires(self, value):
+        import distutils.versionpredicate
+
+        for v in value:
+            distutils.versionpredicate.VersionPredicate(v)
+        self.requires = list(value)
+
+    def get_provides(self):
+        return self.provides or []
+
+    def set_provides(self, value):
+        value = [v.strip() for v in value]
+        for v in value:
+            import distutils.versionpredicate
+
+            distutils.versionpredicate.split_provision(v)
+        self.provides = value
+
+    def get_obsoletes(self):
+        return self.obsoletes or []
+
+    def set_obsoletes(self, value):
+        import distutils.versionpredicate
+
+        for v in value:
+            distutils.versionpredicate.VersionPredicate(v)
+        self.obsoletes = list(value)
+
+
+def fix_help_options(options):
+    """Convert a 4-tuple 'help_options' list as found in various command
+    classes to the 3-tuple form required by FancyGetopt.
+    """
+    return [opt[0:3] for opt in options]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/errors.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/errors.py
new file mode 100644
index 00000000..3196a4f0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/errors.py
@@ -0,0 +1,124 @@
+"""
+Exceptions used by the Distutils modules.
+
+Distutils modules may raise these or standard exceptions,
+including :exc:`SystemExit`.
+"""
+
+
+class DistutilsError(Exception):
+    """The root of all Distutils evil."""
+
+    pass
+
+
+class DistutilsModuleError(DistutilsError):
+    """Unable to load an expected module, or to find an expected class
+    within some module (in particular, command modules and classes)."""
+
+    pass
+
+
+class DistutilsClassError(DistutilsError):
+    """Some command class (or possibly distribution class, if anyone
+    feels a need to subclass Distribution) is found not to be holding
+    up its end of the bargain, ie. implementing some part of the
+    "command "interface."""
+
+    pass
+
+
+class DistutilsGetoptError(DistutilsError):
+    """The option table provided to 'fancy_getopt()' is bogus."""
+
+    pass
+
+
+class DistutilsArgError(DistutilsError):
+    """Raised by fancy_getopt in response to getopt.error -- ie. an
+    error in the command line usage."""
+
+    pass
+
+
+class DistutilsFileError(DistutilsError):
+    """Any problems in the filesystem: expected file not found, etc.
+    Typically this is for problems that we detect before OSError
+    could be raised."""
+
+    pass
+
+
+class DistutilsOptionError(DistutilsError):
+    """Syntactic/semantic errors in command options, such as use of
+    mutually conflicting options, or inconsistent options,
+    badly-spelled values, etc.  No distinction is made between option
+    values originating in the setup script, the command line, config
+    files, or what-have-you -- but if we *know* something originated in
+    the setup script, we'll raise DistutilsSetupError instead."""
+
+    pass
+
+
+class DistutilsSetupError(DistutilsError):
+    """For errors that can be definitely blamed on the setup script,
+    such as invalid keyword arguments to 'setup()'."""
+
+    pass
+
+
+class DistutilsPlatformError(DistutilsError):
+    """We don't know how to do something on the current platform (but
+    we do know how to do it on some platform) -- eg. trying to compile
+    C files on a platform not supported by a CCompiler subclass."""
+
+    pass
+
+
+class DistutilsExecError(DistutilsError):
+    """Any problems executing an external program (such as the C
+    compiler, when compiling C files)."""
+
+    pass
+
+
+class DistutilsInternalError(DistutilsError):
+    """Internal inconsistencies or impossibilities (obviously, this
+    should never be seen if the code is working!)."""
+
+    pass
+
+
+class DistutilsTemplateError(DistutilsError):
+    """Syntax error in a file list template."""
+
+
+class DistutilsByteCompileError(DistutilsError):
+    """Byte compile error."""
+
+
+# Exception classes used by the CCompiler implementation classes
+class CCompilerError(Exception):
+    """Some compile/link operation failed."""
+
+
+class PreprocessError(CCompilerError):
+    """Failure to preprocess one or more C/C++ files."""
+
+
+class CompileError(CCompilerError):
+    """Failure to compile one or more C/C++ source files."""
+
+
+class LibError(CCompilerError):
+    """Failure to create a static library from one or more C/C++ object
+    files."""
+
+
+class LinkError(CCompilerError):
+    """Failure to link one or more C/C++ object files into an executable
+    or shared library file."""
+
+
+class UnknownFileError(CCompilerError):
+    """Attempt to process an unknown file type."""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/extension.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/extension.py
new file mode 100644
index 00000000..e0532734
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/extension.py
@@ -0,0 +1,255 @@
+"""distutils.extension
+
+Provides the Extension class, used to describe C/C++ extension
+modules in setup scripts."""
+
+import os
+import warnings
+
+# This class is really only used by the "build_ext" command, so it might
+# make sense to put it in distutils.command.build_ext.  However, that
+# module is already big enough, and I want to make this class a bit more
+# complex to simplify some common cases ("foo" module in "foo.c") and do
+# better error-checking ("foo.c" actually exists).
+#
+# Also, putting this in build_ext.py means every setup script would have to
+# import that large-ish module (indirectly, through distutils.core) in
+# order to do anything.
+
+
+class Extension:
+    """Just a collection of attributes that describes an extension
+    module and everything needed to build it (hopefully in a portable
+    way, but there are hooks that let you be as unportable as you need).
+
+    Instance attributes:
+      name : string
+        the full name of the extension, including any packages -- ie.
+        *not* a filename or pathname, but Python dotted name
+      sources : Iterable[string | os.PathLike]
+        iterable of source filenames (except strings, which could be misinterpreted
+        as a single filename), relative to the distribution root (where the setup
+        script lives), in Unix form (slash-separated) for portability. Can be any
+        non-string iterable (list, tuple, set, etc.) containing strings or
+        PathLike objects. Source files may be C, C++, SWIG (.i), platform-specific
+        resource files, or whatever else is recognized by the "build_ext" command
+        as source for a Python extension.
+      include_dirs : [string]
+        list of directories to search for C/C++ header files (in Unix
+        form for portability)
+      define_macros : [(name : string, value : string|None)]
+        list of macros to define; each macro is defined using a 2-tuple,
+        where 'value' is either the string to define it to or None to
+        define it without a particular value (equivalent of "#define
+        FOO" in source or -DFOO on Unix C compiler command line)
+      undef_macros : [string]
+        list of macros to undefine explicitly
+      library_dirs : [string]
+        list of directories to search for C/C++ libraries at link time
+      libraries : [string]
+        list of library names (not filenames or paths) to link against
+      runtime_library_dirs : [string]
+        list of directories to search for C/C++ libraries at run time
+        (for shared extensions, this is when the extension is loaded)
+      extra_objects : [string]
+        list of extra files to link with (eg. object files not implied
+        by 'sources', static library that must be explicitly specified,
+        binary resource files, etc.)
+      extra_compile_args : [string]
+        any extra platform- and compiler-specific information to use
+        when compiling the source files in 'sources'.  For platforms and
+        compilers where "command line" makes sense, this is typically a
+        list of command-line arguments, but for other platforms it could
+        be anything.
+      extra_link_args : [string]
+        any extra platform- and compiler-specific information to use
+        when linking object files together to create the extension (or
+        to create a new static Python interpreter).  Similar
+        interpretation as for 'extra_compile_args'.
+      export_symbols : [string]
+        list of symbols to be exported from a shared extension.  Not
+        used on all platforms, and not generally necessary for Python
+        extensions, which typically export exactly one symbol: "init" +
+        extension_name.
+      swig_opts : [string]
+        any extra options to pass to SWIG if a source file has the .i
+        extension.
+      depends : [string]
+        list of files that the extension depends on
+      language : string
+        extension language (i.e. "c", "c++", "objc"). Will be detected
+        from the source extensions if not provided.
+      optional : boolean
+        specifies that a build failure in the extension should not abort the
+        build process, but simply not install the failing extension.
+    """
+
+    # When adding arguments to this constructor, be sure to update
+    # setup_keywords in core.py.
+    def __init__(
+        self,
+        name,
+        sources,
+        include_dirs=None,
+        define_macros=None,
+        undef_macros=None,
+        library_dirs=None,
+        libraries=None,
+        runtime_library_dirs=None,
+        extra_objects=None,
+        extra_compile_args=None,
+        extra_link_args=None,
+        export_symbols=None,
+        swig_opts=None,
+        depends=None,
+        language=None,
+        optional=None,
+        **kw,  # To catch unknown keywords
+    ):
+        if not isinstance(name, str):
+            raise TypeError("'name' must be a string")
+
+        # handle the string case first; since strings are iterable, disallow them
+        if isinstance(sources, str):
+            raise TypeError(
+                "'sources' must be an iterable of strings or PathLike objects, not a string"
+            )
+
+        # now we check if it's iterable and contains valid types
+        try:
+            self.sources = list(map(os.fspath, sources))
+        except TypeError:
+            raise TypeError(
+                "'sources' must be an iterable of strings or PathLike objects"
+            )
+
+        self.name = name
+        self.include_dirs = include_dirs or []
+        self.define_macros = define_macros or []
+        self.undef_macros = undef_macros or []
+        self.library_dirs = library_dirs or []
+        self.libraries = libraries or []
+        self.runtime_library_dirs = runtime_library_dirs or []
+        self.extra_objects = extra_objects or []
+        self.extra_compile_args = extra_compile_args or []
+        self.extra_link_args = extra_link_args or []
+        self.export_symbols = export_symbols or []
+        self.swig_opts = swig_opts or []
+        self.depends = depends or []
+        self.language = language
+        self.optional = optional
+
+        # If there are unknown keyword options, warn about them
+        if len(kw) > 0:
+            options = [repr(option) for option in kw]
+            options = ', '.join(sorted(options))
+            msg = f"Unknown Extension options: {options}"
+            warnings.warn(msg)
+
+    def __repr__(self):
+        return f'<{self.__class__.__module__}.{self.__class__.__qualname__}({self.name!r}) at {id(self):#x}>'
+
+
+def read_setup_file(filename):  # noqa: C901
+    """Reads a Setup file and returns Extension instances."""
+    from distutils.sysconfig import _variable_rx, expand_makefile_vars, parse_makefile
+    from distutils.text_file import TextFile
+    from distutils.util import split_quoted
+
+    # First pass over the file to gather "VAR = VALUE" assignments.
+    vars = parse_makefile(filename)
+
+    # Second pass to gobble up the real content: lines of the form
+    #   <module> ... [<sourcefile> ...] [<cpparg> ...] [<library> ...]
+    file = TextFile(
+        filename,
+        strip_comments=True,
+        skip_blanks=True,
+        join_lines=True,
+        lstrip_ws=True,
+        rstrip_ws=True,
+    )
+    try:
+        extensions = []
+
+        while True:
+            line = file.readline()
+            if line is None:  # eof
+                break
+            if _variable_rx.match(line):  # VAR=VALUE, handled in first pass
+                continue
+
+            if line[0] == line[-1] == "*":
+                file.warn(f"'{line}' lines not handled yet")
+                continue
+
+            line = expand_makefile_vars(line, vars)
+            words = split_quoted(line)
+
+            # NB. this parses a slightly different syntax than the old
+            # makesetup script: here, there must be exactly one extension per
+            # line, and it must be the first word of the line.  I have no idea
+            # why the old syntax supported multiple extensions per line, as
+            # they all wind up being the same.
+
+            module = words[0]
+            ext = Extension(module, [])
+            append_next_word = None
+
+            for word in words[1:]:
+                if append_next_word is not None:
+                    append_next_word.append(word)
+                    append_next_word = None
+                    continue
+
+                suffix = os.path.splitext(word)[1]
+                switch = word[0:2]
+                value = word[2:]
+
+                if suffix in (".c", ".cc", ".cpp", ".cxx", ".c++", ".m", ".mm"):
+                    # hmm, should we do something about C vs. C++ sources?
+                    # or leave it up to the CCompiler implementation to
+                    # worry about?
+                    ext.sources.append(word)
+                elif switch == "-I":
+                    ext.include_dirs.append(value)
+                elif switch == "-D":
+                    equals = value.find("=")
+                    if equals == -1:  # bare "-DFOO" -- no value
+                        ext.define_macros.append((value, None))
+                    else:  # "-DFOO=blah"
+                        ext.define_macros.append((value[0:equals], value[equals + 2 :]))
+                elif switch == "-U":
+                    ext.undef_macros.append(value)
+                elif switch == "-C":  # only here 'cause makesetup has it!
+                    ext.extra_compile_args.append(word)
+                elif switch == "-l":
+                    ext.libraries.append(value)
+                elif switch == "-L":
+                    ext.library_dirs.append(value)
+                elif switch == "-R":
+                    ext.runtime_library_dirs.append(value)
+                elif word == "-rpath":
+                    append_next_word = ext.runtime_library_dirs
+                elif word == "-Xlinker":
+                    append_next_word = ext.extra_link_args
+                elif word == "-Xcompiler":
+                    append_next_word = ext.extra_compile_args
+                elif switch == "-u":
+                    ext.extra_link_args.append(word)
+                    if not value:
+                        append_next_word = ext.extra_link_args
+                elif suffix in (".a", ".so", ".sl", ".o", ".dylib"):
+                    # NB. a really faithful emulation of makesetup would
+                    # append a .o file to extra_objects only if it
+                    # had a slash in it; otherwise, it would s/.o/.c/
+                    # and append it to sources.  Hmmmm.
+                    ext.extra_objects.append(word)
+                else:
+                    file.warn(f"unrecognized argument '{word}'")
+
+            extensions.append(ext)
+    finally:
+        file.close()
+
+    return extensions
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/fancy_getopt.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/fancy_getopt.py
new file mode 100644
index 00000000..1a1d3a05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/fancy_getopt.py
@@ -0,0 +1,471 @@
+"""distutils.fancy_getopt
+
+Wrapper around the standard getopt module that provides the following
+additional features:
+  * short and long options are tied together
+  * options have help strings, so fancy_getopt could potentially
+    create a complete usage summary
+  * options set attributes of a passed-in object
+"""
+
+from __future__ import annotations
+
+import getopt
+import re
+import string
+import sys
+from collections.abc import Sequence
+from typing import Any
+
+from .errors import DistutilsArgError, DistutilsGetoptError
+
+# Much like command_re in distutils.core, this is close to but not quite
+# the same as a Python NAME -- except, in the spirit of most GNU
+# utilities, we use '-' in place of '_'.  (The spirit of LISP lives on!)
+# The similarities to NAME are again not a coincidence...
+longopt_pat = r'[a-zA-Z](?:[a-zA-Z0-9-]*)'
+longopt_re = re.compile(rf'^{longopt_pat}$')
+
+# For recognizing "negative alias" options, eg. "quiet=!verbose"
+neg_alias_re = re.compile(f"^({longopt_pat})=!({longopt_pat})$")
+
+# This is used to translate long options to legitimate Python identifiers
+# (for use as attributes of some object).
+longopt_xlate = str.maketrans('-', '_')
+
+
+class FancyGetopt:
+    """Wrapper around the standard 'getopt()' module that provides some
+    handy extra functionality:
+      * short and long options are tied together
+      * options have help strings, and help text can be assembled
+        from them
+      * options set attributes of a passed-in object
+      * boolean options can have "negative aliases" -- eg. if
+        --quiet is the "negative alias" of --verbose, then "--quiet"
+        on the command line sets 'verbose' to false
+    """
+
+    def __init__(self, option_table=None):
+        # The option table is (currently) a list of tuples.  The
+        # tuples may have 3 or four values:
+        #   (long_option, short_option, help_string [, repeatable])
+        # if an option takes an argument, its long_option should have '='
+        # appended; short_option should just be a single character, no ':'
+        # in any case.  If a long_option doesn't have a corresponding
+        # short_option, short_option should be None.  All option tuples
+        # must have long options.
+        self.option_table = option_table
+
+        # 'option_index' maps long option names to entries in the option
+        # table (ie. those 3-tuples).
+        self.option_index = {}
+        if self.option_table:
+            self._build_index()
+
+        # 'alias' records (duh) alias options; {'foo': 'bar'} means
+        # --foo is an alias for --bar
+        self.alias = {}
+
+        # 'negative_alias' keeps track of options that are the boolean
+        # opposite of some other option
+        self.negative_alias = {}
+
+        # These keep track of the information in the option table.  We
+        # don't actually populate these structures until we're ready to
+        # parse the command-line, since the 'option_table' passed in here
+        # isn't necessarily the final word.
+        self.short_opts = []
+        self.long_opts = []
+        self.short2long = {}
+        self.attr_name = {}
+        self.takes_arg = {}
+
+        # And 'option_order' is filled up in 'getopt()'; it records the
+        # original order of options (and their values) on the command-line,
+        # but expands short options, converts aliases, etc.
+        self.option_order = []
+
+    def _build_index(self):
+        self.option_index.clear()
+        for option in self.option_table:
+            self.option_index[option[0]] = option
+
+    def set_option_table(self, option_table):
+        self.option_table = option_table
+        self._build_index()
+
+    def add_option(self, long_option, short_option=None, help_string=None):
+        if long_option in self.option_index:
+            raise DistutilsGetoptError(
+                f"option conflict: already an option '{long_option}'"
+            )
+        else:
+            option = (long_option, short_option, help_string)
+            self.option_table.append(option)
+            self.option_index[long_option] = option
+
+    def has_option(self, long_option):
+        """Return true if the option table for this parser has an
+        option with long name 'long_option'."""
+        return long_option in self.option_index
+
+    def get_attr_name(self, long_option):
+        """Translate long option name 'long_option' to the form it
+        has as an attribute of some object: ie., translate hyphens
+        to underscores."""
+        return long_option.translate(longopt_xlate)
+
+    def _check_alias_dict(self, aliases, what):
+        assert isinstance(aliases, dict)
+        for alias, opt in aliases.items():
+            if alias not in self.option_index:
+                raise DistutilsGetoptError(
+                    f"invalid {what} '{alias}': option '{alias}' not defined"
+                )
+            if opt not in self.option_index:
+                raise DistutilsGetoptError(
+                    f"invalid {what} '{alias}': aliased option '{opt}' not defined"
+                )
+
+    def set_aliases(self, alias):
+        """Set the aliases for this option parser."""
+        self._check_alias_dict(alias, "alias")
+        self.alias = alias
+
+    def set_negative_aliases(self, negative_alias):
+        """Set the negative aliases for this option parser.
+        'negative_alias' should be a dictionary mapping option names to
+        option names, both the key and value must already be defined
+        in the option table."""
+        self._check_alias_dict(negative_alias, "negative alias")
+        self.negative_alias = negative_alias
+
+    def _grok_option_table(self):  # noqa: C901
+        """Populate the various data structures that keep tabs on the
+        option table.  Called by 'getopt()' before it can do anything
+        worthwhile.
+        """
+        self.long_opts = []
+        self.short_opts = []
+        self.short2long.clear()
+        self.repeat = {}
+
+        for option in self.option_table:
+            if len(option) == 3:
+                long, short, help = option
+                repeat = 0
+            elif len(option) == 4:
+                long, short, help, repeat = option
+            else:
+                # the option table is part of the code, so simply
+                # assert that it is correct
+                raise ValueError(f"invalid option tuple: {option!r}")
+
+            # Type- and value-check the option names
+            if not isinstance(long, str) or len(long) < 2:
+                raise DistutilsGetoptError(
+                    f"invalid long option '{long}': must be a string of length >= 2"
+                )
+
+            if not ((short is None) or (isinstance(short, str) and len(short) == 1)):
+                raise DistutilsGetoptError(
+                    f"invalid short option '{short}': must a single character or None"
+                )
+
+            self.repeat[long] = repeat
+            self.long_opts.append(long)
+
+            if long[-1] == '=':  # option takes an argument?
+                if short:
+                    short = short + ':'
+                long = long[0:-1]
+                self.takes_arg[long] = True
+            else:
+                # Is option is a "negative alias" for some other option (eg.
+                # "quiet" == "!verbose")?
+                alias_to = self.negative_alias.get(long)
+                if alias_to is not None:
+                    if self.takes_arg[alias_to]:
+                        raise DistutilsGetoptError(
+                            f"invalid negative alias '{long}': "
+                            f"aliased option '{alias_to}' takes a value"
+                        )
+
+                    self.long_opts[-1] = long  # XXX redundant?!
+                self.takes_arg[long] = False
+
+            # If this is an alias option, make sure its "takes arg" flag is
+            # the same as the option it's aliased to.
+            alias_to = self.alias.get(long)
+            if alias_to is not None:
+                if self.takes_arg[long] != self.takes_arg[alias_to]:
+                    raise DistutilsGetoptError(
+                        f"invalid alias '{long}': inconsistent with "
+                        f"aliased option '{alias_to}' (one of them takes a value, "
+                        "the other doesn't"
+                    )
+
+            # Now enforce some bondage on the long option name, so we can
+            # later translate it to an attribute name on some object.  Have
+            # to do this a bit late to make sure we've removed any trailing
+            # '='.
+            if not longopt_re.match(long):
+                raise DistutilsGetoptError(
+                    f"invalid long option name '{long}' "
+                    "(must be letters, numbers, hyphens only"
+                )
+
+            self.attr_name[long] = self.get_attr_name(long)
+            if short:
+                self.short_opts.append(short)
+                self.short2long[short[0]] = long
+
+    def getopt(self, args: Sequence[str] | None = None, object=None):  # noqa: C901
+        """Parse command-line options in args. Store as attributes on object.
+
+        If 'args' is None or not supplied, uses 'sys.argv[1:]'.  If
+        'object' is None or not supplied, creates a new OptionDummy
+        object, stores option values there, and returns a tuple (args,
+        object).  If 'object' is supplied, it is modified in place and
+        'getopt()' just returns 'args'; in both cases, the returned
+        'args' is a modified copy of the passed-in 'args' list, which
+        is left untouched.
+        """
+        if args is None:
+            args = sys.argv[1:]
+        if object is None:
+            object = OptionDummy()
+            created_object = True
+        else:
+            created_object = False
+
+        self._grok_option_table()
+
+        short_opts = ' '.join(self.short_opts)
+        try:
+            opts, args = getopt.getopt(args, short_opts, self.long_opts)
+        except getopt.error as msg:
+            raise DistutilsArgError(msg)
+
+        for opt, val in opts:
+            if len(opt) == 2 and opt[0] == '-':  # it's a short option
+                opt = self.short2long[opt[1]]
+            else:
+                assert len(opt) > 2 and opt[:2] == '--'
+                opt = opt[2:]
+
+            alias = self.alias.get(opt)
+            if alias:
+                opt = alias
+
+            if not self.takes_arg[opt]:  # boolean option?
+                assert val == '', "boolean option can't have value"
+                alias = self.negative_alias.get(opt)
+                if alias:
+                    opt = alias
+                    val = 0
+                else:
+                    val = 1
+
+            attr = self.attr_name[opt]
+            # The only repeating option at the moment is 'verbose'.
+            # It has a negative option -q quiet, which should set verbose = False.
+            if val and self.repeat.get(attr) is not None:
+                val = getattr(object, attr, 0) + 1
+            setattr(object, attr, val)
+            self.option_order.append((opt, val))
+
+        # for opts
+        if created_object:
+            return args, object
+        else:
+            return args
+
+    def get_option_order(self):
+        """Returns the list of (option, value) tuples processed by the
+        previous run of 'getopt()'.  Raises RuntimeError if
+        'getopt()' hasn't been called yet.
+        """
+        if self.option_order is None:
+            raise RuntimeError("'getopt()' hasn't been called yet")
+        else:
+            return self.option_order
+
+    def generate_help(self, header=None):  # noqa: C901
+        """Generate help text (a list of strings, one per suggested line of
+        output) from the option table for this FancyGetopt object.
+        """
+        # Blithely assume the option table is good: probably wouldn't call
+        # 'generate_help()' unless you've already called 'getopt()'.
+
+        # First pass: determine maximum length of long option names
+        max_opt = 0
+        for option in self.option_table:
+            long = option[0]
+            short = option[1]
+            ell = len(long)
+            if long[-1] == '=':
+                ell = ell - 1
+            if short is not None:
+                ell = ell + 5  # " (-x)" where short == 'x'
+            if ell > max_opt:
+                max_opt = ell
+
+        opt_width = max_opt + 2 + 2 + 2  # room for indent + dashes + gutter
+
+        # Typical help block looks like this:
+        #   --foo       controls foonabulation
+        # Help block for longest option looks like this:
+        #   --flimflam  set the flim-flam level
+        # and with wrapped text:
+        #   --flimflam  set the flim-flam level (must be between
+        #               0 and 100, except on Tuesdays)
+        # Options with short names will have the short name shown (but
+        # it doesn't contribute to max_opt):
+        #   --foo (-f)  controls foonabulation
+        # If adding the short option would make the left column too wide,
+        # we push the explanation off to the next line
+        #   --flimflam (-l)
+        #               set the flim-flam level
+        # Important parameters:
+        #   - 2 spaces before option block start lines
+        #   - 2 dashes for each long option name
+        #   - min. 2 spaces between option and explanation (gutter)
+        #   - 5 characters (incl. space) for short option name
+
+        # Now generate lines of help text.  (If 80 columns were good enough
+        # for Jesus, then 78 columns are good enough for me!)
+        line_width = 78
+        text_width = line_width - opt_width
+        big_indent = ' ' * opt_width
+        if header:
+            lines = [header]
+        else:
+            lines = ['Option summary:']
+
+        for option in self.option_table:
+            long, short, help = option[:3]
+            text = wrap_text(help, text_width)
+            if long[-1] == '=':
+                long = long[0:-1]
+
+            # Case 1: no short option at all (makes life easy)
+            if short is None:
+                if text:
+                    lines.append(f"  --{long:<{max_opt}}  {text[0]}")
+                else:
+                    lines.append(f"  --{long:<{max_opt}}")
+
+            # Case 2: we have a short option, so we have to include it
+            # just after the long option
+            else:
+                opt_names = f"{long} (-{short})"
+                if text:
+                    lines.append(f"  --{opt_names:<{max_opt}}  {text[0]}")
+                else:
+                    lines.append(f"  --{opt_names:<{max_opt}}")
+
+            for ell in text[1:]:
+                lines.append(big_indent + ell)
+        return lines
+
+    def print_help(self, header=None, file=None):
+        if file is None:
+            file = sys.stdout
+        for line in self.generate_help(header):
+            file.write(line + "\n")
+
+
+def fancy_getopt(options, negative_opt, object, args: Sequence[str] | None):
+    parser = FancyGetopt(options)
+    parser.set_negative_aliases(negative_opt)
+    return parser.getopt(args, object)
+
+
+WS_TRANS = {ord(_wschar): ' ' for _wschar in string.whitespace}
+
+
+def wrap_text(text, width):
+    """wrap_text(text : string, width : int) -> [string]
+
+    Split 'text' into multiple lines of no more than 'width' characters
+    each, and return the list of strings that results.
+    """
+    if text is None:
+        return []
+    if len(text) <= width:
+        return [text]
+
+    text = text.expandtabs()
+    text = text.translate(WS_TRANS)
+    chunks = re.split(r'( +|-+)', text)
+    chunks = [ch for ch in chunks if ch]  # ' - ' results in empty strings
+    lines = []
+
+    while chunks:
+        cur_line = []  # list of chunks (to-be-joined)
+        cur_len = 0  # length of current line
+
+        while chunks:
+            ell = len(chunks[0])
+            if cur_len + ell <= width:  # can squeeze (at least) this chunk in
+                cur_line.append(chunks[0])
+                del chunks[0]
+                cur_len = cur_len + ell
+            else:  # this line is full
+                # drop last chunk if all space
+                if cur_line and cur_line[-1][0] == ' ':
+                    del cur_line[-1]
+                break
+
+        if chunks:  # any chunks left to process?
+            # if the current line is still empty, then we had a single
+            # chunk that's too big too fit on a line -- so we break
+            # down and break it up at the line width
+            if cur_len == 0:
+                cur_line.append(chunks[0][0:width])
+                chunks[0] = chunks[0][width:]
+
+            # all-whitespace chunks at the end of a line can be discarded
+            # (and we know from the re.split above that if a chunk has
+            # *any* whitespace, it is *all* whitespace)
+            if chunks[0][0] == ' ':
+                del chunks[0]
+
+        # and store this line in the list-of-all-lines -- as a single
+        # string, of course!
+        lines.append(''.join(cur_line))
+
+    return lines
+
+
+def translate_longopt(opt):
+    """Convert a long option name to a valid Python identifier by
+    changing "-" to "_".
+    """
+    return opt.translate(longopt_xlate)
+
+
+class OptionDummy:
+    """Dummy class just used as a place to hold command-line option
+    values as instance attributes."""
+
+    def __init__(self, options: Sequence[Any] = []):
+        """Create a new OptionDummy instance.  The attributes listed in
+        'options' will be initialized to None."""
+        for opt in options:
+            setattr(self, opt, None)
+
+
+if __name__ == "__main__":
+    text = """\
+Tra-la-la, supercalifragilisticexpialidocious.
+How *do* you spell that odd word, anyways?
+(Someone ask Mary -- she'll know [or she'll
+say, "How should I know?"].)"""
+
+    for w in (10, 20, 30, 40):
+        print(f"width: {w}")
+        print("\n".join(wrap_text(text, w)))
+        print()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/file_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/file_util.py
new file mode 100644
index 00000000..0acc8cb8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/file_util.py
@@ -0,0 +1,236 @@
+"""distutils.file_util
+
+Utility functions for operating on single files.
+"""
+
+import os
+
+from ._log import log
+from .errors import DistutilsFileError
+
+# for generating verbose output in 'copy_file()'
+_copy_action = {None: 'copying', 'hard': 'hard linking', 'sym': 'symbolically linking'}
+
+
+def _copy_file_contents(src, dst, buffer_size=16 * 1024):  # noqa: C901
+    """Copy the file 'src' to 'dst'; both must be filenames.  Any error
+    opening either file, reading from 'src', or writing to 'dst', raises
+    DistutilsFileError.  Data is read/written in chunks of 'buffer_size'
+    bytes (default 16k).  No attempt is made to handle anything apart from
+    regular files.
+    """
+    # Stolen from shutil module in the standard library, but with
+    # custom error-handling added.
+    fsrc = None
+    fdst = None
+    try:
+        try:
+            fsrc = open(src, 'rb')
+        except OSError as e:
+            raise DistutilsFileError(f"could not open '{src}': {e.strerror}")
+
+        if os.path.exists(dst):
+            try:
+                os.unlink(dst)
+            except OSError as e:
+                raise DistutilsFileError(f"could not delete '{dst}': {e.strerror}")
+
+        try:
+            fdst = open(dst, 'wb')
+        except OSError as e:
+            raise DistutilsFileError(f"could not create '{dst}': {e.strerror}")
+
+        while True:
+            try:
+                buf = fsrc.read(buffer_size)
+            except OSError as e:
+                raise DistutilsFileError(f"could not read from '{src}': {e.strerror}")
+
+            if not buf:
+                break
+
+            try:
+                fdst.write(buf)
+            except OSError as e:
+                raise DistutilsFileError(f"could not write to '{dst}': {e.strerror}")
+    finally:
+        if fdst:
+            fdst.close()
+        if fsrc:
+            fsrc.close()
+
+
+def copy_file(  # noqa: C901
+    src,
+    dst,
+    preserve_mode=True,
+    preserve_times=True,
+    update=False,
+    link=None,
+    verbose=True,
+    dry_run=False,
+):
+    """Copy a file 'src' to 'dst'.  If 'dst' is a directory, then 'src' is
+    copied there with the same name; otherwise, it must be a filename.  (If
+    the file exists, it will be ruthlessly clobbered.)  If 'preserve_mode'
+    is true (the default), the file's mode (type and permission bits, or
+    whatever is analogous on the current platform) is copied.  If
+    'preserve_times' is true (the default), the last-modified and
+    last-access times are copied as well.  If 'update' is true, 'src' will
+    only be copied if 'dst' does not exist, or if 'dst' does exist but is
+    older than 'src'.
+
+    'link' allows you to make hard links (os.link) or symbolic links
+    (os.symlink) instead of copying: set it to "hard" or "sym"; if it is
+    None (the default), files are copied.  Don't set 'link' on systems that
+    don't support it: 'copy_file()' doesn't check if hard or symbolic
+    linking is available. If hardlink fails, falls back to
+    _copy_file_contents().
+
+    Under Mac OS, uses the native file copy function in macostools; on
+    other systems, uses '_copy_file_contents()' to copy file contents.
+
+    Return a tuple (dest_name, copied): 'dest_name' is the actual name of
+    the output file, and 'copied' is true if the file was copied (or would
+    have been copied, if 'dry_run' true).
+    """
+    # XXX if the destination file already exists, we clobber it if
+    # copying, but blow up if linking.  Hmmm.  And I don't know what
+    # macostools.copyfile() does.  Should definitely be consistent, and
+    # should probably blow up if destination exists and we would be
+    # changing it (ie. it's not already a hard/soft link to src OR
+    # (not update) and (src newer than dst).
+
+    from distutils._modified import newer
+    from stat import S_IMODE, ST_ATIME, ST_MODE, ST_MTIME
+
+    if not os.path.isfile(src):
+        raise DistutilsFileError(
+            f"can't copy '{src}': doesn't exist or not a regular file"
+        )
+
+    if os.path.isdir(dst):
+        dir = dst
+        dst = os.path.join(dst, os.path.basename(src))
+    else:
+        dir = os.path.dirname(dst)
+
+    if update and not newer(src, dst):
+        if verbose >= 1:
+            log.debug("not copying %s (output up-to-date)", src)
+        return (dst, False)
+
+    try:
+        action = _copy_action[link]
+    except KeyError:
+        raise ValueError(f"invalid value '{link}' for 'link' argument")
+
+    if verbose >= 1:
+        if os.path.basename(dst) == os.path.basename(src):
+            log.info("%s %s -> %s", action, src, dir)
+        else:
+            log.info("%s %s -> %s", action, src, dst)
+
+    if dry_run:
+        return (dst, True)
+
+    # If linking (hard or symbolic), use the appropriate system call
+    # (Unix only, of course, but that's the caller's responsibility)
+    elif link == 'hard':
+        if not (os.path.exists(dst) and os.path.samefile(src, dst)):
+            try:
+                os.link(src, dst)
+            except OSError:
+                # If hard linking fails, fall back on copying file
+                # (some special filesystems don't support hard linking
+                #  even under Unix, see issue #8876).
+                pass
+            else:
+                return (dst, True)
+    elif link == 'sym':
+        if not (os.path.exists(dst) and os.path.samefile(src, dst)):
+            os.symlink(src, dst)
+            return (dst, True)
+
+    # Otherwise (non-Mac, not linking), copy the file contents and
+    # (optionally) copy the times and mode.
+    _copy_file_contents(src, dst)
+    if preserve_mode or preserve_times:
+        st = os.stat(src)
+
+        # According to David Ascher <da@ski.org>, utime() should be done
+        # before chmod() (at least under NT).
+        if preserve_times:
+            os.utime(dst, (st[ST_ATIME], st[ST_MTIME]))
+        if preserve_mode:
+            os.chmod(dst, S_IMODE(st[ST_MODE]))
+
+    return (dst, True)
+
+
+# XXX I suspect this is Unix-specific -- need porting help!
+def move_file(src, dst, verbose=True, dry_run=False):  # noqa: C901
+    """Move a file 'src' to 'dst'.  If 'dst' is a directory, the file will
+    be moved into it with the same name; otherwise, 'src' is just renamed
+    to 'dst'.  Return the new full name of the file.
+
+    Handles cross-device moves on Unix using 'copy_file()'.  What about
+    other systems???
+    """
+    import errno
+    from os.path import basename, dirname, exists, isdir, isfile
+
+    if verbose >= 1:
+        log.info("moving %s -> %s", src, dst)
+
+    if dry_run:
+        return dst
+
+    if not isfile(src):
+        raise DistutilsFileError(f"can't move '{src}': not a regular file")
+
+    if isdir(dst):
+        dst = os.path.join(dst, basename(src))
+    elif exists(dst):
+        raise DistutilsFileError(
+            f"can't move '{src}': destination '{dst}' already exists"
+        )
+
+    if not isdir(dirname(dst)):
+        raise DistutilsFileError(
+            f"can't move '{src}': destination '{dst}' not a valid path"
+        )
+
+    copy_it = False
+    try:
+        os.rename(src, dst)
+    except OSError as e:
+        (num, msg) = e.args
+        if num == errno.EXDEV:
+            copy_it = True
+        else:
+            raise DistutilsFileError(f"couldn't move '{src}' to '{dst}': {msg}")
+
+    if copy_it:
+        copy_file(src, dst, verbose=verbose)
+        try:
+            os.unlink(src)
+        except OSError as e:
+            (num, msg) = e.args
+            try:
+                os.unlink(dst)
+            except OSError:
+                pass
+            raise DistutilsFileError(
+                f"couldn't move '{src}' to '{dst}' by copy/delete: "
+                f"delete '{src}' failed: {msg}"
+            )
+    return dst
+
+
+def write_file(filename, contents):
+    """Create a file with the specified name and write 'contents' (a
+    sequence of strings without line terminators) to it.
+    """
+    with open(filename, 'w', encoding='utf-8') as f:
+        f.writelines(line + '\n' for line in contents)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/filelist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/filelist.py
new file mode 100644
index 00000000..9857b195
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/filelist.py
@@ -0,0 +1,366 @@
+"""distutils.filelist
+
+Provides the FileList class, used for poking about the filesystem
+and building lists of files.
+"""
+
+import fnmatch
+import functools
+import os
+import re
+
+from ._log import log
+from .errors import DistutilsInternalError, DistutilsTemplateError
+from .util import convert_path
+
+
+class FileList:
+    """A list of files built by on exploring the filesystem and filtered by
+    applying various patterns to what we find there.
+
+    Instance attributes:
+      dir
+        directory from which files will be taken -- only used if
+        'allfiles' not supplied to constructor
+      files
+        list of filenames currently being built/filtered/manipulated
+      allfiles
+        complete list of files under consideration (ie. without any
+        filtering applied)
+    """
+
+    def __init__(self, warn=None, debug_print=None):
+        # ignore argument to FileList, but keep them for backwards
+        # compatibility
+        self.allfiles = None
+        self.files = []
+
+    def set_allfiles(self, allfiles):
+        self.allfiles = allfiles
+
+    def findall(self, dir=os.curdir):
+        self.allfiles = findall(dir)
+
+    def debug_print(self, msg):
+        """Print 'msg' to stdout if the global DEBUG (taken from the
+        DISTUTILS_DEBUG environment variable) flag is true.
+        """
+        from distutils.debug import DEBUG
+
+        if DEBUG:
+            print(msg)
+
+    # Collection methods
+
+    def append(self, item):
+        self.files.append(item)
+
+    def extend(self, items):
+        self.files.extend(items)
+
+    def sort(self):
+        # Not a strict lexical sort!
+        sortable_files = sorted(map(os.path.split, self.files))
+        self.files = []
+        for sort_tuple in sortable_files:
+            self.files.append(os.path.join(*sort_tuple))
+
+    # Other miscellaneous utility methods
+
+    def remove_duplicates(self):
+        # Assumes list has been sorted!
+        for i in range(len(self.files) - 1, 0, -1):
+            if self.files[i] == self.files[i - 1]:
+                del self.files[i]
+
+    # "File template" methods
+
+    def _parse_template_line(self, line):
+        words = line.split()
+        action = words[0]
+
+        patterns = dir = dir_pattern = None
+
+        if action in ('include', 'exclude', 'global-include', 'global-exclude'):
+            if len(words) < 2:
+                raise DistutilsTemplateError(
+                    f"'{action}' expects <pattern1> <pattern2> ..."
+                )
+            patterns = [convert_path(w) for w in words[1:]]
+        elif action in ('recursive-include', 'recursive-exclude'):
+            if len(words) < 3:
+                raise DistutilsTemplateError(
+                    f"'{action}' expects <dir> <pattern1> <pattern2> ..."
+                )
+            dir = convert_path(words[1])
+            patterns = [convert_path(w) for w in words[2:]]
+        elif action in ('graft', 'prune'):
+            if len(words) != 2:
+                raise DistutilsTemplateError(
+                    f"'{action}' expects a single <dir_pattern>"
+                )
+            dir_pattern = convert_path(words[1])
+        else:
+            raise DistutilsTemplateError(f"unknown action '{action}'")
+
+        return (action, patterns, dir, dir_pattern)
+
+    def process_template_line(self, line):  # noqa: C901
+        # Parse the line: split it up, make sure the right number of words
+        # is there, and return the relevant words.  'action' is always
+        # defined: it's the first word of the line.  Which of the other
+        # three are defined depends on the action; it'll be either
+        # patterns, (dir and patterns), or (dir_pattern).
+        (action, patterns, dir, dir_pattern) = self._parse_template_line(line)
+
+        # OK, now we know that the action is valid and we have the
+        # right number of words on the line for that action -- so we
+        # can proceed with minimal error-checking.
+        if action == 'include':
+            self.debug_print("include " + ' '.join(patterns))
+            for pattern in patterns:
+                if not self.include_pattern(pattern, anchor=True):
+                    log.warning("warning: no files found matching '%s'", pattern)
+
+        elif action == 'exclude':
+            self.debug_print("exclude " + ' '.join(patterns))
+            for pattern in patterns:
+                if not self.exclude_pattern(pattern, anchor=True):
+                    log.warning(
+                        "warning: no previously-included files found matching '%s'",
+                        pattern,
+                    )
+
+        elif action == 'global-include':
+            self.debug_print("global-include " + ' '.join(patterns))
+            for pattern in patterns:
+                if not self.include_pattern(pattern, anchor=False):
+                    log.warning(
+                        (
+                            "warning: no files found matching '%s' "
+                            "anywhere in distribution"
+                        ),
+                        pattern,
+                    )
+
+        elif action == 'global-exclude':
+            self.debug_print("global-exclude " + ' '.join(patterns))
+            for pattern in patterns:
+                if not self.exclude_pattern(pattern, anchor=False):
+                    log.warning(
+                        (
+                            "warning: no previously-included files matching "
+                            "'%s' found anywhere in distribution"
+                        ),
+                        pattern,
+                    )
+
+        elif action == 'recursive-include':
+            self.debug_print("recursive-include {} {}".format(dir, ' '.join(patterns)))
+            for pattern in patterns:
+                if not self.include_pattern(pattern, prefix=dir):
+                    msg = "warning: no files found matching '%s' under directory '%s'"
+                    log.warning(msg, pattern, dir)
+
+        elif action == 'recursive-exclude':
+            self.debug_print("recursive-exclude {} {}".format(dir, ' '.join(patterns)))
+            for pattern in patterns:
+                if not self.exclude_pattern(pattern, prefix=dir):
+                    log.warning(
+                        (
+                            "warning: no previously-included files matching "
+                            "'%s' found under directory '%s'"
+                        ),
+                        pattern,
+                        dir,
+                    )
+
+        elif action == 'graft':
+            self.debug_print("graft " + dir_pattern)
+            if not self.include_pattern(None, prefix=dir_pattern):
+                log.warning("warning: no directories found matching '%s'", dir_pattern)
+
+        elif action == 'prune':
+            self.debug_print("prune " + dir_pattern)
+            if not self.exclude_pattern(None, prefix=dir_pattern):
+                log.warning(
+                    ("no previously-included directories found matching '%s'"),
+                    dir_pattern,
+                )
+        else:
+            raise DistutilsInternalError(
+                f"this cannot happen: invalid action '{action}'"
+            )
+
+    # Filtering/selection methods
+
+    def include_pattern(self, pattern, anchor=True, prefix=None, is_regex=False):
+        """Select strings (presumably filenames) from 'self.files' that
+        match 'pattern', a Unix-style wildcard (glob) pattern.  Patterns
+        are not quite the same as implemented by the 'fnmatch' module: '*'
+        and '?'  match non-special characters, where "special" is platform-
+        dependent: slash on Unix; colon, slash, and backslash on
+        DOS/Windows; and colon on Mac OS.
+
+        If 'anchor' is true (the default), then the pattern match is more
+        stringent: "*.py" will match "foo.py" but not "foo/bar.py".  If
+        'anchor' is false, both of these will match.
+
+        If 'prefix' is supplied, then only filenames starting with 'prefix'
+        (itself a pattern) and ending with 'pattern', with anything in between
+        them, will match.  'anchor' is ignored in this case.
+
+        If 'is_regex' is true, 'anchor' and 'prefix' are ignored, and
+        'pattern' is assumed to be either a string containing a regex or a
+        regex object -- no translation is done, the regex is just compiled
+        and used as-is.
+
+        Selected strings will be added to self.files.
+
+        Return True if files are found, False otherwise.
+        """
+        # XXX docstring lying about what the special chars are?
+        files_found = False
+        pattern_re = translate_pattern(pattern, anchor, prefix, is_regex)
+        self.debug_print(f"include_pattern: applying regex r'{pattern_re.pattern}'")
+
+        # delayed loading of allfiles list
+        if self.allfiles is None:
+            self.findall()
+
+        for name in self.allfiles:
+            if pattern_re.search(name):
+                self.debug_print(" adding " + name)
+                self.files.append(name)
+                files_found = True
+        return files_found
+
+    def exclude_pattern(self, pattern, anchor=True, prefix=None, is_regex=False):
+        """Remove strings (presumably filenames) from 'files' that match
+        'pattern'.  Other parameters are the same as for
+        'include_pattern()', above.
+        The list 'self.files' is modified in place.
+        Return True if files are found, False otherwise.
+        """
+        files_found = False
+        pattern_re = translate_pattern(pattern, anchor, prefix, is_regex)
+        self.debug_print(f"exclude_pattern: applying regex r'{pattern_re.pattern}'")
+        for i in range(len(self.files) - 1, -1, -1):
+            if pattern_re.search(self.files[i]):
+                self.debug_print(" removing " + self.files[i])
+                del self.files[i]
+                files_found = True
+        return files_found
+
+
+# Utility functions
+
+
+def _find_all_simple(path):
+    """
+    Find all files under 'path'
+    """
+    all_unique = _UniqueDirs.filter(os.walk(path, followlinks=True))
+    results = (
+        os.path.join(base, file) for base, dirs, files in all_unique for file in files
+    )
+    return filter(os.path.isfile, results)
+
+
+class _UniqueDirs(set):
+    """
+    Exclude previously-seen dirs from walk results,
+    avoiding infinite recursion.
+    Ref https://bugs.python.org/issue44497.
+    """
+
+    def __call__(self, walk_item):
+        """
+        Given an item from an os.walk result, determine
+        if the item represents a unique dir for this instance
+        and if not, prevent further traversal.
+        """
+        base, dirs, files = walk_item
+        stat = os.stat(base)
+        candidate = stat.st_dev, stat.st_ino
+        found = candidate in self
+        if found:
+            del dirs[:]
+        self.add(candidate)
+        return not found
+
+    @classmethod
+    def filter(cls, items):
+        return filter(cls(), items)
+
+
+def findall(dir=os.curdir):
+    """
+    Find all files under 'dir' and return the list of full filenames.
+    Unless dir is '.', return full filenames with dir prepended.
+    """
+    files = _find_all_simple(dir)
+    if dir == os.curdir:
+        make_rel = functools.partial(os.path.relpath, start=dir)
+        files = map(make_rel, files)
+    return list(files)
+
+
+def glob_to_re(pattern):
+    """Translate a shell-like glob pattern to a regular expression; return
+    a string containing the regex.  Differs from 'fnmatch.translate()' in
+    that '*' does not match "special characters" (which are
+    platform-specific).
+    """
+    pattern_re = fnmatch.translate(pattern)
+
+    # '?' and '*' in the glob pattern become '.' and '.*' in the RE, which
+    # IMHO is wrong -- '?' and '*' aren't supposed to match slash in Unix,
+    # and by extension they shouldn't match such "special characters" under
+    # any OS.  So change all non-escaped dots in the RE to match any
+    # character except the special characters (currently: just os.sep).
+    sep = os.sep
+    if os.sep == '\\':
+        # we're using a regex to manipulate a regex, so we need
+        # to escape the backslash twice
+        sep = r'\\\\'
+    escaped = rf'\1[^{sep}]'
+    pattern_re = re.sub(r'((?<!\\)(\\\\)*)\.', escaped, pattern_re)
+    return pattern_re
+
+
+def translate_pattern(pattern, anchor=True, prefix=None, is_regex=False):
+    """Translate a shell-like wildcard pattern to a compiled regular
+    expression.  Return the compiled regex.  If 'is_regex' true,
+    then 'pattern' is directly compiled to a regex (if it's a string)
+    or just returned as-is (assumes it's a regex object).
+    """
+    if is_regex:
+        if isinstance(pattern, str):
+            return re.compile(pattern)
+        else:
+            return pattern
+
+    # ditch start and end characters
+    start, _, end = glob_to_re('_').partition('_')
+
+    if pattern:
+        pattern_re = glob_to_re(pattern)
+        assert pattern_re.startswith(start) and pattern_re.endswith(end)
+    else:
+        pattern_re = ''
+
+    if prefix is not None:
+        prefix_re = glob_to_re(prefix)
+        assert prefix_re.startswith(start) and prefix_re.endswith(end)
+        prefix_re = prefix_re[len(start) : len(prefix_re) - len(end)]
+        sep = os.sep
+        if os.sep == '\\':
+            sep = r'\\'
+        pattern_re = pattern_re[len(start) : len(pattern_re) - len(end)]
+        pattern_re = rf'{start}\A{prefix_re}{sep}.*{pattern_re}{end}'
+    else:  # no prefix -- respect anchor flag
+        if anchor:
+            pattern_re = rf'{start}\A{pattern_re[len(start) :]}'
+
+    return re.compile(pattern_re)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/log.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/log.py
new file mode 100644
index 00000000..8abb09cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/log.py
@@ -0,0 +1,56 @@
+"""
+A simple log mechanism styled after PEP 282.
+
+Retained for compatibility and should not be used.
+"""
+
+import logging
+import warnings
+
+from ._log import log as _global_log
+
+DEBUG = logging.DEBUG
+INFO = logging.INFO
+WARN = logging.WARN
+ERROR = logging.ERROR
+FATAL = logging.FATAL
+
+log = _global_log.log
+debug = _global_log.debug
+info = _global_log.info
+warn = _global_log.warning
+error = _global_log.error
+fatal = _global_log.fatal
+
+
+def set_threshold(level):
+    orig = _global_log.level
+    _global_log.setLevel(level)
+    return orig
+
+
+def set_verbosity(v):
+    if v <= 0:
+        set_threshold(logging.WARN)
+    elif v == 1:
+        set_threshold(logging.INFO)
+    elif v >= 2:
+        set_threshold(logging.DEBUG)
+
+
+class Log(logging.Logger):
+    """distutils.log.Log is deprecated, please use an alternative from `logging`."""
+
+    def __init__(self, threshold=WARN):
+        warnings.warn(Log.__doc__)  # avoid DeprecationWarning to ensure warn is shown
+        super().__init__(__name__, level=threshold)
+
+    @property
+    def threshold(self):
+        return self.level
+
+    @threshold.setter
+    def threshold(self, level):
+        self.setLevel(level)
+
+    warn = logging.Logger.warning
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/spawn.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/spawn.py
new file mode 100644
index 00000000..ba280334
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/spawn.py
@@ -0,0 +1,117 @@
+"""distutils.spawn
+
+Provides the 'spawn()' function, a front-end to various platform-
+specific functions for launching another program in a sub-process.
+"""
+
+from __future__ import annotations
+
+import os
+import platform
+import shutil
+import subprocess
+import sys
+import warnings
+from collections.abc import Mapping
+
+from ._log import log
+from .debug import DEBUG
+from .errors import DistutilsExecError
+
+
+def _debug(cmd):
+    """
+    Render a subprocess command differently depending on DEBUG.
+    """
+    return cmd if DEBUG else cmd[0]
+
+
+def _inject_macos_ver(env: Mapping[str:str] | None) -> Mapping[str:str] | None:
+    if platform.system() != 'Darwin':
+        return env
+
+    from .util import MACOSX_VERSION_VAR, get_macosx_target_ver
+
+    target_ver = get_macosx_target_ver()
+    update = {MACOSX_VERSION_VAR: target_ver} if target_ver else {}
+    return {**_resolve(env), **update}
+
+
+def _resolve(env: Mapping[str:str] | None) -> Mapping[str:str]:
+    return os.environ if env is None else env
+
+
+def spawn(cmd, search_path=True, verbose=False, dry_run=False, env=None):
+    """Run another program, specified as a command list 'cmd', in a new process.
+
+    'cmd' is just the argument list for the new process, ie.
+    cmd[0] is the program to run and cmd[1:] are the rest of its arguments.
+    There is no way to run a program with a name different from that of its
+    executable.
+
+    If 'search_path' is true (the default), the system's executable
+    search path will be used to find the program; otherwise, cmd[0]
+    must be the exact path to the executable.  If 'dry_run' is true,
+    the command will not actually be run.
+
+    Raise DistutilsExecError if running the program fails in any way; just
+    return on success.
+    """
+    log.info(subprocess.list2cmdline(cmd))
+    if dry_run:
+        return
+
+    if search_path:
+        executable = shutil.which(cmd[0])
+        if executable is not None:
+            cmd[0] = executable
+
+    try:
+        subprocess.check_call(cmd, env=_inject_macos_ver(env))
+    except OSError as exc:
+        raise DistutilsExecError(
+            f"command {_debug(cmd)!r} failed: {exc.args[-1]}"
+        ) from exc
+    except subprocess.CalledProcessError as err:
+        raise DistutilsExecError(
+            f"command {_debug(cmd)!r} failed with exit code {err.returncode}"
+        ) from err
+
+
+def find_executable(executable, path=None):
+    """Tries to find 'executable' in the directories listed in 'path'.
+
+    A string listing directories separated by 'os.pathsep'; defaults to
+    os.environ['PATH'].  Returns the complete filename or None if not found.
+    """
+    warnings.warn(
+        'Use shutil.which instead of find_executable', DeprecationWarning, stacklevel=2
+    )
+    _, ext = os.path.splitext(executable)
+    if (sys.platform == 'win32') and (ext != '.exe'):
+        executable = executable + '.exe'
+
+    if os.path.isfile(executable):
+        return executable
+
+    if path is None:
+        path = os.environ.get('PATH', None)
+        # bpo-35755: Don't fall through if PATH is the empty string
+        if path is None:
+            try:
+                path = os.confstr("CS_PATH")
+            except (AttributeError, ValueError):
+                # os.confstr() or CS_PATH is not available
+                path = os.defpath
+
+    # PATH='' doesn't match, whereas PATH=':' looks in the current directory
+    if not path:
+        return None
+
+    paths = path.split(os.pathsep)
+    for p in paths:
+        f = os.path.join(p, executable)
+        if os.path.isfile(f):
+            # the file exists, we have a shot at spawn working
+            return f
+    return None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/sysconfig.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/sysconfig.py
new file mode 100644
index 00000000..ef3def83
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/sysconfig.py
@@ -0,0 +1,583 @@
+"""Provide access to Python's configuration information.  The specific
+configuration variables available depend heavily on the platform and
+configuration.  The values may be retrieved using
+get_config_var(name), and the list of variables is available via
+get_config_vars().keys().  Additional convenience functions are also
+available.
+
+Written by:   Fred L. Drake, Jr.
+Email:        <fdrake@acm.org>
+"""
+
+import functools
+import os
+import pathlib
+import re
+import sys
+import sysconfig
+
+from jaraco.functools import pass_none
+
+from .compat import py39
+from .errors import DistutilsPlatformError
+from .util import is_mingw
+
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+# These are needed in a couple of spots, so just compute them once.
+PREFIX = os.path.normpath(sys.prefix)
+EXEC_PREFIX = os.path.normpath(sys.exec_prefix)
+BASE_PREFIX = os.path.normpath(sys.base_prefix)
+BASE_EXEC_PREFIX = os.path.normpath(sys.base_exec_prefix)
+
+# Path to the base directory of the project. On Windows the binary may
+# live in project/PCbuild/win32 or project/PCbuild/amd64.
+# set for cross builds
+if "_PYTHON_PROJECT_BASE" in os.environ:
+    project_base = os.path.abspath(os.environ["_PYTHON_PROJECT_BASE"])
+else:
+    if sys.executable:
+        project_base = os.path.dirname(os.path.abspath(sys.executable))
+    else:
+        # sys.executable can be empty if argv[0] has been changed and Python is
+        # unable to retrieve the real program name
+        project_base = os.getcwd()
+
+
+def _is_python_source_dir(d):
+    """
+    Return True if the target directory appears to point to an
+    un-installed Python.
+    """
+    modules = pathlib.Path(d).joinpath('Modules')
+    return any(modules.joinpath(fn).is_file() for fn in ('Setup', 'Setup.local'))
+
+
+_sys_home = getattr(sys, '_home', None)
+
+
+def _is_parent(dir_a, dir_b):
+    """
+    Return True if a is a parent of b.
+    """
+    return os.path.normcase(dir_a).startswith(os.path.normcase(dir_b))
+
+
+if os.name == 'nt':
+
+    @pass_none
+    def _fix_pcbuild(d):
+        # In a venv, sys._home will be inside BASE_PREFIX rather than PREFIX.
+        prefixes = PREFIX, BASE_PREFIX
+        matched = (
+            prefix
+            for prefix in prefixes
+            if _is_parent(d, os.path.join(prefix, "PCbuild"))
+        )
+        return next(matched, d)
+
+    project_base = _fix_pcbuild(project_base)
+    _sys_home = _fix_pcbuild(_sys_home)
+
+
+def _python_build():
+    if _sys_home:
+        return _is_python_source_dir(_sys_home)
+    return _is_python_source_dir(project_base)
+
+
+python_build = _python_build()
+
+
+# Calculate the build qualifier flags if they are defined.  Adding the flags
+# to the include and lib directories only makes sense for an installation, not
+# an in-source build.
+build_flags = ''
+try:
+    if not python_build:
+        build_flags = sys.abiflags
+except AttributeError:
+    # It's not a configure-based build, so the sys module doesn't have
+    # this attribute, which is fine.
+    pass
+
+
+def get_python_version():
+    """Return a string containing the major and minor Python version,
+    leaving off the patchlevel.  Sample return values could be '1.5'
+    or '2.2'.
+    """
+    return f'{sys.version_info.major}.{sys.version_info.minor}'
+
+
+def get_python_inc(plat_specific=False, prefix=None):
+    """Return the directory containing installed Python header files.
+
+    If 'plat_specific' is false (the default), this is the path to the
+    non-platform-specific header files, i.e. Python.h and so on;
+    otherwise, this is the path to platform-specific header files
+    (namely pyconfig.h).
+
+    If 'prefix' is supplied, use it instead of sys.base_prefix or
+    sys.base_exec_prefix -- i.e., ignore 'plat_specific'.
+    """
+    default_prefix = BASE_EXEC_PREFIX if plat_specific else BASE_PREFIX
+    resolved_prefix = prefix if prefix is not None else default_prefix
+    # MinGW imitates posix like layout, but os.name != posix
+    os_name = "posix" if is_mingw() else os.name
+    try:
+        getter = globals()[f'_get_python_inc_{os_name}']
+    except KeyError:
+        raise DistutilsPlatformError(
+            "I don't know where Python installs its C header files "
+            f"on platform '{os.name}'"
+        )
+    return getter(resolved_prefix, prefix, plat_specific)
+
+
+@pass_none
+def _extant(path):
+    """
+    Replace path with None if it doesn't exist.
+    """
+    return path if os.path.exists(path) else None
+
+
+def _get_python_inc_posix(prefix, spec_prefix, plat_specific):
+    if IS_PYPY and sys.version_info < (3, 8):
+        return os.path.join(prefix, 'include')
+    return (
+        _get_python_inc_posix_python(plat_specific)
+        or _extant(_get_python_inc_from_config(plat_specific, spec_prefix))
+        or _get_python_inc_posix_prefix(prefix)
+    )
+
+
+def _get_python_inc_posix_python(plat_specific):
+    """
+    Assume the executable is in the build directory. The
+    pyconfig.h file should be in the same directory. Since
+    the build directory may not be the source directory,
+    use "srcdir" from the makefile to find the "Include"
+    directory.
+    """
+    if not python_build:
+        return
+    if plat_specific:
+        return _sys_home or project_base
+    incdir = os.path.join(get_config_var('srcdir'), 'Include')
+    return os.path.normpath(incdir)
+
+
+def _get_python_inc_from_config(plat_specific, spec_prefix):
+    """
+    If no prefix was explicitly specified, provide the include
+    directory from the config vars. Useful when
+    cross-compiling, since the config vars may come from
+    the host
+    platform Python installation, while the current Python
+    executable is from the build platform installation.
+
+    >>> monkeypatch = getfixture('monkeypatch')
+    >>> gpifc = _get_python_inc_from_config
+    >>> monkeypatch.setitem(gpifc.__globals__, 'get_config_var', str.lower)
+    >>> gpifc(False, '/usr/bin/')
+    >>> gpifc(False, '')
+    >>> gpifc(False, None)
+    'includepy'
+    >>> gpifc(True, None)
+    'confincludepy'
+    """
+    if spec_prefix is None:
+        return get_config_var('CONF' * plat_specific + 'INCLUDEPY')
+
+
+def _get_python_inc_posix_prefix(prefix):
+    implementation = 'pypy' if IS_PYPY else 'python'
+    python_dir = implementation + get_python_version() + build_flags
+    return os.path.join(prefix, "include", python_dir)
+
+
+def _get_python_inc_nt(prefix, spec_prefix, plat_specific):
+    if python_build:
+        # Include both include dirs to ensure we can find pyconfig.h
+        return (
+            os.path.join(prefix, "include")
+            + os.path.pathsep
+            + os.path.dirname(sysconfig.get_config_h_filename())
+        )
+    return os.path.join(prefix, "include")
+
+
+# allow this behavior to be monkey-patched. Ref pypa/distutils#2.
+def _posix_lib(standard_lib, libpython, early_prefix, prefix):
+    if standard_lib:
+        return libpython
+    else:
+        return os.path.join(libpython, "site-packages")
+
+
+def get_python_lib(plat_specific=False, standard_lib=False, prefix=None):
+    """Return the directory containing the Python library (standard or
+    site additions).
+
+    If 'plat_specific' is true, return the directory containing
+    platform-specific modules, i.e. any module from a non-pure-Python
+    module distribution; otherwise, return the platform-shared library
+    directory.  If 'standard_lib' is true, return the directory
+    containing standard Python library modules; otherwise, return the
+    directory for site-specific modules.
+
+    If 'prefix' is supplied, use it instead of sys.base_prefix or
+    sys.base_exec_prefix -- i.e., ignore 'plat_specific'.
+    """
+
+    if IS_PYPY and sys.version_info < (3, 8):
+        # PyPy-specific schema
+        if prefix is None:
+            prefix = PREFIX
+        if standard_lib:
+            return os.path.join(prefix, "lib-python", sys.version_info.major)
+        return os.path.join(prefix, 'site-packages')
+
+    early_prefix = prefix
+
+    if prefix is None:
+        if standard_lib:
+            prefix = plat_specific and BASE_EXEC_PREFIX or BASE_PREFIX
+        else:
+            prefix = plat_specific and EXEC_PREFIX or PREFIX
+
+    if os.name == "posix" or is_mingw():
+        if plat_specific or standard_lib:
+            # Platform-specific modules (any module from a non-pure-Python
+            # module distribution) or standard Python library modules.
+            libdir = getattr(sys, "platlibdir", "lib")
+        else:
+            # Pure Python
+            libdir = "lib"
+        implementation = 'pypy' if IS_PYPY else 'python'
+        libpython = os.path.join(prefix, libdir, implementation + get_python_version())
+        return _posix_lib(standard_lib, libpython, early_prefix, prefix)
+    elif os.name == "nt":
+        if standard_lib:
+            return os.path.join(prefix, "Lib")
+        else:
+            return os.path.join(prefix, "Lib", "site-packages")
+    else:
+        raise DistutilsPlatformError(
+            f"I don't know where Python installs its library on platform '{os.name}'"
+        )
+
+
+@functools.lru_cache
+def _customize_macos():
+    """
+    Perform first-time customization of compiler-related
+    config vars on macOS. Use after a compiler is known
+    to be needed. This customization exists primarily to support Pythons
+    from binary installers. The kind and paths to build tools on
+    the user system may vary significantly from the system
+    that Python itself was built on.  Also the user OS
+    version and build tools may not support the same set
+    of CPU architectures for universal builds.
+    """
+
+    sys.platform == "darwin" and __import__('_osx_support').customize_compiler(
+        get_config_vars()
+    )
+
+
+def customize_compiler(compiler):
+    """Do any platform-specific customization of a CCompiler instance.
+
+    Mainly needed on Unix, so we can plug in the information that
+    varies across Unices and is stored in Python's Makefile.
+    """
+    if compiler.compiler_type in ["unix", "cygwin"] or (
+        compiler.compiler_type == "mingw32" and is_mingw()
+    ):
+        _customize_macos()
+
+        (
+            cc,
+            cxx,
+            cflags,
+            ccshared,
+            ldshared,
+            ldcxxshared,
+            shlib_suffix,
+            ar,
+            ar_flags,
+        ) = get_config_vars(
+            'CC',
+            'CXX',
+            'CFLAGS',
+            'CCSHARED',
+            'LDSHARED',
+            'LDCXXSHARED',
+            'SHLIB_SUFFIX',
+            'AR',
+            'ARFLAGS',
+        )
+
+        cxxflags = cflags
+
+        if 'CC' in os.environ:
+            newcc = os.environ['CC']
+            if 'LDSHARED' not in os.environ and ldshared.startswith(cc):
+                # If CC is overridden, use that as the default
+                #       command for LDSHARED as well
+                ldshared = newcc + ldshared[len(cc) :]
+            cc = newcc
+        cxx = os.environ.get('CXX', cxx)
+        ldshared = os.environ.get('LDSHARED', ldshared)
+        ldcxxshared = os.environ.get('LDCXXSHARED', ldcxxshared)
+        cpp = os.environ.get(
+            'CPP',
+            cc + " -E",  # not always
+        )
+
+        ldshared = _add_flags(ldshared, 'LD')
+        ldcxxshared = _add_flags(ldcxxshared, 'LD')
+        cflags = os.environ.get('CFLAGS', cflags)
+        ldshared = _add_flags(ldshared, 'C')
+        cxxflags = os.environ.get('CXXFLAGS', cxxflags)
+        ldcxxshared = _add_flags(ldcxxshared, 'CXX')
+        cpp = _add_flags(cpp, 'CPP')
+        cflags = _add_flags(cflags, 'CPP')
+        cxxflags = _add_flags(cxxflags, 'CPP')
+        ldshared = _add_flags(ldshared, 'CPP')
+        ldcxxshared = _add_flags(ldcxxshared, 'CPP')
+
+        ar = os.environ.get('AR', ar)
+
+        archiver = ar + ' ' + os.environ.get('ARFLAGS', ar_flags)
+        cc_cmd = cc + ' ' + cflags
+        cxx_cmd = cxx + ' ' + cxxflags
+
+        compiler.set_executables(
+            preprocessor=cpp,
+            compiler=cc_cmd,
+            compiler_so=cc_cmd + ' ' + ccshared,
+            compiler_cxx=cxx_cmd,
+            compiler_so_cxx=cxx_cmd + ' ' + ccshared,
+            linker_so=ldshared,
+            linker_so_cxx=ldcxxshared,
+            linker_exe=cc,
+            linker_exe_cxx=cxx,
+            archiver=archiver,
+        )
+
+        if 'RANLIB' in os.environ and compiler.executables.get('ranlib', None):
+            compiler.set_executables(ranlib=os.environ['RANLIB'])
+
+        compiler.shared_lib_extension = shlib_suffix
+
+
+def get_config_h_filename():
+    """Return full pathname of installed pyconfig.h file."""
+    return sysconfig.get_config_h_filename()
+
+
+def get_makefile_filename():
+    """Return full pathname of installed Makefile from the Python build."""
+    return sysconfig.get_makefile_filename()
+
+
+def parse_config_h(fp, g=None):
+    """Parse a config.h-style file.
+
+    A dictionary containing name/value pairs is returned.  If an
+    optional dictionary is passed in as the second argument, it is
+    used instead of a new dictionary.
+    """
+    return sysconfig.parse_config_h(fp, vars=g)
+
+
+# Regexes needed for parsing Makefile (and similar syntaxes,
+# like old-style Setup files).
+_variable_rx = re.compile(r"([a-zA-Z][a-zA-Z0-9_]+)\s*=\s*(.*)")
+_findvar1_rx = re.compile(r"\$\(([A-Za-z][A-Za-z0-9_]*)\)")
+_findvar2_rx = re.compile(r"\${([A-Za-z][A-Za-z0-9_]*)}")
+
+
+def parse_makefile(fn, g=None):  # noqa: C901
+    """Parse a Makefile-style file.
+
+    A dictionary containing name/value pairs is returned.  If an
+    optional dictionary is passed in as the second argument, it is
+    used instead of a new dictionary.
+    """
+    from distutils.text_file import TextFile
+
+    fp = TextFile(
+        fn,
+        strip_comments=True,
+        skip_blanks=True,
+        join_lines=True,
+        errors="surrogateescape",
+    )
+
+    if g is None:
+        g = {}
+    done = {}
+    notdone = {}
+
+    while True:
+        line = fp.readline()
+        if line is None:  # eof
+            break
+        m = _variable_rx.match(line)
+        if m:
+            n, v = m.group(1, 2)
+            v = v.strip()
+            # `$$' is a literal `$' in make
+            tmpv = v.replace('$$', '')
+
+            if "$" in tmpv:
+                notdone[n] = v
+            else:
+                try:
+                    v = int(v)
+                except ValueError:
+                    # insert literal `$'
+                    done[n] = v.replace('$$', '$')
+                else:
+                    done[n] = v
+
+    # Variables with a 'PY_' prefix in the makefile. These need to
+    # be made available without that prefix through sysconfig.
+    # Special care is needed to ensure that variable expansion works, even
+    # if the expansion uses the name without a prefix.
+    renamed_variables = ('CFLAGS', 'LDFLAGS', 'CPPFLAGS')
+
+    # do variable interpolation here
+    while notdone:
+        for name in list(notdone):
+            value = notdone[name]
+            m = _findvar1_rx.search(value) or _findvar2_rx.search(value)
+            if m:
+                n = m.group(1)
+                found = True
+                if n in done:
+                    item = str(done[n])
+                elif n in notdone:
+                    # get it on a subsequent round
+                    found = False
+                elif n in os.environ:
+                    # do it like make: fall back to environment
+                    item = os.environ[n]
+
+                elif n in renamed_variables:
+                    if name.startswith('PY_') and name[3:] in renamed_variables:
+                        item = ""
+
+                    elif 'PY_' + n in notdone:
+                        found = False
+
+                    else:
+                        item = str(done['PY_' + n])
+                else:
+                    done[n] = item = ""
+                if found:
+                    after = value[m.end() :]
+                    value = value[: m.start()] + item + after
+                    if "$" in after:
+                        notdone[name] = value
+                    else:
+                        try:
+                            value = int(value)
+                        except ValueError:
+                            done[name] = value.strip()
+                        else:
+                            done[name] = value
+                        del notdone[name]
+
+                        if name.startswith('PY_') and name[3:] in renamed_variables:
+                            name = name[3:]
+                            if name not in done:
+                                done[name] = value
+            else:
+                # bogus variable reference; just drop it since we can't deal
+                del notdone[name]
+
+    fp.close()
+
+    # strip spurious spaces
+    for k, v in done.items():
+        if isinstance(v, str):
+            done[k] = v.strip()
+
+    # save the results in the global dictionary
+    g.update(done)
+    return g
+
+
+def expand_makefile_vars(s, vars):
+    """Expand Makefile-style variables -- "${foo}" or "$(foo)" -- in
+    'string' according to 'vars' (a dictionary mapping variable names to
+    values).  Variables not present in 'vars' are silently expanded to the
+    empty string.  The variable values in 'vars' should not contain further
+    variable expansions; if 'vars' is the output of 'parse_makefile()',
+    you're fine.  Returns a variable-expanded version of 's'.
+    """
+
+    # This algorithm does multiple expansion, so if vars['foo'] contains
+    # "${bar}", it will expand ${foo} to ${bar}, and then expand
+    # ${bar}... and so forth.  This is fine as long as 'vars' comes from
+    # 'parse_makefile()', which takes care of such expansions eagerly,
+    # according to make's variable expansion semantics.
+
+    while True:
+        m = _findvar1_rx.search(s) or _findvar2_rx.search(s)
+        if m:
+            (beg, end) = m.span()
+            s = s[0:beg] + vars.get(m.group(1)) + s[end:]
+        else:
+            break
+    return s
+
+
+_config_vars = None
+
+
+def get_config_vars(*args):
+    """With no arguments, return a dictionary of all configuration
+    variables relevant for the current platform.  Generally this includes
+    everything needed to build extensions and install both pure modules and
+    extensions.  On Unix, this means every variable defined in Python's
+    installed Makefile; on Windows it's a much smaller set.
+
+    With arguments, return a list of values that result from looking up
+    each argument in the configuration variable dictionary.
+    """
+    global _config_vars
+    if _config_vars is None:
+        _config_vars = sysconfig.get_config_vars().copy()
+        py39.add_ext_suffix(_config_vars)
+
+    return [_config_vars.get(name) for name in args] if args else _config_vars
+
+
+def get_config_var(name):
+    """Return the value of a single variable using the dictionary
+    returned by 'get_config_vars()'.  Equivalent to
+    get_config_vars().get(name)
+    """
+    if name == 'SO':
+        import warnings
+
+        warnings.warn('SO is deprecated, use EXT_SUFFIX', DeprecationWarning, 2)
+    return get_config_vars().get(name)
+
+
+@pass_none
+def _add_flags(value: str, type: str) -> str:
+    """
+    Add any flags from the environment for the given type.
+
+    type is the prefix to FLAGS in the environment key (e.g. "C" for "CFLAGS").
+    """
+    flags = os.environ.get(f'{type}FLAGS')
+    return f'{value} {flags}' if flags else value
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__init__.py
new file mode 100644
index 00000000..5a8ab061
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__init__.py
@@ -0,0 +1,42 @@
+"""
+Test suite for distutils.
+
+Tests for the command classes in the distutils.command package are
+included in distutils.tests as well, instead of using a separate
+distutils.command.tests package, since command identification is done
+by import rather than matching pre-defined names.
+"""
+
+import shutil
+from collections.abc import Sequence
+
+
+def missing_compiler_executable(cmd_names: Sequence[str] = []):  # pragma: no cover
+    """Check if the compiler components used to build the interpreter exist.
+
+    Check for the existence of the compiler executables whose names are listed
+    in 'cmd_names' or all the compiler executables when 'cmd_names' is empty
+    and return the first missing executable or None when none is found
+    missing.
+
+    """
+    from distutils import ccompiler, errors, sysconfig
+
+    compiler = ccompiler.new_compiler()
+    sysconfig.customize_compiler(compiler)
+    if compiler.compiler_type == "msvc":
+        # MSVC has no executables, so check whether initialization succeeds
+        try:
+            compiler.initialize()
+        except errors.DistutilsPlatformError:
+            return "msvc"
+    for name in compiler.executables:
+        if cmd_names and name not in cmd_names:
+            continue
+        cmd = getattr(compiler, name)
+        if cmd_names:
+            assert cmd is not None, f"the '{name}' executable is not configured"
+        elif not cmd:
+            continue
+        if shutil.which(cmd[0]) is None:
+            return cmd[0]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..8f1c410a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/support.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/support.cpython-312.pyc
new file mode 100644
index 00000000..584881bc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/support.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_archive_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_archive_util.cpython-312.pyc
new file mode 100644
index 00000000..8fddf6f7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_archive_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist.cpython-312.pyc
new file mode 100644
index 00000000..68db0bfc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_dumb.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_dumb.cpython-312.pyc
new file mode 100644
index 00000000..56063fd8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_dumb.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_rpm.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_rpm.cpython-312.pyc
new file mode 100644
index 00000000..b79a24fb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_bdist_rpm.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build.cpython-312.pyc
new file mode 100644
index 00000000..c33668fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_clib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_clib.cpython-312.pyc
new file mode 100644
index 00000000..1b406f5d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_clib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_ext.cpython-312.pyc
new file mode 100644
index 00000000..e551a217
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_py.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_py.cpython-312.pyc
new file mode 100644
index 00000000..aacc2d97
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_py.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_scripts.cpython-312.pyc
new file mode 100644
index 00000000..3bc74ea3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_build_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_ccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_ccompiler.cpython-312.pyc
new file mode 100644
index 00000000..8f83bad9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_ccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_check.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_check.cpython-312.pyc
new file mode 100644
index 00000000..bcb293d3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_check.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_clean.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_clean.cpython-312.pyc
new file mode 100644
index 00000000..31b949c3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_clean.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cmd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cmd.cpython-312.pyc
new file mode 100644
index 00000000..7e7d92bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cmd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_config_cmd.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_config_cmd.cpython-312.pyc
new file mode 100644
index 00000000..d2bc2314
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_config_cmd.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_core.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_core.cpython-312.pyc
new file mode 100644
index 00000000..826830df
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_core.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cygwinccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cygwinccompiler.cpython-312.pyc
new file mode 100644
index 00000000..a3794c61
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_cygwinccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dir_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dir_util.cpython-312.pyc
new file mode 100644
index 00000000..c95ddd06
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dir_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dist.cpython-312.pyc
new file mode 100644
index 00000000..e11819b3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_dist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_extension.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_extension.cpython-312.pyc
new file mode 100644
index 00000000..215acfd8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_extension.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_file_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_file_util.cpython-312.pyc
new file mode 100644
index 00000000..0d72fe6a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_file_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_filelist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_filelist.cpython-312.pyc
new file mode 100644
index 00000000..357620a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_filelist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install.cpython-312.pyc
new file mode 100644
index 00000000..fbab039f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_data.cpython-312.pyc
new file mode 100644
index 00000000..d952fe66
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_headers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_headers.cpython-312.pyc
new file mode 100644
index 00000000..0b7c3368
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_headers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_lib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_lib.cpython-312.pyc
new file mode 100644
index 00000000..3d5ea2b4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_lib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_scripts.cpython-312.pyc
new file mode 100644
index 00000000..81d7d4b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_install_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_log.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_log.cpython-312.pyc
new file mode 100644
index 00000000..67f8ec49
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_log.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_mingwccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_mingwccompiler.cpython-312.pyc
new file mode 100644
index 00000000..15dff2c5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_mingwccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_modified.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_modified.cpython-312.pyc
new file mode 100644
index 00000000..9c4ada51
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_modified.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_msvccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_msvccompiler.cpython-312.pyc
new file mode 100644
index 00000000..86b0cf6a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_msvccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sdist.cpython-312.pyc
new file mode 100644
index 00000000..028dc088
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_spawn.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_spawn.cpython-312.pyc
new file mode 100644
index 00000000..4d2ca96b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_spawn.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sysconfig.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sysconfig.cpython-312.pyc
new file mode 100644
index 00000000..ed6d2ed9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_sysconfig.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_text_file.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_text_file.cpython-312.pyc
new file mode 100644
index 00000000..21633b01
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_text_file.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_unixccompiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_unixccompiler.cpython-312.pyc
new file mode 100644
index 00000000..fc546a93
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_unixccompiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_util.cpython-312.pyc
new file mode 100644
index 00000000..1ca9bc69
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_version.cpython-312.pyc
new file mode 100644
index 00000000..8e1e7a52
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_versionpredicate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_versionpredicate.cpython-312.pyc
new file mode 100644
index 00000000..7711c0c8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/test_versionpredicate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/unix_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/unix_compat.cpython-312.pyc
new file mode 100644
index 00000000..67148066
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/__pycache__/unix_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3deb7a96
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/py39.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/py39.cpython-312.pyc
new file mode 100644
index 00000000..31c8da5e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/__pycache__/py39.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/py39.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/py39.py
new file mode 100644
index 00000000..aca3939a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/compat/py39.py
@@ -0,0 +1,40 @@
+import sys
+
+if sys.version_info >= (3, 10):
+    from test.support.import_helper import (
+        CleanImport as CleanImport,
+    )
+    from test.support.import_helper import (
+        DirsOnSysPath as DirsOnSysPath,
+    )
+    from test.support.os_helper import (
+        EnvironmentVarGuard as EnvironmentVarGuard,
+    )
+    from test.support.os_helper import (
+        rmtree as rmtree,
+    )
+    from test.support.os_helper import (
+        skip_unless_symlink as skip_unless_symlink,
+    )
+    from test.support.os_helper import (
+        unlink as unlink,
+    )
+else:
+    from test.support import (
+        CleanImport as CleanImport,
+    )
+    from test.support import (
+        DirsOnSysPath as DirsOnSysPath,
+    )
+    from test.support import (
+        EnvironmentVarGuard as EnvironmentVarGuard,
+    )
+    from test.support import (
+        rmtree as rmtree,
+    )
+    from test.support import (
+        skip_unless_symlink as skip_unless_symlink,
+    )
+    from test.support import (
+        unlink as unlink,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/support.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/support.py
new file mode 100644
index 00000000..9cd2b8a9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/support.py
@@ -0,0 +1,134 @@
+"""Support code for distutils test cases."""
+
+import itertools
+import os
+import pathlib
+import shutil
+import sys
+import sysconfig
+import tempfile
+from distutils.core import Distribution
+
+import pytest
+from more_itertools import always_iterable
+
+
+@pytest.mark.usefixtures('distutils_managed_tempdir')
+class TempdirManager:
+    """
+    Mix-in class that handles temporary directories for test cases.
+    """
+
+    def mkdtemp(self):
+        """Create a temporary directory that will be cleaned up.
+
+        Returns the path of the directory.
+        """
+        d = tempfile.mkdtemp()
+        self.tempdirs.append(d)
+        return d
+
+    def write_file(self, path, content='xxx'):
+        """Writes a file in the given path.
+
+        path can be a string or a sequence.
+        """
+        pathlib.Path(*always_iterable(path)).write_text(content, encoding='utf-8')
+
+    def create_dist(self, pkg_name='foo', **kw):
+        """Will generate a test environment.
+
+        This function creates:
+         - a Distribution instance using keywords
+         - a temporary directory with a package structure
+
+        It returns the package directory and the distribution
+        instance.
+        """
+        tmp_dir = self.mkdtemp()
+        pkg_dir = os.path.join(tmp_dir, pkg_name)
+        os.mkdir(pkg_dir)
+        dist = Distribution(attrs=kw)
+
+        return pkg_dir, dist
+
+
+class DummyCommand:
+    """Class to store options for retrieval via set_undefined_options()."""
+
+    def __init__(self, **kwargs):
+        vars(self).update(kwargs)
+
+    def ensure_finalized(self):
+        pass
+
+
+def copy_xxmodule_c(directory):
+    """Helper for tests that need the xxmodule.c source file.
+
+    Example use:
+
+        def test_compile(self):
+            copy_xxmodule_c(self.tmpdir)
+            self.assertIn('xxmodule.c', os.listdir(self.tmpdir))
+
+    If the source file can be found, it will be copied to *directory*.  If not,
+    the test will be skipped.  Errors during copy are not caught.
+    """
+    shutil.copy(_get_xxmodule_path(), os.path.join(directory, 'xxmodule.c'))
+
+
+def _get_xxmodule_path():
+    source_name = 'xxmodule.c' if sys.version_info > (3, 9) else 'xxmodule-3.8.c'
+    return os.path.join(os.path.dirname(__file__), source_name)
+
+
+def fixup_build_ext(cmd):
+    """Function needed to make build_ext tests pass.
+
+    When Python was built with --enable-shared on Unix, -L. is not enough to
+    find libpython<blah>.so, because regrtest runs in a tempdir, not in the
+    source directory where the .so lives.
+
+    When Python was built with in debug mode on Windows, build_ext commands
+    need their debug attribute set, and it is not done automatically for
+    some reason.
+
+    This function handles both of these things.  Example use:
+
+        cmd = build_ext(dist)
+        support.fixup_build_ext(cmd)
+        cmd.ensure_finalized()
+
+    Unlike most other Unix platforms, Mac OS X embeds absolute paths
+    to shared libraries into executables, so the fixup is not needed there.
+    """
+    if os.name == 'nt':
+        cmd.debug = sys.executable.endswith('_d.exe')
+    elif sysconfig.get_config_var('Py_ENABLE_SHARED'):
+        # To further add to the shared builds fun on Unix, we can't just add
+        # library_dirs to the Extension() instance because that doesn't get
+        # plumbed through to the final compiler command.
+        runshared = sysconfig.get_config_var('RUNSHARED')
+        if runshared is None:
+            cmd.library_dirs = ['.']
+        else:
+            if sys.platform == 'darwin':
+                cmd.library_dirs = []
+            else:
+                name, equals, value = runshared.partition('=')
+                cmd.library_dirs = [d for d in value.split(os.pathsep) if d]
+
+
+def combine_markers(cls):
+    """
+    pytest will honor markers as found on the class, but when
+    markers are on multiple subclasses, only one appears. Use
+    this decorator to combine those markers.
+    """
+    cls.pytestmark = [
+        mark
+        for base in itertools.chain([cls], cls.__bases__)
+        for mark in getattr(base, 'pytestmark', [])
+    ]
+    return cls
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_archive_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_archive_util.py
new file mode 100644
index 00000000..3e4ed75a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_archive_util.py
@@ -0,0 +1,353 @@
+"""Tests for distutils.archive_util."""
+
+import functools
+import operator
+import os
+import pathlib
+import sys
+import tarfile
+from distutils import archive_util
+from distutils.archive_util import (
+    ARCHIVE_FORMATS,
+    check_archive_formats,
+    make_archive,
+    make_tarball,
+    make_zipfile,
+)
+from distutils.spawn import spawn
+from distutils.tests import support
+from os.path import splitdrive
+
+import path
+import pytest
+from test.support import patch
+
+from .unix_compat import UID_0_SUPPORT, grp, pwd, require_uid_0, require_unix_id
+
+
+def can_fs_encode(filename):
+    """
+    Return True if the filename can be saved in the file system.
+    """
+    if os.path.supports_unicode_filenames:
+        return True
+    try:
+        filename.encode(sys.getfilesystemencoding())
+    except UnicodeEncodeError:
+        return False
+    return True
+
+
+def all_equal(values):
+    return functools.reduce(operator.eq, values)
+
+
+def same_drive(*paths):
+    return all_equal(pathlib.Path(path).drive for path in paths)
+
+
+class ArchiveUtilTestCase(support.TempdirManager):
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_make_tarball(self, name='archive'):
+        # creating something to tar
+        tmpdir = self._create_files()
+        self._make_tarball(tmpdir, name, '.tar.gz')
+        # trying an uncompressed one
+        self._make_tarball(tmpdir, name, '.tar', compress=None)
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_make_tarball_gzip(self):
+        tmpdir = self._create_files()
+        self._make_tarball(tmpdir, 'archive', '.tar.gz', compress='gzip')
+
+    def test_make_tarball_bzip2(self):
+        pytest.importorskip('bz2')
+        tmpdir = self._create_files()
+        self._make_tarball(tmpdir, 'archive', '.tar.bz2', compress='bzip2')
+
+    def test_make_tarball_xz(self):
+        pytest.importorskip('lzma')
+        tmpdir = self._create_files()
+        self._make_tarball(tmpdir, 'archive', '.tar.xz', compress='xz')
+
+    @pytest.mark.skipif("not can_fs_encode('årchiv')")
+    def test_make_tarball_latin1(self):
+        """
+        Mirror test_make_tarball, except filename contains latin characters.
+        """
+        self.test_make_tarball('årchiv')  # note this isn't a real word
+
+    @pytest.mark.skipif("not can_fs_encode('のアーカイブ')")
+    def test_make_tarball_extended(self):
+        """
+        Mirror test_make_tarball, except filename contains extended
+        characters outside the latin charset.
+        """
+        self.test_make_tarball('のアーカイブ')  # japanese for archive
+
+    def _make_tarball(self, tmpdir, target_name, suffix, **kwargs):
+        tmpdir2 = self.mkdtemp()
+        if same_drive(tmpdir, tmpdir2):
+            pytest.skip("source and target should be on same drive")
+
+        base_name = os.path.join(tmpdir2, target_name)
+
+        # working with relative paths to avoid tar warnings
+        with path.Path(tmpdir):
+            make_tarball(splitdrive(base_name)[1], 'dist', **kwargs)
+
+        # check if the compressed tarball was created
+        tarball = base_name + suffix
+        assert os.path.exists(tarball)
+        assert self._tarinfo(tarball) == self._created_files
+
+    def _tarinfo(self, path):
+        tar = tarfile.open(path)
+        try:
+            names = tar.getnames()
+            names.sort()
+            return names
+        finally:
+            tar.close()
+
+    _zip_created_files = [
+        'dist/',
+        'dist/file1',
+        'dist/file2',
+        'dist/sub/',
+        'dist/sub/file3',
+        'dist/sub2/',
+    ]
+    _created_files = [p.rstrip('/') for p in _zip_created_files]
+
+    def _create_files(self):
+        # creating something to tar
+        tmpdir = self.mkdtemp()
+        dist = os.path.join(tmpdir, 'dist')
+        os.mkdir(dist)
+        self.write_file([dist, 'file1'], 'xxx')
+        self.write_file([dist, 'file2'], 'xxx')
+        os.mkdir(os.path.join(dist, 'sub'))
+        self.write_file([dist, 'sub', 'file3'], 'xxx')
+        os.mkdir(os.path.join(dist, 'sub2'))
+        return tmpdir
+
+    @pytest.mark.usefixtures('needs_zlib')
+    @pytest.mark.skipif("not (shutil.which('tar') and shutil.which('gzip'))")
+    def test_tarfile_vs_tar(self):
+        tmpdir = self._create_files()
+        tmpdir2 = self.mkdtemp()
+        base_name = os.path.join(tmpdir2, 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist')
+        finally:
+            os.chdir(old_dir)
+
+        # check if the compressed tarball was created
+        tarball = base_name + '.tar.gz'
+        assert os.path.exists(tarball)
+
+        # now create another tarball using `tar`
+        tarball2 = os.path.join(tmpdir, 'archive2.tar.gz')
+        tar_cmd = ['tar', '-cf', 'archive2.tar', 'dist']
+        gzip_cmd = ['gzip', '-f', '-9', 'archive2.tar']
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            spawn(tar_cmd)
+            spawn(gzip_cmd)
+        finally:
+            os.chdir(old_dir)
+
+        assert os.path.exists(tarball2)
+        # let's compare both tarballs
+        assert self._tarinfo(tarball) == self._created_files
+        assert self._tarinfo(tarball2) == self._created_files
+
+        # trying an uncompressed one
+        base_name = os.path.join(tmpdir2, 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist', compress=None)
+        finally:
+            os.chdir(old_dir)
+        tarball = base_name + '.tar'
+        assert os.path.exists(tarball)
+
+        # now for a dry_run
+        base_name = os.path.join(tmpdir2, 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist', compress=None, dry_run=True)
+        finally:
+            os.chdir(old_dir)
+        tarball = base_name + '.tar'
+        assert os.path.exists(tarball)
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_make_zipfile(self):
+        zipfile = pytest.importorskip('zipfile')
+        # creating something to tar
+        tmpdir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        with path.Path(tmpdir):
+            make_zipfile(base_name, 'dist')
+
+        # check if the compressed tarball was created
+        tarball = base_name + '.zip'
+        assert os.path.exists(tarball)
+        with zipfile.ZipFile(tarball) as zf:
+            assert sorted(zf.namelist()) == self._zip_created_files
+
+    def test_make_zipfile_no_zlib(self):
+        zipfile = pytest.importorskip('zipfile')
+        patch(self, archive_util.zipfile, 'zlib', None)  # force zlib ImportError
+
+        called = []
+        zipfile_class = zipfile.ZipFile
+
+        def fake_zipfile(*a, **kw):
+            if kw.get('compression', None) == zipfile.ZIP_STORED:
+                called.append((a, kw))
+            return zipfile_class(*a, **kw)
+
+        patch(self, archive_util.zipfile, 'ZipFile', fake_zipfile)
+
+        # create something to tar and compress
+        tmpdir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        with path.Path(tmpdir):
+            make_zipfile(base_name, 'dist')
+
+        tarball = base_name + '.zip'
+        assert called == [((tarball, "w"), {'compression': zipfile.ZIP_STORED})]
+        assert os.path.exists(tarball)
+        with zipfile.ZipFile(tarball) as zf:
+            assert sorted(zf.namelist()) == self._zip_created_files
+
+    def test_check_archive_formats(self):
+        assert check_archive_formats(['gztar', 'xxx', 'zip']) == 'xxx'
+        assert (
+            check_archive_formats(['gztar', 'bztar', 'xztar', 'ztar', 'tar', 'zip'])
+            is None
+        )
+
+    def test_make_archive(self):
+        tmpdir = self.mkdtemp()
+        base_name = os.path.join(tmpdir, 'archive')
+        with pytest.raises(ValueError):
+            make_archive(base_name, 'xxx')
+
+    def test_make_archive_cwd(self):
+        current_dir = os.getcwd()
+
+        def _breaks(*args, **kw):
+            raise RuntimeError()
+
+        ARCHIVE_FORMATS['xxx'] = (_breaks, [], 'xxx file')
+        try:
+            try:
+                make_archive('xxx', 'xxx', root_dir=self.mkdtemp())
+            except Exception:
+                pass
+            assert os.getcwd() == current_dir
+        finally:
+            ARCHIVE_FORMATS.pop('xxx')
+
+    def test_make_archive_tar(self):
+        base_dir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        res = make_archive(base_name, 'tar', base_dir, 'dist')
+        assert os.path.exists(res)
+        assert os.path.basename(res) == 'archive.tar'
+        assert self._tarinfo(res) == self._created_files
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_make_archive_gztar(self):
+        base_dir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        res = make_archive(base_name, 'gztar', base_dir, 'dist')
+        assert os.path.exists(res)
+        assert os.path.basename(res) == 'archive.tar.gz'
+        assert self._tarinfo(res) == self._created_files
+
+    def test_make_archive_bztar(self):
+        pytest.importorskip('bz2')
+        base_dir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        res = make_archive(base_name, 'bztar', base_dir, 'dist')
+        assert os.path.exists(res)
+        assert os.path.basename(res) == 'archive.tar.bz2'
+        assert self._tarinfo(res) == self._created_files
+
+    def test_make_archive_xztar(self):
+        pytest.importorskip('lzma')
+        base_dir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        res = make_archive(base_name, 'xztar', base_dir, 'dist')
+        assert os.path.exists(res)
+        assert os.path.basename(res) == 'archive.tar.xz'
+        assert self._tarinfo(res) == self._created_files
+
+    def test_make_archive_owner_group(self):
+        # testing make_archive with owner and group, with various combinations
+        # this works even if there's not gid/uid support
+        if UID_0_SUPPORT:
+            group = grp.getgrgid(0)[0]
+            owner = pwd.getpwuid(0)[0]
+        else:
+            group = owner = 'root'
+
+        base_dir = self._create_files()
+        root_dir = self.mkdtemp()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        res = make_archive(
+            base_name, 'zip', root_dir, base_dir, owner=owner, group=group
+        )
+        assert os.path.exists(res)
+
+        res = make_archive(base_name, 'zip', root_dir, base_dir)
+        assert os.path.exists(res)
+
+        res = make_archive(
+            base_name, 'tar', root_dir, base_dir, owner=owner, group=group
+        )
+        assert os.path.exists(res)
+
+        res = make_archive(
+            base_name, 'tar', root_dir, base_dir, owner='kjhkjhkjg', group='oihohoh'
+        )
+        assert os.path.exists(res)
+
+    @pytest.mark.usefixtures('needs_zlib')
+    @require_unix_id
+    @require_uid_0
+    def test_tarfile_root_owner(self):
+        tmpdir = self._create_files()
+        base_name = os.path.join(self.mkdtemp(), 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        group = grp.getgrgid(0)[0]
+        owner = pwd.getpwuid(0)[0]
+        try:
+            archive_name = make_tarball(
+                base_name, 'dist', compress=None, owner=owner, group=group
+            )
+        finally:
+            os.chdir(old_dir)
+
+        # check if the compressed tarball was created
+        assert os.path.exists(archive_name)
+
+        # now checks the rights
+        archive = tarfile.open(archive_name)
+        try:
+            for member in archive.getmembers():
+                assert member.uid == 0
+                assert member.gid == 0
+        finally:
+            archive.close()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist.py
new file mode 100644
index 00000000..d5696fc3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist.py
@@ -0,0 +1,47 @@
+"""Tests for distutils.command.bdist."""
+
+from distutils.command.bdist import bdist
+from distutils.tests import support
+
+
+class TestBuild(support.TempdirManager):
+    def test_formats(self):
+        # let's create a command and make sure
+        # we can set the format
+        dist = self.create_dist()[1]
+        cmd = bdist(dist)
+        cmd.formats = ['gztar']
+        cmd.ensure_finalized()
+        assert cmd.formats == ['gztar']
+
+        # what formats does bdist offer?
+        formats = [
+            'bztar',
+            'gztar',
+            'rpm',
+            'tar',
+            'xztar',
+            'zip',
+            'ztar',
+        ]
+        found = sorted(cmd.format_commands)
+        assert found == formats
+
+    def test_skip_build(self):
+        # bug #10946: bdist --skip-build should trickle down to subcommands
+        dist = self.create_dist()[1]
+        cmd = bdist(dist)
+        cmd.skip_build = True
+        cmd.ensure_finalized()
+        dist.command_obj['bdist'] = cmd
+
+        names = [
+            'bdist_dumb',
+        ]  # bdist_rpm does not support --skip-build
+
+        for name in names:
+            subcmd = cmd.get_finalized_command(name)
+            if getattr(subcmd, '_unsupported', False):
+                # command is not supported on this build
+                continue
+            assert subcmd.skip_build, f'{name} should take --skip-build from bdist'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_dumb.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_dumb.py
new file mode 100644
index 00000000..1fc51d24
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_dumb.py
@@ -0,0 +1,78 @@
+"""Tests for distutils.command.bdist_dumb."""
+
+import os
+import sys
+import zipfile
+from distutils.command.bdist_dumb import bdist_dumb
+from distutils.core import Distribution
+from distutils.tests import support
+
+import pytest
+
+SETUP_PY = """\
+from distutils.core import setup
+import foo
+
+setup(name='foo', version='0.1', py_modules=['foo'],
+      url='xxx', author='xxx', author_email='xxx')
+
+"""
+
+
+@support.combine_markers
+@pytest.mark.usefixtures('save_env')
+@pytest.mark.usefixtures('save_argv')
+@pytest.mark.usefixtures('save_cwd')
+class TestBuildDumb(
+    support.TempdirManager,
+):
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_simple_built(self):
+        # let's create a simple package
+        tmp_dir = self.mkdtemp()
+        pkg_dir = os.path.join(tmp_dir, 'foo')
+        os.mkdir(pkg_dir)
+        self.write_file((pkg_dir, 'setup.py'), SETUP_PY)
+        self.write_file((pkg_dir, 'foo.py'), '#')
+        self.write_file((pkg_dir, 'MANIFEST.in'), 'include foo.py')
+        self.write_file((pkg_dir, 'README'), '')
+
+        dist = Distribution({
+            'name': 'foo',
+            'version': '0.1',
+            'py_modules': ['foo'],
+            'url': 'xxx',
+            'author': 'xxx',
+            'author_email': 'xxx',
+        })
+        dist.script_name = 'setup.py'
+        os.chdir(pkg_dir)
+
+        sys.argv = ['setup.py']
+        cmd = bdist_dumb(dist)
+
+        # so the output is the same no matter
+        # what is the platform
+        cmd.format = 'zip'
+
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # see what we have
+        dist_created = os.listdir(os.path.join(pkg_dir, 'dist'))
+        base = f"{dist.get_fullname()}.{cmd.plat_name}.zip"
+
+        assert dist_created == [base]
+
+        # now let's check what we have in the zip file
+        fp = zipfile.ZipFile(os.path.join('dist', base))
+        try:
+            contents = fp.namelist()
+        finally:
+            fp.close()
+
+        contents = sorted(filter(None, map(os.path.basename, contents)))
+        wanted = ['foo-0.1-py{}.{}.egg-info'.format(*sys.version_info[:2]), 'foo.py']
+        if not sys.dont_write_bytecode:
+            wanted.append(f'foo.{sys.implementation.cache_tag}.pyc')
+        assert contents == sorted(wanted)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_rpm.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_rpm.py
new file mode 100644
index 00000000..75051430
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_bdist_rpm.py
@@ -0,0 +1,127 @@
+"""Tests for distutils.command.bdist_rpm."""
+
+import os
+import shutil  # noqa: F401
+import sys
+from distutils.command.bdist_rpm import bdist_rpm
+from distutils.core import Distribution
+from distutils.tests import support
+
+import pytest
+from test.support import requires_zlib
+
+SETUP_PY = """\
+from distutils.core import setup
+import foo
+
+setup(name='foo', version='0.1', py_modules=['foo'],
+      url='xxx', author='xxx', author_email='xxx')
+
+"""
+
+
+@pytest.fixture(autouse=True)
+def sys_executable_encodable():
+    try:
+        sys.executable.encode('UTF-8')
+    except UnicodeEncodeError:
+        pytest.skip("sys.executable is not encodable to UTF-8")
+
+
+mac_woes = pytest.mark.skipif(
+    "not sys.platform.startswith('linux')",
+    reason='spurious sdtout/stderr output under macOS',
+)
+
+
+@pytest.mark.usefixtures('save_env')
+@pytest.mark.usefixtures('save_argv')
+@pytest.mark.usefixtures('save_cwd')
+class TestBuildRpm(
+    support.TempdirManager,
+):
+    @mac_woes
+    @requires_zlib()
+    @pytest.mark.skipif("not shutil.which('rpm')")
+    @pytest.mark.skipif("not shutil.which('rpmbuild')")
+    def test_quiet(self):
+        # let's create a package
+        tmp_dir = self.mkdtemp()
+        os.environ['HOME'] = tmp_dir  # to confine dir '.rpmdb' creation
+        pkg_dir = os.path.join(tmp_dir, 'foo')
+        os.mkdir(pkg_dir)
+        self.write_file((pkg_dir, 'setup.py'), SETUP_PY)
+        self.write_file((pkg_dir, 'foo.py'), '#')
+        self.write_file((pkg_dir, 'MANIFEST.in'), 'include foo.py')
+        self.write_file((pkg_dir, 'README'), '')
+
+        dist = Distribution({
+            'name': 'foo',
+            'version': '0.1',
+            'py_modules': ['foo'],
+            'url': 'xxx',
+            'author': 'xxx',
+            'author_email': 'xxx',
+        })
+        dist.script_name = 'setup.py'
+        os.chdir(pkg_dir)
+
+        sys.argv = ['setup.py']
+        cmd = bdist_rpm(dist)
+        cmd.fix_python = True
+
+        # running in quiet mode
+        cmd.quiet = True
+        cmd.ensure_finalized()
+        cmd.run()
+
+        dist_created = os.listdir(os.path.join(pkg_dir, 'dist'))
+        assert 'foo-0.1-1.noarch.rpm' in dist_created
+
+        # bug #2945: upload ignores bdist_rpm files
+        assert ('bdist_rpm', 'any', 'dist/foo-0.1-1.src.rpm') in dist.dist_files
+        assert ('bdist_rpm', 'any', 'dist/foo-0.1-1.noarch.rpm') in dist.dist_files
+
+    @mac_woes
+    @requires_zlib()
+    # https://bugs.python.org/issue1533164
+    @pytest.mark.skipif("not shutil.which('rpm')")
+    @pytest.mark.skipif("not shutil.which('rpmbuild')")
+    def test_no_optimize_flag(self):
+        # let's create a package that breaks bdist_rpm
+        tmp_dir = self.mkdtemp()
+        os.environ['HOME'] = tmp_dir  # to confine dir '.rpmdb' creation
+        pkg_dir = os.path.join(tmp_dir, 'foo')
+        os.mkdir(pkg_dir)
+        self.write_file((pkg_dir, 'setup.py'), SETUP_PY)
+        self.write_file((pkg_dir, 'foo.py'), '#')
+        self.write_file((pkg_dir, 'MANIFEST.in'), 'include foo.py')
+        self.write_file((pkg_dir, 'README'), '')
+
+        dist = Distribution({
+            'name': 'foo',
+            'version': '0.1',
+            'py_modules': ['foo'],
+            'url': 'xxx',
+            'author': 'xxx',
+            'author_email': 'xxx',
+        })
+        dist.script_name = 'setup.py'
+        os.chdir(pkg_dir)
+
+        sys.argv = ['setup.py']
+        cmd = bdist_rpm(dist)
+        cmd.fix_python = True
+
+        cmd.quiet = True
+        cmd.ensure_finalized()
+        cmd.run()
+
+        dist_created = os.listdir(os.path.join(pkg_dir, 'dist'))
+        assert 'foo-0.1-1.noarch.rpm' in dist_created
+
+        # bug #2945: upload ignores bdist_rpm files
+        assert ('bdist_rpm', 'any', 'dist/foo-0.1-1.src.rpm') in dist.dist_files
+        assert ('bdist_rpm', 'any', 'dist/foo-0.1-1.noarch.rpm') in dist.dist_files
+
+        os.remove(os.path.join(pkg_dir, 'dist', 'foo-0.1-1.noarch.rpm'))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build.py
new file mode 100644
index 00000000..f7fe69ac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build.py
@@ -0,0 +1,49 @@
+"""Tests for distutils.command.build."""
+
+import os
+import sys
+from distutils.command.build import build
+from distutils.tests import support
+from sysconfig import get_config_var, get_platform
+
+
+class TestBuild(support.TempdirManager):
+    def test_finalize_options(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build(dist)
+        cmd.finalize_options()
+
+        # if not specified, plat_name gets the current platform
+        assert cmd.plat_name == get_platform()
+
+        # build_purelib is build + lib
+        wanted = os.path.join(cmd.build_base, 'lib')
+        assert cmd.build_purelib == wanted
+
+        # build_platlib is 'build/lib.platform-cache_tag[-pydebug]'
+        # examples:
+        #   build/lib.macosx-10.3-i386-cpython39
+        plat_spec = f'.{cmd.plat_name}-{sys.implementation.cache_tag}'
+        if get_config_var('Py_GIL_DISABLED'):
+            plat_spec += 't'
+        if hasattr(sys, 'gettotalrefcount'):
+            assert cmd.build_platlib.endswith('-pydebug')
+            plat_spec += '-pydebug'
+        wanted = os.path.join(cmd.build_base, 'lib' + plat_spec)
+        assert cmd.build_platlib == wanted
+
+        # by default, build_lib = build_purelib
+        assert cmd.build_lib == cmd.build_purelib
+
+        # build_temp is build/temp.<plat>
+        wanted = os.path.join(cmd.build_base, 'temp' + plat_spec)
+        assert cmd.build_temp == wanted
+
+        # build_scripts is build/scripts-x.x
+        wanted = os.path.join(
+            cmd.build_base, f'scripts-{sys.version_info.major}.{sys.version_info.minor}'
+        )
+        assert cmd.build_scripts == wanted
+
+        # executable is os.path.normpath(sys.executable)
+        assert cmd.executable == os.path.normpath(sys.executable)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_clib.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_clib.py
new file mode 100644
index 00000000..f76f26bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_clib.py
@@ -0,0 +1,134 @@
+"""Tests for distutils.command.build_clib."""
+
+import os
+from distutils.command.build_clib import build_clib
+from distutils.errors import DistutilsSetupError
+from distutils.tests import missing_compiler_executable, support
+
+import pytest
+
+
+class TestBuildCLib(support.TempdirManager):
+    def test_check_library_dist(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build_clib(dist)
+
+        # 'libraries' option must be a list
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_library_list('foo')
+
+        # each element of 'libraries' must a 2-tuple
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_library_list(['foo1', 'foo2'])
+
+        # first element of each tuple in 'libraries'
+        # must be a string (the library name)
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_library_list([(1, 'foo1'), ('name', 'foo2')])
+
+        # library name may not contain directory separators
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_library_list(
+                [('name', 'foo1'), ('another/name', 'foo2')],
+            )
+
+        # second element of each tuple must be a dictionary (build info)
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_library_list(
+                [('name', {}), ('another', 'foo2')],
+            )
+
+        # those work
+        libs = [('name', {}), ('name', {'ok': 'good'})]
+        cmd.check_library_list(libs)
+
+    def test_get_source_files(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build_clib(dist)
+
+        # "in 'libraries' option 'sources' must be present and must be
+        # a list of source filenames
+        cmd.libraries = [('name', {})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.get_source_files()
+
+        cmd.libraries = [('name', {'sources': 1})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.get_source_files()
+
+        cmd.libraries = [('name', {'sources': ['a', 'b']})]
+        assert cmd.get_source_files() == ['a', 'b']
+
+        cmd.libraries = [('name', {'sources': ('a', 'b')})]
+        assert cmd.get_source_files() == ['a', 'b']
+
+        cmd.libraries = [
+            ('name', {'sources': ('a', 'b')}),
+            ('name2', {'sources': ['c', 'd']}),
+        ]
+        assert cmd.get_source_files() == ['a', 'b', 'c', 'd']
+
+    def test_build_libraries(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build_clib(dist)
+
+        class FakeCompiler:
+            def compile(*args, **kw):
+                pass
+
+            create_static_lib = compile
+
+        cmd.compiler = FakeCompiler()
+
+        # build_libraries is also doing a bit of typo checking
+        lib = [('name', {'sources': 'notvalid'})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.build_libraries(lib)
+
+        lib = [('name', {'sources': list()})]
+        cmd.build_libraries(lib)
+
+        lib = [('name', {'sources': tuple()})]
+        cmd.build_libraries(lib)
+
+    def test_finalize_options(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build_clib(dist)
+
+        cmd.include_dirs = 'one-dir'
+        cmd.finalize_options()
+        assert cmd.include_dirs == ['one-dir']
+
+        cmd.include_dirs = None
+        cmd.finalize_options()
+        assert cmd.include_dirs == []
+
+        cmd.distribution.libraries = 'WONTWORK'
+        with pytest.raises(DistutilsSetupError):
+            cmd.finalize_options()
+
+    @pytest.mark.skipif('platform.system() == "Windows"')
+    def test_run(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = build_clib(dist)
+
+        foo_c = os.path.join(pkg_dir, 'foo.c')
+        self.write_file(foo_c, 'int main(void) { return 1;}\n')
+        cmd.libraries = [('foo', {'sources': [foo_c]})]
+
+        build_temp = os.path.join(pkg_dir, 'build')
+        os.mkdir(build_temp)
+        cmd.build_temp = build_temp
+        cmd.build_clib = build_temp
+
+        # Before we run the command, we want to make sure
+        # all commands are present on the system.
+        ccmd = missing_compiler_executable()
+        if ccmd is not None:
+            self.skipTest(f'The {ccmd!r} command is not found')
+
+        # this should work
+        cmd.run()
+
+        # let's check the result
+        assert 'libfoo.a' in os.listdir(build_temp)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_ext.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_ext.py
new file mode 100644
index 00000000..beeba485
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_ext.py
@@ -0,0 +1,560 @@
+import contextlib
+import importlib
+import os
+import platform
+import re
+import shutil
+import site
+import sys
+import tempfile
+import textwrap
+from distutils import sysconfig
+from distutils.command.build_ext import build_ext
+from distutils.core import Distribution
+from distutils.errors import (
+    CompileError,
+    DistutilsPlatformError,
+    DistutilsSetupError,
+    UnknownFileError,
+)
+from distutils.extension import Extension
+from distutils.tests import missing_compiler_executable
+from distutils.tests.support import TempdirManager, copy_xxmodule_c, fixup_build_ext
+from io import StringIO
+
+import jaraco.path
+import path
+import pytest
+from test import support
+
+from .compat import py39 as import_helper
+
+
+@pytest.fixture()
+def user_site_dir(request):
+    self = request.instance
+    self.tmp_dir = self.mkdtemp()
+    self.tmp_path = path.Path(self.tmp_dir)
+    from distutils.command import build_ext
+
+    orig_user_base = site.USER_BASE
+
+    site.USER_BASE = self.mkdtemp()
+    build_ext.USER_BASE = site.USER_BASE
+
+    # bpo-30132: On Windows, a .pdb file may be created in the current
+    # working directory. Create a temporary working directory to cleanup
+    # everything at the end of the test.
+    with self.tmp_path:
+        yield
+
+    site.USER_BASE = orig_user_base
+    build_ext.USER_BASE = orig_user_base
+
+
+@contextlib.contextmanager
+def safe_extension_import(name, path):
+    with import_helper.CleanImport(name):
+        with extension_redirect(name, path) as new_path:
+            with import_helper.DirsOnSysPath(new_path):
+                yield
+
+
+@contextlib.contextmanager
+def extension_redirect(mod, path):
+    """
+    Tests will fail to tear down an extension module if it's been imported.
+
+    Before importing, copy the file to a temporary directory that won't
+    be cleaned up. Yield the new path.
+    """
+    if platform.system() != "Windows" and sys.platform != "cygwin":
+        yield path
+        return
+    with import_helper.DirsOnSysPath(path):
+        spec = importlib.util.find_spec(mod)
+    filename = os.path.basename(spec.origin)
+    trash_dir = tempfile.mkdtemp(prefix='deleteme')
+    dest = os.path.join(trash_dir, os.path.basename(filename))
+    shutil.copy(spec.origin, dest)
+    yield trash_dir
+    # TODO: can the file be scheduled for deletion?
+
+
+@pytest.mark.usefixtures('user_site_dir')
+class TestBuildExt(TempdirManager):
+    def build_ext(self, *args, **kwargs):
+        return build_ext(*args, **kwargs)
+
+    def test_build_ext(self):
+        missing_compiler_executable()
+        copy_xxmodule_c(self.tmp_dir)
+        xx_c = os.path.join(self.tmp_dir, 'xxmodule.c')
+        xx_ext = Extension('xx', [xx_c])
+        dist = Distribution({'name': 'xx', 'ext_modules': [xx_ext]})
+        dist.package_dir = self.tmp_dir
+        cmd = self.build_ext(dist)
+        fixup_build_ext(cmd)
+        cmd.build_lib = self.tmp_dir
+        cmd.build_temp = self.tmp_dir
+
+        old_stdout = sys.stdout
+        if not support.verbose:
+            # silence compiler output
+            sys.stdout = StringIO()
+        try:
+            cmd.ensure_finalized()
+            cmd.run()
+        finally:
+            sys.stdout = old_stdout
+
+        with safe_extension_import('xx', self.tmp_dir):
+            self._test_xx()
+
+    @staticmethod
+    def _test_xx():
+        import xx
+
+        for attr in ('error', 'foo', 'new', 'roj'):
+            assert hasattr(xx, attr)
+
+        assert xx.foo(2, 5) == 7
+        assert xx.foo(13, 15) == 28
+        assert xx.new().demo() is None
+        if support.HAVE_DOCSTRINGS:
+            doc = 'This is a template module just for instruction.'
+            assert xx.__doc__ == doc
+        assert isinstance(xx.Null(), xx.Null)
+        assert isinstance(xx.Str(), xx.Str)
+
+    def test_solaris_enable_shared(self):
+        dist = Distribution({'name': 'xx'})
+        cmd = self.build_ext(dist)
+        old = sys.platform
+
+        sys.platform = 'sunos'  # fooling finalize_options
+        from distutils.sysconfig import _config_vars
+
+        old_var = _config_vars.get('Py_ENABLE_SHARED')
+        _config_vars['Py_ENABLE_SHARED'] = True
+        try:
+            cmd.ensure_finalized()
+        finally:
+            sys.platform = old
+            if old_var is None:
+                del _config_vars['Py_ENABLE_SHARED']
+            else:
+                _config_vars['Py_ENABLE_SHARED'] = old_var
+
+        # make sure we get some library dirs under solaris
+        assert len(cmd.library_dirs) > 0
+
+    def test_user_site(self):
+        import site
+
+        dist = Distribution({'name': 'xx'})
+        cmd = self.build_ext(dist)
+
+        # making sure the user option is there
+        options = [name for name, short, label in cmd.user_options]
+        assert 'user' in options
+
+        # setting a value
+        cmd.user = True
+
+        # setting user based lib and include
+        lib = os.path.join(site.USER_BASE, 'lib')
+        incl = os.path.join(site.USER_BASE, 'include')
+        os.mkdir(lib)
+        os.mkdir(incl)
+
+        # let's run finalize
+        cmd.ensure_finalized()
+
+        # see if include_dirs and library_dirs
+        # were set
+        assert lib in cmd.library_dirs
+        assert lib in cmd.rpath
+        assert incl in cmd.include_dirs
+
+    def test_optional_extension(self):
+        # this extension will fail, but let's ignore this failure
+        # with the optional argument.
+        modules = [Extension('foo', ['xxx'], optional=False)]
+        dist = Distribution({'name': 'xx', 'ext_modules': modules})
+        cmd = self.build_ext(dist)
+        cmd.ensure_finalized()
+        with pytest.raises((UnknownFileError, CompileError)):
+            cmd.run()  # should raise an error
+
+        modules = [Extension('foo', ['xxx'], optional=True)]
+        dist = Distribution({'name': 'xx', 'ext_modules': modules})
+        cmd = self.build_ext(dist)
+        cmd.ensure_finalized()
+        cmd.run()  # should pass
+
+    def test_finalize_options(self):
+        # Make sure Python's include directories (for Python.h, pyconfig.h,
+        # etc.) are in the include search path.
+        modules = [Extension('foo', ['xxx'], optional=False)]
+        dist = Distribution({'name': 'xx', 'ext_modules': modules})
+        cmd = self.build_ext(dist)
+        cmd.finalize_options()
+
+        py_include = sysconfig.get_python_inc()
+        for p in py_include.split(os.path.pathsep):
+            assert p in cmd.include_dirs
+
+        plat_py_include = sysconfig.get_python_inc(plat_specific=True)
+        for p in plat_py_include.split(os.path.pathsep):
+            assert p in cmd.include_dirs
+
+        # make sure cmd.libraries is turned into a list
+        # if it's a string
+        cmd = self.build_ext(dist)
+        cmd.libraries = 'my_lib, other_lib lastlib'
+        cmd.finalize_options()
+        assert cmd.libraries == ['my_lib', 'other_lib', 'lastlib']
+
+        # make sure cmd.library_dirs is turned into a list
+        # if it's a string
+        cmd = self.build_ext(dist)
+        cmd.library_dirs = f'my_lib_dir{os.pathsep}other_lib_dir'
+        cmd.finalize_options()
+        assert 'my_lib_dir' in cmd.library_dirs
+        assert 'other_lib_dir' in cmd.library_dirs
+
+        # make sure rpath is turned into a list
+        # if it's a string
+        cmd = self.build_ext(dist)
+        cmd.rpath = f'one{os.pathsep}two'
+        cmd.finalize_options()
+        assert cmd.rpath == ['one', 'two']
+
+        # make sure cmd.link_objects is turned into a list
+        # if it's a string
+        cmd = build_ext(dist)
+        cmd.link_objects = 'one two,three'
+        cmd.finalize_options()
+        assert cmd.link_objects == ['one', 'two', 'three']
+
+        # XXX more tests to perform for win32
+
+        # make sure define is turned into 2-tuples
+        # strings if they are ','-separated strings
+        cmd = self.build_ext(dist)
+        cmd.define = 'one,two'
+        cmd.finalize_options()
+        assert cmd.define == [('one', '1'), ('two', '1')]
+
+        # make sure undef is turned into a list of
+        # strings if they are ','-separated strings
+        cmd = self.build_ext(dist)
+        cmd.undef = 'one,two'
+        cmd.finalize_options()
+        assert cmd.undef == ['one', 'two']
+
+        # make sure swig_opts is turned into a list
+        cmd = self.build_ext(dist)
+        cmd.swig_opts = None
+        cmd.finalize_options()
+        assert cmd.swig_opts == []
+
+        cmd = self.build_ext(dist)
+        cmd.swig_opts = '1 2'
+        cmd.finalize_options()
+        assert cmd.swig_opts == ['1', '2']
+
+    def test_check_extensions_list(self):
+        dist = Distribution()
+        cmd = self.build_ext(dist)
+        cmd.finalize_options()
+
+        # 'extensions' option must be a list of Extension instances
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_extensions_list('foo')
+
+        # each element of 'ext_modules' option must be an
+        # Extension instance or 2-tuple
+        exts = [('bar', 'foo', 'bar'), 'foo']
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_extensions_list(exts)
+
+        # first element of each tuple in 'ext_modules'
+        # must be the extension name (a string) and match
+        # a python dotted-separated name
+        exts = [('foo-bar', '')]
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_extensions_list(exts)
+
+        # second element of each tuple in 'ext_modules'
+        # must be a dictionary (build info)
+        exts = [('foo.bar', '')]
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_extensions_list(exts)
+
+        # ok this one should pass
+        exts = [('foo.bar', {'sources': [''], 'libraries': 'foo', 'some': 'bar'})]
+        cmd.check_extensions_list(exts)
+        ext = exts[0]
+        assert isinstance(ext, Extension)
+
+        # check_extensions_list adds in ext the values passed
+        # when they are in ('include_dirs', 'library_dirs', 'libraries'
+        # 'extra_objects', 'extra_compile_args', 'extra_link_args')
+        assert ext.libraries == 'foo'
+        assert not hasattr(ext, 'some')
+
+        # 'macros' element of build info dict must be 1- or 2-tuple
+        exts = [
+            (
+                'foo.bar',
+                {
+                    'sources': [''],
+                    'libraries': 'foo',
+                    'some': 'bar',
+                    'macros': [('1', '2', '3'), 'foo'],
+                },
+            )
+        ]
+        with pytest.raises(DistutilsSetupError):
+            cmd.check_extensions_list(exts)
+
+        exts[0][1]['macros'] = [('1', '2'), ('3',)]
+        cmd.check_extensions_list(exts)
+        assert exts[0].undef_macros == ['3']
+        assert exts[0].define_macros == [('1', '2')]
+
+    def test_get_source_files(self):
+        modules = [Extension('foo', ['xxx'], optional=False)]
+        dist = Distribution({'name': 'xx', 'ext_modules': modules})
+        cmd = self.build_ext(dist)
+        cmd.ensure_finalized()
+        assert cmd.get_source_files() == ['xxx']
+
+    def test_unicode_module_names(self):
+        modules = [
+            Extension('foo', ['aaa'], optional=False),
+            Extension('föö', ['uuu'], optional=False),
+        ]
+        dist = Distribution({'name': 'xx', 'ext_modules': modules})
+        cmd = self.build_ext(dist)
+        cmd.ensure_finalized()
+        assert re.search(r'foo(_d)?\..*', cmd.get_ext_filename(modules[0].name))
+        assert re.search(r'föö(_d)?\..*', cmd.get_ext_filename(modules[1].name))
+        assert cmd.get_export_symbols(modules[0]) == ['PyInit_foo']
+        assert cmd.get_export_symbols(modules[1]) == ['PyInitU_f_1gaa']
+
+    def test_compiler_option(self):
+        # cmd.compiler is an option and
+        # should not be overridden by a compiler instance
+        # when the command is run
+        dist = Distribution()
+        cmd = self.build_ext(dist)
+        cmd.compiler = 'unix'
+        cmd.ensure_finalized()
+        cmd.run()
+        assert cmd.compiler == 'unix'
+
+    def test_get_outputs(self):
+        missing_compiler_executable()
+        tmp_dir = self.mkdtemp()
+        c_file = os.path.join(tmp_dir, 'foo.c')
+        self.write_file(c_file, 'void PyInit_foo(void) {}\n')
+        ext = Extension('foo', [c_file], optional=False)
+        dist = Distribution({'name': 'xx', 'ext_modules': [ext]})
+        cmd = self.build_ext(dist)
+        fixup_build_ext(cmd)
+        cmd.ensure_finalized()
+        assert len(cmd.get_outputs()) == 1
+
+        cmd.build_lib = os.path.join(self.tmp_dir, 'build')
+        cmd.build_temp = os.path.join(self.tmp_dir, 'tempt')
+
+        # issue #5977 : distutils build_ext.get_outputs
+        # returns wrong result with --inplace
+        other_tmp_dir = os.path.realpath(self.mkdtemp())
+        old_wd = os.getcwd()
+        os.chdir(other_tmp_dir)
+        try:
+            cmd.inplace = True
+            cmd.run()
+            so_file = cmd.get_outputs()[0]
+        finally:
+            os.chdir(old_wd)
+        assert os.path.exists(so_file)
+        ext_suffix = sysconfig.get_config_var('EXT_SUFFIX')
+        assert so_file.endswith(ext_suffix)
+        so_dir = os.path.dirname(so_file)
+        assert so_dir == other_tmp_dir
+
+        cmd.inplace = False
+        cmd.compiler = None
+        cmd.run()
+        so_file = cmd.get_outputs()[0]
+        assert os.path.exists(so_file)
+        assert so_file.endswith(ext_suffix)
+        so_dir = os.path.dirname(so_file)
+        assert so_dir == cmd.build_lib
+
+        # inplace = False, cmd.package = 'bar'
+        build_py = cmd.get_finalized_command('build_py')
+        build_py.package_dir = {'': 'bar'}
+        path = cmd.get_ext_fullpath('foo')
+        # checking that the last directory is the build_dir
+        path = os.path.split(path)[0]
+        assert path == cmd.build_lib
+
+        # inplace = True, cmd.package = 'bar'
+        cmd.inplace = True
+        other_tmp_dir = os.path.realpath(self.mkdtemp())
+        old_wd = os.getcwd()
+        os.chdir(other_tmp_dir)
+        try:
+            path = cmd.get_ext_fullpath('foo')
+        finally:
+            os.chdir(old_wd)
+        # checking that the last directory is bar
+        path = os.path.split(path)[0]
+        lastdir = os.path.split(path)[-1]
+        assert lastdir == 'bar'
+
+    def test_ext_fullpath(self):
+        ext = sysconfig.get_config_var('EXT_SUFFIX')
+        # building lxml.etree inplace
+        # etree_c = os.path.join(self.tmp_dir, 'lxml.etree.c')
+        # etree_ext = Extension('lxml.etree', [etree_c])
+        # dist = Distribution({'name': 'lxml', 'ext_modules': [etree_ext]})
+        dist = Distribution()
+        cmd = self.build_ext(dist)
+        cmd.inplace = True
+        cmd.distribution.package_dir = {'': 'src'}
+        cmd.distribution.packages = ['lxml', 'lxml.html']
+        curdir = os.getcwd()
+        wanted = os.path.join(curdir, 'src', 'lxml', 'etree' + ext)
+        path = cmd.get_ext_fullpath('lxml.etree')
+        assert wanted == path
+
+        # building lxml.etree not inplace
+        cmd.inplace = False
+        cmd.build_lib = os.path.join(curdir, 'tmpdir')
+        wanted = os.path.join(curdir, 'tmpdir', 'lxml', 'etree' + ext)
+        path = cmd.get_ext_fullpath('lxml.etree')
+        assert wanted == path
+
+        # building twisted.runner.portmap not inplace
+        build_py = cmd.get_finalized_command('build_py')
+        build_py.package_dir = {}
+        cmd.distribution.packages = ['twisted', 'twisted.runner.portmap']
+        path = cmd.get_ext_fullpath('twisted.runner.portmap')
+        wanted = os.path.join(curdir, 'tmpdir', 'twisted', 'runner', 'portmap' + ext)
+        assert wanted == path
+
+        # building twisted.runner.portmap inplace
+        cmd.inplace = True
+        path = cmd.get_ext_fullpath('twisted.runner.portmap')
+        wanted = os.path.join(curdir, 'twisted', 'runner', 'portmap' + ext)
+        assert wanted == path
+
+    @pytest.mark.skipif('platform.system() != "Darwin"')
+    @pytest.mark.usefixtures('save_env')
+    def test_deployment_target_default(self):
+        # Issue 9516: Test that, in the absence of the environment variable,
+        # an extension module is compiled with the same deployment target as
+        #  the interpreter.
+        self._try_compile_deployment_target('==', None)
+
+    @pytest.mark.skipif('platform.system() != "Darwin"')
+    @pytest.mark.usefixtures('save_env')
+    def test_deployment_target_too_low(self):
+        # Issue 9516: Test that an extension module is not allowed to be
+        # compiled with a deployment target less than that of the interpreter.
+        with pytest.raises(DistutilsPlatformError):
+            self._try_compile_deployment_target('>', '10.1')
+
+    @pytest.mark.skipif('platform.system() != "Darwin"')
+    @pytest.mark.usefixtures('save_env')
+    def test_deployment_target_higher_ok(self):  # pragma: no cover
+        # Issue 9516: Test that an extension module can be compiled with a
+        # deployment target higher than that of the interpreter: the ext
+        # module may depend on some newer OS feature.
+        deptarget = sysconfig.get_config_var('MACOSX_DEPLOYMENT_TARGET')
+        if deptarget:
+            # increment the minor version number (i.e. 10.6 -> 10.7)
+            deptarget = [int(x) for x in deptarget.split('.')]
+            deptarget[-1] += 1
+            deptarget = '.'.join(str(i) for i in deptarget)
+            self._try_compile_deployment_target('<', deptarget)
+
+    def _try_compile_deployment_target(self, operator, target):  # pragma: no cover
+        if target is None:
+            if os.environ.get('MACOSX_DEPLOYMENT_TARGET'):
+                del os.environ['MACOSX_DEPLOYMENT_TARGET']
+        else:
+            os.environ['MACOSX_DEPLOYMENT_TARGET'] = target
+
+        jaraco.path.build(
+            {
+                'deptargetmodule.c': textwrap.dedent(f"""\
+                    #include <AvailabilityMacros.h>
+
+                    int dummy;
+
+                    #if TARGET {operator} MAC_OS_X_VERSION_MIN_REQUIRED
+                    #else
+                    #error "Unexpected target"
+                    #endif
+
+                    """),
+            },
+            self.tmp_path,
+        )
+
+        # get the deployment target that the interpreter was built with
+        target = sysconfig.get_config_var('MACOSX_DEPLOYMENT_TARGET')
+        target = tuple(map(int, target.split('.')[0:2]))
+        # format the target value as defined in the Apple
+        # Availability Macros.  We can't use the macro names since
+        # at least one value we test with will not exist yet.
+        if target[:2] < (10, 10):
+            # for 10.1 through 10.9.x -> "10n0"
+            tmpl = '{:02}{:01}0'
+        else:
+            # for 10.10 and beyond -> "10nn00"
+            if len(target) >= 2:
+                tmpl = '{:02}{:02}00'
+            else:
+                # 11 and later can have no minor version (11 instead of 11.0)
+                tmpl = '{:02}0000'
+        target = tmpl.format(*target)
+        deptarget_ext = Extension(
+            'deptarget',
+            [self.tmp_path / 'deptargetmodule.c'],
+            extra_compile_args=[f'-DTARGET={target}'],
+        )
+        dist = Distribution({'name': 'deptarget', 'ext_modules': [deptarget_ext]})
+        dist.package_dir = self.tmp_dir
+        cmd = self.build_ext(dist)
+        cmd.build_lib = self.tmp_dir
+        cmd.build_temp = self.tmp_dir
+
+        try:
+            old_stdout = sys.stdout
+            if not support.verbose:
+                # silence compiler output
+                sys.stdout = StringIO()
+            try:
+                cmd.ensure_finalized()
+                cmd.run()
+            finally:
+                sys.stdout = old_stdout
+
+        except CompileError:
+            self.fail("Wrong deployment target during compilation")
+
+
+class TestParallelBuildExt(TestBuildExt):
+    def build_ext(self, *args, **kwargs):
+        build_ext = super().build_ext(*args, **kwargs)
+        build_ext.parallel = True
+        return build_ext
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_py.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_py.py
new file mode 100644
index 00000000..b316ed43
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_py.py
@@ -0,0 +1,196 @@
+"""Tests for distutils.command.build_py."""
+
+import os
+import sys
+from distutils.command.build_py import build_py
+from distutils.core import Distribution
+from distutils.errors import DistutilsFileError
+from distutils.tests import support
+
+import jaraco.path
+import pytest
+
+
+@support.combine_markers
+class TestBuildPy(support.TempdirManager):
+    def test_package_data(self):
+        sources = self.mkdtemp()
+        jaraco.path.build(
+            {
+                '__init__.py': "# Pretend this is a package.",
+                'README.txt': 'Info about this package',
+            },
+            sources,
+        )
+
+        destination = self.mkdtemp()
+
+        dist = Distribution({"packages": ["pkg"], "package_dir": {"pkg": sources}})
+        # script_name need not exist, it just need to be initialized
+        dist.script_name = os.path.join(sources, "setup.py")
+        dist.command_obj["build"] = support.DummyCommand(
+            force=False, build_lib=destination
+        )
+        dist.packages = ["pkg"]
+        dist.package_data = {"pkg": ["README.txt"]}
+        dist.package_dir = {"pkg": sources}
+
+        cmd = build_py(dist)
+        cmd.compile = True
+        cmd.ensure_finalized()
+        assert cmd.package_data == dist.package_data
+
+        cmd.run()
+
+        # This makes sure the list of outputs includes byte-compiled
+        # files for Python modules but not for package data files
+        # (there shouldn't *be* byte-code files for those!).
+        assert len(cmd.get_outputs()) == 3
+        pkgdest = os.path.join(destination, "pkg")
+        files = os.listdir(pkgdest)
+        pycache_dir = os.path.join(pkgdest, "__pycache__")
+        assert "__init__.py" in files
+        assert "README.txt" in files
+        if sys.dont_write_bytecode:
+            assert not os.path.exists(pycache_dir)
+        else:
+            pyc_files = os.listdir(pycache_dir)
+            assert f"__init__.{sys.implementation.cache_tag}.pyc" in pyc_files
+
+    def test_empty_package_dir(self):
+        # See bugs #1668596/#1720897
+        sources = self.mkdtemp()
+        jaraco.path.build({'__init__.py': '', 'doc': {'testfile': ''}}, sources)
+
+        os.chdir(sources)
+        dist = Distribution({
+            "packages": ["pkg"],
+            "package_dir": {"pkg": ""},
+            "package_data": {"pkg": ["doc/*"]},
+        })
+        # script_name need not exist, it just need to be initialized
+        dist.script_name = os.path.join(sources, "setup.py")
+        dist.script_args = ["build"]
+        dist.parse_command_line()
+
+        try:
+            dist.run_commands()
+        except DistutilsFileError:
+            self.fail("failed package_data test when package_dir is ''")
+
+    @pytest.mark.skipif('sys.dont_write_bytecode')
+    def test_byte_compile(self):
+        project_dir, dist = self.create_dist(py_modules=['boiledeggs'])
+        os.chdir(project_dir)
+        self.write_file('boiledeggs.py', 'import antigravity')
+        cmd = build_py(dist)
+        cmd.compile = True
+        cmd.build_lib = 'here'
+        cmd.finalize_options()
+        cmd.run()
+
+        found = os.listdir(cmd.build_lib)
+        assert sorted(found) == ['__pycache__', 'boiledeggs.py']
+        found = os.listdir(os.path.join(cmd.build_lib, '__pycache__'))
+        assert found == [f'boiledeggs.{sys.implementation.cache_tag}.pyc']
+
+    @pytest.mark.skipif('sys.dont_write_bytecode')
+    def test_byte_compile_optimized(self):
+        project_dir, dist = self.create_dist(py_modules=['boiledeggs'])
+        os.chdir(project_dir)
+        self.write_file('boiledeggs.py', 'import antigravity')
+        cmd = build_py(dist)
+        cmd.compile = False
+        cmd.optimize = 1
+        cmd.build_lib = 'here'
+        cmd.finalize_options()
+        cmd.run()
+
+        found = os.listdir(cmd.build_lib)
+        assert sorted(found) == ['__pycache__', 'boiledeggs.py']
+        found = os.listdir(os.path.join(cmd.build_lib, '__pycache__'))
+        expect = f'boiledeggs.{sys.implementation.cache_tag}.opt-1.pyc'
+        assert sorted(found) == [expect]
+
+    def test_dir_in_package_data(self):
+        """
+        A directory in package_data should not be added to the filelist.
+        """
+        # See bug 19286
+        sources = self.mkdtemp()
+        jaraco.path.build(
+            {
+                'pkg': {
+                    '__init__.py': '',
+                    'doc': {
+                        'testfile': '',
+                        # create a directory that could be incorrectly detected as a file
+                        'otherdir': {},
+                    },
+                }
+            },
+            sources,
+        )
+
+        os.chdir(sources)
+        dist = Distribution({"packages": ["pkg"], "package_data": {"pkg": ["doc/*"]}})
+        # script_name need not exist, it just need to be initialized
+        dist.script_name = os.path.join(sources, "setup.py")
+        dist.script_args = ["build"]
+        dist.parse_command_line()
+
+        try:
+            dist.run_commands()
+        except DistutilsFileError:
+            self.fail("failed package_data when data dir includes a dir")
+
+    def test_dont_write_bytecode(self, caplog):
+        # makes sure byte_compile is not used
+        dist = self.create_dist()[1]
+        cmd = build_py(dist)
+        cmd.compile = True
+        cmd.optimize = 1
+
+        old_dont_write_bytecode = sys.dont_write_bytecode
+        sys.dont_write_bytecode = True
+        try:
+            cmd.byte_compile([])
+        finally:
+            sys.dont_write_bytecode = old_dont_write_bytecode
+
+        assert 'byte-compiling is disabled' in caplog.records[0].message
+
+    def test_namespace_package_does_not_warn(self, caplog):
+        """
+        Originally distutils implementation did not account for PEP 420
+        and included warns for package directories that did not contain
+        ``__init__.py`` files.
+        After the acceptance of PEP 420, these warnings don't make more sense
+        so we want to ensure there are not displayed to not confuse the users.
+        """
+        # Create a fake project structure with a package namespace:
+        tmp = self.mkdtemp()
+        jaraco.path.build({'ns': {'pkg': {'module.py': ''}}}, tmp)
+        os.chdir(tmp)
+
+        # Configure the package:
+        attrs = {
+            "name": "ns.pkg",
+            "packages": ["ns", "ns.pkg"],
+            "script_name": "setup.py",
+        }
+        dist = Distribution(attrs)
+
+        # Run code paths that would trigger the trap:
+        cmd = dist.get_command_obj("build_py")
+        cmd.finalize_options()
+        modules = cmd.find_all_modules()
+        assert len(modules) == 1
+        module_path = modules[0][-1]
+        assert module_path.replace(os.sep, "/") == "ns/pkg/module.py"
+
+        cmd.run()
+
+        assert not any(
+            "package init file" in msg and "not found" in msg for msg in caplog.messages
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_scripts.py
new file mode 100644
index 00000000..3582f691
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_build_scripts.py
@@ -0,0 +1,96 @@
+"""Tests for distutils.command.build_scripts."""
+
+import os
+import textwrap
+from distutils import sysconfig
+from distutils.command.build_scripts import build_scripts
+from distutils.core import Distribution
+from distutils.tests import support
+
+import jaraco.path
+
+
+class TestBuildScripts(support.TempdirManager):
+    def test_default_settings(self):
+        cmd = self.get_build_scripts_cmd("/foo/bar", [])
+        assert not cmd.force
+        assert cmd.build_dir is None
+
+        cmd.finalize_options()
+
+        assert cmd.force
+        assert cmd.build_dir == "/foo/bar"
+
+    def test_build(self):
+        source = self.mkdtemp()
+        target = self.mkdtemp()
+        expected = self.write_sample_scripts(source)
+
+        cmd = self.get_build_scripts_cmd(
+            target, [os.path.join(source, fn) for fn in expected]
+        )
+        cmd.finalize_options()
+        cmd.run()
+
+        built = os.listdir(target)
+        for name in expected:
+            assert name in built
+
+    def get_build_scripts_cmd(self, target, scripts):
+        import sys
+
+        dist = Distribution()
+        dist.scripts = scripts
+        dist.command_obj["build"] = support.DummyCommand(
+            build_scripts=target, force=True, executable=sys.executable
+        )
+        return build_scripts(dist)
+
+    @staticmethod
+    def write_sample_scripts(dir):
+        spec = {
+            'script1.py': textwrap.dedent("""
+                #! /usr/bin/env python2.3
+                # bogus script w/ Python sh-bang
+                pass
+                """).lstrip(),
+            'script2.py': textwrap.dedent("""
+                #!/usr/bin/python
+                # bogus script w/ Python sh-bang
+                pass
+                """).lstrip(),
+            'shell.sh': textwrap.dedent("""
+                #!/bin/sh
+                # bogus shell script w/ sh-bang
+                exit 0
+                """).lstrip(),
+        }
+        jaraco.path.build(spec, dir)
+        return list(spec)
+
+    def test_version_int(self):
+        source = self.mkdtemp()
+        target = self.mkdtemp()
+        expected = self.write_sample_scripts(source)
+
+        cmd = self.get_build_scripts_cmd(
+            target, [os.path.join(source, fn) for fn in expected]
+        )
+        cmd.finalize_options()
+
+        # https://bugs.python.org/issue4524
+        #
+        # On linux-g++-32 with command line `./configure --enable-ipv6
+        # --with-suffix=3`, python is compiled okay but the build scripts
+        # failed when writing the name of the executable
+        old = sysconfig.get_config_vars().get('VERSION')
+        sysconfig._config_vars['VERSION'] = 4
+        try:
+            cmd.run()
+        finally:
+            if old is not None:
+                sysconfig._config_vars['VERSION'] = old
+
+        built = os.listdir(target)
+        for name in expected:
+            assert name in built
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_ccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_ccompiler.py
new file mode 100644
index 00000000..7ebfed56
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_ccompiler.py
@@ -0,0 +1,93 @@
+import os
+import platform
+import sys
+import sysconfig
+import textwrap
+from distutils import ccompiler
+
+import pytest
+
+pytestmark = pytest.mark.usefixtures('suppress_path_mangle')
+
+
+def _make_strs(paths):
+    """
+    Convert paths to strings for legacy compatibility.
+    """
+    if sys.version_info > (3, 8) and platform.system() != "Windows":
+        return paths
+    return list(map(os.fspath, paths))
+
+
+@pytest.fixture
+def c_file(tmp_path):
+    c_file = tmp_path / 'foo.c'
+    gen_headers = ('Python.h',)
+    is_windows = platform.system() == "Windows"
+    plat_headers = ('windows.h',) * is_windows
+    all_headers = gen_headers + plat_headers
+    headers = '\n'.join(f'#include <{header}>\n' for header in all_headers)
+    payload = (
+        textwrap.dedent(
+            """
+        #headers
+        void PyInit_foo(void) {}
+        """
+        )
+        .lstrip()
+        .replace('#headers', headers)
+    )
+    c_file.write_text(payload, encoding='utf-8')
+    return c_file
+
+
+def test_set_include_dirs(c_file):
+    """
+    Extensions should build even if set_include_dirs is invoked.
+    In particular, compiler-specific paths should not be overridden.
+    """
+    compiler = ccompiler.new_compiler()
+    python = sysconfig.get_paths()['include']
+    compiler.set_include_dirs([python])
+    compiler.compile(_make_strs([c_file]))
+
+    # do it again, setting include dirs after any initialization
+    compiler.set_include_dirs([python])
+    compiler.compile(_make_strs([c_file]))
+
+
+def test_has_function_prototype():
+    # Issue https://github.com/pypa/setuptools/issues/3648
+    # Test prototype-generating behavior.
+
+    compiler = ccompiler.new_compiler()
+
+    # Every C implementation should have these.
+    assert compiler.has_function('abort')
+    assert compiler.has_function('exit')
+    with pytest.deprecated_call(match='includes is deprecated'):
+        # abort() is a valid expression with the <stdlib.h> prototype.
+        assert compiler.has_function('abort', includes=['stdlib.h'])
+    with pytest.deprecated_call(match='includes is deprecated'):
+        # But exit() is not valid with the actual prototype in scope.
+        assert not compiler.has_function('exit', includes=['stdlib.h'])
+    # And setuptools_does_not_exist is not declared or defined at all.
+    assert not compiler.has_function('setuptools_does_not_exist')
+    with pytest.deprecated_call(match='includes is deprecated'):
+        assert not compiler.has_function(
+            'setuptools_does_not_exist', includes=['stdio.h']
+        )
+
+
+def test_include_dirs_after_multiple_compile_calls(c_file):
+    """
+    Calling compile multiple times should not change the include dirs
+    (regression test for setuptools issue #3591).
+    """
+    compiler = ccompiler.new_compiler()
+    python = sysconfig.get_paths()['include']
+    compiler.set_include_dirs([python])
+    compiler.compile(_make_strs([c_file]))
+    assert compiler.include_dirs == [python]
+    compiler.compile(_make_strs([c_file]))
+    assert compiler.include_dirs == [python]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_check.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_check.py
new file mode 100644
index 00000000..b672b1f9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_check.py
@@ -0,0 +1,194 @@
+"""Tests for distutils.command.check."""
+
+import os
+import textwrap
+from distutils.command.check import check
+from distutils.errors import DistutilsSetupError
+from distutils.tests import support
+
+import pytest
+
+try:
+    import pygments
+except ImportError:
+    pygments = None
+
+
+HERE = os.path.dirname(__file__)
+
+
+@support.combine_markers
+class TestCheck(support.TempdirManager):
+    def _run(self, metadata=None, cwd=None, **options):
+        if metadata is None:
+            metadata = {}
+        if cwd is not None:
+            old_dir = os.getcwd()
+            os.chdir(cwd)
+        pkg_info, dist = self.create_dist(**metadata)
+        cmd = check(dist)
+        cmd.initialize_options()
+        for name, value in options.items():
+            setattr(cmd, name, value)
+        cmd.ensure_finalized()
+        cmd.run()
+        if cwd is not None:
+            os.chdir(old_dir)
+        return cmd
+
+    def test_check_metadata(self):
+        # let's run the command with no metadata at all
+        # by default, check is checking the metadata
+        # should have some warnings
+        cmd = self._run()
+        assert cmd._warnings == 1
+
+        # now let's add the required fields
+        # and run it again, to make sure we don't get
+        # any warning anymore
+        metadata = {
+            'url': 'xxx',
+            'author': 'xxx',
+            'author_email': 'xxx',
+            'name': 'xxx',
+            'version': 'xxx',
+        }
+        cmd = self._run(metadata)
+        assert cmd._warnings == 0
+
+        # now with the strict mode, we should
+        # get an error if there are missing metadata
+        with pytest.raises(DistutilsSetupError):
+            self._run({}, **{'strict': 1})
+
+        # and of course, no error when all metadata are present
+        cmd = self._run(metadata, strict=True)
+        assert cmd._warnings == 0
+
+        # now a test with non-ASCII characters
+        metadata = {
+            'url': 'xxx',
+            'author': '\u00c9ric',
+            'author_email': 'xxx',
+            'name': 'xxx',
+            'version': 'xxx',
+            'description': 'Something about esszet \u00df',
+            'long_description': 'More things about esszet \u00df',
+        }
+        cmd = self._run(metadata)
+        assert cmd._warnings == 0
+
+    def test_check_author_maintainer(self):
+        for kind in ("author", "maintainer"):
+            # ensure no warning when author_email or maintainer_email is given
+            # (the spec allows these fields to take the form "Name <email>")
+            metadata = {
+                'url': 'xxx',
+                kind + '_email': 'Name <name@email.com>',
+                'name': 'xxx',
+                'version': 'xxx',
+            }
+            cmd = self._run(metadata)
+            assert cmd._warnings == 0
+
+            # the check should not warn if only email is given
+            metadata[kind + '_email'] = 'name@email.com'
+            cmd = self._run(metadata)
+            assert cmd._warnings == 0
+
+            # the check should not warn if only the name is given
+            metadata[kind] = "Name"
+            del metadata[kind + '_email']
+            cmd = self._run(metadata)
+            assert cmd._warnings == 0
+
+    def test_check_document(self):
+        pytest.importorskip('docutils')
+        pkg_info, dist = self.create_dist()
+        cmd = check(dist)
+
+        # let's see if it detects broken rest
+        broken_rest = 'title\n===\n\ntest'
+        msgs = cmd._check_rst_data(broken_rest)
+        assert len(msgs) == 1
+
+        # and non-broken rest
+        rest = 'title\n=====\n\ntest'
+        msgs = cmd._check_rst_data(rest)
+        assert len(msgs) == 0
+
+    def test_check_restructuredtext(self):
+        pytest.importorskip('docutils')
+        # let's see if it detects broken rest in long_description
+        broken_rest = 'title\n===\n\ntest'
+        pkg_info, dist = self.create_dist(long_description=broken_rest)
+        cmd = check(dist)
+        cmd.check_restructuredtext()
+        assert cmd._warnings == 1
+
+        # let's see if we have an error with strict=True
+        metadata = {
+            'url': 'xxx',
+            'author': 'xxx',
+            'author_email': 'xxx',
+            'name': 'xxx',
+            'version': 'xxx',
+            'long_description': broken_rest,
+        }
+        with pytest.raises(DistutilsSetupError):
+            self._run(metadata, **{'strict': 1, 'restructuredtext': 1})
+
+        # and non-broken rest, including a non-ASCII character to test #12114
+        metadata['long_description'] = 'title\n=====\n\ntest \u00df'
+        cmd = self._run(metadata, strict=True, restructuredtext=True)
+        assert cmd._warnings == 0
+
+        # check that includes work to test #31292
+        metadata['long_description'] = 'title\n=====\n\n.. include:: includetest.rst'
+        cmd = self._run(metadata, cwd=HERE, strict=True, restructuredtext=True)
+        assert cmd._warnings == 0
+
+    def test_check_restructuredtext_with_syntax_highlight(self):
+        pytest.importorskip('docutils')
+        # Don't fail if there is a `code` or `code-block` directive
+
+        example_rst_docs = [
+            textwrap.dedent(
+                """\
+            Here's some code:
+
+            .. code:: python
+
+                def foo():
+                    pass
+            """
+            ),
+            textwrap.dedent(
+                """\
+            Here's some code:
+
+            .. code-block:: python
+
+                def foo():
+                    pass
+            """
+            ),
+        ]
+
+        for rest_with_code in example_rst_docs:
+            pkg_info, dist = self.create_dist(long_description=rest_with_code)
+            cmd = check(dist)
+            cmd.check_restructuredtext()
+            msgs = cmd._check_rst_data(rest_with_code)
+            if pygments is not None:
+                assert len(msgs) == 0
+            else:
+                assert len(msgs) == 1
+                assert (
+                    str(msgs[0][1])
+                    == 'Cannot analyze code. Pygments package not found.'
+                )
+
+    def test_check_all(self):
+        with pytest.raises(DistutilsSetupError):
+            self._run({}, **{'strict': 1, 'restructuredtext': 1})
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_clean.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_clean.py
new file mode 100644
index 00000000..cc78f30f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_clean.py
@@ -0,0 +1,45 @@
+"""Tests for distutils.command.clean."""
+
+import os
+from distutils.command.clean import clean
+from distutils.tests import support
+
+
+class TestClean(support.TempdirManager):
+    def test_simple_run(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = clean(dist)
+
+        # let's add some elements clean should remove
+        dirs = [
+            (d, os.path.join(pkg_dir, d))
+            for d in (
+                'build_temp',
+                'build_lib',
+                'bdist_base',
+                'build_scripts',
+                'build_base',
+            )
+        ]
+
+        for name, path in dirs:
+            os.mkdir(path)
+            setattr(cmd, name, path)
+            if name == 'build_base':
+                continue
+            for f in ('one', 'two', 'three'):
+                self.write_file(os.path.join(path, f))
+
+        # let's run the command
+        cmd.all = 1
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # make sure the files where removed
+        for _name, path in dirs:
+            assert not os.path.exists(path), f'{path} was not removed'
+
+        # let's run the command again (should spit warnings but succeed)
+        cmd.all = 1
+        cmd.ensure_finalized()
+        cmd.run()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cmd.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cmd.py
new file mode 100644
index 00000000..76e8f598
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cmd.py
@@ -0,0 +1,107 @@
+"""Tests for distutils.cmd."""
+
+import os
+from distutils import debug
+from distutils.cmd import Command
+from distutils.dist import Distribution
+from distutils.errors import DistutilsOptionError
+
+import pytest
+
+
+class MyCmd(Command):
+    def initialize_options(self):
+        pass
+
+
+@pytest.fixture
+def cmd(request):
+    return MyCmd(Distribution())
+
+
+class TestCommand:
+    def test_ensure_string_list(self, cmd):
+        cmd.not_string_list = ['one', 2, 'three']
+        cmd.yes_string_list = ['one', 'two', 'three']
+        cmd.not_string_list2 = object()
+        cmd.yes_string_list2 = 'ok'
+        cmd.ensure_string_list('yes_string_list')
+        cmd.ensure_string_list('yes_string_list2')
+
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_string_list('not_string_list')
+
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_string_list('not_string_list2')
+
+        cmd.option1 = 'ok,dok'
+        cmd.ensure_string_list('option1')
+        assert cmd.option1 == ['ok', 'dok']
+
+        cmd.option2 = ['xxx', 'www']
+        cmd.ensure_string_list('option2')
+
+        cmd.option3 = ['ok', 2]
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_string_list('option3')
+
+    def test_make_file(self, cmd):
+        # making sure it raises when infiles is not a string or a list/tuple
+        with pytest.raises(TypeError):
+            cmd.make_file(infiles=True, outfile='', func='func', args=())
+
+        # making sure execute gets called properly
+        def _execute(func, args, exec_msg, level):
+            assert exec_msg == 'generating out from in'
+
+        cmd.force = True
+        cmd.execute = _execute
+        cmd.make_file(infiles='in', outfile='out', func='func', args=())
+
+    def test_dump_options(self, cmd):
+        msgs = []
+
+        def _announce(msg, level):
+            msgs.append(msg)
+
+        cmd.announce = _announce
+        cmd.option1 = 1
+        cmd.option2 = 1
+        cmd.user_options = [('option1', '', ''), ('option2', '', '')]
+        cmd.dump_options()
+
+        wanted = ["command options for 'MyCmd':", '  option1 = 1', '  option2 = 1']
+        assert msgs == wanted
+
+    def test_ensure_string(self, cmd):
+        cmd.option1 = 'ok'
+        cmd.ensure_string('option1')
+
+        cmd.option2 = None
+        cmd.ensure_string('option2', 'xxx')
+        assert hasattr(cmd, 'option2')
+
+        cmd.option3 = 1
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_string('option3')
+
+    def test_ensure_filename(self, cmd):
+        cmd.option1 = __file__
+        cmd.ensure_filename('option1')
+        cmd.option2 = 'xxx'
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_filename('option2')
+
+    def test_ensure_dirname(self, cmd):
+        cmd.option1 = os.path.dirname(__file__) or os.curdir
+        cmd.ensure_dirname('option1')
+        cmd.option2 = 'xxx'
+        with pytest.raises(DistutilsOptionError):
+            cmd.ensure_dirname('option2')
+
+    def test_debug_print(self, cmd, capsys, monkeypatch):
+        cmd.debug_print('xxx')
+        assert capsys.readouterr().out == ''
+        monkeypatch.setattr(debug, 'DEBUG', True)
+        cmd.debug_print('xxx')
+        assert capsys.readouterr().out == 'xxx\n'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_config_cmd.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_config_cmd.py
new file mode 100644
index 00000000..ebee2ef9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_config_cmd.py
@@ -0,0 +1,87 @@
+"""Tests for distutils.command.config."""
+
+import os
+import sys
+from distutils._log import log
+from distutils.command.config import config, dump_file
+from distutils.tests import missing_compiler_executable, support
+
+import more_itertools
+import path
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def info_log(request, monkeypatch):
+    self = request.instance
+    self._logs = []
+    monkeypatch.setattr(log, 'info', self._info)
+
+
+@support.combine_markers
+class TestConfig(support.TempdirManager):
+    def _info(self, msg, *args):
+        for line in msg.splitlines():
+            self._logs.append(line)
+
+    def test_dump_file(self):
+        this_file = path.Path(__file__).with_suffix('.py')
+        with this_file.open(encoding='utf-8') as f:
+            numlines = more_itertools.ilen(f)
+
+        dump_file(this_file, 'I am the header')
+        assert len(self._logs) == numlines + 1
+
+    @pytest.mark.skipif('platform.system() == "Windows"')
+    def test_search_cpp(self):
+        cmd = missing_compiler_executable(['preprocessor'])
+        if cmd is not None:
+            self.skipTest(f'The {cmd!r} command is not found')
+        pkg_dir, dist = self.create_dist()
+        cmd = config(dist)
+        cmd._check_compiler()
+        compiler = cmd.compiler
+        if sys.platform[:3] == "aix" and "xlc" in compiler.preprocessor[0].lower():
+            self.skipTest(
+                'xlc: The -E option overrides the -P, -o, and -qsyntaxonly options'
+            )
+
+        # simple pattern searches
+        match = cmd.search_cpp(pattern='xxx', body='/* xxx */')
+        assert match == 0
+
+        match = cmd.search_cpp(pattern='_configtest', body='/* xxx */')
+        assert match == 1
+
+    def test_finalize_options(self):
+        # finalize_options does a bit of transformation
+        # on options
+        pkg_dir, dist = self.create_dist()
+        cmd = config(dist)
+        cmd.include_dirs = f'one{os.pathsep}two'
+        cmd.libraries = 'one'
+        cmd.library_dirs = f'three{os.pathsep}four'
+        cmd.ensure_finalized()
+
+        assert cmd.include_dirs == ['one', 'two']
+        assert cmd.libraries == ['one']
+        assert cmd.library_dirs == ['three', 'four']
+
+    def test_clean(self):
+        # _clean removes files
+        tmp_dir = self.mkdtemp()
+        f1 = os.path.join(tmp_dir, 'one')
+        f2 = os.path.join(tmp_dir, 'two')
+
+        self.write_file(f1, 'xxx')
+        self.write_file(f2, 'xxx')
+
+        for f in (f1, f2):
+            assert os.path.exists(f)
+
+        pkg_dir, dist = self.create_dist()
+        cmd = config(dist)
+        cmd._clean(f1, f2)
+
+        for f in (f1, f2):
+            assert not os.path.exists(f)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_core.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_core.py
new file mode 100644
index 00000000..bad3fb7e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_core.py
@@ -0,0 +1,130 @@
+"""Tests for distutils.core."""
+
+import distutils.core
+import io
+import os
+import sys
+from distutils.dist import Distribution
+
+import pytest
+
+# setup script that uses __file__
+setup_using___file__ = """\
+
+__file__
+
+from distutils.core import setup
+setup()
+"""
+
+setup_prints_cwd = """\
+
+import os
+print(os.getcwd())
+
+from distutils.core import setup
+setup()
+"""
+
+setup_does_nothing = """\
+from distutils.core import setup
+setup()
+"""
+
+
+setup_defines_subclass = """\
+from distutils.core import setup
+from distutils.command.install import install as _install
+
+class install(_install):
+    sub_commands = _install.sub_commands + ['cmd']
+
+setup(cmdclass={'install': install})
+"""
+
+setup_within_if_main = """\
+from distutils.core import setup
+
+def main():
+    return setup(name="setup_within_if_main")
+
+if __name__ == "__main__":
+    main()
+"""
+
+
+@pytest.fixture(autouse=True)
+def save_stdout(monkeypatch):
+    monkeypatch.setattr(sys, 'stdout', sys.stdout)
+
+
+@pytest.fixture
+def temp_file(tmp_path):
+    return tmp_path / 'file'
+
+
+@pytest.mark.usefixtures('save_env')
+@pytest.mark.usefixtures('save_argv')
+class TestCore:
+    def test_run_setup_provides_file(self, temp_file):
+        # Make sure the script can use __file__; if that's missing, the test
+        # setup.py script will raise NameError.
+        temp_file.write_text(setup_using___file__, encoding='utf-8')
+        distutils.core.run_setup(temp_file)
+
+    def test_run_setup_preserves_sys_argv(self, temp_file):
+        # Make sure run_setup does not clobber sys.argv
+        argv_copy = sys.argv.copy()
+        temp_file.write_text(setup_does_nothing, encoding='utf-8')
+        distutils.core.run_setup(temp_file)
+        assert sys.argv == argv_copy
+
+    def test_run_setup_defines_subclass(self, temp_file):
+        # Make sure the script can use __file__; if that's missing, the test
+        # setup.py script will raise NameError.
+        temp_file.write_text(setup_defines_subclass, encoding='utf-8')
+        dist = distutils.core.run_setup(temp_file)
+        install = dist.get_command_obj('install')
+        assert 'cmd' in install.sub_commands
+
+    def test_run_setup_uses_current_dir(self, tmp_path):
+        """
+        Test that the setup script is run with the current directory
+        as its own current directory.
+        """
+        sys.stdout = io.StringIO()
+        cwd = os.getcwd()
+
+        # Create a directory and write the setup.py file there:
+        setup_py = tmp_path / 'setup.py'
+        setup_py.write_text(setup_prints_cwd, encoding='utf-8')
+        distutils.core.run_setup(setup_py)
+
+        output = sys.stdout.getvalue()
+        if output.endswith("\n"):
+            output = output[:-1]
+        assert cwd == output
+
+    def test_run_setup_within_if_main(self, temp_file):
+        temp_file.write_text(setup_within_if_main, encoding='utf-8')
+        dist = distutils.core.run_setup(temp_file, stop_after="config")
+        assert isinstance(dist, Distribution)
+        assert dist.get_name() == "setup_within_if_main"
+
+    def test_run_commands(self, temp_file):
+        sys.argv = ['setup.py', 'build']
+        temp_file.write_text(setup_within_if_main, encoding='utf-8')
+        dist = distutils.core.run_setup(temp_file, stop_after="commandline")
+        assert 'build' not in dist.have_run
+        distutils.core.run_commands(dist)
+        assert 'build' in dist.have_run
+
+    def test_debug_mode(self, capsys, monkeypatch):
+        # this covers the code called when DEBUG is set
+        sys.argv = ['setup.py', '--name']
+        distutils.core.setup(name='bar')
+        assert capsys.readouterr().out == 'bar\n'
+        monkeypatch.setattr(distutils.core, 'DEBUG', True)
+        distutils.core.setup(name='bar')
+        wanted = "options (after parsing config files):\n"
+        assert capsys.readouterr().out.startswith(wanted)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cygwinccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cygwinccompiler.py
new file mode 100644
index 00000000..677bc0ac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_cygwinccompiler.py
@@ -0,0 +1,81 @@
+"""Tests for distutils.cygwinccompiler."""
+
+import os
+import sys
+from distutils import sysconfig
+from distutils.cygwinccompiler import (
+    CONFIG_H_NOTOK,
+    CONFIG_H_OK,
+    CONFIG_H_UNCERTAIN,
+    check_config_h,
+    get_msvcr,
+)
+from distutils.tests import support
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def stuff(request, monkeypatch, distutils_managed_tempdir):
+    self = request.instance
+    self.python_h = os.path.join(self.mkdtemp(), 'python.h')
+    monkeypatch.setattr(sysconfig, 'get_config_h_filename', self._get_config_h_filename)
+    monkeypatch.setattr(sys, 'version', sys.version)
+
+
+class TestCygwinCCompiler(support.TempdirManager):
+    def _get_config_h_filename(self):
+        return self.python_h
+
+    @pytest.mark.skipif('sys.platform != "cygwin"')
+    @pytest.mark.skipif('not os.path.exists("/usr/lib/libbash.dll.a")')
+    def test_find_library_file(self):
+        from distutils.cygwinccompiler import CygwinCCompiler
+
+        compiler = CygwinCCompiler()
+        link_name = "bash"
+        linkable_file = compiler.find_library_file(["/usr/lib"], link_name)
+        assert linkable_file is not None
+        assert os.path.exists(linkable_file)
+        assert linkable_file == f"/usr/lib/lib{link_name:s}.dll.a"
+
+    @pytest.mark.skipif('sys.platform != "cygwin"')
+    def test_runtime_library_dir_option(self):
+        from distutils.cygwinccompiler import CygwinCCompiler
+
+        compiler = CygwinCCompiler()
+        assert compiler.runtime_library_dir_option('/foo') == []
+
+    def test_check_config_h(self):
+        # check_config_h looks for "GCC" in sys.version first
+        # returns CONFIG_H_OK if found
+        sys.version = (
+            '2.6.1 (r261:67515, Dec  6 2008, 16:42:21) \n[GCC '
+            '4.0.1 (Apple Computer, Inc. build 5370)]'
+        )
+
+        assert check_config_h()[0] == CONFIG_H_OK
+
+        # then it tries to see if it can find "__GNUC__" in pyconfig.h
+        sys.version = 'something without the *CC word'
+
+        # if the file doesn't exist it returns  CONFIG_H_UNCERTAIN
+        assert check_config_h()[0] == CONFIG_H_UNCERTAIN
+
+        # if it exists but does not contain __GNUC__, it returns CONFIG_H_NOTOK
+        self.write_file(self.python_h, 'xxx')
+        assert check_config_h()[0] == CONFIG_H_NOTOK
+
+        # and CONFIG_H_OK if __GNUC__ is found
+        self.write_file(self.python_h, 'xxx __GNUC__ xxx')
+        assert check_config_h()[0] == CONFIG_H_OK
+
+    def test_get_msvcr(self):
+        assert get_msvcr() == []
+
+    @pytest.mark.skipif('sys.platform != "cygwin"')
+    def test_dll_libraries_not_none(self):
+        from distutils.cygwinccompiler import CygwinCCompiler
+
+        compiler = CygwinCCompiler()
+        assert compiler.dll_libraries is not None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dir_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dir_util.py
new file mode 100644
index 00000000..326cb346
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dir_util.py
@@ -0,0 +1,139 @@
+"""Tests for distutils.dir_util."""
+
+import os
+import pathlib
+import stat
+import sys
+import unittest.mock as mock
+from distutils import dir_util, errors
+from distutils.dir_util import (
+    copy_tree,
+    create_tree,
+    ensure_relative,
+    mkpath,
+    remove_tree,
+)
+from distutils.tests import support
+
+import jaraco.path
+import path
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def stuff(request, monkeypatch, distutils_managed_tempdir):
+    self = request.instance
+    tmp_dir = self.mkdtemp()
+    self.root_target = os.path.join(tmp_dir, 'deep')
+    self.target = os.path.join(self.root_target, 'here')
+    self.target2 = os.path.join(tmp_dir, 'deep2')
+
+
+class TestDirUtil(support.TempdirManager):
+    def test_mkpath_remove_tree_verbosity(self, caplog):
+        mkpath(self.target, verbose=False)
+        assert not caplog.records
+        remove_tree(self.root_target, verbose=False)
+
+        mkpath(self.target, verbose=True)
+        wanted = [f'creating {self.target}']
+        assert caplog.messages == wanted
+        caplog.clear()
+
+        remove_tree(self.root_target, verbose=True)
+        wanted = [f"removing '{self.root_target}' (and everything under it)"]
+        assert caplog.messages == wanted
+
+    @pytest.mark.skipif("platform.system() == 'Windows'")
+    def test_mkpath_with_custom_mode(self):
+        # Get and set the current umask value for testing mode bits.
+        umask = os.umask(0o002)
+        os.umask(umask)
+        mkpath(self.target, 0o700)
+        assert stat.S_IMODE(os.stat(self.target).st_mode) == 0o700 & ~umask
+        mkpath(self.target2, 0o555)
+        assert stat.S_IMODE(os.stat(self.target2).st_mode) == 0o555 & ~umask
+
+    def test_create_tree_verbosity(self, caplog):
+        create_tree(self.root_target, ['one', 'two', 'three'], verbose=False)
+        assert caplog.messages == []
+        remove_tree(self.root_target, verbose=False)
+
+        wanted = [f'creating {self.root_target}']
+        create_tree(self.root_target, ['one', 'two', 'three'], verbose=True)
+        assert caplog.messages == wanted
+
+        remove_tree(self.root_target, verbose=False)
+
+    def test_copy_tree_verbosity(self, caplog):
+        mkpath(self.target, verbose=False)
+
+        copy_tree(self.target, self.target2, verbose=False)
+        assert caplog.messages == []
+
+        remove_tree(self.root_target, verbose=False)
+
+        mkpath(self.target, verbose=False)
+        a_file = path.Path(self.target) / 'ok.txt'
+        jaraco.path.build({'ok.txt': 'some content'}, self.target)
+
+        wanted = [f'copying {a_file} -> {self.target2}']
+        copy_tree(self.target, self.target2, verbose=True)
+        assert caplog.messages == wanted
+
+        remove_tree(self.root_target, verbose=False)
+        remove_tree(self.target2, verbose=False)
+
+    def test_copy_tree_skips_nfs_temp_files(self):
+        mkpath(self.target, verbose=False)
+
+        jaraco.path.build({'ok.txt': 'some content', '.nfs123abc': ''}, self.target)
+
+        copy_tree(self.target, self.target2)
+        assert os.listdir(self.target2) == ['ok.txt']
+
+        remove_tree(self.root_target, verbose=False)
+        remove_tree(self.target2, verbose=False)
+
+    def test_ensure_relative(self):
+        if os.sep == '/':
+            assert ensure_relative('/home/foo') == 'home/foo'
+            assert ensure_relative('some/path') == 'some/path'
+        else:  # \\
+            assert ensure_relative('c:\\home\\foo') == 'c:home\\foo'
+            assert ensure_relative('home\\foo') == 'home\\foo'
+
+    def test_copy_tree_exception_in_listdir(self):
+        """
+        An exception in listdir should raise a DistutilsFileError
+        """
+        with (
+            mock.patch("os.listdir", side_effect=OSError()),
+            pytest.raises(errors.DistutilsFileError),
+        ):
+            src = self.tempdirs[-1]
+            dir_util.copy_tree(src, None)
+
+    def test_mkpath_exception_uncached(self, monkeypatch, tmp_path):
+        """
+        Caching should not remember failed attempts.
+
+        pypa/distutils#304
+        """
+
+        class FailPath(pathlib.Path):
+            def mkdir(self, *args, **kwargs):
+                raise OSError("Failed to create directory")
+
+            if sys.version_info < (3, 12):
+                _flavour = pathlib.Path()._flavour
+
+        target = tmp_path / 'foodir'
+
+        with pytest.raises(errors.DistutilsFileError):
+            mkpath(FailPath(target))
+
+        assert not target.exists()
+
+        mkpath(target)
+        assert target.exists()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dist.py
new file mode 100644
index 00000000..2c5beebe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_dist.py
@@ -0,0 +1,552 @@
+"""Tests for distutils.dist."""
+
+import email
+import email.generator
+import email.policy
+import functools
+import io
+import os
+import sys
+import textwrap
+import unittest.mock as mock
+import warnings
+from distutils.cmd import Command
+from distutils.dist import Distribution, fix_help_options
+from distutils.tests import support
+from typing import ClassVar
+
+import jaraco.path
+import pytest
+
+pydistutils_cfg = '.' * (os.name == 'posix') + 'pydistutils.cfg'
+
+
+class test_dist(Command):
+    """Sample distutils extension command."""
+
+    user_options: ClassVar[list[tuple[str, str, str]]] = [
+        ("sample-option=", "S", "help text"),
+    ]
+
+    def initialize_options(self):
+        self.sample_option = None
+
+
+class TestDistribution(Distribution):
+    """Distribution subclasses that avoids the default search for
+    configuration files.
+
+    The ._config_files attribute must be set before
+    .parse_config_files() is called.
+    """
+
+    def find_config_files(self):
+        return self._config_files
+
+
+@pytest.fixture
+def clear_argv():
+    del sys.argv[1:]
+
+
+@support.combine_markers
+@pytest.mark.usefixtures('save_env')
+@pytest.mark.usefixtures('save_argv')
+class TestDistributionBehavior(support.TempdirManager):
+    def create_distribution(self, configfiles=()):
+        d = TestDistribution()
+        d._config_files = configfiles
+        d.parse_config_files()
+        d.parse_command_line()
+        return d
+
+    def test_command_packages_unspecified(self, clear_argv):
+        sys.argv.append("build")
+        d = self.create_distribution()
+        assert d.get_command_packages() == ["distutils.command"]
+
+    def test_command_packages_cmdline(self, clear_argv):
+        from distutils.tests.test_dist import test_dist
+
+        sys.argv.extend([
+            "--command-packages",
+            "foo.bar,distutils.tests",
+            "test_dist",
+            "-Ssometext",
+        ])
+        d = self.create_distribution()
+        # let's actually try to load our test command:
+        assert d.get_command_packages() == [
+            "distutils.command",
+            "foo.bar",
+            "distutils.tests",
+        ]
+        cmd = d.get_command_obj("test_dist")
+        assert isinstance(cmd, test_dist)
+        assert cmd.sample_option == "sometext"
+
+    @pytest.mark.skipif(
+        'distutils' not in Distribution.parse_config_files.__module__,
+        reason='Cannot test when virtualenv has monkey-patched Distribution',
+    )
+    def test_venv_install_options(self, tmp_path, clear_argv):
+        sys.argv.append("install")
+        file = str(tmp_path / 'file')
+
+        fakepath = '/somedir'
+
+        jaraco.path.build({
+            file: f"""
+                    [install]
+                    install-base = {fakepath}
+                    install-platbase = {fakepath}
+                    install-lib = {fakepath}
+                    install-platlib = {fakepath}
+                    install-purelib = {fakepath}
+                    install-headers = {fakepath}
+                    install-scripts = {fakepath}
+                    install-data = {fakepath}
+                    prefix = {fakepath}
+                    exec-prefix = {fakepath}
+                    home = {fakepath}
+                    user = {fakepath}
+                    root = {fakepath}
+                    """,
+        })
+
+        # Base case: Not in a Virtual Environment
+        with mock.patch.multiple(sys, prefix='/a', base_prefix='/a'):
+            d = self.create_distribution([file])
+
+        option_tuple = (file, fakepath)
+
+        result_dict = {
+            'install_base': option_tuple,
+            'install_platbase': option_tuple,
+            'install_lib': option_tuple,
+            'install_platlib': option_tuple,
+            'install_purelib': option_tuple,
+            'install_headers': option_tuple,
+            'install_scripts': option_tuple,
+            'install_data': option_tuple,
+            'prefix': option_tuple,
+            'exec_prefix': option_tuple,
+            'home': option_tuple,
+            'user': option_tuple,
+            'root': option_tuple,
+        }
+
+        assert sorted(d.command_options.get('install').keys()) == sorted(
+            result_dict.keys()
+        )
+
+        for key, value in d.command_options.get('install').items():
+            assert value == result_dict[key]
+
+        # Test case: In a Virtual Environment
+        with mock.patch.multiple(sys, prefix='/a', base_prefix='/b'):
+            d = self.create_distribution([file])
+
+        for key in result_dict.keys():
+            assert key not in d.command_options.get('install', {})
+
+    def test_command_packages_configfile(self, tmp_path, clear_argv):
+        sys.argv.append("build")
+        file = str(tmp_path / "file")
+        jaraco.path.build({
+            file: """
+                    [global]
+                    command_packages = foo.bar, splat
+                    """,
+        })
+
+        d = self.create_distribution([file])
+        assert d.get_command_packages() == ["distutils.command", "foo.bar", "splat"]
+
+        # ensure command line overrides config:
+        sys.argv[1:] = ["--command-packages", "spork", "build"]
+        d = self.create_distribution([file])
+        assert d.get_command_packages() == ["distutils.command", "spork"]
+
+        # Setting --command-packages to '' should cause the default to
+        # be used even if a config file specified something else:
+        sys.argv[1:] = ["--command-packages", "", "build"]
+        d = self.create_distribution([file])
+        assert d.get_command_packages() == ["distutils.command"]
+
+    def test_empty_options(self, request):
+        # an empty options dictionary should not stay in the
+        # list of attributes
+
+        # catching warnings
+        warns = []
+
+        def _warn(msg):
+            warns.append(msg)
+
+        request.addfinalizer(
+            functools.partial(setattr, warnings, 'warn', warnings.warn)
+        )
+        warnings.warn = _warn
+        dist = Distribution(
+            attrs={
+                'author': 'xxx',
+                'name': 'xxx',
+                'version': 'xxx',
+                'url': 'xxxx',
+                'options': {},
+            }
+        )
+
+        assert len(warns) == 0
+        assert 'options' not in dir(dist)
+
+    def test_finalize_options(self):
+        attrs = {'keywords': 'one,two', 'platforms': 'one,two'}
+
+        dist = Distribution(attrs=attrs)
+        dist.finalize_options()
+
+        # finalize_option splits platforms and keywords
+        assert dist.metadata.platforms == ['one', 'two']
+        assert dist.metadata.keywords == ['one', 'two']
+
+        attrs = {'keywords': 'foo bar', 'platforms': 'foo bar'}
+        dist = Distribution(attrs=attrs)
+        dist.finalize_options()
+        assert dist.metadata.platforms == ['foo bar']
+        assert dist.metadata.keywords == ['foo bar']
+
+    def test_get_command_packages(self):
+        dist = Distribution()
+        assert dist.command_packages is None
+        cmds = dist.get_command_packages()
+        assert cmds == ['distutils.command']
+        assert dist.command_packages == ['distutils.command']
+
+        dist.command_packages = 'one,two'
+        cmds = dist.get_command_packages()
+        assert cmds == ['distutils.command', 'one', 'two']
+
+    def test_announce(self):
+        # make sure the level is known
+        dist = Distribution()
+        with pytest.raises(TypeError):
+            dist.announce('ok', level='ok2')
+
+    def test_find_config_files_disable(self, temp_home):
+        # Ticket #1180: Allow user to disable their home config file.
+        jaraco.path.build({pydistutils_cfg: '[distutils]\n'}, temp_home)
+
+        d = Distribution()
+        all_files = d.find_config_files()
+
+        d = Distribution(attrs={'script_args': ['--no-user-cfg']})
+        files = d.find_config_files()
+
+        # make sure --no-user-cfg disables the user cfg file
+        assert len(all_files) - 1 == len(files)
+
+    def test_script_args_list_coercion(self):
+        d = Distribution(attrs={'script_args': ('build', '--no-user-cfg')})
+
+        # make sure script_args is a list even if it started as a different iterable
+        assert d.script_args == ['build', '--no-user-cfg']
+
+    @pytest.mark.skipif(
+        'platform.system() == "Windows"',
+        reason='Windows does not honor chmod 000',
+    )
+    def test_find_config_files_permission_error(self, fake_home):
+        """
+        Finding config files should not fail when directory is inaccessible.
+        """
+        fake_home.joinpath(pydistutils_cfg).write_text('', encoding='utf-8')
+        fake_home.chmod(0o000)
+        Distribution().find_config_files()
+
+
+@pytest.mark.usefixtures('save_env')
+@pytest.mark.usefixtures('save_argv')
+class TestMetadata(support.TempdirManager):
+    def format_metadata(self, dist):
+        sio = io.StringIO()
+        dist.metadata.write_pkg_file(sio)
+        return sio.getvalue()
+
+    def test_simple_metadata(self):
+        attrs = {"name": "package", "version": "1.0"}
+        dist = Distribution(attrs)
+        meta = self.format_metadata(dist)
+        assert "Metadata-Version: 1.0" in meta
+        assert "provides:" not in meta.lower()
+        assert "requires:" not in meta.lower()
+        assert "obsoletes:" not in meta.lower()
+
+    def test_provides(self):
+        attrs = {
+            "name": "package",
+            "version": "1.0",
+            "provides": ["package", "package.sub"],
+        }
+        dist = Distribution(attrs)
+        assert dist.metadata.get_provides() == ["package", "package.sub"]
+        assert dist.get_provides() == ["package", "package.sub"]
+        meta = self.format_metadata(dist)
+        assert "Metadata-Version: 1.1" in meta
+        assert "requires:" not in meta.lower()
+        assert "obsoletes:" not in meta.lower()
+
+    def test_provides_illegal(self):
+        with pytest.raises(ValueError):
+            Distribution(
+                {"name": "package", "version": "1.0", "provides": ["my.pkg (splat)"]},
+            )
+
+    def test_requires(self):
+        attrs = {
+            "name": "package",
+            "version": "1.0",
+            "requires": ["other", "another (==1.0)"],
+        }
+        dist = Distribution(attrs)
+        assert dist.metadata.get_requires() == ["other", "another (==1.0)"]
+        assert dist.get_requires() == ["other", "another (==1.0)"]
+        meta = self.format_metadata(dist)
+        assert "Metadata-Version: 1.1" in meta
+        assert "provides:" not in meta.lower()
+        assert "Requires: other" in meta
+        assert "Requires: another (==1.0)" in meta
+        assert "obsoletes:" not in meta.lower()
+
+    def test_requires_illegal(self):
+        with pytest.raises(ValueError):
+            Distribution(
+                {"name": "package", "version": "1.0", "requires": ["my.pkg (splat)"]},
+            )
+
+    def test_requires_to_list(self):
+        attrs = {"name": "package", "requires": iter(["other"])}
+        dist = Distribution(attrs)
+        assert isinstance(dist.metadata.requires, list)
+
+    def test_obsoletes(self):
+        attrs = {
+            "name": "package",
+            "version": "1.0",
+            "obsoletes": ["other", "another (<1.0)"],
+        }
+        dist = Distribution(attrs)
+        assert dist.metadata.get_obsoletes() == ["other", "another (<1.0)"]
+        assert dist.get_obsoletes() == ["other", "another (<1.0)"]
+        meta = self.format_metadata(dist)
+        assert "Metadata-Version: 1.1" in meta
+        assert "provides:" not in meta.lower()
+        assert "requires:" not in meta.lower()
+        assert "Obsoletes: other" in meta
+        assert "Obsoletes: another (<1.0)" in meta
+
+    def test_obsoletes_illegal(self):
+        with pytest.raises(ValueError):
+            Distribution(
+                {"name": "package", "version": "1.0", "obsoletes": ["my.pkg (splat)"]},
+            )
+
+    def test_obsoletes_to_list(self):
+        attrs = {"name": "package", "obsoletes": iter(["other"])}
+        dist = Distribution(attrs)
+        assert isinstance(dist.metadata.obsoletes, list)
+
+    def test_classifier(self):
+        attrs = {
+            'name': 'Boa',
+            'version': '3.0',
+            'classifiers': ['Programming Language :: Python :: 3'],
+        }
+        dist = Distribution(attrs)
+        assert dist.get_classifiers() == ['Programming Language :: Python :: 3']
+        meta = self.format_metadata(dist)
+        assert 'Metadata-Version: 1.1' in meta
+
+    def test_classifier_invalid_type(self, caplog):
+        attrs = {
+            'name': 'Boa',
+            'version': '3.0',
+            'classifiers': ('Programming Language :: Python :: 3',),
+        }
+        d = Distribution(attrs)
+        # should have warning about passing a non-list
+        assert 'should be a list' in caplog.messages[0]
+        # should be converted to a list
+        assert isinstance(d.metadata.classifiers, list)
+        assert d.metadata.classifiers == list(attrs['classifiers'])
+
+    def test_keywords(self):
+        attrs = {
+            'name': 'Monty',
+            'version': '1.0',
+            'keywords': ['spam', 'eggs', 'life of brian'],
+        }
+        dist = Distribution(attrs)
+        assert dist.get_keywords() == ['spam', 'eggs', 'life of brian']
+
+    def test_keywords_invalid_type(self, caplog):
+        attrs = {
+            'name': 'Monty',
+            'version': '1.0',
+            'keywords': ('spam', 'eggs', 'life of brian'),
+        }
+        d = Distribution(attrs)
+        # should have warning about passing a non-list
+        assert 'should be a list' in caplog.messages[0]
+        # should be converted to a list
+        assert isinstance(d.metadata.keywords, list)
+        assert d.metadata.keywords == list(attrs['keywords'])
+
+    def test_platforms(self):
+        attrs = {
+            'name': 'Monty',
+            'version': '1.0',
+            'platforms': ['GNU/Linux', 'Some Evil Platform'],
+        }
+        dist = Distribution(attrs)
+        assert dist.get_platforms() == ['GNU/Linux', 'Some Evil Platform']
+
+    def test_platforms_invalid_types(self, caplog):
+        attrs = {
+            'name': 'Monty',
+            'version': '1.0',
+            'platforms': ('GNU/Linux', 'Some Evil Platform'),
+        }
+        d = Distribution(attrs)
+        # should have warning about passing a non-list
+        assert 'should be a list' in caplog.messages[0]
+        # should be converted to a list
+        assert isinstance(d.metadata.platforms, list)
+        assert d.metadata.platforms == list(attrs['platforms'])
+
+    def test_download_url(self):
+        attrs = {
+            'name': 'Boa',
+            'version': '3.0',
+            'download_url': 'http://example.org/boa',
+        }
+        dist = Distribution(attrs)
+        meta = self.format_metadata(dist)
+        assert 'Metadata-Version: 1.1' in meta
+
+    def test_long_description(self):
+        long_desc = textwrap.dedent(
+            """\
+        example::
+              We start here
+            and continue here
+          and end here."""
+        )
+        attrs = {"name": "package", "version": "1.0", "long_description": long_desc}
+
+        dist = Distribution(attrs)
+        meta = self.format_metadata(dist)
+        meta = meta.replace('\n' + 8 * ' ', '\n')
+        assert long_desc in meta
+
+    def test_custom_pydistutils(self, temp_home):
+        """
+        pydistutils.cfg is found
+        """
+        jaraco.path.build({pydistutils_cfg: ''}, temp_home)
+        config_path = temp_home / pydistutils_cfg
+
+        assert str(config_path) in Distribution().find_config_files()
+
+    def test_extra_pydistutils(self, monkeypatch, tmp_path):
+        jaraco.path.build({'overrides.cfg': ''}, tmp_path)
+        filename = tmp_path / 'overrides.cfg'
+        monkeypatch.setenv('DIST_EXTRA_CONFIG', str(filename))
+        assert str(filename) in Distribution().find_config_files()
+
+    def test_fix_help_options(self):
+        help_tuples = [('a', 'b', 'c', 'd'), (1, 2, 3, 4)]
+        fancy_options = fix_help_options(help_tuples)
+        assert fancy_options[0] == ('a', 'b', 'c')
+        assert fancy_options[1] == (1, 2, 3)
+
+    def test_show_help(self, request, capsys):
+        # smoke test, just makes sure some help is displayed
+        dist = Distribution()
+        sys.argv = []
+        dist.help = True
+        dist.script_name = 'setup.py'
+        dist.parse_command_line()
+
+        output = [
+            line for line in capsys.readouterr().out.split('\n') if line.strip() != ''
+        ]
+        assert output
+
+    def test_read_metadata(self):
+        attrs = {
+            "name": "package",
+            "version": "1.0",
+            "long_description": "desc",
+            "description": "xxx",
+            "download_url": "http://example.com",
+            "keywords": ['one', 'two'],
+            "requires": ['foo'],
+        }
+
+        dist = Distribution(attrs)
+        metadata = dist.metadata
+
+        # write it then reloads it
+        PKG_INFO = io.StringIO()
+        metadata.write_pkg_file(PKG_INFO)
+        PKG_INFO.seek(0)
+        metadata.read_pkg_file(PKG_INFO)
+
+        assert metadata.name == "package"
+        assert metadata.version == "1.0"
+        assert metadata.description == "xxx"
+        assert metadata.download_url == 'http://example.com'
+        assert metadata.keywords == ['one', 'two']
+        assert metadata.platforms is None
+        assert metadata.obsoletes is None
+        assert metadata.requires == ['foo']
+
+    def test_round_trip_through_email_generator(self):
+        """
+        In pypa/setuptools#4033, it was shown that once PKG-INFO is
+        re-generated using ``email.generator.Generator``, some control
+        characters might cause problems.
+        """
+        # Given a PKG-INFO file ...
+        attrs = {
+            "name": "package",
+            "version": "1.0",
+            "long_description": "hello\x0b\nworld\n",
+        }
+        dist = Distribution(attrs)
+        metadata = dist.metadata
+
+        with io.StringIO() as buffer:
+            metadata.write_pkg_file(buffer)
+            msg = buffer.getvalue()
+
+        # ... when it is read and re-written using stdlib's email library,
+        orig = email.message_from_string(msg)
+        policy = email.policy.EmailPolicy(
+            utf8=True,
+            mangle_from_=False,
+            max_line_length=0,
+        )
+        with io.StringIO() as buffer:
+            email.generator.Generator(buffer, policy=policy).flatten(orig)
+
+            buffer.seek(0)
+            regen = email.message_from_file(buffer)
+
+        # ... then it should be the same as the original
+        # (except for the specific line break characters)
+        orig_desc = set(orig["Description"].splitlines())
+        regen_desc = set(regen["Description"].splitlines())
+        assert regen_desc == orig_desc
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_extension.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_extension.py
new file mode 100644
index 00000000..5e8e7682
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_extension.py
@@ -0,0 +1,117 @@
+"""Tests for distutils.extension."""
+
+import os
+import pathlib
+import warnings
+from distutils.extension import Extension, read_setup_file
+
+import pytest
+from test.support.warnings_helper import check_warnings
+
+
+class TestExtension:
+    def test_read_setup_file(self):
+        # trying to read a Setup file
+        # (sample extracted from the PyGame project)
+        setup = os.path.join(os.path.dirname(__file__), 'Setup.sample')
+
+        exts = read_setup_file(setup)
+        names = [ext.name for ext in exts]
+        names.sort()
+
+        # here are the extensions read_setup_file should have created
+        # out of the file
+        wanted = [
+            '_arraysurfarray',
+            '_camera',
+            '_numericsndarray',
+            '_numericsurfarray',
+            'base',
+            'bufferproxy',
+            'cdrom',
+            'color',
+            'constants',
+            'display',
+            'draw',
+            'event',
+            'fastevent',
+            'font',
+            'gfxdraw',
+            'image',
+            'imageext',
+            'joystick',
+            'key',
+            'mask',
+            'mixer',
+            'mixer_music',
+            'mouse',
+            'movie',
+            'overlay',
+            'pixelarray',
+            'pypm',
+            'rect',
+            'rwobject',
+            'scrap',
+            'surface',
+            'surflock',
+            'time',
+            'transform',
+        ]
+
+        assert names == wanted
+
+    def test_extension_init(self):
+        # the first argument, which is the name, must be a string
+        with pytest.raises(TypeError):
+            Extension(1, [])
+        ext = Extension('name', [])
+        assert ext.name == 'name'
+
+        # the second argument, which is the list of files, must
+        # be an iterable of strings or PathLike objects, and not a string
+        with pytest.raises(TypeError):
+            Extension('name', 'file')
+        with pytest.raises(TypeError):
+            Extension('name', ['file', 1])
+        ext = Extension('name', ['file1', 'file2'])
+        assert ext.sources == ['file1', 'file2']
+        ext = Extension('name', [pathlib.Path('file1'), pathlib.Path('file2')])
+        assert ext.sources == ['file1', 'file2']
+
+        # any non-string iterable of strings or PathLike objects should work
+        ext = Extension('name', ('file1', 'file2'))  # tuple
+        assert ext.sources == ['file1', 'file2']
+        ext = Extension('name', {'file1', 'file2'})  # set
+        assert sorted(ext.sources) == ['file1', 'file2']
+        ext = Extension('name', iter(['file1', 'file2']))  # iterator
+        assert ext.sources == ['file1', 'file2']
+        ext = Extension('name', [pathlib.Path('file1'), 'file2'])  # mixed types
+        assert ext.sources == ['file1', 'file2']
+
+        # others arguments have defaults
+        for attr in (
+            'include_dirs',
+            'define_macros',
+            'undef_macros',
+            'library_dirs',
+            'libraries',
+            'runtime_library_dirs',
+            'extra_objects',
+            'extra_compile_args',
+            'extra_link_args',
+            'export_symbols',
+            'swig_opts',
+            'depends',
+        ):
+            assert getattr(ext, attr) == []
+
+        assert ext.language is None
+        assert ext.optional is None
+
+        # if there are unknown keyword options, warn about them
+        with check_warnings() as w:
+            warnings.simplefilter('always')
+            ext = Extension('name', ['file1', 'file2'], chic=True)
+
+        assert len(w.warnings) == 1
+        assert str(w.warnings[0].message) == "Unknown Extension options: 'chic'"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_file_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_file_util.py
new file mode 100644
index 00000000..a75d4a03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_file_util.py
@@ -0,0 +1,95 @@
+"""Tests for distutils.file_util."""
+
+import errno
+import os
+import unittest.mock as mock
+from distutils.errors import DistutilsFileError
+from distutils.file_util import copy_file, move_file
+
+import jaraco.path
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def stuff(request, tmp_path):
+    self = request.instance
+    self.source = tmp_path / 'f1'
+    self.target = tmp_path / 'f2'
+    self.target_dir = tmp_path / 'd1'
+
+
+class TestFileUtil:
+    def test_move_file_verbosity(self, caplog):
+        jaraco.path.build({self.source: 'some content'})
+
+        move_file(self.source, self.target, verbose=False)
+        assert not caplog.messages
+
+        # back to original state
+        move_file(self.target, self.source, verbose=False)
+
+        move_file(self.source, self.target, verbose=True)
+        wanted = [f'moving {self.source} -> {self.target}']
+        assert caplog.messages == wanted
+
+        # back to original state
+        move_file(self.target, self.source, verbose=False)
+
+        caplog.clear()
+        # now the target is a dir
+        os.mkdir(self.target_dir)
+        move_file(self.source, self.target_dir, verbose=True)
+        wanted = [f'moving {self.source} -> {self.target_dir}']
+        assert caplog.messages == wanted
+
+    def test_move_file_exception_unpacking_rename(self):
+        # see issue 22182
+        with (
+            mock.patch("os.rename", side_effect=OSError("wrong", 1)),
+            pytest.raises(DistutilsFileError),
+        ):
+            jaraco.path.build({self.source: 'spam eggs'})
+            move_file(self.source, self.target, verbose=False)
+
+    def test_move_file_exception_unpacking_unlink(self):
+        # see issue 22182
+        with (
+            mock.patch("os.rename", side_effect=OSError(errno.EXDEV, "wrong")),
+            mock.patch("os.unlink", side_effect=OSError("wrong", 1)),
+            pytest.raises(DistutilsFileError),
+        ):
+            jaraco.path.build({self.source: 'spam eggs'})
+            move_file(self.source, self.target, verbose=False)
+
+    def test_copy_file_hard_link(self):
+        jaraco.path.build({self.source: 'some content'})
+        # Check first that copy_file() will not fall back on copying the file
+        # instead of creating the hard link.
+        try:
+            os.link(self.source, self.target)
+        except OSError as e:
+            self.skipTest(f'os.link: {e}')
+        else:
+            self.target.unlink()
+        st = os.stat(self.source)
+        copy_file(self.source, self.target, link='hard')
+        st2 = os.stat(self.source)
+        st3 = os.stat(self.target)
+        assert os.path.samestat(st, st2), (st, st2)
+        assert os.path.samestat(st2, st3), (st2, st3)
+        assert self.source.read_text(encoding='utf-8') == 'some content'
+
+    def test_copy_file_hard_link_failure(self):
+        # If hard linking fails, copy_file() falls back on copying file
+        # (some special filesystems don't support hard linking even under
+        #  Unix, see issue #8876).
+        jaraco.path.build({self.source: 'some content'})
+        st = os.stat(self.source)
+        with mock.patch("os.link", side_effect=OSError(0, "linking unsupported")):
+            copy_file(self.source, self.target, link='hard')
+        st2 = os.stat(self.source)
+        st3 = os.stat(self.target)
+        assert os.path.samestat(st, st2), (st, st2)
+        assert not os.path.samestat(st2, st3), (st2, st3)
+        for fn in (self.source, self.target):
+            assert fn.read_text(encoding='utf-8') == 'some content'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_filelist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_filelist.py
new file mode 100644
index 00000000..130e6fb5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_filelist.py
@@ -0,0 +1,336 @@
+"""Tests for distutils.filelist."""
+
+import logging
+import os
+import re
+from distutils import debug, filelist
+from distutils.errors import DistutilsTemplateError
+from distutils.filelist import FileList, glob_to_re, translate_pattern
+
+import jaraco.path
+import pytest
+
+from .compat import py39 as os_helper
+
+MANIFEST_IN = """\
+include ok
+include xo
+exclude xo
+include foo.tmp
+include buildout.cfg
+global-include *.x
+global-include *.txt
+global-exclude *.tmp
+recursive-include f *.oo
+recursive-exclude global *.x
+graft dir
+prune dir3
+"""
+
+
+def make_local_path(s):
+    """Converts '/' in a string to os.sep"""
+    return s.replace('/', os.sep)
+
+
+class TestFileList:
+    def assertNoWarnings(self, caplog):
+        warnings = [rec for rec in caplog.records if rec.levelno == logging.WARNING]
+        assert not warnings
+        caplog.clear()
+
+    def assertWarnings(self, caplog):
+        warnings = [rec for rec in caplog.records if rec.levelno == logging.WARNING]
+        assert warnings
+        caplog.clear()
+
+    def test_glob_to_re(self):
+        sep = os.sep
+        if os.sep == '\\':
+            sep = re.escape(os.sep)
+
+        for glob, regex in (
+            # simple cases
+            ('foo*', r'(?s:foo[^%(sep)s]*)\Z'),
+            ('foo?', r'(?s:foo[^%(sep)s])\Z'),
+            ('foo??', r'(?s:foo[^%(sep)s][^%(sep)s])\Z'),
+            # special cases
+            (r'foo\\*', r'(?s:foo\\\\[^%(sep)s]*)\Z'),
+            (r'foo\\\*', r'(?s:foo\\\\\\[^%(sep)s]*)\Z'),
+            ('foo????', r'(?s:foo[^%(sep)s][^%(sep)s][^%(sep)s][^%(sep)s])\Z'),
+            (r'foo\\??', r'(?s:foo\\\\[^%(sep)s][^%(sep)s])\Z'),
+        ):
+            regex = regex % {'sep': sep}
+            assert glob_to_re(glob) == regex
+
+    def test_process_template_line(self):
+        # testing  all MANIFEST.in template patterns
+        file_list = FileList()
+        mlp = make_local_path
+
+        # simulated file list
+        file_list.allfiles = [
+            'foo.tmp',
+            'ok',
+            'xo',
+            'four.txt',
+            'buildout.cfg',
+            # filelist does not filter out VCS directories,
+            # it's sdist that does
+            mlp('.hg/last-message.txt'),
+            mlp('global/one.txt'),
+            mlp('global/two.txt'),
+            mlp('global/files.x'),
+            mlp('global/here.tmp'),
+            mlp('f/o/f.oo'),
+            mlp('dir/graft-one'),
+            mlp('dir/dir2/graft2'),
+            mlp('dir3/ok'),
+            mlp('dir3/sub/ok.txt'),
+        ]
+
+        for line in MANIFEST_IN.split('\n'):
+            if line.strip() == '':
+                continue
+            file_list.process_template_line(line)
+
+        wanted = [
+            'ok',
+            'buildout.cfg',
+            'four.txt',
+            mlp('.hg/last-message.txt'),
+            mlp('global/one.txt'),
+            mlp('global/two.txt'),
+            mlp('f/o/f.oo'),
+            mlp('dir/graft-one'),
+            mlp('dir/dir2/graft2'),
+        ]
+
+        assert file_list.files == wanted
+
+    def test_debug_print(self, capsys, monkeypatch):
+        file_list = FileList()
+        file_list.debug_print('xxx')
+        assert capsys.readouterr().out == ''
+
+        monkeypatch.setattr(debug, 'DEBUG', True)
+        file_list.debug_print('xxx')
+        assert capsys.readouterr().out == 'xxx\n'
+
+    def test_set_allfiles(self):
+        file_list = FileList()
+        files = ['a', 'b', 'c']
+        file_list.set_allfiles(files)
+        assert file_list.allfiles == files
+
+    def test_remove_duplicates(self):
+        file_list = FileList()
+        file_list.files = ['a', 'b', 'a', 'g', 'c', 'g']
+        # files must be sorted beforehand (sdist does it)
+        file_list.sort()
+        file_list.remove_duplicates()
+        assert file_list.files == ['a', 'b', 'c', 'g']
+
+    def test_translate_pattern(self):
+        # not regex
+        assert hasattr(translate_pattern('a', anchor=True, is_regex=False), 'search')
+
+        # is a regex
+        regex = re.compile('a')
+        assert translate_pattern(regex, anchor=True, is_regex=True) == regex
+
+        # plain string flagged as regex
+        assert hasattr(translate_pattern('a', anchor=True, is_regex=True), 'search')
+
+        # glob support
+        assert translate_pattern('*.py', anchor=True, is_regex=False).search(
+            'filelist.py'
+        )
+
+    def test_exclude_pattern(self):
+        # return False if no match
+        file_list = FileList()
+        assert not file_list.exclude_pattern('*.py')
+
+        # return True if files match
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.py']
+        assert file_list.exclude_pattern('*.py')
+
+        # test excludes
+        file_list = FileList()
+        file_list.files = ['a.py', 'a.txt']
+        file_list.exclude_pattern('*.py')
+        assert file_list.files == ['a.txt']
+
+    def test_include_pattern(self):
+        # return False if no match
+        file_list = FileList()
+        file_list.set_allfiles([])
+        assert not file_list.include_pattern('*.py')
+
+        # return True if files match
+        file_list = FileList()
+        file_list.set_allfiles(['a.py', 'b.txt'])
+        assert file_list.include_pattern('*.py')
+
+        # test * matches all files
+        file_list = FileList()
+        assert file_list.allfiles is None
+        file_list.set_allfiles(['a.py', 'b.txt'])
+        file_list.include_pattern('*')
+        assert file_list.allfiles == ['a.py', 'b.txt']
+
+    def test_process_template(self, caplog):
+        mlp = make_local_path
+        # invalid lines
+        file_list = FileList()
+        for action in (
+            'include',
+            'exclude',
+            'global-include',
+            'global-exclude',
+            'recursive-include',
+            'recursive-exclude',
+            'graft',
+            'prune',
+            'blarg',
+        ):
+            with pytest.raises(DistutilsTemplateError):
+                file_list.process_template_line(action)
+
+        # include
+        file_list = FileList()
+        file_list.set_allfiles(['a.py', 'b.txt', mlp('d/c.py')])
+
+        file_list.process_template_line('include *.py')
+        assert file_list.files == ['a.py']
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('include *.rb')
+        assert file_list.files == ['a.py']
+        self.assertWarnings(caplog)
+
+        # exclude
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.txt', mlp('d/c.py')]
+
+        file_list.process_template_line('exclude *.py')
+        assert file_list.files == ['b.txt', mlp('d/c.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('exclude *.rb')
+        assert file_list.files == ['b.txt', mlp('d/c.py')]
+        self.assertWarnings(caplog)
+
+        # global-include
+        file_list = FileList()
+        file_list.set_allfiles(['a.py', 'b.txt', mlp('d/c.py')])
+
+        file_list.process_template_line('global-include *.py')
+        assert file_list.files == ['a.py', mlp('d/c.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('global-include *.rb')
+        assert file_list.files == ['a.py', mlp('d/c.py')]
+        self.assertWarnings(caplog)
+
+        # global-exclude
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.txt', mlp('d/c.py')]
+
+        file_list.process_template_line('global-exclude *.py')
+        assert file_list.files == ['b.txt']
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('global-exclude *.rb')
+        assert file_list.files == ['b.txt']
+        self.assertWarnings(caplog)
+
+        # recursive-include
+        file_list = FileList()
+        file_list.set_allfiles(['a.py', mlp('d/b.py'), mlp('d/c.txt'), mlp('d/d/e.py')])
+
+        file_list.process_template_line('recursive-include d *.py')
+        assert file_list.files == [mlp('d/b.py'), mlp('d/d/e.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('recursive-include e *.py')
+        assert file_list.files == [mlp('d/b.py'), mlp('d/d/e.py')]
+        self.assertWarnings(caplog)
+
+        # recursive-exclude
+        file_list = FileList()
+        file_list.files = ['a.py', mlp('d/b.py'), mlp('d/c.txt'), mlp('d/d/e.py')]
+
+        file_list.process_template_line('recursive-exclude d *.py')
+        assert file_list.files == ['a.py', mlp('d/c.txt')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('recursive-exclude e *.py')
+        assert file_list.files == ['a.py', mlp('d/c.txt')]
+        self.assertWarnings(caplog)
+
+        # graft
+        file_list = FileList()
+        file_list.set_allfiles(['a.py', mlp('d/b.py'), mlp('d/d/e.py'), mlp('f/f.py')])
+
+        file_list.process_template_line('graft d')
+        assert file_list.files == [mlp('d/b.py'), mlp('d/d/e.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('graft e')
+        assert file_list.files == [mlp('d/b.py'), mlp('d/d/e.py')]
+        self.assertWarnings(caplog)
+
+        # prune
+        file_list = FileList()
+        file_list.files = ['a.py', mlp('d/b.py'), mlp('d/d/e.py'), mlp('f/f.py')]
+
+        file_list.process_template_line('prune d')
+        assert file_list.files == ['a.py', mlp('f/f.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('prune e')
+        assert file_list.files == ['a.py', mlp('f/f.py')]
+        self.assertWarnings(caplog)
+
+
+class TestFindAll:
+    @os_helper.skip_unless_symlink
+    def test_missing_symlink(self, temp_cwd):
+        os.symlink('foo', 'bar')
+        assert filelist.findall() == []
+
+    def test_basic_discovery(self, temp_cwd):
+        """
+        When findall is called with no parameters or with
+        '.' as the parameter, the dot should be omitted from
+        the results.
+        """
+        jaraco.path.build({'foo': {'file1.txt': ''}, 'bar': {'file2.txt': ''}})
+        file1 = os.path.join('foo', 'file1.txt')
+        file2 = os.path.join('bar', 'file2.txt')
+        expected = [file2, file1]
+        assert sorted(filelist.findall()) == expected
+
+    def test_non_local_discovery(self, tmp_path):
+        """
+        When findall is called with another path, the full
+        path name should be returned.
+        """
+        jaraco.path.build({'file1.txt': ''}, tmp_path)
+        expected = [str(tmp_path / 'file1.txt')]
+        assert filelist.findall(tmp_path) == expected
+
+    @os_helper.skip_unless_symlink
+    def test_symlink_loop(self, tmp_path):
+        jaraco.path.build(
+            {
+                'link-to-parent': jaraco.path.Symlink('.'),
+                'somefile': '',
+            },
+            tmp_path,
+        )
+        files = filelist.findall(tmp_path)
+        assert len(files) == 1
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install.py
new file mode 100644
index 00000000..b3ffb2e6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install.py
@@ -0,0 +1,245 @@
+"""Tests for distutils.command.install."""
+
+import logging
+import os
+import pathlib
+import site
+import sys
+from distutils import sysconfig
+from distutils.command import install as install_module
+from distutils.command.build_ext import build_ext
+from distutils.command.install import INSTALL_SCHEMES, install
+from distutils.core import Distribution
+from distutils.errors import DistutilsOptionError
+from distutils.extension import Extension
+from distutils.tests import missing_compiler_executable, support
+from distutils.util import is_mingw
+
+import pytest
+
+
+def _make_ext_name(modname):
+    return modname + sysconfig.get_config_var('EXT_SUFFIX')
+
+
+@support.combine_markers
+@pytest.mark.usefixtures('save_env')
+class TestInstall(
+    support.TempdirManager,
+):
+    @pytest.mark.xfail(
+        'platform.system() == "Windows" and sys.version_info > (3, 11)',
+        reason="pypa/distutils#148",
+    )
+    def test_home_installation_scheme(self):
+        # This ensure two things:
+        # - that --home generates the desired set of directory names
+        # - test --home is supported on all platforms
+        builddir = self.mkdtemp()
+        destination = os.path.join(builddir, "installation")
+
+        dist = Distribution({"name": "foopkg"})
+        # script_name need not exist, it just need to be initialized
+        dist.script_name = os.path.join(builddir, "setup.py")
+        dist.command_obj["build"] = support.DummyCommand(
+            build_base=builddir,
+            build_lib=os.path.join(builddir, "lib"),
+        )
+
+        cmd = install(dist)
+        cmd.home = destination
+        cmd.ensure_finalized()
+
+        assert cmd.install_base == destination
+        assert cmd.install_platbase == destination
+
+        def check_path(got, expected):
+            got = os.path.normpath(got)
+            expected = os.path.normpath(expected)
+            assert got == expected
+
+        impl_name = sys.implementation.name.replace("cpython", "python")
+        libdir = os.path.join(destination, "lib", impl_name)
+        check_path(cmd.install_lib, libdir)
+        _platlibdir = getattr(sys, "platlibdir", "lib")
+        platlibdir = os.path.join(destination, _platlibdir, impl_name)
+        check_path(cmd.install_platlib, platlibdir)
+        check_path(cmd.install_purelib, libdir)
+        check_path(
+            cmd.install_headers,
+            os.path.join(destination, "include", impl_name, "foopkg"),
+        )
+        check_path(cmd.install_scripts, os.path.join(destination, "bin"))
+        check_path(cmd.install_data, destination)
+
+    def test_user_site(self, monkeypatch):
+        # test install with --user
+        # preparing the environment for the test
+        self.tmpdir = self.mkdtemp()
+        orig_site = site.USER_SITE
+        orig_base = site.USER_BASE
+        monkeypatch.setattr(site, 'USER_BASE', os.path.join(self.tmpdir, 'B'))
+        monkeypatch.setattr(site, 'USER_SITE', os.path.join(self.tmpdir, 'S'))
+        monkeypatch.setattr(install_module, 'USER_BASE', site.USER_BASE)
+        monkeypatch.setattr(install_module, 'USER_SITE', site.USER_SITE)
+
+        def _expanduser(path):
+            if path.startswith('~'):
+                return os.path.normpath(self.tmpdir + path[1:])
+            return path
+
+        monkeypatch.setattr(os.path, 'expanduser', _expanduser)
+
+        for key in ('nt_user', 'posix_user'):
+            assert key in INSTALL_SCHEMES
+
+        dist = Distribution({'name': 'xx'})
+        cmd = install(dist)
+
+        # making sure the user option is there
+        options = [name for name, short, label in cmd.user_options]
+        assert 'user' in options
+
+        # setting a value
+        cmd.user = True
+
+        # user base and site shouldn't be created yet
+        assert not os.path.exists(site.USER_BASE)
+        assert not os.path.exists(site.USER_SITE)
+
+        # let's run finalize
+        cmd.ensure_finalized()
+
+        # now they should
+        assert os.path.exists(site.USER_BASE)
+        assert os.path.exists(site.USER_SITE)
+
+        assert 'userbase' in cmd.config_vars
+        assert 'usersite' in cmd.config_vars
+
+        actual_headers = os.path.relpath(cmd.install_headers, site.USER_BASE)
+        if os.name == 'nt' and not is_mingw():
+            site_path = os.path.relpath(os.path.dirname(orig_site), orig_base)
+            include = os.path.join(site_path, 'Include')
+        else:
+            include = sysconfig.get_python_inc(0, '')
+        expect_headers = os.path.join(include, 'xx')
+
+        assert os.path.normcase(actual_headers) == os.path.normcase(expect_headers)
+
+    def test_handle_extra_path(self):
+        dist = Distribution({'name': 'xx', 'extra_path': 'path,dirs'})
+        cmd = install(dist)
+
+        # two elements
+        cmd.handle_extra_path()
+        assert cmd.extra_path == ['path', 'dirs']
+        assert cmd.extra_dirs == 'dirs'
+        assert cmd.path_file == 'path'
+
+        # one element
+        cmd.extra_path = ['path']
+        cmd.handle_extra_path()
+        assert cmd.extra_path == ['path']
+        assert cmd.extra_dirs == 'path'
+        assert cmd.path_file == 'path'
+
+        # none
+        dist.extra_path = cmd.extra_path = None
+        cmd.handle_extra_path()
+        assert cmd.extra_path is None
+        assert cmd.extra_dirs == ''
+        assert cmd.path_file is None
+
+        # three elements (no way !)
+        cmd.extra_path = 'path,dirs,again'
+        with pytest.raises(DistutilsOptionError):
+            cmd.handle_extra_path()
+
+    def test_finalize_options(self):
+        dist = Distribution({'name': 'xx'})
+        cmd = install(dist)
+
+        # must supply either prefix/exec-prefix/home or
+        # install-base/install-platbase -- not both
+        cmd.prefix = 'prefix'
+        cmd.install_base = 'base'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+
+        # must supply either home or prefix/exec-prefix -- not both
+        cmd.install_base = None
+        cmd.home = 'home'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+
+        # can't combine user with prefix/exec_prefix/home or
+        # install_(plat)base
+        cmd.prefix = None
+        cmd.user = 'user'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+
+    def test_record(self):
+        install_dir = self.mkdtemp()
+        project_dir, dist = self.create_dist(py_modules=['hello'], scripts=['sayhi'])
+        os.chdir(project_dir)
+        self.write_file('hello.py', "def main(): print('o hai')")
+        self.write_file('sayhi', 'from hello import main; main()')
+
+        cmd = install(dist)
+        dist.command_obj['install'] = cmd
+        cmd.root = install_dir
+        cmd.record = os.path.join(project_dir, 'filelist')
+        cmd.ensure_finalized()
+        cmd.run()
+
+        content = pathlib.Path(cmd.record).read_text(encoding='utf-8')
+
+        found = [pathlib.Path(line).name for line in content.splitlines()]
+        expected = [
+            'hello.py',
+            f'hello.{sys.implementation.cache_tag}.pyc',
+            'sayhi',
+            'UNKNOWN-0.0.0-py{}.{}.egg-info'.format(*sys.version_info[:2]),
+        ]
+        assert found == expected
+
+    def test_record_extensions(self):
+        cmd = missing_compiler_executable()
+        if cmd is not None:
+            pytest.skip(f'The {cmd!r} command is not found')
+        install_dir = self.mkdtemp()
+        project_dir, dist = self.create_dist(
+            ext_modules=[Extension('xx', ['xxmodule.c'])]
+        )
+        os.chdir(project_dir)
+        support.copy_xxmodule_c(project_dir)
+
+        buildextcmd = build_ext(dist)
+        support.fixup_build_ext(buildextcmd)
+        buildextcmd.ensure_finalized()
+
+        cmd = install(dist)
+        dist.command_obj['install'] = cmd
+        dist.command_obj['build_ext'] = buildextcmd
+        cmd.root = install_dir
+        cmd.record = os.path.join(project_dir, 'filelist')
+        cmd.ensure_finalized()
+        cmd.run()
+
+        content = pathlib.Path(cmd.record).read_text(encoding='utf-8')
+
+        found = [pathlib.Path(line).name for line in content.splitlines()]
+        expected = [
+            _make_ext_name('xx'),
+            'UNKNOWN-0.0.0-py{}.{}.egg-info'.format(*sys.version_info[:2]),
+        ]
+        assert found == expected
+
+    def test_debug_mode(self, caplog, monkeypatch):
+        # this covers the code called when DEBUG is set
+        monkeypatch.setattr(install_module, 'DEBUG', True)
+        caplog.set_level(logging.DEBUG)
+        self.test_record()
+        assert any(rec for rec in caplog.records if rec.levelno == logging.DEBUG)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_data.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_data.py
new file mode 100644
index 00000000..c800f86c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_data.py
@@ -0,0 +1,74 @@
+"""Tests for distutils.command.install_data."""
+
+import os
+import pathlib
+from distutils.command.install_data import install_data
+from distutils.tests import support
+
+import pytest
+
+
+@pytest.mark.usefixtures('save_env')
+class TestInstallData(
+    support.TempdirManager,
+):
+    def test_simple_run(self):
+        pkg_dir, dist = self.create_dist()
+        cmd = install_data(dist)
+        cmd.install_dir = inst = os.path.join(pkg_dir, 'inst')
+
+        # data_files can contain
+        #  - simple files
+        #  - a Path object
+        #  - a tuple with a path, and a list of file
+        one = os.path.join(pkg_dir, 'one')
+        self.write_file(one, 'xxx')
+        inst2 = os.path.join(pkg_dir, 'inst2')
+        two = os.path.join(pkg_dir, 'two')
+        self.write_file(two, 'xxx')
+        three = pathlib.Path(pkg_dir) / 'three'
+        self.write_file(three, 'xxx')
+
+        cmd.data_files = [one, (inst2, [two]), three]
+        assert cmd.get_inputs() == [one, (inst2, [two]), three]
+
+        # let's run the command
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # let's check the result
+        assert len(cmd.get_outputs()) == 3
+        rthree = os.path.split(one)[-1]
+        assert os.path.exists(os.path.join(inst, rthree))
+        rtwo = os.path.split(two)[-1]
+        assert os.path.exists(os.path.join(inst2, rtwo))
+        rone = os.path.split(one)[-1]
+        assert os.path.exists(os.path.join(inst, rone))
+        cmd.outfiles = []
+
+        # let's try with warn_dir one
+        cmd.warn_dir = True
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # let's check the result
+        assert len(cmd.get_outputs()) == 3
+        assert os.path.exists(os.path.join(inst, rthree))
+        assert os.path.exists(os.path.join(inst2, rtwo))
+        assert os.path.exists(os.path.join(inst, rone))
+        cmd.outfiles = []
+
+        # now using root and empty dir
+        cmd.root = os.path.join(pkg_dir, 'root')
+        inst5 = os.path.join(pkg_dir, 'inst5')
+        four = os.path.join(cmd.install_dir, 'four')
+        self.write_file(four, 'xx')
+        cmd.data_files = [one, (inst2, [two]), three, ('inst5', [four]), (inst5, [])]
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # let's check the result
+        assert len(cmd.get_outputs()) == 5
+        assert os.path.exists(os.path.join(inst, rthree))
+        assert os.path.exists(os.path.join(inst2, rtwo))
+        assert os.path.exists(os.path.join(inst, rone))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_headers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_headers.py
new file mode 100644
index 00000000..2c74f06b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_headers.py
@@ -0,0 +1,33 @@
+"""Tests for distutils.command.install_headers."""
+
+import os
+from distutils.command.install_headers import install_headers
+from distutils.tests import support
+
+import pytest
+
+
+@pytest.mark.usefixtures('save_env')
+class TestInstallHeaders(
+    support.TempdirManager,
+):
+    def test_simple_run(self):
+        # we have two headers
+        header_list = self.mkdtemp()
+        header1 = os.path.join(header_list, 'header1')
+        header2 = os.path.join(header_list, 'header2')
+        self.write_file(header1)
+        self.write_file(header2)
+        headers = [header1, header2]
+
+        pkg_dir, dist = self.create_dist(headers=headers)
+        cmd = install_headers(dist)
+        assert cmd.get_inputs() == headers
+
+        # let's run the command
+        cmd.install_dir = os.path.join(pkg_dir, 'inst')
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # let's check the results
+        assert len(cmd.get_outputs()) == 2
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_lib.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_lib.py
new file mode 100644
index 00000000..f685a579
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_lib.py
@@ -0,0 +1,110 @@
+"""Tests for distutils.command.install_data."""
+
+import importlib.util
+import os
+import sys
+from distutils.command.install_lib import install_lib
+from distutils.errors import DistutilsOptionError
+from distutils.extension import Extension
+from distutils.tests import support
+
+import pytest
+
+
+@support.combine_markers
+@pytest.mark.usefixtures('save_env')
+class TestInstallLib(
+    support.TempdirManager,
+):
+    def test_finalize_options(self):
+        dist = self.create_dist()[1]
+        cmd = install_lib(dist)
+
+        cmd.finalize_options()
+        assert cmd.compile == 1
+        assert cmd.optimize == 0
+
+        # optimize must be 0, 1, or 2
+        cmd.optimize = 'foo'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+        cmd.optimize = '4'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+
+        cmd.optimize = '2'
+        cmd.finalize_options()
+        assert cmd.optimize == 2
+
+    @pytest.mark.skipif('sys.dont_write_bytecode')
+    def test_byte_compile(self):
+        project_dir, dist = self.create_dist()
+        os.chdir(project_dir)
+        cmd = install_lib(dist)
+        cmd.compile = cmd.optimize = 1
+
+        f = os.path.join(project_dir, 'foo.py')
+        self.write_file(f, '# python file')
+        cmd.byte_compile([f])
+        pyc_file = importlib.util.cache_from_source('foo.py', optimization='')
+        pyc_opt_file = importlib.util.cache_from_source(
+            'foo.py', optimization=cmd.optimize
+        )
+        assert os.path.exists(pyc_file)
+        assert os.path.exists(pyc_opt_file)
+
+    def test_get_outputs(self):
+        project_dir, dist = self.create_dist()
+        os.chdir(project_dir)
+        os.mkdir('spam')
+        cmd = install_lib(dist)
+
+        # setting up a dist environment
+        cmd.compile = cmd.optimize = 1
+        cmd.install_dir = self.mkdtemp()
+        f = os.path.join(project_dir, 'spam', '__init__.py')
+        self.write_file(f, '# python package')
+        cmd.distribution.ext_modules = [Extension('foo', ['xxx'])]
+        cmd.distribution.packages = ['spam']
+        cmd.distribution.script_name = 'setup.py'
+
+        # get_outputs should return 4 elements: spam/__init__.py and .pyc,
+        # foo.import-tag-abiflags.so / foo.pyd
+        outputs = cmd.get_outputs()
+        assert len(outputs) == 4, outputs
+
+    def test_get_inputs(self):
+        project_dir, dist = self.create_dist()
+        os.chdir(project_dir)
+        os.mkdir('spam')
+        cmd = install_lib(dist)
+
+        # setting up a dist environment
+        cmd.compile = cmd.optimize = 1
+        cmd.install_dir = self.mkdtemp()
+        f = os.path.join(project_dir, 'spam', '__init__.py')
+        self.write_file(f, '# python package')
+        cmd.distribution.ext_modules = [Extension('foo', ['xxx'])]
+        cmd.distribution.packages = ['spam']
+        cmd.distribution.script_name = 'setup.py'
+
+        # get_inputs should return 2 elements: spam/__init__.py and
+        # foo.import-tag-abiflags.so / foo.pyd
+        inputs = cmd.get_inputs()
+        assert len(inputs) == 2, inputs
+
+    def test_dont_write_bytecode(self, caplog):
+        # makes sure byte_compile is not used
+        dist = self.create_dist()[1]
+        cmd = install_lib(dist)
+        cmd.compile = True
+        cmd.optimize = 1
+
+        old_dont_write_bytecode = sys.dont_write_bytecode
+        sys.dont_write_bytecode = True
+        try:
+            cmd.byte_compile([])
+        finally:
+            sys.dont_write_bytecode = old_dont_write_bytecode
+
+        assert 'byte-compiling is disabled' in caplog.messages[0]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_scripts.py
new file mode 100644
index 00000000..868b1c22
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_install_scripts.py
@@ -0,0 +1,52 @@
+"""Tests for distutils.command.install_scripts."""
+
+import os
+from distutils.command.install_scripts import install_scripts
+from distutils.core import Distribution
+from distutils.tests import support
+
+from . import test_build_scripts
+
+
+class TestInstallScripts(support.TempdirManager):
+    def test_default_settings(self):
+        dist = Distribution()
+        dist.command_obj["build"] = support.DummyCommand(build_scripts="/foo/bar")
+        dist.command_obj["install"] = support.DummyCommand(
+            install_scripts="/splat/funk",
+            force=True,
+            skip_build=True,
+        )
+        cmd = install_scripts(dist)
+        assert not cmd.force
+        assert not cmd.skip_build
+        assert cmd.build_dir is None
+        assert cmd.install_dir is None
+
+        cmd.finalize_options()
+
+        assert cmd.force
+        assert cmd.skip_build
+        assert cmd.build_dir == "/foo/bar"
+        assert cmd.install_dir == "/splat/funk"
+
+    def test_installation(self):
+        source = self.mkdtemp()
+
+        expected = test_build_scripts.TestBuildScripts.write_sample_scripts(source)
+
+        target = self.mkdtemp()
+        dist = Distribution()
+        dist.command_obj["build"] = support.DummyCommand(build_scripts=source)
+        dist.command_obj["install"] = support.DummyCommand(
+            install_scripts=target,
+            force=True,
+            skip_build=True,
+        )
+        cmd = install_scripts(dist)
+        cmd.finalize_options()
+        cmd.run()
+
+        installed = os.listdir(target)
+        for name in expected:
+            assert name in installed
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_log.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_log.py
new file mode 100644
index 00000000..d67779fc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_log.py
@@ -0,0 +1,12 @@
+"""Tests for distutils.log"""
+
+import logging
+from distutils._log import log
+
+
+class TestLog:
+    def test_non_ascii(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        log.debug('Dεbug\tMėssãge')
+        log.fatal('Fαtal\tÈrrōr')
+        assert caplog.messages == ['Dεbug\tMėssãge', 'Fαtal\tÈrrōr']
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_mingwccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_mingwccompiler.py
new file mode 100644
index 00000000..3e3ad505
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_mingwccompiler.py
@@ -0,0 +1,56 @@
+from distutils import sysconfig
+from distutils.errors import CCompilerError, DistutilsPlatformError
+from distutils.util import is_mingw, split_quoted
+
+import pytest
+
+
+class TestMingw32CCompiler:
+    @pytest.mark.skipif(not is_mingw(), reason='not on mingw')
+    def test_compiler_type(self):
+        from distutils.cygwinccompiler import Mingw32CCompiler
+
+        compiler = Mingw32CCompiler()
+        assert compiler.compiler_type == 'mingw32'
+
+    @pytest.mark.skipif(not is_mingw(), reason='not on mingw')
+    def test_set_executables(self, monkeypatch):
+        from distutils.cygwinccompiler import Mingw32CCompiler
+
+        monkeypatch.setenv('CC', 'cc')
+        monkeypatch.setenv('CXX', 'c++')
+
+        compiler = Mingw32CCompiler()
+
+        assert compiler.compiler == split_quoted('cc -O -Wall')
+        assert compiler.compiler_so == split_quoted('cc -shared -O -Wall')
+        assert compiler.compiler_cxx == split_quoted('c++ -O -Wall')
+        assert compiler.linker_exe == split_quoted('cc')
+        assert compiler.linker_so == split_quoted('cc -shared')
+
+    @pytest.mark.skipif(not is_mingw(), reason='not on mingw')
+    def test_runtime_library_dir_option(self):
+        from distutils.cygwinccompiler import Mingw32CCompiler
+
+        compiler = Mingw32CCompiler()
+        with pytest.raises(DistutilsPlatformError):
+            compiler.runtime_library_dir_option('/usr/lib')
+
+    @pytest.mark.skipif(not is_mingw(), reason='not on mingw')
+    def test_cygwincc_error(self, monkeypatch):
+        import distutils.cygwinccompiler
+
+        monkeypatch.setattr(distutils.cygwinccompiler, 'is_cygwincc', lambda _: True)
+
+        with pytest.raises(CCompilerError):
+            distutils.cygwinccompiler.Mingw32CCompiler()
+
+    @pytest.mark.skipif('sys.platform == "cygwin"')
+    def test_customize_compiler_with_msvc_python(self):
+        from distutils.cygwinccompiler import Mingw32CCompiler
+
+        # In case we have an MSVC Python build, but still want to use
+        # Mingw32CCompiler, then customize_compiler() shouldn't fail at least.
+        # https://github.com/pypa/setuptools/issues/4456
+        compiler = Mingw32CCompiler()
+        sysconfig.customize_compiler(compiler)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_modified.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_modified.py
new file mode 100644
index 00000000..e35cec2d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_modified.py
@@ -0,0 +1,126 @@
+"""Tests for distutils._modified."""
+
+import os
+import types
+from distutils._modified import newer, newer_group, newer_pairwise, newer_pairwise_group
+from distutils.errors import DistutilsFileError
+from distutils.tests import support
+
+import pytest
+
+
+class TestDepUtil(support.TempdirManager):
+    def test_newer(self):
+        tmpdir = self.mkdtemp()
+        new_file = os.path.join(tmpdir, 'new')
+        old_file = os.path.abspath(__file__)
+
+        # Raise DistutilsFileError if 'new_file' does not exist.
+        with pytest.raises(DistutilsFileError):
+            newer(new_file, old_file)
+
+        # Return true if 'new_file' exists and is more recently modified than
+        # 'old_file', or if 'new_file' exists and 'old_file' doesn't.
+        self.write_file(new_file)
+        assert newer(new_file, 'I_dont_exist')
+        assert newer(new_file, old_file)
+
+        # Return false if both exist and 'old_file' is the same age or younger
+        # than 'new_file'.
+        assert not newer(old_file, new_file)
+
+    def _setup_1234(self):
+        tmpdir = self.mkdtemp()
+        sources = os.path.join(tmpdir, 'sources')
+        targets = os.path.join(tmpdir, 'targets')
+        os.mkdir(sources)
+        os.mkdir(targets)
+        one = os.path.join(sources, 'one')
+        two = os.path.join(sources, 'two')
+        three = os.path.abspath(__file__)  # I am the old file
+        four = os.path.join(targets, 'four')
+        self.write_file(one)
+        self.write_file(two)
+        self.write_file(four)
+        return one, two, three, four
+
+    def test_newer_pairwise(self):
+        one, two, three, four = self._setup_1234()
+
+        assert newer_pairwise([one, two], [three, four]) == ([one], [three])
+
+    def test_newer_pairwise_mismatch(self):
+        one, two, three, four = self._setup_1234()
+
+        with pytest.raises(ValueError):
+            newer_pairwise([one], [three, four])
+
+        with pytest.raises(ValueError):
+            newer_pairwise([one, two], [three])
+
+    def test_newer_pairwise_empty(self):
+        assert newer_pairwise([], []) == ([], [])
+
+    def test_newer_pairwise_fresh(self):
+        one, two, three, four = self._setup_1234()
+
+        assert newer_pairwise([one, three], [two, four]) == ([], [])
+
+    def test_newer_group(self):
+        tmpdir = self.mkdtemp()
+        sources = os.path.join(tmpdir, 'sources')
+        os.mkdir(sources)
+        one = os.path.join(sources, 'one')
+        two = os.path.join(sources, 'two')
+        three = os.path.join(sources, 'three')
+        old_file = os.path.abspath(__file__)
+
+        # return true if 'old_file' is out-of-date with respect to any file
+        # listed in 'sources'.
+        self.write_file(one)
+        self.write_file(two)
+        self.write_file(three)
+        assert newer_group([one, two, three], old_file)
+        assert not newer_group([one, two, old_file], three)
+
+        # missing handling
+        os.remove(one)
+        with pytest.raises(OSError):
+            newer_group([one, two, old_file], three)
+
+        assert not newer_group([one, two, old_file], three, missing='ignore')
+
+        assert newer_group([one, two, old_file], three, missing='newer')
+
+
+@pytest.fixture
+def groups_target(tmp_path):
+    """
+    Set up some older sources, a target, and newer sources.
+
+    Returns a simple namespace with these values.
+    """
+    filenames = ['older.c', 'older.h', 'target.o', 'newer.c', 'newer.h']
+    paths = [tmp_path / name for name in filenames]
+
+    for mtime, path in enumerate(paths):
+        path.write_text('', encoding='utf-8')
+
+        # make sure modification times are sequential
+        os.utime(path, (mtime, mtime))
+
+    return types.SimpleNamespace(older=paths[:2], target=paths[2], newer=paths[3:])
+
+
+def test_newer_pairwise_group(groups_target):
+    older = newer_pairwise_group([groups_target.older], [groups_target.target])
+    newer = newer_pairwise_group([groups_target.newer], [groups_target.target])
+    assert older == ([], [])
+    assert newer == ([groups_target.newer], [groups_target.target])
+
+
+def test_newer_group_no_sources_no_target(tmp_path):
+    """
+    Consider no sources and no target "newer".
+    """
+    assert newer_group([], str(tmp_path / 'does-not-exist'))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_msvccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_msvccompiler.py
new file mode 100644
index 00000000..ceb15d3a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_msvccompiler.py
@@ -0,0 +1,137 @@
+"""Tests for distutils._msvccompiler."""
+
+import os
+import sys
+import sysconfig
+import threading
+import unittest.mock as mock
+from distutils import _msvccompiler
+from distutils.errors import DistutilsPlatformError
+from distutils.tests import support
+from distutils.util import get_platform
+
+import pytest
+
+needs_winreg = pytest.mark.skipif('not hasattr(_msvccompiler, "winreg")')
+
+
+class Testmsvccompiler(support.TempdirManager):
+    def test_no_compiler(self, monkeypatch):
+        # makes sure query_vcvarsall raises
+        # a DistutilsPlatformError if the compiler
+        # is not found
+        def _find_vcvarsall(plat_spec):
+            return None, None
+
+        monkeypatch.setattr(_msvccompiler, '_find_vcvarsall', _find_vcvarsall)
+
+        with pytest.raises(DistutilsPlatformError):
+            _msvccompiler._get_vc_env(
+                'wont find this version',
+            )
+
+    @pytest.mark.skipif(
+        not sysconfig.get_platform().startswith("win"),
+        reason="Only run test for non-mingw Windows platforms",
+    )
+    @pytest.mark.parametrize(
+        "plat_name, expected",
+        [
+            ("win-arm64", "win-arm64"),
+            ("win-amd64", "win-amd64"),
+            (None, get_platform()),
+        ],
+    )
+    def test_cross_platform_compilation_paths(self, monkeypatch, plat_name, expected):
+        """
+        Ensure a specified target platform is passed to _get_vcvars_spec.
+        """
+        compiler = _msvccompiler.MSVCCompiler()
+
+        def _get_vcvars_spec(host_platform, platform):
+            assert platform == expected
+
+        monkeypatch.setattr(_msvccompiler, '_get_vcvars_spec', _get_vcvars_spec)
+        compiler.initialize(plat_name)
+
+    @needs_winreg
+    def test_get_vc_env_unicode(self):
+        test_var = 'ṰḖṤṪ┅ṼẨṜ'
+        test_value = '₃⁴₅'
+
+        # Ensure we don't early exit from _get_vc_env
+        old_distutils_use_sdk = os.environ.pop('DISTUTILS_USE_SDK', None)
+        os.environ[test_var] = test_value
+        try:
+            env = _msvccompiler._get_vc_env('x86')
+            assert test_var.lower() in env
+            assert test_value == env[test_var.lower()]
+        finally:
+            os.environ.pop(test_var)
+            if old_distutils_use_sdk:
+                os.environ['DISTUTILS_USE_SDK'] = old_distutils_use_sdk
+
+    @needs_winreg
+    @pytest.mark.parametrize('ver', (2015, 2017))
+    def test_get_vc(self, ver):
+        # This function cannot be mocked, so pass if VC is found
+        # and skip otherwise.
+        lookup = getattr(_msvccompiler, f'_find_vc{ver}')
+        expected_version = {2015: 14, 2017: 15}[ver]
+        version, path = lookup()
+        if not version:
+            pytest.skip(f"VS {ver} is not installed")
+        assert version >= expected_version
+        assert os.path.isdir(path)
+
+
+class CheckThread(threading.Thread):
+    exc_info = None
+
+    def run(self):
+        try:
+            super().run()
+        except Exception:
+            self.exc_info = sys.exc_info()
+
+    def __bool__(self):
+        return not self.exc_info
+
+
+class TestSpawn:
+    def test_concurrent_safe(self):
+        """
+        Concurrent calls to spawn should have consistent results.
+        """
+        compiler = _msvccompiler.MSVCCompiler()
+        compiler._paths = "expected"
+        inner_cmd = 'import os; assert os.environ["PATH"] == "expected"'
+        command = [sys.executable, '-c', inner_cmd]
+
+        threads = [
+            CheckThread(target=compiler.spawn, args=[command]) for n in range(100)
+        ]
+        for thread in threads:
+            thread.start()
+        for thread in threads:
+            thread.join()
+        assert all(threads)
+
+    def test_concurrent_safe_fallback(self):
+        """
+        If CCompiler.spawn has been monkey-patched without support
+        for an env, it should still execute.
+        """
+        from distutils import ccompiler
+
+        compiler = _msvccompiler.MSVCCompiler()
+        compiler._paths = "expected"
+
+        def CCompiler_spawn(self, cmd):
+            "A spawn without an env argument."
+            assert os.environ["PATH"] == "expected"
+
+        with mock.patch.object(ccompiler.CCompiler, 'spawn', CCompiler_spawn):
+            compiler.spawn(["n/a"])
+
+        assert os.environ.get("PATH") != "expected"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sdist.py
new file mode 100644
index 00000000..5aca43e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sdist.py
@@ -0,0 +1,470 @@
+"""Tests for distutils.command.sdist."""
+
+import os
+import pathlib
+import shutil  # noqa: F401
+import tarfile
+import zipfile
+from distutils.archive_util import ARCHIVE_FORMATS
+from distutils.command.sdist import sdist, show_formats
+from distutils.core import Distribution
+from distutils.errors import DistutilsOptionError
+from distutils.filelist import FileList
+from os.path import join
+from textwrap import dedent
+
+import jaraco.path
+import path
+import pytest
+from more_itertools import ilen
+
+from . import support
+from .unix_compat import grp, pwd, require_uid_0, require_unix_id
+
+SETUP_PY = """
+from distutils.core import setup
+import somecode
+
+setup(name='fake')
+"""
+
+MANIFEST = """\
+# file GENERATED by distutils, do NOT edit
+README
+buildout.cfg
+inroot.txt
+setup.py
+data%(sep)sdata.dt
+scripts%(sep)sscript.py
+some%(sep)sfile.txt
+some%(sep)sother_file.txt
+somecode%(sep)s__init__.py
+somecode%(sep)sdoc.dat
+somecode%(sep)sdoc.txt
+"""
+
+
+@pytest.fixture(autouse=True)
+def project_dir(request, distutils_managed_tempdir):
+    self = request.instance
+    self.tmp_dir = self.mkdtemp()
+    jaraco.path.build(
+        {
+            'somecode': {
+                '__init__.py': '#',
+            },
+            'README': 'xxx',
+            'setup.py': SETUP_PY,
+        },
+        self.tmp_dir,
+    )
+    with path.Path(self.tmp_dir):
+        yield
+
+
+def clean_lines(filepath):
+    with pathlib.Path(filepath).open(encoding='utf-8') as f:
+        yield from filter(None, map(str.strip, f))
+
+
+class TestSDist(support.TempdirManager):
+    def get_cmd(self, metadata=None):
+        """Returns a cmd"""
+        if metadata is None:
+            metadata = {
+                'name': 'ns.fake--pkg',
+                'version': '1.0',
+                'url': 'xxx',
+                'author': 'xxx',
+                'author_email': 'xxx',
+            }
+        dist = Distribution(metadata)
+        dist.script_name = 'setup.py'
+        dist.packages = ['somecode']
+        dist.include_package_data = True
+        cmd = sdist(dist)
+        cmd.dist_dir = 'dist'
+        return dist, cmd
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_prune_file_list(self):
+        # this test creates a project with some VCS dirs and an NFS rename
+        # file, then launches sdist to check they get pruned on all systems
+
+        # creating VCS directories with some files in them
+        os.mkdir(join(self.tmp_dir, 'somecode', '.svn'))
+        self.write_file((self.tmp_dir, 'somecode', '.svn', 'ok.py'), 'xxx')
+
+        os.mkdir(join(self.tmp_dir, 'somecode', '.hg'))
+        self.write_file((self.tmp_dir, 'somecode', '.hg', 'ok'), 'xxx')
+
+        os.mkdir(join(self.tmp_dir, 'somecode', '.git'))
+        self.write_file((self.tmp_dir, 'somecode', '.git', 'ok'), 'xxx')
+
+        self.write_file((self.tmp_dir, 'somecode', '.nfs0001'), 'xxx')
+
+        # now building a sdist
+        dist, cmd = self.get_cmd()
+
+        # zip is available universally
+        # (tar might not be installed under win32)
+        cmd.formats = ['zip']
+
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # now let's check what we have
+        dist_folder = join(self.tmp_dir, 'dist')
+        files = os.listdir(dist_folder)
+        assert files == ['ns_fake_pkg-1.0.zip']
+
+        zip_file = zipfile.ZipFile(join(dist_folder, 'ns_fake_pkg-1.0.zip'))
+        try:
+            content = zip_file.namelist()
+        finally:
+            zip_file.close()
+
+        # making sure everything has been pruned correctly
+        expected = [
+            '',
+            'PKG-INFO',
+            'README',
+            'setup.py',
+            'somecode/',
+            'somecode/__init__.py',
+        ]
+        assert sorted(content) == ['ns_fake_pkg-1.0/' + x for x in expected]
+
+    @pytest.mark.usefixtures('needs_zlib')
+    @pytest.mark.skipif("not shutil.which('tar')")
+    @pytest.mark.skipif("not shutil.which('gzip')")
+    def test_make_distribution(self):
+        # now building a sdist
+        dist, cmd = self.get_cmd()
+
+        # creating a gztar then a tar
+        cmd.formats = ['gztar', 'tar']
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # making sure we have two files
+        dist_folder = join(self.tmp_dir, 'dist')
+        result = os.listdir(dist_folder)
+        result.sort()
+        assert result == ['ns_fake_pkg-1.0.tar', 'ns_fake_pkg-1.0.tar.gz']
+
+        os.remove(join(dist_folder, 'ns_fake_pkg-1.0.tar'))
+        os.remove(join(dist_folder, 'ns_fake_pkg-1.0.tar.gz'))
+
+        # now trying a tar then a gztar
+        cmd.formats = ['tar', 'gztar']
+
+        cmd.ensure_finalized()
+        cmd.run()
+
+        result = os.listdir(dist_folder)
+        result.sort()
+        assert result == ['ns_fake_pkg-1.0.tar', 'ns_fake_pkg-1.0.tar.gz']
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_add_defaults(self):
+        # https://bugs.python.org/issue2279
+
+        # add_default should also include
+        # data_files and package_data
+        dist, cmd = self.get_cmd()
+
+        # filling data_files by pointing files
+        # in package_data
+        dist.package_data = {'': ['*.cfg', '*.dat'], 'somecode': ['*.txt']}
+        self.write_file((self.tmp_dir, 'somecode', 'doc.txt'), '#')
+        self.write_file((self.tmp_dir, 'somecode', 'doc.dat'), '#')
+
+        # adding some data in data_files
+        data_dir = join(self.tmp_dir, 'data')
+        os.mkdir(data_dir)
+        self.write_file((data_dir, 'data.dt'), '#')
+        some_dir = join(self.tmp_dir, 'some')
+        os.mkdir(some_dir)
+        # make sure VCS directories are pruned (#14004)
+        hg_dir = join(self.tmp_dir, '.hg')
+        os.mkdir(hg_dir)
+        self.write_file((hg_dir, 'last-message.txt'), '#')
+        # a buggy regex used to prevent this from working on windows (#6884)
+        self.write_file((self.tmp_dir, 'buildout.cfg'), '#')
+        self.write_file((self.tmp_dir, 'inroot.txt'), '#')
+        self.write_file((some_dir, 'file.txt'), '#')
+        self.write_file((some_dir, 'other_file.txt'), '#')
+
+        dist.data_files = [
+            ('data', ['data/data.dt', 'buildout.cfg', 'inroot.txt', 'notexisting']),
+            'some/file.txt',
+            'some/other_file.txt',
+        ]
+
+        # adding a script
+        script_dir = join(self.tmp_dir, 'scripts')
+        os.mkdir(script_dir)
+        self.write_file((script_dir, 'script.py'), '#')
+        dist.scripts = [join('scripts', 'script.py')]
+
+        cmd.formats = ['zip']
+        cmd.use_defaults = True
+
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # now let's check what we have
+        dist_folder = join(self.tmp_dir, 'dist')
+        files = os.listdir(dist_folder)
+        assert files == ['ns_fake_pkg-1.0.zip']
+
+        zip_file = zipfile.ZipFile(join(dist_folder, 'ns_fake_pkg-1.0.zip'))
+        try:
+            content = zip_file.namelist()
+        finally:
+            zip_file.close()
+
+        # making sure everything was added
+        expected = [
+            '',
+            'PKG-INFO',
+            'README',
+            'buildout.cfg',
+            'data/',
+            'data/data.dt',
+            'inroot.txt',
+            'scripts/',
+            'scripts/script.py',
+            'setup.py',
+            'some/',
+            'some/file.txt',
+            'some/other_file.txt',
+            'somecode/',
+            'somecode/__init__.py',
+            'somecode/doc.dat',
+            'somecode/doc.txt',
+        ]
+        assert sorted(content) == ['ns_fake_pkg-1.0/' + x for x in expected]
+
+        # checking the MANIFEST
+        manifest = pathlib.Path(self.tmp_dir, 'MANIFEST').read_text(encoding='utf-8')
+        assert manifest == MANIFEST % {'sep': os.sep}
+
+    @staticmethod
+    def warnings(messages, prefix='warning: '):
+        return [msg for msg in messages if msg.startswith(prefix)]
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_metadata_check_option(self, caplog):
+        # testing the `medata-check` option
+        dist, cmd = self.get_cmd(metadata={})
+
+        # this should raise some warnings !
+        # with the `check` subcommand
+        cmd.ensure_finalized()
+        cmd.run()
+        assert len(self.warnings(caplog.messages, 'warning: check: ')) == 1
+
+        # trying with a complete set of metadata
+        caplog.clear()
+        dist, cmd = self.get_cmd()
+        cmd.ensure_finalized()
+        cmd.metadata_check = 0
+        cmd.run()
+        assert len(self.warnings(caplog.messages, 'warning: check: ')) == 0
+
+    def test_show_formats(self, capsys):
+        show_formats()
+
+        # the output should be a header line + one line per format
+        num_formats = len(ARCHIVE_FORMATS.keys())
+        output = [
+            line
+            for line in capsys.readouterr().out.split('\n')
+            if line.strip().startswith('--formats=')
+        ]
+        assert len(output) == num_formats
+
+    def test_finalize_options(self):
+        dist, cmd = self.get_cmd()
+        cmd.finalize_options()
+
+        # default options set by finalize
+        assert cmd.manifest == 'MANIFEST'
+        assert cmd.template == 'MANIFEST.in'
+        assert cmd.dist_dir == 'dist'
+
+        # formats has to be a string splitable on (' ', ',') or
+        # a stringlist
+        cmd.formats = 1
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+        cmd.formats = ['zip']
+        cmd.finalize_options()
+
+        # formats has to be known
+        cmd.formats = 'supazipa'
+        with pytest.raises(DistutilsOptionError):
+            cmd.finalize_options()
+
+    # the following tests make sure there is a nice error message instead
+    # of a traceback when parsing an invalid manifest template
+
+    def _check_template(self, content, caplog):
+        dist, cmd = self.get_cmd()
+        os.chdir(self.tmp_dir)
+        self.write_file('MANIFEST.in', content)
+        cmd.ensure_finalized()
+        cmd.filelist = FileList()
+        cmd.read_template()
+        assert len(self.warnings(caplog.messages)) == 1
+
+    def test_invalid_template_unknown_command(self, caplog):
+        self._check_template('taunt knights *', caplog)
+
+    def test_invalid_template_wrong_arguments(self, caplog):
+        # this manifest command takes one argument
+        self._check_template('prune', caplog)
+
+    @pytest.mark.skipif("platform.system() != 'Windows'")
+    def test_invalid_template_wrong_path(self, caplog):
+        # on Windows, trailing slashes are not allowed
+        # this used to crash instead of raising a warning: #8286
+        self._check_template('include examples/', caplog)
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_get_file_list(self):
+        # make sure MANIFEST is recalculated
+        dist, cmd = self.get_cmd()
+
+        # filling data_files by pointing files in package_data
+        dist.package_data = {'somecode': ['*.txt']}
+        self.write_file((self.tmp_dir, 'somecode', 'doc.txt'), '#')
+        cmd.formats = ['gztar']
+        cmd.ensure_finalized()
+        cmd.run()
+
+        assert ilen(clean_lines(cmd.manifest)) == 5
+
+        # adding a file
+        self.write_file((self.tmp_dir, 'somecode', 'doc2.txt'), '#')
+
+        # make sure build_py is reinitialized, like a fresh run
+        build_py = dist.get_command_obj('build_py')
+        build_py.finalized = False
+        build_py.ensure_finalized()
+
+        cmd.run()
+
+        manifest2 = list(clean_lines(cmd.manifest))
+
+        # do we have the new file in MANIFEST ?
+        assert len(manifest2) == 6
+        assert 'doc2.txt' in manifest2[-1]
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_manifest_marker(self):
+        # check that autogenerated MANIFESTs have a marker
+        dist, cmd = self.get_cmd()
+        cmd.ensure_finalized()
+        cmd.run()
+
+        assert (
+            next(clean_lines(cmd.manifest))
+            == '# file GENERATED by distutils, do NOT edit'
+        )
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_manifest_comments(self):
+        # make sure comments don't cause exceptions or wrong includes
+        contents = dedent(
+            """\
+            # bad.py
+            #bad.py
+            good.py
+            """
+        )
+        dist, cmd = self.get_cmd()
+        cmd.ensure_finalized()
+        self.write_file((self.tmp_dir, cmd.manifest), contents)
+        self.write_file((self.tmp_dir, 'good.py'), '# pick me!')
+        self.write_file((self.tmp_dir, 'bad.py'), "# don't pick me!")
+        self.write_file((self.tmp_dir, '#bad.py'), "# don't pick me!")
+        cmd.run()
+        assert cmd.filelist.files == ['good.py']
+
+    @pytest.mark.usefixtures('needs_zlib')
+    def test_manual_manifest(self):
+        # check that a MANIFEST without a marker is left alone
+        dist, cmd = self.get_cmd()
+        cmd.formats = ['gztar']
+        cmd.ensure_finalized()
+        self.write_file((self.tmp_dir, cmd.manifest), 'README.manual')
+        self.write_file(
+            (self.tmp_dir, 'README.manual'),
+            'This project maintains its MANIFEST file itself.',
+        )
+        cmd.run()
+        assert cmd.filelist.files == ['README.manual']
+
+        assert list(clean_lines(cmd.manifest)) == ['README.manual']
+
+        archive_name = join(self.tmp_dir, 'dist', 'ns_fake_pkg-1.0.tar.gz')
+        archive = tarfile.open(archive_name)
+        try:
+            filenames = [tarinfo.name for tarinfo in archive]
+        finally:
+            archive.close()
+        assert sorted(filenames) == [
+            'ns_fake_pkg-1.0',
+            'ns_fake_pkg-1.0/PKG-INFO',
+            'ns_fake_pkg-1.0/README.manual',
+        ]
+
+    @pytest.mark.usefixtures('needs_zlib')
+    @require_unix_id
+    @require_uid_0
+    @pytest.mark.skipif("not shutil.which('tar')")
+    @pytest.mark.skipif("not shutil.which('gzip')")
+    def test_make_distribution_owner_group(self):
+        # now building a sdist
+        dist, cmd = self.get_cmd()
+
+        # creating a gztar and specifying the owner+group
+        cmd.formats = ['gztar']
+        cmd.owner = pwd.getpwuid(0)[0]
+        cmd.group = grp.getgrgid(0)[0]
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # making sure we have the good rights
+        archive_name = join(self.tmp_dir, 'dist', 'ns_fake_pkg-1.0.tar.gz')
+        archive = tarfile.open(archive_name)
+        try:
+            for member in archive.getmembers():
+                assert member.uid == 0
+                assert member.gid == 0
+        finally:
+            archive.close()
+
+        # building a sdist again
+        dist, cmd = self.get_cmd()
+
+        # creating a gztar
+        cmd.formats = ['gztar']
+        cmd.ensure_finalized()
+        cmd.run()
+
+        # making sure we have the good rights
+        archive_name = join(self.tmp_dir, 'dist', 'ns_fake_pkg-1.0.tar.gz')
+        archive = tarfile.open(archive_name)
+
+        # note that we are not testing the group ownership here
+        # because, depending on the platforms and the container
+        # rights (see #7408)
+        try:
+            for member in archive.getmembers():
+                assert member.uid == os.getuid()
+        finally:
+            archive.close()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_spawn.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_spawn.py
new file mode 100644
index 00000000..3b9fc926
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_spawn.py
@@ -0,0 +1,141 @@
+"""Tests for distutils.spawn."""
+
+import os
+import stat
+import sys
+import unittest.mock as mock
+from distutils.errors import DistutilsExecError
+from distutils.spawn import find_executable, spawn
+from distutils.tests import support
+
+import path
+import pytest
+from test.support import unix_shell
+
+from .compat import py39 as os_helper
+
+
+class TestSpawn(support.TempdirManager):
+    @pytest.mark.skipif("os.name not in ('nt', 'posix')")
+    def test_spawn(self):
+        tmpdir = self.mkdtemp()
+
+        # creating something executable
+        # through the shell that returns 1
+        if sys.platform != 'win32':
+            exe = os.path.join(tmpdir, 'foo.sh')
+            self.write_file(exe, f'#!{unix_shell}\nexit 1')
+        else:
+            exe = os.path.join(tmpdir, 'foo.bat')
+            self.write_file(exe, 'exit 1')
+
+        os.chmod(exe, 0o777)
+        with pytest.raises(DistutilsExecError):
+            spawn([exe])
+
+        # now something that works
+        if sys.platform != 'win32':
+            exe = os.path.join(tmpdir, 'foo.sh')
+            self.write_file(exe, f'#!{unix_shell}\nexit 0')
+        else:
+            exe = os.path.join(tmpdir, 'foo.bat')
+            self.write_file(exe, 'exit 0')
+
+        os.chmod(exe, 0o777)
+        spawn([exe])  # should work without any error
+
+    def test_find_executable(self, tmp_path):
+        program_path = self._make_executable(tmp_path, '.exe')
+        program = program_path.name
+        program_noeext = program_path.with_suffix('').name
+        filename = str(program_path)
+        tmp_dir = path.Path(tmp_path)
+
+        # test path parameter
+        rv = find_executable(program, path=tmp_dir)
+        assert rv == filename
+
+        if sys.platform == 'win32':
+            # test without ".exe" extension
+            rv = find_executable(program_noeext, path=tmp_dir)
+            assert rv == filename
+
+        # test find in the current directory
+        with tmp_dir:
+            rv = find_executable(program)
+            assert rv == program
+
+        # test non-existent program
+        dont_exist_program = "dontexist_" + program
+        rv = find_executable(dont_exist_program, path=tmp_dir)
+        assert rv is None
+
+        # PATH='': no match, except in the current directory
+        with os_helper.EnvironmentVarGuard() as env:
+            env['PATH'] = ''
+            with (
+                mock.patch(
+                    'distutils.spawn.os.confstr', return_value=tmp_dir, create=True
+                ),
+                mock.patch('distutils.spawn.os.defpath', tmp_dir),
+            ):
+                rv = find_executable(program)
+                assert rv is None
+
+                # look in current directory
+                with tmp_dir:
+                    rv = find_executable(program)
+                    assert rv == program
+
+        # PATH=':': explicitly looks in the current directory
+        with os_helper.EnvironmentVarGuard() as env:
+            env['PATH'] = os.pathsep
+            with (
+                mock.patch('distutils.spawn.os.confstr', return_value='', create=True),
+                mock.patch('distutils.spawn.os.defpath', ''),
+            ):
+                rv = find_executable(program)
+                assert rv is None
+
+                # look in current directory
+                with tmp_dir:
+                    rv = find_executable(program)
+                    assert rv == program
+
+        # missing PATH: test os.confstr("CS_PATH") and os.defpath
+        with os_helper.EnvironmentVarGuard() as env:
+            env.pop('PATH', None)
+
+            # without confstr
+            with (
+                mock.patch(
+                    'distutils.spawn.os.confstr', side_effect=ValueError, create=True
+                ),
+                mock.patch('distutils.spawn.os.defpath', tmp_dir),
+            ):
+                rv = find_executable(program)
+                assert rv == filename
+
+            # with confstr
+            with (
+                mock.patch(
+                    'distutils.spawn.os.confstr', return_value=tmp_dir, create=True
+                ),
+                mock.patch('distutils.spawn.os.defpath', ''),
+            ):
+                rv = find_executable(program)
+                assert rv == filename
+
+    @staticmethod
+    def _make_executable(tmp_path, ext):
+        # Give the temporary program a suffix regardless of platform.
+        # It's needed on Windows and not harmful on others.
+        program = tmp_path.joinpath('program').with_suffix(ext)
+        program.write_text("", encoding='utf-8')
+        program.chmod(stat.S_IXUSR)
+        return program
+
+    def test_spawn_missing_exe(self):
+        with pytest.raises(DistutilsExecError) as ctx:
+            spawn(['does-not-exist'])
+        assert "command 'does-not-exist' failed" in str(ctx.value)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sysconfig.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sysconfig.py
new file mode 100644
index 00000000..43d77c23
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_sysconfig.py
@@ -0,0 +1,319 @@
+"""Tests for distutils.sysconfig."""
+
+import contextlib
+import distutils
+import os
+import pathlib
+import subprocess
+import sys
+from distutils import sysconfig
+from distutils.ccompiler import new_compiler  # noqa: F401
+from distutils.unixccompiler import UnixCCompiler
+
+import jaraco.envs
+import path
+import pytest
+from jaraco.text import trim
+from test.support import swap_item
+
+
+def _gen_makefile(root, contents):
+    jaraco.path.build({'Makefile': trim(contents)}, root)
+    return root / 'Makefile'
+
+
+@pytest.mark.usefixtures('save_env')
+class TestSysconfig:
+    def test_get_config_h_filename(self):
+        config_h = sysconfig.get_config_h_filename()
+        assert os.path.isfile(config_h)
+
+    @pytest.mark.skipif("platform.system() == 'Windows'")
+    @pytest.mark.skipif("sys.implementation.name != 'cpython'")
+    def test_get_makefile_filename(self):
+        makefile = sysconfig.get_makefile_filename()
+        assert os.path.isfile(makefile)
+
+    def test_get_python_lib(self, tmp_path):
+        assert sysconfig.get_python_lib() != sysconfig.get_python_lib(prefix=tmp_path)
+
+    def test_get_config_vars(self):
+        cvars = sysconfig.get_config_vars()
+        assert isinstance(cvars, dict)
+        assert cvars
+
+    @pytest.mark.skipif('sysconfig.IS_PYPY')
+    @pytest.mark.skipif('sysconfig.python_build')
+    @pytest.mark.xfail('platform.system() == "Windows"')
+    def test_srcdir_simple(self):
+        # See #15364.
+        srcdir = pathlib.Path(sysconfig.get_config_var('srcdir'))
+
+        assert srcdir.absolute()
+        assert srcdir.is_dir()
+
+        makefile = pathlib.Path(sysconfig.get_makefile_filename())
+        assert makefile.parent.samefile(srcdir)
+
+    @pytest.mark.skipif('sysconfig.IS_PYPY')
+    @pytest.mark.skipif('not sysconfig.python_build')
+    def test_srcdir_python_build(self):
+        # See #15364.
+        srcdir = pathlib.Path(sysconfig.get_config_var('srcdir'))
+
+        # The python executable has not been installed so srcdir
+        # should be a full source checkout.
+        Python_h = srcdir.joinpath('Include', 'Python.h')
+        assert Python_h.is_file()
+        assert sysconfig._is_python_source_dir(srcdir)
+        assert sysconfig._is_python_source_dir(str(srcdir))
+
+    def test_srcdir_independent_of_cwd(self):
+        """
+        srcdir should be independent of the current working directory
+        """
+        # See #15364.
+        srcdir = sysconfig.get_config_var('srcdir')
+        with path.Path('..'):
+            srcdir2 = sysconfig.get_config_var('srcdir')
+        assert srcdir == srcdir2
+
+    def customize_compiler(self):
+        # make sure AR gets caught
+        class compiler:
+            compiler_type = 'unix'
+            executables = UnixCCompiler.executables
+
+            def __init__(self):
+                self.exes = {}
+
+            def set_executables(self, **kw):
+                for k, v in kw.items():
+                    self.exes[k] = v
+
+        sysconfig_vars = {
+            'AR': 'sc_ar',
+            'CC': 'sc_cc',
+            'CXX': 'sc_cxx',
+            'ARFLAGS': '--sc-arflags',
+            'CFLAGS': '--sc-cflags',
+            'CCSHARED': '--sc-ccshared',
+            'LDSHARED': 'sc_ldshared',
+            'SHLIB_SUFFIX': 'sc_shutil_suffix',
+        }
+
+        comp = compiler()
+        with contextlib.ExitStack() as cm:
+            for key, value in sysconfig_vars.items():
+                cm.enter_context(swap_item(sysconfig._config_vars, key, value))
+            sysconfig.customize_compiler(comp)
+
+        return comp
+
+    @pytest.mark.skipif("not isinstance(new_compiler(), UnixCCompiler)")
+    @pytest.mark.usefixtures('disable_macos_customization')
+    def test_customize_compiler(self):
+        # Make sure that sysconfig._config_vars is initialized
+        sysconfig.get_config_vars()
+
+        os.environ['AR'] = 'env_ar'
+        os.environ['CC'] = 'env_cc'
+        os.environ['CPP'] = 'env_cpp'
+        os.environ['CXX'] = 'env_cxx --env-cxx-flags'
+        os.environ['LDSHARED'] = 'env_ldshared'
+        os.environ['LDFLAGS'] = '--env-ldflags'
+        os.environ['ARFLAGS'] = '--env-arflags'
+        os.environ['CFLAGS'] = '--env-cflags'
+        os.environ['CPPFLAGS'] = '--env-cppflags'
+        os.environ['RANLIB'] = 'env_ranlib'
+
+        comp = self.customize_compiler()
+        assert comp.exes['archiver'] == 'env_ar --env-arflags'
+        assert comp.exes['preprocessor'] == 'env_cpp --env-cppflags'
+        assert comp.exes['compiler'] == 'env_cc --env-cflags --env-cppflags'
+        assert comp.exes['compiler_so'] == (
+            'env_cc --env-cflags --env-cppflags --sc-ccshared'
+        )
+        assert (
+            comp.exes['compiler_cxx']
+            == 'env_cxx --env-cxx-flags --sc-cflags --env-cppflags'
+        )
+        assert comp.exes['linker_exe'] == 'env_cc'
+        assert comp.exes['linker_so'] == (
+            'env_ldshared --env-ldflags --env-cflags --env-cppflags'
+        )
+        assert comp.shared_lib_extension == 'sc_shutil_suffix'
+
+        if sys.platform == "darwin":
+            assert comp.exes['ranlib'] == 'env_ranlib'
+        else:
+            assert 'ranlib' not in comp.exes
+
+        del os.environ['AR']
+        del os.environ['CC']
+        del os.environ['CPP']
+        del os.environ['CXX']
+        del os.environ['LDSHARED']
+        del os.environ['LDFLAGS']
+        del os.environ['ARFLAGS']
+        del os.environ['CFLAGS']
+        del os.environ['CPPFLAGS']
+        del os.environ['RANLIB']
+
+        comp = self.customize_compiler()
+        assert comp.exes['archiver'] == 'sc_ar --sc-arflags'
+        assert comp.exes['preprocessor'] == 'sc_cc -E'
+        assert comp.exes['compiler'] == 'sc_cc --sc-cflags'
+        assert comp.exes['compiler_so'] == 'sc_cc --sc-cflags --sc-ccshared'
+        assert comp.exes['compiler_cxx'] == 'sc_cxx --sc-cflags'
+        assert comp.exes['linker_exe'] == 'sc_cc'
+        assert comp.exes['linker_so'] == 'sc_ldshared'
+        assert comp.shared_lib_extension == 'sc_shutil_suffix'
+        assert 'ranlib' not in comp.exes
+
+    def test_parse_makefile_base(self, tmp_path):
+        makefile = _gen_makefile(
+            tmp_path,
+            """
+            CONFIG_ARGS=  '--arg1=optarg1' 'ENV=LIB'
+            VAR=$OTHER
+            OTHER=foo
+            """,
+        )
+        d = sysconfig.parse_makefile(makefile)
+        assert d == {'CONFIG_ARGS': "'--arg1=optarg1' 'ENV=LIB'", 'OTHER': 'foo'}
+
+    def test_parse_makefile_literal_dollar(self, tmp_path):
+        makefile = _gen_makefile(
+            tmp_path,
+            """
+            CONFIG_ARGS=  '--arg1=optarg1' 'ENV=\\$$LIB'
+            VAR=$OTHER
+            OTHER=foo
+            """,
+        )
+        d = sysconfig.parse_makefile(makefile)
+        assert d == {'CONFIG_ARGS': r"'--arg1=optarg1' 'ENV=\$LIB'", 'OTHER': 'foo'}
+
+    def test_sysconfig_module(self):
+        import sysconfig as global_sysconfig
+
+        assert global_sysconfig.get_config_var('CFLAGS') == sysconfig.get_config_var(
+            'CFLAGS'
+        )
+        assert global_sysconfig.get_config_var('LDFLAGS') == sysconfig.get_config_var(
+            'LDFLAGS'
+        )
+
+    # On macOS, binary installers support extension module building on
+    # various levels of the operating system with differing Xcode
+    # configurations, requiring customization of some of the
+    # compiler configuration directives to suit the environment on
+    # the installed machine. Some of these customizations may require
+    # running external programs and are thus deferred until needed by
+    # the first extension module build. Only
+    # the Distutils version of sysconfig is used for extension module
+    # builds, which happens earlier in the Distutils tests. This may
+    # cause the following tests to fail since no tests have caused
+    # the global version of sysconfig to call the customization yet.
+    # The solution for now is to simply skip this test in this case.
+    # The longer-term solution is to only have one version of sysconfig.
+    @pytest.mark.skipif("sysconfig.get_config_var('CUSTOMIZED_OSX_COMPILER')")
+    def test_sysconfig_compiler_vars(self):
+        import sysconfig as global_sysconfig
+
+        if sysconfig.get_config_var('CUSTOMIZED_OSX_COMPILER'):
+            pytest.skip('compiler flags customized')
+        assert global_sysconfig.get_config_var('LDSHARED') == sysconfig.get_config_var(
+            'LDSHARED'
+        )
+        assert global_sysconfig.get_config_var('CC') == sysconfig.get_config_var('CC')
+
+    @pytest.mark.skipif("not sysconfig.get_config_var('EXT_SUFFIX')")
+    def test_SO_deprecation(self):
+        with pytest.warns(DeprecationWarning):
+            sysconfig.get_config_var('SO')
+
+    def test_customize_compiler_before_get_config_vars(self, tmp_path):
+        # Issue #21923: test that a Distribution compiler
+        # instance can be called without an explicit call to
+        # get_config_vars().
+        jaraco.path.build(
+            {
+                'file': trim("""
+                    from distutils.core import Distribution
+                    config = Distribution().get_command_obj('config')
+                    # try_compile may pass or it may fail if no compiler
+                    # is found but it should not raise an exception.
+                    rc = config.try_compile('int x;')
+                    """)
+            },
+            tmp_path,
+        )
+        p = subprocess.Popen(
+            [sys.executable, tmp_path / 'file'],
+            stdout=subprocess.PIPE,
+            stderr=subprocess.STDOUT,
+            universal_newlines=True,
+            encoding='utf-8',
+        )
+        outs, errs = p.communicate()
+        assert 0 == p.returncode, "Subprocess failed: " + outs
+
+    def test_parse_config_h(self):
+        config_h = sysconfig.get_config_h_filename()
+        input = {}
+        with open(config_h, encoding="utf-8") as f:
+            result = sysconfig.parse_config_h(f, g=input)
+        assert input is result
+        with open(config_h, encoding="utf-8") as f:
+            result = sysconfig.parse_config_h(f)
+        assert isinstance(result, dict)
+
+    @pytest.mark.skipif("platform.system() != 'Windows'")
+    @pytest.mark.skipif("sys.implementation.name != 'cpython'")
+    def test_win_ext_suffix(self):
+        assert sysconfig.get_config_var("EXT_SUFFIX").endswith(".pyd")
+        assert sysconfig.get_config_var("EXT_SUFFIX") != ".pyd"
+
+    @pytest.mark.skipif("platform.system() != 'Windows'")
+    @pytest.mark.skipif("sys.implementation.name != 'cpython'")
+    @pytest.mark.skipif(
+        '\\PCbuild\\'.casefold() not in sys.executable.casefold(),
+        reason='Need sys.executable to be in a source tree',
+    )
+    def test_win_build_venv_from_source_tree(self, tmp_path):
+        """Ensure distutils.sysconfig detects venvs from source tree builds."""
+        env = jaraco.envs.VEnv()
+        env.create_opts = env.clean_opts
+        env.root = tmp_path
+        env.ensure_env()
+        cmd = [
+            env.exe(),
+            "-c",
+            "import distutils.sysconfig; print(distutils.sysconfig.python_build)",
+        ]
+        distutils_path = os.path.dirname(os.path.dirname(distutils.__file__))
+        out = subprocess.check_output(
+            cmd, env={**os.environ, "PYTHONPATH": distutils_path}
+        )
+        assert out == "True"
+
+    def test_get_python_inc_missing_config_dir(self, monkeypatch):
+        """
+        In portable Python installations, the sysconfig will be broken,
+        pointing to the directories where the installation was built and
+        not where it currently is. In this case, ensure that the missing
+        directory isn't used for get_python_inc.
+
+        See pypa/distutils#178.
+        """
+
+        def override(name):
+            if name == 'INCLUDEPY':
+                return '/does-not-exist'
+            return sysconfig.get_config_var(name)
+
+        monkeypatch.setattr(sysconfig, 'get_config_var', override)
+
+        assert os.path.exists(sysconfig.get_python_inc())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_text_file.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_text_file.py
new file mode 100644
index 00000000..f5111565
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_text_file.py
@@ -0,0 +1,127 @@
+"""Tests for distutils.text_file."""
+
+from distutils.tests import support
+from distutils.text_file import TextFile
+
+import jaraco.path
+import path
+
+TEST_DATA = """# test file
+
+line 3 \\
+# intervening comment
+  continues on next line
+"""
+
+
+class TestTextFile(support.TempdirManager):
+    def test_class(self):
+        # old tests moved from text_file.__main__
+        # so they are really called by the buildbots
+
+        # result 1: no fancy options
+        result1 = [
+            '# test file\n',
+            '\n',
+            'line 3 \\\n',
+            '# intervening comment\n',
+            '  continues on next line\n',
+        ]
+
+        # result 2: just strip comments
+        result2 = ["\n", "line 3 \\\n", "  continues on next line\n"]
+
+        # result 3: just strip blank lines
+        result3 = [
+            "# test file\n",
+            "line 3 \\\n",
+            "# intervening comment\n",
+            "  continues on next line\n",
+        ]
+
+        # result 4: default, strip comments, blank lines,
+        # and trailing whitespace
+        result4 = ["line 3 \\", "  continues on next line"]
+
+        # result 5: strip comments and blanks, plus join lines (but don't
+        # "collapse" joined lines
+        result5 = ["line 3   continues on next line"]
+
+        # result 6: strip comments and blanks, plus join lines (and
+        # "collapse" joined lines
+        result6 = ["line 3 continues on next line"]
+
+        def test_input(count, description, file, expected_result):
+            result = file.readlines()
+            assert result == expected_result
+
+        tmp_path = path.Path(self.mkdtemp())
+        filename = tmp_path / 'test.txt'
+        jaraco.path.build({filename.name: TEST_DATA}, tmp_path)
+
+        in_file = TextFile(
+            filename,
+            strip_comments=False,
+            skip_blanks=False,
+            lstrip_ws=False,
+            rstrip_ws=False,
+        )
+        try:
+            test_input(1, "no processing", in_file, result1)
+        finally:
+            in_file.close()
+
+        in_file = TextFile(
+            filename,
+            strip_comments=True,
+            skip_blanks=False,
+            lstrip_ws=False,
+            rstrip_ws=False,
+        )
+        try:
+            test_input(2, "strip comments", in_file, result2)
+        finally:
+            in_file.close()
+
+        in_file = TextFile(
+            filename,
+            strip_comments=False,
+            skip_blanks=True,
+            lstrip_ws=False,
+            rstrip_ws=False,
+        )
+        try:
+            test_input(3, "strip blanks", in_file, result3)
+        finally:
+            in_file.close()
+
+        in_file = TextFile(filename)
+        try:
+            test_input(4, "default processing", in_file, result4)
+        finally:
+            in_file.close()
+
+        in_file = TextFile(
+            filename,
+            strip_comments=True,
+            skip_blanks=True,
+            join_lines=True,
+            rstrip_ws=True,
+        )
+        try:
+            test_input(5, "join lines without collapsing", in_file, result5)
+        finally:
+            in_file.close()
+
+        in_file = TextFile(
+            filename,
+            strip_comments=True,
+            skip_blanks=True,
+            join_lines=True,
+            rstrip_ws=True,
+            collapse_join=True,
+        )
+        try:
+            test_input(6, "join lines with collapsing", in_file, result6)
+        finally:
+            in_file.close()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_unixccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_unixccompiler.py
new file mode 100644
index 00000000..2c2f4aae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_unixccompiler.py
@@ -0,0 +1,350 @@
+"""Tests for distutils.unixccompiler."""
+
+import os
+import sys
+import unittest.mock as mock
+from distutils import sysconfig
+from distutils.compat import consolidate_linker_args
+from distutils.errors import DistutilsPlatformError
+from distutils.unixccompiler import UnixCCompiler
+from distutils.util import _clear_cached_macosx_ver
+
+import pytest
+
+from . import support
+from .compat.py39 import EnvironmentVarGuard
+
+
+@pytest.fixture(autouse=True)
+def save_values(monkeypatch):
+    monkeypatch.setattr(sys, 'platform', sys.platform)
+    monkeypatch.setattr(sysconfig, 'get_config_var', sysconfig.get_config_var)
+    monkeypatch.setattr(sysconfig, 'get_config_vars', sysconfig.get_config_vars)
+
+
+@pytest.fixture(autouse=True)
+def compiler_wrapper(request):
+    class CompilerWrapper(UnixCCompiler):
+        def rpath_foo(self):
+            return self.runtime_library_dir_option('/foo')
+
+    request.instance.cc = CompilerWrapper()
+
+
+class TestUnixCCompiler(support.TempdirManager):
+    @pytest.mark.skipif('platform.system == "Windows"')
+    def test_runtime_libdir_option(self):  # noqa: C901
+        # Issue #5900; GitHub Issue #37
+        #
+        # Ensure RUNPATH is added to extension modules with RPATH if
+        # GNU ld is used
+
+        # darwin
+        sys.platform = 'darwin'
+        darwin_ver_var = 'MACOSX_DEPLOYMENT_TARGET'
+        darwin_rpath_flag = '-Wl,-rpath,/foo'
+        darwin_lib_flag = '-L/foo'
+
+        # (macOS version from syscfg, macOS version from env var) -> flag
+        # Version value of None generates two tests: as None and as empty string
+        # Expected flag value of None means an mismatch exception is expected
+        darwin_test_cases = [
+            ((None, None), darwin_lib_flag),
+            ((None, '11'), darwin_rpath_flag),
+            (('10', None), darwin_lib_flag),
+            (('10.3', None), darwin_lib_flag),
+            (('10.3.1', None), darwin_lib_flag),
+            (('10.5', None), darwin_rpath_flag),
+            (('10.5.1', None), darwin_rpath_flag),
+            (('10.3', '10.3'), darwin_lib_flag),
+            (('10.3', '10.5'), darwin_rpath_flag),
+            (('10.5', '10.3'), darwin_lib_flag),
+            (('10.5', '11'), darwin_rpath_flag),
+            (('10.4', '10'), None),
+        ]
+
+        def make_darwin_gcv(syscfg_macosx_ver):
+            def gcv(var):
+                if var == darwin_ver_var:
+                    return syscfg_macosx_ver
+                return "xxx"
+
+            return gcv
+
+        def do_darwin_test(syscfg_macosx_ver, env_macosx_ver, expected_flag):
+            env = os.environ
+            msg = f"macOS version = (sysconfig={syscfg_macosx_ver!r}, env={env_macosx_ver!r})"
+
+            # Save
+            old_gcv = sysconfig.get_config_var
+            old_env_macosx_ver = env.get(darwin_ver_var)
+
+            # Setup environment
+            _clear_cached_macosx_ver()
+            sysconfig.get_config_var = make_darwin_gcv(syscfg_macosx_ver)
+            if env_macosx_ver is not None:
+                env[darwin_ver_var] = env_macosx_ver
+            elif darwin_ver_var in env:
+                env.pop(darwin_ver_var)
+
+            # Run the test
+            if expected_flag is not None:
+                assert self.cc.rpath_foo() == expected_flag, msg
+            else:
+                with pytest.raises(
+                    DistutilsPlatformError, match=darwin_ver_var + r' mismatch'
+                ):
+                    self.cc.rpath_foo()
+
+            # Restore
+            if old_env_macosx_ver is not None:
+                env[darwin_ver_var] = old_env_macosx_ver
+            elif darwin_ver_var in env:
+                env.pop(darwin_ver_var)
+            sysconfig.get_config_var = old_gcv
+            _clear_cached_macosx_ver()
+
+        for macosx_vers, expected_flag in darwin_test_cases:
+            syscfg_macosx_ver, env_macosx_ver = macosx_vers
+            do_darwin_test(syscfg_macosx_ver, env_macosx_ver, expected_flag)
+            # Bonus test cases with None interpreted as empty string
+            if syscfg_macosx_ver is None:
+                do_darwin_test("", env_macosx_ver, expected_flag)
+            if env_macosx_ver is None:
+                do_darwin_test(syscfg_macosx_ver, "", expected_flag)
+            if syscfg_macosx_ver is None and env_macosx_ver is None:
+                do_darwin_test("", "", expected_flag)
+
+        old_gcv = sysconfig.get_config_var
+
+        # hp-ux
+        sys.platform = 'hp-ux'
+
+        def gcv(v):
+            return 'xxx'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == ['+s', '-L/foo']
+
+        def gcv(v):
+            return 'gcc'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == ['-Wl,+s', '-L/foo']
+
+        def gcv(v):
+            return 'g++'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == ['-Wl,+s', '-L/foo']
+
+        sysconfig.get_config_var = old_gcv
+
+        # GCC GNULD
+        sys.platform = 'bar'
+
+        def gcv(v):
+            if v == 'CC':
+                return 'gcc'
+            elif v == 'GNULD':
+                return 'yes'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == consolidate_linker_args([
+            '-Wl,--enable-new-dtags',
+            '-Wl,-rpath,/foo',
+        ])
+
+        def gcv(v):
+            if v == 'CC':
+                return 'gcc -pthread -B /bar'
+            elif v == 'GNULD':
+                return 'yes'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == consolidate_linker_args([
+            '-Wl,--enable-new-dtags',
+            '-Wl,-rpath,/foo',
+        ])
+
+        # GCC non-GNULD
+        sys.platform = 'bar'
+
+        def gcv(v):
+            if v == 'CC':
+                return 'gcc'
+            elif v == 'GNULD':
+                return 'no'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == '-Wl,-R/foo'
+
+        # GCC GNULD with fully qualified configuration prefix
+        # see #7617
+        sys.platform = 'bar'
+
+        def gcv(v):
+            if v == 'CC':
+                return 'x86_64-pc-linux-gnu-gcc-4.4.2'
+            elif v == 'GNULD':
+                return 'yes'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == consolidate_linker_args([
+            '-Wl,--enable-new-dtags',
+            '-Wl,-rpath,/foo',
+        ])
+
+        # non-GCC GNULD
+        sys.platform = 'bar'
+
+        def gcv(v):
+            if v == 'CC':
+                return 'cc'
+            elif v == 'GNULD':
+                return 'yes'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == consolidate_linker_args([
+            '-Wl,--enable-new-dtags',
+            '-Wl,-rpath,/foo',
+        ])
+
+        # non-GCC non-GNULD
+        sys.platform = 'bar'
+
+        def gcv(v):
+            if v == 'CC':
+                return 'cc'
+            elif v == 'GNULD':
+                return 'no'
+
+        sysconfig.get_config_var = gcv
+        assert self.cc.rpath_foo() == '-Wl,-R/foo'
+
+    @pytest.mark.skipif('platform.system == "Windows"')
+    def test_cc_overrides_ldshared(self):
+        # Issue #18080:
+        # ensure that setting CC env variable also changes default linker
+        def gcv(v):
+            if v == 'LDSHARED':
+                return 'gcc-4.2 -bundle -undefined dynamic_lookup '
+            return 'gcc-4.2'
+
+        def gcvs(*args, _orig=sysconfig.get_config_vars):
+            if args:
+                return list(map(sysconfig.get_config_var, args))
+            return _orig()
+
+        sysconfig.get_config_var = gcv
+        sysconfig.get_config_vars = gcvs
+        with EnvironmentVarGuard() as env:
+            env['CC'] = 'my_cc'
+            del env['LDSHARED']
+            sysconfig.customize_compiler(self.cc)
+        assert self.cc.linker_so[0] == 'my_cc'
+
+    @pytest.mark.skipif('platform.system == "Windows"')
+    @pytest.mark.usefixtures('disable_macos_customization')
+    def test_cc_overrides_ldshared_for_cxx_correctly(self):
+        """
+        Ensure that setting CC env variable also changes default linker
+        correctly when building C++ extensions.
+
+        pypa/distutils#126
+        """
+
+        def gcv(v):
+            if v == 'LDSHARED':
+                return 'gcc-4.2 -bundle -undefined dynamic_lookup '
+            elif v == 'LDCXXSHARED':
+                return 'g++-4.2 -bundle -undefined dynamic_lookup '
+            elif v == 'CXX':
+                return 'g++-4.2'
+            elif v == 'CC':
+                return 'gcc-4.2'
+            return ''
+
+        def gcvs(*args, _orig=sysconfig.get_config_vars):
+            if args:
+                return list(map(sysconfig.get_config_var, args))
+            return _orig()
+
+        sysconfig.get_config_var = gcv
+        sysconfig.get_config_vars = gcvs
+        with (
+            mock.patch.object(self.cc, 'spawn', return_value=None) as mock_spawn,
+            mock.patch.object(self.cc, '_need_link', return_value=True),
+            mock.patch.object(self.cc, 'mkpath', return_value=None),
+            EnvironmentVarGuard() as env,
+        ):
+            env['CC'] = 'ccache my_cc'
+            env['CXX'] = 'my_cxx'
+            del env['LDSHARED']
+            sysconfig.customize_compiler(self.cc)
+            assert self.cc.linker_so[0:2] == ['ccache', 'my_cc']
+            self.cc.link(None, [], 'a.out', target_lang='c++')
+            call_args = mock_spawn.call_args[0][0]
+            expected = ['my_cxx', '-bundle', '-undefined', 'dynamic_lookup']
+            assert call_args[:4] == expected
+
+    @pytest.mark.skipif('platform.system == "Windows"')
+    def test_explicit_ldshared(self):
+        # Issue #18080:
+        # ensure that setting CC env variable does not change
+        #   explicit LDSHARED setting for linker
+        def gcv(v):
+            if v == 'LDSHARED':
+                return 'gcc-4.2 -bundle -undefined dynamic_lookup '
+            return 'gcc-4.2'
+
+        def gcvs(*args, _orig=sysconfig.get_config_vars):
+            if args:
+                return list(map(sysconfig.get_config_var, args))
+            return _orig()
+
+        sysconfig.get_config_var = gcv
+        sysconfig.get_config_vars = gcvs
+        with EnvironmentVarGuard() as env:
+            env['CC'] = 'my_cc'
+            env['LDSHARED'] = 'my_ld -bundle -dynamic'
+            sysconfig.customize_compiler(self.cc)
+        assert self.cc.linker_so[0] == 'my_ld'
+
+    def test_has_function(self):
+        # Issue https://github.com/pypa/distutils/issues/64:
+        # ensure that setting output_dir does not raise
+        # FileNotFoundError: [Errno 2] No such file or directory: 'a.out'
+        self.cc.output_dir = 'scratch'
+        os.chdir(self.mkdtemp())
+        self.cc.has_function('abort')
+
+    def test_find_library_file(self, monkeypatch):
+        compiler = UnixCCompiler()
+        compiler._library_root = lambda dir: dir
+        monkeypatch.setattr(os.path, 'exists', lambda d: 'existing' in d)
+
+        libname = 'libabc.dylib' if sys.platform != 'cygwin' else 'cygabc.dll'
+        dirs = ('/foo/bar/missing', '/foo/bar/existing')
+        assert (
+            compiler.find_library_file(dirs, 'abc').replace('\\', '/')
+            == f'/foo/bar/existing/{libname}'
+        )
+        assert (
+            compiler.find_library_file(reversed(dirs), 'abc').replace('\\', '/')
+            == f'/foo/bar/existing/{libname}'
+        )
+
+        monkeypatch.setattr(
+            os.path,
+            'exists',
+            lambda d: 'existing' in d and '.a' in d and '.dll.a' not in d,
+        )
+        assert (
+            compiler.find_library_file(dirs, 'abc').replace('\\', '/')
+            == '/foo/bar/existing/libabc.a'
+        )
+        assert (
+            compiler.find_library_file(reversed(dirs), 'abc').replace('\\', '/')
+            == '/foo/bar/existing/libabc.a'
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_util.py
new file mode 100644
index 00000000..00c9743e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_util.py
@@ -0,0 +1,243 @@
+"""Tests for distutils.util."""
+
+import email
+import email.generator
+import email.policy
+import io
+import os
+import pathlib
+import sys
+import sysconfig as stdlib_sysconfig
+import unittest.mock as mock
+from copy import copy
+from distutils import sysconfig, util
+from distutils.errors import DistutilsByteCompileError, DistutilsPlatformError
+from distutils.util import (
+    byte_compile,
+    change_root,
+    check_environ,
+    convert_path,
+    get_host_platform,
+    get_platform,
+    grok_environment_error,
+    rfc822_escape,
+    split_quoted,
+    strtobool,
+)
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def environment(monkeypatch):
+    monkeypatch.setattr(os, 'name', os.name)
+    monkeypatch.setattr(sys, 'platform', sys.platform)
+    monkeypatch.setattr(sys, 'version', sys.version)
+    monkeypatch.setattr(os, 'sep', os.sep)
+    monkeypatch.setattr(os.path, 'join', os.path.join)
+    monkeypatch.setattr(os.path, 'isabs', os.path.isabs)
+    monkeypatch.setattr(os.path, 'splitdrive', os.path.splitdrive)
+    monkeypatch.setattr(sysconfig, '_config_vars', copy(sysconfig._config_vars))
+
+
+@pytest.mark.usefixtures('save_env')
+class TestUtil:
+    def test_get_host_platform(self):
+        with mock.patch('os.name', 'nt'):
+            with mock.patch('sys.version', '... [... (ARM64)]'):
+                assert get_host_platform() == 'win-arm64'
+            with mock.patch('sys.version', '... [... (ARM)]'):
+                assert get_host_platform() == 'win-arm32'
+
+        with mock.patch('sys.version_info', (3, 9, 0, 'final', 0)):
+            assert get_host_platform() == stdlib_sysconfig.get_platform()
+
+    def test_get_platform(self):
+        with mock.patch('os.name', 'nt'):
+            with mock.patch.dict('os.environ', {'VSCMD_ARG_TGT_ARCH': 'x86'}):
+                assert get_platform() == 'win32'
+            with mock.patch.dict('os.environ', {'VSCMD_ARG_TGT_ARCH': 'x64'}):
+                assert get_platform() == 'win-amd64'
+            with mock.patch.dict('os.environ', {'VSCMD_ARG_TGT_ARCH': 'arm'}):
+                assert get_platform() == 'win-arm32'
+            with mock.patch.dict('os.environ', {'VSCMD_ARG_TGT_ARCH': 'arm64'}):
+                assert get_platform() == 'win-arm64'
+
+    def test_convert_path(self):
+        expected = os.sep.join(('', 'home', 'to', 'my', 'stuff'))
+        assert convert_path('/home/to/my/stuff') == expected
+        assert convert_path(pathlib.Path('/home/to/my/stuff')) == expected
+        assert convert_path('.') == os.curdir
+
+    def test_change_root(self):
+        # linux/mac
+        os.name = 'posix'
+
+        def _isabs(path):
+            return path[0] == '/'
+
+        os.path.isabs = _isabs
+
+        def _join(*path):
+            return '/'.join(path)
+
+        os.path.join = _join
+
+        assert change_root('/root', '/old/its/here') == '/root/old/its/here'
+        assert change_root('/root', 'its/here') == '/root/its/here'
+
+        # windows
+        os.name = 'nt'
+        os.sep = '\\'
+
+        def _isabs(path):
+            return path.startswith('c:\\')
+
+        os.path.isabs = _isabs
+
+        def _splitdrive(path):
+            if path.startswith('c:'):
+                return ('', path.replace('c:', ''))
+            return ('', path)
+
+        os.path.splitdrive = _splitdrive
+
+        def _join(*path):
+            return '\\'.join(path)
+
+        os.path.join = _join
+
+        assert (
+            change_root('c:\\root', 'c:\\old\\its\\here') == 'c:\\root\\old\\its\\here'
+        )
+        assert change_root('c:\\root', 'its\\here') == 'c:\\root\\its\\here'
+
+        # BugsBunny os (it's a great os)
+        os.name = 'BugsBunny'
+        with pytest.raises(DistutilsPlatformError):
+            change_root('c:\\root', 'its\\here')
+
+        # XXX platforms to be covered: mac
+
+    def test_check_environ(self):
+        util.check_environ.cache_clear()
+        os.environ.pop('HOME', None)
+
+        check_environ()
+
+        assert os.environ['PLAT'] == get_platform()
+
+    @pytest.mark.skipif("os.name != 'posix'")
+    def test_check_environ_getpwuid(self):
+        util.check_environ.cache_clear()
+        os.environ.pop('HOME', None)
+
+        import pwd
+
+        # only set pw_dir field, other fields are not used
+        result = pwd.struct_passwd((
+            None,
+            None,
+            None,
+            None,
+            None,
+            '/home/distutils',
+            None,
+        ))
+        with mock.patch.object(pwd, 'getpwuid', return_value=result):
+            check_environ()
+            assert os.environ['HOME'] == '/home/distutils'
+
+        util.check_environ.cache_clear()
+        os.environ.pop('HOME', None)
+
+        # bpo-10496: Catch pwd.getpwuid() error
+        with mock.patch.object(pwd, 'getpwuid', side_effect=KeyError):
+            check_environ()
+            assert 'HOME' not in os.environ
+
+    def test_split_quoted(self):
+        assert split_quoted('""one"" "two" \'three\' \\four') == [
+            'one',
+            'two',
+            'three',
+            'four',
+        ]
+
+    def test_strtobool(self):
+        yes = ('y', 'Y', 'yes', 'True', 't', 'true', 'True', 'On', 'on', '1')
+        no = ('n', 'no', 'f', 'false', 'off', '0', 'Off', 'No', 'N')
+
+        for y in yes:
+            assert strtobool(y)
+
+        for n in no:
+            assert not strtobool(n)
+
+    indent = 8 * ' '
+
+    @pytest.mark.parametrize(
+        "given,wanted",
+        [
+            # 0x0b, 0x0c, ..., etc are also considered a line break by Python
+            ("hello\x0b\nworld\n", f"hello\x0b{indent}\n{indent}world\n{indent}"),
+            ("hello\x1eworld", f"hello\x1e{indent}world"),
+            ("", ""),
+            (
+                "I am a\npoor\nlonesome\nheader\n",
+                f"I am a\n{indent}poor\n{indent}lonesome\n{indent}header\n{indent}",
+            ),
+        ],
+    )
+    def test_rfc822_escape(self, given, wanted):
+        """
+        We want to ensure a multi-line header parses correctly.
+
+        For interoperability, the escaped value should also "round-trip" over
+        `email.generator.Generator.flatten` and `email.message_from_*`
+        (see pypa/setuptools#4033).
+
+        The main issue is that internally `email.policy.EmailPolicy` uses
+        `splitlines` which will split on some control chars. If all the new lines
+        are not prefixed with spaces, the parser will interrupt reading
+        the current header and produce an incomplete value, while
+        incorrectly interpreting the rest of the headers as part of the payload.
+        """
+        res = rfc822_escape(given)
+
+        policy = email.policy.EmailPolicy(
+            utf8=True,
+            mangle_from_=False,
+            max_line_length=0,
+        )
+        with io.StringIO() as buffer:
+            raw = f"header: {res}\nother-header: 42\n\npayload\n"
+            orig = email.message_from_string(raw)
+            email.generator.Generator(buffer, policy=policy).flatten(orig)
+            buffer.seek(0)
+            regen = email.message_from_file(buffer)
+
+        for msg in (orig, regen):
+            assert msg.get_payload() == "payload\n"
+            assert msg["other-header"] == "42"
+            # Generator may replace control chars with `\n`
+            assert set(msg["header"].splitlines()) == set(res.splitlines())
+
+        assert res == wanted
+
+    def test_dont_write_bytecode(self):
+        # makes sure byte_compile raise a DistutilsError
+        # if sys.dont_write_bytecode is True
+        old_dont_write_bytecode = sys.dont_write_bytecode
+        sys.dont_write_bytecode = True
+        try:
+            with pytest.raises(DistutilsByteCompileError):
+                byte_compile([])
+        finally:
+            sys.dont_write_bytecode = old_dont_write_bytecode
+
+    def test_grok_environment_error(self):
+        # test obsolete function to ensure backward compat (#4931)
+        exc = OSError("Unable to find batch file")
+        msg = grok_environment_error(exc)
+        assert msg == "error: Unable to find batch file"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_version.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_version.py
new file mode 100644
index 00000000..b68f0977
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_version.py
@@ -0,0 +1,80 @@
+"""Tests for distutils.version."""
+
+import distutils
+from distutils.version import LooseVersion, StrictVersion
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def suppress_deprecation():
+    with distutils.version.suppress_known_deprecation():
+        yield
+
+
+class TestVersion:
+    def test_prerelease(self):
+        version = StrictVersion('1.2.3a1')
+        assert version.version == (1, 2, 3)
+        assert version.prerelease == ('a', 1)
+        assert str(version) == '1.2.3a1'
+
+        version = StrictVersion('1.2.0')
+        assert str(version) == '1.2'
+
+    def test_cmp_strict(self):
+        versions = (
+            ('1.5.1', '1.5.2b2', -1),
+            ('161', '3.10a', ValueError),
+            ('8.02', '8.02', 0),
+            ('3.4j', '1996.07.12', ValueError),
+            ('3.2.pl0', '3.1.1.6', ValueError),
+            ('2g6', '11g', ValueError),
+            ('0.9', '2.2', -1),
+            ('1.2.1', '1.2', 1),
+            ('1.1', '1.2.2', -1),
+            ('1.2', '1.1', 1),
+            ('1.2.1', '1.2.2', -1),
+            ('1.2.2', '1.2', 1),
+            ('1.2', '1.2.2', -1),
+            ('0.4.0', '0.4', 0),
+            ('1.13++', '5.5.kw', ValueError),
+        )
+
+        for v1, v2, wanted in versions:
+            try:
+                res = StrictVersion(v1)._cmp(StrictVersion(v2))
+            except ValueError:
+                if wanted is ValueError:
+                    continue
+                else:
+                    raise AssertionError(f"cmp({v1}, {v2}) shouldn't raise ValueError")
+            assert res == wanted, f'cmp({v1}, {v2}) should be {wanted}, got {res}'
+            res = StrictVersion(v1)._cmp(v2)
+            assert res == wanted, f'cmp({v1}, {v2}) should be {wanted}, got {res}'
+            res = StrictVersion(v1)._cmp(object())
+            assert res is NotImplemented, (
+                f'cmp({v1}, {v2}) should be NotImplemented, got {res}'
+            )
+
+    def test_cmp(self):
+        versions = (
+            ('1.5.1', '1.5.2b2', -1),
+            ('161', '3.10a', 1),
+            ('8.02', '8.02', 0),
+            ('3.4j', '1996.07.12', -1),
+            ('3.2.pl0', '3.1.1.6', 1),
+            ('2g6', '11g', -1),
+            ('0.960923', '2.2beta29', -1),
+            ('1.13++', '5.5.kw', -1),
+        )
+
+        for v1, v2, wanted in versions:
+            res = LooseVersion(v1)._cmp(LooseVersion(v2))
+            assert res == wanted, f'cmp({v1}, {v2}) should be {wanted}, got {res}'
+            res = LooseVersion(v1)._cmp(v2)
+            assert res == wanted, f'cmp({v1}, {v2}) should be {wanted}, got {res}'
+            res = LooseVersion(v1)._cmp(object())
+            assert res is NotImplemented, (
+                f'cmp({v1}, {v2}) should be NotImplemented, got {res}'
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_versionpredicate.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/test_versionpredicate.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/unix_compat.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/unix_compat.py
new file mode 100644
index 00000000..a5d9ee45
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/tests/unix_compat.py
@@ -0,0 +1,17 @@
+import sys
+
+try:
+    import grp
+    import pwd
+except ImportError:
+    grp = pwd = None
+
+import pytest
+
+UNIX_ID_SUPPORT = grp and pwd
+UID_0_SUPPORT = UNIX_ID_SUPPORT and sys.platform != "cygwin"
+
+require_unix_id = pytest.mark.skipif(
+    not UNIX_ID_SUPPORT, reason="Requires grp and pwd support"
+)
+require_uid_0 = pytest.mark.skipif(not UID_0_SUPPORT, reason="Requires UID 0 support")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/text_file.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/text_file.py
new file mode 100644
index 00000000..89d9048d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/text_file.py
@@ -0,0 +1,286 @@
+"""text_file
+
+provides the TextFile class, which gives an interface to text files
+that (optionally) takes care of stripping comments, ignoring blank
+lines, and joining lines with backslashes."""
+
+import sys
+
+
+class TextFile:
+    """Provides a file-like object that takes care of all the things you
+    commonly want to do when processing a text file that has some
+    line-by-line syntax: strip comments (as long as "#" is your
+    comment character), skip blank lines, join adjacent lines by
+    escaping the newline (ie. backslash at end of line), strip
+    leading and/or trailing whitespace.  All of these are optional
+    and independently controllable.
+
+    Provides a 'warn()' method so you can generate warning messages that
+    report physical line number, even if the logical line in question
+    spans multiple physical lines.  Also provides 'unreadline()' for
+    implementing line-at-a-time lookahead.
+
+    Constructor is called as:
+
+        TextFile (filename=None, file=None, **options)
+
+    It bombs (RuntimeError) if both 'filename' and 'file' are None;
+    'filename' should be a string, and 'file' a file object (or
+    something that provides 'readline()' and 'close()' methods).  It is
+    recommended that you supply at least 'filename', so that TextFile
+    can include it in warning messages.  If 'file' is not supplied,
+    TextFile creates its own using 'io.open()'.
+
+    The options are all boolean, and affect the value returned by
+    'readline()':
+      strip_comments [default: true]
+        strip from "#" to end-of-line, as well as any whitespace
+        leading up to the "#" -- unless it is escaped by a backslash
+      lstrip_ws [default: false]
+        strip leading whitespace from each line before returning it
+      rstrip_ws [default: true]
+        strip trailing whitespace (including line terminator!) from
+        each line before returning it
+      skip_blanks [default: true}
+        skip lines that are empty *after* stripping comments and
+        whitespace.  (If both lstrip_ws and rstrip_ws are false,
+        then some lines may consist of solely whitespace: these will
+        *not* be skipped, even if 'skip_blanks' is true.)
+      join_lines [default: false]
+        if a backslash is the last non-newline character on a line
+        after stripping comments and whitespace, join the following line
+        to it to form one "logical line"; if N consecutive lines end
+        with a backslash, then N+1 physical lines will be joined to
+        form one logical line.
+      collapse_join [default: false]
+        strip leading whitespace from lines that are joined to their
+        predecessor; only matters if (join_lines and not lstrip_ws)
+      errors [default: 'strict']
+        error handler used to decode the file content
+
+    Note that since 'rstrip_ws' can strip the trailing newline, the
+    semantics of 'readline()' must differ from those of the builtin file
+    object's 'readline()' method!  In particular, 'readline()' returns
+    None for end-of-file: an empty string might just be a blank line (or
+    an all-whitespace line), if 'rstrip_ws' is true but 'skip_blanks' is
+    not."""
+
+    default_options = {
+        'strip_comments': 1,
+        'skip_blanks': 1,
+        'lstrip_ws': 0,
+        'rstrip_ws': 1,
+        'join_lines': 0,
+        'collapse_join': 0,
+        'errors': 'strict',
+    }
+
+    def __init__(self, filename=None, file=None, **options):
+        """Construct a new TextFile object.  At least one of 'filename'
+        (a string) and 'file' (a file-like object) must be supplied.
+        They keyword argument options are described above and affect
+        the values returned by 'readline()'."""
+        if filename is None and file is None:
+            raise RuntimeError(
+                "you must supply either or both of 'filename' and 'file'"
+            )
+
+        # set values for all options -- either from client option hash
+        # or fallback to default_options
+        for opt in self.default_options.keys():
+            if opt in options:
+                setattr(self, opt, options[opt])
+            else:
+                setattr(self, opt, self.default_options[opt])
+
+        # sanity check client option hash
+        for opt in options.keys():
+            if opt not in self.default_options:
+                raise KeyError(f"invalid TextFile option '{opt}'")
+
+        if file is None:
+            self.open(filename)
+        else:
+            self.filename = filename
+            self.file = file
+            self.current_line = 0  # assuming that file is at BOF!
+
+        # 'linebuf' is a stack of lines that will be emptied before we
+        # actually read from the file; it's only populated by an
+        # 'unreadline()' operation
+        self.linebuf = []
+
+    def open(self, filename):
+        """Open a new file named 'filename'.  This overrides both the
+        'filename' and 'file' arguments to the constructor."""
+        self.filename = filename
+        self.file = open(self.filename, errors=self.errors, encoding='utf-8')
+        self.current_line = 0
+
+    def close(self):
+        """Close the current file and forget everything we know about it
+        (filename, current line number)."""
+        file = self.file
+        self.file = None
+        self.filename = None
+        self.current_line = None
+        file.close()
+
+    def gen_error(self, msg, line=None):
+        outmsg = []
+        if line is None:
+            line = self.current_line
+        outmsg.append(self.filename + ", ")
+        if isinstance(line, (list, tuple)):
+            outmsg.append("lines {}-{}: ".format(*line))
+        else:
+            outmsg.append(f"line {int(line)}: ")
+        outmsg.append(str(msg))
+        return "".join(outmsg)
+
+    def error(self, msg, line=None):
+        raise ValueError("error: " + self.gen_error(msg, line))
+
+    def warn(self, msg, line=None):
+        """Print (to stderr) a warning message tied to the current logical
+        line in the current file.  If the current logical line in the
+        file spans multiple physical lines, the warning refers to the
+        whole range, eg. "lines 3-5".  If 'line' supplied, it overrides
+        the current line number; it may be a list or tuple to indicate a
+        range of physical lines, or an integer for a single physical
+        line."""
+        sys.stderr.write("warning: " + self.gen_error(msg, line) + "\n")
+
+    def readline(self):  # noqa: C901
+        """Read and return a single logical line from the current file (or
+        from an internal buffer if lines have previously been "unread"
+        with 'unreadline()').  If the 'join_lines' option is true, this
+        may involve reading multiple physical lines concatenated into a
+        single string.  Updates the current line number, so calling
+        'warn()' after 'readline()' emits a warning about the physical
+        line(s) just read.  Returns None on end-of-file, since the empty
+        string can occur if 'rstrip_ws' is true but 'strip_blanks' is
+        not."""
+        # If any "unread" lines waiting in 'linebuf', return the top
+        # one.  (We don't actually buffer read-ahead data -- lines only
+        # get put in 'linebuf' if the client explicitly does an
+        # 'unreadline()'.
+        if self.linebuf:
+            line = self.linebuf[-1]
+            del self.linebuf[-1]
+            return line
+
+        buildup_line = ''
+
+        while True:
+            # read the line, make it None if EOF
+            line = self.file.readline()
+            if line == '':
+                line = None
+
+            if self.strip_comments and line:
+                # Look for the first "#" in the line.  If none, never
+                # mind.  If we find one and it's the first character, or
+                # is not preceded by "\", then it starts a comment --
+                # strip the comment, strip whitespace before it, and
+                # carry on.  Otherwise, it's just an escaped "#", so
+                # unescape it (and any other escaped "#"'s that might be
+                # lurking in there) and otherwise leave the line alone.
+
+                pos = line.find("#")
+                if pos == -1:  # no "#" -- no comments
+                    pass
+
+                # It's definitely a comment -- either "#" is the first
+                # character, or it's elsewhere and unescaped.
+                elif pos == 0 or line[pos - 1] != "\\":
+                    # Have to preserve the trailing newline, because it's
+                    # the job of a later step (rstrip_ws) to remove it --
+                    # and if rstrip_ws is false, we'd better preserve it!
+                    # (NB. this means that if the final line is all comment
+                    # and has no trailing newline, we will think that it's
+                    # EOF; I think that's OK.)
+                    eol = (line[-1] == '\n') and '\n' or ''
+                    line = line[0:pos] + eol
+
+                    # If all that's left is whitespace, then skip line
+                    # *now*, before we try to join it to 'buildup_line' --
+                    # that way constructs like
+                    #   hello \\
+                    #   # comment that should be ignored
+                    #   there
+                    # result in "hello there".
+                    if line.strip() == "":
+                        continue
+                else:  # it's an escaped "#"
+                    line = line.replace("\\#", "#")
+
+            # did previous line end with a backslash? then accumulate
+            if self.join_lines and buildup_line:
+                # oops: end of file
+                if line is None:
+                    self.warn("continuation line immediately precedes end-of-file")
+                    return buildup_line
+
+                if self.collapse_join:
+                    line = line.lstrip()
+                line = buildup_line + line
+
+                # careful: pay attention to line number when incrementing it
+                if isinstance(self.current_line, list):
+                    self.current_line[1] = self.current_line[1] + 1
+                else:
+                    self.current_line = [self.current_line, self.current_line + 1]
+            # just an ordinary line, read it as usual
+            else:
+                if line is None:  # eof
+                    return None
+
+                # still have to be careful about incrementing the line number!
+                if isinstance(self.current_line, list):
+                    self.current_line = self.current_line[1] + 1
+                else:
+                    self.current_line = self.current_line + 1
+
+            # strip whitespace however the client wants (leading and
+            # trailing, or one or the other, or neither)
+            if self.lstrip_ws and self.rstrip_ws:
+                line = line.strip()
+            elif self.lstrip_ws:
+                line = line.lstrip()
+            elif self.rstrip_ws:
+                line = line.rstrip()
+
+            # blank line (whether we rstrip'ed or not)? skip to next line
+            # if appropriate
+            if line in ('', '\n') and self.skip_blanks:
+                continue
+
+            if self.join_lines:
+                if line[-1] == '\\':
+                    buildup_line = line[:-1]
+                    continue
+
+                if line[-2:] == '\\\n':
+                    buildup_line = line[0:-2] + '\n'
+                    continue
+
+            # well, I guess there's some actual content there: return it
+            return line
+
+    def readlines(self):
+        """Read and return the list of all logical lines remaining in the
+        current file."""
+        lines = []
+        while True:
+            line = self.readline()
+            if line is None:
+                return lines
+            lines.append(line)
+
+    def unreadline(self, line):
+        """Push 'line' (a string) onto an internal buffer that will be
+        checked by future 'readline()' calls.  Handy for implementing
+        a parser with line-at-a-time lookahead."""
+        self.linebuf.append(line)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/unixccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/unixccompiler.py
new file mode 100644
index 00000000..6c1116ae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/unixccompiler.py
@@ -0,0 +1,402 @@
+"""distutils.unixccompiler
+
+Contains the UnixCCompiler class, a subclass of CCompiler that handles
+the "typical" Unix-style command-line C compiler:
+  * macros defined with -Dname[=value]
+  * macros undefined with -Uname
+  * include search directories specified with -Idir
+  * libraries specified with -lllib
+  * library search directories specified with -Ldir
+  * compile handled by 'cc' (or similar) executable with -c option:
+    compiles .c to .o
+  * link static library handled by 'ar' command (possibly with 'ranlib')
+  * link shared library handled by 'cc -shared'
+"""
+
+from __future__ import annotations
+
+import itertools
+import os
+import re
+import shlex
+import sys
+
+from . import sysconfig
+from ._log import log
+from ._macos_compat import compiler_fixup
+from ._modified import newer
+from .ccompiler import CCompiler, gen_lib_options, gen_preprocess_options
+from .compat import consolidate_linker_args
+from .errors import CompileError, DistutilsExecError, LibError, LinkError
+
+# XXX Things not currently handled:
+#   * optimization/debug/warning flags; we just use whatever's in Python's
+#     Makefile and live with it.  Is this adequate?  If not, we might
+#     have to have a bunch of subclasses GNUCCompiler, SGICCompiler,
+#     SunCCompiler, and I suspect down that road lies madness.
+#   * even if we don't know a warning flag from an optimization flag,
+#     we need some way for outsiders to feed preprocessor/compiler/linker
+#     flags in to us -- eg. a sysadmin might want to mandate certain flags
+#     via a site config file, or a user might want to set something for
+#     compiling this module distribution only via the setup.py command
+#     line, whatever.  As long as these options come from something on the
+#     current system, they can be as system-dependent as they like, and we
+#     should just happily stuff them into the preprocessor/compiler/linker
+#     options and carry on.
+
+
+def _split_env(cmd):
+    """
+    For macOS, split command into 'env' portion (if any)
+    and the rest of the linker command.
+
+    >>> _split_env(['a', 'b', 'c'])
+    ([], ['a', 'b', 'c'])
+    >>> _split_env(['/usr/bin/env', 'A=3', 'gcc'])
+    (['/usr/bin/env', 'A=3'], ['gcc'])
+    """
+    pivot = 0
+    if os.path.basename(cmd[0]) == "env":
+        pivot = 1
+        while '=' in cmd[pivot]:
+            pivot += 1
+    return cmd[:pivot], cmd[pivot:]
+
+
+def _split_aix(cmd):
+    """
+    AIX platforms prefix the compiler with the ld_so_aix
+    script, so split that from the linker command.
+
+    >>> _split_aix(['a', 'b', 'c'])
+    ([], ['a', 'b', 'c'])
+    >>> _split_aix(['/bin/foo/ld_so_aix', 'gcc'])
+    (['/bin/foo/ld_so_aix'], ['gcc'])
+    """
+    pivot = os.path.basename(cmd[0]) == 'ld_so_aix'
+    return cmd[:pivot], cmd[pivot:]
+
+
+def _linker_params(linker_cmd, compiler_cmd):
+    """
+    The linker command usually begins with the compiler
+    command (possibly multiple elements), followed by zero or more
+    params for shared library building.
+
+    If the LDSHARED env variable overrides the linker command,
+    however, the commands may not match.
+
+    Return the best guess of the linker parameters by stripping
+    the linker command. If the compiler command does not
+    match the linker command, assume the linker command is
+    just the first element.
+
+    >>> _linker_params('gcc foo bar'.split(), ['gcc'])
+    ['foo', 'bar']
+    >>> _linker_params('gcc foo bar'.split(), ['other'])
+    ['foo', 'bar']
+    >>> _linker_params('ccache gcc foo bar'.split(), 'ccache gcc'.split())
+    ['foo', 'bar']
+    >>> _linker_params(['gcc'], ['gcc'])
+    []
+    """
+    c_len = len(compiler_cmd)
+    pivot = c_len if linker_cmd[:c_len] == compiler_cmd else 1
+    return linker_cmd[pivot:]
+
+
+class UnixCCompiler(CCompiler):
+    compiler_type = 'unix'
+
+    # These are used by CCompiler in two places: the constructor sets
+    # instance attributes 'preprocessor', 'compiler', etc. from them, and
+    # 'set_executable()' allows any of these to be set.  The defaults here
+    # are pretty generic; they will probably have to be set by an outsider
+    # (eg. using information discovered by the sysconfig about building
+    # Python extensions).
+    executables = {
+        'preprocessor': None,
+        'compiler': ["cc"],
+        'compiler_so': ["cc"],
+        'compiler_cxx': ["c++"],
+        'compiler_so_cxx': ["c++"],
+        'linker_so': ["cc", "-shared"],
+        'linker_so_cxx': ["c++", "-shared"],
+        'linker_exe': ["cc"],
+        'linker_exe_cxx': ["c++", "-shared"],
+        'archiver': ["ar", "-cr"],
+        'ranlib': None,
+    }
+
+    if sys.platform[:6] == "darwin":
+        executables['ranlib'] = ["ranlib"]
+
+    # Needed for the filename generation methods provided by the base
+    # class, CCompiler.  NB. whoever instantiates/uses a particular
+    # UnixCCompiler instance should set 'shared_lib_ext' -- we set a
+    # reasonable common default here, but it's not necessarily used on all
+    # Unices!
+
+    src_extensions = [".c", ".C", ".cc", ".cxx", ".cpp", ".m"]
+    obj_extension = ".o"
+    static_lib_extension = ".a"
+    shared_lib_extension = ".so"
+    dylib_lib_extension = ".dylib"
+    xcode_stub_lib_extension = ".tbd"
+    static_lib_format = shared_lib_format = dylib_lib_format = "lib%s%s"
+    xcode_stub_lib_format = dylib_lib_format
+    if sys.platform == "cygwin":
+        exe_extension = ".exe"
+        shared_lib_extension = ".dll.a"
+        dylib_lib_extension = ".dll"
+        dylib_lib_format = "cyg%s%s"
+
+    def preprocess(
+        self,
+        source,
+        output_file=None,
+        macros=None,
+        include_dirs=None,
+        extra_preargs=None,
+        extra_postargs=None,
+    ):
+        fixed_args = self._fix_compile_args(None, macros, include_dirs)
+        ignore, macros, include_dirs = fixed_args
+        pp_opts = gen_preprocess_options(macros, include_dirs)
+        pp_args = self.preprocessor + pp_opts
+        if output_file:
+            pp_args.extend(['-o', output_file])
+        if extra_preargs:
+            pp_args[:0] = extra_preargs
+        if extra_postargs:
+            pp_args.extend(extra_postargs)
+        pp_args.append(source)
+
+        # reasons to preprocess:
+        # - force is indicated
+        # - output is directed to stdout
+        # - source file is newer than the target
+        preprocess = self.force or output_file is None or newer(source, output_file)
+        if not preprocess:
+            return
+
+        if output_file:
+            self.mkpath(os.path.dirname(output_file))
+
+        try:
+            self.spawn(pp_args)
+        except DistutilsExecError as msg:
+            raise CompileError(msg)
+
+    def _compile(self, obj, src, ext, cc_args, extra_postargs, pp_opts):
+        compiler_so = compiler_fixup(self.compiler_so, cc_args + extra_postargs)
+        compiler_so_cxx = compiler_fixup(self.compiler_so_cxx, cc_args + extra_postargs)
+        try:
+            if self.detect_language(src) == 'c++':
+                self.spawn(
+                    compiler_so_cxx + cc_args + [src, '-o', obj] + extra_postargs
+                )
+            else:
+                self.spawn(compiler_so + cc_args + [src, '-o', obj] + extra_postargs)
+        except DistutilsExecError as msg:
+            raise CompileError(msg)
+
+    def create_static_lib(
+        self, objects, output_libname, output_dir=None, debug=False, target_lang=None
+    ):
+        objects, output_dir = self._fix_object_args(objects, output_dir)
+
+        output_filename = self.library_filename(output_libname, output_dir=output_dir)
+
+        if self._need_link(objects, output_filename):
+            self.mkpath(os.path.dirname(output_filename))
+            self.spawn(self.archiver + [output_filename] + objects + self.objects)
+
+            # Not many Unices required ranlib anymore -- SunOS 4.x is, I
+            # think the only major Unix that does.  Maybe we need some
+            # platform intelligence here to skip ranlib if it's not
+            # needed -- or maybe Python's configure script took care of
+            # it for us, hence the check for leading colon.
+            if self.ranlib:
+                try:
+                    self.spawn(self.ranlib + [output_filename])
+                except DistutilsExecError as msg:
+                    raise LibError(msg)
+        else:
+            log.debug("skipping %s (up-to-date)", output_filename)
+
+    def link(
+        self,
+        target_desc,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        objects, output_dir = self._fix_object_args(objects, output_dir)
+        fixed_args = self._fix_lib_args(libraries, library_dirs, runtime_library_dirs)
+        libraries, library_dirs, runtime_library_dirs = fixed_args
+
+        lib_opts = gen_lib_options(self, library_dirs, runtime_library_dirs, libraries)
+        if not isinstance(output_dir, (str, type(None))):
+            raise TypeError("'output_dir' must be a string or None")
+        if output_dir is not None:
+            output_filename = os.path.join(output_dir, output_filename)
+
+        if self._need_link(objects, output_filename):
+            ld_args = objects + self.objects + lib_opts + ['-o', output_filename]
+            if debug:
+                ld_args[:0] = ['-g']
+            if extra_preargs:
+                ld_args[:0] = extra_preargs
+            if extra_postargs:
+                ld_args.extend(extra_postargs)
+            self.mkpath(os.path.dirname(output_filename))
+            try:
+                # Select a linker based on context: linker_exe when
+                # building an executable or linker_so (with shared options)
+                # when building a shared library.
+                building_exe = target_desc == CCompiler.EXECUTABLE
+                linker = (
+                    self.linker_exe
+                    if building_exe
+                    else (
+                        self.linker_so_cxx if target_lang == "c++" else self.linker_so
+                    )
+                )[:]
+
+                if target_lang == "c++" and self.compiler_cxx:
+                    env, linker_ne = _split_env(linker)
+                    aix, linker_na = _split_aix(linker_ne)
+                    _, compiler_cxx_ne = _split_env(self.compiler_cxx)
+                    _, linker_exe_ne = _split_env(self.linker_exe)
+
+                    params = _linker_params(linker_na, linker_exe_ne)
+                    linker = env + aix + compiler_cxx_ne + params
+
+                linker = compiler_fixup(linker, ld_args)
+
+                self.spawn(linker + ld_args)
+            except DistutilsExecError as msg:
+                raise LinkError(msg)
+        else:
+            log.debug("skipping %s (up-to-date)", output_filename)
+
+    # -- Miscellaneous methods -----------------------------------------
+    # These are all used by the 'gen_lib_options() function, in
+    # ccompiler.py.
+
+    def library_dir_option(self, dir):
+        return "-L" + dir
+
+    def _is_gcc(self):
+        cc_var = sysconfig.get_config_var("CC")
+        compiler = os.path.basename(shlex.split(cc_var)[0])
+        return "gcc" in compiler or "g++" in compiler
+
+    def runtime_library_dir_option(self, dir: str) -> str | list[str]:
+        # XXX Hackish, at the very least.  See Python bug #445902:
+        # https://bugs.python.org/issue445902
+        # Linkers on different platforms need different options to
+        # specify that directories need to be added to the list of
+        # directories searched for dependencies when a dynamic library
+        # is sought.  GCC on GNU systems (Linux, FreeBSD, ...) has to
+        # be told to pass the -R option through to the linker, whereas
+        # other compilers and gcc on other systems just know this.
+        # Other compilers may need something slightly different.  At
+        # this time, there's no way to determine this information from
+        # the configuration data stored in the Python installation, so
+        # we use this hack.
+        if sys.platform[:6] == "darwin":
+            from distutils.util import get_macosx_target_ver, split_version
+
+            macosx_target_ver = get_macosx_target_ver()
+            if macosx_target_ver and split_version(macosx_target_ver) >= [10, 5]:
+                return "-Wl,-rpath," + dir
+            else:  # no support for -rpath on earlier macOS versions
+                return "-L" + dir
+        elif sys.platform[:7] == "freebsd":
+            return "-Wl,-rpath=" + dir
+        elif sys.platform[:5] == "hp-ux":
+            return [
+                "-Wl,+s" if self._is_gcc() else "+s",
+                "-L" + dir,
+            ]
+
+        # For all compilers, `-Wl` is the presumed way to pass a
+        # compiler option to the linker
+        if sysconfig.get_config_var("GNULD") == "yes":
+            return consolidate_linker_args([
+                # Force RUNPATH instead of RPATH
+                "-Wl,--enable-new-dtags",
+                "-Wl,-rpath," + dir,
+            ])
+        else:
+            return "-Wl,-R" + dir
+
+    def library_option(self, lib):
+        return "-l" + lib
+
+    @staticmethod
+    def _library_root(dir):
+        """
+        macOS users can specify an alternate SDK using'-isysroot'.
+        Calculate the SDK root if it is specified.
+
+        Note that, as of Xcode 7, Apple SDKs may contain textual stub
+        libraries with .tbd extensions rather than the normal .dylib
+        shared libraries installed in /.  The Apple compiler tool
+        chain handles this transparently but it can cause problems
+        for programs that are being built with an SDK and searching
+        for specific libraries.  Callers of find_library_file need to
+        keep in mind that the base filename of the returned SDK library
+        file might have a different extension from that of the library
+        file installed on the running system, for example:
+          /Applications/Xcode.app/Contents/Developer/Platforms/
+              MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/
+              usr/lib/libedit.tbd
+        vs
+          /usr/lib/libedit.dylib
+        """
+        cflags = sysconfig.get_config_var('CFLAGS')
+        match = re.search(r'-isysroot\s*(\S+)', cflags)
+
+        apply_root = (
+            sys.platform == 'darwin'
+            and match
+            and (
+                dir.startswith('/System/')
+                or (dir.startswith('/usr/') and not dir.startswith('/usr/local/'))
+            )
+        )
+
+        return os.path.join(match.group(1), dir[1:]) if apply_root else dir
+
+    def find_library_file(self, dirs, lib, debug=False):
+        """
+        Second-guess the linker with not much hard
+        data to go on: GCC seems to prefer the shared library, so
+        assume that *all* Unix C compilers do,
+        ignoring even GCC's "-static" option.
+        """
+        lib_names = (
+            self.library_filename(lib, lib_type=type)
+            for type in 'dylib xcode_stub shared static'.split()
+        )
+
+        roots = map(self._library_root, dirs)
+
+        searched = itertools.starmap(os.path.join, itertools.product(roots, lib_names))
+
+        found = filter(os.path.exists, searched)
+
+        # Return None if it could not be found in any dir.
+        return next(found, None)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/util.py
new file mode 100644
index 00000000..83ad39e9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/util.py
@@ -0,0 +1,502 @@
+"""distutils.util
+
+Miscellaneous utility functions -- anything that doesn't fit into
+one of the other *util.py modules.
+"""
+
+from __future__ import annotations
+
+import functools
+import importlib.util
+import os
+import pathlib
+import re
+import string
+import subprocess
+import sys
+import sysconfig
+import tempfile
+
+from jaraco.functools import pass_none
+
+from ._log import log
+from ._modified import newer
+from .errors import DistutilsByteCompileError, DistutilsPlatformError
+from .spawn import spawn
+
+
+def get_host_platform() -> str:
+    """
+    Return a string that identifies the current platform. Use this
+    function to distinguish platform-specific build directories and
+    platform-specific built distributions.
+    """
+
+    # This function initially exposed platforms as defined in Python 3.9
+    # even with older Python versions when distutils was split out.
+    # Now it delegates to stdlib sysconfig.
+
+    return sysconfig.get_platform()
+
+
+def get_platform():
+    if os.name == 'nt':
+        TARGET_TO_PLAT = {
+            'x86': 'win32',
+            'x64': 'win-amd64',
+            'arm': 'win-arm32',
+            'arm64': 'win-arm64',
+        }
+        target = os.environ.get('VSCMD_ARG_TGT_ARCH')
+        return TARGET_TO_PLAT.get(target) or get_host_platform()
+    return get_host_platform()
+
+
+if sys.platform == 'darwin':
+    _syscfg_macosx_ver = None  # cache the version pulled from sysconfig
+MACOSX_VERSION_VAR = 'MACOSX_DEPLOYMENT_TARGET'
+
+
+def _clear_cached_macosx_ver():
+    """For testing only. Do not call."""
+    global _syscfg_macosx_ver
+    _syscfg_macosx_ver = None
+
+
+def get_macosx_target_ver_from_syscfg():
+    """Get the version of macOS latched in the Python interpreter configuration.
+    Returns the version as a string or None if can't obtain one. Cached."""
+    global _syscfg_macosx_ver
+    if _syscfg_macosx_ver is None:
+        from distutils import sysconfig
+
+        ver = sysconfig.get_config_var(MACOSX_VERSION_VAR) or ''
+        if ver:
+            _syscfg_macosx_ver = ver
+    return _syscfg_macosx_ver
+
+
+def get_macosx_target_ver():
+    """Return the version of macOS for which we are building.
+
+    The target version defaults to the version in sysconfig latched at time
+    the Python interpreter was built, unless overridden by an environment
+    variable. If neither source has a value, then None is returned"""
+
+    syscfg_ver = get_macosx_target_ver_from_syscfg()
+    env_ver = os.environ.get(MACOSX_VERSION_VAR)
+
+    if env_ver:
+        # Validate overridden version against sysconfig version, if have both.
+        # Ensure that the deployment target of the build process is not less
+        # than 10.3 if the interpreter was built for 10.3 or later.  This
+        # ensures extension modules are built with correct compatibility
+        # values, specifically LDSHARED which can use
+        # '-undefined dynamic_lookup' which only works on >= 10.3.
+        if (
+            syscfg_ver
+            and split_version(syscfg_ver) >= [10, 3]
+            and split_version(env_ver) < [10, 3]
+        ):
+            my_msg = (
+                '$' + MACOSX_VERSION_VAR + ' mismatch: '
+                f'now "{env_ver}" but "{syscfg_ver}" during configure; '
+                'must use 10.3 or later'
+            )
+            raise DistutilsPlatformError(my_msg)
+        return env_ver
+    return syscfg_ver
+
+
+def split_version(s):
+    """Convert a dot-separated string into a list of numbers for comparisons"""
+    return [int(n) for n in s.split('.')]
+
+
+@pass_none
+def convert_path(pathname: str | os.PathLike) -> str:
+    r"""
+    Allow for pathlib.Path inputs, coax to a native path string.
+
+    If None is passed, will just pass it through as
+    Setuptools relies on this behavior.
+
+    >>> convert_path(None) is None
+    True
+
+    Removes empty paths.
+
+    >>> convert_path('foo/./bar').replace('\\', '/')
+    'foo/bar'
+    """
+    return os.fspath(pathlib.PurePath(pathname))
+
+
+def change_root(new_root, pathname):
+    """Return 'pathname' with 'new_root' prepended.  If 'pathname' is
+    relative, this is equivalent to "os.path.join(new_root,pathname)".
+    Otherwise, it requires making 'pathname' relative and then joining the
+    two, which is tricky on DOS/Windows and Mac OS.
+    """
+    if os.name == 'posix':
+        if not os.path.isabs(pathname):
+            return os.path.join(new_root, pathname)
+        else:
+            return os.path.join(new_root, pathname[1:])
+
+    elif os.name == 'nt':
+        (drive, path) = os.path.splitdrive(pathname)
+        if path[0] == os.sep:
+            path = path[1:]
+        return os.path.join(new_root, path)
+
+    raise DistutilsPlatformError(f"nothing known about platform '{os.name}'")
+
+
+@functools.lru_cache
+def check_environ():
+    """Ensure that 'os.environ' has all the environment variables we
+    guarantee that users can use in config files, command-line options,
+    etc.  Currently this includes:
+      HOME - user's home directory (Unix only)
+      PLAT - description of the current platform, including hardware
+             and OS (see 'get_platform()')
+    """
+    if os.name == 'posix' and 'HOME' not in os.environ:
+        try:
+            import pwd
+
+            os.environ['HOME'] = pwd.getpwuid(os.getuid())[5]
+        except (ImportError, KeyError):
+            # bpo-10496: if the current user identifier doesn't exist in the
+            # password database, do nothing
+            pass
+
+    if 'PLAT' not in os.environ:
+        os.environ['PLAT'] = get_platform()
+
+
+def subst_vars(s, local_vars):
+    """
+    Perform variable substitution on 'string'.
+    Variables are indicated by format-style braces ("{var}").
+    Variable is substituted by the value found in the 'local_vars'
+    dictionary or in 'os.environ' if it's not in 'local_vars'.
+    'os.environ' is first checked/augmented to guarantee that it contains
+    certain values: see 'check_environ()'.  Raise ValueError for any
+    variables not found in either 'local_vars' or 'os.environ'.
+    """
+    check_environ()
+    lookup = dict(os.environ)
+    lookup.update((name, str(value)) for name, value in local_vars.items())
+    try:
+        return _subst_compat(s).format_map(lookup)
+    except KeyError as var:
+        raise ValueError(f"invalid variable {var}")
+
+
+def _subst_compat(s):
+    """
+    Replace shell/Perl-style variable substitution with
+    format-style. For compatibility.
+    """
+
+    def _subst(match):
+        return f'{{{match.group(1)}}}'
+
+    repl = re.sub(r'\$([a-zA-Z_][a-zA-Z_0-9]*)', _subst, s)
+    if repl != s:
+        import warnings
+
+        warnings.warn(
+            "shell/Perl-style substitutions are deprecated",
+            DeprecationWarning,
+        )
+    return repl
+
+
+def grok_environment_error(exc, prefix="error: "):
+    # Function kept for backward compatibility.
+    # Used to try clever things with EnvironmentErrors,
+    # but nowadays str(exception) produces good messages.
+    return prefix + str(exc)
+
+
+# Needed by 'split_quoted()'
+_wordchars_re = _squote_re = _dquote_re = None
+
+
+def _init_regex():
+    global _wordchars_re, _squote_re, _dquote_re
+    _wordchars_re = re.compile(rf'[^\\\'\"{string.whitespace} ]*')
+    _squote_re = re.compile(r"'(?:[^'\\]|\\.)*'")
+    _dquote_re = re.compile(r'"(?:[^"\\]|\\.)*"')
+
+
+def split_quoted(s):
+    """Split a string up according to Unix shell-like rules for quotes and
+    backslashes.  In short: words are delimited by spaces, as long as those
+    spaces are not escaped by a backslash, or inside a quoted string.
+    Single and double quotes are equivalent, and the quote characters can
+    be backslash-escaped.  The backslash is stripped from any two-character
+    escape sequence, leaving only the escaped character.  The quote
+    characters are stripped from any quoted string.  Returns a list of
+    words.
+    """
+
+    # This is a nice algorithm for splitting up a single string, since it
+    # doesn't require character-by-character examination.  It was a little
+    # bit of a brain-bender to get it working right, though...
+    if _wordchars_re is None:
+        _init_regex()
+
+    s = s.strip()
+    words = []
+    pos = 0
+
+    while s:
+        m = _wordchars_re.match(s, pos)
+        end = m.end()
+        if end == len(s):
+            words.append(s[:end])
+            break
+
+        if s[end] in string.whitespace:
+            # unescaped, unquoted whitespace: now
+            # we definitely have a word delimiter
+            words.append(s[:end])
+            s = s[end:].lstrip()
+            pos = 0
+
+        elif s[end] == '\\':
+            # preserve whatever is being escaped;
+            # will become part of the current word
+            s = s[:end] + s[end + 1 :]
+            pos = end + 1
+
+        else:
+            if s[end] == "'":  # slurp singly-quoted string
+                m = _squote_re.match(s, end)
+            elif s[end] == '"':  # slurp doubly-quoted string
+                m = _dquote_re.match(s, end)
+            else:
+                raise RuntimeError(f"this can't happen (bad char '{s[end]}')")
+
+            if m is None:
+                raise ValueError(f"bad string (mismatched {s[end]} quotes?)")
+
+            (beg, end) = m.span()
+            s = s[:beg] + s[beg + 1 : end - 1] + s[end:]
+            pos = m.end() - 2
+
+        if pos >= len(s):
+            words.append(s)
+            break
+
+    return words
+
+
+# split_quoted ()
+
+
+def execute(func, args, msg=None, verbose=False, dry_run=False):
+    """Perform some action that affects the outside world (eg.  by
+    writing to the filesystem).  Such actions are special because they
+    are disabled by the 'dry_run' flag.  This method takes care of all
+    that bureaucracy for you; all you have to do is supply the
+    function to call and an argument tuple for it (to embody the
+    "external action" being performed), and an optional message to
+    print.
+    """
+    if msg is None:
+        msg = f"{func.__name__}{args!r}"
+        if msg[-2:] == ',)':  # correct for singleton tuple
+            msg = msg[0:-2] + ')'
+
+    log.info(msg)
+    if not dry_run:
+        func(*args)
+
+
+def strtobool(val):
+    """Convert a string representation of truth to true (1) or false (0).
+
+    True values are 'y', 'yes', 't', 'true', 'on', and '1'; false values
+    are 'n', 'no', 'f', 'false', 'off', and '0'.  Raises ValueError if
+    'val' is anything else.
+    """
+    val = val.lower()
+    if val in ('y', 'yes', 't', 'true', 'on', '1'):
+        return 1
+    elif val in ('n', 'no', 'f', 'false', 'off', '0'):
+        return 0
+    else:
+        raise ValueError(f"invalid truth value {val!r}")
+
+
+def byte_compile(  # noqa: C901
+    py_files,
+    optimize=0,
+    force=False,
+    prefix=None,
+    base_dir=None,
+    verbose=True,
+    dry_run=False,
+    direct=None,
+):
+    """Byte-compile a collection of Python source files to .pyc
+    files in a __pycache__ subdirectory.  'py_files' is a list
+    of files to compile; any files that don't end in ".py" are silently
+    skipped.  'optimize' must be one of the following:
+      0 - don't optimize
+      1 - normal optimization (like "python -O")
+      2 - extra optimization (like "python -OO")
+    If 'force' is true, all files are recompiled regardless of
+    timestamps.
+
+    The source filename encoded in each bytecode file defaults to the
+    filenames listed in 'py_files'; you can modify these with 'prefix' and
+    'basedir'.  'prefix' is a string that will be stripped off of each
+    source filename, and 'base_dir' is a directory name that will be
+    prepended (after 'prefix' is stripped).  You can supply either or both
+    (or neither) of 'prefix' and 'base_dir', as you wish.
+
+    If 'dry_run' is true, doesn't actually do anything that would
+    affect the filesystem.
+
+    Byte-compilation is either done directly in this interpreter process
+    with the standard py_compile module, or indirectly by writing a
+    temporary script and executing it.  Normally, you should let
+    'byte_compile()' figure out to use direct compilation or not (see
+    the source for details).  The 'direct' flag is used by the script
+    generated in indirect mode; unless you know what you're doing, leave
+    it set to None.
+    """
+
+    # nothing is done if sys.dont_write_bytecode is True
+    if sys.dont_write_bytecode:
+        raise DistutilsByteCompileError('byte-compiling is disabled.')
+
+    # First, if the caller didn't force us into direct or indirect mode,
+    # figure out which mode we should be in.  We take a conservative
+    # approach: choose direct mode *only* if the current interpreter is
+    # in debug mode and optimize is 0.  If we're not in debug mode (-O
+    # or -OO), we don't know which level of optimization this
+    # interpreter is running with, so we can't do direct
+    # byte-compilation and be certain that it's the right thing.  Thus,
+    # always compile indirectly if the current interpreter is in either
+    # optimize mode, or if either optimization level was requested by
+    # the caller.
+    if direct is None:
+        direct = __debug__ and optimize == 0
+
+    # "Indirect" byte-compilation: write a temporary script and then
+    # run it with the appropriate flags.
+    if not direct:
+        (script_fd, script_name) = tempfile.mkstemp(".py")
+        log.info("writing byte-compilation script '%s'", script_name)
+        if not dry_run:
+            script = os.fdopen(script_fd, "w", encoding='utf-8')
+
+            with script:
+                script.write(
+                    """\
+from distutils.util import byte_compile
+files = [
+"""
+                )
+
+                # XXX would be nice to write absolute filenames, just for
+                # safety's sake (script should be more robust in the face of
+                # chdir'ing before running it).  But this requires abspath'ing
+                # 'prefix' as well, and that breaks the hack in build_lib's
+                # 'byte_compile()' method that carefully tacks on a trailing
+                # slash (os.sep really) to make sure the prefix here is "just
+                # right".  This whole prefix business is rather delicate -- the
+                # problem is that it's really a directory, but I'm treating it
+                # as a dumb string, so trailing slashes and so forth matter.
+
+                script.write(",\n".join(map(repr, py_files)) + "]\n")
+                script.write(
+                    f"""
+byte_compile(files, optimize={optimize!r}, force={force!r},
+             prefix={prefix!r}, base_dir={base_dir!r},
+             verbose={verbose!r}, dry_run=False,
+             direct=True)
+"""
+                )
+
+        cmd = [sys.executable]
+        cmd.extend(subprocess._optim_args_from_interpreter_flags())
+        cmd.append(script_name)
+        spawn(cmd, dry_run=dry_run)
+        execute(os.remove, (script_name,), f"removing {script_name}", dry_run=dry_run)
+
+    # "Direct" byte-compilation: use the py_compile module to compile
+    # right here, right now.  Note that the script generated in indirect
+    # mode simply calls 'byte_compile()' in direct mode, a weird sort of
+    # cross-process recursion.  Hey, it works!
+    else:
+        from py_compile import compile
+
+        for file in py_files:
+            if file[-3:] != ".py":
+                # This lets us be lazy and not filter filenames in
+                # the "install_lib" command.
+                continue
+
+            # Terminology from the py_compile module:
+            #   cfile - byte-compiled file
+            #   dfile - purported source filename (same as 'file' by default)
+            if optimize >= 0:
+                opt = '' if optimize == 0 else optimize
+                cfile = importlib.util.cache_from_source(file, optimization=opt)
+            else:
+                cfile = importlib.util.cache_from_source(file)
+            dfile = file
+            if prefix:
+                if file[: len(prefix)] != prefix:
+                    raise ValueError(
+                        f"invalid prefix: filename {file!r} doesn't start with {prefix!r}"
+                    )
+                dfile = dfile[len(prefix) :]
+            if base_dir:
+                dfile = os.path.join(base_dir, dfile)
+
+            cfile_base = os.path.basename(cfile)
+            if direct:
+                if force or newer(file, cfile):
+                    log.info("byte-compiling %s to %s", file, cfile_base)
+                    if not dry_run:
+                        compile(file, cfile, dfile)
+                else:
+                    log.debug("skipping byte-compilation of %s to %s", file, cfile_base)
+
+
+def rfc822_escape(header):
+    """Return a version of the string escaped for inclusion in an
+    RFC-822 header, by ensuring there are 8 spaces space after each newline.
+    """
+    indent = 8 * " "
+    lines = header.splitlines(keepends=True)
+
+    # Emulate the behaviour of `str.split`
+    # (the terminal line break in `splitlines` does not result in an extra line):
+    ends_in_newline = lines and lines[-1].splitlines()[0] != lines[-1]
+    suffix = indent if ends_in_newline else ""
+
+    return indent.join(lines) + suffix
+
+
+def is_mingw():
+    """Returns True if the current platform is mingw.
+
+    Python compiled with Mingw-w64 has sys.platform == 'win32' and
+    get_platform() starts with 'mingw'.
+    """
+    return sys.platform == 'win32' and get_platform().startswith('mingw')
+
+
+def is_freethreaded():
+    """Return True if the Python interpreter is built with free threading support."""
+    return bool(sysconfig.get_config_var('Py_GIL_DISABLED'))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/version.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/version.py
new file mode 100644
index 00000000..2223ee9c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/version.py
@@ -0,0 +1,348 @@
+#
+# distutils/version.py
+#
+# Implements multiple version numbering conventions for the
+# Python Module Distribution Utilities.
+#
+# $Id$
+#
+
+"""Provides classes to represent module version numbers (one class for
+each style of version numbering).  There are currently two such classes
+implemented: StrictVersion and LooseVersion.
+
+Every version number class implements the following interface:
+  * the 'parse' method takes a string and parses it to some internal
+    representation; if the string is an invalid version number,
+    'parse' raises a ValueError exception
+  * the class constructor takes an optional string argument which,
+    if supplied, is passed to 'parse'
+  * __str__ reconstructs the string that was passed to 'parse' (or
+    an equivalent string -- ie. one that will generate an equivalent
+    version number instance)
+  * __repr__ generates Python code to recreate the version number instance
+  * _cmp compares the current instance with either another instance
+    of the same class or a string (which will be parsed to an instance
+    of the same class, thus must follow the same rules)
+"""
+
+import contextlib
+import re
+import warnings
+
+
+@contextlib.contextmanager
+def suppress_known_deprecation():
+    with warnings.catch_warnings(record=True) as ctx:
+        warnings.filterwarnings(
+            action='default',
+            category=DeprecationWarning,
+            message="distutils Version classes are deprecated.",
+        )
+        yield ctx
+
+
+class Version:
+    """Abstract base class for version numbering classes.  Just provides
+    constructor (__init__) and reproducer (__repr__), because those
+    seem to be the same for all version numbering classes; and route
+    rich comparisons to _cmp.
+    """
+
+    def __init__(self, vstring=None):
+        if vstring:
+            self.parse(vstring)
+        warnings.warn(
+            "distutils Version classes are deprecated. Use packaging.version instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+    def __repr__(self):
+        return f"{self.__class__.__name__} ('{self}')"
+
+    def __eq__(self, other):
+        c = self._cmp(other)
+        if c is NotImplemented:
+            return c
+        return c == 0
+
+    def __lt__(self, other):
+        c = self._cmp(other)
+        if c is NotImplemented:
+            return c
+        return c < 0
+
+    def __le__(self, other):
+        c = self._cmp(other)
+        if c is NotImplemented:
+            return c
+        return c <= 0
+
+    def __gt__(self, other):
+        c = self._cmp(other)
+        if c is NotImplemented:
+            return c
+        return c > 0
+
+    def __ge__(self, other):
+        c = self._cmp(other)
+        if c is NotImplemented:
+            return c
+        return c >= 0
+
+
+# Interface for version-number classes -- must be implemented
+# by the following classes (the concrete ones -- Version should
+# be treated as an abstract class).
+#    __init__ (string) - create and take same action as 'parse'
+#                        (string parameter is optional)
+#    parse (string)    - convert a string representation to whatever
+#                        internal representation is appropriate for
+#                        this style of version numbering
+#    __str__ (self)    - convert back to a string; should be very similar
+#                        (if not identical to) the string supplied to parse
+#    __repr__ (self)   - generate Python code to recreate
+#                        the instance
+#    _cmp (self, other) - compare two version numbers ('other' may
+#                        be an unparsed version string, or another
+#                        instance of your version class)
+
+
+class StrictVersion(Version):
+    """Version numbering for anal retentives and software idealists.
+    Implements the standard interface for version number classes as
+    described above.  A version number consists of two or three
+    dot-separated numeric components, with an optional "pre-release" tag
+    on the end.  The pre-release tag consists of the letter 'a' or 'b'
+    followed by a number.  If the numeric components of two version
+    numbers are equal, then one with a pre-release tag will always
+    be deemed earlier (lesser) than one without.
+
+    The following are valid version numbers (shown in the order that
+    would be obtained by sorting according to the supplied cmp function):
+
+        0.4       0.4.0  (these two are equivalent)
+        0.4.1
+        0.5a1
+        0.5b3
+        0.5
+        0.9.6
+        1.0
+        1.0.4a3
+        1.0.4b1
+        1.0.4
+
+    The following are examples of invalid version numbers:
+
+        1
+        2.7.2.2
+        1.3.a4
+        1.3pl1
+        1.3c4
+
+    The rationale for this version numbering system will be explained
+    in the distutils documentation.
+    """
+
+    version_re = re.compile(
+        r'^(\d+) \. (\d+) (\. (\d+))? ([ab](\d+))?$', re.VERBOSE | re.ASCII
+    )
+
+    def parse(self, vstring):
+        match = self.version_re.match(vstring)
+        if not match:
+            raise ValueError(f"invalid version number '{vstring}'")
+
+        (major, minor, patch, prerelease, prerelease_num) = match.group(1, 2, 4, 5, 6)
+
+        if patch:
+            self.version = tuple(map(int, [major, minor, patch]))
+        else:
+            self.version = tuple(map(int, [major, minor])) + (0,)
+
+        if prerelease:
+            self.prerelease = (prerelease[0], int(prerelease_num))
+        else:
+            self.prerelease = None
+
+    def __str__(self):
+        if self.version[2] == 0:
+            vstring = '.'.join(map(str, self.version[0:2]))
+        else:
+            vstring = '.'.join(map(str, self.version))
+
+        if self.prerelease:
+            vstring = vstring + self.prerelease[0] + str(self.prerelease[1])
+
+        return vstring
+
+    def _cmp(self, other):
+        if isinstance(other, str):
+            with suppress_known_deprecation():
+                other = StrictVersion(other)
+        elif not isinstance(other, StrictVersion):
+            return NotImplemented
+
+        if self.version == other.version:
+            # versions match; pre-release drives the comparison
+            return self._cmp_prerelease(other)
+
+        return -1 if self.version < other.version else 1
+
+    def _cmp_prerelease(self, other):
+        """
+        case 1: self has prerelease, other doesn't; other is greater
+        case 2: self doesn't have prerelease, other does: self is greater
+        case 3: both or neither have prerelease: compare them!
+        """
+        if self.prerelease and not other.prerelease:
+            return -1
+        elif not self.prerelease and other.prerelease:
+            return 1
+
+        if self.prerelease == other.prerelease:
+            return 0
+        elif self.prerelease < other.prerelease:
+            return -1
+        else:
+            return 1
+
+
+# end class StrictVersion
+
+
+# The rules according to Greg Stein:
+# 1) a version number has 1 or more numbers separated by a period or by
+#    sequences of letters. If only periods, then these are compared
+#    left-to-right to determine an ordering.
+# 2) sequences of letters are part of the tuple for comparison and are
+#    compared lexicographically
+# 3) recognize the numeric components may have leading zeroes
+#
+# The LooseVersion class below implements these rules: a version number
+# string is split up into a tuple of integer and string components, and
+# comparison is a simple tuple comparison.  This means that version
+# numbers behave in a predictable and obvious way, but a way that might
+# not necessarily be how people *want* version numbers to behave.  There
+# wouldn't be a problem if people could stick to purely numeric version
+# numbers: just split on period and compare the numbers as tuples.
+# However, people insist on putting letters into their version numbers;
+# the most common purpose seems to be:
+#   - indicating a "pre-release" version
+#     ('alpha', 'beta', 'a', 'b', 'pre', 'p')
+#   - indicating a post-release patch ('p', 'pl', 'patch')
+# but of course this can't cover all version number schemes, and there's
+# no way to know what a programmer means without asking him.
+#
+# The problem is what to do with letters (and other non-numeric
+# characters) in a version number.  The current implementation does the
+# obvious and predictable thing: keep them as strings and compare
+# lexically within a tuple comparison.  This has the desired effect if
+# an appended letter sequence implies something "post-release":
+# eg. "0.99" < "0.99pl14" < "1.0", and "5.001" < "5.001m" < "5.002".
+#
+# However, if letters in a version number imply a pre-release version,
+# the "obvious" thing isn't correct.  Eg. you would expect that
+# "1.5.1" < "1.5.2a2" < "1.5.2", but under the tuple/lexical comparison
+# implemented here, this just isn't so.
+#
+# Two possible solutions come to mind.  The first is to tie the
+# comparison algorithm to a particular set of semantic rules, as has
+# been done in the StrictVersion class above.  This works great as long
+# as everyone can go along with bondage and discipline.  Hopefully a
+# (large) subset of Python module programmers will agree that the
+# particular flavour of bondage and discipline provided by StrictVersion
+# provides enough benefit to be worth using, and will submit their
+# version numbering scheme to its domination.  The free-thinking
+# anarchists in the lot will never give in, though, and something needs
+# to be done to accommodate them.
+#
+# Perhaps a "moderately strict" version class could be implemented that
+# lets almost anything slide (syntactically), and makes some heuristic
+# assumptions about non-digits in version number strings.  This could
+# sink into special-case-hell, though; if I was as talented and
+# idiosyncratic as Larry Wall, I'd go ahead and implement a class that
+# somehow knows that "1.2.1" < "1.2.2a2" < "1.2.2" < "1.2.2pl3", and is
+# just as happy dealing with things like "2g6" and "1.13++".  I don't
+# think I'm smart enough to do it right though.
+#
+# In any case, I've coded the test suite for this module (see
+# ../test/test_version.py) specifically to fail on things like comparing
+# "1.2a2" and "1.2".  That's not because the *code* is doing anything
+# wrong, it's because the simple, obvious design doesn't match my
+# complicated, hairy expectations for real-world version numbers.  It
+# would be a snap to fix the test suite to say, "Yep, LooseVersion does
+# the Right Thing" (ie. the code matches the conception).  But I'd rather
+# have a conception that matches common notions about version numbers.
+
+
+class LooseVersion(Version):
+    """Version numbering for anarchists and software realists.
+    Implements the standard interface for version number classes as
+    described above.  A version number consists of a series of numbers,
+    separated by either periods or strings of letters.  When comparing
+    version numbers, the numeric components will be compared
+    numerically, and the alphabetic components lexically.  The following
+    are all valid version numbers, in no particular order:
+
+        1.5.1
+        1.5.2b2
+        161
+        3.10a
+        8.02
+        3.4j
+        1996.07.12
+        3.2.pl0
+        3.1.1.6
+        2g6
+        11g
+        0.960923
+        2.2beta29
+        1.13++
+        5.5.kw
+        2.0b1pl0
+
+    In fact, there is no such thing as an invalid version number under
+    this scheme; the rules for comparison are simple and predictable,
+    but may not always give the results you want (for some definition
+    of "want").
+    """
+
+    component_re = re.compile(r'(\d+ | [a-z]+ | \.)', re.VERBOSE)
+
+    def parse(self, vstring):
+        # I've given up on thinking I can reconstruct the version string
+        # from the parsed tuple -- so I just store the string here for
+        # use by __str__
+        self.vstring = vstring
+        components = [x for x in self.component_re.split(vstring) if x and x != '.']
+        for i, obj in enumerate(components):
+            try:
+                components[i] = int(obj)
+            except ValueError:
+                pass
+
+        self.version = components
+
+    def __str__(self):
+        return self.vstring
+
+    def __repr__(self):
+        return f"LooseVersion ('{self}')"
+
+    def _cmp(self, other):
+        if isinstance(other, str):
+            other = LooseVersion(other)
+        elif not isinstance(other, LooseVersion):
+            return NotImplemented
+
+        if self.version == other.version:
+            return 0
+        if self.version < other.version:
+            return -1
+        if self.version > other.version:
+            return 1
+
+
+# end class LooseVersion
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/versionpredicate.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/versionpredicate.py
new file mode 100644
index 00000000..fe31b0ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/versionpredicate.py
@@ -0,0 +1,175 @@
+"""Module for parsing and testing package version predicate strings."""
+
+import operator
+import re
+
+from . import version
+
+re_validPackage = re.compile(r"(?i)^\s*([a-z_]\w*(?:\.[a-z_]\w*)*)(.*)", re.ASCII)
+# (package) (rest)
+
+re_paren = re.compile(r"^\s*\((.*)\)\s*$")  # (list) inside of parentheses
+re_splitComparison = re.compile(r"^\s*(<=|>=|<|>|!=|==)\s*([^\s,]+)\s*$")
+# (comp) (version)
+
+
+def splitUp(pred):
+    """Parse a single version comparison.
+
+    Return (comparison string, StrictVersion)
+    """
+    res = re_splitComparison.match(pred)
+    if not res:
+        raise ValueError(f"bad package restriction syntax: {pred!r}")
+    comp, verStr = res.groups()
+    with version.suppress_known_deprecation():
+        other = version.StrictVersion(verStr)
+    return (comp, other)
+
+
+compmap = {
+    "<": operator.lt,
+    "<=": operator.le,
+    "==": operator.eq,
+    ">": operator.gt,
+    ">=": operator.ge,
+    "!=": operator.ne,
+}
+
+
+class VersionPredicate:
+    """Parse and test package version predicates.
+
+    >>> v = VersionPredicate('pyepat.abc (>1.0, <3333.3a1, !=1555.1b3)')
+
+    The `name` attribute provides the full dotted name that is given::
+
+    >>> v.name
+    'pyepat.abc'
+
+    The str() of a `VersionPredicate` provides a normalized
+    human-readable version of the expression::
+
+    >>> print(v)
+    pyepat.abc (> 1.0, < 3333.3a1, != 1555.1b3)
+
+    The `satisfied_by()` method can be used to determine with a given
+    version number is included in the set described by the version
+    restrictions::
+
+    >>> v.satisfied_by('1.1')
+    True
+    >>> v.satisfied_by('1.4')
+    True
+    >>> v.satisfied_by('1.0')
+    False
+    >>> v.satisfied_by('4444.4')
+    False
+    >>> v.satisfied_by('1555.1b3')
+    False
+
+    `VersionPredicate` is flexible in accepting extra whitespace::
+
+    >>> v = VersionPredicate(' pat( ==  0.1  )  ')
+    >>> v.name
+    'pat'
+    >>> v.satisfied_by('0.1')
+    True
+    >>> v.satisfied_by('0.2')
+    False
+
+    If any version numbers passed in do not conform to the
+    restrictions of `StrictVersion`, a `ValueError` is raised::
+
+    >>> v = VersionPredicate('p1.p2.p3.p4(>=1.0, <=1.3a1, !=1.2zb3)')
+    Traceback (most recent call last):
+      ...
+    ValueError: invalid version number '1.2zb3'
+
+    It the module or package name given does not conform to what's
+    allowed as a legal module or package name, `ValueError` is
+    raised::
+
+    >>> v = VersionPredicate('foo-bar')
+    Traceback (most recent call last):
+      ...
+    ValueError: expected parenthesized list: '-bar'
+
+    >>> v = VersionPredicate('foo bar (12.21)')
+    Traceback (most recent call last):
+      ...
+    ValueError: expected parenthesized list: 'bar (12.21)'
+
+    """
+
+    def __init__(self, versionPredicateStr):
+        """Parse a version predicate string."""
+        # Fields:
+        #    name:  package name
+        #    pred:  list of (comparison string, StrictVersion)
+
+        versionPredicateStr = versionPredicateStr.strip()
+        if not versionPredicateStr:
+            raise ValueError("empty package restriction")
+        match = re_validPackage.match(versionPredicateStr)
+        if not match:
+            raise ValueError(f"bad package name in {versionPredicateStr!r}")
+        self.name, paren = match.groups()
+        paren = paren.strip()
+        if paren:
+            match = re_paren.match(paren)
+            if not match:
+                raise ValueError(f"expected parenthesized list: {paren!r}")
+            str = match.groups()[0]
+            self.pred = [splitUp(aPred) for aPred in str.split(",")]
+            if not self.pred:
+                raise ValueError(f"empty parenthesized list in {versionPredicateStr!r}")
+        else:
+            self.pred = []
+
+    def __str__(self):
+        if self.pred:
+            seq = [cond + " " + str(ver) for cond, ver in self.pred]
+            return self.name + " (" + ", ".join(seq) + ")"
+        else:
+            return self.name
+
+    def satisfied_by(self, version):
+        """True if version is compatible with all the predicates in self.
+        The parameter version must be acceptable to the StrictVersion
+        constructor.  It may be either a string or StrictVersion.
+        """
+        for cond, ver in self.pred:
+            if not compmap[cond](version, ver):
+                return False
+        return True
+
+
+_provision_rx = None
+
+
+def split_provision(value):
+    """Return the name and optional version number of a provision.
+
+    The version number, if given, will be returned as a `StrictVersion`
+    instance, otherwise it will be `None`.
+
+    >>> split_provision('mypkg')
+    ('mypkg', None)
+    >>> split_provision(' mypkg( 1.2 ) ')
+    ('mypkg', StrictVersion ('1.2'))
+    """
+    global _provision_rx
+    if _provision_rx is None:
+        _provision_rx = re.compile(
+            r"([a-zA-Z_]\w*(?:\.[a-zA-Z_]\w*)*)(?:\s*\(\s*([^)\s]+)\s*\))?$", re.ASCII
+        )
+    value = value.strip()
+    m = _provision_rx.match(value)
+    if not m:
+        raise ValueError(f"illegal provides specification: {value!r}")
+    ver = m.group(2) or None
+    if ver:
+        with version.suppress_known_deprecation():
+            ver = version.StrictVersion(ver)
+    return m.group(1), ver
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/zosccompiler.py b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/zosccompiler.py
new file mode 100644
index 00000000..af1e7fa5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_distutils/zosccompiler.py
@@ -0,0 +1,229 @@
+"""distutils.zosccompiler
+
+Contains the selection of the c & c++ compilers on z/OS. There are several
+different c compilers on z/OS, all of them are optional, so the correct
+one needs to be chosen based on the users input. This is compatible with
+the following compilers:
+
+IBM C/C++ For Open Enterprise Languages on z/OS 2.0
+IBM Open XL C/C++ 1.1 for z/OS
+IBM XL C/C++ V2.4.1 for z/OS 2.4 and 2.5
+IBM z/OS XL C/C++
+"""
+
+import os
+
+from . import sysconfig
+from .errors import CompileError, DistutilsExecError
+from .unixccompiler import UnixCCompiler
+
+_cc_args = {
+    'ibm-openxl': [
+        '-m64',
+        '-fvisibility=default',
+        '-fzos-le-char-mode=ascii',
+        '-fno-short-enums',
+    ],
+    'ibm-xlclang': [
+        '-q64',
+        '-qexportall',
+        '-qascii',
+        '-qstrict',
+        '-qnocsect',
+        '-Wa,asa,goff',
+        '-Wa,xplink',
+        '-qgonumber',
+        '-qenum=int',
+        '-Wc,DLL',
+    ],
+    'ibm-xlc': [
+        '-q64',
+        '-qexportall',
+        '-qascii',
+        '-qstrict',
+        '-qnocsect',
+        '-Wa,asa,goff',
+        '-Wa,xplink',
+        '-qgonumber',
+        '-qenum=int',
+        '-Wc,DLL',
+        '-qlanglvl=extc99',
+    ],
+}
+
+_cxx_args = {
+    'ibm-openxl': [
+        '-m64',
+        '-fvisibility=default',
+        '-fzos-le-char-mode=ascii',
+        '-fno-short-enums',
+    ],
+    'ibm-xlclang': [
+        '-q64',
+        '-qexportall',
+        '-qascii',
+        '-qstrict',
+        '-qnocsect',
+        '-Wa,asa,goff',
+        '-Wa,xplink',
+        '-qgonumber',
+        '-qenum=int',
+        '-Wc,DLL',
+    ],
+    'ibm-xlc': [
+        '-q64',
+        '-qexportall',
+        '-qascii',
+        '-qstrict',
+        '-qnocsect',
+        '-Wa,asa,goff',
+        '-Wa,xplink',
+        '-qgonumber',
+        '-qenum=int',
+        '-Wc,DLL',
+        '-qlanglvl=extended0x',
+    ],
+}
+
+_asm_args = {
+    'ibm-openxl': ['-fasm', '-fno-integrated-as', '-Wa,--ASA', '-Wa,--GOFF'],
+    'ibm-xlclang': [],
+    'ibm-xlc': [],
+}
+
+_ld_args = {
+    'ibm-openxl': [],
+    'ibm-xlclang': ['-Wl,dll', '-q64'],
+    'ibm-xlc': ['-Wl,dll', '-q64'],
+}
+
+
+# Python on z/OS is built with no compiler specific options in it's CFLAGS.
+# But each compiler requires it's own specific options to build successfully,
+# though some of the options are common between them
+class zOSCCompiler(UnixCCompiler):
+    src_extensions = ['.c', '.C', '.cc', '.cxx', '.cpp', '.m', '.s']
+    _cpp_extensions = ['.cc', '.cpp', '.cxx', '.C']
+    _asm_extensions = ['.s']
+
+    def _get_zos_compiler_name(self):
+        zos_compiler_names = [
+            os.path.basename(binary)
+            for envvar in ('CC', 'CXX', 'LDSHARED')
+            if (binary := os.environ.get(envvar, None))
+        ]
+        if len(zos_compiler_names) == 0:
+            return 'ibm-openxl'
+
+        zos_compilers = {}
+        for compiler in (
+            'ibm-clang',
+            'ibm-clang64',
+            'ibm-clang++',
+            'ibm-clang++64',
+            'clang',
+            'clang++',
+            'clang-14',
+        ):
+            zos_compilers[compiler] = 'ibm-openxl'
+
+        for compiler in ('xlclang', 'xlclang++', 'njsc', 'njsc++'):
+            zos_compilers[compiler] = 'ibm-xlclang'
+
+        for compiler in ('xlc', 'xlC', 'xlc++'):
+            zos_compilers[compiler] = 'ibm-xlc'
+
+        return zos_compilers.get(zos_compiler_names[0], 'ibm-openxl')
+
+    def __init__(self, verbose=False, dry_run=False, force=False):
+        super().__init__(verbose, dry_run, force)
+        self.zos_compiler = self._get_zos_compiler_name()
+        sysconfig.customize_compiler(self)
+
+    def _compile(self, obj, src, ext, cc_args, extra_postargs, pp_opts):
+        local_args = []
+        if ext in self._cpp_extensions:
+            compiler = self.compiler_cxx
+            local_args.extend(_cxx_args[self.zos_compiler])
+        elif ext in self._asm_extensions:
+            compiler = self.compiler_so
+            local_args.extend(_cc_args[self.zos_compiler])
+            local_args.extend(_asm_args[self.zos_compiler])
+        else:
+            compiler = self.compiler_so
+            local_args.extend(_cc_args[self.zos_compiler])
+        local_args.extend(cc_args)
+
+        try:
+            self.spawn(compiler + local_args + [src, '-o', obj] + extra_postargs)
+        except DistutilsExecError as msg:
+            raise CompileError(msg)
+
+    def runtime_library_dir_option(self, dir):
+        return '-L' + dir
+
+    def link(
+        self,
+        target_desc,
+        objects,
+        output_filename,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug=False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ):
+        # For a built module to use functions from cpython, it needs to use Pythons
+        # side deck file. The side deck is located beside the libpython3.xx.so
+        ldversion = sysconfig.get_config_var('LDVERSION')
+        if sysconfig.python_build:
+            side_deck_path = os.path.join(
+                sysconfig.get_config_var('abs_builddir'),
+                f'libpython{ldversion}.x',
+            )
+        else:
+            side_deck_path = os.path.join(
+                sysconfig.get_config_var('installed_base'),
+                sysconfig.get_config_var('platlibdir'),
+                f'libpython{ldversion}.x',
+            )
+
+        if os.path.exists(side_deck_path):
+            if extra_postargs:
+                extra_postargs.append(side_deck_path)
+            else:
+                extra_postargs = [side_deck_path]
+
+        # Check and replace libraries included side deck files
+        if runtime_library_dirs:
+            for dir in runtime_library_dirs:
+                for library in libraries[:]:
+                    library_side_deck = os.path.join(dir, f'{library}.x')
+                    if os.path.exists(library_side_deck):
+                        libraries.remove(library)
+                        extra_postargs.append(library_side_deck)
+                        break
+
+        # Any required ld args for the given compiler
+        extra_postargs.extend(_ld_args[self.zos_compiler])
+
+        super().link(
+            target_desc,
+            objects,
+            output_filename,
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            export_symbols,
+            debug,
+            extra_preargs,
+            extra_postargs,
+            build_temp,
+            target_lang,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_entry_points.py b/backend/venv/lib/python3.12/site-packages/setuptools/_entry_points.py
new file mode 100644
index 00000000..e785fc7d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_entry_points.py
@@ -0,0 +1,90 @@
+import functools
+import itertools
+import operator
+
+from jaraco.functools import pass_none
+from jaraco.text import yield_lines
+from more_itertools import consume
+
+from ._importlib import metadata
+from ._itertools import ensure_unique
+from .errors import OptionError
+
+
+def ensure_valid(ep):
+    """
+    Exercise one of the dynamic properties to trigger
+    the pattern match.
+    """
+    try:
+        ep.extras
+    except (AttributeError, AssertionError) as ex:
+        # Why both? See https://github.com/python/importlib_metadata/issues/488
+        msg = (
+            f"Problems to parse {ep}.\nPlease ensure entry-point follows the spec: "
+            "https://packaging.python.org/en/latest/specifications/entry-points/"
+        )
+        raise OptionError(msg) from ex
+
+
+def load_group(value, group):
+    """
+    Given a value of an entry point or series of entry points,
+    return each as an EntryPoint.
+    """
+    # normalize to a single sequence of lines
+    lines = yield_lines(value)
+    text = f'[{group}]\n' + '\n'.join(lines)
+    return metadata.EntryPoints._from_text(text)
+
+
+def by_group_and_name(ep):
+    return ep.group, ep.name
+
+
+def validate(eps: metadata.EntryPoints):
+    """
+    Ensure entry points are unique by group and name and validate each.
+    """
+    consume(map(ensure_valid, ensure_unique(eps, key=by_group_and_name)))
+    return eps
+
+
+@functools.singledispatch
+def load(eps):
+    """
+    Given a Distribution.entry_points, produce EntryPoints.
+    """
+    groups = itertools.chain.from_iterable(
+        load_group(value, group) for group, value in eps.items()
+    )
+    return validate(metadata.EntryPoints(groups))
+
+
+@load.register(str)
+def _(eps):
+    r"""
+    >>> ep, = load('[console_scripts]\nfoo=bar')
+    >>> ep.group
+    'console_scripts'
+    >>> ep.name
+    'foo'
+    >>> ep.value
+    'bar'
+    """
+    return validate(metadata.EntryPoints(metadata.EntryPoints._from_text(eps)))
+
+
+load.register(type(None), lambda x: x)
+
+
+@pass_none
+def render(eps: metadata.EntryPoints):
+    by_group = operator.attrgetter('group')
+    groups = itertools.groupby(sorted(eps, key=by_group), by_group)
+
+    return '\n'.join(f'[{group}]\n{render_items(items)}\n' for group, items in groups)
+
+
+def render_items(eps):
+    return '\n'.join(f'{ep.name} = {ep.value}' for ep in sorted(eps))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_imp.py b/backend/venv/lib/python3.12/site-packages/setuptools/_imp.py
new file mode 100644
index 00000000..f1d9f292
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_imp.py
@@ -0,0 +1,87 @@
+"""
+Re-implementation of find_module and get_frozen_object
+from the deprecated imp module.
+"""
+
+import importlib.machinery
+import importlib.util
+import os
+import tokenize
+from importlib.util import module_from_spec
+
+PY_SOURCE = 1
+PY_COMPILED = 2
+C_EXTENSION = 3
+C_BUILTIN = 6
+PY_FROZEN = 7
+
+
+def find_spec(module, paths):
+    finder = (
+        importlib.machinery.PathFinder().find_spec
+        if isinstance(paths, list)
+        else importlib.util.find_spec
+    )
+    return finder(module, paths)
+
+
+def find_module(module, paths=None):
+    """Just like 'imp.find_module()', but with package support"""
+    spec = find_spec(module, paths)
+    if spec is None:
+        raise ImportError(f"Can't find {module}")
+    if not spec.has_location and hasattr(spec, 'submodule_search_locations'):
+        spec = importlib.util.spec_from_loader('__init__.py', spec.loader)
+
+    kind = -1
+    file = None
+    static = isinstance(spec.loader, type)
+    if (
+        spec.origin == 'frozen'
+        or static
+        and issubclass(spec.loader, importlib.machinery.FrozenImporter)
+    ):
+        kind = PY_FROZEN
+        path = None  # imp compabilty
+        suffix = mode = ''  # imp compatibility
+    elif (
+        spec.origin == 'built-in'
+        or static
+        and issubclass(spec.loader, importlib.machinery.BuiltinImporter)
+    ):
+        kind = C_BUILTIN
+        path = None  # imp compabilty
+        suffix = mode = ''  # imp compatibility
+    elif spec.has_location:
+        path = spec.origin
+        suffix = os.path.splitext(path)[1]
+        mode = 'r' if suffix in importlib.machinery.SOURCE_SUFFIXES else 'rb'
+
+        if suffix in importlib.machinery.SOURCE_SUFFIXES:
+            kind = PY_SOURCE
+            file = tokenize.open(path)
+        elif suffix in importlib.machinery.BYTECODE_SUFFIXES:
+            kind = PY_COMPILED
+            file = open(path, 'rb')
+        elif suffix in importlib.machinery.EXTENSION_SUFFIXES:
+            kind = C_EXTENSION
+
+    else:
+        path = None
+        suffix = mode = ''
+
+    return file, path, (suffix, mode, kind)
+
+
+def get_frozen_object(module, paths=None):
+    spec = find_spec(module, paths)
+    if not spec:
+        raise ImportError(f"Can't find {module}")
+    return spec.loader.get_code(module)
+
+
+def get_module(module, paths, info):
+    spec = find_spec(module, paths)
+    if not spec:
+        raise ImportError(f"Can't find {module}")
+    return module_from_spec(spec)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_importlib.py b/backend/venv/lib/python3.12/site-packages/setuptools/_importlib.py
new file mode 100644
index 00000000..ce0fd526
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_importlib.py
@@ -0,0 +1,9 @@
+import sys
+
+if sys.version_info < (3, 10):
+    import importlib_metadata as metadata  # pragma: no cover
+else:
+    import importlib.metadata as metadata  # noqa: F401
+
+
+import importlib.resources as resources  # noqa: F401
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_itertools.py b/backend/venv/lib/python3.12/site-packages/setuptools/_itertools.py
new file mode 100644
index 00000000..d6ca8413
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_itertools.py
@@ -0,0 +1,23 @@
+from more_itertools import consume  # noqa: F401
+
+
+# copied from jaraco.itertools 6.1
+def ensure_unique(iterable, key=lambda x: x):
+    """
+    Wrap an iterable to raise a ValueError if non-unique values are encountered.
+
+    >>> list(ensure_unique('abc'))
+    ['a', 'b', 'c']
+    >>> consume(ensure_unique('abca'))
+    Traceback (most recent call last):
+    ...
+    ValueError: Duplicate element 'a' encountered.
+    """
+    seen = set()
+    seen_add = seen.add
+    for element in iterable:
+        k = key(element)
+        if k in seen:
+            raise ValueError(f"Duplicate element {element!r} encountered.")
+        seen_add(k)
+        yield element
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_normalization.py b/backend/venv/lib/python3.12/site-packages/setuptools/_normalization.py
new file mode 100644
index 00000000..467b643d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_normalization.py
@@ -0,0 +1,144 @@
+"""
+Helpers for normalization as expected in wheel/sdist/module file names
+and core metadata
+"""
+
+import re
+
+import packaging
+
+# https://packaging.python.org/en/latest/specifications/core-metadata/#name
+_VALID_NAME = re.compile(r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.I)
+_UNSAFE_NAME_CHARS = re.compile(r"[^A-Z0-9._-]+", re.I)
+_NON_ALPHANUMERIC = re.compile(r"[^A-Z0-9]+", re.I)
+_PEP440_FALLBACK = re.compile(r"^v?(?P<safe>(?:[0-9]+!)?[0-9]+(?:\.[0-9]+)*)", re.I)
+
+
+def safe_identifier(name: str) -> str:
+    """Make a string safe to be used as Python identifier.
+    >>> safe_identifier("12abc")
+    '_12abc'
+    >>> safe_identifier("__editable__.myns.pkg-78.9.3_local")
+    '__editable___myns_pkg_78_9_3_local'
+    """
+    safe = re.sub(r'\W|^(?=\d)', '_', name)
+    assert safe.isidentifier()
+    return safe
+
+
+def safe_name(component: str) -> str:
+    """Escape a component used as a project name according to Core Metadata.
+    >>> safe_name("hello world")
+    'hello-world'
+    >>> safe_name("hello?world")
+    'hello-world'
+    >>> safe_name("hello_world")
+    'hello_world'
+    """
+    # See pkg_resources.safe_name
+    return _UNSAFE_NAME_CHARS.sub("-", component)
+
+
+def safe_version(version: str) -> str:
+    """Convert an arbitrary string into a valid version string.
+    Can still raise an ``InvalidVersion`` exception.
+    To avoid exceptions use ``best_effort_version``.
+    >>> safe_version("1988 12 25")
+    '1988.12.25'
+    >>> safe_version("v0.2.1")
+    '0.2.1'
+    >>> safe_version("v0.2?beta")
+    '0.2b0'
+    >>> safe_version("v0.2 beta")
+    '0.2b0'
+    >>> safe_version("ubuntu lts")
+    Traceback (most recent call last):
+    ...
+    packaging.version.InvalidVersion: Invalid version: 'ubuntu.lts'
+    """
+    v = version.replace(' ', '.')
+    try:
+        return str(packaging.version.Version(v))
+    except packaging.version.InvalidVersion:
+        attempt = _UNSAFE_NAME_CHARS.sub("-", v)
+        return str(packaging.version.Version(attempt))
+
+
+def best_effort_version(version: str) -> str:
+    """Convert an arbitrary string into a version-like string.
+    Fallback when ``safe_version`` is not safe enough.
+    >>> best_effort_version("v0.2 beta")
+    '0.2b0'
+    >>> best_effort_version("ubuntu lts")
+    '0.dev0+sanitized.ubuntu.lts'
+    >>> best_effort_version("0.23ubuntu1")
+    '0.23.dev0+sanitized.ubuntu1'
+    >>> best_effort_version("0.23-")
+    '0.23.dev0+sanitized'
+    >>> best_effort_version("0.-_")
+    '0.dev0+sanitized'
+    >>> best_effort_version("42.+?1")
+    '42.dev0+sanitized.1'
+    """
+    # See pkg_resources._forgiving_version
+    try:
+        return safe_version(version)
+    except packaging.version.InvalidVersion:
+        v = version.replace(' ', '.')
+        match = _PEP440_FALLBACK.search(v)
+        if match:
+            safe = match["safe"]
+            rest = v[len(safe) :]
+        else:
+            safe = "0"
+            rest = version
+        safe_rest = _NON_ALPHANUMERIC.sub(".", rest).strip(".")
+        local = f"sanitized.{safe_rest}".strip(".")
+        return safe_version(f"{safe}.dev0+{local}")
+
+
+def safe_extra(extra: str) -> str:
+    """Normalize extra name according to PEP 685
+    >>> safe_extra("_FrIeNdLy-._.-bArD")
+    'friendly-bard'
+    >>> safe_extra("FrIeNdLy-._.-bArD__._-")
+    'friendly-bard'
+    """
+    return _NON_ALPHANUMERIC.sub("-", extra).strip("-").lower()
+
+
+def filename_component(value: str) -> str:
+    """Normalize each component of a filename (e.g. distribution/version part of wheel)
+    Note: ``value`` needs to be already normalized.
+    >>> filename_component("my-pkg")
+    'my_pkg'
+    """
+    return value.replace("-", "_").strip("_")
+
+
+def filename_component_broken(value: str) -> str:
+    """
+    Produce the incorrect filename component for compatibility.
+
+    See pypa/setuptools#4167 for detailed analysis.
+
+    TODO: replace this with filename_component after pip 24 is
+    nearly-ubiquitous.
+
+    >>> filename_component_broken('foo_bar-baz')
+    'foo-bar-baz'
+    """
+    return value.replace('_', '-')
+
+
+def safer_name(value: str) -> str:
+    """Like ``safe_name`` but can be used as filename component for wheel"""
+    # See bdist_wheel.safer_name
+    return filename_component(safe_name(value))
+
+
+def safer_best_effort_version(value: str) -> str:
+    """Like ``best_effort_version`` but can be used as filename component for wheel"""
+    # See bdist_wheel.safer_verion
+    # TODO: Replace with only safe_version in the future (no need for best effort)
+    return filename_component(best_effort_version(value))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_path.py b/backend/venv/lib/python3.12/site-packages/setuptools/_path.py
new file mode 100644
index 00000000..0d99b0f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_path.py
@@ -0,0 +1,84 @@
+from __future__ import annotations
+
+import contextlib
+import os
+import sys
+from typing import TYPE_CHECKING, TypeVar, Union
+
+from more_itertools import unique_everseen
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+StrPath: TypeAlias = Union[str, os.PathLike[str]]  #  Same as _typeshed.StrPath
+StrPathT = TypeVar("StrPathT", bound=Union[str, os.PathLike[str]])
+
+
+def ensure_directory(path):
+    """Ensure that the parent directory of `path` exists"""
+    dirname = os.path.dirname(path)
+    os.makedirs(dirname, exist_ok=True)
+
+
+def same_path(p1: StrPath, p2: StrPath) -> bool:
+    """Differs from os.path.samefile because it does not require paths to exist.
+    Purely string based (no comparison between i-nodes).
+    >>> same_path("a/b", "./a/b")
+    True
+    >>> same_path("a/b", "a/./b")
+    True
+    >>> same_path("a/b", "././a/b")
+    True
+    >>> same_path("a/b", "./a/b/c/..")
+    True
+    >>> same_path("a/b", "../a/b/c")
+    False
+    >>> same_path("a", "a/b")
+    False
+    """
+    return normpath(p1) == normpath(p2)
+
+
+def normpath(filename: StrPath) -> str:
+    """Normalize a file/dir name for comparison purposes."""
+    # See pkg_resources.normalize_path for notes about cygwin
+    file = os.path.abspath(filename) if sys.platform == 'cygwin' else filename
+    return os.path.normcase(os.path.realpath(os.path.normpath(file)))
+
+
+@contextlib.contextmanager
+def paths_on_pythonpath(paths):
+    """
+    Add the indicated paths to the head of the PYTHONPATH environment
+    variable so that subprocesses will also see the packages at
+    these paths.
+
+    Do this in a context that restores the value on exit.
+
+    >>> getfixture('monkeypatch').setenv('PYTHONPATH', 'anything')
+    >>> with paths_on_pythonpath(['foo', 'bar']):
+    ...     assert 'foo' in os.environ['PYTHONPATH']
+    ...     assert 'anything' in os.environ['PYTHONPATH']
+    >>> os.environ['PYTHONPATH']
+    'anything'
+
+    >>> getfixture('monkeypatch').delenv('PYTHONPATH')
+    >>> with paths_on_pythonpath(['foo', 'bar']):
+    ...     assert 'foo' in os.environ['PYTHONPATH']
+    >>> os.environ.get('PYTHONPATH')
+    """
+    nothing = object()
+    orig_pythonpath = os.environ.get('PYTHONPATH', nothing)
+    current_pythonpath = os.environ.get('PYTHONPATH', '')
+    try:
+        prefix = os.pathsep.join(unique_everseen(paths))
+        to_join = filter(None, [prefix, current_pythonpath])
+        new_path = os.pathsep.join(to_join)
+        if new_path:
+            os.environ['PYTHONPATH'] = new_path
+        yield
+    finally:
+        if orig_pythonpath is nothing:
+            os.environ.pop('PYTHONPATH', None)
+        else:
+            os.environ['PYTHONPATH'] = orig_pythonpath
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_reqs.py b/backend/venv/lib/python3.12/site-packages/setuptools/_reqs.py
new file mode 100644
index 00000000..c793be4d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_reqs.py
@@ -0,0 +1,42 @@
+from __future__ import annotations
+
+from collections.abc import Iterable, Iterator
+from functools import lru_cache
+from typing import TYPE_CHECKING, Callable, TypeVar, Union, overload
+
+import jaraco.text as text
+from packaging.requirements import Requirement
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+_T = TypeVar("_T")
+_StrOrIter: TypeAlias = Union[str, Iterable[str]]
+
+
+parse_req: Callable[[str], Requirement] = lru_cache()(Requirement)
+# Setuptools parses the same requirement many times
+# (e.g. first for validation than for normalisation),
+# so it might be worth to cache.
+
+
+def parse_strings(strs: _StrOrIter) -> Iterator[str]:
+    """
+    Yield requirement strings for each specification in `strs`.
+
+    `strs` must be a string, or a (possibly-nested) iterable thereof.
+    """
+    return text.join_continuation(map(text.drop_comment, text.yield_lines(strs)))
+
+
+# These overloads are only needed because of a mypy false-positive, pyright gets it right
+# https://github.com/python/mypy/issues/3737
+@overload
+def parse(strs: _StrOrIter) -> Iterator[Requirement]: ...
+@overload
+def parse(strs: _StrOrIter, parser: Callable[[str], _T]) -> Iterator[_T]: ...
+def parse(strs: _StrOrIter, parser: Callable[[str], _T] = parse_req) -> Iterator[_T]:  # type: ignore[assignment]
+    """
+    Replacement for ``pkg_resources.parse_requirements`` that uses ``packaging``.
+    """
+    return map(parser, parse_strings(strs))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_shutil.py b/backend/venv/lib/python3.12/site-packages/setuptools/_shutil.py
new file mode 100644
index 00000000..6acbb428
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_shutil.py
@@ -0,0 +1,53 @@
+"""Convenience layer on top of stdlib's shutil and os"""
+
+import os
+import stat
+from typing import Callable, TypeVar
+
+from .compat import py311
+
+from distutils import log
+
+try:
+    from os import chmod  # pyright: ignore[reportAssignmentType]
+    # Losing type-safety w/ pyright, but that's ok
+except ImportError:  # pragma: no cover
+    # Jython compatibility
+    def chmod(*args: object, **kwargs: object) -> None:  # type: ignore[misc] # Mypy reuses the imported definition anyway
+        pass
+
+
+_T = TypeVar("_T")
+
+
+def attempt_chmod_verbose(path, mode):
+    log.debug("changing mode of %s to %o", path, mode)
+    try:
+        chmod(path, mode)
+    except OSError as e:  # pragma: no cover
+        log.debug("chmod failed: %s", e)
+
+
+# Must match shutil._OnExcCallback
+def _auto_chmod(
+    func: Callable[..., _T], arg: str, exc: BaseException
+) -> _T:  # pragma: no cover
+    """shutils onexc callback to automatically call chmod for certain functions."""
+    # Only retry for scenarios known to have an issue
+    if func in [os.unlink, os.remove] and os.name == 'nt':
+        attempt_chmod_verbose(arg, stat.S_IWRITE)
+        return func(arg)
+    raise exc
+
+
+def rmtree(path, ignore_errors=False, onexc=_auto_chmod):
+    """
+    Similar to ``shutil.rmtree`` but automatically executes ``chmod``
+    for well know Windows failure scenarios.
+    """
+    return py311.shutil_rmtree(path, ignore_errors, onexc)
+
+
+def rmdir(path, **opts):
+    if os.path.isdir(path):
+        rmtree(path, **opts)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_static.py b/backend/venv/lib/python3.12/site-packages/setuptools/_static.py
new file mode 100644
index 00000000..075a0bcd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_static.py
@@ -0,0 +1,188 @@
+from functools import wraps
+from typing import TypeVar
+
+import packaging.specifiers
+
+from .warnings import SetuptoolsDeprecationWarning
+
+
+class Static:
+    """
+    Wrapper for built-in object types that are allow setuptools to identify
+    static core metadata (in opposition to ``Dynamic``, as defined :pep:`643`).
+
+    The trick is to mark values with :class:`Static` when they come from
+    ``pyproject.toml`` or ``setup.cfg``, so if any plugin overwrite the value
+    with a built-in, setuptools will be able to recognise the change.
+
+    We inherit from built-in classes, so that we don't need to change the existing
+    code base to deal with the new types.
+    We also should strive for immutability objects to avoid changes after the
+    initial parsing.
+    """
+
+    _mutated_: bool = False  # TODO: Remove after deprecation warning is solved
+
+
+def _prevent_modification(target: type, method: str, copying: str) -> None:
+    """
+    Because setuptools is very flexible we cannot fully prevent
+    plugins and user customisations from modifying static values that were
+    parsed from config files.
+    But we can attempt to block "in-place" mutations and identify when they
+    were done.
+    """
+    fn = getattr(target, method, None)
+    if fn is None:
+        return
+
+    @wraps(fn)
+    def _replacement(self: Static, *args, **kwargs):
+        # TODO: After deprecation period raise NotImplementedError instead of warning
+        #       which obviated the existence and checks of the `_mutated_` attribute.
+        self._mutated_ = True
+        SetuptoolsDeprecationWarning.emit(
+            "Direct modification of value will be disallowed",
+            f"""
+            In an effort to implement PEP 643, direct/in-place changes of static values
+            that come from configuration files are deprecated.
+            If you need to modify this value, please first create a copy with {copying}
+            and make sure conform to all relevant standards when overriding setuptools
+            functionality (https://packaging.python.org/en/latest/specifications/).
+            """,
+            due_date=(2025, 10, 10),  # Initially introduced in 2024-09-06
+        )
+        return fn(self, *args, **kwargs)
+
+    _replacement.__doc__ = ""  # otherwise doctest may fail.
+    setattr(target, method, _replacement)
+
+
+class Str(str, Static):
+    pass
+
+
+class Tuple(tuple, Static):
+    pass
+
+
+class List(list, Static):
+    """
+    :meta private:
+    >>> x = List([1, 2, 3])
+    >>> is_static(x)
+    True
+    >>> x += [0]  # doctest: +IGNORE_EXCEPTION_DETAIL
+    Traceback (most recent call last):
+    SetuptoolsDeprecationWarning: Direct modification ...
+    >>> is_static(x)  # no longer static after modification
+    False
+    >>> y = list(x)
+    >>> y.clear()
+    >>> y
+    []
+    >>> y == x
+    False
+    >>> is_static(List(y))
+    True
+    """
+
+
+# Make `List` immutable-ish
+# (certain places of setuptools/distutils issue a warn if we use tuple instead of list)
+for _method in (
+    '__delitem__',
+    '__iadd__',
+    '__setitem__',
+    'append',
+    'clear',
+    'extend',
+    'insert',
+    'remove',
+    'reverse',
+    'pop',
+):
+    _prevent_modification(List, _method, "`list(value)`")
+
+
+class Dict(dict, Static):
+    """
+    :meta private:
+    >>> x = Dict({'a': 1, 'b': 2})
+    >>> is_static(x)
+    True
+    >>> x['c'] = 0  # doctest: +IGNORE_EXCEPTION_DETAIL
+    Traceback (most recent call last):
+    SetuptoolsDeprecationWarning: Direct modification ...
+    >>> x._mutated_
+    True
+    >>> is_static(x)  # no longer static after modification
+    False
+    >>> y = dict(x)
+    >>> y.popitem()
+    ('b', 2)
+    >>> y == x
+    False
+    >>> is_static(Dict(y))
+    True
+    """
+
+
+# Make `Dict` immutable-ish (we cannot inherit from types.MappingProxyType):
+for _method in (
+    '__delitem__',
+    '__ior__',
+    '__setitem__',
+    'clear',
+    'pop',
+    'popitem',
+    'setdefault',
+    'update',
+):
+    _prevent_modification(Dict, _method, "`dict(value)`")
+
+
+class SpecifierSet(packaging.specifiers.SpecifierSet, Static):
+    """Not exactly a built-in type but useful for ``requires-python``"""
+
+
+T = TypeVar("T")
+
+
+def noop(value: T) -> T:
+    """
+    >>> noop(42)
+    42
+    """
+    return value
+
+
+_CONVERSIONS = {str: Str, tuple: Tuple, list: List, dict: Dict}
+
+
+def attempt_conversion(value: T) -> T:
+    """
+    >>> is_static(attempt_conversion("hello"))
+    True
+    >>> is_static(object())
+    False
+    """
+    return _CONVERSIONS.get(type(value), noop)(value)  # type: ignore[call-overload]
+
+
+def is_static(value: object) -> bool:
+    """
+    >>> is_static(a := Dict({'a': 1}))
+    True
+    >>> is_static(dict(a))
+    False
+    >>> is_static(b := List([1, 2, 3]))
+    True
+    >>> is_static(list(b))
+    False
+    """
+    return isinstance(value, Static) and not value._mutated_
+
+
+EMPTY_LIST = List()
+EMPTY_DICT = Dict()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/__pycache__/typing_extensions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/__pycache__/typing_extensions.cpython-312.pyc
new file mode 100644
index 00000000..906edc04
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/__pycache__/typing_extensions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/LICENSE
new file mode 100644
index 00000000..b49c3af0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/LICENSE
@@ -0,0 +1,166 @@
+GNU LESSER GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+
+  This version of the GNU Lesser General Public License incorporates
+the terms and conditions of version 3 of the GNU General Public
+License, supplemented by the additional permissions listed below.
+
+  0. Additional Definitions.
+
+  As used herein, "this License" refers to version 3 of the GNU Lesser
+General Public License, and the "GNU GPL" refers to version 3 of the GNU
+General Public License.
+
+  "The Library" refers to a covered work governed by this License,
+other than an Application or a Combined Work as defined below.
+
+  An "Application" is any work that makes use of an interface provided
+by the Library, but which is not otherwise based on the Library.
+Defining a subclass of a class defined by the Library is deemed a mode
+of using an interface provided by the Library.
+
+  A "Combined Work" is a work produced by combining or linking an
+Application with the Library.  The particular version of the Library
+with which the Combined Work was made is also called the "Linked
+Version".
+
+  The "Minimal Corresponding Source" for a Combined Work means the
+Corresponding Source for the Combined Work, excluding any source code
+for portions of the Combined Work that, considered in isolation, are
+based on the Application, and not on the Linked Version.
+
+  The "Corresponding Application Code" for a Combined Work means the
+object code and/or source code for the Application, including any data
+and utility programs needed for reproducing the Combined Work from the
+Application, but excluding the System Libraries of the Combined Work.
+
+  1. Exception to Section 3 of the GNU GPL.
+
+  You may convey a covered work under sections 3 and 4 of this License
+without being bound by section 3 of the GNU GPL.
+
+  2. Conveying Modified Versions.
+
+  If you modify a copy of the Library, and, in your modifications, a
+facility refers to a function or data to be supplied by an Application
+that uses the facility (other than as an argument passed when the
+facility is invoked), then you may convey a copy of the modified
+version:
+
+   a) under this License, provided that you make a good faith effort to
+   ensure that, in the event an Application does not supply the
+   function or data, the facility still operates, and performs
+   whatever part of its purpose remains meaningful, or
+
+   b) under the GNU GPL, with none of the additional permissions of
+   this License applicable to that copy.
+
+  3. Object Code Incorporating Material from Library Header Files.
+
+  The object code form of an Application may incorporate material from
+a header file that is part of the Library.  You may convey such object
+code under terms of your choice, provided that, if the incorporated
+material is not limited to numerical parameters, data structure
+layouts and accessors, or small macros, inline functions and templates
+(ten or fewer lines in length), you do both of the following:
+
+   a) Give prominent notice with each copy of the object code that the
+   Library is used in it and that the Library and its use are
+   covered by this License.
+
+   b) Accompany the object code with a copy of the GNU GPL and this license
+   document.
+
+  4. Combined Works.
+
+  You may convey a Combined Work under terms of your choice that,
+taken together, effectively do not restrict modification of the
+portions of the Library contained in the Combined Work and reverse
+engineering for debugging such modifications, if you also do each of
+the following:
+
+   a) Give prominent notice with each copy of the Combined Work that
+   the Library is used in it and that the Library and its use are
+   covered by this License.
+
+   b) Accompany the Combined Work with a copy of the GNU GPL and this license
+   document.
+
+   c) For a Combined Work that displays copyright notices during
+   execution, include the copyright notice for the Library among
+   these notices, as well as a reference directing the user to the
+   copies of the GNU GPL and this license document.
+
+   d) Do one of the following:
+
+       0) Convey the Minimal Corresponding Source under the terms of this
+       License, and the Corresponding Application Code in a form
+       suitable for, and under terms that permit, the user to
+       recombine or relink the Application with a modified version of
+       the Linked Version to produce a modified Combined Work, in the
+       manner specified by section 6 of the GNU GPL for conveying
+       Corresponding Source.
+
+       1) Use a suitable shared library mechanism for linking with the
+       Library.  A suitable mechanism is one that (a) uses at run time
+       a copy of the Library already present on the user's computer
+       system, and (b) will operate properly with a modified version
+       of the Library that is interface-compatible with the Linked
+       Version.
+
+   e) Provide Installation Information, but only if you would otherwise
+   be required to provide such information under section 6 of the
+   GNU GPL, and only to the extent that such information is
+   necessary to install and execute a modified version of the
+   Combined Work produced by recombining or relinking the
+   Application with a modified version of the Linked Version. (If
+   you use option 4d0, the Installation Information must accompany
+   the Minimal Corresponding Source and Corresponding Application
+   Code. If you use option 4d1, you must provide the Installation
+   Information in the manner specified by section 6 of the GNU GPL
+   for conveying Corresponding Source.)
+
+  5. Combined Libraries.
+
+  You may place library facilities that are a work based on the
+Library side by side in a single library together with other library
+facilities that are not Applications and are not covered by this
+License, and convey such a combined library under terms of your
+choice, if you do both of the following:
+
+   a) Accompany the combined library with a copy of the same work based
+   on the Library, uncombined with any other library facilities,
+   conveyed under the terms of this License.
+
+   b) Give prominent notice with the combined library that part of it
+   is a work based on the Library, and explaining where to find the
+   accompanying uncombined form of the same work.
+
+  6. Revised Versions of the GNU Lesser General Public License.
+
+  The Free Software Foundation may publish revised and/or new versions
+of the GNU Lesser General Public License from time to time. Such new
+versions will be similar in spirit to the present version, but may
+differ in detail to address new problems or concerns.
+
+  Each version is given a distinguishing version number. If the
+Library as you received it specifies that a certain numbered version
+of the GNU Lesser General Public License "or any later version"
+applies to it, you have the option of following the terms and
+conditions either of that published version or of any later version
+published by the Free Software Foundation. If the Library as you
+received it does not specify a version number of the GNU Lesser
+General Public License, you may choose any version of the GNU Lesser
+General Public License ever published by the Free Software Foundation.
+
+  If the Library as you received it specifies that a proxy can decide
+whether future versions of the GNU Lesser General Public License shall
+apply, that proxy's public statement of acceptance of any version is
+permanent authorization for you to choose that version for the
+Library.
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/METADATA
new file mode 100644
index 00000000..32214fb4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/METADATA
@@ -0,0 +1,420 @@
+Metadata-Version: 2.1
+Name: autocommand
+Version: 2.2.2
+Summary: A library to create a command-line program from a function
+Home-page: https://github.com/Lucretiel/autocommand
+Author: Nathan West
+License: LGPLv3
+Project-URL: Homepage, https://github.com/Lucretiel/autocommand
+Project-URL: Bug Tracker, https://github.com/Lucretiel/autocommand/issues
+Platform: any
+Classifier: Development Status :: 6 - Mature
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: GNU Lesser General Public License v3 (LGPLv3)
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Software Development
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+License-File: LICENSE
+
+[![PyPI version](https://badge.fury.io/py/autocommand.svg)](https://badge.fury.io/py/autocommand)
+
+# autocommand
+
+A library to automatically generate and run simple argparse parsers from function signatures.
+
+## Installation
+
+Autocommand is installed via pip:
+
+```
+$ pip install autocommand
+```
+
+## Usage
+
+Autocommand turns a function into a command-line program. It converts the function's parameter signature into command-line arguments, and automatically runs the function if the module was called as `__main__`. In effect, it lets your create a smart main function.
+
+```python
+from autocommand import autocommand
+
+# This program takes exactly one argument and echos it.
+@autocommand(__name__)
+def echo(thing):
+    print(thing)
+```
+
+```
+$ python echo.py hello
+hello
+$ python echo.py -h
+usage: echo [-h] thing
+
+positional arguments:
+  thing
+
+optional arguments:
+  -h, --help  show this help message and exit
+$ python echo.py hello world  # too many arguments
+usage: echo.py [-h] thing
+echo.py: error: unrecognized arguments: world
+```
+
+As you can see, autocommand converts the signature of the function into an argument spec. When you run the file as a program, autocommand collects the command-line arguments and turns them into function arguments. The function is executed with these arguments, and then the program exits with the return value of the function, via `sys.exit`. Autocommand also automatically creates a usage message, which can be invoked with `-h` or `--help`, and automatically prints an error message when provided with invalid arguments.
+
+### Types
+
+You can use a type annotation to give an argument a type. Any type (or in fact any callable) that returns an object when given a string argument can be used, though there are a few special cases that are described later.
+
+```python
+@autocommand(__name__)
+def net_client(host, port: int):
+    ...
+```
+
+Autocommand will catch `TypeErrors` raised by the type during argument parsing, so you can supply a callable and do some basic argument validation as well.
+
+### Trailing Arguments
+
+You can add a `*args` parameter to your function to give it trailing arguments. The command will collect 0 or more trailing arguments and supply them to `args` as a tuple. If a type annotation is supplied, the type is applied to each argument.
+
+```python
+# Write the contents of each file, one by one
+@autocommand(__name__)
+def cat(*files):
+    for filename in files:
+        with open(filename) as file:
+            for line in file:
+                print(line.rstrip())
+```
+
+```
+$ python cat.py -h
+usage: ipython [-h] [file [file ...]]
+
+positional arguments:
+  file
+
+optional arguments:
+  -h, --help  show this help message and exit
+```
+
+### Options
+
+To create `--option` switches, just assign a default. Autocommand will automatically create `--long` and `-s`hort switches.
+
+```python
+@autocommand(__name__)
+def do_with_config(argument, config='~/foo.conf'):
+    pass
+```
+
+```
+$ python example.py -h
+usage: example.py [-h] [-c CONFIG] argument
+
+positional arguments:
+  argument
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -c CONFIG, --config CONFIG
+```
+
+The option's type is automatically deduced from the default, unless one is explicitly given in an annotation:
+
+```python
+@autocommand(__name__)
+def http_connect(host, port=80):
+    print('{}:{}'.format(host, port))
+```
+
+```
+$ python http.py -h
+usage: http.py [-h] [-p PORT] host
+
+positional arguments:
+  host
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -p PORT, --port PORT
+$ python http.py localhost
+localhost:80
+$ python http.py localhost -p 8080
+localhost:8080
+$ python http.py localhost -p blah
+usage: http.py [-h] [-p PORT] host
+http.py: error: argument -p/--port: invalid int value: 'blah'
+```
+
+#### None
+
+If an option is given a default value of `None`, it reads in a value as normal, but supplies `None` if the option isn't provided.
+
+#### Switches
+
+If an argument is given a default value of `True` or `False`, or
+given an explicit `bool` type, it becomes an option switch.
+
+```python
+    @autocommand(__name__)
+    def example(verbose=False, quiet=False):
+        pass
+```
+
+```
+$ python example.py -h
+usage: example.py [-h] [-v] [-q]
+
+optional arguments:
+  -h, --help     show this help message and exit
+  -v, --verbose
+  -q, --quiet
+```
+
+Autocommand attempts to do the "correct thing" in these cases- if the default is `True`, then supplying the switch makes the argument `False`; if the type is `bool` and the default is some other `True` value, then supplying the switch makes the argument `False`, while not supplying the switch makes the argument the default value.
+
+Autocommand also supports the creation of switch inverters. Pass `add_nos=True` to `autocommand` to enable this.
+
+```
+    @autocommand(__name__, add_nos=True)
+    def example(verbose=False):
+        pass
+```
+
+```
+$ python example.py -h
+usage: ipython [-h] [-v] [--no-verbose]
+
+optional arguments:
+  -h, --help     show this help message and exit
+  -v, --verbose
+  --no-verbose
+```
+
+Using the `--no-` version of a switch will pass the opposite value in as a function argument. If multiple switches are present, the last one takes precedence.
+
+#### Files
+
+If the default value is a file object, such as `sys.stdout`, then autocommand just looks for a string, for a file path. It doesn't do any special checking on the string, though (such as checking if the file exists); it's better to let the client decide how to handle errors in this case. Instead, it provides a special context manager called `smart_open`, which behaves exactly like `open` if a filename or other openable type is provided, but also lets you use already open files:
+
+```python
+from autocommand import autocommand, smart_open
+import sys
+
+# Write the contents of stdin, or a file, to stdout
+@autocommand(__name__)
+def write_out(infile=sys.stdin):
+    with smart_open(infile) as f:
+        for line in f:
+            print(line.rstrip())
+    # If a file was opened, it is closed here. If it was just stdin, it is untouched.
+```
+
+```
+$ echo "Hello World!" | python write_out.py | tee hello.txt
+Hello World!
+$ python write_out.py --infile hello.txt
+Hello World!
+```
+
+### Descriptions and docstrings
+
+The `autocommand` decorator accepts `description` and `epilog` kwargs, corresponding to the `description <https://docs.python.org/3/library/argparse.html#description>`_ and `epilog <https://docs.python.org/3/library/argparse.html#epilog>`_ of the `ArgumentParser`. If no description is given, but the decorated function has a docstring, then it is taken as the `description` for the `ArgumentParser`. You can also provide both the description and epilog in the docstring by splitting it into two sections with 4 or more - characters.
+
+```python
+@autocommand(__name__)
+def copy(infile=sys.stdin, outfile=sys.stdout):
+    '''
+    Copy an the contents of a file (or stdin) to another file (or stdout)
+    ----------
+    Some extra documentation in the epilog
+    '''
+    with smart_open(infile) as istr:
+        with smart_open(outfile, 'w') as ostr:
+            for line in istr:
+                ostr.write(line)
+```
+
+```
+$ python copy.py -h
+usage: copy.py [-h] [-i INFILE] [-o OUTFILE]
+
+Copy an the contents of a file (or stdin) to another file (or stdout)
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -i INFILE, --infile INFILE
+  -o OUTFILE, --outfile OUTFILE
+
+Some extra documentation in the epilog
+$ echo "Hello World" | python copy.py --outfile hello.txt
+$ python copy.py --infile hello.txt --outfile hello2.txt
+$ python copy.py --infile hello2.txt
+Hello World
+```
+
+### Parameter descriptions
+
+You can also attach description text to individual parameters in the annotation. To attach both a type and a description, supply them both in any order in a tuple
+
+```python
+@autocommand(__name__)
+def copy_net(
+    infile: 'The name of the file to send',
+    host: 'The host to send the file to',
+    port: (int, 'The port to connect to')):
+
+    '''
+    Copy a file over raw TCP to a remote destination.
+    '''
+    # Left as an exercise to the reader
+```
+
+### Decorators and wrappers
+
+Autocommand automatically follows wrapper chains created by `@functools.wraps`. This means that you can apply other wrapping decorators to your main function, and autocommand will still correctly detect the signature.
+
+```python
+from functools import wraps
+from autocommand import autocommand
+
+def print_yielded(func):
+    '''
+    Convert a generator into a function that prints all yielded elements
+    '''
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+        for thing in func(*args, **kwargs):
+            print(thing)
+    return wrapper
+
+@autocommand(__name__,
+    description= 'Print all the values from START to STOP, inclusive, in steps of STEP',
+    epilog=      'STOP and STEP default to 1')
+@print_yielded
+def seq(stop, start=1, step=1):
+    for i in range(start, stop + 1, step):
+        yield i
+```
+
+```
+$ seq.py -h
+usage: seq.py [-h] [-s START] [-S STEP] stop
+
+Print all the values from START to STOP, inclusive, in steps of STEP
+
+positional arguments:
+  stop
+
+optional arguments:
+  -h, --help            show this help message and exit
+  -s START, --start START
+  -S STEP, --step STEP
+
+STOP and STEP default to 1
+```
+
+Even though autocommand is being applied to the `wrapper` returned by `print_yielded`, it still retreives the signature of the underlying `seq` function to create the argument parsing.
+
+### Custom Parser
+
+While autocommand's automatic parser generator is a powerful convenience, it doesn't cover all of the different features that argparse provides. If you need these features, you can provide your own parser as a kwarg to `autocommand`:
+
+```python
+from argparse import ArgumentParser
+from autocommand import autocommand
+
+parser = ArgumentParser()
+# autocommand can't do optional positonal parameters
+parser.add_argument('arg', nargs='?')
+# or mutually exclusive options
+group = parser.add_mutually_exclusive_group()
+group.add_argument('-v', '--verbose', action='store_true')
+group.add_argument('-q', '--quiet', action='store_true')
+
+@autocommand(__name__, parser=parser)
+def main(arg, verbose, quiet):
+    print(arg, verbose, quiet)
+```
+
+```
+$ python parser.py -h
+usage: write_file.py [-h] [-v | -q] [arg]
+
+positional arguments:
+  arg
+
+optional arguments:
+  -h, --help     show this help message and exit
+  -v, --verbose
+  -q, --quiet
+$ python parser.py
+None False False
+$ python parser.py hello
+hello False False
+$ python parser.py -v
+None True False
+$ python parser.py -q
+None False True
+$ python parser.py -vq
+usage: parser.py [-h] [-v | -q] [arg]
+parser.py: error: argument -q/--quiet: not allowed with argument -v/--verbose
+```
+
+Any parser should work fine, so long as each of the parser's arguments has a corresponding parameter in the decorated main function. The order of parameters doesn't matter, as long as they are all present. Note that when using a custom parser, autocommand doesn't modify the parser or the retrieved arguments. This means that no description/epilog will be added, and the function's type annotations and defaults (if present) will be ignored.
+
+## Testing and Library use
+
+The decorated function is only called and exited from if the first argument to `autocommand` is `'__main__'` or `True`. If it is neither of these values, or no argument is given, then a new main function is created by the decorator. This function has the signature `main(argv=None)`, and is intended to be called with arguments as if via `main(sys.argv[1:])`. The function has the attributes `parser` and `main`, which are the generated `ArgumentParser` and the original main function that was decorated. This is to facilitate testing and library use of your main. Calling the function triggers a `parse_args()` with the supplied arguments, and returns the result of the main function. Note that, while it returns instead of calling `sys.exit`, the `parse_args()` function will raise a `SystemExit` in the event of a parsing error or `-h/--help` argument.
+
+```python
+    @autocommand()
+    def test_prog(arg1, arg2: int, quiet=False, verbose=False):
+        if not quiet:
+            print(arg1, arg2)
+            if verbose:
+                print("LOUD NOISES")
+
+        return 0
+
+    print(test_prog(['-v', 'hello', '80']))
+```
+
+```
+$ python test_prog.py
+hello 80
+LOUD NOISES
+0
+```
+
+If the function is called with no arguments, `sys.argv[1:]` is used. This is to allow the autocommand function to be used as a setuptools entry point.
+
+## Exceptions and limitations
+
+- There are a few possible exceptions that `autocommand` can raise. All of them derive from `autocommand.AutocommandError`.
+
+  - If an invalid annotation is given (that is, it isn't a `type`, `str`, `(type, str)`, or `(str, type)`, an `AnnotationError` is raised. The `type` may be any callable, as described in the `Types`_ section.
+  - If the function has a `**kwargs` parameter, a `KWargError` is raised.
+  - If, somehow, the function has a positional-only parameter, a `PositionalArgError` is raised. This means that the argument doesn't have a name, which is currently not possible with a plain `def` or `lambda`, though many built-in functions have this kind of parameter.
+
+- There are a few argparse features that are not supported by autocommand.
+
+  - It isn't possible to have an optional positional argument (as opposed to a `--option`). POSIX thinks this is bad form anyway.
+  - It isn't possible to have mutually exclusive arguments or options
+  - It isn't possible to have subcommands or subparsers, though I'm working on a few solutions involving classes or nested function definitions to allow this.
+
+## Development
+
+Autocommand cannot be important from the project root; this is to enforce separation of concerns and prevent accidental importing of `setup.py` or tests. To develop, install the project in editable mode:
+
+```
+$ python setup.py develop
+```
+
+This will create a link to the source files in the deployment directory, so that any source changes are reflected when it is imported.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/RECORD
new file mode 100644
index 00000000..e6e12ea5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/RECORD
@@ -0,0 +1,18 @@
+autocommand-2.2.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+autocommand-2.2.2.dist-info/LICENSE,sha256=reeNBJgtaZctREqOFKlPh6IzTdOFXMgDSOqOJAqg3y0,7634
+autocommand-2.2.2.dist-info/METADATA,sha256=OADZuR3O6iBlpu1ieTgzYul6w4uOVrk0P0BO5TGGAJk,15006
+autocommand-2.2.2.dist-info/RECORD,,
+autocommand-2.2.2.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+autocommand-2.2.2.dist-info/top_level.txt,sha256=AzfhgKKS8EdAwWUTSF8mgeVQbXOY9kokHB6kSqwwqu0,12
+autocommand/__init__.py,sha256=zko5Rnvolvb-UXjCx_2ArPTGBWwUK5QY4LIQIKYR7As,1037
+autocommand/__pycache__/__init__.cpython-312.pyc,,
+autocommand/__pycache__/autoasync.cpython-312.pyc,,
+autocommand/__pycache__/autocommand.cpython-312.pyc,,
+autocommand/__pycache__/automain.cpython-312.pyc,,
+autocommand/__pycache__/autoparse.cpython-312.pyc,,
+autocommand/__pycache__/errors.cpython-312.pyc,,
+autocommand/autoasync.py,sha256=AMdyrxNS4pqWJfP_xuoOcImOHWD-qT7x06wmKN1Vp-U,5680
+autocommand/autocommand.py,sha256=hmkEmQ72HtL55gnURVjDOnsfYlGd5lLXbvT4KG496Qw,2505
+autocommand/automain.py,sha256=A2b8i754Mxc_DjU9WFr6vqYDWlhz0cn8miu8d8EsxV8,2076
+autocommand/autoparse.py,sha256=WVWmZJPcbzUKXP40raQw_0HD8qPJ2V9VG1eFFmmnFxw,11642
+autocommand/errors.py,sha256=7aa3roh9Herd6nIKpQHNWEslWE8oq7GiHYVUuRqORnA,886
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/WHEEL
new file mode 100644
index 00000000..57e3d840
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.38.4)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/top_level.txt
new file mode 100644
index 00000000..dda5158f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand-2.2.2.dist-info/top_level.txt
@@ -0,0 +1 @@
+autocommand
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__init__.py
new file mode 100644
index 00000000..73fbfca6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__init__.py
@@ -0,0 +1,27 @@
+# Copyright 2014-2016 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+# flake8 flags all these imports as unused, hence the NOQAs everywhere.
+
+from .automain import automain  # NOQA
+from .autoparse import autoparse, smart_open  # NOQA
+from .autocommand import autocommand  # NOQA
+
+try:
+    from .autoasync import autoasync  # NOQA
+except ImportError:  # pragma: no cover
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..4b29e244
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoasync.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoasync.cpython-312.pyc
new file mode 100644
index 00000000..ccc4c32a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoasync.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autocommand.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autocommand.cpython-312.pyc
new file mode 100644
index 00000000..9a2bb5bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autocommand.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/automain.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/automain.cpython-312.pyc
new file mode 100644
index 00000000..32374ba8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/automain.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoparse.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoparse.cpython-312.pyc
new file mode 100644
index 00000000..a05c4f8e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/autoparse.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/errors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/errors.cpython-312.pyc
new file mode 100644
index 00000000..17fe9a14
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/__pycache__/errors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoasync.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoasync.py
new file mode 100644
index 00000000..688f7e05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoasync.py
@@ -0,0 +1,142 @@
+# Copyright 2014-2015 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+from asyncio import get_event_loop, iscoroutine
+from functools import wraps
+from inspect import signature
+
+
+async def _run_forever_coro(coro, args, kwargs, loop):
+    '''
+    This helper function launches an async main function that was tagged with
+    forever=True. There are two possibilities:
+
+    - The function is a normal function, which handles initializing the event
+      loop, which is then run forever
+    - The function is a coroutine, which needs to be scheduled in the event
+      loop, which is then run forever
+      - There is also the possibility that the function is a normal function
+        wrapping a coroutine function
+
+    The function is therefore called unconditionally and scheduled in the event
+    loop if the return value is a coroutine object.
+
+    The reason this is a separate function is to make absolutely sure that all
+    the objects created are garbage collected after all is said and done; we
+    do this to ensure that any exceptions raised in the tasks are collected
+    ASAP.
+    '''
+
+    # Personal note: I consider this an antipattern, as it relies on the use of
+    # unowned resources. The setup function dumps some stuff into the event
+    # loop where it just whirls in the ether without a well defined owner or
+    # lifetime. For this reason, there's a good chance I'll remove the
+    # forever=True feature from autoasync at some point in the future.
+    thing = coro(*args, **kwargs)
+    if iscoroutine(thing):
+        await thing
+
+
+def autoasync(coro=None, *, loop=None, forever=False, pass_loop=False):
+    '''
+    Convert an asyncio coroutine into a function which, when called, is
+    evaluted in an event loop, and the return value returned. This is intented
+    to make it easy to write entry points into asyncio coroutines, which
+    otherwise need to be explictly evaluted with an event loop's
+    run_until_complete.
+
+    If `loop` is given, it is used as the event loop to run the coro in. If it
+    is None (the default), the loop is retreived using asyncio.get_event_loop.
+    This call is defered until the decorated function is called, so that
+    callers can install custom event loops or event loop policies after
+    @autoasync is applied.
+
+    If `forever` is True, the loop is run forever after the decorated coroutine
+    is finished. Use this for servers created with asyncio.start_server and the
+    like.
+
+    If `pass_loop` is True, the event loop object is passed into the coroutine
+    as the `loop` kwarg when the wrapper function is called. In this case, the
+    wrapper function's __signature__ is updated to remove this parameter, so
+    that autoparse can still be used on it without generating a parameter for
+    `loop`.
+
+    This coroutine can be called with ( @autoasync(...) ) or without
+    ( @autoasync ) arguments.
+
+    Examples:
+
+    @autoasync
+    def get_file(host, port):
+        reader, writer = yield from asyncio.open_connection(host, port)
+        data = reader.read()
+        sys.stdout.write(data.decode())
+
+    get_file(host, port)
+
+    @autoasync(forever=True, pass_loop=True)
+    def server(host, port, loop):
+        yield_from loop.create_server(Proto, host, port)
+
+    server('localhost', 8899)
+
+    '''
+    if coro is None:
+        return lambda c: autoasync(
+            c, loop=loop,
+            forever=forever,
+            pass_loop=pass_loop)
+
+    # The old and new signatures are required to correctly bind the loop
+    # parameter in 100% of cases, even if it's a positional parameter.
+    # NOTE: A future release will probably require the loop parameter to be
+    # a kwonly parameter.
+    if pass_loop:
+        old_sig = signature(coro)
+        new_sig = old_sig.replace(parameters=(
+            param for name, param in old_sig.parameters.items()
+            if name != "loop"))
+
+    @wraps(coro)
+    def autoasync_wrapper(*args, **kwargs):
+        # Defer the call to get_event_loop so that, if a custom policy is
+        # installed after the autoasync decorator, it is respected at call time
+        local_loop = get_event_loop() if loop is None else loop
+
+        # Inject the 'loop' argument. We have to use this signature binding to
+        # ensure it's injected in the correct place (positional, keyword, etc)
+        if pass_loop:
+            bound_args = old_sig.bind_partial()
+            bound_args.arguments.update(
+                loop=local_loop,
+                **new_sig.bind(*args, **kwargs).arguments)
+            args, kwargs = bound_args.args, bound_args.kwargs
+
+        if forever:
+            local_loop.create_task(_run_forever_coro(
+                coro, args, kwargs, local_loop
+            ))
+            local_loop.run_forever()
+        else:
+            return local_loop.run_until_complete(coro(*args, **kwargs))
+
+    # Attach the updated signature. This allows 'pass_loop' to be used with
+    # autoparse
+    if pass_loop:
+        autoasync_wrapper.__signature__ = new_sig
+
+    return autoasync_wrapper
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autocommand.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autocommand.py
new file mode 100644
index 00000000..097e86de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autocommand.py
@@ -0,0 +1,70 @@
+# Copyright 2014-2015 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+from .autoparse import autoparse
+from .automain import automain
+try:
+    from .autoasync import autoasync
+except ImportError:  # pragma: no cover
+    pass
+
+
+def autocommand(
+        module, *,
+        description=None,
+        epilog=None,
+        add_nos=False,
+        parser=None,
+        loop=None,
+        forever=False,
+        pass_loop=False):
+
+    if callable(module):
+        raise TypeError('autocommand requires a module name argument')
+
+    def autocommand_decorator(func):
+        # Step 1: if requested, run it all in an asyncio event loop. autoasync
+        # patches the __signature__ of the decorated function, so that in the
+        # event that pass_loop is True, the `loop` parameter of the original
+        # function will *not* be interpreted as a command-line argument by
+        # autoparse
+        if loop is not None or forever or pass_loop:
+            func = autoasync(
+                func,
+                loop=None if loop is True else loop,
+                pass_loop=pass_loop,
+                forever=forever)
+
+        # Step 2: create parser. We do this second so that the arguments are
+        # parsed and passed *before* entering the asyncio event loop, if it
+        # exists. This simplifies the stack trace and ensures errors are
+        # reported earlier. It also ensures that errors raised during parsing &
+        # passing are still raised if `forever` is True.
+        func = autoparse(
+            func,
+            description=description,
+            epilog=epilog,
+            add_nos=add_nos,
+            parser=parser)
+
+        # Step 3: call the function automatically if __name__ == '__main__' (or
+        # if True was provided)
+        func = automain(module)(func)
+
+        return func
+
+    return autocommand_decorator
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/automain.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/automain.py
new file mode 100644
index 00000000..6cc45db6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/automain.py
@@ -0,0 +1,59 @@
+# Copyright 2014-2015 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+import sys
+from .errors import AutocommandError
+
+
+class AutomainRequiresModuleError(AutocommandError, TypeError):
+    pass
+
+
+def automain(module, *, args=(), kwargs=None):
+    '''
+    This decorator automatically invokes a function if the module is being run
+    as the "__main__" module. Optionally, provide args or kwargs with which to
+    call the function. If `module` is "__main__", the function is called, and
+    the program is `sys.exit`ed with the return value. You can also pass `True`
+    to cause the function to be called unconditionally. If the function is not
+    called, it is returned unchanged by the decorator.
+
+    Usage:
+
+    @automain(__name__)  # Pass __name__ to check __name__=="__main__"
+    def main():
+        ...
+
+    If __name__ is "__main__" here, the main function is called, and then
+    sys.exit called with the return value.
+    '''
+
+    # Check that @automain(...) was called, rather than @automain
+    if callable(module):
+        raise AutomainRequiresModuleError(module)
+
+    if module == '__main__' or module is True:
+        if kwargs is None:
+            kwargs = {}
+
+        # Use a function definition instead of a lambda for a neater traceback
+        def automain_decorator(main):
+            sys.exit(main(*args, **kwargs))
+
+        return automain_decorator
+    else:
+        return lambda main: main
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoparse.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoparse.py
new file mode 100644
index 00000000..0276a3fa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/autoparse.py
@@ -0,0 +1,333 @@
+# Copyright 2014-2015 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+import sys
+from re import compile as compile_regex
+from inspect import signature, getdoc, Parameter
+from argparse import ArgumentParser
+from contextlib import contextmanager
+from functools import wraps
+from io import IOBase
+from autocommand.errors import AutocommandError
+
+
+_empty = Parameter.empty
+
+
+class AnnotationError(AutocommandError):
+    '''Annotation error: annotation must be a string, type, or tuple of both'''
+
+
+class PositionalArgError(AutocommandError):
+    '''
+    Postional Arg Error: autocommand can't handle postional-only parameters
+    '''
+
+
+class KWArgError(AutocommandError):
+    '''kwarg Error: autocommand can't handle a **kwargs parameter'''
+
+
+class DocstringError(AutocommandError):
+    '''Docstring error'''
+
+
+class TooManySplitsError(DocstringError):
+    '''
+    The docstring had too many ---- section splits. Currently we only support
+    using up to a single split, to split the docstring into description and
+    epilog parts.
+    '''
+
+
+def _get_type_description(annotation):
+    '''
+    Given an annotation, return the (type, description) for the parameter.
+    If you provide an annotation that is somehow both a string and a callable,
+    the behavior is undefined.
+    '''
+    if annotation is _empty:
+        return None, None
+    elif callable(annotation):
+        return annotation, None
+    elif isinstance(annotation, str):
+        return None, annotation
+    elif isinstance(annotation, tuple):
+        try:
+            arg1, arg2 = annotation
+        except ValueError as e:
+            raise AnnotationError(annotation) from e
+        else:
+            if callable(arg1) and isinstance(arg2, str):
+                return arg1, arg2
+            elif isinstance(arg1, str) and callable(arg2):
+                return arg2, arg1
+
+    raise AnnotationError(annotation)
+
+
+def _add_arguments(param, parser, used_char_args, add_nos):
+    '''
+    Add the argument(s) to an ArgumentParser (using add_argument) for a given
+    parameter. used_char_args is the set of -short options currently already in
+    use, and is updated (if necessary) by this function. If add_nos is True,
+    this will also add an inverse switch for all boolean options. For
+    instance, for the boolean parameter "verbose", this will create --verbose
+    and --no-verbose.
+    '''
+
+    # Impl note: This function is kept separate from make_parser because it's
+    # already very long and I wanted to separate out as much as possible into
+    # its own call scope, to prevent even the possibility of suble mutation
+    # bugs.
+    if param.kind is param.POSITIONAL_ONLY:
+        raise PositionalArgError(param)
+    elif param.kind is param.VAR_KEYWORD:
+        raise KWArgError(param)
+
+    # These are the kwargs for the add_argument function.
+    arg_spec = {}
+    is_option = False
+
+    # Get the type and default from the annotation.
+    arg_type, description = _get_type_description(param.annotation)
+
+    # Get the default value
+    default = param.default
+
+    # If there is no explicit type, and the default is present and not None,
+    # infer the type from the default.
+    if arg_type is None and default not in {_empty, None}:
+        arg_type = type(default)
+
+    # Add default. The presence of a default means this is an option, not an
+    # argument.
+    if default is not _empty:
+        arg_spec['default'] = default
+        is_option = True
+
+    # Add the type
+    if arg_type is not None:
+        # Special case for bool: make it just a --switch
+        if arg_type is bool:
+            if not default or default is _empty:
+                arg_spec['action'] = 'store_true'
+            else:
+                arg_spec['action'] = 'store_false'
+
+            # Switches are always options
+            is_option = True
+
+        # Special case for file types: make it a string type, for filename
+        elif isinstance(default, IOBase):
+            arg_spec['type'] = str
+
+        # TODO: special case for list type.
+        #   - How to specificy type of list members?
+        #       - param: [int]
+        #       - param: int =[]
+        #   - action='append' vs nargs='*'
+
+        else:
+            arg_spec['type'] = arg_type
+
+    # nargs: if the signature includes *args, collect them as trailing CLI
+    # arguments in a list. *args can't have a default value, so it can never be
+    # an option.
+    if param.kind is param.VAR_POSITIONAL:
+        # TODO: consider depluralizing metavar/name here.
+        arg_spec['nargs'] = '*'
+
+    # Add description.
+    if description is not None:
+        arg_spec['help'] = description
+
+    # Get the --flags
+    flags = []
+    name = param.name
+
+    if is_option:
+        # Add the first letter as a -short option.
+        for letter in name[0], name[0].swapcase():
+            if letter not in used_char_args:
+                used_char_args.add(letter)
+                flags.append('-{}'.format(letter))
+                break
+
+        # If the parameter is a --long option, or is a -short option that
+        # somehow failed to get a flag, add it.
+        if len(name) > 1 or not flags:
+            flags.append('--{}'.format(name))
+
+        arg_spec['dest'] = name
+    else:
+        flags.append(name)
+
+    parser.add_argument(*flags, **arg_spec)
+
+    # Create the --no- version for boolean switches
+    if add_nos and arg_type is bool:
+        parser.add_argument(
+            '--no-{}'.format(name),
+            action='store_const',
+            dest=name,
+            const=default if default is not _empty else False)
+
+
+def make_parser(func_sig, description, epilog, add_nos):
+    '''
+    Given the signature of a function, create an ArgumentParser
+    '''
+    parser = ArgumentParser(description=description, epilog=epilog)
+
+    used_char_args = {'h'}
+
+    # Arange the params so that single-character arguments are first. This
+    # esnures they don't have to get --long versions. sorted is stable, so the
+    # parameters will otherwise still be in relative order.
+    params = sorted(
+        func_sig.parameters.values(),
+        key=lambda param: len(param.name) > 1)
+
+    for param in params:
+        _add_arguments(param, parser, used_char_args, add_nos)
+
+    return parser
+
+
+_DOCSTRING_SPLIT = compile_regex(r'\n\s*-{4,}\s*\n')
+
+
+def parse_docstring(docstring):
+    '''
+    Given a docstring, parse it into a description and epilog part
+    '''
+    if docstring is None:
+        return '', ''
+
+    parts = _DOCSTRING_SPLIT.split(docstring)
+
+    if len(parts) == 1:
+        return docstring, ''
+    elif len(parts) == 2:
+        return parts[0], parts[1]
+    else:
+        raise TooManySplitsError()
+
+
+def autoparse(
+        func=None, *,
+        description=None,
+        epilog=None,
+        add_nos=False,
+        parser=None):
+    '''
+    This decorator converts a function that takes normal arguments into a
+    function which takes a single optional argument, argv, parses it using an
+    argparse.ArgumentParser, and calls the underlying function with the parsed
+    arguments. If it is not given, sys.argv[1:] is used. This is so that the
+    function can be used as a setuptools entry point, as well as a normal main
+    function. sys.argv[1:] is not evaluated until the function is called, to
+    allow injecting different arguments for testing.
+
+    It uses the argument signature of the function to create an
+    ArgumentParser. Parameters without defaults become positional parameters,
+    while parameters *with* defaults become --options. Use annotations to set
+    the type of the parameter.
+
+    The `desctiption` and `epilog` parameters corrospond to the same respective
+    argparse parameters. If no description is given, it defaults to the
+    decorated functions's docstring, if present.
+
+    If add_nos is True, every boolean option (that is, every parameter with a
+    default of True/False or a type of bool) will have a --no- version created
+    as well, which inverts the option. For instance, the --verbose option will
+    have a --no-verbose counterpart. These are not mutually exclusive-
+    whichever one appears last in the argument list will have precedence.
+
+    If a parser is given, it is used instead of one generated from the function
+    signature. In this case, no parser is created; instead, the given parser is
+    used to parse the argv argument. The parser's results' argument names must
+    match up with the parameter names of the decorated function.
+
+    The decorated function is attached to the result as the `func` attribute,
+    and the parser is attached as the `parser` attribute.
+    '''
+
+    # If @autoparse(...) is used instead of @autoparse
+    if func is None:
+        return lambda f: autoparse(
+            f, description=description,
+            epilog=epilog,
+            add_nos=add_nos,
+            parser=parser)
+
+    func_sig = signature(func)
+
+    docstr_description, docstr_epilog = parse_docstring(getdoc(func))
+
+    if parser is None:
+        parser = make_parser(
+            func_sig,
+            description or docstr_description,
+            epilog or docstr_epilog,
+            add_nos)
+
+    @wraps(func)
+    def autoparse_wrapper(argv=None):
+        if argv is None:
+            argv = sys.argv[1:]
+
+        # Get empty argument binding, to fill with parsed arguments. This
+        # object does all the heavy lifting of turning named arguments into
+        # into correctly bound *args and **kwargs.
+        parsed_args = func_sig.bind_partial()
+        parsed_args.arguments.update(vars(parser.parse_args(argv)))
+
+        return func(*parsed_args.args, **parsed_args.kwargs)
+
+    # TODO: attach an updated __signature__ to autoparse_wrapper, just in case.
+
+    # Attach the wrapped function and parser, and return the wrapper.
+    autoparse_wrapper.func = func
+    autoparse_wrapper.parser = parser
+    return autoparse_wrapper
+
+
+@contextmanager
+def smart_open(filename_or_file, *args, **kwargs):
+    '''
+    This context manager allows you to open a filename, if you want to default
+    some already-existing file object, like sys.stdout, which shouldn't be
+    closed at the end of the context. If the filename argument is a str, bytes,
+    or int, the file object is created via a call to open with the given *args
+    and **kwargs, sent to the context, and closed at the end of the context,
+    just like "with open(filename) as f:". If it isn't one of the openable
+    types, the object simply sent to the context unchanged, and left unclosed
+    at the end of the context. Example:
+
+        def work_with_file(name=sys.stdout):
+            with smart_open(name) as f:
+                # Works correctly if name is a str filename or sys.stdout
+                print("Some stuff", file=f)
+                # If it was a filename, f is closed at the end here.
+    '''
+    if isinstance(filename_or_file, (str, bytes, int)):
+        with open(filename_or_file, *args, **kwargs) as file:
+            yield file
+    else:
+        yield filename_or_file
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/errors.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/errors.py
new file mode 100644
index 00000000..25706073
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/autocommand/errors.py
@@ -0,0 +1,23 @@
+# Copyright 2014-2016 Nathan West
+#
+# This file is part of autocommand.
+#
+# autocommand is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# autocommand is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with autocommand.  If not, see <http://www.gnu.org/licenses/>.
+
+
+class AutocommandError(Exception):
+    '''Base class for autocommand exceptions'''
+    pass
+
+# Individual modules will define errors specific to that module.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/METADATA
new file mode 100644
index 00000000..db0a2dcd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/METADATA
@@ -0,0 +1,46 @@
+Metadata-Version: 2.1
+Name: backports.tarfile
+Version: 1.2.0
+Summary: Backport of CPython tarfile module
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Homepage, https://github.com/jaraco/backports.tarfile
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Provides-Extra: docs
+Requires-Dist: sphinx >=3.5 ; extra == 'docs'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'docs'
+Requires-Dist: rst.linker >=1.9 ; extra == 'docs'
+Requires-Dist: furo ; extra == 'docs'
+Requires-Dist: sphinx-lint ; extra == 'docs'
+Provides-Extra: testing
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'testing'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'testing'
+Requires-Dist: pytest-cov ; extra == 'testing'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'testing'
+Requires-Dist: jaraco.test ; extra == 'testing'
+Requires-Dist: pytest !=8.0.* ; extra == 'testing'
+
+.. image:: https://img.shields.io/pypi/v/backports.tarfile.svg
+   :target: https://pypi.org/project/backports.tarfile
+
+.. image:: https://img.shields.io/pypi/pyversions/backports.tarfile.svg
+
+.. image:: https://github.com/jaraco/backports.tarfile/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/backports.tarfile/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. .. image:: https://readthedocs.org/projects/backportstarfile/badge/?version=latest
+..    :target: https://backportstarfile.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/RECORD
new file mode 100644
index 00000000..536dc2f0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/RECORD
@@ -0,0 +1,17 @@
+backports.tarfile-1.2.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+backports.tarfile-1.2.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+backports.tarfile-1.2.0.dist-info/METADATA,sha256=ghXFTq132dxaEIolxr3HK1mZqm9iyUmaRANZQSr6WlE,2020
+backports.tarfile-1.2.0.dist-info/RECORD,,
+backports.tarfile-1.2.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+backports.tarfile-1.2.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+backports.tarfile-1.2.0.dist-info/top_level.txt,sha256=cGjaLMOoBR1FK0ApojtzWVmViTtJ7JGIK_HwXiEsvtU,10
+backports/__init__.py,sha256=iOEMwnlORWezdO8-2vxBIPSR37D7JGjluZ8f55vzxls,81
+backports/__pycache__/__init__.cpython-312.pyc,,
+backports/tarfile/__init__.py,sha256=Pwf2qUIfB0SolJPCKcx3vz3UEu_aids4g4sAfxy94qg,108491
+backports/tarfile/__main__.py,sha256=Yw2oGT1afrz2eBskzdPYL8ReB_3liApmhFkN2EbDmc4,59
+backports/tarfile/__pycache__/__init__.cpython-312.pyc,,
+backports/tarfile/__pycache__/__main__.cpython-312.pyc,,
+backports/tarfile/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+backports/tarfile/compat/__pycache__/__init__.cpython-312.pyc,,
+backports/tarfile/compat/__pycache__/py38.cpython-312.pyc,,
+backports/tarfile/compat/py38.py,sha256=iYkyt_gvWjLzGUTJD9TuTfMMjOk-ersXZmRlvQYN2qE,568
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/top_level.txt
new file mode 100644
index 00000000..99d2be5b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports.tarfile-1.2.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+backports
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__init__.py
new file mode 100644
index 00000000..0d1f7edf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__init__.py
@@ -0,0 +1 @@
+__path__ = __import__('pkgutil').extend_path(__path__, __name__)  # type: ignore
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..aa012406
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__init__.py
new file mode 100644
index 00000000..8c16881c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__init__.py
@@ -0,0 +1,2937 @@
+#-------------------------------------------------------------------
+# tarfile.py
+#-------------------------------------------------------------------
+# Copyright (C) 2002 Lars Gustaebel <lars@gustaebel.de>
+# All rights reserved.
+#
+# Permission  is  hereby granted,  free  of charge,  to  any person
+# obtaining a  copy of  this software  and associated documentation
+# files  (the  "Software"),  to   deal  in  the  Software   without
+# restriction,  including  without limitation  the  rights to  use,
+# copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies  of  the  Software,  and to  permit  persons  to  whom the
+# Software  is  furnished  to  do  so,  subject  to  the  following
+# conditions:
+#
+# The above copyright  notice and this  permission notice shall  be
+# included in all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS  IS", WITHOUT WARRANTY OF ANY  KIND,
+# EXPRESS OR IMPLIED, INCLUDING  BUT NOT LIMITED TO  THE WARRANTIES
+# OF  MERCHANTABILITY,  FITNESS   FOR  A  PARTICULAR   PURPOSE  AND
+# NONINFRINGEMENT.  IN  NO  EVENT SHALL  THE  AUTHORS  OR COPYRIGHT
+# HOLDERS  BE LIABLE  FOR ANY  CLAIM, DAMAGES  OR OTHER  LIABILITY,
+# WHETHER  IN AN  ACTION OF  CONTRACT, TORT  OR OTHERWISE,  ARISING
+# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+# OTHER DEALINGS IN THE SOFTWARE.
+#
+"""Read from and write to tar format archives.
+"""
+
+version     = "0.9.0"
+__author__  = "Lars Gust\u00e4bel (lars@gustaebel.de)"
+__credits__ = "Gustavo Niemeyer, Niels Gust\u00e4bel, Richard Townsend."
+
+#---------
+# Imports
+#---------
+from builtins import open as bltn_open
+import sys
+import os
+import io
+import shutil
+import stat
+import time
+import struct
+import copy
+import re
+
+from .compat.py38 import removesuffix
+
+try:
+    import pwd
+except ImportError:
+    pwd = None
+try:
+    import grp
+except ImportError:
+    grp = None
+
+# os.symlink on Windows prior to 6.0 raises NotImplementedError
+# OSError (winerror=1314) will be raised if the caller does not hold the
+# SeCreateSymbolicLinkPrivilege privilege
+symlink_exception = (AttributeError, NotImplementedError, OSError)
+
+# from tarfile import *
+__all__ = ["TarFile", "TarInfo", "is_tarfile", "TarError", "ReadError",
+           "CompressionError", "StreamError", "ExtractError", "HeaderError",
+           "ENCODING", "USTAR_FORMAT", "GNU_FORMAT", "PAX_FORMAT",
+           "DEFAULT_FORMAT", "open","fully_trusted_filter", "data_filter",
+           "tar_filter", "FilterError", "AbsoluteLinkError",
+           "OutsideDestinationError", "SpecialFileError", "AbsolutePathError",
+           "LinkOutsideDestinationError"]
+
+
+#---------------------------------------------------------
+# tar constants
+#---------------------------------------------------------
+NUL = b"\0"                     # the null character
+BLOCKSIZE = 512                 # length of processing blocks
+RECORDSIZE = BLOCKSIZE * 20     # length of records
+GNU_MAGIC = b"ustar  \0"        # magic gnu tar string
+POSIX_MAGIC = b"ustar\x0000"    # magic posix tar string
+
+LENGTH_NAME = 100               # maximum length of a filename
+LENGTH_LINK = 100               # maximum length of a linkname
+LENGTH_PREFIX = 155             # maximum length of the prefix field
+
+REGTYPE = b"0"                  # regular file
+AREGTYPE = b"\0"                # regular file
+LNKTYPE = b"1"                  # link (inside tarfile)
+SYMTYPE = b"2"                  # symbolic link
+CHRTYPE = b"3"                  # character special device
+BLKTYPE = b"4"                  # block special device
+DIRTYPE = b"5"                  # directory
+FIFOTYPE = b"6"                 # fifo special device
+CONTTYPE = b"7"                 # contiguous file
+
+GNUTYPE_LONGNAME = b"L"         # GNU tar longname
+GNUTYPE_LONGLINK = b"K"         # GNU tar longlink
+GNUTYPE_SPARSE = b"S"           # GNU tar sparse file
+
+XHDTYPE = b"x"                  # POSIX.1-2001 extended header
+XGLTYPE = b"g"                  # POSIX.1-2001 global header
+SOLARIS_XHDTYPE = b"X"          # Solaris extended header
+
+USTAR_FORMAT = 0                # POSIX.1-1988 (ustar) format
+GNU_FORMAT = 1                  # GNU tar format
+PAX_FORMAT = 2                  # POSIX.1-2001 (pax) format
+DEFAULT_FORMAT = PAX_FORMAT
+
+#---------------------------------------------------------
+# tarfile constants
+#---------------------------------------------------------
+# File types that tarfile supports:
+SUPPORTED_TYPES = (REGTYPE, AREGTYPE, LNKTYPE,
+                   SYMTYPE, DIRTYPE, FIFOTYPE,
+                   CONTTYPE, CHRTYPE, BLKTYPE,
+                   GNUTYPE_LONGNAME, GNUTYPE_LONGLINK,
+                   GNUTYPE_SPARSE)
+
+# File types that will be treated as a regular file.
+REGULAR_TYPES = (REGTYPE, AREGTYPE,
+                 CONTTYPE, GNUTYPE_SPARSE)
+
+# File types that are part of the GNU tar format.
+GNU_TYPES = (GNUTYPE_LONGNAME, GNUTYPE_LONGLINK,
+             GNUTYPE_SPARSE)
+
+# Fields from a pax header that override a TarInfo attribute.
+PAX_FIELDS = ("path", "linkpath", "size", "mtime",
+              "uid", "gid", "uname", "gname")
+
+# Fields from a pax header that are affected by hdrcharset.
+PAX_NAME_FIELDS = {"path", "linkpath", "uname", "gname"}
+
+# Fields in a pax header that are numbers, all other fields
+# are treated as strings.
+PAX_NUMBER_FIELDS = {
+    "atime": float,
+    "ctime": float,
+    "mtime": float,
+    "uid": int,
+    "gid": int,
+    "size": int
+}
+
+#---------------------------------------------------------
+# initialization
+#---------------------------------------------------------
+if os.name == "nt":
+    ENCODING = "utf-8"
+else:
+    ENCODING = sys.getfilesystemencoding()
+
+#---------------------------------------------------------
+# Some useful functions
+#---------------------------------------------------------
+
+def stn(s, length, encoding, errors):
+    """Convert a string to a null-terminated bytes object.
+    """
+    if s is None:
+        raise ValueError("metadata cannot contain None")
+    s = s.encode(encoding, errors)
+    return s[:length] + (length - len(s)) * NUL
+
+def nts(s, encoding, errors):
+    """Convert a null-terminated bytes object to a string.
+    """
+    p = s.find(b"\0")
+    if p != -1:
+        s = s[:p]
+    return s.decode(encoding, errors)
+
+def nti(s):
+    """Convert a number field to a python number.
+    """
+    # There are two possible encodings for a number field, see
+    # itn() below.
+    if s[0] in (0o200, 0o377):
+        n = 0
+        for i in range(len(s) - 1):
+            n <<= 8
+            n += s[i + 1]
+        if s[0] == 0o377:
+            n = -(256 ** (len(s) - 1) - n)
+    else:
+        try:
+            s = nts(s, "ascii", "strict")
+            n = int(s.strip() or "0", 8)
+        except ValueError:
+            raise InvalidHeaderError("invalid header")
+    return n
+
+def itn(n, digits=8, format=DEFAULT_FORMAT):
+    """Convert a python number to a number field.
+    """
+    # POSIX 1003.1-1988 requires numbers to be encoded as a string of
+    # octal digits followed by a null-byte, this allows values up to
+    # (8**(digits-1))-1. GNU tar allows storing numbers greater than
+    # that if necessary. A leading 0o200 or 0o377 byte indicate this
+    # particular encoding, the following digits-1 bytes are a big-endian
+    # base-256 representation. This allows values up to (256**(digits-1))-1.
+    # A 0o200 byte indicates a positive number, a 0o377 byte a negative
+    # number.
+    original_n = n
+    n = int(n)
+    if 0 <= n < 8 ** (digits - 1):
+        s = bytes("%0*o" % (digits - 1, n), "ascii") + NUL
+    elif format == GNU_FORMAT and -256 ** (digits - 1) <= n < 256 ** (digits - 1):
+        if n >= 0:
+            s = bytearray([0o200])
+        else:
+            s = bytearray([0o377])
+            n = 256 ** digits + n
+
+        for i in range(digits - 1):
+            s.insert(1, n & 0o377)
+            n >>= 8
+    else:
+        raise ValueError("overflow in number field")
+
+    return s
+
+def calc_chksums(buf):
+    """Calculate the checksum for a member's header by summing up all
+       characters except for the chksum field which is treated as if
+       it was filled with spaces. According to the GNU tar sources,
+       some tars (Sun and NeXT) calculate chksum with signed char,
+       which will be different if there are chars in the buffer with
+       the high bit set. So we calculate two checksums, unsigned and
+       signed.
+    """
+    unsigned_chksum = 256 + sum(struct.unpack_from("148B8x356B", buf))
+    signed_chksum = 256 + sum(struct.unpack_from("148b8x356b", buf))
+    return unsigned_chksum, signed_chksum
+
+def copyfileobj(src, dst, length=None, exception=OSError, bufsize=None):
+    """Copy length bytes from fileobj src to fileobj dst.
+       If length is None, copy the entire content.
+    """
+    bufsize = bufsize or 16 * 1024
+    if length == 0:
+        return
+    if length is None:
+        shutil.copyfileobj(src, dst, bufsize)
+        return
+
+    blocks, remainder = divmod(length, bufsize)
+    for b in range(blocks):
+        buf = src.read(bufsize)
+        if len(buf) < bufsize:
+            raise exception("unexpected end of data")
+        dst.write(buf)
+
+    if remainder != 0:
+        buf = src.read(remainder)
+        if len(buf) < remainder:
+            raise exception("unexpected end of data")
+        dst.write(buf)
+    return
+
+def _safe_print(s):
+    encoding = getattr(sys.stdout, 'encoding', None)
+    if encoding is not None:
+        s = s.encode(encoding, 'backslashreplace').decode(encoding)
+    print(s, end=' ')
+
+
+class TarError(Exception):
+    """Base exception."""
+    pass
+class ExtractError(TarError):
+    """General exception for extract errors."""
+    pass
+class ReadError(TarError):
+    """Exception for unreadable tar archives."""
+    pass
+class CompressionError(TarError):
+    """Exception for unavailable compression methods."""
+    pass
+class StreamError(TarError):
+    """Exception for unsupported operations on stream-like TarFiles."""
+    pass
+class HeaderError(TarError):
+    """Base exception for header errors."""
+    pass
+class EmptyHeaderError(HeaderError):
+    """Exception for empty headers."""
+    pass
+class TruncatedHeaderError(HeaderError):
+    """Exception for truncated headers."""
+    pass
+class EOFHeaderError(HeaderError):
+    """Exception for end of file headers."""
+    pass
+class InvalidHeaderError(HeaderError):
+    """Exception for invalid headers."""
+    pass
+class SubsequentHeaderError(HeaderError):
+    """Exception for missing and invalid extended headers."""
+    pass
+
+#---------------------------
+# internal stream interface
+#---------------------------
+class _LowLevelFile:
+    """Low-level file object. Supports reading and writing.
+       It is used instead of a regular file object for streaming
+       access.
+    """
+
+    def __init__(self, name, mode):
+        mode = {
+            "r": os.O_RDONLY,
+            "w": os.O_WRONLY | os.O_CREAT | os.O_TRUNC,
+        }[mode]
+        if hasattr(os, "O_BINARY"):
+            mode |= os.O_BINARY
+        self.fd = os.open(name, mode, 0o666)
+
+    def close(self):
+        os.close(self.fd)
+
+    def read(self, size):
+        return os.read(self.fd, size)
+
+    def write(self, s):
+        os.write(self.fd, s)
+
+class _Stream:
+    """Class that serves as an adapter between TarFile and
+       a stream-like object.  The stream-like object only
+       needs to have a read() or write() method that works with bytes,
+       and the method is accessed blockwise.
+       Use of gzip or bzip2 compression is possible.
+       A stream-like object could be for example: sys.stdin.buffer,
+       sys.stdout.buffer, a socket, a tape device etc.
+
+       _Stream is intended to be used only internally.
+    """
+
+    def __init__(self, name, mode, comptype, fileobj, bufsize,
+                 compresslevel):
+        """Construct a _Stream object.
+        """
+        self._extfileobj = True
+        if fileobj is None:
+            fileobj = _LowLevelFile(name, mode)
+            self._extfileobj = False
+
+        if comptype == '*':
+            # Enable transparent compression detection for the
+            # stream interface
+            fileobj = _StreamProxy(fileobj)
+            comptype = fileobj.getcomptype()
+
+        self.name     = name or ""
+        self.mode     = mode
+        self.comptype = comptype
+        self.fileobj  = fileobj
+        self.bufsize  = bufsize
+        self.buf      = b""
+        self.pos      = 0
+        self.closed   = False
+
+        try:
+            if comptype == "gz":
+                try:
+                    import zlib
+                except ImportError:
+                    raise CompressionError("zlib module is not available") from None
+                self.zlib = zlib
+                self.crc = zlib.crc32(b"")
+                if mode == "r":
+                    self.exception = zlib.error
+                    self._init_read_gz()
+                else:
+                    self._init_write_gz(compresslevel)
+
+            elif comptype == "bz2":
+                try:
+                    import bz2
+                except ImportError:
+                    raise CompressionError("bz2 module is not available") from None
+                if mode == "r":
+                    self.dbuf = b""
+                    self.cmp = bz2.BZ2Decompressor()
+                    self.exception = OSError
+                else:
+                    self.cmp = bz2.BZ2Compressor(compresslevel)
+
+            elif comptype == "xz":
+                try:
+                    import lzma
+                except ImportError:
+                    raise CompressionError("lzma module is not available") from None
+                if mode == "r":
+                    self.dbuf = b""
+                    self.cmp = lzma.LZMADecompressor()
+                    self.exception = lzma.LZMAError
+                else:
+                    self.cmp = lzma.LZMACompressor()
+
+            elif comptype != "tar":
+                raise CompressionError("unknown compression type %r" % comptype)
+
+        except:
+            if not self._extfileobj:
+                self.fileobj.close()
+            self.closed = True
+            raise
+
+    def __del__(self):
+        if hasattr(self, "closed") and not self.closed:
+            self.close()
+
+    def _init_write_gz(self, compresslevel):
+        """Initialize for writing with gzip compression.
+        """
+        self.cmp = self.zlib.compressobj(compresslevel,
+                                         self.zlib.DEFLATED,
+                                         -self.zlib.MAX_WBITS,
+                                         self.zlib.DEF_MEM_LEVEL,
+                                         0)
+        timestamp = struct.pack("<L", int(time.time()))
+        self.__write(b"\037\213\010\010" + timestamp + b"\002\377")
+        if self.name.endswith(".gz"):
+            self.name = self.name[:-3]
+        # Honor "directory components removed" from RFC1952
+        self.name = os.path.basename(self.name)
+        # RFC1952 says we must use ISO-8859-1 for the FNAME field.
+        self.__write(self.name.encode("iso-8859-1", "replace") + NUL)
+
+    def write(self, s):
+        """Write string s to the stream.
+        """
+        if self.comptype == "gz":
+            self.crc = self.zlib.crc32(s, self.crc)
+        self.pos += len(s)
+        if self.comptype != "tar":
+            s = self.cmp.compress(s)
+        self.__write(s)
+
+    def __write(self, s):
+        """Write string s to the stream if a whole new block
+           is ready to be written.
+        """
+        self.buf += s
+        while len(self.buf) > self.bufsize:
+            self.fileobj.write(self.buf[:self.bufsize])
+            self.buf = self.buf[self.bufsize:]
+
+    def close(self):
+        """Close the _Stream object. No operation should be
+           done on it afterwards.
+        """
+        if self.closed:
+            return
+
+        self.closed = True
+        try:
+            if self.mode == "w" and self.comptype != "tar":
+                self.buf += self.cmp.flush()
+
+            if self.mode == "w" and self.buf:
+                self.fileobj.write(self.buf)
+                self.buf = b""
+                if self.comptype == "gz":
+                    self.fileobj.write(struct.pack("<L", self.crc))
+                    self.fileobj.write(struct.pack("<L", self.pos & 0xffffFFFF))
+        finally:
+            if not self._extfileobj:
+                self.fileobj.close()
+
+    def _init_read_gz(self):
+        """Initialize for reading a gzip compressed fileobj.
+        """
+        self.cmp = self.zlib.decompressobj(-self.zlib.MAX_WBITS)
+        self.dbuf = b""
+
+        # taken from gzip.GzipFile with some alterations
+        if self.__read(2) != b"\037\213":
+            raise ReadError("not a gzip file")
+        if self.__read(1) != b"\010":
+            raise CompressionError("unsupported compression method")
+
+        flag = ord(self.__read(1))
+        self.__read(6)
+
+        if flag & 4:
+            xlen = ord(self.__read(1)) + 256 * ord(self.__read(1))
+            self.read(xlen)
+        if flag & 8:
+            while True:
+                s = self.__read(1)
+                if not s or s == NUL:
+                    break
+        if flag & 16:
+            while True:
+                s = self.__read(1)
+                if not s or s == NUL:
+                    break
+        if flag & 2:
+            self.__read(2)
+
+    def tell(self):
+        """Return the stream's file pointer position.
+        """
+        return self.pos
+
+    def seek(self, pos=0):
+        """Set the stream's file pointer to pos. Negative seeking
+           is forbidden.
+        """
+        if pos - self.pos >= 0:
+            blocks, remainder = divmod(pos - self.pos, self.bufsize)
+            for i in range(blocks):
+                self.read(self.bufsize)
+            self.read(remainder)
+        else:
+            raise StreamError("seeking backwards is not allowed")
+        return self.pos
+
+    def read(self, size):
+        """Return the next size number of bytes from the stream."""
+        assert size is not None
+        buf = self._read(size)
+        self.pos += len(buf)
+        return buf
+
+    def _read(self, size):
+        """Return size bytes from the stream.
+        """
+        if self.comptype == "tar":
+            return self.__read(size)
+
+        c = len(self.dbuf)
+        t = [self.dbuf]
+        while c < size:
+            # Skip underlying buffer to avoid unaligned double buffering.
+            if self.buf:
+                buf = self.buf
+                self.buf = b""
+            else:
+                buf = self.fileobj.read(self.bufsize)
+                if not buf:
+                    break
+            try:
+                buf = self.cmp.decompress(buf)
+            except self.exception as e:
+                raise ReadError("invalid compressed data") from e
+            t.append(buf)
+            c += len(buf)
+        t = b"".join(t)
+        self.dbuf = t[size:]
+        return t[:size]
+
+    def __read(self, size):
+        """Return size bytes from stream. If internal buffer is empty,
+           read another block from the stream.
+        """
+        c = len(self.buf)
+        t = [self.buf]
+        while c < size:
+            buf = self.fileobj.read(self.bufsize)
+            if not buf:
+                break
+            t.append(buf)
+            c += len(buf)
+        t = b"".join(t)
+        self.buf = t[size:]
+        return t[:size]
+# class _Stream
+
+class _StreamProxy(object):
+    """Small proxy class that enables transparent compression
+       detection for the Stream interface (mode 'r|*').
+    """
+
+    def __init__(self, fileobj):
+        self.fileobj = fileobj
+        self.buf = self.fileobj.read(BLOCKSIZE)
+
+    def read(self, size):
+        self.read = self.fileobj.read
+        return self.buf
+
+    def getcomptype(self):
+        if self.buf.startswith(b"\x1f\x8b\x08"):
+            return "gz"
+        elif self.buf[0:3] == b"BZh" and self.buf[4:10] == b"1AY&SY":
+            return "bz2"
+        elif self.buf.startswith((b"\x5d\x00\x00\x80", b"\xfd7zXZ")):
+            return "xz"
+        else:
+            return "tar"
+
+    def close(self):
+        self.fileobj.close()
+# class StreamProxy
+
+#------------------------
+# Extraction file object
+#------------------------
+class _FileInFile(object):
+    """A thin wrapper around an existing file object that
+       provides a part of its data as an individual file
+       object.
+    """
+
+    def __init__(self, fileobj, offset, size, name, blockinfo=None):
+        self.fileobj = fileobj
+        self.offset = offset
+        self.size = size
+        self.position = 0
+        self.name = name
+        self.closed = False
+
+        if blockinfo is None:
+            blockinfo = [(0, size)]
+
+        # Construct a map with data and zero blocks.
+        self.map_index = 0
+        self.map = []
+        lastpos = 0
+        realpos = self.offset
+        for offset, size in blockinfo:
+            if offset > lastpos:
+                self.map.append((False, lastpos, offset, None))
+            self.map.append((True, offset, offset + size, realpos))
+            realpos += size
+            lastpos = offset + size
+        if lastpos < self.size:
+            self.map.append((False, lastpos, self.size, None))
+
+    def flush(self):
+        pass
+
+    @property
+    def mode(self):
+        return 'rb'
+
+    def readable(self):
+        return True
+
+    def writable(self):
+        return False
+
+    def seekable(self):
+        return self.fileobj.seekable()
+
+    def tell(self):
+        """Return the current file position.
+        """
+        return self.position
+
+    def seek(self, position, whence=io.SEEK_SET):
+        """Seek to a position in the file.
+        """
+        if whence == io.SEEK_SET:
+            self.position = min(max(position, 0), self.size)
+        elif whence == io.SEEK_CUR:
+            if position < 0:
+                self.position = max(self.position + position, 0)
+            else:
+                self.position = min(self.position + position, self.size)
+        elif whence == io.SEEK_END:
+            self.position = max(min(self.size + position, self.size), 0)
+        else:
+            raise ValueError("Invalid argument")
+        return self.position
+
+    def read(self, size=None):
+        """Read data from the file.
+        """
+        if size is None:
+            size = self.size - self.position
+        else:
+            size = min(size, self.size - self.position)
+
+        buf = b""
+        while size > 0:
+            while True:
+                data, start, stop, offset = self.map[self.map_index]
+                if start <= self.position < stop:
+                    break
+                else:
+                    self.map_index += 1
+                    if self.map_index == len(self.map):
+                        self.map_index = 0
+            length = min(size, stop - self.position)
+            if data:
+                self.fileobj.seek(offset + (self.position - start))
+                b = self.fileobj.read(length)
+                if len(b) != length:
+                    raise ReadError("unexpected end of data")
+                buf += b
+            else:
+                buf += NUL * length
+            size -= length
+            self.position += length
+        return buf
+
+    def readinto(self, b):
+        buf = self.read(len(b))
+        b[:len(buf)] = buf
+        return len(buf)
+
+    def close(self):
+        self.closed = True
+#class _FileInFile
+
+class ExFileObject(io.BufferedReader):
+
+    def __init__(self, tarfile, tarinfo):
+        fileobj = _FileInFile(tarfile.fileobj, tarinfo.offset_data,
+                tarinfo.size, tarinfo.name, tarinfo.sparse)
+        super().__init__(fileobj)
+#class ExFileObject
+
+
+#-----------------------------
+# extraction filters (PEP 706)
+#-----------------------------
+
+class FilterError(TarError):
+    pass
+
+class AbsolutePathError(FilterError):
+    def __init__(self, tarinfo):
+        self.tarinfo = tarinfo
+        super().__init__(f'member {tarinfo.name!r} has an absolute path')
+
+class OutsideDestinationError(FilterError):
+    def __init__(self, tarinfo, path):
+        self.tarinfo = tarinfo
+        self._path = path
+        super().__init__(f'{tarinfo.name!r} would be extracted to {path!r}, '
+                         + 'which is outside the destination')
+
+class SpecialFileError(FilterError):
+    def __init__(self, tarinfo):
+        self.tarinfo = tarinfo
+        super().__init__(f'{tarinfo.name!r} is a special file')
+
+class AbsoluteLinkError(FilterError):
+    def __init__(self, tarinfo):
+        self.tarinfo = tarinfo
+        super().__init__(f'{tarinfo.name!r} is a link to an absolute path')
+
+class LinkOutsideDestinationError(FilterError):
+    def __init__(self, tarinfo, path):
+        self.tarinfo = tarinfo
+        self._path = path
+        super().__init__(f'{tarinfo.name!r} would link to {path!r}, '
+                         + 'which is outside the destination')
+
+def _get_filtered_attrs(member, dest_path, for_data=True):
+    new_attrs = {}
+    name = member.name
+    dest_path = os.path.realpath(dest_path)
+    # Strip leading / (tar's directory separator) from filenames.
+    # Include os.sep (target OS directory separator) as well.
+    if name.startswith(('/', os.sep)):
+        name = new_attrs['name'] = member.path.lstrip('/' + os.sep)
+    if os.path.isabs(name):
+        # Path is absolute even after stripping.
+        # For example, 'C:/foo' on Windows.
+        raise AbsolutePathError(member)
+    # Ensure we stay in the destination
+    target_path = os.path.realpath(os.path.join(dest_path, name))
+    if os.path.commonpath([target_path, dest_path]) != dest_path:
+        raise OutsideDestinationError(member, target_path)
+    # Limit permissions (no high bits, and go-w)
+    mode = member.mode
+    if mode is not None:
+        # Strip high bits & group/other write bits
+        mode = mode & 0o755
+        if for_data:
+            # For data, handle permissions & file types
+            if member.isreg() or member.islnk():
+                if not mode & 0o100:
+                    # Clear executable bits if not executable by user
+                    mode &= ~0o111
+                # Ensure owner can read & write
+                mode |= 0o600
+            elif member.isdir() or member.issym():
+                # Ignore mode for directories & symlinks
+                mode = None
+            else:
+                # Reject special files
+                raise SpecialFileError(member)
+        if mode != member.mode:
+            new_attrs['mode'] = mode
+    if for_data:
+        # Ignore ownership for 'data'
+        if member.uid is not None:
+            new_attrs['uid'] = None
+        if member.gid is not None:
+            new_attrs['gid'] = None
+        if member.uname is not None:
+            new_attrs['uname'] = None
+        if member.gname is not None:
+            new_attrs['gname'] = None
+        # Check link destination for 'data'
+        if member.islnk() or member.issym():
+            if os.path.isabs(member.linkname):
+                raise AbsoluteLinkError(member)
+            if member.issym():
+                target_path = os.path.join(dest_path,
+                                           os.path.dirname(name),
+                                           member.linkname)
+            else:
+                target_path = os.path.join(dest_path,
+                                           member.linkname)
+            target_path = os.path.realpath(target_path)
+            if os.path.commonpath([target_path, dest_path]) != dest_path:
+                raise LinkOutsideDestinationError(member, target_path)
+    return new_attrs
+
+def fully_trusted_filter(member, dest_path):
+    return member
+
+def tar_filter(member, dest_path):
+    new_attrs = _get_filtered_attrs(member, dest_path, False)
+    if new_attrs:
+        return member.replace(**new_attrs, deep=False)
+    return member
+
+def data_filter(member, dest_path):
+    new_attrs = _get_filtered_attrs(member, dest_path, True)
+    if new_attrs:
+        return member.replace(**new_attrs, deep=False)
+    return member
+
+_NAMED_FILTERS = {
+    "fully_trusted": fully_trusted_filter,
+    "tar": tar_filter,
+    "data": data_filter,
+}
+
+#------------------
+# Exported Classes
+#------------------
+
+# Sentinel for replace() defaults, meaning "don't change the attribute"
+_KEEP = object()
+
+class TarInfo(object):
+    """Informational class which holds the details about an
+       archive member given by a tar header block.
+       TarInfo objects are returned by TarFile.getmember(),
+       TarFile.getmembers() and TarFile.gettarinfo() and are
+       usually created internally.
+    """
+
+    __slots__ = dict(
+        name = 'Name of the archive member.',
+        mode = 'Permission bits.',
+        uid = 'User ID of the user who originally stored this member.',
+        gid = 'Group ID of the user who originally stored this member.',
+        size = 'Size in bytes.',
+        mtime = 'Time of last modification.',
+        chksum = 'Header checksum.',
+        type = ('File type. type is usually one of these constants: '
+                'REGTYPE, AREGTYPE, LNKTYPE, SYMTYPE, DIRTYPE, FIFOTYPE, '
+                'CONTTYPE, CHRTYPE, BLKTYPE, GNUTYPE_SPARSE.'),
+        linkname = ('Name of the target file name, which is only present '
+                    'in TarInfo objects of type LNKTYPE and SYMTYPE.'),
+        uname = 'User name.',
+        gname = 'Group name.',
+        devmajor = 'Device major number.',
+        devminor = 'Device minor number.',
+        offset = 'The tar header starts here.',
+        offset_data = "The file's data starts here.",
+        pax_headers = ('A dictionary containing key-value pairs of an '
+                       'associated pax extended header.'),
+        sparse = 'Sparse member information.',
+        _tarfile = None,
+        _sparse_structs = None,
+        _link_target = None,
+        )
+
+    def __init__(self, name=""):
+        """Construct a TarInfo object. name is the optional name
+           of the member.
+        """
+        self.name = name        # member name
+        self.mode = 0o644       # file permissions
+        self.uid = 0            # user id
+        self.gid = 0            # group id
+        self.size = 0           # file size
+        self.mtime = 0          # modification time
+        self.chksum = 0         # header checksum
+        self.type = REGTYPE     # member type
+        self.linkname = ""      # link name
+        self.uname = ""         # user name
+        self.gname = ""         # group name
+        self.devmajor = 0       # device major number
+        self.devminor = 0       # device minor number
+
+        self.offset = 0         # the tar header starts here
+        self.offset_data = 0    # the file's data starts here
+
+        self.sparse = None      # sparse member information
+        self.pax_headers = {}   # pax header information
+
+    @property
+    def tarfile(self):
+        import warnings
+        warnings.warn(
+            'The undocumented "tarfile" attribute of TarInfo objects '
+            + 'is deprecated and will be removed in Python 3.16',
+            DeprecationWarning, stacklevel=2)
+        return self._tarfile
+
+    @tarfile.setter
+    def tarfile(self, tarfile):
+        import warnings
+        warnings.warn(
+            'The undocumented "tarfile" attribute of TarInfo objects '
+            + 'is deprecated and will be removed in Python 3.16',
+            DeprecationWarning, stacklevel=2)
+        self._tarfile = tarfile
+
+    @property
+    def path(self):
+        'In pax headers, "name" is called "path".'
+        return self.name
+
+    @path.setter
+    def path(self, name):
+        self.name = name
+
+    @property
+    def linkpath(self):
+        'In pax headers, "linkname" is called "linkpath".'
+        return self.linkname
+
+    @linkpath.setter
+    def linkpath(self, linkname):
+        self.linkname = linkname
+
+    def __repr__(self):
+        return "<%s %r at %#x>" % (self.__class__.__name__,self.name,id(self))
+
+    def replace(self, *,
+                name=_KEEP, mtime=_KEEP, mode=_KEEP, linkname=_KEEP,
+                uid=_KEEP, gid=_KEEP, uname=_KEEP, gname=_KEEP,
+                deep=True, _KEEP=_KEEP):
+        """Return a deep copy of self with the given attributes replaced.
+        """
+        if deep:
+            result = copy.deepcopy(self)
+        else:
+            result = copy.copy(self)
+        if name is not _KEEP:
+            result.name = name
+        if mtime is not _KEEP:
+            result.mtime = mtime
+        if mode is not _KEEP:
+            result.mode = mode
+        if linkname is not _KEEP:
+            result.linkname = linkname
+        if uid is not _KEEP:
+            result.uid = uid
+        if gid is not _KEEP:
+            result.gid = gid
+        if uname is not _KEEP:
+            result.uname = uname
+        if gname is not _KEEP:
+            result.gname = gname
+        return result
+
+    def get_info(self):
+        """Return the TarInfo's attributes as a dictionary.
+        """
+        if self.mode is None:
+            mode = None
+        else:
+            mode = self.mode & 0o7777
+        info = {
+            "name":     self.name,
+            "mode":     mode,
+            "uid":      self.uid,
+            "gid":      self.gid,
+            "size":     self.size,
+            "mtime":    self.mtime,
+            "chksum":   self.chksum,
+            "type":     self.type,
+            "linkname": self.linkname,
+            "uname":    self.uname,
+            "gname":    self.gname,
+            "devmajor": self.devmajor,
+            "devminor": self.devminor
+        }
+
+        if info["type"] == DIRTYPE and not info["name"].endswith("/"):
+            info["name"] += "/"
+
+        return info
+
+    def tobuf(self, format=DEFAULT_FORMAT, encoding=ENCODING, errors="surrogateescape"):
+        """Return a tar header as a string of 512 byte blocks.
+        """
+        info = self.get_info()
+        for name, value in info.items():
+            if value is None:
+                raise ValueError("%s may not be None" % name)
+
+        if format == USTAR_FORMAT:
+            return self.create_ustar_header(info, encoding, errors)
+        elif format == GNU_FORMAT:
+            return self.create_gnu_header(info, encoding, errors)
+        elif format == PAX_FORMAT:
+            return self.create_pax_header(info, encoding)
+        else:
+            raise ValueError("invalid format")
+
+    def create_ustar_header(self, info, encoding, errors):
+        """Return the object as a ustar header block.
+        """
+        info["magic"] = POSIX_MAGIC
+
+        if len(info["linkname"].encode(encoding, errors)) > LENGTH_LINK:
+            raise ValueError("linkname is too long")
+
+        if len(info["name"].encode(encoding, errors)) > LENGTH_NAME:
+            info["prefix"], info["name"] = self._posix_split_name(info["name"], encoding, errors)
+
+        return self._create_header(info, USTAR_FORMAT, encoding, errors)
+
+    def create_gnu_header(self, info, encoding, errors):
+        """Return the object as a GNU header block sequence.
+        """
+        info["magic"] = GNU_MAGIC
+
+        buf = b""
+        if len(info["linkname"].encode(encoding, errors)) > LENGTH_LINK:
+            buf += self._create_gnu_long_header(info["linkname"], GNUTYPE_LONGLINK, encoding, errors)
+
+        if len(info["name"].encode(encoding, errors)) > LENGTH_NAME:
+            buf += self._create_gnu_long_header(info["name"], GNUTYPE_LONGNAME, encoding, errors)
+
+        return buf + self._create_header(info, GNU_FORMAT, encoding, errors)
+
+    def create_pax_header(self, info, encoding):
+        """Return the object as a ustar header block. If it cannot be
+           represented this way, prepend a pax extended header sequence
+           with supplement information.
+        """
+        info["magic"] = POSIX_MAGIC
+        pax_headers = self.pax_headers.copy()
+
+        # Test string fields for values that exceed the field length or cannot
+        # be represented in ASCII encoding.
+        for name, hname, length in (
+                ("name", "path", LENGTH_NAME), ("linkname", "linkpath", LENGTH_LINK),
+                ("uname", "uname", 32), ("gname", "gname", 32)):
+
+            if hname in pax_headers:
+                # The pax header has priority.
+                continue
+
+            # Try to encode the string as ASCII.
+            try:
+                info[name].encode("ascii", "strict")
+            except UnicodeEncodeError:
+                pax_headers[hname] = info[name]
+                continue
+
+            if len(info[name]) > length:
+                pax_headers[hname] = info[name]
+
+        # Test number fields for values that exceed the field limit or values
+        # that like to be stored as float.
+        for name, digits in (("uid", 8), ("gid", 8), ("size", 12), ("mtime", 12)):
+            needs_pax = False
+
+            val = info[name]
+            val_is_float = isinstance(val, float)
+            val_int = round(val) if val_is_float else val
+            if not 0 <= val_int < 8 ** (digits - 1):
+                # Avoid overflow.
+                info[name] = 0
+                needs_pax = True
+            elif val_is_float:
+                # Put rounded value in ustar header, and full
+                # precision value in pax header.
+                info[name] = val_int
+                needs_pax = True
+
+            # The existing pax header has priority.
+            if needs_pax and name not in pax_headers:
+                pax_headers[name] = str(val)
+
+        # Create a pax extended header if necessary.
+        if pax_headers:
+            buf = self._create_pax_generic_header(pax_headers, XHDTYPE, encoding)
+        else:
+            buf = b""
+
+        return buf + self._create_header(info, USTAR_FORMAT, "ascii", "replace")
+
+    @classmethod
+    def create_pax_global_header(cls, pax_headers):
+        """Return the object as a pax global header block sequence.
+        """
+        return cls._create_pax_generic_header(pax_headers, XGLTYPE, "utf-8")
+
+    def _posix_split_name(self, name, encoding, errors):
+        """Split a name longer than 100 chars into a prefix
+           and a name part.
+        """
+        components = name.split("/")
+        for i in range(1, len(components)):
+            prefix = "/".join(components[:i])
+            name = "/".join(components[i:])
+            if len(prefix.encode(encoding, errors)) <= LENGTH_PREFIX and \
+                    len(name.encode(encoding, errors)) <= LENGTH_NAME:
+                break
+        else:
+            raise ValueError("name is too long")
+
+        return prefix, name
+
+    @staticmethod
+    def _create_header(info, format, encoding, errors):
+        """Return a header block. info is a dictionary with file
+           information, format must be one of the *_FORMAT constants.
+        """
+        has_device_fields = info.get("type") in (CHRTYPE, BLKTYPE)
+        if has_device_fields:
+            devmajor = itn(info.get("devmajor", 0), 8, format)
+            devminor = itn(info.get("devminor", 0), 8, format)
+        else:
+            devmajor = stn("", 8, encoding, errors)
+            devminor = stn("", 8, encoding, errors)
+
+        # None values in metadata should cause ValueError.
+        # itn()/stn() do this for all fields except type.
+        filetype = info.get("type", REGTYPE)
+        if filetype is None:
+            raise ValueError("TarInfo.type must not be None")
+
+        parts = [
+            stn(info.get("name", ""), 100, encoding, errors),
+            itn(info.get("mode", 0) & 0o7777, 8, format),
+            itn(info.get("uid", 0), 8, format),
+            itn(info.get("gid", 0), 8, format),
+            itn(info.get("size", 0), 12, format),
+            itn(info.get("mtime", 0), 12, format),
+            b"        ", # checksum field
+            filetype,
+            stn(info.get("linkname", ""), 100, encoding, errors),
+            info.get("magic", POSIX_MAGIC),
+            stn(info.get("uname", ""), 32, encoding, errors),
+            stn(info.get("gname", ""), 32, encoding, errors),
+            devmajor,
+            devminor,
+            stn(info.get("prefix", ""), 155, encoding, errors)
+        ]
+
+        buf = struct.pack("%ds" % BLOCKSIZE, b"".join(parts))
+        chksum = calc_chksums(buf[-BLOCKSIZE:])[0]
+        buf = buf[:-364] + bytes("%06o\0" % chksum, "ascii") + buf[-357:]
+        return buf
+
+    @staticmethod
+    def _create_payload(payload):
+        """Return the string payload filled with zero bytes
+           up to the next 512 byte border.
+        """
+        blocks, remainder = divmod(len(payload), BLOCKSIZE)
+        if remainder > 0:
+            payload += (BLOCKSIZE - remainder) * NUL
+        return payload
+
+    @classmethod
+    def _create_gnu_long_header(cls, name, type, encoding, errors):
+        """Return a GNUTYPE_LONGNAME or GNUTYPE_LONGLINK sequence
+           for name.
+        """
+        name = name.encode(encoding, errors) + NUL
+
+        info = {}
+        info["name"] = "././@LongLink"
+        info["type"] = type
+        info["size"] = len(name)
+        info["magic"] = GNU_MAGIC
+
+        # create extended header + name blocks.
+        return cls._create_header(info, USTAR_FORMAT, encoding, errors) + \
+                cls._create_payload(name)
+
+    @classmethod
+    def _create_pax_generic_header(cls, pax_headers, type, encoding):
+        """Return a POSIX.1-2008 extended or global header sequence
+           that contains a list of keyword, value pairs. The values
+           must be strings.
+        """
+        # Check if one of the fields contains surrogate characters and thereby
+        # forces hdrcharset=BINARY, see _proc_pax() for more information.
+        binary = False
+        for keyword, value in pax_headers.items():
+            try:
+                value.encode("utf-8", "strict")
+            except UnicodeEncodeError:
+                binary = True
+                break
+
+        records = b""
+        if binary:
+            # Put the hdrcharset field at the beginning of the header.
+            records += b"21 hdrcharset=BINARY\n"
+
+        for keyword, value in pax_headers.items():
+            keyword = keyword.encode("utf-8")
+            if binary:
+                # Try to restore the original byte representation of 'value'.
+                # Needless to say, that the encoding must match the string.
+                value = value.encode(encoding, "surrogateescape")
+            else:
+                value = value.encode("utf-8")
+
+            l = len(keyword) + len(value) + 3   # ' ' + '=' + '\n'
+            n = p = 0
+            while True:
+                n = l + len(str(p))
+                if n == p:
+                    break
+                p = n
+            records += bytes(str(p), "ascii") + b" " + keyword + b"=" + value + b"\n"
+
+        # We use a hardcoded "././@PaxHeader" name like star does
+        # instead of the one that POSIX recommends.
+        info = {}
+        info["name"] = "././@PaxHeader"
+        info["type"] = type
+        info["size"] = len(records)
+        info["magic"] = POSIX_MAGIC
+
+        # Create pax header + record blocks.
+        return cls._create_header(info, USTAR_FORMAT, "ascii", "replace") + \
+                cls._create_payload(records)
+
+    @classmethod
+    def frombuf(cls, buf, encoding, errors):
+        """Construct a TarInfo object from a 512 byte bytes object.
+        """
+        if len(buf) == 0:
+            raise EmptyHeaderError("empty header")
+        if len(buf) != BLOCKSIZE:
+            raise TruncatedHeaderError("truncated header")
+        if buf.count(NUL) == BLOCKSIZE:
+            raise EOFHeaderError("end of file header")
+
+        chksum = nti(buf[148:156])
+        if chksum not in calc_chksums(buf):
+            raise InvalidHeaderError("bad checksum")
+
+        obj = cls()
+        obj.name = nts(buf[0:100], encoding, errors)
+        obj.mode = nti(buf[100:108])
+        obj.uid = nti(buf[108:116])
+        obj.gid = nti(buf[116:124])
+        obj.size = nti(buf[124:136])
+        obj.mtime = nti(buf[136:148])
+        obj.chksum = chksum
+        obj.type = buf[156:157]
+        obj.linkname = nts(buf[157:257], encoding, errors)
+        obj.uname = nts(buf[265:297], encoding, errors)
+        obj.gname = nts(buf[297:329], encoding, errors)
+        obj.devmajor = nti(buf[329:337])
+        obj.devminor = nti(buf[337:345])
+        prefix = nts(buf[345:500], encoding, errors)
+
+        # Old V7 tar format represents a directory as a regular
+        # file with a trailing slash.
+        if obj.type == AREGTYPE and obj.name.endswith("/"):
+            obj.type = DIRTYPE
+
+        # The old GNU sparse format occupies some of the unused
+        # space in the buffer for up to 4 sparse structures.
+        # Save them for later processing in _proc_sparse().
+        if obj.type == GNUTYPE_SPARSE:
+            pos = 386
+            structs = []
+            for i in range(4):
+                try:
+                    offset = nti(buf[pos:pos + 12])
+                    numbytes = nti(buf[pos + 12:pos + 24])
+                except ValueError:
+                    break
+                structs.append((offset, numbytes))
+                pos += 24
+            isextended = bool(buf[482])
+            origsize = nti(buf[483:495])
+            obj._sparse_structs = (structs, isextended, origsize)
+
+        # Remove redundant slashes from directories.
+        if obj.isdir():
+            obj.name = obj.name.rstrip("/")
+
+        # Reconstruct a ustar longname.
+        if prefix and obj.type not in GNU_TYPES:
+            obj.name = prefix + "/" + obj.name
+        return obj
+
+    @classmethod
+    def fromtarfile(cls, tarfile):
+        """Return the next TarInfo object from TarFile object
+           tarfile.
+        """
+        buf = tarfile.fileobj.read(BLOCKSIZE)
+        obj = cls.frombuf(buf, tarfile.encoding, tarfile.errors)
+        obj.offset = tarfile.fileobj.tell() - BLOCKSIZE
+        return obj._proc_member(tarfile)
+
+    #--------------------------------------------------------------------------
+    # The following are methods that are called depending on the type of a
+    # member. The entry point is _proc_member() which can be overridden in a
+    # subclass to add custom _proc_*() methods. A _proc_*() method MUST
+    # implement the following
+    # operations:
+    # 1. Set self.offset_data to the position where the data blocks begin,
+    #    if there is data that follows.
+    # 2. Set tarfile.offset to the position where the next member's header will
+    #    begin.
+    # 3. Return self or another valid TarInfo object.
+    def _proc_member(self, tarfile):
+        """Choose the right processing method depending on
+           the type and call it.
+        """
+        if self.type in (GNUTYPE_LONGNAME, GNUTYPE_LONGLINK):
+            return self._proc_gnulong(tarfile)
+        elif self.type == GNUTYPE_SPARSE:
+            return self._proc_sparse(tarfile)
+        elif self.type in (XHDTYPE, XGLTYPE, SOLARIS_XHDTYPE):
+            return self._proc_pax(tarfile)
+        else:
+            return self._proc_builtin(tarfile)
+
+    def _proc_builtin(self, tarfile):
+        """Process a builtin type or an unknown type which
+           will be treated as a regular file.
+        """
+        self.offset_data = tarfile.fileobj.tell()
+        offset = self.offset_data
+        if self.isreg() or self.type not in SUPPORTED_TYPES:
+            # Skip the following data blocks.
+            offset += self._block(self.size)
+        tarfile.offset = offset
+
+        # Patch the TarInfo object with saved global
+        # header information.
+        self._apply_pax_info(tarfile.pax_headers, tarfile.encoding, tarfile.errors)
+
+        # Remove redundant slashes from directories. This is to be consistent
+        # with frombuf().
+        if self.isdir():
+            self.name = self.name.rstrip("/")
+
+        return self
+
+    def _proc_gnulong(self, tarfile):
+        """Process the blocks that hold a GNU longname
+           or longlink member.
+        """
+        buf = tarfile.fileobj.read(self._block(self.size))
+
+        # Fetch the next header and process it.
+        try:
+            next = self.fromtarfile(tarfile)
+        except HeaderError as e:
+            raise SubsequentHeaderError(str(e)) from None
+
+        # Patch the TarInfo object from the next header with
+        # the longname information.
+        next.offset = self.offset
+        if self.type == GNUTYPE_LONGNAME:
+            next.name = nts(buf, tarfile.encoding, tarfile.errors)
+        elif self.type == GNUTYPE_LONGLINK:
+            next.linkname = nts(buf, tarfile.encoding, tarfile.errors)
+
+        # Remove redundant slashes from directories. This is to be consistent
+        # with frombuf().
+        if next.isdir():
+            next.name = removesuffix(next.name, "/")
+
+        return next
+
+    def _proc_sparse(self, tarfile):
+        """Process a GNU sparse header plus extra headers.
+        """
+        # We already collected some sparse structures in frombuf().
+        structs, isextended, origsize = self._sparse_structs
+        del self._sparse_structs
+
+        # Collect sparse structures from extended header blocks.
+        while isextended:
+            buf = tarfile.fileobj.read(BLOCKSIZE)
+            pos = 0
+            for i in range(21):
+                try:
+                    offset = nti(buf[pos:pos + 12])
+                    numbytes = nti(buf[pos + 12:pos + 24])
+                except ValueError:
+                    break
+                if offset and numbytes:
+                    structs.append((offset, numbytes))
+                pos += 24
+            isextended = bool(buf[504])
+        self.sparse = structs
+
+        self.offset_data = tarfile.fileobj.tell()
+        tarfile.offset = self.offset_data + self._block(self.size)
+        self.size = origsize
+        return self
+
+    def _proc_pax(self, tarfile):
+        """Process an extended or global header as described in
+           POSIX.1-2008.
+        """
+        # Read the header information.
+        buf = tarfile.fileobj.read(self._block(self.size))
+
+        # A pax header stores supplemental information for either
+        # the following file (extended) or all following files
+        # (global).
+        if self.type == XGLTYPE:
+            pax_headers = tarfile.pax_headers
+        else:
+            pax_headers = tarfile.pax_headers.copy()
+
+        # Check if the pax header contains a hdrcharset field. This tells us
+        # the encoding of the path, linkpath, uname and gname fields. Normally,
+        # these fields are UTF-8 encoded but since POSIX.1-2008 tar
+        # implementations are allowed to store them as raw binary strings if
+        # the translation to UTF-8 fails.
+        match = re.search(br"\d+ hdrcharset=([^\n]+)\n", buf)
+        if match is not None:
+            pax_headers["hdrcharset"] = match.group(1).decode("utf-8")
+
+        # For the time being, we don't care about anything other than "BINARY".
+        # The only other value that is currently allowed by the standard is
+        # "ISO-IR 10646 2000 UTF-8" in other words UTF-8.
+        hdrcharset = pax_headers.get("hdrcharset")
+        if hdrcharset == "BINARY":
+            encoding = tarfile.encoding
+        else:
+            encoding = "utf-8"
+
+        # Parse pax header information. A record looks like that:
+        # "%d %s=%s\n" % (length, keyword, value). length is the size
+        # of the complete record including the length field itself and
+        # the newline. keyword and value are both UTF-8 encoded strings.
+        regex = re.compile(br"(\d+) ([^=]+)=")
+        pos = 0
+        while match := regex.match(buf, pos):
+            length, keyword = match.groups()
+            length = int(length)
+            if length == 0:
+                raise InvalidHeaderError("invalid header")
+            value = buf[match.end(2) + 1:match.start(1) + length - 1]
+
+            # Normally, we could just use "utf-8" as the encoding and "strict"
+            # as the error handler, but we better not take the risk. For
+            # example, GNU tar <= 1.23 is known to store filenames it cannot
+            # translate to UTF-8 as raw strings (unfortunately without a
+            # hdrcharset=BINARY header).
+            # We first try the strict standard encoding, and if that fails we
+            # fall back on the user's encoding and error handler.
+            keyword = self._decode_pax_field(keyword, "utf-8", "utf-8",
+                    tarfile.errors)
+            if keyword in PAX_NAME_FIELDS:
+                value = self._decode_pax_field(value, encoding, tarfile.encoding,
+                        tarfile.errors)
+            else:
+                value = self._decode_pax_field(value, "utf-8", "utf-8",
+                        tarfile.errors)
+
+            pax_headers[keyword] = value
+            pos += length
+
+        # Fetch the next header.
+        try:
+            next = self.fromtarfile(tarfile)
+        except HeaderError as e:
+            raise SubsequentHeaderError(str(e)) from None
+
+        # Process GNU sparse information.
+        if "GNU.sparse.map" in pax_headers:
+            # GNU extended sparse format version 0.1.
+            self._proc_gnusparse_01(next, pax_headers)
+
+        elif "GNU.sparse.size" in pax_headers:
+            # GNU extended sparse format version 0.0.
+            self._proc_gnusparse_00(next, pax_headers, buf)
+
+        elif pax_headers.get("GNU.sparse.major") == "1" and pax_headers.get("GNU.sparse.minor") == "0":
+            # GNU extended sparse format version 1.0.
+            self._proc_gnusparse_10(next, pax_headers, tarfile)
+
+        if self.type in (XHDTYPE, SOLARIS_XHDTYPE):
+            # Patch the TarInfo object with the extended header info.
+            next._apply_pax_info(pax_headers, tarfile.encoding, tarfile.errors)
+            next.offset = self.offset
+
+            if "size" in pax_headers:
+                # If the extended header replaces the size field,
+                # we need to recalculate the offset where the next
+                # header starts.
+                offset = next.offset_data
+                if next.isreg() or next.type not in SUPPORTED_TYPES:
+                    offset += next._block(next.size)
+                tarfile.offset = offset
+
+        return next
+
+    def _proc_gnusparse_00(self, next, pax_headers, buf):
+        """Process a GNU tar extended sparse header, version 0.0.
+        """
+        offsets = []
+        for match in re.finditer(br"\d+ GNU.sparse.offset=(\d+)\n", buf):
+            offsets.append(int(match.group(1)))
+        numbytes = []
+        for match in re.finditer(br"\d+ GNU.sparse.numbytes=(\d+)\n", buf):
+            numbytes.append(int(match.group(1)))
+        next.sparse = list(zip(offsets, numbytes))
+
+    def _proc_gnusparse_01(self, next, pax_headers):
+        """Process a GNU tar extended sparse header, version 0.1.
+        """
+        sparse = [int(x) for x in pax_headers["GNU.sparse.map"].split(",")]
+        next.sparse = list(zip(sparse[::2], sparse[1::2]))
+
+    def _proc_gnusparse_10(self, next, pax_headers, tarfile):
+        """Process a GNU tar extended sparse header, version 1.0.
+        """
+        fields = None
+        sparse = []
+        buf = tarfile.fileobj.read(BLOCKSIZE)
+        fields, buf = buf.split(b"\n", 1)
+        fields = int(fields)
+        while len(sparse) < fields * 2:
+            if b"\n" not in buf:
+                buf += tarfile.fileobj.read(BLOCKSIZE)
+            number, buf = buf.split(b"\n", 1)
+            sparse.append(int(number))
+        next.offset_data = tarfile.fileobj.tell()
+        next.sparse = list(zip(sparse[::2], sparse[1::2]))
+
+    def _apply_pax_info(self, pax_headers, encoding, errors):
+        """Replace fields with supplemental information from a previous
+           pax extended or global header.
+        """
+        for keyword, value in pax_headers.items():
+            if keyword == "GNU.sparse.name":
+                setattr(self, "path", value)
+            elif keyword == "GNU.sparse.size":
+                setattr(self, "size", int(value))
+            elif keyword == "GNU.sparse.realsize":
+                setattr(self, "size", int(value))
+            elif keyword in PAX_FIELDS:
+                if keyword in PAX_NUMBER_FIELDS:
+                    try:
+                        value = PAX_NUMBER_FIELDS[keyword](value)
+                    except ValueError:
+                        value = 0
+                if keyword == "path":
+                    value = value.rstrip("/")
+                setattr(self, keyword, value)
+
+        self.pax_headers = pax_headers.copy()
+
+    def _decode_pax_field(self, value, encoding, fallback_encoding, fallback_errors):
+        """Decode a single field from a pax record.
+        """
+        try:
+            return value.decode(encoding, "strict")
+        except UnicodeDecodeError:
+            return value.decode(fallback_encoding, fallback_errors)
+
+    def _block(self, count):
+        """Round up a byte count by BLOCKSIZE and return it,
+           e.g. _block(834) => 1024.
+        """
+        blocks, remainder = divmod(count, BLOCKSIZE)
+        if remainder:
+            blocks += 1
+        return blocks * BLOCKSIZE
+
+    def isreg(self):
+        'Return True if the Tarinfo object is a regular file.'
+        return self.type in REGULAR_TYPES
+
+    def isfile(self):
+        'Return True if the Tarinfo object is a regular file.'
+        return self.isreg()
+
+    def isdir(self):
+        'Return True if it is a directory.'
+        return self.type == DIRTYPE
+
+    def issym(self):
+        'Return True if it is a symbolic link.'
+        return self.type == SYMTYPE
+
+    def islnk(self):
+        'Return True if it is a hard link.'
+        return self.type == LNKTYPE
+
+    def ischr(self):
+        'Return True if it is a character device.'
+        return self.type == CHRTYPE
+
+    def isblk(self):
+        'Return True if it is a block device.'
+        return self.type == BLKTYPE
+
+    def isfifo(self):
+        'Return True if it is a FIFO.'
+        return self.type == FIFOTYPE
+
+    def issparse(self):
+        return self.sparse is not None
+
+    def isdev(self):
+        'Return True if it is one of character device, block device or FIFO.'
+        return self.type in (CHRTYPE, BLKTYPE, FIFOTYPE)
+# class TarInfo
+
+class TarFile(object):
+    """The TarFile Class provides an interface to tar archives.
+    """
+
+    debug = 0                   # May be set from 0 (no msgs) to 3 (all msgs)
+
+    dereference = False         # If true, add content of linked file to the
+                                # tar file, else the link.
+
+    ignore_zeros = False        # If true, skips empty or invalid blocks and
+                                # continues processing.
+
+    errorlevel = 1              # If 0, fatal errors only appear in debug
+                                # messages (if debug >= 0). If > 0, errors
+                                # are passed to the caller as exceptions.
+
+    format = DEFAULT_FORMAT     # The format to use when creating an archive.
+
+    encoding = ENCODING         # Encoding for 8-bit character strings.
+
+    errors = None               # Error handler for unicode conversion.
+
+    tarinfo = TarInfo           # The default TarInfo class to use.
+
+    fileobject = ExFileObject   # The file-object for extractfile().
+
+    extraction_filter = None    # The default filter for extraction.
+
+    def __init__(self, name=None, mode="r", fileobj=None, format=None,
+            tarinfo=None, dereference=None, ignore_zeros=None, encoding=None,
+            errors="surrogateescape", pax_headers=None, debug=None,
+            errorlevel=None, copybufsize=None, stream=False):
+        """Open an (uncompressed) tar archive 'name'. 'mode' is either 'r' to
+           read from an existing archive, 'a' to append data to an existing
+           file or 'w' to create a new file overwriting an existing one. 'mode'
+           defaults to 'r'.
+           If 'fileobj' is given, it is used for reading or writing data. If it
+           can be determined, 'mode' is overridden by 'fileobj's mode.
+           'fileobj' is not closed, when TarFile is closed.
+        """
+        modes = {"r": "rb", "a": "r+b", "w": "wb", "x": "xb"}
+        if mode not in modes:
+            raise ValueError("mode must be 'r', 'a', 'w' or 'x'")
+        self.mode = mode
+        self._mode = modes[mode]
+
+        if not fileobj:
+            if self.mode == "a" and not os.path.exists(name):
+                # Create nonexistent files in append mode.
+                self.mode = "w"
+                self._mode = "wb"
+            fileobj = bltn_open(name, self._mode)
+            self._extfileobj = False
+        else:
+            if (name is None and hasattr(fileobj, "name") and
+                isinstance(fileobj.name, (str, bytes))):
+                name = fileobj.name
+            if hasattr(fileobj, "mode"):
+                self._mode = fileobj.mode
+            self._extfileobj = True
+        self.name = os.path.abspath(name) if name else None
+        self.fileobj = fileobj
+
+        self.stream = stream
+
+        # Init attributes.
+        if format is not None:
+            self.format = format
+        if tarinfo is not None:
+            self.tarinfo = tarinfo
+        if dereference is not None:
+            self.dereference = dereference
+        if ignore_zeros is not None:
+            self.ignore_zeros = ignore_zeros
+        if encoding is not None:
+            self.encoding = encoding
+        self.errors = errors
+
+        if pax_headers is not None and self.format == PAX_FORMAT:
+            self.pax_headers = pax_headers
+        else:
+            self.pax_headers = {}
+
+        if debug is not None:
+            self.debug = debug
+        if errorlevel is not None:
+            self.errorlevel = errorlevel
+
+        # Init datastructures.
+        self.copybufsize = copybufsize
+        self.closed = False
+        self.members = []       # list of members as TarInfo objects
+        self._loaded = False    # flag if all members have been read
+        self.offset = self.fileobj.tell()
+                                # current position in the archive file
+        self.inodes = {}        # dictionary caching the inodes of
+                                # archive members already added
+
+        try:
+            if self.mode == "r":
+                self.firstmember = None
+                self.firstmember = self.next()
+
+            if self.mode == "a":
+                # Move to the end of the archive,
+                # before the first empty block.
+                while True:
+                    self.fileobj.seek(self.offset)
+                    try:
+                        tarinfo = self.tarinfo.fromtarfile(self)
+                        self.members.append(tarinfo)
+                    except EOFHeaderError:
+                        self.fileobj.seek(self.offset)
+                        break
+                    except HeaderError as e:
+                        raise ReadError(str(e)) from None
+
+            if self.mode in ("a", "w", "x"):
+                self._loaded = True
+
+                if self.pax_headers:
+                    buf = self.tarinfo.create_pax_global_header(self.pax_headers.copy())
+                    self.fileobj.write(buf)
+                    self.offset += len(buf)
+        except:
+            if not self._extfileobj:
+                self.fileobj.close()
+            self.closed = True
+            raise
+
+    #--------------------------------------------------------------------------
+    # Below are the classmethods which act as alternate constructors to the
+    # TarFile class. The open() method is the only one that is needed for
+    # public use; it is the "super"-constructor and is able to select an
+    # adequate "sub"-constructor for a particular compression using the mapping
+    # from OPEN_METH.
+    #
+    # This concept allows one to subclass TarFile without losing the comfort of
+    # the super-constructor. A sub-constructor is registered and made available
+    # by adding it to the mapping in OPEN_METH.
+
+    @classmethod
+    def open(cls, name=None, mode="r", fileobj=None, bufsize=RECORDSIZE, **kwargs):
+        r"""Open a tar archive for reading, writing or appending. Return
+           an appropriate TarFile class.
+
+           mode:
+           'r' or 'r:\*' open for reading with transparent compression
+           'r:'         open for reading exclusively uncompressed
+           'r:gz'       open for reading with gzip compression
+           'r:bz2'      open for reading with bzip2 compression
+           'r:xz'       open for reading with lzma compression
+           'a' or 'a:'  open for appending, creating the file if necessary
+           'w' or 'w:'  open for writing without compression
+           'w:gz'       open for writing with gzip compression
+           'w:bz2'      open for writing with bzip2 compression
+           'w:xz'       open for writing with lzma compression
+
+           'x' or 'x:'  create a tarfile exclusively without compression, raise
+                        an exception if the file is already created
+           'x:gz'       create a gzip compressed tarfile, raise an exception
+                        if the file is already created
+           'x:bz2'      create a bzip2 compressed tarfile, raise an exception
+                        if the file is already created
+           'x:xz'       create an lzma compressed tarfile, raise an exception
+                        if the file is already created
+
+           'r|\*'        open a stream of tar blocks with transparent compression
+           'r|'         open an uncompressed stream of tar blocks for reading
+           'r|gz'       open a gzip compressed stream of tar blocks
+           'r|bz2'      open a bzip2 compressed stream of tar blocks
+           'r|xz'       open an lzma compressed stream of tar blocks
+           'w|'         open an uncompressed stream for writing
+           'w|gz'       open a gzip compressed stream for writing
+           'w|bz2'      open a bzip2 compressed stream for writing
+           'w|xz'       open an lzma compressed stream for writing
+        """
+
+        if not name and not fileobj:
+            raise ValueError("nothing to open")
+
+        if mode in ("r", "r:*"):
+            # Find out which *open() is appropriate for opening the file.
+            def not_compressed(comptype):
+                return cls.OPEN_METH[comptype] == 'taropen'
+            error_msgs = []
+            for comptype in sorted(cls.OPEN_METH, key=not_compressed):
+                func = getattr(cls, cls.OPEN_METH[comptype])
+                if fileobj is not None:
+                    saved_pos = fileobj.tell()
+                try:
+                    return func(name, "r", fileobj, **kwargs)
+                except (ReadError, CompressionError) as e:
+                    error_msgs.append(f'- method {comptype}: {e!r}')
+                    if fileobj is not None:
+                        fileobj.seek(saved_pos)
+                    continue
+            error_msgs_summary = '\n'.join(error_msgs)
+            raise ReadError(f"file could not be opened successfully:\n{error_msgs_summary}")
+
+        elif ":" in mode:
+            filemode, comptype = mode.split(":", 1)
+            filemode = filemode or "r"
+            comptype = comptype or "tar"
+
+            # Select the *open() function according to
+            # given compression.
+            if comptype in cls.OPEN_METH:
+                func = getattr(cls, cls.OPEN_METH[comptype])
+            else:
+                raise CompressionError("unknown compression type %r" % comptype)
+            return func(name, filemode, fileobj, **kwargs)
+
+        elif "|" in mode:
+            filemode, comptype = mode.split("|", 1)
+            filemode = filemode or "r"
+            comptype = comptype or "tar"
+
+            if filemode not in ("r", "w"):
+                raise ValueError("mode must be 'r' or 'w'")
+
+            compresslevel = kwargs.pop("compresslevel", 9)
+            stream = _Stream(name, filemode, comptype, fileobj, bufsize,
+                             compresslevel)
+            try:
+                t = cls(name, filemode, stream, **kwargs)
+            except:
+                stream.close()
+                raise
+            t._extfileobj = False
+            return t
+
+        elif mode in ("a", "w", "x"):
+            return cls.taropen(name, mode, fileobj, **kwargs)
+
+        raise ValueError("undiscernible mode")
+
+    @classmethod
+    def taropen(cls, name, mode="r", fileobj=None, **kwargs):
+        """Open uncompressed tar archive name for reading or writing.
+        """
+        if mode not in ("r", "a", "w", "x"):
+            raise ValueError("mode must be 'r', 'a', 'w' or 'x'")
+        return cls(name, mode, fileobj, **kwargs)
+
+    @classmethod
+    def gzopen(cls, name, mode="r", fileobj=None, compresslevel=9, **kwargs):
+        """Open gzip compressed tar archive name for reading or writing.
+           Appending is not allowed.
+        """
+        if mode not in ("r", "w", "x"):
+            raise ValueError("mode must be 'r', 'w' or 'x'")
+
+        try:
+            from gzip import GzipFile
+        except ImportError:
+            raise CompressionError("gzip module is not available") from None
+
+        try:
+            fileobj = GzipFile(name, mode + "b", compresslevel, fileobj)
+        except OSError as e:
+            if fileobj is not None and mode == 'r':
+                raise ReadError("not a gzip file") from e
+            raise
+
+        try:
+            t = cls.taropen(name, mode, fileobj, **kwargs)
+        except OSError as e:
+            fileobj.close()
+            if mode == 'r':
+                raise ReadError("not a gzip file") from e
+            raise
+        except:
+            fileobj.close()
+            raise
+        t._extfileobj = False
+        return t
+
+    @classmethod
+    def bz2open(cls, name, mode="r", fileobj=None, compresslevel=9, **kwargs):
+        """Open bzip2 compressed tar archive name for reading or writing.
+           Appending is not allowed.
+        """
+        if mode not in ("r", "w", "x"):
+            raise ValueError("mode must be 'r', 'w' or 'x'")
+
+        try:
+            from bz2 import BZ2File
+        except ImportError:
+            raise CompressionError("bz2 module is not available") from None
+
+        fileobj = BZ2File(fileobj or name, mode, compresslevel=compresslevel)
+
+        try:
+            t = cls.taropen(name, mode, fileobj, **kwargs)
+        except (OSError, EOFError) as e:
+            fileobj.close()
+            if mode == 'r':
+                raise ReadError("not a bzip2 file") from e
+            raise
+        except:
+            fileobj.close()
+            raise
+        t._extfileobj = False
+        return t
+
+    @classmethod
+    def xzopen(cls, name, mode="r", fileobj=None, preset=None, **kwargs):
+        """Open lzma compressed tar archive name for reading or writing.
+           Appending is not allowed.
+        """
+        if mode not in ("r", "w", "x"):
+            raise ValueError("mode must be 'r', 'w' or 'x'")
+
+        try:
+            from lzma import LZMAFile, LZMAError
+        except ImportError:
+            raise CompressionError("lzma module is not available") from None
+
+        fileobj = LZMAFile(fileobj or name, mode, preset=preset)
+
+        try:
+            t = cls.taropen(name, mode, fileobj, **kwargs)
+        except (LZMAError, EOFError) as e:
+            fileobj.close()
+            if mode == 'r':
+                raise ReadError("not an lzma file") from e
+            raise
+        except:
+            fileobj.close()
+            raise
+        t._extfileobj = False
+        return t
+
+    # All *open() methods are registered here.
+    OPEN_METH = {
+        "tar": "taropen",   # uncompressed tar
+        "gz":  "gzopen",    # gzip compressed tar
+        "bz2": "bz2open",   # bzip2 compressed tar
+        "xz":  "xzopen"     # lzma compressed tar
+    }
+
+    #--------------------------------------------------------------------------
+    # The public methods which TarFile provides:
+
+    def close(self):
+        """Close the TarFile. In write-mode, two finishing zero blocks are
+           appended to the archive.
+        """
+        if self.closed:
+            return
+
+        self.closed = True
+        try:
+            if self.mode in ("a", "w", "x"):
+                self.fileobj.write(NUL * (BLOCKSIZE * 2))
+                self.offset += (BLOCKSIZE * 2)
+                # fill up the end with zero-blocks
+                # (like option -b20 for tar does)
+                blocks, remainder = divmod(self.offset, RECORDSIZE)
+                if remainder > 0:
+                    self.fileobj.write(NUL * (RECORDSIZE - remainder))
+        finally:
+            if not self._extfileobj:
+                self.fileobj.close()
+
+    def getmember(self, name):
+        """Return a TarInfo object for member 'name'. If 'name' can not be
+           found in the archive, KeyError is raised. If a member occurs more
+           than once in the archive, its last occurrence is assumed to be the
+           most up-to-date version.
+        """
+        tarinfo = self._getmember(name.rstrip('/'))
+        if tarinfo is None:
+            raise KeyError("filename %r not found" % name)
+        return tarinfo
+
+    def getmembers(self):
+        """Return the members of the archive as a list of TarInfo objects. The
+           list has the same order as the members in the archive.
+        """
+        self._check()
+        if not self._loaded:    # if we want to obtain a list of
+            self._load()        # all members, we first have to
+                                # scan the whole archive.
+        return self.members
+
+    def getnames(self):
+        """Return the members of the archive as a list of their names. It has
+           the same order as the list returned by getmembers().
+        """
+        return [tarinfo.name for tarinfo in self.getmembers()]
+
+    def gettarinfo(self, name=None, arcname=None, fileobj=None):
+        """Create a TarInfo object from the result of os.stat or equivalent
+           on an existing file. The file is either named by 'name', or
+           specified as a file object 'fileobj' with a file descriptor. If
+           given, 'arcname' specifies an alternative name for the file in the
+           archive, otherwise, the name is taken from the 'name' attribute of
+           'fileobj', or the 'name' argument. The name should be a text
+           string.
+        """
+        self._check("awx")
+
+        # When fileobj is given, replace name by
+        # fileobj's real name.
+        if fileobj is not None:
+            name = fileobj.name
+
+        # Building the name of the member in the archive.
+        # Backward slashes are converted to forward slashes,
+        # Absolute paths are turned to relative paths.
+        if arcname is None:
+            arcname = name
+        drv, arcname = os.path.splitdrive(arcname)
+        arcname = arcname.replace(os.sep, "/")
+        arcname = arcname.lstrip("/")
+
+        # Now, fill the TarInfo object with
+        # information specific for the file.
+        tarinfo = self.tarinfo()
+        tarinfo._tarfile = self  # To be removed in 3.16.
+
+        # Use os.stat or os.lstat, depending on if symlinks shall be resolved.
+        if fileobj is None:
+            if not self.dereference:
+                statres = os.lstat(name)
+            else:
+                statres = os.stat(name)
+        else:
+            statres = os.fstat(fileobj.fileno())
+        linkname = ""
+
+        stmd = statres.st_mode
+        if stat.S_ISREG(stmd):
+            inode = (statres.st_ino, statres.st_dev)
+            if not self.dereference and statres.st_nlink > 1 and \
+                    inode in self.inodes and arcname != self.inodes[inode]:
+                # Is it a hardlink to an already
+                # archived file?
+                type = LNKTYPE
+                linkname = self.inodes[inode]
+            else:
+                # The inode is added only if its valid.
+                # For win32 it is always 0.
+                type = REGTYPE
+                if inode[0]:
+                    self.inodes[inode] = arcname
+        elif stat.S_ISDIR(stmd):
+            type = DIRTYPE
+        elif stat.S_ISFIFO(stmd):
+            type = FIFOTYPE
+        elif stat.S_ISLNK(stmd):
+            type = SYMTYPE
+            linkname = os.readlink(name)
+        elif stat.S_ISCHR(stmd):
+            type = CHRTYPE
+        elif stat.S_ISBLK(stmd):
+            type = BLKTYPE
+        else:
+            return None
+
+        # Fill the TarInfo object with all
+        # information we can get.
+        tarinfo.name = arcname
+        tarinfo.mode = stmd
+        tarinfo.uid = statres.st_uid
+        tarinfo.gid = statres.st_gid
+        if type == REGTYPE:
+            tarinfo.size = statres.st_size
+        else:
+            tarinfo.size = 0
+        tarinfo.mtime = statres.st_mtime
+        tarinfo.type = type
+        tarinfo.linkname = linkname
+        if pwd:
+            try:
+                tarinfo.uname = pwd.getpwuid(tarinfo.uid)[0]
+            except KeyError:
+                pass
+        if grp:
+            try:
+                tarinfo.gname = grp.getgrgid(tarinfo.gid)[0]
+            except KeyError:
+                pass
+
+        if type in (CHRTYPE, BLKTYPE):
+            if hasattr(os, "major") and hasattr(os, "minor"):
+                tarinfo.devmajor = os.major(statres.st_rdev)
+                tarinfo.devminor = os.minor(statres.st_rdev)
+        return tarinfo
+
+    def list(self, verbose=True, *, members=None):
+        """Print a table of contents to sys.stdout. If 'verbose' is False, only
+           the names of the members are printed. If it is True, an 'ls -l'-like
+           output is produced. 'members' is optional and must be a subset of the
+           list returned by getmembers().
+        """
+        # Convert tarinfo type to stat type.
+        type2mode = {REGTYPE: stat.S_IFREG, SYMTYPE: stat.S_IFLNK,
+                     FIFOTYPE: stat.S_IFIFO, CHRTYPE: stat.S_IFCHR,
+                     DIRTYPE: stat.S_IFDIR, BLKTYPE: stat.S_IFBLK}
+        self._check()
+
+        if members is None:
+            members = self
+        for tarinfo in members:
+            if verbose:
+                if tarinfo.mode is None:
+                    _safe_print("??????????")
+                else:
+                    modetype = type2mode.get(tarinfo.type, 0)
+                    _safe_print(stat.filemode(modetype | tarinfo.mode))
+                _safe_print("%s/%s" % (tarinfo.uname or tarinfo.uid,
+                                       tarinfo.gname or tarinfo.gid))
+                if tarinfo.ischr() or tarinfo.isblk():
+                    _safe_print("%10s" %
+                            ("%d,%d" % (tarinfo.devmajor, tarinfo.devminor)))
+                else:
+                    _safe_print("%10d" % tarinfo.size)
+                if tarinfo.mtime is None:
+                    _safe_print("????-??-?? ??:??:??")
+                else:
+                    _safe_print("%d-%02d-%02d %02d:%02d:%02d" \
+                                % time.localtime(tarinfo.mtime)[:6])
+
+            _safe_print(tarinfo.name + ("/" if tarinfo.isdir() else ""))
+
+            if verbose:
+                if tarinfo.issym():
+                    _safe_print("-> " + tarinfo.linkname)
+                if tarinfo.islnk():
+                    _safe_print("link to " + tarinfo.linkname)
+            print()
+
+    def add(self, name, arcname=None, recursive=True, *, filter=None):
+        """Add the file 'name' to the archive. 'name' may be any type of file
+           (directory, fifo, symbolic link, etc.). If given, 'arcname'
+           specifies an alternative name for the file in the archive.
+           Directories are added recursively by default. This can be avoided by
+           setting 'recursive' to False. 'filter' is a function
+           that expects a TarInfo object argument and returns the changed
+           TarInfo object, if it returns None the TarInfo object will be
+           excluded from the archive.
+        """
+        self._check("awx")
+
+        if arcname is None:
+            arcname = name
+
+        # Skip if somebody tries to archive the archive...
+        if self.name is not None and os.path.abspath(name) == self.name:
+            self._dbg(2, "tarfile: Skipped %r" % name)
+            return
+
+        self._dbg(1, name)
+
+        # Create a TarInfo object from the file.
+        tarinfo = self.gettarinfo(name, arcname)
+
+        if tarinfo is None:
+            self._dbg(1, "tarfile: Unsupported type %r" % name)
+            return
+
+        # Change or exclude the TarInfo object.
+        if filter is not None:
+            tarinfo = filter(tarinfo)
+            if tarinfo is None:
+                self._dbg(2, "tarfile: Excluded %r" % name)
+                return
+
+        # Append the tar header and data to the archive.
+        if tarinfo.isreg():
+            with bltn_open(name, "rb") as f:
+                self.addfile(tarinfo, f)
+
+        elif tarinfo.isdir():
+            self.addfile(tarinfo)
+            if recursive:
+                for f in sorted(os.listdir(name)):
+                    self.add(os.path.join(name, f), os.path.join(arcname, f),
+                            recursive, filter=filter)
+
+        else:
+            self.addfile(tarinfo)
+
+    def addfile(self, tarinfo, fileobj=None):
+        """Add the TarInfo object 'tarinfo' to the archive. If 'tarinfo' represents
+           a non zero-size regular file, the 'fileobj' argument should be a binary file,
+           and tarinfo.size bytes are read from it and added to the archive.
+           You can create TarInfo objects directly, or by using gettarinfo().
+        """
+        self._check("awx")
+
+        if fileobj is None and tarinfo.isreg() and tarinfo.size != 0:
+            raise ValueError("fileobj not provided for non zero-size regular file")
+
+        tarinfo = copy.copy(tarinfo)
+
+        buf = tarinfo.tobuf(self.format, self.encoding, self.errors)
+        self.fileobj.write(buf)
+        self.offset += len(buf)
+        bufsize=self.copybufsize
+        # If there's data to follow, append it.
+        if fileobj is not None:
+            copyfileobj(fileobj, self.fileobj, tarinfo.size, bufsize=bufsize)
+            blocks, remainder = divmod(tarinfo.size, BLOCKSIZE)
+            if remainder > 0:
+                self.fileobj.write(NUL * (BLOCKSIZE - remainder))
+                blocks += 1
+            self.offset += blocks * BLOCKSIZE
+
+        self.members.append(tarinfo)
+
+    def _get_filter_function(self, filter):
+        if filter is None:
+            filter = self.extraction_filter
+            if filter is None:
+                import warnings
+                warnings.warn(
+                    'Python 3.14 will, by default, filter extracted tar '
+                    + 'archives and reject files or modify their metadata. '
+                    + 'Use the filter argument to control this behavior.',
+                    DeprecationWarning, stacklevel=3)
+                return fully_trusted_filter
+            if isinstance(filter, str):
+                raise TypeError(
+                    'String names are not supported for '
+                    + 'TarFile.extraction_filter. Use a function such as '
+                    + 'tarfile.data_filter directly.')
+            return filter
+        if callable(filter):
+            return filter
+        try:
+            return _NAMED_FILTERS[filter]
+        except KeyError:
+            raise ValueError(f"filter {filter!r} not found") from None
+
+    def extractall(self, path=".", members=None, *, numeric_owner=False,
+                   filter=None):
+        """Extract all members from the archive to the current working
+           directory and set owner, modification time and permissions on
+           directories afterwards. 'path' specifies a different directory
+           to extract to. 'members' is optional and must be a subset of the
+           list returned by getmembers(). If 'numeric_owner' is True, only
+           the numbers for user/group names are used and not the names.
+
+           The 'filter' function will be called on each member just
+           before extraction.
+           It can return a changed TarInfo or None to skip the member.
+           String names of common filters are accepted.
+        """
+        directories = []
+
+        filter_function = self._get_filter_function(filter)
+        if members is None:
+            members = self
+
+        for member in members:
+            tarinfo = self._get_extract_tarinfo(member, filter_function, path)
+            if tarinfo is None:
+                continue
+            if tarinfo.isdir():
+                # For directories, delay setting attributes until later,
+                # since permissions can interfere with extraction and
+                # extracting contents can reset mtime.
+                directories.append(tarinfo)
+            self._extract_one(tarinfo, path, set_attrs=not tarinfo.isdir(),
+                              numeric_owner=numeric_owner)
+
+        # Reverse sort directories.
+        directories.sort(key=lambda a: a.name, reverse=True)
+
+        # Set correct owner, mtime and filemode on directories.
+        for tarinfo in directories:
+            dirpath = os.path.join(path, tarinfo.name)
+            try:
+                self.chown(tarinfo, dirpath, numeric_owner=numeric_owner)
+                self.utime(tarinfo, dirpath)
+                self.chmod(tarinfo, dirpath)
+            except ExtractError as e:
+                self._handle_nonfatal_error(e)
+
+    def extract(self, member, path="", set_attrs=True, *, numeric_owner=False,
+                filter=None):
+        """Extract a member from the archive to the current working directory,
+           using its full name. Its file information is extracted as accurately
+           as possible. 'member' may be a filename or a TarInfo object. You can
+           specify a different directory using 'path'. File attributes (owner,
+           mtime, mode) are set unless 'set_attrs' is False. If 'numeric_owner'
+           is True, only the numbers for user/group names are used and not
+           the names.
+
+           The 'filter' function will be called before extraction.
+           It can return a changed TarInfo or None to skip the member.
+           String names of common filters are accepted.
+        """
+        filter_function = self._get_filter_function(filter)
+        tarinfo = self._get_extract_tarinfo(member, filter_function, path)
+        if tarinfo is not None:
+            self._extract_one(tarinfo, path, set_attrs, numeric_owner)
+
+    def _get_extract_tarinfo(self, member, filter_function, path):
+        """Get filtered TarInfo (or None) from member, which might be a str"""
+        if isinstance(member, str):
+            tarinfo = self.getmember(member)
+        else:
+            tarinfo = member
+
+        unfiltered = tarinfo
+        try:
+            tarinfo = filter_function(tarinfo, path)
+        except (OSError, FilterError) as e:
+            self._handle_fatal_error(e)
+        except ExtractError as e:
+            self._handle_nonfatal_error(e)
+        if tarinfo is None:
+            self._dbg(2, "tarfile: Excluded %r" % unfiltered.name)
+            return None
+        # Prepare the link target for makelink().
+        if tarinfo.islnk():
+            tarinfo = copy.copy(tarinfo)
+            tarinfo._link_target = os.path.join(path, tarinfo.linkname)
+        return tarinfo
+
+    def _extract_one(self, tarinfo, path, set_attrs, numeric_owner):
+        """Extract from filtered tarinfo to disk"""
+        self._check("r")
+
+        try:
+            self._extract_member(tarinfo, os.path.join(path, tarinfo.name),
+                                 set_attrs=set_attrs,
+                                 numeric_owner=numeric_owner)
+        except OSError as e:
+            self._handle_fatal_error(e)
+        except ExtractError as e:
+            self._handle_nonfatal_error(e)
+
+    def _handle_nonfatal_error(self, e):
+        """Handle non-fatal error (ExtractError) according to errorlevel"""
+        if self.errorlevel > 1:
+            raise
+        else:
+            self._dbg(1, "tarfile: %s" % e)
+
+    def _handle_fatal_error(self, e):
+        """Handle "fatal" error according to self.errorlevel"""
+        if self.errorlevel > 0:
+            raise
+        elif isinstance(e, OSError):
+            if e.filename is None:
+                self._dbg(1, "tarfile: %s" % e.strerror)
+            else:
+                self._dbg(1, "tarfile: %s %r" % (e.strerror, e.filename))
+        else:
+            self._dbg(1, "tarfile: %s %s" % (type(e).__name__, e))
+
+    def extractfile(self, member):
+        """Extract a member from the archive as a file object. 'member' may be
+           a filename or a TarInfo object. If 'member' is a regular file or
+           a link, an io.BufferedReader object is returned. For all other
+           existing members, None is returned. If 'member' does not appear
+           in the archive, KeyError is raised.
+        """
+        self._check("r")
+
+        if isinstance(member, str):
+            tarinfo = self.getmember(member)
+        else:
+            tarinfo = member
+
+        if tarinfo.isreg() or tarinfo.type not in SUPPORTED_TYPES:
+            # Members with unknown types are treated as regular files.
+            return self.fileobject(self, tarinfo)
+
+        elif tarinfo.islnk() or tarinfo.issym():
+            if isinstance(self.fileobj, _Stream):
+                # A small but ugly workaround for the case that someone tries
+                # to extract a (sym)link as a file-object from a non-seekable
+                # stream of tar blocks.
+                raise StreamError("cannot extract (sym)link as file object")
+            else:
+                # A (sym)link's file object is its target's file object.
+                return self.extractfile(self._find_link_target(tarinfo))
+        else:
+            # If there's no data associated with the member (directory, chrdev,
+            # blkdev, etc.), return None instead of a file object.
+            return None
+
+    def _extract_member(self, tarinfo, targetpath, set_attrs=True,
+                        numeric_owner=False):
+        """Extract the TarInfo object tarinfo to a physical
+           file called targetpath.
+        """
+        # Fetch the TarInfo object for the given name
+        # and build the destination pathname, replacing
+        # forward slashes to platform specific separators.
+        targetpath = targetpath.rstrip("/")
+        targetpath = targetpath.replace("/", os.sep)
+
+        # Create all upper directories.
+        upperdirs = os.path.dirname(targetpath)
+        if upperdirs and not os.path.exists(upperdirs):
+            # Create directories that are not part of the archive with
+            # default permissions.
+            os.makedirs(upperdirs, exist_ok=True)
+
+        if tarinfo.islnk() or tarinfo.issym():
+            self._dbg(1, "%s -> %s" % (tarinfo.name, tarinfo.linkname))
+        else:
+            self._dbg(1, tarinfo.name)
+
+        if tarinfo.isreg():
+            self.makefile(tarinfo, targetpath)
+        elif tarinfo.isdir():
+            self.makedir(tarinfo, targetpath)
+        elif tarinfo.isfifo():
+            self.makefifo(tarinfo, targetpath)
+        elif tarinfo.ischr() or tarinfo.isblk():
+            self.makedev(tarinfo, targetpath)
+        elif tarinfo.islnk() or tarinfo.issym():
+            self.makelink(tarinfo, targetpath)
+        elif tarinfo.type not in SUPPORTED_TYPES:
+            self.makeunknown(tarinfo, targetpath)
+        else:
+            self.makefile(tarinfo, targetpath)
+
+        if set_attrs:
+            self.chown(tarinfo, targetpath, numeric_owner)
+            if not tarinfo.issym():
+                self.chmod(tarinfo, targetpath)
+                self.utime(tarinfo, targetpath)
+
+    #--------------------------------------------------------------------------
+    # Below are the different file methods. They are called via
+    # _extract_member() when extract() is called. They can be replaced in a
+    # subclass to implement other functionality.
+
+    def makedir(self, tarinfo, targetpath):
+        """Make a directory called targetpath.
+        """
+        try:
+            if tarinfo.mode is None:
+                # Use the system's default mode
+                os.mkdir(targetpath)
+            else:
+                # Use a safe mode for the directory, the real mode is set
+                # later in _extract_member().
+                os.mkdir(targetpath, 0o700)
+        except FileExistsError:
+            if not os.path.isdir(targetpath):
+                raise
+
+    def makefile(self, tarinfo, targetpath):
+        """Make a file called targetpath.
+        """
+        source = self.fileobj
+        source.seek(tarinfo.offset_data)
+        bufsize = self.copybufsize
+        with bltn_open(targetpath, "wb") as target:
+            if tarinfo.sparse is not None:
+                for offset, size in tarinfo.sparse:
+                    target.seek(offset)
+                    copyfileobj(source, target, size, ReadError, bufsize)
+                target.seek(tarinfo.size)
+                target.truncate()
+            else:
+                copyfileobj(source, target, tarinfo.size, ReadError, bufsize)
+
+    def makeunknown(self, tarinfo, targetpath):
+        """Make a file from a TarInfo object with an unknown type
+           at targetpath.
+        """
+        self.makefile(tarinfo, targetpath)
+        self._dbg(1, "tarfile: Unknown file type %r, " \
+                     "extracted as regular file." % tarinfo.type)
+
+    def makefifo(self, tarinfo, targetpath):
+        """Make a fifo called targetpath.
+        """
+        if hasattr(os, "mkfifo"):
+            os.mkfifo(targetpath)
+        else:
+            raise ExtractError("fifo not supported by system")
+
+    def makedev(self, tarinfo, targetpath):
+        """Make a character or block device called targetpath.
+        """
+        if not hasattr(os, "mknod") or not hasattr(os, "makedev"):
+            raise ExtractError("special devices not supported by system")
+
+        mode = tarinfo.mode
+        if mode is None:
+            # Use mknod's default
+            mode = 0o600
+        if tarinfo.isblk():
+            mode |= stat.S_IFBLK
+        else:
+            mode |= stat.S_IFCHR
+
+        os.mknod(targetpath, mode,
+                 os.makedev(tarinfo.devmajor, tarinfo.devminor))
+
+    def makelink(self, tarinfo, targetpath):
+        """Make a (symbolic) link called targetpath. If it cannot be created
+          (platform limitation), we try to make a copy of the referenced file
+          instead of a link.
+        """
+        try:
+            # For systems that support symbolic and hard links.
+            if tarinfo.issym():
+                if os.path.lexists(targetpath):
+                    # Avoid FileExistsError on following os.symlink.
+                    os.unlink(targetpath)
+                os.symlink(tarinfo.linkname, targetpath)
+            else:
+                if os.path.exists(tarinfo._link_target):
+                    os.link(tarinfo._link_target, targetpath)
+                else:
+                    self._extract_member(self._find_link_target(tarinfo),
+                                         targetpath)
+        except symlink_exception:
+            try:
+                self._extract_member(self._find_link_target(tarinfo),
+                                     targetpath)
+            except KeyError:
+                raise ExtractError("unable to resolve link inside archive") from None
+
+    def chown(self, tarinfo, targetpath, numeric_owner):
+        """Set owner of targetpath according to tarinfo. If numeric_owner
+           is True, use .gid/.uid instead of .gname/.uname. If numeric_owner
+           is False, fall back to .gid/.uid when the search based on name
+           fails.
+        """
+        if hasattr(os, "geteuid") and os.geteuid() == 0:
+            # We have to be root to do so.
+            g = tarinfo.gid
+            u = tarinfo.uid
+            if not numeric_owner:
+                try:
+                    if grp and tarinfo.gname:
+                        g = grp.getgrnam(tarinfo.gname)[2]
+                except KeyError:
+                    pass
+                try:
+                    if pwd and tarinfo.uname:
+                        u = pwd.getpwnam(tarinfo.uname)[2]
+                except KeyError:
+                    pass
+            if g is None:
+                g = -1
+            if u is None:
+                u = -1
+            try:
+                if tarinfo.issym() and hasattr(os, "lchown"):
+                    os.lchown(targetpath, u, g)
+                else:
+                    os.chown(targetpath, u, g)
+            except (OSError, OverflowError) as e:
+                # OverflowError can be raised if an ID doesn't fit in 'id_t'
+                raise ExtractError("could not change owner") from e
+
+    def chmod(self, tarinfo, targetpath):
+        """Set file permissions of targetpath according to tarinfo.
+        """
+        if tarinfo.mode is None:
+            return
+        try:
+            os.chmod(targetpath, tarinfo.mode)
+        except OSError as e:
+            raise ExtractError("could not change mode") from e
+
+    def utime(self, tarinfo, targetpath):
+        """Set modification time of targetpath according to tarinfo.
+        """
+        mtime = tarinfo.mtime
+        if mtime is None:
+            return
+        if not hasattr(os, 'utime'):
+            return
+        try:
+            os.utime(targetpath, (mtime, mtime))
+        except OSError as e:
+            raise ExtractError("could not change modification time") from e
+
+    #--------------------------------------------------------------------------
+    def next(self):
+        """Return the next member of the archive as a TarInfo object, when
+           TarFile is opened for reading. Return None if there is no more
+           available.
+        """
+        self._check("ra")
+        if self.firstmember is not None:
+            m = self.firstmember
+            self.firstmember = None
+            return m
+
+        # Advance the file pointer.
+        if self.offset != self.fileobj.tell():
+            if self.offset == 0:
+                return None
+            self.fileobj.seek(self.offset - 1)
+            if not self.fileobj.read(1):
+                raise ReadError("unexpected end of data")
+
+        # Read the next block.
+        tarinfo = None
+        while True:
+            try:
+                tarinfo = self.tarinfo.fromtarfile(self)
+            except EOFHeaderError as e:
+                if self.ignore_zeros:
+                    self._dbg(2, "0x%X: %s" % (self.offset, e))
+                    self.offset += BLOCKSIZE
+                    continue
+            except InvalidHeaderError as e:
+                if self.ignore_zeros:
+                    self._dbg(2, "0x%X: %s" % (self.offset, e))
+                    self.offset += BLOCKSIZE
+                    continue
+                elif self.offset == 0:
+                    raise ReadError(str(e)) from None
+            except EmptyHeaderError:
+                if self.offset == 0:
+                    raise ReadError("empty file") from None
+            except TruncatedHeaderError as e:
+                if self.offset == 0:
+                    raise ReadError(str(e)) from None
+            except SubsequentHeaderError as e:
+                raise ReadError(str(e)) from None
+            except Exception as e:
+                try:
+                    import zlib
+                    if isinstance(e, zlib.error):
+                        raise ReadError(f'zlib error: {e}') from None
+                    else:
+                        raise e
+                except ImportError:
+                    raise e
+            break
+
+        if tarinfo is not None:
+            # if streaming the file we do not want to cache the tarinfo
+            if not self.stream:
+                self.members.append(tarinfo)
+        else:
+            self._loaded = True
+
+        return tarinfo
+
+    #--------------------------------------------------------------------------
+    # Little helper methods:
+
+    def _getmember(self, name, tarinfo=None, normalize=False):
+        """Find an archive member by name from bottom to top.
+           If tarinfo is given, it is used as the starting point.
+        """
+        # Ensure that all members have been loaded.
+        members = self.getmembers()
+
+        # Limit the member search list up to tarinfo.
+        skipping = False
+        if tarinfo is not None:
+            try:
+                index = members.index(tarinfo)
+            except ValueError:
+                # The given starting point might be a (modified) copy.
+                # We'll later skip members until we find an equivalent.
+                skipping = True
+            else:
+                # Happy fast path
+                members = members[:index]
+
+        if normalize:
+            name = os.path.normpath(name)
+
+        for member in reversed(members):
+            if skipping:
+                if tarinfo.offset == member.offset:
+                    skipping = False
+                continue
+            if normalize:
+                member_name = os.path.normpath(member.name)
+            else:
+                member_name = member.name
+
+            if name == member_name:
+                return member
+
+        if skipping:
+            # Starting point was not found
+            raise ValueError(tarinfo)
+
+    def _load(self):
+        """Read through the entire archive file and look for readable
+           members. This should not run if the file is set to stream.
+        """
+        if not self.stream:
+            while self.next() is not None:
+                pass
+            self._loaded = True
+
+    def _check(self, mode=None):
+        """Check if TarFile is still open, and if the operation's mode
+           corresponds to TarFile's mode.
+        """
+        if self.closed:
+            raise OSError("%s is closed" % self.__class__.__name__)
+        if mode is not None and self.mode not in mode:
+            raise OSError("bad operation for mode %r" % self.mode)
+
+    def _find_link_target(self, tarinfo):
+        """Find the target member of a symlink or hardlink member in the
+           archive.
+        """
+        if tarinfo.issym():
+            # Always search the entire archive.
+            linkname = "/".join(filter(None, (os.path.dirname(tarinfo.name), tarinfo.linkname)))
+            limit = None
+        else:
+            # Search the archive before the link, because a hard link is
+            # just a reference to an already archived file.
+            linkname = tarinfo.linkname
+            limit = tarinfo
+
+        member = self._getmember(linkname, tarinfo=limit, normalize=True)
+        if member is None:
+            raise KeyError("linkname %r not found" % linkname)
+        return member
+
+    def __iter__(self):
+        """Provide an iterator object.
+        """
+        if self._loaded:
+            yield from self.members
+            return
+
+        # Yield items using TarFile's next() method.
+        # When all members have been read, set TarFile as _loaded.
+        index = 0
+        # Fix for SF #1100429: Under rare circumstances it can
+        # happen that getmembers() is called during iteration,
+        # which will have already exhausted the next() method.
+        if self.firstmember is not None:
+            tarinfo = self.next()
+            index += 1
+            yield tarinfo
+
+        while True:
+            if index < len(self.members):
+                tarinfo = self.members[index]
+            elif not self._loaded:
+                tarinfo = self.next()
+                if not tarinfo:
+                    self._loaded = True
+                    return
+            else:
+                return
+            index += 1
+            yield tarinfo
+
+    def _dbg(self, level, msg):
+        """Write debugging output to sys.stderr.
+        """
+        if level <= self.debug:
+            print(msg, file=sys.stderr)
+
+    def __enter__(self):
+        self._check()
+        return self
+
+    def __exit__(self, type, value, traceback):
+        if type is None:
+            self.close()
+        else:
+            # An exception occurred. We must not call close() because
+            # it would try to write end-of-archive blocks and padding.
+            if not self._extfileobj:
+                self.fileobj.close()
+            self.closed = True
+
+#--------------------
+# exported functions
+#--------------------
+
+def is_tarfile(name):
+    """Return True if name points to a tar archive that we
+       are able to handle, else return False.
+
+       'name' should be a string, file, or file-like object.
+    """
+    try:
+        if hasattr(name, "read"):
+            pos = name.tell()
+            t = open(fileobj=name)
+            name.seek(pos)
+        else:
+            t = open(name)
+        t.close()
+        return True
+    except TarError:
+        return False
+
+open = TarFile.open
+
+
+def main():
+    import argparse
+
+    description = 'A simple command-line interface for tarfile module.'
+    parser = argparse.ArgumentParser(description=description)
+    parser.add_argument('-v', '--verbose', action='store_true', default=False,
+                        help='Verbose output')
+    parser.add_argument('--filter', metavar='<filtername>',
+                        choices=_NAMED_FILTERS,
+                        help='Filter for extraction')
+
+    group = parser.add_mutually_exclusive_group(required=True)
+    group.add_argument('-l', '--list', metavar='<tarfile>',
+                       help='Show listing of a tarfile')
+    group.add_argument('-e', '--extract', nargs='+',
+                       metavar=('<tarfile>', '<output_dir>'),
+                       help='Extract tarfile into target dir')
+    group.add_argument('-c', '--create', nargs='+',
+                       metavar=('<name>', '<file>'),
+                       help='Create tarfile from sources')
+    group.add_argument('-t', '--test', metavar='<tarfile>',
+                       help='Test if a tarfile is valid')
+
+    args = parser.parse_args()
+
+    if args.filter and args.extract is None:
+        parser.exit(1, '--filter is only valid for extraction\n')
+
+    if args.test is not None:
+        src = args.test
+        if is_tarfile(src):
+            with open(src, 'r') as tar:
+                tar.getmembers()
+                print(tar.getmembers(), file=sys.stderr)
+            if args.verbose:
+                print('{!r} is a tar archive.'.format(src))
+        else:
+            parser.exit(1, '{!r} is not a tar archive.\n'.format(src))
+
+    elif args.list is not None:
+        src = args.list
+        if is_tarfile(src):
+            with TarFile.open(src, 'r:*') as tf:
+                tf.list(verbose=args.verbose)
+        else:
+            parser.exit(1, '{!r} is not a tar archive.\n'.format(src))
+
+    elif args.extract is not None:
+        if len(args.extract) == 1:
+            src = args.extract[0]
+            curdir = os.curdir
+        elif len(args.extract) == 2:
+            src, curdir = args.extract
+        else:
+            parser.exit(1, parser.format_help())
+
+        if is_tarfile(src):
+            with TarFile.open(src, 'r:*') as tf:
+                tf.extractall(path=curdir, filter=args.filter)
+            if args.verbose:
+                if curdir == '.':
+                    msg = '{!r} file is extracted.'.format(src)
+                else:
+                    msg = ('{!r} file is extracted '
+                           'into {!r} directory.').format(src, curdir)
+                print(msg)
+        else:
+            parser.exit(1, '{!r} is not a tar archive.\n'.format(src))
+
+    elif args.create is not None:
+        tar_name = args.create.pop(0)
+        _, ext = os.path.splitext(tar_name)
+        compressions = {
+            # gz
+            '.gz': 'gz',
+            '.tgz': 'gz',
+            # xz
+            '.xz': 'xz',
+            '.txz': 'xz',
+            # bz2
+            '.bz2': 'bz2',
+            '.tbz': 'bz2',
+            '.tbz2': 'bz2',
+            '.tb2': 'bz2',
+        }
+        tar_mode = 'w:' + compressions[ext] if ext in compressions else 'w'
+        tar_files = args.create
+
+        with TarFile.open(tar_name, tar_mode) as tf:
+            for file_name in tar_files:
+                tf.add(file_name)
+
+        if args.verbose:
+            print('{!r} file created.'.format(tar_name))
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__main__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__main__.py
new file mode 100644
index 00000000..daf55090
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__main__.py
@@ -0,0 +1,5 @@
+from . import main
+
+
+if __name__ == '__main__':
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2370ad82
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..a1407be1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9cfb2d98
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/py38.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/py38.cpython-312.pyc
new file mode 100644
index 00000000..903e94d9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/__pycache__/py38.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/py38.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/py38.py
new file mode 100644
index 00000000..20fbbfc1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/backports/tarfile/compat/py38.py
@@ -0,0 +1,24 @@
+import sys
+
+
+if sys.version_info < (3, 9):
+
+    def removesuffix(self, suffix):
+        # suffix='' should not call self[:-0].
+        if suffix and self.endswith(suffix):
+            return self[: -len(suffix)]
+        else:
+            return self[:]
+
+    def removeprefix(self, prefix):
+        if self.startswith(prefix):
+            return self[len(prefix) :]
+        else:
+            return self[:]
+else:
+
+    def removesuffix(self, suffix):
+        return self.removesuffix(suffix)
+
+    def removeprefix(self, prefix):
+        return self.removeprefix(prefix)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/LICENSE
new file mode 100644
index 00000000..d6456956
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/LICENSE
@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/METADATA
new file mode 100644
index 00000000..85513e8a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/METADATA
@@ -0,0 +1,129 @@
+Metadata-Version: 2.1
+Name: importlib_metadata
+Version: 8.0.0
+Summary: Read metadata from Python packages
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Source, https://github.com/python/importlib_metadata
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: zipp >=0.5
+Requires-Dist: typing-extensions >=3.6.4 ; python_version < "3.8"
+Provides-Extra: doc
+Requires-Dist: sphinx >=3.5 ; extra == 'doc'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'doc'
+Requires-Dist: rst.linker >=1.9 ; extra == 'doc'
+Requires-Dist: furo ; extra == 'doc'
+Requires-Dist: sphinx-lint ; extra == 'doc'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'doc'
+Provides-Extra: perf
+Requires-Dist: ipython ; extra == 'perf'
+Provides-Extra: test
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'test'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'test'
+Requires-Dist: pytest-cov ; extra == 'test'
+Requires-Dist: pytest-mypy ; extra == 'test'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'test'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'test'
+Requires-Dist: packaging ; extra == 'test'
+Requires-Dist: pyfakefs ; extra == 'test'
+Requires-Dist: flufl.flake8 ; extra == 'test'
+Requires-Dist: pytest-perf >=0.9.2 ; extra == 'test'
+Requires-Dist: jaraco.test >=5.4 ; extra == 'test'
+Requires-Dist: importlib-resources >=1.3 ; (python_version < "3.9") and extra == 'test'
+
+.. image:: https://img.shields.io/pypi/v/importlib_metadata.svg
+   :target: https://pypi.org/project/importlib_metadata
+
+.. image:: https://img.shields.io/pypi/pyversions/importlib_metadata.svg
+
+.. image:: https://github.com/python/importlib_metadata/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/python/importlib_metadata/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/importlib-metadata/badge/?version=latest
+   :target: https://importlib-metadata.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/importlib-metadata
+   :target: https://tidelift.com/subscription/pkg/pypi-importlib-metadata?utm_source=pypi-importlib-metadata&utm_medium=readme
+
+Library to access the metadata for a Python package.
+
+This package supplies third-party access to the functionality of
+`importlib.metadata <https://docs.python.org/3/library/importlib.metadata.html>`_
+including improvements added to subsequent Python versions.
+
+
+Compatibility
+=============
+
+New features are introduced in this third-party library and later merged
+into CPython. The following table indicates which versions of this library
+were contributed to different versions in the standard library:
+
+.. list-table::
+   :header-rows: 1
+
+   * - importlib_metadata
+     - stdlib
+   * - 7.0
+     - 3.13
+   * - 6.5
+     - 3.12
+   * - 4.13
+     - 3.11
+   * - 4.6
+     - 3.10
+   * - 1.4
+     - 3.8
+
+
+Usage
+=====
+
+See the `online documentation <https://importlib-metadata.readthedocs.io/>`_
+for usage details.
+
+`Finder authors
+<https://docs.python.org/3/reference/import.html#finders-and-loaders>`_ can
+also add support for custom package installers.  See the above documentation
+for details.
+
+
+Caveats
+=======
+
+This project primarily supports third-party packages installed by PyPA
+tools (or other conforming packages). It does not support:
+
+- Packages in the stdlib.
+- Packages installed without metadata.
+
+Project details
+===============
+
+ * Project home: https://github.com/python/importlib_metadata
+ * Report bugs at: https://github.com/python/importlib_metadata/issues
+ * Code hosting: https://github.com/python/importlib_metadata
+ * Documentation: https://importlib-metadata.readthedocs.io/
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-importlib-metadata?utm_source=pypi-importlib-metadata&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/RECORD
new file mode 100644
index 00000000..07b7dc51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/RECORD
@@ -0,0 +1,32 @@
+importlib_metadata-8.0.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+importlib_metadata-8.0.0.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+importlib_metadata-8.0.0.dist-info/METADATA,sha256=anuQ7_7h4J1bSEzfcjIBakPi2cyVQ7y7jklLHsBeH1k,4648
+importlib_metadata-8.0.0.dist-info/RECORD,,
+importlib_metadata-8.0.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+importlib_metadata-8.0.0.dist-info/WHEEL,sha256=mguMlWGMX-VHnMpKOjjQidIo1ssRlCFu4a4mBpz1s2M,91
+importlib_metadata-8.0.0.dist-info/top_level.txt,sha256=CO3fD9yylANiXkrMo4qHLV_mqXL2sC5JFKgt1yWAT-A,19
+importlib_metadata/__init__.py,sha256=tZNB-23h8Bixi9uCrQqj9Yf0aeC--Josdy3IZRIQeB0,33798
+importlib_metadata/__pycache__/__init__.cpython-312.pyc,,
+importlib_metadata/__pycache__/_adapters.cpython-312.pyc,,
+importlib_metadata/__pycache__/_collections.cpython-312.pyc,,
+importlib_metadata/__pycache__/_compat.cpython-312.pyc,,
+importlib_metadata/__pycache__/_functools.cpython-312.pyc,,
+importlib_metadata/__pycache__/_itertools.cpython-312.pyc,,
+importlib_metadata/__pycache__/_meta.cpython-312.pyc,,
+importlib_metadata/__pycache__/_text.cpython-312.pyc,,
+importlib_metadata/__pycache__/diagnose.cpython-312.pyc,,
+importlib_metadata/_adapters.py,sha256=rIhWTwBvYA1bV7i-5FfVX38qEXDTXFeS5cb5xJtP3ks,2317
+importlib_metadata/_collections.py,sha256=CJ0OTCHIjWA0ZIVS4voORAsn2R4R2cQBEtPsZEJpASY,743
+importlib_metadata/_compat.py,sha256=73QKrN9KNoaZzhbX5yPCCZa-FaALwXe8TPlDR72JgBU,1314
+importlib_metadata/_functools.py,sha256=PsY2-4rrKX4RVeRC1oGp1lB1pmC9eKN88_f-bD9uOoA,2895
+importlib_metadata/_itertools.py,sha256=cvr_2v8BRbxcIl5x5ldfqdHjhI8Yi8s8yk50G_nm6jQ,2068
+importlib_metadata/_meta.py,sha256=nxZ7C8GVlcBFAKWyVOn_dn7ot_twBcbm1NmvjIetBHI,1801
+importlib_metadata/_text.py,sha256=HCsFksZpJLeTP3NEk_ngrAeXVRRtTrtyh9eOABoRP4A,2166
+importlib_metadata/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+importlib_metadata/compat/__pycache__/__init__.cpython-312.pyc,,
+importlib_metadata/compat/__pycache__/py311.cpython-312.pyc,,
+importlib_metadata/compat/__pycache__/py39.cpython-312.pyc,,
+importlib_metadata/compat/py311.py,sha256=uqm-K-uohyj1042TH4a9Er_I5o7667DvulcD-gC_fSA,608
+importlib_metadata/compat/py39.py,sha256=cPkMv6-0ilK-0Jw_Tkn0xYbOKJZc4WJKQHow0c2T44w,1102
+importlib_metadata/diagnose.py,sha256=nkSRMiowlmkhLYhKhvCg9glmt_11Cox-EmLzEbqYTa8,379
+importlib_metadata/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/WHEEL
new file mode 100644
index 00000000..edf4ec7c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (70.1.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/top_level.txt
new file mode 100644
index 00000000..bbb07547
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata-8.0.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+importlib_metadata
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__init__.py
new file mode 100644
index 00000000..ed481355
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__init__.py
@@ -0,0 +1,1083 @@
+from __future__ import annotations
+
+import os
+import re
+import abc
+import sys
+import json
+import zipp
+import email
+import types
+import inspect
+import pathlib
+import operator
+import textwrap
+import functools
+import itertools
+import posixpath
+import collections
+
+from . import _meta
+from .compat import py39, py311
+from ._collections import FreezableDefaultDict, Pair
+from ._compat import (
+    NullFinder,
+    install,
+)
+from ._functools import method_cache, pass_none
+from ._itertools import always_iterable, unique_everseen
+from ._meta import PackageMetadata, SimplePath
+
+from contextlib import suppress
+from importlib import import_module
+from importlib.abc import MetaPathFinder
+from itertools import starmap
+from typing import Any, Iterable, List, Mapping, Match, Optional, Set, cast
+
+__all__ = [
+    'Distribution',
+    'DistributionFinder',
+    'PackageMetadata',
+    'PackageNotFoundError',
+    'distribution',
+    'distributions',
+    'entry_points',
+    'files',
+    'metadata',
+    'packages_distributions',
+    'requires',
+    'version',
+]
+
+
+class PackageNotFoundError(ModuleNotFoundError):
+    """The package was not found."""
+
+    def __str__(self) -> str:
+        return f"No package metadata was found for {self.name}"
+
+    @property
+    def name(self) -> str:  # type: ignore[override]
+        (name,) = self.args
+        return name
+
+
+class Sectioned:
+    """
+    A simple entry point config parser for performance
+
+    >>> for item in Sectioned.read(Sectioned._sample):
+    ...     print(item)
+    Pair(name='sec1', value='# comments ignored')
+    Pair(name='sec1', value='a = 1')
+    Pair(name='sec1', value='b = 2')
+    Pair(name='sec2', value='a = 2')
+
+    >>> res = Sectioned.section_pairs(Sectioned._sample)
+    >>> item = next(res)
+    >>> item.name
+    'sec1'
+    >>> item.value
+    Pair(name='a', value='1')
+    >>> item = next(res)
+    >>> item.value
+    Pair(name='b', value='2')
+    >>> item = next(res)
+    >>> item.name
+    'sec2'
+    >>> item.value
+    Pair(name='a', value='2')
+    >>> list(res)
+    []
+    """
+
+    _sample = textwrap.dedent(
+        """
+        [sec1]
+        # comments ignored
+        a = 1
+        b = 2
+
+        [sec2]
+        a = 2
+        """
+    ).lstrip()
+
+    @classmethod
+    def section_pairs(cls, text):
+        return (
+            section._replace(value=Pair.parse(section.value))
+            for section in cls.read(text, filter_=cls.valid)
+            if section.name is not None
+        )
+
+    @staticmethod
+    def read(text, filter_=None):
+        lines = filter(filter_, map(str.strip, text.splitlines()))
+        name = None
+        for value in lines:
+            section_match = value.startswith('[') and value.endswith(']')
+            if section_match:
+                name = value.strip('[]')
+                continue
+            yield Pair(name, value)
+
+    @staticmethod
+    def valid(line: str):
+        return line and not line.startswith('#')
+
+
+class EntryPoint:
+    """An entry point as defined by Python packaging conventions.
+
+    See `the packaging docs on entry points
+    <https://packaging.python.org/specifications/entry-points/>`_
+    for more information.
+
+    >>> ep = EntryPoint(
+    ...     name=None, group=None, value='package.module:attr [extra1, extra2]')
+    >>> ep.module
+    'package.module'
+    >>> ep.attr
+    'attr'
+    >>> ep.extras
+    ['extra1', 'extra2']
+    """
+
+    pattern = re.compile(
+        r'(?P<module>[\w.]+)\s*'
+        r'(:\s*(?P<attr>[\w.]+)\s*)?'
+        r'((?P<extras>\[.*\])\s*)?$'
+    )
+    """
+    A regular expression describing the syntax for an entry point,
+    which might look like:
+
+        - module
+        - package.module
+        - package.module:attribute
+        - package.module:object.attribute
+        - package.module:attr [extra1, extra2]
+
+    Other combinations are possible as well.
+
+    The expression is lenient about whitespace around the ':',
+    following the attr, and following any extras.
+    """
+
+    name: str
+    value: str
+    group: str
+
+    dist: Optional[Distribution] = None
+
+    def __init__(self, name: str, value: str, group: str) -> None:
+        vars(self).update(name=name, value=value, group=group)
+
+    def load(self) -> Any:
+        """Load the entry point from its definition. If only a module
+        is indicated by the value, return that module. Otherwise,
+        return the named object.
+        """
+        match = cast(Match, self.pattern.match(self.value))
+        module = import_module(match.group('module'))
+        attrs = filter(None, (match.group('attr') or '').split('.'))
+        return functools.reduce(getattr, attrs, module)
+
+    @property
+    def module(self) -> str:
+        match = self.pattern.match(self.value)
+        assert match is not None
+        return match.group('module')
+
+    @property
+    def attr(self) -> str:
+        match = self.pattern.match(self.value)
+        assert match is not None
+        return match.group('attr')
+
+    @property
+    def extras(self) -> List[str]:
+        match = self.pattern.match(self.value)
+        assert match is not None
+        return re.findall(r'\w+', match.group('extras') or '')
+
+    def _for(self, dist):
+        vars(self).update(dist=dist)
+        return self
+
+    def matches(self, **params):
+        """
+        EntryPoint matches the given parameters.
+
+        >>> ep = EntryPoint(group='foo', name='bar', value='bing:bong [extra1, extra2]')
+        >>> ep.matches(group='foo')
+        True
+        >>> ep.matches(name='bar', value='bing:bong [extra1, extra2]')
+        True
+        >>> ep.matches(group='foo', name='other')
+        False
+        >>> ep.matches()
+        True
+        >>> ep.matches(extras=['extra1', 'extra2'])
+        True
+        >>> ep.matches(module='bing')
+        True
+        >>> ep.matches(attr='bong')
+        True
+        """
+        attrs = (getattr(self, param) for param in params)
+        return all(map(operator.eq, params.values(), attrs))
+
+    def _key(self):
+        return self.name, self.value, self.group
+
+    def __lt__(self, other):
+        return self._key() < other._key()
+
+    def __eq__(self, other):
+        return self._key() == other._key()
+
+    def __setattr__(self, name, value):
+        raise AttributeError("EntryPoint objects are immutable.")
+
+    def __repr__(self):
+        return (
+            f'EntryPoint(name={self.name!r}, value={self.value!r}, '
+            f'group={self.group!r})'
+        )
+
+    def __hash__(self) -> int:
+        return hash(self._key())
+
+
+class EntryPoints(tuple):
+    """
+    An immutable collection of selectable EntryPoint objects.
+    """
+
+    __slots__ = ()
+
+    def __getitem__(self, name: str) -> EntryPoint:  # type: ignore[override]
+        """
+        Get the EntryPoint in self matching name.
+        """
+        try:
+            return next(iter(self.select(name=name)))
+        except StopIteration:
+            raise KeyError(name)
+
+    def __repr__(self):
+        """
+        Repr with classname and tuple constructor to
+        signal that we deviate from regular tuple behavior.
+        """
+        return '%s(%r)' % (self.__class__.__name__, tuple(self))
+
+    def select(self, **params) -> EntryPoints:
+        """
+        Select entry points from self that match the
+        given parameters (typically group and/or name).
+        """
+        return EntryPoints(ep for ep in self if py39.ep_matches(ep, **params))
+
+    @property
+    def names(self) -> Set[str]:
+        """
+        Return the set of all names of all entry points.
+        """
+        return {ep.name for ep in self}
+
+    @property
+    def groups(self) -> Set[str]:
+        """
+        Return the set of all groups of all entry points.
+        """
+        return {ep.group for ep in self}
+
+    @classmethod
+    def _from_text_for(cls, text, dist):
+        return cls(ep._for(dist) for ep in cls._from_text(text))
+
+    @staticmethod
+    def _from_text(text):
+        return (
+            EntryPoint(name=item.value.name, value=item.value.value, group=item.name)
+            for item in Sectioned.section_pairs(text or '')
+        )
+
+
+class PackagePath(pathlib.PurePosixPath):
+    """A reference to a path in a package"""
+
+    hash: Optional[FileHash]
+    size: int
+    dist: Distribution
+
+    def read_text(self, encoding: str = 'utf-8') -> str:  # type: ignore[override]
+        return self.locate().read_text(encoding=encoding)
+
+    def read_binary(self) -> bytes:
+        return self.locate().read_bytes()
+
+    def locate(self) -> SimplePath:
+        """Return a path-like object for this path"""
+        return self.dist.locate_file(self)
+
+
+class FileHash:
+    def __init__(self, spec: str) -> None:
+        self.mode, _, self.value = spec.partition('=')
+
+    def __repr__(self) -> str:
+        return f'<FileHash mode: {self.mode} value: {self.value}>'
+
+
+class Distribution(metaclass=abc.ABCMeta):
+    """
+    An abstract Python distribution package.
+
+    Custom providers may derive from this class and define
+    the abstract methods to provide a concrete implementation
+    for their environment. Some providers may opt to override
+    the default implementation of some properties to bypass
+    the file-reading mechanism.
+    """
+
+    @abc.abstractmethod
+    def read_text(self, filename) -> Optional[str]:
+        """Attempt to load metadata file given by the name.
+
+        Python distribution metadata is organized by blobs of text
+        typically represented as "files" in the metadata directory
+        (e.g. package-1.0.dist-info). These files include things
+        like:
+
+        - METADATA: The distribution metadata including fields
+          like Name and Version and Description.
+        - entry_points.txt: A series of entry points as defined in
+          `the entry points spec <https://packaging.python.org/en/latest/specifications/entry-points/#file-format>`_.
+        - RECORD: A record of files according to
+          `this recording spec <https://packaging.python.org/en/latest/specifications/recording-installed-packages/#the-record-file>`_.
+
+        A package may provide any set of files, including those
+        not listed here or none at all.
+
+        :param filename: The name of the file in the distribution info.
+        :return: The text if found, otherwise None.
+        """
+
+    @abc.abstractmethod
+    def locate_file(self, path: str | os.PathLike[str]) -> SimplePath:
+        """
+        Given a path to a file in this distribution, return a SimplePath
+        to it.
+        """
+
+    @classmethod
+    def from_name(cls, name: str) -> Distribution:
+        """Return the Distribution for the given package name.
+
+        :param name: The name of the distribution package to search for.
+        :return: The Distribution instance (or subclass thereof) for the named
+            package, if found.
+        :raises PackageNotFoundError: When the named package's distribution
+            metadata cannot be found.
+        :raises ValueError: When an invalid value is supplied for name.
+        """
+        if not name:
+            raise ValueError("A distribution name is required.")
+        try:
+            return next(iter(cls.discover(name=name)))
+        except StopIteration:
+            raise PackageNotFoundError(name)
+
+    @classmethod
+    def discover(
+        cls, *, context: Optional[DistributionFinder.Context] = None, **kwargs
+    ) -> Iterable[Distribution]:
+        """Return an iterable of Distribution objects for all packages.
+
+        Pass a ``context`` or pass keyword arguments for constructing
+        a context.
+
+        :context: A ``DistributionFinder.Context`` object.
+        :return: Iterable of Distribution objects for packages matching
+          the context.
+        """
+        if context and kwargs:
+            raise ValueError("cannot accept context and kwargs")
+        context = context or DistributionFinder.Context(**kwargs)
+        return itertools.chain.from_iterable(
+            resolver(context) for resolver in cls._discover_resolvers()
+        )
+
+    @staticmethod
+    def at(path: str | os.PathLike[str]) -> Distribution:
+        """Return a Distribution for the indicated metadata path.
+
+        :param path: a string or path-like object
+        :return: a concrete Distribution instance for the path
+        """
+        return PathDistribution(pathlib.Path(path))
+
+    @staticmethod
+    def _discover_resolvers():
+        """Search the meta_path for resolvers (MetadataPathFinders)."""
+        declared = (
+            getattr(finder, 'find_distributions', None) for finder in sys.meta_path
+        )
+        return filter(None, declared)
+
+    @property
+    def metadata(self) -> _meta.PackageMetadata:
+        """Return the parsed metadata for this Distribution.
+
+        The returned object will have keys that name the various bits of
+        metadata per the
+        `Core metadata specifications <https://packaging.python.org/en/latest/specifications/core-metadata/#core-metadata>`_.
+
+        Custom providers may provide the METADATA file or override this
+        property.
+        """
+        # deferred for performance (python/cpython#109829)
+        from . import _adapters
+
+        opt_text = (
+            self.read_text('METADATA')
+            or self.read_text('PKG-INFO')
+            # This last clause is here to support old egg-info files.  Its
+            # effect is to just end up using the PathDistribution's self._path
+            # (which points to the egg-info file) attribute unchanged.
+            or self.read_text('')
+        )
+        text = cast(str, opt_text)
+        return _adapters.Message(email.message_from_string(text))
+
+    @property
+    def name(self) -> str:
+        """Return the 'Name' metadata for the distribution package."""
+        return self.metadata['Name']
+
+    @property
+    def _normalized_name(self):
+        """Return a normalized version of the name."""
+        return Prepared.normalize(self.name)
+
+    @property
+    def version(self) -> str:
+        """Return the 'Version' metadata for the distribution package."""
+        return self.metadata['Version']
+
+    @property
+    def entry_points(self) -> EntryPoints:
+        """
+        Return EntryPoints for this distribution.
+
+        Custom providers may provide the ``entry_points.txt`` file
+        or override this property.
+        """
+        return EntryPoints._from_text_for(self.read_text('entry_points.txt'), self)
+
+    @property
+    def files(self) -> Optional[List[PackagePath]]:
+        """Files in this distribution.
+
+        :return: List of PackagePath for this distribution or None
+
+        Result is `None` if the metadata file that enumerates files
+        (i.e. RECORD for dist-info, or installed-files.txt or
+        SOURCES.txt for egg-info) is missing.
+        Result may be empty if the metadata exists but is empty.
+
+        Custom providers are recommended to provide a "RECORD" file (in
+        ``read_text``) or override this property to allow for callers to be
+        able to resolve filenames provided by the package.
+        """
+
+        def make_file(name, hash=None, size_str=None):
+            result = PackagePath(name)
+            result.hash = FileHash(hash) if hash else None
+            result.size = int(size_str) if size_str else None
+            result.dist = self
+            return result
+
+        @pass_none
+        def make_files(lines):
+            # Delay csv import, since Distribution.files is not as widely used
+            # as other parts of importlib.metadata
+            import csv
+
+            return starmap(make_file, csv.reader(lines))
+
+        @pass_none
+        def skip_missing_files(package_paths):
+            return list(filter(lambda path: path.locate().exists(), package_paths))
+
+        return skip_missing_files(
+            make_files(
+                self._read_files_distinfo()
+                or self._read_files_egginfo_installed()
+                or self._read_files_egginfo_sources()
+            )
+        )
+
+    def _read_files_distinfo(self):
+        """
+        Read the lines of RECORD.
+        """
+        text = self.read_text('RECORD')
+        return text and text.splitlines()
+
+    def _read_files_egginfo_installed(self):
+        """
+        Read installed-files.txt and return lines in a similar
+        CSV-parsable format as RECORD: each file must be placed
+        relative to the site-packages directory and must also be
+        quoted (since file names can contain literal commas).
+
+        This file is written when the package is installed by pip,
+        but it might not be written for other installation methods.
+        Assume the file is accurate if it exists.
+        """
+        text = self.read_text('installed-files.txt')
+        # Prepend the .egg-info/ subdir to the lines in this file.
+        # But this subdir is only available from PathDistribution's
+        # self._path.
+        subdir = getattr(self, '_path', None)
+        if not text or not subdir:
+            return
+
+        paths = (
+            py311.relative_fix((subdir / name).resolve())
+            .relative_to(self.locate_file('').resolve(), walk_up=True)
+            .as_posix()
+            for name in text.splitlines()
+        )
+        return map('"{}"'.format, paths)
+
+    def _read_files_egginfo_sources(self):
+        """
+        Read SOURCES.txt and return lines in a similar CSV-parsable
+        format as RECORD: each file name must be quoted (since it
+        might contain literal commas).
+
+        Note that SOURCES.txt is not a reliable source for what
+        files are installed by a package. This file is generated
+        for a source archive, and the files that are present
+        there (e.g. setup.py) may not correctly reflect the files
+        that are present after the package has been installed.
+        """
+        text = self.read_text('SOURCES.txt')
+        return text and map('"{}"'.format, text.splitlines())
+
+    @property
+    def requires(self) -> Optional[List[str]]:
+        """Generated requirements specified for this Distribution"""
+        reqs = self._read_dist_info_reqs() or self._read_egg_info_reqs()
+        return reqs and list(reqs)
+
+    def _read_dist_info_reqs(self):
+        return self.metadata.get_all('Requires-Dist')
+
+    def _read_egg_info_reqs(self):
+        source = self.read_text('requires.txt')
+        return pass_none(self._deps_from_requires_text)(source)
+
+    @classmethod
+    def _deps_from_requires_text(cls, source):
+        return cls._convert_egg_info_reqs_to_simple_reqs(Sectioned.read(source))
+
+    @staticmethod
+    def _convert_egg_info_reqs_to_simple_reqs(sections):
+        """
+        Historically, setuptools would solicit and store 'extra'
+        requirements, including those with environment markers,
+        in separate sections. More modern tools expect each
+        dependency to be defined separately, with any relevant
+        extras and environment markers attached directly to that
+        requirement. This method converts the former to the
+        latter. See _test_deps_from_requires_text for an example.
+        """
+
+        def make_condition(name):
+            return name and f'extra == "{name}"'
+
+        def quoted_marker(section):
+            section = section or ''
+            extra, sep, markers = section.partition(':')
+            if extra and markers:
+                markers = f'({markers})'
+            conditions = list(filter(None, [markers, make_condition(extra)]))
+            return '; ' + ' and '.join(conditions) if conditions else ''
+
+        def url_req_space(req):
+            """
+            PEP 508 requires a space between the url_spec and the quoted_marker.
+            Ref python/importlib_metadata#357.
+            """
+            # '@' is uniquely indicative of a url_req.
+            return ' ' * ('@' in req)
+
+        for section in sections:
+            space = url_req_space(section.value)
+            yield section.value + space + quoted_marker(section.name)
+
+    @property
+    def origin(self):
+        return self._load_json('direct_url.json')
+
+    def _load_json(self, filename):
+        return pass_none(json.loads)(
+            self.read_text(filename),
+            object_hook=lambda data: types.SimpleNamespace(**data),
+        )
+
+
+class DistributionFinder(MetaPathFinder):
+    """
+    A MetaPathFinder capable of discovering installed distributions.
+
+    Custom providers should implement this interface in order to
+    supply metadata.
+    """
+
+    class Context:
+        """
+        Keyword arguments presented by the caller to
+        ``distributions()`` or ``Distribution.discover()``
+        to narrow the scope of a search for distributions
+        in all DistributionFinders.
+
+        Each DistributionFinder may expect any parameters
+        and should attempt to honor the canonical
+        parameters defined below when appropriate.
+
+        This mechanism gives a custom provider a means to
+        solicit additional details from the caller beyond
+        "name" and "path" when searching distributions.
+        For example, imagine a provider that exposes suites
+        of packages in either a "public" or "private" ``realm``.
+        A caller may wish to query only for distributions in
+        a particular realm and could call
+        ``distributions(realm="private")`` to signal to the
+        custom provider to only include distributions from that
+        realm.
+        """
+
+        name = None
+        """
+        Specific name for which a distribution finder should match.
+        A name of ``None`` matches all distributions.
+        """
+
+        def __init__(self, **kwargs):
+            vars(self).update(kwargs)
+
+        @property
+        def path(self) -> List[str]:
+            """
+            The sequence of directory path that a distribution finder
+            should search.
+
+            Typically refers to Python installed package paths such as
+            "site-packages" directories and defaults to ``sys.path``.
+            """
+            return vars(self).get('path', sys.path)
+
+    @abc.abstractmethod
+    def find_distributions(self, context=Context()) -> Iterable[Distribution]:
+        """
+        Find distributions.
+
+        Return an iterable of all Distribution instances capable of
+        loading the metadata for packages matching the ``context``,
+        a DistributionFinder.Context instance.
+        """
+
+
+class FastPath:
+    """
+    Micro-optimized class for searching a root for children.
+
+    Root is a path on the file system that may contain metadata
+    directories either as natural directories or within a zip file.
+
+    >>> FastPath('').children()
+    ['...']
+
+    FastPath objects are cached and recycled for any given root.
+
+    >>> FastPath('foobar') is FastPath('foobar')
+    True
+    """
+
+    @functools.lru_cache()  # type: ignore
+    def __new__(cls, root):
+        return super().__new__(cls)
+
+    def __init__(self, root):
+        self.root = root
+
+    def joinpath(self, child):
+        return pathlib.Path(self.root, child)
+
+    def children(self):
+        with suppress(Exception):
+            return os.listdir(self.root or '.')
+        with suppress(Exception):
+            return self.zip_children()
+        return []
+
+    def zip_children(self):
+        zip_path = zipp.Path(self.root)
+        names = zip_path.root.namelist()
+        self.joinpath = zip_path.joinpath
+
+        return dict.fromkeys(child.split(posixpath.sep, 1)[0] for child in names)
+
+    def search(self, name):
+        return self.lookup(self.mtime).search(name)
+
+    @property
+    def mtime(self):
+        with suppress(OSError):
+            return os.stat(self.root).st_mtime
+        self.lookup.cache_clear()
+
+    @method_cache
+    def lookup(self, mtime):
+        return Lookup(self)
+
+
+class Lookup:
+    """
+    A micro-optimized class for searching a (fast) path for metadata.
+    """
+
+    def __init__(self, path: FastPath):
+        """
+        Calculate all of the children representing metadata.
+
+        From the children in the path, calculate early all of the
+        children that appear to represent metadata (infos) or legacy
+        metadata (eggs).
+        """
+
+        base = os.path.basename(path.root).lower()
+        base_is_egg = base.endswith(".egg")
+        self.infos = FreezableDefaultDict(list)
+        self.eggs = FreezableDefaultDict(list)
+
+        for child in path.children():
+            low = child.lower()
+            if low.endswith((".dist-info", ".egg-info")):
+                # rpartition is faster than splitext and suitable for this purpose.
+                name = low.rpartition(".")[0].partition("-")[0]
+                normalized = Prepared.normalize(name)
+                self.infos[normalized].append(path.joinpath(child))
+            elif base_is_egg and low == "egg-info":
+                name = base.rpartition(".")[0].partition("-")[0]
+                legacy_normalized = Prepared.legacy_normalize(name)
+                self.eggs[legacy_normalized].append(path.joinpath(child))
+
+        self.infos.freeze()
+        self.eggs.freeze()
+
+    def search(self, prepared: Prepared):
+        """
+        Yield all infos and eggs matching the Prepared query.
+        """
+        infos = (
+            self.infos[prepared.normalized]
+            if prepared
+            else itertools.chain.from_iterable(self.infos.values())
+        )
+        eggs = (
+            self.eggs[prepared.legacy_normalized]
+            if prepared
+            else itertools.chain.from_iterable(self.eggs.values())
+        )
+        return itertools.chain(infos, eggs)
+
+
+class Prepared:
+    """
+    A prepared search query for metadata on a possibly-named package.
+
+    Pre-calculates the normalization to prevent repeated operations.
+
+    >>> none = Prepared(None)
+    >>> none.normalized
+    >>> none.legacy_normalized
+    >>> bool(none)
+    False
+    >>> sample = Prepared('Sample__Pkg-name.foo')
+    >>> sample.normalized
+    'sample_pkg_name_foo'
+    >>> sample.legacy_normalized
+    'sample__pkg_name.foo'
+    >>> bool(sample)
+    True
+    """
+
+    normalized = None
+    legacy_normalized = None
+
+    def __init__(self, name: Optional[str]):
+        self.name = name
+        if name is None:
+            return
+        self.normalized = self.normalize(name)
+        self.legacy_normalized = self.legacy_normalize(name)
+
+    @staticmethod
+    def normalize(name):
+        """
+        PEP 503 normalization plus dashes as underscores.
+        """
+        return re.sub(r"[-_.]+", "-", name).lower().replace('-', '_')
+
+    @staticmethod
+    def legacy_normalize(name):
+        """
+        Normalize the package name as found in the convention in
+        older packaging tools versions and specs.
+        """
+        return name.lower().replace('-', '_')
+
+    def __bool__(self):
+        return bool(self.name)
+
+
+@install
+class MetadataPathFinder(NullFinder, DistributionFinder):
+    """A degenerate finder for distribution packages on the file system.
+
+    This finder supplies only a find_distributions() method for versions
+    of Python that do not have a PathFinder find_distributions().
+    """
+
+    @classmethod
+    def find_distributions(
+        cls, context=DistributionFinder.Context()
+    ) -> Iterable[PathDistribution]:
+        """
+        Find distributions.
+
+        Return an iterable of all Distribution instances capable of
+        loading the metadata for packages matching ``context.name``
+        (or all names if ``None`` indicated) along the paths in the list
+        of directories ``context.path``.
+        """
+        found = cls._search_paths(context.name, context.path)
+        return map(PathDistribution, found)
+
+    @classmethod
+    def _search_paths(cls, name, paths):
+        """Find metadata directories in paths heuristically."""
+        prepared = Prepared(name)
+        return itertools.chain.from_iterable(
+            path.search(prepared) for path in map(FastPath, paths)
+        )
+
+    @classmethod
+    def invalidate_caches(cls) -> None:
+        FastPath.__new__.cache_clear()
+
+
+class PathDistribution(Distribution):
+    def __init__(self, path: SimplePath) -> None:
+        """Construct a distribution.
+
+        :param path: SimplePath indicating the metadata directory.
+        """
+        self._path = path
+
+    def read_text(self, filename: str | os.PathLike[str]) -> Optional[str]:
+        with suppress(
+            FileNotFoundError,
+            IsADirectoryError,
+            KeyError,
+            NotADirectoryError,
+            PermissionError,
+        ):
+            return self._path.joinpath(filename).read_text(encoding='utf-8')
+
+        return None
+
+    read_text.__doc__ = Distribution.read_text.__doc__
+
+    def locate_file(self, path: str | os.PathLike[str]) -> SimplePath:
+        return self._path.parent / path
+
+    @property
+    def _normalized_name(self):
+        """
+        Performance optimization: where possible, resolve the
+        normalized name from the file system path.
+        """
+        stem = os.path.basename(str(self._path))
+        return (
+            pass_none(Prepared.normalize)(self._name_from_stem(stem))
+            or super()._normalized_name
+        )
+
+    @staticmethod
+    def _name_from_stem(stem):
+        """
+        >>> PathDistribution._name_from_stem('foo-3.0.egg-info')
+        'foo'
+        >>> PathDistribution._name_from_stem('CherryPy-3.0.dist-info')
+        'CherryPy'
+        >>> PathDistribution._name_from_stem('face.egg-info')
+        'face'
+        >>> PathDistribution._name_from_stem('foo.bar')
+        """
+        filename, ext = os.path.splitext(stem)
+        if ext not in ('.dist-info', '.egg-info'):
+            return
+        name, sep, rest = filename.partition('-')
+        return name
+
+
+def distribution(distribution_name: str) -> Distribution:
+    """Get the ``Distribution`` instance for the named package.
+
+    :param distribution_name: The name of the distribution package as a string.
+    :return: A ``Distribution`` instance (or subclass thereof).
+    """
+    return Distribution.from_name(distribution_name)
+
+
+def distributions(**kwargs) -> Iterable[Distribution]:
+    """Get all ``Distribution`` instances in the current environment.
+
+    :return: An iterable of ``Distribution`` instances.
+    """
+    return Distribution.discover(**kwargs)
+
+
+def metadata(distribution_name: str) -> _meta.PackageMetadata:
+    """Get the metadata for the named package.
+
+    :param distribution_name: The name of the distribution package to query.
+    :return: A PackageMetadata containing the parsed metadata.
+    """
+    return Distribution.from_name(distribution_name).metadata
+
+
+def version(distribution_name: str) -> str:
+    """Get the version string for the named package.
+
+    :param distribution_name: The name of the distribution package to query.
+    :return: The version string for the package as defined in the package's
+        "Version" metadata key.
+    """
+    return distribution(distribution_name).version
+
+
+_unique = functools.partial(
+    unique_everseen,
+    key=py39.normalized_name,
+)
+"""
+Wrapper for ``distributions`` to return unique distributions by name.
+"""
+
+
+def entry_points(**params) -> EntryPoints:
+    """Return EntryPoint objects for all installed packages.
+
+    Pass selection parameters (group or name) to filter the
+    result to entry points matching those properties (see
+    EntryPoints.select()).
+
+    :return: EntryPoints for all installed packages.
+    """
+    eps = itertools.chain.from_iterable(
+        dist.entry_points for dist in _unique(distributions())
+    )
+    return EntryPoints(eps).select(**params)
+
+
+def files(distribution_name: str) -> Optional[List[PackagePath]]:
+    """Return a list of files for the named package.
+
+    :param distribution_name: The name of the distribution package to query.
+    :return: List of files composing the distribution.
+    """
+    return distribution(distribution_name).files
+
+
+def requires(distribution_name: str) -> Optional[List[str]]:
+    """
+    Return a list of requirements for the named package.
+
+    :return: An iterable of requirements, suitable for
+        packaging.requirement.Requirement.
+    """
+    return distribution(distribution_name).requires
+
+
+def packages_distributions() -> Mapping[str, List[str]]:
+    """
+    Return a mapping of top-level packages to their
+    distributions.
+
+    >>> import collections.abc
+    >>> pkgs = packages_distributions()
+    >>> all(isinstance(dist, collections.abc.Sequence) for dist in pkgs.values())
+    True
+    """
+    pkg_to_dist = collections.defaultdict(list)
+    for dist in distributions():
+        for pkg in _top_level_declared(dist) or _top_level_inferred(dist):
+            pkg_to_dist[pkg].append(dist.metadata['Name'])
+    return dict(pkg_to_dist)
+
+
+def _top_level_declared(dist):
+    return (dist.read_text('top_level.txt') or '').split()
+
+
+def _topmost(name: PackagePath) -> Optional[str]:
+    """
+    Return the top-most parent as long as there is a parent.
+    """
+    top, *rest = name.parts
+    return top if rest else None
+
+
+def _get_toplevel_name(name: PackagePath) -> str:
+    """
+    Infer a possibly importable module name from a name presumed on
+    sys.path.
+
+    >>> _get_toplevel_name(PackagePath('foo.py'))
+    'foo'
+    >>> _get_toplevel_name(PackagePath('foo'))
+    'foo'
+    >>> _get_toplevel_name(PackagePath('foo.pyc'))
+    'foo'
+    >>> _get_toplevel_name(PackagePath('foo/__init__.py'))
+    'foo'
+    >>> _get_toplevel_name(PackagePath('foo.pth'))
+    'foo.pth'
+    >>> _get_toplevel_name(PackagePath('foo.dist-info'))
+    'foo.dist-info'
+    """
+    return _topmost(name) or (
+        # python/typeshed#10328
+        inspect.getmodulename(name)  # type: ignore
+        or str(name)
+    )
+
+
+def _top_level_inferred(dist):
+    opt_names = set(map(_get_toplevel_name, always_iterable(dist.files)))
+
+    def importable_name(name):
+        return '.' not in name
+
+    return filter(importable_name, opt_names)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0d859dcf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_adapters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_adapters.cpython-312.pyc
new file mode 100644
index 00000000..05b41f17
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_adapters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_collections.cpython-312.pyc
new file mode 100644
index 00000000..535e84e3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_compat.cpython-312.pyc
new file mode 100644
index 00000000..3e6c5719
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_functools.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_functools.cpython-312.pyc
new file mode 100644
index 00000000..0063c01e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_functools.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_itertools.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_itertools.cpython-312.pyc
new file mode 100644
index 00000000..43ecb90a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_itertools.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_meta.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_meta.cpython-312.pyc
new file mode 100644
index 00000000..7b3ea5e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_meta.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_text.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_text.cpython-312.pyc
new file mode 100644
index 00000000..5a35ef91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/_text.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/diagnose.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/diagnose.cpython-312.pyc
new file mode 100644
index 00000000..f12db331
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/__pycache__/diagnose.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_adapters.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_adapters.py
new file mode 100644
index 00000000..6223263e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_adapters.py
@@ -0,0 +1,83 @@
+import re
+import textwrap
+import email.message
+
+from ._text import FoldedCase
+
+
+class Message(email.message.Message):
+    multiple_use_keys = set(
+        map(
+            FoldedCase,
+            [
+                'Classifier',
+                'Obsoletes-Dist',
+                'Platform',
+                'Project-URL',
+                'Provides-Dist',
+                'Provides-Extra',
+                'Requires-Dist',
+                'Requires-External',
+                'Supported-Platform',
+                'Dynamic',
+            ],
+        )
+    )
+    """
+    Keys that may be indicated multiple times per PEP 566.
+    """
+
+    def __new__(cls, orig: email.message.Message):
+        res = super().__new__(cls)
+        vars(res).update(vars(orig))
+        return res
+
+    def __init__(self, *args, **kwargs):
+        self._headers = self._repair_headers()
+
+    # suppress spurious error from mypy
+    def __iter__(self):
+        return super().__iter__()
+
+    def __getitem__(self, item):
+        """
+        Override parent behavior to typical dict behavior.
+
+        ``email.message.Message`` will emit None values for missing
+        keys. Typical mappings, including this ``Message``, will raise
+        a key error for missing keys.
+
+        Ref python/importlib_metadata#371.
+        """
+        res = super().__getitem__(item)
+        if res is None:
+            raise KeyError(item)
+        return res
+
+    def _repair_headers(self):
+        def redent(value):
+            "Correct for RFC822 indentation"
+            if not value or '\n' not in value:
+                return value
+            return textwrap.dedent(' ' * 8 + value)
+
+        headers = [(key, redent(value)) for key, value in vars(self)['_headers']]
+        if self._payload:
+            headers.append(('Description', self.get_payload()))
+        return headers
+
+    @property
+    def json(self):
+        """
+        Convert PackageMetadata to a JSON-compatible format
+        per PEP 0566.
+        """
+
+        def transform(key):
+            value = self.get_all(key) if key in self.multiple_use_keys else self[key]
+            if key == 'Keywords':
+                value = re.split(r'\s+', value)
+            tk = key.lower().replace('-', '_')
+            return tk, value
+
+        return dict(map(transform, map(FoldedCase, self)))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_collections.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_collections.py
new file mode 100644
index 00000000..cf0954e1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_collections.py
@@ -0,0 +1,30 @@
+import collections
+
+
+# from jaraco.collections 3.3
+class FreezableDefaultDict(collections.defaultdict):
+    """
+    Often it is desirable to prevent the mutation of
+    a default dict after its initial construction, such
+    as to prevent mutation during iteration.
+
+    >>> dd = FreezableDefaultDict(list)
+    >>> dd[0].append('1')
+    >>> dd.freeze()
+    >>> dd[1]
+    []
+    >>> len(dd)
+    1
+    """
+
+    def __missing__(self, key):
+        return getattr(self, '_frozen', super().__missing__)(key)
+
+    def freeze(self):
+        self._frozen = lambda key: self.default_factory()
+
+
+class Pair(collections.namedtuple('Pair', 'name value')):
+    @classmethod
+    def parse(cls, text):
+        return cls(*map(str.strip, text.split("=", 1)))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_compat.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_compat.py
new file mode 100644
index 00000000..df312b1c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_compat.py
@@ -0,0 +1,57 @@
+import sys
+import platform
+
+
+__all__ = ['install', 'NullFinder']
+
+
+def install(cls):
+    """
+    Class decorator for installation on sys.meta_path.
+
+    Adds the backport DistributionFinder to sys.meta_path and
+    attempts to disable the finder functionality of the stdlib
+    DistributionFinder.
+    """
+    sys.meta_path.append(cls())
+    disable_stdlib_finder()
+    return cls
+
+
+def disable_stdlib_finder():
+    """
+    Give the backport primacy for discovering path-based distributions
+    by monkey-patching the stdlib O_O.
+
+    See #91 for more background for rationale on this sketchy
+    behavior.
+    """
+
+    def matches(finder):
+        return getattr(
+            finder, '__module__', None
+        ) == '_frozen_importlib_external' and hasattr(finder, 'find_distributions')
+
+    for finder in filter(matches, sys.meta_path):  # pragma: nocover
+        del finder.find_distributions
+
+
+class NullFinder:
+    """
+    A "Finder" (aka "MetaPathFinder") that never finds any modules,
+    but may find distributions.
+    """
+
+    @staticmethod
+    def find_spec(*args, **kwargs):
+        return None
+
+
+def pypy_partial(val):
+    """
+    Adjust for variable stacklevel on partial under PyPy.
+
+    Workaround for #327.
+    """
+    is_pypy = platform.python_implementation() == 'PyPy'
+    return val + is_pypy
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_functools.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_functools.py
new file mode 100644
index 00000000..71f66bd0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_functools.py
@@ -0,0 +1,104 @@
+import types
+import functools
+
+
+# from jaraco.functools 3.3
+def method_cache(method, cache_wrapper=None):
+    """
+    Wrap lru_cache to support storing the cache data in the object instances.
+
+    Abstracts the common paradigm where the method explicitly saves an
+    underscore-prefixed protected property on first call and returns that
+    subsequently.
+
+    >>> class MyClass:
+    ...     calls = 0
+    ...
+    ...     @method_cache
+    ...     def method(self, value):
+    ...         self.calls += 1
+    ...         return value
+
+    >>> a = MyClass()
+    >>> a.method(3)
+    3
+    >>> for x in range(75):
+    ...     res = a.method(x)
+    >>> a.calls
+    75
+
+    Note that the apparent behavior will be exactly like that of lru_cache
+    except that the cache is stored on each instance, so values in one
+    instance will not flush values from another, and when an instance is
+    deleted, so are the cached values for that instance.
+
+    >>> b = MyClass()
+    >>> for x in range(35):
+    ...     res = b.method(x)
+    >>> b.calls
+    35
+    >>> a.method(0)
+    0
+    >>> a.calls
+    75
+
+    Note that if method had been decorated with ``functools.lru_cache()``,
+    a.calls would have been 76 (due to the cached value of 0 having been
+    flushed by the 'b' instance).
+
+    Clear the cache with ``.cache_clear()``
+
+    >>> a.method.cache_clear()
+
+    Same for a method that hasn't yet been called.
+
+    >>> c = MyClass()
+    >>> c.method.cache_clear()
+
+    Another cache wrapper may be supplied:
+
+    >>> cache = functools.lru_cache(maxsize=2)
+    >>> MyClass.method2 = method_cache(lambda self: 3, cache_wrapper=cache)
+    >>> a = MyClass()
+    >>> a.method2()
+    3
+
+    Caution - do not subsequently wrap the method with another decorator, such
+    as ``@property``, which changes the semantics of the function.
+
+    See also
+    http://code.activestate.com/recipes/577452-a-memoize-decorator-for-instance-methods/
+    for another implementation and additional justification.
+    """
+    cache_wrapper = cache_wrapper or functools.lru_cache()
+
+    def wrapper(self, *args, **kwargs):
+        # it's the first call, replace the method with a cached, bound method
+        bound_method = types.MethodType(method, self)
+        cached_method = cache_wrapper(bound_method)
+        setattr(self, method.__name__, cached_method)
+        return cached_method(*args, **kwargs)
+
+    # Support cache clear even before cache has been created.
+    wrapper.cache_clear = lambda: None
+
+    return wrapper
+
+
+# From jaraco.functools 3.3
+def pass_none(func):
+    """
+    Wrap func so it's not called if its first param is None
+
+    >>> print_text = pass_none(print)
+    >>> print_text('text')
+    text
+    >>> print_text(None)
+    """
+
+    @functools.wraps(func)
+    def wrapper(param, *args, **kwargs):
+        if param is not None:
+            return func(param, *args, **kwargs)
+
+    return wrapper
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_itertools.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_itertools.py
new file mode 100644
index 00000000..d4ca9b91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_itertools.py
@@ -0,0 +1,73 @@
+from itertools import filterfalse
+
+
+def unique_everseen(iterable, key=None):
+    "List unique elements, preserving order. Remember all elements ever seen."
+    # unique_everseen('AAAABBBCCDAABBB') --> A B C D
+    # unique_everseen('ABBCcAD', str.lower) --> A B C D
+    seen = set()
+    seen_add = seen.add
+    if key is None:
+        for element in filterfalse(seen.__contains__, iterable):
+            seen_add(element)
+            yield element
+    else:
+        for element in iterable:
+            k = key(element)
+            if k not in seen:
+                seen_add(k)
+                yield element
+
+
+# copied from more_itertools 8.8
+def always_iterable(obj, base_type=(str, bytes)):
+    """If *obj* is iterable, return an iterator over its items::
+
+        >>> obj = (1, 2, 3)
+        >>> list(always_iterable(obj))
+        [1, 2, 3]
+
+    If *obj* is not iterable, return a one-item iterable containing *obj*::
+
+        >>> obj = 1
+        >>> list(always_iterable(obj))
+        [1]
+
+    If *obj* is ``None``, return an empty iterable:
+
+        >>> obj = None
+        >>> list(always_iterable(None))
+        []
+
+    By default, binary and text strings are not considered iterable::
+
+        >>> obj = 'foo'
+        >>> list(always_iterable(obj))
+        ['foo']
+
+    If *base_type* is set, objects for which ``isinstance(obj, base_type)``
+    returns ``True`` won't be considered iterable.
+
+        >>> obj = {'a': 1}
+        >>> list(always_iterable(obj))  # Iterate over the dict's keys
+        ['a']
+        >>> list(always_iterable(obj, base_type=dict))  # Treat dicts as a unit
+        [{'a': 1}]
+
+    Set *base_type* to ``None`` to avoid any special handling and treat objects
+    Python considers iterable as iterable:
+
+        >>> obj = 'foo'
+        >>> list(always_iterable(obj, base_type=None))
+        ['f', 'o', 'o']
+    """
+    if obj is None:
+        return iter(())
+
+    if (base_type is not None) and isinstance(obj, base_type):
+        return iter((obj,))
+
+    try:
+        return iter(obj)
+    except TypeError:
+        return iter((obj,))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_meta.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_meta.py
new file mode 100644
index 00000000..1927d0f6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_meta.py
@@ -0,0 +1,67 @@
+from __future__ import annotations
+
+import os
+from typing import Protocol
+from typing import Any, Dict, Iterator, List, Optional, TypeVar, Union, overload
+
+
+_T = TypeVar("_T")
+
+
+class PackageMetadata(Protocol):
+    def __len__(self) -> int: ...  # pragma: no cover
+
+    def __contains__(self, item: str) -> bool: ...  # pragma: no cover
+
+    def __getitem__(self, key: str) -> str: ...  # pragma: no cover
+
+    def __iter__(self) -> Iterator[str]: ...  # pragma: no cover
+
+    @overload
+    def get(
+        self, name: str, failobj: None = None
+    ) -> Optional[str]: ...  # pragma: no cover
+
+    @overload
+    def get(self, name: str, failobj: _T) -> Union[str, _T]: ...  # pragma: no cover
+
+    # overload per python/importlib_metadata#435
+    @overload
+    def get_all(
+        self, name: str, failobj: None = None
+    ) -> Optional[List[Any]]: ...  # pragma: no cover
+
+    @overload
+    def get_all(self, name: str, failobj: _T) -> Union[List[Any], _T]:
+        """
+        Return all values associated with a possibly multi-valued key.
+        """
+
+    @property
+    def json(self) -> Dict[str, Union[str, List[str]]]:
+        """
+        A JSON-compatible form of the metadata.
+        """
+
+
+class SimplePath(Protocol):
+    """
+    A minimal subset of pathlib.Path required by Distribution.
+    """
+
+    def joinpath(
+        self, other: Union[str, os.PathLike[str]]
+    ) -> SimplePath: ...  # pragma: no cover
+
+    def __truediv__(
+        self, other: Union[str, os.PathLike[str]]
+    ) -> SimplePath: ...  # pragma: no cover
+
+    @property
+    def parent(self) -> SimplePath: ...  # pragma: no cover
+
+    def read_text(self, encoding=None) -> str: ...  # pragma: no cover
+
+    def read_bytes(self) -> bytes: ...  # pragma: no cover
+
+    def exists(self) -> bool: ...  # pragma: no cover
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_text.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_text.py
new file mode 100644
index 00000000..c88cfbb2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/_text.py
@@ -0,0 +1,99 @@
+import re
+
+from ._functools import method_cache
+
+
+# from jaraco.text 3.5
+class FoldedCase(str):
+    """
+    A case insensitive string class; behaves just like str
+    except compares equal when the only variation is case.
+
+    >>> s = FoldedCase('hello world')
+
+    >>> s == 'Hello World'
+    True
+
+    >>> 'Hello World' == s
+    True
+
+    >>> s != 'Hello World'
+    False
+
+    >>> s.index('O')
+    4
+
+    >>> s.split('O')
+    ['hell', ' w', 'rld']
+
+    >>> sorted(map(FoldedCase, ['GAMMA', 'alpha', 'Beta']))
+    ['alpha', 'Beta', 'GAMMA']
+
+    Sequence membership is straightforward.
+
+    >>> "Hello World" in [s]
+    True
+    >>> s in ["Hello World"]
+    True
+
+    You may test for set inclusion, but candidate and elements
+    must both be folded.
+
+    >>> FoldedCase("Hello World") in {s}
+    True
+    >>> s in {FoldedCase("Hello World")}
+    True
+
+    String inclusion works as long as the FoldedCase object
+    is on the right.
+
+    >>> "hello" in FoldedCase("Hello World")
+    True
+
+    But not if the FoldedCase object is on the left:
+
+    >>> FoldedCase('hello') in 'Hello World'
+    False
+
+    In that case, use in_:
+
+    >>> FoldedCase('hello').in_('Hello World')
+    True
+
+    >>> FoldedCase('hello') > FoldedCase('Hello')
+    False
+    """
+
+    def __lt__(self, other):
+        return self.lower() < other.lower()
+
+    def __gt__(self, other):
+        return self.lower() > other.lower()
+
+    def __eq__(self, other):
+        return self.lower() == other.lower()
+
+    def __ne__(self, other):
+        return self.lower() != other.lower()
+
+    def __hash__(self):
+        return hash(self.lower())
+
+    def __contains__(self, other):
+        return super().lower().__contains__(other.lower())
+
+    def in_(self, other):
+        "Does self appear in other?"
+        return self in FoldedCase(other)
+
+    # cache lower since it's likely to be called frequently.
+    @method_cache
+    def lower(self):
+        return super().lower()
+
+    def index(self, sub):
+        return self.lower().index(sub.lower())
+
+    def split(self, splitter=' ', maxsplit=0):
+        pattern = re.compile(re.escape(splitter), re.I)
+        return pattern.split(self, maxsplit)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..11a81ba9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py311.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py311.cpython-312.pyc
new file mode 100644
index 00000000..3c978eca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py311.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py39.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py39.cpython-312.pyc
new file mode 100644
index 00000000..663cc56a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/__pycache__/py39.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py311.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py311.py
new file mode 100644
index 00000000..3a532743
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py311.py
@@ -0,0 +1,22 @@
+import os
+import pathlib
+import sys
+import types
+
+
+def wrap(path):  # pragma: no cover
+    """
+    Workaround for https://github.com/python/cpython/issues/84538
+    to add backward compatibility for walk_up=True.
+    An example affected package is dask-labextension, which uses
+    jupyter-packaging to install JupyterLab javascript files outside
+    of site-packages.
+    """
+
+    def relative_to(root, *, walk_up=False):
+        return pathlib.Path(os.path.relpath(path, root))
+
+    return types.SimpleNamespace(relative_to=relative_to)
+
+
+relative_fix = wrap if sys.version_info < (3, 12) else lambda x: x
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py39.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py39.py
new file mode 100644
index 00000000..1f15bd97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/compat/py39.py
@@ -0,0 +1,36 @@
+"""
+Compatibility layer with Python 3.8/3.9
+"""
+
+from typing import TYPE_CHECKING, Any, Optional
+
+if TYPE_CHECKING:  # pragma: no cover
+    # Prevent circular imports on runtime.
+    from .. import Distribution, EntryPoint
+else:
+    Distribution = EntryPoint = Any
+
+
+def normalized_name(dist: Distribution) -> Optional[str]:
+    """
+    Honor name normalization for distributions that don't provide ``_normalized_name``.
+    """
+    try:
+        return dist._normalized_name
+    except AttributeError:
+        from .. import Prepared  # -> delay to prevent circular imports.
+
+        return Prepared.normalize(getattr(dist, "name", None) or dist.metadata['Name'])
+
+
+def ep_matches(ep: EntryPoint, **params) -> bool:
+    """
+    Workaround for ``EntryPoint`` objects without the ``matches`` method.
+    """
+    try:
+        return ep.matches(**params)
+    except AttributeError:
+        from .. import EntryPoint  # -> delay to prevent circular imports.
+
+        # Reconstruct the EntryPoint object to make sure it is compatible.
+        return EntryPoint(ep.name, ep.value, ep.group).matches(**params)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/diagnose.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/diagnose.py
new file mode 100644
index 00000000..e405471a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/diagnose.py
@@ -0,0 +1,21 @@
+import sys
+
+from . import Distribution
+
+
+def inspect(path):
+    print("Inspecting", path)
+    dists = list(Distribution.discover(path=[path]))
+    if not dists:
+        return
+    print("Found", len(dists), "packages:", end=' ')
+    print(', '.join(dist.name for dist in dists))
+
+
+def run():
+    for path in sys.path:
+        inspect(path)
+
+
+if __name__ == '__main__':
+    run()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/importlib_metadata/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/METADATA
new file mode 100644
index 00000000..9a2097a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/METADATA
@@ -0,0 +1,591 @@
+Metadata-Version: 2.1
+Name: inflect
+Version: 7.3.1
+Summary: Correctly generate plurals, singular nouns, ordinals, indefinite articles
+Author-email: Paul Dyson <pwdyson@yahoo.com>
+Maintainer-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Source, https://github.com/jaraco/inflect
+Keywords: plural,inflect,participle
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Natural Language :: English
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Text Processing :: Linguistic
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: more-itertools >=8.5.0
+Requires-Dist: typeguard >=4.0.1
+Requires-Dist: typing-extensions ; python_version < "3.9"
+Provides-Extra: doc
+Requires-Dist: sphinx >=3.5 ; extra == 'doc'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'doc'
+Requires-Dist: rst.linker >=1.9 ; extra == 'doc'
+Requires-Dist: furo ; extra == 'doc'
+Requires-Dist: sphinx-lint ; extra == 'doc'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'doc'
+Provides-Extra: test
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'test'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'test'
+Requires-Dist: pytest-cov ; extra == 'test'
+Requires-Dist: pytest-mypy ; extra == 'test'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'test'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'test'
+Requires-Dist: pygments ; extra == 'test'
+
+.. image:: https://img.shields.io/pypi/v/inflect.svg
+   :target: https://pypi.org/project/inflect
+
+.. image:: https://img.shields.io/pypi/pyversions/inflect.svg
+
+.. image:: https://github.com/jaraco/inflect/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/inflect/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/inflect/badge/?version=latest
+   :target: https://inflect.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/inflect
+   :target: https://tidelift.com/subscription/pkg/pypi-inflect?utm_source=pypi-inflect&utm_medium=readme
+
+NAME
+====
+
+inflect.py - Correctly generate plurals, singular nouns, ordinals, indefinite articles; convert numbers to words.
+
+SYNOPSIS
+========
+
+.. code-block:: python
+
+    import inflect
+
+    p = inflect.engine()
+
+    # METHODS:
+
+    # plural plural_noun plural_verb plural_adj singular_noun no num
+    # compare compare_nouns compare_nouns compare_adjs
+    # a an
+    # present_participle
+    # ordinal number_to_words
+    # join
+    # inflect classical gender
+    # defnoun defverb defadj defa defan
+
+
+    # UNCONDITIONALLY FORM THE PLURAL
+
+    print("The plural of ", word, " is ", p.plural(word))
+
+
+    # CONDITIONALLY FORM THE PLURAL
+
+    print("I saw", cat_count, p.plural("cat", cat_count))
+
+
+    # FORM PLURALS FOR SPECIFIC PARTS OF SPEECH
+
+    print(
+        p.plural_noun("I", N1),
+        p.plural_verb("saw", N1),
+        p.plural_adj("my", N2),
+        p.plural_noun("saw", N2),
+    )
+
+
+    # FORM THE SINGULAR OF PLURAL NOUNS
+
+    print("The singular of ", word, " is ", p.singular_noun(word))
+
+    # SELECT THE GENDER OF SINGULAR PRONOUNS
+
+    print(p.singular_noun("they"))  # 'it'
+    p.gender("feminine")
+    print(p.singular_noun("they"))  # 'she'
+
+
+    # DEAL WITH "0/1/N" -> "no/1/N" TRANSLATION:
+
+    print("There ", p.plural_verb("was", errors), p.no(" error", errors))
+
+
+    # USE DEFAULT COUNTS:
+
+    print(
+        p.num(N1, ""),
+        p.plural("I"),
+        p.plural_verb(" saw"),
+        p.num(N2),
+        p.plural_noun(" saw"),
+    )
+    print("There ", p.num(errors, ""), p.plural_verb("was"), p.no(" error"))
+
+
+    # COMPARE TWO WORDS "NUMBER-INSENSITIVELY":
+
+    if p.compare(word1, word2):
+        print("same")
+    if p.compare_nouns(word1, word2):
+        print("same noun")
+    if p.compare_verbs(word1, word2):
+        print("same verb")
+    if p.compare_adjs(word1, word2):
+        print("same adj.")
+
+
+    # ADD CORRECT "a" OR "an" FOR A GIVEN WORD:
+
+    print("Did you want ", p.a(thing), " or ", p.an(idea))
+
+
+    # CONVERT NUMERALS INTO ORDINALS (i.e. 1->1st, 2->2nd, 3->3rd, etc.)
+
+    print("It was", p.ordinal(position), " from the left\n")
+
+    # CONVERT NUMERALS TO WORDS (i.e. 1->"one", 101->"one hundred and one", etc.)
+    # RETURNS A SINGLE STRING...
+
+    words = p.number_to_words(1234)
+    # "one thousand, two hundred and thirty-four"
+    words = p.number_to_words(p.ordinal(1234))
+    # "one thousand, two hundred and thirty-fourth"
+
+
+    # GET BACK A LIST OF STRINGS, ONE FOR EACH "CHUNK"...
+
+    words = p.number_to_words(1234, wantlist=True)
+    # ("one thousand","two hundred and thirty-four")
+
+
+    # OPTIONAL PARAMETERS CHANGE TRANSLATION:
+
+    words = p.number_to_words(12345, group=1)
+    # "one, two, three, four, five"
+
+    words = p.number_to_words(12345, group=2)
+    # "twelve, thirty-four, five"
+
+    words = p.number_to_words(12345, group=3)
+    # "one twenty-three, forty-five"
+
+    words = p.number_to_words(1234, andword="")
+    # "one thousand, two hundred thirty-four"
+
+    words = p.number_to_words(1234, andword=", plus")
+    # "one thousand, two hundred, plus thirty-four"
+    # TODO: I get no comma before plus: check perl
+
+    words = p.number_to_words(555_1202, group=1, zero="oh")
+    # "five, five, five, one, two, oh, two"
+
+    words = p.number_to_words(555_1202, group=1, one="unity")
+    # "five, five, five, unity, two, oh, two"
+
+    words = p.number_to_words(123.456, group=1, decimal="mark")
+    # "one two three mark four five six"
+    # TODO: DOCBUG: perl gives commas here as do I
+
+    # LITERAL STYLE ONLY NAMES NUMBERS LESS THAN A CERTAIN THRESHOLD...
+
+    words = p.number_to_words(9, threshold=10)  # "nine"
+    words = p.number_to_words(10, threshold=10)  # "ten"
+    words = p.number_to_words(11, threshold=10)  # "11"
+    words = p.number_to_words(1000, threshold=10)  # "1,000"
+
+    # JOIN WORDS INTO A LIST:
+
+    mylist = p.join(("apple", "banana", "carrot"))
+    # "apple, banana, and carrot"
+
+    mylist = p.join(("apple", "banana"))
+    # "apple and banana"
+
+    mylist = p.join(("apple", "banana", "carrot"), final_sep="")
+    # "apple, banana and carrot"
+
+
+    # REQUIRE "CLASSICAL" PLURALS (EG: "focus"->"foci", "cherub"->"cherubim")
+
+    p.classical()  # USE ALL CLASSICAL PLURALS
+
+    p.classical(all=True)  # USE ALL CLASSICAL PLURALS
+    p.classical(all=False)  # SWITCH OFF CLASSICAL MODE
+
+    p.classical(zero=True)  #  "no error" INSTEAD OF "no errors"
+    p.classical(zero=False)  #  "no errors" INSTEAD OF "no error"
+
+    p.classical(herd=True)  #  "2 buffalo" INSTEAD OF "2 buffalos"
+    p.classical(herd=False)  #  "2 buffalos" INSTEAD OF "2 buffalo"
+
+    p.classical(persons=True)  # "2 chairpersons" INSTEAD OF "2 chairpeople"
+    p.classical(persons=False)  # "2 chairpeople" INSTEAD OF "2 chairpersons"
+
+    p.classical(ancient=True)  # "2 formulae" INSTEAD OF "2 formulas"
+    p.classical(ancient=False)  # "2 formulas" INSTEAD OF "2 formulae"
+
+
+    # INTERPOLATE "plural()", "plural_noun()", "plural_verb()", "plural_adj()", "singular_noun()",
+    # a()", "an()", "num()" AND "ordinal()" WITHIN STRINGS:
+
+    print(p.inflect("The plural of {0} is plural('{0}')".format(word)))
+    print(p.inflect("The singular of {0} is singular_noun('{0}')".format(word)))
+    print(p.inflect("I saw {0} plural('cat',{0})".format(cat_count)))
+    print(
+        p.inflect(
+            "plural('I',{0}) "
+            "plural_verb('saw',{0}) "
+            "plural('a',{1}) "
+            "plural_noun('saw',{1})".format(N1, N2)
+        )
+    )
+    print(
+        p.inflect(
+            "num({0}, False)plural('I') "
+            "plural_verb('saw') "
+            "num({1}, False)plural('a') "
+            "plural_noun('saw')".format(N1, N2)
+        )
+    )
+    print(p.inflect("I saw num({0}) plural('cat')\nnum()".format(cat_count)))
+    print(p.inflect("There plural_verb('was',{0}) no('error',{0})".format(errors)))
+    print(p.inflect("There num({0}, False)plural_verb('was') no('error')".format(errors)))
+    print(p.inflect("Did you want a('{0}') or an('{1}')".format(thing, idea)))
+    print(p.inflect("It was ordinal('{0}') from the left".format(position)))
+
+
+    # ADD USER-DEFINED INFLECTIONS (OVERRIDING INBUILT RULES):
+
+    p.defnoun("VAX", "VAXen")  # SINGULAR => PLURAL
+
+    p.defverb(
+        "will",  # 1ST PERSON SINGULAR
+        "shall",  # 1ST PERSON PLURAL
+        "will",  # 2ND PERSON SINGULAR
+        "will",  # 2ND PERSON PLURAL
+        "will",  # 3RD PERSON SINGULAR
+        "will",  # 3RD PERSON PLURAL
+    )
+
+    p.defadj("hir", "their")  # SINGULAR => PLURAL
+
+    p.defa("h")  # "AY HALWAYS SEZ 'HAITCH'!"
+
+    p.defan("horrendous.*")  # "AN HORRENDOUS AFFECTATION"
+
+
+DESCRIPTION
+===========
+
+The methods of the class ``engine`` in module ``inflect.py`` provide plural
+inflections, singular noun inflections, "a"/"an" selection for English words,
+and manipulation of numbers as words.
+
+Plural forms of all nouns, most verbs, and some adjectives are
+provided. Where appropriate, "classical" variants (for example: "brother" ->
+"brethren", "dogma" -> "dogmata", etc.) are also provided.
+
+Single forms of nouns are also provided. The gender of singular pronouns
+can be chosen (for example "they" -> "it" or "she" or "he" or "they").
+
+Pronunciation-based "a"/"an" selection is provided for all English
+words, and most initialisms.
+
+It is also possible to inflect numerals (1,2,3) to ordinals (1st, 2nd, 3rd)
+and to English words ("one", "two", "three").
+
+In generating these inflections, ``inflect.py`` follows the Oxford
+English Dictionary and the guidelines in Fowler's Modern English
+Usage, preferring the former where the two disagree.
+
+The module is built around standard British spelling, but is designed
+to cope with common American variants as well. Slang, jargon, and
+other English dialects are *not* explicitly catered for.
+
+Where two or more inflected forms exist for a single word (typically a
+"classical" form and a "modern" form), ``inflect.py`` prefers the
+more common form (typically the "modern" one), unless "classical"
+processing has been specified
+(see `MODERN VS CLASSICAL INFLECTIONS`).
+
+FORMING PLURALS AND SINGULARS
+=============================
+
+Inflecting Plurals and Singulars
+--------------------------------
+
+All of the ``plural...`` plural inflection methods take the word to be
+inflected as their first argument and return the corresponding inflection.
+Note that all such methods expect the *singular* form of the word. The
+results of passing a plural form are undefined (and unlikely to be correct).
+Similarly, the ``si...`` singular inflection method expects the *plural*
+form of the word.
+
+The ``plural...`` methods also take an optional second argument,
+which indicates the grammatical "number" of the word (or of another word
+with which the word being inflected must agree). If the "number" argument is
+supplied and is not ``1`` (or ``"one"`` or ``"a"``, or some other adjective that
+implies the singular), the plural form of the word is returned. If the
+"number" argument *does* indicate singularity, the (uninflected) word
+itself is returned. If the number argument is omitted, the plural form
+is returned unconditionally.
+
+The ``si...`` method takes a second argument in a similar fashion. If it is
+some form of the number ``1``, or is omitted, the singular form is returned.
+Otherwise the plural is returned unaltered.
+
+
+The various methods of ``inflect.engine`` are:
+
+
+
+``plural_noun(word, count=None)``
+
+ The method ``plural_noun()`` takes a *singular* English noun or
+ pronoun and returns its plural. Pronouns in the nominative ("I" ->
+ "we") and accusative ("me" -> "us") cases are handled, as are
+ possessive pronouns ("mine" -> "ours").
+
+
+``plural_verb(word, count=None)``
+
+ The method ``plural_verb()`` takes the *singular* form of a
+ conjugated verb (that is, one which is already in the correct "person"
+ and "mood") and returns the corresponding plural conjugation.
+
+
+``plural_adj(word, count=None)``
+
+ The method ``plural_adj()`` takes the *singular* form of
+ certain types of adjectives and returns the corresponding plural form.
+ Adjectives that are correctly handled include: "numerical" adjectives
+ ("a" -> "some"), demonstrative adjectives ("this" -> "these", "that" ->
+ "those"), and possessives ("my" -> "our", "cat's" -> "cats'", "child's"
+ -> "childrens'", etc.)
+
+
+``plural(word, count=None)``
+
+ The method ``plural()`` takes a *singular* English noun,
+ pronoun, verb, or adjective and returns its plural form. Where a word
+ has more than one inflection depending on its part of speech (for
+ example, the noun "thought" inflects to "thoughts", the verb "thought"
+ to "thought"), the (singular) noun sense is preferred to the (singular)
+ verb sense.
+
+ Hence ``plural("knife")`` will return "knives" ("knife" having been treated
+ as a singular noun), whereas ``plural("knifes")`` will return "knife"
+ ("knifes" having been treated as a 3rd person singular verb).
+
+ The inherent ambiguity of such cases suggests that,
+ where the part of speech is known, ``plural_noun``, ``plural_verb``, and
+ ``plural_adj`` should be used in preference to ``plural``.
+
+
+``singular_noun(word, count=None)``
+
+ The method ``singular_noun()`` takes a *plural* English noun or
+ pronoun and returns its singular. Pronouns in the nominative ("we" ->
+ "I") and accusative ("us" -> "me") cases are handled, as are
+ possessive pronouns ("ours" -> "mine"). When third person
+ singular pronouns are returned they take the neuter gender by default
+ ("they" -> "it"), not ("they"-> "she") nor ("they" -> "he"). This can be
+ changed with ``gender()``.
+
+Note that all these methods ignore any whitespace surrounding the
+word being inflected, but preserve that whitespace when the result is
+returned. For example, ``plural(" cat  ")`` returns " cats  ".
+
+
+``gender(genderletter)``
+
+ The third person plural pronoun takes the same form for the female, male and
+ neuter (e.g. "they"). The singular however, depends upon gender (e.g. "she",
+ "he", "it" and "they" -- "they" being the gender neutral form.) By default
+ ``singular_noun`` returns the neuter form, however, the gender can be selected with
+ the ``gender`` method. Pass the first letter of the gender to
+ ``gender`` to return the f(eminine), m(asculine), n(euter) or t(hey)
+ form of the singular. e.g.
+ gender('f') followed by singular_noun('themselves') returns 'herself'.
+
+Numbered plurals
+----------------
+
+The ``plural...`` methods return only the inflected word, not the count that
+was used to inflect it. Thus, in order to produce "I saw 3 ducks", it
+is necessary to use:
+
+.. code-block:: python
+
+    print("I saw", N, p.plural_noun(animal, N))
+
+Since the usual purpose of producing a plural is to make it agree with
+a preceding count, inflect.py provides a method
+(``no(word, count)``) which, given a word and a(n optional) count, returns the
+count followed by the correctly inflected word. Hence the previous
+example can be rewritten:
+
+.. code-block:: python
+
+    print("I saw ", p.no(animal, N))
+
+In addition, if the count is zero (or some other term which implies
+zero, such as ``"zero"``, ``"nil"``, etc.) the count is replaced by the
+word "no". Hence, if ``N`` had the value zero, the previous example
+would print (the somewhat more elegant)::
+
+    I saw no animals
+
+rather than::
+
+    I saw 0 animals
+
+Note that the name of the method is a pun: the method
+returns either a number (a *No.*) or a ``"no"``, in front of the
+inflected word.
+
+
+Reducing the number of counts required
+--------------------------------------
+
+In some contexts, the need to supply an explicit count to the various
+``plural...`` methods makes for tiresome repetition. For example:
+
+.. code-block:: python
+
+    print(
+        plural_adj("This", errors),
+        plural_noun(" error", errors),
+        plural_verb(" was", errors),
+        " fatal.",
+    )
+
+inflect.py therefore provides a method
+(``num(count=None, show=None)``) which may be used to set a persistent "default number"
+value. If such a value is set, it is subsequently used whenever an
+optional second "number" argument is omitted. The default value thus set
+can subsequently be removed by calling ``num()`` with no arguments.
+Hence we could rewrite the previous example:
+
+.. code-block:: python
+
+    p.num(errors)
+    print(p.plural_adj("This"), p.plural_noun(" error"), p.plural_verb(" was"), "fatal.")
+    p.num()
+
+Normally, ``num()`` returns its first argument, so that it may also
+be "inlined" in contexts like:
+
+.. code-block:: python
+
+    print(p.num(errors), p.plural_noun(" error"), p.plural_verb(" was"), " detected.")
+    if severity > 1:
+        print(
+            p.plural_adj("This"), p.plural_noun(" error"), p.plural_verb(" was"), "fatal."
+        )
+
+However, in certain contexts (see `INTERPOLATING INFLECTIONS IN STRINGS`)
+it is preferable that ``num()`` return an empty string. Hence ``num()``
+provides an optional second argument. If that argument is supplied (that is, if
+it is defined) and evaluates to false, ``num`` returns an empty string
+instead of its first argument. For example:
+
+.. code-block:: python
+
+    print(p.num(errors, 0), p.no("error"), p.plural_verb(" was"), " detected.")
+    if severity > 1:
+        print(
+            p.plural_adj("This"), p.plural_noun(" error"), p.plural_verb(" was"), "fatal."
+        )
+
+
+
+Number-insensitive equality
+---------------------------
+
+inflect.py also provides a solution to the problem
+of comparing words of differing plurality through the methods
+``compare(word1, word2)``, ``compare_nouns(word1, word2)``,
+``compare_verbs(word1, word2)``, and ``compare_adjs(word1, word2)``.
+Each  of these methods takes two strings, and  compares them
+using the corresponding plural-inflection method (``plural()``, ``plural_noun()``,
+``plural_verb()``, and ``plural_adj()`` respectively).
+
+The comparison returns true if:
+
+- the strings are equal, or
+- one string is equal to a plural form of the other, or
+- the strings are two different plural forms of the one word.
+
+
+Hence all of the following return true:
+
+.. code-block:: python
+
+    p.compare("index", "index")  # RETURNS "eq"
+    p.compare("index", "indexes")  # RETURNS "s:p"
+    p.compare("index", "indices")  # RETURNS "s:p"
+    p.compare("indexes", "index")  # RETURNS "p:s"
+    p.compare("indices", "index")  # RETURNS "p:s"
+    p.compare("indices", "indexes")  # RETURNS "p:p"
+    p.compare("indexes", "indices")  # RETURNS "p:p"
+    p.compare("indices", "indices")  # RETURNS "eq"
+
+As indicated by the comments in the previous example, the actual value
+returned by the various ``compare`` methods encodes which of the
+three equality rules succeeded: "eq" is returned if the strings were
+identical, "s:p" if the strings were singular and plural respectively,
+"p:s" for plural and singular, and "p:p" for two distinct plurals.
+Inequality is indicated by returning an empty string.
+
+It should be noted that two distinct singular words which happen to take
+the same plural form are *not* considered equal, nor are cases where
+one (singular) word's plural is the other (plural) word's singular.
+Hence all of the following return false:
+
+.. code-block:: python
+
+    p.compare("base", "basis")  # ALTHOUGH BOTH -> "bases"
+    p.compare("syrinx", "syringe")  # ALTHOUGH BOTH -> "syringes"
+    p.compare("she", "he")  # ALTHOUGH BOTH -> "they"
+
+    p.compare("opus", "operas")  # ALTHOUGH "opus" -> "opera" -> "operas"
+    p.compare("taxi", "taxes")  # ALTHOUGH "taxi" -> "taxis" -> "taxes"
+
+Note too that, although the comparison is "number-insensitive" it is *not*
+case-insensitive (that is, ``plural("time","Times")`` returns false. To obtain
+both number and case insensitivity, use the ``lower()`` method on both strings
+(that is, ``plural("time".lower(), "Times".lower())`` returns true).
+
+Related Functionality
+=====================
+
+Shout out to these libraries that provide related functionality:
+
+* `WordSet <https://jaracotext.readthedocs.io/en/latest/#jaraco.text.WordSet>`_
+  parses identifiers like variable names into sets of words suitable for re-assembling
+  in another form.
+
+* `word2number <https://pypi.org/project/word2number/>`_ converts words to
+  a number.
+
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-PROJECT?utm_source=pypi-PROJECT&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/RECORD
new file mode 100644
index 00000000..73ff576b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/RECORD
@@ -0,0 +1,13 @@
+inflect-7.3.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+inflect-7.3.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+inflect-7.3.1.dist-info/METADATA,sha256=ZgMNY0WAZRs-U8wZiV2SMfjSKqBrMngXyDMs_CAwMwg,21079
+inflect-7.3.1.dist-info/RECORD,,
+inflect-7.3.1.dist-info/WHEEL,sha256=y4mX-SOX4fYIkonsAGA5N0Oy-8_gI4FXw5HNI1xqvWg,91
+inflect-7.3.1.dist-info/top_level.txt,sha256=m52ujdp10CqT6jh1XQxZT6kEntcnv-7Tl7UiGNTzWZA,8
+inflect/__init__.py,sha256=Jxy1HJXZiZ85kHeLAhkmvz6EMTdFqBe-duvt34R6IOc,103796
+inflect/__pycache__/__init__.cpython-312.pyc,,
+inflect/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+inflect/compat/__pycache__/__init__.cpython-312.pyc,,
+inflect/compat/__pycache__/py38.cpython-312.pyc,,
+inflect/compat/py38.py,sha256=oObVfVnWX9_OpnOuEJn1mFbJxVhwyR5epbiTNXDDaso,160
+inflect/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/WHEEL
new file mode 100644
index 00000000..564c6724
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (70.2.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/top_level.txt
new file mode 100644
index 00000000..0fd75fab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect-7.3.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+inflect
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__init__.py
new file mode 100644
index 00000000..3eec27f4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__init__.py
@@ -0,0 +1,3986 @@
+"""
+inflect: english language inflection
+ - correctly generate plurals, ordinals, indefinite articles
+ - convert numbers to words
+
+Copyright (C) 2010 Paul Dyson
+
+Based upon the Perl module
+`Lingua::EN::Inflect <https://metacpan.org/pod/Lingua::EN::Inflect>`_.
+
+methods:
+    classical inflect
+    plural plural_noun plural_verb plural_adj singular_noun no num a an
+    compare compare_nouns compare_verbs compare_adjs
+    present_participle
+    ordinal
+    number_to_words
+    join
+    defnoun defverb defadj defa defan
+
+INFLECTIONS:
+    classical inflect
+    plural plural_noun plural_verb plural_adj singular_noun compare
+    no num a an present_participle
+
+PLURALS:
+    classical inflect
+    plural plural_noun plural_verb plural_adj singular_noun no num
+    compare compare_nouns compare_verbs compare_adjs
+
+COMPARISONS:
+    classical
+    compare compare_nouns compare_verbs compare_adjs
+
+ARTICLES:
+    classical inflect num a an
+
+NUMERICAL:
+    ordinal number_to_words
+
+USER_DEFINED:
+    defnoun defverb defadj defa defan
+
+Exceptions:
+ UnknownClassicalModeError
+ BadNumValueError
+ BadChunkingOptionError
+ NumOutOfRangeError
+ BadUserDefinedPatternError
+ BadRcFileError
+ BadGenderError
+
+"""
+
+from __future__ import annotations
+
+import ast
+import collections
+import contextlib
+import functools
+import itertools
+import re
+from numbers import Number
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    Literal,
+    Match,
+    Optional,
+    Sequence,
+    Tuple,
+    Union,
+    cast,
+)
+
+from more_itertools import windowed_complete
+from typeguard import typechecked
+
+from .compat.py38 import Annotated
+
+
+class UnknownClassicalModeError(Exception):
+    pass
+
+
+class BadNumValueError(Exception):
+    pass
+
+
+class BadChunkingOptionError(Exception):
+    pass
+
+
+class NumOutOfRangeError(Exception):
+    pass
+
+
+class BadUserDefinedPatternError(Exception):
+    pass
+
+
+class BadRcFileError(Exception):
+    pass
+
+
+class BadGenderError(Exception):
+    pass
+
+
+def enclose(s: str) -> str:
+    return f"(?:{s})"
+
+
+def joinstem(cutpoint: Optional[int] = 0, words: Optional[Iterable[str]] = None) -> str:
+    """
+    Join stem of each word in words into a string for regex.
+
+    Each word is truncated at cutpoint.
+
+    Cutpoint is usually negative indicating the number of letters to remove
+    from the end of each word.
+
+    >>> joinstem(-2, ["ephemeris", "iris", ".*itis"])
+    '(?:ephemer|ir|.*it)'
+
+    >>> joinstem(None, ["ephemeris"])
+    '(?:ephemeris)'
+
+    >>> joinstem(5, None)
+    '(?:)'
+    """
+    return enclose("|".join(w[:cutpoint] for w in words or []))
+
+
+def bysize(words: Iterable[str]) -> Dict[int, set]:
+    """
+    From a list of words, return a dict of sets sorted by word length.
+
+    >>> words = ['ant', 'cat', 'dog', 'pig', 'frog', 'goat', 'horse', 'elephant']
+    >>> ret = bysize(words)
+    >>> sorted(ret[3])
+    ['ant', 'cat', 'dog', 'pig']
+    >>> ret[5]
+    {'horse'}
+    """
+    res: Dict[int, set] = collections.defaultdict(set)
+    for w in words:
+        res[len(w)].add(w)
+    return res
+
+
+def make_pl_si_lists(
+    lst: Iterable[str],
+    plending: str,
+    siendingsize: Optional[int],
+    dojoinstem: bool = True,
+):
+    """
+    given a list of singular words: lst
+
+    an ending to append to make the plural: plending
+
+    the number of characters to remove from the singular
+    before appending plending: siendingsize
+
+    a flag whether to create a joinstem: dojoinstem
+
+    return:
+    a list of pluralised words: si_list (called si because this is what you need to
+    look for to make the singular)
+
+    the pluralised words as a dict of sets sorted by word length: si_bysize
+    the singular words as a dict of sets sorted by word length: pl_bysize
+    if dojoinstem is True: a regular expression that matches any of the stems: stem
+    """
+    if siendingsize is not None:
+        siendingsize = -siendingsize
+    si_list = [w[:siendingsize] + plending for w in lst]
+    pl_bysize = bysize(lst)
+    si_bysize = bysize(si_list)
+    if dojoinstem:
+        stem = joinstem(siendingsize, lst)
+        return si_list, si_bysize, pl_bysize, stem
+    else:
+        return si_list, si_bysize, pl_bysize
+
+
+# 1. PLURALS
+
+pl_sb_irregular_s = {
+    "corpus": "corpuses|corpora",
+    "opus": "opuses|opera",
+    "genus": "genera",
+    "mythos": "mythoi",
+    "penis": "penises|penes",
+    "testis": "testes",
+    "atlas": "atlases|atlantes",
+    "yes": "yeses",
+}
+
+pl_sb_irregular = {
+    "child": "children",
+    "chili": "chilis|chilies",
+    "brother": "brothers|brethren",
+    "infinity": "infinities|infinity",
+    "loaf": "loaves",
+    "lore": "lores|lore",
+    "hoof": "hoofs|hooves",
+    "beef": "beefs|beeves",
+    "thief": "thiefs|thieves",
+    "money": "monies",
+    "mongoose": "mongooses",
+    "ox": "oxen",
+    "cow": "cows|kine",
+    "graffito": "graffiti",
+    "octopus": "octopuses|octopodes",
+    "genie": "genies|genii",
+    "ganglion": "ganglions|ganglia",
+    "trilby": "trilbys",
+    "turf": "turfs|turves",
+    "numen": "numina",
+    "atman": "atmas",
+    "occiput": "occiputs|occipita",
+    "sabretooth": "sabretooths",
+    "sabertooth": "sabertooths",
+    "lowlife": "lowlifes",
+    "flatfoot": "flatfoots",
+    "tenderfoot": "tenderfoots",
+    "romany": "romanies",
+    "jerry": "jerries",
+    "mary": "maries",
+    "talouse": "talouses",
+    "rom": "roma",
+    "carmen": "carmina",
+}
+
+pl_sb_irregular.update(pl_sb_irregular_s)
+# pl_sb_irregular_keys = enclose('|'.join(pl_sb_irregular.keys()))
+
+pl_sb_irregular_caps = {
+    "Romany": "Romanies",
+    "Jerry": "Jerrys",
+    "Mary": "Marys",
+    "Rom": "Roma",
+}
+
+pl_sb_irregular_compound = {"prima donna": "prima donnas|prime donne"}
+
+si_sb_irregular = {v: k for (k, v) in pl_sb_irregular.items()}
+for k in list(si_sb_irregular):
+    if "|" in k:
+        k1, k2 = k.split("|")
+        si_sb_irregular[k1] = si_sb_irregular[k2] = si_sb_irregular[k]
+        del si_sb_irregular[k]
+si_sb_irregular_caps = {v: k for (k, v) in pl_sb_irregular_caps.items()}
+si_sb_irregular_compound = {v: k for (k, v) in pl_sb_irregular_compound.items()}
+for k in list(si_sb_irregular_compound):
+    if "|" in k:
+        k1, k2 = k.split("|")
+        si_sb_irregular_compound[k1] = si_sb_irregular_compound[k2] = (
+            si_sb_irregular_compound[k]
+        )
+        del si_sb_irregular_compound[k]
+
+# si_sb_irregular_keys = enclose('|'.join(si_sb_irregular.keys()))
+
+# Z's that don't double
+
+pl_sb_z_zes_list = ("quartz", "topaz")
+pl_sb_z_zes_bysize = bysize(pl_sb_z_zes_list)
+
+pl_sb_ze_zes_list = ("snooze",)
+pl_sb_ze_zes_bysize = bysize(pl_sb_ze_zes_list)
+
+
+# CLASSICAL "..is" -> "..ides"
+
+pl_sb_C_is_ides_complete = [
+    # GENERAL WORDS...
+    "ephemeris",
+    "iris",
+    "clitoris",
+    "chrysalis",
+    "epididymis",
+]
+
+pl_sb_C_is_ides_endings = [
+    # INFLAMATIONS...
+    "itis"
+]
+
+pl_sb_C_is_ides = joinstem(
+    -2, pl_sb_C_is_ides_complete + [f".*{w}" for w in pl_sb_C_is_ides_endings]
+)
+
+pl_sb_C_is_ides_list = pl_sb_C_is_ides_complete + pl_sb_C_is_ides_endings
+
+(
+    si_sb_C_is_ides_list,
+    si_sb_C_is_ides_bysize,
+    pl_sb_C_is_ides_bysize,
+) = make_pl_si_lists(pl_sb_C_is_ides_list, "ides", 2, dojoinstem=False)
+
+
+# CLASSICAL "..a" -> "..ata"
+
+pl_sb_C_a_ata_list = (
+    "anathema",
+    "bema",
+    "carcinoma",
+    "charisma",
+    "diploma",
+    "dogma",
+    "drama",
+    "edema",
+    "enema",
+    "enigma",
+    "lemma",
+    "lymphoma",
+    "magma",
+    "melisma",
+    "miasma",
+    "oedema",
+    "sarcoma",
+    "schema",
+    "soma",
+    "stigma",
+    "stoma",
+    "trauma",
+    "gumma",
+    "pragma",
+)
+
+(
+    si_sb_C_a_ata_list,
+    si_sb_C_a_ata_bysize,
+    pl_sb_C_a_ata_bysize,
+    pl_sb_C_a_ata,
+) = make_pl_si_lists(pl_sb_C_a_ata_list, "ata", 1)
+
+# UNCONDITIONAL "..a" -> "..ae"
+
+pl_sb_U_a_ae_list = (
+    "alumna",
+    "alga",
+    "vertebra",
+    "persona",
+    "vita",
+)
+(
+    si_sb_U_a_ae_list,
+    si_sb_U_a_ae_bysize,
+    pl_sb_U_a_ae_bysize,
+    pl_sb_U_a_ae,
+) = make_pl_si_lists(pl_sb_U_a_ae_list, "e", None)
+
+# CLASSICAL "..a" -> "..ae"
+
+pl_sb_C_a_ae_list = (
+    "amoeba",
+    "antenna",
+    "formula",
+    "hyperbola",
+    "medusa",
+    "nebula",
+    "parabola",
+    "abscissa",
+    "hydra",
+    "nova",
+    "lacuna",
+    "aurora",
+    "umbra",
+    "flora",
+    "fauna",
+)
+(
+    si_sb_C_a_ae_list,
+    si_sb_C_a_ae_bysize,
+    pl_sb_C_a_ae_bysize,
+    pl_sb_C_a_ae,
+) = make_pl_si_lists(pl_sb_C_a_ae_list, "e", None)
+
+
+# CLASSICAL "..en" -> "..ina"
+
+pl_sb_C_en_ina_list = ("stamen", "foramen", "lumen")
+
+(
+    si_sb_C_en_ina_list,
+    si_sb_C_en_ina_bysize,
+    pl_sb_C_en_ina_bysize,
+    pl_sb_C_en_ina,
+) = make_pl_si_lists(pl_sb_C_en_ina_list, "ina", 2)
+
+
+# UNCONDITIONAL "..um" -> "..a"
+
+pl_sb_U_um_a_list = (
+    "bacterium",
+    "agendum",
+    "desideratum",
+    "erratum",
+    "stratum",
+    "datum",
+    "ovum",
+    "extremum",
+    "candelabrum",
+)
+(
+    si_sb_U_um_a_list,
+    si_sb_U_um_a_bysize,
+    pl_sb_U_um_a_bysize,
+    pl_sb_U_um_a,
+) = make_pl_si_lists(pl_sb_U_um_a_list, "a", 2)
+
+# CLASSICAL "..um" -> "..a"
+
+pl_sb_C_um_a_list = (
+    "maximum",
+    "minimum",
+    "momentum",
+    "optimum",
+    "quantum",
+    "cranium",
+    "curriculum",
+    "dictum",
+    "phylum",
+    "aquarium",
+    "compendium",
+    "emporium",
+    "encomium",
+    "gymnasium",
+    "honorarium",
+    "interregnum",
+    "lustrum",
+    "memorandum",
+    "millennium",
+    "rostrum",
+    "spectrum",
+    "speculum",
+    "stadium",
+    "trapezium",
+    "ultimatum",
+    "medium",
+    "vacuum",
+    "velum",
+    "consortium",
+    "arboretum",
+)
+
+(
+    si_sb_C_um_a_list,
+    si_sb_C_um_a_bysize,
+    pl_sb_C_um_a_bysize,
+    pl_sb_C_um_a,
+) = make_pl_si_lists(pl_sb_C_um_a_list, "a", 2)
+
+
+# UNCONDITIONAL "..us" -> "i"
+
+pl_sb_U_us_i_list = (
+    "alumnus",
+    "alveolus",
+    "bacillus",
+    "bronchus",
+    "locus",
+    "nucleus",
+    "stimulus",
+    "meniscus",
+    "sarcophagus",
+)
+(
+    si_sb_U_us_i_list,
+    si_sb_U_us_i_bysize,
+    pl_sb_U_us_i_bysize,
+    pl_sb_U_us_i,
+) = make_pl_si_lists(pl_sb_U_us_i_list, "i", 2)
+
+# CLASSICAL "..us" -> "..i"
+
+pl_sb_C_us_i_list = (
+    "focus",
+    "radius",
+    "genius",
+    "incubus",
+    "succubus",
+    "nimbus",
+    "fungus",
+    "nucleolus",
+    "stylus",
+    "torus",
+    "umbilicus",
+    "uterus",
+    "hippopotamus",
+    "cactus",
+)
+
+(
+    si_sb_C_us_i_list,
+    si_sb_C_us_i_bysize,
+    pl_sb_C_us_i_bysize,
+    pl_sb_C_us_i,
+) = make_pl_si_lists(pl_sb_C_us_i_list, "i", 2)
+
+
+# CLASSICAL "..us" -> "..us"  (ASSIMILATED 4TH DECLENSION LATIN NOUNS)
+
+pl_sb_C_us_us = (
+    "status",
+    "apparatus",
+    "prospectus",
+    "sinus",
+    "hiatus",
+    "impetus",
+    "plexus",
+)
+pl_sb_C_us_us_bysize = bysize(pl_sb_C_us_us)
+
+# UNCONDITIONAL "..on" -> "a"
+
+pl_sb_U_on_a_list = (
+    "criterion",
+    "perihelion",
+    "aphelion",
+    "phenomenon",
+    "prolegomenon",
+    "noumenon",
+    "organon",
+    "asyndeton",
+    "hyperbaton",
+)
+(
+    si_sb_U_on_a_list,
+    si_sb_U_on_a_bysize,
+    pl_sb_U_on_a_bysize,
+    pl_sb_U_on_a,
+) = make_pl_si_lists(pl_sb_U_on_a_list, "a", 2)
+
+# CLASSICAL "..on" -> "..a"
+
+pl_sb_C_on_a_list = ("oxymoron",)
+
+(
+    si_sb_C_on_a_list,
+    si_sb_C_on_a_bysize,
+    pl_sb_C_on_a_bysize,
+    pl_sb_C_on_a,
+) = make_pl_si_lists(pl_sb_C_on_a_list, "a", 2)
+
+
+# CLASSICAL "..o" -> "..i"  (BUT NORMALLY -> "..os")
+
+pl_sb_C_o_i = [
+    "solo",
+    "soprano",
+    "basso",
+    "alto",
+    "contralto",
+    "tempo",
+    "piano",
+    "virtuoso",
+]  # list not tuple so can concat for pl_sb_U_o_os
+
+pl_sb_C_o_i_bysize = bysize(pl_sb_C_o_i)
+si_sb_C_o_i_bysize = bysize([f"{w[:-1]}i" for w in pl_sb_C_o_i])
+
+pl_sb_C_o_i_stems = joinstem(-1, pl_sb_C_o_i)
+
+# ALWAYS "..o" -> "..os"
+
+pl_sb_U_o_os_complete = {"ado", "ISO", "NATO", "NCO", "NGO", "oto"}
+si_sb_U_o_os_complete = {f"{w}s" for w in pl_sb_U_o_os_complete}
+
+
+pl_sb_U_o_os_endings = [
+    "aficionado",
+    "aggro",
+    "albino",
+    "allegro",
+    "ammo",
+    "Antananarivo",
+    "archipelago",
+    "armadillo",
+    "auto",
+    "avocado",
+    "Bamako",
+    "Barquisimeto",
+    "bimbo",
+    "bingo",
+    "Biro",
+    "bolero",
+    "Bolzano",
+    "bongo",
+    "Boto",
+    "burro",
+    "Cairo",
+    "canto",
+    "cappuccino",
+    "casino",
+    "cello",
+    "Chicago",
+    "Chimango",
+    "cilantro",
+    "cochito",
+    "coco",
+    "Colombo",
+    "Colorado",
+    "commando",
+    "concertino",
+    "contango",
+    "credo",
+    "crescendo",
+    "cyano",
+    "demo",
+    "ditto",
+    "Draco",
+    "dynamo",
+    "embryo",
+    "Esperanto",
+    "espresso",
+    "euro",
+    "falsetto",
+    "Faro",
+    "fiasco",
+    "Filipino",
+    "flamenco",
+    "furioso",
+    "generalissimo",
+    "Gestapo",
+    "ghetto",
+    "gigolo",
+    "gizmo",
+    "Greensboro",
+    "gringo",
+    "Guaiabero",
+    "guano",
+    "gumbo",
+    "gyro",
+    "hairdo",
+    "hippo",
+    "Idaho",
+    "impetigo",
+    "inferno",
+    "info",
+    "intermezzo",
+    "intertrigo",
+    "Iquico",
+    "jumbo",
+    "junto",
+    "Kakapo",
+    "kilo",
+    "Kinkimavo",
+    "Kokako",
+    "Kosovo",
+    "Lesotho",
+    "libero",
+    "libido",
+    "libretto",
+    "lido",
+    "Lilo",
+    "limbo",
+    "limo",
+    "lineno",
+    "lingo",
+    "lino",
+    "livedo",
+    "loco",
+    "logo",
+    "lumbago",
+    "macho",
+    "macro",
+    "mafioso",
+    "magneto",
+    "magnifico",
+    "Majuro",
+    "Malabo",
+    "manifesto",
+    "Maputo",
+    "Maracaibo",
+    "medico",
+    "memo",
+    "metro",
+    "Mexico",
+    "micro",
+    "Milano",
+    "Monaco",
+    "mono",
+    "Montenegro",
+    "Morocco",
+    "Muqdisho",
+    "myo",
+    "neutrino",
+    "Ningbo",
+    "octavo",
+    "oregano",
+    "Orinoco",
+    "Orlando",
+    "Oslo",
+    "panto",
+    "Paramaribo",
+    "Pardusco",
+    "pedalo",
+    "photo",
+    "pimento",
+    "pinto",
+    "pleco",
+    "Pluto",
+    "pogo",
+    "polo",
+    "poncho",
+    "Porto-Novo",
+    "Porto",
+    "pro",
+    "psycho",
+    "pueblo",
+    "quarto",
+    "Quito",
+    "repo",
+    "rhino",
+    "risotto",
+    "rococo",
+    "rondo",
+    "Sacramento",
+    "saddo",
+    "sago",
+    "salvo",
+    "Santiago",
+    "Sapporo",
+    "Sarajevo",
+    "scherzando",
+    "scherzo",
+    "silo",
+    "sirocco",
+    "sombrero",
+    "staccato",
+    "sterno",
+    "stucco",
+    "stylo",
+    "sumo",
+    "Taiko",
+    "techno",
+    "terrazzo",
+    "testudo",
+    "timpano",
+    "tiro",
+    "tobacco",
+    "Togo",
+    "Tokyo",
+    "torero",
+    "Torino",
+    "Toronto",
+    "torso",
+    "tremolo",
+    "typo",
+    "tyro",
+    "ufo",
+    "UNESCO",
+    "vaquero",
+    "vermicello",
+    "verso",
+    "vibrato",
+    "violoncello",
+    "Virgo",
+    "weirdo",
+    "WHO",
+    "WTO",
+    "Yamoussoukro",
+    "yo-yo",
+    "zero",
+    "Zibo",
+] + pl_sb_C_o_i
+
+pl_sb_U_o_os_bysize = bysize(pl_sb_U_o_os_endings)
+si_sb_U_o_os_bysize = bysize([f"{w}s" for w in pl_sb_U_o_os_endings])
+
+
+# UNCONDITIONAL "..ch" -> "..chs"
+
+pl_sb_U_ch_chs_list = ("czech", "eunuch", "stomach")
+
+(
+    si_sb_U_ch_chs_list,
+    si_sb_U_ch_chs_bysize,
+    pl_sb_U_ch_chs_bysize,
+    pl_sb_U_ch_chs,
+) = make_pl_si_lists(pl_sb_U_ch_chs_list, "s", None)
+
+
+# UNCONDITIONAL "..[ei]x" -> "..ices"
+
+pl_sb_U_ex_ices_list = ("codex", "murex", "silex")
+(
+    si_sb_U_ex_ices_list,
+    si_sb_U_ex_ices_bysize,
+    pl_sb_U_ex_ices_bysize,
+    pl_sb_U_ex_ices,
+) = make_pl_si_lists(pl_sb_U_ex_ices_list, "ices", 2)
+
+pl_sb_U_ix_ices_list = ("radix", "helix")
+(
+    si_sb_U_ix_ices_list,
+    si_sb_U_ix_ices_bysize,
+    pl_sb_U_ix_ices_bysize,
+    pl_sb_U_ix_ices,
+) = make_pl_si_lists(pl_sb_U_ix_ices_list, "ices", 2)
+
+# CLASSICAL "..[ei]x" -> "..ices"
+
+pl_sb_C_ex_ices_list = (
+    "vortex",
+    "vertex",
+    "cortex",
+    "latex",
+    "pontifex",
+    "apex",
+    "index",
+    "simplex",
+)
+
+(
+    si_sb_C_ex_ices_list,
+    si_sb_C_ex_ices_bysize,
+    pl_sb_C_ex_ices_bysize,
+    pl_sb_C_ex_ices,
+) = make_pl_si_lists(pl_sb_C_ex_ices_list, "ices", 2)
+
+
+pl_sb_C_ix_ices_list = ("appendix",)
+
+(
+    si_sb_C_ix_ices_list,
+    si_sb_C_ix_ices_bysize,
+    pl_sb_C_ix_ices_bysize,
+    pl_sb_C_ix_ices,
+) = make_pl_si_lists(pl_sb_C_ix_ices_list, "ices", 2)
+
+
+# ARABIC: ".." -> "..i"
+
+pl_sb_C_i_list = ("afrit", "afreet", "efreet")
+
+(si_sb_C_i_list, si_sb_C_i_bysize, pl_sb_C_i_bysize, pl_sb_C_i) = make_pl_si_lists(
+    pl_sb_C_i_list, "i", None
+)
+
+
+# HEBREW: ".." -> "..im"
+
+pl_sb_C_im_list = ("goy", "seraph", "cherub")
+
+(si_sb_C_im_list, si_sb_C_im_bysize, pl_sb_C_im_bysize, pl_sb_C_im) = make_pl_si_lists(
+    pl_sb_C_im_list, "im", None
+)
+
+
+# UNCONDITIONAL "..man" -> "..mans"
+
+pl_sb_U_man_mans_list = """
+    ataman caiman cayman ceriman
+    desman dolman farman harman hetman
+    human leman ottoman shaman talisman
+""".split()
+pl_sb_U_man_mans_caps_list = """
+    Alabaman Bahaman Burman German
+    Hiroshiman Liman Nakayaman Norman Oklahoman
+    Panaman Roman Selman Sonaman Tacoman Yakiman
+    Yokohaman Yuman
+""".split()
+
+(
+    si_sb_U_man_mans_list,
+    si_sb_U_man_mans_bysize,
+    pl_sb_U_man_mans_bysize,
+) = make_pl_si_lists(pl_sb_U_man_mans_list, "s", None, dojoinstem=False)
+(
+    si_sb_U_man_mans_caps_list,
+    si_sb_U_man_mans_caps_bysize,
+    pl_sb_U_man_mans_caps_bysize,
+) = make_pl_si_lists(pl_sb_U_man_mans_caps_list, "s", None, dojoinstem=False)
+
+# UNCONDITIONAL "..louse" -> "..lice"
+pl_sb_U_louse_lice_list = ("booklouse", "grapelouse", "louse", "woodlouse")
+
+(
+    si_sb_U_louse_lice_list,
+    si_sb_U_louse_lice_bysize,
+    pl_sb_U_louse_lice_bysize,
+) = make_pl_si_lists(pl_sb_U_louse_lice_list, "lice", 5, dojoinstem=False)
+
+pl_sb_uninflected_s_complete = [
+    # PAIRS OR GROUPS SUBSUMED TO A SINGULAR...
+    "breeches",
+    "britches",
+    "pajamas",
+    "pyjamas",
+    "clippers",
+    "gallows",
+    "hijinks",
+    "headquarters",
+    "pliers",
+    "scissors",
+    "testes",
+    "herpes",
+    "pincers",
+    "shears",
+    "proceedings",
+    "trousers",
+    # UNASSIMILATED LATIN 4th DECLENSION
+    "cantus",
+    "coitus",
+    "nexus",
+    # RECENT IMPORTS...
+    "contretemps",
+    "corps",
+    "debris",
+    "siemens",
+    # DISEASES
+    "mumps",
+    # MISCELLANEOUS OTHERS...
+    "diabetes",
+    "jackanapes",
+    "series",
+    "species",
+    "subspecies",
+    "rabies",
+    "chassis",
+    "innings",
+    "news",
+    "mews",
+    "haggis",
+]
+
+pl_sb_uninflected_s_endings = [
+    # RECENT IMPORTS...
+    "ois",
+    # DISEASES
+    "measles",
+]
+
+pl_sb_uninflected_s = pl_sb_uninflected_s_complete + [
+    f".*{w}" for w in pl_sb_uninflected_s_endings
+]
+
+pl_sb_uninflected_herd = (
+    # DON'T INFLECT IN CLASSICAL MODE, OTHERWISE NORMAL INFLECTION
+    "wildebeest",
+    "swine",
+    "eland",
+    "bison",
+    "buffalo",
+    "cattle",
+    "elk",
+    "rhinoceros",
+    "zucchini",
+    "caribou",
+    "dace",
+    "grouse",
+    "guinea fowl",
+    "guinea-fowl",
+    "haddock",
+    "hake",
+    "halibut",
+    "herring",
+    "mackerel",
+    "pickerel",
+    "pike",
+    "roe",
+    "seed",
+    "shad",
+    "snipe",
+    "teal",
+    "turbot",
+    "water fowl",
+    "water-fowl",
+)
+
+pl_sb_uninflected_complete = [
+    # SOME FISH AND HERD ANIMALS
+    "tuna",
+    "salmon",
+    "mackerel",
+    "trout",
+    "bream",
+    "sea-bass",
+    "sea bass",
+    "carp",
+    "cod",
+    "flounder",
+    "whiting",
+    "moose",
+    # OTHER ODDITIES
+    "graffiti",
+    "djinn",
+    "samuri",
+    "offspring",
+    "pence",
+    "quid",
+    "hertz",
+] + pl_sb_uninflected_s_complete
+# SOME WORDS ENDING IN ...s (OFTEN PAIRS TAKEN AS A WHOLE)
+
+pl_sb_uninflected_caps = [
+    # ALL NATIONALS ENDING IN -ese
+    "Portuguese",
+    "Amoyese",
+    "Borghese",
+    "Congoese",
+    "Faroese",
+    "Foochowese",
+    "Genevese",
+    "Genoese",
+    "Gilbertese",
+    "Hottentotese",
+    "Kiplingese",
+    "Kongoese",
+    "Lucchese",
+    "Maltese",
+    "Nankingese",
+    "Niasese",
+    "Pekingese",
+    "Piedmontese",
+    "Pistoiese",
+    "Sarawakese",
+    "Shavese",
+    "Vermontese",
+    "Wenchowese",
+    "Yengeese",
+]
+
+
+pl_sb_uninflected_endings = [
+    # UNCOUNTABLE NOUNS
+    "butter",
+    "cash",
+    "furniture",
+    "information",
+    # SOME FISH AND HERD ANIMALS
+    "fish",
+    "deer",
+    "sheep",
+    # ALL NATIONALS ENDING IN -ese
+    "nese",
+    "rese",
+    "lese",
+    "mese",
+    # DISEASES
+    "pox",
+    # OTHER ODDITIES
+    "craft",
+] + pl_sb_uninflected_s_endings
+# SOME WORDS ENDING IN ...s (OFTEN PAIRS TAKEN AS A WHOLE)
+
+
+pl_sb_uninflected_bysize = bysize(pl_sb_uninflected_endings)
+
+
+# SINGULAR WORDS ENDING IN ...s (ALL INFLECT WITH ...es)
+
+pl_sb_singular_s_complete = [
+    "acropolis",
+    "aegis",
+    "alias",
+    "asbestos",
+    "bathos",
+    "bias",
+    "bronchitis",
+    "bursitis",
+    "caddis",
+    "cannabis",
+    "canvas",
+    "chaos",
+    "cosmos",
+    "dais",
+    "digitalis",
+    "epidermis",
+    "ethos",
+    "eyas",
+    "gas",
+    "glottis",
+    "hubris",
+    "ibis",
+    "lens",
+    "mantis",
+    "marquis",
+    "metropolis",
+    "pathos",
+    "pelvis",
+    "polis",
+    "rhinoceros",
+    "sassafras",
+    "trellis",
+] + pl_sb_C_is_ides_complete
+
+
+pl_sb_singular_s_endings = ["ss", "us"] + pl_sb_C_is_ides_endings
+
+pl_sb_singular_s_bysize = bysize(pl_sb_singular_s_endings)
+
+si_sb_singular_s_complete = [f"{w}es" for w in pl_sb_singular_s_complete]
+si_sb_singular_s_endings = [f"{w}es" for w in pl_sb_singular_s_endings]
+si_sb_singular_s_bysize = bysize(si_sb_singular_s_endings)
+
+pl_sb_singular_s_es = ["[A-Z].*es"]
+
+pl_sb_singular_s = enclose(
+    "|".join(
+        pl_sb_singular_s_complete
+        + [f".*{w}" for w in pl_sb_singular_s_endings]
+        + pl_sb_singular_s_es
+    )
+)
+
+
+# PLURALS ENDING IN uses -> use
+
+
+si_sb_ois_oi_case = ("Bolshois", "Hanois")
+
+si_sb_uses_use_case = ("Betelgeuses", "Duses", "Meuses", "Syracuses", "Toulouses")
+
+si_sb_uses_use = (
+    "abuses",
+    "applauses",
+    "blouses",
+    "carouses",
+    "causes",
+    "chartreuses",
+    "clauses",
+    "contuses",
+    "douses",
+    "excuses",
+    "fuses",
+    "grouses",
+    "hypotenuses",
+    "masseuses",
+    "menopauses",
+    "misuses",
+    "muses",
+    "overuses",
+    "pauses",
+    "peruses",
+    "profuses",
+    "recluses",
+    "reuses",
+    "ruses",
+    "souses",
+    "spouses",
+    "suffuses",
+    "transfuses",
+    "uses",
+)
+
+si_sb_ies_ie_case = (
+    "Addies",
+    "Aggies",
+    "Allies",
+    "Amies",
+    "Angies",
+    "Annies",
+    "Annmaries",
+    "Archies",
+    "Arties",
+    "Aussies",
+    "Barbies",
+    "Barries",
+    "Basies",
+    "Bennies",
+    "Bernies",
+    "Berties",
+    "Bessies",
+    "Betties",
+    "Billies",
+    "Blondies",
+    "Bobbies",
+    "Bonnies",
+    "Bowies",
+    "Brandies",
+    "Bries",
+    "Brownies",
+    "Callies",
+    "Carnegies",
+    "Carries",
+    "Cassies",
+    "Charlies",
+    "Cheries",
+    "Christies",
+    "Connies",
+    "Curies",
+    "Dannies",
+    "Debbies",
+    "Dixies",
+    "Dollies",
+    "Donnies",
+    "Drambuies",
+    "Eddies",
+    "Effies",
+    "Ellies",
+    "Elsies",
+    "Eries",
+    "Ernies",
+    "Essies",
+    "Eugenies",
+    "Fannies",
+    "Flossies",
+    "Frankies",
+    "Freddies",
+    "Gillespies",
+    "Goldies",
+    "Gracies",
+    "Guthries",
+    "Hallies",
+    "Hatties",
+    "Hetties",
+    "Hollies",
+    "Jackies",
+    "Jamies",
+    "Janies",
+    "Jannies",
+    "Jeanies",
+    "Jeannies",
+    "Jennies",
+    "Jessies",
+    "Jimmies",
+    "Jodies",
+    "Johnies",
+    "Johnnies",
+    "Josies",
+    "Julies",
+    "Kalgoorlies",
+    "Kathies",
+    "Katies",
+    "Kellies",
+    "Kewpies",
+    "Kristies",
+    "Laramies",
+    "Lassies",
+    "Lauries",
+    "Leslies",
+    "Lessies",
+    "Lillies",
+    "Lizzies",
+    "Lonnies",
+    "Lories",
+    "Lorries",
+    "Lotties",
+    "Louies",
+    "Mackenzies",
+    "Maggies",
+    "Maisies",
+    "Mamies",
+    "Marcies",
+    "Margies",
+    "Maries",
+    "Marjories",
+    "Matties",
+    "McKenzies",
+    "Melanies",
+    "Mickies",
+    "Millies",
+    "Minnies",
+    "Mollies",
+    "Mounties",
+    "Nannies",
+    "Natalies",
+    "Nellies",
+    "Netties",
+    "Ollies",
+    "Ozzies",
+    "Pearlies",
+    "Pottawatomies",
+    "Reggies",
+    "Richies",
+    "Rickies",
+    "Robbies",
+    "Ronnies",
+    "Rosalies",
+    "Rosemaries",
+    "Rosies",
+    "Roxies",
+    "Rushdies",
+    "Ruthies",
+    "Sadies",
+    "Sallies",
+    "Sammies",
+    "Scotties",
+    "Selassies",
+    "Sherries",
+    "Sophies",
+    "Stacies",
+    "Stefanies",
+    "Stephanies",
+    "Stevies",
+    "Susies",
+    "Sylvies",
+    "Tammies",
+    "Terries",
+    "Tessies",
+    "Tommies",
+    "Tracies",
+    "Trekkies",
+    "Valaries",
+    "Valeries",
+    "Valkyries",
+    "Vickies",
+    "Virgies",
+    "Willies",
+    "Winnies",
+    "Wylies",
+    "Yorkies",
+)
+
+si_sb_ies_ie = (
+    "aeries",
+    "baggies",
+    "belies",
+    "biggies",
+    "birdies",
+    "bogies",
+    "bonnies",
+    "boogies",
+    "bookies",
+    "bourgeoisies",
+    "brownies",
+    "budgies",
+    "caddies",
+    "calories",
+    "camaraderies",
+    "cockamamies",
+    "collies",
+    "cookies",
+    "coolies",
+    "cooties",
+    "coteries",
+    "crappies",
+    "curies",
+    "cutesies",
+    "dogies",
+    "eyries",
+    "floozies",
+    "footsies",
+    "freebies",
+    "genies",
+    "goalies",
+    "groupies",
+    "hies",
+    "jalousies",
+    "junkies",
+    "kiddies",
+    "laddies",
+    "lassies",
+    "lies",
+    "lingeries",
+    "magpies",
+    "menageries",
+    "mommies",
+    "movies",
+    "neckties",
+    "newbies",
+    "nighties",
+    "oldies",
+    "organdies",
+    "overlies",
+    "pies",
+    "pinkies",
+    "pixies",
+    "potpies",
+    "prairies",
+    "quickies",
+    "reveries",
+    "rookies",
+    "rotisseries",
+    "softies",
+    "sorties",
+    "species",
+    "stymies",
+    "sweeties",
+    "ties",
+    "underlies",
+    "unties",
+    "veggies",
+    "vies",
+    "yuppies",
+    "zombies",
+)
+
+
+si_sb_oes_oe_case = (
+    "Chloes",
+    "Crusoes",
+    "Defoes",
+    "Faeroes",
+    "Ivanhoes",
+    "Joes",
+    "McEnroes",
+    "Moes",
+    "Monroes",
+    "Noes",
+    "Poes",
+    "Roscoes",
+    "Tahoes",
+    "Tippecanoes",
+    "Zoes",
+)
+
+si_sb_oes_oe = (
+    "aloes",
+    "backhoes",
+    "canoes",
+    "does",
+    "floes",
+    "foes",
+    "hoes",
+    "mistletoes",
+    "oboes",
+    "pekoes",
+    "roes",
+    "sloes",
+    "throes",
+    "tiptoes",
+    "toes",
+    "woes",
+)
+
+si_sb_z_zes = ("quartzes", "topazes")
+
+si_sb_zzes_zz = ("buzzes", "fizzes", "frizzes", "razzes")
+
+si_sb_ches_che_case = (
+    "Andromaches",
+    "Apaches",
+    "Blanches",
+    "Comanches",
+    "Nietzsches",
+    "Porsches",
+    "Roches",
+)
+
+si_sb_ches_che = (
+    "aches",
+    "avalanches",
+    "backaches",
+    "bellyaches",
+    "caches",
+    "cloches",
+    "creches",
+    "douches",
+    "earaches",
+    "fiches",
+    "headaches",
+    "heartaches",
+    "microfiches",
+    "niches",
+    "pastiches",
+    "psyches",
+    "quiches",
+    "stomachaches",
+    "toothaches",
+    "tranches",
+)
+
+si_sb_xes_xe = ("annexes", "axes", "deluxes", "pickaxes")
+
+si_sb_sses_sse_case = ("Hesses", "Jesses", "Larousses", "Matisses")
+si_sb_sses_sse = (
+    "bouillabaisses",
+    "crevasses",
+    "demitasses",
+    "impasses",
+    "mousses",
+    "posses",
+)
+
+si_sb_ves_ve_case = (
+    # *[nwl]ives -> [nwl]live
+    "Clives",
+    "Palmolives",
+)
+si_sb_ves_ve = (
+    # *[^d]eaves -> eave
+    "interweaves",
+    "weaves",
+    # *[nwl]ives -> [nwl]live
+    "olives",
+    # *[eoa]lves -> [eoa]lve
+    "bivalves",
+    "dissolves",
+    "resolves",
+    "salves",
+    "twelves",
+    "valves",
+)
+
+
+plverb_special_s = enclose(
+    "|".join(
+        [pl_sb_singular_s]
+        + pl_sb_uninflected_s
+        + list(pl_sb_irregular_s)
+        + ["(.*[csx])is", "(.*)ceps", "[A-Z].*s"]
+    )
+)
+
+_pl_sb_postfix_adj_defn = (
+    ("general", enclose(r"(?!major|lieutenant|brigadier|adjutant|.*star)\S+")),
+    ("martial", enclose("court")),
+    ("force", enclose("pound")),
+)
+
+pl_sb_postfix_adj: Iterable[str] = (
+    enclose(val + f"(?=(?:-|\\s+){key})") for key, val in _pl_sb_postfix_adj_defn
+)
+
+pl_sb_postfix_adj_stems = f"({'|'.join(pl_sb_postfix_adj)})(.*)"
+
+
+# PLURAL WORDS ENDING IS es GO TO SINGULAR is
+
+si_sb_es_is = (
+    "amanuenses",
+    "amniocenteses",
+    "analyses",
+    "antitheses",
+    "apotheoses",
+    "arterioscleroses",
+    "atheroscleroses",
+    "axes",
+    # 'bases', # bases -> basis
+    "catalyses",
+    "catharses",
+    "chasses",
+    "cirrhoses",
+    "cocces",
+    "crises",
+    "diagnoses",
+    "dialyses",
+    "diereses",
+    "electrolyses",
+    "emphases",
+    "exegeses",
+    "geneses",
+    "halitoses",
+    "hydrolyses",
+    "hypnoses",
+    "hypotheses",
+    "hystereses",
+    "metamorphoses",
+    "metastases",
+    "misdiagnoses",
+    "mitoses",
+    "mononucleoses",
+    "narcoses",
+    "necroses",
+    "nemeses",
+    "neuroses",
+    "oases",
+    "osmoses",
+    "osteoporoses",
+    "paralyses",
+    "parentheses",
+    "parthenogeneses",
+    "periphrases",
+    "photosyntheses",
+    "probosces",
+    "prognoses",
+    "prophylaxes",
+    "prostheses",
+    "preces",
+    "psoriases",
+    "psychoanalyses",
+    "psychokineses",
+    "psychoses",
+    "scleroses",
+    "scolioses",
+    "sepses",
+    "silicoses",
+    "symbioses",
+    "synopses",
+    "syntheses",
+    "taxes",
+    "telekineses",
+    "theses",
+    "thromboses",
+    "tuberculoses",
+    "urinalyses",
+)
+
+pl_prep_list = """
+    about above across after among around at athwart before behind
+    below beneath beside besides between betwixt beyond but by
+    during except for from in into near of off on onto out over
+    since till to under until unto upon with""".split()
+
+pl_prep_list_da = pl_prep_list + ["de", "du", "da"]
+
+pl_prep_bysize = bysize(pl_prep_list_da)
+
+pl_prep = enclose("|".join(pl_prep_list_da))
+
+pl_sb_prep_dual_compound = rf"(.*?)((?:-|\s+)(?:{pl_prep})(?:-|\s+))a(?:-|\s+)(.*)"
+
+
+singular_pronoun_genders = {
+    "neuter",
+    "feminine",
+    "masculine",
+    "gender-neutral",
+    "feminine or masculine",
+    "masculine or feminine",
+}
+
+pl_pron_nom = {
+    # NOMINATIVE    REFLEXIVE
+    "i": "we",
+    "myself": "ourselves",
+    "you": "you",
+    "yourself": "yourselves",
+    "she": "they",
+    "herself": "themselves",
+    "he": "they",
+    "himself": "themselves",
+    "it": "they",
+    "itself": "themselves",
+    "they": "they",
+    "themself": "themselves",
+    #   POSSESSIVE
+    "mine": "ours",
+    "yours": "yours",
+    "hers": "theirs",
+    "his": "theirs",
+    "its": "theirs",
+    "theirs": "theirs",
+}
+
+si_pron: Dict[str, Dict[str, Union[str, Dict[str, str]]]] = {
+    "nom": {v: k for (k, v) in pl_pron_nom.items()}
+}
+si_pron["nom"]["we"] = "I"
+
+
+pl_pron_acc = {
+    # ACCUSATIVE    REFLEXIVE
+    "me": "us",
+    "myself": "ourselves",
+    "you": "you",
+    "yourself": "yourselves",
+    "her": "them",
+    "herself": "themselves",
+    "him": "them",
+    "himself": "themselves",
+    "it": "them",
+    "itself": "themselves",
+    "them": "them",
+    "themself": "themselves",
+}
+
+pl_pron_acc_keys = enclose("|".join(pl_pron_acc))
+pl_pron_acc_keys_bysize = bysize(pl_pron_acc)
+
+si_pron["acc"] = {v: k for (k, v) in pl_pron_acc.items()}
+
+for _thecase, _plur, _gend, _sing in (
+    ("nom", "they", "neuter", "it"),
+    ("nom", "they", "feminine", "she"),
+    ("nom", "they", "masculine", "he"),
+    ("nom", "they", "gender-neutral", "they"),
+    ("nom", "they", "feminine or masculine", "she or he"),
+    ("nom", "they", "masculine or feminine", "he or she"),
+    ("nom", "themselves", "neuter", "itself"),
+    ("nom", "themselves", "feminine", "herself"),
+    ("nom", "themselves", "masculine", "himself"),
+    ("nom", "themselves", "gender-neutral", "themself"),
+    ("nom", "themselves", "feminine or masculine", "herself or himself"),
+    ("nom", "themselves", "masculine or feminine", "himself or herself"),
+    ("nom", "theirs", "neuter", "its"),
+    ("nom", "theirs", "feminine", "hers"),
+    ("nom", "theirs", "masculine", "his"),
+    ("nom", "theirs", "gender-neutral", "theirs"),
+    ("nom", "theirs", "feminine or masculine", "hers or his"),
+    ("nom", "theirs", "masculine or feminine", "his or hers"),
+    ("acc", "them", "neuter", "it"),
+    ("acc", "them", "feminine", "her"),
+    ("acc", "them", "masculine", "him"),
+    ("acc", "them", "gender-neutral", "them"),
+    ("acc", "them", "feminine or masculine", "her or him"),
+    ("acc", "them", "masculine or feminine", "him or her"),
+    ("acc", "themselves", "neuter", "itself"),
+    ("acc", "themselves", "feminine", "herself"),
+    ("acc", "themselves", "masculine", "himself"),
+    ("acc", "themselves", "gender-neutral", "themself"),
+    ("acc", "themselves", "feminine or masculine", "herself or himself"),
+    ("acc", "themselves", "masculine or feminine", "himself or herself"),
+):
+    try:
+        si_pron[_thecase][_plur][_gend] = _sing  # type: ignore
+    except TypeError:
+        si_pron[_thecase][_plur] = {}
+        si_pron[_thecase][_plur][_gend] = _sing  # type: ignore
+
+
+si_pron_acc_keys = enclose("|".join(si_pron["acc"]))
+si_pron_acc_keys_bysize = bysize(si_pron["acc"])
+
+
+def get_si_pron(thecase, word, gender) -> str:
+    try:
+        sing = si_pron[thecase][word]
+    except KeyError:
+        raise  # not a pronoun
+    try:
+        return sing[gender]  # has several types due to gender
+    except TypeError:
+        return cast(str, sing)  # answer independent of gender
+
+
+# These dictionaries group verbs by first, second and third person
+# conjugations.
+
+plverb_irregular_pres = {
+    "am": "are",
+    "are": "are",
+    "is": "are",
+    "was": "were",
+    "were": "were",
+    "have": "have",
+    "has": "have",
+    "do": "do",
+    "does": "do",
+}
+
+plverb_ambiguous_pres = {
+    "act": "act",
+    "acts": "act",
+    "blame": "blame",
+    "blames": "blame",
+    "can": "can",
+    "must": "must",
+    "fly": "fly",
+    "flies": "fly",
+    "copy": "copy",
+    "copies": "copy",
+    "drink": "drink",
+    "drinks": "drink",
+    "fight": "fight",
+    "fights": "fight",
+    "fire": "fire",
+    "fires": "fire",
+    "like": "like",
+    "likes": "like",
+    "look": "look",
+    "looks": "look",
+    "make": "make",
+    "makes": "make",
+    "reach": "reach",
+    "reaches": "reach",
+    "run": "run",
+    "runs": "run",
+    "sink": "sink",
+    "sinks": "sink",
+    "sleep": "sleep",
+    "sleeps": "sleep",
+    "view": "view",
+    "views": "view",
+}
+
+plverb_ambiguous_pres_keys = re.compile(
+    rf"^({enclose('|'.join(plverb_ambiguous_pres))})((\s.*)?)$", re.IGNORECASE
+)
+
+
+plverb_irregular_non_pres = (
+    "did",
+    "had",
+    "ate",
+    "made",
+    "put",
+    "spent",
+    "fought",
+    "sank",
+    "gave",
+    "sought",
+    "shall",
+    "could",
+    "ought",
+    "should",
+)
+
+plverb_ambiguous_non_pres = re.compile(
+    r"^((?:thought|saw|bent|will|might|cut))((\s.*)?)$", re.IGNORECASE
+)
+
+# "..oes" -> "..oe" (the rest are "..oes" -> "o")
+
+pl_v_oes_oe = ("canoes", "floes", "oboes", "roes", "throes", "woes")
+pl_v_oes_oe_endings_size4 = ("hoes", "toes")
+pl_v_oes_oe_endings_size5 = ("shoes",)
+
+
+pl_count_zero = ("0", "no", "zero", "nil")
+
+
+pl_count_one = ("1", "a", "an", "one", "each", "every", "this", "that")
+
+pl_adj_special = {"a": "some", "an": "some", "this": "these", "that": "those"}
+
+pl_adj_special_keys = re.compile(
+    rf"^({enclose('|'.join(pl_adj_special))})$", re.IGNORECASE
+)
+
+pl_adj_poss = {
+    "my": "our",
+    "your": "your",
+    "its": "their",
+    "her": "their",
+    "his": "their",
+    "their": "their",
+}
+
+pl_adj_poss_keys = re.compile(rf"^({enclose('|'.join(pl_adj_poss))})$", re.IGNORECASE)
+
+
+# 2. INDEFINITE ARTICLES
+
+# THIS PATTERN MATCHES STRINGS OF CAPITALS STARTING WITH A "VOWEL-SOUND"
+# CONSONANT FOLLOWED BY ANOTHER CONSONANT, AND WHICH ARE NOT LIKELY
+# TO BE REAL WORDS (OH, ALL RIGHT THEN, IT'S JUST MAGIC!)
+
+A_abbrev = re.compile(
+    r"""
+^(?! FJO | [HLMNS]Y.  | RY[EO] | SQU
+  | ( F[LR]? | [HL] | MN? | N | RH? | S[CHKLMNPTVW]? | X(YL)?) [AEIOU])
+[FHLMNRSX][A-Z]
+""",
+    re.VERBOSE,
+)
+
+# THIS PATTERN CODES THE BEGINNINGS OF ALL ENGLISH WORDS BEGINING WITH A
+# 'y' FOLLOWED BY A CONSONANT. ANY OTHER Y-CONSONANT PREFIX THEREFORE
+# IMPLIES AN ABBREVIATION.
+
+A_y_cons = re.compile(r"^(y(b[lor]|cl[ea]|fere|gg|p[ios]|rou|tt))", re.IGNORECASE)
+
+# EXCEPTIONS TO EXCEPTIONS
+
+A_explicit_a = re.compile(r"^((?:unabomber|unanimous|US))", re.IGNORECASE)
+
+A_explicit_an = re.compile(
+    r"^((?:euler|hour(?!i)|heir|honest|hono[ur]|mpeg))", re.IGNORECASE
+)
+
+A_ordinal_an = re.compile(r"^([aefhilmnorsx]-?th)", re.IGNORECASE)
+
+A_ordinal_a = re.compile(r"^([bcdgjkpqtuvwyz]-?th)", re.IGNORECASE)
+
+
+# NUMERICAL INFLECTIONS
+
+nth = {
+    0: "th",
+    1: "st",
+    2: "nd",
+    3: "rd",
+    4: "th",
+    5: "th",
+    6: "th",
+    7: "th",
+    8: "th",
+    9: "th",
+    11: "th",
+    12: "th",
+    13: "th",
+}
+nth_suff = set(nth.values())
+
+ordinal = dict(
+    ty="tieth",
+    one="first",
+    two="second",
+    three="third",
+    five="fifth",
+    eight="eighth",
+    nine="ninth",
+    twelve="twelfth",
+)
+
+ordinal_suff = re.compile(rf"({'|'.join(ordinal)})\Z")
+
+
+# NUMBERS
+
+unit = ["", "one", "two", "three", "four", "five", "six", "seven", "eight", "nine"]
+teen = [
+    "ten",
+    "eleven",
+    "twelve",
+    "thirteen",
+    "fourteen",
+    "fifteen",
+    "sixteen",
+    "seventeen",
+    "eighteen",
+    "nineteen",
+]
+ten = [
+    "",
+    "",
+    "twenty",
+    "thirty",
+    "forty",
+    "fifty",
+    "sixty",
+    "seventy",
+    "eighty",
+    "ninety",
+]
+mill = [
+    " ",
+    " thousand",
+    " million",
+    " billion",
+    " trillion",
+    " quadrillion",
+    " quintillion",
+    " sextillion",
+    " septillion",
+    " octillion",
+    " nonillion",
+    " decillion",
+]
+
+
+# SUPPORT CLASSICAL PLURALIZATIONS
+
+def_classical = dict(
+    all=False, zero=False, herd=False, names=True, persons=False, ancient=False
+)
+
+all_classical = {k: True for k in def_classical}
+no_classical = {k: False for k in def_classical}
+
+
+# Maps strings to built-in constant types
+string_to_constant = {"True": True, "False": False, "None": None}
+
+
+# Pre-compiled regular expression objects
+DOLLAR_DIGITS = re.compile(r"\$(\d+)")
+FUNCTION_CALL = re.compile(r"((\w+)\([^)]*\)*)", re.IGNORECASE)
+PARTITION_WORD = re.compile(r"\A(\s*)(.+?)(\s*)\Z")
+PL_SB_POSTFIX_ADJ_STEMS_RE = re.compile(
+    rf"^(?:{pl_sb_postfix_adj_stems})$", re.IGNORECASE
+)
+PL_SB_PREP_DUAL_COMPOUND_RE = re.compile(
+    rf"^(?:{pl_sb_prep_dual_compound})$", re.IGNORECASE
+)
+DENOMINATOR = re.compile(r"(?P<denominator>.+)( (per|a) .+)")
+PLVERB_SPECIAL_S_RE = re.compile(rf"^({plverb_special_s})$")
+WHITESPACE = re.compile(r"\s")
+ENDS_WITH_S = re.compile(r"^(.*[^s])s$", re.IGNORECASE)
+ENDS_WITH_APOSTROPHE_S = re.compile(r"^(.*)'s?$")
+INDEFINITE_ARTICLE_TEST = re.compile(r"\A(\s*)(?:an?\s+)?(.+?)(\s*)\Z", re.IGNORECASE)
+SPECIAL_AN = re.compile(r"^[aefhilmnorsx]$", re.IGNORECASE)
+SPECIAL_A = re.compile(r"^[bcdgjkpqtuvwyz]$", re.IGNORECASE)
+SPECIAL_ABBREV_AN = re.compile(r"^[aefhilmnorsx][.-]", re.IGNORECASE)
+SPECIAL_ABBREV_A = re.compile(r"^[a-z][.-]", re.IGNORECASE)
+CONSONANTS = re.compile(r"^[^aeiouy]", re.IGNORECASE)
+ARTICLE_SPECIAL_EU = re.compile(r"^e[uw]", re.IGNORECASE)
+ARTICLE_SPECIAL_ONCE = re.compile(r"^onc?e\b", re.IGNORECASE)
+ARTICLE_SPECIAL_ONETIME = re.compile(r"^onetime\b", re.IGNORECASE)
+ARTICLE_SPECIAL_UNIT = re.compile(r"^uni([^nmd]|mo)", re.IGNORECASE)
+ARTICLE_SPECIAL_UBA = re.compile(r"^u[bcfghjkqrst][aeiou]", re.IGNORECASE)
+ARTICLE_SPECIAL_UKR = re.compile(r"^ukr", re.IGNORECASE)
+SPECIAL_CAPITALS = re.compile(r"^U[NK][AIEO]?")
+VOWELS = re.compile(r"^[aeiou]", re.IGNORECASE)
+
+DIGIT_GROUP = re.compile(r"(\d)")
+TWO_DIGITS = re.compile(r"(\d)(\d)")
+THREE_DIGITS = re.compile(r"(\d)(\d)(\d)")
+THREE_DIGITS_WORD = re.compile(r"(\d)(\d)(\d)(?=\D*\Z)")
+TWO_DIGITS_WORD = re.compile(r"(\d)(\d)(?=\D*\Z)")
+ONE_DIGIT_WORD = re.compile(r"(\d)(?=\D*\Z)")
+
+FOUR_DIGIT_COMMA = re.compile(r"(\d)(\d{3}(?:,|\Z))")
+NON_DIGIT = re.compile(r"\D")
+WHITESPACES_COMMA = re.compile(r"\s+,")
+COMMA_WORD = re.compile(r", (\S+)\s+\Z")
+WHITESPACES = re.compile(r"\s+")
+
+
+PRESENT_PARTICIPLE_REPLACEMENTS = (
+    (re.compile(r"ie$"), r"y"),
+    (
+        re.compile(r"ue$"),
+        r"u",
+    ),  # TODO: isn't ue$ -> u encompassed in the following rule?
+    (re.compile(r"([auy])e$"), r"\g<1>"),
+    (re.compile(r"ski$"), r"ski"),
+    (re.compile(r"[^b]i$"), r""),
+    (re.compile(r"^(are|were)$"), r"be"),
+    (re.compile(r"^(had)$"), r"hav"),
+    (re.compile(r"^(hoe)$"), r"\g<1>"),
+    (re.compile(r"([^e])e$"), r"\g<1>"),
+    (re.compile(r"er$"), r"er"),
+    (re.compile(r"([^aeiou][aeiouy]([bdgmnprst]))$"), r"\g<1>\g<2>"),
+)
+
+DIGIT = re.compile(r"\d")
+
+
+class Words(str):
+    lowered: str
+    split_: List[str]
+    first: str
+    last: str
+
+    def __init__(self, orig) -> None:
+        self.lowered = self.lower()
+        self.split_ = self.split()
+        self.first = self.split_[0]
+        self.last = self.split_[-1]
+
+
+Falsish = Any  # ideally, falsish would only validate on bool(value) is False
+
+
+_STATIC_TYPE_CHECKING = TYPE_CHECKING
+# ^-- Workaround for typeguard AST manipulation:
+#     https://github.com/agronholm/typeguard/issues/353#issuecomment-1556306554
+
+if _STATIC_TYPE_CHECKING:  # pragma: no cover
+    Word = Annotated[str, "String with at least 1 character"]
+else:
+
+    class _WordMeta(type):  # Too dynamic to be supported by mypy...
+        def __instancecheck__(self, instance: Any) -> bool:
+            return isinstance(instance, str) and len(instance) >= 1
+
+    class Word(metaclass=_WordMeta):  # type: ignore[no-redef]
+        """String with at least 1 character"""
+
+
+class engine:
+    def __init__(self) -> None:
+        self.classical_dict = def_classical.copy()
+        self.persistent_count: Optional[int] = None
+        self.mill_count = 0
+        self.pl_sb_user_defined: List[Optional[Word]] = []
+        self.pl_v_user_defined: List[Optional[Word]] = []
+        self.pl_adj_user_defined: List[Optional[Word]] = []
+        self.si_sb_user_defined: List[Optional[Word]] = []
+        self.A_a_user_defined: List[Optional[Word]] = []
+        self.thegender = "neuter"
+        self.__number_args: Optional[Dict[str, str]] = None
+
+    @property
+    def _number_args(self):
+        return cast(Dict[str, str], self.__number_args)
+
+    @_number_args.setter
+    def _number_args(self, val):
+        self.__number_args = val
+
+    @typechecked
+    def defnoun(self, singular: Optional[Word], plural: Optional[Word]) -> int:
+        """
+        Set the noun plural of singular to plural.
+
+        """
+        self.checkpat(singular)
+        self.checkpatplural(plural)
+        self.pl_sb_user_defined.extend((singular, plural))
+        self.si_sb_user_defined.extend((plural, singular))
+        return 1
+
+    @typechecked
+    def defverb(
+        self,
+        s1: Optional[Word],
+        p1: Optional[Word],
+        s2: Optional[Word],
+        p2: Optional[Word],
+        s3: Optional[Word],
+        p3: Optional[Word],
+    ) -> int:
+        """
+        Set the verb plurals for s1, s2 and s3 to p1, p2 and p3 respectively.
+
+        Where 1, 2 and 3 represent the 1st, 2nd and 3rd person forms of the verb.
+
+        """
+        self.checkpat(s1)
+        self.checkpat(s2)
+        self.checkpat(s3)
+        self.checkpatplural(p1)
+        self.checkpatplural(p2)
+        self.checkpatplural(p3)
+        self.pl_v_user_defined.extend((s1, p1, s2, p2, s3, p3))
+        return 1
+
+    @typechecked
+    def defadj(self, singular: Optional[Word], plural: Optional[Word]) -> int:
+        """
+        Set the adjective plural of singular to plural.
+
+        """
+        self.checkpat(singular)
+        self.checkpatplural(plural)
+        self.pl_adj_user_defined.extend((singular, plural))
+        return 1
+
+    @typechecked
+    def defa(self, pattern: Optional[Word]) -> int:
+        """
+        Define the indefinite article as 'a' for words matching pattern.
+
+        """
+        self.checkpat(pattern)
+        self.A_a_user_defined.extend((pattern, "a"))
+        return 1
+
+    @typechecked
+    def defan(self, pattern: Optional[Word]) -> int:
+        """
+        Define the indefinite article as 'an' for words matching pattern.
+
+        """
+        self.checkpat(pattern)
+        self.A_a_user_defined.extend((pattern, "an"))
+        return 1
+
+    def checkpat(self, pattern: Optional[Word]) -> None:
+        """
+        check for errors in a regex pattern
+        """
+        if pattern is None:
+            return
+        try:
+            re.match(pattern, "")
+        except re.error as err:
+            raise BadUserDefinedPatternError(pattern) from err
+
+    def checkpatplural(self, pattern: Optional[Word]) -> None:
+        """
+        check for errors in a regex replace pattern
+        """
+        return
+
+    @typechecked
+    def ud_match(self, word: Word, wordlist: Sequence[Optional[Word]]) -> Optional[str]:
+        for i in range(len(wordlist) - 2, -2, -2):  # backwards through even elements
+            mo = re.search(rf"^{wordlist[i]}$", word, re.IGNORECASE)
+            if mo:
+                if wordlist[i + 1] is None:
+                    return None
+                pl = DOLLAR_DIGITS.sub(
+                    r"\\1", cast(Word, wordlist[i + 1])
+                )  # change $n to \n for expand
+                return mo.expand(pl)
+        return None
+
+    def classical(self, **kwargs) -> None:
+        """
+        turn classical mode on and off for various categories
+
+        turn on all classical modes:
+        classical()
+        classical(all=True)
+
+        turn on or off specific claassical modes:
+        e.g.
+        classical(herd=True)
+        classical(names=False)
+
+        By default all classical modes are off except names.
+
+        unknown value in args or key in kwargs raises
+        exception: UnknownClasicalModeError
+
+        """
+        if not kwargs:
+            self.classical_dict = all_classical.copy()
+            return
+        if "all" in kwargs:
+            if kwargs["all"]:
+                self.classical_dict = all_classical.copy()
+            else:
+                self.classical_dict = no_classical.copy()
+
+        for k, v in kwargs.items():
+            if k in def_classical:
+                self.classical_dict[k] = v
+            else:
+                raise UnknownClassicalModeError
+
+    def num(
+        self, count: Optional[int] = None, show: Optional[int] = None
+    ) -> str:  # (;$count,$show)
+        """
+        Set the number to be used in other method calls.
+
+        Returns count.
+
+        Set show to False to return '' instead.
+
+        """
+        if count is not None:
+            try:
+                self.persistent_count = int(count)
+            except ValueError as err:
+                raise BadNumValueError from err
+            if (show is None) or show:
+                return str(count)
+        else:
+            self.persistent_count = None
+        return ""
+
+    def gender(self, gender: str) -> None:
+        """
+        set the gender for the singular of plural pronouns
+
+        can be one of:
+        'neuter'                ('they' -> 'it')
+        'feminine'              ('they' -> 'she')
+        'masculine'             ('they' -> 'he')
+        'gender-neutral'        ('they' -> 'they')
+        'feminine or masculine' ('they' -> 'she or he')
+        'masculine or feminine' ('they' -> 'he or she')
+        """
+        if gender in singular_pronoun_genders:
+            self.thegender = gender
+        else:
+            raise BadGenderError
+
+    def _get_value_from_ast(self, obj):
+        """
+        Return the value of the ast object.
+        """
+        if isinstance(obj, ast.Num):
+            return obj.n
+        elif isinstance(obj, ast.Str):
+            return obj.s
+        elif isinstance(obj, ast.List):
+            return [self._get_value_from_ast(e) for e in obj.elts]
+        elif isinstance(obj, ast.Tuple):
+            return tuple([self._get_value_from_ast(e) for e in obj.elts])
+
+        # None, True and False are NameConstants in Py3.4 and above.
+        elif isinstance(obj, ast.NameConstant):
+            return obj.value
+
+        # Probably passed a variable name.
+        # Or passed a single word without wrapping it in quotes as an argument
+        # ex: p.inflect("I plural(see)") instead of p.inflect("I plural('see')")
+        raise NameError(f"name '{obj.id}' is not defined")
+
+    def _string_to_substitute(
+        self, mo: Match, methods_dict: Dict[str, Callable]
+    ) -> str:
+        """
+        Return the string to be substituted for the match.
+        """
+        matched_text, f_name = mo.groups()
+        # matched_text is the complete match string. e.g. plural_noun(cat)
+        # f_name is the function name. e.g. plural_noun
+
+        # Return matched_text if function name is not in methods_dict
+        if f_name not in methods_dict:
+            return matched_text
+
+        # Parse the matched text
+        a_tree = ast.parse(matched_text)
+
+        # get the args and kwargs from ast objects
+        args_list = [
+            self._get_value_from_ast(a)
+            for a in a_tree.body[0].value.args  # type: ignore[attr-defined]
+        ]
+        kwargs_list = {
+            kw.arg: self._get_value_from_ast(kw.value)
+            for kw in a_tree.body[0].value.keywords  # type: ignore[attr-defined]
+        }
+
+        # Call the corresponding function
+        return methods_dict[f_name](*args_list, **kwargs_list)
+
+    # 0. PERFORM GENERAL INFLECTIONS IN A STRING
+
+    @typechecked
+    def inflect(self, text: Word) -> str:
+        """
+        Perform inflections in a string.
+
+        e.g. inflect('The plural of cat is plural(cat)') returns
+        'The plural of cat is cats'
+
+        can use plural, plural_noun, plural_verb, plural_adj,
+        singular_noun, a, an, no, ordinal, number_to_words,
+        and prespart
+
+        """
+        save_persistent_count = self.persistent_count
+
+        # Dictionary of allowed methods
+        methods_dict: Dict[str, Callable] = {
+            "plural": self.plural,
+            "plural_adj": self.plural_adj,
+            "plural_noun": self.plural_noun,
+            "plural_verb": self.plural_verb,
+            "singular_noun": self.singular_noun,
+            "a": self.a,
+            "an": self.a,
+            "no": self.no,
+            "ordinal": self.ordinal,
+            "number_to_words": self.number_to_words,
+            "present_participle": self.present_participle,
+            "num": self.num,
+        }
+
+        # Regular expression to find Python's function call syntax
+        output = FUNCTION_CALL.sub(
+            lambda mo: self._string_to_substitute(mo, methods_dict), text
+        )
+        self.persistent_count = save_persistent_count
+        return output
+
+    # ## PLURAL SUBROUTINES
+
+    def postprocess(self, orig: str, inflected) -> str:
+        inflected = str(inflected)
+        if "|" in inflected:
+            word_options = inflected.split("|")
+            # When two parts of a noun need to be pluralized
+            if len(word_options[0].split(" ")) == len(word_options[1].split(" ")):
+                result = inflected.split("|")[self.classical_dict["all"]].split(" ")
+            # When only the last part of the noun needs to be pluralized
+            else:
+                result = inflected.split(" ")
+                for index, word in enumerate(result):
+                    if "|" in word:
+                        result[index] = word.split("|")[self.classical_dict["all"]]
+        else:
+            result = inflected.split(" ")
+
+        # Try to fix word wise capitalization
+        for index, word in enumerate(orig.split(" ")):
+            if word == "I":
+                # Is this the only word for exceptions like this
+                # Where the original is fully capitalized
+                # without 'meaning' capitalization?
+                # Also this fails to handle a capitalizaion in context
+                continue
+            if word.capitalize() == word:
+                result[index] = result[index].capitalize()
+            if word == word.upper():
+                result[index] = result[index].upper()
+        return " ".join(result)
+
+    def partition_word(self, text: str) -> Tuple[str, str, str]:
+        mo = PARTITION_WORD.search(text)
+        if mo:
+            return mo.group(1), mo.group(2), mo.group(3)
+        else:
+            return "", "", ""
+
+    @typechecked
+    def plural(self, text: Word, count: Optional[Union[str, int, Any]] = None) -> str:
+        """
+        Return the plural of text.
+
+        If count supplied, then return text if count is one of:
+            1, a, an, one, each, every, this, that
+
+        otherwise return the plural.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        pre, word, post = self.partition_word(text)
+        if not word:
+            return text
+        plural = self.postprocess(
+            word,
+            self._pl_special_adjective(word, count)
+            or self._pl_special_verb(word, count)
+            or self._plnoun(word, count),
+        )
+        return f"{pre}{plural}{post}"
+
+    @typechecked
+    def plural_noun(
+        self, text: Word, count: Optional[Union[str, int, Any]] = None
+    ) -> str:
+        """
+        Return the plural of text, where text is a noun.
+
+        If count supplied, then return text if count is one of:
+            1, a, an, one, each, every, this, that
+
+        otherwise return the plural.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        pre, word, post = self.partition_word(text)
+        if not word:
+            return text
+        plural = self.postprocess(word, self._plnoun(word, count))
+        return f"{pre}{plural}{post}"
+
+    @typechecked
+    def plural_verb(
+        self, text: Word, count: Optional[Union[str, int, Any]] = None
+    ) -> str:
+        """
+        Return the plural of text, where text is a verb.
+
+        If count supplied, then return text if count is one of:
+            1, a, an, one, each, every, this, that
+
+        otherwise return the plural.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        pre, word, post = self.partition_word(text)
+        if not word:
+            return text
+        plural = self.postprocess(
+            word,
+            self._pl_special_verb(word, count) or self._pl_general_verb(word, count),
+        )
+        return f"{pre}{plural}{post}"
+
+    @typechecked
+    def plural_adj(
+        self, text: Word, count: Optional[Union[str, int, Any]] = None
+    ) -> str:
+        """
+        Return the plural of text, where text is an adjective.
+
+        If count supplied, then return text if count is one of:
+            1, a, an, one, each, every, this, that
+
+        otherwise return the plural.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        pre, word, post = self.partition_word(text)
+        if not word:
+            return text
+        plural = self.postprocess(word, self._pl_special_adjective(word, count) or word)
+        return f"{pre}{plural}{post}"
+
+    @typechecked
+    def compare(self, word1: Word, word2: Word) -> Union[str, bool]:
+        """
+        compare word1 and word2 for equality regardless of plurality
+
+        return values:
+        eq - the strings are equal
+        p:s - word1 is the plural of word2
+        s:p - word2 is the plural of word1
+        p:p - word1 and word2 are two different plural forms of the one word
+        False - otherwise
+
+        >>> compare = engine().compare
+        >>> compare("egg", "eggs")
+        's:p'
+        >>> compare('egg', 'egg')
+        'eq'
+
+        Words should not be empty.
+
+        >>> compare('egg', '')
+        Traceback (most recent call last):
+        ...
+        typeguard.TypeCheckError:...is not an instance of inflect.Word
+        """
+        norms = self.plural_noun, self.plural_verb, self.plural_adj
+        results = (self._plequal(word1, word2, norm) for norm in norms)
+        return next(filter(None, results), False)
+
+    @typechecked
+    def compare_nouns(self, word1: Word, word2: Word) -> Union[str, bool]:
+        """
+        compare word1 and word2 for equality regardless of plurality
+        word1 and word2 are to be treated as nouns
+
+        return values:
+        eq - the strings are equal
+        p:s - word1 is the plural of word2
+        s:p - word2 is the plural of word1
+        p:p - word1 and word2 are two different plural forms of the one word
+        False - otherwise
+
+        """
+        return self._plequal(word1, word2, self.plural_noun)
+
+    @typechecked
+    def compare_verbs(self, word1: Word, word2: Word) -> Union[str, bool]:
+        """
+        compare word1 and word2 for equality regardless of plurality
+        word1 and word2 are to be treated as verbs
+
+        return values:
+        eq - the strings are equal
+        p:s - word1 is the plural of word2
+        s:p - word2 is the plural of word1
+        p:p - word1 and word2 are two different plural forms of the one word
+        False - otherwise
+
+        """
+        return self._plequal(word1, word2, self.plural_verb)
+
+    @typechecked
+    def compare_adjs(self, word1: Word, word2: Word) -> Union[str, bool]:
+        """
+        compare word1 and word2 for equality regardless of plurality
+        word1 and word2 are to be treated as adjectives
+
+        return values:
+        eq - the strings are equal
+        p:s - word1 is the plural of word2
+        s:p - word2 is the plural of word1
+        p:p - word1 and word2 are two different plural forms of the one word
+        False - otherwise
+
+        """
+        return self._plequal(word1, word2, self.plural_adj)
+
+    @typechecked
+    def singular_noun(
+        self,
+        text: Word,
+        count: Optional[Union[int, str, Any]] = None,
+        gender: Optional[str] = None,
+    ) -> Union[str, Literal[False]]:
+        """
+        Return the singular of text, where text is a plural noun.
+
+        If count supplied, then return the singular if count is one of:
+            1, a, an, one, each, every, this, that or if count is None
+
+        otherwise return text unchanged.
+
+        Whitespace at the start and end is preserved.
+
+        >>> p = engine()
+        >>> p.singular_noun('horses')
+        'horse'
+        >>> p.singular_noun('knights')
+        'knight'
+
+        Returns False when a singular noun is passed.
+
+        >>> p.singular_noun('horse')
+        False
+        >>> p.singular_noun('knight')
+        False
+        >>> p.singular_noun('soldier')
+        False
+
+        """
+        pre, word, post = self.partition_word(text)
+        if not word:
+            return text
+        sing = self._sinoun(word, count=count, gender=gender)
+        if sing is not False:
+            plural = self.postprocess(word, sing)
+            return f"{pre}{plural}{post}"
+        return False
+
+    def _plequal(self, word1: str, word2: str, pl) -> Union[str, bool]:  # noqa: C901
+        classval = self.classical_dict.copy()
+        self.classical_dict = all_classical.copy()
+        if word1 == word2:
+            return "eq"
+        if word1 == pl(word2):
+            return "p:s"
+        if pl(word1) == word2:
+            return "s:p"
+        self.classical_dict = no_classical.copy()
+        if word1 == pl(word2):
+            return "p:s"
+        if pl(word1) == word2:
+            return "s:p"
+        self.classical_dict = classval.copy()
+
+        if pl == self.plural or pl == self.plural_noun:
+            if self._pl_check_plurals_N(word1, word2):
+                return "p:p"
+            if self._pl_check_plurals_N(word2, word1):
+                return "p:p"
+        if pl == self.plural or pl == self.plural_adj:
+            if self._pl_check_plurals_adj(word1, word2):
+                return "p:p"
+        return False
+
+    def _pl_reg_plurals(self, pair: str, stems: str, end1: str, end2: str) -> bool:
+        pattern = rf"({stems})({end1}\|\1{end2}|{end2}\|\1{end1})"
+        return bool(re.search(pattern, pair))
+
+    def _pl_check_plurals_N(self, word1: str, word2: str) -> bool:
+        stem_endings = (
+            (pl_sb_C_a_ata, "as", "ata"),
+            (pl_sb_C_is_ides, "is", "ides"),
+            (pl_sb_C_a_ae, "s", "e"),
+            (pl_sb_C_en_ina, "ens", "ina"),
+            (pl_sb_C_um_a, "ums", "a"),
+            (pl_sb_C_us_i, "uses", "i"),
+            (pl_sb_C_on_a, "ons", "a"),
+            (pl_sb_C_o_i_stems, "os", "i"),
+            (pl_sb_C_ex_ices, "exes", "ices"),
+            (pl_sb_C_ix_ices, "ixes", "ices"),
+            (pl_sb_C_i, "s", "i"),
+            (pl_sb_C_im, "s", "im"),
+            (".*eau", "s", "x"),
+            (".*ieu", "s", "x"),
+            (".*tri", "xes", "ces"),
+            (".{2,}[yia]n", "xes", "ges"),
+        )
+
+        words = map(Words, (word1, word2))
+        pair = "|".join(word.last for word in words)
+
+        return (
+            pair in pl_sb_irregular_s.values()
+            or pair in pl_sb_irregular.values()
+            or pair in pl_sb_irregular_caps.values()
+            or any(
+                self._pl_reg_plurals(pair, stems, end1, end2)
+                for stems, end1, end2 in stem_endings
+            )
+        )
+
+    def _pl_check_plurals_adj(self, word1: str, word2: str) -> bool:
+        word1a = word1[: word1.rfind("'")] if word1.endswith(("'s", "'")) else ""
+        word2a = word2[: word2.rfind("'")] if word2.endswith(("'s", "'")) else ""
+
+        return (
+            bool(word1a)
+            and bool(word2a)
+            and (
+                self._pl_check_plurals_N(word1a, word2a)
+                or self._pl_check_plurals_N(word2a, word1a)
+            )
+        )
+
+    def get_count(self, count: Optional[Union[str, int]] = None) -> Union[str, int]:
+        if count is None and self.persistent_count is not None:
+            count = self.persistent_count
+
+        if count is not None:
+            count = (
+                1
+                if (
+                    (str(count) in pl_count_one)
+                    or (
+                        self.classical_dict["zero"]
+                        and str(count).lower() in pl_count_zero
+                    )
+                )
+                else 2
+            )
+        else:
+            count = ""
+        return count
+
+    # @profile
+    def _plnoun(  # noqa: C901
+        self, word: str, count: Optional[Union[str, int]] = None
+    ) -> str:
+        count = self.get_count(count)
+
+        # DEFAULT TO PLURAL
+
+        if count == 1:
+            return word
+
+        # HANDLE USER-DEFINED NOUNS
+
+        value = self.ud_match(word, self.pl_sb_user_defined)
+        if value is not None:
+            return value
+
+        # HANDLE EMPTY WORD, SINGULAR COUNT AND UNINFLECTED PLURALS
+
+        if word == "":
+            return word
+
+        word = Words(word)
+
+        if word.last.lower() in pl_sb_uninflected_complete:
+            if len(word.split_) >= 3:
+                return self._handle_long_compounds(word, count=2) or word
+            return word
+
+        if word in pl_sb_uninflected_caps:
+            return word
+
+        for k, v in pl_sb_uninflected_bysize.items():
+            if word.lowered[-k:] in v:
+                return word
+
+        if self.classical_dict["herd"] and word.last.lower() in pl_sb_uninflected_herd:
+            return word
+
+        # HANDLE COMPOUNDS ("Governor General", "mother-in-law", "aide-de-camp", ETC.)
+
+        mo = PL_SB_POSTFIX_ADJ_STEMS_RE.search(word)
+        if mo and mo.group(2) != "":
+            return f"{self._plnoun(mo.group(1), 2)}{mo.group(2)}"
+
+        if " a " in word.lowered or "-a-" in word.lowered:
+            mo = PL_SB_PREP_DUAL_COMPOUND_RE.search(word)
+            if mo and mo.group(2) != "" and mo.group(3) != "":
+                return (
+                    f"{self._plnoun(mo.group(1), 2)}"
+                    f"{mo.group(2)}"
+                    f"{self._plnoun(mo.group(3))}"
+                )
+
+        if len(word.split_) >= 3:
+            handled_words = self._handle_long_compounds(word, count=2)
+            if handled_words is not None:
+                return handled_words
+
+        # only pluralize denominators in units
+        mo = DENOMINATOR.search(word.lowered)
+        if mo:
+            index = len(mo.group("denominator"))
+            return f"{self._plnoun(word[:index])}{word[index:]}"
+
+        # handle units given in degrees (only accept if
+        # there is no more than one word following)
+        # degree Celsius => degrees Celsius but degree
+        # fahrenheit hour => degree fahrenheit hours
+        if len(word.split_) >= 2 and word.split_[-2] == "degree":
+            return " ".join([self._plnoun(word.first)] + word.split_[1:])
+
+        with contextlib.suppress(ValueError):
+            return self._handle_prepositional_phrase(
+                word.lowered,
+                functools.partial(self._plnoun, count=2),
+                '-',
+            )
+
+        # HANDLE PRONOUNS
+
+        for k, v in pl_pron_acc_keys_bysize.items():
+            if word.lowered[-k:] in v:  # ends with accusative pronoun
+                for pk, pv in pl_prep_bysize.items():
+                    if word.lowered[:pk] in pv:  # starts with a prep
+                        if word.lowered.split() == [
+                            word.lowered[:pk],
+                            word.lowered[-k:],
+                        ]:
+                            # only whitespace in between
+                            return word.lowered[:-k] + pl_pron_acc[word.lowered[-k:]]
+
+        try:
+            return pl_pron_nom[word.lowered]
+        except KeyError:
+            pass
+
+        try:
+            return pl_pron_acc[word.lowered]
+        except KeyError:
+            pass
+
+        # HANDLE ISOLATED IRREGULAR PLURALS
+
+        if word.last in pl_sb_irregular_caps:
+            llen = len(word.last)
+            return f"{word[:-llen]}{pl_sb_irregular_caps[word.last]}"
+
+        lowered_last = word.last.lower()
+        if lowered_last in pl_sb_irregular:
+            llen = len(lowered_last)
+            return f"{word[:-llen]}{pl_sb_irregular[lowered_last]}"
+
+        dash_split = word.lowered.split('-')
+        if (" ".join(dash_split[-2:])).lower() in pl_sb_irregular_compound:
+            llen = len(
+                " ".join(dash_split[-2:])
+            )  # TODO: what if 2 spaces between these words?
+            return (
+                f"{word[:-llen]}"
+                f"{pl_sb_irregular_compound[(' '.join(dash_split[-2:])).lower()]}"
+            )
+
+        if word.lowered[-3:] == "quy":
+            return f"{word[:-1]}ies"
+
+        if word.lowered[-6:] == "person":
+            if self.classical_dict["persons"]:
+                return f"{word}s"
+            else:
+                return f"{word[:-4]}ople"
+
+        # HANDLE FAMILIES OF IRREGULAR PLURALS
+
+        if word.lowered[-3:] == "man":
+            for k, v in pl_sb_U_man_mans_bysize.items():
+                if word.lowered[-k:] in v:
+                    return f"{word}s"
+            for k, v in pl_sb_U_man_mans_caps_bysize.items():
+                if word[-k:] in v:
+                    return f"{word}s"
+            return f"{word[:-3]}men"
+        if word.lowered[-5:] == "mouse":
+            return f"{word[:-5]}mice"
+        if word.lowered[-5:] == "louse":
+            v = pl_sb_U_louse_lice_bysize.get(len(word))
+            if v and word.lowered in v:
+                return f"{word[:-5]}lice"
+            return f"{word}s"
+        if word.lowered[-5:] == "goose":
+            return f"{word[:-5]}geese"
+        if word.lowered[-5:] == "tooth":
+            return f"{word[:-5]}teeth"
+        if word.lowered[-4:] == "foot":
+            return f"{word[:-4]}feet"
+        if word.lowered[-4:] == "taco":
+            return f"{word[:-5]}tacos"
+
+        if word.lowered == "die":
+            return "dice"
+
+        # HANDLE UNASSIMILATED IMPORTS
+
+        if word.lowered[-4:] == "ceps":
+            return word
+        if word.lowered[-4:] == "zoon":
+            return f"{word[:-2]}a"
+        if word.lowered[-3:] in ("cis", "sis", "xis"):
+            return f"{word[:-2]}es"
+
+        for lastlet, d, numend, post in (
+            ("h", pl_sb_U_ch_chs_bysize, None, "s"),
+            ("x", pl_sb_U_ex_ices_bysize, -2, "ices"),
+            ("x", pl_sb_U_ix_ices_bysize, -2, "ices"),
+            ("m", pl_sb_U_um_a_bysize, -2, "a"),
+            ("s", pl_sb_U_us_i_bysize, -2, "i"),
+            ("n", pl_sb_U_on_a_bysize, -2, "a"),
+            ("a", pl_sb_U_a_ae_bysize, None, "e"),
+        ):
+            if word.lowered[-1] == lastlet:  # this test to add speed
+                for k, v in d.items():
+                    if word.lowered[-k:] in v:
+                        return word[:numend] + post
+
+        # HANDLE INCOMPLETELY ASSIMILATED IMPORTS
+
+        if self.classical_dict["ancient"]:
+            if word.lowered[-4:] == "trix":
+                return f"{word[:-1]}ces"
+            if word.lowered[-3:] in ("eau", "ieu"):
+                return f"{word}x"
+            if word.lowered[-3:] in ("ynx", "inx", "anx") and len(word) > 4:
+                return f"{word[:-1]}ges"
+
+            for lastlet, d, numend, post in (
+                ("n", pl_sb_C_en_ina_bysize, -2, "ina"),
+                ("x", pl_sb_C_ex_ices_bysize, -2, "ices"),
+                ("x", pl_sb_C_ix_ices_bysize, -2, "ices"),
+                ("m", pl_sb_C_um_a_bysize, -2, "a"),
+                ("s", pl_sb_C_us_i_bysize, -2, "i"),
+                ("s", pl_sb_C_us_us_bysize, None, ""),
+                ("a", pl_sb_C_a_ae_bysize, None, "e"),
+                ("a", pl_sb_C_a_ata_bysize, None, "ta"),
+                ("s", pl_sb_C_is_ides_bysize, -1, "des"),
+                ("o", pl_sb_C_o_i_bysize, -1, "i"),
+                ("n", pl_sb_C_on_a_bysize, -2, "a"),
+            ):
+                if word.lowered[-1] == lastlet:  # this test to add speed
+                    for k, v in d.items():
+                        if word.lowered[-k:] in v:
+                            return word[:numend] + post
+
+            for d, numend, post in (
+                (pl_sb_C_i_bysize, None, "i"),
+                (pl_sb_C_im_bysize, None, "im"),
+            ):
+                for k, v in d.items():
+                    if word.lowered[-k:] in v:
+                        return word[:numend] + post
+
+        # HANDLE SINGULAR NOUNS ENDING IN ...s OR OTHER SILIBANTS
+
+        if lowered_last in pl_sb_singular_s_complete:
+            return f"{word}es"
+
+        for k, v in pl_sb_singular_s_bysize.items():
+            if word.lowered[-k:] in v:
+                return f"{word}es"
+
+        if word.lowered[-2:] == "es" and word[0] == word[0].upper():
+            return f"{word}es"
+
+        if word.lowered[-1] == "z":
+            for k, v in pl_sb_z_zes_bysize.items():
+                if word.lowered[-k:] in v:
+                    return f"{word}es"
+
+            if word.lowered[-2:-1] != "z":
+                return f"{word}zes"
+
+        if word.lowered[-2:] == "ze":
+            for k, v in pl_sb_ze_zes_bysize.items():
+                if word.lowered[-k:] in v:
+                    return f"{word}s"
+
+        if word.lowered[-2:] in ("ch", "sh", "zz", "ss") or word.lowered[-1] == "x":
+            return f"{word}es"
+
+        # HANDLE ...f -> ...ves
+
+        if word.lowered[-3:] in ("elf", "alf", "olf"):
+            return f"{word[:-1]}ves"
+        if word.lowered[-3:] == "eaf" and word.lowered[-4:-3] != "d":
+            return f"{word[:-1]}ves"
+        if word.lowered[-4:] in ("nife", "life", "wife"):
+            return f"{word[:-2]}ves"
+        if word.lowered[-3:] == "arf":
+            return f"{word[:-1]}ves"
+
+        # HANDLE ...y
+
+        if word.lowered[-1] == "y":
+            if word.lowered[-2:-1] in "aeiou" or len(word) == 1:
+                return f"{word}s"
+
+            if self.classical_dict["names"]:
+                if word.lowered[-1] == "y" and word[0] == word[0].upper():
+                    return f"{word}s"
+
+            return f"{word[:-1]}ies"
+
+        # HANDLE ...o
+
+        if lowered_last in pl_sb_U_o_os_complete:
+            return f"{word}s"
+
+        for k, v in pl_sb_U_o_os_bysize.items():
+            if word.lowered[-k:] in v:
+                return f"{word}s"
+
+        if word.lowered[-2:] in ("ao", "eo", "io", "oo", "uo"):
+            return f"{word}s"
+
+        if word.lowered[-1] == "o":
+            return f"{word}es"
+
+        # OTHERWISE JUST ADD ...s
+
+        return f"{word}s"
+
+    @classmethod
+    def _handle_prepositional_phrase(cls, phrase, transform, sep):
+        """
+        Given a word or phrase possibly separated by sep, parse out
+        the prepositional phrase and apply the transform to the word
+        preceding the prepositional phrase.
+
+        Raise ValueError if the pivot is not found or if at least two
+        separators are not found.
+
+        >>> engine._handle_prepositional_phrase("man-of-war", str.upper, '-')
+        'MAN-of-war'
+        >>> engine._handle_prepositional_phrase("man of war", str.upper, ' ')
+        'MAN of war'
+        """
+        parts = phrase.split(sep)
+        if len(parts) < 3:
+            raise ValueError("Cannot handle words with fewer than two separators")
+
+        pivot = cls._find_pivot(parts, pl_prep_list_da)
+
+        transformed = transform(parts[pivot - 1]) or parts[pivot - 1]
+        return " ".join(
+            parts[: pivot - 1] + [sep.join([transformed, parts[pivot], ''])]
+        ) + " ".join(parts[(pivot + 1) :])
+
+    def _handle_long_compounds(self, word: Words, count: int) -> Union[str, None]:
+        """
+        Handles the plural and singular for compound `Words` that
+        have three or more words, based on the given count.
+
+        >>> engine()._handle_long_compounds(Words("pair of scissors"), 2)
+        'pairs of scissors'
+        >>> engine()._handle_long_compounds(Words("men beyond hills"), 1)
+        'man beyond hills'
+        """
+        inflection = self._sinoun if count == 1 else self._plnoun
+        solutions = (  # type: ignore
+            " ".join(
+                itertools.chain(
+                    leader,
+                    [inflection(cand, count), prep],  # type: ignore
+                    trailer,
+                )
+            )
+            for leader, (cand, prep), trailer in windowed_complete(word.split_, 2)
+            if prep in pl_prep_list_da  # type: ignore
+        )
+        return next(solutions, None)
+
+    @staticmethod
+    def _find_pivot(words, candidates):
+        pivots = (
+            index for index in range(1, len(words) - 1) if words[index] in candidates
+        )
+        try:
+            return next(pivots)
+        except StopIteration:
+            raise ValueError("No pivot found") from None
+
+    def _pl_special_verb(  # noqa: C901
+        self, word: str, count: Optional[Union[str, int]] = None
+    ) -> Union[str, bool]:
+        if self.classical_dict["zero"] and str(count).lower() in pl_count_zero:
+            return False
+        count = self.get_count(count)
+
+        if count == 1:
+            return word
+
+        # HANDLE USER-DEFINED VERBS
+
+        value = self.ud_match(word, self.pl_v_user_defined)
+        if value is not None:
+            return value
+
+        # HANDLE IRREGULAR PRESENT TENSE (SIMPLE AND COMPOUND)
+
+        try:
+            words = Words(word)
+        except IndexError:
+            return False  # word is ''
+
+        if words.first in plverb_irregular_pres:
+            return f"{plverb_irregular_pres[words.first]}{words[len(words.first) :]}"
+
+        # HANDLE IRREGULAR FUTURE, PRETERITE AND PERFECT TENSES
+
+        if words.first in plverb_irregular_non_pres:
+            return word
+
+        # HANDLE PRESENT NEGATIONS (SIMPLE AND COMPOUND)
+
+        if words.first.endswith("n't") and words.first[:-3] in plverb_irregular_pres:
+            return (
+                f"{plverb_irregular_pres[words.first[:-3]]}n't"
+                f"{words[len(words.first) :]}"
+            )
+
+        if words.first.endswith("n't"):
+            return word
+
+        # HANDLE SPECIAL CASES
+
+        mo = PLVERB_SPECIAL_S_RE.search(word)
+        if mo:
+            return False
+        if WHITESPACE.search(word):
+            return False
+
+        if words.lowered == "quizzes":
+            return "quiz"
+
+        # HANDLE STANDARD 3RD PERSON (CHOP THE ...(e)s OFF SINGLE WORDS)
+
+        if (
+            words.lowered[-4:] in ("ches", "shes", "zzes", "sses")
+            or words.lowered[-3:] == "xes"
+        ):
+            return words[:-2]
+
+        if words.lowered[-3:] == "ies" and len(words) > 3:
+            return words.lowered[:-3] + "y"
+
+        if (
+            words.last.lower() in pl_v_oes_oe
+            or words.lowered[-4:] in pl_v_oes_oe_endings_size4
+            or words.lowered[-5:] in pl_v_oes_oe_endings_size5
+        ):
+            return words[:-1]
+
+        if words.lowered.endswith("oes") and len(words) > 3:
+            return words.lowered[:-2]
+
+        mo = ENDS_WITH_S.search(words)
+        if mo:
+            return mo.group(1)
+
+        # OTHERWISE, A REGULAR VERB (HANDLE ELSEWHERE)
+
+        return False
+
+    def _pl_general_verb(
+        self, word: str, count: Optional[Union[str, int]] = None
+    ) -> str:
+        count = self.get_count(count)
+
+        if count == 1:
+            return word
+
+        # HANDLE AMBIGUOUS PRESENT TENSES  (SIMPLE AND COMPOUND)
+
+        mo = plverb_ambiguous_pres_keys.search(word)
+        if mo:
+            return f"{plverb_ambiguous_pres[mo.group(1).lower()]}{mo.group(2)}"
+
+        # HANDLE AMBIGUOUS PRETERITE AND PERFECT TENSES
+
+        mo = plverb_ambiguous_non_pres.search(word)
+        if mo:
+            return word
+
+        # OTHERWISE, 1st OR 2ND PERSON IS UNINFLECTED
+
+        return word
+
+    def _pl_special_adjective(
+        self, word: str, count: Optional[Union[str, int]] = None
+    ) -> Union[str, bool]:
+        count = self.get_count(count)
+
+        if count == 1:
+            return word
+
+        # HANDLE USER-DEFINED ADJECTIVES
+
+        value = self.ud_match(word, self.pl_adj_user_defined)
+        if value is not None:
+            return value
+
+        # HANDLE KNOWN CASES
+
+        mo = pl_adj_special_keys.search(word)
+        if mo:
+            return pl_adj_special[mo.group(1).lower()]
+
+        # HANDLE POSSESSIVES
+
+        mo = pl_adj_poss_keys.search(word)
+        if mo:
+            return pl_adj_poss[mo.group(1).lower()]
+
+        mo = ENDS_WITH_APOSTROPHE_S.search(word)
+        if mo:
+            pl = self.plural_noun(mo.group(1))
+            trailing_s = "" if pl[-1] == "s" else "s"
+            return f"{pl}'{trailing_s}"
+
+        # OTHERWISE, NO IDEA
+
+        return False
+
+    # @profile
+    def _sinoun(  # noqa: C901
+        self,
+        word: str,
+        count: Optional[Union[str, int]] = None,
+        gender: Optional[str] = None,
+    ) -> Union[str, bool]:
+        count = self.get_count(count)
+
+        # DEFAULT TO PLURAL
+
+        if count == 2:
+            return word
+
+        # SET THE GENDER
+
+        try:
+            if gender is None:
+                gender = self.thegender
+            elif gender not in singular_pronoun_genders:
+                raise BadGenderError
+        except (TypeError, IndexError) as err:
+            raise BadGenderError from err
+
+        # HANDLE USER-DEFINED NOUNS
+
+        value = self.ud_match(word, self.si_sb_user_defined)
+        if value is not None:
+            return value
+
+        # HANDLE EMPTY WORD, SINGULAR COUNT AND UNINFLECTED PLURALS
+
+        if word == "":
+            return word
+
+        if word in si_sb_ois_oi_case:
+            return word[:-1]
+
+        words = Words(word)
+
+        if words.last.lower() in pl_sb_uninflected_complete:
+            if len(words.split_) >= 3:
+                return self._handle_long_compounds(words, count=1) or word
+            return word
+
+        if word in pl_sb_uninflected_caps:
+            return word
+
+        for k, v in pl_sb_uninflected_bysize.items():
+            if words.lowered[-k:] in v:
+                return word
+
+        if self.classical_dict["herd"] and words.last.lower() in pl_sb_uninflected_herd:
+            return word
+
+        if words.last.lower() in pl_sb_C_us_us:
+            return word if self.classical_dict["ancient"] else False
+
+        # HANDLE COMPOUNDS ("Governor General", "mother-in-law", "aide-de-camp", ETC.)
+
+        mo = PL_SB_POSTFIX_ADJ_STEMS_RE.search(word)
+        if mo and mo.group(2) != "":
+            return f"{self._sinoun(mo.group(1), 1, gender=gender)}{mo.group(2)}"
+
+        with contextlib.suppress(ValueError):
+            return self._handle_prepositional_phrase(
+                words.lowered,
+                functools.partial(self._sinoun, count=1, gender=gender),
+                ' ',
+            )
+
+        with contextlib.suppress(ValueError):
+            return self._handle_prepositional_phrase(
+                words.lowered,
+                functools.partial(self._sinoun, count=1, gender=gender),
+                '-',
+            )
+
+        # HANDLE PRONOUNS
+
+        for k, v in si_pron_acc_keys_bysize.items():
+            if words.lowered[-k:] in v:  # ends with accusative pronoun
+                for pk, pv in pl_prep_bysize.items():
+                    if words.lowered[:pk] in pv:  # starts with a prep
+                        if words.lowered.split() == [
+                            words.lowered[:pk],
+                            words.lowered[-k:],
+                        ]:
+                            # only whitespace in between
+                            return words.lowered[:-k] + get_si_pron(
+                                "acc", words.lowered[-k:], gender
+                            )
+
+        try:
+            return get_si_pron("nom", words.lowered, gender)
+        except KeyError:
+            pass
+
+        try:
+            return get_si_pron("acc", words.lowered, gender)
+        except KeyError:
+            pass
+
+        # HANDLE ISOLATED IRREGULAR PLURALS
+
+        if words.last in si_sb_irregular_caps:
+            llen = len(words.last)
+            return f"{word[:-llen]}{si_sb_irregular_caps[words.last]}"
+
+        if words.last.lower() in si_sb_irregular:
+            llen = len(words.last.lower())
+            return f"{word[:-llen]}{si_sb_irregular[words.last.lower()]}"
+
+        dash_split = words.lowered.split("-")
+        if (" ".join(dash_split[-2:])).lower() in si_sb_irregular_compound:
+            llen = len(
+                " ".join(dash_split[-2:])
+            )  # TODO: what if 2 spaces between these words?
+            return "{}{}".format(
+                word[:-llen],
+                si_sb_irregular_compound[(" ".join(dash_split[-2:])).lower()],
+            )
+
+        if words.lowered[-5:] == "quies":
+            return word[:-3] + "y"
+
+        if words.lowered[-7:] == "persons":
+            return word[:-1]
+        if words.lowered[-6:] == "people":
+            return word[:-4] + "rson"
+
+        # HANDLE FAMILIES OF IRREGULAR PLURALS
+
+        if words.lowered[-4:] == "mans":
+            for k, v in si_sb_U_man_mans_bysize.items():
+                if words.lowered[-k:] in v:
+                    return word[:-1]
+            for k, v in si_sb_U_man_mans_caps_bysize.items():
+                if word[-k:] in v:
+                    return word[:-1]
+        if words.lowered[-3:] == "men":
+            return word[:-3] + "man"
+        if words.lowered[-4:] == "mice":
+            return word[:-4] + "mouse"
+        if words.lowered[-4:] == "lice":
+            v = si_sb_U_louse_lice_bysize.get(len(word))
+            if v and words.lowered in v:
+                return word[:-4] + "louse"
+        if words.lowered[-5:] == "geese":
+            return word[:-5] + "goose"
+        if words.lowered[-5:] == "teeth":
+            return word[:-5] + "tooth"
+        if words.lowered[-4:] == "feet":
+            return word[:-4] + "foot"
+
+        if words.lowered == "dice":
+            return "die"
+
+        # HANDLE UNASSIMILATED IMPORTS
+
+        if words.lowered[-4:] == "ceps":
+            return word
+        if words.lowered[-3:] == "zoa":
+            return word[:-1] + "on"
+
+        for lastlet, d, unass_numend, post in (
+            ("s", si_sb_U_ch_chs_bysize, -1, ""),
+            ("s", si_sb_U_ex_ices_bysize, -4, "ex"),
+            ("s", si_sb_U_ix_ices_bysize, -4, "ix"),
+            ("a", si_sb_U_um_a_bysize, -1, "um"),
+            ("i", si_sb_U_us_i_bysize, -1, "us"),
+            ("a", si_sb_U_on_a_bysize, -1, "on"),
+            ("e", si_sb_U_a_ae_bysize, -1, ""),
+        ):
+            if words.lowered[-1] == lastlet:  # this test to add speed
+                for k, v in d.items():
+                    if words.lowered[-k:] in v:
+                        return word[:unass_numend] + post
+
+        # HANDLE INCOMPLETELY ASSIMILATED IMPORTS
+
+        if self.classical_dict["ancient"]:
+            if words.lowered[-6:] == "trices":
+                return word[:-3] + "x"
+            if words.lowered[-4:] in ("eaux", "ieux"):
+                return word[:-1]
+            if words.lowered[-5:] in ("ynges", "inges", "anges") and len(word) > 6:
+                return word[:-3] + "x"
+
+            for lastlet, d, class_numend, post in (
+                ("a", si_sb_C_en_ina_bysize, -3, "en"),
+                ("s", si_sb_C_ex_ices_bysize, -4, "ex"),
+                ("s", si_sb_C_ix_ices_bysize, -4, "ix"),
+                ("a", si_sb_C_um_a_bysize, -1, "um"),
+                ("i", si_sb_C_us_i_bysize, -1, "us"),
+                ("s", pl_sb_C_us_us_bysize, None, ""),
+                ("e", si_sb_C_a_ae_bysize, -1, ""),
+                ("a", si_sb_C_a_ata_bysize, -2, ""),
+                ("s", si_sb_C_is_ides_bysize, -3, "s"),
+                ("i", si_sb_C_o_i_bysize, -1, "o"),
+                ("a", si_sb_C_on_a_bysize, -1, "on"),
+                ("m", si_sb_C_im_bysize, -2, ""),
+                ("i", si_sb_C_i_bysize, -1, ""),
+            ):
+                if words.lowered[-1] == lastlet:  # this test to add speed
+                    for k, v in d.items():
+                        if words.lowered[-k:] in v:
+                            return word[:class_numend] + post
+
+        # HANDLE PLURLS ENDING IN uses -> use
+
+        if (
+            words.lowered[-6:] == "houses"
+            or word in si_sb_uses_use_case
+            or words.last.lower() in si_sb_uses_use
+        ):
+            return word[:-1]
+
+        # HANDLE PLURLS ENDING IN ies -> ie
+
+        if word in si_sb_ies_ie_case or words.last.lower() in si_sb_ies_ie:
+            return word[:-1]
+
+        # HANDLE PLURLS ENDING IN oes -> oe
+
+        if (
+            words.lowered[-5:] == "shoes"
+            or word in si_sb_oes_oe_case
+            or words.last.lower() in si_sb_oes_oe
+        ):
+            return word[:-1]
+
+        # HANDLE SINGULAR NOUNS ENDING IN ...s OR OTHER SILIBANTS
+
+        if word in si_sb_sses_sse_case or words.last.lower() in si_sb_sses_sse:
+            return word[:-1]
+
+        if words.last.lower() in si_sb_singular_s_complete:
+            return word[:-2]
+
+        for k, v in si_sb_singular_s_bysize.items():
+            if words.lowered[-k:] in v:
+                return word[:-2]
+
+        if words.lowered[-4:] == "eses" and word[0] == word[0].upper():
+            return word[:-2]
+
+        if words.last.lower() in si_sb_z_zes:
+            return word[:-2]
+
+        if words.last.lower() in si_sb_zzes_zz:
+            return word[:-2]
+
+        if words.lowered[-4:] == "zzes":
+            return word[:-3]
+
+        if word in si_sb_ches_che_case or words.last.lower() in si_sb_ches_che:
+            return word[:-1]
+
+        if words.lowered[-4:] in ("ches", "shes"):
+            return word[:-2]
+
+        if words.last.lower() in si_sb_xes_xe:
+            return word[:-1]
+
+        if words.lowered[-3:] == "xes":
+            return word[:-2]
+
+        # HANDLE ...f -> ...ves
+
+        if word in si_sb_ves_ve_case or words.last.lower() in si_sb_ves_ve:
+            return word[:-1]
+
+        if words.lowered[-3:] == "ves":
+            if words.lowered[-5:-3] in ("el", "al", "ol"):
+                return word[:-3] + "f"
+            if words.lowered[-5:-3] == "ea" and word[-6:-5] != "d":
+                return word[:-3] + "f"
+            if words.lowered[-5:-3] in ("ni", "li", "wi"):
+                return word[:-3] + "fe"
+            if words.lowered[-5:-3] == "ar":
+                return word[:-3] + "f"
+
+        # HANDLE ...y
+
+        if words.lowered[-2:] == "ys":
+            if len(words.lowered) > 2 and words.lowered[-3] in "aeiou":
+                return word[:-1]
+
+            if self.classical_dict["names"]:
+                if words.lowered[-2:] == "ys" and word[0] == word[0].upper():
+                    return word[:-1]
+
+        if words.lowered[-3:] == "ies":
+            return word[:-3] + "y"
+
+        # HANDLE ...o
+
+        if words.lowered[-2:] == "os":
+            if words.last.lower() in si_sb_U_o_os_complete:
+                return word[:-1]
+
+            for k, v in si_sb_U_o_os_bysize.items():
+                if words.lowered[-k:] in v:
+                    return word[:-1]
+
+            if words.lowered[-3:] in ("aos", "eos", "ios", "oos", "uos"):
+                return word[:-1]
+
+        if words.lowered[-3:] == "oes":
+            return word[:-2]
+
+        # UNASSIMILATED IMPORTS FINAL RULE
+
+        if word in si_sb_es_is:
+            return word[:-2] + "is"
+
+        # OTHERWISE JUST REMOVE ...s
+
+        if words.lowered[-1] == "s":
+            return word[:-1]
+
+        # COULD NOT FIND SINGULAR
+
+        return False
+
+    # ADJECTIVES
+
+    @typechecked
+    def a(self, text: Word, count: Optional[Union[int, str, Any]] = 1) -> str:
+        """
+        Return the appropriate indefinite article followed by text.
+
+        The indefinite article is either 'a' or 'an'.
+
+        If count is not one, then return count followed by text
+        instead of 'a' or 'an'.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        mo = INDEFINITE_ARTICLE_TEST.search(text)
+        if mo:
+            word = mo.group(2)
+            if not word:
+                return text
+            pre = mo.group(1)
+            post = mo.group(3)
+            result = self._indef_article(word, count)
+            return f"{pre}{result}{post}"
+        return ""
+
+    an = a
+
+    _indef_article_cases = (
+        # HANDLE ORDINAL FORMS
+        (A_ordinal_a, "a"),
+        (A_ordinal_an, "an"),
+        # HANDLE SPECIAL CASES
+        (A_explicit_an, "an"),
+        (SPECIAL_AN, "an"),
+        (SPECIAL_A, "a"),
+        # HANDLE ABBREVIATIONS
+        (A_abbrev, "an"),
+        (SPECIAL_ABBREV_AN, "an"),
+        (SPECIAL_ABBREV_A, "a"),
+        # HANDLE CONSONANTS
+        (CONSONANTS, "a"),
+        # HANDLE SPECIAL VOWEL-FORMS
+        (ARTICLE_SPECIAL_EU, "a"),
+        (ARTICLE_SPECIAL_ONCE, "a"),
+        (ARTICLE_SPECIAL_ONETIME, "a"),
+        (ARTICLE_SPECIAL_UNIT, "a"),
+        (ARTICLE_SPECIAL_UBA, "a"),
+        (ARTICLE_SPECIAL_UKR, "a"),
+        (A_explicit_a, "a"),
+        # HANDLE SPECIAL CAPITALS
+        (SPECIAL_CAPITALS, "a"),
+        # HANDLE VOWELS
+        (VOWELS, "an"),
+        # HANDLE y...
+        # (BEFORE CERTAIN CONSONANTS IMPLIES (UNNATURALIZED) "i.." SOUND)
+        (A_y_cons, "an"),
+    )
+
+    def _indef_article(self, word: str, count: Union[int, str, Any]) -> str:
+        mycount = self.get_count(count)
+
+        if mycount != 1:
+            return f"{count} {word}"
+
+        # HANDLE USER-DEFINED VARIANTS
+
+        value = self.ud_match(word, self.A_a_user_defined)
+        if value is not None:
+            return f"{value} {word}"
+
+        matches = (
+            f'{article} {word}'
+            for regexen, article in self._indef_article_cases
+            if regexen.search(word)
+        )
+
+        # OTHERWISE, GUESS "a"
+        fallback = f'a {word}'
+        return next(matches, fallback)
+
+    # 2. TRANSLATE ZERO-QUANTIFIED $word TO "no plural($word)"
+
+    @typechecked
+    def no(self, text: Word, count: Optional[Union[int, str]] = None) -> str:
+        """
+        If count is 0, no, zero or nil, return 'no' followed by the plural
+        of text.
+
+        If count is one of:
+            1, a, an, one, each, every, this, that
+            return count followed by text.
+
+        Otherwise return count follow by the plural of text.
+
+        In the return value count is always followed by a space.
+
+        Whitespace at the start and end is preserved.
+
+        """
+        if count is None and self.persistent_count is not None:
+            count = self.persistent_count
+
+        if count is None:
+            count = 0
+        mo = PARTITION_WORD.search(text)
+        if mo:
+            pre = mo.group(1)
+            word = mo.group(2)
+            post = mo.group(3)
+        else:
+            pre = ""
+            word = ""
+            post = ""
+
+        if str(count).lower() in pl_count_zero:
+            count = 'no'
+        return f"{pre}{count} {self.plural(word, count)}{post}"
+
+    # PARTICIPLES
+
+    @typechecked
+    def present_participle(self, word: Word) -> str:
+        """
+        Return the present participle for word.
+
+        word is the 3rd person singular verb.
+
+        """
+        plv = self.plural_verb(word, 2)
+        ans = plv
+
+        for regexen, repl in PRESENT_PARTICIPLE_REPLACEMENTS:
+            ans, num = regexen.subn(repl, plv)
+            if num:
+                return f"{ans}ing"
+        return f"{ans}ing"
+
+    # NUMERICAL INFLECTIONS
+
+    @typechecked
+    def ordinal(self, num: Union[Number, Word]) -> str:
+        """
+        Return the ordinal of num.
+
+        >>> ordinal = engine().ordinal
+        >>> ordinal(1)
+        '1st'
+        >>> ordinal('one')
+        'first'
+        """
+        if DIGIT.match(str(num)):
+            if isinstance(num, (float, int)) and int(num) == num:
+                n = int(num)
+            else:
+                if "." in str(num):
+                    try:
+                        # numbers after decimal,
+                        # so only need last one for ordinal
+                        n = int(str(num)[-1])
+
+                    except ValueError:  # ends with '.', so need to use whole string
+                        n = int(str(num)[:-1])
+                else:
+                    n = int(num)  # type: ignore
+            try:
+                post = nth[n % 100]
+            except KeyError:
+                post = nth[n % 10]
+            return f"{num}{post}"
+        else:
+            return self._sub_ord(num)
+
+    def millfn(self, ind: int = 0) -> str:
+        if ind > len(mill) - 1:
+            raise NumOutOfRangeError
+        return mill[ind]
+
+    def unitfn(self, units: int, mindex: int = 0) -> str:
+        return f"{unit[units]}{self.millfn(mindex)}"
+
+    def tenfn(self, tens, units, mindex=0) -> str:
+        if tens != 1:
+            tens_part = ten[tens]
+            if tens and units:
+                hyphen = "-"
+            else:
+                hyphen = ""
+            unit_part = unit[units]
+            mill_part = self.millfn(mindex)
+            return f"{tens_part}{hyphen}{unit_part}{mill_part}"
+        return f"{teen[units]}{mill[mindex]}"
+
+    def hundfn(self, hundreds: int, tens: int, units: int, mindex: int) -> str:
+        if hundreds:
+            andword = f" {self._number_args['andword']} " if tens or units else ""
+            # use unit not unitfn as simpler
+            return (
+                f"{unit[hundreds]} hundred{andword}"
+                f"{self.tenfn(tens, units)}{self.millfn(mindex)}, "
+            )
+        if tens or units:
+            return f"{self.tenfn(tens, units)}{self.millfn(mindex)}, "
+        return ""
+
+    def group1sub(self, mo: Match) -> str:
+        units = int(mo.group(1))
+        if units == 1:
+            return f" {self._number_args['one']}, "
+        elif units:
+            return f"{unit[units]}, "
+        else:
+            return f" {self._number_args['zero']}, "
+
+    def group1bsub(self, mo: Match) -> str:
+        units = int(mo.group(1))
+        if units:
+            return f"{unit[units]}, "
+        else:
+            return f" {self._number_args['zero']}, "
+
+    def group2sub(self, mo: Match) -> str:
+        tens = int(mo.group(1))
+        units = int(mo.group(2))
+        if tens:
+            return f"{self.tenfn(tens, units)}, "
+        if units:
+            return f" {self._number_args['zero']} {unit[units]}, "
+        return f" {self._number_args['zero']} {self._number_args['zero']}, "
+
+    def group3sub(self, mo: Match) -> str:
+        hundreds = int(mo.group(1))
+        tens = int(mo.group(2))
+        units = int(mo.group(3))
+        if hundreds == 1:
+            hunword = f" {self._number_args['one']}"
+        elif hundreds:
+            hunword = str(unit[hundreds])
+        else:
+            hunword = f" {self._number_args['zero']}"
+        if tens:
+            tenword = self.tenfn(tens, units)
+        elif units:
+            tenword = f" {self._number_args['zero']} {unit[units]}"
+        else:
+            tenword = f" {self._number_args['zero']} {self._number_args['zero']}"
+        return f"{hunword} {tenword}, "
+
+    def hundsub(self, mo: Match) -> str:
+        ret = self.hundfn(
+            int(mo.group(1)), int(mo.group(2)), int(mo.group(3)), self.mill_count
+        )
+        self.mill_count += 1
+        return ret
+
+    def tensub(self, mo: Match) -> str:
+        return f"{self.tenfn(int(mo.group(1)), int(mo.group(2)), self.mill_count)}, "
+
+    def unitsub(self, mo: Match) -> str:
+        return f"{self.unitfn(int(mo.group(1)), self.mill_count)}, "
+
+    def enword(self, num: str, group: int) -> str:
+        # import pdb
+        # pdb.set_trace()
+
+        if group == 1:
+            num = DIGIT_GROUP.sub(self.group1sub, num)
+        elif group == 2:
+            num = TWO_DIGITS.sub(self.group2sub, num)
+            num = DIGIT_GROUP.sub(self.group1bsub, num, 1)
+        elif group == 3:
+            num = THREE_DIGITS.sub(self.group3sub, num)
+            num = TWO_DIGITS.sub(self.group2sub, num, 1)
+            num = DIGIT_GROUP.sub(self.group1sub, num, 1)
+        elif int(num) == 0:
+            num = self._number_args["zero"]
+        elif int(num) == 1:
+            num = self._number_args["one"]
+        else:
+            num = num.lstrip().lstrip("0")
+            self.mill_count = 0
+            # surely there's a better way to do the next bit
+            mo = THREE_DIGITS_WORD.search(num)
+            while mo:
+                num = THREE_DIGITS_WORD.sub(self.hundsub, num, 1)
+                mo = THREE_DIGITS_WORD.search(num)
+            num = TWO_DIGITS_WORD.sub(self.tensub, num, 1)
+            num = ONE_DIGIT_WORD.sub(self.unitsub, num, 1)
+        return num
+
+    @staticmethod
+    def _sub_ord(val):
+        new = ordinal_suff.sub(lambda match: ordinal[match.group(1)], val)
+        return new + "th" * (new == val)
+
+    @classmethod
+    def _chunk_num(cls, num, decimal, group):
+        if decimal:
+            max_split = -1 if group != 0 else 1
+            chunks = num.split(".", max_split)
+        else:
+            chunks = [num]
+        return cls._remove_last_blank(chunks)
+
+    @staticmethod
+    def _remove_last_blank(chunks):
+        """
+        Remove the last item from chunks if it's a blank string.
+
+        Return the resultant chunks and whether the last item was removed.
+        """
+        removed = chunks[-1] == ""
+        result = chunks[:-1] if removed else chunks
+        return result, removed
+
+    @staticmethod
+    def _get_sign(num):
+        return {'+': 'plus', '-': 'minus'}.get(num.lstrip()[0], '')
+
+    @typechecked
+    def number_to_words(  # noqa: C901
+        self,
+        num: Union[Number, Word],
+        wantlist: bool = False,
+        group: int = 0,
+        comma: Union[Falsish, str] = ",",
+        andword: str = "and",
+        zero: str = "zero",
+        one: str = "one",
+        decimal: Union[Falsish, str] = "point",
+        threshold: Optional[int] = None,
+    ) -> Union[str, List[str]]:
+        """
+        Return a number in words.
+
+        group = 1, 2 or 3 to group numbers before turning into words
+        comma: define comma
+
+        andword:
+            word for 'and'. Can be set to ''.
+            e.g. "one hundred and one" vs "one hundred one"
+
+        zero: word for '0'
+        one: word for '1'
+        decimal: word for decimal point
+        threshold: numbers above threshold not turned into words
+
+        parameters not remembered from last call. Departure from Perl version.
+        """
+        self._number_args = {"andword": andword, "zero": zero, "one": one}
+        num = str(num)
+
+        # Handle "stylistic" conversions (up to a given threshold)...
+        if threshold is not None and float(num) > threshold:
+            spnum = num.split(".", 1)
+            while comma:
+                (spnum[0], n) = FOUR_DIGIT_COMMA.subn(r"\1,\2", spnum[0])
+                if n == 0:
+                    break
+            try:
+                return f"{spnum[0]}.{spnum[1]}"
+            except IndexError:
+                return str(spnum[0])
+
+        if group < 0 or group > 3:
+            raise BadChunkingOptionError
+
+        sign = self._get_sign(num)
+
+        if num in nth_suff:
+            num = zero
+
+        myord = num[-2:] in nth_suff
+        if myord:
+            num = num[:-2]
+
+        chunks, finalpoint = self._chunk_num(num, decimal, group)
+
+        loopstart = chunks[0] == ""
+        first: bool | None = not loopstart
+
+        def _handle_chunk(chunk):
+            nonlocal first
+
+            # remove all non numeric \D
+            chunk = NON_DIGIT.sub("", chunk)
+            if chunk == "":
+                chunk = "0"
+
+            if group == 0 and not first:
+                chunk = self.enword(chunk, 1)
+            else:
+                chunk = self.enword(chunk, group)
+
+            if chunk[-2:] == ", ":
+                chunk = chunk[:-2]
+            chunk = WHITESPACES_COMMA.sub(",", chunk)
+
+            if group == 0 and first:
+                chunk = COMMA_WORD.sub(f" {andword} \\1", chunk)
+            chunk = WHITESPACES.sub(" ", chunk)
+            # chunk = re.sub(r"(\A\s|\s\Z)", self.blankfn, chunk)
+            chunk = chunk.strip()
+            if first:
+                first = None
+            return chunk
+
+        chunks[loopstart:] = map(_handle_chunk, chunks[loopstart:])
+
+        numchunks = []
+        if first != 0:
+            numchunks = chunks[0].split(f"{comma} ")
+
+        if myord and numchunks:
+            numchunks[-1] = self._sub_ord(numchunks[-1])
+
+        for chunk in chunks[1:]:
+            numchunks.append(decimal)
+            numchunks.extend(chunk.split(f"{comma} "))
+
+        if finalpoint:
+            numchunks.append(decimal)
+
+        if wantlist:
+            return [sign] * bool(sign) + numchunks
+
+        signout = f"{sign} " if sign else ""
+        valout = (
+            ', '.join(numchunks)
+            if group
+            else ''.join(self._render(numchunks, decimal, comma))
+        )
+        return signout + valout
+
+    @staticmethod
+    def _render(chunks, decimal, comma):
+        first_item = chunks.pop(0)
+        yield first_item
+        first = decimal is None or not first_item.endswith(decimal)
+        for nc in chunks:
+            if nc == decimal:
+                first = False
+            elif first:
+                yield comma
+            yield f" {nc}"
+
+    @typechecked
+    def join(
+        self,
+        words: Optional[Sequence[Word]],
+        sep: Optional[str] = None,
+        sep_spaced: bool = True,
+        final_sep: Optional[str] = None,
+        conj: str = "and",
+        conj_spaced: bool = True,
+    ) -> str:
+        """
+        Join words into a list.
+
+        e.g. join(['ant', 'bee', 'fly']) returns 'ant, bee, and fly'
+
+        options:
+        conj: replacement for 'and'
+        sep: separator. default ',', unless ',' is in the list then ';'
+        final_sep: final separator. default ',', unless ',' is in the list then ';'
+        conj_spaced: boolean. Should conj have spaces around it
+
+        """
+        if not words:
+            return ""
+        if len(words) == 1:
+            return words[0]
+
+        if conj_spaced:
+            if conj == "":
+                conj = " "
+            else:
+                conj = f" {conj} "
+
+        if len(words) == 2:
+            return f"{words[0]}{conj}{words[1]}"
+
+        if sep is None:
+            if "," in "".join(words):
+                sep = ";"
+            else:
+                sep = ","
+        if final_sep is None:
+            final_sep = sep
+
+        final_sep = f"{final_sep}{conj}"
+
+        if sep_spaced:
+            sep += " "
+
+        return f"{sep.join(words[0:-1])}{final_sep}{words[-1]}"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6028e457
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d051a91f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/py38.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/py38.cpython-312.pyc
new file mode 100644
index 00000000..c615d6cf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/__pycache__/py38.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/py38.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/py38.py
new file mode 100644
index 00000000..a2d01bd9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/compat/py38.py
@@ -0,0 +1,7 @@
+import sys
+
+
+if sys.version_info > (3, 9):
+    from typing import Annotated
+else:  # pragma: no cover
+    from typing_extensions import Annotated  # noqa: F401
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/inflect/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/METADATA
new file mode 100644
index 00000000..fe6ca5ad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/METADATA
@@ -0,0 +1,85 @@
+Metadata-Version: 2.1
+Name: jaraco.collections
+Version: 5.1.0
+Summary: Collection objects similar to those in stdlib by jaraco
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Source, https://github.com/jaraco/jaraco.collections
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: jaraco.text
+Provides-Extra: check
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'check'
+Requires-Dist: pytest-ruff >=0.2.1 ; (sys_platform != "cygwin") and extra == 'check'
+Provides-Extra: cover
+Requires-Dist: pytest-cov ; extra == 'cover'
+Provides-Extra: doc
+Requires-Dist: sphinx >=3.5 ; extra == 'doc'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'doc'
+Requires-Dist: rst.linker >=1.9 ; extra == 'doc'
+Requires-Dist: furo ; extra == 'doc'
+Requires-Dist: sphinx-lint ; extra == 'doc'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'doc'
+Provides-Extra: enabler
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'enabler'
+Provides-Extra: test
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'test'
+Provides-Extra: type
+Requires-Dist: pytest-mypy ; extra == 'type'
+
+.. image:: https://img.shields.io/pypi/v/jaraco.collections.svg
+   :target: https://pypi.org/project/jaraco.collections
+
+.. image:: https://img.shields.io/pypi/pyversions/jaraco.collections.svg
+
+.. image:: https://github.com/jaraco/jaraco.collections/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/jaraco.collections/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/jaracocollections/badge/?version=latest
+   :target: https://jaracocollections.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/jaraco.collections
+   :target: https://tidelift.com/subscription/pkg/pypi-jaraco.collections?utm_source=pypi-jaraco.collections&utm_medium=readme
+
+Models and classes to supplement the stdlib 'collections' module.
+
+See the docs, linked above, for descriptions and usage examples.
+
+Highlights include:
+
+- RangeMap: A mapping that accepts a range of values for keys.
+- Projection: A subset over an existing mapping.
+- KeyTransformingDict: Generalized mapping with keys transformed by a function.
+- FoldedCaseKeyedDict: A dict whose string keys are case-insensitive.
+- BijectiveMap: A map where keys map to values and values back to their keys.
+- ItemsAsAttributes: A mapping mix-in exposing items as attributes.
+- IdentityOverrideMap: A map whose keys map by default to themselves unless overridden.
+- FrozenDict: A hashable, immutable map.
+- Enumeration: An object whose keys are enumerated.
+- Everything: A container that contains all things.
+- Least, Greatest: Objects that are always less than or greater than any other.
+- pop_all: Return all items from the mutable sequence and remove them from that sequence.
+- DictStack: A stack of dicts, great for sharing scopes.
+- WeightedLookup: A specialized RangeMap for selecting an item by weights.
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-jaraco.collections?utm_source=pypi-jaraco.collections&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/RECORD
new file mode 100644
index 00000000..48b957ec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/RECORD
@@ -0,0 +1,10 @@
+jaraco.collections-5.1.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+jaraco.collections-5.1.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+jaraco.collections-5.1.0.dist-info/METADATA,sha256=IMUaliNsA5X1Ox9MXUWOagch5R4Wwb_3M7erp29dBtg,3933
+jaraco.collections-5.1.0.dist-info/RECORD,,
+jaraco.collections-5.1.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+jaraco.collections-5.1.0.dist-info/WHEEL,sha256=Mdi9PDNwEZptOjTlUcAth7XJDFtKrHYaQMPulZeBCiQ,91
+jaraco.collections-5.1.0.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+jaraco/collections/__init__.py,sha256=Pc1-SqjWm81ad1P0-GttpkwO_LWlnaY6gUq8gcKh2v0,26640
+jaraco/collections/__pycache__/__init__.cpython-312.pyc,,
+jaraco/collections/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/WHEEL
new file mode 100644
index 00000000..50e1e84e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (73.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/top_level.txt
new file mode 100644
index 00000000..f6205a5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.collections-5.1.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+jaraco
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/METADATA
new file mode 100644
index 00000000..a36f7c5e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/METADATA
@@ -0,0 +1,75 @@
+Metadata-Version: 2.1
+Name: jaraco.context
+Version: 5.3.0
+Summary: Useful decorators and context managers
+Home-page: https://github.com/jaraco/jaraco.context
+Author: Jason R. Coombs
+Author-email: jaraco@jaraco.com
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+License-File: LICENSE
+Requires-Dist: backports.tarfile ; python_version < "3.12"
+Provides-Extra: docs
+Requires-Dist: sphinx >=3.5 ; extra == 'docs'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'docs'
+Requires-Dist: rst.linker >=1.9 ; extra == 'docs'
+Requires-Dist: furo ; extra == 'docs'
+Requires-Dist: sphinx-lint ; extra == 'docs'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'docs'
+Provides-Extra: testing
+Requires-Dist: pytest !=8.1.1,>=6 ; extra == 'testing'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'testing'
+Requires-Dist: pytest-cov ; extra == 'testing'
+Requires-Dist: pytest-mypy ; extra == 'testing'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'testing'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'testing'
+Requires-Dist: portend ; extra == 'testing'
+
+.. image:: https://img.shields.io/pypi/v/jaraco.context.svg
+   :target: https://pypi.org/project/jaraco.context
+
+.. image:: https://img.shields.io/pypi/pyversions/jaraco.context.svg
+
+.. image:: https://github.com/jaraco/jaraco.context/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/jaraco.context/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/jaracocontext/badge/?version=latest
+   :target: https://jaracocontext.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/jaraco.context
+   :target: https://tidelift.com/subscription/pkg/pypi-jaraco.context?utm_source=pypi-jaraco.context&utm_medium=readme
+
+
+Highlights
+==========
+
+See the docs linked from the badge above for the full details, but here are some features that may be of interest.
+
+- ``ExceptionTrap`` provides a general-purpose wrapper for trapping exceptions and then acting on the outcome. Includes ``passes`` and ``raises`` decorators to replace the result of a wrapped function by a boolean indicating the outcome of the exception trap. See `this keyring commit <https://github.com/jaraco/keyring/commit/a85a7cbc6c909f8121660ed1f7b487f99a1c2bf7>`_ for an example of it in production.
+- ``suppress`` simply enables ``contextlib.suppress`` as a decorator.
+- ``on_interrupt`` is a decorator used by CLI entry points to affect the handling of a ``KeyboardInterrupt``. Inspired by `Lucretiel/autocommand#18 <https://github.com/Lucretiel/autocommand/issues/18>`_.
+- ``pushd`` is similar to pytest's ``monkeypatch.chdir`` or path's `default context <https://path.readthedocs.io/en/latest/api.html>`_, changes the current working directory for the duration of the context.
+- ``tarball`` will download a tarball, extract it, change directory, yield, then clean up after. Convenient when working with web assets.
+- ``null`` is there for those times when one code branch needs a context and the other doesn't; this null context provides symmetry across those branches.
+
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-jaraco.context?utm_source=pypi-jaraco.context&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/RECORD
new file mode 100644
index 00000000..09d191f2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/RECORD
@@ -0,0 +1,8 @@
+jaraco.context-5.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+jaraco.context-5.3.0.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+jaraco.context-5.3.0.dist-info/METADATA,sha256=xDtguJej0tN9iEXCUvxEJh2a7xceIRVBEakBLSr__tY,4020
+jaraco.context-5.3.0.dist-info/RECORD,,
+jaraco.context-5.3.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+jaraco.context-5.3.0.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+jaraco/__pycache__/context.cpython-312.pyc,,
+jaraco/context.py,sha256=REoLIxDkO5MfEYowt_WoupNCRoxBS5v7YX2PbW8lIcs,9552
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/top_level.txt
new file mode 100644
index 00000000..f6205a5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.context-5.3.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+jaraco
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/METADATA
new file mode 100644
index 00000000..c865140a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/METADATA
@@ -0,0 +1,64 @@
+Metadata-Version: 2.1
+Name: jaraco.functools
+Version: 4.0.1
+Summary: Functools like those found in stdlib
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Homepage, https://github.com/jaraco/jaraco.functools
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: more-itertools
+Provides-Extra: docs
+Requires-Dist: sphinx >=3.5 ; extra == 'docs'
+Requires-Dist: sphinx <7.2.5 ; extra == 'docs'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'docs'
+Requires-Dist: rst.linker >=1.9 ; extra == 'docs'
+Requires-Dist: furo ; extra == 'docs'
+Requires-Dist: sphinx-lint ; extra == 'docs'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'docs'
+Provides-Extra: testing
+Requires-Dist: pytest >=6 ; extra == 'testing'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'testing'
+Requires-Dist: pytest-cov ; extra == 'testing'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'testing'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'testing'
+Requires-Dist: jaraco.classes ; extra == 'testing'
+Requires-Dist: pytest-mypy ; (platform_python_implementation != "PyPy") and extra == 'testing'
+
+.. image:: https://img.shields.io/pypi/v/jaraco.functools.svg
+   :target: https://pypi.org/project/jaraco.functools
+
+.. image:: https://img.shields.io/pypi/pyversions/jaraco.functools.svg
+
+.. image:: https://github.com/jaraco/jaraco.functools/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/jaraco.functools/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/jaracofunctools/badge/?version=latest
+   :target: https://jaracofunctools.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/jaraco.functools
+   :target: https://tidelift.com/subscription/pkg/pypi-jaraco.functools?utm_source=pypi-jaraco.functools&utm_medium=readme
+
+Additional functools in the spirit of stdlib's functools.
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-jaraco.functools?utm_source=pypi-jaraco.functools&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/RECORD
new file mode 100644
index 00000000..ef3bc21e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/RECORD
@@ -0,0 +1,10 @@
+jaraco.functools-4.0.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+jaraco.functools-4.0.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+jaraco.functools-4.0.1.dist-info/METADATA,sha256=i4aUaQDX-jjdEQK5wevhegyx8JyLfin2HyvaSk3FHso,2891
+jaraco.functools-4.0.1.dist-info/RECORD,,
+jaraco.functools-4.0.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+jaraco.functools-4.0.1.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+jaraco/functools/__init__.py,sha256=hEAJaS2uSZRuF_JY4CxCHIYh79ZpxaPp9OiHyr9EJ1w,16642
+jaraco/functools/__init__.pyi,sha256=gk3dsgHzo5F_U74HzAvpNivFAPCkPJ1b2-yCd62dfnw,3878
+jaraco/functools/__pycache__/__init__.cpython-312.pyc,,
+jaraco/functools/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/top_level.txt
new file mode 100644
index 00000000..f6205a5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.functools-4.0.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+jaraco
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/METADATA
new file mode 100644
index 00000000..0258a380
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/METADATA
@@ -0,0 +1,95 @@
+Metadata-Version: 2.1
+Name: jaraco.text
+Version: 3.12.1
+Summary: Module for text manipulation
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Homepage, https://github.com/jaraco/jaraco.text
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: jaraco.functools
+Requires-Dist: jaraco.context >=4.1
+Requires-Dist: autocommand
+Requires-Dist: inflect
+Requires-Dist: more-itertools
+Requires-Dist: importlib-resources ; python_version < "3.9"
+Provides-Extra: doc
+Requires-Dist: sphinx >=3.5 ; extra == 'doc'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'doc'
+Requires-Dist: rst.linker >=1.9 ; extra == 'doc'
+Requires-Dist: furo ; extra == 'doc'
+Requires-Dist: sphinx-lint ; extra == 'doc'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'doc'
+Provides-Extra: test
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'test'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'test'
+Requires-Dist: pytest-cov ; extra == 'test'
+Requires-Dist: pytest-mypy ; extra == 'test'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'test'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'test'
+Requires-Dist: pathlib2 ; (python_version < "3.10") and extra == 'test'
+
+.. image:: https://img.shields.io/pypi/v/jaraco.text.svg
+   :target: https://pypi.org/project/jaraco.text
+
+.. image:: https://img.shields.io/pypi/pyversions/jaraco.text.svg
+
+.. image:: https://github.com/jaraco/jaraco.text/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/jaraco.text/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. image:: https://readthedocs.org/projects/jaracotext/badge/?version=latest
+   :target: https://jaracotext.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/jaraco.text
+   :target: https://tidelift.com/subscription/pkg/pypi-jaraco.text?utm_source=pypi-jaraco.text&utm_medium=readme
+
+
+This package provides handy routines for dealing with text, such as
+wrapping, substitution, trimming, stripping, prefix and suffix removal,
+line continuation, indentation, comment processing, identifier processing,
+values parsing, case insensitive comparison, and more. See the docs
+(linked in the badge above) for the detailed documentation and examples.
+
+Layouts
+=======
+
+One of the features of this package is the layouts module, which
+provides a simple example of translating keystrokes from one keyboard
+layout to another::
+
+    echo qwerty | python -m jaraco.text.to-dvorak
+    ',.pyf
+    echo  "',.pyf" | python -m jaraco.text.to-qwerty
+    qwerty
+
+Newline Reporting
+=================
+
+Need to know what newlines appear in a file?
+
+::
+
+    $ python -m jaraco.text.show-newlines README.rst
+    newline is '\n'
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-jaraco.text?utm_source=pypi-jaraco.text&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/RECORD
new file mode 100644
index 00000000..19e2d840
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/RECORD
@@ -0,0 +1,20 @@
+jaraco.text-3.12.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+jaraco.text-3.12.1.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+jaraco.text-3.12.1.dist-info/METADATA,sha256=AzWdm6ViMfDOPoQMfLWn2zgBQSGJScyqeN29TcuWXVI,3658
+jaraco.text-3.12.1.dist-info/RECORD,,
+jaraco.text-3.12.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+jaraco.text-3.12.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+jaraco.text-3.12.1.dist-info/top_level.txt,sha256=0JnN3LfXH4LIRfXL-QFOGCJzQWZO3ELx4R1d_louoQM,7
+jaraco/text/Lorem ipsum.txt,sha256=N_7c_79zxOufBY9HZ3yzMgOkNv-TkOTTio4BydrSjgs,1335
+jaraco/text/__init__.py,sha256=Y2YUqXR_orUoDaY4SkPRe6ZZhb5HUHB_Ah9RCNsVyho,16250
+jaraco/text/__pycache__/__init__.cpython-312.pyc,,
+jaraco/text/__pycache__/layouts.cpython-312.pyc,,
+jaraco/text/__pycache__/show-newlines.cpython-312.pyc,,
+jaraco/text/__pycache__/strip-prefix.cpython-312.pyc,,
+jaraco/text/__pycache__/to-dvorak.cpython-312.pyc,,
+jaraco/text/__pycache__/to-qwerty.cpython-312.pyc,,
+jaraco/text/layouts.py,sha256=HTC8aSTLZ7uXipyOXapRMC158juecjK6RVwitfmZ9_w,643
+jaraco/text/show-newlines.py,sha256=WGQa65e8lyhb92LUOLqVn6KaCtoeVgVws6WtSRmLk6w,904
+jaraco/text/strip-prefix.py,sha256=NfVXV8JVNo6nqcuYASfMV7_y4Eo8zMQqlCOGvAnRIVw,412
+jaraco/text/to-dvorak.py,sha256=1SNcbSsvISpXXg-LnybIHHY-RUFOQr36zcHkY1pWFqw,119
+jaraco/text/to-qwerty.py,sha256=s4UMQUnPwFn_dB5uZC27BurHOQcYondBfzIpVL5pEzw,119
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/top_level.txt
new file mode 100644
index 00000000..f6205a5f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco.text-3.12.1.dist-info/top_level.txt
@@ -0,0 +1 @@
+jaraco
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/__pycache__/context.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/__pycache__/context.cpython-312.pyc
new file mode 100644
index 00000000..49ebbcad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/__pycache__/context.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__init__.py
new file mode 100644
index 00000000..0d501cf9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__init__.py
@@ -0,0 +1,1091 @@
+from __future__ import annotations
+
+import collections.abc
+import copy
+import functools
+import itertools
+import operator
+import random
+import re
+from collections.abc import Container, Iterable, Mapping
+from typing import TYPE_CHECKING, Any, Callable, Dict, TypeVar, Union, overload
+
+import jaraco.text
+
+if TYPE_CHECKING:
+    from _operator import _SupportsComparison
+
+    from _typeshed import SupportsKeysAndGetItem
+    from typing_extensions import Self
+
+    _RangeMapKT = TypeVar('_RangeMapKT', bound=_SupportsComparison)
+else:
+    # _SupportsComparison doesn't exist at runtime,
+    # but _RangeMapKT is used in RangeMap's superclass' type parameters
+    _RangeMapKT = TypeVar('_RangeMapKT')
+
+_T = TypeVar('_T')
+_VT = TypeVar('_VT')
+
+_Matchable = Union[Callable, Container, Iterable, re.Pattern]
+
+
+def _dispatch(obj: _Matchable) -> Callable:
+    # can't rely on singledispatch for Union[Container, Iterable]
+    # due to ambiguity
+    # (https://peps.python.org/pep-0443/#abstract-base-classes).
+    if isinstance(obj, re.Pattern):
+        return obj.fullmatch
+    # mypy issue: https://github.com/python/mypy/issues/11071
+    if not isinstance(obj, Callable):  # type: ignore[arg-type]
+        if not isinstance(obj, Container):
+            obj = set(obj)  # type: ignore[arg-type]
+        obj = obj.__contains__
+    return obj  # type: ignore[return-value]
+
+
+class Projection(collections.abc.Mapping):
+    """
+    Project a set of keys over a mapping
+
+    >>> sample = {'a': 1, 'b': 2, 'c': 3}
+    >>> prj = Projection(['a', 'c', 'd'], sample)
+    >>> dict(prj)
+    {'a': 1, 'c': 3}
+
+    Projection also accepts an iterable or callable or pattern.
+
+    >>> iter_prj = Projection(iter('acd'), sample)
+    >>> call_prj = Projection(lambda k: ord(k) in (97, 99, 100), sample)
+    >>> pat_prj = Projection(re.compile(r'[acd]'), sample)
+    >>> prj == iter_prj == call_prj == pat_prj
+    True
+
+    Keys should only appear if they were specified and exist in the space.
+    Order is retained.
+
+    >>> list(prj)
+    ['a', 'c']
+
+    Attempting to access a key not in the projection
+    results in a KeyError.
+
+    >>> prj['b']
+    Traceback (most recent call last):
+    ...
+    KeyError: 'b'
+
+    Use the projection to update another dict.
+
+    >>> target = {'a': 2, 'b': 2}
+    >>> target.update(prj)
+    >>> target
+    {'a': 1, 'b': 2, 'c': 3}
+
+    Projection keeps a reference to the original dict, so
+    modifying the original dict may modify the Projection.
+
+    >>> del sample['a']
+    >>> dict(prj)
+    {'c': 3}
+    """
+
+    def __init__(self, keys: _Matchable, space: Mapping):
+        self._match = _dispatch(keys)
+        self._space = space
+
+    def __getitem__(self, key):
+        if not self._match(key):
+            raise KeyError(key)
+        return self._space[key]
+
+    def _keys_resolved(self):
+        return filter(self._match, self._space)
+
+    def __iter__(self):
+        return self._keys_resolved()
+
+    def __len__(self):
+        return len(tuple(self._keys_resolved()))
+
+
+class Mask(Projection):
+    """
+    The inverse of a :class:`Projection`, masking out keys.
+
+    >>> sample = {'a': 1, 'b': 2, 'c': 3}
+    >>> msk = Mask(['a', 'c', 'd'], sample)
+    >>> dict(msk)
+    {'b': 2}
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        # self._match = compose(operator.not_, self._match)
+        self._match = lambda key, orig=self._match: not orig(key)
+
+
+def dict_map(function, dictionary):
+    """
+    Return a new dict with function applied to values of dictionary.
+
+    >>> dict_map(lambda x: x+1, dict(a=1, b=2))
+    {'a': 2, 'b': 3}
+    """
+    return dict((key, function(value)) for key, value in dictionary.items())
+
+
+class RangeMap(Dict[_RangeMapKT, _VT]):
+    """
+    A dictionary-like object that uses the keys as bounds for a range.
+    Inclusion of the value for that range is determined by the
+    key_match_comparator, which defaults to less-than-or-equal.
+    A value is returned for a key if it is the first key that matches in
+    the sorted list of keys.
+
+    One may supply keyword parameters to be passed to the sort function used
+    to sort keys (i.e. key, reverse) as sort_params.
+
+    Create a map that maps 1-3 -> 'a', 4-6 -> 'b'
+
+    >>> r = RangeMap({3: 'a', 6: 'b'})  # boy, that was easy
+    >>> r[1], r[2], r[3], r[4], r[5], r[6]
+    ('a', 'a', 'a', 'b', 'b', 'b')
+
+    Even float values should work so long as the comparison operator
+    supports it.
+
+    >>> r[4.5]
+    'b'
+
+    Notice that the way rangemap is defined, it must be open-ended
+    on one side.
+
+    >>> r[0]
+    'a'
+    >>> r[-1]
+    'a'
+
+    One can close the open-end of the RangeMap by using undefined_value
+
+    >>> r = RangeMap({0: RangeMap.undefined_value, 3: 'a', 6: 'b'})
+    >>> r[0]
+    Traceback (most recent call last):
+    ...
+    KeyError: 0
+
+    One can get the first or last elements in the range by using RangeMap.Item
+
+    >>> last_item = RangeMap.Item(-1)
+    >>> r[last_item]
+    'b'
+
+    .last_item is a shortcut for Item(-1)
+
+    >>> r[RangeMap.last_item]
+    'b'
+
+    Sometimes it's useful to find the bounds for a RangeMap
+
+    >>> r.bounds()
+    (0, 6)
+
+    RangeMap supports .get(key, default)
+
+    >>> r.get(0, 'not found')
+    'not found'
+
+    >>> r.get(7, 'not found')
+    'not found'
+
+    One often wishes to define the ranges by their left-most values,
+    which requires use of sort params and a key_match_comparator.
+
+    >>> r = RangeMap({1: 'a', 4: 'b'},
+    ...     sort_params=dict(reverse=True),
+    ...     key_match_comparator=operator.ge)
+    >>> r[1], r[2], r[3], r[4], r[5], r[6]
+    ('a', 'a', 'a', 'b', 'b', 'b')
+
+    That wasn't nearly as easy as before, so an alternate constructor
+    is provided:
+
+    >>> r = RangeMap.left({1: 'a', 4: 'b', 7: RangeMap.undefined_value})
+    >>> r[1], r[2], r[3], r[4], r[5], r[6]
+    ('a', 'a', 'a', 'b', 'b', 'b')
+
+    """
+
+    def __init__(
+        self,
+        source: (
+            SupportsKeysAndGetItem[_RangeMapKT, _VT] | Iterable[tuple[_RangeMapKT, _VT]]
+        ),
+        sort_params: Mapping[str, Any] = {},
+        key_match_comparator: Callable[[_RangeMapKT, _RangeMapKT], bool] = operator.le,
+    ):
+        dict.__init__(self, source)
+        self.sort_params = sort_params
+        self.match = key_match_comparator
+
+    @classmethod
+    def left(
+        cls,
+        source: (
+            SupportsKeysAndGetItem[_RangeMapKT, _VT] | Iterable[tuple[_RangeMapKT, _VT]]
+        ),
+    ) -> Self:
+        return cls(
+            source, sort_params=dict(reverse=True), key_match_comparator=operator.ge
+        )
+
+    def __getitem__(self, item: _RangeMapKT) -> _VT:
+        sorted_keys = sorted(self.keys(), **self.sort_params)
+        if isinstance(item, RangeMap.Item):
+            result = self.__getitem__(sorted_keys[item])
+        else:
+            key = self._find_first_match_(sorted_keys, item)
+            result = dict.__getitem__(self, key)
+            if result is RangeMap.undefined_value:
+                raise KeyError(key)
+        return result
+
+    @overload  # type: ignore[override] # Signature simplified over dict and Mapping
+    def get(self, key: _RangeMapKT, default: _T) -> _VT | _T: ...
+    @overload
+    def get(self, key: _RangeMapKT, default: None = None) -> _VT | None: ...
+    def get(self, key: _RangeMapKT, default: _T | None = None) -> _VT | _T | None:
+        """
+        Return the value for key if key is in the dictionary, else default.
+        If default is not given, it defaults to None, so that this method
+        never raises a KeyError.
+        """
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def _find_first_match_(
+        self, keys: Iterable[_RangeMapKT], item: _RangeMapKT
+    ) -> _RangeMapKT:
+        is_match = functools.partial(self.match, item)
+        matches = filter(is_match, keys)
+        try:
+            return next(matches)
+        except StopIteration:
+            raise KeyError(item) from None
+
+    def bounds(self) -> tuple[_RangeMapKT, _RangeMapKT]:
+        sorted_keys = sorted(self.keys(), **self.sort_params)
+        return (sorted_keys[RangeMap.first_item], sorted_keys[RangeMap.last_item])
+
+    # some special values for the RangeMap
+    undefined_value = type('RangeValueUndefined', (), {})()
+
+    class Item(int):
+        """RangeMap Item"""
+
+    first_item = Item(0)
+    last_item = Item(-1)
+
+
+def __identity(x):
+    return x
+
+
+def sorted_items(d, key=__identity, reverse=False):
+    """
+    Return the items of the dictionary sorted by the keys.
+
+    >>> sample = dict(foo=20, bar=42, baz=10)
+    >>> tuple(sorted_items(sample))
+    (('bar', 42), ('baz', 10), ('foo', 20))
+
+    >>> reverse_string = lambda s: ''.join(reversed(s))
+    >>> tuple(sorted_items(sample, key=reverse_string))
+    (('foo', 20), ('bar', 42), ('baz', 10))
+
+    >>> tuple(sorted_items(sample, reverse=True))
+    (('foo', 20), ('baz', 10), ('bar', 42))
+    """
+
+    # wrap the key func so it operates on the first element of each item
+    def pairkey_key(item):
+        return key(item[0])
+
+    return sorted(d.items(), key=pairkey_key, reverse=reverse)
+
+
+class KeyTransformingDict(dict):
+    """
+    A dict subclass that transforms the keys before they're used.
+    Subclasses may override the default transform_key to customize behavior.
+    """
+
+    @staticmethod
+    def transform_key(key):  # pragma: nocover
+        return key
+
+    def __init__(self, *args, **kargs):
+        super().__init__()
+        # build a dictionary using the default constructs
+        d = dict(*args, **kargs)
+        # build this dictionary using transformed keys.
+        for item in d.items():
+            self.__setitem__(*item)
+
+    def __setitem__(self, key, val):
+        key = self.transform_key(key)
+        super().__setitem__(key, val)
+
+    def __getitem__(self, key):
+        key = self.transform_key(key)
+        return super().__getitem__(key)
+
+    def __contains__(self, key):
+        key = self.transform_key(key)
+        return super().__contains__(key)
+
+    def __delitem__(self, key):
+        key = self.transform_key(key)
+        return super().__delitem__(key)
+
+    def get(self, key, *args, **kwargs):
+        key = self.transform_key(key)
+        return super().get(key, *args, **kwargs)
+
+    def setdefault(self, key, *args, **kwargs):
+        key = self.transform_key(key)
+        return super().setdefault(key, *args, **kwargs)
+
+    def pop(self, key, *args, **kwargs):
+        key = self.transform_key(key)
+        return super().pop(key, *args, **kwargs)
+
+    def matching_key_for(self, key):
+        """
+        Given a key, return the actual key stored in self that matches.
+        Raise KeyError if the key isn't found.
+        """
+        try:
+            return next(e_key for e_key in self.keys() if e_key == key)
+        except StopIteration as err:
+            raise KeyError(key) from err
+
+
+class FoldedCaseKeyedDict(KeyTransformingDict):
+    """
+    A case-insensitive dictionary (keys are compared as insensitive
+    if they are strings).
+
+    >>> d = FoldedCaseKeyedDict()
+    >>> d['heLlo'] = 'world'
+    >>> list(d.keys()) == ['heLlo']
+    True
+    >>> list(d.values()) == ['world']
+    True
+    >>> d['hello'] == 'world'
+    True
+    >>> 'hello' in d
+    True
+    >>> 'HELLO' in d
+    True
+    >>> print(repr(FoldedCaseKeyedDict({'heLlo': 'world'})))
+    {'heLlo': 'world'}
+    >>> d = FoldedCaseKeyedDict({'heLlo': 'world'})
+    >>> print(d['hello'])
+    world
+    >>> print(d['Hello'])
+    world
+    >>> list(d.keys())
+    ['heLlo']
+    >>> d = FoldedCaseKeyedDict({'heLlo': 'world', 'Hello': 'world'})
+    >>> list(d.values())
+    ['world']
+    >>> key, = d.keys()
+    >>> key in ['heLlo', 'Hello']
+    True
+    >>> del d['HELLO']
+    >>> d
+    {}
+
+    get should work
+
+    >>> d['Sumthin'] = 'else'
+    >>> d.get('SUMTHIN')
+    'else'
+    >>> d.get('OTHER', 'thing')
+    'thing'
+    >>> del d['sumthin']
+
+    setdefault should also work
+
+    >>> d['This'] = 'that'
+    >>> print(d.setdefault('this', 'other'))
+    that
+    >>> len(d)
+    1
+    >>> print(d['this'])
+    that
+    >>> print(d.setdefault('That', 'other'))
+    other
+    >>> print(d['THAT'])
+    other
+
+    Make it pop!
+
+    >>> print(d.pop('THAT'))
+    other
+
+    To retrieve the key in its originally-supplied form, use matching_key_for
+
+    >>> print(d.matching_key_for('this'))
+    This
+
+    >>> d.matching_key_for('missing')
+    Traceback (most recent call last):
+    ...
+    KeyError: 'missing'
+    """
+
+    @staticmethod
+    def transform_key(key):
+        return jaraco.text.FoldedCase(key)
+
+
+class DictAdapter:
+    """
+    Provide a getitem interface for attributes of an object.
+
+    Let's say you want to get at the string.lowercase property in a formatted
+    string. It's easy with DictAdapter.
+
+    >>> import string
+    >>> print("lowercase is %(ascii_lowercase)s" % DictAdapter(string))
+    lowercase is abcdefghijklmnopqrstuvwxyz
+    """
+
+    def __init__(self, wrapped_ob):
+        self.object = wrapped_ob
+
+    def __getitem__(self, name):
+        return getattr(self.object, name)
+
+
+class ItemsAsAttributes:
+    """
+    Mix-in class to enable a mapping object to provide items as
+    attributes.
+
+    >>> C = type('C', (dict, ItemsAsAttributes), dict())
+    >>> i = C()
+    >>> i['foo'] = 'bar'
+    >>> i.foo
+    'bar'
+
+    Natural attribute access takes precedence
+
+    >>> i.foo = 'henry'
+    >>> i.foo
+    'henry'
+
+    But as you might expect, the mapping functionality is preserved.
+
+    >>> i['foo']
+    'bar'
+
+    A normal attribute error should be raised if an attribute is
+    requested that doesn't exist.
+
+    >>> i.missing
+    Traceback (most recent call last):
+    ...
+    AttributeError: 'C' object has no attribute 'missing'
+
+    It also works on dicts that customize __getitem__
+
+    >>> missing_func = lambda self, key: 'missing item'
+    >>> C = type(
+    ...     'C',
+    ...     (dict, ItemsAsAttributes),
+    ...     dict(__missing__ = missing_func),
+    ... )
+    >>> i = C()
+    >>> i.missing
+    'missing item'
+    >>> i.foo
+    'missing item'
+    """
+
+    def __getattr__(self, key):
+        try:
+            return getattr(super(), key)
+        except AttributeError as e:
+            # attempt to get the value from the mapping (return self[key])
+            #  but be careful not to lose the original exception context.
+            noval = object()
+
+            def _safe_getitem(cont, key, missing_result):
+                try:
+                    return cont[key]
+                except KeyError:
+                    return missing_result
+
+            result = _safe_getitem(self, key, noval)
+            if result is not noval:
+                return result
+            # raise the original exception, but use the original class
+            #  name, not 'super'.
+            (message,) = e.args
+            message = message.replace('super', self.__class__.__name__, 1)
+            e.args = (message,)
+            raise
+
+
+def invert_map(map):
+    """
+    Given a dictionary, return another dictionary with keys and values
+    switched. If any of the values resolve to the same key, raises
+    a ValueError.
+
+    >>> numbers = dict(a=1, b=2, c=3)
+    >>> letters = invert_map(numbers)
+    >>> letters[1]
+    'a'
+    >>> numbers['d'] = 3
+    >>> invert_map(numbers)
+    Traceback (most recent call last):
+    ...
+    ValueError: Key conflict in inverted mapping
+    """
+    res = dict((v, k) for k, v in map.items())
+    if not len(res) == len(map):
+        raise ValueError('Key conflict in inverted mapping')
+    return res
+
+
+class IdentityOverrideMap(dict):
+    """
+    A dictionary that by default maps each key to itself, but otherwise
+    acts like a normal dictionary.
+
+    >>> d = IdentityOverrideMap()
+    >>> d[42]
+    42
+    >>> d['speed'] = 'speedo'
+    >>> print(d['speed'])
+    speedo
+    """
+
+    def __missing__(self, key):
+        return key
+
+
+class DictStack(list, collections.abc.MutableMapping):
+    """
+    A stack of dictionaries that behaves as a view on those dictionaries,
+    giving preference to the last.
+
+    >>> stack = DictStack([dict(a=1, c=2), dict(b=2, a=2)])
+    >>> stack['a']
+    2
+    >>> stack['b']
+    2
+    >>> stack['c']
+    2
+    >>> len(stack)
+    3
+    >>> stack.push(dict(a=3))
+    >>> stack['a']
+    3
+    >>> stack['a'] = 4
+    >>> set(stack.keys()) == set(['a', 'b', 'c'])
+    True
+    >>> set(stack.items()) == set([('a', 4), ('b', 2), ('c', 2)])
+    True
+    >>> dict(**stack) == dict(stack) == dict(a=4, c=2, b=2)
+    True
+    >>> d = stack.pop()
+    >>> stack['a']
+    2
+    >>> d = stack.pop()
+    >>> stack['a']
+    1
+    >>> stack.get('b', None)
+    >>> 'c' in stack
+    True
+    >>> del stack['c']
+    >>> dict(stack)
+    {'a': 1}
+    """
+
+    def __iter__(self):
+        dicts = list.__iter__(self)
+        return iter(set(itertools.chain.from_iterable(c.keys() for c in dicts)))
+
+    def __getitem__(self, key):
+        for scope in reversed(tuple(list.__iter__(self))):
+            if key in scope:
+                return scope[key]
+        raise KeyError(key)
+
+    push = list.append
+
+    def __contains__(self, other):
+        return collections.abc.Mapping.__contains__(self, other)
+
+    def __len__(self):
+        return len(list(iter(self)))
+
+    def __setitem__(self, key, item):
+        last = list.__getitem__(self, -1)
+        return last.__setitem__(key, item)
+
+    def __delitem__(self, key):
+        last = list.__getitem__(self, -1)
+        return last.__delitem__(key)
+
+    # workaround for mypy confusion
+    def pop(self, *args, **kwargs):
+        return list.pop(self, *args, **kwargs)
+
+
+class BijectiveMap(dict):
+    """
+    A Bijective Map (two-way mapping).
+
+    Implemented as a simple dictionary of 2x the size, mapping values back
+    to keys.
+
+    Note, this implementation may be incomplete. If there's not a test for
+    your use case below, it's likely to fail, so please test and send pull
+    requests or patches for additional functionality needed.
+
+
+    >>> m = BijectiveMap()
+    >>> m['a'] = 'b'
+    >>> m == {'a': 'b', 'b': 'a'}
+    True
+    >>> print(m['b'])
+    a
+
+    >>> m['c'] = 'd'
+    >>> len(m)
+    2
+
+    Some weird things happen if you map an item to itself or overwrite a
+    single key of a pair, so it's disallowed.
+
+    >>> m['e'] = 'e'
+    Traceback (most recent call last):
+    ValueError: Key cannot map to itself
+
+    >>> m['d'] = 'e'
+    Traceback (most recent call last):
+    ValueError: Key/Value pairs may not overlap
+
+    >>> m['e'] = 'd'
+    Traceback (most recent call last):
+    ValueError: Key/Value pairs may not overlap
+
+    >>> print(m.pop('d'))
+    c
+
+    >>> 'c' in m
+    False
+
+    >>> m = BijectiveMap(dict(a='b'))
+    >>> len(m)
+    1
+    >>> print(m['b'])
+    a
+
+    >>> m = BijectiveMap()
+    >>> m.update(a='b')
+    >>> m['b']
+    'a'
+
+    >>> del m['b']
+    >>> len(m)
+    0
+    >>> 'a' in m
+    False
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__()
+        self.update(*args, **kwargs)
+
+    def __setitem__(self, item, value):
+        if item == value:
+            raise ValueError("Key cannot map to itself")
+        overlap = (
+            item in self
+            and self[item] != value
+            or value in self
+            and self[value] != item
+        )
+        if overlap:
+            raise ValueError("Key/Value pairs may not overlap")
+        super().__setitem__(item, value)
+        super().__setitem__(value, item)
+
+    def __delitem__(self, item):
+        self.pop(item)
+
+    def __len__(self):
+        return super().__len__() // 2
+
+    def pop(self, key, *args, **kwargs):
+        mirror = self[key]
+        super().__delitem__(mirror)
+        return super().pop(key, *args, **kwargs)
+
+    def update(self, *args, **kwargs):
+        # build a dictionary using the default constructs
+        d = dict(*args, **kwargs)
+        # build this dictionary using transformed keys.
+        for item in d.items():
+            self.__setitem__(*item)
+
+
+class FrozenDict(collections.abc.Mapping, collections.abc.Hashable):
+    """
+    An immutable mapping.
+
+    >>> a = FrozenDict(a=1, b=2)
+    >>> b = FrozenDict(a=1, b=2)
+    >>> a == b
+    True
+
+    >>> a == dict(a=1, b=2)
+    True
+    >>> dict(a=1, b=2) == a
+    True
+    >>> 'a' in a
+    True
+    >>> type(hash(a)) is type(0)
+    True
+    >>> set(iter(a)) == {'a', 'b'}
+    True
+    >>> len(a)
+    2
+    >>> a['a'] == a.get('a') == 1
+    True
+
+    >>> a['c'] = 3
+    Traceback (most recent call last):
+    ...
+    TypeError: 'FrozenDict' object does not support item assignment
+
+    >>> a.update(y=3)
+    Traceback (most recent call last):
+    ...
+    AttributeError: 'FrozenDict' object has no attribute 'update'
+
+    Copies should compare equal
+
+    >>> copy.copy(a) == a
+    True
+
+    Copies should be the same type
+
+    >>> isinstance(copy.copy(a), FrozenDict)
+    True
+
+    FrozenDict supplies .copy(), even though
+    collections.abc.Mapping doesn't demand it.
+
+    >>> a.copy() == a
+    True
+    >>> a.copy() is not a
+    True
+    """
+
+    __slots__ = ['__data']
+
+    def __new__(cls, *args, **kwargs):
+        self = super().__new__(cls)
+        self.__data = dict(*args, **kwargs)
+        return self
+
+    # Container
+    def __contains__(self, key):
+        return key in self.__data
+
+    # Hashable
+    def __hash__(self):
+        return hash(tuple(sorted(self.__data.items())))
+
+    # Mapping
+    def __iter__(self):
+        return iter(self.__data)
+
+    def __len__(self):
+        return len(self.__data)
+
+    def __getitem__(self, key):
+        return self.__data[key]
+
+    # override get for efficiency provided by dict
+    def get(self, *args, **kwargs):
+        return self.__data.get(*args, **kwargs)
+
+    # override eq to recognize underlying implementation
+    def __eq__(self, other):
+        if isinstance(other, FrozenDict):
+            other = other.__data
+        return self.__data.__eq__(other)
+
+    def copy(self):
+        "Return a shallow copy of self"
+        return copy.copy(self)
+
+
+class Enumeration(ItemsAsAttributes, BijectiveMap):
+    """
+    A convenient way to provide enumerated values
+
+    >>> e = Enumeration('a b c')
+    >>> e['a']
+    0
+
+    >>> e.a
+    0
+
+    >>> e[1]
+    'b'
+
+    >>> set(e.names) == set('abc')
+    True
+
+    >>> set(e.codes) == set(range(3))
+    True
+
+    >>> e.get('d') is None
+    True
+
+    Codes need not start with 0
+
+    >>> e = Enumeration('a b c', range(1, 4))
+    >>> e['a']
+    1
+
+    >>> e[3]
+    'c'
+    """
+
+    def __init__(self, names, codes=None):
+        if isinstance(names, str):
+            names = names.split()
+        if codes is None:
+            codes = itertools.count()
+        super().__init__(zip(names, codes))
+
+    @property
+    def names(self):
+        return (key for key in self if isinstance(key, str))
+
+    @property
+    def codes(self):
+        return (self[name] for name in self.names)
+
+
+class Everything:
+    """
+    A collection "containing" every possible thing.
+
+    >>> 'foo' in Everything()
+    True
+
+    >>> import random
+    >>> random.randint(1, 999) in Everything()
+    True
+
+    >>> random.choice([None, 'foo', 42, ('a', 'b', 'c')]) in Everything()
+    True
+    """
+
+    def __contains__(self, other):
+        return True
+
+
+class InstrumentedDict(collections.UserDict):
+    """
+    Instrument an existing dictionary with additional
+    functionality, but always reference and mutate
+    the original dictionary.
+
+    >>> orig = {'a': 1, 'b': 2}
+    >>> inst = InstrumentedDict(orig)
+    >>> inst['a']
+    1
+    >>> inst['c'] = 3
+    >>> orig['c']
+    3
+    >>> inst.keys() == orig.keys()
+    True
+    """
+
+    def __init__(self, data):
+        super().__init__()
+        self.data = data
+
+
+class Least:
+    """
+    A value that is always lesser than any other
+
+    >>> least = Least()
+    >>> 3 < least
+    False
+    >>> 3 > least
+    True
+    >>> least < 3
+    True
+    >>> least <= 3
+    True
+    >>> least > 3
+    False
+    >>> 'x' > least
+    True
+    >>> None > least
+    True
+    """
+
+    def __le__(self, other):
+        return True
+
+    __lt__ = __le__
+
+    def __ge__(self, other):
+        return False
+
+    __gt__ = __ge__
+
+
+class Greatest:
+    """
+    A value that is always greater than any other
+
+    >>> greatest = Greatest()
+    >>> 3 < greatest
+    True
+    >>> 3 > greatest
+    False
+    >>> greatest < 3
+    False
+    >>> greatest > 3
+    True
+    >>> greatest >= 3
+    True
+    >>> 'x' > greatest
+    False
+    >>> None > greatest
+    False
+    """
+
+    def __ge__(self, other):
+        return True
+
+    __gt__ = __ge__
+
+    def __le__(self, other):
+        return False
+
+    __lt__ = __le__
+
+
+def pop_all(items):
+    """
+    Clear items in place and return a copy of items.
+
+    >>> items = [1, 2, 3]
+    >>> popped = pop_all(items)
+    >>> popped is items
+    False
+    >>> popped
+    [1, 2, 3]
+    >>> items
+    []
+    """
+    result, items[:] = items[:], []
+    return result
+
+
+class FreezableDefaultDict(collections.defaultdict):
+    """
+    Often it is desirable to prevent the mutation of
+    a default dict after its initial construction, such
+    as to prevent mutation during iteration.
+
+    >>> dd = FreezableDefaultDict(list)
+    >>> dd[0].append('1')
+    >>> dd.freeze()
+    >>> dd[1]
+    []
+    >>> len(dd)
+    1
+    """
+
+    def __missing__(self, key):
+        return getattr(self, '_frozen', super().__missing__)(key)
+
+    def freeze(self):
+        self._frozen = lambda key: self.default_factory()
+
+
+class Accumulator:
+    def __init__(self, initial=0):
+        self.val = initial
+
+    def __call__(self, val):
+        self.val += val
+        return self.val
+
+
+class WeightedLookup(RangeMap):
+    """
+    Given parameters suitable for a dict representing keys
+    and a weighted proportion, return a RangeMap representing
+    spans of values proportial to the weights:
+
+    >>> even = WeightedLookup(a=1, b=1)
+
+    [0, 1) -> a
+    [1, 2) -> b
+
+    >>> lk = WeightedLookup(a=1, b=2)
+
+    [0, 1) -> a
+    [1, 3) -> b
+
+    >>> lk[.5]
+    'a'
+    >>> lk[1.5]
+    'b'
+
+    Adds ``.random()`` to select a random weighted value:
+
+    >>> lk.random() in ['a', 'b']
+    True
+
+    >>> choices = [lk.random() for x in range(1000)]
+
+    Statistically speaking, choices should be .5 a:b
+    >>> ratio = choices.count('a') / choices.count('b')
+    >>> .4 < ratio < .6
+    True
+    """
+
+    def __init__(self, *args, **kwargs):
+        raw = dict(*args, **kwargs)
+
+        # allocate keys by weight
+        indexes = map(Accumulator(), raw.values())
+        super().__init__(zip(indexes, raw.keys()), key_match_comparator=operator.lt)
+
+    def random(self):
+        lower, upper = self.bounds()
+        selector = random.random() * upper
+        return self[selector]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6e4cf5c1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/collections/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/context.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/context.py
new file mode 100644
index 00000000..61b27135
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/context.py
@@ -0,0 +1,361 @@
+from __future__ import annotations
+
+import contextlib
+import functools
+import operator
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+import urllib.request
+import warnings
+from typing import Iterator
+
+
+if sys.version_info < (3, 12):
+    from backports import tarfile
+else:
+    import tarfile
+
+
+@contextlib.contextmanager
+def pushd(dir: str | os.PathLike) -> Iterator[str | os.PathLike]:
+    """
+    >>> tmp_path = getfixture('tmp_path')
+    >>> with pushd(tmp_path):
+    ...     assert os.getcwd() == os.fspath(tmp_path)
+    >>> assert os.getcwd() != os.fspath(tmp_path)
+    """
+
+    orig = os.getcwd()
+    os.chdir(dir)
+    try:
+        yield dir
+    finally:
+        os.chdir(orig)
+
+
+@contextlib.contextmanager
+def tarball(
+    url, target_dir: str | os.PathLike | None = None
+) -> Iterator[str | os.PathLike]:
+    """
+    Get a tarball, extract it, yield, then clean up.
+
+    >>> import urllib.request
+    >>> url = getfixture('tarfile_served')
+    >>> target = getfixture('tmp_path') / 'out'
+    >>> tb = tarball(url, target_dir=target)
+    >>> import pathlib
+    >>> with tb as extracted:
+    ...     contents = pathlib.Path(extracted, 'contents.txt').read_text(encoding='utf-8')
+    >>> assert not os.path.exists(extracted)
+    """
+    if target_dir is None:
+        target_dir = os.path.basename(url).replace('.tar.gz', '').replace('.tgz', '')
+    # In the tar command, use --strip-components=1 to strip the first path and
+    #  then
+    #  use -C to cause the files to be extracted to {target_dir}. This ensures
+    #  that we always know where the files were extracted.
+    os.mkdir(target_dir)
+    try:
+        req = urllib.request.urlopen(url)
+        with tarfile.open(fileobj=req, mode='r|*') as tf:
+            tf.extractall(path=target_dir, filter=strip_first_component)
+        yield target_dir
+    finally:
+        shutil.rmtree(target_dir)
+
+
+def strip_first_component(
+    member: tarfile.TarInfo,
+    path,
+) -> tarfile.TarInfo:
+    _, member.name = member.name.split('/', 1)
+    return member
+
+
+def _compose(*cmgrs):
+    """
+    Compose any number of dependent context managers into a single one.
+
+    The last, innermost context manager may take arbitrary arguments, but
+    each successive context manager should accept the result from the
+    previous as a single parameter.
+
+    Like :func:`jaraco.functools.compose`, behavior works from right to
+    left, so the context manager should be indicated from outermost to
+    innermost.
+
+    Example, to create a context manager to change to a temporary
+    directory:
+
+    >>> temp_dir_as_cwd = _compose(pushd, temp_dir)
+    >>> with temp_dir_as_cwd() as dir:
+    ...     assert os.path.samefile(os.getcwd(), dir)
+    """
+
+    def compose_two(inner, outer):
+        def composed(*args, **kwargs):
+            with inner(*args, **kwargs) as saved, outer(saved) as res:
+                yield res
+
+        return contextlib.contextmanager(composed)
+
+    return functools.reduce(compose_two, reversed(cmgrs))
+
+
+tarball_cwd = _compose(pushd, tarball)
+
+
+@contextlib.contextmanager
+def tarball_context(*args, **kwargs):
+    warnings.warn(
+        "tarball_context is deprecated. Use tarball or tarball_cwd instead.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    pushd_ctx = kwargs.pop('pushd', pushd)
+    with tarball(*args, **kwargs) as tball, pushd_ctx(tball) as dir:
+        yield dir
+
+
+def infer_compression(url):
+    """
+    Given a URL or filename, infer the compression code for tar.
+
+    >>> infer_compression('http://foo/bar.tar.gz')
+    'z'
+    >>> infer_compression('http://foo/bar.tgz')
+    'z'
+    >>> infer_compression('file.bz')
+    'j'
+    >>> infer_compression('file.xz')
+    'J'
+    """
+    warnings.warn(
+        "infer_compression is deprecated with no replacement",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    # cheat and just assume it's the last two characters
+    compression_indicator = url[-2:]
+    mapping = dict(gz='z', bz='j', xz='J')
+    # Assume 'z' (gzip) if no match
+    return mapping.get(compression_indicator, 'z')
+
+
+@contextlib.contextmanager
+def temp_dir(remover=shutil.rmtree):
+    """
+    Create a temporary directory context. Pass a custom remover
+    to override the removal behavior.
+
+    >>> import pathlib
+    >>> with temp_dir() as the_dir:
+    ...     assert os.path.isdir(the_dir)
+    ...     _ = pathlib.Path(the_dir).joinpath('somefile').write_text('contents', encoding='utf-8')
+    >>> assert not os.path.exists(the_dir)
+    """
+    temp_dir = tempfile.mkdtemp()
+    try:
+        yield temp_dir
+    finally:
+        remover(temp_dir)
+
+
+@contextlib.contextmanager
+def repo_context(url, branch=None, quiet=True, dest_ctx=temp_dir):
+    """
+    Check out the repo indicated by url.
+
+    If dest_ctx is supplied, it should be a context manager
+    to yield the target directory for the check out.
+    """
+    exe = 'git' if 'git' in url else 'hg'
+    with dest_ctx() as repo_dir:
+        cmd = [exe, 'clone', url, repo_dir]
+        if branch:
+            cmd.extend(['--branch', branch])
+        devnull = open(os.path.devnull, 'w')
+        stdout = devnull if quiet else None
+        subprocess.check_call(cmd, stdout=stdout)
+        yield repo_dir
+
+
+def null():
+    """
+    A null context suitable to stand in for a meaningful context.
+
+    >>> with null() as value:
+    ...     assert value is None
+
+    This context is most useful when dealing with two or more code
+    branches but only some need a context. Wrap the others in a null
+    context to provide symmetry across all options.
+    """
+    warnings.warn(
+        "null is deprecated. Use contextlib.nullcontext",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    return contextlib.nullcontext()
+
+
+class ExceptionTrap:
+    """
+    A context manager that will catch certain exceptions and provide an
+    indication they occurred.
+
+    >>> with ExceptionTrap() as trap:
+    ...     raise Exception()
+    >>> bool(trap)
+    True
+
+    >>> with ExceptionTrap() as trap:
+    ...     pass
+    >>> bool(trap)
+    False
+
+    >>> with ExceptionTrap(ValueError) as trap:
+    ...     raise ValueError("1 + 1 is not 3")
+    >>> bool(trap)
+    True
+    >>> trap.value
+    ValueError('1 + 1 is not 3')
+    >>> trap.tb
+    <traceback object at ...>
+
+    >>> with ExceptionTrap(ValueError) as trap:
+    ...     raise Exception()
+    Traceback (most recent call last):
+    ...
+    Exception
+
+    >>> bool(trap)
+    False
+    """
+
+    exc_info = None, None, None
+
+    def __init__(self, exceptions=(Exception,)):
+        self.exceptions = exceptions
+
+    def __enter__(self):
+        return self
+
+    @property
+    def type(self):
+        return self.exc_info[0]
+
+    @property
+    def value(self):
+        return self.exc_info[1]
+
+    @property
+    def tb(self):
+        return self.exc_info[2]
+
+    def __exit__(self, *exc_info):
+        type = exc_info[0]
+        matches = type and issubclass(type, self.exceptions)
+        if matches:
+            self.exc_info = exc_info
+        return matches
+
+    def __bool__(self):
+        return bool(self.type)
+
+    def raises(self, func, *, _test=bool):
+        """
+        Wrap func and replace the result with the truth
+        value of the trap (True if an exception occurred).
+
+        First, give the decorator an alias to support Python 3.8
+        Syntax.
+
+        >>> raises = ExceptionTrap(ValueError).raises
+
+        Now decorate a function that always fails.
+
+        >>> @raises
+        ... def fail():
+        ...     raise ValueError('failed')
+        >>> fail()
+        True
+        """
+
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            with ExceptionTrap(self.exceptions) as trap:
+                func(*args, **kwargs)
+            return _test(trap)
+
+        return wrapper
+
+    def passes(self, func):
+        """
+        Wrap func and replace the result with the truth
+        value of the trap (True if no exception).
+
+        First, give the decorator an alias to support Python 3.8
+        Syntax.
+
+        >>> passes = ExceptionTrap(ValueError).passes
+
+        Now decorate a function that always fails.
+
+        >>> @passes
+        ... def fail():
+        ...     raise ValueError('failed')
+
+        >>> fail()
+        False
+        """
+        return self.raises(func, _test=operator.not_)
+
+
+class suppress(contextlib.suppress, contextlib.ContextDecorator):
+    """
+    A version of contextlib.suppress with decorator support.
+
+    >>> @suppress(KeyError)
+    ... def key_error():
+    ...     {}['']
+    >>> key_error()
+    """
+
+
+class on_interrupt(contextlib.ContextDecorator):
+    """
+    Replace a KeyboardInterrupt with SystemExit(1)
+
+    >>> def do_interrupt():
+    ...     raise KeyboardInterrupt()
+    >>> on_interrupt('error')(do_interrupt)()
+    Traceback (most recent call last):
+    ...
+    SystemExit: 1
+    >>> on_interrupt('error', code=255)(do_interrupt)()
+    Traceback (most recent call last):
+    ...
+    SystemExit: 255
+    >>> on_interrupt('suppress')(do_interrupt)()
+    >>> with __import__('pytest').raises(KeyboardInterrupt):
+    ...     on_interrupt('ignore')(do_interrupt)()
+    """
+
+    def __init__(self, action='error', /, code=1):
+        self.action = action
+        self.code = code
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exctype, excinst, exctb):
+        if exctype is not KeyboardInterrupt or self.action == 'ignore':
+            return
+        elif self.action == 'error':
+            raise SystemExit(self.code) from excinst
+        return self.action == 'suppress'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.py
new file mode 100644
index 00000000..ca6c22fa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.py
@@ -0,0 +1,633 @@
+import collections.abc
+import functools
+import inspect
+import itertools
+import operator
+import time
+import types
+import warnings
+
+import more_itertools
+
+
+def compose(*funcs):
+    """
+    Compose any number of unary functions into a single unary function.
+
+    >>> import textwrap
+    >>> expected = str.strip(textwrap.dedent(compose.__doc__))
+    >>> strip_and_dedent = compose(str.strip, textwrap.dedent)
+    >>> strip_and_dedent(compose.__doc__) == expected
+    True
+
+    Compose also allows the innermost function to take arbitrary arguments.
+
+    >>> round_three = lambda x: round(x, ndigits=3)
+    >>> f = compose(round_three, int.__truediv__)
+    >>> [f(3*x, x+1) for x in range(1,10)]
+    [1.5, 2.0, 2.25, 2.4, 2.5, 2.571, 2.625, 2.667, 2.7]
+    """
+
+    def compose_two(f1, f2):
+        return lambda *args, **kwargs: f1(f2(*args, **kwargs))
+
+    return functools.reduce(compose_two, funcs)
+
+
+def once(func):
+    """
+    Decorate func so it's only ever called the first time.
+
+    This decorator can ensure that an expensive or non-idempotent function
+    will not be expensive on subsequent calls and is idempotent.
+
+    >>> add_three = once(lambda a: a+3)
+    >>> add_three(3)
+    6
+    >>> add_three(9)
+    6
+    >>> add_three('12')
+    6
+
+    To reset the stored value, simply clear the property ``saved_result``.
+
+    >>> del add_three.saved_result
+    >>> add_three(9)
+    12
+    >>> add_three(8)
+    12
+
+    Or invoke 'reset()' on it.
+
+    >>> add_three.reset()
+    >>> add_three(-3)
+    0
+    >>> add_three(0)
+    0
+    """
+
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        if not hasattr(wrapper, 'saved_result'):
+            wrapper.saved_result = func(*args, **kwargs)
+        return wrapper.saved_result
+
+    wrapper.reset = lambda: vars(wrapper).__delitem__('saved_result')
+    return wrapper
+
+
+def method_cache(method, cache_wrapper=functools.lru_cache()):
+    """
+    Wrap lru_cache to support storing the cache data in the object instances.
+
+    Abstracts the common paradigm where the method explicitly saves an
+    underscore-prefixed protected property on first call and returns that
+    subsequently.
+
+    >>> class MyClass:
+    ...     calls = 0
+    ...
+    ...     @method_cache
+    ...     def method(self, value):
+    ...         self.calls += 1
+    ...         return value
+
+    >>> a = MyClass()
+    >>> a.method(3)
+    3
+    >>> for x in range(75):
+    ...     res = a.method(x)
+    >>> a.calls
+    75
+
+    Note that the apparent behavior will be exactly like that of lru_cache
+    except that the cache is stored on each instance, so values in one
+    instance will not flush values from another, and when an instance is
+    deleted, so are the cached values for that instance.
+
+    >>> b = MyClass()
+    >>> for x in range(35):
+    ...     res = b.method(x)
+    >>> b.calls
+    35
+    >>> a.method(0)
+    0
+    >>> a.calls
+    75
+
+    Note that if method had been decorated with ``functools.lru_cache()``,
+    a.calls would have been 76 (due to the cached value of 0 having been
+    flushed by the 'b' instance).
+
+    Clear the cache with ``.cache_clear()``
+
+    >>> a.method.cache_clear()
+
+    Same for a method that hasn't yet been called.
+
+    >>> c = MyClass()
+    >>> c.method.cache_clear()
+
+    Another cache wrapper may be supplied:
+
+    >>> cache = functools.lru_cache(maxsize=2)
+    >>> MyClass.method2 = method_cache(lambda self: 3, cache_wrapper=cache)
+    >>> a = MyClass()
+    >>> a.method2()
+    3
+
+    Caution - do not subsequently wrap the method with another decorator, such
+    as ``@property``, which changes the semantics of the function.
+
+    See also
+    http://code.activestate.com/recipes/577452-a-memoize-decorator-for-instance-methods/
+    for another implementation and additional justification.
+    """
+
+    def wrapper(self, *args, **kwargs):
+        # it's the first call, replace the method with a cached, bound method
+        bound_method = types.MethodType(method, self)
+        cached_method = cache_wrapper(bound_method)
+        setattr(self, method.__name__, cached_method)
+        return cached_method(*args, **kwargs)
+
+    # Support cache clear even before cache has been created.
+    wrapper.cache_clear = lambda: None
+
+    return _special_method_cache(method, cache_wrapper) or wrapper
+
+
+def _special_method_cache(method, cache_wrapper):
+    """
+    Because Python treats special methods differently, it's not
+    possible to use instance attributes to implement the cached
+    methods.
+
+    Instead, install the wrapper method under a different name
+    and return a simple proxy to that wrapper.
+
+    https://github.com/jaraco/jaraco.functools/issues/5
+    """
+    name = method.__name__
+    special_names = '__getattr__', '__getitem__'
+
+    if name not in special_names:
+        return None
+
+    wrapper_name = '__cached' + name
+
+    def proxy(self, /, *args, **kwargs):
+        if wrapper_name not in vars(self):
+            bound = types.MethodType(method, self)
+            cache = cache_wrapper(bound)
+            setattr(self, wrapper_name, cache)
+        else:
+            cache = getattr(self, wrapper_name)
+        return cache(*args, **kwargs)
+
+    return proxy
+
+
+def apply(transform):
+    """
+    Decorate a function with a transform function that is
+    invoked on results returned from the decorated function.
+
+    >>> @apply(reversed)
+    ... def get_numbers(start):
+    ...     "doc for get_numbers"
+    ...     return range(start, start+3)
+    >>> list(get_numbers(4))
+    [6, 5, 4]
+    >>> get_numbers.__doc__
+    'doc for get_numbers'
+    """
+
+    def wrap(func):
+        return functools.wraps(func)(compose(transform, func))
+
+    return wrap
+
+
+def result_invoke(action):
+    r"""
+    Decorate a function with an action function that is
+    invoked on the results returned from the decorated
+    function (for its side effect), then return the original
+    result.
+
+    >>> @result_invoke(print)
+    ... def add_two(a, b):
+    ...     return a + b
+    >>> x = add_two(2, 3)
+    5
+    >>> x
+    5
+    """
+
+    def wrap(func):
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            result = func(*args, **kwargs)
+            action(result)
+            return result
+
+        return wrapper
+
+    return wrap
+
+
+def invoke(f, /, *args, **kwargs):
+    """
+    Call a function for its side effect after initialization.
+
+    The benefit of using the decorator instead of simply invoking a function
+    after defining it is that it makes explicit the author's intent for the
+    function to be called immediately. Whereas if one simply calls the
+    function immediately, it's less obvious if that was intentional or
+    incidental. It also avoids repeating the name - the two actions, defining
+    the function and calling it immediately are modeled separately, but linked
+    by the decorator construct.
+
+    The benefit of having a function construct (opposed to just invoking some
+    behavior inline) is to serve as a scope in which the behavior occurs. It
+    avoids polluting the global namespace with local variables, provides an
+    anchor on which to attach documentation (docstring), keeps the behavior
+    logically separated (instead of conceptually separated or not separated at
+    all), and provides potential to re-use the behavior for testing or other
+    purposes.
+
+    This function is named as a pithy way to communicate, "call this function
+    primarily for its side effect", or "while defining this function, also
+    take it aside and call it". It exists because there's no Python construct
+    for "define and call" (nor should there be, as decorators serve this need
+    just fine). The behavior happens immediately and synchronously.
+
+    >>> @invoke
+    ... def func(): print("called")
+    called
+    >>> func()
+    called
+
+    Use functools.partial to pass parameters to the initial call
+
+    >>> @functools.partial(invoke, name='bingo')
+    ... def func(name): print('called with', name)
+    called with bingo
+    """
+    f(*args, **kwargs)
+    return f
+
+
+class Throttler:
+    """Rate-limit a function (or other callable)."""
+
+    def __init__(self, func, max_rate=float('Inf')):
+        if isinstance(func, Throttler):
+            func = func.func
+        self.func = func
+        self.max_rate = max_rate
+        self.reset()
+
+    def reset(self):
+        self.last_called = 0
+
+    def __call__(self, *args, **kwargs):
+        self._wait()
+        return self.func(*args, **kwargs)
+
+    def _wait(self):
+        """Ensure at least 1/max_rate seconds from last call."""
+        elapsed = time.time() - self.last_called
+        must_wait = 1 / self.max_rate - elapsed
+        time.sleep(max(0, must_wait))
+        self.last_called = time.time()
+
+    def __get__(self, obj, owner=None):
+        return first_invoke(self._wait, functools.partial(self.func, obj))
+
+
+def first_invoke(func1, func2):
+    """
+    Return a function that when invoked will invoke func1 without
+    any parameters (for its side effect) and then invoke func2
+    with whatever parameters were passed, returning its result.
+    """
+
+    def wrapper(*args, **kwargs):
+        func1()
+        return func2(*args, **kwargs)
+
+    return wrapper
+
+
+method_caller = first_invoke(
+    lambda: warnings.warn(
+        '`jaraco.functools.method_caller` is deprecated, '
+        'use `operator.methodcaller` instead',
+        DeprecationWarning,
+        stacklevel=3,
+    ),
+    operator.methodcaller,
+)
+
+
+def retry_call(func, cleanup=lambda: None, retries=0, trap=()):
+    """
+    Given a callable func, trap the indicated exceptions
+    for up to 'retries' times, invoking cleanup on the
+    exception. On the final attempt, allow any exceptions
+    to propagate.
+    """
+    attempts = itertools.count() if retries == float('inf') else range(retries)
+    for _ in attempts:
+        try:
+            return func()
+        except trap:
+            cleanup()
+
+    return func()
+
+
+def retry(*r_args, **r_kwargs):
+    """
+    Decorator wrapper for retry_call. Accepts arguments to retry_call
+    except func and then returns a decorator for the decorated function.
+
+    Ex:
+
+    >>> @retry(retries=3)
+    ... def my_func(a, b):
+    ...     "this is my funk"
+    ...     print(a, b)
+    >>> my_func.__doc__
+    'this is my funk'
+    """
+
+    def decorate(func):
+        @functools.wraps(func)
+        def wrapper(*f_args, **f_kwargs):
+            bound = functools.partial(func, *f_args, **f_kwargs)
+            return retry_call(bound, *r_args, **r_kwargs)
+
+        return wrapper
+
+    return decorate
+
+
+def print_yielded(func):
+    """
+    Convert a generator into a function that prints all yielded elements.
+
+    >>> @print_yielded
+    ... def x():
+    ...     yield 3; yield None
+    >>> x()
+    3
+    None
+    """
+    print_all = functools.partial(map, print)
+    print_results = compose(more_itertools.consume, print_all, func)
+    return functools.wraps(func)(print_results)
+
+
+def pass_none(func):
+    """
+    Wrap func so it's not called if its first param is None.
+
+    >>> print_text = pass_none(print)
+    >>> print_text('text')
+    text
+    >>> print_text(None)
+    """
+
+    @functools.wraps(func)
+    def wrapper(param, /, *args, **kwargs):
+        if param is not None:
+            return func(param, *args, **kwargs)
+        return None
+
+    return wrapper
+
+
+def assign_params(func, namespace):
+    """
+    Assign parameters from namespace where func solicits.
+
+    >>> def func(x, y=3):
+    ...     print(x, y)
+    >>> assigned = assign_params(func, dict(x=2, z=4))
+    >>> assigned()
+    2 3
+
+    The usual errors are raised if a function doesn't receive
+    its required parameters:
+
+    >>> assigned = assign_params(func, dict(y=3, z=4))
+    >>> assigned()
+    Traceback (most recent call last):
+    TypeError: func() ...argument...
+
+    It even works on methods:
+
+    >>> class Handler:
+    ...     def meth(self, arg):
+    ...         print(arg)
+    >>> assign_params(Handler().meth, dict(arg='crystal', foo='clear'))()
+    crystal
+    """
+    sig = inspect.signature(func)
+    params = sig.parameters.keys()
+    call_ns = {k: namespace[k] for k in params if k in namespace}
+    return functools.partial(func, **call_ns)
+
+
+def save_method_args(method):
+    """
+    Wrap a method such that when it is called, the args and kwargs are
+    saved on the method.
+
+    >>> class MyClass:
+    ...     @save_method_args
+    ...     def method(self, a, b):
+    ...         print(a, b)
+    >>> my_ob = MyClass()
+    >>> my_ob.method(1, 2)
+    1 2
+    >>> my_ob._saved_method.args
+    (1, 2)
+    >>> my_ob._saved_method.kwargs
+    {}
+    >>> my_ob.method(a=3, b='foo')
+    3 foo
+    >>> my_ob._saved_method.args
+    ()
+    >>> my_ob._saved_method.kwargs == dict(a=3, b='foo')
+    True
+
+    The arguments are stored on the instance, allowing for
+    different instance to save different args.
+
+    >>> your_ob = MyClass()
+    >>> your_ob.method({str('x'): 3}, b=[4])
+    {'x': 3} [4]
+    >>> your_ob._saved_method.args
+    ({'x': 3},)
+    >>> my_ob._saved_method.args
+    ()
+    """
+    args_and_kwargs = collections.namedtuple('args_and_kwargs', 'args kwargs')
+
+    @functools.wraps(method)
+    def wrapper(self, /, *args, **kwargs):
+        attr_name = '_saved_' + method.__name__
+        attr = args_and_kwargs(args, kwargs)
+        setattr(self, attr_name, attr)
+        return method(self, *args, **kwargs)
+
+    return wrapper
+
+
+def except_(*exceptions, replace=None, use=None):
+    """
+    Replace the indicated exceptions, if raised, with the indicated
+    literal replacement or evaluated expression (if present).
+
+    >>> safe_int = except_(ValueError)(int)
+    >>> safe_int('five')
+    >>> safe_int('5')
+    5
+
+    Specify a literal replacement with ``replace``.
+
+    >>> safe_int_r = except_(ValueError, replace=0)(int)
+    >>> safe_int_r('five')
+    0
+
+    Provide an expression to ``use`` to pass through particular parameters.
+
+    >>> safe_int_pt = except_(ValueError, use='args[0]')(int)
+    >>> safe_int_pt('five')
+    'five'
+
+    """
+
+    def decorate(func):
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            try:
+                return func(*args, **kwargs)
+            except exceptions:
+                try:
+                    return eval(use)
+                except TypeError:
+                    return replace
+
+        return wrapper
+
+    return decorate
+
+
+def identity(x):
+    """
+    Return the argument.
+
+    >>> o = object()
+    >>> identity(o) is o
+    True
+    """
+    return x
+
+
+def bypass_when(check, *, _op=identity):
+    """
+    Decorate a function to return its parameter when ``check``.
+
+    >>> bypassed = []  # False
+
+    >>> @bypass_when(bypassed)
+    ... def double(x):
+    ...     return x * 2
+    >>> double(2)
+    4
+    >>> bypassed[:] = [object()]  # True
+    >>> double(2)
+    2
+    """
+
+    def decorate(func):
+        @functools.wraps(func)
+        def wrapper(param, /):
+            return param if _op(check) else func(param)
+
+        return wrapper
+
+    return decorate
+
+
+def bypass_unless(check):
+    """
+    Decorate a function to return its parameter unless ``check``.
+
+    >>> enabled = [object()]  # True
+
+    >>> @bypass_unless(enabled)
+    ... def double(x):
+    ...     return x * 2
+    >>> double(2)
+    4
+    >>> del enabled[:]  # False
+    >>> double(2)
+    2
+    """
+    return bypass_when(check, _op=operator.not_)
+
+
+@functools.singledispatch
+def _splat_inner(args, func):
+    """Splat args to func."""
+    return func(*args)
+
+
+@_splat_inner.register
+def _(args: collections.abc.Mapping, func):
+    """Splat kargs to func as kwargs."""
+    return func(**args)
+
+
+def splat(func):
+    """
+    Wrap func to expect its parameters to be passed positionally in a tuple.
+
+    Has a similar effect to that of ``itertools.starmap`` over
+    simple ``map``.
+
+    >>> pairs = [(-1, 1), (0, 2)]
+    >>> more_itertools.consume(itertools.starmap(print, pairs))
+    -1 1
+    0 2
+    >>> more_itertools.consume(map(splat(print), pairs))
+    -1 1
+    0 2
+
+    The approach generalizes to other iterators that don't have a "star"
+    equivalent, such as a "starfilter".
+
+    >>> list(filter(splat(operator.add), pairs))
+    [(0, 2)]
+
+    Splat also accepts a mapping argument.
+
+    >>> def is_nice(msg, code):
+    ...     return "smile" in msg or code == 0
+    >>> msgs = [
+    ...     dict(msg='smile!', code=20),
+    ...     dict(msg='error :(', code=1),
+    ...     dict(msg='unknown', code=0),
+    ... ]
+    >>> for msg in filter(splat(is_nice), msgs):
+    ...     print(msg)
+    {'msg': 'smile!', 'code': 20}
+    {'msg': 'unknown', 'code': 0}
+    """
+    return functools.wraps(func)(functools.partial(_splat_inner, func=func))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.pyi b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.pyi
new file mode 100644
index 00000000..19191bf9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__init__.pyi
@@ -0,0 +1,125 @@
+from collections.abc import Callable, Hashable, Iterator
+from functools import partial
+from operator import methodcaller
+import sys
+from typing import (
+    Any,
+    Generic,
+    Protocol,
+    TypeVar,
+    overload,
+)
+
+if sys.version_info >= (3, 10):
+    from typing import Concatenate, ParamSpec
+else:
+    from typing_extensions import Concatenate, ParamSpec
+
+_P = ParamSpec('_P')
+_R = TypeVar('_R')
+_T = TypeVar('_T')
+_R1 = TypeVar('_R1')
+_R2 = TypeVar('_R2')
+_V = TypeVar('_V')
+_S = TypeVar('_S')
+_R_co = TypeVar('_R_co', covariant=True)
+
+class _OnceCallable(Protocol[_P, _R]):
+    saved_result: _R
+    reset: Callable[[], None]
+    def __call__(self, *args: _P.args, **kwargs: _P.kwargs) -> _R: ...
+
+class _ProxyMethodCacheWrapper(Protocol[_R_co]):
+    cache_clear: Callable[[], None]
+    def __call__(self, *args: Hashable, **kwargs: Hashable) -> _R_co: ...
+
+class _MethodCacheWrapper(Protocol[_R_co]):
+    def cache_clear(self) -> None: ...
+    def __call__(self, *args: Hashable, **kwargs: Hashable) -> _R_co: ...
+
+# `compose()` overloads below will cover most use cases.
+
+@overload
+def compose(
+    __func1: Callable[[_R], _T],
+    __func2: Callable[_P, _R],
+    /,
+) -> Callable[_P, _T]: ...
+@overload
+def compose(
+    __func1: Callable[[_R], _T],
+    __func2: Callable[[_R1], _R],
+    __func3: Callable[_P, _R1],
+    /,
+) -> Callable[_P, _T]: ...
+@overload
+def compose(
+    __func1: Callable[[_R], _T],
+    __func2: Callable[[_R2], _R],
+    __func3: Callable[[_R1], _R2],
+    __func4: Callable[_P, _R1],
+    /,
+) -> Callable[_P, _T]: ...
+def once(func: Callable[_P, _R]) -> _OnceCallable[_P, _R]: ...
+def method_cache(
+    method: Callable[..., _R],
+    cache_wrapper: Callable[[Callable[..., _R]], _MethodCacheWrapper[_R]] = ...,
+) -> _MethodCacheWrapper[_R] | _ProxyMethodCacheWrapper[_R]: ...
+def apply(
+    transform: Callable[[_R], _T]
+) -> Callable[[Callable[_P, _R]], Callable[_P, _T]]: ...
+def result_invoke(
+    action: Callable[[_R], Any]
+) -> Callable[[Callable[_P, _R]], Callable[_P, _R]]: ...
+def invoke(
+    f: Callable[_P, _R], /, *args: _P.args, **kwargs: _P.kwargs
+) -> Callable[_P, _R]: ...
+
+class Throttler(Generic[_R]):
+    last_called: float
+    func: Callable[..., _R]
+    max_rate: float
+    def __init__(
+        self, func: Callable[..., _R] | Throttler[_R], max_rate: float = ...
+    ) -> None: ...
+    def reset(self) -> None: ...
+    def __call__(self, *args: Any, **kwargs: Any) -> _R: ...
+    def __get__(self, obj: Any, owner: type[Any] | None = ...) -> Callable[..., _R]: ...
+
+def first_invoke(
+    func1: Callable[..., Any], func2: Callable[_P, _R]
+) -> Callable[_P, _R]: ...
+
+method_caller: Callable[..., methodcaller]
+
+def retry_call(
+    func: Callable[..., _R],
+    cleanup: Callable[..., None] = ...,
+    retries: int | float = ...,
+    trap: type[BaseException] | tuple[type[BaseException], ...] = ...,
+) -> _R: ...
+def retry(
+    cleanup: Callable[..., None] = ...,
+    retries: int | float = ...,
+    trap: type[BaseException] | tuple[type[BaseException], ...] = ...,
+) -> Callable[[Callable[..., _R]], Callable[..., _R]]: ...
+def print_yielded(func: Callable[_P, Iterator[Any]]) -> Callable[_P, None]: ...
+def pass_none(
+    func: Callable[Concatenate[_T, _P], _R]
+) -> Callable[Concatenate[_T, _P], _R]: ...
+def assign_params(
+    func: Callable[..., _R], namespace: dict[str, Any]
+) -> partial[_R]: ...
+def save_method_args(
+    method: Callable[Concatenate[_S, _P], _R]
+) -> Callable[Concatenate[_S, _P], _R]: ...
+def except_(
+    *exceptions: type[BaseException], replace: Any = ..., use: Any = ...
+) -> Callable[[Callable[_P, Any]], Callable[_P, Any]]: ...
+def identity(x: _T) -> _T: ...
+def bypass_when(
+    check: _V, *, _op: Callable[[_V], Any] = ...
+) -> Callable[[Callable[[_T], _R]], Callable[[_T], _T | _R]]: ...
+def bypass_unless(
+    check: Any,
+) -> Callable[[Callable[[_T], _R]], Callable[[_T], _T | _R]]: ...
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..370b3efd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/functools/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/Lorem ipsum.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/Lorem ipsum.txt
new file mode 100644
index 00000000..986f944b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/Lorem ipsum.txt	
@@ -0,0 +1,2 @@
+Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
+Curabitur pretium tincidunt lacus. Nulla gravida orci a odio. Nullam varius, turpis et commodo pharetra, est eros bibendum elit, nec luctus magna felis sollicitudin mauris. Integer in mauris eu nibh euismod gravida. Duis ac tellus et risus vulputate vehicula. Donec lobortis risus a elit. Etiam tempor. Ut ullamcorper, ligula eu tempor congue, eros est euismod turpis, id tincidunt sapien risus a quam. Maecenas fermentum consequat mi. Donec fermentum. Pellentesque malesuada nulla a mi. Duis sapien sem, aliquet nec, commodo eget, consequat quis, neque. Aliquam faucibus, elit ut dictum aliquet, felis nisl adipiscing sapien, sed malesuada diam lacus eget erat. Cras mollis scelerisque nunc. Nullam arcu. Aliquam consequat. Curabitur augue lorem, dapibus quis, laoreet et, pretium ac, nisi. Aenean magna nisl, mollis quis, molestie eu, feugiat in, orci. In hac habitasse platea dictumst.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__init__.py
new file mode 100644
index 00000000..0fabd0c3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__init__.py
@@ -0,0 +1,624 @@
+import re
+import itertools
+import textwrap
+import functools
+
+try:
+    from importlib.resources import files  # type: ignore
+except ImportError:  # pragma: nocover
+    from importlib_resources import files  # type: ignore
+
+from jaraco.functools import compose, method_cache
+from jaraco.context import ExceptionTrap
+
+
+def substitution(old, new):
+    """
+    Return a function that will perform a substitution on a string
+    """
+    return lambda s: s.replace(old, new)
+
+
+def multi_substitution(*substitutions):
+    """
+    Take a sequence of pairs specifying substitutions, and create
+    a function that performs those substitutions.
+
+    >>> multi_substitution(('foo', 'bar'), ('bar', 'baz'))('foo')
+    'baz'
+    """
+    substitutions = itertools.starmap(substitution, substitutions)
+    # compose function applies last function first, so reverse the
+    #  substitutions to get the expected order.
+    substitutions = reversed(tuple(substitutions))
+    return compose(*substitutions)
+
+
+class FoldedCase(str):
+    """
+    A case insensitive string class; behaves just like str
+    except compares equal when the only variation is case.
+
+    >>> s = FoldedCase('hello world')
+
+    >>> s == 'Hello World'
+    True
+
+    >>> 'Hello World' == s
+    True
+
+    >>> s != 'Hello World'
+    False
+
+    >>> s.index('O')
+    4
+
+    >>> s.split('O')
+    ['hell', ' w', 'rld']
+
+    >>> sorted(map(FoldedCase, ['GAMMA', 'alpha', 'Beta']))
+    ['alpha', 'Beta', 'GAMMA']
+
+    Sequence membership is straightforward.
+
+    >>> "Hello World" in [s]
+    True
+    >>> s in ["Hello World"]
+    True
+
+    Allows testing for set inclusion, but candidate and elements
+    must both be folded.
+
+    >>> FoldedCase("Hello World") in {s}
+    True
+    >>> s in {FoldedCase("Hello World")}
+    True
+
+    String inclusion works as long as the FoldedCase object
+    is on the right.
+
+    >>> "hello" in FoldedCase("Hello World")
+    True
+
+    But not if the FoldedCase object is on the left:
+
+    >>> FoldedCase('hello') in 'Hello World'
+    False
+
+    In that case, use ``in_``:
+
+    >>> FoldedCase('hello').in_('Hello World')
+    True
+
+    >>> FoldedCase('hello') > FoldedCase('Hello')
+    False
+
+    >>> FoldedCase('ß') == FoldedCase('ss')
+    True
+    """
+
+    def __lt__(self, other):
+        return self.casefold() < other.casefold()
+
+    def __gt__(self, other):
+        return self.casefold() > other.casefold()
+
+    def __eq__(self, other):
+        return self.casefold() == other.casefold()
+
+    def __ne__(self, other):
+        return self.casefold() != other.casefold()
+
+    def __hash__(self):
+        return hash(self.casefold())
+
+    def __contains__(self, other):
+        return super().casefold().__contains__(other.casefold())
+
+    def in_(self, other):
+        "Does self appear in other?"
+        return self in FoldedCase(other)
+
+    # cache casefold since it's likely to be called frequently.
+    @method_cache
+    def casefold(self):
+        return super().casefold()
+
+    def index(self, sub):
+        return self.casefold().index(sub.casefold())
+
+    def split(self, splitter=' ', maxsplit=0):
+        pattern = re.compile(re.escape(splitter), re.I)
+        return pattern.split(self, maxsplit)
+
+
+# Python 3.8 compatibility
+_unicode_trap = ExceptionTrap(UnicodeDecodeError)
+
+
+@_unicode_trap.passes
+def is_decodable(value):
+    r"""
+    Return True if the supplied value is decodable (using the default
+    encoding).
+
+    >>> is_decodable(b'\xff')
+    False
+    >>> is_decodable(b'\x32')
+    True
+    """
+    value.decode()
+
+
+def is_binary(value):
+    r"""
+    Return True if the value appears to be binary (that is, it's a byte
+    string and isn't decodable).
+
+    >>> is_binary(b'\xff')
+    True
+    >>> is_binary('\xff')
+    False
+    """
+    return isinstance(value, bytes) and not is_decodable(value)
+
+
+def trim(s):
+    r"""
+    Trim something like a docstring to remove the whitespace that
+    is common due to indentation and formatting.
+
+    >>> trim("\n\tfoo = bar\n\t\tbar = baz\n")
+    'foo = bar\n\tbar = baz'
+    """
+    return textwrap.dedent(s).strip()
+
+
+def wrap(s):
+    """
+    Wrap lines of text, retaining existing newlines as
+    paragraph markers.
+
+    >>> print(wrap(lorem_ipsum))
+    Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do
+    eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad
+    minim veniam, quis nostrud exercitation ullamco laboris nisi ut
+    aliquip ex ea commodo consequat. Duis aute irure dolor in
+    reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla
+    pariatur. Excepteur sint occaecat cupidatat non proident, sunt in
+    culpa qui officia deserunt mollit anim id est laborum.
+    <BLANKLINE>
+    Curabitur pretium tincidunt lacus. Nulla gravida orci a odio. Nullam
+    varius, turpis et commodo pharetra, est eros bibendum elit, nec luctus
+    magna felis sollicitudin mauris. Integer in mauris eu nibh euismod
+    gravida. Duis ac tellus et risus vulputate vehicula. Donec lobortis
+    risus a elit. Etiam tempor. Ut ullamcorper, ligula eu tempor congue,
+    eros est euismod turpis, id tincidunt sapien risus a quam. Maecenas
+    fermentum consequat mi. Donec fermentum. Pellentesque malesuada nulla
+    a mi. Duis sapien sem, aliquet nec, commodo eget, consequat quis,
+    neque. Aliquam faucibus, elit ut dictum aliquet, felis nisl adipiscing
+    sapien, sed malesuada diam lacus eget erat. Cras mollis scelerisque
+    nunc. Nullam arcu. Aliquam consequat. Curabitur augue lorem, dapibus
+    quis, laoreet et, pretium ac, nisi. Aenean magna nisl, mollis quis,
+    molestie eu, feugiat in, orci. In hac habitasse platea dictumst.
+    """
+    paragraphs = s.splitlines()
+    wrapped = ('\n'.join(textwrap.wrap(para)) for para in paragraphs)
+    return '\n\n'.join(wrapped)
+
+
+def unwrap(s):
+    r"""
+    Given a multi-line string, return an unwrapped version.
+
+    >>> wrapped = wrap(lorem_ipsum)
+    >>> wrapped.count('\n')
+    20
+    >>> unwrapped = unwrap(wrapped)
+    >>> unwrapped.count('\n')
+    1
+    >>> print(unwrapped)
+    Lorem ipsum dolor sit amet, consectetur adipiscing ...
+    Curabitur pretium tincidunt lacus. Nulla gravida orci ...
+
+    """
+    paragraphs = re.split(r'\n\n+', s)
+    cleaned = (para.replace('\n', ' ') for para in paragraphs)
+    return '\n'.join(cleaned)
+
+
+lorem_ipsum: str = (
+    files(__name__).joinpath('Lorem ipsum.txt').read_text(encoding='utf-8')
+)
+
+
+class Splitter:
+    """object that will split a string with the given arguments for each call
+
+    >>> s = Splitter(',')
+    >>> s('hello, world, this is your, master calling')
+    ['hello', ' world', ' this is your', ' master calling']
+    """
+
+    def __init__(self, *args):
+        self.args = args
+
+    def __call__(self, s):
+        return s.split(*self.args)
+
+
+def indent(string, prefix=' ' * 4):
+    """
+    >>> indent('foo')
+    '    foo'
+    """
+    return prefix + string
+
+
+class WordSet(tuple):
+    """
+    Given an identifier, return the words that identifier represents,
+    whether in camel case, underscore-separated, etc.
+
+    >>> WordSet.parse("camelCase")
+    ('camel', 'Case')
+
+    >>> WordSet.parse("under_sep")
+    ('under', 'sep')
+
+    Acronyms should be retained
+
+    >>> WordSet.parse("firstSNL")
+    ('first', 'SNL')
+
+    >>> WordSet.parse("you_and_I")
+    ('you', 'and', 'I')
+
+    >>> WordSet.parse("A simple test")
+    ('A', 'simple', 'test')
+
+    Multiple caps should not interfere with the first cap of another word.
+
+    >>> WordSet.parse("myABCClass")
+    ('my', 'ABC', 'Class')
+
+    The result is a WordSet, providing access to other forms.
+
+    >>> WordSet.parse("myABCClass").underscore_separated()
+    'my_ABC_Class'
+
+    >>> WordSet.parse('a-command').camel_case()
+    'ACommand'
+
+    >>> WordSet.parse('someIdentifier').lowered().space_separated()
+    'some identifier'
+
+    Slices of the result should return another WordSet.
+
+    >>> WordSet.parse('taken-out-of-context')[1:].underscore_separated()
+    'out_of_context'
+
+    >>> WordSet.from_class_name(WordSet()).lowered().space_separated()
+    'word set'
+
+    >>> example = WordSet.parse('figured it out')
+    >>> example.headless_camel_case()
+    'figuredItOut'
+    >>> example.dash_separated()
+    'figured-it-out'
+
+    """
+
+    _pattern = re.compile('([A-Z]?[a-z]+)|([A-Z]+(?![a-z]))')
+
+    def capitalized(self):
+        return WordSet(word.capitalize() for word in self)
+
+    def lowered(self):
+        return WordSet(word.lower() for word in self)
+
+    def camel_case(self):
+        return ''.join(self.capitalized())
+
+    def headless_camel_case(self):
+        words = iter(self)
+        first = next(words).lower()
+        new_words = itertools.chain((first,), WordSet(words).camel_case())
+        return ''.join(new_words)
+
+    def underscore_separated(self):
+        return '_'.join(self)
+
+    def dash_separated(self):
+        return '-'.join(self)
+
+    def space_separated(self):
+        return ' '.join(self)
+
+    def trim_right(self, item):
+        """
+        Remove the item from the end of the set.
+
+        >>> WordSet.parse('foo bar').trim_right('foo')
+        ('foo', 'bar')
+        >>> WordSet.parse('foo bar').trim_right('bar')
+        ('foo',)
+        >>> WordSet.parse('').trim_right('bar')
+        ()
+        """
+        return self[:-1] if self and self[-1] == item else self
+
+    def trim_left(self, item):
+        """
+        Remove the item from the beginning of the set.
+
+        >>> WordSet.parse('foo bar').trim_left('foo')
+        ('bar',)
+        >>> WordSet.parse('foo bar').trim_left('bar')
+        ('foo', 'bar')
+        >>> WordSet.parse('').trim_left('bar')
+        ()
+        """
+        return self[1:] if self and self[0] == item else self
+
+    def trim(self, item):
+        """
+        >>> WordSet.parse('foo bar').trim('foo')
+        ('bar',)
+        """
+        return self.trim_left(item).trim_right(item)
+
+    def __getitem__(self, item):
+        result = super().__getitem__(item)
+        if isinstance(item, slice):
+            result = WordSet(result)
+        return result
+
+    @classmethod
+    def parse(cls, identifier):
+        matches = cls._pattern.finditer(identifier)
+        return WordSet(match.group(0) for match in matches)
+
+    @classmethod
+    def from_class_name(cls, subject):
+        return cls.parse(subject.__class__.__name__)
+
+
+# for backward compatibility
+words = WordSet.parse
+
+
+def simple_html_strip(s):
+    r"""
+    Remove HTML from the string `s`.
+
+    >>> str(simple_html_strip(''))
+    ''
+
+    >>> print(simple_html_strip('A <bold>stormy</bold> day in paradise'))
+    A stormy day in paradise
+
+    >>> print(simple_html_strip('Somebody <!-- do not --> tell the truth.'))
+    Somebody  tell the truth.
+
+    >>> print(simple_html_strip('What about<br/>\nmultiple lines?'))
+    What about
+    multiple lines?
+    """
+    html_stripper = re.compile('(<!--.*?-->)|(<[^>]*>)|([^<]+)', re.DOTALL)
+    texts = (match.group(3) or '' for match in html_stripper.finditer(s))
+    return ''.join(texts)
+
+
+class SeparatedValues(str):
+    """
+    A string separated by a separator. Overrides __iter__ for getting
+    the values.
+
+    >>> list(SeparatedValues('a,b,c'))
+    ['a', 'b', 'c']
+
+    Whitespace is stripped and empty values are discarded.
+
+    >>> list(SeparatedValues(' a,   b   , c,  '))
+    ['a', 'b', 'c']
+    """
+
+    separator = ','
+
+    def __iter__(self):
+        parts = self.split(self.separator)
+        return filter(None, (part.strip() for part in parts))
+
+
+class Stripper:
+    r"""
+    Given a series of lines, find the common prefix and strip it from them.
+
+    >>> lines = [
+    ...     'abcdefg\n',
+    ...     'abc\n',
+    ...     'abcde\n',
+    ... ]
+    >>> res = Stripper.strip_prefix(lines)
+    >>> res.prefix
+    'abc'
+    >>> list(res.lines)
+    ['defg\n', '\n', 'de\n']
+
+    If no prefix is common, nothing should be stripped.
+
+    >>> lines = [
+    ...     'abcd\n',
+    ...     '1234\n',
+    ... ]
+    >>> res = Stripper.strip_prefix(lines)
+    >>> res.prefix = ''
+    >>> list(res.lines)
+    ['abcd\n', '1234\n']
+    """
+
+    def __init__(self, prefix, lines):
+        self.prefix = prefix
+        self.lines = map(self, lines)
+
+    @classmethod
+    def strip_prefix(cls, lines):
+        prefix_lines, lines = itertools.tee(lines)
+        prefix = functools.reduce(cls.common_prefix, prefix_lines)
+        return cls(prefix, lines)
+
+    def __call__(self, line):
+        if not self.prefix:
+            return line
+        null, prefix, rest = line.partition(self.prefix)
+        return rest
+
+    @staticmethod
+    def common_prefix(s1, s2):
+        """
+        Return the common prefix of two lines.
+        """
+        index = min(len(s1), len(s2))
+        while s1[:index] != s2[:index]:
+            index -= 1
+        return s1[:index]
+
+
+def remove_prefix(text, prefix):
+    """
+    Remove the prefix from the text if it exists.
+
+    >>> remove_prefix('underwhelming performance', 'underwhelming ')
+    'performance'
+
+    >>> remove_prefix('something special', 'sample')
+    'something special'
+    """
+    null, prefix, rest = text.rpartition(prefix)
+    return rest
+
+
+def remove_suffix(text, suffix):
+    """
+    Remove the suffix from the text if it exists.
+
+    >>> remove_suffix('name.git', '.git')
+    'name'
+
+    >>> remove_suffix('something special', 'sample')
+    'something special'
+    """
+    rest, suffix, null = text.partition(suffix)
+    return rest
+
+
+def normalize_newlines(text):
+    r"""
+    Replace alternate newlines with the canonical newline.
+
+    >>> normalize_newlines('Lorem Ipsum\u2029')
+    'Lorem Ipsum\n'
+    >>> normalize_newlines('Lorem Ipsum\r\n')
+    'Lorem Ipsum\n'
+    >>> normalize_newlines('Lorem Ipsum\x85')
+    'Lorem Ipsum\n'
+    """
+    newlines = ['\r\n', '\r', '\n', '\u0085', '\u2028', '\u2029']
+    pattern = '|'.join(newlines)
+    return re.sub(pattern, '\n', text)
+
+
+def _nonblank(str):
+    return str and not str.startswith('#')
+
+
+@functools.singledispatch
+def yield_lines(iterable):
+    r"""
+    Yield valid lines of a string or iterable.
+
+    >>> list(yield_lines(''))
+    []
+    >>> list(yield_lines(['foo', 'bar']))
+    ['foo', 'bar']
+    >>> list(yield_lines('foo\nbar'))
+    ['foo', 'bar']
+    >>> list(yield_lines('\nfoo\n#bar\nbaz #comment'))
+    ['foo', 'baz #comment']
+    >>> list(yield_lines(['foo\nbar', 'baz', 'bing\n\n\n']))
+    ['foo', 'bar', 'baz', 'bing']
+    """
+    return itertools.chain.from_iterable(map(yield_lines, iterable))
+
+
+@yield_lines.register(str)
+def _(text):
+    return filter(_nonblank, map(str.strip, text.splitlines()))
+
+
+def drop_comment(line):
+    """
+    Drop comments.
+
+    >>> drop_comment('foo # bar')
+    'foo'
+
+    A hash without a space may be in a URL.
+
+    >>> drop_comment('http://example.com/foo#bar')
+    'http://example.com/foo#bar'
+    """
+    return line.partition(' #')[0]
+
+
+def join_continuation(lines):
+    r"""
+    Join lines continued by a trailing backslash.
+
+    >>> list(join_continuation(['foo \\', 'bar', 'baz']))
+    ['foobar', 'baz']
+    >>> list(join_continuation(['foo \\', 'bar', 'baz']))
+    ['foobar', 'baz']
+    >>> list(join_continuation(['foo \\', 'bar \\', 'baz']))
+    ['foobarbaz']
+
+    Not sure why, but...
+    The character preceding the backslash is also elided.
+
+    >>> list(join_continuation(['goo\\', 'dly']))
+    ['godly']
+
+    A terrible idea, but...
+    If no line is available to continue, suppress the lines.
+
+    >>> list(join_continuation(['foo', 'bar\\', 'baz\\']))
+    ['foo']
+    """
+    lines = iter(lines)
+    for item in lines:
+        while item.endswith('\\'):
+            try:
+                item = item[:-2].strip() + next(lines)
+            except StopIteration:
+                return
+        yield item
+
+
+def read_newlines(filename, limit=1024):
+    r"""
+    >>> tmp_path = getfixture('tmp_path')
+    >>> filename = tmp_path / 'out.txt'
+    >>> _ = filename.write_text('foo\n', newline='', encoding='utf-8')
+    >>> read_newlines(filename)
+    '\n'
+    >>> _ = filename.write_text('foo\r\n', newline='', encoding='utf-8')
+    >>> read_newlines(filename)
+    '\r\n'
+    >>> _ = filename.write_text('foo\r\nbar\nbing\r', newline='', encoding='utf-8')
+    >>> read_newlines(filename)
+    ('\r', '\n', '\r\n')
+    """
+    with open(filename, encoding='utf-8') as fp:
+        fp.read(limit)
+    return fp.newlines
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7b7a27a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/layouts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/layouts.cpython-312.pyc
new file mode 100644
index 00000000..c0c0e135
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/layouts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/show-newlines.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/show-newlines.cpython-312.pyc
new file mode 100644
index 00000000..e10f13d6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/show-newlines.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/strip-prefix.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/strip-prefix.cpython-312.pyc
new file mode 100644
index 00000000..99758078
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/strip-prefix.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-dvorak.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-dvorak.cpython-312.pyc
new file mode 100644
index 00000000..1a8e2b34
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-dvorak.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-qwerty.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-qwerty.cpython-312.pyc
new file mode 100644
index 00000000..4ce5a774
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/__pycache__/to-qwerty.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/layouts.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/layouts.py
new file mode 100644
index 00000000..9636f0f7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/layouts.py
@@ -0,0 +1,25 @@
+qwerty = "-=qwertyuiop[]asdfghjkl;'zxcvbnm,./_+QWERTYUIOP{}ASDFGHJKL:\"ZXCVBNM<>?"
+dvorak = "[]',.pyfgcrl/=aoeuidhtns-;qjkxbmwvz{}\"<>PYFGCRL?+AOEUIDHTNS_:QJKXBMWVZ"
+
+
+to_dvorak = str.maketrans(qwerty, dvorak)
+to_qwerty = str.maketrans(dvorak, qwerty)
+
+
+def translate(input, translation):
+    """
+    >>> translate('dvorak', to_dvorak)
+    'ekrpat'
+    >>> translate('qwerty', to_qwerty)
+    'x,dokt'
+    """
+    return input.translate(translation)
+
+
+def _translate_stream(stream, translation):
+    """
+    >>> import io
+    >>> _translate_stream(io.StringIO('foo'), to_dvorak)
+    urr
+    """
+    print(translate(stream.read(), translation))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/show-newlines.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/show-newlines.py
new file mode 100644
index 00000000..e11d1ba4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/show-newlines.py
@@ -0,0 +1,33 @@
+import autocommand
+import inflect
+
+from more_itertools import always_iterable
+
+import jaraco.text
+
+
+def report_newlines(filename):
+    r"""
+    Report the newlines in the indicated file.
+
+    >>> tmp_path = getfixture('tmp_path')
+    >>> filename = tmp_path / 'out.txt'
+    >>> _ = filename.write_text('foo\nbar\n', newline='', encoding='utf-8')
+    >>> report_newlines(filename)
+    newline is '\n'
+    >>> filename = tmp_path / 'out.txt'
+    >>> _ = filename.write_text('foo\nbar\r\n', newline='', encoding='utf-8')
+    >>> report_newlines(filename)
+    newlines are ('\n', '\r\n')
+    """
+    newlines = jaraco.text.read_newlines(filename)
+    count = len(tuple(always_iterable(newlines)))
+    engine = inflect.engine()
+    print(
+        engine.plural_noun("newline", count),
+        engine.plural_verb("is", count),
+        repr(newlines),
+    )
+
+
+autocommand.autocommand(__name__)(report_newlines)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/strip-prefix.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/strip-prefix.py
new file mode 100644
index 00000000..761717a9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/strip-prefix.py
@@ -0,0 +1,21 @@
+import sys
+
+import autocommand
+
+from jaraco.text import Stripper
+
+
+def strip_prefix():
+    r"""
+    Strip any common prefix from stdin.
+
+    >>> import io, pytest
+    >>> getfixture('monkeypatch').setattr('sys.stdin', io.StringIO('abcdef\nabc123'))
+    >>> strip_prefix()
+    def
+    123
+    """
+    sys.stdout.writelines(Stripper.strip_prefix(sys.stdin).lines)
+
+
+autocommand.autocommand(__name__)(strip_prefix)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-dvorak.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-dvorak.py
new file mode 100644
index 00000000..a6d5da80
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-dvorak.py
@@ -0,0 +1,6 @@
+import sys
+
+from . import layouts
+
+
+__name__ == '__main__' and layouts._translate_stream(sys.stdin, layouts.to_dvorak)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-qwerty.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-qwerty.py
new file mode 100644
index 00000000..abe27286
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/jaraco/text/to-qwerty.py
@@ -0,0 +1,6 @@
+import sys
+
+from . import layouts
+
+
+__name__ == '__main__' and layouts._translate_stream(sys.stdin, layouts.to_qwerty)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/LICENSE
new file mode 100644
index 00000000..0a523bec
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/LICENSE
@@ -0,0 +1,19 @@
+Copyright (c) 2012 Erik Rose
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+of the Software, and to permit persons to whom the Software is furnished to do
+so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/METADATA
new file mode 100644
index 00000000..fb41b0cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/METADATA
@@ -0,0 +1,266 @@
+Metadata-Version: 2.1
+Name: more-itertools
+Version: 10.3.0
+Summary: More routines for operating on iterables, beyond itertools
+Keywords: itertools,iterator,iteration,filter,peek,peekable,chunk,chunked
+Author-email: Erik Rose <erikrose@grinchcentral.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Natural Language :: English
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Software Development :: Libraries
+Project-URL: Homepage, https://github.com/more-itertools/more-itertools
+
+==============
+More Itertools
+==============
+
+.. image:: https://readthedocs.org/projects/more-itertools/badge/?version=latest
+  :target: https://more-itertools.readthedocs.io/en/stable/
+
+Python's ``itertools`` library is a gem - you can compose elegant solutions
+for a variety of problems with the functions it provides. In ``more-itertools``
+we collect additional building blocks, recipes, and routines for working with
+Python iterables.
+
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Grouping               | `chunked <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.chunked>`_,                                                                               |
+|                        | `ichunked <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.ichunked>`_,                                                                             |
+|                        | `chunked_even <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.chunked_even>`_,                                                                     |
+|                        | `sliced <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sliced>`_,                                                                                 |
+|                        | `constrained_batches <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.constrained_batches>`_,                                                       |
+|                        | `distribute <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.distribute>`_,                                                                         |
+|                        | `divide <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.divide>`_,                                                                                 |
+|                        | `split_at <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.split_at>`_,                                                                             |
+|                        | `split_before <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.split_before>`_,                                                                     |
+|                        | `split_after <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.split_after>`_,                                                                       |
+|                        | `split_into <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.split_into>`_,                                                                         |
+|                        | `split_when <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.split_when>`_,                                                                         |
+|                        | `bucket <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.bucket>`_,                                                                                 |
+|                        | `unzip <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unzip>`_,                                                                                   |
+|                        | `batched <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.batched>`_,                                                                               |
+|                        | `grouper <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.grouper>`_,                                                                               |
+|                        | `partition <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.partition>`_,                                                                           |
+|                        | `transpose <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.transpose>`_                                                                            |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Lookahead and lookback | `spy <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.spy>`_,                                                                                       |
+|                        | `peekable <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.peekable>`_,                                                                             |
+|                        | `seekable <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.seekable>`_                                                                              |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Windowing              | `windowed <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.windowed>`_,                                                                             |
+|                        | `substrings <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.substrings>`_,                                                                         |
+|                        | `substrings_indexes <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.substrings_indexes>`_,                                                         |
+|                        | `stagger <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.stagger>`_,                                                                               |
+|                        | `windowed_complete <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.windowed_complete>`_,                                                           |
+|                        | `pairwise <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.pairwise>`_,                                                                             |
+|                        | `triplewise <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.triplewise>`_,                                                                         |
+|                        | `sliding_window <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sliding_window>`_,                                                                 |
+|                        | `subslices <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.subslices>`_                                                                            |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Augmenting             | `count_cycle <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.count_cycle>`_,                                                                       |
+|                        | `intersperse <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.intersperse>`_,                                                                       |
+|                        | `padded <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.padded>`_,                                                                                 |
+|                        | `repeat_each <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.repeat_each>`_,                                                                       |
+|                        | `mark_ends <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.mark_ends>`_,                                                                           |
+|                        | `repeat_last <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.repeat_last>`_,                                                                       |
+|                        | `adjacent <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.adjacent>`_,                                                                             |
+|                        | `groupby_transform <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.groupby_transform>`_,                                                           |
+|                        | `pad_none <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.pad_none>`_,                                                                             |
+|                        | `ncycles <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.ncycles>`_                                                                                |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Combining              | `collapse <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.collapse>`_,                                                                             |
+|                        | `sort_together <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sort_together>`_,                                                                   |
+|                        | `interleave <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.interleave>`_,                                                                         |
+|                        | `interleave_longest <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.interleave_longest>`_,                                                         |
+|                        | `interleave_evenly <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.interleave_evenly>`_,                                                           |
+|                        | `zip_offset <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.zip_offset>`_,                                                                         |
+|                        | `zip_equal <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.zip_equal>`_,                                                                           |
+|                        | `zip_broadcast <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.zip_broadcast>`_,                                                                   |
+|                        | `flatten <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.flatten>`_,                                                                               |
+|                        | `roundrobin <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.roundrobin>`_,                                                                         |
+|                        | `prepend <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.prepend>`_,                                                                               |
+|                        | `value_chain <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.value_chain>`_,                                                                       |
+|                        | `partial_product <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.partial_product>`_                                                                |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Summarizing            | `ilen <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.ilen>`_,                                                                                     |
+|                        | `unique_to_each <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unique_to_each>`_,                                                                 |
+|                        | `sample <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sample>`_,                                                                                 |
+|                        | `consecutive_groups <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.consecutive_groups>`_,                                                         |
+|                        | `run_length <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.run_length>`_,                                                                         |
+|                        | `map_reduce <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.map_reduce>`_,                                                                         |
+|                        | `join_mappings <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.join_mappings>`_,                                                                   |
+|                        | `exactly_n <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.exactly_n>`_,                                                                           |
+|                        | `is_sorted <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.is_sorted>`_,                                                                           |
+|                        | `all_equal <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.all_equal>`_,                                                                           |
+|                        | `all_unique <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.all_unique>`_,                                                                         |
+|                        | `minmax <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.minmax>`_,                                                                                 |
+|                        | `first_true <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.first_true>`_,                                                                         |
+|                        | `quantify <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.quantify>`_,                                                                             |
+|                        | `iequals <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.iequals>`_                                                                                |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Selecting              | `islice_extended <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.islice_extended>`_,                                                               |
+|                        | `first <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.first>`_,                                                                                   |
+|                        | `last <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.last>`_,                                                                                     |
+|                        | `one <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.one>`_,                                                                                       |
+|                        | `only <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.only>`_,                                                                                     |
+|                        | `strictly_n <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.strictly_n>`_,                                                                         |
+|                        | `strip <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.strip>`_,                                                                                   |
+|                        | `lstrip <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.lstrip>`_,                                                                                 |
+|                        | `rstrip <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.rstrip>`_,                                                                                 |
+|                        | `filter_except <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.filter_except>`_,                                                                   |
+|                        | `map_except <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.map_except>`_,                                                                         |
+|                        | `filter_map <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.filter_map>`_,                                                                         |
+|                        | `iter_suppress <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.iter_suppress>`_,                                                                   |
+|                        | `nth_or_last <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth_or_last>`_,                                                                       |
+|                        | `unique_in_window <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unique_in_window>`_,                                                             |
+|                        | `before_and_after <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.before_and_after>`_,                                                             |
+|                        | `nth <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth>`_,                                                                                       |
+|                        | `take <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.take>`_,                                                                                     |
+|                        | `tail <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.tail>`_,                                                                                     |
+|                        | `unique_everseen <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unique_everseen>`_,                                                               |
+|                        | `unique_justseen <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unique_justseen>`_,                                                               |
+|                        | `unique <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.unique>`_,                                                                                 |
+|                        | `duplicates_everseen <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.duplicates_everseen>`_,                                                       |
+|                        | `duplicates_justseen <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.duplicates_justseen>`_,                                                       |
+|                        | `classify_unique <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.classify_unique>`_,                                                               |
+|                        | `longest_common_prefix <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.longest_common_prefix>`_,                                                   |
+|                        | `takewhile_inclusive <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.takewhile_inclusive>`_                                                        |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Math                   | `dft <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.dft>`_,                                                                                       |
+|                        | `idft <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.idft>`_,                                                                                     |
+|                        | `convolve <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.convolve>`_,                                                                             |
+|                        | `dotproduct <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.dotproduct>`_,                                                                         |
+|                        | `factor <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.factor>`_,                                                                                 |
+|                        | `matmul <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.matmul>`_,                                                                                 |
+|                        | `polynomial_from_roots <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.polynomial_from_roots>`_,                                                   |
+|                        | `polynomial_derivative <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.polynomial_derivative>`_,                                                   |
+|                        | `polynomial_eval <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.polynomial_eval>`_,                                                               |
+|                        | `sieve <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sieve>`_,                                                                                   |
+|                        | `sum_of_squares <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.sum_of_squares>`_,                                                                 |
+|                        | `totient <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.totient>`_                                                                                |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Combinatorics          | `distinct_permutations <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.distinct_permutations>`_,                                                   |
+|                        | `distinct_combinations <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.distinct_combinations>`_,                                                   |
+|                        | `circular_shifts <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.circular_shifts>`_,                                                               |
+|                        | `partitions <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.partitions>`_,                                                                         |
+|                        | `set_partitions <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.set_partitions>`_,                                                                 |
+|                        | `product_index <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.product_index>`_,                                                                   |
+|                        | `combination_index <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.combination_index>`_,                                                           |
+|                        | `permutation_index <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.permutation_index>`_,                                                           |
+|                        | `combination_with_replacement_index <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.combination_with_replacement_index>`_,                         |
+|                        | `gray_product  <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.gray_product>`_,                                                                    |
+|                        | `outer_product  <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.outer_product>`_,                                                                  |
+|                        | `powerset <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.powerset>`_,                                                                             |
+|                        | `powerset_of_sets <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.powerset_of_sets>`_,                                                             |
+|                        | `random_product <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.random_product>`_,                                                                 |
+|                        | `random_permutation <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.random_permutation>`_,                                                         |
+|                        | `random_combination <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.random_combination>`_,                                                         |
+|                        | `random_combination_with_replacement <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.random_combination_with_replacement>`_,                       |
+|                        | `nth_product <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth_product>`_,                                                                       |
+|                        | `nth_permutation <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth_permutation>`_,                                                               |
+|                        | `nth_combination <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth_combination>`_,                                                               |
+|                        | `nth_combination_with_replacement <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.nth_combination_with_replacement>`_                              |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Wrapping               | `always_iterable <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.always_iterable>`_,                                                               |
+|                        | `always_reversible <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.always_reversible>`_,                                                           |
+|                        | `countable <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.countable>`_,                                                                           |
+|                        | `consumer <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.consumer>`_,                                                                             |
+|                        | `with_iter <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.with_iter>`_,                                                                           |
+|                        | `iter_except <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.iter_except>`_                                                                        |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+| Others                 | `locate <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.locate>`_,                                                                                 |
+|                        | `rlocate <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.rlocate>`_,                                                                               |
+|                        | `replace <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.replace>`_,                                                                               |
+|                        | `numeric_range <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.numeric_range>`_,                                                                   |
+|                        | `side_effect <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.side_effect>`_,                                                                       |
+|                        | `iterate <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.iterate>`_,                                                                               |
+|                        | `difference <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.difference>`_,                                                                         |
+|                        | `make_decorator <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.make_decorator>`_,                                                                 |
+|                        | `SequenceView <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.SequenceView>`_,                                                                     |
+|                        | `time_limited <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.time_limited>`_,                                                                     |
+|                        | `map_if <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.map_if>`_,                                                                                 |
+|                        | `iter_index <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.iter_index>`_,                                                                         |
+|                        | `consume <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.consume>`_,                                                                               |
+|                        | `tabulate <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.tabulate>`_,                                                                             |
+|                        | `repeatfunc <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.repeatfunc>`_,                                                                         |
+|                        | `reshape <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.reshape>`_                                                                                |
+|                        | `doublestarmap <https://more-itertools.readthedocs.io/en/stable/api.html#more_itertools.doublestarmap>`_                                                                    |
++------------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
+
+
+Getting started
+===============
+
+To get started, install the library with `pip <https://pip.pypa.io/en/stable/>`_:
+
+.. code-block:: shell
+
+    pip install more-itertools
+
+The recipes from the `itertools docs <https://docs.python.org/3/library/itertools.html#itertools-recipes>`_
+are included in the top-level package:
+
+.. code-block:: python
+
+    >>> from more_itertools import flatten
+    >>> iterable = [(0, 1), (2, 3)]
+    >>> list(flatten(iterable))
+    [0, 1, 2, 3]
+
+Several new recipes are available as well:
+
+.. code-block:: python
+
+    >>> from more_itertools import chunked
+    >>> iterable = [0, 1, 2, 3, 4, 5, 6, 7, 8]
+    >>> list(chunked(iterable, 3))
+    [[0, 1, 2], [3, 4, 5], [6, 7, 8]]
+
+    >>> from more_itertools import spy
+    >>> iterable = (x * x for x in range(1, 6))
+    >>> head, iterable = spy(iterable, n=3)
+    >>> list(head)
+    [1, 4, 9]
+    >>> list(iterable)
+    [1, 4, 9, 16, 25]
+
+
+
+For the full listing of functions, see the `API documentation <https://more-itertools.readthedocs.io/en/stable/api.html>`_.
+
+
+Links elsewhere
+===============
+
+Blog posts about ``more-itertools``:
+
+* `Yo, I heard you like decorators <https://www.bbayles.com/index/decorator_factory>`__
+* `Tour of Python Itertools <https://martinheinz.dev/blog/16>`__ (`Alternate <https://dev.to/martinheinz/tour-of-python-itertools-4122>`__)
+* `Real-World Python More Itertools <https://www.gidware.com/real-world-more-itertools/>`_
+
+
+Development
+===========
+
+``more-itertools`` is maintained by `@erikrose <https://github.com/erikrose>`_
+and `@bbayles <https://github.com/bbayles>`_, with help from `many others <https://github.com/more-itertools/more-itertools/graphs/contributors>`_.
+If you have a problem or suggestion, please file a bug or pull request in this
+repository. Thanks for contributing!
+
+
+Version History
+===============
+
+The version history can be found in `documentation <https://more-itertools.readthedocs.io/en/stable/versions.html>`_.
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/RECORD
new file mode 100644
index 00000000..f15f3fcd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/RECORD
@@ -0,0 +1,16 @@
+more_itertools-10.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+more_itertools-10.3.0.dist-info/LICENSE,sha256=CfHIyelBrz5YTVlkHqm4fYPAyw_QB-te85Gn4mQ8GkY,1053
+more_itertools-10.3.0.dist-info/METADATA,sha256=BFO90O-fLNiVQMpj7oIS5ztzgJUUQZ3TA32P5HH3N-A,36293
+more_itertools-10.3.0.dist-info/RECORD,,
+more_itertools-10.3.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+more_itertools-10.3.0.dist-info/WHEEL,sha256=rSgq_JpHF9fHR1lx53qwg_1-2LypZE_qmcuXbVUq948,81
+more_itertools/__init__.py,sha256=dtAbGjTDmn_ghiU5YXfhyDy0phAlXVdt5klZA5fUa-Q,149
+more_itertools/__init__.pyi,sha256=5B3eTzON1BBuOLob1vCflyEb2lSd6usXQQ-Cv-hXkeA,43
+more_itertools/__pycache__/__init__.cpython-312.pyc,,
+more_itertools/__pycache__/more.cpython-312.pyc,,
+more_itertools/__pycache__/recipes.cpython-312.pyc,,
+more_itertools/more.py,sha256=1E5kzFncRKTDw0cYv1yRXMgDdunstLQd1QStcnL6U90,148370
+more_itertools/more.pyi,sha256=iXXeqt48Nxe8VGmIWpkVXuKpR2FYNuu2DU8nQLWCCu0,21484
+more_itertools/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+more_itertools/recipes.py,sha256=WedhhfhGVgr6zii8fIbGJVmRTw0ZKRiLKnYBDGJv4nY,28591
+more_itertools/recipes.pyi,sha256=T_mdGpcFdfrP3JSWbwzYP9JyNV-Go-7RPfpxfftAWlA,4617
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/WHEEL
new file mode 100644
index 00000000..db4a255f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools-10.3.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.8.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.py
new file mode 100644
index 00000000..9c4662fc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.py
@@ -0,0 +1,6 @@
+"""More routines for operating on iterables, beyond itertools"""
+
+from .more import *  # noqa
+from .recipes import *  # noqa
+
+__version__ = '10.3.0'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.pyi b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.pyi
new file mode 100644
index 00000000..96f6e36c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__init__.pyi
@@ -0,0 +1,2 @@
+from .more import *
+from .recipes import *
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..80c843be
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/more.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/more.cpython-312.pyc
new file mode 100644
index 00000000..ab923d26
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/more.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/recipes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/recipes.cpython-312.pyc
new file mode 100644
index 00000000..46cb88df
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/__pycache__/recipes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.py
new file mode 100644
index 00000000..7b481907
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.py
@@ -0,0 +1,4806 @@
+import math
+import warnings
+
+from collections import Counter, defaultdict, deque, abc
+from collections.abc import Sequence
+from functools import cached_property, partial, reduce, wraps
+from heapq import heapify, heapreplace, heappop
+from itertools import (
+    chain,
+    combinations,
+    compress,
+    count,
+    cycle,
+    dropwhile,
+    groupby,
+    islice,
+    repeat,
+    starmap,
+    takewhile,
+    tee,
+    zip_longest,
+    product,
+)
+from math import comb, e, exp, factorial, floor, fsum, log, perm, tau
+from queue import Empty, Queue
+from random import random, randrange, uniform
+from operator import itemgetter, mul, sub, gt, lt, ge, le
+from sys import hexversion, maxsize
+from time import monotonic
+
+from .recipes import (
+    _marker,
+    _zip_equal,
+    UnequalIterablesError,
+    consume,
+    flatten,
+    pairwise,
+    powerset,
+    take,
+    unique_everseen,
+    all_equal,
+    batched,
+)
+
+__all__ = [
+    'AbortThread',
+    'SequenceView',
+    'UnequalIterablesError',
+    'adjacent',
+    'all_unique',
+    'always_iterable',
+    'always_reversible',
+    'bucket',
+    'callback_iter',
+    'chunked',
+    'chunked_even',
+    'circular_shifts',
+    'collapse',
+    'combination_index',
+    'combination_with_replacement_index',
+    'consecutive_groups',
+    'constrained_batches',
+    'consumer',
+    'count_cycle',
+    'countable',
+    'dft',
+    'difference',
+    'distinct_combinations',
+    'distinct_permutations',
+    'distribute',
+    'divide',
+    'doublestarmap',
+    'duplicates_everseen',
+    'duplicates_justseen',
+    'classify_unique',
+    'exactly_n',
+    'filter_except',
+    'filter_map',
+    'first',
+    'gray_product',
+    'groupby_transform',
+    'ichunked',
+    'iequals',
+    'idft',
+    'ilen',
+    'interleave',
+    'interleave_evenly',
+    'interleave_longest',
+    'intersperse',
+    'is_sorted',
+    'islice_extended',
+    'iterate',
+    'iter_suppress',
+    'join_mappings',
+    'last',
+    'locate',
+    'longest_common_prefix',
+    'lstrip',
+    'make_decorator',
+    'map_except',
+    'map_if',
+    'map_reduce',
+    'mark_ends',
+    'minmax',
+    'nth_or_last',
+    'nth_permutation',
+    'nth_product',
+    'nth_combination_with_replacement',
+    'numeric_range',
+    'one',
+    'only',
+    'outer_product',
+    'padded',
+    'partial_product',
+    'partitions',
+    'peekable',
+    'permutation_index',
+    'powerset_of_sets',
+    'product_index',
+    'raise_',
+    'repeat_each',
+    'repeat_last',
+    'replace',
+    'rlocate',
+    'rstrip',
+    'run_length',
+    'sample',
+    'seekable',
+    'set_partitions',
+    'side_effect',
+    'sliced',
+    'sort_together',
+    'split_after',
+    'split_at',
+    'split_before',
+    'split_into',
+    'split_when',
+    'spy',
+    'stagger',
+    'strip',
+    'strictly_n',
+    'substrings',
+    'substrings_indexes',
+    'takewhile_inclusive',
+    'time_limited',
+    'unique_in_window',
+    'unique_to_each',
+    'unzip',
+    'value_chain',
+    'windowed',
+    'windowed_complete',
+    'with_iter',
+    'zip_broadcast',
+    'zip_equal',
+    'zip_offset',
+]
+
+# math.sumprod is available for Python 3.12+
+_fsumprod = getattr(math, 'sumprod', lambda x, y: fsum(map(mul, x, y)))
+
+
+def chunked(iterable, n, strict=False):
+    """Break *iterable* into lists of length *n*:
+
+        >>> list(chunked([1, 2, 3, 4, 5, 6], 3))
+        [[1, 2, 3], [4, 5, 6]]
+
+    By the default, the last yielded list will have fewer than *n* elements
+    if the length of *iterable* is not divisible by *n*:
+
+        >>> list(chunked([1, 2, 3, 4, 5, 6, 7, 8], 3))
+        [[1, 2, 3], [4, 5, 6], [7, 8]]
+
+    To use a fill-in value instead, see the :func:`grouper` recipe.
+
+    If the length of *iterable* is not divisible by *n* and *strict* is
+    ``True``, then ``ValueError`` will be raised before the last
+    list is yielded.
+
+    """
+    iterator = iter(partial(take, n, iter(iterable)), [])
+    if strict:
+        if n is None:
+            raise ValueError('n must not be None when using strict mode.')
+
+        def ret():
+            for chunk in iterator:
+                if len(chunk) != n:
+                    raise ValueError('iterable is not divisible by n.')
+                yield chunk
+
+        return iter(ret())
+    else:
+        return iterator
+
+
+def first(iterable, default=_marker):
+    """Return the first item of *iterable*, or *default* if *iterable* is
+    empty.
+
+        >>> first([0, 1, 2, 3])
+        0
+        >>> first([], 'some default')
+        'some default'
+
+    If *default* is not provided and there are no items in the iterable,
+    raise ``ValueError``.
+
+    :func:`first` is useful when you have a generator of expensive-to-retrieve
+    values and want any arbitrary one. It is marginally shorter than
+    ``next(iter(iterable), default)``.
+
+    """
+    for item in iterable:
+        return item
+    if default is _marker:
+        raise ValueError(
+            'first() was called on an empty iterable, and no '
+            'default value was provided.'
+        )
+    return default
+
+
+def last(iterable, default=_marker):
+    """Return the last item of *iterable*, or *default* if *iterable* is
+    empty.
+
+        >>> last([0, 1, 2, 3])
+        3
+        >>> last([], 'some default')
+        'some default'
+
+    If *default* is not provided and there are no items in the iterable,
+    raise ``ValueError``.
+    """
+    try:
+        if isinstance(iterable, Sequence):
+            return iterable[-1]
+        # Work around https://bugs.python.org/issue38525
+        elif hasattr(iterable, '__reversed__') and (hexversion != 0x030800F0):
+            return next(reversed(iterable))
+        else:
+            return deque(iterable, maxlen=1)[-1]
+    except (IndexError, TypeError, StopIteration):
+        if default is _marker:
+            raise ValueError(
+                'last() was called on an empty iterable, and no default was '
+                'provided.'
+            )
+        return default
+
+
+def nth_or_last(iterable, n, default=_marker):
+    """Return the nth or the last item of *iterable*,
+    or *default* if *iterable* is empty.
+
+        >>> nth_or_last([0, 1, 2, 3], 2)
+        2
+        >>> nth_or_last([0, 1], 2)
+        1
+        >>> nth_or_last([], 0, 'some default')
+        'some default'
+
+    If *default* is not provided and there are no items in the iterable,
+    raise ``ValueError``.
+    """
+    return last(islice(iterable, n + 1), default=default)
+
+
+class peekable:
+    """Wrap an iterator to allow lookahead and prepending elements.
+
+    Call :meth:`peek` on the result to get the value that will be returned
+    by :func:`next`. This won't advance the iterator:
+
+        >>> p = peekable(['a', 'b'])
+        >>> p.peek()
+        'a'
+        >>> next(p)
+        'a'
+
+    Pass :meth:`peek` a default value to return that instead of raising
+    ``StopIteration`` when the iterator is exhausted.
+
+        >>> p = peekable([])
+        >>> p.peek('hi')
+        'hi'
+
+    peekables also offer a :meth:`prepend` method, which "inserts" items
+    at the head of the iterable:
+
+        >>> p = peekable([1, 2, 3])
+        >>> p.prepend(10, 11, 12)
+        >>> next(p)
+        10
+        >>> p.peek()
+        11
+        >>> list(p)
+        [11, 12, 1, 2, 3]
+
+    peekables can be indexed. Index 0 is the item that will be returned by
+    :func:`next`, index 1 is the item after that, and so on:
+    The values up to the given index will be cached.
+
+        >>> p = peekable(['a', 'b', 'c', 'd'])
+        >>> p[0]
+        'a'
+        >>> p[1]
+        'b'
+        >>> next(p)
+        'a'
+
+    Negative indexes are supported, but be aware that they will cache the
+    remaining items in the source iterator, which may require significant
+    storage.
+
+    To check whether a peekable is exhausted, check its truth value:
+
+        >>> p = peekable(['a', 'b'])
+        >>> if p:  # peekable has items
+        ...     list(p)
+        ['a', 'b']
+        >>> if not p:  # peekable is exhausted
+        ...     list(p)
+        []
+
+    """
+
+    def __init__(self, iterable):
+        self._it = iter(iterable)
+        self._cache = deque()
+
+    def __iter__(self):
+        return self
+
+    def __bool__(self):
+        try:
+            self.peek()
+        except StopIteration:
+            return False
+        return True
+
+    def peek(self, default=_marker):
+        """Return the item that will be next returned from ``next()``.
+
+        Return ``default`` if there are no items left. If ``default`` is not
+        provided, raise ``StopIteration``.
+
+        """
+        if not self._cache:
+            try:
+                self._cache.append(next(self._it))
+            except StopIteration:
+                if default is _marker:
+                    raise
+                return default
+        return self._cache[0]
+
+    def prepend(self, *items):
+        """Stack up items to be the next ones returned from ``next()`` or
+        ``self.peek()``. The items will be returned in
+        first in, first out order::
+
+            >>> p = peekable([1, 2, 3])
+            >>> p.prepend(10, 11, 12)
+            >>> next(p)
+            10
+            >>> list(p)
+            [11, 12, 1, 2, 3]
+
+        It is possible, by prepending items, to "resurrect" a peekable that
+        previously raised ``StopIteration``.
+
+            >>> p = peekable([])
+            >>> next(p)
+            Traceback (most recent call last):
+              ...
+            StopIteration
+            >>> p.prepend(1)
+            >>> next(p)
+            1
+            >>> next(p)
+            Traceback (most recent call last):
+              ...
+            StopIteration
+
+        """
+        self._cache.extendleft(reversed(items))
+
+    def __next__(self):
+        if self._cache:
+            return self._cache.popleft()
+
+        return next(self._it)
+
+    def _get_slice(self, index):
+        # Normalize the slice's arguments
+        step = 1 if (index.step is None) else index.step
+        if step > 0:
+            start = 0 if (index.start is None) else index.start
+            stop = maxsize if (index.stop is None) else index.stop
+        elif step < 0:
+            start = -1 if (index.start is None) else index.start
+            stop = (-maxsize - 1) if (index.stop is None) else index.stop
+        else:
+            raise ValueError('slice step cannot be zero')
+
+        # If either the start or stop index is negative, we'll need to cache
+        # the rest of the iterable in order to slice from the right side.
+        if (start < 0) or (stop < 0):
+            self._cache.extend(self._it)
+        # Otherwise we'll need to find the rightmost index and cache to that
+        # point.
+        else:
+            n = min(max(start, stop) + 1, maxsize)
+            cache_len = len(self._cache)
+            if n >= cache_len:
+                self._cache.extend(islice(self._it, n - cache_len))
+
+        return list(self._cache)[index]
+
+    def __getitem__(self, index):
+        if isinstance(index, slice):
+            return self._get_slice(index)
+
+        cache_len = len(self._cache)
+        if index < 0:
+            self._cache.extend(self._it)
+        elif index >= cache_len:
+            self._cache.extend(islice(self._it, index + 1 - cache_len))
+
+        return self._cache[index]
+
+
+def consumer(func):
+    """Decorator that automatically advances a PEP-342-style "reverse iterator"
+    to its first yield point so you don't have to call ``next()`` on it
+    manually.
+
+        >>> @consumer
+        ... def tally():
+        ...     i = 0
+        ...     while True:
+        ...         print('Thing number %s is %s.' % (i, (yield)))
+        ...         i += 1
+        ...
+        >>> t = tally()
+        >>> t.send('red')
+        Thing number 0 is red.
+        >>> t.send('fish')
+        Thing number 1 is fish.
+
+    Without the decorator, you would have to call ``next(t)`` before
+    ``t.send()`` could be used.
+
+    """
+
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+        gen = func(*args, **kwargs)
+        next(gen)
+        return gen
+
+    return wrapper
+
+
+def ilen(iterable):
+    """Return the number of items in *iterable*.
+
+        >>> ilen(x for x in range(1000000) if x % 3 == 0)
+        333334
+
+    This consumes the iterable, so handle with care.
+
+    """
+    # This approach was selected because benchmarks showed it's likely the
+    # fastest of the known implementations at the time of writing.
+    # See GitHub tracker: #236, #230.
+    counter = count()
+    deque(zip(iterable, counter), maxlen=0)
+    return next(counter)
+
+
+def iterate(func, start):
+    """Return ``start``, ``func(start)``, ``func(func(start))``, ...
+
+    >>> from itertools import islice
+    >>> list(islice(iterate(lambda x: 2*x, 1), 10))
+    [1, 2, 4, 8, 16, 32, 64, 128, 256, 512]
+
+    """
+    while True:
+        yield start
+        try:
+            start = func(start)
+        except StopIteration:
+            break
+
+
+def with_iter(context_manager):
+    """Wrap an iterable in a ``with`` statement, so it closes once exhausted.
+
+    For example, this will close the file when the iterator is exhausted::
+
+        upper_lines = (line.upper() for line in with_iter(open('foo')))
+
+    Any context manager which returns an iterable is a candidate for
+    ``with_iter``.
+
+    """
+    with context_manager as iterable:
+        yield from iterable
+
+
+def one(iterable, too_short=None, too_long=None):
+    """Return the first item from *iterable*, which is expected to contain only
+    that item. Raise an exception if *iterable* is empty or has more than one
+    item.
+
+    :func:`one` is useful for ensuring that an iterable contains only one item.
+    For example, it can be used to retrieve the result of a database query
+    that is expected to return a single row.
+
+    If *iterable* is empty, ``ValueError`` will be raised. You may specify a
+    different exception with the *too_short* keyword:
+
+        >>> it = []
+        >>> one(it)  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        ValueError: too many items in iterable (expected 1)'
+        >>> too_short = IndexError('too few items')
+        >>> one(it, too_short=too_short)  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        IndexError: too few items
+
+    Similarly, if *iterable* contains more than one item, ``ValueError`` will
+    be raised. You may specify a different exception with the *too_long*
+    keyword:
+
+        >>> it = ['too', 'many']
+        >>> one(it)  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        ValueError: Expected exactly one item in iterable, but got 'too',
+        'many', and perhaps more.
+        >>> too_long = RuntimeError
+        >>> one(it, too_long=too_long)  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        RuntimeError
+
+    Note that :func:`one` attempts to advance *iterable* twice to ensure there
+    is only one item. See :func:`spy` or :func:`peekable` to check iterable
+    contents less destructively.
+
+    """
+    it = iter(iterable)
+
+    try:
+        first_value = next(it)
+    except StopIteration as exc:
+        raise (
+            too_short or ValueError('too few items in iterable (expected 1)')
+        ) from exc
+
+    try:
+        second_value = next(it)
+    except StopIteration:
+        pass
+    else:
+        msg = (
+            'Expected exactly one item in iterable, but got {!r}, {!r}, '
+            'and perhaps more.'.format(first_value, second_value)
+        )
+        raise too_long or ValueError(msg)
+
+    return first_value
+
+
+def raise_(exception, *args):
+    raise exception(*args)
+
+
+def strictly_n(iterable, n, too_short=None, too_long=None):
+    """Validate that *iterable* has exactly *n* items and return them if
+    it does. If it has fewer than *n* items, call function *too_short*
+    with those items. If it has more than *n* items, call function
+    *too_long* with the first ``n + 1`` items.
+
+        >>> iterable = ['a', 'b', 'c', 'd']
+        >>> n = 4
+        >>> list(strictly_n(iterable, n))
+        ['a', 'b', 'c', 'd']
+
+    Note that the returned iterable must be consumed in order for the check to
+    be made.
+
+    By default, *too_short* and *too_long* are functions that raise
+    ``ValueError``.
+
+        >>> list(strictly_n('ab', 3))  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        ValueError: too few items in iterable (got 2)
+
+        >>> list(strictly_n('abc', 2))  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        ValueError: too many items in iterable (got at least 3)
+
+    You can instead supply functions that do something else.
+    *too_short* will be called with the number of items in *iterable*.
+    *too_long* will be called with `n + 1`.
+
+        >>> def too_short(item_count):
+        ...     raise RuntimeError
+        >>> it = strictly_n('abcd', 6, too_short=too_short)
+        >>> list(it)  # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        RuntimeError
+
+        >>> def too_long(item_count):
+        ...     print('The boss is going to hear about this')
+        >>> it = strictly_n('abcdef', 4, too_long=too_long)
+        >>> list(it)
+        The boss is going to hear about this
+        ['a', 'b', 'c', 'd']
+
+    """
+    if too_short is None:
+        too_short = lambda item_count: raise_(
+            ValueError,
+            'Too few items in iterable (got {})'.format(item_count),
+        )
+
+    if too_long is None:
+        too_long = lambda item_count: raise_(
+            ValueError,
+            'Too many items in iterable (got at least {})'.format(item_count),
+        )
+
+    it = iter(iterable)
+    for i in range(n):
+        try:
+            item = next(it)
+        except StopIteration:
+            too_short(i)
+            return
+        else:
+            yield item
+
+    try:
+        next(it)
+    except StopIteration:
+        pass
+    else:
+        too_long(n + 1)
+
+
+def distinct_permutations(iterable, r=None):
+    """Yield successive distinct permutations of the elements in *iterable*.
+
+        >>> sorted(distinct_permutations([1, 0, 1]))
+        [(0, 1, 1), (1, 0, 1), (1, 1, 0)]
+
+    Equivalent to ``set(permutations(iterable))``, except duplicates are not
+    generated and thrown away. For larger input sequences this is much more
+    efficient.
+
+    Duplicate permutations arise when there are duplicated elements in the
+    input iterable. The number of items returned is
+    `n! / (x_1! * x_2! * ... * x_n!)`, where `n` is the total number of
+    items input, and each `x_i` is the count of a distinct item in the input
+    sequence.
+
+    If *r* is given, only the *r*-length permutations are yielded.
+
+        >>> sorted(distinct_permutations([1, 0, 1], r=2))
+        [(0, 1), (1, 0), (1, 1)]
+        >>> sorted(distinct_permutations(range(3), r=2))
+        [(0, 1), (0, 2), (1, 0), (1, 2), (2, 0), (2, 1)]
+
+    """
+
+    # Algorithm: https://w.wiki/Qai
+    def _full(A):
+        while True:
+            # Yield the permutation we have
+            yield tuple(A)
+
+            # Find the largest index i such that A[i] < A[i + 1]
+            for i in range(size - 2, -1, -1):
+                if A[i] < A[i + 1]:
+                    break
+            #  If no such index exists, this permutation is the last one
+            else:
+                return
+
+            # Find the largest index j greater than j such that A[i] < A[j]
+            for j in range(size - 1, i, -1):
+                if A[i] < A[j]:
+                    break
+
+            # Swap the value of A[i] with that of A[j], then reverse the
+            # sequence from A[i + 1] to form the new permutation
+            A[i], A[j] = A[j], A[i]
+            A[i + 1 :] = A[: i - size : -1]  # A[i + 1:][::-1]
+
+    # Algorithm: modified from the above
+    def _partial(A, r):
+        # Split A into the first r items and the last r items
+        head, tail = A[:r], A[r:]
+        right_head_indexes = range(r - 1, -1, -1)
+        left_tail_indexes = range(len(tail))
+
+        while True:
+            # Yield the permutation we have
+            yield tuple(head)
+
+            # Starting from the right, find the first index of the head with
+            # value smaller than the maximum value of the tail - call it i.
+            pivot = tail[-1]
+            for i in right_head_indexes:
+                if head[i] < pivot:
+                    break
+                pivot = head[i]
+            else:
+                return
+
+            # Starting from the left, find the first value of the tail
+            # with a value greater than head[i] and swap.
+            for j in left_tail_indexes:
+                if tail[j] > head[i]:
+                    head[i], tail[j] = tail[j], head[i]
+                    break
+            # If we didn't find one, start from the right and find the first
+            # index of the head with a value greater than head[i] and swap.
+            else:
+                for j in right_head_indexes:
+                    if head[j] > head[i]:
+                        head[i], head[j] = head[j], head[i]
+                        break
+
+            # Reverse head[i + 1:] and swap it with tail[:r - (i + 1)]
+            tail += head[: i - r : -1]  # head[i + 1:][::-1]
+            i += 1
+            head[i:], tail[:] = tail[: r - i], tail[r - i :]
+
+    items = sorted(iterable)
+
+    size = len(items)
+    if r is None:
+        r = size
+
+    if 0 < r <= size:
+        return _full(items) if (r == size) else _partial(items, r)
+
+    return iter(() if r else ((),))
+
+
+def intersperse(e, iterable, n=1):
+    """Intersperse filler element *e* among the items in *iterable*, leaving
+    *n* items between each filler element.
+
+        >>> list(intersperse('!', [1, 2, 3, 4, 5]))
+        [1, '!', 2, '!', 3, '!', 4, '!', 5]
+
+        >>> list(intersperse(None, [1, 2, 3, 4, 5], n=2))
+        [1, 2, None, 3, 4, None, 5]
+
+    """
+    if n == 0:
+        raise ValueError('n must be > 0')
+    elif n == 1:
+        # interleave(repeat(e), iterable) -> e, x_0, e, x_1, e, x_2...
+        # islice(..., 1, None) -> x_0, e, x_1, e, x_2...
+        return islice(interleave(repeat(e), iterable), 1, None)
+    else:
+        # interleave(filler, chunks) -> [e], [x_0, x_1], [e], [x_2, x_3]...
+        # islice(..., 1, None) -> [x_0, x_1], [e], [x_2, x_3]...
+        # flatten(...) -> x_0, x_1, e, x_2, x_3...
+        filler = repeat([e])
+        chunks = chunked(iterable, n)
+        return flatten(islice(interleave(filler, chunks), 1, None))
+
+
+def unique_to_each(*iterables):
+    """Return the elements from each of the input iterables that aren't in the
+    other input iterables.
+
+    For example, suppose you have a set of packages, each with a set of
+    dependencies::
+
+        {'pkg_1': {'A', 'B'}, 'pkg_2': {'B', 'C'}, 'pkg_3': {'B', 'D'}}
+
+    If you remove one package, which dependencies can also be removed?
+
+    If ``pkg_1`` is removed, then ``A`` is no longer necessary - it is not
+    associated with ``pkg_2`` or ``pkg_3``. Similarly, ``C`` is only needed for
+    ``pkg_2``, and ``D`` is only needed for ``pkg_3``::
+
+        >>> unique_to_each({'A', 'B'}, {'B', 'C'}, {'B', 'D'})
+        [['A'], ['C'], ['D']]
+
+    If there are duplicates in one input iterable that aren't in the others
+    they will be duplicated in the output. Input order is preserved::
+
+        >>> unique_to_each("mississippi", "missouri")
+        [['p', 'p'], ['o', 'u', 'r']]
+
+    It is assumed that the elements of each iterable are hashable.
+
+    """
+    pool = [list(it) for it in iterables]
+    counts = Counter(chain.from_iterable(map(set, pool)))
+    uniques = {element for element in counts if counts[element] == 1}
+    return [list(filter(uniques.__contains__, it)) for it in pool]
+
+
+def windowed(seq, n, fillvalue=None, step=1):
+    """Return a sliding window of width *n* over the given iterable.
+
+        >>> all_windows = windowed([1, 2, 3, 4, 5], 3)
+        >>> list(all_windows)
+        [(1, 2, 3), (2, 3, 4), (3, 4, 5)]
+
+    When the window is larger than the iterable, *fillvalue* is used in place
+    of missing values:
+
+        >>> list(windowed([1, 2, 3], 4))
+        [(1, 2, 3, None)]
+
+    Each window will advance in increments of *step*:
+
+        >>> list(windowed([1, 2, 3, 4, 5, 6], 3, fillvalue='!', step=2))
+        [(1, 2, 3), (3, 4, 5), (5, 6, '!')]
+
+    To slide into the iterable's items, use :func:`chain` to add filler items
+    to the left:
+
+        >>> iterable = [1, 2, 3, 4]
+        >>> n = 3
+        >>> padding = [None] * (n - 1)
+        >>> list(windowed(chain(padding, iterable), 3))
+        [(None, None, 1), (None, 1, 2), (1, 2, 3), (2, 3, 4)]
+    """
+    if n < 0:
+        raise ValueError('n must be >= 0')
+    if n == 0:
+        yield ()
+        return
+    if step < 1:
+        raise ValueError('step must be >= 1')
+
+    iterable = iter(seq)
+
+    # Generate first window
+    window = deque(islice(iterable, n), maxlen=n)
+
+    # Deal with the first window not being full
+    if not window:
+        return
+    if len(window) < n:
+        yield tuple(window) + ((fillvalue,) * (n - len(window)))
+        return
+    yield tuple(window)
+
+    # Create the filler for the next windows. The padding ensures
+    # we have just enough elements to fill the last window.
+    padding = (fillvalue,) * (n - 1 if step >= n else step - 1)
+    filler = map(window.append, chain(iterable, padding))
+
+    # Generate the rest of the windows
+    for _ in islice(filler, step - 1, None, step):
+        yield tuple(window)
+
+
+def substrings(iterable):
+    """Yield all of the substrings of *iterable*.
+
+        >>> [''.join(s) for s in substrings('more')]
+        ['m', 'o', 'r', 'e', 'mo', 'or', 're', 'mor', 'ore', 'more']
+
+    Note that non-string iterables can also be subdivided.
+
+        >>> list(substrings([0, 1, 2]))
+        [(0,), (1,), (2,), (0, 1), (1, 2), (0, 1, 2)]
+
+    """
+    # The length-1 substrings
+    seq = []
+    for item in iter(iterable):
+        seq.append(item)
+        yield (item,)
+    seq = tuple(seq)
+    item_count = len(seq)
+
+    # And the rest
+    for n in range(2, item_count + 1):
+        for i in range(item_count - n + 1):
+            yield seq[i : i + n]
+
+
+def substrings_indexes(seq, reverse=False):
+    """Yield all substrings and their positions in *seq*
+
+    The items yielded will be a tuple of the form ``(substr, i, j)``, where
+    ``substr == seq[i:j]``.
+
+    This function only works for iterables that support slicing, such as
+    ``str`` objects.
+
+    >>> for item in substrings_indexes('more'):
+    ...    print(item)
+    ('m', 0, 1)
+    ('o', 1, 2)
+    ('r', 2, 3)
+    ('e', 3, 4)
+    ('mo', 0, 2)
+    ('or', 1, 3)
+    ('re', 2, 4)
+    ('mor', 0, 3)
+    ('ore', 1, 4)
+    ('more', 0, 4)
+
+    Set *reverse* to ``True`` to yield the same items in the opposite order.
+
+
+    """
+    r = range(1, len(seq) + 1)
+    if reverse:
+        r = reversed(r)
+    return (
+        (seq[i : i + L], i, i + L) for L in r for i in range(len(seq) - L + 1)
+    )
+
+
+class bucket:
+    """Wrap *iterable* and return an object that buckets the iterable into
+    child iterables based on a *key* function.
+
+        >>> iterable = ['a1', 'b1', 'c1', 'a2', 'b2', 'c2', 'b3']
+        >>> s = bucket(iterable, key=lambda x: x[0])  # Bucket by 1st character
+        >>> sorted(list(s))  # Get the keys
+        ['a', 'b', 'c']
+        >>> a_iterable = s['a']
+        >>> next(a_iterable)
+        'a1'
+        >>> next(a_iterable)
+        'a2'
+        >>> list(s['b'])
+        ['b1', 'b2', 'b3']
+
+    The original iterable will be advanced and its items will be cached until
+    they are used by the child iterables. This may require significant storage.
+
+    By default, attempting to select a bucket to which no items belong  will
+    exhaust the iterable and cache all values.
+    If you specify a *validator* function, selected buckets will instead be
+    checked against it.
+
+        >>> from itertools import count
+        >>> it = count(1, 2)  # Infinite sequence of odd numbers
+        >>> key = lambda x: x % 10  # Bucket by last digit
+        >>> validator = lambda x: x in {1, 3, 5, 7, 9}  # Odd digits only
+        >>> s = bucket(it, key=key, validator=validator)
+        >>> 2 in s
+        False
+        >>> list(s[2])
+        []
+
+    """
+
+    def __init__(self, iterable, key, validator=None):
+        self._it = iter(iterable)
+        self._key = key
+        self._cache = defaultdict(deque)
+        self._validator = validator or (lambda x: True)
+
+    def __contains__(self, value):
+        if not self._validator(value):
+            return False
+
+        try:
+            item = next(self[value])
+        except StopIteration:
+            return False
+        else:
+            self._cache[value].appendleft(item)
+
+        return True
+
+    def _get_values(self, value):
+        """
+        Helper to yield items from the parent iterator that match *value*.
+        Items that don't match are stored in the local cache as they
+        are encountered.
+        """
+        while True:
+            # If we've cached some items that match the target value, emit
+            # the first one and evict it from the cache.
+            if self._cache[value]:
+                yield self._cache[value].popleft()
+            # Otherwise we need to advance the parent iterator to search for
+            # a matching item, caching the rest.
+            else:
+                while True:
+                    try:
+                        item = next(self._it)
+                    except StopIteration:
+                        return
+                    item_value = self._key(item)
+                    if item_value == value:
+                        yield item
+                        break
+                    elif self._validator(item_value):
+                        self._cache[item_value].append(item)
+
+    def __iter__(self):
+        for item in self._it:
+            item_value = self._key(item)
+            if self._validator(item_value):
+                self._cache[item_value].append(item)
+
+        yield from self._cache.keys()
+
+    def __getitem__(self, value):
+        if not self._validator(value):
+            return iter(())
+
+        return self._get_values(value)
+
+
+def spy(iterable, n=1):
+    """Return a 2-tuple with a list containing the first *n* elements of
+    *iterable*, and an iterator with the same items as *iterable*.
+    This allows you to "look ahead" at the items in the iterable without
+    advancing it.
+
+    There is one item in the list by default:
+
+        >>> iterable = 'abcdefg'
+        >>> head, iterable = spy(iterable)
+        >>> head
+        ['a']
+        >>> list(iterable)
+        ['a', 'b', 'c', 'd', 'e', 'f', 'g']
+
+    You may use unpacking to retrieve items instead of lists:
+
+        >>> (head,), iterable = spy('abcdefg')
+        >>> head
+        'a'
+        >>> (first, second), iterable = spy('abcdefg', 2)
+        >>> first
+        'a'
+        >>> second
+        'b'
+
+    The number of items requested can be larger than the number of items in
+    the iterable:
+
+        >>> iterable = [1, 2, 3, 4, 5]
+        >>> head, iterable = spy(iterable, 10)
+        >>> head
+        [1, 2, 3, 4, 5]
+        >>> list(iterable)
+        [1, 2, 3, 4, 5]
+
+    """
+    it = iter(iterable)
+    head = take(n, it)
+
+    return head.copy(), chain(head, it)
+
+
+def interleave(*iterables):
+    """Return a new iterable yielding from each iterable in turn,
+    until the shortest is exhausted.
+
+        >>> list(interleave([1, 2, 3], [4, 5], [6, 7, 8]))
+        [1, 4, 6, 2, 5, 7]
+
+    For a version that doesn't terminate after the shortest iterable is
+    exhausted, see :func:`interleave_longest`.
+
+    """
+    return chain.from_iterable(zip(*iterables))
+
+
+def interleave_longest(*iterables):
+    """Return a new iterable yielding from each iterable in turn,
+    skipping any that are exhausted.
+
+        >>> list(interleave_longest([1, 2, 3], [4, 5], [6, 7, 8]))
+        [1, 4, 6, 2, 5, 7, 3, 8]
+
+    This function produces the same output as :func:`roundrobin`, but may
+    perform better for some inputs (in particular when the number of iterables
+    is large).
+
+    """
+    i = chain.from_iterable(zip_longest(*iterables, fillvalue=_marker))
+    return (x for x in i if x is not _marker)
+
+
+def interleave_evenly(iterables, lengths=None):
+    """
+    Interleave multiple iterables so that their elements are evenly distributed
+    throughout the output sequence.
+
+    >>> iterables = [1, 2, 3, 4, 5], ['a', 'b']
+    >>> list(interleave_evenly(iterables))
+    [1, 2, 'a', 3, 4, 'b', 5]
+
+    >>> iterables = [[1, 2, 3], [4, 5], [6, 7, 8]]
+    >>> list(interleave_evenly(iterables))
+    [1, 6, 4, 2, 7, 3, 8, 5]
+
+    This function requires iterables of known length. Iterables without
+    ``__len__()`` can be used by manually specifying lengths with *lengths*:
+
+    >>> from itertools import combinations, repeat
+    >>> iterables = [combinations(range(4), 2), ['a', 'b', 'c']]
+    >>> lengths = [4 * (4 - 1) // 2, 3]
+    >>> list(interleave_evenly(iterables, lengths=lengths))
+    [(0, 1), (0, 2), 'a', (0, 3), (1, 2), 'b', (1, 3), (2, 3), 'c']
+
+    Based on Bresenham's algorithm.
+    """
+    if lengths is None:
+        try:
+            lengths = [len(it) for it in iterables]
+        except TypeError:
+            raise ValueError(
+                'Iterable lengths could not be determined automatically. '
+                'Specify them with the lengths keyword.'
+            )
+    elif len(iterables) != len(lengths):
+        raise ValueError('Mismatching number of iterables and lengths.')
+
+    dims = len(lengths)
+
+    # sort iterables by length, descending
+    lengths_permute = sorted(
+        range(dims), key=lambda i: lengths[i], reverse=True
+    )
+    lengths_desc = [lengths[i] for i in lengths_permute]
+    iters_desc = [iter(iterables[i]) for i in lengths_permute]
+
+    # the longest iterable is the primary one (Bresenham: the longest
+    # distance along an axis)
+    delta_primary, deltas_secondary = lengths_desc[0], lengths_desc[1:]
+    iter_primary, iters_secondary = iters_desc[0], iters_desc[1:]
+    errors = [delta_primary // dims] * len(deltas_secondary)
+
+    to_yield = sum(lengths)
+    while to_yield:
+        yield next(iter_primary)
+        to_yield -= 1
+        # update errors for each secondary iterable
+        errors = [e - delta for e, delta in zip(errors, deltas_secondary)]
+
+        # those iterables for which the error is negative are yielded
+        # ("diagonal step" in Bresenham)
+        for i, e_ in enumerate(errors):
+            if e_ < 0:
+                yield next(iters_secondary[i])
+                to_yield -= 1
+                errors[i] += delta_primary
+
+
+def collapse(iterable, base_type=None, levels=None):
+    """Flatten an iterable with multiple levels of nesting (e.g., a list of
+    lists of tuples) into non-iterable types.
+
+        >>> iterable = [(1, 2), ([3, 4], [[5], [6]])]
+        >>> list(collapse(iterable))
+        [1, 2, 3, 4, 5, 6]
+
+    Binary and text strings are not considered iterable and
+    will not be collapsed.
+
+    To avoid collapsing other types, specify *base_type*:
+
+        >>> iterable = ['ab', ('cd', 'ef'), ['gh', 'ij']]
+        >>> list(collapse(iterable, base_type=tuple))
+        ['ab', ('cd', 'ef'), 'gh', 'ij']
+
+    Specify *levels* to stop flattening after a certain level:
+
+    >>> iterable = [('a', ['b']), ('c', ['d'])]
+    >>> list(collapse(iterable))  # Fully flattened
+    ['a', 'b', 'c', 'd']
+    >>> list(collapse(iterable, levels=1))  # Only one level flattened
+    ['a', ['b'], 'c', ['d']]
+
+    """
+    stack = deque()
+    # Add our first node group, treat the iterable as a single node
+    stack.appendleft((0, repeat(iterable, 1)))
+
+    while stack:
+        node_group = stack.popleft()
+        level, nodes = node_group
+
+        # Check if beyond max level
+        if levels is not None and level > levels:
+            yield from nodes
+            continue
+
+        for node in nodes:
+            # Check if done iterating
+            if isinstance(node, (str, bytes)) or (
+                (base_type is not None) and isinstance(node, base_type)
+            ):
+                yield node
+            # Otherwise try to create child nodes
+            else:
+                try:
+                    tree = iter(node)
+                except TypeError:
+                    yield node
+                else:
+                    # Save our current location
+                    stack.appendleft(node_group)
+                    # Append the new child node
+                    stack.appendleft((level + 1, tree))
+                    # Break to process child node
+                    break
+
+
+def side_effect(func, iterable, chunk_size=None, before=None, after=None):
+    """Invoke *func* on each item in *iterable* (or on each *chunk_size* group
+    of items) before yielding the item.
+
+    `func` must be a function that takes a single argument. Its return value
+    will be discarded.
+
+    *before* and *after* are optional functions that take no arguments. They
+    will be executed before iteration starts and after it ends, respectively.
+
+    `side_effect` can be used for logging, updating progress bars, or anything
+    that is not functionally "pure."
+
+    Emitting a status message:
+
+        >>> from more_itertools import consume
+        >>> func = lambda item: print('Received {}'.format(item))
+        >>> consume(side_effect(func, range(2)))
+        Received 0
+        Received 1
+
+    Operating on chunks of items:
+
+        >>> pair_sums = []
+        >>> func = lambda chunk: pair_sums.append(sum(chunk))
+        >>> list(side_effect(func, [0, 1, 2, 3, 4, 5], 2))
+        [0, 1, 2, 3, 4, 5]
+        >>> list(pair_sums)
+        [1, 5, 9]
+
+    Writing to a file-like object:
+
+        >>> from io import StringIO
+        >>> from more_itertools import consume
+        >>> f = StringIO()
+        >>> func = lambda x: print(x, file=f)
+        >>> before = lambda: print(u'HEADER', file=f)
+        >>> after = f.close
+        >>> it = [u'a', u'b', u'c']
+        >>> consume(side_effect(func, it, before=before, after=after))
+        >>> f.closed
+        True
+
+    """
+    try:
+        if before is not None:
+            before()
+
+        if chunk_size is None:
+            for item in iterable:
+                func(item)
+                yield item
+        else:
+            for chunk in chunked(iterable, chunk_size):
+                func(chunk)
+                yield from chunk
+    finally:
+        if after is not None:
+            after()
+
+
+def sliced(seq, n, strict=False):
+    """Yield slices of length *n* from the sequence *seq*.
+
+    >>> list(sliced((1, 2, 3, 4, 5, 6), 3))
+    [(1, 2, 3), (4, 5, 6)]
+
+    By the default, the last yielded slice will have fewer than *n* elements
+    if the length of *seq* is not divisible by *n*:
+
+    >>> list(sliced((1, 2, 3, 4, 5, 6, 7, 8), 3))
+    [(1, 2, 3), (4, 5, 6), (7, 8)]
+
+    If the length of *seq* is not divisible by *n* and *strict* is
+    ``True``, then ``ValueError`` will be raised before the last
+    slice is yielded.
+
+    This function will only work for iterables that support slicing.
+    For non-sliceable iterables, see :func:`chunked`.
+
+    """
+    iterator = takewhile(len, (seq[i : i + n] for i in count(0, n)))
+    if strict:
+
+        def ret():
+            for _slice in iterator:
+                if len(_slice) != n:
+                    raise ValueError("seq is not divisible by n.")
+                yield _slice
+
+        return iter(ret())
+    else:
+        return iterator
+
+
+def split_at(iterable, pred, maxsplit=-1, keep_separator=False):
+    """Yield lists of items from *iterable*, where each list is delimited by
+    an item where callable *pred* returns ``True``.
+
+        >>> list(split_at('abcdcba', lambda x: x == 'b'))
+        [['a'], ['c', 'd', 'c'], ['a']]
+
+        >>> list(split_at(range(10), lambda n: n % 2 == 1))
+        [[0], [2], [4], [6], [8], []]
+
+    At most *maxsplit* splits are done. If *maxsplit* is not specified or -1,
+    then there is no limit on the number of splits:
+
+        >>> list(split_at(range(10), lambda n: n % 2 == 1, maxsplit=2))
+        [[0], [2], [4, 5, 6, 7, 8, 9]]
+
+    By default, the delimiting items are not included in the output.
+    To include them, set *keep_separator* to ``True``.
+
+        >>> list(split_at('abcdcba', lambda x: x == 'b', keep_separator=True))
+        [['a'], ['b'], ['c', 'd', 'c'], ['b'], ['a']]
+
+    """
+    if maxsplit == 0:
+        yield list(iterable)
+        return
+
+    buf = []
+    it = iter(iterable)
+    for item in it:
+        if pred(item):
+            yield buf
+            if keep_separator:
+                yield [item]
+            if maxsplit == 1:
+                yield list(it)
+                return
+            buf = []
+            maxsplit -= 1
+        else:
+            buf.append(item)
+    yield buf
+
+
+def split_before(iterable, pred, maxsplit=-1):
+    """Yield lists of items from *iterable*, where each list ends just before
+    an item for which callable *pred* returns ``True``:
+
+        >>> list(split_before('OneTwo', lambda s: s.isupper()))
+        [['O', 'n', 'e'], ['T', 'w', 'o']]
+
+        >>> list(split_before(range(10), lambda n: n % 3 == 0))
+        [[0, 1, 2], [3, 4, 5], [6, 7, 8], [9]]
+
+    At most *maxsplit* splits are done. If *maxsplit* is not specified or -1,
+    then there is no limit on the number of splits:
+
+        >>> list(split_before(range(10), lambda n: n % 3 == 0, maxsplit=2))
+        [[0, 1, 2], [3, 4, 5], [6, 7, 8, 9]]
+    """
+    if maxsplit == 0:
+        yield list(iterable)
+        return
+
+    buf = []
+    it = iter(iterable)
+    for item in it:
+        if pred(item) and buf:
+            yield buf
+            if maxsplit == 1:
+                yield [item] + list(it)
+                return
+            buf = []
+            maxsplit -= 1
+        buf.append(item)
+    if buf:
+        yield buf
+
+
+def split_after(iterable, pred, maxsplit=-1):
+    """Yield lists of items from *iterable*, where each list ends with an
+    item where callable *pred* returns ``True``:
+
+        >>> list(split_after('one1two2', lambda s: s.isdigit()))
+        [['o', 'n', 'e', '1'], ['t', 'w', 'o', '2']]
+
+        >>> list(split_after(range(10), lambda n: n % 3 == 0))
+        [[0], [1, 2, 3], [4, 5, 6], [7, 8, 9]]
+
+    At most *maxsplit* splits are done. If *maxsplit* is not specified or -1,
+    then there is no limit on the number of splits:
+
+        >>> list(split_after(range(10), lambda n: n % 3 == 0, maxsplit=2))
+        [[0], [1, 2, 3], [4, 5, 6, 7, 8, 9]]
+
+    """
+    if maxsplit == 0:
+        yield list(iterable)
+        return
+
+    buf = []
+    it = iter(iterable)
+    for item in it:
+        buf.append(item)
+        if pred(item) and buf:
+            yield buf
+            if maxsplit == 1:
+                buf = list(it)
+                if buf:
+                    yield buf
+                return
+            buf = []
+            maxsplit -= 1
+    if buf:
+        yield buf
+
+
+def split_when(iterable, pred, maxsplit=-1):
+    """Split *iterable* into pieces based on the output of *pred*.
+    *pred* should be a function that takes successive pairs of items and
+    returns ``True`` if the iterable should be split in between them.
+
+    For example, to find runs of increasing numbers, split the iterable when
+    element ``i`` is larger than element ``i + 1``:
+
+        >>> list(split_when([1, 2, 3, 3, 2, 5, 2, 4, 2], lambda x, y: x > y))
+        [[1, 2, 3, 3], [2, 5], [2, 4], [2]]
+
+    At most *maxsplit* splits are done. If *maxsplit* is not specified or -1,
+    then there is no limit on the number of splits:
+
+        >>> list(split_when([1, 2, 3, 3, 2, 5, 2, 4, 2],
+        ...                 lambda x, y: x > y, maxsplit=2))
+        [[1, 2, 3, 3], [2, 5], [2, 4, 2]]
+
+    """
+    if maxsplit == 0:
+        yield list(iterable)
+        return
+
+    it = iter(iterable)
+    try:
+        cur_item = next(it)
+    except StopIteration:
+        return
+
+    buf = [cur_item]
+    for next_item in it:
+        if pred(cur_item, next_item):
+            yield buf
+            if maxsplit == 1:
+                yield [next_item] + list(it)
+                return
+            buf = []
+            maxsplit -= 1
+
+        buf.append(next_item)
+        cur_item = next_item
+
+    yield buf
+
+
+def split_into(iterable, sizes):
+    """Yield a list of sequential items from *iterable* of length 'n' for each
+    integer 'n' in *sizes*.
+
+        >>> list(split_into([1,2,3,4,5,6], [1,2,3]))
+        [[1], [2, 3], [4, 5, 6]]
+
+    If the sum of *sizes* is smaller than the length of *iterable*, then the
+    remaining items of *iterable* will not be returned.
+
+        >>> list(split_into([1,2,3,4,5,6], [2,3]))
+        [[1, 2], [3, 4, 5]]
+
+    If the sum of *sizes* is larger than the length of *iterable*, fewer items
+    will be returned in the iteration that overruns *iterable* and further
+    lists will be empty:
+
+        >>> list(split_into([1,2,3,4], [1,2,3,4]))
+        [[1], [2, 3], [4], []]
+
+    When a ``None`` object is encountered in *sizes*, the returned list will
+    contain items up to the end of *iterable* the same way that itertools.slice
+    does:
+
+        >>> list(split_into([1,2,3,4,5,6,7,8,9,0], [2,3,None]))
+        [[1, 2], [3, 4, 5], [6, 7, 8, 9, 0]]
+
+    :func:`split_into` can be useful for grouping a series of items where the
+    sizes of the groups are not uniform. An example would be where in a row
+    from a table, multiple columns represent elements of the same feature
+    (e.g. a point represented by x,y,z) but, the format is not the same for
+    all columns.
+    """
+    # convert the iterable argument into an iterator so its contents can
+    # be consumed by islice in case it is a generator
+    it = iter(iterable)
+
+    for size in sizes:
+        if size is None:
+            yield list(it)
+            return
+        else:
+            yield list(islice(it, size))
+
+
+def padded(iterable, fillvalue=None, n=None, next_multiple=False):
+    """Yield the elements from *iterable*, followed by *fillvalue*, such that
+    at least *n* items are emitted.
+
+        >>> list(padded([1, 2, 3], '?', 5))
+        [1, 2, 3, '?', '?']
+
+    If *next_multiple* is ``True``, *fillvalue* will be emitted until the
+    number of items emitted is a multiple of *n*:
+
+        >>> list(padded([1, 2, 3, 4], n=3, next_multiple=True))
+        [1, 2, 3, 4, None, None]
+
+    If *n* is ``None``, *fillvalue* will be emitted indefinitely.
+
+    To create an *iterable* of exactly size *n*, you can truncate with
+    :func:`islice`.
+
+        >>> list(islice(padded([1, 2, 3], '?'), 5))
+        [1, 2, 3, '?', '?']
+        >>> list(islice(padded([1, 2, 3, 4, 5, 6, 7, 8], '?'), 5))
+        [1, 2, 3, 4, 5]
+
+    """
+    iterable = iter(iterable)
+    iterable_with_repeat = chain(iterable, repeat(fillvalue))
+
+    if n is None:
+        return iterable_with_repeat
+    elif n < 1:
+        raise ValueError('n must be at least 1')
+    elif next_multiple:
+
+        def slice_generator():
+            for first in iterable:
+                yield (first,)
+                yield islice(iterable_with_repeat, n - 1)
+
+        # While elements exist produce slices of size n
+        return chain.from_iterable(slice_generator())
+    else:
+        # Ensure the first batch is at least size n then iterate
+        return chain(islice(iterable_with_repeat, n), iterable)
+
+
+def repeat_each(iterable, n=2):
+    """Repeat each element in *iterable* *n* times.
+
+    >>> list(repeat_each('ABC', 3))
+    ['A', 'A', 'A', 'B', 'B', 'B', 'C', 'C', 'C']
+    """
+    return chain.from_iterable(map(repeat, iterable, repeat(n)))
+
+
+def repeat_last(iterable, default=None):
+    """After the *iterable* is exhausted, keep yielding its last element.
+
+        >>> list(islice(repeat_last(range(3)), 5))
+        [0, 1, 2, 2, 2]
+
+    If the iterable is empty, yield *default* forever::
+
+        >>> list(islice(repeat_last(range(0), 42), 5))
+        [42, 42, 42, 42, 42]
+
+    """
+    item = _marker
+    for item in iterable:
+        yield item
+    final = default if item is _marker else item
+    yield from repeat(final)
+
+
+def distribute(n, iterable):
+    """Distribute the items from *iterable* among *n* smaller iterables.
+
+        >>> group_1, group_2 = distribute(2, [1, 2, 3, 4, 5, 6])
+        >>> list(group_1)
+        [1, 3, 5]
+        >>> list(group_2)
+        [2, 4, 6]
+
+    If the length of *iterable* is not evenly divisible by *n*, then the
+    length of the returned iterables will not be identical:
+
+        >>> children = distribute(3, [1, 2, 3, 4, 5, 6, 7])
+        >>> [list(c) for c in children]
+        [[1, 4, 7], [2, 5], [3, 6]]
+
+    If the length of *iterable* is smaller than *n*, then the last returned
+    iterables will be empty:
+
+        >>> children = distribute(5, [1, 2, 3])
+        >>> [list(c) for c in children]
+        [[1], [2], [3], [], []]
+
+    This function uses :func:`itertools.tee` and may require significant
+    storage.
+
+    If you need the order items in the smaller iterables to match the
+    original iterable, see :func:`divide`.
+
+    """
+    if n < 1:
+        raise ValueError('n must be at least 1')
+
+    children = tee(iterable, n)
+    return [islice(it, index, None, n) for index, it in enumerate(children)]
+
+
+def stagger(iterable, offsets=(-1, 0, 1), longest=False, fillvalue=None):
+    """Yield tuples whose elements are offset from *iterable*.
+    The amount by which the `i`-th item in each tuple is offset is given by
+    the `i`-th item in *offsets*.
+
+        >>> list(stagger([0, 1, 2, 3]))
+        [(None, 0, 1), (0, 1, 2), (1, 2, 3)]
+        >>> list(stagger(range(8), offsets=(0, 2, 4)))
+        [(0, 2, 4), (1, 3, 5), (2, 4, 6), (3, 5, 7)]
+
+    By default, the sequence will end when the final element of a tuple is the
+    last item in the iterable. To continue until the first element of a tuple
+    is the last item in the iterable, set *longest* to ``True``::
+
+        >>> list(stagger([0, 1, 2, 3], longest=True))
+        [(None, 0, 1), (0, 1, 2), (1, 2, 3), (2, 3, None), (3, None, None)]
+
+    By default, ``None`` will be used to replace offsets beyond the end of the
+    sequence. Specify *fillvalue* to use some other value.
+
+    """
+    children = tee(iterable, len(offsets))
+
+    return zip_offset(
+        *children, offsets=offsets, longest=longest, fillvalue=fillvalue
+    )
+
+
+def zip_equal(*iterables):
+    """``zip`` the input *iterables* together, but raise
+    ``UnequalIterablesError`` if they aren't all the same length.
+
+        >>> it_1 = range(3)
+        >>> it_2 = iter('abc')
+        >>> list(zip_equal(it_1, it_2))
+        [(0, 'a'), (1, 'b'), (2, 'c')]
+
+        >>> it_1 = range(3)
+        >>> it_2 = iter('abcd')
+        >>> list(zip_equal(it_1, it_2)) # doctest: +IGNORE_EXCEPTION_DETAIL
+        Traceback (most recent call last):
+        ...
+        more_itertools.more.UnequalIterablesError: Iterables have different
+        lengths
+
+    """
+    if hexversion >= 0x30A00A6:
+        warnings.warn(
+            (
+                'zip_equal will be removed in a future version of '
+                'more-itertools. Use the builtin zip function with '
+                'strict=True instead.'
+            ),
+            DeprecationWarning,
+        )
+
+    return _zip_equal(*iterables)
+
+
+def zip_offset(*iterables, offsets, longest=False, fillvalue=None):
+    """``zip`` the input *iterables* together, but offset the `i`-th iterable
+    by the `i`-th item in *offsets*.
+
+        >>> list(zip_offset('0123', 'abcdef', offsets=(0, 1)))
+        [('0', 'b'), ('1', 'c'), ('2', 'd'), ('3', 'e')]
+
+    This can be used as a lightweight alternative to SciPy or pandas to analyze
+    data sets in which some series have a lead or lag relationship.
+
+    By default, the sequence will end when the shortest iterable is exhausted.
+    To continue until the longest iterable is exhausted, set *longest* to
+    ``True``.
+
+        >>> list(zip_offset('0123', 'abcdef', offsets=(0, 1), longest=True))
+        [('0', 'b'), ('1', 'c'), ('2', 'd'), ('3', 'e'), (None, 'f')]
+
+    By default, ``None`` will be used to replace offsets beyond the end of the
+    sequence. Specify *fillvalue* to use some other value.
+
+    """
+    if len(iterables) != len(offsets):
+        raise ValueError("Number of iterables and offsets didn't match")
+
+    staggered = []
+    for it, n in zip(iterables, offsets):
+        if n < 0:
+            staggered.append(chain(repeat(fillvalue, -n), it))
+        elif n > 0:
+            staggered.append(islice(it, n, None))
+        else:
+            staggered.append(it)
+
+    if longest:
+        return zip_longest(*staggered, fillvalue=fillvalue)
+
+    return zip(*staggered)
+
+
+def sort_together(iterables, key_list=(0,), key=None, reverse=False):
+    """Return the input iterables sorted together, with *key_list* as the
+    priority for sorting. All iterables are trimmed to the length of the
+    shortest one.
+
+    This can be used like the sorting function in a spreadsheet. If each
+    iterable represents a column of data, the key list determines which
+    columns are used for sorting.
+
+    By default, all iterables are sorted using the ``0``-th iterable::
+
+        >>> iterables = [(4, 3, 2, 1), ('a', 'b', 'c', 'd')]
+        >>> sort_together(iterables)
+        [(1, 2, 3, 4), ('d', 'c', 'b', 'a')]
+
+    Set a different key list to sort according to another iterable.
+    Specifying multiple keys dictates how ties are broken::
+
+        >>> iterables = [(3, 1, 2), (0, 1, 0), ('c', 'b', 'a')]
+        >>> sort_together(iterables, key_list=(1, 2))
+        [(2, 3, 1), (0, 0, 1), ('a', 'c', 'b')]
+
+    To sort by a function of the elements of the iterable, pass a *key*
+    function. Its arguments are the elements of the iterables corresponding to
+    the key list::
+
+        >>> names = ('a', 'b', 'c')
+        >>> lengths = (1, 2, 3)
+        >>> widths = (5, 2, 1)
+        >>> def area(length, width):
+        ...     return length * width
+        >>> sort_together([names, lengths, widths], key_list=(1, 2), key=area)
+        [('c', 'b', 'a'), (3, 2, 1), (1, 2, 5)]
+
+    Set *reverse* to ``True`` to sort in descending order.
+
+        >>> sort_together([(1, 2, 3), ('c', 'b', 'a')], reverse=True)
+        [(3, 2, 1), ('a', 'b', 'c')]
+
+    """
+    if key is None:
+        # if there is no key function, the key argument to sorted is an
+        # itemgetter
+        key_argument = itemgetter(*key_list)
+    else:
+        # if there is a key function, call it with the items at the offsets
+        # specified by the key function as arguments
+        key_list = list(key_list)
+        if len(key_list) == 1:
+            # if key_list contains a single item, pass the item at that offset
+            # as the only argument to the key function
+            key_offset = key_list[0]
+            key_argument = lambda zipped_items: key(zipped_items[key_offset])
+        else:
+            # if key_list contains multiple items, use itemgetter to return a
+            # tuple of items, which we pass as *args to the key function
+            get_key_items = itemgetter(*key_list)
+            key_argument = lambda zipped_items: key(
+                *get_key_items(zipped_items)
+            )
+
+    return list(
+        zip(*sorted(zip(*iterables), key=key_argument, reverse=reverse))
+    )
+
+
+def unzip(iterable):
+    """The inverse of :func:`zip`, this function disaggregates the elements
+    of the zipped *iterable*.
+
+    The ``i``-th iterable contains the ``i``-th element from each element
+    of the zipped iterable. The first element is used to determine the
+    length of the remaining elements.
+
+        >>> iterable = [('a', 1), ('b', 2), ('c', 3), ('d', 4)]
+        >>> letters, numbers = unzip(iterable)
+        >>> list(letters)
+        ['a', 'b', 'c', 'd']
+        >>> list(numbers)
+        [1, 2, 3, 4]
+
+    This is similar to using ``zip(*iterable)``, but it avoids reading
+    *iterable* into memory. Note, however, that this function uses
+    :func:`itertools.tee` and thus may require significant storage.
+
+    """
+    head, iterable = spy(iter(iterable))
+    if not head:
+        # empty iterable, e.g. zip([], [], [])
+        return ()
+    # spy returns a one-length iterable as head
+    head = head[0]
+    iterables = tee(iterable, len(head))
+
+    def itemgetter(i):
+        def getter(obj):
+            try:
+                return obj[i]
+            except IndexError:
+                # basically if we have an iterable like
+                # iter([(1, 2, 3), (4, 5), (6,)])
+                # the second unzipped iterable would fail at the third tuple
+                # since it would try to access tup[1]
+                # same with the third unzipped iterable and the second tuple
+                # to support these "improperly zipped" iterables,
+                # we create a custom itemgetter
+                # which just stops the unzipped iterables
+                # at first length mismatch
+                raise StopIteration
+
+        return getter
+
+    return tuple(map(itemgetter(i), it) for i, it in enumerate(iterables))
+
+
+def divide(n, iterable):
+    """Divide the elements from *iterable* into *n* parts, maintaining
+    order.
+
+        >>> group_1, group_2 = divide(2, [1, 2, 3, 4, 5, 6])
+        >>> list(group_1)
+        [1, 2, 3]
+        >>> list(group_2)
+        [4, 5, 6]
+
+    If the length of *iterable* is not evenly divisible by *n*, then the
+    length of the returned iterables will not be identical:
+
+        >>> children = divide(3, [1, 2, 3, 4, 5, 6, 7])
+        >>> [list(c) for c in children]
+        [[1, 2, 3], [4, 5], [6, 7]]
+
+    If the length of the iterable is smaller than n, then the last returned
+    iterables will be empty:
+
+        >>> children = divide(5, [1, 2, 3])
+        >>> [list(c) for c in children]
+        [[1], [2], [3], [], []]
+
+    This function will exhaust the iterable before returning.
+    If order is not important, see :func:`distribute`, which does not first
+    pull the iterable into memory.
+
+    """
+    if n < 1:
+        raise ValueError('n must be at least 1')
+
+    try:
+        iterable[:0]
+    except TypeError:
+        seq = tuple(iterable)
+    else:
+        seq = iterable
+
+    q, r = divmod(len(seq), n)
+
+    ret = []
+    stop = 0
+    for i in range(1, n + 1):
+        start = stop
+        stop += q + 1 if i <= r else q
+        ret.append(iter(seq[start:stop]))
+
+    return ret
+
+
+def always_iterable(obj, base_type=(str, bytes)):
+    """If *obj* is iterable, return an iterator over its items::
+
+        >>> obj = (1, 2, 3)
+        >>> list(always_iterable(obj))
+        [1, 2, 3]
+
+    If *obj* is not iterable, return a one-item iterable containing *obj*::
+
+        >>> obj = 1
+        >>> list(always_iterable(obj))
+        [1]
+
+    If *obj* is ``None``, return an empty iterable:
+
+        >>> obj = None
+        >>> list(always_iterable(None))
+        []
+
+    By default, binary and text strings are not considered iterable::
+
+        >>> obj = 'foo'
+        >>> list(always_iterable(obj))
+        ['foo']
+
+    If *base_type* is set, objects for which ``isinstance(obj, base_type)``
+    returns ``True`` won't be considered iterable.
+
+        >>> obj = {'a': 1}
+        >>> list(always_iterable(obj))  # Iterate over the dict's keys
+        ['a']
+        >>> list(always_iterable(obj, base_type=dict))  # Treat dicts as a unit
+        [{'a': 1}]
+
+    Set *base_type* to ``None`` to avoid any special handling and treat objects
+    Python considers iterable as iterable:
+
+        >>> obj = 'foo'
+        >>> list(always_iterable(obj, base_type=None))
+        ['f', 'o', 'o']
+    """
+    if obj is None:
+        return iter(())
+
+    if (base_type is not None) and isinstance(obj, base_type):
+        return iter((obj,))
+
+    try:
+        return iter(obj)
+    except TypeError:
+        return iter((obj,))
+
+
+def adjacent(predicate, iterable, distance=1):
+    """Return an iterable over `(bool, item)` tuples where the `item` is
+    drawn from *iterable* and the `bool` indicates whether
+    that item satisfies the *predicate* or is adjacent to an item that does.
+
+    For example, to find whether items are adjacent to a ``3``::
+
+        >>> list(adjacent(lambda x: x == 3, range(6)))
+        [(False, 0), (False, 1), (True, 2), (True, 3), (True, 4), (False, 5)]
+
+    Set *distance* to change what counts as adjacent. For example, to find
+    whether items are two places away from a ``3``:
+
+        >>> list(adjacent(lambda x: x == 3, range(6), distance=2))
+        [(False, 0), (True, 1), (True, 2), (True, 3), (True, 4), (True, 5)]
+
+    This is useful for contextualizing the results of a search function.
+    For example, a code comparison tool might want to identify lines that
+    have changed, but also surrounding lines to give the viewer of the diff
+    context.
+
+    The predicate function will only be called once for each item in the
+    iterable.
+
+    See also :func:`groupby_transform`, which can be used with this function
+    to group ranges of items with the same `bool` value.
+
+    """
+    # Allow distance=0 mainly for testing that it reproduces results with map()
+    if distance < 0:
+        raise ValueError('distance must be at least 0')
+
+    i1, i2 = tee(iterable)
+    padding = [False] * distance
+    selected = chain(padding, map(predicate, i1), padding)
+    adjacent_to_selected = map(any, windowed(selected, 2 * distance + 1))
+    return zip(adjacent_to_selected, i2)
+
+
+def groupby_transform(iterable, keyfunc=None, valuefunc=None, reducefunc=None):
+    """An extension of :func:`itertools.groupby` that can apply transformations
+    to the grouped data.
+
+    * *keyfunc* is a function computing a key value for each item in *iterable*
+    * *valuefunc* is a function that transforms the individual items from
+      *iterable* after grouping
+    * *reducefunc* is a function that transforms each group of items
+
+    >>> iterable = 'aAAbBBcCC'
+    >>> keyfunc = lambda k: k.upper()
+    >>> valuefunc = lambda v: v.lower()
+    >>> reducefunc = lambda g: ''.join(g)
+    >>> list(groupby_transform(iterable, keyfunc, valuefunc, reducefunc))
+    [('A', 'aaa'), ('B', 'bbb'), ('C', 'ccc')]
+
+    Each optional argument defaults to an identity function if not specified.
+
+    :func:`groupby_transform` is useful when grouping elements of an iterable
+    using a separate iterable as the key. To do this, :func:`zip` the iterables
+    and pass a *keyfunc* that extracts the first element and a *valuefunc*
+    that extracts the second element::
+
+        >>> from operator import itemgetter
+        >>> keys = [0, 0, 1, 1, 1, 2, 2, 2, 3]
+        >>> values = 'abcdefghi'
+        >>> iterable = zip(keys, values)
+        >>> grouper = groupby_transform(iterable, itemgetter(0), itemgetter(1))
+        >>> [(k, ''.join(g)) for k, g in grouper]
+        [(0, 'ab'), (1, 'cde'), (2, 'fgh'), (3, 'i')]
+
+    Note that the order of items in the iterable is significant.
+    Only adjacent items are grouped together, so if you don't want any
+    duplicate groups, you should sort the iterable by the key function.
+
+    """
+    ret = groupby(iterable, keyfunc)
+    if valuefunc:
+        ret = ((k, map(valuefunc, g)) for k, g in ret)
+    if reducefunc:
+        ret = ((k, reducefunc(g)) for k, g in ret)
+
+    return ret
+
+
+class numeric_range(abc.Sequence, abc.Hashable):
+    """An extension of the built-in ``range()`` function whose arguments can
+    be any orderable numeric type.
+
+    With only *stop* specified, *start* defaults to ``0`` and *step*
+    defaults to ``1``. The output items will match the type of *stop*:
+
+        >>> list(numeric_range(3.5))
+        [0.0, 1.0, 2.0, 3.0]
+
+    With only *start* and *stop* specified, *step* defaults to ``1``. The
+    output items will match the type of *start*:
+
+        >>> from decimal import Decimal
+        >>> start = Decimal('2.1')
+        >>> stop = Decimal('5.1')
+        >>> list(numeric_range(start, stop))
+        [Decimal('2.1'), Decimal('3.1'), Decimal('4.1')]
+
+    With *start*, *stop*, and *step*  specified the output items will match
+    the type of ``start + step``:
+
+        >>> from fractions import Fraction
+        >>> start = Fraction(1, 2)  # Start at 1/2
+        >>> stop = Fraction(5, 2)  # End at 5/2
+        >>> step = Fraction(1, 2)  # Count by 1/2
+        >>> list(numeric_range(start, stop, step))
+        [Fraction(1, 2), Fraction(1, 1), Fraction(3, 2), Fraction(2, 1)]
+
+    If *step* is zero, ``ValueError`` is raised. Negative steps are supported:
+
+        >>> list(numeric_range(3, -1, -1.0))
+        [3.0, 2.0, 1.0, 0.0]
+
+    Be aware of the limitations of floating point numbers; the representation
+    of the yielded numbers may be surprising.
+
+    ``datetime.datetime`` objects can be used for *start* and *stop*, if *step*
+    is a ``datetime.timedelta`` object:
+
+        >>> import datetime
+        >>> start = datetime.datetime(2019, 1, 1)
+        >>> stop = datetime.datetime(2019, 1, 3)
+        >>> step = datetime.timedelta(days=1)
+        >>> items = iter(numeric_range(start, stop, step))
+        >>> next(items)
+        datetime.datetime(2019, 1, 1, 0, 0)
+        >>> next(items)
+        datetime.datetime(2019, 1, 2, 0, 0)
+
+    """
+
+    _EMPTY_HASH = hash(range(0, 0))
+
+    def __init__(self, *args):
+        argc = len(args)
+        if argc == 1:
+            (self._stop,) = args
+            self._start = type(self._stop)(0)
+            self._step = type(self._stop - self._start)(1)
+        elif argc == 2:
+            self._start, self._stop = args
+            self._step = type(self._stop - self._start)(1)
+        elif argc == 3:
+            self._start, self._stop, self._step = args
+        elif argc == 0:
+            raise TypeError(
+                'numeric_range expected at least '
+                '1 argument, got {}'.format(argc)
+            )
+        else:
+            raise TypeError(
+                'numeric_range expected at most '
+                '3 arguments, got {}'.format(argc)
+            )
+
+        self._zero = type(self._step)(0)
+        if self._step == self._zero:
+            raise ValueError('numeric_range() arg 3 must not be zero')
+        self._growing = self._step > self._zero
+
+    def __bool__(self):
+        if self._growing:
+            return self._start < self._stop
+        else:
+            return self._start > self._stop
+
+    def __contains__(self, elem):
+        if self._growing:
+            if self._start <= elem < self._stop:
+                return (elem - self._start) % self._step == self._zero
+        else:
+            if self._start >= elem > self._stop:
+                return (self._start - elem) % (-self._step) == self._zero
+
+        return False
+
+    def __eq__(self, other):
+        if isinstance(other, numeric_range):
+            empty_self = not bool(self)
+            empty_other = not bool(other)
+            if empty_self or empty_other:
+                return empty_self and empty_other  # True if both empty
+            else:
+                return (
+                    self._start == other._start
+                    and self._step == other._step
+                    and self._get_by_index(-1) == other._get_by_index(-1)
+                )
+        else:
+            return False
+
+    def __getitem__(self, key):
+        if isinstance(key, int):
+            return self._get_by_index(key)
+        elif isinstance(key, slice):
+            step = self._step if key.step is None else key.step * self._step
+
+            if key.start is None or key.start <= -self._len:
+                start = self._start
+            elif key.start >= self._len:
+                start = self._stop
+            else:  # -self._len < key.start < self._len
+                start = self._get_by_index(key.start)
+
+            if key.stop is None or key.stop >= self._len:
+                stop = self._stop
+            elif key.stop <= -self._len:
+                stop = self._start
+            else:  # -self._len < key.stop < self._len
+                stop = self._get_by_index(key.stop)
+
+            return numeric_range(start, stop, step)
+        else:
+            raise TypeError(
+                'numeric range indices must be '
+                'integers or slices, not {}'.format(type(key).__name__)
+            )
+
+    def __hash__(self):
+        if self:
+            return hash((self._start, self._get_by_index(-1), self._step))
+        else:
+            return self._EMPTY_HASH
+
+    def __iter__(self):
+        values = (self._start + (n * self._step) for n in count())
+        if self._growing:
+            return takewhile(partial(gt, self._stop), values)
+        else:
+            return takewhile(partial(lt, self._stop), values)
+
+    def __len__(self):
+        return self._len
+
+    @cached_property
+    def _len(self):
+        if self._growing:
+            start = self._start
+            stop = self._stop
+            step = self._step
+        else:
+            start = self._stop
+            stop = self._start
+            step = -self._step
+        distance = stop - start
+        if distance <= self._zero:
+            return 0
+        else:  # distance > 0 and step > 0: regular euclidean division
+            q, r = divmod(distance, step)
+            return int(q) + int(r != self._zero)
+
+    def __reduce__(self):
+        return numeric_range, (self._start, self._stop, self._step)
+
+    def __repr__(self):
+        if self._step == 1:
+            return "numeric_range({}, {})".format(
+                repr(self._start), repr(self._stop)
+            )
+        else:
+            return "numeric_range({}, {}, {})".format(
+                repr(self._start), repr(self._stop), repr(self._step)
+            )
+
+    def __reversed__(self):
+        return iter(
+            numeric_range(
+                self._get_by_index(-1), self._start - self._step, -self._step
+            )
+        )
+
+    def count(self, value):
+        return int(value in self)
+
+    def index(self, value):
+        if self._growing:
+            if self._start <= value < self._stop:
+                q, r = divmod(value - self._start, self._step)
+                if r == self._zero:
+                    return int(q)
+        else:
+            if self._start >= value > self._stop:
+                q, r = divmod(self._start - value, -self._step)
+                if r == self._zero:
+                    return int(q)
+
+        raise ValueError("{} is not in numeric range".format(value))
+
+    def _get_by_index(self, i):
+        if i < 0:
+            i += self._len
+        if i < 0 or i >= self._len:
+            raise IndexError("numeric range object index out of range")
+        return self._start + i * self._step
+
+
+def count_cycle(iterable, n=None):
+    """Cycle through the items from *iterable* up to *n* times, yielding
+    the number of completed cycles along with each item. If *n* is omitted the
+    process repeats indefinitely.
+
+    >>> list(count_cycle('AB', 3))
+    [(0, 'A'), (0, 'B'), (1, 'A'), (1, 'B'), (2, 'A'), (2, 'B')]
+
+    """
+    iterable = tuple(iterable)
+    if not iterable:
+        return iter(())
+    counter = count() if n is None else range(n)
+    return ((i, item) for i in counter for item in iterable)
+
+
+def mark_ends(iterable):
+    """Yield 3-tuples of the form ``(is_first, is_last, item)``.
+
+    >>> list(mark_ends('ABC'))
+    [(True, False, 'A'), (False, False, 'B'), (False, True, 'C')]
+
+    Use this when looping over an iterable to take special action on its first
+    and/or last items:
+
+    >>> iterable = ['Header', 100, 200, 'Footer']
+    >>> total = 0
+    >>> for is_first, is_last, item in mark_ends(iterable):
+    ...     if is_first:
+    ...         continue  # Skip the header
+    ...     if is_last:
+    ...         continue  # Skip the footer
+    ...     total += item
+    >>> print(total)
+    300
+    """
+    it = iter(iterable)
+
+    try:
+        b = next(it)
+    except StopIteration:
+        return
+
+    try:
+        for i in count():
+            a = b
+            b = next(it)
+            yield i == 0, False, a
+
+    except StopIteration:
+        yield i == 0, True, a
+
+
+def locate(iterable, pred=bool, window_size=None):
+    """Yield the index of each item in *iterable* for which *pred* returns
+    ``True``.
+
+    *pred* defaults to :func:`bool`, which will select truthy items:
+
+        >>> list(locate([0, 1, 1, 0, 1, 0, 0]))
+        [1, 2, 4]
+
+    Set *pred* to a custom function to, e.g., find the indexes for a particular
+    item.
+
+        >>> list(locate(['a', 'b', 'c', 'b'], lambda x: x == 'b'))
+        [1, 3]
+
+    If *window_size* is given, then the *pred* function will be called with
+    that many items. This enables searching for sub-sequences:
+
+        >>> iterable = [0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3]
+        >>> pred = lambda *args: args == (1, 2, 3)
+        >>> list(locate(iterable, pred=pred, window_size=3))
+        [1, 5, 9]
+
+    Use with :func:`seekable` to find indexes and then retrieve the associated
+    items:
+
+        >>> from itertools import count
+        >>> from more_itertools import seekable
+        >>> source = (3 * n + 1 if (n % 2) else n // 2 for n in count())
+        >>> it = seekable(source)
+        >>> pred = lambda x: x > 100
+        >>> indexes = locate(it, pred=pred)
+        >>> i = next(indexes)
+        >>> it.seek(i)
+        >>> next(it)
+        106
+
+    """
+    if window_size is None:
+        return compress(count(), map(pred, iterable))
+
+    if window_size < 1:
+        raise ValueError('window size must be at least 1')
+
+    it = windowed(iterable, window_size, fillvalue=_marker)
+    return compress(count(), starmap(pred, it))
+
+
+def longest_common_prefix(iterables):
+    """Yield elements of the longest common prefix amongst given *iterables*.
+
+    >>> ''.join(longest_common_prefix(['abcd', 'abc', 'abf']))
+    'ab'
+
+    """
+    return (c[0] for c in takewhile(all_equal, zip(*iterables)))
+
+
+def lstrip(iterable, pred):
+    """Yield the items from *iterable*, but strip any from the beginning
+    for which *pred* returns ``True``.
+
+    For example, to remove a set of items from the start of an iterable:
+
+        >>> iterable = (None, False, None, 1, 2, None, 3, False, None)
+        >>> pred = lambda x: x in {None, False, ''}
+        >>> list(lstrip(iterable, pred))
+        [1, 2, None, 3, False, None]
+
+    This function is analogous to to :func:`str.lstrip`, and is essentially
+    an wrapper for :func:`itertools.dropwhile`.
+
+    """
+    return dropwhile(pred, iterable)
+
+
+def rstrip(iterable, pred):
+    """Yield the items from *iterable*, but strip any from the end
+    for which *pred* returns ``True``.
+
+    For example, to remove a set of items from the end of an iterable:
+
+        >>> iterable = (None, False, None, 1, 2, None, 3, False, None)
+        >>> pred = lambda x: x in {None, False, ''}
+        >>> list(rstrip(iterable, pred))
+        [None, False, None, 1, 2, None, 3]
+
+    This function is analogous to :func:`str.rstrip`.
+
+    """
+    cache = []
+    cache_append = cache.append
+    cache_clear = cache.clear
+    for x in iterable:
+        if pred(x):
+            cache_append(x)
+        else:
+            yield from cache
+            cache_clear()
+            yield x
+
+
+def strip(iterable, pred):
+    """Yield the items from *iterable*, but strip any from the
+    beginning and end for which *pred* returns ``True``.
+
+    For example, to remove a set of items from both ends of an iterable:
+
+        >>> iterable = (None, False, None, 1, 2, None, 3, False, None)
+        >>> pred = lambda x: x in {None, False, ''}
+        >>> list(strip(iterable, pred))
+        [1, 2, None, 3]
+
+    This function is analogous to :func:`str.strip`.
+
+    """
+    return rstrip(lstrip(iterable, pred), pred)
+
+
+class islice_extended:
+    """An extension of :func:`itertools.islice` that supports negative values
+    for *stop*, *start*, and *step*.
+
+        >>> iterable = iter('abcdefgh')
+        >>> list(islice_extended(iterable, -4, -1))
+        ['e', 'f', 'g']
+
+    Slices with negative values require some caching of *iterable*, but this
+    function takes care to minimize the amount of memory required.
+
+    For example, you can use a negative step with an infinite iterator:
+
+        >>> from itertools import count
+        >>> list(islice_extended(count(), 110, 99, -2))
+        [110, 108, 106, 104, 102, 100]
+
+    You can also use slice notation directly:
+
+        >>> iterable = map(str, count())
+        >>> it = islice_extended(iterable)[10:20:2]
+        >>> list(it)
+        ['10', '12', '14', '16', '18']
+
+    """
+
+    def __init__(self, iterable, *args):
+        it = iter(iterable)
+        if args:
+            self._iterable = _islice_helper(it, slice(*args))
+        else:
+            self._iterable = it
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        return next(self._iterable)
+
+    def __getitem__(self, key):
+        if isinstance(key, slice):
+            return islice_extended(_islice_helper(self._iterable, key))
+
+        raise TypeError('islice_extended.__getitem__ argument must be a slice')
+
+
+def _islice_helper(it, s):
+    start = s.start
+    stop = s.stop
+    if s.step == 0:
+        raise ValueError('step argument must be a non-zero integer or None.')
+    step = s.step or 1
+
+    if step > 0:
+        start = 0 if (start is None) else start
+
+        if start < 0:
+            # Consume all but the last -start items
+            cache = deque(enumerate(it, 1), maxlen=-start)
+            len_iter = cache[-1][0] if cache else 0
+
+            # Adjust start to be positive
+            i = max(len_iter + start, 0)
+
+            # Adjust stop to be positive
+            if stop is None:
+                j = len_iter
+            elif stop >= 0:
+                j = min(stop, len_iter)
+            else:
+                j = max(len_iter + stop, 0)
+
+            # Slice the cache
+            n = j - i
+            if n <= 0:
+                return
+
+            for index, item in islice(cache, 0, n, step):
+                yield item
+        elif (stop is not None) and (stop < 0):
+            # Advance to the start position
+            next(islice(it, start, start), None)
+
+            # When stop is negative, we have to carry -stop items while
+            # iterating
+            cache = deque(islice(it, -stop), maxlen=-stop)
+
+            for index, item in enumerate(it):
+                cached_item = cache.popleft()
+                if index % step == 0:
+                    yield cached_item
+                cache.append(item)
+        else:
+            # When both start and stop are positive we have the normal case
+            yield from islice(it, start, stop, step)
+    else:
+        start = -1 if (start is None) else start
+
+        if (stop is not None) and (stop < 0):
+            # Consume all but the last items
+            n = -stop - 1
+            cache = deque(enumerate(it, 1), maxlen=n)
+            len_iter = cache[-1][0] if cache else 0
+
+            # If start and stop are both negative they are comparable and
+            # we can just slice. Otherwise we can adjust start to be negative
+            # and then slice.
+            if start < 0:
+                i, j = start, stop
+            else:
+                i, j = min(start - len_iter, -1), None
+
+            for index, item in list(cache)[i:j:step]:
+                yield item
+        else:
+            # Advance to the stop position
+            if stop is not None:
+                m = stop + 1
+                next(islice(it, m, m), None)
+
+            # stop is positive, so if start is negative they are not comparable
+            # and we need the rest of the items.
+            if start < 0:
+                i = start
+                n = None
+            # stop is None and start is positive, so we just need items up to
+            # the start index.
+            elif stop is None:
+                i = None
+                n = start + 1
+            # Both stop and start are positive, so they are comparable.
+            else:
+                i = None
+                n = start - stop
+                if n <= 0:
+                    return
+
+            cache = list(islice(it, n))
+
+            yield from cache[i::step]
+
+
+def always_reversible(iterable):
+    """An extension of :func:`reversed` that supports all iterables, not
+    just those which implement the ``Reversible`` or ``Sequence`` protocols.
+
+        >>> print(*always_reversible(x for x in range(3)))
+        2 1 0
+
+    If the iterable is already reversible, this function returns the
+    result of :func:`reversed()`. If the iterable is not reversible,
+    this function will cache the remaining items in the iterable and
+    yield them in reverse order, which may require significant storage.
+    """
+    try:
+        return reversed(iterable)
+    except TypeError:
+        return reversed(list(iterable))
+
+
+def consecutive_groups(iterable, ordering=lambda x: x):
+    """Yield groups of consecutive items using :func:`itertools.groupby`.
+    The *ordering* function determines whether two items are adjacent by
+    returning their position.
+
+    By default, the ordering function is the identity function. This is
+    suitable for finding runs of numbers:
+
+        >>> iterable = [1, 10, 11, 12, 20, 30, 31, 32, 33, 40]
+        >>> for group in consecutive_groups(iterable):
+        ...     print(list(group))
+        [1]
+        [10, 11, 12]
+        [20]
+        [30, 31, 32, 33]
+        [40]
+
+    For finding runs of adjacent letters, try using the :meth:`index` method
+    of a string of letters:
+
+        >>> from string import ascii_lowercase
+        >>> iterable = 'abcdfgilmnop'
+        >>> ordering = ascii_lowercase.index
+        >>> for group in consecutive_groups(iterable, ordering):
+        ...     print(list(group))
+        ['a', 'b', 'c', 'd']
+        ['f', 'g']
+        ['i']
+        ['l', 'm', 'n', 'o', 'p']
+
+    Each group of consecutive items is an iterator that shares it source with
+    *iterable*. When an an output group is advanced, the previous group is
+    no longer available unless its elements are copied (e.g., into a ``list``).
+
+        >>> iterable = [1, 2, 11, 12, 21, 22]
+        >>> saved_groups = []
+        >>> for group in consecutive_groups(iterable):
+        ...     saved_groups.append(list(group))  # Copy group elements
+        >>> saved_groups
+        [[1, 2], [11, 12], [21, 22]]
+
+    """
+    for k, g in groupby(
+        enumerate(iterable), key=lambda x: x[0] - ordering(x[1])
+    ):
+        yield map(itemgetter(1), g)
+
+
+def difference(iterable, func=sub, *, initial=None):
+    """This function is the inverse of :func:`itertools.accumulate`. By default
+    it will compute the first difference of *iterable* using
+    :func:`operator.sub`:
+
+        >>> from itertools import accumulate
+        >>> iterable = accumulate([0, 1, 2, 3, 4])  # produces 0, 1, 3, 6, 10
+        >>> list(difference(iterable))
+        [0, 1, 2, 3, 4]
+
+    *func* defaults to :func:`operator.sub`, but other functions can be
+    specified. They will be applied as follows::
+
+        A, B, C, D, ... --> A, func(B, A), func(C, B), func(D, C), ...
+
+    For example, to do progressive division:
+
+        >>> iterable = [1, 2, 6, 24, 120]
+        >>> func = lambda x, y: x // y
+        >>> list(difference(iterable, func))
+        [1, 2, 3, 4, 5]
+
+    If the *initial* keyword is set, the first element will be skipped when
+    computing successive differences.
+
+        >>> it = [10, 11, 13, 16]  # from accumulate([1, 2, 3], initial=10)
+        >>> list(difference(it, initial=10))
+        [1, 2, 3]
+
+    """
+    a, b = tee(iterable)
+    try:
+        first = [next(b)]
+    except StopIteration:
+        return iter([])
+
+    if initial is not None:
+        first = []
+
+    return chain(first, map(func, b, a))
+
+
+class SequenceView(Sequence):
+    """Return a read-only view of the sequence object *target*.
+
+    :class:`SequenceView` objects are analogous to Python's built-in
+    "dictionary view" types. They provide a dynamic view of a sequence's items,
+    meaning that when the sequence updates, so does the view.
+
+        >>> seq = ['0', '1', '2']
+        >>> view = SequenceView(seq)
+        >>> view
+        SequenceView(['0', '1', '2'])
+        >>> seq.append('3')
+        >>> view
+        SequenceView(['0', '1', '2', '3'])
+
+    Sequence views support indexing, slicing, and length queries. They act
+    like the underlying sequence, except they don't allow assignment:
+
+        >>> view[1]
+        '1'
+        >>> view[1:-1]
+        ['1', '2']
+        >>> len(view)
+        4
+
+    Sequence views are useful as an alternative to copying, as they don't
+    require (much) extra storage.
+
+    """
+
+    def __init__(self, target):
+        if not isinstance(target, Sequence):
+            raise TypeError
+        self._target = target
+
+    def __getitem__(self, index):
+        return self._target[index]
+
+    def __len__(self):
+        return len(self._target)
+
+    def __repr__(self):
+        return '{}({})'.format(self.__class__.__name__, repr(self._target))
+
+
+class seekable:
+    """Wrap an iterator to allow for seeking backward and forward. This
+    progressively caches the items in the source iterable so they can be
+    re-visited.
+
+    Call :meth:`seek` with an index to seek to that position in the source
+    iterable.
+
+    To "reset" an iterator, seek to ``0``:
+
+        >>> from itertools import count
+        >>> it = seekable((str(n) for n in count()))
+        >>> next(it), next(it), next(it)
+        ('0', '1', '2')
+        >>> it.seek(0)
+        >>> next(it), next(it), next(it)
+        ('0', '1', '2')
+        >>> next(it)
+        '3'
+
+    You can also seek forward:
+
+        >>> it = seekable((str(n) for n in range(20)))
+        >>> it.seek(10)
+        >>> next(it)
+        '10'
+        >>> it.relative_seek(-2)  # Seeking relative to the current position
+        >>> next(it)
+        '9'
+        >>> it.seek(20)  # Seeking past the end of the source isn't a problem
+        >>> list(it)
+        []
+        >>> it.seek(0)  # Resetting works even after hitting the end
+        >>> next(it), next(it), next(it)
+        ('0', '1', '2')
+
+    Call :meth:`peek` to look ahead one item without advancing the iterator:
+
+        >>> it = seekable('1234')
+        >>> it.peek()
+        '1'
+        >>> list(it)
+        ['1', '2', '3', '4']
+        >>> it.peek(default='empty')
+        'empty'
+
+    Before the iterator is at its end, calling :func:`bool` on it will return
+    ``True``. After it will return ``False``:
+
+        >>> it = seekable('5678')
+        >>> bool(it)
+        True
+        >>> list(it)
+        ['5', '6', '7', '8']
+        >>> bool(it)
+        False
+
+    You may view the contents of the cache with the :meth:`elements` method.
+    That returns a :class:`SequenceView`, a view that updates automatically:
+
+        >>> it = seekable((str(n) for n in range(10)))
+        >>> next(it), next(it), next(it)
+        ('0', '1', '2')
+        >>> elements = it.elements()
+        >>> elements
+        SequenceView(['0', '1', '2'])
+        >>> next(it)
+        '3'
+        >>> elements
+        SequenceView(['0', '1', '2', '3'])
+
+    By default, the cache grows as the source iterable progresses, so beware of
+    wrapping very large or infinite iterables. Supply *maxlen* to limit the
+    size of the cache (this of course limits how far back you can seek).
+
+        >>> from itertools import count
+        >>> it = seekable((str(n) for n in count()), maxlen=2)
+        >>> next(it), next(it), next(it), next(it)
+        ('0', '1', '2', '3')
+        >>> list(it.elements())
+        ['2', '3']
+        >>> it.seek(0)
+        >>> next(it), next(it), next(it), next(it)
+        ('2', '3', '4', '5')
+        >>> next(it)
+        '6'
+
+    """
+
+    def __init__(self, iterable, maxlen=None):
+        self._source = iter(iterable)
+        if maxlen is None:
+            self._cache = []
+        else:
+            self._cache = deque([], maxlen)
+        self._index = None
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self._index is not None:
+            try:
+                item = self._cache[self._index]
+            except IndexError:
+                self._index = None
+            else:
+                self._index += 1
+                return item
+
+        item = next(self._source)
+        self._cache.append(item)
+        return item
+
+    def __bool__(self):
+        try:
+            self.peek()
+        except StopIteration:
+            return False
+        return True
+
+    def peek(self, default=_marker):
+        try:
+            peeked = next(self)
+        except StopIteration:
+            if default is _marker:
+                raise
+            return default
+        if self._index is None:
+            self._index = len(self._cache)
+        self._index -= 1
+        return peeked
+
+    def elements(self):
+        return SequenceView(self._cache)
+
+    def seek(self, index):
+        self._index = index
+        remainder = index - len(self._cache)
+        if remainder > 0:
+            consume(self, remainder)
+
+    def relative_seek(self, count):
+        index = len(self._cache)
+        self.seek(max(index + count, 0))
+
+
+class run_length:
+    """
+    :func:`run_length.encode` compresses an iterable with run-length encoding.
+    It yields groups of repeated items with the count of how many times they
+    were repeated:
+
+        >>> uncompressed = 'abbcccdddd'
+        >>> list(run_length.encode(uncompressed))
+        [('a', 1), ('b', 2), ('c', 3), ('d', 4)]
+
+    :func:`run_length.decode` decompresses an iterable that was previously
+    compressed with run-length encoding. It yields the items of the
+    decompressed iterable:
+
+        >>> compressed = [('a', 1), ('b', 2), ('c', 3), ('d', 4)]
+        >>> list(run_length.decode(compressed))
+        ['a', 'b', 'b', 'c', 'c', 'c', 'd', 'd', 'd', 'd']
+
+    """
+
+    @staticmethod
+    def encode(iterable):
+        return ((k, ilen(g)) for k, g in groupby(iterable))
+
+    @staticmethod
+    def decode(iterable):
+        return chain.from_iterable(repeat(k, n) for k, n in iterable)
+
+
+def exactly_n(iterable, n, predicate=bool):
+    """Return ``True`` if exactly ``n`` items in the iterable are ``True``
+    according to the *predicate* function.
+
+        >>> exactly_n([True, True, False], 2)
+        True
+        >>> exactly_n([True, True, False], 1)
+        False
+        >>> exactly_n([0, 1, 2, 3, 4, 5], 3, lambda x: x < 3)
+        True
+
+    The iterable will be advanced until ``n + 1`` truthy items are encountered,
+    so avoid calling it on infinite iterables.
+
+    """
+    return len(take(n + 1, filter(predicate, iterable))) == n
+
+
+def circular_shifts(iterable):
+    """Return a list of circular shifts of *iterable*.
+
+    >>> circular_shifts(range(4))
+    [(0, 1, 2, 3), (1, 2, 3, 0), (2, 3, 0, 1), (3, 0, 1, 2)]
+    """
+    lst = list(iterable)
+    return take(len(lst), windowed(cycle(lst), len(lst)))
+
+
+def make_decorator(wrapping_func, result_index=0):
+    """Return a decorator version of *wrapping_func*, which is a function that
+    modifies an iterable. *result_index* is the position in that function's
+    signature where the iterable goes.
+
+    This lets you use itertools on the "production end," i.e. at function
+    definition. This can augment what the function returns without changing the
+    function's code.
+
+    For example, to produce a decorator version of :func:`chunked`:
+
+        >>> from more_itertools import chunked
+        >>> chunker = make_decorator(chunked, result_index=0)
+        >>> @chunker(3)
+        ... def iter_range(n):
+        ...     return iter(range(n))
+        ...
+        >>> list(iter_range(9))
+        [[0, 1, 2], [3, 4, 5], [6, 7, 8]]
+
+    To only allow truthy items to be returned:
+
+        >>> truth_serum = make_decorator(filter, result_index=1)
+        >>> @truth_serum(bool)
+        ... def boolean_test():
+        ...     return [0, 1, '', ' ', False, True]
+        ...
+        >>> list(boolean_test())
+        [1, ' ', True]
+
+    The :func:`peekable` and :func:`seekable` wrappers make for practical
+    decorators:
+
+        >>> from more_itertools import peekable
+        >>> peekable_function = make_decorator(peekable)
+        >>> @peekable_function()
+        ... def str_range(*args):
+        ...     return (str(x) for x in range(*args))
+        ...
+        >>> it = str_range(1, 20, 2)
+        >>> next(it), next(it), next(it)
+        ('1', '3', '5')
+        >>> it.peek()
+        '7'
+        >>> next(it)
+        '7'
+
+    """
+
+    # See https://sites.google.com/site/bbayles/index/decorator_factory for
+    # notes on how this works.
+    def decorator(*wrapping_args, **wrapping_kwargs):
+        def outer_wrapper(f):
+            def inner_wrapper(*args, **kwargs):
+                result = f(*args, **kwargs)
+                wrapping_args_ = list(wrapping_args)
+                wrapping_args_.insert(result_index, result)
+                return wrapping_func(*wrapping_args_, **wrapping_kwargs)
+
+            return inner_wrapper
+
+        return outer_wrapper
+
+    return decorator
+
+
+def map_reduce(iterable, keyfunc, valuefunc=None, reducefunc=None):
+    """Return a dictionary that maps the items in *iterable* to categories
+    defined by *keyfunc*, transforms them with *valuefunc*, and
+    then summarizes them by category with *reducefunc*.
+
+    *valuefunc* defaults to the identity function if it is unspecified.
+    If *reducefunc* is unspecified, no summarization takes place:
+
+        >>> keyfunc = lambda x: x.upper()
+        >>> result = map_reduce('abbccc', keyfunc)
+        >>> sorted(result.items())
+        [('A', ['a']), ('B', ['b', 'b']), ('C', ['c', 'c', 'c'])]
+
+    Specifying *valuefunc* transforms the categorized items:
+
+        >>> keyfunc = lambda x: x.upper()
+        >>> valuefunc = lambda x: 1
+        >>> result = map_reduce('abbccc', keyfunc, valuefunc)
+        >>> sorted(result.items())
+        [('A', [1]), ('B', [1, 1]), ('C', [1, 1, 1])]
+
+    Specifying *reducefunc* summarizes the categorized items:
+
+        >>> keyfunc = lambda x: x.upper()
+        >>> valuefunc = lambda x: 1
+        >>> reducefunc = sum
+        >>> result = map_reduce('abbccc', keyfunc, valuefunc, reducefunc)
+        >>> sorted(result.items())
+        [('A', 1), ('B', 2), ('C', 3)]
+
+    You may want to filter the input iterable before applying the map/reduce
+    procedure:
+
+        >>> all_items = range(30)
+        >>> items = [x for x in all_items if 10 <= x <= 20]  # Filter
+        >>> keyfunc = lambda x: x % 2  # Evens map to 0; odds to 1
+        >>> categories = map_reduce(items, keyfunc=keyfunc)
+        >>> sorted(categories.items())
+        [(0, [10, 12, 14, 16, 18, 20]), (1, [11, 13, 15, 17, 19])]
+        >>> summaries = map_reduce(items, keyfunc=keyfunc, reducefunc=sum)
+        >>> sorted(summaries.items())
+        [(0, 90), (1, 75)]
+
+    Note that all items in the iterable are gathered into a list before the
+    summarization step, which may require significant storage.
+
+    The returned object is a :obj:`collections.defaultdict` with the
+    ``default_factory`` set to ``None``, such that it behaves like a normal
+    dictionary.
+
+    """
+    valuefunc = (lambda x: x) if (valuefunc is None) else valuefunc
+
+    ret = defaultdict(list)
+    for item in iterable:
+        key = keyfunc(item)
+        value = valuefunc(item)
+        ret[key].append(value)
+
+    if reducefunc is not None:
+        for key, value_list in ret.items():
+            ret[key] = reducefunc(value_list)
+
+    ret.default_factory = None
+    return ret
+
+
+def rlocate(iterable, pred=bool, window_size=None):
+    """Yield the index of each item in *iterable* for which *pred* returns
+    ``True``, starting from the right and moving left.
+
+    *pred* defaults to :func:`bool`, which will select truthy items:
+
+        >>> list(rlocate([0, 1, 1, 0, 1, 0, 0]))  # Truthy at 1, 2, and 4
+        [4, 2, 1]
+
+    Set *pred* to a custom function to, e.g., find the indexes for a particular
+    item:
+
+        >>> iterable = iter('abcb')
+        >>> pred = lambda x: x == 'b'
+        >>> list(rlocate(iterable, pred))
+        [3, 1]
+
+    If *window_size* is given, then the *pred* function will be called with
+    that many items. This enables searching for sub-sequences:
+
+        >>> iterable = [0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3]
+        >>> pred = lambda *args: args == (1, 2, 3)
+        >>> list(rlocate(iterable, pred=pred, window_size=3))
+        [9, 5, 1]
+
+    Beware, this function won't return anything for infinite iterables.
+    If *iterable* is reversible, ``rlocate`` will reverse it and search from
+    the right. Otherwise, it will search from the left and return the results
+    in reverse order.
+
+    See :func:`locate` to for other example applications.
+
+    """
+    if window_size is None:
+        try:
+            len_iter = len(iterable)
+            return (len_iter - i - 1 for i in locate(reversed(iterable), pred))
+        except TypeError:
+            pass
+
+    return reversed(list(locate(iterable, pred, window_size)))
+
+
+def replace(iterable, pred, substitutes, count=None, window_size=1):
+    """Yield the items from *iterable*, replacing the items for which *pred*
+    returns ``True`` with the items from the iterable *substitutes*.
+
+        >>> iterable = [1, 1, 0, 1, 1, 0, 1, 1]
+        >>> pred = lambda x: x == 0
+        >>> substitutes = (2, 3)
+        >>> list(replace(iterable, pred, substitutes))
+        [1, 1, 2, 3, 1, 1, 2, 3, 1, 1]
+
+    If *count* is given, the number of replacements will be limited:
+
+        >>> iterable = [1, 1, 0, 1, 1, 0, 1, 1, 0]
+        >>> pred = lambda x: x == 0
+        >>> substitutes = [None]
+        >>> list(replace(iterable, pred, substitutes, count=2))
+        [1, 1, None, 1, 1, None, 1, 1, 0]
+
+    Use *window_size* to control the number of items passed as arguments to
+    *pred*. This allows for locating and replacing subsequences.
+
+        >>> iterable = [0, 1, 2, 5, 0, 1, 2, 5]
+        >>> window_size = 3
+        >>> pred = lambda *args: args == (0, 1, 2)  # 3 items passed to pred
+        >>> substitutes = [3, 4] # Splice in these items
+        >>> list(replace(iterable, pred, substitutes, window_size=window_size))
+        [3, 4, 5, 3, 4, 5]
+
+    """
+    if window_size < 1:
+        raise ValueError('window_size must be at least 1')
+
+    # Save the substitutes iterable, since it's used more than once
+    substitutes = tuple(substitutes)
+
+    # Add padding such that the number of windows matches the length of the
+    # iterable
+    it = chain(iterable, [_marker] * (window_size - 1))
+    windows = windowed(it, window_size)
+
+    n = 0
+    for w in windows:
+        # If the current window matches our predicate (and we haven't hit
+        # our maximum number of replacements), splice in the substitutes
+        # and then consume the following windows that overlap with this one.
+        # For example, if the iterable is (0, 1, 2, 3, 4...)
+        # and the window size is 2, we have (0, 1), (1, 2), (2, 3)...
+        # If the predicate matches on (0, 1), we need to zap (0, 1) and (1, 2)
+        if pred(*w):
+            if (count is None) or (n < count):
+                n += 1
+                yield from substitutes
+                consume(windows, window_size - 1)
+                continue
+
+        # If there was no match (or we've reached the replacement limit),
+        # yield the first item from the window.
+        if w and (w[0] is not _marker):
+            yield w[0]
+
+
+def partitions(iterable):
+    """Yield all possible order-preserving partitions of *iterable*.
+
+    >>> iterable = 'abc'
+    >>> for part in partitions(iterable):
+    ...     print([''.join(p) for p in part])
+    ['abc']
+    ['a', 'bc']
+    ['ab', 'c']
+    ['a', 'b', 'c']
+
+    This is unrelated to :func:`partition`.
+
+    """
+    sequence = list(iterable)
+    n = len(sequence)
+    for i in powerset(range(1, n)):
+        yield [sequence[i:j] for i, j in zip((0,) + i, i + (n,))]
+
+
+def set_partitions(iterable, k=None):
+    """
+    Yield the set partitions of *iterable* into *k* parts. Set partitions are
+    not order-preserving.
+
+    >>> iterable = 'abc'
+    >>> for part in set_partitions(iterable, 2):
+    ...     print([''.join(p) for p in part])
+    ['a', 'bc']
+    ['ab', 'c']
+    ['b', 'ac']
+
+
+    If *k* is not given, every set partition is generated.
+
+    >>> iterable = 'abc'
+    >>> for part in set_partitions(iterable):
+    ...     print([''.join(p) for p in part])
+    ['abc']
+    ['a', 'bc']
+    ['ab', 'c']
+    ['b', 'ac']
+    ['a', 'b', 'c']
+
+    """
+    L = list(iterable)
+    n = len(L)
+    if k is not None:
+        if k < 1:
+            raise ValueError(
+                "Can't partition in a negative or zero number of groups"
+            )
+        elif k > n:
+            return
+
+    def set_partitions_helper(L, k):
+        n = len(L)
+        if k == 1:
+            yield [L]
+        elif n == k:
+            yield [[s] for s in L]
+        else:
+            e, *M = L
+            for p in set_partitions_helper(M, k - 1):
+                yield [[e], *p]
+            for p in set_partitions_helper(M, k):
+                for i in range(len(p)):
+                    yield p[:i] + [[e] + p[i]] + p[i + 1 :]
+
+    if k is None:
+        for k in range(1, n + 1):
+            yield from set_partitions_helper(L, k)
+    else:
+        yield from set_partitions_helper(L, k)
+
+
+class time_limited:
+    """
+    Yield items from *iterable* until *limit_seconds* have passed.
+    If the time limit expires before all items have been yielded, the
+    ``timed_out`` parameter will be set to ``True``.
+
+    >>> from time import sleep
+    >>> def generator():
+    ...     yield 1
+    ...     yield 2
+    ...     sleep(0.2)
+    ...     yield 3
+    >>> iterable = time_limited(0.1, generator())
+    >>> list(iterable)
+    [1, 2]
+    >>> iterable.timed_out
+    True
+
+    Note that the time is checked before each item is yielded, and iteration
+    stops if  the time elapsed is greater than *limit_seconds*. If your time
+    limit is 1 second, but it takes 2 seconds to generate the first item from
+    the iterable, the function will run for 2 seconds and not yield anything.
+    As a special case, when *limit_seconds* is zero, the iterator never
+    returns anything.
+
+    """
+
+    def __init__(self, limit_seconds, iterable):
+        if limit_seconds < 0:
+            raise ValueError('limit_seconds must be positive')
+        self.limit_seconds = limit_seconds
+        self._iterable = iter(iterable)
+        self._start_time = monotonic()
+        self.timed_out = False
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self.limit_seconds == 0:
+            self.timed_out = True
+            raise StopIteration
+        item = next(self._iterable)
+        if monotonic() - self._start_time > self.limit_seconds:
+            self.timed_out = True
+            raise StopIteration
+
+        return item
+
+
+def only(iterable, default=None, too_long=None):
+    """If *iterable* has only one item, return it.
+    If it has zero items, return *default*.
+    If it has more than one item, raise the exception given by *too_long*,
+    which is ``ValueError`` by default.
+
+    >>> only([], default='missing')
+    'missing'
+    >>> only([1])
+    1
+    >>> only([1, 2])  # doctest: +IGNORE_EXCEPTION_DETAIL
+    Traceback (most recent call last):
+    ...
+    ValueError: Expected exactly one item in iterable, but got 1, 2,
+     and perhaps more.'
+    >>> only([1, 2], too_long=TypeError)  # doctest: +IGNORE_EXCEPTION_DETAIL
+    Traceback (most recent call last):
+    ...
+    TypeError
+
+    Note that :func:`only` attempts to advance *iterable* twice to ensure there
+    is only one item.  See :func:`spy` or :func:`peekable` to check
+    iterable contents less destructively.
+    """
+    it = iter(iterable)
+    first_value = next(it, default)
+
+    try:
+        second_value = next(it)
+    except StopIteration:
+        pass
+    else:
+        msg = (
+            'Expected exactly one item in iterable, but got {!r}, {!r}, '
+            'and perhaps more.'.format(first_value, second_value)
+        )
+        raise too_long or ValueError(msg)
+
+    return first_value
+
+
+def _ichunk(iterable, n):
+    cache = deque()
+    chunk = islice(iterable, n)
+
+    def generator():
+        while True:
+            if cache:
+                yield cache.popleft()
+            else:
+                try:
+                    item = next(chunk)
+                except StopIteration:
+                    return
+                else:
+                    yield item
+
+    def materialize_next(n=1):
+        # if n not specified materialize everything
+        if n is None:
+            cache.extend(chunk)
+            return len(cache)
+
+        to_cache = n - len(cache)
+
+        # materialize up to n
+        if to_cache > 0:
+            cache.extend(islice(chunk, to_cache))
+
+        # return number materialized up to n
+        return min(n, len(cache))
+
+    return (generator(), materialize_next)
+
+
+def ichunked(iterable, n):
+    """Break *iterable* into sub-iterables with *n* elements each.
+    :func:`ichunked` is like :func:`chunked`, but it yields iterables
+    instead of lists.
+
+    If the sub-iterables are read in order, the elements of *iterable*
+    won't be stored in memory.
+    If they are read out of order, :func:`itertools.tee` is used to cache
+    elements as necessary.
+
+    >>> from itertools import count
+    >>> all_chunks = ichunked(count(), 4)
+    >>> c_1, c_2, c_3 = next(all_chunks), next(all_chunks), next(all_chunks)
+    >>> list(c_2)  # c_1's elements have been cached; c_3's haven't been
+    [4, 5, 6, 7]
+    >>> list(c_1)
+    [0, 1, 2, 3]
+    >>> list(c_3)
+    [8, 9, 10, 11]
+
+    """
+    iterable = iter(iterable)
+    while True:
+        # Create new chunk
+        chunk, materialize_next = _ichunk(iterable, n)
+
+        # Check to see whether we're at the end of the source iterable
+        if not materialize_next():
+            return
+
+        yield chunk
+
+        # Fill previous chunk's cache
+        materialize_next(None)
+
+
+def iequals(*iterables):
+    """Return ``True`` if all given *iterables* are equal to each other,
+    which means that they contain the same elements in the same order.
+
+    The function is useful for comparing iterables of different data types
+    or iterables that do not support equality checks.
+
+    >>> iequals("abc", ['a', 'b', 'c'], ('a', 'b', 'c'), iter("abc"))
+    True
+
+    >>> iequals("abc", "acb")
+    False
+
+    Not to be confused with :func:`all_equal`, which checks whether all
+    elements of iterable are equal to each other.
+
+    """
+    return all(map(all_equal, zip_longest(*iterables, fillvalue=object())))
+
+
+def distinct_combinations(iterable, r):
+    """Yield the distinct combinations of *r* items taken from *iterable*.
+
+        >>> list(distinct_combinations([0, 0, 1], 2))
+        [(0, 0), (0, 1)]
+
+    Equivalent to ``set(combinations(iterable))``, except duplicates are not
+    generated and thrown away. For larger input sequences this is much more
+    efficient.
+
+    """
+    if r < 0:
+        raise ValueError('r must be non-negative')
+    elif r == 0:
+        yield ()
+        return
+    pool = tuple(iterable)
+    generators = [unique_everseen(enumerate(pool), key=itemgetter(1))]
+    current_combo = [None] * r
+    level = 0
+    while generators:
+        try:
+            cur_idx, p = next(generators[-1])
+        except StopIteration:
+            generators.pop()
+            level -= 1
+            continue
+        current_combo[level] = p
+        if level + 1 == r:
+            yield tuple(current_combo)
+        else:
+            generators.append(
+                unique_everseen(
+                    enumerate(pool[cur_idx + 1 :], cur_idx + 1),
+                    key=itemgetter(1),
+                )
+            )
+            level += 1
+
+
+def filter_except(validator, iterable, *exceptions):
+    """Yield the items from *iterable* for which the *validator* function does
+    not raise one of the specified *exceptions*.
+
+    *validator* is called for each item in *iterable*.
+    It should be a function that accepts one argument and raises an exception
+    if that item is not valid.
+
+    >>> iterable = ['1', '2', 'three', '4', None]
+    >>> list(filter_except(int, iterable, ValueError, TypeError))
+    ['1', '2', '4']
+
+    If an exception other than one given by *exceptions* is raised by
+    *validator*, it is raised like normal.
+    """
+    for item in iterable:
+        try:
+            validator(item)
+        except exceptions:
+            pass
+        else:
+            yield item
+
+
+def map_except(function, iterable, *exceptions):
+    """Transform each item from *iterable* with *function* and yield the
+    result, unless *function* raises one of the specified *exceptions*.
+
+    *function* is called to transform each item in *iterable*.
+    It should accept one argument.
+
+    >>> iterable = ['1', '2', 'three', '4', None]
+    >>> list(map_except(int, iterable, ValueError, TypeError))
+    [1, 2, 4]
+
+    If an exception other than one given by *exceptions* is raised by
+    *function*, it is raised like normal.
+    """
+    for item in iterable:
+        try:
+            yield function(item)
+        except exceptions:
+            pass
+
+
+def map_if(iterable, pred, func, func_else=lambda x: x):
+    """Evaluate each item from *iterable* using *pred*. If the result is
+    equivalent to ``True``, transform the item with *func* and yield it.
+    Otherwise, transform the item with *func_else* and yield it.
+
+    *pred*, *func*, and *func_else* should each be functions that accept
+    one argument. By default, *func_else* is the identity function.
+
+    >>> from math import sqrt
+    >>> iterable = list(range(-5, 5))
+    >>> iterable
+    [-5, -4, -3, -2, -1, 0, 1, 2, 3, 4]
+    >>> list(map_if(iterable, lambda x: x > 3, lambda x: 'toobig'))
+    [-5, -4, -3, -2, -1, 0, 1, 2, 3, 'toobig']
+    >>> list(map_if(iterable, lambda x: x >= 0,
+    ... lambda x: f'{sqrt(x):.2f}', lambda x: None))
+    [None, None, None, None, None, '0.00', '1.00', '1.41', '1.73', '2.00']
+    """
+    for item in iterable:
+        yield func(item) if pred(item) else func_else(item)
+
+
+def _sample_unweighted(iterable, k):
+    # Implementation of "Algorithm L" from the 1994 paper by Kim-Hung Li:
+    # "Reservoir-Sampling Algorithms of Time Complexity O(n(1+log(N/n)))".
+
+    # Fill up the reservoir (collection of samples) with the first `k` samples
+    reservoir = take(k, iterable)
+
+    # Generate random number that's the largest in a sample of k U(0,1) numbers
+    # Largest order statistic: https://en.wikipedia.org/wiki/Order_statistic
+    W = exp(log(random()) / k)
+
+    # The number of elements to skip before changing the reservoir is a random
+    # number with a geometric distribution. Sample it using random() and logs.
+    next_index = k + floor(log(random()) / log(1 - W))
+
+    for index, element in enumerate(iterable, k):
+        if index == next_index:
+            reservoir[randrange(k)] = element
+            # The new W is the largest in a sample of k U(0, `old_W`) numbers
+            W *= exp(log(random()) / k)
+            next_index += floor(log(random()) / log(1 - W)) + 1
+
+    return reservoir
+
+
+def _sample_weighted(iterable, k, weights):
+    # Implementation of "A-ExpJ" from the 2006 paper by Efraimidis et al. :
+    # "Weighted random sampling with a reservoir".
+
+    # Log-transform for numerical stability for weights that are small/large
+    weight_keys = (log(random()) / weight for weight in weights)
+
+    # Fill up the reservoir (collection of samples) with the first `k`
+    # weight-keys and elements, then heapify the list.
+    reservoir = take(k, zip(weight_keys, iterable))
+    heapify(reservoir)
+
+    # The number of jumps before changing the reservoir is a random variable
+    # with an exponential distribution. Sample it using random() and logs.
+    smallest_weight_key, _ = reservoir[0]
+    weights_to_skip = log(random()) / smallest_weight_key
+
+    for weight, element in zip(weights, iterable):
+        if weight >= weights_to_skip:
+            # The notation here is consistent with the paper, but we store
+            # the weight-keys in log-space for better numerical stability.
+            smallest_weight_key, _ = reservoir[0]
+            t_w = exp(weight * smallest_weight_key)
+            r_2 = uniform(t_w, 1)  # generate U(t_w, 1)
+            weight_key = log(r_2) / weight
+            heapreplace(reservoir, (weight_key, element))
+            smallest_weight_key, _ = reservoir[0]
+            weights_to_skip = log(random()) / smallest_weight_key
+        else:
+            weights_to_skip -= weight
+
+    # Equivalent to [element for weight_key, element in sorted(reservoir)]
+    return [heappop(reservoir)[1] for _ in range(k)]
+
+
+def sample(iterable, k, weights=None):
+    """Return a *k*-length list of elements chosen (without replacement)
+    from the *iterable*. Like :func:`random.sample`, but works on iterables
+    of unknown length.
+
+    >>> iterable = range(100)
+    >>> sample(iterable, 5)  # doctest: +SKIP
+    [81, 60, 96, 16, 4]
+
+    An iterable with *weights* may also be given:
+
+    >>> iterable = range(100)
+    >>> weights = (i * i + 1 for i in range(100))
+    >>> sampled = sample(iterable, 5, weights=weights)  # doctest: +SKIP
+    [79, 67, 74, 66, 78]
+
+    The algorithm can also be used to generate weighted random permutations.
+    The relative weight of each item determines the probability that it
+    appears late in the permutation.
+
+    >>> data = "abcdefgh"
+    >>> weights = range(1, len(data) + 1)
+    >>> sample(data, k=len(data), weights=weights)  # doctest: +SKIP
+    ['c', 'a', 'b', 'e', 'g', 'd', 'h', 'f']
+    """
+    if k == 0:
+        return []
+
+    iterable = iter(iterable)
+    if weights is None:
+        return _sample_unweighted(iterable, k)
+    else:
+        weights = iter(weights)
+        return _sample_weighted(iterable, k, weights)
+
+
+def is_sorted(iterable, key=None, reverse=False, strict=False):
+    """Returns ``True`` if the items of iterable are in sorted order, and
+    ``False`` otherwise. *key* and *reverse* have the same meaning that they do
+    in the built-in :func:`sorted` function.
+
+    >>> is_sorted(['1', '2', '3', '4', '5'], key=int)
+    True
+    >>> is_sorted([5, 4, 3, 1, 2], reverse=True)
+    False
+
+    If *strict*, tests for strict sorting, that is, returns ``False`` if equal
+    elements are found:
+
+    >>> is_sorted([1, 2, 2])
+    True
+    >>> is_sorted([1, 2, 2], strict=True)
+    False
+
+    The function returns ``False`` after encountering the first out-of-order
+    item. If there are no out-of-order items, the iterable is exhausted.
+    """
+
+    compare = (le if reverse else ge) if strict else (lt if reverse else gt)
+    it = iterable if key is None else map(key, iterable)
+    return not any(starmap(compare, pairwise(it)))
+
+
+class AbortThread(BaseException):
+    pass
+
+
+class callback_iter:
+    """Convert a function that uses callbacks to an iterator.
+
+    Let *func* be a function that takes a `callback` keyword argument.
+    For example:
+
+    >>> def func(callback=None):
+    ...     for i, c in [(1, 'a'), (2, 'b'), (3, 'c')]:
+    ...         if callback:
+    ...             callback(i, c)
+    ...     return 4
+
+
+    Use ``with callback_iter(func)`` to get an iterator over the parameters
+    that are delivered to the callback.
+
+    >>> with callback_iter(func) as it:
+    ...     for args, kwargs in it:
+    ...         print(args)
+    (1, 'a')
+    (2, 'b')
+    (3, 'c')
+
+    The function will be called in a background thread. The ``done`` property
+    indicates whether it has completed execution.
+
+    >>> it.done
+    True
+
+    If it completes successfully, its return value will be available
+    in the ``result`` property.
+
+    >>> it.result
+    4
+
+    Notes:
+
+    * If the function uses some keyword argument besides ``callback``, supply
+      *callback_kwd*.
+    * If it finished executing, but raised an exception, accessing the
+      ``result`` property will raise the same exception.
+    * If it hasn't finished executing, accessing the ``result``
+      property from within the ``with`` block will raise ``RuntimeError``.
+    * If it hasn't finished executing, accessing the ``result`` property from
+      outside the ``with`` block will raise a
+      ``more_itertools.AbortThread`` exception.
+    * Provide *wait_seconds* to adjust how frequently the it is polled for
+      output.
+
+    """
+
+    def __init__(self, func, callback_kwd='callback', wait_seconds=0.1):
+        self._func = func
+        self._callback_kwd = callback_kwd
+        self._aborted = False
+        self._future = None
+        self._wait_seconds = wait_seconds
+        # Lazily import concurrent.future
+        self._executor = __import__(
+            'concurrent.futures'
+        ).futures.ThreadPoolExecutor(max_workers=1)
+        self._iterator = self._reader()
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_value, traceback):
+        self._aborted = True
+        self._executor.shutdown()
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        return next(self._iterator)
+
+    @property
+    def done(self):
+        if self._future is None:
+            return False
+        return self._future.done()
+
+    @property
+    def result(self):
+        if not self.done:
+            raise RuntimeError('Function has not yet completed')
+
+        return self._future.result()
+
+    def _reader(self):
+        q = Queue()
+
+        def callback(*args, **kwargs):
+            if self._aborted:
+                raise AbortThread('canceled by user')
+
+            q.put((args, kwargs))
+
+        self._future = self._executor.submit(
+            self._func, **{self._callback_kwd: callback}
+        )
+
+        while True:
+            try:
+                item = q.get(timeout=self._wait_seconds)
+            except Empty:
+                pass
+            else:
+                q.task_done()
+                yield item
+
+            if self._future.done():
+                break
+
+        remaining = []
+        while True:
+            try:
+                item = q.get_nowait()
+            except Empty:
+                break
+            else:
+                q.task_done()
+                remaining.append(item)
+        q.join()
+        yield from remaining
+
+
+def windowed_complete(iterable, n):
+    """
+    Yield ``(beginning, middle, end)`` tuples, where:
+
+    * Each ``middle`` has *n* items from *iterable*
+    * Each ``beginning`` has the items before the ones in ``middle``
+    * Each ``end`` has the items after the ones in ``middle``
+
+    >>> iterable = range(7)
+    >>> n = 3
+    >>> for beginning, middle, end in windowed_complete(iterable, n):
+    ...     print(beginning, middle, end)
+    () (0, 1, 2) (3, 4, 5, 6)
+    (0,) (1, 2, 3) (4, 5, 6)
+    (0, 1) (2, 3, 4) (5, 6)
+    (0, 1, 2) (3, 4, 5) (6,)
+    (0, 1, 2, 3) (4, 5, 6) ()
+
+    Note that *n* must be at least 0 and most equal to the length of
+    *iterable*.
+
+    This function will exhaust the iterable and may require significant
+    storage.
+    """
+    if n < 0:
+        raise ValueError('n must be >= 0')
+
+    seq = tuple(iterable)
+    size = len(seq)
+
+    if n > size:
+        raise ValueError('n must be <= len(seq)')
+
+    for i in range(size - n + 1):
+        beginning = seq[:i]
+        middle = seq[i : i + n]
+        end = seq[i + n :]
+        yield beginning, middle, end
+
+
+def all_unique(iterable, key=None):
+    """
+    Returns ``True`` if all the elements of *iterable* are unique (no two
+    elements are equal).
+
+        >>> all_unique('ABCB')
+        False
+
+    If a *key* function is specified, it will be used to make comparisons.
+
+        >>> all_unique('ABCb')
+        True
+        >>> all_unique('ABCb', str.lower)
+        False
+
+    The function returns as soon as the first non-unique element is
+    encountered. Iterables with a mix of hashable and unhashable items can
+    be used, but the function will be slower for unhashable items.
+    """
+    seenset = set()
+    seenset_add = seenset.add
+    seenlist = []
+    seenlist_add = seenlist.append
+    for element in map(key, iterable) if key else iterable:
+        try:
+            if element in seenset:
+                return False
+            seenset_add(element)
+        except TypeError:
+            if element in seenlist:
+                return False
+            seenlist_add(element)
+    return True
+
+
+def nth_product(index, *args):
+    """Equivalent to ``list(product(*args))[index]``.
+
+    The products of *args* can be ordered lexicographically.
+    :func:`nth_product` computes the product at sort position *index* without
+    computing the previous products.
+
+        >>> nth_product(8, range(2), range(2), range(2), range(2))
+        (1, 0, 0, 0)
+
+    ``IndexError`` will be raised if the given *index* is invalid.
+    """
+    pools = list(map(tuple, reversed(args)))
+    ns = list(map(len, pools))
+
+    c = reduce(mul, ns)
+
+    if index < 0:
+        index += c
+
+    if not 0 <= index < c:
+        raise IndexError
+
+    result = []
+    for pool, n in zip(pools, ns):
+        result.append(pool[index % n])
+        index //= n
+
+    return tuple(reversed(result))
+
+
+def nth_permutation(iterable, r, index):
+    """Equivalent to ``list(permutations(iterable, r))[index]```
+
+    The subsequences of *iterable* that are of length *r* where order is
+    important can be ordered lexicographically. :func:`nth_permutation`
+    computes the subsequence at sort position *index* directly, without
+    computing the previous subsequences.
+
+        >>> nth_permutation('ghijk', 2, 5)
+        ('h', 'i')
+
+    ``ValueError`` will be raised If *r* is negative or greater than the length
+    of *iterable*.
+    ``IndexError`` will be raised if the given *index* is invalid.
+    """
+    pool = list(iterable)
+    n = len(pool)
+
+    if r is None or r == n:
+        r, c = n, factorial(n)
+    elif not 0 <= r < n:
+        raise ValueError
+    else:
+        c = perm(n, r)
+    assert c > 0  # factortial(n)>0, and r<n so perm(n,r) is never zero
+
+    if index < 0:
+        index += c
+
+    if not 0 <= index < c:
+        raise IndexError
+
+    result = [0] * r
+    q = index * factorial(n) // c if r < n else index
+    for d in range(1, n + 1):
+        q, i = divmod(q, d)
+        if 0 <= n - d < r:
+            result[n - d] = i
+        if q == 0:
+            break
+
+    return tuple(map(pool.pop, result))
+
+
+def nth_combination_with_replacement(iterable, r, index):
+    """Equivalent to
+    ``list(combinations_with_replacement(iterable, r))[index]``.
+
+
+    The subsequences with repetition of *iterable* that are of length *r* can
+    be ordered lexicographically. :func:`nth_combination_with_replacement`
+    computes the subsequence at sort position *index* directly, without
+    computing the previous subsequences with replacement.
+
+        >>> nth_combination_with_replacement(range(5), 3, 5)
+        (0, 1, 1)
+
+    ``ValueError`` will be raised If *r* is negative or greater than the length
+    of *iterable*.
+    ``IndexError`` will be raised if the given *index* is invalid.
+    """
+    pool = tuple(iterable)
+    n = len(pool)
+    if (r < 0) or (r > n):
+        raise ValueError
+
+    c = comb(n + r - 1, r)
+
+    if index < 0:
+        index += c
+
+    if (index < 0) or (index >= c):
+        raise IndexError
+
+    result = []
+    i = 0
+    while r:
+        r -= 1
+        while n >= 0:
+            num_combs = comb(n + r - 1, r)
+            if index < num_combs:
+                break
+            n -= 1
+            i += 1
+            index -= num_combs
+        result.append(pool[i])
+
+    return tuple(result)
+
+
+def value_chain(*args):
+    """Yield all arguments passed to the function in the same order in which
+    they were passed. If an argument itself is iterable then iterate over its
+    values.
+
+        >>> list(value_chain(1, 2, 3, [4, 5, 6]))
+        [1, 2, 3, 4, 5, 6]
+
+    Binary and text strings are not considered iterable and are emitted
+    as-is:
+
+        >>> list(value_chain('12', '34', ['56', '78']))
+        ['12', '34', '56', '78']
+
+    Pre- or postpend a single element to an iterable:
+
+        >>> list(value_chain(1, [2, 3, 4, 5, 6]))
+        [1, 2, 3, 4, 5, 6]
+        >>> list(value_chain([1, 2, 3, 4, 5], 6))
+        [1, 2, 3, 4, 5, 6]
+
+    Multiple levels of nesting are not flattened.
+
+    """
+    for value in args:
+        if isinstance(value, (str, bytes)):
+            yield value
+            continue
+        try:
+            yield from value
+        except TypeError:
+            yield value
+
+
+def product_index(element, *args):
+    """Equivalent to ``list(product(*args)).index(element)``
+
+    The products of *args* can be ordered lexicographically.
+    :func:`product_index` computes the first index of *element* without
+    computing the previous products.
+
+        >>> product_index([8, 2], range(10), range(5))
+        42
+
+    ``ValueError`` will be raised if the given *element* isn't in the product
+    of *args*.
+    """
+    index = 0
+
+    for x, pool in zip_longest(element, args, fillvalue=_marker):
+        if x is _marker or pool is _marker:
+            raise ValueError('element is not a product of args')
+
+        pool = tuple(pool)
+        index = index * len(pool) + pool.index(x)
+
+    return index
+
+
+def combination_index(element, iterable):
+    """Equivalent to ``list(combinations(iterable, r)).index(element)``
+
+    The subsequences of *iterable* that are of length *r* can be ordered
+    lexicographically. :func:`combination_index` computes the index of the
+    first *element*, without computing the previous combinations.
+
+        >>> combination_index('adf', 'abcdefg')
+        10
+
+    ``ValueError`` will be raised if the given *element* isn't one of the
+    combinations of *iterable*.
+    """
+    element = enumerate(element)
+    k, y = next(element, (None, None))
+    if k is None:
+        return 0
+
+    indexes = []
+    pool = enumerate(iterable)
+    for n, x in pool:
+        if x == y:
+            indexes.append(n)
+            tmp, y = next(element, (None, None))
+            if tmp is None:
+                break
+            else:
+                k = tmp
+    else:
+        raise ValueError('element is not a combination of iterable')
+
+    n, _ = last(pool, default=(n, None))
+
+    # Python versions below 3.8 don't have math.comb
+    index = 1
+    for i, j in enumerate(reversed(indexes), start=1):
+        j = n - j
+        if i <= j:
+            index += comb(j, i)
+
+    return comb(n + 1, k + 1) - index
+
+
+def combination_with_replacement_index(element, iterable):
+    """Equivalent to
+    ``list(combinations_with_replacement(iterable, r)).index(element)``
+
+    The subsequences with repetition of *iterable* that are of length *r* can
+    be ordered lexicographically. :func:`combination_with_replacement_index`
+    computes the index of the first *element*, without computing the previous
+    combinations with replacement.
+
+        >>> combination_with_replacement_index('adf', 'abcdefg')
+        20
+
+    ``ValueError`` will be raised if the given *element* isn't one of the
+    combinations with replacement of *iterable*.
+    """
+    element = tuple(element)
+    l = len(element)
+    element = enumerate(element)
+
+    k, y = next(element, (None, None))
+    if k is None:
+        return 0
+
+    indexes = []
+    pool = tuple(iterable)
+    for n, x in enumerate(pool):
+        while x == y:
+            indexes.append(n)
+            tmp, y = next(element, (None, None))
+            if tmp is None:
+                break
+            else:
+                k = tmp
+        if y is None:
+            break
+    else:
+        raise ValueError(
+            'element is not a combination with replacement of iterable'
+        )
+
+    n = len(pool)
+    occupations = [0] * n
+    for p in indexes:
+        occupations[p] += 1
+
+    index = 0
+    cumulative_sum = 0
+    for k in range(1, n):
+        cumulative_sum += occupations[k - 1]
+        j = l + n - 1 - k - cumulative_sum
+        i = n - k
+        if i <= j:
+            index += comb(j, i)
+
+    return index
+
+
+def permutation_index(element, iterable):
+    """Equivalent to ``list(permutations(iterable, r)).index(element)```
+
+    The subsequences of *iterable* that are of length *r* where order is
+    important can be ordered lexicographically. :func:`permutation_index`
+    computes the index of the first *element* directly, without computing
+    the previous permutations.
+
+        >>> permutation_index([1, 3, 2], range(5))
+        19
+
+    ``ValueError`` will be raised if the given *element* isn't one of the
+    permutations of *iterable*.
+    """
+    index = 0
+    pool = list(iterable)
+    for i, x in zip(range(len(pool), -1, -1), element):
+        r = pool.index(x)
+        index = index * i + r
+        del pool[r]
+
+    return index
+
+
+class countable:
+    """Wrap *iterable* and keep a count of how many items have been consumed.
+
+    The ``items_seen`` attribute starts at ``0`` and increments as the iterable
+    is consumed:
+
+        >>> iterable = map(str, range(10))
+        >>> it = countable(iterable)
+        >>> it.items_seen
+        0
+        >>> next(it), next(it)
+        ('0', '1')
+        >>> list(it)
+        ['2', '3', '4', '5', '6', '7', '8', '9']
+        >>> it.items_seen
+        10
+    """
+
+    def __init__(self, iterable):
+        self._it = iter(iterable)
+        self.items_seen = 0
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        item = next(self._it)
+        self.items_seen += 1
+
+        return item
+
+
+def chunked_even(iterable, n):
+    """Break *iterable* into lists of approximately length *n*.
+    Items are distributed such the lengths of the lists differ by at most
+    1 item.
+
+    >>> iterable = [1, 2, 3, 4, 5, 6, 7]
+    >>> n = 3
+    >>> list(chunked_even(iterable, n))  # List lengths: 3, 2, 2
+    [[1, 2, 3], [4, 5], [6, 7]]
+    >>> list(chunked(iterable, n))  # List lengths: 3, 3, 1
+    [[1, 2, 3], [4, 5, 6], [7]]
+
+    """
+    iterable = iter(iterable)
+
+    # Initialize a buffer to process the chunks while keeping
+    # some back to fill any underfilled chunks
+    min_buffer = (n - 1) * (n - 2)
+    buffer = list(islice(iterable, min_buffer))
+
+    # Append items until we have a completed chunk
+    for _ in islice(map(buffer.append, iterable), n, None, n):
+        yield buffer[:n]
+        del buffer[:n]
+
+    # Check if any chunks need addition processing
+    if not buffer:
+        return
+    length = len(buffer)
+
+    # Chunks are either size `full_size <= n` or `partial_size = full_size - 1`
+    q, r = divmod(length, n)
+    num_lists = q + (1 if r > 0 else 0)
+    q, r = divmod(length, num_lists)
+    full_size = q + (1 if r > 0 else 0)
+    partial_size = full_size - 1
+    num_full = length - partial_size * num_lists
+
+    # Yield chunks of full size
+    partial_start_idx = num_full * full_size
+    if full_size > 0:
+        for i in range(0, partial_start_idx, full_size):
+            yield buffer[i : i + full_size]
+
+    # Yield chunks of partial size
+    if partial_size > 0:
+        for i in range(partial_start_idx, length, partial_size):
+            yield buffer[i : i + partial_size]
+
+
+def zip_broadcast(*objects, scalar_types=(str, bytes), strict=False):
+    """A version of :func:`zip` that "broadcasts" any scalar
+    (i.e., non-iterable) items into output tuples.
+
+    >>> iterable_1 = [1, 2, 3]
+    >>> iterable_2 = ['a', 'b', 'c']
+    >>> scalar = '_'
+    >>> list(zip_broadcast(iterable_1, iterable_2, scalar))
+    [(1, 'a', '_'), (2, 'b', '_'), (3, 'c', '_')]
+
+    The *scalar_types* keyword argument determines what types are considered
+    scalar. It is set to ``(str, bytes)`` by default. Set it to ``None`` to
+    treat strings and byte strings as iterable:
+
+    >>> list(zip_broadcast('abc', 0, 'xyz', scalar_types=None))
+    [('a', 0, 'x'), ('b', 0, 'y'), ('c', 0, 'z')]
+
+    If the *strict* keyword argument is ``True``, then
+    ``UnequalIterablesError`` will be raised if any of the iterables have
+    different lengths.
+    """
+
+    def is_scalar(obj):
+        if scalar_types and isinstance(obj, scalar_types):
+            return True
+        try:
+            iter(obj)
+        except TypeError:
+            return True
+        else:
+            return False
+
+    size = len(objects)
+    if not size:
+        return
+
+    new_item = [None] * size
+    iterables, iterable_positions = [], []
+    for i, obj in enumerate(objects):
+        if is_scalar(obj):
+            new_item[i] = obj
+        else:
+            iterables.append(iter(obj))
+            iterable_positions.append(i)
+
+    if not iterables:
+        yield tuple(objects)
+        return
+
+    zipper = _zip_equal if strict else zip
+    for item in zipper(*iterables):
+        for i, new_item[i] in zip(iterable_positions, item):
+            pass
+        yield tuple(new_item)
+
+
+def unique_in_window(iterable, n, key=None):
+    """Yield the items from *iterable* that haven't been seen recently.
+    *n* is the size of the lookback window.
+
+        >>> iterable = [0, 1, 0, 2, 3, 0]
+        >>> n = 3
+        >>> list(unique_in_window(iterable, n))
+        [0, 1, 2, 3, 0]
+
+    The *key* function, if provided, will be used to determine uniqueness:
+
+        >>> list(unique_in_window('abAcda', 3, key=lambda x: x.lower()))
+        ['a', 'b', 'c', 'd', 'a']
+
+    The items in *iterable* must be hashable.
+
+    """
+    if n <= 0:
+        raise ValueError('n must be greater than 0')
+
+    window = deque(maxlen=n)
+    counts = defaultdict(int)
+    use_key = key is not None
+
+    for item in iterable:
+        if len(window) == n:
+            to_discard = window[0]
+            if counts[to_discard] == 1:
+                del counts[to_discard]
+            else:
+                counts[to_discard] -= 1
+
+        k = key(item) if use_key else item
+        if k not in counts:
+            yield item
+        counts[k] += 1
+        window.append(k)
+
+
+def duplicates_everseen(iterable, key=None):
+    """Yield duplicate elements after their first appearance.
+
+    >>> list(duplicates_everseen('mississippi'))
+    ['s', 'i', 's', 's', 'i', 'p', 'i']
+    >>> list(duplicates_everseen('AaaBbbCccAaa', str.lower))
+    ['a', 'a', 'b', 'b', 'c', 'c', 'A', 'a', 'a']
+
+    This function is analogous to :func:`unique_everseen` and is subject to
+    the same performance considerations.
+
+    """
+    seen_set = set()
+    seen_list = []
+    use_key = key is not None
+
+    for element in iterable:
+        k = key(element) if use_key else element
+        try:
+            if k not in seen_set:
+                seen_set.add(k)
+            else:
+                yield element
+        except TypeError:
+            if k not in seen_list:
+                seen_list.append(k)
+            else:
+                yield element
+
+
+def duplicates_justseen(iterable, key=None):
+    """Yields serially-duplicate elements after their first appearance.
+
+    >>> list(duplicates_justseen('mississippi'))
+    ['s', 's', 'p']
+    >>> list(duplicates_justseen('AaaBbbCccAaa', str.lower))
+    ['a', 'a', 'b', 'b', 'c', 'c', 'a', 'a']
+
+    This function is analogous to :func:`unique_justseen`.
+
+    """
+    return flatten(g for _, g in groupby(iterable, key) for _ in g)
+
+
+def classify_unique(iterable, key=None):
+    """Classify each element in terms of its uniqueness.
+
+    For each element in the input iterable, return a 3-tuple consisting of:
+
+    1. The element itself
+    2. ``False`` if the element is equal to the one preceding it in the input,
+       ``True`` otherwise (i.e. the equivalent of :func:`unique_justseen`)
+    3. ``False`` if this element has been seen anywhere in the input before,
+       ``True`` otherwise (i.e. the equivalent of :func:`unique_everseen`)
+
+    >>> list(classify_unique('otto'))    # doctest: +NORMALIZE_WHITESPACE
+    [('o', True,  True),
+     ('t', True,  True),
+     ('t', False, False),
+     ('o', True,  False)]
+
+    This function is analogous to :func:`unique_everseen` and is subject to
+    the same performance considerations.
+
+    """
+    seen_set = set()
+    seen_list = []
+    use_key = key is not None
+    previous = None
+
+    for i, element in enumerate(iterable):
+        k = key(element) if use_key else element
+        is_unique_justseen = not i or previous != k
+        previous = k
+        is_unique_everseen = False
+        try:
+            if k not in seen_set:
+                seen_set.add(k)
+                is_unique_everseen = True
+        except TypeError:
+            if k not in seen_list:
+                seen_list.append(k)
+                is_unique_everseen = True
+        yield element, is_unique_justseen, is_unique_everseen
+
+
+def minmax(iterable_or_value, *others, key=None, default=_marker):
+    """Returns both the smallest and largest items in an iterable
+    or the largest of two or more arguments.
+
+        >>> minmax([3, 1, 5])
+        (1, 5)
+
+        >>> minmax(4, 2, 6)
+        (2, 6)
+
+    If a *key* function is provided, it will be used to transform the input
+    items for comparison.
+
+        >>> minmax([5, 30], key=str)  # '30' sorts before '5'
+        (30, 5)
+
+    If a *default* value is provided, it will be returned if there are no
+    input items.
+
+        >>> minmax([], default=(0, 0))
+        (0, 0)
+
+    Otherwise ``ValueError`` is raised.
+
+    This function is based on the
+    `recipe <http://code.activestate.com/recipes/577916/>`__ by
+    Raymond Hettinger and takes care to minimize the number of comparisons
+    performed.
+    """
+    iterable = (iterable_or_value, *others) if others else iterable_or_value
+
+    it = iter(iterable)
+
+    try:
+        lo = hi = next(it)
+    except StopIteration as exc:
+        if default is _marker:
+            raise ValueError(
+                '`minmax()` argument is an empty iterable. '
+                'Provide a `default` value to suppress this error.'
+            ) from exc
+        return default
+
+    # Different branches depending on the presence of key. This saves a lot
+    # of unimportant copies which would slow the "key=None" branch
+    # significantly down.
+    if key is None:
+        for x, y in zip_longest(it, it, fillvalue=lo):
+            if y < x:
+                x, y = y, x
+            if x < lo:
+                lo = x
+            if hi < y:
+                hi = y
+
+    else:
+        lo_key = hi_key = key(lo)
+
+        for x, y in zip_longest(it, it, fillvalue=lo):
+            x_key, y_key = key(x), key(y)
+
+            if y_key < x_key:
+                x, y, x_key, y_key = y, x, y_key, x_key
+            if x_key < lo_key:
+                lo, lo_key = x, x_key
+            if hi_key < y_key:
+                hi, hi_key = y, y_key
+
+    return lo, hi
+
+
+def constrained_batches(
+    iterable, max_size, max_count=None, get_len=len, strict=True
+):
+    """Yield batches of items from *iterable* with a combined size limited by
+    *max_size*.
+
+    >>> iterable = [b'12345', b'123', b'12345678', b'1', b'1', b'12', b'1']
+    >>> list(constrained_batches(iterable, 10))
+    [(b'12345', b'123'), (b'12345678', b'1', b'1'), (b'12', b'1')]
+
+    If a *max_count* is supplied, the number of items per batch is also
+    limited:
+
+    >>> iterable = [b'12345', b'123', b'12345678', b'1', b'1', b'12', b'1']
+    >>> list(constrained_batches(iterable, 10, max_count = 2))
+    [(b'12345', b'123'), (b'12345678', b'1'), (b'1', b'12'), (b'1',)]
+
+    If a *get_len* function is supplied, use that instead of :func:`len` to
+    determine item size.
+
+    If *strict* is ``True``, raise ``ValueError`` if any single item is bigger
+    than *max_size*. Otherwise, allow single items to exceed *max_size*.
+    """
+    if max_size <= 0:
+        raise ValueError('maximum size must be greater than zero')
+
+    batch = []
+    batch_size = 0
+    batch_count = 0
+    for item in iterable:
+        item_len = get_len(item)
+        if strict and item_len > max_size:
+            raise ValueError('item size exceeds maximum size')
+
+        reached_count = batch_count == max_count
+        reached_size = item_len + batch_size > max_size
+        if batch_count and (reached_size or reached_count):
+            yield tuple(batch)
+            batch.clear()
+            batch_size = 0
+            batch_count = 0
+
+        batch.append(item)
+        batch_size += item_len
+        batch_count += 1
+
+    if batch:
+        yield tuple(batch)
+
+
+def gray_product(*iterables):
+    """Like :func:`itertools.product`, but return tuples in an order such
+    that only one element in the generated tuple changes from one iteration
+    to the next.
+
+        >>> list(gray_product('AB','CD'))
+        [('A', 'C'), ('B', 'C'), ('B', 'D'), ('A', 'D')]
+
+    This function consumes all of the input iterables before producing output.
+    If any of the input iterables have fewer than two items, ``ValueError``
+    is raised.
+
+    For information on the algorithm, see
+    `this section <https://www-cs-faculty.stanford.edu/~knuth/fasc2a.ps.gz>`__
+    of Donald Knuth's *The Art of Computer Programming*.
+    """
+    all_iterables = tuple(tuple(x) for x in iterables)
+    iterable_count = len(all_iterables)
+    for iterable in all_iterables:
+        if len(iterable) < 2:
+            raise ValueError("each iterable must have two or more items")
+
+    # This is based on "Algorithm H" from section 7.2.1.1, page 20.
+    # a holds the indexes of the source iterables for the n-tuple to be yielded
+    # f is the array of "focus pointers"
+    # o is the array of "directions"
+    a = [0] * iterable_count
+    f = list(range(iterable_count + 1))
+    o = [1] * iterable_count
+    while True:
+        yield tuple(all_iterables[i][a[i]] for i in range(iterable_count))
+        j = f[0]
+        f[0] = 0
+        if j == iterable_count:
+            break
+        a[j] = a[j] + o[j]
+        if a[j] == 0 or a[j] == len(all_iterables[j]) - 1:
+            o[j] = -o[j]
+            f[j] = f[j + 1]
+            f[j + 1] = j + 1
+
+
+def partial_product(*iterables):
+    """Yields tuples containing one item from each iterator, with subsequent
+    tuples changing a single item at a time by advancing each iterator until it
+    is exhausted. This sequence guarantees every value in each iterable is
+    output at least once without generating all possible combinations.
+
+    This may be useful, for example, when testing an expensive function.
+
+        >>> list(partial_product('AB', 'C', 'DEF'))
+        [('A', 'C', 'D'), ('B', 'C', 'D'), ('B', 'C', 'E'), ('B', 'C', 'F')]
+    """
+
+    iterators = list(map(iter, iterables))
+
+    try:
+        prod = [next(it) for it in iterators]
+    except StopIteration:
+        return
+    yield tuple(prod)
+
+    for i, it in enumerate(iterators):
+        for prod[i] in it:
+            yield tuple(prod)
+
+
+def takewhile_inclusive(predicate, iterable):
+    """A variant of :func:`takewhile` that yields one additional element.
+
+        >>> list(takewhile_inclusive(lambda x: x < 5, [1, 4, 6, 4, 1]))
+        [1, 4, 6]
+
+    :func:`takewhile` would return ``[1, 4]``.
+    """
+    for x in iterable:
+        yield x
+        if not predicate(x):
+            break
+
+
+def outer_product(func, xs, ys, *args, **kwargs):
+    """A generalized outer product that applies a binary function to all
+    pairs of items. Returns a 2D matrix with ``len(xs)`` rows and ``len(ys)``
+    columns.
+    Also accepts ``*args`` and ``**kwargs`` that are passed to ``func``.
+
+    Multiplication table:
+
+    >>> list(outer_product(mul, range(1, 4), range(1, 6)))
+    [(1, 2, 3, 4, 5), (2, 4, 6, 8, 10), (3, 6, 9, 12, 15)]
+
+    Cross tabulation:
+
+    >>> xs = ['A', 'B', 'A', 'A', 'B', 'B', 'A', 'A', 'B', 'B']
+    >>> ys = ['X', 'X', 'X', 'Y', 'Z', 'Z', 'Y', 'Y', 'Z', 'Z']
+    >>> rows = list(zip(xs, ys))
+    >>> count_rows = lambda x, y: rows.count((x, y))
+    >>> list(outer_product(count_rows, sorted(set(xs)), sorted(set(ys))))
+    [(2, 3, 0), (1, 0, 4)]
+
+    Usage with ``*args`` and ``**kwargs``:
+
+    >>> animals = ['cat', 'wolf', 'mouse']
+    >>> list(outer_product(min, animals, animals, key=len))
+    [('cat', 'cat', 'cat'), ('cat', 'wolf', 'wolf'), ('cat', 'wolf', 'mouse')]
+    """
+    ys = tuple(ys)
+    return batched(
+        starmap(lambda x, y: func(x, y, *args, **kwargs), product(xs, ys)),
+        n=len(ys),
+    )
+
+
+def iter_suppress(iterable, *exceptions):
+    """Yield each of the items from *iterable*. If the iteration raises one of
+    the specified *exceptions*, that exception will be suppressed and iteration
+    will stop.
+
+    >>> from itertools import chain
+    >>> def breaks_at_five(x):
+    ...     while True:
+    ...         if x >= 5:
+    ...             raise RuntimeError
+    ...         yield x
+    ...         x += 1
+    >>> it_1 = iter_suppress(breaks_at_five(1), RuntimeError)
+    >>> it_2 = iter_suppress(breaks_at_five(2), RuntimeError)
+    >>> list(chain(it_1, it_2))
+    [1, 2, 3, 4, 2, 3, 4]
+    """
+    try:
+        yield from iterable
+    except exceptions:
+        return
+
+
+def filter_map(func, iterable):
+    """Apply *func* to every element of *iterable*, yielding only those which
+    are not ``None``.
+
+    >>> elems = ['1', 'a', '2', 'b', '3']
+    >>> list(filter_map(lambda s: int(s) if s.isnumeric() else None, elems))
+    [1, 2, 3]
+    """
+    for x in iterable:
+        y = func(x)
+        if y is not None:
+            yield y
+
+
+def powerset_of_sets(iterable):
+    """Yields all possible subsets of the iterable.
+
+        >>> list(powerset_of_sets([1, 2, 3]))  # doctest: +SKIP
+        [set(), {1}, {2}, {3}, {1, 2}, {1, 3}, {2, 3}, {1, 2, 3}]
+        >>> list(powerset_of_sets([1, 1, 0]))  # doctest: +SKIP
+        [set(), {1}, {0}, {0, 1}]
+
+    :func:`powerset_of_sets` takes care to minimize the number
+    of hash operations performed.
+    """
+    sets = tuple(map(set, dict.fromkeys(map(frozenset, zip(iterable)))))
+    for r in range(len(sets) + 1):
+        yield from starmap(set().union, combinations(sets, r))
+
+
+def join_mappings(**field_to_map):
+    """
+    Joins multiple mappings together using their common keys.
+
+    >>> user_scores = {'elliot': 50, 'claris': 60}
+    >>> user_times = {'elliot': 30, 'claris': 40}
+    >>> join_mappings(score=user_scores, time=user_times)
+    {'elliot': {'score': 50, 'time': 30}, 'claris': {'score': 60, 'time': 40}}
+    """
+    ret = defaultdict(dict)
+
+    for field_name, mapping in field_to_map.items():
+        for key, value in mapping.items():
+            ret[key][field_name] = value
+
+    return dict(ret)
+
+
+def _complex_sumprod(v1, v2):
+    """High precision sumprod() for complex numbers.
+    Used by :func:`dft` and :func:`idft`.
+    """
+
+    r1 = chain((p.real for p in v1), (-p.imag for p in v1))
+    r2 = chain((q.real for q in v2), (q.imag for q in v2))
+    i1 = chain((p.real for p in v1), (p.imag for p in v1))
+    i2 = chain((q.imag for q in v2), (q.real for q in v2))
+    return complex(_fsumprod(r1, r2), _fsumprod(i1, i2))
+
+
+def dft(xarr):
+    """Discrete Fourier Tranform. *xarr* is a sequence of complex numbers.
+    Yields the components of the corresponding transformed output vector.
+
+    >>> import cmath
+    >>> xarr = [1, 2-1j, -1j, -1+2j]
+    >>> Xarr = [2, -2-2j, -2j, 4+4j]
+    >>> all(map(cmath.isclose, dft(xarr), Xarr))
+    True
+
+    See :func:`idft` for the inverse Discrete Fourier Transform.
+    """
+    N = len(xarr)
+    roots_of_unity = [e ** (n / N * tau * -1j) for n in range(N)]
+    for k in range(N):
+        coeffs = [roots_of_unity[k * n % N] for n in range(N)]
+        yield _complex_sumprod(xarr, coeffs)
+
+
+def idft(Xarr):
+    """Inverse Discrete Fourier Tranform. *Xarr* is a sequence of
+    complex numbers. Yields the components of the corresponding
+    inverse-transformed output vector.
+
+    >>> import cmath
+    >>> xarr = [1, 2-1j, -1j, -1+2j]
+    >>> Xarr = [2, -2-2j, -2j, 4+4j]
+    >>> all(map(cmath.isclose, idft(Xarr), xarr))
+    True
+
+    See :func:`dft` for the Discrete Fourier Transform.
+    """
+    N = len(Xarr)
+    roots_of_unity = [e ** (n / N * tau * 1j) for n in range(N)]
+    for k in range(N):
+        coeffs = [roots_of_unity[k * n % N] for n in range(N)]
+        yield _complex_sumprod(Xarr, coeffs) / N
+
+
+def doublestarmap(func, iterable):
+    """Apply *func* to every item of *iterable* by dictionary unpacking
+    the item into *func*.
+
+    The difference between :func:`itertools.starmap` and :func:`doublestarmap`
+    parallels the distinction between ``func(*a)`` and ``func(**a)``.
+
+    >>> iterable = [{'a': 1, 'b': 2}, {'a': 40, 'b': 60}]
+    >>> list(doublestarmap(lambda a, b: a + b, iterable))
+    [3, 100]
+
+    ``TypeError`` will be raised if *func*'s signature doesn't match the
+    mapping contained in *iterable* or if *iterable* does not contain mappings.
+    """
+    for item in iterable:
+        yield func(**item)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.pyi b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.pyi
new file mode 100644
index 00000000..e9460232
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/more.pyi
@@ -0,0 +1,709 @@
+"""Stubs for more_itertools.more"""
+
+from __future__ import annotations
+
+from types import TracebackType
+from typing import (
+    Any,
+    Callable,
+    Container,
+    ContextManager,
+    Generic,
+    Hashable,
+    Mapping,
+    Iterable,
+    Iterator,
+    Mapping,
+    overload,
+    Reversible,
+    Sequence,
+    Sized,
+    Type,
+    TypeVar,
+    type_check_only,
+)
+from typing_extensions import Protocol
+
+# Type and type variable definitions
+_T = TypeVar('_T')
+_T1 = TypeVar('_T1')
+_T2 = TypeVar('_T2')
+_U = TypeVar('_U')
+_V = TypeVar('_V')
+_W = TypeVar('_W')
+_T_co = TypeVar('_T_co', covariant=True)
+_GenFn = TypeVar('_GenFn', bound=Callable[..., Iterator[Any]])
+_Raisable = BaseException | Type[BaseException]
+
+@type_check_only
+class _SizedIterable(Protocol[_T_co], Sized, Iterable[_T_co]): ...
+
+@type_check_only
+class _SizedReversible(Protocol[_T_co], Sized, Reversible[_T_co]): ...
+
+@type_check_only
+class _SupportsSlicing(Protocol[_T_co]):
+    def __getitem__(self, __k: slice) -> _T_co: ...
+
+def chunked(
+    iterable: Iterable[_T], n: int | None, strict: bool = ...
+) -> Iterator[list[_T]]: ...
+@overload
+def first(iterable: Iterable[_T]) -> _T: ...
+@overload
+def first(iterable: Iterable[_T], default: _U) -> _T | _U: ...
+@overload
+def last(iterable: Iterable[_T]) -> _T: ...
+@overload
+def last(iterable: Iterable[_T], default: _U) -> _T | _U: ...
+@overload
+def nth_or_last(iterable: Iterable[_T], n: int) -> _T: ...
+@overload
+def nth_or_last(iterable: Iterable[_T], n: int, default: _U) -> _T | _U: ...
+
+class peekable(Generic[_T], Iterator[_T]):
+    def __init__(self, iterable: Iterable[_T]) -> None: ...
+    def __iter__(self) -> peekable[_T]: ...
+    def __bool__(self) -> bool: ...
+    @overload
+    def peek(self) -> _T: ...
+    @overload
+    def peek(self, default: _U) -> _T | _U: ...
+    def prepend(self, *items: _T) -> None: ...
+    def __next__(self) -> _T: ...
+    @overload
+    def __getitem__(self, index: int) -> _T: ...
+    @overload
+    def __getitem__(self, index: slice) -> list[_T]: ...
+
+def consumer(func: _GenFn) -> _GenFn: ...
+def ilen(iterable: Iterable[_T]) -> int: ...
+def iterate(func: Callable[[_T], _T], start: _T) -> Iterator[_T]: ...
+def with_iter(
+    context_manager: ContextManager[Iterable[_T]],
+) -> Iterator[_T]: ...
+def one(
+    iterable: Iterable[_T],
+    too_short: _Raisable | None = ...,
+    too_long: _Raisable | None = ...,
+) -> _T: ...
+def raise_(exception: _Raisable, *args: Any) -> None: ...
+def strictly_n(
+    iterable: Iterable[_T],
+    n: int,
+    too_short: _GenFn | None = ...,
+    too_long: _GenFn | None = ...,
+) -> list[_T]: ...
+def distinct_permutations(
+    iterable: Iterable[_T], r: int | None = ...
+) -> Iterator[tuple[_T, ...]]: ...
+def intersperse(
+    e: _U, iterable: Iterable[_T], n: int = ...
+) -> Iterator[_T | _U]: ...
+def unique_to_each(*iterables: Iterable[_T]) -> list[list[_T]]: ...
+@overload
+def windowed(
+    seq: Iterable[_T], n: int, *, step: int = ...
+) -> Iterator[tuple[_T | None, ...]]: ...
+@overload
+def windowed(
+    seq: Iterable[_T], n: int, fillvalue: _U, step: int = ...
+) -> Iterator[tuple[_T | _U, ...]]: ...
+def substrings(iterable: Iterable[_T]) -> Iterator[tuple[_T, ...]]: ...
+def substrings_indexes(
+    seq: Sequence[_T], reverse: bool = ...
+) -> Iterator[tuple[Sequence[_T], int, int]]: ...
+
+class bucket(Generic[_T, _U], Container[_U]):
+    def __init__(
+        self,
+        iterable: Iterable[_T],
+        key: Callable[[_T], _U],
+        validator: Callable[[_U], object] | None = ...,
+    ) -> None: ...
+    def __contains__(self, value: object) -> bool: ...
+    def __iter__(self) -> Iterator[_U]: ...
+    def __getitem__(self, value: object) -> Iterator[_T]: ...
+
+def spy(
+    iterable: Iterable[_T], n: int = ...
+) -> tuple[list[_T], Iterator[_T]]: ...
+def interleave(*iterables: Iterable[_T]) -> Iterator[_T]: ...
+def interleave_longest(*iterables: Iterable[_T]) -> Iterator[_T]: ...
+def interleave_evenly(
+    iterables: list[Iterable[_T]], lengths: list[int] | None = ...
+) -> Iterator[_T]: ...
+def collapse(
+    iterable: Iterable[Any],
+    base_type: type | None = ...,
+    levels: int | None = ...,
+) -> Iterator[Any]: ...
+@overload
+def side_effect(
+    func: Callable[[_T], object],
+    iterable: Iterable[_T],
+    chunk_size: None = ...,
+    before: Callable[[], object] | None = ...,
+    after: Callable[[], object] | None = ...,
+) -> Iterator[_T]: ...
+@overload
+def side_effect(
+    func: Callable[[list[_T]], object],
+    iterable: Iterable[_T],
+    chunk_size: int,
+    before: Callable[[], object] | None = ...,
+    after: Callable[[], object] | None = ...,
+) -> Iterator[_T]: ...
+def sliced(
+    seq: _SupportsSlicing[_T], n: int, strict: bool = ...
+) -> Iterator[_T]: ...
+def split_at(
+    iterable: Iterable[_T],
+    pred: Callable[[_T], object],
+    maxsplit: int = ...,
+    keep_separator: bool = ...,
+) -> Iterator[list[_T]]: ...
+def split_before(
+    iterable: Iterable[_T], pred: Callable[[_T], object], maxsplit: int = ...
+) -> Iterator[list[_T]]: ...
+def split_after(
+    iterable: Iterable[_T], pred: Callable[[_T], object], maxsplit: int = ...
+) -> Iterator[list[_T]]: ...
+def split_when(
+    iterable: Iterable[_T],
+    pred: Callable[[_T, _T], object],
+    maxsplit: int = ...,
+) -> Iterator[list[_T]]: ...
+def split_into(
+    iterable: Iterable[_T], sizes: Iterable[int | None]
+) -> Iterator[list[_T]]: ...
+@overload
+def padded(
+    iterable: Iterable[_T],
+    *,
+    n: int | None = ...,
+    next_multiple: bool = ...,
+) -> Iterator[_T | None]: ...
+@overload
+def padded(
+    iterable: Iterable[_T],
+    fillvalue: _U,
+    n: int | None = ...,
+    next_multiple: bool = ...,
+) -> Iterator[_T | _U]: ...
+@overload
+def repeat_last(iterable: Iterable[_T]) -> Iterator[_T]: ...
+@overload
+def repeat_last(iterable: Iterable[_T], default: _U) -> Iterator[_T | _U]: ...
+def distribute(n: int, iterable: Iterable[_T]) -> list[Iterator[_T]]: ...
+@overload
+def stagger(
+    iterable: Iterable[_T],
+    offsets: _SizedIterable[int] = ...,
+    longest: bool = ...,
+) -> Iterator[tuple[_T | None, ...]]: ...
+@overload
+def stagger(
+    iterable: Iterable[_T],
+    offsets: _SizedIterable[int] = ...,
+    longest: bool = ...,
+    fillvalue: _U = ...,
+) -> Iterator[tuple[_T | _U, ...]]: ...
+
+class UnequalIterablesError(ValueError):
+    def __init__(self, details: tuple[int, int, int] | None = ...) -> None: ...
+
+@overload
+def zip_equal(__iter1: Iterable[_T1]) -> Iterator[tuple[_T1]]: ...
+@overload
+def zip_equal(
+    __iter1: Iterable[_T1], __iter2: Iterable[_T2]
+) -> Iterator[tuple[_T1, _T2]]: ...
+@overload
+def zip_equal(
+    __iter1: Iterable[_T],
+    __iter2: Iterable[_T],
+    __iter3: Iterable[_T],
+    *iterables: Iterable[_T],
+) -> Iterator[tuple[_T, ...]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T1],
+    *,
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: None = None,
+) -> Iterator[tuple[_T1 | None]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T1],
+    __iter2: Iterable[_T2],
+    *,
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: None = None,
+) -> Iterator[tuple[_T1 | None, _T2 | None]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T],
+    __iter2: Iterable[_T],
+    __iter3: Iterable[_T],
+    *iterables: Iterable[_T],
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: None = None,
+) -> Iterator[tuple[_T | None, ...]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T1],
+    *,
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: _U,
+) -> Iterator[tuple[_T1 | _U]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T1],
+    __iter2: Iterable[_T2],
+    *,
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: _U,
+) -> Iterator[tuple[_T1 | _U, _T2 | _U]]: ...
+@overload
+def zip_offset(
+    __iter1: Iterable[_T],
+    __iter2: Iterable[_T],
+    __iter3: Iterable[_T],
+    *iterables: Iterable[_T],
+    offsets: _SizedIterable[int],
+    longest: bool = ...,
+    fillvalue: _U,
+) -> Iterator[tuple[_T | _U, ...]]: ...
+def sort_together(
+    iterables: Iterable[Iterable[_T]],
+    key_list: Iterable[int] = ...,
+    key: Callable[..., Any] | None = ...,
+    reverse: bool = ...,
+) -> list[tuple[_T, ...]]: ...
+def unzip(iterable: Iterable[Sequence[_T]]) -> tuple[Iterator[_T], ...]: ...
+def divide(n: int, iterable: Iterable[_T]) -> list[Iterator[_T]]: ...
+def always_iterable(
+    obj: object,
+    base_type: type | tuple[type | tuple[Any, ...], ...] | None = ...,
+) -> Iterator[Any]: ...
+def adjacent(
+    predicate: Callable[[_T], bool],
+    iterable: Iterable[_T],
+    distance: int = ...,
+) -> Iterator[tuple[bool, _T]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: None = None,
+    valuefunc: None = None,
+    reducefunc: None = None,
+) -> Iterator[tuple[_T, Iterator[_T]]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: None,
+    reducefunc: None,
+) -> Iterator[tuple[_U, Iterator[_T]]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: None,
+    valuefunc: Callable[[_T], _V],
+    reducefunc: None,
+) -> Iterable[tuple[_T, Iterable[_V]]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: Callable[[_T], _V],
+    reducefunc: None,
+) -> Iterable[tuple[_U, Iterator[_V]]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: None,
+    valuefunc: None,
+    reducefunc: Callable[[Iterator[_T]], _W],
+) -> Iterable[tuple[_T, _W]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: None,
+    reducefunc: Callable[[Iterator[_T]], _W],
+) -> Iterable[tuple[_U, _W]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: None,
+    valuefunc: Callable[[_T], _V],
+    reducefunc: Callable[[Iterable[_V]], _W],
+) -> Iterable[tuple[_T, _W]]: ...
+@overload
+def groupby_transform(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: Callable[[_T], _V],
+    reducefunc: Callable[[Iterable[_V]], _W],
+) -> Iterable[tuple[_U, _W]]: ...
+
+class numeric_range(Generic[_T, _U], Sequence[_T], Hashable, Reversible[_T]):
+    @overload
+    def __init__(self, __stop: _T) -> None: ...
+    @overload
+    def __init__(self, __start: _T, __stop: _T) -> None: ...
+    @overload
+    def __init__(self, __start: _T, __stop: _T, __step: _U) -> None: ...
+    def __bool__(self) -> bool: ...
+    def __contains__(self, elem: object) -> bool: ...
+    def __eq__(self, other: object) -> bool: ...
+    @overload
+    def __getitem__(self, key: int) -> _T: ...
+    @overload
+    def __getitem__(self, key: slice) -> numeric_range[_T, _U]: ...
+    def __hash__(self) -> int: ...
+    def __iter__(self) -> Iterator[_T]: ...
+    def __len__(self) -> int: ...
+    def __reduce__(
+        self,
+    ) -> tuple[Type[numeric_range[_T, _U]], tuple[_T, _T, _U]]: ...
+    def __repr__(self) -> str: ...
+    def __reversed__(self) -> Iterator[_T]: ...
+    def count(self, value: _T) -> int: ...
+    def index(self, value: _T) -> int: ...  # type: ignore
+
+def count_cycle(
+    iterable: Iterable[_T], n: int | None = ...
+) -> Iterable[tuple[int, _T]]: ...
+def mark_ends(
+    iterable: Iterable[_T],
+) -> Iterable[tuple[bool, bool, _T]]: ...
+def locate(
+    iterable: Iterable[_T],
+    pred: Callable[..., Any] = ...,
+    window_size: int | None = ...,
+) -> Iterator[int]: ...
+def lstrip(
+    iterable: Iterable[_T], pred: Callable[[_T], object]
+) -> Iterator[_T]: ...
+def rstrip(
+    iterable: Iterable[_T], pred: Callable[[_T], object]
+) -> Iterator[_T]: ...
+def strip(
+    iterable: Iterable[_T], pred: Callable[[_T], object]
+) -> Iterator[_T]: ...
+
+class islice_extended(Generic[_T], Iterator[_T]):
+    def __init__(self, iterable: Iterable[_T], *args: int | None) -> None: ...
+    def __iter__(self) -> islice_extended[_T]: ...
+    def __next__(self) -> _T: ...
+    def __getitem__(self, index: slice) -> islice_extended[_T]: ...
+
+def always_reversible(iterable: Iterable[_T]) -> Iterator[_T]: ...
+def consecutive_groups(
+    iterable: Iterable[_T], ordering: Callable[[_T], int] = ...
+) -> Iterator[Iterator[_T]]: ...
+@overload
+def difference(
+    iterable: Iterable[_T],
+    func: Callable[[_T, _T], _U] = ...,
+    *,
+    initial: None = ...,
+) -> Iterator[_T | _U]: ...
+@overload
+def difference(
+    iterable: Iterable[_T], func: Callable[[_T, _T], _U] = ..., *, initial: _U
+) -> Iterator[_U]: ...
+
+class SequenceView(Generic[_T], Sequence[_T]):
+    def __init__(self, target: Sequence[_T]) -> None: ...
+    @overload
+    def __getitem__(self, index: int) -> _T: ...
+    @overload
+    def __getitem__(self, index: slice) -> Sequence[_T]: ...
+    def __len__(self) -> int: ...
+
+class seekable(Generic[_T], Iterator[_T]):
+    def __init__(
+        self, iterable: Iterable[_T], maxlen: int | None = ...
+    ) -> None: ...
+    def __iter__(self) -> seekable[_T]: ...
+    def __next__(self) -> _T: ...
+    def __bool__(self) -> bool: ...
+    @overload
+    def peek(self) -> _T: ...
+    @overload
+    def peek(self, default: _U) -> _T | _U: ...
+    def elements(self) -> SequenceView[_T]: ...
+    def seek(self, index: int) -> None: ...
+    def relative_seek(self, count: int) -> None: ...
+
+class run_length:
+    @staticmethod
+    def encode(iterable: Iterable[_T]) -> Iterator[tuple[_T, int]]: ...
+    @staticmethod
+    def decode(iterable: Iterable[tuple[_T, int]]) -> Iterator[_T]: ...
+
+def exactly_n(
+    iterable: Iterable[_T], n: int, predicate: Callable[[_T], object] = ...
+) -> bool: ...
+def circular_shifts(iterable: Iterable[_T]) -> list[tuple[_T, ...]]: ...
+def make_decorator(
+    wrapping_func: Callable[..., _U], result_index: int = ...
+) -> Callable[..., Callable[[Callable[..., Any]], Callable[..., _U]]]: ...
+@overload
+def map_reduce(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: None = ...,
+    reducefunc: None = ...,
+) -> dict[_U, list[_T]]: ...
+@overload
+def map_reduce(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: Callable[[_T], _V],
+    reducefunc: None = ...,
+) -> dict[_U, list[_V]]: ...
+@overload
+def map_reduce(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: None = ...,
+    reducefunc: Callable[[list[_T]], _W] = ...,
+) -> dict[_U, _W]: ...
+@overload
+def map_reduce(
+    iterable: Iterable[_T],
+    keyfunc: Callable[[_T], _U],
+    valuefunc: Callable[[_T], _V],
+    reducefunc: Callable[[list[_V]], _W],
+) -> dict[_U, _W]: ...
+def rlocate(
+    iterable: Iterable[_T],
+    pred: Callable[..., object] = ...,
+    window_size: int | None = ...,
+) -> Iterator[int]: ...
+def replace(
+    iterable: Iterable[_T],
+    pred: Callable[..., object],
+    substitutes: Iterable[_U],
+    count: int | None = ...,
+    window_size: int = ...,
+) -> Iterator[_T | _U]: ...
+def partitions(iterable: Iterable[_T]) -> Iterator[list[list[_T]]]: ...
+def set_partitions(
+    iterable: Iterable[_T], k: int | None = ...
+) -> Iterator[list[list[_T]]]: ...
+
+class time_limited(Generic[_T], Iterator[_T]):
+    def __init__(
+        self, limit_seconds: float, iterable: Iterable[_T]
+    ) -> None: ...
+    def __iter__(self) -> islice_extended[_T]: ...
+    def __next__(self) -> _T: ...
+
+@overload
+def only(
+    iterable: Iterable[_T], *, too_long: _Raisable | None = ...
+) -> _T | None: ...
+@overload
+def only(
+    iterable: Iterable[_T], default: _U, too_long: _Raisable | None = ...
+) -> _T | _U: ...
+def ichunked(iterable: Iterable[_T], n: int) -> Iterator[Iterator[_T]]: ...
+def distinct_combinations(
+    iterable: Iterable[_T], r: int
+) -> Iterator[tuple[_T, ...]]: ...
+def filter_except(
+    validator: Callable[[Any], object],
+    iterable: Iterable[_T],
+    *exceptions: Type[BaseException],
+) -> Iterator[_T]: ...
+def map_except(
+    function: Callable[[Any], _U],
+    iterable: Iterable[_T],
+    *exceptions: Type[BaseException],
+) -> Iterator[_U]: ...
+def map_if(
+    iterable: Iterable[Any],
+    pred: Callable[[Any], bool],
+    func: Callable[[Any], Any],
+    func_else: Callable[[Any], Any] | None = ...,
+) -> Iterator[Any]: ...
+def sample(
+    iterable: Iterable[_T],
+    k: int,
+    weights: Iterable[float] | None = ...,
+) -> list[_T]: ...
+def is_sorted(
+    iterable: Iterable[_T],
+    key: Callable[[_T], _U] | None = ...,
+    reverse: bool = False,
+    strict: bool = False,
+) -> bool: ...
+
+class AbortThread(BaseException):
+    pass
+
+class callback_iter(Generic[_T], Iterator[_T]):
+    def __init__(
+        self,
+        func: Callable[..., Any],
+        callback_kwd: str = ...,
+        wait_seconds: float = ...,
+    ) -> None: ...
+    def __enter__(self) -> callback_iter[_T]: ...
+    def __exit__(
+        self,
+        exc_type: Type[BaseException] | None,
+        exc_value: BaseException | None,
+        traceback: TracebackType | None,
+    ) -> bool | None: ...
+    def __iter__(self) -> callback_iter[_T]: ...
+    def __next__(self) -> _T: ...
+    def _reader(self) -> Iterator[_T]: ...
+    @property
+    def done(self) -> bool: ...
+    @property
+    def result(self) -> Any: ...
+
+def windowed_complete(
+    iterable: Iterable[_T], n: int
+) -> Iterator[tuple[_T, ...]]: ...
+def all_unique(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> bool: ...
+def nth_product(index: int, *args: Iterable[_T]) -> tuple[_T, ...]: ...
+def nth_combination_with_replacement(
+    iterable: Iterable[_T], r: int, index: int
+) -> tuple[_T, ...]: ...
+def nth_permutation(
+    iterable: Iterable[_T], r: int, index: int
+) -> tuple[_T, ...]: ...
+def value_chain(*args: _T | Iterable[_T]) -> Iterable[_T]: ...
+def product_index(element: Iterable[_T], *args: Iterable[_T]) -> int: ...
+def combination_index(
+    element: Iterable[_T], iterable: Iterable[_T]
+) -> int: ...
+def combination_with_replacement_index(
+    element: Iterable[_T], iterable: Iterable[_T]
+) -> int: ...
+def permutation_index(
+    element: Iterable[_T], iterable: Iterable[_T]
+) -> int: ...
+def repeat_each(iterable: Iterable[_T], n: int = ...) -> Iterator[_T]: ...
+
+class countable(Generic[_T], Iterator[_T]):
+    def __init__(self, iterable: Iterable[_T]) -> None: ...
+    def __iter__(self) -> countable[_T]: ...
+    def __next__(self) -> _T: ...
+    items_seen: int
+
+def chunked_even(iterable: Iterable[_T], n: int) -> Iterator[list[_T]]: ...
+def zip_broadcast(
+    *objects: _T | Iterable[_T],
+    scalar_types: type | tuple[type | tuple[Any, ...], ...] | None = ...,
+    strict: bool = ...,
+) -> Iterable[tuple[_T, ...]]: ...
+def unique_in_window(
+    iterable: Iterable[_T], n: int, key: Callable[[_T], _U] | None = ...
+) -> Iterator[_T]: ...
+def duplicates_everseen(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> Iterator[_T]: ...
+def duplicates_justseen(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> Iterator[_T]: ...
+def classify_unique(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> Iterator[tuple[_T, bool, bool]]: ...
+
+class _SupportsLessThan(Protocol):
+    def __lt__(self, __other: Any) -> bool: ...
+
+_SupportsLessThanT = TypeVar("_SupportsLessThanT", bound=_SupportsLessThan)
+
+@overload
+def minmax(
+    iterable_or_value: Iterable[_SupportsLessThanT], *, key: None = None
+) -> tuple[_SupportsLessThanT, _SupportsLessThanT]: ...
+@overload
+def minmax(
+    iterable_or_value: Iterable[_T], *, key: Callable[[_T], _SupportsLessThan]
+) -> tuple[_T, _T]: ...
+@overload
+def minmax(
+    iterable_or_value: Iterable[_SupportsLessThanT],
+    *,
+    key: None = None,
+    default: _U,
+) -> _U | tuple[_SupportsLessThanT, _SupportsLessThanT]: ...
+@overload
+def minmax(
+    iterable_or_value: Iterable[_T],
+    *,
+    key: Callable[[_T], _SupportsLessThan],
+    default: _U,
+) -> _U | tuple[_T, _T]: ...
+@overload
+def minmax(
+    iterable_or_value: _SupportsLessThanT,
+    __other: _SupportsLessThanT,
+    *others: _SupportsLessThanT,
+) -> tuple[_SupportsLessThanT, _SupportsLessThanT]: ...
+@overload
+def minmax(
+    iterable_or_value: _T,
+    __other: _T,
+    *others: _T,
+    key: Callable[[_T], _SupportsLessThan],
+) -> tuple[_T, _T]: ...
+def longest_common_prefix(
+    iterables: Iterable[Iterable[_T]],
+) -> Iterator[_T]: ...
+def iequals(*iterables: Iterable[Any]) -> bool: ...
+def constrained_batches(
+    iterable: Iterable[_T],
+    max_size: int,
+    max_count: int | None = ...,
+    get_len: Callable[[_T], object] = ...,
+    strict: bool = ...,
+) -> Iterator[tuple[_T]]: ...
+def gray_product(*iterables: Iterable[_T]) -> Iterator[tuple[_T, ...]]: ...
+def partial_product(*iterables: Iterable[_T]) -> Iterator[tuple[_T, ...]]: ...
+def takewhile_inclusive(
+    predicate: Callable[[_T], bool], iterable: Iterable[_T]
+) -> Iterator[_T]: ...
+def outer_product(
+    func: Callable[[_T, _U], _V],
+    xs: Iterable[_T],
+    ys: Iterable[_U],
+    *args: Any,
+    **kwargs: Any,
+) -> Iterator[tuple[_V, ...]]: ...
+def iter_suppress(
+    iterable: Iterable[_T],
+    *exceptions: Type[BaseException],
+) -> Iterator[_T]: ...
+def filter_map(
+    func: Callable[[_T], _V | None],
+    iterable: Iterable[_T],
+) -> Iterator[_V]: ...
+def powerset_of_sets(iterable: Iterable[_T]) -> Iterator[set[_T]]: ...
+def join_mappings(
+    **field_to_map: Mapping[_T, _V]
+) -> dict[_T, dict[str, _V]]: ...
+def doublestarmap(
+    func: Callable[..., _T],
+    iterable: Iterable[Mapping[str, Any]],
+) -> Iterator[_T]: ...
+def dft(xarr: Sequence[complex]) -> Iterator[complex]: ...
+def idft(Xarr: Sequence[complex]) -> Iterator[complex]: ...
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.py
new file mode 100644
index 00000000..b32fa955
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.py
@@ -0,0 +1,1046 @@
+"""Imported from the recipes section of the itertools documentation.
+
+All functions taken from the recipes section of the itertools library docs
+[1]_.
+Some backward-compatible usability improvements have been made.
+
+.. [1] http://docs.python.org/library/itertools.html#recipes
+
+"""
+
+import math
+import operator
+
+from collections import deque
+from collections.abc import Sized
+from functools import partial, reduce
+from itertools import (
+    chain,
+    combinations,
+    compress,
+    count,
+    cycle,
+    groupby,
+    islice,
+    product,
+    repeat,
+    starmap,
+    tee,
+    zip_longest,
+)
+from random import randrange, sample, choice
+from sys import hexversion
+
+__all__ = [
+    'all_equal',
+    'batched',
+    'before_and_after',
+    'consume',
+    'convolve',
+    'dotproduct',
+    'first_true',
+    'factor',
+    'flatten',
+    'grouper',
+    'iter_except',
+    'iter_index',
+    'matmul',
+    'ncycles',
+    'nth',
+    'nth_combination',
+    'padnone',
+    'pad_none',
+    'pairwise',
+    'partition',
+    'polynomial_eval',
+    'polynomial_from_roots',
+    'polynomial_derivative',
+    'powerset',
+    'prepend',
+    'quantify',
+    'reshape',
+    'random_combination_with_replacement',
+    'random_combination',
+    'random_permutation',
+    'random_product',
+    'repeatfunc',
+    'roundrobin',
+    'sieve',
+    'sliding_window',
+    'subslices',
+    'sum_of_squares',
+    'tabulate',
+    'tail',
+    'take',
+    'totient',
+    'transpose',
+    'triplewise',
+    'unique',
+    'unique_everseen',
+    'unique_justseen',
+]
+
+_marker = object()
+
+
+# zip with strict is available for Python 3.10+
+try:
+    zip(strict=True)
+except TypeError:
+    _zip_strict = zip
+else:
+    _zip_strict = partial(zip, strict=True)
+
+# math.sumprod is available for Python 3.12+
+_sumprod = getattr(math, 'sumprod', lambda x, y: dotproduct(x, y))
+
+
+def take(n, iterable):
+    """Return first *n* items of the iterable as a list.
+
+        >>> take(3, range(10))
+        [0, 1, 2]
+
+    If there are fewer than *n* items in the iterable, all of them are
+    returned.
+
+        >>> take(10, range(3))
+        [0, 1, 2]
+
+    """
+    return list(islice(iterable, n))
+
+
+def tabulate(function, start=0):
+    """Return an iterator over the results of ``func(start)``,
+    ``func(start + 1)``, ``func(start + 2)``...
+
+    *func* should be a function that accepts one integer argument.
+
+    If *start* is not specified it defaults to 0. It will be incremented each
+    time the iterator is advanced.
+
+        >>> square = lambda x: x ** 2
+        >>> iterator = tabulate(square, -3)
+        >>> take(4, iterator)
+        [9, 4, 1, 0]
+
+    """
+    return map(function, count(start))
+
+
+def tail(n, iterable):
+    """Return an iterator over the last *n* items of *iterable*.
+
+    >>> t = tail(3, 'ABCDEFG')
+    >>> list(t)
+    ['E', 'F', 'G']
+
+    """
+    # If the given iterable has a length, then we can use islice to get its
+    # final elements. Note that if the iterable is not actually Iterable,
+    # either islice or deque will throw a TypeError. This is why we don't
+    # check if it is Iterable.
+    if isinstance(iterable, Sized):
+        yield from islice(iterable, max(0, len(iterable) - n), None)
+    else:
+        yield from iter(deque(iterable, maxlen=n))
+
+
+def consume(iterator, n=None):
+    """Advance *iterable* by *n* steps. If *n* is ``None``, consume it
+    entirely.
+
+    Efficiently exhausts an iterator without returning values. Defaults to
+    consuming the whole iterator, but an optional second argument may be
+    provided to limit consumption.
+
+        >>> i = (x for x in range(10))
+        >>> next(i)
+        0
+        >>> consume(i, 3)
+        >>> next(i)
+        4
+        >>> consume(i)
+        >>> next(i)
+        Traceback (most recent call last):
+          File "<stdin>", line 1, in <module>
+        StopIteration
+
+    If the iterator has fewer items remaining than the provided limit, the
+    whole iterator will be consumed.
+
+        >>> i = (x for x in range(3))
+        >>> consume(i, 5)
+        >>> next(i)
+        Traceback (most recent call last):
+          File "<stdin>", line 1, in <module>
+        StopIteration
+
+    """
+    # Use functions that consume iterators at C speed.
+    if n is None:
+        # feed the entire iterator into a zero-length deque
+        deque(iterator, maxlen=0)
+    else:
+        # advance to the empty slice starting at position n
+        next(islice(iterator, n, n), None)
+
+
+def nth(iterable, n, default=None):
+    """Returns the nth item or a default value.
+
+    >>> l = range(10)
+    >>> nth(l, 3)
+    3
+    >>> nth(l, 20, "zebra")
+    'zebra'
+
+    """
+    return next(islice(iterable, n, None), default)
+
+
+def all_equal(iterable, key=None):
+    """
+    Returns ``True`` if all the elements are equal to each other.
+
+        >>> all_equal('aaaa')
+        True
+        >>> all_equal('aaab')
+        False
+
+    A function that accepts a single argument and returns a transformed version
+    of each input item can be specified with *key*:
+
+        >>> all_equal('AaaA', key=str.casefold)
+        True
+        >>> all_equal([1, 2, 3], key=lambda x: x < 10)
+        True
+
+    """
+    return len(list(islice(groupby(iterable, key), 2))) <= 1
+
+
+def quantify(iterable, pred=bool):
+    """Return the how many times the predicate is true.
+
+    >>> quantify([True, False, True])
+    2
+
+    """
+    return sum(map(pred, iterable))
+
+
+def pad_none(iterable):
+    """Returns the sequence of elements and then returns ``None`` indefinitely.
+
+        >>> take(5, pad_none(range(3)))
+        [0, 1, 2, None, None]
+
+    Useful for emulating the behavior of the built-in :func:`map` function.
+
+    See also :func:`padded`.
+
+    """
+    return chain(iterable, repeat(None))
+
+
+padnone = pad_none
+
+
+def ncycles(iterable, n):
+    """Returns the sequence elements *n* times
+
+    >>> list(ncycles(["a", "b"], 3))
+    ['a', 'b', 'a', 'b', 'a', 'b']
+
+    """
+    return chain.from_iterable(repeat(tuple(iterable), n))
+
+
+def dotproduct(vec1, vec2):
+    """Returns the dot product of the two iterables.
+
+    >>> dotproduct([10, 10], [20, 20])
+    400
+
+    """
+    return sum(map(operator.mul, vec1, vec2))
+
+
+def flatten(listOfLists):
+    """Return an iterator flattening one level of nesting in a list of lists.
+
+        >>> list(flatten([[0, 1], [2, 3]]))
+        [0, 1, 2, 3]
+
+    See also :func:`collapse`, which can flatten multiple levels of nesting.
+
+    """
+    return chain.from_iterable(listOfLists)
+
+
+def repeatfunc(func, times=None, *args):
+    """Call *func* with *args* repeatedly, returning an iterable over the
+    results.
+
+    If *times* is specified, the iterable will terminate after that many
+    repetitions:
+
+        >>> from operator import add
+        >>> times = 4
+        >>> args = 3, 5
+        >>> list(repeatfunc(add, times, *args))
+        [8, 8, 8, 8]
+
+    If *times* is ``None`` the iterable will not terminate:
+
+        >>> from random import randrange
+        >>> times = None
+        >>> args = 1, 11
+        >>> take(6, repeatfunc(randrange, times, *args))  # doctest:+SKIP
+        [2, 4, 8, 1, 8, 4]
+
+    """
+    if times is None:
+        return starmap(func, repeat(args))
+    return starmap(func, repeat(args, times))
+
+
+def _pairwise(iterable):
+    """Returns an iterator of paired items, overlapping, from the original
+
+    >>> take(4, pairwise(count()))
+    [(0, 1), (1, 2), (2, 3), (3, 4)]
+
+    On Python 3.10 and above, this is an alias for :func:`itertools.pairwise`.
+
+    """
+    a, b = tee(iterable)
+    next(b, None)
+    return zip(a, b)
+
+
+try:
+    from itertools import pairwise as itertools_pairwise
+except ImportError:
+    pairwise = _pairwise
+else:
+
+    def pairwise(iterable):
+        return itertools_pairwise(iterable)
+
+    pairwise.__doc__ = _pairwise.__doc__
+
+
+class UnequalIterablesError(ValueError):
+    def __init__(self, details=None):
+        msg = 'Iterables have different lengths'
+        if details is not None:
+            msg += (': index 0 has length {}; index {} has length {}').format(
+                *details
+            )
+
+        super().__init__(msg)
+
+
+def _zip_equal_generator(iterables):
+    for combo in zip_longest(*iterables, fillvalue=_marker):
+        for val in combo:
+            if val is _marker:
+                raise UnequalIterablesError()
+        yield combo
+
+
+def _zip_equal(*iterables):
+    # Check whether the iterables are all the same size.
+    try:
+        first_size = len(iterables[0])
+        for i, it in enumerate(iterables[1:], 1):
+            size = len(it)
+            if size != first_size:
+                raise UnequalIterablesError(details=(first_size, i, size))
+        # All sizes are equal, we can use the built-in zip.
+        return zip(*iterables)
+    # If any one of the iterables didn't have a length, start reading
+    # them until one runs out.
+    except TypeError:
+        return _zip_equal_generator(iterables)
+
+
+def grouper(iterable, n, incomplete='fill', fillvalue=None):
+    """Group elements from *iterable* into fixed-length groups of length *n*.
+
+    >>> list(grouper('ABCDEF', 3))
+    [('A', 'B', 'C'), ('D', 'E', 'F')]
+
+    The keyword arguments *incomplete* and *fillvalue* control what happens for
+    iterables whose length is not a multiple of *n*.
+
+    When *incomplete* is `'fill'`, the last group will contain instances of
+    *fillvalue*.
+
+    >>> list(grouper('ABCDEFG', 3, incomplete='fill', fillvalue='x'))
+    [('A', 'B', 'C'), ('D', 'E', 'F'), ('G', 'x', 'x')]
+
+    When *incomplete* is `'ignore'`, the last group will not be emitted.
+
+    >>> list(grouper('ABCDEFG', 3, incomplete='ignore', fillvalue='x'))
+    [('A', 'B', 'C'), ('D', 'E', 'F')]
+
+    When *incomplete* is `'strict'`, a subclass of `ValueError` will be raised.
+
+    >>> it = grouper('ABCDEFG', 3, incomplete='strict')
+    >>> list(it)  # doctest: +IGNORE_EXCEPTION_DETAIL
+    Traceback (most recent call last):
+    ...
+    UnequalIterablesError
+
+    """
+    args = [iter(iterable)] * n
+    if incomplete == 'fill':
+        return zip_longest(*args, fillvalue=fillvalue)
+    if incomplete == 'strict':
+        return _zip_equal(*args)
+    if incomplete == 'ignore':
+        return zip(*args)
+    else:
+        raise ValueError('Expected fill, strict, or ignore')
+
+
+def roundrobin(*iterables):
+    """Yields an item from each iterable, alternating between them.
+
+        >>> list(roundrobin('ABC', 'D', 'EF'))
+        ['A', 'D', 'E', 'B', 'F', 'C']
+
+    This function produces the same output as :func:`interleave_longest`, but
+    may perform better for some inputs (in particular when the number of
+    iterables is small).
+
+    """
+    # Algorithm credited to George Sakkis
+    iterators = map(iter, iterables)
+    for num_active in range(len(iterables), 0, -1):
+        iterators = cycle(islice(iterators, num_active))
+        yield from map(next, iterators)
+
+
+def partition(pred, iterable):
+    """
+    Returns a 2-tuple of iterables derived from the input iterable.
+    The first yields the items that have ``pred(item) == False``.
+    The second yields the items that have ``pred(item) == True``.
+
+        >>> is_odd = lambda x: x % 2 != 0
+        >>> iterable = range(10)
+        >>> even_items, odd_items = partition(is_odd, iterable)
+        >>> list(even_items), list(odd_items)
+        ([0, 2, 4, 6, 8], [1, 3, 5, 7, 9])
+
+    If *pred* is None, :func:`bool` is used.
+
+        >>> iterable = [0, 1, False, True, '', ' ']
+        >>> false_items, true_items = partition(None, iterable)
+        >>> list(false_items), list(true_items)
+        ([0, False, ''], [1, True, ' '])
+
+    """
+    if pred is None:
+        pred = bool
+
+    t1, t2, p = tee(iterable, 3)
+    p1, p2 = tee(map(pred, p))
+    return (compress(t1, map(operator.not_, p1)), compress(t2, p2))
+
+
+def powerset(iterable):
+    """Yields all possible subsets of the iterable.
+
+        >>> list(powerset([1, 2, 3]))
+        [(), (1,), (2,), (3,), (1, 2), (1, 3), (2, 3), (1, 2, 3)]
+
+    :func:`powerset` will operate on iterables that aren't :class:`set`
+    instances, so repeated elements in the input will produce repeated elements
+    in the output.
+
+        >>> seq = [1, 1, 0]
+        >>> list(powerset(seq))
+        [(), (1,), (1,), (0,), (1, 1), (1, 0), (1, 0), (1, 1, 0)]
+
+    For a variant that efficiently yields actual :class:`set` instances, see
+    :func:`powerset_of_sets`.
+    """
+    s = list(iterable)
+    return chain.from_iterable(combinations(s, r) for r in range(len(s) + 1))
+
+
+def unique_everseen(iterable, key=None):
+    """
+    Yield unique elements, preserving order.
+
+        >>> list(unique_everseen('AAAABBBCCDAABBB'))
+        ['A', 'B', 'C', 'D']
+        >>> list(unique_everseen('ABBCcAD', str.lower))
+        ['A', 'B', 'C', 'D']
+
+    Sequences with a mix of hashable and unhashable items can be used.
+    The function will be slower (i.e., `O(n^2)`) for unhashable items.
+
+    Remember that ``list`` objects are unhashable - you can use the *key*
+    parameter to transform the list to a tuple (which is hashable) to
+    avoid a slowdown.
+
+        >>> iterable = ([1, 2], [2, 3], [1, 2])
+        >>> list(unique_everseen(iterable))  # Slow
+        [[1, 2], [2, 3]]
+        >>> list(unique_everseen(iterable, key=tuple))  # Faster
+        [[1, 2], [2, 3]]
+
+    Similarly, you may want to convert unhashable ``set`` objects with
+    ``key=frozenset``. For ``dict`` objects,
+    ``key=lambda x: frozenset(x.items())`` can be used.
+
+    """
+    seenset = set()
+    seenset_add = seenset.add
+    seenlist = []
+    seenlist_add = seenlist.append
+    use_key = key is not None
+
+    for element in iterable:
+        k = key(element) if use_key else element
+        try:
+            if k not in seenset:
+                seenset_add(k)
+                yield element
+        except TypeError:
+            if k not in seenlist:
+                seenlist_add(k)
+                yield element
+
+
+def unique_justseen(iterable, key=None):
+    """Yields elements in order, ignoring serial duplicates
+
+    >>> list(unique_justseen('AAAABBBCCDAABBB'))
+    ['A', 'B', 'C', 'D', 'A', 'B']
+    >>> list(unique_justseen('ABBCcAD', str.lower))
+    ['A', 'B', 'C', 'A', 'D']
+
+    """
+    if key is None:
+        return map(operator.itemgetter(0), groupby(iterable))
+
+    return map(next, map(operator.itemgetter(1), groupby(iterable, key)))
+
+
+def unique(iterable, key=None, reverse=False):
+    """Yields unique elements in sorted order.
+
+    >>> list(unique([[1, 2], [3, 4], [1, 2]]))
+    [[1, 2], [3, 4]]
+
+    *key* and *reverse* are passed to :func:`sorted`.
+
+    >>> list(unique('ABBcCAD', str.casefold))
+    ['A', 'B', 'c', 'D']
+    >>> list(unique('ABBcCAD', str.casefold, reverse=True))
+    ['D', 'c', 'B', 'A']
+
+    The elements in *iterable* need not be hashable, but they must be
+    comparable for sorting to work.
+    """
+    return unique_justseen(sorted(iterable, key=key, reverse=reverse), key=key)
+
+
+def iter_except(func, exception, first=None):
+    """Yields results from a function repeatedly until an exception is raised.
+
+    Converts a call-until-exception interface to an iterator interface.
+    Like ``iter(func, sentinel)``, but uses an exception instead of a sentinel
+    to end the loop.
+
+        >>> l = [0, 1, 2]
+        >>> list(iter_except(l.pop, IndexError))
+        [2, 1, 0]
+
+    Multiple exceptions can be specified as a stopping condition:
+
+        >>> l = [1, 2, 3, '...', 4, 5, 6]
+        >>> list(iter_except(lambda: 1 + l.pop(), (IndexError, TypeError)))
+        [7, 6, 5]
+        >>> list(iter_except(lambda: 1 + l.pop(), (IndexError, TypeError)))
+        [4, 3, 2]
+        >>> list(iter_except(lambda: 1 + l.pop(), (IndexError, TypeError)))
+        []
+
+    """
+    try:
+        if first is not None:
+            yield first()
+        while 1:
+            yield func()
+    except exception:
+        pass
+
+
+def first_true(iterable, default=None, pred=None):
+    """
+    Returns the first true value in the iterable.
+
+    If no true value is found, returns *default*
+
+    If *pred* is not None, returns the first item for which
+    ``pred(item) == True`` .
+
+        >>> first_true(range(10))
+        1
+        >>> first_true(range(10), pred=lambda x: x > 5)
+        6
+        >>> first_true(range(10), default='missing', pred=lambda x: x > 9)
+        'missing'
+
+    """
+    return next(filter(pred, iterable), default)
+
+
+def random_product(*args, repeat=1):
+    """Draw an item at random from each of the input iterables.
+
+        >>> random_product('abc', range(4), 'XYZ')  # doctest:+SKIP
+        ('c', 3, 'Z')
+
+    If *repeat* is provided as a keyword argument, that many items will be
+    drawn from each iterable.
+
+        >>> random_product('abcd', range(4), repeat=2)  # doctest:+SKIP
+        ('a', 2, 'd', 3)
+
+    This equivalent to taking a random selection from
+    ``itertools.product(*args, **kwarg)``.
+
+    """
+    pools = [tuple(pool) for pool in args] * repeat
+    return tuple(choice(pool) for pool in pools)
+
+
+def random_permutation(iterable, r=None):
+    """Return a random *r* length permutation of the elements in *iterable*.
+
+    If *r* is not specified or is ``None``, then *r* defaults to the length of
+    *iterable*.
+
+        >>> random_permutation(range(5))  # doctest:+SKIP
+        (3, 4, 0, 1, 2)
+
+    This equivalent to taking a random selection from
+    ``itertools.permutations(iterable, r)``.
+
+    """
+    pool = tuple(iterable)
+    r = len(pool) if r is None else r
+    return tuple(sample(pool, r))
+
+
+def random_combination(iterable, r):
+    """Return a random *r* length subsequence of the elements in *iterable*.
+
+        >>> random_combination(range(5), 3)  # doctest:+SKIP
+        (2, 3, 4)
+
+    This equivalent to taking a random selection from
+    ``itertools.combinations(iterable, r)``.
+
+    """
+    pool = tuple(iterable)
+    n = len(pool)
+    indices = sorted(sample(range(n), r))
+    return tuple(pool[i] for i in indices)
+
+
+def random_combination_with_replacement(iterable, r):
+    """Return a random *r* length subsequence of elements in *iterable*,
+    allowing individual elements to be repeated.
+
+        >>> random_combination_with_replacement(range(3), 5) # doctest:+SKIP
+        (0, 0, 1, 2, 2)
+
+    This equivalent to taking a random selection from
+    ``itertools.combinations_with_replacement(iterable, r)``.
+
+    """
+    pool = tuple(iterable)
+    n = len(pool)
+    indices = sorted(randrange(n) for i in range(r))
+    return tuple(pool[i] for i in indices)
+
+
+def nth_combination(iterable, r, index):
+    """Equivalent to ``list(combinations(iterable, r))[index]``.
+
+    The subsequences of *iterable* that are of length *r* can be ordered
+    lexicographically. :func:`nth_combination` computes the subsequence at
+    sort position *index* directly, without computing the previous
+    subsequences.
+
+        >>> nth_combination(range(5), 3, 5)
+        (0, 3, 4)
+
+    ``ValueError`` will be raised If *r* is negative or greater than the length
+    of *iterable*.
+    ``IndexError`` will be raised if the given *index* is invalid.
+    """
+    pool = tuple(iterable)
+    n = len(pool)
+    if (r < 0) or (r > n):
+        raise ValueError
+
+    c = 1
+    k = min(r, n - r)
+    for i in range(1, k + 1):
+        c = c * (n - k + i) // i
+
+    if index < 0:
+        index += c
+
+    if (index < 0) or (index >= c):
+        raise IndexError
+
+    result = []
+    while r:
+        c, n, r = c * r // n, n - 1, r - 1
+        while index >= c:
+            index -= c
+            c, n = c * (n - r) // n, n - 1
+        result.append(pool[-1 - n])
+
+    return tuple(result)
+
+
+def prepend(value, iterator):
+    """Yield *value*, followed by the elements in *iterator*.
+
+        >>> value = '0'
+        >>> iterator = ['1', '2', '3']
+        >>> list(prepend(value, iterator))
+        ['0', '1', '2', '3']
+
+    To prepend multiple values, see :func:`itertools.chain`
+    or :func:`value_chain`.
+
+    """
+    return chain([value], iterator)
+
+
+def convolve(signal, kernel):
+    """Convolve the iterable *signal* with the iterable *kernel*.
+
+        >>> signal = (1, 2, 3, 4, 5)
+        >>> kernel = [3, 2, 1]
+        >>> list(convolve(signal, kernel))
+        [3, 8, 14, 20, 26, 14, 5]
+
+    Note: the input arguments are not interchangeable, as the *kernel*
+    is immediately consumed and stored.
+
+    """
+    # This implementation intentionally doesn't match the one in the itertools
+    # documentation.
+    kernel = tuple(kernel)[::-1]
+    n = len(kernel)
+    window = deque([0], maxlen=n) * n
+    for x in chain(signal, repeat(0, n - 1)):
+        window.append(x)
+        yield _sumprod(kernel, window)
+
+
+def before_and_after(predicate, it):
+    """A variant of :func:`takewhile` that allows complete access to the
+    remainder of the iterator.
+
+         >>> it = iter('ABCdEfGhI')
+         >>> all_upper, remainder = before_and_after(str.isupper, it)
+         >>> ''.join(all_upper)
+         'ABC'
+         >>> ''.join(remainder) # takewhile() would lose the 'd'
+         'dEfGhI'
+
+    Note that the first iterator must be fully consumed before the second
+    iterator can generate valid results.
+    """
+    it = iter(it)
+    transition = []
+
+    def true_iterator():
+        for elem in it:
+            if predicate(elem):
+                yield elem
+            else:
+                transition.append(elem)
+                return
+
+    # Note: this is different from itertools recipes to allow nesting
+    # before_and_after remainders into before_and_after again. See tests
+    # for an example.
+    remainder_iterator = chain(transition, it)
+
+    return true_iterator(), remainder_iterator
+
+
+def triplewise(iterable):
+    """Return overlapping triplets from *iterable*.
+
+    >>> list(triplewise('ABCDE'))
+    [('A', 'B', 'C'), ('B', 'C', 'D'), ('C', 'D', 'E')]
+
+    """
+    for (a, _), (b, c) in pairwise(pairwise(iterable)):
+        yield a, b, c
+
+
+def sliding_window(iterable, n):
+    """Return a sliding window of width *n* over *iterable*.
+
+        >>> list(sliding_window(range(6), 4))
+        [(0, 1, 2, 3), (1, 2, 3, 4), (2, 3, 4, 5)]
+
+    If *iterable* has fewer than *n* items, then nothing is yielded:
+
+        >>> list(sliding_window(range(3), 4))
+        []
+
+    For a variant with more features, see :func:`windowed`.
+    """
+    it = iter(iterable)
+    window = deque(islice(it, n - 1), maxlen=n)
+    for x in it:
+        window.append(x)
+        yield tuple(window)
+
+
+def subslices(iterable):
+    """Return all contiguous non-empty subslices of *iterable*.
+
+        >>> list(subslices('ABC'))
+        [['A'], ['A', 'B'], ['A', 'B', 'C'], ['B'], ['B', 'C'], ['C']]
+
+    This is similar to :func:`substrings`, but emits items in a different
+    order.
+    """
+    seq = list(iterable)
+    slices = starmap(slice, combinations(range(len(seq) + 1), 2))
+    return map(operator.getitem, repeat(seq), slices)
+
+
+def polynomial_from_roots(roots):
+    """Compute a polynomial's coefficients from its roots.
+
+    >>> roots = [5, -4, 3]  # (x - 5) * (x + 4) * (x - 3)
+    >>> polynomial_from_roots(roots)  # x^3 - 4 * x^2 - 17 * x + 60
+    [1, -4, -17, 60]
+    """
+    factors = zip(repeat(1), map(operator.neg, roots))
+    return list(reduce(convolve, factors, [1]))
+
+
+def iter_index(iterable, value, start=0, stop=None):
+    """Yield the index of each place in *iterable* that *value* occurs,
+    beginning with index *start* and ending before index *stop*.
+
+
+    >>> list(iter_index('AABCADEAF', 'A'))
+    [0, 1, 4, 7]
+    >>> list(iter_index('AABCADEAF', 'A', 1))  # start index is inclusive
+    [1, 4, 7]
+    >>> list(iter_index('AABCADEAF', 'A', 1, 7))  # stop index is not inclusive
+    [1, 4]
+
+    The behavior for non-scalar *values* matches the built-in Python types.
+
+    >>> list(iter_index('ABCDABCD', 'AB'))
+    [0, 4]
+    >>> list(iter_index([0, 1, 2, 3, 0, 1, 2, 3], [0, 1]))
+    []
+    >>> list(iter_index([[0, 1], [2, 3], [0, 1], [2, 3]], [0, 1]))
+    [0, 2]
+
+    See :func:`locate` for a more general means of finding the indexes
+    associated with particular values.
+
+    """
+    seq_index = getattr(iterable, 'index', None)
+    if seq_index is None:
+        # Slow path for general iterables
+        it = islice(iterable, start, stop)
+        for i, element in enumerate(it, start):
+            if element is value or element == value:
+                yield i
+    else:
+        # Fast path for sequences
+        stop = len(iterable) if stop is None else stop
+        i = start - 1
+        try:
+            while True:
+                yield (i := seq_index(value, i + 1, stop))
+        except ValueError:
+            pass
+
+
+def sieve(n):
+    """Yield the primes less than n.
+
+    >>> list(sieve(30))
+    [2, 3, 5, 7, 11, 13, 17, 19, 23, 29]
+    """
+    if n > 2:
+        yield 2
+    start = 3
+    data = bytearray((0, 1)) * (n // 2)
+    limit = math.isqrt(n) + 1
+    for p in iter_index(data, 1, start, limit):
+        yield from iter_index(data, 1, start, p * p)
+        data[p * p : n : p + p] = bytes(len(range(p * p, n, p + p)))
+        start = p * p
+    yield from iter_index(data, 1, start)
+
+
+def _batched(iterable, n, *, strict=False):
+    """Batch data into tuples of length *n*. If the number of items in
+    *iterable* is not divisible by *n*:
+    * The last batch will be shorter if *strict* is ``False``.
+    * :exc:`ValueError` will be raised if *strict* is ``True``.
+
+    >>> list(batched('ABCDEFG', 3))
+    [('A', 'B', 'C'), ('D', 'E', 'F'), ('G',)]
+
+    On Python 3.13 and above, this is an alias for :func:`itertools.batched`.
+    """
+    if n < 1:
+        raise ValueError('n must be at least one')
+    it = iter(iterable)
+    while batch := tuple(islice(it, n)):
+        if strict and len(batch) != n:
+            raise ValueError('batched(): incomplete batch')
+        yield batch
+
+
+if hexversion >= 0x30D00A2:
+    from itertools import batched as itertools_batched
+
+    def batched(iterable, n, *, strict=False):
+        return itertools_batched(iterable, n, strict=strict)
+
+else:
+    batched = _batched
+
+    batched.__doc__ = _batched.__doc__
+
+
+def transpose(it):
+    """Swap the rows and columns of the input matrix.
+
+    >>> list(transpose([(1, 2, 3), (11, 22, 33)]))
+    [(1, 11), (2, 22), (3, 33)]
+
+    The caller should ensure that the dimensions of the input are compatible.
+    If the input is empty, no output will be produced.
+    """
+    return _zip_strict(*it)
+
+
+def reshape(matrix, cols):
+    """Reshape the 2-D input *matrix* to have a column count given by *cols*.
+
+    >>> matrix = [(0, 1), (2, 3), (4, 5)]
+    >>> cols = 3
+    >>> list(reshape(matrix, cols))
+    [(0, 1, 2), (3, 4, 5)]
+    """
+    return batched(chain.from_iterable(matrix), cols)
+
+
+def matmul(m1, m2):
+    """Multiply two matrices.
+
+    >>> list(matmul([(7, 5), (3, 5)], [(2, 5), (7, 9)]))
+    [(49, 80), (41, 60)]
+
+    The caller should ensure that the dimensions of the input matrices are
+    compatible with each other.
+    """
+    n = len(m2[0])
+    return batched(starmap(_sumprod, product(m1, transpose(m2))), n)
+
+
+def factor(n):
+    """Yield the prime factors of n.
+
+    >>> list(factor(360))
+    [2, 2, 2, 3, 3, 5]
+    """
+    for prime in sieve(math.isqrt(n) + 1):
+        while not n % prime:
+            yield prime
+            n //= prime
+            if n == 1:
+                return
+    if n > 1:
+        yield n
+
+
+def polynomial_eval(coefficients, x):
+    """Evaluate a polynomial at a specific value.
+
+    Example: evaluating x^3 - 4 * x^2 - 17 * x + 60 at x = 2.5:
+
+    >>> coefficients = [1, -4, -17, 60]
+    >>> x = 2.5
+    >>> polynomial_eval(coefficients, x)
+    8.125
+    """
+    n = len(coefficients)
+    if n == 0:
+        return x * 0  # coerce zero to the type of x
+    powers = map(pow, repeat(x), reversed(range(n)))
+    return _sumprod(coefficients, powers)
+
+
+def sum_of_squares(it):
+    """Return the sum of the squares of the input values.
+
+    >>> sum_of_squares([10, 20, 30])
+    1400
+    """
+    return _sumprod(*tee(it))
+
+
+def polynomial_derivative(coefficients):
+    """Compute the first derivative of a polynomial.
+
+    Example: evaluating the derivative of x^3 - 4 * x^2 - 17 * x + 60
+
+    >>> coefficients = [1, -4, -17, 60]
+    >>> derivative_coefficients = polynomial_derivative(coefficients)
+    >>> derivative_coefficients
+    [3, -8, -17]
+    """
+    n = len(coefficients)
+    powers = reversed(range(1, n))
+    return list(map(operator.mul, coefficients, powers))
+
+
+def totient(n):
+    """Return the count of natural numbers up to *n* that are coprime with *n*.
+
+    >>> totient(9)
+    6
+    >>> totient(12)
+    4
+    """
+    # The itertools docs use unique_justseen instead of set; see
+    # https://github.com/more-itertools/more-itertools/issues/823
+    for p in set(factor(n)):
+        n = n // p * (p - 1)
+
+    return n
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.pyi b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.pyi
new file mode 100644
index 00000000..739acec0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/more_itertools/recipes.pyi
@@ -0,0 +1,136 @@
+"""Stubs for more_itertools.recipes"""
+
+from __future__ import annotations
+
+from typing import (
+    Any,
+    Callable,
+    Iterable,
+    Iterator,
+    overload,
+    Sequence,
+    Type,
+    TypeVar,
+)
+
+# Type and type variable definitions
+_T = TypeVar('_T')
+_T1 = TypeVar('_T1')
+_T2 = TypeVar('_T2')
+_U = TypeVar('_U')
+
+def take(n: int, iterable: Iterable[_T]) -> list[_T]: ...
+def tabulate(
+    function: Callable[[int], _T], start: int = ...
+) -> Iterator[_T]: ...
+def tail(n: int, iterable: Iterable[_T]) -> Iterator[_T]: ...
+def consume(iterator: Iterable[_T], n: int | None = ...) -> None: ...
+@overload
+def nth(iterable: Iterable[_T], n: int) -> _T | None: ...
+@overload
+def nth(iterable: Iterable[_T], n: int, default: _U) -> _T | _U: ...
+def all_equal(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> bool: ...
+def quantify(
+    iterable: Iterable[_T], pred: Callable[[_T], bool] = ...
+) -> int: ...
+def pad_none(iterable: Iterable[_T]) -> Iterator[_T | None]: ...
+def padnone(iterable: Iterable[_T]) -> Iterator[_T | None]: ...
+def ncycles(iterable: Iterable[_T], n: int) -> Iterator[_T]: ...
+def dotproduct(vec1: Iterable[_T1], vec2: Iterable[_T2]) -> Any: ...
+def flatten(listOfLists: Iterable[Iterable[_T]]) -> Iterator[_T]: ...
+def repeatfunc(
+    func: Callable[..., _U], times: int | None = ..., *args: Any
+) -> Iterator[_U]: ...
+def pairwise(iterable: Iterable[_T]) -> Iterator[tuple[_T, _T]]: ...
+def grouper(
+    iterable: Iterable[_T],
+    n: int,
+    incomplete: str = ...,
+    fillvalue: _U = ...,
+) -> Iterator[tuple[_T | _U, ...]]: ...
+def roundrobin(*iterables: Iterable[_T]) -> Iterator[_T]: ...
+def partition(
+    pred: Callable[[_T], object] | None, iterable: Iterable[_T]
+) -> tuple[Iterator[_T], Iterator[_T]]: ...
+def powerset(iterable: Iterable[_T]) -> Iterator[tuple[_T, ...]]: ...
+def unique_everseen(
+    iterable: Iterable[_T], key: Callable[[_T], _U] | None = ...
+) -> Iterator[_T]: ...
+def unique_justseen(
+    iterable: Iterable[_T], key: Callable[[_T], object] | None = ...
+) -> Iterator[_T]: ...
+def unique(
+    iterable: Iterable[_T],
+    key: Callable[[_T], object] | None = ...,
+    reverse: bool = False,
+) -> Iterator[_T]: ...
+@overload
+def iter_except(
+    func: Callable[[], _T],
+    exception: Type[BaseException] | tuple[Type[BaseException], ...],
+    first: None = ...,
+) -> Iterator[_T]: ...
+@overload
+def iter_except(
+    func: Callable[[], _T],
+    exception: Type[BaseException] | tuple[Type[BaseException], ...],
+    first: Callable[[], _U],
+) -> Iterator[_T | _U]: ...
+@overload
+def first_true(
+    iterable: Iterable[_T], *, pred: Callable[[_T], object] | None = ...
+) -> _T | None: ...
+@overload
+def first_true(
+    iterable: Iterable[_T],
+    default: _U,
+    pred: Callable[[_T], object] | None = ...,
+) -> _T | _U: ...
+def random_product(
+    *args: Iterable[_T], repeat: int = ...
+) -> tuple[_T, ...]: ...
+def random_permutation(
+    iterable: Iterable[_T], r: int | None = ...
+) -> tuple[_T, ...]: ...
+def random_combination(iterable: Iterable[_T], r: int) -> tuple[_T, ...]: ...
+def random_combination_with_replacement(
+    iterable: Iterable[_T], r: int
+) -> tuple[_T, ...]: ...
+def nth_combination(
+    iterable: Iterable[_T], r: int, index: int
+) -> tuple[_T, ...]: ...
+def prepend(value: _T, iterator: Iterable[_U]) -> Iterator[_T | _U]: ...
+def convolve(signal: Iterable[_T], kernel: Iterable[_T]) -> Iterator[_T]: ...
+def before_and_after(
+    predicate: Callable[[_T], bool], it: Iterable[_T]
+) -> tuple[Iterator[_T], Iterator[_T]]: ...
+def triplewise(iterable: Iterable[_T]) -> Iterator[tuple[_T, _T, _T]]: ...
+def sliding_window(
+    iterable: Iterable[_T], n: int
+) -> Iterator[tuple[_T, ...]]: ...
+def subslices(iterable: Iterable[_T]) -> Iterator[list[_T]]: ...
+def polynomial_from_roots(roots: Sequence[_T]) -> list[_T]: ...
+def iter_index(
+    iterable: Iterable[_T],
+    value: Any,
+    start: int | None = ...,
+    stop: int | None = ...,
+) -> Iterator[int]: ...
+def sieve(n: int) -> Iterator[int]: ...
+def batched(
+    iterable: Iterable[_T], n: int, *, strict: bool = False
+) -> Iterator[tuple[_T]]: ...
+def transpose(
+    it: Iterable[Iterable[_T]],
+) -> Iterator[tuple[_T, ...]]: ...
+def reshape(
+    matrix: Iterable[Iterable[_T]], cols: int
+) -> Iterator[tuple[_T, ...]]: ...
+def matmul(m1: Sequence[_T], m2: Sequence[_T]) -> Iterator[tuple[_T]]: ...
+def factor(n: int) -> Iterator[int]: ...
+def polynomial_eval(coefficients: Sequence[_T], x: _U) -> _U: ...
+def sum_of_squares(it: Iterable[_T]) -> _T: ...
+def polynomial_derivative(coefficients: Sequence[_T]) -> list[_T]: ...
+def totient(n: int) -> int: ...
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/INSTALLER
new file mode 100644
index 00000000..5c69047b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+uv
\ No newline at end of file
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE
new file mode 100644
index 00000000..6f62d44e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE
@@ -0,0 +1,3 @@
+This software is made available under the terms of *either* of the licenses
+found in LICENSE.APACHE or LICENSE.BSD. Contributions to this software is made
+under the terms of *both* these licenses.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.APACHE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.APACHE
new file mode 100644
index 00000000..f433b1a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.APACHE
@@ -0,0 +1,177 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.BSD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.BSD
new file mode 100644
index 00000000..42ce7b75
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/LICENSE.BSD
@@ -0,0 +1,23 @@
+Copyright (c) Donald Stufft and individual contributors.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+    1. Redistributions of source code must retain the above copyright notice,
+       this list of conditions and the following disclaimer.
+
+    2. Redistributions in binary form must reproduce the above copyright
+       notice, this list of conditions and the following disclaimer in the
+       documentation and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/METADATA
new file mode 100644
index 00000000..1479c869
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/METADATA
@@ -0,0 +1,102 @@
+Metadata-Version: 2.3
+Name: packaging
+Version: 24.2
+Summary: Core utilities for Python packages
+Author-email: Donald Stufft <donald@stufft.io>
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Typing :: Typed
+Project-URL: Documentation, https://packaging.pypa.io/
+Project-URL: Source, https://github.com/pypa/packaging
+
+packaging
+=========
+
+.. start-intro
+
+Reusable core utilities for various Python Packaging
+`interoperability specifications <https://packaging.python.org/specifications/>`_.
+
+This library provides utilities that implement the interoperability
+specifications which have clearly one correct behaviour (eg: :pep:`440`)
+or benefit greatly from having a single shared implementation (eg: :pep:`425`).
+
+.. end-intro
+
+The ``packaging`` project includes the following: version handling, specifiers,
+markers, requirements, tags, utilities.
+
+Documentation
+-------------
+
+The `documentation`_ provides information and the API for the following:
+
+- Version Handling
+- Specifiers
+- Markers
+- Requirements
+- Tags
+- Utilities
+
+Installation
+------------
+
+Use ``pip`` to install these utilities::
+
+    pip install packaging
+
+The ``packaging`` library uses calendar-based versioning (``YY.N``).
+
+Discussion
+----------
+
+If you run into bugs, you can file them in our `issue tracker`_.
+
+You can also join ``#pypa`` on Freenode to ask questions or get involved.
+
+
+.. _`documentation`: https://packaging.pypa.io/
+.. _`issue tracker`: https://github.com/pypa/packaging/issues
+
+
+Code of Conduct
+---------------
+
+Everyone interacting in the packaging project's codebases, issue trackers, chat
+rooms, and mailing lists is expected to follow the `PSF Code of Conduct`_.
+
+.. _PSF Code of Conduct: https://github.com/pypa/.github/blob/main/CODE_OF_CONDUCT.md
+
+Contributing
+------------
+
+The ``CONTRIBUTING.rst`` file outlines how to contribute to this project as
+well as how to report a potential security issue. The documentation for this
+project also covers information about `project development`_ and `security`_.
+
+.. _`project development`: https://packaging.pypa.io/en/latest/development/
+.. _`security`: https://packaging.pypa.io/en/latest/security/
+
+Project History
+---------------
+
+Please review the ``CHANGELOG.rst`` file or the `Changelog documentation`_ for
+recent changes and project history.
+
+.. _`Changelog documentation`: https://packaging.pypa.io/en/latest/changelog/
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/RECORD
new file mode 100644
index 00000000..678aa5a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/RECORD
@@ -0,0 +1,25 @@
+packaging-24.2.dist-info/INSTALLER,sha256=5hhM4Q4mYTT9z6QB6PGpUAW81PGNFrYrdXMj4oM_6ak,2
+packaging-24.2.dist-info/LICENSE,sha256=ytHvW9NA1z4HS6YU0m996spceUDD2MNIUuZcSQlobEg,197
+packaging-24.2.dist-info/LICENSE.APACHE,sha256=DVQuDIgE45qn836wDaWnYhSdxoLXgpRRKH4RuTjpRZQ,10174
+packaging-24.2.dist-info/LICENSE.BSD,sha256=tw5-m3QvHMb5SLNMFqo5_-zpQZY2S8iP8NIYDwAo-sU,1344
+packaging-24.2.dist-info/METADATA,sha256=ohH86s6k5mIfQxY2TS0LcSfADeOFa4BiCC-bxZV-pNs,3204
+packaging-24.2.dist-info/RECORD,,
+packaging-24.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+packaging-24.2.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+packaging/__init__.py,sha256=dk4Ta_vmdVJxYHDcfyhvQNw8V3PgSBomKNXqg-D2JDY,494
+packaging/_elffile.py,sha256=cflAQAkE25tzhYmq_aCi72QfbT_tn891tPzfpbeHOwE,3306
+packaging/_manylinux.py,sha256=vl5OCoz4kx80H5rwXKeXWjl9WNISGmr4ZgTpTP9lU9c,9612
+packaging/_musllinux.py,sha256=p9ZqNYiOItGee8KcZFeHF_YcdhVwGHdK6r-8lgixvGQ,2694
+packaging/_parser.py,sha256=s_TvTvDNK0NrM2QB3VKThdWFM4Nc0P6JnkObkl3MjpM,10236
+packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+packaging/_tokenizer.py,sha256=J6v5H7Jzvb-g81xp_2QACKwO7LxHQA6ikryMU7zXwN8,5273
+packaging/licenses/__init__.py,sha256=1x5M1nEYjcgwEbLt0dXwz2ukjr18DiCzC0sraQqJ-Ww,5715
+packaging/licenses/_spdx.py,sha256=oAm1ztPFwlsmCKe7lAAsv_OIOfS1cWDu9bNBkeu-2ns,48398
+packaging/markers.py,sha256=c89TNzB7ZdGYhkovm6PYmqGyHxXlYVaLW591PHUNKD8,10561
+packaging/metadata.py,sha256=YJibM7GYe4re8-0a3OlXmGS-XDgTEoO4tlBt2q25Bng,34762
+packaging/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+packaging/requirements.py,sha256=gYyRSAdbrIyKDY66ugIDUQjRMvxkH2ALioTmX3tnL6o,2947
+packaging/specifiers.py,sha256=GG1wPNMcL0fMJO68vF53wKMdwnfehDcaI-r9NpTfilA,40074
+packaging/tags.py,sha256=CFqrJzAzc2XNGexerH__T-Y5Iwq7WbsYXsiLERLWxY0,21014
+packaging/utils.py,sha256=0F3Hh9OFuRgrhTgGZUl5K22Fv1YP2tZl1z_2gO6kJiA,5050
+packaging/version.py,sha256=olfyuk_DPbflNkJ4wBWetXQ17c74x3DB501degUv7DY,16676
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging-24.2.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__init__.py
new file mode 100644
index 00000000..d79f73c5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__init__.py
@@ -0,0 +1,15 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+__title__ = "packaging"
+__summary__ = "Core utilities for Python packages"
+__uri__ = "https://github.com/pypa/packaging"
+
+__version__ = "24.2"
+
+__author__ = "Donald Stufft and individual contributors"
+__email__ = "donald@stufft.io"
+
+__license__ = "BSD-2-Clause or Apache-2.0"
+__copyright__ = f"2014 {__author__}"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6fa62f82
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc
new file mode 100644
index 00000000..28e7e549
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_elffile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc
new file mode 100644
index 00000000..f9b46df8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_manylinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc
new file mode 100644
index 00000000..7c40332e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_musllinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..2d18445d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_structures.cpython-312.pyc
new file mode 100644
index 00000000..1b60e377
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc
new file mode 100644
index 00000000..6e538e63
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/_tokenizer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/markers.cpython-312.pyc
new file mode 100644
index 00000000..edabc8f0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/metadata.cpython-312.pyc
new file mode 100644
index 00000000..69d0b7f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..07809ac6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc
new file mode 100644
index 00000000..63743d46
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/specifiers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/tags.cpython-312.pyc
new file mode 100644
index 00000000..d630febb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..9a94a5f8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..98717546
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_elffile.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_elffile.py
new file mode 100644
index 00000000..25f4282c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_elffile.py
@@ -0,0 +1,110 @@
+"""
+ELF file parser.
+
+This provides a class ``ELFFile`` that parses an ELF executable in a similar
+interface to ``ZipFile``. Only the read interface is implemented.
+
+Based on: https://gist.github.com/lyssdod/f51579ae8d93c8657a5564aefc2ffbca
+ELF header: https://refspecs.linuxfoundation.org/elf/gabi4+/ch4.eheader.html
+"""
+
+from __future__ import annotations
+
+import enum
+import os
+import struct
+from typing import IO
+
+
+class ELFInvalid(ValueError):
+    pass
+
+
+class EIClass(enum.IntEnum):
+    C32 = 1
+    C64 = 2
+
+
+class EIData(enum.IntEnum):
+    Lsb = 1
+    Msb = 2
+
+
+class EMachine(enum.IntEnum):
+    I386 = 3
+    S390 = 22
+    Arm = 40
+    X8664 = 62
+    AArc64 = 183
+
+
+class ELFFile:
+    """
+    Representation of an ELF executable.
+    """
+
+    def __init__(self, f: IO[bytes]) -> None:
+        self._f = f
+
+        try:
+            ident = self._read("16B")
+        except struct.error as e:
+            raise ELFInvalid("unable to parse identification") from e
+        magic = bytes(ident[:4])
+        if magic != b"\x7fELF":
+            raise ELFInvalid(f"invalid magic: {magic!r}")
+
+        self.capacity = ident[4]  # Format for program header (bitness).
+        self.encoding = ident[5]  # Data structure encoding (endianness).
+
+        try:
+            # e_fmt: Format for program header.
+            # p_fmt: Format for section header.
+            # p_idx: Indexes to find p_type, p_offset, and p_filesz.
+            e_fmt, self._p_fmt, self._p_idx = {
+                (1, 1): ("<HHIIIIIHHH", "<IIIIIIII", (0, 1, 4)),  # 32-bit LSB.
+                (1, 2): (">HHIIIIIHHH", ">IIIIIIII", (0, 1, 4)),  # 32-bit MSB.
+                (2, 1): ("<HHIQQQIHHH", "<IIQQQQQQ", (0, 2, 5)),  # 64-bit LSB.
+                (2, 2): (">HHIQQQIHHH", ">IIQQQQQQ", (0, 2, 5)),  # 64-bit MSB.
+            }[(self.capacity, self.encoding)]
+        except KeyError as e:
+            raise ELFInvalid(
+                f"unrecognized capacity ({self.capacity}) or "
+                f"encoding ({self.encoding})"
+            ) from e
+
+        try:
+            (
+                _,
+                self.machine,  # Architecture type.
+                _,
+                _,
+                self._e_phoff,  # Offset of program header.
+                _,
+                self.flags,  # Processor-specific flags.
+                _,
+                self._e_phentsize,  # Size of section.
+                self._e_phnum,  # Number of sections.
+            ) = self._read(e_fmt)
+        except struct.error as e:
+            raise ELFInvalid("unable to parse machine and section information") from e
+
+    def _read(self, fmt: str) -> tuple[int, ...]:
+        return struct.unpack(fmt, self._f.read(struct.calcsize(fmt)))
+
+    @property
+    def interpreter(self) -> str | None:
+        """
+        The path recorded in the ``PT_INTERP`` section header.
+        """
+        for index in range(self._e_phnum):
+            self._f.seek(self._e_phoff + self._e_phentsize * index)
+            try:
+                data = self._read(self._p_fmt)
+            except struct.error:
+                continue
+            if data[self._p_idx[0]] != 3:  # Not PT_INTERP.
+                continue
+            self._f.seek(data[self._p_idx[1]])
+            return os.fsdecode(self._f.read(data[self._p_idx[2]])).strip("\0")
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_manylinux.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_manylinux.py
new file mode 100644
index 00000000..61339a6f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_manylinux.py
@@ -0,0 +1,263 @@
+from __future__ import annotations
+
+import collections
+import contextlib
+import functools
+import os
+import re
+import sys
+import warnings
+from typing import Generator, Iterator, NamedTuple, Sequence
+
+from ._elffile import EIClass, EIData, ELFFile, EMachine
+
+EF_ARM_ABIMASK = 0xFF000000
+EF_ARM_ABI_VER5 = 0x05000000
+EF_ARM_ABI_FLOAT_HARD = 0x00000400
+
+
+# `os.PathLike` not a generic type until Python 3.9, so sticking with `str`
+# as the type for `path` until then.
+@contextlib.contextmanager
+def _parse_elf(path: str) -> Generator[ELFFile | None, None, None]:
+    try:
+        with open(path, "rb") as f:
+            yield ELFFile(f)
+    except (OSError, TypeError, ValueError):
+        yield None
+
+
+def _is_linux_armhf(executable: str) -> bool:
+    # hard-float ABI can be detected from the ELF header of the running
+    # process
+    # https://static.docs.arm.com/ihi0044/g/aaelf32.pdf
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.Arm
+            and f.flags & EF_ARM_ABIMASK == EF_ARM_ABI_VER5
+            and f.flags & EF_ARM_ABI_FLOAT_HARD == EF_ARM_ABI_FLOAT_HARD
+        )
+
+
+def _is_linux_i686(executable: str) -> bool:
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.I386
+        )
+
+
+def _have_compatible_abi(executable: str, archs: Sequence[str]) -> bool:
+    if "armv7l" in archs:
+        return _is_linux_armhf(executable)
+    if "i686" in archs:
+        return _is_linux_i686(executable)
+    allowed_archs = {
+        "x86_64",
+        "aarch64",
+        "ppc64",
+        "ppc64le",
+        "s390x",
+        "loongarch64",
+        "riscv64",
+    }
+    return any(arch in allowed_archs for arch in archs)
+
+
+# If glibc ever changes its major version, we need to know what the last
+# minor version was, so we can build the complete list of all versions.
+# For now, guess what the highest minor version might be, assume it will
+# be 50 for testing. Once this actually happens, update the dictionary
+# with the actual value.
+_LAST_GLIBC_MINOR: dict[int, int] = collections.defaultdict(lambda: 50)
+
+
+class _GLibCVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _glibc_version_string_confstr() -> str | None:
+    """
+    Primary implementation of glibc_version_string using os.confstr.
+    """
+    # os.confstr is quite a bit faster than ctypes.DLL. It's also less likely
+    # to be broken or missing. This strategy is used in the standard library
+    # platform module.
+    # https://github.com/python/cpython/blob/fcf1d003bf4f0100c/Lib/platform.py#L175-L183
+    try:
+        # Should be a string like "glibc 2.17".
+        version_string: str | None = os.confstr("CS_GNU_LIBC_VERSION")
+        assert version_string is not None
+        _, version = version_string.rsplit()
+    except (AssertionError, AttributeError, OSError, ValueError):
+        # os.confstr() or CS_GNU_LIBC_VERSION not available (or a bad value)...
+        return None
+    return version
+
+
+def _glibc_version_string_ctypes() -> str | None:
+    """
+    Fallback implementation of glibc_version_string using ctypes.
+    """
+    try:
+        import ctypes
+    except ImportError:
+        return None
+
+    # ctypes.CDLL(None) internally calls dlopen(NULL), and as the dlopen
+    # manpage says, "If filename is NULL, then the returned handle is for the
+    # main program". This way we can let the linker do the work to figure out
+    # which libc our process is actually using.
+    #
+    # We must also handle the special case where the executable is not a
+    # dynamically linked executable. This can occur when using musl libc,
+    # for example. In this situation, dlopen() will error, leading to an
+    # OSError. Interestingly, at least in the case of musl, there is no
+    # errno set on the OSError. The single string argument used to construct
+    # OSError comes from libc itself and is therefore not portable to
+    # hard code here. In any case, failure to call dlopen() means we
+    # can proceed, so we bail on our attempt.
+    try:
+        process_namespace = ctypes.CDLL(None)
+    except OSError:
+        return None
+
+    try:
+        gnu_get_libc_version = process_namespace.gnu_get_libc_version
+    except AttributeError:
+        # Symbol doesn't exist -> therefore, we are not linked to
+        # glibc.
+        return None
+
+    # Call gnu_get_libc_version, which returns a string like "2.5"
+    gnu_get_libc_version.restype = ctypes.c_char_p
+    version_str: str = gnu_get_libc_version()
+    # py2 / py3 compatibility:
+    if not isinstance(version_str, str):
+        version_str = version_str.decode("ascii")
+
+    return version_str
+
+
+def _glibc_version_string() -> str | None:
+    """Returns glibc version string, or None if not using glibc."""
+    return _glibc_version_string_confstr() or _glibc_version_string_ctypes()
+
+
+def _parse_glibc_version(version_str: str) -> tuple[int, int]:
+    """Parse glibc version.
+
+    We use a regexp instead of str.split because we want to discard any
+    random junk that might come after the minor version -- this might happen
+    in patched/forked versions of glibc (e.g. Linaro's version of glibc
+    uses version strings like "2.20-2014.11"). See gh-3588.
+    """
+    m = re.match(r"(?P<major>[0-9]+)\.(?P<minor>[0-9]+)", version_str)
+    if not m:
+        warnings.warn(
+            f"Expected glibc version with 2 components major.minor,"
+            f" got: {version_str}",
+            RuntimeWarning,
+            stacklevel=2,
+        )
+        return -1, -1
+    return int(m.group("major")), int(m.group("minor"))
+
+
+@functools.lru_cache
+def _get_glibc_version() -> tuple[int, int]:
+    version_str = _glibc_version_string()
+    if version_str is None:
+        return (-1, -1)
+    return _parse_glibc_version(version_str)
+
+
+# From PEP 513, PEP 600
+def _is_compatible(arch: str, version: _GLibCVersion) -> bool:
+    sys_glibc = _get_glibc_version()
+    if sys_glibc < version:
+        return False
+    # Check for presence of _manylinux module.
+    try:
+        import _manylinux
+    except ImportError:
+        return True
+    if hasattr(_manylinux, "manylinux_compatible"):
+        result = _manylinux.manylinux_compatible(version[0], version[1], arch)
+        if result is not None:
+            return bool(result)
+        return True
+    if version == _GLibCVersion(2, 5):
+        if hasattr(_manylinux, "manylinux1_compatible"):
+            return bool(_manylinux.manylinux1_compatible)
+    if version == _GLibCVersion(2, 12):
+        if hasattr(_manylinux, "manylinux2010_compatible"):
+            return bool(_manylinux.manylinux2010_compatible)
+    if version == _GLibCVersion(2, 17):
+        if hasattr(_manylinux, "manylinux2014_compatible"):
+            return bool(_manylinux.manylinux2014_compatible)
+    return True
+
+
+_LEGACY_MANYLINUX_MAP = {
+    # CentOS 7 w/ glibc 2.17 (PEP 599)
+    (2, 17): "manylinux2014",
+    # CentOS 6 w/ glibc 2.12 (PEP 571)
+    (2, 12): "manylinux2010",
+    # CentOS 5 w/ glibc 2.5 (PEP 513)
+    (2, 5): "manylinux1",
+}
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate manylinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be manylinux-compatible.
+
+    :returns: An iterator of compatible manylinux tags.
+    """
+    if not _have_compatible_abi(sys.executable, archs):
+        return
+    # Oldest glibc to be supported regardless of architecture is (2, 17).
+    too_old_glibc2 = _GLibCVersion(2, 16)
+    if set(archs) & {"x86_64", "i686"}:
+        # On x86/i686 also oldest glibc to be supported is (2, 5).
+        too_old_glibc2 = _GLibCVersion(2, 4)
+    current_glibc = _GLibCVersion(*_get_glibc_version())
+    glibc_max_list = [current_glibc]
+    # We can assume compatibility across glibc major versions.
+    # https://sourceware.org/bugzilla/show_bug.cgi?id=24636
+    #
+    # Build a list of maximum glibc versions so that we can
+    # output the canonical list of all glibc from current_glibc
+    # down to too_old_glibc2, including all intermediary versions.
+    for glibc_major in range(current_glibc.major - 1, 1, -1):
+        glibc_minor = _LAST_GLIBC_MINOR[glibc_major]
+        glibc_max_list.append(_GLibCVersion(glibc_major, glibc_minor))
+    for arch in archs:
+        for glibc_max in glibc_max_list:
+            if glibc_max.major == too_old_glibc2.major:
+                min_minor = too_old_glibc2.minor
+            else:
+                # For other glibc major versions oldest supported is (x, 0).
+                min_minor = -1
+            for glibc_minor in range(glibc_max.minor, min_minor, -1):
+                glibc_version = _GLibCVersion(glibc_max.major, glibc_minor)
+                tag = "manylinux_{}_{}".format(*glibc_version)
+                if _is_compatible(arch, glibc_version):
+                    yield f"{tag}_{arch}"
+                # Handle the legacy manylinux1, manylinux2010, manylinux2014 tags.
+                if glibc_version in _LEGACY_MANYLINUX_MAP:
+                    legacy_tag = _LEGACY_MANYLINUX_MAP[glibc_version]
+                    if _is_compatible(arch, glibc_version):
+                        yield f"{legacy_tag}_{arch}"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_musllinux.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_musllinux.py
new file mode 100644
index 00000000..d2bf30b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_musllinux.py
@@ -0,0 +1,85 @@
+"""PEP 656 support.
+
+This module implements logic to detect if the currently running Python is
+linked against musl, and what musl version is used.
+"""
+
+from __future__ import annotations
+
+import functools
+import re
+import subprocess
+import sys
+from typing import Iterator, NamedTuple, Sequence
+
+from ._elffile import ELFFile
+
+
+class _MuslVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _parse_musl_version(output: str) -> _MuslVersion | None:
+    lines = [n for n in (n.strip() for n in output.splitlines()) if n]
+    if len(lines) < 2 or lines[0][:4] != "musl":
+        return None
+    m = re.match(r"Version (\d+)\.(\d+)", lines[1])
+    if not m:
+        return None
+    return _MuslVersion(major=int(m.group(1)), minor=int(m.group(2)))
+
+
+@functools.lru_cache
+def _get_musl_version(executable: str) -> _MuslVersion | None:
+    """Detect currently-running musl runtime version.
+
+    This is done by checking the specified executable's dynamic linking
+    information, and invoking the loader to parse its output for a version
+    string. If the loader is musl, the output would be something like::
+
+        musl libc (x86_64)
+        Version 1.2.2
+        Dynamic Program Loader
+    """
+    try:
+        with open(executable, "rb") as f:
+            ld = ELFFile(f).interpreter
+    except (OSError, TypeError, ValueError):
+        return None
+    if ld is None or "musl" not in ld:
+        return None
+    proc = subprocess.run([ld], stderr=subprocess.PIPE, text=True)
+    return _parse_musl_version(proc.stderr)
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate musllinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be musllinux-compatible.
+
+    :returns: An iterator of compatible musllinux tags.
+    """
+    sys_musl = _get_musl_version(sys.executable)
+    if sys_musl is None:  # Python not dynamically linked against musl.
+        return
+    for arch in archs:
+        for minor in range(sys_musl.minor, -1, -1):
+            yield f"musllinux_{sys_musl.major}_{minor}_{arch}"
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sysconfig
+
+    plat = sysconfig.get_platform()
+    assert plat.startswith("linux-"), "not linux"
+
+    print("plat:", plat)
+    print("musl:", _get_musl_version(sys.executable))
+    print("tags:", end=" ")
+    for t in platform_tags(re.sub(r"[.-]", "_", plat.split("-", 1)[-1])):
+        print(t, end="\n      ")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_parser.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_parser.py
new file mode 100644
index 00000000..c1238c06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_parser.py
@@ -0,0 +1,354 @@
+"""Handwritten parser of dependency specifiers.
+
+The docstring for each __parse_* function contains EBNF-inspired grammar representing
+the implementation.
+"""
+
+from __future__ import annotations
+
+import ast
+from typing import NamedTuple, Sequence, Tuple, Union
+
+from ._tokenizer import DEFAULT_RULES, Tokenizer
+
+
+class Node:
+    def __init__(self, value: str) -> None:
+        self.value = value
+
+    def __str__(self) -> str:
+        return self.value
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__}('{self}')>"
+
+    def serialize(self) -> str:
+        raise NotImplementedError
+
+
+class Variable(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+class Value(Node):
+    def serialize(self) -> str:
+        return f'"{self}"'
+
+
+class Op(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+MarkerVar = Union[Variable, Value]
+MarkerItem = Tuple[MarkerVar, Op, MarkerVar]
+MarkerAtom = Union[MarkerItem, Sequence["MarkerAtom"]]
+MarkerList = Sequence[Union["MarkerList", MarkerAtom, str]]
+
+
+class ParsedRequirement(NamedTuple):
+    name: str
+    url: str
+    extras: list[str]
+    specifier: str
+    marker: MarkerList | None
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for dependency specifier
+# --------------------------------------------------------------------------------------
+def parse_requirement(source: str) -> ParsedRequirement:
+    return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_requirement(tokenizer: Tokenizer) -> ParsedRequirement:
+    """
+    requirement = WS? IDENTIFIER WS? extras WS? requirement_details
+    """
+    tokenizer.consume("WS")
+
+    name_token = tokenizer.expect(
+        "IDENTIFIER", expected="package name at the start of dependency specifier"
+    )
+    name = name_token.text
+    tokenizer.consume("WS")
+
+    extras = _parse_extras(tokenizer)
+    tokenizer.consume("WS")
+
+    url, specifier, marker = _parse_requirement_details(tokenizer)
+    tokenizer.expect("END", expected="end of dependency specifier")
+
+    return ParsedRequirement(name, url, extras, specifier, marker)
+
+
+def _parse_requirement_details(
+    tokenizer: Tokenizer,
+) -> tuple[str, str, MarkerList | None]:
+    """
+    requirement_details = AT URL (WS requirement_marker?)?
+                        | specifier WS? (requirement_marker)?
+    """
+
+    specifier = ""
+    url = ""
+    marker = None
+
+    if tokenizer.check("AT"):
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        url_start = tokenizer.position
+        url = tokenizer.expect("URL", expected="URL after @").text
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        tokenizer.expect("WS", expected="whitespace after URL")
+
+        # The input might end after whitespace.
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer, span_start=url_start, after="URL and whitespace"
+        )
+    else:
+        specifier_start = tokenizer.position
+        specifier = _parse_specifier(tokenizer)
+        tokenizer.consume("WS")
+
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer,
+            span_start=specifier_start,
+            after=(
+                "version specifier"
+                if specifier
+                else "name and no valid version specifier"
+            ),
+        )
+
+    return (url, specifier, marker)
+
+
+def _parse_requirement_marker(
+    tokenizer: Tokenizer, *, span_start: int, after: str
+) -> MarkerList:
+    """
+    requirement_marker = SEMICOLON marker WS?
+    """
+
+    if not tokenizer.check("SEMICOLON"):
+        tokenizer.raise_syntax_error(
+            f"Expected end or semicolon (after {after})",
+            span_start=span_start,
+        )
+    tokenizer.read()
+
+    marker = _parse_marker(tokenizer)
+    tokenizer.consume("WS")
+
+    return marker
+
+
+def _parse_extras(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras = (LEFT_BRACKET wsp* extras_list? wsp* RIGHT_BRACKET)?
+    """
+    if not tokenizer.check("LEFT_BRACKET", peek=True):
+        return []
+
+    with tokenizer.enclosing_tokens(
+        "LEFT_BRACKET",
+        "RIGHT_BRACKET",
+        around="extras",
+    ):
+        tokenizer.consume("WS")
+        extras = _parse_extras_list(tokenizer)
+        tokenizer.consume("WS")
+
+    return extras
+
+
+def _parse_extras_list(tokenizer: Tokenizer) -> list[str]:
+    """
+    extras_list = identifier (wsp* ',' wsp* identifier)*
+    """
+    extras: list[str] = []
+
+    if not tokenizer.check("IDENTIFIER"):
+        return extras
+
+    extras.append(tokenizer.read().text)
+
+    while True:
+        tokenizer.consume("WS")
+        if tokenizer.check("IDENTIFIER", peek=True):
+            tokenizer.raise_syntax_error("Expected comma between extra names")
+        elif not tokenizer.check("COMMA"):
+            break
+
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        extra_token = tokenizer.expect("IDENTIFIER", expected="extra name after comma")
+        extras.append(extra_token.text)
+
+    return extras
+
+
+def _parse_specifier(tokenizer: Tokenizer) -> str:
+    """
+    specifier = LEFT_PARENTHESIS WS? version_many WS? RIGHT_PARENTHESIS
+              | WS? version_many WS?
+    """
+    with tokenizer.enclosing_tokens(
+        "LEFT_PARENTHESIS",
+        "RIGHT_PARENTHESIS",
+        around="version specifier",
+    ):
+        tokenizer.consume("WS")
+        parsed_specifiers = _parse_version_many(tokenizer)
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+def _parse_version_many(tokenizer: Tokenizer) -> str:
+    """
+    version_many = (SPECIFIER (WS? COMMA WS? SPECIFIER)*)?
+    """
+    parsed_specifiers = ""
+    while tokenizer.check("SPECIFIER"):
+        span_start = tokenizer.position
+        parsed_specifiers += tokenizer.read().text
+        if tokenizer.check("VERSION_PREFIX_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                ".* suffix can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position + 1,
+            )
+        if tokenizer.check("VERSION_LOCAL_LABEL_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                "Local version label can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position,
+            )
+        tokenizer.consume("WS")
+        if not tokenizer.check("COMMA"):
+            break
+        parsed_specifiers += tokenizer.read().text
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for marker expression
+# --------------------------------------------------------------------------------------
+def parse_marker(source: str) -> MarkerList:
+    return _parse_full_marker(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_full_marker(tokenizer: Tokenizer) -> MarkerList:
+    retval = _parse_marker(tokenizer)
+    tokenizer.expect("END", expected="end of marker expression")
+    return retval
+
+
+def _parse_marker(tokenizer: Tokenizer) -> MarkerList:
+    """
+    marker = marker_atom (BOOLOP marker_atom)+
+    """
+    expression = [_parse_marker_atom(tokenizer)]
+    while tokenizer.check("BOOLOP"):
+        token = tokenizer.read()
+        expr_right = _parse_marker_atom(tokenizer)
+        expression.extend((token.text, expr_right))
+    return expression
+
+
+def _parse_marker_atom(tokenizer: Tokenizer) -> MarkerAtom:
+    """
+    marker_atom = WS? LEFT_PARENTHESIS WS? marker WS? RIGHT_PARENTHESIS WS?
+                | WS? marker_item WS?
+    """
+
+    tokenizer.consume("WS")
+    if tokenizer.check("LEFT_PARENTHESIS", peek=True):
+        with tokenizer.enclosing_tokens(
+            "LEFT_PARENTHESIS",
+            "RIGHT_PARENTHESIS",
+            around="marker expression",
+        ):
+            tokenizer.consume("WS")
+            marker: MarkerAtom = _parse_marker(tokenizer)
+            tokenizer.consume("WS")
+    else:
+        marker = _parse_marker_item(tokenizer)
+    tokenizer.consume("WS")
+    return marker
+
+
+def _parse_marker_item(tokenizer: Tokenizer) -> MarkerItem:
+    """
+    marker_item = WS? marker_var WS? marker_op WS? marker_var WS?
+    """
+    tokenizer.consume("WS")
+    marker_var_left = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    marker_op = _parse_marker_op(tokenizer)
+    tokenizer.consume("WS")
+    marker_var_right = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    return (marker_var_left, marker_op, marker_var_right)
+
+
+def _parse_marker_var(tokenizer: Tokenizer) -> MarkerVar:
+    """
+    marker_var = VARIABLE | QUOTED_STRING
+    """
+    if tokenizer.check("VARIABLE"):
+        return process_env_var(tokenizer.read().text.replace(".", "_"))
+    elif tokenizer.check("QUOTED_STRING"):
+        return process_python_str(tokenizer.read().text)
+    else:
+        tokenizer.raise_syntax_error(
+            message="Expected a marker variable or quoted string"
+        )
+
+
+def process_env_var(env_var: str) -> Variable:
+    if env_var in ("platform_python_implementation", "python_implementation"):
+        return Variable("platform_python_implementation")
+    else:
+        return Variable(env_var)
+
+
+def process_python_str(python_str: str) -> Value:
+    value = ast.literal_eval(python_str)
+    return Value(str(value))
+
+
+def _parse_marker_op(tokenizer: Tokenizer) -> Op:
+    """
+    marker_op = IN | NOT IN | OP
+    """
+    if tokenizer.check("IN"):
+        tokenizer.read()
+        return Op("in")
+    elif tokenizer.check("NOT"):
+        tokenizer.read()
+        tokenizer.expect("WS", expected="whitespace after 'not'")
+        tokenizer.expect("IN", expected="'in' after 'not'")
+        return Op("not in")
+    elif tokenizer.check("OP"):
+        return Op(tokenizer.read().text)
+    else:
+        return tokenizer.raise_syntax_error(
+            "Expected marker operator, one of "
+            "<=, <, !=, ==, >=, >, ~=, ===, in, not in"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_structures.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_structures.py
new file mode 100644
index 00000000..90a6465f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_structures.py
@@ -0,0 +1,61 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+
+class InfinityType:
+    def __repr__(self) -> str:
+        return "Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return False
+
+    def __le__(self, other: object) -> bool:
+        return False
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return True
+
+    def __ge__(self, other: object) -> bool:
+        return True
+
+    def __neg__(self: object) -> "NegativeInfinityType":
+        return NegativeInfinity
+
+
+Infinity = InfinityType()
+
+
+class NegativeInfinityType:
+    def __repr__(self) -> str:
+        return "-Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return True
+
+    def __le__(self, other: object) -> bool:
+        return True
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return False
+
+    def __ge__(self, other: object) -> bool:
+        return False
+
+    def __neg__(self: object) -> InfinityType:
+        return Infinity
+
+
+NegativeInfinity = NegativeInfinityType()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_tokenizer.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_tokenizer.py
new file mode 100644
index 00000000..89d04160
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/_tokenizer.py
@@ -0,0 +1,194 @@
+from __future__ import annotations
+
+import contextlib
+import re
+from dataclasses import dataclass
+from typing import Iterator, NoReturn
+
+from .specifiers import Specifier
+
+
+@dataclass
+class Token:
+    name: str
+    text: str
+    position: int
+
+
+class ParserSyntaxError(Exception):
+    """The provided source text could not be parsed correctly."""
+
+    def __init__(
+        self,
+        message: str,
+        *,
+        source: str,
+        span: tuple[int, int],
+    ) -> None:
+        self.span = span
+        self.message = message
+        self.source = source
+
+        super().__init__()
+
+    def __str__(self) -> str:
+        marker = " " * self.span[0] + "~" * (self.span[1] - self.span[0]) + "^"
+        return "\n    ".join([self.message, self.source, marker])
+
+
+DEFAULT_RULES: dict[str, str | re.Pattern[str]] = {
+    "LEFT_PARENTHESIS": r"\(",
+    "RIGHT_PARENTHESIS": r"\)",
+    "LEFT_BRACKET": r"\[",
+    "RIGHT_BRACKET": r"\]",
+    "SEMICOLON": r";",
+    "COMMA": r",",
+    "QUOTED_STRING": re.compile(
+        r"""
+            (
+                ('[^']*')
+                |
+                ("[^"]*")
+            )
+        """,
+        re.VERBOSE,
+    ),
+    "OP": r"(===|==|~=|!=|<=|>=|<|>)",
+    "BOOLOP": r"\b(or|and)\b",
+    "IN": r"\bin\b",
+    "NOT": r"\bnot\b",
+    "VARIABLE": re.compile(
+        r"""
+            \b(
+                python_version
+                |python_full_version
+                |os[._]name
+                |sys[._]platform
+                |platform_(release|system)
+                |platform[._](version|machine|python_implementation)
+                |python_implementation
+                |implementation_(name|version)
+                |extra
+            )\b
+        """,
+        re.VERBOSE,
+    ),
+    "SPECIFIER": re.compile(
+        Specifier._operator_regex_str + Specifier._version_regex_str,
+        re.VERBOSE | re.IGNORECASE,
+    ),
+    "AT": r"\@",
+    "URL": r"[^ \t]+",
+    "IDENTIFIER": r"\b[a-zA-Z0-9][a-zA-Z0-9._-]*\b",
+    "VERSION_PREFIX_TRAIL": r"\.\*",
+    "VERSION_LOCAL_LABEL_TRAIL": r"\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*",
+    "WS": r"[ \t]+",
+    "END": r"$",
+}
+
+
+class Tokenizer:
+    """Context-sensitive token parsing.
+
+    Provides methods to examine the input stream to check whether the next token
+    matches.
+    """
+
+    def __init__(
+        self,
+        source: str,
+        *,
+        rules: dict[str, str | re.Pattern[str]],
+    ) -> None:
+        self.source = source
+        self.rules: dict[str, re.Pattern[str]] = {
+            name: re.compile(pattern) for name, pattern in rules.items()
+        }
+        self.next_token: Token | None = None
+        self.position = 0
+
+    def consume(self, name: str) -> None:
+        """Move beyond provided token name, if at current position."""
+        if self.check(name):
+            self.read()
+
+    def check(self, name: str, *, peek: bool = False) -> bool:
+        """Check whether the next token has the provided name.
+
+        By default, if the check succeeds, the token *must* be read before
+        another check. If `peek` is set to `True`, the token is not loaded and
+        would need to be checked again.
+        """
+        assert (
+            self.next_token is None
+        ), f"Cannot check for {name!r}, already have {self.next_token!r}"
+        assert name in self.rules, f"Unknown token name: {name!r}"
+
+        expression = self.rules[name]
+
+        match = expression.match(self.source, self.position)
+        if match is None:
+            return False
+        if not peek:
+            self.next_token = Token(name, match[0], self.position)
+        return True
+
+    def expect(self, name: str, *, expected: str) -> Token:
+        """Expect a certain token name next, failing with a syntax error otherwise.
+
+        The token is *not* read.
+        """
+        if not self.check(name):
+            raise self.raise_syntax_error(f"Expected {expected}")
+        return self.read()
+
+    def read(self) -> Token:
+        """Consume the next token and return it."""
+        token = self.next_token
+        assert token is not None
+
+        self.position += len(token.text)
+        self.next_token = None
+
+        return token
+
+    def raise_syntax_error(
+        self,
+        message: str,
+        *,
+        span_start: int | None = None,
+        span_end: int | None = None,
+    ) -> NoReturn:
+        """Raise ParserSyntaxError at the given position."""
+        span = (
+            self.position if span_start is None else span_start,
+            self.position if span_end is None else span_end,
+        )
+        raise ParserSyntaxError(
+            message,
+            source=self.source,
+            span=span,
+        )
+
+    @contextlib.contextmanager
+    def enclosing_tokens(
+        self, open_token: str, close_token: str, *, around: str
+    ) -> Iterator[None]:
+        if self.check(open_token):
+            open_position = self.position
+            self.read()
+        else:
+            open_position = None
+
+        yield
+
+        if open_position is None:
+            return
+
+        if not self.check(close_token):
+            self.raise_syntax_error(
+                f"Expected matching {close_token} for {open_token}, after {around}",
+                span_start=open_position,
+            )
+
+        self.read()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__init__.py
new file mode 100644
index 00000000..569156d6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__init__.py
@@ -0,0 +1,145 @@
+#######################################################################################
+#
+# Adapted from:
+#  https://github.com/pypa/hatch/blob/5352e44/backend/src/hatchling/licenses/parse.py
+#
+# MIT License
+#
+# Copyright (c) 2017-present Ofek Lev <oss@ofek.dev>
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy of this
+# software and associated documentation files (the "Software"), to deal in the Software
+# without restriction, including without limitation the rights to use, copy, modify,
+# merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to the following
+# conditions:
+#
+# The above copyright notice and this permission notice shall be included in all copies
+# or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+# PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
+# CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
+# OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+#
+#
+# With additional allowance of arbitrary `LicenseRef-` identifiers, not just
+# `LicenseRef-Public-Domain` and `LicenseRef-Proprietary`.
+#
+#######################################################################################
+from __future__ import annotations
+
+import re
+from typing import NewType, cast
+
+from packaging.licenses._spdx import EXCEPTIONS, LICENSES
+
+__all__ = [
+    "NormalizedLicenseExpression",
+    "InvalidLicenseExpression",
+    "canonicalize_license_expression",
+]
+
+license_ref_allowed = re.compile("^[A-Za-z0-9.-]*$")
+
+NormalizedLicenseExpression = NewType("NormalizedLicenseExpression", str)
+
+
+class InvalidLicenseExpression(ValueError):
+    """Raised when a license-expression string is invalid
+
+    >>> canonicalize_license_expression("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.licenses.InvalidLicenseExpression: Invalid license expression: 'invalid'
+    """
+
+
+def canonicalize_license_expression(
+    raw_license_expression: str,
+) -> NormalizedLicenseExpression:
+    if not raw_license_expression:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message)
+
+    # Pad any parentheses so tokenization can be achieved by merely splitting on
+    # whitespace.
+    license_expression = raw_license_expression.replace("(", " ( ").replace(")", " ) ")
+    licenseref_prefix = "LicenseRef-"
+    license_refs = {
+        ref.lower(): "LicenseRef-" + ref[len(licenseref_prefix) :]
+        for ref in license_expression.split()
+        if ref.lower().startswith(licenseref_prefix.lower())
+    }
+
+    # Normalize to lower case so we can look up licenses/exceptions
+    # and so boolean operators are Python-compatible.
+    license_expression = license_expression.lower()
+
+    tokens = license_expression.split()
+
+    # Rather than implementing boolean logic, we create an expression that Python can
+    # parse. Everything that is not involved with the grammar itself is treated as
+    # `False` and the expression should evaluate as such.
+    python_tokens = []
+    for token in tokens:
+        if token not in {"or", "and", "with", "(", ")"}:
+            python_tokens.append("False")
+        elif token == "with":
+            python_tokens.append("or")
+        elif token == "(" and python_tokens and python_tokens[-1] not in {"or", "and"}:
+            message = f"Invalid license expression: {raw_license_expression!r}"
+            raise InvalidLicenseExpression(message)
+        else:
+            python_tokens.append(token)
+
+    python_expression = " ".join(python_tokens)
+    try:
+        invalid = eval(python_expression, globals(), locals())
+    except Exception:
+        invalid = True
+
+    if invalid is not False:
+        message = f"Invalid license expression: {raw_license_expression!r}"
+        raise InvalidLicenseExpression(message) from None
+
+    # Take a final pass to check for unknown licenses/exceptions.
+    normalized_tokens = []
+    for token in tokens:
+        if token in {"or", "and", "with", "(", ")"}:
+            normalized_tokens.append(token.upper())
+            continue
+
+        if normalized_tokens and normalized_tokens[-1] == "WITH":
+            if token not in EXCEPTIONS:
+                message = f"Unknown license exception: {token!r}"
+                raise InvalidLicenseExpression(message)
+
+            normalized_tokens.append(EXCEPTIONS[token]["id"])
+        else:
+            if token.endswith("+"):
+                final_token = token[:-1]
+                suffix = "+"
+            else:
+                final_token = token
+                suffix = ""
+
+            if final_token.startswith("licenseref-"):
+                if not license_ref_allowed.match(final_token):
+                    message = f"Invalid licenseref: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(license_refs[final_token] + suffix)
+            else:
+                if final_token not in LICENSES:
+                    message = f"Unknown license: {final_token!r}"
+                    raise InvalidLicenseExpression(message)
+                normalized_tokens.append(LICENSES[final_token]["id"] + suffix)
+
+    normalized_expression = " ".join(normalized_tokens)
+
+    return cast(
+        NormalizedLicenseExpression,
+        normalized_expression.replace("( ", "(").replace(" )", ")"),
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..6f46895d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc
new file mode 100644
index 00000000..c163b953
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/__pycache__/_spdx.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/_spdx.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/_spdx.py
new file mode 100644
index 00000000..eac22276
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/licenses/_spdx.py
@@ -0,0 +1,759 @@
+
+from __future__ import annotations
+
+from typing import TypedDict
+
+class SPDXLicense(TypedDict):
+    id: str
+    deprecated: bool
+
+class SPDXException(TypedDict):
+    id: str
+    deprecated: bool
+
+
+VERSION = '3.25.0'
+
+LICENSES: dict[str, SPDXLicense] = {
+    '0bsd': {'id': '0BSD', 'deprecated': False},
+    '3d-slicer-1.0': {'id': '3D-Slicer-1.0', 'deprecated': False},
+    'aal': {'id': 'AAL', 'deprecated': False},
+    'abstyles': {'id': 'Abstyles', 'deprecated': False},
+    'adacore-doc': {'id': 'AdaCore-doc', 'deprecated': False},
+    'adobe-2006': {'id': 'Adobe-2006', 'deprecated': False},
+    'adobe-display-postscript': {'id': 'Adobe-Display-PostScript', 'deprecated': False},
+    'adobe-glyph': {'id': 'Adobe-Glyph', 'deprecated': False},
+    'adobe-utopia': {'id': 'Adobe-Utopia', 'deprecated': False},
+    'adsl': {'id': 'ADSL', 'deprecated': False},
+    'afl-1.1': {'id': 'AFL-1.1', 'deprecated': False},
+    'afl-1.2': {'id': 'AFL-1.2', 'deprecated': False},
+    'afl-2.0': {'id': 'AFL-2.0', 'deprecated': False},
+    'afl-2.1': {'id': 'AFL-2.1', 'deprecated': False},
+    'afl-3.0': {'id': 'AFL-3.0', 'deprecated': False},
+    'afmparse': {'id': 'Afmparse', 'deprecated': False},
+    'agpl-1.0': {'id': 'AGPL-1.0', 'deprecated': True},
+    'agpl-1.0-only': {'id': 'AGPL-1.0-only', 'deprecated': False},
+    'agpl-1.0-or-later': {'id': 'AGPL-1.0-or-later', 'deprecated': False},
+    'agpl-3.0': {'id': 'AGPL-3.0', 'deprecated': True},
+    'agpl-3.0-only': {'id': 'AGPL-3.0-only', 'deprecated': False},
+    'agpl-3.0-or-later': {'id': 'AGPL-3.0-or-later', 'deprecated': False},
+    'aladdin': {'id': 'Aladdin', 'deprecated': False},
+    'amd-newlib': {'id': 'AMD-newlib', 'deprecated': False},
+    'amdplpa': {'id': 'AMDPLPA', 'deprecated': False},
+    'aml': {'id': 'AML', 'deprecated': False},
+    'aml-glslang': {'id': 'AML-glslang', 'deprecated': False},
+    'ampas': {'id': 'AMPAS', 'deprecated': False},
+    'antlr-pd': {'id': 'ANTLR-PD', 'deprecated': False},
+    'antlr-pd-fallback': {'id': 'ANTLR-PD-fallback', 'deprecated': False},
+    'any-osi': {'id': 'any-OSI', 'deprecated': False},
+    'apache-1.0': {'id': 'Apache-1.0', 'deprecated': False},
+    'apache-1.1': {'id': 'Apache-1.1', 'deprecated': False},
+    'apache-2.0': {'id': 'Apache-2.0', 'deprecated': False},
+    'apafml': {'id': 'APAFML', 'deprecated': False},
+    'apl-1.0': {'id': 'APL-1.0', 'deprecated': False},
+    'app-s2p': {'id': 'App-s2p', 'deprecated': False},
+    'apsl-1.0': {'id': 'APSL-1.0', 'deprecated': False},
+    'apsl-1.1': {'id': 'APSL-1.1', 'deprecated': False},
+    'apsl-1.2': {'id': 'APSL-1.2', 'deprecated': False},
+    'apsl-2.0': {'id': 'APSL-2.0', 'deprecated': False},
+    'arphic-1999': {'id': 'Arphic-1999', 'deprecated': False},
+    'artistic-1.0': {'id': 'Artistic-1.0', 'deprecated': False},
+    'artistic-1.0-cl8': {'id': 'Artistic-1.0-cl8', 'deprecated': False},
+    'artistic-1.0-perl': {'id': 'Artistic-1.0-Perl', 'deprecated': False},
+    'artistic-2.0': {'id': 'Artistic-2.0', 'deprecated': False},
+    'aswf-digital-assets-1.0': {'id': 'ASWF-Digital-Assets-1.0', 'deprecated': False},
+    'aswf-digital-assets-1.1': {'id': 'ASWF-Digital-Assets-1.1', 'deprecated': False},
+    'baekmuk': {'id': 'Baekmuk', 'deprecated': False},
+    'bahyph': {'id': 'Bahyph', 'deprecated': False},
+    'barr': {'id': 'Barr', 'deprecated': False},
+    'bcrypt-solar-designer': {'id': 'bcrypt-Solar-Designer', 'deprecated': False},
+    'beerware': {'id': 'Beerware', 'deprecated': False},
+    'bitstream-charter': {'id': 'Bitstream-Charter', 'deprecated': False},
+    'bitstream-vera': {'id': 'Bitstream-Vera', 'deprecated': False},
+    'bittorrent-1.0': {'id': 'BitTorrent-1.0', 'deprecated': False},
+    'bittorrent-1.1': {'id': 'BitTorrent-1.1', 'deprecated': False},
+    'blessing': {'id': 'blessing', 'deprecated': False},
+    'blueoak-1.0.0': {'id': 'BlueOak-1.0.0', 'deprecated': False},
+    'boehm-gc': {'id': 'Boehm-GC', 'deprecated': False},
+    'borceux': {'id': 'Borceux', 'deprecated': False},
+    'brian-gladman-2-clause': {'id': 'Brian-Gladman-2-Clause', 'deprecated': False},
+    'brian-gladman-3-clause': {'id': 'Brian-Gladman-3-Clause', 'deprecated': False},
+    'bsd-1-clause': {'id': 'BSD-1-Clause', 'deprecated': False},
+    'bsd-2-clause': {'id': 'BSD-2-Clause', 'deprecated': False},
+    'bsd-2-clause-darwin': {'id': 'BSD-2-Clause-Darwin', 'deprecated': False},
+    'bsd-2-clause-first-lines': {'id': 'BSD-2-Clause-first-lines', 'deprecated': False},
+    'bsd-2-clause-freebsd': {'id': 'BSD-2-Clause-FreeBSD', 'deprecated': True},
+    'bsd-2-clause-netbsd': {'id': 'BSD-2-Clause-NetBSD', 'deprecated': True},
+    'bsd-2-clause-patent': {'id': 'BSD-2-Clause-Patent', 'deprecated': False},
+    'bsd-2-clause-views': {'id': 'BSD-2-Clause-Views', 'deprecated': False},
+    'bsd-3-clause': {'id': 'BSD-3-Clause', 'deprecated': False},
+    'bsd-3-clause-acpica': {'id': 'BSD-3-Clause-acpica', 'deprecated': False},
+    'bsd-3-clause-attribution': {'id': 'BSD-3-Clause-Attribution', 'deprecated': False},
+    'bsd-3-clause-clear': {'id': 'BSD-3-Clause-Clear', 'deprecated': False},
+    'bsd-3-clause-flex': {'id': 'BSD-3-Clause-flex', 'deprecated': False},
+    'bsd-3-clause-hp': {'id': 'BSD-3-Clause-HP', 'deprecated': False},
+    'bsd-3-clause-lbnl': {'id': 'BSD-3-Clause-LBNL', 'deprecated': False},
+    'bsd-3-clause-modification': {'id': 'BSD-3-Clause-Modification', 'deprecated': False},
+    'bsd-3-clause-no-military-license': {'id': 'BSD-3-Clause-No-Military-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license': {'id': 'BSD-3-Clause-No-Nuclear-License', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-license-2014': {'id': 'BSD-3-Clause-No-Nuclear-License-2014', 'deprecated': False},
+    'bsd-3-clause-no-nuclear-warranty': {'id': 'BSD-3-Clause-No-Nuclear-Warranty', 'deprecated': False},
+    'bsd-3-clause-open-mpi': {'id': 'BSD-3-Clause-Open-MPI', 'deprecated': False},
+    'bsd-3-clause-sun': {'id': 'BSD-3-Clause-Sun', 'deprecated': False},
+    'bsd-4-clause': {'id': 'BSD-4-Clause', 'deprecated': False},
+    'bsd-4-clause-shortened': {'id': 'BSD-4-Clause-Shortened', 'deprecated': False},
+    'bsd-4-clause-uc': {'id': 'BSD-4-Clause-UC', 'deprecated': False},
+    'bsd-4.3reno': {'id': 'BSD-4.3RENO', 'deprecated': False},
+    'bsd-4.3tahoe': {'id': 'BSD-4.3TAHOE', 'deprecated': False},
+    'bsd-advertising-acknowledgement': {'id': 'BSD-Advertising-Acknowledgement', 'deprecated': False},
+    'bsd-attribution-hpnd-disclaimer': {'id': 'BSD-Attribution-HPND-disclaimer', 'deprecated': False},
+    'bsd-inferno-nettverk': {'id': 'BSD-Inferno-Nettverk', 'deprecated': False},
+    'bsd-protection': {'id': 'BSD-Protection', 'deprecated': False},
+    'bsd-source-beginning-file': {'id': 'BSD-Source-beginning-file', 'deprecated': False},
+    'bsd-source-code': {'id': 'BSD-Source-Code', 'deprecated': False},
+    'bsd-systemics': {'id': 'BSD-Systemics', 'deprecated': False},
+    'bsd-systemics-w3works': {'id': 'BSD-Systemics-W3Works', 'deprecated': False},
+    'bsl-1.0': {'id': 'BSL-1.0', 'deprecated': False},
+    'busl-1.1': {'id': 'BUSL-1.1', 'deprecated': False},
+    'bzip2-1.0.5': {'id': 'bzip2-1.0.5', 'deprecated': True},
+    'bzip2-1.0.6': {'id': 'bzip2-1.0.6', 'deprecated': False},
+    'c-uda-1.0': {'id': 'C-UDA-1.0', 'deprecated': False},
+    'cal-1.0': {'id': 'CAL-1.0', 'deprecated': False},
+    'cal-1.0-combined-work-exception': {'id': 'CAL-1.0-Combined-Work-Exception', 'deprecated': False},
+    'caldera': {'id': 'Caldera', 'deprecated': False},
+    'caldera-no-preamble': {'id': 'Caldera-no-preamble', 'deprecated': False},
+    'catharon': {'id': 'Catharon', 'deprecated': False},
+    'catosl-1.1': {'id': 'CATOSL-1.1', 'deprecated': False},
+    'cc-by-1.0': {'id': 'CC-BY-1.0', 'deprecated': False},
+    'cc-by-2.0': {'id': 'CC-BY-2.0', 'deprecated': False},
+    'cc-by-2.5': {'id': 'CC-BY-2.5', 'deprecated': False},
+    'cc-by-2.5-au': {'id': 'CC-BY-2.5-AU', 'deprecated': False},
+    'cc-by-3.0': {'id': 'CC-BY-3.0', 'deprecated': False},
+    'cc-by-3.0-at': {'id': 'CC-BY-3.0-AT', 'deprecated': False},
+    'cc-by-3.0-au': {'id': 'CC-BY-3.0-AU', 'deprecated': False},
+    'cc-by-3.0-de': {'id': 'CC-BY-3.0-DE', 'deprecated': False},
+    'cc-by-3.0-igo': {'id': 'CC-BY-3.0-IGO', 'deprecated': False},
+    'cc-by-3.0-nl': {'id': 'CC-BY-3.0-NL', 'deprecated': False},
+    'cc-by-3.0-us': {'id': 'CC-BY-3.0-US', 'deprecated': False},
+    'cc-by-4.0': {'id': 'CC-BY-4.0', 'deprecated': False},
+    'cc-by-nc-1.0': {'id': 'CC-BY-NC-1.0', 'deprecated': False},
+    'cc-by-nc-2.0': {'id': 'CC-BY-NC-2.0', 'deprecated': False},
+    'cc-by-nc-2.5': {'id': 'CC-BY-NC-2.5', 'deprecated': False},
+    'cc-by-nc-3.0': {'id': 'CC-BY-NC-3.0', 'deprecated': False},
+    'cc-by-nc-3.0-de': {'id': 'CC-BY-NC-3.0-DE', 'deprecated': False},
+    'cc-by-nc-4.0': {'id': 'CC-BY-NC-4.0', 'deprecated': False},
+    'cc-by-nc-nd-1.0': {'id': 'CC-BY-NC-ND-1.0', 'deprecated': False},
+    'cc-by-nc-nd-2.0': {'id': 'CC-BY-NC-ND-2.0', 'deprecated': False},
+    'cc-by-nc-nd-2.5': {'id': 'CC-BY-NC-ND-2.5', 'deprecated': False},
+    'cc-by-nc-nd-3.0': {'id': 'CC-BY-NC-ND-3.0', 'deprecated': False},
+    'cc-by-nc-nd-3.0-de': {'id': 'CC-BY-NC-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nc-nd-3.0-igo': {'id': 'CC-BY-NC-ND-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-nd-4.0': {'id': 'CC-BY-NC-ND-4.0', 'deprecated': False},
+    'cc-by-nc-sa-1.0': {'id': 'CC-BY-NC-SA-1.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0': {'id': 'CC-BY-NC-SA-2.0', 'deprecated': False},
+    'cc-by-nc-sa-2.0-de': {'id': 'CC-BY-NC-SA-2.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-2.0-fr': {'id': 'CC-BY-NC-SA-2.0-FR', 'deprecated': False},
+    'cc-by-nc-sa-2.0-uk': {'id': 'CC-BY-NC-SA-2.0-UK', 'deprecated': False},
+    'cc-by-nc-sa-2.5': {'id': 'CC-BY-NC-SA-2.5', 'deprecated': False},
+    'cc-by-nc-sa-3.0': {'id': 'CC-BY-NC-SA-3.0', 'deprecated': False},
+    'cc-by-nc-sa-3.0-de': {'id': 'CC-BY-NC-SA-3.0-DE', 'deprecated': False},
+    'cc-by-nc-sa-3.0-igo': {'id': 'CC-BY-NC-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-nc-sa-4.0': {'id': 'CC-BY-NC-SA-4.0', 'deprecated': False},
+    'cc-by-nd-1.0': {'id': 'CC-BY-ND-1.0', 'deprecated': False},
+    'cc-by-nd-2.0': {'id': 'CC-BY-ND-2.0', 'deprecated': False},
+    'cc-by-nd-2.5': {'id': 'CC-BY-ND-2.5', 'deprecated': False},
+    'cc-by-nd-3.0': {'id': 'CC-BY-ND-3.0', 'deprecated': False},
+    'cc-by-nd-3.0-de': {'id': 'CC-BY-ND-3.0-DE', 'deprecated': False},
+    'cc-by-nd-4.0': {'id': 'CC-BY-ND-4.0', 'deprecated': False},
+    'cc-by-sa-1.0': {'id': 'CC-BY-SA-1.0', 'deprecated': False},
+    'cc-by-sa-2.0': {'id': 'CC-BY-SA-2.0', 'deprecated': False},
+    'cc-by-sa-2.0-uk': {'id': 'CC-BY-SA-2.0-UK', 'deprecated': False},
+    'cc-by-sa-2.1-jp': {'id': 'CC-BY-SA-2.1-JP', 'deprecated': False},
+    'cc-by-sa-2.5': {'id': 'CC-BY-SA-2.5', 'deprecated': False},
+    'cc-by-sa-3.0': {'id': 'CC-BY-SA-3.0', 'deprecated': False},
+    'cc-by-sa-3.0-at': {'id': 'CC-BY-SA-3.0-AT', 'deprecated': False},
+    'cc-by-sa-3.0-de': {'id': 'CC-BY-SA-3.0-DE', 'deprecated': False},
+    'cc-by-sa-3.0-igo': {'id': 'CC-BY-SA-3.0-IGO', 'deprecated': False},
+    'cc-by-sa-4.0': {'id': 'CC-BY-SA-4.0', 'deprecated': False},
+    'cc-pddc': {'id': 'CC-PDDC', 'deprecated': False},
+    'cc0-1.0': {'id': 'CC0-1.0', 'deprecated': False},
+    'cddl-1.0': {'id': 'CDDL-1.0', 'deprecated': False},
+    'cddl-1.1': {'id': 'CDDL-1.1', 'deprecated': False},
+    'cdl-1.0': {'id': 'CDL-1.0', 'deprecated': False},
+    'cdla-permissive-1.0': {'id': 'CDLA-Permissive-1.0', 'deprecated': False},
+    'cdla-permissive-2.0': {'id': 'CDLA-Permissive-2.0', 'deprecated': False},
+    'cdla-sharing-1.0': {'id': 'CDLA-Sharing-1.0', 'deprecated': False},
+    'cecill-1.0': {'id': 'CECILL-1.0', 'deprecated': False},
+    'cecill-1.1': {'id': 'CECILL-1.1', 'deprecated': False},
+    'cecill-2.0': {'id': 'CECILL-2.0', 'deprecated': False},
+    'cecill-2.1': {'id': 'CECILL-2.1', 'deprecated': False},
+    'cecill-b': {'id': 'CECILL-B', 'deprecated': False},
+    'cecill-c': {'id': 'CECILL-C', 'deprecated': False},
+    'cern-ohl-1.1': {'id': 'CERN-OHL-1.1', 'deprecated': False},
+    'cern-ohl-1.2': {'id': 'CERN-OHL-1.2', 'deprecated': False},
+    'cern-ohl-p-2.0': {'id': 'CERN-OHL-P-2.0', 'deprecated': False},
+    'cern-ohl-s-2.0': {'id': 'CERN-OHL-S-2.0', 'deprecated': False},
+    'cern-ohl-w-2.0': {'id': 'CERN-OHL-W-2.0', 'deprecated': False},
+    'cfitsio': {'id': 'CFITSIO', 'deprecated': False},
+    'check-cvs': {'id': 'check-cvs', 'deprecated': False},
+    'checkmk': {'id': 'checkmk', 'deprecated': False},
+    'clartistic': {'id': 'ClArtistic', 'deprecated': False},
+    'clips': {'id': 'Clips', 'deprecated': False},
+    'cmu-mach': {'id': 'CMU-Mach', 'deprecated': False},
+    'cmu-mach-nodoc': {'id': 'CMU-Mach-nodoc', 'deprecated': False},
+    'cnri-jython': {'id': 'CNRI-Jython', 'deprecated': False},
+    'cnri-python': {'id': 'CNRI-Python', 'deprecated': False},
+    'cnri-python-gpl-compatible': {'id': 'CNRI-Python-GPL-Compatible', 'deprecated': False},
+    'coil-1.0': {'id': 'COIL-1.0', 'deprecated': False},
+    'community-spec-1.0': {'id': 'Community-Spec-1.0', 'deprecated': False},
+    'condor-1.1': {'id': 'Condor-1.1', 'deprecated': False},
+    'copyleft-next-0.3.0': {'id': 'copyleft-next-0.3.0', 'deprecated': False},
+    'copyleft-next-0.3.1': {'id': 'copyleft-next-0.3.1', 'deprecated': False},
+    'cornell-lossless-jpeg': {'id': 'Cornell-Lossless-JPEG', 'deprecated': False},
+    'cpal-1.0': {'id': 'CPAL-1.0', 'deprecated': False},
+    'cpl-1.0': {'id': 'CPL-1.0', 'deprecated': False},
+    'cpol-1.02': {'id': 'CPOL-1.02', 'deprecated': False},
+    'cronyx': {'id': 'Cronyx', 'deprecated': False},
+    'crossword': {'id': 'Crossword', 'deprecated': False},
+    'crystalstacker': {'id': 'CrystalStacker', 'deprecated': False},
+    'cua-opl-1.0': {'id': 'CUA-OPL-1.0', 'deprecated': False},
+    'cube': {'id': 'Cube', 'deprecated': False},
+    'curl': {'id': 'curl', 'deprecated': False},
+    'cve-tou': {'id': 'cve-tou', 'deprecated': False},
+    'd-fsl-1.0': {'id': 'D-FSL-1.0', 'deprecated': False},
+    'dec-3-clause': {'id': 'DEC-3-Clause', 'deprecated': False},
+    'diffmark': {'id': 'diffmark', 'deprecated': False},
+    'dl-de-by-2.0': {'id': 'DL-DE-BY-2.0', 'deprecated': False},
+    'dl-de-zero-2.0': {'id': 'DL-DE-ZERO-2.0', 'deprecated': False},
+    'doc': {'id': 'DOC', 'deprecated': False},
+    'docbook-schema': {'id': 'DocBook-Schema', 'deprecated': False},
+    'docbook-xml': {'id': 'DocBook-XML', 'deprecated': False},
+    'dotseqn': {'id': 'Dotseqn', 'deprecated': False},
+    'drl-1.0': {'id': 'DRL-1.0', 'deprecated': False},
+    'drl-1.1': {'id': 'DRL-1.1', 'deprecated': False},
+    'dsdp': {'id': 'DSDP', 'deprecated': False},
+    'dtoa': {'id': 'dtoa', 'deprecated': False},
+    'dvipdfm': {'id': 'dvipdfm', 'deprecated': False},
+    'ecl-1.0': {'id': 'ECL-1.0', 'deprecated': False},
+    'ecl-2.0': {'id': 'ECL-2.0', 'deprecated': False},
+    'ecos-2.0': {'id': 'eCos-2.0', 'deprecated': True},
+    'efl-1.0': {'id': 'EFL-1.0', 'deprecated': False},
+    'efl-2.0': {'id': 'EFL-2.0', 'deprecated': False},
+    'egenix': {'id': 'eGenix', 'deprecated': False},
+    'elastic-2.0': {'id': 'Elastic-2.0', 'deprecated': False},
+    'entessa': {'id': 'Entessa', 'deprecated': False},
+    'epics': {'id': 'EPICS', 'deprecated': False},
+    'epl-1.0': {'id': 'EPL-1.0', 'deprecated': False},
+    'epl-2.0': {'id': 'EPL-2.0', 'deprecated': False},
+    'erlpl-1.1': {'id': 'ErlPL-1.1', 'deprecated': False},
+    'etalab-2.0': {'id': 'etalab-2.0', 'deprecated': False},
+    'eudatagrid': {'id': 'EUDatagrid', 'deprecated': False},
+    'eupl-1.0': {'id': 'EUPL-1.0', 'deprecated': False},
+    'eupl-1.1': {'id': 'EUPL-1.1', 'deprecated': False},
+    'eupl-1.2': {'id': 'EUPL-1.2', 'deprecated': False},
+    'eurosym': {'id': 'Eurosym', 'deprecated': False},
+    'fair': {'id': 'Fair', 'deprecated': False},
+    'fbm': {'id': 'FBM', 'deprecated': False},
+    'fdk-aac': {'id': 'FDK-AAC', 'deprecated': False},
+    'ferguson-twofish': {'id': 'Ferguson-Twofish', 'deprecated': False},
+    'frameworx-1.0': {'id': 'Frameworx-1.0', 'deprecated': False},
+    'freebsd-doc': {'id': 'FreeBSD-DOC', 'deprecated': False},
+    'freeimage': {'id': 'FreeImage', 'deprecated': False},
+    'fsfap': {'id': 'FSFAP', 'deprecated': False},
+    'fsfap-no-warranty-disclaimer': {'id': 'FSFAP-no-warranty-disclaimer', 'deprecated': False},
+    'fsful': {'id': 'FSFUL', 'deprecated': False},
+    'fsfullr': {'id': 'FSFULLR', 'deprecated': False},
+    'fsfullrwd': {'id': 'FSFULLRWD', 'deprecated': False},
+    'ftl': {'id': 'FTL', 'deprecated': False},
+    'furuseth': {'id': 'Furuseth', 'deprecated': False},
+    'fwlw': {'id': 'fwlw', 'deprecated': False},
+    'gcr-docs': {'id': 'GCR-docs', 'deprecated': False},
+    'gd': {'id': 'GD', 'deprecated': False},
+    'gfdl-1.1': {'id': 'GFDL-1.1', 'deprecated': True},
+    'gfdl-1.1-invariants-only': {'id': 'GFDL-1.1-invariants-only', 'deprecated': False},
+    'gfdl-1.1-invariants-or-later': {'id': 'GFDL-1.1-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-no-invariants-only': {'id': 'GFDL-1.1-no-invariants-only', 'deprecated': False},
+    'gfdl-1.1-no-invariants-or-later': {'id': 'GFDL-1.1-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.1-only': {'id': 'GFDL-1.1-only', 'deprecated': False},
+    'gfdl-1.1-or-later': {'id': 'GFDL-1.1-or-later', 'deprecated': False},
+    'gfdl-1.2': {'id': 'GFDL-1.2', 'deprecated': True},
+    'gfdl-1.2-invariants-only': {'id': 'GFDL-1.2-invariants-only', 'deprecated': False},
+    'gfdl-1.2-invariants-or-later': {'id': 'GFDL-1.2-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-no-invariants-only': {'id': 'GFDL-1.2-no-invariants-only', 'deprecated': False},
+    'gfdl-1.2-no-invariants-or-later': {'id': 'GFDL-1.2-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.2-only': {'id': 'GFDL-1.2-only', 'deprecated': False},
+    'gfdl-1.2-or-later': {'id': 'GFDL-1.2-or-later', 'deprecated': False},
+    'gfdl-1.3': {'id': 'GFDL-1.3', 'deprecated': True},
+    'gfdl-1.3-invariants-only': {'id': 'GFDL-1.3-invariants-only', 'deprecated': False},
+    'gfdl-1.3-invariants-or-later': {'id': 'GFDL-1.3-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-no-invariants-only': {'id': 'GFDL-1.3-no-invariants-only', 'deprecated': False},
+    'gfdl-1.3-no-invariants-or-later': {'id': 'GFDL-1.3-no-invariants-or-later', 'deprecated': False},
+    'gfdl-1.3-only': {'id': 'GFDL-1.3-only', 'deprecated': False},
+    'gfdl-1.3-or-later': {'id': 'GFDL-1.3-or-later', 'deprecated': False},
+    'giftware': {'id': 'Giftware', 'deprecated': False},
+    'gl2ps': {'id': 'GL2PS', 'deprecated': False},
+    'glide': {'id': 'Glide', 'deprecated': False},
+    'glulxe': {'id': 'Glulxe', 'deprecated': False},
+    'glwtpl': {'id': 'GLWTPL', 'deprecated': False},
+    'gnuplot': {'id': 'gnuplot', 'deprecated': False},
+    'gpl-1.0': {'id': 'GPL-1.0', 'deprecated': True},
+    'gpl-1.0+': {'id': 'GPL-1.0+', 'deprecated': True},
+    'gpl-1.0-only': {'id': 'GPL-1.0-only', 'deprecated': False},
+    'gpl-1.0-or-later': {'id': 'GPL-1.0-or-later', 'deprecated': False},
+    'gpl-2.0': {'id': 'GPL-2.0', 'deprecated': True},
+    'gpl-2.0+': {'id': 'GPL-2.0+', 'deprecated': True},
+    'gpl-2.0-only': {'id': 'GPL-2.0-only', 'deprecated': False},
+    'gpl-2.0-or-later': {'id': 'GPL-2.0-or-later', 'deprecated': False},
+    'gpl-2.0-with-autoconf-exception': {'id': 'GPL-2.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-2.0-with-bison-exception': {'id': 'GPL-2.0-with-bison-exception', 'deprecated': True},
+    'gpl-2.0-with-classpath-exception': {'id': 'GPL-2.0-with-classpath-exception', 'deprecated': True},
+    'gpl-2.0-with-font-exception': {'id': 'GPL-2.0-with-font-exception', 'deprecated': True},
+    'gpl-2.0-with-gcc-exception': {'id': 'GPL-2.0-with-GCC-exception', 'deprecated': True},
+    'gpl-3.0': {'id': 'GPL-3.0', 'deprecated': True},
+    'gpl-3.0+': {'id': 'GPL-3.0+', 'deprecated': True},
+    'gpl-3.0-only': {'id': 'GPL-3.0-only', 'deprecated': False},
+    'gpl-3.0-or-later': {'id': 'GPL-3.0-or-later', 'deprecated': False},
+    'gpl-3.0-with-autoconf-exception': {'id': 'GPL-3.0-with-autoconf-exception', 'deprecated': True},
+    'gpl-3.0-with-gcc-exception': {'id': 'GPL-3.0-with-GCC-exception', 'deprecated': True},
+    'graphics-gems': {'id': 'Graphics-Gems', 'deprecated': False},
+    'gsoap-1.3b': {'id': 'gSOAP-1.3b', 'deprecated': False},
+    'gtkbook': {'id': 'gtkbook', 'deprecated': False},
+    'gutmann': {'id': 'Gutmann', 'deprecated': False},
+    'haskellreport': {'id': 'HaskellReport', 'deprecated': False},
+    'hdparm': {'id': 'hdparm', 'deprecated': False},
+    'hidapi': {'id': 'HIDAPI', 'deprecated': False},
+    'hippocratic-2.1': {'id': 'Hippocratic-2.1', 'deprecated': False},
+    'hp-1986': {'id': 'HP-1986', 'deprecated': False},
+    'hp-1989': {'id': 'HP-1989', 'deprecated': False},
+    'hpnd': {'id': 'HPND', 'deprecated': False},
+    'hpnd-dec': {'id': 'HPND-DEC', 'deprecated': False},
+    'hpnd-doc': {'id': 'HPND-doc', 'deprecated': False},
+    'hpnd-doc-sell': {'id': 'HPND-doc-sell', 'deprecated': False},
+    'hpnd-export-us': {'id': 'HPND-export-US', 'deprecated': False},
+    'hpnd-export-us-acknowledgement': {'id': 'HPND-export-US-acknowledgement', 'deprecated': False},
+    'hpnd-export-us-modify': {'id': 'HPND-export-US-modify', 'deprecated': False},
+    'hpnd-export2-us': {'id': 'HPND-export2-US', 'deprecated': False},
+    'hpnd-fenneberg-livingston': {'id': 'HPND-Fenneberg-Livingston', 'deprecated': False},
+    'hpnd-inria-imag': {'id': 'HPND-INRIA-IMAG', 'deprecated': False},
+    'hpnd-intel': {'id': 'HPND-Intel', 'deprecated': False},
+    'hpnd-kevlin-henney': {'id': 'HPND-Kevlin-Henney', 'deprecated': False},
+    'hpnd-markus-kuhn': {'id': 'HPND-Markus-Kuhn', 'deprecated': False},
+    'hpnd-merchantability-variant': {'id': 'HPND-merchantability-variant', 'deprecated': False},
+    'hpnd-mit-disclaimer': {'id': 'HPND-MIT-disclaimer', 'deprecated': False},
+    'hpnd-netrek': {'id': 'HPND-Netrek', 'deprecated': False},
+    'hpnd-pbmplus': {'id': 'HPND-Pbmplus', 'deprecated': False},
+    'hpnd-sell-mit-disclaimer-xserver': {'id': 'HPND-sell-MIT-disclaimer-xserver', 'deprecated': False},
+    'hpnd-sell-regexpr': {'id': 'HPND-sell-regexpr', 'deprecated': False},
+    'hpnd-sell-variant': {'id': 'HPND-sell-variant', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer': {'id': 'HPND-sell-variant-MIT-disclaimer', 'deprecated': False},
+    'hpnd-sell-variant-mit-disclaimer-rev': {'id': 'HPND-sell-variant-MIT-disclaimer-rev', 'deprecated': False},
+    'hpnd-uc': {'id': 'HPND-UC', 'deprecated': False},
+    'hpnd-uc-export-us': {'id': 'HPND-UC-export-US', 'deprecated': False},
+    'htmltidy': {'id': 'HTMLTIDY', 'deprecated': False},
+    'ibm-pibs': {'id': 'IBM-pibs', 'deprecated': False},
+    'icu': {'id': 'ICU', 'deprecated': False},
+    'iec-code-components-eula': {'id': 'IEC-Code-Components-EULA', 'deprecated': False},
+    'ijg': {'id': 'IJG', 'deprecated': False},
+    'ijg-short': {'id': 'IJG-short', 'deprecated': False},
+    'imagemagick': {'id': 'ImageMagick', 'deprecated': False},
+    'imatix': {'id': 'iMatix', 'deprecated': False},
+    'imlib2': {'id': 'Imlib2', 'deprecated': False},
+    'info-zip': {'id': 'Info-ZIP', 'deprecated': False},
+    'inner-net-2.0': {'id': 'Inner-Net-2.0', 'deprecated': False},
+    'intel': {'id': 'Intel', 'deprecated': False},
+    'intel-acpi': {'id': 'Intel-ACPI', 'deprecated': False},
+    'interbase-1.0': {'id': 'Interbase-1.0', 'deprecated': False},
+    'ipa': {'id': 'IPA', 'deprecated': False},
+    'ipl-1.0': {'id': 'IPL-1.0', 'deprecated': False},
+    'isc': {'id': 'ISC', 'deprecated': False},
+    'isc-veillard': {'id': 'ISC-Veillard', 'deprecated': False},
+    'jam': {'id': 'Jam', 'deprecated': False},
+    'jasper-2.0': {'id': 'JasPer-2.0', 'deprecated': False},
+    'jpl-image': {'id': 'JPL-image', 'deprecated': False},
+    'jpnic': {'id': 'JPNIC', 'deprecated': False},
+    'json': {'id': 'JSON', 'deprecated': False},
+    'kastrup': {'id': 'Kastrup', 'deprecated': False},
+    'kazlib': {'id': 'Kazlib', 'deprecated': False},
+    'knuth-ctan': {'id': 'Knuth-CTAN', 'deprecated': False},
+    'lal-1.2': {'id': 'LAL-1.2', 'deprecated': False},
+    'lal-1.3': {'id': 'LAL-1.3', 'deprecated': False},
+    'latex2e': {'id': 'Latex2e', 'deprecated': False},
+    'latex2e-translated-notice': {'id': 'Latex2e-translated-notice', 'deprecated': False},
+    'leptonica': {'id': 'Leptonica', 'deprecated': False},
+    'lgpl-2.0': {'id': 'LGPL-2.0', 'deprecated': True},
+    'lgpl-2.0+': {'id': 'LGPL-2.0+', 'deprecated': True},
+    'lgpl-2.0-only': {'id': 'LGPL-2.0-only', 'deprecated': False},
+    'lgpl-2.0-or-later': {'id': 'LGPL-2.0-or-later', 'deprecated': False},
+    'lgpl-2.1': {'id': 'LGPL-2.1', 'deprecated': True},
+    'lgpl-2.1+': {'id': 'LGPL-2.1+', 'deprecated': True},
+    'lgpl-2.1-only': {'id': 'LGPL-2.1-only', 'deprecated': False},
+    'lgpl-2.1-or-later': {'id': 'LGPL-2.1-or-later', 'deprecated': False},
+    'lgpl-3.0': {'id': 'LGPL-3.0', 'deprecated': True},
+    'lgpl-3.0+': {'id': 'LGPL-3.0+', 'deprecated': True},
+    'lgpl-3.0-only': {'id': 'LGPL-3.0-only', 'deprecated': False},
+    'lgpl-3.0-or-later': {'id': 'LGPL-3.0-or-later', 'deprecated': False},
+    'lgpllr': {'id': 'LGPLLR', 'deprecated': False},
+    'libpng': {'id': 'Libpng', 'deprecated': False},
+    'libpng-2.0': {'id': 'libpng-2.0', 'deprecated': False},
+    'libselinux-1.0': {'id': 'libselinux-1.0', 'deprecated': False},
+    'libtiff': {'id': 'libtiff', 'deprecated': False},
+    'libutil-david-nugent': {'id': 'libutil-David-Nugent', 'deprecated': False},
+    'liliq-p-1.1': {'id': 'LiLiQ-P-1.1', 'deprecated': False},
+    'liliq-r-1.1': {'id': 'LiLiQ-R-1.1', 'deprecated': False},
+    'liliq-rplus-1.1': {'id': 'LiLiQ-Rplus-1.1', 'deprecated': False},
+    'linux-man-pages-1-para': {'id': 'Linux-man-pages-1-para', 'deprecated': False},
+    'linux-man-pages-copyleft': {'id': 'Linux-man-pages-copyleft', 'deprecated': False},
+    'linux-man-pages-copyleft-2-para': {'id': 'Linux-man-pages-copyleft-2-para', 'deprecated': False},
+    'linux-man-pages-copyleft-var': {'id': 'Linux-man-pages-copyleft-var', 'deprecated': False},
+    'linux-openib': {'id': 'Linux-OpenIB', 'deprecated': False},
+    'loop': {'id': 'LOOP', 'deprecated': False},
+    'lpd-document': {'id': 'LPD-document', 'deprecated': False},
+    'lpl-1.0': {'id': 'LPL-1.0', 'deprecated': False},
+    'lpl-1.02': {'id': 'LPL-1.02', 'deprecated': False},
+    'lppl-1.0': {'id': 'LPPL-1.0', 'deprecated': False},
+    'lppl-1.1': {'id': 'LPPL-1.1', 'deprecated': False},
+    'lppl-1.2': {'id': 'LPPL-1.2', 'deprecated': False},
+    'lppl-1.3a': {'id': 'LPPL-1.3a', 'deprecated': False},
+    'lppl-1.3c': {'id': 'LPPL-1.3c', 'deprecated': False},
+    'lsof': {'id': 'lsof', 'deprecated': False},
+    'lucida-bitmap-fonts': {'id': 'Lucida-Bitmap-Fonts', 'deprecated': False},
+    'lzma-sdk-9.11-to-9.20': {'id': 'LZMA-SDK-9.11-to-9.20', 'deprecated': False},
+    'lzma-sdk-9.22': {'id': 'LZMA-SDK-9.22', 'deprecated': False},
+    'mackerras-3-clause': {'id': 'Mackerras-3-Clause', 'deprecated': False},
+    'mackerras-3-clause-acknowledgment': {'id': 'Mackerras-3-Clause-acknowledgment', 'deprecated': False},
+    'magaz': {'id': 'magaz', 'deprecated': False},
+    'mailprio': {'id': 'mailprio', 'deprecated': False},
+    'makeindex': {'id': 'MakeIndex', 'deprecated': False},
+    'martin-birgmeier': {'id': 'Martin-Birgmeier', 'deprecated': False},
+    'mcphee-slideshow': {'id': 'McPhee-slideshow', 'deprecated': False},
+    'metamail': {'id': 'metamail', 'deprecated': False},
+    'minpack': {'id': 'Minpack', 'deprecated': False},
+    'miros': {'id': 'MirOS', 'deprecated': False},
+    'mit': {'id': 'MIT', 'deprecated': False},
+    'mit-0': {'id': 'MIT-0', 'deprecated': False},
+    'mit-advertising': {'id': 'MIT-advertising', 'deprecated': False},
+    'mit-cmu': {'id': 'MIT-CMU', 'deprecated': False},
+    'mit-enna': {'id': 'MIT-enna', 'deprecated': False},
+    'mit-feh': {'id': 'MIT-feh', 'deprecated': False},
+    'mit-festival': {'id': 'MIT-Festival', 'deprecated': False},
+    'mit-khronos-old': {'id': 'MIT-Khronos-old', 'deprecated': False},
+    'mit-modern-variant': {'id': 'MIT-Modern-Variant', 'deprecated': False},
+    'mit-open-group': {'id': 'MIT-open-group', 'deprecated': False},
+    'mit-testregex': {'id': 'MIT-testregex', 'deprecated': False},
+    'mit-wu': {'id': 'MIT-Wu', 'deprecated': False},
+    'mitnfa': {'id': 'MITNFA', 'deprecated': False},
+    'mmixware': {'id': 'MMIXware', 'deprecated': False},
+    'motosoto': {'id': 'Motosoto', 'deprecated': False},
+    'mpeg-ssg': {'id': 'MPEG-SSG', 'deprecated': False},
+    'mpi-permissive': {'id': 'mpi-permissive', 'deprecated': False},
+    'mpich2': {'id': 'mpich2', 'deprecated': False},
+    'mpl-1.0': {'id': 'MPL-1.0', 'deprecated': False},
+    'mpl-1.1': {'id': 'MPL-1.1', 'deprecated': False},
+    'mpl-2.0': {'id': 'MPL-2.0', 'deprecated': False},
+    'mpl-2.0-no-copyleft-exception': {'id': 'MPL-2.0-no-copyleft-exception', 'deprecated': False},
+    'mplus': {'id': 'mplus', 'deprecated': False},
+    'ms-lpl': {'id': 'MS-LPL', 'deprecated': False},
+    'ms-pl': {'id': 'MS-PL', 'deprecated': False},
+    'ms-rl': {'id': 'MS-RL', 'deprecated': False},
+    'mtll': {'id': 'MTLL', 'deprecated': False},
+    'mulanpsl-1.0': {'id': 'MulanPSL-1.0', 'deprecated': False},
+    'mulanpsl-2.0': {'id': 'MulanPSL-2.0', 'deprecated': False},
+    'multics': {'id': 'Multics', 'deprecated': False},
+    'mup': {'id': 'Mup', 'deprecated': False},
+    'naist-2003': {'id': 'NAIST-2003', 'deprecated': False},
+    'nasa-1.3': {'id': 'NASA-1.3', 'deprecated': False},
+    'naumen': {'id': 'Naumen', 'deprecated': False},
+    'nbpl-1.0': {'id': 'NBPL-1.0', 'deprecated': False},
+    'ncbi-pd': {'id': 'NCBI-PD', 'deprecated': False},
+    'ncgl-uk-2.0': {'id': 'NCGL-UK-2.0', 'deprecated': False},
+    'ncl': {'id': 'NCL', 'deprecated': False},
+    'ncsa': {'id': 'NCSA', 'deprecated': False},
+    'net-snmp': {'id': 'Net-SNMP', 'deprecated': True},
+    'netcdf': {'id': 'NetCDF', 'deprecated': False},
+    'newsletr': {'id': 'Newsletr', 'deprecated': False},
+    'ngpl': {'id': 'NGPL', 'deprecated': False},
+    'nicta-1.0': {'id': 'NICTA-1.0', 'deprecated': False},
+    'nist-pd': {'id': 'NIST-PD', 'deprecated': False},
+    'nist-pd-fallback': {'id': 'NIST-PD-fallback', 'deprecated': False},
+    'nist-software': {'id': 'NIST-Software', 'deprecated': False},
+    'nlod-1.0': {'id': 'NLOD-1.0', 'deprecated': False},
+    'nlod-2.0': {'id': 'NLOD-2.0', 'deprecated': False},
+    'nlpl': {'id': 'NLPL', 'deprecated': False},
+    'nokia': {'id': 'Nokia', 'deprecated': False},
+    'nosl': {'id': 'NOSL', 'deprecated': False},
+    'noweb': {'id': 'Noweb', 'deprecated': False},
+    'npl-1.0': {'id': 'NPL-1.0', 'deprecated': False},
+    'npl-1.1': {'id': 'NPL-1.1', 'deprecated': False},
+    'nposl-3.0': {'id': 'NPOSL-3.0', 'deprecated': False},
+    'nrl': {'id': 'NRL', 'deprecated': False},
+    'ntp': {'id': 'NTP', 'deprecated': False},
+    'ntp-0': {'id': 'NTP-0', 'deprecated': False},
+    'nunit': {'id': 'Nunit', 'deprecated': True},
+    'o-uda-1.0': {'id': 'O-UDA-1.0', 'deprecated': False},
+    'oar': {'id': 'OAR', 'deprecated': False},
+    'occt-pl': {'id': 'OCCT-PL', 'deprecated': False},
+    'oclc-2.0': {'id': 'OCLC-2.0', 'deprecated': False},
+    'odbl-1.0': {'id': 'ODbL-1.0', 'deprecated': False},
+    'odc-by-1.0': {'id': 'ODC-By-1.0', 'deprecated': False},
+    'offis': {'id': 'OFFIS', 'deprecated': False},
+    'ofl-1.0': {'id': 'OFL-1.0', 'deprecated': False},
+    'ofl-1.0-no-rfn': {'id': 'OFL-1.0-no-RFN', 'deprecated': False},
+    'ofl-1.0-rfn': {'id': 'OFL-1.0-RFN', 'deprecated': False},
+    'ofl-1.1': {'id': 'OFL-1.1', 'deprecated': False},
+    'ofl-1.1-no-rfn': {'id': 'OFL-1.1-no-RFN', 'deprecated': False},
+    'ofl-1.1-rfn': {'id': 'OFL-1.1-RFN', 'deprecated': False},
+    'ogc-1.0': {'id': 'OGC-1.0', 'deprecated': False},
+    'ogdl-taiwan-1.0': {'id': 'OGDL-Taiwan-1.0', 'deprecated': False},
+    'ogl-canada-2.0': {'id': 'OGL-Canada-2.0', 'deprecated': False},
+    'ogl-uk-1.0': {'id': 'OGL-UK-1.0', 'deprecated': False},
+    'ogl-uk-2.0': {'id': 'OGL-UK-2.0', 'deprecated': False},
+    'ogl-uk-3.0': {'id': 'OGL-UK-3.0', 'deprecated': False},
+    'ogtsl': {'id': 'OGTSL', 'deprecated': False},
+    'oldap-1.1': {'id': 'OLDAP-1.1', 'deprecated': False},
+    'oldap-1.2': {'id': 'OLDAP-1.2', 'deprecated': False},
+    'oldap-1.3': {'id': 'OLDAP-1.3', 'deprecated': False},
+    'oldap-1.4': {'id': 'OLDAP-1.4', 'deprecated': False},
+    'oldap-2.0': {'id': 'OLDAP-2.0', 'deprecated': False},
+    'oldap-2.0.1': {'id': 'OLDAP-2.0.1', 'deprecated': False},
+    'oldap-2.1': {'id': 'OLDAP-2.1', 'deprecated': False},
+    'oldap-2.2': {'id': 'OLDAP-2.2', 'deprecated': False},
+    'oldap-2.2.1': {'id': 'OLDAP-2.2.1', 'deprecated': False},
+    'oldap-2.2.2': {'id': 'OLDAP-2.2.2', 'deprecated': False},
+    'oldap-2.3': {'id': 'OLDAP-2.3', 'deprecated': False},
+    'oldap-2.4': {'id': 'OLDAP-2.4', 'deprecated': False},
+    'oldap-2.5': {'id': 'OLDAP-2.5', 'deprecated': False},
+    'oldap-2.6': {'id': 'OLDAP-2.6', 'deprecated': False},
+    'oldap-2.7': {'id': 'OLDAP-2.7', 'deprecated': False},
+    'oldap-2.8': {'id': 'OLDAP-2.8', 'deprecated': False},
+    'olfl-1.3': {'id': 'OLFL-1.3', 'deprecated': False},
+    'oml': {'id': 'OML', 'deprecated': False},
+    'openpbs-2.3': {'id': 'OpenPBS-2.3', 'deprecated': False},
+    'openssl': {'id': 'OpenSSL', 'deprecated': False},
+    'openssl-standalone': {'id': 'OpenSSL-standalone', 'deprecated': False},
+    'openvision': {'id': 'OpenVision', 'deprecated': False},
+    'opl-1.0': {'id': 'OPL-1.0', 'deprecated': False},
+    'opl-uk-3.0': {'id': 'OPL-UK-3.0', 'deprecated': False},
+    'opubl-1.0': {'id': 'OPUBL-1.0', 'deprecated': False},
+    'oset-pl-2.1': {'id': 'OSET-PL-2.1', 'deprecated': False},
+    'osl-1.0': {'id': 'OSL-1.0', 'deprecated': False},
+    'osl-1.1': {'id': 'OSL-1.1', 'deprecated': False},
+    'osl-2.0': {'id': 'OSL-2.0', 'deprecated': False},
+    'osl-2.1': {'id': 'OSL-2.1', 'deprecated': False},
+    'osl-3.0': {'id': 'OSL-3.0', 'deprecated': False},
+    'padl': {'id': 'PADL', 'deprecated': False},
+    'parity-6.0.0': {'id': 'Parity-6.0.0', 'deprecated': False},
+    'parity-7.0.0': {'id': 'Parity-7.0.0', 'deprecated': False},
+    'pddl-1.0': {'id': 'PDDL-1.0', 'deprecated': False},
+    'php-3.0': {'id': 'PHP-3.0', 'deprecated': False},
+    'php-3.01': {'id': 'PHP-3.01', 'deprecated': False},
+    'pixar': {'id': 'Pixar', 'deprecated': False},
+    'pkgconf': {'id': 'pkgconf', 'deprecated': False},
+    'plexus': {'id': 'Plexus', 'deprecated': False},
+    'pnmstitch': {'id': 'pnmstitch', 'deprecated': False},
+    'polyform-noncommercial-1.0.0': {'id': 'PolyForm-Noncommercial-1.0.0', 'deprecated': False},
+    'polyform-small-business-1.0.0': {'id': 'PolyForm-Small-Business-1.0.0', 'deprecated': False},
+    'postgresql': {'id': 'PostgreSQL', 'deprecated': False},
+    'ppl': {'id': 'PPL', 'deprecated': False},
+    'psf-2.0': {'id': 'PSF-2.0', 'deprecated': False},
+    'psfrag': {'id': 'psfrag', 'deprecated': False},
+    'psutils': {'id': 'psutils', 'deprecated': False},
+    'python-2.0': {'id': 'Python-2.0', 'deprecated': False},
+    'python-2.0.1': {'id': 'Python-2.0.1', 'deprecated': False},
+    'python-ldap': {'id': 'python-ldap', 'deprecated': False},
+    'qhull': {'id': 'Qhull', 'deprecated': False},
+    'qpl-1.0': {'id': 'QPL-1.0', 'deprecated': False},
+    'qpl-1.0-inria-2004': {'id': 'QPL-1.0-INRIA-2004', 'deprecated': False},
+    'radvd': {'id': 'radvd', 'deprecated': False},
+    'rdisc': {'id': 'Rdisc', 'deprecated': False},
+    'rhecos-1.1': {'id': 'RHeCos-1.1', 'deprecated': False},
+    'rpl-1.1': {'id': 'RPL-1.1', 'deprecated': False},
+    'rpl-1.5': {'id': 'RPL-1.5', 'deprecated': False},
+    'rpsl-1.0': {'id': 'RPSL-1.0', 'deprecated': False},
+    'rsa-md': {'id': 'RSA-MD', 'deprecated': False},
+    'rscpl': {'id': 'RSCPL', 'deprecated': False},
+    'ruby': {'id': 'Ruby', 'deprecated': False},
+    'ruby-pty': {'id': 'Ruby-pty', 'deprecated': False},
+    'sax-pd': {'id': 'SAX-PD', 'deprecated': False},
+    'sax-pd-2.0': {'id': 'SAX-PD-2.0', 'deprecated': False},
+    'saxpath': {'id': 'Saxpath', 'deprecated': False},
+    'scea': {'id': 'SCEA', 'deprecated': False},
+    'schemereport': {'id': 'SchemeReport', 'deprecated': False},
+    'sendmail': {'id': 'Sendmail', 'deprecated': False},
+    'sendmail-8.23': {'id': 'Sendmail-8.23', 'deprecated': False},
+    'sgi-b-1.0': {'id': 'SGI-B-1.0', 'deprecated': False},
+    'sgi-b-1.1': {'id': 'SGI-B-1.1', 'deprecated': False},
+    'sgi-b-2.0': {'id': 'SGI-B-2.0', 'deprecated': False},
+    'sgi-opengl': {'id': 'SGI-OpenGL', 'deprecated': False},
+    'sgp4': {'id': 'SGP4', 'deprecated': False},
+    'shl-0.5': {'id': 'SHL-0.5', 'deprecated': False},
+    'shl-0.51': {'id': 'SHL-0.51', 'deprecated': False},
+    'simpl-2.0': {'id': 'SimPL-2.0', 'deprecated': False},
+    'sissl': {'id': 'SISSL', 'deprecated': False},
+    'sissl-1.2': {'id': 'SISSL-1.2', 'deprecated': False},
+    'sl': {'id': 'SL', 'deprecated': False},
+    'sleepycat': {'id': 'Sleepycat', 'deprecated': False},
+    'smlnj': {'id': 'SMLNJ', 'deprecated': False},
+    'smppl': {'id': 'SMPPL', 'deprecated': False},
+    'snia': {'id': 'SNIA', 'deprecated': False},
+    'snprintf': {'id': 'snprintf', 'deprecated': False},
+    'softsurfer': {'id': 'softSurfer', 'deprecated': False},
+    'soundex': {'id': 'Soundex', 'deprecated': False},
+    'spencer-86': {'id': 'Spencer-86', 'deprecated': False},
+    'spencer-94': {'id': 'Spencer-94', 'deprecated': False},
+    'spencer-99': {'id': 'Spencer-99', 'deprecated': False},
+    'spl-1.0': {'id': 'SPL-1.0', 'deprecated': False},
+    'ssh-keyscan': {'id': 'ssh-keyscan', 'deprecated': False},
+    'ssh-openssh': {'id': 'SSH-OpenSSH', 'deprecated': False},
+    'ssh-short': {'id': 'SSH-short', 'deprecated': False},
+    'ssleay-standalone': {'id': 'SSLeay-standalone', 'deprecated': False},
+    'sspl-1.0': {'id': 'SSPL-1.0', 'deprecated': False},
+    'standardml-nj': {'id': 'StandardML-NJ', 'deprecated': True},
+    'sugarcrm-1.1.3': {'id': 'SugarCRM-1.1.3', 'deprecated': False},
+    'sun-ppp': {'id': 'Sun-PPP', 'deprecated': False},
+    'sun-ppp-2000': {'id': 'Sun-PPP-2000', 'deprecated': False},
+    'sunpro': {'id': 'SunPro', 'deprecated': False},
+    'swl': {'id': 'SWL', 'deprecated': False},
+    'swrule': {'id': 'swrule', 'deprecated': False},
+    'symlinks': {'id': 'Symlinks', 'deprecated': False},
+    'tapr-ohl-1.0': {'id': 'TAPR-OHL-1.0', 'deprecated': False},
+    'tcl': {'id': 'TCL', 'deprecated': False},
+    'tcp-wrappers': {'id': 'TCP-wrappers', 'deprecated': False},
+    'termreadkey': {'id': 'TermReadKey', 'deprecated': False},
+    'tgppl-1.0': {'id': 'TGPPL-1.0', 'deprecated': False},
+    'threeparttable': {'id': 'threeparttable', 'deprecated': False},
+    'tmate': {'id': 'TMate', 'deprecated': False},
+    'torque-1.1': {'id': 'TORQUE-1.1', 'deprecated': False},
+    'tosl': {'id': 'TOSL', 'deprecated': False},
+    'tpdl': {'id': 'TPDL', 'deprecated': False},
+    'tpl-1.0': {'id': 'TPL-1.0', 'deprecated': False},
+    'ttwl': {'id': 'TTWL', 'deprecated': False},
+    'ttyp0': {'id': 'TTYP0', 'deprecated': False},
+    'tu-berlin-1.0': {'id': 'TU-Berlin-1.0', 'deprecated': False},
+    'tu-berlin-2.0': {'id': 'TU-Berlin-2.0', 'deprecated': False},
+    'ubuntu-font-1.0': {'id': 'Ubuntu-font-1.0', 'deprecated': False},
+    'ucar': {'id': 'UCAR', 'deprecated': False},
+    'ucl-1.0': {'id': 'UCL-1.0', 'deprecated': False},
+    'ulem': {'id': 'ulem', 'deprecated': False},
+    'umich-merit': {'id': 'UMich-Merit', 'deprecated': False},
+    'unicode-3.0': {'id': 'Unicode-3.0', 'deprecated': False},
+    'unicode-dfs-2015': {'id': 'Unicode-DFS-2015', 'deprecated': False},
+    'unicode-dfs-2016': {'id': 'Unicode-DFS-2016', 'deprecated': False},
+    'unicode-tou': {'id': 'Unicode-TOU', 'deprecated': False},
+    'unixcrypt': {'id': 'UnixCrypt', 'deprecated': False},
+    'unlicense': {'id': 'Unlicense', 'deprecated': False},
+    'upl-1.0': {'id': 'UPL-1.0', 'deprecated': False},
+    'urt-rle': {'id': 'URT-RLE', 'deprecated': False},
+    'vim': {'id': 'Vim', 'deprecated': False},
+    'vostrom': {'id': 'VOSTROM', 'deprecated': False},
+    'vsl-1.0': {'id': 'VSL-1.0', 'deprecated': False},
+    'w3c': {'id': 'W3C', 'deprecated': False},
+    'w3c-19980720': {'id': 'W3C-19980720', 'deprecated': False},
+    'w3c-20150513': {'id': 'W3C-20150513', 'deprecated': False},
+    'w3m': {'id': 'w3m', 'deprecated': False},
+    'watcom-1.0': {'id': 'Watcom-1.0', 'deprecated': False},
+    'widget-workshop': {'id': 'Widget-Workshop', 'deprecated': False},
+    'wsuipa': {'id': 'Wsuipa', 'deprecated': False},
+    'wtfpl': {'id': 'WTFPL', 'deprecated': False},
+    'wxwindows': {'id': 'wxWindows', 'deprecated': True},
+    'x11': {'id': 'X11', 'deprecated': False},
+    'x11-distribute-modifications-variant': {'id': 'X11-distribute-modifications-variant', 'deprecated': False},
+    'x11-swapped': {'id': 'X11-swapped', 'deprecated': False},
+    'xdebug-1.03': {'id': 'Xdebug-1.03', 'deprecated': False},
+    'xerox': {'id': 'Xerox', 'deprecated': False},
+    'xfig': {'id': 'Xfig', 'deprecated': False},
+    'xfree86-1.1': {'id': 'XFree86-1.1', 'deprecated': False},
+    'xinetd': {'id': 'xinetd', 'deprecated': False},
+    'xkeyboard-config-zinoviev': {'id': 'xkeyboard-config-Zinoviev', 'deprecated': False},
+    'xlock': {'id': 'xlock', 'deprecated': False},
+    'xnet': {'id': 'Xnet', 'deprecated': False},
+    'xpp': {'id': 'xpp', 'deprecated': False},
+    'xskat': {'id': 'XSkat', 'deprecated': False},
+    'xzoom': {'id': 'xzoom', 'deprecated': False},
+    'ypl-1.0': {'id': 'YPL-1.0', 'deprecated': False},
+    'ypl-1.1': {'id': 'YPL-1.1', 'deprecated': False},
+    'zed': {'id': 'Zed', 'deprecated': False},
+    'zeeff': {'id': 'Zeeff', 'deprecated': False},
+    'zend-2.0': {'id': 'Zend-2.0', 'deprecated': False},
+    'zimbra-1.3': {'id': 'Zimbra-1.3', 'deprecated': False},
+    'zimbra-1.4': {'id': 'Zimbra-1.4', 'deprecated': False},
+    'zlib': {'id': 'Zlib', 'deprecated': False},
+    'zlib-acknowledgement': {'id': 'zlib-acknowledgement', 'deprecated': False},
+    'zpl-1.1': {'id': 'ZPL-1.1', 'deprecated': False},
+    'zpl-2.0': {'id': 'ZPL-2.0', 'deprecated': False},
+    'zpl-2.1': {'id': 'ZPL-2.1', 'deprecated': False},
+}
+
+EXCEPTIONS: dict[str, SPDXException] = {
+    '389-exception': {'id': '389-exception', 'deprecated': False},
+    'asterisk-exception': {'id': 'Asterisk-exception', 'deprecated': False},
+    'asterisk-linking-protocols-exception': {'id': 'Asterisk-linking-protocols-exception', 'deprecated': False},
+    'autoconf-exception-2.0': {'id': 'Autoconf-exception-2.0', 'deprecated': False},
+    'autoconf-exception-3.0': {'id': 'Autoconf-exception-3.0', 'deprecated': False},
+    'autoconf-exception-generic': {'id': 'Autoconf-exception-generic', 'deprecated': False},
+    'autoconf-exception-generic-3.0': {'id': 'Autoconf-exception-generic-3.0', 'deprecated': False},
+    'autoconf-exception-macro': {'id': 'Autoconf-exception-macro', 'deprecated': False},
+    'bison-exception-1.24': {'id': 'Bison-exception-1.24', 'deprecated': False},
+    'bison-exception-2.2': {'id': 'Bison-exception-2.2', 'deprecated': False},
+    'bootloader-exception': {'id': 'Bootloader-exception', 'deprecated': False},
+    'classpath-exception-2.0': {'id': 'Classpath-exception-2.0', 'deprecated': False},
+    'clisp-exception-2.0': {'id': 'CLISP-exception-2.0', 'deprecated': False},
+    'cryptsetup-openssl-exception': {'id': 'cryptsetup-OpenSSL-exception', 'deprecated': False},
+    'digirule-foss-exception': {'id': 'DigiRule-FOSS-exception', 'deprecated': False},
+    'ecos-exception-2.0': {'id': 'eCos-exception-2.0', 'deprecated': False},
+    'erlang-otp-linking-exception': {'id': 'erlang-otp-linking-exception', 'deprecated': False},
+    'fawkes-runtime-exception': {'id': 'Fawkes-Runtime-exception', 'deprecated': False},
+    'fltk-exception': {'id': 'FLTK-exception', 'deprecated': False},
+    'fmt-exception': {'id': 'fmt-exception', 'deprecated': False},
+    'font-exception-2.0': {'id': 'Font-exception-2.0', 'deprecated': False},
+    'freertos-exception-2.0': {'id': 'freertos-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0': {'id': 'GCC-exception-2.0', 'deprecated': False},
+    'gcc-exception-2.0-note': {'id': 'GCC-exception-2.0-note', 'deprecated': False},
+    'gcc-exception-3.1': {'id': 'GCC-exception-3.1', 'deprecated': False},
+    'gmsh-exception': {'id': 'Gmsh-exception', 'deprecated': False},
+    'gnat-exception': {'id': 'GNAT-exception', 'deprecated': False},
+    'gnome-examples-exception': {'id': 'GNOME-examples-exception', 'deprecated': False},
+    'gnu-compiler-exception': {'id': 'GNU-compiler-exception', 'deprecated': False},
+    'gnu-javamail-exception': {'id': 'gnu-javamail-exception', 'deprecated': False},
+    'gpl-3.0-interface-exception': {'id': 'GPL-3.0-interface-exception', 'deprecated': False},
+    'gpl-3.0-linking-exception': {'id': 'GPL-3.0-linking-exception', 'deprecated': False},
+    'gpl-3.0-linking-source-exception': {'id': 'GPL-3.0-linking-source-exception', 'deprecated': False},
+    'gpl-cc-1.0': {'id': 'GPL-CC-1.0', 'deprecated': False},
+    'gstreamer-exception-2005': {'id': 'GStreamer-exception-2005', 'deprecated': False},
+    'gstreamer-exception-2008': {'id': 'GStreamer-exception-2008', 'deprecated': False},
+    'i2p-gpl-java-exception': {'id': 'i2p-gpl-java-exception', 'deprecated': False},
+    'kicad-libraries-exception': {'id': 'KiCad-libraries-exception', 'deprecated': False},
+    'lgpl-3.0-linking-exception': {'id': 'LGPL-3.0-linking-exception', 'deprecated': False},
+    'libpri-openh323-exception': {'id': 'libpri-OpenH323-exception', 'deprecated': False},
+    'libtool-exception': {'id': 'Libtool-exception', 'deprecated': False},
+    'linux-syscall-note': {'id': 'Linux-syscall-note', 'deprecated': False},
+    'llgpl': {'id': 'LLGPL', 'deprecated': False},
+    'llvm-exception': {'id': 'LLVM-exception', 'deprecated': False},
+    'lzma-exception': {'id': 'LZMA-exception', 'deprecated': False},
+    'mif-exception': {'id': 'mif-exception', 'deprecated': False},
+    'nokia-qt-exception-1.1': {'id': 'Nokia-Qt-exception-1.1', 'deprecated': True},
+    'ocaml-lgpl-linking-exception': {'id': 'OCaml-LGPL-linking-exception', 'deprecated': False},
+    'occt-exception-1.0': {'id': 'OCCT-exception-1.0', 'deprecated': False},
+    'openjdk-assembly-exception-1.0': {'id': 'OpenJDK-assembly-exception-1.0', 'deprecated': False},
+    'openvpn-openssl-exception': {'id': 'openvpn-openssl-exception', 'deprecated': False},
+    'pcre2-exception': {'id': 'PCRE2-exception', 'deprecated': False},
+    'ps-or-pdf-font-exception-20170817': {'id': 'PS-or-PDF-font-exception-20170817', 'deprecated': False},
+    'qpl-1.0-inria-2004-exception': {'id': 'QPL-1.0-INRIA-2004-exception', 'deprecated': False},
+    'qt-gpl-exception-1.0': {'id': 'Qt-GPL-exception-1.0', 'deprecated': False},
+    'qt-lgpl-exception-1.1': {'id': 'Qt-LGPL-exception-1.1', 'deprecated': False},
+    'qwt-exception-1.0': {'id': 'Qwt-exception-1.0', 'deprecated': False},
+    'romic-exception': {'id': 'romic-exception', 'deprecated': False},
+    'rrdtool-floss-exception-2.0': {'id': 'RRDtool-FLOSS-exception-2.0', 'deprecated': False},
+    'sane-exception': {'id': 'SANE-exception', 'deprecated': False},
+    'shl-2.0': {'id': 'SHL-2.0', 'deprecated': False},
+    'shl-2.1': {'id': 'SHL-2.1', 'deprecated': False},
+    'stunnel-exception': {'id': 'stunnel-exception', 'deprecated': False},
+    'swi-exception': {'id': 'SWI-exception', 'deprecated': False},
+    'swift-exception': {'id': 'Swift-exception', 'deprecated': False},
+    'texinfo-exception': {'id': 'Texinfo-exception', 'deprecated': False},
+    'u-boot-exception-2.0': {'id': 'u-boot-exception-2.0', 'deprecated': False},
+    'ubdl-exception': {'id': 'UBDL-exception', 'deprecated': False},
+    'universal-foss-exception-1.0': {'id': 'Universal-FOSS-exception-1.0', 'deprecated': False},
+    'vsftpd-openssl-exception': {'id': 'vsftpd-openssl-exception', 'deprecated': False},
+    'wxwindows-exception-3.1': {'id': 'WxWindows-exception-3.1', 'deprecated': False},
+    'x11vnc-openssl-exception': {'id': 'x11vnc-openssl-exception', 'deprecated': False},
+}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/markers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/markers.py
new file mode 100644
index 00000000..fb7f49cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/markers.py
@@ -0,0 +1,331 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import operator
+import os
+import platform
+import sys
+from typing import Any, Callable, TypedDict, cast
+
+from ._parser import MarkerAtom, MarkerList, Op, Value, Variable
+from ._parser import parse_marker as _parse_marker
+from ._tokenizer import ParserSyntaxError
+from .specifiers import InvalidSpecifier, Specifier
+from .utils import canonicalize_name
+
+__all__ = [
+    "InvalidMarker",
+    "Marker",
+    "UndefinedComparison",
+    "UndefinedEnvironmentName",
+    "default_environment",
+]
+
+Operator = Callable[[str, str], bool]
+
+
+class InvalidMarker(ValueError):
+    """
+    An invalid marker was found, users should refer to PEP 508.
+    """
+
+
+class UndefinedComparison(ValueError):
+    """
+    An invalid operation was attempted on a value that doesn't support it.
+    """
+
+
+class UndefinedEnvironmentName(ValueError):
+    """
+    A name was attempted to be used that does not exist inside of the
+    environment.
+    """
+
+
+class Environment(TypedDict):
+    implementation_name: str
+    """The implementation's identifier, e.g. ``'cpython'``."""
+
+    implementation_version: str
+    """
+    The implementation's version, e.g. ``'3.13.0a2'`` for CPython 3.13.0a2, or
+    ``'7.3.13'`` for PyPy3.10 v7.3.13.
+    """
+
+    os_name: str
+    """
+    The value of :py:data:`os.name`. The name of the operating system dependent module
+    imported, e.g. ``'posix'``.
+    """
+
+    platform_machine: str
+    """
+    Returns the machine type, e.g. ``'i386'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_release: str
+    """
+    The system's release, e.g. ``'2.2.0'`` or ``'NT'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_system: str
+    """
+    The system/OS name, e.g. ``'Linux'``, ``'Windows'`` or ``'Java'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    platform_version: str
+    """
+    The system's release version, e.g. ``'#3 on degas'``.
+
+    An empty string if the value cannot be determined.
+    """
+
+    python_full_version: str
+    """
+    The Python version as string ``'major.minor.patchlevel'``.
+
+    Note that unlike the Python :py:data:`sys.version`, this value will always include
+    the patchlevel (it defaults to 0).
+    """
+
+    platform_python_implementation: str
+    """
+    A string identifying the Python implementation, e.g. ``'CPython'``.
+    """
+
+    python_version: str
+    """The Python version as string ``'major.minor'``."""
+
+    sys_platform: str
+    """
+    This string contains a platform identifier that can be used to append
+    platform-specific components to :py:data:`sys.path`, for instance.
+
+    For Unix systems, except on Linux and AIX, this is the lowercased OS name as
+    returned by ``uname -s`` with the first part of the version as returned by
+    ``uname -r`` appended, e.g. ``'sunos5'`` or ``'freebsd8'``, at the time when Python
+    was built.
+    """
+
+
+def _normalize_extra_values(results: Any) -> Any:
+    """
+    Normalize extra values.
+    """
+    if isinstance(results[0], tuple):
+        lhs, op, rhs = results[0]
+        if isinstance(lhs, Variable) and lhs.value == "extra":
+            normalized_extra = canonicalize_name(rhs.value)
+            rhs = Value(normalized_extra)
+        elif isinstance(rhs, Variable) and rhs.value == "extra":
+            normalized_extra = canonicalize_name(lhs.value)
+            lhs = Value(normalized_extra)
+        results[0] = lhs, op, rhs
+    return results
+
+
+def _format_marker(
+    marker: list[str] | MarkerAtom | str, first: bool | None = True
+) -> str:
+    assert isinstance(marker, (list, tuple, str))
+
+    # Sometimes we have a structure like [[...]] which is a single item list
+    # where the single item is itself it's own list. In that case we want skip
+    # the rest of this function so that we don't get extraneous () on the
+    # outside.
+    if (
+        isinstance(marker, list)
+        and len(marker) == 1
+        and isinstance(marker[0], (list, tuple))
+    ):
+        return _format_marker(marker[0])
+
+    if isinstance(marker, list):
+        inner = (_format_marker(m, first=False) for m in marker)
+        if first:
+            return " ".join(inner)
+        else:
+            return "(" + " ".join(inner) + ")"
+    elif isinstance(marker, tuple):
+        return " ".join([m.serialize() for m in marker])
+    else:
+        return marker
+
+
+_operators: dict[str, Operator] = {
+    "in": lambda lhs, rhs: lhs in rhs,
+    "not in": lambda lhs, rhs: lhs not in rhs,
+    "<": operator.lt,
+    "<=": operator.le,
+    "==": operator.eq,
+    "!=": operator.ne,
+    ">=": operator.ge,
+    ">": operator.gt,
+}
+
+
+def _eval_op(lhs: str, op: Op, rhs: str) -> bool:
+    try:
+        spec = Specifier("".join([op.serialize(), rhs]))
+    except InvalidSpecifier:
+        pass
+    else:
+        return spec.contains(lhs, prereleases=True)
+
+    oper: Operator | None = _operators.get(op.serialize())
+    if oper is None:
+        raise UndefinedComparison(f"Undefined {op!r} on {lhs!r} and {rhs!r}.")
+
+    return oper(lhs, rhs)
+
+
+def _normalize(*values: str, key: str) -> tuple[str, ...]:
+    # PEP 685 – Comparison of extra names for optional distribution dependencies
+    # https://peps.python.org/pep-0685/
+    # > When comparing extra names, tools MUST normalize the names being
+    # > compared using the semantics outlined in PEP 503 for names
+    if key == "extra":
+        return tuple(canonicalize_name(v) for v in values)
+
+    # other environment markers don't have such standards
+    return values
+
+
+def _evaluate_markers(markers: MarkerList, environment: dict[str, str]) -> bool:
+    groups: list[list[bool]] = [[]]
+
+    for marker in markers:
+        assert isinstance(marker, (list, tuple, str))
+
+        if isinstance(marker, list):
+            groups[-1].append(_evaluate_markers(marker, environment))
+        elif isinstance(marker, tuple):
+            lhs, op, rhs = marker
+
+            if isinstance(lhs, Variable):
+                environment_key = lhs.value
+                lhs_value = environment[environment_key]
+                rhs_value = rhs.value
+            else:
+                lhs_value = lhs.value
+                environment_key = rhs.value
+                rhs_value = environment[environment_key]
+
+            lhs_value, rhs_value = _normalize(lhs_value, rhs_value, key=environment_key)
+            groups[-1].append(_eval_op(lhs_value, op, rhs_value))
+        else:
+            assert marker in ["and", "or"]
+            if marker == "or":
+                groups.append([])
+
+    return any(all(item) for item in groups)
+
+
+def format_full_version(info: sys._version_info) -> str:
+    version = f"{info.major}.{info.minor}.{info.micro}"
+    kind = info.releaselevel
+    if kind != "final":
+        version += kind[0] + str(info.serial)
+    return version
+
+
+def default_environment() -> Environment:
+    iver = format_full_version(sys.implementation.version)
+    implementation_name = sys.implementation.name
+    return {
+        "implementation_name": implementation_name,
+        "implementation_version": iver,
+        "os_name": os.name,
+        "platform_machine": platform.machine(),
+        "platform_release": platform.release(),
+        "platform_system": platform.system(),
+        "platform_version": platform.version(),
+        "python_full_version": platform.python_version(),
+        "platform_python_implementation": platform.python_implementation(),
+        "python_version": ".".join(platform.python_version_tuple()[:2]),
+        "sys_platform": sys.platform,
+    }
+
+
+class Marker:
+    def __init__(self, marker: str) -> None:
+        # Note: We create a Marker object without calling this constructor in
+        #       packaging.requirements.Requirement. If any additional logic is
+        #       added here, make sure to mirror/adapt Requirement.
+        try:
+            self._markers = _normalize_extra_values(_parse_marker(marker))
+            # The attribute `_markers` can be described in terms of a recursive type:
+            # MarkerList = List[Union[Tuple[Node, ...], str, MarkerList]]
+            #
+            # For example, the following expression:
+            # python_version > "3.6" or (python_version == "3.6" and os_name == "unix")
+            #
+            # is parsed into:
+            # [
+            #     (<Variable('python_version')>, <Op('>')>, <Value('3.6')>),
+            #     'and',
+            #     [
+            #         (<Variable('python_version')>, <Op('==')>, <Value('3.6')>),
+            #         'or',
+            #         (<Variable('os_name')>, <Op('==')>, <Value('unix')>)
+            #     ]
+            # ]
+        except ParserSyntaxError as e:
+            raise InvalidMarker(str(e)) from e
+
+    def __str__(self) -> str:
+        return _format_marker(self._markers)
+
+    def __repr__(self) -> str:
+        return f"<Marker('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash((self.__class__.__name__, str(self)))
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Marker):
+            return NotImplemented
+
+        return str(self) == str(other)
+
+    def evaluate(self, environment: dict[str, str] | None = None) -> bool:
+        """Evaluate a marker.
+
+        Return the boolean from evaluating the given marker against the
+        environment. environment is an optional argument to override all or
+        part of the determined environment.
+
+        The environment is determined from the current Python process.
+        """
+        current_environment = cast("dict[str, str]", default_environment())
+        current_environment["extra"] = ""
+        if environment is not None:
+            current_environment.update(environment)
+            # The API used to allow setting extra to None. We need to handle this
+            # case for backwards compatibility.
+            if current_environment["extra"] is None:
+                current_environment["extra"] = ""
+
+        return _evaluate_markers(
+            self._markers, _repair_python_full_version(current_environment)
+        )
+
+
+def _repair_python_full_version(env: dict[str, str]) -> dict[str, str]:
+    """
+    Work around platform.python_version() returning something that is not PEP 440
+    compliant for non-tagged Python builds.
+    """
+    if env["python_full_version"].endswith("+"):
+        env["python_full_version"] += "local"
+    return env
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/metadata.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/metadata.py
new file mode 100644
index 00000000..721f411c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/metadata.py
@@ -0,0 +1,863 @@
+from __future__ import annotations
+
+import email.feedparser
+import email.header
+import email.message
+import email.parser
+import email.policy
+import pathlib
+import sys
+import typing
+from typing import (
+    Any,
+    Callable,
+    Generic,
+    Literal,
+    TypedDict,
+    cast,
+)
+
+from . import licenses, requirements, specifiers, utils
+from . import version as version_module
+from .licenses import NormalizedLicenseExpression
+
+T = typing.TypeVar("T")
+
+
+if sys.version_info >= (3, 11):  # pragma: no cover
+    ExceptionGroup = ExceptionGroup
+else:  # pragma: no cover
+
+    class ExceptionGroup(Exception):
+        """A minimal implementation of :external:exc:`ExceptionGroup` from Python 3.11.
+
+        If :external:exc:`ExceptionGroup` is already defined by Python itself,
+        that version is used instead.
+        """
+
+        message: str
+        exceptions: list[Exception]
+
+        def __init__(self, message: str, exceptions: list[Exception]) -> None:
+            self.message = message
+            self.exceptions = exceptions
+
+        def __repr__(self) -> str:
+            return f"{self.__class__.__name__}({self.message!r}, {self.exceptions!r})"
+
+
+class InvalidMetadata(ValueError):
+    """A metadata field contains invalid data."""
+
+    field: str
+    """The name of the field that contains invalid data."""
+
+    def __init__(self, field: str, message: str) -> None:
+        self.field = field
+        super().__init__(message)
+
+
+# The RawMetadata class attempts to make as few assumptions about the underlying
+# serialization formats as possible. The idea is that as long as a serialization
+# formats offer some very basic primitives in *some* way then we can support
+# serializing to and from that format.
+class RawMetadata(TypedDict, total=False):
+    """A dictionary of raw core metadata.
+
+    Each field in core metadata maps to a key of this dictionary (when data is
+    provided). The key is lower-case and underscores are used instead of dashes
+    compared to the equivalent core metadata field. Any core metadata field that
+    can be specified multiple times or can hold multiple values in a single
+    field have a key with a plural name. See :class:`Metadata` whose attributes
+    match the keys of this dictionary.
+
+    Core metadata fields that can be specified multiple times are stored as a
+    list or dict depending on which is appropriate for the field. Any fields
+    which hold multiple values in a single field are stored as a list.
+
+    """
+
+    # Metadata 1.0 - PEP 241
+    metadata_version: str
+    name: str
+    version: str
+    platforms: list[str]
+    summary: str
+    description: str
+    keywords: list[str]
+    home_page: str
+    author: str
+    author_email: str
+    license: str
+
+    # Metadata 1.1 - PEP 314
+    supported_platforms: list[str]
+    download_url: str
+    classifiers: list[str]
+    requires: list[str]
+    provides: list[str]
+    obsoletes: list[str]
+
+    # Metadata 1.2 - PEP 345
+    maintainer: str
+    maintainer_email: str
+    requires_dist: list[str]
+    provides_dist: list[str]
+    obsoletes_dist: list[str]
+    requires_python: str
+    requires_external: list[str]
+    project_urls: dict[str, str]
+
+    # Metadata 2.0
+    # PEP 426 attempted to completely revamp the metadata format
+    # but got stuck without ever being able to build consensus on
+    # it and ultimately ended up withdrawn.
+    #
+    # However, a number of tools had started emitting METADATA with
+    # `2.0` Metadata-Version, so for historical reasons, this version
+    # was skipped.
+
+    # Metadata 2.1 - PEP 566
+    description_content_type: str
+    provides_extra: list[str]
+
+    # Metadata 2.2 - PEP 643
+    dynamic: list[str]
+
+    # Metadata 2.3 - PEP 685
+    # No new fields were added in PEP 685, just some edge case were
+    # tightened up to provide better interoptability.
+
+    # Metadata 2.4 - PEP 639
+    license_expression: str
+    license_files: list[str]
+
+
+_STRING_FIELDS = {
+    "author",
+    "author_email",
+    "description",
+    "description_content_type",
+    "download_url",
+    "home_page",
+    "license",
+    "license_expression",
+    "maintainer",
+    "maintainer_email",
+    "metadata_version",
+    "name",
+    "requires_python",
+    "summary",
+    "version",
+}
+
+_LIST_FIELDS = {
+    "classifiers",
+    "dynamic",
+    "license_files",
+    "obsoletes",
+    "obsoletes_dist",
+    "platforms",
+    "provides",
+    "provides_dist",
+    "provides_extra",
+    "requires",
+    "requires_dist",
+    "requires_external",
+    "supported_platforms",
+}
+
+_DICT_FIELDS = {
+    "project_urls",
+}
+
+
+def _parse_keywords(data: str) -> list[str]:
+    """Split a string of comma-separated keywords into a list of keywords."""
+    return [k.strip() for k in data.split(",")]
+
+
+def _parse_project_urls(data: list[str]) -> dict[str, str]:
+    """Parse a list of label/URL string pairings separated by a comma."""
+    urls = {}
+    for pair in data:
+        # Our logic is slightly tricky here as we want to try and do
+        # *something* reasonable with malformed data.
+        #
+        # The main thing that we have to worry about, is data that does
+        # not have a ',' at all to split the label from the Value. There
+        # isn't a singular right answer here, and we will fail validation
+        # later on (if the caller is validating) so it doesn't *really*
+        # matter, but since the missing value has to be an empty str
+        # and our return value is dict[str, str], if we let the key
+        # be the missing value, then they'd have multiple '' values that
+        # overwrite each other in a accumulating dict.
+        #
+        # The other potentional issue is that it's possible to have the
+        # same label multiple times in the metadata, with no solid "right"
+        # answer with what to do in that case. As such, we'll do the only
+        # thing we can, which is treat the field as unparseable and add it
+        # to our list of unparsed fields.
+        parts = [p.strip() for p in pair.split(",", 1)]
+        parts.extend([""] * (max(0, 2 - len(parts))))  # Ensure 2 items
+
+        # TODO: The spec doesn't say anything about if the keys should be
+        #       considered case sensitive or not... logically they should
+        #       be case-preserving and case-insensitive, but doing that
+        #       would open up more cases where we might have duplicate
+        #       entries.
+        label, url = parts
+        if label in urls:
+            # The label already exists in our set of urls, so this field
+            # is unparseable, and we can just add the whole thing to our
+            # unparseable data and stop processing it.
+            raise KeyError("duplicate labels in project urls")
+        urls[label] = url
+
+    return urls
+
+
+def _get_payload(msg: email.message.Message, source: bytes | str) -> str:
+    """Get the body of the message."""
+    # If our source is a str, then our caller has managed encodings for us,
+    # and we don't need to deal with it.
+    if isinstance(source, str):
+        payload = msg.get_payload()
+        assert isinstance(payload, str)
+        return payload
+    # If our source is a bytes, then we're managing the encoding and we need
+    # to deal with it.
+    else:
+        bpayload = msg.get_payload(decode=True)
+        assert isinstance(bpayload, bytes)
+        try:
+            return bpayload.decode("utf8", "strict")
+        except UnicodeDecodeError as exc:
+            raise ValueError("payload in an invalid encoding") from exc
+
+
+# The various parse_FORMAT functions here are intended to be as lenient as
+# possible in their parsing, while still returning a correctly typed
+# RawMetadata.
+#
+# To aid in this, we also generally want to do as little touching of the
+# data as possible, except where there are possibly some historic holdovers
+# that make valid data awkward to work with.
+#
+# While this is a lower level, intermediate format than our ``Metadata``
+# class, some light touch ups can make a massive difference in usability.
+
+# Map METADATA fields to RawMetadata.
+_EMAIL_TO_RAW_MAPPING = {
+    "author": "author",
+    "author-email": "author_email",
+    "classifier": "classifiers",
+    "description": "description",
+    "description-content-type": "description_content_type",
+    "download-url": "download_url",
+    "dynamic": "dynamic",
+    "home-page": "home_page",
+    "keywords": "keywords",
+    "license": "license",
+    "license-expression": "license_expression",
+    "license-file": "license_files",
+    "maintainer": "maintainer",
+    "maintainer-email": "maintainer_email",
+    "metadata-version": "metadata_version",
+    "name": "name",
+    "obsoletes": "obsoletes",
+    "obsoletes-dist": "obsoletes_dist",
+    "platform": "platforms",
+    "project-url": "project_urls",
+    "provides": "provides",
+    "provides-dist": "provides_dist",
+    "provides-extra": "provides_extra",
+    "requires": "requires",
+    "requires-dist": "requires_dist",
+    "requires-external": "requires_external",
+    "requires-python": "requires_python",
+    "summary": "summary",
+    "supported-platform": "supported_platforms",
+    "version": "version",
+}
+_RAW_TO_EMAIL_MAPPING = {raw: email for email, raw in _EMAIL_TO_RAW_MAPPING.items()}
+
+
+def parse_email(data: bytes | str) -> tuple[RawMetadata, dict[str, list[str]]]:
+    """Parse a distribution's metadata stored as email headers (e.g. from ``METADATA``).
+
+    This function returns a two-item tuple of dicts. The first dict is of
+    recognized fields from the core metadata specification. Fields that can be
+    parsed and translated into Python's built-in types are converted
+    appropriately. All other fields are left as-is. Fields that are allowed to
+    appear multiple times are stored as lists.
+
+    The second dict contains all other fields from the metadata. This includes
+    any unrecognized fields. It also includes any fields which are expected to
+    be parsed into a built-in type but were not formatted appropriately. Finally,
+    any fields that are expected to appear only once but are repeated are
+    included in this dict.
+
+    """
+    raw: dict[str, str | list[str] | dict[str, str]] = {}
+    unparsed: dict[str, list[str]] = {}
+
+    if isinstance(data, str):
+        parsed = email.parser.Parser(policy=email.policy.compat32).parsestr(data)
+    else:
+        parsed = email.parser.BytesParser(policy=email.policy.compat32).parsebytes(data)
+
+    # We have to wrap parsed.keys() in a set, because in the case of multiple
+    # values for a key (a list), the key will appear multiple times in the
+    # list of keys, but we're avoiding that by using get_all().
+    for name in frozenset(parsed.keys()):
+        # Header names in RFC are case insensitive, so we'll normalize to all
+        # lower case to make comparisons easier.
+        name = name.lower()
+
+        # We use get_all() here, even for fields that aren't multiple use,
+        # because otherwise someone could have e.g. two Name fields, and we
+        # would just silently ignore it rather than doing something about it.
+        headers = parsed.get_all(name) or []
+
+        # The way the email module works when parsing bytes is that it
+        # unconditionally decodes the bytes as ascii using the surrogateescape
+        # handler. When you pull that data back out (such as with get_all() ),
+        # it looks to see if the str has any surrogate escapes, and if it does
+        # it wraps it in a Header object instead of returning the string.
+        #
+        # As such, we'll look for those Header objects, and fix up the encoding.
+        value = []
+        # Flag if we have run into any issues processing the headers, thus
+        # signalling that the data belongs in 'unparsed'.
+        valid_encoding = True
+        for h in headers:
+            # It's unclear if this can return more types than just a Header or
+            # a str, so we'll just assert here to make sure.
+            assert isinstance(h, (email.header.Header, str))
+
+            # If it's a header object, we need to do our little dance to get
+            # the real data out of it. In cases where there is invalid data
+            # we're going to end up with mojibake, but there's no obvious, good
+            # way around that without reimplementing parts of the Header object
+            # ourselves.
+            #
+            # That should be fine since, if mojibacked happens, this key is
+            # going into the unparsed dict anyways.
+            if isinstance(h, email.header.Header):
+                # The Header object stores it's data as chunks, and each chunk
+                # can be independently encoded, so we'll need to check each
+                # of them.
+                chunks: list[tuple[bytes, str | None]] = []
+                for bin, encoding in email.header.decode_header(h):
+                    try:
+                        bin.decode("utf8", "strict")
+                    except UnicodeDecodeError:
+                        # Enable mojibake.
+                        encoding = "latin1"
+                        valid_encoding = False
+                    else:
+                        encoding = "utf8"
+                    chunks.append((bin, encoding))
+
+                # Turn our chunks back into a Header object, then let that
+                # Header object do the right thing to turn them into a
+                # string for us.
+                value.append(str(email.header.make_header(chunks)))
+            # This is already a string, so just add it.
+            else:
+                value.append(h)
+
+        # We've processed all of our values to get them into a list of str,
+        # but we may have mojibake data, in which case this is an unparsed
+        # field.
+        if not valid_encoding:
+            unparsed[name] = value
+            continue
+
+        raw_name = _EMAIL_TO_RAW_MAPPING.get(name)
+        if raw_name is None:
+            # This is a bit of a weird situation, we've encountered a key that
+            # we don't know what it means, so we don't know whether it's meant
+            # to be a list or not.
+            #
+            # Since we can't really tell one way or another, we'll just leave it
+            # as a list, even though it may be a single item list, because that's
+            # what makes the most sense for email headers.
+            unparsed[name] = value
+            continue
+
+        # If this is one of our string fields, then we'll check to see if our
+        # value is a list of a single item. If it is then we'll assume that
+        # it was emitted as a single string, and unwrap the str from inside
+        # the list.
+        #
+        # If it's any other kind of data, then we haven't the faintest clue
+        # what we should parse it as, and we have to just add it to our list
+        # of unparsed stuff.
+        if raw_name in _STRING_FIELDS and len(value) == 1:
+            raw[raw_name] = value[0]
+        # If this is one of our list of string fields, then we can just assign
+        # the value, since email *only* has strings, and our get_all() call
+        # above ensures that this is a list.
+        elif raw_name in _LIST_FIELDS:
+            raw[raw_name] = value
+        # Special Case: Keywords
+        # The keywords field is implemented in the metadata spec as a str,
+        # but it conceptually is a list of strings, and is serialized using
+        # ", ".join(keywords), so we'll do some light data massaging to turn
+        # this into what it logically is.
+        elif raw_name == "keywords" and len(value) == 1:
+            raw[raw_name] = _parse_keywords(value[0])
+        # Special Case: Project-URL
+        # The project urls is implemented in the metadata spec as a list of
+        # specially-formatted strings that represent a key and a value, which
+        # is fundamentally a mapping, however the email format doesn't support
+        # mappings in a sane way, so it was crammed into a list of strings
+        # instead.
+        #
+        # We will do a little light data massaging to turn this into a map as
+        # it logically should be.
+        elif raw_name == "project_urls":
+            try:
+                raw[raw_name] = _parse_project_urls(value)
+            except KeyError:
+                unparsed[name] = value
+        # Nothing that we've done has managed to parse this, so it'll just
+        # throw it in our unparseable data and move on.
+        else:
+            unparsed[name] = value
+
+    # We need to support getting the Description from the message payload in
+    # addition to getting it from the the headers. This does mean, though, there
+    # is the possibility of it being set both ways, in which case we put both
+    # in 'unparsed' since we don't know which is right.
+    try:
+        payload = _get_payload(parsed, data)
+    except ValueError:
+        unparsed.setdefault("description", []).append(
+            parsed.get_payload(decode=isinstance(data, bytes))  # type: ignore[call-overload]
+        )
+    else:
+        if payload:
+            # Check to see if we've already got a description, if so then both
+            # it, and this body move to unparseable.
+            if "description" in raw:
+                description_header = cast(str, raw.pop("description"))
+                unparsed.setdefault("description", []).extend(
+                    [description_header, payload]
+                )
+            elif "description" in unparsed:
+                unparsed["description"].append(payload)
+            else:
+                raw["description"] = payload
+
+    # We need to cast our `raw` to a metadata, because a TypedDict only support
+    # literal key names, but we're computing our key names on purpose, but the
+    # way this function is implemented, our `TypedDict` can only have valid key
+    # names.
+    return cast(RawMetadata, raw), unparsed
+
+
+_NOT_FOUND = object()
+
+
+# Keep the two values in sync.
+_VALID_METADATA_VERSIONS = ["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+_MetadataVersion = Literal["1.0", "1.1", "1.2", "2.1", "2.2", "2.3", "2.4"]
+
+_REQUIRED_ATTRS = frozenset(["metadata_version", "name", "version"])
+
+
+class _Validator(Generic[T]):
+    """Validate a metadata field.
+
+    All _process_*() methods correspond to a core metadata field. The method is
+    called with the field's raw value. If the raw value is valid it is returned
+    in its "enriched" form (e.g. ``version.Version`` for the ``Version`` field).
+    If the raw value is invalid, :exc:`InvalidMetadata` is raised (with a cause
+    as appropriate).
+    """
+
+    name: str
+    raw_name: str
+    added: _MetadataVersion
+
+    def __init__(
+        self,
+        *,
+        added: _MetadataVersion = "1.0",
+    ) -> None:
+        self.added = added
+
+    def __set_name__(self, _owner: Metadata, name: str) -> None:
+        self.name = name
+        self.raw_name = _RAW_TO_EMAIL_MAPPING[name]
+
+    def __get__(self, instance: Metadata, _owner: type[Metadata]) -> T:
+        # With Python 3.8, the caching can be replaced with functools.cached_property().
+        # No need to check the cache as attribute lookup will resolve into the
+        # instance's __dict__ before __get__ is called.
+        cache = instance.__dict__
+        value = instance._raw.get(self.name)
+
+        # To make the _process_* methods easier, we'll check if the value is None
+        # and if this field is NOT a required attribute, and if both of those
+        # things are true, we'll skip the the converter. This will mean that the
+        # converters never have to deal with the None union.
+        if self.name in _REQUIRED_ATTRS or value is not None:
+            try:
+                converter: Callable[[Any], T] = getattr(self, f"_process_{self.name}")
+            except AttributeError:
+                pass
+            else:
+                value = converter(value)
+
+        cache[self.name] = value
+        try:
+            del instance._raw[self.name]  # type: ignore[misc]
+        except KeyError:
+            pass
+
+        return cast(T, value)
+
+    def _invalid_metadata(
+        self, msg: str, cause: Exception | None = None
+    ) -> InvalidMetadata:
+        exc = InvalidMetadata(
+            self.raw_name, msg.format_map({"field": repr(self.raw_name)})
+        )
+        exc.__cause__ = cause
+        return exc
+
+    def _process_metadata_version(self, value: str) -> _MetadataVersion:
+        # Implicitly makes Metadata-Version required.
+        if value not in _VALID_METADATA_VERSIONS:
+            raise self._invalid_metadata(f"{value!r} is not a valid metadata version")
+        return cast(_MetadataVersion, value)
+
+    def _process_name(self, value: str) -> str:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        # Validate the name as a side-effect.
+        try:
+            utils.canonicalize_name(value, validate=True)
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return value
+
+    def _process_version(self, value: str) -> version_module.Version:
+        if not value:
+            raise self._invalid_metadata("{field} is a required field")
+        try:
+            return version_module.parse(value)
+        except version_module.InvalidVersion as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_summary(self, value: str) -> str:
+        """Check the field contains no newlines."""
+        if "\n" in value:
+            raise self._invalid_metadata("{field} must be a single line")
+        return value
+
+    def _process_description_content_type(self, value: str) -> str:
+        content_types = {"text/plain", "text/x-rst", "text/markdown"}
+        message = email.message.EmailMessage()
+        message["content-type"] = value
+
+        content_type, parameters = (
+            # Defaults to `text/plain` if parsing failed.
+            message.get_content_type().lower(),
+            message["content-type"].params,
+        )
+        # Check if content-type is valid or defaulted to `text/plain` and thus was
+        # not parseable.
+        if content_type not in content_types or content_type not in value.lower():
+            raise self._invalid_metadata(
+                f"{{field}} must be one of {list(content_types)}, not {value!r}"
+            )
+
+        charset = parameters.get("charset", "UTF-8")
+        if charset != "UTF-8":
+            raise self._invalid_metadata(
+                f"{{field}} can only specify the UTF-8 charset, not {list(charset)}"
+            )
+
+        markdown_variants = {"GFM", "CommonMark"}
+        variant = parameters.get("variant", "GFM")  # Use an acceptable default.
+        if content_type == "text/markdown" and variant not in markdown_variants:
+            raise self._invalid_metadata(
+                f"valid Markdown variants for {{field}} are {list(markdown_variants)}, "
+                f"not {variant!r}",
+            )
+        return value
+
+    def _process_dynamic(self, value: list[str]) -> list[str]:
+        for dynamic_field in map(str.lower, value):
+            if dynamic_field in {"name", "version", "metadata-version"}:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not allowed as a dynamic field"
+                )
+            elif dynamic_field not in _EMAIL_TO_RAW_MAPPING:
+                raise self._invalid_metadata(
+                    f"{dynamic_field!r} is not a valid dynamic field"
+                )
+        return list(map(str.lower, value))
+
+    def _process_provides_extra(
+        self,
+        value: list[str],
+    ) -> list[utils.NormalizedName]:
+        normalized_names = []
+        try:
+            for name in value:
+                normalized_names.append(utils.canonicalize_name(name, validate=True))
+        except utils.InvalidName as exc:
+            raise self._invalid_metadata(
+                f"{name!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return normalized_names
+
+    def _process_requires_python(self, value: str) -> specifiers.SpecifierSet:
+        try:
+            return specifiers.SpecifierSet(value)
+        except specifiers.InvalidSpecifier as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_requires_dist(
+        self,
+        value: list[str],
+    ) -> list[requirements.Requirement]:
+        reqs = []
+        try:
+            for req in value:
+                reqs.append(requirements.Requirement(req))
+        except requirements.InvalidRequirement as exc:
+            raise self._invalid_metadata(
+                f"{req!r} is invalid for {{field}}", cause=exc
+            ) from exc
+        else:
+            return reqs
+
+    def _process_license_expression(
+        self, value: str
+    ) -> NormalizedLicenseExpression | None:
+        try:
+            return licenses.canonicalize_license_expression(value)
+        except ValueError as exc:
+            raise self._invalid_metadata(
+                f"{value!r} is invalid for {{field}}", cause=exc
+            ) from exc
+
+    def _process_license_files(self, value: list[str]) -> list[str]:
+        paths = []
+        for path in value:
+            if ".." in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "parent directory indicators are not allowed"
+                )
+            if "*" in path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be resolved"
+                )
+            if (
+                pathlib.PurePosixPath(path).is_absolute()
+                or pathlib.PureWindowsPath(path).is_absolute()
+            ):
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, paths must be relative"
+                )
+            if pathlib.PureWindowsPath(path).as_posix() != path:
+                raise self._invalid_metadata(
+                    f"{path!r} is invalid for {{field}}, "
+                    "paths must use '/' delimiter"
+                )
+            paths.append(path)
+        return paths
+
+
+class Metadata:
+    """Representation of distribution metadata.
+
+    Compared to :class:`RawMetadata`, this class provides objects representing
+    metadata fields instead of only using built-in types. Any invalid metadata
+    will cause :exc:`InvalidMetadata` to be raised (with a
+    :py:attr:`~BaseException.__cause__` attribute as appropriate).
+    """
+
+    _raw: RawMetadata
+
+    @classmethod
+    def from_raw(cls, data: RawMetadata, *, validate: bool = True) -> Metadata:
+        """Create an instance from :class:`RawMetadata`.
+
+        If *validate* is true, all metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        ins = cls()
+        ins._raw = data.copy()  # Mutations occur due to caching enriched values.
+
+        if validate:
+            exceptions: list[Exception] = []
+            try:
+                metadata_version = ins.metadata_version
+                metadata_age = _VALID_METADATA_VERSIONS.index(metadata_version)
+            except InvalidMetadata as metadata_version_exc:
+                exceptions.append(metadata_version_exc)
+                metadata_version = None
+
+            # Make sure to check for the fields that are present, the required
+            # fields (so their absence can be reported).
+            fields_to_check = frozenset(ins._raw) | _REQUIRED_ATTRS
+            # Remove fields that have already been checked.
+            fields_to_check -= {"metadata_version"}
+
+            for key in fields_to_check:
+                try:
+                    if metadata_version:
+                        # Can't use getattr() as that triggers descriptor protocol which
+                        # will fail due to no value for the instance argument.
+                        try:
+                            field_metadata_version = cls.__dict__[key].added
+                        except KeyError:
+                            exc = InvalidMetadata(key, f"unrecognized field: {key!r}")
+                            exceptions.append(exc)
+                            continue
+                        field_age = _VALID_METADATA_VERSIONS.index(
+                            field_metadata_version
+                        )
+                        if field_age > metadata_age:
+                            field = _RAW_TO_EMAIL_MAPPING[key]
+                            exc = InvalidMetadata(
+                                field,
+                                f"{field} introduced in metadata version "
+                                f"{field_metadata_version}, not {metadata_version}",
+                            )
+                            exceptions.append(exc)
+                            continue
+                    getattr(ins, key)
+                except InvalidMetadata as exc:
+                    exceptions.append(exc)
+
+            if exceptions:
+                raise ExceptionGroup("invalid metadata", exceptions)
+
+        return ins
+
+    @classmethod
+    def from_email(cls, data: bytes | str, *, validate: bool = True) -> Metadata:
+        """Parse metadata from email headers.
+
+        If *validate* is true, the metadata will be validated. All exceptions
+        related to validation will be gathered and raised as an :class:`ExceptionGroup`.
+        """
+        raw, unparsed = parse_email(data)
+
+        if validate:
+            exceptions: list[Exception] = []
+            for unparsed_key in unparsed:
+                if unparsed_key in _EMAIL_TO_RAW_MAPPING:
+                    message = f"{unparsed_key!r} has invalid data"
+                else:
+                    message = f"unrecognized field: {unparsed_key!r}"
+                exceptions.append(InvalidMetadata(unparsed_key, message))
+
+            if exceptions:
+                raise ExceptionGroup("unparsed", exceptions)
+
+        try:
+            return cls.from_raw(raw, validate=validate)
+        except ExceptionGroup as exc_group:
+            raise ExceptionGroup(
+                "invalid or unparsed metadata", exc_group.exceptions
+            ) from None
+
+    metadata_version: _Validator[_MetadataVersion] = _Validator()
+    """:external:ref:`core-metadata-metadata-version`
+    (required; validated to be a valid metadata version)"""
+    # `name` is not normalized/typed to NormalizedName so as to provide access to
+    # the original/raw name.
+    name: _Validator[str] = _Validator()
+    """:external:ref:`core-metadata-name`
+    (required; validated using :func:`~packaging.utils.canonicalize_name` and its
+    *validate* parameter)"""
+    version: _Validator[version_module.Version] = _Validator()
+    """:external:ref:`core-metadata-version` (required)"""
+    dynamic: _Validator[list[str] | None] = _Validator(
+        added="2.2",
+    )
+    """:external:ref:`core-metadata-dynamic`
+    (validated against core metadata field names and lowercased)"""
+    platforms: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-platform`"""
+    supported_platforms: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-supported-platform`"""
+    summary: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-summary` (validated to contain no newlines)"""
+    description: _Validator[str | None] = _Validator()  # TODO 2.1: can be in body
+    """:external:ref:`core-metadata-description`"""
+    description_content_type: _Validator[str | None] = _Validator(added="2.1")
+    """:external:ref:`core-metadata-description-content-type` (validated)"""
+    keywords: _Validator[list[str] | None] = _Validator()
+    """:external:ref:`core-metadata-keywords`"""
+    home_page: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-home-page`"""
+    download_url: _Validator[str | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-download-url`"""
+    author: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author`"""
+    author_email: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-author-email`"""
+    maintainer: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer`"""
+    maintainer_email: _Validator[str | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-maintainer-email`"""
+    license: _Validator[str | None] = _Validator()
+    """:external:ref:`core-metadata-license`"""
+    license_expression: _Validator[NormalizedLicenseExpression | None] = _Validator(
+        added="2.4"
+    )
+    """:external:ref:`core-metadata-license-expression`"""
+    license_files: _Validator[list[str] | None] = _Validator(added="2.4")
+    """:external:ref:`core-metadata-license-file`"""
+    classifiers: _Validator[list[str] | None] = _Validator(added="1.1")
+    """:external:ref:`core-metadata-classifier`"""
+    requires_dist: _Validator[list[requirements.Requirement] | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-dist`"""
+    requires_python: _Validator[specifiers.SpecifierSet | None] = _Validator(
+        added="1.2"
+    )
+    """:external:ref:`core-metadata-requires-python`"""
+    # Because `Requires-External` allows for non-PEP 440 version specifiers, we
+    # don't do any processing on the values.
+    requires_external: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-requires-external`"""
+    project_urls: _Validator[dict[str, str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-project-url`"""
+    # PEP 685 lets us raise an error if an extra doesn't pass `Name` validation
+    # regardless of metadata version.
+    provides_extra: _Validator[list[utils.NormalizedName] | None] = _Validator(
+        added="2.1",
+    )
+    """:external:ref:`core-metadata-provides-extra`"""
+    provides_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-provides-dist`"""
+    obsoletes_dist: _Validator[list[str] | None] = _Validator(added="1.2")
+    """:external:ref:`core-metadata-obsoletes-dist`"""
+    requires: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Requires`` (deprecated)"""
+    provides: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Provides`` (deprecated)"""
+    obsoletes: _Validator[list[str] | None] = _Validator(added="1.1")
+    """``Obsoletes`` (deprecated)"""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/requirements.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/requirements.py
new file mode 100644
index 00000000..4e068c95
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/requirements.py
@@ -0,0 +1,91 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+from __future__ import annotations
+
+from typing import Any, Iterator
+
+from ._parser import parse_requirement as _parse_requirement
+from ._tokenizer import ParserSyntaxError
+from .markers import Marker, _normalize_extra_values
+from .specifiers import SpecifierSet
+from .utils import canonicalize_name
+
+
+class InvalidRequirement(ValueError):
+    """
+    An invalid requirement was found, users should refer to PEP 508.
+    """
+
+
+class Requirement:
+    """Parse a requirement.
+
+    Parse a given requirement string into its parts, such as name, specifier,
+    URL, and extras. Raises InvalidRequirement on a badly-formed requirement
+    string.
+    """
+
+    # TODO: Can we test whether something is contained within a requirement?
+    #       If so how do we do that? Do we need to test against the _name_ of
+    #       the thing as well as the version? What about the markers?
+    # TODO: Can we normalize the name and extra name?
+
+    def __init__(self, requirement_string: str) -> None:
+        try:
+            parsed = _parse_requirement(requirement_string)
+        except ParserSyntaxError as e:
+            raise InvalidRequirement(str(e)) from e
+
+        self.name: str = parsed.name
+        self.url: str | None = parsed.url or None
+        self.extras: set[str] = set(parsed.extras or [])
+        self.specifier: SpecifierSet = SpecifierSet(parsed.specifier)
+        self.marker: Marker | None = None
+        if parsed.marker is not None:
+            self.marker = Marker.__new__(Marker)
+            self.marker._markers = _normalize_extra_values(parsed.marker)
+
+    def _iter_parts(self, name: str) -> Iterator[str]:
+        yield name
+
+        if self.extras:
+            formatted_extras = ",".join(sorted(self.extras))
+            yield f"[{formatted_extras}]"
+
+        if self.specifier:
+            yield str(self.specifier)
+
+        if self.url:
+            yield f"@ {self.url}"
+            if self.marker:
+                yield " "
+
+        if self.marker:
+            yield f"; {self.marker}"
+
+    def __str__(self) -> str:
+        return "".join(self._iter_parts(self.name))
+
+    def __repr__(self) -> str:
+        return f"<Requirement('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash(
+            (
+                self.__class__.__name__,
+                *self._iter_parts(canonicalize_name(self.name)),
+            )
+        )
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Requirement):
+            return NotImplemented
+
+        return (
+            canonicalize_name(self.name) == canonicalize_name(other.name)
+            and self.extras == other.extras
+            and self.specifier == other.specifier
+            and self.url == other.url
+            and self.marker == other.marker
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/specifiers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/specifiers.py
new file mode 100644
index 00000000..b30926af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/specifiers.py
@@ -0,0 +1,1020 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.specifiers import Specifier, SpecifierSet, InvalidSpecifier
+    from packaging.version import Version
+"""
+
+from __future__ import annotations
+
+import abc
+import itertools
+import re
+from typing import Callable, Iterable, Iterator, TypeVar, Union
+
+from .utils import canonicalize_version
+from .version import Version
+
+UnparsedVersion = Union[Version, str]
+UnparsedVersionVar = TypeVar("UnparsedVersionVar", bound=UnparsedVersion)
+CallableOperator = Callable[[Version, str], bool]
+
+
+def _coerce_version(version: UnparsedVersion) -> Version:
+    if not isinstance(version, Version):
+        version = Version(version)
+    return version
+
+
+class InvalidSpecifier(ValueError):
+    """
+    Raised when attempting to create a :class:`Specifier` with a specifier
+    string that is invalid.
+
+    >>> Specifier("lolwat")
+    Traceback (most recent call last):
+        ...
+    packaging.specifiers.InvalidSpecifier: Invalid specifier: 'lolwat'
+    """
+
+
+class BaseSpecifier(metaclass=abc.ABCMeta):
+    @abc.abstractmethod
+    def __str__(self) -> str:
+        """
+        Returns the str representation of this Specifier-like object. This
+        should be representative of the Specifier itself.
+        """
+
+    @abc.abstractmethod
+    def __hash__(self) -> int:
+        """
+        Returns a hash value for this Specifier-like object.
+        """
+
+    @abc.abstractmethod
+    def __eq__(self, other: object) -> bool:
+        """
+        Returns a boolean representing whether or not the two Specifier-like
+        objects are equal.
+
+        :param other: The other object to check against.
+        """
+
+    @property
+    @abc.abstractmethod
+    def prereleases(self) -> bool | None:
+        """Whether or not pre-releases as a whole are allowed.
+
+        This can be set to either ``True`` or ``False`` to explicitly enable or disable
+        prereleases or it can be set to ``None`` (the default) to use default semantics.
+        """
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        """Setter for :attr:`prereleases`.
+
+        :param value: The value to set.
+        """
+
+    @abc.abstractmethod
+    def contains(self, item: str, prereleases: bool | None = None) -> bool:
+        """
+        Determines if the given item is contained within this specifier.
+        """
+
+    @abc.abstractmethod
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """
+        Takes an iterable of items and filters them so that only items which
+        are contained within this specifier are allowed in it.
+        """
+
+
+class Specifier(BaseSpecifier):
+    """This class abstracts handling of version specifiers.
+
+    .. tip::
+
+        It is generally not required to instantiate this manually. You should instead
+        prefer to work with :class:`SpecifierSet` instead, which can parse
+        comma-separated version specifiers (which is what package metadata contains).
+    """
+
+    _operator_regex_str = r"""
+        (?P<operator>(~=|==|!=|<=|>=|<|>|===))
+        """
+    _version_regex_str = r"""
+        (?P<version>
+            (?:
+                # The identity operators allow for an escape hatch that will
+                # do an exact string match of the version you wish to install.
+                # This will not be parsed by PEP 440 and we cannot determine
+                # any semantic meaning from it. This operator is discouraged
+                # but included entirely as an escape hatch.
+                (?<====)  # Only match for the identity operator
+                \s*
+                [^\s;)]*  # The arbitrary version can be just about anything,
+                          # we match everything except for whitespace, a
+                          # semi-colon for marker support, and a closing paren
+                          # since versions can be enclosed in them.
+            )
+            |
+            (?:
+                # The (non)equality operators allow for wild card and local
+                # versions to be specified so we have to define these two
+                # operators separately to enable that.
+                (?<===|!=)            # Only match for equals and not equals
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+
+                # You cannot use a wild card and a pre-release, post-release, a dev or
+                # local version together so group them with a | and make them optional.
+                (?:
+                    \.\*  # Wild card syntax of .*
+                    |
+                    (?:                                  # pre release
+                        [-_\.]?
+                        (alpha|beta|preview|pre|a|b|c|rc)
+                        [-_\.]?
+                        [0-9]*
+                    )?
+                    (?:                                  # post release
+                        (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                    )?
+                    (?:[-_\.]?dev[-_\.]?[0-9]*)?         # dev release
+                    (?:\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*)? # local
+                )?
+            )
+            |
+            (?:
+                # The compatible operator requires at least two digits in the
+                # release segment.
+                (?<=~=)               # Only match for the compatible operator
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)+   # release  (We have a + instead of a *)
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+            |
+            (?:
+                # All other operators only allow a sub set of what the
+                # (non)equality operators do. Specifically they do not allow
+                # local versions to be specified nor do they allow the prefix
+                # matching wild cards.
+                (?<!==|!=|~=)         # We have special cases for these
+                                      # operators so we want to make sure they
+                                      # don't match here.
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+        )
+        """
+
+    _regex = re.compile(
+        r"^\s*" + _operator_regex_str + _version_regex_str + r"\s*$",
+        re.VERBOSE | re.IGNORECASE,
+    )
+
+    _operators = {
+        "~=": "compatible",
+        "==": "equal",
+        "!=": "not_equal",
+        "<=": "less_than_equal",
+        ">=": "greater_than_equal",
+        "<": "less_than",
+        ">": "greater_than",
+        "===": "arbitrary",
+    }
+
+    def __init__(self, spec: str = "", prereleases: bool | None = None) -> None:
+        """Initialize a Specifier instance.
+
+        :param spec:
+            The string representation of a specifier which will be parsed and
+            normalized before use.
+        :param prereleases:
+            This tells the specifier if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+        :raises InvalidSpecifier:
+            If the given specifier is invalid (i.e. bad syntax).
+        """
+        match = self._regex.search(spec)
+        if not match:
+            raise InvalidSpecifier(f"Invalid specifier: {spec!r}")
+
+        self._spec: tuple[str, str] = (
+            match.group("operator").strip(),
+            match.group("version").strip(),
+        )
+
+        # Store whether or not this Specifier should accept prereleases
+        self._prereleases = prereleases
+
+    # https://github.com/python/mypy/pull/13475#pullrequestreview-1079784515
+    @property  # type: ignore[override]
+    def prereleases(self) -> bool:
+        # If there is an explicit prereleases set for this, then we'll just
+        # blindly use that.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # Look at all of our specifiers and determine if they are inclusive
+        # operators, and if they are if they are including an explicit
+        # prerelease.
+        operator, version = self._spec
+        if operator in ["==", ">=", "<=", "~=", "===", ">", "<"]:
+            # The == specifier can include a trailing .*, if it does we
+            # want to remove before parsing.
+            if operator == "==" and version.endswith(".*"):
+                version = version[:-2]
+
+            # Parse the version, and if it is a pre-release than this
+            # specifier allows pre-releases.
+            if Version(version).is_prerelease:
+                return True
+
+        return False
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    @property
+    def operator(self) -> str:
+        """The operator of this specifier.
+
+        >>> Specifier("==1.2.3").operator
+        '=='
+        """
+        return self._spec[0]
+
+    @property
+    def version(self) -> str:
+        """The version of this specifier.
+
+        >>> Specifier("==1.2.3").version
+        '1.2.3'
+        """
+        return self._spec[1]
+
+    def __repr__(self) -> str:
+        """A representation of the Specifier that shows all internal state.
+
+        >>> Specifier('>=1.0.0')
+        <Specifier('>=1.0.0')>
+        >>> Specifier('>=1.0.0', prereleases=False)
+        <Specifier('>=1.0.0', prereleases=False)>
+        >>> Specifier('>=1.0.0', prereleases=True)
+        <Specifier('>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<{self.__class__.__name__}({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the Specifier that can be round-tripped.
+
+        >>> str(Specifier('>=1.0.0'))
+        '>=1.0.0'
+        >>> str(Specifier('>=1.0.0', prereleases=False))
+        '>=1.0.0'
+        """
+        return "{}{}".format(*self._spec)
+
+    @property
+    def _canonical_spec(self) -> tuple[str, str]:
+        canonical_version = canonicalize_version(
+            self._spec[1],
+            strip_trailing_zero=(self._spec[0] != "~="),
+        )
+        return self._spec[0], canonical_version
+
+    def __hash__(self) -> int:
+        return hash(self._canonical_spec)
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two Specifier-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> Specifier("==1.2.3") == Specifier("== 1.2.3.0")
+        True
+        >>> (Specifier("==1.2.3", prereleases=False) ==
+        ...  Specifier("==1.2.3", prereleases=True))
+        True
+        >>> Specifier("==1.2.3") == "==1.2.3"
+        True
+        >>> Specifier("==1.2.3") == Specifier("==1.2.4")
+        False
+        >>> Specifier("==1.2.3") == Specifier("~=1.2.3")
+        False
+        """
+        if isinstance(other, str):
+            try:
+                other = self.__class__(str(other))
+            except InvalidSpecifier:
+                return NotImplemented
+        elif not isinstance(other, self.__class__):
+            return NotImplemented
+
+        return self._canonical_spec == other._canonical_spec
+
+    def _get_operator(self, op: str) -> CallableOperator:
+        operator_callable: CallableOperator = getattr(
+            self, f"_compare_{self._operators[op]}"
+        )
+        return operator_callable
+
+    def _compare_compatible(self, prospective: Version, spec: str) -> bool:
+        # Compatible releases have an equivalent combination of >= and ==. That
+        # is that ~=2.2 is equivalent to >=2.2,==2.*. This allows us to
+        # implement this in terms of the other specifiers instead of
+        # implementing it ourselves. The only thing we need to do is construct
+        # the other specifiers.
+
+        # We want everything but the last item in the version, but we want to
+        # ignore suffix segments.
+        prefix = _version_join(
+            list(itertools.takewhile(_is_not_suffix, _version_split(spec)))[:-1]
+        )
+
+        # Add the prefix notation to the end of our string
+        prefix += ".*"
+
+        return self._get_operator(">=")(prospective, spec) and self._get_operator("==")(
+            prospective, prefix
+        )
+
+    def _compare_equal(self, prospective: Version, spec: str) -> bool:
+        # We need special logic to handle prefix matching
+        if spec.endswith(".*"):
+            # In the case of prefix matching we want to ignore local segment.
+            normalized_prospective = canonicalize_version(
+                prospective.public, strip_trailing_zero=False
+            )
+            # Get the normalized version string ignoring the trailing .*
+            normalized_spec = canonicalize_version(spec[:-2], strip_trailing_zero=False)
+            # Split the spec out by bangs and dots, and pretend that there is
+            # an implicit dot in between a release segment and a pre-release segment.
+            split_spec = _version_split(normalized_spec)
+
+            # Split the prospective version out by bangs and dots, and pretend
+            # that there is an implicit dot in between a release segment and
+            # a pre-release segment.
+            split_prospective = _version_split(normalized_prospective)
+
+            # 0-pad the prospective version before shortening it to get the correct
+            # shortened version.
+            padded_prospective, _ = _pad_version(split_prospective, split_spec)
+
+            # Shorten the prospective version to be the same length as the spec
+            # so that we can determine if the specifier is a prefix of the
+            # prospective version or not.
+            shortened_prospective = padded_prospective[: len(split_spec)]
+
+            return shortened_prospective == split_spec
+        else:
+            # Convert our spec string into a Version
+            spec_version = Version(spec)
+
+            # If the specifier does not have a local segment, then we want to
+            # act as if the prospective version also does not have a local
+            # segment.
+            if not spec_version.local:
+                prospective = Version(prospective.public)
+
+            return prospective == spec_version
+
+    def _compare_not_equal(self, prospective: Version, spec: str) -> bool:
+        return not self._compare_equal(prospective, spec)
+
+    def _compare_less_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) <= Version(spec)
+
+    def _compare_greater_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) >= Version(spec)
+
+    def _compare_less_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is less than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective < spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a pre-release version, that we do not accept pre-release
+        # versions for the version mentioned in the specifier (e.g. <3.1 should
+        # not match 3.1.dev0, but should match 3.0.dev0).
+        if not spec.is_prerelease and prospective.is_prerelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # less than the spec version *and* it's not a pre-release of the same
+        # version in the spec.
+        return True
+
+    def _compare_greater_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is greater than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective > spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a post-release version, that we do not accept
+        # post-release versions for the version mentioned in the specifier
+        # (e.g. >3.1 should not match 3.0.post0, but should match 3.2.post0).
+        if not spec.is_postrelease and prospective.is_postrelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # Ensure that we do not allow a local version of the version mentioned
+        # in the specifier, which is technically greater than, to match.
+        if prospective.local is not None:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # greater than the spec version *and* it's not a pre-release of the
+        # same version in the spec.
+        return True
+
+    def _compare_arbitrary(self, prospective: Version, spec: str) -> bool:
+        return str(prospective).lower() == str(spec).lower()
+
+    def __contains__(self, item: str | Version) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in Specifier(">=1.2.3")
+        True
+        >>> Version("1.2.3") in Specifier(">=1.2.3")
+        True
+        >>> "1.0.0" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(self, item: UnparsedVersion, prereleases: bool | None = None) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this Specifier. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> Specifier(">=1.2.3").contains("1.2.3")
+        True
+        >>> Specifier(">=1.2.3").contains(Version("1.2.3"))
+        True
+        >>> Specifier(">=1.2.3").contains("1.0.0")
+        False
+        >>> Specifier(">=1.2.3").contains("1.3.0a1")
+        False
+        >>> Specifier(">=1.2.3", prereleases=True).contains("1.3.0a1")
+        True
+        >>> Specifier(">=1.2.3").contains("1.3.0a1", prereleases=True)
+        True
+        """
+
+        # Determine if prereleases are to be allowed or not.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # Normalize item to a Version, this allows us to have a shortcut for
+        # "2.0" in Specifier(">=2")
+        normalized_item = _coerce_version(item)
+
+        # Determine if we should be supporting prereleases in this specifier
+        # or not, if we do not support prereleases than we can short circuit
+        # logic if this version is a prereleases.
+        if normalized_item.is_prerelease and not prereleases:
+            return False
+
+        # Actually do the comparison to determine if this item is contained
+        # within this Specifier or not.
+        operator_callable: CallableOperator = self._get_operator(self.operator)
+        return operator_callable(normalized_item, self.version)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifier.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(Specifier().contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.2.3", "1.3", Version("1.4")]))
+        ['1.2.3', '1.3', <Version('1.4')>]
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.5a1"]))
+        ['1.5a1']
+        >>> list(Specifier(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(Specifier(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        """
+
+        yielded = False
+        found_prereleases = []
+
+        kw = {"prereleases": prereleases if prereleases is not None else True}
+
+        # Attempt to iterate over all the values in the iterable and if any of
+        # them match, yield them.
+        for version in iterable:
+            parsed_version = _coerce_version(version)
+
+            if self.contains(parsed_version, **kw):
+                # If our version is a prerelease, and we were not set to allow
+                # prereleases, then we'll store it for later in case nothing
+                # else matches this specifier.
+                if parsed_version.is_prerelease and not (
+                    prereleases or self.prereleases
+                ):
+                    found_prereleases.append(version)
+                # Either this is not a prerelease, or we should have been
+                # accepting prereleases from the beginning.
+                else:
+                    yielded = True
+                    yield version
+
+        # Now that we've iterated over everything, determine if we've yielded
+        # any values, and if we have not and we have any prereleases stored up
+        # then we will go ahead and yield the prereleases.
+        if not yielded and found_prereleases:
+            for version in found_prereleases:
+                yield version
+
+
+_prefix_regex = re.compile(r"^([0-9]+)((?:a|b|c|rc)[0-9]+)$")
+
+
+def _version_split(version: str) -> list[str]:
+    """Split version into components.
+
+    The split components are intended for version comparison. The logic does
+    not attempt to retain the original version string, so joining the
+    components back with :func:`_version_join` may not produce the original
+    version string.
+    """
+    result: list[str] = []
+
+    epoch, _, rest = version.rpartition("!")
+    result.append(epoch or "0")
+
+    for item in rest.split("."):
+        match = _prefix_regex.search(item)
+        if match:
+            result.extend(match.groups())
+        else:
+            result.append(item)
+    return result
+
+
+def _version_join(components: list[str]) -> str:
+    """Join split version components into a version string.
+
+    This function assumes the input came from :func:`_version_split`, where the
+    first component must be the epoch (either empty or numeric), and all other
+    components numeric.
+    """
+    epoch, *rest = components
+    return f"{epoch}!{'.'.join(rest)}"
+
+
+def _is_not_suffix(segment: str) -> bool:
+    return not any(
+        segment.startswith(prefix) for prefix in ("dev", "a", "b", "rc", "post")
+    )
+
+
+def _pad_version(left: list[str], right: list[str]) -> tuple[list[str], list[str]]:
+    left_split, right_split = [], []
+
+    # Get the release segment of our versions
+    left_split.append(list(itertools.takewhile(lambda x: x.isdigit(), left)))
+    right_split.append(list(itertools.takewhile(lambda x: x.isdigit(), right)))
+
+    # Get the rest of our versions
+    left_split.append(left[len(left_split[0]) :])
+    right_split.append(right[len(right_split[0]) :])
+
+    # Insert our padding
+    left_split.insert(1, ["0"] * max(0, len(right_split[0]) - len(left_split[0])))
+    right_split.insert(1, ["0"] * max(0, len(left_split[0]) - len(right_split[0])))
+
+    return (
+        list(itertools.chain.from_iterable(left_split)),
+        list(itertools.chain.from_iterable(right_split)),
+    )
+
+
+class SpecifierSet(BaseSpecifier):
+    """This class abstracts handling of a set of version specifiers.
+
+    It can be passed a single specifier (``>=3.0``), a comma-separated list of
+    specifiers (``>=3.0,!=3.1``), or no specifier at all.
+    """
+
+    def __init__(
+        self,
+        specifiers: str | Iterable[Specifier] = "",
+        prereleases: bool | None = None,
+    ) -> None:
+        """Initialize a SpecifierSet instance.
+
+        :param specifiers:
+            The string representation of a specifier or a comma-separated list of
+            specifiers which will be parsed and normalized before use.
+            May also be an iterable of ``Specifier`` instances, which will be used
+            as is.
+        :param prereleases:
+            This tells the SpecifierSet if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+
+        :raises InvalidSpecifier:
+            If the given ``specifiers`` are not parseable than this exception will be
+            raised.
+        """
+
+        if isinstance(specifiers, str):
+            # Split on `,` to break each individual specifier into its own item, and
+            # strip each item to remove leading/trailing whitespace.
+            split_specifiers = [s.strip() for s in specifiers.split(",") if s.strip()]
+
+            # Make each individual specifier a Specifier and save in a frozen set
+            # for later.
+            self._specs = frozenset(map(Specifier, split_specifiers))
+        else:
+            # Save the supplied specifiers in a frozen set.
+            self._specs = frozenset(specifiers)
+
+        # Store our prereleases value so we can use it later to determine if
+        # we accept prereleases or not.
+        self._prereleases = prereleases
+
+    @property
+    def prereleases(self) -> bool | None:
+        # If we have been given an explicit prerelease modifier, then we'll
+        # pass that through here.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # If we don't have any specifiers, and we don't have a forced value,
+        # then we'll just return None since we don't know if this should have
+        # pre-releases or not.
+        if not self._specs:
+            return None
+
+        # Otherwise we'll see if any of the given specifiers accept
+        # prereleases, if any of them do we'll return True, otherwise False.
+        return any(s.prereleases for s in self._specs)
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    def __repr__(self) -> str:
+        """A representation of the specifier set that shows all internal state.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> SpecifierSet('>=1.0.0,!=2.0.0')
+        <SpecifierSet('!=2.0.0,>=1.0.0')>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=False)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=False)>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=True)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<SpecifierSet({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the specifier set that can be round-tripped.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1"))
+        '!=1.0.1,>=1.0.0'
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False))
+        '!=1.0.1,>=1.0.0'
+        """
+        return ",".join(sorted(str(s) for s in self._specs))
+
+    def __hash__(self) -> int:
+        return hash(self._specs)
+
+    def __and__(self, other: SpecifierSet | str) -> SpecifierSet:
+        """Return a SpecifierSet which is a combination of the two sets.
+
+        :param other: The other object to combine with.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & '<=2.0.0,!=2.0.1'
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & SpecifierSet('<=2.0.0,!=2.0.1')
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        """
+        if isinstance(other, str):
+            other = SpecifierSet(other)
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        specifier = SpecifierSet()
+        specifier._specs = frozenset(self._specs | other._specs)
+
+        if self._prereleases is None and other._prereleases is not None:
+            specifier._prereleases = other._prereleases
+        elif self._prereleases is not None and other._prereleases is None:
+            specifier._prereleases = self._prereleases
+        elif self._prereleases == other._prereleases:
+            specifier._prereleases = self._prereleases
+        else:
+            raise ValueError(
+                "Cannot combine SpecifierSets with True and False prerelease "
+                "overrides."
+            )
+
+        return specifier
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two SpecifierSet-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> (SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False) ==
+        ...  SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == ">=1.0.0,!=1.0.1"
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.2")
+        False
+        """
+        if isinstance(other, (str, Specifier)):
+            other = SpecifierSet(str(other))
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        return self._specs == other._specs
+
+    def __len__(self) -> int:
+        """Returns the number of specifiers in this specifier set."""
+        return len(self._specs)
+
+    def __iter__(self) -> Iterator[Specifier]:
+        """
+        Returns an iterator over all the underlying :class:`Specifier` instances
+        in this specifier set.
+
+        >>> sorted(SpecifierSet(">=1.0.0,!=1.0.1"), key=str)
+        [<Specifier('!=1.0.1')>, <Specifier('>=1.0.0')>]
+        """
+        return iter(self._specs)
+
+    def __contains__(self, item: UnparsedVersion) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> Version("1.2.3") in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> "1.0.1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(
+        self,
+        item: UnparsedVersion,
+        prereleases: bool | None = None,
+        installed: bool | None = None,
+    ) -> bool:
+        """Return whether or not the item is contained in this SpecifierSet.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this SpecifierSet. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.2.3")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains(Version("1.2.3"))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.0.1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True).contains("1.3.0a1")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1", prereleases=True)
+        True
+        """
+        # Ensure that our item is a Version instance.
+        if not isinstance(item, Version):
+            item = Version(item)
+
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # We can determine if we're going to allow pre-releases by looking to
+        # see if any of the underlying items supports them. If none of them do
+        # and this item is a pre-release then we do not allow it and we can
+        # short circuit that here.
+        # Note: This means that 1.0.dev1 would not be contained in something
+        #       like >=1.0.devabc however it would be in >=1.0.debabc,>0.0.dev0
+        if not prereleases and item.is_prerelease:
+            return False
+
+        if installed and item.is_prerelease:
+            item = Version(item.base_version)
+
+        # We simply dispatch to the underlying specs here to make sure that the
+        # given version is contained within all of them.
+        # Note: This use of all() here means that an empty set of specifiers
+        #       will always return True, this is an explicit design decision.
+        return all(s.contains(item, prereleases=prereleases) for s in self._specs)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: bool | None = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifiers in this set.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(SpecifierSet(...).contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", Version("1.4")]))
+        ['1.3', <Version('1.4')>]
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.5a1"]))
+        []
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+
+        An "empty" SpecifierSet will filter items based on the presence of prerelease
+        versions in the set.
+
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet("").filter(["1.5a1"]))
+        ['1.5a1']
+        >>> list(SpecifierSet("", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        """
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # If we have any specifiers, then we want to wrap our iterable in the
+        # filter method for each one, this will act as a logical AND amongst
+        # each specifier.
+        if self._specs:
+            for spec in self._specs:
+                iterable = spec.filter(iterable, prereleases=bool(prereleases))
+            return iter(iterable)
+        # If we do not have any specifiers, then we need to have a rough filter
+        # which will filter out any pre-releases, unless there are no final
+        # releases.
+        else:
+            filtered: list[UnparsedVersionVar] = []
+            found_prereleases: list[UnparsedVersionVar] = []
+
+            for item in iterable:
+                parsed_version = _coerce_version(item)
+
+                # Store any item which is a pre-release for later unless we've
+                # already found a final version or we are accepting prereleases
+                if parsed_version.is_prerelease and not prereleases:
+                    if not filtered:
+                        found_prereleases.append(item)
+                else:
+                    filtered.append(item)
+
+            # If we've found no items except for pre-releases, then we'll go
+            # ahead and use the pre-releases
+            if not filtered and found_prereleases and prereleases is None:
+                return iter(found_prereleases)
+
+            return iter(filtered)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/tags.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/tags.py
new file mode 100644
index 00000000..f5903402
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/tags.py
@@ -0,0 +1,617 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import logging
+import platform
+import re
+import struct
+import subprocess
+import sys
+import sysconfig
+from importlib.machinery import EXTENSION_SUFFIXES
+from typing import (
+    Iterable,
+    Iterator,
+    Sequence,
+    Tuple,
+    cast,
+)
+
+from . import _manylinux, _musllinux
+
+logger = logging.getLogger(__name__)
+
+PythonVersion = Sequence[int]
+AppleVersion = Tuple[int, int]
+
+INTERPRETER_SHORT_NAMES: dict[str, str] = {
+    "python": "py",  # Generic.
+    "cpython": "cp",
+    "pypy": "pp",
+    "ironpython": "ip",
+    "jython": "jy",
+}
+
+
+_32_BIT_INTERPRETER = struct.calcsize("P") == 4
+
+
+class Tag:
+    """
+    A representation of the tag triple for a wheel.
+
+    Instances are considered immutable and thus are hashable. Equality checking
+    is also supported.
+    """
+
+    __slots__ = ["_abi", "_hash", "_interpreter", "_platform"]
+
+    def __init__(self, interpreter: str, abi: str, platform: str) -> None:
+        self._interpreter = interpreter.lower()
+        self._abi = abi.lower()
+        self._platform = platform.lower()
+        # The __hash__ of every single element in a Set[Tag] will be evaluated each time
+        # that a set calls its `.disjoint()` method, which may be called hundreds of
+        # times when scanning a page of links for packages with tags matching that
+        # Set[Tag]. Pre-computing the value here produces significant speedups for
+        # downstream consumers.
+        self._hash = hash((self._interpreter, self._abi, self._platform))
+
+    @property
+    def interpreter(self) -> str:
+        return self._interpreter
+
+    @property
+    def abi(self) -> str:
+        return self._abi
+
+    @property
+    def platform(self) -> str:
+        return self._platform
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Tag):
+            return NotImplemented
+
+        return (
+            (self._hash == other._hash)  # Short-circuit ASAP for perf reasons.
+            and (self._platform == other._platform)
+            and (self._abi == other._abi)
+            and (self._interpreter == other._interpreter)
+        )
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __str__(self) -> str:
+        return f"{self._interpreter}-{self._abi}-{self._platform}"
+
+    def __repr__(self) -> str:
+        return f"<{self} @ {id(self)}>"
+
+
+def parse_tag(tag: str) -> frozenset[Tag]:
+    """
+    Parses the provided tag (e.g. `py3-none-any`) into a frozenset of Tag instances.
+
+    Returning a set is required due to the possibility that the tag is a
+    compressed tag set.
+    """
+    tags = set()
+    interpreters, abis, platforms = tag.split("-")
+    for interpreter in interpreters.split("."):
+        for abi in abis.split("."):
+            for platform_ in platforms.split("."):
+                tags.add(Tag(interpreter, abi, platform_))
+    return frozenset(tags)
+
+
+def _get_config_var(name: str, warn: bool = False) -> int | str | None:
+    value: int | str | None = sysconfig.get_config_var(name)
+    if value is None and warn:
+        logger.debug(
+            "Config variable '%s' is unset, Python ABI tag may be incorrect", name
+        )
+    return value
+
+
+def _normalize_string(string: str) -> str:
+    return string.replace(".", "_").replace("-", "_").replace(" ", "_")
+
+
+def _is_threaded_cpython(abis: list[str]) -> bool:
+    """
+    Determine if the ABI corresponds to a threaded (`--disable-gil`) build.
+
+    The threaded builds are indicated by a "t" in the abiflags.
+    """
+    if len(abis) == 0:
+        return False
+    # expect e.g., cp313
+    m = re.match(r"cp\d+(.*)", abis[0])
+    if not m:
+        return False
+    abiflags = m.group(1)
+    return "t" in abiflags
+
+
+def _abi3_applies(python_version: PythonVersion, threading: bool) -> bool:
+    """
+    Determine if the Python version supports abi3.
+
+    PEP 384 was first implemented in Python 3.2. The threaded (`--disable-gil`)
+    builds do not support abi3.
+    """
+    return len(python_version) > 1 and tuple(python_version) >= (3, 2) and not threading
+
+
+def _cpython_abis(py_version: PythonVersion, warn: bool = False) -> list[str]:
+    py_version = tuple(py_version)  # To allow for version comparison.
+    abis = []
+    version = _version_nodot(py_version[:2])
+    threading = debug = pymalloc = ucs4 = ""
+    with_debug = _get_config_var("Py_DEBUG", warn)
+    has_refcount = hasattr(sys, "gettotalrefcount")
+    # Windows doesn't set Py_DEBUG, so checking for support of debug-compiled
+    # extension modules is the best option.
+    # https://github.com/pypa/pip/issues/3383#issuecomment-173267692
+    has_ext = "_d.pyd" in EXTENSION_SUFFIXES
+    if with_debug or (with_debug is None and (has_refcount or has_ext)):
+        debug = "d"
+    if py_version >= (3, 13) and _get_config_var("Py_GIL_DISABLED", warn):
+        threading = "t"
+    if py_version < (3, 8):
+        with_pymalloc = _get_config_var("WITH_PYMALLOC", warn)
+        if with_pymalloc or with_pymalloc is None:
+            pymalloc = "m"
+        if py_version < (3, 3):
+            unicode_size = _get_config_var("Py_UNICODE_SIZE", warn)
+            if unicode_size == 4 or (
+                unicode_size is None and sys.maxunicode == 0x10FFFF
+            ):
+                ucs4 = "u"
+    elif debug:
+        # Debug builds can also load "normal" extension modules.
+        # We can also assume no UCS-4 or pymalloc requirement.
+        abis.append(f"cp{version}{threading}")
+    abis.insert(0, f"cp{version}{threading}{debug}{pymalloc}{ucs4}")
+    return abis
+
+
+def cpython_tags(
+    python_version: PythonVersion | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a CPython interpreter.
+
+    The tags consist of:
+    - cp<python_version>-<abi>-<platform>
+    - cp<python_version>-abi3-<platform>
+    - cp<python_version>-none-<platform>
+    - cp<less than python_version>-abi3-<platform>  # Older Python versions down to 3.2.
+
+    If python_version only specifies a major version then user-provided ABIs and
+    the 'none' ABItag will be used.
+
+    If 'abi3' or 'none' are specified in 'abis' then they will be yielded at
+    their normal position and not at the beginning.
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+
+    interpreter = f"cp{_version_nodot(python_version[:2])}"
+
+    if abis is None:
+        if len(python_version) > 1:
+            abis = _cpython_abis(python_version, warn)
+        else:
+            abis = []
+    abis = list(abis)
+    # 'abi3' and 'none' are explicitly handled later.
+    for explicit_abi in ("abi3", "none"):
+        try:
+            abis.remove(explicit_abi)
+        except ValueError:
+            pass
+
+    platforms = list(platforms or platform_tags())
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+    threading = _is_threaded_cpython(abis)
+    use_abi3 = _abi3_applies(python_version, threading)
+    if use_abi3:
+        yield from (Tag(interpreter, "abi3", platform_) for platform_ in platforms)
+    yield from (Tag(interpreter, "none", platform_) for platform_ in platforms)
+
+    if use_abi3:
+        for minor_version in range(python_version[1] - 1, 1, -1):
+            for platform_ in platforms:
+                version = _version_nodot((python_version[0], minor_version))
+                interpreter = f"cp{version}"
+                yield Tag(interpreter, "abi3", platform_)
+
+
+def _generic_abi() -> list[str]:
+    """
+    Return the ABI tag based on EXT_SUFFIX.
+    """
+    # The following are examples of `EXT_SUFFIX`.
+    # We want to keep the parts which are related to the ABI and remove the
+    # parts which are related to the platform:
+    # - linux:   '.cpython-310-x86_64-linux-gnu.so' => cp310
+    # - mac:     '.cpython-310-darwin.so'           => cp310
+    # - win:     '.cp310-win_amd64.pyd'             => cp310
+    # - win:     '.pyd'                             => cp37 (uses _cpython_abis())
+    # - pypy:    '.pypy38-pp73-x86_64-linux-gnu.so' => pypy38_pp73
+    # - graalpy: '.graalpy-38-native-x86_64-darwin.dylib'
+    #                                               => graalpy_38_native
+
+    ext_suffix = _get_config_var("EXT_SUFFIX", warn=True)
+    if not isinstance(ext_suffix, str) or ext_suffix[0] != ".":
+        raise SystemError("invalid sysconfig.get_config_var('EXT_SUFFIX')")
+    parts = ext_suffix.split(".")
+    if len(parts) < 3:
+        # CPython3.7 and earlier uses ".pyd" on Windows.
+        return _cpython_abis(sys.version_info[:2])
+    soabi = parts[1]
+    if soabi.startswith("cpython"):
+        # non-windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi.startswith("cp"):
+        # windows
+        abi = soabi.split("-")[0]
+    elif soabi.startswith("pypy"):
+        abi = "-".join(soabi.split("-")[:2])
+    elif soabi.startswith("graalpy"):
+        abi = "-".join(soabi.split("-")[:3])
+    elif soabi:
+        # pyston, ironpython, others?
+        abi = soabi
+    else:
+        return []
+    return [_normalize_string(abi)]
+
+
+def generic_tags(
+    interpreter: str | None = None,
+    abis: Iterable[str] | None = None,
+    platforms: Iterable[str] | None = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a generic interpreter.
+
+    The tags consist of:
+    - <interpreter>-<abi>-<platform>
+
+    The "none" ABI will be added if it was not explicitly provided.
+    """
+    if not interpreter:
+        interp_name = interpreter_name()
+        interp_version = interpreter_version(warn=warn)
+        interpreter = "".join([interp_name, interp_version])
+    if abis is None:
+        abis = _generic_abi()
+    else:
+        abis = list(abis)
+    platforms = list(platforms or platform_tags())
+    if "none" not in abis:
+        abis.append("none")
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+
+def _py_interpreter_range(py_version: PythonVersion) -> Iterator[str]:
+    """
+    Yields Python versions in descending order.
+
+    After the latest version, the major-only version will be yielded, and then
+    all previous versions of that major version.
+    """
+    if len(py_version) > 1:
+        yield f"py{_version_nodot(py_version[:2])}"
+    yield f"py{py_version[0]}"
+    if len(py_version) > 1:
+        for minor in range(py_version[1] - 1, -1, -1):
+            yield f"py{_version_nodot((py_version[0], minor))}"
+
+
+def compatible_tags(
+    python_version: PythonVersion | None = None,
+    interpreter: str | None = None,
+    platforms: Iterable[str] | None = None,
+) -> Iterator[Tag]:
+    """
+    Yields the sequence of tags that are compatible with a specific version of Python.
+
+    The tags consist of:
+    - py*-none-<platform>
+    - <interpreter>-none-any  # ... if `interpreter` is provided.
+    - py*-none-any
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+    platforms = list(platforms or platform_tags())
+    for version in _py_interpreter_range(python_version):
+        for platform_ in platforms:
+            yield Tag(version, "none", platform_)
+    if interpreter:
+        yield Tag(interpreter, "none", "any")
+    for version in _py_interpreter_range(python_version):
+        yield Tag(version, "none", "any")
+
+
+def _mac_arch(arch: str, is_32bit: bool = _32_BIT_INTERPRETER) -> str:
+    if not is_32bit:
+        return arch
+
+    if arch.startswith("ppc"):
+        return "ppc"
+
+    return "i386"
+
+
+def _mac_binary_formats(version: AppleVersion, cpu_arch: str) -> list[str]:
+    formats = [cpu_arch]
+    if cpu_arch == "x86_64":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat64", "fat32"])
+
+    elif cpu_arch == "i386":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat32", "fat"])
+
+    elif cpu_arch == "ppc64":
+        # TODO: Need to care about 32-bit PPC for ppc64 through 10.2?
+        if version > (10, 5) or version < (10, 4):
+            return []
+        formats.append("fat64")
+
+    elif cpu_arch == "ppc":
+        if version > (10, 6):
+            return []
+        formats.extend(["fat32", "fat"])
+
+    if cpu_arch in {"arm64", "x86_64"}:
+        formats.append("universal2")
+
+    if cpu_arch in {"x86_64", "i386", "ppc64", "ppc", "intel"}:
+        formats.append("universal")
+
+    return formats
+
+
+def mac_platforms(
+    version: AppleVersion | None = None, arch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for a macOS system.
+
+    The `version` parameter is a two-item tuple specifying the macOS version to
+    generate platform tags for. The `arch` parameter is the CPU architecture to
+    generate platform tags for. Both parameters default to the appropriate value
+    for the current system.
+    """
+    version_str, _, cpu_arch = platform.mac_ver()
+    if version is None:
+        version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+        if version == (10, 16):
+            # When built against an older macOS SDK, Python will report macOS 10.16
+            # instead of the real version.
+            version_str = subprocess.run(
+                [
+                    sys.executable,
+                    "-sS",
+                    "-c",
+                    "import platform; print(platform.mac_ver()[0])",
+                ],
+                check=True,
+                env={"SYSTEM_VERSION_COMPAT": "0"},
+                stdout=subprocess.PIPE,
+                text=True,
+            ).stdout
+            version = cast("AppleVersion", tuple(map(int, version_str.split(".")[:2])))
+    else:
+        version = version
+    if arch is None:
+        arch = _mac_arch(cpu_arch)
+    else:
+        arch = arch
+
+    if (10, 0) <= version and version < (11, 0):
+        # Prior to Mac OS 11, each yearly release of Mac OS bumped the
+        # "minor" version number.  The major version was always 10.
+        major_version = 10
+        for minor_version in range(version[1], -1, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Starting with Mac OS 11, each yearly release bumps the major version
+        # number.   The minor versions are now the midyear updates.
+        minor_version = 0
+        for major_version in range(version[0], 10, -1):
+            compat_version = major_version, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+    if version >= (11, 0):
+        # Mac OS 11 on x86_64 is compatible with binaries from previous releases.
+        # Arm64 support was introduced in 11.0, so no Arm binaries from previous
+        # releases exist.
+        #
+        # However, the "universal2" binary format can have a
+        # macOS version earlier than 11.0 when the x86_64 part of the binary supports
+        # that version of macOS.
+        major_version = 10
+        if arch == "x86_64":
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_formats = _mac_binary_formats(compat_version, arch)
+                for binary_format in binary_formats:
+                    yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+        else:
+            for minor_version in range(16, 3, -1):
+                compat_version = major_version, minor_version
+                binary_format = "universal2"
+                yield f"macosx_{major_version}_{minor_version}_{binary_format}"
+
+
+def ios_platforms(
+    version: AppleVersion | None = None, multiarch: str | None = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for an iOS system.
+
+    :param version: A two-item tuple specifying the iOS version to generate
+        platform tags for. Defaults to the current iOS version.
+    :param multiarch: The CPU architecture+ABI to generate platform tags for -
+        (the value used by `sys.implementation._multiarch` e.g.,
+        `arm64_iphoneos` or `x84_64_iphonesimulator`). Defaults to the current
+        multiarch value.
+    """
+    if version is None:
+        # if iOS is the current platform, ios_ver *must* be defined. However,
+        # it won't exist for CPython versions before 3.13, which causes a mypy
+        # error.
+        _, release, _, _ = platform.ios_ver()  # type: ignore[attr-defined, unused-ignore]
+        version = cast("AppleVersion", tuple(map(int, release.split(".")[:2])))
+
+    if multiarch is None:
+        multiarch = sys.implementation._multiarch
+    multiarch = multiarch.replace("-", "_")
+
+    ios_platform_template = "ios_{major}_{minor}_{multiarch}"
+
+    # Consider any iOS major.minor version from the version requested, down to
+    # 12.0. 12.0 is the first iOS version that is known to have enough features
+    # to support CPython. Consider every possible minor release up to X.9. There
+    # highest the minor has ever gone is 8 (14.8 and 15.8) but having some extra
+    # candidates that won't ever match doesn't really hurt, and it saves us from
+    # having to keep an explicit list of known iOS versions in the code. Return
+    # the results descending order of version number.
+
+    # If the requested major version is less than 12, there won't be any matches.
+    if version[0] < 12:
+        return
+
+    # Consider the actual X.Y version that was requested.
+    yield ios_platform_template.format(
+        major=version[0], minor=version[1], multiarch=multiarch
+    )
+
+    # Consider every minor version from X.0 to the minor version prior to the
+    # version requested by the platform.
+    for minor in range(version[1] - 1, -1, -1):
+        yield ios_platform_template.format(
+            major=version[0], minor=minor, multiarch=multiarch
+        )
+
+    for major in range(version[0] - 1, 11, -1):
+        for minor in range(9, -1, -1):
+            yield ios_platform_template.format(
+                major=major, minor=minor, multiarch=multiarch
+            )
+
+
+def _linux_platforms(is_32bit: bool = _32_BIT_INTERPRETER) -> Iterator[str]:
+    linux = _normalize_string(sysconfig.get_platform())
+    if not linux.startswith("linux_"):
+        # we should never be here, just yield the sysconfig one and return
+        yield linux
+        return
+    if is_32bit:
+        if linux == "linux_x86_64":
+            linux = "linux_i686"
+        elif linux == "linux_aarch64":
+            linux = "linux_armv8l"
+    _, arch = linux.split("_", 1)
+    archs = {"armv8l": ["armv8l", "armv7l"]}.get(arch, [arch])
+    yield from _manylinux.platform_tags(archs)
+    yield from _musllinux.platform_tags(archs)
+    for arch in archs:
+        yield f"linux_{arch}"
+
+
+def _generic_platforms() -> Iterator[str]:
+    yield _normalize_string(sysconfig.get_platform())
+
+
+def platform_tags() -> Iterator[str]:
+    """
+    Provides the platform tags for this installation.
+    """
+    if platform.system() == "Darwin":
+        return mac_platforms()
+    elif platform.system() == "iOS":
+        return ios_platforms()
+    elif platform.system() == "Linux":
+        return _linux_platforms()
+    else:
+        return _generic_platforms()
+
+
+def interpreter_name() -> str:
+    """
+    Returns the name of the running interpreter.
+
+    Some implementations have a reserved, two-letter abbreviation which will
+    be returned when appropriate.
+    """
+    name = sys.implementation.name
+    return INTERPRETER_SHORT_NAMES.get(name) or name
+
+
+def interpreter_version(*, warn: bool = False) -> str:
+    """
+    Returns the version of the running interpreter.
+    """
+    version = _get_config_var("py_version_nodot", warn=warn)
+    if version:
+        version = str(version)
+    else:
+        version = _version_nodot(sys.version_info[:2])
+    return version
+
+
+def _version_nodot(version: PythonVersion) -> str:
+    return "".join(map(str, version))
+
+
+def sys_tags(*, warn: bool = False) -> Iterator[Tag]:
+    """
+    Returns the sequence of tag triples for the running interpreter.
+
+    The order of the sequence corresponds to priority order for the
+    interpreter, from most to least important.
+    """
+
+    interp_name = interpreter_name()
+    if interp_name == "cp":
+        yield from cpython_tags(warn=warn)
+    else:
+        yield from generic_tags()
+
+    if interp_name == "pp":
+        interp = "pp3"
+    elif interp_name == "cp":
+        interp = "cp" + interpreter_version(warn=warn)
+    else:
+        interp = None
+    yield from compatible_tags(interpreter=interp)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/utils.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/utils.py
new file mode 100644
index 00000000..23450953
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/utils.py
@@ -0,0 +1,163 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from __future__ import annotations
+
+import functools
+import re
+from typing import NewType, Tuple, Union, cast
+
+from .tags import Tag, parse_tag
+from .version import InvalidVersion, Version, _TrimmedRelease
+
+BuildTag = Union[Tuple[()], Tuple[int, str]]
+NormalizedName = NewType("NormalizedName", str)
+
+
+class InvalidName(ValueError):
+    """
+    An invalid distribution name; users should refer to the packaging user guide.
+    """
+
+
+class InvalidWheelFilename(ValueError):
+    """
+    An invalid wheel filename was found, users should refer to PEP 427.
+    """
+
+
+class InvalidSdistFilename(ValueError):
+    """
+    An invalid sdist filename was found, users should refer to the packaging user guide.
+    """
+
+
+# Core metadata spec for `Name`
+_validate_regex = re.compile(
+    r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.IGNORECASE
+)
+_canonicalize_regex = re.compile(r"[-_.]+")
+_normalized_regex = re.compile(r"^([a-z0-9]|[a-z0-9]([a-z0-9-](?!--))*[a-z0-9])$")
+# PEP 427: The build number must start with a digit.
+_build_tag_regex = re.compile(r"(\d+)(.*)")
+
+
+def canonicalize_name(name: str, *, validate: bool = False) -> NormalizedName:
+    if validate and not _validate_regex.match(name):
+        raise InvalidName(f"name is invalid: {name!r}")
+    # This is taken from PEP 503.
+    value = _canonicalize_regex.sub("-", name).lower()
+    return cast(NormalizedName, value)
+
+
+def is_normalized_name(name: str) -> bool:
+    return _normalized_regex.match(name) is not None
+
+
+@functools.singledispatch
+def canonicalize_version(
+    version: Version | str, *, strip_trailing_zero: bool = True
+) -> str:
+    """
+    Return a canonical form of a version as a string.
+
+    >>> canonicalize_version('1.0.1')
+    '1.0.1'
+
+    Per PEP 625, versions may have multiple canonical forms, differing
+    only by trailing zeros.
+
+    >>> canonicalize_version('1.0.0')
+    '1'
+    >>> canonicalize_version('1.0.0', strip_trailing_zero=False)
+    '1.0.0'
+
+    Invalid versions are returned unaltered.
+
+    >>> canonicalize_version('foo bar baz')
+    'foo bar baz'
+    """
+    return str(_TrimmedRelease(str(version)) if strip_trailing_zero else version)
+
+
+@canonicalize_version.register
+def _(version: str, *, strip_trailing_zero: bool = True) -> str:
+    try:
+        parsed = Version(version)
+    except InvalidVersion:
+        # Legacy versions cannot be normalized
+        return version
+    return canonicalize_version(parsed, strip_trailing_zero=strip_trailing_zero)
+
+
+def parse_wheel_filename(
+    filename: str,
+) -> tuple[NormalizedName, Version, BuildTag, frozenset[Tag]]:
+    if not filename.endswith(".whl"):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (extension must be '.whl'): {filename!r}"
+        )
+
+    filename = filename[:-4]
+    dashes = filename.count("-")
+    if dashes not in (4, 5):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (wrong number of parts): {filename!r}"
+        )
+
+    parts = filename.split("-", dashes - 2)
+    name_part = parts[0]
+    # See PEP 427 for the rules on escaping the project name.
+    if "__" in name_part or re.match(r"^[\w\d._]*$", name_part, re.UNICODE) is None:
+        raise InvalidWheelFilename(f"Invalid project name: {filename!r}")
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(parts[1])
+    except InvalidVersion as e:
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (invalid version): {filename!r}"
+        ) from e
+
+    if dashes == 5:
+        build_part = parts[2]
+        build_match = _build_tag_regex.match(build_part)
+        if build_match is None:
+            raise InvalidWheelFilename(
+                f"Invalid build number: {build_part} in {filename!r}"
+            )
+        build = cast(BuildTag, (int(build_match.group(1)), build_match.group(2)))
+    else:
+        build = ()
+    tags = parse_tag(parts[-1])
+    return (name, version, build, tags)
+
+
+def parse_sdist_filename(filename: str) -> tuple[NormalizedName, Version]:
+    if filename.endswith(".tar.gz"):
+        file_stem = filename[: -len(".tar.gz")]
+    elif filename.endswith(".zip"):
+        file_stem = filename[: -len(".zip")]
+    else:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (extension must be '.tar.gz' or '.zip'):"
+            f" {filename!r}"
+        )
+
+    # We are requiring a PEP 440 version, which cannot contain dashes,
+    # so we split on the last dash.
+    name_part, sep, version_part = file_stem.rpartition("-")
+    if not sep:
+        raise InvalidSdistFilename(f"Invalid sdist filename: {filename!r}")
+
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(version_part)
+    except InvalidVersion as e:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (invalid version): {filename!r}"
+        ) from e
+
+    return (name, version)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/version.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/version.py
new file mode 100644
index 00000000..c9bbda20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/packaging/version.py
@@ -0,0 +1,582 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.version import parse, Version
+"""
+
+from __future__ import annotations
+
+import itertools
+import re
+from typing import Any, Callable, NamedTuple, SupportsInt, Tuple, Union
+
+from ._structures import Infinity, InfinityType, NegativeInfinity, NegativeInfinityType
+
+__all__ = ["VERSION_PATTERN", "InvalidVersion", "Version", "parse"]
+
+LocalType = Tuple[Union[int, str], ...]
+
+CmpPrePostDevType = Union[InfinityType, NegativeInfinityType, Tuple[str, int]]
+CmpLocalType = Union[
+    NegativeInfinityType,
+    Tuple[Union[Tuple[int, str], Tuple[NegativeInfinityType, Union[int, str]]], ...],
+]
+CmpKey = Tuple[
+    int,
+    Tuple[int, ...],
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpLocalType,
+]
+VersionComparisonMethod = Callable[[CmpKey, CmpKey], bool]
+
+
+class _Version(NamedTuple):
+    epoch: int
+    release: tuple[int, ...]
+    dev: tuple[str, int] | None
+    pre: tuple[str, int] | None
+    post: tuple[str, int] | None
+    local: LocalType | None
+
+
+def parse(version: str) -> Version:
+    """Parse the given version string.
+
+    >>> parse('1.0.dev1')
+    <Version('1.0.dev1')>
+
+    :param version: The version string to parse.
+    :raises InvalidVersion: When the version string is not a valid version.
+    """
+    return Version(version)
+
+
+class InvalidVersion(ValueError):
+    """Raised when a version string is not a valid version.
+
+    >>> Version("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.version.InvalidVersion: Invalid version: 'invalid'
+    """
+
+
+class _BaseVersion:
+    _key: tuple[Any, ...]
+
+    def __hash__(self) -> int:
+        return hash(self._key)
+
+    # Please keep the duplicated `isinstance` check
+    # in the six comparisons hereunder
+    # unless you find a way to avoid adding overhead function calls.
+    def __lt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key < other._key
+
+    def __le__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key <= other._key
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key == other._key
+
+    def __ge__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key >= other._key
+
+    def __gt__(self, other: _BaseVersion) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key > other._key
+
+    def __ne__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key != other._key
+
+
+# Deliberately not anchored to the start and end of the string, to make it
+# easier for 3rd party code to reuse
+_VERSION_PATTERN = r"""
+    v?
+    (?:
+        (?:(?P<epoch>[0-9]+)!)?                           # epoch
+        (?P<release>[0-9]+(?:\.[0-9]+)*)                  # release segment
+        (?P<pre>                                          # pre-release
+            [-_\.]?
+            (?P<pre_l>alpha|a|beta|b|preview|pre|c|rc)
+            [-_\.]?
+            (?P<pre_n>[0-9]+)?
+        )?
+        (?P<post>                                         # post release
+            (?:-(?P<post_n1>[0-9]+))
+            |
+            (?:
+                [-_\.]?
+                (?P<post_l>post|rev|r)
+                [-_\.]?
+                (?P<post_n2>[0-9]+)?
+            )
+        )?
+        (?P<dev>                                          # dev release
+            [-_\.]?
+            (?P<dev_l>dev)
+            [-_\.]?
+            (?P<dev_n>[0-9]+)?
+        )?
+    )
+    (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
+"""
+
+VERSION_PATTERN = _VERSION_PATTERN
+"""
+A string containing the regular expression used to match a valid version.
+
+The pattern is not anchored at either end, and is intended for embedding in larger
+expressions (for example, matching a version number as part of a file name). The
+regular expression should be compiled with the ``re.VERBOSE`` and ``re.IGNORECASE``
+flags set.
+
+:meta hide-value:
+"""
+
+
+class Version(_BaseVersion):
+    """This class abstracts handling of a project's versions.
+
+    A :class:`Version` instance is comparison aware and can be compared and
+    sorted using the standard Python interfaces.
+
+    >>> v1 = Version("1.0a5")
+    >>> v2 = Version("1.0")
+    >>> v1
+    <Version('1.0a5')>
+    >>> v2
+    <Version('1.0')>
+    >>> v1 < v2
+    True
+    >>> v1 == v2
+    False
+    >>> v1 > v2
+    False
+    >>> v1 >= v2
+    False
+    >>> v1 <= v2
+    True
+    """
+
+    _regex = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.VERBOSE | re.IGNORECASE)
+    _key: CmpKey
+
+    def __init__(self, version: str) -> None:
+        """Initialize a Version object.
+
+        :param version:
+            The string representation of a version which will be parsed and normalized
+            before use.
+        :raises InvalidVersion:
+            If the ``version`` does not conform to PEP 440 in any way then this
+            exception will be raised.
+        """
+
+        # Validate the version and parse it into pieces
+        match = self._regex.search(version)
+        if not match:
+            raise InvalidVersion(f"Invalid version: {version!r}")
+
+        # Store the parsed out pieces of the version
+        self._version = _Version(
+            epoch=int(match.group("epoch")) if match.group("epoch") else 0,
+            release=tuple(int(i) for i in match.group("release").split(".")),
+            pre=_parse_letter_version(match.group("pre_l"), match.group("pre_n")),
+            post=_parse_letter_version(
+                match.group("post_l"), match.group("post_n1") or match.group("post_n2")
+            ),
+            dev=_parse_letter_version(match.group("dev_l"), match.group("dev_n")),
+            local=_parse_local_version(match.group("local")),
+        )
+
+        # Generate a key which will be used for sorting
+        self._key = _cmpkey(
+            self._version.epoch,
+            self._version.release,
+            self._version.pre,
+            self._version.post,
+            self._version.dev,
+            self._version.local,
+        )
+
+    def __repr__(self) -> str:
+        """A representation of the Version that shows all internal state.
+
+        >>> Version('1.0.0')
+        <Version('1.0.0')>
+        """
+        return f"<Version('{self}')>"
+
+    def __str__(self) -> str:
+        """A string representation of the version that can be round-tripped.
+
+        >>> str(Version("1.0a5"))
+        '1.0a5'
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        # Pre-release
+        if self.pre is not None:
+            parts.append("".join(str(x) for x in self.pre))
+
+        # Post-release
+        if self.post is not None:
+            parts.append(f".post{self.post}")
+
+        # Development release
+        if self.dev is not None:
+            parts.append(f".dev{self.dev}")
+
+        # Local version segment
+        if self.local is not None:
+            parts.append(f"+{self.local}")
+
+        return "".join(parts)
+
+    @property
+    def epoch(self) -> int:
+        """The epoch of the version.
+
+        >>> Version("2.0.0").epoch
+        0
+        >>> Version("1!2.0.0").epoch
+        1
+        """
+        return self._version.epoch
+
+    @property
+    def release(self) -> tuple[int, ...]:
+        """The components of the "release" segment of the version.
+
+        >>> Version("1.2.3").release
+        (1, 2, 3)
+        >>> Version("2.0.0").release
+        (2, 0, 0)
+        >>> Version("1!2.0.0.post0").release
+        (2, 0, 0)
+
+        Includes trailing zeroes but not the epoch or any pre-release / development /
+        post-release suffixes.
+        """
+        return self._version.release
+
+    @property
+    def pre(self) -> tuple[str, int] | None:
+        """The pre-release segment of the version.
+
+        >>> print(Version("1.2.3").pre)
+        None
+        >>> Version("1.2.3a1").pre
+        ('a', 1)
+        >>> Version("1.2.3b1").pre
+        ('b', 1)
+        >>> Version("1.2.3rc1").pre
+        ('rc', 1)
+        """
+        return self._version.pre
+
+    @property
+    def post(self) -> int | None:
+        """The post-release number of the version.
+
+        >>> print(Version("1.2.3").post)
+        None
+        >>> Version("1.2.3.post1").post
+        1
+        """
+        return self._version.post[1] if self._version.post else None
+
+    @property
+    def dev(self) -> int | None:
+        """The development number of the version.
+
+        >>> print(Version("1.2.3").dev)
+        None
+        >>> Version("1.2.3.dev1").dev
+        1
+        """
+        return self._version.dev[1] if self._version.dev else None
+
+    @property
+    def local(self) -> str | None:
+        """The local version segment of the version.
+
+        >>> print(Version("1.2.3").local)
+        None
+        >>> Version("1.2.3+abc").local
+        'abc'
+        """
+        if self._version.local:
+            return ".".join(str(x) for x in self._version.local)
+        else:
+            return None
+
+    @property
+    def public(self) -> str:
+        """The public portion of the version.
+
+        >>> Version("1.2.3").public
+        '1.2.3'
+        >>> Version("1.2.3+abc").public
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").public
+        '1!1.2.3.dev1'
+        """
+        return str(self).split("+", 1)[0]
+
+    @property
+    def base_version(self) -> str:
+        """The "base version" of the version.
+
+        >>> Version("1.2.3").base_version
+        '1.2.3'
+        >>> Version("1.2.3+abc").base_version
+        '1.2.3'
+        >>> Version("1!1.2.3dev1+abc").base_version
+        '1!1.2.3'
+
+        The "base version" is the public version of the project without any pre or post
+        release markers.
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        return "".join(parts)
+
+    @property
+    def is_prerelease(self) -> bool:
+        """Whether this version is a pre-release.
+
+        >>> Version("1.2.3").is_prerelease
+        False
+        >>> Version("1.2.3a1").is_prerelease
+        True
+        >>> Version("1.2.3b1").is_prerelease
+        True
+        >>> Version("1.2.3rc1").is_prerelease
+        True
+        >>> Version("1.2.3dev1").is_prerelease
+        True
+        """
+        return self.dev is not None or self.pre is not None
+
+    @property
+    def is_postrelease(self) -> bool:
+        """Whether this version is a post-release.
+
+        >>> Version("1.2.3").is_postrelease
+        False
+        >>> Version("1.2.3.post1").is_postrelease
+        True
+        """
+        return self.post is not None
+
+    @property
+    def is_devrelease(self) -> bool:
+        """Whether this version is a development release.
+
+        >>> Version("1.2.3").is_devrelease
+        False
+        >>> Version("1.2.3.dev1").is_devrelease
+        True
+        """
+        return self.dev is not None
+
+    @property
+    def major(self) -> int:
+        """The first item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").major
+        1
+        """
+        return self.release[0] if len(self.release) >= 1 else 0
+
+    @property
+    def minor(self) -> int:
+        """The second item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").minor
+        2
+        >>> Version("1").minor
+        0
+        """
+        return self.release[1] if len(self.release) >= 2 else 0
+
+    @property
+    def micro(self) -> int:
+        """The third item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").micro
+        3
+        >>> Version("1").micro
+        0
+        """
+        return self.release[2] if len(self.release) >= 3 else 0
+
+
+class _TrimmedRelease(Version):
+    @property
+    def release(self) -> tuple[int, ...]:
+        """
+        Release segment without any trailing zeros.
+
+        >>> _TrimmedRelease('1.0.0').release
+        (1,)
+        >>> _TrimmedRelease('0.0').release
+        (0,)
+        """
+        rel = super().release
+        nonzeros = (index for index, val in enumerate(rel) if val)
+        last_nonzero = max(nonzeros, default=0)
+        return rel[: last_nonzero + 1]
+
+
+def _parse_letter_version(
+    letter: str | None, number: str | bytes | SupportsInt | None
+) -> tuple[str, int] | None:
+    if letter:
+        # We consider there to be an implicit 0 in a pre-release if there is
+        # not a numeral associated with it.
+        if number is None:
+            number = 0
+
+        # We normalize any letters to their lower case form
+        letter = letter.lower()
+
+        # We consider some words to be alternate spellings of other words and
+        # in those cases we want to normalize the spellings to our preferred
+        # spelling.
+        if letter == "alpha":
+            letter = "a"
+        elif letter == "beta":
+            letter = "b"
+        elif letter in ["c", "pre", "preview"]:
+            letter = "rc"
+        elif letter in ["rev", "r"]:
+            letter = "post"
+
+        return letter, int(number)
+
+    assert not letter
+    if number:
+        # We assume if we are given a number, but we are not given a letter
+        # then this is using the implicit post release syntax (e.g. 1.0-1)
+        letter = "post"
+
+        return letter, int(number)
+
+    return None
+
+
+_local_version_separators = re.compile(r"[\._-]")
+
+
+def _parse_local_version(local: str | None) -> LocalType | None:
+    """
+    Takes a string like abc.1.twelve and turns it into ("abc", 1, "twelve").
+    """
+    if local is not None:
+        return tuple(
+            part.lower() if not part.isdigit() else int(part)
+            for part in _local_version_separators.split(local)
+        )
+    return None
+
+
+def _cmpkey(
+    epoch: int,
+    release: tuple[int, ...],
+    pre: tuple[str, int] | None,
+    post: tuple[str, int] | None,
+    dev: tuple[str, int] | None,
+    local: LocalType | None,
+) -> CmpKey:
+    # When we compare a release version, we want to compare it with all of the
+    # trailing zeros removed. So we'll use a reverse the list, drop all the now
+    # leading zeros until we come to something non zero, then take the rest
+    # re-reverse it back into the correct order and make it a tuple and use
+    # that for our sorting key.
+    _release = tuple(
+        reversed(list(itertools.dropwhile(lambda x: x == 0, reversed(release))))
+    )
+
+    # We need to "trick" the sorting algorithm to put 1.0.dev0 before 1.0a0.
+    # We'll do this by abusing the pre segment, but we _only_ want to do this
+    # if there is not a pre or a post segment. If we have one of those then
+    # the normal sorting rules will handle this case correctly.
+    if pre is None and post is None and dev is not None:
+        _pre: CmpPrePostDevType = NegativeInfinity
+    # Versions without a pre-release (except as noted above) should sort after
+    # those with one.
+    elif pre is None:
+        _pre = Infinity
+    else:
+        _pre = pre
+
+    # Versions without a post segment should sort before those with one.
+    if post is None:
+        _post: CmpPrePostDevType = NegativeInfinity
+
+    else:
+        _post = post
+
+    # Versions without a development segment should sort after those with one.
+    if dev is None:
+        _dev: CmpPrePostDevType = Infinity
+
+    else:
+        _dev = dev
+
+    if local is None:
+        # Versions without a local segment should sort before those with one.
+        _local: CmpLocalType = NegativeInfinity
+    else:
+        # Versions with a local segment need that segment parsed to implement
+        # the sorting rules in PEP440.
+        # - Alpha numeric segments sort before numeric segments
+        # - Alpha numeric segments sort lexicographically
+        # - Numeric segments sort numerically
+        # - Shorter versions sort before longer versions when the prefixes
+        #   match exactly
+        _local = tuple(
+            (i, "") if isinstance(i, int) else (NegativeInfinity, i) for i in local
+        )
+
+    return epoch, _release, _pre, _post, _dev, _local
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/METADATA
new file mode 100644
index 00000000..ab51ef36
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/METADATA
@@ -0,0 +1,319 @@
+Metadata-Version: 2.3
+Name: platformdirs
+Version: 4.2.2
+Summary: A small Python package for determining appropriate platform-specific dirs, e.g. a `user data dir`.
+Project-URL: Documentation, https://platformdirs.readthedocs.io
+Project-URL: Homepage, https://github.com/platformdirs/platformdirs
+Project-URL: Source, https://github.com/platformdirs/platformdirs
+Project-URL: Tracker, https://github.com/platformdirs/platformdirs/issues
+Maintainer-email: Bernát Gábor <gaborjbernat@gmail.com>, Julian Berman <Julian@GrayVines.com>, Ofek Lev <oss@ofek.dev>, Ronny Pfannschmidt <opensource@ronnypfannschmidt.de>
+License-Expression: MIT
+License-File: LICENSE
+Keywords: appdirs,application,cache,directory,log,user
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.8
+Provides-Extra: docs
+Requires-Dist: furo>=2023.9.10; extra == 'docs'
+Requires-Dist: proselint>=0.13; extra == 'docs'
+Requires-Dist: sphinx-autodoc-typehints>=1.25.2; extra == 'docs'
+Requires-Dist: sphinx>=7.2.6; extra == 'docs'
+Provides-Extra: test
+Requires-Dist: appdirs==1.4.4; extra == 'test'
+Requires-Dist: covdefaults>=2.3; extra == 'test'
+Requires-Dist: pytest-cov>=4.1; extra == 'test'
+Requires-Dist: pytest-mock>=3.12; extra == 'test'
+Requires-Dist: pytest>=7.4.3; extra == 'test'
+Provides-Extra: type
+Requires-Dist: mypy>=1.8; extra == 'type'
+Description-Content-Type: text/x-rst
+
+The problem
+===========
+
+.. image:: https://github.com/platformdirs/platformdirs/actions/workflows/check.yml/badge.svg
+   :target: https://github.com/platformdirs/platformdirs/actions
+
+When writing desktop application, finding the right location to store user data
+and configuration varies per platform. Even for single-platform apps, there
+may by plenty of nuances in figuring out the right location.
+
+For example, if running on macOS, you should use::
+
+    ~/Library/Application Support/<AppName>
+
+If on Windows (at least English Win) that should be::
+
+    C:\Documents and Settings\<User>\Application Data\Local Settings\<AppAuthor>\<AppName>
+
+or possibly::
+
+    C:\Documents and Settings\<User>\Application Data\<AppAuthor>\<AppName>
+
+for `roaming profiles <https://docs.microsoft.com/en-us/previous-versions/windows/it-pro/windows-vista/cc766489(v=ws.10)>`_ but that is another story.
+
+On Linux (and other Unices), according to the `XDG Basedir Spec`_, it should be::
+
+    ~/.local/share/<AppName>
+
+.. _XDG Basedir Spec: https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html
+
+``platformdirs`` to the rescue
+==============================
+
+This kind of thing is what the ``platformdirs`` package is for.
+``platformdirs`` will help you choose an appropriate:
+
+- user data dir (``user_data_dir``)
+- user config dir (``user_config_dir``)
+- user cache dir (``user_cache_dir``)
+- site data dir (``site_data_dir``)
+- site config dir (``site_config_dir``)
+- user log dir (``user_log_dir``)
+- user documents dir (``user_documents_dir``)
+- user downloads dir (``user_downloads_dir``)
+- user pictures dir (``user_pictures_dir``)
+- user videos dir (``user_videos_dir``)
+- user music dir (``user_music_dir``)
+- user desktop dir (``user_desktop_dir``)
+- user runtime dir (``user_runtime_dir``)
+
+And also:
+
+- Is slightly opinionated on the directory names used. Look for "OPINION" in
+  documentation and code for when an opinion is being applied.
+
+Example output
+==============
+
+On macOS:
+
+.. code-block:: pycon
+
+    >>> from platformdirs import *
+    >>> appname = "SuperApp"
+    >>> appauthor = "Acme"
+    >>> user_data_dir(appname, appauthor)
+    '/Users/trentm/Library/Application Support/SuperApp'
+    >>> site_data_dir(appname, appauthor)
+    '/Library/Application Support/SuperApp'
+    >>> user_cache_dir(appname, appauthor)
+    '/Users/trentm/Library/Caches/SuperApp'
+    >>> user_log_dir(appname, appauthor)
+    '/Users/trentm/Library/Logs/SuperApp'
+    >>> user_documents_dir()
+    '/Users/trentm/Documents'
+    >>> user_downloads_dir()
+    '/Users/trentm/Downloads'
+    >>> user_pictures_dir()
+    '/Users/trentm/Pictures'
+    >>> user_videos_dir()
+    '/Users/trentm/Movies'
+    >>> user_music_dir()
+    '/Users/trentm/Music'
+    >>> user_desktop_dir()
+    '/Users/trentm/Desktop'
+    >>> user_runtime_dir(appname, appauthor)
+    '/Users/trentm/Library/Caches/TemporaryItems/SuperApp'
+
+On Windows:
+
+.. code-block:: pycon
+
+    >>> from platformdirs import *
+    >>> appname = "SuperApp"
+    >>> appauthor = "Acme"
+    >>> user_data_dir(appname, appauthor)
+    'C:\\Users\\trentm\\AppData\\Local\\Acme\\SuperApp'
+    >>> user_data_dir(appname, appauthor, roaming=True)
+    'C:\\Users\\trentm\\AppData\\Roaming\\Acme\\SuperApp'
+    >>> user_cache_dir(appname, appauthor)
+    'C:\\Users\\trentm\\AppData\\Local\\Acme\\SuperApp\\Cache'
+    >>> user_log_dir(appname, appauthor)
+    'C:\\Users\\trentm\\AppData\\Local\\Acme\\SuperApp\\Logs'
+    >>> user_documents_dir()
+    'C:\\Users\\trentm\\Documents'
+    >>> user_downloads_dir()
+    'C:\\Users\\trentm\\Downloads'
+    >>> user_pictures_dir()
+    'C:\\Users\\trentm\\Pictures'
+    >>> user_videos_dir()
+    'C:\\Users\\trentm\\Videos'
+    >>> user_music_dir()
+    'C:\\Users\\trentm\\Music'
+    >>> user_desktop_dir()
+    'C:\\Users\\trentm\\Desktop'
+    >>> user_runtime_dir(appname, appauthor)
+    'C:\\Users\\trentm\\AppData\\Local\\Temp\\Acme\\SuperApp'
+
+On Linux:
+
+.. code-block:: pycon
+
+    >>> from platformdirs import *
+    >>> appname = "SuperApp"
+    >>> appauthor = "Acme"
+    >>> user_data_dir(appname, appauthor)
+    '/home/trentm/.local/share/SuperApp'
+    >>> site_data_dir(appname, appauthor)
+    '/usr/local/share/SuperApp'
+    >>> site_data_dir(appname, appauthor, multipath=True)
+    '/usr/local/share/SuperApp:/usr/share/SuperApp'
+    >>> user_cache_dir(appname, appauthor)
+    '/home/trentm/.cache/SuperApp'
+    >>> user_log_dir(appname, appauthor)
+    '/home/trentm/.local/state/SuperApp/log'
+    >>> user_config_dir(appname)
+    '/home/trentm/.config/SuperApp'
+    >>> user_documents_dir()
+    '/home/trentm/Documents'
+    >>> user_downloads_dir()
+    '/home/trentm/Downloads'
+    >>> user_pictures_dir()
+    '/home/trentm/Pictures'
+    >>> user_videos_dir()
+    '/home/trentm/Videos'
+    >>> user_music_dir()
+    '/home/trentm/Music'
+    >>> user_desktop_dir()
+    '/home/trentm/Desktop'
+    >>> user_runtime_dir(appname, appauthor)
+    '/run/user/{os.getuid()}/SuperApp'
+    >>> site_config_dir(appname)
+    '/etc/xdg/SuperApp'
+    >>> os.environ["XDG_CONFIG_DIRS"] = "/etc:/usr/local/etc"
+    >>> site_config_dir(appname, multipath=True)
+    '/etc/SuperApp:/usr/local/etc/SuperApp'
+
+On Android::
+
+    >>> from platformdirs import *
+    >>> appname = "SuperApp"
+    >>> appauthor = "Acme"
+    >>> user_data_dir(appname, appauthor)
+    '/data/data/com.myApp/files/SuperApp'
+    >>> user_cache_dir(appname, appauthor)
+    '/data/data/com.myApp/cache/SuperApp'
+    >>> user_log_dir(appname, appauthor)
+    '/data/data/com.myApp/cache/SuperApp/log'
+    >>> user_config_dir(appname)
+    '/data/data/com.myApp/shared_prefs/SuperApp'
+    >>> user_documents_dir()
+    '/storage/emulated/0/Documents'
+    >>> user_downloads_dir()
+    '/storage/emulated/0/Downloads'
+    >>> user_pictures_dir()
+    '/storage/emulated/0/Pictures'
+    >>> user_videos_dir()
+    '/storage/emulated/0/DCIM/Camera'
+    >>> user_music_dir()
+    '/storage/emulated/0/Music'
+    >>> user_desktop_dir()
+    '/storage/emulated/0/Desktop'
+    >>> user_runtime_dir(appname, appauthor)
+    '/data/data/com.myApp/cache/SuperApp/tmp'
+
+Note: Some android apps like Termux and Pydroid are used as shells. These
+apps are used by the end user to emulate Linux environment. Presence of
+``SHELL`` environment variable is used by Platformdirs to differentiate
+between general android apps and android apps used as shells. Shell android
+apps also support ``XDG_*`` environment variables.
+
+
+``PlatformDirs`` for convenience
+================================
+
+.. code-block:: pycon
+
+    >>> from platformdirs import PlatformDirs
+    >>> dirs = PlatformDirs("SuperApp", "Acme")
+    >>> dirs.user_data_dir
+    '/Users/trentm/Library/Application Support/SuperApp'
+    >>> dirs.site_data_dir
+    '/Library/Application Support/SuperApp'
+    >>> dirs.user_cache_dir
+    '/Users/trentm/Library/Caches/SuperApp'
+    >>> dirs.user_log_dir
+    '/Users/trentm/Library/Logs/SuperApp'
+    >>> dirs.user_documents_dir
+    '/Users/trentm/Documents'
+    >>> dirs.user_downloads_dir
+    '/Users/trentm/Downloads'
+    >>> dirs.user_pictures_dir
+    '/Users/trentm/Pictures'
+    >>> dirs.user_videos_dir
+    '/Users/trentm/Movies'
+    >>> dirs.user_music_dir
+    '/Users/trentm/Music'
+    >>> dirs.user_desktop_dir
+    '/Users/trentm/Desktop'
+    >>> dirs.user_runtime_dir
+    '/Users/trentm/Library/Caches/TemporaryItems/SuperApp'
+
+Per-version isolation
+=====================
+
+If you have multiple versions of your app in use that you want to be
+able to run side-by-side, then you may want version-isolation for these
+dirs::
+
+    >>> from platformdirs import PlatformDirs
+    >>> dirs = PlatformDirs("SuperApp", "Acme", version="1.0")
+    >>> dirs.user_data_dir
+    '/Users/trentm/Library/Application Support/SuperApp/1.0'
+    >>> dirs.site_data_dir
+    '/Library/Application Support/SuperApp/1.0'
+    >>> dirs.user_cache_dir
+    '/Users/trentm/Library/Caches/SuperApp/1.0'
+    >>> dirs.user_log_dir
+    '/Users/trentm/Library/Logs/SuperApp/1.0'
+    >>> dirs.user_documents_dir
+    '/Users/trentm/Documents'
+    >>> dirs.user_downloads_dir
+    '/Users/trentm/Downloads'
+    >>> dirs.user_pictures_dir
+    '/Users/trentm/Pictures'
+    >>> dirs.user_videos_dir
+    '/Users/trentm/Movies'
+    >>> dirs.user_music_dir
+    '/Users/trentm/Music'
+    >>> dirs.user_desktop_dir
+    '/Users/trentm/Desktop'
+    >>> dirs.user_runtime_dir
+    '/Users/trentm/Library/Caches/TemporaryItems/SuperApp/1.0'
+
+Be wary of using this for configuration files though; you'll need to handle
+migrating configuration files manually.
+
+Why this Fork?
+==============
+
+This repository is a friendly fork of the wonderful work started by
+`ActiveState <https://github.com/ActiveState/appdirs>`_ who created
+``appdirs``, this package's ancestor.
+
+Maintaining an open source project is no easy task, particularly
+from within an organization, and the Python community is indebted
+to ``appdirs`` (and to Trent Mick and Jeff Rouse in particular) for
+creating an incredibly useful simple module, as evidenced by the wide
+number of users it has attracted over the years.
+
+Nonetheless, given the number of long-standing open issues
+and pull requests, and no clear path towards `ensuring
+that maintenance of the package would continue or grow
+<https://github.com/ActiveState/appdirs/issues/79>`_, this fork was
+created.
+
+Contributions are most welcome.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/RECORD
new file mode 100644
index 00000000..64c0c8ea
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/RECORD
@@ -0,0 +1,23 @@
+platformdirs-4.2.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+platformdirs-4.2.2.dist-info/METADATA,sha256=zmsie01G1MtXR0wgIv5XpVeTO7idr0WWvfmxKsKWuGk,11429
+platformdirs-4.2.2.dist-info/RECORD,,
+platformdirs-4.2.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+platformdirs-4.2.2.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
+platformdirs-4.2.2.dist-info/licenses/LICENSE,sha256=KeD9YukphQ6G6yjD_czwzv30-pSHkBHP-z0NS-1tTbY,1089
+platformdirs/__init__.py,sha256=EMGE8qeHRR9CzDFr8kL3tA8hdZZniYjXBVZd0UGTWK0,22225
+platformdirs/__main__.py,sha256=HnsUQHpiBaiTxwcmwVw-nFaPdVNZtQIdi1eWDtI-MzI,1493
+platformdirs/__pycache__/__init__.cpython-312.pyc,,
+platformdirs/__pycache__/__main__.cpython-312.pyc,,
+platformdirs/__pycache__/android.cpython-312.pyc,,
+platformdirs/__pycache__/api.cpython-312.pyc,,
+platformdirs/__pycache__/macos.cpython-312.pyc,,
+platformdirs/__pycache__/unix.cpython-312.pyc,,
+platformdirs/__pycache__/version.cpython-312.pyc,,
+platformdirs/__pycache__/windows.cpython-312.pyc,,
+platformdirs/android.py,sha256=xZXY9Jd46WOsxT2U6-5HsNtDZ-IQqxcEUrBLl3hYk4o,9016
+platformdirs/api.py,sha256=QBYdUac2eC521ek_y53uD1Dcq-lJX8IgSRVd4InC6uc,8996
+platformdirs/macos.py,sha256=wftsbsvq6nZ0WORXSiCrZNkRHz_WKuktl0a6mC7MFkI,5580
+platformdirs/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+platformdirs/unix.py,sha256=Cci9Wqt35dAMsg6HT9nRGHSBW5obb0pR3AE1JJnsCXg,10643
+platformdirs/version.py,sha256=r7F76tZRjgQKzrpx_I0_ZMQOMU-PS7eGnHD7zEK3KB0,411
+platformdirs/windows.py,sha256=IFpiohUBwxPtCzlyKwNtxyW4Jk8haa6W8o59mfrDXVo,10125
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/WHEEL
new file mode 100644
index 00000000..516596c7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.24.2
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/licenses/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/licenses/LICENSE
new file mode 100644
index 00000000..f35fed91
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs-4.2.2.dist-info/licenses/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2010-202x The platformdirs developers
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__init__.py
new file mode 100644
index 00000000..3f7d9490
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__init__.py
@@ -0,0 +1,627 @@
+"""
+Utilities for determining application-specific dirs.
+
+See <https://github.com/platformdirs/platformdirs> for details and usage.
+
+"""
+
+from __future__ import annotations
+
+import os
+import sys
+from typing import TYPE_CHECKING
+
+from .api import PlatformDirsABC
+from .version import __version__
+from .version import __version_tuple__ as __version_info__
+
+if TYPE_CHECKING:
+    from pathlib import Path
+    from typing import Literal
+
+
+def _set_platform_dir_class() -> type[PlatformDirsABC]:
+    if sys.platform == "win32":
+        from platformdirs.windows import Windows as Result  # noqa: PLC0415
+    elif sys.platform == "darwin":
+        from platformdirs.macos import MacOS as Result  # noqa: PLC0415
+    else:
+        from platformdirs.unix import Unix as Result  # noqa: PLC0415
+
+    if os.getenv("ANDROID_DATA") == "/data" and os.getenv("ANDROID_ROOT") == "/system":
+        if os.getenv("SHELL") or os.getenv("PREFIX"):
+            return Result
+
+        from platformdirs.android import _android_folder  # noqa: PLC0415
+
+        if _android_folder() is not None:
+            from platformdirs.android import Android  # noqa: PLC0415
+
+            return Android  # return to avoid redefinition of a result
+
+    return Result
+
+
+PlatformDirs = _set_platform_dir_class()  #: Currently active platform
+AppDirs = PlatformDirs  #: Backwards compatibility with appdirs
+
+
+def user_data_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_data_dir
+
+
+def site_data_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data directory shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_data_dir
+
+
+def user_config_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_config_dir
+
+
+def site_config_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config directory shared by the users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_config_dir
+
+
+def user_cache_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_cache_dir
+
+
+def site_cache_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_cache_dir
+
+
+def user_state_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: state directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_state_dir
+
+
+def user_log_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: log directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_log_dir
+
+
+def user_documents_dir() -> str:
+    """:returns: documents directory tied to the user"""
+    return PlatformDirs().user_documents_dir
+
+
+def user_downloads_dir() -> str:
+    """:returns: downloads directory tied to the user"""
+    return PlatformDirs().user_downloads_dir
+
+
+def user_pictures_dir() -> str:
+    """:returns: pictures directory tied to the user"""
+    return PlatformDirs().user_pictures_dir
+
+
+def user_videos_dir() -> str:
+    """:returns: videos directory tied to the user"""
+    return PlatformDirs().user_videos_dir
+
+
+def user_music_dir() -> str:
+    """:returns: music directory tied to the user"""
+    return PlatformDirs().user_music_dir
+
+
+def user_desktop_dir() -> str:
+    """:returns: desktop directory tied to the user"""
+    return PlatformDirs().user_desktop_dir
+
+
+def user_runtime_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_runtime_dir
+
+
+def site_runtime_dir(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> str:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime directory shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_runtime_dir
+
+
+def user_data_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_data_path
+
+
+def site_data_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `multipath <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: data path shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_data_path
+
+
+def user_config_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_config_path
+
+
+def site_config_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    multipath: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param multipath: See `roaming <platformdirs.api.PlatformDirsABC.multipath>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: config path shared by the users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        multipath=multipath,
+        ensure_exists=ensure_exists,
+    ).site_config_path
+
+
+def site_cache_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache directory tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_cache_path
+
+
+def user_cache_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: cache path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_cache_path
+
+
+def user_state_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    roaming: bool = False,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param roaming: See `roaming <platformdirs.api.PlatformDirsABC.roaming>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: state path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        roaming=roaming,
+        ensure_exists=ensure_exists,
+    ).user_state_path
+
+
+def user_log_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `roaming <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: log path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_log_path
+
+
+def user_documents_path() -> Path:
+    """:returns: documents a path tied to the user"""
+    return PlatformDirs().user_documents_path
+
+
+def user_downloads_path() -> Path:
+    """:returns: downloads path tied to the user"""
+    return PlatformDirs().user_downloads_path
+
+
+def user_pictures_path() -> Path:
+    """:returns: pictures path tied to the user"""
+    return PlatformDirs().user_pictures_path
+
+
+def user_videos_path() -> Path:
+    """:returns: videos path tied to the user"""
+    return PlatformDirs().user_videos_path
+
+
+def user_music_path() -> Path:
+    """:returns: music path tied to the user"""
+    return PlatformDirs().user_music_path
+
+
+def user_desktop_path() -> Path:
+    """:returns: desktop path tied to the user"""
+    return PlatformDirs().user_desktop_path
+
+
+def user_runtime_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime path tied to the user
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).user_runtime_path
+
+
+def site_runtime_path(
+    appname: str | None = None,
+    appauthor: str | None | Literal[False] = None,
+    version: str | None = None,
+    opinion: bool = True,  # noqa: FBT001, FBT002
+    ensure_exists: bool = False,  # noqa: FBT001, FBT002
+) -> Path:
+    """
+    :param appname: See `appname <platformdirs.api.PlatformDirsABC.appname>`.
+    :param appauthor: See `appauthor <platformdirs.api.PlatformDirsABC.appauthor>`.
+    :param version: See `version <platformdirs.api.PlatformDirsABC.version>`.
+    :param opinion: See `opinion <platformdirs.api.PlatformDirsABC.opinion>`.
+    :param ensure_exists: See `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+    :returns: runtime path shared by users
+    """
+    return PlatformDirs(
+        appname=appname,
+        appauthor=appauthor,
+        version=version,
+        opinion=opinion,
+        ensure_exists=ensure_exists,
+    ).site_runtime_path
+
+
+__all__ = [
+    "AppDirs",
+    "PlatformDirs",
+    "PlatformDirsABC",
+    "__version__",
+    "__version_info__",
+    "site_cache_dir",
+    "site_cache_path",
+    "site_config_dir",
+    "site_config_path",
+    "site_data_dir",
+    "site_data_path",
+    "site_runtime_dir",
+    "site_runtime_path",
+    "user_cache_dir",
+    "user_cache_path",
+    "user_config_dir",
+    "user_config_path",
+    "user_data_dir",
+    "user_data_path",
+    "user_desktop_dir",
+    "user_desktop_path",
+    "user_documents_dir",
+    "user_documents_path",
+    "user_downloads_dir",
+    "user_downloads_path",
+    "user_log_dir",
+    "user_log_path",
+    "user_music_dir",
+    "user_music_path",
+    "user_pictures_dir",
+    "user_pictures_path",
+    "user_runtime_dir",
+    "user_runtime_path",
+    "user_state_dir",
+    "user_state_path",
+    "user_videos_dir",
+    "user_videos_path",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__main__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__main__.py
new file mode 100644
index 00000000..922c5213
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__main__.py
@@ -0,0 +1,55 @@
+"""Main entry point."""
+
+from __future__ import annotations
+
+from platformdirs import PlatformDirs, __version__
+
+PROPS = (
+    "user_data_dir",
+    "user_config_dir",
+    "user_cache_dir",
+    "user_state_dir",
+    "user_log_dir",
+    "user_documents_dir",
+    "user_downloads_dir",
+    "user_pictures_dir",
+    "user_videos_dir",
+    "user_music_dir",
+    "user_runtime_dir",
+    "site_data_dir",
+    "site_config_dir",
+    "site_cache_dir",
+    "site_runtime_dir",
+)
+
+
+def main() -> None:
+    """Run the main entry point."""
+    app_name = "MyApp"
+    app_author = "MyCompany"
+
+    print(f"-- platformdirs {__version__} --")  # noqa: T201
+
+    print("-- app dirs (with optional 'version')")  # noqa: T201
+    dirs = PlatformDirs(app_name, app_author, version="1.0")
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (without optional 'version')")  # noqa: T201
+    dirs = PlatformDirs(app_name, app_author)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (without optional 'appauthor')")  # noqa: T201
+    dirs = PlatformDirs(app_name)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+    print("\n-- app dirs (with disabled 'appauthor')")  # noqa: T201
+    dirs = PlatformDirs(app_name, appauthor=False)
+    for prop in PROPS:
+        print(f"{prop}: {getattr(dirs, prop)}")  # noqa: T201
+
+
+if __name__ == "__main__":
+    main()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..cd83c254
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..6afd0403
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/android.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/android.cpython-312.pyc
new file mode 100644
index 00000000..dfa10f75
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/android.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..63f21829
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc
new file mode 100644
index 00000000..a70e4548
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/macos.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc
new file mode 100644
index 00000000..1c470e9b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/unix.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..496ed408
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc
new file mode 100644
index 00000000..1a671824
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/__pycache__/windows.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/android.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/android.py
new file mode 100644
index 00000000..afd3141c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/android.py
@@ -0,0 +1,249 @@
+"""Android."""
+
+from __future__ import annotations
+
+import os
+import re
+import sys
+from functools import lru_cache
+from typing import TYPE_CHECKING, cast
+
+from .api import PlatformDirsABC
+
+
+class Android(PlatformDirsABC):
+    """
+    Follows the guidance `from here <https://android.stackexchange.com/a/216132>`_.
+
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`, `version
+    <platformdirs.api.PlatformDirsABC.version>`, `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user, e.g. ``/data/user/<userid>/<packagename>/files/<AppName>``"""
+        return self._append_app_name_and_version(cast(str, _android_folder()), "files")
+
+    @property
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_config_dir(self) -> str:
+        """
+        :return: config directory tied to the user, e.g. \
+        ``/data/user/<userid>/<packagename>/shared_prefs/<AppName>``
+        """
+        return self._append_app_name_and_version(cast(str, _android_folder()), "shared_prefs")
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `user_config_dir`"""
+        return self.user_config_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user, e.g.,``/data/user/<userid>/<packagename>/cache/<AppName>``"""
+        return self._append_app_name_and_version(cast(str, _android_folder()), "cache")
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, same as `user_cache_dir`"""
+        return self.user_cache_dir
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """
+        :return: log directory tied to the user, same as `user_cache_dir` if not opinionated else ``log`` in it,
+          e.g. ``/data/user/<userid>/<packagename>/cache/<AppName>/log``
+        """
+        path = self.user_cache_dir
+        if self.opinion:
+            path = os.path.join(path, "log")  # noqa: PTH118
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user e.g. ``/storage/emulated/0/Documents``"""
+        return _android_documents_folder()
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user e.g. ``/storage/emulated/0/Downloads``"""
+        return _android_downloads_folder()
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user e.g. ``/storage/emulated/0/Pictures``"""
+        return _android_pictures_folder()
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user e.g. ``/storage/emulated/0/DCIM/Camera``"""
+        return _android_videos_folder()
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user e.g. ``/storage/emulated/0/Music``"""
+        return _android_music_folder()
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user e.g. ``/storage/emulated/0/Desktop``"""
+        return "/storage/emulated/0/Desktop"
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, same as `user_cache_dir` if not opinionated else ``tmp`` in it,
+          e.g. ``/data/user/<userid>/<packagename>/cache/<AppName>/tmp``
+        """
+        path = self.user_cache_dir
+        if self.opinion:
+            path = os.path.join(path, "tmp")  # noqa: PTH118
+        return path
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+@lru_cache(maxsize=1)
+def _android_folder() -> str | None:  # noqa: C901, PLR0912
+    """:return: base folder for the Android OS or None if it cannot be found"""
+    result: str | None = None
+    # type checker isn't happy with our "import android", just don't do this when type checking see
+    # https://stackoverflow.com/a/61394121
+    if not TYPE_CHECKING:
+        try:
+            # First try to get a path to android app using python4android (if available)...
+            from android import mActivity  # noqa: PLC0415
+
+            context = cast("android.content.Context", mActivity.getApplicationContext())  # noqa: F821
+            result = context.getFilesDir().getParentFile().getAbsolutePath()
+        except Exception:  # noqa: BLE001
+            result = None
+    if result is None:
+        try:
+            # ...and fall back to using plain pyjnius, if python4android isn't available or doesn't deliver any useful
+            # result...
+            from jnius import autoclass  # noqa: PLC0415
+
+            context = autoclass("android.content.Context")
+            result = context.getFilesDir().getParentFile().getAbsolutePath()
+        except Exception:  # noqa: BLE001
+            result = None
+    if result is None:
+        # and if that fails, too, find an android folder looking at path on the sys.path
+        # warning: only works for apps installed under /data, not adopted storage etc.
+        pattern = re.compile(r"/data/(data|user/\d+)/(.+)/files")
+        for path in sys.path:
+            if pattern.match(path):
+                result = path.split("/files")[0]
+                break
+        else:
+            result = None
+    if result is None:
+        # one last try: find an android folder looking at path on the sys.path taking adopted storage paths into
+        # account
+        pattern = re.compile(r"/mnt/expand/[a-fA-F0-9-]{36}/(data|user/\d+)/(.+)/files")
+        for path in sys.path:
+            if pattern.match(path):
+                result = path.split("/files")[0]
+                break
+        else:
+            result = None
+    return result
+
+
+@lru_cache(maxsize=1)
+def _android_documents_folder() -> str:
+    """:return: documents folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        documents_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DOCUMENTS).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        documents_dir = "/storage/emulated/0/Documents"
+
+    return documents_dir
+
+
+@lru_cache(maxsize=1)
+def _android_downloads_folder() -> str:
+    """:return: downloads folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        downloads_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DOWNLOADS).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        downloads_dir = "/storage/emulated/0/Downloads"
+
+    return downloads_dir
+
+
+@lru_cache(maxsize=1)
+def _android_pictures_folder() -> str:
+    """:return: pictures folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        pictures_dir: str = context.getExternalFilesDir(environment.DIRECTORY_PICTURES).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        pictures_dir = "/storage/emulated/0/Pictures"
+
+    return pictures_dir
+
+
+@lru_cache(maxsize=1)
+def _android_videos_folder() -> str:
+    """:return: videos folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        videos_dir: str = context.getExternalFilesDir(environment.DIRECTORY_DCIM).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        videos_dir = "/storage/emulated/0/DCIM/Camera"
+
+    return videos_dir
+
+
+@lru_cache(maxsize=1)
+def _android_music_folder() -> str:
+    """:return: music folder for the Android OS"""
+    # Get directories with pyjnius
+    try:
+        from jnius import autoclass  # noqa: PLC0415
+
+        context = autoclass("android.content.Context")
+        environment = autoclass("android.os.Environment")
+        music_dir: str = context.getExternalFilesDir(environment.DIRECTORY_MUSIC).getAbsolutePath()
+    except Exception:  # noqa: BLE001
+        music_dir = "/storage/emulated/0/Music"
+
+    return music_dir
+
+
+__all__ = [
+    "Android",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/api.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/api.py
new file mode 100644
index 00000000..c50caa64
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/api.py
@@ -0,0 +1,292 @@
+"""Base API."""
+
+from __future__ import annotations
+
+import os
+from abc import ABC, abstractmethod
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing import Iterator, Literal
+
+
+class PlatformDirsABC(ABC):  # noqa: PLR0904
+    """Abstract base class for platform directories."""
+
+    def __init__(  # noqa: PLR0913, PLR0917
+        self,
+        appname: str | None = None,
+        appauthor: str | None | Literal[False] = None,
+        version: str | None = None,
+        roaming: bool = False,  # noqa: FBT001, FBT002
+        multipath: bool = False,  # noqa: FBT001, FBT002
+        opinion: bool = True,  # noqa: FBT001, FBT002
+        ensure_exists: bool = False,  # noqa: FBT001, FBT002
+    ) -> None:
+        """
+        Create a new platform directory.
+
+        :param appname: See `appname`.
+        :param appauthor: See `appauthor`.
+        :param version: See `version`.
+        :param roaming: See `roaming`.
+        :param multipath: See `multipath`.
+        :param opinion: See `opinion`.
+        :param ensure_exists: See `ensure_exists`.
+
+        """
+        self.appname = appname  #: The name of application.
+        self.appauthor = appauthor
+        """
+        The name of the app author or distributing body for this application.
+
+        Typically, it is the owning company name. Defaults to `appname`. You may pass ``False`` to disable it.
+
+        """
+        self.version = version
+        """
+        An optional version path element to append to the path.
+
+        You might want to use this if you want multiple versions of your app to be able to run independently. If used,
+        this would typically be ``<major>.<minor>``.
+
+        """
+        self.roaming = roaming
+        """
+        Whether to use the roaming appdata directory on Windows.
+
+        That means that for users on a Windows network setup for roaming profiles, this user data will be synced on
+        login (see
+        `here <https://technet.microsoft.com/en-us/library/cc766489(WS.10).aspx>`_).
+
+        """
+        self.multipath = multipath
+        """
+        An optional parameter which indicates that the entire list of data dirs should be returned.
+
+        By default, the first item would only be returned.
+
+        """
+        self.opinion = opinion  #: A flag to indicating to use opinionated values.
+        self.ensure_exists = ensure_exists
+        """
+        Optionally create the directory (and any missing parents) upon access if it does not exist.
+
+        By default, no directories are created.
+
+        """
+
+    def _append_app_name_and_version(self, *base: str) -> str:
+        params = list(base[1:])
+        if self.appname:
+            params.append(self.appname)
+            if self.version:
+                params.append(self.version)
+        path = os.path.join(base[0], *params)  # noqa: PTH118
+        self._optionally_create_directory(path)
+        return path
+
+    def _optionally_create_directory(self, path: str) -> None:
+        if self.ensure_exists:
+            Path(path).mkdir(parents=True, exist_ok=True)
+
+    @property
+    @abstractmethod
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users"""
+
+    @property
+    @abstractmethod
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users"""
+
+    @property
+    @abstractmethod
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users"""
+
+    @property
+    @abstractmethod
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def user_runtime_dir(self) -> str:
+        """:return: runtime directory tied to the user"""
+
+    @property
+    @abstractmethod
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users"""
+
+    @property
+    def user_data_path(self) -> Path:
+        """:return: data path tied to the user"""
+        return Path(self.user_data_dir)
+
+    @property
+    def site_data_path(self) -> Path:
+        """:return: data path shared by users"""
+        return Path(self.site_data_dir)
+
+    @property
+    def user_config_path(self) -> Path:
+        """:return: config path tied to the user"""
+        return Path(self.user_config_dir)
+
+    @property
+    def site_config_path(self) -> Path:
+        """:return: config path shared by the users"""
+        return Path(self.site_config_dir)
+
+    @property
+    def user_cache_path(self) -> Path:
+        """:return: cache path tied to the user"""
+        return Path(self.user_cache_dir)
+
+    @property
+    def site_cache_path(self) -> Path:
+        """:return: cache path shared by users"""
+        return Path(self.site_cache_dir)
+
+    @property
+    def user_state_path(self) -> Path:
+        """:return: state path tied to the user"""
+        return Path(self.user_state_dir)
+
+    @property
+    def user_log_path(self) -> Path:
+        """:return: log path tied to the user"""
+        return Path(self.user_log_dir)
+
+    @property
+    def user_documents_path(self) -> Path:
+        """:return: documents a path tied to the user"""
+        return Path(self.user_documents_dir)
+
+    @property
+    def user_downloads_path(self) -> Path:
+        """:return: downloads path tied to the user"""
+        return Path(self.user_downloads_dir)
+
+    @property
+    def user_pictures_path(self) -> Path:
+        """:return: pictures path tied to the user"""
+        return Path(self.user_pictures_dir)
+
+    @property
+    def user_videos_path(self) -> Path:
+        """:return: videos path tied to the user"""
+        return Path(self.user_videos_dir)
+
+    @property
+    def user_music_path(self) -> Path:
+        """:return: music path tied to the user"""
+        return Path(self.user_music_dir)
+
+    @property
+    def user_desktop_path(self) -> Path:
+        """:return: desktop path tied to the user"""
+        return Path(self.user_desktop_dir)
+
+    @property
+    def user_runtime_path(self) -> Path:
+        """:return: runtime path tied to the user"""
+        return Path(self.user_runtime_dir)
+
+    @property
+    def site_runtime_path(self) -> Path:
+        """:return: runtime path shared by users"""
+        return Path(self.site_runtime_dir)
+
+    def iter_config_dirs(self) -> Iterator[str]:
+        """:yield: all user and site configuration directories."""
+        yield self.user_config_dir
+        yield self.site_config_dir
+
+    def iter_data_dirs(self) -> Iterator[str]:
+        """:yield: all user and site data directories."""
+        yield self.user_data_dir
+        yield self.site_data_dir
+
+    def iter_cache_dirs(self) -> Iterator[str]:
+        """:yield: all user and site cache directories."""
+        yield self.user_cache_dir
+        yield self.site_cache_dir
+
+    def iter_runtime_dirs(self) -> Iterator[str]:
+        """:yield: all user and site runtime directories."""
+        yield self.user_runtime_dir
+        yield self.site_runtime_dir
+
+    def iter_config_paths(self) -> Iterator[Path]:
+        """:yield: all user and site configuration paths."""
+        for path in self.iter_config_dirs():
+            yield Path(path)
+
+    def iter_data_paths(self) -> Iterator[Path]:
+        """:yield: all user and site data paths."""
+        for path in self.iter_data_dirs():
+            yield Path(path)
+
+    def iter_cache_paths(self) -> Iterator[Path]:
+        """:yield: all user and site cache paths."""
+        for path in self.iter_cache_dirs():
+            yield Path(path)
+
+    def iter_runtime_paths(self) -> Iterator[Path]:
+        """:yield: all user and site runtime paths."""
+        for path in self.iter_runtime_dirs():
+            yield Path(path)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/macos.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/macos.py
new file mode 100644
index 00000000..eb1ba5df
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/macos.py
@@ -0,0 +1,130 @@
+"""macOS."""
+
+from __future__ import annotations
+
+import os.path
+import sys
+
+from .api import PlatformDirsABC
+
+
+class MacOS(PlatformDirsABC):
+    """
+    Platform directories for the macOS operating system.
+
+    Follows the guidance from
+    `Apple documentation <https://developer.apple.com/library/archive/documentation/FileManagement/Conceptual/FileSystemProgrammingGuide/MacOSXDirectories/MacOSXDirectories.html>`_.
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`,
+    `version <platformdirs.api.PlatformDirsABC.version>`,
+    `ensure_exists <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """:return: data directory tied to the user, e.g. ``~/Library/Application Support/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Application Support"))  # noqa: PTH111
+
+    @property
+    def site_data_dir(self) -> str:
+        """
+        :return: data directory shared by users, e.g. ``/Library/Application Support/$appname/$version``.
+          If we're using a Python binary managed by `Homebrew <https://brew.sh>`_, the directory
+          will be under the Homebrew prefix, e.g. ``/opt/homebrew/share/$appname/$version``.
+          If `multipath <platformdirs.api.PlatformDirsABC.multipath>` is enabled, and we're in Homebrew,
+          the response is a multi-path string separated by ":", e.g.
+          ``/opt/homebrew/share/$appname/$version:/Library/Application Support/$appname/$version``
+        """
+        is_homebrew = sys.prefix.startswith("/opt/homebrew")
+        path_list = [self._append_app_name_and_version("/opt/homebrew/share")] if is_homebrew else []
+        path_list.append(self._append_app_name_and_version("/Library/Application Support"))
+        if self.multipath:
+            return os.pathsep.join(path_list)
+        return path_list[0]
+
+    @property
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `site_data_dir`"""
+        return self.site_data_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """:return: cache directory tied to the user, e.g. ``~/Library/Caches/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Caches"))  # noqa: PTH111
+
+    @property
+    def site_cache_dir(self) -> str:
+        """
+        :return: cache directory shared by users, e.g. ``/Library/Caches/$appname/$version``.
+          If we're using a Python binary managed by `Homebrew <https://brew.sh>`_, the directory
+          will be under the Homebrew prefix, e.g. ``/opt/homebrew/var/cache/$appname/$version``.
+          If `multipath <platformdirs.api.PlatformDirsABC.multipath>` is enabled, and we're in Homebrew,
+          the response is a multi-path string separated by ":", e.g.
+          ``/opt/homebrew/var/cache/$appname/$version:/Library/Caches/$appname/$version``
+        """
+        is_homebrew = sys.prefix.startswith("/opt/homebrew")
+        path_list = [self._append_app_name_and_version("/opt/homebrew/var/cache")] if is_homebrew else []
+        path_list.append(self._append_app_name_and_version("/Library/Caches"))
+        if self.multipath:
+            return os.pathsep.join(path_list)
+        return path_list[0]
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, e.g. ``~/Library/Logs/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Logs"))  # noqa: PTH111
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user, e.g. ``~/Documents``"""
+        return os.path.expanduser("~/Documents")  # noqa: PTH111
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user, e.g. ``~/Downloads``"""
+        return os.path.expanduser("~/Downloads")  # noqa: PTH111
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user, e.g. ``~/Pictures``"""
+        return os.path.expanduser("~/Pictures")  # noqa: PTH111
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user, e.g. ``~/Movies``"""
+        return os.path.expanduser("~/Movies")  # noqa: PTH111
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user, e.g. ``~/Music``"""
+        return os.path.expanduser("~/Music")  # noqa: PTH111
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``~/Desktop``"""
+        return os.path.expanduser("~/Desktop")  # noqa: PTH111
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """:return: runtime directory tied to the user, e.g. ``~/Library/Caches/TemporaryItems/$appname/$version``"""
+        return self._append_app_name_and_version(os.path.expanduser("~/Library/Caches/TemporaryItems"))  # noqa: PTH111
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+__all__ = [
+    "MacOS",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/unix.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/unix.py
new file mode 100644
index 00000000..9500ade6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/unix.py
@@ -0,0 +1,275 @@
+"""Unix."""
+
+from __future__ import annotations
+
+import os
+import sys
+from configparser import ConfigParser
+from pathlib import Path
+from typing import Iterator, NoReturn
+
+from .api import PlatformDirsABC
+
+if sys.platform == "win32":
+
+    def getuid() -> NoReturn:
+        msg = "should only be used on Unix"
+        raise RuntimeError(msg)
+
+else:
+    from os import getuid
+
+
+class Unix(PlatformDirsABC):  # noqa: PLR0904
+    """
+    On Unix/Linux, we follow the `XDG Basedir Spec <https://specifications.freedesktop.org/basedir-spec/basedir-spec-
+    latest.html>`_.
+
+    The spec allows overriding directories with environment variables. The examples shown are the default values,
+    alongside the name of the environment variable that overrides them. Makes use of the `appname
+    <platformdirs.api.PlatformDirsABC.appname>`, `version <platformdirs.api.PlatformDirsABC.version>`, `multipath
+    <platformdirs.api.PlatformDirsABC.multipath>`, `opinion <platformdirs.api.PlatformDirsABC.opinion>`, `ensure_exists
+    <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """
+        :return: data directory tied to the user, e.g. ``~/.local/share/$appname/$version`` or
+         ``$XDG_DATA_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_DATA_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.local/share")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def _site_data_dirs(self) -> list[str]:
+        path = os.environ.get("XDG_DATA_DIRS", "")
+        if not path.strip():
+            path = f"/usr/local/share{os.pathsep}/usr/share"
+        return [self._append_app_name_and_version(p) for p in path.split(os.pathsep)]
+
+    @property
+    def site_data_dir(self) -> str:
+        """
+        :return: data directories shared by users (if `multipath <platformdirs.api.PlatformDirsABC.multipath>` is
+         enabled and ``XDG_DATA_DIRS`` is set and a multi path the response is also a multi path separated by the
+         OS path separator), e.g. ``/usr/local/share/$appname/$version`` or ``/usr/share/$appname/$version``
+        """
+        # XDG default for $XDG_DATA_DIRS; only first, if multipath is False
+        dirs = self._site_data_dirs
+        if not self.multipath:
+            return dirs[0]
+        return os.pathsep.join(dirs)
+
+    @property
+    def user_config_dir(self) -> str:
+        """
+        :return: config directory tied to the user, e.g. ``~/.config/$appname/$version`` or
+         ``$XDG_CONFIG_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_CONFIG_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.config")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def _site_config_dirs(self) -> list[str]:
+        path = os.environ.get("XDG_CONFIG_DIRS", "")
+        if not path.strip():
+            path = "/etc/xdg"
+        return [self._append_app_name_and_version(p) for p in path.split(os.pathsep)]
+
+    @property
+    def site_config_dir(self) -> str:
+        """
+        :return: config directories shared by users (if `multipath <platformdirs.api.PlatformDirsABC.multipath>`
+         is enabled and ``XDG_CONFIG_DIRS`` is set and a multi path the response is also a multi path separated by
+         the OS path separator), e.g. ``/etc/xdg/$appname/$version``
+        """
+        # XDG default for $XDG_CONFIG_DIRS only first, if multipath is False
+        dirs = self._site_config_dirs
+        if not self.multipath:
+            return dirs[0]
+        return os.pathsep.join(dirs)
+
+    @property
+    def user_cache_dir(self) -> str:
+        """
+        :return: cache directory tied to the user, e.g. ``~/.cache/$appname/$version`` or
+         ``~/$XDG_CACHE_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_CACHE_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.cache")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, e.g. ``/var/cache/$appname/$version``"""
+        return self._append_app_name_and_version("/var/cache")
+
+    @property
+    def user_state_dir(self) -> str:
+        """
+        :return: state directory tied to the user, e.g. ``~/.local/state/$appname/$version`` or
+         ``$XDG_STATE_HOME/$appname/$version``
+        """
+        path = os.environ.get("XDG_STATE_HOME", "")
+        if not path.strip():
+            path = os.path.expanduser("~/.local/state")  # noqa: PTH111
+        return self._append_app_name_and_version(path)
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, same as `user_state_dir` if not opinionated else ``log`` in it"""
+        path = self.user_state_dir
+        if self.opinion:
+            path = os.path.join(path, "log")  # noqa: PTH118
+            self._optionally_create_directory(path)
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user, e.g. ``~/Documents``"""
+        return _get_user_media_dir("XDG_DOCUMENTS_DIR", "~/Documents")
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user, e.g. ``~/Downloads``"""
+        return _get_user_media_dir("XDG_DOWNLOAD_DIR", "~/Downloads")
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user, e.g. ``~/Pictures``"""
+        return _get_user_media_dir("XDG_PICTURES_DIR", "~/Pictures")
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user, e.g. ``~/Videos``"""
+        return _get_user_media_dir("XDG_VIDEOS_DIR", "~/Videos")
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user, e.g. ``~/Music``"""
+        return _get_user_media_dir("XDG_MUSIC_DIR", "~/Music")
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``~/Desktop``"""
+        return _get_user_media_dir("XDG_DESKTOP_DIR", "~/Desktop")
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, e.g. ``/run/user/$(id -u)/$appname/$version`` or
+         ``$XDG_RUNTIME_DIR/$appname/$version``.
+
+         For FreeBSD/OpenBSD/NetBSD, it would return ``/var/run/user/$(id -u)/$appname/$version`` if
+         exists, otherwise ``/tmp/runtime-$(id -u)/$appname/$version``, if``$XDG_RUNTIME_DIR``
+         is not set.
+        """
+        path = os.environ.get("XDG_RUNTIME_DIR", "")
+        if not path.strip():
+            if sys.platform.startswith(("freebsd", "openbsd", "netbsd")):
+                path = f"/var/run/user/{getuid()}"
+                if not Path(path).exists():
+                    path = f"/tmp/runtime-{getuid()}"  # noqa: S108
+            else:
+                path = f"/run/user/{getuid()}"
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """
+        :return: runtime directory shared by users, e.g. ``/run/$appname/$version`` or \
+        ``$XDG_RUNTIME_DIR/$appname/$version``.
+
+        Note that this behaves almost exactly like `user_runtime_dir` if ``$XDG_RUNTIME_DIR`` is set, but will
+        fall back to paths associated to the root user instead of a regular logged-in user if it's not set.
+
+        If you wish to ensure that a logged-in root user path is returned e.g. ``/run/user/0``, use `user_runtime_dir`
+        instead.
+
+        For FreeBSD/OpenBSD/NetBSD, it would return ``/var/run/$appname/$version`` if ``$XDG_RUNTIME_DIR`` is not set.
+        """
+        path = os.environ.get("XDG_RUNTIME_DIR", "")
+        if not path.strip():
+            if sys.platform.startswith(("freebsd", "openbsd", "netbsd")):
+                path = "/var/run"
+            else:
+                path = "/run"
+        return self._append_app_name_and_version(path)
+
+    @property
+    def site_data_path(self) -> Path:
+        """:return: data path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_data_dir)
+
+    @property
+    def site_config_path(self) -> Path:
+        """:return: config path shared by the users, returns the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_config_dir)
+
+    @property
+    def site_cache_path(self) -> Path:
+        """:return: cache path shared by users. Only return the first item, even if ``multipath`` is set to ``True``"""
+        return self._first_item_as_path_if_multipath(self.site_cache_dir)
+
+    def _first_item_as_path_if_multipath(self, directory: str) -> Path:
+        if self.multipath:
+            # If multipath is True, the first path is returned.
+            directory = directory.split(os.pathsep)[0]
+        return Path(directory)
+
+    def iter_config_dirs(self) -> Iterator[str]:
+        """:yield: all user and site configuration directories."""
+        yield self.user_config_dir
+        yield from self._site_config_dirs
+
+    def iter_data_dirs(self) -> Iterator[str]:
+        """:yield: all user and site data directories."""
+        yield self.user_data_dir
+        yield from self._site_data_dirs
+
+
+def _get_user_media_dir(env_var: str, fallback_tilde_path: str) -> str:
+    media_dir = _get_user_dirs_folder(env_var)
+    if media_dir is None:
+        media_dir = os.environ.get(env_var, "").strip()
+        if not media_dir:
+            media_dir = os.path.expanduser(fallback_tilde_path)  # noqa: PTH111
+
+    return media_dir
+
+
+def _get_user_dirs_folder(key: str) -> str | None:
+    """
+    Return directory from user-dirs.dirs config file.
+
+    See https://freedesktop.org/wiki/Software/xdg-user-dirs/.
+
+    """
+    user_dirs_config_path = Path(Unix().user_config_dir) / "user-dirs.dirs"
+    if user_dirs_config_path.exists():
+        parser = ConfigParser()
+
+        with user_dirs_config_path.open() as stream:
+            # Add fake section header, so ConfigParser doesn't complain
+            parser.read_string(f"[top]\n{stream.read()}")
+
+        if key not in parser["top"]:
+            return None
+
+        path = parser["top"][key].strip('"')
+        # Handle relative home paths
+        return path.replace("$HOME", os.path.expanduser("~"))  # noqa: PTH111
+
+    return None
+
+
+__all__ = [
+    "Unix",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/version.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/version.py
new file mode 100644
index 00000000..6483ddce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '4.2.2'
+__version_tuple__ = version_tuple = (4, 2, 2)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/windows.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/windows.py
new file mode 100644
index 00000000..d7bc9609
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/platformdirs/windows.py
@@ -0,0 +1,272 @@
+"""Windows."""
+
+from __future__ import annotations
+
+import os
+import sys
+from functools import lru_cache
+from typing import TYPE_CHECKING
+
+from .api import PlatformDirsABC
+
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+
+class Windows(PlatformDirsABC):
+    """
+    `MSDN on where to store app data files <https://learn.microsoft.com/en-us/windows/win32/shell/knownfolderid>`_.
+
+    Makes use of the `appname <platformdirs.api.PlatformDirsABC.appname>`, `appauthor
+    <platformdirs.api.PlatformDirsABC.appauthor>`, `version <platformdirs.api.PlatformDirsABC.version>`, `roaming
+    <platformdirs.api.PlatformDirsABC.roaming>`, `opinion <platformdirs.api.PlatformDirsABC.opinion>`, `ensure_exists
+    <platformdirs.api.PlatformDirsABC.ensure_exists>`.
+
+    """
+
+    @property
+    def user_data_dir(self) -> str:
+        """
+        :return: data directory tied to the user, e.g.
+         ``%USERPROFILE%\\AppData\\Local\\$appauthor\\$appname`` (not roaming) or
+         ``%USERPROFILE%\\AppData\\Roaming\\$appauthor\\$appname`` (roaming)
+        """
+        const = "CSIDL_APPDATA" if self.roaming else "CSIDL_LOCAL_APPDATA"
+        path = os.path.normpath(get_win_folder(const))
+        return self._append_parts(path)
+
+    def _append_parts(self, path: str, *, opinion_value: str | None = None) -> str:
+        params = []
+        if self.appname:
+            if self.appauthor is not False:
+                author = self.appauthor or self.appname
+                params.append(author)
+            params.append(self.appname)
+            if opinion_value is not None and self.opinion:
+                params.append(opinion_value)
+            if self.version:
+                params.append(self.version)
+        path = os.path.join(path, *params)  # noqa: PTH118
+        self._optionally_create_directory(path)
+        return path
+
+    @property
+    def site_data_dir(self) -> str:
+        """:return: data directory shared by users, e.g. ``C:\\ProgramData\\$appauthor\\$appname``"""
+        path = os.path.normpath(get_win_folder("CSIDL_COMMON_APPDATA"))
+        return self._append_parts(path)
+
+    @property
+    def user_config_dir(self) -> str:
+        """:return: config directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def site_config_dir(self) -> str:
+        """:return: config directory shared by the users, same as `site_data_dir`"""
+        return self.site_data_dir
+
+    @property
+    def user_cache_dir(self) -> str:
+        """
+        :return: cache directory tied to the user (if opinionated with ``Cache`` folder within ``$appname``) e.g.
+         ``%USERPROFILE%\\AppData\\Local\\$appauthor\\$appname\\Cache\\$version``
+        """
+        path = os.path.normpath(get_win_folder("CSIDL_LOCAL_APPDATA"))
+        return self._append_parts(path, opinion_value="Cache")
+
+    @property
+    def site_cache_dir(self) -> str:
+        """:return: cache directory shared by users, e.g. ``C:\\ProgramData\\$appauthor\\$appname\\Cache\\$version``"""
+        path = os.path.normpath(get_win_folder("CSIDL_COMMON_APPDATA"))
+        return self._append_parts(path, opinion_value="Cache")
+
+    @property
+    def user_state_dir(self) -> str:
+        """:return: state directory tied to the user, same as `user_data_dir`"""
+        return self.user_data_dir
+
+    @property
+    def user_log_dir(self) -> str:
+        """:return: log directory tied to the user, same as `user_data_dir` if not opinionated else ``Logs`` in it"""
+        path = self.user_data_dir
+        if self.opinion:
+            path = os.path.join(path, "Logs")  # noqa: PTH118
+            self._optionally_create_directory(path)
+        return path
+
+    @property
+    def user_documents_dir(self) -> str:
+        """:return: documents directory tied to the user e.g. ``%USERPROFILE%\\Documents``"""
+        return os.path.normpath(get_win_folder("CSIDL_PERSONAL"))
+
+    @property
+    def user_downloads_dir(self) -> str:
+        """:return: downloads directory tied to the user e.g. ``%USERPROFILE%\\Downloads``"""
+        return os.path.normpath(get_win_folder("CSIDL_DOWNLOADS"))
+
+    @property
+    def user_pictures_dir(self) -> str:
+        """:return: pictures directory tied to the user e.g. ``%USERPROFILE%\\Pictures``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYPICTURES"))
+
+    @property
+    def user_videos_dir(self) -> str:
+        """:return: videos directory tied to the user e.g. ``%USERPROFILE%\\Videos``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYVIDEO"))
+
+    @property
+    def user_music_dir(self) -> str:
+        """:return: music directory tied to the user e.g. ``%USERPROFILE%\\Music``"""
+        return os.path.normpath(get_win_folder("CSIDL_MYMUSIC"))
+
+    @property
+    def user_desktop_dir(self) -> str:
+        """:return: desktop directory tied to the user, e.g. ``%USERPROFILE%\\Desktop``"""
+        return os.path.normpath(get_win_folder("CSIDL_DESKTOPDIRECTORY"))
+
+    @property
+    def user_runtime_dir(self) -> str:
+        """
+        :return: runtime directory tied to the user, e.g.
+         ``%USERPROFILE%\\AppData\\Local\\Temp\\$appauthor\\$appname``
+        """
+        path = os.path.normpath(os.path.join(get_win_folder("CSIDL_LOCAL_APPDATA"), "Temp"))  # noqa: PTH118
+        return self._append_parts(path)
+
+    @property
+    def site_runtime_dir(self) -> str:
+        """:return: runtime directory shared by users, same as `user_runtime_dir`"""
+        return self.user_runtime_dir
+
+
+def get_win_folder_from_env_vars(csidl_name: str) -> str:
+    """Get folder from environment variables."""
+    result = get_win_folder_if_csidl_name_not_env_var(csidl_name)
+    if result is not None:
+        return result
+
+    env_var_name = {
+        "CSIDL_APPDATA": "APPDATA",
+        "CSIDL_COMMON_APPDATA": "ALLUSERSPROFILE",
+        "CSIDL_LOCAL_APPDATA": "LOCALAPPDATA",
+    }.get(csidl_name)
+    if env_var_name is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+    result = os.environ.get(env_var_name)
+    if result is None:
+        msg = f"Unset environment variable: {env_var_name}"
+        raise ValueError(msg)
+    return result
+
+
+def get_win_folder_if_csidl_name_not_env_var(csidl_name: str) -> str | None:
+    """Get a folder for a CSIDL name that does not exist as an environment variable."""
+    if csidl_name == "CSIDL_PERSONAL":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Documents")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_DOWNLOADS":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Downloads")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYPICTURES":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Pictures")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYVIDEO":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Videos")  # noqa: PTH118
+
+    if csidl_name == "CSIDL_MYMUSIC":
+        return os.path.join(os.path.normpath(os.environ["USERPROFILE"]), "Music")  # noqa: PTH118
+    return None
+
+
+def get_win_folder_from_registry(csidl_name: str) -> str:
+    """
+    Get folder from the registry.
+
+    This is a fallback technique at best. I'm not sure if using the registry for these guarantees us the correct answer
+    for all CSIDL_* names.
+
+    """
+    shell_folder_name = {
+        "CSIDL_APPDATA": "AppData",
+        "CSIDL_COMMON_APPDATA": "Common AppData",
+        "CSIDL_LOCAL_APPDATA": "Local AppData",
+        "CSIDL_PERSONAL": "Personal",
+        "CSIDL_DOWNLOADS": "{374DE290-123F-4565-9164-39C4925E467B}",
+        "CSIDL_MYPICTURES": "My Pictures",
+        "CSIDL_MYVIDEO": "My Video",
+        "CSIDL_MYMUSIC": "My Music",
+    }.get(csidl_name)
+    if shell_folder_name is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+    if sys.platform != "win32":  # only needed for mypy type checker to know that this code runs only on Windows
+        raise NotImplementedError
+    import winreg  # noqa: PLC0415
+
+    key = winreg.OpenKey(winreg.HKEY_CURRENT_USER, r"Software\Microsoft\Windows\CurrentVersion\Explorer\Shell Folders")
+    directory, _ = winreg.QueryValueEx(key, shell_folder_name)
+    return str(directory)
+
+
+def get_win_folder_via_ctypes(csidl_name: str) -> str:
+    """Get folder with ctypes."""
+    # There is no 'CSIDL_DOWNLOADS'.
+    # Use 'CSIDL_PROFILE' (40) and append the default folder 'Downloads' instead.
+    # https://learn.microsoft.com/en-us/windows/win32/shell/knownfolderid
+
+    import ctypes  # noqa: PLC0415
+
+    csidl_const = {
+        "CSIDL_APPDATA": 26,
+        "CSIDL_COMMON_APPDATA": 35,
+        "CSIDL_LOCAL_APPDATA": 28,
+        "CSIDL_PERSONAL": 5,
+        "CSIDL_MYPICTURES": 39,
+        "CSIDL_MYVIDEO": 14,
+        "CSIDL_MYMUSIC": 13,
+        "CSIDL_DOWNLOADS": 40,
+        "CSIDL_DESKTOPDIRECTORY": 16,
+    }.get(csidl_name)
+    if csidl_const is None:
+        msg = f"Unknown CSIDL name: {csidl_name}"
+        raise ValueError(msg)
+
+    buf = ctypes.create_unicode_buffer(1024)
+    windll = getattr(ctypes, "windll")  # noqa: B009 # using getattr to avoid false positive with mypy type checker
+    windll.shell32.SHGetFolderPathW(None, csidl_const, None, 0, buf)
+
+    # Downgrade to short path name if it has high-bit chars.
+    if any(ord(c) > 255 for c in buf):  # noqa: PLR2004
+        buf2 = ctypes.create_unicode_buffer(1024)
+        if windll.kernel32.GetShortPathNameW(buf.value, buf2, 1024):
+            buf = buf2
+
+    if csidl_name == "CSIDL_DOWNLOADS":
+        return os.path.join(buf.value, "Downloads")  # noqa: PTH118
+
+    return buf.value
+
+
+def _pick_get_win_folder() -> Callable[[str], str]:
+    try:
+        import ctypes  # noqa: PLC0415
+    except ImportError:
+        pass
+    else:
+        if hasattr(ctypes, "windll"):
+            return get_win_folder_via_ctypes
+    try:
+        import winreg  # noqa: PLC0415, F401
+    except ImportError:
+        return get_win_folder_from_env_vars
+    else:
+        return get_win_folder_from_registry
+
+
+get_win_folder = lru_cache(maxsize=None)(_pick_get_win_folder())
+
+__all__ = [
+    "Windows",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/LICENSE
new file mode 100644
index 00000000..e859590f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2021 Taneli Hukkinen
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/METADATA
new file mode 100644
index 00000000..efd87ecc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/METADATA
@@ -0,0 +1,206 @@
+Metadata-Version: 2.1
+Name: tomli
+Version: 2.0.1
+Summary: A lil' TOML parser
+Keywords: toml
+Author-email: Taneli Hukkinen <hukkin@users.noreply.github.com>
+Requires-Python: >=3.7
+Description-Content-Type: text/markdown
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Typing :: Typed
+Project-URL: Changelog, https://github.com/hukkin/tomli/blob/master/CHANGELOG.md
+Project-URL: Homepage, https://github.com/hukkin/tomli
+
+[![Build Status](https://github.com/hukkin/tomli/workflows/Tests/badge.svg?branch=master)](https://github.com/hukkin/tomli/actions?query=workflow%3ATests+branch%3Amaster+event%3Apush)
+[![codecov.io](https://codecov.io/gh/hukkin/tomli/branch/master/graph/badge.svg)](https://codecov.io/gh/hukkin/tomli)
+[![PyPI version](https://img.shields.io/pypi/v/tomli)](https://pypi.org/project/tomli)
+
+# Tomli
+
+> A lil' TOML parser
+
+**Table of Contents**  *generated with [mdformat-toc](https://github.com/hukkin/mdformat-toc)*
+
+<!-- mdformat-toc start --slug=github --maxlevel=6 --minlevel=2 -->
+
+- [Intro](#intro)
+- [Installation](#installation)
+- [Usage](#usage)
+  - [Parse a TOML string](#parse-a-toml-string)
+  - [Parse a TOML file](#parse-a-toml-file)
+  - [Handle invalid TOML](#handle-invalid-toml)
+  - [Construct `decimal.Decimal`s from TOML floats](#construct-decimaldecimals-from-toml-floats)
+- [FAQ](#faq)
+  - [Why this parser?](#why-this-parser)
+  - [Is comment preserving round-trip parsing supported?](#is-comment-preserving-round-trip-parsing-supported)
+  - [Is there a `dumps`, `write` or `encode` function?](#is-there-a-dumps-write-or-encode-function)
+  - [How do TOML types map into Python types?](#how-do-toml-types-map-into-python-types)
+- [Performance](#performance)
+
+<!-- mdformat-toc end -->
+
+## Intro<a name="intro"></a>
+
+Tomli is a Python library for parsing [TOML](https://toml.io).
+Tomli is fully compatible with [TOML v1.0.0](https://toml.io/en/v1.0.0).
+
+## Installation<a name="installation"></a>
+
+```bash
+pip install tomli
+```
+
+## Usage<a name="usage"></a>
+
+### Parse a TOML string<a name="parse-a-toml-string"></a>
+
+```python
+import tomli
+
+toml_str = """
+           gretzky = 99
+
+           [kurri]
+           jari = 17
+           """
+
+toml_dict = tomli.loads(toml_str)
+assert toml_dict == {"gretzky": 99, "kurri": {"jari": 17}}
+```
+
+### Parse a TOML file<a name="parse-a-toml-file"></a>
+
+```python
+import tomli
+
+with open("path_to_file/conf.toml", "rb") as f:
+    toml_dict = tomli.load(f)
+```
+
+The file must be opened in binary mode (with the `"rb"` flag).
+Binary mode will enforce decoding the file as UTF-8 with universal newlines disabled,
+both of which are required to correctly parse TOML.
+
+### Handle invalid TOML<a name="handle-invalid-toml"></a>
+
+```python
+import tomli
+
+try:
+    toml_dict = tomli.loads("]] this is invalid TOML [[")
+except tomli.TOMLDecodeError:
+    print("Yep, definitely not valid.")
+```
+
+Note that error messages are considered informational only.
+They should not be assumed to stay constant across Tomli versions.
+
+### Construct `decimal.Decimal`s from TOML floats<a name="construct-decimaldecimals-from-toml-floats"></a>
+
+```python
+from decimal import Decimal
+import tomli
+
+toml_dict = tomli.loads("precision-matters = 0.982492", parse_float=Decimal)
+assert toml_dict["precision-matters"] == Decimal("0.982492")
+```
+
+Note that `decimal.Decimal` can be replaced with another callable that converts a TOML float from string to a Python type.
+The `decimal.Decimal` is, however, a practical choice for use cases where float inaccuracies can not be tolerated.
+
+Illegal types are `dict` and `list`, and their subtypes.
+A `ValueError` will be raised if `parse_float` produces illegal types.
+
+## FAQ<a name="faq"></a>
+
+### Why this parser?<a name="why-this-parser"></a>
+
+- it's lil'
+- pure Python with zero dependencies
+- the fastest pure Python parser [\*](#performance):
+  15x as fast as [tomlkit](https://pypi.org/project/tomlkit/),
+  2.4x as fast as [toml](https://pypi.org/project/toml/)
+- outputs [basic data types](#how-do-toml-types-map-into-python-types) only
+- 100% spec compliant: passes all tests in
+  [a test set](https://github.com/toml-lang/compliance/pull/8)
+  soon to be merged to the official
+  [compliance tests for TOML](https://github.com/toml-lang/compliance)
+  repository
+- thoroughly tested: 100% branch coverage
+
+### Is comment preserving round-trip parsing supported?<a name="is-comment-preserving-round-trip-parsing-supported"></a>
+
+No.
+
+The `tomli.loads` function returns a plain `dict` that is populated with builtin types and types from the standard library only.
+Preserving comments requires a custom type to be returned so will not be supported,
+at least not by the `tomli.loads` and `tomli.load` functions.
+
+Look into [TOML Kit](https://github.com/sdispater/tomlkit) if preservation of style is what you need.
+
+### Is there a `dumps`, `write` or `encode` function?<a name="is-there-a-dumps-write-or-encode-function"></a>
+
+[Tomli-W](https://github.com/hukkin/tomli-w) is the write-only counterpart of Tomli, providing `dump` and `dumps` functions.
+
+The core library does not include write capability, as most TOML use cases are read-only, and Tomli intends to be minimal.
+
+### How do TOML types map into Python types?<a name="how-do-toml-types-map-into-python-types"></a>
+
+| TOML type        | Python type         | Details                                                      |
+| ---------------- | ------------------- | ------------------------------------------------------------ |
+| Document Root    | `dict`              |                                                              |
+| Key              | `str`               |                                                              |
+| String           | `str`               |                                                              |
+| Integer          | `int`               |                                                              |
+| Float            | `float`             |                                                              |
+| Boolean          | `bool`              |                                                              |
+| Offset Date-Time | `datetime.datetime` | `tzinfo` attribute set to an instance of `datetime.timezone` |
+| Local Date-Time  | `datetime.datetime` | `tzinfo` attribute set to `None`                             |
+| Local Date       | `datetime.date`     |                                                              |
+| Local Time       | `datetime.time`     |                                                              |
+| Array            | `list`              |                                                              |
+| Table            | `dict`              |                                                              |
+| Inline Table     | `dict`              |                                                              |
+
+## Performance<a name="performance"></a>
+
+The `benchmark/` folder in this repository contains a performance benchmark for comparing the various Python TOML parsers.
+The benchmark can be run with `tox -e benchmark-pypi`.
+Running the benchmark on my personal computer output the following:
+
+```console
+foo@bar:~/dev/tomli$ tox -e benchmark-pypi
+benchmark-pypi installed: attrs==19.3.0,click==7.1.2,pytomlpp==1.0.2,qtoml==0.3.0,rtoml==0.7.0,toml==0.10.2,tomli==1.1.0,tomlkit==0.7.2
+benchmark-pypi run-test-pre: PYTHONHASHSEED='2658546909'
+benchmark-pypi run-test: commands[0] | python -c 'import datetime; print(datetime.date.today())'
+2021-07-23
+benchmark-pypi run-test: commands[1] | python --version
+Python 3.8.10
+benchmark-pypi run-test: commands[2] | python benchmark/run.py
+Parsing data.toml 5000 times:
+------------------------------------------------------
+    parser |  exec time | performance (more is better)
+-----------+------------+-----------------------------
+     rtoml |    0.901 s | baseline (100%)
+  pytomlpp |     1.08 s | 83.15%
+     tomli |     3.89 s | 23.15%
+      toml |     9.36 s | 9.63%
+     qtoml |     11.5 s | 7.82%
+   tomlkit |     56.8 s | 1.59%
+```
+
+The parsers are ordered from fastest to slowest, using the fastest parser as baseline.
+Tomli performed the best out of all pure Python TOML parsers,
+losing only to pytomlpp (wraps C++) and rtoml (wraps Rust).
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/RECORD
new file mode 100644
index 00000000..1db8063e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/RECORD
@@ -0,0 +1,15 @@
+tomli-2.0.1.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+tomli-2.0.1.dist-info/LICENSE,sha256=uAgWsNUwuKzLTCIReDeQmEpuO2GSLCte6S8zcqsnQv4,1072
+tomli-2.0.1.dist-info/METADATA,sha256=zPDceKmPwJGLWtZykrHixL7WVXWmJGzZ1jyRT5lCoPI,8875
+tomli-2.0.1.dist-info/RECORD,,
+tomli-2.0.1.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tomli-2.0.1.dist-info/WHEEL,sha256=jPMR_Dzkc4X4icQtmz81lnNY_kAsfog7ry7qoRvYLXw,81
+tomli/__init__.py,sha256=JhUwV66DB1g4Hvt1UQCVMdfCu-IgAV8FXmvDU9onxd4,396
+tomli/__pycache__/__init__.cpython-312.pyc,,
+tomli/__pycache__/_parser.cpython-312.pyc,,
+tomli/__pycache__/_re.cpython-312.pyc,,
+tomli/__pycache__/_types.cpython-312.pyc,,
+tomli/_parser.py,sha256=g9-ENaALS-B8dokYpCuzUFalWlog7T-SIYMjLZSWrtM,22633
+tomli/_re.py,sha256=dbjg5ChZT23Ka9z9DHOXfdtSpPwUfdgMXnj8NOoly-w,2943
+tomli/_types.py,sha256=-GTG2VUqkpxwMqzmVO4F7ybKddIbAnuAHXfmWQcTi3Q,254
+tomli/py.typed,sha256=8PjyZ1aVoQpRVvt71muvuq5qE-jTFZkK-GLHkhdebmc,26
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/WHEEL
new file mode 100644
index 00000000..c727d148
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli-2.0.1.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.6.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__init__.py
new file mode 100644
index 00000000..4c6ec97e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__init__.py
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+__all__ = ("loads", "load", "TOMLDecodeError")
+__version__ = "2.0.1"  # DO NOT EDIT THIS LINE MANUALLY. LET bump2version UTILITY DO IT
+
+from ._parser import TOMLDecodeError, load, loads
+
+# Pretend this exception was created here.
+TOMLDecodeError.__module__ = __name__
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..685af38e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..bf1302af
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_re.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_re.cpython-312.pyc
new file mode 100644
index 00000000..314d1a21
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_re.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_types.cpython-312.pyc
new file mode 100644
index 00000000..75d6c6e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/__pycache__/_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_parser.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_parser.py
new file mode 100644
index 00000000..f1bb0aa1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_parser.py
@@ -0,0 +1,691 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from __future__ import annotations
+
+from collections.abc import Iterable
+import string
+from types import MappingProxyType
+from typing import Any, BinaryIO, NamedTuple
+
+from ._re import (
+    RE_DATETIME,
+    RE_LOCALTIME,
+    RE_NUMBER,
+    match_to_datetime,
+    match_to_localtime,
+    match_to_number,
+)
+from ._types import Key, ParseFloat, Pos
+
+ASCII_CTRL = frozenset(chr(i) for i in range(32)) | frozenset(chr(127))
+
+# Neither of these sets include quotation mark or backslash. They are
+# currently handled as separate cases in the parser functions.
+ILLEGAL_BASIC_STR_CHARS = ASCII_CTRL - frozenset("\t")
+ILLEGAL_MULTILINE_BASIC_STR_CHARS = ASCII_CTRL - frozenset("\t\n")
+
+ILLEGAL_LITERAL_STR_CHARS = ILLEGAL_BASIC_STR_CHARS
+ILLEGAL_MULTILINE_LITERAL_STR_CHARS = ILLEGAL_MULTILINE_BASIC_STR_CHARS
+
+ILLEGAL_COMMENT_CHARS = ILLEGAL_BASIC_STR_CHARS
+
+TOML_WS = frozenset(" \t")
+TOML_WS_AND_NEWLINE = TOML_WS | frozenset("\n")
+BARE_KEY_CHARS = frozenset(string.ascii_letters + string.digits + "-_")
+KEY_INITIAL_CHARS = BARE_KEY_CHARS | frozenset("\"'")
+HEXDIGIT_CHARS = frozenset(string.hexdigits)
+
+BASIC_STR_ESCAPE_REPLACEMENTS = MappingProxyType(
+    {
+        "\\b": "\u0008",  # backspace
+        "\\t": "\u0009",  # tab
+        "\\n": "\u000A",  # linefeed
+        "\\f": "\u000C",  # form feed
+        "\\r": "\u000D",  # carriage return
+        '\\"': "\u0022",  # quote
+        "\\\\": "\u005C",  # backslash
+    }
+)
+
+
+class TOMLDecodeError(ValueError):
+    """An error raised if a document is not valid TOML."""
+
+
+def load(__fp: BinaryIO, *, parse_float: ParseFloat = float) -> dict[str, Any]:
+    """Parse TOML from a binary file object."""
+    b = __fp.read()
+    try:
+        s = b.decode()
+    except AttributeError:
+        raise TypeError(
+            "File must be opened in binary mode, e.g. use `open('foo.toml', 'rb')`"
+        ) from None
+    return loads(s, parse_float=parse_float)
+
+
+def loads(__s: str, *, parse_float: ParseFloat = float) -> dict[str, Any]:  # noqa: C901
+    """Parse TOML from a string."""
+
+    # The spec allows converting "\r\n" to "\n", even in string
+    # literals. Let's do so to simplify parsing.
+    src = __s.replace("\r\n", "\n")
+    pos = 0
+    out = Output(NestedDict(), Flags())
+    header: Key = ()
+    parse_float = make_safe_parse_float(parse_float)
+
+    # Parse one statement at a time
+    # (typically means one line in TOML source)
+    while True:
+        # 1. Skip line leading whitespace
+        pos = skip_chars(src, pos, TOML_WS)
+
+        # 2. Parse rules. Expect one of the following:
+        #    - end of file
+        #    - end of line
+        #    - comment
+        #    - key/value pair
+        #    - append dict to list (and move to its namespace)
+        #    - create dict (and move to its namespace)
+        # Skip trailing whitespace when applicable.
+        try:
+            char = src[pos]
+        except IndexError:
+            break
+        if char == "\n":
+            pos += 1
+            continue
+        if char in KEY_INITIAL_CHARS:
+            pos = key_value_rule(src, pos, out, header, parse_float)
+            pos = skip_chars(src, pos, TOML_WS)
+        elif char == "[":
+            try:
+                second_char: str | None = src[pos + 1]
+            except IndexError:
+                second_char = None
+            out.flags.finalize_pending()
+            if second_char == "[":
+                pos, header = create_list_rule(src, pos, out)
+            else:
+                pos, header = create_dict_rule(src, pos, out)
+            pos = skip_chars(src, pos, TOML_WS)
+        elif char != "#":
+            raise suffixed_err(src, pos, "Invalid statement")
+
+        # 3. Skip comment
+        pos = skip_comment(src, pos)
+
+        # 4. Expect end of line or end of file
+        try:
+            char = src[pos]
+        except IndexError:
+            break
+        if char != "\n":
+            raise suffixed_err(
+                src, pos, "Expected newline or end of document after a statement"
+            )
+        pos += 1
+
+    return out.data.dict
+
+
+class Flags:
+    """Flags that map to parsed keys/namespaces."""
+
+    # Marks an immutable namespace (inline array or inline table).
+    FROZEN = 0
+    # Marks a nest that has been explicitly created and can no longer
+    # be opened using the "[table]" syntax.
+    EXPLICIT_NEST = 1
+
+    def __init__(self) -> None:
+        self._flags: dict[str, dict] = {}
+        self._pending_flags: set[tuple[Key, int]] = set()
+
+    def add_pending(self, key: Key, flag: int) -> None:
+        self._pending_flags.add((key, flag))
+
+    def finalize_pending(self) -> None:
+        for key, flag in self._pending_flags:
+            self.set(key, flag, recursive=False)
+        self._pending_flags.clear()
+
+    def unset_all(self, key: Key) -> None:
+        cont = self._flags
+        for k in key[:-1]:
+            if k not in cont:
+                return
+            cont = cont[k]["nested"]
+        cont.pop(key[-1], None)
+
+    def set(self, key: Key, flag: int, *, recursive: bool) -> None:  # noqa: A003
+        cont = self._flags
+        key_parent, key_stem = key[:-1], key[-1]
+        for k in key_parent:
+            if k not in cont:
+                cont[k] = {"flags": set(), "recursive_flags": set(), "nested": {}}
+            cont = cont[k]["nested"]
+        if key_stem not in cont:
+            cont[key_stem] = {"flags": set(), "recursive_flags": set(), "nested": {}}
+        cont[key_stem]["recursive_flags" if recursive else "flags"].add(flag)
+
+    def is_(self, key: Key, flag: int) -> bool:
+        if not key:
+            return False  # document root has no flags
+        cont = self._flags
+        for k in key[:-1]:
+            if k not in cont:
+                return False
+            inner_cont = cont[k]
+            if flag in inner_cont["recursive_flags"]:
+                return True
+            cont = inner_cont["nested"]
+        key_stem = key[-1]
+        if key_stem in cont:
+            cont = cont[key_stem]
+            return flag in cont["flags"] or flag in cont["recursive_flags"]
+        return False
+
+
+class NestedDict:
+    def __init__(self) -> None:
+        # The parsed content of the TOML document
+        self.dict: dict[str, Any] = {}
+
+    def get_or_create_nest(
+        self,
+        key: Key,
+        *,
+        access_lists: bool = True,
+    ) -> dict:
+        cont: Any = self.dict
+        for k in key:
+            if k not in cont:
+                cont[k] = {}
+            cont = cont[k]
+            if access_lists and isinstance(cont, list):
+                cont = cont[-1]
+            if not isinstance(cont, dict):
+                raise KeyError("There is no nest behind this key")
+        return cont
+
+    def append_nest_to_list(self, key: Key) -> None:
+        cont = self.get_or_create_nest(key[:-1])
+        last_key = key[-1]
+        if last_key in cont:
+            list_ = cont[last_key]
+            if not isinstance(list_, list):
+                raise KeyError("An object other than list found behind this key")
+            list_.append({})
+        else:
+            cont[last_key] = [{}]
+
+
+class Output(NamedTuple):
+    data: NestedDict
+    flags: Flags
+
+
+def skip_chars(src: str, pos: Pos, chars: Iterable[str]) -> Pos:
+    try:
+        while src[pos] in chars:
+            pos += 1
+    except IndexError:
+        pass
+    return pos
+
+
+def skip_until(
+    src: str,
+    pos: Pos,
+    expect: str,
+    *,
+    error_on: frozenset[str],
+    error_on_eof: bool,
+) -> Pos:
+    try:
+        new_pos = src.index(expect, pos)
+    except ValueError:
+        new_pos = len(src)
+        if error_on_eof:
+            raise suffixed_err(src, new_pos, f"Expected {expect!r}") from None
+
+    if not error_on.isdisjoint(src[pos:new_pos]):
+        while src[pos] not in error_on:
+            pos += 1
+        raise suffixed_err(src, pos, f"Found invalid character {src[pos]!r}")
+    return new_pos
+
+
+def skip_comment(src: str, pos: Pos) -> Pos:
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char == "#":
+        return skip_until(
+            src, pos + 1, "\n", error_on=ILLEGAL_COMMENT_CHARS, error_on_eof=False
+        )
+    return pos
+
+
+def skip_comments_and_array_ws(src: str, pos: Pos) -> Pos:
+    while True:
+        pos_before_skip = pos
+        pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
+        pos = skip_comment(src, pos)
+        if pos == pos_before_skip:
+            return pos
+
+
+def create_dict_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
+    pos += 1  # Skip "["
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, key = parse_key(src, pos)
+
+    if out.flags.is_(key, Flags.EXPLICIT_NEST) or out.flags.is_(key, Flags.FROZEN):
+        raise suffixed_err(src, pos, f"Cannot declare {key} twice")
+    out.flags.set(key, Flags.EXPLICIT_NEST, recursive=False)
+    try:
+        out.data.get_or_create_nest(key)
+    except KeyError:
+        raise suffixed_err(src, pos, "Cannot overwrite a value") from None
+
+    if not src.startswith("]", pos):
+        raise suffixed_err(src, pos, "Expected ']' at the end of a table declaration")
+    return pos + 1, key
+
+
+def create_list_rule(src: str, pos: Pos, out: Output) -> tuple[Pos, Key]:
+    pos += 2  # Skip "[["
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, key = parse_key(src, pos)
+
+    if out.flags.is_(key, Flags.FROZEN):
+        raise suffixed_err(src, pos, f"Cannot mutate immutable namespace {key}")
+    # Free the namespace now that it points to another empty list item...
+    out.flags.unset_all(key)
+    # ...but this key precisely is still prohibited from table declaration
+    out.flags.set(key, Flags.EXPLICIT_NEST, recursive=False)
+    try:
+        out.data.append_nest_to_list(key)
+    except KeyError:
+        raise suffixed_err(src, pos, "Cannot overwrite a value") from None
+
+    if not src.startswith("]]", pos):
+        raise suffixed_err(src, pos, "Expected ']]' at the end of an array declaration")
+    return pos + 2, key
+
+
+def key_value_rule(
+    src: str, pos: Pos, out: Output, header: Key, parse_float: ParseFloat
+) -> Pos:
+    pos, key, value = parse_key_value_pair(src, pos, parse_float)
+    key_parent, key_stem = key[:-1], key[-1]
+    abs_key_parent = header + key_parent
+
+    relative_path_cont_keys = (header + key[:i] for i in range(1, len(key)))
+    for cont_key in relative_path_cont_keys:
+        # Check that dotted key syntax does not redefine an existing table
+        if out.flags.is_(cont_key, Flags.EXPLICIT_NEST):
+            raise suffixed_err(src, pos, f"Cannot redefine namespace {cont_key}")
+        # Containers in the relative path can't be opened with the table syntax or
+        # dotted key/value syntax in following table sections.
+        out.flags.add_pending(cont_key, Flags.EXPLICIT_NEST)
+
+    if out.flags.is_(abs_key_parent, Flags.FROZEN):
+        raise suffixed_err(
+            src, pos, f"Cannot mutate immutable namespace {abs_key_parent}"
+        )
+
+    try:
+        nest = out.data.get_or_create_nest(abs_key_parent)
+    except KeyError:
+        raise suffixed_err(src, pos, "Cannot overwrite a value") from None
+    if key_stem in nest:
+        raise suffixed_err(src, pos, "Cannot overwrite a value")
+    # Mark inline table and array namespaces recursively immutable
+    if isinstance(value, (dict, list)):
+        out.flags.set(header + key, Flags.FROZEN, recursive=True)
+    nest[key_stem] = value
+    return pos
+
+
+def parse_key_value_pair(
+    src: str, pos: Pos, parse_float: ParseFloat
+) -> tuple[Pos, Key, Any]:
+    pos, key = parse_key(src, pos)
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char != "=":
+        raise suffixed_err(src, pos, "Expected '=' after a key in a key/value pair")
+    pos += 1
+    pos = skip_chars(src, pos, TOML_WS)
+    pos, value = parse_value(src, pos, parse_float)
+    return pos, key, value
+
+
+def parse_key(src: str, pos: Pos) -> tuple[Pos, Key]:
+    pos, key_part = parse_key_part(src, pos)
+    key: Key = (key_part,)
+    pos = skip_chars(src, pos, TOML_WS)
+    while True:
+        try:
+            char: str | None = src[pos]
+        except IndexError:
+            char = None
+        if char != ".":
+            return pos, key
+        pos += 1
+        pos = skip_chars(src, pos, TOML_WS)
+        pos, key_part = parse_key_part(src, pos)
+        key += (key_part,)
+        pos = skip_chars(src, pos, TOML_WS)
+
+
+def parse_key_part(src: str, pos: Pos) -> tuple[Pos, str]:
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+    if char in BARE_KEY_CHARS:
+        start_pos = pos
+        pos = skip_chars(src, pos, BARE_KEY_CHARS)
+        return pos, src[start_pos:pos]
+    if char == "'":
+        return parse_literal_str(src, pos)
+    if char == '"':
+        return parse_one_line_basic_str(src, pos)
+    raise suffixed_err(src, pos, "Invalid initial character for a key part")
+
+
+def parse_one_line_basic_str(src: str, pos: Pos) -> tuple[Pos, str]:
+    pos += 1
+    return parse_basic_str(src, pos, multiline=False)
+
+
+def parse_array(src: str, pos: Pos, parse_float: ParseFloat) -> tuple[Pos, list]:
+    pos += 1
+    array: list = []
+
+    pos = skip_comments_and_array_ws(src, pos)
+    if src.startswith("]", pos):
+        return pos + 1, array
+    while True:
+        pos, val = parse_value(src, pos, parse_float)
+        array.append(val)
+        pos = skip_comments_and_array_ws(src, pos)
+
+        c = src[pos : pos + 1]
+        if c == "]":
+            return pos + 1, array
+        if c != ",":
+            raise suffixed_err(src, pos, "Unclosed array")
+        pos += 1
+
+        pos = skip_comments_and_array_ws(src, pos)
+        if src.startswith("]", pos):
+            return pos + 1, array
+
+
+def parse_inline_table(src: str, pos: Pos, parse_float: ParseFloat) -> tuple[Pos, dict]:
+    pos += 1
+    nested_dict = NestedDict()
+    flags = Flags()
+
+    pos = skip_chars(src, pos, TOML_WS)
+    if src.startswith("}", pos):
+        return pos + 1, nested_dict.dict
+    while True:
+        pos, key, value = parse_key_value_pair(src, pos, parse_float)
+        key_parent, key_stem = key[:-1], key[-1]
+        if flags.is_(key, Flags.FROZEN):
+            raise suffixed_err(src, pos, f"Cannot mutate immutable namespace {key}")
+        try:
+            nest = nested_dict.get_or_create_nest(key_parent, access_lists=False)
+        except KeyError:
+            raise suffixed_err(src, pos, "Cannot overwrite a value") from None
+        if key_stem in nest:
+            raise suffixed_err(src, pos, f"Duplicate inline table key {key_stem!r}")
+        nest[key_stem] = value
+        pos = skip_chars(src, pos, TOML_WS)
+        c = src[pos : pos + 1]
+        if c == "}":
+            return pos + 1, nested_dict.dict
+        if c != ",":
+            raise suffixed_err(src, pos, "Unclosed inline table")
+        if isinstance(value, (dict, list)):
+            flags.set(key, Flags.FROZEN, recursive=True)
+        pos += 1
+        pos = skip_chars(src, pos, TOML_WS)
+
+
+def parse_basic_str_escape(
+    src: str, pos: Pos, *, multiline: bool = False
+) -> tuple[Pos, str]:
+    escape_id = src[pos : pos + 2]
+    pos += 2
+    if multiline and escape_id in {"\\ ", "\\\t", "\\\n"}:
+        # Skip whitespace until next non-whitespace character or end of
+        # the doc. Error if non-whitespace is found before newline.
+        if escape_id != "\\\n":
+            pos = skip_chars(src, pos, TOML_WS)
+            try:
+                char = src[pos]
+            except IndexError:
+                return pos, ""
+            if char != "\n":
+                raise suffixed_err(src, pos, "Unescaped '\\' in a string")
+            pos += 1
+        pos = skip_chars(src, pos, TOML_WS_AND_NEWLINE)
+        return pos, ""
+    if escape_id == "\\u":
+        return parse_hex_char(src, pos, 4)
+    if escape_id == "\\U":
+        return parse_hex_char(src, pos, 8)
+    try:
+        return pos, BASIC_STR_ESCAPE_REPLACEMENTS[escape_id]
+    except KeyError:
+        raise suffixed_err(src, pos, "Unescaped '\\' in a string") from None
+
+
+def parse_basic_str_escape_multiline(src: str, pos: Pos) -> tuple[Pos, str]:
+    return parse_basic_str_escape(src, pos, multiline=True)
+
+
+def parse_hex_char(src: str, pos: Pos, hex_len: int) -> tuple[Pos, str]:
+    hex_str = src[pos : pos + hex_len]
+    if len(hex_str) != hex_len or not HEXDIGIT_CHARS.issuperset(hex_str):
+        raise suffixed_err(src, pos, "Invalid hex value")
+    pos += hex_len
+    hex_int = int(hex_str, 16)
+    if not is_unicode_scalar_value(hex_int):
+        raise suffixed_err(src, pos, "Escaped character is not a Unicode scalar value")
+    return pos, chr(hex_int)
+
+
+def parse_literal_str(src: str, pos: Pos) -> tuple[Pos, str]:
+    pos += 1  # Skip starting apostrophe
+    start_pos = pos
+    pos = skip_until(
+        src, pos, "'", error_on=ILLEGAL_LITERAL_STR_CHARS, error_on_eof=True
+    )
+    return pos + 1, src[start_pos:pos]  # Skip ending apostrophe
+
+
+def parse_multiline_str(src: str, pos: Pos, *, literal: bool) -> tuple[Pos, str]:
+    pos += 3
+    if src.startswith("\n", pos):
+        pos += 1
+
+    if literal:
+        delim = "'"
+        end_pos = skip_until(
+            src,
+            pos,
+            "'''",
+            error_on=ILLEGAL_MULTILINE_LITERAL_STR_CHARS,
+            error_on_eof=True,
+        )
+        result = src[pos:end_pos]
+        pos = end_pos + 3
+    else:
+        delim = '"'
+        pos, result = parse_basic_str(src, pos, multiline=True)
+
+    # Add at maximum two extra apostrophes/quotes if the end sequence
+    # is 4 or 5 chars long instead of just 3.
+    if not src.startswith(delim, pos):
+        return pos, result
+    pos += 1
+    if not src.startswith(delim, pos):
+        return pos, result + delim
+    pos += 1
+    return pos, result + (delim * 2)
+
+
+def parse_basic_str(src: str, pos: Pos, *, multiline: bool) -> tuple[Pos, str]:
+    if multiline:
+        error_on = ILLEGAL_MULTILINE_BASIC_STR_CHARS
+        parse_escapes = parse_basic_str_escape_multiline
+    else:
+        error_on = ILLEGAL_BASIC_STR_CHARS
+        parse_escapes = parse_basic_str_escape
+    result = ""
+    start_pos = pos
+    while True:
+        try:
+            char = src[pos]
+        except IndexError:
+            raise suffixed_err(src, pos, "Unterminated string") from None
+        if char == '"':
+            if not multiline:
+                return pos + 1, result + src[start_pos:pos]
+            if src.startswith('"""', pos):
+                return pos + 3, result + src[start_pos:pos]
+            pos += 1
+            continue
+        if char == "\\":
+            result += src[start_pos:pos]
+            pos, parsed_escape = parse_escapes(src, pos)
+            result += parsed_escape
+            start_pos = pos
+            continue
+        if char in error_on:
+            raise suffixed_err(src, pos, f"Illegal character {char!r}")
+        pos += 1
+
+
+def parse_value(  # noqa: C901
+    src: str, pos: Pos, parse_float: ParseFloat
+) -> tuple[Pos, Any]:
+    try:
+        char: str | None = src[pos]
+    except IndexError:
+        char = None
+
+    # IMPORTANT: order conditions based on speed of checking and likelihood
+
+    # Basic strings
+    if char == '"':
+        if src.startswith('"""', pos):
+            return parse_multiline_str(src, pos, literal=False)
+        return parse_one_line_basic_str(src, pos)
+
+    # Literal strings
+    if char == "'":
+        if src.startswith("'''", pos):
+            return parse_multiline_str(src, pos, literal=True)
+        return parse_literal_str(src, pos)
+
+    # Booleans
+    if char == "t":
+        if src.startswith("true", pos):
+            return pos + 4, True
+    if char == "f":
+        if src.startswith("false", pos):
+            return pos + 5, False
+
+    # Arrays
+    if char == "[":
+        return parse_array(src, pos, parse_float)
+
+    # Inline tables
+    if char == "{":
+        return parse_inline_table(src, pos, parse_float)
+
+    # Dates and times
+    datetime_match = RE_DATETIME.match(src, pos)
+    if datetime_match:
+        try:
+            datetime_obj = match_to_datetime(datetime_match)
+        except ValueError as e:
+            raise suffixed_err(src, pos, "Invalid date or datetime") from e
+        return datetime_match.end(), datetime_obj
+    localtime_match = RE_LOCALTIME.match(src, pos)
+    if localtime_match:
+        return localtime_match.end(), match_to_localtime(localtime_match)
+
+    # Integers and "normal" floats.
+    # The regex will greedily match any type starting with a decimal
+    # char, so needs to be located after handling of dates and times.
+    number_match = RE_NUMBER.match(src, pos)
+    if number_match:
+        return number_match.end(), match_to_number(number_match, parse_float)
+
+    # Special floats
+    first_three = src[pos : pos + 3]
+    if first_three in {"inf", "nan"}:
+        return pos + 3, parse_float(first_three)
+    first_four = src[pos : pos + 4]
+    if first_four in {"-inf", "+inf", "-nan", "+nan"}:
+        return pos + 4, parse_float(first_four)
+
+    raise suffixed_err(src, pos, "Invalid value")
+
+
+def suffixed_err(src: str, pos: Pos, msg: str) -> TOMLDecodeError:
+    """Return a `TOMLDecodeError` where error message is suffixed with
+    coordinates in source."""
+
+    def coord_repr(src: str, pos: Pos) -> str:
+        if pos >= len(src):
+            return "end of document"
+        line = src.count("\n", 0, pos) + 1
+        if line == 1:
+            column = pos + 1
+        else:
+            column = pos - src.rindex("\n", 0, pos)
+        return f"line {line}, column {column}"
+
+    return TOMLDecodeError(f"{msg} (at {coord_repr(src, pos)})")
+
+
+def is_unicode_scalar_value(codepoint: int) -> bool:
+    return (0 <= codepoint <= 55295) or (57344 <= codepoint <= 1114111)
+
+
+def make_safe_parse_float(parse_float: ParseFloat) -> ParseFloat:
+    """A decorator to make `parse_float` safe.
+
+    `parse_float` must not return dicts or lists, because these types
+    would be mixed with parsed TOML tables and arrays, thus confusing
+    the parser. The returned decorated callable raises `ValueError`
+    instead of returning illegal types.
+    """
+    # The default `float` callable never returns illegal types. Optimize it.
+    if parse_float is float:  # type: ignore[comparison-overlap]
+        return float
+
+    def safe_parse_float(float_str: str) -> Any:
+        float_value = parse_float(float_str)
+        if isinstance(float_value, (dict, list)):
+            raise ValueError("parse_float must not return dicts or lists")
+        return float_value
+
+    return safe_parse_float
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_re.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_re.py
new file mode 100644
index 00000000..994bb749
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_re.py
@@ -0,0 +1,107 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from __future__ import annotations
+
+from datetime import date, datetime, time, timedelta, timezone, tzinfo
+from functools import lru_cache
+import re
+from typing import Any
+
+from ._types import ParseFloat
+
+# E.g.
+# - 00:32:00.999999
+# - 00:32:00
+_TIME_RE_STR = r"([01][0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(?:\.([0-9]{1,6})[0-9]*)?"
+
+RE_NUMBER = re.compile(
+    r"""
+0
+(?:
+    x[0-9A-Fa-f](?:_?[0-9A-Fa-f])*   # hex
+    |
+    b[01](?:_?[01])*                 # bin
+    |
+    o[0-7](?:_?[0-7])*               # oct
+)
+|
+[+-]?(?:0|[1-9](?:_?[0-9])*)         # dec, integer part
+(?P<floatpart>
+    (?:\.[0-9](?:_?[0-9])*)?         # optional fractional part
+    (?:[eE][+-]?[0-9](?:_?[0-9])*)?  # optional exponent part
+)
+""",
+    flags=re.VERBOSE,
+)
+RE_LOCALTIME = re.compile(_TIME_RE_STR)
+RE_DATETIME = re.compile(
+    rf"""
+([0-9]{{4}})-(0[1-9]|1[0-2])-(0[1-9]|[12][0-9]|3[01])  # date, e.g. 1988-10-27
+(?:
+    [Tt ]
+    {_TIME_RE_STR}
+    (?:([Zz])|([+-])([01][0-9]|2[0-3]):([0-5][0-9]))?  # optional time offset
+)?
+""",
+    flags=re.VERBOSE,
+)
+
+
+def match_to_datetime(match: re.Match) -> datetime | date:
+    """Convert a `RE_DATETIME` match to `datetime.datetime` or `datetime.date`.
+
+    Raises ValueError if the match does not correspond to a valid date
+    or datetime.
+    """
+    (
+        year_str,
+        month_str,
+        day_str,
+        hour_str,
+        minute_str,
+        sec_str,
+        micros_str,
+        zulu_time,
+        offset_sign_str,
+        offset_hour_str,
+        offset_minute_str,
+    ) = match.groups()
+    year, month, day = int(year_str), int(month_str), int(day_str)
+    if hour_str is None:
+        return date(year, month, day)
+    hour, minute, sec = int(hour_str), int(minute_str), int(sec_str)
+    micros = int(micros_str.ljust(6, "0")) if micros_str else 0
+    if offset_sign_str:
+        tz: tzinfo | None = cached_tz(
+            offset_hour_str, offset_minute_str, offset_sign_str
+        )
+    elif zulu_time:
+        tz = timezone.utc
+    else:  # local date-time
+        tz = None
+    return datetime(year, month, day, hour, minute, sec, micros, tzinfo=tz)
+
+
+@lru_cache(maxsize=None)
+def cached_tz(hour_str: str, minute_str: str, sign_str: str) -> timezone:
+    sign = 1 if sign_str == "+" else -1
+    return timezone(
+        timedelta(
+            hours=sign * int(hour_str),
+            minutes=sign * int(minute_str),
+        )
+    )
+
+
+def match_to_localtime(match: re.Match) -> time:
+    hour_str, minute_str, sec_str, micros_str = match.groups()
+    micros = int(micros_str.ljust(6, "0")) if micros_str else 0
+    return time(int(hour_str), int(minute_str), int(sec_str), micros)
+
+
+def match_to_number(match: re.Match, parse_float: ParseFloat) -> Any:
+    if match.group("floatpart"):
+        return parse_float(match.group())
+    return int(match.group(), 0)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_types.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_types.py
new file mode 100644
index 00000000..d949412e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/_types.py
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: MIT
+# SPDX-FileCopyrightText: 2021 Taneli Hukkinen
+# Licensed to PSF under a Contributor Agreement.
+
+from typing import Any, Callable, Tuple
+
+# Type annotations
+ParseFloat = Callable[[str], Any]
+Key = Tuple[str, ...]
+Pos = int
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/py.typed
new file mode 100644
index 00000000..7632ecf7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/tomli/py.typed
@@ -0,0 +1 @@
+# Marker file for PEP 561
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/LICENSE
new file mode 100644
index 00000000..07806f8a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/LICENSE
@@ -0,0 +1,19 @@
+This is the MIT license: http://www.opensource.org/licenses/mit-license.php
+
+Copyright (c) Alex Grönholm
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this
+software and associated documentation files (the "Software"), to deal in the Software
+without restriction, including without limitation the rights to use, copy, modify, merge,
+publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons
+to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or
+substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE
+FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/METADATA
new file mode 100644
index 00000000..6e5750b4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/METADATA
@@ -0,0 +1,81 @@
+Metadata-Version: 2.1
+Name: typeguard
+Version: 4.3.0
+Summary: Run-time type checker for Python
+Author-email: Alex Grönholm <alex.gronholm@nextday.fi>
+License: MIT
+Project-URL: Documentation, https://typeguard.readthedocs.io/en/latest/
+Project-URL: Change log, https://typeguard.readthedocs.io/en/latest/versionhistory.html
+Project-URL: Source code, https://github.com/agronholm/typeguard
+Project-URL: Issue tracker, https://github.com/agronholm/typeguard/issues
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Requires-Dist: typing-extensions >=4.10.0
+Requires-Dist: importlib-metadata >=3.6 ; python_version < "3.10"
+Provides-Extra: doc
+Requires-Dist: packaging ; extra == 'doc'
+Requires-Dist: Sphinx >=7 ; extra == 'doc'
+Requires-Dist: sphinx-autodoc-typehints >=1.2.0 ; extra == 'doc'
+Requires-Dist: sphinx-rtd-theme >=1.3.0 ; extra == 'doc'
+Provides-Extra: test
+Requires-Dist: coverage[toml] >=7 ; extra == 'test'
+Requires-Dist: pytest >=7 ; extra == 'test'
+Requires-Dist: mypy >=1.2.0 ; (platform_python_implementation != "PyPy") and extra == 'test'
+
+.. image:: https://github.com/agronholm/typeguard/actions/workflows/test.yml/badge.svg
+  :target: https://github.com/agronholm/typeguard/actions/workflows/test.yml
+  :alt: Build Status
+.. image:: https://coveralls.io/repos/agronholm/typeguard/badge.svg?branch=master&service=github
+  :target: https://coveralls.io/github/agronholm/typeguard?branch=master
+  :alt: Code Coverage
+.. image:: https://readthedocs.org/projects/typeguard/badge/?version=latest
+  :target: https://typeguard.readthedocs.io/en/latest/?badge=latest
+  :alt: Documentation
+
+This library provides run-time type checking for functions defined with
+`PEP 484 <https://www.python.org/dev/peps/pep-0484/>`_ argument (and return) type
+annotations, and any arbitrary objects. It can be used together with static type
+checkers as an additional layer of type safety, to catch type violations that could only
+be detected at run time.
+
+Two principal ways to do type checking are provided:
+
+#. The ``check_type`` function:
+
+   * like ``isinstance()``, but supports arbitrary type annotations (within limits)
+   * can be used as a ``cast()`` replacement, but with actual checking of the value
+#. Code instrumentation:
+
+   * entire modules, or individual functions (via ``@typechecked``) are recompiled, with
+     type checking code injected into them
+   * automatically checks function arguments, return values and assignments to annotated
+     local variables
+   * for generator functions (regular and async), checks yield and send values
+   * requires the original source code of the instrumented module(s) to be accessible
+
+Two options are provided for code instrumentation:
+
+#. the ``@typechecked`` function:
+
+   * can be applied to functions individually
+#. the import hook (``typeguard.install_import_hook()``):
+
+   * automatically instruments targeted modules on import
+   * no manual code changes required in the target modules
+   * requires the import hook to be installed before the targeted modules are imported
+   * may clash with other import hooks
+
+See the documentation_ for further information.
+
+.. _documentation: https://typeguard.readthedocs.io/en/latest/
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/RECORD
new file mode 100644
index 00000000..801e7334
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/RECORD
@@ -0,0 +1,34 @@
+typeguard-4.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+typeguard-4.3.0.dist-info/LICENSE,sha256=YWP3mH37ONa8MgzitwsvArhivEESZRbVUu8c1DJH51g,1130
+typeguard-4.3.0.dist-info/METADATA,sha256=z2dcHAp0TwhYCFU5Deh8x31nazElgujUz9tbuP0pjSE,3717
+typeguard-4.3.0.dist-info/RECORD,,
+typeguard-4.3.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+typeguard-4.3.0.dist-info/entry_points.txt,sha256=qp7NQ1aLtiSgMQqo6gWlfGpy0IIXzoMJmeQTLpzqFZQ,48
+typeguard-4.3.0.dist-info/top_level.txt,sha256=4z28AhuDodwRS_c1J_l8H51t5QuwfTseskYzlxp6grs,10
+typeguard/__init__.py,sha256=Onh4w38elPCjtlcU3JY9k3h70NjsxXIkAflmQn-Z0FY,2071
+typeguard/__pycache__/__init__.cpython-312.pyc,,
+typeguard/__pycache__/_checkers.cpython-312.pyc,,
+typeguard/__pycache__/_config.cpython-312.pyc,,
+typeguard/__pycache__/_decorators.cpython-312.pyc,,
+typeguard/__pycache__/_exceptions.cpython-312.pyc,,
+typeguard/__pycache__/_functions.cpython-312.pyc,,
+typeguard/__pycache__/_importhook.cpython-312.pyc,,
+typeguard/__pycache__/_memo.cpython-312.pyc,,
+typeguard/__pycache__/_pytest_plugin.cpython-312.pyc,,
+typeguard/__pycache__/_suppression.cpython-312.pyc,,
+typeguard/__pycache__/_transformer.cpython-312.pyc,,
+typeguard/__pycache__/_union_transformer.cpython-312.pyc,,
+typeguard/__pycache__/_utils.cpython-312.pyc,,
+typeguard/_checkers.py,sha256=JRrgKicdOEfIBoNEtegYCEIlhpad-a1u1Em7GCj0WCI,31360
+typeguard/_config.py,sha256=nIz8QwDa-oFO3L9O8_6srzlmd99pSby2wOM4Wb7F_B0,2846
+typeguard/_decorators.py,sha256=v6dsIeWvPhExGLP_wXF-RmDUyjZf_Ak28g7gBJ_v0-0,9033
+typeguard/_exceptions.py,sha256=ZIPeiV-FBd5Emw2EaWd2Fvlsrwi4ocwT2fVGBIAtHcQ,1121
+typeguard/_functions.py,sha256=ibgSAKa5ptIm1eR9ARG0BSozAFJPFNASZqhPVyQeqig,10393
+typeguard/_importhook.py,sha256=ugjCDvFcdWMU7UugqlJG91IpVNpEIxtRr-99s0h1k7M,6389
+typeguard/_memo.py,sha256=1juQV_vxnD2JYKbSrebiQuj4oKHz6n67v9pYA-CCISg,1303
+typeguard/_pytest_plugin.py,sha256=-fcSqkv54rIfIF8pDavY5YQPkj4OX8GMt_lL7CQSD4I,4416
+typeguard/_suppression.py,sha256=VQfzxcwIbu3if0f7VBkKM7hkYOA7tNFw9a7jMBsmMg4,2266
+typeguard/_transformer.py,sha256=9Ha7_QhdwoUni_6hvdY-hZbuEergowHrNL2vzHIakFY,44937
+typeguard/_union_transformer.py,sha256=v_42r7-6HuRX2SoFwnyJ-E5PlxXpVeUJPJR1-HU9qSo,1354
+typeguard/_utils.py,sha256=5HhO1rPn5f1M6ymkVAEv7Xmlz1cX-j0OnTMlyHqqrR8,5270
+typeguard/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/entry_points.txt
new file mode 100644
index 00000000..47c9d0bd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/entry_points.txt
@@ -0,0 +1,2 @@
+[pytest11]
+typeguard = typeguard._pytest_plugin
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/top_level.txt
new file mode 100644
index 00000000..be5ec23e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard-4.3.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+typeguard
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__init__.py
new file mode 100644
index 00000000..6781cad0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__init__.py
@@ -0,0 +1,48 @@
+import os
+from typing import Any
+
+from ._checkers import TypeCheckerCallable as TypeCheckerCallable
+from ._checkers import TypeCheckLookupCallback as TypeCheckLookupCallback
+from ._checkers import check_type_internal as check_type_internal
+from ._checkers import checker_lookup_functions as checker_lookup_functions
+from ._checkers import load_plugins as load_plugins
+from ._config import CollectionCheckStrategy as CollectionCheckStrategy
+from ._config import ForwardRefPolicy as ForwardRefPolicy
+from ._config import TypeCheckConfiguration as TypeCheckConfiguration
+from ._decorators import typechecked as typechecked
+from ._decorators import typeguard_ignore as typeguard_ignore
+from ._exceptions import InstrumentationWarning as InstrumentationWarning
+from ._exceptions import TypeCheckError as TypeCheckError
+from ._exceptions import TypeCheckWarning as TypeCheckWarning
+from ._exceptions import TypeHintWarning as TypeHintWarning
+from ._functions import TypeCheckFailCallback as TypeCheckFailCallback
+from ._functions import check_type as check_type
+from ._functions import warn_on_error as warn_on_error
+from ._importhook import ImportHookManager as ImportHookManager
+from ._importhook import TypeguardFinder as TypeguardFinder
+from ._importhook import install_import_hook as install_import_hook
+from ._memo import TypeCheckMemo as TypeCheckMemo
+from ._suppression import suppress_type_checks as suppress_type_checks
+from ._utils import Unset as Unset
+
+# Re-export imports so they look like they live directly in this package
+for value in list(locals().values()):
+    if getattr(value, "__module__", "").startswith(f"{__name__}."):
+        value.__module__ = __name__
+
+
+config: TypeCheckConfiguration
+
+
+def __getattr__(name: str) -> Any:
+    if name == "config":
+        from ._config import global_config
+
+        return global_config
+
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+
+# Automatically load checker lookup functions unless explicitly disabled
+if "TYPEGUARD_DISABLE_PLUGIN_AUTOLOAD" not in os.environ:
+    load_plugins()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e55a6107
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_checkers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_checkers.cpython-312.pyc
new file mode 100644
index 00000000..9e031d9d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_checkers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_config.cpython-312.pyc
new file mode 100644
index 00000000..ef1a03d8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_decorators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_decorators.cpython-312.pyc
new file mode 100644
index 00000000..8a30e0d7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_decorators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_exceptions.cpython-312.pyc
new file mode 100644
index 00000000..c85af9ff
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_functions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_functions.cpython-312.pyc
new file mode 100644
index 00000000..302e56c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_functions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_importhook.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_importhook.cpython-312.pyc
new file mode 100644
index 00000000..472373a6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_importhook.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_memo.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_memo.cpython-312.pyc
new file mode 100644
index 00000000..5b89753c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_memo.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_pytest_plugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_pytest_plugin.cpython-312.pyc
new file mode 100644
index 00000000..6cbefc43
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_pytest_plugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_suppression.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_suppression.cpython-312.pyc
new file mode 100644
index 00000000..5c57c17d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_suppression.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_transformer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_transformer.cpython-312.pyc
new file mode 100644
index 00000000..2a6e71c4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_transformer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_union_transformer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_union_transformer.cpython-312.pyc
new file mode 100644
index 00000000..841d8a70
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_union_transformer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_utils.cpython-312.pyc
new file mode 100644
index 00000000..f8302003
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/__pycache__/_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_checkers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_checkers.py
new file mode 100644
index 00000000..67dd5ad4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_checkers.py
@@ -0,0 +1,993 @@
+from __future__ import annotations
+
+import collections.abc
+import inspect
+import sys
+import types
+import typing
+import warnings
+from enum import Enum
+from inspect import Parameter, isclass, isfunction
+from io import BufferedIOBase, IOBase, RawIOBase, TextIOBase
+from textwrap import indent
+from typing import (
+    IO,
+    AbstractSet,
+    Any,
+    BinaryIO,
+    Callable,
+    Dict,
+    ForwardRef,
+    List,
+    Mapping,
+    MutableMapping,
+    NewType,
+    Optional,
+    Sequence,
+    Set,
+    TextIO,
+    Tuple,
+    Type,
+    TypeVar,
+    Union,
+)
+from unittest.mock import Mock
+from weakref import WeakKeyDictionary
+
+try:
+    import typing_extensions
+except ImportError:
+    typing_extensions = None  # type: ignore[assignment]
+
+# Must use this because typing.is_typeddict does not recognize
+# TypedDict from typing_extensions, and as of version 4.12.0
+# typing_extensions.TypedDict is different from typing.TypedDict
+# on all versions.
+from typing_extensions import is_typeddict
+
+from ._config import ForwardRefPolicy
+from ._exceptions import TypeCheckError, TypeHintWarning
+from ._memo import TypeCheckMemo
+from ._utils import evaluate_forwardref, get_stacklevel, get_type_name, qualified_name
+
+if sys.version_info >= (3, 11):
+    from typing import (
+        Annotated,
+        NotRequired,
+        TypeAlias,
+        get_args,
+        get_origin,
+    )
+
+    SubclassableAny = Any
+else:
+    from typing_extensions import (
+        Annotated,
+        NotRequired,
+        TypeAlias,
+        get_args,
+        get_origin,
+    )
+    from typing_extensions import Any as SubclassableAny
+
+if sys.version_info >= (3, 10):
+    from importlib.metadata import entry_points
+    from typing import ParamSpec
+else:
+    from importlib_metadata import entry_points
+    from typing_extensions import ParamSpec
+
+TypeCheckerCallable: TypeAlias = Callable[
+    [Any, Any, Tuple[Any, ...], TypeCheckMemo], Any
+]
+TypeCheckLookupCallback: TypeAlias = Callable[
+    [Any, Tuple[Any, ...], Tuple[Any, ...]], Optional[TypeCheckerCallable]
+]
+
+checker_lookup_functions: list[TypeCheckLookupCallback] = []
+generic_alias_types: tuple[type, ...] = (type(List), type(List[Any]))
+if sys.version_info >= (3, 9):
+    generic_alias_types += (types.GenericAlias,)
+
+protocol_check_cache: WeakKeyDictionary[
+    type[Any], dict[type[Any], TypeCheckError | None]
+] = WeakKeyDictionary()
+
+# Sentinel
+_missing = object()
+
+# Lifted from mypy.sharedparse
+BINARY_MAGIC_METHODS = {
+    "__add__",
+    "__and__",
+    "__cmp__",
+    "__divmod__",
+    "__div__",
+    "__eq__",
+    "__floordiv__",
+    "__ge__",
+    "__gt__",
+    "__iadd__",
+    "__iand__",
+    "__idiv__",
+    "__ifloordiv__",
+    "__ilshift__",
+    "__imatmul__",
+    "__imod__",
+    "__imul__",
+    "__ior__",
+    "__ipow__",
+    "__irshift__",
+    "__isub__",
+    "__itruediv__",
+    "__ixor__",
+    "__le__",
+    "__lshift__",
+    "__lt__",
+    "__matmul__",
+    "__mod__",
+    "__mul__",
+    "__ne__",
+    "__or__",
+    "__pow__",
+    "__radd__",
+    "__rand__",
+    "__rdiv__",
+    "__rfloordiv__",
+    "__rlshift__",
+    "__rmatmul__",
+    "__rmod__",
+    "__rmul__",
+    "__ror__",
+    "__rpow__",
+    "__rrshift__",
+    "__rshift__",
+    "__rsub__",
+    "__rtruediv__",
+    "__rxor__",
+    "__sub__",
+    "__truediv__",
+    "__xor__",
+}
+
+
+def check_callable(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not callable(value):
+        raise TypeCheckError("is not callable")
+
+    if args:
+        try:
+            signature = inspect.signature(value)
+        except (TypeError, ValueError):
+            return
+
+        argument_types = args[0]
+        if isinstance(argument_types, list) and not any(
+            type(item) is ParamSpec for item in argument_types
+        ):
+            # The callable must not have keyword-only arguments without defaults
+            unfulfilled_kwonlyargs = [
+                param.name
+                for param in signature.parameters.values()
+                if param.kind == Parameter.KEYWORD_ONLY
+                and param.default == Parameter.empty
+            ]
+            if unfulfilled_kwonlyargs:
+                raise TypeCheckError(
+                    f"has mandatory keyword-only arguments in its declaration: "
+                    f'{", ".join(unfulfilled_kwonlyargs)}'
+                )
+
+            num_positional_args = num_mandatory_pos_args = 0
+            has_varargs = False
+            for param in signature.parameters.values():
+                if param.kind in (
+                    Parameter.POSITIONAL_ONLY,
+                    Parameter.POSITIONAL_OR_KEYWORD,
+                ):
+                    num_positional_args += 1
+                    if param.default is Parameter.empty:
+                        num_mandatory_pos_args += 1
+                elif param.kind == Parameter.VAR_POSITIONAL:
+                    has_varargs = True
+
+            if num_mandatory_pos_args > len(argument_types):
+                raise TypeCheckError(
+                    f"has too many mandatory positional arguments in its declaration; "
+                    f"expected {len(argument_types)} but {num_mandatory_pos_args} "
+                    f"mandatory positional argument(s) declared"
+                )
+            elif not has_varargs and num_positional_args < len(argument_types):
+                raise TypeCheckError(
+                    f"has too few arguments in its declaration; expected "
+                    f"{len(argument_types)} but {num_positional_args} argument(s) "
+                    f"declared"
+                )
+
+
+def check_mapping(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if origin_type is Dict or origin_type is dict:
+        if not isinstance(value, dict):
+            raise TypeCheckError("is not a dict")
+    if origin_type is MutableMapping or origin_type is collections.abc.MutableMapping:
+        if not isinstance(value, collections.abc.MutableMapping):
+            raise TypeCheckError("is not a mutable mapping")
+    elif not isinstance(value, collections.abc.Mapping):
+        raise TypeCheckError("is not a mapping")
+
+    if args:
+        key_type, value_type = args
+        if key_type is not Any or value_type is not Any:
+            samples = memo.config.collection_check_strategy.iterate_samples(
+                value.items()
+            )
+            for k, v in samples:
+                try:
+                    check_type_internal(k, key_type, memo)
+                except TypeCheckError as exc:
+                    exc.append_path_element(f"key {k!r}")
+                    raise
+
+                try:
+                    check_type_internal(v, value_type, memo)
+                except TypeCheckError as exc:
+                    exc.append_path_element(f"value of key {k!r}")
+                    raise
+
+
+def check_typed_dict(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, dict):
+        raise TypeCheckError("is not a dict")
+
+    declared_keys = frozenset(origin_type.__annotations__)
+    if hasattr(origin_type, "__required_keys__"):
+        required_keys = set(origin_type.__required_keys__)
+    else:  # py3.8 and lower
+        required_keys = set(declared_keys) if origin_type.__total__ else set()
+
+    existing_keys = set(value)
+    extra_keys = existing_keys - declared_keys
+    if extra_keys:
+        keys_formatted = ", ".join(f'"{key}"' for key in sorted(extra_keys, key=repr))
+        raise TypeCheckError(f"has unexpected extra key(s): {keys_formatted}")
+
+    # Detect NotRequired fields which are hidden by get_type_hints()
+    type_hints: dict[str, type] = {}
+    for key, annotation in origin_type.__annotations__.items():
+        if isinstance(annotation, ForwardRef):
+            annotation = evaluate_forwardref(annotation, memo)
+            if get_origin(annotation) is NotRequired:
+                required_keys.discard(key)
+                annotation = get_args(annotation)[0]
+
+        type_hints[key] = annotation
+
+    missing_keys = required_keys - existing_keys
+    if missing_keys:
+        keys_formatted = ", ".join(f'"{key}"' for key in sorted(missing_keys, key=repr))
+        raise TypeCheckError(f"is missing required key(s): {keys_formatted}")
+
+    for key, argtype in type_hints.items():
+        argvalue = value.get(key, _missing)
+        if argvalue is not _missing:
+            try:
+                check_type_internal(argvalue, argtype, memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"value of key {key!r}")
+                raise
+
+
+def check_list(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, list):
+        raise TypeCheckError("is not a list")
+
+    if args and args != (Any,):
+        samples = memo.config.collection_check_strategy.iterate_samples(value)
+        for i, v in enumerate(samples):
+            try:
+                check_type_internal(v, args[0], memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"item {i}")
+                raise
+
+
+def check_sequence(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, collections.abc.Sequence):
+        raise TypeCheckError("is not a sequence")
+
+    if args and args != (Any,):
+        samples = memo.config.collection_check_strategy.iterate_samples(value)
+        for i, v in enumerate(samples):
+            try:
+                check_type_internal(v, args[0], memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"item {i}")
+                raise
+
+
+def check_set(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if origin_type is frozenset:
+        if not isinstance(value, frozenset):
+            raise TypeCheckError("is not a frozenset")
+    elif not isinstance(value, AbstractSet):
+        raise TypeCheckError("is not a set")
+
+    if args and args != (Any,):
+        samples = memo.config.collection_check_strategy.iterate_samples(value)
+        for v in samples:
+            try:
+                check_type_internal(v, args[0], memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"[{v}]")
+                raise
+
+
+def check_tuple(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    # Specialized check for NamedTuples
+    if field_types := getattr(origin_type, "__annotations__", None):
+        if not isinstance(value, origin_type):
+            raise TypeCheckError(
+                f"is not a named tuple of type {qualified_name(origin_type)}"
+            )
+
+        for name, field_type in field_types.items():
+            try:
+                check_type_internal(getattr(value, name), field_type, memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"attribute {name!r}")
+                raise
+
+        return
+    elif not isinstance(value, tuple):
+        raise TypeCheckError("is not a tuple")
+
+    if args:
+        use_ellipsis = args[-1] is Ellipsis
+        tuple_params = args[: -1 if use_ellipsis else None]
+    else:
+        # Unparametrized Tuple or plain tuple
+        return
+
+    if use_ellipsis:
+        element_type = tuple_params[0]
+        samples = memo.config.collection_check_strategy.iterate_samples(value)
+        for i, element in enumerate(samples):
+            try:
+                check_type_internal(element, element_type, memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"item {i}")
+                raise
+    elif tuple_params == ((),):
+        if value != ():
+            raise TypeCheckError("is not an empty tuple")
+    else:
+        if len(value) != len(tuple_params):
+            raise TypeCheckError(
+                f"has wrong number of elements (expected {len(tuple_params)}, got "
+                f"{len(value)} instead)"
+            )
+
+        for i, (element, element_type) in enumerate(zip(value, tuple_params)):
+            try:
+                check_type_internal(element, element_type, memo)
+            except TypeCheckError as exc:
+                exc.append_path_element(f"item {i}")
+                raise
+
+
+def check_union(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    errors: dict[str, TypeCheckError] = {}
+    try:
+        for type_ in args:
+            try:
+                check_type_internal(value, type_, memo)
+                return
+            except TypeCheckError as exc:
+                errors[get_type_name(type_)] = exc
+
+        formatted_errors = indent(
+            "\n".join(f"{key}: {error}" for key, error in errors.items()), "  "
+        )
+    finally:
+        del errors  # avoid creating ref cycle
+    raise TypeCheckError(f"did not match any element in the union:\n{formatted_errors}")
+
+
+def check_uniontype(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    errors: dict[str, TypeCheckError] = {}
+    for type_ in args:
+        try:
+            check_type_internal(value, type_, memo)
+            return
+        except TypeCheckError as exc:
+            errors[get_type_name(type_)] = exc
+
+    formatted_errors = indent(
+        "\n".join(f"{key}: {error}" for key, error in errors.items()), "  "
+    )
+    raise TypeCheckError(f"did not match any element in the union:\n{formatted_errors}")
+
+
+def check_class(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isclass(value) and not isinstance(value, generic_alias_types):
+        raise TypeCheckError("is not a class")
+
+    if not args:
+        return
+
+    if isinstance(args[0], ForwardRef):
+        expected_class = evaluate_forwardref(args[0], memo)
+    else:
+        expected_class = args[0]
+
+    if expected_class is Any:
+        return
+    elif getattr(expected_class, "_is_protocol", False):
+        check_protocol(value, expected_class, (), memo)
+    elif isinstance(expected_class, TypeVar):
+        check_typevar(value, expected_class, (), memo, subclass_check=True)
+    elif get_origin(expected_class) is Union:
+        errors: dict[str, TypeCheckError] = {}
+        for arg in get_args(expected_class):
+            if arg is Any:
+                return
+
+            try:
+                check_class(value, type, (arg,), memo)
+                return
+            except TypeCheckError as exc:
+                errors[get_type_name(arg)] = exc
+        else:
+            formatted_errors = indent(
+                "\n".join(f"{key}: {error}" for key, error in errors.items()), "  "
+            )
+            raise TypeCheckError(
+                f"did not match any element in the union:\n{formatted_errors}"
+            )
+    elif not issubclass(value, expected_class):  # type: ignore[arg-type]
+        raise TypeCheckError(f"is not a subclass of {qualified_name(expected_class)}")
+
+
+def check_newtype(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    check_type_internal(value, origin_type.__supertype__, memo)
+
+
+def check_instance(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, origin_type):
+        raise TypeCheckError(f"is not an instance of {qualified_name(origin_type)}")
+
+
+def check_typevar(
+    value: Any,
+    origin_type: TypeVar,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+    *,
+    subclass_check: bool = False,
+) -> None:
+    if origin_type.__bound__ is not None:
+        annotation = (
+            Type[origin_type.__bound__] if subclass_check else origin_type.__bound__
+        )
+        check_type_internal(value, annotation, memo)
+    elif origin_type.__constraints__:
+        for constraint in origin_type.__constraints__:
+            annotation = Type[constraint] if subclass_check else constraint
+            try:
+                check_type_internal(value, annotation, memo)
+            except TypeCheckError:
+                pass
+            else:
+                break
+        else:
+            formatted_constraints = ", ".join(
+                get_type_name(constraint) for constraint in origin_type.__constraints__
+            )
+            raise TypeCheckError(
+                f"does not match any of the constraints " f"({formatted_constraints})"
+            )
+
+
+if typing_extensions is None:
+
+    def _is_literal_type(typ: object) -> bool:
+        return typ is typing.Literal
+
+else:
+
+    def _is_literal_type(typ: object) -> bool:
+        return typ is typing.Literal or typ is typing_extensions.Literal
+
+
+def check_literal(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    def get_literal_args(literal_args: tuple[Any, ...]) -> tuple[Any, ...]:
+        retval: list[Any] = []
+        for arg in literal_args:
+            if _is_literal_type(get_origin(arg)):
+                retval.extend(get_literal_args(arg.__args__))
+            elif arg is None or isinstance(arg, (int, str, bytes, bool, Enum)):
+                retval.append(arg)
+            else:
+                raise TypeError(
+                    f"Illegal literal value: {arg}"
+                )  # TypeError here is deliberate
+
+        return tuple(retval)
+
+    final_args = tuple(get_literal_args(args))
+    try:
+        index = final_args.index(value)
+    except ValueError:
+        pass
+    else:
+        if type(final_args[index]) is type(value):
+            return
+
+    formatted_args = ", ".join(repr(arg) for arg in final_args)
+    raise TypeCheckError(f"is not any of ({formatted_args})") from None
+
+
+def check_literal_string(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    check_type_internal(value, str, memo)
+
+
+def check_typeguard(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    check_type_internal(value, bool, memo)
+
+
+def check_none(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if value is not None:
+        raise TypeCheckError("is not None")
+
+
+def check_number(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if origin_type is complex and not isinstance(value, (complex, float, int)):
+        raise TypeCheckError("is neither complex, float or int")
+    elif origin_type is float and not isinstance(value, (float, int)):
+        raise TypeCheckError("is neither float or int")
+
+
+def check_io(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if origin_type is TextIO or (origin_type is IO and args == (str,)):
+        if not isinstance(value, TextIOBase):
+            raise TypeCheckError("is not a text based I/O object")
+    elif origin_type is BinaryIO or (origin_type is IO and args == (bytes,)):
+        if not isinstance(value, (RawIOBase, BufferedIOBase)):
+            raise TypeCheckError("is not a binary I/O object")
+    elif not isinstance(value, IOBase):
+        raise TypeCheckError("is not an I/O object")
+
+
+def check_protocol(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    subject: type[Any] = value if isclass(value) else type(value)
+
+    if subject in protocol_check_cache:
+        result_map = protocol_check_cache[subject]
+        if origin_type in result_map:
+            if exc := result_map[origin_type]:
+                raise exc
+            else:
+                return
+
+    # Collect a set of methods and non-method attributes present in the protocol
+    ignored_attrs = set(dir(typing.Protocol)) | {
+        "__annotations__",
+        "__non_callable_proto_members__",
+    }
+    expected_methods: dict[str, tuple[Any, Any]] = {}
+    expected_noncallable_members: dict[str, Any] = {}
+    for attrname in dir(origin_type):
+        # Skip attributes present in typing.Protocol
+        if attrname in ignored_attrs:
+            continue
+
+        member = getattr(origin_type, attrname)
+        if callable(member):
+            signature = inspect.signature(member)
+            argtypes = [
+                (p.annotation if p.annotation is not Parameter.empty else Any)
+                for p in signature.parameters.values()
+                if p.kind is not Parameter.KEYWORD_ONLY
+            ] or Ellipsis
+            return_annotation = (
+                signature.return_annotation
+                if signature.return_annotation is not Parameter.empty
+                else Any
+            )
+            expected_methods[attrname] = argtypes, return_annotation
+        else:
+            expected_noncallable_members[attrname] = member
+
+    for attrname, annotation in typing.get_type_hints(origin_type).items():
+        expected_noncallable_members[attrname] = annotation
+
+    subject_annotations = typing.get_type_hints(subject)
+
+    # Check that all required methods are present and their signatures are compatible
+    result_map = protocol_check_cache.setdefault(subject, {})
+    try:
+        for attrname, callable_args in expected_methods.items():
+            try:
+                method = getattr(subject, attrname)
+            except AttributeError:
+                if attrname in subject_annotations:
+                    raise TypeCheckError(
+                        f"is not compatible with the {origin_type.__qualname__} protocol "
+                        f"because its {attrname!r} attribute is not a method"
+                    ) from None
+                else:
+                    raise TypeCheckError(
+                        f"is not compatible with the {origin_type.__qualname__} protocol "
+                        f"because it has no method named {attrname!r}"
+                    ) from None
+
+            if not callable(method):
+                raise TypeCheckError(
+                    f"is not compatible with the {origin_type.__qualname__} protocol "
+                    f"because its {attrname!r} attribute is not a callable"
+                )
+
+            # TODO: raise exception on added keyword-only arguments without defaults
+            try:
+                check_callable(method, Callable, callable_args, memo)
+            except TypeCheckError as exc:
+                raise TypeCheckError(
+                    f"is not compatible with the {origin_type.__qualname__} protocol "
+                    f"because its {attrname!r} method {exc}"
+                ) from None
+
+        # Check that all required non-callable members are present
+        for attrname in expected_noncallable_members:
+            # TODO: implement assignability checks for non-callable members
+            if attrname not in subject_annotations and not hasattr(subject, attrname):
+                raise TypeCheckError(
+                    f"is not compatible with the {origin_type.__qualname__} protocol "
+                    f"because it has no attribute named {attrname!r}"
+                )
+    except TypeCheckError as exc:
+        result_map[origin_type] = exc
+        raise
+    else:
+        result_map[origin_type] = None
+
+
+def check_byteslike(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, (bytearray, bytes, memoryview)):
+        raise TypeCheckError("is not bytes-like")
+
+
+def check_self(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if memo.self_type is None:
+        raise TypeCheckError("cannot be checked against Self outside of a method call")
+
+    if isclass(value):
+        if not issubclass(value, memo.self_type):
+            raise TypeCheckError(
+                f"is not an instance of the self type "
+                f"({qualified_name(memo.self_type)})"
+            )
+    elif not isinstance(value, memo.self_type):
+        raise TypeCheckError(
+            f"is not an instance of the self type ({qualified_name(memo.self_type)})"
+        )
+
+
+def check_paramspec(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    pass  # No-op for now
+
+
+def check_instanceof(
+    value: Any,
+    origin_type: Any,
+    args: tuple[Any, ...],
+    memo: TypeCheckMemo,
+) -> None:
+    if not isinstance(value, origin_type):
+        raise TypeCheckError(f"is not an instance of {qualified_name(origin_type)}")
+
+
+def check_type_internal(
+    value: Any,
+    annotation: Any,
+    memo: TypeCheckMemo,
+) -> None:
+    """
+    Check that the given object is compatible with the given type annotation.
+
+    This function should only be used by type checker callables. Applications should use
+    :func:`~.check_type` instead.
+
+    :param value: the value to check
+    :param annotation: the type annotation to check against
+    :param memo: a memo object containing configuration and information necessary for
+        looking up forward references
+    """
+
+    if isinstance(annotation, ForwardRef):
+        try:
+            annotation = evaluate_forwardref(annotation, memo)
+        except NameError:
+            if memo.config.forward_ref_policy is ForwardRefPolicy.ERROR:
+                raise
+            elif memo.config.forward_ref_policy is ForwardRefPolicy.WARN:
+                warnings.warn(
+                    f"Cannot resolve forward reference {annotation.__forward_arg__!r}",
+                    TypeHintWarning,
+                    stacklevel=get_stacklevel(),
+                )
+
+            return
+
+    if annotation is Any or annotation is SubclassableAny or isinstance(value, Mock):
+        return
+
+    # Skip type checks if value is an instance of a class that inherits from Any
+    if not isclass(value) and SubclassableAny in type(value).__bases__:
+        return
+
+    extras: tuple[Any, ...]
+    origin_type = get_origin(annotation)
+    if origin_type is Annotated:
+        annotation, *extras_ = get_args(annotation)
+        extras = tuple(extras_)
+        origin_type = get_origin(annotation)
+    else:
+        extras = ()
+
+    if origin_type is not None:
+        args = get_args(annotation)
+
+        # Compatibility hack to distinguish between unparametrized and empty tuple
+        # (tuple[()]), necessary due to https://github.com/python/cpython/issues/91137
+        if origin_type in (tuple, Tuple) and annotation is not Tuple and not args:
+            args = ((),)
+    else:
+        origin_type = annotation
+        args = ()
+
+    for lookup_func in checker_lookup_functions:
+        checker = lookup_func(origin_type, args, extras)
+        if checker:
+            checker(value, origin_type, args, memo)
+            return
+
+    if isclass(origin_type):
+        if not isinstance(value, origin_type):
+            raise TypeCheckError(f"is not an instance of {qualified_name(origin_type)}")
+    elif type(origin_type) is str:  # noqa: E721
+        warnings.warn(
+            f"Skipping type check against {origin_type!r}; this looks like a "
+            f"string-form forward reference imported from another module",
+            TypeHintWarning,
+            stacklevel=get_stacklevel(),
+        )
+
+
+# Equality checks are applied to these
+origin_type_checkers = {
+    bytes: check_byteslike,
+    AbstractSet: check_set,
+    BinaryIO: check_io,
+    Callable: check_callable,
+    collections.abc.Callable: check_callable,
+    complex: check_number,
+    dict: check_mapping,
+    Dict: check_mapping,
+    float: check_number,
+    frozenset: check_set,
+    IO: check_io,
+    list: check_list,
+    List: check_list,
+    typing.Literal: check_literal,
+    Mapping: check_mapping,
+    MutableMapping: check_mapping,
+    None: check_none,
+    collections.abc.Mapping: check_mapping,
+    collections.abc.MutableMapping: check_mapping,
+    Sequence: check_sequence,
+    collections.abc.Sequence: check_sequence,
+    collections.abc.Set: check_set,
+    set: check_set,
+    Set: check_set,
+    TextIO: check_io,
+    tuple: check_tuple,
+    Tuple: check_tuple,
+    type: check_class,
+    Type: check_class,
+    Union: check_union,
+}
+if sys.version_info >= (3, 10):
+    origin_type_checkers[types.UnionType] = check_uniontype
+    origin_type_checkers[typing.TypeGuard] = check_typeguard
+if sys.version_info >= (3, 11):
+    origin_type_checkers.update(
+        {typing.LiteralString: check_literal_string, typing.Self: check_self}
+    )
+if typing_extensions is not None:
+    # On some Python versions, these may simply be re-exports from typing,
+    # but exactly which Python versions is subject to change,
+    # so it's best to err on the safe side
+    # and update the dictionary on all Python versions
+    # if typing_extensions is installed
+    origin_type_checkers[typing_extensions.Literal] = check_literal
+    origin_type_checkers[typing_extensions.LiteralString] = check_literal_string
+    origin_type_checkers[typing_extensions.Self] = check_self
+    origin_type_checkers[typing_extensions.TypeGuard] = check_typeguard
+
+
+def builtin_checker_lookup(
+    origin_type: Any, args: tuple[Any, ...], extras: tuple[Any, ...]
+) -> TypeCheckerCallable | None:
+    checker = origin_type_checkers.get(origin_type)
+    if checker is not None:
+        return checker
+    elif is_typeddict(origin_type):
+        return check_typed_dict
+    elif isclass(origin_type) and issubclass(
+        origin_type,
+        Tuple,  # type: ignore[arg-type]
+    ):
+        # NamedTuple
+        return check_tuple
+    elif getattr(origin_type, "_is_protocol", False):
+        return check_protocol
+    elif isinstance(origin_type, ParamSpec):
+        return check_paramspec
+    elif isinstance(origin_type, TypeVar):
+        return check_typevar
+    elif origin_type.__class__ is NewType:
+        # typing.NewType on Python 3.10+
+        return check_newtype
+    elif (
+        isfunction(origin_type)
+        and getattr(origin_type, "__module__", None) == "typing"
+        and getattr(origin_type, "__qualname__", "").startswith("NewType.")
+        and hasattr(origin_type, "__supertype__")
+    ):
+        # typing.NewType on Python 3.9 and below
+        return check_newtype
+
+    return None
+
+
+checker_lookup_functions.append(builtin_checker_lookup)
+
+
+def load_plugins() -> None:
+    """
+    Load all type checker lookup functions from entry points.
+
+    All entry points from the ``typeguard.checker_lookup`` group are loaded, and the
+    returned lookup functions are added to :data:`typeguard.checker_lookup_functions`.
+
+    .. note:: This function is called implicitly on import, unless the
+        ``TYPEGUARD_DISABLE_PLUGIN_AUTOLOAD`` environment variable is present.
+    """
+
+    for ep in entry_points(group="typeguard.checker_lookup"):
+        try:
+            plugin = ep.load()
+        except Exception as exc:
+            warnings.warn(
+                f"Failed to load plugin {ep.name!r}: " f"{qualified_name(exc)}: {exc}",
+                stacklevel=2,
+            )
+            continue
+
+        if not callable(plugin):
+            warnings.warn(
+                f"Plugin {ep} returned a non-callable object: {plugin!r}", stacklevel=2
+            )
+            continue
+
+        checker_lookup_functions.insert(0, plugin)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_config.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_config.py
new file mode 100644
index 00000000..36efad53
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_config.py
@@ -0,0 +1,108 @@
+from __future__ import annotations
+
+from collections.abc import Iterable
+from dataclasses import dataclass
+from enum import Enum, auto
+from typing import TYPE_CHECKING, TypeVar
+
+if TYPE_CHECKING:
+    from ._functions import TypeCheckFailCallback
+
+T = TypeVar("T")
+
+
+class ForwardRefPolicy(Enum):
+    """
+    Defines how unresolved forward references are handled.
+
+    Members:
+
+    * ``ERROR``: propagate the :exc:`NameError` when the forward reference lookup fails
+    * ``WARN``: emit a :class:`~.TypeHintWarning` if the forward reference lookup fails
+    * ``IGNORE``: silently skip checks for unresolveable forward references
+    """
+
+    ERROR = auto()
+    WARN = auto()
+    IGNORE = auto()
+
+
+class CollectionCheckStrategy(Enum):
+    """
+    Specifies how thoroughly the contents of collections are type checked.
+
+    This has an effect on the following built-in checkers:
+
+    * ``AbstractSet``
+    * ``Dict``
+    * ``List``
+    * ``Mapping``
+    * ``Set``
+    * ``Tuple[<type>, ...]`` (arbitrarily sized tuples)
+
+    Members:
+
+    * ``FIRST_ITEM``: check only the first item
+    * ``ALL_ITEMS``: check all items
+    """
+
+    FIRST_ITEM = auto()
+    ALL_ITEMS = auto()
+
+    def iterate_samples(self, collection: Iterable[T]) -> Iterable[T]:
+        if self is CollectionCheckStrategy.FIRST_ITEM:
+            try:
+                return [next(iter(collection))]
+            except StopIteration:
+                return ()
+        else:
+            return collection
+
+
+@dataclass
+class TypeCheckConfiguration:
+    """
+     You can change Typeguard's behavior with these settings.
+
+    .. attribute:: typecheck_fail_callback
+       :type: Callable[[TypeCheckError, TypeCheckMemo], Any]
+
+         Callable that is called when type checking fails.
+
+         Default: ``None`` (the :exc:`~.TypeCheckError` is raised directly)
+
+    .. attribute:: forward_ref_policy
+       :type: ForwardRefPolicy
+
+         Specifies what to do when a forward reference fails to resolve.
+
+         Default: ``WARN``
+
+    .. attribute:: collection_check_strategy
+       :type: CollectionCheckStrategy
+
+         Specifies how thoroughly the contents of collections (list, dict, etc.) are
+         type checked.
+
+         Default: ``FIRST_ITEM``
+
+    .. attribute:: debug_instrumentation
+       :type: bool
+
+         If set to ``True``, the code of modules or functions instrumented by typeguard
+         is printed to ``sys.stderr`` after the instrumentation is done
+
+         Requires Python 3.9 or newer.
+
+         Default: ``False``
+    """
+
+    forward_ref_policy: ForwardRefPolicy = ForwardRefPolicy.WARN
+    typecheck_fail_callback: TypeCheckFailCallback | None = None
+    collection_check_strategy: CollectionCheckStrategy = (
+        CollectionCheckStrategy.FIRST_ITEM
+    )
+    debug_instrumentation: bool = False
+
+
+global_config = TypeCheckConfiguration()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_decorators.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_decorators.py
new file mode 100644
index 00000000..cf325335
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_decorators.py
@@ -0,0 +1,235 @@
+from __future__ import annotations
+
+import ast
+import inspect
+import sys
+from collections.abc import Sequence
+from functools import partial
+from inspect import isclass, isfunction
+from types import CodeType, FrameType, FunctionType
+from typing import TYPE_CHECKING, Any, Callable, ForwardRef, TypeVar, cast, overload
+from warnings import warn
+
+from ._config import CollectionCheckStrategy, ForwardRefPolicy, global_config
+from ._exceptions import InstrumentationWarning
+from ._functions import TypeCheckFailCallback
+from ._transformer import TypeguardTransformer
+from ._utils import Unset, function_name, get_stacklevel, is_method_of, unset
+
+if TYPE_CHECKING:
+    from typeshed.stdlib.types import _Cell
+
+    _F = TypeVar("_F")
+
+    def typeguard_ignore(f: _F) -> _F:
+        """This decorator is a noop during static type-checking."""
+        return f
+
+else:
+    from typing import no_type_check as typeguard_ignore  # noqa: F401
+
+T_CallableOrType = TypeVar("T_CallableOrType", bound=Callable[..., Any])
+
+
+def make_cell(value: object) -> _Cell:
+    return (lambda: value).__closure__[0]  # type: ignore[index]
+
+
+def find_target_function(
+    new_code: CodeType, target_path: Sequence[str], firstlineno: int
+) -> CodeType | None:
+    target_name = target_path[0]
+    for const in new_code.co_consts:
+        if isinstance(const, CodeType):
+            if const.co_name == target_name:
+                if const.co_firstlineno == firstlineno:
+                    return const
+                elif len(target_path) > 1:
+                    target_code = find_target_function(
+                        const, target_path[1:], firstlineno
+                    )
+                    if target_code:
+                        return target_code
+
+    return None
+
+
+def instrument(f: T_CallableOrType) -> FunctionType | str:
+    if not getattr(f, "__code__", None):
+        return "no code associated"
+    elif not getattr(f, "__module__", None):
+        return "__module__ attribute is not set"
+    elif f.__code__.co_filename == "<stdin>":
+        return "cannot instrument functions defined in a REPL"
+    elif hasattr(f, "__wrapped__"):
+        return (
+            "@typechecked only supports instrumenting functions wrapped with "
+            "@classmethod, @staticmethod or @property"
+        )
+
+    target_path = [item for item in f.__qualname__.split(".") if item != "<locals>"]
+    module_source = inspect.getsource(sys.modules[f.__module__])
+    module_ast = ast.parse(module_source)
+    instrumentor = TypeguardTransformer(target_path, f.__code__.co_firstlineno)
+    instrumentor.visit(module_ast)
+
+    if not instrumentor.target_node or instrumentor.target_lineno is None:
+        return "instrumentor did not find the target function"
+
+    module_code = compile(module_ast, f.__code__.co_filename, "exec", dont_inherit=True)
+    new_code = find_target_function(
+        module_code, target_path, instrumentor.target_lineno
+    )
+    if not new_code:
+        return "cannot find the target function in the AST"
+
+    if global_config.debug_instrumentation and sys.version_info >= (3, 9):
+        # Find the matching AST node, then unparse it to source and print to stdout
+        print(
+            f"Source code of {f.__qualname__}() after instrumentation:"
+            "\n----------------------------------------------",
+            file=sys.stderr,
+        )
+        print(ast.unparse(instrumentor.target_node), file=sys.stderr)
+        print(
+            "----------------------------------------------",
+            file=sys.stderr,
+        )
+
+    closure = f.__closure__
+    if new_code.co_freevars != f.__code__.co_freevars:
+        # Create a new closure and find values for the new free variables
+        frame = cast(FrameType, inspect.currentframe())
+        frame = cast(FrameType, frame.f_back)
+        frame_locals = cast(FrameType, frame.f_back).f_locals
+        cells: list[_Cell] = []
+        for key in new_code.co_freevars:
+            if key in instrumentor.names_used_in_annotations:
+                # Find the value and make a new cell from it
+                value = frame_locals.get(key) or ForwardRef(key)
+                cells.append(make_cell(value))
+            else:
+                # Reuse the cell from the existing closure
+                assert f.__closure__
+                cells.append(f.__closure__[f.__code__.co_freevars.index(key)])
+
+        closure = tuple(cells)
+
+    new_function = FunctionType(new_code, f.__globals__, f.__name__, closure=closure)
+    new_function.__module__ = f.__module__
+    new_function.__name__ = f.__name__
+    new_function.__qualname__ = f.__qualname__
+    new_function.__annotations__ = f.__annotations__
+    new_function.__doc__ = f.__doc__
+    new_function.__defaults__ = f.__defaults__
+    new_function.__kwdefaults__ = f.__kwdefaults__
+    return new_function
+
+
+@overload
+def typechecked(
+    *,
+    forward_ref_policy: ForwardRefPolicy | Unset = unset,
+    typecheck_fail_callback: TypeCheckFailCallback | Unset = unset,
+    collection_check_strategy: CollectionCheckStrategy | Unset = unset,
+    debug_instrumentation: bool | Unset = unset,
+) -> Callable[[T_CallableOrType], T_CallableOrType]: ...
+
+
+@overload
+def typechecked(target: T_CallableOrType) -> T_CallableOrType: ...
+
+
+def typechecked(
+    target: T_CallableOrType | None = None,
+    *,
+    forward_ref_policy: ForwardRefPolicy | Unset = unset,
+    typecheck_fail_callback: TypeCheckFailCallback | Unset = unset,
+    collection_check_strategy: CollectionCheckStrategy | Unset = unset,
+    debug_instrumentation: bool | Unset = unset,
+) -> Any:
+    """
+    Instrument the target function to perform run-time type checking.
+
+    This decorator recompiles the target function, injecting code to type check
+    arguments, return values, yield values (excluding ``yield from``) and assignments to
+    annotated local variables.
+
+    This can also be used as a class decorator. This will instrument all type annotated
+    methods, including :func:`@classmethod <classmethod>`,
+    :func:`@staticmethod <staticmethod>`,  and :class:`@property <property>` decorated
+    methods in the class.
+
+    .. note:: When Python is run in optimized mode (``-O`` or ``-OO``, this decorator
+        is a no-op). This is a feature meant for selectively introducing type checking
+        into a code base where the checks aren't meant to be run in production.
+
+    :param target: the function or class to enable type checking for
+    :param forward_ref_policy: override for
+        :attr:`.TypeCheckConfiguration.forward_ref_policy`
+    :param typecheck_fail_callback: override for
+        :attr:`.TypeCheckConfiguration.typecheck_fail_callback`
+    :param collection_check_strategy: override for
+        :attr:`.TypeCheckConfiguration.collection_check_strategy`
+    :param debug_instrumentation: override for
+        :attr:`.TypeCheckConfiguration.debug_instrumentation`
+
+    """
+    if target is None:
+        return partial(
+            typechecked,
+            forward_ref_policy=forward_ref_policy,
+            typecheck_fail_callback=typecheck_fail_callback,
+            collection_check_strategy=collection_check_strategy,
+            debug_instrumentation=debug_instrumentation,
+        )
+
+    if not __debug__:
+        return target
+
+    if isclass(target):
+        for key, attr in target.__dict__.items():
+            if is_method_of(attr, target):
+                retval = instrument(attr)
+                if isfunction(retval):
+                    setattr(target, key, retval)
+            elif isinstance(attr, (classmethod, staticmethod)):
+                if is_method_of(attr.__func__, target):
+                    retval = instrument(attr.__func__)
+                    if isfunction(retval):
+                        wrapper = attr.__class__(retval)
+                        setattr(target, key, wrapper)
+            elif isinstance(attr, property):
+                kwargs: dict[str, Any] = dict(doc=attr.__doc__)
+                for name in ("fset", "fget", "fdel"):
+                    property_func = kwargs[name] = getattr(attr, name)
+                    if is_method_of(property_func, target):
+                        retval = instrument(property_func)
+                        if isfunction(retval):
+                            kwargs[name] = retval
+
+                setattr(target, key, attr.__class__(**kwargs))
+
+        return target
+
+    # Find either the first Python wrapper or the actual function
+    wrapper_class: (
+        type[classmethod[Any, Any, Any]] | type[staticmethod[Any, Any]] | None
+    ) = None
+    if isinstance(target, (classmethod, staticmethod)):
+        wrapper_class = target.__class__
+        target = target.__func__
+
+    retval = instrument(target)
+    if isinstance(retval, str):
+        warn(
+            f"{retval} -- not typechecking {function_name(target)}",
+            InstrumentationWarning,
+            stacklevel=get_stacklevel(),
+        )
+        return target
+
+    if wrapper_class is None:
+        return retval
+    else:
+        return wrapper_class(retval)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_exceptions.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_exceptions.py
new file mode 100644
index 00000000..625437a6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_exceptions.py
@@ -0,0 +1,42 @@
+from collections import deque
+from typing import Deque
+
+
+class TypeHintWarning(UserWarning):
+    """
+    A warning that is emitted when a type hint in string form could not be resolved to
+    an actual type.
+    """
+
+
+class TypeCheckWarning(UserWarning):
+    """Emitted by typeguard's type checkers when a type mismatch is detected."""
+
+    def __init__(self, message: str):
+        super().__init__(message)
+
+
+class InstrumentationWarning(UserWarning):
+    """Emitted when there's a problem with instrumenting a function for type checks."""
+
+    def __init__(self, message: str):
+        super().__init__(message)
+
+
+class TypeCheckError(Exception):
+    """
+    Raised by typeguard's type checkers when a type mismatch is detected.
+    """
+
+    def __init__(self, message: str):
+        super().__init__(message)
+        self._path: Deque[str] = deque()
+
+    def append_path_element(self, element: str) -> None:
+        self._path.append(element)
+
+    def __str__(self) -> str:
+        if self._path:
+            return " of ".join(self._path) + " " + str(self.args[0])
+        else:
+            return str(self.args[0])
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_functions.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_functions.py
new file mode 100644
index 00000000..28497856
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_functions.py
@@ -0,0 +1,308 @@
+from __future__ import annotations
+
+import sys
+import warnings
+from typing import Any, Callable, NoReturn, TypeVar, Union, overload
+
+from . import _suppression
+from ._checkers import BINARY_MAGIC_METHODS, check_type_internal
+from ._config import (
+    CollectionCheckStrategy,
+    ForwardRefPolicy,
+    TypeCheckConfiguration,
+)
+from ._exceptions import TypeCheckError, TypeCheckWarning
+from ._memo import TypeCheckMemo
+from ._utils import get_stacklevel, qualified_name
+
+if sys.version_info >= (3, 11):
+    from typing import Literal, Never, TypeAlias
+else:
+    from typing_extensions import Literal, Never, TypeAlias
+
+T = TypeVar("T")
+TypeCheckFailCallback: TypeAlias = Callable[[TypeCheckError, TypeCheckMemo], Any]
+
+
+@overload
+def check_type(
+    value: object,
+    expected_type: type[T],
+    *,
+    forward_ref_policy: ForwardRefPolicy = ...,
+    typecheck_fail_callback: TypeCheckFailCallback | None = ...,
+    collection_check_strategy: CollectionCheckStrategy = ...,
+) -> T: ...
+
+
+@overload
+def check_type(
+    value: object,
+    expected_type: Any,
+    *,
+    forward_ref_policy: ForwardRefPolicy = ...,
+    typecheck_fail_callback: TypeCheckFailCallback | None = ...,
+    collection_check_strategy: CollectionCheckStrategy = ...,
+) -> Any: ...
+
+
+def check_type(
+    value: object,
+    expected_type: Any,
+    *,
+    forward_ref_policy: ForwardRefPolicy = TypeCheckConfiguration().forward_ref_policy,
+    typecheck_fail_callback: TypeCheckFailCallback | None = (
+        TypeCheckConfiguration().typecheck_fail_callback
+    ),
+    collection_check_strategy: CollectionCheckStrategy = (
+        TypeCheckConfiguration().collection_check_strategy
+    ),
+) -> Any:
+    """
+    Ensure that ``value`` matches ``expected_type``.
+
+    The types from the :mod:`typing` module do not support :func:`isinstance` or
+    :func:`issubclass` so a number of type specific checks are required. This function
+    knows which checker to call for which type.
+
+    This function wraps :func:`~.check_type_internal` in the following ways:
+
+    * Respects type checking suppression (:func:`~.suppress_type_checks`)
+    * Forms a :class:`~.TypeCheckMemo` from the current stack frame
+    * Calls the configured type check fail callback if the check fails
+
+    Note that this function is independent of the globally shared configuration in
+    :data:`typeguard.config`. This means that usage within libraries is safe from being
+    affected configuration changes made by other libraries or by the integrating
+    application. Instead, configuration options have the same default values as their
+    corresponding fields in :class:`TypeCheckConfiguration`.
+
+    :param value: value to be checked against ``expected_type``
+    :param expected_type: a class or generic type instance, or a tuple of such things
+    :param forward_ref_policy: see :attr:`TypeCheckConfiguration.forward_ref_policy`
+    :param typecheck_fail_callback:
+        see :attr`TypeCheckConfiguration.typecheck_fail_callback`
+    :param collection_check_strategy:
+        see :attr:`TypeCheckConfiguration.collection_check_strategy`
+    :return: ``value``, unmodified
+    :raises TypeCheckError: if there is a type mismatch
+
+    """
+    if type(expected_type) is tuple:
+        expected_type = Union[expected_type]
+
+    config = TypeCheckConfiguration(
+        forward_ref_policy=forward_ref_policy,
+        typecheck_fail_callback=typecheck_fail_callback,
+        collection_check_strategy=collection_check_strategy,
+    )
+
+    if _suppression.type_checks_suppressed or expected_type is Any:
+        return value
+
+    frame = sys._getframe(1)
+    memo = TypeCheckMemo(frame.f_globals, frame.f_locals, config=config)
+    try:
+        check_type_internal(value, expected_type, memo)
+    except TypeCheckError as exc:
+        exc.append_path_element(qualified_name(value, add_class_prefix=True))
+        if config.typecheck_fail_callback:
+            config.typecheck_fail_callback(exc, memo)
+        else:
+            raise
+
+    return value
+
+
+def check_argument_types(
+    func_name: str,
+    arguments: dict[str, tuple[Any, Any]],
+    memo: TypeCheckMemo,
+) -> Literal[True]:
+    if _suppression.type_checks_suppressed:
+        return True
+
+    for argname, (value, annotation) in arguments.items():
+        if annotation is NoReturn or annotation is Never:
+            exc = TypeCheckError(
+                f"{func_name}() was declared never to be called but it was"
+            )
+            if memo.config.typecheck_fail_callback:
+                memo.config.typecheck_fail_callback(exc, memo)
+            else:
+                raise exc
+
+        try:
+            check_type_internal(value, annotation, memo)
+        except TypeCheckError as exc:
+            qualname = qualified_name(value, add_class_prefix=True)
+            exc.append_path_element(f'argument "{argname}" ({qualname})')
+            if memo.config.typecheck_fail_callback:
+                memo.config.typecheck_fail_callback(exc, memo)
+            else:
+                raise
+
+    return True
+
+
+def check_return_type(
+    func_name: str,
+    retval: T,
+    annotation: Any,
+    memo: TypeCheckMemo,
+) -> T:
+    if _suppression.type_checks_suppressed:
+        return retval
+
+    if annotation is NoReturn or annotation is Never:
+        exc = TypeCheckError(f"{func_name}() was declared never to return but it did")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise exc
+
+    try:
+        check_type_internal(retval, annotation, memo)
+    except TypeCheckError as exc:
+        # Allow NotImplemented if this is a binary magic method (__eq__() et al)
+        if retval is NotImplemented and annotation is bool:
+            # This does (and cannot) not check if it's actually a method
+            func_name = func_name.rsplit(".", 1)[-1]
+            if func_name in BINARY_MAGIC_METHODS:
+                return retval
+
+        qualname = qualified_name(retval, add_class_prefix=True)
+        exc.append_path_element(f"the return value ({qualname})")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise
+
+    return retval
+
+
+def check_send_type(
+    func_name: str,
+    sendval: T,
+    annotation: Any,
+    memo: TypeCheckMemo,
+) -> T:
+    if _suppression.type_checks_suppressed:
+        return sendval
+
+    if annotation is NoReturn or annotation is Never:
+        exc = TypeCheckError(
+            f"{func_name}() was declared never to be sent a value to but it was"
+        )
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise exc
+
+    try:
+        check_type_internal(sendval, annotation, memo)
+    except TypeCheckError as exc:
+        qualname = qualified_name(sendval, add_class_prefix=True)
+        exc.append_path_element(f"the value sent to generator ({qualname})")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise
+
+    return sendval
+
+
+def check_yield_type(
+    func_name: str,
+    yieldval: T,
+    annotation: Any,
+    memo: TypeCheckMemo,
+) -> T:
+    if _suppression.type_checks_suppressed:
+        return yieldval
+
+    if annotation is NoReturn or annotation is Never:
+        exc = TypeCheckError(f"{func_name}() was declared never to yield but it did")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise exc
+
+    try:
+        check_type_internal(yieldval, annotation, memo)
+    except TypeCheckError as exc:
+        qualname = qualified_name(yieldval, add_class_prefix=True)
+        exc.append_path_element(f"the yielded value ({qualname})")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise
+
+    return yieldval
+
+
+def check_variable_assignment(
+    value: object, varname: str, annotation: Any, memo: TypeCheckMemo
+) -> Any:
+    if _suppression.type_checks_suppressed:
+        return value
+
+    try:
+        check_type_internal(value, annotation, memo)
+    except TypeCheckError as exc:
+        qualname = qualified_name(value, add_class_prefix=True)
+        exc.append_path_element(f"value assigned to {varname} ({qualname})")
+        if memo.config.typecheck_fail_callback:
+            memo.config.typecheck_fail_callback(exc, memo)
+        else:
+            raise
+
+    return value
+
+
+def check_multi_variable_assignment(
+    value: Any, targets: list[dict[str, Any]], memo: TypeCheckMemo
+) -> Any:
+    if max(len(target) for target in targets) == 1:
+        iterated_values = [value]
+    else:
+        iterated_values = list(value)
+
+    if not _suppression.type_checks_suppressed:
+        for expected_types in targets:
+            value_index = 0
+            for ann_index, (varname, expected_type) in enumerate(
+                expected_types.items()
+            ):
+                if varname.startswith("*"):
+                    varname = varname[1:]
+                    keys_left = len(expected_types) - 1 - ann_index
+                    next_value_index = len(iterated_values) - keys_left
+                    obj: object = iterated_values[value_index:next_value_index]
+                    value_index = next_value_index
+                else:
+                    obj = iterated_values[value_index]
+                    value_index += 1
+
+                try:
+                    check_type_internal(obj, expected_type, memo)
+                except TypeCheckError as exc:
+                    qualname = qualified_name(obj, add_class_prefix=True)
+                    exc.append_path_element(f"value assigned to {varname} ({qualname})")
+                    if memo.config.typecheck_fail_callback:
+                        memo.config.typecheck_fail_callback(exc, memo)
+                    else:
+                        raise
+
+    return iterated_values[0] if len(iterated_values) == 1 else iterated_values
+
+
+def warn_on_error(exc: TypeCheckError, memo: TypeCheckMemo) -> None:
+    """
+    Emit a warning on a type mismatch.
+
+    This is intended to be used as an error handler in
+    :attr:`TypeCheckConfiguration.typecheck_fail_callback`.
+
+    """
+    warnings.warn(TypeCheckWarning(str(exc)), stacklevel=get_stacklevel())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_importhook.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_importhook.py
new file mode 100644
index 00000000..8590540a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_importhook.py
@@ -0,0 +1,213 @@
+from __future__ import annotations
+
+import ast
+import sys
+import types
+from collections.abc import Callable, Iterable
+from importlib.abc import MetaPathFinder
+from importlib.machinery import ModuleSpec, SourceFileLoader
+from importlib.util import cache_from_source, decode_source
+from inspect import isclass
+from os import PathLike
+from types import CodeType, ModuleType, TracebackType
+from typing import Sequence, TypeVar
+from unittest.mock import patch
+
+from ._config import global_config
+from ._transformer import TypeguardTransformer
+
+if sys.version_info >= (3, 12):
+    from collections.abc import Buffer
+else:
+    from typing_extensions import Buffer
+
+if sys.version_info >= (3, 11):
+    from typing import ParamSpec
+else:
+    from typing_extensions import ParamSpec
+
+if sys.version_info >= (3, 10):
+    from importlib.metadata import PackageNotFoundError, version
+else:
+    from importlib_metadata import PackageNotFoundError, version
+
+try:
+    OPTIMIZATION = "typeguard" + "".join(version("typeguard").split(".")[:3])
+except PackageNotFoundError:
+    OPTIMIZATION = "typeguard"
+
+P = ParamSpec("P")
+T = TypeVar("T")
+
+
+# The name of this function is magical
+def _call_with_frames_removed(
+    f: Callable[P, T], *args: P.args, **kwargs: P.kwargs
+) -> T:
+    return f(*args, **kwargs)
+
+
+def optimized_cache_from_source(path: str, debug_override: bool | None = None) -> str:
+    return cache_from_source(path, debug_override, optimization=OPTIMIZATION)
+
+
+class TypeguardLoader(SourceFileLoader):
+    @staticmethod
+    def source_to_code(
+        data: Buffer | str | ast.Module | ast.Expression | ast.Interactive,
+        path: Buffer | str | PathLike[str] = "<string>",
+    ) -> CodeType:
+        if isinstance(data, (ast.Module, ast.Expression, ast.Interactive)):
+            tree = data
+        else:
+            if isinstance(data, str):
+                source = data
+            else:
+                source = decode_source(data)
+
+            tree = _call_with_frames_removed(
+                ast.parse,
+                source,
+                path,
+                "exec",
+            )
+
+        tree = TypeguardTransformer().visit(tree)
+        ast.fix_missing_locations(tree)
+
+        if global_config.debug_instrumentation and sys.version_info >= (3, 9):
+            print(
+                f"Source code of {path!r} after instrumentation:\n"
+                "----------------------------------------------",
+                file=sys.stderr,
+            )
+            print(ast.unparse(tree), file=sys.stderr)
+            print("----------------------------------------------", file=sys.stderr)
+
+        return _call_with_frames_removed(
+            compile, tree, path, "exec", 0, dont_inherit=True
+        )
+
+    def exec_module(self, module: ModuleType) -> None:
+        # Use a custom optimization marker – the import lock should make this monkey
+        # patch safe
+        with patch(
+            "importlib._bootstrap_external.cache_from_source",
+            optimized_cache_from_source,
+        ):
+            super().exec_module(module)
+
+
+class TypeguardFinder(MetaPathFinder):
+    """
+    Wraps another path finder and instruments the module with
+    :func:`@typechecked <typeguard.typechecked>` if :meth:`should_instrument` returns
+    ``True``.
+
+    Should not be used directly, but rather via :func:`~.install_import_hook`.
+
+    .. versionadded:: 2.6
+    """
+
+    def __init__(self, packages: list[str] | None, original_pathfinder: MetaPathFinder):
+        self.packages = packages
+        self._original_pathfinder = original_pathfinder
+
+    def find_spec(
+        self,
+        fullname: str,
+        path: Sequence[str] | None,
+        target: types.ModuleType | None = None,
+    ) -> ModuleSpec | None:
+        if self.should_instrument(fullname):
+            spec = self._original_pathfinder.find_spec(fullname, path, target)
+            if spec is not None and isinstance(spec.loader, SourceFileLoader):
+                spec.loader = TypeguardLoader(spec.loader.name, spec.loader.path)
+                return spec
+
+        return None
+
+    def should_instrument(self, module_name: str) -> bool:
+        """
+        Determine whether the module with the given name should be instrumented.
+
+        :param module_name: full name of the module that is about to be imported (e.g.
+            ``xyz.abc``)
+
+        """
+        if self.packages is None:
+            return True
+
+        for package in self.packages:
+            if module_name == package or module_name.startswith(package + "."):
+                return True
+
+        return False
+
+
+class ImportHookManager:
+    """
+    A handle that can be used to uninstall the Typeguard import hook.
+    """
+
+    def __init__(self, hook: MetaPathFinder):
+        self.hook = hook
+
+    def __enter__(self) -> None:
+        pass
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException],
+        exc_val: BaseException,
+        exc_tb: TracebackType,
+    ) -> None:
+        self.uninstall()
+
+    def uninstall(self) -> None:
+        """Uninstall the import hook."""
+        try:
+            sys.meta_path.remove(self.hook)
+        except ValueError:
+            pass  # already removed
+
+
+def install_import_hook(
+    packages: Iterable[str] | None = None,
+    *,
+    cls: type[TypeguardFinder] = TypeguardFinder,
+) -> ImportHookManager:
+    """
+    Install an import hook that instruments functions for automatic type checking.
+
+    This only affects modules loaded **after** this hook has been installed.
+
+    :param packages: an iterable of package names to instrument, or ``None`` to
+        instrument all packages
+    :param cls: a custom meta path finder class
+    :return: a context manager that uninstalls the hook on exit (or when you call
+        ``.uninstall()``)
+
+    .. versionadded:: 2.6
+
+    """
+    if packages is None:
+        target_packages: list[str] | None = None
+    elif isinstance(packages, str):
+        target_packages = [packages]
+    else:
+        target_packages = list(packages)
+
+    for finder in sys.meta_path:
+        if (
+            isclass(finder)
+            and finder.__name__ == "PathFinder"
+            and hasattr(finder, "find_spec")
+        ):
+            break
+    else:
+        raise RuntimeError("Cannot find a PathFinder in sys.meta_path")
+
+    hook = cls(target_packages, finder)
+    sys.meta_path.insert(0, hook)
+    return ImportHookManager(hook)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_memo.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_memo.py
new file mode 100644
index 00000000..1d0d80c6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_memo.py
@@ -0,0 +1,48 @@
+from __future__ import annotations
+
+from typing import Any
+
+from typeguard._config import TypeCheckConfiguration, global_config
+
+
+class TypeCheckMemo:
+    """
+    Contains information necessary for type checkers to do their work.
+
+    .. attribute:: globals
+       :type: dict[str, Any]
+
+        Dictionary of global variables to use for resolving forward references.
+
+    .. attribute:: locals
+       :type: dict[str, Any]
+
+        Dictionary of local variables to use for resolving forward references.
+
+    .. attribute:: self_type
+       :type: type | None
+
+        When running type checks within an instance method or class method, this is the
+        class object that the first argument (usually named ``self`` or ``cls``) refers
+        to.
+
+    .. attribute:: config
+       :type: TypeCheckConfiguration
+
+         Contains the configuration for a particular set of type checking operations.
+    """
+
+    __slots__ = "globals", "locals", "self_type", "config"
+
+    def __init__(
+        self,
+        globals: dict[str, Any],
+        locals: dict[str, Any],
+        *,
+        self_type: type | None = None,
+        config: TypeCheckConfiguration = global_config,
+    ):
+        self.globals = globals
+        self.locals = locals
+        self.self_type = self_type
+        self.config = config
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_pytest_plugin.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_pytest_plugin.py
new file mode 100644
index 00000000..7b2f494e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_pytest_plugin.py
@@ -0,0 +1,127 @@
+from __future__ import annotations
+
+import sys
+import warnings
+from typing import TYPE_CHECKING, Any, Literal
+
+from typeguard._config import CollectionCheckStrategy, ForwardRefPolicy, global_config
+from typeguard._exceptions import InstrumentationWarning
+from typeguard._importhook import install_import_hook
+from typeguard._utils import qualified_name, resolve_reference
+
+if TYPE_CHECKING:
+    from pytest import Config, Parser
+
+
+def pytest_addoption(parser: Parser) -> None:
+    def add_ini_option(
+        opt_type: (
+            Literal["string", "paths", "pathlist", "args", "linelist", "bool"] | None
+        ),
+    ) -> None:
+        parser.addini(
+            group.options[-1].names()[0][2:],
+            group.options[-1].attrs()["help"],
+            opt_type,
+        )
+
+    group = parser.getgroup("typeguard")
+    group.addoption(
+        "--typeguard-packages",
+        action="store",
+        help="comma separated name list of packages and modules to instrument for "
+        "type checking, or :all: to instrument all modules loaded after typeguard",
+    )
+    add_ini_option("linelist")
+
+    group.addoption(
+        "--typeguard-debug-instrumentation",
+        action="store_true",
+        help="print all instrumented code to stderr",
+    )
+    add_ini_option("bool")
+
+    group.addoption(
+        "--typeguard-typecheck-fail-callback",
+        action="store",
+        help=(
+            "a module:varname (e.g. typeguard:warn_on_error) reference to a function "
+            "that is called (with the exception, and memo object as arguments) to "
+            "handle a TypeCheckError"
+        ),
+    )
+    add_ini_option("string")
+
+    group.addoption(
+        "--typeguard-forward-ref-policy",
+        action="store",
+        choices=list(ForwardRefPolicy.__members__),
+        help=(
+            "determines how to deal with unresolveable forward references in type "
+            "annotations"
+        ),
+    )
+    add_ini_option("string")
+
+    group.addoption(
+        "--typeguard-collection-check-strategy",
+        action="store",
+        choices=list(CollectionCheckStrategy.__members__),
+        help="determines how thoroughly to check collections (list, dict, etc)",
+    )
+    add_ini_option("string")
+
+
+def pytest_configure(config: Config) -> None:
+    def getoption(name: str) -> Any:
+        return config.getoption(name.replace("-", "_")) or config.getini(name)
+
+    packages: list[str] | None = []
+    if packages_option := config.getoption("typeguard_packages"):
+        packages = [pkg.strip() for pkg in packages_option.split(",")]
+    elif packages_ini := config.getini("typeguard-packages"):
+        packages = packages_ini
+
+    if packages:
+        if packages == [":all:"]:
+            packages = None
+        else:
+            already_imported_packages = sorted(
+                package for package in packages if package in sys.modules
+            )
+            if already_imported_packages:
+                warnings.warn(
+                    f"typeguard cannot check these packages because they are already "
+                    f"imported: {', '.join(already_imported_packages)}",
+                    InstrumentationWarning,
+                    stacklevel=1,
+                )
+
+        install_import_hook(packages=packages)
+
+    debug_option = getoption("typeguard-debug-instrumentation")
+    if debug_option:
+        global_config.debug_instrumentation = True
+
+    fail_callback_option = getoption("typeguard-typecheck-fail-callback")
+    if fail_callback_option:
+        callback = resolve_reference(fail_callback_option)
+        if not callable(callback):
+            raise TypeError(
+                f"{fail_callback_option} ({qualified_name(callback.__class__)}) is not "
+                f"a callable"
+            )
+
+        global_config.typecheck_fail_callback = callback
+
+    forward_ref_policy_option = getoption("typeguard-forward-ref-policy")
+    if forward_ref_policy_option:
+        forward_ref_policy = ForwardRefPolicy.__members__[forward_ref_policy_option]
+        global_config.forward_ref_policy = forward_ref_policy
+
+    collection_check_strategy_option = getoption("typeguard-collection-check-strategy")
+    if collection_check_strategy_option:
+        collection_check_strategy = CollectionCheckStrategy.__members__[
+            collection_check_strategy_option
+        ]
+        global_config.collection_check_strategy = collection_check_strategy
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_suppression.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_suppression.py
new file mode 100644
index 00000000..bbbfbfbe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_suppression.py
@@ -0,0 +1,86 @@
+from __future__ import annotations
+
+import sys
+from collections.abc import Callable, Generator
+from contextlib import contextmanager
+from functools import update_wrapper
+from threading import Lock
+from typing import ContextManager, TypeVar, overload
+
+if sys.version_info >= (3, 10):
+    from typing import ParamSpec
+else:
+    from typing_extensions import ParamSpec
+
+P = ParamSpec("P")
+T = TypeVar("T")
+
+type_checks_suppressed = 0
+type_checks_suppress_lock = Lock()
+
+
+@overload
+def suppress_type_checks(func: Callable[P, T]) -> Callable[P, T]: ...
+
+
+@overload
+def suppress_type_checks() -> ContextManager[None]: ...
+
+
+def suppress_type_checks(
+    func: Callable[P, T] | None = None,
+) -> Callable[P, T] | ContextManager[None]:
+    """
+    Temporarily suppress all type checking.
+
+    This function has two operating modes, based on how it's used:
+
+    #. as a context manager (``with suppress_type_checks(): ...``)
+    #. as a decorator (``@suppress_type_checks``)
+
+    When used as a context manager, :func:`check_type` and any automatically
+    instrumented functions skip the actual type checking. These context managers can be
+    nested.
+
+    When used as a decorator, all type checking is suppressed while the function is
+    running.
+
+    Type checking will resume once no more context managers are active and no decorated
+    functions are running.
+
+    Both operating modes are thread-safe.
+
+    """
+
+    def wrapper(*args: P.args, **kwargs: P.kwargs) -> T:
+        global type_checks_suppressed
+
+        with type_checks_suppress_lock:
+            type_checks_suppressed += 1
+
+        assert func is not None
+        try:
+            return func(*args, **kwargs)
+        finally:
+            with type_checks_suppress_lock:
+                type_checks_suppressed -= 1
+
+    def cm() -> Generator[None, None, None]:
+        global type_checks_suppressed
+
+        with type_checks_suppress_lock:
+            type_checks_suppressed += 1
+
+        try:
+            yield
+        finally:
+            with type_checks_suppress_lock:
+                type_checks_suppressed -= 1
+
+    if func is None:
+        # Context manager mode
+        return contextmanager(cm)()
+    else:
+        # Decorator mode
+        update_wrapper(wrapper, func)
+        return wrapper
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_transformer.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_transformer.py
new file mode 100644
index 00000000..13ac3630
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_transformer.py
@@ -0,0 +1,1229 @@
+from __future__ import annotations
+
+import ast
+import builtins
+import sys
+import typing
+from ast import (
+    AST,
+    Add,
+    AnnAssign,
+    Assign,
+    AsyncFunctionDef,
+    Attribute,
+    AugAssign,
+    BinOp,
+    BitAnd,
+    BitOr,
+    BitXor,
+    Call,
+    ClassDef,
+    Constant,
+    Dict,
+    Div,
+    Expr,
+    Expression,
+    FloorDiv,
+    FunctionDef,
+    If,
+    Import,
+    ImportFrom,
+    Index,
+    List,
+    Load,
+    LShift,
+    MatMult,
+    Mod,
+    Module,
+    Mult,
+    Name,
+    NamedExpr,
+    NodeTransformer,
+    NodeVisitor,
+    Pass,
+    Pow,
+    Return,
+    RShift,
+    Starred,
+    Store,
+    Sub,
+    Subscript,
+    Tuple,
+    Yield,
+    YieldFrom,
+    alias,
+    copy_location,
+    expr,
+    fix_missing_locations,
+    keyword,
+    walk,
+)
+from collections import defaultdict
+from collections.abc import Generator, Sequence
+from contextlib import contextmanager
+from copy import deepcopy
+from dataclasses import dataclass, field
+from typing import Any, ClassVar, cast, overload
+
+generator_names = (
+    "typing.Generator",
+    "collections.abc.Generator",
+    "typing.Iterator",
+    "collections.abc.Iterator",
+    "typing.Iterable",
+    "collections.abc.Iterable",
+    "typing.AsyncIterator",
+    "collections.abc.AsyncIterator",
+    "typing.AsyncIterable",
+    "collections.abc.AsyncIterable",
+    "typing.AsyncGenerator",
+    "collections.abc.AsyncGenerator",
+)
+anytype_names = (
+    "typing.Any",
+    "typing_extensions.Any",
+)
+literal_names = (
+    "typing.Literal",
+    "typing_extensions.Literal",
+)
+annotated_names = (
+    "typing.Annotated",
+    "typing_extensions.Annotated",
+)
+ignore_decorators = (
+    "typing.no_type_check",
+    "typeguard.typeguard_ignore",
+)
+aug_assign_functions = {
+    Add: "iadd",
+    Sub: "isub",
+    Mult: "imul",
+    MatMult: "imatmul",
+    Div: "itruediv",
+    FloorDiv: "ifloordiv",
+    Mod: "imod",
+    Pow: "ipow",
+    LShift: "ilshift",
+    RShift: "irshift",
+    BitAnd: "iand",
+    BitXor: "ixor",
+    BitOr: "ior",
+}
+
+
+@dataclass
+class TransformMemo:
+    node: Module | ClassDef | FunctionDef | AsyncFunctionDef | None
+    parent: TransformMemo | None
+    path: tuple[str, ...]
+    joined_path: Constant = field(init=False)
+    return_annotation: expr | None = None
+    yield_annotation: expr | None = None
+    send_annotation: expr | None = None
+    is_async: bool = False
+    local_names: set[str] = field(init=False, default_factory=set)
+    imported_names: dict[str, str] = field(init=False, default_factory=dict)
+    ignored_names: set[str] = field(init=False, default_factory=set)
+    load_names: defaultdict[str, dict[str, Name]] = field(
+        init=False, default_factory=lambda: defaultdict(dict)
+    )
+    has_yield_expressions: bool = field(init=False, default=False)
+    has_return_expressions: bool = field(init=False, default=False)
+    memo_var_name: Name | None = field(init=False, default=None)
+    should_instrument: bool = field(init=False, default=True)
+    variable_annotations: dict[str, expr] = field(init=False, default_factory=dict)
+    configuration_overrides: dict[str, Any] = field(init=False, default_factory=dict)
+    code_inject_index: int = field(init=False, default=0)
+
+    def __post_init__(self) -> None:
+        elements: list[str] = []
+        memo = self
+        while isinstance(memo.node, (ClassDef, FunctionDef, AsyncFunctionDef)):
+            elements.insert(0, memo.node.name)
+            if not memo.parent:
+                break
+
+            memo = memo.parent
+            if isinstance(memo.node, (FunctionDef, AsyncFunctionDef)):
+                elements.insert(0, "<locals>")
+
+        self.joined_path = Constant(".".join(elements))
+
+        # Figure out where to insert instrumentation code
+        if self.node:
+            for index, child in enumerate(self.node.body):
+                if isinstance(child, ImportFrom) and child.module == "__future__":
+                    # (module only) __future__ imports must come first
+                    continue
+                elif (
+                    isinstance(child, Expr)
+                    and isinstance(child.value, Constant)
+                    and isinstance(child.value.value, str)
+                ):
+                    continue  # docstring
+
+                self.code_inject_index = index
+                break
+
+    def get_unused_name(self, name: str) -> str:
+        memo: TransformMemo | None = self
+        while memo is not None:
+            if name in memo.local_names:
+                memo = self
+                name += "_"
+            else:
+                memo = memo.parent
+
+        self.local_names.add(name)
+        return name
+
+    def is_ignored_name(self, expression: expr | Expr | None) -> bool:
+        top_expression = (
+            expression.value if isinstance(expression, Expr) else expression
+        )
+
+        if isinstance(top_expression, Attribute) and isinstance(
+            top_expression.value, Name
+        ):
+            name = top_expression.value.id
+        elif isinstance(top_expression, Name):
+            name = top_expression.id
+        else:
+            return False
+
+        memo: TransformMemo | None = self
+        while memo is not None:
+            if name in memo.ignored_names:
+                return True
+
+            memo = memo.parent
+
+        return False
+
+    def get_memo_name(self) -> Name:
+        if not self.memo_var_name:
+            self.memo_var_name = Name(id="memo", ctx=Load())
+
+        return self.memo_var_name
+
+    def get_import(self, module: str, name: str) -> Name:
+        if module in self.load_names and name in self.load_names[module]:
+            return self.load_names[module][name]
+
+        qualified_name = f"{module}.{name}"
+        if name in self.imported_names and self.imported_names[name] == qualified_name:
+            return Name(id=name, ctx=Load())
+
+        alias = self.get_unused_name(name)
+        node = self.load_names[module][name] = Name(id=alias, ctx=Load())
+        self.imported_names[name] = qualified_name
+        return node
+
+    def insert_imports(self, node: Module | FunctionDef | AsyncFunctionDef) -> None:
+        """Insert imports needed by injected code."""
+        if not self.load_names:
+            return
+
+        # Insert imports after any "from __future__ ..." imports and any docstring
+        for modulename, names in self.load_names.items():
+            aliases = [
+                alias(orig_name, new_name.id if orig_name != new_name.id else None)
+                for orig_name, new_name in sorted(names.items())
+            ]
+            node.body.insert(self.code_inject_index, ImportFrom(modulename, aliases, 0))
+
+    def name_matches(self, expression: expr | Expr | None, *names: str) -> bool:
+        if expression is None:
+            return False
+
+        path: list[str] = []
+        top_expression = (
+            expression.value if isinstance(expression, Expr) else expression
+        )
+
+        if isinstance(top_expression, Subscript):
+            top_expression = top_expression.value
+        elif isinstance(top_expression, Call):
+            top_expression = top_expression.func
+
+        while isinstance(top_expression, Attribute):
+            path.insert(0, top_expression.attr)
+            top_expression = top_expression.value
+
+        if not isinstance(top_expression, Name):
+            return False
+
+        if top_expression.id in self.imported_names:
+            translated = self.imported_names[top_expression.id]
+        elif hasattr(builtins, top_expression.id):
+            translated = "builtins." + top_expression.id
+        else:
+            translated = top_expression.id
+
+        path.insert(0, translated)
+        joined_path = ".".join(path)
+        if joined_path in names:
+            return True
+        elif self.parent:
+            return self.parent.name_matches(expression, *names)
+        else:
+            return False
+
+    def get_config_keywords(self) -> list[keyword]:
+        if self.parent and isinstance(self.parent.node, ClassDef):
+            overrides = self.parent.configuration_overrides.copy()
+        else:
+            overrides = {}
+
+        overrides.update(self.configuration_overrides)
+        return [keyword(key, value) for key, value in overrides.items()]
+
+
+class NameCollector(NodeVisitor):
+    def __init__(self) -> None:
+        self.names: set[str] = set()
+
+    def visit_Import(self, node: Import) -> None:
+        for name in node.names:
+            self.names.add(name.asname or name.name)
+
+    def visit_ImportFrom(self, node: ImportFrom) -> None:
+        for name in node.names:
+            self.names.add(name.asname or name.name)
+
+    def visit_Assign(self, node: Assign) -> None:
+        for target in node.targets:
+            if isinstance(target, Name):
+                self.names.add(target.id)
+
+    def visit_NamedExpr(self, node: NamedExpr) -> Any:
+        if isinstance(node.target, Name):
+            self.names.add(node.target.id)
+
+    def visit_FunctionDef(self, node: FunctionDef) -> None:
+        pass
+
+    def visit_ClassDef(self, node: ClassDef) -> None:
+        pass
+
+
+class GeneratorDetector(NodeVisitor):
+    """Detects if a function node is a generator function."""
+
+    contains_yields: bool = False
+    in_root_function: bool = False
+
+    def visit_Yield(self, node: Yield) -> Any:
+        self.contains_yields = True
+
+    def visit_YieldFrom(self, node: YieldFrom) -> Any:
+        self.contains_yields = True
+
+    def visit_ClassDef(self, node: ClassDef) -> Any:
+        pass
+
+    def visit_FunctionDef(self, node: FunctionDef | AsyncFunctionDef) -> Any:
+        if not self.in_root_function:
+            self.in_root_function = True
+            self.generic_visit(node)
+            self.in_root_function = False
+
+    def visit_AsyncFunctionDef(self, node: AsyncFunctionDef) -> Any:
+        self.visit_FunctionDef(node)
+
+
+class AnnotationTransformer(NodeTransformer):
+    type_substitutions: ClassVar[dict[str, tuple[str, str]]] = {
+        "builtins.dict": ("typing", "Dict"),
+        "builtins.list": ("typing", "List"),
+        "builtins.tuple": ("typing", "Tuple"),
+        "builtins.set": ("typing", "Set"),
+        "builtins.frozenset": ("typing", "FrozenSet"),
+    }
+
+    def __init__(self, transformer: TypeguardTransformer):
+        self.transformer = transformer
+        self._memo = transformer._memo
+        self._level = 0
+
+    def visit(self, node: AST) -> Any:
+        # Don't process Literals
+        if isinstance(node, expr) and self._memo.name_matches(node, *literal_names):
+            return node
+
+        self._level += 1
+        new_node = super().visit(node)
+        self._level -= 1
+
+        if isinstance(new_node, Expression) and not hasattr(new_node, "body"):
+            return None
+
+        # Return None if this new node matches a variation of typing.Any
+        if (
+            self._level == 0
+            and isinstance(new_node, expr)
+            and self._memo.name_matches(new_node, *anytype_names)
+        ):
+            return None
+
+        return new_node
+
+    def visit_BinOp(self, node: BinOp) -> Any:
+        self.generic_visit(node)
+
+        if isinstance(node.op, BitOr):
+            # If either branch of the BinOp has been transformed to `None`, it means
+            # that a type in the union was ignored, so the entire annotation should e
+            # ignored
+            if not hasattr(node, "left") or not hasattr(node, "right"):
+                return None
+
+            # Return Any if either side is Any
+            if self._memo.name_matches(node.left, *anytype_names):
+                return node.left
+            elif self._memo.name_matches(node.right, *anytype_names):
+                return node.right
+
+            if sys.version_info < (3, 10):
+                union_name = self.transformer._get_import("typing", "Union")
+                return Subscript(
+                    value=union_name,
+                    slice=Index(
+                        Tuple(elts=[node.left, node.right], ctx=Load()), ctx=Load()
+                    ),
+                    ctx=Load(),
+                )
+
+        return node
+
+    def visit_Attribute(self, node: Attribute) -> Any:
+        if self._memo.is_ignored_name(node):
+            return None
+
+        return node
+
+    def visit_Subscript(self, node: Subscript) -> Any:
+        if self._memo.is_ignored_name(node.value):
+            return None
+
+        # The subscript of typing(_extensions).Literal can be any arbitrary string, so
+        # don't try to evaluate it as code
+        if node.slice:
+            if isinstance(node.slice, Index):
+                # Python 3.8
+                slice_value = node.slice.value  # type: ignore[attr-defined]
+            else:
+                slice_value = node.slice
+
+            if isinstance(slice_value, Tuple):
+                if self._memo.name_matches(node.value, *annotated_names):
+                    # Only treat the first argument to typing.Annotated as a potential
+                    # forward reference
+                    items = cast(
+                        typing.List[expr],
+                        [self.visit(slice_value.elts[0])] + slice_value.elts[1:],
+                    )
+                else:
+                    items = cast(
+                        typing.List[expr],
+                        [self.visit(item) for item in slice_value.elts],
+                    )
+
+                # If this is a Union and any of the items is Any, erase the entire
+                # annotation
+                if self._memo.name_matches(node.value, "typing.Union") and any(
+                    item is None
+                    or (
+                        isinstance(item, expr)
+                        and self._memo.name_matches(item, *anytype_names)
+                    )
+                    for item in items
+                ):
+                    return None
+
+                # If all items in the subscript were Any, erase the subscript entirely
+                if all(item is None for item in items):
+                    return node.value
+
+                for index, item in enumerate(items):
+                    if item is None:
+                        items[index] = self.transformer._get_import("typing", "Any")
+
+                slice_value.elts = items
+            else:
+                self.generic_visit(node)
+
+                # If the transformer erased the slice entirely, just return the node
+                # value without the subscript (unless it's Optional, in which case erase
+                # the node entirely
+                if self._memo.name_matches(
+                    node.value, "typing.Optional"
+                ) and not hasattr(node, "slice"):
+                    return None
+                if sys.version_info >= (3, 9) and not hasattr(node, "slice"):
+                    return node.value
+                elif sys.version_info < (3, 9) and not hasattr(node.slice, "value"):
+                    return node.value
+
+        return node
+
+    def visit_Name(self, node: Name) -> Any:
+        if self._memo.is_ignored_name(node):
+            return None
+
+        if sys.version_info < (3, 9):
+            for typename, substitute in self.type_substitutions.items():
+                if self._memo.name_matches(node, typename):
+                    new_node = self.transformer._get_import(*substitute)
+                    return copy_location(new_node, node)
+
+        return node
+
+    def visit_Call(self, node: Call) -> Any:
+        # Don't recurse into calls
+        return node
+
+    def visit_Constant(self, node: Constant) -> Any:
+        if isinstance(node.value, str):
+            expression = ast.parse(node.value, mode="eval")
+            new_node = self.visit(expression)
+            if new_node:
+                return copy_location(new_node.body, node)
+            else:
+                return None
+
+        return node
+
+
+class TypeguardTransformer(NodeTransformer):
+    def __init__(
+        self, target_path: Sequence[str] | None = None, target_lineno: int | None = None
+    ) -> None:
+        self._target_path = tuple(target_path) if target_path else None
+        self._memo = self._module_memo = TransformMemo(None, None, ())
+        self.names_used_in_annotations: set[str] = set()
+        self.target_node: FunctionDef | AsyncFunctionDef | None = None
+        self.target_lineno = target_lineno
+
+    def generic_visit(self, node: AST) -> AST:
+        has_non_empty_body_initially = bool(getattr(node, "body", None))
+        initial_type = type(node)
+
+        node = super().generic_visit(node)
+
+        if (
+            type(node) is initial_type
+            and has_non_empty_body_initially
+            and hasattr(node, "body")
+            and not node.body
+        ):
+            # If we have still the same node type after transformation
+            # but we've optimised it's body away, we add a `pass` statement.
+            node.body = [Pass()]
+
+        return node
+
+    @contextmanager
+    def _use_memo(
+        self, node: ClassDef | FunctionDef | AsyncFunctionDef
+    ) -> Generator[None, Any, None]:
+        new_memo = TransformMemo(node, self._memo, self._memo.path + (node.name,))
+        old_memo = self._memo
+        self._memo = new_memo
+
+        if isinstance(node, (FunctionDef, AsyncFunctionDef)):
+            new_memo.should_instrument = (
+                self._target_path is None or new_memo.path == self._target_path
+            )
+            if new_memo.should_instrument:
+                # Check if the function is a generator function
+                detector = GeneratorDetector()
+                detector.visit(node)
+
+                # Extract yield, send and return types where possible from a subscripted
+                # annotation like Generator[int, str, bool]
+                return_annotation = deepcopy(node.returns)
+                if detector.contains_yields and new_memo.name_matches(
+                    return_annotation, *generator_names
+                ):
+                    if isinstance(return_annotation, Subscript):
+                        annotation_slice = return_annotation.slice
+
+                        # Python < 3.9
+                        if isinstance(annotation_slice, Index):
+                            annotation_slice = (
+                                annotation_slice.value  # type: ignore[attr-defined]
+                            )
+
+                        if isinstance(annotation_slice, Tuple):
+                            items = annotation_slice.elts
+                        else:
+                            items = [annotation_slice]
+
+                        if len(items) > 0:
+                            new_memo.yield_annotation = self._convert_annotation(
+                                items[0]
+                            )
+
+                        if len(items) > 1:
+                            new_memo.send_annotation = self._convert_annotation(
+                                items[1]
+                            )
+
+                        if len(items) > 2:
+                            new_memo.return_annotation = self._convert_annotation(
+                                items[2]
+                            )
+                else:
+                    new_memo.return_annotation = self._convert_annotation(
+                        return_annotation
+                    )
+
+        if isinstance(node, AsyncFunctionDef):
+            new_memo.is_async = True
+
+        yield
+        self._memo = old_memo
+
+    def _get_import(self, module: str, name: str) -> Name:
+        memo = self._memo if self._target_path else self._module_memo
+        return memo.get_import(module, name)
+
+    @overload
+    def _convert_annotation(self, annotation: None) -> None: ...
+
+    @overload
+    def _convert_annotation(self, annotation: expr) -> expr: ...
+
+    def _convert_annotation(self, annotation: expr | None) -> expr | None:
+        if annotation is None:
+            return None
+
+        # Convert PEP 604 unions (x | y) and generic built-in collections where
+        # necessary, and undo forward references
+        new_annotation = cast(expr, AnnotationTransformer(self).visit(annotation))
+        if isinstance(new_annotation, expr):
+            new_annotation = ast.copy_location(new_annotation, annotation)
+
+            # Store names used in the annotation
+            names = {node.id for node in walk(new_annotation) if isinstance(node, Name)}
+            self.names_used_in_annotations.update(names)
+
+        return new_annotation
+
+    def visit_Name(self, node: Name) -> Name:
+        self._memo.local_names.add(node.id)
+        return node
+
+    def visit_Module(self, node: Module) -> Module:
+        self._module_memo = self._memo = TransformMemo(node, None, ())
+        self.generic_visit(node)
+        self._module_memo.insert_imports(node)
+
+        fix_missing_locations(node)
+        return node
+
+    def visit_Import(self, node: Import) -> Import:
+        for name in node.names:
+            self._memo.local_names.add(name.asname or name.name)
+            self._memo.imported_names[name.asname or name.name] = name.name
+
+        return node
+
+    def visit_ImportFrom(self, node: ImportFrom) -> ImportFrom:
+        for name in node.names:
+            if name.name != "*":
+                alias = name.asname or name.name
+                self._memo.local_names.add(alias)
+                self._memo.imported_names[alias] = f"{node.module}.{name.name}"
+
+        return node
+
+    def visit_ClassDef(self, node: ClassDef) -> ClassDef | None:
+        self._memo.local_names.add(node.name)
+
+        # Eliminate top level classes not belonging to the target path
+        if (
+            self._target_path is not None
+            and not self._memo.path
+            and node.name != self._target_path[0]
+        ):
+            return None
+
+        with self._use_memo(node):
+            for decorator in node.decorator_list.copy():
+                if self._memo.name_matches(decorator, "typeguard.typechecked"):
+                    # Remove the decorator to prevent duplicate instrumentation
+                    node.decorator_list.remove(decorator)
+
+                    # Store any configuration overrides
+                    if isinstance(decorator, Call) and decorator.keywords:
+                        self._memo.configuration_overrides.update(
+                            {kw.arg: kw.value for kw in decorator.keywords if kw.arg}
+                        )
+
+            self.generic_visit(node)
+            return node
+
+    def visit_FunctionDef(
+        self, node: FunctionDef | AsyncFunctionDef
+    ) -> FunctionDef | AsyncFunctionDef | None:
+        """
+        Injects type checks for function arguments, and for a return of None if the
+        function is annotated to return something else than Any or None, and the body
+        ends without an explicit "return".
+
+        """
+        self._memo.local_names.add(node.name)
+
+        # Eliminate top level functions not belonging to the target path
+        if (
+            self._target_path is not None
+            and not self._memo.path
+            and node.name != self._target_path[0]
+        ):
+            return None
+
+        # Skip instrumentation if we're instrumenting the whole module and the function
+        # contains either @no_type_check or @typeguard_ignore
+        if self._target_path is None:
+            for decorator in node.decorator_list:
+                if self._memo.name_matches(decorator, *ignore_decorators):
+                    return node
+
+        with self._use_memo(node):
+            arg_annotations: dict[str, Any] = {}
+            if self._target_path is None or self._memo.path == self._target_path:
+                # Find line number we're supposed to match against
+                if node.decorator_list:
+                    first_lineno = node.decorator_list[0].lineno
+                else:
+                    first_lineno = node.lineno
+
+                for decorator in node.decorator_list.copy():
+                    if self._memo.name_matches(decorator, "typing.overload"):
+                        # Remove overloads entirely
+                        return None
+                    elif self._memo.name_matches(decorator, "typeguard.typechecked"):
+                        # Remove the decorator to prevent duplicate instrumentation
+                        node.decorator_list.remove(decorator)
+
+                        # Store any configuration overrides
+                        if isinstance(decorator, Call) and decorator.keywords:
+                            self._memo.configuration_overrides = {
+                                kw.arg: kw.value for kw in decorator.keywords if kw.arg
+                            }
+
+                if self.target_lineno == first_lineno:
+                    assert self.target_node is None
+                    self.target_node = node
+                    if node.decorator_list:
+                        self.target_lineno = node.decorator_list[0].lineno
+                    else:
+                        self.target_lineno = node.lineno
+
+                all_args = node.args.args + node.args.kwonlyargs + node.args.posonlyargs
+
+                # Ensure that any type shadowed by the positional or keyword-only
+                # argument names are ignored in this function
+                for arg in all_args:
+                    self._memo.ignored_names.add(arg.arg)
+
+                # Ensure that any type shadowed by the variable positional argument name
+                # (e.g. "args" in *args) is ignored this function
+                if node.args.vararg:
+                    self._memo.ignored_names.add(node.args.vararg.arg)
+
+                # Ensure that any type shadowed by the variable keywrod argument name
+                # (e.g. "kwargs" in *kwargs) is ignored this function
+                if node.args.kwarg:
+                    self._memo.ignored_names.add(node.args.kwarg.arg)
+
+                for arg in all_args:
+                    annotation = self._convert_annotation(deepcopy(arg.annotation))
+                    if annotation:
+                        arg_annotations[arg.arg] = annotation
+
+                if node.args.vararg:
+                    annotation_ = self._convert_annotation(node.args.vararg.annotation)
+                    if annotation_:
+                        if sys.version_info >= (3, 9):
+                            container = Name("tuple", ctx=Load())
+                        else:
+                            container = self._get_import("typing", "Tuple")
+
+                        subscript_slice: Tuple | Index = Tuple(
+                            [
+                                annotation_,
+                                Constant(Ellipsis),
+                            ],
+                            ctx=Load(),
+                        )
+                        if sys.version_info < (3, 9):
+                            subscript_slice = Index(subscript_slice, ctx=Load())
+
+                        arg_annotations[node.args.vararg.arg] = Subscript(
+                            container, subscript_slice, ctx=Load()
+                        )
+
+                if node.args.kwarg:
+                    annotation_ = self._convert_annotation(node.args.kwarg.annotation)
+                    if annotation_:
+                        if sys.version_info >= (3, 9):
+                            container = Name("dict", ctx=Load())
+                        else:
+                            container = self._get_import("typing", "Dict")
+
+                        subscript_slice = Tuple(
+                            [
+                                Name("str", ctx=Load()),
+                                annotation_,
+                            ],
+                            ctx=Load(),
+                        )
+                        if sys.version_info < (3, 9):
+                            subscript_slice = Index(subscript_slice, ctx=Load())
+
+                        arg_annotations[node.args.kwarg.arg] = Subscript(
+                            container, subscript_slice, ctx=Load()
+                        )
+
+                if arg_annotations:
+                    self._memo.variable_annotations.update(arg_annotations)
+
+            self.generic_visit(node)
+
+            if arg_annotations:
+                annotations_dict = Dict(
+                    keys=[Constant(key) for key in arg_annotations.keys()],
+                    values=[
+                        Tuple([Name(key, ctx=Load()), annotation], ctx=Load())
+                        for key, annotation in arg_annotations.items()
+                    ],
+                )
+                func_name = self._get_import(
+                    "typeguard._functions", "check_argument_types"
+                )
+                args = [
+                    self._memo.joined_path,
+                    annotations_dict,
+                    self._memo.get_memo_name(),
+                ]
+                node.body.insert(
+                    self._memo.code_inject_index, Expr(Call(func_name, args, []))
+                )
+
+            # Add a checked "return None" to the end if there's no explicit return
+            # Skip if the return annotation is None or Any
+            if (
+                self._memo.return_annotation
+                and (not self._memo.is_async or not self._memo.has_yield_expressions)
+                and not isinstance(node.body[-1], Return)
+                and (
+                    not isinstance(self._memo.return_annotation, Constant)
+                    or self._memo.return_annotation.value is not None
+                )
+            ):
+                func_name = self._get_import(
+                    "typeguard._functions", "check_return_type"
+                )
+                return_node = Return(
+                    Call(
+                        func_name,
+                        [
+                            self._memo.joined_path,
+                            Constant(None),
+                            self._memo.return_annotation,
+                            self._memo.get_memo_name(),
+                        ],
+                        [],
+                    )
+                )
+
+                # Replace a placeholder "pass" at the end
+                if isinstance(node.body[-1], Pass):
+                    copy_location(return_node, node.body[-1])
+                    del node.body[-1]
+
+                node.body.append(return_node)
+
+            # Insert code to create the call memo, if it was ever needed for this
+            # function
+            if self._memo.memo_var_name:
+                memo_kwargs: dict[str, Any] = {}
+                if self._memo.parent and isinstance(self._memo.parent.node, ClassDef):
+                    for decorator in node.decorator_list:
+                        if (
+                            isinstance(decorator, Name)
+                            and decorator.id == "staticmethod"
+                        ):
+                            break
+                        elif (
+                            isinstance(decorator, Name)
+                            and decorator.id == "classmethod"
+                        ):
+                            memo_kwargs["self_type"] = Name(
+                                id=node.args.args[0].arg, ctx=Load()
+                            )
+                            break
+                    else:
+                        if node.args.args:
+                            if node.name == "__new__":
+                                memo_kwargs["self_type"] = Name(
+                                    id=node.args.args[0].arg, ctx=Load()
+                                )
+                            else:
+                                memo_kwargs["self_type"] = Attribute(
+                                    Name(id=node.args.args[0].arg, ctx=Load()),
+                                    "__class__",
+                                    ctx=Load(),
+                                )
+
+                # Construct the function reference
+                # Nested functions get special treatment: the function name is added
+                # to free variables (and the closure of the resulting function)
+                names: list[str] = [node.name]
+                memo = self._memo.parent
+                while memo:
+                    if isinstance(memo.node, (FunctionDef, AsyncFunctionDef)):
+                        # This is a nested function. Use the function name as-is.
+                        del names[:-1]
+                        break
+                    elif not isinstance(memo.node, ClassDef):
+                        break
+
+                    names.insert(0, memo.node.name)
+                    memo = memo.parent
+
+                config_keywords = self._memo.get_config_keywords()
+                if config_keywords:
+                    memo_kwargs["config"] = Call(
+                        self._get_import("dataclasses", "replace"),
+                        [self._get_import("typeguard._config", "global_config")],
+                        config_keywords,
+                    )
+
+                self._memo.memo_var_name.id = self._memo.get_unused_name("memo")
+                memo_store_name = Name(id=self._memo.memo_var_name.id, ctx=Store())
+                globals_call = Call(Name(id="globals", ctx=Load()), [], [])
+                locals_call = Call(Name(id="locals", ctx=Load()), [], [])
+                memo_expr = Call(
+                    self._get_import("typeguard", "TypeCheckMemo"),
+                    [globals_call, locals_call],
+                    [keyword(key, value) for key, value in memo_kwargs.items()],
+                )
+                node.body.insert(
+                    self._memo.code_inject_index,
+                    Assign([memo_store_name], memo_expr),
+                )
+
+                self._memo.insert_imports(node)
+
+                # Special case the __new__() method to create a local alias from the
+                # class name to the first argument (usually "cls")
+                if (
+                    isinstance(node, FunctionDef)
+                    and node.args
+                    and self._memo.parent is not None
+                    and isinstance(self._memo.parent.node, ClassDef)
+                    and node.name == "__new__"
+                ):
+                    first_args_expr = Name(node.args.args[0].arg, ctx=Load())
+                    cls_name = Name(self._memo.parent.node.name, ctx=Store())
+                    node.body.insert(
+                        self._memo.code_inject_index,
+                        Assign([cls_name], first_args_expr),
+                    )
+
+                # Rmove any placeholder "pass" at the end
+                if isinstance(node.body[-1], Pass):
+                    del node.body[-1]
+
+        return node
+
+    def visit_AsyncFunctionDef(
+        self, node: AsyncFunctionDef
+    ) -> FunctionDef | AsyncFunctionDef | None:
+        return self.visit_FunctionDef(node)
+
+    def visit_Return(self, node: Return) -> Return:
+        """This injects type checks into "return" statements."""
+        self.generic_visit(node)
+        if (
+            self._memo.return_annotation
+            and self._memo.should_instrument
+            and not self._memo.is_ignored_name(self._memo.return_annotation)
+        ):
+            func_name = self._get_import("typeguard._functions", "check_return_type")
+            old_node = node
+            retval = old_node.value or Constant(None)
+            node = Return(
+                Call(
+                    func_name,
+                    [
+                        self._memo.joined_path,
+                        retval,
+                        self._memo.return_annotation,
+                        self._memo.get_memo_name(),
+                    ],
+                    [],
+                )
+            )
+            copy_location(node, old_node)
+
+        return node
+
+    def visit_Yield(self, node: Yield) -> Yield | Call:
+        """
+        This injects type checks into "yield" expressions, checking both the yielded
+        value and the value sent back to the generator, when appropriate.
+
+        """
+        self._memo.has_yield_expressions = True
+        self.generic_visit(node)
+
+        if (
+            self._memo.yield_annotation
+            and self._memo.should_instrument
+            and not self._memo.is_ignored_name(self._memo.yield_annotation)
+        ):
+            func_name = self._get_import("typeguard._functions", "check_yield_type")
+            yieldval = node.value or Constant(None)
+            node.value = Call(
+                func_name,
+                [
+                    self._memo.joined_path,
+                    yieldval,
+                    self._memo.yield_annotation,
+                    self._memo.get_memo_name(),
+                ],
+                [],
+            )
+
+        if (
+            self._memo.send_annotation
+            and self._memo.should_instrument
+            and not self._memo.is_ignored_name(self._memo.send_annotation)
+        ):
+            func_name = self._get_import("typeguard._functions", "check_send_type")
+            old_node = node
+            call_node = Call(
+                func_name,
+                [
+                    self._memo.joined_path,
+                    old_node,
+                    self._memo.send_annotation,
+                    self._memo.get_memo_name(),
+                ],
+                [],
+            )
+            copy_location(call_node, old_node)
+            return call_node
+
+        return node
+
+    def visit_AnnAssign(self, node: AnnAssign) -> Any:
+        """
+        This injects a type check into a local variable annotation-assignment within a
+        function body.
+
+        """
+        self.generic_visit(node)
+
+        if (
+            isinstance(self._memo.node, (FunctionDef, AsyncFunctionDef))
+            and node.annotation
+            and isinstance(node.target, Name)
+        ):
+            self._memo.ignored_names.add(node.target.id)
+            annotation = self._convert_annotation(deepcopy(node.annotation))
+            if annotation:
+                self._memo.variable_annotations[node.target.id] = annotation
+                if node.value:
+                    func_name = self._get_import(
+                        "typeguard._functions", "check_variable_assignment"
+                    )
+                    node.value = Call(
+                        func_name,
+                        [
+                            node.value,
+                            Constant(node.target.id),
+                            annotation,
+                            self._memo.get_memo_name(),
+                        ],
+                        [],
+                    )
+
+        return node
+
+    def visit_Assign(self, node: Assign) -> Any:
+        """
+        This injects a type check into a local variable assignment within a function
+        body. The variable must have been annotated earlier in the function body.
+
+        """
+        self.generic_visit(node)
+
+        # Only instrument function-local assignments
+        if isinstance(self._memo.node, (FunctionDef, AsyncFunctionDef)):
+            targets: list[dict[Constant, expr | None]] = []
+            check_required = False
+            for target in node.targets:
+                elts: Sequence[expr]
+                if isinstance(target, Name):
+                    elts = [target]
+                elif isinstance(target, Tuple):
+                    elts = target.elts
+                else:
+                    continue
+
+                annotations_: dict[Constant, expr | None] = {}
+                for exp in elts:
+                    prefix = ""
+                    if isinstance(exp, Starred):
+                        exp = exp.value
+                        prefix = "*"
+
+                    if isinstance(exp, Name):
+                        self._memo.ignored_names.add(exp.id)
+                        name = prefix + exp.id
+                        annotation = self._memo.variable_annotations.get(exp.id)
+                        if annotation:
+                            annotations_[Constant(name)] = annotation
+                            check_required = True
+                        else:
+                            annotations_[Constant(name)] = None
+
+                targets.append(annotations_)
+
+            if check_required:
+                # Replace missing annotations with typing.Any
+                for item in targets:
+                    for key, expression in item.items():
+                        if expression is None:
+                            item[key] = self._get_import("typing", "Any")
+
+                if len(targets) == 1 and len(targets[0]) == 1:
+                    func_name = self._get_import(
+                        "typeguard._functions", "check_variable_assignment"
+                    )
+                    target_varname = next(iter(targets[0]))
+                    node.value = Call(
+                        func_name,
+                        [
+                            node.value,
+                            target_varname,
+                            targets[0][target_varname],
+                            self._memo.get_memo_name(),
+                        ],
+                        [],
+                    )
+                elif targets:
+                    func_name = self._get_import(
+                        "typeguard._functions", "check_multi_variable_assignment"
+                    )
+                    targets_arg = List(
+                        [
+                            Dict(keys=list(target), values=list(target.values()))
+                            for target in targets
+                        ],
+                        ctx=Load(),
+                    )
+                    node.value = Call(
+                        func_name,
+                        [node.value, targets_arg, self._memo.get_memo_name()],
+                        [],
+                    )
+
+        return node
+
+    def visit_NamedExpr(self, node: NamedExpr) -> Any:
+        """This injects a type check into an assignment expression (a := foo())."""
+        self.generic_visit(node)
+
+        # Only instrument function-local assignments
+        if isinstance(self._memo.node, (FunctionDef, AsyncFunctionDef)) and isinstance(
+            node.target, Name
+        ):
+            self._memo.ignored_names.add(node.target.id)
+
+            # Bail out if no matching annotation is found
+            annotation = self._memo.variable_annotations.get(node.target.id)
+            if annotation is None:
+                return node
+
+            func_name = self._get_import(
+                "typeguard._functions", "check_variable_assignment"
+            )
+            node.value = Call(
+                func_name,
+                [
+                    node.value,
+                    Constant(node.target.id),
+                    annotation,
+                    self._memo.get_memo_name(),
+                ],
+                [],
+            )
+
+        return node
+
+    def visit_AugAssign(self, node: AugAssign) -> Any:
+        """
+        This injects a type check into an augmented assignment expression (a += 1).
+
+        """
+        self.generic_visit(node)
+
+        # Only instrument function-local assignments
+        if isinstance(self._memo.node, (FunctionDef, AsyncFunctionDef)) and isinstance(
+            node.target, Name
+        ):
+            # Bail out if no matching annotation is found
+            annotation = self._memo.variable_annotations.get(node.target.id)
+            if annotation is None:
+                return node
+
+            # Bail out if the operator is not found (newer Python version?)
+            try:
+                operator_func_name = aug_assign_functions[node.op.__class__]
+            except KeyError:
+                return node
+
+            operator_func = self._get_import("operator", operator_func_name)
+            operator_call = Call(
+                operator_func, [Name(node.target.id, ctx=Load()), node.value], []
+            )
+            check_call = Call(
+                self._get_import("typeguard._functions", "check_variable_assignment"),
+                [
+                    operator_call,
+                    Constant(node.target.id),
+                    annotation,
+                    self._memo.get_memo_name(),
+                ],
+                [],
+            )
+            return Assign(targets=[node.target], value=check_call)
+
+        return node
+
+    def visit_If(self, node: If) -> Any:
+        """
+        This blocks names from being collected from a module-level
+        "if typing.TYPE_CHECKING:" block, so that they won't be type checked.
+
+        """
+        self.generic_visit(node)
+
+        if (
+            self._memo is self._module_memo
+            and isinstance(node.test, Name)
+            and self._memo.name_matches(node.test, "typing.TYPE_CHECKING")
+        ):
+            collector = NameCollector()
+            collector.visit(node)
+            self._memo.ignored_names.update(collector.names)
+
+        return node
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_union_transformer.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_union_transformer.py
new file mode 100644
index 00000000..19617e6a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_union_transformer.py
@@ -0,0 +1,55 @@
+"""
+Transforms lazily evaluated PEP 604 unions into typing.Unions, for compatibility with
+Python versions older than 3.10.
+"""
+
+from __future__ import annotations
+
+from ast import (
+    BinOp,
+    BitOr,
+    Index,
+    Load,
+    Name,
+    NodeTransformer,
+    Subscript,
+    fix_missing_locations,
+    parse,
+)
+from ast import Tuple as ASTTuple
+from types import CodeType
+from typing import Any, Dict, FrozenSet, List, Set, Tuple, Union
+
+type_substitutions = {
+    "dict": Dict,
+    "list": List,
+    "tuple": Tuple,
+    "set": Set,
+    "frozenset": FrozenSet,
+    "Union": Union,
+}
+
+
+class UnionTransformer(NodeTransformer):
+    def __init__(self, union_name: Name | None = None):
+        self.union_name = union_name or Name(id="Union", ctx=Load())
+
+    def visit_BinOp(self, node: BinOp) -> Any:
+        self.generic_visit(node)
+        if isinstance(node.op, BitOr):
+            return Subscript(
+                value=self.union_name,
+                slice=Index(
+                    ASTTuple(elts=[node.left, node.right], ctx=Load()), ctx=Load()
+                ),
+                ctx=Load(),
+            )
+
+        return node
+
+
+def compile_type_hint(hint: str) -> CodeType:
+    parsed = parse(hint, "<string>", "eval")
+    UnionTransformer().visit(parsed)
+    fix_missing_locations(parsed)
+    return compile(parsed, "<string>", "eval", flags=0)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_utils.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_utils.py
new file mode 100644
index 00000000..9bcc8417
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/_utils.py
@@ -0,0 +1,173 @@
+from __future__ import annotations
+
+import inspect
+import sys
+from importlib import import_module
+from inspect import currentframe
+from types import CodeType, FrameType, FunctionType
+from typing import TYPE_CHECKING, Any, Callable, ForwardRef, Union, cast, final
+from weakref import WeakValueDictionary
+
+if TYPE_CHECKING:
+    from ._memo import TypeCheckMemo
+
+if sys.version_info >= (3, 13):
+    from typing import get_args, get_origin
+
+    def evaluate_forwardref(forwardref: ForwardRef, memo: TypeCheckMemo) -> Any:
+        return forwardref._evaluate(
+            memo.globals, memo.locals, type_params=(), recursive_guard=frozenset()
+        )
+
+elif sys.version_info >= (3, 10):
+    from typing import get_args, get_origin
+
+    def evaluate_forwardref(forwardref: ForwardRef, memo: TypeCheckMemo) -> Any:
+        return forwardref._evaluate(
+            memo.globals, memo.locals, recursive_guard=frozenset()
+        )
+
+else:
+    from typing_extensions import get_args, get_origin
+
+    evaluate_extra_args: tuple[frozenset[Any], ...] = (
+        (frozenset(),) if sys.version_info >= (3, 9) else ()
+    )
+
+    def evaluate_forwardref(forwardref: ForwardRef, memo: TypeCheckMemo) -> Any:
+        from ._union_transformer import compile_type_hint, type_substitutions
+
+        if not forwardref.__forward_evaluated__:
+            forwardref.__forward_code__ = compile_type_hint(forwardref.__forward_arg__)
+
+        try:
+            return forwardref._evaluate(memo.globals, memo.locals, *evaluate_extra_args)
+        except NameError:
+            if sys.version_info < (3, 10):
+                # Try again, with the type substitutions (list -> List etc.) in place
+                new_globals = memo.globals.copy()
+                new_globals.setdefault("Union", Union)
+                if sys.version_info < (3, 9):
+                    new_globals.update(type_substitutions)
+
+                return forwardref._evaluate(
+                    new_globals, memo.locals or new_globals, *evaluate_extra_args
+                )
+
+            raise
+
+
+_functions_map: WeakValueDictionary[CodeType, FunctionType] = WeakValueDictionary()
+
+
+def get_type_name(type_: Any) -> str:
+    name: str
+    for attrname in "__name__", "_name", "__forward_arg__":
+        candidate = getattr(type_, attrname, None)
+        if isinstance(candidate, str):
+            name = candidate
+            break
+    else:
+        origin = get_origin(type_)
+        candidate = getattr(origin, "_name", None)
+        if candidate is None:
+            candidate = type_.__class__.__name__.strip("_")
+
+        if isinstance(candidate, str):
+            name = candidate
+        else:
+            return "(unknown)"
+
+    args = get_args(type_)
+    if args:
+        if name == "Literal":
+            formatted_args = ", ".join(repr(arg) for arg in args)
+        else:
+            formatted_args = ", ".join(get_type_name(arg) for arg in args)
+
+        name += f"[{formatted_args}]"
+
+    module = getattr(type_, "__module__", None)
+    if module and module not in (None, "typing", "typing_extensions", "builtins"):
+        name = module + "." + name
+
+    return name
+
+
+def qualified_name(obj: Any, *, add_class_prefix: bool = False) -> str:
+    """
+    Return the qualified name (e.g. package.module.Type) for the given object.
+
+    Builtins and types from the :mod:`typing` package get special treatment by having
+    the module name stripped from the generated name.
+
+    """
+    if obj is None:
+        return "None"
+    elif inspect.isclass(obj):
+        prefix = "class " if add_class_prefix else ""
+        type_ = obj
+    else:
+        prefix = ""
+        type_ = type(obj)
+
+    module = type_.__module__
+    qualname = type_.__qualname__
+    name = qualname if module in ("typing", "builtins") else f"{module}.{qualname}"
+    return prefix + name
+
+
+def function_name(func: Callable[..., Any]) -> str:
+    """
+    Return the qualified name of the given function.
+
+    Builtins and types from the :mod:`typing` package get special treatment by having
+    the module name stripped from the generated name.
+
+    """
+    # For partial functions and objects with __call__ defined, __qualname__ does not
+    # exist
+    module = getattr(func, "__module__", "")
+    qualname = (module + ".") if module not in ("builtins", "") else ""
+    return qualname + getattr(func, "__qualname__", repr(func))
+
+
+def resolve_reference(reference: str) -> Any:
+    modulename, varname = reference.partition(":")[::2]
+    if not modulename or not varname:
+        raise ValueError(f"{reference!r} is not a module:varname reference")
+
+    obj = import_module(modulename)
+    for attr in varname.split("."):
+        obj = getattr(obj, attr)
+
+    return obj
+
+
+def is_method_of(obj: object, cls: type) -> bool:
+    return (
+        inspect.isfunction(obj)
+        and obj.__module__ == cls.__module__
+        and obj.__qualname__.startswith(cls.__qualname__ + ".")
+    )
+
+
+def get_stacklevel() -> int:
+    level = 1
+    frame = cast(FrameType, currentframe()).f_back
+    while frame and frame.f_globals.get("__name__", "").startswith("typeguard."):
+        level += 1
+        frame = frame.f_back
+
+    return level
+
+
+@final
+class Unset:
+    __slots__ = ()
+
+    def __repr__(self) -> str:
+        return "<unset>"
+
+
+unset = Unset()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/py.typed b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typeguard/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/LICENSE
new file mode 100644
index 00000000..f26bcf4d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/LICENSE
@@ -0,0 +1,279 @@
+A. HISTORY OF THE SOFTWARE
+==========================
+
+Python was created in the early 1990s by Guido van Rossum at Stichting
+Mathematisch Centrum (CWI, see https://www.cwi.nl) in the Netherlands
+as a successor of a language called ABC.  Guido remains Python's
+principal author, although it includes many contributions from others.
+
+In 1995, Guido continued his work on Python at the Corporation for
+National Research Initiatives (CNRI, see https://www.cnri.reston.va.us)
+in Reston, Virginia where he released several versions of the
+software.
+
+In May 2000, Guido and the Python core development team moved to
+BeOpen.com to form the BeOpen PythonLabs team.  In October of the same
+year, the PythonLabs team moved to Digital Creations, which became
+Zope Corporation.  In 2001, the Python Software Foundation (PSF, see
+https://www.python.org/psf/) was formed, a non-profit organization
+created specifically to own Python-related Intellectual Property.
+Zope Corporation was a sponsoring member of the PSF.
+
+All Python releases are Open Source (see https://opensource.org for
+the Open Source Definition).  Historically, most, but not all, Python
+releases have also been GPL-compatible; the table below summarizes
+the various releases.
+
+    Release         Derived     Year        Owner       GPL-
+                    from                                compatible? (1)
+
+    0.9.0 thru 1.2              1991-1995   CWI         yes
+    1.3 thru 1.5.2  1.2         1995-1999   CNRI        yes
+    1.6             1.5.2       2000        CNRI        no
+    2.0             1.6         2000        BeOpen.com  no
+    1.6.1           1.6         2001        CNRI        yes (2)
+    2.1             2.0+1.6.1   2001        PSF         no
+    2.0.1           2.0+1.6.1   2001        PSF         yes
+    2.1.1           2.1+2.0.1   2001        PSF         yes
+    2.1.2           2.1.1       2002        PSF         yes
+    2.1.3           2.1.2       2002        PSF         yes
+    2.2 and above   2.1.1       2001-now    PSF         yes
+
+Footnotes:
+
+(1) GPL-compatible doesn't mean that we're distributing Python under
+    the GPL.  All Python licenses, unlike the GPL, let you distribute
+    a modified version without making your changes open source.  The
+    GPL-compatible licenses make it possible to combine Python with
+    other software that is released under the GPL; the others don't.
+
+(2) According to Richard Stallman, 1.6.1 is not GPL-compatible,
+    because its license has a choice of law clause.  According to
+    CNRI, however, Stallman's lawyer has told CNRI's lawyer that 1.6.1
+    is "not incompatible" with the GPL.
+
+Thanks to the many outside volunteers who have worked under Guido's
+direction to make these releases possible.
+
+
+B. TERMS AND CONDITIONS FOR ACCESSING OR OTHERWISE USING PYTHON
+===============================================================
+
+Python software and documentation are licensed under the
+Python Software Foundation License Version 2.
+
+Starting with Python 3.8.6, examples, recipes, and other code in
+the documentation are dual licensed under the PSF License Version 2
+and the Zero-Clause BSD license.
+
+Some software incorporated into Python is under different licenses.
+The licenses are listed with code falling under that license.
+
+
+PYTHON SOFTWARE FOUNDATION LICENSE VERSION 2
+--------------------------------------------
+
+1. This LICENSE AGREEMENT is between the Python Software Foundation
+("PSF"), and the Individual or Organization ("Licensee") accessing and
+otherwise using this software ("Python") in source or binary form and
+its associated documentation.
+
+2. Subject to the terms and conditions of this License Agreement, PSF hereby
+grants Licensee a nonexclusive, royalty-free, world-wide license to reproduce,
+analyze, test, perform and/or display publicly, prepare derivative works,
+distribute, and otherwise use Python alone or in any derivative version,
+provided, however, that PSF's License Agreement and PSF's notice of copyright,
+i.e., "Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
+2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023 Python Software Foundation;
+All Rights Reserved" are retained in Python alone or in any derivative version
+prepared by Licensee.
+
+3. In the event Licensee prepares a derivative work that is based on
+or incorporates Python or any part thereof, and wants to make
+the derivative work available to others as provided herein, then
+Licensee hereby agrees to include in any such work a brief summary of
+the changes made to Python.
+
+4. PSF is making Python available to Licensee on an "AS IS"
+basis.  PSF MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, PSF MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF PYTHON WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+5. PSF SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF PYTHON
+FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS AS
+A RESULT OF MODIFYING, DISTRIBUTING, OR OTHERWISE USING PYTHON,
+OR ANY DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+6. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+7. Nothing in this License Agreement shall be deemed to create any
+relationship of agency, partnership, or joint venture between PSF and
+Licensee.  This License Agreement does not grant permission to use PSF
+trademarks or trade name in a trademark sense to endorse or promote
+products or services of Licensee, or any third party.
+
+8. By copying, installing or otherwise using Python, Licensee
+agrees to be bound by the terms and conditions of this License
+Agreement.
+
+
+BEOPEN.COM LICENSE AGREEMENT FOR PYTHON 2.0
+-------------------------------------------
+
+BEOPEN PYTHON OPEN SOURCE LICENSE AGREEMENT VERSION 1
+
+1. This LICENSE AGREEMENT is between BeOpen.com ("BeOpen"), having an
+office at 160 Saratoga Avenue, Santa Clara, CA 95051, and the
+Individual or Organization ("Licensee") accessing and otherwise using
+this software in source or binary form and its associated
+documentation ("the Software").
+
+2. Subject to the terms and conditions of this BeOpen Python License
+Agreement, BeOpen hereby grants Licensee a non-exclusive,
+royalty-free, world-wide license to reproduce, analyze, test, perform
+and/or display publicly, prepare derivative works, distribute, and
+otherwise use the Software alone or in any derivative version,
+provided, however, that the BeOpen Python License is retained in the
+Software, alone or in any derivative version prepared by Licensee.
+
+3. BeOpen is making the Software available to Licensee on an "AS IS"
+basis.  BEOPEN MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, BEOPEN MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF THE SOFTWARE WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+4. BEOPEN SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF THE
+SOFTWARE FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS
+AS A RESULT OF USING, MODIFYING OR DISTRIBUTING THE SOFTWARE, OR ANY
+DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+5. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+6. This License Agreement shall be governed by and interpreted in all
+respects by the law of the State of California, excluding conflict of
+law provisions.  Nothing in this License Agreement shall be deemed to
+create any relationship of agency, partnership, or joint venture
+between BeOpen and Licensee.  This License Agreement does not grant
+permission to use BeOpen trademarks or trade names in a trademark
+sense to endorse or promote products or services of Licensee, or any
+third party.  As an exception, the "BeOpen Python" logos available at
+http://www.pythonlabs.com/logos.html may be used according to the
+permissions granted on that web page.
+
+7. By copying, installing or otherwise using the software, Licensee
+agrees to be bound by the terms and conditions of this License
+Agreement.
+
+
+CNRI LICENSE AGREEMENT FOR PYTHON 1.6.1
+---------------------------------------
+
+1. This LICENSE AGREEMENT is between the Corporation for National
+Research Initiatives, having an office at 1895 Preston White Drive,
+Reston, VA 20191 ("CNRI"), and the Individual or Organization
+("Licensee") accessing and otherwise using Python 1.6.1 software in
+source or binary form and its associated documentation.
+
+2. Subject to the terms and conditions of this License Agreement, CNRI
+hereby grants Licensee a nonexclusive, royalty-free, world-wide
+license to reproduce, analyze, test, perform and/or display publicly,
+prepare derivative works, distribute, and otherwise use Python 1.6.1
+alone or in any derivative version, provided, however, that CNRI's
+License Agreement and CNRI's notice of copyright, i.e., "Copyright (c)
+1995-2001 Corporation for National Research Initiatives; All Rights
+Reserved" are retained in Python 1.6.1 alone or in any derivative
+version prepared by Licensee.  Alternately, in lieu of CNRI's License
+Agreement, Licensee may substitute the following text (omitting the
+quotes): "Python 1.6.1 is made available subject to the terms and
+conditions in CNRI's License Agreement.  This Agreement together with
+Python 1.6.1 may be located on the internet using the following
+unique, persistent identifier (known as a handle): 1895.22/1013.  This
+Agreement may also be obtained from a proxy server on the internet
+using the following URL: http://hdl.handle.net/1895.22/1013".
+
+3. In the event Licensee prepares a derivative work that is based on
+or incorporates Python 1.6.1 or any part thereof, and wants to make
+the derivative work available to others as provided herein, then
+Licensee hereby agrees to include in any such work a brief summary of
+the changes made to Python 1.6.1.
+
+4. CNRI is making Python 1.6.1 available to Licensee on an "AS IS"
+basis.  CNRI MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, CNRI MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF PYTHON 1.6.1 WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+5. CNRI SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF PYTHON
+1.6.1 FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS AS
+A RESULT OF MODIFYING, DISTRIBUTING, OR OTHERWISE USING PYTHON 1.6.1,
+OR ANY DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+6. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+7. This License Agreement shall be governed by the federal
+intellectual property law of the United States, including without
+limitation the federal copyright law, and, to the extent such
+U.S. federal law does not apply, by the law of the Commonwealth of
+Virginia, excluding Virginia's conflict of law provisions.
+Notwithstanding the foregoing, with regard to derivative works based
+on Python 1.6.1 that incorporate non-separable material that was
+previously distributed under the GNU General Public License (GPL), the
+law of the Commonwealth of Virginia shall govern this License
+Agreement only as to issues arising under or with respect to
+Paragraphs 4, 5, and 7 of this License Agreement.  Nothing in this
+License Agreement shall be deemed to create any relationship of
+agency, partnership, or joint venture between CNRI and Licensee.  This
+License Agreement does not grant permission to use CNRI trademarks or
+trade name in a trademark sense to endorse or promote products or
+services of Licensee, or any third party.
+
+8. By clicking on the "ACCEPT" button where indicated, or by copying,
+installing or otherwise using Python 1.6.1, Licensee agrees to be
+bound by the terms and conditions of this License Agreement.
+
+        ACCEPT
+
+
+CWI LICENSE AGREEMENT FOR PYTHON 0.9.0 THROUGH 1.2
+--------------------------------------------------
+
+Copyright (c) 1991 - 1995, Stichting Mathematisch Centrum Amsterdam,
+The Netherlands.  All rights reserved.
+
+Permission to use, copy, modify, and distribute this software and its
+documentation for any purpose and without fee is hereby granted,
+provided that the above copyright notice appear in all copies and that
+both that copyright notice and this permission notice appear in
+supporting documentation, and that the name of Stichting Mathematisch
+Centrum or CWI not be used in advertising or publicity pertaining to
+distribution of the software without specific, written prior
+permission.
+
+STICHTING MATHEMATISCH CENTRUM DISCLAIMS ALL WARRANTIES WITH REGARD TO
+THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
+FITNESS, IN NO EVENT SHALL STICHTING MATHEMATISCH CENTRUM BE LIABLE
+FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
+OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+
+ZERO-CLAUSE BSD LICENSE FOR CODE IN THE PYTHON DOCUMENTATION
+----------------------------------------------------------------------
+
+Permission to use, copy, modify, and/or distribute this software for any
+purpose with or without fee is hereby granted.
+
+THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH
+REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
+AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
+INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+PERFORMANCE OF THIS SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/METADATA
new file mode 100644
index 00000000..f15e2b38
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/METADATA
@@ -0,0 +1,67 @@
+Metadata-Version: 2.1
+Name: typing_extensions
+Version: 4.12.2
+Summary: Backported and Experimental Type Hints for Python 3.8+
+Keywords: annotations,backport,checker,checking,function,hinting,hints,type,typechecking,typehinting,typehints,typing
+Author-email: "Guido van Rossum, Jukka Lehtosalo, Łukasz Langa, Michael Lee" <levkivskyi@gmail.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Python Software Foundation License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development
+Project-URL: Bug Tracker, https://github.com/python/typing_extensions/issues
+Project-URL: Changes, https://github.com/python/typing_extensions/blob/main/CHANGELOG.md
+Project-URL: Documentation, https://typing-extensions.readthedocs.io/
+Project-URL: Home, https://github.com/python/typing_extensions
+Project-URL: Q & A, https://github.com/python/typing/discussions
+Project-URL: Repository, https://github.com/python/typing_extensions
+
+# Typing Extensions
+
+[![Chat at https://gitter.im/python/typing](https://badges.gitter.im/python/typing.svg)](https://gitter.im/python/typing)
+
+[Documentation](https://typing-extensions.readthedocs.io/en/latest/#) –
+[PyPI](https://pypi.org/project/typing-extensions/)
+
+## Overview
+
+The `typing_extensions` module serves two related purposes:
+
+- Enable use of new type system features on older Python versions. For example,
+  `typing.TypeGuard` is new in Python 3.10, but `typing_extensions` allows
+  users on previous Python versions to use it too.
+- Enable experimentation with new type system PEPs before they are accepted and
+  added to the `typing` module.
+
+`typing_extensions` is treated specially by static type checkers such as
+mypy and pyright. Objects defined in `typing_extensions` are treated the same
+way as equivalent forms in `typing`.
+
+`typing_extensions` uses
+[Semantic Versioning](https://semver.org/). The
+major version will be incremented only for backwards-incompatible changes.
+Therefore, it's safe to depend
+on `typing_extensions` like this: `typing_extensions >=x.y, <(x+1)`,
+where `x.y` is the first version that includes all features you need.
+
+## Included items
+
+See [the documentation](https://typing-extensions.readthedocs.io/en/latest/#) for a
+complete listing of module contents.
+
+## Contributing
+
+See [CONTRIBUTING.md](https://github.com/python/typing_extensions/blob/main/CONTRIBUTING.md)
+for how to contribute to `typing_extensions`.
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/RECORD
new file mode 100644
index 00000000..bc7b4533
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/RECORD
@@ -0,0 +1,7 @@
+__pycache__/typing_extensions.cpython-312.pyc,,
+typing_extensions-4.12.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+typing_extensions-4.12.2.dist-info/LICENSE,sha256=Oy-B_iHRgcSZxZolbI4ZaEVdZonSaaqFNzv7avQdo78,13936
+typing_extensions-4.12.2.dist-info/METADATA,sha256=BeUQIa8cnYbrjWx-N8TOznM9UGW5Gm2DicVpDtRA8W0,3018
+typing_extensions-4.12.2.dist-info/RECORD,,
+typing_extensions-4.12.2.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+typing_extensions.py,sha256=gwekpyG9DVG3lxWKX4ni8u7nk3We5slG98mA9F3DJQw,134451
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions-4.12.2.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions.py
new file mode 100644
index 00000000..dec429ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/typing_extensions.py
@@ -0,0 +1,3641 @@
+import abc
+import collections
+import collections.abc
+import contextlib
+import functools
+import inspect
+import operator
+import sys
+import types as _types
+import typing
+import warnings
+
+__all__ = [
+    # Super-special typing primitives.
+    'Any',
+    'ClassVar',
+    'Concatenate',
+    'Final',
+    'LiteralString',
+    'ParamSpec',
+    'ParamSpecArgs',
+    'ParamSpecKwargs',
+    'Self',
+    'Type',
+    'TypeVar',
+    'TypeVarTuple',
+    'Unpack',
+
+    # ABCs (from collections.abc).
+    'Awaitable',
+    'AsyncIterator',
+    'AsyncIterable',
+    'Coroutine',
+    'AsyncGenerator',
+    'AsyncContextManager',
+    'Buffer',
+    'ChainMap',
+
+    # Concrete collection types.
+    'ContextManager',
+    'Counter',
+    'Deque',
+    'DefaultDict',
+    'NamedTuple',
+    'OrderedDict',
+    'TypedDict',
+
+    # Structural checks, a.k.a. protocols.
+    'SupportsAbs',
+    'SupportsBytes',
+    'SupportsComplex',
+    'SupportsFloat',
+    'SupportsIndex',
+    'SupportsInt',
+    'SupportsRound',
+
+    # One-off things.
+    'Annotated',
+    'assert_never',
+    'assert_type',
+    'clear_overloads',
+    'dataclass_transform',
+    'deprecated',
+    'Doc',
+    'get_overloads',
+    'final',
+    'get_args',
+    'get_origin',
+    'get_original_bases',
+    'get_protocol_members',
+    'get_type_hints',
+    'IntVar',
+    'is_protocol',
+    'is_typeddict',
+    'Literal',
+    'NewType',
+    'overload',
+    'override',
+    'Protocol',
+    'reveal_type',
+    'runtime',
+    'runtime_checkable',
+    'Text',
+    'TypeAlias',
+    'TypeAliasType',
+    'TypeGuard',
+    'TypeIs',
+    'TYPE_CHECKING',
+    'Never',
+    'NoReturn',
+    'ReadOnly',
+    'Required',
+    'NotRequired',
+
+    # Pure aliases, have always been in typing
+    'AbstractSet',
+    'AnyStr',
+    'BinaryIO',
+    'Callable',
+    'Collection',
+    'Container',
+    'Dict',
+    'ForwardRef',
+    'FrozenSet',
+    'Generator',
+    'Generic',
+    'Hashable',
+    'IO',
+    'ItemsView',
+    'Iterable',
+    'Iterator',
+    'KeysView',
+    'List',
+    'Mapping',
+    'MappingView',
+    'Match',
+    'MutableMapping',
+    'MutableSequence',
+    'MutableSet',
+    'NoDefault',
+    'Optional',
+    'Pattern',
+    'Reversible',
+    'Sequence',
+    'Set',
+    'Sized',
+    'TextIO',
+    'Tuple',
+    'Union',
+    'ValuesView',
+    'cast',
+    'no_type_check',
+    'no_type_check_decorator',
+]
+
+# for backward compatibility
+PEP_560 = True
+GenericMeta = type
+_PEP_696_IMPLEMENTED = sys.version_info >= (3, 13, 0, "beta")
+
+# The functions below are modified copies of typing internal helpers.
+# They are needed by _ProtocolMeta and they provide support for PEP 646.
+
+
+class _Sentinel:
+    def __repr__(self):
+        return "<sentinel>"
+
+
+_marker = _Sentinel()
+
+
+if sys.version_info >= (3, 10):
+    def _should_collect_from_parameters(t):
+        return isinstance(
+            t, (typing._GenericAlias, _types.GenericAlias, _types.UnionType)
+        )
+elif sys.version_info >= (3, 9):
+    def _should_collect_from_parameters(t):
+        return isinstance(t, (typing._GenericAlias, _types.GenericAlias))
+else:
+    def _should_collect_from_parameters(t):
+        return isinstance(t, typing._GenericAlias) and not t._special
+
+
+NoReturn = typing.NoReturn
+
+# Some unconstrained type variables.  These are used by the container types.
+# (These are not for export.)
+T = typing.TypeVar('T')  # Any type.
+KT = typing.TypeVar('KT')  # Key type.
+VT = typing.TypeVar('VT')  # Value type.
+T_co = typing.TypeVar('T_co', covariant=True)  # Any type covariant containers.
+T_contra = typing.TypeVar('T_contra', contravariant=True)  # Ditto contravariant.
+
+
+if sys.version_info >= (3, 11):
+    from typing import Any
+else:
+
+    class _AnyMeta(type):
+        def __instancecheck__(self, obj):
+            if self is Any:
+                raise TypeError("typing_extensions.Any cannot be used with isinstance()")
+            return super().__instancecheck__(obj)
+
+        def __repr__(self):
+            if self is Any:
+                return "typing_extensions.Any"
+            return super().__repr__()
+
+    class Any(metaclass=_AnyMeta):
+        """Special type indicating an unconstrained type.
+        - Any is compatible with every type.
+        - Any assumed to have all methods.
+        - All values assumed to be instances of Any.
+        Note that all the above statements are true from the point of view of
+        static type checkers. At runtime, Any should not be used with instance
+        checks.
+        """
+        def __new__(cls, *args, **kwargs):
+            if cls is Any:
+                raise TypeError("Any cannot be instantiated")
+            return super().__new__(cls, *args, **kwargs)
+
+
+ClassVar = typing.ClassVar
+
+
+class _ExtensionsSpecialForm(typing._SpecialForm, _root=True):
+    def __repr__(self):
+        return 'typing_extensions.' + self._name
+
+
+Final = typing.Final
+
+if sys.version_info >= (3, 11):
+    final = typing.final
+else:
+    # @final exists in 3.8+, but we backport it for all versions
+    # before 3.11 to keep support for the __final__ attribute.
+    # See https://bugs.python.org/issue46342
+    def final(f):
+        """This decorator can be used to indicate to type checkers that
+        the decorated method cannot be overridden, and decorated class
+        cannot be subclassed. For example:
+
+            class Base:
+                @final
+                def done(self) -> None:
+                    ...
+            class Sub(Base):
+                def done(self) -> None:  # Error reported by type checker
+                    ...
+            @final
+            class Leaf:
+                ...
+            class Other(Leaf):  # Error reported by type checker
+                ...
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__final__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+        """
+        try:
+            f.__final__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return f
+
+
+def IntVar(name):
+    return typing.TypeVar(name)
+
+
+# A Literal bug was fixed in 3.11.0, 3.10.1 and 3.9.8
+if sys.version_info >= (3, 10, 1):
+    Literal = typing.Literal
+else:
+    def _flatten_literal_params(parameters):
+        """An internal helper for Literal creation: flatten Literals among parameters"""
+        params = []
+        for p in parameters:
+            if isinstance(p, _LiteralGenericAlias):
+                params.extend(p.__args__)
+            else:
+                params.append(p)
+        return tuple(params)
+
+    def _value_and_type_iter(params):
+        for p in params:
+            yield p, type(p)
+
+    class _LiteralGenericAlias(typing._GenericAlias, _root=True):
+        def __eq__(self, other):
+            if not isinstance(other, _LiteralGenericAlias):
+                return NotImplemented
+            these_args_deduped = set(_value_and_type_iter(self.__args__))
+            other_args_deduped = set(_value_and_type_iter(other.__args__))
+            return these_args_deduped == other_args_deduped
+
+        def __hash__(self):
+            return hash(frozenset(_value_and_type_iter(self.__args__)))
+
+    class _LiteralForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, doc: str):
+            self._name = 'Literal'
+            self._doc = self.__doc__ = doc
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+
+            parameters = _flatten_literal_params(parameters)
+
+            val_type_pairs = list(_value_and_type_iter(parameters))
+            try:
+                deduped_pairs = set(val_type_pairs)
+            except TypeError:
+                # unhashable parameters
+                pass
+            else:
+                # similar logic to typing._deduplicate on Python 3.9+
+                if len(deduped_pairs) < len(val_type_pairs):
+                    new_parameters = []
+                    for pair in val_type_pairs:
+                        if pair in deduped_pairs:
+                            new_parameters.append(pair[0])
+                            deduped_pairs.remove(pair)
+                    assert not deduped_pairs, deduped_pairs
+                    parameters = tuple(new_parameters)
+
+            return _LiteralGenericAlias(self, parameters)
+
+    Literal = _LiteralForm(doc="""\
+                           A type that can be used to indicate to type checkers
+                           that the corresponding value has a value literally equivalent
+                           to the provided parameter. For example:
+
+                               var: Literal[4] = 4
+
+                           The type checker understands that 'var' is literally equal to
+                           the value 4 and no other value.
+
+                           Literal[...] cannot be subclassed. There is no runtime
+                           checking verifying that the parameter is actually a value
+                           instead of a type.""")
+
+
+_overload_dummy = typing._overload_dummy
+
+
+if hasattr(typing, "get_overloads"):  # 3.11+
+    overload = typing.overload
+    get_overloads = typing.get_overloads
+    clear_overloads = typing.clear_overloads
+else:
+    # {module: {qualname: {firstlineno: func}}}
+    _overload_registry = collections.defaultdict(
+        functools.partial(collections.defaultdict, dict)
+    )
+
+    def overload(func):
+        """Decorator for overloaded functions/methods.
+
+        In a stub file, place two or more stub definitions for the same
+        function in a row, each decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+
+        In a non-stub file (i.e. a regular .py file), do the same but
+        follow it with an implementation.  The implementation should *not*
+        be decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+        def utf8(value):
+            # implementation goes here
+
+        The overloads for a function can be retrieved at runtime using the
+        get_overloads() function.
+        """
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        try:
+            _overload_registry[f.__module__][f.__qualname__][
+                f.__code__.co_firstlineno
+            ] = func
+        except AttributeError:
+            # Not a normal function; ignore.
+            pass
+        return _overload_dummy
+
+    def get_overloads(func):
+        """Return all defined overloads for *func* as a sequence."""
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        if f.__module__ not in _overload_registry:
+            return []
+        mod_dict = _overload_registry[f.__module__]
+        if f.__qualname__ not in mod_dict:
+            return []
+        return list(mod_dict[f.__qualname__].values())
+
+    def clear_overloads():
+        """Clear all overloads in the registry."""
+        _overload_registry.clear()
+
+
+# This is not a real generic class.  Don't use outside annotations.
+Type = typing.Type
+
+# Various ABCs mimicking those in collections.abc.
+# A few are simply re-exported for completeness.
+Awaitable = typing.Awaitable
+Coroutine = typing.Coroutine
+AsyncIterable = typing.AsyncIterable
+AsyncIterator = typing.AsyncIterator
+Deque = typing.Deque
+DefaultDict = typing.DefaultDict
+OrderedDict = typing.OrderedDict
+Counter = typing.Counter
+ChainMap = typing.ChainMap
+Text = typing.Text
+TYPE_CHECKING = typing.TYPE_CHECKING
+
+
+if sys.version_info >= (3, 13, 0, "beta"):
+    from typing import AsyncContextManager, AsyncGenerator, ContextManager, Generator
+else:
+    def _is_dunder(attr):
+        return attr.startswith('__') and attr.endswith('__')
+
+    # Python <3.9 doesn't have typing._SpecialGenericAlias
+    _special_generic_alias_base = getattr(
+        typing, "_SpecialGenericAlias", typing._GenericAlias
+    )
+
+    class _SpecialGenericAlias(_special_generic_alias_base, _root=True):
+        def __init__(self, origin, nparams, *, inst=True, name=None, defaults=()):
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                self.__origin__ = origin
+                self._nparams = nparams
+                super().__init__(origin, nparams, special=True, inst=inst, name=name)
+            else:
+                # Python >= 3.9
+                super().__init__(origin, nparams, inst=inst, name=name)
+            self._defaults = defaults
+
+        def __setattr__(self, attr, val):
+            allowed_attrs = {'_name', '_inst', '_nparams', '_defaults'}
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                allowed_attrs.add("__origin__")
+            if _is_dunder(attr) or attr in allowed_attrs:
+                object.__setattr__(self, attr, val)
+            else:
+                setattr(self.__origin__, attr, val)
+
+        @typing._tp_cache
+        def __getitem__(self, params):
+            if not isinstance(params, tuple):
+                params = (params,)
+            msg = "Parameters to generic types must be types."
+            params = tuple(typing._type_check(p, msg) for p in params)
+            if (
+                self._defaults
+                and len(params) < self._nparams
+                and len(params) + len(self._defaults) >= self._nparams
+            ):
+                params = (*params, *self._defaults[len(params) - self._nparams:])
+            actual_len = len(params)
+
+            if actual_len != self._nparams:
+                if self._defaults:
+                    expected = f"at least {self._nparams - len(self._defaults)}"
+                else:
+                    expected = str(self._nparams)
+                if not self._nparams:
+                    raise TypeError(f"{self} is not a generic class")
+                raise TypeError(
+                    f"Too {'many' if actual_len > self._nparams else 'few'}"
+                    f" arguments for {self};"
+                    f" actual {actual_len}, expected {expected}"
+                )
+            return self.copy_with(params)
+
+    _NoneType = type(None)
+    Generator = _SpecialGenericAlias(
+        collections.abc.Generator, 3, defaults=(_NoneType, _NoneType)
+    )
+    AsyncGenerator = _SpecialGenericAlias(
+        collections.abc.AsyncGenerator, 2, defaults=(_NoneType,)
+    )
+    ContextManager = _SpecialGenericAlias(
+        contextlib.AbstractContextManager,
+        2,
+        name="ContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+    AsyncContextManager = _SpecialGenericAlias(
+        contextlib.AbstractAsyncContextManager,
+        2,
+        name="AsyncContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+
+
+_PROTO_ALLOWLIST = {
+    'collections.abc': [
+        'Callable', 'Awaitable', 'Iterable', 'Iterator', 'AsyncIterable',
+        'Hashable', 'Sized', 'Container', 'Collection', 'Reversible', 'Buffer',
+    ],
+    'contextlib': ['AbstractContextManager', 'AbstractAsyncContextManager'],
+    'typing_extensions': ['Buffer'],
+}
+
+
+_EXCLUDED_ATTRS = frozenset(typing.EXCLUDED_ATTRIBUTES) | {
+    "__match_args__", "__protocol_attrs__", "__non_callable_proto_members__",
+    "__final__",
+}
+
+
+def _get_protocol_attrs(cls):
+    attrs = set()
+    for base in cls.__mro__[:-1]:  # without object
+        if base.__name__ in {'Protocol', 'Generic'}:
+            continue
+        annotations = getattr(base, '__annotations__', {})
+        for attr in (*base.__dict__, *annotations):
+            if (not attr.startswith('_abc_') and attr not in _EXCLUDED_ATTRS):
+                attrs.add(attr)
+    return attrs
+
+
+def _caller(depth=2):
+    try:
+        return sys._getframe(depth).f_globals.get('__name__', '__main__')
+    except (AttributeError, ValueError):  # For platforms without _getframe()
+        return None
+
+
+# `__match_args__` attribute was removed from protocol members in 3.13,
+# we want to backport this change to older Python versions.
+if sys.version_info >= (3, 13):
+    Protocol = typing.Protocol
+else:
+    def _allow_reckless_class_checks(depth=3):
+        """Allow instance and class checks for special stdlib modules.
+        The abc and functools modules indiscriminately call isinstance() and
+        issubclass() on the whole MRO of a user class, which may contain protocols.
+        """
+        return _caller(depth) in {'abc', 'functools', None}
+
+    def _no_init(self, *args, **kwargs):
+        if type(self)._is_protocol:
+            raise TypeError('Protocols cannot be instantiated')
+
+    def _type_check_issubclass_arg_1(arg):
+        """Raise TypeError if `arg` is not an instance of `type`
+        in `issubclass(arg, <protocol>)`.
+
+        In most cases, this is verified by type.__subclasscheck__.
+        Checking it again unnecessarily would slow down issubclass() checks,
+        so, we don't perform this check unless we absolutely have to.
+
+        For various error paths, however,
+        we want to ensure that *this* error message is shown to the user
+        where relevant, rather than a typing.py-specific error message.
+        """
+        if not isinstance(arg, type):
+            # Same error message as for issubclass(1, int).
+            raise TypeError('issubclass() arg 1 must be a class')
+
+    # Inheriting from typing._ProtocolMeta isn't actually desirable,
+    # but is necessary to allow typing.Protocol and typing_extensions.Protocol
+    # to mix without getting TypeErrors about "metaclass conflict"
+    class _ProtocolMeta(type(typing.Protocol)):
+        # This metaclass is somewhat unfortunate,
+        # but is necessary for several reasons...
+        #
+        # NOTE: DO NOT call super() in any methods in this class
+        # That would call the methods on typing._ProtocolMeta on Python 3.8-3.11
+        # and those are slow
+        def __new__(mcls, name, bases, namespace, **kwargs):
+            if name == "Protocol" and len(bases) < 2:
+                pass
+            elif {Protocol, typing.Protocol} & set(bases):
+                for base in bases:
+                    if not (
+                        base in {object, typing.Generic, Protocol, typing.Protocol}
+                        or base.__name__ in _PROTO_ALLOWLIST.get(base.__module__, [])
+                        or is_protocol(base)
+                    ):
+                        raise TypeError(
+                            f"Protocols can only inherit from other protocols, "
+                            f"got {base!r}"
+                        )
+            return abc.ABCMeta.__new__(mcls, name, bases, namespace, **kwargs)
+
+        def __init__(cls, *args, **kwargs):
+            abc.ABCMeta.__init__(cls, *args, **kwargs)
+            if getattr(cls, "_is_protocol", False):
+                cls.__protocol_attrs__ = _get_protocol_attrs(cls)
+
+        def __subclasscheck__(cls, other):
+            if cls is Protocol:
+                return type.__subclasscheck__(cls, other)
+            if (
+                getattr(cls, '_is_protocol', False)
+                and not _allow_reckless_class_checks()
+            ):
+                if not getattr(cls, '_is_runtime_protocol', False):
+                    _type_check_issubclass_arg_1(other)
+                    raise TypeError(
+                        "Instance and class checks can only be used with "
+                        "@runtime_checkable protocols"
+                    )
+                if (
+                    # this attribute is set by @runtime_checkable:
+                    cls.__non_callable_proto_members__
+                    and cls.__dict__.get("__subclasshook__") is _proto_hook
+                ):
+                    _type_check_issubclass_arg_1(other)
+                    non_method_attrs = sorted(cls.__non_callable_proto_members__)
+                    raise TypeError(
+                        "Protocols with non-method members don't support issubclass()."
+                        f" Non-method members: {str(non_method_attrs)[1:-1]}."
+                    )
+            return abc.ABCMeta.__subclasscheck__(cls, other)
+
+        def __instancecheck__(cls, instance):
+            # We need this method for situations where attributes are
+            # assigned in __init__.
+            if cls is Protocol:
+                return type.__instancecheck__(cls, instance)
+            if not getattr(cls, "_is_protocol", False):
+                # i.e., it's a concrete subclass of a protocol
+                return abc.ABCMeta.__instancecheck__(cls, instance)
+
+            if (
+                not getattr(cls, '_is_runtime_protocol', False) and
+                not _allow_reckless_class_checks()
+            ):
+                raise TypeError("Instance and class checks can only be used with"
+                                " @runtime_checkable protocols")
+
+            if abc.ABCMeta.__instancecheck__(cls, instance):
+                return True
+
+            for attr in cls.__protocol_attrs__:
+                try:
+                    val = inspect.getattr_static(instance, attr)
+                except AttributeError:
+                    break
+                # this attribute is set by @runtime_checkable:
+                if val is None and attr not in cls.__non_callable_proto_members__:
+                    break
+            else:
+                return True
+
+            return False
+
+        def __eq__(cls, other):
+            # Hack so that typing.Generic.__class_getitem__
+            # treats typing_extensions.Protocol
+            # as equivalent to typing.Protocol
+            if abc.ABCMeta.__eq__(cls, other) is True:
+                return True
+            return cls is Protocol and other is typing.Protocol
+
+        # This has to be defined, or the abc-module cache
+        # complains about classes with this metaclass being unhashable,
+        # if we define only __eq__!
+        def __hash__(cls) -> int:
+            return type.__hash__(cls)
+
+    @classmethod
+    def _proto_hook(cls, other):
+        if not cls.__dict__.get('_is_protocol', False):
+            return NotImplemented
+
+        for attr in cls.__protocol_attrs__:
+            for base in other.__mro__:
+                # Check if the members appears in the class dictionary...
+                if attr in base.__dict__:
+                    if base.__dict__[attr] is None:
+                        return NotImplemented
+                    break
+
+                # ...or in annotations, if it is a sub-protocol.
+                annotations = getattr(base, '__annotations__', {})
+                if (
+                    isinstance(annotations, collections.abc.Mapping)
+                    and attr in annotations
+                    and is_protocol(other)
+                ):
+                    break
+            else:
+                return NotImplemented
+        return True
+
+    class Protocol(typing.Generic, metaclass=_ProtocolMeta):
+        __doc__ = typing.Protocol.__doc__
+        __slots__ = ()
+        _is_protocol = True
+        _is_runtime_protocol = False
+
+        def __init_subclass__(cls, *args, **kwargs):
+            super().__init_subclass__(*args, **kwargs)
+
+            # Determine if this is a protocol or a concrete subclass.
+            if not cls.__dict__.get('_is_protocol', False):
+                cls._is_protocol = any(b is Protocol for b in cls.__bases__)
+
+            # Set (or override) the protocol subclass hook.
+            if '__subclasshook__' not in cls.__dict__:
+                cls.__subclasshook__ = _proto_hook
+
+            # Prohibit instantiation for protocol classes
+            if cls._is_protocol and cls.__init__ is Protocol.__init__:
+                cls.__init__ = _no_init
+
+
+if sys.version_info >= (3, 13):
+    runtime_checkable = typing.runtime_checkable
+else:
+    def runtime_checkable(cls):
+        """Mark a protocol class as a runtime protocol.
+
+        Such protocol can be used with isinstance() and issubclass().
+        Raise TypeError if applied to a non-protocol class.
+        This allows a simple-minded structural check very similar to
+        one trick ponies in collections.abc such as Iterable.
+
+        For example::
+
+            @runtime_checkable
+            class Closable(Protocol):
+                def close(self): ...
+
+            assert isinstance(open('/some/file'), Closable)
+
+        Warning: this will check only the presence of the required methods,
+        not their type signatures!
+        """
+        if not issubclass(cls, typing.Generic) or not getattr(cls, '_is_protocol', False):
+            raise TypeError(f'@runtime_checkable can be only applied to protocol classes,'
+                            f' got {cls!r}')
+        cls._is_runtime_protocol = True
+
+        # typing.Protocol classes on <=3.11 break if we execute this block,
+        # because typing.Protocol classes on <=3.11 don't have a
+        # `__protocol_attrs__` attribute, and this block relies on the
+        # `__protocol_attrs__` attribute. Meanwhile, typing.Protocol classes on 3.12.2+
+        # break if we *don't* execute this block, because *they* assume that all
+        # protocol classes have a `__non_callable_proto_members__` attribute
+        # (which this block sets)
+        if isinstance(cls, _ProtocolMeta) or sys.version_info >= (3, 12, 2):
+            # PEP 544 prohibits using issubclass()
+            # with protocols that have non-method members.
+            # See gh-113320 for why we compute this attribute here,
+            # rather than in `_ProtocolMeta.__init__`
+            cls.__non_callable_proto_members__ = set()
+            for attr in cls.__protocol_attrs__:
+                try:
+                    is_callable = callable(getattr(cls, attr, None))
+                except Exception as e:
+                    raise TypeError(
+                        f"Failed to determine whether protocol member {attr!r} "
+                        "is a method member"
+                    ) from e
+                else:
+                    if not is_callable:
+                        cls.__non_callable_proto_members__.add(attr)
+
+        return cls
+
+
+# The "runtime" alias exists for backwards compatibility.
+runtime = runtime_checkable
+
+
+# Our version of runtime-checkable protocols is faster on Python 3.8-3.11
+if sys.version_info >= (3, 12):
+    SupportsInt = typing.SupportsInt
+    SupportsFloat = typing.SupportsFloat
+    SupportsComplex = typing.SupportsComplex
+    SupportsBytes = typing.SupportsBytes
+    SupportsIndex = typing.SupportsIndex
+    SupportsAbs = typing.SupportsAbs
+    SupportsRound = typing.SupportsRound
+else:
+    @runtime_checkable
+    class SupportsInt(Protocol):
+        """An ABC with one abstract method __int__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __int__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsFloat(Protocol):
+        """An ABC with one abstract method __float__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __float__(self) -> float:
+            pass
+
+    @runtime_checkable
+    class SupportsComplex(Protocol):
+        """An ABC with one abstract method __complex__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __complex__(self) -> complex:
+            pass
+
+    @runtime_checkable
+    class SupportsBytes(Protocol):
+        """An ABC with one abstract method __bytes__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __bytes__(self) -> bytes:
+            pass
+
+    @runtime_checkable
+    class SupportsIndex(Protocol):
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __index__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsAbs(Protocol[T_co]):
+        """
+        An ABC with one abstract method __abs__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __abs__(self) -> T_co:
+            pass
+
+    @runtime_checkable
+    class SupportsRound(Protocol[T_co]):
+        """
+        An ABC with one abstract method __round__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __round__(self, ndigits: int = 0) -> T_co:
+            pass
+
+
+def _ensure_subclassable(mro_entries):
+    def inner(func):
+        if sys.implementation.name == "pypy" and sys.version_info < (3, 9):
+            cls_dict = {
+                "__call__": staticmethod(func),
+                "__mro_entries__": staticmethod(mro_entries)
+            }
+            t = type(func.__name__, (), cls_dict)
+            return functools.update_wrapper(t(), func)
+        else:
+            func.__mro_entries__ = mro_entries
+            return func
+    return inner
+
+
+# Update this to something like >=3.13.0b1 if and when
+# PEP 728 is implemented in CPython
+_PEP_728_IMPLEMENTED = False
+
+if _PEP_728_IMPLEMENTED:
+    # The standard library TypedDict in Python 3.8 does not store runtime information
+    # about which (if any) keys are optional.  See https://bugs.python.org/issue38834
+    # The standard library TypedDict in Python 3.9.0/1 does not honour the "total"
+    # keyword with old-style TypedDict().  See https://bugs.python.org/issue42059
+    # The standard library TypedDict below Python 3.11 does not store runtime
+    # information about optional and required keys when using Required or NotRequired.
+    # Generic TypedDicts are also impossible using typing.TypedDict on Python <3.11.
+    # Aaaand on 3.12 we add __orig_bases__ to TypedDict
+    # to enable better runtime introspection.
+    # On 3.13 we deprecate some odd ways of creating TypedDicts.
+    # Also on 3.13, PEP 705 adds the ReadOnly[] qualifier.
+    # PEP 728 (still pending) makes more changes.
+    TypedDict = typing.TypedDict
+    _TypedDictMeta = typing._TypedDictMeta
+    is_typeddict = typing.is_typeddict
+else:
+    # 3.10.0 and later
+    _TAKES_MODULE = "module" in inspect.signature(typing._type_check).parameters
+
+    def _get_typeddict_qualifiers(annotation_type):
+        while True:
+            annotation_origin = get_origin(annotation_type)
+            if annotation_origin is Annotated:
+                annotation_args = get_args(annotation_type)
+                if annotation_args:
+                    annotation_type = annotation_args[0]
+                else:
+                    break
+            elif annotation_origin is Required:
+                yield Required
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is NotRequired:
+                yield NotRequired
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is ReadOnly:
+                yield ReadOnly
+                annotation_type, = get_args(annotation_type)
+            else:
+                break
+
+    class _TypedDictMeta(type):
+        def __new__(cls, name, bases, ns, *, total=True, closed=False):
+            """Create new typed dict class object.
+
+            This method is called when TypedDict is subclassed,
+            or when TypedDict is instantiated. This way
+            TypedDict supports all three syntax forms described in its docstring.
+            Subclasses and instances of TypedDict return actual dictionaries.
+            """
+            for base in bases:
+                if type(base) is not _TypedDictMeta and base is not typing.Generic:
+                    raise TypeError('cannot inherit from both a TypedDict type '
+                                    'and a non-TypedDict base class')
+
+            if any(issubclass(b, typing.Generic) for b in bases):
+                generic_base = (typing.Generic,)
+            else:
+                generic_base = ()
+
+            # typing.py generally doesn't let you inherit from plain Generic, unless
+            # the name of the class happens to be "Protocol"
+            tp_dict = type.__new__(_TypedDictMeta, "Protocol", (*generic_base, dict), ns)
+            tp_dict.__name__ = name
+            if tp_dict.__qualname__ == "Protocol":
+                tp_dict.__qualname__ = name
+
+            if not hasattr(tp_dict, '__orig_bases__'):
+                tp_dict.__orig_bases__ = bases
+
+            annotations = {}
+            if "__annotations__" in ns:
+                own_annotations = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                own_annotations = ns["__annotate__"](1)
+            else:
+                own_annotations = {}
+            msg = "TypedDict('Name', {f0: t0, f1: t1, ...}); each t must be a type"
+            if _TAKES_MODULE:
+                own_annotations = {
+                    n: typing._type_check(tp, msg, module=tp_dict.__module__)
+                    for n, tp in own_annotations.items()
+                }
+            else:
+                own_annotations = {
+                    n: typing._type_check(tp, msg)
+                    for n, tp in own_annotations.items()
+                }
+            required_keys = set()
+            optional_keys = set()
+            readonly_keys = set()
+            mutable_keys = set()
+            extra_items_type = None
+
+            for base in bases:
+                base_dict = base.__dict__
+
+                annotations.update(base_dict.get('__annotations__', {}))
+                required_keys.update(base_dict.get('__required_keys__', ()))
+                optional_keys.update(base_dict.get('__optional_keys__', ()))
+                readonly_keys.update(base_dict.get('__readonly_keys__', ()))
+                mutable_keys.update(base_dict.get('__mutable_keys__', ()))
+                base_extra_items_type = base_dict.get('__extra_items__', None)
+                if base_extra_items_type is not None:
+                    extra_items_type = base_extra_items_type
+
+            if closed and extra_items_type is None:
+                extra_items_type = Never
+            if closed and "__extra_items__" in own_annotations:
+                annotation_type = own_annotations.pop("__extra_items__")
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+                if Required in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "Required"
+                    )
+                if NotRequired in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "NotRequired"
+                    )
+                extra_items_type = annotation_type
+
+            annotations.update(own_annotations)
+            for annotation_key, annotation_type in own_annotations.items():
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+
+                if Required in qualifiers:
+                    required_keys.add(annotation_key)
+                elif NotRequired in qualifiers:
+                    optional_keys.add(annotation_key)
+                elif total:
+                    required_keys.add(annotation_key)
+                else:
+                    optional_keys.add(annotation_key)
+                if ReadOnly in qualifiers:
+                    mutable_keys.discard(annotation_key)
+                    readonly_keys.add(annotation_key)
+                else:
+                    mutable_keys.add(annotation_key)
+                    readonly_keys.discard(annotation_key)
+
+            tp_dict.__annotations__ = annotations
+            tp_dict.__required_keys__ = frozenset(required_keys)
+            tp_dict.__optional_keys__ = frozenset(optional_keys)
+            tp_dict.__readonly_keys__ = frozenset(readonly_keys)
+            tp_dict.__mutable_keys__ = frozenset(mutable_keys)
+            if not hasattr(tp_dict, '__total__'):
+                tp_dict.__total__ = total
+            tp_dict.__closed__ = closed
+            tp_dict.__extra_items__ = extra_items_type
+            return tp_dict
+
+        __call__ = dict  # static method
+
+        def __subclasscheck__(cls, other):
+            # Typed dicts are only for static structural subtyping.
+            raise TypeError('TypedDict does not support instance and class checks')
+
+        __instancecheck__ = __subclasscheck__
+
+    _TypedDict = type.__new__(_TypedDictMeta, 'TypedDict', (), {})
+
+    @_ensure_subclassable(lambda bases: (_TypedDict,))
+    def TypedDict(typename, fields=_marker, /, *, total=True, closed=False, **kwargs):
+        """A simple typed namespace. At runtime it is equivalent to a plain dict.
+
+        TypedDict creates a dictionary type such that a type checker will expect all
+        instances to have a certain set of keys, where each key is
+        associated with a value of a consistent type. This expectation
+        is not checked at runtime.
+
+        Usage::
+
+            class Point2D(TypedDict):
+                x: int
+                y: int
+                label: str
+
+            a: Point2D = {'x': 1, 'y': 2, 'label': 'good'}  # OK
+            b: Point2D = {'z': 3, 'label': 'bad'}           # Fails type check
+
+            assert Point2D(x=1, y=2, label='first') == dict(x=1, y=2, label='first')
+
+        The type info can be accessed via the Point2D.__annotations__ dict, and
+        the Point2D.__required_keys__ and Point2D.__optional_keys__ frozensets.
+        TypedDict supports an additional equivalent form::
+
+            Point2D = TypedDict('Point2D', {'x': int, 'y': int, 'label': str})
+
+        By default, all keys must be present in a TypedDict. It is possible
+        to override this by specifying totality::
+
+            class Point2D(TypedDict, total=False):
+                x: int
+                y: int
+
+        This means that a Point2D TypedDict can have any of the keys omitted. A type
+        checker is only expected to support a literal False or True as the value of
+        the total argument. True is the default, and makes all items defined in the
+        class body be required.
+
+        The Required and NotRequired special forms can also be used to mark
+        individual keys as being required or not required::
+
+            class Point2D(TypedDict):
+                x: int  # the "x" key must always be present (Required is the default)
+                y: NotRequired[int]  # the "y" key can be omitted
+
+        See PEP 655 for more details on Required and NotRequired.
+        """
+        if fields is _marker or fields is None:
+            if fields is _marker:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+
+            example = f"`{typename} = TypedDict({typename!r}, {{}})`"
+            deprecation_msg = (
+                f"{deprecated_thing} is deprecated and will be disallowed in "
+                "Python 3.15. To create a TypedDict class with 0 fields "
+                "using the functional syntax, pass an empty dictionary, e.g. "
+            ) + example + "."
+            warnings.warn(deprecation_msg, DeprecationWarning, stacklevel=2)
+            if closed is not False and closed is not True:
+                kwargs["closed"] = closed
+                closed = False
+            fields = kwargs
+        elif kwargs:
+            raise TypeError("TypedDict takes either a dict or keyword arguments,"
+                            " but not both")
+        if kwargs:
+            if sys.version_info >= (3, 13):
+                raise TypeError("TypedDict takes no keyword arguments")
+            warnings.warn(
+                "The kwargs-based syntax for TypedDict definitions is deprecated "
+                "in Python 3.11, will be removed in Python 3.13, and may not be "
+                "understood by third-party type checkers.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+
+        ns = {'__annotations__': dict(fields)}
+        module = _caller()
+        if module is not None:
+            # Setting correct module is necessary to make typed dict classes pickleable.
+            ns['__module__'] = module
+
+        td = _TypedDictMeta(typename, (), ns, total=total, closed=closed)
+        td.__orig_bases__ = (TypedDict,)
+        return td
+
+    if hasattr(typing, "_TypedDictMeta"):
+        _TYPEDDICT_TYPES = (typing._TypedDictMeta, _TypedDictMeta)
+    else:
+        _TYPEDDICT_TYPES = (_TypedDictMeta,)
+
+    def is_typeddict(tp):
+        """Check if an annotation is a TypedDict class
+
+        For example::
+            class Film(TypedDict):
+                title: str
+                year: int
+
+            is_typeddict(Film)  # => True
+            is_typeddict(Union[list, str])  # => False
+        """
+        # On 3.8, this would otherwise return True
+        if hasattr(typing, "TypedDict") and tp is typing.TypedDict:
+            return False
+        return isinstance(tp, _TYPEDDICT_TYPES)
+
+
+if hasattr(typing, "assert_type"):
+    assert_type = typing.assert_type
+
+else:
+    def assert_type(val, typ, /):
+        """Assert (to the type checker) that the value is of the given type.
+
+        When the type checker encounters a call to assert_type(), it
+        emits an error if the value is not of the specified type::
+
+            def greet(name: str) -> None:
+                assert_type(name, str)  # ok
+                assert_type(name, int)  # type checker error
+
+        At runtime this returns the first argument unchanged and otherwise
+        does nothing.
+        """
+        return val
+
+
+if hasattr(typing, "ReadOnly"):  # 3.13+
+    get_type_hints = typing.get_type_hints
+else:  # <=3.13
+    # replaces _strip_annotations()
+    def _strip_extras(t):
+        """Strips Annotated, Required and NotRequired from a given type."""
+        if isinstance(t, _AnnotatedAlias):
+            return _strip_extras(t.__origin__)
+        if hasattr(t, "__origin__") and t.__origin__ in (Required, NotRequired, ReadOnly):
+            return _strip_extras(t.__args__[0])
+        if isinstance(t, typing._GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return t.copy_with(stripped_args)
+        if hasattr(_types, "GenericAlias") and isinstance(t, _types.GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return _types.GenericAlias(t.__origin__, stripped_args)
+        if hasattr(_types, "UnionType") and isinstance(t, _types.UnionType):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return functools.reduce(operator.or_, stripped_args)
+
+        return t
+
+    def get_type_hints(obj, globalns=None, localns=None, include_extras=False):
+        """Return type hints for an object.
+
+        This is often the same as obj.__annotations__, but it handles
+        forward references encoded as string literals, adds Optional[t] if a
+        default value equal to None is set and recursively replaces all
+        'Annotated[T, ...]', 'Required[T]' or 'NotRequired[T]' with 'T'
+        (unless 'include_extras=True').
+
+        The argument may be a module, class, method, or function. The annotations
+        are returned as a dictionary. For classes, annotations include also
+        inherited members.
+
+        TypeError is raised if the argument is not of a type that can contain
+        annotations, and an empty dictionary is returned if no annotations are
+        present.
+
+        BEWARE -- the behavior of globalns and localns is counterintuitive
+        (unless you are familiar with how eval() and exec() work).  The
+        search order is locals first, then globals.
+
+        - If no dict arguments are passed, an attempt is made to use the
+          globals from obj (or the respective module's globals for classes),
+          and these are also used as the locals.  If the object does not appear
+          to have globals, an empty dictionary is used.
+
+        - If one dict argument is passed, it is used for both globals and
+          locals.
+
+        - If two dict arguments are passed, they specify globals and
+          locals, respectively.
+        """
+        if hasattr(typing, "Annotated"):  # 3.9+
+            hint = typing.get_type_hints(
+                obj, globalns=globalns, localns=localns, include_extras=True
+            )
+        else:  # 3.8
+            hint = typing.get_type_hints(obj, globalns=globalns, localns=localns)
+        if include_extras:
+            return hint
+        return {k: _strip_extras(t) for k, t in hint.items()}
+
+
+# Python 3.9+ has PEP 593 (Annotated)
+if hasattr(typing, 'Annotated'):
+    Annotated = typing.Annotated
+    # Not exported and not a public API, but needed for get_origin() and get_args()
+    # to work.
+    _AnnotatedAlias = typing._AnnotatedAlias
+# 3.8
+else:
+    class _AnnotatedAlias(typing._GenericAlias, _root=True):
+        """Runtime representation of an annotated type.
+
+        At its core 'Annotated[t, dec1, dec2, ...]' is an alias for the type 't'
+        with extra annotations. The alias behaves like a normal typing alias,
+        instantiating is the same as instantiating the underlying type, binding
+        it to types is also the same.
+        """
+        def __init__(self, origin, metadata):
+            if isinstance(origin, _AnnotatedAlias):
+                metadata = origin.__metadata__ + metadata
+                origin = origin.__origin__
+            super().__init__(origin, origin)
+            self.__metadata__ = metadata
+
+        def copy_with(self, params):
+            assert len(params) == 1
+            new_type = params[0]
+            return _AnnotatedAlias(new_type, self.__metadata__)
+
+        def __repr__(self):
+            return (f"typing_extensions.Annotated[{typing._type_repr(self.__origin__)}, "
+                    f"{', '.join(repr(a) for a in self.__metadata__)}]")
+
+        def __reduce__(self):
+            return operator.getitem, (
+                Annotated, (self.__origin__, *self.__metadata__)
+            )
+
+        def __eq__(self, other):
+            if not isinstance(other, _AnnotatedAlias):
+                return NotImplemented
+            if self.__origin__ != other.__origin__:
+                return False
+            return self.__metadata__ == other.__metadata__
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__metadata__))
+
+    class Annotated:
+        """Add context specific metadata to a type.
+
+        Example: Annotated[int, runtime_check.Unsigned] indicates to the
+        hypothetical runtime_check module that this type is an unsigned int.
+        Every other consumer of this type can ignore this metadata and treat
+        this type as int.
+
+        The first argument to Annotated must be a valid type (and will be in
+        the __origin__ field), the remaining arguments are kept as a tuple in
+        the __extra__ field.
+
+        Details:
+
+        - It's an error to call `Annotated` with less than two arguments.
+        - Nested Annotated are flattened::
+
+            Annotated[Annotated[T, Ann1, Ann2], Ann3] == Annotated[T, Ann1, Ann2, Ann3]
+
+        - Instantiating an annotated type is equivalent to instantiating the
+        underlying type::
+
+            Annotated[C, Ann1](5) == C(5)
+
+        - Annotated can be used as a generic type alias::
+
+            Optimized = Annotated[T, runtime.Optimize()]
+            Optimized[int] == Annotated[int, runtime.Optimize()]
+
+            OptimizedList = Annotated[List[T], runtime.Optimize()]
+            OptimizedList[int] == Annotated[List[int], runtime.Optimize()]
+        """
+
+        __slots__ = ()
+
+        def __new__(cls, *args, **kwargs):
+            raise TypeError("Type Annotated cannot be instantiated.")
+
+        @typing._tp_cache
+        def __class_getitem__(cls, params):
+            if not isinstance(params, tuple) or len(params) < 2:
+                raise TypeError("Annotated[...] should be used "
+                                "with at least two arguments (a type and an "
+                                "annotation).")
+            allowed_special_forms = (ClassVar, Final)
+            if get_origin(params[0]) in allowed_special_forms:
+                origin = params[0]
+            else:
+                msg = "Annotated[t, ...]: t must be a type."
+                origin = typing._type_check(params[0], msg)
+            metadata = tuple(params[1:])
+            return _AnnotatedAlias(origin, metadata)
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                f"Cannot subclass {cls.__module__}.Annotated"
+            )
+
+# Python 3.8 has get_origin() and get_args() but those implementations aren't
+# Annotated-aware, so we can't use those. Python 3.9's versions don't support
+# ParamSpecArgs and ParamSpecKwargs, so only Python 3.10's versions will do.
+if sys.version_info[:2] >= (3, 10):
+    get_origin = typing.get_origin
+    get_args = typing.get_args
+# 3.8-3.9
+else:
+    try:
+        # 3.9+
+        from typing import _BaseGenericAlias
+    except ImportError:
+        _BaseGenericAlias = typing._GenericAlias
+    try:
+        # 3.9+
+        from typing import GenericAlias as _typing_GenericAlias
+    except ImportError:
+        _typing_GenericAlias = typing._GenericAlias
+
+    def get_origin(tp):
+        """Get the unsubscripted version of a type.
+
+        This supports generic types, Callable, Tuple, Union, Literal, Final, ClassVar
+        and Annotated. Return None for unsupported types. Examples::
+
+            get_origin(Literal[42]) is Literal
+            get_origin(int) is None
+            get_origin(ClassVar[int]) is ClassVar
+            get_origin(Generic) is Generic
+            get_origin(Generic[T]) is Generic
+            get_origin(Union[T, int]) is Union
+            get_origin(List[Tuple[T, T]][int]) == list
+            get_origin(P.args) is P
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return Annotated
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias, _BaseGenericAlias,
+                           ParamSpecArgs, ParamSpecKwargs)):
+            return tp.__origin__
+        if tp is typing.Generic:
+            return typing.Generic
+        return None
+
+    def get_args(tp):
+        """Get type arguments with all substitutions performed.
+
+        For unions, basic simplifications used by Union constructor are performed.
+        Examples::
+            get_args(Dict[str, int]) == (str, int)
+            get_args(int) == ()
+            get_args(Union[int, Union[T, int], str][int]) == (int, str)
+            get_args(Union[int, Tuple[T, int]][str]) == (int, Tuple[str, int])
+            get_args(Callable[[], T][int]) == ([], int)
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return (tp.__origin__, *tp.__metadata__)
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias)):
+            if getattr(tp, "_special", False):
+                return ()
+            res = tp.__args__
+            if get_origin(tp) is collections.abc.Callable and res[0] is not Ellipsis:
+                res = (list(res[:-1]), res[-1])
+            return res
+        return ()
+
+
+# 3.10+
+if hasattr(typing, 'TypeAlias'):
+    TypeAlias = typing.TypeAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeAlias(self, parameters):
+        """Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example above.
+        """
+        raise TypeError(f"{self} is not subscriptable")
+# 3.8
+else:
+    TypeAlias = _ExtensionsSpecialForm(
+        'TypeAlias',
+        doc="""Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example
+        above."""
+    )
+
+
+if hasattr(typing, "NoDefault"):
+    NoDefault = typing.NoDefault
+else:
+    class NoDefaultTypeMeta(type):
+        def __setattr__(cls, attr, value):
+            # TypeError is consistent with the behavior of NoneType
+            raise TypeError(
+                f"cannot set {attr!r} attribute of immutable type {cls.__name__!r}"
+            )
+
+    class NoDefaultType(metaclass=NoDefaultTypeMeta):
+        """The type of the NoDefault singleton."""
+
+        __slots__ = ()
+
+        def __new__(cls):
+            return globals().get("NoDefault") or object.__new__(cls)
+
+        def __repr__(self):
+            return "typing_extensions.NoDefault"
+
+        def __reduce__(self):
+            return "NoDefault"
+
+    NoDefault = NoDefaultType()
+    del NoDefaultType, NoDefaultTypeMeta
+
+
+def _set_default(type_param, default):
+    type_param.has_default = lambda: default is not NoDefault
+    type_param.__default__ = default
+
+
+def _set_module(typevarlike):
+    # for pickling:
+    def_mod = _caller(depth=3)
+    if def_mod != 'typing_extensions':
+        typevarlike.__module__ = def_mod
+
+
+class _DefaultMixin:
+    """Mixin for TypeVarLike defaults."""
+
+    __slots__ = ()
+    __init__ = _set_default
+
+
+# Classes using this metaclass must provide a _backported_typevarlike ClassVar
+class _TypeVarLikeMeta(type):
+    def __instancecheck__(cls, __instance: Any) -> bool:
+        return isinstance(__instance, cls._backported_typevarlike)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVar
+else:
+    # Add default and infer_variance parameters from PEP 696 and 695
+    class TypeVar(metaclass=_TypeVarLikeMeta):
+        """Type variable."""
+
+        _backported_typevarlike = typing.TypeVar
+
+        def __new__(cls, name, *constraints, bound=None,
+                    covariant=False, contravariant=False,
+                    default=NoDefault, infer_variance=False):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented (3.12+), can pass infer_variance to typing.TypeVar
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant,
+                                         infer_variance=infer_variance)
+            else:
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant)
+                if infer_variance and (covariant or contravariant):
+                    raise ValueError("Variance cannot be specified with infer_variance.")
+                typevar.__infer_variance__ = infer_variance
+
+            _set_default(typevar, default)
+            _set_module(typevar)
+
+            def _tvar_prepare_subst(alias, args):
+                if (
+                    typevar.has_default()
+                    and alias.__parameters__.index(typevar) == len(args)
+                ):
+                    args += (typevar.__default__,)
+                return args
+
+            typevar.__typing_prepare_subst__ = _tvar_prepare_subst
+            return typevar
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.TypeVar' is not an acceptable base type")
+
+
+# Python 3.10+ has PEP 612
+if hasattr(typing, 'ParamSpecArgs'):
+    ParamSpecArgs = typing.ParamSpecArgs
+    ParamSpecKwargs = typing.ParamSpecKwargs
+# 3.8-3.9
+else:
+    class _Immutable:
+        """Mixin to indicate that object should not be copied."""
+        __slots__ = ()
+
+        def __copy__(self):
+            return self
+
+        def __deepcopy__(self, memo):
+            return self
+
+    class ParamSpecArgs(_Immutable):
+        """The args for a ParamSpec object.
+
+        Given a ParamSpec object P, P.args is an instance of ParamSpecArgs.
+
+        ParamSpecArgs objects have a reference back to their ParamSpec:
+
+        P.args.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.args"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecArgs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+    class ParamSpecKwargs(_Immutable):
+        """The kwargs for a ParamSpec object.
+
+        Given a ParamSpec object P, P.kwargs is an instance of ParamSpecKwargs.
+
+        ParamSpecKwargs objects have a reference back to their ParamSpec:
+
+        P.kwargs.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.kwargs"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecKwargs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import ParamSpec
+
+# 3.10+
+elif hasattr(typing, 'ParamSpec'):
+
+    # Add default parameter - PEP 696
+    class ParamSpec(metaclass=_TypeVarLikeMeta):
+        """Parameter specification."""
+
+        _backported_typevarlike = typing.ParamSpec
+
+        def __new__(cls, name, *, bound=None,
+                    covariant=False, contravariant=False,
+                    infer_variance=False, default=NoDefault):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented, can pass infer_variance to typing.TypeVar
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant,
+                                             infer_variance=infer_variance)
+            else:
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant)
+                paramspec.__infer_variance__ = infer_variance
+
+            _set_default(paramspec, default)
+            _set_module(paramspec)
+
+            def _paramspec_prepare_subst(alias, args):
+                params = alias.__parameters__
+                i = params.index(paramspec)
+                if i == len(args) and paramspec.has_default():
+                    args = [*args, paramspec.__default__]
+                if i >= len(args):
+                    raise TypeError(f"Too few arguments for {alias}")
+                # Special case where Z[[int, str, bool]] == Z[int, str, bool] in PEP 612.
+                if len(params) == 1 and not typing._is_param_expr(args[0]):
+                    assert i == 0
+                    args = (args,)
+                # Convert lists to tuples to help other libraries cache the results.
+                elif isinstance(args[i], list):
+                    args = (*args[:i], tuple(args[i]), *args[i + 1:])
+                return args
+
+            paramspec.__typing_prepare_subst__ = _paramspec_prepare_subst
+            return paramspec
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.ParamSpec' is not an acceptable base type")
+
+# 3.8-3.9
+else:
+
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class ParamSpec(list, _DefaultMixin):
+        """Parameter specification variable.
+
+        Usage::
+
+           P = ParamSpec('P')
+
+        Parameter specification variables exist primarily for the benefit of static
+        type checkers.  They are used to forward the parameter types of one
+        callable to another callable, a pattern commonly found in higher order
+        functions and decorators.  They are only valid when used in ``Concatenate``,
+        or s the first argument to ``Callable``. In Python 3.10 and higher,
+        they are also supported in user-defined Generics at runtime.
+        See class Generic for more information on generic types.  An
+        example for annotating a decorator::
+
+           T = TypeVar('T')
+           P = ParamSpec('P')
+
+           def add_logging(f: Callable[P, T]) -> Callable[P, T]:
+               '''A type-safe decorator to add logging to a function.'''
+               def inner(*args: P.args, **kwargs: P.kwargs) -> T:
+                   logging.info(f'{f.__name__} was called')
+                   return f(*args, **kwargs)
+               return inner
+
+           @add_logging
+           def add_two(x: float, y: float) -> float:
+               '''Add two numbers together.'''
+               return x + y
+
+        Parameter specification variables defined with covariant=True or
+        contravariant=True can be used to declare covariant or contravariant
+        generic types.  These keyword arguments are valid, but their actual semantics
+        are yet to be decided.  See PEP 612 for details.
+
+        Parameter specification variables can be introspected. e.g.:
+
+           P.__name__ == 'T'
+           P.__bound__ == None
+           P.__covariant__ == False
+           P.__contravariant__ == False
+
+        Note that only parameter specification variables defined in global scope can
+        be pickled.
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        @property
+        def args(self):
+            return ParamSpecArgs(self)
+
+        @property
+        def kwargs(self):
+            return ParamSpecKwargs(self)
+
+        def __init__(self, name, *, bound=None, covariant=False, contravariant=False,
+                     infer_variance=False, default=NoDefault):
+            list.__init__(self, [self])
+            self.__name__ = name
+            self.__covariant__ = bool(covariant)
+            self.__contravariant__ = bool(contravariant)
+            self.__infer_variance__ = bool(infer_variance)
+            if bound:
+                self.__bound__ = typing._type_check(bound, 'Bound must be a type.')
+            else:
+                self.__bound__ = None
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __repr__(self):
+            if self.__infer_variance__:
+                prefix = ''
+            elif self.__covariant__:
+                prefix = '+'
+            elif self.__contravariant__:
+                prefix = '-'
+            else:
+                prefix = '~'
+            return prefix + self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        # Hack to get typing._type_check to pass.
+        def __call__(self, *args, **kwargs):
+            pass
+
+
+# 3.8-3.9
+if not hasattr(typing, 'Concatenate'):
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class _ConcatenateGenericAlias(list):
+
+        # Trick Generic into looking into this for __parameters__.
+        __class__ = typing._GenericAlias
+
+        # Flag in 3.8.
+        _special = False
+
+        def __init__(self, origin, args):
+            super().__init__(args)
+            self.__origin__ = origin
+            self.__args__ = args
+
+        def __repr__(self):
+            _type_repr = typing._type_repr
+            return (f'{_type_repr(self.__origin__)}'
+                    f'[{", ".join(_type_repr(arg) for arg in self.__args__)}]')
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__args__))
+
+        # Hack to get typing._type_check to pass in Generic.
+        def __call__(self, *args, **kwargs):
+            pass
+
+        @property
+        def __parameters__(self):
+            return tuple(
+                tp for tp in self.__args__ if isinstance(tp, (typing.TypeVar, ParamSpec))
+            )
+
+
+# 3.8-3.9
+@typing._tp_cache
+def _concatenate_getitem(self, parameters):
+    if parameters == ():
+        raise TypeError("Cannot take a Concatenate of no types.")
+    if not isinstance(parameters, tuple):
+        parameters = (parameters,)
+    if not isinstance(parameters[-1], ParamSpec):
+        raise TypeError("The last parameter to Concatenate should be a "
+                        "ParamSpec variable.")
+    msg = "Concatenate[arg, ...]: each arg must be a type."
+    parameters = tuple(typing._type_check(p, msg) for p in parameters)
+    return _ConcatenateGenericAlias(self, parameters)
+
+
+# 3.10+
+if hasattr(typing, 'Concatenate'):
+    Concatenate = typing.Concatenate
+    _ConcatenateGenericAlias = typing._ConcatenateGenericAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def Concatenate(self, parameters):
+        """Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """
+        return _concatenate_getitem(self, parameters)
+# 3.8
+else:
+    class _ConcatenateForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            return _concatenate_getitem(self, parameters)
+
+    Concatenate = _ConcatenateForm(
+        'Concatenate',
+        doc="""Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """)
+
+# 3.10+
+if hasattr(typing, 'TypeGuard'):
+    TypeGuard = typing.TypeGuard
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeGuard(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeGuardForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeGuard = _TypeGuardForm(
+        'TypeGuard',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """)
+
+# 3.13+
+if hasattr(typing, 'TypeIs'):
+    TypeIs = typing.TypeIs
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeIs(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeIsForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeIs = _TypeIsForm(
+        'TypeIs',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """)
+
+
+# Vendored from cpython typing._SpecialFrom
+class _SpecialForm(typing._Final, _root=True):
+    __slots__ = ('_name', '__doc__', '_getitem')
+
+    def __init__(self, getitem):
+        self._getitem = getitem
+        self._name = getitem.__name__
+        self.__doc__ = getitem.__doc__
+
+    def __getattr__(self, item):
+        if item in {'__name__', '__qualname__'}:
+            return self._name
+
+        raise AttributeError(item)
+
+    def __mro_entries__(self, bases):
+        raise TypeError(f"Cannot subclass {self!r}")
+
+    def __repr__(self):
+        return f'typing_extensions.{self._name}'
+
+    def __reduce__(self):
+        return self._name
+
+    def __call__(self, *args, **kwds):
+        raise TypeError(f"Cannot instantiate {self!r}")
+
+    def __or__(self, other):
+        return typing.Union[self, other]
+
+    def __ror__(self, other):
+        return typing.Union[other, self]
+
+    def __instancecheck__(self, obj):
+        raise TypeError(f"{self} cannot be used with isinstance()")
+
+    def __subclasscheck__(self, cls):
+        raise TypeError(f"{self} cannot be used with issubclass()")
+
+    @typing._tp_cache
+    def __getitem__(self, parameters):
+        return self._getitem(self, parameters)
+
+
+if hasattr(typing, "LiteralString"):  # 3.11+
+    LiteralString = typing.LiteralString
+else:
+    @_SpecialForm
+    def LiteralString(self, params):
+        """Represents an arbitrary literal string.
+
+        Example::
+
+          from typing_extensions import LiteralString
+
+          def query(sql: LiteralString) -> ...:
+              ...
+
+          query("SELECT * FROM table")  # ok
+          query(f"SELECT * FROM {input()}")  # not ok
+
+        See PEP 675 for details.
+
+        """
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Self"):  # 3.11+
+    Self = typing.Self
+else:
+    @_SpecialForm
+    def Self(self, params):
+        """Used to spell the type of "self" in classes.
+
+        Example::
+
+          from typing import Self
+
+          class ReturnsSelf:
+              def parse(self, data: bytes) -> Self:
+                  ...
+                  return self
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Never"):  # 3.11+
+    Never = typing.Never
+else:
+    @_SpecialForm
+    def Never(self, params):
+        """The bottom type, a type that has no members.
+
+        This can be used to define a function that should never be
+        called, or a function that never returns::
+
+            from typing_extensions import Never
+
+            def never_call_me(arg: Never) -> None:
+                pass
+
+            def int_or_str(arg: int | str) -> None:
+                never_call_me(arg)  # type checker error
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        never_call_me(arg)  # ok, arg is of type Never
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, 'Required'):  # 3.11+
+    Required = typing.Required
+    NotRequired = typing.NotRequired
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.10
+    @_ExtensionsSpecialForm
+    def Required(self, parameters):
+        """A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+    @_ExtensionsSpecialForm
+    def NotRequired(self, parameters):
+        """A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _RequiredForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    Required = _RequiredForm(
+        'Required',
+        doc="""A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """)
+    NotRequired = _RequiredForm(
+        'NotRequired',
+        doc="""A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """)
+
+
+if hasattr(typing, 'ReadOnly'):
+    ReadOnly = typing.ReadOnly
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.12
+    @_ExtensionsSpecialForm
+    def ReadOnly(self, parameters):
+        """A special typing construct to mark an item of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this property.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _ReadOnlyForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    ReadOnly = _ReadOnlyForm(
+        'ReadOnly',
+        doc="""A special typing construct to mark a key of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this propery.
+        """)
+
+
+_UNPACK_DOC = """\
+Type unpack operator.
+
+The type unpack operator takes the child types from some container type,
+such as `tuple[int, str]` or a `TypeVarTuple`, and 'pulls them out'. For
+example:
+
+  # For some generic class `Foo`:
+  Foo[Unpack[tuple[int, str]]]  # Equivalent to Foo[int, str]
+
+  Ts = TypeVarTuple('Ts')
+  # Specifies that `Bar` is generic in an arbitrary number of types.
+  # (Think of `Ts` as a tuple of an arbitrary number of individual
+  #  `TypeVar`s, which the `Unpack` is 'pulling out' directly into the
+  #  `Generic[]`.)
+  class Bar(Generic[Unpack[Ts]]): ...
+  Bar[int]  # Valid
+  Bar[int, str]  # Also valid
+
+From Python 3.11, this can also be done using the `*` operator:
+
+    Foo[*tuple[int, str]]
+    class Bar(Generic[*Ts]): ...
+
+The operator can also be used along with a `TypedDict` to annotate
+`**kwargs` in a function signature. For instance:
+
+  class Movie(TypedDict):
+    name: str
+    year: int
+
+  # This function expects two keyword arguments - *name* of type `str` and
+  # *year* of type `int`.
+  def foo(**kwargs: Unpack[Movie]): ...
+
+Note that there is only some runtime checking of this operator. Not
+everything the runtime allows may be accepted by static type checkers.
+
+For more information, see PEP 646 and PEP 692.
+"""
+
+
+if sys.version_info >= (3, 12):  # PEP 692 changed the repr of Unpack[]
+    Unpack = typing.Unpack
+
+    def _is_unpack(obj):
+        return get_origin(obj) is Unpack
+
+elif sys.version_info[:2] >= (3, 9):  # 3.9+
+    class _UnpackSpecialForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, getitem):
+            super().__init__(getitem)
+            self.__doc__ = _UNPACK_DOC
+
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+        @property
+        def __typing_unpacked_tuple_args__(self):
+            assert self.__origin__ is Unpack
+            assert len(self.__args__) == 1
+            arg, = self.__args__
+            if isinstance(arg, (typing._GenericAlias, _types.GenericAlias)):
+                if arg.__origin__ is not tuple:
+                    raise TypeError("Unpack[...] must be used with a tuple type")
+                return arg.__args__
+            return None
+
+    @_UnpackSpecialForm
+    def Unpack(self, parameters):
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return _UnpackAlias(self, (item,))
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+else:  # 3.8
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+    class _UnpackForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return _UnpackAlias(self, (item,))
+
+    Unpack = _UnpackForm('Unpack', doc=_UNPACK_DOC)
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVarTuple
+
+elif hasattr(typing, "TypeVarTuple"):  # 3.11+
+
+    def _unpack_args(*args):
+        newargs = []
+        for arg in args:
+            subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+            if subargs is not None and not (subargs and subargs[-1] is ...):
+                newargs.extend(subargs)
+            else:
+                newargs.append(arg)
+        return newargs
+
+    # Add default parameter - PEP 696
+    class TypeVarTuple(metaclass=_TypeVarLikeMeta):
+        """Type variable tuple."""
+
+        _backported_typevarlike = typing.TypeVarTuple
+
+        def __new__(cls, name, *, default=NoDefault):
+            tvt = typing.TypeVarTuple(name)
+            _set_default(tvt, default)
+            _set_module(tvt)
+
+            def _typevartuple_prepare_subst(alias, args):
+                params = alias.__parameters__
+                typevartuple_index = params.index(tvt)
+                for param in params[typevartuple_index + 1:]:
+                    if isinstance(param, TypeVarTuple):
+                        raise TypeError(
+                            f"More than one TypeVarTuple parameter in {alias}"
+                        )
+
+                alen = len(args)
+                plen = len(params)
+                left = typevartuple_index
+                right = plen - typevartuple_index - 1
+                var_tuple_index = None
+                fillarg = None
+                for k, arg in enumerate(args):
+                    if not isinstance(arg, type):
+                        subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+                        if subargs and len(subargs) == 2 and subargs[-1] is ...:
+                            if var_tuple_index is not None:
+                                raise TypeError(
+                                    "More than one unpacked "
+                                    "arbitrary-length tuple argument"
+                                )
+                            var_tuple_index = k
+                            fillarg = subargs[0]
+                if var_tuple_index is not None:
+                    left = min(left, var_tuple_index)
+                    right = min(right, alen - var_tuple_index - 1)
+                elif left + right > alen:
+                    raise TypeError(f"Too few arguments for {alias};"
+                                    f" actual {alen}, expected at least {plen - 1}")
+                if left == alen - right and tvt.has_default():
+                    replacement = _unpack_args(tvt.__default__)
+                else:
+                    replacement = args[left: alen - right]
+
+                return (
+                    *args[:left],
+                    *([fillarg] * (typevartuple_index - left)),
+                    replacement,
+                    *([fillarg] * (plen - right - left - typevartuple_index - 1)),
+                    *args[alen - right:],
+                )
+
+            tvt.__typing_prepare_subst__ = _typevartuple_prepare_subst
+            return tvt
+
+        def __init_subclass__(self, *args, **kwds):
+            raise TypeError("Cannot subclass special typing classes")
+
+else:  # <=3.10
+    class TypeVarTuple(_DefaultMixin):
+        """Type variable tuple.
+
+        Usage::
+
+            Ts = TypeVarTuple('Ts')
+
+        In the same way that a normal type variable is a stand-in for a single
+        type such as ``int``, a type variable *tuple* is a stand-in for a *tuple*
+        type such as ``Tuple[int, str]``.
+
+        Type variable tuples can be used in ``Generic`` declarations.
+        Consider the following example::
+
+            class Array(Generic[*Ts]): ...
+
+        The ``Ts`` type variable tuple here behaves like ``tuple[T1, T2]``,
+        where ``T1`` and ``T2`` are type variables. To use these type variables
+        as type parameters of ``Array``, we must *unpack* the type variable tuple using
+        the star operator: ``*Ts``. The signature of ``Array`` then behaves
+        as if we had simply written ``class Array(Generic[T1, T2]): ...``.
+        In contrast to ``Generic[T1, T2]``, however, ``Generic[*Shape]`` allows
+        us to parameterise the class with an *arbitrary* number of type parameters.
+
+        Type variable tuples can be used anywhere a normal ``TypeVar`` can.
+        This includes class definitions, as shown above, as well as function
+        signatures and variable annotations::
+
+            class Array(Generic[*Ts]):
+
+                def __init__(self, shape: Tuple[*Ts]):
+                    self._shape: Tuple[*Ts] = shape
+
+                def get_shape(self) -> Tuple[*Ts]:
+                    return self._shape
+
+            shape = (Height(480), Width(640))
+            x: Array[Height, Width] = Array(shape)
+            y = abs(x)  # Inferred type is Array[Height, Width]
+            z = x + x   #        ...    is Array[Height, Width]
+            x.get_shape()  #     ...    is tuple[Height, Width]
+
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        def __iter__(self):
+            yield self.__unpacked__
+
+        def __init__(self, name, *, default=NoDefault):
+            self.__name__ = name
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+            self.__unpacked__ = Unpack[self]
+
+        def __repr__(self):
+            return self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(self, *args, **kwds):
+            if '_root' not in kwds:
+                raise TypeError("Cannot subclass special typing classes")
+
+
+if hasattr(typing, "reveal_type"):  # 3.11+
+    reveal_type = typing.reveal_type
+else:  # <=3.10
+    def reveal_type(obj: T, /) -> T:
+        """Reveal the inferred type of a variable.
+
+        When a static type checker encounters a call to ``reveal_type()``,
+        it will emit the inferred type of the argument::
+
+            x: int = 1
+            reveal_type(x)
+
+        Running a static type checker (e.g., ``mypy``) on this example
+        will produce output similar to 'Revealed type is "builtins.int"'.
+
+        At runtime, the function prints the runtime type of the
+        argument and returns it unchanged.
+
+        """
+        print(f"Runtime type is {type(obj).__name__!r}", file=sys.stderr)
+        return obj
+
+
+if hasattr(typing, "_ASSERT_NEVER_REPR_MAX_LENGTH"):  # 3.11+
+    _ASSERT_NEVER_REPR_MAX_LENGTH = typing._ASSERT_NEVER_REPR_MAX_LENGTH
+else:  # <=3.10
+    _ASSERT_NEVER_REPR_MAX_LENGTH = 100
+
+
+if hasattr(typing, "assert_never"):  # 3.11+
+    assert_never = typing.assert_never
+else:  # <=3.10
+    def assert_never(arg: Never, /) -> Never:
+        """Assert to the type checker that a line of code is unreachable.
+
+        Example::
+
+            def int_or_str(arg: int | str) -> None:
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        assert_never(arg)
+
+        If a type checker finds that a call to assert_never() is
+        reachable, it will emit an error.
+
+        At runtime, this throws an exception when called.
+
+        """
+        value = repr(arg)
+        if len(value) > _ASSERT_NEVER_REPR_MAX_LENGTH:
+            value = value[:_ASSERT_NEVER_REPR_MAX_LENGTH] + '...'
+        raise AssertionError(f"Expected code to be unreachable, but got: {value}")
+
+
+if sys.version_info >= (3, 12):  # 3.12+
+    # dataclass_transform exists in 3.11 but lacks the frozen_default parameter
+    dataclass_transform = typing.dataclass_transform
+else:  # <=3.11
+    def dataclass_transform(
+        *,
+        eq_default: bool = True,
+        order_default: bool = False,
+        kw_only_default: bool = False,
+        frozen_default: bool = False,
+        field_specifiers: typing.Tuple[
+            typing.Union[typing.Type[typing.Any], typing.Callable[..., typing.Any]],
+            ...
+        ] = (),
+        **kwargs: typing.Any,
+    ) -> typing.Callable[[T], T]:
+        """Decorator that marks a function, class, or metaclass as providing
+        dataclass-like behavior.
+
+        Example:
+
+            from typing_extensions import dataclass_transform
+
+            _T = TypeVar("_T")
+
+            # Used on a decorator function
+            @dataclass_transform()
+            def create_model(cls: type[_T]) -> type[_T]:
+                ...
+                return cls
+
+            @create_model
+            class CustomerModel:
+                id: int
+                name: str
+
+            # Used on a base class
+            @dataclass_transform()
+            class ModelBase: ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+            # Used on a metaclass
+            @dataclass_transform()
+            class ModelMeta(type): ...
+
+            class ModelBase(metaclass=ModelMeta): ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+        Each of the ``CustomerModel`` classes defined in this example will now
+        behave similarly to a dataclass created with the ``@dataclasses.dataclass``
+        decorator. For example, the type checker will synthesize an ``__init__``
+        method.
+
+        The arguments to this decorator can be used to customize this behavior:
+        - ``eq_default`` indicates whether the ``eq`` parameter is assumed to be
+          True or False if it is omitted by the caller.
+        - ``order_default`` indicates whether the ``order`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``kw_only_default`` indicates whether the ``kw_only`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``frozen_default`` indicates whether the ``frozen`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``field_specifiers`` specifies a static list of supported classes
+          or functions that describe fields, similar to ``dataclasses.field()``.
+
+        At runtime, this decorator records its arguments in the
+        ``__dataclass_transform__`` attribute on the decorated object.
+
+        See PEP 681 for details.
+
+        """
+        def decorator(cls_or_fn):
+            cls_or_fn.__dataclass_transform__ = {
+                "eq_default": eq_default,
+                "order_default": order_default,
+                "kw_only_default": kw_only_default,
+                "frozen_default": frozen_default,
+                "field_specifiers": field_specifiers,
+                "kwargs": kwargs,
+            }
+            return cls_or_fn
+        return decorator
+
+
+if hasattr(typing, "override"):  # 3.12+
+    override = typing.override
+else:  # <=3.11
+    _F = typing.TypeVar("_F", bound=typing.Callable[..., typing.Any])
+
+    def override(arg: _F, /) -> _F:
+        """Indicate that a method is intended to override a method in a base class.
+
+        Usage:
+
+            class Base:
+                def method(self) -> None:
+                    pass
+
+            class Child(Base):
+                @override
+                def method(self) -> None:
+                    super().method()
+
+        When this decorator is applied to a method, the type checker will
+        validate that it overrides a method with the same name on a base class.
+        This helps prevent bugs that may occur when a base class is changed
+        without an equivalent change to a child class.
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__override__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+
+        See PEP 698 for details.
+
+        """
+        try:
+            arg.__override__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return arg
+
+
+if hasattr(warnings, "deprecated"):
+    deprecated = warnings.deprecated
+else:
+    _T = typing.TypeVar("_T")
+
+    class deprecated:
+        """Indicate that a class, function or overload is deprecated.
+
+        When this decorator is applied to an object, the type checker
+        will generate a diagnostic on usage of the deprecated object.
+
+        Usage:
+
+            @deprecated("Use B instead")
+            class A:
+                pass
+
+            @deprecated("Use g instead")
+            def f():
+                pass
+
+            @overload
+            @deprecated("int support is deprecated")
+            def g(x: int) -> int: ...
+            @overload
+            def g(x: str) -> int: ...
+
+        The warning specified by *category* will be emitted at runtime
+        on use of deprecated objects. For functions, that happens on calls;
+        for classes, on instantiation and on creation of subclasses.
+        If the *category* is ``None``, no warning is emitted at runtime.
+        The *stacklevel* determines where the
+        warning is emitted. If it is ``1`` (the default), the warning
+        is emitted at the direct caller of the deprecated object; if it
+        is higher, it is emitted further up the stack.
+        Static type checker behavior is not affected by the *category*
+        and *stacklevel* arguments.
+
+        The deprecation message passed to the decorator is saved in the
+        ``__deprecated__`` attribute on the decorated object.
+        If applied to an overload, the decorator
+        must be after the ``@overload`` decorator for the attribute to
+        exist on the overload as returned by ``get_overloads()``.
+
+        See PEP 702 for details.
+
+        """
+        def __init__(
+            self,
+            message: str,
+            /,
+            *,
+            category: typing.Optional[typing.Type[Warning]] = DeprecationWarning,
+            stacklevel: int = 1,
+        ) -> None:
+            if not isinstance(message, str):
+                raise TypeError(
+                    "Expected an object of type str for 'message', not "
+                    f"{type(message).__name__!r}"
+                )
+            self.message = message
+            self.category = category
+            self.stacklevel = stacklevel
+
+        def __call__(self, arg: _T, /) -> _T:
+            # Make sure the inner functions created below don't
+            # retain a reference to self.
+            msg = self.message
+            category = self.category
+            stacklevel = self.stacklevel
+            if category is None:
+                arg.__deprecated__ = msg
+                return arg
+            elif isinstance(arg, type):
+                import functools
+                from types import MethodType
+
+                original_new = arg.__new__
+
+                @functools.wraps(original_new)
+                def __new__(cls, *args, **kwargs):
+                    if cls is arg:
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    if original_new is not object.__new__:
+                        return original_new(cls, *args, **kwargs)
+                    # Mirrors a similar check in object.__new__.
+                    elif cls.__init__ is object.__init__ and (args or kwargs):
+                        raise TypeError(f"{cls.__name__}() takes no arguments")
+                    else:
+                        return original_new(cls)
+
+                arg.__new__ = staticmethod(__new__)
+
+                original_init_subclass = arg.__init_subclass__
+                # We need slightly different behavior if __init_subclass__
+                # is a bound method (likely if it was implemented in Python)
+                if isinstance(original_init_subclass, MethodType):
+                    original_init_subclass = original_init_subclass.__func__
+
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = classmethod(__init_subclass__)
+                # Or otherwise, which likely means it's a builtin such as
+                # object's implementation of __init_subclass__.
+                else:
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = __init_subclass__
+
+                arg.__deprecated__ = __new__.__deprecated__ = msg
+                __init_subclass__.__deprecated__ = msg
+                return arg
+            elif callable(arg):
+                import functools
+
+                @functools.wraps(arg)
+                def wrapper(*args, **kwargs):
+                    warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    return arg(*args, **kwargs)
+
+                arg.__deprecated__ = wrapper.__deprecated__ = msg
+                return wrapper
+            else:
+                raise TypeError(
+                    "@deprecated decorator with non-None category must be applied to "
+                    f"a class or callable, not {arg!r}"
+                )
+
+
+# We have to do some monkey patching to deal with the dual nature of
+# Unpack/TypeVarTuple:
+# - We want Unpack to be a kind of TypeVar so it gets accepted in
+#   Generic[Unpack[Ts]]
+# - We want it to *not* be treated as a TypeVar for the purposes of
+#   counting generic parameters, so that when we subscript a generic,
+#   the runtime doesn't try to substitute the Unpack with the subscripted type.
+if not hasattr(typing, "TypeVarTuple"):
+    def _check_generic(cls, parameters, elen=_marker):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        if elen is _marker:
+            if not hasattr(cls, "__parameters__") or not cls.__parameters__:
+                raise TypeError(f"{cls} is not a generic class")
+            elen = len(cls.__parameters__)
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+                num_tv_tuples = sum(isinstance(p, TypeVarTuple) for p in parameters)
+                if (num_tv_tuples > 0) and (alen >= elen - num_tv_tuples):
+                    return
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            things = "arguments" if sys.version_info >= (3, 10) else "parameters"
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} {things}"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+else:
+    # Python 3.11+
+
+    def _check_generic(cls, parameters, elen):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} arguments"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+
+if not _PEP_696_IMPLEMENTED:
+    typing._check_generic = _check_generic
+
+
+def _has_generic_or_protocol_as_origin() -> bool:
+    try:
+        frame = sys._getframe(2)
+    # - Catch AttributeError: not all Python implementations have sys._getframe()
+    # - Catch ValueError: maybe we're called from an unexpected module
+    #   and the call stack isn't deep enough
+    except (AttributeError, ValueError):
+        return False  # err on the side of leniency
+    else:
+        # If we somehow get invoked from outside typing.py,
+        # also err on the side of leniency
+        if frame.f_globals.get("__name__") != "typing":
+            return False
+        origin = frame.f_locals.get("origin")
+        # Cannot use "in" because origin may be an object with a buggy __eq__ that
+        # throws an error.
+        return origin is typing.Generic or origin is Protocol or origin is typing.Protocol
+
+
+_TYPEVARTUPLE_TYPES = {TypeVarTuple, getattr(typing, "TypeVarTuple", None)}
+
+
+def _is_unpacked_typevartuple(x) -> bool:
+    if get_origin(x) is not Unpack:
+        return False
+    args = get_args(x)
+    return (
+        bool(args)
+        and len(args) == 1
+        and type(args[0]) in _TYPEVARTUPLE_TYPES
+    )
+
+
+# Python 3.11+ _collect_type_vars was renamed to _collect_parameters
+if hasattr(typing, '_collect_type_vars'):
+    def _collect_type_vars(types, typevar_types=None):
+        """Collect all type variable contained in types in order of
+        first appearance (lexicographic order). For example::
+
+            _collect_type_vars((T, List[S, T])) == (T, S)
+        """
+        if typevar_types is None:
+            typevar_types = typing.TypeVar
+        tvars = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with a default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in types:
+            if _is_unpacked_typevartuple(t):
+                type_var_tuple_encountered = True
+            elif isinstance(t, typevar_types) and t not in tvars:
+                if enforce_default_ordering:
+                    has_default = getattr(t, '__default__', NoDefault) is not NoDefault
+                    if has_default:
+                        if type_var_tuple_encountered:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+                        default_encountered = True
+                    elif default_encountered:
+                        raise TypeError(f'Type parameter {t!r} without a default'
+                                        ' follows type parameter with a default')
+
+                tvars.append(t)
+            if _should_collect_from_parameters(t):
+                tvars.extend([t for t in t.__parameters__ if t not in tvars])
+        return tuple(tvars)
+
+    typing._collect_type_vars = _collect_type_vars
+else:
+    def _collect_parameters(args):
+        """Collect all type variables and parameter specifications in args
+        in order of first appearance (lexicographic order).
+
+        For example::
+
+            assert _collect_parameters((T, Callable[P, T])) == (T, P)
+        """
+        parameters = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in args:
+            if isinstance(t, type):
+                # We don't want __parameters__ descriptor of a bare Python class.
+                pass
+            elif isinstance(t, tuple):
+                # `t` might be a tuple, when `ParamSpec` is substituted with
+                # `[T, int]`, or `[int, *Ts]`, etc.
+                for x in t:
+                    for collected in _collect_parameters([x]):
+                        if collected not in parameters:
+                            parameters.append(collected)
+            elif hasattr(t, '__typing_subst__'):
+                if t not in parameters:
+                    if enforce_default_ordering:
+                        has_default = (
+                            getattr(t, '__default__', NoDefault) is not NoDefault
+                        )
+
+                        if type_var_tuple_encountered and has_default:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+
+                        if has_default:
+                            default_encountered = True
+                        elif default_encountered:
+                            raise TypeError(f'Type parameter {t!r} without a default'
+                                            ' follows type parameter with a default')
+
+                    parameters.append(t)
+            else:
+                if _is_unpacked_typevartuple(t):
+                    type_var_tuple_encountered = True
+                for x in getattr(t, '__parameters__', ()):
+                    if x not in parameters:
+                        parameters.append(x)
+
+        return tuple(parameters)
+
+    if not _PEP_696_IMPLEMENTED:
+        typing._collect_parameters = _collect_parameters
+
+# Backport typing.NamedTuple as it exists in Python 3.13.
+# In 3.11, the ability to define generic `NamedTuple`s was supported.
+# This was explicitly disallowed in 3.9-3.10, and only half-worked in <=3.8.
+# On 3.12, we added __orig_bases__ to call-based NamedTuples
+# On 3.13, we deprecated kwargs-based NamedTuples
+if sys.version_info >= (3, 13):
+    NamedTuple = typing.NamedTuple
+else:
+    def _make_nmtuple(name, types, module, defaults=()):
+        fields = [n for n, t in types]
+        annotations = {n: typing._type_check(t, f"field {n} annotation must be a type")
+                       for n, t in types}
+        nm_tpl = collections.namedtuple(name, fields,
+                                        defaults=defaults, module=module)
+        nm_tpl.__annotations__ = nm_tpl.__new__.__annotations__ = annotations
+        # The `_field_types` attribute was removed in 3.9;
+        # in earlier versions, it is the same as the `__annotations__` attribute
+        if sys.version_info < (3, 9):
+            nm_tpl._field_types = annotations
+        return nm_tpl
+
+    _prohibited_namedtuple_fields = typing._prohibited
+    _special_namedtuple_fields = frozenset({'__module__', '__name__', '__annotations__'})
+
+    class _NamedTupleMeta(type):
+        def __new__(cls, typename, bases, ns):
+            assert _NamedTuple in bases
+            for base in bases:
+                if base is not _NamedTuple and base is not typing.Generic:
+                    raise TypeError(
+                        'can only inherit from a NamedTuple type and Generic')
+            bases = tuple(tuple if base is _NamedTuple else base for base in bases)
+            if "__annotations__" in ns:
+                types = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                types = ns["__annotate__"](1)
+            else:
+                types = {}
+            default_names = []
+            for field_name in types:
+                if field_name in ns:
+                    default_names.append(field_name)
+                elif default_names:
+                    raise TypeError(f"Non-default namedtuple field {field_name} "
+                                    f"cannot follow default field"
+                                    f"{'s' if len(default_names) > 1 else ''} "
+                                    f"{', '.join(default_names)}")
+            nm_tpl = _make_nmtuple(
+                typename, types.items(),
+                defaults=[ns[n] for n in default_names],
+                module=ns['__module__']
+            )
+            nm_tpl.__bases__ = bases
+            if typing.Generic in bases:
+                if hasattr(typing, '_generic_class_getitem'):  # 3.12+
+                    nm_tpl.__class_getitem__ = classmethod(typing._generic_class_getitem)
+                else:
+                    class_getitem = typing.Generic.__class_getitem__.__func__
+                    nm_tpl.__class_getitem__ = classmethod(class_getitem)
+            # update from user namespace without overriding special namedtuple attributes
+            for key, val in ns.items():
+                if key in _prohibited_namedtuple_fields:
+                    raise AttributeError("Cannot overwrite NamedTuple attribute " + key)
+                elif key not in _special_namedtuple_fields:
+                    if key not in nm_tpl._fields:
+                        setattr(nm_tpl, key, ns[key])
+                    try:
+                        set_name = type(val).__set_name__
+                    except AttributeError:
+                        pass
+                    else:
+                        try:
+                            set_name(val, nm_tpl, key)
+                        except BaseException as e:
+                            msg = (
+                                f"Error calling __set_name__ on {type(val).__name__!r} "
+                                f"instance {key!r} in {typename!r}"
+                            )
+                            # BaseException.add_note() existed on py311,
+                            # but the __set_name__ machinery didn't start
+                            # using add_note() until py312.
+                            # Making sure exceptions are raised in the same way
+                            # as in "normal" classes seems most important here.
+                            if sys.version_info >= (3, 12):
+                                e.add_note(msg)
+                                raise
+                            else:
+                                raise RuntimeError(msg) from e
+
+            if typing.Generic in bases:
+                nm_tpl.__init_subclass__()
+            return nm_tpl
+
+    _NamedTuple = type.__new__(_NamedTupleMeta, 'NamedTuple', (), {})
+
+    def _namedtuple_mro_entries(bases):
+        assert NamedTuple in bases
+        return (_NamedTuple,)
+
+    @_ensure_subclassable(_namedtuple_mro_entries)
+    def NamedTuple(typename, fields=_marker, /, **kwargs):
+        """Typed version of namedtuple.
+
+        Usage::
+
+            class Employee(NamedTuple):
+                name: str
+                id: int
+
+        This is equivalent to::
+
+            Employee = collections.namedtuple('Employee', ['name', 'id'])
+
+        The resulting class has an extra __annotations__ attribute, giving a
+        dict that maps field names to types.  (The field names are also in
+        the _fields attribute, which is part of the namedtuple API.)
+        An alternative equivalent functional syntax is also accepted::
+
+            Employee = NamedTuple('Employee', [('name', str), ('id', int)])
+        """
+        if fields is _marker:
+            if kwargs:
+                deprecated_thing = "Creating NamedTuple classes using keyword arguments"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "Use the class-based or functional syntax instead."
+                )
+            else:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif fields is None:
+            if kwargs:
+                raise TypeError(
+                    "Cannot pass `None` as the 'fields' parameter "
+                    "and also specify fields using keyword arguments"
+                )
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif kwargs:
+            raise TypeError("Either list of fields or keywords"
+                            " can be provided to NamedTuple, not both")
+        if fields is _marker or fields is None:
+            warnings.warn(
+                deprecation_msg.format(name=deprecated_thing, remove="3.15"),
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            fields = kwargs.items()
+        nt = _make_nmtuple(typename, fields, module=_caller())
+        nt.__orig_bases__ = (NamedTuple,)
+        return nt
+
+
+if hasattr(collections.abc, "Buffer"):
+    Buffer = collections.abc.Buffer
+else:
+    class Buffer(abc.ABC):  # noqa: B024
+        """Base class for classes that implement the buffer protocol.
+
+        The buffer protocol allows Python objects to expose a low-level
+        memory buffer interface. Before Python 3.12, it is not possible
+        to implement the buffer protocol in pure Python code, or even
+        to check whether a class implements the buffer protocol. In
+        Python 3.12 and higher, the ``__buffer__`` method allows access
+        to the buffer protocol from Python code, and the
+        ``collections.abc.Buffer`` ABC allows checking whether a class
+        implements the buffer protocol.
+
+        To indicate support for the buffer protocol in earlier versions,
+        inherit from this ABC, either in a stub file or at runtime,
+        or use ABC registration. This ABC provides no methods, because
+        there is no Python-accessible methods shared by pre-3.12 buffer
+        classes. It is useful primarily for static checks.
+
+        """
+
+    # As a courtesy, register the most common stdlib buffer classes.
+    Buffer.register(memoryview)
+    Buffer.register(bytearray)
+    Buffer.register(bytes)
+
+
+# Backport of types.get_original_bases, available on 3.12+ in CPython
+if hasattr(_types, "get_original_bases"):
+    get_original_bases = _types.get_original_bases
+else:
+    def get_original_bases(cls, /):
+        """Return the class's "original" bases prior to modification by `__mro_entries__`.
+
+        Examples::
+
+            from typing import TypeVar, Generic
+            from typing_extensions import NamedTuple, TypedDict
+
+            T = TypeVar("T")
+            class Foo(Generic[T]): ...
+            class Bar(Foo[int], float): ...
+            class Baz(list[str]): ...
+            Eggs = NamedTuple("Eggs", [("a", int), ("b", str)])
+            Spam = TypedDict("Spam", {"a": int, "b": str})
+
+            assert get_original_bases(Bar) == (Foo[int], float)
+            assert get_original_bases(Baz) == (list[str],)
+            assert get_original_bases(Eggs) == (NamedTuple,)
+            assert get_original_bases(Spam) == (TypedDict,)
+            assert get_original_bases(int) == (object,)
+        """
+        try:
+            return cls.__dict__.get("__orig_bases__", cls.__bases__)
+        except AttributeError:
+            raise TypeError(
+                f'Expected an instance of type, not {type(cls).__name__!r}'
+            ) from None
+
+
+# NewType is a class on Python 3.10+, making it pickleable
+# The error message for subclassing instances of NewType was improved on 3.11+
+if sys.version_info >= (3, 11):
+    NewType = typing.NewType
+else:
+    class NewType:
+        """NewType creates simple unique types with almost zero
+        runtime overhead. NewType(name, tp) is considered a subtype of tp
+        by static type checkers. At runtime, NewType(name, tp) returns
+        a dummy callable that simply returns its argument. Usage::
+            UserId = NewType('UserId', int)
+            def name_by_id(user_id: UserId) -> str:
+                ...
+            UserId('user')          # Fails type check
+            name_by_id(42)          # Fails type check
+            name_by_id(UserId(42))  # OK
+            num = UserId(5) + 1     # type: int
+        """
+
+        def __call__(self, obj, /):
+            return obj
+
+        def __init__(self, name, tp):
+            self.__qualname__ = name
+            if '.' in name:
+                name = name.rpartition('.')[-1]
+            self.__name__ = name
+            self.__supertype__ = tp
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __mro_entries__(self, bases):
+            # We defined __mro_entries__ to get a better error message
+            # if a user attempts to subclass a NewType instance. bpo-46170
+            supercls_name = self.__name__
+
+            class Dummy:
+                def __init_subclass__(cls):
+                    subcls_name = cls.__name__
+                    raise TypeError(
+                        f"Cannot subclass an instance of NewType. "
+                        f"Perhaps you were looking for: "
+                        f"`{subcls_name} = NewType({subcls_name!r}, {supercls_name})`"
+                    )
+
+            return (Dummy,)
+
+        def __repr__(self):
+            return f'{self.__module__}.{self.__qualname__}'
+
+        def __reduce__(self):
+            return self.__qualname__
+
+        if sys.version_info >= (3, 10):
+            # PEP 604 methods
+            # It doesn't make sense to have these methods on Python <3.10
+
+            def __or__(self, other):
+                return typing.Union[self, other]
+
+            def __ror__(self, other):
+                return typing.Union[other, self]
+
+
+if hasattr(typing, "TypeAliasType"):
+    TypeAliasType = typing.TypeAliasType
+else:
+    def _is_unionable(obj):
+        """Corresponds to is_unionable() in unionobject.c in CPython."""
+        return obj is None or isinstance(obj, (
+            type,
+            _types.GenericAlias,
+            _types.UnionType,
+            TypeAliasType,
+        ))
+
+    class TypeAliasType:
+        """Create named, parameterized type aliases.
+
+        This provides a backport of the new `type` statement in Python 3.12:
+
+            type ListOrSet[T] = list[T] | set[T]
+
+        is equivalent to:
+
+            T = TypeVar("T")
+            ListOrSet = TypeAliasType("ListOrSet", list[T] | set[T], type_params=(T,))
+
+        The name ListOrSet can then be used as an alias for the type it refers to.
+
+        The type_params argument should contain all the type parameters used
+        in the value of the type alias. If the alias is not generic, this
+        argument is omitted.
+
+        Static type checkers should only support type aliases declared using
+        TypeAliasType that follow these rules:
+
+        - The first argument (the name) must be a string literal.
+        - The TypeAliasType instance must be immediately assigned to a variable
+          of the same name. (For example, 'X = TypeAliasType("Y", int)' is invalid,
+          as is 'X, Y = TypeAliasType("X", int), TypeAliasType("Y", int)').
+
+        """
+
+        def __init__(self, name: str, value, *, type_params=()):
+            if not isinstance(name, str):
+                raise TypeError("TypeAliasType name must be a string")
+            self.__value__ = value
+            self.__type_params__ = type_params
+
+            parameters = []
+            for type_param in type_params:
+                if isinstance(type_param, TypeVarTuple):
+                    parameters.extend(type_param)
+                else:
+                    parameters.append(type_param)
+            self.__parameters__ = tuple(parameters)
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+            # Setting this attribute closes the TypeAliasType from further modification
+            self.__name__ = name
+
+        def __setattr__(self, name: str, value: object, /) -> None:
+            if hasattr(self, "__name__"):
+                self._raise_attribute_error(name)
+            super().__setattr__(name, value)
+
+        def __delattr__(self, name: str, /) -> Never:
+            self._raise_attribute_error(name)
+
+        def _raise_attribute_error(self, name: str) -> Never:
+            # Match the Python 3.12 error messages exactly
+            if name == "__name__":
+                raise AttributeError("readonly attribute")
+            elif name in {"__value__", "__type_params__", "__parameters__", "__module__"}:
+                raise AttributeError(
+                    f"attribute '{name}' of 'typing.TypeAliasType' objects "
+                    "is not writable"
+                )
+            else:
+                raise AttributeError(
+                    f"'typing.TypeAliasType' object has no attribute '{name}'"
+                )
+
+        def __repr__(self) -> str:
+            return self.__name__
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+            parameters = [
+                typing._type_check(
+                    item, f'Subscripting {self.__name__} requires a type.'
+                )
+                for item in parameters
+            ]
+            return typing._GenericAlias(self, tuple(parameters))
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                "type 'typing_extensions.TypeAliasType' is not an acceptable base type"
+            )
+
+        # The presence of this method convinces typing._type_check
+        # that TypeAliasTypes are types.
+        def __call__(self):
+            raise TypeError("Type alias is not callable")
+
+        if sys.version_info >= (3, 10):
+            def __or__(self, right):
+                # For forward compatibility with 3.12, reject Unions
+                # that are not accepted by the built-in Union.
+                if not _is_unionable(right):
+                    return NotImplemented
+                return typing.Union[self, right]
+
+            def __ror__(self, left):
+                if not _is_unionable(left):
+                    return NotImplemented
+                return typing.Union[left, self]
+
+
+if hasattr(typing, "is_protocol"):
+    is_protocol = typing.is_protocol
+    get_protocol_members = typing.get_protocol_members
+else:
+    def is_protocol(tp: type, /) -> bool:
+        """Return True if the given type is a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, is_protocol
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> is_protocol(P)
+            True
+            >>> is_protocol(int)
+            False
+        """
+        return (
+            isinstance(tp, type)
+            and getattr(tp, '_is_protocol', False)
+            and tp is not Protocol
+            and tp is not typing.Protocol
+        )
+
+    def get_protocol_members(tp: type, /) -> typing.FrozenSet[str]:
+        """Return the set of members defined in a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, get_protocol_members
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> get_protocol_members(P)
+            frozenset({'a', 'b'})
+
+        Raise a TypeError for arguments that are not Protocols.
+        """
+        if not is_protocol(tp):
+            raise TypeError(f'{tp!r} is not a Protocol')
+        if hasattr(tp, '__protocol_attrs__'):
+            return frozenset(tp.__protocol_attrs__)
+        return frozenset(_get_protocol_attrs(tp))
+
+
+if hasattr(typing, "Doc"):
+    Doc = typing.Doc
+else:
+    class Doc:
+        """Define the documentation of a type annotation using ``Annotated``, to be
+         used in class attributes, function and method parameters, return values,
+         and variables.
+
+        The value should be a positional-only string literal to allow static tools
+        like editors and documentation generators to use it.
+
+        This complements docstrings.
+
+        The string value passed is available in the attribute ``documentation``.
+
+        Example::
+
+            >>> from typing_extensions import Annotated, Doc
+            >>> def hi(to: Annotated[str, Doc("Who to say hi to")]) -> None: ...
+        """
+        def __init__(self, documentation: str, /) -> None:
+            self.documentation = documentation
+
+        def __repr__(self) -> str:
+            return f"Doc({self.documentation!r})"
+
+        def __hash__(self) -> int:
+            return hash(self.documentation)
+
+        def __eq__(self, other: object) -> bool:
+            if not isinstance(other, Doc):
+                return NotImplemented
+            return self.documentation == other.documentation
+
+
+_CapsuleType = getattr(_types, "CapsuleType", None)
+
+if _CapsuleType is None:
+    try:
+        import _socket
+    except ImportError:
+        pass
+    else:
+        _CAPI = getattr(_socket, "CAPI", None)
+        if _CAPI is not None:
+            _CapsuleType = type(_CAPI)
+
+if _CapsuleType is not None:
+    CapsuleType = _CapsuleType
+    __all__.append("CapsuleType")
+
+
+# Aliases for items that have always been in typing.
+# Explicitly assign these (rather than using `from typing import *` at the top),
+# so that we get a CI error if one of these is deleted from typing.py
+# in a future version of Python
+AbstractSet = typing.AbstractSet
+AnyStr = typing.AnyStr
+BinaryIO = typing.BinaryIO
+Callable = typing.Callable
+Collection = typing.Collection
+Container = typing.Container
+Dict = typing.Dict
+ForwardRef = typing.ForwardRef
+FrozenSet = typing.FrozenSet
+Generic = typing.Generic
+Hashable = typing.Hashable
+IO = typing.IO
+ItemsView = typing.ItemsView
+Iterable = typing.Iterable
+Iterator = typing.Iterator
+KeysView = typing.KeysView
+List = typing.List
+Mapping = typing.Mapping
+MappingView = typing.MappingView
+Match = typing.Match
+MutableMapping = typing.MutableMapping
+MutableSequence = typing.MutableSequence
+MutableSet = typing.MutableSet
+Optional = typing.Optional
+Pattern = typing.Pattern
+Reversible = typing.Reversible
+Sequence = typing.Sequence
+Set = typing.Set
+Sized = typing.Sized
+TextIO = typing.TextIO
+Tuple = typing.Tuple
+Union = typing.Union
+ValuesView = typing.ValuesView
+cast = typing.cast
+no_type_check = typing.no_type_check
+no_type_check_decorator = typing.no_type_check_decorator
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/LICENSE.txt
new file mode 100644
index 00000000..a31470f1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/LICENSE.txt
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2012 Daniel Holth <dholth@fastmail.fm> and contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and associated documentation files (the "Software"),
+to deal in the Software without restriction, including without limitation
+the rights to use, copy, modify, merge, publish, distribute, sublicense,
+and/or sell copies of the Software, and to permit persons to whom the
+Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included
+in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+OTHER DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/METADATA
new file mode 100644
index 00000000..e3722c00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/METADATA
@@ -0,0 +1,61 @@
+Metadata-Version: 2.1
+Name: wheel
+Version: 0.43.0
+Summary: A built-package format for Python
+Keywords: wheel,packaging
+Author-email: Daniel Holth <dholth@fastmail.fm>
+Maintainer-email: Alex Grönholm <alex.gronholm@nextday.fi>
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Topic :: System :: Archiving :: Packaging
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Dist: pytest >= 6.0.0 ; extra == "test"
+Requires-Dist: setuptools >= 65 ; extra == "test"
+Project-URL: Changelog, https://wheel.readthedocs.io/en/stable/news.html
+Project-URL: Documentation, https://wheel.readthedocs.io/
+Project-URL: Issue Tracker, https://github.com/pypa/wheel/issues
+Project-URL: Source, https://github.com/pypa/wheel
+Provides-Extra: test
+
+wheel
+=====
+
+This library is the reference implementation of the Python wheel packaging
+standard, as defined in `PEP 427`_.
+
+It has two different roles:
+
+#. A setuptools_ extension for building wheels that provides the
+   ``bdist_wheel`` setuptools command
+#. A command line tool for working with wheel files
+
+It should be noted that wheel is **not** intended to be used as a library, and
+as such there is no stable, public API.
+
+.. _PEP 427: https://www.python.org/dev/peps/pep-0427/
+.. _setuptools: https://pypi.org/project/setuptools/
+
+Documentation
+-------------
+
+The documentation_ can be found on Read The Docs.
+
+.. _documentation: https://wheel.readthedocs.io/
+
+Code of Conduct
+---------------
+
+Everyone interacting in the wheel project's codebases, issue trackers, chat
+rooms, and mailing lists is expected to follow the `PSF Code of Conduct`_.
+
+.. _PSF Code of Conduct: https://github.com/pypa/.github/blob/main/CODE_OF_CONDUCT.md
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/RECORD
new file mode 100644
index 00000000..a3c6c3ea
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/RECORD
@@ -0,0 +1,63 @@
+../../bin/wheel,sha256=cT2EHbrv-J-UyUXu26cDY-0I7RgcruysJeHFanT1Xfo,249
+wheel-0.43.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+wheel-0.43.0.dist-info/LICENSE.txt,sha256=MMI2GGeRCPPo6h0qZYx8pBe9_IkcmO8aifpP8MmChlQ,1107
+wheel-0.43.0.dist-info/METADATA,sha256=WbrCKwClnT5WCKVrjPjvxDgxo2tyeS7kOJyc1GaceEE,2153
+wheel-0.43.0.dist-info/RECORD,,
+wheel-0.43.0.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+wheel-0.43.0.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+wheel-0.43.0.dist-info/entry_points.txt,sha256=rTY1BbkPHhkGMm4Q3F0pIzJBzW2kMxoG1oriffvGdA0,104
+wheel/__init__.py,sha256=D6jhH00eMzbgrXGAeOwVfD5i-lCAMMycuG1L0useDlo,59
+wheel/__main__.py,sha256=NkMUnuTCGcOkgY0IBLgBCVC_BGGcWORx2K8jYGS12UE,455
+wheel/__pycache__/__init__.cpython-312.pyc,,
+wheel/__pycache__/__main__.cpython-312.pyc,,
+wheel/__pycache__/_setuptools_logging.cpython-312.pyc,,
+wheel/__pycache__/bdist_wheel.cpython-312.pyc,,
+wheel/__pycache__/macosx_libfile.cpython-312.pyc,,
+wheel/__pycache__/metadata.cpython-312.pyc,,
+wheel/__pycache__/util.cpython-312.pyc,,
+wheel/__pycache__/wheelfile.cpython-312.pyc,,
+wheel/_setuptools_logging.py,sha256=NoCnjJ4DFEZ45Eo-2BdXLsWJCwGkait1tp_17paleVw,746
+wheel/bdist_wheel.py,sha256=OKJyp9E831zJrxoRfmM9AgOjByG1CB-pzF5kXQFmaKk,20938
+wheel/cli/__init__.py,sha256=eBNhnPwWTtdKAJHy77lvz7gOQ5Eu3GavGugXxhSsn-U,4264
+wheel/cli/__pycache__/__init__.cpython-312.pyc,,
+wheel/cli/__pycache__/convert.cpython-312.pyc,,
+wheel/cli/__pycache__/pack.cpython-312.pyc,,
+wheel/cli/__pycache__/tags.cpython-312.pyc,,
+wheel/cli/__pycache__/unpack.cpython-312.pyc,,
+wheel/cli/convert.py,sha256=qJcpYGKqdfw1P6BelgN1Hn_suNgM6bvyEWFlZeuSWx0,9439
+wheel/cli/pack.py,sha256=CAFcHdBVulvsHYJlndKVO7KMI9JqBTZz5ii0PKxxCOs,3103
+wheel/cli/tags.py,sha256=lHw-LaWrkS5Jy_qWcw-6pSjeNM6yAjDnqKI3E5JTTCU,4760
+wheel/cli/unpack.py,sha256=Y_J7ynxPSoFFTT7H0fMgbBlVErwyDGcObgme5MBuz58,1021
+wheel/macosx_libfile.py,sha256=HnW6OPdN993psStvwl49xtx2kw7hoVbe6nvwmf8WsKI,16103
+wheel/metadata.py,sha256=q-xCCqSAK7HzyZxK9A6_HAWmhqS1oB4BFw1-rHQxBiQ,5884
+wheel/util.py,sha256=e0jpnsbbM9QhaaMSyap-_ZgUxcxwpyLDk6RHcrduPLg,621
+wheel/vendored/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+wheel/vendored/__pycache__/__init__.cpython-312.pyc,,
+wheel/vendored/packaging/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+wheel/vendored/packaging/__pycache__/__init__.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_elffile.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_manylinux.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_musllinux.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_parser.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_structures.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/_tokenizer.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/markers.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/requirements.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/specifiers.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/tags.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/utils.cpython-312.pyc,,
+wheel/vendored/packaging/__pycache__/version.cpython-312.pyc,,
+wheel/vendored/packaging/_elffile.py,sha256=hbmK8OD6Z7fY6hwinHEUcD1by7czkGiNYu7ShnFEk2k,3266
+wheel/vendored/packaging/_manylinux.py,sha256=P7sdR5_7XBY09LVYYPhHmydMJIIwPXWsh4olk74Uuj4,9588
+wheel/vendored/packaging/_musllinux.py,sha256=z1s8To2hQ0vpn_d-O2i5qxGwEK8WmGlLt3d_26V7NeY,2674
+wheel/vendored/packaging/_parser.py,sha256=4tT4emSl2qTaU7VTQE1Xa9o1jMPCsBezsYBxyNMUN-s,10347
+wheel/vendored/packaging/_structures.py,sha256=q3eVNmbWJGG_S0Dit_S3Ao8qQqz_5PYTXFAKBZe5yr4,1431
+wheel/vendored/packaging/_tokenizer.py,sha256=alCtbwXhOFAmFGZ6BQ-wCTSFoRAJ2z-ysIf7__MTJ_k,5292
+wheel/vendored/packaging/markers.py,sha256=_TSPI1BhJYO7Bp9AzTmHQxIqHEVXaTjmDh9G-w8qzPA,8232
+wheel/vendored/packaging/requirements.py,sha256=dgoBeVprPu2YE6Q8nGfwOPTjATHbRa_ZGLyXhFEln6Q,2933
+wheel/vendored/packaging/specifiers.py,sha256=IWSt0SrLSP72heWhAC8UL0eGvas7XIQHjqiViVfmPKE,39778
+wheel/vendored/packaging/tags.py,sha256=fedHXiOHkBxNZTXotXv8uXPmMFU9ae-TKBujgYHigcA,18950
+wheel/vendored/packaging/utils.py,sha256=XgdmP3yx9-wQEFjO7OvMj9RjEf5JlR5HFFR69v7SQ9E,5268
+wheel/vendored/packaging/version.py,sha256=PFJaYZDxBgyxkfYhH3SQw4qfE9ICCWrTmitvq14y3bs,16234
+wheel/vendored/vendor.txt,sha256=Z2ENjB1i5prfez8CdM1Sdr3c6Zxv2rRRolMpLmBncAE,16
+wheel/wheelfile.py,sha256=DtJDWoZMvnBh4leNMDPGOprQU9d_dp6q-MmV0U--4xc,7694
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/entry_points.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/entry_points.txt
new file mode 100644
index 00000000..06c9f69d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel-0.43.0.dist-info/entry_points.txt
@@ -0,0 +1,6 @@
+[console_scripts]
+wheel=wheel.cli:main
+
+[distutils.commands]
+bdist_wheel=wheel.bdist_wheel:bdist_wheel
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__init__.py
new file mode 100644
index 00000000..a773bbbc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__init__.py
@@ -0,0 +1,3 @@
+from __future__ import annotations
+
+__version__ = "0.43.0"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__main__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__main__.py
new file mode 100644
index 00000000..0be74537
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__main__.py
@@ -0,0 +1,23 @@
+"""
+Wheel command line tool (enable python -m wheel syntax)
+"""
+
+from __future__ import annotations
+
+import sys
+
+
+def main():  # needed for console script
+    if __package__ == "":
+        # To be able to run 'python wheel-0.9.whl/wheel':
+        import os.path
+
+        path = os.path.dirname(os.path.dirname(__file__))
+        sys.path[0:0] = [path]
+    import wheel.cli
+
+    sys.exit(wheel.cli.main())
+
+
+if __name__ == "__main__":
+    sys.exit(main())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d4c3d5b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__main__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__main__.cpython-312.pyc
new file mode 100644
index 00000000..e6de9eab
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/__main__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/_setuptools_logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/_setuptools_logging.cpython-312.pyc
new file mode 100644
index 00000000..e8074d1b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/_setuptools_logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/bdist_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/bdist_wheel.cpython-312.pyc
new file mode 100644
index 00000000..77a8b41c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/bdist_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/macosx_libfile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/macosx_libfile.cpython-312.pyc
new file mode 100644
index 00000000..5a641a7f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/macosx_libfile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/metadata.cpython-312.pyc
new file mode 100644
index 00000000..2d951cb7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..52384f93
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/wheelfile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/wheelfile.cpython-312.pyc
new file mode 100644
index 00000000..879324f9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/__pycache__/wheelfile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/_setuptools_logging.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/_setuptools_logging.py
new file mode 100644
index 00000000..006c0985
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/_setuptools_logging.py
@@ -0,0 +1,26 @@
+# copied from setuptools.logging, omitting monkeypatching
+from __future__ import annotations
+
+import logging
+import sys
+
+
+def _not_warning(record):
+    return record.levelno < logging.WARNING
+
+
+def configure():
+    """
+    Configure logging to emit warning and above to stderr
+    and everything else to stdout. This behavior is provided
+    for compatibility with distutils.log but may change in
+    the future.
+    """
+    err_handler = logging.StreamHandler()
+    err_handler.setLevel(logging.WARNING)
+    out_handler = logging.StreamHandler(sys.stdout)
+    out_handler.addFilter(_not_warning)
+    handlers = err_handler, out_handler
+    logging.basicConfig(
+        format="{message}", style="{", handlers=handlers, level=logging.DEBUG
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/bdist_wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/bdist_wheel.py
new file mode 100644
index 00000000..6b811ee3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/bdist_wheel.py
@@ -0,0 +1,595 @@
+"""
+Create a wheel (.whl) distribution.
+
+A wheel is a built archive format.
+"""
+
+from __future__ import annotations
+
+import os
+import re
+import shutil
+import stat
+import struct
+import sys
+import sysconfig
+import warnings
+from email.generator import BytesGenerator, Generator
+from email.policy import EmailPolicy
+from glob import iglob
+from shutil import rmtree
+from zipfile import ZIP_DEFLATED, ZIP_STORED
+
+import setuptools
+from setuptools import Command
+
+from . import __version__ as wheel_version
+from .macosx_libfile import calculate_macosx_platform_tag
+from .metadata import pkginfo_to_metadata
+from .util import log
+from .vendored.packaging import tags
+from .vendored.packaging import version as _packaging_version
+from .wheelfile import WheelFile
+
+
+def safe_name(name):
+    """Convert an arbitrary string to a standard distribution name
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub("[^A-Za-z0-9.]+", "-", name)
+
+
+def safe_version(version):
+    """
+    Convert an arbitrary string to a standard version string
+    """
+    try:
+        # normalize the version
+        return str(_packaging_version.Version(version))
+    except _packaging_version.InvalidVersion:
+        version = version.replace(" ", ".")
+        return re.sub("[^A-Za-z0-9.]+", "-", version)
+
+
+setuptools_major_version = int(setuptools.__version__.split(".")[0])
+
+PY_LIMITED_API_PATTERN = r"cp3\d"
+
+
+def _is_32bit_interpreter():
+    return struct.calcsize("P") == 4
+
+
+def python_tag():
+    return f"py{sys.version_info[0]}"
+
+
+def get_platform(archive_root):
+    """Return our platform name 'win32', 'linux_x86_64'"""
+    result = sysconfig.get_platform()
+    if result.startswith("macosx") and archive_root is not None:
+        result = calculate_macosx_platform_tag(archive_root, result)
+    elif _is_32bit_interpreter():
+        if result == "linux-x86_64":
+            # pip pull request #3497
+            result = "linux-i686"
+        elif result == "linux-aarch64":
+            # packaging pull request #234
+            # TODO armv8l, packaging pull request #690 => this did not land
+            # in pip/packaging yet
+            result = "linux-armv7l"
+
+    return result.replace("-", "_")
+
+
+def get_flag(var, fallback, expected=True, warn=True):
+    """Use a fallback value for determining SOABI flags if the needed config
+    var is unset or unavailable."""
+    val = sysconfig.get_config_var(var)
+    if val is None:
+        if warn:
+            warnings.warn(
+                f"Config variable '{var}' is unset, Python ABI tag may " "be incorrect",
+                RuntimeWarning,
+                stacklevel=2,
+            )
+        return fallback
+    return val == expected
+
+
+def get_abi_tag():
+    """Return the ABI tag based on SOABI (if available) or emulate SOABI (PyPy2)."""
+    soabi = sysconfig.get_config_var("SOABI")
+    impl = tags.interpreter_name()
+    if not soabi and impl in ("cp", "pp") and hasattr(sys, "maxunicode"):
+        d = ""
+        m = ""
+        u = ""
+        if get_flag("Py_DEBUG", hasattr(sys, "gettotalrefcount"), warn=(impl == "cp")):
+            d = "d"
+
+        if get_flag(
+            "WITH_PYMALLOC",
+            impl == "cp",
+            warn=(impl == "cp" and sys.version_info < (3, 8)),
+        ) and sys.version_info < (3, 8):
+            m = "m"
+
+        abi = f"{impl}{tags.interpreter_version()}{d}{m}{u}"
+    elif soabi and impl == "cp" and soabi.startswith("cpython"):
+        # non-Windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi and impl == "cp" and soabi.startswith("cp"):
+        # Windows
+        abi = soabi.split("-")[0]
+    elif soabi and impl == "pp":
+        # we want something like pypy36-pp73
+        abi = "-".join(soabi.split("-")[:2])
+        abi = abi.replace(".", "_").replace("-", "_")
+    elif soabi and impl == "graalpy":
+        abi = "-".join(soabi.split("-")[:3])
+        abi = abi.replace(".", "_").replace("-", "_")
+    elif soabi:
+        abi = soabi.replace(".", "_").replace("-", "_")
+    else:
+        abi = None
+
+    return abi
+
+
+def safer_name(name):
+    return safe_name(name).replace("-", "_")
+
+
+def safer_version(version):
+    return safe_version(version).replace("-", "_")
+
+
+def remove_readonly(func, path, excinfo):
+    remove_readonly_exc(func, path, excinfo[1])
+
+
+def remove_readonly_exc(func, path, exc):
+    os.chmod(path, stat.S_IWRITE)
+    func(path)
+
+
+class bdist_wheel(Command):
+    description = "create a wheel distribution"
+
+    supported_compressions = {
+        "stored": ZIP_STORED,
+        "deflated": ZIP_DEFLATED,
+    }
+
+    user_options = [
+        ("bdist-dir=", "b", "temporary directory for creating the distribution"),
+        (
+            "plat-name=",
+            "p",
+            "platform name to embed in generated filenames "
+            "(default: %s)" % get_platform(None),
+        ),
+        (
+            "keep-temp",
+            "k",
+            "keep the pseudo-installation tree around after "
+            "creating the distribution archive",
+        ),
+        ("dist-dir=", "d", "directory to put final built distributions in"),
+        ("skip-build", None, "skip rebuilding everything (for testing/debugging)"),
+        (
+            "relative",
+            None,
+            "build the archive using relative paths " "(default: false)",
+        ),
+        (
+            "owner=",
+            "u",
+            "Owner name used when creating a tar file" " [default: current user]",
+        ),
+        (
+            "group=",
+            "g",
+            "Group name used when creating a tar file" " [default: current group]",
+        ),
+        ("universal", None, "make a universal wheel" " (default: false)"),
+        (
+            "compression=",
+            None,
+            "zipfile compression (one of: {})" " (default: 'deflated')".format(
+                ", ".join(supported_compressions)
+            ),
+        ),
+        (
+            "python-tag=",
+            None,
+            "Python implementation compatibility tag"
+            " (default: '%s')" % (python_tag()),
+        ),
+        (
+            "build-number=",
+            None,
+            "Build number for this particular version. "
+            "As specified in PEP-0427, this must start with a digit. "
+            "[default: None]",
+        ),
+        (
+            "py-limited-api=",
+            None,
+            "Python tag (cp32|cp33|cpNN) for abi3 wheel tag" " (default: false)",
+        ),
+    ]
+
+    boolean_options = ["keep-temp", "skip-build", "relative", "universal"]
+
+    def initialize_options(self):
+        self.bdist_dir = None
+        self.data_dir = None
+        self.plat_name = None
+        self.plat_tag = None
+        self.format = "zip"
+        self.keep_temp = False
+        self.dist_dir = None
+        self.egginfo_dir = None
+        self.root_is_pure = None
+        self.skip_build = None
+        self.relative = False
+        self.owner = None
+        self.group = None
+        self.universal = False
+        self.compression = "deflated"
+        self.python_tag = python_tag()
+        self.build_number = None
+        self.py_limited_api = False
+        self.plat_name_supplied = False
+
+    def finalize_options(self):
+        if self.bdist_dir is None:
+            bdist_base = self.get_finalized_command("bdist").bdist_base
+            self.bdist_dir = os.path.join(bdist_base, "wheel")
+
+        egg_info = self.distribution.get_command_obj("egg_info")
+        egg_info.ensure_finalized()  # needed for correct `wheel_dist_name`
+
+        self.data_dir = self.wheel_dist_name + ".data"
+        self.plat_name_supplied = self.plat_name is not None
+
+        try:
+            self.compression = self.supported_compressions[self.compression]
+        except KeyError:
+            raise ValueError(f"Unsupported compression: {self.compression}") from None
+
+        need_options = ("dist_dir", "plat_name", "skip_build")
+
+        self.set_undefined_options("bdist", *zip(need_options, need_options))
+
+        self.root_is_pure = not (
+            self.distribution.has_ext_modules() or self.distribution.has_c_libraries()
+        )
+
+        if self.py_limited_api and not re.match(
+            PY_LIMITED_API_PATTERN, self.py_limited_api
+        ):
+            raise ValueError("py-limited-api must match '%s'" % PY_LIMITED_API_PATTERN)
+
+        # Support legacy [wheel] section for setting universal
+        wheel = self.distribution.get_option_dict("wheel")
+        if "universal" in wheel:
+            # please don't define this in your global configs
+            log.warning(
+                "The [wheel] section is deprecated. Use [bdist_wheel] instead.",
+            )
+            val = wheel["universal"][1].strip()
+            if val.lower() in ("1", "true", "yes"):
+                self.universal = True
+
+        if self.build_number is not None and not self.build_number[:1].isdigit():
+            raise ValueError("Build tag (build-number) must start with a digit.")
+
+    @property
+    def wheel_dist_name(self):
+        """Return distribution full name with - replaced with _"""
+        components = (
+            safer_name(self.distribution.get_name()),
+            safer_version(self.distribution.get_version()),
+        )
+        if self.build_number:
+            components += (self.build_number,)
+        return "-".join(components)
+
+    def get_tag(self):
+        # bdist sets self.plat_name if unset, we should only use it for purepy
+        # wheels if the user supplied it.
+        if self.plat_name_supplied:
+            plat_name = self.plat_name
+        elif self.root_is_pure:
+            plat_name = "any"
+        else:
+            # macosx contains system version in platform name so need special handle
+            if self.plat_name and not self.plat_name.startswith("macosx"):
+                plat_name = self.plat_name
+            else:
+                # on macosx always limit the platform name to comply with any
+                # c-extension modules in bdist_dir, since the user can specify
+                # a higher MACOSX_DEPLOYMENT_TARGET via tools like CMake
+
+                # on other platforms, and on macosx if there are no c-extension
+                # modules, use the default platform name.
+                plat_name = get_platform(self.bdist_dir)
+
+            if _is_32bit_interpreter():
+                if plat_name in ("linux-x86_64", "linux_x86_64"):
+                    plat_name = "linux_i686"
+                if plat_name in ("linux-aarch64", "linux_aarch64"):
+                    # TODO armv8l, packaging pull request #690 => this did not land
+                    # in pip/packaging yet
+                    plat_name = "linux_armv7l"
+
+        plat_name = (
+            plat_name.lower().replace("-", "_").replace(".", "_").replace(" ", "_")
+        )
+
+        if self.root_is_pure:
+            if self.universal:
+                impl = "py2.py3"
+            else:
+                impl = self.python_tag
+            tag = (impl, "none", plat_name)
+        else:
+            impl_name = tags.interpreter_name()
+            impl_ver = tags.interpreter_version()
+            impl = impl_name + impl_ver
+            # We don't work on CPython 3.1, 3.0.
+            if self.py_limited_api and (impl_name + impl_ver).startswith("cp3"):
+                impl = self.py_limited_api
+                abi_tag = "abi3"
+            else:
+                abi_tag = str(get_abi_tag()).lower()
+            tag = (impl, abi_tag, plat_name)
+            # issue gh-374: allow overriding plat_name
+            supported_tags = [
+                (t.interpreter, t.abi, plat_name) for t in tags.sys_tags()
+            ]
+            assert (
+                tag in supported_tags
+            ), f"would build wheel with unsupported tag {tag}"
+        return tag
+
+    def run(self):
+        build_scripts = self.reinitialize_command("build_scripts")
+        build_scripts.executable = "python"
+        build_scripts.force = True
+
+        build_ext = self.reinitialize_command("build_ext")
+        build_ext.inplace = False
+
+        if not self.skip_build:
+            self.run_command("build")
+
+        install = self.reinitialize_command("install", reinit_subcommands=True)
+        install.root = self.bdist_dir
+        install.compile = False
+        install.skip_build = self.skip_build
+        install.warn_dir = False
+
+        # A wheel without setuptools scripts is more cross-platform.
+        # Use the (undocumented) `no_ep` option to setuptools'
+        # install_scripts command to avoid creating entry point scripts.
+        install_scripts = self.reinitialize_command("install_scripts")
+        install_scripts.no_ep = True
+
+        # Use a custom scheme for the archive, because we have to decide
+        # at installation time which scheme to use.
+        for key in ("headers", "scripts", "data", "purelib", "platlib"):
+            setattr(install, "install_" + key, os.path.join(self.data_dir, key))
+
+        basedir_observed = ""
+
+        if os.name == "nt":
+            # win32 barfs if any of these are ''; could be '.'?
+            # (distutils.command.install:change_roots bug)
+            basedir_observed = os.path.normpath(os.path.join(self.data_dir, ".."))
+            self.install_libbase = self.install_lib = basedir_observed
+
+        setattr(
+            install,
+            "install_purelib" if self.root_is_pure else "install_platlib",
+            basedir_observed,
+        )
+
+        log.info(f"installing to {self.bdist_dir}")
+
+        self.run_command("install")
+
+        impl_tag, abi_tag, plat_tag = self.get_tag()
+        archive_basename = f"{self.wheel_dist_name}-{impl_tag}-{abi_tag}-{plat_tag}"
+        if not self.relative:
+            archive_root = self.bdist_dir
+        else:
+            archive_root = os.path.join(
+                self.bdist_dir, self._ensure_relative(install.install_base)
+            )
+
+        self.set_undefined_options("install_egg_info", ("target", "egginfo_dir"))
+        distinfo_dirname = (
+            f"{safer_name(self.distribution.get_name())}-"
+            f"{safer_version(self.distribution.get_version())}.dist-info"
+        )
+        distinfo_dir = os.path.join(self.bdist_dir, distinfo_dirname)
+        self.egg2dist(self.egginfo_dir, distinfo_dir)
+
+        self.write_wheelfile(distinfo_dir)
+
+        # Make the archive
+        if not os.path.exists(self.dist_dir):
+            os.makedirs(self.dist_dir)
+
+        wheel_path = os.path.join(self.dist_dir, archive_basename + ".whl")
+        with WheelFile(wheel_path, "w", self.compression) as wf:
+            wf.write_files(archive_root)
+
+        # Add to 'Distribution.dist_files' so that the "upload" command works
+        getattr(self.distribution, "dist_files", []).append(
+            (
+                "bdist_wheel",
+                "{}.{}".format(*sys.version_info[:2]),  # like 3.7
+                wheel_path,
+            )
+        )
+
+        if not self.keep_temp:
+            log.info(f"removing {self.bdist_dir}")
+            if not self.dry_run:
+                if sys.version_info < (3, 12):
+                    rmtree(self.bdist_dir, onerror=remove_readonly)
+                else:
+                    rmtree(self.bdist_dir, onexc=remove_readonly_exc)
+
+    def write_wheelfile(
+        self, wheelfile_base, generator="bdist_wheel (" + wheel_version + ")"
+    ):
+        from email.message import Message
+
+        msg = Message()
+        msg["Wheel-Version"] = "1.0"  # of the spec
+        msg["Generator"] = generator
+        msg["Root-Is-Purelib"] = str(self.root_is_pure).lower()
+        if self.build_number is not None:
+            msg["Build"] = self.build_number
+
+        # Doesn't work for bdist_wininst
+        impl_tag, abi_tag, plat_tag = self.get_tag()
+        for impl in impl_tag.split("."):
+            for abi in abi_tag.split("."):
+                for plat in plat_tag.split("."):
+                    msg["Tag"] = "-".join((impl, abi, plat))
+
+        wheelfile_path = os.path.join(wheelfile_base, "WHEEL")
+        log.info(f"creating {wheelfile_path}")
+        with open(wheelfile_path, "wb") as f:
+            BytesGenerator(f, maxheaderlen=0).flatten(msg)
+
+    def _ensure_relative(self, path):
+        # copied from dir_util, deleted
+        drive, path = os.path.splitdrive(path)
+        if path[0:1] == os.sep:
+            path = drive + path[1:]
+        return path
+
+    @property
+    def license_paths(self):
+        if setuptools_major_version >= 57:
+            # Setuptools has resolved any patterns to actual file names
+            return self.distribution.metadata.license_files or ()
+
+        files = set()
+        metadata = self.distribution.get_option_dict("metadata")
+        if setuptools_major_version >= 42:
+            # Setuptools recognizes the license_files option but does not do globbing
+            patterns = self.distribution.metadata.license_files
+        else:
+            # Prior to those, wheel is entirely responsible for handling license files
+            if "license_files" in metadata:
+                patterns = metadata["license_files"][1].split()
+            else:
+                patterns = ()
+
+        if "license_file" in metadata:
+            warnings.warn(
+                'The "license_file" option is deprecated. Use "license_files" instead.',
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            files.add(metadata["license_file"][1])
+
+        if not files and not patterns and not isinstance(patterns, list):
+            patterns = ("LICEN[CS]E*", "COPYING*", "NOTICE*", "AUTHORS*")
+
+        for pattern in patterns:
+            for path in iglob(pattern):
+                if path.endswith("~"):
+                    log.debug(
+                        f'ignoring license file "{path}" as it looks like a backup'
+                    )
+                    continue
+
+                if path not in files and os.path.isfile(path):
+                    log.info(
+                        f'adding license file "{path}" (matched pattern "{pattern}")'
+                    )
+                    files.add(path)
+
+        return files
+
+    def egg2dist(self, egginfo_path, distinfo_path):
+        """Convert an .egg-info directory into a .dist-info directory"""
+
+        def adios(p):
+            """Appropriately delete directory, file or link."""
+            if os.path.exists(p) and not os.path.islink(p) and os.path.isdir(p):
+                shutil.rmtree(p)
+            elif os.path.exists(p):
+                os.unlink(p)
+
+        adios(distinfo_path)
+
+        if not os.path.exists(egginfo_path):
+            # There is no egg-info. This is probably because the egg-info
+            # file/directory is not named matching the distribution name used
+            # to name the archive file. Check for this case and report
+            # accordingly.
+            import glob
+
+            pat = os.path.join(os.path.dirname(egginfo_path), "*.egg-info")
+            possible = glob.glob(pat)
+            err = f"Egg metadata expected at {egginfo_path} but not found"
+            if possible:
+                alt = os.path.basename(possible[0])
+                err += f" ({alt} found - possible misnamed archive file?)"
+
+            raise ValueError(err)
+
+        if os.path.isfile(egginfo_path):
+            # .egg-info is a single file
+            pkginfo_path = egginfo_path
+            pkg_info = pkginfo_to_metadata(egginfo_path, egginfo_path)
+            os.mkdir(distinfo_path)
+        else:
+            # .egg-info is a directory
+            pkginfo_path = os.path.join(egginfo_path, "PKG-INFO")
+            pkg_info = pkginfo_to_metadata(egginfo_path, pkginfo_path)
+
+            # ignore common egg metadata that is useless to wheel
+            shutil.copytree(
+                egginfo_path,
+                distinfo_path,
+                ignore=lambda x, y: {
+                    "PKG-INFO",
+                    "requires.txt",
+                    "SOURCES.txt",
+                    "not-zip-safe",
+                },
+            )
+
+            # delete dependency_links if it is only whitespace
+            dependency_links_path = os.path.join(distinfo_path, "dependency_links.txt")
+            with open(dependency_links_path, encoding="utf-8") as dependency_links_file:
+                dependency_links = dependency_links_file.read().strip()
+            if not dependency_links:
+                adios(dependency_links_path)
+
+        pkg_info_path = os.path.join(distinfo_path, "METADATA")
+        serialization_policy = EmailPolicy(
+            utf8=True,
+            mangle_from_=False,
+            max_line_length=0,
+        )
+        with open(pkg_info_path, "w", encoding="utf-8") as out:
+            Generator(out, policy=serialization_policy).flatten(pkg_info)
+
+        for license_path in self.license_paths:
+            filename = os.path.basename(license_path)
+            shutil.copy(license_path, os.path.join(distinfo_path, filename))
+
+        adios(egginfo_path)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__init__.py
new file mode 100644
index 00000000..a38860f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__init__.py
@@ -0,0 +1,155 @@
+"""
+Wheel command-line utility.
+"""
+
+from __future__ import annotations
+
+import argparse
+import os
+import sys
+from argparse import ArgumentTypeError
+
+
+class WheelError(Exception):
+    pass
+
+
+def unpack_f(args):
+    from .unpack import unpack
+
+    unpack(args.wheelfile, args.dest)
+
+
+def pack_f(args):
+    from .pack import pack
+
+    pack(args.directory, args.dest_dir, args.build_number)
+
+
+def convert_f(args):
+    from .convert import convert
+
+    convert(args.files, args.dest_dir, args.verbose)
+
+
+def tags_f(args):
+    from .tags import tags
+
+    names = (
+        tags(
+            wheel,
+            args.python_tag,
+            args.abi_tag,
+            args.platform_tag,
+            args.build,
+            args.remove,
+        )
+        for wheel in args.wheel
+    )
+
+    for name in names:
+        print(name)
+
+
+def version_f(args):
+    from .. import __version__
+
+    print("wheel %s" % __version__)
+
+
+def parse_build_tag(build_tag: str) -> str:
+    if build_tag and not build_tag[0].isdigit():
+        raise ArgumentTypeError("build tag must begin with a digit")
+    elif "-" in build_tag:
+        raise ArgumentTypeError("invalid character ('-') in build tag")
+
+    return build_tag
+
+
+TAGS_HELP = """\
+Make a new wheel with given tags. Any tags unspecified will remain the same.
+Starting the tags with a "+" will append to the existing tags. Starting with a
+"-" will remove a tag (use --option=-TAG syntax). Multiple tags can be
+separated by ".". The original file will remain unless --remove is given.  The
+output filename(s) will be displayed on stdout for further processing.
+"""
+
+
+def parser():
+    p = argparse.ArgumentParser()
+    s = p.add_subparsers(help="commands")
+
+    unpack_parser = s.add_parser("unpack", help="Unpack wheel")
+    unpack_parser.add_argument(
+        "--dest", "-d", help="Destination directory", default="."
+    )
+    unpack_parser.add_argument("wheelfile", help="Wheel file")
+    unpack_parser.set_defaults(func=unpack_f)
+
+    repack_parser = s.add_parser("pack", help="Repack wheel")
+    repack_parser.add_argument("directory", help="Root directory of the unpacked wheel")
+    repack_parser.add_argument(
+        "--dest-dir",
+        "-d",
+        default=os.path.curdir,
+        help="Directory to store the wheel (default %(default)s)",
+    )
+    repack_parser.add_argument(
+        "--build-number", help="Build tag to use in the wheel name"
+    )
+    repack_parser.set_defaults(func=pack_f)
+
+    convert_parser = s.add_parser("convert", help="Convert egg or wininst to wheel")
+    convert_parser.add_argument("files", nargs="*", help="Files to convert")
+    convert_parser.add_argument(
+        "--dest-dir",
+        "-d",
+        default=os.path.curdir,
+        help="Directory to store wheels (default %(default)s)",
+    )
+    convert_parser.add_argument("--verbose", "-v", action="store_true")
+    convert_parser.set_defaults(func=convert_f)
+
+    tags_parser = s.add_parser(
+        "tags", help="Add or replace the tags on a wheel", description=TAGS_HELP
+    )
+    tags_parser.add_argument("wheel", nargs="*", help="Existing wheel(s) to retag")
+    tags_parser.add_argument(
+        "--remove",
+        action="store_true",
+        help="Remove the original files, keeping only the renamed ones",
+    )
+    tags_parser.add_argument(
+        "--python-tag", metavar="TAG", help="Specify an interpreter tag(s)"
+    )
+    tags_parser.add_argument("--abi-tag", metavar="TAG", help="Specify an ABI tag(s)")
+    tags_parser.add_argument(
+        "--platform-tag", metavar="TAG", help="Specify a platform tag(s)"
+    )
+    tags_parser.add_argument(
+        "--build", type=parse_build_tag, metavar="BUILD", help="Specify a build tag"
+    )
+    tags_parser.set_defaults(func=tags_f)
+
+    version_parser = s.add_parser("version", help="Print version and exit")
+    version_parser.set_defaults(func=version_f)
+
+    help_parser = s.add_parser("help", help="Show this help")
+    help_parser.set_defaults(func=lambda args: p.print_help())
+
+    return p
+
+
+def main():
+    p = parser()
+    args = p.parse_args()
+    if not hasattr(args, "func"):
+        p.print_help()
+    else:
+        try:
+            args.func(args)
+            return 0
+        except WheelError as e:
+            print(e, file=sys.stderr)
+
+    return 1
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..8037d8ca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/convert.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/convert.cpython-312.pyc
new file mode 100644
index 00000000..6bc6e1de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/convert.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/pack.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/pack.cpython-312.pyc
new file mode 100644
index 00000000..df109e2e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/pack.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/tags.cpython-312.pyc
new file mode 100644
index 00000000..a6389e50
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/unpack.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/unpack.cpython-312.pyc
new file mode 100644
index 00000000..a722d6ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/__pycache__/unpack.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/convert.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/convert.py
new file mode 100644
index 00000000..29153404
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/convert.py
@@ -0,0 +1,273 @@
+from __future__ import annotations
+
+import os.path
+import re
+import shutil
+import tempfile
+import zipfile
+from glob import iglob
+
+from ..bdist_wheel import bdist_wheel
+from ..wheelfile import WheelFile
+from . import WheelError
+
+try:
+    from setuptools import Distribution
+except ImportError:
+    from distutils.dist import Distribution
+
+egg_info_re = re.compile(
+    r"""
+    (?P<name>.+?)-(?P<ver>.+?)
+    (-(?P<pyver>py\d\.\d+)
+     (-(?P<arch>.+?))?
+    )?.egg$""",
+    re.VERBOSE,
+)
+
+
+class _bdist_wheel_tag(bdist_wheel):
+    # allow the client to override the default generated wheel tag
+    # The default bdist_wheel implementation uses python and abi tags
+    # of the running python process. This is not suitable for
+    # generating/repackaging prebuild binaries.
+
+    full_tag_supplied = False
+    full_tag = None  # None or a (pytag, soabitag, plattag) triple
+
+    def get_tag(self):
+        if self.full_tag_supplied and self.full_tag is not None:
+            return self.full_tag
+        else:
+            return bdist_wheel.get_tag(self)
+
+
+def egg2wheel(egg_path: str, dest_dir: str) -> None:
+    filename = os.path.basename(egg_path)
+    match = egg_info_re.match(filename)
+    if not match:
+        raise WheelError(f"Invalid egg file name: {filename}")
+
+    egg_info = match.groupdict()
+    dir = tempfile.mkdtemp(suffix="_e2w")
+    if os.path.isfile(egg_path):
+        # assume we have a bdist_egg otherwise
+        with zipfile.ZipFile(egg_path) as egg:
+            egg.extractall(dir)
+    else:
+        # support buildout-style installed eggs directories
+        for pth in os.listdir(egg_path):
+            src = os.path.join(egg_path, pth)
+            if os.path.isfile(src):
+                shutil.copy2(src, dir)
+            else:
+                shutil.copytree(src, os.path.join(dir, pth))
+
+    pyver = egg_info["pyver"]
+    if pyver:
+        pyver = egg_info["pyver"] = pyver.replace(".", "")
+
+    arch = (egg_info["arch"] or "any").replace(".", "_").replace("-", "_")
+
+    # assume all binary eggs are for CPython
+    abi = "cp" + pyver[2:] if arch != "any" else "none"
+
+    root_is_purelib = egg_info["arch"] is None
+    if root_is_purelib:
+        bw = bdist_wheel(Distribution())
+    else:
+        bw = _bdist_wheel_tag(Distribution())
+
+    bw.root_is_pure = root_is_purelib
+    bw.python_tag = pyver
+    bw.plat_name_supplied = True
+    bw.plat_name = egg_info["arch"] or "any"
+    if not root_is_purelib:
+        bw.full_tag_supplied = True
+        bw.full_tag = (pyver, abi, arch)
+
+    dist_info_dir = os.path.join(dir, "{name}-{ver}.dist-info".format(**egg_info))
+    bw.egg2dist(os.path.join(dir, "EGG-INFO"), dist_info_dir)
+    bw.write_wheelfile(dist_info_dir, generator="egg2wheel")
+    wheel_name = "{name}-{ver}-{pyver}-{}-{}.whl".format(abi, arch, **egg_info)
+    with WheelFile(os.path.join(dest_dir, wheel_name), "w") as wf:
+        wf.write_files(dir)
+
+    shutil.rmtree(dir)
+
+
+def parse_wininst_info(wininfo_name, egginfo_name):
+    """Extract metadata from filenames.
+
+    Extracts the 4 metadataitems needed (name, version, pyversion, arch) from
+    the installer filename and the name of the egg-info directory embedded in
+    the zipfile (if any).
+
+    The egginfo filename has the format::
+
+        name-ver(-pyver)(-arch).egg-info
+
+    The installer filename has the format::
+
+        name-ver.arch(-pyver).exe
+
+    Some things to note:
+
+    1. The installer filename is not definitive. An installer can be renamed
+       and work perfectly well as an installer. So more reliable data should
+       be used whenever possible.
+    2. The egg-info data should be preferred for the name and version, because
+       these come straight from the distutils metadata, and are mandatory.
+    3. The pyver from the egg-info data should be ignored, as it is
+       constructed from the version of Python used to build the installer,
+       which is irrelevant - the installer filename is correct here (even to
+       the point that when it's not there, any version is implied).
+    4. The architecture must be taken from the installer filename, as it is
+       not included in the egg-info data.
+    5. Architecture-neutral installers still have an architecture because the
+       installer format itself (being executable) is architecture-specific. We
+       should therefore ignore the architecture if the content is pure-python.
+    """
+
+    egginfo = None
+    if egginfo_name:
+        egginfo = egg_info_re.search(egginfo_name)
+        if not egginfo:
+            raise ValueError(f"Egg info filename {egginfo_name} is not valid")
+
+    # Parse the wininst filename
+    # 1. Distribution name (up to the first '-')
+    w_name, sep, rest = wininfo_name.partition("-")
+    if not sep:
+        raise ValueError(f"Installer filename {wininfo_name} is not valid")
+
+    # Strip '.exe'
+    rest = rest[:-4]
+    # 2. Python version (from the last '-', must start with 'py')
+    rest2, sep, w_pyver = rest.rpartition("-")
+    if sep and w_pyver.startswith("py"):
+        rest = rest2
+        w_pyver = w_pyver.replace(".", "")
+    else:
+        # Not version specific - use py2.py3. While it is possible that
+        # pure-Python code is not compatible with both Python 2 and 3, there
+        # is no way of knowing from the wininst format, so we assume the best
+        # here (the user can always manually rename the wheel to be more
+        # restrictive if needed).
+        w_pyver = "py2.py3"
+    # 3. Version and architecture
+    w_ver, sep, w_arch = rest.rpartition(".")
+    if not sep:
+        raise ValueError(f"Installer filename {wininfo_name} is not valid")
+
+    if egginfo:
+        w_name = egginfo.group("name")
+        w_ver = egginfo.group("ver")
+
+    return {"name": w_name, "ver": w_ver, "arch": w_arch, "pyver": w_pyver}
+
+
+def wininst2wheel(path, dest_dir):
+    with zipfile.ZipFile(path) as bdw:
+        # Search for egg-info in the archive
+        egginfo_name = None
+        for filename in bdw.namelist():
+            if ".egg-info" in filename:
+                egginfo_name = filename
+                break
+
+        info = parse_wininst_info(os.path.basename(path), egginfo_name)
+
+        root_is_purelib = True
+        for zipinfo in bdw.infolist():
+            if zipinfo.filename.startswith("PLATLIB"):
+                root_is_purelib = False
+                break
+        if root_is_purelib:
+            paths = {"purelib": ""}
+        else:
+            paths = {"platlib": ""}
+
+        dist_info = "{name}-{ver}".format(**info)
+        datadir = "%s.data/" % dist_info
+
+        # rewrite paths to trick ZipFile into extracting an egg
+        # XXX grab wininst .ini - between .exe, padding, and first zip file.
+        members = []
+        egginfo_name = ""
+        for zipinfo in bdw.infolist():
+            key, basename = zipinfo.filename.split("/", 1)
+            key = key.lower()
+            basepath = paths.get(key, None)
+            if basepath is None:
+                basepath = datadir + key.lower() + "/"
+            oldname = zipinfo.filename
+            newname = basepath + basename
+            zipinfo.filename = newname
+            del bdw.NameToInfo[oldname]
+            bdw.NameToInfo[newname] = zipinfo
+            # Collect member names, but omit '' (from an entry like "PLATLIB/"
+            if newname:
+                members.append(newname)
+            # Remember egg-info name for the egg2dist call below
+            if not egginfo_name:
+                if newname.endswith(".egg-info"):
+                    egginfo_name = newname
+                elif ".egg-info/" in newname:
+                    egginfo_name, sep, _ = newname.rpartition("/")
+        dir = tempfile.mkdtemp(suffix="_b2w")
+        bdw.extractall(dir, members)
+
+    # egg2wheel
+    abi = "none"
+    pyver = info["pyver"]
+    arch = (info["arch"] or "any").replace(".", "_").replace("-", "_")
+    # Wininst installers always have arch even if they are not
+    # architecture-specific (because the format itself is).
+    # So, assume the content is architecture-neutral if root is purelib.
+    if root_is_purelib:
+        arch = "any"
+    # If the installer is architecture-specific, it's almost certainly also
+    # CPython-specific.
+    if arch != "any":
+        pyver = pyver.replace("py", "cp")
+    wheel_name = "-".join((dist_info, pyver, abi, arch))
+    if root_is_purelib:
+        bw = bdist_wheel(Distribution())
+    else:
+        bw = _bdist_wheel_tag(Distribution())
+
+    bw.root_is_pure = root_is_purelib
+    bw.python_tag = pyver
+    bw.plat_name_supplied = True
+    bw.plat_name = info["arch"] or "any"
+
+    if not root_is_purelib:
+        bw.full_tag_supplied = True
+        bw.full_tag = (pyver, abi, arch)
+
+    dist_info_dir = os.path.join(dir, "%s.dist-info" % dist_info)
+    bw.egg2dist(os.path.join(dir, egginfo_name), dist_info_dir)
+    bw.write_wheelfile(dist_info_dir, generator="wininst2wheel")
+
+    wheel_path = os.path.join(dest_dir, wheel_name)
+    with WheelFile(wheel_path, "w") as wf:
+        wf.write_files(dir)
+
+    shutil.rmtree(dir)
+
+
+def convert(files, dest_dir, verbose):
+    for pat in files:
+        for installer in iglob(pat):
+            if os.path.splitext(installer)[1] == ".egg":
+                conv = egg2wheel
+            else:
+                conv = wininst2wheel
+
+            if verbose:
+                print(f"{installer}... ", flush=True)
+
+            conv(installer, dest_dir)
+            if verbose:
+                print("OK")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/pack.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/pack.py
new file mode 100644
index 00000000..64469c0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/pack.py
@@ -0,0 +1,85 @@
+from __future__ import annotations
+
+import email.policy
+import os.path
+import re
+from email.generator import BytesGenerator
+from email.parser import BytesParser
+
+from wheel.cli import WheelError
+from wheel.wheelfile import WheelFile
+
+DIST_INFO_RE = re.compile(r"^(?P<namever>(?P<name>.+?)-(?P<ver>\d.*?))\.dist-info$")
+
+
+def pack(directory: str, dest_dir: str, build_number: str | None) -> None:
+    """Repack a previously unpacked wheel directory into a new wheel file.
+
+    The .dist-info/WHEEL file must contain one or more tags so that the target
+    wheel file name can be determined.
+
+    :param directory: The unpacked wheel directory
+    :param dest_dir: Destination directory (defaults to the current directory)
+    """
+    # Find the .dist-info directory
+    dist_info_dirs = [
+        fn
+        for fn in os.listdir(directory)
+        if os.path.isdir(os.path.join(directory, fn)) and DIST_INFO_RE.match(fn)
+    ]
+    if len(dist_info_dirs) > 1:
+        raise WheelError(f"Multiple .dist-info directories found in {directory}")
+    elif not dist_info_dirs:
+        raise WheelError(f"No .dist-info directories found in {directory}")
+
+    # Determine the target wheel filename
+    dist_info_dir = dist_info_dirs[0]
+    name_version = DIST_INFO_RE.match(dist_info_dir).group("namever")
+
+    # Read the tags and the existing build number from .dist-info/WHEEL
+    wheel_file_path = os.path.join(directory, dist_info_dir, "WHEEL")
+    with open(wheel_file_path, "rb") as f:
+        info = BytesParser(policy=email.policy.compat32).parse(f)
+        tags: list[str] = info.get_all("Tag", [])
+        existing_build_number = info.get("Build")
+
+        if not tags:
+            raise WheelError(
+                f"No tags present in {dist_info_dir}/WHEEL; cannot determine target "
+                f"wheel filename"
+            )
+
+    # Set the wheel file name and add/replace/remove the Build tag in .dist-info/WHEEL
+    build_number = build_number if build_number is not None else existing_build_number
+    if build_number is not None:
+        del info["Build"]
+        if build_number:
+            info["Build"] = build_number
+            name_version += "-" + build_number
+
+        if build_number != existing_build_number:
+            with open(wheel_file_path, "wb") as f:
+                BytesGenerator(f, maxheaderlen=0).flatten(info)
+
+    # Reassemble the tags for the wheel file
+    tagline = compute_tagline(tags)
+
+    # Repack the wheel
+    wheel_path = os.path.join(dest_dir, f"{name_version}-{tagline}.whl")
+    with WheelFile(wheel_path, "w") as wf:
+        print(f"Repacking wheel as {wheel_path}...", end="", flush=True)
+        wf.write_files(directory)
+
+    print("OK")
+
+
+def compute_tagline(tags: list[str]) -> str:
+    """Compute a tagline from a list of tags.
+
+    :param tags: A list of tags
+    :return: A tagline
+    """
+    impls = sorted({tag.split("-")[0] for tag in tags})
+    abivers = sorted({tag.split("-")[1] for tag in tags})
+    platforms = sorted({tag.split("-")[2] for tag in tags})
+    return "-".join([".".join(impls), ".".join(abivers), ".".join(platforms)])
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/tags.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/tags.py
new file mode 100644
index 00000000..88da72e9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/tags.py
@@ -0,0 +1,139 @@
+from __future__ import annotations
+
+import email.policy
+import itertools
+import os
+from collections.abc import Iterable
+from email.parser import BytesParser
+
+from ..wheelfile import WheelFile
+
+
+def _compute_tags(original_tags: Iterable[str], new_tags: str | None) -> set[str]:
+    """Add or replace tags. Supports dot-separated tags"""
+    if new_tags is None:
+        return set(original_tags)
+
+    if new_tags.startswith("+"):
+        return {*original_tags, *new_tags[1:].split(".")}
+
+    if new_tags.startswith("-"):
+        return set(original_tags) - set(new_tags[1:].split("."))
+
+    return set(new_tags.split("."))
+
+
+def tags(
+    wheel: str,
+    python_tags: str | None = None,
+    abi_tags: str | None = None,
+    platform_tags: str | None = None,
+    build_tag: str | None = None,
+    remove: bool = False,
+) -> str:
+    """Change the tags on a wheel file.
+
+    The tags are left unchanged if they are not specified. To specify "none",
+    use ["none"]. To append to the previous tags, a tag should start with a
+    "+".  If a tag starts with "-", it will be removed from existing tags.
+    Processing is done left to right.
+
+    :param wheel: The paths to the wheels
+    :param python_tags: The Python tags to set
+    :param abi_tags: The ABI tags to set
+    :param platform_tags: The platform tags to set
+    :param build_tag: The build tag to set
+    :param remove: Remove the original wheel
+    """
+    with WheelFile(wheel, "r") as f:
+        assert f.filename, f"{f.filename} must be available"
+
+        wheel_info = f.read(f.dist_info_path + "/WHEEL")
+        info = BytesParser(policy=email.policy.compat32).parsebytes(wheel_info)
+
+        original_wheel_name = os.path.basename(f.filename)
+        namever = f.parsed_filename.group("namever")
+        build = f.parsed_filename.group("build")
+        original_python_tags = f.parsed_filename.group("pyver").split(".")
+        original_abi_tags = f.parsed_filename.group("abi").split(".")
+        original_plat_tags = f.parsed_filename.group("plat").split(".")
+
+    tags: list[str] = info.get_all("Tag", [])
+    existing_build_tag = info.get("Build")
+
+    impls = {tag.split("-")[0] for tag in tags}
+    abivers = {tag.split("-")[1] for tag in tags}
+    platforms = {tag.split("-")[2] for tag in tags}
+
+    if impls != set(original_python_tags):
+        msg = f"Wheel internal tags {impls!r} != filename tags {original_python_tags!r}"
+        raise AssertionError(msg)
+
+    if abivers != set(original_abi_tags):
+        msg = f"Wheel internal tags {abivers!r} != filename tags {original_abi_tags!r}"
+        raise AssertionError(msg)
+
+    if platforms != set(original_plat_tags):
+        msg = (
+            f"Wheel internal tags {platforms!r} != filename tags {original_plat_tags!r}"
+        )
+        raise AssertionError(msg)
+
+    if existing_build_tag != build:
+        msg = (
+            f"Incorrect filename '{build}' "
+            f"& *.dist-info/WHEEL '{existing_build_tag}' build numbers"
+        )
+        raise AssertionError(msg)
+
+    # Start changing as needed
+    if build_tag is not None:
+        build = build_tag
+
+    final_python_tags = sorted(_compute_tags(original_python_tags, python_tags))
+    final_abi_tags = sorted(_compute_tags(original_abi_tags, abi_tags))
+    final_plat_tags = sorted(_compute_tags(original_plat_tags, platform_tags))
+
+    final_tags = [
+        namever,
+        ".".join(final_python_tags),
+        ".".join(final_abi_tags),
+        ".".join(final_plat_tags),
+    ]
+    if build:
+        final_tags.insert(1, build)
+
+    final_wheel_name = "-".join(final_tags) + ".whl"
+
+    if original_wheel_name != final_wheel_name:
+        del info["Tag"], info["Build"]
+        for a, b, c in itertools.product(
+            final_python_tags, final_abi_tags, final_plat_tags
+        ):
+            info["Tag"] = f"{a}-{b}-{c}"
+        if build:
+            info["Build"] = build
+
+        original_wheel_path = os.path.join(
+            os.path.dirname(f.filename), original_wheel_name
+        )
+        final_wheel_path = os.path.join(os.path.dirname(f.filename), final_wheel_name)
+
+        with WheelFile(original_wheel_path, "r") as fin, WheelFile(
+            final_wheel_path, "w"
+        ) as fout:
+            fout.comment = fin.comment  # preserve the comment
+            for item in fin.infolist():
+                if item.is_dir():
+                    continue
+                if item.filename == f.dist_info_path + "/RECORD":
+                    continue
+                if item.filename == f.dist_info_path + "/WHEEL":
+                    fout.writestr(item, info.as_bytes())
+                else:
+                    fout.writestr(item, fin.read(item))
+
+        if remove:
+            os.remove(original_wheel_path)
+
+    return final_wheel_name
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/unpack.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/unpack.py
new file mode 100644
index 00000000..d48840e6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/cli/unpack.py
@@ -0,0 +1,30 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+from ..wheelfile import WheelFile
+
+
+def unpack(path: str, dest: str = ".") -> None:
+    """Unpack a wheel.
+
+    Wheel content will be unpacked to {dest}/{name}-{ver}, where {name}
+    is the package name and {ver} its version.
+
+    :param path: The path to the wheel.
+    :param dest: Destination directory (default to current directory).
+    """
+    with WheelFile(path) as wf:
+        namever = wf.parsed_filename.group("namever")
+        destination = Path(dest) / namever
+        print(f"Unpacking to: {destination}...", end="", flush=True)
+        for zinfo in wf.filelist:
+            wf.extract(zinfo, destination)
+
+            # Set permissions to the same values as they were set in the archive
+            # We have to do this manually due to
+            # https://github.com/python/cpython/issues/59999
+            permissions = zinfo.external_attr >> 16 & 0o777
+            destination.joinpath(zinfo.filename).chmod(permissions)
+
+    print("OK")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/macosx_libfile.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/macosx_libfile.py
new file mode 100644
index 00000000..8953c3f8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/macosx_libfile.py
@@ -0,0 +1,469 @@
+"""
+This module contains function to analyse dynamic library
+headers to extract system information
+
+Currently only for MacOSX
+
+Library file on macosx system starts with Mach-O or Fat field.
+This can be distinguish by first 32 bites and it is called magic number.
+Proper value of magic number is with suffix _MAGIC. Suffix _CIGAM means
+reversed bytes order.
+Both fields can occur in two types: 32 and 64 bytes.
+
+FAT field inform that this library contains few version of library
+(typically for different types version). It contains
+information where Mach-O headers starts.
+
+Each section started with Mach-O header contains one library
+(So if file starts with this field it contains only one version).
+
+After filed Mach-O there are section fields.
+Each of them starts with two fields:
+cmd - magic number for this command
+cmdsize - total size occupied by this section information.
+
+In this case only sections LC_VERSION_MIN_MACOSX (for macosx 10.13 and earlier)
+and LC_BUILD_VERSION (for macosx 10.14 and newer) are interesting,
+because them contains information about minimal system version.
+
+Important remarks:
+- For fat files this implementation looks for maximum number version.
+  It not check if it is 32 or 64 and do not compare it with currently built package.
+  So it is possible to false report higher version that needed.
+- All structures signatures are taken form macosx header files.
+- I think that binary format will be more stable than `otool` output.
+  and if apple introduce some changes both implementation will need to be updated.
+- The system compile will set the deployment target no lower than
+  11.0 for arm64 builds. For "Universal 2" builds use the x86_64 deployment
+  target when the arm64 target is 11.0.
+"""
+
+from __future__ import annotations
+
+import ctypes
+import os
+import sys
+
+"""here the needed const and struct from mach-o header files"""
+
+FAT_MAGIC = 0xCAFEBABE
+FAT_CIGAM = 0xBEBAFECA
+FAT_MAGIC_64 = 0xCAFEBABF
+FAT_CIGAM_64 = 0xBFBAFECA
+MH_MAGIC = 0xFEEDFACE
+MH_CIGAM = 0xCEFAEDFE
+MH_MAGIC_64 = 0xFEEDFACF
+MH_CIGAM_64 = 0xCFFAEDFE
+
+LC_VERSION_MIN_MACOSX = 0x24
+LC_BUILD_VERSION = 0x32
+
+CPU_TYPE_ARM64 = 0x0100000C
+
+mach_header_fields = [
+    ("magic", ctypes.c_uint32),
+    ("cputype", ctypes.c_int),
+    ("cpusubtype", ctypes.c_int),
+    ("filetype", ctypes.c_uint32),
+    ("ncmds", ctypes.c_uint32),
+    ("sizeofcmds", ctypes.c_uint32),
+    ("flags", ctypes.c_uint32),
+]
+"""
+struct mach_header {
+    uint32_t	magic;		/* mach magic number identifier */
+    cpu_type_t	cputype;	/* cpu specifier */
+    cpu_subtype_t	cpusubtype;	/* machine specifier */
+    uint32_t	filetype;	/* type of file */
+    uint32_t	ncmds;		/* number of load commands */
+    uint32_t	sizeofcmds;	/* the size of all the load commands */
+    uint32_t	flags;		/* flags */
+};
+typedef integer_t cpu_type_t;
+typedef integer_t cpu_subtype_t;
+"""
+
+mach_header_fields_64 = mach_header_fields + [("reserved", ctypes.c_uint32)]
+"""
+struct mach_header_64 {
+    uint32_t	magic;		/* mach magic number identifier */
+    cpu_type_t	cputype;	/* cpu specifier */
+    cpu_subtype_t	cpusubtype;	/* machine specifier */
+    uint32_t	filetype;	/* type of file */
+    uint32_t	ncmds;		/* number of load commands */
+    uint32_t	sizeofcmds;	/* the size of all the load commands */
+    uint32_t	flags;		/* flags */
+    uint32_t	reserved;	/* reserved */
+};
+"""
+
+fat_header_fields = [("magic", ctypes.c_uint32), ("nfat_arch", ctypes.c_uint32)]
+"""
+struct fat_header {
+    uint32_t	magic;		/* FAT_MAGIC or FAT_MAGIC_64 */
+    uint32_t	nfat_arch;	/* number of structs that follow */
+};
+"""
+
+fat_arch_fields = [
+    ("cputype", ctypes.c_int),
+    ("cpusubtype", ctypes.c_int),
+    ("offset", ctypes.c_uint32),
+    ("size", ctypes.c_uint32),
+    ("align", ctypes.c_uint32),
+]
+"""
+struct fat_arch {
+    cpu_type_t	cputype;	/* cpu specifier (int) */
+    cpu_subtype_t	cpusubtype;	/* machine specifier (int) */
+    uint32_t	offset;		/* file offset to this object file */
+    uint32_t	size;		/* size of this object file */
+    uint32_t	align;		/* alignment as a power of 2 */
+};
+"""
+
+fat_arch_64_fields = [
+    ("cputype", ctypes.c_int),
+    ("cpusubtype", ctypes.c_int),
+    ("offset", ctypes.c_uint64),
+    ("size", ctypes.c_uint64),
+    ("align", ctypes.c_uint32),
+    ("reserved", ctypes.c_uint32),
+]
+"""
+struct fat_arch_64 {
+    cpu_type_t	cputype;	/* cpu specifier (int) */
+    cpu_subtype_t	cpusubtype;	/* machine specifier (int) */
+    uint64_t	offset;		/* file offset to this object file */
+    uint64_t	size;		/* size of this object file */
+    uint32_t	align;		/* alignment as a power of 2 */
+    uint32_t	reserved;	/* reserved */
+};
+"""
+
+segment_base_fields = [("cmd", ctypes.c_uint32), ("cmdsize", ctypes.c_uint32)]
+"""base for reading segment info"""
+
+segment_command_fields = [
+    ("cmd", ctypes.c_uint32),
+    ("cmdsize", ctypes.c_uint32),
+    ("segname", ctypes.c_char * 16),
+    ("vmaddr", ctypes.c_uint32),
+    ("vmsize", ctypes.c_uint32),
+    ("fileoff", ctypes.c_uint32),
+    ("filesize", ctypes.c_uint32),
+    ("maxprot", ctypes.c_int),
+    ("initprot", ctypes.c_int),
+    ("nsects", ctypes.c_uint32),
+    ("flags", ctypes.c_uint32),
+]
+"""
+struct segment_command { /* for 32-bit architectures */
+    uint32_t	cmd;		/* LC_SEGMENT */
+    uint32_t	cmdsize;	/* includes sizeof section structs */
+    char		segname[16];	/* segment name */
+    uint32_t	vmaddr;		/* memory address of this segment */
+    uint32_t	vmsize;		/* memory size of this segment */
+    uint32_t	fileoff;	/* file offset of this segment */
+    uint32_t	filesize;	/* amount to map from the file */
+    vm_prot_t	maxprot;	/* maximum VM protection */
+    vm_prot_t	initprot;	/* initial VM protection */
+    uint32_t	nsects;		/* number of sections in segment */
+    uint32_t	flags;		/* flags */
+};
+typedef int vm_prot_t;
+"""
+
+segment_command_fields_64 = [
+    ("cmd", ctypes.c_uint32),
+    ("cmdsize", ctypes.c_uint32),
+    ("segname", ctypes.c_char * 16),
+    ("vmaddr", ctypes.c_uint64),
+    ("vmsize", ctypes.c_uint64),
+    ("fileoff", ctypes.c_uint64),
+    ("filesize", ctypes.c_uint64),
+    ("maxprot", ctypes.c_int),
+    ("initprot", ctypes.c_int),
+    ("nsects", ctypes.c_uint32),
+    ("flags", ctypes.c_uint32),
+]
+"""
+struct segment_command_64 { /* for 64-bit architectures */
+    uint32_t	cmd;		/* LC_SEGMENT_64 */
+    uint32_t	cmdsize;	/* includes sizeof section_64 structs */
+    char		segname[16];	/* segment name */
+    uint64_t	vmaddr;		/* memory address of this segment */
+    uint64_t	vmsize;		/* memory size of this segment */
+    uint64_t	fileoff;	/* file offset of this segment */
+    uint64_t	filesize;	/* amount to map from the file */
+    vm_prot_t	maxprot;	/* maximum VM protection */
+    vm_prot_t	initprot;	/* initial VM protection */
+    uint32_t	nsects;		/* number of sections in segment */
+    uint32_t	flags;		/* flags */
+};
+"""
+
+version_min_command_fields = segment_base_fields + [
+    ("version", ctypes.c_uint32),
+    ("sdk", ctypes.c_uint32),
+]
+"""
+struct version_min_command {
+    uint32_t	cmd;		/* LC_VERSION_MIN_MACOSX or
+                               LC_VERSION_MIN_IPHONEOS or
+                               LC_VERSION_MIN_WATCHOS or
+                               LC_VERSION_MIN_TVOS */
+    uint32_t	cmdsize;	/* sizeof(struct min_version_command) */
+    uint32_t	version;	/* X.Y.Z is encoded in nibbles xxxx.yy.zz */
+    uint32_t	sdk;		/* X.Y.Z is encoded in nibbles xxxx.yy.zz */
+};
+"""
+
+build_version_command_fields = segment_base_fields + [
+    ("platform", ctypes.c_uint32),
+    ("minos", ctypes.c_uint32),
+    ("sdk", ctypes.c_uint32),
+    ("ntools", ctypes.c_uint32),
+]
+"""
+struct build_version_command {
+    uint32_t	cmd;		/* LC_BUILD_VERSION */
+    uint32_t	cmdsize;	/* sizeof(struct build_version_command) plus */
+                                /* ntools * sizeof(struct build_tool_version) */
+    uint32_t	platform;	/* platform */
+    uint32_t	minos;		/* X.Y.Z is encoded in nibbles xxxx.yy.zz */
+    uint32_t	sdk;		/* X.Y.Z is encoded in nibbles xxxx.yy.zz */
+    uint32_t	ntools;		/* number of tool entries following this */
+};
+"""
+
+
+def swap32(x):
+    return (
+        ((x << 24) & 0xFF000000)
+        | ((x << 8) & 0x00FF0000)
+        | ((x >> 8) & 0x0000FF00)
+        | ((x >> 24) & 0x000000FF)
+    )
+
+
+def get_base_class_and_magic_number(lib_file, seek=None):
+    if seek is None:
+        seek = lib_file.tell()
+    else:
+        lib_file.seek(seek)
+    magic_number = ctypes.c_uint32.from_buffer_copy(
+        lib_file.read(ctypes.sizeof(ctypes.c_uint32))
+    ).value
+
+    # Handle wrong byte order
+    if magic_number in [FAT_CIGAM, FAT_CIGAM_64, MH_CIGAM, MH_CIGAM_64]:
+        if sys.byteorder == "little":
+            BaseClass = ctypes.BigEndianStructure
+        else:
+            BaseClass = ctypes.LittleEndianStructure
+
+        magic_number = swap32(magic_number)
+    else:
+        BaseClass = ctypes.Structure
+
+    lib_file.seek(seek)
+    return BaseClass, magic_number
+
+
+def read_data(struct_class, lib_file):
+    return struct_class.from_buffer_copy(lib_file.read(ctypes.sizeof(struct_class)))
+
+
+def extract_macosx_min_system_version(path_to_lib):
+    with open(path_to_lib, "rb") as lib_file:
+        BaseClass, magic_number = get_base_class_and_magic_number(lib_file, 0)
+        if magic_number not in [FAT_MAGIC, FAT_MAGIC_64, MH_MAGIC, MH_MAGIC_64]:
+            return
+
+        if magic_number in [FAT_MAGIC, FAT_CIGAM_64]:
+
+            class FatHeader(BaseClass):
+                _fields_ = fat_header_fields
+
+            fat_header = read_data(FatHeader, lib_file)
+            if magic_number == FAT_MAGIC:
+
+                class FatArch(BaseClass):
+                    _fields_ = fat_arch_fields
+
+            else:
+
+                class FatArch(BaseClass):
+                    _fields_ = fat_arch_64_fields
+
+            fat_arch_list = [
+                read_data(FatArch, lib_file) for _ in range(fat_header.nfat_arch)
+            ]
+
+            versions_list = []
+            for el in fat_arch_list:
+                try:
+                    version = read_mach_header(lib_file, el.offset)
+                    if version is not None:
+                        if el.cputype == CPU_TYPE_ARM64 and len(fat_arch_list) != 1:
+                            # Xcode will not set the deployment target below 11.0.0
+                            # for the arm64 architecture. Ignore the arm64 deployment
+                            # in fat binaries when the target is 11.0.0, that way
+                            # the other architectures can select a lower deployment
+                            # target.
+                            # This is safe because there is no arm64 variant for
+                            # macOS 10.15 or earlier.
+                            if version == (11, 0, 0):
+                                continue
+                        versions_list.append(version)
+                except ValueError:
+                    pass
+
+            if len(versions_list) > 0:
+                return max(versions_list)
+            else:
+                return None
+
+        else:
+            try:
+                return read_mach_header(lib_file, 0)
+            except ValueError:
+                """when some error during read library files"""
+                return None
+
+
+def read_mach_header(lib_file, seek=None):
+    """
+    This function parses a Mach-O header and extracts
+    information about the minimal macOS version.
+
+    :param lib_file: reference to opened library file with pointer
+    """
+    base_class, magic_number = get_base_class_and_magic_number(lib_file, seek)
+    arch = "32" if magic_number == MH_MAGIC else "64"
+
+    class SegmentBase(base_class):
+        _fields_ = segment_base_fields
+
+    if arch == "32":
+
+        class MachHeader(base_class):
+            _fields_ = mach_header_fields
+
+    else:
+
+        class MachHeader(base_class):
+            _fields_ = mach_header_fields_64
+
+    mach_header = read_data(MachHeader, lib_file)
+    for _i in range(mach_header.ncmds):
+        pos = lib_file.tell()
+        segment_base = read_data(SegmentBase, lib_file)
+        lib_file.seek(pos)
+        if segment_base.cmd == LC_VERSION_MIN_MACOSX:
+
+            class VersionMinCommand(base_class):
+                _fields_ = version_min_command_fields
+
+            version_info = read_data(VersionMinCommand, lib_file)
+            return parse_version(version_info.version)
+        elif segment_base.cmd == LC_BUILD_VERSION:
+
+            class VersionBuild(base_class):
+                _fields_ = build_version_command_fields
+
+            version_info = read_data(VersionBuild, lib_file)
+            return parse_version(version_info.minos)
+        else:
+            lib_file.seek(pos + segment_base.cmdsize)
+            continue
+
+
+def parse_version(version):
+    x = (version & 0xFFFF0000) >> 16
+    y = (version & 0x0000FF00) >> 8
+    z = version & 0x000000FF
+    return x, y, z
+
+
+def calculate_macosx_platform_tag(archive_root, platform_tag):
+    """
+    Calculate proper macosx platform tag basing on files which are included to wheel
+
+    Example platform tag `macosx-10.14-x86_64`
+    """
+    prefix, base_version, suffix = platform_tag.split("-")
+    base_version = tuple(int(x) for x in base_version.split("."))
+    base_version = base_version[:2]
+    if base_version[0] > 10:
+        base_version = (base_version[0], 0)
+    assert len(base_version) == 2
+    if "MACOSX_DEPLOYMENT_TARGET" in os.environ:
+        deploy_target = tuple(
+            int(x) for x in os.environ["MACOSX_DEPLOYMENT_TARGET"].split(".")
+        )
+        deploy_target = deploy_target[:2]
+        if deploy_target[0] > 10:
+            deploy_target = (deploy_target[0], 0)
+        if deploy_target < base_version:
+            sys.stderr.write(
+                "[WARNING] MACOSX_DEPLOYMENT_TARGET is set to a lower value ({}) than "
+                "the version on which the Python interpreter was compiled ({}), and "
+                "will be ignored.\n".format(
+                    ".".join(str(x) for x in deploy_target),
+                    ".".join(str(x) for x in base_version),
+                )
+            )
+        else:
+            base_version = deploy_target
+
+    assert len(base_version) == 2
+    start_version = base_version
+    versions_dict = {}
+    for dirpath, _dirnames, filenames in os.walk(archive_root):
+        for filename in filenames:
+            if filename.endswith(".dylib") or filename.endswith(".so"):
+                lib_path = os.path.join(dirpath, filename)
+                min_ver = extract_macosx_min_system_version(lib_path)
+                if min_ver is not None:
+                    min_ver = min_ver[0:2]
+                    if min_ver[0] > 10:
+                        min_ver = (min_ver[0], 0)
+                    versions_dict[lib_path] = min_ver
+
+    if len(versions_dict) > 0:
+        base_version = max(base_version, max(versions_dict.values()))
+
+    # macosx platform tag do not support minor bugfix release
+    fin_base_version = "_".join([str(x) for x in base_version])
+    if start_version < base_version:
+        problematic_files = [k for k, v in versions_dict.items() if v > start_version]
+        problematic_files = "\n".join(problematic_files)
+        if len(problematic_files) == 1:
+            files_form = "this file"
+        else:
+            files_form = "these files"
+        error_message = (
+            "[WARNING] This wheel needs a higher macOS version than {}  "
+            "To silence this warning, set MACOSX_DEPLOYMENT_TARGET to at least "
+            + fin_base_version
+            + " or recreate "
+            + files_form
+            + " with lower "
+            "MACOSX_DEPLOYMENT_TARGET:  \n" + problematic_files
+        )
+
+        if "MACOSX_DEPLOYMENT_TARGET" in os.environ:
+            error_message = error_message.format(
+                "is set in MACOSX_DEPLOYMENT_TARGET variable."
+            )
+        else:
+            error_message = error_message.format(
+                "the version your Python interpreter is compiled against."
+            )
+
+        sys.stderr.write(error_message)
+
+    platform_tag = prefix + "_" + fin_base_version + "_" + suffix
+    return platform_tag
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/metadata.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/metadata.py
new file mode 100644
index 00000000..6aa43628
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/metadata.py
@@ -0,0 +1,180 @@
+"""
+Tools for converting old- to new-style metadata.
+"""
+
+from __future__ import annotations
+
+import functools
+import itertools
+import os.path
+import re
+import textwrap
+from email.message import Message
+from email.parser import Parser
+from typing import Iterator
+
+from .vendored.packaging.requirements import Requirement
+
+
+def _nonblank(str):
+    return str and not str.startswith("#")
+
+
+@functools.singledispatch
+def yield_lines(iterable):
+    r"""
+    Yield valid lines of a string or iterable.
+    >>> list(yield_lines(''))
+    []
+    >>> list(yield_lines(['foo', 'bar']))
+    ['foo', 'bar']
+    >>> list(yield_lines('foo\nbar'))
+    ['foo', 'bar']
+    >>> list(yield_lines('\nfoo\n#bar\nbaz #comment'))
+    ['foo', 'baz #comment']
+    >>> list(yield_lines(['foo\nbar', 'baz', 'bing\n\n\n']))
+    ['foo', 'bar', 'baz', 'bing']
+    """
+    return itertools.chain.from_iterable(map(yield_lines, iterable))
+
+
+@yield_lines.register(str)
+def _(text):
+    return filter(_nonblank, map(str.strip, text.splitlines()))
+
+
+def split_sections(s):
+    """Split a string or iterable thereof into (section, content) pairs
+    Each ``section`` is a stripped version of the section header ("[section]")
+    and each ``content`` is a list of stripped lines excluding blank lines and
+    comment-only lines.  If there are any such lines before the first section
+    header, they're returned in a first ``section`` of ``None``.
+    """
+    section = None
+    content = []
+    for line in yield_lines(s):
+        if line.startswith("["):
+            if line.endswith("]"):
+                if section or content:
+                    yield section, content
+                section = line[1:-1].strip()
+                content = []
+            else:
+                raise ValueError("Invalid section heading", line)
+        else:
+            content.append(line)
+
+    # wrap up last segment
+    yield section, content
+
+
+def safe_extra(extra):
+    """Convert an arbitrary string to a standard 'extra' name
+    Any runs of non-alphanumeric characters are replaced with a single '_',
+    and the result is always lowercased.
+    """
+    return re.sub("[^A-Za-z0-9.-]+", "_", extra).lower()
+
+
+def safe_name(name):
+    """Convert an arbitrary string to a standard distribution name
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub("[^A-Za-z0-9.]+", "-", name)
+
+
+def requires_to_requires_dist(requirement: Requirement) -> str:
+    """Return the version specifier for a requirement in PEP 345/566 fashion."""
+    if getattr(requirement, "url", None):
+        return " @ " + requirement.url
+
+    requires_dist = []
+    for spec in requirement.specifier:
+        requires_dist.append(spec.operator + spec.version)
+
+    if requires_dist:
+        return " " + ",".join(sorted(requires_dist))
+    else:
+        return ""
+
+
+def convert_requirements(requirements: list[str]) -> Iterator[str]:
+    """Yield Requires-Dist: strings for parsed requirements strings."""
+    for req in requirements:
+        parsed_requirement = Requirement(req)
+        spec = requires_to_requires_dist(parsed_requirement)
+        extras = ",".join(sorted(safe_extra(e) for e in parsed_requirement.extras))
+        if extras:
+            extras = f"[{extras}]"
+
+        yield safe_name(parsed_requirement.name) + extras + spec
+
+
+def generate_requirements(
+    extras_require: dict[str, list[str]],
+) -> Iterator[tuple[str, str]]:
+    """
+    Convert requirements from a setup()-style dictionary to
+    ('Requires-Dist', 'requirement') and ('Provides-Extra', 'extra') tuples.
+
+    extras_require is a dictionary of {extra: [requirements]} as passed to setup(),
+    using the empty extra {'': [requirements]} to hold install_requires.
+    """
+    for extra, depends in extras_require.items():
+        condition = ""
+        extra = extra or ""
+        if ":" in extra:  # setuptools extra:condition syntax
+            extra, condition = extra.split(":", 1)
+
+        extra = safe_extra(extra)
+        if extra:
+            yield "Provides-Extra", extra
+            if condition:
+                condition = "(" + condition + ") and "
+            condition += "extra == '%s'" % extra
+
+        if condition:
+            condition = " ; " + condition
+
+        for new_req in convert_requirements(depends):
+            yield "Requires-Dist", new_req + condition
+
+
+def pkginfo_to_metadata(egg_info_path: str, pkginfo_path: str) -> Message:
+    """
+    Convert .egg-info directory with PKG-INFO to the Metadata 2.1 format
+    """
+    with open(pkginfo_path, encoding="utf-8") as headers:
+        pkg_info = Parser().parse(headers)
+
+    pkg_info.replace_header("Metadata-Version", "2.1")
+    # Those will be regenerated from `requires.txt`.
+    del pkg_info["Provides-Extra"]
+    del pkg_info["Requires-Dist"]
+    requires_path = os.path.join(egg_info_path, "requires.txt")
+    if os.path.exists(requires_path):
+        with open(requires_path, encoding="utf-8") as requires_file:
+            requires = requires_file.read()
+
+        parsed_requirements = sorted(split_sections(requires), key=lambda x: x[0] or "")
+        for extra, reqs in parsed_requirements:
+            for key, value in generate_requirements({extra: reqs}):
+                if (key, value) not in pkg_info.items():
+                    pkg_info[key] = value
+
+    description = pkg_info["Description"]
+    if description:
+        description_lines = pkg_info["Description"].splitlines()
+        dedented_description = "\n".join(
+            # if the first line of long_description is blank,
+            # the first line here will be indented.
+            (
+                description_lines[0].lstrip(),
+                textwrap.dedent("\n".join(description_lines[1:])),
+                "\n",
+            )
+        )
+        pkg_info.set_payload(dedented_description)
+        del pkg_info["Description"]
+
+    return pkg_info
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/util.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/util.py
new file mode 100644
index 00000000..d98d98cb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/util.py
@@ -0,0 +1,26 @@
+from __future__ import annotations
+
+import base64
+import logging
+
+log = logging.getLogger("wheel")
+
+# ensure Python logging is configured
+try:
+    __import__("setuptools.logging")
+except ImportError:
+    # setuptools < ??
+    from . import _setuptools_logging
+
+    _setuptools_logging.configure()
+
+
+def urlsafe_b64encode(data: bytes) -> bytes:
+    """urlsafe_b64encode without padding"""
+    return base64.urlsafe_b64encode(data).rstrip(b"=")
+
+
+def urlsafe_b64decode(data: bytes) -> bytes:
+    """urlsafe_b64decode without padding"""
+    pad = b"=" * (4 - (len(data) & 3))
+    return base64.urlsafe_b64decode(data + pad)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3d4440b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..74cd6c15
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_elffile.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_elffile.cpython-312.pyc
new file mode 100644
index 00000000..4fecc5a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_elffile.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_manylinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_manylinux.cpython-312.pyc
new file mode 100644
index 00000000..269184ba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_manylinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_musllinux.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_musllinux.cpython-312.pyc
new file mode 100644
index 00000000..61b92ef5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_musllinux.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_parser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_parser.cpython-312.pyc
new file mode 100644
index 00000000..4a7b546f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_parser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_structures.cpython-312.pyc
new file mode 100644
index 00000000..65f215ba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_tokenizer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_tokenizer.cpython-312.pyc
new file mode 100644
index 00000000..eabf62d5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/_tokenizer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/markers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/markers.cpython-312.pyc
new file mode 100644
index 00000000..8a1ff177
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/markers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..29ebbd91
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/specifiers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/specifiers.cpython-312.pyc
new file mode 100644
index 00000000..8cb2a6e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/specifiers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/tags.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/tags.cpython-312.pyc
new file mode 100644
index 00000000..33d210a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/tags.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..f47936f7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..46fca2de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_elffile.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_elffile.py
new file mode 100644
index 00000000..6fb19b30
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_elffile.py
@@ -0,0 +1,108 @@
+"""
+ELF file parser.
+
+This provides a class ``ELFFile`` that parses an ELF executable in a similar
+interface to ``ZipFile``. Only the read interface is implemented.
+
+Based on: https://gist.github.com/lyssdod/f51579ae8d93c8657a5564aefc2ffbca
+ELF header: https://refspecs.linuxfoundation.org/elf/gabi4+/ch4.eheader.html
+"""
+
+import enum
+import os
+import struct
+from typing import IO, Optional, Tuple
+
+
+class ELFInvalid(ValueError):
+    pass
+
+
+class EIClass(enum.IntEnum):
+    C32 = 1
+    C64 = 2
+
+
+class EIData(enum.IntEnum):
+    Lsb = 1
+    Msb = 2
+
+
+class EMachine(enum.IntEnum):
+    I386 = 3
+    S390 = 22
+    Arm = 40
+    X8664 = 62
+    AArc64 = 183
+
+
+class ELFFile:
+    """
+    Representation of an ELF executable.
+    """
+
+    def __init__(self, f: IO[bytes]) -> None:
+        self._f = f
+
+        try:
+            ident = self._read("16B")
+        except struct.error:
+            raise ELFInvalid("unable to parse identification")
+        magic = bytes(ident[:4])
+        if magic != b"\x7fELF":
+            raise ELFInvalid(f"invalid magic: {magic!r}")
+
+        self.capacity = ident[4]  # Format for program header (bitness).
+        self.encoding = ident[5]  # Data structure encoding (endianness).
+
+        try:
+            # e_fmt: Format for program header.
+            # p_fmt: Format for section header.
+            # p_idx: Indexes to find p_type, p_offset, and p_filesz.
+            e_fmt, self._p_fmt, self._p_idx = {
+                (1, 1): ("<HHIIIIIHHH", "<IIIIIIII", (0, 1, 4)),  # 32-bit LSB.
+                (1, 2): (">HHIIIIIHHH", ">IIIIIIII", (0, 1, 4)),  # 32-bit MSB.
+                (2, 1): ("<HHIQQQIHHH", "<IIQQQQQQ", (0, 2, 5)),  # 64-bit LSB.
+                (2, 2): (">HHIQQQIHHH", ">IIQQQQQQ", (0, 2, 5)),  # 64-bit MSB.
+            }[(self.capacity, self.encoding)]
+        except KeyError:
+            raise ELFInvalid(
+                f"unrecognized capacity ({self.capacity}) or "
+                f"encoding ({self.encoding})"
+            )
+
+        try:
+            (
+                _,
+                self.machine,  # Architecture type.
+                _,
+                _,
+                self._e_phoff,  # Offset of program header.
+                _,
+                self.flags,  # Processor-specific flags.
+                _,
+                self._e_phentsize,  # Size of section.
+                self._e_phnum,  # Number of sections.
+            ) = self._read(e_fmt)
+        except struct.error as e:
+            raise ELFInvalid("unable to parse machine and section information") from e
+
+    def _read(self, fmt: str) -> Tuple[int, ...]:
+        return struct.unpack(fmt, self._f.read(struct.calcsize(fmt)))
+
+    @property
+    def interpreter(self) -> Optional[str]:
+        """
+        The path recorded in the ``PT_INTERP`` section header.
+        """
+        for index in range(self._e_phnum):
+            self._f.seek(self._e_phoff + self._e_phentsize * index)
+            try:
+                data = self._read(self._p_fmt)
+            except struct.error:
+                continue
+            if data[self._p_idx[0]] != 3:  # Not PT_INTERP.
+                continue
+            self._f.seek(data[self._p_idx[1]])
+            return os.fsdecode(self._f.read(data[self._p_idx[2]])).strip("\0")
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_manylinux.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_manylinux.py
new file mode 100644
index 00000000..1f5f4ab3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_manylinux.py
@@ -0,0 +1,260 @@
+import collections
+import contextlib
+import functools
+import os
+import re
+import sys
+import warnings
+from typing import Dict, Generator, Iterator, NamedTuple, Optional, Sequence, Tuple
+
+from ._elffile import EIClass, EIData, ELFFile, EMachine
+
+EF_ARM_ABIMASK = 0xFF000000
+EF_ARM_ABI_VER5 = 0x05000000
+EF_ARM_ABI_FLOAT_HARD = 0x00000400
+
+
+# `os.PathLike` not a generic type until Python 3.9, so sticking with `str`
+# as the type for `path` until then.
+@contextlib.contextmanager
+def _parse_elf(path: str) -> Generator[Optional[ELFFile], None, None]:
+    try:
+        with open(path, "rb") as f:
+            yield ELFFile(f)
+    except (OSError, TypeError, ValueError):
+        yield None
+
+
+def _is_linux_armhf(executable: str) -> bool:
+    # hard-float ABI can be detected from the ELF header of the running
+    # process
+    # https://static.docs.arm.com/ihi0044/g/aaelf32.pdf
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.Arm
+            and f.flags & EF_ARM_ABIMASK == EF_ARM_ABI_VER5
+            and f.flags & EF_ARM_ABI_FLOAT_HARD == EF_ARM_ABI_FLOAT_HARD
+        )
+
+
+def _is_linux_i686(executable: str) -> bool:
+    with _parse_elf(executable) as f:
+        return (
+            f is not None
+            and f.capacity == EIClass.C32
+            and f.encoding == EIData.Lsb
+            and f.machine == EMachine.I386
+        )
+
+
+def _have_compatible_abi(executable: str, archs: Sequence[str]) -> bool:
+    if "armv7l" in archs:
+        return _is_linux_armhf(executable)
+    if "i686" in archs:
+        return _is_linux_i686(executable)
+    allowed_archs = {
+        "x86_64",
+        "aarch64",
+        "ppc64",
+        "ppc64le",
+        "s390x",
+        "loongarch64",
+        "riscv64",
+    }
+    return any(arch in allowed_archs for arch in archs)
+
+
+# If glibc ever changes its major version, we need to know what the last
+# minor version was, so we can build the complete list of all versions.
+# For now, guess what the highest minor version might be, assume it will
+# be 50 for testing. Once this actually happens, update the dictionary
+# with the actual value.
+_LAST_GLIBC_MINOR: Dict[int, int] = collections.defaultdict(lambda: 50)
+
+
+class _GLibCVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _glibc_version_string_confstr() -> Optional[str]:
+    """
+    Primary implementation of glibc_version_string using os.confstr.
+    """
+    # os.confstr is quite a bit faster than ctypes.DLL. It's also less likely
+    # to be broken or missing. This strategy is used in the standard library
+    # platform module.
+    # https://github.com/python/cpython/blob/fcf1d003bf4f0100c/Lib/platform.py#L175-L183
+    try:
+        # Should be a string like "glibc 2.17".
+        version_string: Optional[str] = os.confstr("CS_GNU_LIBC_VERSION")
+        assert version_string is not None
+        _, version = version_string.rsplit()
+    except (AssertionError, AttributeError, OSError, ValueError):
+        # os.confstr() or CS_GNU_LIBC_VERSION not available (or a bad value)...
+        return None
+    return version
+
+
+def _glibc_version_string_ctypes() -> Optional[str]:
+    """
+    Fallback implementation of glibc_version_string using ctypes.
+    """
+    try:
+        import ctypes
+    except ImportError:
+        return None
+
+    # ctypes.CDLL(None) internally calls dlopen(NULL), and as the dlopen
+    # manpage says, "If filename is NULL, then the returned handle is for the
+    # main program". This way we can let the linker do the work to figure out
+    # which libc our process is actually using.
+    #
+    # We must also handle the special case where the executable is not a
+    # dynamically linked executable. This can occur when using musl libc,
+    # for example. In this situation, dlopen() will error, leading to an
+    # OSError. Interestingly, at least in the case of musl, there is no
+    # errno set on the OSError. The single string argument used to construct
+    # OSError comes from libc itself and is therefore not portable to
+    # hard code here. In any case, failure to call dlopen() means we
+    # can proceed, so we bail on our attempt.
+    try:
+        process_namespace = ctypes.CDLL(None)
+    except OSError:
+        return None
+
+    try:
+        gnu_get_libc_version = process_namespace.gnu_get_libc_version
+    except AttributeError:
+        # Symbol doesn't exist -> therefore, we are not linked to
+        # glibc.
+        return None
+
+    # Call gnu_get_libc_version, which returns a string like "2.5"
+    gnu_get_libc_version.restype = ctypes.c_char_p
+    version_str: str = gnu_get_libc_version()
+    # py2 / py3 compatibility:
+    if not isinstance(version_str, str):
+        version_str = version_str.decode("ascii")
+
+    return version_str
+
+
+def _glibc_version_string() -> Optional[str]:
+    """Returns glibc version string, or None if not using glibc."""
+    return _glibc_version_string_confstr() or _glibc_version_string_ctypes()
+
+
+def _parse_glibc_version(version_str: str) -> Tuple[int, int]:
+    """Parse glibc version.
+
+    We use a regexp instead of str.split because we want to discard any
+    random junk that might come after the minor version -- this might happen
+    in patched/forked versions of glibc (e.g. Linaro's version of glibc
+    uses version strings like "2.20-2014.11"). See gh-3588.
+    """
+    m = re.match(r"(?P<major>[0-9]+)\.(?P<minor>[0-9]+)", version_str)
+    if not m:
+        warnings.warn(
+            f"Expected glibc version with 2 components major.minor,"
+            f" got: {version_str}",
+            RuntimeWarning,
+        )
+        return -1, -1
+    return int(m.group("major")), int(m.group("minor"))
+
+
+@functools.lru_cache
+def _get_glibc_version() -> Tuple[int, int]:
+    version_str = _glibc_version_string()
+    if version_str is None:
+        return (-1, -1)
+    return _parse_glibc_version(version_str)
+
+
+# From PEP 513, PEP 600
+def _is_compatible(arch: str, version: _GLibCVersion) -> bool:
+    sys_glibc = _get_glibc_version()
+    if sys_glibc < version:
+        return False
+    # Check for presence of _manylinux module.
+    try:
+        import _manylinux
+    except ImportError:
+        return True
+    if hasattr(_manylinux, "manylinux_compatible"):
+        result = _manylinux.manylinux_compatible(version[0], version[1], arch)
+        if result is not None:
+            return bool(result)
+        return True
+    if version == _GLibCVersion(2, 5):
+        if hasattr(_manylinux, "manylinux1_compatible"):
+            return bool(_manylinux.manylinux1_compatible)
+    if version == _GLibCVersion(2, 12):
+        if hasattr(_manylinux, "manylinux2010_compatible"):
+            return bool(_manylinux.manylinux2010_compatible)
+    if version == _GLibCVersion(2, 17):
+        if hasattr(_manylinux, "manylinux2014_compatible"):
+            return bool(_manylinux.manylinux2014_compatible)
+    return True
+
+
+_LEGACY_MANYLINUX_MAP = {
+    # CentOS 7 w/ glibc 2.17 (PEP 599)
+    (2, 17): "manylinux2014",
+    # CentOS 6 w/ glibc 2.12 (PEP 571)
+    (2, 12): "manylinux2010",
+    # CentOS 5 w/ glibc 2.5 (PEP 513)
+    (2, 5): "manylinux1",
+}
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate manylinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be manylinux-compatible.
+
+    :returns: An iterator of compatible manylinux tags.
+    """
+    if not _have_compatible_abi(sys.executable, archs):
+        return
+    # Oldest glibc to be supported regardless of architecture is (2, 17).
+    too_old_glibc2 = _GLibCVersion(2, 16)
+    if set(archs) & {"x86_64", "i686"}:
+        # On x86/i686 also oldest glibc to be supported is (2, 5).
+        too_old_glibc2 = _GLibCVersion(2, 4)
+    current_glibc = _GLibCVersion(*_get_glibc_version())
+    glibc_max_list = [current_glibc]
+    # We can assume compatibility across glibc major versions.
+    # https://sourceware.org/bugzilla/show_bug.cgi?id=24636
+    #
+    # Build a list of maximum glibc versions so that we can
+    # output the canonical list of all glibc from current_glibc
+    # down to too_old_glibc2, including all intermediary versions.
+    for glibc_major in range(current_glibc.major - 1, 1, -1):
+        glibc_minor = _LAST_GLIBC_MINOR[glibc_major]
+        glibc_max_list.append(_GLibCVersion(glibc_major, glibc_minor))
+    for arch in archs:
+        for glibc_max in glibc_max_list:
+            if glibc_max.major == too_old_glibc2.major:
+                min_minor = too_old_glibc2.minor
+            else:
+                # For other glibc major versions oldest supported is (x, 0).
+                min_minor = -1
+            for glibc_minor in range(glibc_max.minor, min_minor, -1):
+                glibc_version = _GLibCVersion(glibc_max.major, glibc_minor)
+                tag = "manylinux_{}_{}".format(*glibc_version)
+                if _is_compatible(arch, glibc_version):
+                    yield f"{tag}_{arch}"
+                # Handle the legacy manylinux1, manylinux2010, manylinux2014 tags.
+                if glibc_version in _LEGACY_MANYLINUX_MAP:
+                    legacy_tag = _LEGACY_MANYLINUX_MAP[glibc_version]
+                    if _is_compatible(arch, glibc_version):
+                        yield f"{legacy_tag}_{arch}"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_musllinux.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_musllinux.py
new file mode 100644
index 00000000..eb4251b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_musllinux.py
@@ -0,0 +1,83 @@
+"""PEP 656 support.
+
+This module implements logic to detect if the currently running Python is
+linked against musl, and what musl version is used.
+"""
+
+import functools
+import re
+import subprocess
+import sys
+from typing import Iterator, NamedTuple, Optional, Sequence
+
+from ._elffile import ELFFile
+
+
+class _MuslVersion(NamedTuple):
+    major: int
+    minor: int
+
+
+def _parse_musl_version(output: str) -> Optional[_MuslVersion]:
+    lines = [n for n in (n.strip() for n in output.splitlines()) if n]
+    if len(lines) < 2 or lines[0][:4] != "musl":
+        return None
+    m = re.match(r"Version (\d+)\.(\d+)", lines[1])
+    if not m:
+        return None
+    return _MuslVersion(major=int(m.group(1)), minor=int(m.group(2)))
+
+
+@functools.lru_cache
+def _get_musl_version(executable: str) -> Optional[_MuslVersion]:
+    """Detect currently-running musl runtime version.
+
+    This is done by checking the specified executable's dynamic linking
+    information, and invoking the loader to parse its output for a version
+    string. If the loader is musl, the output would be something like::
+
+        musl libc (x86_64)
+        Version 1.2.2
+        Dynamic Program Loader
+    """
+    try:
+        with open(executable, "rb") as f:
+            ld = ELFFile(f).interpreter
+    except (OSError, TypeError, ValueError):
+        return None
+    if ld is None or "musl" not in ld:
+        return None
+    proc = subprocess.run([ld], stderr=subprocess.PIPE, text=True)
+    return _parse_musl_version(proc.stderr)
+
+
+def platform_tags(archs: Sequence[str]) -> Iterator[str]:
+    """Generate musllinux tags compatible to the current platform.
+
+    :param archs: Sequence of compatible architectures.
+        The first one shall be the closest to the actual architecture and be the part of
+        platform tag after the ``linux_`` prefix, e.g. ``x86_64``.
+        The ``linux_`` prefix is assumed as a prerequisite for the current platform to
+        be musllinux-compatible.
+
+    :returns: An iterator of compatible musllinux tags.
+    """
+    sys_musl = _get_musl_version(sys.executable)
+    if sys_musl is None:  # Python not dynamically linked against musl.
+        return
+    for arch in archs:
+        for minor in range(sys_musl.minor, -1, -1):
+            yield f"musllinux_{sys_musl.major}_{minor}_{arch}"
+
+
+if __name__ == "__main__":  # pragma: no cover
+    import sysconfig
+
+    plat = sysconfig.get_platform()
+    assert plat.startswith("linux-"), "not linux"
+
+    print("plat:", plat)
+    print("musl:", _get_musl_version(sys.executable))
+    print("tags:", end=" ")
+    for t in platform_tags(re.sub(r"[.-]", "_", plat.split("-", 1)[-1])):
+        print(t, end="\n      ")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_parser.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_parser.py
new file mode 100644
index 00000000..513686a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_parser.py
@@ -0,0 +1,356 @@
+"""Handwritten parser of dependency specifiers.
+
+The docstring for each __parse_* function contains EBNF-inspired grammar representing
+the implementation.
+"""
+
+import ast
+from typing import Any, List, NamedTuple, Optional, Tuple, Union
+
+from ._tokenizer import DEFAULT_RULES, Tokenizer
+
+
+class Node:
+    def __init__(self, value: str) -> None:
+        self.value = value
+
+    def __str__(self) -> str:
+        return self.value
+
+    def __repr__(self) -> str:
+        return f"<{self.__class__.__name__}('{self}')>"
+
+    def serialize(self) -> str:
+        raise NotImplementedError
+
+
+class Variable(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+class Value(Node):
+    def serialize(self) -> str:
+        return f'"{self}"'
+
+
+class Op(Node):
+    def serialize(self) -> str:
+        return str(self)
+
+
+MarkerVar = Union[Variable, Value]
+MarkerItem = Tuple[MarkerVar, Op, MarkerVar]
+# MarkerAtom = Union[MarkerItem, List["MarkerAtom"]]
+# MarkerList = List[Union["MarkerList", MarkerAtom, str]]
+# mypy does not support recursive type definition
+# https://github.com/python/mypy/issues/731
+MarkerAtom = Any
+MarkerList = List[Any]
+
+
+class ParsedRequirement(NamedTuple):
+    name: str
+    url: str
+    extras: List[str]
+    specifier: str
+    marker: Optional[MarkerList]
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for dependency specifier
+# --------------------------------------------------------------------------------------
+def parse_requirement(source: str) -> ParsedRequirement:
+    return _parse_requirement(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_requirement(tokenizer: Tokenizer) -> ParsedRequirement:
+    """
+    requirement = WS? IDENTIFIER WS? extras WS? requirement_details
+    """
+    tokenizer.consume("WS")
+
+    name_token = tokenizer.expect(
+        "IDENTIFIER", expected="package name at the start of dependency specifier"
+    )
+    name = name_token.text
+    tokenizer.consume("WS")
+
+    extras = _parse_extras(tokenizer)
+    tokenizer.consume("WS")
+
+    url, specifier, marker = _parse_requirement_details(tokenizer)
+    tokenizer.expect("END", expected="end of dependency specifier")
+
+    return ParsedRequirement(name, url, extras, specifier, marker)
+
+
+def _parse_requirement_details(
+    tokenizer: Tokenizer,
+) -> Tuple[str, str, Optional[MarkerList]]:
+    """
+    requirement_details = AT URL (WS requirement_marker?)?
+                        | specifier WS? (requirement_marker)?
+    """
+
+    specifier = ""
+    url = ""
+    marker = None
+
+    if tokenizer.check("AT"):
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        url_start = tokenizer.position
+        url = tokenizer.expect("URL", expected="URL after @").text
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        tokenizer.expect("WS", expected="whitespace after URL")
+
+        # The input might end after whitespace.
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer, span_start=url_start, after="URL and whitespace"
+        )
+    else:
+        specifier_start = tokenizer.position
+        specifier = _parse_specifier(tokenizer)
+        tokenizer.consume("WS")
+
+        if tokenizer.check("END", peek=True):
+            return (url, specifier, marker)
+
+        marker = _parse_requirement_marker(
+            tokenizer,
+            span_start=specifier_start,
+            after=(
+                "version specifier"
+                if specifier
+                else "name and no valid version specifier"
+            ),
+        )
+
+    return (url, specifier, marker)
+
+
+def _parse_requirement_marker(
+    tokenizer: Tokenizer, *, span_start: int, after: str
+) -> MarkerList:
+    """
+    requirement_marker = SEMICOLON marker WS?
+    """
+
+    if not tokenizer.check("SEMICOLON"):
+        tokenizer.raise_syntax_error(
+            f"Expected end or semicolon (after {after})",
+            span_start=span_start,
+        )
+    tokenizer.read()
+
+    marker = _parse_marker(tokenizer)
+    tokenizer.consume("WS")
+
+    return marker
+
+
+def _parse_extras(tokenizer: Tokenizer) -> List[str]:
+    """
+    extras = (LEFT_BRACKET wsp* extras_list? wsp* RIGHT_BRACKET)?
+    """
+    if not tokenizer.check("LEFT_BRACKET", peek=True):
+        return []
+
+    with tokenizer.enclosing_tokens(
+        "LEFT_BRACKET",
+        "RIGHT_BRACKET",
+        around="extras",
+    ):
+        tokenizer.consume("WS")
+        extras = _parse_extras_list(tokenizer)
+        tokenizer.consume("WS")
+
+    return extras
+
+
+def _parse_extras_list(tokenizer: Tokenizer) -> List[str]:
+    """
+    extras_list = identifier (wsp* ',' wsp* identifier)*
+    """
+    extras: List[str] = []
+
+    if not tokenizer.check("IDENTIFIER"):
+        return extras
+
+    extras.append(tokenizer.read().text)
+
+    while True:
+        tokenizer.consume("WS")
+        if tokenizer.check("IDENTIFIER", peek=True):
+            tokenizer.raise_syntax_error("Expected comma between extra names")
+        elif not tokenizer.check("COMMA"):
+            break
+
+        tokenizer.read()
+        tokenizer.consume("WS")
+
+        extra_token = tokenizer.expect("IDENTIFIER", expected="extra name after comma")
+        extras.append(extra_token.text)
+
+    return extras
+
+
+def _parse_specifier(tokenizer: Tokenizer) -> str:
+    """
+    specifier = LEFT_PARENTHESIS WS? version_many WS? RIGHT_PARENTHESIS
+              | WS? version_many WS?
+    """
+    with tokenizer.enclosing_tokens(
+        "LEFT_PARENTHESIS",
+        "RIGHT_PARENTHESIS",
+        around="version specifier",
+    ):
+        tokenizer.consume("WS")
+        parsed_specifiers = _parse_version_many(tokenizer)
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+def _parse_version_many(tokenizer: Tokenizer) -> str:
+    """
+    version_many = (SPECIFIER (WS? COMMA WS? SPECIFIER)*)?
+    """
+    parsed_specifiers = ""
+    while tokenizer.check("SPECIFIER"):
+        span_start = tokenizer.position
+        parsed_specifiers += tokenizer.read().text
+        if tokenizer.check("VERSION_PREFIX_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                ".* suffix can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position + 1,
+            )
+        if tokenizer.check("VERSION_LOCAL_LABEL_TRAIL", peek=True):
+            tokenizer.raise_syntax_error(
+                "Local version label can only be used with `==` or `!=` operators",
+                span_start=span_start,
+                span_end=tokenizer.position,
+            )
+        tokenizer.consume("WS")
+        if not tokenizer.check("COMMA"):
+            break
+        parsed_specifiers += tokenizer.read().text
+        tokenizer.consume("WS")
+
+    return parsed_specifiers
+
+
+# --------------------------------------------------------------------------------------
+# Recursive descent parser for marker expression
+# --------------------------------------------------------------------------------------
+def parse_marker(source: str) -> MarkerList:
+    return _parse_full_marker(Tokenizer(source, rules=DEFAULT_RULES))
+
+
+def _parse_full_marker(tokenizer: Tokenizer) -> MarkerList:
+    retval = _parse_marker(tokenizer)
+    tokenizer.expect("END", expected="end of marker expression")
+    return retval
+
+
+def _parse_marker(tokenizer: Tokenizer) -> MarkerList:
+    """
+    marker = marker_atom (BOOLOP marker_atom)+
+    """
+    expression = [_parse_marker_atom(tokenizer)]
+    while tokenizer.check("BOOLOP"):
+        token = tokenizer.read()
+        expr_right = _parse_marker_atom(tokenizer)
+        expression.extend((token.text, expr_right))
+    return expression
+
+
+def _parse_marker_atom(tokenizer: Tokenizer) -> MarkerAtom:
+    """
+    marker_atom = WS? LEFT_PARENTHESIS WS? marker WS? RIGHT_PARENTHESIS WS?
+                | WS? marker_item WS?
+    """
+
+    tokenizer.consume("WS")
+    if tokenizer.check("LEFT_PARENTHESIS", peek=True):
+        with tokenizer.enclosing_tokens(
+            "LEFT_PARENTHESIS",
+            "RIGHT_PARENTHESIS",
+            around="marker expression",
+        ):
+            tokenizer.consume("WS")
+            marker: MarkerAtom = _parse_marker(tokenizer)
+            tokenizer.consume("WS")
+    else:
+        marker = _parse_marker_item(tokenizer)
+    tokenizer.consume("WS")
+    return marker
+
+
+def _parse_marker_item(tokenizer: Tokenizer) -> MarkerItem:
+    """
+    marker_item = WS? marker_var WS? marker_op WS? marker_var WS?
+    """
+    tokenizer.consume("WS")
+    marker_var_left = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    marker_op = _parse_marker_op(tokenizer)
+    tokenizer.consume("WS")
+    marker_var_right = _parse_marker_var(tokenizer)
+    tokenizer.consume("WS")
+    return (marker_var_left, marker_op, marker_var_right)
+
+
+def _parse_marker_var(tokenizer: Tokenizer) -> MarkerVar:
+    """
+    marker_var = VARIABLE | QUOTED_STRING
+    """
+    if tokenizer.check("VARIABLE"):
+        return process_env_var(tokenizer.read().text.replace(".", "_"))
+    elif tokenizer.check("QUOTED_STRING"):
+        return process_python_str(tokenizer.read().text)
+    else:
+        tokenizer.raise_syntax_error(
+            message="Expected a marker variable or quoted string"
+        )
+
+
+def process_env_var(env_var: str) -> Variable:
+    if env_var in ("platform_python_implementation", "python_implementation"):
+        return Variable("platform_python_implementation")
+    else:
+        return Variable(env_var)
+
+
+def process_python_str(python_str: str) -> Value:
+    value = ast.literal_eval(python_str)
+    return Value(str(value))
+
+
+def _parse_marker_op(tokenizer: Tokenizer) -> Op:
+    """
+    marker_op = IN | NOT IN | OP
+    """
+    if tokenizer.check("IN"):
+        tokenizer.read()
+        return Op("in")
+    elif tokenizer.check("NOT"):
+        tokenizer.read()
+        tokenizer.expect("WS", expected="whitespace after 'not'")
+        tokenizer.expect("IN", expected="'in' after 'not'")
+        return Op("not in")
+    elif tokenizer.check("OP"):
+        return Op(tokenizer.read().text)
+    else:
+        return tokenizer.raise_syntax_error(
+            "Expected marker operator, one of "
+            "<=, <, !=, ==, >=, >, ~=, ===, in, not in"
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_structures.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_structures.py
new file mode 100644
index 00000000..90a6465f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_structures.py
@@ -0,0 +1,61 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+
+class InfinityType:
+    def __repr__(self) -> str:
+        return "Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return False
+
+    def __le__(self, other: object) -> bool:
+        return False
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return True
+
+    def __ge__(self, other: object) -> bool:
+        return True
+
+    def __neg__(self: object) -> "NegativeInfinityType":
+        return NegativeInfinity
+
+
+Infinity = InfinityType()
+
+
+class NegativeInfinityType:
+    def __repr__(self) -> str:
+        return "-Infinity"
+
+    def __hash__(self) -> int:
+        return hash(repr(self))
+
+    def __lt__(self, other: object) -> bool:
+        return True
+
+    def __le__(self, other: object) -> bool:
+        return True
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__)
+
+    def __gt__(self, other: object) -> bool:
+        return False
+
+    def __ge__(self, other: object) -> bool:
+        return False
+
+    def __neg__(self: object) -> InfinityType:
+        return Infinity
+
+
+NegativeInfinity = NegativeInfinityType()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_tokenizer.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_tokenizer.py
new file mode 100644
index 00000000..dd0d648d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/_tokenizer.py
@@ -0,0 +1,192 @@
+import contextlib
+import re
+from dataclasses import dataclass
+from typing import Dict, Iterator, NoReturn, Optional, Tuple, Union
+
+from .specifiers import Specifier
+
+
+@dataclass
+class Token:
+    name: str
+    text: str
+    position: int
+
+
+class ParserSyntaxError(Exception):
+    """The provided source text could not be parsed correctly."""
+
+    def __init__(
+        self,
+        message: str,
+        *,
+        source: str,
+        span: Tuple[int, int],
+    ) -> None:
+        self.span = span
+        self.message = message
+        self.source = source
+
+        super().__init__()
+
+    def __str__(self) -> str:
+        marker = " " * self.span[0] + "~" * (self.span[1] - self.span[0]) + "^"
+        return "\n    ".join([self.message, self.source, marker])
+
+
+DEFAULT_RULES: "Dict[str, Union[str, re.Pattern[str]]]" = {
+    "LEFT_PARENTHESIS": r"\(",
+    "RIGHT_PARENTHESIS": r"\)",
+    "LEFT_BRACKET": r"\[",
+    "RIGHT_BRACKET": r"\]",
+    "SEMICOLON": r";",
+    "COMMA": r",",
+    "QUOTED_STRING": re.compile(
+        r"""
+            (
+                ('[^']*')
+                |
+                ("[^"]*")
+            )
+        """,
+        re.VERBOSE,
+    ),
+    "OP": r"(===|==|~=|!=|<=|>=|<|>)",
+    "BOOLOP": r"\b(or|and)\b",
+    "IN": r"\bin\b",
+    "NOT": r"\bnot\b",
+    "VARIABLE": re.compile(
+        r"""
+            \b(
+                python_version
+                |python_full_version
+                |os[._]name
+                |sys[._]platform
+                |platform_(release|system)
+                |platform[._](version|machine|python_implementation)
+                |python_implementation
+                |implementation_(name|version)
+                |extra
+            )\b
+        """,
+        re.VERBOSE,
+    ),
+    "SPECIFIER": re.compile(
+        Specifier._operator_regex_str + Specifier._version_regex_str,
+        re.VERBOSE | re.IGNORECASE,
+    ),
+    "AT": r"\@",
+    "URL": r"[^ \t]+",
+    "IDENTIFIER": r"\b[a-zA-Z0-9][a-zA-Z0-9._-]*\b",
+    "VERSION_PREFIX_TRAIL": r"\.\*",
+    "VERSION_LOCAL_LABEL_TRAIL": r"\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*",
+    "WS": r"[ \t]+",
+    "END": r"$",
+}
+
+
+class Tokenizer:
+    """Context-sensitive token parsing.
+
+    Provides methods to examine the input stream to check whether the next token
+    matches.
+    """
+
+    def __init__(
+        self,
+        source: str,
+        *,
+        rules: "Dict[str, Union[str, re.Pattern[str]]]",
+    ) -> None:
+        self.source = source
+        self.rules: Dict[str, re.Pattern[str]] = {
+            name: re.compile(pattern) for name, pattern in rules.items()
+        }
+        self.next_token: Optional[Token] = None
+        self.position = 0
+
+    def consume(self, name: str) -> None:
+        """Move beyond provided token name, if at current position."""
+        if self.check(name):
+            self.read()
+
+    def check(self, name: str, *, peek: bool = False) -> bool:
+        """Check whether the next token has the provided name.
+
+        By default, if the check succeeds, the token *must* be read before
+        another check. If `peek` is set to `True`, the token is not loaded and
+        would need to be checked again.
+        """
+        assert (
+            self.next_token is None
+        ), f"Cannot check for {name!r}, already have {self.next_token!r}"
+        assert name in self.rules, f"Unknown token name: {name!r}"
+
+        expression = self.rules[name]
+
+        match = expression.match(self.source, self.position)
+        if match is None:
+            return False
+        if not peek:
+            self.next_token = Token(name, match[0], self.position)
+        return True
+
+    def expect(self, name: str, *, expected: str) -> Token:
+        """Expect a certain token name next, failing with a syntax error otherwise.
+
+        The token is *not* read.
+        """
+        if not self.check(name):
+            raise self.raise_syntax_error(f"Expected {expected}")
+        return self.read()
+
+    def read(self) -> Token:
+        """Consume the next token and return it."""
+        token = self.next_token
+        assert token is not None
+
+        self.position += len(token.text)
+        self.next_token = None
+
+        return token
+
+    def raise_syntax_error(
+        self,
+        message: str,
+        *,
+        span_start: Optional[int] = None,
+        span_end: Optional[int] = None,
+    ) -> NoReturn:
+        """Raise ParserSyntaxError at the given position."""
+        span = (
+            self.position if span_start is None else span_start,
+            self.position if span_end is None else span_end,
+        )
+        raise ParserSyntaxError(
+            message,
+            source=self.source,
+            span=span,
+        )
+
+    @contextlib.contextmanager
+    def enclosing_tokens(
+        self, open_token: str, close_token: str, *, around: str
+    ) -> Iterator[None]:
+        if self.check(open_token):
+            open_position = self.position
+            self.read()
+        else:
+            open_position = None
+
+        yield
+
+        if open_position is None:
+            return
+
+        if not self.check(close_token):
+            self.raise_syntax_error(
+                f"Expected matching {close_token} for {open_token}, after {around}",
+                span_start=open_position,
+            )
+
+        self.read()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/markers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/markers.py
new file mode 100644
index 00000000..c96d22a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/markers.py
@@ -0,0 +1,253 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+import operator
+import os
+import platform
+import sys
+from typing import Any, Callable, Dict, List, Optional, Tuple, Union
+
+from ._parser import (
+    MarkerAtom,
+    MarkerList,
+    Op,
+    Value,
+    Variable,
+)
+from ._parser import (
+    parse_marker as _parse_marker,
+)
+from ._tokenizer import ParserSyntaxError
+from .specifiers import InvalidSpecifier, Specifier
+from .utils import canonicalize_name
+
+__all__ = [
+    "InvalidMarker",
+    "UndefinedComparison",
+    "UndefinedEnvironmentName",
+    "Marker",
+    "default_environment",
+]
+
+Operator = Callable[[str, str], bool]
+
+
+class InvalidMarker(ValueError):
+    """
+    An invalid marker was found, users should refer to PEP 508.
+    """
+
+
+class UndefinedComparison(ValueError):
+    """
+    An invalid operation was attempted on a value that doesn't support it.
+    """
+
+
+class UndefinedEnvironmentName(ValueError):
+    """
+    A name was attempted to be used that does not exist inside of the
+    environment.
+    """
+
+
+def _normalize_extra_values(results: Any) -> Any:
+    """
+    Normalize extra values.
+    """
+    if isinstance(results[0], tuple):
+        lhs, op, rhs = results[0]
+        if isinstance(lhs, Variable) and lhs.value == "extra":
+            normalized_extra = canonicalize_name(rhs.value)
+            rhs = Value(normalized_extra)
+        elif isinstance(rhs, Variable) and rhs.value == "extra":
+            normalized_extra = canonicalize_name(lhs.value)
+            lhs = Value(normalized_extra)
+        results[0] = lhs, op, rhs
+    return results
+
+
+def _format_marker(
+    marker: Union[List[str], MarkerAtom, str], first: Optional[bool] = True
+) -> str:
+    assert isinstance(marker, (list, tuple, str))
+
+    # Sometimes we have a structure like [[...]] which is a single item list
+    # where the single item is itself it's own list. In that case we want skip
+    # the rest of this function so that we don't get extraneous () on the
+    # outside.
+    if (
+        isinstance(marker, list)
+        and len(marker) == 1
+        and isinstance(marker[0], (list, tuple))
+    ):
+        return _format_marker(marker[0])
+
+    if isinstance(marker, list):
+        inner = (_format_marker(m, first=False) for m in marker)
+        if first:
+            return " ".join(inner)
+        else:
+            return "(" + " ".join(inner) + ")"
+    elif isinstance(marker, tuple):
+        return " ".join([m.serialize() for m in marker])
+    else:
+        return marker
+
+
+_operators: Dict[str, Operator] = {
+    "in": lambda lhs, rhs: lhs in rhs,
+    "not in": lambda lhs, rhs: lhs not in rhs,
+    "<": operator.lt,
+    "<=": operator.le,
+    "==": operator.eq,
+    "!=": operator.ne,
+    ">=": operator.ge,
+    ">": operator.gt,
+}
+
+
+def _eval_op(lhs: str, op: Op, rhs: str) -> bool:
+    try:
+        spec = Specifier("".join([op.serialize(), rhs]))
+    except InvalidSpecifier:
+        pass
+    else:
+        return spec.contains(lhs, prereleases=True)
+
+    oper: Optional[Operator] = _operators.get(op.serialize())
+    if oper is None:
+        raise UndefinedComparison(f"Undefined {op!r} on {lhs!r} and {rhs!r}.")
+
+    return oper(lhs, rhs)
+
+
+def _normalize(*values: str, key: str) -> Tuple[str, ...]:
+    # PEP 685 – Comparison of extra names for optional distribution dependencies
+    # https://peps.python.org/pep-0685/
+    # > When comparing extra names, tools MUST normalize the names being
+    # > compared using the semantics outlined in PEP 503 for names
+    if key == "extra":
+        return tuple(canonicalize_name(v) for v in values)
+
+    # other environment markers don't have such standards
+    return values
+
+
+def _evaluate_markers(markers: MarkerList, environment: Dict[str, str]) -> bool:
+    groups: List[List[bool]] = [[]]
+
+    for marker in markers:
+        assert isinstance(marker, (list, tuple, str))
+
+        if isinstance(marker, list):
+            groups[-1].append(_evaluate_markers(marker, environment))
+        elif isinstance(marker, tuple):
+            lhs, op, rhs = marker
+
+            if isinstance(lhs, Variable):
+                environment_key = lhs.value
+                lhs_value = environment[environment_key]
+                rhs_value = rhs.value
+            else:
+                lhs_value = lhs.value
+                environment_key = rhs.value
+                rhs_value = environment[environment_key]
+
+            lhs_value, rhs_value = _normalize(lhs_value, rhs_value, key=environment_key)
+            groups[-1].append(_eval_op(lhs_value, op, rhs_value))
+        else:
+            assert marker in ["and", "or"]
+            if marker == "or":
+                groups.append([])
+
+    return any(all(item) for item in groups)
+
+
+def format_full_version(info: "sys._version_info") -> str:
+    version = "{0.major}.{0.minor}.{0.micro}".format(info)
+    kind = info.releaselevel
+    if kind != "final":
+        version += kind[0] + str(info.serial)
+    return version
+
+
+def default_environment() -> Dict[str, str]:
+    iver = format_full_version(sys.implementation.version)
+    implementation_name = sys.implementation.name
+    return {
+        "implementation_name": implementation_name,
+        "implementation_version": iver,
+        "os_name": os.name,
+        "platform_machine": platform.machine(),
+        "platform_release": platform.release(),
+        "platform_system": platform.system(),
+        "platform_version": platform.version(),
+        "python_full_version": platform.python_version(),
+        "platform_python_implementation": platform.python_implementation(),
+        "python_version": ".".join(platform.python_version_tuple()[:2]),
+        "sys_platform": sys.platform,
+    }
+
+
+class Marker:
+    def __init__(self, marker: str) -> None:
+        # Note: We create a Marker object without calling this constructor in
+        #       packaging.requirements.Requirement. If any additional logic is
+        #       added here, make sure to mirror/adapt Requirement.
+        try:
+            self._markers = _normalize_extra_values(_parse_marker(marker))
+            # The attribute `_markers` can be described in terms of a recursive type:
+            # MarkerList = List[Union[Tuple[Node, ...], str, MarkerList]]
+            #
+            # For example, the following expression:
+            # python_version > "3.6" or (python_version == "3.6" and os_name == "unix")
+            #
+            # is parsed into:
+            # [
+            #     (<Variable('python_version')>, <Op('>')>, <Value('3.6')>),
+            #     'and',
+            #     [
+            #         (<Variable('python_version')>, <Op('==')>, <Value('3.6')>),
+            #         'or',
+            #         (<Variable('os_name')>, <Op('==')>, <Value('unix')>)
+            #     ]
+            # ]
+        except ParserSyntaxError as e:
+            raise InvalidMarker(str(e)) from e
+
+    def __str__(self) -> str:
+        return _format_marker(self._markers)
+
+    def __repr__(self) -> str:
+        return f"<Marker('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash((self.__class__.__name__, str(self)))
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Marker):
+            return NotImplemented
+
+        return str(self) == str(other)
+
+    def evaluate(self, environment: Optional[Dict[str, str]] = None) -> bool:
+        """Evaluate a marker.
+
+        Return the boolean from evaluating the given marker against the
+        environment. environment is an optional argument to override all or
+        part of the determined environment.
+
+        The environment is determined from the current Python process.
+        """
+        current_environment = default_environment()
+        current_environment["extra"] = ""
+        if environment is not None:
+            current_environment.update(environment)
+            # The API used to allow setting extra to None. We need to handle this
+            # case for backwards compatibility.
+            if current_environment["extra"] is None:
+                current_environment["extra"] = ""
+
+        return _evaluate_markers(self._markers, current_environment)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/requirements.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/requirements.py
new file mode 100644
index 00000000..bdc43a7e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/requirements.py
@@ -0,0 +1,90 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+from typing import Any, Iterator, Optional, Set
+
+from ._parser import parse_requirement as _parse_requirement
+from ._tokenizer import ParserSyntaxError
+from .markers import Marker, _normalize_extra_values
+from .specifiers import SpecifierSet
+from .utils import canonicalize_name
+
+
+class InvalidRequirement(ValueError):
+    """
+    An invalid requirement was found, users should refer to PEP 508.
+    """
+
+
+class Requirement:
+    """Parse a requirement.
+
+    Parse a given requirement string into its parts, such as name, specifier,
+    URL, and extras. Raises InvalidRequirement on a badly-formed requirement
+    string.
+    """
+
+    # TODO: Can we test whether something is contained within a requirement?
+    #       If so how do we do that? Do we need to test against the _name_ of
+    #       the thing as well as the version? What about the markers?
+    # TODO: Can we normalize the name and extra name?
+
+    def __init__(self, requirement_string: str) -> None:
+        try:
+            parsed = _parse_requirement(requirement_string)
+        except ParserSyntaxError as e:
+            raise InvalidRequirement(str(e)) from e
+
+        self.name: str = parsed.name
+        self.url: Optional[str] = parsed.url or None
+        self.extras: Set[str] = set(parsed.extras or [])
+        self.specifier: SpecifierSet = SpecifierSet(parsed.specifier)
+        self.marker: Optional[Marker] = None
+        if parsed.marker is not None:
+            self.marker = Marker.__new__(Marker)
+            self.marker._markers = _normalize_extra_values(parsed.marker)
+
+    def _iter_parts(self, name: str) -> Iterator[str]:
+        yield name
+
+        if self.extras:
+            formatted_extras = ",".join(sorted(self.extras))
+            yield f"[{formatted_extras}]"
+
+        if self.specifier:
+            yield str(self.specifier)
+
+        if self.url:
+            yield f"@ {self.url}"
+            if self.marker:
+                yield " "
+
+        if self.marker:
+            yield f"; {self.marker}"
+
+    def __str__(self) -> str:
+        return "".join(self._iter_parts(self.name))
+
+    def __repr__(self) -> str:
+        return f"<Requirement('{self}')>"
+
+    def __hash__(self) -> int:
+        return hash(
+            (
+                self.__class__.__name__,
+                *self._iter_parts(canonicalize_name(self.name)),
+            )
+        )
+
+    def __eq__(self, other: Any) -> bool:
+        if not isinstance(other, Requirement):
+            return NotImplemented
+
+        return (
+            canonicalize_name(self.name) == canonicalize_name(other.name)
+            and self.extras == other.extras
+            and self.specifier == other.specifier
+            and self.url == other.url
+            and self.marker == other.marker
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/specifiers.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/specifiers.py
new file mode 100644
index 00000000..6d4066ae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/specifiers.py
@@ -0,0 +1,1011 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.specifiers import Specifier, SpecifierSet, InvalidSpecifier
+    from packaging.version import Version
+"""
+
+import abc
+import itertools
+import re
+from typing import Callable, Iterable, Iterator, List, Optional, Tuple, TypeVar, Union
+
+from .utils import canonicalize_version
+from .version import Version
+
+UnparsedVersion = Union[Version, str]
+UnparsedVersionVar = TypeVar("UnparsedVersionVar", bound=UnparsedVersion)
+CallableOperator = Callable[[Version, str], bool]
+
+
+def _coerce_version(version: UnparsedVersion) -> Version:
+    if not isinstance(version, Version):
+        version = Version(version)
+    return version
+
+
+class InvalidSpecifier(ValueError):
+    """
+    Raised when attempting to create a :class:`Specifier` with a specifier
+    string that is invalid.
+
+    >>> Specifier("lolwat")
+    Traceback (most recent call last):
+        ...
+    packaging.specifiers.InvalidSpecifier: Invalid specifier: 'lolwat'
+    """
+
+
+class BaseSpecifier(metaclass=abc.ABCMeta):
+    @abc.abstractmethod
+    def __str__(self) -> str:
+        """
+        Returns the str representation of this Specifier-like object. This
+        should be representative of the Specifier itself.
+        """
+
+    @abc.abstractmethod
+    def __hash__(self) -> int:
+        """
+        Returns a hash value for this Specifier-like object.
+        """
+
+    @abc.abstractmethod
+    def __eq__(self, other: object) -> bool:
+        """
+        Returns a boolean representing whether or not the two Specifier-like
+        objects are equal.
+
+        :param other: The other object to check against.
+        """
+
+    @property
+    @abc.abstractmethod
+    def prereleases(self) -> Optional[bool]:
+        """Whether or not pre-releases as a whole are allowed.
+
+        This can be set to either ``True`` or ``False`` to explicitly enable or disable
+        prereleases or it can be set to ``None`` (the default) to use default semantics.
+        """
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        """Setter for :attr:`prereleases`.
+
+        :param value: The value to set.
+        """
+
+    @abc.abstractmethod
+    def contains(self, item: str, prereleases: Optional[bool] = None) -> bool:
+        """
+        Determines if the given item is contained within this specifier.
+        """
+
+    @abc.abstractmethod
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: Optional[bool] = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """
+        Takes an iterable of items and filters them so that only items which
+        are contained within this specifier are allowed in it.
+        """
+
+
+class Specifier(BaseSpecifier):
+    """This class abstracts handling of version specifiers.
+
+    .. tip::
+
+        It is generally not required to instantiate this manually. You should instead
+        prefer to work with :class:`SpecifierSet` instead, which can parse
+        comma-separated version specifiers (which is what package metadata contains).
+    """
+
+    _operator_regex_str = r"""
+        (?P<operator>(~=|==|!=|<=|>=|<|>|===))
+        """
+    _version_regex_str = r"""
+        (?P<version>
+            (?:
+                # The identity operators allow for an escape hatch that will
+                # do an exact string match of the version you wish to install.
+                # This will not be parsed by PEP 440 and we cannot determine
+                # any semantic meaning from it. This operator is discouraged
+                # but included entirely as an escape hatch.
+                (?<====)  # Only match for the identity operator
+                \s*
+                [^\s;)]*  # The arbitrary version can be just about anything,
+                          # we match everything except for whitespace, a
+                          # semi-colon for marker support, and a closing paren
+                          # since versions can be enclosed in them.
+            )
+            |
+            (?:
+                # The (non)equality operators allow for wild card and local
+                # versions to be specified so we have to define these two
+                # operators separately to enable that.
+                (?<===|!=)            # Only match for equals and not equals
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+
+                # You cannot use a wild card and a pre-release, post-release, a dev or
+                # local version together so group them with a | and make them optional.
+                (?:
+                    \.\*  # Wild card syntax of .*
+                    |
+                    (?:                                  # pre release
+                        [-_\.]?
+                        (alpha|beta|preview|pre|a|b|c|rc)
+                        [-_\.]?
+                        [0-9]*
+                    )?
+                    (?:                                  # post release
+                        (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                    )?
+                    (?:[-_\.]?dev[-_\.]?[0-9]*)?         # dev release
+                    (?:\+[a-z0-9]+(?:[-_\.][a-z0-9]+)*)? # local
+                )?
+            )
+            |
+            (?:
+                # The compatible operator requires at least two digits in the
+                # release segment.
+                (?<=~=)               # Only match for the compatible operator
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)+   # release  (We have a + instead of a *)
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+            |
+            (?:
+                # All other operators only allow a sub set of what the
+                # (non)equality operators do. Specifically they do not allow
+                # local versions to be specified nor do they allow the prefix
+                # matching wild cards.
+                (?<!==|!=|~=)         # We have special cases for these
+                                      # operators so we want to make sure they
+                                      # don't match here.
+
+                \s*
+                v?
+                (?:[0-9]+!)?          # epoch
+                [0-9]+(?:\.[0-9]+)*   # release
+                (?:                   # pre release
+                    [-_\.]?
+                    (alpha|beta|preview|pre|a|b|c|rc)
+                    [-_\.]?
+                    [0-9]*
+                )?
+                (?:                                   # post release
+                    (?:-[0-9]+)|(?:[-_\.]?(post|rev|r)[-_\.]?[0-9]*)
+                )?
+                (?:[-_\.]?dev[-_\.]?[0-9]*)?          # dev release
+            )
+        )
+        """
+
+    _regex = re.compile(
+        r"^\s*" + _operator_regex_str + _version_regex_str + r"\s*$",
+        re.VERBOSE | re.IGNORECASE,
+    )
+
+    _operators = {
+        "~=": "compatible",
+        "==": "equal",
+        "!=": "not_equal",
+        "<=": "less_than_equal",
+        ">=": "greater_than_equal",
+        "<": "less_than",
+        ">": "greater_than",
+        "===": "arbitrary",
+    }
+
+    def __init__(self, spec: str = "", prereleases: Optional[bool] = None) -> None:
+        """Initialize a Specifier instance.
+
+        :param spec:
+            The string representation of a specifier which will be parsed and
+            normalized before use.
+        :param prereleases:
+            This tells the specifier if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+        :raises InvalidSpecifier:
+            If the given specifier is invalid (i.e. bad syntax).
+        """
+        match = self._regex.search(spec)
+        if not match:
+            raise InvalidSpecifier(f"Invalid specifier: '{spec}'")
+
+        self._spec: Tuple[str, str] = (
+            match.group("operator").strip(),
+            match.group("version").strip(),
+        )
+
+        # Store whether or not this Specifier should accept prereleases
+        self._prereleases = prereleases
+
+    # https://github.com/python/mypy/pull/13475#pullrequestreview-1079784515
+    @property  # type: ignore[override]
+    def prereleases(self) -> bool:
+        # If there is an explicit prereleases set for this, then we'll just
+        # blindly use that.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # Look at all of our specifiers and determine if they are inclusive
+        # operators, and if they are if they are including an explicit
+        # prerelease.
+        operator, version = self._spec
+        if operator in ["==", ">=", "<=", "~=", "==="]:
+            # The == specifier can include a trailing .*, if it does we
+            # want to remove before parsing.
+            if operator == "==" and version.endswith(".*"):
+                version = version[:-2]
+
+            # Parse the version, and if it is a pre-release than this
+            # specifier allows pre-releases.
+            if Version(version).is_prerelease:
+                return True
+
+        return False
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    @property
+    def operator(self) -> str:
+        """The operator of this specifier.
+
+        >>> Specifier("==1.2.3").operator
+        '=='
+        """
+        return self._spec[0]
+
+    @property
+    def version(self) -> str:
+        """The version of this specifier.
+
+        >>> Specifier("==1.2.3").version
+        '1.2.3'
+        """
+        return self._spec[1]
+
+    def __repr__(self) -> str:
+        """A representation of the Specifier that shows all internal state.
+
+        >>> Specifier('>=1.0.0')
+        <Specifier('>=1.0.0')>
+        >>> Specifier('>=1.0.0', prereleases=False)
+        <Specifier('>=1.0.0', prereleases=False)>
+        >>> Specifier('>=1.0.0', prereleases=True)
+        <Specifier('>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<{self.__class__.__name__}({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the Specifier that can be round-tripped.
+
+        >>> str(Specifier('>=1.0.0'))
+        '>=1.0.0'
+        >>> str(Specifier('>=1.0.0', prereleases=False))
+        '>=1.0.0'
+        """
+        return "{}{}".format(*self._spec)
+
+    @property
+    def _canonical_spec(self) -> Tuple[str, str]:
+        canonical_version = canonicalize_version(
+            self._spec[1],
+            strip_trailing_zero=(self._spec[0] != "~="),
+        )
+        return self._spec[0], canonical_version
+
+    def __hash__(self) -> int:
+        return hash(self._canonical_spec)
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two Specifier-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> Specifier("==1.2.3") == Specifier("== 1.2.3.0")
+        True
+        >>> (Specifier("==1.2.3", prereleases=False) ==
+        ...  Specifier("==1.2.3", prereleases=True))
+        True
+        >>> Specifier("==1.2.3") == "==1.2.3"
+        True
+        >>> Specifier("==1.2.3") == Specifier("==1.2.4")
+        False
+        >>> Specifier("==1.2.3") == Specifier("~=1.2.3")
+        False
+        """
+        if isinstance(other, str):
+            try:
+                other = self.__class__(str(other))
+            except InvalidSpecifier:
+                return NotImplemented
+        elif not isinstance(other, self.__class__):
+            return NotImplemented
+
+        return self._canonical_spec == other._canonical_spec
+
+    def _get_operator(self, op: str) -> CallableOperator:
+        operator_callable: CallableOperator = getattr(
+            self, f"_compare_{self._operators[op]}"
+        )
+        return operator_callable
+
+    def _compare_compatible(self, prospective: Version, spec: str) -> bool:
+        # Compatible releases have an equivalent combination of >= and ==. That
+        # is that ~=2.2 is equivalent to >=2.2,==2.*. This allows us to
+        # implement this in terms of the other specifiers instead of
+        # implementing it ourselves. The only thing we need to do is construct
+        # the other specifiers.
+
+        # We want everything but the last item in the version, but we want to
+        # ignore suffix segments.
+        prefix = _version_join(
+            list(itertools.takewhile(_is_not_suffix, _version_split(spec)))[:-1]
+        )
+
+        # Add the prefix notation to the end of our string
+        prefix += ".*"
+
+        return self._get_operator(">=")(prospective, spec) and self._get_operator("==")(
+            prospective, prefix
+        )
+
+    def _compare_equal(self, prospective: Version, spec: str) -> bool:
+        # We need special logic to handle prefix matching
+        if spec.endswith(".*"):
+            # In the case of prefix matching we want to ignore local segment.
+            normalized_prospective = canonicalize_version(
+                prospective.public, strip_trailing_zero=False
+            )
+            # Get the normalized version string ignoring the trailing .*
+            normalized_spec = canonicalize_version(spec[:-2], strip_trailing_zero=False)
+            # Split the spec out by bangs and dots, and pretend that there is
+            # an implicit dot in between a release segment and a pre-release segment.
+            split_spec = _version_split(normalized_spec)
+
+            # Split the prospective version out by bangs and dots, and pretend
+            # that there is an implicit dot in between a release segment and
+            # a pre-release segment.
+            split_prospective = _version_split(normalized_prospective)
+
+            # 0-pad the prospective version before shortening it to get the correct
+            # shortened version.
+            padded_prospective, _ = _pad_version(split_prospective, split_spec)
+
+            # Shorten the prospective version to be the same length as the spec
+            # so that we can determine if the specifier is a prefix of the
+            # prospective version or not.
+            shortened_prospective = padded_prospective[: len(split_spec)]
+
+            return shortened_prospective == split_spec
+        else:
+            # Convert our spec string into a Version
+            spec_version = Version(spec)
+
+            # If the specifier does not have a local segment, then we want to
+            # act as if the prospective version also does not have a local
+            # segment.
+            if not spec_version.local:
+                prospective = Version(prospective.public)
+
+            return prospective == spec_version
+
+    def _compare_not_equal(self, prospective: Version, spec: str) -> bool:
+        return not self._compare_equal(prospective, spec)
+
+    def _compare_less_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) <= Version(spec)
+
+    def _compare_greater_than_equal(self, prospective: Version, spec: str) -> bool:
+        # NB: Local version identifiers are NOT permitted in the version
+        # specifier, so local version labels can be universally removed from
+        # the prospective version.
+        return Version(prospective.public) >= Version(spec)
+
+    def _compare_less_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is less than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective < spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a pre-release version, that we do not accept pre-release
+        # versions for the version mentioned in the specifier (e.g. <3.1 should
+        # not match 3.1.dev0, but should match 3.0.dev0).
+        if not spec.is_prerelease and prospective.is_prerelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # less than the spec version *and* it's not a pre-release of the same
+        # version in the spec.
+        return True
+
+    def _compare_greater_than(self, prospective: Version, spec_str: str) -> bool:
+        # Convert our spec to a Version instance, since we'll want to work with
+        # it as a version.
+        spec = Version(spec_str)
+
+        # Check to see if the prospective version is greater than the spec
+        # version. If it's not we can short circuit and just return False now
+        # instead of doing extra unneeded work.
+        if not prospective > spec:
+            return False
+
+        # This special case is here so that, unless the specifier itself
+        # includes is a post-release version, that we do not accept
+        # post-release versions for the version mentioned in the specifier
+        # (e.g. >3.1 should not match 3.0.post0, but should match 3.2.post0).
+        if not spec.is_postrelease and prospective.is_postrelease:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # Ensure that we do not allow a local version of the version mentioned
+        # in the specifier, which is technically greater than, to match.
+        if prospective.local is not None:
+            if Version(prospective.base_version) == Version(spec.base_version):
+                return False
+
+        # If we've gotten to here, it means that prospective version is both
+        # greater than the spec version *and* it's not a pre-release of the
+        # same version in the spec.
+        return True
+
+    def _compare_arbitrary(self, prospective: Version, spec: str) -> bool:
+        return str(prospective).lower() == str(spec).lower()
+
+    def __contains__(self, item: Union[str, Version]) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in Specifier(">=1.2.3")
+        True
+        >>> Version("1.2.3") in Specifier(">=1.2.3")
+        True
+        >>> "1.0.0" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3")
+        False
+        >>> "1.3.0a1" in Specifier(">=1.2.3", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(
+        self, item: UnparsedVersion, prereleases: Optional[bool] = None
+    ) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this Specifier. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> Specifier(">=1.2.3").contains("1.2.3")
+        True
+        >>> Specifier(">=1.2.3").contains(Version("1.2.3"))
+        True
+        >>> Specifier(">=1.2.3").contains("1.0.0")
+        False
+        >>> Specifier(">=1.2.3").contains("1.3.0a1")
+        False
+        >>> Specifier(">=1.2.3", prereleases=True).contains("1.3.0a1")
+        True
+        >>> Specifier(">=1.2.3").contains("1.3.0a1", prereleases=True)
+        True
+        """
+
+        # Determine if prereleases are to be allowed or not.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # Normalize item to a Version, this allows us to have a shortcut for
+        # "2.0" in Specifier(">=2")
+        normalized_item = _coerce_version(item)
+
+        # Determine if we should be supporting prereleases in this specifier
+        # or not, if we do not support prereleases than we can short circuit
+        # logic if this version is a prereleases.
+        if normalized_item.is_prerelease and not prereleases:
+            return False
+
+        # Actually do the comparison to determine if this item is contained
+        # within this Specifier or not.
+        operator_callable: CallableOperator = self._get_operator(self.operator)
+        return operator_callable(normalized_item, self.version)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: Optional[bool] = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifier.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(Specifier().contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.2.3", "1.3", Version("1.4")]))
+        ['1.2.3', '1.3', <Version('1.4')>]
+        >>> list(Specifier(">=1.2.3").filter(["1.2", "1.5a1"]))
+        ['1.5a1']
+        >>> list(Specifier(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(Specifier(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        """
+
+        yielded = False
+        found_prereleases = []
+
+        kw = {"prereleases": prereleases if prereleases is not None else True}
+
+        # Attempt to iterate over all the values in the iterable and if any of
+        # them match, yield them.
+        for version in iterable:
+            parsed_version = _coerce_version(version)
+
+            if self.contains(parsed_version, **kw):
+                # If our version is a prerelease, and we were not set to allow
+                # prereleases, then we'll store it for later in case nothing
+                # else matches this specifier.
+                if parsed_version.is_prerelease and not (
+                    prereleases or self.prereleases
+                ):
+                    found_prereleases.append(version)
+                # Either this is not a prerelease, or we should have been
+                # accepting prereleases from the beginning.
+                else:
+                    yielded = True
+                    yield version
+
+        # Now that we've iterated over everything, determine if we've yielded
+        # any values, and if we have not and we have any prereleases stored up
+        # then we will go ahead and yield the prereleases.
+        if not yielded and found_prereleases:
+            for version in found_prereleases:
+                yield version
+
+
+_prefix_regex = re.compile(r"^([0-9]+)((?:a|b|c|rc)[0-9]+)$")
+
+
+def _version_split(version: str) -> List[str]:
+    """Split version into components.
+
+    The split components are intended for version comparison. The logic does
+    not attempt to retain the original version string, so joining the
+    components back with :func:`_version_join` may not produce the original
+    version string.
+    """
+    result: List[str] = []
+
+    epoch, _, rest = version.rpartition("!")
+    result.append(epoch or "0")
+
+    for item in rest.split("."):
+        match = _prefix_regex.search(item)
+        if match:
+            result.extend(match.groups())
+        else:
+            result.append(item)
+    return result
+
+
+def _version_join(components: List[str]) -> str:
+    """Join split version components into a version string.
+
+    This function assumes the input came from :func:`_version_split`, where the
+    first component must be the epoch (either empty or numeric), and all other
+    components numeric.
+    """
+    epoch, *rest = components
+    return f"{epoch}!{'.'.join(rest)}"
+
+
+def _is_not_suffix(segment: str) -> bool:
+    return not any(
+        segment.startswith(prefix) for prefix in ("dev", "a", "b", "rc", "post")
+    )
+
+
+def _pad_version(left: List[str], right: List[str]) -> Tuple[List[str], List[str]]:
+    left_split, right_split = [], []
+
+    # Get the release segment of our versions
+    left_split.append(list(itertools.takewhile(lambda x: x.isdigit(), left)))
+    right_split.append(list(itertools.takewhile(lambda x: x.isdigit(), right)))
+
+    # Get the rest of our versions
+    left_split.append(left[len(left_split[0]) :])
+    right_split.append(right[len(right_split[0]) :])
+
+    # Insert our padding
+    left_split.insert(1, ["0"] * max(0, len(right_split[0]) - len(left_split[0])))
+    right_split.insert(1, ["0"] * max(0, len(left_split[0]) - len(right_split[0])))
+
+    return (
+        list(itertools.chain.from_iterable(left_split)),
+        list(itertools.chain.from_iterable(right_split)),
+    )
+
+
+class SpecifierSet(BaseSpecifier):
+    """This class abstracts handling of a set of version specifiers.
+
+    It can be passed a single specifier (``>=3.0``), a comma-separated list of
+    specifiers (``>=3.0,!=3.1``), or no specifier at all.
+    """
+
+    def __init__(
+        self, specifiers: str = "", prereleases: Optional[bool] = None
+    ) -> None:
+        """Initialize a SpecifierSet instance.
+
+        :param specifiers:
+            The string representation of a specifier or a comma-separated list of
+            specifiers which will be parsed and normalized before use.
+        :param prereleases:
+            This tells the SpecifierSet if it should accept prerelease versions if
+            applicable or not. The default of ``None`` will autodetect it from the
+            given specifiers.
+
+        :raises InvalidSpecifier:
+            If the given ``specifiers`` are not parseable than this exception will be
+            raised.
+        """
+
+        # Split on `,` to break each individual specifier into it's own item, and
+        # strip each item to remove leading/trailing whitespace.
+        split_specifiers = [s.strip() for s in specifiers.split(",") if s.strip()]
+
+        # Make each individual specifier a Specifier and save in a frozen set for later.
+        self._specs = frozenset(map(Specifier, split_specifiers))
+
+        # Store our prereleases value so we can use it later to determine if
+        # we accept prereleases or not.
+        self._prereleases = prereleases
+
+    @property
+    def prereleases(self) -> Optional[bool]:
+        # If we have been given an explicit prerelease modifier, then we'll
+        # pass that through here.
+        if self._prereleases is not None:
+            return self._prereleases
+
+        # If we don't have any specifiers, and we don't have a forced value,
+        # then we'll just return None since we don't know if this should have
+        # pre-releases or not.
+        if not self._specs:
+            return None
+
+        # Otherwise we'll see if any of the given specifiers accept
+        # prereleases, if any of them do we'll return True, otherwise False.
+        return any(s.prereleases for s in self._specs)
+
+    @prereleases.setter
+    def prereleases(self, value: bool) -> None:
+        self._prereleases = value
+
+    def __repr__(self) -> str:
+        """A representation of the specifier set that shows all internal state.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> SpecifierSet('>=1.0.0,!=2.0.0')
+        <SpecifierSet('!=2.0.0,>=1.0.0')>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=False)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=False)>
+        >>> SpecifierSet('>=1.0.0,!=2.0.0', prereleases=True)
+        <SpecifierSet('!=2.0.0,>=1.0.0', prereleases=True)>
+        """
+        pre = (
+            f", prereleases={self.prereleases!r}"
+            if self._prereleases is not None
+            else ""
+        )
+
+        return f"<SpecifierSet({str(self)!r}{pre})>"
+
+    def __str__(self) -> str:
+        """A string representation of the specifier set that can be round-tripped.
+
+        Note that the ordering of the individual specifiers within the set may not
+        match the input string.
+
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1"))
+        '!=1.0.1,>=1.0.0'
+        >>> str(SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False))
+        '!=1.0.1,>=1.0.0'
+        """
+        return ",".join(sorted(str(s) for s in self._specs))
+
+    def __hash__(self) -> int:
+        return hash(self._specs)
+
+    def __and__(self, other: Union["SpecifierSet", str]) -> "SpecifierSet":
+        """Return a SpecifierSet which is a combination of the two sets.
+
+        :param other: The other object to combine with.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & '<=2.0.0,!=2.0.1'
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") & SpecifierSet('<=2.0.0,!=2.0.1')
+        <SpecifierSet('!=1.0.1,!=2.0.1,<=2.0.0,>=1.0.0')>
+        """
+        if isinstance(other, str):
+            other = SpecifierSet(other)
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        specifier = SpecifierSet()
+        specifier._specs = frozenset(self._specs | other._specs)
+
+        if self._prereleases is None and other._prereleases is not None:
+            specifier._prereleases = other._prereleases
+        elif self._prereleases is not None and other._prereleases is None:
+            specifier._prereleases = self._prereleases
+        elif self._prereleases == other._prereleases:
+            specifier._prereleases = self._prereleases
+        else:
+            raise ValueError(
+                "Cannot combine SpecifierSets with True and False prerelease "
+                "overrides."
+            )
+
+        return specifier
+
+    def __eq__(self, other: object) -> bool:
+        """Whether or not the two SpecifierSet-like objects are equal.
+
+        :param other: The other object to check against.
+
+        The value of :attr:`prereleases` is ignored.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> (SpecifierSet(">=1.0.0,!=1.0.1", prereleases=False) ==
+        ...  SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == ">=1.0.0,!=1.0.1"
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1") == SpecifierSet(">=1.0.0,!=1.0.2")
+        False
+        """
+        if isinstance(other, (str, Specifier)):
+            other = SpecifierSet(str(other))
+        elif not isinstance(other, SpecifierSet):
+            return NotImplemented
+
+        return self._specs == other._specs
+
+    def __len__(self) -> int:
+        """Returns the number of specifiers in this specifier set."""
+        return len(self._specs)
+
+    def __iter__(self) -> Iterator[Specifier]:
+        """
+        Returns an iterator over all the underlying :class:`Specifier` instances
+        in this specifier set.
+
+        >>> sorted(SpecifierSet(">=1.0.0,!=1.0.1"), key=str)
+        [<Specifier('!=1.0.1')>, <Specifier('>=1.0.0')>]
+        """
+        return iter(self._specs)
+
+    def __contains__(self, item: UnparsedVersion) -> bool:
+        """Return whether or not the item is contained in this specifier.
+
+        :param item: The item to check for.
+
+        This is used for the ``in`` operator and behaves the same as
+        :meth:`contains` with no ``prereleases`` argument passed.
+
+        >>> "1.2.3" in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> Version("1.2.3") in SpecifierSet(">=1.0.0,!=1.0.1")
+        True
+        >>> "1.0.1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1")
+        False
+        >>> "1.3.0a1" in SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True)
+        True
+        """
+        return self.contains(item)
+
+    def contains(
+        self,
+        item: UnparsedVersion,
+        prereleases: Optional[bool] = None,
+        installed: Optional[bool] = None,
+    ) -> bool:
+        """Return whether or not the item is contained in this SpecifierSet.
+
+        :param item:
+            The item to check for, which can be a version string or a
+            :class:`Version` instance.
+        :param prereleases:
+            Whether or not to match prereleases with this SpecifierSet. If set to
+            ``None`` (the default), it uses :attr:`prereleases` to determine
+            whether or not prereleases are allowed.
+
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.2.3")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains(Version("1.2.3"))
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.0.1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1")
+        False
+        >>> SpecifierSet(">=1.0.0,!=1.0.1", prereleases=True).contains("1.3.0a1")
+        True
+        >>> SpecifierSet(">=1.0.0,!=1.0.1").contains("1.3.0a1", prereleases=True)
+        True
+        """
+        # Ensure that our item is a Version instance.
+        if not isinstance(item, Version):
+            item = Version(item)
+
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # We can determine if we're going to allow pre-releases by looking to
+        # see if any of the underlying items supports them. If none of them do
+        # and this item is a pre-release then we do not allow it and we can
+        # short circuit that here.
+        # Note: This means that 1.0.dev1 would not be contained in something
+        #       like >=1.0.devabc however it would be in >=1.0.debabc,>0.0.dev0
+        if not prereleases and item.is_prerelease:
+            return False
+
+        if installed and item.is_prerelease:
+            item = Version(item.base_version)
+
+        # We simply dispatch to the underlying specs here to make sure that the
+        # given version is contained within all of them.
+        # Note: This use of all() here means that an empty set of specifiers
+        #       will always return True, this is an explicit design decision.
+        return all(s.contains(item, prereleases=prereleases) for s in self._specs)
+
+    def filter(
+        self, iterable: Iterable[UnparsedVersionVar], prereleases: Optional[bool] = None
+    ) -> Iterator[UnparsedVersionVar]:
+        """Filter items in the given iterable, that match the specifiers in this set.
+
+        :param iterable:
+            An iterable that can contain version strings and :class:`Version` instances.
+            The items in the iterable will be filtered according to the specifier.
+        :param prereleases:
+            Whether or not to allow prereleases in the returned iterator. If set to
+            ``None`` (the default), it will be intelligently decide whether to allow
+            prereleases or not (based on the :attr:`prereleases` attribute, and
+            whether the only versions matching are prereleases).
+
+        This method is smarter than just ``filter(SpecifierSet(...).contains, [...])``
+        because it implements the rule from :pep:`440` that a prerelease item
+        SHOULD be accepted if no other versions match the given specifier.
+
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.3", Version("1.4")]))
+        ['1.3', <Version('1.4')>]
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.2", "1.5a1"]))
+        []
+        >>> list(SpecifierSet(">=1.2.3").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet(">=1.2.3", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+
+        An "empty" SpecifierSet will filter items based on the presence of prerelease
+        versions in the set.
+
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"]))
+        ['1.3']
+        >>> list(SpecifierSet("").filter(["1.5a1"]))
+        ['1.5a1']
+        >>> list(SpecifierSet("", prereleases=True).filter(["1.3", "1.5a1"]))
+        ['1.3', '1.5a1']
+        >>> list(SpecifierSet("").filter(["1.3", "1.5a1"], prereleases=True))
+        ['1.3', '1.5a1']
+        """
+        # Determine if we're forcing a prerelease or not, if we're not forcing
+        # one for this particular filter call, then we'll use whatever the
+        # SpecifierSet thinks for whether or not we should support prereleases.
+        if prereleases is None:
+            prereleases = self.prereleases
+
+        # If we have any specifiers, then we want to wrap our iterable in the
+        # filter method for each one, this will act as a logical AND amongst
+        # each specifier.
+        if self._specs:
+            for spec in self._specs:
+                iterable = spec.filter(iterable, prereleases=bool(prereleases))
+            return iter(iterable)
+        # If we do not have any specifiers, then we need to have a rough filter
+        # which will filter out any pre-releases, unless there are no final
+        # releases.
+        else:
+            filtered: List[UnparsedVersionVar] = []
+            found_prereleases: List[UnparsedVersionVar] = []
+
+            for item in iterable:
+                parsed_version = _coerce_version(item)
+
+                # Store any item which is a pre-release for later unless we've
+                # already found a final version or we are accepting prereleases
+                if parsed_version.is_prerelease and not prereleases:
+                    if not filtered:
+                        found_prereleases.append(item)
+                else:
+                    filtered.append(item)
+
+            # If we've found no items except for pre-releases, then we'll go
+            # ahead and use the pre-releases
+            if not filtered and found_prereleases and prereleases is None:
+                return iter(found_prereleases)
+
+            return iter(filtered)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/tags.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/tags.py
new file mode 100644
index 00000000..89f19261
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/tags.py
@@ -0,0 +1,571 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+import logging
+import platform
+import re
+import struct
+import subprocess
+import sys
+import sysconfig
+from importlib.machinery import EXTENSION_SUFFIXES
+from typing import (
+    Dict,
+    FrozenSet,
+    Iterable,
+    Iterator,
+    List,
+    Optional,
+    Sequence,
+    Tuple,
+    Union,
+    cast,
+)
+
+from . import _manylinux, _musllinux
+
+logger = logging.getLogger(__name__)
+
+PythonVersion = Sequence[int]
+MacVersion = Tuple[int, int]
+
+INTERPRETER_SHORT_NAMES: Dict[str, str] = {
+    "python": "py",  # Generic.
+    "cpython": "cp",
+    "pypy": "pp",
+    "ironpython": "ip",
+    "jython": "jy",
+}
+
+
+_32_BIT_INTERPRETER = struct.calcsize("P") == 4
+
+
+class Tag:
+    """
+    A representation of the tag triple for a wheel.
+
+    Instances are considered immutable and thus are hashable. Equality checking
+    is also supported.
+    """
+
+    __slots__ = ["_interpreter", "_abi", "_platform", "_hash"]
+
+    def __init__(self, interpreter: str, abi: str, platform: str) -> None:
+        self._interpreter = interpreter.lower()
+        self._abi = abi.lower()
+        self._platform = platform.lower()
+        # The __hash__ of every single element in a Set[Tag] will be evaluated each time
+        # that a set calls its `.disjoint()` method, which may be called hundreds of
+        # times when scanning a page of links for packages with tags matching that
+        # Set[Tag]. Pre-computing the value here produces significant speedups for
+        # downstream consumers.
+        self._hash = hash((self._interpreter, self._abi, self._platform))
+
+    @property
+    def interpreter(self) -> str:
+        return self._interpreter
+
+    @property
+    def abi(self) -> str:
+        return self._abi
+
+    @property
+    def platform(self) -> str:
+        return self._platform
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Tag):
+            return NotImplemented
+
+        return (
+            (self._hash == other._hash)  # Short-circuit ASAP for perf reasons.
+            and (self._platform == other._platform)
+            and (self._abi == other._abi)
+            and (self._interpreter == other._interpreter)
+        )
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __str__(self) -> str:
+        return f"{self._interpreter}-{self._abi}-{self._platform}"
+
+    def __repr__(self) -> str:
+        return f"<{self} @ {id(self)}>"
+
+
+def parse_tag(tag: str) -> FrozenSet[Tag]:
+    """
+    Parses the provided tag (e.g. `py3-none-any`) into a frozenset of Tag instances.
+
+    Returning a set is required due to the possibility that the tag is a
+    compressed tag set.
+    """
+    tags = set()
+    interpreters, abis, platforms = tag.split("-")
+    for interpreter in interpreters.split("."):
+        for abi in abis.split("."):
+            for platform_ in platforms.split("."):
+                tags.add(Tag(interpreter, abi, platform_))
+    return frozenset(tags)
+
+
+def _get_config_var(name: str, warn: bool = False) -> Union[int, str, None]:
+    value: Union[int, str, None] = sysconfig.get_config_var(name)
+    if value is None and warn:
+        logger.debug(
+            "Config variable '%s' is unset, Python ABI tag may be incorrect", name
+        )
+    return value
+
+
+def _normalize_string(string: str) -> str:
+    return string.replace(".", "_").replace("-", "_").replace(" ", "_")
+
+
+def _is_threaded_cpython(abis: List[str]) -> bool:
+    """
+    Determine if the ABI corresponds to a threaded (`--disable-gil`) build.
+
+    The threaded builds are indicated by a "t" in the abiflags.
+    """
+    if len(abis) == 0:
+        return False
+    # expect e.g., cp313
+    m = re.match(r"cp\d+(.*)", abis[0])
+    if not m:
+        return False
+    abiflags = m.group(1)
+    return "t" in abiflags
+
+
+def _abi3_applies(python_version: PythonVersion, threading: bool) -> bool:
+    """
+    Determine if the Python version supports abi3.
+
+    PEP 384 was first implemented in Python 3.2. The threaded (`--disable-gil`)
+    builds do not support abi3.
+    """
+    return len(python_version) > 1 and tuple(python_version) >= (3, 2) and not threading
+
+
+def _cpython_abis(py_version: PythonVersion, warn: bool = False) -> List[str]:
+    py_version = tuple(py_version)  # To allow for version comparison.
+    abis = []
+    version = _version_nodot(py_version[:2])
+    threading = debug = pymalloc = ucs4 = ""
+    with_debug = _get_config_var("Py_DEBUG", warn)
+    has_refcount = hasattr(sys, "gettotalrefcount")
+    # Windows doesn't set Py_DEBUG, so checking for support of debug-compiled
+    # extension modules is the best option.
+    # https://github.com/pypa/pip/issues/3383#issuecomment-173267692
+    has_ext = "_d.pyd" in EXTENSION_SUFFIXES
+    if with_debug or (with_debug is None and (has_refcount or has_ext)):
+        debug = "d"
+    if py_version >= (3, 13) and _get_config_var("Py_GIL_DISABLED", warn):
+        threading = "t"
+    if py_version < (3, 8):
+        with_pymalloc = _get_config_var("WITH_PYMALLOC", warn)
+        if with_pymalloc or with_pymalloc is None:
+            pymalloc = "m"
+        if py_version < (3, 3):
+            unicode_size = _get_config_var("Py_UNICODE_SIZE", warn)
+            if unicode_size == 4 or (
+                unicode_size is None and sys.maxunicode == 0x10FFFF
+            ):
+                ucs4 = "u"
+    elif debug:
+        # Debug builds can also load "normal" extension modules.
+        # We can also assume no UCS-4 or pymalloc requirement.
+        abis.append(f"cp{version}{threading}")
+    abis.insert(0, f"cp{version}{threading}{debug}{pymalloc}{ucs4}")
+    return abis
+
+
+def cpython_tags(
+    python_version: Optional[PythonVersion] = None,
+    abis: Optional[Iterable[str]] = None,
+    platforms: Optional[Iterable[str]] = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a CPython interpreter.
+
+    The tags consist of:
+    - cp<python_version>-<abi>-<platform>
+    - cp<python_version>-abi3-<platform>
+    - cp<python_version>-none-<platform>
+    - cp<less than python_version>-abi3-<platform>  # Older Python versions down to 3.2.
+
+    If python_version only specifies a major version then user-provided ABIs and
+    the 'none' ABItag will be used.
+
+    If 'abi3' or 'none' are specified in 'abis' then they will be yielded at
+    their normal position and not at the beginning.
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+
+    interpreter = f"cp{_version_nodot(python_version[:2])}"
+
+    if abis is None:
+        if len(python_version) > 1:
+            abis = _cpython_abis(python_version, warn)
+        else:
+            abis = []
+    abis = list(abis)
+    # 'abi3' and 'none' are explicitly handled later.
+    for explicit_abi in ("abi3", "none"):
+        try:
+            abis.remove(explicit_abi)
+        except ValueError:
+            pass
+
+    platforms = list(platforms or platform_tags())
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+    threading = _is_threaded_cpython(abis)
+    use_abi3 = _abi3_applies(python_version, threading)
+    if use_abi3:
+        yield from (Tag(interpreter, "abi3", platform_) for platform_ in platforms)
+    yield from (Tag(interpreter, "none", platform_) for platform_ in platforms)
+
+    if use_abi3:
+        for minor_version in range(python_version[1] - 1, 1, -1):
+            for platform_ in platforms:
+                interpreter = "cp{version}".format(
+                    version=_version_nodot((python_version[0], minor_version))
+                )
+                yield Tag(interpreter, "abi3", platform_)
+
+
+def _generic_abi() -> List[str]:
+    """
+    Return the ABI tag based on EXT_SUFFIX.
+    """
+    # The following are examples of `EXT_SUFFIX`.
+    # We want to keep the parts which are related to the ABI and remove the
+    # parts which are related to the platform:
+    # - linux:   '.cpython-310-x86_64-linux-gnu.so' => cp310
+    # - mac:     '.cpython-310-darwin.so'           => cp310
+    # - win:     '.cp310-win_amd64.pyd'             => cp310
+    # - win:     '.pyd'                             => cp37 (uses _cpython_abis())
+    # - pypy:    '.pypy38-pp73-x86_64-linux-gnu.so' => pypy38_pp73
+    # - graalpy: '.graalpy-38-native-x86_64-darwin.dylib'
+    #                                               => graalpy_38_native
+
+    ext_suffix = _get_config_var("EXT_SUFFIX", warn=True)
+    if not isinstance(ext_suffix, str) or ext_suffix[0] != ".":
+        raise SystemError("invalid sysconfig.get_config_var('EXT_SUFFIX')")
+    parts = ext_suffix.split(".")
+    if len(parts) < 3:
+        # CPython3.7 and earlier uses ".pyd" on Windows.
+        return _cpython_abis(sys.version_info[:2])
+    soabi = parts[1]
+    if soabi.startswith("cpython"):
+        # non-windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi.startswith("cp"):
+        # windows
+        abi = soabi.split("-")[0]
+    elif soabi.startswith("pypy"):
+        abi = "-".join(soabi.split("-")[:2])
+    elif soabi.startswith("graalpy"):
+        abi = "-".join(soabi.split("-")[:3])
+    elif soabi:
+        # pyston, ironpython, others?
+        abi = soabi
+    else:
+        return []
+    return [_normalize_string(abi)]
+
+
+def generic_tags(
+    interpreter: Optional[str] = None,
+    abis: Optional[Iterable[str]] = None,
+    platforms: Optional[Iterable[str]] = None,
+    *,
+    warn: bool = False,
+) -> Iterator[Tag]:
+    """
+    Yields the tags for a generic interpreter.
+
+    The tags consist of:
+    - <interpreter>-<abi>-<platform>
+
+    The "none" ABI will be added if it was not explicitly provided.
+    """
+    if not interpreter:
+        interp_name = interpreter_name()
+        interp_version = interpreter_version(warn=warn)
+        interpreter = "".join([interp_name, interp_version])
+    if abis is None:
+        abis = _generic_abi()
+    else:
+        abis = list(abis)
+    platforms = list(platforms or platform_tags())
+    if "none" not in abis:
+        abis.append("none")
+    for abi in abis:
+        for platform_ in platforms:
+            yield Tag(interpreter, abi, platform_)
+
+
+def _py_interpreter_range(py_version: PythonVersion) -> Iterator[str]:
+    """
+    Yields Python versions in descending order.
+
+    After the latest version, the major-only version will be yielded, and then
+    all previous versions of that major version.
+    """
+    if len(py_version) > 1:
+        yield f"py{_version_nodot(py_version[:2])}"
+    yield f"py{py_version[0]}"
+    if len(py_version) > 1:
+        for minor in range(py_version[1] - 1, -1, -1):
+            yield f"py{_version_nodot((py_version[0], minor))}"
+
+
+def compatible_tags(
+    python_version: Optional[PythonVersion] = None,
+    interpreter: Optional[str] = None,
+    platforms: Optional[Iterable[str]] = None,
+) -> Iterator[Tag]:
+    """
+    Yields the sequence of tags that are compatible with a specific version of Python.
+
+    The tags consist of:
+    - py*-none-<platform>
+    - <interpreter>-none-any  # ... if `interpreter` is provided.
+    - py*-none-any
+    """
+    if not python_version:
+        python_version = sys.version_info[:2]
+    platforms = list(platforms or platform_tags())
+    for version in _py_interpreter_range(python_version):
+        for platform_ in platforms:
+            yield Tag(version, "none", platform_)
+    if interpreter:
+        yield Tag(interpreter, "none", "any")
+    for version in _py_interpreter_range(python_version):
+        yield Tag(version, "none", "any")
+
+
+def _mac_arch(arch: str, is_32bit: bool = _32_BIT_INTERPRETER) -> str:
+    if not is_32bit:
+        return arch
+
+    if arch.startswith("ppc"):
+        return "ppc"
+
+    return "i386"
+
+
+def _mac_binary_formats(version: MacVersion, cpu_arch: str) -> List[str]:
+    formats = [cpu_arch]
+    if cpu_arch == "x86_64":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat64", "fat32"])
+
+    elif cpu_arch == "i386":
+        if version < (10, 4):
+            return []
+        formats.extend(["intel", "fat32", "fat"])
+
+    elif cpu_arch == "ppc64":
+        # TODO: Need to care about 32-bit PPC for ppc64 through 10.2?
+        if version > (10, 5) or version < (10, 4):
+            return []
+        formats.append("fat64")
+
+    elif cpu_arch == "ppc":
+        if version > (10, 6):
+            return []
+        formats.extend(["fat32", "fat"])
+
+    if cpu_arch in {"arm64", "x86_64"}:
+        formats.append("universal2")
+
+    if cpu_arch in {"x86_64", "i386", "ppc64", "ppc", "intel"}:
+        formats.append("universal")
+
+    return formats
+
+
+def mac_platforms(
+    version: Optional[MacVersion] = None, arch: Optional[str] = None
+) -> Iterator[str]:
+    """
+    Yields the platform tags for a macOS system.
+
+    The `version` parameter is a two-item tuple specifying the macOS version to
+    generate platform tags for. The `arch` parameter is the CPU architecture to
+    generate platform tags for. Both parameters default to the appropriate value
+    for the current system.
+    """
+    version_str, _, cpu_arch = platform.mac_ver()
+    if version is None:
+        version = cast("MacVersion", tuple(map(int, version_str.split(".")[:2])))
+        if version == (10, 16):
+            # When built against an older macOS SDK, Python will report macOS 10.16
+            # instead of the real version.
+            version_str = subprocess.run(
+                [
+                    sys.executable,
+                    "-sS",
+                    "-c",
+                    "import platform; print(platform.mac_ver()[0])",
+                ],
+                check=True,
+                env={"SYSTEM_VERSION_COMPAT": "0"},
+                stdout=subprocess.PIPE,
+                text=True,
+            ).stdout
+            version = cast("MacVersion", tuple(map(int, version_str.split(".")[:2])))
+    else:
+        version = version
+    if arch is None:
+        arch = _mac_arch(cpu_arch)
+    else:
+        arch = arch
+
+    if (10, 0) <= version and version < (11, 0):
+        # Prior to Mac OS 11, each yearly release of Mac OS bumped the
+        # "minor" version number.  The major version was always 10.
+        for minor_version in range(version[1], -1, -1):
+            compat_version = 10, minor_version
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield "macosx_{major}_{minor}_{binary_format}".format(
+                    major=10, minor=minor_version, binary_format=binary_format
+                )
+
+    if version >= (11, 0):
+        # Starting with Mac OS 11, each yearly release bumps the major version
+        # number.   The minor versions are now the midyear updates.
+        for major_version in range(version[0], 10, -1):
+            compat_version = major_version, 0
+            binary_formats = _mac_binary_formats(compat_version, arch)
+            for binary_format in binary_formats:
+                yield "macosx_{major}_{minor}_{binary_format}".format(
+                    major=major_version, minor=0, binary_format=binary_format
+                )
+
+    if version >= (11, 0):
+        # Mac OS 11 on x86_64 is compatible with binaries from previous releases.
+        # Arm64 support was introduced in 11.0, so no Arm binaries from previous
+        # releases exist.
+        #
+        # However, the "universal2" binary format can have a
+        # macOS version earlier than 11.0 when the x86_64 part of the binary supports
+        # that version of macOS.
+        if arch == "x86_64":
+            for minor_version in range(16, 3, -1):
+                compat_version = 10, minor_version
+                binary_formats = _mac_binary_formats(compat_version, arch)
+                for binary_format in binary_formats:
+                    yield "macosx_{major}_{minor}_{binary_format}".format(
+                        major=compat_version[0],
+                        minor=compat_version[1],
+                        binary_format=binary_format,
+                    )
+        else:
+            for minor_version in range(16, 3, -1):
+                compat_version = 10, minor_version
+                binary_format = "universal2"
+                yield "macosx_{major}_{minor}_{binary_format}".format(
+                    major=compat_version[0],
+                    minor=compat_version[1],
+                    binary_format=binary_format,
+                )
+
+
+def _linux_platforms(is_32bit: bool = _32_BIT_INTERPRETER) -> Iterator[str]:
+    linux = _normalize_string(sysconfig.get_platform())
+    if not linux.startswith("linux_"):
+        # we should never be here, just yield the sysconfig one and return
+        yield linux
+        return
+    if is_32bit:
+        if linux == "linux_x86_64":
+            linux = "linux_i686"
+        elif linux == "linux_aarch64":
+            linux = "linux_armv8l"
+    _, arch = linux.split("_", 1)
+    archs = {"armv8l": ["armv8l", "armv7l"]}.get(arch, [arch])
+    yield from _manylinux.platform_tags(archs)
+    yield from _musllinux.platform_tags(archs)
+    for arch in archs:
+        yield f"linux_{arch}"
+
+
+def _generic_platforms() -> Iterator[str]:
+    yield _normalize_string(sysconfig.get_platform())
+
+
+def platform_tags() -> Iterator[str]:
+    """
+    Provides the platform tags for this installation.
+    """
+    if platform.system() == "Darwin":
+        return mac_platforms()
+    elif platform.system() == "Linux":
+        return _linux_platforms()
+    else:
+        return _generic_platforms()
+
+
+def interpreter_name() -> str:
+    """
+    Returns the name of the running interpreter.
+
+    Some implementations have a reserved, two-letter abbreviation which will
+    be returned when appropriate.
+    """
+    name = sys.implementation.name
+    return INTERPRETER_SHORT_NAMES.get(name) or name
+
+
+def interpreter_version(*, warn: bool = False) -> str:
+    """
+    Returns the version of the running interpreter.
+    """
+    version = _get_config_var("py_version_nodot", warn=warn)
+    if version:
+        version = str(version)
+    else:
+        version = _version_nodot(sys.version_info[:2])
+    return version
+
+
+def _version_nodot(version: PythonVersion) -> str:
+    return "".join(map(str, version))
+
+
+def sys_tags(*, warn: bool = False) -> Iterator[Tag]:
+    """
+    Returns the sequence of tag triples for the running interpreter.
+
+    The order of the sequence corresponds to priority order for the
+    interpreter, from most to least important.
+    """
+
+    interp_name = interpreter_name()
+    if interp_name == "cp":
+        yield from cpython_tags(warn=warn)
+    else:
+        yield from generic_tags()
+
+    if interp_name == "pp":
+        interp = "pp3"
+    elif interp_name == "cp":
+        interp = "cp" + interpreter_version(warn=warn)
+    else:
+        interp = None
+    yield from compatible_tags(interpreter=interp)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/utils.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/utils.py
new file mode 100644
index 00000000..c2c2f75a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/utils.py
@@ -0,0 +1,172 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+
+import re
+from typing import FrozenSet, NewType, Tuple, Union, cast
+
+from .tags import Tag, parse_tag
+from .version import InvalidVersion, Version
+
+BuildTag = Union[Tuple[()], Tuple[int, str]]
+NormalizedName = NewType("NormalizedName", str)
+
+
+class InvalidName(ValueError):
+    """
+    An invalid distribution name; users should refer to the packaging user guide.
+    """
+
+
+class InvalidWheelFilename(ValueError):
+    """
+    An invalid wheel filename was found, users should refer to PEP 427.
+    """
+
+
+class InvalidSdistFilename(ValueError):
+    """
+    An invalid sdist filename was found, users should refer to the packaging user guide.
+    """
+
+
+# Core metadata spec for `Name`
+_validate_regex = re.compile(
+    r"^([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])$", re.IGNORECASE
+)
+_canonicalize_regex = re.compile(r"[-_.]+")
+_normalized_regex = re.compile(r"^([a-z0-9]|[a-z0-9]([a-z0-9-](?!--))*[a-z0-9])$")
+# PEP 427: The build number must start with a digit.
+_build_tag_regex = re.compile(r"(\d+)(.*)")
+
+
+def canonicalize_name(name: str, *, validate: bool = False) -> NormalizedName:
+    if validate and not _validate_regex.match(name):
+        raise InvalidName(f"name is invalid: {name!r}")
+    # This is taken from PEP 503.
+    value = _canonicalize_regex.sub("-", name).lower()
+    return cast(NormalizedName, value)
+
+
+def is_normalized_name(name: str) -> bool:
+    return _normalized_regex.match(name) is not None
+
+
+def canonicalize_version(
+    version: Union[Version, str], *, strip_trailing_zero: bool = True
+) -> str:
+    """
+    This is very similar to Version.__str__, but has one subtle difference
+    with the way it handles the release segment.
+    """
+    if isinstance(version, str):
+        try:
+            parsed = Version(version)
+        except InvalidVersion:
+            # Legacy versions cannot be normalized
+            return version
+    else:
+        parsed = version
+
+    parts = []
+
+    # Epoch
+    if parsed.epoch != 0:
+        parts.append(f"{parsed.epoch}!")
+
+    # Release segment
+    release_segment = ".".join(str(x) for x in parsed.release)
+    if strip_trailing_zero:
+        # NB: This strips trailing '.0's to normalize
+        release_segment = re.sub(r"(\.0)+$", "", release_segment)
+    parts.append(release_segment)
+
+    # Pre-release
+    if parsed.pre is not None:
+        parts.append("".join(str(x) for x in parsed.pre))
+
+    # Post-release
+    if parsed.post is not None:
+        parts.append(f".post{parsed.post}")
+
+    # Development release
+    if parsed.dev is not None:
+        parts.append(f".dev{parsed.dev}")
+
+    # Local version segment
+    if parsed.local is not None:
+        parts.append(f"+{parsed.local}")
+
+    return "".join(parts)
+
+
+def parse_wheel_filename(
+    filename: str,
+) -> Tuple[NormalizedName, Version, BuildTag, FrozenSet[Tag]]:
+    if not filename.endswith(".whl"):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (extension must be '.whl'): {filename}"
+        )
+
+    filename = filename[:-4]
+    dashes = filename.count("-")
+    if dashes not in (4, 5):
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (wrong number of parts): {filename}"
+        )
+
+    parts = filename.split("-", dashes - 2)
+    name_part = parts[0]
+    # See PEP 427 for the rules on escaping the project name.
+    if "__" in name_part or re.match(r"^[\w\d._]*$", name_part, re.UNICODE) is None:
+        raise InvalidWheelFilename(f"Invalid project name: {filename}")
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(parts[1])
+    except InvalidVersion as e:
+        raise InvalidWheelFilename(
+            f"Invalid wheel filename (invalid version): {filename}"
+        ) from e
+
+    if dashes == 5:
+        build_part = parts[2]
+        build_match = _build_tag_regex.match(build_part)
+        if build_match is None:
+            raise InvalidWheelFilename(
+                f"Invalid build number: {build_part} in '{filename}'"
+            )
+        build = cast(BuildTag, (int(build_match.group(1)), build_match.group(2)))
+    else:
+        build = ()
+    tags = parse_tag(parts[-1])
+    return (name, version, build, tags)
+
+
+def parse_sdist_filename(filename: str) -> Tuple[NormalizedName, Version]:
+    if filename.endswith(".tar.gz"):
+        file_stem = filename[: -len(".tar.gz")]
+    elif filename.endswith(".zip"):
+        file_stem = filename[: -len(".zip")]
+    else:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (extension must be '.tar.gz' or '.zip'):"
+            f" {filename}"
+        )
+
+    # We are requiring a PEP 440 version, which cannot contain dashes,
+    # so we split on the last dash.
+    name_part, sep, version_part = file_stem.rpartition("-")
+    if not sep:
+        raise InvalidSdistFilename(f"Invalid sdist filename: {filename}")
+
+    name = canonicalize_name(name_part)
+
+    try:
+        version = Version(version_part)
+    except InvalidVersion as e:
+        raise InvalidSdistFilename(
+            f"Invalid sdist filename (invalid version): {filename}"
+        ) from e
+
+    return (name, version)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/version.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/version.py
new file mode 100644
index 00000000..cda8e999
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/packaging/version.py
@@ -0,0 +1,561 @@
+# This file is dual licensed under the terms of the Apache License, Version
+# 2.0, and the BSD License. See the LICENSE file in the root of this repository
+# for complete details.
+"""
+.. testsetup::
+
+    from packaging.version import parse, Version
+"""
+
+import itertools
+import re
+from typing import Any, Callable, NamedTuple, Optional, SupportsInt, Tuple, Union
+
+from ._structures import Infinity, InfinityType, NegativeInfinity, NegativeInfinityType
+
+__all__ = ["VERSION_PATTERN", "parse", "Version", "InvalidVersion"]
+
+LocalType = Tuple[Union[int, str], ...]
+
+CmpPrePostDevType = Union[InfinityType, NegativeInfinityType, Tuple[str, int]]
+CmpLocalType = Union[
+    NegativeInfinityType,
+    Tuple[Union[Tuple[int, str], Tuple[NegativeInfinityType, Union[int, str]]], ...],
+]
+CmpKey = Tuple[
+    int,
+    Tuple[int, ...],
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpPrePostDevType,
+    CmpLocalType,
+]
+VersionComparisonMethod = Callable[[CmpKey, CmpKey], bool]
+
+
+class _Version(NamedTuple):
+    epoch: int
+    release: Tuple[int, ...]
+    dev: Optional[Tuple[str, int]]
+    pre: Optional[Tuple[str, int]]
+    post: Optional[Tuple[str, int]]
+    local: Optional[LocalType]
+
+
+def parse(version: str) -> "Version":
+    """Parse the given version string.
+
+    >>> parse('1.0.dev1')
+    <Version('1.0.dev1')>
+
+    :param version: The version string to parse.
+    :raises InvalidVersion: When the version string is not a valid version.
+    """
+    return Version(version)
+
+
+class InvalidVersion(ValueError):
+    """Raised when a version string is not a valid version.
+
+    >>> Version("invalid")
+    Traceback (most recent call last):
+        ...
+    packaging.version.InvalidVersion: Invalid version: 'invalid'
+    """
+
+
+class _BaseVersion:
+    _key: Tuple[Any, ...]
+
+    def __hash__(self) -> int:
+        return hash(self._key)
+
+    # Please keep the duplicated `isinstance` check
+    # in the six comparisons hereunder
+    # unless you find a way to avoid adding overhead function calls.
+    def __lt__(self, other: "_BaseVersion") -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key < other._key
+
+    def __le__(self, other: "_BaseVersion") -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key <= other._key
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key == other._key
+
+    def __ge__(self, other: "_BaseVersion") -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key >= other._key
+
+    def __gt__(self, other: "_BaseVersion") -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key > other._key
+
+    def __ne__(self, other: object) -> bool:
+        if not isinstance(other, _BaseVersion):
+            return NotImplemented
+
+        return self._key != other._key
+
+
+# Deliberately not anchored to the start and end of the string, to make it
+# easier for 3rd party code to reuse
+_VERSION_PATTERN = r"""
+    v?
+    (?:
+        (?:(?P<epoch>[0-9]+)!)?                           # epoch
+        (?P<release>[0-9]+(?:\.[0-9]+)*)                  # release segment
+        (?P<pre>                                          # pre-release
+            [-_\.]?
+            (?P<pre_l>alpha|a|beta|b|preview|pre|c|rc)
+            [-_\.]?
+            (?P<pre_n>[0-9]+)?
+        )?
+        (?P<post>                                         # post release
+            (?:-(?P<post_n1>[0-9]+))
+            |
+            (?:
+                [-_\.]?
+                (?P<post_l>post|rev|r)
+                [-_\.]?
+                (?P<post_n2>[0-9]+)?
+            )
+        )?
+        (?P<dev>                                          # dev release
+            [-_\.]?
+            (?P<dev_l>dev)
+            [-_\.]?
+            (?P<dev_n>[0-9]+)?
+        )?
+    )
+    (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
+"""
+
+VERSION_PATTERN = _VERSION_PATTERN
+"""
+A string containing the regular expression used to match a valid version.
+
+The pattern is not anchored at either end, and is intended for embedding in larger
+expressions (for example, matching a version number as part of a file name). The
+regular expression should be compiled with the ``re.VERBOSE`` and ``re.IGNORECASE``
+flags set.
+
+:meta hide-value:
+"""
+
+
+class Version(_BaseVersion):
+    """This class abstracts handling of a project's versions.
+
+    A :class:`Version` instance is comparison aware and can be compared and
+    sorted using the standard Python interfaces.
+
+    >>> v1 = Version("1.0a5")
+    >>> v2 = Version("1.0")
+    >>> v1
+    <Version('1.0a5')>
+    >>> v2
+    <Version('1.0')>
+    >>> v1 < v2
+    True
+    >>> v1 == v2
+    False
+    >>> v1 > v2
+    False
+    >>> v1 >= v2
+    False
+    >>> v1 <= v2
+    True
+    """
+
+    _regex = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.VERBOSE | re.IGNORECASE)
+    _key: CmpKey
+
+    def __init__(self, version: str) -> None:
+        """Initialize a Version object.
+
+        :param version:
+            The string representation of a version which will be parsed and normalized
+            before use.
+        :raises InvalidVersion:
+            If the ``version`` does not conform to PEP 440 in any way then this
+            exception will be raised.
+        """
+
+        # Validate the version and parse it into pieces
+        match = self._regex.search(version)
+        if not match:
+            raise InvalidVersion(f"Invalid version: '{version}'")
+
+        # Store the parsed out pieces of the version
+        self._version = _Version(
+            epoch=int(match.group("epoch")) if match.group("epoch") else 0,
+            release=tuple(int(i) for i in match.group("release").split(".")),
+            pre=_parse_letter_version(match.group("pre_l"), match.group("pre_n")),
+            post=_parse_letter_version(
+                match.group("post_l"), match.group("post_n1") or match.group("post_n2")
+            ),
+            dev=_parse_letter_version(match.group("dev_l"), match.group("dev_n")),
+            local=_parse_local_version(match.group("local")),
+        )
+
+        # Generate a key which will be used for sorting
+        self._key = _cmpkey(
+            self._version.epoch,
+            self._version.release,
+            self._version.pre,
+            self._version.post,
+            self._version.dev,
+            self._version.local,
+        )
+
+    def __repr__(self) -> str:
+        """A representation of the Version that shows all internal state.
+
+        >>> Version('1.0.0')
+        <Version('1.0.0')>
+        """
+        return f"<Version('{self}')>"
+
+    def __str__(self) -> str:
+        """A string representation of the version that can be rounded-tripped.
+
+        >>> str(Version("1.0a5"))
+        '1.0a5'
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        # Pre-release
+        if self.pre is not None:
+            parts.append("".join(str(x) for x in self.pre))
+
+        # Post-release
+        if self.post is not None:
+            parts.append(f".post{self.post}")
+
+        # Development release
+        if self.dev is not None:
+            parts.append(f".dev{self.dev}")
+
+        # Local version segment
+        if self.local is not None:
+            parts.append(f"+{self.local}")
+
+        return "".join(parts)
+
+    @property
+    def epoch(self) -> int:
+        """The epoch of the version.
+
+        >>> Version("2.0.0").epoch
+        0
+        >>> Version("1!2.0.0").epoch
+        1
+        """
+        return self._version.epoch
+
+    @property
+    def release(self) -> Tuple[int, ...]:
+        """The components of the "release" segment of the version.
+
+        >>> Version("1.2.3").release
+        (1, 2, 3)
+        >>> Version("2.0.0").release
+        (2, 0, 0)
+        >>> Version("1!2.0.0.post0").release
+        (2, 0, 0)
+
+        Includes trailing zeroes but not the epoch or any pre-release / development /
+        post-release suffixes.
+        """
+        return self._version.release
+
+    @property
+    def pre(self) -> Optional[Tuple[str, int]]:
+        """The pre-release segment of the version.
+
+        >>> print(Version("1.2.3").pre)
+        None
+        >>> Version("1.2.3a1").pre
+        ('a', 1)
+        >>> Version("1.2.3b1").pre
+        ('b', 1)
+        >>> Version("1.2.3rc1").pre
+        ('rc', 1)
+        """
+        return self._version.pre
+
+    @property
+    def post(self) -> Optional[int]:
+        """The post-release number of the version.
+
+        >>> print(Version("1.2.3").post)
+        None
+        >>> Version("1.2.3.post1").post
+        1
+        """
+        return self._version.post[1] if self._version.post else None
+
+    @property
+    def dev(self) -> Optional[int]:
+        """The development number of the version.
+
+        >>> print(Version("1.2.3").dev)
+        None
+        >>> Version("1.2.3.dev1").dev
+        1
+        """
+        return self._version.dev[1] if self._version.dev else None
+
+    @property
+    def local(self) -> Optional[str]:
+        """The local version segment of the version.
+
+        >>> print(Version("1.2.3").local)
+        None
+        >>> Version("1.2.3+abc").local
+        'abc'
+        """
+        if self._version.local:
+            return ".".join(str(x) for x in self._version.local)
+        else:
+            return None
+
+    @property
+    def public(self) -> str:
+        """The public portion of the version.
+
+        >>> Version("1.2.3").public
+        '1.2.3'
+        >>> Version("1.2.3+abc").public
+        '1.2.3'
+        >>> Version("1.2.3+abc.dev1").public
+        '1.2.3'
+        """
+        return str(self).split("+", 1)[0]
+
+    @property
+    def base_version(self) -> str:
+        """The "base version" of the version.
+
+        >>> Version("1.2.3").base_version
+        '1.2.3'
+        >>> Version("1.2.3+abc").base_version
+        '1.2.3'
+        >>> Version("1!1.2.3+abc.dev1").base_version
+        '1!1.2.3'
+
+        The "base version" is the public version of the project without any pre or post
+        release markers.
+        """
+        parts = []
+
+        # Epoch
+        if self.epoch != 0:
+            parts.append(f"{self.epoch}!")
+
+        # Release segment
+        parts.append(".".join(str(x) for x in self.release))
+
+        return "".join(parts)
+
+    @property
+    def is_prerelease(self) -> bool:
+        """Whether this version is a pre-release.
+
+        >>> Version("1.2.3").is_prerelease
+        False
+        >>> Version("1.2.3a1").is_prerelease
+        True
+        >>> Version("1.2.3b1").is_prerelease
+        True
+        >>> Version("1.2.3rc1").is_prerelease
+        True
+        >>> Version("1.2.3dev1").is_prerelease
+        True
+        """
+        return self.dev is not None or self.pre is not None
+
+    @property
+    def is_postrelease(self) -> bool:
+        """Whether this version is a post-release.
+
+        >>> Version("1.2.3").is_postrelease
+        False
+        >>> Version("1.2.3.post1").is_postrelease
+        True
+        """
+        return self.post is not None
+
+    @property
+    def is_devrelease(self) -> bool:
+        """Whether this version is a development release.
+
+        >>> Version("1.2.3").is_devrelease
+        False
+        >>> Version("1.2.3.dev1").is_devrelease
+        True
+        """
+        return self.dev is not None
+
+    @property
+    def major(self) -> int:
+        """The first item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").major
+        1
+        """
+        return self.release[0] if len(self.release) >= 1 else 0
+
+    @property
+    def minor(self) -> int:
+        """The second item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").minor
+        2
+        >>> Version("1").minor
+        0
+        """
+        return self.release[1] if len(self.release) >= 2 else 0
+
+    @property
+    def micro(self) -> int:
+        """The third item of :attr:`release` or ``0`` if unavailable.
+
+        >>> Version("1.2.3").micro
+        3
+        >>> Version("1").micro
+        0
+        """
+        return self.release[2] if len(self.release) >= 3 else 0
+
+
+def _parse_letter_version(
+    letter: Optional[str], number: Union[str, bytes, SupportsInt, None]
+) -> Optional[Tuple[str, int]]:
+    if letter:
+        # We consider there to be an implicit 0 in a pre-release if there is
+        # not a numeral associated with it.
+        if number is None:
+            number = 0
+
+        # We normalize any letters to their lower case form
+        letter = letter.lower()
+
+        # We consider some words to be alternate spellings of other words and
+        # in those cases we want to normalize the spellings to our preferred
+        # spelling.
+        if letter == "alpha":
+            letter = "a"
+        elif letter == "beta":
+            letter = "b"
+        elif letter in ["c", "pre", "preview"]:
+            letter = "rc"
+        elif letter in ["rev", "r"]:
+            letter = "post"
+
+        return letter, int(number)
+    if not letter and number:
+        # We assume if we are given a number, but we are not given a letter
+        # then this is using the implicit post release syntax (e.g. 1.0-1)
+        letter = "post"
+
+        return letter, int(number)
+
+    return None
+
+
+_local_version_separators = re.compile(r"[\._-]")
+
+
+def _parse_local_version(local: Optional[str]) -> Optional[LocalType]:
+    """
+    Takes a string like abc.1.twelve and turns it into ("abc", 1, "twelve").
+    """
+    if local is not None:
+        return tuple(
+            part.lower() if not part.isdigit() else int(part)
+            for part in _local_version_separators.split(local)
+        )
+    return None
+
+
+def _cmpkey(
+    epoch: int,
+    release: Tuple[int, ...],
+    pre: Optional[Tuple[str, int]],
+    post: Optional[Tuple[str, int]],
+    dev: Optional[Tuple[str, int]],
+    local: Optional[LocalType],
+) -> CmpKey:
+    # When we compare a release version, we want to compare it with all of the
+    # trailing zeros removed. So we'll use a reverse the list, drop all the now
+    # leading zeros until we come to something non zero, then take the rest
+    # re-reverse it back into the correct order and make it a tuple and use
+    # that for our sorting key.
+    _release = tuple(
+        reversed(list(itertools.dropwhile(lambda x: x == 0, reversed(release))))
+    )
+
+    # We need to "trick" the sorting algorithm to put 1.0.dev0 before 1.0a0.
+    # We'll do this by abusing the pre segment, but we _only_ want to do this
+    # if there is not a pre or a post segment. If we have one of those then
+    # the normal sorting rules will handle this case correctly.
+    if pre is None and post is None and dev is not None:
+        _pre: CmpPrePostDevType = NegativeInfinity
+    # Versions without a pre-release (except as noted above) should sort after
+    # those with one.
+    elif pre is None:
+        _pre = Infinity
+    else:
+        _pre = pre
+
+    # Versions without a post segment should sort before those with one.
+    if post is None:
+        _post: CmpPrePostDevType = NegativeInfinity
+
+    else:
+        _post = post
+
+    # Versions without a development segment should sort after those with one.
+    if dev is None:
+        _dev: CmpPrePostDevType = Infinity
+
+    else:
+        _dev = dev
+
+    if local is None:
+        # Versions without a local segment should sort before those with one.
+        _local: CmpLocalType = NegativeInfinity
+    else:
+        # Versions with a local segment need that segment parsed to implement
+        # the sorting rules in PEP440.
+        # - Alpha numeric segments sort before numeric segments
+        # - Alpha numeric segments sort lexicographically
+        # - Numeric segments sort numerically
+        # - Shorter versions sort before longer versions when the prefixes
+        #   match exactly
+        _local = tuple(
+            (i, "") if isinstance(i, int) else (NegativeInfinity, i) for i in local
+        )
+
+    return epoch, _release, _pre, _post, _dev, _local
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/vendor.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/vendor.txt
new file mode 100644
index 00000000..14666103
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/vendored/vendor.txt
@@ -0,0 +1 @@
+packaging==24.0
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/wheelfile.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/wheelfile.py
new file mode 100644
index 00000000..6440e90a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/wheel/wheelfile.py
@@ -0,0 +1,196 @@
+from __future__ import annotations
+
+import csv
+import hashlib
+import os.path
+import re
+import stat
+import time
+from io import StringIO, TextIOWrapper
+from zipfile import ZIP_DEFLATED, ZipFile, ZipInfo
+
+from wheel.cli import WheelError
+from wheel.util import log, urlsafe_b64decode, urlsafe_b64encode
+
+# Non-greedy matching of an optional build number may be too clever (more
+# invalid wheel filenames will match). Separate regex for .dist-info?
+WHEEL_INFO_RE = re.compile(
+    r"""^(?P<namever>(?P<name>[^\s-]+?)-(?P<ver>[^\s-]+?))(-(?P<build>\d[^\s-]*))?
+     -(?P<pyver>[^\s-]+?)-(?P<abi>[^\s-]+?)-(?P<plat>\S+)\.whl$""",
+    re.VERBOSE,
+)
+MINIMUM_TIMESTAMP = 315532800  # 1980-01-01 00:00:00 UTC
+
+
+def get_zipinfo_datetime(timestamp=None):
+    # Some applications need reproducible .whl files, but they can't do this without
+    # forcing the timestamp of the individual ZipInfo objects. See issue #143.
+    timestamp = int(os.environ.get("SOURCE_DATE_EPOCH", timestamp or time.time()))
+    timestamp = max(timestamp, MINIMUM_TIMESTAMP)
+    return time.gmtime(timestamp)[0:6]
+
+
+class WheelFile(ZipFile):
+    """A ZipFile derivative class that also reads SHA-256 hashes from
+    .dist-info/RECORD and checks any read files against those.
+    """
+
+    _default_algorithm = hashlib.sha256
+
+    def __init__(self, file, mode="r", compression=ZIP_DEFLATED):
+        basename = os.path.basename(file)
+        self.parsed_filename = WHEEL_INFO_RE.match(basename)
+        if not basename.endswith(".whl") or self.parsed_filename is None:
+            raise WheelError(f"Bad wheel filename {basename!r}")
+
+        ZipFile.__init__(self, file, mode, compression=compression, allowZip64=True)
+
+        self.dist_info_path = "{}.dist-info".format(
+            self.parsed_filename.group("namever")
+        )
+        self.record_path = self.dist_info_path + "/RECORD"
+        self._file_hashes = {}
+        self._file_sizes = {}
+        if mode == "r":
+            # Ignore RECORD and any embedded wheel signatures
+            self._file_hashes[self.record_path] = None, None
+            self._file_hashes[self.record_path + ".jws"] = None, None
+            self._file_hashes[self.record_path + ".p7s"] = None, None
+
+            # Fill in the expected hashes by reading them from RECORD
+            try:
+                record = self.open(self.record_path)
+            except KeyError:
+                raise WheelError(f"Missing {self.record_path} file") from None
+
+            with record:
+                for line in csv.reader(
+                    TextIOWrapper(record, newline="", encoding="utf-8")
+                ):
+                    path, hash_sum, size = line
+                    if not hash_sum:
+                        continue
+
+                    algorithm, hash_sum = hash_sum.split("=")
+                    try:
+                        hashlib.new(algorithm)
+                    except ValueError:
+                        raise WheelError(
+                            f"Unsupported hash algorithm: {algorithm}"
+                        ) from None
+
+                    if algorithm.lower() in {"md5", "sha1"}:
+                        raise WheelError(
+                            f"Weak hash algorithm ({algorithm}) is not permitted by "
+                            f"PEP 427"
+                        )
+
+                    self._file_hashes[path] = (
+                        algorithm,
+                        urlsafe_b64decode(hash_sum.encode("ascii")),
+                    )
+
+    def open(self, name_or_info, mode="r", pwd=None):
+        def _update_crc(newdata):
+            eof = ef._eof
+            update_crc_orig(newdata)
+            running_hash.update(newdata)
+            if eof and running_hash.digest() != expected_hash:
+                raise WheelError(f"Hash mismatch for file '{ef_name}'")
+
+        ef_name = (
+            name_or_info.filename if isinstance(name_or_info, ZipInfo) else name_or_info
+        )
+        if (
+            mode == "r"
+            and not ef_name.endswith("/")
+            and ef_name not in self._file_hashes
+        ):
+            raise WheelError(f"No hash found for file '{ef_name}'")
+
+        ef = ZipFile.open(self, name_or_info, mode, pwd)
+        if mode == "r" and not ef_name.endswith("/"):
+            algorithm, expected_hash = self._file_hashes[ef_name]
+            if expected_hash is not None:
+                # Monkey patch the _update_crc method to also check for the hash from
+                # RECORD
+                running_hash = hashlib.new(algorithm)
+                update_crc_orig, ef._update_crc = ef._update_crc, _update_crc
+
+        return ef
+
+    def write_files(self, base_dir):
+        log.info(f"creating '{self.filename}' and adding '{base_dir}' to it")
+        deferred = []
+        for root, dirnames, filenames in os.walk(base_dir):
+            # Sort the directory names so that `os.walk` will walk them in a
+            # defined order on the next iteration.
+            dirnames.sort()
+            for name in sorted(filenames):
+                path = os.path.normpath(os.path.join(root, name))
+                if os.path.isfile(path):
+                    arcname = os.path.relpath(path, base_dir).replace(os.path.sep, "/")
+                    if arcname == self.record_path:
+                        pass
+                    elif root.endswith(".dist-info"):
+                        deferred.append((path, arcname))
+                    else:
+                        self.write(path, arcname)
+
+        deferred.sort()
+        for path, arcname in deferred:
+            self.write(path, arcname)
+
+    def write(self, filename, arcname=None, compress_type=None):
+        with open(filename, "rb") as f:
+            st = os.fstat(f.fileno())
+            data = f.read()
+
+        zinfo = ZipInfo(
+            arcname or filename, date_time=get_zipinfo_datetime(st.st_mtime)
+        )
+        zinfo.external_attr = (stat.S_IMODE(st.st_mode) | stat.S_IFMT(st.st_mode)) << 16
+        zinfo.compress_type = compress_type or self.compression
+        self.writestr(zinfo, data, compress_type)
+
+    def writestr(self, zinfo_or_arcname, data, compress_type=None):
+        if isinstance(zinfo_or_arcname, str):
+            zinfo_or_arcname = ZipInfo(
+                zinfo_or_arcname, date_time=get_zipinfo_datetime()
+            )
+            zinfo_or_arcname.compress_type = self.compression
+            zinfo_or_arcname.external_attr = (0o664 | stat.S_IFREG) << 16
+
+        if isinstance(data, str):
+            data = data.encode("utf-8")
+
+        ZipFile.writestr(self, zinfo_or_arcname, data, compress_type)
+        fname = (
+            zinfo_or_arcname.filename
+            if isinstance(zinfo_or_arcname, ZipInfo)
+            else zinfo_or_arcname
+        )
+        log.info(f"adding '{fname}'")
+        if fname != self.record_path:
+            hash_ = self._default_algorithm(data)
+            self._file_hashes[fname] = (
+                hash_.name,
+                urlsafe_b64encode(hash_.digest()).decode("ascii"),
+            )
+            self._file_sizes[fname] = len(data)
+
+    def close(self):
+        # Write RECORD
+        if self.fp is not None and self.mode == "w" and self._file_hashes:
+            data = StringIO()
+            writer = csv.writer(data, delimiter=",", quotechar='"', lineterminator="\n")
+            writer.writerows(
+                (
+                    (fname, algorithm + "=" + hash_, self._file_sizes[fname])
+                    for fname, (algorithm, hash_) in self._file_hashes.items()
+                )
+            )
+            writer.writerow((format(self.record_path), "", ""))
+            self.writestr(self.record_path, data.getvalue())
+
+        ZipFile.close(self)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/LICENSE
new file mode 100644
index 00000000..1bb5a443
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/LICENSE
@@ -0,0 +1,17 @@
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/METADATA
new file mode 100644
index 00000000..13992817
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/METADATA
@@ -0,0 +1,102 @@
+Metadata-Version: 2.1
+Name: zipp
+Version: 3.19.2
+Summary: Backport of pathlib-compatible object wrapper for zip files
+Author-email: "Jason R. Coombs" <jaraco@jaraco.com>
+Project-URL: Homepage, https://github.com/jaraco/zipp
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Requires-Python: >=3.8
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+Provides-Extra: doc
+Requires-Dist: sphinx >=3.5 ; extra == 'doc'
+Requires-Dist: jaraco.packaging >=9.3 ; extra == 'doc'
+Requires-Dist: rst.linker >=1.9 ; extra == 'doc'
+Requires-Dist: furo ; extra == 'doc'
+Requires-Dist: sphinx-lint ; extra == 'doc'
+Requires-Dist: jaraco.tidelift >=1.4 ; extra == 'doc'
+Provides-Extra: test
+Requires-Dist: pytest !=8.1.*,>=6 ; extra == 'test'
+Requires-Dist: pytest-checkdocs >=2.4 ; extra == 'test'
+Requires-Dist: pytest-cov ; extra == 'test'
+Requires-Dist: pytest-mypy ; extra == 'test'
+Requires-Dist: pytest-enabler >=2.2 ; extra == 'test'
+Requires-Dist: pytest-ruff >=0.2.1 ; extra == 'test'
+Requires-Dist: jaraco.itertools ; extra == 'test'
+Requires-Dist: jaraco.functools ; extra == 'test'
+Requires-Dist: more-itertools ; extra == 'test'
+Requires-Dist: big-O ; extra == 'test'
+Requires-Dist: pytest-ignore-flaky ; extra == 'test'
+Requires-Dist: jaraco.test ; extra == 'test'
+Requires-Dist: importlib-resources ; (python_version < "3.9") and extra == 'test'
+
+.. image:: https://img.shields.io/pypi/v/zipp.svg
+   :target: https://pypi.org/project/zipp
+
+.. image:: https://img.shields.io/pypi/pyversions/zipp.svg
+
+.. image:: https://github.com/jaraco/zipp/actions/workflows/main.yml/badge.svg
+   :target: https://github.com/jaraco/zipp/actions?query=workflow%3A%22tests%22
+   :alt: tests
+
+.. image:: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/charliermarsh/ruff/main/assets/badge/v2.json
+    :target: https://github.com/astral-sh/ruff
+    :alt: Ruff
+
+.. .. image:: https://readthedocs.org/projects/PROJECT_RTD/badge/?version=latest
+..    :target: https://PROJECT_RTD.readthedocs.io/en/latest/?badge=latest
+
+.. image:: https://img.shields.io/badge/skeleton-2024-informational
+   :target: https://blog.jaraco.com/skeleton
+
+.. image:: https://tidelift.com/badges/package/pypi/zipp
+   :target: https://tidelift.com/subscription/pkg/pypi-zipp?utm_source=pypi-zipp&utm_medium=readme
+
+
+A pathlib-compatible Zipfile object wrapper. Official backport of the standard library
+`Path object <https://docs.python.org/3.8/library/zipfile.html#path-objects>`_.
+
+
+Compatibility
+=============
+
+New features are introduced in this third-party library and later merged
+into CPython. The following table indicates which versions of this library
+were contributed to different versions in the standard library:
+
+.. list-table::
+   :header-rows: 1
+
+   * - zipp
+     - stdlib
+   * - 3.18
+     - 3.13
+   * - 3.16
+     - 3.12
+   * - 3.5
+     - 3.11
+   * - 3.2
+     - 3.10
+   * - 3.3 ??
+     - 3.9
+   * - 1.0
+     - 3.8
+
+
+Usage
+=====
+
+Use ``zipp.Path`` in place of ``zipfile.Path`` on any Python.
+
+For Enterprise
+==============
+
+Available as part of the Tidelift Subscription.
+
+This project and the maintainers of thousands of other packages are working with Tidelift to deliver one enterprise subscription that covers all of the open source you use.
+
+`Learn more <https://tidelift.com/subscription/pkg/pypi-zipp?utm_source=pypi-zipp&utm_medium=referral&utm_campaign=github>`_.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/RECORD
new file mode 100644
index 00000000..77c02835
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/RECORD
@@ -0,0 +1,15 @@
+zipp-3.19.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+zipp-3.19.2.dist-info/LICENSE,sha256=htoPAa6uRjSKPD1GUZXcHOzN55956HdppkuNoEsqR0E,1023
+zipp-3.19.2.dist-info/METADATA,sha256=UIrk_kMIHGSwsKKChYizqMw0MMZpPRZ2ZiVpQAsN_bE,3575
+zipp-3.19.2.dist-info/RECORD,,
+zipp-3.19.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+zipp-3.19.2.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+zipp-3.19.2.dist-info/top_level.txt,sha256=iAbdoSHfaGqBfVb2XuR9JqSQHCoOsOtG6y9C_LSpqFw,5
+zipp/__init__.py,sha256=QuI1g00G4fRAcGt-HqbV0oWIkmSgedCGGYsHHYzNa8A,13412
+zipp/__pycache__/__init__.cpython-312.pyc,,
+zipp/__pycache__/glob.cpython-312.pyc,,
+zipp/compat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+zipp/compat/__pycache__/__init__.cpython-312.pyc,,
+zipp/compat/__pycache__/py310.cpython-312.pyc,,
+zipp/compat/py310.py,sha256=eZpkW0zRtunkhEh8jjX3gCGe22emoKCBJw72Zt4RkhA,219
+zipp/glob.py,sha256=etWpnfEoRyfUvrUsi6sTiGmErvPwe6HzY6pT8jg_lUI,3082
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/WHEEL
new file mode 100644
index 00000000..bab98d67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.43.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/top_level.txt
new file mode 100644
index 00000000..e82f676f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp-3.19.2.dist-info/top_level.txt
@@ -0,0 +1 @@
+zipp
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__init__.py
new file mode 100644
index 00000000..d65297b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__init__.py
@@ -0,0 +1,501 @@
+import io
+import posixpath
+import zipfile
+import itertools
+import contextlib
+import pathlib
+import re
+import stat
+import sys
+
+from .compat.py310 import text_encoding
+from .glob import Translator
+
+
+__all__ = ['Path']
+
+
+def _parents(path):
+    """
+    Given a path with elements separated by
+    posixpath.sep, generate all parents of that path.
+
+    >>> list(_parents('b/d'))
+    ['b']
+    >>> list(_parents('/b/d/'))
+    ['/b']
+    >>> list(_parents('b/d/f/'))
+    ['b/d', 'b']
+    >>> list(_parents('b'))
+    []
+    >>> list(_parents(''))
+    []
+    """
+    return itertools.islice(_ancestry(path), 1, None)
+
+
+def _ancestry(path):
+    """
+    Given a path with elements separated by
+    posixpath.sep, generate all elements of that path
+
+    >>> list(_ancestry('b/d'))
+    ['b/d', 'b']
+    >>> list(_ancestry('/b/d/'))
+    ['/b/d', '/b']
+    >>> list(_ancestry('b/d/f/'))
+    ['b/d/f', 'b/d', 'b']
+    >>> list(_ancestry('b'))
+    ['b']
+    >>> list(_ancestry(''))
+    []
+    """
+    path = path.rstrip(posixpath.sep)
+    while path and path != posixpath.sep:
+        yield path
+        path, tail = posixpath.split(path)
+
+
+_dedupe = dict.fromkeys
+"""Deduplicate an iterable in original order"""
+
+
+def _difference(minuend, subtrahend):
+    """
+    Return items in minuend not in subtrahend, retaining order
+    with O(1) lookup.
+    """
+    return itertools.filterfalse(set(subtrahend).__contains__, minuend)
+
+
+class InitializedState:
+    """
+    Mix-in to save the initialization state for pickling.
+    """
+
+    def __init__(self, *args, **kwargs):
+        self.__args = args
+        self.__kwargs = kwargs
+        super().__init__(*args, **kwargs)
+
+    def __getstate__(self):
+        return self.__args, self.__kwargs
+
+    def __setstate__(self, state):
+        args, kwargs = state
+        super().__init__(*args, **kwargs)
+
+
+class SanitizedNames:
+    """
+    ZipFile mix-in to ensure names are sanitized.
+    """
+
+    def namelist(self):
+        return list(map(self._sanitize, super().namelist()))
+
+    @staticmethod
+    def _sanitize(name):
+        r"""
+        Ensure a relative path with posix separators and no dot names.
+
+        Modeled after
+        https://github.com/python/cpython/blob/bcc1be39cb1d04ad9fc0bd1b9193d3972835a57c/Lib/zipfile/__init__.py#L1799-L1813
+        but provides consistent cross-platform behavior.
+
+        >>> san = SanitizedNames._sanitize
+        >>> san('/foo/bar')
+        'foo/bar'
+        >>> san('//foo.txt')
+        'foo.txt'
+        >>> san('foo/.././bar.txt')
+        'foo/bar.txt'
+        >>> san('foo../.bar.txt')
+        'foo../.bar.txt'
+        >>> san('\\foo\\bar.txt')
+        'foo/bar.txt'
+        >>> san('D:\\foo.txt')
+        'D/foo.txt'
+        >>> san('\\\\server\\share\\file.txt')
+        'server/share/file.txt'
+        >>> san('\\\\?\\GLOBALROOT\\Volume3')
+        '?/GLOBALROOT/Volume3'
+        >>> san('\\\\.\\PhysicalDrive1\\root')
+        'PhysicalDrive1/root'
+
+        Retain any trailing slash.
+        >>> san('abc/')
+        'abc/'
+
+        Raises a ValueError if the result is empty.
+        >>> san('../..')
+        Traceback (most recent call last):
+        ...
+        ValueError: Empty filename
+        """
+
+        def allowed(part):
+            return part and part not in {'..', '.'}
+
+        # Remove the drive letter.
+        # Don't use ntpath.splitdrive, because that also strips UNC paths
+        bare = re.sub('^([A-Z]):', r'\1', name, flags=re.IGNORECASE)
+        clean = bare.replace('\\', '/')
+        parts = clean.split('/')
+        joined = '/'.join(filter(allowed, parts))
+        if not joined:
+            raise ValueError("Empty filename")
+        return joined + '/' * name.endswith('/')
+
+
+class CompleteDirs(InitializedState, SanitizedNames, zipfile.ZipFile):
+    """
+    A ZipFile subclass that ensures that implied directories
+    are always included in the namelist.
+
+    >>> list(CompleteDirs._implied_dirs(['foo/bar.txt', 'foo/bar/baz.txt']))
+    ['foo/', 'foo/bar/']
+    >>> list(CompleteDirs._implied_dirs(['foo/bar.txt', 'foo/bar/baz.txt', 'foo/bar/']))
+    ['foo/']
+    """
+
+    @staticmethod
+    def _implied_dirs(names):
+        parents = itertools.chain.from_iterable(map(_parents, names))
+        as_dirs = (p + posixpath.sep for p in parents)
+        return _dedupe(_difference(as_dirs, names))
+
+    def namelist(self):
+        names = super().namelist()
+        return names + list(self._implied_dirs(names))
+
+    def _name_set(self):
+        return set(self.namelist())
+
+    def resolve_dir(self, name):
+        """
+        If the name represents a directory, return that name
+        as a directory (with the trailing slash).
+        """
+        names = self._name_set()
+        dirname = name + '/'
+        dir_match = name not in names and dirname in names
+        return dirname if dir_match else name
+
+    def getinfo(self, name):
+        """
+        Supplement getinfo for implied dirs.
+        """
+        try:
+            return super().getinfo(name)
+        except KeyError:
+            if not name.endswith('/') or name not in self._name_set():
+                raise
+            return zipfile.ZipInfo(filename=name)
+
+    @classmethod
+    def make(cls, source):
+        """
+        Given a source (filename or zipfile), return an
+        appropriate CompleteDirs subclass.
+        """
+        if isinstance(source, CompleteDirs):
+            return source
+
+        if not isinstance(source, zipfile.ZipFile):
+            return cls(source)
+
+        # Only allow for FastLookup when supplied zipfile is read-only
+        if 'r' not in source.mode:
+            cls = CompleteDirs
+
+        source.__class__ = cls
+        return source
+
+    @classmethod
+    def inject(cls, zf: zipfile.ZipFile) -> zipfile.ZipFile:
+        """
+        Given a writable zip file zf, inject directory entries for
+        any directories implied by the presence of children.
+        """
+        for name in cls._implied_dirs(zf.namelist()):
+            zf.writestr(name, b"")
+        return zf
+
+
+class FastLookup(CompleteDirs):
+    """
+    ZipFile subclass to ensure implicit
+    dirs exist and are resolved rapidly.
+    """
+
+    def namelist(self):
+        with contextlib.suppress(AttributeError):
+            return self.__names
+        self.__names = super().namelist()
+        return self.__names
+
+    def _name_set(self):
+        with contextlib.suppress(AttributeError):
+            return self.__lookup
+        self.__lookup = super()._name_set()
+        return self.__lookup
+
+
+def _extract_text_encoding(encoding=None, *args, **kwargs):
+    # compute stack level so that the caller of the caller sees any warning.
+    is_pypy = sys.implementation.name == 'pypy'
+    stack_level = 3 + is_pypy
+    return text_encoding(encoding, stack_level), args, kwargs
+
+
+class Path:
+    """
+    A :class:`importlib.resources.abc.Traversable` interface for zip files.
+
+    Implements many of the features users enjoy from
+    :class:`pathlib.Path`.
+
+    Consider a zip file with this structure::
+
+        .
+        ├── a.txt
+        └── b
+            ├── c.txt
+            └── d
+                └── e.txt
+
+    >>> data = io.BytesIO()
+    >>> zf = zipfile.ZipFile(data, 'w')
+    >>> zf.writestr('a.txt', 'content of a')
+    >>> zf.writestr('b/c.txt', 'content of c')
+    >>> zf.writestr('b/d/e.txt', 'content of e')
+    >>> zf.filename = 'mem/abcde.zip'
+
+    Path accepts the zipfile object itself or a filename
+
+    >>> path = Path(zf)
+
+    From there, several path operations are available.
+
+    Directory iteration (including the zip file itself):
+
+    >>> a, b = path.iterdir()
+    >>> a
+    Path('mem/abcde.zip', 'a.txt')
+    >>> b
+    Path('mem/abcde.zip', 'b/')
+
+    name property:
+
+    >>> b.name
+    'b'
+
+    join with divide operator:
+
+    >>> c = b / 'c.txt'
+    >>> c
+    Path('mem/abcde.zip', 'b/c.txt')
+    >>> c.name
+    'c.txt'
+
+    Read text:
+
+    >>> c.read_text(encoding='utf-8')
+    'content of c'
+
+    existence:
+
+    >>> c.exists()
+    True
+    >>> (b / 'missing.txt').exists()
+    False
+
+    Coercion to string:
+
+    >>> import os
+    >>> str(c).replace(os.sep, posixpath.sep)
+    'mem/abcde.zip/b/c.txt'
+
+    At the root, ``name``, ``filename``, and ``parent``
+    resolve to the zipfile.
+
+    >>> str(path)
+    'mem/abcde.zip/'
+    >>> path.name
+    'abcde.zip'
+    >>> path.filename == pathlib.Path('mem/abcde.zip')
+    True
+    >>> str(path.parent)
+    'mem'
+
+    If the zipfile has no filename, such attributes are not
+    valid and accessing them will raise an Exception.
+
+    >>> zf.filename = None
+    >>> path.name
+    Traceback (most recent call last):
+    ...
+    TypeError: ...
+
+    >>> path.filename
+    Traceback (most recent call last):
+    ...
+    TypeError: ...
+
+    >>> path.parent
+    Traceback (most recent call last):
+    ...
+    TypeError: ...
+
+    # workaround python/cpython#106763
+    >>> pass
+    """
+
+    __repr = "{self.__class__.__name__}({self.root.filename!r}, {self.at!r})"
+
+    def __init__(self, root, at=""):
+        """
+        Construct a Path from a ZipFile or filename.
+
+        Note: When the source is an existing ZipFile object,
+        its type (__class__) will be mutated to a
+        specialized type. If the caller wishes to retain the
+        original type, the caller should either create a
+        separate ZipFile object or pass a filename.
+        """
+        self.root = FastLookup.make(root)
+        self.at = at
+
+    def __eq__(self, other):
+        """
+        >>> Path(zipfile.ZipFile(io.BytesIO(), 'w')) == 'foo'
+        False
+        """
+        if self.__class__ is not other.__class__:
+            return NotImplemented
+        return (self.root, self.at) == (other.root, other.at)
+
+    def __hash__(self):
+        return hash((self.root, self.at))
+
+    def open(self, mode='r', *args, pwd=None, **kwargs):
+        """
+        Open this entry as text or binary following the semantics
+        of ``pathlib.Path.open()`` by passing arguments through
+        to io.TextIOWrapper().
+        """
+        if self.is_dir():
+            raise IsADirectoryError(self)
+        zip_mode = mode[0]
+        if not self.exists() and zip_mode == 'r':
+            raise FileNotFoundError(self)
+        stream = self.root.open(self.at, zip_mode, pwd=pwd)
+        if 'b' in mode:
+            if args or kwargs:
+                raise ValueError("encoding args invalid for binary operation")
+            return stream
+        # Text mode:
+        encoding, args, kwargs = _extract_text_encoding(*args, **kwargs)
+        return io.TextIOWrapper(stream, encoding, *args, **kwargs)
+
+    def _base(self):
+        return pathlib.PurePosixPath(self.at or self.root.filename)
+
+    @property
+    def name(self):
+        return self._base().name
+
+    @property
+    def suffix(self):
+        return self._base().suffix
+
+    @property
+    def suffixes(self):
+        return self._base().suffixes
+
+    @property
+    def stem(self):
+        return self._base().stem
+
+    @property
+    def filename(self):
+        return pathlib.Path(self.root.filename).joinpath(self.at)
+
+    def read_text(self, *args, **kwargs):
+        encoding, args, kwargs = _extract_text_encoding(*args, **kwargs)
+        with self.open('r', encoding, *args, **kwargs) as strm:
+            return strm.read()
+
+    def read_bytes(self):
+        with self.open('rb') as strm:
+            return strm.read()
+
+    def _is_child(self, path):
+        return posixpath.dirname(path.at.rstrip("/")) == self.at.rstrip("/")
+
+    def _next(self, at):
+        return self.__class__(self.root, at)
+
+    def is_dir(self):
+        return not self.at or self.at.endswith("/")
+
+    def is_file(self):
+        return self.exists() and not self.is_dir()
+
+    def exists(self):
+        return self.at in self.root._name_set()
+
+    def iterdir(self):
+        if not self.is_dir():
+            raise ValueError("Can't listdir a file")
+        subs = map(self._next, self.root.namelist())
+        return filter(self._is_child, subs)
+
+    def match(self, path_pattern):
+        return pathlib.PurePosixPath(self.at).match(path_pattern)
+
+    def is_symlink(self):
+        """
+        Return whether this path is a symlink.
+        """
+        info = self.root.getinfo(self.at)
+        mode = info.external_attr >> 16
+        return stat.S_ISLNK(mode)
+
+    def glob(self, pattern):
+        if not pattern:
+            raise ValueError(f"Unacceptable pattern: {pattern!r}")
+
+        prefix = re.escape(self.at)
+        tr = Translator(seps='/')
+        matches = re.compile(prefix + tr.translate(pattern)).fullmatch
+        names = (data.filename for data in self.root.filelist)
+        return map(self._next, filter(matches, names))
+
+    def rglob(self, pattern):
+        return self.glob(f'**/{pattern}')
+
+    def relative_to(self, other, *extra):
+        return posixpath.relpath(str(self), str(other.joinpath(*extra)))
+
+    def __str__(self):
+        return posixpath.join(self.root.filename, self.at)
+
+    def __repr__(self):
+        return self.__repr.format(self=self)
+
+    def joinpath(self, *other):
+        next = posixpath.join(self.at, *other)
+        return self._next(self.root.resolve_dir(next))
+
+    __truediv__ = joinpath
+
+    @property
+    def parent(self):
+        if not self.at:
+            return self.filename.parent
+        parent_at = posixpath.dirname(self.at.rstrip('/'))
+        if parent_at:
+            parent_at += '/'
+        return self._next(parent_at)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b72dcb53
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/glob.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/glob.cpython-312.pyc
new file mode 100644
index 00000000..610d1ca9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/__pycache__/glob.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..cc651c6d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/py310.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/py310.cpython-312.pyc
new file mode 100644
index 00000000..6ee35e65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/__pycache__/py310.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/py310.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/py310.py
new file mode 100644
index 00000000..d5ca53e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/compat/py310.py
@@ -0,0 +1,11 @@
+import sys
+import io
+
+
+def _text_encoding(encoding, stacklevel=2, /):  # pragma: no cover
+    return encoding
+
+
+text_encoding = (
+    io.text_encoding if sys.version_info > (3, 10) else _text_encoding  # type: ignore
+)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/glob.py b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/glob.py
new file mode 100644
index 00000000..69c41d77
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/_vendor/zipp/glob.py
@@ -0,0 +1,106 @@
+import os
+import re
+
+
+_default_seps = os.sep + str(os.altsep) * bool(os.altsep)
+
+
+class Translator:
+    """
+    >>> Translator('xyz')
+    Traceback (most recent call last):
+    ...
+    AssertionError: Invalid separators
+
+    >>> Translator('')
+    Traceback (most recent call last):
+    ...
+    AssertionError: Invalid separators
+    """
+
+    seps: str
+
+    def __init__(self, seps: str = _default_seps):
+        assert seps and set(seps) <= set(_default_seps), "Invalid separators"
+        self.seps = seps
+
+    def translate(self, pattern):
+        """
+        Given a glob pattern, produce a regex that matches it.
+        """
+        return self.extend(self.translate_core(pattern))
+
+    def extend(self, pattern):
+        r"""
+        Extend regex for pattern-wide concerns.
+
+        Apply '(?s:)' to create a non-matching group that
+        matches newlines (valid on Unix).
+
+        Append '\Z' to imply fullmatch even when match is used.
+        """
+        return rf'(?s:{pattern})\Z'
+
+    def translate_core(self, pattern):
+        r"""
+        Given a glob pattern, produce a regex that matches it.
+
+        >>> t = Translator()
+        >>> t.translate_core('*.txt').replace('\\\\', '')
+        '[^/]*\\.txt'
+        >>> t.translate_core('a?txt')
+        'a[^/]txt'
+        >>> t.translate_core('**/*').replace('\\\\', '')
+        '.*/[^/][^/]*'
+        """
+        self.restrict_rglob(pattern)
+        return ''.join(map(self.replace, separate(self.star_not_empty(pattern))))
+
+    def replace(self, match):
+        """
+        Perform the replacements for a match from :func:`separate`.
+        """
+        return match.group('set') or (
+            re.escape(match.group(0))
+            .replace('\\*\\*', r'.*')
+            .replace('\\*', rf'[^{re.escape(self.seps)}]*')
+            .replace('\\?', r'[^/]')
+        )
+
+    def restrict_rglob(self, pattern):
+        """
+        Raise ValueError if ** appears in anything but a full path segment.
+
+        >>> Translator().translate('**foo')
+        Traceback (most recent call last):
+        ...
+        ValueError: ** must appear alone in a path segment
+        """
+        seps_pattern = rf'[{re.escape(self.seps)}]+'
+        segments = re.split(seps_pattern, pattern)
+        if any('**' in segment and segment != '**' for segment in segments):
+            raise ValueError("** must appear alone in a path segment")
+
+    def star_not_empty(self, pattern):
+        """
+        Ensure that * will not match an empty segment.
+        """
+
+        def handle_segment(match):
+            segment = match.group(0)
+            return '?*' if segment == '*' else segment
+
+        not_seps_pattern = rf'[^{re.escape(self.seps)}]+'
+        return re.sub(not_seps_pattern, handle_segment, pattern)
+
+
+def separate(pattern):
+    """
+    Separate out character sets to avoid translating their contents.
+
+    >>> [m.group(0) for m in separate('*.txt')]
+    ['*.txt']
+    >>> [m.group(0) for m in separate('a[?]txt')]
+    ['a', '[?]', 'txt']
+    """
+    return re.finditer(r'([^\[]+)|(?P<set>[\[].*?[\]])|([\[][^\]]*$)', pattern)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/archive_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/archive_util.py
new file mode 100644
index 00000000..1a02010b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/archive_util.py
@@ -0,0 +1,219 @@
+"""Utilities for extracting common archive formats"""
+
+import contextlib
+import os
+import posixpath
+import shutil
+import tarfile
+import zipfile
+
+from ._path import ensure_directory
+
+from distutils.errors import DistutilsError
+
+__all__ = [
+    "unpack_archive",
+    "unpack_zipfile",
+    "unpack_tarfile",
+    "default_filter",
+    "UnrecognizedFormat",
+    "extraction_drivers",
+    "unpack_directory",
+]
+
+
+class UnrecognizedFormat(DistutilsError):
+    """Couldn't recognize the archive type"""
+
+
+def default_filter(src, dst):
+    """The default progress/filter callback; returns True for all files"""
+    return dst
+
+
+def unpack_archive(
+    filename, extract_dir, progress_filter=default_filter, drivers=None
+) -> None:
+    """Unpack `filename` to `extract_dir`, or raise ``UnrecognizedFormat``
+
+    `progress_filter` is a function taking two arguments: a source path
+    internal to the archive ('/'-separated), and a filesystem path where it
+    will be extracted.  The callback must return the desired extract path
+    (which may be the same as the one passed in), or else ``None`` to skip
+    that file or directory.  The callback can thus be used to report on the
+    progress of the extraction, as well as to filter the items extracted or
+    alter their extraction paths.
+
+    `drivers`, if supplied, must be a non-empty sequence of functions with the
+    same signature as this function (minus the `drivers` argument), that raise
+    ``UnrecognizedFormat`` if they do not support extracting the designated
+    archive type.  The `drivers` are tried in sequence until one is found that
+    does not raise an error, or until all are exhausted (in which case
+    ``UnrecognizedFormat`` is raised).  If you do not supply a sequence of
+    drivers, the module's ``extraction_drivers`` constant will be used, which
+    means that ``unpack_zipfile`` and ``unpack_tarfile`` will be tried, in that
+    order.
+    """
+    for driver in drivers or extraction_drivers:
+        try:
+            driver(filename, extract_dir, progress_filter)
+        except UnrecognizedFormat:
+            continue
+        else:
+            return
+    else:
+        raise UnrecognizedFormat(f"Not a recognized archive type: {filename}")
+
+
+def unpack_directory(filename, extract_dir, progress_filter=default_filter) -> None:
+    """ "Unpack" a directory, using the same interface as for archives
+
+    Raises ``UnrecognizedFormat`` if `filename` is not a directory
+    """
+    if not os.path.isdir(filename):
+        raise UnrecognizedFormat(f"{filename} is not a directory")
+
+    paths = {
+        filename: ('', extract_dir),
+    }
+    for base, dirs, files in os.walk(filename):
+        src, dst = paths[base]
+        for d in dirs:
+            paths[os.path.join(base, d)] = src + d + '/', os.path.join(dst, d)
+        for f in files:
+            target = os.path.join(dst, f)
+            target = progress_filter(src + f, target)
+            if not target:
+                # skip non-files
+                continue
+            ensure_directory(target)
+            f = os.path.join(base, f)
+            shutil.copyfile(f, target)
+            shutil.copystat(f, target)
+
+
+def unpack_zipfile(filename, extract_dir, progress_filter=default_filter) -> None:
+    """Unpack zip `filename` to `extract_dir`
+
+    Raises ``UnrecognizedFormat`` if `filename` is not a zipfile (as determined
+    by ``zipfile.is_zipfile()``).  See ``unpack_archive()`` for an explanation
+    of the `progress_filter` argument.
+    """
+
+    if not zipfile.is_zipfile(filename):
+        raise UnrecognizedFormat(f"{filename} is not a zip file")
+
+    with zipfile.ZipFile(filename) as z:
+        _unpack_zipfile_obj(z, extract_dir, progress_filter)
+
+
+def _unpack_zipfile_obj(zipfile_obj, extract_dir, progress_filter=default_filter):
+    """Internal/private API used by other parts of setuptools.
+    Similar to ``unpack_zipfile``, but receives an already opened :obj:`zipfile.ZipFile`
+    object instead of a filename.
+    """
+    for info in zipfile_obj.infolist():
+        name = info.filename
+
+        # don't extract absolute paths or ones with .. in them
+        if name.startswith('/') or '..' in name.split('/'):
+            continue
+
+        target = os.path.join(extract_dir, *name.split('/'))
+        target = progress_filter(name, target)
+        if not target:
+            continue
+        if name.endswith('/'):
+            # directory
+            ensure_directory(target)
+        else:
+            # file
+            ensure_directory(target)
+            data = zipfile_obj.read(info.filename)
+            with open(target, 'wb') as f:
+                f.write(data)
+        unix_attributes = info.external_attr >> 16
+        if unix_attributes:
+            os.chmod(target, unix_attributes)
+
+
+def _resolve_tar_file_or_dir(tar_obj, tar_member_obj):
+    """Resolve any links and extract link targets as normal files."""
+    while tar_member_obj is not None and (
+        tar_member_obj.islnk() or tar_member_obj.issym()
+    ):
+        linkpath = tar_member_obj.linkname
+        if tar_member_obj.issym():
+            base = posixpath.dirname(tar_member_obj.name)
+            linkpath = posixpath.join(base, linkpath)
+            linkpath = posixpath.normpath(linkpath)
+        tar_member_obj = tar_obj._getmember(linkpath)
+
+    is_file_or_dir = tar_member_obj is not None and (
+        tar_member_obj.isfile() or tar_member_obj.isdir()
+    )
+    if is_file_or_dir:
+        return tar_member_obj
+
+    raise LookupError('Got unknown file type')
+
+
+def _iter_open_tar(tar_obj, extract_dir, progress_filter):
+    """Emit member-destination pairs from a tar archive."""
+    # don't do any chowning!
+    tar_obj.chown = lambda *args: None
+
+    with contextlib.closing(tar_obj):
+        for member in tar_obj:
+            name = member.name
+            # don't extract absolute paths or ones with .. in them
+            if name.startswith('/') or '..' in name.split('/'):
+                continue
+
+            prelim_dst = os.path.join(extract_dir, *name.split('/'))
+
+            try:
+                member = _resolve_tar_file_or_dir(tar_obj, member)
+            except LookupError:
+                continue
+
+            final_dst = progress_filter(name, prelim_dst)
+            if not final_dst:
+                continue
+
+            if final_dst.endswith(os.sep):
+                final_dst = final_dst[:-1]
+
+            yield member, final_dst
+
+
+def unpack_tarfile(filename, extract_dir, progress_filter=default_filter) -> bool:
+    """Unpack tar/tar.gz/tar.bz2 `filename` to `extract_dir`
+
+    Raises ``UnrecognizedFormat`` if `filename` is not a tarfile (as determined
+    by ``tarfile.open()``).  See ``unpack_archive()`` for an explanation
+    of the `progress_filter` argument.
+    """
+    try:
+        tarobj = tarfile.open(filename)
+    except tarfile.TarError as e:
+        raise UnrecognizedFormat(
+            f"{filename} is not a compressed or uncompressed tar file"
+        ) from e
+
+    for member, final_dst in _iter_open_tar(
+        tarobj,
+        extract_dir,
+        progress_filter,
+    ):
+        try:
+            # XXX Ugh
+            tarobj._extract_member(member, final_dst)
+        except tarfile.ExtractError:
+            # chown/chmod/mkfifo/mknode/makedev failed
+            pass
+
+    return True
+
+
+extraction_drivers = unpack_directory, unpack_zipfile, unpack_tarfile
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/build_meta.py b/backend/venv/lib/python3.12/site-packages/setuptools/build_meta.py
new file mode 100644
index 00000000..00fa5e1f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/build_meta.py
@@ -0,0 +1,560 @@
+"""A PEP 517 interface to setuptools
+
+Previously, when a user or a command line tool (let's call it a "frontend")
+needed to make a request of setuptools to take a certain action, for
+example, generating a list of installation requirements, the frontend
+would call "setup.py egg_info" or "setup.py bdist_wheel" on the command line.
+
+PEP 517 defines a different method of interfacing with setuptools. Rather
+than calling "setup.py" directly, the frontend should:
+
+  1. Set the current directory to the directory with a setup.py file
+  2. Import this module into a safe python interpreter (one in which
+     setuptools can potentially set global variables or crash hard).
+  3. Call one of the functions defined in PEP 517.
+
+What each function does is defined in PEP 517. However, here is a "casual"
+definition of the functions (this definition should not be relied on for
+bug reports or API stability):
+
+  - `build_wheel`: build a wheel in the folder and return the basename
+  - `get_requires_for_build_wheel`: get the `setup_requires` to build
+  - `prepare_metadata_for_build_wheel`: get the `install_requires`
+  - `build_sdist`: build an sdist in the folder and return the basename
+  - `get_requires_for_build_sdist`: get the `setup_requires` to build
+
+Again, this is not a formal definition! Just a "taste" of the module.
+"""
+
+from __future__ import annotations
+
+import contextlib
+import io
+import os
+import shlex
+import shutil
+import sys
+import tempfile
+import tokenize
+import warnings
+from collections.abc import Iterable, Iterator, Mapping
+from pathlib import Path
+from typing import TYPE_CHECKING, Union
+
+import setuptools
+
+from . import errors
+from ._path import StrPath, same_path
+from ._reqs import parse_strings
+from .warnings import SetuptoolsDeprecationWarning
+
+import distutils
+from distutils.util import strtobool
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+__all__ = [
+    'get_requires_for_build_sdist',
+    'get_requires_for_build_wheel',
+    'prepare_metadata_for_build_wheel',
+    'build_wheel',
+    'build_sdist',
+    'get_requires_for_build_editable',
+    'prepare_metadata_for_build_editable',
+    'build_editable',
+    '__legacy__',
+    'SetupRequirementsError',
+]
+
+SETUPTOOLS_ENABLE_FEATURES = os.getenv("SETUPTOOLS_ENABLE_FEATURES", "").lower()
+LEGACY_EDITABLE = "legacy-editable" in SETUPTOOLS_ENABLE_FEATURES.replace("_", "-")
+
+
+class SetupRequirementsError(BaseException):
+    def __init__(self, specifiers) -> None:
+        self.specifiers = specifiers
+
+
+class Distribution(setuptools.dist.Distribution):
+    def fetch_build_eggs(self, specifiers):
+        specifier_list = list(parse_strings(specifiers))
+
+        raise SetupRequirementsError(specifier_list)
+
+    @classmethod
+    @contextlib.contextmanager
+    def patch(cls):
+        """
+        Replace
+        distutils.dist.Distribution with this class
+        for the duration of this context.
+        """
+        orig = distutils.core.Distribution
+        distutils.core.Distribution = cls  # type: ignore[misc] # monkeypatching
+        try:
+            yield
+        finally:
+            distutils.core.Distribution = orig  # type: ignore[misc] # monkeypatching
+
+
+@contextlib.contextmanager
+def no_install_setup_requires():
+    """Temporarily disable installing setup_requires
+
+    Under PEP 517, the backend reports build dependencies to the frontend,
+    and the frontend is responsible for ensuring they're installed.
+    So setuptools (acting as a backend) should not try to install them.
+    """
+    orig = setuptools._install_setup_requires
+    setuptools._install_setup_requires = lambda attrs: None
+    try:
+        yield
+    finally:
+        setuptools._install_setup_requires = orig
+
+
+def _get_immediate_subdirectories(a_dir):
+    return [
+        name for name in os.listdir(a_dir) if os.path.isdir(os.path.join(a_dir, name))
+    ]
+
+
+def _file_with_extension(directory: StrPath, extension: str | tuple[str, ...]):
+    matching = (f for f in os.listdir(directory) if f.endswith(extension))
+    try:
+        (file,) = matching
+    except ValueError:
+        raise ValueError(
+            'No distribution was found. Ensure that `setup.py` '
+            'is not empty and that it calls `setup()`.'
+        ) from None
+    return file
+
+
+def _open_setup_script(setup_script):
+    if not os.path.exists(setup_script):
+        # Supply a default setup.py
+        return io.StringIO("from setuptools import setup; setup()")
+
+    return tokenize.open(setup_script)
+
+
+@contextlib.contextmanager
+def suppress_known_deprecation():
+    with warnings.catch_warnings():
+        warnings.filterwarnings('ignore', 'setup.py install is deprecated')
+        yield
+
+
+_ConfigSettings: TypeAlias = Union[Mapping[str, Union[str, list[str], None]], None]
+"""
+Currently the user can run::
+
+    pip install -e . --config-settings key=value
+    python -m build -C--key=value -C key=value
+
+- pip will pass both key and value as strings and overwriting repeated keys
+  (pypa/pip#11059).
+- build will accumulate values associated with repeated keys in a list.
+  It will also accept keys with no associated value.
+  This means that an option passed by build can be ``str | list[str] | None``.
+- PEP 517 specifies that ``config_settings`` is an optional dict.
+"""
+
+
+class _ConfigSettingsTranslator:
+    """Translate ``config_settings`` into distutils-style command arguments.
+    Only a limited number of options is currently supported.
+    """
+
+    # See pypa/setuptools#1928 pypa/setuptools#2491
+
+    def _get_config(self, key: str, config_settings: _ConfigSettings) -> list[str]:
+        """
+        Get the value of a specific key in ``config_settings`` as a list of strings.
+
+        >>> fn = _ConfigSettingsTranslator()._get_config
+        >>> fn("--global-option", None)
+        []
+        >>> fn("--global-option", {})
+        []
+        >>> fn("--global-option", {'--global-option': 'foo'})
+        ['foo']
+        >>> fn("--global-option", {'--global-option': ['foo']})
+        ['foo']
+        >>> fn("--global-option", {'--global-option': 'foo'})
+        ['foo']
+        >>> fn("--global-option", {'--global-option': 'foo bar'})
+        ['foo', 'bar']
+        """
+        cfg = config_settings or {}
+        opts = cfg.get(key) or []
+        return shlex.split(opts) if isinstance(opts, str) else opts
+
+    def _global_args(self, config_settings: _ConfigSettings) -> Iterator[str]:
+        """
+        Let the user specify ``verbose`` or ``quiet`` + escape hatch via
+        ``--global-option``.
+        Note: ``-v``, ``-vv``, ``-vvv`` have similar effects in setuptools,
+        so we just have to cover the basic scenario ``-v``.
+
+        >>> fn = _ConfigSettingsTranslator()._global_args
+        >>> list(fn(None))
+        []
+        >>> list(fn({"verbose": "False"}))
+        ['-q']
+        >>> list(fn({"verbose": "1"}))
+        ['-v']
+        >>> list(fn({"--verbose": None}))
+        ['-v']
+        >>> list(fn({"verbose": "true", "--global-option": "-q --no-user-cfg"}))
+        ['-v', '-q', '--no-user-cfg']
+        >>> list(fn({"--quiet": None}))
+        ['-q']
+        """
+        cfg = config_settings or {}
+        falsey = {"false", "no", "0", "off"}
+        if "verbose" in cfg or "--verbose" in cfg:
+            level = str(cfg.get("verbose") or cfg.get("--verbose") or "1")
+            yield ("-q" if level.lower() in falsey else "-v")
+        if "quiet" in cfg or "--quiet" in cfg:
+            level = str(cfg.get("quiet") or cfg.get("--quiet") or "1")
+            yield ("-v" if level.lower() in falsey else "-q")
+
+        yield from self._get_config("--global-option", config_settings)
+
+    def __dist_info_args(self, config_settings: _ConfigSettings) -> Iterator[str]:
+        """
+        The ``dist_info`` command accepts ``tag-date`` and ``tag-build``.
+
+        .. warning::
+           We cannot use this yet as it requires the ``sdist`` and ``bdist_wheel``
+           commands run in ``build_sdist`` and ``build_wheel`` to reuse the egg-info
+           directory created in ``prepare_metadata_for_build_wheel``.
+
+        >>> fn = _ConfigSettingsTranslator()._ConfigSettingsTranslator__dist_info_args
+        >>> list(fn(None))
+        []
+        >>> list(fn({"tag-date": "False"}))
+        ['--no-date']
+        >>> list(fn({"tag-date": None}))
+        ['--no-date']
+        >>> list(fn({"tag-date": "true", "tag-build": ".a"}))
+        ['--tag-date', '--tag-build', '.a']
+        """
+        cfg = config_settings or {}
+        if "tag-date" in cfg:
+            val = strtobool(str(cfg["tag-date"] or "false"))
+            yield ("--tag-date" if val else "--no-date")
+        if "tag-build" in cfg:
+            yield from ["--tag-build", str(cfg["tag-build"])]
+
+    def _editable_args(self, config_settings: _ConfigSettings) -> Iterator[str]:
+        """
+        The ``editable_wheel`` command accepts ``editable-mode=strict``.
+
+        >>> fn = _ConfigSettingsTranslator()._editable_args
+        >>> list(fn(None))
+        []
+        >>> list(fn({"editable-mode": "strict"}))
+        ['--mode', 'strict']
+        """
+        cfg = config_settings or {}
+        mode = cfg.get("editable-mode") or cfg.get("editable_mode")
+        if not mode:
+            return
+        yield from ["--mode", str(mode)]
+
+    def _arbitrary_args(self, config_settings: _ConfigSettings) -> Iterator[str]:
+        """
+        Users may expect to pass arbitrary lists of arguments to a command
+        via "--global-option" (example provided in PEP 517 of a "escape hatch").
+
+        >>> fn = _ConfigSettingsTranslator()._arbitrary_args
+        >>> list(fn(None))
+        []
+        >>> list(fn({}))
+        []
+        >>> list(fn({'--build-option': 'foo'}))
+        ['foo']
+        >>> list(fn({'--build-option': ['foo']}))
+        ['foo']
+        >>> list(fn({'--build-option': 'foo'}))
+        ['foo']
+        >>> list(fn({'--build-option': 'foo bar'}))
+        ['foo', 'bar']
+        >>> list(fn({'--global-option': 'foo'}))
+        []
+        """
+        yield from self._get_config("--build-option", config_settings)
+
+
+class _BuildMetaBackend(_ConfigSettingsTranslator):
+    def _get_build_requires(
+        self, config_settings: _ConfigSettings, requirements: list[str]
+    ):
+        sys.argv = [
+            *sys.argv[:1],
+            *self._global_args(config_settings),
+            "egg_info",
+        ]
+        try:
+            with Distribution.patch():
+                self.run_setup()
+        except SetupRequirementsError as e:
+            requirements += e.specifiers
+
+        return requirements
+
+    def run_setup(self, setup_script: str = 'setup.py'):
+        # Note that we can reuse our build directory between calls
+        # Correctness comes first, then optimization later
+        __file__ = os.path.abspath(setup_script)
+        __name__ = '__main__'
+
+        with _open_setup_script(__file__) as f:
+            code = f.read().replace(r'\r\n', r'\n')
+
+        try:
+            exec(code, locals())
+        except SystemExit as e:
+            if e.code:
+                raise
+            # We ignore exit code indicating success
+            SetuptoolsDeprecationWarning.emit(
+                "Running `setup.py` directly as CLI tool is deprecated.",
+                "Please avoid using `sys.exit(0)` or similar statements "
+                "that don't fit in the paradigm of a configuration file.",
+                see_url="https://blog.ganssle.io/articles/2021/10/"
+                "setup-py-deprecated.html",
+            )
+
+    def get_requires_for_build_wheel(self, config_settings: _ConfigSettings = None):
+        return self._get_build_requires(config_settings, requirements=[])
+
+    def get_requires_for_build_sdist(self, config_settings: _ConfigSettings = None):
+        return self._get_build_requires(config_settings, requirements=[])
+
+    def _bubble_up_info_directory(
+        self, metadata_directory: StrPath, suffix: str
+    ) -> str:
+        """
+        PEP 517 requires that the .dist-info directory be placed in the
+        metadata_directory. To comply, we MUST copy the directory to the root.
+
+        Returns the basename of the info directory, e.g. `proj-0.0.0.dist-info`.
+        """
+        info_dir = self._find_info_directory(metadata_directory, suffix)
+        if not same_path(info_dir.parent, metadata_directory):
+            shutil.move(str(info_dir), metadata_directory)
+            # PEP 517 allow other files and dirs to exist in metadata_directory
+        return info_dir.name
+
+    def _find_info_directory(self, metadata_directory: StrPath, suffix: str) -> Path:
+        for parent, dirs, _ in os.walk(metadata_directory):
+            candidates = [f for f in dirs if f.endswith(suffix)]
+
+            if len(candidates) != 0 or len(dirs) != 1:
+                assert len(candidates) == 1, f"Multiple {suffix} directories found"
+                return Path(parent, candidates[0])
+
+        msg = f"No {suffix} directory found in {metadata_directory}"
+        raise errors.InternalError(msg)
+
+    def prepare_metadata_for_build_wheel(
+        self, metadata_directory: StrPath, config_settings: _ConfigSettings = None
+    ):
+        sys.argv = [
+            *sys.argv[:1],
+            *self._global_args(config_settings),
+            "dist_info",
+            "--output-dir",
+            str(metadata_directory),
+            "--keep-egg-info",
+        ]
+        with no_install_setup_requires():
+            self.run_setup()
+
+        self._bubble_up_info_directory(metadata_directory, ".egg-info")
+        return self._bubble_up_info_directory(metadata_directory, ".dist-info")
+
+    def _build_with_temp_dir(
+        self,
+        setup_command: Iterable[str],
+        result_extension: str | tuple[str, ...],
+        result_directory: StrPath,
+        config_settings: _ConfigSettings,
+        arbitrary_args: Iterable[str] = (),
+    ):
+        result_directory = os.path.abspath(result_directory)
+
+        # Build in a temporary directory, then copy to the target.
+        os.makedirs(result_directory, exist_ok=True)
+
+        with tempfile.TemporaryDirectory(
+            prefix=".tmp-", dir=result_directory
+        ) as tmp_dist_dir:
+            sys.argv = [
+                *sys.argv[:1],
+                *self._global_args(config_settings),
+                *setup_command,
+                "--dist-dir",
+                tmp_dist_dir,
+                *arbitrary_args,
+            ]
+            with no_install_setup_requires():
+                self.run_setup()
+
+            result_basename = _file_with_extension(tmp_dist_dir, result_extension)
+            result_path = os.path.join(result_directory, result_basename)
+            if os.path.exists(result_path):
+                # os.rename will fail overwriting on non-Unix.
+                os.remove(result_path)
+            os.rename(os.path.join(tmp_dist_dir, result_basename), result_path)
+
+        return result_basename
+
+    def build_wheel(
+        self,
+        wheel_directory: StrPath,
+        config_settings: _ConfigSettings = None,
+        metadata_directory: StrPath | None = None,
+    ):
+        def _build(cmd: list[str]):
+            with suppress_known_deprecation():
+                return self._build_with_temp_dir(
+                    cmd,
+                    '.whl',
+                    wheel_directory,
+                    config_settings,
+                    self._arbitrary_args(config_settings),
+                )
+
+        if metadata_directory is None:
+            return _build(['bdist_wheel'])
+
+        try:
+            return _build(['bdist_wheel', '--dist-info-dir', str(metadata_directory)])
+        except SystemExit as ex:  # pragma: nocover
+            # pypa/setuptools#4683
+            if "--dist-info-dir not recognized" not in str(ex):
+                raise
+            _IncompatibleBdistWheel.emit()
+            return _build(['bdist_wheel'])
+
+    def build_sdist(
+        self, sdist_directory: StrPath, config_settings: _ConfigSettings = None
+    ):
+        return self._build_with_temp_dir(
+            ['sdist', '--formats', 'gztar'], '.tar.gz', sdist_directory, config_settings
+        )
+
+    def _get_dist_info_dir(self, metadata_directory: StrPath | None) -> str | None:
+        if not metadata_directory:
+            return None
+        dist_info_candidates = list(Path(metadata_directory).glob("*.dist-info"))
+        assert len(dist_info_candidates) <= 1
+        return str(dist_info_candidates[0]) if dist_info_candidates else None
+
+    if not LEGACY_EDITABLE:
+        # PEP660 hooks:
+        # build_editable
+        # get_requires_for_build_editable
+        # prepare_metadata_for_build_editable
+        def build_editable(
+            self,
+            wheel_directory: StrPath,
+            config_settings: _ConfigSettings = None,
+            metadata_directory: StrPath | None = None,
+        ):
+            # XXX can or should we hide our editable_wheel command normally?
+            info_dir = self._get_dist_info_dir(metadata_directory)
+            opts = ["--dist-info-dir", info_dir] if info_dir else []
+            cmd = ["editable_wheel", *opts, *self._editable_args(config_settings)]
+            with suppress_known_deprecation():
+                return self._build_with_temp_dir(
+                    cmd, ".whl", wheel_directory, config_settings
+                )
+
+        def get_requires_for_build_editable(
+            self, config_settings: _ConfigSettings = None
+        ):
+            return self.get_requires_for_build_wheel(config_settings)
+
+        def prepare_metadata_for_build_editable(
+            self, metadata_directory: StrPath, config_settings: _ConfigSettings = None
+        ):
+            return self.prepare_metadata_for_build_wheel(
+                metadata_directory, config_settings
+            )
+
+
+class _BuildMetaLegacyBackend(_BuildMetaBackend):
+    """Compatibility backend for setuptools
+
+    This is a version of setuptools.build_meta that endeavors
+    to maintain backwards
+    compatibility with pre-PEP 517 modes of invocation. It
+    exists as a temporary
+    bridge between the old packaging mechanism and the new
+    packaging mechanism,
+    and will eventually be removed.
+    """
+
+    def run_setup(self, setup_script: str = 'setup.py'):
+        # In order to maintain compatibility with scripts assuming that
+        # the setup.py script is in a directory on the PYTHONPATH, inject
+        # '' into sys.path. (pypa/setuptools#1642)
+        sys_path = list(sys.path)  # Save the original path
+
+        script_dir = os.path.dirname(os.path.abspath(setup_script))
+        if script_dir not in sys.path:
+            sys.path.insert(0, script_dir)
+
+        # Some setup.py scripts (e.g. in pygame and numpy) use sys.argv[0] to
+        # get the directory of the source code. They expect it to refer to the
+        # setup.py script.
+        sys_argv_0 = sys.argv[0]
+        sys.argv[0] = setup_script
+
+        try:
+            super().run_setup(setup_script=setup_script)
+        finally:
+            # While PEP 517 frontends should be calling each hook in a fresh
+            # subprocess according to the standard (and thus it should not be
+            # strictly necessary to restore the old sys.path), we'll restore
+            # the original path so that the path manipulation does not persist
+            # within the hook after run_setup is called.
+            sys.path[:] = sys_path
+            sys.argv[0] = sys_argv_0
+
+
+class _IncompatibleBdistWheel(SetuptoolsDeprecationWarning):
+    _SUMMARY = "wheel.bdist_wheel is deprecated, please import it from setuptools"
+    _DETAILS = """
+    Ensure that any custom bdist_wheel implementation is a subclass of
+    setuptools.command.bdist_wheel.bdist_wheel.
+    """
+    _DUE_DATE = (2025, 10, 15)
+    # Initially introduced in 2024/10/15, but maybe too disruptive to be enforced?
+    _SEE_URL = "https://github.com/pypa/wheel/pull/631"
+
+
+# The primary backend
+_BACKEND = _BuildMetaBackend()
+
+get_requires_for_build_wheel = _BACKEND.get_requires_for_build_wheel
+get_requires_for_build_sdist = _BACKEND.get_requires_for_build_sdist
+prepare_metadata_for_build_wheel = _BACKEND.prepare_metadata_for_build_wheel
+build_wheel = _BACKEND.build_wheel
+build_sdist = _BACKEND.build_sdist
+
+if not LEGACY_EDITABLE:
+    get_requires_for_build_editable = _BACKEND.get_requires_for_build_editable
+    prepare_metadata_for_build_editable = _BACKEND.prepare_metadata_for_build_editable
+    build_editable = _BACKEND.build_editable
+
+
+# The legacy backend
+__legacy__ = _BuildMetaLegacyBackend()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/cli-32.exe b/backend/venv/lib/python3.12/site-packages/setuptools/cli-32.exe
new file mode 100644
index 00000000..65c3cd99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/cli-32.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/cli-64.exe b/backend/venv/lib/python3.12/site-packages/setuptools/cli-64.exe
new file mode 100644
index 00000000..3ea50eeb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/cli-64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/cli-arm64.exe b/backend/venv/lib/python3.12/site-packages/setuptools/cli-arm64.exe
new file mode 100644
index 00000000..da96455a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/cli-arm64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/cli.exe b/backend/venv/lib/python3.12/site-packages/setuptools/cli.exe
new file mode 100644
index 00000000..65c3cd99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/cli.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/__init__.py
new file mode 100644
index 00000000..50e6c2f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/__init__.py
@@ -0,0 +1,21 @@
+# mypy: disable_error_code=call-overload
+# pyright: reportCallIssue=false, reportArgumentType=false
+# Can't disable on the exact line because distutils doesn't exists on Python 3.12
+# and type-checkers aren't aware of distutils_hack,
+# causing distutils.command.bdist.bdist.format_commands to be Any.
+
+import sys
+
+from distutils.command.bdist import bdist
+
+if 'egg' not in bdist.format_commands:
+    try:
+        # format_commands is a dict in vendored distutils
+        # It used to be a list in older (stdlib) distutils
+        # We support both for backwards compatibility
+        bdist.format_commands['egg'] = ('bdist_egg', "Python .egg file")
+    except TypeError:
+        bdist.format_command['egg'] = ('bdist_egg', "Python .egg file")
+        bdist.format_commands.append('egg')
+
+del bdist, sys
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..cd34dcb6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/_requirestxt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/_requirestxt.cpython-312.pyc
new file mode 100644
index 00000000..8993f66a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/_requirestxt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/alias.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/alias.cpython-312.pyc
new file mode 100644
index 00000000..da7aff30
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/alias.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_egg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_egg.cpython-312.pyc
new file mode 100644
index 00000000..e780964b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_egg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_rpm.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_rpm.cpython-312.pyc
new file mode 100644
index 00000000..20878785
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_rpm.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_wheel.cpython-312.pyc
new file mode 100644
index 00000000..263a0385
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/bdist_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build.cpython-312.pyc
new file mode 100644
index 00000000..dd753da5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_clib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_clib.cpython-312.pyc
new file mode 100644
index 00000000..02328578
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_clib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_ext.cpython-312.pyc
new file mode 100644
index 00000000..d7b0969b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_py.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_py.cpython-312.pyc
new file mode 100644
index 00000000..672b808d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/build_py.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/develop.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/develop.cpython-312.pyc
new file mode 100644
index 00000000..0d55eeda
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/develop.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/dist_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/dist_info.cpython-312.pyc
new file mode 100644
index 00000000..7b601887
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/dist_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/easy_install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/easy_install.cpython-312.pyc
new file mode 100644
index 00000000..eb979108
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/easy_install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/editable_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/editable_wheel.cpython-312.pyc
new file mode 100644
index 00000000..26873b53
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/editable_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/egg_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/egg_info.cpython-312.pyc
new file mode 100644
index 00000000..8661f6d1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/egg_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install.cpython-312.pyc
new file mode 100644
index 00000000..8c631060
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_egg_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_egg_info.cpython-312.pyc
new file mode 100644
index 00000000..80f3f878
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_egg_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_lib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_lib.cpython-312.pyc
new file mode 100644
index 00000000..63880e2a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_lib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_scripts.cpython-312.pyc
new file mode 100644
index 00000000..72b3c39b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/install_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/rotate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/rotate.cpython-312.pyc
new file mode 100644
index 00000000..0d8ca16c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/rotate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/saveopts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/saveopts.cpython-312.pyc
new file mode 100644
index 00000000..235d4cea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/saveopts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/sdist.cpython-312.pyc
new file mode 100644
index 00000000..fc441d20
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/setopt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/setopt.cpython-312.pyc
new file mode 100644
index 00000000..da48924a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/setopt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/test.cpython-312.pyc
new file mode 100644
index 00000000..99fc56a9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/command/__pycache__/test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/_requirestxt.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/_requirestxt.py
new file mode 100644
index 00000000..9029b125
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/_requirestxt.py
@@ -0,0 +1,131 @@
+"""Helper code used to generate ``requires.txt`` files in the egg-info directory.
+
+The ``requires.txt`` file has an specific format:
+    - Environment markers need to be part of the section headers and
+      should not be part of the requirement spec itself.
+
+See https://setuptools.pypa.io/en/latest/deprecated/python_eggs.html#requires-txt
+"""
+
+from __future__ import annotations
+
+import io
+from collections import defaultdict
+from collections.abc import Mapping
+from itertools import filterfalse
+from typing import TypeVar
+
+from jaraco.text import yield_lines
+from packaging.requirements import Requirement
+
+from .. import _reqs
+from .._reqs import _StrOrIter
+
+# dict can work as an ordered set
+_T = TypeVar("_T")
+_Ordered = dict[_T, None]
+
+
+def _prepare(
+    install_requires: _StrOrIter, extras_require: Mapping[str, _StrOrIter]
+) -> tuple[list[str], dict[str, list[str]]]:
+    """Given values for ``install_requires`` and ``extras_require``
+    create modified versions in a way that can be written in ``requires.txt``
+    """
+    extras = _convert_extras_requirements(extras_require)
+    return _move_install_requirements_markers(install_requires, extras)
+
+
+def _convert_extras_requirements(
+    extras_require: Mapping[str, _StrOrIter],
+) -> defaultdict[str, _Ordered[Requirement]]:
+    """
+    Convert requirements in `extras_require` of the form
+    `"extra": ["barbazquux; {marker}"]` to
+    `"extra:{marker}": ["barbazquux"]`.
+    """
+    output = defaultdict[str, _Ordered[Requirement]](dict)
+    for section, v in extras_require.items():
+        # Do not strip empty sections.
+        output[section]
+        for r in _reqs.parse(v):
+            output[section + _suffix_for(r)].setdefault(r)
+
+    return output
+
+
+def _move_install_requirements_markers(
+    install_requires: _StrOrIter, extras_require: Mapping[str, _Ordered[Requirement]]
+) -> tuple[list[str], dict[str, list[str]]]:
+    """
+    The ``requires.txt`` file has an specific format:
+        - Environment markers need to be part of the section headers and
+          should not be part of the requirement spec itself.
+
+    Move requirements in ``install_requires`` that are using environment
+    markers ``extras_require``.
+    """
+
+    # divide the install_requires into two sets, simple ones still
+    # handled by install_requires and more complex ones handled by extras_require.
+
+    inst_reqs = list(_reqs.parse(install_requires))
+    simple_reqs = filter(_no_marker, inst_reqs)
+    complex_reqs = filterfalse(_no_marker, inst_reqs)
+    simple_install_requires = list(map(str, simple_reqs))
+
+    for r in complex_reqs:
+        extras_require[':' + str(r.marker)].setdefault(r)
+
+    expanded_extras = dict(
+        # list(dict.fromkeys(...))  ensures a list of unique strings
+        (k, list(dict.fromkeys(str(r) for r in map(_clean_req, v))))
+        for k, v in extras_require.items()
+    )
+
+    return simple_install_requires, expanded_extras
+
+
+def _suffix_for(req):
+    """Return the 'extras_require' suffix for a given requirement."""
+    return ':' + str(req.marker) if req.marker else ''
+
+
+def _clean_req(req):
+    """Given a Requirement, remove environment markers and return it"""
+    r = Requirement(str(req))  # create a copy before modifying
+    r.marker = None
+    return r
+
+
+def _no_marker(req):
+    return not req.marker
+
+
+def _write_requirements(stream, reqs):
+    lines = yield_lines(reqs or ())
+
+    def append_cr(line):
+        return line + '\n'
+
+    lines = map(append_cr, lines)
+    stream.writelines(lines)
+
+
+def write_requirements(cmd, basename, filename):
+    dist = cmd.distribution
+    data = io.StringIO()
+    install_requires, extras_require = _prepare(
+        dist.install_requires or (), dist.extras_require or {}
+    )
+    _write_requirements(data, install_requires)
+    for extra in sorted(extras_require):
+        data.write('\n[{extra}]\n'.format(**vars()))
+        _write_requirements(data, extras_require[extra])
+    cmd.write_or_delete_file("requirements", filename, data.getvalue())
+
+
+def write_setup_requirements(cmd, basename, filename):
+    data = io.StringIO()
+    _write_requirements(data, cmd.distribution.setup_requires)
+    cmd.write_or_delete_file("setup-requirements", filename, data.getvalue())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/alias.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/alias.py
new file mode 100644
index 00000000..b8d74af7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/alias.py
@@ -0,0 +1,77 @@
+from setuptools.command.setopt import config_file, edit_config, option_base
+
+from distutils.errors import DistutilsOptionError
+
+
+def shquote(arg):
+    """Quote an argument for later parsing by shlex.split()"""
+    for c in '"', "'", "\\", "#":
+        if c in arg:
+            return repr(arg)
+    if arg.split() != [arg]:
+        return repr(arg)
+    return arg
+
+
+class alias(option_base):
+    """Define a shortcut that invokes one or more commands"""
+
+    description = "define a shortcut to invoke one or more commands"
+    command_consumes_arguments = True
+
+    user_options = [
+        ('remove', 'r', 'remove (unset) the alias'),
+    ] + option_base.user_options
+
+    boolean_options = option_base.boolean_options + ['remove']
+
+    def initialize_options(self):
+        option_base.initialize_options(self)
+        self.args = None
+        self.remove = None
+
+    def finalize_options(self) -> None:
+        option_base.finalize_options(self)
+        if self.remove and len(self.args) != 1:
+            raise DistutilsOptionError(
+                "Must specify exactly one argument (the alias name) when using --remove"
+            )
+
+    def run(self) -> None:
+        aliases = self.distribution.get_option_dict('aliases')
+
+        if not self.args:
+            print("Command Aliases")
+            print("---------------")
+            for alias in aliases:
+                print("setup.py alias", format_alias(alias, aliases))
+            return
+
+        elif len(self.args) == 1:
+            (alias,) = self.args
+            if self.remove:
+                command = None
+            elif alias in aliases:
+                print("setup.py alias", format_alias(alias, aliases))
+                return
+            else:
+                print(f"No alias definition found for {alias!r}")
+                return
+        else:
+            alias = self.args[0]
+            command = ' '.join(map(shquote, self.args[1:]))
+
+        edit_config(self.filename, {'aliases': {alias: command}}, self.dry_run)
+
+
+def format_alias(name, aliases):
+    source, command = aliases[name]
+    if source == config_file('global'):
+        source = '--global-config '
+    elif source == config_file('user'):
+        source = '--user-config '
+    elif source == config_file('local'):
+        source = ''
+    else:
+        source = f'--filename={source!r}'
+    return source + name + ' ' + command
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_egg.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_egg.py
new file mode 100644
index 00000000..7f66c3ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_egg.py
@@ -0,0 +1,479 @@
+"""setuptools.command.bdist_egg
+
+Build .egg distributions"""
+
+from __future__ import annotations
+
+import marshal
+import os
+import re
+import sys
+import textwrap
+from sysconfig import get_path, get_python_version
+from types import CodeType
+from typing import TYPE_CHECKING, Literal
+
+from setuptools import Command
+from setuptools.extension import Library
+
+from .._path import StrPathT, ensure_directory
+
+from distutils import log
+from distutils.dir_util import mkpath, remove_tree
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+# Same as zipfile._ZipFileMode from typeshed
+_ZipFileMode: TypeAlias = Literal["r", "w", "x", "a"]
+
+
+def _get_purelib():
+    return get_path("purelib")
+
+
+def strip_module(filename):
+    if '.' in filename:
+        filename = os.path.splitext(filename)[0]
+    if filename.endswith('module'):
+        filename = filename[:-6]
+    return filename
+
+
+def sorted_walk(dir):
+    """Do os.walk in a reproducible way,
+    independent of indeterministic filesystem readdir order
+    """
+    for base, dirs, files in os.walk(dir):
+        dirs.sort()
+        files.sort()
+        yield base, dirs, files
+
+
+def write_stub(resource, pyfile) -> None:
+    _stub_template = textwrap.dedent(
+        """
+        def __bootstrap__():
+            global __bootstrap__, __loader__, __file__
+            import sys, pkg_resources, importlib.util
+            __file__ = pkg_resources.resource_filename(__name__, %r)
+            __loader__ = None; del __bootstrap__, __loader__
+            spec = importlib.util.spec_from_file_location(__name__,__file__)
+            mod = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(mod)
+        __bootstrap__()
+        """
+    ).lstrip()
+    with open(pyfile, 'w', encoding="utf-8") as f:
+        f.write(_stub_template % resource)
+
+
+class bdist_egg(Command):
+    description = 'create an "egg" distribution'
+
+    user_options = [
+        ('bdist-dir=', 'b', "temporary directory for creating the distribution"),
+        (
+            'plat-name=',
+            'p',
+            "platform name to embed in generated filenames "
+            "(by default uses `pkg_resources.get_build_platform()`)",
+        ),
+        ('exclude-source-files', None, "remove all .py files from the generated egg"),
+        (
+            'keep-temp',
+            'k',
+            "keep the pseudo-installation tree around after "
+            "creating the distribution archive",
+        ),
+        ('dist-dir=', 'd', "directory to put final built distributions in"),
+        ('skip-build', None, "skip rebuilding everything (for testing/debugging)"),
+    ]
+
+    boolean_options = ['keep-temp', 'skip-build', 'exclude-source-files']
+
+    def initialize_options(self):
+        self.bdist_dir = None
+        self.plat_name = None
+        self.keep_temp = False
+        self.dist_dir = None
+        self.skip_build = False
+        self.egg_output = None
+        self.exclude_source_files = None
+
+    def finalize_options(self) -> None:
+        ei_cmd = self.ei_cmd = self.get_finalized_command("egg_info")
+        self.egg_info = ei_cmd.egg_info
+
+        if self.bdist_dir is None:
+            bdist_base = self.get_finalized_command('bdist').bdist_base
+            self.bdist_dir = os.path.join(bdist_base, 'egg')
+
+        if self.plat_name is None:
+            from pkg_resources import get_build_platform
+
+            self.plat_name = get_build_platform()
+
+        self.set_undefined_options('bdist', ('dist_dir', 'dist_dir'))
+
+        if self.egg_output is None:
+            # Compute filename of the output egg
+            basename = ei_cmd._get_egg_basename(
+                py_version=get_python_version(),
+                platform=self.distribution.has_ext_modules() and self.plat_name,
+            )
+
+            self.egg_output = os.path.join(self.dist_dir, basename + '.egg')
+
+    def do_install_data(self) -> None:
+        # Hack for packages that install data to install's --install-lib
+        self.get_finalized_command('install').install_lib = self.bdist_dir
+
+        site_packages = os.path.normcase(os.path.realpath(_get_purelib()))
+        old, self.distribution.data_files = self.distribution.data_files, []
+
+        for item in old:
+            if isinstance(item, tuple) and len(item) == 2:
+                if os.path.isabs(item[0]):
+                    realpath = os.path.realpath(item[0])
+                    normalized = os.path.normcase(realpath)
+                    if normalized == site_packages or normalized.startswith(
+                        site_packages + os.sep
+                    ):
+                        item = realpath[len(site_packages) + 1 :], item[1]
+                        # XXX else: raise ???
+            self.distribution.data_files.append(item)
+
+        try:
+            log.info("installing package data to %s", self.bdist_dir)
+            self.call_command('install_data', force=False, root=None)
+        finally:
+            self.distribution.data_files = old
+
+    def get_outputs(self):
+        return [self.egg_output]
+
+    def call_command(self, cmdname, **kw):
+        """Invoke reinitialized command `cmdname` with keyword args"""
+        for dirname in INSTALL_DIRECTORY_ATTRS:
+            kw.setdefault(dirname, self.bdist_dir)
+        kw.setdefault('skip_build', self.skip_build)
+        kw.setdefault('dry_run', self.dry_run)
+        cmd = self.reinitialize_command(cmdname, **kw)
+        self.run_command(cmdname)
+        return cmd
+
+    def run(self):  # noqa: C901  # is too complex (14)  # FIXME
+        # Generate metadata first
+        self.run_command("egg_info")
+        # We run install_lib before install_data, because some data hacks
+        # pull their data path from the install_lib command.
+        log.info("installing library code to %s", self.bdist_dir)
+        instcmd = self.get_finalized_command('install')
+        old_root = instcmd.root
+        instcmd.root = None
+        if self.distribution.has_c_libraries() and not self.skip_build:
+            self.run_command('build_clib')
+        cmd = self.call_command('install_lib', warn_dir=False)
+        instcmd.root = old_root
+
+        all_outputs, ext_outputs = self.get_ext_outputs()
+        self.stubs = []
+        to_compile = []
+        for p, ext_name in enumerate(ext_outputs):
+            filename, _ext = os.path.splitext(ext_name)
+            pyfile = os.path.join(self.bdist_dir, strip_module(filename) + '.py')
+            self.stubs.append(pyfile)
+            log.info("creating stub loader for %s", ext_name)
+            if not self.dry_run:
+                write_stub(os.path.basename(ext_name), pyfile)
+            to_compile.append(pyfile)
+            ext_outputs[p] = ext_name.replace(os.sep, '/')
+
+        if to_compile:
+            cmd.byte_compile(to_compile)
+        if self.distribution.data_files:
+            self.do_install_data()
+
+        # Make the EGG-INFO directory
+        archive_root = self.bdist_dir
+        egg_info = os.path.join(archive_root, 'EGG-INFO')
+        self.mkpath(egg_info)
+        if self.distribution.scripts:
+            script_dir = os.path.join(egg_info, 'scripts')
+            log.info("installing scripts to %s", script_dir)
+            self.call_command('install_scripts', install_dir=script_dir, no_ep=True)
+
+        self.copy_metadata_to(egg_info)
+        native_libs = os.path.join(egg_info, "native_libs.txt")
+        if all_outputs:
+            log.info("writing %s", native_libs)
+            if not self.dry_run:
+                ensure_directory(native_libs)
+                with open(native_libs, 'wt', encoding="utf-8") as libs_file:
+                    libs_file.write('\n'.join(all_outputs))
+                    libs_file.write('\n')
+        elif os.path.isfile(native_libs):
+            log.info("removing %s", native_libs)
+            if not self.dry_run:
+                os.unlink(native_libs)
+
+        write_safety_flag(os.path.join(archive_root, 'EGG-INFO'), self.zip_safe())
+
+        if os.path.exists(os.path.join(self.egg_info, 'depends.txt')):
+            log.warn(
+                "WARNING: 'depends.txt' will not be used by setuptools 0.6!\n"
+                "Use the install_requires/extras_require setup() args instead."
+            )
+
+        if self.exclude_source_files:
+            self.zap_pyfiles()
+
+        # Make the archive
+        make_zipfile(
+            self.egg_output,
+            archive_root,
+            verbose=self.verbose,
+            dry_run=self.dry_run,
+            mode=self.gen_header(),
+        )
+        if not self.keep_temp:
+            remove_tree(self.bdist_dir, dry_run=self.dry_run)
+
+        # Add to 'Distribution.dist_files' so that the "upload" command works
+        getattr(self.distribution, 'dist_files', []).append((
+            'bdist_egg',
+            get_python_version(),
+            self.egg_output,
+        ))
+
+    def zap_pyfiles(self):
+        log.info("Removing .py files from temporary directory")
+        for base, dirs, files in walk_egg(self.bdist_dir):
+            for name in files:
+                path = os.path.join(base, name)
+
+                if name.endswith('.py'):
+                    log.debug("Deleting %s", path)
+                    os.unlink(path)
+
+                if base.endswith('__pycache__'):
+                    path_old = path
+
+                    pattern = r'(?P<name>.+)\.(?P<magic>[^.]+)\.pyc'
+                    m = re.match(pattern, name)
+                    path_new = os.path.join(base, os.pardir, m.group('name') + '.pyc')
+                    log.info(f"Renaming file from [{path_old}] to [{path_new}]")
+                    try:
+                        os.remove(path_new)
+                    except OSError:
+                        pass
+                    os.rename(path_old, path_new)
+
+    def zip_safe(self):
+        safe = getattr(self.distribution, 'zip_safe', None)
+        if safe is not None:
+            return safe
+        log.warn("zip_safe flag not set; analyzing archive contents...")
+        return analyze_egg(self.bdist_dir, self.stubs)
+
+    def gen_header(self) -> Literal["w"]:
+        return 'w'
+
+    def copy_metadata_to(self, target_dir) -> None:
+        "Copy metadata (egg info) to the target_dir"
+        # normalize the path (so that a forward-slash in egg_info will
+        # match using startswith below)
+        norm_egg_info = os.path.normpath(self.egg_info)
+        prefix = os.path.join(norm_egg_info, '')
+        for path in self.ei_cmd.filelist.files:
+            if path.startswith(prefix):
+                target = os.path.join(target_dir, path[len(prefix) :])
+                ensure_directory(target)
+                self.copy_file(path, target)
+
+    def get_ext_outputs(self):
+        """Get a list of relative paths to C extensions in the output distro"""
+
+        all_outputs = []
+        ext_outputs = []
+
+        paths = {self.bdist_dir: ''}
+        for base, dirs, files in sorted_walk(self.bdist_dir):
+            all_outputs.extend(
+                paths[base] + filename
+                for filename in files
+                if os.path.splitext(filename)[1].lower() in NATIVE_EXTENSIONS
+            )
+            for filename in dirs:
+                paths[os.path.join(base, filename)] = paths[base] + filename + '/'
+
+        if self.distribution.has_ext_modules():
+            build_cmd = self.get_finalized_command('build_ext')
+            for ext in build_cmd.extensions:
+                if isinstance(ext, Library):
+                    continue
+                fullname = build_cmd.get_ext_fullname(ext.name)
+                filename = build_cmd.get_ext_filename(fullname)
+                if not os.path.basename(filename).startswith('dl-'):
+                    if os.path.exists(os.path.join(self.bdist_dir, filename)):
+                        ext_outputs.append(filename)
+
+        return all_outputs, ext_outputs
+
+
+NATIVE_EXTENSIONS: dict[str, None] = dict.fromkeys('.dll .so .dylib .pyd'.split())
+
+
+def walk_egg(egg_dir):
+    """Walk an unpacked egg's contents, skipping the metadata directory"""
+    walker = sorted_walk(egg_dir)
+    base, dirs, files = next(walker)
+    if 'EGG-INFO' in dirs:
+        dirs.remove('EGG-INFO')
+    yield base, dirs, files
+    yield from walker
+
+
+def analyze_egg(egg_dir, stubs):
+    # check for existing flag in EGG-INFO
+    for flag, fn in safety_flags.items():
+        if os.path.exists(os.path.join(egg_dir, 'EGG-INFO', fn)):
+            return flag
+    if not can_scan():
+        return False
+    safe = True
+    for base, dirs, files in walk_egg(egg_dir):
+        for name in files:
+            if name.endswith('.py') or name.endswith('.pyw'):
+                continue
+            elif name.endswith('.pyc') or name.endswith('.pyo'):
+                # always scan, even if we already know we're not safe
+                safe = scan_module(egg_dir, base, name, stubs) and safe
+    return safe
+
+
+def write_safety_flag(egg_dir, safe) -> None:
+    # Write or remove zip safety flag file(s)
+    for flag, fn in safety_flags.items():
+        fn = os.path.join(egg_dir, fn)
+        if os.path.exists(fn):
+            if safe is None or bool(safe) != flag:
+                os.unlink(fn)
+        elif safe is not None and bool(safe) == flag:
+            with open(fn, 'wt', encoding="utf-8") as f:
+                f.write('\n')
+
+
+safety_flags = {
+    True: 'zip-safe',
+    False: 'not-zip-safe',
+}
+
+
+def scan_module(egg_dir, base, name, stubs):
+    """Check whether module possibly uses unsafe-for-zipfile stuff"""
+
+    filename = os.path.join(base, name)
+    if filename[:-1] in stubs:
+        return True  # Extension module
+    pkg = base[len(egg_dir) + 1 :].replace(os.sep, '.')
+    module = pkg + (pkg and '.' or '') + os.path.splitext(name)[0]
+    skip = 16  # skip magic & reserved? & date & file size
+    f = open(filename, 'rb')
+    f.read(skip)
+    code = marshal.load(f)
+    f.close()
+    safe = True
+    symbols = dict.fromkeys(iter_symbols(code))
+    for bad in ['__file__', '__path__']:
+        if bad in symbols:
+            log.warn("%s: module references %s", module, bad)
+            safe = False
+    if 'inspect' in symbols:
+        for bad in [
+            'getsource',
+            'getabsfile',
+            'getfile',
+            'getsourcefile',
+            'getsourcelines',
+            'findsource',
+            'getcomments',
+            'getframeinfo',
+            'getinnerframes',
+            'getouterframes',
+            'stack',
+            'trace',
+        ]:
+            if bad in symbols:
+                log.warn("%s: module MAY be using inspect.%s", module, bad)
+                safe = False
+    return safe
+
+
+def iter_symbols(code):
+    """Yield names and strings used by `code` and its nested code objects"""
+    yield from code.co_names
+    for const in code.co_consts:
+        if isinstance(const, str):
+            yield const
+        elif isinstance(const, CodeType):
+            yield from iter_symbols(const)
+
+
+def can_scan() -> bool:
+    if not sys.platform.startswith('java') and sys.platform != 'cli':
+        # CPython, PyPy, etc.
+        return True
+    log.warn("Unable to analyze compiled code on this platform.")
+    log.warn(
+        "Please ask the author to include a 'zip_safe'"
+        " setting (either True or False) in the package's setup.py"
+    )
+    return False
+
+
+# Attribute names of options for commands that might need to be convinced to
+# install to the egg build directory
+
+INSTALL_DIRECTORY_ATTRS = ['install_lib', 'install_dir', 'install_data', 'install_base']
+
+
+def make_zipfile(
+    zip_filename: StrPathT,
+    base_dir,
+    verbose: bool = False,
+    dry_run: bool = False,
+    compress=True,
+    mode: _ZipFileMode = 'w',
+) -> StrPathT:
+    """Create a zip file from all the files under 'base_dir'.  The output
+    zip file will be named 'base_dir' + ".zip".  Uses either the "zipfile"
+    Python module (if available) or the InfoZIP "zip" utility (if installed
+    and found on the default search path).  If neither tool is available,
+    raises DistutilsExecError.  Returns the name of the output zip file.
+    """
+    import zipfile
+
+    mkpath(os.path.dirname(zip_filename), dry_run=dry_run)  # type: ignore[arg-type] # python/mypy#18075
+    log.info("creating '%s' and adding '%s' to it", zip_filename, base_dir)
+
+    def visit(z, dirname, names):
+        for name in names:
+            path = os.path.normpath(os.path.join(dirname, name))
+            if os.path.isfile(path):
+                p = path[len(base_dir) + 1 :]
+                if not dry_run:
+                    z.write(path, p)
+                log.debug("adding '%s'", p)
+
+    compression = zipfile.ZIP_DEFLATED if compress else zipfile.ZIP_STORED
+    if not dry_run:
+        z = zipfile.ZipFile(zip_filename, mode, compression=compression)
+        for dirname, dirs, files in sorted_walk(base_dir):
+            visit(z, dirname, files)
+        z.close()
+    else:
+        for dirname, dirs, files in sorted_walk(base_dir):
+            visit(None, dirname, files)
+    return zip_filename
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_rpm.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_rpm.py
new file mode 100644
index 00000000..6dbb2700
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_rpm.py
@@ -0,0 +1,42 @@
+from ..dist import Distribution
+from ..warnings import SetuptoolsDeprecationWarning
+
+import distutils.command.bdist_rpm as orig
+
+
+class bdist_rpm(orig.bdist_rpm):
+    """
+    Override the default bdist_rpm behavior to do the following:
+
+    1. Run egg_info to ensure the name and version are properly calculated.
+    2. Always run 'install' using --single-version-externally-managed to
+       disable eggs in RPM distributions.
+    """
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    def run(self) -> None:
+        SetuptoolsDeprecationWarning.emit(
+            "Deprecated command",
+            """
+            bdist_rpm is deprecated and will be removed in a future version.
+            Use bdist_wheel (wheel packages) instead.
+            """,
+            see_url="https://github.com/pypa/setuptools/issues/1988",
+            due_date=(2023, 10, 30),  # Deprecation introduced in 22 Oct 2021.
+        )
+
+        # ensure distro name is up-to-date
+        self.run_command('egg_info')
+
+        orig.bdist_rpm.run(self)
+
+    def _make_spec_file(self):
+        spec = orig.bdist_rpm._make_spec_file(self)
+        return [
+            line.replace(
+                "setup.py install ",
+                "setup.py install --single-version-externally-managed ",
+            ).replace("%setup", "%setup -n %{name}-%{unmangled_version}")
+            for line in spec
+        ]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_wheel.py
new file mode 100644
index 00000000..d9e1eb97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/bdist_wheel.py
@@ -0,0 +1,610 @@
+"""
+Create a wheel (.whl) distribution.
+
+A wheel is a built archive format.
+"""
+
+from __future__ import annotations
+
+import os
+import re
+import shutil
+import struct
+import sys
+import sysconfig
+import warnings
+from collections.abc import Iterable, Sequence
+from email.generator import BytesGenerator
+from glob import iglob
+from typing import Literal, cast
+from zipfile import ZIP_DEFLATED, ZIP_STORED
+
+from packaging import tags, version as _packaging_version
+from wheel.wheelfile import WheelFile
+
+from .. import Command, __version__, _shutil
+from ..warnings import SetuptoolsDeprecationWarning
+from .egg_info import egg_info as egg_info_cls
+
+from distutils import log
+
+
+def safe_name(name: str) -> str:
+    """Convert an arbitrary string to a standard distribution name
+    Any runs of non-alphanumeric/. characters are replaced with a single '-'.
+    """
+    return re.sub("[^A-Za-z0-9.]+", "-", name)
+
+
+def safe_version(version: str) -> str:
+    """
+    Convert an arbitrary string to a standard version string
+    """
+    try:
+        # normalize the version
+        return str(_packaging_version.Version(version))
+    except _packaging_version.InvalidVersion:
+        version = version.replace(" ", ".")
+        return re.sub("[^A-Za-z0-9.]+", "-", version)
+
+
+setuptools_major_version = int(__version__.split(".")[0])
+
+PY_LIMITED_API_PATTERN = r"cp3\d"
+
+
+def _is_32bit_interpreter() -> bool:
+    return struct.calcsize("P") == 4
+
+
+def python_tag() -> str:
+    return f"py{sys.version_info.major}"
+
+
+def get_platform(archive_root: str | None) -> str:
+    """Return our platform name 'win32', 'linux_x86_64'"""
+    result = sysconfig.get_platform()
+    if result.startswith("macosx") and archive_root is not None:  # pragma: no cover
+        from wheel.macosx_libfile import calculate_macosx_platform_tag
+
+        result = calculate_macosx_platform_tag(archive_root, result)
+    elif _is_32bit_interpreter():
+        if result == "linux-x86_64":
+            # pip pull request #3497
+            result = "linux-i686"
+        elif result == "linux-aarch64":
+            # packaging pull request #234
+            # TODO armv8l, packaging pull request #690 => this did not land
+            # in pip/packaging yet
+            result = "linux-armv7l"
+
+    return result.replace("-", "_")
+
+
+def get_flag(
+    var: str, fallback: bool, expected: bool = True, warn: bool = True
+) -> bool:
+    """Use a fallback value for determining SOABI flags if the needed config
+    var is unset or unavailable."""
+    val = sysconfig.get_config_var(var)
+    if val is None:
+        if warn:
+            warnings.warn(
+                f"Config variable '{var}' is unset, Python ABI tag may be incorrect",
+                RuntimeWarning,
+                stacklevel=2,
+            )
+        return fallback
+    return val == expected
+
+
+def get_abi_tag() -> str | None:
+    """Return the ABI tag based on SOABI (if available) or emulate SOABI (PyPy2)."""
+    soabi: str = sysconfig.get_config_var("SOABI")
+    impl = tags.interpreter_name()
+    if not soabi and impl in ("cp", "pp") and hasattr(sys, "maxunicode"):
+        d = ""
+        u = ""
+        if get_flag("Py_DEBUG", hasattr(sys, "gettotalrefcount"), warn=(impl == "cp")):
+            d = "d"
+
+        abi = f"{impl}{tags.interpreter_version()}{d}{u}"
+    elif soabi and impl == "cp" and soabi.startswith("cpython"):
+        # non-Windows
+        abi = "cp" + soabi.split("-")[1]
+    elif soabi and impl == "cp" and soabi.startswith("cp"):
+        # Windows
+        abi = soabi.split("-")[0]
+        if hasattr(sys, "gettotalrefcount"):
+            # using debug build; append "d" flag
+            abi += "d"
+    elif soabi and impl == "pp":
+        # we want something like pypy36-pp73
+        abi = "-".join(soabi.split("-")[:2])
+        abi = abi.replace(".", "_").replace("-", "_")
+    elif soabi and impl == "graalpy":
+        abi = "-".join(soabi.split("-")[:3])
+        abi = abi.replace(".", "_").replace("-", "_")
+    elif soabi:
+        abi = soabi.replace(".", "_").replace("-", "_")
+    else:
+        abi = None
+
+    return abi
+
+
+def safer_name(name: str) -> str:
+    return safe_name(name).replace("-", "_")
+
+
+def safer_version(version: str) -> str:
+    return safe_version(version).replace("-", "_")
+
+
+class bdist_wheel(Command):
+    description = "create a wheel distribution"
+
+    supported_compressions = {
+        "stored": ZIP_STORED,
+        "deflated": ZIP_DEFLATED,
+    }
+
+    user_options = [
+        ("bdist-dir=", "b", "temporary directory for creating the distribution"),
+        (
+            "plat-name=",
+            "p",
+            "platform name to embed in generated filenames "
+            f"[default: {get_platform(None)}]",
+        ),
+        (
+            "keep-temp",
+            "k",
+            "keep the pseudo-installation tree around after "
+            "creating the distribution archive",
+        ),
+        ("dist-dir=", "d", "directory to put final built distributions in"),
+        ("skip-build", None, "skip rebuilding everything (for testing/debugging)"),
+        (
+            "relative",
+            None,
+            "build the archive using relative paths [default: false]",
+        ),
+        (
+            "owner=",
+            "u",
+            "Owner name used when creating a tar file [default: current user]",
+        ),
+        (
+            "group=",
+            "g",
+            "Group name used when creating a tar file [default: current group]",
+        ),
+        ("universal", None, "*DEPRECATED* make a universal wheel [default: false]"),
+        (
+            "compression=",
+            None,
+            f"zipfile compression (one of: {', '.join(supported_compressions)}) [default: 'deflated']",
+        ),
+        (
+            "python-tag=",
+            None,
+            f"Python implementation compatibility tag [default: '{python_tag()}']",
+        ),
+        (
+            "build-number=",
+            None,
+            "Build number for this particular version. "
+            "As specified in PEP-0427, this must start with a digit. "
+            "[default: None]",
+        ),
+        (
+            "py-limited-api=",
+            None,
+            "Python tag (cp32|cp33|cpNN) for abi3 wheel tag [default: false]",
+        ),
+        (
+            "dist-info-dir=",
+            None,
+            "directory where a pre-generated dist-info can be found (e.g. as a "
+            "result of calling the PEP517 'prepare_metadata_for_build_wheel' "
+            "method)",
+        ),
+    ]
+
+    boolean_options = ["keep-temp", "skip-build", "relative", "universal"]
+
+    def initialize_options(self) -> None:
+        self.bdist_dir: str | None = None
+        self.data_dir = ""
+        self.plat_name: str | None = None
+        self.plat_tag: str | None = None
+        self.format = "zip"
+        self.keep_temp = False
+        self.dist_dir: str | None = None
+        self.dist_info_dir = None
+        self.egginfo_dir: str | None = None
+        self.root_is_pure: bool | None = None
+        self.skip_build = False
+        self.relative = False
+        self.owner = None
+        self.group = None
+        self.universal = False
+        self.compression: str | int = "deflated"
+        self.python_tag = python_tag()
+        self.build_number: str | None = None
+        self.py_limited_api: str | Literal[False] = False
+        self.plat_name_supplied = False
+
+    def finalize_options(self) -> None:
+        if not self.bdist_dir:
+            bdist_base = self.get_finalized_command("bdist").bdist_base
+            self.bdist_dir = os.path.join(bdist_base, "wheel")
+
+        if self.dist_info_dir is None:
+            egg_info = cast(egg_info_cls, self.distribution.get_command_obj("egg_info"))
+            egg_info.ensure_finalized()  # needed for correct `wheel_dist_name`
+
+        self.data_dir = self.wheel_dist_name + ".data"
+        self.plat_name_supplied = bool(self.plat_name)
+
+        need_options = ("dist_dir", "plat_name", "skip_build")
+
+        self.set_undefined_options("bdist", *zip(need_options, need_options))
+
+        self.root_is_pure = not (
+            self.distribution.has_ext_modules() or self.distribution.has_c_libraries()
+        )
+
+        self._validate_py_limited_api()
+
+        # Support legacy [wheel] section for setting universal
+        wheel = self.distribution.get_option_dict("wheel")
+        if "universal" in wheel:  # pragma: no cover
+            # please don't define this in your global configs
+            log.warn("The [wheel] section is deprecated. Use [bdist_wheel] instead.")
+            val = wheel["universal"][1].strip()
+            if val.lower() in ("1", "true", "yes"):
+                self.universal = True
+
+        if self.universal:
+            SetuptoolsDeprecationWarning.emit(
+                "bdist_wheel.universal is deprecated",
+                """
+                With Python 2.7 end-of-life, support for building universal wheels
+                (i.e., wheels that support both Python 2 and Python 3)
+                is being obviated.
+                Please discontinue using this option, or if you still need it,
+                file an issue with pypa/setuptools describing your use case.
+                """,
+                due_date=(2025, 8, 30),  # Introduced in 2024-08-30
+            )
+
+        if self.build_number is not None and not self.build_number[:1].isdigit():
+            raise ValueError("Build tag (build-number) must start with a digit.")
+
+    def _validate_py_limited_api(self) -> None:
+        if not self.py_limited_api:
+            return
+
+        if not re.match(PY_LIMITED_API_PATTERN, self.py_limited_api):
+            raise ValueError(f"py-limited-api must match '{PY_LIMITED_API_PATTERN}'")
+
+        if sysconfig.get_config_var("Py_GIL_DISABLED"):
+            raise ValueError(
+                f"`py_limited_api={self.py_limited_api!r}` not supported. "
+                "`Py_LIMITED_API` is currently incompatible with "
+                f"`Py_GIL_DISABLED` ({sys.abiflags=!r}). "
+                "See https://github.com/python/cpython/issues/111506."
+            )
+
+    @property
+    def wheel_dist_name(self) -> str:
+        """Return distribution full name with - replaced with _"""
+        components = [
+            safer_name(self.distribution.get_name()),
+            safer_version(self.distribution.get_version()),
+        ]
+        if self.build_number:
+            components.append(self.build_number)
+        return "-".join(components)
+
+    def get_tag(self) -> tuple[str, str, str]:
+        # bdist sets self.plat_name if unset, we should only use it for purepy
+        # wheels if the user supplied it.
+        if self.plat_name_supplied and self.plat_name:
+            plat_name = self.plat_name
+        elif self.root_is_pure:
+            plat_name = "any"
+        else:
+            # macosx contains system version in platform name so need special handle
+            if self.plat_name and not self.plat_name.startswith("macosx"):
+                plat_name = self.plat_name
+            else:
+                # on macosx always limit the platform name to comply with any
+                # c-extension modules in bdist_dir, since the user can specify
+                # a higher MACOSX_DEPLOYMENT_TARGET via tools like CMake
+
+                # on other platforms, and on macosx if there are no c-extension
+                # modules, use the default platform name.
+                plat_name = get_platform(self.bdist_dir)
+
+            if _is_32bit_interpreter():
+                if plat_name in ("linux-x86_64", "linux_x86_64"):
+                    plat_name = "linux_i686"
+                if plat_name in ("linux-aarch64", "linux_aarch64"):
+                    # TODO armv8l, packaging pull request #690 => this did not land
+                    # in pip/packaging yet
+                    plat_name = "linux_armv7l"
+
+        plat_name = (
+            plat_name.lower().replace("-", "_").replace(".", "_").replace(" ", "_")
+        )
+
+        if self.root_is_pure:
+            if self.universal:
+                impl = "py2.py3"
+            else:
+                impl = self.python_tag
+            tag = (impl, "none", plat_name)
+        else:
+            impl_name = tags.interpreter_name()
+            impl_ver = tags.interpreter_version()
+            impl = impl_name + impl_ver
+            # We don't work on CPython 3.1, 3.0.
+            if self.py_limited_api and (impl_name + impl_ver).startswith("cp3"):
+                impl = self.py_limited_api
+                abi_tag = "abi3"
+            else:
+                abi_tag = str(get_abi_tag()).lower()
+            tag = (impl, abi_tag, plat_name)
+            # issue gh-374: allow overriding plat_name
+            supported_tags = [
+                (t.interpreter, t.abi, plat_name) for t in tags.sys_tags()
+            ]
+            assert tag in supported_tags, (
+                f"would build wheel with unsupported tag {tag}"
+            )
+        return tag
+
+    def run(self):
+        build_scripts = self.reinitialize_command("build_scripts")
+        build_scripts.executable = "python"
+        build_scripts.force = True
+
+        build_ext = self.reinitialize_command("build_ext")
+        build_ext.inplace = False
+
+        if not self.skip_build:
+            self.run_command("build")
+
+        install = self.reinitialize_command("install", reinit_subcommands=True)
+        install.root = self.bdist_dir
+        install.compile = False
+        install.skip_build = self.skip_build
+        install.warn_dir = False
+
+        # A wheel without setuptools scripts is more cross-platform.
+        # Use the (undocumented) `no_ep` option to setuptools'
+        # install_scripts command to avoid creating entry point scripts.
+        install_scripts = self.reinitialize_command("install_scripts")
+        install_scripts.no_ep = True
+
+        # Use a custom scheme for the archive, because we have to decide
+        # at installation time which scheme to use.
+        for key in ("headers", "scripts", "data", "purelib", "platlib"):
+            setattr(install, "install_" + key, os.path.join(self.data_dir, key))
+
+        basedir_observed = ""
+
+        if os.name == "nt":
+            # win32 barfs if any of these are ''; could be '.'?
+            # (distutils.command.install:change_roots bug)
+            basedir_observed = os.path.normpath(os.path.join(self.data_dir, ".."))
+            self.install_libbase = self.install_lib = basedir_observed
+
+        setattr(
+            install,
+            "install_purelib" if self.root_is_pure else "install_platlib",
+            basedir_observed,
+        )
+
+        log.info(f"installing to {self.bdist_dir}")
+
+        self.run_command("install")
+
+        impl_tag, abi_tag, plat_tag = self.get_tag()
+        archive_basename = f"{self.wheel_dist_name}-{impl_tag}-{abi_tag}-{plat_tag}"
+        if not self.relative:
+            archive_root = self.bdist_dir
+        else:
+            archive_root = os.path.join(
+                self.bdist_dir, self._ensure_relative(install.install_base)
+            )
+
+        self.set_undefined_options("install_egg_info", ("target", "egginfo_dir"))
+        distinfo_dirname = (
+            f"{safer_name(self.distribution.get_name())}-"
+            f"{safer_version(self.distribution.get_version())}.dist-info"
+        )
+        distinfo_dir = os.path.join(self.bdist_dir, distinfo_dirname)
+        if self.dist_info_dir:
+            # Use the given dist-info directly.
+            log.debug(f"reusing {self.dist_info_dir}")
+            shutil.copytree(self.dist_info_dir, distinfo_dir)
+            # Egg info is still generated, so remove it now to avoid it getting
+            # copied into the wheel.
+            _shutil.rmtree(self.egginfo_dir)
+        else:
+            # Convert the generated egg-info into dist-info.
+            self.egg2dist(self.egginfo_dir, distinfo_dir)
+
+        self.write_wheelfile(distinfo_dir)
+
+        # Make the archive
+        if not os.path.exists(self.dist_dir):
+            os.makedirs(self.dist_dir)
+
+        wheel_path = os.path.join(self.dist_dir, archive_basename + ".whl")
+        with WheelFile(wheel_path, "w", self._zip_compression()) as wf:
+            wf.write_files(archive_root)
+
+        # Add to 'Distribution.dist_files' so that the "upload" command works
+        getattr(self.distribution, "dist_files", []).append((
+            "bdist_wheel",
+            f"{sys.version_info.major}.{sys.version_info.minor}",
+            wheel_path,
+        ))
+
+        if not self.keep_temp:
+            log.info(f"removing {self.bdist_dir}")
+            if not self.dry_run:
+                _shutil.rmtree(self.bdist_dir)
+
+    def write_wheelfile(
+        self, wheelfile_base: str, generator: str = f"setuptools ({__version__})"
+    ) -> None:
+        from email.message import Message
+
+        msg = Message()
+        msg["Wheel-Version"] = "1.0"  # of the spec
+        msg["Generator"] = generator
+        msg["Root-Is-Purelib"] = str(self.root_is_pure).lower()
+        if self.build_number is not None:
+            msg["Build"] = self.build_number
+
+        # Doesn't work for bdist_wininst
+        impl_tag, abi_tag, plat_tag = self.get_tag()
+        for impl in impl_tag.split("."):
+            for abi in abi_tag.split("."):
+                for plat in plat_tag.split("."):
+                    msg["Tag"] = "-".join((impl, abi, plat))
+
+        wheelfile_path = os.path.join(wheelfile_base, "WHEEL")
+        log.info(f"creating {wheelfile_path}")
+        with open(wheelfile_path, "wb") as f:
+            BytesGenerator(f, maxheaderlen=0).flatten(msg)
+
+    def _ensure_relative(self, path: str) -> str:
+        # copied from dir_util, deleted
+        drive, path = os.path.splitdrive(path)
+        if path[0:1] == os.sep:
+            path = drive + path[1:]
+        return path
+
+    @property
+    def license_paths(self) -> Iterable[str]:
+        if setuptools_major_version >= 57:
+            # Setuptools has resolved any patterns to actual file names
+            return self.distribution.metadata.license_files or ()
+
+        files = set[str]()
+        metadata = self.distribution.get_option_dict("metadata")
+        if setuptools_major_version >= 42:
+            # Setuptools recognizes the license_files option but does not do globbing
+            patterns = cast(Sequence[str], self.distribution.metadata.license_files)
+        else:
+            # Prior to those, wheel is entirely responsible for handling license files
+            if "license_files" in metadata:
+                patterns = metadata["license_files"][1].split()
+            else:
+                patterns = ()
+
+        if "license_file" in metadata:
+            warnings.warn(
+                'The "license_file" option is deprecated. Use "license_files" instead.',
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            files.add(metadata["license_file"][1])
+
+        if not files and not patterns and not isinstance(patterns, list):
+            patterns = ("LICEN[CS]E*", "COPYING*", "NOTICE*", "AUTHORS*")
+
+        for pattern in patterns:
+            for path in iglob(pattern):
+                if path.endswith("~"):
+                    log.debug(
+                        f'ignoring license file "{path}" as it looks like a backup'
+                    )
+                    continue
+
+                if path not in files and os.path.isfile(path):
+                    log.info(
+                        f'adding license file "{path}" (matched pattern "{pattern}")'
+                    )
+                    files.add(path)
+
+        return files
+
+    def egg2dist(self, egginfo_path: str, distinfo_path: str) -> None:
+        """Convert an .egg-info directory into a .dist-info directory"""
+
+        def adios(p: str) -> None:
+            """Appropriately delete directory, file or link."""
+            if os.path.exists(p) and not os.path.islink(p) and os.path.isdir(p):
+                _shutil.rmtree(p)
+            elif os.path.exists(p):
+                os.unlink(p)
+
+        adios(distinfo_path)
+
+        if not os.path.exists(egginfo_path):
+            # There is no egg-info. This is probably because the egg-info
+            # file/directory is not named matching the distribution name used
+            # to name the archive file. Check for this case and report
+            # accordingly.
+            import glob
+
+            pat = os.path.join(os.path.dirname(egginfo_path), "*.egg-info")
+            possible = glob.glob(pat)
+            err = f"Egg metadata expected at {egginfo_path} but not found"
+            if possible:
+                alt = os.path.basename(possible[0])
+                err += f" ({alt} found - possible misnamed archive file?)"
+
+            raise ValueError(err)
+
+        # .egg-info is a directory
+        pkginfo_path = os.path.join(egginfo_path, "PKG-INFO")
+
+        # ignore common egg metadata that is useless to wheel
+        shutil.copytree(
+            egginfo_path,
+            distinfo_path,
+            ignore=lambda x, y: {
+                "PKG-INFO",
+                "requires.txt",
+                "SOURCES.txt",
+                "not-zip-safe",
+            },
+        )
+
+        # delete dependency_links if it is only whitespace
+        dependency_links_path = os.path.join(distinfo_path, "dependency_links.txt")
+        with open(dependency_links_path, encoding="utf-8") as dependency_links_file:
+            dependency_links = dependency_links_file.read().strip()
+        if not dependency_links:
+            adios(dependency_links_path)
+
+        metadata_path = os.path.join(distinfo_path, "METADATA")
+        shutil.copy(pkginfo_path, metadata_path)
+
+        for license_path in self.license_paths:
+            filename = os.path.basename(license_path)
+            shutil.copy(license_path, os.path.join(distinfo_path, filename))
+
+        adios(egginfo_path)
+
+    def _zip_compression(self) -> int:
+        if (
+            isinstance(self.compression, int)
+            and self.compression in self.supported_compressions.values()
+        ):
+            return self.compression
+
+        compression = self.supported_compressions.get(str(self.compression))
+        if compression is not None:
+            return compression
+
+        raise ValueError(f"Unsupported compression: {self.compression!r}")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/build.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/build.py
new file mode 100644
index 00000000..54cbb8d2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/build.py
@@ -0,0 +1,135 @@
+from __future__ import annotations
+
+from typing import Protocol
+
+from ..dist import Distribution
+
+from distutils.command.build import build as _build
+
+_ORIGINAL_SUBCOMMANDS = {"build_py", "build_clib", "build_ext", "build_scripts"}
+
+
+class build(_build):
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    # copy to avoid sharing the object with parent class
+    sub_commands = _build.sub_commands[:]
+
+
+class SubCommand(Protocol):
+    """In order to support editable installations (see :pep:`660`) all
+    build subcommands **SHOULD** implement this protocol. They also **MUST** inherit
+    from ``setuptools.Command``.
+
+    When creating an :pep:`editable wheel <660>`, ``setuptools`` will try to evaluate
+    custom ``build`` subcommands using the following procedure:
+
+    1. ``setuptools`` will set the ``editable_mode`` attribute to ``True``
+    2. ``setuptools`` will execute the ``run()`` command.
+
+       .. important::
+          Subcommands **SHOULD** take advantage of ``editable_mode=True`` to adequate
+          its behaviour or perform optimisations.
+
+          For example, if a subcommand doesn't need to generate an extra file and
+          all it does is to copy a source file into the build directory,
+          ``run()`` **SHOULD** simply "early return".
+
+          Similarly, if the subcommand creates files that would be placed alongside
+          Python files in the final distribution, during an editable install
+          the command **SHOULD** generate these files "in place" (i.e. write them to
+          the original source directory, instead of using the build directory).
+          Note that ``get_output_mapping()`` should reflect that and include mappings
+          for "in place" builds accordingly.
+
+    3. ``setuptools`` use any knowledge it can derive from the return values of
+       ``get_outputs()`` and ``get_output_mapping()`` to create an editable wheel.
+       When relevant ``setuptools`` **MAY** attempt to use file links based on the value
+       of ``get_output_mapping()``. Alternatively, ``setuptools`` **MAY** attempt to use
+       :doc:`import hooks <python:reference/import>` to redirect any attempt to import
+       to the directory with the original source code and other files built in place.
+
+    Please note that custom sub-commands **SHOULD NOT** rely on ``run()`` being
+    executed (or not) to provide correct return values for ``get_outputs()``,
+    ``get_output_mapping()`` or ``get_source_files()``. The ``get_*`` methods should
+    work independently of ``run()``.
+    """
+
+    editable_mode: bool = False
+    """Boolean flag that will be set to ``True`` when setuptools is used for an
+    editable installation (see :pep:`660`).
+    Implementations **SHOULD** explicitly set the default value of this attribute to
+    ``False``.
+    When subcommands run, they can use this flag to perform optimizations or change
+    their behaviour accordingly.
+    """
+
+    build_lib: str
+    """String representing the directory where the build artifacts should be stored,
+    e.g. ``build/lib``.
+    For example, if a distribution wants to provide a Python module named ``pkg.mod``,
+    then a corresponding file should be written to ``{build_lib}/package/module.py``.
+    A way of thinking about this is that the files saved under ``build_lib``
+    would be eventually copied to one of the directories in :obj:`site.PREFIXES`
+    upon installation.
+
+    A command that produces platform-independent files (e.g. compiling text templates
+    into Python functions), **CAN** initialize ``build_lib`` by copying its value from
+    the ``build_py`` command. On the other hand, a command that produces
+    platform-specific files **CAN** initialize ``build_lib`` by copying its value from
+    the ``build_ext`` command. In general this is done inside the ``finalize_options``
+    method with the help of the ``set_undefined_options`` command::
+
+        def finalize_options(self):
+            self.set_undefined_options("build_py", ("build_lib", "build_lib"))
+            ...
+    """
+
+    def initialize_options(self) -> None:
+        """(Required by the original :class:`setuptools.Command` interface)"""
+        ...
+
+    def finalize_options(self) -> None:
+        """(Required by the original :class:`setuptools.Command` interface)"""
+        ...
+
+    def run(self) -> None:
+        """(Required by the original :class:`setuptools.Command` interface)"""
+        ...
+
+    def get_source_files(self) -> list[str]:
+        """
+        Return a list of all files that are used by the command to create the expected
+        outputs.
+        For example, if your build command transpiles Java files into Python, you should
+        list here all the Java files.
+        The primary purpose of this function is to help populating the ``sdist``
+        with all the files necessary to build the distribution.
+        All files should be strings relative to the project root directory.
+        """
+        ...
+
+    def get_outputs(self) -> list[str]:
+        """
+        Return a list of files intended for distribution as they would have been
+        produced by the build.
+        These files should be strings in the form of
+        ``"{build_lib}/destination/file/path"``.
+
+        .. note::
+           The return value of ``get_output()`` should include all files used as keys
+           in ``get_output_mapping()`` plus files that are generated during the build
+           and don't correspond to any source file already present in the project.
+        """
+        ...
+
+    def get_output_mapping(self) -> dict[str, str]:
+        """
+        Return a mapping between destination files as they would be produced by the
+        build (dict keys) into the respective existing (source) files (dict values).
+        Existing (source) files should be represented as strings relative to the project
+        root directory.
+        Destination files should be strings in the form of
+        ``"{build_lib}/destination/file/path"``.
+        """
+        ...
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/build_clib.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_clib.py
new file mode 100644
index 00000000..f376f4ce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_clib.py
@@ -0,0 +1,103 @@
+from ..dist import Distribution
+from ..modified import newer_pairwise_group
+
+import distutils.command.build_clib as orig
+from distutils import log
+from distutils.errors import DistutilsSetupError
+
+
+class build_clib(orig.build_clib):
+    """
+    Override the default build_clib behaviour to do the following:
+
+    1. Implement a rudimentary timestamp-based dependency system
+       so 'compile()' doesn't run every time.
+    2. Add more keys to the 'build_info' dictionary:
+        * obj_deps - specify dependencies for each object compiled.
+                     this should be a dictionary mapping a key
+                     with the source filename to a list of
+                     dependencies. Use an empty string for global
+                     dependencies.
+        * cflags   - specify a list of additional flags to pass to
+                     the compiler.
+    """
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    def build_libraries(self, libraries) -> None:
+        for lib_name, build_info in libraries:
+            sources = build_info.get('sources')
+            if sources is None or not isinstance(sources, (list, tuple)):
+                raise DistutilsSetupError(
+                    f"in 'libraries' option (library '{lib_name}'), "
+                    "'sources' must be present and must be "
+                    "a list of source filenames"
+                )
+            sources = sorted(list(sources))
+
+            log.info("building '%s' library", lib_name)
+
+            # Make sure everything is the correct type.
+            # obj_deps should be a dictionary of keys as sources
+            # and a list/tuple of files that are its dependencies.
+            obj_deps = build_info.get('obj_deps', dict())
+            if not isinstance(obj_deps, dict):
+                raise DistutilsSetupError(
+                    f"in 'libraries' option (library '{lib_name}'), "
+                    "'obj_deps' must be a dictionary of "
+                    "type 'source: list'"
+                )
+            dependencies = []
+
+            # Get the global dependencies that are specified by the '' key.
+            # These will go into every source's dependency list.
+            global_deps = obj_deps.get('', list())
+            if not isinstance(global_deps, (list, tuple)):
+                raise DistutilsSetupError(
+                    f"in 'libraries' option (library '{lib_name}'), "
+                    "'obj_deps' must be a dictionary of "
+                    "type 'source: list'"
+                )
+
+            # Build the list to be used by newer_pairwise_group
+            # each source will be auto-added to its dependencies.
+            for source in sources:
+                src_deps = [source]
+                src_deps.extend(global_deps)
+                extra_deps = obj_deps.get(source, list())
+                if not isinstance(extra_deps, (list, tuple)):
+                    raise DistutilsSetupError(
+                        f"in 'libraries' option (library '{lib_name}'), "
+                        "'obj_deps' must be a dictionary of "
+                        "type 'source: list'"
+                    )
+                src_deps.extend(extra_deps)
+                dependencies.append(src_deps)
+
+            expected_objects = self.compiler.object_filenames(
+                sources,
+                output_dir=self.build_temp,
+            )
+
+            if newer_pairwise_group(dependencies, expected_objects) != ([], []):
+                # First, compile the source code to object files in the library
+                # directory.  (This should probably change to putting object
+                # files in a temporary build directory.)
+                macros = build_info.get('macros')
+                include_dirs = build_info.get('include_dirs')
+                cflags = build_info.get('cflags')
+                self.compiler.compile(
+                    sources,
+                    output_dir=self.build_temp,
+                    macros=macros,
+                    include_dirs=include_dirs,
+                    extra_postargs=cflags,
+                    debug=self.debug,
+                )
+
+            # Now "link" the object files together into a static library.
+            # (On Unix at least, this isn't really linking -- it just
+            # builds an archive.  Whatever.)
+            self.compiler.create_static_lib(
+                expected_objects, lib_name, output_dir=self.build_clib, debug=self.debug
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/build_ext.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_ext.py
new file mode 100644
index 00000000..be833a37
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_ext.py
@@ -0,0 +1,469 @@
+from __future__ import annotations
+
+import itertools
+import os
+import sys
+from collections.abc import Iterator
+from importlib.machinery import EXTENSION_SUFFIXES
+from importlib.util import cache_from_source as _compiled_file_name
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+from setuptools.dist import Distribution
+from setuptools.errors import BaseError
+from setuptools.extension import Extension, Library
+
+from distutils import log
+from distutils.ccompiler import new_compiler
+from distutils.sysconfig import customize_compiler, get_config_var
+
+if TYPE_CHECKING:
+    # Cython not installed on CI tests, causing _build_ext to be `Any`
+    from distutils.command.build_ext import build_ext as _build_ext
+else:
+    try:
+        # Attempt to use Cython for building extensions, if available
+        from Cython.Distutils.build_ext import build_ext as _build_ext
+
+        # Additionally, assert that the compiler module will load
+        # also. Ref #1229.
+        __import__('Cython.Compiler.Main')
+    except ImportError:
+        from distutils.command.build_ext import build_ext as _build_ext
+
+# make sure _config_vars is initialized
+get_config_var("LDSHARED")
+# Not publicly exposed in typeshed distutils stubs, but this is done on purpose
+# See https://github.com/pypa/setuptools/pull/4228#issuecomment-1959856400
+from distutils.sysconfig import _config_vars as _CONFIG_VARS  # noqa: E402
+
+
+def _customize_compiler_for_shlib(compiler):
+    if sys.platform == "darwin":
+        # building .dylib requires additional compiler flags on OSX; here we
+        # temporarily substitute the pyconfig.h variables so that distutils'
+        # 'customize_compiler' uses them before we build the shared libraries.
+        tmp = _CONFIG_VARS.copy()
+        try:
+            # XXX Help!  I don't have any idea whether these are right...
+            _CONFIG_VARS['LDSHARED'] = (
+                "gcc -Wl,-x -dynamiclib -undefined dynamic_lookup"
+            )
+            _CONFIG_VARS['CCSHARED'] = " -dynamiclib"
+            _CONFIG_VARS['SO'] = ".dylib"
+            customize_compiler(compiler)
+        finally:
+            _CONFIG_VARS.clear()
+            _CONFIG_VARS.update(tmp)
+    else:
+        customize_compiler(compiler)
+
+
+have_rtld = False
+use_stubs = False
+libtype = 'shared'
+
+if sys.platform == "darwin":
+    use_stubs = True
+elif os.name != 'nt':
+    try:
+        import dl  # type: ignore[import-not-found] # https://github.com/python/mypy/issues/13002
+
+        use_stubs = have_rtld = hasattr(dl, 'RTLD_NOW')
+    except ImportError:
+        pass
+
+
+def if_dl(s):
+    return s if have_rtld else ''
+
+
+def get_abi3_suffix():
+    """Return the file extension for an abi3-compliant Extension()"""
+    for suffix in EXTENSION_SUFFIXES:
+        if '.abi3' in suffix:  # Unix
+            return suffix
+        elif suffix == '.pyd':  # Windows
+            return suffix
+    return None
+
+
+class build_ext(_build_ext):
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+    editable_mode = False
+    inplace = False
+
+    def run(self):
+        """Build extensions in build directory, then copy if --inplace"""
+        old_inplace, self.inplace = self.inplace, False
+        _build_ext.run(self)
+        self.inplace = old_inplace
+        if old_inplace:
+            self.copy_extensions_to_source()
+
+    def _get_inplace_equivalent(self, build_py, ext: Extension) -> tuple[str, str]:
+        fullname = self.get_ext_fullname(ext.name)
+        filename = self.get_ext_filename(fullname)
+        modpath = fullname.split('.')
+        package = '.'.join(modpath[:-1])
+        package_dir = build_py.get_package_dir(package)
+        inplace_file = os.path.join(package_dir, os.path.basename(filename))
+        regular_file = os.path.join(self.build_lib, filename)
+        return (inplace_file, regular_file)
+
+    def copy_extensions_to_source(self) -> None:
+        build_py = self.get_finalized_command('build_py')
+        for ext in self.extensions:
+            inplace_file, regular_file = self._get_inplace_equivalent(build_py, ext)
+
+            # Always copy, even if source is older than destination, to ensure
+            # that the right extensions for the current Python/platform are
+            # used.
+            if os.path.exists(regular_file) or not ext.optional:
+                self.copy_file(regular_file, inplace_file, level=self.verbose)
+
+            if ext._needs_stub:
+                inplace_stub = self._get_equivalent_stub(ext, inplace_file)
+                self._write_stub_file(inplace_stub, ext, compile=True)
+                # Always compile stub and remove the original (leave the cache behind)
+                # (this behaviour was observed in previous iterations of the code)
+
+    def _get_equivalent_stub(self, ext: Extension, output_file: str) -> str:
+        dir_ = os.path.dirname(output_file)
+        _, _, name = ext.name.rpartition(".")
+        return f"{os.path.join(dir_, name)}.py"
+
+    def _get_output_mapping(self) -> Iterator[tuple[str, str]]:
+        if not self.inplace:
+            return
+
+        build_py = self.get_finalized_command('build_py')
+        opt = self.get_finalized_command('install_lib').optimize or ""
+
+        for ext in self.extensions:
+            inplace_file, regular_file = self._get_inplace_equivalent(build_py, ext)
+            yield (regular_file, inplace_file)
+
+            if ext._needs_stub:
+                # This version of `build_ext` always builds artifacts in another dir,
+                # when "inplace=True" is given it just copies them back.
+                # This is done in the `copy_extensions_to_source` function, which
+                # always compile stub files via `_compile_and_remove_stub`.
+                # At the end of the process, a `.pyc` stub file is created without the
+                # corresponding `.py`.
+
+                inplace_stub = self._get_equivalent_stub(ext, inplace_file)
+                regular_stub = self._get_equivalent_stub(ext, regular_file)
+                inplace_cache = _compiled_file_name(inplace_stub, optimization=opt)
+                output_cache = _compiled_file_name(regular_stub, optimization=opt)
+                yield (output_cache, inplace_cache)
+
+    def get_ext_filename(self, fullname: str) -> str:
+        so_ext = os.getenv('SETUPTOOLS_EXT_SUFFIX')
+        if so_ext:
+            filename = os.path.join(*fullname.split('.')) + so_ext
+        else:
+            filename = _build_ext.get_ext_filename(self, fullname)
+            ext_suffix = get_config_var('EXT_SUFFIX')
+            if not isinstance(ext_suffix, str):
+                raise OSError(
+                    "Configuration variable EXT_SUFFIX not found for this platform "
+                    "and environment variable SETUPTOOLS_EXT_SUFFIX is missing"
+                )
+            so_ext = ext_suffix
+
+        if fullname in self.ext_map:
+            ext = self.ext_map[fullname]
+            abi3_suffix = get_abi3_suffix()
+            if ext.py_limited_api and abi3_suffix:  # Use abi3
+                filename = filename[: -len(so_ext)] + abi3_suffix
+            if isinstance(ext, Library):
+                fn, ext = os.path.splitext(filename)
+                return self.shlib_compiler.library_filename(fn, libtype)
+            elif use_stubs and ext._links_to_dynamic:
+                d, fn = os.path.split(filename)
+                return os.path.join(d, 'dl-' + fn)
+        return filename
+
+    def initialize_options(self):
+        _build_ext.initialize_options(self)
+        self.shlib_compiler = None
+        self.shlibs = []
+        self.ext_map = {}
+        self.editable_mode = False
+
+    def finalize_options(self) -> None:
+        _build_ext.finalize_options(self)
+        self.extensions = self.extensions or []
+        self.check_extensions_list(self.extensions)
+        self.shlibs = [ext for ext in self.extensions if isinstance(ext, Library)]
+        if self.shlibs:
+            self.setup_shlib_compiler()
+        for ext in self.extensions:
+            ext._full_name = self.get_ext_fullname(ext.name)
+        for ext in self.extensions:
+            fullname = ext._full_name
+            self.ext_map[fullname] = ext
+
+            # distutils 3.1 will also ask for module names
+            # XXX what to do with conflicts?
+            self.ext_map[fullname.split('.')[-1]] = ext
+
+            ltd = self.shlibs and self.links_to_dynamic(ext) or False
+            ns = ltd and use_stubs and not isinstance(ext, Library)
+            ext._links_to_dynamic = ltd
+            ext._needs_stub = ns
+            filename = ext._file_name = self.get_ext_filename(fullname)
+            libdir = os.path.dirname(os.path.join(self.build_lib, filename))
+            if ltd and libdir not in ext.library_dirs:
+                ext.library_dirs.append(libdir)
+            if ltd and use_stubs and os.curdir not in ext.runtime_library_dirs:
+                ext.runtime_library_dirs.append(os.curdir)
+
+        if self.editable_mode:
+            self.inplace = True
+
+    def setup_shlib_compiler(self):
+        compiler = self.shlib_compiler = new_compiler(
+            compiler=self.compiler, dry_run=self.dry_run, force=self.force
+        )
+        _customize_compiler_for_shlib(compiler)
+
+        if self.include_dirs is not None:
+            compiler.set_include_dirs(self.include_dirs)
+        if self.define is not None:
+            # 'define' option is a list of (name,value) tuples
+            for name, value in self.define:
+                compiler.define_macro(name, value)
+        if self.undef is not None:
+            for macro in self.undef:
+                compiler.undefine_macro(macro)
+        if self.libraries is not None:
+            compiler.set_libraries(self.libraries)
+        if self.library_dirs is not None:
+            compiler.set_library_dirs(self.library_dirs)
+        if self.rpath is not None:
+            compiler.set_runtime_library_dirs(self.rpath)
+        if self.link_objects is not None:
+            compiler.set_link_objects(self.link_objects)
+
+        # hack so distutils' build_extension() builds a library instead
+        compiler.link_shared_object = link_shared_object.__get__(compiler)  # type: ignore[method-assign]
+
+    def get_export_symbols(self, ext):
+        if isinstance(ext, Library):
+            return ext.export_symbols
+        return _build_ext.get_export_symbols(self, ext)
+
+    def build_extension(self, ext) -> None:
+        ext._convert_pyx_sources_to_lang()
+        _compiler = self.compiler
+        try:
+            if isinstance(ext, Library):
+                self.compiler = self.shlib_compiler
+            _build_ext.build_extension(self, ext)
+            if ext._needs_stub:
+                build_lib = self.get_finalized_command('build_py').build_lib
+                self.write_stub(build_lib, ext)
+        finally:
+            self.compiler = _compiler
+
+    def links_to_dynamic(self, ext):
+        """Return true if 'ext' links to a dynamic lib in the same package"""
+        # XXX this should check to ensure the lib is actually being built
+        # XXX as dynamic, and not just using a locally-found version or a
+        # XXX static-compiled version
+        libnames = dict.fromkeys([lib._full_name for lib in self.shlibs])
+        pkg = '.'.join(ext._full_name.split('.')[:-1] + [''])
+        return any(pkg + libname in libnames for libname in ext.libraries)
+
+    def get_source_files(self) -> list[str]:
+        return [*_build_ext.get_source_files(self), *self._get_internal_depends()]
+
+    def _get_internal_depends(self) -> Iterator[str]:
+        """Yield ``ext.depends`` that are contained by the project directory"""
+        project_root = Path(self.distribution.src_root or os.curdir).resolve()
+        depends = (dep for ext in self.extensions for dep in ext.depends)
+
+        def skip(orig_path: str, reason: str) -> None:
+            log.info(
+                "dependency %s won't be automatically "
+                "included in the manifest: the path %s",
+                orig_path,
+                reason,
+            )
+
+        for dep in depends:
+            path = Path(dep)
+
+            if path.is_absolute():
+                skip(dep, "must be relative")
+                continue
+
+            if ".." in path.parts:
+                skip(dep, "can't have `..` segments")
+                continue
+
+            try:
+                resolved = (project_root / path).resolve(strict=True)
+            except OSError:
+                skip(dep, "doesn't exist")
+                continue
+
+            try:
+                resolved.relative_to(project_root)
+            except ValueError:
+                skip(dep, "must be inside the project root")
+                continue
+
+            yield path.as_posix()
+
+    def get_outputs(self) -> list[str]:
+        if self.inplace:
+            return list(self.get_output_mapping().keys())
+        return sorted(_build_ext.get_outputs(self) + self.__get_stubs_outputs())
+
+    def get_output_mapping(self) -> dict[str, str]:
+        """See :class:`setuptools.commands.build.SubCommand`"""
+        mapping = self._get_output_mapping()
+        return dict(sorted(mapping, key=lambda x: x[0]))
+
+    def __get_stubs_outputs(self):
+        # assemble the base name for each extension that needs a stub
+        ns_ext_bases = (
+            os.path.join(self.build_lib, *ext._full_name.split('.'))
+            for ext in self.extensions
+            if ext._needs_stub
+        )
+        # pair each base with the extension
+        pairs = itertools.product(ns_ext_bases, self.__get_output_extensions())
+        return list(base + fnext for base, fnext in pairs)
+
+    def __get_output_extensions(self):
+        yield '.py'
+        yield '.pyc'
+        if self.get_finalized_command('build_py').optimize:
+            yield '.pyo'
+
+    def write_stub(self, output_dir, ext, compile=False) -> None:
+        stub_file = os.path.join(output_dir, *ext._full_name.split('.')) + '.py'
+        self._write_stub_file(stub_file, ext, compile)
+
+    def _write_stub_file(self, stub_file: str, ext: Extension, compile=False):
+        log.info("writing stub loader for %s to %s", ext._full_name, stub_file)
+        if compile and os.path.exists(stub_file):
+            raise BaseError(stub_file + " already exists! Please delete.")
+        if not self.dry_run:
+            with open(stub_file, 'w', encoding="utf-8") as f:
+                content = '\n'.join([
+                    "def __bootstrap__():",
+                    "   global __bootstrap__, __file__, __loader__",
+                    "   import sys, os, pkg_resources, importlib.util" + if_dl(", dl"),
+                    "   __file__ = pkg_resources.resource_filename"
+                    f"(__name__,{os.path.basename(ext._file_name)!r})",
+                    "   del __bootstrap__",
+                    "   if '__loader__' in globals():",
+                    "       del __loader__",
+                    if_dl("   old_flags = sys.getdlopenflags()"),
+                    "   old_dir = os.getcwd()",
+                    "   try:",
+                    "     os.chdir(os.path.dirname(__file__))",
+                    if_dl("     sys.setdlopenflags(dl.RTLD_NOW)"),
+                    "     spec = importlib.util.spec_from_file_location(",
+                    "                __name__, __file__)",
+                    "     mod = importlib.util.module_from_spec(spec)",
+                    "     spec.loader.exec_module(mod)",
+                    "   finally:",
+                    if_dl("     sys.setdlopenflags(old_flags)"),
+                    "     os.chdir(old_dir)",
+                    "__bootstrap__()",
+                    "",  # terminal \n
+                ])
+                f.write(content)
+        if compile:
+            self._compile_and_remove_stub(stub_file)
+
+    def _compile_and_remove_stub(self, stub_file: str):
+        from distutils.util import byte_compile
+
+        byte_compile([stub_file], optimize=0, force=True, dry_run=self.dry_run)
+        optimize = self.get_finalized_command('install_lib').optimize
+        if optimize > 0:
+            byte_compile(
+                [stub_file],
+                optimize=optimize,
+                force=True,
+                dry_run=self.dry_run,
+            )
+        if os.path.exists(stub_file) and not self.dry_run:
+            os.unlink(stub_file)
+
+
+if use_stubs or os.name == 'nt':
+    # Build shared libraries
+    #
+    def link_shared_object(
+        self,
+        objects,
+        output_libname,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug: bool = False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ) -> None:
+        self.link(
+            self.SHARED_LIBRARY,
+            objects,
+            output_libname,
+            output_dir,
+            libraries,
+            library_dirs,
+            runtime_library_dirs,
+            export_symbols,
+            debug,
+            extra_preargs,
+            extra_postargs,
+            build_temp,
+            target_lang,
+        )
+
+else:
+    # Build static libraries everywhere else
+    libtype = 'static'
+
+    def link_shared_object(
+        self,
+        objects,
+        output_libname,
+        output_dir=None,
+        libraries=None,
+        library_dirs=None,
+        runtime_library_dirs=None,
+        export_symbols=None,
+        debug: bool = False,
+        extra_preargs=None,
+        extra_postargs=None,
+        build_temp=None,
+        target_lang=None,
+    ) -> None:
+        # XXX we need to either disallow these attrs on Library instances,
+        # or warn/abort here if set, or something...
+        # libraries=None, library_dirs=None, runtime_library_dirs=None,
+        # export_symbols=None, extra_preargs=None, extra_postargs=None,
+        # build_temp=None
+
+        assert output_dir is None  # distutils build_ext doesn't pass this
+        output_dir, filename = os.path.split(output_libname)
+        basename, _ext = os.path.splitext(filename)
+        if self.library_filename("x").startswith('lib'):
+            # strip 'lib' prefix; this is kludgy if some platform uses
+            # a different prefix
+            basename = basename[3:]
+
+        self.create_static_lib(objects, basename, output_dir, debug, target_lang)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/build_py.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_py.py
new file mode 100644
index 00000000..2f6fcb7c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/build_py.py
@@ -0,0 +1,400 @@
+from __future__ import annotations
+
+import fnmatch
+import itertools
+import os
+import stat
+import textwrap
+from collections.abc import Iterable, Iterator
+from functools import partial
+from glob import glob
+from pathlib import Path
+
+from more_itertools import unique_everseen
+
+from .._path import StrPath, StrPathT
+from ..dist import Distribution
+from ..warnings import SetuptoolsDeprecationWarning
+
+import distutils.command.build_py as orig
+import distutils.errors
+from distutils.util import convert_path
+
+_IMPLICIT_DATA_FILES = ('*.pyi', 'py.typed')
+
+
+def make_writable(target) -> None:
+    os.chmod(target, os.stat(target).st_mode | stat.S_IWRITE)
+
+
+class build_py(orig.build_py):
+    """Enhanced 'build_py' command that includes data files with packages
+
+    The data files are specified via a 'package_data' argument to 'setup()'.
+    See 'setuptools.dist.Distribution' for more details.
+
+    Also, this version of the 'build_py' command allows you to specify both
+    'py_modules' and 'packages' in the same setup operation.
+    """
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+    editable_mode: bool = False
+    existing_egg_info_dir: StrPath | None = None  #: Private API, internal use only.
+
+    def finalize_options(self):
+        orig.build_py.finalize_options(self)
+        self.package_data = self.distribution.package_data
+        self.exclude_package_data = self.distribution.exclude_package_data or {}
+        if 'data_files' in self.__dict__:
+            del self.__dict__['data_files']
+
+    def copy_file(  # type: ignore[override] # No overload, no bytes support
+        self,
+        infile: StrPath,
+        outfile: StrPathT,
+        preserve_mode: bool = True,
+        preserve_times: bool = True,
+        link: str | None = None,
+        level: object = 1,
+    ) -> tuple[StrPathT | str, bool]:
+        # Overwrite base class to allow using links
+        if link:
+            infile = str(Path(infile).resolve())
+            outfile = str(Path(outfile).resolve())  # type: ignore[assignment] # Re-assigning a str when outfile is StrPath is ok
+        return super().copy_file(  # pyright: ignore[reportReturnType] # pypa/distutils#309
+            infile, outfile, preserve_mode, preserve_times, link, level
+        )
+
+    def run(self) -> None:
+        """Build modules, packages, and copy data files to build directory"""
+        if not (self.py_modules or self.packages) or self.editable_mode:
+            return
+
+        if self.py_modules:
+            self.build_modules()
+
+        if self.packages:
+            self.build_packages()
+            self.build_package_data()
+
+        # Only compile actual .py files, using our base class' idea of what our
+        # output files are.
+        self.byte_compile(orig.build_py.get_outputs(self, include_bytecode=False))
+
+    def __getattr__(self, attr: str):
+        "lazily compute data files"
+        if attr == 'data_files':
+            self.data_files = self._get_data_files()
+            return self.data_files
+        return orig.build_py.__getattr__(self, attr)
+
+    def _get_data_files(self):
+        """Generate list of '(package,src_dir,build_dir,filenames)' tuples"""
+        self.analyze_manifest()
+        return list(map(self._get_pkg_data_files, self.packages or ()))
+
+    def get_data_files_without_manifest(self):
+        """
+        Generate list of ``(package,src_dir,build_dir,filenames)`` tuples,
+        but without triggering any attempt to analyze or build the manifest.
+        """
+        # Prevent eventual errors from unset `manifest_files`
+        # (that would otherwise be set by `analyze_manifest`)
+        self.__dict__.setdefault('manifest_files', {})
+        return list(map(self._get_pkg_data_files, self.packages or ()))
+
+    def _get_pkg_data_files(self, package):
+        # Locate package source directory
+        src_dir = self.get_package_dir(package)
+
+        # Compute package build directory
+        build_dir = os.path.join(*([self.build_lib] + package.split('.')))
+
+        # Strip directory from globbed filenames
+        filenames = [
+            os.path.relpath(file, src_dir)
+            for file in self.find_data_files(package, src_dir)
+        ]
+        return package, src_dir, build_dir, filenames
+
+    def find_data_files(self, package, src_dir):
+        """Return filenames for package's data files in 'src_dir'"""
+        patterns = self._get_platform_patterns(
+            self.package_data,
+            package,
+            src_dir,
+            extra_patterns=_IMPLICIT_DATA_FILES,
+        )
+        globs_expanded = map(partial(glob, recursive=True), patterns)
+        # flatten the expanded globs into an iterable of matches
+        globs_matches = itertools.chain.from_iterable(globs_expanded)
+        glob_files = filter(os.path.isfile, globs_matches)
+        files = itertools.chain(
+            self.manifest_files.get(package, []),
+            glob_files,
+        )
+        return self.exclude_data_files(package, src_dir, files)
+
+    def get_outputs(self, include_bytecode: bool = True) -> list[str]:  # type: ignore[override] # Using a real boolean instead of 0|1
+        """See :class:`setuptools.commands.build.SubCommand`"""
+        if self.editable_mode:
+            return list(self.get_output_mapping().keys())
+        return super().get_outputs(include_bytecode)
+
+    def get_output_mapping(self) -> dict[str, str]:
+        """See :class:`setuptools.commands.build.SubCommand`"""
+        mapping = itertools.chain(
+            self._get_package_data_output_mapping(),
+            self._get_module_mapping(),
+        )
+        return dict(sorted(mapping, key=lambda x: x[0]))
+
+    def _get_module_mapping(self) -> Iterator[tuple[str, str]]:
+        """Iterate over all modules producing (dest, src) pairs."""
+        for package, module, module_file in self.find_all_modules():
+            package = package.split('.')
+            filename = self.get_module_outfile(self.build_lib, package, module)
+            yield (filename, module_file)
+
+    def _get_package_data_output_mapping(self) -> Iterator[tuple[str, str]]:
+        """Iterate over package data producing (dest, src) pairs."""
+        for package, src_dir, build_dir, filenames in self.data_files:
+            for filename in filenames:
+                target = os.path.join(build_dir, filename)
+                srcfile = os.path.join(src_dir, filename)
+                yield (target, srcfile)
+
+    def build_package_data(self) -> None:
+        """Copy data files into build directory"""
+        for target, srcfile in self._get_package_data_output_mapping():
+            self.mkpath(os.path.dirname(target))
+            _outf, _copied = self.copy_file(srcfile, target)
+            make_writable(target)
+
+    def analyze_manifest(self) -> None:
+        self.manifest_files: dict[str, list[str]] = {}
+        if not self.distribution.include_package_data:
+            return
+        src_dirs: dict[str, str] = {}
+        for package in self.packages or ():
+            # Locate package source directory
+            src_dirs[assert_relative(self.get_package_dir(package))] = package
+
+        if (
+            self.existing_egg_info_dir
+            and Path(self.existing_egg_info_dir, "SOURCES.txt").exists()
+        ):
+            egg_info_dir = self.existing_egg_info_dir
+            manifest = Path(egg_info_dir, "SOURCES.txt")
+            files = manifest.read_text(encoding="utf-8").splitlines()
+        else:
+            self.run_command('egg_info')
+            ei_cmd = self.get_finalized_command('egg_info')
+            egg_info_dir = ei_cmd.egg_info
+            files = ei_cmd.filelist.files
+
+        check = _IncludePackageDataAbuse()
+        for path in self._filter_build_files(files, egg_info_dir):
+            d, f = os.path.split(assert_relative(path))
+            prev = None
+            oldf = f
+            while d and d != prev and d not in src_dirs:
+                prev = d
+                d, df = os.path.split(d)
+                f = os.path.join(df, f)
+            if d in src_dirs:
+                if f == oldf:
+                    if check.is_module(f):
+                        continue  # it's a module, not data
+                else:
+                    importable = check.importable_subpackage(src_dirs[d], f)
+                    if importable:
+                        check.warn(importable)
+                self.manifest_files.setdefault(src_dirs[d], []).append(path)
+
+    def _filter_build_files(
+        self, files: Iterable[str], egg_info: StrPath
+    ) -> Iterator[str]:
+        """
+        ``build_meta`` may try to create egg_info outside of the project directory,
+        and this can be problematic for certain plugins (reported in issue #3500).
+
+        Extensions might also include between their sources files created on the
+        ``build_lib`` and ``build_temp`` directories.
+
+        This function should filter this case of invalid files out.
+        """
+        build = self.get_finalized_command("build")
+        build_dirs = (egg_info, self.build_lib, build.build_temp, build.build_base)
+        norm_dirs = [os.path.normpath(p) for p in build_dirs if p]
+
+        for file in files:
+            norm_path = os.path.normpath(file)
+            if not os.path.isabs(file) or all(d not in norm_path for d in norm_dirs):
+                yield file
+
+    def get_data_files(self) -> None:
+        pass  # Lazily compute data files in _get_data_files() function.
+
+    def check_package(self, package, package_dir):
+        """Check namespace packages' __init__ for declare_namespace"""
+        try:
+            return self.packages_checked[package]
+        except KeyError:
+            pass
+
+        init_py = orig.build_py.check_package(self, package, package_dir)
+        self.packages_checked[package] = init_py
+
+        if not init_py or not self.distribution.namespace_packages:
+            return init_py
+
+        for pkg in self.distribution.namespace_packages:
+            if pkg == package or pkg.startswith(package + '.'):
+                break
+        else:
+            return init_py
+
+        with open(init_py, 'rb') as f:
+            contents = f.read()
+        if b'declare_namespace' not in contents:
+            raise distutils.errors.DistutilsError(
+                f"Namespace package problem: {package} is a namespace package, but "
+                "its\n__init__.py does not call declare_namespace()! Please "
+                'fix it.\n(See the setuptools manual under '
+                '"Namespace Packages" for details.)\n"'
+            )
+        return init_py
+
+    def initialize_options(self):
+        self.packages_checked = {}
+        orig.build_py.initialize_options(self)
+        self.editable_mode = False
+        self.existing_egg_info_dir = None
+
+    def get_package_dir(self, package):
+        res = orig.build_py.get_package_dir(self, package)
+        if self.distribution.src_root is not None:
+            return os.path.join(self.distribution.src_root, res)
+        return res
+
+    def exclude_data_files(self, package, src_dir, files):
+        """Filter filenames for package's data files in 'src_dir'"""
+        files = list(files)
+        patterns = self._get_platform_patterns(
+            self.exclude_package_data,
+            package,
+            src_dir,
+        )
+        match_groups = (fnmatch.filter(files, pattern) for pattern in patterns)
+        # flatten the groups of matches into an iterable of matches
+        matches = itertools.chain.from_iterable(match_groups)
+        bad = set(matches)
+        keepers = (fn for fn in files if fn not in bad)
+        # ditch dupes
+        return list(unique_everseen(keepers))
+
+    @staticmethod
+    def _get_platform_patterns(spec, package, src_dir, extra_patterns=()):
+        """
+        yield platform-specific path patterns (suitable for glob
+        or fn_match) from a glob-based spec (such as
+        self.package_data or self.exclude_package_data)
+        matching package in src_dir.
+        """
+        raw_patterns = itertools.chain(
+            extra_patterns,
+            spec.get('', []),
+            spec.get(package, []),
+        )
+        return (
+            # Each pattern has to be converted to a platform-specific path
+            os.path.join(src_dir, convert_path(pattern))
+            for pattern in raw_patterns
+        )
+
+
+def assert_relative(path):
+    if not os.path.isabs(path):
+        return path
+    from distutils.errors import DistutilsSetupError
+
+    msg = (
+        textwrap.dedent(
+            """
+        Error: setup script specifies an absolute path:
+
+            %s
+
+        setup() arguments must *always* be /-separated paths relative to the
+        setup.py directory, *never* absolute paths.
+        """
+        ).lstrip()
+        % path
+    )
+    raise DistutilsSetupError(msg)
+
+
+class _IncludePackageDataAbuse:
+    """Inform users that package or module is included as 'data file'"""
+
+    class _Warning(SetuptoolsDeprecationWarning):
+        _SUMMARY = """
+        Package {importable!r} is absent from the `packages` configuration.
+        """
+
+        _DETAILS = """
+        ############################
+        # Package would be ignored #
+        ############################
+        Python recognizes {importable!r} as an importable package[^1],
+        but it is absent from setuptools' `packages` configuration.
+
+        This leads to an ambiguous overall configuration. If you want to distribute this
+        package, please make sure that {importable!r} is explicitly added
+        to the `packages` configuration field.
+
+        Alternatively, you can also rely on setuptools' discovery methods
+        (for example by using `find_namespace_packages(...)`/`find_namespace:`
+        instead of `find_packages(...)`/`find:`).
+
+        You can read more about "package discovery" on setuptools documentation page:
+
+        - https://setuptools.pypa.io/en/latest/userguide/package_discovery.html
+
+        If you don't want {importable!r} to be distributed and are
+        already explicitly excluding {importable!r} via
+        `find_namespace_packages(...)/find_namespace` or `find_packages(...)/find`,
+        you can try to use `exclude_package_data`, or `include-package-data=False` in
+        combination with a more fine grained `package-data` configuration.
+
+        You can read more about "package data files" on setuptools documentation page:
+
+        - https://setuptools.pypa.io/en/latest/userguide/datafiles.html
+
+
+        [^1]: For Python, any directory (with suitable naming) can be imported,
+              even if it does not contain any `.py` files.
+              On the other hand, currently there is no concept of package data
+              directory, all directories are treated like packages.
+        """
+        # _DUE_DATE: still not defined as this is particularly controversial.
+        # Warning initially introduced in May 2022. See issue #3340 for discussion.
+
+    def __init__(self):
+        self._already_warned = set()
+
+    def is_module(self, file):
+        return file.endswith(".py") and file[: -len(".py")].isidentifier()
+
+    def importable_subpackage(self, parent, file):
+        pkg = Path(file).parent
+        parts = list(itertools.takewhile(str.isidentifier, pkg.parts))
+        if parts:
+            return ".".join([parent, *parts])
+        return None
+
+    def warn(self, importable):
+        if importable not in self._already_warned:
+            self._Warning.emit(importable=importable)
+            self._already_warned.add(importable)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/develop.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/develop.py
new file mode 100644
index 00000000..7eee29d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/develop.py
@@ -0,0 +1,195 @@
+import glob
+import os
+
+import setuptools
+from setuptools import _normalization, _path, namespaces
+from setuptools.command.easy_install import easy_install
+
+from ..unicode_utils import _read_utf8_with_fallback
+
+from distutils import log
+from distutils.errors import DistutilsOptionError
+from distutils.util import convert_path
+
+
+class develop(namespaces.DevelopInstaller, easy_install):
+    """Set up package for development"""
+
+    description = "install package in 'development mode'"
+
+    user_options = easy_install.user_options + [
+        ("uninstall", "u", "Uninstall this source package"),
+        ("egg-path=", None, "Set the path to be used in the .egg-link file"),
+    ]
+
+    boolean_options = easy_install.boolean_options + ['uninstall']
+
+    command_consumes_arguments = False  # override base
+
+    def run(self):
+        if self.uninstall:
+            self.multi_version = True
+            self.uninstall_link()
+            self.uninstall_namespaces()
+        else:
+            self.install_for_development()
+        self.warn_deprecated_options()
+
+    def initialize_options(self):
+        self.uninstall = None
+        self.egg_path = None
+        easy_install.initialize_options(self)
+        self.setup_path = None
+        self.always_copy_from = '.'  # always copy eggs installed in curdir
+
+    def finalize_options(self) -> None:
+        import pkg_resources
+
+        ei = self.get_finalized_command("egg_info")
+        self.args = [ei.egg_name]
+
+        easy_install.finalize_options(self)
+        self.expand_basedirs()
+        self.expand_dirs()
+        # pick up setup-dir .egg files only: no .egg-info
+        self.package_index.scan(glob.glob('*.egg'))
+
+        egg_link_fn = (
+            _normalization.filename_component_broken(ei.egg_name) + '.egg-link'
+        )
+        self.egg_link = os.path.join(self.install_dir, egg_link_fn)
+        self.egg_base = ei.egg_base
+        if self.egg_path is None:
+            self.egg_path = os.path.abspath(ei.egg_base)
+
+        target = _path.normpath(self.egg_base)
+        egg_path = _path.normpath(os.path.join(self.install_dir, self.egg_path))
+        if egg_path != target:
+            raise DistutilsOptionError(
+                "--egg-path must be a relative path from the install"
+                " directory to " + target
+            )
+
+        # Make a distribution for the package's source
+        self.dist = pkg_resources.Distribution(
+            target,
+            pkg_resources.PathMetadata(target, os.path.abspath(ei.egg_info)),
+            project_name=ei.egg_name,
+        )
+
+        self.setup_path = self._resolve_setup_path(
+            self.egg_base,
+            self.install_dir,
+            self.egg_path,
+        )
+
+    @staticmethod
+    def _resolve_setup_path(egg_base, install_dir, egg_path):
+        """
+        Generate a path from egg_base back to '.' where the
+        setup script resides and ensure that path points to the
+        setup path from $install_dir/$egg_path.
+        """
+        path_to_setup = egg_base.replace(os.sep, '/').rstrip('/')
+        if path_to_setup != os.curdir:
+            path_to_setup = '../' * (path_to_setup.count('/') + 1)
+        resolved = _path.normpath(os.path.join(install_dir, egg_path, path_to_setup))
+        curdir = _path.normpath(os.curdir)
+        if resolved != curdir:
+            raise DistutilsOptionError(
+                "Can't get a consistent path to setup script from"
+                " installation directory",
+                resolved,
+                curdir,
+            )
+        return path_to_setup
+
+    def install_for_development(self) -> None:
+        self.run_command('egg_info')
+
+        # Build extensions in-place
+        self.reinitialize_command('build_ext', inplace=True)
+        self.run_command('build_ext')
+
+        if setuptools.bootstrap_install_from:
+            self.easy_install(setuptools.bootstrap_install_from)
+            setuptools.bootstrap_install_from = None
+
+        self.install_namespaces()
+
+        # create an .egg-link in the installation dir, pointing to our egg
+        log.info("Creating %s (link to %s)", self.egg_link, self.egg_base)
+        if not self.dry_run:
+            with open(self.egg_link, "w", encoding="utf-8") as f:
+                f.write(self.egg_path + "\n" + self.setup_path)
+        # postprocess the installed distro, fixing up .pth, installing scripts,
+        # and handling requirements
+        self.process_distribution(None, self.dist, not self.no_deps)
+
+    def uninstall_link(self) -> None:
+        if os.path.exists(self.egg_link):
+            log.info("Removing %s (link to %s)", self.egg_link, self.egg_base)
+
+            contents = [
+                line.rstrip()
+                for line in _read_utf8_with_fallback(self.egg_link).splitlines()
+            ]
+
+            if contents not in ([self.egg_path], [self.egg_path, self.setup_path]):
+                log.warn("Link points to %s: uninstall aborted", contents)
+                return
+            if not self.dry_run:
+                os.unlink(self.egg_link)
+        if not self.dry_run:
+            self.update_pth(self.dist)  # remove any .pth link to us
+        if self.distribution.scripts:
+            # XXX should also check for entry point scripts!
+            log.warn("Note: you must uninstall or replace scripts manually!")
+
+    def install_egg_scripts(self, dist):
+        if dist is not self.dist:
+            # Installing a dependency, so fall back to normal behavior
+            return easy_install.install_egg_scripts(self, dist)
+
+        # create wrapper scripts in the script dir, pointing to dist.scripts
+
+        # new-style...
+        self.install_wrapper_scripts(dist)
+
+        # ...and old-style
+        for script_name in self.distribution.scripts or []:
+            script_path = os.path.abspath(convert_path(script_name))
+            script_name = os.path.basename(script_path)
+            script_text = _read_utf8_with_fallback(script_path)
+            self.install_script(dist, script_name, script_text, script_path)
+
+        return None
+
+    def install_wrapper_scripts(self, dist):
+        dist = VersionlessRequirement(dist)
+        return easy_install.install_wrapper_scripts(self, dist)
+
+
+class VersionlessRequirement:
+    """
+    Adapt a pkg_resources.Distribution to simply return the project
+    name as the 'requirement' so that scripts will work across
+    multiple versions.
+
+    >>> from pkg_resources import Distribution
+    >>> dist = Distribution(project_name='foo', version='1.0')
+    >>> str(dist.as_requirement())
+    'foo==1.0'
+    >>> adapted_dist = VersionlessRequirement(dist)
+    >>> str(adapted_dist.as_requirement())
+    'foo'
+    """
+
+    def __init__(self, dist) -> None:
+        self.__dist = dist
+
+    def __getattr__(self, name: str):
+        return getattr(self.__dist, name)
+
+    def as_requirement(self):
+        return self.project_name
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/dist_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/dist_info.py
new file mode 100644
index 00000000..dca01ff0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/dist_info.py
@@ -0,0 +1,103 @@
+"""
+Create a dist_info directory
+As defined in the wheel specification
+"""
+
+import os
+import shutil
+from contextlib import contextmanager
+from pathlib import Path
+from typing import cast
+
+from .. import _normalization
+from .._shutil import rmdir as _rm
+from .egg_info import egg_info as egg_info_cls
+
+from distutils import log
+from distutils.core import Command
+
+
+class dist_info(Command):
+    """
+    This command is private and reserved for internal use of setuptools,
+    users should rely on ``setuptools.build_meta`` APIs.
+    """
+
+    description = "DO NOT CALL DIRECTLY, INTERNAL ONLY: create .dist-info directory"
+
+    user_options = [
+        (
+            'output-dir=',
+            'o',
+            "directory inside of which the .dist-info will be"
+            "created [default: top of the source tree]",
+        ),
+        ('tag-date', 'd', "Add date stamp (e.g. 20050528) to version number"),
+        ('tag-build=', 'b', "Specify explicit tag to add to version number"),
+        ('no-date', 'D', "Don't include date stamp [default]"),
+        ('keep-egg-info', None, "*TRANSITIONAL* will be removed in the future"),
+    ]
+
+    boolean_options = ['tag-date', 'keep-egg-info']
+    negative_opt = {'no-date': 'tag-date'}
+
+    def initialize_options(self):
+        self.output_dir = None
+        self.name = None
+        self.dist_info_dir = None
+        self.tag_date = None
+        self.tag_build = None
+        self.keep_egg_info = False
+
+    def finalize_options(self) -> None:
+        dist = self.distribution
+        project_dir = dist.src_root or os.curdir
+        self.output_dir = Path(self.output_dir or project_dir)
+
+        egg_info = cast(egg_info_cls, self.reinitialize_command("egg_info"))
+        egg_info.egg_base = str(self.output_dir)
+
+        if self.tag_date:
+            egg_info.tag_date = self.tag_date
+        else:
+            self.tag_date = egg_info.tag_date
+
+        if self.tag_build:
+            egg_info.tag_build = self.tag_build
+        else:
+            self.tag_build = egg_info.tag_build
+
+        egg_info.finalize_options()
+        self.egg_info = egg_info
+
+        name = _normalization.safer_name(dist.get_name())
+        version = _normalization.safer_best_effort_version(dist.get_version())
+        self.name = f"{name}-{version}"
+        self.dist_info_dir = os.path.join(self.output_dir, f"{self.name}.dist-info")
+
+    @contextmanager
+    def _maybe_bkp_dir(self, dir_path: str, requires_bkp: bool):
+        if requires_bkp:
+            bkp_name = f"{dir_path}.__bkp__"
+            _rm(bkp_name, ignore_errors=True)
+            shutil.copytree(dir_path, bkp_name, dirs_exist_ok=True, symlinks=True)
+            try:
+                yield
+            finally:
+                _rm(dir_path, ignore_errors=True)
+                shutil.move(bkp_name, dir_path)
+        else:
+            yield
+
+    def run(self) -> None:
+        self.output_dir.mkdir(parents=True, exist_ok=True)
+        self.egg_info.run()
+        egg_info_dir = self.egg_info.egg_info
+        assert os.path.isdir(egg_info_dir), ".egg-info dir should have been created"
+
+        log.info(f"creating '{os.path.abspath(self.dist_info_dir)}'")
+        bdist_wheel = self.get_finalized_command('bdist_wheel')
+
+        # TODO: if bdist_wheel if merged into setuptools, just add "keep_egg_info" there
+        with self._maybe_bkp_dir(egg_info_dir, self.keep_egg_info):
+            bdist_wheel.egg2dist(egg_info_dir, self.dist_info_dir)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/easy_install.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/easy_install.py
new file mode 100644
index 00000000..eb1b4c1f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/easy_install.py
@@ -0,0 +1,2365 @@
+"""
+Easy Install
+------------
+
+A tool for doing automatic download/extract/build of distutils-based Python
+packages.  For detailed documentation, see the accompanying EasyInstall.txt
+file, or visit the `EasyInstall home page`__.
+
+__ https://setuptools.pypa.io/en/latest/deprecated/easy_install.html
+
+"""
+
+from __future__ import annotations
+
+import configparser
+import contextlib
+import io
+import os
+import random
+import re
+import shlex
+import shutil
+import site
+import stat
+import struct
+import subprocess
+import sys
+import sysconfig
+import tempfile
+import textwrap
+import warnings
+import zipfile
+import zipimport
+from collections.abc import Iterable
+from glob import glob
+from sysconfig import get_path
+from typing import TYPE_CHECKING, NoReturn, TypedDict
+
+from jaraco.text import yield_lines
+
+import pkg_resources
+from pkg_resources import (
+    DEVELOP_DIST,
+    Distribution,
+    DistributionNotFound,
+    EggMetadata,
+    Environment,
+    PathMetadata,
+    Requirement,
+    VersionConflict,
+    WorkingSet,
+    find_distributions,
+    get_distribution,
+    normalize_path,
+    resource_string,
+)
+from setuptools import Command
+from setuptools.archive_util import unpack_archive
+from setuptools.command import bdist_egg, egg_info, setopt
+from setuptools.package_index import URL_SCHEME, PackageIndex, parse_requirement_arg
+from setuptools.sandbox import run_setup
+from setuptools.warnings import SetuptoolsDeprecationWarning, SetuptoolsWarning
+from setuptools.wheel import Wheel
+
+from .._path import ensure_directory
+from .._shutil import attempt_chmod_verbose as chmod, rmtree as _rmtree
+from ..compat import py39, py312
+
+from distutils import dir_util, log
+from distutils.command import install
+from distutils.command.build_scripts import first_line_re
+from distutils.errors import (
+    DistutilsArgError,
+    DistutilsError,
+    DistutilsOptionError,
+    DistutilsPlatformError,
+)
+from distutils.util import convert_path, get_platform, subst_vars
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+# Turn on PEP440Warnings
+warnings.filterwarnings("default", category=pkg_resources.PEP440Warning)
+
+__all__ = [
+    'easy_install',
+    'PthDistributions',
+    'extract_wininst_cfg',
+    'get_exe_prefixes',
+]
+
+
+def is_64bit():
+    return struct.calcsize("P") == 8
+
+
+def _to_bytes(s):
+    return s.encode('utf8')
+
+
+def isascii(s):
+    try:
+        s.encode('ascii')
+    except UnicodeError:
+        return False
+    return True
+
+
+def _one_liner(text):
+    return textwrap.dedent(text).strip().replace('\n', '; ')
+
+
+class easy_install(Command):
+    """Manage a download/build/install process"""
+
+    description = "Find/get/install Python packages"
+    command_consumes_arguments = True
+
+    user_options = [
+        ('prefix=', None, "installation prefix"),
+        ("zip-ok", "z", "install package as a zipfile"),
+        ("multi-version", "m", "make apps have to require() a version"),
+        ("upgrade", "U", "force upgrade (searches PyPI for latest versions)"),
+        ("install-dir=", "d", "install package to DIR"),
+        ("script-dir=", "s", "install scripts to DIR"),
+        ("exclude-scripts", "x", "Don't install scripts"),
+        ("always-copy", "a", "Copy all needed packages to install dir"),
+        ("index-url=", "i", "base URL of Python Package Index"),
+        ("find-links=", "f", "additional URL(s) to search for packages"),
+        ("build-directory=", "b", "download/extract/build in DIR; keep the results"),
+        (
+            'optimize=',
+            'O',
+            'also compile with optimization: -O1 for "python -O", '
+            '-O2 for "python -OO", and -O0 to disable [default: -O0]',
+        ),
+        ('record=', None, "filename in which to record list of installed files"),
+        ('always-unzip', 'Z', "don't install as a zipfile, no matter what"),
+        ('site-dirs=', 'S', "list of directories where .pth files work"),
+        ('editable', 'e', "Install specified packages in editable form"),
+        ('no-deps', 'N', "don't install dependencies"),
+        ('allow-hosts=', 'H', "pattern(s) that hostnames must match"),
+        ('local-snapshots-ok', 'l', "allow building eggs from local checkouts"),
+        ('version', None, "print version information and exit"),
+        (
+            'no-find-links',
+            None,
+            "Don't load find-links defined in packages being installed",
+        ),
+        ('user', None, f"install in user site-package '{site.USER_SITE}'"),
+    ]
+    boolean_options = [
+        'zip-ok',
+        'multi-version',
+        'exclude-scripts',
+        'upgrade',
+        'always-copy',
+        'editable',
+        'no-deps',
+        'local-snapshots-ok',
+        'version',
+        'user',
+    ]
+
+    negative_opt = {'always-unzip': 'zip-ok'}
+    create_index = PackageIndex
+
+    def initialize_options(self):
+        EasyInstallDeprecationWarning.emit()
+
+        # the --user option seems to be an opt-in one,
+        # so the default should be False.
+        self.user = False
+        self.zip_ok = self.local_snapshots_ok = None
+        self.install_dir = self.script_dir = self.exclude_scripts = None
+        self.index_url = None
+        self.find_links = None
+        self.build_directory = None
+        self.args = None
+        self.optimize = self.record = None
+        self.upgrade = self.always_copy = self.multi_version = None
+        self.editable = self.no_deps = self.allow_hosts = None
+        self.root = self.prefix = self.no_report = None
+        self.version = None
+        self.install_purelib = None  # for pure module distributions
+        self.install_platlib = None  # non-pure (dists w/ extensions)
+        self.install_headers = None  # for C/C++ headers
+        self.install_lib = None  # set to either purelib or platlib
+        self.install_scripts = None
+        self.install_data = None
+        self.install_base = None
+        self.install_platbase = None
+        self.install_userbase = site.USER_BASE
+        self.install_usersite = site.USER_SITE
+        self.no_find_links = None
+
+        # Options not specifiable via command line
+        self.package_index = None
+        self.pth_file = self.always_copy_from = None
+        self.site_dirs = None
+        self.installed_projects = {}
+        # Always read easy_install options, even if we are subclassed, or have
+        # an independent instance created.  This ensures that defaults will
+        # always come from the standard configuration file(s)' "easy_install"
+        # section, even if this is a "develop" or "install" command, or some
+        # other embedding.
+        self._dry_run = None
+        self.verbose = self.distribution.verbose
+        self.distribution._set_command_options(
+            self, self.distribution.get_option_dict('easy_install')
+        )
+
+    def delete_blockers(self, blockers) -> None:
+        extant_blockers = (
+            filename
+            for filename in blockers
+            if os.path.exists(filename) or os.path.islink(filename)
+        )
+        list(map(self._delete_path, extant_blockers))
+
+    def _delete_path(self, path):
+        log.info("Deleting %s", path)
+        if self.dry_run:
+            return
+
+        is_tree = os.path.isdir(path) and not os.path.islink(path)
+        remover = _rmtree if is_tree else os.unlink
+        remover(path)
+
+    @staticmethod
+    def _render_version():
+        """
+        Render the Setuptools version and installation details, then exit.
+        """
+        ver = f'{sys.version_info.major}.{sys.version_info.minor}'
+        dist = get_distribution('setuptools')
+        print(f'setuptools {dist.version} from {dist.location} (Python {ver})')
+        raise SystemExit
+
+    def finalize_options(self) -> None:  # noqa: C901  # is too complex (25)  # FIXME
+        self.version and self._render_version()
+
+        py_version = sys.version.split()[0]
+
+        self.config_vars = dict(sysconfig.get_config_vars())
+
+        self.config_vars.update({
+            'dist_name': self.distribution.get_name(),
+            'dist_version': self.distribution.get_version(),
+            'dist_fullname': self.distribution.get_fullname(),
+            'py_version': py_version,
+            'py_version_short': f'{sys.version_info.major}.{sys.version_info.minor}',
+            'py_version_nodot': f'{sys.version_info.major}{sys.version_info.minor}',
+            'sys_prefix': self.config_vars['prefix'],
+            'sys_exec_prefix': self.config_vars['exec_prefix'],
+            # Only POSIX systems have abiflags
+            'abiflags': getattr(sys, 'abiflags', ''),
+            # Only python 3.9+ has platlibdir
+            'platlibdir': getattr(sys, 'platlibdir', 'lib'),
+        })
+        with contextlib.suppress(AttributeError):
+            # only for distutils outside stdlib
+            self.config_vars.update({
+                'implementation_lower': install._get_implementation().lower(),
+                'implementation': install._get_implementation(),
+            })
+
+        # pypa/distutils#113 Python 3.9 compat
+        self.config_vars.setdefault(
+            'py_version_nodot_plat',
+            getattr(sys, 'windir', '').replace('.', ''),
+        )
+
+        self.config_vars['userbase'] = self.install_userbase
+        self.config_vars['usersite'] = self.install_usersite
+        if self.user and not site.ENABLE_USER_SITE:
+            log.warn("WARNING: The user site-packages directory is disabled.")
+
+        self._fix_install_dir_for_user_site()
+
+        self.expand_basedirs()
+        self.expand_dirs()
+
+        self._expand(
+            'install_dir',
+            'script_dir',
+            'build_directory',
+            'site_dirs',
+        )
+        # If a non-default installation directory was specified, default the
+        # script directory to match it.
+        if self.script_dir is None:
+            self.script_dir = self.install_dir
+
+        if self.no_find_links is None:
+            self.no_find_links = False
+
+        # Let install_dir get set by install_lib command, which in turn
+        # gets its info from the install command, and takes into account
+        # --prefix and --home and all that other crud.
+        self.set_undefined_options('install_lib', ('install_dir', 'install_dir'))
+        # Likewise, set default script_dir from 'install_scripts.install_dir'
+        self.set_undefined_options('install_scripts', ('install_dir', 'script_dir'))
+
+        if self.user and self.install_purelib:
+            self.install_dir = self.install_purelib
+            self.script_dir = self.install_scripts
+        # default --record from the install command
+        self.set_undefined_options('install', ('record', 'record'))
+        self.all_site_dirs = get_site_dirs()
+        self.all_site_dirs.extend(self._process_site_dirs(self.site_dirs))
+
+        if not self.editable:
+            self.check_site_dir()
+        default_index = os.getenv("__EASYINSTALL_INDEX", "https://pypi.org/simple/")
+        # ^ Private API for testing purposes only
+        self.index_url = self.index_url or default_index
+        self.shadow_path = self.all_site_dirs[:]
+        for path_item in self.install_dir, normalize_path(self.script_dir):
+            if path_item not in self.shadow_path:
+                self.shadow_path.insert(0, path_item)
+
+        if self.allow_hosts is not None:
+            hosts = [s.strip() for s in self.allow_hosts.split(',')]
+        else:
+            hosts = ['*']
+        if self.package_index is None:
+            self.package_index = self.create_index(
+                self.index_url,
+                search_path=self.shadow_path,
+                hosts=hosts,
+            )
+        self.local_index = Environment(self.shadow_path + sys.path)
+
+        if self.find_links is not None:
+            if isinstance(self.find_links, str):
+                self.find_links = self.find_links.split()
+        else:
+            self.find_links = []
+        if self.local_snapshots_ok:
+            self.package_index.scan_egg_links(self.shadow_path + sys.path)
+        if not self.no_find_links:
+            self.package_index.add_find_links(self.find_links)
+        self.set_undefined_options('install_lib', ('optimize', 'optimize'))
+        self.optimize = self._validate_optimize(self.optimize)
+
+        if self.editable and not self.build_directory:
+            raise DistutilsArgError(
+                "Must specify a build directory (-b) when using --editable"
+            )
+        if not self.args:
+            raise DistutilsArgError(
+                "No urls, filenames, or requirements specified (see --help)"
+            )
+
+        self.outputs: list[str] = []
+
+    @staticmethod
+    def _process_site_dirs(site_dirs):
+        if site_dirs is None:
+            return
+
+        normpath = map(normalize_path, sys.path)
+        site_dirs = [os.path.expanduser(s.strip()) for s in site_dirs.split(',')]
+        for d in site_dirs:
+            if not os.path.isdir(d):
+                log.warn("%s (in --site-dirs) does not exist", d)
+            elif normalize_path(d) not in normpath:
+                raise DistutilsOptionError(d + " (in --site-dirs) is not on sys.path")
+            else:
+                yield normalize_path(d)
+
+    @staticmethod
+    def _validate_optimize(value):
+        try:
+            value = int(value)
+            if value not in range(3):
+                raise ValueError
+        except ValueError as e:
+            raise DistutilsOptionError("--optimize must be 0, 1, or 2") from e
+
+        return value
+
+    def _fix_install_dir_for_user_site(self):
+        """
+        Fix the install_dir if "--user" was used.
+        """
+        if not self.user:
+            return
+
+        self.create_home_path()
+        if self.install_userbase is None:
+            msg = "User base directory is not specified"
+            raise DistutilsPlatformError(msg)
+        self.install_base = self.install_platbase = self.install_userbase
+        scheme_name = f'{os.name}_user'
+        self.select_scheme(scheme_name)
+
+    def _expand_attrs(self, attrs):
+        for attr in attrs:
+            val = getattr(self, attr)
+            if val is not None:
+                if os.name == 'posix' or os.name == 'nt':
+                    val = os.path.expanduser(val)
+                val = subst_vars(val, self.config_vars)
+                setattr(self, attr, val)
+
+    def expand_basedirs(self) -> None:
+        """Calls `os.path.expanduser` on install_base, install_platbase and
+        root."""
+        self._expand_attrs(['install_base', 'install_platbase', 'root'])
+
+    def expand_dirs(self) -> None:
+        """Calls `os.path.expanduser` on install dirs."""
+        dirs = [
+            'install_purelib',
+            'install_platlib',
+            'install_lib',
+            'install_headers',
+            'install_scripts',
+            'install_data',
+        ]
+        self._expand_attrs(dirs)
+
+    def run(self, show_deprecation: bool = True) -> None:
+        if show_deprecation:
+            self.announce(
+                "WARNING: The easy_install command is deprecated "
+                "and will be removed in a future version.",
+                log.WARN,
+            )
+        if self.verbose != self.distribution.verbose:
+            log.set_verbosity(self.verbose)
+        try:
+            for spec in self.args:
+                self.easy_install(spec, not self.no_deps)
+            if self.record:
+                outputs = self.outputs
+                if self.root:  # strip any package prefix
+                    root_len = len(self.root)
+                    for counter in range(len(outputs)):
+                        outputs[counter] = outputs[counter][root_len:]
+                from distutils import file_util
+
+                self.execute(
+                    file_util.write_file,
+                    (self.record, outputs),
+                    f"writing list of installed files to '{self.record}'",
+                )
+            self.warn_deprecated_options()
+        finally:
+            log.set_verbosity(self.distribution.verbose)
+
+    def pseudo_tempname(self):
+        """Return a pseudo-tempname base in the install directory.
+        This code is intentionally naive; if a malicious party can write to
+        the target directory you're already in deep doodoo.
+        """
+        try:
+            pid = os.getpid()
+        except Exception:
+            pid = random.randint(0, sys.maxsize)
+        return os.path.join(self.install_dir, f"test-easy-install-{pid}")
+
+    def warn_deprecated_options(self) -> None:
+        pass
+
+    def check_site_dir(self) -> None:  # is too complex (12)  # FIXME
+        """Verify that self.install_dir is .pth-capable dir, if needed"""
+
+        instdir = normalize_path(self.install_dir)
+        pth_file = os.path.join(instdir, 'easy-install.pth')
+
+        if not os.path.exists(instdir):
+            try:
+                os.makedirs(instdir)
+            except OSError:
+                self.cant_write_to_target()
+
+        # Is it a configured, PYTHONPATH, implicit, or explicit site dir?
+        is_site_dir = instdir in self.all_site_dirs
+
+        if not is_site_dir and not self.multi_version:
+            # No?  Then directly test whether it does .pth file processing
+            is_site_dir = self.check_pth_processing()
+        else:
+            # make sure we can write to target dir
+            testfile = self.pseudo_tempname() + '.write-test'
+            test_exists = os.path.exists(testfile)
+            try:
+                if test_exists:
+                    os.unlink(testfile)
+                open(testfile, 'wb').close()
+                os.unlink(testfile)
+            except OSError:
+                self.cant_write_to_target()
+
+        if not is_site_dir and not self.multi_version:
+            # Can't install non-multi to non-site dir with easy_install
+            pythonpath = os.environ.get('PYTHONPATH', '')
+            log.warn(self.__no_default_msg, self.install_dir, pythonpath)
+
+        if is_site_dir:
+            if self.pth_file is None:
+                self.pth_file = PthDistributions(pth_file, self.all_site_dirs)
+        else:
+            self.pth_file = None
+
+        if self.multi_version and not os.path.exists(pth_file):
+            self.pth_file = None  # don't create a .pth file
+        self.install_dir = instdir
+
+    __cant_write_msg = textwrap.dedent(
+        """
+        can't create or remove files in install directory
+
+        The following error occurred while trying to add or remove files in the
+        installation directory:
+
+            %s
+
+        The installation directory you specified (via --install-dir, --prefix, or
+        the distutils default setting) was:
+
+            %s
+        """
+    ).lstrip()
+
+    __not_exists_id = textwrap.dedent(
+        """
+        This directory does not currently exist.  Please create it and try again, or
+        choose a different installation directory (using the -d or --install-dir
+        option).
+        """
+    ).lstrip()
+
+    __access_msg = textwrap.dedent(
+        """
+        Perhaps your account does not have write access to this directory?  If the
+        installation directory is a system-owned directory, you may need to sign in
+        as the administrator or "root" account.  If you do not have administrative
+        access to this machine, you may wish to choose a different installation
+        directory, preferably one that is listed in your PYTHONPATH environment
+        variable.
+
+        For information on other options, you may wish to consult the
+        documentation at:
+
+          https://setuptools.pypa.io/en/latest/deprecated/easy_install.html
+
+        Please make the appropriate changes for your system and try again.
+        """
+    ).lstrip()
+
+    def cant_write_to_target(self) -> NoReturn:
+        msg = self.__cant_write_msg % (
+            sys.exc_info()[1],
+            self.install_dir,
+        )
+
+        if not os.path.exists(self.install_dir):
+            msg += '\n' + self.__not_exists_id
+        else:
+            msg += '\n' + self.__access_msg
+        raise DistutilsError(msg)
+
+    def check_pth_processing(self):  # noqa: C901
+        """Empirically verify whether .pth files are supported in inst. dir"""
+        instdir = self.install_dir
+        log.info("Checking .pth file support in %s", instdir)
+        pth_file = self.pseudo_tempname() + ".pth"
+        ok_file = pth_file + '.ok'
+        ok_exists = os.path.exists(ok_file)
+        tmpl = (
+            _one_liner(
+                """
+            import os
+            f = open({ok_file!r}, 'w', encoding="utf-8")
+            f.write('OK')
+            f.close()
+            """
+            )
+            + '\n'
+        )
+        try:
+            if ok_exists:
+                os.unlink(ok_file)
+            dirname = os.path.dirname(ok_file)
+            os.makedirs(dirname, exist_ok=True)
+            f = open(pth_file, 'w', encoding=py312.PTH_ENCODING)
+            # ^-- Python<3.13 require encoding="locale" instead of "utf-8",
+            #     see python/cpython#77102.
+        except OSError:
+            self.cant_write_to_target()
+        else:
+            try:
+                f.write(tmpl.format(**locals()))
+                f.close()
+                f = None
+                executable = sys.executable
+                if os.name == 'nt':
+                    dirname, basename = os.path.split(executable)
+                    alt = os.path.join(dirname, 'pythonw.exe')
+                    use_alt = basename.lower() == 'python.exe' and os.path.exists(alt)
+                    if use_alt:
+                        # use pythonw.exe to avoid opening a console window
+                        executable = alt
+
+                from distutils.spawn import spawn
+
+                spawn([executable, '-E', '-c', 'pass'], 0)
+
+                if os.path.exists(ok_file):
+                    log.info("TEST PASSED: %s appears to support .pth files", instdir)
+                    return True
+            finally:
+                if f:
+                    f.close()
+                if os.path.exists(ok_file):
+                    os.unlink(ok_file)
+                if os.path.exists(pth_file):
+                    os.unlink(pth_file)
+        if not self.multi_version:
+            log.warn("TEST FAILED: %s does NOT support .pth files", instdir)
+        return False
+
+    def install_egg_scripts(self, dist) -> None:
+        """Write all the scripts for `dist`, unless scripts are excluded"""
+        if not self.exclude_scripts and dist.metadata_isdir('scripts'):
+            for script_name in dist.metadata_listdir('scripts'):
+                if dist.metadata_isdir('scripts/' + script_name):
+                    # The "script" is a directory, likely a Python 3
+                    # __pycache__ directory, so skip it.
+                    continue
+                self.install_script(
+                    dist, script_name, dist.get_metadata('scripts/' + script_name)
+                )
+        self.install_wrapper_scripts(dist)
+
+    def add_output(self, path) -> None:
+        if os.path.isdir(path):
+            for base, dirs, files in os.walk(path):
+                for filename in files:
+                    self.outputs.append(os.path.join(base, filename))
+        else:
+            self.outputs.append(path)
+
+    def not_editable(self, spec) -> None:
+        if self.editable:
+            raise DistutilsArgError(
+                f"Invalid argument {spec!r}: you can't use filenames or URLs "
+                "with --editable (except via the --find-links option)."
+            )
+
+    def check_editable(self, spec) -> None:
+        if not self.editable:
+            return
+
+        if os.path.exists(os.path.join(self.build_directory, spec.key)):
+            raise DistutilsArgError(
+                f"{spec.key!r} already exists in {self.build_directory}; can't do a checkout there"
+            )
+
+    @contextlib.contextmanager
+    def _tmpdir(self):
+        tmpdir = tempfile.mkdtemp(prefix="easy_install-")
+        try:
+            # cast to str as workaround for #709 and #710 and #712
+            yield str(tmpdir)
+        finally:
+            os.path.exists(tmpdir) and _rmtree(tmpdir)
+
+    def easy_install(self, spec, deps: bool = False) -> Distribution | None:
+        with self._tmpdir() as tmpdir:
+            if not isinstance(spec, Requirement):
+                if URL_SCHEME(spec):
+                    # It's a url, download it to tmpdir and process
+                    self.not_editable(spec)
+                    dl = self.package_index.download(spec, tmpdir)
+                    return self.install_item(None, dl, tmpdir, deps, True)
+
+                elif os.path.exists(spec):
+                    # Existing file or directory, just process it directly
+                    self.not_editable(spec)
+                    return self.install_item(None, spec, tmpdir, deps, True)
+                else:
+                    spec = parse_requirement_arg(spec)
+
+            self.check_editable(spec)
+            dist = self.package_index.fetch_distribution(
+                spec,
+                tmpdir,
+                self.upgrade,
+                self.editable,
+                not self.always_copy,
+                self.local_index,
+            )
+            if dist is None:
+                msg = f"Could not find suitable distribution for {spec!r}"
+                if self.always_copy:
+                    msg += " (--always-copy skips system and development eggs)"
+                raise DistutilsError(msg)
+            elif dist.precedence == DEVELOP_DIST:
+                # .egg-info dists don't need installing, just process deps
+                self.process_distribution(spec, dist, deps, "Using")
+                return dist
+            else:
+                return self.install_item(spec, dist.location, tmpdir, deps)
+
+    def install_item(
+        self, spec, download, tmpdir, deps, install_needed: bool = False
+    ) -> Distribution | None:
+        # Installation is also needed if file in tmpdir or is not an egg
+        install_needed = install_needed or bool(self.always_copy)
+        install_needed = install_needed or os.path.dirname(download) == tmpdir
+        install_needed = install_needed or not download.endswith('.egg')
+        install_needed = install_needed or (
+            self.always_copy_from is not None
+            and os.path.dirname(normalize_path(download))
+            == normalize_path(self.always_copy_from)
+        )
+
+        if spec and not install_needed:
+            # at this point, we know it's a local .egg, we just don't know if
+            # it's already installed.
+            for dist in self.local_index[spec.project_name]:
+                if dist.location == download:
+                    break
+            else:
+                install_needed = True  # it's not in the local index
+
+        log.info("Processing %s", os.path.basename(download))
+
+        if install_needed:
+            dists = self.install_eggs(spec, download, tmpdir)
+            for dist in dists:
+                self.process_distribution(spec, dist, deps)
+        else:
+            dists = [self.egg_distribution(download)]
+            self.process_distribution(spec, dists[0], deps, "Using")
+
+        if spec is not None:
+            for dist in dists:
+                if dist in spec:
+                    return dist
+        return None
+
+    def select_scheme(self, name):
+        try:
+            install._select_scheme(self, name)
+        except AttributeError:
+            # stdlib distutils
+            install.install.select_scheme(self, name.replace('posix', 'unix'))
+
+    # FIXME: 'easy_install.process_distribution' is too complex (12)
+    def process_distribution(  # noqa: C901
+        self,
+        requirement,
+        dist,
+        deps: bool = True,
+        *info,
+    ) -> None:
+        self.update_pth(dist)
+        self.package_index.add(dist)
+        if dist in self.local_index[dist.key]:
+            self.local_index.remove(dist)
+        self.local_index.add(dist)
+        self.install_egg_scripts(dist)
+        self.installed_projects[dist.key] = dist
+        log.info(self.installation_report(requirement, dist, *info))
+        if dist.has_metadata('dependency_links.txt') and not self.no_find_links:
+            self.package_index.add_find_links(
+                dist.get_metadata_lines('dependency_links.txt')
+            )
+        if not deps and not self.always_copy:
+            return
+        elif requirement is not None and dist.key != requirement.key:
+            log.warn("Skipping dependencies for %s", dist)
+            return  # XXX this is not the distribution we were looking for
+        elif requirement is None or dist not in requirement:
+            # if we wound up with a different version, resolve what we've got
+            distreq = dist.as_requirement()
+            requirement = Requirement(str(distreq))
+        log.info("Processing dependencies for %s", requirement)
+        try:
+            distros = WorkingSet([]).resolve(
+                [requirement], self.local_index, self.easy_install
+            )
+        except DistributionNotFound as e:
+            raise DistutilsError(str(e)) from e
+        except VersionConflict as e:
+            raise DistutilsError(e.report()) from e
+        if self.always_copy or self.always_copy_from:
+            # Force all the relevant distros to be copied or activated
+            for dist in distros:
+                if dist.key not in self.installed_projects:
+                    self.easy_install(dist.as_requirement())
+        log.info("Finished processing dependencies for %s", requirement)
+
+    def should_unzip(self, dist) -> bool:
+        if self.zip_ok is not None:
+            return not self.zip_ok
+        if dist.has_metadata('not-zip-safe'):
+            return True
+        if not dist.has_metadata('zip-safe'):
+            return True
+        return False
+
+    def maybe_move(self, spec, dist_filename, setup_base):
+        dst = os.path.join(self.build_directory, spec.key)
+        if os.path.exists(dst):
+            msg = "%r already exists in %s; build directory %s will not be kept"
+            log.warn(msg, spec.key, self.build_directory, setup_base)
+            return setup_base
+        if os.path.isdir(dist_filename):
+            setup_base = dist_filename
+        else:
+            if os.path.dirname(dist_filename) == setup_base:
+                os.unlink(dist_filename)  # get it out of the tmp dir
+            contents = os.listdir(setup_base)
+            if len(contents) == 1:
+                dist_filename = os.path.join(setup_base, contents[0])
+                if os.path.isdir(dist_filename):
+                    # if the only thing there is a directory, move it instead
+                    setup_base = dist_filename
+        ensure_directory(dst)
+        shutil.move(setup_base, dst)
+        return dst
+
+    def install_wrapper_scripts(self, dist) -> None:
+        if self.exclude_scripts:
+            return
+        for args in ScriptWriter.best().get_args(dist):
+            self.write_script(*args)
+
+    def install_script(self, dist, script_name, script_text, dev_path=None) -> None:
+        """Generate a legacy script wrapper and install it"""
+        spec = str(dist.as_requirement())
+        is_script = is_python_script(script_text, script_name)
+
+        if is_script:
+            body = self._load_template(dev_path) % locals()
+            script_text = ScriptWriter.get_header(script_text) + body
+        self.write_script(script_name, _to_bytes(script_text), 'b')
+
+    @staticmethod
+    def _load_template(dev_path):
+        """
+        There are a couple of template scripts in the package. This
+        function loads one of them and prepares it for use.
+        """
+        # See https://github.com/pypa/setuptools/issues/134 for info
+        # on script file naming and downstream issues with SVR4
+        name = 'script.tmpl'
+        if dev_path:
+            name = name.replace('.tmpl', ' (dev).tmpl')
+
+        raw_bytes = resource_string('setuptools', name)
+        return raw_bytes.decode('utf-8')
+
+    def write_script(self, script_name, contents, mode: str = "t", blockers=()) -> None:
+        """Write an executable file to the scripts directory"""
+        self.delete_blockers(  # clean up old .py/.pyw w/o a script
+            [os.path.join(self.script_dir, x) for x in blockers]
+        )
+        log.info("Installing %s script to %s", script_name, self.script_dir)
+        target = os.path.join(self.script_dir, script_name)
+        self.add_output(target)
+
+        if self.dry_run:
+            return
+
+        mask = current_umask()
+        ensure_directory(target)
+        if os.path.exists(target):
+            os.unlink(target)
+
+        encoding = None if "b" in mode else "utf-8"
+        with open(target, "w" + mode, encoding=encoding) as f:
+            f.write(contents)
+        chmod(target, 0o777 - mask)
+
+    def install_eggs(self, spec, dist_filename, tmpdir) -> list[Distribution]:
+        # .egg dirs or files are already built, so just return them
+        installer_map = {
+            '.egg': self.install_egg,
+            '.exe': self.install_exe,
+            '.whl': self.install_wheel,
+        }
+        try:
+            install_dist = installer_map[dist_filename.lower()[-4:]]
+        except KeyError:
+            pass
+        else:
+            return [install_dist(dist_filename, tmpdir)]
+
+        # Anything else, try to extract and build
+        setup_base = tmpdir
+        if os.path.isfile(dist_filename) and not dist_filename.endswith('.py'):
+            unpack_archive(dist_filename, tmpdir, self.unpack_progress)
+        elif os.path.isdir(dist_filename):
+            setup_base = os.path.abspath(dist_filename)
+
+        if (
+            setup_base.startswith(tmpdir)  # something we downloaded
+            and self.build_directory
+            and spec is not None
+        ):
+            setup_base = self.maybe_move(spec, dist_filename, setup_base)
+
+        # Find the setup.py file
+        setup_script = os.path.join(setup_base, 'setup.py')
+
+        if not os.path.exists(setup_script):
+            setups = glob(os.path.join(setup_base, '*', 'setup.py'))
+            if not setups:
+                raise DistutilsError(
+                    f"Couldn't find a setup script in {os.path.abspath(dist_filename)}"
+                )
+            if len(setups) > 1:
+                raise DistutilsError(
+                    f"Multiple setup scripts in {os.path.abspath(dist_filename)}"
+                )
+            setup_script = setups[0]
+
+        # Now run it, and return the result
+        if self.editable:
+            log.info(self.report_editable(spec, setup_script))
+            return []
+        else:
+            return self.build_and_install(setup_script, setup_base)
+
+    def egg_distribution(self, egg_path):
+        if os.path.isdir(egg_path):
+            metadata = PathMetadata(egg_path, os.path.join(egg_path, 'EGG-INFO'))
+        else:
+            metadata = EggMetadata(zipimport.zipimporter(egg_path))
+        return Distribution.from_filename(egg_path, metadata=metadata)
+
+    # FIXME: 'easy_install.install_egg' is too complex (11)
+    def install_egg(self, egg_path, tmpdir):
+        destination = os.path.join(
+            self.install_dir,
+            os.path.basename(egg_path),
+        )
+        destination = os.path.abspath(destination)
+        if not self.dry_run:
+            ensure_directory(destination)
+
+        dist = self.egg_distribution(egg_path)
+        if not (
+            os.path.exists(destination) and os.path.samefile(egg_path, destination)
+        ):
+            if os.path.isdir(destination) and not os.path.islink(destination):
+                dir_util.remove_tree(destination, dry_run=self.dry_run)
+            elif os.path.exists(destination):
+                self.execute(
+                    os.unlink,
+                    (destination,),
+                    "Removing " + destination,
+                )
+            try:
+                new_dist_is_zipped = False
+                if os.path.isdir(egg_path):
+                    if egg_path.startswith(tmpdir):
+                        f, m = shutil.move, "Moving"
+                    else:
+                        f, m = shutil.copytree, "Copying"
+                elif self.should_unzip(dist):
+                    self.mkpath(destination)
+                    f, m = self.unpack_and_compile, "Extracting"
+                else:
+                    new_dist_is_zipped = True
+                    if egg_path.startswith(tmpdir):
+                        f, m = shutil.move, "Moving"
+                    else:
+                        f, m = shutil.copy2, "Copying"
+                self.execute(
+                    f,
+                    (egg_path, destination),
+                    (m + " %s to %s")
+                    % (os.path.basename(egg_path), os.path.dirname(destination)),
+                )
+                update_dist_caches(
+                    destination,
+                    fix_zipimporter_caches=new_dist_is_zipped,
+                )
+            except Exception:
+                update_dist_caches(destination, fix_zipimporter_caches=False)
+                raise
+
+        self.add_output(destination)
+        return self.egg_distribution(destination)
+
+    def install_exe(self, dist_filename, tmpdir):
+        # See if it's valid, get data
+        cfg = extract_wininst_cfg(dist_filename)
+        if cfg is None:
+            raise DistutilsError(
+                f"{dist_filename} is not a valid distutils Windows .exe"
+            )
+        # Create a dummy distribution object until we build the real distro
+        dist = Distribution(
+            None,
+            project_name=cfg.get('metadata', 'name'),
+            version=cfg.get('metadata', 'version'),
+            platform=get_platform(),
+        )
+
+        # Convert the .exe to an unpacked egg
+        egg_path = os.path.join(tmpdir, dist.egg_name() + '.egg')
+        dist.location = egg_path
+        egg_tmp = egg_path + '.tmp'
+        _egg_info = os.path.join(egg_tmp, 'EGG-INFO')
+        pkg_inf = os.path.join(_egg_info, 'PKG-INFO')
+        ensure_directory(pkg_inf)  # make sure EGG-INFO dir exists
+        dist._provider = PathMetadata(egg_tmp, _egg_info)  # XXX
+        self.exe_to_egg(dist_filename, egg_tmp)
+
+        # Write EGG-INFO/PKG-INFO
+        if not os.path.exists(pkg_inf):
+            with open(pkg_inf, 'w', encoding="utf-8") as f:
+                f.write('Metadata-Version: 1.0\n')
+                for k, v in cfg.items('metadata'):
+                    if k != 'target_version':
+                        k = k.replace('_', '-').title()
+                        f.write(f'{k}: {v}\n')
+        script_dir = os.path.join(_egg_info, 'scripts')
+        # delete entry-point scripts to avoid duping
+        self.delete_blockers([
+            os.path.join(script_dir, args[0]) for args in ScriptWriter.get_args(dist)
+        ])
+        # Build .egg file from tmpdir
+        bdist_egg.make_zipfile(
+            egg_path,
+            egg_tmp,
+            verbose=self.verbose,
+            dry_run=self.dry_run,
+        )
+        # install the .egg
+        return self.install_egg(egg_path, tmpdir)
+
+    # FIXME: 'easy_install.exe_to_egg' is too complex (12)
+    def exe_to_egg(self, dist_filename, egg_tmp) -> None:  # noqa: C901
+        """Extract a bdist_wininst to the directories an egg would use"""
+        # Check for .pth file and set up prefix translations
+        prefixes = get_exe_prefixes(dist_filename)
+        to_compile = []
+        native_libs = []
+        top_level = set()
+
+        def process(src, dst):
+            s = src.lower()
+            for old, new in prefixes:
+                if s.startswith(old):
+                    src = new + src[len(old) :]
+                    parts = src.split('/')
+                    dst = os.path.join(egg_tmp, *parts)
+                    dl = dst.lower()
+                    if dl.endswith('.pyd') or dl.endswith('.dll'):
+                        parts[-1] = bdist_egg.strip_module(parts[-1])
+                        top_level.add([os.path.splitext(parts[0])[0]])
+                        native_libs.append(src)
+                    elif dl.endswith('.py') and old != 'SCRIPTS/':
+                        top_level.add([os.path.splitext(parts[0])[0]])
+                        to_compile.append(dst)
+                    return dst
+            if not src.endswith('.pth'):
+                log.warn("WARNING: can't process %s", src)
+            return None
+
+        # extract, tracking .pyd/.dll->native_libs and .py -> to_compile
+        unpack_archive(dist_filename, egg_tmp, process)
+        stubs = []
+        for res in native_libs:
+            if res.lower().endswith('.pyd'):  # create stubs for .pyd's
+                parts = res.split('/')
+                resource = parts[-1]
+                parts[-1] = bdist_egg.strip_module(parts[-1]) + '.py'
+                pyfile = os.path.join(egg_tmp, *parts)
+                to_compile.append(pyfile)
+                stubs.append(pyfile)
+                bdist_egg.write_stub(resource, pyfile)
+        self.byte_compile(to_compile)  # compile .py's
+        bdist_egg.write_safety_flag(
+            os.path.join(egg_tmp, 'EGG-INFO'), bdist_egg.analyze_egg(egg_tmp, stubs)
+        )  # write zip-safety flag
+
+        for name in 'top_level', 'native_libs':
+            if locals()[name]:
+                txt = os.path.join(egg_tmp, 'EGG-INFO', name + '.txt')
+                if not os.path.exists(txt):
+                    with open(txt, 'w', encoding="utf-8") as f:
+                        f.write('\n'.join(locals()[name]) + '\n')
+
+    def install_wheel(self, wheel_path, tmpdir):
+        wheel = Wheel(wheel_path)
+        assert wheel.is_compatible()
+        destination = os.path.join(self.install_dir, wheel.egg_name())
+        destination = os.path.abspath(destination)
+        if not self.dry_run:
+            ensure_directory(destination)
+        if os.path.isdir(destination) and not os.path.islink(destination):
+            dir_util.remove_tree(destination, dry_run=self.dry_run)
+        elif os.path.exists(destination):
+            self.execute(
+                os.unlink,
+                (destination,),
+                "Removing " + destination,
+            )
+        try:
+            self.execute(
+                wheel.install_as_egg,
+                (destination,),
+                (
+                    f"Installing {os.path.basename(wheel_path)} to {os.path.dirname(destination)}"
+                ),
+            )
+        finally:
+            update_dist_caches(destination, fix_zipimporter_caches=False)
+        self.add_output(destination)
+        return self.egg_distribution(destination)
+
+    __mv_warning = textwrap.dedent(
+        """
+        Because this distribution was installed --multi-version, before you can
+        import modules from this package in an application, you will need to
+        'import pkg_resources' and then use a 'require()' call similar to one of
+        these examples, in order to select the desired version:
+
+            pkg_resources.require("%(name)s")  # latest installed version
+            pkg_resources.require("%(name)s==%(version)s")  # this exact version
+            pkg_resources.require("%(name)s>=%(version)s")  # this version or higher
+        """
+    ).lstrip()
+
+    __id_warning = textwrap.dedent(
+        """
+        Note also that the installation directory must be on sys.path at runtime for
+        this to work.  (e.g. by being the application's script directory, by being on
+        PYTHONPATH, or by being added to sys.path by your code.)
+        """
+    )
+
+    def installation_report(self, req, dist, what: str = "Installed") -> str:
+        """Helpful installation message for display to package users"""
+        msg = "\n%(what)s %(eggloc)s%(extras)s"
+        if self.multi_version and not self.no_report:
+            msg += '\n' + self.__mv_warning
+            if self.install_dir not in map(normalize_path, sys.path):
+                msg += '\n' + self.__id_warning
+
+        eggloc = dist.location
+        name = dist.project_name
+        version = dist.version
+        extras = ''  # TODO: self.report_extras(req, dist)
+        return msg % locals()
+
+    __editable_msg = textwrap.dedent(
+        """
+        Extracted editable version of %(spec)s to %(dirname)s
+
+        If it uses setuptools in its setup script, you can activate it in
+        "development" mode by going to that directory and running::
+
+            %(python)s setup.py develop
+
+        See the setuptools documentation for the "develop" command for more info.
+        """
+    ).lstrip()
+
+    def report_editable(self, spec, setup_script):
+        dirname = os.path.dirname(setup_script)
+        python = sys.executable
+        return '\n' + self.__editable_msg % locals()
+
+    def run_setup(self, setup_script, setup_base, args) -> None:
+        sys.modules.setdefault('distutils.command.bdist_egg', bdist_egg)
+        sys.modules.setdefault('distutils.command.egg_info', egg_info)
+
+        args = list(args)
+        if self.verbose > 2:
+            v = 'v' * (self.verbose - 1)
+            args.insert(0, '-' + v)
+        elif self.verbose < 2:
+            args.insert(0, '-q')
+        if self.dry_run:
+            args.insert(0, '-n')
+        log.info("Running %s %s", setup_script[len(setup_base) + 1 :], ' '.join(args))
+        try:
+            run_setup(setup_script, args)
+        except SystemExit as v:
+            raise DistutilsError(f"Setup script exited with {v.args[0]}") from v
+
+    def build_and_install(self, setup_script, setup_base):
+        args = ['bdist_egg', '--dist-dir']
+
+        dist_dir = tempfile.mkdtemp(
+            prefix='egg-dist-tmp-', dir=os.path.dirname(setup_script)
+        )
+        try:
+            self._set_fetcher_options(os.path.dirname(setup_script))
+            args.append(dist_dir)
+
+            self.run_setup(setup_script, setup_base, args)
+            all_eggs = Environment([dist_dir])
+            eggs = [
+                self.install_egg(dist.location, setup_base)
+                for key in all_eggs
+                for dist in all_eggs[key]
+            ]
+            if not eggs and not self.dry_run:
+                log.warn("No eggs found in %s (setup script problem?)", dist_dir)
+            return eggs
+        finally:
+            _rmtree(dist_dir)
+            log.set_verbosity(self.verbose)  # restore our log verbosity
+
+    def _set_fetcher_options(self, base):
+        """
+        When easy_install is about to run bdist_egg on a source dist, that
+        source dist might have 'setup_requires' directives, requiring
+        additional fetching. Ensure the fetcher options given to easy_install
+        are available to that command as well.
+        """
+        # find the fetch options from easy_install and write them out
+        # to the setup.cfg file.
+        ei_opts = self.distribution.get_option_dict('easy_install').copy()
+        fetch_directives = (
+            'find_links',
+            'site_dirs',
+            'index_url',
+            'optimize',
+            'allow_hosts',
+        )
+        fetch_options = {}
+        for key, val in ei_opts.items():
+            if key not in fetch_directives:
+                continue
+            fetch_options[key] = val[1]
+        # create a settings dictionary suitable for `edit_config`
+        settings = dict(easy_install=fetch_options)
+        cfg_filename = os.path.join(base, 'setup.cfg')
+        setopt.edit_config(cfg_filename, settings)
+
+    def update_pth(self, dist) -> None:  # noqa: C901  # is too complex (11)  # FIXME
+        if self.pth_file is None:
+            return
+
+        for d in self.pth_file[dist.key]:  # drop old entries
+            if not self.multi_version and d.location == dist.location:
+                continue
+
+            log.info("Removing %s from easy-install.pth file", d)
+            self.pth_file.remove(d)
+            if d.location in self.shadow_path:
+                self.shadow_path.remove(d.location)
+
+        if not self.multi_version:
+            if dist.location in self.pth_file.paths:
+                log.info(
+                    "%s is already the active version in easy-install.pth",
+                    dist,
+                )
+            else:
+                log.info("Adding %s to easy-install.pth file", dist)
+                self.pth_file.add(dist)  # add new entry
+                if dist.location not in self.shadow_path:
+                    self.shadow_path.append(dist.location)
+
+        if self.dry_run:
+            return
+
+        self.pth_file.save()
+
+        if dist.key != 'setuptools':
+            return
+
+        # Ensure that setuptools itself never becomes unavailable!
+        # XXX should this check for latest version?
+        filename = os.path.join(self.install_dir, 'setuptools.pth')
+        if os.path.islink(filename):
+            os.unlink(filename)
+
+        with open(filename, 'wt', encoding=py312.PTH_ENCODING) as f:
+            # ^-- Python<3.13 require encoding="locale" instead of "utf-8",
+            #     see python/cpython#77102.
+            f.write(self.pth_file.make_relative(dist.location) + '\n')
+
+    def unpack_progress(self, src, dst):
+        # Progress filter for unpacking
+        log.debug("Unpacking %s to %s", src, dst)
+        return dst  # only unpack-and-compile skips files for dry run
+
+    def unpack_and_compile(self, egg_path, destination) -> None:
+        to_compile = []
+        to_chmod = []
+
+        def pf(src, dst):
+            if dst.endswith('.py') and not src.startswith('EGG-INFO/'):
+                to_compile.append(dst)
+            elif dst.endswith('.dll') or dst.endswith('.so'):
+                to_chmod.append(dst)
+            self.unpack_progress(src, dst)
+            return not self.dry_run and dst or None
+
+        unpack_archive(egg_path, destination, pf)
+        self.byte_compile(to_compile)
+        if not self.dry_run:
+            for f in to_chmod:
+                mode = ((os.stat(f)[stat.ST_MODE]) | 0o555) & 0o7755
+                chmod(f, mode)
+
+    def byte_compile(self, to_compile) -> None:
+        if sys.dont_write_bytecode:
+            return
+
+        from distutils.util import byte_compile
+
+        try:
+            # try to make the byte compile messages quieter
+            log.set_verbosity(self.verbose - 1)
+
+            byte_compile(to_compile, optimize=0, force=True, dry_run=self.dry_run)
+            if self.optimize:
+                byte_compile(
+                    to_compile,
+                    optimize=self.optimize,
+                    force=True,
+                    dry_run=self.dry_run,
+                )
+        finally:
+            log.set_verbosity(self.verbose)  # restore original verbosity
+
+    __no_default_msg = textwrap.dedent(
+        """
+        bad install directory or PYTHONPATH
+
+        You are attempting to install a package to a directory that is not
+        on PYTHONPATH and which Python does not read ".pth" files from.  The
+        installation directory you specified (via --install-dir, --prefix, or
+        the distutils default setting) was:
+
+            %s
+
+        and your PYTHONPATH environment variable currently contains:
+
+            %r
+
+        Here are some of your options for correcting the problem:
+
+        * You can choose a different installation directory, i.e., one that is
+          on PYTHONPATH or supports .pth files
+
+        * You can add the installation directory to the PYTHONPATH environment
+          variable.  (It must then also be on PYTHONPATH whenever you run
+          Python and want to use the package(s) you are installing.)
+
+        * You can set up the installation directory to support ".pth" files by
+          using one of the approaches described here:
+
+          https://setuptools.pypa.io/en/latest/deprecated/easy_install.html#custom-installation-locations
+
+
+        Please make the appropriate changes for your system and try again.
+        """
+    ).strip()
+
+    def create_home_path(self) -> None:
+        """Create directories under ~."""
+        if not self.user:
+            return
+        home = convert_path(os.path.expanduser("~"))
+        for path in only_strs(self.config_vars.values()):
+            if path.startswith(home) and not os.path.isdir(path):
+                self.debug_print(f"os.makedirs('{path}', 0o700)")
+                os.makedirs(path, 0o700)
+
+    INSTALL_SCHEMES = dict(
+        posix=dict(
+            install_dir='$base/lib/python$py_version_short/site-packages',
+            script_dir='$base/bin',
+        ),
+    )
+
+    DEFAULT_SCHEME = dict(
+        install_dir='$base/Lib/site-packages',
+        script_dir='$base/Scripts',
+    )
+
+    def _expand(self, *attrs):
+        config_vars = self.get_finalized_command('install').config_vars
+
+        if self.prefix:
+            # Set default install_dir/scripts from --prefix
+            config_vars = dict(config_vars)
+            config_vars['base'] = self.prefix
+            scheme = self.INSTALL_SCHEMES.get(os.name, self.DEFAULT_SCHEME)
+            for attr, val in scheme.items():
+                if getattr(self, attr, None) is None:
+                    setattr(self, attr, val)
+
+        from distutils.util import subst_vars
+
+        for attr in attrs:
+            val = getattr(self, attr)
+            if val is not None:
+                val = subst_vars(val, config_vars)
+                if os.name == 'posix':
+                    val = os.path.expanduser(val)
+                setattr(self, attr, val)
+
+
+def _pythonpath():
+    items = os.environ.get('PYTHONPATH', '').split(os.pathsep)
+    return filter(None, items)
+
+
+def get_site_dirs():
+    """
+    Return a list of 'site' dirs
+    """
+
+    sitedirs = []
+
+    # start with PYTHONPATH
+    sitedirs.extend(_pythonpath())
+
+    prefixes = [sys.prefix]
+    if sys.exec_prefix != sys.prefix:
+        prefixes.append(sys.exec_prefix)
+    for prefix in prefixes:
+        if not prefix:
+            continue
+
+        if sys.platform in ('os2emx', 'riscos'):
+            sitedirs.append(os.path.join(prefix, "Lib", "site-packages"))
+        elif os.sep == '/':
+            sitedirs.extend([
+                os.path.join(
+                    prefix,
+                    "lib",
+                    f"python{sys.version_info.major}.{sys.version_info.minor}",
+                    "site-packages",
+                ),
+                os.path.join(prefix, "lib", "site-python"),
+            ])
+        else:
+            sitedirs.extend([
+                prefix,
+                os.path.join(prefix, "lib", "site-packages"),
+            ])
+        if sys.platform != 'darwin':
+            continue
+
+        # for framework builds *only* we add the standard Apple
+        # locations. Currently only per-user, but /Library and
+        # /Network/Library could be added too
+        if 'Python.framework' not in prefix:
+            continue
+
+        home = os.environ.get('HOME')
+        if not home:
+            continue
+
+        home_sp = os.path.join(
+            home,
+            'Library',
+            'Python',
+            f'{sys.version_info.major}.{sys.version_info.minor}',
+            'site-packages',
+        )
+        sitedirs.append(home_sp)
+    lib_paths = get_path('purelib'), get_path('platlib')
+
+    sitedirs.extend(s for s in lib_paths if s not in sitedirs)
+
+    if site.ENABLE_USER_SITE:
+        sitedirs.append(site.USER_SITE)
+
+    with contextlib.suppress(AttributeError):
+        sitedirs.extend(site.getsitepackages())
+
+    return list(map(normalize_path, sitedirs))
+
+
+def expand_paths(inputs):  # noqa: C901  # is too complex (11)  # FIXME
+    """Yield sys.path directories that might contain "old-style" packages"""
+
+    seen = set()
+
+    for dirname in inputs:
+        dirname = normalize_path(dirname)
+        if dirname in seen:
+            continue
+
+        seen.add(dirname)
+        if not os.path.isdir(dirname):
+            continue
+
+        files = os.listdir(dirname)
+        yield dirname, files
+
+        for name in files:
+            if not name.endswith('.pth'):
+                # We only care about the .pth files
+                continue
+            if name in ('easy-install.pth', 'setuptools.pth'):
+                # Ignore .pth files that we control
+                continue
+
+            # Read the .pth file
+            content = _read_pth(os.path.join(dirname, name))
+            lines = list(yield_lines(content))
+
+            # Yield existing non-dupe, non-import directory lines from it
+            for line in lines:
+                if line.startswith("import"):
+                    continue
+
+                line = normalize_path(line.rstrip())
+                if line in seen:
+                    continue
+
+                seen.add(line)
+                if not os.path.isdir(line):
+                    continue
+
+                yield line, os.listdir(line)
+
+
+def extract_wininst_cfg(dist_filename):
+    """Extract configuration data from a bdist_wininst .exe
+
+    Returns a configparser.RawConfigParser, or None
+    """
+    f = open(dist_filename, 'rb')
+    try:
+        endrec = zipfile._EndRecData(f)
+        if endrec is None:
+            return None
+
+        prepended = (endrec[9] - endrec[5]) - endrec[6]
+        if prepended < 12:  # no wininst data here
+            return None
+        f.seek(prepended - 12)
+
+        tag, cfglen, _bmlen = struct.unpack("<iii", f.read(12))
+        if tag not in (0x1234567A, 0x1234567B):
+            return None  # not a valid tag
+
+        f.seek(prepended - (12 + cfglen))
+        init = {'version': '', 'target_version': ''}
+        cfg = configparser.RawConfigParser(init)
+        try:
+            part = f.read(cfglen)
+            # Read up to the first null byte.
+            config = part.split(b'\0', 1)[0]
+            # Now the config is in bytes, but for RawConfigParser, it should
+            #  be text, so decode it.
+            config = config.decode(sys.getfilesystemencoding())
+            cfg.read_file(io.StringIO(config))
+        except configparser.Error:
+            return None
+        if not cfg.has_section('metadata') or not cfg.has_section('Setup'):
+            return None
+        return cfg
+
+    finally:
+        f.close()
+
+
+def get_exe_prefixes(exe_filename):
+    """Get exe->egg path translations for a given .exe file"""
+
+    prefixes = [
+        ('PURELIB/', ''),
+        ('PLATLIB/pywin32_system32', ''),
+        ('PLATLIB/', ''),
+        ('SCRIPTS/', 'EGG-INFO/scripts/'),
+        ('DATA/lib/site-packages', ''),
+    ]
+    z = zipfile.ZipFile(exe_filename)
+    try:
+        for info in z.infolist():
+            name = info.filename
+            parts = name.split('/')
+            if len(parts) == 3 and parts[2] == 'PKG-INFO':
+                if parts[1].endswith('.egg-info'):
+                    prefixes.insert(0, ('/'.join(parts[:2]), 'EGG-INFO/'))
+                    break
+            if len(parts) != 2 or not name.endswith('.pth'):
+                continue
+            if name.endswith('-nspkg.pth'):
+                continue
+            if parts[0].upper() in ('PURELIB', 'PLATLIB'):
+                contents = z.read(name).decode()
+                for pth in yield_lines(contents):
+                    pth = pth.strip().replace('\\', '/')
+                    if not pth.startswith('import'):
+                        prefixes.append(((f'{parts[0]}/{pth}/'), ''))
+    finally:
+        z.close()
+    prefixes = [(x.lower(), y) for x, y in prefixes]
+    prefixes.sort()
+    prefixes.reverse()
+    return prefixes
+
+
+class PthDistributions(Environment):
+    """A .pth file with Distribution paths in it"""
+
+    def __init__(self, filename, sitedirs=()) -> None:
+        self.filename = filename
+        self.sitedirs = list(map(normalize_path, sitedirs))
+        self.basedir = normalize_path(os.path.dirname(self.filename))
+        self.paths, self.dirty = self._load()
+        # keep a copy if someone manually updates the paths attribute on the instance
+        self._init_paths = self.paths[:]
+        super().__init__([], None, None)
+        for path in yield_lines(self.paths):
+            list(map(self.add, find_distributions(path, True)))
+
+    def _load_raw(self):
+        paths = []
+        dirty = saw_import = False
+        seen = set(self.sitedirs)
+        content = _read_pth(self.filename)
+        for line in content.splitlines():
+            path = line.rstrip()
+            # still keep imports and empty/commented lines for formatting
+            paths.append(path)
+            if line.startswith(('import ', 'from ')):
+                saw_import = True
+                continue
+            stripped_path = path.strip()
+            if not stripped_path or stripped_path.startswith('#'):
+                continue
+            # skip non-existent paths, in case somebody deleted a package
+            # manually, and duplicate paths as well
+            normalized_path = normalize_path(os.path.join(self.basedir, path))
+            if normalized_path in seen or not os.path.exists(normalized_path):
+                log.debug("cleaned up dirty or duplicated %r", path)
+                dirty = True
+                paths.pop()
+                continue
+            seen.add(normalized_path)
+        # remove any trailing empty/blank line
+        while paths and not paths[-1].strip():
+            paths.pop()
+            dirty = True
+        return paths, dirty or (paths and saw_import)
+
+    def _load(self):
+        if os.path.isfile(self.filename):
+            return self._load_raw()
+        return [], False
+
+    def save(self) -> None:
+        """Write changed .pth file back to disk"""
+        # first reload the file
+        last_paths, last_dirty = self._load()
+        # and check that there are no difference with what we have.
+        # there can be difference if someone else has written to the file
+        # since we first loaded it.
+        # we don't want to lose the eventual new paths added since then.
+        for path in last_paths[:]:
+            if path not in self.paths:
+                self.paths.append(path)
+                log.info("detected new path %r", path)
+                last_dirty = True
+            else:
+                last_paths.remove(path)
+        # also, re-check that all paths are still valid before saving them
+        for path in self.paths[:]:
+            if path not in last_paths and not path.startswith((
+                'import ',
+                'from ',
+                '#',
+            )):
+                absolute_path = os.path.join(self.basedir, path)
+                if not os.path.exists(absolute_path):
+                    self.paths.remove(path)
+                    log.info("removing now non-existent path %r", path)
+                    last_dirty = True
+
+        self.dirty |= last_dirty or self.paths != self._init_paths
+        if not self.dirty:
+            return
+
+        rel_paths = list(map(self.make_relative, self.paths))
+        if rel_paths:
+            log.debug("Saving %s", self.filename)
+            lines = self._wrap_lines(rel_paths)
+            data = '\n'.join(lines) + '\n'
+            if os.path.islink(self.filename):
+                os.unlink(self.filename)
+            with open(self.filename, 'wt', encoding=py312.PTH_ENCODING) as f:
+                # ^-- Python<3.13 require encoding="locale" instead of "utf-8",
+                #     see python/cpython#77102.
+                f.write(data)
+        elif os.path.exists(self.filename):
+            log.debug("Deleting empty %s", self.filename)
+            os.unlink(self.filename)
+
+        self.dirty = False
+        self._init_paths[:] = self.paths[:]
+
+    @staticmethod
+    def _wrap_lines(lines):
+        return lines
+
+    def add(self, dist) -> None:
+        """Add `dist` to the distribution map"""
+        new_path = dist.location not in self.paths and (
+            dist.location not in self.sitedirs
+            or
+            # account for '.' being in PYTHONPATH
+            dist.location == os.getcwd()
+        )
+        if new_path:
+            self.paths.append(dist.location)
+            self.dirty = True
+        super().add(dist)
+
+    def remove(self, dist) -> None:
+        """Remove `dist` from the distribution map"""
+        while dist.location in self.paths:
+            self.paths.remove(dist.location)
+            self.dirty = True
+        super().remove(dist)
+
+    def make_relative(self, path):
+        npath, last = os.path.split(normalize_path(path))
+        baselen = len(self.basedir)
+        parts = [last]
+        sep = os.altsep == '/' and '/' or os.sep
+        while len(npath) >= baselen:
+            if npath == self.basedir:
+                parts.append(os.curdir)
+                parts.reverse()
+                return sep.join(parts)
+            npath, last = os.path.split(npath)
+            parts.append(last)
+        else:
+            return path
+
+
+class RewritePthDistributions(PthDistributions):
+    @classmethod
+    def _wrap_lines(cls, lines):
+        yield cls.prelude
+        yield from lines
+        yield cls.postlude
+
+    prelude = _one_liner(
+        """
+        import sys
+        sys.__plen = len(sys.path)
+        """
+    )
+    postlude = _one_liner(
+        """
+        import sys
+        new = sys.path[sys.__plen:]
+        del sys.path[sys.__plen:]
+        p = getattr(sys, '__egginsert', 0)
+        sys.path[p:p] = new
+        sys.__egginsert = p + len(new)
+        """
+    )
+
+
+if os.environ.get('SETUPTOOLS_SYS_PATH_TECHNIQUE', 'raw') == 'rewrite':
+    PthDistributions = RewritePthDistributions  # type: ignore[misc]  # Overwriting type
+
+
+def _first_line_re():
+    """
+    Return a regular expression based on first_line_re suitable for matching
+    strings.
+    """
+    if isinstance(first_line_re.pattern, str):
+        return first_line_re
+
+    # first_line_re in Python >=3.1.4 and >=3.2.1 is a bytes pattern.
+    return re.compile(first_line_re.pattern.decode())
+
+
+def update_dist_caches(dist_path, fix_zipimporter_caches):
+    """
+    Fix any globally cached `dist_path` related data
+
+    `dist_path` should be a path of a newly installed egg distribution (zipped
+    or unzipped).
+
+    sys.path_importer_cache contains finder objects that have been cached when
+    importing data from the original distribution. Any such finders need to be
+    cleared since the replacement distribution might be packaged differently,
+    e.g. a zipped egg distribution might get replaced with an unzipped egg
+    folder or vice versa. Having the old finders cached may then cause Python
+    to attempt loading modules from the replacement distribution using an
+    incorrect loader.
+
+    zipimport.zipimporter objects are Python loaders charged with importing
+    data packaged inside zip archives. If stale loaders referencing the
+    original distribution, are left behind, they can fail to load modules from
+    the replacement distribution. E.g. if an old zipimport.zipimporter instance
+    is used to load data from a new zipped egg archive, it may cause the
+    operation to attempt to locate the requested data in the wrong location -
+    one indicated by the original distribution's zip archive directory
+    information. Such an operation may then fail outright, e.g. report having
+    read a 'bad local file header', or even worse, it may fail silently &
+    return invalid data.
+
+    zipimport._zip_directory_cache contains cached zip archive directory
+    information for all existing zipimport.zipimporter instances and all such
+    instances connected to the same archive share the same cached directory
+    information.
+
+    If asked, and the underlying Python implementation allows it, we can fix
+    all existing zipimport.zipimporter instances instead of having to track
+    them down and remove them one by one, by updating their shared cached zip
+    archive directory information. This, of course, assumes that the
+    replacement distribution is packaged as a zipped egg.
+
+    If not asked to fix existing zipimport.zipimporter instances, we still do
+    our best to clear any remaining zipimport.zipimporter related cached data
+    that might somehow later get used when attempting to load data from the new
+    distribution and thus cause such load operations to fail. Note that when
+    tracking down such remaining stale data, we can not catch every conceivable
+    usage from here, and we clear only those that we know of and have found to
+    cause problems if left alive. Any remaining caches should be updated by
+    whomever is in charge of maintaining them, i.e. they should be ready to
+    handle us replacing their zip archives with new distributions at runtime.
+
+    """
+    # There are several other known sources of stale zipimport.zipimporter
+    # instances that we do not clear here, but might if ever given a reason to
+    # do so:
+    # * Global setuptools pkg_resources.working_set (a.k.a. 'master working
+    # set') may contain distributions which may in turn contain their
+    #   zipimport.zipimporter loaders.
+    # * Several zipimport.zipimporter loaders held by local variables further
+    #   up the function call stack when running the setuptools installation.
+    # * Already loaded modules may have their __loader__ attribute set to the
+    #   exact loader instance used when importing them. Python 3.4 docs state
+    #   that this information is intended mostly for introspection and so is
+    #   not expected to cause us problems.
+    normalized_path = normalize_path(dist_path)
+    _uncache(normalized_path, sys.path_importer_cache)
+    if fix_zipimporter_caches:
+        _replace_zip_directory_cache_data(normalized_path)
+    else:
+        # Here, even though we do not want to fix existing and now stale
+        # zipimporter cache information, we still want to remove it. Related to
+        # Python's zip archive directory information cache, we clear each of
+        # its stale entries in two phases:
+        #   1. Clear the entry so attempting to access zip archive information
+        #      via any existing stale zipimport.zipimporter instances fails.
+        #   2. Remove the entry from the cache so any newly constructed
+        #      zipimport.zipimporter instances do not end up using old stale
+        #      zip archive directory information.
+        # This whole stale data removal step does not seem strictly necessary,
+        # but has been left in because it was done before we started replacing
+        # the zip archive directory information cache content if possible, and
+        # there are no relevant unit tests that we can depend on to tell us if
+        # this is really needed.
+        _remove_and_clear_zip_directory_cache_data(normalized_path)
+
+
+def _collect_zipimporter_cache_entries(normalized_path, cache):
+    """
+    Return zipimporter cache entry keys related to a given normalized path.
+
+    Alternative path spellings (e.g. those using different character case or
+    those using alternative path separators) related to the same path are
+    included. Any sub-path entries are included as well, i.e. those
+    corresponding to zip archives embedded in other zip archives.
+
+    """
+    result = []
+    prefix_len = len(normalized_path)
+    for p in cache:
+        np = normalize_path(p)
+        if np.startswith(normalized_path) and np[prefix_len : prefix_len + 1] in (
+            os.sep,
+            '',
+        ):
+            result.append(p)
+    return result
+
+
+def _update_zipimporter_cache(normalized_path, cache, updater=None):
+    """
+    Update zipimporter cache data for a given normalized path.
+
+    Any sub-path entries are processed as well, i.e. those corresponding to zip
+    archives embedded in other zip archives.
+
+    Given updater is a callable taking a cache entry key and the original entry
+    (after already removing the entry from the cache), and expected to update
+    the entry and possibly return a new one to be inserted in its place.
+    Returning None indicates that the entry should not be replaced with a new
+    one. If no updater is given, the cache entries are simply removed without
+    any additional processing, the same as if the updater simply returned None.
+
+    """
+    for p in _collect_zipimporter_cache_entries(normalized_path, cache):
+        # N.B. pypy's custom zipimport._zip_directory_cache implementation does
+        # not support the complete dict interface:
+        # * Does not support item assignment, thus not allowing this function
+        #    to be used only for removing existing cache entries.
+        #  * Does not support the dict.pop() method, forcing us to use the
+        #    get/del patterns instead. For more detailed information see the
+        #    following links:
+        #      https://github.com/pypa/setuptools/issues/202#issuecomment-202913420
+        #      https://foss.heptapod.net/pypy/pypy/-/blob/144c4e65cb6accb8e592f3a7584ea38265d1873c/pypy/module/zipimport/interp_zipimport.py
+        old_entry = cache[p]
+        del cache[p]
+        new_entry = updater and updater(p, old_entry)
+        if new_entry is not None:
+            cache[p] = new_entry
+
+
+def _uncache(normalized_path, cache):
+    _update_zipimporter_cache(normalized_path, cache)
+
+
+def _remove_and_clear_zip_directory_cache_data(normalized_path):
+    def clear_and_remove_cached_zip_archive_directory_data(path, old_entry):
+        old_entry.clear()
+
+    _update_zipimporter_cache(
+        normalized_path,
+        zipimport._zip_directory_cache,
+        updater=clear_and_remove_cached_zip_archive_directory_data,
+    )
+
+
+# PyPy Python implementation does not allow directly writing to the
+# zipimport._zip_directory_cache and so prevents us from attempting to correct
+# its content. The best we can do there is clear the problematic cache content
+# and have PyPy repopulate it as needed. The downside is that if there are any
+# stale zipimport.zipimporter instances laying around, attempting to use them
+# will fail due to not having its zip archive directory information available
+# instead of being automatically corrected to use the new correct zip archive
+# directory information.
+if '__pypy__' in sys.builtin_module_names:
+    _replace_zip_directory_cache_data = _remove_and_clear_zip_directory_cache_data
+else:
+
+    def _replace_zip_directory_cache_data(normalized_path):
+        def replace_cached_zip_archive_directory_data(path, old_entry):
+            # N.B. In theory, we could load the zip directory information just
+            # once for all updated path spellings, and then copy it locally and
+            # update its contained path strings to contain the correct
+            # spelling, but that seems like a way too invasive move (this cache
+            # structure is not officially documented anywhere and could in
+            # theory change with new Python releases) for no significant
+            # benefit.
+            old_entry.clear()
+            zipimport.zipimporter(path)
+            old_entry.update(zipimport._zip_directory_cache[path])
+            return old_entry
+
+        _update_zipimporter_cache(
+            normalized_path,
+            zipimport._zip_directory_cache,
+            updater=replace_cached_zip_archive_directory_data,
+        )
+
+
+def is_python(text, filename='<string>'):
+    "Is this string a valid Python script?"
+    try:
+        compile(text, filename, 'exec')
+    except (SyntaxError, TypeError):
+        return False
+    else:
+        return True
+
+
+def is_sh(executable):
+    """Determine if the specified executable is a .sh (contains a #! line)"""
+    try:
+        with open(executable, encoding='latin-1') as fp:
+            magic = fp.read(2)
+    except OSError:
+        return executable
+    return magic == '#!'
+
+
+def nt_quote_arg(arg):
+    """Quote a command line argument according to Windows parsing rules"""
+    return subprocess.list2cmdline([arg])
+
+
+def is_python_script(script_text, filename):
+    """Is this text, as a whole, a Python script? (as opposed to shell/bat/etc."""
+    if filename.endswith('.py') or filename.endswith('.pyw'):
+        return True  # extension says it's Python
+    if is_python(script_text, filename):
+        return True  # it's syntactically valid Python
+    if script_text.startswith('#!'):
+        # It begins with a '#!' line, so check if 'python' is in it somewhere
+        return 'python' in script_text.splitlines()[0].lower()
+
+    return False  # Not any Python I can recognize
+
+
+class _SplitArgs(TypedDict, total=False):
+    comments: bool
+    posix: bool
+
+
+class CommandSpec(list):
+    """
+    A command spec for a #! header, specified as a list of arguments akin to
+    those passed to Popen.
+    """
+
+    options: list[str] = []
+    split_args = _SplitArgs()
+
+    @classmethod
+    def best(cls):
+        """
+        Choose the best CommandSpec class based on environmental conditions.
+        """
+        return cls
+
+    @classmethod
+    def _sys_executable(cls):
+        _default = os.path.normpath(sys.executable)
+        return os.environ.get('__PYVENV_LAUNCHER__', _default)
+
+    @classmethod
+    def from_param(cls, param: Self | str | Iterable[str] | None) -> Self:
+        """
+        Construct a CommandSpec from a parameter to build_scripts, which may
+        be None.
+        """
+        if isinstance(param, cls):
+            return param
+        if isinstance(param, str):
+            return cls.from_string(param)
+        if isinstance(param, Iterable):
+            return cls(param)
+        if param is None:
+            return cls.from_environment()
+        raise TypeError(f"Argument has an unsupported type {type(param)}")
+
+    @classmethod
+    def from_environment(cls):
+        return cls([cls._sys_executable()])
+
+    @classmethod
+    def from_string(cls, string: str) -> Self:
+        """
+        Construct a command spec from a simple string representing a command
+        line parseable by shlex.split.
+        """
+        items = shlex.split(string, **cls.split_args)
+        return cls(items)
+
+    def install_options(self, script_text: str):
+        self.options = shlex.split(self._extract_options(script_text))
+        cmdline = subprocess.list2cmdline(self)
+        if not isascii(cmdline):
+            self.options[:0] = ['-x']
+
+    @staticmethod
+    def _extract_options(orig_script):
+        """
+        Extract any options from the first line of the script.
+        """
+        first = (orig_script + '\n').splitlines()[0]
+        match = _first_line_re().match(first)
+        options = match.group(1) or '' if match else ''
+        return options.strip()
+
+    def as_header(self):
+        return self._render(self + list(self.options))
+
+    @staticmethod
+    def _strip_quotes(item):
+        _QUOTES = '"\''
+        for q in _QUOTES:
+            if item.startswith(q) and item.endswith(q):
+                return item[1:-1]
+        return item
+
+    @staticmethod
+    def _render(items):
+        cmdline = subprocess.list2cmdline(
+            CommandSpec._strip_quotes(item.strip()) for item in items
+        )
+        return '#!' + cmdline + '\n'
+
+
+# For pbr compat; will be removed in a future version.
+sys_executable = CommandSpec._sys_executable()
+
+
+class WindowsCommandSpec(CommandSpec):
+    split_args = _SplitArgs(posix=False)
+
+
+class ScriptWriter:
+    """
+    Encapsulates behavior around writing entry point scripts for console and
+    gui apps.
+    """
+
+    template = textwrap.dedent(
+        r"""
+        # EASY-INSTALL-ENTRY-SCRIPT: %(spec)r,%(group)r,%(name)r
+        import re
+        import sys
+
+        # for compatibility with easy_install; see #2198
+        __requires__ = %(spec)r
+
+        try:
+            from importlib.metadata import distribution
+        except ImportError:
+            try:
+                from importlib_metadata import distribution
+            except ImportError:
+                from pkg_resources import load_entry_point
+
+
+        def importlib_load_entry_point(spec, group, name):
+            dist_name, _, _ = spec.partition('==')
+            matches = (
+                entry_point
+                for entry_point in distribution(dist_name).entry_points
+                if entry_point.group == group and entry_point.name == name
+            )
+            return next(matches).load()
+
+
+        globals().setdefault('load_entry_point', importlib_load_entry_point)
+
+
+        if __name__ == '__main__':
+            sys.argv[0] = re.sub(r'(-script\.pyw?|\.exe)?$', '', sys.argv[0])
+            sys.exit(load_entry_point(%(spec)r, %(group)r, %(name)r)())
+        """
+    ).lstrip()
+
+    command_spec_class = CommandSpec
+
+    @classmethod
+    def get_args(cls, dist, header=None):
+        """
+        Yield write_script() argument tuples for a distribution's
+        console_scripts and gui_scripts entry points.
+        """
+        if header is None:
+            header = cls.get_header()
+        spec = str(dist.as_requirement())
+        for type_ in 'console', 'gui':
+            group = type_ + '_scripts'
+            for name in dist.get_entry_map(group).keys():
+                cls._ensure_safe_name(name)
+                script_text = cls.template % locals()
+                args = cls._get_script_args(type_, name, header, script_text)
+                yield from args
+
+    @staticmethod
+    def _ensure_safe_name(name):
+        """
+        Prevent paths in *_scripts entry point names.
+        """
+        has_path_sep = re.search(r'[\\/]', name)
+        if has_path_sep:
+            raise ValueError("Path separators not allowed in script names")
+
+    @classmethod
+    def best(cls):
+        """
+        Select the best ScriptWriter for this environment.
+        """
+        if sys.platform == 'win32' or (os.name == 'java' and os._name == 'nt'):
+            return WindowsScriptWriter.best()
+        else:
+            return cls
+
+    @classmethod
+    def _get_script_args(cls, type_, name, header, script_text):
+        # Simply write the stub with no extension.
+        yield (name, header + script_text)
+
+    @classmethod
+    def get_header(
+        cls,
+        script_text: str = "",
+        executable: str | CommandSpec | Iterable[str] | None = None,
+    ) -> str:
+        """Create a #! line, getting options (if any) from script_text"""
+        cmd = cls.command_spec_class.best().from_param(executable)
+        cmd.install_options(script_text)
+        return cmd.as_header()
+
+
+class WindowsScriptWriter(ScriptWriter):
+    command_spec_class = WindowsCommandSpec
+
+    @classmethod
+    def best(cls):
+        """
+        Select the best ScriptWriter suitable for Windows
+        """
+        writer_lookup = dict(
+            executable=WindowsExecutableLauncherWriter,
+            natural=cls,
+        )
+        # for compatibility, use the executable launcher by default
+        launcher = os.environ.get('SETUPTOOLS_LAUNCHER', 'executable')
+        return writer_lookup[launcher]
+
+    @classmethod
+    def _get_script_args(cls, type_, name, header, script_text):
+        "For Windows, add a .py extension"
+        ext = dict(console='.pya', gui='.pyw')[type_]
+        if ext not in os.environ['PATHEXT'].lower().split(';'):
+            msg = (
+                "{ext} not listed in PATHEXT; scripts will not be "
+                "recognized as executables."
+            ).format(**locals())
+            SetuptoolsWarning.emit(msg)
+        old = ['.pya', '.py', '-script.py', '.pyc', '.pyo', '.pyw', '.exe']
+        old.remove(ext)
+        header = cls._adjust_header(type_, header)
+        blockers = [name + x for x in old]
+        yield name + ext, header + script_text, 't', blockers
+
+    @classmethod
+    def _adjust_header(cls, type_, orig_header):
+        """
+        Make sure 'pythonw' is used for gui and 'python' is used for
+        console (regardless of what sys.executable is).
+        """
+        pattern = 'pythonw.exe'
+        repl = 'python.exe'
+        if type_ == 'gui':
+            pattern, repl = repl, pattern
+        pattern_ob = re.compile(re.escape(pattern), re.IGNORECASE)
+        new_header = pattern_ob.sub(string=orig_header, repl=repl)
+        return new_header if cls._use_header(new_header) else orig_header
+
+    @staticmethod
+    def _use_header(new_header):
+        """
+        Should _adjust_header use the replaced header?
+
+        On non-windows systems, always use. On
+        Windows systems, only use the replaced header if it resolves
+        to an executable on the system.
+        """
+        clean_header = new_header[2:-1].strip('"')
+        return sys.platform != 'win32' or shutil.which(clean_header)
+
+
+class WindowsExecutableLauncherWriter(WindowsScriptWriter):
+    @classmethod
+    def _get_script_args(cls, type_, name, header, script_text):
+        """
+        For Windows, add a .py extension and an .exe launcher
+        """
+        if type_ == 'gui':
+            launcher_type = 'gui'
+            ext = '-script.pyw'
+            old = ['.pyw']
+        else:
+            launcher_type = 'cli'
+            ext = '-script.py'
+            old = ['.py', '.pyc', '.pyo']
+        hdr = cls._adjust_header(type_, header)
+        blockers = [name + x for x in old]
+        yield (name + ext, hdr + script_text, 't', blockers)
+        yield (
+            name + '.exe',
+            get_win_launcher(launcher_type),
+            'b',  # write in binary mode
+        )
+        if not is_64bit():
+            # install a manifest for the launcher to prevent Windows
+            # from detecting it as an installer (which it will for
+            #  launchers like easy_install.exe). Consider only
+            #  adding a manifest for launchers detected as installers.
+            #  See Distribute #143 for details.
+            m_name = name + '.exe.manifest'
+            yield (m_name, load_launcher_manifest(name), 't')
+
+
+def get_win_launcher(type):
+    """
+    Load the Windows launcher (executable) suitable for launching a script.
+
+    `type` should be either 'cli' or 'gui'
+
+    Returns the executable as a byte string.
+    """
+    launcher_fn = f'{type}.exe'
+    if is_64bit():
+        if get_platform() == "win-arm64":
+            launcher_fn = launcher_fn.replace(".", "-arm64.")
+        else:
+            launcher_fn = launcher_fn.replace(".", "-64.")
+    else:
+        launcher_fn = launcher_fn.replace(".", "-32.")
+    return resource_string('setuptools', launcher_fn)
+
+
+def load_launcher_manifest(name):
+    manifest = pkg_resources.resource_string(__name__, 'launcher manifest.xml')
+    return manifest.decode('utf-8') % vars()
+
+
+def current_umask():
+    tmp = os.umask(0o022)
+    os.umask(tmp)
+    return tmp
+
+
+def only_strs(values):
+    """
+    Exclude non-str values. Ref #3063.
+    """
+    return filter(lambda val: isinstance(val, str), values)
+
+
+def _read_pth(fullname: str) -> str:
+    # Python<3.13 require encoding="locale" instead of "utf-8", see python/cpython#77102
+    # In the case old versions of setuptools are producing `pth` files with
+    # different encodings that might be problematic... So we fallback to "locale".
+
+    try:
+        with open(fullname, encoding=py312.PTH_ENCODING) as f:
+            return f.read()
+    except UnicodeDecodeError:  # pragma: no cover
+        # This error may only happen for Python >= 3.13
+        # TODO: Possible deprecation warnings to be added in the future:
+        #       ``.pth file {fullname!r} is not UTF-8.``
+        #       Your environment contain {fullname!r} that cannot be read as UTF-8.
+        #       This is likely to have been produced with an old version of setuptools.
+        #       Please be mindful that this is deprecated and in the future, non-utf8
+        #       .pth files may cause setuptools to fail.
+        with open(fullname, encoding=py39.LOCALE_ENCODING) as f:
+            return f.read()
+
+
+class EasyInstallDeprecationWarning(SetuptoolsDeprecationWarning):
+    _SUMMARY = "easy_install command is deprecated."
+    _DETAILS = """
+    Please avoid running ``setup.py`` and ``easy_install``.
+    Instead, use pypa/build, pypa/installer or other
+    standards-based tools.
+    """
+    _SEE_URL = "https://github.com/pypa/setuptools/issues/917"
+    # _DUE_DATE not defined yet
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/editable_wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/editable_wheel.py
new file mode 100644
index 00000000..b03e6777
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/editable_wheel.py
@@ -0,0 +1,925 @@
+"""
+Create a wheel that, when installed, will make the source package 'editable'
+(add it to the interpreter's path, including metadata) per PEP 660. Replaces
+'setup.py develop'.
+
+.. note::
+   One of the mechanisms briefly mentioned in PEP 660 to implement editable installs is
+   to create a separated directory inside ``build`` and use a .pth file to point to that
+   directory. In the context of this file such directory is referred as
+   *auxiliary build directory* or ``auxiliary_dir``.
+"""
+
+from __future__ import annotations
+
+import io
+import logging
+import os
+import shutil
+import traceback
+from collections.abc import Iterable, Iterator, Mapping
+from contextlib import suppress
+from enum import Enum
+from inspect import cleandoc
+from itertools import chain, starmap
+from pathlib import Path
+from tempfile import TemporaryDirectory
+from types import TracebackType
+from typing import TYPE_CHECKING, Protocol, TypeVar, cast
+
+from .. import Command, _normalization, _path, _shutil, errors, namespaces
+from .._path import StrPath
+from ..compat import py312
+from ..discovery import find_package_path
+from ..dist import Distribution
+from ..warnings import InformationOnly, SetuptoolsDeprecationWarning, SetuptoolsWarning
+from .build import build as build_cls
+from .build_py import build_py as build_py_cls
+from .dist_info import dist_info as dist_info_cls
+from .egg_info import egg_info as egg_info_cls
+from .install import install as install_cls
+from .install_scripts import install_scripts as install_scripts_cls
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from .._vendor.wheel.wheelfile import WheelFile
+
+_P = TypeVar("_P", bound=StrPath)
+_logger = logging.getLogger(__name__)
+
+
+class _EditableMode(Enum):
+    """
+    Possible editable installation modes:
+    `lenient` (new files automatically added to the package - DEFAULT);
+    `strict` (requires a new installation when files are added/removed); or
+    `compat` (attempts to emulate `python setup.py develop` - DEPRECATED).
+    """
+
+    STRICT = "strict"
+    LENIENT = "lenient"
+    COMPAT = "compat"  # TODO: Remove `compat` after Dec/2022.
+
+    @classmethod
+    def convert(cls, mode: str | None) -> _EditableMode:
+        if not mode:
+            return _EditableMode.LENIENT  # default
+
+        _mode = mode.upper()
+        if _mode not in _EditableMode.__members__:
+            raise errors.OptionError(f"Invalid editable mode: {mode!r}. Try: 'strict'.")
+
+        if _mode == "COMPAT":
+            SetuptoolsDeprecationWarning.emit(
+                "Compat editable installs",
+                """
+                The 'compat' editable mode is transitional and will be removed
+                in future versions of `setuptools`.
+                Please adapt your code accordingly to use either the 'strict' or the
+                'lenient' modes.
+                """,
+                see_docs="userguide/development_mode.html",
+                # TODO: define due_date
+                # There is a series of shortcomings with the available editable install
+                # methods, and they are very controversial. This is something that still
+                # needs work.
+                # Moreover, `pip` is still hiding this warning, so users are not aware.
+            )
+
+        return _EditableMode[_mode]
+
+
+_STRICT_WARNING = """
+New or renamed files may not be automatically picked up without a new installation.
+"""
+
+_LENIENT_WARNING = """
+Options like `package-data`, `include/exclude-package-data` or
+`packages.find.exclude/include` may have no effect.
+"""
+
+
+class editable_wheel(Command):
+    """Build 'editable' wheel for development.
+    This command is private and reserved for internal use of setuptools,
+    users should rely on ``setuptools.build_meta`` APIs.
+    """
+
+    description = "DO NOT CALL DIRECTLY, INTERNAL ONLY: create PEP 660 editable wheel"
+
+    user_options = [
+        ("dist-dir=", "d", "directory to put final built distributions in"),
+        ("dist-info-dir=", "I", "path to a pre-build .dist-info directory"),
+        ("mode=", None, cleandoc(_EditableMode.__doc__ or "")),
+    ]
+
+    def initialize_options(self):
+        self.dist_dir = None
+        self.dist_info_dir = None
+        self.project_dir = None
+        self.mode = None
+
+    def finalize_options(self) -> None:
+        dist = self.distribution
+        self.project_dir = dist.src_root or os.curdir
+        self.package_dir = dist.package_dir or {}
+        self.dist_dir = Path(self.dist_dir or os.path.join(self.project_dir, "dist"))
+
+    def run(self) -> None:
+        try:
+            self.dist_dir.mkdir(exist_ok=True)
+            self._ensure_dist_info()
+
+            # Add missing dist_info files
+            self.reinitialize_command("bdist_wheel")
+            bdist_wheel = self.get_finalized_command("bdist_wheel")
+            bdist_wheel.write_wheelfile(self.dist_info_dir)
+
+            self._create_wheel_file(bdist_wheel)
+        except Exception:
+            traceback.print_exc()
+            project = self.distribution.name or self.distribution.get_name()
+            _DebuggingTips.emit(project=project)
+            raise
+
+    def _ensure_dist_info(self):
+        if self.dist_info_dir is None:
+            dist_info = cast(dist_info_cls, self.reinitialize_command("dist_info"))
+            dist_info.output_dir = self.dist_dir
+            dist_info.ensure_finalized()
+            dist_info.run()
+            self.dist_info_dir = dist_info.dist_info_dir
+        else:
+            assert str(self.dist_info_dir).endswith(".dist-info")
+            assert Path(self.dist_info_dir, "METADATA").exists()
+
+    def _install_namespaces(self, installation_dir, pth_prefix):
+        # XXX: Only required to support the deprecated namespace practice
+        dist = self.distribution
+        if not dist.namespace_packages:
+            return
+
+        src_root = Path(self.project_dir, self.package_dir.get("", ".")).resolve()
+        installer = _NamespaceInstaller(dist, installation_dir, pth_prefix, src_root)
+        installer.install_namespaces()
+
+    def _find_egg_info_dir(self) -> str | None:
+        parent_dir = Path(self.dist_info_dir).parent if self.dist_info_dir else Path()
+        candidates = map(str, parent_dir.glob("*.egg-info"))
+        return next(candidates, None)
+
+    def _configure_build(
+        self, name: str, unpacked_wheel: StrPath, build_lib: StrPath, tmp_dir: StrPath
+    ):
+        """Configure commands to behave in the following ways:
+
+        - Build commands can write to ``build_lib`` if they really want to...
+          (but this folder is expected to be ignored and modules are expected to live
+          in the project directory...)
+        - Binary extensions should be built in-place (editable_mode = True)
+        - Data/header/script files are not part of the "editable" specification
+          so they are written directly to the unpacked_wheel directory.
+        """
+        # Non-editable files (data, headers, scripts) are written directly to the
+        # unpacked_wheel
+
+        dist = self.distribution
+        wheel = str(unpacked_wheel)
+        build_lib = str(build_lib)
+        data = str(Path(unpacked_wheel, f"{name}.data", "data"))
+        headers = str(Path(unpacked_wheel, f"{name}.data", "headers"))
+        scripts = str(Path(unpacked_wheel, f"{name}.data", "scripts"))
+
+        # egg-info may be generated again to create a manifest (used for package data)
+        egg_info = cast(
+            egg_info_cls, dist.reinitialize_command("egg_info", reinit_subcommands=True)
+        )
+        egg_info.egg_base = str(tmp_dir)
+        egg_info.ignore_egg_info_in_manifest = True
+
+        build = cast(
+            build_cls, dist.reinitialize_command("build", reinit_subcommands=True)
+        )
+        install = cast(
+            install_cls, dist.reinitialize_command("install", reinit_subcommands=True)
+        )
+
+        build.build_platlib = build.build_purelib = build.build_lib = build_lib
+        install.install_purelib = install.install_platlib = install.install_lib = wheel
+        install.install_scripts = build.build_scripts = scripts
+        install.install_headers = headers
+        install.install_data = data
+
+        install_scripts = cast(
+            install_scripts_cls, dist.get_command_obj("install_scripts")
+        )
+        install_scripts.no_ep = True
+
+        build.build_temp = str(tmp_dir)
+
+        build_py = cast(build_py_cls, dist.get_command_obj("build_py"))
+        build_py.compile = False
+        build_py.existing_egg_info_dir = self._find_egg_info_dir()
+
+        self._set_editable_mode()
+
+        build.ensure_finalized()
+        install.ensure_finalized()
+
+    def _set_editable_mode(self):
+        """Set the ``editable_mode`` flag in the build sub-commands"""
+        dist = self.distribution
+        build = dist.get_command_obj("build")
+        for cmd_name in build.get_sub_commands():
+            cmd = dist.get_command_obj(cmd_name)
+            if hasattr(cmd, "editable_mode"):
+                cmd.editable_mode = True
+            elif hasattr(cmd, "inplace"):
+                cmd.inplace = True  # backward compatibility with distutils
+
+    def _collect_build_outputs(self) -> tuple[list[str], dict[str, str]]:
+        files: list[str] = []
+        mapping: dict[str, str] = {}
+        build = self.get_finalized_command("build")
+
+        for cmd_name in build.get_sub_commands():
+            cmd = self.get_finalized_command(cmd_name)
+            if hasattr(cmd, "get_outputs"):
+                files.extend(cmd.get_outputs() or [])
+            if hasattr(cmd, "get_output_mapping"):
+                mapping.update(cmd.get_output_mapping() or {})
+
+        return files, mapping
+
+    def _run_build_commands(
+        self,
+        dist_name: str,
+        unpacked_wheel: StrPath,
+        build_lib: StrPath,
+        tmp_dir: StrPath,
+    ) -> tuple[list[str], dict[str, str]]:
+        self._configure_build(dist_name, unpacked_wheel, build_lib, tmp_dir)
+        self._run_build_subcommands()
+        files, mapping = self._collect_build_outputs()
+        self._run_install("headers")
+        self._run_install("scripts")
+        self._run_install("data")
+        return files, mapping
+
+    def _run_build_subcommands(self) -> None:
+        """
+        Issue #3501 indicates that some plugins/customizations might rely on:
+
+        1. ``build_py`` not running
+        2. ``build_py`` always copying files to ``build_lib``
+
+        However both these assumptions may be false in editable_wheel.
+        This method implements a temporary workaround to support the ecosystem
+        while the implementations catch up.
+        """
+        # TODO: Once plugins/customisations had the chance to catch up, replace
+        #       `self._run_build_subcommands()` with `self.run_command("build")`.
+        #       Also remove _safely_run, TestCustomBuildPy. Suggested date: Aug/2023.
+        build = self.get_finalized_command("build")
+        for name in build.get_sub_commands():
+            cmd = self.get_finalized_command(name)
+            if name == "build_py" and type(cmd) is not build_py_cls:
+                self._safely_run(name)
+            else:
+                self.run_command(name)
+
+    def _safely_run(self, cmd_name: str):
+        try:
+            return self.run_command(cmd_name)
+        except Exception:
+            SetuptoolsDeprecationWarning.emit(
+                "Customization incompatible with editable install",
+                f"""
+                {traceback.format_exc()}
+
+                If you are seeing this warning it is very likely that a setuptools
+                plugin or customization overrides the `{cmd_name}` command, without
+                taking into consideration how editable installs run build steps
+                starting from setuptools v64.0.0.
+
+                Plugin authors and developers relying on custom build steps are
+                encouraged to update their `{cmd_name}` implementation considering the
+                information about editable installs in
+                https://setuptools.pypa.io/en/latest/userguide/extension.html.
+
+                For the time being `setuptools` will silence this error and ignore
+                the faulty command, but this behaviour will change in future versions.
+                """,
+                # TODO: define due_date
+                # There is a series of shortcomings with the available editable install
+                # methods, and they are very controversial. This is something that still
+                # needs work.
+            )
+
+    def _create_wheel_file(self, bdist_wheel):
+        from wheel.wheelfile import WheelFile
+
+        dist_info = self.get_finalized_command("dist_info")
+        dist_name = dist_info.name
+        tag = "-".join(bdist_wheel.get_tag())
+        build_tag = "0.editable"  # According to PEP 427 needs to start with digit
+        archive_name = f"{dist_name}-{build_tag}-{tag}.whl"
+        wheel_path = Path(self.dist_dir, archive_name)
+        if wheel_path.exists():
+            wheel_path.unlink()
+
+        unpacked_wheel = TemporaryDirectory(suffix=archive_name)
+        build_lib = TemporaryDirectory(suffix=".build-lib")
+        build_tmp = TemporaryDirectory(suffix=".build-temp")
+
+        with unpacked_wheel as unpacked, build_lib as lib, build_tmp as tmp:
+            unpacked_dist_info = Path(unpacked, Path(self.dist_info_dir).name)
+            shutil.copytree(self.dist_info_dir, unpacked_dist_info)
+            self._install_namespaces(unpacked, dist_name)
+            files, mapping = self._run_build_commands(dist_name, unpacked, lib, tmp)
+            strategy = self._select_strategy(dist_name, tag, lib)
+            with strategy, WheelFile(wheel_path, "w") as wheel_obj:
+                strategy(wheel_obj, files, mapping)
+                wheel_obj.write_files(unpacked)
+
+        return wheel_path
+
+    def _run_install(self, category: str):
+        has_category = getattr(self.distribution, f"has_{category}", None)
+        if has_category and has_category():
+            _logger.info(f"Installing {category} as non editable")
+            self.run_command(f"install_{category}")
+
+    def _select_strategy(
+        self,
+        name: str,
+        tag: str,
+        build_lib: StrPath,
+    ) -> EditableStrategy:
+        """Decides which strategy to use to implement an editable installation."""
+        build_name = f"__editable__.{name}-{tag}"
+        project_dir = Path(self.project_dir)
+        mode = _EditableMode.convert(self.mode)
+
+        if mode is _EditableMode.STRICT:
+            auxiliary_dir = _empty_dir(Path(self.project_dir, "build", build_name))
+            return _LinkTree(self.distribution, name, auxiliary_dir, build_lib)
+
+        packages = _find_packages(self.distribution)
+        has_simple_layout = _simple_layout(packages, self.package_dir, project_dir)
+        is_compat_mode = mode is _EditableMode.COMPAT
+        if set(self.package_dir) == {""} and has_simple_layout or is_compat_mode:
+            # src-layout(ish) is relatively safe for a simple pth file
+            src_dir = self.package_dir.get("", ".")
+            return _StaticPth(self.distribution, name, [Path(project_dir, src_dir)])
+
+        # Use a MetaPathFinder to avoid adding accidental top-level packages/modules
+        return _TopLevelFinder(self.distribution, name)
+
+
+class EditableStrategy(Protocol):
+    def __call__(
+        self, wheel: WheelFile, files: list[str], mapping: Mapping[str, str]
+    ) -> object: ...
+    def __enter__(self) -> Self: ...
+    def __exit__(
+        self,
+        _exc_type: type[BaseException] | None,
+        _exc_value: BaseException | None,
+        _traceback: TracebackType | None,
+    ) -> object: ...
+
+
+class _StaticPth:
+    def __init__(self, dist: Distribution, name: str, path_entries: list[Path]) -> None:
+        self.dist = dist
+        self.name = name
+        self.path_entries = path_entries
+
+    def __call__(self, wheel: WheelFile, files: list[str], mapping: Mapping[str, str]):
+        entries = "\n".join(str(p.resolve()) for p in self.path_entries)
+        contents = _encode_pth(f"{entries}\n")
+        wheel.writestr(f"__editable__.{self.name}.pth", contents)
+
+    def __enter__(self) -> Self:
+        msg = f"""
+        Editable install will be performed using .pth file to extend `sys.path` with:
+        {list(map(os.fspath, self.path_entries))!r}
+        """
+        _logger.warning(msg + _LENIENT_WARNING)
+        return self
+
+    def __exit__(
+        self,
+        _exc_type: object,
+        _exc_value: object,
+        _traceback: object,
+    ) -> None:
+        pass
+
+
+class _LinkTree(_StaticPth):
+    """
+    Creates a ``.pth`` file that points to a link tree in the ``auxiliary_dir``.
+
+    This strategy will only link files (not dirs), so it can be implemented in
+    any OS, even if that means using hardlinks instead of symlinks.
+
+    By collocating ``auxiliary_dir`` and the original source code, limitations
+    with hardlinks should be avoided.
+    """
+
+    def __init__(
+        self,
+        dist: Distribution,
+        name: str,
+        auxiliary_dir: StrPath,
+        build_lib: StrPath,
+    ) -> None:
+        self.auxiliary_dir = Path(auxiliary_dir)
+        self.build_lib = Path(build_lib).resolve()
+        self._file = dist.get_command_obj("build_py").copy_file
+        super().__init__(dist, name, [self.auxiliary_dir])
+
+    def __call__(self, wheel: WheelFile, files: list[str], mapping: Mapping[str, str]):
+        self._create_links(files, mapping)
+        super().__call__(wheel, files, mapping)
+
+    def _normalize_output(self, file: str) -> str | None:
+        # Files relative to build_lib will be normalized to None
+        with suppress(ValueError):
+            path = Path(file).resolve().relative_to(self.build_lib)
+            return str(path).replace(os.sep, '/')
+        return None
+
+    def _create_file(self, relative_output: str, src_file: str, link=None):
+        dest = self.auxiliary_dir / relative_output
+        if not dest.parent.is_dir():
+            dest.parent.mkdir(parents=True)
+        self._file(src_file, dest, link=link)
+
+    def _create_links(self, outputs, output_mapping: Mapping[str, str]):
+        self.auxiliary_dir.mkdir(parents=True, exist_ok=True)
+        link_type = "sym" if _can_symlink_files(self.auxiliary_dir) else "hard"
+        normalised = ((self._normalize_output(k), v) for k, v in output_mapping.items())
+        # remove files that are not relative to build_lib
+        mappings = {k: v for k, v in normalised if k is not None}
+
+        for output in outputs:
+            relative = self._normalize_output(output)
+            if relative and relative not in mappings:
+                self._create_file(relative, output)
+
+        for relative, src in mappings.items():
+            self._create_file(relative, src, link=link_type)
+
+    def __enter__(self) -> Self:
+        msg = "Strict editable install will be performed using a link tree.\n"
+        _logger.warning(msg + _STRICT_WARNING)
+        return self
+
+    def __exit__(
+        self,
+        _exc_type: object,
+        _exc_value: object,
+        _traceback: object,
+    ) -> None:
+        msg = f"""\n
+        Strict editable installation performed using the auxiliary directory:
+            {self.auxiliary_dir}
+
+        Please be careful to not remove this directory, otherwise you might not be able
+        to import/use your package.
+        """
+        InformationOnly.emit("Editable installation.", msg)
+
+
+class _TopLevelFinder:
+    def __init__(self, dist: Distribution, name: str) -> None:
+        self.dist = dist
+        self.name = name
+
+    def template_vars(self) -> tuple[str, str, dict[str, str], dict[str, list[str]]]:
+        src_root = self.dist.src_root or os.curdir
+        top_level = chain(_find_packages(self.dist), _find_top_level_modules(self.dist))
+        package_dir = self.dist.package_dir or {}
+        roots = _find_package_roots(top_level, package_dir, src_root)
+
+        namespaces_ = dict(
+            chain(
+                _find_namespaces(self.dist.packages or [], roots),
+                ((ns, []) for ns in _find_virtual_namespaces(roots)),
+            )
+        )
+
+        legacy_namespaces = {
+            pkg: find_package_path(pkg, roots, self.dist.src_root or "")
+            for pkg in self.dist.namespace_packages or []
+        }
+
+        mapping = {**roots, **legacy_namespaces}
+        # ^-- We need to explicitly add the legacy_namespaces to the mapping to be
+        #     able to import their modules even if another package sharing the same
+        #     namespace is installed in a conventional (non-editable) way.
+
+        name = f"__editable__.{self.name}.finder"
+        finder = _normalization.safe_identifier(name)
+        return finder, name, mapping, namespaces_
+
+    def get_implementation(self) -> Iterator[tuple[str, bytes]]:
+        finder, name, mapping, namespaces_ = self.template_vars()
+
+        content = bytes(_finder_template(name, mapping, namespaces_), "utf-8")
+        yield (f"{finder}.py", content)
+
+        content = _encode_pth(f"import {finder}; {finder}.install()")
+        yield (f"__editable__.{self.name}.pth", content)
+
+    def __call__(self, wheel: WheelFile, files: list[str], mapping: Mapping[str, str]):
+        for file, content in self.get_implementation():
+            wheel.writestr(file, content)
+
+    def __enter__(self) -> Self:
+        msg = "Editable install will be performed using a meta path finder.\n"
+        _logger.warning(msg + _LENIENT_WARNING)
+        return self
+
+    def __exit__(
+        self,
+        _exc_type: object,
+        _exc_value: object,
+        _traceback: object,
+    ) -> None:
+        msg = """\n
+        Please be careful with folders in your working directory with the same
+        name as your package as they may take precedence during imports.
+        """
+        InformationOnly.emit("Editable installation.", msg)
+
+
+def _encode_pth(content: str) -> bytes:
+    """
+    Prior to Python 3.13 (see https://github.com/python/cpython/issues/77102),
+    .pth files are always read with 'locale' encoding, the recommendation
+    from the cpython core developers is to write them as ``open(path, "w")``
+    and ignore warnings (see python/cpython#77102, pypa/setuptools#3937).
+    This function tries to simulate this behaviour without having to create an
+    actual file, in a way that supports a range of active Python versions.
+    (There seems to be some variety in the way different version of Python handle
+    ``encoding=None``, not all of them use ``locale.getpreferredencoding(False)``
+    or ``locale.getencoding()``).
+    """
+    with io.BytesIO() as buffer:
+        wrapper = io.TextIOWrapper(buffer, encoding=py312.PTH_ENCODING)
+        # TODO: Python 3.13 replace the whole function with `bytes(content, "utf-8")`
+        wrapper.write(content)
+        wrapper.flush()
+        buffer.seek(0)
+        return buffer.read()
+
+
+def _can_symlink_files(base_dir: Path) -> bool:
+    with TemporaryDirectory(dir=str(base_dir.resolve())) as tmp:
+        path1, path2 = Path(tmp, "file1.txt"), Path(tmp, "file2.txt")
+        path1.write_text("file1", encoding="utf-8")
+        with suppress(AttributeError, NotImplementedError, OSError):
+            os.symlink(path1, path2)
+            if path2.is_symlink() and path2.read_text(encoding="utf-8") == "file1":
+                return True
+
+        try:
+            os.link(path1, path2)  # Ensure hard links can be created
+        except Exception as ex:
+            msg = (
+                "File system does not seem to support either symlinks or hard links. "
+                "Strict editable installs require one of them to be supported."
+            )
+            raise LinksNotSupported(msg) from ex
+        return False
+
+
+def _simple_layout(
+    packages: Iterable[str], package_dir: dict[str, str], project_dir: StrPath
+) -> bool:
+    """Return ``True`` if:
+    - all packages are contained by the same parent directory, **and**
+    - all packages become importable if the parent directory is added to ``sys.path``.
+
+    >>> _simple_layout(['a'], {"": "src"}, "/tmp/myproj")
+    True
+    >>> _simple_layout(['a', 'a.b'], {"": "src"}, "/tmp/myproj")
+    True
+    >>> _simple_layout(['a', 'a.b'], {}, "/tmp/myproj")
+    True
+    >>> _simple_layout(['a', 'a.a1', 'a.a1.a2', 'b'], {"": "src"}, "/tmp/myproj")
+    True
+    >>> _simple_layout(['a', 'a.a1', 'a.a1.a2', 'b'], {"a": "a", "b": "b"}, ".")
+    True
+    >>> _simple_layout(['a', 'a.a1', 'a.a1.a2', 'b'], {"a": "_a", "b": "_b"}, ".")
+    False
+    >>> _simple_layout(['a', 'a.a1', 'a.a1.a2', 'b'], {"a": "_a"}, "/tmp/myproj")
+    False
+    >>> _simple_layout(['a', 'a.a1', 'a.a1.a2', 'b'], {"a.a1.a2": "_a2"}, ".")
+    False
+    >>> _simple_layout(['a', 'a.b'], {"": "src", "a.b": "_ab"}, "/tmp/myproj")
+    False
+    >>> # Special cases, no packages yet:
+    >>> _simple_layout([], {"": "src"}, "/tmp/myproj")
+    True
+    >>> _simple_layout([], {"a": "_a", "": "src"}, "/tmp/myproj")
+    False
+    """
+    layout = {pkg: find_package_path(pkg, package_dir, project_dir) for pkg in packages}
+    if not layout:
+        return set(package_dir) in ({}, {""})
+    parent = os.path.commonpath(starmap(_parent_path, layout.items()))
+    return all(
+        _path.same_path(Path(parent, *key.split('.')), value)
+        for key, value in layout.items()
+    )
+
+
+def _parent_path(pkg, pkg_path):
+    """Infer the parent path containing a package, that if added to ``sys.path`` would
+    allow importing that package.
+    When ``pkg`` is directly mapped into a directory with a different name, return its
+    own path.
+    >>> _parent_path("a", "src/a")
+    'src'
+    >>> _parent_path("b", "src/c")
+    'src/c'
+    """
+    parent = pkg_path[: -len(pkg)] if pkg_path.endswith(pkg) else pkg_path
+    return parent.rstrip("/" + os.sep)
+
+
+def _find_packages(dist: Distribution) -> Iterator[str]:
+    yield from iter(dist.packages or [])
+
+    py_modules = dist.py_modules or []
+    nested_modules = [mod for mod in py_modules if "." in mod]
+    if dist.ext_package:
+        yield dist.ext_package
+    else:
+        ext_modules = dist.ext_modules or []
+        nested_modules += [x.name for x in ext_modules if "." in x.name]
+
+    for module in nested_modules:
+        package, _, _ = module.rpartition(".")
+        yield package
+
+
+def _find_top_level_modules(dist: Distribution) -> Iterator[str]:
+    py_modules = dist.py_modules or []
+    yield from (mod for mod in py_modules if "." not in mod)
+
+    if not dist.ext_package:
+        ext_modules = dist.ext_modules or []
+        yield from (x.name for x in ext_modules if "." not in x.name)
+
+
+def _find_package_roots(
+    packages: Iterable[str],
+    package_dir: Mapping[str, str],
+    src_root: StrPath,
+) -> dict[str, str]:
+    pkg_roots: dict[str, str] = {
+        pkg: _absolute_root(find_package_path(pkg, package_dir, src_root))
+        for pkg in sorted(packages)
+    }
+
+    return _remove_nested(pkg_roots)
+
+
+def _absolute_root(path: StrPath) -> str:
+    """Works for packages and top-level modules"""
+    path_ = Path(path)
+    parent = path_.parent
+
+    if path_.exists():
+        return str(path_.resolve())
+    else:
+        return str(parent.resolve() / path_.name)
+
+
+def _find_virtual_namespaces(pkg_roots: dict[str, str]) -> Iterator[str]:
+    """By carefully designing ``package_dir``, it is possible to implement the logical
+    structure of PEP 420 in a package without the corresponding directories.
+
+    Moreover a parent package can be purposefully/accidentally skipped in the discovery
+    phase (e.g. ``find_packages(include=["mypkg.*"])``, when ``mypkg.foo`` is included
+    by ``mypkg`` itself is not).
+    We consider this case to also be a virtual namespace (ignoring the original
+    directory) to emulate a non-editable installation.
+
+    This function will try to find these kinds of namespaces.
+    """
+    for pkg in pkg_roots:
+        if "." not in pkg:
+            continue
+        parts = pkg.split(".")
+        for i in range(len(parts) - 1, 0, -1):
+            partial_name = ".".join(parts[:i])
+            path = Path(find_package_path(partial_name, pkg_roots, ""))
+            if not path.exists() or partial_name not in pkg_roots:
+                # partial_name not in pkg_roots ==> purposefully/accidentally skipped
+                yield partial_name
+
+
+def _find_namespaces(
+    packages: list[str], pkg_roots: dict[str, str]
+) -> Iterator[tuple[str, list[str]]]:
+    for pkg in packages:
+        path = find_package_path(pkg, pkg_roots, "")
+        if Path(path).exists() and not Path(path, "__init__.py").exists():
+            yield (pkg, [path])
+
+
+def _remove_nested(pkg_roots: dict[str, str]) -> dict[str, str]:
+    output = dict(pkg_roots.copy())
+
+    for pkg, path in reversed(list(pkg_roots.items())):
+        if any(
+            pkg != other and _is_nested(pkg, path, other, other_path)
+            for other, other_path in pkg_roots.items()
+        ):
+            output.pop(pkg)
+
+    return output
+
+
+def _is_nested(pkg: str, pkg_path: str, parent: str, parent_path: str) -> bool:
+    """
+    Return ``True`` if ``pkg`` is nested inside ``parent`` both logically and in the
+    file system.
+    >>> _is_nested("a.b", "path/a/b", "a", "path/a")
+    True
+    >>> _is_nested("a.b", "path/a/b", "a", "otherpath/a")
+    False
+    >>> _is_nested("a.b", "path/a/b", "c", "path/c")
+    False
+    >>> _is_nested("a.a", "path/a/a", "a", "path/a")
+    True
+    >>> _is_nested("b.a", "path/b/a", "a", "path/a")
+    False
+    """
+    norm_pkg_path = _path.normpath(pkg_path)
+    rest = pkg.replace(parent, "", 1).strip(".").split(".")
+    return pkg.startswith(parent) and norm_pkg_path == _path.normpath(
+        Path(parent_path, *rest)
+    )
+
+
+def _empty_dir(dir_: _P) -> _P:
+    """Create a directory ensured to be empty. Existing files may be removed."""
+    _shutil.rmtree(dir_, ignore_errors=True)
+    os.makedirs(dir_)
+    return dir_
+
+
+class _NamespaceInstaller(namespaces.Installer):
+    def __init__(self, distribution, installation_dir, editable_name, src_root) -> None:
+        self.distribution = distribution
+        self.src_root = src_root
+        self.installation_dir = installation_dir
+        self.editable_name = editable_name
+        self.outputs: list[str] = []
+        self.dry_run = False
+
+    def _get_nspkg_file(self):
+        """Installation target."""
+        return os.path.join(self.installation_dir, self.editable_name + self.nspkg_ext)
+
+    def _get_root(self):
+        """Where the modules/packages should be loaded from."""
+        return repr(str(self.src_root))
+
+
+_FINDER_TEMPLATE = """\
+from __future__ import annotations
+import sys
+from importlib.machinery import ModuleSpec, PathFinder
+from importlib.machinery import all_suffixes as module_suffixes
+from importlib.util import spec_from_file_location
+from itertools import chain
+from pathlib import Path
+
+MAPPING: dict[str, str] = {mapping!r}
+NAMESPACES: dict[str, list[str]] = {namespaces!r}
+PATH_PLACEHOLDER = {name!r} + ".__path_hook__"
+
+
+class _EditableFinder:  # MetaPathFinder
+    @classmethod
+    def find_spec(cls, fullname: str, path=None, target=None) -> ModuleSpec | None:  # type: ignore
+        # Top-level packages and modules (we know these exist in the FS)
+        if fullname in MAPPING:
+            pkg_path = MAPPING[fullname]
+            return cls._find_spec(fullname, Path(pkg_path))
+
+        # Handle immediate children modules (required for namespaces to work)
+        # To avoid problems with case sensitivity in the file system we delegate
+        # to the importlib.machinery implementation.
+        parent, _, child = fullname.rpartition(".")
+        if parent and parent in MAPPING:
+            return PathFinder.find_spec(fullname, path=[MAPPING[parent]])
+
+        # Other levels of nesting should be handled automatically by importlib
+        # using the parent path.
+        return None
+
+    @classmethod
+    def _find_spec(cls, fullname: str, candidate_path: Path) -> ModuleSpec | None:
+        init = candidate_path / "__init__.py"
+        candidates = (candidate_path.with_suffix(x) for x in module_suffixes())
+        for candidate in chain([init], candidates):
+            if candidate.exists():
+                return spec_from_file_location(fullname, candidate)
+        return None
+
+
+class _EditableNamespaceFinder:  # PathEntryFinder
+    @classmethod
+    def _path_hook(cls, path) -> type[_EditableNamespaceFinder]:
+        if path == PATH_PLACEHOLDER:
+            return cls
+        raise ImportError
+
+    @classmethod
+    def _paths(cls, fullname: str) -> list[str]:
+        paths = NAMESPACES[fullname]
+        if not paths and fullname in MAPPING:
+            paths = [MAPPING[fullname]]
+        # Always add placeholder, for 2 reasons:
+        # 1. __path__ cannot be empty for the spec to be considered namespace.
+        # 2. In the case of nested namespaces, we need to force
+        #    import machinery to query _EditableNamespaceFinder again.
+        return [*paths, PATH_PLACEHOLDER]
+
+    @classmethod
+    def find_spec(cls, fullname: str, target=None) -> ModuleSpec | None:  # type: ignore
+        if fullname in NAMESPACES:
+            spec = ModuleSpec(fullname, None, is_package=True)
+            spec.submodule_search_locations = cls._paths(fullname)
+            return spec
+        return None
+
+    @classmethod
+    def find_module(cls, _fullname) -> None:
+        return None
+
+
+def install():
+    if not any(finder == _EditableFinder for finder in sys.meta_path):
+        sys.meta_path.append(_EditableFinder)
+
+    if not NAMESPACES:
+        return
+
+    if not any(hook == _EditableNamespaceFinder._path_hook for hook in sys.path_hooks):
+        # PathEntryFinder is needed to create NamespaceSpec without private APIS
+        sys.path_hooks.append(_EditableNamespaceFinder._path_hook)
+    if PATH_PLACEHOLDER not in sys.path:
+        sys.path.append(PATH_PLACEHOLDER)  # Used just to trigger the path hook
+"""
+
+
+def _finder_template(
+    name: str, mapping: Mapping[str, str], namespaces: dict[str, list[str]]
+) -> str:
+    """Create a string containing the code for the``MetaPathFinder`` and
+    ``PathEntryFinder``.
+    """
+    mapping = dict(sorted(mapping.items(), key=lambda p: p[0]))
+    return _FINDER_TEMPLATE.format(name=name, mapping=mapping, namespaces=namespaces)
+
+
+class LinksNotSupported(errors.FileError):
+    """File system does not seem to support either symlinks or hard links."""
+
+
+class _DebuggingTips(SetuptoolsWarning):
+    _SUMMARY = "Problem in editable installation."
+    _DETAILS = """
+    An error happened while installing `{project}` in editable mode.
+
+    The following steps are recommended to help debug this problem:
+
+    - Try to install the project normally, without using the editable mode.
+      Does the error still persist?
+      (If it does, try fixing the problem before attempting the editable mode).
+    - If you are using binary extensions, make sure you have all OS-level
+      dependencies installed (e.g. compilers, toolchains, binary libraries, ...).
+    - Try the latest version of setuptools (maybe the error was already fixed).
+    - If you (or your project dependencies) are using any setuptools extension
+      or customization, make sure they support the editable mode.
+
+    After following the steps above, if the problem still persists and
+    you think this is related to how setuptools handles editable installations,
+    please submit a reproducible example
+    (see https://stackoverflow.com/help/minimal-reproducible-example) to:
+
+        https://github.com/pypa/setuptools/issues
+    """
+    _SEE_DOCS = "userguide/development_mode.html"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/egg_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/egg_info.py
new file mode 100644
index 00000000..f7763116
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/egg_info.py
@@ -0,0 +1,720 @@
+"""setuptools.command.egg_info
+
+Create a distribution's .egg-info directory and contents"""
+
+import functools
+import os
+import re
+import sys
+import time
+from collections.abc import Callable
+
+import packaging
+import packaging.requirements
+import packaging.version
+
+import setuptools.unicode_utils as unicode_utils
+from setuptools import Command
+from setuptools.command import bdist_egg
+from setuptools.command.sdist import sdist, walk_revctrl
+from setuptools.command.setopt import edit_config
+from setuptools.glob import glob
+
+from .. import _entry_points, _normalization
+from .._importlib import metadata
+from ..warnings import SetuptoolsDeprecationWarning
+from . import _requirestxt
+
+import distutils.errors
+import distutils.filelist
+from distutils import log
+from distutils.errors import DistutilsInternalError
+from distutils.filelist import FileList as _FileList
+from distutils.util import convert_path
+
+PY_MAJOR = f'{sys.version_info.major}.{sys.version_info.minor}'
+
+
+def translate_pattern(glob):  # noqa: C901  # is too complex (14)  # FIXME
+    """
+    Translate a file path glob like '*.txt' in to a regular expression.
+    This differs from fnmatch.translate which allows wildcards to match
+    directory separators. It also knows about '**/' which matches any number of
+    directories.
+    """
+    pat = ''
+
+    # This will split on '/' within [character classes]. This is deliberate.
+    chunks = glob.split(os.path.sep)
+
+    sep = re.escape(os.sep)
+    valid_char = f'[^{sep}]'
+
+    for c, chunk in enumerate(chunks):
+        last_chunk = c == len(chunks) - 1
+
+        # Chunks that are a literal ** are globstars. They match anything.
+        if chunk == '**':
+            if last_chunk:
+                # Match anything if this is the last component
+                pat += '.*'
+            else:
+                # Match '(name/)*'
+                pat += f'(?:{valid_char}+{sep})*'
+            continue  # Break here as the whole path component has been handled
+
+        # Find any special characters in the remainder
+        i = 0
+        chunk_len = len(chunk)
+        while i < chunk_len:
+            char = chunk[i]
+            if char == '*':
+                # Match any number of name characters
+                pat += valid_char + '*'
+            elif char == '?':
+                # Match a name character
+                pat += valid_char
+            elif char == '[':
+                # Character class
+                inner_i = i + 1
+                # Skip initial !/] chars
+                if inner_i < chunk_len and chunk[inner_i] == '!':
+                    inner_i = inner_i + 1
+                if inner_i < chunk_len and chunk[inner_i] == ']':
+                    inner_i = inner_i + 1
+
+                # Loop till the closing ] is found
+                while inner_i < chunk_len and chunk[inner_i] != ']':
+                    inner_i = inner_i + 1
+
+                if inner_i >= chunk_len:
+                    # Got to the end of the string without finding a closing ]
+                    # Do not treat this as a matching group, but as a literal [
+                    pat += re.escape(char)
+                else:
+                    # Grab the insides of the [brackets]
+                    inner = chunk[i + 1 : inner_i]
+                    char_class = ''
+
+                    # Class negation
+                    if inner[0] == '!':
+                        char_class = '^'
+                        inner = inner[1:]
+
+                    char_class += re.escape(inner)
+                    pat += f'[{char_class}]'
+
+                    # Skip to the end ]
+                    i = inner_i
+            else:
+                pat += re.escape(char)
+            i += 1
+
+        # Join each chunk with the dir separator
+        if not last_chunk:
+            pat += sep
+
+    pat += r'\Z'
+    return re.compile(pat, flags=re.MULTILINE | re.DOTALL)
+
+
+class InfoCommon:
+    tag_build = None
+    tag_date = None
+
+    @property
+    def name(self):
+        return _normalization.safe_name(self.distribution.get_name())
+
+    def tagged_version(self):
+        tagged = self._maybe_tag(self.distribution.get_version())
+        return _normalization.safe_version(tagged)
+
+    def _maybe_tag(self, version):
+        """
+        egg_info may be called more than once for a distribution,
+        in which case the version string already contains all tags.
+        """
+        return (
+            version
+            if self.vtags and self._already_tagged(version)
+            else version + self.vtags
+        )
+
+    def _already_tagged(self, version: str) -> bool:
+        # Depending on their format, tags may change with version normalization.
+        # So in addition the regular tags, we have to search for the normalized ones.
+        return version.endswith(self.vtags) or version.endswith(self._safe_tags())
+
+    def _safe_tags(self) -> str:
+        # To implement this we can rely on `safe_version` pretending to be version 0
+        # followed by tags. Then we simply discard the starting 0 (fake version number)
+        try:
+            return _normalization.safe_version(f"0{self.vtags}")[1:]
+        except packaging.version.InvalidVersion:
+            return _normalization.safe_name(self.vtags.replace(' ', '.'))
+
+    def tags(self) -> str:
+        version = ''
+        if self.tag_build:
+            version += self.tag_build
+        if self.tag_date:
+            version += time.strftime("%Y%m%d")
+        return version
+
+    vtags = property(tags)
+
+
+class egg_info(InfoCommon, Command):
+    description = "create a distribution's .egg-info directory"
+
+    user_options = [
+        (
+            'egg-base=',
+            'e',
+            "directory containing .egg-info directories"
+            " [default: top of the source tree]",
+        ),
+        ('tag-date', 'd', "Add date stamp (e.g. 20050528) to version number"),
+        ('tag-build=', 'b', "Specify explicit tag to add to version number"),
+        ('no-date', 'D', "Don't include date stamp [default]"),
+    ]
+
+    boolean_options = ['tag-date']
+    negative_opt = {
+        'no-date': 'tag-date',
+    }
+
+    def initialize_options(self):
+        self.egg_base = None
+        self.egg_name = None
+        self.egg_info = None
+        self.egg_version = None
+        self.ignore_egg_info_in_manifest = False
+
+    ####################################
+    # allow the 'tag_svn_revision' to be detected and
+    # set, supporting sdists built on older Setuptools.
+    @property
+    def tag_svn_revision(self) -> None:
+        pass
+
+    @tag_svn_revision.setter
+    def tag_svn_revision(self, value):
+        pass
+
+    ####################################
+
+    def save_version_info(self, filename) -> None:
+        """
+        Materialize the value of date into the
+        build tag. Install build keys in a deterministic order
+        to avoid arbitrary reordering on subsequent builds.
+        """
+        # follow the order these keys would have been added
+        # when PYTHONHASHSEED=0
+        egg_info = dict(tag_build=self.tags(), tag_date=0)
+        edit_config(filename, dict(egg_info=egg_info))
+
+    def finalize_options(self) -> None:
+        # Note: we need to capture the current value returned
+        # by `self.tagged_version()`, so we can later update
+        # `self.distribution.metadata.version` without
+        # repercussions.
+        self.egg_name = self.name
+        self.egg_version = self.tagged_version()
+        parsed_version = packaging.version.Version(self.egg_version)
+
+        try:
+            is_version = isinstance(parsed_version, packaging.version.Version)
+            spec = "%s==%s" if is_version else "%s===%s"
+            packaging.requirements.Requirement(spec % (self.egg_name, self.egg_version))
+        except ValueError as e:
+            raise distutils.errors.DistutilsOptionError(
+                f"Invalid distribution name or version syntax: {self.egg_name}-{self.egg_version}"
+            ) from e
+
+        if self.egg_base is None:
+            dirs = self.distribution.package_dir
+            self.egg_base = (dirs or {}).get('', os.curdir)
+
+        self.ensure_dirname('egg_base')
+        self.egg_info = _normalization.filename_component(self.egg_name) + '.egg-info'
+        if self.egg_base != os.curdir:
+            self.egg_info = os.path.join(self.egg_base, self.egg_info)
+
+        # Set package version for the benefit of dumber commands
+        # (e.g. sdist, bdist_wininst, etc.)
+        #
+        self.distribution.metadata.version = self.egg_version
+
+    def _get_egg_basename(self, py_version=PY_MAJOR, platform=None):
+        """Compute filename of the output egg. Private API."""
+        return _egg_basename(self.egg_name, self.egg_version, py_version, platform)
+
+    def write_or_delete_file(self, what, filename, data, force: bool = False) -> None:
+        """Write `data` to `filename` or delete if empty
+
+        If `data` is non-empty, this routine is the same as ``write_file()``.
+        If `data` is empty but not ``None``, this is the same as calling
+        ``delete_file(filename)`.  If `data` is ``None``, then this is a no-op
+        unless `filename` exists, in which case a warning is issued about the
+        orphaned file (if `force` is false), or deleted (if `force` is true).
+        """
+        if data:
+            self.write_file(what, filename, data)
+        elif os.path.exists(filename):
+            if data is None and not force:
+                log.warn("%s not set in setup(), but %s exists", what, filename)
+                return
+            else:
+                self.delete_file(filename)
+
+    def write_file(self, what, filename, data) -> None:
+        """Write `data` to `filename` (if not a dry run) after announcing it
+
+        `what` is used in a log message to identify what is being written
+        to the file.
+        """
+        log.info("writing %s to %s", what, filename)
+        data = data.encode("utf-8")
+        if not self.dry_run:
+            f = open(filename, 'wb')
+            f.write(data)
+            f.close()
+
+    def delete_file(self, filename) -> None:
+        """Delete `filename` (if not a dry run) after announcing it"""
+        log.info("deleting %s", filename)
+        if not self.dry_run:
+            os.unlink(filename)
+
+    def run(self) -> None:
+        # Pre-load to avoid iterating over entry-points while an empty .egg-info
+        # exists in sys.path. See pypa/pyproject-hooks#206
+        writers = list(metadata.entry_points(group='egg_info.writers'))
+
+        self.mkpath(self.egg_info)
+        try:
+            os.utime(self.egg_info, None)
+        except OSError as e:
+            msg = f"Cannot update time stamp of directory '{self.egg_info}'"
+            raise distutils.errors.DistutilsFileError(msg) from e
+        for ep in writers:
+            writer = ep.load()
+            writer(self, ep.name, os.path.join(self.egg_info, ep.name))
+
+        # Get rid of native_libs.txt if it was put there by older bdist_egg
+        nl = os.path.join(self.egg_info, "native_libs.txt")
+        if os.path.exists(nl):
+            self.delete_file(nl)
+
+        self.find_sources()
+
+    def find_sources(self) -> None:
+        """Generate SOURCES.txt manifest file"""
+        manifest_filename = os.path.join(self.egg_info, "SOURCES.txt")
+        mm = manifest_maker(self.distribution)
+        mm.ignore_egg_info_dir = self.ignore_egg_info_in_manifest
+        mm.manifest = manifest_filename
+        mm.run()
+        self.filelist = mm.filelist
+
+
+class FileList(_FileList):
+    # Implementations of the various MANIFEST.in commands
+
+    def __init__(
+        self, warn=None, debug_print=None, ignore_egg_info_dir: bool = False
+    ) -> None:
+        super().__init__(warn, debug_print)
+        self.ignore_egg_info_dir = ignore_egg_info_dir
+
+    def process_template_line(self, line) -> None:
+        # Parse the line: split it up, make sure the right number of words
+        # is there, and return the relevant words.  'action' is always
+        # defined: it's the first word of the line.  Which of the other
+        # three are defined depends on the action; it'll be either
+        # patterns, (dir and patterns), or (dir_pattern).
+        (action, patterns, dir, dir_pattern) = self._parse_template_line(line)
+
+        action_map: dict[str, Callable] = {
+            'include': self.include,
+            'exclude': self.exclude,
+            'global-include': self.global_include,
+            'global-exclude': self.global_exclude,
+            'recursive-include': functools.partial(
+                self.recursive_include,
+                dir,
+            ),
+            'recursive-exclude': functools.partial(
+                self.recursive_exclude,
+                dir,
+            ),
+            'graft': self.graft,
+            'prune': self.prune,
+        }
+        log_map = {
+            'include': "warning: no files found matching '%s'",
+            'exclude': ("warning: no previously-included files found matching '%s'"),
+            'global-include': (
+                "warning: no files found matching '%s' anywhere in distribution"
+            ),
+            'global-exclude': (
+                "warning: no previously-included files matching "
+                "'%s' found anywhere in distribution"
+            ),
+            'recursive-include': (
+                "warning: no files found matching '%s' under directory '%s'"
+            ),
+            'recursive-exclude': (
+                "warning: no previously-included files matching "
+                "'%s' found under directory '%s'"
+            ),
+            'graft': "warning: no directories found matching '%s'",
+            'prune': "no previously-included directories found matching '%s'",
+        }
+
+        try:
+            process_action = action_map[action]
+        except KeyError:
+            msg = f"Invalid MANIFEST.in: unknown action {action!r} in {line!r}"
+            raise DistutilsInternalError(msg) from None
+
+        # OK, now we know that the action is valid and we have the
+        # right number of words on the line for that action -- so we
+        # can proceed with minimal error-checking.
+
+        action_is_recursive = action.startswith('recursive-')
+        if action in {'graft', 'prune'}:
+            patterns = [dir_pattern]
+        extra_log_args = (dir,) if action_is_recursive else ()
+        log_tmpl = log_map[action]
+
+        self.debug_print(
+            ' '.join(
+                [action] + ([dir] if action_is_recursive else []) + patterns,
+            )
+        )
+        for pattern in patterns:
+            if not process_action(pattern):
+                log.warn(log_tmpl, pattern, *extra_log_args)
+
+    def _remove_files(self, predicate):
+        """
+        Remove all files from the file list that match the predicate.
+        Return True if any matching files were removed
+        """
+        found = False
+        for i in range(len(self.files) - 1, -1, -1):
+            if predicate(self.files[i]):
+                self.debug_print(" removing " + self.files[i])
+                del self.files[i]
+                found = True
+        return found
+
+    def include(self, pattern):
+        """Include files that match 'pattern'."""
+        found = [f for f in glob(pattern) if not os.path.isdir(f)]
+        self.extend(found)
+        return bool(found)
+
+    def exclude(self, pattern):
+        """Exclude files that match 'pattern'."""
+        match = translate_pattern(pattern)
+        return self._remove_files(match.match)
+
+    def recursive_include(self, dir, pattern):
+        """
+        Include all files anywhere in 'dir/' that match the pattern.
+        """
+        full_pattern = os.path.join(dir, '**', pattern)
+        found = [f for f in glob(full_pattern, recursive=True) if not os.path.isdir(f)]
+        self.extend(found)
+        return bool(found)
+
+    def recursive_exclude(self, dir, pattern):
+        """
+        Exclude any file anywhere in 'dir/' that match the pattern.
+        """
+        match = translate_pattern(os.path.join(dir, '**', pattern))
+        return self._remove_files(match.match)
+
+    def graft(self, dir):
+        """Include all files from 'dir/'."""
+        found = [
+            item
+            for match_dir in glob(dir)
+            for item in distutils.filelist.findall(match_dir)
+        ]
+        self.extend(found)
+        return bool(found)
+
+    def prune(self, dir):
+        """Filter out files from 'dir/'."""
+        match = translate_pattern(os.path.join(dir, '**'))
+        return self._remove_files(match.match)
+
+    def global_include(self, pattern):
+        """
+        Include all files anywhere in the current directory that match the
+        pattern. This is very inefficient on large file trees.
+        """
+        if self.allfiles is None:
+            self.findall()
+        match = translate_pattern(os.path.join('**', pattern))
+        found = [f for f in self.allfiles if match.match(f)]
+        self.extend(found)
+        return bool(found)
+
+    def global_exclude(self, pattern):
+        """
+        Exclude all files anywhere that match the pattern.
+        """
+        match = translate_pattern(os.path.join('**', pattern))
+        return self._remove_files(match.match)
+
+    def append(self, item) -> None:
+        if item.endswith('\r'):  # Fix older sdists built on Windows
+            item = item[:-1]
+        path = convert_path(item)
+
+        if self._safe_path(path):
+            self.files.append(path)
+
+    def extend(self, paths) -> None:
+        self.files.extend(filter(self._safe_path, paths))
+
+    def _repair(self):
+        """
+        Replace self.files with only safe paths
+
+        Because some owners of FileList manipulate the underlying
+        ``files`` attribute directly, this method must be called to
+        repair those paths.
+        """
+        self.files = list(filter(self._safe_path, self.files))
+
+    def _safe_path(self, path):
+        enc_warn = "'%s' not %s encodable -- skipping"
+
+        # To avoid accidental trans-codings errors, first to unicode
+        u_path = unicode_utils.filesys_decode(path)
+        if u_path is None:
+            log.warn(f"'{path}' in unexpected encoding -- skipping")
+            return False
+
+        # Must ensure utf-8 encodability
+        utf8_path = unicode_utils.try_encode(u_path, "utf-8")
+        if utf8_path is None:
+            log.warn(enc_warn, path, 'utf-8')
+            return False
+
+        try:
+            # ignore egg-info paths
+            is_egg_info = ".egg-info" in u_path or b".egg-info" in utf8_path
+            if self.ignore_egg_info_dir and is_egg_info:
+                return False
+            # accept is either way checks out
+            if os.path.exists(u_path) or os.path.exists(utf8_path):
+                return True
+        # this will catch any encode errors decoding u_path
+        except UnicodeEncodeError:
+            log.warn(enc_warn, path, sys.getfilesystemencoding())
+
+
+class manifest_maker(sdist):
+    template = "MANIFEST.in"
+
+    def initialize_options(self) -> None:
+        self.use_defaults = True
+        self.prune = True
+        self.manifest_only = True
+        self.force_manifest = True
+        self.ignore_egg_info_dir = False
+
+    def finalize_options(self) -> None:
+        pass
+
+    def run(self) -> None:
+        self.filelist = FileList(ignore_egg_info_dir=self.ignore_egg_info_dir)
+        if not os.path.exists(self.manifest):
+            self.write_manifest()  # it must exist so it'll get in the list
+        self.add_defaults()
+        if os.path.exists(self.template):
+            self.read_template()
+        self.add_license_files()
+        self._add_referenced_files()
+        self.prune_file_list()
+        self.filelist.sort()
+        self.filelist.remove_duplicates()
+        self.write_manifest()
+
+    def _manifest_normalize(self, path):
+        path = unicode_utils.filesys_decode(path)
+        return path.replace(os.sep, '/')
+
+    def write_manifest(self) -> None:
+        """
+        Write the file list in 'self.filelist' to the manifest file
+        named by 'self.manifest'.
+        """
+        self.filelist._repair()
+
+        # Now _repairs should encodability, but not unicode
+        files = [self._manifest_normalize(f) for f in self.filelist.files]
+        msg = f"writing manifest file '{self.manifest}'"
+        self.execute(write_file, (self.manifest, files), msg)
+
+    def warn(self, msg) -> None:
+        if not self._should_suppress_warning(msg):
+            sdist.warn(self, msg)
+
+    @staticmethod
+    def _should_suppress_warning(msg):
+        """
+        suppress missing-file warnings from sdist
+        """
+        return re.match(r"standard file .*not found", msg)
+
+    def add_defaults(self) -> None:
+        sdist.add_defaults(self)
+        self.filelist.append(self.template)
+        self.filelist.append(self.manifest)
+        rcfiles = list(walk_revctrl())
+        if rcfiles:
+            self.filelist.extend(rcfiles)
+        elif os.path.exists(self.manifest):
+            self.read_manifest()
+
+        if os.path.exists("setup.py"):
+            # setup.py should be included by default, even if it's not
+            # the script called to create the sdist
+            self.filelist.append("setup.py")
+
+        ei_cmd = self.get_finalized_command('egg_info')
+        self.filelist.graft(ei_cmd.egg_info)
+
+    def add_license_files(self) -> None:
+        license_files = self.distribution.metadata.license_files or []
+        for lf in license_files:
+            log.info("adding license file '%s'", lf)
+        self.filelist.extend(license_files)
+
+    def _add_referenced_files(self):
+        """Add files referenced by the config (e.g. `file:` directive) to filelist"""
+        referenced = getattr(self.distribution, '_referenced_files', [])
+        # ^-- fallback if dist comes from distutils or is a custom class
+        for rf in referenced:
+            log.debug("adding file referenced by config '%s'", rf)
+        self.filelist.extend(referenced)
+
+    def _safe_data_files(self, build_py):
+        """
+        The parent class implementation of this method
+        (``sdist``) will try to include data files, which
+        might cause recursion problems when
+        ``include_package_data=True``.
+
+        Therefore, avoid triggering any attempt of
+        analyzing/building the manifest again.
+        """
+        if hasattr(build_py, 'get_data_files_without_manifest'):
+            return build_py.get_data_files_without_manifest()
+
+        SetuptoolsDeprecationWarning.emit(
+            "`build_py` command does not inherit from setuptools' `build_py`.",
+            """
+            Custom 'build_py' does not implement 'get_data_files_without_manifest'.
+            Please extend command classes from setuptools instead of distutils.
+            """,
+            see_url="https://peps.python.org/pep-0632/",
+            # due_date not defined yet, old projects might still do it?
+        )
+        return build_py.get_data_files()
+
+
+def write_file(filename, contents) -> None:
+    """Create a file with the specified name and write 'contents' (a
+    sequence of strings without line terminators) to it.
+    """
+    contents = "\n".join(contents)
+
+    # assuming the contents has been vetted for utf-8 encoding
+    contents = contents.encode("utf-8")
+
+    with open(filename, "wb") as f:  # always write POSIX-style manifest
+        f.write(contents)
+
+
+def write_pkg_info(cmd, basename, filename) -> None:
+    log.info("writing %s", filename)
+    if not cmd.dry_run:
+        metadata = cmd.distribution.metadata
+        metadata.version, oldver = cmd.egg_version, metadata.version
+        metadata.name, oldname = cmd.egg_name, metadata.name
+
+        try:
+            # write unescaped data to PKG-INFO, so older pkg_resources
+            # can still parse it
+            metadata.write_pkg_info(cmd.egg_info)
+        finally:
+            metadata.name, metadata.version = oldname, oldver
+
+        safe = getattr(cmd.distribution, 'zip_safe', None)
+
+        bdist_egg.write_safety_flag(cmd.egg_info, safe)
+
+
+def warn_depends_obsolete(cmd, basename, filename) -> None:
+    """
+    Unused: left to avoid errors when updating (from source) from <= 67.8.
+    Old installations have a .dist-info directory with the entry-point
+    ``depends.txt = setuptools.command.egg_info:warn_depends_obsolete``.
+    This may trigger errors when running the first egg_info in build_meta.
+    TODO: Remove this function in a version sufficiently > 68.
+    """
+
+
+# Export API used in entry_points
+write_requirements = _requirestxt.write_requirements
+write_setup_requirements = _requirestxt.write_setup_requirements
+
+
+def write_toplevel_names(cmd, basename, filename) -> None:
+    pkgs = dict.fromkeys([
+        k.split('.', 1)[0] for k in cmd.distribution.iter_distribution_names()
+    ])
+    cmd.write_file("top-level names", filename, '\n'.join(sorted(pkgs)) + '\n')
+
+
+def overwrite_arg(cmd, basename, filename) -> None:
+    write_arg(cmd, basename, filename, True)
+
+
+def write_arg(cmd, basename, filename, force: bool = False) -> None:
+    argname = os.path.splitext(basename)[0]
+    value = getattr(cmd.distribution, argname, None)
+    if value is not None:
+        value = '\n'.join(value) + '\n'
+    cmd.write_or_delete_file(argname, filename, value, force)
+
+
+def write_entries(cmd, basename, filename) -> None:
+    eps = _entry_points.load(cmd.distribution.entry_points)
+    defn = _entry_points.render(eps)
+    cmd.write_or_delete_file('entry points', filename, defn, True)
+
+
+def _egg_basename(egg_name, egg_version, py_version=None, platform=None):
+    """Compute filename of the output egg. Private API."""
+    name = _normalization.filename_component(egg_name)
+    version = _normalization.filename_component(egg_version)
+    egg = f"{name}-{version}-py{py_version or PY_MAJOR}"
+    if platform:
+        egg += f"-{platform}"
+    return egg
+
+
+class EggInfoDeprecationWarning(SetuptoolsDeprecationWarning):
+    """Deprecated behavior warning for EggInfo, bypassing suppression."""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/install.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/install.py
new file mode 100644
index 00000000..741b140c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/install.py
@@ -0,0 +1,183 @@
+from __future__ import annotations
+
+import glob
+import inspect
+import platform
+from collections.abc import Callable
+from typing import TYPE_CHECKING, Any, ClassVar, cast
+
+import setuptools
+
+from ..dist import Distribution
+from ..warnings import SetuptoolsDeprecationWarning, SetuptoolsWarning
+from .bdist_egg import bdist_egg as bdist_egg_cls
+
+import distutils.command.install as orig
+from distutils.errors import DistutilsArgError
+
+if TYPE_CHECKING:
+    # This is only used for a type-cast, don't import at runtime or it'll cause deprecation warnings
+    from .easy_install import easy_install as easy_install_cls
+else:
+    easy_install_cls = None
+
+
+def __getattr__(name: str):  # pragma: no cover
+    if name == "_install":
+        SetuptoolsDeprecationWarning.emit(
+            "`setuptools.command._install` was an internal implementation detail "
+            + "that was left in for numpy<1.9 support.",
+            due_date=(2025, 5, 2),  # Originally added on 2024-11-01
+        )
+        return orig.install
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+
+class install(orig.install):
+    """Use easy_install to install the package, w/dependencies"""
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    user_options = orig.install.user_options + [
+        ('old-and-unmanageable', None, "Try not to use this!"),
+        (
+            'single-version-externally-managed',
+            None,
+            "used by system package builders to create 'flat' eggs",
+        ),
+    ]
+    boolean_options = orig.install.boolean_options + [
+        'old-and-unmanageable',
+        'single-version-externally-managed',
+    ]
+    # Type the same as distutils.command.install.install.sub_commands
+    # Must keep the second tuple item potentially None due to invariance
+    new_commands: ClassVar[list[tuple[str, Callable[[Any], bool] | None]]] = [
+        ('install_egg_info', lambda self: True),
+        ('install_scripts', lambda self: True),
+    ]
+    _nc = dict(new_commands)
+
+    def initialize_options(self):
+        SetuptoolsDeprecationWarning.emit(
+            "setup.py install is deprecated.",
+            """
+            Please avoid running ``setup.py`` directly.
+            Instead, use pypa/build, pypa/installer or other
+            standards-based tools.
+            """,
+            see_url="https://blog.ganssle.io/articles/2021/10/setup-py-deprecated.html",
+            # TODO: Document how to bootstrap setuptools without install
+            #       (e.g. by unziping the wheel file)
+            #       and then add a due_date to this warning.
+        )
+
+        super().initialize_options()
+        self.old_and_unmanageable = None
+        self.single_version_externally_managed = None
+
+    def finalize_options(self) -> None:
+        super().finalize_options()
+        if self.root:
+            self.single_version_externally_managed = True
+        elif self.single_version_externally_managed:
+            if not self.root and not self.record:
+                raise DistutilsArgError(
+                    "You must specify --record or --root when building system packages"
+                )
+
+    def handle_extra_path(self):
+        if self.root or self.single_version_externally_managed:
+            # explicit backward-compatibility mode, allow extra_path to work
+            return orig.install.handle_extra_path(self)
+
+        # Ignore extra_path when installing an egg (or being run by another
+        # command without --root or --single-version-externally-managed
+        self.path_file = None
+        self.extra_dirs = ''
+        return None
+
+    def run(self):
+        # Explicit request for old-style install?  Just do it
+        if self.old_and_unmanageable or self.single_version_externally_managed:
+            return super().run()
+
+        if not self._called_from_setup(inspect.currentframe()):
+            # Run in backward-compatibility mode to support bdist_* commands.
+            super().run()
+        else:
+            self.do_egg_install()
+
+        return None
+
+    @staticmethod
+    def _called_from_setup(run_frame):
+        """
+        Attempt to detect whether run() was called from setup() or by another
+        command.  If called by setup(), the parent caller will be the
+        'run_command' method in 'distutils.dist', and *its* caller will be
+        the 'run_commands' method.  If called any other way, the
+        immediate caller *might* be 'run_command', but it won't have been
+        called by 'run_commands'. Return True in that case or if a call stack
+        is unavailable. Return False otherwise.
+        """
+        if run_frame is None:
+            msg = "Call stack not available. bdist_* commands may fail."
+            SetuptoolsWarning.emit(msg)
+            if platform.python_implementation() == 'IronPython':
+                msg = "For best results, pass -X:Frames to enable call stack."
+                SetuptoolsWarning.emit(msg)
+            return True
+
+        frames = inspect.getouterframes(run_frame)
+        for frame in frames[2:4]:
+            (caller,) = frame[:1]
+            info = inspect.getframeinfo(caller)
+            caller_module = caller.f_globals.get('__name__', '')
+
+            if caller_module == "setuptools.dist" and info.function == "run_command":
+                # Starting from v61.0.0 setuptools overwrites dist.run_command
+                continue
+
+            return caller_module == 'distutils.dist' and info.function == 'run_commands'
+
+        return False
+
+    def do_egg_install(self) -> None:
+        easy_install = self.distribution.get_command_class('easy_install')
+
+        cmd = cast(
+            # We'd want to cast easy_install as type[easy_install_cls] but a bug in
+            # mypy makes it think easy_install() returns a Command on Python 3.12+
+            # https://github.com/python/mypy/issues/18088
+            easy_install_cls,
+            easy_install(  # type: ignore[call-arg]
+                self.distribution,
+                args="x",
+                root=self.root,
+                record=self.record,
+            ),
+        )
+        cmd.ensure_finalized()  # finalize before bdist_egg munges install cmd
+        cmd.always_copy_from = '.'  # make sure local-dir eggs get installed
+
+        # pick up setup-dir .egg files only: no .egg-info
+        cmd.package_index.scan(glob.glob('*.egg'))
+
+        self.run_command('bdist_egg')
+        bdist_egg = cast(bdist_egg_cls, self.distribution.get_command_obj('bdist_egg'))
+        args = [bdist_egg.egg_output]
+
+        if setuptools.bootstrap_install_from:
+            # Bootstrap self-installation of setuptools
+            args.insert(0, setuptools.bootstrap_install_from)
+
+        cmd.args = args
+        cmd.run(show_deprecation=False)
+        setuptools.bootstrap_install_from = None
+
+
+# XXX Python 3.1 doesn't see _nc if this is inside the class
+install.sub_commands = [
+    cmd for cmd in orig.install.sub_commands if cmd[0] not in install._nc
+] + install.new_commands
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/install_egg_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_egg_info.py
new file mode 100644
index 00000000..44f22ccf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_egg_info.py
@@ -0,0 +1,58 @@
+import os
+
+from setuptools import Command, namespaces
+from setuptools.archive_util import unpack_archive
+
+from .._path import ensure_directory
+
+from distutils import dir_util, log
+
+
+class install_egg_info(namespaces.Installer, Command):
+    """Install an .egg-info directory for the package"""
+
+    description = "Install an .egg-info directory for the package"
+
+    user_options = [
+        ('install-dir=', 'd', "directory to install to"),
+    ]
+
+    def initialize_options(self):
+        self.install_dir = None
+
+    def finalize_options(self) -> None:
+        self.set_undefined_options('install_lib', ('install_dir', 'install_dir'))
+        ei_cmd = self.get_finalized_command("egg_info")
+        basename = f"{ei_cmd._get_egg_basename()}.egg-info"
+        self.source = ei_cmd.egg_info
+        self.target = os.path.join(self.install_dir, basename)
+        self.outputs: list[str] = []
+
+    def run(self) -> None:
+        self.run_command('egg_info')
+        if os.path.isdir(self.target) and not os.path.islink(self.target):
+            dir_util.remove_tree(self.target, dry_run=self.dry_run)
+        elif os.path.exists(self.target):
+            self.execute(os.unlink, (self.target,), "Removing " + self.target)
+        if not self.dry_run:
+            ensure_directory(self.target)
+        self.execute(self.copytree, (), f"Copying {self.source} to {self.target}")
+        self.install_namespaces()
+
+    def get_outputs(self):
+        return self.outputs
+
+    def copytree(self) -> None:
+        # Copy the .egg-info tree to site-packages
+        def skimmer(src, dst):
+            # filter out source-control directories; note that 'src' is always
+            # a '/'-separated path, regardless of platform.  'dst' is a
+            # platform-specific path.
+            for skip in '.svn/', 'CVS/':
+                if src.startswith(skip) or '/' + skip in src:
+                    return None
+            self.outputs.append(dst)
+            log.debug("Copying %s to %s", src, dst)
+            return dst
+
+        unpack_archive(self.source, self.target, skimmer)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/install_lib.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_lib.py
new file mode 100644
index 00000000..8e1e0727
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_lib.py
@@ -0,0 +1,137 @@
+from __future__ import annotations
+
+import os
+import sys
+from itertools import product, starmap
+
+from .._path import StrPath
+from ..dist import Distribution
+
+import distutils.command.install_lib as orig
+
+
+class install_lib(orig.install_lib):
+    """Don't add compiled flags to filenames of non-Python files"""
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    def run(self) -> None:
+        self.build()
+        outfiles = self.install()
+        if outfiles is not None:
+            # always compile, in case we have any extension stubs to deal with
+            self.byte_compile(outfiles)
+
+    def get_exclusions(self):
+        """
+        Return a collections.Sized collections.Container of paths to be
+        excluded for single_version_externally_managed installations.
+        """
+        all_packages = (
+            pkg
+            for ns_pkg in self._get_SVEM_NSPs()
+            for pkg in self._all_packages(ns_pkg)
+        )
+
+        excl_specs = product(all_packages, self._gen_exclusion_paths())
+        return set(starmap(self._exclude_pkg_path, excl_specs))
+
+    def _exclude_pkg_path(self, pkg, exclusion_path):
+        """
+        Given a package name and exclusion path within that package,
+        compute the full exclusion path.
+        """
+        parts = pkg.split('.') + [exclusion_path]
+        return os.path.join(self.install_dir, *parts)
+
+    @staticmethod
+    def _all_packages(pkg_name):
+        """
+        >>> list(install_lib._all_packages('foo.bar.baz'))
+        ['foo.bar.baz', 'foo.bar', 'foo']
+        """
+        while pkg_name:
+            yield pkg_name
+            pkg_name, _sep, _child = pkg_name.rpartition('.')
+
+    def _get_SVEM_NSPs(self):
+        """
+        Get namespace packages (list) but only for
+        single_version_externally_managed installations and empty otherwise.
+        """
+        # TODO: is it necessary to short-circuit here? i.e. what's the cost
+        # if get_finalized_command is called even when namespace_packages is
+        # False?
+        if not self.distribution.namespace_packages:
+            return []
+
+        install_cmd = self.get_finalized_command('install')
+        svem = install_cmd.single_version_externally_managed
+
+        return self.distribution.namespace_packages if svem else []
+
+    @staticmethod
+    def _gen_exclusion_paths():
+        """
+        Generate file paths to be excluded for namespace packages (bytecode
+        cache files).
+        """
+        # always exclude the package module itself
+        yield '__init__.py'
+
+        yield '__init__.pyc'
+        yield '__init__.pyo'
+
+        if not hasattr(sys, 'implementation'):
+            return
+
+        base = os.path.join('__pycache__', '__init__.' + sys.implementation.cache_tag)
+        yield base + '.pyc'
+        yield base + '.pyo'
+        yield base + '.opt-1.pyc'
+        yield base + '.opt-2.pyc'
+
+    def copy_tree(
+        self,
+        infile: StrPath,
+        outfile: str,
+        # override: Using actual booleans
+        preserve_mode: bool = True,  # type: ignore[override]
+        preserve_times: bool = True,  # type: ignore[override]
+        preserve_symlinks: bool = False,  # type: ignore[override]
+        level: object = 1,
+    ) -> list[str]:
+        assert preserve_mode
+        assert preserve_times
+        assert not preserve_symlinks
+        exclude = self.get_exclusions()
+
+        if not exclude:
+            return orig.install_lib.copy_tree(self, infile, outfile)
+
+        # Exclude namespace package __init__.py* files from the output
+
+        from setuptools.archive_util import unpack_directory
+
+        from distutils import log
+
+        outfiles: list[str] = []
+
+        def pf(src: str, dst: str):
+            if dst in exclude:
+                log.warn("Skipping installation of %s (namespace package)", dst)
+                return False
+
+            log.info("copying %s -> %s", src, os.path.dirname(dst))
+            outfiles.append(dst)
+            return dst
+
+        unpack_directory(infile, outfile, pf)
+        return outfiles
+
+    def get_outputs(self):
+        outputs = orig.install_lib.get_outputs(self)
+        exclude = self.get_exclusions()
+        if exclude:
+            return [f for f in outputs if f not in exclude]
+        return outputs
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/install_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_scripts.py
new file mode 100644
index 00000000..4401cf69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/install_scripts.py
@@ -0,0 +1,73 @@
+from __future__ import annotations
+
+import os
+import sys
+
+from .._path import ensure_directory
+from ..dist import Distribution
+
+import distutils.command.install_scripts as orig
+from distutils import log
+
+
+class install_scripts(orig.install_scripts):
+    """Do normal script install, plus any egg_info wrapper scripts"""
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+
+    def initialize_options(self) -> None:
+        orig.install_scripts.initialize_options(self)
+        self.no_ep = False
+
+    def run(self) -> None:
+        self.run_command("egg_info")
+        if self.distribution.scripts:
+            orig.install_scripts.run(self)  # run first to set up self.outfiles
+        else:
+            self.outfiles: list[str] = []
+        if self.no_ep:
+            # don't install entry point scripts into .egg file!
+            return
+        self._install_ep_scripts()
+
+    def _install_ep_scripts(self):
+        # Delay import side-effects
+        from pkg_resources import Distribution, PathMetadata
+
+        from . import easy_install as ei
+
+        ei_cmd = self.get_finalized_command("egg_info")
+        dist = Distribution(
+            ei_cmd.egg_base,
+            PathMetadata(ei_cmd.egg_base, ei_cmd.egg_info),
+            ei_cmd.egg_name,
+            ei_cmd.egg_version,
+        )
+        bs_cmd = self.get_finalized_command('build_scripts')
+        exec_param = getattr(bs_cmd, 'executable', None)
+        writer = ei.ScriptWriter
+        if exec_param == sys.executable:
+            # In case the path to the Python executable contains a space, wrap
+            # it so it's not split up.
+            exec_param = [exec_param]
+        # resolve the writer to the environment
+        writer = writer.best()
+        cmd = writer.command_spec_class.best().from_param(exec_param)
+        for args in writer.get_args(dist, cmd.as_header()):
+            self.write_script(*args)
+
+    def write_script(self, script_name, contents, mode: str = "t", *ignored) -> None:
+        """Write an executable file to the scripts directory"""
+        from setuptools.command.easy_install import chmod, current_umask
+
+        log.info("Installing %s script to %s", script_name, self.install_dir)
+        target = os.path.join(self.install_dir, script_name)
+        self.outfiles.append(target)
+
+        encoding = None if "b" in mode else "utf-8"
+        mask = current_umask()
+        if not self.dry_run:
+            ensure_directory(target)
+            with open(target, "w" + mode, encoding=encoding) as f:
+                f.write(contents)
+            chmod(target, 0o777 - mask)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/launcher manifest.xml b/backend/venv/lib/python3.12/site-packages/setuptools/command/launcher manifest.xml
new file mode 100644
index 00000000..5972a96d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/launcher manifest.xml	
@@ -0,0 +1,15 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">
+    <assemblyIdentity version="1.0.0.0"
+                      processorArchitecture="X86"
+                      name="%(name)s"
+                      type="win32"/>
+    <!-- Identify the application security requirements. -->
+    <trustInfo xmlns="urn:schemas-microsoft-com:asm.v3">
+        <security>
+            <requestedPrivileges>
+                <requestedExecutionLevel level="asInvoker" uiAccess="false"/>
+            </requestedPrivileges>
+        </security>
+    </trustInfo>
+</assembly>
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/rotate.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/rotate.py
new file mode 100644
index 00000000..acdce07b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/rotate.py
@@ -0,0 +1,65 @@
+from __future__ import annotations
+
+import os
+from typing import ClassVar
+
+from .. import Command, _shutil
+
+from distutils import log
+from distutils.errors import DistutilsOptionError
+from distutils.util import convert_path
+
+
+class rotate(Command):
+    """Delete older distributions"""
+
+    description = "delete older distributions, keeping N newest files"
+    user_options = [
+        ('match=', 'm', "patterns to match (required)"),
+        ('dist-dir=', 'd', "directory where the distributions are"),
+        ('keep=', 'k', "number of matching distributions to keep"),
+    ]
+
+    boolean_options: ClassVar[list[str]] = []
+
+    def initialize_options(self):
+        self.match = None
+        self.dist_dir = None
+        self.keep = None
+
+    def finalize_options(self) -> None:
+        if self.match is None:
+            raise DistutilsOptionError(
+                "Must specify one or more (comma-separated) match patterns "
+                "(e.g. '.zip' or '.egg')"
+            )
+        if self.keep is None:
+            raise DistutilsOptionError("Must specify number of files to keep")
+        try:
+            self.keep = int(self.keep)
+        except ValueError as e:
+            raise DistutilsOptionError("--keep must be an integer") from e
+        if isinstance(self.match, str):
+            self.match = [convert_path(p.strip()) for p in self.match.split(',')]
+        self.set_undefined_options('bdist', ('dist_dir', 'dist_dir'))
+
+    def run(self) -> None:
+        self.run_command("egg_info")
+        from glob import glob
+
+        for pattern in self.match:
+            pattern = self.distribution.get_name() + '*' + pattern
+            files = glob(os.path.join(self.dist_dir, pattern))
+            files = [(os.path.getmtime(f), f) for f in files]
+            files.sort()
+            files.reverse()
+
+            log.info("%d file(s) matching %s", len(files), pattern)
+            files = files[self.keep :]
+            for t, f in files:
+                log.info("Deleting %s", f)
+                if not self.dry_run:
+                    if os.path.isdir(f):
+                        _shutil.rmtree(f)
+                    else:
+                        os.unlink(f)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/saveopts.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/saveopts.py
new file mode 100644
index 00000000..2a2cbce6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/saveopts.py
@@ -0,0 +1,21 @@
+from setuptools.command.setopt import edit_config, option_base
+
+
+class saveopts(option_base):
+    """Save command-line options to a file"""
+
+    description = "save supplied options to setup.cfg or other config file"
+
+    def run(self) -> None:
+        dist = self.distribution
+        settings: dict[str, dict[str, str]] = {}
+
+        for cmd in dist.command_options:
+            if cmd == 'saveopts':
+                continue  # don't save our own options!
+
+            for opt, (src, val) in dist.get_option_dict(cmd).items():
+                if src == "command line":
+                    settings.setdefault(cmd, {})[opt] = val
+
+        edit_config(self.filename, settings, self.dry_run)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/sdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/sdist.py
new file mode 100644
index 00000000..9631cf31
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/sdist.py
@@ -0,0 +1,217 @@
+from __future__ import annotations
+
+import contextlib
+import os
+import re
+from itertools import chain
+from typing import ClassVar
+
+from .._importlib import metadata
+from ..dist import Distribution
+from .build import _ORIGINAL_SUBCOMMANDS
+
+import distutils.command.sdist as orig
+from distutils import log
+
+_default_revctrl = list
+
+
+def walk_revctrl(dirname=''):
+    """Find all files under revision control"""
+    for ep in metadata.entry_points(group='setuptools.file_finders'):
+        yield from ep.load()(dirname)
+
+
+class sdist(orig.sdist):
+    """Smart sdist that finds anything supported by revision control"""
+
+    user_options = [
+        ('formats=', None, "formats for source distribution (comma-separated list)"),
+        (
+            'keep-temp',
+            'k',
+            "keep the distribution tree around after creating " + "archive file(s)",
+        ),
+        (
+            'dist-dir=',
+            'd',
+            "directory to put the source distribution archive(s) in [default: dist]",
+        ),
+        (
+            'owner=',
+            'u',
+            "Owner name used when creating a tar file [default: current user]",
+        ),
+        (
+            'group=',
+            'g',
+            "Group name used when creating a tar file [default: current group]",
+        ),
+    ]
+
+    distribution: Distribution  # override distutils.dist.Distribution with setuptools.dist.Distribution
+    negative_opt: ClassVar[dict[str, str]] = {}
+
+    README_EXTENSIONS = ['', '.rst', '.txt', '.md']
+    READMES = tuple(f'README{ext}' for ext in README_EXTENSIONS)
+
+    def run(self) -> None:
+        self.run_command('egg_info')
+        ei_cmd = self.get_finalized_command('egg_info')
+        self.filelist = ei_cmd.filelist
+        self.filelist.append(os.path.join(ei_cmd.egg_info, 'SOURCES.txt'))
+        self.check_readme()
+
+        # Run sub commands
+        for cmd_name in self.get_sub_commands():
+            self.run_command(cmd_name)
+
+        self.make_distribution()
+
+        dist_files = getattr(self.distribution, 'dist_files', [])
+        for file in self.archive_files:
+            data = ('sdist', '', file)
+            if data not in dist_files:
+                dist_files.append(data)
+
+    def initialize_options(self) -> None:
+        orig.sdist.initialize_options(self)
+
+    def make_distribution(self) -> None:
+        """
+        Workaround for #516
+        """
+        with self._remove_os_link():
+            orig.sdist.make_distribution(self)
+
+    @staticmethod
+    @contextlib.contextmanager
+    def _remove_os_link():
+        """
+        In a context, remove and restore os.link if it exists
+        """
+
+        class NoValue:
+            pass
+
+        orig_val = getattr(os, 'link', NoValue)
+        try:
+            del os.link
+        except Exception:
+            pass
+        try:
+            yield
+        finally:
+            if orig_val is not NoValue:
+                os.link = orig_val
+
+    def add_defaults(self) -> None:
+        super().add_defaults()
+        self._add_defaults_build_sub_commands()
+
+    def _add_defaults_optional(self):
+        super()._add_defaults_optional()
+        if os.path.isfile('pyproject.toml'):
+            self.filelist.append('pyproject.toml')
+
+    def _add_defaults_python(self):
+        """getting python files"""
+        if self.distribution.has_pure_modules():
+            build_py = self.get_finalized_command('build_py')
+            self.filelist.extend(build_py.get_source_files())
+            self._add_data_files(self._safe_data_files(build_py))
+
+    def _add_defaults_build_sub_commands(self):
+        build = self.get_finalized_command("build")
+        missing_cmds = set(build.get_sub_commands()) - _ORIGINAL_SUBCOMMANDS
+        # ^-- the original built-in sub-commands are already handled by default.
+        cmds = (self.get_finalized_command(c) for c in missing_cmds)
+        files = (c.get_source_files() for c in cmds if hasattr(c, "get_source_files"))
+        self.filelist.extend(chain.from_iterable(files))
+
+    def _safe_data_files(self, build_py):
+        """
+        Since the ``sdist`` class is also used to compute the MANIFEST
+        (via :obj:`setuptools.command.egg_info.manifest_maker`),
+        there might be recursion problems when trying to obtain the list of
+        data_files and ``include_package_data=True`` (which in turn depends on
+        the files included in the MANIFEST).
+
+        To avoid that, ``manifest_maker`` should be able to overwrite this
+        method and avoid recursive attempts to build/analyze the MANIFEST.
+        """
+        return build_py.data_files
+
+    def _add_data_files(self, data_files):
+        """
+        Add data files as found in build_py.data_files.
+        """
+        self.filelist.extend(
+            os.path.join(src_dir, name)
+            for _, src_dir, _, filenames in data_files
+            for name in filenames
+        )
+
+    def _add_defaults_data_files(self):
+        try:
+            super()._add_defaults_data_files()
+        except TypeError:
+            log.warn("data_files contains unexpected objects")
+
+    def prune_file_list(self) -> None:
+        super().prune_file_list()
+        # Prevent accidental inclusion of test-related cache dirs at the project root
+        sep = re.escape(os.sep)
+        self.filelist.exclude_pattern(r"^(\.tox|\.nox|\.venv)" + sep, is_regex=True)
+
+    def check_readme(self) -> None:
+        for f in self.READMES:
+            if os.path.exists(f):
+                return
+        else:
+            self.warn(
+                "standard file not found: should have one of " + ', '.join(self.READMES)
+            )
+
+    def make_release_tree(self, base_dir, files) -> None:
+        orig.sdist.make_release_tree(self, base_dir, files)
+
+        # Save any egg_info command line options used to create this sdist
+        dest = os.path.join(base_dir, 'setup.cfg')
+        if hasattr(os, 'link') and os.path.exists(dest):
+            # unlink and re-copy, since it might be hard-linked, and
+            # we don't want to change the source version
+            os.unlink(dest)
+            self.copy_file('setup.cfg', dest)
+
+        self.get_finalized_command('egg_info').save_version_info(dest)
+
+    def _manifest_is_not_generated(self):
+        # check for special comment used in 2.7.1 and higher
+        if not os.path.isfile(self.manifest):
+            return False
+
+        with open(self.manifest, 'rb') as fp:
+            first_line = fp.readline()
+        return first_line != b'# file GENERATED by distutils, do NOT edit\n'
+
+    def read_manifest(self):
+        """Read the manifest file (named by 'self.manifest') and use it to
+        fill in 'self.filelist', the list of files to include in the source
+        distribution.
+        """
+        log.info("reading manifest file '%s'", self.manifest)
+        manifest = open(self.manifest, 'rb')
+        for bytes_line in manifest:
+            # The manifest must contain UTF-8. See #303.
+            try:
+                line = bytes_line.decode('UTF-8')
+            except UnicodeDecodeError:
+                log.warn(f"{line!r} not UTF-8 decodable -- skipping")
+                continue
+            # ignore comments and blank lines
+            line = line.strip()
+            if line.startswith('#') or not line:
+                continue
+            self.filelist.append(line)
+        manifest.close()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/setopt.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/setopt.py
new file mode 100644
index 00000000..678a0593
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/setopt.py
@@ -0,0 +1,141 @@
+import configparser
+import os
+
+from .. import Command
+from ..unicode_utils import _cfg_read_utf8_with_fallback
+
+import distutils
+from distutils import log
+from distutils.errors import DistutilsOptionError
+from distutils.util import convert_path
+
+__all__ = ['config_file', 'edit_config', 'option_base', 'setopt']
+
+
+def config_file(kind="local"):
+    """Get the filename of the distutils, local, global, or per-user config
+
+    `kind` must be one of "local", "global", or "user"
+    """
+    if kind == 'local':
+        return 'setup.cfg'
+    if kind == 'global':
+        return os.path.join(os.path.dirname(distutils.__file__), 'distutils.cfg')
+    if kind == 'user':
+        dot = os.name == 'posix' and '.' or ''
+        return os.path.expanduser(convert_path(f"~/{dot}pydistutils.cfg"))
+    raise ValueError("config_file() type must be 'local', 'global', or 'user'", kind)
+
+
+def edit_config(filename, settings, dry_run=False):
+    """Edit a configuration file to include `settings`
+
+    `settings` is a dictionary of dictionaries or ``None`` values, keyed by
+    command/section name.  A ``None`` value means to delete the entire section,
+    while a dictionary lists settings to be changed or deleted in that section.
+    A setting of ``None`` means to delete that setting.
+    """
+    log.debug("Reading configuration from %s", filename)
+    opts = configparser.RawConfigParser()
+    opts.optionxform = lambda optionstr: optionstr  # type: ignore[method-assign] # overriding method
+    _cfg_read_utf8_with_fallback(opts, filename)
+
+    for section, options in settings.items():
+        if options is None:
+            log.info("Deleting section [%s] from %s", section, filename)
+            opts.remove_section(section)
+        else:
+            if not opts.has_section(section):
+                log.debug("Adding new section [%s] to %s", section, filename)
+                opts.add_section(section)
+            for option, value in options.items():
+                if value is None:
+                    log.debug("Deleting %s.%s from %s", section, option, filename)
+                    opts.remove_option(section, option)
+                    if not opts.options(section):
+                        log.info(
+                            "Deleting empty [%s] section from %s", section, filename
+                        )
+                        opts.remove_section(section)
+                else:
+                    log.debug(
+                        "Setting %s.%s to %r in %s", section, option, value, filename
+                    )
+                    opts.set(section, option, value)
+
+    log.info("Writing %s", filename)
+    if not dry_run:
+        with open(filename, 'w', encoding="utf-8") as f:
+            opts.write(f)
+
+
+class option_base(Command):
+    """Abstract base class for commands that mess with config files"""
+
+    user_options = [
+        ('global-config', 'g', "save options to the site-wide distutils.cfg file"),
+        ('user-config', 'u', "save options to the current user's pydistutils.cfg file"),
+        ('filename=', 'f', "configuration file to use (default=setup.cfg)"),
+    ]
+
+    boolean_options = [
+        'global-config',
+        'user-config',
+    ]
+
+    def initialize_options(self):
+        self.global_config = None
+        self.user_config = None
+        self.filename = None
+
+    def finalize_options(self):
+        filenames = []
+        if self.global_config:
+            filenames.append(config_file('global'))
+        if self.user_config:
+            filenames.append(config_file('user'))
+        if self.filename is not None:
+            filenames.append(self.filename)
+        if not filenames:
+            filenames.append(config_file('local'))
+        if len(filenames) > 1:
+            raise DistutilsOptionError(
+                "Must specify only one configuration file option", filenames
+            )
+        (self.filename,) = filenames
+
+
+class setopt(option_base):
+    """Save command-line options to a file"""
+
+    description = "set an option in setup.cfg or another config file"
+
+    user_options = [
+        ('command=', 'c', 'command to set an option for'),
+        ('option=', 'o', 'option to set'),
+        ('set-value=', 's', 'value of the option'),
+        ('remove', 'r', 'remove (unset) the value'),
+    ] + option_base.user_options
+
+    boolean_options = option_base.boolean_options + ['remove']
+
+    def initialize_options(self):
+        option_base.initialize_options(self)
+        self.command = None
+        self.option = None
+        self.set_value = None
+        self.remove = None
+
+    def finalize_options(self) -> None:
+        option_base.finalize_options(self)
+        if self.command is None or self.option is None:
+            raise DistutilsOptionError("Must specify --command *and* --option")
+        if self.set_value is None and not self.remove:
+            raise DistutilsOptionError("Must specify --set-value or --remove")
+
+    def run(self) -> None:
+        edit_config(
+            self.filename,
+            {self.command: {self.option.replace('-', '_'): self.set_value}},
+            self.dry_run,
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/command/test.py b/backend/venv/lib/python3.12/site-packages/setuptools/command/test.py
new file mode 100644
index 00000000..341b11a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/command/test.py
@@ -0,0 +1,45 @@
+from __future__ import annotations
+
+from setuptools import Command
+from setuptools.warnings import SetuptoolsDeprecationWarning
+
+
+# Would restrict to Literal["test"], but mypy doesn't support it: https://github.com/python/mypy/issues/8203
+def __getattr__(name: str) -> type[_test]:
+    if name == 'test':
+        SetuptoolsDeprecationWarning.emit(
+            "The test command is disabled and references to it are deprecated.",
+            "Please remove any references to `setuptools.command.test` in all "
+            "supported versions of the affected package.",
+            due_date=(2024, 11, 15),
+            stacklevel=2,
+        )
+        return _test
+    raise AttributeError(name)
+
+
+class _test(Command):
+    """
+    Stub to warn when test command is referenced or used.
+    """
+
+    description = "stub for old test command (do not use)"
+
+    user_options = [
+        ('test-module=', 'm', "Run 'test_suite' in specified module"),
+        (
+            'test-suite=',
+            's',
+            "Run single test, case or suite (e.g. 'module.test_suite')",
+        ),
+        ('test-runner=', 'r', "Test runner to use"),
+    ]
+
+    def initialize_options(self):
+        pass
+
+    def finalize_options(self):
+        pass
+
+    def run(self):
+        raise RuntimeError("Support for the test command was removed in Setuptools 72")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..3b204de7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py310.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py310.cpython-312.pyc
new file mode 100644
index 00000000..492e9816
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py310.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py311.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py311.cpython-312.pyc
new file mode 100644
index 00000000..59390194
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py311.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py312.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py312.cpython-312.pyc
new file mode 100644
index 00000000..60cae84d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py312.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py39.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py39.cpython-312.pyc
new file mode 100644
index 00000000..45127a31
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/compat/__pycache__/py39.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/py310.py b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py310.py
new file mode 100644
index 00000000..b3912f8e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py310.py
@@ -0,0 +1,9 @@
+import sys
+
+__all__ = ['tomllib']
+
+
+if sys.version_info >= (3, 11):
+    import tomllib
+else:  # pragma: no cover
+    import tomli as tomllib
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/py311.py b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py311.py
new file mode 100644
index 00000000..52b58af3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py311.py
@@ -0,0 +1,27 @@
+from __future__ import annotations
+
+import shutil
+import sys
+from typing import TYPE_CHECKING, Any, Callable
+
+if TYPE_CHECKING:
+    from _typeshed import ExcInfo, StrOrBytesPath
+    from typing_extensions import TypeAlias
+
+# Same as shutil._OnExcCallback from typeshed
+_OnExcCallback: TypeAlias = Callable[[Callable[..., Any], str, BaseException], object]
+
+
+def shutil_rmtree(
+    path: StrOrBytesPath,
+    ignore_errors: bool = False,
+    onexc: _OnExcCallback | None = None,
+) -> None:
+    if sys.version_info >= (3, 12):
+        return shutil.rmtree(path, ignore_errors, onexc=onexc)
+
+    def _handler(fn: Callable[..., Any], path: str, excinfo: ExcInfo) -> None:
+        if onexc:
+            onexc(fn, path, excinfo[1])
+
+    return shutil.rmtree(path, ignore_errors, onerror=_handler)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/py312.py b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py312.py
new file mode 100644
index 00000000..b20c5f69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py312.py
@@ -0,0 +1,13 @@
+from __future__ import annotations
+
+import sys
+
+if sys.version_info >= (3, 12, 4):
+    # Python 3.13 should support `.pth` files encoded in UTF-8
+    # See discussion in https://github.com/python/cpython/issues/77102
+    PTH_ENCODING: str | None = "utf-8"
+else:
+    from .py39 import LOCALE_ENCODING
+
+    # PTH_ENCODING = "locale"
+    PTH_ENCODING = LOCALE_ENCODING
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/compat/py39.py b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py39.py
new file mode 100644
index 00000000..04a4abe5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/compat/py39.py
@@ -0,0 +1,9 @@
+import sys
+
+# Explicitly use the ``"locale"`` encoding in versions that support it,
+# otherwise just rely on the implicit handling of ``encoding=None``.
+# Since all platforms that support ``EncodingWarning`` also support
+# ``encoding="locale"``, this can be used to suppress the warning.
+# However, please try to use UTF-8 when possible
+# (.pth files are the notorious exception: python/cpython#77102, pypa/setuptools#3937).
+LOCALE_ENCODING = "locale" if sys.version_info >= (3, 10) else None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/NOTICE b/backend/venv/lib/python3.12/site-packages/setuptools/config/NOTICE
new file mode 100644
index 00000000..01864511
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/NOTICE
@@ -0,0 +1,10 @@
+The following files include code from opensource projects
+(either as direct copies or modified versions):
+
+- `setuptools.schema.json`, `distutils.schema.json`:
+    - project: `validate-pyproject` - licensed under MPL-2.0
+      (https://github.com/abravalheri/validate-pyproject):
+
+      This Source Code Form is subject to the terms of the Mozilla Public
+      License, v. 2.0. If a copy of the MPL was not distributed with this file,
+      You can obtain one at https://mozilla.org/MPL/2.0/.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/__init__.py
new file mode 100644
index 00000000..fcc7d008
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/__init__.py
@@ -0,0 +1,43 @@
+"""For backward compatibility, expose main functions from
+``setuptools.config.setupcfg``
+"""
+
+from functools import wraps
+from typing import Callable, TypeVar, cast
+
+from ..warnings import SetuptoolsDeprecationWarning
+from . import setupcfg
+
+Fn = TypeVar("Fn", bound=Callable)
+
+__all__ = ('parse_configuration', 'read_configuration')
+
+
+def _deprecation_notice(fn: Fn) -> Fn:
+    @wraps(fn)
+    def _wrapper(*args, **kwargs):
+        SetuptoolsDeprecationWarning.emit(
+            "Deprecated API usage.",
+            f"""
+            As setuptools moves its configuration towards `pyproject.toml`,
+            `{__name__}.{fn.__name__}` became deprecated.
+
+            For the time being, you can use the `{setupcfg.__name__}` module
+            to access a backward compatible API, but this module is provisional
+            and might be removed in the future.
+
+            To read project metadata, consider using
+            ``build.util.project_wheel_metadata`` (https://pypi.org/project/build/).
+            For simple scenarios, you can also try parsing the file directly
+            with the help of ``configparser``.
+            """,
+            # due_date not defined yet, because the community still heavily relies on it
+            # Warning introduced in 24 Mar 2022
+        )
+        return fn(*args, **kwargs)
+
+    return cast(Fn, _wrapper)
+
+
+read_configuration = _deprecation_notice(setupcfg.read_configuration)
+parse_configuration = _deprecation_notice(setupcfg.parse_configuration)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..38eb7108
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/_apply_pyprojecttoml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/_apply_pyprojecttoml.cpython-312.pyc
new file mode 100644
index 00000000..6fb96d08
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/_apply_pyprojecttoml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/expand.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/expand.cpython-312.pyc
new file mode 100644
index 00000000..2c0f2761
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/expand.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/pyprojecttoml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/pyprojecttoml.cpython-312.pyc
new file mode 100644
index 00000000..a8642d99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/pyprojecttoml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/setupcfg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/setupcfg.cpython-312.pyc
new file mode 100644
index 00000000..276468f1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/__pycache__/setupcfg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_apply_pyprojecttoml.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_apply_pyprojecttoml.py
new file mode 100644
index 00000000..331596bd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_apply_pyprojecttoml.py
@@ -0,0 +1,488 @@
+"""Translation layer between pyproject config and setuptools distribution and
+metadata objects.
+
+The distribution and metadata objects are modeled after (an old version of)
+core metadata, therefore configs in the format specified for ``pyproject.toml``
+need to be processed before being applied.
+
+**PRIVATE MODULE**: API reserved for setuptools internal usage only.
+"""
+
+from __future__ import annotations
+
+import logging
+import os
+from collections.abc import Mapping
+from email.headerregistry import Address
+from functools import partial, reduce
+from inspect import cleandoc
+from itertools import chain
+from types import MappingProxyType
+from typing import TYPE_CHECKING, Any, Callable, TypeVar, Union
+
+from .. import _static
+from .._path import StrPath
+from ..errors import RemovedConfigError
+from ..extension import Extension
+from ..warnings import SetuptoolsWarning
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+    from setuptools._importlib import metadata
+    from setuptools.dist import Distribution
+
+    from distutils.dist import _OptionsList  # Comes from typeshed
+
+
+EMPTY: Mapping = MappingProxyType({})  # Immutable dict-like
+_ProjectReadmeValue: TypeAlias = Union[str, dict[str, str]]
+_Correspondence: TypeAlias = Callable[["Distribution", Any, Union[StrPath, None]], None]
+_T = TypeVar("_T")
+
+_logger = logging.getLogger(__name__)
+
+
+def apply(dist: Distribution, config: dict, filename: StrPath) -> Distribution:
+    """Apply configuration dict read with :func:`read_configuration`"""
+
+    if not config:
+        return dist  # short-circuit unrelated pyproject.toml file
+
+    root_dir = os.path.dirname(filename) or "."
+
+    _apply_project_table(dist, config, root_dir)
+    _apply_tool_table(dist, config, filename)
+
+    current_directory = os.getcwd()
+    os.chdir(root_dir)
+    try:
+        dist._finalize_requires()
+        dist._finalize_license_files()
+    finally:
+        os.chdir(current_directory)
+
+    return dist
+
+
+def _apply_project_table(dist: Distribution, config: dict, root_dir: StrPath):
+    orig_config = config.get("project", {})
+    if not orig_config:
+        return  # short-circuit
+
+    project_table = {k: _static.attempt_conversion(v) for k, v in orig_config.items()}
+    _handle_missing_dynamic(dist, project_table)
+    _unify_entry_points(project_table)
+
+    for field, value in project_table.items():
+        norm_key = json_compatible_key(field)
+        corresp = PYPROJECT_CORRESPONDENCE.get(norm_key, norm_key)
+        if callable(corresp):
+            corresp(dist, value, root_dir)
+        else:
+            _set_config(dist, corresp, value)
+
+
+def _apply_tool_table(dist: Distribution, config: dict, filename: StrPath):
+    tool_table = config.get("tool", {}).get("setuptools", {})
+    if not tool_table:
+        return  # short-circuit
+
+    for field, value in tool_table.items():
+        norm_key = json_compatible_key(field)
+
+        if norm_key in TOOL_TABLE_REMOVALS:
+            suggestion = cleandoc(TOOL_TABLE_REMOVALS[norm_key])
+            msg = f"""
+            The parameter `tool.setuptools.{field}` was long deprecated
+            and has been removed from `pyproject.toml`.
+            """
+            raise RemovedConfigError("\n".join([cleandoc(msg), suggestion]))
+
+        norm_key = TOOL_TABLE_RENAMES.get(norm_key, norm_key)
+        corresp = TOOL_TABLE_CORRESPONDENCE.get(norm_key, norm_key)
+        if callable(corresp):
+            corresp(dist, value)
+        else:
+            _set_config(dist, corresp, value)
+
+    _copy_command_options(config, dist, filename)
+
+
+def _handle_missing_dynamic(dist: Distribution, project_table: dict):
+    """Be temporarily forgiving with ``dynamic`` fields not listed in ``dynamic``"""
+    dynamic = set(project_table.get("dynamic", []))
+    for field, getter in _PREVIOUSLY_DEFINED.items():
+        if not (field in project_table or field in dynamic):
+            value = getter(dist)
+            if value:
+                _MissingDynamic.emit(field=field, value=value)
+                project_table[field] = _RESET_PREVIOUSLY_DEFINED.get(field)
+
+
+def json_compatible_key(key: str) -> str:
+    """As defined in :pep:`566#json-compatible-metadata`"""
+    return key.lower().replace("-", "_")
+
+
+def _set_config(dist: Distribution, field: str, value: Any):
+    val = _PREPROCESS.get(field, _noop)(dist, value)
+    setter = getattr(dist.metadata, f"set_{field}", None)
+    if setter:
+        setter(val)
+    elif hasattr(dist.metadata, field) or field in SETUPTOOLS_PATCHES:
+        setattr(dist.metadata, field, val)
+    else:
+        setattr(dist, field, val)
+
+
+_CONTENT_TYPES = {
+    ".md": "text/markdown",
+    ".rst": "text/x-rst",
+    ".txt": "text/plain",
+}
+
+
+def _guess_content_type(file: str) -> str | None:
+    _, ext = os.path.splitext(file.lower())
+    if not ext:
+        return None
+
+    if ext in _CONTENT_TYPES:
+        return _static.Str(_CONTENT_TYPES[ext])
+
+    valid = ", ".join(f"{k} ({v})" for k, v in _CONTENT_TYPES.items())
+    msg = f"only the following file extensions are recognized: {valid}."
+    raise ValueError(f"Undefined content type for {file}, {msg}")
+
+
+def _long_description(
+    dist: Distribution, val: _ProjectReadmeValue, root_dir: StrPath | None
+):
+    from setuptools.config import expand
+
+    file: str | tuple[()]
+    if isinstance(val, str):
+        file = val
+        text = expand.read_files(file, root_dir)
+        ctype = _guess_content_type(file)
+    else:
+        file = val.get("file") or ()
+        text = val.get("text") or expand.read_files(file, root_dir)
+        ctype = val["content-type"]
+
+    # XXX: Is it completely safe to assume static?
+    _set_config(dist, "long_description", _static.Str(text))
+
+    if ctype:
+        _set_config(dist, "long_description_content_type", _static.Str(ctype))
+
+    if file:
+        dist._referenced_files.add(file)
+
+
+def _license(dist: Distribution, val: dict, root_dir: StrPath | None):
+    from setuptools.config import expand
+
+    if "file" in val:
+        # XXX: Is it completely safe to assume static?
+        value = expand.read_files([val["file"]], root_dir)
+        _set_config(dist, "license", _static.Str(value))
+        dist._referenced_files.add(val["file"])
+    else:
+        _set_config(dist, "license", _static.Str(val["text"]))
+
+
+def _people(dist: Distribution, val: list[dict], _root_dir: StrPath | None, kind: str):
+    field = []
+    email_field = []
+    for person in val:
+        if "name" not in person:
+            email_field.append(person["email"])
+        elif "email" not in person:
+            field.append(person["name"])
+        else:
+            addr = Address(display_name=person["name"], addr_spec=person["email"])
+            email_field.append(str(addr))
+
+    if field:
+        _set_config(dist, kind, _static.Str(", ".join(field)))
+    if email_field:
+        _set_config(dist, f"{kind}_email", _static.Str(", ".join(email_field)))
+
+
+def _project_urls(dist: Distribution, val: dict, _root_dir: StrPath | None):
+    _set_config(dist, "project_urls", val)
+
+
+def _python_requires(dist: Distribution, val: str, _root_dir: StrPath | None):
+    _set_config(dist, "python_requires", _static.SpecifierSet(val))
+
+
+def _dependencies(dist: Distribution, val: list, _root_dir: StrPath | None):
+    if getattr(dist, "install_requires", []):
+        msg = "`install_requires` overwritten in `pyproject.toml` (dependencies)"
+        SetuptoolsWarning.emit(msg)
+    dist.install_requires = val
+
+
+def _optional_dependencies(dist: Distribution, val: dict, _root_dir: StrPath | None):
+    if getattr(dist, "extras_require", None):
+        msg = "`extras_require` overwritten in `pyproject.toml` (optional-dependencies)"
+        SetuptoolsWarning.emit(msg)
+    dist.extras_require = val
+
+
+def _ext_modules(dist: Distribution, val: list[dict]) -> list[Extension]:
+    existing = dist.ext_modules or []
+    args = ({k.replace("-", "_"): v for k, v in x.items()} for x in val)
+    new = [Extension(**kw) for kw in args]
+    return [*existing, *new]
+
+
+def _noop(_dist: Distribution, val: _T) -> _T:
+    return val
+
+
+def _identity(val: _T) -> _T:
+    return val
+
+
+def _unify_entry_points(project_table: dict):
+    project = project_table
+    given = project.pop("entry-points", project.pop("entry_points", {}))
+    entry_points = dict(given)  # Avoid problems with static
+    renaming = {"scripts": "console_scripts", "gui_scripts": "gui_scripts"}
+    for key, value in list(project.items()):  # eager to allow modifications
+        norm_key = json_compatible_key(key)
+        if norm_key in renaming:
+            # Don't skip even if value is empty (reason: reset missing `dynamic`)
+            entry_points[renaming[norm_key]] = project.pop(key)
+
+    if entry_points:
+        project["entry-points"] = {
+            name: [f"{k} = {v}" for k, v in group.items()]
+            for name, group in entry_points.items()
+            if group  # now we can skip empty groups
+        }
+        # Sometimes this will set `project["entry-points"] = {}`, and that is
+        # intentional (for resetting configurations that are missing `dynamic`).
+
+
+def _copy_command_options(pyproject: dict, dist: Distribution, filename: StrPath):
+    tool_table = pyproject.get("tool", {})
+    cmdclass = tool_table.get("setuptools", {}).get("cmdclass", {})
+    valid_options = _valid_command_options(cmdclass)
+
+    cmd_opts = dist.command_options
+    for cmd, config in pyproject.get("tool", {}).get("distutils", {}).items():
+        cmd = json_compatible_key(cmd)
+        valid = valid_options.get(cmd, set())
+        cmd_opts.setdefault(cmd, {})
+        for key, value in config.items():
+            key = json_compatible_key(key)
+            cmd_opts[cmd][key] = (str(filename), value)
+            if key not in valid:
+                # To avoid removing options that are specified dynamically we
+                # just log a warn...
+                _logger.warning(f"Command option {cmd}.{key} is not defined")
+
+
+def _valid_command_options(cmdclass: Mapping = EMPTY) -> dict[str, set[str]]:
+    from setuptools.dist import Distribution
+
+    from .._importlib import metadata
+
+    valid_options = {"global": _normalise_cmd_options(Distribution.global_options)}
+
+    unloaded_entry_points = metadata.entry_points(group='distutils.commands')
+    loaded_entry_points = (_load_ep(ep) for ep in unloaded_entry_points)
+    entry_points = (ep for ep in loaded_entry_points if ep)
+    for cmd, cmd_class in chain(entry_points, cmdclass.items()):
+        opts = valid_options.get(cmd, set())
+        opts = opts | _normalise_cmd_options(getattr(cmd_class, "user_options", []))
+        valid_options[cmd] = opts
+
+    return valid_options
+
+
+def _load_ep(ep: metadata.EntryPoint) -> tuple[str, type] | None:
+    if ep.value.startswith("wheel.bdist_wheel"):
+        # Ignore deprecated entrypoint from wheel and avoid warning pypa/wheel#631
+        # TODO: remove check when `bdist_wheel` has been fully removed from pypa/wheel
+        return None
+
+    # Ignore all the errors
+    try:
+        return (ep.name, ep.load())
+    except Exception as ex:
+        msg = f"{ex.__class__.__name__} while trying to load entry-point {ep.name}"
+        _logger.warning(f"{msg}: {ex}")
+        return None
+
+
+def _normalise_cmd_option_key(name: str) -> str:
+    return json_compatible_key(name).strip("_=")
+
+
+def _normalise_cmd_options(desc: _OptionsList) -> set[str]:
+    return {_normalise_cmd_option_key(fancy_option[0]) for fancy_option in desc}
+
+
+def _get_previous_entrypoints(dist: Distribution) -> dict[str, list]:
+    ignore = ("console_scripts", "gui_scripts")
+    value = getattr(dist, "entry_points", None) or {}
+    return {k: v for k, v in value.items() if k not in ignore}
+
+
+def _get_previous_scripts(dist: Distribution) -> list | None:
+    value = getattr(dist, "entry_points", None) or {}
+    return value.get("console_scripts")
+
+
+def _get_previous_gui_scripts(dist: Distribution) -> list | None:
+    value = getattr(dist, "entry_points", None) or {}
+    return value.get("gui_scripts")
+
+
+def _set_static_list_metadata(attr: str, dist: Distribution, val: list) -> None:
+    """Apply distutils metadata validation but preserve "static" behaviour"""
+    meta = dist.metadata
+    setter, getter = getattr(meta, f"set_{attr}"), getattr(meta, f"get_{attr}")
+    setter(val)
+    setattr(meta, attr, _static.List(getter()))
+
+
+def _attrgetter(attr):
+    """
+    Similar to ``operator.attrgetter`` but returns None if ``attr`` is not found
+    >>> from types import SimpleNamespace
+    >>> obj = SimpleNamespace(a=42, b=SimpleNamespace(c=13))
+    >>> _attrgetter("a")(obj)
+    42
+    >>> _attrgetter("b.c")(obj)
+    13
+    >>> _attrgetter("d")(obj) is None
+    True
+    """
+    return partial(reduce, lambda acc, x: getattr(acc, x, None), attr.split("."))
+
+
+def _some_attrgetter(*items):
+    """
+    Return the first "truth-y" attribute or None
+    >>> from types import SimpleNamespace
+    >>> obj = SimpleNamespace(a=42, b=SimpleNamespace(c=13))
+    >>> _some_attrgetter("d", "a", "b.c")(obj)
+    42
+    >>> _some_attrgetter("d", "e", "b.c", "a")(obj)
+    13
+    >>> _some_attrgetter("d", "e", "f")(obj) is None
+    True
+    """
+
+    def _acessor(obj):
+        values = (_attrgetter(i)(obj) for i in items)
+        return next((i for i in values if i is not None), None)
+
+    return _acessor
+
+
+PYPROJECT_CORRESPONDENCE: dict[str, _Correspondence] = {
+    "readme": _long_description,
+    "license": _license,
+    "authors": partial(_people, kind="author"),
+    "maintainers": partial(_people, kind="maintainer"),
+    "urls": _project_urls,
+    "dependencies": _dependencies,
+    "optional_dependencies": _optional_dependencies,
+    "requires_python": _python_requires,
+}
+
+TOOL_TABLE_RENAMES = {"script_files": "scripts"}
+TOOL_TABLE_REMOVALS = {
+    "namespace_packages": """
+        Please migrate to implicit native namespaces instead.
+        See https://packaging.python.org/en/latest/guides/packaging-namespace-packages/.
+        """,
+}
+TOOL_TABLE_CORRESPONDENCE = {
+    # Fields with corresponding core metadata need to be marked as static:
+    "obsoletes": partial(_set_static_list_metadata, "obsoletes"),
+    "provides": partial(_set_static_list_metadata, "provides"),
+    "platforms": partial(_set_static_list_metadata, "platforms"),
+}
+
+SETUPTOOLS_PATCHES = {
+    "long_description_content_type",
+    "project_urls",
+    "provides_extras",
+    "license_file",
+    "license_files",
+}
+
+_PREPROCESS = {
+    "ext_modules": _ext_modules,
+}
+
+_PREVIOUSLY_DEFINED = {
+    "name": _attrgetter("metadata.name"),
+    "version": _attrgetter("metadata.version"),
+    "description": _attrgetter("metadata.description"),
+    "readme": _attrgetter("metadata.long_description"),
+    "requires-python": _some_attrgetter("python_requires", "metadata.python_requires"),
+    "license": _attrgetter("metadata.license"),
+    "authors": _some_attrgetter("metadata.author", "metadata.author_email"),
+    "maintainers": _some_attrgetter("metadata.maintainer", "metadata.maintainer_email"),
+    "keywords": _attrgetter("metadata.keywords"),
+    "classifiers": _attrgetter("metadata.classifiers"),
+    "urls": _attrgetter("metadata.project_urls"),
+    "entry-points": _get_previous_entrypoints,
+    "scripts": _get_previous_scripts,
+    "gui-scripts": _get_previous_gui_scripts,
+    "dependencies": _attrgetter("install_requires"),
+    "optional-dependencies": _attrgetter("extras_require"),
+}
+
+
+_RESET_PREVIOUSLY_DEFINED: dict = {
+    # Fix improper setting: given in `setup.py`, but not listed in `dynamic`
+    # dict: pyproject name => value to which reset
+    "license": _static.EMPTY_DICT,
+    "authors": _static.EMPTY_LIST,
+    "maintainers": _static.EMPTY_LIST,
+    "keywords": _static.EMPTY_LIST,
+    "classifiers": _static.EMPTY_LIST,
+    "urls": _static.EMPTY_DICT,
+    "entry-points": _static.EMPTY_DICT,
+    "scripts": _static.EMPTY_DICT,
+    "gui-scripts": _static.EMPTY_DICT,
+    "dependencies": _static.EMPTY_LIST,
+    "optional-dependencies": _static.EMPTY_DICT,
+}
+
+
+class _MissingDynamic(SetuptoolsWarning):
+    _SUMMARY = "`{field}` defined outside of `pyproject.toml` is ignored."
+
+    _DETAILS = """
+    The following seems to be defined outside of `pyproject.toml`:
+
+    `{field} = {value!r}`
+
+    According to the spec (see the link below), however, setuptools CANNOT
+    consider this value unless `{field}` is listed as `dynamic`.
+
+    https://packaging.python.org/en/latest/specifications/pyproject-toml/#declaring-project-metadata-the-project-table
+
+    To prevent this problem, you can list `{field}` under `dynamic` or alternatively
+    remove the `[project]` table from your file and rely entirely on other means of
+    configuration.
+    """
+    # TODO: Consider removing this check in the future?
+    #       There is a trade-off here between improving "debug-ability" and the cost
+    #       of running/testing/maintaining these unnecessary checks...
+
+    @classmethod
+    def details(cls, field: str, value: Any) -> str:
+        return cls._DETAILS.format(field=field, value=value)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/NOTICE b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/NOTICE
new file mode 100644
index 00000000..74e8821f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/NOTICE
@@ -0,0 +1,438 @@
+The code contained in this directory was automatically generated using the
+following command:
+
+    python -m validate_pyproject.pre_compile --output-dir=setuptools/config/_validate_pyproject --enable-plugins setuptools distutils --very-verbose -t distutils=setuptools/config/distutils.schema.json -t setuptools=setuptools/config/setuptools.schema.json
+
+Please avoid changing it manually.
+
+
+You can report issues or suggest changes directly to `validate-pyproject`
+(or to the relevant plugin repository)
+
+- https://github.com/abravalheri/validate-pyproject/issues
+
+
+***
+
+The following files include code from opensource projects
+(either as direct copies or modified versions):
+
+- `fastjsonschema_exceptions.py`:
+    - project: `fastjsonschema` - licensed under BSD-3-Clause
+      (https://github.com/horejsek/python-fastjsonschema)
+- `extra_validations.py` and `format.py`, `error_reporting.py`:
+    - project: `validate-pyproject` - licensed under MPL-2.0
+      (https://github.com/abravalheri/validate-pyproject)
+
+
+Additionally the following files are automatically generated by tools provided
+by the same projects:
+
+- `__init__.py`
+- `fastjsonschema_validations.py`
+
+The relevant copyright notes and licenses are included below.
+
+
+***
+
+`fastjsonschema`
+================
+
+Copyright (c) 2018, Michal Horejsek
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification,
+are permitted provided that the following conditions are met:
+
+  Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+
+  Redistributions in binary form must reproduce the above copyright notice, this
+  list of conditions and the following disclaimer in the documentation and/or
+  other materials provided with the distribution.
+
+  Neither the name of the {organization} nor the names of its
+  contributors may be used to endorse or promote products derived from
+  this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+
+***
+
+`validate-pyproject`
+====================
+
+Mozilla Public License, version 2.0
+
+1. Definitions
+
+1.1. "Contributor"
+
+     means each individual or legal entity that creates, contributes to the
+     creation of, or owns Covered Software.
+
+1.2. "Contributor Version"
+
+     means the combination of the Contributions of others (if any) used by a
+     Contributor and that particular Contributor's Contribution.
+
+1.3. "Contribution"
+
+     means Covered Software of a particular Contributor.
+
+1.4. "Covered Software"
+
+     means Source Code Form to which the initial Contributor has attached the
+     notice in Exhibit A, the Executable Form of such Source Code Form, and
+     Modifications of such Source Code Form, in each case including portions
+     thereof.
+
+1.5. "Incompatible With Secondary Licenses"
+     means
+
+     a. that the initial Contributor has attached the notice described in
+        Exhibit B to the Covered Software; or
+
+     b. that the Covered Software was made available under the terms of
+        version 1.1 or earlier of the License, but not also under the terms of
+        a Secondary License.
+
+1.6. "Executable Form"
+
+     means any form of the work other than Source Code Form.
+
+1.7. "Larger Work"
+
+     means a work that combines Covered Software with other material, in a
+     separate file or files, that is not Covered Software.
+
+1.8. "License"
+
+     means this document.
+
+1.9. "Licensable"
+
+     means having the right to grant, to the maximum extent possible, whether
+     at the time of the initial grant or subsequently, any and all of the
+     rights conveyed by this License.
+
+1.10. "Modifications"
+
+     means any of the following:
+
+     a. any file in Source Code Form that results from an addition to,
+        deletion from, or modification of the contents of Covered Software; or
+
+     b. any new file in Source Code Form that contains any Covered Software.
+
+1.11. "Patent Claims" of a Contributor
+
+      means any patent claim(s), including without limitation, method,
+      process, and apparatus claims, in any patent Licensable by such
+      Contributor that would be infringed, but for the grant of the License,
+      by the making, using, selling, offering for sale, having made, import,
+      or transfer of either its Contributions or its Contributor Version.
+
+1.12. "Secondary License"
+
+      means either the GNU General Public License, Version 2.0, the GNU Lesser
+      General Public License, Version 2.1, the GNU Affero General Public
+      License, Version 3.0, or any later versions of those licenses.
+
+1.13. "Source Code Form"
+
+      means the form of the work preferred for making modifications.
+
+1.14. "You" (or "Your")
+
+      means an individual or a legal entity exercising rights under this
+      License. For legal entities, "You" includes any entity that controls, is
+      controlled by, or is under common control with You. For purposes of this
+      definition, "control" means (a) the power, direct or indirect, to cause
+      the direction or management of such entity, whether by contract or
+      otherwise, or (b) ownership of more than fifty percent (50%) of the
+      outstanding shares or beneficial ownership of such entity.
+
+
+2. License Grants and Conditions
+
+2.1. Grants
+
+     Each Contributor hereby grants You a world-wide, royalty-free,
+     non-exclusive license:
+
+     a. under intellectual property rights (other than patent or trademark)
+        Licensable by such Contributor to use, reproduce, make available,
+        modify, display, perform, distribute, and otherwise exploit its
+        Contributions, either on an unmodified basis, with Modifications, or
+        as part of a Larger Work; and
+
+     b. under Patent Claims of such Contributor to make, use, sell, offer for
+        sale, have made, import, and otherwise transfer either its
+        Contributions or its Contributor Version.
+
+2.2. Effective Date
+
+     The licenses granted in Section 2.1 with respect to any Contribution
+     become effective for each Contribution on the date the Contributor first
+     distributes such Contribution.
+
+2.3. Limitations on Grant Scope
+
+     The licenses granted in this Section 2 are the only rights granted under
+     this License. No additional rights or licenses will be implied from the
+     distribution or licensing of Covered Software under this License.
+     Notwithstanding Section 2.1(b) above, no patent license is granted by a
+     Contributor:
+
+     a. for any code that a Contributor has removed from Covered Software; or
+
+     b. for infringements caused by: (i) Your and any other third party's
+        modifications of Covered Software, or (ii) the combination of its
+        Contributions with other software (except as part of its Contributor
+        Version); or
+
+     c. under Patent Claims infringed by Covered Software in the absence of
+        its Contributions.
+
+     This License does not grant any rights in the trademarks, service marks,
+     or logos of any Contributor (except as may be necessary to comply with
+     the notice requirements in Section 3.4).
+
+2.4. Subsequent Licenses
+
+     No Contributor makes additional grants as a result of Your choice to
+     distribute the Covered Software under a subsequent version of this
+     License (see Section 10.2) or under the terms of a Secondary License (if
+     permitted under the terms of Section 3.3).
+
+2.5. Representation
+
+     Each Contributor represents that the Contributor believes its
+     Contributions are its original creation(s) or it has sufficient rights to
+     grant the rights to its Contributions conveyed by this License.
+
+2.6. Fair Use
+
+     This License is not intended to limit any rights You have under
+     applicable copyright doctrines of fair use, fair dealing, or other
+     equivalents.
+
+2.7. Conditions
+
+     Sections 3.1, 3.2, 3.3, and 3.4 are conditions of the licenses granted in
+     Section 2.1.
+
+
+3. Responsibilities
+
+3.1. Distribution of Source Form
+
+     All distribution of Covered Software in Source Code Form, including any
+     Modifications that You create or to which You contribute, must be under
+     the terms of this License. You must inform recipients that the Source
+     Code Form of the Covered Software is governed by the terms of this
+     License, and how they can obtain a copy of this License. You may not
+     attempt to alter or restrict the recipients' rights in the Source Code
+     Form.
+
+3.2. Distribution of Executable Form
+
+     If You distribute Covered Software in Executable Form then:
+
+     a. such Covered Software must also be made available in Source Code Form,
+        as described in Section 3.1, and You must inform recipients of the
+        Executable Form how they can obtain a copy of such Source Code Form by
+        reasonable means in a timely manner, at a charge no more than the cost
+        of distribution to the recipient; and
+
+     b. You may distribute such Executable Form under the terms of this
+        License, or sublicense it under different terms, provided that the
+        license for the Executable Form does not attempt to limit or alter the
+        recipients' rights in the Source Code Form under this License.
+
+3.3. Distribution of a Larger Work
+
+     You may create and distribute a Larger Work under terms of Your choice,
+     provided that You also comply with the requirements of this License for
+     the Covered Software. If the Larger Work is a combination of Covered
+     Software with a work governed by one or more Secondary Licenses, and the
+     Covered Software is not Incompatible With Secondary Licenses, this
+     License permits You to additionally distribute such Covered Software
+     under the terms of such Secondary License(s), so that the recipient of
+     the Larger Work may, at their option, further distribute the Covered
+     Software under the terms of either this License or such Secondary
+     License(s).
+
+3.4. Notices
+
+     You may not remove or alter the substance of any license notices
+     (including copyright notices, patent notices, disclaimers of warranty, or
+     limitations of liability) contained within the Source Code Form of the
+     Covered Software, except that You may alter any license notices to the
+     extent required to remedy known factual inaccuracies.
+
+3.5. Application of Additional Terms
+
+     You may choose to offer, and to charge a fee for, warranty, support,
+     indemnity or liability obligations to one or more recipients of Covered
+     Software. However, You may do so only on Your own behalf, and not on
+     behalf of any Contributor. You must make it absolutely clear that any
+     such warranty, support, indemnity, or liability obligation is offered by
+     You alone, and You hereby agree to indemnify every Contributor for any
+     liability incurred by such Contributor as a result of warranty, support,
+     indemnity or liability terms You offer. You may include additional
+     disclaimers of warranty and limitations of liability specific to any
+     jurisdiction.
+
+4. Inability to Comply Due to Statute or Regulation
+
+   If it is impossible for You to comply with any of the terms of this License
+   with respect to some or all of the Covered Software due to statute,
+   judicial order, or regulation then You must: (a) comply with the terms of
+   this License to the maximum extent possible; and (b) describe the
+   limitations and the code they affect. Such description must be placed in a
+   text file included with all distributions of the Covered Software under
+   this License. Except to the extent prohibited by statute or regulation,
+   such description must be sufficiently detailed for a recipient of ordinary
+   skill to be able to understand it.
+
+5. Termination
+
+5.1. The rights granted under this License will terminate automatically if You
+     fail to comply with any of its terms. However, if You become compliant,
+     then the rights granted under this License from a particular Contributor
+     are reinstated (a) provisionally, unless and until such Contributor
+     explicitly and finally terminates Your grants, and (b) on an ongoing
+     basis, if such Contributor fails to notify You of the non-compliance by
+     some reasonable means prior to 60 days after You have come back into
+     compliance. Moreover, Your grants from a particular Contributor are
+     reinstated on an ongoing basis if such Contributor notifies You of the
+     non-compliance by some reasonable means, this is the first time You have
+     received notice of non-compliance with this License from such
+     Contributor, and You become compliant prior to 30 days after Your receipt
+     of the notice.
+
+5.2. If You initiate litigation against any entity by asserting a patent
+     infringement claim (excluding declaratory judgment actions,
+     counter-claims, and cross-claims) alleging that a Contributor Version
+     directly or indirectly infringes any patent, then the rights granted to
+     You by any and all Contributors for the Covered Software under Section
+     2.1 of this License shall terminate.
+
+5.3. In the event of termination under Sections 5.1 or 5.2 above, all end user
+     license agreements (excluding distributors and resellers) which have been
+     validly granted by You or Your distributors under this License prior to
+     termination shall survive termination.
+
+6. Disclaimer of Warranty
+
+   Covered Software is provided under this License on an "as is" basis,
+   without warranty of any kind, either expressed, implied, or statutory,
+   including, without limitation, warranties that the Covered Software is free
+   of defects, merchantable, fit for a particular purpose or non-infringing.
+   The entire risk as to the quality and performance of the Covered Software
+   is with You. Should any Covered Software prove defective in any respect,
+   You (not any Contributor) assume the cost of any necessary servicing,
+   repair, or correction. This disclaimer of warranty constitutes an essential
+   part of this License. No use of  any Covered Software is authorized under
+   this License except under this disclaimer.
+
+7. Limitation of Liability
+
+   Under no circumstances and under no legal theory, whether tort (including
+   negligence), contract, or otherwise, shall any Contributor, or anyone who
+   distributes Covered Software as permitted above, be liable to You for any
+   direct, indirect, special, incidental, or consequential damages of any
+   character including, without limitation, damages for lost profits, loss of
+   goodwill, work stoppage, computer failure or malfunction, or any and all
+   other commercial damages or losses, even if such party shall have been
+   informed of the possibility of such damages. This limitation of liability
+   shall not apply to liability for death or personal injury resulting from
+   such party's negligence to the extent applicable law prohibits such
+   limitation. Some jurisdictions do not allow the exclusion or limitation of
+   incidental or consequential damages, so this exclusion and limitation may
+   not apply to You.
+
+8. Litigation
+
+   Any litigation relating to this License may be brought only in the courts
+   of a jurisdiction where the defendant maintains its principal place of
+   business and such litigation shall be governed by laws of that
+   jurisdiction, without reference to its conflict-of-law provisions. Nothing
+   in this Section shall prevent a party's ability to bring cross-claims or
+   counter-claims.
+
+9. Miscellaneous
+
+   This License represents the complete agreement concerning the subject
+   matter hereof. If any provision of this License is held to be
+   unenforceable, such provision shall be reformed only to the extent
+   necessary to make it enforceable. Any law or regulation which provides that
+   the language of a contract shall be construed against the drafter shall not
+   be used to construe this License against a Contributor.
+
+
+10. Versions of the License
+
+10.1. New Versions
+
+      Mozilla Foundation is the license steward. Except as provided in Section
+      10.3, no one other than the license steward has the right to modify or
+      publish new versions of this License. Each version will be given a
+      distinguishing version number.
+
+10.2. Effect of New Versions
+
+      You may distribute the Covered Software under the terms of the version
+      of the License under which You originally received the Covered Software,
+      or under the terms of any subsequent version published by the license
+      steward.
+
+10.3. Modified Versions
+
+      If you create software not governed by this License, and you want to
+      create a new license for such software, you may create and use a
+      modified version of this License if you rename the license and remove
+      any references to the name of the license steward (except to note that
+      such modified license differs from this License).
+
+10.4. Distributing Source Code Form that is Incompatible With Secondary
+      Licenses If You choose to distribute Source Code Form that is
+      Incompatible With Secondary Licenses under the terms of this version of
+      the License, the notice described in Exhibit B of this License must be
+      attached.
+
+Exhibit A - Source Code Form License Notice
+
+      This Source Code Form is subject to the
+      terms of the Mozilla Public License, v.
+      2.0. If a copy of the MPL was not
+      distributed with this file, You can
+      obtain one at
+      https://mozilla.org/MPL/2.0/.
+
+If it is not possible or desirable to put the notice in a particular file,
+then You may include the notice in a location (such as a LICENSE file in a
+relevant directory) where a recipient would be likely to look for such a
+notice.
+
+You may add additional accurate notices of copyright ownership.
+
+Exhibit B - "Incompatible With Secondary Licenses" Notice
+
+      This Source Code Form is "Incompatible
+      With Secondary Licenses", as defined by
+      the Mozilla Public License, v. 2.0.
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__init__.py
new file mode 100644
index 00000000..4f612bd5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__init__.py
@@ -0,0 +1,34 @@
+from functools import reduce
+from typing import Any, Callable, Dict
+
+from . import formats
+from .error_reporting import detailed_errors, ValidationError
+from .extra_validations import EXTRA_VALIDATIONS
+from .fastjsonschema_exceptions import JsonSchemaException, JsonSchemaValueException
+from .fastjsonschema_validations import validate as _validate
+
+__all__ = [
+    "validate",
+    "FORMAT_FUNCTIONS",
+    "EXTRA_VALIDATIONS",
+    "ValidationError",
+    "JsonSchemaException",
+    "JsonSchemaValueException",
+]
+
+
+FORMAT_FUNCTIONS: Dict[str, Callable[[str], bool]] = {
+    fn.__name__.replace("_", "-"): fn
+    for fn in formats.__dict__.values()
+    if callable(fn) and not fn.__name__.startswith("_")
+}
+
+
+def validate(data: Any) -> bool:
+    """Validate the given ``data`` object using JSON Schema
+    This function raises ``ValidationError`` if ``data`` is invalid.
+    """
+    with detailed_errors():
+        _validate(data, custom_formats=FORMAT_FUNCTIONS)
+        reduce(lambda acc, fn: fn(acc), EXTRA_VALIDATIONS, data)
+    return True
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..07e1e0de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/error_reporting.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/error_reporting.cpython-312.pyc
new file mode 100644
index 00000000..29c0e419
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/error_reporting.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/extra_validations.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/extra_validations.cpython-312.pyc
new file mode 100644
index 00000000..6a53807e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/extra_validations.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_exceptions.cpython-312.pyc
new file mode 100644
index 00000000..d0eae552
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_validations.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_validations.cpython-312.pyc
new file mode 100644
index 00000000..08710471
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/fastjsonschema_validations.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/formats.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/formats.cpython-312.pyc
new file mode 100644
index 00000000..d37e125a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/__pycache__/formats.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/error_reporting.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/error_reporting.py
new file mode 100644
index 00000000..3591231c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/error_reporting.py
@@ -0,0 +1,336 @@
+import io
+import json
+import logging
+import os
+import re
+import typing
+from contextlib import contextmanager
+from textwrap import indent, wrap
+from typing import Any, Dict, Generator, Iterator, List, Optional, Sequence, Union
+
+from .fastjsonschema_exceptions import JsonSchemaValueException
+
+if typing.TYPE_CHECKING:
+    import sys
+
+    if sys.version_info < (3, 11):
+        from typing_extensions import Self
+    else:
+        from typing import Self
+
+_logger = logging.getLogger(__name__)
+
+_MESSAGE_REPLACEMENTS = {
+    "must be named by propertyName definition": "keys must be named by",
+    "one of contains definition": "at least one item that matches",
+    " same as const definition:": "",
+    "only specified items": "only items matching the definition",
+}
+
+_SKIP_DETAILS = (
+    "must not be empty",
+    "is always invalid",
+    "must not be there",
+)
+
+_NEED_DETAILS = {"anyOf", "oneOf", "allOf", "contains", "propertyNames", "not", "items"}
+
+_CAMEL_CASE_SPLITTER = re.compile(r"\W+|([A-Z][^A-Z\W]*)")
+_IDENTIFIER = re.compile(r"^[\w_]+$", re.I)
+
+_TOML_JARGON = {
+    "object": "table",
+    "property": "key",
+    "properties": "keys",
+    "property names": "keys",
+}
+
+_FORMATS_HELP = """
+For more details about `format` see
+https://validate-pyproject.readthedocs.io/en/latest/api/validate_pyproject.formats.html
+"""
+
+
+class ValidationError(JsonSchemaValueException):
+    """Report violations of a given JSON schema.
+
+    This class extends :exc:`~fastjsonschema.JsonSchemaValueException`
+    by adding the following properties:
+
+    - ``summary``: an improved version of the ``JsonSchemaValueException`` error message
+      with only the necessary information)
+
+    - ``details``: more contextual information about the error like the failing schema
+      itself and the value that violates the schema.
+
+    Depending on the level of the verbosity of the ``logging`` configuration
+    the exception message will be only ``summary`` (default) or a combination of
+    ``summary`` and ``details`` (when the logging level is set to :obj:`logging.DEBUG`).
+    """
+
+    summary = ""
+    details = ""
+    _original_message = ""
+
+    @classmethod
+    def _from_jsonschema(cls, ex: JsonSchemaValueException) -> "Self":
+        formatter = _ErrorFormatting(ex)
+        obj = cls(str(formatter), ex.value, formatter.name, ex.definition, ex.rule)
+        debug_code = os.getenv("JSONSCHEMA_DEBUG_CODE_GENERATION", "false").lower()
+        if debug_code != "false":  # pragma: no cover
+            obj.__cause__, obj.__traceback__ = ex.__cause__, ex.__traceback__
+        obj._original_message = ex.message
+        obj.summary = formatter.summary
+        obj.details = formatter.details
+        return obj
+
+
+@contextmanager
+def detailed_errors() -> Generator[None, None, None]:
+    try:
+        yield
+    except JsonSchemaValueException as ex:
+        raise ValidationError._from_jsonschema(ex) from None
+
+
+class _ErrorFormatting:
+    def __init__(self, ex: JsonSchemaValueException):
+        self.ex = ex
+        self.name = f"`{self._simplify_name(ex.name)}`"
+        self._original_message: str = self.ex.message.replace(ex.name, self.name)
+        self._summary = ""
+        self._details = ""
+
+    def __str__(self) -> str:
+        if _logger.getEffectiveLevel() <= logging.DEBUG and self.details:
+            return f"{self.summary}\n\n{self.details}"
+
+        return self.summary
+
+    @property
+    def summary(self) -> str:
+        if not self._summary:
+            self._summary = self._expand_summary()
+
+        return self._summary
+
+    @property
+    def details(self) -> str:
+        if not self._details:
+            self._details = self._expand_details()
+
+        return self._details
+
+    @staticmethod
+    def _simplify_name(name: str) -> str:
+        x = len("data.")
+        return name[x:] if name.startswith("data.") else name
+
+    def _expand_summary(self) -> str:
+        msg = self._original_message
+
+        for bad, repl in _MESSAGE_REPLACEMENTS.items():
+            msg = msg.replace(bad, repl)
+
+        if any(substring in msg for substring in _SKIP_DETAILS):
+            return msg
+
+        schema = self.ex.rule_definition
+        if self.ex.rule in _NEED_DETAILS and schema:
+            summary = _SummaryWriter(_TOML_JARGON)
+            return f"{msg}:\n\n{indent(summary(schema), '    ')}"
+
+        return msg
+
+    def _expand_details(self) -> str:
+        optional = []
+        definition = self.ex.definition or {}
+        desc_lines = definition.pop("$$description", [])
+        desc = definition.pop("description", None) or " ".join(desc_lines)
+        if desc:
+            description = "\n".join(
+                wrap(
+                    desc,
+                    width=80,
+                    initial_indent="    ",
+                    subsequent_indent="    ",
+                    break_long_words=False,
+                )
+            )
+            optional.append(f"DESCRIPTION:\n{description}")
+        schema = json.dumps(definition, indent=4)
+        value = json.dumps(self.ex.value, indent=4)
+        defaults = [
+            f"GIVEN VALUE:\n{indent(value, '    ')}",
+            f"OFFENDING RULE: {self.ex.rule!r}",
+            f"DEFINITION:\n{indent(schema, '    ')}",
+        ]
+        msg = "\n\n".join(optional + defaults)
+        epilog = f"\n{_FORMATS_HELP}" if "format" in msg.lower() else ""
+        return msg + epilog
+
+
+class _SummaryWriter:
+    _IGNORE = frozenset(("description", "default", "title", "examples"))
+
+    def __init__(self, jargon: Optional[Dict[str, str]] = None):
+        self.jargon: Dict[str, str] = jargon or {}
+        # Clarify confusing terms
+        self._terms = {
+            "anyOf": "at least one of the following",
+            "oneOf": "exactly one of the following",
+            "allOf": "all of the following",
+            "not": "(*NOT* the following)",
+            "prefixItems": f"{self._jargon('items')} (in order)",
+            "items": "items",
+            "contains": "contains at least one of",
+            "propertyNames": (
+                f"non-predefined acceptable {self._jargon('property names')}"
+            ),
+            "patternProperties": f"{self._jargon('properties')} named via pattern",
+            "const": "predefined value",
+            "enum": "one of",
+        }
+        # Attributes that indicate that the definition is easy and can be done
+        # inline (e.g. string and number)
+        self._guess_inline_defs = [
+            "enum",
+            "const",
+            "maxLength",
+            "minLength",
+            "pattern",
+            "format",
+            "minimum",
+            "maximum",
+            "exclusiveMinimum",
+            "exclusiveMaximum",
+            "multipleOf",
+        ]
+
+    def _jargon(self, term: Union[str, List[str]]) -> Union[str, List[str]]:
+        if isinstance(term, list):
+            return [self.jargon.get(t, t) for t in term]
+        return self.jargon.get(term, term)
+
+    def __call__(
+        self,
+        schema: Union[dict, List[dict]],
+        prefix: str = "",
+        *,
+        _path: Sequence[str] = (),
+    ) -> str:
+        if isinstance(schema, list):
+            return self._handle_list(schema, prefix, _path)
+
+        filtered = self._filter_unecessary(schema, _path)
+        simple = self._handle_simple_dict(filtered, _path)
+        if simple:
+            return f"{prefix}{simple}"
+
+        child_prefix = self._child_prefix(prefix, "  ")
+        item_prefix = self._child_prefix(prefix, "- ")
+        indent = len(prefix) * " "
+        with io.StringIO() as buffer:
+            for i, (key, value) in enumerate(filtered.items()):
+                child_path = [*_path, key]
+                line_prefix = prefix if i == 0 else indent
+                buffer.write(f"{line_prefix}{self._label(child_path)}:")
+                # ^  just the first item should receive the complete prefix
+                if isinstance(value, dict):
+                    filtered = self._filter_unecessary(value, child_path)
+                    simple = self._handle_simple_dict(filtered, child_path)
+                    buffer.write(
+                        f" {simple}"
+                        if simple
+                        else f"\n{self(value, child_prefix, _path=child_path)}"
+                    )
+                elif isinstance(value, list) and (
+                    key != "type" or self._is_property(child_path)
+                ):
+                    children = self._handle_list(value, item_prefix, child_path)
+                    sep = " " if children.startswith("[") else "\n"
+                    buffer.write(f"{sep}{children}")
+                else:
+                    buffer.write(f" {self._value(value, child_path)}\n")
+            return buffer.getvalue()
+
+    def _is_unecessary(self, path: Sequence[str]) -> bool:
+        if self._is_property(path) or not path:  # empty path => instruction @ root
+            return False
+        key = path[-1]
+        return any(key.startswith(k) for k in "$_") or key in self._IGNORE
+
+    def _filter_unecessary(
+        self, schema: Dict[str, Any], path: Sequence[str]
+    ) -> Dict[str, Any]:
+        return {
+            key: value
+            for key, value in schema.items()
+            if not self._is_unecessary([*path, key])
+        }
+
+    def _handle_simple_dict(self, value: dict, path: Sequence[str]) -> Optional[str]:
+        inline = any(p in value for p in self._guess_inline_defs)
+        simple = not any(isinstance(v, (list, dict)) for v in value.values())
+        if inline or simple:
+            return f"{{{', '.join(self._inline_attrs(value, path))}}}\n"
+        return None
+
+    def _handle_list(
+        self, schemas: list, prefix: str = "", path: Sequence[str] = ()
+    ) -> str:
+        if self._is_unecessary(path):
+            return ""
+
+        repr_ = repr(schemas)
+        if all(not isinstance(e, (dict, list)) for e in schemas) and len(repr_) < 60:
+            return f"{repr_}\n"
+
+        item_prefix = self._child_prefix(prefix, "- ")
+        return "".join(
+            self(v, item_prefix, _path=[*path, f"[{i}]"]) for i, v in enumerate(schemas)
+        )
+
+    def _is_property(self, path: Sequence[str]) -> bool:
+        """Check if the given path can correspond to an arbitrarily named property"""
+        counter = 0
+        for key in path[-2::-1]:
+            if key not in {"properties", "patternProperties"}:
+                break
+            counter += 1
+
+        # If the counter if even, the path correspond to a JSON Schema keyword
+        # otherwise it can be any arbitrary string naming a property
+        return counter % 2 == 1
+
+    def _label(self, path: Sequence[str]) -> str:
+        *parents, key = path
+        if not self._is_property(path):
+            norm_key = _separate_terms(key)
+            return self._terms.get(key) or " ".join(self._jargon(norm_key))
+
+        if parents[-1] == "patternProperties":
+            return f"(regex {key!r})"
+        return repr(key)  # property name
+
+    def _value(self, value: Any, path: Sequence[str]) -> str:
+        if path[-1] == "type" and not self._is_property(path):
+            type_ = self._jargon(value)
+            return f"[{', '.join(type_)}]" if isinstance(type_, list) else type_
+        return repr(value)
+
+    def _inline_attrs(self, schema: dict, path: Sequence[str]) -> Iterator[str]:
+        for key, value in schema.items():
+            child_path = [*path, key]
+            yield f"{self._label(child_path)}: {self._value(value, child_path)}"
+
+    def _child_prefix(self, parent_prefix: str, child_prefix: str) -> str:
+        return len(parent_prefix) * " " + child_prefix
+
+
+def _separate_terms(word: str) -> List[str]:
+    """
+    >>> _separate_terms("FooBar-foo")
+    ['foo', 'bar', 'foo']
+    """
+    return [w.lower() for w in _CAMEL_CASE_SPLITTER.split(word) if w]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/extra_validations.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/extra_validations.py
new file mode 100644
index 00000000..c4ffe651
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/extra_validations.py
@@ -0,0 +1,52 @@
+"""The purpose of this module is implement PEP 621 validations that are
+difficult to express as a JSON Schema (or that are not supported by the current
+JSON Schema library).
+"""
+
+from inspect import cleandoc
+from typing import Mapping, TypeVar
+
+from .error_reporting import ValidationError
+
+T = TypeVar("T", bound=Mapping)
+
+
+class RedefiningStaticFieldAsDynamic(ValidationError):
+    _DESC = """According to PEP 621:
+
+    Build back-ends MUST raise an error if the metadata specifies a field
+    statically as well as being listed in dynamic.
+    """
+    __doc__ = _DESC
+    _URL = (
+        "https://packaging.python.org/en/latest/specifications/"
+        "pyproject-toml/#dynamic"
+    )
+
+
+def validate_project_dynamic(pyproject: T) -> T:
+    project_table = pyproject.get("project", {})
+    dynamic = project_table.get("dynamic", [])
+
+    for field in dynamic:
+        if field in project_table:
+            raise RedefiningStaticFieldAsDynamic(
+                message=f"You cannot provide a value for `project.{field}` and "
+                "list it under `project.dynamic` at the same time",
+                value={
+                    field: project_table[field],
+                    "...": " # ...",
+                    "dynamic": dynamic,
+                },
+                name=f"data.project.{field}",
+                definition={
+                    "description": cleandoc(RedefiningStaticFieldAsDynamic._DESC),
+                    "see": RedefiningStaticFieldAsDynamic._URL,
+                },
+                rule="PEP 621",
+            )
+
+    return pyproject
+
+
+EXTRA_VALIDATIONS = (validate_project_dynamic,)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_exceptions.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_exceptions.py
new file mode 100644
index 00000000..d2dddd6a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_exceptions.py
@@ -0,0 +1,51 @@
+import re
+
+
+SPLIT_RE = re.compile(r'[\.\[\]]+')
+
+
+class JsonSchemaException(ValueError):
+    """
+    Base exception of ``fastjsonschema`` library.
+    """
+
+
+class JsonSchemaValueException(JsonSchemaException):
+    """
+    Exception raised by validation function. Available properties:
+
+     * ``message`` containing human-readable information what is wrong (e.g. ``data.property[index] must be smaller than or equal to 42``),
+     * invalid ``value`` (e.g. ``60``),
+     * ``name`` of a path in the data structure (e.g. ``data.property[index]``),
+     * ``path`` as an array in the data structure (e.g. ``['data', 'property', 'index']``),
+     * the whole ``definition`` which the ``value`` has to fulfil (e.g. ``{'type': 'number', 'maximum': 42}``),
+     * ``rule`` which the ``value`` is breaking (e.g. ``maximum``)
+     * and ``rule_definition`` (e.g. ``42``).
+
+    .. versionchanged:: 2.14.0
+        Added all extra properties.
+    """
+
+    def __init__(self, message, value=None, name=None, definition=None, rule=None):
+        super().__init__(message)
+        self.message = message
+        self.value = value
+        self.name = name
+        self.definition = definition
+        self.rule = rule
+
+    @property
+    def path(self):
+        return [item for item in SPLIT_RE.split(self.name) if item != '']
+
+    @property
+    def rule_definition(self):
+        if not self.rule or not self.definition:
+            return None
+        return self.definition.get(self.rule)
+
+
+class JsonSchemaDefinitionException(JsonSchemaException):
+    """
+    Exception raised by generator of validation function.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_validations.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_validations.py
new file mode 100644
index 00000000..42e7aa5e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/fastjsonschema_validations.py
@@ -0,0 +1,1319 @@
+# noqa
+# ruff: noqa
+# flake8: noqa
+# pylint: skip-file
+# mypy: ignore-errors
+# yapf: disable
+# pylama:skip=1
+
+
+# *** PLEASE DO NOT MODIFY DIRECTLY: Automatically generated code *** 
+
+
+VERSION = "2.20.0"
+from decimal import Decimal
+import re
+from .fastjsonschema_exceptions import JsonSchemaValueException
+
+
+REGEX_PATTERNS = {
+    '^.*$': re.compile('^.*$'),
+    '.+': re.compile('.+'),
+    '^.+$': re.compile('^.+$'),
+    'idn-email_re_pattern': re.compile('^[^@]+@[^@]+\\.[^@]+\\Z')
+}
+
+NoneType = type(None)
+
+def validate(data, custom_formats={}, name_prefix=None):
+    validate_https___packaging_python_org_en_latest_specifications_declaring_build_dependencies(data, custom_formats, (name_prefix or "data") + "")
+    return data
+
+def validate_https___packaging_python_org_en_latest_specifications_declaring_build_dependencies(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/declaring-build-dependencies/', 'title': 'Data structure for ``pyproject.toml`` files', '$$description': ['File format containing build-time configurations for the Python ecosystem. ', ':pep:`517` initially defined a build-system independent format for source trees', 'which was complemented by :pep:`518` to provide a way of specifying dependencies ', 'for building Python projects.', 'Please notice the ``project`` table (as initially defined in  :pep:`621`) is not included', 'in this schema and should be considered separately.'], 'type': 'object', 'additionalProperties': False, 'properties': {'build-system': {'type': 'object', 'description': 'Table used to store build-related data', 'additionalProperties': False, 'properties': {'requires': {'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, 'build-backend': {'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, 'backend-path': {'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}}, 'required': ['requires']}, 'project': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create command-line wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.']}, 'gui-scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create GUI wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.', 'The difference between ``scripts`` and ``gui-scripts`` is only relevant in', 'Windows.']}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$ref': '#/definitions/entry-point-group'}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$ref': '#/definitions/dependency'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$ref': '#/definitions/dependency'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}, 'tool': {'type': 'object', 'properties': {'distutils': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html', 'title': '``tool.distutils`` table', '$$description': ['**EXPERIMENTAL** (NOT OFFICIALLY SUPPORTED): Use ``tool.distutils``', 'subtables to configure arguments for ``distutils`` commands.', 'Originally, ``distutils`` allowed developers to configure arguments for', '``setup.py`` commands via `distutils configuration files', '<https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html>`_.', 'See also `the old Python docs <https://docs.python.org/3.11/install/>_`.'], 'type': 'object', 'properties': {'global': {'type': 'object', 'description': 'Global options applied to all ``distutils`` commands'}}, 'patternProperties': {'.+': {'type': 'object'}}, '$comment': 'TODO: Is there a practical way of making this schema more specific?'}, 'setuptools': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html', 'title': '``tool.setuptools`` table', '$$description': ['``setuptools``-specific configurations that can be set by users that require', 'customization.', 'These configurations are completely optional and probably can be skipped when', 'creating simple packages. They are equivalent to some of the `Keywords', '<https://setuptools.pypa.io/en/latest/references/keywords.html>`_', 'used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.', 'It considers only ``setuptools`` `parameters', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_', 'that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``', 'and ``setup_requires`` (incompatible with modern workflows/standards).'], 'type': 'object', 'additionalProperties': False, 'properties': {'platforms': {'type': 'array', 'items': {'type': 'string'}}, 'provides': {'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'obsoletes': {'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'zip-safe': {'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, 'script-files': {'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, 'eager-resources': {'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, 'packages': {'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$ref': '#/definitions/package-name'}}, {'$ref': '#/definitions/find-directive'}]}, 'package-dir': {'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$ref': '#/definitions/package-name'}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, 'package-data': {'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'include-package-data': {'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, 'exclude-package-data': {'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'namespace-packages': {'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, 'py-modules': {'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, 'ext-modules': {'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$ref': '#/definitions/ext-module'}}, 'data-files': {'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'cmdclass': {'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, 'license-files': {'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, 'dynamic': {'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'$ref': '#/definitions/attr-directive'}, {'$ref': '#/definitions/file-directive'}]}, 'classifiers': {'$ref': '#/definitions/file-directive'}, 'description': {'$ref': '#/definitions/file-directive'}, 'entry-points': {'$ref': '#/definitions/file-directive'}, 'dependencies': {'$ref': '#/definitions/file-directive-for-dependencies'}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'$ref': '#/definitions/file-directive-for-dependencies'}}}, 'readme': {'type': 'object', 'anyOf': [{'$ref': '#/definitions/file-directive'}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'$ref': '#/definitions/file-directive/properties/file'}}, 'additionalProperties': False}], 'required': ['file']}}}}, 'definitions': {'package-name': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, 'ext-module': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, 'file-directive': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'file-directive-for-dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'attr-directive': {'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, 'find-directive': {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}}}}}}, 'project': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create command-line wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.']}, 'gui-scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create GUI wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.', 'The difference between ``scripts`` and ``gui-scripts`` is only relevant in', 'Windows.']}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$ref': '#/definitions/entry-point-group'}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$ref': '#/definitions/dependency'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$ref': '#/definitions/dependency'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        if "build-system" in data_keys:
+            data_keys.remove("build-system")
+            data__buildsystem = data["build-system"]
+            if not isinstance(data__buildsystem, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system must be object", value=data__buildsystem, name="" + (name_prefix or "data") + ".build-system", definition={'type': 'object', 'description': 'Table used to store build-related data', 'additionalProperties': False, 'properties': {'requires': {'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, 'build-backend': {'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, 'backend-path': {'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}}, 'required': ['requires']}, rule='type')
+            data__buildsystem_is_dict = isinstance(data__buildsystem, dict)
+            if data__buildsystem_is_dict:
+                data__buildsystem__missing_keys = set(['requires']) - data__buildsystem.keys()
+                if data__buildsystem__missing_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system must contain " + (str(sorted(data__buildsystem__missing_keys)) + " properties"), value=data__buildsystem, name="" + (name_prefix or "data") + ".build-system", definition={'type': 'object', 'description': 'Table used to store build-related data', 'additionalProperties': False, 'properties': {'requires': {'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, 'build-backend': {'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, 'backend-path': {'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}}, 'required': ['requires']}, rule='required')
+                data__buildsystem_keys = set(data__buildsystem.keys())
+                if "requires" in data__buildsystem_keys:
+                    data__buildsystem_keys.remove("requires")
+                    data__buildsystem__requires = data__buildsystem["requires"]
+                    if not isinstance(data__buildsystem__requires, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.requires must be array", value=data__buildsystem__requires, name="" + (name_prefix or "data") + ".build-system.requires", definition={'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, rule='type')
+                    data__buildsystem__requires_is_list = isinstance(data__buildsystem__requires, (list, tuple))
+                    if data__buildsystem__requires_is_list:
+                        data__buildsystem__requires_len = len(data__buildsystem__requires)
+                        for data__buildsystem__requires_x, data__buildsystem__requires_item in enumerate(data__buildsystem__requires):
+                            if not isinstance(data__buildsystem__requires_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.requires[{data__buildsystem__requires_x}]".format(**locals()) + " must be string", value=data__buildsystem__requires_item, name="" + (name_prefix or "data") + ".build-system.requires[{data__buildsystem__requires_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if "build-backend" in data__buildsystem_keys:
+                    data__buildsystem_keys.remove("build-backend")
+                    data__buildsystem__buildbackend = data__buildsystem["build-backend"]
+                    if not isinstance(data__buildsystem__buildbackend, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.build-backend must be string", value=data__buildsystem__buildbackend, name="" + (name_prefix or "data") + ".build-system.build-backend", definition={'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, rule='type')
+                    if isinstance(data__buildsystem__buildbackend, str):
+                        if not custom_formats["pep517-backend-reference"](data__buildsystem__buildbackend):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.build-backend must be pep517-backend-reference", value=data__buildsystem__buildbackend, name="" + (name_prefix or "data") + ".build-system.build-backend", definition={'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, rule='format')
+                if "backend-path" in data__buildsystem_keys:
+                    data__buildsystem_keys.remove("backend-path")
+                    data__buildsystem__backendpath = data__buildsystem["backend-path"]
+                    if not isinstance(data__buildsystem__backendpath, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.backend-path must be array", value=data__buildsystem__backendpath, name="" + (name_prefix or "data") + ".build-system.backend-path", definition={'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}, rule='type')
+                    data__buildsystem__backendpath_is_list = isinstance(data__buildsystem__backendpath, (list, tuple))
+                    if data__buildsystem__backendpath_is_list:
+                        data__buildsystem__backendpath_len = len(data__buildsystem__backendpath)
+                        for data__buildsystem__backendpath_x, data__buildsystem__backendpath_item in enumerate(data__buildsystem__backendpath):
+                            if not isinstance(data__buildsystem__backendpath_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system.backend-path[{data__buildsystem__backendpath_x}]".format(**locals()) + " must be string", value=data__buildsystem__backendpath_item, name="" + (name_prefix or "data") + ".build-system.backend-path[{data__buildsystem__backendpath_x}]".format(**locals()) + "", definition={'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}, rule='type')
+                if data__buildsystem_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".build-system must not contain "+str(data__buildsystem_keys)+" properties", value=data__buildsystem, name="" + (name_prefix or "data") + ".build-system", definition={'type': 'object', 'description': 'Table used to store build-related data', 'additionalProperties': False, 'properties': {'requires': {'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, 'build-backend': {'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, 'backend-path': {'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}}, 'required': ['requires']}, rule='additionalProperties')
+        if "project" in data_keys:
+            data_keys.remove("project")
+            data__project = data["project"]
+            validate_https___packaging_python_org_en_latest_specifications_pyproject_toml(data__project, custom_formats, (name_prefix or "data") + ".project")
+        if "tool" in data_keys:
+            data_keys.remove("tool")
+            data__tool = data["tool"]
+            if not isinstance(data__tool, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".tool must be object", value=data__tool, name="" + (name_prefix or "data") + ".tool", definition={'type': 'object', 'properties': {'distutils': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html', 'title': '``tool.distutils`` table', '$$description': ['**EXPERIMENTAL** (NOT OFFICIALLY SUPPORTED): Use ``tool.distutils``', 'subtables to configure arguments for ``distutils`` commands.', 'Originally, ``distutils`` allowed developers to configure arguments for', '``setup.py`` commands via `distutils configuration files', '<https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html>`_.', 'See also `the old Python docs <https://docs.python.org/3.11/install/>_`.'], 'type': 'object', 'properties': {'global': {'type': 'object', 'description': 'Global options applied to all ``distutils`` commands'}}, 'patternProperties': {'.+': {'type': 'object'}}, '$comment': 'TODO: Is there a practical way of making this schema more specific?'}, 'setuptools': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html', 'title': '``tool.setuptools`` table', '$$description': ['``setuptools``-specific configurations that can be set by users that require', 'customization.', 'These configurations are completely optional and probably can be skipped when', 'creating simple packages. They are equivalent to some of the `Keywords', '<https://setuptools.pypa.io/en/latest/references/keywords.html>`_', 'used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.', 'It considers only ``setuptools`` `parameters', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_', 'that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``', 'and ``setup_requires`` (incompatible with modern workflows/standards).'], 'type': 'object', 'additionalProperties': False, 'properties': {'platforms': {'type': 'array', 'items': {'type': 'string'}}, 'provides': {'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'obsoletes': {'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'zip-safe': {'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, 'script-files': {'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, 'eager-resources': {'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, 'packages': {'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$ref': '#/definitions/package-name'}}, {'$ref': '#/definitions/find-directive'}]}, 'package-dir': {'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$ref': '#/definitions/package-name'}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, 'package-data': {'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'include-package-data': {'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, 'exclude-package-data': {'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'namespace-packages': {'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, 'py-modules': {'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, 'ext-modules': {'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$ref': '#/definitions/ext-module'}}, 'data-files': {'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'cmdclass': {'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, 'license-files': {'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, 'dynamic': {'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'$ref': '#/definitions/attr-directive'}, {'$ref': '#/definitions/file-directive'}]}, 'classifiers': {'$ref': '#/definitions/file-directive'}, 'description': {'$ref': '#/definitions/file-directive'}, 'entry-points': {'$ref': '#/definitions/file-directive'}, 'dependencies': {'$ref': '#/definitions/file-directive-for-dependencies'}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'$ref': '#/definitions/file-directive-for-dependencies'}}}, 'readme': {'type': 'object', 'anyOf': [{'$ref': '#/definitions/file-directive'}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'$ref': '#/definitions/file-directive/properties/file'}}, 'additionalProperties': False}], 'required': ['file']}}}}, 'definitions': {'package-name': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, 'ext-module': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, 'file-directive': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'file-directive-for-dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'attr-directive': {'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, 'find-directive': {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}}}}}, rule='type')
+            data__tool_is_dict = isinstance(data__tool, dict)
+            if data__tool_is_dict:
+                data__tool_keys = set(data__tool.keys())
+                if "distutils" in data__tool_keys:
+                    data__tool_keys.remove("distutils")
+                    data__tool__distutils = data__tool["distutils"]
+                    validate_https___setuptools_pypa_io_en_latest_deprecated_distutils_configfile_html(data__tool__distutils, custom_formats, (name_prefix or "data") + ".tool.distutils")
+                if "setuptools" in data__tool_keys:
+                    data__tool_keys.remove("setuptools")
+                    data__tool__setuptools = data__tool["setuptools"]
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html(data__tool__setuptools, custom_formats, (name_prefix or "data") + ".tool.setuptools")
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/declaring-build-dependencies/', 'title': 'Data structure for ``pyproject.toml`` files', '$$description': ['File format containing build-time configurations for the Python ecosystem. ', ':pep:`517` initially defined a build-system independent format for source trees', 'which was complemented by :pep:`518` to provide a way of specifying dependencies ', 'for building Python projects.', 'Please notice the ``project`` table (as initially defined in  :pep:`621`) is not included', 'in this schema and should be considered separately.'], 'type': 'object', 'additionalProperties': False, 'properties': {'build-system': {'type': 'object', 'description': 'Table used to store build-related data', 'additionalProperties': False, 'properties': {'requires': {'type': 'array', '$$description': ['List of dependencies in the :pep:`508` format required to execute the build', 'system. Please notice that the resulting dependency graph', '**MUST NOT contain cycles**'], 'items': {'type': 'string'}}, 'build-backend': {'type': 'string', 'description': 'Python object that will be used to perform the build according to :pep:`517`', 'format': 'pep517-backend-reference'}, 'backend-path': {'type': 'array', '$$description': ['List of directories to be prepended to ``sys.path`` when loading the', 'back-end, and running its hooks'], 'items': {'type': 'string', '$comment': 'Should be a path (TODO: enforce it with format?)'}}}, 'required': ['requires']}, 'project': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create command-line wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.']}, 'gui-scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create GUI wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.', 'The difference between ``scripts`` and ``gui-scripts`` is only relevant in', 'Windows.']}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$ref': '#/definitions/entry-point-group'}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$ref': '#/definitions/dependency'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$ref': '#/definitions/dependency'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}, 'tool': {'type': 'object', 'properties': {'distutils': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html', 'title': '``tool.distutils`` table', '$$description': ['**EXPERIMENTAL** (NOT OFFICIALLY SUPPORTED): Use ``tool.distutils``', 'subtables to configure arguments for ``distutils`` commands.', 'Originally, ``distutils`` allowed developers to configure arguments for', '``setup.py`` commands via `distutils configuration files', '<https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html>`_.', 'See also `the old Python docs <https://docs.python.org/3.11/install/>_`.'], 'type': 'object', 'properties': {'global': {'type': 'object', 'description': 'Global options applied to all ``distutils`` commands'}}, 'patternProperties': {'.+': {'type': 'object'}}, '$comment': 'TODO: Is there a practical way of making this schema more specific?'}, 'setuptools': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html', 'title': '``tool.setuptools`` table', '$$description': ['``setuptools``-specific configurations that can be set by users that require', 'customization.', 'These configurations are completely optional and probably can be skipped when', 'creating simple packages. They are equivalent to some of the `Keywords', '<https://setuptools.pypa.io/en/latest/references/keywords.html>`_', 'used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.', 'It considers only ``setuptools`` `parameters', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_', 'that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``', 'and ``setup_requires`` (incompatible with modern workflows/standards).'], 'type': 'object', 'additionalProperties': False, 'properties': {'platforms': {'type': 'array', 'items': {'type': 'string'}}, 'provides': {'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'obsoletes': {'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'zip-safe': {'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, 'script-files': {'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, 'eager-resources': {'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, 'packages': {'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$ref': '#/definitions/package-name'}}, {'$ref': '#/definitions/find-directive'}]}, 'package-dir': {'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$ref': '#/definitions/package-name'}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, 'package-data': {'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'include-package-data': {'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, 'exclude-package-data': {'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'namespace-packages': {'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, 'py-modules': {'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, 'ext-modules': {'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$ref': '#/definitions/ext-module'}}, 'data-files': {'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'cmdclass': {'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, 'license-files': {'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, 'dynamic': {'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'$ref': '#/definitions/attr-directive'}, {'$ref': '#/definitions/file-directive'}]}, 'classifiers': {'$ref': '#/definitions/file-directive'}, 'description': {'$ref': '#/definitions/file-directive'}, 'entry-points': {'$ref': '#/definitions/file-directive'}, 'dependencies': {'$ref': '#/definitions/file-directive-for-dependencies'}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'$ref': '#/definitions/file-directive-for-dependencies'}}}, 'readme': {'type': 'object', 'anyOf': [{'$ref': '#/definitions/file-directive'}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'$ref': '#/definitions/file-directive/properties/file'}}, 'additionalProperties': False}], 'required': ['file']}}}}, 'definitions': {'package-name': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, 'ext-module': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, 'file-directive': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'file-directive-for-dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'attr-directive': {'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, 'find-directive': {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}}}}}}, 'project': {'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$ref': '#/definitions/author'}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create command-line wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.']}, 'gui-scripts': {'$ref': '#/definitions/entry-point-group', '$$description': ['Instruct the installer to create GUI wrappers for the given', '`entry points <https://packaging.python.org/specifications/entry-points/>`_.', 'The difference between ``scripts`` and ``gui-scripts`` is only relevant in', 'Windows.']}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$ref': '#/definitions/entry-point-group'}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$ref': '#/definitions/dependency'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$ref': '#/definitions/dependency'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html', 'title': '``tool.setuptools`` table', '$$description': ['``setuptools``-specific configurations that can be set by users that require', 'customization.', 'These configurations are completely optional and probably can be skipped when', 'creating simple packages. They are equivalent to some of the `Keywords', '<https://setuptools.pypa.io/en/latest/references/keywords.html>`_', 'used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.', 'It considers only ``setuptools`` `parameters', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_', 'that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``', 'and ``setup_requires`` (incompatible with modern workflows/standards).'], 'type': 'object', 'additionalProperties': False, 'properties': {'platforms': {'type': 'array', 'items': {'type': 'string'}}, 'provides': {'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'obsoletes': {'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'zip-safe': {'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, 'script-files': {'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, 'eager-resources': {'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, 'packages': {'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}}, {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}]}, 'package-dir': {'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, 'package-data': {'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'include-package-data': {'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, 'exclude-package-data': {'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'namespace-packages': {'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, 'py-modules': {'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, 'ext-modules': {'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}}, 'data-files': {'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'cmdclass': {'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, 'license-files': {'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, 'dynamic': {'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'classifiers': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'description': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'entry-points': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, 'readme': {'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}}}}, 'definitions': {'package-name': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, 'ext-module': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, 'file-directive': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'file-directive-for-dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'attr-directive': {'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, 'find-directive': {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        if "platforms" in data_keys:
+            data_keys.remove("platforms")
+            data__platforms = data["platforms"]
+            if not isinstance(data__platforms, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".platforms must be array", value=data__platforms, name="" + (name_prefix or "data") + ".platforms", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__platforms_is_list = isinstance(data__platforms, (list, tuple))
+            if data__platforms_is_list:
+                data__platforms_len = len(data__platforms)
+                for data__platforms_x, data__platforms_item in enumerate(data__platforms):
+                    if not isinstance(data__platforms_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".platforms[{data__platforms_x}]".format(**locals()) + " must be string", value=data__platforms_item, name="" + (name_prefix or "data") + ".platforms[{data__platforms_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "provides" in data_keys:
+            data_keys.remove("provides")
+            data__provides = data["provides"]
+            if not isinstance(data__provides, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".provides must be array", value=data__provides, name="" + (name_prefix or "data") + ".provides", definition={'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, rule='type')
+            data__provides_is_list = isinstance(data__provides, (list, tuple))
+            if data__provides_is_list:
+                data__provides_len = len(data__provides)
+                for data__provides_x, data__provides_item in enumerate(data__provides):
+                    if not isinstance(data__provides_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".provides[{data__provides_x}]".format(**locals()) + " must be string", value=data__provides_item, name="" + (name_prefix or "data") + ".provides[{data__provides_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='type')
+                    if isinstance(data__provides_item, str):
+                        if not custom_formats["pep508-identifier"](data__provides_item):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".provides[{data__provides_x}]".format(**locals()) + " must be pep508-identifier", value=data__provides_item, name="" + (name_prefix or "data") + ".provides[{data__provides_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='format')
+        if "obsoletes" in data_keys:
+            data_keys.remove("obsoletes")
+            data__obsoletes = data["obsoletes"]
+            if not isinstance(data__obsoletes, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".obsoletes must be array", value=data__obsoletes, name="" + (name_prefix or "data") + ".obsoletes", definition={'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, rule='type')
+            data__obsoletes_is_list = isinstance(data__obsoletes, (list, tuple))
+            if data__obsoletes_is_list:
+                data__obsoletes_len = len(data__obsoletes)
+                for data__obsoletes_x, data__obsoletes_item in enumerate(data__obsoletes):
+                    if not isinstance(data__obsoletes_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".obsoletes[{data__obsoletes_x}]".format(**locals()) + " must be string", value=data__obsoletes_item, name="" + (name_prefix or "data") + ".obsoletes[{data__obsoletes_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='type')
+                    if isinstance(data__obsoletes_item, str):
+                        if not custom_formats["pep508-identifier"](data__obsoletes_item):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".obsoletes[{data__obsoletes_x}]".format(**locals()) + " must be pep508-identifier", value=data__obsoletes_item, name="" + (name_prefix or "data") + ".obsoletes[{data__obsoletes_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='format')
+        if "zip-safe" in data_keys:
+            data_keys.remove("zip-safe")
+            data__zipsafe = data["zip-safe"]
+            if not isinstance(data__zipsafe, (bool)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".zip-safe must be boolean", value=data__zipsafe, name="" + (name_prefix or "data") + ".zip-safe", definition={'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, rule='type')
+        if "script-files" in data_keys:
+            data_keys.remove("script-files")
+            data__scriptfiles = data["script-files"]
+            if not isinstance(data__scriptfiles, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".script-files must be array", value=data__scriptfiles, name="" + (name_prefix or "data") + ".script-files", definition={'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, rule='type')
+            data__scriptfiles_is_list = isinstance(data__scriptfiles, (list, tuple))
+            if data__scriptfiles_is_list:
+                data__scriptfiles_len = len(data__scriptfiles)
+                for data__scriptfiles_x, data__scriptfiles_item in enumerate(data__scriptfiles):
+                    if not isinstance(data__scriptfiles_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".script-files[{data__scriptfiles_x}]".format(**locals()) + " must be string", value=data__scriptfiles_item, name="" + (name_prefix or "data") + ".script-files[{data__scriptfiles_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "eager-resources" in data_keys:
+            data_keys.remove("eager-resources")
+            data__eagerresources = data["eager-resources"]
+            if not isinstance(data__eagerresources, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".eager-resources must be array", value=data__eagerresources, name="" + (name_prefix or "data") + ".eager-resources", definition={'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__eagerresources_is_list = isinstance(data__eagerresources, (list, tuple))
+            if data__eagerresources_is_list:
+                data__eagerresources_len = len(data__eagerresources)
+                for data__eagerresources_x, data__eagerresources_item in enumerate(data__eagerresources):
+                    if not isinstance(data__eagerresources_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".eager-resources[{data__eagerresources_x}]".format(**locals()) + " must be string", value=data__eagerresources_item, name="" + (name_prefix or "data") + ".eager-resources[{data__eagerresources_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "packages" in data_keys:
+            data_keys.remove("packages")
+            data__packages = data["packages"]
+            data__packages_one_of_count1 = 0
+            if data__packages_one_of_count1 < 2:
+                try:
+                    if not isinstance(data__packages, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".packages must be array", value=data__packages, name="" + (name_prefix or "data") + ".packages", definition={'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}}, rule='type')
+                    data__packages_is_list = isinstance(data__packages, (list, tuple))
+                    if data__packages_is_list:
+                        data__packages_len = len(data__packages)
+                        for data__packages_x, data__packages_item in enumerate(data__packages):
+                            validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_package_name(data__packages_item, custom_formats, (name_prefix or "data") + ".packages[{data__packages_x}]".format(**locals()))
+                    data__packages_one_of_count1 += 1
+                except JsonSchemaValueException: pass
+            if data__packages_one_of_count1 < 2:
+                try:
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_find_directive(data__packages, custom_formats, (name_prefix or "data") + ".packages")
+                    data__packages_one_of_count1 += 1
+                except JsonSchemaValueException: pass
+            if data__packages_one_of_count1 != 1:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".packages must be valid exactly by one definition" + (" (" + str(data__packages_one_of_count1) + " matches found)"), value=data__packages, name="" + (name_prefix or "data") + ".packages", definition={'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}}, {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}]}, rule='oneOf')
+        if "package-dir" in data_keys:
+            data_keys.remove("package-dir")
+            data__packagedir = data["package-dir"]
+            if not isinstance(data__packagedir, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir must be object", value=data__packagedir, name="" + (name_prefix or "data") + ".package-dir", definition={'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, rule='type')
+            data__packagedir_is_dict = isinstance(data__packagedir, dict)
+            if data__packagedir_is_dict:
+                data__packagedir_keys = set(data__packagedir.keys())
+                for data__packagedir_key, data__packagedir_val in data__packagedir.items():
+                    if REGEX_PATTERNS['^.*$'].search(data__packagedir_key):
+                        if data__packagedir_key in data__packagedir_keys:
+                            data__packagedir_keys.remove(data__packagedir_key)
+                        if not isinstance(data__packagedir_val, (str)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir.{data__packagedir_key}".format(**locals()) + " must be string", value=data__packagedir_val, name="" + (name_prefix or "data") + ".package-dir.{data__packagedir_key}".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if data__packagedir_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir must not contain "+str(data__packagedir_keys)+" properties", value=data__packagedir, name="" + (name_prefix or "data") + ".package-dir", definition={'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, rule='additionalProperties')
+                data__packagedir_len = len(data__packagedir)
+                if data__packagedir_len != 0:
+                    data__packagedir_property_names = True
+                    for data__packagedir_key in data__packagedir:
+                        try:
+                            data__packagedir_key_any_of_count2 = 0
+                            if not data__packagedir_key_any_of_count2:
+                                try:
+                                    if data__packagedir_key != "":
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir must be same as const definition: ", value=data__packagedir_key, name="" + (name_prefix or "data") + ".package-dir", definition={'const': ''}, rule='const')
+                                    data__packagedir_key_any_of_count2 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__packagedir_key_any_of_count2:
+                                try:
+                                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_package_name(data__packagedir_key, custom_formats, (name_prefix or "data") + ".package-dir")
+                                    data__packagedir_key_any_of_count2 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__packagedir_key_any_of_count2:
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir cannot be validated by any definition", value=data__packagedir_key, name="" + (name_prefix or "data") + ".package-dir", definition={'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, rule='anyOf')
+                        except JsonSchemaValueException:
+                            data__packagedir_property_names = False
+                    if not data__packagedir_property_names:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-dir must be named by propertyName definition", value=data__packagedir, name="" + (name_prefix or "data") + ".package-dir", definition={'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, rule='propertyNames')
+        if "package-data" in data_keys:
+            data_keys.remove("package-data")
+            data__packagedata = data["package-data"]
+            if not isinstance(data__packagedata, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must be object", value=data__packagedata, name="" + (name_prefix or "data") + ".package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='type')
+            data__packagedata_is_dict = isinstance(data__packagedata, dict)
+            if data__packagedata_is_dict:
+                data__packagedata_keys = set(data__packagedata.keys())
+                for data__packagedata_key, data__packagedata_val in data__packagedata.items():
+                    if REGEX_PATTERNS['^.*$'].search(data__packagedata_key):
+                        if data__packagedata_key in data__packagedata_keys:
+                            data__packagedata_keys.remove(data__packagedata_key)
+                        if not isinstance(data__packagedata_val, (list, tuple)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data.{data__packagedata_key}".format(**locals()) + " must be array", value=data__packagedata_val, name="" + (name_prefix or "data") + ".package-data.{data__packagedata_key}".format(**locals()) + "", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+                        data__packagedata_val_is_list = isinstance(data__packagedata_val, (list, tuple))
+                        if data__packagedata_val_is_list:
+                            data__packagedata_val_len = len(data__packagedata_val)
+                            for data__packagedata_val_x, data__packagedata_val_item in enumerate(data__packagedata_val):
+                                if not isinstance(data__packagedata_val_item, (str)):
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data.{data__packagedata_key}[{data__packagedata_val_x}]".format(**locals()) + " must be string", value=data__packagedata_val_item, name="" + (name_prefix or "data") + ".package-data.{data__packagedata_key}[{data__packagedata_val_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if data__packagedata_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must not contain "+str(data__packagedata_keys)+" properties", value=data__packagedata, name="" + (name_prefix or "data") + ".package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='additionalProperties')
+                data__packagedata_len = len(data__packagedata)
+                if data__packagedata_len != 0:
+                    data__packagedata_property_names = True
+                    for data__packagedata_key in data__packagedata:
+                        try:
+                            data__packagedata_key_any_of_count3 = 0
+                            if not data__packagedata_key_any_of_count3:
+                                try:
+                                    if not isinstance(data__packagedata_key, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must be string", value=data__packagedata_key, name="" + (name_prefix or "data") + ".package-data", definition={'type': 'string', 'format': 'python-module-name'}, rule='type')
+                                    if isinstance(data__packagedata_key, str):
+                                        if not custom_formats["python-module-name"](data__packagedata_key):
+                                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must be python-module-name", value=data__packagedata_key, name="" + (name_prefix or "data") + ".package-data", definition={'type': 'string', 'format': 'python-module-name'}, rule='format')
+                                    data__packagedata_key_any_of_count3 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__packagedata_key_any_of_count3:
+                                try:
+                                    if data__packagedata_key != "*":
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must be same as const definition: *", value=data__packagedata_key, name="" + (name_prefix or "data") + ".package-data", definition={'const': '*'}, rule='const')
+                                    data__packagedata_key_any_of_count3 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__packagedata_key_any_of_count3:
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data cannot be validated by any definition", value=data__packagedata_key, name="" + (name_prefix or "data") + ".package-data", definition={'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, rule='anyOf')
+                        except JsonSchemaValueException:
+                            data__packagedata_property_names = False
+                    if not data__packagedata_property_names:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".package-data must be named by propertyName definition", value=data__packagedata, name="" + (name_prefix or "data") + ".package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='propertyNames')
+        if "include-package-data" in data_keys:
+            data_keys.remove("include-package-data")
+            data__includepackagedata = data["include-package-data"]
+            if not isinstance(data__includepackagedata, (bool)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".include-package-data must be boolean", value=data__includepackagedata, name="" + (name_prefix or "data") + ".include-package-data", definition={'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, rule='type')
+        if "exclude-package-data" in data_keys:
+            data_keys.remove("exclude-package-data")
+            data__excludepackagedata = data["exclude-package-data"]
+            if not isinstance(data__excludepackagedata, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must be object", value=data__excludepackagedata, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='type')
+            data__excludepackagedata_is_dict = isinstance(data__excludepackagedata, dict)
+            if data__excludepackagedata_is_dict:
+                data__excludepackagedata_keys = set(data__excludepackagedata.keys())
+                for data__excludepackagedata_key, data__excludepackagedata_val in data__excludepackagedata.items():
+                    if REGEX_PATTERNS['^.*$'].search(data__excludepackagedata_key):
+                        if data__excludepackagedata_key in data__excludepackagedata_keys:
+                            data__excludepackagedata_keys.remove(data__excludepackagedata_key)
+                        if not isinstance(data__excludepackagedata_val, (list, tuple)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data.{data__excludepackagedata_key}".format(**locals()) + " must be array", value=data__excludepackagedata_val, name="" + (name_prefix or "data") + ".exclude-package-data.{data__excludepackagedata_key}".format(**locals()) + "", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+                        data__excludepackagedata_val_is_list = isinstance(data__excludepackagedata_val, (list, tuple))
+                        if data__excludepackagedata_val_is_list:
+                            data__excludepackagedata_val_len = len(data__excludepackagedata_val)
+                            for data__excludepackagedata_val_x, data__excludepackagedata_val_item in enumerate(data__excludepackagedata_val):
+                                if not isinstance(data__excludepackagedata_val_item, (str)):
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data.{data__excludepackagedata_key}[{data__excludepackagedata_val_x}]".format(**locals()) + " must be string", value=data__excludepackagedata_val_item, name="" + (name_prefix or "data") + ".exclude-package-data.{data__excludepackagedata_key}[{data__excludepackagedata_val_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if data__excludepackagedata_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must not contain "+str(data__excludepackagedata_keys)+" properties", value=data__excludepackagedata, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='additionalProperties')
+                data__excludepackagedata_len = len(data__excludepackagedata)
+                if data__excludepackagedata_len != 0:
+                    data__excludepackagedata_property_names = True
+                    for data__excludepackagedata_key in data__excludepackagedata:
+                        try:
+                            data__excludepackagedata_key_any_of_count4 = 0
+                            if not data__excludepackagedata_key_any_of_count4:
+                                try:
+                                    if not isinstance(data__excludepackagedata_key, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must be string", value=data__excludepackagedata_key, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'type': 'string', 'format': 'python-module-name'}, rule='type')
+                                    if isinstance(data__excludepackagedata_key, str):
+                                        if not custom_formats["python-module-name"](data__excludepackagedata_key):
+                                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must be python-module-name", value=data__excludepackagedata_key, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'type': 'string', 'format': 'python-module-name'}, rule='format')
+                                    data__excludepackagedata_key_any_of_count4 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__excludepackagedata_key_any_of_count4:
+                                try:
+                                    if data__excludepackagedata_key != "*":
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must be same as const definition: *", value=data__excludepackagedata_key, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'const': '*'}, rule='const')
+                                    data__excludepackagedata_key_any_of_count4 += 1
+                                except JsonSchemaValueException: pass
+                            if not data__excludepackagedata_key_any_of_count4:
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data cannot be validated by any definition", value=data__excludepackagedata_key, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, rule='anyOf')
+                        except JsonSchemaValueException:
+                            data__excludepackagedata_property_names = False
+                    if not data__excludepackagedata_property_names:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".exclude-package-data must be named by propertyName definition", value=data__excludepackagedata, name="" + (name_prefix or "data") + ".exclude-package-data", definition={'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='propertyNames')
+        if "namespace-packages" in data_keys:
+            data_keys.remove("namespace-packages")
+            data__namespacepackages = data["namespace-packages"]
+            if not isinstance(data__namespacepackages, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".namespace-packages must be array", value=data__namespacepackages, name="" + (name_prefix or "data") + ".namespace-packages", definition={'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, rule='type')
+            data__namespacepackages_is_list = isinstance(data__namespacepackages, (list, tuple))
+            if data__namespacepackages_is_list:
+                data__namespacepackages_len = len(data__namespacepackages)
+                for data__namespacepackages_x, data__namespacepackages_item in enumerate(data__namespacepackages):
+                    if not isinstance(data__namespacepackages_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".namespace-packages[{data__namespacepackages_x}]".format(**locals()) + " must be string", value=data__namespacepackages_item, name="" + (name_prefix or "data") + ".namespace-packages[{data__namespacepackages_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='type')
+                    if isinstance(data__namespacepackages_item, str):
+                        if not custom_formats["python-module-name-relaxed"](data__namespacepackages_item):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".namespace-packages[{data__namespacepackages_x}]".format(**locals()) + " must be python-module-name-relaxed", value=data__namespacepackages_item, name="" + (name_prefix or "data") + ".namespace-packages[{data__namespacepackages_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='format')
+        if "py-modules" in data_keys:
+            data_keys.remove("py-modules")
+            data__pymodules = data["py-modules"]
+            if not isinstance(data__pymodules, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".py-modules must be array", value=data__pymodules, name="" + (name_prefix or "data") + ".py-modules", definition={'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, rule='type')
+            data__pymodules_is_list = isinstance(data__pymodules, (list, tuple))
+            if data__pymodules_is_list:
+                data__pymodules_len = len(data__pymodules)
+                for data__pymodules_x, data__pymodules_item in enumerate(data__pymodules):
+                    if not isinstance(data__pymodules_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".py-modules[{data__pymodules_x}]".format(**locals()) + " must be string", value=data__pymodules_item, name="" + (name_prefix or "data") + ".py-modules[{data__pymodules_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='type')
+                    if isinstance(data__pymodules_item, str):
+                        if not custom_formats["python-module-name-relaxed"](data__pymodules_item):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".py-modules[{data__pymodules_x}]".format(**locals()) + " must be python-module-name-relaxed", value=data__pymodules_item, name="" + (name_prefix or "data") + ".py-modules[{data__pymodules_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='format')
+        if "ext-modules" in data_keys:
+            data_keys.remove("ext-modules")
+            data__extmodules = data["ext-modules"]
+            if not isinstance(data__extmodules, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".ext-modules must be array", value=data__extmodules, name="" + (name_prefix or "data") + ".ext-modules", definition={'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}}, rule='type')
+            data__extmodules_is_list = isinstance(data__extmodules, (list, tuple))
+            if data__extmodules_is_list:
+                data__extmodules_len = len(data__extmodules)
+                for data__extmodules_x, data__extmodules_item in enumerate(data__extmodules):
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_ext_module(data__extmodules_item, custom_formats, (name_prefix or "data") + ".ext-modules[{data__extmodules_x}]".format(**locals()))
+        if "data-files" in data_keys:
+            data_keys.remove("data-files")
+            data__datafiles = data["data-files"]
+            if not isinstance(data__datafiles, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".data-files must be object", value=data__datafiles, name="" + (name_prefix or "data") + ".data-files", definition={'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, rule='type')
+            data__datafiles_is_dict = isinstance(data__datafiles, dict)
+            if data__datafiles_is_dict:
+                data__datafiles_keys = set(data__datafiles.keys())
+                for data__datafiles_key, data__datafiles_val in data__datafiles.items():
+                    if REGEX_PATTERNS['^.*$'].search(data__datafiles_key):
+                        if data__datafiles_key in data__datafiles_keys:
+                            data__datafiles_keys.remove(data__datafiles_key)
+                        if not isinstance(data__datafiles_val, (list, tuple)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".data-files.{data__datafiles_key}".format(**locals()) + " must be array", value=data__datafiles_val, name="" + (name_prefix or "data") + ".data-files.{data__datafiles_key}".format(**locals()) + "", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+                        data__datafiles_val_is_list = isinstance(data__datafiles_val, (list, tuple))
+                        if data__datafiles_val_is_list:
+                            data__datafiles_val_len = len(data__datafiles_val)
+                            for data__datafiles_val_x, data__datafiles_val_item in enumerate(data__datafiles_val):
+                                if not isinstance(data__datafiles_val_item, (str)):
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".data-files.{data__datafiles_key}[{data__datafiles_val_x}]".format(**locals()) + " must be string", value=data__datafiles_val_item, name="" + (name_prefix or "data") + ".data-files.{data__datafiles_key}[{data__datafiles_val_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "cmdclass" in data_keys:
+            data_keys.remove("cmdclass")
+            data__cmdclass = data["cmdclass"]
+            if not isinstance(data__cmdclass, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".cmdclass must be object", value=data__cmdclass, name="" + (name_prefix or "data") + ".cmdclass", definition={'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, rule='type')
+            data__cmdclass_is_dict = isinstance(data__cmdclass, dict)
+            if data__cmdclass_is_dict:
+                data__cmdclass_keys = set(data__cmdclass.keys())
+                for data__cmdclass_key, data__cmdclass_val in data__cmdclass.items():
+                    if REGEX_PATTERNS['^.*$'].search(data__cmdclass_key):
+                        if data__cmdclass_key in data__cmdclass_keys:
+                            data__cmdclass_keys.remove(data__cmdclass_key)
+                        if not isinstance(data__cmdclass_val, (str)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".cmdclass.{data__cmdclass_key}".format(**locals()) + " must be string", value=data__cmdclass_val, name="" + (name_prefix or "data") + ".cmdclass.{data__cmdclass_key}".format(**locals()) + "", definition={'type': 'string', 'format': 'python-qualified-identifier'}, rule='type')
+                        if isinstance(data__cmdclass_val, str):
+                            if not custom_formats["python-qualified-identifier"](data__cmdclass_val):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".cmdclass.{data__cmdclass_key}".format(**locals()) + " must be python-qualified-identifier", value=data__cmdclass_val, name="" + (name_prefix or "data") + ".cmdclass.{data__cmdclass_key}".format(**locals()) + "", definition={'type': 'string', 'format': 'python-qualified-identifier'}, rule='format')
+        if "license-files" in data_keys:
+            data_keys.remove("license-files")
+            data__licensefiles = data["license-files"]
+            if not isinstance(data__licensefiles, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".license-files must be array", value=data__licensefiles, name="" + (name_prefix or "data") + ".license-files", definition={'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, rule='type')
+            data__licensefiles_is_list = isinstance(data__licensefiles, (list, tuple))
+            if data__licensefiles_is_list:
+                data__licensefiles_len = len(data__licensefiles)
+                for data__licensefiles_x, data__licensefiles_item in enumerate(data__licensefiles):
+                    if not isinstance(data__licensefiles_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".license-files[{data__licensefiles_x}]".format(**locals()) + " must be string", value=data__licensefiles_item, name="" + (name_prefix or "data") + ".license-files[{data__licensefiles_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "dynamic" in data_keys:
+            data_keys.remove("dynamic")
+            data__dynamic = data["dynamic"]
+            if not isinstance(data__dynamic, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic must be object", value=data__dynamic, name="" + (name_prefix or "data") + ".dynamic", definition={'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'classifiers': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'description': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'entry-points': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, 'readme': {'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}}}, rule='type')
+            data__dynamic_is_dict = isinstance(data__dynamic, dict)
+            if data__dynamic_is_dict:
+                data__dynamic_keys = set(data__dynamic.keys())
+                if "version" in data__dynamic_keys:
+                    data__dynamic_keys.remove("version")
+                    data__dynamic__version = data__dynamic["version"]
+                    data__dynamic__version_one_of_count5 = 0
+                    if data__dynamic__version_one_of_count5 < 2:
+                        try:
+                            validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_attr_directive(data__dynamic__version, custom_formats, (name_prefix or "data") + ".dynamic.version")
+                            data__dynamic__version_one_of_count5 += 1
+                        except JsonSchemaValueException: pass
+                    if data__dynamic__version_one_of_count5 < 2:
+                        try:
+                            validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data__dynamic__version, custom_formats, (name_prefix or "data") + ".dynamic.version")
+                            data__dynamic__version_one_of_count5 += 1
+                        except JsonSchemaValueException: pass
+                    if data__dynamic__version_one_of_count5 != 1:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.version must be valid exactly by one definition" + (" (" + str(data__dynamic__version_one_of_count5) + " matches found)"), value=data__dynamic__version, name="" + (name_prefix or "data") + ".dynamic.version", definition={'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, rule='oneOf')
+                if "classifiers" in data__dynamic_keys:
+                    data__dynamic_keys.remove("classifiers")
+                    data__dynamic__classifiers = data__dynamic["classifiers"]
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data__dynamic__classifiers, custom_formats, (name_prefix or "data") + ".dynamic.classifiers")
+                if "description" in data__dynamic_keys:
+                    data__dynamic_keys.remove("description")
+                    data__dynamic__description = data__dynamic["description"]
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data__dynamic__description, custom_formats, (name_prefix or "data") + ".dynamic.description")
+                if "entry-points" in data__dynamic_keys:
+                    data__dynamic_keys.remove("entry-points")
+                    data__dynamic__entrypoints = data__dynamic["entry-points"]
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data__dynamic__entrypoints, custom_formats, (name_prefix or "data") + ".dynamic.entry-points")
+                if "dependencies" in data__dynamic_keys:
+                    data__dynamic_keys.remove("dependencies")
+                    data__dynamic__dependencies = data__dynamic["dependencies"]
+                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive_for_dependencies(data__dynamic__dependencies, custom_formats, (name_prefix or "data") + ".dynamic.dependencies")
+                if "optional-dependencies" in data__dynamic_keys:
+                    data__dynamic_keys.remove("optional-dependencies")
+                    data__dynamic__optionaldependencies = data__dynamic["optional-dependencies"]
+                    if not isinstance(data__dynamic__optionaldependencies, (dict)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.optional-dependencies must be object", value=data__dynamic__optionaldependencies, name="" + (name_prefix or "data") + ".dynamic.optional-dependencies", definition={'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, rule='type')
+                    data__dynamic__optionaldependencies_is_dict = isinstance(data__dynamic__optionaldependencies, dict)
+                    if data__dynamic__optionaldependencies_is_dict:
+                        data__dynamic__optionaldependencies_keys = set(data__dynamic__optionaldependencies.keys())
+                        for data__dynamic__optionaldependencies_key, data__dynamic__optionaldependencies_val in data__dynamic__optionaldependencies.items():
+                            if REGEX_PATTERNS['.+'].search(data__dynamic__optionaldependencies_key):
+                                if data__dynamic__optionaldependencies_key in data__dynamic__optionaldependencies_keys:
+                                    data__dynamic__optionaldependencies_keys.remove(data__dynamic__optionaldependencies_key)
+                                validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive_for_dependencies(data__dynamic__optionaldependencies_val, custom_formats, (name_prefix or "data") + ".dynamic.optional-dependencies.{data__dynamic__optionaldependencies_key}".format(**locals()))
+                        if data__dynamic__optionaldependencies_keys:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.optional-dependencies must not contain "+str(data__dynamic__optionaldependencies_keys)+" properties", value=data__dynamic__optionaldependencies, name="" + (name_prefix or "data") + ".dynamic.optional-dependencies", definition={'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, rule='additionalProperties')
+                        data__dynamic__optionaldependencies_len = len(data__dynamic__optionaldependencies)
+                        if data__dynamic__optionaldependencies_len != 0:
+                            data__dynamic__optionaldependencies_property_names = True
+                            for data__dynamic__optionaldependencies_key in data__dynamic__optionaldependencies:
+                                try:
+                                    if not isinstance(data__dynamic__optionaldependencies_key, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.optional-dependencies must be string", value=data__dynamic__optionaldependencies_key, name="" + (name_prefix or "data") + ".dynamic.optional-dependencies", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='type')
+                                    if isinstance(data__dynamic__optionaldependencies_key, str):
+                                        if not custom_formats["pep508-identifier"](data__dynamic__optionaldependencies_key):
+                                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.optional-dependencies must be pep508-identifier", value=data__dynamic__optionaldependencies_key, name="" + (name_prefix or "data") + ".dynamic.optional-dependencies", definition={'type': 'string', 'format': 'pep508-identifier'}, rule='format')
+                                except JsonSchemaValueException:
+                                    data__dynamic__optionaldependencies_property_names = False
+                            if not data__dynamic__optionaldependencies_property_names:
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.optional-dependencies must be named by propertyName definition", value=data__dynamic__optionaldependencies, name="" + (name_prefix or "data") + ".dynamic.optional-dependencies", definition={'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, rule='propertyNames')
+                if "readme" in data__dynamic_keys:
+                    data__dynamic_keys.remove("readme")
+                    data__dynamic__readme = data__dynamic["readme"]
+                    if not isinstance(data__dynamic__readme, (dict)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme must be object", value=data__dynamic__readme, name="" + (name_prefix or "data") + ".dynamic.readme", definition={'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}, rule='type')
+                    data__dynamic__readme_any_of_count6 = 0
+                    if not data__dynamic__readme_any_of_count6:
+                        try:
+                            validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data__dynamic__readme, custom_formats, (name_prefix or "data") + ".dynamic.readme")
+                            data__dynamic__readme_any_of_count6 += 1
+                        except JsonSchemaValueException: pass
+                    if not data__dynamic__readme_any_of_count6:
+                        try:
+                            if not isinstance(data__dynamic__readme, (dict)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme must be object", value=data__dynamic__readme, name="" + (name_prefix or "data") + ".dynamic.readme", definition={'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}, rule='type')
+                            data__dynamic__readme_is_dict = isinstance(data__dynamic__readme, dict)
+                            if data__dynamic__readme_is_dict:
+                                data__dynamic__readme_keys = set(data__dynamic__readme.keys())
+                                if "content-type" in data__dynamic__readme_keys:
+                                    data__dynamic__readme_keys.remove("content-type")
+                                    data__dynamic__readme__contenttype = data__dynamic__readme["content-type"]
+                                    if not isinstance(data__dynamic__readme__contenttype, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme.content-type must be string", value=data__dynamic__readme__contenttype, name="" + (name_prefix or "data") + ".dynamic.readme.content-type", definition={'type': 'string'}, rule='type')
+                                if "file" in data__dynamic__readme_keys:
+                                    data__dynamic__readme_keys.remove("file")
+                                    data__dynamic__readme__file = data__dynamic__readme["file"]
+                                    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive_properties_file(data__dynamic__readme__file, custom_formats, (name_prefix or "data") + ".dynamic.readme.file")
+                                if data__dynamic__readme_keys:
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme must not contain "+str(data__dynamic__readme_keys)+" properties", value=data__dynamic__readme, name="" + (name_prefix or "data") + ".dynamic.readme", definition={'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}, rule='additionalProperties')
+                            data__dynamic__readme_any_of_count6 += 1
+                        except JsonSchemaValueException: pass
+                    if not data__dynamic__readme_any_of_count6:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme cannot be validated by any definition", value=data__dynamic__readme, name="" + (name_prefix or "data") + ".dynamic.readme", definition={'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}, rule='anyOf')
+                    data__dynamic__readme_is_dict = isinstance(data__dynamic__readme, dict)
+                    if data__dynamic__readme_is_dict:
+                        data__dynamic__readme__missing_keys = set(['file']) - data__dynamic__readme.keys()
+                        if data__dynamic__readme__missing_keys:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic.readme must contain " + (str(sorted(data__dynamic__readme__missing_keys)) + " properties"), value=data__dynamic__readme, name="" + (name_prefix or "data") + ".dynamic.readme", definition={'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}, rule='required')
+                if data__dynamic_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic must not contain "+str(data__dynamic_keys)+" properties", value=data__dynamic, name="" + (name_prefix or "data") + ".dynamic", definition={'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'classifiers': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'description': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'entry-points': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, 'readme': {'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}}}, rule='additionalProperties')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html', 'title': '``tool.setuptools`` table', '$$description': ['``setuptools``-specific configurations that can be set by users that require', 'customization.', 'These configurations are completely optional and probably can be skipped when', 'creating simple packages. They are equivalent to some of the `Keywords', '<https://setuptools.pypa.io/en/latest/references/keywords.html>`_', 'used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.', 'It considers only ``setuptools`` `parameters', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_', 'that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``', 'and ``setup_requires`` (incompatible with modern workflows/standards).'], 'type': 'object', 'additionalProperties': False, 'properties': {'platforms': {'type': 'array', 'items': {'type': 'string'}}, 'provides': {'$$description': ['Package and virtual package names contained within this package', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'obsoletes': {'$$description': ['Packages which this package renders obsolete', '**(not supported by pip)**'], 'type': 'array', 'items': {'type': 'string', 'format': 'pep508-identifier'}}, 'zip-safe': {'$$description': ['Whether the project can be safely installed and run from a zip file.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'boolean'}, 'script-files': {'$$description': ['Legacy way of defining scripts (entry-points are preferred).', 'Equivalent to the ``script`` keyword in ``setup.py``', '(it was renamed to avoid confusion with entry-point based ``project.scripts``', 'defined in :pep:`621`).', '**DISCOURAGED**: generic script wrappers are tricky and may not work properly.', 'Whenever possible, please use ``project.scripts`` instead.'], 'type': 'array', 'items': {'type': 'string'}, '$comment': 'TODO: is this field deprecated/should be removed?'}, 'eager-resources': {'$$description': ['Resources that should be extracted together, if any of them is needed,', 'or if any C extensions included in the project are imported.', '**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and', '``setup.py install`` in the context of ``eggs`` (**DEPRECATED**).'], 'type': 'array', 'items': {'type': 'string'}}, 'packages': {'$$description': ['Packages that should be included in the distribution.', 'It can be given either as a list of package identifiers', 'or as a ``dict``-like structure with a single key ``find``', 'which corresponds to a dynamic call to', '``setuptools.config.expand.find_packages`` function.', 'The ``find`` key is associated with a nested ``dict``-like structure that can', 'contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,', 'mimicking the keyword arguments of the associated function.'], 'oneOf': [{'title': 'Array of Python package identifiers', 'type': 'array', 'items': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}}, {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}]}, 'package-dir': {'$$description': [':class:`dict`-like structure mapping from package names to directories where their', 'code can be found.', 'The empty string (as key) means that all packages are contained inside', 'the given directory will be included in the distribution.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'const': ''}, {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}]}, 'patternProperties': {'^.*$': {'type': 'string'}}}, 'package-data': {'$$description': ['Mapping from package names to lists of glob patterns.', 'Usually this option is not needed when using ``include-package-data = true``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'include-package-data': {'$$description': ['Automatically include any data files inside the package directories', 'that are specified by ``MANIFEST.in``', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'boolean'}, 'exclude-package-data': {'$$description': ['Mapping from package names to lists of glob patterns that should be excluded', 'For more information on how to include data files, check ``setuptools`` `docs', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'additionalProperties': False, 'propertyNames': {'anyOf': [{'type': 'string', 'format': 'python-module-name'}, {'const': '*'}]}, 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'namespace-packages': {'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'https://setuptools.pypa.io/en/latest/userguide/package_discovery.html', 'description': '**DEPRECATED**: use implicit namespaces instead (:pep:`420`).'}, 'py-modules': {'description': 'Modules that setuptools will manipulate', 'type': 'array', 'items': {'type': 'string', 'format': 'python-module-name-relaxed'}, '$comment': 'TODO: clarify the relationship with ``packages``'}, 'ext-modules': {'description': 'Extension modules to be compiled by setuptools', 'type': 'array', 'items': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}}, 'data-files': {'$$description': ['``dict``-like structure where each key represents a directory and', 'the value is a list of glob patterns that should be installed in them.', '**DISCOURAGED**: please notice this might not work as expected with wheels.', 'Whenever possible, consider using data files inside the package directories', '(or create a new namespace package that only contains data files).', 'See `data files support', '<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_.'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'array', 'items': {'type': 'string'}}}}, 'cmdclass': {'$$description': ['Mapping of distutils-style command names to ``setuptools.Command`` subclasses', 'which in turn should be represented by strings with a qualified class name', '(i.e., "dotted" form with module), e.g.::\n\n', '    cmdclass = {mycmd = "pkg.subpkg.module.CommandClass"}\n\n', 'The command class should be a directly defined at the top-level of the', 'containing module (no class nesting).'], 'type': 'object', 'patternProperties': {'^.*$': {'type': 'string', 'format': 'python-qualified-identifier'}}}, 'license-files': {'type': 'array', 'items': {'type': 'string'}, '$$description': ['**PROVISIONAL**: list of glob patterns for all license files being distributed.', '(likely to become standard with :pep:`639`).', "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"], '$comment': 'TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?'}, 'dynamic': {'type': 'object', 'description': 'Instructions for loading :pep:`621`-related metadata dynamically', 'additionalProperties': False, 'properties': {'version': {'$$description': ['A version dynamically loaded via either the ``attr:`` or ``file:``', 'directives. Please make sure the given file or attribute respects :pep:`440`.', 'Also ensure to set ``project.dynamic`` accordingly.'], 'oneOf': [{'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'classifiers': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'description': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'entry-points': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}, 'optional-dependencies': {'type': 'object', 'propertyNames': {'type': 'string', 'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'.+': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$ref': '#/definitions/file-directive'}]}}}, 'readme': {'type': 'object', 'anyOf': [{'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, {'type': 'object', 'properties': {'content-type': {'type': 'string'}, 'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'additionalProperties': False}], 'required': ['file']}}}}, 'definitions': {'package-name': {'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, 'ext-module': {'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, 'file-directive': {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, 'file-directive-for-dependencies': {'title': "'file:' directive for dependencies", 'allOf': [{'$$description': ['**BETA**: subset of the ``requirements.txt`` format', 'without ``pip`` flags and options', '(one :pep:`508`-compliant string per line,', 'lines that are blank or start with ``#`` are excluded).', 'See `dynamic metadata', '<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_.']}, {'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}]}, 'attr-directive': {'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, 'find-directive': {'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}}}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive_properties_file(data, custom_formats={}, name_prefix=None):
+    data_one_of_count7 = 0
+    if data_one_of_count7 < 2:
+        try:
+            if not isinstance(data, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must be string", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'string'}, rule='type')
+            data_one_of_count7 += 1
+        except JsonSchemaValueException: pass
+    if data_one_of_count7 < 2:
+        try:
+            if not isinstance(data, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must be array", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data_is_list = isinstance(data, (list, tuple))
+            if data_is_list:
+                data_len = len(data)
+                for data_x, data_item in enumerate(data):
+                    if not isinstance(data_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + "[{data_x}]".format(**locals()) + " must be string", value=data_item, name="" + (name_prefix or "data") + "[{data_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+            data_one_of_count7 += 1
+        except JsonSchemaValueException: pass
+    if data_one_of_count7 != 1:
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be valid exactly by one definition" + (" (" + str(data_one_of_count7) + " matches found)"), value=data, name="" + (name_prefix or "data") + "", definition={'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}, rule='oneOf')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive_for_dependencies(data, custom_formats={}, name_prefix=None):
+    validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data, custom_formats, (name_prefix or "data") + "")
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_file_directive(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data__missing_keys = set(['file']) - data.keys()
+        if data__missing_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, rule='required')
+        data_keys = set(data.keys())
+        if "file" in data_keys:
+            data_keys.remove("file")
+            data__file = data["file"]
+            data__file_one_of_count8 = 0
+            if data__file_one_of_count8 < 2:
+                try:
+                    if not isinstance(data__file, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".file must be string", value=data__file, name="" + (name_prefix or "data") + ".file", definition={'type': 'string'}, rule='type')
+                    data__file_one_of_count8 += 1
+                except JsonSchemaValueException: pass
+            if data__file_one_of_count8 < 2:
+                try:
+                    if not isinstance(data__file, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".file must be array", value=data__file, name="" + (name_prefix or "data") + ".file", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+                    data__file_is_list = isinstance(data__file, (list, tuple))
+                    if data__file_is_list:
+                        data__file_len = len(data__file)
+                        for data__file_x, data__file_item in enumerate(data__file):
+                            if not isinstance(data__file_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".file[{data__file_x}]".format(**locals()) + " must be string", value=data__file_item, name="" + (name_prefix or "data") + ".file[{data__file_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                    data__file_one_of_count8 += 1
+                except JsonSchemaValueException: pass
+            if data__file_one_of_count8 != 1:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".file must be valid exactly by one definition" + (" (" + str(data__file_one_of_count8) + " matches found)"), value=data__file, name="" + (name_prefix or "data") + ".file", definition={'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}, rule='oneOf')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/file-directive', 'title': "'file:' directive", 'description': 'Value is read from a file (or list of files and then concatenated)', 'type': 'object', 'additionalProperties': False, 'properties': {'file': {'oneOf': [{'type': 'string'}, {'type': 'array', 'items': {'type': 'string'}}]}}, 'required': ['file']}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_attr_directive(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data__missing_keys = set(['attr']) - data.keys()
+        if data__missing_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, rule='required')
+        data_keys = set(data.keys())
+        if "attr" in data_keys:
+            data_keys.remove("attr")
+            data__attr = data["attr"]
+            if not isinstance(data__attr, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".attr must be string", value=data__attr, name="" + (name_prefix or "data") + ".attr", definition={'type': 'string', 'format': 'python-qualified-identifier'}, rule='type')
+            if isinstance(data__attr, str):
+                if not custom_formats["python-qualified-identifier"](data__attr):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".attr must be python-qualified-identifier", value=data__attr, name="" + (name_prefix or "data") + ".attr", definition={'type': 'string', 'format': 'python-qualified-identifier'}, rule='format')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'title': "'attr:' directive", '$id': '#/definitions/attr-directive', '$$description': ['Value is read from a module attribute. Supports callables and iterables;', 'unsupported types are cast via ``str()``'], 'type': 'object', 'additionalProperties': False, 'properties': {'attr': {'type': 'string', 'format': 'python-qualified-identifier'}}, 'required': ['attr']}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_ext_module(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data__missing_keys = set(['name', 'sources']) - data.keys()
+        if data__missing_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, rule='required')
+        data_keys = set(data.keys())
+        if "name" in data_keys:
+            data_keys.remove("name")
+            data__name = data["name"]
+            if not isinstance(data__name, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".name must be string", value=data__name, name="" + (name_prefix or "data") + ".name", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='type')
+            if isinstance(data__name, str):
+                if not custom_formats["python-module-name-relaxed"](data__name):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".name must be python-module-name-relaxed", value=data__name, name="" + (name_prefix or "data") + ".name", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='format')
+        if "sources" in data_keys:
+            data_keys.remove("sources")
+            data__sources = data["sources"]
+            if not isinstance(data__sources, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".sources must be array", value=data__sources, name="" + (name_prefix or "data") + ".sources", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__sources_is_list = isinstance(data__sources, (list, tuple))
+            if data__sources_is_list:
+                data__sources_len = len(data__sources)
+                for data__sources_x, data__sources_item in enumerate(data__sources):
+                    if not isinstance(data__sources_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".sources[{data__sources_x}]".format(**locals()) + " must be string", value=data__sources_item, name="" + (name_prefix or "data") + ".sources[{data__sources_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "include-dirs" in data_keys:
+            data_keys.remove("include-dirs")
+            data__includedirs = data["include-dirs"]
+            if not isinstance(data__includedirs, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".include-dirs must be array", value=data__includedirs, name="" + (name_prefix or "data") + ".include-dirs", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__includedirs_is_list = isinstance(data__includedirs, (list, tuple))
+            if data__includedirs_is_list:
+                data__includedirs_len = len(data__includedirs)
+                for data__includedirs_x, data__includedirs_item in enumerate(data__includedirs):
+                    if not isinstance(data__includedirs_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".include-dirs[{data__includedirs_x}]".format(**locals()) + " must be string", value=data__includedirs_item, name="" + (name_prefix or "data") + ".include-dirs[{data__includedirs_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "define-macros" in data_keys:
+            data_keys.remove("define-macros")
+            data__definemacros = data["define-macros"]
+            if not isinstance(data__definemacros, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros must be array", value=data__definemacros, name="" + (name_prefix or "data") + ".define-macros", definition={'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, rule='type')
+            data__definemacros_is_list = isinstance(data__definemacros, (list, tuple))
+            if data__definemacros_is_list:
+                data__definemacros_len = len(data__definemacros)
+                for data__definemacros_x, data__definemacros_item in enumerate(data__definemacros):
+                    if not isinstance(data__definemacros_item, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}]".format(**locals()) + " must be array", value=data__definemacros_item, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}]".format(**locals()) + "", definition={'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}, rule='type')
+                    data__definemacros_item_is_list = isinstance(data__definemacros_item, (list, tuple))
+                    if data__definemacros_item_is_list:
+                        data__definemacros_item_len = len(data__definemacros_item)
+                        if data__definemacros_item_len > 0:
+                            data__definemacros_item__0 = data__definemacros_item[0]
+                            if not isinstance(data__definemacros_item__0, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][0]".format(**locals()) + " must be string", value=data__definemacros_item__0, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][0]".format(**locals()) + "", definition={'description': 'macro name', 'type': 'string'}, rule='type')
+                        if data__definemacros_item_len > 1:
+                            data__definemacros_item__1 = data__definemacros_item[1]
+                            data__definemacros_item__1_one_of_count9 = 0
+                            if data__definemacros_item__1_one_of_count9 < 2:
+                                try:
+                                    if not isinstance(data__definemacros_item__1, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + " must be string", value=data__definemacros_item__1, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                                    data__definemacros_item__1_one_of_count9 += 1
+                                except JsonSchemaValueException: pass
+                            if data__definemacros_item__1_one_of_count9 < 2:
+                                try:
+                                    if not isinstance(data__definemacros_item__1, (NoneType)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + " must be null", value=data__definemacros_item__1, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + "", definition={'type': 'null'}, rule='type')
+                                    data__definemacros_item__1_one_of_count9 += 1
+                                except JsonSchemaValueException: pass
+                            if data__definemacros_item__1_one_of_count9 != 1:
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + " must be valid exactly by one definition" + (" (" + str(data__definemacros_item__1_one_of_count9) + " matches found)"), value=data__definemacros_item__1, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}][1]".format(**locals()) + "", definition={'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}, rule='oneOf')
+                        if data__definemacros_item_len > 2:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}]".format(**locals()) + " must contain only specified items", value=data__definemacros_item, name="" + (name_prefix or "data") + ".define-macros[{data__definemacros_x}]".format(**locals()) + "", definition={'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}, rule='items')
+        if "undef-macros" in data_keys:
+            data_keys.remove("undef-macros")
+            data__undefmacros = data["undef-macros"]
+            if not isinstance(data__undefmacros, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".undef-macros must be array", value=data__undefmacros, name="" + (name_prefix or "data") + ".undef-macros", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__undefmacros_is_list = isinstance(data__undefmacros, (list, tuple))
+            if data__undefmacros_is_list:
+                data__undefmacros_len = len(data__undefmacros)
+                for data__undefmacros_x, data__undefmacros_item in enumerate(data__undefmacros):
+                    if not isinstance(data__undefmacros_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".undef-macros[{data__undefmacros_x}]".format(**locals()) + " must be string", value=data__undefmacros_item, name="" + (name_prefix or "data") + ".undef-macros[{data__undefmacros_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "library-dirs" in data_keys:
+            data_keys.remove("library-dirs")
+            data__librarydirs = data["library-dirs"]
+            if not isinstance(data__librarydirs, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".library-dirs must be array", value=data__librarydirs, name="" + (name_prefix or "data") + ".library-dirs", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__librarydirs_is_list = isinstance(data__librarydirs, (list, tuple))
+            if data__librarydirs_is_list:
+                data__librarydirs_len = len(data__librarydirs)
+                for data__librarydirs_x, data__librarydirs_item in enumerate(data__librarydirs):
+                    if not isinstance(data__librarydirs_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".library-dirs[{data__librarydirs_x}]".format(**locals()) + " must be string", value=data__librarydirs_item, name="" + (name_prefix or "data") + ".library-dirs[{data__librarydirs_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "libraries" in data_keys:
+            data_keys.remove("libraries")
+            data__libraries = data["libraries"]
+            if not isinstance(data__libraries, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".libraries must be array", value=data__libraries, name="" + (name_prefix or "data") + ".libraries", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__libraries_is_list = isinstance(data__libraries, (list, tuple))
+            if data__libraries_is_list:
+                data__libraries_len = len(data__libraries)
+                for data__libraries_x, data__libraries_item in enumerate(data__libraries):
+                    if not isinstance(data__libraries_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".libraries[{data__libraries_x}]".format(**locals()) + " must be string", value=data__libraries_item, name="" + (name_prefix or "data") + ".libraries[{data__libraries_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "runtime-library-dirs" in data_keys:
+            data_keys.remove("runtime-library-dirs")
+            data__runtimelibrarydirs = data["runtime-library-dirs"]
+            if not isinstance(data__runtimelibrarydirs, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".runtime-library-dirs must be array", value=data__runtimelibrarydirs, name="" + (name_prefix or "data") + ".runtime-library-dirs", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__runtimelibrarydirs_is_list = isinstance(data__runtimelibrarydirs, (list, tuple))
+            if data__runtimelibrarydirs_is_list:
+                data__runtimelibrarydirs_len = len(data__runtimelibrarydirs)
+                for data__runtimelibrarydirs_x, data__runtimelibrarydirs_item in enumerate(data__runtimelibrarydirs):
+                    if not isinstance(data__runtimelibrarydirs_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".runtime-library-dirs[{data__runtimelibrarydirs_x}]".format(**locals()) + " must be string", value=data__runtimelibrarydirs_item, name="" + (name_prefix or "data") + ".runtime-library-dirs[{data__runtimelibrarydirs_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "extra-objects" in data_keys:
+            data_keys.remove("extra-objects")
+            data__extraobjects = data["extra-objects"]
+            if not isinstance(data__extraobjects, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-objects must be array", value=data__extraobjects, name="" + (name_prefix or "data") + ".extra-objects", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__extraobjects_is_list = isinstance(data__extraobjects, (list, tuple))
+            if data__extraobjects_is_list:
+                data__extraobjects_len = len(data__extraobjects)
+                for data__extraobjects_x, data__extraobjects_item in enumerate(data__extraobjects):
+                    if not isinstance(data__extraobjects_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-objects[{data__extraobjects_x}]".format(**locals()) + " must be string", value=data__extraobjects_item, name="" + (name_prefix or "data") + ".extra-objects[{data__extraobjects_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "extra-compile-args" in data_keys:
+            data_keys.remove("extra-compile-args")
+            data__extracompileargs = data["extra-compile-args"]
+            if not isinstance(data__extracompileargs, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-compile-args must be array", value=data__extracompileargs, name="" + (name_prefix or "data") + ".extra-compile-args", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__extracompileargs_is_list = isinstance(data__extracompileargs, (list, tuple))
+            if data__extracompileargs_is_list:
+                data__extracompileargs_len = len(data__extracompileargs)
+                for data__extracompileargs_x, data__extracompileargs_item in enumerate(data__extracompileargs):
+                    if not isinstance(data__extracompileargs_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-compile-args[{data__extracompileargs_x}]".format(**locals()) + " must be string", value=data__extracompileargs_item, name="" + (name_prefix or "data") + ".extra-compile-args[{data__extracompileargs_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "extra-link-args" in data_keys:
+            data_keys.remove("extra-link-args")
+            data__extralinkargs = data["extra-link-args"]
+            if not isinstance(data__extralinkargs, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-link-args must be array", value=data__extralinkargs, name="" + (name_prefix or "data") + ".extra-link-args", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__extralinkargs_is_list = isinstance(data__extralinkargs, (list, tuple))
+            if data__extralinkargs_is_list:
+                data__extralinkargs_len = len(data__extralinkargs)
+                for data__extralinkargs_x, data__extralinkargs_item in enumerate(data__extralinkargs):
+                    if not isinstance(data__extralinkargs_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".extra-link-args[{data__extralinkargs_x}]".format(**locals()) + " must be string", value=data__extralinkargs_item, name="" + (name_prefix or "data") + ".extra-link-args[{data__extralinkargs_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "export-symbols" in data_keys:
+            data_keys.remove("export-symbols")
+            data__exportsymbols = data["export-symbols"]
+            if not isinstance(data__exportsymbols, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".export-symbols must be array", value=data__exportsymbols, name="" + (name_prefix or "data") + ".export-symbols", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__exportsymbols_is_list = isinstance(data__exportsymbols, (list, tuple))
+            if data__exportsymbols_is_list:
+                data__exportsymbols_len = len(data__exportsymbols)
+                for data__exportsymbols_x, data__exportsymbols_item in enumerate(data__exportsymbols):
+                    if not isinstance(data__exportsymbols_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".export-symbols[{data__exportsymbols_x}]".format(**locals()) + " must be string", value=data__exportsymbols_item, name="" + (name_prefix or "data") + ".export-symbols[{data__exportsymbols_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "swig-opts" in data_keys:
+            data_keys.remove("swig-opts")
+            data__swigopts = data["swig-opts"]
+            if not isinstance(data__swigopts, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".swig-opts must be array", value=data__swigopts, name="" + (name_prefix or "data") + ".swig-opts", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__swigopts_is_list = isinstance(data__swigopts, (list, tuple))
+            if data__swigopts_is_list:
+                data__swigopts_len = len(data__swigopts)
+                for data__swigopts_x, data__swigopts_item in enumerate(data__swigopts):
+                    if not isinstance(data__swigopts_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".swig-opts[{data__swigopts_x}]".format(**locals()) + " must be string", value=data__swigopts_item, name="" + (name_prefix or "data") + ".swig-opts[{data__swigopts_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "depends" in data_keys:
+            data_keys.remove("depends")
+            data__depends = data["depends"]
+            if not isinstance(data__depends, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".depends must be array", value=data__depends, name="" + (name_prefix or "data") + ".depends", definition={'type': 'array', 'items': {'type': 'string'}}, rule='type')
+            data__depends_is_list = isinstance(data__depends, (list, tuple))
+            if data__depends_is_list:
+                data__depends_len = len(data__depends)
+                for data__depends_x, data__depends_item in enumerate(data__depends):
+                    if not isinstance(data__depends_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".depends[{data__depends_x}]".format(**locals()) + " must be string", value=data__depends_item, name="" + (name_prefix or "data") + ".depends[{data__depends_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "language" in data_keys:
+            data_keys.remove("language")
+            data__language = data["language"]
+            if not isinstance(data__language, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".language must be string", value=data__language, name="" + (name_prefix or "data") + ".language", definition={'type': 'string'}, rule='type')
+        if "optional" in data_keys:
+            data_keys.remove("optional")
+            data__optional = data["optional"]
+            if not isinstance(data__optional, (bool)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional must be boolean", value=data__optional, name="" + (name_prefix or "data") + ".optional", definition={'type': 'boolean'}, rule='type')
+        if "py-limited-api" in data_keys:
+            data_keys.remove("py-limited-api")
+            data__pylimitedapi = data["py-limited-api"]
+            if not isinstance(data__pylimitedapi, (bool)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".py-limited-api must be boolean", value=data__pylimitedapi, name="" + (name_prefix or "data") + ".py-limited-api", definition={'type': 'boolean'}, rule='type')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/ext-module', 'title': 'Extension module', 'description': 'Parameters to construct a :class:`setuptools.Extension` object', 'type': 'object', 'required': ['name', 'sources'], 'additionalProperties': False, 'properties': {'name': {'type': 'string', 'format': 'python-module-name-relaxed'}, 'sources': {'type': 'array', 'items': {'type': 'string'}}, 'include-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'define-macros': {'type': 'array', 'items': {'type': 'array', 'items': [{'description': 'macro name', 'type': 'string'}, {'description': 'macro value', 'oneOf': [{'type': 'string'}, {'type': 'null'}]}], 'additionalItems': False}}, 'undef-macros': {'type': 'array', 'items': {'type': 'string'}}, 'library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'libraries': {'type': 'array', 'items': {'type': 'string'}}, 'runtime-library-dirs': {'type': 'array', 'items': {'type': 'string'}}, 'extra-objects': {'type': 'array', 'items': {'type': 'string'}}, 'extra-compile-args': {'type': 'array', 'items': {'type': 'string'}}, 'extra-link-args': {'type': 'array', 'items': {'type': 'string'}}, 'export-symbols': {'type': 'array', 'items': {'type': 'string'}}, 'swig-opts': {'type': 'array', 'items': {'type': 'string'}}, 'depends': {'type': 'array', 'items': {'type': 'string'}}, 'language': {'type': 'string'}, 'optional': {'type': 'boolean'}, 'py-limited-api': {'type': 'boolean'}}}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_find_directive(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        if "find" in data_keys:
+            data_keys.remove("find")
+            data__find = data["find"]
+            if not isinstance(data__find, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".find must be object", value=data__find, name="" + (name_prefix or "data") + ".find", definition={'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}, rule='type')
+            data__find_is_dict = isinstance(data__find, dict)
+            if data__find_is_dict:
+                data__find_keys = set(data__find.keys())
+                if "where" in data__find_keys:
+                    data__find_keys.remove("where")
+                    data__find__where = data__find["where"]
+                    if not isinstance(data__find__where, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.where must be array", value=data__find__where, name="" + (name_prefix or "data") + ".find.where", definition={'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, rule='type')
+                    data__find__where_is_list = isinstance(data__find__where, (list, tuple))
+                    if data__find__where_is_list:
+                        data__find__where_len = len(data__find__where)
+                        for data__find__where_x, data__find__where_item in enumerate(data__find__where):
+                            if not isinstance(data__find__where_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.where[{data__find__where_x}]".format(**locals()) + " must be string", value=data__find__where_item, name="" + (name_prefix or "data") + ".find.where[{data__find__where_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if "exclude" in data__find_keys:
+                    data__find_keys.remove("exclude")
+                    data__find__exclude = data__find["exclude"]
+                    if not isinstance(data__find__exclude, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.exclude must be array", value=data__find__exclude, name="" + (name_prefix or "data") + ".find.exclude", definition={'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, rule='type')
+                    data__find__exclude_is_list = isinstance(data__find__exclude, (list, tuple))
+                    if data__find__exclude_is_list:
+                        data__find__exclude_len = len(data__find__exclude)
+                        for data__find__exclude_x, data__find__exclude_item in enumerate(data__find__exclude):
+                            if not isinstance(data__find__exclude_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.exclude[{data__find__exclude_x}]".format(**locals()) + " must be string", value=data__find__exclude_item, name="" + (name_prefix or "data") + ".find.exclude[{data__find__exclude_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if "include" in data__find_keys:
+                    data__find_keys.remove("include")
+                    data__find__include = data__find["include"]
+                    if not isinstance(data__find__include, (list, tuple)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.include must be array", value=data__find__include, name="" + (name_prefix or "data") + ".find.include", definition={'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, rule='type')
+                    data__find__include_is_list = isinstance(data__find__include, (list, tuple))
+                    if data__find__include_is_list:
+                        data__find__include_len = len(data__find__include)
+                        for data__find__include_x, data__find__include_item in enumerate(data__find__include):
+                            if not isinstance(data__find__include_item, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.include[{data__find__include_x}]".format(**locals()) + " must be string", value=data__find__include_item, name="" + (name_prefix or "data") + ".find.include[{data__find__include_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+                if "namespaces" in data__find_keys:
+                    data__find_keys.remove("namespaces")
+                    data__find__namespaces = data__find["namespaces"]
+                    if not isinstance(data__find__namespaces, (bool)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".find.namespaces must be boolean", value=data__find__namespaces, name="" + (name_prefix or "data") + ".find.namespaces", definition={'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}, rule='type')
+                if data__find_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".find must not contain "+str(data__find_keys)+" properties", value=data__find, name="" + (name_prefix or "data") + ".find", definition={'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}, rule='additionalProperties')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/find-directive', 'title': "'find:' directive", 'type': 'object', 'additionalProperties': False, 'properties': {'find': {'type': 'object', '$$description': ['Dynamic `package discovery', '<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_.'], 'additionalProperties': False, 'properties': {'where': {'description': 'Directories to be searched for packages (Unix-style relative path)', 'type': 'array', 'items': {'type': 'string'}}, 'exclude': {'type': 'array', '$$description': ['Exclude packages that match the values listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'include': {'type': 'array', '$$description': ['Restrict the found packages to just the ones listed in this field.', "Can container shell-style wildcards (e.g. ``'pkg.*'``)"], 'items': {'type': 'string'}}, 'namespaces': {'type': 'boolean', '$$description': ['When ``True``, directories without a ``__init__.py`` file will also', 'be scanned for :pep:`420`-style implicit namespaces']}}}}}, rule='additionalProperties')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_userguide_pyproject_config_html__definitions_package_name(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (str)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be string", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, rule='type')
+    data_any_of_count10 = 0
+    if not data_any_of_count10:
+        try:
+            if not isinstance(data, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must be string", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='type')
+            if isinstance(data, str):
+                if not custom_formats["python-module-name-relaxed"](data):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + " must be python-module-name-relaxed", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'string', 'format': 'python-module-name-relaxed'}, rule='format')
+            data_any_of_count10 += 1
+        except JsonSchemaValueException: pass
+    if not data_any_of_count10:
+        try:
+            if not isinstance(data, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must be string", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'string', 'format': 'pep561-stub-name'}, rule='type')
+            if isinstance(data, str):
+                if not custom_formats["pep561-stub-name"](data):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + " must be pep561-stub-name", value=data, name="" + (name_prefix or "data") + "", definition={'type': 'string', 'format': 'pep561-stub-name'}, rule='format')
+            data_any_of_count10 += 1
+        except JsonSchemaValueException: pass
+    if not data_any_of_count10:
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " cannot be validated by any definition", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/package-name', 'title': 'Valid package name', 'description': 'Valid package name (importable or :pep:`561`).', 'type': 'string', 'anyOf': [{'type': 'string', 'format': 'python-module-name-relaxed'}, {'type': 'string', 'format': 'pep561-stub-name'}]}, rule='anyOf')
+    return data
+
+def validate_https___setuptools_pypa_io_en_latest_deprecated_distutils_configfile_html(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html', 'title': '``tool.distutils`` table', '$$description': ['**EXPERIMENTAL** (NOT OFFICIALLY SUPPORTED): Use ``tool.distutils``', 'subtables to configure arguments for ``distutils`` commands.', 'Originally, ``distutils`` allowed developers to configure arguments for', '``setup.py`` commands via `distutils configuration files', '<https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html>`_.', 'See also `the old Python docs <https://docs.python.org/3.11/install/>_`.'], 'type': 'object', 'properties': {'global': {'type': 'object', 'description': 'Global options applied to all ``distutils`` commands'}}, 'patternProperties': {'.+': {'type': 'object'}}, '$comment': 'TODO: Is there a practical way of making this schema more specific?'}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        if "global" in data_keys:
+            data_keys.remove("global")
+            data__global = data["global"]
+            if not isinstance(data__global, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".global must be object", value=data__global, name="" + (name_prefix or "data") + ".global", definition={'type': 'object', 'description': 'Global options applied to all ``distutils`` commands'}, rule='type')
+        for data_key, data_val in data.items():
+            if REGEX_PATTERNS['.+'].search(data_key):
+                if data_key in data_keys:
+                    data_keys.remove(data_key)
+                if not isinstance(data_val, (dict)):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".{data_key}".format(**locals()) + " must be object", value=data_val, name="" + (name_prefix or "data") + ".{data_key}".format(**locals()) + "", definition={'type': 'object'}, rule='type')
+    return data
+
+def validate_https___packaging_python_org_en_latest_specifications_pyproject_toml(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'gui-scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data__missing_keys = set(['name']) - data.keys()
+        if data__missing_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'gui-scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}, rule='required')
+        data_keys = set(data.keys())
+        if "name" in data_keys:
+            data_keys.remove("name")
+            data__name = data["name"]
+            if not isinstance(data__name, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".name must be string", value=data__name, name="" + (name_prefix or "data") + ".name", definition={'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, rule='type')
+            if isinstance(data__name, str):
+                if not custom_formats["pep508-identifier"](data__name):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".name must be pep508-identifier", value=data__name, name="" + (name_prefix or "data") + ".name", definition={'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, rule='format')
+        if "version" in data_keys:
+            data_keys.remove("version")
+            data__version = data["version"]
+            if not isinstance(data__version, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".version must be string", value=data__version, name="" + (name_prefix or "data") + ".version", definition={'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, rule='type')
+            if isinstance(data__version, str):
+                if not custom_formats["pep440"](data__version):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".version must be pep440", value=data__version, name="" + (name_prefix or "data") + ".version", definition={'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, rule='format')
+        if "description" in data_keys:
+            data_keys.remove("description")
+            data__description = data["description"]
+            if not isinstance(data__description, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".description must be string", value=data__description, name="" + (name_prefix or "data") + ".description", definition={'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, rule='type')
+        if "readme" in data_keys:
+            data_keys.remove("readme")
+            data__readme = data["readme"]
+            data__readme_one_of_count11 = 0
+            if data__readme_one_of_count11 < 2:
+                try:
+                    if not isinstance(data__readme, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must be string", value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, rule='type')
+                    data__readme_one_of_count11 += 1
+                except JsonSchemaValueException: pass
+            if data__readme_one_of_count11 < 2:
+                try:
+                    if not isinstance(data__readme, (dict)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must be object", value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}, rule='type')
+                    data__readme_any_of_count12 = 0
+                    if not data__readme_any_of_count12:
+                        try:
+                            data__readme_is_dict = isinstance(data__readme, dict)
+                            if data__readme_is_dict:
+                                data__readme__missing_keys = set(['file']) - data__readme.keys()
+                                if data__readme__missing_keys:
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must contain " + (str(sorted(data__readme__missing_keys)) + " properties"), value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, rule='required')
+                                data__readme_keys = set(data__readme.keys())
+                                if "file" in data__readme_keys:
+                                    data__readme_keys.remove("file")
+                                    data__readme__file = data__readme["file"]
+                                    if not isinstance(data__readme__file, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme.file must be string", value=data__readme__file, name="" + (name_prefix or "data") + ".readme.file", definition={'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}, rule='type')
+                            data__readme_any_of_count12 += 1
+                        except JsonSchemaValueException: pass
+                    if not data__readme_any_of_count12:
+                        try:
+                            data__readme_is_dict = isinstance(data__readme, dict)
+                            if data__readme_is_dict:
+                                data__readme__missing_keys = set(['text']) - data__readme.keys()
+                                if data__readme__missing_keys:
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must contain " + (str(sorted(data__readme__missing_keys)) + " properties"), value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}, rule='required')
+                                data__readme_keys = set(data__readme.keys())
+                                if "text" in data__readme_keys:
+                                    data__readme_keys.remove("text")
+                                    data__readme__text = data__readme["text"]
+                                    if not isinstance(data__readme__text, (str)):
+                                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme.text must be string", value=data__readme__text, name="" + (name_prefix or "data") + ".readme.text", definition={'type': 'string', 'description': 'Full text describing the project.'}, rule='type')
+                            data__readme_any_of_count12 += 1
+                        except JsonSchemaValueException: pass
+                    if not data__readme_any_of_count12:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme cannot be validated by any definition", value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, rule='anyOf')
+                    data__readme_is_dict = isinstance(data__readme, dict)
+                    if data__readme_is_dict:
+                        data__readme__missing_keys = set(['content-type']) - data__readme.keys()
+                        if data__readme__missing_keys:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must contain " + (str(sorted(data__readme__missing_keys)) + " properties"), value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}, rule='required')
+                        data__readme_keys = set(data__readme.keys())
+                        if "content-type" in data__readme_keys:
+                            data__readme_keys.remove("content-type")
+                            data__readme__contenttype = data__readme["content-type"]
+                            if not isinstance(data__readme__contenttype, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme.content-type must be string", value=data__readme__contenttype, name="" + (name_prefix or "data") + ".readme.content-type", definition={'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}, rule='type')
+                    data__readme_one_of_count11 += 1
+                except JsonSchemaValueException: pass
+            if data__readme_one_of_count11 != 1:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".readme must be valid exactly by one definition" + (" (" + str(data__readme_one_of_count11) + " matches found)"), value=data__readme, name="" + (name_prefix or "data") + ".readme", definition={'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, rule='oneOf')
+        if "requires-python" in data_keys:
+            data_keys.remove("requires-python")
+            data__requirespython = data["requires-python"]
+            if not isinstance(data__requirespython, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".requires-python must be string", value=data__requirespython, name="" + (name_prefix or "data") + ".requires-python", definition={'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, rule='type')
+            if isinstance(data__requirespython, str):
+                if not custom_formats["pep508-versionspec"](data__requirespython):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".requires-python must be pep508-versionspec", value=data__requirespython, name="" + (name_prefix or "data") + ".requires-python", definition={'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, rule='format')
+        if "license" in data_keys:
+            data_keys.remove("license")
+            data__license = data["license"]
+            data__license_one_of_count13 = 0
+            if data__license_one_of_count13 < 2:
+                try:
+                    data__license_is_dict = isinstance(data__license, dict)
+                    if data__license_is_dict:
+                        data__license__missing_keys = set(['file']) - data__license.keys()
+                        if data__license__missing_keys:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".license must contain " + (str(sorted(data__license__missing_keys)) + " properties"), value=data__license, name="" + (name_prefix or "data") + ".license", definition={'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, rule='required')
+                        data__license_keys = set(data__license.keys())
+                        if "file" in data__license_keys:
+                            data__license_keys.remove("file")
+                            data__license__file = data__license["file"]
+                            if not isinstance(data__license__file, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".license.file must be string", value=data__license__file, name="" + (name_prefix or "data") + ".license.file", definition={'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}, rule='type')
+                    data__license_one_of_count13 += 1
+                except JsonSchemaValueException: pass
+            if data__license_one_of_count13 < 2:
+                try:
+                    data__license_is_dict = isinstance(data__license, dict)
+                    if data__license_is_dict:
+                        data__license__missing_keys = set(['text']) - data__license.keys()
+                        if data__license__missing_keys:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".license must contain " + (str(sorted(data__license__missing_keys)) + " properties"), value=data__license, name="" + (name_prefix or "data") + ".license", definition={'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}, rule='required')
+                        data__license_keys = set(data__license.keys())
+                        if "text" in data__license_keys:
+                            data__license_keys.remove("text")
+                            data__license__text = data__license["text"]
+                            if not isinstance(data__license__text, (str)):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".license.text must be string", value=data__license__text, name="" + (name_prefix or "data") + ".license.text", definition={'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}, rule='type')
+                    data__license_one_of_count13 += 1
+                except JsonSchemaValueException: pass
+            if data__license_one_of_count13 != 1:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".license must be valid exactly by one definition" + (" (" + str(data__license_one_of_count13) + " matches found)"), value=data__license, name="" + (name_prefix or "data") + ".license", definition={'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, rule='oneOf')
+        if "authors" in data_keys:
+            data_keys.remove("authors")
+            data__authors = data["authors"]
+            if not isinstance(data__authors, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".authors must be array", value=data__authors, name="" + (name_prefix or "data") + ".authors", definition={'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, rule='type')
+            data__authors_is_list = isinstance(data__authors, (list, tuple))
+            if data__authors_is_list:
+                data__authors_len = len(data__authors)
+                for data__authors_x, data__authors_item in enumerate(data__authors):
+                    validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_author(data__authors_item, custom_formats, (name_prefix or "data") + ".authors[{data__authors_x}]".format(**locals()))
+        if "maintainers" in data_keys:
+            data_keys.remove("maintainers")
+            data__maintainers = data["maintainers"]
+            if not isinstance(data__maintainers, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".maintainers must be array", value=data__maintainers, name="" + (name_prefix or "data") + ".maintainers", definition={'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, rule='type')
+            data__maintainers_is_list = isinstance(data__maintainers, (list, tuple))
+            if data__maintainers_is_list:
+                data__maintainers_len = len(data__maintainers)
+                for data__maintainers_x, data__maintainers_item in enumerate(data__maintainers):
+                    validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_author(data__maintainers_item, custom_formats, (name_prefix or "data") + ".maintainers[{data__maintainers_x}]".format(**locals()))
+        if "keywords" in data_keys:
+            data_keys.remove("keywords")
+            data__keywords = data["keywords"]
+            if not isinstance(data__keywords, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".keywords must be array", value=data__keywords, name="" + (name_prefix or "data") + ".keywords", definition={'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, rule='type')
+            data__keywords_is_list = isinstance(data__keywords, (list, tuple))
+            if data__keywords_is_list:
+                data__keywords_len = len(data__keywords)
+                for data__keywords_x, data__keywords_item in enumerate(data__keywords):
+                    if not isinstance(data__keywords_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".keywords[{data__keywords_x}]".format(**locals()) + " must be string", value=data__keywords_item, name="" + (name_prefix or "data") + ".keywords[{data__keywords_x}]".format(**locals()) + "", definition={'type': 'string'}, rule='type')
+        if "classifiers" in data_keys:
+            data_keys.remove("classifiers")
+            data__classifiers = data["classifiers"]
+            if not isinstance(data__classifiers, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".classifiers must be array", value=data__classifiers, name="" + (name_prefix or "data") + ".classifiers", definition={'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, rule='type')
+            data__classifiers_is_list = isinstance(data__classifiers, (list, tuple))
+            if data__classifiers_is_list:
+                data__classifiers_len = len(data__classifiers)
+                for data__classifiers_x, data__classifiers_item in enumerate(data__classifiers):
+                    if not isinstance(data__classifiers_item, (str)):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".classifiers[{data__classifiers_x}]".format(**locals()) + " must be string", value=data__classifiers_item, name="" + (name_prefix or "data") + ".classifiers[{data__classifiers_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, rule='type')
+                    if isinstance(data__classifiers_item, str):
+                        if not custom_formats["trove-classifier"](data__classifiers_item):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".classifiers[{data__classifiers_x}]".format(**locals()) + " must be trove-classifier", value=data__classifiers_item, name="" + (name_prefix or "data") + ".classifiers[{data__classifiers_x}]".format(**locals()) + "", definition={'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, rule='format')
+        if "urls" in data_keys:
+            data_keys.remove("urls")
+            data__urls = data["urls"]
+            if not isinstance(data__urls, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".urls must be object", value=data__urls, name="" + (name_prefix or "data") + ".urls", definition={'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, rule='type')
+            data__urls_is_dict = isinstance(data__urls, dict)
+            if data__urls_is_dict:
+                data__urls_keys = set(data__urls.keys())
+                for data__urls_key, data__urls_val in data__urls.items():
+                    if REGEX_PATTERNS['^.+$'].search(data__urls_key):
+                        if data__urls_key in data__urls_keys:
+                            data__urls_keys.remove(data__urls_key)
+                        if not isinstance(data__urls_val, (str)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".urls.{data__urls_key}".format(**locals()) + " must be string", value=data__urls_val, name="" + (name_prefix or "data") + ".urls.{data__urls_key}".format(**locals()) + "", definition={'type': 'string', 'format': 'url'}, rule='type')
+                        if isinstance(data__urls_val, str):
+                            if not custom_formats["url"](data__urls_val):
+                                raise JsonSchemaValueException("" + (name_prefix or "data") + ".urls.{data__urls_key}".format(**locals()) + " must be url", value=data__urls_val, name="" + (name_prefix or "data") + ".urls.{data__urls_key}".format(**locals()) + "", definition={'type': 'string', 'format': 'url'}, rule='format')
+                if data__urls_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".urls must not contain "+str(data__urls_keys)+" properties", value=data__urls, name="" + (name_prefix or "data") + ".urls", definition={'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, rule='additionalProperties')
+        if "scripts" in data_keys:
+            data_keys.remove("scripts")
+            data__scripts = data["scripts"]
+            validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_entry_point_group(data__scripts, custom_formats, (name_prefix or "data") + ".scripts")
+        if "gui-scripts" in data_keys:
+            data_keys.remove("gui-scripts")
+            data__guiscripts = data["gui-scripts"]
+            validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_entry_point_group(data__guiscripts, custom_formats, (name_prefix or "data") + ".gui-scripts")
+        if "entry-points" in data_keys:
+            data_keys.remove("entry-points")
+            data__entrypoints = data["entry-points"]
+            data__entrypoints_is_dict = isinstance(data__entrypoints, dict)
+            if data__entrypoints_is_dict:
+                data__entrypoints_keys = set(data__entrypoints.keys())
+                for data__entrypoints_key, data__entrypoints_val in data__entrypoints.items():
+                    if REGEX_PATTERNS['^.+$'].search(data__entrypoints_key):
+                        if data__entrypoints_key in data__entrypoints_keys:
+                            data__entrypoints_keys.remove(data__entrypoints_key)
+                        validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_entry_point_group(data__entrypoints_val, custom_formats, (name_prefix or "data") + ".entry-points.{data__entrypoints_key}".format(**locals()))
+                if data__entrypoints_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".entry-points must not contain "+str(data__entrypoints_keys)+" properties", value=data__entrypoints, name="" + (name_prefix or "data") + ".entry-points", definition={'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}}}, rule='additionalProperties')
+                data__entrypoints_len = len(data__entrypoints)
+                if data__entrypoints_len != 0:
+                    data__entrypoints_property_names = True
+                    for data__entrypoints_key in data__entrypoints:
+                        try:
+                            if isinstance(data__entrypoints_key, str):
+                                if not custom_formats["python-entrypoint-group"](data__entrypoints_key):
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".entry-points must be python-entrypoint-group", value=data__entrypoints_key, name="" + (name_prefix or "data") + ".entry-points", definition={'format': 'python-entrypoint-group'}, rule='format')
+                        except JsonSchemaValueException:
+                            data__entrypoints_property_names = False
+                    if not data__entrypoints_property_names:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".entry-points must be named by propertyName definition", value=data__entrypoints, name="" + (name_prefix or "data") + ".entry-points", definition={'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}}}, rule='propertyNames')
+        if "dependencies" in data_keys:
+            data_keys.remove("dependencies")
+            data__dependencies = data["dependencies"]
+            if not isinstance(data__dependencies, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".dependencies must be array", value=data__dependencies, name="" + (name_prefix or "data") + ".dependencies", definition={'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}, rule='type')
+            data__dependencies_is_list = isinstance(data__dependencies, (list, tuple))
+            if data__dependencies_is_list:
+                data__dependencies_len = len(data__dependencies)
+                for data__dependencies_x, data__dependencies_item in enumerate(data__dependencies):
+                    validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_dependency(data__dependencies_item, custom_formats, (name_prefix or "data") + ".dependencies[{data__dependencies_x}]".format(**locals()))
+        if "optional-dependencies" in data_keys:
+            data_keys.remove("optional-dependencies")
+            data__optionaldependencies = data["optional-dependencies"]
+            if not isinstance(data__optionaldependencies, (dict)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional-dependencies must be object", value=data__optionaldependencies, name="" + (name_prefix or "data") + ".optional-dependencies", definition={'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, rule='type')
+            data__optionaldependencies_is_dict = isinstance(data__optionaldependencies, dict)
+            if data__optionaldependencies_is_dict:
+                data__optionaldependencies_keys = set(data__optionaldependencies.keys())
+                for data__optionaldependencies_key, data__optionaldependencies_val in data__optionaldependencies.items():
+                    if REGEX_PATTERNS['^.+$'].search(data__optionaldependencies_key):
+                        if data__optionaldependencies_key in data__optionaldependencies_keys:
+                            data__optionaldependencies_keys.remove(data__optionaldependencies_key)
+                        if not isinstance(data__optionaldependencies_val, (list, tuple)):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional-dependencies.{data__optionaldependencies_key}".format(**locals()) + " must be array", value=data__optionaldependencies_val, name="" + (name_prefix or "data") + ".optional-dependencies.{data__optionaldependencies_key}".format(**locals()) + "", definition={'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}, rule='type')
+                        data__optionaldependencies_val_is_list = isinstance(data__optionaldependencies_val, (list, tuple))
+                        if data__optionaldependencies_val_is_list:
+                            data__optionaldependencies_val_len = len(data__optionaldependencies_val)
+                            for data__optionaldependencies_val_x, data__optionaldependencies_val_item in enumerate(data__optionaldependencies_val):
+                                validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_dependency(data__optionaldependencies_val_item, custom_formats, (name_prefix or "data") + ".optional-dependencies.{data__optionaldependencies_key}[{data__optionaldependencies_val_x}]".format(**locals()))
+                if data__optionaldependencies_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional-dependencies must not contain "+str(data__optionaldependencies_keys)+" properties", value=data__optionaldependencies, name="" + (name_prefix or "data") + ".optional-dependencies", definition={'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, rule='additionalProperties')
+                data__optionaldependencies_len = len(data__optionaldependencies)
+                if data__optionaldependencies_len != 0:
+                    data__optionaldependencies_property_names = True
+                    for data__optionaldependencies_key in data__optionaldependencies:
+                        try:
+                            if isinstance(data__optionaldependencies_key, str):
+                                if not custom_formats["pep508-identifier"](data__optionaldependencies_key):
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional-dependencies must be pep508-identifier", value=data__optionaldependencies_key, name="" + (name_prefix or "data") + ".optional-dependencies", definition={'format': 'pep508-identifier'}, rule='format')
+                        except JsonSchemaValueException:
+                            data__optionaldependencies_property_names = False
+                    if not data__optionaldependencies_property_names:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".optional-dependencies must be named by propertyName definition", value=data__optionaldependencies, name="" + (name_prefix or "data") + ".optional-dependencies", definition={'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, rule='propertyNames')
+        if "dynamic" in data_keys:
+            data_keys.remove("dynamic")
+            data__dynamic = data["dynamic"]
+            if not isinstance(data__dynamic, (list, tuple)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic must be array", value=data__dynamic, name="" + (name_prefix or "data") + ".dynamic", definition={'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}, rule='type')
+            data__dynamic_is_list = isinstance(data__dynamic, (list, tuple))
+            if data__dynamic_is_list:
+                data__dynamic_len = len(data__dynamic)
+                for data__dynamic_x, data__dynamic_item in enumerate(data__dynamic):
+                    if data__dynamic_item not in ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']:
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic[{data__dynamic_x}]".format(**locals()) + " must be one of ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']", value=data__dynamic_item, name="" + (name_prefix or "data") + ".dynamic[{data__dynamic_x}]".format(**locals()) + "", definition={'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}, rule='enum')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$schema': 'http://json-schema.org/draft-07/schema#', '$id': 'https://packaging.python.org/en/latest/specifications/pyproject-toml/', 'title': 'Package metadata stored in the ``project`` table', '$$description': ['Data structure for the **project** table inside ``pyproject.toml``', '(as initially defined in :pep:`621`)'], 'type': 'object', 'properties': {'name': {'type': 'string', 'description': 'The name (primary identifier) of the project. MUST be statically defined.', 'format': 'pep508-identifier'}, 'version': {'type': 'string', 'description': 'The version of the project as supported by :pep:`440`.', 'format': 'pep440'}, 'description': {'type': 'string', '$$description': ['The `summary description of the project', '<https://packaging.python.org/specifications/core-metadata/#summary>`_']}, 'readme': {'$$description': ['`Full/detailed description of the project in the form of a README', '<https://peps.python.org/pep-0621/#readme>`_', "with meaning similar to the one defined in `core metadata's Description", '<https://packaging.python.org/specifications/core-metadata/#description>`_'], 'oneOf': [{'type': 'string', '$$description': ['Relative path to a text file (UTF-8) containing the full description', 'of the project. If the file path ends in case-insensitive ``.md`` or', '``.rst`` suffixes, then the content-type is respectively', '``text/markdown`` or ``text/x-rst``']}, {'type': 'object', 'allOf': [{'anyOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to a text file containing the full description', 'of the project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', 'description': 'Full text describing the project.'}}, 'required': ['text']}]}, {'properties': {'content-type': {'type': 'string', '$$description': ['Content-type (:rfc:`1341`) of the full description', '(e.g. ``text/markdown``). The ``charset`` parameter is assumed', 'UTF-8 when not present.'], '$comment': 'TODO: add regex pattern or format?'}}, 'required': ['content-type']}]}]}, 'requires-python': {'type': 'string', 'format': 'pep508-versionspec', '$$description': ['`The Python version requirements of the project', '<https://packaging.python.org/specifications/core-metadata/#requires-python>`_.']}, 'license': {'description': '`Project license <https://peps.python.org/pep-0621/#license>`_.', 'oneOf': [{'properties': {'file': {'type': 'string', '$$description': ['Relative path to the file (UTF-8) which contains the license for the', 'project.']}}, 'required': ['file']}, {'properties': {'text': {'type': 'string', '$$description': ['The license of the project whose meaning is that of the', '`License field from the core metadata', '<https://packaging.python.org/specifications/core-metadata/#license>`_.']}}, 'required': ['text']}]}, 'authors': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'authors' of the project.", 'The exact meaning is open to interpretation (e.g. original or primary authors,', 'current maintainers, or owners of the package).']}, 'maintainers': {'type': 'array', 'items': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, '$$description': ["The people or organizations considered to be the 'maintainers' of the project.", 'Similarly to ``authors``, the exact meaning is open to interpretation.']}, 'keywords': {'type': 'array', 'items': {'type': 'string'}, 'description': 'List of keywords to assist searching for the distribution in a larger catalog.'}, 'classifiers': {'type': 'array', 'items': {'type': 'string', 'format': 'trove-classifier', 'description': '`PyPI classifier <https://pypi.org/classifiers/>`_.'}, '$$description': ['`Trove classifiers <https://pypi.org/classifiers/>`_', 'which apply to the project.']}, 'urls': {'type': 'object', 'description': 'URLs associated with the project in the form ``label => value``.', 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', 'format': 'url'}}}, 'scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'gui-scripts': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'entry-points': {'$$description': ['Instruct the installer to expose the given modules/functions via', '``entry-point`` discovery mechanism (useful for plugins).', 'More information available in the `Python packaging guide', '<https://packaging.python.org/specifications/entry-points/>`_.'], 'propertyNames': {'format': 'python-entrypoint-group'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}}}, 'dependencies': {'type': 'array', 'description': 'Project (mandatory) dependencies.', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}, 'optional-dependencies': {'type': 'object', 'description': 'Optional dependency for the project', 'propertyNames': {'format': 'pep508-identifier'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'array', 'items': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}}, 'dynamic': {'type': 'array', '$$description': ['Specifies which fields are intentionally unspecified and expected to be', 'dynamically provided by build tools'], 'items': {'enum': ['version', 'description', 'readme', 'requires-python', 'license', 'authors', 'maintainers', 'keywords', 'classifiers', 'urls', 'scripts', 'gui-scripts', 'entry-points', 'dependencies', 'optional-dependencies']}}}, 'required': ['name'], 'additionalProperties': False, 'if': {'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, 'then': {'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, 'definitions': {'author': {'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, 'entry-point-group': {'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, 'dependency': {'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}}}, rule='additionalProperties')
+    try:
+        try:
+            data_is_dict = isinstance(data, dict)
+            if data_is_dict:
+                data__missing_keys = set(['dynamic']) - data.keys()
+                if data__missing_keys:
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, rule='required')
+                data_keys = set(data.keys())
+                if "dynamic" in data_keys:
+                    data_keys.remove("dynamic")
+                    data__dynamic = data["dynamic"]
+                    data__dynamic_is_list = isinstance(data__dynamic, (list, tuple))
+                    if data__dynamic_is_list:
+                        data__dynamic_contains = False
+                        for data__dynamic_key in data__dynamic:
+                            try:
+                                if data__dynamic_key != "version":
+                                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic must be same as const definition: version", value=data__dynamic_key, name="" + (name_prefix or "data") + ".dynamic", definition={'const': 'version'}, rule='const')
+                                data__dynamic_contains = True
+                                break
+                            except JsonSchemaValueException: pass
+                        if not data__dynamic_contains:
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + ".dynamic must contain one of contains definition", value=data__dynamic, name="" + (name_prefix or "data") + ".dynamic", definition={'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}, rule='contains')
+        except JsonSchemaValueException: pass
+        else:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must NOT match a disallowed definition", value=data, name="" + (name_prefix or "data") + "", definition={'not': {'required': ['dynamic'], 'properties': {'dynamic': {'contains': {'const': 'version'}, '$$description': ['version is listed in ``dynamic``']}}}, '$$comment': ['According to :pep:`621`:', '    If the core metadata specification lists a field as "Required", then', '    the metadata MUST specify the field statically or list it in dynamic', 'In turn, `core metadata`_ defines:', '    The required fields are: Metadata-Version, Name, Version.', '    All the other fields are optional.', 'Since ``Metadata-Version`` is defined by the build back-end, ``name`` and', '``version`` are the only mandatory information in ``pyproject.toml``.', '.. _core metadata: https://packaging.python.org/specifications/core-metadata/']}, rule='not')
+    except JsonSchemaValueException:
+        pass
+    else:
+        data_is_dict = isinstance(data, dict)
+        if data_is_dict:
+            data__missing_keys = set(['version']) - data.keys()
+            if data__missing_keys:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must contain " + (str(sorted(data__missing_keys)) + " properties"), value=data, name="" + (name_prefix or "data") + "", definition={'required': ['version'], '$$description': ['version should be statically defined in the ``version`` field']}, rule='required')
+    return data
+
+def validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_dependency(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (str)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be string", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}, rule='type')
+    if isinstance(data, str):
+        if not custom_formats["pep508"](data):
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must be pep508", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/dependency', 'title': 'Dependency', 'type': 'string', 'description': 'Project dependency specification according to PEP 508', 'format': 'pep508'}, rule='format')
+    return data
+
+def validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_entry_point_group(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        for data_key, data_val in data.items():
+            if REGEX_PATTERNS['^.+$'].search(data_key):
+                if data_key in data_keys:
+                    data_keys.remove(data_key)
+                if not isinstance(data_val, (str)):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".{data_key}".format(**locals()) + " must be string", value=data_val, name="" + (name_prefix or "data") + ".{data_key}".format(**locals()) + "", definition={'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}, rule='type')
+                if isinstance(data_val, str):
+                    if not custom_formats["python-entrypoint-reference"](data_val):
+                        raise JsonSchemaValueException("" + (name_prefix or "data") + ".{data_key}".format(**locals()) + " must be python-entrypoint-reference", value=data_val, name="" + (name_prefix or "data") + ".{data_key}".format(**locals()) + "", definition={'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}, rule='format')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, rule='additionalProperties')
+        data_len = len(data)
+        if data_len != 0:
+            data_property_names = True
+            for data_key in data:
+                try:
+                    if isinstance(data_key, str):
+                        if not custom_formats["python-entrypoint-name"](data_key):
+                            raise JsonSchemaValueException("" + (name_prefix or "data") + " must be python-entrypoint-name", value=data_key, name="" + (name_prefix or "data") + "", definition={'format': 'python-entrypoint-name'}, rule='format')
+                except JsonSchemaValueException:
+                    data_property_names = False
+            if not data_property_names:
+                raise JsonSchemaValueException("" + (name_prefix or "data") + " must be named by propertyName definition", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/entry-point-group', 'title': 'Entry-points', 'type': 'object', '$$description': ['Entry-points are grouped together to indicate what sort of capabilities they', 'provide.', 'See the `packaging guides', '<https://packaging.python.org/specifications/entry-points/>`_', 'and `setuptools docs', '<https://setuptools.pypa.io/en/latest/userguide/entry_point.html>`_', 'for more information.'], 'propertyNames': {'format': 'python-entrypoint-name'}, 'additionalProperties': False, 'patternProperties': {'^.+$': {'type': 'string', '$$description': ['Reference to a Python object. It is either in the form', '``importable.module``, or ``importable.module:object.attr``.'], 'format': 'python-entrypoint-reference', '$comment': 'https://packaging.python.org/specifications/entry-points/'}}}, rule='propertyNames')
+    return data
+
+def validate_https___packaging_python_org_en_latest_specifications_pyproject_toml___definitions_author(data, custom_formats={}, name_prefix=None):
+    if not isinstance(data, (dict)):
+        raise JsonSchemaValueException("" + (name_prefix or "data") + " must be object", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, rule='type')
+    data_is_dict = isinstance(data, dict)
+    if data_is_dict:
+        data_keys = set(data.keys())
+        if "name" in data_keys:
+            data_keys.remove("name")
+            data__name = data["name"]
+            if not isinstance(data__name, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".name must be string", value=data__name, name="" + (name_prefix or "data") + ".name", definition={'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, rule='type')
+        if "email" in data_keys:
+            data_keys.remove("email")
+            data__email = data["email"]
+            if not isinstance(data__email, (str)):
+                raise JsonSchemaValueException("" + (name_prefix or "data") + ".email must be string", value=data__email, name="" + (name_prefix or "data") + ".email", definition={'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}, rule='type')
+            if isinstance(data__email, str):
+                if not REGEX_PATTERNS["idn-email_re_pattern"].match(data__email):
+                    raise JsonSchemaValueException("" + (name_prefix or "data") + ".email must be idn-email", value=data__email, name="" + (name_prefix or "data") + ".email", definition={'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}, rule='format')
+        if data_keys:
+            raise JsonSchemaValueException("" + (name_prefix or "data") + " must not contain "+str(data_keys)+" properties", value=data, name="" + (name_prefix or "data") + "", definition={'$id': '#/definitions/author', 'title': 'Author or Maintainer', '$comment': 'https://peps.python.org/pep-0621/#authors-maintainers', 'type': 'object', 'additionalProperties': False, 'properties': {'name': {'type': 'string', '$$description': ['MUST be a valid email name, i.e. whatever can be put as a name, before an', 'email, in :rfc:`822`.']}, 'email': {'type': 'string', 'format': 'idn-email', 'description': 'MUST be a valid email address'}}}, rule='additionalProperties')
+    return data
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/formats.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/formats.py
new file mode 100644
index 00000000..153b1f0b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/_validate_pyproject/formats.py
@@ -0,0 +1,375 @@
+"""
+The functions in this module are used to validate schemas with the
+`format JSON Schema keyword
+<https://json-schema.org/understanding-json-schema/reference/string#format>`_.
+
+The correspondence is given by replacing the ``_`` character in the name of the
+function with a ``-`` to obtain the format name and vice versa.
+"""
+
+import builtins
+import logging
+import os
+import re
+import string
+import typing
+from itertools import chain as _chain
+
+if typing.TYPE_CHECKING:
+    from typing_extensions import Literal
+
+_logger = logging.getLogger(__name__)
+
+# -------------------------------------------------------------------------------------
+# PEP 440
+
+VERSION_PATTERN = r"""
+    v?
+    (?:
+        (?:(?P<epoch>[0-9]+)!)?                           # epoch
+        (?P<release>[0-9]+(?:\.[0-9]+)*)                  # release segment
+        (?P<pre>                                          # pre-release
+            [-_\.]?
+            (?P<pre_l>alpha|a|beta|b|preview|pre|c|rc)
+            [-_\.]?
+            (?P<pre_n>[0-9]+)?
+        )?
+        (?P<post>                                         # post release
+            (?:-(?P<post_n1>[0-9]+))
+            |
+            (?:
+                [-_\.]?
+                (?P<post_l>post|rev|r)
+                [-_\.]?
+                (?P<post_n2>[0-9]+)?
+            )
+        )?
+        (?P<dev>                                          # dev release
+            [-_\.]?
+            (?P<dev_l>dev)
+            [-_\.]?
+            (?P<dev_n>[0-9]+)?
+        )?
+    )
+    (?:\+(?P<local>[a-z0-9]+(?:[-_\.][a-z0-9]+)*))?       # local version
+"""
+
+VERSION_REGEX = re.compile(r"^\s*" + VERSION_PATTERN + r"\s*$", re.X | re.I)
+
+
+def pep440(version: str) -> bool:
+    """See :ref:`PyPA's version specification <pypa:version-specifiers>`
+    (initially introduced in :pep:`440`).
+    """
+    return VERSION_REGEX.match(version) is not None
+
+
+# -------------------------------------------------------------------------------------
+# PEP 508
+
+PEP508_IDENTIFIER_PATTERN = r"([A-Z0-9]|[A-Z0-9][A-Z0-9._-]*[A-Z0-9])"
+PEP508_IDENTIFIER_REGEX = re.compile(f"^{PEP508_IDENTIFIER_PATTERN}$", re.I)
+
+
+def pep508_identifier(name: str) -> bool:
+    """See :ref:`PyPA's name specification <pypa:name-format>`
+    (initially introduced in :pep:`508#names`).
+    """
+    return PEP508_IDENTIFIER_REGEX.match(name) is not None
+
+
+try:
+    try:
+        from packaging import requirements as _req
+    except ImportError:  # pragma: no cover
+        # let's try setuptools vendored version
+        from setuptools._vendor.packaging import (  # type: ignore[no-redef]
+            requirements as _req,
+        )
+
+    def pep508(value: str) -> bool:
+        """See :ref:`PyPA's dependency specifiers <pypa:dependency-specifiers>`
+        (initially introduced in :pep:`508`).
+        """
+        try:
+            _req.Requirement(value)
+            return True
+        except _req.InvalidRequirement:
+            return False
+
+except ImportError:  # pragma: no cover
+    _logger.warning(
+        "Could not find an installation of `packaging`. Requirements, dependencies and "
+        "versions might not be validated. "
+        "To enforce validation, please install `packaging`."
+    )
+
+    def pep508(value: str) -> bool:
+        return True
+
+
+def pep508_versionspec(value: str) -> bool:
+    """Expression that can be used to specify/lock versions (including ranges)
+    See ``versionspec`` in :ref:`PyPA's dependency specifiers
+    <pypa:dependency-specifiers>` (initially introduced in :pep:`508`).
+    """
+    if any(c in value for c in (";", "]", "@")):
+        # In PEP 508:
+        # conditional markers, extras and URL specs are not included in the
+        # versionspec
+        return False
+    # Let's pretend we have a dependency called `requirement` with the given
+    # version spec, then we can reuse the pep508 function for validation:
+    return pep508(f"requirement{value}")
+
+
+# -------------------------------------------------------------------------------------
+# PEP 517
+
+
+def pep517_backend_reference(value: str) -> bool:
+    """See PyPA's specification for defining build-backend references
+    introduced in :pep:`517#source-trees`.
+
+    This is similar to an entry-point reference (e.g., ``package.module:object``).
+    """
+    module, _, obj = value.partition(":")
+    identifiers = (i.strip() for i in _chain(module.split("."), obj.split(".")))
+    return all(python_identifier(i) for i in identifiers if i)
+
+
+# -------------------------------------------------------------------------------------
+# Classifiers - PEP 301
+
+
+def _download_classifiers() -> str:
+    import ssl
+    from email.message import Message
+    from urllib.request import urlopen
+
+    url = "https://pypi.org/pypi?:action=list_classifiers"
+    context = ssl.create_default_context()
+    with urlopen(url, context=context) as response:  # noqa: S310 (audit URLs)
+        headers = Message()
+        headers["content_type"] = response.getheader("content-type", "text/plain")
+        return response.read().decode(headers.get_param("charset", "utf-8"))  # type: ignore[no-any-return]
+
+
+class _TroveClassifier:
+    """The ``trove_classifiers`` package is the official way of validating classifiers,
+    however this package might not be always available.
+    As a workaround we can still download a list from PyPI.
+    We also don't want to be over strict about it, so simply skipping silently is an
+    option (classifiers will be validated anyway during the upload to PyPI).
+    """
+
+    downloaded: typing.Union[None, "Literal[False]", typing.Set[str]]
+
+    def __init__(self) -> None:
+        self.downloaded = None
+        self._skip_download = False
+        # None => not cached yet
+        # False => cache not available
+        self.__name__ = "trove_classifier"  # Emulate a public function
+
+    def _disable_download(self) -> None:
+        # This is a private API. Only setuptools has the consent of using it.
+        self._skip_download = True
+
+    def __call__(self, value: str) -> bool:
+        if self.downloaded is False or self._skip_download is True:
+            return True
+
+        if os.getenv("NO_NETWORK") or os.getenv("VALIDATE_PYPROJECT_NO_NETWORK"):
+            self.downloaded = False
+            msg = (
+                "Install ``trove-classifiers`` to ensure proper validation. "
+                "Skipping download of classifiers list from PyPI (NO_NETWORK)."
+            )
+            _logger.debug(msg)
+            return True
+
+        if self.downloaded is None:
+            msg = (
+                "Install ``trove-classifiers`` to ensure proper validation. "
+                "Meanwhile a list of classifiers will be downloaded from PyPI."
+            )
+            _logger.debug(msg)
+            try:
+                self.downloaded = set(_download_classifiers().splitlines())
+            except Exception:
+                self.downloaded = False
+                _logger.debug("Problem with download, skipping validation")
+                return True
+
+        return value in self.downloaded or value.lower().startswith("private ::")
+
+
+try:
+    from trove_classifiers import classifiers as _trove_classifiers
+
+    def trove_classifier(value: str) -> bool:
+        """See https://pypi.org/classifiers/"""
+        return value in _trove_classifiers or value.lower().startswith("private ::")
+
+except ImportError:  # pragma: no cover
+    trove_classifier = _TroveClassifier()
+
+
+# -------------------------------------------------------------------------------------
+# Stub packages - PEP 561
+
+
+def pep561_stub_name(value: str) -> bool:
+    """Name of a directory containing type stubs.
+    It must follow the name scheme ``<package>-stubs`` as defined in
+    :pep:`561#stub-only-packages`.
+    """
+    top, *children = value.split(".")
+    if not top.endswith("-stubs"):
+        return False
+    return python_module_name(".".join([top[: -len("-stubs")], *children]))
+
+
+# -------------------------------------------------------------------------------------
+# Non-PEP related
+
+
+def url(value: str) -> bool:
+    """Valid URL (validation uses :obj:`urllib.parse`).
+    For maximum compatibility please make sure to include a ``scheme`` prefix
+    in your URL (e.g. ``http://``).
+    """
+    from urllib.parse import urlparse
+
+    try:
+        parts = urlparse(value)
+        if not parts.scheme:
+            _logger.warning(
+                "For maximum compatibility please make sure to include a "
+                "`scheme` prefix in your URL (e.g. 'http://'). "
+                f"Given value: {value}"
+            )
+            if not (value.startswith("/") or value.startswith("\\") or "@" in value):
+                parts = urlparse(f"http://{value}")
+
+        return bool(parts.scheme and parts.netloc)
+    except Exception:
+        return False
+
+
+# https://packaging.python.org/specifications/entry-points/
+ENTRYPOINT_PATTERN = r"[^\[\s=]([^=]*[^\s=])?"
+ENTRYPOINT_REGEX = re.compile(f"^{ENTRYPOINT_PATTERN}$", re.I)
+RECOMMEDED_ENTRYPOINT_PATTERN = r"[\w.-]+"
+RECOMMEDED_ENTRYPOINT_REGEX = re.compile(f"^{RECOMMEDED_ENTRYPOINT_PATTERN}$", re.I)
+ENTRYPOINT_GROUP_PATTERN = r"\w+(\.\w+)*"
+ENTRYPOINT_GROUP_REGEX = re.compile(f"^{ENTRYPOINT_GROUP_PATTERN}$", re.I)
+
+
+def python_identifier(value: str) -> bool:
+    """Can be used as identifier in Python.
+    (Validation uses :obj:`str.isidentifier`).
+    """
+    return value.isidentifier()
+
+
+def python_qualified_identifier(value: str) -> bool:
+    """
+    Python "dotted identifier", i.e. a sequence of :obj:`python_identifier`
+    concatenated with ``"."`` (e.g.: ``package.module.submodule``).
+    """
+    if value.startswith(".") or value.endswith("."):
+        return False
+    return all(python_identifier(m) for m in value.split("."))
+
+
+def python_module_name(value: str) -> bool:
+    """Module name that can be used in an ``import``-statement in Python.
+    See :obj:`python_qualified_identifier`.
+    """
+    return python_qualified_identifier(value)
+
+
+def python_module_name_relaxed(value: str) -> bool:
+    """Similar to :obj:`python_module_name`, but relaxed to also accept
+    dash characters (``-``) and cover special cases like ``pip-run``.
+
+    It is recommended, however, that beginners avoid dash characters,
+    as they require advanced knowledge about Python internals.
+
+    The following are disallowed:
+
+    * names starting/ending in dashes,
+    * names ending in ``-stubs`` (potentially collide with :obj:`pep561_stub_name`).
+    """
+    if value.startswith("-") or value.endswith("-"):
+        return False
+    if value.endswith("-stubs"):
+        return False  # Avoid collision with PEP 561
+    return python_module_name(value.replace("-", "_"))
+
+
+def python_entrypoint_group(value: str) -> bool:
+    """See ``Data model > group`` in the :ref:`PyPA's entry-points specification
+    <pypa:entry-points>`.
+    """
+    return ENTRYPOINT_GROUP_REGEX.match(value) is not None
+
+
+def python_entrypoint_name(value: str) -> bool:
+    """See ``Data model > name`` in the :ref:`PyPA's entry-points specification
+    <pypa:entry-points>`.
+    """
+    if not ENTRYPOINT_REGEX.match(value):
+        return False
+    if not RECOMMEDED_ENTRYPOINT_REGEX.match(value):
+        msg = f"Entry point `{value}` does not follow recommended pattern: "
+        msg += RECOMMEDED_ENTRYPOINT_PATTERN
+        _logger.warning(msg)
+    return True
+
+
+def python_entrypoint_reference(value: str) -> bool:
+    """Reference to a Python object using in the format::
+
+        importable.module:object.attr
+
+    See ``Data model >object reference`` in the :ref:`PyPA's entry-points specification
+    <pypa:entry-points>`.
+    """
+    module, _, rest = value.partition(":")
+    if "[" in rest:
+        obj, _, extras_ = rest.partition("[")
+        if extras_.strip()[-1] != "]":
+            return False
+        extras = (x.strip() for x in extras_.strip(string.whitespace + "[]").split(","))
+        if not all(pep508_identifier(e) for e in extras):
+            return False
+        _logger.warning(f"`{value}` - using extras for entry points is not recommended")
+    else:
+        obj = rest
+
+    module_parts = module.split(".")
+    identifiers = _chain(module_parts, obj.split(".")) if rest else module_parts
+    return all(python_identifier(i.strip()) for i in identifiers)
+
+
+def uint8(value: builtins.int) -> bool:
+    r"""Unsigned 8-bit integer (:math:`0 \leq x < 2^8`)"""
+    return 0 <= value < 2**8
+
+
+def uint16(value: builtins.int) -> bool:
+    r"""Unsigned 16-bit integer (:math:`0 \leq x < 2^{16}`)"""
+    return 0 <= value < 2**16
+
+
+def uint(value: builtins.int) -> bool:
+    r"""Unsigned 64-bit integer (:math:`0 \leq x < 2^{64}`)"""
+    return 0 <= value < 2**64
+
+
+def int(value: builtins.int) -> bool:
+    r"""Signed 64-bit integer (:math:`-2^{63} \leq x < 2^{63}`)"""
+    return -(2**63) <= value < 2**63
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/distutils.schema.json b/backend/venv/lib/python3.12/site-packages/setuptools/config/distutils.schema.json
new file mode 100644
index 00000000..93cd2e86
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/distutils.schema.json
@@ -0,0 +1,26 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+
+  "$id": "https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html",
+  "title": "``tool.distutils`` table",
+  "$$description": [
+    "**EXPERIMENTAL** (NOT OFFICIALLY SUPPORTED): Use ``tool.distutils``",
+    "subtables to configure arguments for ``distutils`` commands.",
+    "Originally, ``distutils`` allowed developers to configure arguments for",
+    "``setup.py`` commands via `distutils configuration files",
+    "<https://setuptools.pypa.io/en/latest/deprecated/distutils/configfile.html>`_.",
+    "See also `the old Python docs <https://docs.python.org/3.11/install/>_`."
+  ],
+
+  "type": "object",
+  "properties": {
+    "global": {
+      "type": "object",
+      "description": "Global options applied to all ``distutils`` commands"
+    }
+  },
+  "patternProperties": {
+    ".+": {"type": "object"}
+  },
+  "$comment": "TODO: Is there a practical way of making this schema more specific?"
+}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/expand.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/expand.py
new file mode 100644
index 00000000..531f9650
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/expand.py
@@ -0,0 +1,452 @@
+"""Utility functions to expand configuration directives or special values
+(such glob patterns).
+
+We can split the process of interpreting configuration files into 2 steps:
+
+1. The parsing the file contents from strings to value objects
+   that can be understand by Python (for example a string with a comma
+   separated list of keywords into an actual Python list of strings).
+
+2. The expansion (or post-processing) of these values according to the
+   semantics ``setuptools`` assign to them (for example a configuration field
+   with the ``file:`` directive should be expanded from a list of file paths to
+   a single string with the contents of those files concatenated)
+
+This module focus on the second step, and therefore allow sharing the expansion
+functions among several configuration file formats.
+
+**PRIVATE MODULE**: API reserved for setuptools internal usage only.
+"""
+
+from __future__ import annotations
+
+import ast
+import importlib
+import os
+import pathlib
+import sys
+from collections.abc import Iterable, Iterator, Mapping
+from configparser import ConfigParser
+from glob import iglob
+from importlib.machinery import ModuleSpec, all_suffixes
+from itertools import chain
+from pathlib import Path
+from types import ModuleType, TracebackType
+from typing import TYPE_CHECKING, Any, Callable, TypeVar
+
+from .. import _static
+from .._path import StrPath, same_path as _same_path
+from ..discovery import find_package_path
+from ..warnings import SetuptoolsWarning
+
+from distutils.errors import DistutilsOptionError
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from setuptools.dist import Distribution
+
+_K = TypeVar("_K")
+_V_co = TypeVar("_V_co", covariant=True)
+
+
+class StaticModule:
+    """Proxy to a module object that avoids executing arbitrary code."""
+
+    def __init__(self, name: str, spec: ModuleSpec) -> None:
+        module = ast.parse(pathlib.Path(spec.origin).read_bytes())  # type: ignore[arg-type] # Let it raise an error on None
+        vars(self).update(locals())
+        del self.self
+
+    def _find_assignments(self) -> Iterator[tuple[ast.AST, ast.AST]]:
+        for statement in self.module.body:
+            if isinstance(statement, ast.Assign):
+                yield from ((target, statement.value) for target in statement.targets)
+            elif isinstance(statement, ast.AnnAssign) and statement.value:
+                yield (statement.target, statement.value)
+
+    def __getattr__(self, attr: str):
+        """Attempt to load an attribute "statically", via :func:`ast.literal_eval`."""
+        try:
+            return next(
+                ast.literal_eval(value)
+                for target, value in self._find_assignments()
+                if isinstance(target, ast.Name) and target.id == attr
+            )
+        except Exception as e:
+            raise AttributeError(f"{self.name} has no attribute {attr}") from e
+
+
+def glob_relative(
+    patterns: Iterable[str], root_dir: StrPath | None = None
+) -> list[str]:
+    """Expand the list of glob patterns, but preserving relative paths.
+
+    :param list[str] patterns: List of glob patterns
+    :param str root_dir: Path to which globs should be relative
+                         (current directory by default)
+    :rtype: list
+    """
+    glob_characters = {'*', '?', '[', ']', '{', '}'}
+    expanded_values = []
+    root_dir = root_dir or os.getcwd()
+    for value in patterns:
+        # Has globby characters?
+        if any(char in value for char in glob_characters):
+            # then expand the glob pattern while keeping paths *relative*:
+            glob_path = os.path.abspath(os.path.join(root_dir, value))
+            expanded_values.extend(
+                sorted(
+                    os.path.relpath(path, root_dir).replace(os.sep, "/")
+                    for path in iglob(glob_path, recursive=True)
+                )
+            )
+
+        else:
+            # take the value as-is
+            path = os.path.relpath(value, root_dir).replace(os.sep, "/")
+            expanded_values.append(path)
+
+    return expanded_values
+
+
+def read_files(
+    filepaths: StrPath | Iterable[StrPath], root_dir: StrPath | None = None
+) -> str:
+    """Return the content of the files concatenated using ``\n`` as str
+
+    This function is sandboxed and won't reach anything outside ``root_dir``
+
+    (By default ``root_dir`` is the current directory).
+    """
+    from more_itertools import always_iterable
+
+    root_dir = os.path.abspath(root_dir or os.getcwd())
+    _filepaths = (os.path.join(root_dir, path) for path in always_iterable(filepaths))
+    return '\n'.join(
+        _read_file(path)
+        for path in _filter_existing_files(_filepaths)
+        if _assert_local(path, root_dir)
+    )
+
+
+def _filter_existing_files(filepaths: Iterable[StrPath]) -> Iterator[StrPath]:
+    for path in filepaths:
+        if os.path.isfile(path):
+            yield path
+        else:
+            SetuptoolsWarning.emit(f"File {path!r} cannot be found")
+
+
+def _read_file(filepath: bytes | StrPath) -> str:
+    with open(filepath, encoding='utf-8') as f:
+        return f.read()
+
+
+def _assert_local(filepath: StrPath, root_dir: str):
+    if Path(os.path.abspath(root_dir)) not in Path(os.path.abspath(filepath)).parents:
+        msg = f"Cannot access {filepath!r} (or anything outside {root_dir!r})"
+        raise DistutilsOptionError(msg)
+
+    return True
+
+
+def read_attr(
+    attr_desc: str,
+    package_dir: Mapping[str, str] | None = None,
+    root_dir: StrPath | None = None,
+) -> Any:
+    """Reads the value of an attribute from a module.
+
+    This function will try to read the attributed statically first
+    (via :func:`ast.literal_eval`), and only evaluate the module if it fails.
+
+    Examples:
+        read_attr("package.attr")
+        read_attr("package.module.attr")
+
+    :param str attr_desc: Dot-separated string describing how to reach the
+        attribute (see examples above)
+    :param dict[str, str] package_dir: Mapping of package names to their
+        location in disk (represented by paths relative to ``root_dir``).
+    :param str root_dir: Path to directory containing all the packages in
+        ``package_dir`` (current directory by default).
+    :rtype: str
+    """
+    root_dir = root_dir or os.getcwd()
+    attrs_path = attr_desc.strip().split('.')
+    attr_name = attrs_path.pop()
+    module_name = '.'.join(attrs_path)
+    module_name = module_name or '__init__'
+    path = _find_module(module_name, package_dir, root_dir)
+    spec = _find_spec(module_name, path)
+
+    try:
+        value = getattr(StaticModule(module_name, spec), attr_name)
+        # XXX: Is marking as static contents coming from modules too optimistic?
+        return _static.attempt_conversion(value)
+    except Exception:
+        # fallback to evaluate module
+        module = _load_spec(spec, module_name)
+        return getattr(module, attr_name)
+
+
+def _find_spec(module_name: str, module_path: StrPath | None) -> ModuleSpec:
+    spec = importlib.util.spec_from_file_location(module_name, module_path)
+    spec = spec or importlib.util.find_spec(module_name)
+
+    if spec is None:
+        raise ModuleNotFoundError(module_name)
+
+    return spec
+
+
+def _load_spec(spec: ModuleSpec, module_name: str) -> ModuleType:
+    name = getattr(spec, "__name__", module_name)
+    if name in sys.modules:
+        return sys.modules[name]
+    module = importlib.util.module_from_spec(spec)
+    sys.modules[name] = module  # cache (it also ensures `==` works on loaded items)
+    assert spec.loader is not None
+    spec.loader.exec_module(module)
+    return module
+
+
+def _find_module(
+    module_name: str, package_dir: Mapping[str, str] | None, root_dir: StrPath
+) -> str | None:
+    """Find the path to the module named ``module_name``,
+    considering the ``package_dir`` in the build configuration and ``root_dir``.
+
+    >>> tmp = getfixture('tmpdir')
+    >>> _ = tmp.ensure("a/b/c.py")
+    >>> _ = tmp.ensure("a/b/d/__init__.py")
+    >>> r = lambda x: x.replace(str(tmp), "tmp").replace(os.sep, "/")
+    >>> r(_find_module("a.b.c", None, tmp))
+    'tmp/a/b/c.py'
+    >>> r(_find_module("f.g.h", {"": "1", "f": "2", "f.g": "3", "f.g.h": "a/b/d"}, tmp))
+    'tmp/a/b/d/__init__.py'
+    """
+    path_start = find_package_path(module_name, package_dir or {}, root_dir)
+    candidates = chain.from_iterable(
+        (f"{path_start}{ext}", os.path.join(path_start, f"__init__{ext}"))
+        for ext in all_suffixes()
+    )
+    return next((x for x in candidates if os.path.isfile(x)), None)
+
+
+def resolve_class(
+    qualified_class_name: str,
+    package_dir: Mapping[str, str] | None = None,
+    root_dir: StrPath | None = None,
+) -> Callable:
+    """Given a qualified class name, return the associated class object"""
+    root_dir = root_dir or os.getcwd()
+    idx = qualified_class_name.rfind('.')
+    class_name = qualified_class_name[idx + 1 :]
+    pkg_name = qualified_class_name[:idx]
+
+    path = _find_module(pkg_name, package_dir, root_dir)
+    module = _load_spec(_find_spec(pkg_name, path), pkg_name)
+    return getattr(module, class_name)
+
+
+def cmdclass(
+    values: dict[str, str],
+    package_dir: Mapping[str, str] | None = None,
+    root_dir: StrPath | None = None,
+) -> dict[str, Callable]:
+    """Given a dictionary mapping command names to strings for qualified class
+    names, apply :func:`resolve_class` to the dict values.
+    """
+    return {k: resolve_class(v, package_dir, root_dir) for k, v in values.items()}
+
+
+def find_packages(
+    *,
+    namespaces=True,
+    fill_package_dir: dict[str, str] | None = None,
+    root_dir: StrPath | None = None,
+    **kwargs,
+) -> list[str]:
+    """Works similarly to :func:`setuptools.find_packages`, but with all
+    arguments given as keyword arguments. Moreover, ``where`` can be given
+    as a list (the results will be simply concatenated).
+
+    When the additional keyword argument ``namespaces`` is ``True``, it will
+    behave like :func:`setuptools.find_namespace_packages`` (i.e. include
+    implicit namespaces as per :pep:`420`).
+
+    The ``where`` argument will be considered relative to ``root_dir`` (or the current
+    working directory when ``root_dir`` is not given).
+
+    If the ``fill_package_dir`` argument is passed, this function will consider it as a
+    similar data structure to the ``package_dir`` configuration parameter add fill-in
+    any missing package location.
+
+    :rtype: list
+    """
+    from more_itertools import always_iterable, unique_everseen
+
+    from setuptools.discovery import construct_package_dir
+
+    # check "not namespaces" first due to python/mypy#6232
+    if not namespaces:
+        from setuptools.discovery import PackageFinder
+    else:
+        from setuptools.discovery import PEP420PackageFinder as PackageFinder
+
+    root_dir = root_dir or os.curdir
+    where = kwargs.pop('where', ['.'])
+    packages: list[str] = []
+    fill_package_dir = {} if fill_package_dir is None else fill_package_dir
+    search = list(unique_everseen(always_iterable(where)))
+
+    if len(search) == 1 and all(not _same_path(search[0], x) for x in (".", root_dir)):
+        fill_package_dir.setdefault("", search[0])
+
+    for path in search:
+        package_path = _nest_path(root_dir, path)
+        pkgs = PackageFinder.find(package_path, **kwargs)
+        packages.extend(pkgs)
+        if pkgs and not (
+            fill_package_dir.get("") == path or os.path.samefile(package_path, root_dir)
+        ):
+            fill_package_dir.update(construct_package_dir(pkgs, path))
+
+    return packages
+
+
+def _nest_path(parent: StrPath, path: StrPath) -> str:
+    path = parent if path in {".", ""} else os.path.join(parent, path)
+    return os.path.normpath(path)
+
+
+def version(value: Callable | Iterable[str | int] | str) -> str:
+    """When getting the version directly from an attribute,
+    it should be normalised to string.
+    """
+    _value = value() if callable(value) else value
+
+    if isinstance(_value, str):
+        return _value
+    if hasattr(_value, '__iter__'):
+        return '.'.join(map(str, _value))
+    return f'{_value}'
+
+
+def canonic_package_data(package_data: dict) -> dict:
+    if "*" in package_data:
+        package_data[""] = package_data.pop("*")
+    return package_data
+
+
+def canonic_data_files(
+    data_files: list | dict, root_dir: StrPath | None = None
+) -> list[tuple[str, list[str]]]:
+    """For compatibility with ``setup.py``, ``data_files`` should be a list
+    of pairs instead of a dict.
+
+    This function also expands glob patterns.
+    """
+    if isinstance(data_files, list):
+        return data_files
+
+    return [
+        (dest, glob_relative(patterns, root_dir))
+        for dest, patterns in data_files.items()
+    ]
+
+
+def entry_points(
+    text: str, text_source: str = "entry-points"
+) -> dict[str, dict[str, str]]:
+    """Given the contents of entry-points file,
+    process it into a 2-level dictionary (``dict[str, dict[str, str]]``).
+    The first level keys are entry-point groups, the second level keys are
+    entry-point names, and the second level values are references to objects
+    (that correspond to the entry-point value).
+    """
+    # Using undocumented behaviour, see python/typeshed#12700
+    parser = ConfigParser(default_section=None, delimiters=("=",))  # type: ignore[call-overload]
+    parser.optionxform = str  # case sensitive
+    parser.read_string(text, text_source)
+    groups = {k: dict(v.items()) for k, v in parser.items()}
+    groups.pop(parser.default_section, None)
+    return groups
+
+
+class EnsurePackagesDiscovered:
+    """Some expand functions require all the packages to already be discovered before
+    they run, e.g. :func:`read_attr`, :func:`resolve_class`, :func:`cmdclass`.
+
+    Therefore in some cases we will need to run autodiscovery during the evaluation of
+    the configuration. However, it is better to postpone calling package discovery as
+    much as possible, because some parameters can influence it (e.g. ``package_dir``),
+    and those might not have been processed yet.
+    """
+
+    def __init__(self, distribution: Distribution) -> None:
+        self._dist = distribution
+        self._called = False
+
+    def __call__(self):
+        """Trigger the automatic package discovery, if it is still necessary."""
+        if not self._called:
+            self._called = True
+            self._dist.set_defaults(name=False)  # Skip name, we can still be parsing
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_value: BaseException | None,
+        traceback: TracebackType | None,
+    ):
+        if self._called:
+            self._dist.set_defaults.analyse_name()  # Now we can set a default name
+
+    def _get_package_dir(self) -> Mapping[str, str]:
+        self()
+        pkg_dir = self._dist.package_dir
+        return {} if pkg_dir is None else pkg_dir
+
+    @property
+    def package_dir(self) -> Mapping[str, str]:
+        """Proxy to ``package_dir`` that may trigger auto-discovery when used."""
+        return LazyMappingProxy(self._get_package_dir)
+
+
+class LazyMappingProxy(Mapping[_K, _V_co]):
+    """Mapping proxy that delays resolving the target object, until really needed.
+
+    >>> def obtain_mapping():
+    ...     print("Running expensive function!")
+    ...     return {"key": "value", "other key": "other value"}
+    >>> mapping = LazyMappingProxy(obtain_mapping)
+    >>> mapping["key"]
+    Running expensive function!
+    'value'
+    >>> mapping["other key"]
+    'other value'
+    """
+
+    def __init__(self, obtain_mapping_value: Callable[[], Mapping[_K, _V_co]]) -> None:
+        self._obtain = obtain_mapping_value
+        self._value: Mapping[_K, _V_co] | None = None
+
+    def _target(self) -> Mapping[_K, _V_co]:
+        if self._value is None:
+            self._value = self._obtain()
+        return self._value
+
+    def __getitem__(self, key: _K) -> _V_co:
+        return self._target()[key]
+
+    def __len__(self) -> int:
+        return len(self._target())
+
+    def __iter__(self) -> Iterator[_K]:
+        return iter(self._target())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/pyprojecttoml.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/pyprojecttoml.py
new file mode 100644
index 00000000..fd6c5968
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/pyprojecttoml.py
@@ -0,0 +1,468 @@
+"""
+Load setuptools configuration from ``pyproject.toml`` files.
+
+**PRIVATE MODULE**: API reserved for setuptools internal usage only.
+
+To read project metadata, consider using
+``build.util.project_wheel_metadata`` (https://pypi.org/project/build/).
+For simple scenarios, you can also try parsing the file directly
+with the help of ``tomllib`` or ``tomli``.
+"""
+
+from __future__ import annotations
+
+import logging
+import os
+from collections.abc import Mapping
+from contextlib import contextmanager
+from functools import partial
+from types import TracebackType
+from typing import TYPE_CHECKING, Any, Callable
+
+from .._path import StrPath
+from ..errors import FileError, InvalidConfigError
+from ..warnings import SetuptoolsWarning
+from . import expand as _expand
+from ._apply_pyprojecttoml import _PREVIOUSLY_DEFINED, _MissingDynamic, apply as _apply
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from setuptools.dist import Distribution
+
+_logger = logging.getLogger(__name__)
+
+
+def load_file(filepath: StrPath) -> dict:
+    from ..compat.py310 import tomllib
+
+    with open(filepath, "rb") as file:
+        return tomllib.load(file)
+
+
+def validate(config: dict, filepath: StrPath) -> bool:
+    from . import _validate_pyproject as validator
+
+    trove_classifier = validator.FORMAT_FUNCTIONS.get("trove-classifier")
+    if hasattr(trove_classifier, "_disable_download"):
+        # Improve reproducibility by default. See abravalheri/validate-pyproject#31
+        trove_classifier._disable_download()  # type: ignore[union-attr]
+
+    try:
+        return validator.validate(config)
+    except validator.ValidationError as ex:
+        summary = f"configuration error: {ex.summary}"
+        if ex.name.strip("`") != "project":
+            # Probably it is just a field missing/misnamed, not worthy the verbosity...
+            _logger.debug(summary)
+            _logger.debug(ex.details)
+
+        error = f"invalid pyproject.toml config: {ex.name}."
+        raise ValueError(f"{error}\n{summary}") from None
+
+
+def apply_configuration(
+    dist: Distribution,
+    filepath: StrPath,
+    ignore_option_errors: bool = False,
+) -> Distribution:
+    """Apply the configuration from a ``pyproject.toml`` file into an existing
+    distribution object.
+    """
+    config = read_configuration(filepath, True, ignore_option_errors, dist)
+    return _apply(dist, config, filepath)
+
+
+def read_configuration(
+    filepath: StrPath,
+    expand: bool = True,
+    ignore_option_errors: bool = False,
+    dist: Distribution | None = None,
+) -> dict[str, Any]:
+    """Read given configuration file and returns options from it as a dict.
+
+    :param str|unicode filepath: Path to configuration file in the ``pyproject.toml``
+        format.
+
+    :param bool expand: Whether to expand directives and other computed values
+        (i.e. post-process the given configuration)
+
+    :param bool ignore_option_errors: Whether to silently ignore
+        options, values of which could not be resolved (e.g. due to exceptions
+        in directives such as file:, attr:, etc.).
+        If False exceptions are propagated as expected.
+
+    :param Distribution|None: Distribution object to which the configuration refers.
+        If not given a dummy object will be created and discarded after the
+        configuration is read. This is used for auto-discovery of packages and in the
+        case a dynamic configuration (e.g. ``attr`` or ``cmdclass``) is expanded.
+        When ``expand=False`` this object is simply ignored.
+
+    :rtype: dict
+    """
+    filepath = os.path.abspath(filepath)
+
+    if not os.path.isfile(filepath):
+        raise FileError(f"Configuration file {filepath!r} does not exist.")
+
+    asdict = load_file(filepath) or {}
+    project_table = asdict.get("project", {})
+    tool_table = asdict.get("tool", {})
+    setuptools_table = tool_table.get("setuptools", {})
+    if not asdict or not (project_table or setuptools_table):
+        return {}  # User is not using pyproject to configure setuptools
+
+    if "setuptools" in asdict.get("tools", {}):
+        # let the user know they probably have a typo in their metadata
+        _ToolsTypoInMetadata.emit()
+
+    if "distutils" in tool_table:
+        _ExperimentalConfiguration.emit(subject="[tool.distutils]")
+
+    # There is an overall sense in the community that making include_package_data=True
+    # the default would be an improvement.
+    # `ini2toml` backfills include_package_data=False when nothing is explicitly given,
+    # therefore setting a default here is backwards compatible.
+    if dist and dist.include_package_data is not None:
+        setuptools_table.setdefault("include-package-data", dist.include_package_data)
+    else:
+        setuptools_table.setdefault("include-package-data", True)
+    # Persist changes:
+    asdict["tool"] = tool_table
+    tool_table["setuptools"] = setuptools_table
+
+    if "ext-modules" in setuptools_table:
+        _ExperimentalConfiguration.emit(subject="[tool.setuptools.ext-modules]")
+
+    with _ignore_errors(ignore_option_errors):
+        # Don't complain about unrelated errors (e.g. tools not using the "tool" table)
+        subset = {"project": project_table, "tool": {"setuptools": setuptools_table}}
+        validate(subset, filepath)
+
+    if expand:
+        root_dir = os.path.dirname(filepath)
+        return expand_configuration(asdict, root_dir, ignore_option_errors, dist)
+
+    return asdict
+
+
+def expand_configuration(
+    config: dict,
+    root_dir: StrPath | None = None,
+    ignore_option_errors: bool = False,
+    dist: Distribution | None = None,
+) -> dict:
+    """Given a configuration with unresolved fields (e.g. dynamic, cmdclass, ...)
+    find their final values.
+
+    :param dict config: Dict containing the configuration for the distribution
+    :param str root_dir: Top-level directory for the distribution/project
+        (the same directory where ``pyproject.toml`` is place)
+    :param bool ignore_option_errors: see :func:`read_configuration`
+    :param Distribution|None: Distribution object to which the configuration refers.
+        If not given a dummy object will be created and discarded after the
+        configuration is read. Used in the case a dynamic configuration
+        (e.g. ``attr`` or ``cmdclass``).
+
+    :rtype: dict
+    """
+    return _ConfigExpander(config, root_dir, ignore_option_errors, dist).expand()
+
+
+class _ConfigExpander:
+    def __init__(
+        self,
+        config: dict,
+        root_dir: StrPath | None = None,
+        ignore_option_errors: bool = False,
+        dist: Distribution | None = None,
+    ) -> None:
+        self.config = config
+        self.root_dir = root_dir or os.getcwd()
+        self.project_cfg = config.get("project", {})
+        self.dynamic = self.project_cfg.get("dynamic", [])
+        self.setuptools_cfg = config.get("tool", {}).get("setuptools", {})
+        self.dynamic_cfg = self.setuptools_cfg.get("dynamic", {})
+        self.ignore_option_errors = ignore_option_errors
+        self._dist = dist
+        self._referenced_files = set[str]()
+
+    def _ensure_dist(self) -> Distribution:
+        from setuptools.dist import Distribution
+
+        attrs = {"src_root": self.root_dir, "name": self.project_cfg.get("name", None)}
+        return self._dist or Distribution(attrs)
+
+    def _process_field(self, container: dict, field: str, fn: Callable):
+        if field in container:
+            with _ignore_errors(self.ignore_option_errors):
+                container[field] = fn(container[field])
+
+    def _canonic_package_data(self, field="package-data"):
+        package_data = self.setuptools_cfg.get(field, {})
+        return _expand.canonic_package_data(package_data)
+
+    def expand(self):
+        self._expand_packages()
+        self._canonic_package_data()
+        self._canonic_package_data("exclude-package-data")
+
+        # A distribution object is required for discovering the correct package_dir
+        dist = self._ensure_dist()
+        ctx = _EnsurePackagesDiscovered(dist, self.project_cfg, self.setuptools_cfg)
+        with ctx as ensure_discovered:
+            package_dir = ensure_discovered.package_dir
+            self._expand_data_files()
+            self._expand_cmdclass(package_dir)
+            self._expand_all_dynamic(dist, package_dir)
+
+        dist._referenced_files.update(self._referenced_files)
+        return self.config
+
+    def _expand_packages(self):
+        packages = self.setuptools_cfg.get("packages")
+        if packages is None or isinstance(packages, (list, tuple)):
+            return
+
+        find = packages.get("find")
+        if isinstance(find, dict):
+            find["root_dir"] = self.root_dir
+            find["fill_package_dir"] = self.setuptools_cfg.setdefault("package-dir", {})
+            with _ignore_errors(self.ignore_option_errors):
+                self.setuptools_cfg["packages"] = _expand.find_packages(**find)
+
+    def _expand_data_files(self):
+        data_files = partial(_expand.canonic_data_files, root_dir=self.root_dir)
+        self._process_field(self.setuptools_cfg, "data-files", data_files)
+
+    def _expand_cmdclass(self, package_dir: Mapping[str, str]):
+        root_dir = self.root_dir
+        cmdclass = partial(_expand.cmdclass, package_dir=package_dir, root_dir=root_dir)
+        self._process_field(self.setuptools_cfg, "cmdclass", cmdclass)
+
+    def _expand_all_dynamic(self, dist: Distribution, package_dir: Mapping[str, str]):
+        special = (  # need special handling
+            "version",
+            "readme",
+            "entry-points",
+            "scripts",
+            "gui-scripts",
+            "classifiers",
+            "dependencies",
+            "optional-dependencies",
+        )
+        # `_obtain` functions are assumed to raise appropriate exceptions/warnings.
+        obtained_dynamic = {
+            field: self._obtain(dist, field, package_dir)
+            for field in self.dynamic
+            if field not in special
+        }
+        obtained_dynamic.update(
+            self._obtain_entry_points(dist, package_dir) or {},
+            version=self._obtain_version(dist, package_dir),
+            readme=self._obtain_readme(dist),
+            classifiers=self._obtain_classifiers(dist),
+            dependencies=self._obtain_dependencies(dist),
+            optional_dependencies=self._obtain_optional_dependencies(dist),
+        )
+        # `None` indicates there is nothing in `tool.setuptools.dynamic` but the value
+        # might have already been set by setup.py/extensions, so avoid overwriting.
+        updates = {k: v for k, v in obtained_dynamic.items() if v is not None}
+        self.project_cfg.update(updates)
+
+    def _ensure_previously_set(self, dist: Distribution, field: str):
+        previous = _PREVIOUSLY_DEFINED[field](dist)
+        if previous is None and not self.ignore_option_errors:
+            msg = (
+                f"No configuration found for dynamic {field!r}.\n"
+                "Some dynamic fields need to be specified via `tool.setuptools.dynamic`"
+                "\nothers must be specified via the equivalent attribute in `setup.py`."
+            )
+            raise InvalidConfigError(msg)
+
+    def _expand_directive(
+        self, specifier: str, directive, package_dir: Mapping[str, str]
+    ):
+        from more_itertools import always_iterable
+
+        with _ignore_errors(self.ignore_option_errors):
+            root_dir = self.root_dir
+            if "file" in directive:
+                self._referenced_files.update(always_iterable(directive["file"]))
+                return _expand.read_files(directive["file"], root_dir)
+            if "attr" in directive:
+                return _expand.read_attr(directive["attr"], package_dir, root_dir)
+            raise ValueError(f"invalid `{specifier}`: {directive!r}")
+        return None
+
+    def _obtain(self, dist: Distribution, field: str, package_dir: Mapping[str, str]):
+        if field in self.dynamic_cfg:
+            return self._expand_directive(
+                f"tool.setuptools.dynamic.{field}",
+                self.dynamic_cfg[field],
+                package_dir,
+            )
+        self._ensure_previously_set(dist, field)
+        return None
+
+    def _obtain_version(self, dist: Distribution, package_dir: Mapping[str, str]):
+        # Since plugins can set version, let's silently skip if it cannot be obtained
+        if "version" in self.dynamic and "version" in self.dynamic_cfg:
+            return _expand.version(
+                # We already do an early check for the presence of "version"
+                self._obtain(dist, "version", package_dir)  # pyright: ignore[reportArgumentType]
+            )
+        return None
+
+    def _obtain_readme(self, dist: Distribution) -> dict[str, str] | None:
+        if "readme" not in self.dynamic:
+            return None
+
+        dynamic_cfg = self.dynamic_cfg
+        if "readme" in dynamic_cfg:
+            return {
+                # We already do an early check for the presence of "readme"
+                "text": self._obtain(dist, "readme", {}),
+                "content-type": dynamic_cfg["readme"].get("content-type", "text/x-rst"),
+            }  # pyright: ignore[reportReturnType]
+
+        self._ensure_previously_set(dist, "readme")
+        return None
+
+    def _obtain_entry_points(
+        self, dist: Distribution, package_dir: Mapping[str, str]
+    ) -> dict[str, dict[str, Any]] | None:
+        fields = ("entry-points", "scripts", "gui-scripts")
+        if not any(field in self.dynamic for field in fields):
+            return None
+
+        text = self._obtain(dist, "entry-points", package_dir)
+        if text is None:
+            return None
+
+        groups = _expand.entry_points(text)
+        # Any is str | dict[str, str], but causes variance issues
+        expanded: dict[str, dict[str, Any]] = {"entry-points": groups}
+
+        def _set_scripts(field: str, group: str):
+            if group in groups:
+                value = groups.pop(group)
+                if field not in self.dynamic:
+                    raise InvalidConfigError(_MissingDynamic.details(field, value))
+                expanded[field] = value
+
+        _set_scripts("scripts", "console_scripts")
+        _set_scripts("gui-scripts", "gui_scripts")
+
+        return expanded
+
+    def _obtain_classifiers(self, dist: Distribution):
+        if "classifiers" in self.dynamic:
+            value = self._obtain(dist, "classifiers", {})
+            if value:
+                return value.splitlines()
+        return None
+
+    def _obtain_dependencies(self, dist: Distribution):
+        if "dependencies" in self.dynamic:
+            value = self._obtain(dist, "dependencies", {})
+            if value:
+                return _parse_requirements_list(value)
+        return None
+
+    def _obtain_optional_dependencies(self, dist: Distribution):
+        if "optional-dependencies" not in self.dynamic:
+            return None
+        if "optional-dependencies" in self.dynamic_cfg:
+            optional_dependencies_map = self.dynamic_cfg["optional-dependencies"]
+            assert isinstance(optional_dependencies_map, dict)
+            return {
+                group: _parse_requirements_list(
+                    self._expand_directive(
+                        f"tool.setuptools.dynamic.optional-dependencies.{group}",
+                        directive,
+                        {},
+                    )
+                )
+                for group, directive in optional_dependencies_map.items()
+            }
+        self._ensure_previously_set(dist, "optional-dependencies")
+        return None
+
+
+def _parse_requirements_list(value):
+    return [
+        line
+        for line in value.splitlines()
+        if line.strip() and not line.strip().startswith("#")
+    ]
+
+
+@contextmanager
+def _ignore_errors(ignore_option_errors: bool):
+    if not ignore_option_errors:
+        yield
+        return
+
+    try:
+        yield
+    except Exception as ex:
+        _logger.debug(f"ignored error: {ex.__class__.__name__} - {ex}")
+
+
+class _EnsurePackagesDiscovered(_expand.EnsurePackagesDiscovered):
+    def __init__(
+        self, distribution: Distribution, project_cfg: dict, setuptools_cfg: dict
+    ) -> None:
+        super().__init__(distribution)
+        self._project_cfg = project_cfg
+        self._setuptools_cfg = setuptools_cfg
+
+    def __enter__(self) -> Self:
+        """When entering the context, the values of ``packages``, ``py_modules`` and
+        ``package_dir`` that are missing in ``dist`` are copied from ``setuptools_cfg``.
+        """
+        dist, cfg = self._dist, self._setuptools_cfg
+        package_dir: dict[str, str] = cfg.setdefault("package-dir", {})
+        package_dir.update(dist.package_dir or {})
+        dist.package_dir = package_dir  # needs to be the same object
+
+        dist.set_defaults._ignore_ext_modules()  # pyproject.toml-specific behaviour
+
+        # Set `name`, `py_modules` and `packages` in dist to short-circuit
+        # auto-discovery, but avoid overwriting empty lists purposefully set by users.
+        if dist.metadata.name is None:
+            dist.metadata.name = self._project_cfg.get("name")
+        if dist.py_modules is None:
+            dist.py_modules = cfg.get("py-modules")
+        if dist.packages is None:
+            dist.packages = cfg.get("packages")
+
+        return super().__enter__()
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_value: BaseException | None,
+        traceback: TracebackType | None,
+    ) -> None:
+        """When exiting the context, if values of ``packages``, ``py_modules`` and
+        ``package_dir`` are missing in ``setuptools_cfg``, copy from ``dist``.
+        """
+        # If anything was discovered set them back, so they count in the final config.
+        self._setuptools_cfg.setdefault("packages", self._dist.packages)
+        self._setuptools_cfg.setdefault("py-modules", self._dist.py_modules)
+        return super().__exit__(exc_type, exc_value, traceback)
+
+
+class _ExperimentalConfiguration(SetuptoolsWarning):
+    _SUMMARY = (
+        "`{subject}` in `pyproject.toml` is still *experimental* "
+        "and likely to change in future releases."
+    )
+
+
+class _ToolsTypoInMetadata(SetuptoolsWarning):
+    _SUMMARY = (
+        "Ignoring [tools.setuptools] in pyproject.toml, did you mean [tool.setuptools]?"
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/setupcfg.py b/backend/venv/lib/python3.12/site-packages/setuptools/config/setupcfg.py
new file mode 100644
index 00000000..633aa9d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/setupcfg.py
@@ -0,0 +1,780 @@
+"""
+Load setuptools configuration from ``setup.cfg`` files.
+
+**API will be made private in the future**
+
+To read project metadata, consider using
+``build.util.project_wheel_metadata`` (https://pypi.org/project/build/).
+For simple scenarios, you can also try parsing the file directly
+with the help of ``configparser``.
+"""
+
+from __future__ import annotations
+
+import contextlib
+import functools
+import os
+from collections import defaultdict
+from collections.abc import Iterable, Iterator
+from functools import partial, wraps
+from typing import TYPE_CHECKING, Any, Callable, ClassVar, Generic, TypeVar, cast
+
+from packaging.markers import default_environment as marker_env
+from packaging.requirements import InvalidRequirement, Requirement
+from packaging.version import InvalidVersion, Version
+
+from .. import _static
+from .._path import StrPath
+from ..errors import FileError, OptionError
+from ..warnings import SetuptoolsDeprecationWarning
+from . import expand
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+    from setuptools.dist import Distribution
+
+    from distutils.dist import DistributionMetadata
+
+SingleCommandOptions: TypeAlias = dict[str, tuple[str, Any]]
+"""Dict that associate the name of the options of a particular command to a
+tuple. The first element of the tuple indicates the origin of the option value
+(e.g. the name of the configuration file where it was read from),
+while the second element of the tuple is the option value itself
+"""
+AllCommandOptions: TypeAlias = dict[str, SingleCommandOptions]
+"""cmd name => its options"""
+Target = TypeVar("Target", "Distribution", "DistributionMetadata")
+
+
+def read_configuration(
+    filepath: StrPath, find_others: bool = False, ignore_option_errors: bool = False
+) -> dict:
+    """Read given configuration file and returns options from it as a dict.
+
+    :param str|unicode filepath: Path to configuration file
+        to get options from.
+
+    :param bool find_others: Whether to search for other configuration files
+        which could be on in various places.
+
+    :param bool ignore_option_errors: Whether to silently ignore
+        options, values of which could not be resolved (e.g. due to exceptions
+        in directives such as file:, attr:, etc.).
+        If False exceptions are propagated as expected.
+
+    :rtype: dict
+    """
+    from setuptools.dist import Distribution
+
+    dist = Distribution()
+    filenames = dist.find_config_files() if find_others else []
+    handlers = _apply(dist, filepath, filenames, ignore_option_errors)
+    return configuration_to_dict(handlers)
+
+
+def apply_configuration(dist: Distribution, filepath: StrPath) -> Distribution:
+    """Apply the configuration from a ``setup.cfg`` file into an existing
+    distribution object.
+    """
+    _apply(dist, filepath)
+    dist._finalize_requires()
+    return dist
+
+
+def _apply(
+    dist: Distribution,
+    filepath: StrPath,
+    other_files: Iterable[StrPath] = (),
+    ignore_option_errors: bool = False,
+) -> tuple[ConfigMetadataHandler, ConfigOptionsHandler]:
+    """Read configuration from ``filepath`` and applies to the ``dist`` object."""
+    from setuptools.dist import _Distribution
+
+    filepath = os.path.abspath(filepath)
+
+    if not os.path.isfile(filepath):
+        raise FileError(f'Configuration file {filepath} does not exist.')
+
+    current_directory = os.getcwd()
+    os.chdir(os.path.dirname(filepath))
+    filenames = [*other_files, filepath]
+
+    try:
+        # TODO: Temporary cast until mypy 1.12 is released with upstream fixes from typeshed
+        _Distribution.parse_config_files(dist, filenames=cast(list[str], filenames))
+        handlers = parse_configuration(
+            dist, dist.command_options, ignore_option_errors=ignore_option_errors
+        )
+        dist._finalize_license_files()
+    finally:
+        os.chdir(current_directory)
+
+    return handlers
+
+
+def _get_option(target_obj: Distribution | DistributionMetadata, key: str):
+    """
+    Given a target object and option key, get that option from
+    the target object, either through a get_{key} method or
+    from an attribute directly.
+    """
+    getter_name = f'get_{key}'
+    by_attribute = functools.partial(getattr, target_obj, key)
+    getter = getattr(target_obj, getter_name, by_attribute)
+    return getter()
+
+
+def configuration_to_dict(
+    handlers: Iterable[
+        ConfigHandler[Distribution] | ConfigHandler[DistributionMetadata]
+    ],
+) -> dict:
+    """Returns configuration data gathered by given handlers as a dict.
+
+    :param Iterable[ConfigHandler] handlers: Handlers list,
+        usually from parse_configuration()
+
+    :rtype: dict
+    """
+    config_dict: dict = defaultdict(dict)
+
+    for handler in handlers:
+        for option in handler.set_options:
+            value = _get_option(handler.target_obj, option)
+            config_dict[handler.section_prefix][option] = value
+
+    return config_dict
+
+
+def parse_configuration(
+    distribution: Distribution,
+    command_options: AllCommandOptions,
+    ignore_option_errors: bool = False,
+) -> tuple[ConfigMetadataHandler, ConfigOptionsHandler]:
+    """Performs additional parsing of configuration options
+    for a distribution.
+
+    Returns a list of used option handlers.
+
+    :param Distribution distribution:
+    :param dict command_options:
+    :param bool ignore_option_errors: Whether to silently ignore
+        options, values of which could not be resolved (e.g. due to exceptions
+        in directives such as file:, attr:, etc.).
+        If False exceptions are propagated as expected.
+    :rtype: list
+    """
+    with expand.EnsurePackagesDiscovered(distribution) as ensure_discovered:
+        options = ConfigOptionsHandler(
+            distribution,
+            command_options,
+            ignore_option_errors,
+            ensure_discovered,
+        )
+
+        options.parse()
+        if not distribution.package_dir:
+            distribution.package_dir = options.package_dir  # Filled by `find_packages`
+
+        meta = ConfigMetadataHandler(
+            distribution.metadata,
+            command_options,
+            ignore_option_errors,
+            ensure_discovered,
+            distribution.package_dir,
+            distribution.src_root,
+        )
+        meta.parse()
+        distribution._referenced_files.update(
+            options._referenced_files, meta._referenced_files
+        )
+
+    return meta, options
+
+
+def _warn_accidental_env_marker_misconfig(label: str, orig_value: str, parsed: list):
+    """Because users sometimes misinterpret this configuration:
+
+    [options.extras_require]
+    foo = bar;python_version<"4"
+
+    It looks like one requirement with an environment marker
+    but because there is no newline, it's parsed as two requirements
+    with a semicolon as separator.
+
+    Therefore, if:
+        * input string does not contain a newline AND
+        * parsed result contains two requirements AND
+        * parsing of the two parts from the result ("<first>;<second>")
+        leads in a valid Requirement with a valid marker
+    a UserWarning is shown to inform the user about the possible problem.
+    """
+    if "\n" in orig_value or len(parsed) != 2:
+        return
+
+    markers = marker_env().keys()
+
+    try:
+        req = Requirement(parsed[1])
+        if req.name in markers:
+            _AmbiguousMarker.emit(field=label, req=parsed[1])
+    except InvalidRequirement as ex:
+        if any(parsed[1].startswith(marker) for marker in markers):
+            msg = _AmbiguousMarker.message(field=label, req=parsed[1])
+            raise InvalidRequirement(msg) from ex
+
+
+class ConfigHandler(Generic[Target]):
+    """Handles metadata supplied in configuration files."""
+
+    section_prefix: str
+    """Prefix for config sections handled by this handler.
+    Must be provided by class heirs.
+
+    """
+
+    aliases: ClassVar[dict[str, str]] = {}
+    """Options aliases.
+    For compatibility with various packages. E.g.: d2to1 and pbr.
+    Note: `-` in keys is replaced with `_` by config parser.
+
+    """
+
+    def __init__(
+        self,
+        target_obj: Target,
+        options: AllCommandOptions,
+        ignore_option_errors,
+        ensure_discovered: expand.EnsurePackagesDiscovered,
+    ) -> None:
+        self.ignore_option_errors = ignore_option_errors
+        self.target_obj: Target = target_obj
+        self.sections = dict(self._section_options(options))
+        self.set_options: list[str] = []
+        self.ensure_discovered = ensure_discovered
+        self._referenced_files = set[str]()
+        """After parsing configurations, this property will enumerate
+        all files referenced by the "file:" directive. Private API for setuptools only.
+        """
+
+    @classmethod
+    def _section_options(
+        cls, options: AllCommandOptions
+    ) -> Iterator[tuple[str, SingleCommandOptions]]:
+        for full_name, value in options.items():
+            pre, _sep, name = full_name.partition(cls.section_prefix)
+            if pre:
+                continue
+            yield name.lstrip('.'), value
+
+    @property
+    def parsers(self):
+        """Metadata item name to parser function mapping."""
+        raise NotImplementedError(
+            f'{self.__class__.__name__} must provide .parsers property'
+        )
+
+    def __setitem__(self, option_name, value) -> None:
+        target_obj = self.target_obj
+
+        # Translate alias into real name.
+        option_name = self.aliases.get(option_name, option_name)
+
+        try:
+            current_value = getattr(target_obj, option_name)
+        except AttributeError as e:
+            raise KeyError(option_name) from e
+
+        if current_value:
+            # Already inhabited. Skipping.
+            return
+
+        try:
+            parsed = self.parsers.get(option_name, lambda x: x)(value)
+        except (Exception,) * self.ignore_option_errors:
+            return
+
+        simple_setter = functools.partial(target_obj.__setattr__, option_name)
+        setter = getattr(target_obj, f"set_{option_name}", simple_setter)
+        setter(parsed)
+
+        self.set_options.append(option_name)
+
+    @classmethod
+    def _parse_list(cls, value, separator=','):
+        """Represents value as a list.
+
+        Value is split either by separator (defaults to comma) or by lines.
+
+        :param value:
+        :param separator: List items separator character.
+        :rtype: list
+        """
+        if isinstance(value, list):  # _get_parser_compound case
+            return value
+
+        if '\n' in value:
+            value = value.splitlines()
+        else:
+            value = value.split(separator)
+
+        return [chunk.strip() for chunk in value if chunk.strip()]
+
+    @classmethod
+    def _parse_dict(cls, value):
+        """Represents value as a dict.
+
+        :param value:
+        :rtype: dict
+        """
+        separator = '='
+        result = {}
+        for line in cls._parse_list(value):
+            key, sep, val = line.partition(separator)
+            if sep != separator:
+                raise OptionError(f"Unable to parse option value to dict: {value}")
+            result[key.strip()] = val.strip()
+
+        return result
+
+    @classmethod
+    def _parse_bool(cls, value):
+        """Represents value as boolean.
+
+        :param value:
+        :rtype: bool
+        """
+        value = value.lower()
+        return value in ('1', 'true', 'yes')
+
+    @classmethod
+    def _exclude_files_parser(cls, key):
+        """Returns a parser function to make sure field inputs
+        are not files.
+
+        Parses a value after getting the key so error messages are
+        more informative.
+
+        :param key:
+        :rtype: callable
+        """
+
+        def parser(value):
+            exclude_directive = 'file:'
+            if value.startswith(exclude_directive):
+                raise ValueError(
+                    f'Only strings are accepted for the {key} field, '
+                    'files are not accepted'
+                )
+            return _static.Str(value)
+
+        return parser
+
+    def _parse_file(self, value, root_dir: StrPath | None):
+        """Represents value as a string, allowing including text
+        from nearest files using `file:` directive.
+
+        Directive is sandboxed and won't reach anything outside
+        directory with setup.py.
+
+        Examples:
+            file: README.rst, CHANGELOG.md, src/file.txt
+
+        :param str value:
+        :rtype: str
+        """
+        include_directive = 'file:'
+
+        if not isinstance(value, str):
+            return value
+
+        if not value.startswith(include_directive):
+            return _static.Str(value)
+
+        spec = value[len(include_directive) :]
+        filepaths = [path.strip() for path in spec.split(',')]
+        self._referenced_files.update(filepaths)
+        # XXX: Is marking as static contents coming from files too optimistic?
+        return _static.Str(expand.read_files(filepaths, root_dir))
+
+    def _parse_attr(self, value, package_dir, root_dir: StrPath):
+        """Represents value as a module attribute.
+
+        Examples:
+            attr: package.attr
+            attr: package.module.attr
+
+        :param str value:
+        :rtype: str
+        """
+        attr_directive = 'attr:'
+        if not value.startswith(attr_directive):
+            return _static.Str(value)
+
+        attr_desc = value.replace(attr_directive, '')
+
+        # Make sure package_dir is populated correctly, so `attr:` directives can work
+        package_dir.update(self.ensure_discovered.package_dir)
+        return expand.read_attr(attr_desc, package_dir, root_dir)
+
+    @classmethod
+    def _get_parser_compound(cls, *parse_methods):
+        """Returns parser function to represents value as a list.
+
+        Parses a value applying given methods one after another.
+
+        :param parse_methods:
+        :rtype: callable
+        """
+
+        def parse(value):
+            parsed = value
+
+            for method in parse_methods:
+                parsed = method(parsed)
+
+            return parsed
+
+        return parse
+
+    @classmethod
+    def _parse_section_to_dict_with_key(cls, section_options, values_parser):
+        """Parses section options into a dictionary.
+
+        Applies a given parser to each option in a section.
+
+        :param dict section_options:
+        :param callable values_parser: function with 2 args corresponding to key, value
+        :rtype: dict
+        """
+        value = {}
+        for key, (_, val) in section_options.items():
+            value[key] = values_parser(key, val)
+        return value
+
+    @classmethod
+    def _parse_section_to_dict(cls, section_options, values_parser=None):
+        """Parses section options into a dictionary.
+
+        Optionally applies a given parser to each value.
+
+        :param dict section_options:
+        :param callable values_parser: function with 1 arg corresponding to option value
+        :rtype: dict
+        """
+        parser = (lambda _, v: values_parser(v)) if values_parser else (lambda _, v: v)
+        return cls._parse_section_to_dict_with_key(section_options, parser)
+
+    def parse_section(self, section_options) -> None:
+        """Parses configuration file section.
+
+        :param dict section_options:
+        """
+        for name, (_, value) in section_options.items():
+            with contextlib.suppress(KeyError):
+                # Keep silent for a new option may appear anytime.
+                self[name] = value
+
+    def parse(self) -> None:
+        """Parses configuration file items from one
+        or more related sections.
+
+        """
+        for section_name, section_options in self.sections.items():
+            method_postfix = ''
+            if section_name:  # [section.option] variant
+                method_postfix = f"_{section_name}"
+
+            section_parser_method: Callable | None = getattr(
+                self,
+                # Dots in section names are translated into dunderscores.
+                f'parse_section{method_postfix}'.replace('.', '__'),
+                None,
+            )
+
+            if section_parser_method is None:
+                raise OptionError(
+                    "Unsupported distribution option section: "
+                    f"[{self.section_prefix}.{section_name}]"
+                )
+
+            section_parser_method(section_options)
+
+    def _deprecated_config_handler(self, func, msg, **kw):
+        """this function will wrap around parameters that are deprecated
+
+        :param msg: deprecation message
+        :param func: function to be wrapped around
+        """
+
+        @wraps(func)
+        def config_handler(*args, **kwargs):
+            kw.setdefault("stacklevel", 2)
+            _DeprecatedConfig.emit("Deprecated config in `setup.cfg`", msg, **kw)
+            return func(*args, **kwargs)
+
+        return config_handler
+
+
+class ConfigMetadataHandler(ConfigHandler["DistributionMetadata"]):
+    section_prefix = 'metadata'
+
+    aliases = {
+        'home_page': 'url',
+        'summary': 'description',
+        'classifier': 'classifiers',
+        'platform': 'platforms',
+    }
+
+    strict_mode = False
+    """We need to keep it loose, to be partially compatible with
+    `pbr` and `d2to1` packages which also uses `metadata` section.
+
+    """
+
+    def __init__(
+        self,
+        target_obj: DistributionMetadata,
+        options: AllCommandOptions,
+        ignore_option_errors: bool,
+        ensure_discovered: expand.EnsurePackagesDiscovered,
+        package_dir: dict | None = None,
+        root_dir: StrPath | None = os.curdir,
+    ) -> None:
+        super().__init__(target_obj, options, ignore_option_errors, ensure_discovered)
+        self.package_dir = package_dir
+        self.root_dir = root_dir
+
+    @property
+    def parsers(self):
+        """Metadata item name to parser function mapping."""
+        parse_list_static = self._get_parser_compound(self._parse_list, _static.List)
+        parse_dict_static = self._get_parser_compound(self._parse_dict, _static.Dict)
+        parse_file = partial(self._parse_file, root_dir=self.root_dir)
+        exclude_files_parser = self._exclude_files_parser
+
+        return {
+            'author': _static.Str,
+            'author_email': _static.Str,
+            'maintainer': _static.Str,
+            'maintainer_email': _static.Str,
+            'platforms': parse_list_static,
+            'keywords': parse_list_static,
+            'provides': parse_list_static,
+            'obsoletes': parse_list_static,
+            'classifiers': self._get_parser_compound(parse_file, parse_list_static),
+            'license': exclude_files_parser('license'),
+            'license_files': parse_list_static,
+            'description': parse_file,
+            'long_description': parse_file,
+            'long_description_content_type': _static.Str,
+            'version': self._parse_version,  # Cannot be marked as dynamic
+            'url': _static.Str,
+            'project_urls': parse_dict_static,
+        }
+
+    def _parse_version(self, value):
+        """Parses `version` option value.
+
+        :param value:
+        :rtype: str
+
+        """
+        version = self._parse_file(value, self.root_dir)
+
+        if version != value:
+            version = version.strip()
+            # Be strict about versions loaded from file because it's easy to
+            # accidentally include newlines and other unintended content
+            try:
+                Version(version)
+            except InvalidVersion as e:
+                raise OptionError(
+                    f'Version loaded from {value} does not '
+                    f'comply with PEP 440: {version}'
+                ) from e
+
+            return version
+
+        return expand.version(self._parse_attr(value, self.package_dir, self.root_dir))
+
+
+class ConfigOptionsHandler(ConfigHandler["Distribution"]):
+    section_prefix = 'options'
+
+    def __init__(
+        self,
+        target_obj: Distribution,
+        options: AllCommandOptions,
+        ignore_option_errors: bool,
+        ensure_discovered: expand.EnsurePackagesDiscovered,
+    ) -> None:
+        super().__init__(target_obj, options, ignore_option_errors, ensure_discovered)
+        self.root_dir = target_obj.src_root
+        self.package_dir: dict[str, str] = {}  # To be filled by `find_packages`
+
+    @classmethod
+    def _parse_list_semicolon(cls, value):
+        return cls._parse_list(value, separator=';')
+
+    def _parse_file_in_root(self, value):
+        return self._parse_file(value, root_dir=self.root_dir)
+
+    def _parse_requirements_list(self, label: str, value: str):
+        # Parse a requirements list, either by reading in a `file:`, or a list.
+        parsed = self._parse_list_semicolon(self._parse_file_in_root(value))
+        _warn_accidental_env_marker_misconfig(label, value, parsed)
+        # Filter it to only include lines that are not comments. `parse_list`
+        # will have stripped each line and filtered out empties.
+        return _static.List(line for line in parsed if not line.startswith("#"))
+        # ^-- Use `_static.List` to mark a non-`Dynamic` Core Metadata
+
+    @property
+    def parsers(self):
+        """Metadata item name to parser function mapping."""
+        parse_list = self._parse_list
+        parse_bool = self._parse_bool
+        parse_cmdclass = self._parse_cmdclass
+
+        return {
+            'zip_safe': parse_bool,
+            'include_package_data': parse_bool,
+            'package_dir': self._parse_dict,
+            'scripts': parse_list,
+            'eager_resources': parse_list,
+            'dependency_links': parse_list,
+            'namespace_packages': self._deprecated_config_handler(
+                parse_list,
+                "The namespace_packages parameter is deprecated, "
+                "consider using implicit namespaces instead (PEP 420).",
+                # TODO: define due date, see setuptools.dist:check_nsp.
+            ),
+            'install_requires': partial(  # Core Metadata
+                self._parse_requirements_list, "install_requires"
+            ),
+            'setup_requires': self._parse_list_semicolon,
+            'packages': self._parse_packages,
+            'entry_points': self._parse_file_in_root,
+            'py_modules': parse_list,
+            'python_requires': _static.SpecifierSet,  # Core Metadata
+            'cmdclass': parse_cmdclass,
+        }
+
+    def _parse_cmdclass(self, value):
+        package_dir = self.ensure_discovered.package_dir
+        return expand.cmdclass(self._parse_dict(value), package_dir, self.root_dir)
+
+    def _parse_packages(self, value):
+        """Parses `packages` option value.
+
+        :param value:
+        :rtype: list
+        """
+        find_directives = ['find:', 'find_namespace:']
+        trimmed_value = value.strip()
+
+        if trimmed_value not in find_directives:
+            return self._parse_list(value)
+
+        # Read function arguments from a dedicated section.
+        find_kwargs = self.parse_section_packages__find(
+            self.sections.get('packages.find', {})
+        )
+
+        find_kwargs.update(
+            namespaces=(trimmed_value == find_directives[1]),
+            root_dir=self.root_dir,
+            fill_package_dir=self.package_dir,
+        )
+
+        return expand.find_packages(**find_kwargs)
+
+    def parse_section_packages__find(self, section_options):
+        """Parses `packages.find` configuration file section.
+
+        To be used in conjunction with _parse_packages().
+
+        :param dict section_options:
+        """
+        section_data = self._parse_section_to_dict(section_options, self._parse_list)
+
+        valid_keys = ['where', 'include', 'exclude']
+        find_kwargs = {k: v for k, v in section_data.items() if k in valid_keys and v}
+
+        where = find_kwargs.get('where')
+        if where is not None:
+            find_kwargs['where'] = where[0]  # cast list to single val
+
+        return find_kwargs
+
+    def parse_section_entry_points(self, section_options) -> None:
+        """Parses `entry_points` configuration file section.
+
+        :param dict section_options:
+        """
+        parsed = self._parse_section_to_dict(section_options, self._parse_list)
+        self['entry_points'] = parsed
+
+    def _parse_package_data(self, section_options):
+        package_data = self._parse_section_to_dict(section_options, self._parse_list)
+        return expand.canonic_package_data(package_data)
+
+    def parse_section_package_data(self, section_options) -> None:
+        """Parses `package_data` configuration file section.
+
+        :param dict section_options:
+        """
+        self['package_data'] = self._parse_package_data(section_options)
+
+    def parse_section_exclude_package_data(self, section_options) -> None:
+        """Parses `exclude_package_data` configuration file section.
+
+        :param dict section_options:
+        """
+        self['exclude_package_data'] = self._parse_package_data(section_options)
+
+    def parse_section_extras_require(self, section_options) -> None:  # Core Metadata
+        """Parses `extras_require` configuration file section.
+
+        :param dict section_options:
+        """
+        parsed = self._parse_section_to_dict_with_key(
+            section_options,
+            lambda k, v: self._parse_requirements_list(f"extras_require[{k}]", v),
+        )
+
+        self['extras_require'] = _static.Dict(parsed)
+        # ^-- Use `_static.Dict` to mark a non-`Dynamic` Core Metadata
+
+    def parse_section_data_files(self, section_options) -> None:
+        """Parses `data_files` configuration file section.
+
+        :param dict section_options:
+        """
+        parsed = self._parse_section_to_dict(section_options, self._parse_list)
+        self['data_files'] = expand.canonic_data_files(parsed, self.root_dir)
+
+
+class _AmbiguousMarker(SetuptoolsDeprecationWarning):
+    _SUMMARY = "Ambiguous requirement marker."
+    _DETAILS = """
+    One of the parsed requirements in `{field}` looks like a valid environment marker:
+
+        {req!r}
+
+    Please make sure that the configuration file is correct.
+    You can use dangling lines to avoid this problem.
+    """
+    _SEE_DOCS = "userguide/declarative_config.html#opt-2"
+    # TODO: should we include due_date here? Initially introduced in 6 Aug 2022.
+    # Does this make sense with latest version of packaging?
+
+    @classmethod
+    def message(cls, **kw):
+        docs = f"https://setuptools.pypa.io/en/latest/{cls._SEE_DOCS}"
+        return cls._format(cls._SUMMARY, cls._DETAILS, see_url=docs, format_args=kw)
+
+
+class _DeprecatedConfig(SetuptoolsDeprecationWarning):
+    _SEE_DOCS = "userguide/declarative_config.html"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/config/setuptools.schema.json b/backend/venv/lib/python3.12/site-packages/setuptools/config/setuptools.schema.json
new file mode 100644
index 00000000..ec887b35
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/config/setuptools.schema.json
@@ -0,0 +1,433 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+
+  "$id": "https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html",
+  "title": "``tool.setuptools`` table",
+  "$$description": [
+    "``setuptools``-specific configurations that can be set by users that require",
+    "customization.",
+    "These configurations are completely optional and probably can be skipped when",
+    "creating simple packages. They are equivalent to some of the `Keywords",
+    "<https://setuptools.pypa.io/en/latest/references/keywords.html>`_",
+    "used by the ``setup.py`` file, and can be set via the ``tool.setuptools`` table.",
+    "It considers only ``setuptools`` `parameters",
+    "<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#setuptools-specific-configuration>`_",
+    "that are not covered by :pep:`621`; and intentionally excludes ``dependency_links``",
+    "and ``setup_requires`` (incompatible with modern workflows/standards)."
+  ],
+
+  "type": "object",
+  "additionalProperties": false,
+  "properties": {
+    "platforms": {
+      "type": "array",
+      "items": {"type": "string"}
+    },
+    "provides": {
+      "$$description": [
+        "Package and virtual package names contained within this package",
+        "**(not supported by pip)**"
+      ],
+      "type": "array",
+      "items": {"type": "string", "format": "pep508-identifier"}
+    },
+    "obsoletes": {
+      "$$description": [
+        "Packages which this package renders obsolete",
+        "**(not supported by pip)**"
+      ],
+      "type": "array",
+      "items": {"type": "string", "format": "pep508-identifier"}
+    },
+    "zip-safe": {
+      "$$description": [
+        "Whether the project can be safely installed and run from a zip file.",
+        "**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and",
+        "``setup.py install`` in the context of ``eggs`` (**DEPRECATED**)."
+      ],
+      "type": "boolean"
+    },
+    "script-files": {
+      "$$description": [
+        "Legacy way of defining scripts (entry-points are preferred).",
+        "Equivalent to the ``script`` keyword in ``setup.py``",
+        "(it was renamed to avoid confusion with entry-point based ``project.scripts``",
+        "defined in :pep:`621`).",
+        "**DISCOURAGED**: generic script wrappers are tricky and may not work properly.",
+        "Whenever possible, please use ``project.scripts`` instead."
+      ],
+      "type": "array",
+      "items": {"type": "string"},
+      "$comment": "TODO: is this field deprecated/should be removed?"
+    },
+    "eager-resources": {
+      "$$description": [
+        "Resources that should be extracted together, if any of them is needed,",
+        "or if any C extensions included in the project are imported.",
+        "**OBSOLETE**: only relevant for ``pkg_resources``, ``easy_install`` and",
+        "``setup.py install`` in the context of ``eggs`` (**DEPRECATED**)."
+      ],
+      "type": "array",
+      "items": {"type": "string"}
+    },
+    "packages": {
+      "$$description": [
+        "Packages that should be included in the distribution.",
+        "It can be given either as a list of package identifiers",
+        "or as a ``dict``-like structure with a single key ``find``",
+        "which corresponds to a dynamic call to",
+        "``setuptools.config.expand.find_packages`` function.",
+        "The ``find`` key is associated with a nested ``dict``-like structure that can",
+        "contain ``where``, ``include``, ``exclude`` and ``namespaces`` keys,",
+        "mimicking the keyword arguments of the associated function."
+      ],
+      "oneOf": [
+        {
+          "title": "Array of Python package identifiers",
+          "type": "array",
+          "items": {"$ref": "#/definitions/package-name"}
+        },
+        {"$ref": "#/definitions/find-directive"}
+      ]
+    },
+    "package-dir": {
+      "$$description": [
+        ":class:`dict`-like structure mapping from package names to directories where their",
+        "code can be found.",
+        "The empty string (as key) means that all packages are contained inside",
+        "the given directory will be included in the distribution."
+      ],
+      "type": "object",
+      "additionalProperties": false,
+      "propertyNames": {
+        "anyOf": [{"const": ""}, {"$ref": "#/definitions/package-name"}]
+      },
+      "patternProperties": {
+        "^.*$": {"type": "string" }
+      }
+    },
+    "package-data": {
+      "$$description": [
+        "Mapping from package names to lists of glob patterns.",
+        "Usually this option is not needed when using ``include-package-data = true``",
+        "For more information on how to include data files, check ``setuptools`` `docs",
+        "<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_."
+      ],
+      "type": "object",
+      "additionalProperties": false,
+      "propertyNames": {
+        "anyOf": [{"type": "string", "format": "python-module-name"}, {"const": "*"}]
+      },
+      "patternProperties": {
+        "^.*$": {"type": "array", "items": {"type": "string"}}
+      }
+    },
+    "include-package-data": {
+      "$$description": [
+        "Automatically include any data files inside the package directories",
+        "that are specified by ``MANIFEST.in``",
+        "For more information on how to include data files, check ``setuptools`` `docs",
+        "<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_."
+      ],
+      "type": "boolean"
+    },
+    "exclude-package-data": {
+      "$$description": [
+        "Mapping from package names to lists of glob patterns that should be excluded",
+        "For more information on how to include data files, check ``setuptools`` `docs",
+        "<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_."
+      ],
+      "type": "object",
+      "additionalProperties": false,
+      "propertyNames": {
+        "anyOf": [{"type": "string", "format": "python-module-name"}, {"const": "*"}]
+      },
+      "patternProperties": {
+          "^.*$": {"type": "array", "items": {"type": "string"}}
+      }
+    },
+    "namespace-packages": {
+      "type": "array",
+      "items": {"type": "string", "format": "python-module-name-relaxed"},
+      "$comment": "https://setuptools.pypa.io/en/latest/userguide/package_discovery.html",
+      "description": "**DEPRECATED**: use implicit namespaces instead (:pep:`420`)."
+    },
+    "py-modules": {
+      "description": "Modules that setuptools will manipulate",
+      "type": "array",
+      "items": {"type": "string", "format": "python-module-name-relaxed"},
+      "$comment": "TODO: clarify the relationship with ``packages``"
+    },
+    "ext-modules": {
+      "description": "Extension modules to be compiled by setuptools",
+      "type": "array",
+      "items": {"$ref": "#/definitions/ext-module"}
+    },
+    "data-files": {
+      "$$description": [
+        "``dict``-like structure where each key represents a directory and",
+        "the value is a list of glob patterns that should be installed in them.",
+        "**DISCOURAGED**: please notice this might not work as expected with wheels.",
+        "Whenever possible, consider using data files inside the package directories",
+        "(or create a new namespace package that only contains data files).",
+        "See `data files support",
+        "<https://setuptools.pypa.io/en/latest/userguide/datafiles.html>`_."
+      ],
+      "type": "object",
+      "patternProperties": {
+          "^.*$": {"type": "array", "items": {"type": "string"}}
+      }
+    },
+    "cmdclass": {
+      "$$description": [
+        "Mapping of distutils-style command names to ``setuptools.Command`` subclasses",
+        "which in turn should be represented by strings with a qualified class name",
+        "(i.e., \"dotted\" form with module), e.g.::\n\n",
+        "    cmdclass = {mycmd = \"pkg.subpkg.module.CommandClass\"}\n\n",
+        "The command class should be a directly defined at the top-level of the",
+        "containing module (no class nesting)."
+      ],
+      "type": "object",
+      "patternProperties": {
+          "^.*$": {"type": "string", "format": "python-qualified-identifier"}
+      }
+    },
+    "license-files": {
+      "type": "array",
+      "items": {"type": "string"},
+      "$$description": [
+        "**PROVISIONAL**: list of glob patterns for all license files being distributed.",
+        "(likely to become standard with :pep:`639`).",
+        "By default: ``['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']``"
+      ],
+      "$comment": "TODO: revise if PEP 639 is accepted. Probably ``project.license-files``?"
+    },
+    "dynamic": {
+      "type": "object",
+      "description": "Instructions for loading :pep:`621`-related metadata dynamically",
+      "additionalProperties": false,
+      "properties": {
+        "version": {
+          "$$description": [
+            "A version dynamically loaded via either the ``attr:`` or ``file:``",
+            "directives. Please make sure the given file or attribute respects :pep:`440`.",
+            "Also ensure to set ``project.dynamic`` accordingly."
+          ],
+          "oneOf": [
+            {"$ref": "#/definitions/attr-directive"},
+            {"$ref": "#/definitions/file-directive"}
+          ]
+        },
+        "classifiers": {"$ref": "#/definitions/file-directive"},
+        "description": {"$ref": "#/definitions/file-directive"},
+        "entry-points": {"$ref": "#/definitions/file-directive"},
+        "dependencies": {"$ref": "#/definitions/file-directive-for-dependencies"},
+        "optional-dependencies": {
+          "type": "object",
+          "propertyNames": {"type": "string", "format": "pep508-identifier"},
+          "additionalProperties": false,
+          "patternProperties": {
+            ".+": {"$ref": "#/definitions/file-directive-for-dependencies"}
+          }
+        },
+        "readme": {
+          "type": "object",
+          "anyOf": [
+            {"$ref": "#/definitions/file-directive"},
+            {
+              "type": "object",
+              "properties": {
+                "content-type": {"type": "string"},
+                "file": { "$ref": "#/definitions/file-directive/properties/file" }
+              },
+              "additionalProperties": false}
+          ],
+          "required": ["file"]
+        }
+      }
+    }
+  },
+
+  "definitions": {
+    "package-name": {
+      "$id": "#/definitions/package-name",
+      "title": "Valid package name",
+      "description": "Valid package name (importable or :pep:`561`).",
+      "type": "string",
+      "anyOf": [
+        {"type": "string", "format": "python-module-name-relaxed"},
+        {"type": "string", "format": "pep561-stub-name"}
+      ]
+    },
+    "ext-module": {
+      "$id": "#/definitions/ext-module",
+      "title": "Extension module",
+      "description": "Parameters to construct a :class:`setuptools.Extension` object",
+      "type": "object",
+      "required": ["name", "sources"],
+      "additionalProperties": false,
+      "properties": {
+        "name": {
+          "type": "string",
+          "format": "python-module-name-relaxed"
+        },
+        "sources": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "include-dirs":{
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "define-macros": {
+          "type": "array",
+          "items": {
+            "type": "array",
+            "items": [
+              {"description": "macro name", "type": "string"},
+              {"description": "macro value", "oneOf": [{"type": "string"}, {"type": "null"}]}
+            ],
+            "additionalItems": false
+          }
+        },
+        "undef-macros": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "library-dirs": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "libraries": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "runtime-library-dirs": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "extra-objects": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "extra-compile-args": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "extra-link-args": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "export-symbols": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "swig-opts": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "depends": {
+          "type": "array",
+          "items": {"type": "string"}
+        },
+        "language": {"type": "string"},
+        "optional": {"type": "boolean"},
+        "py-limited-api": {"type": "boolean"}
+      }
+    },
+    "file-directive": {
+      "$id": "#/definitions/file-directive",
+      "title": "'file:' directive",
+      "description":
+        "Value is read from a file (or list of files and then concatenated)",
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "file": {
+          "oneOf": [
+            {"type": "string"},
+            {"type": "array", "items": {"type": "string"}}
+          ]
+        }
+      },
+      "required": ["file"]
+    },
+    "file-directive-for-dependencies": {
+      "title": "'file:' directive for dependencies",
+      "allOf": [
+        {
+          "$$description": [
+            "**BETA**: subset of the ``requirements.txt`` format",
+            "without ``pip`` flags and options",
+            "(one :pep:`508`-compliant string per line,",
+            "lines that are blank or start with ``#`` are excluded).",
+            "See `dynamic metadata",
+            "<https://setuptools.pypa.io/en/latest/userguide/pyproject_config.html#dynamic-metadata>`_."
+          ]
+        },
+        {"$ref": "#/definitions/file-directive"}
+      ]
+    },
+    "attr-directive": {
+      "title": "'attr:' directive",
+      "$id": "#/definitions/attr-directive",
+      "$$description": [
+        "Value is read from a module attribute. Supports callables and iterables;",
+        "unsupported types are cast via ``str()``"
+      ],
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "attr": {"type": "string", "format": "python-qualified-identifier"}
+      },
+      "required": ["attr"]
+    },
+    "find-directive": {
+      "$id": "#/definitions/find-directive",
+      "title": "'find:' directive",
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "find": {
+          "type": "object",
+          "$$description": [
+            "Dynamic `package discovery",
+            "<https://setuptools.pypa.io/en/latest/userguide/package_discovery.html>`_."
+          ],
+          "additionalProperties": false,
+          "properties": {
+            "where": {
+              "description":
+                "Directories to be searched for packages (Unix-style relative path)",
+              "type": "array",
+              "items": {"type": "string"}
+            },
+            "exclude": {
+              "type": "array",
+              "$$description": [
+                "Exclude packages that match the values listed in this field.",
+                "Can container shell-style wildcards (e.g. ``'pkg.*'``)"
+              ],
+              "items": {"type": "string"}
+            },
+            "include": {
+              "type": "array",
+              "$$description": [
+                "Restrict the found packages to just the ones listed in this field.",
+                "Can container shell-style wildcards (e.g. ``'pkg.*'``)"
+              ],
+              "items": {"type": "string"}
+            },
+            "namespaces": {
+              "type": "boolean",
+              "$$description": [
+                "When ``True``, directories without a ``__init__.py`` file will also",
+                "be scanned for :pep:`420`-style implicit namespaces"
+              ]
+            }
+          }
+        }
+      }
+    }
+  }
+}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/depends.py b/backend/venv/lib/python3.12/site-packages/setuptools/depends.py
new file mode 100644
index 00000000..e5223b79
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/depends.py
@@ -0,0 +1,185 @@
+from __future__ import annotations
+
+import contextlib
+import dis
+import marshal
+import sys
+from types import CodeType
+from typing import Any, Literal, TypeVar
+
+from packaging.version import Version
+
+from . import _imp
+from ._imp import PY_COMPILED, PY_FROZEN, PY_SOURCE, find_module
+
+_T = TypeVar("_T")
+
+__all__ = ['Require', 'find_module']
+
+
+class Require:
+    """A prerequisite to building or installing a distribution"""
+
+    def __init__(
+        self,
+        name,
+        requested_version,
+        module,
+        homepage: str = '',
+        attribute=None,
+        format=None,
+    ) -> None:
+        if format is None and requested_version is not None:
+            format = Version
+
+        if format is not None:
+            requested_version = format(requested_version)
+            if attribute is None:
+                attribute = '__version__'
+
+        self.__dict__.update(locals())
+        del self.self
+
+    def full_name(self):
+        """Return full package/distribution name, w/version"""
+        if self.requested_version is not None:
+            return f'{self.name}-{self.requested_version}'
+        return self.name
+
+    def version_ok(self, version):
+        """Is 'version' sufficiently up-to-date?"""
+        return (
+            self.attribute is None
+            or self.format is None
+            or str(version) != "unknown"
+            and self.format(version) >= self.requested_version
+        )
+
+    def get_version(
+        self, paths=None, default: _T | Literal["unknown"] = "unknown"
+    ) -> _T | Literal["unknown"] | None | Any:
+        """Get version number of installed module, 'None', or 'default'
+
+        Search 'paths' for module.  If not found, return 'None'.  If found,
+        return the extracted version attribute, or 'default' if no version
+        attribute was specified, or the value cannot be determined without
+        importing the module.  The version is formatted according to the
+        requirement's version format (if any), unless it is 'None' or the
+        supplied 'default'.
+        """
+
+        if self.attribute is None:
+            try:
+                f, _p, _i = find_module(self.module, paths)
+            except ImportError:
+                return None
+            if f:
+                f.close()
+            return default
+
+        v = get_module_constant(self.module, self.attribute, default, paths)
+
+        if v is not None and v is not default and self.format is not None:
+            return self.format(v)
+
+        return v
+
+    def is_present(self, paths=None):
+        """Return true if dependency is present on 'paths'"""
+        return self.get_version(paths) is not None
+
+    def is_current(self, paths=None):
+        """Return true if dependency is present and up-to-date on 'paths'"""
+        version = self.get_version(paths)
+        if version is None:
+            return False
+        return self.version_ok(str(version))
+
+
+def maybe_close(f):
+    @contextlib.contextmanager
+    def empty():
+        yield
+        return
+
+    if not f:
+        return empty()
+
+    return contextlib.closing(f)
+
+
+# Some objects are not available on some platforms.
+# XXX it'd be better to test assertions about bytecode instead.
+if not sys.platform.startswith('java') and sys.platform != 'cli':
+
+    def get_module_constant(
+        module, symbol, default: _T | int = -1, paths=None
+    ) -> _T | int | None | Any:
+        """Find 'module' by searching 'paths', and extract 'symbol'
+
+        Return 'None' if 'module' does not exist on 'paths', or it does not define
+        'symbol'.  If the module defines 'symbol' as a constant, return the
+        constant.  Otherwise, return 'default'."""
+
+        try:
+            f, path, (_suffix, _mode, kind) = info = find_module(module, paths)
+        except ImportError:
+            # Module doesn't exist
+            return None
+
+        with maybe_close(f):
+            if kind == PY_COMPILED:
+                f.read(8)  # skip magic & date
+                code = marshal.load(f)
+            elif kind == PY_FROZEN:
+                code = _imp.get_frozen_object(module, paths)
+            elif kind == PY_SOURCE:
+                code = compile(f.read(), path, 'exec')
+            else:
+                # Not something we can parse; we'll have to import it.  :(
+                imported = _imp.get_module(module, paths, info)
+                return getattr(imported, symbol, None)
+
+        return extract_constant(code, symbol, default)
+
+    def extract_constant(
+        code: CodeType, symbol: str, default: _T | int = -1
+    ) -> _T | int | None | Any:
+        """Extract the constant value of 'symbol' from 'code'
+
+        If the name 'symbol' is bound to a constant value by the Python code
+        object 'code', return that value.  If 'symbol' is bound to an expression,
+        return 'default'.  Otherwise, return 'None'.
+
+        Return value is based on the first assignment to 'symbol'.  'symbol' must
+        be a global, or at least a non-"fast" local in the code block.  That is,
+        only 'STORE_NAME' and 'STORE_GLOBAL' opcodes are checked, and 'symbol'
+        must be present in 'code.co_names'.
+        """
+        if symbol not in code.co_names:
+            # name's not there, can't possibly be an assignment
+            return None
+
+        name_idx = list(code.co_names).index(symbol)
+
+        STORE_NAME = dis.opmap['STORE_NAME']
+        STORE_GLOBAL = dis.opmap['STORE_GLOBAL']
+        LOAD_CONST = dis.opmap['LOAD_CONST']
+
+        const = default
+
+        for byte_code in dis.Bytecode(code):
+            op = byte_code.opcode
+            arg = byte_code.arg
+
+            if op == LOAD_CONST:
+                assert arg is not None
+                const = code.co_consts[arg]
+            elif arg == name_idx and (op == STORE_NAME or op == STORE_GLOBAL):
+                return const
+            else:
+                const = default
+
+        return None
+
+    __all__ += ['get_module_constant', 'extract_constant']
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/discovery.py b/backend/venv/lib/python3.12/site-packages/setuptools/discovery.py
new file mode 100644
index 00000000..c8883991
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/discovery.py
@@ -0,0 +1,614 @@
+"""Automatic discovery of Python modules and packages (for inclusion in the
+distribution) and other config values.
+
+For the purposes of this module, the following nomenclature is used:
+
+- "src-layout": a directory representing a Python project that contains a "src"
+  folder. Everything under the "src" folder is meant to be included in the
+  distribution when packaging the project. Example::
+
+    .
+    ├── tox.ini
+    ├── pyproject.toml
+    └── src/
+        └── mypkg/
+            ├── __init__.py
+            ├── mymodule.py
+            └── my_data_file.txt
+
+- "flat-layout": a Python project that does not use "src-layout" but instead
+  have a directory under the project root for each package::
+
+    .
+    ├── tox.ini
+    ├── pyproject.toml
+    └── mypkg/
+        ├── __init__.py
+        ├── mymodule.py
+        └── my_data_file.txt
+
+- "single-module": a project that contains a single Python script direct under
+  the project root (no directory used)::
+
+    .
+    ├── tox.ini
+    ├── pyproject.toml
+    └── mymodule.py
+
+"""
+
+from __future__ import annotations
+
+import itertools
+import os
+from collections.abc import Iterable, Iterator, Mapping
+from fnmatch import fnmatchcase
+from glob import glob
+from pathlib import Path
+from typing import TYPE_CHECKING, ClassVar
+
+import _distutils_hack.override  # noqa: F401
+
+from ._path import StrPath
+
+from distutils import log
+from distutils.util import convert_path
+
+if TYPE_CHECKING:
+    from setuptools import Distribution
+
+chain_iter = itertools.chain.from_iterable
+
+
+def _valid_name(path: StrPath) -> bool:
+    # Ignore invalid names that cannot be imported directly
+    return os.path.basename(path).isidentifier()
+
+
+class _Filter:
+    """
+    Given a list of patterns, create a callable that will be true only if
+    the input matches at least one of the patterns.
+    """
+
+    def __init__(self, *patterns: str) -> None:
+        self._patterns = dict.fromkeys(patterns)
+
+    def __call__(self, item: str) -> bool:
+        return any(fnmatchcase(item, pat) for pat in self._patterns)
+
+    def __contains__(self, item: str) -> bool:
+        return item in self._patterns
+
+
+class _Finder:
+    """Base class that exposes functionality for module/package finders"""
+
+    ALWAYS_EXCLUDE: ClassVar[tuple[str, ...]] = ()
+    DEFAULT_EXCLUDE: ClassVar[tuple[str, ...]] = ()
+
+    @classmethod
+    def find(
+        cls,
+        where: StrPath = '.',
+        exclude: Iterable[str] = (),
+        include: Iterable[str] = ('*',),
+    ) -> list[str]:
+        """Return a list of all Python items (packages or modules, depending on
+        the finder implementation) found within directory 'where'.
+
+        'where' is the root directory which will be searched.
+        It should be supplied as a "cross-platform" (i.e. URL-style) path;
+        it will be converted to the appropriate local path syntax.
+
+        'exclude' is a sequence of names to exclude; '*' can be used
+        as a wildcard in the names.
+        When finding packages, 'foo.*' will exclude all subpackages of 'foo'
+        (but not 'foo' itself).
+
+        'include' is a sequence of names to include.
+        If it's specified, only the named items will be included.
+        If it's not specified, all found items will be included.
+        'include' can contain shell style wildcard patterns just like
+        'exclude'.
+        """
+
+        exclude = exclude or cls.DEFAULT_EXCLUDE
+        return list(
+            cls._find_iter(
+                convert_path(str(where)),
+                _Filter(*cls.ALWAYS_EXCLUDE, *exclude),
+                _Filter(*include),
+            )
+        )
+
+    @classmethod
+    def _find_iter(
+        cls, where: StrPath, exclude: _Filter, include: _Filter
+    ) -> Iterator[str]:
+        raise NotImplementedError
+
+
+class PackageFinder(_Finder):
+    """
+    Generate a list of all Python packages found within a directory
+    """
+
+    ALWAYS_EXCLUDE = ("ez_setup", "*__pycache__")
+
+    @classmethod
+    def _find_iter(
+        cls, where: StrPath, exclude: _Filter, include: _Filter
+    ) -> Iterator[str]:
+        """
+        All the packages found in 'where' that pass the 'include' filter, but
+        not the 'exclude' filter.
+        """
+        for root, dirs, files in os.walk(str(where), followlinks=True):
+            # Copy dirs to iterate over it, then empty dirs.
+            all_dirs = dirs[:]
+            dirs[:] = []
+
+            for dir in all_dirs:
+                full_path = os.path.join(root, dir)
+                rel_path = os.path.relpath(full_path, where)
+                package = rel_path.replace(os.path.sep, '.')
+
+                # Skip directory trees that are not valid packages
+                if '.' in dir or not cls._looks_like_package(full_path, package):
+                    continue
+
+                # Should this package be included?
+                if include(package) and not exclude(package):
+                    yield package
+
+                # Early pruning if there is nothing else to be scanned
+                if f"{package}*" in exclude or f"{package}.*" in exclude:
+                    continue
+
+                # Keep searching subdirectories, as there may be more packages
+                # down there, even if the parent was excluded.
+                dirs.append(dir)
+
+    @staticmethod
+    def _looks_like_package(path: StrPath, _package_name: str) -> bool:
+        """Does a directory look like a package?"""
+        return os.path.isfile(os.path.join(path, '__init__.py'))
+
+
+class PEP420PackageFinder(PackageFinder):
+    @staticmethod
+    def _looks_like_package(_path: StrPath, _package_name: str) -> bool:
+        return True
+
+
+class ModuleFinder(_Finder):
+    """Find isolated Python modules.
+    This function will **not** recurse subdirectories.
+    """
+
+    @classmethod
+    def _find_iter(
+        cls, where: StrPath, exclude: _Filter, include: _Filter
+    ) -> Iterator[str]:
+        for file in glob(os.path.join(where, "*.py")):
+            module, _ext = os.path.splitext(os.path.basename(file))
+
+            if not cls._looks_like_module(module):
+                continue
+
+            if include(module) and not exclude(module):
+                yield module
+
+    _looks_like_module = staticmethod(_valid_name)
+
+
+# We have to be extra careful in the case of flat layout to not include files
+# and directories not meant for distribution (e.g. tool-related)
+
+
+class FlatLayoutPackageFinder(PEP420PackageFinder):
+    _EXCLUDE = (
+        "ci",
+        "bin",
+        "debian",
+        "doc",
+        "docs",
+        "documentation",
+        "manpages",
+        "news",
+        "newsfragments",
+        "changelog",
+        "test",
+        "tests",
+        "unit_test",
+        "unit_tests",
+        "example",
+        "examples",
+        "scripts",
+        "tools",
+        "util",
+        "utils",
+        "python",
+        "build",
+        "dist",
+        "venv",
+        "env",
+        "requirements",
+        # ---- Task runners / Build tools ----
+        "tasks",  # invoke
+        "fabfile",  # fabric
+        "site_scons",  # SCons
+        # ---- Other tools ----
+        "benchmark",
+        "benchmarks",
+        "exercise",
+        "exercises",
+        "htmlcov",  # Coverage.py
+        # ---- Hidden directories/Private packages ----
+        "[._]*",
+    )
+
+    DEFAULT_EXCLUDE = tuple(chain_iter((p, f"{p}.*") for p in _EXCLUDE))
+    """Reserved package names"""
+
+    @staticmethod
+    def _looks_like_package(_path: StrPath, package_name: str) -> bool:
+        names = package_name.split('.')
+        # Consider PEP 561
+        root_pkg_is_valid = names[0].isidentifier() or names[0].endswith("-stubs")
+        return root_pkg_is_valid and all(name.isidentifier() for name in names[1:])
+
+
+class FlatLayoutModuleFinder(ModuleFinder):
+    DEFAULT_EXCLUDE = (
+        "setup",
+        "conftest",
+        "test",
+        "tests",
+        "example",
+        "examples",
+        "build",
+        # ---- Task runners ----
+        "toxfile",
+        "noxfile",
+        "pavement",
+        "dodo",
+        "tasks",
+        "fabfile",
+        # ---- Other tools ----
+        "[Ss][Cc]onstruct",  # SCons
+        "conanfile",  # Connan: C/C++ build tool
+        "manage",  # Django
+        "benchmark",
+        "benchmarks",
+        "exercise",
+        "exercises",
+        # ---- Hidden files/Private modules ----
+        "[._]*",
+    )
+    """Reserved top-level module names"""
+
+
+def _find_packages_within(root_pkg: str, pkg_dir: StrPath) -> list[str]:
+    nested = PEP420PackageFinder.find(pkg_dir)
+    return [root_pkg] + [".".join((root_pkg, n)) for n in nested]
+
+
+class ConfigDiscovery:
+    """Fill-in metadata and options that can be automatically derived
+    (from other metadata/options, the file system or conventions)
+    """
+
+    def __init__(self, distribution: Distribution) -> None:
+        self.dist = distribution
+        self._called = False
+        self._disabled = False
+        self._skip_ext_modules = False
+
+    def _disable(self):
+        """Internal API to disable automatic discovery"""
+        self._disabled = True
+
+    def _ignore_ext_modules(self):
+        """Internal API to disregard ext_modules.
+
+        Normally auto-discovery would not be triggered if ``ext_modules`` are set
+        (this is done for backward compatibility with existing packages relying on
+        ``setup.py`` or ``setup.cfg``). However, ``setuptools`` can call this function
+        to ignore given ``ext_modules`` and proceed with the auto-discovery if
+        ``packages`` and ``py_modules`` are not given (e.g. when using pyproject.toml
+        metadata).
+        """
+        self._skip_ext_modules = True
+
+    @property
+    def _root_dir(self) -> StrPath:
+        # The best is to wait until `src_root` is set in dist, before using _root_dir.
+        return self.dist.src_root or os.curdir
+
+    @property
+    def _package_dir(self) -> dict[str, str]:
+        if self.dist.package_dir is None:
+            return {}
+        return self.dist.package_dir
+
+    def __call__(
+        self, force: bool = False, name: bool = True, ignore_ext_modules: bool = False
+    ):
+        """Automatically discover missing configuration fields
+        and modifies the given ``distribution`` object in-place.
+
+        Note that by default this will only have an effect the first time the
+        ``ConfigDiscovery`` object is called.
+
+        To repeatedly invoke automatic discovery (e.g. when the project
+        directory changes), please use ``force=True`` (or create a new
+        ``ConfigDiscovery`` instance).
+        """
+        if force is False and (self._called or self._disabled):
+            # Avoid overhead of multiple calls
+            return
+
+        self._analyse_package_layout(ignore_ext_modules)
+        if name:
+            self.analyse_name()  # depends on ``packages`` and ``py_modules``
+
+        self._called = True
+
+    def _explicitly_specified(self, ignore_ext_modules: bool) -> bool:
+        """``True`` if the user has specified some form of package/module listing"""
+        ignore_ext_modules = ignore_ext_modules or self._skip_ext_modules
+        ext_modules = not (self.dist.ext_modules is None or ignore_ext_modules)
+        return (
+            self.dist.packages is not None
+            or self.dist.py_modules is not None
+            or ext_modules
+            or hasattr(self.dist, "configuration")
+            and self.dist.configuration
+            # ^ Some projects use numpy.distutils.misc_util.Configuration
+        )
+
+    def _analyse_package_layout(self, ignore_ext_modules: bool) -> bool:
+        if self._explicitly_specified(ignore_ext_modules):
+            # For backward compatibility, just try to find modules/packages
+            # when nothing is given
+            return True
+
+        log.debug(
+            "No `packages` or `py_modules` configuration, performing "
+            "automatic discovery."
+        )
+
+        return (
+            self._analyse_explicit_layout()
+            or self._analyse_src_layout()
+            # flat-layout is the trickiest for discovery so it should be last
+            or self._analyse_flat_layout()
+        )
+
+    def _analyse_explicit_layout(self) -> bool:
+        """The user can explicitly give a package layout via ``package_dir``"""
+        package_dir = self._package_dir.copy()  # don't modify directly
+        package_dir.pop("", None)  # This falls under the "src-layout" umbrella
+        root_dir = self._root_dir
+
+        if not package_dir:
+            return False
+
+        log.debug(f"`explicit-layout` detected -- analysing {package_dir}")
+        pkgs = chain_iter(
+            _find_packages_within(pkg, os.path.join(root_dir, parent_dir))
+            for pkg, parent_dir in package_dir.items()
+        )
+        self.dist.packages = list(pkgs)
+        log.debug(f"discovered packages -- {self.dist.packages}")
+        return True
+
+    def _analyse_src_layout(self) -> bool:
+        """Try to find all packages or modules under the ``src`` directory
+        (or anything pointed by ``package_dir[""]``).
+
+        The "src-layout" is relatively safe for automatic discovery.
+        We assume that everything within is meant to be included in the
+        distribution.
+
+        If ``package_dir[""]`` is not given, but the ``src`` directory exists,
+        this function will set ``package_dir[""] = "src"``.
+        """
+        package_dir = self._package_dir
+        src_dir = os.path.join(self._root_dir, package_dir.get("", "src"))
+        if not os.path.isdir(src_dir):
+            return False
+
+        log.debug(f"`src-layout` detected -- analysing {src_dir}")
+        package_dir.setdefault("", os.path.basename(src_dir))
+        self.dist.package_dir = package_dir  # persist eventual modifications
+        self.dist.packages = PEP420PackageFinder.find(src_dir)
+        self.dist.py_modules = ModuleFinder.find(src_dir)
+        log.debug(f"discovered packages -- {self.dist.packages}")
+        log.debug(f"discovered py_modules -- {self.dist.py_modules}")
+        return True
+
+    def _analyse_flat_layout(self) -> bool:
+        """Try to find all packages and modules under the project root.
+
+        Since the ``flat-layout`` is more dangerous in terms of accidentally including
+        extra files/directories, this function is more conservative and will raise an
+        error if multiple packages or modules are found.
+
+        This assumes that multi-package dists are uncommon and refuse to support that
+        use case in order to be able to prevent unintended errors.
+        """
+        log.debug(f"`flat-layout` detected -- analysing {self._root_dir}")
+        return self._analyse_flat_packages() or self._analyse_flat_modules()
+
+    def _analyse_flat_packages(self) -> bool:
+        self.dist.packages = FlatLayoutPackageFinder.find(self._root_dir)
+        top_level = remove_nested_packages(remove_stubs(self.dist.packages))
+        log.debug(f"discovered packages -- {self.dist.packages}")
+        self._ensure_no_accidental_inclusion(top_level, "packages")
+        return bool(top_level)
+
+    def _analyse_flat_modules(self) -> bool:
+        self.dist.py_modules = FlatLayoutModuleFinder.find(self._root_dir)
+        log.debug(f"discovered py_modules -- {self.dist.py_modules}")
+        self._ensure_no_accidental_inclusion(self.dist.py_modules, "modules")
+        return bool(self.dist.py_modules)
+
+    def _ensure_no_accidental_inclusion(self, detected: list[str], kind: str):
+        if len(detected) > 1:
+            from inspect import cleandoc
+
+            from setuptools.errors import PackageDiscoveryError
+
+            msg = f"""Multiple top-level {kind} discovered in a flat-layout: {detected}.
+
+            To avoid accidental inclusion of unwanted files or directories,
+            setuptools will not proceed with this build.
+
+            If you are trying to create a single distribution with multiple {kind}
+            on purpose, you should not rely on automatic discovery.
+            Instead, consider the following options:
+
+            1. set up custom discovery (`find` directive with `include` or `exclude`)
+            2. use a `src-layout`
+            3. explicitly set `py_modules` or `packages` with a list of names
+
+            To find more information, look for "package discovery" on setuptools docs.
+            """
+            raise PackageDiscoveryError(cleandoc(msg))
+
+    def analyse_name(self) -> None:
+        """The packages/modules are the essential contribution of the author.
+        Therefore the name of the distribution can be derived from them.
+        """
+        if self.dist.metadata.name or self.dist.name:
+            # get_name() is not reliable (can return "UNKNOWN")
+            return
+
+        log.debug("No `name` configuration, performing automatic discovery")
+
+        name = (
+            self._find_name_single_package_or_module()
+            or self._find_name_from_packages()
+        )
+        if name:
+            self.dist.metadata.name = name
+
+    def _find_name_single_package_or_module(self) -> str | None:
+        """Exactly one module or package"""
+        for field in ('packages', 'py_modules'):
+            items = getattr(self.dist, field, None) or []
+            if items and len(items) == 1:
+                log.debug(f"Single module/package detected, name: {items[0]}")
+                return items[0]
+
+        return None
+
+    def _find_name_from_packages(self) -> str | None:
+        """Try to find the root package that is not a PEP 420 namespace"""
+        if not self.dist.packages:
+            return None
+
+        packages = remove_stubs(sorted(self.dist.packages, key=len))
+        package_dir = self.dist.package_dir or {}
+
+        parent_pkg = find_parent_package(packages, package_dir, self._root_dir)
+        if parent_pkg:
+            log.debug(f"Common parent package detected, name: {parent_pkg}")
+            return parent_pkg
+
+        log.warn("No parent package detected, impossible to derive `name`")
+        return None
+
+
+def remove_nested_packages(packages: list[str]) -> list[str]:
+    """Remove nested packages from a list of packages.
+
+    >>> remove_nested_packages(["a", "a.b1", "a.b2", "a.b1.c1"])
+    ['a']
+    >>> remove_nested_packages(["a", "b", "c.d", "c.d.e.f", "g.h", "a.a1"])
+    ['a', 'b', 'c.d', 'g.h']
+    """
+    pkgs = sorted(packages, key=len)
+    top_level = pkgs[:]
+    size = len(pkgs)
+    for i, name in enumerate(reversed(pkgs)):
+        if any(name.startswith(f"{other}.") for other in top_level):
+            top_level.pop(size - i - 1)
+
+    return top_level
+
+
+def remove_stubs(packages: list[str]) -> list[str]:
+    """Remove type stubs (:pep:`561`) from a list of packages.
+
+    >>> remove_stubs(["a", "a.b", "a-stubs", "a-stubs.b.c", "b", "c-stubs"])
+    ['a', 'a.b', 'b']
+    """
+    return [pkg for pkg in packages if not pkg.split(".")[0].endswith("-stubs")]
+
+
+def find_parent_package(
+    packages: list[str], package_dir: Mapping[str, str], root_dir: StrPath
+) -> str | None:
+    """Find the parent package that is not a namespace."""
+    packages = sorted(packages, key=len)
+    common_ancestors = []
+    for i, name in enumerate(packages):
+        if not all(n.startswith(f"{name}.") for n in packages[i + 1 :]):
+            # Since packages are sorted by length, this condition is able
+            # to find a list of all common ancestors.
+            # When there is divergence (e.g. multiple root packages)
+            # the list will be empty
+            break
+        common_ancestors.append(name)
+
+    for name in common_ancestors:
+        pkg_path = find_package_path(name, package_dir, root_dir)
+        init = os.path.join(pkg_path, "__init__.py")
+        if os.path.isfile(init):
+            return name
+
+    return None
+
+
+def find_package_path(
+    name: str, package_dir: Mapping[str, str], root_dir: StrPath
+) -> str:
+    """Given a package name, return the path where it should be found on
+    disk, considering the ``package_dir`` option.
+
+    >>> path = find_package_path("my.pkg", {"": "root/is/nested"}, ".")
+    >>> path.replace(os.sep, "/")
+    './root/is/nested/my/pkg'
+
+    >>> path = find_package_path("my.pkg", {"my": "root/is/nested"}, ".")
+    >>> path.replace(os.sep, "/")
+    './root/is/nested/pkg'
+
+    >>> path = find_package_path("my.pkg", {"my.pkg": "root/is/nested"}, ".")
+    >>> path.replace(os.sep, "/")
+    './root/is/nested'
+
+    >>> path = find_package_path("other.pkg", {"my.pkg": "root/is/nested"}, ".")
+    >>> path.replace(os.sep, "/")
+    './other/pkg'
+    """
+    parts = name.split(".")
+    for i in range(len(parts), 0, -1):
+        # Look backwards, the most specific package_dir first
+        partial_name = ".".join(parts[:i])
+        if partial_name in package_dir:
+            parent = package_dir[partial_name]
+            return os.path.join(root_dir, parent, *parts[i:])
+
+    parent = package_dir.get("") or ""
+    return os.path.join(root_dir, *parent.split("/"), *parts)
+
+
+def construct_package_dir(packages: list[str], package_path: StrPath) -> dict[str, str]:
+    parent_pkgs = remove_nested_packages(packages)
+    prefix = Path(package_path).parts
+    return {pkg: "/".join([*prefix, *pkg.split(".")]) for pkg in parent_pkgs}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/dist.py b/backend/venv/lib/python3.12/site-packages/setuptools/dist.py
new file mode 100644
index 00000000..02496512
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/dist.py
@@ -0,0 +1,1004 @@
+from __future__ import annotations
+
+import io
+import itertools
+import numbers
+import os
+import re
+import sys
+from collections.abc import Iterable, MutableMapping, Sequence
+from glob import iglob
+from pathlib import Path
+from typing import TYPE_CHECKING, Any, Union
+
+from more_itertools import partition, unique_everseen
+from packaging.markers import InvalidMarker, Marker
+from packaging.specifiers import InvalidSpecifier, SpecifierSet
+from packaging.version import Version
+
+from . import (
+    _entry_points,
+    _reqs,
+    _static,
+    command as _,  # noqa: F401 # imported for side-effects
+)
+from ._importlib import metadata
+from ._path import StrPath
+from ._reqs import _StrOrIter
+from .config import pyprojecttoml, setupcfg
+from .discovery import ConfigDiscovery
+from .monkey import get_unpatched
+from .warnings import InformationOnly, SetuptoolsDeprecationWarning
+
+import distutils.cmd
+import distutils.command
+import distutils.core
+import distutils.dist
+import distutils.log
+from distutils.debug import DEBUG
+from distutils.errors import DistutilsOptionError, DistutilsSetupError
+from distutils.fancy_getopt import translate_longopt
+from distutils.util import strtobool
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+    from pkg_resources import Distribution as _pkg_resources_Distribution
+
+
+__all__ = ['Distribution']
+
+_sequence = tuple, list
+"""
+:meta private:
+
+Supported iterable types that are known to be:
+- ordered (which `set` isn't)
+- not match a str (which `Sequence[str]` does)
+- not imply a nested type (like `dict`)
+for use with `isinstance`.
+"""
+_Sequence: TypeAlias = Union[tuple[str, ...], list[str]]
+# This is how stringifying _Sequence would look in Python 3.10
+_sequence_type_repr = "tuple[str, ...] | list[str]"
+_OrderedStrSequence: TypeAlias = Union[str, dict[str, Any], Sequence[str]]
+"""
+:meta private:
+Avoid single-use iterable. Disallow sets.
+A poor approximation of an OrderedSequence (dict doesn't match a Sequence).
+"""
+
+
+def __getattr__(name: str) -> Any:  # pragma: no cover
+    if name == "sequence":
+        SetuptoolsDeprecationWarning.emit(
+            "`setuptools.dist.sequence` is an internal implementation detail.",
+            "Please define your own `sequence = tuple, list` instead.",
+            due_date=(2025, 8, 28),  # Originally added on 2024-08-27
+        )
+        return _sequence
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+
+def check_importable(dist, attr, value):
+    try:
+        ep = metadata.EntryPoint(value=value, name=None, group=None)
+        assert not ep.extras
+    except (TypeError, ValueError, AttributeError, AssertionError) as e:
+        raise DistutilsSetupError(
+            f"{attr!r} must be importable 'module:attrs' string (got {value!r})"
+        ) from e
+
+
+def assert_string_list(dist, attr: str, value: _Sequence) -> None:
+    """Verify that value is a string list"""
+    try:
+        # verify that value is a list or tuple to exclude unordered
+        # or single-use iterables
+        assert isinstance(value, _sequence)
+        # verify that elements of value are strings
+        assert ''.join(value) != value
+    except (TypeError, ValueError, AttributeError, AssertionError) as e:
+        raise DistutilsSetupError(
+            f"{attr!r} must be of type <{_sequence_type_repr}> (got {value!r})"
+        ) from e
+
+
+def check_nsp(dist, attr, value):
+    """Verify that namespace packages are valid"""
+    ns_packages = value
+    assert_string_list(dist, attr, ns_packages)
+    for nsp in ns_packages:
+        if not dist.has_contents_for(nsp):
+            raise DistutilsSetupError(
+                f"Distribution contains no modules or packages for namespace package {nsp!r}"
+            )
+        parent, _sep, _child = nsp.rpartition('.')
+        if parent and parent not in ns_packages:
+            distutils.log.warn(
+                "WARNING: %r is declared as a package namespace, but %r"
+                " is not: please correct this in setup.py",
+                nsp,
+                parent,
+            )
+        SetuptoolsDeprecationWarning.emit(
+            "The namespace_packages parameter is deprecated.",
+            "Please replace its usage with implicit namespaces (PEP 420).",
+            see_docs="references/keywords.html#keyword-namespace-packages",
+            # TODO: define due_date, it may break old packages that are no longer
+            # maintained (e.g. sphinxcontrib extensions) when installed from source.
+            # Warning officially introduced in May 2022, however the deprecation
+            # was mentioned much earlier in the docs (May 2020, see #2149).
+        )
+
+
+def check_extras(dist, attr, value):
+    """Verify that extras_require mapping is valid"""
+    try:
+        list(itertools.starmap(_check_extra, value.items()))
+    except (TypeError, ValueError, AttributeError) as e:
+        raise DistutilsSetupError(
+            "'extras_require' must be a dictionary whose values are "
+            "strings or lists of strings containing valid project/version "
+            "requirement specifiers."
+        ) from e
+
+
+def _check_extra(extra, reqs):
+    _name, _sep, marker = extra.partition(':')
+    try:
+        _check_marker(marker)
+    except InvalidMarker:
+        msg = f"Invalid environment marker: {marker} ({extra!r})"
+        raise DistutilsSetupError(msg) from None
+    list(_reqs.parse(reqs))
+
+
+def _check_marker(marker):
+    if not marker:
+        return
+    m = Marker(marker)
+    m.evaluate()
+
+
+def assert_bool(dist, attr, value):
+    """Verify that value is True, False, 0, or 1"""
+    if bool(value) != value:
+        raise DistutilsSetupError(f"{attr!r} must be a boolean value (got {value!r})")
+
+
+def invalid_unless_false(dist, attr, value):
+    if not value:
+        DistDeprecationWarning.emit(f"{attr} is ignored.")
+        # TODO: should there be a `due_date` here?
+        return
+    raise DistutilsSetupError(f"{attr} is invalid.")
+
+
+def check_requirements(dist, attr: str, value: _OrderedStrSequence) -> None:
+    """Verify that install_requires is a valid requirements list"""
+    try:
+        list(_reqs.parse(value))
+        if isinstance(value, set):
+            raise TypeError("Unordered types are not allowed")
+    except (TypeError, ValueError) as error:
+        msg = (
+            f"{attr!r} must be a string or iterable of strings "
+            f"containing valid project/version requirement specifiers; {error}"
+        )
+        raise DistutilsSetupError(msg) from error
+
+
+def check_specifier(dist, attr, value):
+    """Verify that value is a valid version specifier"""
+    try:
+        SpecifierSet(value)
+    except (InvalidSpecifier, AttributeError) as error:
+        msg = f"{attr!r} must be a string containing valid version specifiers; {error}"
+        raise DistutilsSetupError(msg) from error
+
+
+def check_entry_points(dist, attr, value):
+    """Verify that entry_points map is parseable"""
+    try:
+        _entry_points.load(value)
+    except Exception as e:
+        raise DistutilsSetupError(e) from e
+
+
+def check_package_data(dist, attr, value):
+    """Verify that value is a dictionary of package names to glob lists"""
+    if not isinstance(value, dict):
+        raise DistutilsSetupError(
+            f"{attr!r} must be a dictionary mapping package names to lists of "
+            "string wildcard patterns"
+        )
+    for k, v in value.items():
+        if not isinstance(k, str):
+            raise DistutilsSetupError(
+                f"keys of {attr!r} dict must be strings (got {k!r})"
+            )
+        assert_string_list(dist, f'values of {attr!r} dict', v)
+
+
+def check_packages(dist, attr, value):
+    for pkgname in value:
+        if not re.match(r'\w+(\.\w+)*', pkgname):
+            distutils.log.warn(
+                "WARNING: %r not a valid package name; please use only "
+                ".-separated package names in setup.py",
+                pkgname,
+            )
+
+
+if TYPE_CHECKING:
+    # Work around a mypy issue where type[T] can't be used as a base: https://github.com/python/mypy/issues/10962
+    from distutils.core import Distribution as _Distribution
+else:
+    _Distribution = get_unpatched(distutils.core.Distribution)
+
+
+class Distribution(_Distribution):
+    """Distribution with support for tests and package data
+
+    This is an enhanced version of 'distutils.dist.Distribution' that
+    effectively adds the following new optional keyword arguments to 'setup()':
+
+     'install_requires' -- a string or sequence of strings specifying project
+        versions that the distribution requires when installed, in the format
+        used by 'pkg_resources.require()'.  They will be installed
+        automatically when the package is installed.  If you wish to use
+        packages that are not available in PyPI, or want to give your users an
+        alternate download location, you can add a 'find_links' option to the
+        '[easy_install]' section of your project's 'setup.cfg' file, and then
+        setuptools will scan the listed web pages for links that satisfy the
+        requirements.
+
+     'extras_require' -- a dictionary mapping names of optional "extras" to the
+        additional requirement(s) that using those extras incurs. For example,
+        this::
+
+            extras_require = dict(reST = ["docutils>=0.3", "reSTedit"])
+
+        indicates that the distribution can optionally provide an extra
+        capability called "reST", but it can only be used if docutils and
+        reSTedit are installed.  If the user installs your package using
+        EasyInstall and requests one of your extras, the corresponding
+        additional requirements will be installed if needed.
+
+     'package_data' -- a dictionary mapping package names to lists of filenames
+        or globs to use to find data files contained in the named packages.
+        If the dictionary has filenames or globs listed under '""' (the empty
+        string), those names will be searched for in every package, in addition
+        to any names for the specific package.  Data files found using these
+        names/globs will be installed along with the package, in the same
+        location as the package.  Note that globs are allowed to reference
+        the contents of non-package subdirectories, as long as you use '/' as
+        a path separator.  (Globs are automatically converted to
+        platform-specific paths at runtime.)
+
+    In addition to these new keywords, this class also has several new methods
+    for manipulating the distribution's contents.  For example, the 'include()'
+    and 'exclude()' methods can be thought of as in-place add and subtract
+    commands that add or remove packages, modules, extensions, and so on from
+    the distribution.
+    """
+
+    _DISTUTILS_UNSUPPORTED_METADATA = {
+        'long_description_content_type': lambda: None,
+        'project_urls': dict,
+        'provides_extras': dict,  # behaves like an ordered set
+        'license_file': lambda: None,
+        'license_files': lambda: None,
+        'install_requires': list,
+        'extras_require': dict,
+    }
+
+    # Used by build_py, editable_wheel and install_lib commands for legacy namespaces
+    namespace_packages: list[str]  #: :meta private: DEPRECATED
+
+    # Any: Dynamic assignment results in Incompatible types in assignment
+    def __init__(self, attrs: MutableMapping[str, Any] | None = None) -> None:
+        have_package_data = hasattr(self, "package_data")
+        if not have_package_data:
+            self.package_data: dict[str, list[str]] = {}
+        attrs = attrs or {}
+        self.dist_files: list[tuple[str, str, str]] = []
+        self.include_package_data: bool | None = None
+        self.exclude_package_data: dict[str, list[str]] | None = None
+        # Filter-out setuptools' specific options.
+        self.src_root: str | None = attrs.pop("src_root", None)
+        self.dependency_links: list[str] = attrs.pop('dependency_links', [])
+        self.setup_requires: list[str] = attrs.pop('setup_requires', [])
+        for ep in metadata.entry_points(group='distutils.setup_keywords'):
+            vars(self).setdefault(ep.name, None)
+
+        metadata_only = set(self._DISTUTILS_UNSUPPORTED_METADATA)
+        metadata_only -= {"install_requires", "extras_require"}
+        dist_attrs = {k: v for k, v in attrs.items() if k not in metadata_only}
+        _Distribution.__init__(self, dist_attrs)
+
+        # Private API (setuptools-use only, not restricted to Distribution)
+        # Stores files that are referenced by the configuration and need to be in the
+        # sdist (e.g. `version = file: VERSION.txt`)
+        self._referenced_files = set[str]()
+
+        self.set_defaults = ConfigDiscovery(self)
+
+        self._set_metadata_defaults(attrs)
+
+        self.metadata.version = self._normalize_version(self.metadata.version)
+        self._finalize_requires()
+
+    def _validate_metadata(self):
+        required = {"name"}
+        provided = {
+            key
+            for key in vars(self.metadata)
+            if getattr(self.metadata, key, None) is not None
+        }
+        missing = required - provided
+
+        if missing:
+            msg = f"Required package metadata is missing: {missing}"
+            raise DistutilsSetupError(msg)
+
+    def _set_metadata_defaults(self, attrs):
+        """
+        Fill-in missing metadata fields not supported by distutils.
+        Some fields may have been set by other tools (e.g. pbr).
+        Those fields (vars(self.metadata)) take precedence to
+        supplied attrs.
+        """
+        for option, default in self._DISTUTILS_UNSUPPORTED_METADATA.items():
+            vars(self.metadata).setdefault(option, attrs.get(option, default()))
+
+    @staticmethod
+    def _normalize_version(version):
+        from . import sic
+
+        if isinstance(version, numbers.Number):
+            # Some people apparently take "version number" too literally :)
+            version = str(version)
+        elif isinstance(version, sic) or version is None:
+            return version
+
+        normalized = str(Version(version))
+        if version != normalized:
+            InformationOnly.emit(f"Normalizing '{version}' to '{normalized}'")
+            return normalized
+        return version
+
+    def _finalize_requires(self):
+        """
+        Set `metadata.python_requires` and fix environment markers
+        in `install_requires` and `extras_require`.
+        """
+        if getattr(self, 'python_requires', None):
+            self.metadata.python_requires = self.python_requires
+
+        self._normalize_requires()
+        self.metadata.install_requires = self.install_requires
+        self.metadata.extras_require = self.extras_require
+
+        if self.extras_require:
+            for extra in self.extras_require.keys():
+                # Setuptools allows a weird "<name>:<env markers> syntax for extras
+                extra = extra.split(':')[0]
+                if extra:
+                    self.metadata.provides_extras.setdefault(extra)
+
+    def _normalize_requires(self):
+        """Make sure requirement-related attributes exist and are normalized"""
+        install_requires = getattr(self, "install_requires", None) or []
+        extras_require = getattr(self, "extras_require", None) or {}
+
+        # Preserve the "static"-ness of values parsed from config files
+        list_ = _static.List if _static.is_static(install_requires) else list
+        self.install_requires = list_(map(str, _reqs.parse(install_requires)))
+
+        dict_ = _static.Dict if _static.is_static(extras_require) else dict
+        self.extras_require = dict_(
+            (k, list(map(str, _reqs.parse(v or [])))) for k, v in extras_require.items()
+        )
+
+    def _finalize_license_files(self) -> None:
+        """Compute names of all license files which should be included."""
+        license_files: list[str] | None = self.metadata.license_files
+        patterns = license_files or []
+
+        license_file: str | None = self.metadata.license_file
+        if license_file and license_file not in patterns:
+            patterns.append(license_file)
+
+        if license_files is None and license_file is None:
+            # Default patterns match the ones wheel uses
+            # See https://wheel.readthedocs.io/en/stable/user_guide.html
+            # -> 'Including license files in the generated wheel file'
+            patterns = ['LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*']
+
+        self.metadata.license_files = list(
+            unique_everseen(self._expand_patterns(patterns))
+        )
+
+    @staticmethod
+    def _expand_patterns(patterns):
+        """
+        >>> list(Distribution._expand_patterns(['LICENSE']))
+        ['LICENSE']
+        >>> list(Distribution._expand_patterns(['pyproject.toml', 'LIC*']))
+        ['pyproject.toml', 'LICENSE']
+        """
+        return (
+            path
+            for pattern in patterns
+            for path in sorted(iglob(pattern))
+            if not path.endswith('~') and os.path.isfile(path)
+        )
+
+    # FIXME: 'Distribution._parse_config_files' is too complex (14)
+    def _parse_config_files(self, filenames=None):  # noqa: C901
+        """
+        Adapted from distutils.dist.Distribution.parse_config_files,
+        this method provides the same functionality in subtly-improved
+        ways.
+        """
+        from configparser import ConfigParser
+
+        # Ignore install directory options if we have a venv
+        ignore_options = (
+            []
+            if sys.prefix == sys.base_prefix
+            else [
+                'install-base',
+                'install-platbase',
+                'install-lib',
+                'install-platlib',
+                'install-purelib',
+                'install-headers',
+                'install-scripts',
+                'install-data',
+                'prefix',
+                'exec-prefix',
+                'home',
+                'user',
+                'root',
+            ]
+        )
+
+        ignore_options = frozenset(ignore_options)
+
+        if filenames is None:
+            filenames = self.find_config_files()
+
+        if DEBUG:
+            self.announce("Distribution.parse_config_files():")
+
+        parser = ConfigParser()
+        parser.optionxform = str
+        for filename in filenames:
+            with open(filename, encoding='utf-8') as reader:
+                if DEBUG:
+                    self.announce("  reading {filename}".format(**locals()))
+                parser.read_file(reader)
+            for section in parser.sections():
+                options = parser.options(section)
+                opt_dict = self.get_option_dict(section)
+
+                for opt in options:
+                    if opt == '__name__' or opt in ignore_options:
+                        continue
+
+                    val = parser.get(section, opt)
+                    opt = self.warn_dash_deprecation(opt, section)
+                    opt = self.make_option_lowercase(opt, section)
+                    opt_dict[opt] = (filename, val)
+
+            # Make the ConfigParser forget everything (so we retain
+            # the original filenames that options come from)
+            parser.__init__()
+
+        if 'global' not in self.command_options:
+            return
+
+        # If there was a "global" section in the config file, use it
+        # to set Distribution options.
+
+        for opt, (src, val) in self.command_options['global'].items():
+            alias = self.negative_opt.get(opt)
+            if alias:
+                val = not strtobool(val)
+            elif opt in ('verbose', 'dry_run'):  # ugh!
+                val = strtobool(val)
+
+            try:
+                setattr(self, alias or opt, val)
+            except ValueError as e:
+                raise DistutilsOptionError(e) from e
+
+    def warn_dash_deprecation(self, opt: str, section: str) -> str:
+        if section in (
+            'options.extras_require',
+            'options.data_files',
+        ):
+            return opt
+
+        underscore_opt = opt.replace('-', '_')
+        commands = list(
+            itertools.chain(
+                distutils.command.__all__,
+                self._setuptools_commands(),
+            )
+        )
+        if (
+            not section.startswith('options')
+            and section != 'metadata'
+            and section not in commands
+        ):
+            return underscore_opt
+
+        if '-' in opt:
+            SetuptoolsDeprecationWarning.emit(
+                "Invalid dash-separated options",
+                f"""
+                Usage of dash-separated {opt!r} will not be supported in future
+                versions. Please use the underscore name {underscore_opt!r} instead.
+                """,
+                see_docs="userguide/declarative_config.html",
+                due_date=(2025, 3, 3),
+                # Warning initially introduced in 3 Mar 2021
+            )
+        return underscore_opt
+
+    def _setuptools_commands(self):
+        try:
+            entry_points = metadata.distribution('setuptools').entry_points
+            return {ep.name for ep in entry_points}  # Avoid newer API for compatibility
+        except metadata.PackageNotFoundError:
+            # during bootstrapping, distribution doesn't exist
+            return []
+
+    def make_option_lowercase(self, opt: str, section: str) -> str:
+        if section != 'metadata' or opt.islower():
+            return opt
+
+        lowercase_opt = opt.lower()
+        SetuptoolsDeprecationWarning.emit(
+            "Invalid uppercase configuration",
+            f"""
+            Usage of uppercase key {opt!r} in {section!r} will not be supported in
+            future versions. Please use lowercase {lowercase_opt!r} instead.
+            """,
+            see_docs="userguide/declarative_config.html",
+            due_date=(2025, 3, 3),
+            # Warning initially introduced in 6 Mar 2021
+        )
+        return lowercase_opt
+
+    # FIXME: 'Distribution._set_command_options' is too complex (14)
+    def _set_command_options(self, command_obj, option_dict=None):  # noqa: C901
+        """
+        Set the options for 'command_obj' from 'option_dict'.  Basically
+        this means copying elements of a dictionary ('option_dict') to
+        attributes of an instance ('command').
+
+        'command_obj' must be a Command instance.  If 'option_dict' is not
+        supplied, uses the standard option dictionary for this command
+        (from 'self.command_options').
+
+        (Adopted from distutils.dist.Distribution._set_command_options)
+        """
+        command_name = command_obj.get_command_name()
+        if option_dict is None:
+            option_dict = self.get_option_dict(command_name)
+
+        if DEBUG:
+            self.announce(f"  setting options for '{command_name}' command:")
+        for option, (source, value) in option_dict.items():
+            if DEBUG:
+                self.announce(f"    {option} = {value} (from {source})")
+            try:
+                bool_opts = [translate_longopt(o) for o in command_obj.boolean_options]
+            except AttributeError:
+                bool_opts = []
+            try:
+                neg_opt = command_obj.negative_opt
+            except AttributeError:
+                neg_opt = {}
+
+            try:
+                is_string = isinstance(value, str)
+                if option in neg_opt and is_string:
+                    setattr(command_obj, neg_opt[option], not strtobool(value))
+                elif option in bool_opts and is_string:
+                    setattr(command_obj, option, strtobool(value))
+                elif hasattr(command_obj, option):
+                    setattr(command_obj, option, value)
+                else:
+                    raise DistutilsOptionError(
+                        f"error in {source}: command '{command_name}' has no such option '{option}'"
+                    )
+            except ValueError as e:
+                raise DistutilsOptionError(e) from e
+
+    def _get_project_config_files(self, filenames: Iterable[StrPath] | None):
+        """Add default file and split between INI and TOML"""
+        tomlfiles = []
+        standard_project_metadata = Path(self.src_root or os.curdir, "pyproject.toml")
+        if filenames is not None:
+            parts = partition(lambda f: Path(f).suffix == ".toml", filenames)
+            filenames = list(parts[0])  # 1st element => predicate is False
+            tomlfiles = list(parts[1])  # 2nd element => predicate is True
+        elif standard_project_metadata.exists():
+            tomlfiles = [standard_project_metadata]
+        return filenames, tomlfiles
+
+    def parse_config_files(
+        self,
+        filenames: Iterable[StrPath] | None = None,
+        ignore_option_errors: bool = False,
+    ) -> None:
+        """Parses configuration files from various levels
+        and loads configuration.
+        """
+        inifiles, tomlfiles = self._get_project_config_files(filenames)
+
+        self._parse_config_files(filenames=inifiles)
+
+        setupcfg.parse_configuration(
+            self, self.command_options, ignore_option_errors=ignore_option_errors
+        )
+        for filename in tomlfiles:
+            pyprojecttoml.apply_configuration(self, filename, ignore_option_errors)
+
+        self._finalize_requires()
+        self._finalize_license_files()
+
+    def fetch_build_eggs(
+        self, requires: _StrOrIter
+    ) -> list[_pkg_resources_Distribution]:
+        """Resolve pre-setup requirements"""
+        from .installer import _fetch_build_eggs
+
+        return _fetch_build_eggs(self, requires)
+
+    def finalize_options(self) -> None:
+        """
+        Allow plugins to apply arbitrary operations to the
+        distribution. Each hook may optionally define a 'order'
+        to influence the order of execution. Smaller numbers
+        go first and the default is 0.
+        """
+        group = 'setuptools.finalize_distribution_options'
+
+        def by_order(hook):
+            return getattr(hook, 'order', 0)
+
+        defined = metadata.entry_points(group=group)
+        filtered = itertools.filterfalse(self._removed, defined)
+        loaded = map(lambda e: e.load(), filtered)
+        for ep in sorted(loaded, key=by_order):
+            ep(self)
+
+    @staticmethod
+    def _removed(ep):
+        """
+        When removing an entry point, if metadata is loaded
+        from an older version of Setuptools, that removed
+        entry point will attempt to be loaded and will fail.
+        See #2765 for more details.
+        """
+        removed = {
+            # removed 2021-09-05
+            '2to3_doctests',
+        }
+        return ep.name in removed
+
+    def _finalize_setup_keywords(self):
+        for ep in metadata.entry_points(group='distutils.setup_keywords'):
+            value = getattr(self, ep.name, None)
+            if value is not None:
+                ep.load()(self, ep.name, value)
+
+    def get_egg_cache_dir(self):
+        from . import windows_support
+
+        egg_cache_dir = os.path.join(os.curdir, '.eggs')
+        if not os.path.exists(egg_cache_dir):
+            os.mkdir(egg_cache_dir)
+            windows_support.hide_file(egg_cache_dir)
+            readme_txt_filename = os.path.join(egg_cache_dir, 'README.txt')
+            with open(readme_txt_filename, 'w', encoding="utf-8") as f:
+                f.write(
+                    'This directory contains eggs that were downloaded '
+                    'by setuptools to build, test, and run plug-ins.\n\n'
+                )
+                f.write(
+                    'This directory caches those eggs to prevent '
+                    'repeated downloads.\n\n'
+                )
+                f.write('However, it is safe to delete this directory.\n\n')
+
+        return egg_cache_dir
+
+    def fetch_build_egg(self, req):
+        """Fetch an egg needed for building"""
+        from .installer import fetch_build_egg
+
+        return fetch_build_egg(self, req)
+
+    def get_command_class(self, command: str) -> type[distutils.cmd.Command]:  # type: ignore[override] # Not doing complex overrides yet
+        """Pluggable version of get_command_class()"""
+        if command in self.cmdclass:
+            return self.cmdclass[command]
+
+        # Special case bdist_wheel so it's never loaded from "wheel"
+        if command == 'bdist_wheel':
+            from .command.bdist_wheel import bdist_wheel
+
+            return bdist_wheel
+
+        eps = metadata.entry_points(group='distutils.commands', name=command)
+        for ep in eps:
+            self.cmdclass[command] = cmdclass = ep.load()
+            return cmdclass
+        else:
+            return _Distribution.get_command_class(self, command)
+
+    def print_commands(self):
+        for ep in metadata.entry_points(group='distutils.commands'):
+            if ep.name not in self.cmdclass:
+                cmdclass = ep.load()
+                self.cmdclass[ep.name] = cmdclass
+        return _Distribution.print_commands(self)
+
+    def get_command_list(self):
+        for ep in metadata.entry_points(group='distutils.commands'):
+            if ep.name not in self.cmdclass:
+                cmdclass = ep.load()
+                self.cmdclass[ep.name] = cmdclass
+        return _Distribution.get_command_list(self)
+
+    def include(self, **attrs) -> None:
+        """Add items to distribution that are named in keyword arguments
+
+        For example, 'dist.include(py_modules=["x"])' would add 'x' to
+        the distribution's 'py_modules' attribute, if it was not already
+        there.
+
+        Currently, this method only supports inclusion for attributes that are
+        lists or tuples.  If you need to add support for adding to other
+        attributes in this or a subclass, you can add an '_include_X' method,
+        where 'X' is the name of the attribute.  The method will be called with
+        the value passed to 'include()'.  So, 'dist.include(foo={"bar":"baz"})'
+        will try to call 'dist._include_foo({"bar":"baz"})', which can then
+        handle whatever special inclusion logic is needed.
+        """
+        for k, v in attrs.items():
+            include = getattr(self, '_include_' + k, None)
+            if include:
+                include(v)
+            else:
+                self._include_misc(k, v)
+
+    def exclude_package(self, package: str) -> None:
+        """Remove packages, modules, and extensions in named package"""
+
+        pfx = package + '.'
+        if self.packages:
+            self.packages = [
+                p for p in self.packages if p != package and not p.startswith(pfx)
+            ]
+
+        if self.py_modules:
+            self.py_modules = [
+                p for p in self.py_modules if p != package and not p.startswith(pfx)
+            ]
+
+        if self.ext_modules:
+            self.ext_modules = [
+                p
+                for p in self.ext_modules
+                if p.name != package and not p.name.startswith(pfx)
+            ]
+
+    def has_contents_for(self, package: str) -> bool:
+        """Return true if 'exclude_package(package)' would do something"""
+
+        pfx = package + '.'
+
+        for p in self.iter_distribution_names():
+            if p == package or p.startswith(pfx):
+                return True
+
+        return False
+
+    def _exclude_misc(self, name: str, value: _Sequence) -> None:
+        """Handle 'exclude()' for list/tuple attrs without a special handler"""
+        if not isinstance(value, _sequence):
+            raise DistutilsSetupError(
+                f"{name}: setting must be of type <{_sequence_type_repr}> (got {value!r})"
+            )
+        try:
+            old = getattr(self, name)
+        except AttributeError as e:
+            raise DistutilsSetupError(f"{name}: No such distribution setting") from e
+        if old is not None and not isinstance(old, _sequence):
+            raise DistutilsSetupError(
+                name + ": this setting cannot be changed via include/exclude"
+            )
+        elif old:
+            setattr(self, name, [item for item in old if item not in value])
+
+    def _include_misc(self, name: str, value: _Sequence) -> None:
+        """Handle 'include()' for list/tuple attrs without a special handler"""
+
+        if not isinstance(value, _sequence):
+            raise DistutilsSetupError(
+                f"{name}: setting must be of type <{_sequence_type_repr}> (got {value!r})"
+            )
+        try:
+            old = getattr(self, name)
+        except AttributeError as e:
+            raise DistutilsSetupError(f"{name}: No such distribution setting") from e
+        if old is None:
+            setattr(self, name, value)
+        elif not isinstance(old, _sequence):
+            raise DistutilsSetupError(
+                name + ": this setting cannot be changed via include/exclude"
+            )
+        else:
+            new = [item for item in value if item not in old]
+            setattr(self, name, list(old) + new)
+
+    def exclude(self, **attrs) -> None:
+        """Remove items from distribution that are named in keyword arguments
+
+        For example, 'dist.exclude(py_modules=["x"])' would remove 'x' from
+        the distribution's 'py_modules' attribute.  Excluding packages uses
+        the 'exclude_package()' method, so all of the package's contained
+        packages, modules, and extensions are also excluded.
+
+        Currently, this method only supports exclusion from attributes that are
+        lists or tuples.  If you need to add support for excluding from other
+        attributes in this or a subclass, you can add an '_exclude_X' method,
+        where 'X' is the name of the attribute.  The method will be called with
+        the value passed to 'exclude()'.  So, 'dist.exclude(foo={"bar":"baz"})'
+        will try to call 'dist._exclude_foo({"bar":"baz"})', which can then
+        handle whatever special exclusion logic is needed.
+        """
+        for k, v in attrs.items():
+            exclude = getattr(self, '_exclude_' + k, None)
+            if exclude:
+                exclude(v)
+            else:
+                self._exclude_misc(k, v)
+
+    def _exclude_packages(self, packages: _Sequence) -> None:
+        if not isinstance(packages, _sequence):
+            raise DistutilsSetupError(
+                f"packages: setting must be of type <{_sequence_type_repr}> (got {packages!r})"
+            )
+        list(map(self.exclude_package, packages))
+
+    def _parse_command_opts(self, parser, args):
+        # Remove --with-X/--without-X options when processing command args
+        self.global_options = self.__class__.global_options
+        self.negative_opt = self.__class__.negative_opt
+
+        # First, expand any aliases
+        command = args[0]
+        aliases = self.get_option_dict('aliases')
+        while command in aliases:
+            _src, alias = aliases[command]
+            del aliases[command]  # ensure each alias can expand only once!
+            import shlex
+
+            args[:1] = shlex.split(alias, True)
+            command = args[0]
+
+        nargs = _Distribution._parse_command_opts(self, parser, args)
+
+        # Handle commands that want to consume all remaining arguments
+        cmd_class = self.get_command_class(command)
+        if getattr(cmd_class, 'command_consumes_arguments', None):
+            self.get_option_dict(command)['args'] = ("command line", nargs)
+            if nargs is not None:
+                return []
+
+        return nargs
+
+    def get_cmdline_options(self) -> dict[str, dict[str, str | None]]:
+        """Return a '{cmd: {opt:val}}' map of all command-line options
+
+        Option names are all long, but do not include the leading '--', and
+        contain dashes rather than underscores.  If the option doesn't take
+        an argument (e.g. '--quiet'), the 'val' is 'None'.
+
+        Note that options provided by config files are intentionally excluded.
+        """
+
+        d: dict[str, dict[str, str | None]] = {}
+
+        for cmd, opts in self.command_options.items():
+            val: str | None
+            for opt, (src, val) in opts.items():
+                if src != "command line":
+                    continue
+
+                opt = opt.replace('_', '-')
+
+                if val == 0:
+                    cmdobj = self.get_command_obj(cmd)
+                    neg_opt = self.negative_opt.copy()
+                    neg_opt.update(getattr(cmdobj, 'negative_opt', {}))
+                    for neg, pos in neg_opt.items():
+                        if pos == opt:
+                            opt = neg
+                            val = None
+                            break
+                    else:
+                        raise AssertionError("Shouldn't be able to get here")
+
+                elif val == 1:
+                    val = None
+
+                d.setdefault(cmd, {})[opt] = val
+
+        return d
+
+    def iter_distribution_names(self):
+        """Yield all packages, modules, and extension names in distribution"""
+
+        yield from self.packages or ()
+
+        yield from self.py_modules or ()
+
+        for ext in self.ext_modules or ():
+            if isinstance(ext, tuple):
+                name, _buildinfo = ext
+            else:
+                name = ext.name
+            if name.endswith('module'):
+                name = name[:-6]
+            yield name
+
+    def handle_display_options(self, option_order):
+        """If there were any non-global "display-only" options
+        (--help-commands or the metadata display options) on the command
+        line, display the requested info and return true; else return
+        false.
+        """
+        import sys
+
+        if self.help_commands:
+            return _Distribution.handle_display_options(self, option_order)
+
+        # Stdout may be StringIO (e.g. in tests)
+        if not isinstance(sys.stdout, io.TextIOWrapper):
+            return _Distribution.handle_display_options(self, option_order)
+
+        # Don't wrap stdout if utf-8 is already the encoding. Provides
+        #  workaround for #334.
+        if sys.stdout.encoding.lower() in ('utf-8', 'utf8'):
+            return _Distribution.handle_display_options(self, option_order)
+
+        # Print metadata in UTF-8 no matter the platform
+        encoding = sys.stdout.encoding
+        sys.stdout.reconfigure(encoding='utf-8')
+        try:
+            return _Distribution.handle_display_options(self, option_order)
+        finally:
+            sys.stdout.reconfigure(encoding=encoding)
+
+    def run_command(self, command) -> None:
+        self.set_defaults()
+        # Postpone defaults until all explicit configuration is considered
+        # (setup() args, config files, command line and plugins)
+
+        super().run_command(command)
+
+
+class DistDeprecationWarning(SetuptoolsDeprecationWarning):
+    """Class for warning about deprecations in dist in
+    setuptools. Not ignored by default, unlike DeprecationWarning."""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/errors.py b/backend/venv/lib/python3.12/site-packages/setuptools/errors.py
new file mode 100644
index 00000000..990ecbf4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/errors.py
@@ -0,0 +1,67 @@
+"""setuptools.errors
+
+Provides exceptions used by setuptools modules.
+"""
+
+from __future__ import annotations
+
+from distutils import errors as _distutils_errors
+
+# Re-export errors from distutils to facilitate the migration to PEP632
+
+ByteCompileError = _distutils_errors.DistutilsByteCompileError
+CCompilerError = _distutils_errors.CCompilerError
+ClassError = _distutils_errors.DistutilsClassError
+CompileError = _distutils_errors.CompileError
+ExecError = _distutils_errors.DistutilsExecError
+FileError = _distutils_errors.DistutilsFileError
+InternalError = _distutils_errors.DistutilsInternalError
+LibError = _distutils_errors.LibError
+LinkError = _distutils_errors.LinkError
+ModuleError = _distutils_errors.DistutilsModuleError
+OptionError = _distutils_errors.DistutilsOptionError
+PlatformError = _distutils_errors.DistutilsPlatformError
+PreprocessError = _distutils_errors.PreprocessError
+SetupError = _distutils_errors.DistutilsSetupError
+TemplateError = _distutils_errors.DistutilsTemplateError
+UnknownFileError = _distutils_errors.UnknownFileError
+
+# The root error class in the hierarchy
+BaseError = _distutils_errors.DistutilsError
+
+
+class InvalidConfigError(OptionError):  # type: ignore[valid-type, misc] # distutils imports are `Any` on python 3.12+
+    """Error used for invalid configurations."""
+
+
+class RemovedConfigError(OptionError):  # type: ignore[valid-type, misc] # distutils imports are `Any` on python 3.12+
+    """Error used for configurations that were deprecated and removed."""
+
+
+class RemovedCommandError(BaseError, RuntimeError):  # type: ignore[valid-type, misc] # distutils imports are `Any` on python 3.12+
+    """Error used for commands that have been removed in setuptools.
+
+    Since ``setuptools`` is built on ``distutils``, simply removing a command
+    from ``setuptools`` will make the behavior fall back to ``distutils``; this
+    error is raised if a command exists in ``distutils`` but has been actively
+    removed in ``setuptools``.
+    """
+
+
+class PackageDiscoveryError(BaseError, RuntimeError):  # type: ignore[valid-type, misc] # distutils imports are `Any` on python 3.12+
+    """Impossible to perform automatic discovery of packages and/or modules.
+
+    The current project layout or given discovery options can lead to problems when
+    scanning the project directory.
+
+    Setuptools might also refuse to complete auto-discovery if an error prone condition
+    is detected (e.g. when a project is organised as a flat-layout but contains
+    multiple directories that can be taken as top-level packages inside a single
+    distribution [*]_). In these situations the users are encouraged to be explicit
+    about which packages to include or to make the discovery parameters more specific.
+
+    .. [*] Since multi-package distributions are uncommon it is very likely that the
+       developers did not intend for all the directories to be packaged, and are just
+       leaving auxiliary code in the repository top-level, such as maintenance-related
+       scripts.
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/extension.py b/backend/venv/lib/python3.12/site-packages/setuptools/extension.py
new file mode 100644
index 00000000..76e03d9d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/extension.py
@@ -0,0 +1,177 @@
+from __future__ import annotations
+
+import functools
+import re
+from typing import TYPE_CHECKING
+
+from setuptools._path import StrPath
+
+from .monkey import get_unpatched
+
+import distutils.core
+import distutils.errors
+import distutils.extension
+
+
+def _have_cython():
+    """
+    Return True if Cython can be imported.
+    """
+    cython_impl = 'Cython.Distutils.build_ext'
+    try:
+        # from (cython_impl) import build_ext
+        __import__(cython_impl, fromlist=['build_ext']).build_ext
+    except Exception:
+        return False
+    return True
+
+
+# for compatibility
+have_pyrex = _have_cython
+if TYPE_CHECKING:
+    # Work around a mypy issue where type[T] can't be used as a base: https://github.com/python/mypy/issues/10962
+    from distutils.core import Extension as _Extension
+else:
+    _Extension = get_unpatched(distutils.core.Extension)
+
+
+class Extension(_Extension):
+    """
+    Describes a single extension module.
+
+    This means that all source files will be compiled into a single binary file
+    ``<module path>.<suffix>`` (with ``<module path>`` derived from ``name`` and
+    ``<suffix>`` defined by one of the values in
+    ``importlib.machinery.EXTENSION_SUFFIXES``).
+
+    In the case ``.pyx`` files are passed as ``sources and`` ``Cython`` is **not**
+    installed in the build environment, ``setuptools`` may also try to look for the
+    equivalent ``.cpp`` or ``.c`` files.
+
+    :arg str name:
+      the full name of the extension, including any packages -- ie.
+      *not* a filename or pathname, but Python dotted name
+
+    :arg list[str|os.PathLike[str]] sources:
+      list of source filenames, relative to the distribution root
+      (where the setup script lives), in Unix form (slash-separated)
+      for portability.  Source files may be C, C++, SWIG (.i),
+      platform-specific resource files, or whatever else is recognized
+      by the "build_ext" command as source for a Python extension.
+
+    :keyword list[str] include_dirs:
+      list of directories to search for C/C++ header files (in Unix
+      form for portability)
+
+    :keyword list[tuple[str, str|None]] define_macros:
+      list of macros to define; each macro is defined using a 2-tuple:
+      the first item corresponding to the name of the macro and the second
+      item either a string with its value or None to
+      define it without a particular value (equivalent of "#define
+      FOO" in source or -DFOO on Unix C compiler command line)
+
+    :keyword list[str] undef_macros:
+      list of macros to undefine explicitly
+
+    :keyword list[str] library_dirs:
+      list of directories to search for C/C++ libraries at link time
+
+    :keyword list[str] libraries:
+      list of library names (not filenames or paths) to link against
+
+    :keyword list[str] runtime_library_dirs:
+      list of directories to search for C/C++ libraries at run time
+      (for shared extensions, this is when the extension is loaded).
+      Setting this will cause an exception during build on Windows
+      platforms.
+
+    :keyword list[str] extra_objects:
+      list of extra files to link with (eg. object files not implied
+      by 'sources', static library that must be explicitly specified,
+      binary resource files, etc.)
+
+    :keyword list[str] extra_compile_args:
+      any extra platform- and compiler-specific information to use
+      when compiling the source files in 'sources'.  For platforms and
+      compilers where "command line" makes sense, this is typically a
+      list of command-line arguments, but for other platforms it could
+      be anything.
+
+    :keyword list[str] extra_link_args:
+      any extra platform- and compiler-specific information to use
+      when linking object files together to create the extension (or
+      to create a new static Python interpreter).  Similar
+      interpretation as for 'extra_compile_args'.
+
+    :keyword list[str] export_symbols:
+      list of symbols to be exported from a shared extension.  Not
+      used on all platforms, and not generally necessary for Python
+      extensions, which typically export exactly one symbol: "init" +
+      extension_name.
+
+    :keyword list[str] swig_opts:
+      any extra options to pass to SWIG if a source file has the .i
+      extension.
+
+    :keyword list[str] depends:
+      list of files that the extension depends on
+
+    :keyword str language:
+      extension language (i.e. "c", "c++", "objc"). Will be detected
+      from the source extensions if not provided.
+
+    :keyword bool optional:
+      specifies that a build failure in the extension should not abort the
+      build process, but simply not install the failing extension.
+
+    :keyword bool py_limited_api:
+      opt-in flag for the usage of :doc:`Python's limited API <python:c-api/stable>`.
+
+    :raises setuptools.errors.PlatformError: if ``runtime_library_dirs`` is
+      specified on Windows. (since v63)
+    """
+
+    # These 4 are set and used in setuptools/command/build_ext.py
+    # The lack of a default value and risk of `AttributeError` is purposeful
+    # to avoid people forgetting to call finalize_options if they modify the extension list.
+    # See example/rationale in https://github.com/pypa/setuptools/issues/4529.
+    _full_name: str  #: Private API, internal use only.
+    _links_to_dynamic: bool  #: Private API, internal use only.
+    _needs_stub: bool  #: Private API, internal use only.
+    _file_name: str  #: Private API, internal use only.
+
+    def __init__(
+        self,
+        name: str,
+        sources: list[StrPath],
+        *args,
+        py_limited_api: bool = False,
+        **kw,
+    ) -> None:
+        # The *args is needed for compatibility as calls may use positional
+        # arguments. py_limited_api may be set only via keyword.
+        self.py_limited_api = py_limited_api
+        super().__init__(
+            name,
+            sources,  # type: ignore[arg-type] # Vendored version of setuptools supports PathLike
+            *args,
+            **kw,
+        )
+
+    def _convert_pyx_sources_to_lang(self):
+        """
+        Replace sources with .pyx extensions to sources with the target
+        language extension. This mechanism allows language authors to supply
+        pre-converted sources but to prefer the .pyx sources.
+        """
+        if _have_cython():
+            # the build has Cython, so allow it to compile the .pyx files
+            return
+        lang = self.language or ''
+        target_ext = '.cpp' if lang.lower() == 'c++' else '.c'
+        sub = functools.partial(re.sub, '.pyx$', target_ext)
+        self.sources = list(map(sub, self.sources))
+
+
+class Library(Extension):
+    """Just like a regular Extension, but built as a library instead"""
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/glob.py b/backend/venv/lib/python3.12/site-packages/setuptools/glob.py
new file mode 100644
index 00000000..1dfff2cd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/glob.py
@@ -0,0 +1,185 @@
+"""
+Filename globbing utility. Mostly a copy of `glob` from Python 3.5.
+
+Changes include:
+ * `yield from` and PEP3102 `*` removed.
+ * Hidden files are not ignored.
+"""
+
+from __future__ import annotations
+
+import fnmatch
+import os
+import re
+from collections.abc import Iterable, Iterator
+from typing import TYPE_CHECKING, AnyStr, overload
+
+if TYPE_CHECKING:
+    from _typeshed import BytesPath, StrOrBytesPath, StrPath
+
+__all__ = ["glob", "iglob", "escape"]
+
+
+def glob(pathname: AnyStr, recursive: bool = False) -> list[AnyStr]:
+    """Return a list of paths matching a pathname pattern.
+
+    The pattern may contain simple shell-style wildcards a la
+    fnmatch. However, unlike fnmatch, filenames starting with a
+    dot are special cases that are not matched by '*' and '?'
+    patterns.
+
+    If recursive is true, the pattern '**' will match any files and
+    zero or more directories and subdirectories.
+    """
+    return list(iglob(pathname, recursive=recursive))
+
+
+def iglob(pathname: AnyStr, recursive: bool = False) -> Iterator[AnyStr]:
+    """Return an iterator which yields the paths matching a pathname pattern.
+
+    The pattern may contain simple shell-style wildcards a la
+    fnmatch. However, unlike fnmatch, filenames starting with a
+    dot are special cases that are not matched by '*' and '?'
+    patterns.
+
+    If recursive is true, the pattern '**' will match any files and
+    zero or more directories and subdirectories.
+    """
+    it = _iglob(pathname, recursive)
+    if recursive and _isrecursive(pathname):
+        s = next(it)  # skip empty string
+        assert not s
+    return it
+
+
+def _iglob(pathname: AnyStr, recursive: bool) -> Iterator[AnyStr]:
+    dirname, basename = os.path.split(pathname)
+    glob_in_dir = glob2 if recursive and _isrecursive(basename) else glob1
+
+    if not has_magic(pathname):
+        if basename:
+            if os.path.lexists(pathname):
+                yield pathname
+        else:
+            # Patterns ending with a slash should match only directories
+            if os.path.isdir(dirname):
+                yield pathname
+        return
+
+    if not dirname:
+        yield from glob_in_dir(dirname, basename)
+        return
+    # `os.path.split()` returns the argument itself as a dirname if it is a
+    # drive or UNC path.  Prevent an infinite recursion if a drive or UNC path
+    # contains magic characters (i.e. r'\\?\C:').
+    if dirname != pathname and has_magic(dirname):
+        dirs: Iterable[AnyStr] = _iglob(dirname, recursive)
+    else:
+        dirs = [dirname]
+    if not has_magic(basename):
+        glob_in_dir = glob0
+    for dirname in dirs:
+        for name in glob_in_dir(dirname, basename):
+            yield os.path.join(dirname, name)
+
+
+# These 2 helper functions non-recursively glob inside a literal directory.
+# They return a list of basenames. `glob1` accepts a pattern while `glob0`
+# takes a literal basename (so it only has to check for its existence).
+
+
+@overload
+def glob1(dirname: StrPath, pattern: str) -> list[str]: ...
+@overload
+def glob1(dirname: BytesPath, pattern: bytes) -> list[bytes]: ...
+def glob1(dirname: StrOrBytesPath, pattern: str | bytes) -> list[str] | list[bytes]:
+    if not dirname:
+        if isinstance(pattern, bytes):
+            dirname = os.curdir.encode('ASCII')
+        else:
+            dirname = os.curdir
+    try:
+        names = os.listdir(dirname)
+    except OSError:
+        return []
+    # mypy false-positives: str or bytes type possibility is always kept in sync
+    return fnmatch.filter(names, pattern)  # type: ignore[type-var, return-value]
+
+
+def glob0(dirname, basename):
+    if not basename:
+        # `os.path.split()` returns an empty basename for paths ending with a
+        # directory separator.  'q*x/' should match only directories.
+        if os.path.isdir(dirname):
+            return [basename]
+    else:
+        if os.path.lexists(os.path.join(dirname, basename)):
+            return [basename]
+    return []
+
+
+# This helper function recursively yields relative pathnames inside a literal
+# directory.
+
+
+@overload
+def glob2(dirname: StrPath, pattern: str) -> Iterator[str]: ...
+@overload
+def glob2(dirname: BytesPath, pattern: bytes) -> Iterator[bytes]: ...
+def glob2(dirname: StrOrBytesPath, pattern: str | bytes) -> Iterator[str | bytes]:
+    assert _isrecursive(pattern)
+    yield pattern[:0]
+    yield from _rlistdir(dirname)
+
+
+# Recursively yields relative pathnames inside a literal directory.
+@overload
+def _rlistdir(dirname: StrPath) -> Iterator[str]: ...
+@overload
+def _rlistdir(dirname: BytesPath) -> Iterator[bytes]: ...
+def _rlistdir(dirname: StrOrBytesPath) -> Iterator[str | bytes]:
+    if not dirname:
+        if isinstance(dirname, bytes):
+            dirname = os.curdir.encode('ASCII')
+        else:
+            dirname = os.curdir
+    try:
+        names = os.listdir(dirname)
+    except OSError:
+        return
+    for x in names:
+        yield x
+        # mypy false-positives: str or bytes type possibility is always kept in sync
+        path = os.path.join(dirname, x) if dirname else x  # type: ignore[arg-type]
+        for y in _rlistdir(path):
+            yield os.path.join(x, y)  # type: ignore[arg-type]
+
+
+magic_check = re.compile('([*?[])')
+magic_check_bytes = re.compile(b'([*?[])')
+
+
+def has_magic(s: str | bytes) -> bool:
+    if isinstance(s, bytes):
+        return magic_check_bytes.search(s) is not None
+    else:
+        return magic_check.search(s) is not None
+
+
+def _isrecursive(pattern: str | bytes) -> bool:
+    if isinstance(pattern, bytes):
+        return pattern == b'**'
+    else:
+        return pattern == '**'
+
+
+def escape(pathname):
+    """Escape all special characters."""
+    # Escaping is done by wrapping any of "*?[" between square brackets.
+    # Metacharacters do not work in the drive part and shouldn't be escaped.
+    drive, pathname = os.path.splitdrive(pathname)
+    if isinstance(pathname, bytes):
+        pathname = magic_check_bytes.sub(rb'[\1]', pathname)
+    else:
+        pathname = magic_check.sub(r'[\1]', pathname)
+    return drive + pathname
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/gui-32.exe b/backend/venv/lib/python3.12/site-packages/setuptools/gui-32.exe
new file mode 100644
index 00000000..1eb430c6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/gui-32.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/gui-64.exe b/backend/venv/lib/python3.12/site-packages/setuptools/gui-64.exe
new file mode 100644
index 00000000..031cb77c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/gui-64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/gui-arm64.exe b/backend/venv/lib/python3.12/site-packages/setuptools/gui-arm64.exe
new file mode 100644
index 00000000..1e00ffac
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/gui-arm64.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/gui.exe b/backend/venv/lib/python3.12/site-packages/setuptools/gui.exe
new file mode 100644
index 00000000..1eb430c6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/gui.exe differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/installer.py b/backend/venv/lib/python3.12/site-packages/setuptools/installer.py
new file mode 100644
index 00000000..64bc2def
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/installer.py
@@ -0,0 +1,150 @@
+from __future__ import annotations
+
+import glob
+import os
+import subprocess
+import sys
+import tempfile
+from functools import partial
+
+from pkg_resources import Distribution
+
+from . import _reqs
+from ._reqs import _StrOrIter
+from .warnings import SetuptoolsDeprecationWarning
+from .wheel import Wheel
+
+from distutils import log
+from distutils.errors import DistutilsError
+
+
+def _fixup_find_links(find_links):
+    """Ensure find-links option end-up being a list of strings."""
+    if isinstance(find_links, str):
+        return find_links.split()
+    assert isinstance(find_links, (tuple, list))
+    return find_links
+
+
+def fetch_build_egg(dist, req):
+    """Fetch an egg needed for building.
+
+    Use pip/wheel to fetch/build a wheel."""
+    _DeprecatedInstaller.emit()
+    _warn_wheel_not_available(dist)
+    return _fetch_build_egg_no_warn(dist, req)
+
+
+def _fetch_build_eggs(dist, requires: _StrOrIter) -> list[Distribution]:
+    import pkg_resources  # Delay import to avoid unnecessary side-effects
+
+    _DeprecatedInstaller.emit(stacklevel=3)
+    _warn_wheel_not_available(dist)
+
+    resolved_dists = pkg_resources.working_set.resolve(
+        _reqs.parse(requires, pkg_resources.Requirement),  # required for compatibility
+        installer=partial(_fetch_build_egg_no_warn, dist),  # avoid warning twice
+        replace_conflicting=True,
+    )
+    for dist in resolved_dists:
+        pkg_resources.working_set.add(dist, replace=True)
+    return resolved_dists
+
+
+def _fetch_build_egg_no_warn(dist, req):  # noqa: C901  # is too complex (16)  # FIXME
+    import pkg_resources  # Delay import to avoid unnecessary side-effects
+
+    # Ignore environment markers; if supplied, it is required.
+    req = strip_marker(req)
+    # Take easy_install options into account, but do not override relevant
+    # pip environment variables (like PIP_INDEX_URL or PIP_QUIET); they'll
+    # take precedence.
+    opts = dist.get_option_dict('easy_install')
+    if 'allow_hosts' in opts:
+        raise DistutilsError(
+            'the `allow-hosts` option is not supported '
+            'when using pip to install requirements.'
+        )
+    quiet = 'PIP_QUIET' not in os.environ and 'PIP_VERBOSE' not in os.environ
+    if 'PIP_INDEX_URL' in os.environ:
+        index_url = None
+    elif 'index_url' in opts:
+        index_url = opts['index_url'][1]
+    else:
+        index_url = None
+    find_links = (
+        _fixup_find_links(opts['find_links'][1])[:] if 'find_links' in opts else []
+    )
+    if dist.dependency_links:
+        find_links.extend(dist.dependency_links)
+    eggs_dir = os.path.realpath(dist.get_egg_cache_dir())
+    environment = pkg_resources.Environment()
+    for egg_dist in pkg_resources.find_distributions(eggs_dir):
+        if egg_dist in req and environment.can_add(egg_dist):
+            return egg_dist
+    with tempfile.TemporaryDirectory() as tmpdir:
+        cmd = [
+            sys.executable,
+            '-m',
+            'pip',
+            '--disable-pip-version-check',
+            'wheel',
+            '--no-deps',
+            '-w',
+            tmpdir,
+        ]
+        if quiet:
+            cmd.append('--quiet')
+        if index_url is not None:
+            cmd.extend(('--index-url', index_url))
+        for link in find_links or []:
+            cmd.extend(('--find-links', link))
+        # If requirement is a PEP 508 direct URL, directly pass
+        # the URL to pip, as `req @ url` does not work on the
+        # command line.
+        cmd.append(req.url or str(req))
+        try:
+            subprocess.check_call(cmd)
+        except subprocess.CalledProcessError as e:
+            raise DistutilsError(str(e)) from e
+        wheel = Wheel(glob.glob(os.path.join(tmpdir, '*.whl'))[0])
+        dist_location = os.path.join(eggs_dir, wheel.egg_name())
+        wheel.install_as_egg(dist_location)
+        dist_metadata = pkg_resources.PathMetadata(
+            dist_location, os.path.join(dist_location, 'EGG-INFO')
+        )
+        return pkg_resources.Distribution.from_filename(
+            dist_location, metadata=dist_metadata
+        )
+
+
+def strip_marker(req):
+    """
+    Return a new requirement without the environment marker to avoid
+    calling pip with something like `babel; extra == "i18n"`, which
+    would always be ignored.
+    """
+    import pkg_resources  # Delay import to avoid unnecessary side-effects
+
+    # create a copy to avoid mutating the input
+    req = pkg_resources.Requirement.parse(str(req))
+    req.marker = None
+    return req
+
+
+def _warn_wheel_not_available(dist):
+    import pkg_resources  # Delay import to avoid unnecessary side-effects
+
+    try:
+        pkg_resources.get_distribution('wheel')
+    except pkg_resources.DistributionNotFound:
+        dist.announce('WARNING: The wheel package is not available.', log.WARN)
+
+
+class _DeprecatedInstaller(SetuptoolsDeprecationWarning):
+    _SUMMARY = "setuptools.installer and fetch_build_eggs are deprecated."
+    _DETAILS = """
+    Requirements should be satisfied by a PEP 517 installer.
+    If you are using pip, you can try `pip install --use-pep517`.
+    """
+    # _DUE_DATE not decided yet
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/launch.py b/backend/venv/lib/python3.12/site-packages/setuptools/launch.py
new file mode 100644
index 00000000..0d162647
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/launch.py
@@ -0,0 +1,36 @@
+"""
+Launch the Python script on the command line after
+setuptools is bootstrapped via import.
+"""
+
+# Note that setuptools gets imported implicitly by the
+# invocation of this script using python -m setuptools.launch
+
+import sys
+import tokenize
+
+
+def run() -> None:
+    """
+    Run the script in sys.argv[1] as if it had
+    been invoked naturally.
+    """
+    __builtins__
+    script_name = sys.argv[1]
+    namespace = dict(
+        __file__=script_name,
+        __name__='__main__',
+        __doc__=None,
+    )
+    sys.argv[:] = sys.argv[1:]
+
+    open_ = getattr(tokenize, 'open', open)
+    with open_(script_name) as fid:
+        script = fid.read()
+    norm_script = script.replace('\\r\\n', '\\n')
+    code = compile(norm_script, script_name, 'exec')
+    exec(code, namespace)
+
+
+if __name__ == '__main__':
+    run()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/logging.py b/backend/venv/lib/python3.12/site-packages/setuptools/logging.py
new file mode 100644
index 00000000..532da899
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/logging.py
@@ -0,0 +1,40 @@
+import inspect
+import logging
+import sys
+
+from . import monkey
+
+import distutils.log
+
+
+def _not_warning(record):
+    return record.levelno < logging.WARNING
+
+
+def configure() -> None:
+    """
+    Configure logging to emit warning and above to stderr
+    and everything else to stdout. This behavior is provided
+    for compatibility with distutils.log but may change in
+    the future.
+    """
+    err_handler = logging.StreamHandler()
+    err_handler.setLevel(logging.WARNING)
+    out_handler = logging.StreamHandler(sys.stdout)
+    out_handler.addFilter(_not_warning)
+    handlers = err_handler, out_handler
+    logging.basicConfig(
+        format="{message}", style='{', handlers=handlers, level=logging.DEBUG
+    )
+    if inspect.ismodule(distutils.dist.log):
+        monkey.patch_func(set_threshold, distutils.log, 'set_threshold')
+        # For some reason `distutils.log` module is getting cached in `distutils.dist`
+        # and then loaded again when patched,
+        # implying: id(distutils.log) != id(distutils.dist.log).
+        # Make sure the same module object is used everywhere:
+        distutils.dist.log = distutils.log
+
+
+def set_threshold(level: int) -> int:
+    logging.root.setLevel(level * 10)
+    return set_threshold.unpatched(level)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/modified.py b/backend/venv/lib/python3.12/site-packages/setuptools/modified.py
new file mode 100644
index 00000000..6ba02fab
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/modified.py
@@ -0,0 +1,18 @@
+try:
+    # Ensure a DistutilsError raised by these methods is the same as distutils.errors.DistutilsError
+    from distutils._modified import (
+        newer,
+        newer_group,
+        newer_pairwise,
+        newer_pairwise_group,
+    )
+except ImportError:
+    # fallback for SETUPTOOLS_USE_DISTUTILS=stdlib, because _modified never existed in stdlib
+    from ._distutils._modified import (
+        newer,
+        newer_group,
+        newer_pairwise,
+        newer_pairwise_group,
+    )
+
+__all__ = ['newer', 'newer_pairwise', 'newer_group', 'newer_pairwise_group']
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/monkey.py b/backend/venv/lib/python3.12/site-packages/setuptools/monkey.py
new file mode 100644
index 00000000..6ad1abac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/monkey.py
@@ -0,0 +1,126 @@
+"""
+Monkey patching of distutils.
+"""
+
+from __future__ import annotations
+
+import inspect
+import platform
+import sys
+import types
+from typing import TypeVar, cast, overload
+
+import distutils.filelist
+
+_T = TypeVar("_T")
+_UnpatchT = TypeVar("_UnpatchT", type, types.FunctionType)
+
+
+__all__: list[str] = []
+"""
+Everything is private. Contact the project team
+if you think you need this functionality.
+"""
+
+
+def _get_mro(cls):
+    """
+    Returns the bases classes for cls sorted by the MRO.
+
+    Works around an issue on Jython where inspect.getmro will not return all
+    base classes if multiple classes share the same name. Instead, this
+    function will return a tuple containing the class itself, and the contents
+    of cls.__bases__. See https://github.com/pypa/setuptools/issues/1024.
+    """
+    if platform.python_implementation() == "Jython":
+        return (cls,) + cls.__bases__
+    return inspect.getmro(cls)
+
+
+@overload
+def get_unpatched(item: _UnpatchT) -> _UnpatchT: ...
+@overload
+def get_unpatched(item: object) -> None: ...
+def get_unpatched(
+    item: type | types.FunctionType | object,
+) -> type | types.FunctionType | None:
+    if isinstance(item, type):
+        return get_unpatched_class(item)
+    if isinstance(item, types.FunctionType):
+        return get_unpatched_function(item)
+    return None
+
+
+def get_unpatched_class(cls: type[_T]) -> type[_T]:
+    """Protect against re-patching the distutils if reloaded
+
+    Also ensures that no other distutils extension monkeypatched the distutils
+    first.
+    """
+    external_bases = (
+        cast(type[_T], cls)
+        for cls in _get_mro(cls)
+        if not cls.__module__.startswith('setuptools')
+    )
+    base = next(external_bases)
+    if not base.__module__.startswith('distutils'):
+        msg = f"distutils has already been patched by {cls!r}"
+        raise AssertionError(msg)
+    return base
+
+
+def patch_all():
+    import setuptools
+
+    # we can't patch distutils.cmd, alas
+    distutils.core.Command = setuptools.Command  # type: ignore[misc,assignment] # monkeypatching
+
+    _patch_distribution_metadata()
+
+    # Install Distribution throughout the distutils
+    for module in distutils.dist, distutils.core, distutils.cmd:
+        module.Distribution = setuptools.dist.Distribution
+
+    # Install the patched Extension
+    distutils.core.Extension = setuptools.extension.Extension  # type: ignore[misc,assignment] # monkeypatching
+    distutils.extension.Extension = setuptools.extension.Extension  # type: ignore[misc,assignment] # monkeypatching
+    if 'distutils.command.build_ext' in sys.modules:
+        sys.modules[
+            'distutils.command.build_ext'
+        ].Extension = setuptools.extension.Extension
+
+
+def _patch_distribution_metadata():
+    from . import _core_metadata
+
+    """Patch write_pkg_file and read_pkg_file for higher metadata standards"""
+    for attr in (
+        'write_pkg_info',
+        'write_pkg_file',
+        'read_pkg_file',
+        'get_metadata_version',
+        'get_fullname',
+    ):
+        new_val = getattr(_core_metadata, attr)
+        setattr(distutils.dist.DistributionMetadata, attr, new_val)
+
+
+def patch_func(replacement, target_mod, func_name):
+    """
+    Patch func_name in target_mod with replacement
+
+    Important - original must be resolved by name to avoid
+    patching an already patched function.
+    """
+    original = getattr(target_mod, func_name)
+
+    # set the 'unpatched' attribute on the replacement to
+    # point to the original.
+    vars(replacement).setdefault('unpatched', original)
+
+    # replace the function in the original module
+    setattr(target_mod, func_name, replacement)
+
+
+def get_unpatched_function(candidate):
+    return candidate.unpatched
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/msvc.py b/backend/venv/lib/python3.12/site-packages/setuptools/msvc.py
new file mode 100644
index 00000000..9c9a6356
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/msvc.py
@@ -0,0 +1,1527 @@
+"""
+Environment info about Microsoft Compilers.
+
+>>> getfixture('windows_only')
+>>> ei = EnvironmentInfo('amd64')
+"""
+
+from __future__ import annotations
+
+import contextlib
+import itertools
+import json
+import os
+import os.path
+import platform
+from typing import TYPE_CHECKING, TypedDict
+
+from more_itertools import unique_everseen
+
+import distutils.errors
+
+if TYPE_CHECKING:
+    from typing_extensions import LiteralString, NotRequired
+
+# https://github.com/python/mypy/issues/8166
+if not TYPE_CHECKING and platform.system() == 'Windows':
+    import winreg
+    from os import environ
+else:
+    # Mock winreg and environ so the module can be imported on this platform.
+
+    class winreg:
+        HKEY_USERS = None
+        HKEY_CURRENT_USER = None
+        HKEY_LOCAL_MACHINE = None
+        HKEY_CLASSES_ROOT = None
+
+    environ: dict[str, str] = dict()
+
+
+class PlatformInfo:
+    """
+    Current and Target Architectures information.
+
+    Parameters
+    ----------
+    arch: str
+        Target architecture.
+    """
+
+    current_cpu = environ.get('processor_architecture', '').lower()
+
+    def __init__(self, arch) -> None:
+        self.arch = arch.lower().replace('x64', 'amd64')
+
+    @property
+    def target_cpu(self):
+        """
+        Return Target CPU architecture.
+
+        Return
+        ------
+        str
+            Target CPU
+        """
+        return self.arch[self.arch.find('_') + 1 :]
+
+    def target_is_x86(self):
+        """
+        Return True if target CPU is x86 32 bits..
+
+        Return
+        ------
+        bool
+            CPU is x86 32 bits
+        """
+        return self.target_cpu == 'x86'
+
+    def current_is_x86(self):
+        """
+        Return True if current CPU is x86 32 bits..
+
+        Return
+        ------
+        bool
+            CPU is x86 32 bits
+        """
+        return self.current_cpu == 'x86'
+
+    def current_dir(self, hidex86=False, x64=False) -> str:
+        """
+        Current platform specific subfolder.
+
+        Parameters
+        ----------
+        hidex86: bool
+            return '' and not '\x86' if architecture is x86.
+        x64: bool
+            return '\x64' and not '\amd64' if architecture is amd64.
+
+        Return
+        ------
+        str
+            subfolder: '\target', or '' (see hidex86 parameter)
+        """
+        return (
+            ''
+            if (self.current_cpu == 'x86' and hidex86)
+            else r'\x64'
+            if (self.current_cpu == 'amd64' and x64)
+            else rf'\{self.current_cpu}'
+        )
+
+    def target_dir(self, hidex86=False, x64=False) -> str:
+        r"""
+        Target platform specific subfolder.
+
+        Parameters
+        ----------
+        hidex86: bool
+            return '' and not '\x86' if architecture is x86.
+        x64: bool
+            return '\x64' and not '\amd64' if architecture is amd64.
+
+        Return
+        ------
+        str
+            subfolder: '\current', or '' (see hidex86 parameter)
+        """
+        return (
+            ''
+            if (self.target_cpu == 'x86' and hidex86)
+            else r'\x64'
+            if (self.target_cpu == 'amd64' and x64)
+            else rf'\{self.target_cpu}'
+        )
+
+    def cross_dir(self, forcex86=False):
+        r"""
+        Cross platform specific subfolder.
+
+        Parameters
+        ----------
+        forcex86: bool
+            Use 'x86' as current architecture even if current architecture is
+            not x86.
+
+        Return
+        ------
+        str
+            subfolder: '' if target architecture is current architecture,
+            '\current_target' if not.
+        """
+        current = 'x86' if forcex86 else self.current_cpu
+        return (
+            ''
+            if self.target_cpu == current
+            else self.target_dir().replace('\\', f'\\{current}_')
+        )
+
+
+class RegistryInfo:
+    """
+    Microsoft Visual Studio related registry information.
+
+    Parameters
+    ----------
+    platform_info: PlatformInfo
+        "PlatformInfo" instance.
+    """
+
+    HKEYS = (
+        winreg.HKEY_USERS,
+        winreg.HKEY_CURRENT_USER,
+        winreg.HKEY_LOCAL_MACHINE,
+        winreg.HKEY_CLASSES_ROOT,
+    )
+
+    def __init__(self, platform_info) -> None:
+        self.pi = platform_info
+
+    @property
+    def visualstudio(self) -> str:
+        """
+        Microsoft Visual Studio root registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return 'VisualStudio'
+
+    @property
+    def sxs(self):
+        """
+        Microsoft Visual Studio SxS registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return os.path.join(self.visualstudio, 'SxS')
+
+    @property
+    def vc(self):
+        """
+        Microsoft Visual C++ VC7 registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return os.path.join(self.sxs, 'VC7')
+
+    @property
+    def vs(self):
+        """
+        Microsoft Visual Studio VS7 registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return os.path.join(self.sxs, 'VS7')
+
+    @property
+    def vc_for_python(self) -> str:
+        """
+        Microsoft Visual C++ for Python registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return r'DevDiv\VCForPython'
+
+    @property
+    def microsoft_sdk(self) -> str:
+        """
+        Microsoft SDK registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return 'Microsoft SDKs'
+
+    @property
+    def windows_sdk(self):
+        """
+        Microsoft Windows/Platform SDK registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return os.path.join(self.microsoft_sdk, 'Windows')
+
+    @property
+    def netfx_sdk(self):
+        """
+        Microsoft .NET Framework SDK registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return os.path.join(self.microsoft_sdk, 'NETFXSDK')
+
+    @property
+    def windows_kits_roots(self) -> str:
+        """
+        Microsoft Windows Kits Roots registry key.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        return r'Windows Kits\Installed Roots'
+
+    def microsoft(self, key, x86=False):
+        """
+        Return key in Microsoft software registry.
+
+        Parameters
+        ----------
+        key: str
+            Registry key path where look.
+        x86: str
+            Force x86 software registry.
+
+        Return
+        ------
+        str
+            Registry key
+        """
+        node64 = '' if self.pi.current_is_x86() or x86 else 'Wow6432Node'
+        return os.path.join('Software', node64, 'Microsoft', key)
+
+    def lookup(self, key, name):
+        """
+        Look for values in registry in Microsoft software registry.
+
+        Parameters
+        ----------
+        key: str
+            Registry key path where look.
+        name: str
+            Value name to find.
+
+        Return
+        ------
+        str
+            value
+        """
+        key_read = winreg.KEY_READ
+        openkey = winreg.OpenKey
+        closekey = winreg.CloseKey
+        ms = self.microsoft
+        for hkey in self.HKEYS:
+            bkey = None
+            try:
+                bkey = openkey(hkey, ms(key), 0, key_read)
+            except OSError:
+                if not self.pi.current_is_x86():
+                    try:
+                        bkey = openkey(hkey, ms(key, True), 0, key_read)
+                    except OSError:
+                        continue
+                else:
+                    continue
+            try:
+                return winreg.QueryValueEx(bkey, name)[0]
+            except OSError:
+                pass
+            finally:
+                if bkey:
+                    closekey(bkey)
+        return None
+
+
+class SystemInfo:
+    """
+    Microsoft Windows and Visual Studio related system information.
+
+    Parameters
+    ----------
+    registry_info: RegistryInfo
+        "RegistryInfo" instance.
+    vc_ver: float
+        Required Microsoft Visual C++ version.
+    """
+
+    # Variables and properties in this class use originals CamelCase variables
+    # names from Microsoft source files for more easy comparison.
+    WinDir = environ.get('WinDir', '')
+    ProgramFiles = environ.get('ProgramFiles', '')
+    ProgramFilesx86 = environ.get('ProgramFiles(x86)', ProgramFiles)
+
+    def __init__(self, registry_info, vc_ver=None) -> None:
+        self.ri = registry_info
+        self.pi = self.ri.pi
+
+        self.known_vs_paths = self.find_programdata_vs_vers()
+
+        # Except for VS15+, VC version is aligned with VS version
+        self.vs_ver = self.vc_ver = vc_ver or self._find_latest_available_vs_ver()
+
+    def _find_latest_available_vs_ver(self):
+        """
+        Find the latest VC version
+
+        Return
+        ------
+        float
+            version
+        """
+        reg_vc_vers = self.find_reg_vs_vers()
+
+        if not (reg_vc_vers or self.known_vs_paths):
+            raise distutils.errors.DistutilsPlatformError(
+                'No Microsoft Visual C++ version found'
+            )
+
+        vc_vers = set(reg_vc_vers)
+        vc_vers.update(self.known_vs_paths)
+        return sorted(vc_vers)[-1]
+
+    def find_reg_vs_vers(self):
+        """
+        Find Microsoft Visual Studio versions available in registry.
+
+        Return
+        ------
+        list of float
+            Versions
+        """
+        ms = self.ri.microsoft
+        vckeys = (self.ri.vc, self.ri.vc_for_python, self.ri.vs)
+        vs_vers = []
+        for hkey, key in itertools.product(self.ri.HKEYS, vckeys):
+            try:
+                bkey = winreg.OpenKey(hkey, ms(key), 0, winreg.KEY_READ)
+            except OSError:
+                continue
+            with bkey:
+                subkeys, values, _ = winreg.QueryInfoKey(bkey)
+                for i in range(values):
+                    with contextlib.suppress(ValueError):
+                        ver = float(winreg.EnumValue(bkey, i)[0])
+                        if ver not in vs_vers:
+                            vs_vers.append(ver)
+                for i in range(subkeys):
+                    with contextlib.suppress(ValueError):
+                        ver = float(winreg.EnumKey(bkey, i))
+                        if ver not in vs_vers:
+                            vs_vers.append(ver)
+        return sorted(vs_vers)
+
+    def find_programdata_vs_vers(self) -> dict[float, str]:
+        r"""
+        Find Visual studio 2017+ versions from information in
+        "C:\ProgramData\Microsoft\VisualStudio\Packages\_Instances".
+
+        Return
+        ------
+        dict
+            float version as key, path as value.
+        """
+        vs_versions: dict[float, str] = {}
+        instances_dir = r'C:\ProgramData\Microsoft\VisualStudio\Packages\_Instances'
+
+        try:
+            hashed_names = os.listdir(instances_dir)
+
+        except OSError:
+            # Directory not exists with all Visual Studio versions
+            return vs_versions
+
+        for name in hashed_names:
+            try:
+                # Get VS installation path from "state.json" file
+                state_path = os.path.join(instances_dir, name, 'state.json')
+                with open(state_path, 'rt', encoding='utf-8') as state_file:
+                    state = json.load(state_file)
+                vs_path = state['installationPath']
+
+                # Raises OSError if this VS installation does not contain VC
+                os.listdir(os.path.join(vs_path, r'VC\Tools\MSVC'))
+
+                # Store version and path
+                vs_versions[self._as_float_version(state['installationVersion'])] = (
+                    vs_path
+                )
+
+            except (OSError, KeyError):
+                # Skip if "state.json" file is missing or bad format
+                continue
+
+        return vs_versions
+
+    @staticmethod
+    def _as_float_version(version):
+        """
+        Return a string version as a simplified float version (major.minor)
+
+        Parameters
+        ----------
+        version: str
+            Version.
+
+        Return
+        ------
+        float
+            version
+        """
+        return float('.'.join(version.split('.')[:2]))
+
+    @property
+    def VSInstallDir(self):
+        """
+        Microsoft Visual Studio directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        # Default path
+        default = os.path.join(
+            self.ProgramFilesx86, f'Microsoft Visual Studio {self.vs_ver:0.1f}'
+        )
+
+        # Try to get path from registry, if fail use default path
+        return self.ri.lookup(self.ri.vs, f'{self.vs_ver:0.1f}') or default
+
+    @property
+    def VCInstallDir(self):
+        """
+        Microsoft Visual C++ directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        path = self._guess_vc() or self._guess_vc_legacy()
+
+        if not os.path.isdir(path):
+            msg = 'Microsoft Visual C++ directory not found'
+            raise distutils.errors.DistutilsPlatformError(msg)
+
+        return path
+
+    def _guess_vc(self):
+        """
+        Locate Visual C++ for VS2017+.
+
+        Return
+        ------
+        str
+            path
+        """
+        if self.vs_ver <= 14.0:
+            return ''
+
+        try:
+            # First search in known VS paths
+            vs_dir = self.known_vs_paths[self.vs_ver]
+        except KeyError:
+            # Else, search with path from registry
+            vs_dir = self.VSInstallDir
+
+        guess_vc = os.path.join(vs_dir, r'VC\Tools\MSVC')
+
+        # Subdir with VC exact version as name
+        try:
+            # Update the VC version with real one instead of VS version
+            vc_ver = os.listdir(guess_vc)[-1]
+            self.vc_ver = self._as_float_version(vc_ver)
+            return os.path.join(guess_vc, vc_ver)
+        except (OSError, IndexError):
+            return ''
+
+    def _guess_vc_legacy(self):
+        """
+        Locate Visual C++ for versions prior to 2017.
+
+        Return
+        ------
+        str
+            path
+        """
+        default = os.path.join(
+            self.ProgramFilesx86,
+            rf'Microsoft Visual Studio {self.vs_ver:0.1f}\VC',
+        )
+
+        # Try to get "VC++ for Python" path from registry as default path
+        reg_path = os.path.join(self.ri.vc_for_python, f'{self.vs_ver:0.1f}')
+        python_vc = self.ri.lookup(reg_path, 'installdir')
+        default_vc = os.path.join(python_vc, 'VC') if python_vc else default
+
+        # Try to get path from registry, if fail use default path
+        return self.ri.lookup(self.ri.vc, f'{self.vs_ver:0.1f}') or default_vc
+
+    @property
+    def WindowsSdkVersion(self) -> tuple[LiteralString, ...]:
+        """
+        Microsoft Windows SDK versions for specified MSVC++ version.
+
+        Return
+        ------
+        tuple of str
+            versions
+        """
+        if self.vs_ver <= 9.0:
+            return '7.0', '6.1', '6.0a'
+        elif self.vs_ver == 10.0:
+            return '7.1', '7.0a'
+        elif self.vs_ver == 11.0:
+            return '8.0', '8.0a'
+        elif self.vs_ver == 12.0:
+            return '8.1', '8.1a'
+        elif self.vs_ver >= 14.0:
+            return '10.0', '8.1'
+        return ()
+
+    @property
+    def WindowsSdkLastVersion(self):
+        """
+        Microsoft Windows SDK last version.
+
+        Return
+        ------
+        str
+            version
+        """
+        return self._use_last_dir_name(os.path.join(self.WindowsSdkDir, 'lib'))
+
+    @property
+    def WindowsSdkDir(self) -> str | None:  # noqa: C901  # is too complex (12)  # FIXME
+        """
+        Microsoft Windows SDK directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        sdkdir: str | None = ''
+        for ver in self.WindowsSdkVersion:
+            # Try to get it from registry
+            loc = os.path.join(self.ri.windows_sdk, f'v{ver}')
+            sdkdir = self.ri.lookup(loc, 'installationfolder')
+            if sdkdir:
+                break
+        if not sdkdir or not os.path.isdir(sdkdir):
+            # Try to get "VC++ for Python" version from registry
+            path = os.path.join(self.ri.vc_for_python, f'{self.vc_ver:0.1f}')
+            install_base = self.ri.lookup(path, 'installdir')
+            if install_base:
+                sdkdir = os.path.join(install_base, 'WinSDK')
+        if not sdkdir or not os.path.isdir(sdkdir):
+            # If fail, use default new path
+            for ver in self.WindowsSdkVersion:
+                intver = ver[: ver.rfind('.')]
+                path = rf'Microsoft SDKs\Windows Kits\{intver}'
+                d = os.path.join(self.ProgramFiles, path)
+                if os.path.isdir(d):
+                    sdkdir = d
+        if not sdkdir or not os.path.isdir(sdkdir):
+            # If fail, use default old path
+            for ver in self.WindowsSdkVersion:
+                path = rf'Microsoft SDKs\Windows\v{ver}'
+                d = os.path.join(self.ProgramFiles, path)
+                if os.path.isdir(d):
+                    sdkdir = d
+        if not sdkdir:
+            # If fail, use Platform SDK
+            sdkdir = os.path.join(self.VCInstallDir, 'PlatformSDK')
+        return sdkdir
+
+    @property
+    def WindowsSDKExecutablePath(self):
+        """
+        Microsoft Windows SDK executable directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        # Find WinSDK NetFx Tools registry dir name
+        if self.vs_ver <= 11.0:
+            netfxver = 35
+            arch = ''
+        else:
+            netfxver = 40
+            hidex86 = True if self.vs_ver <= 12.0 else False
+            arch = self.pi.current_dir(x64=True, hidex86=hidex86).replace('\\', '-')
+        fx = f'WinSDK-NetFx{netfxver}Tools{arch}'
+
+        # list all possibles registry paths
+        regpaths = []
+        if self.vs_ver >= 14.0:
+            for ver in self.NetFxSdkVersion:
+                regpaths += [os.path.join(self.ri.netfx_sdk, ver, fx)]
+
+        for ver in self.WindowsSdkVersion:
+            regpaths += [os.path.join(self.ri.windows_sdk, f'v{ver}A', fx)]
+
+        # Return installation folder from the more recent path
+        for path in regpaths:
+            execpath = self.ri.lookup(path, 'installationfolder')
+            if execpath:
+                return execpath
+
+        return None
+
+    @property
+    def FSharpInstallDir(self):
+        """
+        Microsoft Visual F# directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        path = os.path.join(self.ri.visualstudio, rf'{self.vs_ver:0.1f}\Setup\F#')
+        return self.ri.lookup(path, 'productdir') or ''
+
+    @property
+    def UniversalCRTSdkDir(self):
+        """
+        Microsoft Universal CRT SDK directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        # Set Kit Roots versions for specified MSVC++ version
+        vers = ('10', '81') if self.vs_ver >= 14.0 else ()
+
+        # Find path of the more recent Kit
+        for ver in vers:
+            sdkdir = self.ri.lookup(self.ri.windows_kits_roots, f'kitsroot{ver}')
+            if sdkdir:
+                return sdkdir or ''
+
+        return None
+
+    @property
+    def UniversalCRTSdkLastVersion(self):
+        """
+        Microsoft Universal C Runtime SDK last version.
+
+        Return
+        ------
+        str
+            version
+        """
+        return self._use_last_dir_name(os.path.join(self.UniversalCRTSdkDir, 'lib'))
+
+    @property
+    def NetFxSdkVersion(self):
+        """
+        Microsoft .NET Framework SDK versions.
+
+        Return
+        ------
+        tuple of str
+            versions
+        """
+        # Set FxSdk versions for specified VS version
+        return (
+            ('4.7.2', '4.7.1', '4.7', '4.6.2', '4.6.1', '4.6', '4.5.2', '4.5.1', '4.5')
+            if self.vs_ver >= 14.0
+            else ()
+        )
+
+    @property
+    def NetFxSdkDir(self):
+        """
+        Microsoft .NET Framework SDK directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        sdkdir = ''
+        for ver in self.NetFxSdkVersion:
+            loc = os.path.join(self.ri.netfx_sdk, ver)
+            sdkdir = self.ri.lookup(loc, 'kitsinstallationfolder')
+            if sdkdir:
+                break
+        return sdkdir
+
+    @property
+    def FrameworkDir32(self):
+        """
+        Microsoft .NET Framework 32bit directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        # Default path
+        guess_fw = os.path.join(self.WinDir, r'Microsoft.NET\Framework')
+
+        # Try to get path from registry, if fail use default path
+        return self.ri.lookup(self.ri.vc, 'frameworkdir32') or guess_fw
+
+    @property
+    def FrameworkDir64(self):
+        """
+        Microsoft .NET Framework 64bit directory.
+
+        Return
+        ------
+        str
+            path
+        """
+        # Default path
+        guess_fw = os.path.join(self.WinDir, r'Microsoft.NET\Framework64')
+
+        # Try to get path from registry, if fail use default path
+        return self.ri.lookup(self.ri.vc, 'frameworkdir64') or guess_fw
+
+    @property
+    def FrameworkVersion32(self) -> tuple[str, ...]:
+        """
+        Microsoft .NET Framework 32bit versions.
+
+        Return
+        ------
+        tuple of str
+            versions
+        """
+        return self._find_dot_net_versions(32)
+
+    @property
+    def FrameworkVersion64(self) -> tuple[str, ...]:
+        """
+        Microsoft .NET Framework 64bit versions.
+
+        Return
+        ------
+        tuple of str
+            versions
+        """
+        return self._find_dot_net_versions(64)
+
+    def _find_dot_net_versions(self, bits) -> tuple[str, ...]:
+        """
+        Find Microsoft .NET Framework versions.
+
+        Parameters
+        ----------
+        bits: int
+            Platform number of bits: 32 or 64.
+
+        Return
+        ------
+        tuple of str
+            versions
+        """
+        # Find actual .NET version in registry
+        reg_ver = self.ri.lookup(self.ri.vc, f'frameworkver{bits}')
+        dot_net_dir = getattr(self, f'FrameworkDir{bits}')
+        ver = reg_ver or self._use_last_dir_name(dot_net_dir, 'v') or ''
+
+        # Set .NET versions for specified MSVC++ version
+        if self.vs_ver >= 12.0:
+            return ver, 'v4.0'
+        elif self.vs_ver >= 10.0:
+            return 'v4.0.30319' if ver.lower()[:2] != 'v4' else ver, 'v3.5'
+        elif self.vs_ver == 9.0:
+            return 'v3.5', 'v2.0.50727'
+        elif self.vs_ver == 8.0:
+            return 'v3.0', 'v2.0.50727'
+        return ()
+
+    @staticmethod
+    def _use_last_dir_name(path, prefix=''):
+        """
+        Return name of the last dir in path or '' if no dir found.
+
+        Parameters
+        ----------
+        path: str
+            Use dirs in this path
+        prefix: str
+            Use only dirs starting by this prefix
+
+        Return
+        ------
+        str
+            name
+        """
+        matching_dirs = (
+            dir_name
+            for dir_name in reversed(os.listdir(path))
+            if os.path.isdir(os.path.join(path, dir_name))
+            and dir_name.startswith(prefix)
+        )
+        return next(matching_dirs, None) or ''
+
+
+class _EnvironmentDict(TypedDict):
+    include: str
+    lib: str
+    libpath: str
+    path: str
+    py_vcruntime_redist: NotRequired[str | None]
+
+
+class EnvironmentInfo:
+    """
+    Return environment variables for specified Microsoft Visual C++ version
+    and platform : Lib, Include, Path and libpath.
+
+    This function is compatible with Microsoft Visual C++ 9.0 to 14.X.
+
+    Script created by analysing Microsoft environment configuration files like
+    "vcvars[...].bat", "SetEnv.Cmd", "vcbuildtools.bat", ...
+
+    Parameters
+    ----------
+    arch: str
+        Target architecture.
+    vc_ver: float
+        Required Microsoft Visual C++ version. If not set, autodetect the last
+        version.
+    vc_min_ver: float
+        Minimum Microsoft Visual C++ version.
+    """
+
+    # Variables and properties in this class use originals CamelCase variables
+    # names from Microsoft source files for more easy comparison.
+
+    def __init__(self, arch, vc_ver=None, vc_min_ver=0) -> None:
+        self.pi = PlatformInfo(arch)
+        self.ri = RegistryInfo(self.pi)
+        self.si = SystemInfo(self.ri, vc_ver)
+
+        if self.vc_ver < vc_min_ver:
+            err = 'No suitable Microsoft Visual C++ version found'
+            raise distutils.errors.DistutilsPlatformError(err)
+
+    @property
+    def vs_ver(self):
+        """
+        Microsoft Visual Studio.
+
+        Return
+        ------
+        float
+            version
+        """
+        return self.si.vs_ver
+
+    @property
+    def vc_ver(self):
+        """
+        Microsoft Visual C++ version.
+
+        Return
+        ------
+        float
+            version
+        """
+        return self.si.vc_ver
+
+    @property
+    def VSTools(self):
+        """
+        Microsoft Visual Studio Tools.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        paths = [r'Common7\IDE', r'Common7\Tools']
+
+        if self.vs_ver >= 14.0:
+            arch_subdir = self.pi.current_dir(hidex86=True, x64=True)
+            paths += [r'Common7\IDE\CommonExtensions\Microsoft\TestWindow']
+            paths += [r'Team Tools\Performance Tools']
+            paths += [rf'Team Tools\Performance Tools{arch_subdir}']
+
+        return [os.path.join(self.si.VSInstallDir, path) for path in paths]
+
+    @property
+    def VCIncludes(self):
+        """
+        Microsoft Visual C++ & Microsoft Foundation Class Includes.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        return [
+            os.path.join(self.si.VCInstallDir, 'Include'),
+            os.path.join(self.si.VCInstallDir, r'ATLMFC\Include'),
+        ]
+
+    @property
+    def VCLibraries(self):
+        """
+        Microsoft Visual C++ & Microsoft Foundation Class Libraries.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver >= 15.0:
+            arch_subdir = self.pi.target_dir(x64=True)
+        else:
+            arch_subdir = self.pi.target_dir(hidex86=True)
+        paths = [f'Lib{arch_subdir}', rf'ATLMFC\Lib{arch_subdir}']
+
+        if self.vs_ver >= 14.0:
+            paths += [rf'Lib\store{arch_subdir}']
+
+        return [os.path.join(self.si.VCInstallDir, path) for path in paths]
+
+    @property
+    def VCStoreRefs(self):
+        """
+        Microsoft Visual C++ store references Libraries.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 14.0:
+            return []
+        return [os.path.join(self.si.VCInstallDir, r'Lib\store\references')]
+
+    @property
+    def VCTools(self):
+        """
+        Microsoft Visual C++ Tools.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        si = self.si
+        tools = [os.path.join(si.VCInstallDir, 'VCPackages')]
+
+        forcex86 = True if self.vs_ver <= 10.0 else False
+        arch_subdir = self.pi.cross_dir(forcex86)
+        if arch_subdir:
+            tools += [os.path.join(si.VCInstallDir, f'Bin{arch_subdir}')]
+
+        if self.vs_ver == 14.0:
+            path = f'Bin{self.pi.current_dir(hidex86=True)}'
+            tools += [os.path.join(si.VCInstallDir, path)]
+
+        elif self.vs_ver >= 15.0:
+            host_dir = (
+                r'bin\HostX86%s' if self.pi.current_is_x86() else r'bin\HostX64%s'
+            )
+            tools += [
+                os.path.join(si.VCInstallDir, host_dir % self.pi.target_dir(x64=True))
+            ]
+
+            if self.pi.current_cpu != self.pi.target_cpu:
+                tools += [
+                    os.path.join(
+                        si.VCInstallDir, host_dir % self.pi.current_dir(x64=True)
+                    )
+                ]
+
+        else:
+            tools += [os.path.join(si.VCInstallDir, 'Bin')]
+
+        return tools
+
+    @property
+    def OSLibraries(self):
+        """
+        Microsoft Windows SDK Libraries.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver <= 10.0:
+            arch_subdir = self.pi.target_dir(hidex86=True, x64=True)
+            return [os.path.join(self.si.WindowsSdkDir, f'Lib{arch_subdir}')]
+
+        else:
+            arch_subdir = self.pi.target_dir(x64=True)
+            lib = os.path.join(self.si.WindowsSdkDir, 'lib')
+            libver = self._sdk_subdir
+            return [os.path.join(lib, f'{libver}um{arch_subdir}')]
+
+    @property
+    def OSIncludes(self):
+        """
+        Microsoft Windows SDK Include.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        include = os.path.join(self.si.WindowsSdkDir, 'include')
+
+        if self.vs_ver <= 10.0:
+            return [include, os.path.join(include, 'gl')]
+
+        else:
+            if self.vs_ver >= 14.0:
+                sdkver = self._sdk_subdir
+            else:
+                sdkver = ''
+            return [
+                os.path.join(include, f'{sdkver}shared'),
+                os.path.join(include, f'{sdkver}um'),
+                os.path.join(include, f'{sdkver}winrt'),
+            ]
+
+    @property
+    def OSLibpath(self):
+        """
+        Microsoft Windows SDK Libraries Paths.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        ref = os.path.join(self.si.WindowsSdkDir, 'References')
+        libpath = []
+
+        if self.vs_ver <= 9.0:
+            libpath += self.OSLibraries
+
+        if self.vs_ver >= 11.0:
+            libpath += [os.path.join(ref, r'CommonConfiguration\Neutral')]
+
+        if self.vs_ver >= 14.0:
+            libpath += [
+                ref,
+                os.path.join(self.si.WindowsSdkDir, 'UnionMetadata'),
+                os.path.join(ref, 'Windows.Foundation.UniversalApiContract', '1.0.0.0'),
+                os.path.join(ref, 'Windows.Foundation.FoundationContract', '1.0.0.0'),
+                os.path.join(
+                    ref, 'Windows.Networking.Connectivity.WwanContract', '1.0.0.0'
+                ),
+                os.path.join(
+                    self.si.WindowsSdkDir,
+                    'ExtensionSDKs',
+                    'Microsoft.VCLibs',
+                    f'{self.vs_ver:0.1f}',
+                    'References',
+                    'CommonConfiguration',
+                    'neutral',
+                ),
+            ]
+        return libpath
+
+    @property
+    def SdkTools(self):
+        """
+        Microsoft Windows SDK Tools.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        return list(self._sdk_tools())
+
+    def _sdk_tools(self):
+        """
+        Microsoft Windows SDK Tools paths generator.
+
+        Return
+        ------
+        generator of str
+            paths
+        """
+        if self.vs_ver < 15.0:
+            bin_dir = 'Bin' if self.vs_ver <= 11.0 else r'Bin\x86'
+            yield os.path.join(self.si.WindowsSdkDir, bin_dir)
+
+        if not self.pi.current_is_x86():
+            arch_subdir = self.pi.current_dir(x64=True)
+            path = f'Bin{arch_subdir}'
+            yield os.path.join(self.si.WindowsSdkDir, path)
+
+        if self.vs_ver in (10.0, 11.0):
+            if self.pi.target_is_x86():
+                arch_subdir = ''
+            else:
+                arch_subdir = self.pi.current_dir(hidex86=True, x64=True)
+            path = rf'Bin\NETFX 4.0 Tools{arch_subdir}'
+            yield os.path.join(self.si.WindowsSdkDir, path)
+
+        elif self.vs_ver >= 15.0:
+            path = os.path.join(self.si.WindowsSdkDir, 'Bin')
+            arch_subdir = self.pi.current_dir(x64=True)
+            sdkver = self.si.WindowsSdkLastVersion
+            yield os.path.join(path, f'{sdkver}{arch_subdir}')
+
+        if self.si.WindowsSDKExecutablePath:
+            yield self.si.WindowsSDKExecutablePath
+
+    @property
+    def _sdk_subdir(self):
+        """
+        Microsoft Windows SDK version subdir.
+
+        Return
+        ------
+        str
+            subdir
+        """
+        ucrtver = self.si.WindowsSdkLastVersion
+        return (f'{ucrtver}\\') if ucrtver else ''
+
+    @property
+    def SdkSetup(self):
+        """
+        Microsoft Windows SDK Setup.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver > 9.0:
+            return []
+
+        return [os.path.join(self.si.WindowsSdkDir, 'Setup')]
+
+    @property
+    def FxTools(self):
+        """
+        Microsoft .NET Framework Tools.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        pi = self.pi
+        si = self.si
+
+        if self.vs_ver <= 10.0:
+            include32 = True
+            include64 = not pi.target_is_x86() and not pi.current_is_x86()
+        else:
+            include32 = pi.target_is_x86() or pi.current_is_x86()
+            include64 = pi.current_cpu == 'amd64' or pi.target_cpu == 'amd64'
+
+        tools = []
+        if include32:
+            tools += [
+                os.path.join(si.FrameworkDir32, ver) for ver in si.FrameworkVersion32
+            ]
+        if include64:
+            tools += [
+                os.path.join(si.FrameworkDir64, ver) for ver in si.FrameworkVersion64
+            ]
+        return tools
+
+    @property
+    def NetFxSDKLibraries(self):
+        """
+        Microsoft .Net Framework SDK Libraries.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 14.0 or not self.si.NetFxSdkDir:
+            return []
+
+        arch_subdir = self.pi.target_dir(x64=True)
+        return [os.path.join(self.si.NetFxSdkDir, rf'lib\um{arch_subdir}')]
+
+    @property
+    def NetFxSDKIncludes(self):
+        """
+        Microsoft .Net Framework SDK Includes.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 14.0 or not self.si.NetFxSdkDir:
+            return []
+
+        return [os.path.join(self.si.NetFxSdkDir, r'include\um')]
+
+    @property
+    def VsTDb(self):
+        """
+        Microsoft Visual Studio Team System Database.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        return [os.path.join(self.si.VSInstallDir, r'VSTSDB\Deploy')]
+
+    @property
+    def MSBuild(self):
+        """
+        Microsoft Build Engine.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 12.0:
+            return []
+        elif self.vs_ver < 15.0:
+            base_path = self.si.ProgramFilesx86
+            arch_subdir = self.pi.current_dir(hidex86=True)
+        else:
+            base_path = self.si.VSInstallDir
+            arch_subdir = ''
+
+        path = rf'MSBuild\{self.vs_ver:0.1f}\bin{arch_subdir}'
+        build = [os.path.join(base_path, path)]
+
+        if self.vs_ver >= 15.0:
+            # Add Roslyn C# & Visual Basic Compiler
+            build += [os.path.join(base_path, path, 'Roslyn')]
+
+        return build
+
+    @property
+    def HTMLHelpWorkshop(self):
+        """
+        Microsoft HTML Help Workshop.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 11.0:
+            return []
+
+        return [os.path.join(self.si.ProgramFilesx86, 'HTML Help Workshop')]
+
+    @property
+    def UCRTLibraries(self):
+        """
+        Microsoft Universal C Runtime SDK Libraries.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 14.0:
+            return []
+
+        arch_subdir = self.pi.target_dir(x64=True)
+        lib = os.path.join(self.si.UniversalCRTSdkDir, 'lib')
+        ucrtver = self._ucrt_subdir
+        return [os.path.join(lib, f'{ucrtver}ucrt{arch_subdir}')]
+
+    @property
+    def UCRTIncludes(self):
+        """
+        Microsoft Universal C Runtime SDK Include.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if self.vs_ver < 14.0:
+            return []
+
+        include = os.path.join(self.si.UniversalCRTSdkDir, 'include')
+        return [os.path.join(include, f'{self._ucrt_subdir}ucrt')]
+
+    @property
+    def _ucrt_subdir(self):
+        """
+        Microsoft Universal C Runtime SDK version subdir.
+
+        Return
+        ------
+        str
+            subdir
+        """
+        ucrtver = self.si.UniversalCRTSdkLastVersion
+        return (f'{ucrtver}\\') if ucrtver else ''
+
+    @property
+    def FSharp(self):
+        """
+        Microsoft Visual F#.
+
+        Return
+        ------
+        list of str
+            paths
+        """
+        if 11.0 > self.vs_ver > 12.0:
+            return []
+
+        return [self.si.FSharpInstallDir]
+
+    @property
+    def VCRuntimeRedist(self) -> str | None:
+        """
+        Microsoft Visual C++ runtime redistributable dll.
+
+        Returns the first suitable path found or None.
+        """
+        vcruntime = f'vcruntime{self.vc_ver}0.dll'
+        arch_subdir = self.pi.target_dir(x64=True).strip('\\')
+
+        # Installation prefixes candidates
+        prefixes = []
+        tools_path = self.si.VCInstallDir
+        redist_path = os.path.dirname(tools_path.replace(r'\Tools', r'\Redist'))
+        if os.path.isdir(redist_path):
+            # Redist version may not be exactly the same as tools
+            redist_path = os.path.join(redist_path, os.listdir(redist_path)[-1])
+            prefixes += [redist_path, os.path.join(redist_path, 'onecore')]
+
+        prefixes += [os.path.join(tools_path, 'redist')]  # VS14 legacy path
+
+        # CRT directory
+        crt_dirs = (
+            f'Microsoft.VC{self.vc_ver * 10}.CRT',
+            # Sometime store in directory with VS version instead of VC
+            f'Microsoft.VC{int(self.vs_ver) * 10}.CRT',
+        )
+
+        # vcruntime path
+        candidate_paths = (
+            os.path.join(prefix, arch_subdir, crt_dir, vcruntime)
+            for (prefix, crt_dir) in itertools.product(prefixes, crt_dirs)
+        )
+        return next(filter(os.path.isfile, candidate_paths), None)  # type: ignore[arg-type] #python/mypy#12682
+
+    def return_env(self, exists: bool = True) -> _EnvironmentDict:
+        """
+        Return environment dict.
+
+        Parameters
+        ----------
+        exists: bool
+            It True, only return existing paths.
+
+        Return
+        ------
+        dict
+            environment
+        """
+        env = _EnvironmentDict(
+            include=self._build_paths(
+                'include',
+                [
+                    self.VCIncludes,
+                    self.OSIncludes,
+                    self.UCRTIncludes,
+                    self.NetFxSDKIncludes,
+                ],
+                exists,
+            ),
+            lib=self._build_paths(
+                'lib',
+                [
+                    self.VCLibraries,
+                    self.OSLibraries,
+                    self.FxTools,
+                    self.UCRTLibraries,
+                    self.NetFxSDKLibraries,
+                ],
+                exists,
+            ),
+            libpath=self._build_paths(
+                'libpath',
+                [self.VCLibraries, self.FxTools, self.VCStoreRefs, self.OSLibpath],
+                exists,
+            ),
+            path=self._build_paths(
+                'path',
+                [
+                    self.VCTools,
+                    self.VSTools,
+                    self.VsTDb,
+                    self.SdkTools,
+                    self.SdkSetup,
+                    self.FxTools,
+                    self.MSBuild,
+                    self.HTMLHelpWorkshop,
+                    self.FSharp,
+                ],
+                exists,
+            ),
+        )
+        if self.vs_ver >= 14 and self.VCRuntimeRedist:
+            env['py_vcruntime_redist'] = self.VCRuntimeRedist
+        return env
+
+    def _build_paths(self, name, spec_path_lists, exists):
+        """
+        Given an environment variable name and specified paths,
+        return a pathsep-separated string of paths containing
+        unique, extant, directories from those paths and from
+        the environment variable. Raise an error if no paths
+        are resolved.
+
+        Parameters
+        ----------
+        name: str
+            Environment variable name
+        spec_path_lists: list of str
+            Paths
+        exists: bool
+            It True, only return existing paths.
+
+        Return
+        ------
+        str
+            Pathsep-separated paths
+        """
+        # flatten spec_path_lists
+        spec_paths = itertools.chain.from_iterable(spec_path_lists)
+        env_paths = environ.get(name, '').split(os.pathsep)
+        paths = itertools.chain(spec_paths, env_paths)
+        extant_paths = list(filter(os.path.isdir, paths)) if exists else paths
+        if not extant_paths:
+            msg = f"{name.upper()} environment variable is empty"
+            raise distutils.errors.DistutilsPlatformError(msg)
+        unique_paths = unique_everseen(extant_paths)
+        return os.pathsep.join(unique_paths)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/namespaces.py b/backend/venv/lib/python3.12/site-packages/setuptools/namespaces.py
new file mode 100644
index 00000000..85ea2ebd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/namespaces.py
@@ -0,0 +1,106 @@
+import itertools
+import os
+
+from .compat import py312
+
+from distutils import log
+
+flatten = itertools.chain.from_iterable
+
+
+class Installer:
+    nspkg_ext = '-nspkg.pth'
+
+    def install_namespaces(self) -> None:
+        nsp = self._get_all_ns_packages()
+        if not nsp:
+            return
+        filename = self._get_nspkg_file()
+        self.outputs.append(filename)
+        log.info("Installing %s", filename)
+        lines = map(self._gen_nspkg_line, nsp)
+
+        if self.dry_run:
+            # always generate the lines, even in dry run
+            list(lines)
+            return
+
+        with open(filename, 'wt', encoding=py312.PTH_ENCODING) as f:
+            # Python<3.13 requires encoding="locale" instead of "utf-8"
+            # See: python/cpython#77102
+            f.writelines(lines)
+
+    def uninstall_namespaces(self) -> None:
+        filename = self._get_nspkg_file()
+        if not os.path.exists(filename):
+            return
+        log.info("Removing %s", filename)
+        os.remove(filename)
+
+    def _get_nspkg_file(self):
+        filename, _ = os.path.splitext(self._get_target())
+        return filename + self.nspkg_ext
+
+    def _get_target(self):
+        return self.target
+
+    _nspkg_tmpl = (
+        "import sys, types, os",
+        "p = os.path.join(%(root)s, *%(pth)r)",
+        "importlib = __import__('importlib.util')",
+        "__import__('importlib.machinery')",
+        (
+            "m = "
+            "sys.modules.setdefault(%(pkg)r, "
+            "importlib.util.module_from_spec("
+            "importlib.machinery.PathFinder.find_spec(%(pkg)r, "
+            "[os.path.dirname(p)])))"
+        ),
+        ("m = m or sys.modules.setdefault(%(pkg)r, types.ModuleType(%(pkg)r))"),
+        "mp = (m or []) and m.__dict__.setdefault('__path__',[])",
+        "(p not in mp) and mp.append(p)",
+    )
+    "lines for the namespace installer"
+
+    _nspkg_tmpl_multi = ('m and setattr(sys.modules[%(parent)r], %(child)r, m)',)
+    "additional line(s) when a parent package is indicated"
+
+    def _get_root(self):
+        return "sys._getframe(1).f_locals['sitedir']"
+
+    def _gen_nspkg_line(self, pkg):
+        pth = tuple(pkg.split('.'))
+        root = self._get_root()
+        tmpl_lines = self._nspkg_tmpl
+        parent, sep, child = pkg.rpartition('.')
+        if parent:
+            tmpl_lines += self._nspkg_tmpl_multi
+        return ';'.join(tmpl_lines) % locals() + '\n'
+
+    def _get_all_ns_packages(self):
+        """Return sorted list of all package namespaces"""
+        pkgs = self.distribution.namespace_packages or []
+        return sorted(set(flatten(map(self._pkg_names, pkgs))))
+
+    @staticmethod
+    def _pkg_names(pkg):
+        """
+        Given a namespace package, yield the components of that
+        package.
+
+        >>> names = Installer._pkg_names('a.b.c')
+        >>> set(names) == set(['a', 'a.b', 'a.b.c'])
+        True
+        """
+        parts = pkg.split('.')
+        while parts:
+            yield '.'.join(parts)
+            parts.pop()
+
+
+class DevelopInstaller(Installer):
+    def _get_root(self):
+        return repr(str(self.egg_path))
+
+    def _get_target(self):
+        return self.egg_link
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/package_index.py b/backend/venv/lib/python3.12/site-packages/setuptools/package_index.py
new file mode 100644
index 00000000..1a6abebc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/package_index.py
@@ -0,0 +1,1137 @@
+"""PyPI and direct package downloading."""
+
+from __future__ import annotations
+
+import base64
+import configparser
+import hashlib
+import html
+import http.client
+import io
+import itertools
+import os
+import re
+import shutil
+import socket
+import subprocess
+import sys
+import urllib.error
+import urllib.parse
+import urllib.request
+from fnmatch import translate
+from functools import wraps
+from typing import NamedTuple
+
+from more_itertools import unique_everseen
+
+import setuptools
+from pkg_resources import (
+    BINARY_DIST,
+    CHECKOUT_DIST,
+    DEVELOP_DIST,
+    EGG_DIST,
+    SOURCE_DIST,
+    Distribution,
+    Environment,
+    Requirement,
+    find_distributions,
+    normalize_path,
+    parse_version,
+    safe_name,
+    safe_version,
+    to_filename,
+)
+from setuptools.wheel import Wheel
+
+from .unicode_utils import _cfg_read_utf8_with_fallback, _read_utf8_with_fallback
+
+from distutils import log
+from distutils.errors import DistutilsError
+
+EGG_FRAGMENT = re.compile(r'^egg=([-A-Za-z0-9_.+!]+)$')
+HREF = re.compile(r"""href\s*=\s*['"]?([^'"> ]+)""", re.I)
+PYPI_MD5 = re.compile(
+    r'<a href="([^"#]+)">([^<]+)</a>\n\s+\(<a (?:title="MD5 hash"\n\s+)'
+    r'href="[^?]+\?:action=show_md5&amp;digest=([0-9a-f]{32})">md5</a>\)'
+)
+URL_SCHEME = re.compile('([-+.a-z0-9]{2,}):', re.I).match
+EXTENSIONS = ".tar.gz .tar.bz2 .tar .zip .tgz".split()
+
+__all__ = [
+    'PackageIndex',
+    'distros_for_url',
+    'parse_bdist_wininst',
+    'interpret_distro_name',
+]
+
+_SOCKET_TIMEOUT = 15
+
+user_agent = f"setuptools/{setuptools.__version__} Python-urllib/{sys.version_info.major}.{sys.version_info.minor}"
+
+
+def parse_requirement_arg(spec):
+    try:
+        return Requirement.parse(spec)
+    except ValueError as e:
+        raise DistutilsError(
+            f"Not a URL, existing file, or requirement spec: {spec!r}"
+        ) from e
+
+
+def parse_bdist_wininst(name):
+    """Return (base,pyversion) or (None,None) for possible .exe name"""
+
+    lower = name.lower()
+    base, py_ver, plat = None, None, None
+
+    if lower.endswith('.exe'):
+        if lower.endswith('.win32.exe'):
+            base = name[:-10]
+            plat = 'win32'
+        elif lower.startswith('.win32-py', -16):
+            py_ver = name[-7:-4]
+            base = name[:-16]
+            plat = 'win32'
+        elif lower.endswith('.win-amd64.exe'):
+            base = name[:-14]
+            plat = 'win-amd64'
+        elif lower.startswith('.win-amd64-py', -20):
+            py_ver = name[-7:-4]
+            base = name[:-20]
+            plat = 'win-amd64'
+    return base, py_ver, plat
+
+
+def egg_info_for_url(url):
+    parts = urllib.parse.urlparse(url)
+    _scheme, server, path, _parameters, _query, fragment = parts
+    base = urllib.parse.unquote(path.split('/')[-1])
+    if server == 'sourceforge.net' and base == 'download':  # XXX Yuck
+        base = urllib.parse.unquote(path.split('/')[-2])
+    if '#' in base:
+        base, fragment = base.split('#', 1)
+    return base, fragment
+
+
+def distros_for_url(url, metadata=None):
+    """Yield egg or source distribution objects that might be found at a URL"""
+    base, fragment = egg_info_for_url(url)
+    yield from distros_for_location(url, base, metadata)
+    if fragment:
+        match = EGG_FRAGMENT.match(fragment)
+        if match:
+            yield from interpret_distro_name(
+                url, match.group(1), metadata, precedence=CHECKOUT_DIST
+            )
+
+
+def distros_for_location(location, basename, metadata=None):
+    """Yield egg or source distribution objects based on basename"""
+    if basename.endswith('.egg.zip'):
+        basename = basename[:-4]  # strip the .zip
+    if basename.endswith('.egg') and '-' in basename:
+        # only one, unambiguous interpretation
+        return [Distribution.from_location(location, basename, metadata)]
+    if basename.endswith('.whl') and '-' in basename:
+        wheel = Wheel(basename)
+        if not wheel.is_compatible():
+            return []
+        return [
+            Distribution(
+                location=location,
+                project_name=wheel.project_name,
+                version=wheel.version,
+                # Increase priority over eggs.
+                precedence=EGG_DIST + 1,
+            )
+        ]
+    if basename.endswith('.exe'):
+        win_base, py_ver, platform = parse_bdist_wininst(basename)
+        if win_base is not None:
+            return interpret_distro_name(
+                location, win_base, metadata, py_ver, BINARY_DIST, platform
+            )
+    # Try source distro extensions (.zip, .tgz, etc.)
+    #
+    for ext in EXTENSIONS:
+        if basename.endswith(ext):
+            basename = basename[: -len(ext)]
+            return interpret_distro_name(location, basename, metadata)
+    return []  # no extension matched
+
+
+def distros_for_filename(filename, metadata=None):
+    """Yield possible egg or source distribution objects based on a filename"""
+    return distros_for_location(
+        normalize_path(filename), os.path.basename(filename), metadata
+    )
+
+
+def interpret_distro_name(
+    location, basename, metadata, py_version=None, precedence=SOURCE_DIST, platform=None
+):
+    """Generate the interpretation of a source distro name
+
+    Note: if `location` is a filesystem filename, you should call
+    ``pkg_resources.normalize_path()`` on it before passing it to this
+    routine!
+    """
+
+    parts = basename.split('-')
+    if not py_version and any(re.match(r'py\d\.\d$', p) for p in parts[2:]):
+        # it is a bdist_dumb, not an sdist -- bail out
+        return
+
+    # find the pivot (p) that splits the name from the version.
+    # infer the version as the first item that has a digit.
+    for p in range(len(parts)):
+        if parts[p][:1].isdigit():
+            break
+    else:
+        p = len(parts)
+
+    yield Distribution(
+        location,
+        metadata,
+        '-'.join(parts[:p]),
+        '-'.join(parts[p:]),
+        py_version=py_version,
+        precedence=precedence,
+        platform=platform,
+    )
+
+
+def unique_values(func):
+    """
+    Wrap a function returning an iterable such that the resulting iterable
+    only ever yields unique items.
+    """
+
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+        return unique_everseen(func(*args, **kwargs))
+
+    return wrapper
+
+
+REL = re.compile(r"""<([^>]*\srel\s{0,10}=\s{0,10}['"]?([^'" >]+)[^>]*)>""", re.I)
+"""
+Regex for an HTML tag with 'rel="val"' attributes.
+"""
+
+
+@unique_values
+def find_external_links(url, page):
+    """Find rel="homepage" and rel="download" links in `page`, yielding URLs"""
+
+    for match in REL.finditer(page):
+        tag, rel = match.groups()
+        rels = set(map(str.strip, rel.lower().split(',')))
+        if 'homepage' in rels or 'download' in rels:
+            for match in HREF.finditer(tag):
+                yield urllib.parse.urljoin(url, htmldecode(match.group(1)))
+
+    for tag in ("<th>Home Page", "<th>Download URL"):
+        pos = page.find(tag)
+        if pos != -1:
+            match = HREF.search(page, pos)
+            if match:
+                yield urllib.parse.urljoin(url, htmldecode(match.group(1)))
+
+
+class ContentChecker:
+    """
+    A null content checker that defines the interface for checking content
+    """
+
+    def feed(self, block):
+        """
+        Feed a block of data to the hash.
+        """
+        return
+
+    def is_valid(self):
+        """
+        Check the hash. Return False if validation fails.
+        """
+        return True
+
+    def report(self, reporter, template):
+        """
+        Call reporter with information about the checker (hash name)
+        substituted into the template.
+        """
+        return
+
+
+class HashChecker(ContentChecker):
+    pattern = re.compile(
+        r'(?P<hash_name>sha1|sha224|sha384|sha256|sha512|md5)='
+        r'(?P<expected>[a-f0-9]+)'
+    )
+
+    def __init__(self, hash_name, expected) -> None:
+        self.hash_name = hash_name
+        self.hash = hashlib.new(hash_name)
+        self.expected = expected
+
+    @classmethod
+    def from_url(cls, url):
+        "Construct a (possibly null) ContentChecker from a URL"
+        fragment = urllib.parse.urlparse(url)[-1]
+        if not fragment:
+            return ContentChecker()
+        match = cls.pattern.search(fragment)
+        if not match:
+            return ContentChecker()
+        return cls(**match.groupdict())
+
+    def feed(self, block):
+        self.hash.update(block)
+
+    def is_valid(self):
+        return self.hash.hexdigest() == self.expected
+
+    def report(self, reporter, template):
+        msg = template % self.hash_name
+        return reporter(msg)
+
+
+class PackageIndex(Environment):
+    """A distribution index that scans web pages for download URLs"""
+
+    def __init__(
+        self,
+        index_url: str = "https://pypi.org/simple/",
+        hosts=('*',),
+        ca_bundle=None,
+        verify_ssl: bool = True,
+        *args,
+        **kw,
+    ) -> None:
+        super().__init__(*args, **kw)
+        self.index_url = index_url + "/"[: not index_url.endswith('/')]
+        self.scanned_urls: dict = {}
+        self.fetched_urls: dict = {}
+        self.package_pages: dict = {}
+        self.allows = re.compile('|'.join(map(translate, hosts))).match
+        self.to_scan: list = []
+        self.opener = urllib.request.urlopen
+
+    def add(self, dist):
+        # ignore invalid versions
+        try:
+            parse_version(dist.version)
+        except Exception:
+            return None
+        return super().add(dist)
+
+    # FIXME: 'PackageIndex.process_url' is too complex (14)
+    def process_url(self, url, retrieve: bool = False) -> None:  # noqa: C901
+        """Evaluate a URL as a possible download, and maybe retrieve it"""
+        if url in self.scanned_urls and not retrieve:
+            return
+        self.scanned_urls[url] = True
+        if not URL_SCHEME(url):
+            self.process_filename(url)
+            return
+        else:
+            dists = list(distros_for_url(url))
+            if dists:
+                if not self.url_ok(url):
+                    return
+                self.debug("Found link: %s", url)
+
+        if dists or not retrieve or url in self.fetched_urls:
+            list(map(self.add, dists))
+            return  # don't need the actual page
+
+        if not self.url_ok(url):
+            self.fetched_urls[url] = True
+            return
+
+        self.info("Reading %s", url)
+        self.fetched_urls[url] = True  # prevent multiple fetch attempts
+        tmpl = "Download error on %s: %%s -- Some packages may not be found!"
+        f = self.open_url(url, tmpl % url)
+        if f is None:
+            return
+        if isinstance(f, urllib.error.HTTPError) and f.code == 401:
+            self.info(f"Authentication error: {f.msg}")
+        self.fetched_urls[f.url] = True
+        if 'html' not in f.headers.get('content-type', '').lower():
+            f.close()  # not html, we can't process it
+            return
+
+        base = f.url  # handle redirects
+        page = f.read()
+        if not isinstance(page, str):
+            # In Python 3 and got bytes but want str.
+            if isinstance(f, urllib.error.HTTPError):
+                # Errors have no charset, assume latin1:
+                charset = 'latin-1'
+            else:
+                charset = f.headers.get_param('charset') or 'latin-1'
+            page = page.decode(charset, "ignore")
+        f.close()
+        for match in HREF.finditer(page):
+            link = urllib.parse.urljoin(base, htmldecode(match.group(1)))
+            self.process_url(link)
+        if url.startswith(self.index_url) and getattr(f, 'code', None) != 404:
+            page = self.process_index(url, page)
+
+    def process_filename(self, fn, nested: bool = False) -> None:
+        # process filenames or directories
+        if not os.path.exists(fn):
+            self.warn("Not found: %s", fn)
+            return
+
+        if os.path.isdir(fn) and not nested:
+            path = os.path.realpath(fn)
+            for item in os.listdir(path):
+                self.process_filename(os.path.join(path, item), True)
+
+        dists = distros_for_filename(fn)
+        if dists:
+            self.debug("Found: %s", fn)
+            list(map(self.add, dists))
+
+    def url_ok(self, url, fatal: bool = False) -> bool:
+        s = URL_SCHEME(url)
+        is_file = s and s.group(1).lower() == 'file'
+        if is_file or self.allows(urllib.parse.urlparse(url)[1]):
+            return True
+        msg = (
+            "\nNote: Bypassing %s (disallowed host; see "
+            "https://setuptools.pypa.io/en/latest/deprecated/"
+            "easy_install.html#restricting-downloads-with-allow-hosts for details).\n"
+        )
+        if fatal:
+            raise DistutilsError(msg % url)
+        else:
+            self.warn(msg, url)
+            return False
+
+    def scan_egg_links(self, search_path) -> None:
+        dirs = filter(os.path.isdir, search_path)
+        egg_links = (
+            (path, entry)
+            for path in dirs
+            for entry in os.listdir(path)
+            if entry.endswith('.egg-link')
+        )
+        list(itertools.starmap(self.scan_egg_link, egg_links))
+
+    def scan_egg_link(self, path, entry) -> None:
+        content = _read_utf8_with_fallback(os.path.join(path, entry))
+        # filter non-empty lines
+        lines = list(filter(None, map(str.strip, content.splitlines())))
+
+        if len(lines) != 2:
+            # format is not recognized; punt
+            return
+
+        egg_path, _setup_path = lines
+
+        for dist in find_distributions(os.path.join(path, egg_path)):
+            dist.location = os.path.join(path, *lines)
+            dist.precedence = SOURCE_DIST
+            self.add(dist)
+
+    def _scan(self, link):
+        # Process a URL to see if it's for a package page
+        NO_MATCH_SENTINEL = None, None
+        if not link.startswith(self.index_url):
+            return NO_MATCH_SENTINEL
+
+        parts = list(map(urllib.parse.unquote, link[len(self.index_url) :].split('/')))
+        if len(parts) != 2 or '#' in parts[1]:
+            return NO_MATCH_SENTINEL
+
+        # it's a package page, sanitize and index it
+        pkg = safe_name(parts[0])
+        ver = safe_version(parts[1])
+        self.package_pages.setdefault(pkg.lower(), {})[link] = True
+        return to_filename(pkg), to_filename(ver)
+
+    def process_index(self, url, page):
+        """Process the contents of a PyPI page"""
+
+        # process an index page into the package-page index
+        for match in HREF.finditer(page):
+            try:
+                self._scan(urllib.parse.urljoin(url, htmldecode(match.group(1))))
+            except ValueError:
+                pass
+
+        pkg, ver = self._scan(url)  # ensure this page is in the page index
+        if not pkg:
+            return ""  # no sense double-scanning non-package pages
+
+        # process individual package page
+        for new_url in find_external_links(url, page):
+            # Process the found URL
+            base, frag = egg_info_for_url(new_url)
+            if base.endswith('.py') and not frag:
+                if ver:
+                    new_url += f'#egg={pkg}-{ver}'
+                else:
+                    self.need_version_info(url)
+            self.scan_url(new_url)
+
+        return PYPI_MD5.sub(
+            lambda m: '<a href="{}#md5={}">{}</a>'.format(*m.group(1, 3, 2)), page
+        )
+
+    def need_version_info(self, url) -> None:
+        self.scan_all(
+            "Page at %s links to .py file(s) without version info; an index "
+            "scan is required.",
+            url,
+        )
+
+    def scan_all(self, msg=None, *args) -> None:
+        if self.index_url not in self.fetched_urls:
+            if msg:
+                self.warn(msg, *args)
+            self.info("Scanning index of all packages (this may take a while)")
+        self.scan_url(self.index_url)
+
+    def find_packages(self, requirement) -> None:
+        self.scan_url(self.index_url + requirement.unsafe_name + '/')
+
+        if not self.package_pages.get(requirement.key):
+            # Fall back to safe version of the name
+            self.scan_url(self.index_url + requirement.project_name + '/')
+
+        if not self.package_pages.get(requirement.key):
+            # We couldn't find the target package, so search the index page too
+            self.not_found_in_index(requirement)
+
+        for url in list(self.package_pages.get(requirement.key, ())):
+            # scan each page that might be related to the desired package
+            self.scan_url(url)
+
+    def obtain(self, requirement, installer=None):
+        self.prescan()
+        self.find_packages(requirement)
+        for dist in self[requirement.key]:
+            if dist in requirement:
+                return dist
+            self.debug("%s does not match %s", requirement, dist)
+        return super().obtain(requirement, installer)
+
+    def check_hash(self, checker, filename, tfp) -> None:
+        """
+        checker is a ContentChecker
+        """
+        checker.report(self.debug, f"Validating %s checksum for {filename}")
+        if not checker.is_valid():
+            tfp.close()
+            os.unlink(filename)
+            raise DistutilsError(
+                f"{checker.hash.name} validation failed for {os.path.basename(filename)}; "
+                "possible download problem?"
+            )
+
+    def add_find_links(self, urls) -> None:
+        """Add `urls` to the list that will be prescanned for searches"""
+        for url in urls:
+            if (
+                self.to_scan is None  # if we have already "gone online"
+                or not URL_SCHEME(url)  # or it's a local file/directory
+                or url.startswith('file:')
+                or list(distros_for_url(url))  # or a direct package link
+            ):
+                # then go ahead and process it now
+                self.scan_url(url)
+            else:
+                # otherwise, defer retrieval till later
+                self.to_scan.append(url)
+
+    def prescan(self):
+        """Scan urls scheduled for prescanning (e.g. --find-links)"""
+        if self.to_scan:
+            list(map(self.scan_url, self.to_scan))
+        self.to_scan = None  # from now on, go ahead and process immediately
+
+    def not_found_in_index(self, requirement) -> None:
+        if self[requirement.key]:  # we've seen at least one distro
+            meth, msg = self.info, "Couldn't retrieve index page for %r"
+        else:  # no distros seen for this name, might be misspelled
+            meth, msg = self.warn, "Couldn't find index page for %r (maybe misspelled?)"
+        meth(msg, requirement.unsafe_name)
+        self.scan_all()
+
+    def download(self, spec, tmpdir):
+        """Locate and/or download `spec` to `tmpdir`, returning a local path
+
+        `spec` may be a ``Requirement`` object, or a string containing a URL,
+        an existing local filename, or a project/version requirement spec
+        (i.e. the string form of a ``Requirement`` object).  If it is the URL
+        of a .py file with an unambiguous ``#egg=name-version`` tag (i.e., one
+        that escapes ``-`` as ``_`` throughout), a trivial ``setup.py`` is
+        automatically created alongside the downloaded file.
+
+        If `spec` is a ``Requirement`` object or a string containing a
+        project/version requirement spec, this method returns the location of
+        a matching distribution (possibly after downloading it to `tmpdir`).
+        If `spec` is a locally existing file or directory name, it is simply
+        returned unchanged.  If `spec` is a URL, it is downloaded to a subpath
+        of `tmpdir`, and the local filename is returned.  Various errors may be
+        raised if a problem occurs during downloading.
+        """
+        if not isinstance(spec, Requirement):
+            scheme = URL_SCHEME(spec)
+            if scheme:
+                # It's a url, download it to tmpdir
+                found = self._download_url(spec, tmpdir)
+                base, fragment = egg_info_for_url(spec)
+                if base.endswith('.py'):
+                    found = self.gen_setup(found, fragment, tmpdir)
+                return found
+            elif os.path.exists(spec):
+                # Existing file or directory, just return it
+                return spec
+            else:
+                spec = parse_requirement_arg(spec)
+        return getattr(self.fetch_distribution(spec, tmpdir), 'location', None)
+
+    def fetch_distribution(  # noqa: C901  # is too complex (14)  # FIXME
+        self,
+        requirement,
+        tmpdir,
+        force_scan: bool = False,
+        source: bool = False,
+        develop_ok: bool = False,
+        local_index=None,
+    ) -> Distribution | None:
+        """Obtain a distribution suitable for fulfilling `requirement`
+
+        `requirement` must be a ``pkg_resources.Requirement`` instance.
+        If necessary, or if the `force_scan` flag is set, the requirement is
+        searched for in the (online) package index as well as the locally
+        installed packages.  If a distribution matching `requirement` is found,
+        the returned distribution's ``location`` is the value you would have
+        gotten from calling the ``download()`` method with the matching
+        distribution's URL or filename.  If no matching distribution is found,
+        ``None`` is returned.
+
+        If the `source` flag is set, only source distributions and source
+        checkout links will be considered.  Unless the `develop_ok` flag is
+        set, development and system eggs (i.e., those using the ``.egg-info``
+        format) will be ignored.
+        """
+        # process a Requirement
+        self.info("Searching for %s", requirement)
+        skipped = set()
+        dist = None
+
+        def find(req, env: Environment | None = None):
+            if env is None:
+                env = self
+            # Find a matching distribution; may be called more than once
+
+            for dist in env[req.key]:
+                if dist.precedence == DEVELOP_DIST and not develop_ok:
+                    if dist not in skipped:
+                        self.warn(
+                            "Skipping development or system egg: %s",
+                            dist,
+                        )
+                        skipped.add(dist)
+                    continue
+
+                test = dist in req and (dist.precedence <= SOURCE_DIST or not source)
+                if test:
+                    loc = self.download(dist.location, tmpdir)
+                    dist.download_location = loc
+                    if os.path.exists(dist.download_location):
+                        return dist
+
+            return None
+
+        if force_scan:
+            self.prescan()
+            self.find_packages(requirement)
+            dist = find(requirement)
+
+        if not dist and local_index is not None:
+            dist = find(requirement, local_index)
+
+        if dist is None:
+            if self.to_scan is not None:
+                self.prescan()
+            dist = find(requirement)
+
+        if dist is None and not force_scan:
+            self.find_packages(requirement)
+            dist = find(requirement)
+
+        if dist is None:
+            self.warn(
+                "No local packages or working download links found for %s%s",
+                (source and "a source distribution of " or ""),
+                requirement,
+            )
+            return None
+        else:
+            self.info("Best match: %s", dist)
+            return dist.clone(location=dist.download_location)
+
+    def fetch(
+        self, requirement, tmpdir, force_scan: bool = False, source: bool = False
+    ) -> str | None:
+        """Obtain a file suitable for fulfilling `requirement`
+
+        DEPRECATED; use the ``fetch_distribution()`` method now instead.  For
+        backward compatibility, this routine is identical but returns the
+        ``location`` of the downloaded distribution instead of a distribution
+        object.
+        """
+        dist = self.fetch_distribution(requirement, tmpdir, force_scan, source)
+        if dist is not None:
+            return dist.location
+        return None
+
+    def gen_setup(self, filename, fragment, tmpdir):
+        match = EGG_FRAGMENT.match(fragment)
+        dists = (
+            match
+            and [
+                d
+                for d in interpret_distro_name(filename, match.group(1), None)
+                if d.version
+            ]
+            or []
+        )
+
+        if len(dists) == 1:  # unambiguous ``#egg`` fragment
+            basename = os.path.basename(filename)
+
+            # Make sure the file has been downloaded to the temp dir.
+            if os.path.dirname(filename) != tmpdir:
+                dst = os.path.join(tmpdir, basename)
+                if not (os.path.exists(dst) and os.path.samefile(filename, dst)):
+                    shutil.copy2(filename, dst)
+                    filename = dst
+
+            with open(os.path.join(tmpdir, 'setup.py'), 'w', encoding="utf-8") as file:
+                file.write(
+                    "from setuptools import setup\n"
+                    f"setup(name={dists[0].project_name!r}, version={dists[0].version!r}, py_modules=[{os.path.splitext(basename)[0]!r}])\n"
+                )
+            return filename
+
+        elif match:
+            raise DistutilsError(
+                f"Can't unambiguously interpret project/version identifier {fragment!r}; "
+                "any dashes in the name or version should be escaped using "
+                f"underscores. {dists!r}"
+            )
+        else:
+            raise DistutilsError(
+                "Can't process plain .py files without an '#egg=name-version'"
+                " suffix to enable automatic setup script generation."
+            )
+
+    dl_blocksize = 8192
+
+    def _download_to(self, url, filename):
+        self.info("Downloading %s", url)
+        # Download the file
+        fp = None
+        try:
+            checker = HashChecker.from_url(url)
+            fp = self.open_url(url)
+            if isinstance(fp, urllib.error.HTTPError):
+                raise DistutilsError(f"Can't download {url}: {fp.code} {fp.msg}")
+            headers = fp.info()
+            blocknum = 0
+            bs = self.dl_blocksize
+            size = -1
+            if "content-length" in headers:
+                # Some servers return multiple Content-Length headers :(
+                sizes = headers.get_all('Content-Length')
+                size = max(map(int, sizes))
+                self.reporthook(url, filename, blocknum, bs, size)
+            with open(filename, 'wb') as tfp:
+                while True:
+                    block = fp.read(bs)
+                    if block:
+                        checker.feed(block)
+                        tfp.write(block)
+                        blocknum += 1
+                        self.reporthook(url, filename, blocknum, bs, size)
+                    else:
+                        break
+                self.check_hash(checker, filename, tfp)
+            return headers
+        finally:
+            if fp:
+                fp.close()
+
+    def reporthook(self, url, filename, blocknum, blksize, size) -> None:
+        pass  # no-op
+
+    # FIXME:
+    def open_url(self, url, warning=None):  # noqa: C901  # is too complex (12)
+        if url.startswith('file:'):
+            return local_open(url)
+        try:
+            return open_with_auth(url, self.opener)
+        except (ValueError, http.client.InvalidURL) as v:
+            msg = ' '.join([str(arg) for arg in v.args])
+            if warning:
+                self.warn(warning, msg)
+            else:
+                raise DistutilsError(f'{url} {msg}') from v
+        except urllib.error.HTTPError as v:
+            return v
+        except urllib.error.URLError as v:
+            if warning:
+                self.warn(warning, v.reason)
+            else:
+                raise DistutilsError(f"Download error for {url}: {v.reason}") from v
+        except http.client.BadStatusLine as v:
+            if warning:
+                self.warn(warning, v.line)
+            else:
+                raise DistutilsError(
+                    f'{url} returned a bad status line. The server might be '
+                    f'down, {v.line}'
+                ) from v
+        except (http.client.HTTPException, OSError) as v:
+            if warning:
+                self.warn(warning, v)
+            else:
+                raise DistutilsError(f"Download error for {url}: {v}") from v
+
+    def _download_url(self, url, tmpdir):
+        # Determine download filename
+        #
+        name, _fragment = egg_info_for_url(url)
+        if name:
+            while '..' in name:
+                name = name.replace('..', '.').replace('\\', '_')
+        else:
+            name = "__downloaded__"  # default if URL has no path contents
+
+        if name.endswith('.egg.zip'):
+            name = name[:-4]  # strip the extra .zip before download
+
+        filename = os.path.join(tmpdir, name)
+
+        return self._download_vcs(url, filename) or self._download_other(url, filename)
+
+    @staticmethod
+    def _resolve_vcs(url):
+        """
+        >>> rvcs = PackageIndex._resolve_vcs
+        >>> rvcs('git+http://foo/bar')
+        'git'
+        >>> rvcs('hg+https://foo/bar')
+        'hg'
+        >>> rvcs('git:myhost')
+        'git'
+        >>> rvcs('hg:myhost')
+        >>> rvcs('http://foo/bar')
+        """
+        scheme = urllib.parse.urlsplit(url).scheme
+        pre, sep, _post = scheme.partition('+')
+        # svn and git have their own protocol; hg does not
+        allowed = set(['svn', 'git'] + ['hg'] * bool(sep))
+        return next(iter({pre} & allowed), None)
+
+    def _download_vcs(self, url, spec_filename):
+        vcs = self._resolve_vcs(url)
+        if not vcs:
+            return None
+        if vcs == 'svn':
+            raise DistutilsError(
+                f"Invalid config, SVN download is not supported: {url}"
+            )
+
+        filename, _, _ = spec_filename.partition('#')
+        url, rev = self._vcs_split_rev_from_url(url)
+
+        self.info(f"Doing {vcs} clone from {url} to {filename}")
+        subprocess.check_call([vcs, 'clone', '--quiet', url, filename])
+
+        co_commands = dict(
+            git=[vcs, '-C', filename, 'checkout', '--quiet', rev],
+            hg=[vcs, '--cwd', filename, 'up', '-C', '-r', rev, '-q'],
+        )
+        if rev is not None:
+            self.info(f"Checking out {rev}")
+            subprocess.check_call(co_commands[vcs])
+
+        return filename
+
+    def _download_other(self, url, filename):
+        scheme = urllib.parse.urlsplit(url).scheme
+        if scheme == 'file':  # pragma: no cover
+            return urllib.request.url2pathname(urllib.parse.urlparse(url).path)
+        # raise error if not allowed
+        self.url_ok(url, True)
+        return self._attempt_download(url, filename)
+
+    def scan_url(self, url) -> None:
+        self.process_url(url, True)
+
+    def _attempt_download(self, url, filename):
+        headers = self._download_to(url, filename)
+        if 'html' in headers.get('content-type', '').lower():
+            return self._invalid_download_html(url, headers, filename)
+        else:
+            return filename
+
+    def _invalid_download_html(self, url, headers, filename):
+        os.unlink(filename)
+        raise DistutilsError(f"Unexpected HTML page found at {url}")
+
+    @staticmethod
+    def _vcs_split_rev_from_url(url):
+        """
+        Given a possible VCS URL, return a clean URL and resolved revision if any.
+
+        >>> vsrfu = PackageIndex._vcs_split_rev_from_url
+        >>> vsrfu('git+https://github.com/pypa/setuptools@v69.0.0#egg-info=setuptools')
+        ('https://github.com/pypa/setuptools', 'v69.0.0')
+        >>> vsrfu('git+https://github.com/pypa/setuptools#egg-info=setuptools')
+        ('https://github.com/pypa/setuptools', None)
+        >>> vsrfu('http://foo/bar')
+        ('http://foo/bar', None)
+        """
+        parts = urllib.parse.urlsplit(url)
+
+        clean_scheme = parts.scheme.split('+', 1)[-1]
+
+        # Some fragment identification fails
+        no_fragment_path, _, _ = parts.path.partition('#')
+
+        pre, sep, post = no_fragment_path.rpartition('@')
+        clean_path, rev = (pre, post) if sep else (post, None)
+
+        resolved = parts._replace(
+            scheme=clean_scheme,
+            path=clean_path,
+            # discard the fragment
+            fragment='',
+        ).geturl()
+
+        return resolved, rev
+
+    def debug(self, msg, *args) -> None:
+        log.debug(msg, *args)
+
+    def info(self, msg, *args) -> None:
+        log.info(msg, *args)
+
+    def warn(self, msg, *args) -> None:
+        log.warn(msg, *args)
+
+
+# This pattern matches a character entity reference (a decimal numeric
+# references, a hexadecimal numeric reference, or a named reference).
+entity_sub = re.compile(r'&(#(\d+|x[\da-fA-F]+)|[\w.:-]+);?').sub
+
+
+def decode_entity(match):
+    what = match.group(0)
+    return html.unescape(what)
+
+
+def htmldecode(text):
+    """
+    Decode HTML entities in the given text.
+
+    >>> htmldecode(
+    ...     'https://../package_name-0.1.2.tar.gz'
+    ...     '?tokena=A&amp;tokenb=B">package_name-0.1.2.tar.gz')
+    'https://../package_name-0.1.2.tar.gz?tokena=A&tokenb=B">package_name-0.1.2.tar.gz'
+    """
+    return entity_sub(decode_entity, text)
+
+
+def socket_timeout(timeout=15):
+    def _socket_timeout(func):
+        def _socket_timeout(*args, **kwargs):
+            old_timeout = socket.getdefaulttimeout()
+            socket.setdefaulttimeout(timeout)
+            try:
+                return func(*args, **kwargs)
+            finally:
+                socket.setdefaulttimeout(old_timeout)
+
+        return _socket_timeout
+
+    return _socket_timeout
+
+
+def _encode_auth(auth):
+    """
+    Encode auth from a URL suitable for an HTTP header.
+    >>> str(_encode_auth('username%3Apassword'))
+    'dXNlcm5hbWU6cGFzc3dvcmQ='
+
+    Long auth strings should not cause a newline to be inserted.
+    >>> long_auth = 'username:' + 'password'*10
+    >>> chr(10) in str(_encode_auth(long_auth))
+    False
+    """
+    auth_s = urllib.parse.unquote(auth)
+    # convert to bytes
+    auth_bytes = auth_s.encode()
+    encoded_bytes = base64.b64encode(auth_bytes)
+    # convert back to a string
+    encoded = encoded_bytes.decode()
+    # strip the trailing carriage return
+    return encoded.replace('\n', '')
+
+
+class Credential(NamedTuple):
+    """
+    A username/password pair.
+
+    Displayed separated by `:`.
+    >>> str(Credential('username', 'password'))
+    'username:password'
+    """
+
+    username: str
+    password: str
+
+    def __str__(self) -> str:
+        return f'{self.username}:{self.password}'
+
+
+class PyPIConfig(configparser.RawConfigParser):
+    def __init__(self):
+        """
+        Load from ~/.pypirc
+        """
+        defaults = dict.fromkeys(['username', 'password', 'repository'], '')
+        super().__init__(defaults)
+
+        rc = os.path.join(os.path.expanduser('~'), '.pypirc')
+        if os.path.exists(rc):
+            _cfg_read_utf8_with_fallback(self, rc)
+
+    @property
+    def creds_by_repository(self):
+        sections_with_repositories = [
+            section
+            for section in self.sections()
+            if self.get(section, 'repository').strip()
+        ]
+
+        return dict(map(self._get_repo_cred, sections_with_repositories))
+
+    def _get_repo_cred(self, section):
+        repo = self.get(section, 'repository').strip()
+        return repo, Credential(
+            self.get(section, 'username').strip(),
+            self.get(section, 'password').strip(),
+        )
+
+    def find_credential(self, url):
+        """
+        If the URL indicated appears to be a repository defined in this
+        config, return the credential for that repository.
+        """
+        for repository, cred in self.creds_by_repository.items():
+            if url.startswith(repository):
+                return cred
+        return None
+
+
+def open_with_auth(url, opener=urllib.request.urlopen):
+    """Open a urllib2 request, handling HTTP authentication"""
+
+    parsed = urllib.parse.urlparse(url)
+    scheme, netloc, path, params, query, frag = parsed
+
+    # Double scheme does not raise on macOS as revealed by a
+    # failing test. We would expect "nonnumeric port". Refs #20.
+    if netloc.endswith(':'):
+        raise http.client.InvalidURL("nonnumeric port: ''")
+
+    if scheme in ('http', 'https'):
+        auth, address = _splituser(netloc)
+    else:
+        auth, address = (None, None)
+
+    if not auth:
+        cred = PyPIConfig().find_credential(url)
+        if cred:
+            auth = str(cred)
+            info = cred.username, url
+            log.info('Authenticating as %s for %s (from .pypirc)', *info)
+
+    if auth:
+        auth = "Basic " + _encode_auth(auth)
+        parts = scheme, address, path, params, query, frag
+        new_url = urllib.parse.urlunparse(parts)
+        request = urllib.request.Request(new_url)
+        request.add_header("Authorization", auth)
+    else:
+        request = urllib.request.Request(url)
+
+    request.add_header('User-Agent', user_agent)
+    fp = opener(request)
+
+    if auth:
+        # Put authentication info back into request URL if same host,
+        # so that links found on the page will work
+        s2, h2, path2, param2, query2, frag2 = urllib.parse.urlparse(fp.url)
+        if s2 == scheme and h2 == address:
+            parts = s2, netloc, path2, param2, query2, frag2
+            fp.url = urllib.parse.urlunparse(parts)
+
+    return fp
+
+
+# copy of urllib.parse._splituser from Python 3.8
+# See https://github.com/python/cpython/issues/80072.
+def _splituser(host):
+    """splituser('user[:passwd]@host[:port]')
+    --> 'user[:passwd]', 'host[:port]'."""
+    user, delim, host = host.rpartition('@')
+    return (user if delim else None), host
+
+
+# adding a timeout to avoid freezing package_index
+open_with_auth = socket_timeout(_SOCKET_TIMEOUT)(open_with_auth)
+
+
+def fix_sf_url(url):
+    return url  # backward compatibility
+
+
+def local_open(url):
+    """Read a local path, with special support for directories"""
+    _scheme, _server, path, _param, _query, _frag = urllib.parse.urlparse(url)
+    filename = urllib.request.url2pathname(path)
+    if os.path.isfile(filename):
+        return urllib.request.urlopen(url)
+    elif path.endswith('/') and os.path.isdir(filename):
+        files = []
+        for f in os.listdir(filename):
+            filepath = os.path.join(filename, f)
+            if f == 'index.html':
+                body = _read_utf8_with_fallback(filepath)
+                break
+            elif os.path.isdir(filepath):
+                f += '/'
+            files.append(f'<a href="{f}">{f}</a>')
+        else:
+            tmpl = "<html><head><title>{url}</title></head><body>{files}</body></html>"
+            body = tmpl.format(url=url, files='\n'.join(files))
+        status, message = 200, "OK"
+    else:
+        status, message, body = 404, "Path not found", "Not found"
+
+    headers = {'content-type': 'text/html'}
+    body_stream = io.StringIO(body)
+    return urllib.error.HTTPError(url, status, message, headers, body_stream)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/sandbox.py b/backend/venv/lib/python3.12/site-packages/setuptools/sandbox.py
new file mode 100644
index 00000000..2d84242d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/sandbox.py
@@ -0,0 +1,536 @@
+from __future__ import annotations
+
+import builtins
+import contextlib
+import functools
+import itertools
+import operator
+import os
+import pickle
+import re
+import sys
+import tempfile
+import textwrap
+from types import TracebackType
+from typing import TYPE_CHECKING, Any, ClassVar
+
+import pkg_resources
+from pkg_resources import working_set
+
+from distutils.errors import DistutilsError
+
+if TYPE_CHECKING:
+    import os as _os
+elif sys.platform.startswith('java'):
+    import org.python.modules.posix.PosixModule as _os  # pyright: ignore[reportMissingImports]
+else:
+    _os = sys.modules[os.name]
+_open = open
+
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+__all__ = [
+    "AbstractSandbox",
+    "DirectorySandbox",
+    "SandboxViolation",
+    "run_setup",
+]
+
+
+def _execfile(filename, globals, locals=None):
+    """
+    Python 3 implementation of execfile.
+    """
+    mode = 'rb'
+    with open(filename, mode) as stream:
+        script = stream.read()
+    if locals is None:
+        locals = globals
+    code = compile(script, filename, 'exec')
+    exec(code, globals, locals)
+
+
+@contextlib.contextmanager
+def save_argv(repl=None):
+    saved = sys.argv[:]
+    if repl is not None:
+        sys.argv[:] = repl
+    try:
+        yield saved
+    finally:
+        sys.argv[:] = saved
+
+
+@contextlib.contextmanager
+def save_path():
+    saved = sys.path[:]
+    try:
+        yield saved
+    finally:
+        sys.path[:] = saved
+
+
+@contextlib.contextmanager
+def override_temp(replacement):
+    """
+    Monkey-patch tempfile.tempdir with replacement, ensuring it exists
+    """
+    os.makedirs(replacement, exist_ok=True)
+
+    saved = tempfile.tempdir
+
+    tempfile.tempdir = replacement
+
+    try:
+        yield
+    finally:
+        tempfile.tempdir = saved
+
+
+@contextlib.contextmanager
+def pushd(target):
+    saved = os.getcwd()
+    os.chdir(target)
+    try:
+        yield saved
+    finally:
+        os.chdir(saved)
+
+
+class UnpickleableException(Exception):
+    """
+    An exception representing another Exception that could not be pickled.
+    """
+
+    @staticmethod
+    def dump(type, exc):
+        """
+        Always return a dumped (pickled) type and exc. If exc can't be pickled,
+        wrap it in UnpickleableException first.
+        """
+        try:
+            return pickle.dumps(type), pickle.dumps(exc)
+        except Exception:
+            # get UnpickleableException inside the sandbox
+            from setuptools.sandbox import UnpickleableException as cls
+
+            return cls.dump(cls, cls(repr(exc)))
+
+
+class ExceptionSaver:
+    """
+    A Context Manager that will save an exception, serialize, and restore it
+    later.
+    """
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        type: type[BaseException] | None,
+        exc: BaseException | None,
+        tb: TracebackType | None,
+    ) -> bool:
+        if not exc:
+            return False
+
+        # dump the exception
+        self._saved = UnpickleableException.dump(type, exc)
+        self._tb = tb
+
+        # suppress the exception
+        return True
+
+    def resume(self):
+        "restore and re-raise any exception"
+
+        if '_saved' not in vars(self):
+            return
+
+        _type, exc = map(pickle.loads, self._saved)
+        raise exc.with_traceback(self._tb)
+
+
+@contextlib.contextmanager
+def save_modules():
+    """
+    Context in which imported modules are saved.
+
+    Translates exceptions internal to the context into the equivalent exception
+    outside the context.
+    """
+    saved = sys.modules.copy()
+    with ExceptionSaver() as saved_exc:
+        yield saved
+
+    sys.modules.update(saved)
+    # remove any modules imported since
+    del_modules = (
+        mod_name
+        for mod_name in sys.modules
+        if mod_name not in saved
+        # exclude any encodings modules. See #285
+        and not mod_name.startswith('encodings.')
+    )
+    _clear_modules(del_modules)
+
+    saved_exc.resume()
+
+
+def _clear_modules(module_names):
+    for mod_name in list(module_names):
+        del sys.modules[mod_name]
+
+
+@contextlib.contextmanager
+def save_pkg_resources_state():
+    saved = pkg_resources.__getstate__()
+    try:
+        yield saved
+    finally:
+        pkg_resources.__setstate__(saved)
+
+
+@contextlib.contextmanager
+def setup_context(setup_dir):
+    temp_dir = os.path.join(setup_dir, 'temp')
+    with save_pkg_resources_state():
+        with save_modules():
+            with save_path():
+                hide_setuptools()
+                with save_argv():
+                    with override_temp(temp_dir):
+                        with pushd(setup_dir):
+                            # ensure setuptools commands are available
+                            __import__('setuptools')
+                            yield
+
+
+_MODULES_TO_HIDE = {
+    'setuptools',
+    'distutils',
+    'pkg_resources',
+    'Cython',
+    '_distutils_hack',
+}
+
+
+def _needs_hiding(mod_name):
+    """
+    >>> _needs_hiding('setuptools')
+    True
+    >>> _needs_hiding('pkg_resources')
+    True
+    >>> _needs_hiding('setuptools_plugin')
+    False
+    >>> _needs_hiding('setuptools.__init__')
+    True
+    >>> _needs_hiding('distutils')
+    True
+    >>> _needs_hiding('os')
+    False
+    >>> _needs_hiding('Cython')
+    True
+    """
+    base_module = mod_name.split('.', 1)[0]
+    return base_module in _MODULES_TO_HIDE
+
+
+def hide_setuptools():
+    """
+    Remove references to setuptools' modules from sys.modules to allow the
+    invocation to import the most appropriate setuptools. This technique is
+    necessary to avoid issues such as #315 where setuptools upgrading itself
+    would fail to find a function declared in the metadata.
+    """
+    _distutils_hack = sys.modules.get('_distutils_hack', None)
+    if _distutils_hack is not None:
+        _distutils_hack._remove_shim()
+
+    modules = filter(_needs_hiding, sys.modules)
+    _clear_modules(modules)
+
+
+def run_setup(setup_script, args):
+    """Run a distutils setup script, sandboxed in its directory"""
+    setup_dir = os.path.abspath(os.path.dirname(setup_script))
+    with setup_context(setup_dir):
+        try:
+            sys.argv[:] = [setup_script] + list(args)
+            sys.path.insert(0, setup_dir)
+            # reset to include setup dir, w/clean callback list
+            working_set.__init__()
+            working_set.callbacks.append(lambda dist: dist.activate())
+
+            with DirectorySandbox(setup_dir):
+                ns = dict(__file__=setup_script, __name__='__main__')
+                _execfile(setup_script, ns)
+        except SystemExit as v:
+            if v.args and v.args[0]:
+                raise
+            # Normal exit, just return
+
+
+class AbstractSandbox:
+    """Wrap 'os' module and 'open()' builtin for virtualizing setup scripts"""
+
+    _active = False
+
+    def __init__(self) -> None:
+        self._attrs = [
+            name
+            for name in dir(_os)
+            if not name.startswith('_') and hasattr(self, name)
+        ]
+
+    def _copy(self, source):
+        for name in self._attrs:
+            setattr(os, name, getattr(source, name))
+
+    def __enter__(self) -> None:
+        self._copy(self)
+        builtins.open = self._open
+        self._active = True
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_value: BaseException | None,
+        traceback: TracebackType | None,
+    ):
+        self._active = False
+        builtins.open = _open
+        self._copy(_os)
+
+    def run(self, func):
+        """Run 'func' under os sandboxing"""
+        with self:
+            return func()
+
+    def _mk_dual_path_wrapper(name: str):  # type: ignore[misc] # https://github.com/pypa/setuptools/pull/4099
+        original = getattr(_os, name)
+
+        def wrap(self, src, dst, *args, **kw):
+            if self._active:
+                src, dst = self._remap_pair(name, src, dst, *args, **kw)
+            return original(src, dst, *args, **kw)
+
+        return wrap
+
+    for __name in ["rename", "link", "symlink"]:
+        if hasattr(_os, __name):
+            locals()[__name] = _mk_dual_path_wrapper(__name)
+
+    def _mk_single_path_wrapper(name: str, original=None):  # type: ignore[misc] # https://github.com/pypa/setuptools/pull/4099
+        original = original or getattr(_os, name)
+
+        def wrap(self, path, *args, **kw):
+            if self._active:
+                path = self._remap_input(name, path, *args, **kw)
+            return original(path, *args, **kw)
+
+        return wrap
+
+    _open = _mk_single_path_wrapper('open', _open)
+    for __name in [
+        "stat",
+        "listdir",
+        "chdir",
+        "open",
+        "chmod",
+        "chown",
+        "mkdir",
+        "remove",
+        "unlink",
+        "rmdir",
+        "utime",
+        "lchown",
+        "chroot",
+        "lstat",
+        "startfile",
+        "mkfifo",
+        "mknod",
+        "pathconf",
+        "access",
+    ]:
+        if hasattr(_os, __name):
+            locals()[__name] = _mk_single_path_wrapper(__name)
+
+    def _mk_single_with_return(name: str):  # type: ignore[misc] # https://github.com/pypa/setuptools/pull/4099
+        original = getattr(_os, name)
+
+        def wrap(self, path, *args, **kw):
+            if self._active:
+                path = self._remap_input(name, path, *args, **kw)
+                return self._remap_output(name, original(path, *args, **kw))
+            return original(path, *args, **kw)
+
+        return wrap
+
+    for __name in ['readlink', 'tempnam']:
+        if hasattr(_os, __name):
+            locals()[__name] = _mk_single_with_return(__name)
+
+    def _mk_query(name: str):  # type: ignore[misc] # https://github.com/pypa/setuptools/pull/4099
+        original = getattr(_os, name)
+
+        def wrap(self, *args, **kw):
+            retval = original(*args, **kw)
+            if self._active:
+                return self._remap_output(name, retval)
+            return retval
+
+        return wrap
+
+    for __name in ['getcwd', 'tmpnam']:
+        if hasattr(_os, __name):
+            locals()[__name] = _mk_query(__name)
+
+    def _validate_path(self, path):
+        """Called to remap or validate any path, whether input or output"""
+        return path
+
+    def _remap_input(self, operation, path, *args, **kw):
+        """Called for path inputs"""
+        return self._validate_path(path)
+
+    def _remap_output(self, operation, path):
+        """Called for path outputs"""
+        return self._validate_path(path)
+
+    def _remap_pair(self, operation, src, dst, *args, **kw):
+        """Called for path pairs like rename, link, and symlink operations"""
+        return (
+            self._remap_input(operation + '-from', src, *args, **kw),
+            self._remap_input(operation + '-to', dst, *args, **kw),
+        )
+
+    if TYPE_CHECKING:
+        # This is a catch-all for all the dynamically created attributes.
+        # This isn't public API anyway
+        def __getattribute__(self, name: str) -> Any: ...
+
+
+if hasattr(os, 'devnull'):
+    _EXCEPTIONS = [os.devnull]
+else:
+    _EXCEPTIONS = []
+
+
+class DirectorySandbox(AbstractSandbox):
+    """Restrict operations to a single subdirectory - pseudo-chroot"""
+
+    write_ops: ClassVar[dict[str, None]] = dict.fromkeys([
+        "open",
+        "chmod",
+        "chown",
+        "mkdir",
+        "remove",
+        "unlink",
+        "rmdir",
+        "utime",
+        "lchown",
+        "chroot",
+        "mkfifo",
+        "mknod",
+        "tempnam",
+    ])
+
+    _exception_patterns: list[str | re.Pattern] = []
+    "exempt writing to paths that match the pattern"
+
+    def __init__(self, sandbox, exceptions=_EXCEPTIONS) -> None:
+        self._sandbox = os.path.normcase(os.path.realpath(sandbox))
+        self._prefix = os.path.join(self._sandbox, '')
+        self._exceptions = [
+            os.path.normcase(os.path.realpath(path)) for path in exceptions
+        ]
+        AbstractSandbox.__init__(self)
+
+    def _violation(self, operation, *args, **kw):
+        from setuptools.sandbox import SandboxViolation
+
+        raise SandboxViolation(operation, args, kw)
+
+    def _open(self, path, mode='r', *args, **kw):
+        if mode not in ('r', 'rt', 'rb', 'rU', 'U') and not self._ok(path):
+            self._violation("open", path, mode, *args, **kw)
+        return _open(path, mode, *args, **kw)
+
+    def tmpnam(self) -> None:
+        self._violation("tmpnam")
+
+    def _ok(self, path):
+        active = self._active
+        try:
+            self._active = False
+            realpath = os.path.normcase(os.path.realpath(path))
+            return (
+                self._exempted(realpath)
+                or realpath == self._sandbox
+                or realpath.startswith(self._prefix)
+            )
+        finally:
+            self._active = active
+
+    def _exempted(self, filepath):
+        start_matches = (
+            filepath.startswith(exception) for exception in self._exceptions
+        )
+        pattern_matches = (
+            re.match(pattern, filepath) for pattern in self._exception_patterns
+        )
+        candidates = itertools.chain(start_matches, pattern_matches)
+        return any(candidates)
+
+    def _remap_input(self, operation, path, *args, **kw):
+        """Called for path inputs"""
+        if operation in self.write_ops and not self._ok(path):
+            self._violation(operation, os.path.realpath(path), *args, **kw)
+        return path
+
+    def _remap_pair(self, operation, src, dst, *args, **kw):
+        """Called for path pairs like rename, link, and symlink operations"""
+        if not self._ok(src) or not self._ok(dst):
+            self._violation(operation, src, dst, *args, **kw)
+        return (src, dst)
+
+    def open(self, file, flags, mode: int = 0o777, *args, **kw) -> int:
+        """Called for low-level os.open()"""
+        if flags & WRITE_FLAGS and not self._ok(file):
+            self._violation("os.open", file, flags, mode, *args, **kw)
+        return _os.open(file, flags, mode, *args, **kw)
+
+
+WRITE_FLAGS = functools.reduce(
+    operator.or_,
+    [
+        getattr(_os, a, 0)
+        for a in "O_WRONLY O_RDWR O_APPEND O_CREAT O_TRUNC O_TEMPORARY".split()
+    ],
+)
+
+
+class SandboxViolation(DistutilsError):
+    """A setup script attempted to modify the filesystem outside the sandbox"""
+
+    tmpl = textwrap.dedent(
+        """
+        SandboxViolation: {cmd}{args!r} {kwargs}
+
+        The package setup script has attempted to modify files on your system
+        that are not within the EasyInstall build area, and has been aborted.
+
+        This package cannot be safely installed by EasyInstall, and may not
+        support alternate installation locations even if you run its setup
+        script by hand.  Please inform the package's author and the EasyInstall
+        maintainers to find out if a fix or workaround is available.
+        """
+    ).lstrip()
+
+    def __str__(self) -> str:
+        cmd, args, kwargs = self.args
+        return self.tmpl.format(**locals())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/script (dev).tmpl b/backend/venv/lib/python3.12/site-packages/setuptools/script (dev).tmpl
new file mode 100644
index 00000000..39a24b04
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/script (dev).tmpl	
@@ -0,0 +1,6 @@
+# EASY-INSTALL-DEV-SCRIPT: %(spec)r,%(script_name)r
+__requires__ = %(spec)r
+__import__('pkg_resources').require(%(spec)r)
+__file__ = %(dev_path)r
+with open(__file__) as f:
+    exec(compile(f.read(), __file__, 'exec'))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/script.tmpl b/backend/venv/lib/python3.12/site-packages/setuptools/script.tmpl
new file mode 100644
index 00000000..ff5efbca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/script.tmpl
@@ -0,0 +1,3 @@
+# EASY-INSTALL-SCRIPT: %(spec)r,%(script_name)r
+__requires__ = %(spec)r
+__import__('pkg_resources').run_script(%(spec)r, %(script_name)r)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__init__.py
new file mode 100644
index 00000000..eb70bfb7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__init__.py
@@ -0,0 +1,13 @@
+import locale
+import sys
+
+import pytest
+
+__all__ = ['fail_on_ascii']
+
+if sys.version_info >= (3, 11):
+    locale_encoding = locale.getencoding()
+else:
+    locale_encoding = locale.getpreferredencoding(False)
+is_ascii = locale_encoding == 'ANSI_X3.4-1968'
+fail_on_ascii = pytest.mark.xfail(is_ascii, reason="Test fails in this locale")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2458f8ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/contexts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/contexts.cpython-312.pyc
new file mode 100644
index 00000000..575e6af7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/contexts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/environment.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/environment.cpython-312.pyc
new file mode 100644
index 00000000..6eaca8e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/environment.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/fixtures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/fixtures.cpython-312.pyc
new file mode 100644
index 00000000..790985d8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/fixtures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/mod_with_constant.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/mod_with_constant.cpython-312.pyc
new file mode 100644
index 00000000..74ff2a2d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/mod_with_constant.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/namespaces.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/namespaces.cpython-312.pyc
new file mode 100644
index 00000000..aa46e36f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/namespaces.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/script-with-bom.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/script-with-bom.cpython-312.pyc
new file mode 100644
index 00000000..a99f8822
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/script-with-bom.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/server.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/server.cpython-312.pyc
new file mode 100644
index 00000000..8479d970
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/server.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_archive_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_archive_util.cpython-312.pyc
new file mode 100644
index 00000000..c26d1d84
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_archive_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_deprecations.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_deprecations.cpython-312.pyc
new file mode 100644
index 00000000..90c74141
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_deprecations.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_egg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_egg.cpython-312.pyc
new file mode 100644
index 00000000..93005e58
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_egg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_wheel.cpython-312.pyc
new file mode 100644
index 00000000..ce75e7bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_bdist_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build.cpython-312.pyc
new file mode 100644
index 00000000..6b024fcf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_clib.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_clib.cpython-312.pyc
new file mode 100644
index 00000000..7982a6bf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_clib.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_ext.cpython-312.pyc
new file mode 100644
index 00000000..026f5a28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_meta.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_meta.cpython-312.pyc
new file mode 100644
index 00000000..18ca4ce8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_meta.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_py.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_py.cpython-312.pyc
new file mode 100644
index 00000000..703205e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_build_py.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_config_discovery.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_config_discovery.cpython-312.pyc
new file mode 100644
index 00000000..adde1ba1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_config_discovery.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_core_metadata.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_core_metadata.cpython-312.pyc
new file mode 100644
index 00000000..971620be
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_core_metadata.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_depends.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_depends.cpython-312.pyc
new file mode 100644
index 00000000..770de71d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_depends.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_develop.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_develop.cpython-312.pyc
new file mode 100644
index 00000000..c77f1835
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_develop.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist.cpython-312.pyc
new file mode 100644
index 00000000..99f89de4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist_info.cpython-312.pyc
new file mode 100644
index 00000000..85113ff6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_dist_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_distutils_adoption.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_distutils_adoption.cpython-312.pyc
new file mode 100644
index 00000000..600a2cb3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_distutils_adoption.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_easy_install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_easy_install.cpython-312.pyc
new file mode 100644
index 00000000..e13ba66f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_easy_install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_editable_install.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_editable_install.cpython-312.pyc
new file mode 100644
index 00000000..736ac5d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_editable_install.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_egg_info.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_egg_info.cpython-312.pyc
new file mode 100644
index 00000000..9177628f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_egg_info.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_extern.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_extern.cpython-312.pyc
new file mode 100644
index 00000000..f552294a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_extern.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_packages.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_packages.cpython-312.pyc
new file mode 100644
index 00000000..d7b1720b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_packages.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_py_modules.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_py_modules.cpython-312.pyc
new file mode 100644
index 00000000..b91c35cf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_find_py_modules.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_glob.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_glob.cpython-312.pyc
new file mode 100644
index 00000000..1bb0e85e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_glob.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_install_scripts.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_install_scripts.cpython-312.pyc
new file mode 100644
index 00000000..4a7666de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_install_scripts.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_logging.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_logging.cpython-312.pyc
new file mode 100644
index 00000000..61700a8c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_logging.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_manifest.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_manifest.cpython-312.pyc
new file mode 100644
index 00000000..ffc49db4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_manifest.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_namespaces.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_namespaces.cpython-312.pyc
new file mode 100644
index 00000000..54cf7128
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_namespaces.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_packageindex.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_packageindex.cpython-312.pyc
new file mode 100644
index 00000000..d1a3dd9f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_packageindex.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sandbox.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sandbox.cpython-312.pyc
new file mode 100644
index 00000000..88c777bf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sandbox.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sdist.cpython-312.pyc
new file mode 100644
index 00000000..9e62f0dd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setopt.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setopt.cpython-312.pyc
new file mode 100644
index 00000000..c98dfbe8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setopt.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setuptools.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setuptools.cpython-312.pyc
new file mode 100644
index 00000000..8dffb7ba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_setuptools.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_shutil_wrapper.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_shutil_wrapper.cpython-312.pyc
new file mode 100644
index 00000000..ab70df5d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_shutil_wrapper.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_unicode_utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_unicode_utils.cpython-312.pyc
new file mode 100644
index 00000000..f3cb8a05
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_unicode_utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_virtualenv.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_virtualenv.cpython-312.pyc
new file mode 100644
index 00000000..d64365f9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_virtualenv.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_warnings.cpython-312.pyc
new file mode 100644
index 00000000..90532ed1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_wheel.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_wheel.cpython-312.pyc
new file mode 100644
index 00000000..b35895a6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_wheel.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_windows_wrappers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_windows_wrappers.cpython-312.pyc
new file mode 100644
index 00000000..3dfb9325
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/test_windows_wrappers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/text.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/text.cpython-312.pyc
new file mode 100644
index 00000000..13f770b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/text.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/textwrap.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/textwrap.cpython-312.pyc
new file mode 100644
index 00000000..7b609b08
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/__pycache__/textwrap.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..711d8ea4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/py39.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/py39.cpython-312.pyc
new file mode 100644
index 00000000..4915bfc3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/__pycache__/py39.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/py39.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/py39.py
new file mode 100644
index 00000000..1fdb9dac
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/compat/py39.py
@@ -0,0 +1,3 @@
+from jaraco.test.cpython import from_test_support, try_import
+
+os_helper = try_import('os_helper') or from_test_support('can_symlink')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ff991e64
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_apply_pyprojecttoml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_apply_pyprojecttoml.cpython-312.pyc
new file mode 100644
index 00000000..7ca157eb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_apply_pyprojecttoml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_expand.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_expand.cpython-312.pyc
new file mode 100644
index 00000000..69c24462
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_expand.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml.cpython-312.pyc
new file mode 100644
index 00000000..7e60ece5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml_dynamic_deps.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml_dynamic_deps.cpython-312.pyc
new file mode 100644
index 00000000..46414d22
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_pyprojecttoml_dynamic_deps.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_setupcfg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_setupcfg.cpython-312.pyc
new file mode 100644
index 00000000..c80f5b3c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/__pycache__/test_setupcfg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__init__.py
new file mode 100644
index 00000000..00a16423
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__init__.py
@@ -0,0 +1,59 @@
+from __future__ import annotations
+
+import re
+import time
+from pathlib import Path
+from urllib.error import HTTPError
+from urllib.request import urlopen
+
+__all__ = ["DOWNLOAD_DIR", "retrieve_file", "output_file", "urls_from_file"]
+
+
+NAME_REMOVE = ("http://", "https://", "github.com/", "/raw/")
+DOWNLOAD_DIR = Path(__file__).parent
+
+
+# ----------------------------------------------------------------------
+# Please update ./preload.py accordingly when modifying this file
+# ----------------------------------------------------------------------
+
+
+def output_file(url: str, download_dir: Path = DOWNLOAD_DIR) -> Path:
+    file_name = url.strip()
+    for part in NAME_REMOVE:
+        file_name = file_name.replace(part, '').strip().strip('/:').strip()
+    return Path(download_dir, re.sub(r"[^\-_\.\w\d]+", "_", file_name))
+
+
+def retrieve_file(url: str, download_dir: Path = DOWNLOAD_DIR, wait: float = 5) -> Path:
+    path = output_file(url, download_dir)
+    if path.exists():
+        print(f"Skipping {url} (already exists: {path})")
+    else:
+        download_dir.mkdir(exist_ok=True, parents=True)
+        print(f"Downloading {url} to {path}")
+        try:
+            download(url, path)
+        except HTTPError:
+            time.sleep(wait)  # wait a few seconds and try again.
+            download(url, path)
+    return path
+
+
+def urls_from_file(list_file: Path) -> list[str]:
+    """``list_file`` should be a text file where each line corresponds to a URL to
+    download.
+    """
+    print(f"file: {list_file}")
+    content = list_file.read_text(encoding="utf-8")
+    return [url for url in content.splitlines() if not url.startswith("#")]
+
+
+def download(url: str, dest: Path):
+    with urlopen(url) as f:
+        data = f.read()
+
+    with open(dest, "wb") as f:
+        f.write(data)
+
+    assert Path(dest).exists()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..19d49505
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/preload.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/preload.cpython-312.pyc
new file mode 100644
index 00000000..67029701
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/__pycache__/preload.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/preload.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/preload.py
new file mode 100644
index 00000000..8eeb5dd7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/downloads/preload.py
@@ -0,0 +1,18 @@
+"""This file can be used to preload files needed for testing.
+
+For example you can use::
+
+    cd setuptools/tests/config
+    python -m downloads.preload setupcfg_examples.txt
+
+to make sure the `setup.cfg` examples are downloaded before starting the tests.
+"""
+
+import sys
+from pathlib import Path
+
+from . import retrieve_file, urls_from_file
+
+if __name__ == "__main__":
+    urls = urls_from_file(Path(sys.argv[1]))
+    list(map(retrieve_file, urls))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/setupcfg_examples.txt b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/setupcfg_examples.txt
new file mode 100644
index 00000000..6aab887f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/setupcfg_examples.txt
@@ -0,0 +1,22 @@
+# ====================================================================
+# Some popular packages that use setup.cfg (and others not so popular)
+# Reference: https://hugovk.github.io/top-pypi-packages/
+# ====================================================================
+https://github.com/pypa/setuptools/raw/52c990172fec37766b3566679724aa8bf70ae06d/setup.cfg
+https://github.com/pypa/wheel/raw/0acd203cd896afec7f715aa2ff5980a403459a3b/setup.cfg
+https://github.com/python/importlib_metadata/raw/2f05392ca980952a6960d82b2f2d2ea10aa53239/setup.cfg
+https://github.com/jaraco/skeleton/raw/d9008b5c510cd6969127a6a2ab6f832edddef296/setup.cfg
+https://github.com/jaraco/zipp/raw/700d3a96390e970b6b962823bfea78b4f7e1c537/setup.cfg
+https://github.com/pallets/jinja/raw/7d72eb7fefb7dce065193967f31f805180508448/setup.cfg
+https://github.com/tkem/cachetools/raw/2fd87a94b8d3861d80e9e4236cd480bfdd21c90d/setup.cfg
+https://github.com/aio-libs/aiohttp/raw/5e0e6b7080f2408d5f1dd544c0e1cf88378b7b10/setup.cfg
+https://github.com/pallets/flask/raw/9486b6cf57bd6a8a261f67091aca8ca78eeec1e3/setup.cfg
+https://github.com/pallets/click/raw/6411f425fae545f42795665af4162006b36c5e4a/setup.cfg
+https://github.com/sqlalchemy/sqlalchemy/raw/533f5718904b620be8d63f2474229945d6f8ba5d/setup.cfg
+https://github.com/pytest-dev/pluggy/raw/461ef63291d13589c4e21aa182cd1529257e9a0a/setup.cfg
+https://github.com/pytest-dev/pytest/raw/c7be96dae487edbd2f55b561b31b68afac1dabe6/setup.cfg
+https://github.com/platformdirs/platformdirs/raw/7b7852128dd6f07511b618d6edea35046bd0c6ff/setup.cfg
+https://github.com/pandas-dev/pandas/raw/bc17343f934a33dc231c8c74be95d8365537c376/setup.cfg
+https://github.com/django/django/raw/4e249d11a6e56ca8feb4b055b681cec457ef3a3d/setup.cfg
+https://github.com/pyscaffold/pyscaffold/raw/de7aa5dc059fbd04307419c667cc4961bc9df4b8/setup.cfg
+https://github.com/pypa/virtualenv/raw/f92eda6e3da26a4d28c2663ffb85c4960bdb990c/setup.cfg
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_apply_pyprojecttoml.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_apply_pyprojecttoml.py
new file mode 100644
index 00000000..20146b4a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_apply_pyprojecttoml.py
@@ -0,0 +1,539 @@
+"""Make sure that applying the configuration from pyproject.toml is equivalent to
+applying a similar configuration from setup.cfg
+
+To run these tests offline, please have a look on ``./downloads/preload.py``
+"""
+
+from __future__ import annotations
+
+import io
+import re
+import tarfile
+from inspect import cleandoc
+from pathlib import Path
+from unittest.mock import Mock
+
+import pytest
+from ini2toml.api import LiteTranslator
+from packaging.metadata import Metadata
+
+import setuptools  # noqa: F401 # ensure monkey patch to metadata
+from setuptools._static import is_static
+from setuptools.command.egg_info import write_requirements
+from setuptools.config import expand, pyprojecttoml, setupcfg
+from setuptools.config._apply_pyprojecttoml import _MissingDynamic, _some_attrgetter
+from setuptools.dist import Distribution
+from setuptools.errors import RemovedConfigError
+
+from .downloads import retrieve_file, urls_from_file
+
+HERE = Path(__file__).parent
+EXAMPLES_FILE = "setupcfg_examples.txt"
+
+
+def makedist(path, **attrs):
+    return Distribution({"src_root": path, **attrs})
+
+
+@pytest.mark.parametrize("url", urls_from_file(HERE / EXAMPLES_FILE))
+@pytest.mark.filterwarnings("ignore")
+@pytest.mark.uses_network
+def test_apply_pyproject_equivalent_to_setupcfg(url, monkeypatch, tmp_path):
+    monkeypatch.setattr(expand, "read_attr", Mock(return_value="0.0.1"))
+    setupcfg_example = retrieve_file(url)
+    pyproject_example = Path(tmp_path, "pyproject.toml")
+    setupcfg_text = setupcfg_example.read_text(encoding="utf-8")
+    toml_config = LiteTranslator().translate(setupcfg_text, "setup.cfg")
+    pyproject_example.write_text(toml_config, encoding="utf-8")
+
+    dist_toml = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject_example)
+    dist_cfg = setupcfg.apply_configuration(makedist(tmp_path), setupcfg_example)
+
+    pkg_info_toml = core_metadata(dist_toml)
+    pkg_info_cfg = core_metadata(dist_cfg)
+    assert pkg_info_toml == pkg_info_cfg
+
+    if any(getattr(d, "license_files", None) for d in (dist_toml, dist_cfg)):
+        assert set(dist_toml.license_files) == set(dist_cfg.license_files)
+
+    if any(getattr(d, "entry_points", None) for d in (dist_toml, dist_cfg)):
+        print(dist_cfg.entry_points)
+        ep_toml = {
+            (k, *sorted(i.replace(" ", "") for i in v))
+            for k, v in dist_toml.entry_points.items()
+        }
+        ep_cfg = {
+            (k, *sorted(i.replace(" ", "") for i in v))
+            for k, v in dist_cfg.entry_points.items()
+        }
+        assert ep_toml == ep_cfg
+
+    if any(getattr(d, "package_data", None) for d in (dist_toml, dist_cfg)):
+        pkg_data_toml = {(k, *sorted(v)) for k, v in dist_toml.package_data.items()}
+        pkg_data_cfg = {(k, *sorted(v)) for k, v in dist_cfg.package_data.items()}
+        assert pkg_data_toml == pkg_data_cfg
+
+    if any(getattr(d, "data_files", None) for d in (dist_toml, dist_cfg)):
+        data_files_toml = {(k, *sorted(v)) for k, v in dist_toml.data_files}
+        data_files_cfg = {(k, *sorted(v)) for k, v in dist_cfg.data_files}
+        assert data_files_toml == data_files_cfg
+
+    assert set(dist_toml.install_requires) == set(dist_cfg.install_requires)
+    if any(getattr(d, "extras_require", None) for d in (dist_toml, dist_cfg)):
+        extra_req_toml = {(k, *sorted(v)) for k, v in dist_toml.extras_require.items()}
+        extra_req_cfg = {(k, *sorted(v)) for k, v in dist_cfg.extras_require.items()}
+        assert extra_req_toml == extra_req_cfg
+
+
+PEP621_EXAMPLE = """\
+[project]
+name = "spam"
+version = "2020.0.0"
+description = "Lovely Spam! Wonderful Spam!"
+readme = "README.rst"
+requires-python = ">=3.8"
+license = {file = "LICENSE.txt"}
+keywords = ["egg", "bacon", "sausage", "tomatoes", "Lobster Thermidor"]
+authors = [
+  {email = "hi@pradyunsg.me"},
+  {name = "Tzu-Ping Chung"}
+]
+maintainers = [
+  {name = "Brett Cannon", email = "brett@python.org"},
+  {name = "John X. Ãørçeč", email = "john@utf8.org"},
+  {name = "Γαμα קּ 東", email = "gama@utf8.org"},
+]
+classifiers = [
+  "Development Status :: 4 - Beta",
+  "Programming Language :: Python"
+]
+
+dependencies = [
+  "httpx",
+  "gidgethub[httpx]>4.0.0",
+  "django>2.1; os_name != 'nt'",
+  "django>2.0; os_name == 'nt'"
+]
+
+[project.optional-dependencies]
+test = [
+  "pytest < 5.0.0",
+  "pytest-cov[all]"
+]
+
+[project.urls]
+homepage = "http://example.com"
+documentation = "http://readthedocs.org"
+repository = "http://github.com"
+changelog = "http://github.com/me/spam/blob/master/CHANGELOG.md"
+
+[project.scripts]
+spam-cli = "spam:main_cli"
+
+[project.gui-scripts]
+spam-gui = "spam:main_gui"
+
+[project.entry-points."spam.magical"]
+tomatoes = "spam:main_tomatoes"
+"""
+
+PEP621_INTERNATIONAL_EMAIL_EXAMPLE = """\
+[project]
+name = "spam"
+version = "2020.0.0"
+authors = [
+  {email = "hi@pradyunsg.me"},
+  {name = "Tzu-Ping Chung"}
+]
+maintainers = [
+  {name = "Степан Бандера", email = "криївка@оун-упа.укр"},
+]
+"""
+
+PEP621_EXAMPLE_SCRIPT = """
+def main_cli(): pass
+def main_gui(): pass
+def main_tomatoes(): pass
+"""
+
+
+def _pep621_example_project(
+    tmp_path,
+    readme="README.rst",
+    pyproject_text=PEP621_EXAMPLE,
+):
+    pyproject = tmp_path / "pyproject.toml"
+    text = pyproject_text
+    replacements = {'readme = "README.rst"': f'readme = "{readme}"'}
+    for orig, subst in replacements.items():
+        text = text.replace(orig, subst)
+    pyproject.write_text(text, encoding="utf-8")
+
+    (tmp_path / readme).write_text("hello world", encoding="utf-8")
+    (tmp_path / "LICENSE.txt").write_text("--- LICENSE stub ---", encoding="utf-8")
+    (tmp_path / "spam.py").write_text(PEP621_EXAMPLE_SCRIPT, encoding="utf-8")
+    return pyproject
+
+
+def test_pep621_example(tmp_path):
+    """Make sure the example in PEP 621 works"""
+    pyproject = _pep621_example_project(tmp_path)
+    dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+    assert dist.metadata.license == "--- LICENSE stub ---"
+    assert set(dist.metadata.license_files) == {"LICENSE.txt"}
+
+
+@pytest.mark.parametrize(
+    ("readme", "ctype"),
+    [
+        ("Readme.txt", "text/plain"),
+        ("readme.md", "text/markdown"),
+        ("text.rst", "text/x-rst"),
+    ],
+)
+def test_readme_content_type(tmp_path, readme, ctype):
+    pyproject = _pep621_example_project(tmp_path, readme)
+    dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+    assert dist.metadata.long_description_content_type == ctype
+
+
+def test_undefined_content_type(tmp_path):
+    pyproject = _pep621_example_project(tmp_path, "README.tex")
+    with pytest.raises(ValueError, match="Undefined content type for README.tex"):
+        pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+
+
+def test_no_explicit_content_type_for_missing_extension(tmp_path):
+    pyproject = _pep621_example_project(tmp_path, "README")
+    dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+    assert dist.metadata.long_description_content_type is None
+
+
+@pytest.mark.parametrize(
+    ("pyproject_text", "expected_maintainers_meta_value"),
+    (
+        pytest.param(
+            PEP621_EXAMPLE,
+            (
+                'Brett Cannon <brett@python.org>, "John X. Ãørçeč" <john@utf8.org>, '
+                'Γαμα קּ 東 <gama@utf8.org>'
+            ),
+            id='non-international-emails',
+        ),
+        pytest.param(
+            PEP621_INTERNATIONAL_EMAIL_EXAMPLE,
+            'Степан Бандера <криївка@оун-упа.укр>',
+            marks=pytest.mark.xfail(
+                reason="CPython's `email.headerregistry.Address` only supports "
+                'RFC 5322, as of Nov 10, 2022 and latest Python 3.11.0',
+                strict=True,
+            ),
+            id='international-email',
+        ),
+    ),
+)
+def test_utf8_maintainer_in_metadata(  # issue-3663
+    expected_maintainers_meta_value,
+    pyproject_text,
+    tmp_path,
+):
+    pyproject = _pep621_example_project(
+        tmp_path,
+        "README",
+        pyproject_text=pyproject_text,
+    )
+    dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+    assert dist.metadata.maintainer_email == expected_maintainers_meta_value
+    pkg_file = tmp_path / "PKG-FILE"
+    with open(pkg_file, "w", encoding="utf-8") as fh:
+        dist.metadata.write_pkg_file(fh)
+    content = pkg_file.read_text(encoding="utf-8")
+    assert f"Maintainer-email: {expected_maintainers_meta_value}" in content
+
+
+class TestLicenseFiles:
+    # TODO: After PEP 639 is accepted, we have to move the license-files
+    #       to the `project` table instead of `tool.setuptools`
+
+    def base_pyproject(self, tmp_path, additional_text):
+        pyproject = _pep621_example_project(tmp_path, "README")
+        text = pyproject.read_text(encoding="utf-8")
+
+        # Sanity-check
+        assert 'license = {file = "LICENSE.txt"}' in text
+        assert "[tool.setuptools]" not in text
+
+        text = f"{text}\n{additional_text}\n"
+        pyproject.write_text(text, encoding="utf-8")
+        return pyproject
+
+    def test_both_license_and_license_files_defined(self, tmp_path):
+        setuptools_config = '[tool.setuptools]\nlicense-files = ["_FILE*"]'
+        pyproject = self.base_pyproject(tmp_path, setuptools_config)
+
+        (tmp_path / "_FILE.txt").touch()
+        (tmp_path / "_FILE.rst").touch()
+
+        # Would normally match the `license_files` patterns, but we want to exclude it
+        # by being explicit. On the other hand, contents should be added to `license`
+        license = tmp_path / "LICENSE.txt"
+        license.write_text("LicenseRef-Proprietary\n", encoding="utf-8")
+
+        dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+        assert set(dist.metadata.license_files) == {"_FILE.rst", "_FILE.txt"}
+        assert dist.metadata.license == "LicenseRef-Proprietary\n"
+
+    def test_default_patterns(self, tmp_path):
+        setuptools_config = '[tool.setuptools]\nzip-safe = false'
+        # ^ used just to trigger section validation
+        pyproject = self.base_pyproject(tmp_path, setuptools_config)
+
+        license_files = "LICENCE-a.html COPYING-abc.txt AUTHORS-xyz NOTICE,def".split()
+
+        for fname in license_files:
+            (tmp_path / fname).write_text(f"{fname}\n", encoding="utf-8")
+
+        dist = pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+        assert (tmp_path / "LICENSE.txt").exists()  # from base example
+        assert set(dist.metadata.license_files) == {*license_files, "LICENSE.txt"}
+
+
+class TestPyModules:
+    # https://github.com/pypa/setuptools/issues/4316
+
+    def dist(self, name):
+        toml_config = f"""
+        [project]
+        name = "test"
+        version = "42.0"
+        [tool.setuptools]
+        py-modules = [{name!r}]
+        """
+        pyproject = Path("pyproject.toml")
+        pyproject.write_text(cleandoc(toml_config), encoding="utf-8")
+        return pyprojecttoml.apply_configuration(Distribution({}), pyproject)
+
+    @pytest.mark.parametrize("module", ["pip-run", "abc-d.λ-xyz-e"])
+    def test_valid_module_name(self, tmp_path, monkeypatch, module):
+        monkeypatch.chdir(tmp_path)
+        assert module in self.dist(module).py_modules
+
+    @pytest.mark.parametrize("module", ["pip run", "-pip-run", "pip-run-stubs"])
+    def test_invalid_module_name(self, tmp_path, monkeypatch, module):
+        monkeypatch.chdir(tmp_path)
+        with pytest.raises(ValueError, match="py-modules"):
+            self.dist(module).py_modules
+
+
+class TestExtModules:
+    def test_pyproject_sets_attribute(self, tmp_path, monkeypatch):
+        monkeypatch.chdir(tmp_path)
+        pyproject = Path("pyproject.toml")
+        toml_config = """
+        [project]
+        name = "test"
+        version = "42.0"
+        [tool.setuptools]
+        ext-modules = [
+          {name = "my.ext", sources = ["hello.c", "world.c"]}
+        ]
+        """
+        pyproject.write_text(cleandoc(toml_config), encoding="utf-8")
+        with pytest.warns(pyprojecttoml._ExperimentalConfiguration):
+            dist = pyprojecttoml.apply_configuration(Distribution({}), pyproject)
+        assert len(dist.ext_modules) == 1
+        assert dist.ext_modules[0].name == "my.ext"
+        assert set(dist.ext_modules[0].sources) == {"hello.c", "world.c"}
+
+
+class TestDeprecatedFields:
+    def test_namespace_packages(self, tmp_path):
+        pyproject = tmp_path / "pyproject.toml"
+        config = """
+        [project]
+        name = "myproj"
+        version = "42"
+        [tool.setuptools]
+        namespace-packages = ["myproj.pkg"]
+        """
+        pyproject.write_text(cleandoc(config), encoding="utf-8")
+        with pytest.raises(RemovedConfigError, match="namespace-packages"):
+            pyprojecttoml.apply_configuration(makedist(tmp_path), pyproject)
+
+
+class TestPresetField:
+    def pyproject(self, tmp_path, dynamic, extra_content=""):
+        content = f"[project]\nname = 'proj'\ndynamic = {dynamic!r}\n"
+        if "version" not in dynamic:
+            content += "version = '42'\n"
+        file = tmp_path / "pyproject.toml"
+        file.write_text(content + extra_content, encoding="utf-8")
+        return file
+
+    @pytest.mark.parametrize(
+        ("attr", "field", "value"),
+        [
+            ("classifiers", "classifiers", ["Private :: Classifier"]),
+            ("entry_points", "scripts", {"console_scripts": ["foobar=foobar:main"]}),
+            ("entry_points", "gui-scripts", {"gui_scripts": ["bazquux=bazquux:main"]}),
+            pytest.param(
+                *("install_requires", "dependencies", ["six"]),
+                marks=[
+                    pytest.mark.filterwarnings("ignore:.*install_requires. overwritten")
+                ],
+            ),
+        ],
+    )
+    def test_not_listed_in_dynamic(self, tmp_path, attr, field, value):
+        """Setuptools cannot set a field if not listed in ``dynamic``"""
+        pyproject = self.pyproject(tmp_path, [])
+        dist = makedist(tmp_path, **{attr: value})
+        msg = re.compile(f"defined outside of `pyproject.toml`:.*{field}", re.S)
+        with pytest.warns(_MissingDynamic, match=msg):
+            dist = pyprojecttoml.apply_configuration(dist, pyproject)
+
+        dist_value = _some_attrgetter(f"metadata.{attr}", attr)(dist)
+        assert not dist_value
+
+    @pytest.mark.parametrize(
+        ("attr", "field", "value"),
+        [
+            ("install_requires", "dependencies", []),
+            ("extras_require", "optional-dependencies", {}),
+            ("install_requires", "dependencies", ["six"]),
+            ("classifiers", "classifiers", ["Private :: Classifier"]),
+        ],
+    )
+    def test_listed_in_dynamic(self, tmp_path, attr, field, value):
+        pyproject = self.pyproject(tmp_path, [field])
+        dist = makedist(tmp_path, **{attr: value})
+        dist = pyprojecttoml.apply_configuration(dist, pyproject)
+        dist_value = _some_attrgetter(f"metadata.{attr}", attr)(dist)
+        assert dist_value == value
+
+    def test_warning_overwritten_dependencies(self, tmp_path):
+        src = "[project]\nname='pkg'\nversion='0.1'\ndependencies=['click']\n"
+        pyproject = tmp_path / "pyproject.toml"
+        pyproject.write_text(src, encoding="utf-8")
+        dist = makedist(tmp_path, install_requires=["wheel"])
+        with pytest.warns(match="`install_requires` overwritten"):
+            dist = pyprojecttoml.apply_configuration(dist, pyproject)
+        assert "wheel" not in dist.install_requires
+
+    def test_optional_dependencies_dont_remove_env_markers(self, tmp_path):
+        """
+        Internally setuptools converts dependencies with markers to "extras".
+        If ``install_requires`` is given by ``setup.py``, we have to ensure that
+        applying ``optional-dependencies`` does not overwrite the mandatory
+        dependencies with markers (see #3204).
+        """
+        # If setuptools replace its internal mechanism that uses `requires.txt`
+        # this test has to be rewritten to adapt accordingly
+        extra = "\n[project.optional-dependencies]\nfoo = ['bar>1']\n"
+        pyproject = self.pyproject(tmp_path, ["dependencies"], extra)
+        install_req = ['importlib-resources (>=3.0.0) ; python_version < "3.7"']
+        dist = makedist(tmp_path, install_requires=install_req)
+        dist = pyprojecttoml.apply_configuration(dist, pyproject)
+        assert "foo" in dist.extras_require
+        egg_info = dist.get_command_obj("egg_info")
+        write_requirements(egg_info, tmp_path, tmp_path / "requires.txt")
+        reqs = (tmp_path / "requires.txt").read_text(encoding="utf-8")
+        assert "importlib-resources" in reqs
+        assert "bar" in reqs
+        assert ':python_version < "3.7"' in reqs
+
+    @pytest.mark.parametrize(
+        ("field", "group"),
+        [("scripts", "console_scripts"), ("gui-scripts", "gui_scripts")],
+    )
+    @pytest.mark.filterwarnings("error")
+    def test_scripts_dont_require_dynamic_entry_points(self, tmp_path, field, group):
+        # Issue 3862
+        pyproject = self.pyproject(tmp_path, [field])
+        dist = makedist(tmp_path, entry_points={group: ["foobar=foobar:main"]})
+        dist = pyprojecttoml.apply_configuration(dist, pyproject)
+        assert group in dist.entry_points
+
+
+class TestMeta:
+    def test_example_file_in_sdist(self, setuptools_sdist):
+        """Meta test to ensure tests can run from sdist"""
+        with tarfile.open(setuptools_sdist) as tar:
+            assert any(name.endswith(EXAMPLES_FILE) for name in tar.getnames())
+
+
+class TestInteropCommandLineParsing:
+    def test_version(self, tmp_path, monkeypatch, capsys):
+        # See pypa/setuptools#4047
+        # This test can be removed once the CLI interface of setup.py is removed
+        monkeypatch.chdir(tmp_path)
+        toml_config = """
+        [project]
+        name = "test"
+        version = "42.0"
+        """
+        pyproject = Path(tmp_path, "pyproject.toml")
+        pyproject.write_text(cleandoc(toml_config), encoding="utf-8")
+        opts = {"script_args": ["--version"]}
+        dist = pyprojecttoml.apply_configuration(Distribution(opts), pyproject)
+        dist.parse_command_line()  # <-- there should be no exception here.
+        captured = capsys.readouterr()
+        assert "42.0" in captured.out
+
+
+class TestStaticConfig:
+    def test_mark_static_fields(self, tmp_path, monkeypatch):
+        monkeypatch.chdir(tmp_path)
+        toml_config = """
+        [project]
+        name = "test"
+        version = "42.0"
+        dependencies = ["hello"]
+        keywords = ["world"]
+        classifiers = ["private :: hello world"]
+        [tool.setuptools]
+        obsoletes = ["abcd"]
+        provides = ["abcd"]
+        platforms = ["abcd"]
+        """
+        pyproject = Path(tmp_path, "pyproject.toml")
+        pyproject.write_text(cleandoc(toml_config), encoding="utf-8")
+        dist = pyprojecttoml.apply_configuration(Distribution({}), pyproject)
+        assert is_static(dist.install_requires)
+        assert is_static(dist.metadata.keywords)
+        assert is_static(dist.metadata.classifiers)
+        assert is_static(dist.metadata.obsoletes)
+        assert is_static(dist.metadata.provides)
+        assert is_static(dist.metadata.platforms)
+
+
+# --- Auxiliary Functions ---
+
+
+def core_metadata(dist) -> str:
+    with io.StringIO() as buffer:
+        dist.metadata.write_pkg_file(buffer)
+        pkg_file_txt = buffer.getvalue()
+
+    # Make sure core metadata is valid
+    Metadata.from_email(pkg_file_txt, validate=True)  # can raise exceptions
+
+    skip_prefixes: tuple[str, ...] = ()
+    skip_lines = set()
+    # ---- DIFF NORMALISATION ----
+    # PEP 621 is very particular about author/maintainer metadata conversion, so skip
+    skip_prefixes += ("Author:", "Author-email:", "Maintainer:", "Maintainer-email:")
+    # May be redundant with Home-page
+    skip_prefixes += ("Project-URL: Homepage,", "Home-page:")
+    # May be missing in original (relying on default) but backfilled in the TOML
+    skip_prefixes += ("Description-Content-Type:",)
+    # Remove empty lines
+    skip_lines.add("")
+
+    result = []
+    for line in pkg_file_txt.splitlines():
+        if line.startswith(skip_prefixes) or line in skip_lines:
+            continue
+        result.append(line + "\n")
+
+    return "".join(result)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_expand.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_expand.py
new file mode 100644
index 00000000..c5710ec6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_expand.py
@@ -0,0 +1,247 @@
+import os
+import sys
+from pathlib import Path
+
+import pytest
+
+from setuptools._static import is_static
+from setuptools.config import expand
+from setuptools.discovery import find_package_path
+
+from distutils.errors import DistutilsOptionError
+
+
+def write_files(files, root_dir):
+    for file, content in files.items():
+        path = root_dir / file
+        path.parent.mkdir(exist_ok=True, parents=True)
+        path.write_text(content, encoding="utf-8")
+
+
+def test_glob_relative(tmp_path, monkeypatch):
+    files = {
+        "dir1/dir2/dir3/file1.txt",
+        "dir1/dir2/file2.txt",
+        "dir1/file3.txt",
+        "a.ini",
+        "b.ini",
+        "dir1/c.ini",
+        "dir1/dir2/a.ini",
+    }
+
+    write_files({k: "" for k in files}, tmp_path)
+    patterns = ["**/*.txt", "[ab].*", "**/[ac].ini"]
+    monkeypatch.chdir(tmp_path)
+    assert set(expand.glob_relative(patterns)) == files
+    # Make sure the same APIs work outside cwd
+    assert set(expand.glob_relative(patterns, tmp_path)) == files
+
+
+def test_read_files(tmp_path, monkeypatch):
+    dir_ = tmp_path / "dir_"
+    (tmp_path / "_dir").mkdir(exist_ok=True)
+    (tmp_path / "a.txt").touch()
+    files = {"a.txt": "a", "dir1/b.txt": "b", "dir1/dir2/c.txt": "c"}
+    write_files(files, dir_)
+
+    secrets = Path(str(dir_) + "secrets")
+    secrets.mkdir(exist_ok=True)
+    write_files({"secrets.txt": "secret keys"}, secrets)
+
+    with monkeypatch.context() as m:
+        m.chdir(dir_)
+        assert expand.read_files(list(files)) == "a\nb\nc"
+
+        cannot_access_msg = r"Cannot access '.*\.\..a\.txt'"
+        with pytest.raises(DistutilsOptionError, match=cannot_access_msg):
+            expand.read_files(["../a.txt"])
+
+        cannot_access_secrets_msg = r"Cannot access '.*secrets\.txt'"
+        with pytest.raises(DistutilsOptionError, match=cannot_access_secrets_msg):
+            expand.read_files(["../dir_secrets/secrets.txt"])
+
+    # Make sure the same APIs work outside cwd
+    assert expand.read_files(list(files), dir_) == "a\nb\nc"
+    with pytest.raises(DistutilsOptionError, match=cannot_access_msg):
+        expand.read_files(["../a.txt"], dir_)
+
+
+class TestReadAttr:
+    @pytest.mark.parametrize(
+        "example",
+        [
+            # No cookie means UTF-8:
+            b"__version__ = '\xc3\xa9'\nraise SystemExit(1)\n",
+            # If a cookie is present, honor it:
+            b"# -*- coding: utf-8 -*-\n__version__ = '\xc3\xa9'\nraise SystemExit(1)\n",
+            b"# -*- coding: latin1 -*-\n__version__ = '\xe9'\nraise SystemExit(1)\n",
+        ],
+    )
+    def test_read_attr_encoding_cookie(self, example, tmp_path):
+        (tmp_path / "mod.py").write_bytes(example)
+        assert expand.read_attr('mod.__version__', root_dir=tmp_path) == 'é'
+
+    def test_read_attr(self, tmp_path, monkeypatch):
+        files = {
+            "pkg/__init__.py": "",
+            "pkg/sub/__init__.py": "VERSION = '0.1.1'",
+            "pkg/sub/mod.py": (
+                "VALUES = {'a': 0, 'b': {42}, 'c': (0, 1, 1)}\nraise SystemExit(1)"
+            ),
+        }
+        write_files(files, tmp_path)
+
+        with monkeypatch.context() as m:
+            m.chdir(tmp_path)
+            # Make sure it can read the attr statically without evaluating the module
+            version = expand.read_attr('pkg.sub.VERSION')
+            values = expand.read_attr('lib.mod.VALUES', {'lib': 'pkg/sub'})
+
+        assert version == '0.1.1'
+        assert is_static(values)
+
+        assert values['a'] == 0
+        assert values['b'] == {42}
+        assert is_static(values)
+
+        # Make sure the same APIs work outside cwd
+        assert expand.read_attr('pkg.sub.VERSION', root_dir=tmp_path) == '0.1.1'
+        values = expand.read_attr('lib.mod.VALUES', {'lib': 'pkg/sub'}, tmp_path)
+        assert values['c'] == (0, 1, 1)
+
+    @pytest.mark.parametrize(
+        "example",
+        [
+            "VERSION: str\nVERSION = '0.1.1'\nraise SystemExit(1)\n",
+            "VERSION: str = '0.1.1'\nraise SystemExit(1)\n",
+        ],
+    )
+    def test_read_annotated_attr(self, tmp_path, example):
+        files = {
+            "pkg/__init__.py": "",
+            "pkg/sub/__init__.py": example,
+        }
+        write_files(files, tmp_path)
+        # Make sure this attribute can be read statically
+        version = expand.read_attr('pkg.sub.VERSION', root_dir=tmp_path)
+        assert version == '0.1.1'
+        assert is_static(version)
+
+    @pytest.mark.parametrize(
+        "example",
+        [
+            "VERSION = (lambda: '0.1.1')()\n",
+            "def fn(): return '0.1.1'\nVERSION = fn()\n",
+            "VERSION: str = (lambda: '0.1.1')()\n",
+        ],
+    )
+    def test_read_dynamic_attr(self, tmp_path, monkeypatch, example):
+        files = {
+            "pkg/__init__.py": "",
+            "pkg/sub/__init__.py": example,
+        }
+        write_files(files, tmp_path)
+        monkeypatch.chdir(tmp_path)
+        version = expand.read_attr('pkg.sub.VERSION')
+        assert version == '0.1.1'
+        assert not is_static(version)
+
+    def test_import_order(self, tmp_path):
+        """
+        Sometimes the import machinery will import the parent package of a nested
+        module, which triggers side-effects and might create problems (see issue #3176)
+
+        ``read_attr`` should bypass these limitations by resolving modules statically
+        (via ast.literal_eval).
+        """
+        files = {
+            "src/pkg/__init__.py": "from .main import func\nfrom .about import version",
+            "src/pkg/main.py": "import super_complicated_dep\ndef func(): return 42",
+            "src/pkg/about.py": "version = '42'",
+        }
+        write_files(files, tmp_path)
+        attr_desc = "pkg.about.version"
+        package_dir = {"": "src"}
+        # `import super_complicated_dep` should not run, otherwise the build fails
+        assert expand.read_attr(attr_desc, package_dir, tmp_path) == "42"
+
+
+@pytest.mark.parametrize(
+    ("package_dir", "file", "module", "return_value"),
+    [
+        ({"": "src"}, "src/pkg/main.py", "pkg.main", 42),
+        ({"pkg": "lib"}, "lib/main.py", "pkg.main", 13),
+        ({}, "single_module.py", "single_module", 70),
+        ({}, "flat_layout/pkg.py", "flat_layout.pkg", 836),
+    ],
+)
+def test_resolve_class(monkeypatch, tmp_path, package_dir, file, module, return_value):
+    monkeypatch.setattr(sys, "modules", {})  # reproducibility
+    files = {file: f"class Custom:\n    def testing(self): return {return_value}"}
+    write_files(files, tmp_path)
+    cls = expand.resolve_class(f"{module}.Custom", package_dir, tmp_path)
+    assert cls().testing() == return_value
+
+
+@pytest.mark.parametrize(
+    ("args", "pkgs"),
+    [
+        ({"where": ["."], "namespaces": False}, {"pkg", "other"}),
+        ({"where": [".", "dir1"], "namespaces": False}, {"pkg", "other", "dir2"}),
+        ({"namespaces": True}, {"pkg", "other", "dir1", "dir1.dir2"}),
+        ({}, {"pkg", "other", "dir1", "dir1.dir2"}),  # default value for `namespaces`
+    ],
+)
+def test_find_packages(tmp_path, args, pkgs):
+    files = {
+        "pkg/__init__.py",
+        "other/__init__.py",
+        "dir1/dir2/__init__.py",
+    }
+    write_files({k: "" for k in files}, tmp_path)
+
+    package_dir = {}
+    kwargs = {"root_dir": tmp_path, "fill_package_dir": package_dir, **args}
+    where = kwargs.get("where", ["."])
+    assert set(expand.find_packages(**kwargs)) == pkgs
+    for pkg in pkgs:
+        pkg_path = find_package_path(pkg, package_dir, tmp_path)
+        assert os.path.exists(pkg_path)
+
+    # Make sure the same APIs work outside cwd
+    where = [
+        str((tmp_path / p).resolve()).replace(os.sep, "/")  # ensure posix-style paths
+        for p in args.pop("where", ["."])
+    ]
+
+    assert set(expand.find_packages(where=where, **args)) == pkgs
+
+
+@pytest.mark.parametrize(
+    ("files", "where", "expected_package_dir"),
+    [
+        (["pkg1/__init__.py", "pkg1/other.py"], ["."], {}),
+        (["pkg1/__init__.py", "pkg2/__init__.py"], ["."], {}),
+        (["src/pkg1/__init__.py", "src/pkg1/other.py"], ["src"], {"": "src"}),
+        (["src/pkg1/__init__.py", "src/pkg2/__init__.py"], ["src"], {"": "src"}),
+        (
+            ["src1/pkg1/__init__.py", "src2/pkg2/__init__.py"],
+            ["src1", "src2"],
+            {"pkg1": "src1/pkg1", "pkg2": "src2/pkg2"},
+        ),
+        (
+            ["src/pkg1/__init__.py", "pkg2/__init__.py"],
+            ["src", "."],
+            {"pkg1": "src/pkg1"},
+        ),
+    ],
+)
+def test_fill_package_dir(tmp_path, files, where, expected_package_dir):
+    write_files({k: "" for k in files}, tmp_path)
+    pkg_dir = {}
+    kwargs = {"root_dir": tmp_path, "fill_package_dir": pkg_dir, "namespaces": False}
+    pkgs = expand.find_packages(where=where, **kwargs)
+    assert set(pkg_dir.items()) == set(expected_package_dir.items())
+    for pkg in pkgs:
+        pkg_path = find_package_path(pkg, pkg_dir, tmp_path)
+        assert os.path.exists(pkg_path)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml.py
new file mode 100644
index 00000000..db40fcd2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml.py
@@ -0,0 +1,396 @@
+import re
+from configparser import ConfigParser
+from inspect import cleandoc
+
+import jaraco.path
+import pytest
+import tomli_w
+from path import Path
+
+import setuptools  # noqa: F401 # force distutils.core to be patched
+from setuptools.config.pyprojecttoml import (
+    _ToolsTypoInMetadata,
+    apply_configuration,
+    expand_configuration,
+    read_configuration,
+    validate,
+)
+from setuptools.dist import Distribution
+from setuptools.errors import OptionError
+
+import distutils.core
+
+EXAMPLE = """
+[project]
+name = "myproj"
+keywords = ["some", "key", "words"]
+dynamic = ["version", "readme"]
+requires-python = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
+dependencies = [
+    'importlib-metadata>=0.12;python_version<"3.8"',
+    'importlib-resources>=1.0;python_version<"3.7"',
+    'pathlib2>=2.3.3,<3;python_version < "3.4" and sys.platform != "win32"',
+]
+
+[project.optional-dependencies]
+docs = [
+    "sphinx>=3",
+    "sphinx-argparse>=0.2.5",
+    "sphinx-rtd-theme>=0.4.3",
+]
+testing = [
+    "pytest>=1",
+    "coverage>=3,<5",
+]
+
+[project.scripts]
+exec = "pkg.__main__:exec"
+
+[build-system]
+requires = ["setuptools", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[tool.setuptools]
+package-dir = {"" = "src"}
+zip-safe = true
+platforms = ["any"]
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.setuptools.cmdclass]
+sdist = "pkg.mod.CustomSdist"
+
+[tool.setuptools.dynamic.version]
+attr = "pkg.__version__.VERSION"
+
+[tool.setuptools.dynamic.readme]
+file = ["README.md"]
+content-type = "text/markdown"
+
+[tool.setuptools.package-data]
+"*" = ["*.txt"]
+
+[tool.setuptools.data-files]
+"data" = ["_files/*.txt"]
+
+[tool.distutils.sdist]
+formats = "gztar"
+
+[tool.distutils.bdist_wheel]
+universal = true
+"""
+
+
+def create_example(path, pkg_root):
+    files = {
+        "pyproject.toml": EXAMPLE,
+        "README.md": "hello world",
+        "_files": {
+            "file.txt": "",
+        },
+    }
+    packages = {
+        "pkg": {
+            "__init__.py": "",
+            "mod.py": "class CustomSdist: pass",
+            "__version__.py": "VERSION = (3, 10)",
+            "__main__.py": "def exec(): print('hello')",
+        },
+    }
+
+    assert pkg_root  # Meta-test: cannot be empty string.
+
+    if pkg_root == ".":
+        files = {**files, **packages}
+        # skip other files: flat-layout will raise error for multi-package dist
+    else:
+        # Use this opportunity to ensure namespaces are discovered
+        files[pkg_root] = {**packages, "other": {"nested": {"__init__.py": ""}}}
+
+    jaraco.path.build(files, prefix=path)
+
+
+def verify_example(config, path, pkg_root):
+    pyproject = path / "pyproject.toml"
+    pyproject.write_text(tomli_w.dumps(config), encoding="utf-8")
+    expanded = expand_configuration(config, path)
+    expanded_project = expanded["project"]
+    assert read_configuration(pyproject, expand=True) == expanded
+    assert expanded_project["version"] == "3.10"
+    assert expanded_project["readme"]["text"] == "hello world"
+    assert "packages" in expanded["tool"]["setuptools"]
+    if pkg_root == ".":
+        # Auto-discovery will raise error for multi-package dist
+        assert set(expanded["tool"]["setuptools"]["packages"]) == {"pkg"}
+    else:
+        assert set(expanded["tool"]["setuptools"]["packages"]) == {
+            "pkg",
+            "other",
+            "other.nested",
+        }
+    assert expanded["tool"]["setuptools"]["include-package-data"] is True
+    assert "" in expanded["tool"]["setuptools"]["package-data"]
+    assert "*" not in expanded["tool"]["setuptools"]["package-data"]
+    assert expanded["tool"]["setuptools"]["data-files"] == [
+        ("data", ["_files/file.txt"])
+    ]
+
+
+def test_read_configuration(tmp_path):
+    create_example(tmp_path, "src")
+    pyproject = tmp_path / "pyproject.toml"
+
+    config = read_configuration(pyproject, expand=False)
+    assert config["project"].get("version") is None
+    assert config["project"].get("readme") is None
+
+    verify_example(config, tmp_path, "src")
+
+
+@pytest.mark.parametrize(
+    ("pkg_root", "opts"),
+    [
+        (".", {}),
+        ("src", {}),
+        ("lib", {"packages": {"find": {"where": ["lib"]}}}),
+    ],
+)
+def test_discovered_package_dir_with_attr_directive_in_config(tmp_path, pkg_root, opts):
+    create_example(tmp_path, pkg_root)
+
+    pyproject = tmp_path / "pyproject.toml"
+
+    config = read_configuration(pyproject, expand=False)
+    assert config["project"].get("version") is None
+    assert config["project"].get("readme") is None
+    config["tool"]["setuptools"].pop("packages", None)
+    config["tool"]["setuptools"].pop("package-dir", None)
+
+    config["tool"]["setuptools"].update(opts)
+    verify_example(config, tmp_path, pkg_root)
+
+
+ENTRY_POINTS = {
+    "console_scripts": {"a": "mod.a:func"},
+    "gui_scripts": {"b": "mod.b:func"},
+    "other": {"c": "mod.c:func [extra]"},
+}
+
+
+class TestEntryPoints:
+    def write_entry_points(self, tmp_path):
+        entry_points = ConfigParser()
+        entry_points.read_dict(ENTRY_POINTS)
+        with open(tmp_path / "entry-points.txt", "w", encoding="utf-8") as f:
+            entry_points.write(f)
+
+    def pyproject(self, dynamic=None):
+        project = {"dynamic": dynamic or ["scripts", "gui-scripts", "entry-points"]}
+        tool = {"dynamic": {"entry-points": {"file": "entry-points.txt"}}}
+        return {"project": project, "tool": {"setuptools": tool}}
+
+    def test_all_listed_in_dynamic(self, tmp_path):
+        self.write_entry_points(tmp_path)
+        expanded = expand_configuration(self.pyproject(), tmp_path)
+        expanded_project = expanded["project"]
+        assert len(expanded_project["scripts"]) == 1
+        assert expanded_project["scripts"]["a"] == "mod.a:func"
+        assert len(expanded_project["gui-scripts"]) == 1
+        assert expanded_project["gui-scripts"]["b"] == "mod.b:func"
+        assert len(expanded_project["entry-points"]) == 1
+        assert expanded_project["entry-points"]["other"]["c"] == "mod.c:func [extra]"
+
+    @pytest.mark.parametrize("missing_dynamic", ("scripts", "gui-scripts"))
+    def test_scripts_not_listed_in_dynamic(self, tmp_path, missing_dynamic):
+        self.write_entry_points(tmp_path)
+        dynamic = {"scripts", "gui-scripts", "entry-points"} - {missing_dynamic}
+
+        msg = f"defined outside of `pyproject.toml`:.*{missing_dynamic}"
+        with pytest.raises(OptionError, match=re.compile(msg, re.S)):
+            expand_configuration(self.pyproject(dynamic), tmp_path)
+
+
+class TestClassifiers:
+    def test_dynamic(self, tmp_path):
+        # Let's create a project example that has dynamic classifiers
+        # coming from a txt file.
+        create_example(tmp_path, "src")
+        classifiers = cleandoc(
+            """
+            Framework :: Flask
+            Programming Language :: Haskell
+            """
+        )
+        (tmp_path / "classifiers.txt").write_text(classifiers, encoding="utf-8")
+
+        pyproject = tmp_path / "pyproject.toml"
+        config = read_configuration(pyproject, expand=False)
+        dynamic = config["project"]["dynamic"]
+        config["project"]["dynamic"] = list({*dynamic, "classifiers"})
+        dynamic_config = config["tool"]["setuptools"]["dynamic"]
+        dynamic_config["classifiers"] = {"file": "classifiers.txt"}
+
+        # When the configuration is expanded,
+        # each line of the file should be an different classifier.
+        validate(config, pyproject)
+        expanded = expand_configuration(config, tmp_path)
+
+        assert set(expanded["project"]["classifiers"]) == {
+            "Framework :: Flask",
+            "Programming Language :: Haskell",
+        }
+
+    def test_dynamic_without_config(self, tmp_path):
+        config = """
+        [project]
+        name = "myproj"
+        version = '42'
+        dynamic = ["classifiers"]
+        """
+
+        pyproject = tmp_path / "pyproject.toml"
+        pyproject.write_text(cleandoc(config), encoding="utf-8")
+        with pytest.raises(OptionError, match="No configuration .* .classifiers."):
+            read_configuration(pyproject)
+
+    def test_dynamic_readme_from_setup_script_args(self, tmp_path):
+        config = """
+        [project]
+        name = "myproj"
+        version = '42'
+        dynamic = ["readme"]
+        """
+        pyproject = tmp_path / "pyproject.toml"
+        pyproject.write_text(cleandoc(config), encoding="utf-8")
+        dist = Distribution(attrs={"long_description": "42"})
+        # No error should occur because of missing `readme`
+        dist = apply_configuration(dist, pyproject)
+        assert dist.metadata.long_description == "42"
+
+    def test_dynamic_without_file(self, tmp_path):
+        config = """
+        [project]
+        name = "myproj"
+        version = '42'
+        dynamic = ["classifiers"]
+
+        [tool.setuptools.dynamic]
+        classifiers = {file = ["classifiers.txt"]}
+        """
+
+        pyproject = tmp_path / "pyproject.toml"
+        pyproject.write_text(cleandoc(config), encoding="utf-8")
+        with pytest.warns(UserWarning, match="File .*classifiers.txt. cannot be found"):
+            expanded = read_configuration(pyproject)
+        assert "classifiers" not in expanded["project"]
+
+
+@pytest.mark.parametrize(
+    "example",
+    (
+        """
+        [project]
+        name = "myproj"
+        version = "1.2"
+
+        [my-tool.that-disrespect.pep518]
+        value = 42
+        """,
+    ),
+)
+def test_ignore_unrelated_config(tmp_path, example):
+    pyproject = tmp_path / "pyproject.toml"
+    pyproject.write_text(cleandoc(example), encoding="utf-8")
+
+    # Make sure no error is raised due to 3rd party configs in pyproject.toml
+    assert read_configuration(pyproject) is not None
+
+
+@pytest.mark.parametrize(
+    ("example", "error_msg"),
+    [
+        (
+            """
+            [project]
+            name = "myproj"
+            version = "1.2"
+            requires = ['pywin32; platform_system=="Windows"' ]
+            """,
+            "configuration error: .project. must not contain ..requires.. properties",
+        ),
+    ],
+)
+def test_invalid_example(tmp_path, example, error_msg):
+    pyproject = tmp_path / "pyproject.toml"
+    pyproject.write_text(cleandoc(example), encoding="utf-8")
+
+    pattern = re.compile(f"invalid pyproject.toml.*{error_msg}.*", re.M | re.S)
+    with pytest.raises(ValueError, match=pattern):
+        read_configuration(pyproject)
+
+
+@pytest.mark.parametrize("config", ("", "[tool.something]\nvalue = 42"))
+def test_empty(tmp_path, config):
+    pyproject = tmp_path / "pyproject.toml"
+    pyproject.write_text(config, encoding="utf-8")
+
+    # Make sure no error is raised
+    assert read_configuration(pyproject) == {}
+
+
+@pytest.mark.parametrize("config", ("[project]\nname = 'myproj'\nversion='42'\n",))
+def test_include_package_data_by_default(tmp_path, config):
+    """Builds with ``pyproject.toml`` should consider ``include-package-data=True`` as
+    default.
+    """
+    pyproject = tmp_path / "pyproject.toml"
+    pyproject.write_text(config, encoding="utf-8")
+
+    config = read_configuration(pyproject)
+    assert config["tool"]["setuptools"]["include-package-data"] is True
+
+
+def test_include_package_data_in_setuppy(tmp_path):
+    """Builds with ``pyproject.toml`` should consider ``include_package_data`` set in
+    ``setup.py``.
+
+    See https://github.com/pypa/setuptools/issues/3197#issuecomment-1079023889
+    """
+    files = {
+        "pyproject.toml": "[project]\nname = 'myproj'\nversion='42'\n",
+        "setup.py": "__import__('setuptools').setup(include_package_data=False)",
+    }
+    jaraco.path.build(files, prefix=tmp_path)
+
+    with Path(tmp_path):
+        dist = distutils.core.run_setup("setup.py", {}, stop_after="config")
+
+    assert dist.get_name() == "myproj"
+    assert dist.get_version() == "42"
+    assert dist.include_package_data is False
+
+
+def test_warn_tools_typo(tmp_path):
+    """Test that the common ``tools.setuptools`` typo in ``pyproject.toml`` issues a warning
+
+    See https://github.com/pypa/setuptools/issues/4150
+    """
+    config = """
+    [build-system]
+    requires = ["setuptools"]
+    build-backend = "setuptools.build_meta"
+
+    [project]
+    name = "myproj"
+    version = '42'
+
+    [tools.setuptools]
+    packages = ["package"]
+    """
+
+    pyproject = tmp_path / "pyproject.toml"
+    pyproject.write_text(cleandoc(config), encoding="utf-8")
+
+    with pytest.warns(_ToolsTypoInMetadata):
+        read_configuration(pyproject)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml_dynamic_deps.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml_dynamic_deps.py
new file mode 100644
index 00000000..e42f28ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_pyprojecttoml_dynamic_deps.py
@@ -0,0 +1,109 @@
+from inspect import cleandoc
+
+import pytest
+from jaraco import path
+
+from setuptools.config.pyprojecttoml import apply_configuration
+from setuptools.dist import Distribution
+from setuptools.warnings import SetuptoolsWarning
+
+
+def test_dynamic_dependencies(tmp_path):
+    files = {
+        "requirements.txt": "six\n  # comment\n",
+        "pyproject.toml": cleandoc(
+            """
+            [project]
+            name = "myproj"
+            version = "1.0"
+            dynamic = ["dependencies"]
+
+            [build-system]
+            requires = ["setuptools", "wheel"]
+            build-backend = "setuptools.build_meta"
+
+            [tool.setuptools.dynamic.dependencies]
+            file = ["requirements.txt"]
+            """
+        ),
+    }
+    path.build(files, prefix=tmp_path)
+    dist = Distribution()
+    dist = apply_configuration(dist, tmp_path / "pyproject.toml")
+    assert dist.install_requires == ["six"]
+
+
+def test_dynamic_optional_dependencies(tmp_path):
+    files = {
+        "requirements-docs.txt": "sphinx\n  # comment\n",
+        "pyproject.toml": cleandoc(
+            """
+            [project]
+            name = "myproj"
+            version = "1.0"
+            dynamic = ["optional-dependencies"]
+
+            [tool.setuptools.dynamic.optional-dependencies.docs]
+            file = ["requirements-docs.txt"]
+
+            [build-system]
+            requires = ["setuptools", "wheel"]
+            build-backend = "setuptools.build_meta"
+            """
+        ),
+    }
+    path.build(files, prefix=tmp_path)
+    dist = Distribution()
+    dist = apply_configuration(dist, tmp_path / "pyproject.toml")
+    assert dist.extras_require == {"docs": ["sphinx"]}
+
+
+def test_mixed_dynamic_optional_dependencies(tmp_path):
+    """
+    Test that if PEP 621 was loosened to allow mixing of dynamic and static
+    configurations in the case of fields containing sub-fields (groups),
+    things would work out.
+    """
+    files = {
+        "requirements-images.txt": "pillow~=42.0\n  # comment\n",
+        "pyproject.toml": cleandoc(
+            """
+            [project]
+            name = "myproj"
+            version = "1.0"
+            dynamic = ["optional-dependencies"]
+
+            [project.optional-dependencies]
+            docs = ["sphinx"]
+
+            [tool.setuptools.dynamic.optional-dependencies.images]
+            file = ["requirements-images.txt"]
+            """
+        ),
+    }
+
+    path.build(files, prefix=tmp_path)
+    pyproject = tmp_path / "pyproject.toml"
+    with pytest.raises(ValueError, match="project.optional-dependencies"):
+        apply_configuration(Distribution(), pyproject)
+
+
+def test_mixed_extras_require_optional_dependencies(tmp_path):
+    files = {
+        "pyproject.toml": cleandoc(
+            """
+            [project]
+            name = "myproj"
+            version = "1.0"
+            optional-dependencies.docs = ["sphinx"]
+            """
+        ),
+    }
+
+    path.build(files, prefix=tmp_path)
+    pyproject = tmp_path / "pyproject.toml"
+
+    with pytest.warns(SetuptoolsWarning, match=".extras_require. overwritten"):
+        dist = Distribution({"extras_require": {"hello": ["world"]}})
+        dist = apply_configuration(dist, pyproject)
+        assert dist.extras_require == {"docs": ["sphinx"]}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_setupcfg.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_setupcfg.py
new file mode 100644
index 00000000..adadc02d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/config/test_setupcfg.py
@@ -0,0 +1,965 @@
+import configparser
+import contextlib
+import inspect
+from pathlib import Path
+from unittest.mock import Mock, patch
+
+import pytest
+from packaging.requirements import InvalidRequirement
+
+from setuptools.config.setupcfg import ConfigHandler, Target, read_configuration
+from setuptools.dist import Distribution, _Distribution
+from setuptools.warnings import SetuptoolsDeprecationWarning
+
+from ..textwrap import DALS
+
+from distutils.errors import DistutilsFileError, DistutilsOptionError
+
+
+class ErrConfigHandler(ConfigHandler[Target]):
+    """Erroneous handler. Fails to implement required methods."""
+
+    section_prefix = "**err**"
+
+
+def make_package_dir(name, base_dir, ns=False):
+    dir_package = base_dir
+    for dir_name in name.split('/'):
+        dir_package = dir_package.mkdir(dir_name)
+    init_file = None
+    if not ns:
+        init_file = dir_package.join('__init__.py')
+        init_file.write('')
+    return dir_package, init_file
+
+
+def fake_env(
+    tmpdir, setup_cfg, setup_py=None, encoding='ascii', package_path='fake_package'
+):
+    if setup_py is None:
+        setup_py = 'from setuptools import setup\nsetup()\n'
+
+    tmpdir.join('setup.py').write(setup_py)
+    config = tmpdir.join('setup.cfg')
+    config.write(setup_cfg.encode(encoding), mode='wb')
+
+    package_dir, init_file = make_package_dir(package_path, tmpdir)
+
+    init_file.write(
+        'VERSION = (1, 2, 3)\n'
+        '\n'
+        'VERSION_MAJOR = 1'
+        '\n'
+        'def get_version():\n'
+        '    return [3, 4, 5, "dev"]\n'
+        '\n'
+    )
+
+    return package_dir, config
+
+
+@contextlib.contextmanager
+def get_dist(tmpdir, kwargs_initial=None, parse=True):
+    kwargs_initial = kwargs_initial or {}
+
+    with tmpdir.as_cwd():
+        dist = Distribution(kwargs_initial)
+        dist.script_name = 'setup.py'
+        parse and dist.parse_config_files()
+
+        yield dist
+
+
+def test_parsers_implemented():
+    with pytest.raises(NotImplementedError):
+        handler = ErrConfigHandler(None, {}, False, Mock())
+        handler.parsers
+
+
+class TestConfigurationReader:
+    def test_basic(self, tmpdir):
+        _, config = fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'version = 10.1.1\n'
+            'keywords = one, two\n'
+            '\n'
+            '[options]\n'
+            'scripts = bin/a.py, bin/b.py\n',
+        )
+        config_dict = read_configuration(str(config))
+        assert config_dict['metadata']['version'] == '10.1.1'
+        assert config_dict['metadata']['keywords'] == ['one', 'two']
+        assert config_dict['options']['scripts'] == ['bin/a.py', 'bin/b.py']
+
+    def test_no_config(self, tmpdir):
+        with pytest.raises(DistutilsFileError):
+            read_configuration(str(tmpdir.join('setup.cfg')))
+
+    def test_ignore_errors(self, tmpdir):
+        _, config = fake_env(
+            tmpdir,
+            '[metadata]\nversion = attr: none.VERSION\nkeywords = one, two\n',
+        )
+        with pytest.raises(ImportError):
+            read_configuration(str(config))
+
+        config_dict = read_configuration(str(config), ignore_option_errors=True)
+
+        assert config_dict['metadata']['keywords'] == ['one', 'two']
+        assert 'version' not in config_dict['metadata']
+
+        config.remove()
+
+
+class TestMetadata:
+    def test_basic(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'version = 10.1.1\n'
+            'description = Some description\n'
+            'long_description_content_type = text/something\n'
+            'long_description = file: README\n'
+            'name = fake_name\n'
+            'keywords = one, two\n'
+            'provides = package, package.sub\n'
+            'license = otherlic\n'
+            'download_url = http://test.test.com/test/\n'
+            'maintainer_email = test@test.com\n',
+        )
+
+        tmpdir.join('README').write('readme contents\nline2')
+
+        meta_initial = {
+            # This will be used so `otherlic` won't replace it.
+            'license': 'BSD 3-Clause License',
+        }
+
+        with get_dist(tmpdir, meta_initial) as dist:
+            metadata = dist.metadata
+
+            assert metadata.version == '10.1.1'
+            assert metadata.description == 'Some description'
+            assert metadata.long_description_content_type == 'text/something'
+            assert metadata.long_description == 'readme contents\nline2'
+            assert metadata.provides == ['package', 'package.sub']
+            assert metadata.license == 'BSD 3-Clause License'
+            assert metadata.name == 'fake_name'
+            assert metadata.keywords == ['one', 'two']
+            assert metadata.download_url == 'http://test.test.com/test/'
+            assert metadata.maintainer_email == 'test@test.com'
+
+    def test_license_cfg(self, tmpdir):
+        fake_env(
+            tmpdir,
+            DALS(
+                """
+            [metadata]
+            name=foo
+            version=0.0.1
+            license=Apache 2.0
+            """
+            ),
+        )
+
+        with get_dist(tmpdir) as dist:
+            metadata = dist.metadata
+
+            assert metadata.name == "foo"
+            assert metadata.version == "0.0.1"
+            assert metadata.license == "Apache 2.0"
+
+    def test_file_mixed(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\nlong_description = file: README.rst, CHANGES.rst\n\n',
+        )
+
+        tmpdir.join('README.rst').write('readme contents\nline2')
+        tmpdir.join('CHANGES.rst').write('changelog contents\nand stuff')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.long_description == (
+                'readme contents\nline2\nchangelog contents\nand stuff'
+            )
+
+    def test_file_sandboxed(self, tmpdir):
+        tmpdir.ensure("README")
+        project = tmpdir.join('depth1', 'depth2')
+        project.ensure(dir=True)
+        fake_env(project, '[metadata]\nlong_description = file: ../../README\n')
+
+        with get_dist(project, parse=False) as dist:
+            with pytest.raises(DistutilsOptionError):
+                dist.parse_config_files()  # file: out of sandbox
+
+    def test_aliases(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'author_email = test@test.com\n'
+            'home_page = http://test.test.com/test/\n'
+            'summary = Short summary\n'
+            'platform = a, b\n'
+            'classifier =\n'
+            '  Framework :: Django\n'
+            '  Programming Language :: Python :: 3.5\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            metadata = dist.metadata
+            assert metadata.author_email == 'test@test.com'
+            assert metadata.url == 'http://test.test.com/test/'
+            assert metadata.description == 'Short summary'
+            assert metadata.platforms == ['a', 'b']
+            assert metadata.classifiers == [
+                'Framework :: Django',
+                'Programming Language :: Python :: 3.5',
+            ]
+
+    def test_multiline(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'name = fake_name\n'
+            'keywords =\n'
+            '  one\n'
+            '  two\n'
+            'classifiers =\n'
+            '  Framework :: Django\n'
+            '  Programming Language :: Python :: 3.5\n',
+        )
+        with get_dist(tmpdir) as dist:
+            metadata = dist.metadata
+            assert metadata.keywords == ['one', 'two']
+            assert metadata.classifiers == [
+                'Framework :: Django',
+                'Programming Language :: Python :: 3.5',
+            ]
+
+    def test_dict(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'project_urls =\n'
+            '  Link One = https://example.com/one/\n'
+            '  Link Two = https://example.com/two/\n',
+        )
+        with get_dist(tmpdir) as dist:
+            metadata = dist.metadata
+            assert metadata.project_urls == {
+                'Link One': 'https://example.com/one/',
+                'Link Two': 'https://example.com/two/',
+            }
+
+    def test_version(self, tmpdir):
+        package_dir, config = fake_env(
+            tmpdir, '[metadata]\nversion = attr: fake_package.VERSION\n'
+        )
+
+        sub_a = package_dir.mkdir('subpkg_a')
+        sub_a.join('__init__.py').write('')
+        sub_a.join('mod.py').write('VERSION = (2016, 11, 26)')
+
+        sub_b = package_dir.mkdir('subpkg_b')
+        sub_b.join('__init__.py').write('')
+        sub_b.join('mod.py').write(
+            'import third_party_module\nVERSION = (2016, 11, 26)'
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1.2.3'
+
+        config.write('[metadata]\nversion = attr: fake_package.get_version\n')
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '3.4.5.dev'
+
+        config.write('[metadata]\nversion = attr: fake_package.VERSION_MAJOR\n')
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1'
+
+        config.write('[metadata]\nversion = attr: fake_package.subpkg_a.mod.VERSION\n')
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '2016.11.26'
+
+        config.write('[metadata]\nversion = attr: fake_package.subpkg_b.mod.VERSION\n')
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '2016.11.26'
+
+    def test_version_file(self, tmpdir):
+        fake_env(tmpdir, '[metadata]\nversion = file: fake_package/version.txt\n')
+        tmpdir.join('fake_package', 'version.txt').write('1.2.3\n')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1.2.3'
+
+        tmpdir.join('fake_package', 'version.txt').write('1.2.3\n4.5.6\n')
+        with pytest.raises(DistutilsOptionError):
+            with get_dist(tmpdir) as dist:
+                dist.metadata.version
+
+    def test_version_with_package_dir_simple(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'version = attr: fake_package_simple.VERSION\n'
+            '[options]\n'
+            'package_dir =\n'
+            '    = src\n',
+            package_path='src/fake_package_simple',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1.2.3'
+
+    def test_version_with_package_dir_rename(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'version = attr: fake_package_rename.VERSION\n'
+            '[options]\n'
+            'package_dir =\n'
+            '    fake_package_rename = fake_dir\n',
+            package_path='fake_dir',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1.2.3'
+
+    def test_version_with_package_dir_complex(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'version = attr: fake_package_complex.VERSION\n'
+            '[options]\n'
+            'package_dir =\n'
+            '    fake_package_complex = src/fake_dir\n',
+            package_path='src/fake_dir',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.version == '1.2.3'
+
+    def test_unknown_meta_item(self, tmpdir):
+        fake_env(tmpdir, '[metadata]\nname = fake_name\nunknown = some\n')
+        with get_dist(tmpdir, parse=False) as dist:
+            dist.parse_config_files()  # Skip unknown.
+
+    def test_usupported_section(self, tmpdir):
+        fake_env(tmpdir, '[metadata.some]\nkey = val\n')
+        with get_dist(tmpdir, parse=False) as dist:
+            with pytest.raises(DistutilsOptionError):
+                dist.parse_config_files()
+
+    def test_classifiers(self, tmpdir):
+        expected = set([
+            'Framework :: Django',
+            'Programming Language :: Python :: 3',
+            'Programming Language :: Python :: 3.5',
+        ])
+
+        # From file.
+        _, config = fake_env(tmpdir, '[metadata]\nclassifiers = file: classifiers\n')
+
+        tmpdir.join('classifiers').write(
+            'Framework :: Django\n'
+            'Programming Language :: Python :: 3\n'
+            'Programming Language :: Python :: 3.5\n'
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert set(dist.metadata.classifiers) == expected
+
+        # From list notation
+        config.write(
+            '[metadata]\n'
+            'classifiers =\n'
+            '    Framework :: Django\n'
+            '    Programming Language :: Python :: 3\n'
+            '    Programming Language :: Python :: 3.5\n'
+        )
+        with get_dist(tmpdir) as dist:
+            assert set(dist.metadata.classifiers) == expected
+
+    def test_interpolation(self, tmpdir):
+        fake_env(tmpdir, '[metadata]\ndescription = %(message)s\n')
+        with pytest.raises(configparser.InterpolationMissingOptionError):
+            with get_dist(tmpdir):
+                pass
+
+    def test_non_ascii_1(self, tmpdir):
+        fake_env(tmpdir, '[metadata]\ndescription = éàïôñ\n', encoding='utf-8')
+        with get_dist(tmpdir):
+            pass
+
+    def test_non_ascii_3(self, tmpdir):
+        fake_env(tmpdir, '\n# -*- coding: invalid\n')
+        with get_dist(tmpdir):
+            pass
+
+    def test_non_ascii_4(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '# -*- coding: utf-8\n[metadata]\ndescription = éàïôñ\n',
+            encoding='utf-8',
+        )
+        with get_dist(tmpdir) as dist:
+            assert dist.metadata.description == 'éàïôñ'
+
+    def test_not_utf8(self, tmpdir):
+        """
+        Config files encoded not in UTF-8 will fail
+        """
+        fake_env(
+            tmpdir,
+            '# vim: set fileencoding=iso-8859-15 :\n[metadata]\ndescription = éàïôñ\n',
+            encoding='iso-8859-15',
+        )
+        with pytest.raises(UnicodeDecodeError):
+            with get_dist(tmpdir):
+                pass
+
+    def test_warn_dash_deprecation(self, tmpdir):
+        # warn_dash_deprecation() is a method in setuptools.dist
+        # remove this test and the method when no longer needed
+        fake_env(
+            tmpdir,
+            '[metadata]\n'
+            'author-email = test@test.com\n'
+            'maintainer_email = foo@foo.com\n',
+        )
+        msg = "Usage of dash-separated 'author-email' will not be supported"
+        with pytest.warns(SetuptoolsDeprecationWarning, match=msg):
+            with get_dist(tmpdir) as dist:
+                metadata = dist.metadata
+
+        assert metadata.author_email == 'test@test.com'
+        assert metadata.maintainer_email == 'foo@foo.com'
+
+    def test_make_option_lowercase(self, tmpdir):
+        # remove this test and the method make_option_lowercase() in setuptools.dist
+        # when no longer needed
+        fake_env(tmpdir, '[metadata]\nName = foo\ndescription = Some description\n')
+        msg = "Usage of uppercase key 'Name' in 'metadata' will not be supported"
+        with pytest.warns(SetuptoolsDeprecationWarning, match=msg):
+            with get_dist(tmpdir) as dist:
+                metadata = dist.metadata
+
+        assert metadata.name == 'foo'
+        assert metadata.description == 'Some description'
+
+
+class TestOptions:
+    def test_basic(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options]\n'
+            'zip_safe = True\n'
+            'include_package_data = yes\n'
+            'package_dir = b=c, =src\n'
+            'packages = pack_a, pack_b.subpack\n'
+            'namespace_packages = pack1, pack2\n'
+            'scripts = bin/one.py, bin/two.py\n'
+            'eager_resources = bin/one.py, bin/two.py\n'
+            'install_requires = docutils>=0.3; pack ==1.1, ==1.3; hey\n'
+            'setup_requires = docutils>=0.3; spack ==1.1, ==1.3; there\n'
+            'dependency_links = http://some.com/here/1, '
+            'http://some.com/there/2\n'
+            'python_requires = >=1.0, !=2.8\n'
+            'py_modules = module1, module2\n',
+        )
+        deprec = pytest.warns(SetuptoolsDeprecationWarning, match="namespace_packages")
+        with deprec, get_dist(tmpdir) as dist:
+            assert dist.zip_safe
+            assert dist.include_package_data
+            assert dist.package_dir == {'': 'src', 'b': 'c'}
+            assert dist.packages == ['pack_a', 'pack_b.subpack']
+            assert dist.namespace_packages == ['pack1', 'pack2']
+            assert dist.scripts == ['bin/one.py', 'bin/two.py']
+            assert dist.dependency_links == ([
+                'http://some.com/here/1',
+                'http://some.com/there/2',
+            ])
+            assert dist.install_requires == ([
+                'docutils>=0.3',
+                'pack==1.1,==1.3',
+                'hey',
+            ])
+            assert dist.setup_requires == ([
+                'docutils>=0.3',
+                'spack ==1.1, ==1.3',
+                'there',
+            ])
+            assert dist.python_requires == '>=1.0, !=2.8'
+            assert dist.py_modules == ['module1', 'module2']
+
+    def test_multiline(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options]\n'
+            'package_dir = \n'
+            '  b=c\n'
+            '  =src\n'
+            'packages = \n'
+            '  pack_a\n'
+            '  pack_b.subpack\n'
+            'namespace_packages = \n'
+            '  pack1\n'
+            '  pack2\n'
+            'scripts = \n'
+            '  bin/one.py\n'
+            '  bin/two.py\n'
+            'eager_resources = \n'
+            '  bin/one.py\n'
+            '  bin/two.py\n'
+            'install_requires = \n'
+            '  docutils>=0.3\n'
+            '  pack ==1.1, ==1.3\n'
+            '  hey\n'
+            'setup_requires = \n'
+            '  docutils>=0.3\n'
+            '  spack ==1.1, ==1.3\n'
+            '  there\n'
+            'dependency_links = \n'
+            '  http://some.com/here/1\n'
+            '  http://some.com/there/2\n',
+        )
+        deprec = pytest.warns(SetuptoolsDeprecationWarning, match="namespace_packages")
+        with deprec, get_dist(tmpdir) as dist:
+            assert dist.package_dir == {'': 'src', 'b': 'c'}
+            assert dist.packages == ['pack_a', 'pack_b.subpack']
+            assert dist.namespace_packages == ['pack1', 'pack2']
+            assert dist.scripts == ['bin/one.py', 'bin/two.py']
+            assert dist.dependency_links == ([
+                'http://some.com/here/1',
+                'http://some.com/there/2',
+            ])
+            assert dist.install_requires == ([
+                'docutils>=0.3',
+                'pack==1.1,==1.3',
+                'hey',
+            ])
+            assert dist.setup_requires == ([
+                'docutils>=0.3',
+                'spack ==1.1, ==1.3',
+                'there',
+            ])
+
+    def test_package_dir_fail(self, tmpdir):
+        fake_env(tmpdir, '[options]\npackage_dir = a b\n')
+        with get_dist(tmpdir, parse=False) as dist:
+            with pytest.raises(DistutilsOptionError):
+                dist.parse_config_files()
+
+    def test_package_data(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options.package_data]\n'
+            '* = *.txt, *.rst\n'
+            'hello = *.msg\n'
+            '\n'
+            '[options.exclude_package_data]\n'
+            '* = fake1.txt, fake2.txt\n'
+            'hello = *.dat\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.package_data == {
+                '': ['*.txt', '*.rst'],
+                'hello': ['*.msg'],
+            }
+            assert dist.exclude_package_data == {
+                '': ['fake1.txt', 'fake2.txt'],
+                'hello': ['*.dat'],
+            }
+
+    def test_packages(self, tmpdir):
+        fake_env(tmpdir, '[options]\npackages = find:\n')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.packages == ['fake_package']
+
+    def test_find_directive(self, tmpdir):
+        dir_package, config = fake_env(tmpdir, '[options]\npackages = find:\n')
+
+        make_package_dir('sub_one', dir_package)
+        make_package_dir('sub_two', dir_package)
+
+        with get_dist(tmpdir) as dist:
+            assert set(dist.packages) == set([
+                'fake_package',
+                'fake_package.sub_two',
+                'fake_package.sub_one',
+            ])
+
+        config.write(
+            '[options]\n'
+            'packages = find:\n'
+            '\n'
+            '[options.packages.find]\n'
+            'where = .\n'
+            'include =\n'
+            '    fake_package.sub_one\n'
+            '    two\n'
+        )
+        with get_dist(tmpdir) as dist:
+            assert dist.packages == ['fake_package.sub_one']
+
+        config.write(
+            '[options]\n'
+            'packages = find:\n'
+            '\n'
+            '[options.packages.find]\n'
+            'exclude =\n'
+            '    fake_package.sub_one\n'
+        )
+        with get_dist(tmpdir) as dist:
+            assert set(dist.packages) == set(['fake_package', 'fake_package.sub_two'])
+
+    def test_find_namespace_directive(self, tmpdir):
+        dir_package, config = fake_env(
+            tmpdir, '[options]\npackages = find_namespace:\n'
+        )
+
+        make_package_dir('sub_one', dir_package)
+        make_package_dir('sub_two', dir_package, ns=True)
+
+        with get_dist(tmpdir) as dist:
+            assert set(dist.packages) == {
+                'fake_package',
+                'fake_package.sub_two',
+                'fake_package.sub_one',
+            }
+
+        config.write(
+            '[options]\n'
+            'packages = find_namespace:\n'
+            '\n'
+            '[options.packages.find]\n'
+            'where = .\n'
+            'include =\n'
+            '    fake_package.sub_one\n'
+            '    two\n'
+        )
+        with get_dist(tmpdir) as dist:
+            assert dist.packages == ['fake_package.sub_one']
+
+        config.write(
+            '[options]\n'
+            'packages = find_namespace:\n'
+            '\n'
+            '[options.packages.find]\n'
+            'exclude =\n'
+            '    fake_package.sub_one\n'
+        )
+        with get_dist(tmpdir) as dist:
+            assert set(dist.packages) == {'fake_package', 'fake_package.sub_two'}
+
+    def test_extras_require(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options.extras_require]\n'
+            'pdf = ReportLab>=1.2; RXP\n'
+            'rest = \n'
+            '  docutils>=0.3\n'
+            '  pack ==1.1, ==1.3\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.extras_require == {
+                'pdf': ['ReportLab>=1.2', 'RXP'],
+                'rest': ['docutils>=0.3', 'pack==1.1,==1.3'],
+            }
+            assert set(dist.metadata.provides_extras) == {'pdf', 'rest'}
+
+    @pytest.mark.parametrize(
+        "config",
+        [
+            "[options.extras_require]\nfoo = bar;python_version<'3'",
+            "[options.extras_require]\nfoo = bar;os_name=='linux'",
+            "[options.extras_require]\nfoo = bar;python_version<'3'\n",
+            "[options.extras_require]\nfoo = bar;os_name=='linux'\n",
+            "[options]\ninstall_requires = bar;python_version<'3'",
+            "[options]\ninstall_requires = bar;os_name=='linux'",
+            "[options]\ninstall_requires = bar;python_version<'3'\n",
+            "[options]\ninstall_requires = bar;os_name=='linux'\n",
+        ],
+    )
+    def test_raises_accidental_env_marker_misconfig(self, config, tmpdir):
+        fake_env(tmpdir, config)
+        match = (
+            r"One of the parsed requirements in `(install_requires|extras_require.+)` "
+            "looks like a valid environment marker.*"
+        )
+        with pytest.raises(InvalidRequirement, match=match):
+            with get_dist(tmpdir) as _:
+                pass
+
+    @pytest.mark.parametrize(
+        "config",
+        [
+            "[options.extras_require]\nfoo = bar;python_version<3",
+            "[options.extras_require]\nfoo = bar;python_version<3\n",
+            "[options]\ninstall_requires = bar;python_version<3",
+            "[options]\ninstall_requires = bar;python_version<3\n",
+        ],
+    )
+    def test_warn_accidental_env_marker_misconfig(self, config, tmpdir):
+        fake_env(tmpdir, config)
+        match = (
+            r"One of the parsed requirements in `(install_requires|extras_require.+)` "
+            "looks like a valid environment marker.*"
+        )
+        with pytest.warns(SetuptoolsDeprecationWarning, match=match):
+            with get_dist(tmpdir) as _:
+                pass
+
+    @pytest.mark.parametrize(
+        "config",
+        [
+            "[options.extras_require]\nfoo =\n    bar;python_version<'3'",
+            "[options.extras_require]\nfoo = bar;baz\nboo = xxx;yyy",
+            "[options.extras_require]\nfoo =\n    bar;python_version<'3'\n",
+            "[options.extras_require]\nfoo = bar;baz\nboo = xxx;yyy\n",
+            "[options.extras_require]\nfoo =\n    bar\n    python_version<3\n",
+            "[options]\ninstall_requires =\n    bar;python_version<'3'",
+            "[options]\ninstall_requires = bar;baz\nboo = xxx;yyy",
+            "[options]\ninstall_requires =\n    bar;python_version<'3'\n",
+            "[options]\ninstall_requires = bar;baz\nboo = xxx;yyy\n",
+            "[options]\ninstall_requires =\n    bar\n    python_version<3\n",
+        ],
+    )
+    @pytest.mark.filterwarnings("error::setuptools.SetuptoolsDeprecationWarning")
+    def test_nowarn_accidental_env_marker_misconfig(self, config, tmpdir, recwarn):
+        fake_env(tmpdir, config)
+        num_warnings = len(recwarn)
+        with get_dist(tmpdir) as _:
+            pass
+        # The examples are valid, no warnings shown
+        assert len(recwarn) == num_warnings
+
+    def test_dash_preserved_extras_require(self, tmpdir):
+        fake_env(tmpdir, '[options.extras_require]\nfoo-a = foo\nfoo_b = test\n')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.extras_require == {'foo-a': ['foo'], 'foo_b': ['test']}
+
+    def test_entry_points(self, tmpdir):
+        _, config = fake_env(
+            tmpdir,
+            '[options.entry_points]\n'
+            'group1 = point1 = pack.module:func, '
+            '.point2 = pack.module2:func_rest [rest]\n'
+            'group2 = point3 = pack.module:func2\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.entry_points == {
+                'group1': [
+                    'point1 = pack.module:func',
+                    '.point2 = pack.module2:func_rest [rest]',
+                ],
+                'group2': ['point3 = pack.module:func2'],
+            }
+
+        expected = (
+            '[blogtool.parsers]\n'
+            '.rst = some.nested.module:SomeClass.some_classmethod[reST]\n'
+        )
+
+        tmpdir.join('entry_points').write(expected)
+
+        # From file.
+        config.write('[options]\nentry_points = file: entry_points\n')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.entry_points == expected
+
+    def test_case_sensitive_entry_points(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options.entry_points]\n'
+            'GROUP1 = point1 = pack.module:func, '
+            '.point2 = pack.module2:func_rest [rest]\n'
+            'group2 = point3 = pack.module:func2\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            assert dist.entry_points == {
+                'GROUP1': [
+                    'point1 = pack.module:func',
+                    '.point2 = pack.module2:func_rest [rest]',
+                ],
+                'group2': ['point3 = pack.module:func2'],
+            }
+
+    def test_data_files(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options.data_files]\n'
+            'cfg =\n'
+            '      a/b.conf\n'
+            '      c/d.conf\n'
+            'data = e/f.dat, g/h.dat\n',
+        )
+
+        with get_dist(tmpdir) as dist:
+            expected = [
+                ('cfg', ['a/b.conf', 'c/d.conf']),
+                ('data', ['e/f.dat', 'g/h.dat']),
+            ]
+            assert sorted(dist.data_files) == sorted(expected)
+
+    def test_data_files_globby(self, tmpdir):
+        fake_env(
+            tmpdir,
+            '[options.data_files]\n'
+            'cfg =\n'
+            '      a/b.conf\n'
+            '      c/d.conf\n'
+            'data = *.dat\n'
+            'icons = \n'
+            '      *.ico\n'
+            'audio = \n'
+            '      *.wav\n'
+            '      sounds.db\n',
+        )
+
+        # Create dummy files for glob()'s sake:
+        tmpdir.join('a.dat').write('')
+        tmpdir.join('b.dat').write('')
+        tmpdir.join('c.dat').write('')
+        tmpdir.join('a.ico').write('')
+        tmpdir.join('b.ico').write('')
+        tmpdir.join('c.ico').write('')
+        tmpdir.join('beep.wav').write('')
+        tmpdir.join('boop.wav').write('')
+        tmpdir.join('sounds.db').write('')
+
+        with get_dist(tmpdir) as dist:
+            expected = [
+                ('cfg', ['a/b.conf', 'c/d.conf']),
+                ('data', ['a.dat', 'b.dat', 'c.dat']),
+                ('icons', ['a.ico', 'b.ico', 'c.ico']),
+                ('audio', ['beep.wav', 'boop.wav', 'sounds.db']),
+            ]
+            assert sorted(dist.data_files) == sorted(expected)
+
+    def test_python_requires_simple(self, tmpdir):
+        fake_env(
+            tmpdir,
+            DALS(
+                """
+            [options]
+            python_requires=>=2.7
+            """
+            ),
+        )
+        with get_dist(tmpdir) as dist:
+            dist.parse_config_files()
+
+    def test_python_requires_compound(self, tmpdir):
+        fake_env(
+            tmpdir,
+            DALS(
+                """
+            [options]
+            python_requires=>=2.7,!=3.0.*
+            """
+            ),
+        )
+        with get_dist(tmpdir) as dist:
+            dist.parse_config_files()
+
+    def test_python_requires_invalid(self, tmpdir):
+        fake_env(
+            tmpdir,
+            DALS(
+                """
+            [options]
+            python_requires=invalid
+            """
+            ),
+        )
+        with pytest.raises(Exception):
+            with get_dist(tmpdir) as dist:
+                dist.parse_config_files()
+
+    def test_cmdclass(self, tmpdir):
+        module_path = Path(tmpdir, "src/custom_build.py")  # auto discovery for src
+        module_path.parent.mkdir(parents=True, exist_ok=True)
+        module_path.write_text(
+            "from distutils.core import Command\nclass CustomCmd(Command): pass\n",
+            encoding="utf-8",
+        )
+
+        setup_cfg = """
+            [options]
+            cmdclass =
+                customcmd = custom_build.CustomCmd
+        """
+        fake_env(tmpdir, inspect.cleandoc(setup_cfg))
+
+        with get_dist(tmpdir) as dist:
+            cmdclass = dist.cmdclass['customcmd']
+            assert cmdclass.__name__ == "CustomCmd"
+            assert cmdclass.__module__ == "custom_build"
+            assert module_path.samefile(inspect.getfile(cmdclass))
+
+    def test_requirements_file(self, tmpdir):
+        fake_env(
+            tmpdir,
+            DALS(
+                """
+            [options]
+            install_requires = file:requirements.txt
+            [options.extras_require]
+            colors = file:requirements-extra.txt
+            """
+            ),
+        )
+
+        tmpdir.join('requirements.txt').write('\ndocutils>=0.3\n\n')
+        tmpdir.join('requirements-extra.txt').write('colorama')
+
+        with get_dist(tmpdir) as dist:
+            assert dist.install_requires == ['docutils>=0.3']
+            assert dist.extras_require == {'colors': ['colorama']}
+
+
+saved_dist_init = _Distribution.__init__
+
+
+class TestExternalSetters:
+    # During creation of the setuptools Distribution() object, we call
+    # the init of the parent distutils Distribution object via
+    # _Distribution.__init__ ().
+    #
+    # It's possible distutils calls out to various keyword
+    # implementations (i.e. distutils.setup_keywords entry points)
+    # that may set a range of variables.
+    #
+    # This wraps distutil's Distribution.__init__ and simulates
+    # pbr or something else setting these values.
+    def _fake_distribution_init(self, dist, attrs):
+        saved_dist_init(dist, attrs)
+        # see self._DISTUTILS_UNSUPPORTED_METADATA
+        dist.metadata.long_description_content_type = 'text/something'
+        # Test overwrite setup() args
+        dist.metadata.project_urls = {
+            'Link One': 'https://example.com/one/',
+            'Link Two': 'https://example.com/two/',
+        }
+
+    @patch.object(_Distribution, '__init__', autospec=True)
+    def test_external_setters(self, mock_parent_init, tmpdir):
+        mock_parent_init.side_effect = self._fake_distribution_init
+
+        dist = Distribution(attrs={'project_urls': {'will_be': 'ignored'}})
+
+        assert dist.metadata.long_description_content_type == 'text/something'
+        assert dist.metadata.project_urls == {
+            'Link One': 'https://example.com/one/',
+            'Link Two': 'https://example.com/two/',
+        }
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/contexts.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/contexts.py
new file mode 100644
index 00000000..97cceea0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/contexts.py
@@ -0,0 +1,145 @@
+import contextlib
+import io
+import os
+import shutil
+import site
+import sys
+import tempfile
+
+from filelock import FileLock
+
+
+@contextlib.contextmanager
+def tempdir(cd=lambda dir: None, **kwargs):
+    temp_dir = tempfile.mkdtemp(**kwargs)
+    orig_dir = os.getcwd()
+    try:
+        cd(temp_dir)
+        yield temp_dir
+    finally:
+        cd(orig_dir)
+        shutil.rmtree(temp_dir)
+
+
+@contextlib.contextmanager
+def environment(**replacements):
+    """
+    In a context, patch the environment with replacements. Pass None values
+    to clear the values.
+    """
+    saved = dict((key, os.environ[key]) for key in replacements if key in os.environ)
+
+    # remove values that are null
+    remove = (key for (key, value) in replacements.items() if value is None)
+    for key in list(remove):
+        os.environ.pop(key, None)
+        replacements.pop(key)
+
+    os.environ.update(replacements)
+
+    try:
+        yield saved
+    finally:
+        for key in replacements:
+            os.environ.pop(key, None)
+        os.environ.update(saved)
+
+
+@contextlib.contextmanager
+def quiet():
+    """
+    Redirect stdout/stderr to StringIO objects to prevent console output from
+    distutils commands.
+    """
+
+    old_stdout = sys.stdout
+    old_stderr = sys.stderr
+    new_stdout = sys.stdout = io.StringIO()
+    new_stderr = sys.stderr = io.StringIO()
+    try:
+        yield new_stdout, new_stderr
+    finally:
+        new_stdout.seek(0)
+        new_stderr.seek(0)
+        sys.stdout = old_stdout
+        sys.stderr = old_stderr
+
+
+@contextlib.contextmanager
+def save_user_site_setting():
+    saved = site.ENABLE_USER_SITE
+    try:
+        yield saved
+    finally:
+        site.ENABLE_USER_SITE = saved
+
+
+@contextlib.contextmanager
+def save_pkg_resources_state():
+    import pkg_resources
+
+    pr_state = pkg_resources.__getstate__()
+    # also save sys.path
+    sys_path = sys.path[:]
+    try:
+        yield pr_state, sys_path
+    finally:
+        sys.path[:] = sys_path
+        pkg_resources.__setstate__(pr_state)
+
+
+@contextlib.contextmanager
+def suppress_exceptions(*excs):
+    try:
+        yield
+    except excs:
+        pass
+
+
+def multiproc(request):
+    """
+    Return True if running under xdist and multiple
+    workers are used.
+    """
+    try:
+        worker_id = request.getfixturevalue('worker_id')
+    except Exception:
+        return False
+    return worker_id != 'master'
+
+
+@contextlib.contextmanager
+def session_locked_tmp_dir(request, tmp_path_factory, name):
+    """Uses a file lock to guarantee only one worker can access a temp dir"""
+    # get the temp directory shared by all workers
+    base = tmp_path_factory.getbasetemp()
+    shared_dir = base.parent if multiproc(request) else base
+
+    locked_dir = shared_dir / name
+    with FileLock(locked_dir.with_suffix(".lock")):
+        # ^-- prevent multiple workers to access the directory at once
+        locked_dir.mkdir(exist_ok=True, parents=True)
+        yield locked_dir
+
+
+@contextlib.contextmanager
+def save_paths():
+    """Make sure ``sys.path``, ``sys.meta_path`` and ``sys.path_hooks`` are preserved"""
+    prev = sys.path[:], sys.meta_path[:], sys.path_hooks[:]
+
+    try:
+        yield
+    finally:
+        sys.path, sys.meta_path, sys.path_hooks = prev
+
+
+@contextlib.contextmanager
+def save_sys_modules():
+    """Make sure initial ``sys.modules`` is preserved"""
+    prev_modules = sys.modules
+
+    try:
+        sys.modules = sys.modules.copy()
+        yield
+    finally:
+        sys.modules = prev_modules
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/environment.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/environment.py
new file mode 100644
index 00000000..ed5499ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/environment.py
@@ -0,0 +1,95 @@
+import os
+import subprocess
+import sys
+import unicodedata
+from subprocess import PIPE as _PIPE, Popen as _Popen
+
+import jaraco.envs
+
+
+class VirtualEnv(jaraco.envs.VirtualEnv):
+    name = '.env'
+    # Some version of PyPy will import distutils on startup, implicitly
+    # importing setuptools, and thus leading to BackendInvalid errors
+    # when upgrading Setuptools. Bypass this behavior by avoiding the
+    # early availability and need to upgrade.
+    create_opts = ['--no-setuptools']
+
+    def run(self, cmd, *args, **kwargs):
+        cmd = [self.exe(cmd[0])] + cmd[1:]
+        kwargs = {"cwd": self.root, "encoding": "utf-8", **kwargs}  # Allow overriding
+        # In some environments (eg. downstream distro packaging), where:
+        # - tox isn't used to run tests and
+        # - PYTHONPATH is set to point to a specific setuptools codebase and
+        # - no custom env is explicitly set by a test
+        # PYTHONPATH will leak into the spawned processes.
+        # In that case tests look for module in the wrong place (on PYTHONPATH).
+        # Unless the test sets its own special env, pass a copy of the existing
+        # environment with removed PYTHONPATH to the subprocesses.
+        if "env" not in kwargs:
+            env = dict(os.environ)
+            if "PYTHONPATH" in env:
+                del env["PYTHONPATH"]
+            kwargs["env"] = env
+        return subprocess.check_output(cmd, *args, **kwargs)
+
+
+def _which_dirs(cmd):
+    result = set()
+    for path in os.environ.get('PATH', '').split(os.pathsep):
+        filename = os.path.join(path, cmd)
+        if os.access(filename, os.X_OK):
+            result.add(path)
+    return result
+
+
+def run_setup_py(cmd, pypath=None, path=None, data_stream=0, env=None):
+    """
+    Execution command for tests, separate from those used by the
+    code directly to prevent accidental behavior issues
+    """
+    if env is None:
+        env = dict()
+        for envname in os.environ:
+            env[envname] = os.environ[envname]
+
+    # override the python path if needed
+    if pypath is not None:
+        env["PYTHONPATH"] = pypath
+
+    # override the execution path if needed
+    if path is not None:
+        env["PATH"] = path
+    if not env.get("PATH", ""):
+        env["PATH"] = _which_dirs("tar").union(_which_dirs("gzip"))
+        env["PATH"] = os.pathsep.join(env["PATH"])
+
+    cmd = [sys.executable, "setup.py"] + list(cmd)
+
+    # https://bugs.python.org/issue8557
+    shell = sys.platform == 'win32'
+
+    try:
+        proc = _Popen(
+            cmd,
+            stdout=_PIPE,
+            stderr=_PIPE,
+            shell=shell,
+            env=env,
+            encoding="utf-8",
+        )
+
+        if isinstance(data_stream, tuple):
+            data_stream = slice(*data_stream)
+        data = proc.communicate()[data_stream]
+    except OSError:
+        return 1, ''
+
+    # decode the console string if needed
+    if hasattr(data, "decode"):
+        # use the default encoding
+        data = data.decode()
+        data = unicodedata.normalize('NFC', data)
+
+    # communicate calls wait()
+    return proc.returncode, data
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/fixtures.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/fixtures.py
new file mode 100644
index 00000000..a5472984
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/fixtures.py
@@ -0,0 +1,157 @@
+import contextlib
+import os
+import subprocess
+import sys
+from pathlib import Path
+
+import path
+import pytest
+
+from . import contexts, environment
+
+
+@pytest.fixture
+def user_override(monkeypatch):
+    """
+    Override site.USER_BASE and site.USER_SITE with temporary directories in
+    a context.
+    """
+    with contexts.tempdir() as user_base:
+        monkeypatch.setattr('site.USER_BASE', user_base)
+        with contexts.tempdir() as user_site:
+            monkeypatch.setattr('site.USER_SITE', user_site)
+            with contexts.save_user_site_setting():
+                yield
+
+
+@pytest.fixture
+def tmpdir_cwd(tmpdir):
+    with tmpdir.as_cwd() as orig:
+        yield orig
+
+
+@pytest.fixture(autouse=True, scope="session")
+def workaround_xdist_376(request):
+    """
+    Workaround pytest-dev/pytest-xdist#376
+
+    ``pytest-xdist`` tends to inject '' into ``sys.path``,
+    which may break certain isolation expectations.
+    Remove the entry so the import
+    machinery behaves the same irrespective of xdist.
+    """
+    if not request.config.pluginmanager.has_plugin('xdist'):
+        return
+
+    with contextlib.suppress(ValueError):
+        sys.path.remove('')
+
+
+@pytest.fixture
+def sample_project(tmp_path):
+    """
+    Clone the 'sampleproject' and return a path to it.
+    """
+    cmd = ['git', 'clone', 'https://github.com/pypa/sampleproject']
+    try:
+        subprocess.check_call(cmd, cwd=str(tmp_path))
+    except Exception:
+        pytest.skip("Unable to clone sampleproject")
+    return tmp_path / 'sampleproject'
+
+
+# sdist and wheel artifacts should be stable across a round of tests
+# so we can build them once per session and use the files as "readonly"
+
+# In the case of setuptools, building the wheel without sdist may cause
+# it to contain the `build` directory, and therefore create situations with
+# `setuptools/build/lib/build/lib/...`. To avoid that, build both artifacts at once.
+
+
+def _build_distributions(tmp_path_factory, request):
+    with contexts.session_locked_tmp_dir(
+        request, tmp_path_factory, "dist_build"
+    ) as tmp:  # pragma: no cover
+        sdist = next(tmp.glob("*.tar.gz"), None)
+        wheel = next(tmp.glob("*.whl"), None)
+        if sdist and wheel:
+            return (sdist, wheel)
+
+        # Sanity check: should not create recursive setuptools/build/lib/build/lib/...
+        assert not Path(request.config.rootdir, "build/lib/build").exists()
+
+        subprocess.check_output([
+            sys.executable,
+            "-m",
+            "build",
+            "--outdir",
+            str(tmp),
+            str(request.config.rootdir),
+        ])
+
+        # Sanity check: should not create recursive setuptools/build/lib/build/lib/...
+        assert not Path(request.config.rootdir, "build/lib/build").exists()
+
+        return next(tmp.glob("*.tar.gz")), next(tmp.glob("*.whl"))
+
+
+@pytest.fixture(scope="session")
+def setuptools_sdist(tmp_path_factory, request):
+    prebuilt = os.getenv("PRE_BUILT_SETUPTOOLS_SDIST")
+    if prebuilt and os.path.exists(prebuilt):  # pragma: no cover
+        return Path(prebuilt).resolve()
+
+    sdist, _ = _build_distributions(tmp_path_factory, request)
+    return sdist
+
+
+@pytest.fixture(scope="session")
+def setuptools_wheel(tmp_path_factory, request):
+    prebuilt = os.getenv("PRE_BUILT_SETUPTOOLS_WHEEL")
+    if prebuilt and os.path.exists(prebuilt):  # pragma: no cover
+        return Path(prebuilt).resolve()
+
+    _, wheel = _build_distributions(tmp_path_factory, request)
+    return wheel
+
+
+@pytest.fixture
+def venv(tmp_path, setuptools_wheel):
+    """Virtual env with the version of setuptools under test installed"""
+    env = environment.VirtualEnv()
+    env.root = path.Path(tmp_path / 'venv')
+    env.create_opts = ['--no-setuptools', '--wheel=bundle']
+    # TODO: Use `--no-wheel` when setuptools implements its own bdist_wheel
+    env.req = str(setuptools_wheel)
+    # In some environments (eg. downstream distro packaging),
+    # where tox isn't used to run tests and PYTHONPATH is set to point to
+    # a specific setuptools codebase, PYTHONPATH will leak into the spawned
+    # processes.
+    # env.create() should install the just created setuptools
+    # wheel, but it doesn't if it finds another existing matching setuptools
+    # installation present on PYTHONPATH:
+    # `setuptools is already installed with the same version as the provided
+    # wheel. Use --force-reinstall to force an installation of the wheel.`
+    # This prevents leaking PYTHONPATH to the created environment.
+    with contexts.environment(PYTHONPATH=None):
+        return env.create()
+
+
+@pytest.fixture
+def venv_without_setuptools(tmp_path):
+    """Virtual env without any version of setuptools installed"""
+    env = environment.VirtualEnv()
+    env.root = path.Path(tmp_path / 'venv_without_setuptools')
+    env.create_opts = ['--no-setuptools', '--no-wheel']
+    env.ensure_env()
+    return env
+
+
+@pytest.fixture
+def bare_venv(tmp_path):
+    """Virtual env without any common packages installed"""
+    env = environment.VirtualEnv()
+    env.root = path.Path(tmp_path / 'bare_venv')
+    env.create_opts = ['--no-setuptools', '--no-pip', '--no-wheel', '--no-seed']
+    env.ensure_env()
+    return env
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/external.html b/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/external.html
new file mode 100644
index 00000000..92e4702f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/external.html
@@ -0,0 +1,3 @@
+<html><body>
+<a href="/foobar-0.1.tar.gz#md5=1__bad_md5___">bad old link</a>
+</body></html>
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/simple/foobar/index.html b/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/simple/foobar/index.html
new file mode 100644
index 00000000..fefb028b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/indexes/test_links_priority/simple/foobar/index.html
@@ -0,0 +1,4 @@
+<html><body>
+<a href="/foobar-0.1.tar.gz#md5=0_correct_md5">foobar-0.1.tar.gz</a><br/>
+<a href="../../external.html" rel="homepage">external homepage</a><br/>
+</body></html>
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__init__.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b08c4a11
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/helpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/helpers.cpython-312.pyc
new file mode 100644
index 00000000..b0ef9e08
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/helpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/test_pip_install_sdist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/test_pip_install_sdist.cpython-312.pyc
new file mode 100644
index 00000000..6bca54a7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/__pycache__/test_pip_install_sdist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/helpers.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/helpers.py
new file mode 100644
index 00000000..77b196e0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/helpers.py
@@ -0,0 +1,77 @@
+"""Reusable functions and classes for different types of integration tests.
+
+For example ``Archive`` can be used to check the contents of distribution built
+with setuptools, and ``run`` will always try to be as verbose as possible to
+facilitate debugging.
+"""
+
+import os
+import subprocess
+import tarfile
+from pathlib import Path
+from zipfile import ZipFile
+
+
+def run(cmd, env=None):
+    r = subprocess.run(
+        cmd,
+        capture_output=True,
+        text=True,
+        encoding="utf-8",
+        env={**os.environ, **(env or {})},
+        # ^-- allow overwriting instead of discarding the current env
+    )
+
+    out = r.stdout + "\n" + r.stderr
+    # pytest omits stdout/err by default, if the test fails they help debugging
+    print("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
+    print(f"Command: {cmd}\nreturn code: {r.returncode}\n\n{out}")
+
+    if r.returncode == 0:
+        return out
+    raise subprocess.CalledProcessError(r.returncode, cmd, r.stdout, r.stderr)
+
+
+class Archive:
+    """Compatibility layer for ZipFile/Info and TarFile/Info"""
+
+    def __init__(self, filename):
+        self._filename = filename
+        if filename.endswith("tar.gz"):
+            self._obj = tarfile.open(filename, "r:gz")
+        elif filename.endswith("zip"):
+            self._obj = ZipFile(filename)
+        else:
+            raise ValueError(f"{filename} doesn't seem to be a zip or tar.gz")
+
+    def __iter__(self):
+        if hasattr(self._obj, "infolist"):
+            return iter(self._obj.infolist())
+        return iter(self._obj)
+
+    def get_name(self, zip_or_tar_info):
+        if hasattr(zip_or_tar_info, "filename"):
+            return zip_or_tar_info.filename
+        return zip_or_tar_info.name
+
+    def get_content(self, zip_or_tar_info):
+        if hasattr(self._obj, "extractfile"):
+            content = self._obj.extractfile(zip_or_tar_info)
+            if content is None:
+                msg = f"Invalid {zip_or_tar_info.name} in {self._filename}"
+                raise ValueError(msg)
+            return str(content.read(), "utf-8")
+        return str(self._obj.read(zip_or_tar_info), "utf-8")
+
+
+def get_sdist_members(sdist_path):
+    with tarfile.open(sdist_path, "r:gz") as tar:
+        files = [Path(f) for f in tar.getnames()]
+    # remove root folder
+    relative_files = ("/".join(f.parts[1:]) for f in files)
+    return {f for f in relative_files if f}
+
+
+def get_wheel_members(wheel_path):
+    with ZipFile(wheel_path) as zipfile:
+        return set(zipfile.namelist())
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/test_pip_install_sdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/test_pip_install_sdist.py
new file mode 100644
index 00000000..b2f1c080
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/integration/test_pip_install_sdist.py
@@ -0,0 +1,223 @@
+# https://github.com/python/mypy/issues/16936
+# mypy: disable-error-code="has-type"
+"""Integration tests for setuptools that focus on building packages via pip.
+
+The idea behind these tests is not to exhaustively check all the possible
+combinations of packages, operating systems, supporting libraries, etc, but
+rather check a limited number of popular packages and how they interact with
+the exposed public API. This way if any change in API is introduced, we hope to
+identify backward compatibility problems before publishing a release.
+
+The number of tested packages is purposefully kept small, to minimise duration
+and the associated maintenance cost (changes in the way these packages define
+their build process may require changes in the tests).
+"""
+
+import json
+import os
+import shutil
+import sys
+from enum import Enum
+from glob import glob
+from hashlib import md5
+from urllib.request import urlopen
+
+import pytest
+from packaging.requirements import Requirement
+
+from .helpers import Archive, run
+
+pytestmark = pytest.mark.integration
+
+
+(LATEST,) = Enum("v", "LATEST")  # type: ignore[misc] # https://github.com/python/mypy/issues/16936
+"""Default version to be checked"""
+# There are positive and negative aspects of checking the latest version of the
+# packages.
+# The main positive aspect is that the latest version might have already
+# removed the use of APIs deprecated in previous releases of setuptools.
+
+
+# Packages to be tested:
+# (Please notice the test environment cannot support EVERY library required for
+# compiling binary extensions. In Ubuntu/Debian nomenclature, we only assume
+# that `build-essential`, `gfortran` and `libopenblas-dev` are installed,
+# due to their relevance to the numerical/scientific programming ecosystem)
+EXAMPLES = [
+    ("pip", LATEST),  # just in case...
+    ("pytest", LATEST),  # uses setuptools_scm
+    ("mypy", LATEST),  # custom build_py + ext_modules
+    # --- Popular packages: https://hugovk.github.io/top-pypi-packages/ ---
+    ("botocore", LATEST),
+    ("kiwisolver", LATEST),  # build_ext
+    ("brotli", LATEST),  # not in the list but used by urllib3
+    ("pyyaml", LATEST),  # cython + custom build_ext + custom distclass
+    ("charset-normalizer", LATEST),  # uses mypyc, used by aiohttp
+    ("protobuf", LATEST),
+    ("requests", LATEST),
+    ("celery", LATEST),
+    # When adding packages to this list, make sure they expose a `__version__`
+    # attribute, or modify the tests below
+]
+
+
+# Some packages have "optional" dependencies that modify their build behaviour
+# and are not listed in pyproject.toml, others still use `setup_requires`
+EXTRA_BUILD_DEPS = {
+    "pyyaml": ("Cython<3.0",),  # constraint to avoid errors
+    "charset-normalizer": ("mypy>=1.4.1",),  # no pyproject.toml available
+}
+
+EXTRA_ENV_VARS = {
+    "pyyaml": {"PYYAML_FORCE_CYTHON": "1"},
+    "charset-normalizer": {"CHARSET_NORMALIZER_USE_MYPYC": "1"},
+}
+
+IMPORT_NAME = {
+    "pyyaml": "yaml",
+    "protobuf": "google.protobuf",
+}
+
+
+VIRTUALENV = (sys.executable, "-m", "virtualenv")
+
+
+# By default, pip will try to build packages in isolation (PEP 517), which
+# means it will download the previous stable version of setuptools.
+# `pip` flags can avoid that (the version of setuptools under test
+# should be the one to be used)
+INSTALL_OPTIONS = (
+    "--ignore-installed",
+    "--no-build-isolation",
+    # Omit "--no-binary :all:" the sdist is supplied directly.
+    # Allows dependencies as wheels.
+)
+# The downside of `--no-build-isolation` is that pip will not download build
+# dependencies. The test script will have to also handle that.
+
+
+@pytest.fixture
+def venv_python(tmp_path):
+    run([*VIRTUALENV, str(tmp_path / ".venv")])
+    possible_path = (str(p.parent) for p in tmp_path.glob(".venv/*/python*"))
+    return shutil.which("python", path=os.pathsep.join(possible_path))
+
+
+@pytest.fixture(autouse=True)
+def _prepare(tmp_path, venv_python, monkeypatch):
+    download_path = os.getenv("DOWNLOAD_PATH", str(tmp_path))
+    os.makedirs(download_path, exist_ok=True)
+
+    # Environment vars used for building some of the packages
+    monkeypatch.setenv("USE_MYPYC", "1")
+
+    yield
+
+    # Let's provide the maximum amount of information possible in the case
+    # it is necessary to debug the tests directly from the CI logs.
+    print("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
+    print("Temporary directory:")
+    map(print, tmp_path.glob("*"))
+    print("Virtual environment:")
+    run([venv_python, "-m", "pip", "freeze"])
+
+
+@pytest.mark.parametrize(("package", "version"), EXAMPLES)
+@pytest.mark.uses_network
+def test_install_sdist(package, version, tmp_path, venv_python, setuptools_wheel):
+    venv_pip = (venv_python, "-m", "pip")
+    sdist = retrieve_sdist(package, version, tmp_path)
+    deps = build_deps(package, sdist)
+    if deps:
+        print("~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~")
+        print("Dependencies:", deps)
+        run([*venv_pip, "install", *deps])
+
+    # Use a virtualenv to simulate PEP 517 isolation
+    # but install fresh setuptools wheel to ensure the version under development
+    env = EXTRA_ENV_VARS.get(package, {})
+    run([*venv_pip, "install", "--force-reinstall", setuptools_wheel])
+    run([*venv_pip, "install", *INSTALL_OPTIONS, sdist], env)
+
+    # Execute a simple script to make sure the package was installed correctly
+    pkg = IMPORT_NAME.get(package, package).replace("-", "_")
+    script = f"import {pkg}; print(getattr({pkg}, '__version__', 0))"
+    run([venv_python, "-c", script])
+
+
+# ---- Helper Functions ----
+
+
+def retrieve_sdist(package, version, tmp_path):
+    """Either use cached sdist file or download it from PyPI"""
+    # `pip download` cannot be used due to
+    # https://github.com/pypa/pip/issues/1884
+    # https://discuss.python.org/t/pep-625-file-name-of-a-source-distribution/4686
+    # We have to find the correct distribution file and download it
+    download_path = os.getenv("DOWNLOAD_PATH", str(tmp_path))
+    dist = retrieve_pypi_sdist_metadata(package, version)
+
+    # Remove old files to prevent cache to grow indefinitely
+    for file in glob(os.path.join(download_path, f"{package}*")):
+        if dist["filename"] != file:
+            os.unlink(file)
+
+    dist_file = os.path.join(download_path, dist["filename"])
+    if not os.path.exists(dist_file):
+        download(dist["url"], dist_file, dist["md5_digest"])
+    return dist_file
+
+
+def retrieve_pypi_sdist_metadata(package, version):
+    # https://warehouse.pypa.io/api-reference/json.html
+    id_ = package if version is LATEST else f"{package}/{version}"
+    with urlopen(f"https://pypi.org/pypi/{id_}/json") as f:
+        metadata = json.load(f)
+
+    if metadata["info"]["yanked"]:
+        raise ValueError(f"Release for {package} {version} was yanked")
+
+    version = metadata["info"]["version"]
+    release = metadata["releases"][version] if version is LATEST else metadata["urls"]
+    (sdist,) = filter(lambda d: d["packagetype"] == "sdist", release)
+    return sdist
+
+
+def download(url, dest, md5_digest):
+    with urlopen(url) as f:
+        data = f.read()
+
+    assert md5(data).hexdigest() == md5_digest
+
+    with open(dest, "wb") as f:
+        f.write(data)
+
+    assert os.path.exists(dest)
+
+
+def build_deps(package, sdist_file):
+    """Find out what are the build dependencies for a package.
+
+    "Manually" install them, since pip will not install build
+    deps with `--no-build-isolation`.
+    """
+    # delay importing, since pytest discovery phase may hit this file from a
+    # testenv without tomli
+    from setuptools.compat.py310 import tomllib
+
+    archive = Archive(sdist_file)
+    info = tomllib.loads(_read_pyproject(archive))
+    deps = info.get("build-system", {}).get("requires", [])
+    deps += EXTRA_BUILD_DEPS.get(package, [])
+    # Remove setuptools from requirements (and deduplicate)
+    requirements = {Requirement(d).name: d for d in deps}
+    return [v for k, v in requirements.items() if k != "setuptools"]
+
+
+def _read_pyproject(archive):
+    contents = (
+        archive.get_content(member)
+        for member in archive
+        if os.path.basename(archive.get_name(member)) == "pyproject.toml"
+    )
+    return next(contents, "")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/mod_with_constant.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/mod_with_constant.py
new file mode 100644
index 00000000..ef755dd1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/mod_with_constant.py
@@ -0,0 +1 @@
+value = 'three, sir!'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/namespaces.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/namespaces.py
new file mode 100644
index 00000000..248db98f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/namespaces.py
@@ -0,0 +1,90 @@
+import ast
+import json
+import textwrap
+from pathlib import Path
+
+
+def iter_namespace_pkgs(namespace):
+    parts = namespace.split(".")
+    for i in range(len(parts)):
+        yield ".".join(parts[: i + 1])
+
+
+def build_namespace_package(tmpdir, name, version="1.0", impl="pkg_resources"):
+    src_dir = tmpdir / name
+    src_dir.mkdir()
+    setup_py = src_dir / 'setup.py'
+    namespace, _, rest = name.rpartition('.')
+    namespaces = list(iter_namespace_pkgs(namespace))
+    setup_args = {
+        "name": name,
+        "version": version,
+        "packages": namespaces,
+    }
+
+    if impl == "pkg_resources":
+        tmpl = '__import__("pkg_resources").declare_namespace(__name__)'
+        setup_args["namespace_packages"] = namespaces
+    elif impl == "pkgutil":
+        tmpl = '__path__ = __import__("pkgutil").extend_path(__path__, __name__)'
+    else:
+        raise ValueError(f"Cannot recognise {impl=} when creating namespaces")
+
+    args = json.dumps(setup_args, indent=4)
+    assert ast.literal_eval(args)  # ensure it is valid Python
+
+    script = textwrap.dedent(
+        """\
+        import setuptools
+        args = {args}
+        setuptools.setup(**args)
+        """
+    ).format(args=args)
+    setup_py.write_text(script, encoding='utf-8')
+
+    ns_pkg_dir = Path(src_dir, namespace.replace(".", "/"))
+    ns_pkg_dir.mkdir(parents=True)
+
+    for ns in namespaces:
+        pkg_init = src_dir / ns.replace(".", "/") / '__init__.py'
+        pkg_init.write_text(tmpl, encoding='utf-8')
+
+    pkg_mod = ns_pkg_dir / (rest + '.py')
+    some_functionality = 'name = {rest!r}'.format(**locals())
+    pkg_mod.write_text(some_functionality, encoding='utf-8')
+    return src_dir
+
+
+def build_pep420_namespace_package(tmpdir, name):
+    src_dir = tmpdir / name
+    src_dir.mkdir()
+    pyproject = src_dir / "pyproject.toml"
+    namespace, _, rest = name.rpartition(".")
+    script = f"""\
+        [build-system]
+        requires = ["setuptools"]
+        build-backend = "setuptools.build_meta"
+
+        [project]
+        name = "{name}"
+        version = "3.14159"
+        """
+    pyproject.write_text(textwrap.dedent(script), encoding='utf-8')
+    ns_pkg_dir = Path(src_dir, namespace.replace(".", "/"))
+    ns_pkg_dir.mkdir(parents=True)
+    pkg_mod = ns_pkg_dir / (rest + ".py")
+    some_functionality = f"name = {rest!r}"
+    pkg_mod.write_text(some_functionality, encoding='utf-8')
+    return src_dir
+
+
+def make_site_dir(target):
+    """
+    Add a sitecustomize.py module in target to cause
+    target to be added to site dirs such that .pth files
+    are processed there.
+    """
+    sc = target / 'sitecustomize.py'
+    target_str = str(target)
+    tmpl = '__import__("site").addsitedir({target_str!r})'
+    sc.write_text(tmpl.format(**locals()), encoding='utf-8')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/script-with-bom.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/script-with-bom.py
new file mode 100644
index 00000000..c074d263
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/script-with-bom.py
@@ -0,0 +1 @@
+result = 'passed'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/server.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/server.py
new file mode 100644
index 00000000..623a49a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/server.py
@@ -0,0 +1,86 @@
+"""Basic http server for tests to simulate PyPI or custom indexes"""
+
+import http.server
+import os
+import threading
+import time
+import urllib.parse
+import urllib.request
+
+
+class IndexServer(http.server.HTTPServer):
+    """Basic single-threaded http server simulating a package index
+
+    You can use this server in unittest like this::
+        s = IndexServer()
+        s.start()
+        index_url = s.base_url() + 'mytestindex'
+        # do some test requests to the index
+        # The index files should be located in setuptools/tests/indexes
+        s.stop()
+    """
+
+    def __init__(
+        self,
+        server_address=('', 0),
+        RequestHandlerClass=http.server.SimpleHTTPRequestHandler,
+    ):
+        http.server.HTTPServer.__init__(self, server_address, RequestHandlerClass)
+        self._run = True
+
+    def start(self):
+        self.thread = threading.Thread(target=self.serve_forever)
+        self.thread.start()
+
+    def stop(self):
+        "Stop the server"
+
+        # Let the server finish the last request and wait for a new one.
+        time.sleep(0.1)
+
+        self.shutdown()
+        self.thread.join()
+        self.socket.close()
+
+    def base_url(self):
+        port = self.server_port
+        return f'http://127.0.0.1:{port}/setuptools/tests/indexes/'
+
+
+class RequestRecorder(http.server.BaseHTTPRequestHandler):
+    def do_GET(self):
+        requests = vars(self.server).setdefault('requests', [])
+        requests.append(self)
+        self.send_response(200, 'OK')
+
+
+class MockServer(http.server.HTTPServer, threading.Thread):
+    """
+    A simple HTTP Server that records the requests made to it.
+    """
+
+    def __init__(self, server_address=('', 0), RequestHandlerClass=RequestRecorder):
+        http.server.HTTPServer.__init__(self, server_address, RequestHandlerClass)
+        threading.Thread.__init__(self)
+        self.daemon = True
+        self.requests = []
+
+    def run(self):
+        self.serve_forever()
+
+    @property
+    def netloc(self):
+        return f'localhost:{self.server_port}'
+
+    @property
+    def url(self):
+        return f'http://{self.netloc}/'
+
+
+def path_to_url(path, authority=None):
+    """Convert a path to a file: URL."""
+    path = os.path.normpath(os.path.abspath(path))
+    base = 'file:'
+    if authority is not None:
+        base += '//' + authority
+    return urllib.parse.urljoin(base, urllib.request.pathname2url(path))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_archive_util.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_archive_util.py
new file mode 100644
index 00000000..e3efc628
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_archive_util.py
@@ -0,0 +1,36 @@
+import io
+import tarfile
+
+import pytest
+
+from setuptools import archive_util
+
+
+@pytest.fixture
+def tarfile_with_unicode(tmpdir):
+    """
+    Create a tarfile containing only a file whose name is
+    a zero byte file called testimäge.png.
+    """
+    tarobj = io.BytesIO()
+
+    with tarfile.open(fileobj=tarobj, mode="w:gz") as tgz:
+        data = b""
+
+        filename = "testimäge.png"
+
+        t = tarfile.TarInfo(filename)
+        t.size = len(data)
+
+        tgz.addfile(t, io.BytesIO(data))
+
+    target = tmpdir / 'unicode-pkg-1.0.tar.gz'
+    with open(str(target), mode='wb') as tf:
+        tf.write(tarobj.getvalue())
+    return str(target)
+
+
+@pytest.mark.xfail(reason="#710 and #712")
+def test_unicode_files(tarfile_with_unicode, tmpdir):
+    target = tmpdir / 'out'
+    archive_util.unpack_archive(tarfile_with_unicode, str(target))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_deprecations.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_deprecations.py
new file mode 100644
index 00000000..d9d67b06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_deprecations.py
@@ -0,0 +1,28 @@
+"""develop tests"""
+
+import sys
+from unittest import mock
+
+import pytest
+
+from setuptools import SetuptoolsDeprecationWarning
+from setuptools.dist import Distribution
+
+
+@pytest.mark.skipif(sys.platform == 'win32', reason='non-Windows only')
+@pytest.mark.xfail(reason="bdist_rpm is long deprecated, should we remove it? #1988")
+@mock.patch('distutils.command.bdist_rpm.bdist_rpm')
+def test_bdist_rpm_warning(distutils_cmd, tmpdir_cwd):
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['bdist_rpm'],
+            name='foo',
+            py_modules=['hi'],
+        )
+    )
+    dist.parse_command_line()
+    with pytest.warns(SetuptoolsDeprecationWarning):
+        dist.run_commands()
+
+    distutils_cmd.run.assert_called_once()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_egg.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_egg.py
new file mode 100644
index 00000000..036167dd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_egg.py
@@ -0,0 +1,73 @@
+"""develop tests"""
+
+import os
+import re
+import zipfile
+
+import pytest
+
+from setuptools.dist import Distribution
+
+from . import contexts
+
+SETUP_PY = """\
+from setuptools import setup
+
+setup(py_modules=['hi'])
+"""
+
+
+@pytest.fixture
+def setup_context(tmpdir):
+    with (tmpdir / 'setup.py').open('w') as f:
+        f.write(SETUP_PY)
+    with (tmpdir / 'hi.py').open('w') as f:
+        f.write('1\n')
+    with tmpdir.as_cwd():
+        yield tmpdir
+
+
+class Test:
+    @pytest.mark.usefixtures("user_override")
+    @pytest.mark.usefixtures("setup_context")
+    def test_bdist_egg(self):
+        dist = Distribution(
+            dict(
+                script_name='setup.py',
+                script_args=['bdist_egg'],
+                name='foo',
+                py_modules=['hi'],
+            )
+        )
+        os.makedirs(os.path.join('build', 'src'))
+        with contexts.quiet():
+            dist.parse_command_line()
+            dist.run_commands()
+
+        # let's see if we got our egg link at the right place
+        [content] = os.listdir('dist')
+        assert re.match(r'foo-0.0.0-py[23].\d+.egg$', content)
+
+    @pytest.mark.xfail(
+        os.environ.get('PYTHONDONTWRITEBYTECODE', False),
+        reason="Byte code disabled",
+    )
+    @pytest.mark.usefixtures("user_override")
+    @pytest.mark.usefixtures("setup_context")
+    def test_exclude_source_files(self):
+        dist = Distribution(
+            dict(
+                script_name='setup.py',
+                script_args=['bdist_egg', '--exclude-source-files'],
+                py_modules=['hi'],
+            )
+        )
+        with contexts.quiet():
+            dist.parse_command_line()
+            dist.run_commands()
+        [dist_name] = os.listdir('dist')
+        dist_filename = os.path.join('dist', dist_name)
+        zip = zipfile.ZipFile(dist_filename)
+        names = list(zi.filename for zi in zip.filelist)
+        assert 'hi.pyc' in names
+        assert 'hi.py' not in names
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_wheel.py
new file mode 100644
index 00000000..d51dfbeb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_bdist_wheel.py
@@ -0,0 +1,623 @@
+from __future__ import annotations
+
+import builtins
+import importlib
+import os.path
+import platform
+import shutil
+import stat
+import struct
+import sys
+import sysconfig
+from contextlib import suppress
+from inspect import cleandoc
+from zipfile import ZipFile
+
+import jaraco.path
+import pytest
+from packaging import tags
+
+import setuptools
+from setuptools.command.bdist_wheel import bdist_wheel, get_abi_tag
+from setuptools.dist import Distribution
+from setuptools.warnings import SetuptoolsDeprecationWarning
+
+from distutils.core import run_setup
+
+DEFAULT_FILES = {
+    "dummy_dist-1.0.dist-info/top_level.txt",
+    "dummy_dist-1.0.dist-info/METADATA",
+    "dummy_dist-1.0.dist-info/WHEEL",
+    "dummy_dist-1.0.dist-info/RECORD",
+}
+DEFAULT_LICENSE_FILES = {
+    "LICENSE",
+    "LICENSE.txt",
+    "LICENCE",
+    "LICENCE.txt",
+    "COPYING",
+    "COPYING.md",
+    "NOTICE",
+    "NOTICE.rst",
+    "AUTHORS",
+    "AUTHORS.txt",
+}
+OTHER_IGNORED_FILES = {
+    "LICENSE~",
+    "AUTHORS~",
+}
+SETUPPY_EXAMPLE = """\
+from setuptools import setup
+
+setup(
+    name='dummy_dist',
+    version='1.0',
+)
+"""
+
+
+EXAMPLES = {
+    "dummy-dist": {
+        "setup.py": SETUPPY_EXAMPLE,
+        "licenses": {"DUMMYFILE": ""},
+        **dict.fromkeys(DEFAULT_LICENSE_FILES | OTHER_IGNORED_FILES, ""),
+    },
+    "simple-dist": {
+        "setup.py": cleandoc(
+            """
+            from setuptools import setup
+
+            setup(
+                name="simple.dist",
+                version="0.1",
+                description="A testing distribution \N{SNOWMAN}",
+                extras_require={"voting": ["beaglevote"]},
+            )
+            """
+        ),
+        "simpledist": "",
+    },
+    "complex-dist": {
+        "setup.py": cleandoc(
+            """
+            from setuptools import setup
+
+            setup(
+                name="complex-dist",
+                version="0.1",
+                description="Another testing distribution \N{SNOWMAN}",
+                long_description="Another testing distribution \N{SNOWMAN}",
+                author="Illustrious Author",
+                author_email="illustrious@example.org",
+                url="http://example.org/exemplary",
+                packages=["complexdist"],
+                setup_requires=["setuptools"],
+                install_requires=["quux", "splort"],
+                extras_require={"simple": ["simple.dist"]},
+                entry_points={
+                    "console_scripts": [
+                        "complex-dist=complexdist:main",
+                        "complex-dist2=complexdist:main",
+                    ],
+                },
+            )
+            """
+        ),
+        "complexdist": {"__init__.py": "def main(): return"},
+    },
+    "headers-dist": {
+        "setup.py": cleandoc(
+            """
+            from setuptools import setup
+
+            setup(
+                name="headers.dist",
+                version="0.1",
+                description="A distribution with headers",
+                headers=["header.h"],
+            )
+            """
+        ),
+        "headersdist.py": "",
+        "header.h": "",
+    },
+    "commasinfilenames-dist": {
+        "setup.py": cleandoc(
+            """
+            from setuptools import setup
+
+            setup(
+                name="testrepo",
+                version="0.1",
+                packages=["mypackage"],
+                description="A test package with commas in file names",
+                include_package_data=True,
+                package_data={"mypackage.data": ["*"]},
+            )
+            """
+        ),
+        "mypackage": {
+            "__init__.py": "",
+            "data": {"__init__.py": "", "1,2,3.txt": ""},
+        },
+        "testrepo-0.1.0": {
+            "mypackage": {"__init__.py": ""},
+        },
+    },
+    "unicode-dist": {
+        "setup.py": cleandoc(
+            """
+            from setuptools import setup
+
+            setup(
+                name="unicode.dist",
+                version="0.1",
+                description="A testing distribution \N{SNOWMAN}",
+                packages=["unicodedist"],
+                zip_safe=True,
+            )
+            """
+        ),
+        "unicodedist": {"__init__.py": "", "åäö_日本語.py": ""},
+    },
+    "utf8-metadata-dist": {
+        "setup.cfg": cleandoc(
+            """
+            [metadata]
+            name = utf8-metadata-dist
+            version = 42
+            author_email = "John X. Ãørçeč" <john@utf8.org>, Γαμα קּ 東 <gama@utf8.org>
+            long_description = file: README.rst
+            """
+        ),
+        "README.rst": "UTF-8 描述 説明",
+    },
+}
+
+
+if sys.platform != "win32":
+    # ABI3 extensions don't really work on Windows
+    EXAMPLES["abi3extension-dist"] = {
+        "setup.py": cleandoc(
+            """
+            from setuptools import Extension, setup
+
+            setup(
+                name="extension.dist",
+                version="0.1",
+                description="A testing distribution \N{SNOWMAN}",
+                ext_modules=[
+                    Extension(
+                        name="extension", sources=["extension.c"], py_limited_api=True
+                    )
+                ],
+            )
+            """
+        ),
+        "setup.cfg": "[bdist_wheel]\npy_limited_api=cp32",
+        "extension.c": "#define Py_LIMITED_API 0x03020000\n#include <Python.h>",
+    }
+
+
+def bdist_wheel_cmd(**kwargs):
+    """Run command in the same process so that it is easier to collect coverage"""
+    dist_obj = (
+        run_setup("setup.py", stop_after="init")
+        if os.path.exists("setup.py")
+        else Distribution({"script_name": "%%build_meta%%"})
+    )
+    dist_obj.parse_config_files()
+    cmd = bdist_wheel(dist_obj)
+    for attr, value in kwargs.items():
+        setattr(cmd, attr, value)
+    cmd.finalize_options()
+    return cmd
+
+
+def mkexample(tmp_path_factory, name):
+    basedir = tmp_path_factory.mktemp(name)
+    jaraco.path.build(EXAMPLES[name], prefix=str(basedir))
+    return basedir
+
+
+@pytest.fixture(scope="session")
+def wheel_paths(tmp_path_factory):
+    build_base = tmp_path_factory.mktemp("build")
+    dist_dir = tmp_path_factory.mktemp("dist")
+    for name in EXAMPLES:
+        example_dir = mkexample(tmp_path_factory, name)
+        build_dir = build_base / name
+        with jaraco.path.DirectoryStack().context(example_dir):
+            bdist_wheel_cmd(bdist_dir=str(build_dir), dist_dir=str(dist_dir)).run()
+
+    return sorted(str(fname) for fname in dist_dir.glob("*.whl"))
+
+
+@pytest.fixture
+def dummy_dist(tmp_path_factory):
+    return mkexample(tmp_path_factory, "dummy-dist")
+
+
+def test_no_scripts(wheel_paths):
+    """Make sure entry point scripts are not generated."""
+    path = next(path for path in wheel_paths if "complex_dist" in path)
+    for entry in ZipFile(path).infolist():
+        assert ".data/scripts/" not in entry.filename
+
+
+def test_unicode_record(wheel_paths):
+    path = next(path for path in wheel_paths if "unicode.dist" in path)
+    with ZipFile(path) as zf:
+        record = zf.read("unicode.dist-0.1.dist-info/RECORD")
+
+    assert "åäö_日本語.py".encode() in record
+
+
+UTF8_PKG_INFO = """\
+Metadata-Version: 2.1
+Name: helloworld
+Version: 42
+Author-email: "John X. Ãørçeč" <john@utf8.org>, Γαμα קּ 東 <gama@utf8.org>
+
+
+UTF-8 描述 説明
+"""
+
+
+def test_preserve_unicode_metadata(monkeypatch, tmp_path):
+    monkeypatch.chdir(tmp_path)
+    egginfo = tmp_path / "dummy_dist.egg-info"
+    distinfo = tmp_path / "dummy_dist.dist-info"
+
+    egginfo.mkdir()
+    (egginfo / "PKG-INFO").write_text(UTF8_PKG_INFO, encoding="utf-8")
+    (egginfo / "dependency_links.txt").touch()
+
+    class simpler_bdist_wheel(bdist_wheel):
+        """Avoid messing with setuptools/distutils internals"""
+
+        def __init__(self):
+            pass
+
+        @property
+        def license_paths(self):
+            return []
+
+    cmd_obj = simpler_bdist_wheel()
+    cmd_obj.egg2dist(egginfo, distinfo)
+
+    metadata = (distinfo / "METADATA").read_text(encoding="utf-8")
+    assert 'Author-email: "John X. Ãørçeč"' in metadata
+    assert "Γαμα קּ 東 " in metadata
+    assert "UTF-8 描述 説明" in metadata
+
+
+def test_licenses_default(dummy_dist, monkeypatch, tmp_path):
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path)).run()
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        license_files = {
+            "dummy_dist-1.0.dist-info/" + fname for fname in DEFAULT_LICENSE_FILES
+        }
+        assert set(wf.namelist()) == DEFAULT_FILES | license_files
+
+
+def test_licenses_deprecated(dummy_dist, monkeypatch, tmp_path):
+    dummy_dist.joinpath("setup.cfg").write_text(
+        "[metadata]\nlicense_file=licenses/DUMMYFILE", encoding="utf-8"
+    )
+    monkeypatch.chdir(dummy_dist)
+
+    bdist_wheel_cmd(bdist_dir=str(tmp_path)).run()
+
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        license_files = {"dummy_dist-1.0.dist-info/DUMMYFILE"}
+        assert set(wf.namelist()) == DEFAULT_FILES | license_files
+
+
+@pytest.mark.parametrize(
+    ("config_file", "config"),
+    [
+        ("setup.cfg", "[metadata]\nlicense_files=licenses/*\n  LICENSE"),
+        ("setup.cfg", "[metadata]\nlicense_files=licenses/*, LICENSE"),
+        (
+            "setup.py",
+            SETUPPY_EXAMPLE.replace(
+                ")", "  license_files=['licenses/DUMMYFILE', 'LICENSE'])"
+            ),
+        ),
+    ],
+)
+def test_licenses_override(dummy_dist, monkeypatch, tmp_path, config_file, config):
+    dummy_dist.joinpath(config_file).write_text(config, encoding="utf-8")
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path)).run()
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        license_files = {
+            "dummy_dist-1.0.dist-info/" + fname for fname in {"DUMMYFILE", "LICENSE"}
+        }
+        assert set(wf.namelist()) == DEFAULT_FILES | license_files
+
+
+def test_licenses_disabled(dummy_dist, monkeypatch, tmp_path):
+    dummy_dist.joinpath("setup.cfg").write_text(
+        "[metadata]\nlicense_files=\n", encoding="utf-8"
+    )
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path)).run()
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        assert set(wf.namelist()) == DEFAULT_FILES
+
+
+def test_build_number(dummy_dist, monkeypatch, tmp_path):
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path), build_number="2").run()
+    with ZipFile("dist/dummy_dist-1.0-2-py3-none-any.whl") as wf:
+        filenames = set(wf.namelist())
+        assert "dummy_dist-1.0.dist-info/RECORD" in filenames
+        assert "dummy_dist-1.0.dist-info/METADATA" in filenames
+
+
+def test_universal_deprecated(dummy_dist, monkeypatch, tmp_path):
+    monkeypatch.chdir(dummy_dist)
+    with pytest.warns(SetuptoolsDeprecationWarning, match=".*universal is deprecated"):
+        bdist_wheel_cmd(bdist_dir=str(tmp_path), universal=True).run()
+
+    # For now we still respect the option
+    assert os.path.exists("dist/dummy_dist-1.0-py2.py3-none-any.whl")
+
+
+EXTENSION_EXAMPLE = """\
+#include <Python.h>
+
+static PyMethodDef methods[] = {
+  { NULL, NULL, 0, NULL }
+};
+
+static struct PyModuleDef module_def = {
+  PyModuleDef_HEAD_INIT,
+  "extension",
+  "Dummy extension module",
+  -1,
+  methods
+};
+
+PyMODINIT_FUNC PyInit_extension(void) {
+  return PyModule_Create(&module_def);
+}
+"""
+EXTENSION_SETUPPY = """\
+from __future__ import annotations
+
+from setuptools import Extension, setup
+
+setup(
+    name="extension.dist",
+    version="0.1",
+    description="A testing distribution \N{SNOWMAN}",
+    ext_modules=[Extension(name="extension", sources=["extension.c"])],
+)
+"""
+
+
+@pytest.mark.filterwarnings(
+    "once:Config variable '.*' is unset.*, Python ABI tag may be incorrect"
+)
+def test_limited_abi(monkeypatch, tmp_path, tmp_path_factory):
+    """Test that building a binary wheel with the limited ABI works."""
+    source_dir = tmp_path_factory.mktemp("extension_dist")
+    (source_dir / "setup.py").write_text(EXTENSION_SETUPPY, encoding="utf-8")
+    (source_dir / "extension.c").write_text(EXTENSION_EXAMPLE, encoding="utf-8")
+    build_dir = tmp_path.joinpath("build")
+    dist_dir = tmp_path.joinpath("dist")
+    monkeypatch.chdir(source_dir)
+    bdist_wheel_cmd(bdist_dir=str(build_dir), dist_dir=str(dist_dir)).run()
+
+
+def test_build_from_readonly_tree(dummy_dist, monkeypatch, tmp_path):
+    basedir = str(tmp_path.joinpath("dummy"))
+    shutil.copytree(str(dummy_dist), basedir)
+    monkeypatch.chdir(basedir)
+
+    # Make the tree read-only
+    for root, _dirs, files in os.walk(basedir):
+        for fname in files:
+            os.chmod(os.path.join(root, fname), stat.S_IREAD)
+
+    bdist_wheel_cmd().run()
+
+
+@pytest.mark.parametrize(
+    ("option", "compress_type"),
+    list(bdist_wheel.supported_compressions.items()),
+    ids=list(bdist_wheel.supported_compressions),
+)
+def test_compression(dummy_dist, monkeypatch, tmp_path, option, compress_type):
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path), compression=option).run()
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        filenames = set(wf.namelist())
+        assert "dummy_dist-1.0.dist-info/RECORD" in filenames
+        assert "dummy_dist-1.0.dist-info/METADATA" in filenames
+        for zinfo in wf.filelist:
+            assert zinfo.compress_type == compress_type
+
+
+def test_wheelfile_line_endings(wheel_paths):
+    for path in wheel_paths:
+        with ZipFile(path) as wf:
+            wheelfile = next(fn for fn in wf.filelist if fn.filename.endswith("WHEEL"))
+            wheelfile_contents = wf.read(wheelfile)
+            assert b"\r" not in wheelfile_contents
+
+
+def test_unix_epoch_timestamps(dummy_dist, monkeypatch, tmp_path):
+    monkeypatch.setenv("SOURCE_DATE_EPOCH", "0")
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(bdist_dir=str(tmp_path), build_number="2a").run()
+    with ZipFile("dist/dummy_dist-1.0-2a-py3-none-any.whl") as wf:
+        for zinfo in wf.filelist:
+            assert zinfo.date_time >= (1980, 1, 1, 0, 0, 0)  # min epoch is used
+
+
+def test_get_abi_tag_windows(monkeypatch):
+    monkeypatch.setattr(tags, "interpreter_name", lambda: "cp")
+    monkeypatch.setattr(sysconfig, "get_config_var", lambda x: "cp313-win_amd64")
+    assert get_abi_tag() == "cp313"
+    monkeypatch.setattr(sys, "gettotalrefcount", lambda: 1, False)
+    assert get_abi_tag() == "cp313d"
+    monkeypatch.setattr(sysconfig, "get_config_var", lambda x: "cp313t-win_amd64")
+    assert get_abi_tag() == "cp313td"
+    monkeypatch.delattr(sys, "gettotalrefcount")
+    assert get_abi_tag() == "cp313t"
+
+
+def test_get_abi_tag_pypy_old(monkeypatch):
+    monkeypatch.setattr(tags, "interpreter_name", lambda: "pp")
+    monkeypatch.setattr(sysconfig, "get_config_var", lambda x: "pypy36-pp73")
+    assert get_abi_tag() == "pypy36_pp73"
+
+
+def test_get_abi_tag_pypy_new(monkeypatch):
+    monkeypatch.setattr(sysconfig, "get_config_var", lambda x: "pypy37-pp73-darwin")
+    monkeypatch.setattr(tags, "interpreter_name", lambda: "pp")
+    assert get_abi_tag() == "pypy37_pp73"
+
+
+def test_get_abi_tag_graalpy(monkeypatch):
+    monkeypatch.setattr(
+        sysconfig, "get_config_var", lambda x: "graalpy231-310-native-x86_64-linux"
+    )
+    monkeypatch.setattr(tags, "interpreter_name", lambda: "graalpy")
+    assert get_abi_tag() == "graalpy231_310_native"
+
+
+def test_get_abi_tag_fallback(monkeypatch):
+    monkeypatch.setattr(sysconfig, "get_config_var", lambda x: "unknown-python-310")
+    monkeypatch.setattr(tags, "interpreter_name", lambda: "unknown-python")
+    assert get_abi_tag() == "unknown_python_310"
+
+
+def test_platform_with_space(dummy_dist, monkeypatch):
+    """Ensure building on platforms with a space in the name succeed."""
+    monkeypatch.chdir(dummy_dist)
+    bdist_wheel_cmd(plat_name="isilon onefs").run()
+
+
+def test_data_dir_with_tag_build(monkeypatch, tmp_path):
+    """
+    Setuptools allow authors to set PEP 440's local version segments
+    using ``egg_info.tag_build``. This should be reflected not only in the
+    ``.whl`` file name, but also in the ``.dist-info`` and ``.data`` dirs.
+    See pypa/setuptools#3997.
+    """
+    monkeypatch.chdir(tmp_path)
+    files = {
+        "setup.py": """
+            from setuptools import setup
+            setup(headers=["hello.h"])
+            """,
+        "setup.cfg": """
+            [metadata]
+            name = test
+            version = 1.0
+
+            [options.data_files]
+            hello/world = file.txt
+
+            [egg_info]
+            tag_build = +what
+            tag_date = 0
+            """,
+        "file.txt": "",
+        "hello.h": "",
+    }
+    for file, content in files.items():
+        with open(file, "w", encoding="utf-8") as fh:
+            fh.write(cleandoc(content))
+
+    bdist_wheel_cmd().run()
+
+    # Ensure .whl, .dist-info and .data contain the local segment
+    wheel_path = "dist/test-1.0+what-py3-none-any.whl"
+    assert os.path.exists(wheel_path)
+    entries = set(ZipFile(wheel_path).namelist())
+    for expected in (
+        "test-1.0+what.data/headers/hello.h",
+        "test-1.0+what.data/data/hello/world/file.txt",
+        "test-1.0+what.dist-info/METADATA",
+        "test-1.0+what.dist-info/WHEEL",
+    ):
+        assert expected in entries
+
+    for not_expected in (
+        "test.data/headers/hello.h",
+        "test-1.0.data/data/hello/world/file.txt",
+        "test.dist-info/METADATA",
+        "test-1.0.dist-info/WHEEL",
+    ):
+        assert not_expected not in entries
+
+
+@pytest.mark.parametrize(
+    ("reported", "expected"),
+    [("linux-x86_64", "linux_i686"), ("linux-aarch64", "linux_armv7l")],
+)
+@pytest.mark.skipif(
+    platform.system() != "Linux", reason="Only makes sense to test on Linux"
+)
+def test_platform_linux32(reported, expected, monkeypatch):
+    monkeypatch.setattr(struct, "calcsize", lambda x: 4)
+    dist = setuptools.Distribution()
+    cmd = bdist_wheel(dist)
+    cmd.plat_name = reported
+    cmd.root_is_pure = False
+    _, _, actual = cmd.get_tag()
+    assert actual == expected
+
+
+def test_no_ctypes(monkeypatch) -> None:
+    def _fake_import(name: str, *args, **kwargs):
+        if name == "ctypes":
+            raise ModuleNotFoundError(f"No module named {name}")
+
+        return importlib.__import__(name, *args, **kwargs)
+
+    with suppress(KeyError):
+        monkeypatch.delitem(sys.modules, "wheel.macosx_libfile")
+
+    # Install an importer shim that refuses to load ctypes
+    monkeypatch.setattr(builtins, "__import__", _fake_import)
+    with pytest.raises(ModuleNotFoundError, match="No module named ctypes"):
+        import wheel.macosx_libfile  # noqa: F401
+
+    # Unload and reimport the bdist_wheel command module to make sure it won't try to
+    # import ctypes
+    monkeypatch.delitem(sys.modules, "setuptools.command.bdist_wheel")
+
+    import setuptools.command.bdist_wheel  # noqa: F401
+
+
+def test_dist_info_provided(dummy_dist, monkeypatch, tmp_path):
+    monkeypatch.chdir(dummy_dist)
+    distinfo = tmp_path / "dummy_dist.dist-info"
+
+    distinfo.mkdir()
+    (distinfo / "METADATA").write_text("name: helloworld", encoding="utf-8")
+
+    # We don't control the metadata. According to PEP-517, "The hook MAY also
+    # create other files inside this directory, and a build frontend MUST
+    # preserve".
+    (distinfo / "FOO").write_text("bar", encoding="utf-8")
+
+    bdist_wheel_cmd(bdist_dir=str(tmp_path), dist_info_dir=str(distinfo)).run()
+    expected = {
+        "dummy_dist-1.0.dist-info/FOO",
+        "dummy_dist-1.0.dist-info/RECORD",
+    }
+    with ZipFile("dist/dummy_dist-1.0-py3-none-any.whl") as wf:
+        files_found = set(wf.namelist())
+    # Check that all expected files are there.
+    assert expected - files_found == set()
+    # Make sure there is no accidental egg-info bleeding into the wheel.
+    assert not [path for path in files_found if 'egg-info' in str(path)]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build.py
new file mode 100644
index 00000000..f0f1d9dc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build.py
@@ -0,0 +1,33 @@
+from setuptools import Command
+from setuptools.command.build import build
+from setuptools.dist import Distribution
+
+
+def test_distribution_gives_setuptools_build_obj(tmpdir_cwd):
+    """
+    Check that the setuptools Distribution uses the
+    setuptools specific build object.
+    """
+
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['build'],
+            packages=[],
+            package_data={'': ['path/*']},
+        )
+    )
+    assert isinstance(dist.get_command_obj("build"), build)
+
+
+class Subcommand(Command):
+    """Dummy command to be used in tests"""
+
+    def initialize_options(self):
+        pass
+
+    def finalize_options(self):
+        pass
+
+    def run(self):
+        raise NotImplementedError("just to check if the command runs")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_clib.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_clib.py
new file mode 100644
index 00000000..b5315df4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_clib.py
@@ -0,0 +1,84 @@
+import random
+from unittest import mock
+
+import pytest
+
+from setuptools.command.build_clib import build_clib
+from setuptools.dist import Distribution
+
+from distutils.errors import DistutilsSetupError
+
+
+class TestBuildCLib:
+    @mock.patch('setuptools.command.build_clib.newer_pairwise_group')
+    def test_build_libraries(self, mock_newer):
+        dist = Distribution()
+        cmd = build_clib(dist)
+
+        # this will be a long section, just making sure all
+        # exceptions are properly raised
+        libs = [('example', {'sources': 'broken.c'})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.build_libraries(libs)
+
+        obj_deps = 'some_string'
+        libs = [('example', {'sources': ['source.c'], 'obj_deps': obj_deps})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.build_libraries(libs)
+
+        obj_deps = {'': ''}
+        libs = [('example', {'sources': ['source.c'], 'obj_deps': obj_deps})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.build_libraries(libs)
+
+        obj_deps = {'source.c': ''}
+        libs = [('example', {'sources': ['source.c'], 'obj_deps': obj_deps})]
+        with pytest.raises(DistutilsSetupError):
+            cmd.build_libraries(libs)
+
+        # with that out of the way, let's see if the crude dependency
+        # system works
+        cmd.compiler = mock.MagicMock(spec=cmd.compiler)
+        mock_newer.return_value = ([], [])
+
+        obj_deps = {'': ('global.h',), 'example.c': ('example.h',)}
+        libs = [('example', {'sources': ['example.c'], 'obj_deps': obj_deps})]
+
+        cmd.build_libraries(libs)
+        assert [['example.c', 'global.h', 'example.h']] in mock_newer.call_args[0]
+        assert not cmd.compiler.compile.called
+        assert cmd.compiler.create_static_lib.call_count == 1
+
+        # reset the call numbers so we can test again
+        cmd.compiler.reset_mock()
+
+        mock_newer.return_value = ''  # anything as long as it's not ([],[])
+        cmd.build_libraries(libs)
+        assert cmd.compiler.compile.call_count == 1
+        assert cmd.compiler.create_static_lib.call_count == 1
+
+    @mock.patch('setuptools.command.build_clib.newer_pairwise_group')
+    def test_build_libraries_reproducible(self, mock_newer):
+        dist = Distribution()
+        cmd = build_clib(dist)
+
+        # with that out of the way, let's see if the crude dependency
+        # system works
+        cmd.compiler = mock.MagicMock(spec=cmd.compiler)
+        mock_newer.return_value = ([], [])
+
+        original_sources = ['a-example.c', 'example.c']
+        sources = original_sources
+
+        obj_deps = {'': ('global.h',), 'example.c': ('example.h',)}
+        libs = [('example', {'sources': sources, 'obj_deps': obj_deps})]
+
+        cmd.build_libraries(libs)
+        computed_call_args = mock_newer.call_args[0]
+
+        while sources == original_sources:
+            sources = random.sample(original_sources, len(original_sources))
+        libs = [('example', {'sources': sources, 'obj_deps': obj_deps})]
+
+        cmd.build_libraries(libs)
+        assert computed_call_args == mock_newer.call_args[0]
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_ext.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_ext.py
new file mode 100644
index 00000000..c7b60ac3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_ext.py
@@ -0,0 +1,293 @@
+from __future__ import annotations
+
+import os
+import sys
+from importlib.util import cache_from_source as _compiled_file_name
+
+import pytest
+from jaraco import path
+
+from setuptools.command.build_ext import build_ext, get_abi3_suffix
+from setuptools.dist import Distribution
+from setuptools.errors import CompileError
+from setuptools.extension import Extension
+
+from . import environment
+from .textwrap import DALS
+
+import distutils.command.build_ext as orig
+from distutils.sysconfig import get_config_var
+
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+
+class TestBuildExt:
+    def test_get_ext_filename(self):
+        """
+        Setuptools needs to give back the same
+        result as distutils, even if the fullname
+        is not in ext_map.
+        """
+        dist = Distribution()
+        cmd = build_ext(dist)
+        cmd.ext_map['foo/bar'] = ''
+        res = cmd.get_ext_filename('foo')
+        wanted = orig.build_ext.get_ext_filename(cmd, 'foo')
+        assert res == wanted
+
+    def test_abi3_filename(self):
+        """
+        Filename needs to be loadable by several versions
+        of Python 3 if 'is_abi3' is truthy on Extension()
+        """
+        print(get_abi3_suffix())
+
+        extension = Extension('spam.eggs', ['eggs.c'], py_limited_api=True)
+        dist = Distribution(dict(ext_modules=[extension]))
+        cmd = build_ext(dist)
+        cmd.finalize_options()
+        assert 'spam.eggs' in cmd.ext_map
+        res = cmd.get_ext_filename('spam.eggs')
+
+        if not get_abi3_suffix():
+            assert res.endswith(get_config_var('EXT_SUFFIX'))
+        elif sys.platform == 'win32':
+            assert res.endswith('eggs.pyd')
+        else:
+            assert 'abi3' in res
+
+    def test_ext_suffix_override(self):
+        """
+        SETUPTOOLS_EXT_SUFFIX variable always overrides
+        default extension options.
+        """
+        dist = Distribution()
+        cmd = build_ext(dist)
+        cmd.ext_map['for_abi3'] = ext = Extension(
+            'for_abi3',
+            ['s.c'],
+            # Override shouldn't affect abi3 modules
+            py_limited_api=True,
+        )
+        # Mock value needed to pass tests
+        ext._links_to_dynamic = False
+
+        if not IS_PYPY:
+            expect = cmd.get_ext_filename('for_abi3')
+        else:
+            # PyPy builds do not use ABI3 tag, so they will
+            # also get the overridden suffix.
+            expect = 'for_abi3.test-suffix'
+
+        try:
+            os.environ['SETUPTOOLS_EXT_SUFFIX'] = '.test-suffix'
+            res = cmd.get_ext_filename('normal')
+            assert 'normal.test-suffix' == res
+            res = cmd.get_ext_filename('for_abi3')
+            assert expect == res
+        finally:
+            del os.environ['SETUPTOOLS_EXT_SUFFIX']
+
+    def dist_with_example(self):
+        files = {
+            "src": {"mypkg": {"subpkg": {"ext2.c": ""}}},
+            "c-extensions": {"ext1": {"main.c": ""}},
+        }
+
+        ext1 = Extension("mypkg.ext1", ["c-extensions/ext1/main.c"])
+        ext2 = Extension("mypkg.subpkg.ext2", ["src/mypkg/subpkg/ext2.c"])
+        ext3 = Extension("ext3", ["c-extension/ext3.c"])
+
+        path.build(files)
+        return Distribution({
+            "script_name": "%test%",
+            "ext_modules": [ext1, ext2, ext3],
+            "package_dir": {"": "src"},
+        })
+
+    def test_get_outputs(self, tmpdir_cwd, monkeypatch):
+        monkeypatch.setenv('SETUPTOOLS_EXT_SUFFIX', '.mp3')  # make test OS-independent
+        monkeypatch.setattr('setuptools.command.build_ext.use_stubs', False)
+        dist = self.dist_with_example()
+
+        # Regular build: get_outputs not empty, but get_output_mappings is empty
+        build_ext = dist.get_command_obj("build_ext")
+        build_ext.editable_mode = False
+        build_ext.ensure_finalized()
+        build_lib = build_ext.build_lib.replace(os.sep, "/")
+        outputs = [x.replace(os.sep, "/") for x in build_ext.get_outputs()]
+        assert outputs == [
+            f"{build_lib}/ext3.mp3",
+            f"{build_lib}/mypkg/ext1.mp3",
+            f"{build_lib}/mypkg/subpkg/ext2.mp3",
+        ]
+        assert build_ext.get_output_mapping() == {}
+
+        # Editable build: get_output_mappings should contain everything in get_outputs
+        dist.reinitialize_command("build_ext")
+        build_ext.editable_mode = True
+        build_ext.ensure_finalized()
+        mapping = {
+            k.replace(os.sep, "/"): v.replace(os.sep, "/")
+            for k, v in build_ext.get_output_mapping().items()
+        }
+        assert mapping == {
+            f"{build_lib}/ext3.mp3": "src/ext3.mp3",
+            f"{build_lib}/mypkg/ext1.mp3": "src/mypkg/ext1.mp3",
+            f"{build_lib}/mypkg/subpkg/ext2.mp3": "src/mypkg/subpkg/ext2.mp3",
+        }
+
+    def test_get_output_mapping_with_stub(self, tmpdir_cwd, monkeypatch):
+        monkeypatch.setenv('SETUPTOOLS_EXT_SUFFIX', '.mp3')  # make test OS-independent
+        monkeypatch.setattr('setuptools.command.build_ext.use_stubs', True)
+        dist = self.dist_with_example()
+
+        # Editable build should create compiled stubs (.pyc files only, no .py)
+        build_ext = dist.get_command_obj("build_ext")
+        build_ext.editable_mode = True
+        build_ext.ensure_finalized()
+        for ext in build_ext.extensions:
+            monkeypatch.setattr(ext, "_needs_stub", True)
+
+        build_lib = build_ext.build_lib.replace(os.sep, "/")
+        mapping = {
+            k.replace(os.sep, "/"): v.replace(os.sep, "/")
+            for k, v in build_ext.get_output_mapping().items()
+        }
+
+        def C(file):
+            """Make it possible to do comparisons and tests in a OS-independent way"""
+            return _compiled_file_name(file).replace(os.sep, "/")
+
+        assert mapping == {
+            C(f"{build_lib}/ext3.py"): C("src/ext3.py"),
+            f"{build_lib}/ext3.mp3": "src/ext3.mp3",
+            C(f"{build_lib}/mypkg/ext1.py"): C("src/mypkg/ext1.py"),
+            f"{build_lib}/mypkg/ext1.mp3": "src/mypkg/ext1.mp3",
+            C(f"{build_lib}/mypkg/subpkg/ext2.py"): C("src/mypkg/subpkg/ext2.py"),
+            f"{build_lib}/mypkg/subpkg/ext2.mp3": "src/mypkg/subpkg/ext2.mp3",
+        }
+
+        # Ensure only the compiled stubs are present not the raw .py stub
+        assert f"{build_lib}/mypkg/ext1.py" not in mapping
+        assert f"{build_lib}/mypkg/subpkg/ext2.py" not in mapping
+
+        # Visualize what the cached stub files look like
+        example_stub = C(f"{build_lib}/mypkg/ext1.py")
+        assert example_stub in mapping
+        assert example_stub.startswith(f"{build_lib}/mypkg/__pycache__/ext1")
+        assert example_stub.endswith(".pyc")
+
+
+class TestBuildExtInplace:
+    def get_build_ext_cmd(self, optional: bool, **opts) -> build_ext:
+        files: dict[str, str | dict[str, dict[str, str]]] = {
+            "eggs.c": "#include missingheader.h\n",
+            ".build": {"lib": {}, "tmp": {}},
+        }
+        path.build(files)
+        extension = Extension('spam.eggs', ['eggs.c'], optional=optional)
+        dist = Distribution(dict(ext_modules=[extension]))
+        dist.script_name = 'setup.py'
+        cmd = build_ext(dist)
+        vars(cmd).update(build_lib=".build/lib", build_temp=".build/tmp", **opts)
+        cmd.ensure_finalized()
+        return cmd
+
+    def get_log_messages(self, caplog, capsys):
+        """
+        Historically, distutils "logged" by printing to sys.std*.
+        Later versions adopted the logging framework. Grab
+        messages regardless of how they were captured.
+        """
+        std = capsys.readouterr()
+        return std.out.splitlines() + std.err.splitlines() + caplog.messages
+
+    def test_optional(self, tmpdir_cwd, caplog, capsys):
+        """
+        If optional extensions fail to build, setuptools should show the error
+        in the logs but not fail to build
+        """
+        cmd = self.get_build_ext_cmd(optional=True, inplace=True)
+        cmd.run()
+        assert any(
+            'build_ext: building extension "spam.eggs" failed'
+            for msg in self.get_log_messages(caplog, capsys)
+        )
+        # No compile error exception should be raised
+
+    def test_non_optional(self, tmpdir_cwd):
+        # Non-optional extensions should raise an exception
+        cmd = self.get_build_ext_cmd(optional=False, inplace=True)
+        with pytest.raises(CompileError):
+            cmd.run()
+
+
+def test_build_ext_config_handling(tmpdir_cwd):
+    files = {
+        'setup.py': DALS(
+            """
+            from setuptools import Extension, setup
+            setup(
+                name='foo',
+                version='0.0.0',
+                ext_modules=[Extension('foo', ['foo.c'])],
+            )
+            """
+        ),
+        'foo.c': DALS(
+            """
+            #include "Python.h"
+
+            #if PY_MAJOR_VERSION >= 3
+
+            static struct PyModuleDef moduledef = {
+                    PyModuleDef_HEAD_INIT,
+                    "foo",
+                    NULL,
+                    0,
+                    NULL,
+                    NULL,
+                    NULL,
+                    NULL,
+                    NULL
+            };
+
+            #define INITERROR return NULL
+
+            PyMODINIT_FUNC PyInit_foo(void)
+
+            #else
+
+            #define INITERROR return
+
+            void initfoo(void)
+
+            #endif
+            {
+            #if PY_MAJOR_VERSION >= 3
+                PyObject *module = PyModule_Create(&moduledef);
+            #else
+                PyObject *module = Py_InitModule("extension", NULL);
+            #endif
+                if (module == NULL)
+                    INITERROR;
+            #if PY_MAJOR_VERSION >= 3
+                return module;
+            #endif
+            }
+            """
+        ),
+        'setup.cfg': DALS(
+            """
+            [build]
+            build_base = foo_build
+            """
+        ),
+    }
+    path.build(files)
+    code, (stdout, stderr) = environment.run_setup_py(
+        cmd=['build'],
+        data_stream=(0, 2),
+    )
+    assert code == 0, f'\nSTDOUT:\n{stdout}\nSTDERR:\n{stderr}'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_meta.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_meta.py
new file mode 100644
index 00000000..121f4090
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_meta.py
@@ -0,0 +1,970 @@
+import contextlib
+import importlib
+import os
+import re
+import shutil
+import signal
+import sys
+import tarfile
+from concurrent import futures
+from pathlib import Path
+from typing import Any, Callable
+from zipfile import ZipFile
+
+import pytest
+from jaraco import path
+from packaging.requirements import Requirement
+
+from .textwrap import DALS
+
+SETUP_SCRIPT_STUB = "__import__('setuptools').setup()"
+
+
+TIMEOUT = int(os.getenv("TIMEOUT_BACKEND_TEST", "180"))  # in seconds
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+
+pytestmark = pytest.mark.skipif(
+    sys.platform == "win32" and IS_PYPY,
+    reason="The combination of PyPy + Windows + pytest-xdist + ProcessPoolExecutor "
+    "is flaky and problematic",
+)
+
+
+class BuildBackendBase:
+    def __init__(self, cwd='.', env=None, backend_name='setuptools.build_meta'):
+        self.cwd = cwd
+        self.env = env or {}
+        self.backend_name = backend_name
+
+
+class BuildBackend(BuildBackendBase):
+    """PEP 517 Build Backend"""
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.pool = futures.ProcessPoolExecutor(max_workers=1)
+
+    def __getattr__(self, name: str) -> Callable[..., Any]:
+        """Handles arbitrary function invocations on the build backend."""
+
+        def method(*args, **kw):
+            root = os.path.abspath(self.cwd)
+            caller = BuildBackendCaller(root, self.env, self.backend_name)
+            pid = None
+            try:
+                pid = self.pool.submit(os.getpid).result(TIMEOUT)
+                return self.pool.submit(caller, name, *args, **kw).result(TIMEOUT)
+            except futures.TimeoutError:
+                self.pool.shutdown(wait=False)  # doesn't stop already running processes
+                self._kill(pid)
+                pytest.xfail(f"Backend did not respond before timeout ({TIMEOUT} s)")
+            except (futures.process.BrokenProcessPool, MemoryError, OSError):
+                if IS_PYPY:
+                    pytest.xfail("PyPy frequently fails tests with ProcessPoolExector")
+                raise
+
+        return method
+
+    def _kill(self, pid):
+        if pid is None:
+            return
+        with contextlib.suppress(ProcessLookupError, OSError):
+            os.kill(pid, signal.SIGTERM if os.name == "nt" else signal.SIGKILL)
+
+
+class BuildBackendCaller(BuildBackendBase):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        (self.backend_name, _, self.backend_obj) = self.backend_name.partition(':')
+
+    def __call__(self, name, *args, **kw):
+        """Handles arbitrary function invocations on the build backend."""
+        os.chdir(self.cwd)
+        os.environ.update(self.env)
+        mod = importlib.import_module(self.backend_name)
+
+        if self.backend_obj:
+            backend = getattr(mod, self.backend_obj)
+        else:
+            backend = mod
+
+        return getattr(backend, name)(*args, **kw)
+
+
+defns = [
+    {  # simple setup.py script
+        'setup.py': DALS(
+            """
+            __import__('setuptools').setup(
+                name='foo',
+                version='0.0.0',
+                py_modules=['hello'],
+                setup_requires=['six'],
+            )
+            """
+        ),
+        'hello.py': DALS(
+            """
+            def run():
+                print('hello')
+            """
+        ),
+    },
+    {  # setup.py that relies on __name__
+        'setup.py': DALS(
+            """
+            assert __name__ == '__main__'
+            __import__('setuptools').setup(
+                name='foo',
+                version='0.0.0',
+                py_modules=['hello'],
+                setup_requires=['six'],
+            )
+            """
+        ),
+        'hello.py': DALS(
+            """
+            def run():
+                print('hello')
+            """
+        ),
+    },
+    {  # setup.py script that runs arbitrary code
+        'setup.py': DALS(
+            """
+            variable = True
+            def function():
+                return variable
+            assert variable
+            __import__('setuptools').setup(
+                name='foo',
+                version='0.0.0',
+                py_modules=['hello'],
+                setup_requires=['six'],
+            )
+            """
+        ),
+        'hello.py': DALS(
+            """
+            def run():
+                print('hello')
+            """
+        ),
+    },
+    {  # setup.py script that constructs temp files to be included in the distribution
+        'setup.py': DALS(
+            """
+            # Some packages construct files on the fly, include them in the package,
+            # and immediately remove them after `setup()` (e.g. pybind11==2.9.1).
+            # Therefore, we cannot use `distutils.core.run_setup(..., stop_after=...)`
+            # to obtain a distribution object first, and then run the distutils
+            # commands later, because these files will be removed in the meantime.
+
+            with open('world.py', 'w', encoding="utf-8") as f:
+                f.write('x = 42')
+
+            try:
+                __import__('setuptools').setup(
+                    name='foo',
+                    version='0.0.0',
+                    py_modules=['world'],
+                    setup_requires=['six'],
+                )
+            finally:
+                # Some packages will clean temporary files
+                __import__('os').unlink('world.py')
+            """
+        ),
+    },
+    {  # setup.cfg only
+        'setup.cfg': DALS(
+            """
+        [metadata]
+        name = foo
+        version = 0.0.0
+
+        [options]
+        py_modules=hello
+        setup_requires=six
+        """
+        ),
+        'hello.py': DALS(
+            """
+        def run():
+            print('hello')
+        """
+        ),
+    },
+    {  # setup.cfg and setup.py
+        'setup.cfg': DALS(
+            """
+        [metadata]
+        name = foo
+        version = 0.0.0
+
+        [options]
+        py_modules=hello
+        setup_requires=six
+        """
+        ),
+        'setup.py': "__import__('setuptools').setup()",
+        'hello.py': DALS(
+            """
+        def run():
+            print('hello')
+        """
+        ),
+    },
+]
+
+
+class TestBuildMetaBackend:
+    backend_name = 'setuptools.build_meta'
+
+    def get_build_backend(self):
+        return BuildBackend(backend_name=self.backend_name)
+
+    @pytest.fixture(params=defns)
+    def build_backend(self, tmpdir, request):
+        path.build(request.param, prefix=str(tmpdir))
+        with tmpdir.as_cwd():
+            yield self.get_build_backend()
+
+    def test_get_requires_for_build_wheel(self, build_backend):
+        actual = build_backend.get_requires_for_build_wheel()
+        expected = ['six']
+        assert sorted(actual) == sorted(expected)
+
+    def test_get_requires_for_build_sdist(self, build_backend):
+        actual = build_backend.get_requires_for_build_sdist()
+        expected = ['six']
+        assert sorted(actual) == sorted(expected)
+
+    def test_build_wheel(self, build_backend):
+        dist_dir = os.path.abspath('pip-wheel')
+        os.makedirs(dist_dir)
+        wheel_name = build_backend.build_wheel(dist_dir)
+
+        wheel_file = os.path.join(dist_dir, wheel_name)
+        assert os.path.isfile(wheel_file)
+
+        # Temporary files should be removed
+        assert not os.path.isfile('world.py')
+
+        with ZipFile(wheel_file) as zipfile:
+            wheel_contents = set(zipfile.namelist())
+
+        # Each one of the examples have a single module
+        # that should be included in the distribution
+        python_scripts = (f for f in wheel_contents if f.endswith('.py'))
+        modules = [f for f in python_scripts if not f.endswith('setup.py')]
+        assert len(modules) == 1
+
+    @pytest.mark.parametrize('build_type', ('wheel', 'sdist'))
+    def test_build_with_existing_file_present(self, build_type, tmpdir_cwd):
+        # Building a sdist/wheel should still succeed if there's
+        # already a sdist/wheel in the destination directory.
+        files = {
+            'setup.py': "from setuptools import setup\nsetup()",
+            'VERSION': "0.0.1",
+            'setup.cfg': DALS(
+                """
+                [metadata]
+                name = foo
+                version = file: VERSION
+                """
+            ),
+            'pyproject.toml': DALS(
+                """
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+                """
+            ),
+        }
+
+        path.build(files)
+
+        dist_dir = os.path.abspath('preexisting-' + build_type)
+
+        build_backend = self.get_build_backend()
+        build_method = getattr(build_backend, 'build_' + build_type)
+
+        # Build a first sdist/wheel.
+        # Note: this also check the destination directory is
+        # successfully created if it does not exist already.
+        first_result = build_method(dist_dir)
+
+        # Change version.
+        with open("VERSION", "wt", encoding="utf-8") as version_file:
+            version_file.write("0.0.2")
+
+        # Build a *second* sdist/wheel.
+        second_result = build_method(dist_dir)
+
+        assert os.path.isfile(os.path.join(dist_dir, first_result))
+        assert first_result != second_result
+
+        # And if rebuilding the exact same sdist/wheel?
+        open(os.path.join(dist_dir, second_result), 'wb').close()
+        third_result = build_method(dist_dir)
+        assert third_result == second_result
+        assert os.path.getsize(os.path.join(dist_dir, third_result)) > 0
+
+    @pytest.mark.parametrize("setup_script", [None, SETUP_SCRIPT_STUB])
+    def test_build_with_pyproject_config(self, tmpdir, setup_script):
+        files = {
+            'pyproject.toml': DALS(
+                """
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+
+                [project]
+                name = "foo"
+                license = {text = "MIT"}
+                description = "This is a Python package"
+                dynamic = ["version", "readme"]
+                classifiers = [
+                    "Development Status :: 5 - Production/Stable",
+                    "Intended Audience :: Developers"
+                ]
+                urls = {Homepage = "http://github.com"}
+                dependencies = [
+                    "appdirs",
+                ]
+
+                [project.optional-dependencies]
+                all = [
+                    "tomli>=1",
+                    "pyscaffold>=4,<5",
+                    'importlib; python_version == "2.6"',
+                ]
+
+                [project.scripts]
+                foo = "foo.cli:main"
+
+                [tool.setuptools]
+                zip-safe = false
+                package-dir = {"" = "src"}
+                packages = {find = {where = ["src"]}}
+                license-files = ["LICENSE*"]
+
+                [tool.setuptools.dynamic]
+                version = {attr = "foo.__version__"}
+                readme = {file = "README.rst"}
+
+                [tool.distutils.sdist]
+                formats = "gztar"
+                """
+            ),
+            "MANIFEST.in": DALS(
+                """
+                global-include *.py *.txt
+                global-exclude *.py[cod]
+                """
+            ),
+            "README.rst": "This is a ``README``",
+            "LICENSE.txt": "---- placeholder MIT license ----",
+            "src": {
+                "foo": {
+                    "__init__.py": "__version__ = '0.1'",
+                    "__init__.pyi": "__version__: str",
+                    "cli.py": "def main(): print('hello world')",
+                    "data.txt": "def main(): print('hello world')",
+                    "py.typed": "",
+                }
+            },
+        }
+        if setup_script:
+            files["setup.py"] = setup_script
+
+        build_backend = self.get_build_backend()
+        with tmpdir.as_cwd():
+            path.build(files)
+            sdist_path = build_backend.build_sdist("temp")
+            wheel_file = build_backend.build_wheel("temp")
+
+        with tarfile.open(os.path.join(tmpdir, "temp", sdist_path)) as tar:
+            sdist_contents = set(tar.getnames())
+
+        with ZipFile(os.path.join(tmpdir, "temp", wheel_file)) as zipfile:
+            wheel_contents = set(zipfile.namelist())
+            metadata = str(zipfile.read("foo-0.1.dist-info/METADATA"), "utf-8")
+            license = str(zipfile.read("foo-0.1.dist-info/LICENSE.txt"), "utf-8")
+            epoints = str(zipfile.read("foo-0.1.dist-info/entry_points.txt"), "utf-8")
+
+        assert sdist_contents - {"foo-0.1/setup.py"} == {
+            'foo-0.1',
+            'foo-0.1/LICENSE.txt',
+            'foo-0.1/MANIFEST.in',
+            'foo-0.1/PKG-INFO',
+            'foo-0.1/README.rst',
+            'foo-0.1/pyproject.toml',
+            'foo-0.1/setup.cfg',
+            'foo-0.1/src',
+            'foo-0.1/src/foo',
+            'foo-0.1/src/foo/__init__.py',
+            'foo-0.1/src/foo/__init__.pyi',
+            'foo-0.1/src/foo/cli.py',
+            'foo-0.1/src/foo/data.txt',
+            'foo-0.1/src/foo/py.typed',
+            'foo-0.1/src/foo.egg-info',
+            'foo-0.1/src/foo.egg-info/PKG-INFO',
+            'foo-0.1/src/foo.egg-info/SOURCES.txt',
+            'foo-0.1/src/foo.egg-info/dependency_links.txt',
+            'foo-0.1/src/foo.egg-info/entry_points.txt',
+            'foo-0.1/src/foo.egg-info/requires.txt',
+            'foo-0.1/src/foo.egg-info/top_level.txt',
+            'foo-0.1/src/foo.egg-info/not-zip-safe',
+        }
+        assert wheel_contents == {
+            "foo/__init__.py",
+            "foo/__init__.pyi",  # include type information by default
+            "foo/cli.py",
+            "foo/data.txt",  # include_package_data defaults to True
+            "foo/py.typed",  # include type information by default
+            "foo-0.1.dist-info/LICENSE.txt",
+            "foo-0.1.dist-info/METADATA",
+            "foo-0.1.dist-info/WHEEL",
+            "foo-0.1.dist-info/entry_points.txt",
+            "foo-0.1.dist-info/top_level.txt",
+            "foo-0.1.dist-info/RECORD",
+        }
+        assert license == "---- placeholder MIT license ----"
+
+        for line in (
+            "Summary: This is a Python package",
+            "License: MIT",
+            "Classifier: Intended Audience :: Developers",
+            "Requires-Dist: appdirs",
+            "Requires-Dist: " + str(Requirement('tomli>=1 ; extra == "all"')),
+            "Requires-Dist: "
+            + str(Requirement('importlib; python_version=="2.6" and extra =="all"')),
+        ):
+            assert line in metadata, (line, metadata)
+
+        assert metadata.strip().endswith("This is a ``README``")
+        assert epoints.strip() == "[console_scripts]\nfoo = foo.cli:main"
+
+    def test_static_metadata_in_pyproject_config(self, tmpdir):
+        # Make sure static metadata in pyproject.toml is not overwritten by setup.py
+        # as required by PEP 621
+        files = {
+            'pyproject.toml': DALS(
+                """
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+
+                [project]
+                name = "foo"
+                description = "This is a Python package"
+                version = "42"
+                dependencies = ["six"]
+                """
+            ),
+            'hello.py': DALS(
+                """
+                def run():
+                    print('hello')
+                """
+            ),
+            'setup.py': DALS(
+                """
+                __import__('setuptools').setup(
+                    name='bar',
+                    version='13',
+                )
+                """
+            ),
+        }
+        build_backend = self.get_build_backend()
+        with tmpdir.as_cwd():
+            path.build(files)
+            sdist_path = build_backend.build_sdist("temp")
+            wheel_file = build_backend.build_wheel("temp")
+
+        assert (tmpdir / "temp/foo-42.tar.gz").exists()
+        assert (tmpdir / "temp/foo-42-py3-none-any.whl").exists()
+        assert not (tmpdir / "temp/bar-13.tar.gz").exists()
+        assert not (tmpdir / "temp/bar-42.tar.gz").exists()
+        assert not (tmpdir / "temp/foo-13.tar.gz").exists()
+        assert not (tmpdir / "temp/bar-13-py3-none-any.whl").exists()
+        assert not (tmpdir / "temp/bar-42-py3-none-any.whl").exists()
+        assert not (tmpdir / "temp/foo-13-py3-none-any.whl").exists()
+
+        with tarfile.open(os.path.join(tmpdir, "temp", sdist_path)) as tar:
+            pkg_info = str(tar.extractfile('foo-42/PKG-INFO').read(), "utf-8")
+            members = tar.getnames()
+            assert "bar-13/PKG-INFO" not in members
+
+        with ZipFile(os.path.join(tmpdir, "temp", wheel_file)) as zipfile:
+            metadata = str(zipfile.read("foo-42.dist-info/METADATA"), "utf-8")
+            members = zipfile.namelist()
+            assert "bar-13.dist-info/METADATA" not in members
+
+        for file in pkg_info, metadata:
+            for line in ("Name: foo", "Version: 42"):
+                assert line in file
+            for line in ("Name: bar", "Version: 13"):
+                assert line not in file
+
+    def test_build_sdist(self, build_backend):
+        dist_dir = os.path.abspath('pip-sdist')
+        os.makedirs(dist_dir)
+        sdist_name = build_backend.build_sdist(dist_dir)
+
+        assert os.path.isfile(os.path.join(dist_dir, sdist_name))
+
+    def test_prepare_metadata_for_build_wheel(self, build_backend):
+        dist_dir = os.path.abspath('pip-dist-info')
+        os.makedirs(dist_dir)
+
+        dist_info = build_backend.prepare_metadata_for_build_wheel(dist_dir)
+
+        assert os.path.isfile(os.path.join(dist_dir, dist_info, 'METADATA'))
+
+    def test_prepare_metadata_inplace(self, build_backend):
+        """
+        Some users might pass metadata_directory pre-populated with `.tox` or `.venv`.
+        See issue #3523.
+        """
+        for pre_existing in [
+            ".tox/python/lib/python3.10/site-packages/attrs-22.1.0.dist-info",
+            ".tox/python/lib/python3.10/site-packages/autocommand-2.2.1.dist-info",
+            ".nox/python/lib/python3.10/site-packages/build-0.8.0.dist-info",
+            ".venv/python3.10/site-packages/click-8.1.3.dist-info",
+            "venv/python3.10/site-packages/distlib-0.3.5.dist-info",
+            "env/python3.10/site-packages/docutils-0.19.dist-info",
+        ]:
+            os.makedirs(pre_existing, exist_ok=True)
+        dist_info = build_backend.prepare_metadata_for_build_wheel(".")
+        assert os.path.isfile(os.path.join(dist_info, 'METADATA'))
+
+    def test_build_sdist_explicit_dist(self, build_backend):
+        # explicitly specifying the dist folder should work
+        # the folder sdist_directory and the ``--dist-dir`` can be the same
+        dist_dir = os.path.abspath('dist')
+        sdist_name = build_backend.build_sdist(dist_dir)
+        assert os.path.isfile(os.path.join(dist_dir, sdist_name))
+
+    def test_build_sdist_version_change(self, build_backend):
+        sdist_into_directory = os.path.abspath("out_sdist")
+        os.makedirs(sdist_into_directory)
+
+        sdist_name = build_backend.build_sdist(sdist_into_directory)
+        assert os.path.isfile(os.path.join(sdist_into_directory, sdist_name))
+
+        # if the setup.py changes subsequent call of the build meta
+        # should still succeed, given the
+        # sdist_directory the frontend specifies is empty
+        setup_loc = os.path.abspath("setup.py")
+        if not os.path.exists(setup_loc):
+            setup_loc = os.path.abspath("setup.cfg")
+
+        with open(setup_loc, 'rt', encoding="utf-8") as file_handler:
+            content = file_handler.read()
+        with open(setup_loc, 'wt', encoding="utf-8") as file_handler:
+            file_handler.write(content.replace("version='0.0.0'", "version='0.0.1'"))
+
+        shutil.rmtree(sdist_into_directory)
+        os.makedirs(sdist_into_directory)
+
+        sdist_name = build_backend.build_sdist("out_sdist")
+        assert os.path.isfile(os.path.join(os.path.abspath("out_sdist"), sdist_name))
+
+    def test_build_sdist_pyproject_toml_exists(self, tmpdir_cwd):
+        files = {
+            'setup.py': DALS(
+                """
+                __import__('setuptools').setup(
+                    name='foo',
+                    version='0.0.0',
+                    py_modules=['hello']
+                )"""
+            ),
+            'hello.py': '',
+            'pyproject.toml': DALS(
+                """
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+                """
+            ),
+        }
+        path.build(files)
+        build_backend = self.get_build_backend()
+        targz_path = build_backend.build_sdist("temp")
+        with tarfile.open(os.path.join("temp", targz_path)) as tar:
+            assert any('pyproject.toml' in name for name in tar.getnames())
+
+    def test_build_sdist_setup_py_exists(self, tmpdir_cwd):
+        # If build_sdist is called from a script other than setup.py,
+        # ensure setup.py is included
+        path.build(defns[0])
+
+        build_backend = self.get_build_backend()
+        targz_path = build_backend.build_sdist("temp")
+        with tarfile.open(os.path.join("temp", targz_path)) as tar:
+            assert any('setup.py' in name for name in tar.getnames())
+
+    def test_build_sdist_setup_py_manifest_excluded(self, tmpdir_cwd):
+        # Ensure that MANIFEST.in can exclude setup.py
+        files = {
+            'setup.py': DALS(
+                """
+        __import__('setuptools').setup(
+            name='foo',
+            version='0.0.0',
+            py_modules=['hello']
+        )"""
+            ),
+            'hello.py': '',
+            'MANIFEST.in': DALS(
+                """
+        exclude setup.py
+        """
+            ),
+        }
+
+        path.build(files)
+
+        build_backend = self.get_build_backend()
+        targz_path = build_backend.build_sdist("temp")
+        with tarfile.open(os.path.join("temp", targz_path)) as tar:
+            assert not any('setup.py' in name for name in tar.getnames())
+
+    def test_build_sdist_builds_targz_even_if_zip_indicated(self, tmpdir_cwd):
+        files = {
+            'setup.py': DALS(
+                """
+                __import__('setuptools').setup(
+                    name='foo',
+                    version='0.0.0',
+                    py_modules=['hello']
+                )"""
+            ),
+            'hello.py': '',
+            'setup.cfg': DALS(
+                """
+                [sdist]
+                formats=zip
+                """
+            ),
+        }
+
+        path.build(files)
+
+        build_backend = self.get_build_backend()
+        build_backend.build_sdist("temp")
+
+    _relative_path_import_files = {
+        'setup.py': DALS(
+            """
+            __import__('setuptools').setup(
+                name='foo',
+                version=__import__('hello').__version__,
+                py_modules=['hello']
+            )"""
+        ),
+        'hello.py': '__version__ = "0.0.0"',
+        'setup.cfg': DALS(
+            """
+            [sdist]
+            formats=zip
+            """
+        ),
+    }
+
+    def test_build_sdist_relative_path_import(self, tmpdir_cwd):
+        path.build(self._relative_path_import_files)
+        build_backend = self.get_build_backend()
+        with pytest.raises(ImportError, match="^No module named 'hello'$"):
+            build_backend.build_sdist("temp")
+
+    _simple_pyproject_example = {
+        "pyproject.toml": DALS(
+            """
+            [project]
+            name = "proj"
+            version = "42"
+            """
+        ),
+        "src": {"proj": {"__init__.py": ""}},
+    }
+
+    def _assert_link_tree(self, parent_dir):
+        """All files in the directory should be either links or hard links"""
+        files = list(Path(parent_dir).glob("**/*"))
+        assert files  # Should not be empty
+        for file in files:
+            assert file.is_symlink() or os.stat(file).st_nlink > 0
+
+    def test_editable_without_config_settings(self, tmpdir_cwd):
+        """
+        Sanity check to ensure tests with --mode=strict are different from the ones
+        without --mode.
+
+        --mode=strict should create a local directory with a package tree.
+        The directory should not get created otherwise.
+        """
+        path.build(self._simple_pyproject_example)
+        build_backend = self.get_build_backend()
+        assert not Path("build").exists()
+        build_backend.build_editable("temp")
+        assert not Path("build").exists()
+
+    def test_build_wheel_inplace(self, tmpdir_cwd):
+        config_settings = {"--build-option": ["build_ext", "--inplace"]}
+        path.build(self._simple_pyproject_example)
+        build_backend = self.get_build_backend()
+        assert not Path("build").exists()
+        Path("build").mkdir()
+        build_backend.prepare_metadata_for_build_wheel("build", config_settings)
+        build_backend.build_wheel("build", config_settings)
+        assert Path("build/proj-42-py3-none-any.whl").exists()
+
+    @pytest.mark.parametrize("config_settings", [{"editable-mode": "strict"}])
+    def test_editable_with_config_settings(self, tmpdir_cwd, config_settings):
+        path.build({**self._simple_pyproject_example, '_meta': {}})
+        assert not Path("build").exists()
+        build_backend = self.get_build_backend()
+        build_backend.prepare_metadata_for_build_editable("_meta", config_settings)
+        build_backend.build_editable("temp", config_settings, "_meta")
+        self._assert_link_tree(next(Path("build").glob("__editable__.*")))
+
+    @pytest.mark.parametrize(
+        ("setup_literal", "requirements"),
+        [
+            ("'foo'", ['foo']),
+            ("['foo']", ['foo']),
+            (r"'foo\n'", ['foo']),
+            (r"'foo\n\n'", ['foo']),
+            ("['foo', 'bar']", ['foo', 'bar']),
+            (r"'# Has a comment line\nfoo'", ['foo']),
+            (r"'foo # Has an inline comment'", ['foo']),
+            (r"'foo \\\n >=3.0'", ['foo>=3.0']),
+            (r"'foo\nbar'", ['foo', 'bar']),
+            (r"'foo\nbar\n'", ['foo', 'bar']),
+            (r"['foo\n', 'bar\n']", ['foo', 'bar']),
+        ],
+    )
+    @pytest.mark.parametrize('use_wheel', [True, False])
+    def test_setup_requires(self, setup_literal, requirements, use_wheel, tmpdir_cwd):
+        files = {
+            'setup.py': DALS(
+                """
+                from setuptools import setup
+
+                setup(
+                    name="qux",
+                    version="0.0.0",
+                    py_modules=["hello"],
+                    setup_requires={setup_literal},
+                )
+            """
+            ).format(setup_literal=setup_literal),
+            'hello.py': DALS(
+                """
+            def run():
+                print('hello')
+            """
+            ),
+        }
+
+        path.build(files)
+
+        build_backend = self.get_build_backend()
+
+        if use_wheel:
+            get_requires = build_backend.get_requires_for_build_wheel
+        else:
+            get_requires = build_backend.get_requires_for_build_sdist
+
+        # Ensure that the build requirements are properly parsed
+        expected = sorted(requirements)
+        actual = get_requires()
+
+        assert expected == sorted(actual)
+
+    def test_setup_requires_with_auto_discovery(self, tmpdir_cwd):
+        # Make sure patches introduced to retrieve setup_requires don't accidentally
+        # activate auto-discovery and cause problems due to the incomplete set of
+        # attributes passed to MinimalDistribution
+        files = {
+            'pyproject.toml': DALS(
+                """
+                [project]
+                name = "proj"
+                version = "42"
+            """
+            ),
+            "setup.py": DALS(
+                """
+                __import__('setuptools').setup(
+                    setup_requires=["foo"],
+                    py_modules = ["hello", "world"]
+                )
+            """
+            ),
+            'hello.py': "'hello'",
+            'world.py': "'world'",
+        }
+        path.build(files)
+        build_backend = self.get_build_backend()
+        setup_requires = build_backend.get_requires_for_build_wheel()
+        assert setup_requires == ["foo"]
+
+    def test_dont_install_setup_requires(self, tmpdir_cwd):
+        files = {
+            'setup.py': DALS(
+                """
+                        from setuptools import setup
+
+                        setup(
+                            name="qux",
+                            version="0.0.0",
+                            py_modules=["hello"],
+                            setup_requires=["does-not-exist >99"],
+                        )
+                    """
+            ),
+            'hello.py': DALS(
+                """
+                    def run():
+                        print('hello')
+                    """
+            ),
+        }
+
+        path.build(files)
+
+        build_backend = self.get_build_backend()
+
+        dist_dir = os.path.abspath('pip-dist-info')
+        os.makedirs(dist_dir)
+
+        # does-not-exist can't be satisfied, so if it attempts to install
+        # setup_requires, it will fail.
+        build_backend.prepare_metadata_for_build_wheel(dist_dir)
+
+    _sys_argv_0_passthrough = {
+        'setup.py': DALS(
+            """
+            import os
+            import sys
+
+            __import__('setuptools').setup(
+                name='foo',
+                version='0.0.0',
+            )
+
+            sys_argv = os.path.abspath(sys.argv[0])
+            file_path = os.path.abspath('setup.py')
+            assert sys_argv == file_path
+            """
+        )
+    }
+
+    def test_sys_argv_passthrough(self, tmpdir_cwd):
+        path.build(self._sys_argv_0_passthrough)
+        build_backend = self.get_build_backend()
+        with pytest.raises(AssertionError):
+            build_backend.build_sdist("temp")
+
+    _setup_py_file_abspath = {
+        'setup.py': DALS(
+            """
+            import os
+            assert os.path.isabs(__file__)
+            __import__('setuptools').setup(
+                name='foo',
+                version='0.0.0',
+                py_modules=['hello'],
+                setup_requires=['six'],
+            )
+            """
+        )
+    }
+
+    def test_setup_py_file_abspath(self, tmpdir_cwd):
+        path.build(self._setup_py_file_abspath)
+        build_backend = self.get_build_backend()
+        build_backend.build_sdist("temp")
+
+    @pytest.mark.parametrize('build_hook', ('build_sdist', 'build_wheel'))
+    def test_build_with_empty_setuppy(self, build_backend, build_hook):
+        files = {'setup.py': ''}
+        path.build(files)
+
+        msg = re.escape('No distribution was found.')
+        with pytest.raises(ValueError, match=msg):
+            getattr(build_backend, build_hook)("temp")
+
+
+class TestBuildMetaLegacyBackend(TestBuildMetaBackend):
+    backend_name = 'setuptools.build_meta:__legacy__'
+
+    # build_meta_legacy-specific tests
+    def test_build_sdist_relative_path_import(self, tmpdir_cwd):
+        # This must fail in build_meta, but must pass in build_meta_legacy
+        path.build(self._relative_path_import_files)
+
+        build_backend = self.get_build_backend()
+        build_backend.build_sdist("temp")
+
+    def test_sys_argv_passthrough(self, tmpdir_cwd):
+        path.build(self._sys_argv_0_passthrough)
+
+        build_backend = self.get_build_backend()
+        build_backend.build_sdist("temp")
+
+
+def test_legacy_editable_install(venv, tmpdir, tmpdir_cwd):
+    pyproject = """
+    [build-system]
+    requires = ["setuptools"]
+    build-backend = "setuptools.build_meta"
+    [project]
+    name = "myproj"
+    version = "42"
+    """
+    path.build({"pyproject.toml": DALS(pyproject), "mymod.py": ""})
+
+    # First: sanity check
+    cmd = ["pip", "install", "--no-build-isolation", "-e", "."]
+    output = venv.run(cmd, cwd=tmpdir).lower()
+    assert "running setup.py develop for myproj" not in output
+    assert "created wheel for myproj" in output
+
+    # Then: real test
+    env = {**os.environ, "SETUPTOOLS_ENABLE_FEATURES": "legacy-editable"}
+    cmd = ["pip", "install", "--no-build-isolation", "-e", "."]
+    output = venv.run(cmd, cwd=tmpdir, env=env).lower()
+    assert "running setup.py develop for myproj" in output
+
+
+@pytest.mark.filterwarnings("ignore::setuptools.SetuptoolsDeprecationWarning")
+def test_sys_exit_0_in_setuppy(monkeypatch, tmp_path):
+    """Setuptools should be resilient to setup.py with ``sys.exit(0)`` (#3973)."""
+    monkeypatch.chdir(tmp_path)
+    setuppy = """
+        import sys, setuptools
+        setuptools.setup(name='foo', version='0.0.0')
+        sys.exit(0)
+        """
+    (tmp_path / "setup.py").write_text(DALS(setuppy), encoding="utf-8")
+    backend = BuildBackend(backend_name="setuptools.build_meta")
+    assert backend.get_requires_for_build_wheel() == []
+
+
+def test_system_exit_in_setuppy(monkeypatch, tmp_path):
+    monkeypatch.chdir(tmp_path)
+    setuppy = "import sys; sys.exit('some error')"
+    (tmp_path / "setup.py").write_text(setuppy, encoding="utf-8")
+    with pytest.raises(SystemExit, match="some error"):
+        backend = BuildBackend(backend_name="setuptools.build_meta")
+        backend.get_requires_for_build_wheel()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_py.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_py.py
new file mode 100644
index 00000000..e64cfa2e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_build_py.py
@@ -0,0 +1,480 @@
+import os
+import shutil
+import stat
+import warnings
+from pathlib import Path
+from unittest.mock import Mock
+
+import jaraco.path
+import pytest
+
+from setuptools import SetuptoolsDeprecationWarning
+from setuptools.dist import Distribution
+
+from .textwrap import DALS
+
+
+def test_directories_in_package_data_glob(tmpdir_cwd):
+    """
+    Directories matching the glob in package_data should
+    not be included in the package data.
+
+    Regression test for #261.
+    """
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['build_py'],
+            packages=[''],
+            package_data={'': ['path/*']},
+        )
+    )
+    os.makedirs('path/subpath')
+    dist.parse_command_line()
+    dist.run_commands()
+
+
+def test_recursive_in_package_data_glob(tmpdir_cwd):
+    """
+    Files matching recursive globs (**) in package_data should
+    be included in the package data.
+
+    #1806
+    """
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['build_py'],
+            packages=[''],
+            package_data={'': ['path/**/data']},
+        )
+    )
+    os.makedirs('path/subpath/subsubpath')
+    open('path/subpath/subsubpath/data', 'wb').close()
+
+    dist.parse_command_line()
+    dist.run_commands()
+
+    assert stat.S_ISREG(os.stat('build/lib/path/subpath/subsubpath/data').st_mode), (
+        "File is not included"
+    )
+
+
+def test_read_only(tmpdir_cwd):
+    """
+    Ensure read-only flag is not preserved in copy
+    for package modules and package data, as that
+    causes problems with deleting read-only files on
+    Windows.
+
+    #1451
+    """
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['build_py'],
+            packages=['pkg'],
+            package_data={'pkg': ['data.dat']},
+        )
+    )
+    os.makedirs('pkg')
+    open('pkg/__init__.py', 'wb').close()
+    open('pkg/data.dat', 'wb').close()
+    os.chmod('pkg/__init__.py', stat.S_IREAD)
+    os.chmod('pkg/data.dat', stat.S_IREAD)
+    dist.parse_command_line()
+    dist.run_commands()
+    shutil.rmtree('build')
+
+
+@pytest.mark.xfail(
+    'platform.system() == "Windows"',
+    reason="On Windows, files do not have executable bits",
+    raises=AssertionError,
+    strict=True,
+)
+def test_executable_data(tmpdir_cwd):
+    """
+    Ensure executable bit is preserved in copy for
+    package data, as users rely on it for scripts.
+
+    #2041
+    """
+    dist = Distribution(
+        dict(
+            script_name='setup.py',
+            script_args=['build_py'],
+            packages=['pkg'],
+            package_data={'pkg': ['run-me']},
+        )
+    )
+    os.makedirs('pkg')
+    open('pkg/__init__.py', 'wb').close()
+    open('pkg/run-me', 'wb').close()
+    os.chmod('pkg/run-me', 0o700)
+
+    dist.parse_command_line()
+    dist.run_commands()
+
+    assert os.stat('build/lib/pkg/run-me').st_mode & stat.S_IEXEC, (
+        "Script is not executable"
+    )
+
+
+EXAMPLE_WITH_MANIFEST = {
+    "setup.cfg": DALS(
+        """
+        [metadata]
+        name = mypkg
+        version = 42
+
+        [options]
+        include_package_data = True
+        packages = find:
+
+        [options.packages.find]
+        exclude = *.tests*
+        """
+    ),
+    "mypkg": {
+        "__init__.py": "",
+        "resource_file.txt": "",
+        "tests": {
+            "__init__.py": "",
+            "test_mypkg.py": "",
+            "test_file.txt": "",
+        },
+    },
+    "MANIFEST.in": DALS(
+        """
+        global-include *.py *.txt
+        global-exclude *.py[cod]
+        prune dist
+        prune build
+        prune *.egg-info
+        """
+    ),
+}
+
+
+def test_excluded_subpackages(tmpdir_cwd):
+    jaraco.path.build(EXAMPLE_WITH_MANIFEST)
+    dist = Distribution({"script_name": "%PEP 517%"})
+    dist.parse_config_files()
+
+    build_py = dist.get_command_obj("build_py")
+
+    msg = r"Python recognizes 'mypkg\.tests' as an importable package"
+    with pytest.warns(SetuptoolsDeprecationWarning, match=msg):
+        # TODO: To fix #3260 we need some transition period to deprecate the
+        # existing behavior of `include_package_data`. After the transition, we
+        # should remove the warning and fix the behaviour.
+
+        if os.getenv("SETUPTOOLS_USE_DISTUTILS") == "stdlib":
+            # pytest.warns reset the warning filter temporarily
+            # https://github.com/pytest-dev/pytest/issues/4011#issuecomment-423494810
+            warnings.filterwarnings(
+                "ignore",
+                "'encoding' argument not specified",
+                module="distutils.text_file",
+                # This warning is already fixed in pypa/distutils but not in stdlib
+            )
+
+        build_py.finalize_options()
+        build_py.run()
+
+    build_dir = Path(dist.get_command_obj("build_py").build_lib)
+    assert (build_dir / "mypkg/__init__.py").exists()
+    assert (build_dir / "mypkg/resource_file.txt").exists()
+
+    # Setuptools is configured to ignore `mypkg.tests`, therefore the following
+    # files/dirs should not be included in the distribution.
+    for f in [
+        "mypkg/tests/__init__.py",
+        "mypkg/tests/test_mypkg.py",
+        "mypkg/tests/test_file.txt",
+        "mypkg/tests",
+    ]:
+        with pytest.raises(AssertionError):
+            # TODO: Enforce the following assertion once #3260 is fixed
+            # (remove context manager and the following xfail).
+            assert not (build_dir / f).exists()
+
+    pytest.xfail("#3260")
+
+
+@pytest.mark.filterwarnings("ignore::setuptools.SetuptoolsDeprecationWarning")
+def test_existing_egg_info(tmpdir_cwd, monkeypatch):
+    """When provided with the ``existing_egg_info_dir`` attribute, build_py should not
+    attempt to run egg_info again.
+    """
+    # == Pre-condition ==
+    # Generate an egg-info dir
+    jaraco.path.build(EXAMPLE_WITH_MANIFEST)
+    dist = Distribution({"script_name": "%PEP 517%"})
+    dist.parse_config_files()
+    assert dist.include_package_data
+
+    egg_info = dist.get_command_obj("egg_info")
+    dist.run_command("egg_info")
+    egg_info_dir = next(Path(egg_info.egg_base).glob("*.egg-info"))
+    assert egg_info_dir.is_dir()
+
+    # == Setup ==
+    build_py = dist.get_command_obj("build_py")
+    build_py.finalize_options()
+    egg_info = dist.get_command_obj("egg_info")
+    egg_info_run = Mock(side_effect=egg_info.run)
+    monkeypatch.setattr(egg_info, "run", egg_info_run)
+
+    # == Remove caches ==
+    # egg_info is called when build_py looks for data_files, which gets cached.
+    # We need to ensure it is not cached yet, otherwise it may impact on the tests
+    build_py.__dict__.pop('data_files', None)
+    dist.reinitialize_command(egg_info)
+
+    # == Sanity check ==
+    # Ensure that if existing_egg_info is not given, build_py attempts to run egg_info
+    build_py.existing_egg_info_dir = None
+    build_py.run()
+    egg_info_run.assert_called()
+
+    # == Remove caches ==
+    egg_info_run.reset_mock()
+    build_py.__dict__.pop('data_files', None)
+    dist.reinitialize_command(egg_info)
+
+    # == Actual test ==
+    # Ensure that if existing_egg_info_dir is given, egg_info doesn't run
+    build_py.existing_egg_info_dir = egg_info_dir
+    build_py.run()
+    egg_info_run.assert_not_called()
+    assert build_py.data_files
+
+    # Make sure the list of outputs is actually OK
+    outputs = map(lambda x: x.replace(os.sep, "/"), build_py.get_outputs())
+    assert outputs
+    example = str(Path(build_py.build_lib, "mypkg/__init__.py")).replace(os.sep, "/")
+    assert example in outputs
+
+
+EXAMPLE_ARBITRARY_MAPPING = {
+    "pyproject.toml": DALS(
+        """
+        [project]
+        name = "mypkg"
+        version = "42"
+
+        [tool.setuptools]
+        packages = ["mypkg", "mypkg.sub1", "mypkg.sub2", "mypkg.sub2.nested"]
+
+        [tool.setuptools.package-dir]
+        "" = "src"
+        "mypkg.sub2" = "src/mypkg/_sub2"
+        "mypkg.sub2.nested" = "other"
+        """
+    ),
+    "src": {
+        "mypkg": {
+            "__init__.py": "",
+            "resource_file.txt": "",
+            "sub1": {
+                "__init__.py": "",
+                "mod1.py": "",
+            },
+            "_sub2": {
+                "mod2.py": "",
+            },
+        },
+    },
+    "other": {
+        "__init__.py": "",
+        "mod3.py": "",
+    },
+    "MANIFEST.in": DALS(
+        """
+        global-include *.py *.txt
+        global-exclude *.py[cod]
+        """
+    ),
+}
+
+
+def test_get_outputs(tmpdir_cwd):
+    jaraco.path.build(EXAMPLE_ARBITRARY_MAPPING)
+    dist = Distribution({"script_name": "%test%"})
+    dist.parse_config_files()
+
+    build_py = dist.get_command_obj("build_py")
+    build_py.editable_mode = True
+    build_py.ensure_finalized()
+    build_lib = build_py.build_lib.replace(os.sep, "/")
+    outputs = {x.replace(os.sep, "/") for x in build_py.get_outputs()}
+    assert outputs == {
+        f"{build_lib}/mypkg/__init__.py",
+        f"{build_lib}/mypkg/resource_file.txt",
+        f"{build_lib}/mypkg/sub1/__init__.py",
+        f"{build_lib}/mypkg/sub1/mod1.py",
+        f"{build_lib}/mypkg/sub2/mod2.py",
+        f"{build_lib}/mypkg/sub2/nested/__init__.py",
+        f"{build_lib}/mypkg/sub2/nested/mod3.py",
+    }
+    mapping = {
+        k.replace(os.sep, "/"): v.replace(os.sep, "/")
+        for k, v in build_py.get_output_mapping().items()
+    }
+    assert mapping == {
+        f"{build_lib}/mypkg/__init__.py": "src/mypkg/__init__.py",
+        f"{build_lib}/mypkg/resource_file.txt": "src/mypkg/resource_file.txt",
+        f"{build_lib}/mypkg/sub1/__init__.py": "src/mypkg/sub1/__init__.py",
+        f"{build_lib}/mypkg/sub1/mod1.py": "src/mypkg/sub1/mod1.py",
+        f"{build_lib}/mypkg/sub2/mod2.py": "src/mypkg/_sub2/mod2.py",
+        f"{build_lib}/mypkg/sub2/nested/__init__.py": "other/__init__.py",
+        f"{build_lib}/mypkg/sub2/nested/mod3.py": "other/mod3.py",
+    }
+
+
+class TestTypeInfoFiles:
+    PYPROJECTS = {
+        "default_pyproject": DALS(
+            """
+            [project]
+            name = "foo"
+            version = "1"
+            """
+        ),
+        "dont_include_package_data": DALS(
+            """
+            [project]
+            name = "foo"
+            version = "1"
+
+            [tool.setuptools]
+            include-package-data = false
+            """
+        ),
+        "exclude_type_info": DALS(
+            """
+            [project]
+            name = "foo"
+            version = "1"
+
+            [tool.setuptools]
+            include-package-data = false
+
+            [tool.setuptools.exclude-package-data]
+            "*" = ["py.typed", "*.pyi"]
+            """
+        ),
+    }
+
+    EXAMPLES = {
+        "simple_namespace": {
+            "directory_structure": {
+                "foo": {
+                    "bar.pyi": "",
+                    "py.typed": "",
+                    "__init__.py": "",
+                }
+            },
+            "expected_type_files": {"foo/bar.pyi", "foo/py.typed"},
+        },
+        "nested_inside_namespace": {
+            "directory_structure": {
+                "foo": {
+                    "bar": {
+                        "py.typed": "",
+                        "mod.pyi": "",
+                    }
+                }
+            },
+            "expected_type_files": {"foo/bar/mod.pyi", "foo/bar/py.typed"},
+        },
+        "namespace_nested_inside_regular": {
+            "directory_structure": {
+                "foo": {
+                    "namespace": {
+                        "foo.pyi": "",
+                    },
+                    "__init__.pyi": "",
+                    "py.typed": "",
+                }
+            },
+            "expected_type_files": {
+                "foo/namespace/foo.pyi",
+                "foo/__init__.pyi",
+                "foo/py.typed",
+            },
+        },
+    }
+
+    @pytest.mark.parametrize(
+        "pyproject",
+        [
+            "default_pyproject",
+            pytest.param(
+                "dont_include_package_data",
+                marks=pytest.mark.xfail(reason="pypa/setuptools#4350"),
+            ),
+        ],
+    )
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_type_files_included_by_default(self, tmpdir_cwd, pyproject, example):
+        structure = {
+            **self.EXAMPLES[example]["directory_structure"],
+            "pyproject.toml": self.PYPROJECTS[pyproject],
+        }
+        expected_type_files = self.EXAMPLES[example]["expected_type_files"]
+        jaraco.path.build(structure)
+
+        build_py = get_finalized_build_py()
+        outputs = get_outputs(build_py)
+        assert expected_type_files <= outputs
+
+    @pytest.mark.parametrize("pyproject", ["exclude_type_info"])
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_type_files_can_be_excluded(self, tmpdir_cwd, pyproject, example):
+        structure = {
+            **self.EXAMPLES[example]["directory_structure"],
+            "pyproject.toml": self.PYPROJECTS[pyproject],
+        }
+        expected_type_files = self.EXAMPLES[example]["expected_type_files"]
+        jaraco.path.build(structure)
+
+        build_py = get_finalized_build_py()
+        outputs = get_outputs(build_py)
+        assert expected_type_files.isdisjoint(outputs)
+
+    def test_stub_only_package(self, tmpdir_cwd):
+        structure = {
+            "pyproject.toml": DALS(
+                """
+                [project]
+                name = "foo-stubs"
+                version = "1"
+                """
+            ),
+            "foo-stubs": {"__init__.pyi": "", "bar.pyi": ""},
+        }
+        expected_type_files = {"foo-stubs/__init__.pyi", "foo-stubs/bar.pyi"}
+        jaraco.path.build(structure)
+
+        build_py = get_finalized_build_py()
+        outputs = get_outputs(build_py)
+        assert expected_type_files <= outputs
+
+
+def get_finalized_build_py(script_name="%build_py-test%"):
+    dist = Distribution({"script_name": script_name})
+    dist.parse_config_files()
+    build_py = dist.get_command_obj("build_py")
+    build_py.finalize_options()
+    return build_py
+
+
+def get_outputs(build_py):
+    build_dir = Path(build_py.build_lib)
+    return {
+        os.path.relpath(x, build_dir).replace(os.sep, "/")
+        for x in build_py.get_outputs()
+    }
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_config_discovery.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_config_discovery.py
new file mode 100644
index 00000000..b5df8203
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_config_discovery.py
@@ -0,0 +1,647 @@
+import os
+import sys
+from configparser import ConfigParser
+from itertools import product
+from typing import cast
+
+import jaraco.path
+import pytest
+from path import Path
+
+import setuptools  # noqa: F401 # force distutils.core to be patched
+from setuptools.command.sdist import sdist
+from setuptools.discovery import find_package_path, find_parent_package
+from setuptools.dist import Distribution
+from setuptools.errors import PackageDiscoveryError
+
+from .contexts import quiet
+from .integration.helpers import get_sdist_members, get_wheel_members, run
+from .textwrap import DALS
+
+import distutils.core
+
+
+class TestFindParentPackage:
+    def test_single_package(self, tmp_path):
+        # find_parent_package should find a non-namespace parent package
+        (tmp_path / "src/namespace/pkg/nested").mkdir(exist_ok=True, parents=True)
+        (tmp_path / "src/namespace/pkg/nested/__init__.py").touch()
+        (tmp_path / "src/namespace/pkg/__init__.py").touch()
+        packages = ["namespace", "namespace.pkg", "namespace.pkg.nested"]
+        assert find_parent_package(packages, {"": "src"}, tmp_path) == "namespace.pkg"
+
+    def test_multiple_toplevel(self, tmp_path):
+        # find_parent_package should return null if the given list of packages does not
+        # have a single parent package
+        multiple = ["pkg", "pkg1", "pkg2"]
+        for name in multiple:
+            (tmp_path / f"src/{name}").mkdir(exist_ok=True, parents=True)
+            (tmp_path / f"src/{name}/__init__.py").touch()
+        assert find_parent_package(multiple, {"": "src"}, tmp_path) is None
+
+
+class TestDiscoverPackagesAndPyModules:
+    """Make sure discovered values for ``packages`` and ``py_modules`` work
+    similarly to explicit configuration for the simple scenarios.
+    """
+
+    OPTIONS = {
+        # Different options according to the circumstance being tested
+        "explicit-src": {"package_dir": {"": "src"}, "packages": ["pkg"]},
+        "variation-lib": {
+            "package_dir": {"": "lib"},  # variation of the source-layout
+        },
+        "explicit-flat": {"packages": ["pkg"]},
+        "explicit-single_module": {"py_modules": ["pkg"]},
+        "explicit-namespace": {"packages": ["ns", "ns.pkg"]},
+        "automatic-src": {},
+        "automatic-flat": {},
+        "automatic-single_module": {},
+        "automatic-namespace": {},
+    }
+    FILES = {
+        "src": ["src/pkg/__init__.py", "src/pkg/main.py"],
+        "lib": ["lib/pkg/__init__.py", "lib/pkg/main.py"],
+        "flat": ["pkg/__init__.py", "pkg/main.py"],
+        "single_module": ["pkg.py"],
+        "namespace": ["ns/pkg/__init__.py"],
+    }
+
+    def _get_info(self, circumstance):
+        _, _, layout = circumstance.partition("-")
+        files = self.FILES[layout]
+        options = self.OPTIONS[circumstance]
+        return files, options
+
+    @pytest.mark.parametrize("circumstance", OPTIONS.keys())
+    def test_sdist_filelist(self, tmp_path, circumstance):
+        files, options = self._get_info(circumstance)
+        _populate_project_dir(tmp_path, files, options)
+
+        _, cmd = _run_sdist_programatically(tmp_path, options)
+
+        manifest = [f.replace(os.sep, "/") for f in cmd.filelist.files]
+        for file in files:
+            assert any(f.endswith(file) for f in manifest)
+
+    @pytest.mark.parametrize("circumstance", OPTIONS.keys())
+    def test_project(self, tmp_path, circumstance):
+        files, options = self._get_info(circumstance)
+        _populate_project_dir(tmp_path, files, options)
+
+        # Simulate a pre-existing `build` directory
+        (tmp_path / "build").mkdir()
+        (tmp_path / "build/lib").mkdir()
+        (tmp_path / "build/bdist.linux-x86_64").mkdir()
+        (tmp_path / "build/bdist.linux-x86_64/file.py").touch()
+        (tmp_path / "build/lib/__init__.py").touch()
+        (tmp_path / "build/lib/file.py").touch()
+        (tmp_path / "dist").mkdir()
+        (tmp_path / "dist/file.py").touch()
+
+        _run_build(tmp_path)
+
+        sdist_files = get_sdist_members(next(tmp_path.glob("dist/*.tar.gz")))
+        print("~~~~~ sdist_members ~~~~~")
+        print('\n'.join(sdist_files))
+        assert sdist_files >= set(files)
+
+        wheel_files = get_wheel_members(next(tmp_path.glob("dist/*.whl")))
+        print("~~~~~ wheel_members ~~~~~")
+        print('\n'.join(wheel_files))
+        orig_files = {f.replace("src/", "").replace("lib/", "") for f in files}
+        assert wheel_files >= orig_files
+
+        # Make sure build files are not included by mistake
+        for file in wheel_files:
+            assert "build" not in files
+            assert "dist" not in files
+
+    PURPOSEFULLY_EMPY = {
+        "setup.cfg": DALS(
+            """
+            [metadata]
+            name = myproj
+            version = 0.0.0
+
+            [options]
+            {param} =
+            """
+        ),
+        "setup.py": DALS(
+            """
+            __import__('setuptools').setup(
+                name="myproj",
+                version="0.0.0",
+                {param}=[]
+            )
+            """
+        ),
+        "pyproject.toml": DALS(
+            """
+            [build-system]
+            requires = []
+            build-backend = 'setuptools.build_meta'
+
+            [project]
+            name = "myproj"
+            version = "0.0.0"
+
+            [tool.setuptools]
+            {param} = []
+            """
+        ),
+        "template-pyproject.toml": DALS(
+            """
+            [build-system]
+            requires = []
+            build-backend = 'setuptools.build_meta'
+            """
+        ),
+    }
+
+    @pytest.mark.parametrize(
+        ("config_file", "param", "circumstance"),
+        product(
+            ["setup.cfg", "setup.py", "pyproject.toml"],
+            ["packages", "py_modules"],
+            FILES.keys(),
+        ),
+    )
+    def test_purposefully_empty(self, tmp_path, config_file, param, circumstance):
+        files = self.FILES[circumstance] + ["mod.py", "other.py", "src/pkg/__init__.py"]
+        _populate_project_dir(tmp_path, files, {})
+
+        if config_file == "pyproject.toml":
+            template_param = param.replace("_", "-")
+        else:
+            # Make sure build works with or without setup.cfg
+            pyproject = self.PURPOSEFULLY_EMPY["template-pyproject.toml"]
+            (tmp_path / "pyproject.toml").write_text(pyproject, encoding="utf-8")
+            template_param = param
+
+        config = self.PURPOSEFULLY_EMPY[config_file].format(param=template_param)
+        (tmp_path / config_file).write_text(config, encoding="utf-8")
+
+        dist = _get_dist(tmp_path, {})
+        # When either parameter package or py_modules is an empty list,
+        # then there should be no discovery
+        assert getattr(dist, param) == []
+        other = {"py_modules": "packages", "packages": "py_modules"}[param]
+        assert getattr(dist, other) is None
+
+    @pytest.mark.parametrize(
+        ("extra_files", "pkgs"),
+        [
+            (["venv/bin/simulate_venv"], {"pkg"}),
+            (["pkg-stubs/__init__.pyi"], {"pkg", "pkg-stubs"}),
+            (["other-stubs/__init__.pyi"], {"pkg", "other-stubs"}),
+            (
+                # Type stubs can also be namespaced
+                ["namespace-stubs/pkg/__init__.pyi"],
+                {"pkg", "namespace-stubs", "namespace-stubs.pkg"},
+            ),
+            (
+                # Just the top-level package can have `-stubs`, ignore nested ones
+                ["namespace-stubs/pkg-stubs/__init__.pyi"],
+                {"pkg", "namespace-stubs"},
+            ),
+            (["_hidden/file.py"], {"pkg"}),
+            (["news/finalize.py"], {"pkg"}),
+        ],
+    )
+    def test_flat_layout_with_extra_files(self, tmp_path, extra_files, pkgs):
+        files = self.FILES["flat"] + extra_files
+        _populate_project_dir(tmp_path, files, {})
+        dist = _get_dist(tmp_path, {})
+        assert set(dist.packages) == pkgs
+
+    @pytest.mark.parametrize(
+        "extra_files",
+        [
+            ["other/__init__.py"],
+            ["other/finalize.py"],
+        ],
+    )
+    def test_flat_layout_with_dangerous_extra_files(self, tmp_path, extra_files):
+        files = self.FILES["flat"] + extra_files
+        _populate_project_dir(tmp_path, files, {})
+        with pytest.raises(PackageDiscoveryError, match="multiple (packages|modules)"):
+            _get_dist(tmp_path, {})
+
+    def test_flat_layout_with_single_module(self, tmp_path):
+        files = self.FILES["single_module"] + ["invalid-module-name.py"]
+        _populate_project_dir(tmp_path, files, {})
+        dist = _get_dist(tmp_path, {})
+        assert set(dist.py_modules) == {"pkg"}
+
+    def test_flat_layout_with_multiple_modules(self, tmp_path):
+        files = self.FILES["single_module"] + ["valid_module_name.py"]
+        _populate_project_dir(tmp_path, files, {})
+        with pytest.raises(PackageDiscoveryError, match="multiple (packages|modules)"):
+            _get_dist(tmp_path, {})
+
+    def test_py_modules_when_wheel_dir_is_cwd(self, tmp_path):
+        """Regression for issue 3692"""
+        from setuptools import build_meta
+
+        pyproject = '[project]\nname = "test"\nversion = "1"'
+        (tmp_path / "pyproject.toml").write_text(DALS(pyproject), encoding="utf-8")
+        (tmp_path / "foo.py").touch()
+        with jaraco.path.DirectoryStack().context(tmp_path):
+            build_meta.build_wheel(".")
+        # Ensure py_modules are found
+        wheel_files = get_wheel_members(next(tmp_path.glob("*.whl")))
+        assert "foo.py" in wheel_files
+
+
+class TestNoConfig:
+    DEFAULT_VERSION = "0.0.0"  # Default version given by setuptools
+
+    EXAMPLES = {
+        "pkg1": ["src/pkg1.py"],
+        "pkg2": ["src/pkg2/__init__.py"],
+        "pkg3": ["src/pkg3/__init__.py", "src/pkg3-stubs/__init__.py"],
+        "pkg4": ["pkg4/__init__.py", "pkg4-stubs/__init__.py"],
+        "ns.nested.pkg1": ["src/ns/nested/pkg1/__init__.py"],
+        "ns.nested.pkg2": ["ns/nested/pkg2/__init__.py"],
+    }
+
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_discover_name(self, tmp_path, example):
+        _populate_project_dir(tmp_path, self.EXAMPLES[example], {})
+        dist = _get_dist(tmp_path, {})
+        assert dist.get_name() == example
+
+    def test_build_with_discovered_name(self, tmp_path):
+        files = ["src/ns/nested/pkg/__init__.py"]
+        _populate_project_dir(tmp_path, files, {})
+        _run_build(tmp_path, "--sdist")
+        # Expected distribution file
+        dist_file = tmp_path / f"dist/ns_nested_pkg-{self.DEFAULT_VERSION}.tar.gz"
+        assert dist_file.is_file()
+
+
+class TestWithAttrDirective:
+    @pytest.mark.parametrize(
+        ("folder", "opts"),
+        [
+            ("src", {}),
+            ("lib", {"packages": "find:", "packages.find": {"where": "lib"}}),
+        ],
+    )
+    def test_setupcfg_metadata(self, tmp_path, folder, opts):
+        files = [f"{folder}/pkg/__init__.py", "setup.cfg"]
+        _populate_project_dir(tmp_path, files, opts)
+
+        config = (tmp_path / "setup.cfg").read_text(encoding="utf-8")
+        overwrite = {
+            folder: {"pkg": {"__init__.py": "version = 42"}},
+            "setup.cfg": "[metadata]\nversion = attr: pkg.version\n" + config,
+        }
+        jaraco.path.build(overwrite, prefix=tmp_path)
+
+        dist = _get_dist(tmp_path, {})
+        assert dist.get_name() == "pkg"
+        assert dist.get_version() == "42"
+        assert dist.package_dir
+        package_path = find_package_path("pkg", dist.package_dir, tmp_path)
+        assert os.path.exists(package_path)
+        assert folder in Path(package_path).parts()
+
+        _run_build(tmp_path, "--sdist")
+        dist_file = tmp_path / "dist/pkg-42.tar.gz"
+        assert dist_file.is_file()
+
+    def test_pyproject_metadata(self, tmp_path):
+        _populate_project_dir(tmp_path, ["src/pkg/__init__.py"], {})
+
+        overwrite = {
+            "src": {"pkg": {"__init__.py": "version = 42"}},
+            "pyproject.toml": (
+                "[project]\nname = 'pkg'\ndynamic = ['version']\n"
+                "[tool.setuptools.dynamic]\nversion = {attr = 'pkg.version'}\n"
+            ),
+        }
+        jaraco.path.build(overwrite, prefix=tmp_path)
+
+        dist = _get_dist(tmp_path, {})
+        assert dist.get_version() == "42"
+        assert dist.package_dir == {"": "src"}
+
+
+class TestWithCExtension:
+    def _simulate_package_with_extension(self, tmp_path):
+        # This example is based on: https://github.com/nucleic/kiwi/tree/1.4.0
+        files = [
+            "benchmarks/file.py",
+            "docs/Makefile",
+            "docs/requirements.txt",
+            "docs/source/conf.py",
+            "proj/header.h",
+            "proj/file.py",
+            "py/proj.cpp",
+            "py/other.cpp",
+            "py/file.py",
+            "py/py.typed",
+            "py/tests/test_proj.py",
+            "README.rst",
+        ]
+        _populate_project_dir(tmp_path, files, {})
+
+        setup_script = """
+            from setuptools import Extension, setup
+
+            ext_modules = [
+                Extension(
+                    "proj",
+                    ["py/proj.cpp", "py/other.cpp"],
+                    include_dirs=["."],
+                    language="c++",
+                ),
+            ]
+            setup(ext_modules=ext_modules)
+        """
+        (tmp_path / "setup.py").write_text(DALS(setup_script), encoding="utf-8")
+
+    def test_skip_discovery_with_setupcfg_metadata(self, tmp_path):
+        """Ensure that auto-discovery is not triggered when the project is based on
+        C-extensions only, for backward compatibility.
+        """
+        self._simulate_package_with_extension(tmp_path)
+
+        pyproject = """
+            [build-system]
+            requires = []
+            build-backend = 'setuptools.build_meta'
+        """
+        (tmp_path / "pyproject.toml").write_text(DALS(pyproject), encoding="utf-8")
+
+        setupcfg = """
+            [metadata]
+            name = proj
+            version = 42
+        """
+        (tmp_path / "setup.cfg").write_text(DALS(setupcfg), encoding="utf-8")
+
+        dist = _get_dist(tmp_path, {})
+        assert dist.get_name() == "proj"
+        assert dist.get_version() == "42"
+        assert dist.py_modules is None
+        assert dist.packages is None
+        assert len(dist.ext_modules) == 1
+        assert dist.ext_modules[0].name == "proj"
+
+    def test_dont_skip_discovery_with_pyproject_metadata(self, tmp_path):
+        """When opting-in to pyproject.toml metadata, auto-discovery will be active if
+        the package lists C-extensions, but does not configure py-modules or packages.
+
+        This way we ensure users with complex package layouts that would lead to the
+        discovery of multiple top-level modules/packages see errors and are forced to
+        explicitly set ``packages`` or ``py-modules``.
+        """
+        self._simulate_package_with_extension(tmp_path)
+
+        pyproject = """
+            [project]
+            name = 'proj'
+            version = '42'
+        """
+        (tmp_path / "pyproject.toml").write_text(DALS(pyproject), encoding="utf-8")
+        with pytest.raises(PackageDiscoveryError, match="multiple (packages|modules)"):
+            _get_dist(tmp_path, {})
+
+
+class TestWithPackageData:
+    def _simulate_package_with_data_files(self, tmp_path, src_root):
+        files = [
+            f"{src_root}/proj/__init__.py",
+            f"{src_root}/proj/file1.txt",
+            f"{src_root}/proj/nested/file2.txt",
+        ]
+        _populate_project_dir(tmp_path, files, {})
+
+        manifest = """
+            global-include *.py *.txt
+        """
+        (tmp_path / "MANIFEST.in").write_text(DALS(manifest), encoding="utf-8")
+
+    EXAMPLE_SETUPCFG = """
+    [metadata]
+    name = proj
+    version = 42
+
+    [options]
+    include_package_data = True
+    """
+    EXAMPLE_PYPROJECT = """
+    [project]
+    name = "proj"
+    version = "42"
+    """
+
+    PYPROJECT_PACKAGE_DIR = """
+    [tool.setuptools]
+    package-dir = {"" = "src"}
+    """
+
+    @pytest.mark.parametrize(
+        ("src_root", "files"),
+        [
+            (".", {"setup.cfg": DALS(EXAMPLE_SETUPCFG)}),
+            (".", {"pyproject.toml": DALS(EXAMPLE_PYPROJECT)}),
+            ("src", {"setup.cfg": DALS(EXAMPLE_SETUPCFG)}),
+            ("src", {"pyproject.toml": DALS(EXAMPLE_PYPROJECT)}),
+            (
+                "src",
+                {
+                    "setup.cfg": DALS(EXAMPLE_SETUPCFG)
+                    + DALS(
+                        """
+                        packages = find:
+                        package_dir =
+                            =src
+
+                        [options.packages.find]
+                        where = src
+                        """
+                    )
+                },
+            ),
+            (
+                "src",
+                {
+                    "pyproject.toml": DALS(EXAMPLE_PYPROJECT)
+                    + DALS(
+                        """
+                        [tool.setuptools]
+                        package-dir = {"" = "src"}
+                        """
+                    )
+                },
+            ),
+        ],
+    )
+    def test_include_package_data(self, tmp_path, src_root, files):
+        """
+        Make sure auto-discovery does not affect package include_package_data.
+        See issue #3196.
+        """
+        jaraco.path.build(files, prefix=str(tmp_path))
+        self._simulate_package_with_data_files(tmp_path, src_root)
+
+        expected = {
+            os.path.normpath(f"{src_root}/proj/file1.txt").replace(os.sep, "/"),
+            os.path.normpath(f"{src_root}/proj/nested/file2.txt").replace(os.sep, "/"),
+        }
+
+        _run_build(tmp_path)
+
+        sdist_files = get_sdist_members(next(tmp_path.glob("dist/*.tar.gz")))
+        print("~~~~~ sdist_members ~~~~~")
+        print('\n'.join(sdist_files))
+        assert sdist_files >= expected
+
+        wheel_files = get_wheel_members(next(tmp_path.glob("dist/*.whl")))
+        print("~~~~~ wheel_members ~~~~~")
+        print('\n'.join(wheel_files))
+        orig_files = {f.replace("src/", "").replace("lib/", "") for f in expected}
+        assert wheel_files >= orig_files
+
+
+def test_compatible_with_numpy_configuration(tmp_path):
+    files = [
+        "dir1/__init__.py",
+        "dir2/__init__.py",
+        "file.py",
+    ]
+    _populate_project_dir(tmp_path, files, {})
+    dist = Distribution({})
+    dist.configuration = object()
+    dist.set_defaults()
+    assert dist.py_modules is None
+    assert dist.packages is None
+
+
+def test_name_discovery_doesnt_break_cli(tmpdir_cwd):
+    jaraco.path.build({"pkg.py": ""})
+    dist = Distribution({})
+    dist.script_args = ["--name"]
+    dist.set_defaults()
+    dist.parse_command_line()  # <-- no exception should be raised here.
+    assert dist.get_name() == "pkg"
+
+
+def test_preserve_explicit_name_with_dynamic_version(tmpdir_cwd, monkeypatch):
+    """According to #3545 it seems that ``name`` discovery is running,
+    even when the project already explicitly sets it.
+    This seems to be related to parsing of dynamic versions (via ``attr`` directive),
+    which requires the auto-discovery of ``package_dir``.
+    """
+    files = {
+        "src": {
+            "pkg": {"__init__.py": "__version__ = 42\n"},
+        },
+        "pyproject.toml": DALS(
+            """
+            [project]
+            name = "myproj"  # purposefully different from package name
+            dynamic = ["version"]
+            [tool.setuptools.dynamic]
+            version = {"attr" = "pkg.__version__"}
+            """
+        ),
+    }
+    jaraco.path.build(files)
+    dist = Distribution({})
+    orig_analyse_name = dist.set_defaults.analyse_name
+
+    def spy_analyse_name():
+        # We can check if name discovery was triggered by ensuring the original
+        # name remains instead of the package name.
+        orig_analyse_name()
+        assert dist.get_name() == "myproj"
+
+    monkeypatch.setattr(dist.set_defaults, "analyse_name", spy_analyse_name)
+    dist.parse_config_files()
+    assert dist.get_version() == "42"
+    assert set(dist.packages) == {"pkg"}
+
+
+def _populate_project_dir(root, files, options):
+    # NOTE: Currently pypa/build will refuse to build the project if no
+    # `pyproject.toml` or `setup.py` is found. So it is impossible to do
+    # completely "config-less" projects.
+    basic = {
+        "setup.py": "import setuptools\nsetuptools.setup()",
+        "README.md": "# Example Package",
+        "LICENSE": "Copyright (c) 2018",
+    }
+    jaraco.path.build(basic, prefix=root)
+    _write_setupcfg(root, options)
+    paths = (root / f for f in files)
+    for path in paths:
+        path.parent.mkdir(exist_ok=True, parents=True)
+        path.touch()
+
+
+def _write_setupcfg(root, options):
+    if not options:
+        print("~~~~~ **NO** setup.cfg ~~~~~")
+        return
+    setupcfg = ConfigParser()
+    setupcfg.add_section("options")
+    for key, value in options.items():
+        if key == "packages.find":
+            setupcfg.add_section(f"options.{key}")
+            setupcfg[f"options.{key}"].update(value)
+        elif isinstance(value, list):
+            setupcfg["options"][key] = ", ".join(value)
+        elif isinstance(value, dict):
+            str_value = "\n".join(f"\t{k} = {v}" for k, v in value.items())
+            setupcfg["options"][key] = "\n" + str_value
+        else:
+            setupcfg["options"][key] = str(value)
+    with open(root / "setup.cfg", "w", encoding="utf-8") as f:
+        setupcfg.write(f)
+    print("~~~~~ setup.cfg ~~~~~")
+    print((root / "setup.cfg").read_text(encoding="utf-8"))
+
+
+def _run_build(path, *flags):
+    cmd = [sys.executable, "-m", "build", "--no-isolation", *flags, str(path)]
+    return run(cmd, env={'DISTUTILS_DEBUG': ''})
+
+
+def _get_dist(dist_path, attrs):
+    root = "/".join(os.path.split(dist_path))  # POSIX-style
+
+    script = dist_path / 'setup.py'
+    if script.exists():
+        with Path(dist_path):
+            dist = cast(
+                Distribution,
+                distutils.core.run_setup("setup.py", {}, stop_after="init"),
+            )
+    else:
+        dist = Distribution(attrs)
+
+    dist.src_root = root
+    dist.script_name = "setup.py"
+    with Path(dist_path):
+        dist.parse_config_files()
+
+    dist.set_defaults()
+    return dist
+
+
+def _run_sdist_programatically(dist_path, attrs):
+    dist = _get_dist(dist_path, attrs)
+    cmd = sdist(dist)
+    cmd.ensure_finalized()
+    assert cmd.distribution.packages or cmd.distribution.py_modules
+
+    with quiet(), Path(dist_path):
+        cmd.run()
+
+    return dist, cmd
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_core_metadata.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_core_metadata.py
new file mode 100644
index 00000000..b1edb79b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_core_metadata.py
@@ -0,0 +1,577 @@
+from __future__ import annotations
+
+import functools
+import importlib
+import io
+from email import message_from_string
+from email.generator import Generator
+from email.message import EmailMessage, Message
+from email.parser import Parser
+from email.policy import EmailPolicy
+from inspect import cleandoc
+from pathlib import Path
+from unittest.mock import Mock
+
+import pytest
+from packaging.metadata import Metadata
+from packaging.requirements import Requirement
+
+from setuptools import _reqs, sic
+from setuptools._core_metadata import rfc822_escape, rfc822_unescape
+from setuptools.command.egg_info import egg_info, write_requirements
+from setuptools.config import expand, setupcfg
+from setuptools.dist import Distribution
+
+from .config.downloads import retrieve_file, urls_from_file
+
+EXAMPLE_BASE_INFO = dict(
+    name="package",
+    version="0.0.1",
+    author="Foo Bar",
+    author_email="foo@bar.net",
+    long_description="Long\ndescription",
+    description="Short description",
+    keywords=["one", "two"],
+)
+
+
+@pytest.mark.parametrize(
+    ("content", "result"),
+    (
+        pytest.param(
+            "Just a single line",
+            None,
+            id="single_line",
+        ),
+        pytest.param(
+            "Multiline\nText\nwithout\nextra indents\n",
+            None,
+            id="multiline",
+        ),
+        pytest.param(
+            "Multiline\n    With\n\nadditional\n  indentation",
+            None,
+            id="multiline_with_indentation",
+        ),
+        pytest.param(
+            "  Leading whitespace",
+            "Leading whitespace",
+            id="remove_leading_whitespace",
+        ),
+        pytest.param(
+            "  Leading whitespace\nIn\n    Multiline comment",
+            "Leading whitespace\nIn\n    Multiline comment",
+            id="remove_leading_whitespace_multiline",
+        ),
+    ),
+)
+def test_rfc822_unescape(content, result):
+    assert (result or content) == rfc822_unescape(rfc822_escape(content))
+
+
+def __read_test_cases():
+    base = EXAMPLE_BASE_INFO
+
+    params = functools.partial(dict, base)
+
+    return [
+        ('Metadata version 1.0', params()),
+        (
+            'Metadata Version 1.0: Short long description',
+            params(
+                long_description='Short long description',
+            ),
+        ),
+        (
+            'Metadata version 1.1: Classifiers',
+            params(
+                classifiers=[
+                    'Programming Language :: Python :: 3',
+                    'Programming Language :: Python :: 3.7',
+                    'License :: OSI Approved :: MIT License',
+                ],
+            ),
+        ),
+        (
+            'Metadata version 1.1: Download URL',
+            params(
+                download_url='https://example.com',
+            ),
+        ),
+        (
+            'Metadata Version 1.2: Requires-Python',
+            params(
+                python_requires='>=3.7',
+            ),
+        ),
+        pytest.param(
+            'Metadata Version 1.2: Project-Url',
+            params(project_urls=dict(Foo='https://example.bar')),
+            marks=pytest.mark.xfail(
+                reason="Issue #1578: project_urls not read",
+            ),
+        ),
+        (
+            'Metadata Version 2.1: Long Description Content Type',
+            params(
+                long_description_content_type='text/x-rst; charset=UTF-8',
+            ),
+        ),
+        (
+            'License',
+            params(
+                license='MIT',
+            ),
+        ),
+        (
+            'License multiline',
+            params(
+                license='This is a long license \nover multiple lines',
+            ),
+        ),
+        pytest.param(
+            'Metadata Version 2.1: Provides Extra',
+            params(provides_extras=['foo', 'bar']),
+            marks=pytest.mark.xfail(reason="provides_extras not read"),
+        ),
+        (
+            'Missing author',
+            dict(
+                name='foo',
+                version='1.0.0',
+                author_email='snorri@sturluson.name',
+            ),
+        ),
+        (
+            'Missing author e-mail',
+            dict(
+                name='foo',
+                version='1.0.0',
+                author='Snorri Sturluson',
+            ),
+        ),
+        (
+            'Missing author and e-mail',
+            dict(
+                name='foo',
+                version='1.0.0',
+            ),
+        ),
+        (
+            'Bypass normalized version',
+            dict(
+                name='foo',
+                version=sic('1.0.0a'),
+            ),
+        ),
+    ]
+
+
+@pytest.mark.parametrize(("name", "attrs"), __read_test_cases())
+def test_read_metadata(name, attrs):
+    dist = Distribution(attrs)
+    metadata_out = dist.metadata
+    dist_class = metadata_out.__class__
+
+    # Write to PKG_INFO and then load into a new metadata object
+    PKG_INFO = io.StringIO()
+
+    metadata_out.write_pkg_file(PKG_INFO)
+    PKG_INFO.seek(0)
+    pkg_info = PKG_INFO.read()
+    assert _valid_metadata(pkg_info)
+
+    PKG_INFO.seek(0)
+    metadata_in = dist_class()
+    metadata_in.read_pkg_file(PKG_INFO)
+
+    tested_attrs = [
+        ('name', dist_class.get_name),
+        ('version', dist_class.get_version),
+        ('author', dist_class.get_contact),
+        ('author_email', dist_class.get_contact_email),
+        ('metadata_version', dist_class.get_metadata_version),
+        ('provides', dist_class.get_provides),
+        ('description', dist_class.get_description),
+        ('long_description', dist_class.get_long_description),
+        ('download_url', dist_class.get_download_url),
+        ('keywords', dist_class.get_keywords),
+        ('platforms', dist_class.get_platforms),
+        ('obsoletes', dist_class.get_obsoletes),
+        ('requires', dist_class.get_requires),
+        ('classifiers', dist_class.get_classifiers),
+        ('project_urls', lambda s: getattr(s, 'project_urls', {})),
+        ('provides_extras', lambda s: getattr(s, 'provides_extras', {})),
+    ]
+
+    for attr, getter in tested_attrs:
+        assert getter(metadata_in) == getter(metadata_out)
+
+
+def __maintainer_test_cases():
+    attrs = {"name": "package", "version": "1.0", "description": "xxx"}
+
+    def merge_dicts(d1, d2):
+        d1 = d1.copy()
+        d1.update(d2)
+
+        return d1
+
+    return [
+        ('No author, no maintainer', attrs.copy()),
+        (
+            'Author (no e-mail), no maintainer',
+            merge_dicts(attrs, {'author': 'Author Name'}),
+        ),
+        (
+            'Author (e-mail), no maintainer',
+            merge_dicts(
+                attrs, {'author': 'Author Name', 'author_email': 'author@name.com'}
+            ),
+        ),
+        (
+            'No author, maintainer (no e-mail)',
+            merge_dicts(attrs, {'maintainer': 'Maintainer Name'}),
+        ),
+        (
+            'No author, maintainer (e-mail)',
+            merge_dicts(
+                attrs,
+                {
+                    'maintainer': 'Maintainer Name',
+                    'maintainer_email': 'maintainer@name.com',
+                },
+            ),
+        ),
+        (
+            'Author (no e-mail), Maintainer (no-email)',
+            merge_dicts(
+                attrs, {'author': 'Author Name', 'maintainer': 'Maintainer Name'}
+            ),
+        ),
+        (
+            'Author (e-mail), Maintainer (e-mail)',
+            merge_dicts(
+                attrs,
+                {
+                    'author': 'Author Name',
+                    'author_email': 'author@name.com',
+                    'maintainer': 'Maintainer Name',
+                    'maintainer_email': 'maintainer@name.com',
+                },
+            ),
+        ),
+        (
+            'No author (e-mail), no maintainer (e-mail)',
+            merge_dicts(
+                attrs,
+                {
+                    'author_email': 'author@name.com',
+                    'maintainer_email': 'maintainer@name.com',
+                },
+            ),
+        ),
+        ('Author unicode', merge_dicts(attrs, {'author': '鉄沢寛'})),
+        ('Maintainer unicode', merge_dicts(attrs, {'maintainer': 'Jan Łukasiewicz'})),
+    ]
+
+
+@pytest.mark.parametrize(("name", "attrs"), __maintainer_test_cases())
+def test_maintainer_author(name, attrs, tmpdir):
+    tested_keys = {
+        'author': 'Author',
+        'author_email': 'Author-email',
+        'maintainer': 'Maintainer',
+        'maintainer_email': 'Maintainer-email',
+    }
+
+    # Generate a PKG-INFO file
+    dist = Distribution(attrs)
+    fn = tmpdir.mkdir('pkg_info')
+    fn_s = str(fn)
+
+    dist.metadata.write_pkg_info(fn_s)
+
+    with open(str(fn.join('PKG-INFO')), 'r', encoding='utf-8') as f:
+        pkg_info = f.read()
+
+    assert _valid_metadata(pkg_info)
+
+    # Drop blank lines and strip lines from default description
+    raw_pkg_lines = pkg_info.splitlines()
+    pkg_lines = list(filter(None, raw_pkg_lines[:-2]))
+
+    pkg_lines_set = set(pkg_lines)
+
+    # Duplicate lines should not be generated
+    assert len(pkg_lines) == len(pkg_lines_set)
+
+    for fkey, dkey in tested_keys.items():
+        val = attrs.get(dkey, None)
+        if val is None:
+            for line in pkg_lines:
+                assert not line.startswith(fkey + ':')
+        else:
+            line = f'{fkey}: {val}'
+            assert line in pkg_lines_set
+
+
+class TestParityWithMetadataFromPyPaWheel:
+    def base_example(self):
+        attrs = dict(
+            **EXAMPLE_BASE_INFO,
+            # Example with complex requirement definition
+            python_requires=">=3.8",
+            install_requires="""
+            packaging==23.2
+            more-itertools==8.8.0; extra == "other"
+            jaraco.text==3.7.0
+            importlib-resources==5.10.2; python_version<"3.8"
+            importlib-metadata==6.0.0 ; python_version<"3.8"
+            colorama>=0.4.4; sys_platform == "win32"
+            """,
+            extras_require={
+                "testing": """
+                    pytest >= 6
+                    pytest-checkdocs >= 2.4
+                    tomli ; \\
+                            # Using stdlib when possible
+                            python_version < "3.11"
+                    ini2toml[lite]>=0.9
+                    """,
+                "other": [],
+            },
+        )
+        # Generate a PKG-INFO file using setuptools
+        return Distribution(attrs)
+
+    def test_requires_dist(self, tmp_path):
+        dist = self.base_example()
+        pkg_info = _get_pkginfo(dist)
+        assert _valid_metadata(pkg_info)
+
+        # Ensure Requires-Dist is present
+        expected = [
+            'Metadata-Version:',
+            'Requires-Python: >=3.8',
+            'Provides-Extra: other',
+            'Provides-Extra: testing',
+            'Requires-Dist: tomli; python_version < "3.11" and extra == "testing"',
+            'Requires-Dist: more-itertools==8.8.0; extra == "other"',
+            'Requires-Dist: ini2toml[lite]>=0.9; extra == "testing"',
+        ]
+        for line in expected:
+            assert line in pkg_info
+
+    HERE = Path(__file__).parent
+    EXAMPLES_FILE = HERE / "config/setupcfg_examples.txt"
+
+    @pytest.fixture(params=[None, *urls_from_file(EXAMPLES_FILE)])
+    def dist(self, request, monkeypatch, tmp_path):
+        """Example of distribution with arbitrary configuration"""
+        monkeypatch.chdir(tmp_path)
+        monkeypatch.setattr(expand, "read_attr", Mock(return_value="0.42"))
+        monkeypatch.setattr(expand, "read_files", Mock(return_value="hello world"))
+        if request.param is None:
+            yield self.base_example()
+        else:
+            # Real-world usage
+            config = retrieve_file(request.param)
+            yield setupcfg.apply_configuration(Distribution({}), config)
+
+    @pytest.mark.uses_network
+    def test_equivalent_output(self, tmp_path, dist):
+        """Ensure output from setuptools is equivalent to the one from `pypa/wheel`"""
+        # Generate a METADATA file using pypa/wheel for comparison
+        wheel_metadata = importlib.import_module("wheel.metadata")
+        pkginfo_to_metadata = getattr(wheel_metadata, "pkginfo_to_metadata", None)
+
+        if pkginfo_to_metadata is None:  # pragma: nocover
+            pytest.xfail(
+                "wheel.metadata.pkginfo_to_metadata is undefined, "
+                "(this is likely to be caused by API changes in pypa/wheel"
+            )
+
+        # Generate an simplified "egg-info" dir for pypa/wheel to convert
+        pkg_info = _get_pkginfo(dist)
+        egg_info_dir = tmp_path / "pkg.egg-info"
+        egg_info_dir.mkdir(parents=True)
+        (egg_info_dir / "PKG-INFO").write_text(pkg_info, encoding="utf-8")
+        write_requirements(egg_info(dist), egg_info_dir, egg_info_dir / "requires.txt")
+
+        # Get pypa/wheel generated METADATA but normalize requirements formatting
+        metadata_msg = pkginfo_to_metadata(egg_info_dir, egg_info_dir / "PKG-INFO")
+        metadata_str = _normalize_metadata(metadata_msg)
+        pkg_info_msg = message_from_string(pkg_info)
+        pkg_info_str = _normalize_metadata(pkg_info_msg)
+
+        # Compare setuptools PKG-INFO x pypa/wheel METADATA
+        assert metadata_str == pkg_info_str
+
+        # Make sure it parses/serializes well in pypa/wheel
+        _assert_roundtrip_message(pkg_info)
+
+
+class TestPEP643:
+    STATIC_CONFIG = {
+        "setup.cfg": cleandoc(
+            """
+            [metadata]
+            name = package
+            version = 0.0.1
+            author = Foo Bar
+            author_email = foo@bar.net
+            long_description = Long
+                               description
+            description = Short description
+            keywords = one, two
+            platforms = abcd
+            [options]
+            install_requires = requests
+            """
+        ),
+        "pyproject.toml": cleandoc(
+            """
+            [project]
+            name = "package"
+            version = "0.0.1"
+            authors = [
+              {name = "Foo Bar", email = "foo@bar.net"}
+            ]
+            description = "Short description"
+            readme = {text = "Long\\ndescription", content-type = "text/plain"}
+            keywords = ["one", "two"]
+            dependencies = ["requests"]
+            [tool.setuptools]
+            provides = ["abcd"]
+            obsoletes = ["abcd"]
+            """
+        ),
+    }
+
+    @pytest.mark.parametrize("file", STATIC_CONFIG.keys())
+    def test_static_config_has_no_dynamic(self, file, tmpdir_cwd):
+        Path(file).write_text(self.STATIC_CONFIG[file], encoding="utf-8")
+        metadata = _get_metadata()
+        assert metadata.get_all("Dynamic") is None
+        assert metadata.get_all("dynamic") is None
+
+    @pytest.mark.parametrize("file", STATIC_CONFIG.keys())
+    @pytest.mark.parametrize(
+        "fields",
+        [
+            # Single dynamic field
+            {"requires-python": ("python_requires", ">=3.12")},
+            {"author-email": ("author_email", "snoopy@peanuts.com")},
+            {"keywords": ("keywords", ["hello", "world"])},
+            {"platform": ("platforms", ["abcd"])},
+            # Multiple dynamic fields
+            {
+                "summary": ("description", "hello world"),
+                "description": ("long_description", "bla bla bla bla"),
+                "requires-dist": ("install_requires", ["hello-world"]),
+            },
+        ],
+    )
+    def test_modified_fields_marked_as_dynamic(self, file, fields, tmpdir_cwd):
+        # We start with a static config
+        Path(file).write_text(self.STATIC_CONFIG[file], encoding="utf-8")
+        dist = _makedist()
+
+        # ... but then we simulate the effects of a plugin modifying the distribution
+        for attr, value in fields.values():
+            # `dist` and `dist.metadata` are complicated...
+            # Some attributes work when set on `dist`, others on `dist.metadata`...
+            # Here we set in both just in case (this also avoids calling `_finalize_*`)
+            setattr(dist, attr, value)
+            setattr(dist.metadata, attr, value)
+
+        # Then we should be able to list the modified fields as Dynamic
+        metadata = _get_metadata(dist)
+        assert set(metadata.get_all("Dynamic")) == set(fields)
+
+
+def _makedist(**attrs):
+    dist = Distribution(attrs)
+    dist.parse_config_files()
+    return dist
+
+
+def _assert_roundtrip_message(metadata: str) -> None:
+    """Emulate the way wheel.bdist_wheel parses and regenerates the message,
+    then ensures the metadata generated by setuptools is compatible.
+    """
+    with io.StringIO(metadata) as buffer:
+        msg = Parser(EmailMessage).parse(buffer)
+
+    serialization_policy = EmailPolicy(
+        utf8=True,
+        mangle_from_=False,
+        max_line_length=0,
+    )
+    with io.BytesIO() as buffer:
+        out = io.TextIOWrapper(buffer, encoding="utf-8")
+        Generator(out, policy=serialization_policy).flatten(msg)
+        out.flush()
+        regenerated = buffer.getvalue()
+
+    raw_metadata = bytes(metadata, "utf-8")
+    # Normalise newlines to avoid test errors on Windows:
+    raw_metadata = b"\n".join(raw_metadata.splitlines())
+    regenerated = b"\n".join(regenerated.splitlines())
+    assert regenerated == raw_metadata
+
+
+def _normalize_metadata(msg: Message) -> str:
+    """Allow equivalent metadata to be compared directly"""
+    # The main challenge regards the requirements and extras.
+    # Both setuptools and wheel already apply some level of normalization
+    # but they differ regarding which character is chosen, according to the
+    # following spec it should be "-":
+    # https://packaging.python.org/en/latest/specifications/name-normalization/
+
+    # Related issues:
+    # https://github.com/pypa/packaging/issues/845
+    # https://github.com/pypa/packaging/issues/644#issuecomment-2429813968
+
+    extras = {x.replace("_", "-"): x for x in msg.get_all("Provides-Extra", [])}
+    reqs = [
+        _normalize_req(req, extras)
+        for req in _reqs.parse(msg.get_all("Requires-Dist", []))
+    ]
+    del msg["Requires-Dist"]
+    del msg["Provides-Extra"]
+
+    # Ensure consistent ord
+    for req in sorted(reqs):
+        msg["Requires-Dist"] = req
+    for extra in sorted(extras):
+        msg["Provides-Extra"] = extra
+
+    # TODO: Handle lack of PEP 643 implementation in pypa/wheel?
+    del msg["Metadata-Version"]
+
+    return msg.as_string()
+
+
+def _normalize_req(req: Requirement, extras: dict[str, str]) -> str:
+    """Allow equivalent requirement objects to be compared directly"""
+    as_str = str(req).replace(req.name, req.name.replace("_", "-"))
+    for norm, orig in extras.items():
+        as_str = as_str.replace(orig, norm)
+    return as_str
+
+
+def _get_pkginfo(dist: Distribution):
+    with io.StringIO() as fp:
+        dist.metadata.write_pkg_file(fp)
+        return fp.getvalue()
+
+
+def _get_metadata(dist: Distribution | None = None):
+    return message_from_string(_get_pkginfo(dist or _makedist()))
+
+
+def _valid_metadata(text: str) -> bool:
+    metadata = Metadata.from_email(text, validate=True)  # can raise exceptions
+    return metadata is not None
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_depends.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_depends.py
new file mode 100644
index 00000000..1714c041
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_depends.py
@@ -0,0 +1,15 @@
+import sys
+
+from setuptools import depends
+
+
+class TestGetModuleConstant:
+    def test_basic(self):
+        """
+        Invoke get_module_constant on a module in
+        the test package.
+        """
+        mod_name = 'setuptools.tests.mod_with_constant'
+        val = depends.get_module_constant(mod_name, 'value')
+        assert val == 'three, sir!'
+        assert 'setuptools.tests.mod_with_constant' not in sys.modules
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_develop.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_develop.py
new file mode 100644
index 00000000..929fa9c2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_develop.py
@@ -0,0 +1,175 @@
+"""develop tests"""
+
+import os
+import pathlib
+import platform
+import subprocess
+import sys
+
+import pytest
+
+from setuptools._path import paths_on_pythonpath
+from setuptools.command.develop import develop
+from setuptools.dist import Distribution
+
+from . import contexts, namespaces
+
+SETUP_PY = """\
+from setuptools import setup
+
+setup(name='foo',
+    packages=['foo'],
+)
+"""
+
+INIT_PY = """print "foo"
+"""
+
+
+@pytest.fixture
+def temp_user(monkeypatch):
+    with contexts.tempdir() as user_base:
+        with contexts.tempdir() as user_site:
+            monkeypatch.setattr('site.USER_BASE', user_base)
+            monkeypatch.setattr('site.USER_SITE', user_site)
+            yield
+
+
+@pytest.fixture
+def test_env(tmpdir, temp_user):
+    target = tmpdir
+    foo = target.mkdir('foo')
+    setup = target / 'setup.py'
+    if setup.isfile():
+        raise ValueError(dir(target))
+    with setup.open('w') as f:
+        f.write(SETUP_PY)
+    init = foo / '__init__.py'
+    with init.open('w') as f:
+        f.write(INIT_PY)
+    with target.as_cwd():
+        yield target
+
+
+class TestDevelop:
+    in_virtualenv = hasattr(sys, 'real_prefix')
+    in_venv = hasattr(sys, 'base_prefix') and sys.base_prefix != sys.prefix
+
+    def test_console_scripts(self, tmpdir):
+        """
+        Test that console scripts are installed and that they reference
+        only the project by name and not the current version.
+        """
+        pytest.skip(
+            "TODO: needs a fixture to cause 'develop' "
+            "to be invoked without mutating environment."
+        )
+        settings = dict(
+            name='foo',
+            packages=['foo'],
+            version='0.0',
+            entry_points={
+                'console_scripts': [
+                    'foocmd = foo:foo',
+                ],
+            },
+        )
+        dist = Distribution(settings)
+        dist.script_name = 'setup.py'
+        cmd = develop(dist)
+        cmd.ensure_finalized()
+        cmd.install_dir = tmpdir
+        cmd.run()
+        # assert '0.0' not in foocmd_text
+
+    @pytest.mark.xfail(reason="legacy behavior retained for compatibility #4167")
+    def test_egg_link_filename(self):
+        settings = dict(
+            name='Foo $$$ Bar_baz-bing',
+        )
+        dist = Distribution(settings)
+        cmd = develop(dist)
+        cmd.ensure_finalized()
+        link = pathlib.Path(cmd.egg_link)
+        assert link.suffix == '.egg-link'
+        assert link.stem == 'Foo_Bar_baz_bing'
+
+
+class TestResolver:
+    """
+    TODO: These tests were written with a minimal understanding
+    of what _resolve_setup_path is intending to do. Come up with
+    more meaningful cases that look like real-world scenarios.
+    """
+
+    def test_resolve_setup_path_cwd(self):
+        assert develop._resolve_setup_path('.', '.', '.') == '.'
+
+    def test_resolve_setup_path_one_dir(self):
+        assert develop._resolve_setup_path('pkgs', '.', 'pkgs') == '../'
+
+    def test_resolve_setup_path_one_dir_trailing_slash(self):
+        assert develop._resolve_setup_path('pkgs/', '.', 'pkgs') == '../'
+
+
+class TestNamespaces:
+    @staticmethod
+    def install_develop(src_dir, target):
+        develop_cmd = [
+            sys.executable,
+            'setup.py',
+            'develop',
+            '--install-dir',
+            str(target),
+        ]
+        with src_dir.as_cwd():
+            with paths_on_pythonpath([str(target)]):
+                subprocess.check_call(develop_cmd)
+
+    @pytest.mark.skipif(
+        bool(os.environ.get("APPVEYOR")),
+        reason="https://github.com/pypa/setuptools/issues/851",
+    )
+    @pytest.mark.skipif(
+        platform.python_implementation() == 'PyPy',
+        reason="https://github.com/pypa/setuptools/issues/1202",
+    )
+    def test_namespace_package_importable(self, tmpdir):
+        """
+        Installing two packages sharing the same namespace, one installed
+        naturally using pip or `--single-version-externally-managed`
+        and the other installed using `develop` should leave the namespace
+        in tact and both packages reachable by import.
+        """
+        pkg_A = namespaces.build_namespace_package(tmpdir, 'myns.pkgA')
+        pkg_B = namespaces.build_namespace_package(tmpdir, 'myns.pkgB')
+        target = tmpdir / 'packages'
+        # use pip to install to the target directory
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip',
+            'install',
+            str(pkg_A),
+            '-t',
+            str(target),
+        ]
+        subprocess.check_call(install_cmd)
+        self.install_develop(pkg_B, target)
+        namespaces.make_site_dir(target)
+        try_import = [
+            sys.executable,
+            '-c',
+            'import myns.pkgA; import myns.pkgB',
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(try_import)
+
+        # additionally ensure that pkg_resources import works
+        pkg_resources_imp = [
+            sys.executable,
+            '-c',
+            'import pkg_resources',
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(pkg_resources_imp)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist.py
new file mode 100644
index 00000000..533eb9f4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist.py
@@ -0,0 +1,278 @@
+import os
+import re
+import urllib.parse
+import urllib.request
+
+import pytest
+
+from setuptools import Distribution
+from setuptools.dist import check_package_data, check_specifier
+
+from .test_easy_install import make_nspkg_sdist
+from .test_find_packages import ensure_files
+from .textwrap import DALS
+
+from distutils.errors import DistutilsSetupError
+
+
+def test_dist_fetch_build_egg(tmpdir):
+    """
+    Check multiple calls to `Distribution.fetch_build_egg` work as expected.
+    """
+    index = tmpdir.mkdir('index')
+    index_url = urllib.parse.urljoin('file://', urllib.request.pathname2url(str(index)))
+
+    def sdist_with_index(distname, version):
+        dist_dir = index.mkdir(distname)
+        dist_sdist = f'{distname}-{version}.tar.gz'
+        make_nspkg_sdist(str(dist_dir.join(dist_sdist)), distname, version)
+        with dist_dir.join('index.html').open('w') as fp:
+            fp.write(
+                DALS(
+                    """
+                <!DOCTYPE html><html><body>
+                <a href="{dist_sdist}" rel="internal">{dist_sdist}</a><br/>
+                </body></html>
+                """
+                ).format(dist_sdist=dist_sdist)
+            )
+
+    sdist_with_index('barbazquux', '3.2.0')
+    sdist_with_index('barbazquux-runner', '2.11.1')
+    with tmpdir.join('setup.cfg').open('w') as fp:
+        fp.write(
+            DALS(
+                """
+            [easy_install]
+            index_url = {index_url}
+            """
+            ).format(index_url=index_url)
+        )
+    reqs = """
+    barbazquux-runner
+    barbazquux
+    """.split()
+    with tmpdir.as_cwd():
+        dist = Distribution()
+        dist.parse_config_files()
+        resolved_dists = [dist.fetch_build_egg(r) for r in reqs]
+    assert [dist.key for dist in resolved_dists if dist] == reqs
+
+
+EXAMPLE_BASE_INFO = dict(
+    name="package",
+    version="0.0.1",
+    author="Foo Bar",
+    author_email="foo@bar.net",
+    long_description="Long\ndescription",
+    description="Short description",
+    keywords=["one", "two"],
+)
+
+
+def test_provides_extras_deterministic_order():
+    attrs = dict(extras_require=dict(a=['foo'], b=['bar']))
+    dist = Distribution(attrs)
+    assert list(dist.metadata.provides_extras) == ['a', 'b']
+    attrs['extras_require'] = dict(reversed(attrs['extras_require'].items()))
+    dist = Distribution(attrs)
+    assert list(dist.metadata.provides_extras) == ['b', 'a']
+
+
+CHECK_PACKAGE_DATA_TESTS = (
+    # Valid.
+    (
+        {
+            '': ['*.txt', '*.rst'],
+            'hello': ['*.msg'],
+        },
+        None,
+    ),
+    # Not a dictionary.
+    (
+        (
+            ('', ['*.txt', '*.rst']),
+            ('hello', ['*.msg']),
+        ),
+        (
+            "'package_data' must be a dictionary mapping package"
+            " names to lists of string wildcard patterns"
+        ),
+    ),
+    # Invalid key type.
+    (
+        {
+            400: ['*.txt', '*.rst'],
+        },
+        ("keys of 'package_data' dict must be strings (got 400)"),
+    ),
+    # Invalid value type.
+    (
+        {
+            'hello': '*.msg',
+        },
+        (
+            "\"values of 'package_data' dict\" must be of type <tuple[str, ...] | list[str]>"
+            " (got '*.msg')"
+        ),
+    ),
+    # Invalid value type (generators are single use)
+    (
+        {
+            'hello': (x for x in "generator"),
+        },
+        (
+            "\"values of 'package_data' dict\" must be of type <tuple[str, ...] | list[str]>"
+            " (got <generator object"
+        ),
+    ),
+)
+
+
+@pytest.mark.parametrize(('package_data', 'expected_message'), CHECK_PACKAGE_DATA_TESTS)
+def test_check_package_data(package_data, expected_message):
+    if expected_message is None:
+        assert check_package_data(None, 'package_data', package_data) is None
+    else:
+        with pytest.raises(DistutilsSetupError, match=re.escape(expected_message)):
+            check_package_data(None, 'package_data', package_data)
+
+
+def test_check_specifier():
+    # valid specifier value
+    attrs = {'name': 'foo', 'python_requires': '>=3.0, !=3.1'}
+    dist = Distribution(attrs)
+    check_specifier(dist, attrs, attrs['python_requires'])
+
+    attrs = {'name': 'foo', 'python_requires': ['>=3.0', '!=3.1']}
+    dist = Distribution(attrs)
+    check_specifier(dist, attrs, attrs['python_requires'])
+
+    # invalid specifier value
+    attrs = {'name': 'foo', 'python_requires': '>=invalid-version'}
+    with pytest.raises(DistutilsSetupError):
+        dist = Distribution(attrs)
+
+
+def test_metadata_name():
+    with pytest.raises(DistutilsSetupError, match='missing.*name'):
+        Distribution()._validate_metadata()
+
+
+@pytest.mark.parametrize(
+    ('dist_name', 'py_module'),
+    [
+        ("my.pkg", "my_pkg"),
+        ("my-pkg", "my_pkg"),
+        ("my_pkg", "my_pkg"),
+        ("pkg", "pkg"),
+    ],
+)
+def test_dist_default_py_modules(tmp_path, dist_name, py_module):
+    (tmp_path / f"{py_module}.py").touch()
+
+    (tmp_path / "setup.py").touch()
+    (tmp_path / "noxfile.py").touch()
+    # ^-- make sure common tool files are ignored
+
+    attrs = {**EXAMPLE_BASE_INFO, "name": dist_name, "src_root": str(tmp_path)}
+    # Find `py_modules` corresponding to dist_name if not given
+    dist = Distribution(attrs)
+    dist.set_defaults()
+    assert dist.py_modules == [py_module]
+    # When `py_modules` is given, don't do anything
+    dist = Distribution({**attrs, "py_modules": ["explicity_py_module"]})
+    dist.set_defaults()
+    assert dist.py_modules == ["explicity_py_module"]
+    # When `packages` is given, don't do anything
+    dist = Distribution({**attrs, "packages": ["explicity_package"]})
+    dist.set_defaults()
+    assert not dist.py_modules
+
+
+@pytest.mark.parametrize(
+    ('dist_name', 'package_dir', 'package_files', 'packages'),
+    [
+        ("my.pkg", None, ["my_pkg/__init__.py", "my_pkg/mod.py"], ["my_pkg"]),
+        ("my-pkg", None, ["my_pkg/__init__.py", "my_pkg/mod.py"], ["my_pkg"]),
+        ("my_pkg", None, ["my_pkg/__init__.py", "my_pkg/mod.py"], ["my_pkg"]),
+        ("my.pkg", None, ["my/pkg/__init__.py"], ["my", "my.pkg"]),
+        (
+            "my_pkg",
+            None,
+            ["src/my_pkg/__init__.py", "src/my_pkg2/__init__.py"],
+            ["my_pkg", "my_pkg2"],
+        ),
+        (
+            "my_pkg",
+            {"pkg": "lib", "pkg2": "lib2"},
+            ["lib/__init__.py", "lib/nested/__init__.pyt", "lib2/__init__.py"],
+            ["pkg", "pkg.nested", "pkg2"],
+        ),
+    ],
+)
+def test_dist_default_packages(
+    tmp_path, dist_name, package_dir, package_files, packages
+):
+    ensure_files(tmp_path, package_files)
+
+    (tmp_path / "setup.py").touch()
+    (tmp_path / "noxfile.py").touch()
+    # ^-- should not be included by default
+
+    attrs = {
+        **EXAMPLE_BASE_INFO,
+        "name": dist_name,
+        "src_root": str(tmp_path),
+        "package_dir": package_dir,
+    }
+    # Find `packages` either corresponding to dist_name or inside src
+    dist = Distribution(attrs)
+    dist.set_defaults()
+    assert not dist.py_modules
+    assert not dist.py_modules
+    assert set(dist.packages) == set(packages)
+    # When `py_modules` is given, don't do anything
+    dist = Distribution({**attrs, "py_modules": ["explicit_py_module"]})
+    dist.set_defaults()
+    assert not dist.packages
+    assert set(dist.py_modules) == {"explicit_py_module"}
+    # When `packages` is given, don't do anything
+    dist = Distribution({**attrs, "packages": ["explicit_package"]})
+    dist.set_defaults()
+    assert not dist.py_modules
+    assert set(dist.packages) == {"explicit_package"}
+
+
+@pytest.mark.parametrize(
+    ('dist_name', 'package_dir', 'package_files'),
+    [
+        ("my.pkg.nested", None, ["my/pkg/nested/__init__.py"]),
+        ("my.pkg", None, ["my/pkg/__init__.py", "my/pkg/file.py"]),
+        ("my_pkg", None, ["my_pkg.py"]),
+        ("my_pkg", None, ["my_pkg/__init__.py", "my_pkg/nested/__init__.py"]),
+        ("my_pkg", None, ["src/my_pkg/__init__.py", "src/my_pkg/nested/__init__.py"]),
+        (
+            "my_pkg",
+            {"my_pkg": "lib", "my_pkg.lib2": "lib2"},
+            ["lib/__init__.py", "lib/nested/__init__.pyt", "lib2/__init__.py"],
+        ),
+        # Should not try to guess a name from multiple py_modules/packages
+        ("UNKNOWN", None, ["src/mod1.py", "src/mod2.py"]),
+        ("UNKNOWN", None, ["src/pkg1/__ini__.py", "src/pkg2/__init__.py"]),
+    ],
+)
+def test_dist_default_name(tmp_path, dist_name, package_dir, package_files):
+    """Make sure dist.name is discovered from packages/py_modules"""
+    ensure_files(tmp_path, package_files)
+    attrs = {
+        **EXAMPLE_BASE_INFO,
+        "src_root": "/".join(os.path.split(tmp_path)),  # POSIX-style
+        "package_dir": package_dir,
+    }
+    del attrs["name"]
+
+    dist = Distribution(attrs)
+    dist.set_defaults()
+    assert dist.py_modules or dist.packages
+    assert dist.get_name() == dist_name
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist_info.py
new file mode 100644
index 00000000..31e6e95a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_dist_info.py
@@ -0,0 +1,210 @@
+"""Test .dist-info style distributions."""
+
+import pathlib
+import re
+import shutil
+import subprocess
+import sys
+from functools import partial
+
+import pytest
+
+import pkg_resources
+from setuptools.archive_util import unpack_archive
+
+from .textwrap import DALS
+
+read = partial(pathlib.Path.read_text, encoding="utf-8")
+
+
+class TestDistInfo:
+    metadata_base = DALS(
+        """
+        Metadata-Version: 1.2
+        Requires-Dist: splort (==4)
+        Provides-Extra: baz
+        Requires-Dist: quux (>=1.1); extra == 'baz'
+        """
+    )
+
+    @classmethod
+    def build_metadata(cls, **kwargs):
+        lines = ('{key}: {value}\n'.format(**locals()) for key, value in kwargs.items())
+        return cls.metadata_base + ''.join(lines)
+
+    @pytest.fixture
+    def metadata(self, tmpdir):
+        dist_info_name = 'VersionedDistribution-2.718.dist-info'
+        versioned = tmpdir / dist_info_name
+        versioned.mkdir()
+        filename = versioned / 'METADATA'
+        content = self.build_metadata(
+            Name='VersionedDistribution',
+        )
+        filename.write_text(content, encoding='utf-8')
+
+        dist_info_name = 'UnversionedDistribution.dist-info'
+        unversioned = tmpdir / dist_info_name
+        unversioned.mkdir()
+        filename = unversioned / 'METADATA'
+        content = self.build_metadata(
+            Name='UnversionedDistribution',
+            Version='0.3',
+        )
+        filename.write_text(content, encoding='utf-8')
+
+        return str(tmpdir)
+
+    def test_distinfo(self, metadata):
+        dists = dict(
+            (d.project_name, d) for d in pkg_resources.find_distributions(metadata)
+        )
+
+        assert len(dists) == 2, dists
+
+        unversioned = dists['UnversionedDistribution']
+        versioned = dists['VersionedDistribution']
+
+        assert versioned.version == '2.718'  # from filename
+        assert unversioned.version == '0.3'  # from METADATA
+
+    def test_conditional_dependencies(self, metadata):
+        specs = 'splort==4', 'quux>=1.1'
+        requires = list(map(pkg_resources.Requirement.parse, specs))
+
+        for d in pkg_resources.find_distributions(metadata):
+            assert d.requires() == requires[:1]
+            assert d.requires(extras=('baz',)) == [
+                requires[0],
+                pkg_resources.Requirement.parse('quux>=1.1;extra=="baz"'),
+            ]
+            assert d.extras == ['baz']
+
+    def test_invalid_version(self, tmp_path):
+        """
+        Supplying an invalid version crashes dist_info.
+        """
+        config = "[metadata]\nname=proj\nversion=42\n[egg_info]\ntag_build=invalid!!!\n"
+        (tmp_path / "setup.cfg").write_text(config, encoding="utf-8")
+        msg = re.compile("invalid version", re.M | re.I)
+        proc = run_command_inner("dist_info", cwd=tmp_path, check=False)
+        assert proc.returncode
+        assert msg.search(proc.stdout)
+        assert not list(tmp_path.glob("*.dist-info"))
+
+    def test_tag_arguments(self, tmp_path):
+        config = """
+        [metadata]
+        name=proj
+        version=42
+        [egg_info]
+        tag_date=1
+        tag_build=.post
+        """
+        (tmp_path / "setup.cfg").write_text(config, encoding="utf-8")
+
+        print(run_command("dist_info", "--no-date", cwd=tmp_path))
+        dist_info = next(tmp_path.glob("*.dist-info"))
+        assert dist_info.name.startswith("proj-42")
+        shutil.rmtree(dist_info)
+
+        print(run_command("dist_info", "--tag-build", ".a", cwd=tmp_path))
+        dist_info = next(tmp_path.glob("*.dist-info"))
+        assert dist_info.name.startswith("proj-42a")
+
+    @pytest.mark.parametrize("keep_egg_info", (False, True))
+    def test_output_dir(self, tmp_path, keep_egg_info):
+        config = "[metadata]\nname=proj\nversion=42\n"
+        (tmp_path / "setup.cfg").write_text(config, encoding="utf-8")
+        out = tmp_path / "__out"
+        out.mkdir()
+        opts = ["--keep-egg-info"] if keep_egg_info else []
+        run_command("dist_info", "--output-dir", out, *opts, cwd=tmp_path)
+        assert len(list(out.glob("*.dist-info"))) == 1
+        assert len(list(tmp_path.glob("*.dist-info"))) == 0
+        expected_egg_info = int(keep_egg_info)
+        assert len(list(out.glob("*.egg-info"))) == expected_egg_info
+        assert len(list(tmp_path.glob("*.egg-info"))) == 0
+        assert len(list(out.glob("*.__bkp__"))) == 0
+        assert len(list(tmp_path.glob("*.__bkp__"))) == 0
+
+
+class TestWheelCompatibility:
+    """Make sure the .dist-info directory produced with the ``dist_info`` command
+    is the same as the one produced by ``bdist_wheel``.
+    """
+
+    SETUPCFG = DALS(
+        """
+    [metadata]
+    name = {name}
+    version = {version}
+
+    [options]
+    install_requires =
+        foo>=12; sys_platform != "linux"
+
+    [options.extras_require]
+    test = pytest
+
+    [options.entry_points]
+    console_scripts =
+        executable-name = my_package.module:function
+    discover =
+        myproj = my_package.other_module:function
+    """
+    )
+
+    EGG_INFO_OPTS = [
+        # Related: #3088 #2872
+        ("", ""),
+        (".post", "[egg_info]\ntag_build = post\n"),
+        (".post", "[egg_info]\ntag_build = .post\n"),
+        (".post", "[egg_info]\ntag_build = post\ntag_date = 1\n"),
+        (".dev", "[egg_info]\ntag_build = .dev\n"),
+        (".dev", "[egg_info]\ntag_build = .dev\ntag_date = 1\n"),
+        ("a1", "[egg_info]\ntag_build = .a1\n"),
+        ("+local", "[egg_info]\ntag_build = +local\n"),
+    ]
+
+    @pytest.mark.parametrize("name", "my-proj my_proj my.proj My.Proj".split())
+    @pytest.mark.parametrize("version", ["0.42.13"])
+    @pytest.mark.parametrize(("suffix", "cfg"), EGG_INFO_OPTS)
+    def test_dist_info_is_the_same_as_in_wheel(
+        self, name, version, tmp_path, suffix, cfg
+    ):
+        config = self.SETUPCFG.format(name=name, version=version) + cfg
+
+        for i in "dir_wheel", "dir_dist":
+            (tmp_path / i).mkdir()
+            (tmp_path / i / "setup.cfg").write_text(config, encoding="utf-8")
+
+        run_command("bdist_wheel", cwd=tmp_path / "dir_wheel")
+        wheel = next(tmp_path.glob("dir_wheel/dist/*.whl"))
+        unpack_archive(wheel, tmp_path / "unpack")
+        wheel_dist_info = next(tmp_path.glob("unpack/*.dist-info"))
+
+        run_command("dist_info", cwd=tmp_path / "dir_dist")
+        dist_info = next(tmp_path.glob("dir_dist/*.dist-info"))
+
+        assert dist_info.name == wheel_dist_info.name
+        assert dist_info.name.startswith(f"{name.replace('-', '_')}-{version}{suffix}")
+        for file in "METADATA", "entry_points.txt":
+            assert read(dist_info / file) == read(wheel_dist_info / file)
+
+
+def run_command_inner(*cmd, **kwargs):
+    opts = {
+        "stderr": subprocess.STDOUT,
+        "stdout": subprocess.PIPE,
+        "text": True,
+        "encoding": "utf-8",
+        "check": True,
+        **kwargs,
+    }
+    cmd = [sys.executable, "-c", "__import__('setuptools').setup()", *map(str, cmd)]
+    return subprocess.run(cmd, **opts)
+
+
+def run_command(*args, **kwargs):
+    return run_command_inner(*args, **kwargs).stdout
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_distutils_adoption.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_distutils_adoption.py
new file mode 100644
index 00000000..f99a5884
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_distutils_adoption.py
@@ -0,0 +1,198 @@
+import os
+import platform
+import sys
+import textwrap
+
+import pytest
+
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+_TEXT_KWARGS = {"text": True, "encoding": "utf-8"}  # For subprocess.run
+
+
+def win_sr(env):
+    """
+    On Windows, SYSTEMROOT must be present to avoid
+
+    > Fatal Python error: _Py_HashRandomization_Init: failed to
+    > get random numbers to initialize Python
+    """
+    if env and platform.system() == 'Windows':
+        env['SYSTEMROOT'] = os.environ['SYSTEMROOT']
+    return env
+
+
+def find_distutils(venv, imports='distutils', env=None, **kwargs):
+    py_cmd = 'import {imports}; print(distutils.__file__)'.format(**locals())
+    cmd = ['python', '-c', py_cmd]
+    return venv.run(cmd, env=win_sr(env), **_TEXT_KWARGS, **kwargs)
+
+
+def count_meta_path(venv, env=None):
+    py_cmd = textwrap.dedent(
+        """
+        import sys
+        is_distutils = lambda finder: finder.__class__.__name__ == "DistutilsMetaFinder"
+        print(len(list(filter(is_distutils, sys.meta_path))))
+        """
+    )
+    cmd = ['python', '-c', py_cmd]
+    return int(venv.run(cmd, env=win_sr(env), **_TEXT_KWARGS))
+
+
+skip_without_stdlib_distutils = pytest.mark.skipif(
+    sys.version_info >= (3, 12),
+    reason='stdlib distutils is removed from Python 3.12+',
+)
+
+
+@skip_without_stdlib_distutils
+def test_distutils_stdlib(venv):
+    """
+    Ensure stdlib distutils is used when appropriate.
+    """
+    env = dict(SETUPTOOLS_USE_DISTUTILS='stdlib')
+    assert venv.name not in find_distutils(venv, env=env).split(os.sep)
+    assert count_meta_path(venv, env=env) == 0
+
+
+def test_distutils_local_with_setuptools(venv):
+    """
+    Ensure local distutils is used when appropriate.
+    """
+    env = dict(SETUPTOOLS_USE_DISTUTILS='local')
+    loc = find_distutils(venv, imports='setuptools, distutils', env=env)
+    assert venv.name in loc.split(os.sep)
+    assert count_meta_path(venv, env=env) <= 1
+
+
+@pytest.mark.xfail('IS_PYPY', reason='pypy imports distutils on startup')
+def test_distutils_local(venv):
+    """
+    Even without importing, the setuptools-local copy of distutils is
+    preferred.
+    """
+    env = dict(SETUPTOOLS_USE_DISTUTILS='local')
+    assert venv.name in find_distutils(venv, env=env).split(os.sep)
+    assert count_meta_path(venv, env=env) <= 1
+
+
+def test_pip_import(venv):
+    """
+    Ensure pip can be imported.
+    Regression test for #3002.
+    """
+    cmd = ['python', '-c', 'import pip']
+    venv.run(cmd, **_TEXT_KWARGS)
+
+
+def test_distutils_has_origin():
+    """
+    Distutils module spec should have an origin. #2990.
+    """
+    assert __import__('distutils').__spec__.origin
+
+
+ENSURE_IMPORTS_ARE_NOT_DUPLICATED = r"""
+# Depending on the importlib machinery and _distutils_hack, some imports are
+# duplicated resulting in different module objects being loaded, which prevents
+# patches as shown in #3042.
+# This script provides a way of verifying if this duplication is happening.
+
+from distutils import cmd
+import distutils.command.sdist as sdist
+
+# import last to prevent caching
+from distutils import {imported_module}
+
+for mod in (cmd, sdist):
+    assert mod.{imported_module} == {imported_module}, (
+        f"\n{{mod.dir_util}}\n!=\n{{{imported_module}}}"
+    )
+
+print("success")
+"""
+
+
+@pytest.mark.usefixtures("tmpdir_cwd")
+@pytest.mark.parametrize(
+    ('distutils_version', 'imported_module'),
+    [
+        pytest.param("stdlib", "dir_util", marks=skip_without_stdlib_distutils),
+        pytest.param("stdlib", "file_util", marks=skip_without_stdlib_distutils),
+        pytest.param("stdlib", "archive_util", marks=skip_without_stdlib_distutils),
+        ("local", "dir_util"),
+        ("local", "file_util"),
+        ("local", "archive_util"),
+    ],
+)
+def test_modules_are_not_duplicated_on_import(distutils_version, imported_module, venv):
+    env = dict(SETUPTOOLS_USE_DISTUTILS=distutils_version)
+    script = ENSURE_IMPORTS_ARE_NOT_DUPLICATED.format(imported_module=imported_module)
+    cmd = ['python', '-c', script]
+    output = venv.run(cmd, env=win_sr(env), **_TEXT_KWARGS).strip()
+    assert output == "success"
+
+
+ENSURE_LOG_IMPORT_IS_NOT_DUPLICATED = r"""
+import types
+import distutils.dist as dist
+from distutils import log
+if isinstance(dist.log, types.ModuleType):
+    assert dist.log == log, f"\n{dist.log}\n!=\n{log}"
+print("success")
+"""
+
+
+@pytest.mark.usefixtures("tmpdir_cwd")
+@pytest.mark.parametrize(
+    "distutils_version",
+    [
+        "local",
+        pytest.param("stdlib", marks=skip_without_stdlib_distutils),
+    ],
+)
+def test_log_module_is_not_duplicated_on_import(distutils_version, venv):
+    env = dict(SETUPTOOLS_USE_DISTUTILS=distutils_version)
+    cmd = ['python', '-c', ENSURE_LOG_IMPORT_IS_NOT_DUPLICATED]
+    output = venv.run(cmd, env=win_sr(env), **_TEXT_KWARGS).strip()
+    assert output == "success"
+
+
+ENSURE_CONSISTENT_ERROR_FROM_MODIFIED_PY = r"""
+from setuptools.modified import newer
+from {imported_module}.errors import DistutilsError
+
+# Can't use pytest.raises in this context
+try:
+    newer("", "")
+except DistutilsError:
+    print("success")
+else:
+    raise AssertionError("Expected to raise")
+"""
+
+
+@pytest.mark.usefixtures("tmpdir_cwd")
+@pytest.mark.parametrize(
+    ('distutils_version', 'imported_module'),
+    [
+        ("local", "distutils"),
+        # Unfortunately we still get ._distutils.errors.DistutilsError with SETUPTOOLS_USE_DISTUTILS=stdlib
+        # But that's a deprecated use-case we don't mind not fully supporting in newer code
+        pytest.param(
+            "stdlib", "setuptools._distutils", marks=skip_without_stdlib_distutils
+        ),
+    ],
+)
+def test_consistent_error_from_modified_py(distutils_version, imported_module, venv):
+    env = dict(SETUPTOOLS_USE_DISTUTILS=distutils_version)
+    cmd = [
+        'python',
+        '-c',
+        ENSURE_CONSISTENT_ERROR_FROM_MODIFIED_PY.format(
+            imported_module=imported_module
+        ),
+    ]
+    output = venv.run(cmd, env=win_sr(env), **_TEXT_KWARGS).strip()
+    assert output == "success"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_easy_install.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_easy_install.py
new file mode 100644
index 00000000..e9b96027
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_easy_install.py
@@ -0,0 +1,1472 @@
+"""Easy install Tests"""
+
+import contextlib
+import io
+import itertools
+import logging
+import os
+import pathlib
+import re
+import site
+import subprocess
+import sys
+import tarfile
+import tempfile
+import time
+import warnings
+import zipfile
+from pathlib import Path
+from typing import NamedTuple
+from unittest import mock
+
+import pytest
+from jaraco import path
+
+import pkg_resources
+import setuptools.command.easy_install as ei
+from pkg_resources import Distribution as PRDistribution, normalize_path, working_set
+from setuptools import sandbox
+from setuptools.command.easy_install import PthDistributions
+from setuptools.dist import Distribution
+from setuptools.sandbox import run_setup
+from setuptools.tests import fail_on_ascii
+from setuptools.tests.server import MockServer, path_to_url
+
+from . import contexts
+from .textwrap import DALS
+
+import distutils.errors
+
+
+@pytest.fixture(autouse=True)
+def pip_disable_index(monkeypatch):
+    """
+    Important: Disable the default index for pip to avoid
+    querying packages in the index and potentially resolving
+    and installing packages there.
+    """
+    monkeypatch.setenv('PIP_NO_INDEX', 'true')
+
+
+class FakeDist:
+    def get_entry_map(self, group):
+        if group != 'console_scripts':
+            return {}
+        return {'name': 'ep'}
+
+    def as_requirement(self):
+        return 'spec'
+
+
+SETUP_PY = DALS(
+    """
+    from setuptools import setup
+
+    setup()
+    """
+)
+
+
+class TestEasyInstallTest:
+    def test_get_script_args(self):
+        header = ei.CommandSpec.best().from_environment().as_header()
+        dist = FakeDist()
+        args = next(ei.ScriptWriter.get_args(dist))
+        _name, script = itertools.islice(args, 2)
+        assert script.startswith(header)
+        assert "'spec'" in script
+        assert "'console_scripts'" in script
+        assert "'name'" in script
+        assert re.search('^# EASY-INSTALL-ENTRY-SCRIPT', script, flags=re.MULTILINE)
+
+    def test_no_find_links(self):
+        # new option '--no-find-links', that blocks find-links added at
+        # the project level
+        dist = Distribution()
+        cmd = ei.easy_install(dist)
+        cmd.check_pth_processing = lambda: True
+        cmd.no_find_links = True
+        cmd.find_links = ['link1', 'link2']
+        cmd.install_dir = os.path.join(tempfile.mkdtemp(), 'ok')
+        cmd.args = ['ok']
+        cmd.ensure_finalized()
+        assert cmd.package_index.scanned_urls == {}
+
+        # let's try without it (default behavior)
+        cmd = ei.easy_install(dist)
+        cmd.check_pth_processing = lambda: True
+        cmd.find_links = ['link1', 'link2']
+        cmd.install_dir = os.path.join(tempfile.mkdtemp(), 'ok')
+        cmd.args = ['ok']
+        cmd.ensure_finalized()
+        keys = sorted(cmd.package_index.scanned_urls.keys())
+        assert keys == ['link1', 'link2']
+
+    def test_write_exception(self):
+        """
+        Test that `cant_write_to_target` is rendered as a DistutilsError.
+        """
+        dist = Distribution()
+        cmd = ei.easy_install(dist)
+        cmd.install_dir = os.getcwd()
+        with pytest.raises(distutils.errors.DistutilsError):
+            cmd.cant_write_to_target()
+
+    def test_all_site_dirs(self, monkeypatch):
+        """
+        get_site_dirs should always return site dirs reported by
+        site.getsitepackages.
+        """
+        path = normalize_path('/setuptools/test/site-packages')
+
+        def mock_gsp():
+            return [path]
+
+        monkeypatch.setattr(site, 'getsitepackages', mock_gsp, raising=False)
+        assert path in ei.get_site_dirs()
+
+    def test_all_site_dirs_works_without_getsitepackages(self, monkeypatch):
+        monkeypatch.delattr(site, 'getsitepackages', raising=False)
+        assert ei.get_site_dirs()
+
+    @pytest.fixture
+    def sdist_unicode(self, tmpdir):
+        files = [
+            (
+                'setup.py',
+                DALS(
+                    """
+                    import setuptools
+                    setuptools.setup(
+                        name="setuptools-test-unicode",
+                        version="1.0",
+                        packages=["mypkg"],
+                        include_package_data=True,
+                    )
+                    """
+                ),
+            ),
+            (
+                'mypkg/__init__.py',
+                "",
+            ),
+            (
+                'mypkg/☃.txt',
+                "",
+            ),
+        ]
+        sdist_name = 'setuptools-test-unicode-1.0.zip'
+        sdist = tmpdir / sdist_name
+        # can't use make_sdist, because the issue only occurs
+        #  with zip sdists.
+        sdist_zip = zipfile.ZipFile(str(sdist), 'w')
+        for filename, content in files:
+            sdist_zip.writestr(filename, content)
+        sdist_zip.close()
+        return str(sdist)
+
+    @fail_on_ascii
+    def test_unicode_filename_in_sdist(self, sdist_unicode, tmpdir, monkeypatch):
+        """
+        The install command should execute correctly even if
+        the package has unicode filenames.
+        """
+        dist = Distribution({'script_args': ['easy_install']})
+        target = (tmpdir / 'target').ensure_dir()
+        cmd = ei.easy_install(
+            dist,
+            install_dir=str(target),
+            args=['x'],
+        )
+        monkeypatch.setitem(os.environ, 'PYTHONPATH', str(target))
+        cmd.ensure_finalized()
+        cmd.easy_install(sdist_unicode)
+
+    @pytest.fixture
+    def sdist_unicode_in_script(self, tmpdir):
+        files = [
+            (
+                "setup.py",
+                DALS(
+                    """
+                    import setuptools
+                    setuptools.setup(
+                        name="setuptools-test-unicode",
+                        version="1.0",
+                        packages=["mypkg"],
+                        include_package_data=True,
+                        scripts=['mypkg/unicode_in_script'],
+                    )
+                    """
+                ),
+            ),
+            ("mypkg/__init__.py", ""),
+            (
+                "mypkg/unicode_in_script",
+                DALS(
+                    """
+                    #!/bin/sh
+                    # á
+
+                    non_python_fn() {
+                    }
+                """
+                ),
+            ),
+        ]
+        sdist_name = "setuptools-test-unicode-script-1.0.zip"
+        sdist = tmpdir / sdist_name
+        # can't use make_sdist, because the issue only occurs
+        #  with zip sdists.
+        sdist_zip = zipfile.ZipFile(str(sdist), "w")
+        for filename, content in files:
+            sdist_zip.writestr(filename, content.encode('utf-8'))
+        sdist_zip.close()
+        return str(sdist)
+
+    @fail_on_ascii
+    def test_unicode_content_in_sdist(
+        self, sdist_unicode_in_script, tmpdir, monkeypatch
+    ):
+        """
+        The install command should execute correctly even if
+        the package has unicode in scripts.
+        """
+        dist = Distribution({"script_args": ["easy_install"]})
+        target = (tmpdir / "target").ensure_dir()
+        cmd = ei.easy_install(dist, install_dir=str(target), args=["x"])
+        monkeypatch.setitem(os.environ, "PYTHONPATH", str(target))
+        cmd.ensure_finalized()
+        cmd.easy_install(sdist_unicode_in_script)
+
+    @pytest.fixture
+    def sdist_script(self, tmpdir):
+        files = [
+            (
+                'setup.py',
+                DALS(
+                    """
+                    import setuptools
+                    setuptools.setup(
+                        name="setuptools-test-script",
+                        version="1.0",
+                        scripts=["mypkg_script"],
+                    )
+                    """
+                ),
+            ),
+            (
+                'mypkg_script',
+                DALS(
+                    """
+                     #/usr/bin/python
+                     print('mypkg_script')
+                     """
+                ),
+            ),
+        ]
+        sdist_name = 'setuptools-test-script-1.0.zip'
+        sdist = str(tmpdir / sdist_name)
+        make_sdist(sdist, files)
+        return sdist
+
+    @pytest.mark.skipif(
+        not sys.platform.startswith('linux'), reason="Test can only be run on Linux"
+    )
+    def test_script_install(self, sdist_script, tmpdir, monkeypatch):
+        """
+        Check scripts are installed.
+        """
+        dist = Distribution({'script_args': ['easy_install']})
+        target = (tmpdir / 'target').ensure_dir()
+        cmd = ei.easy_install(
+            dist,
+            install_dir=str(target),
+            args=['x'],
+        )
+        monkeypatch.setitem(os.environ, 'PYTHONPATH', str(target))
+        cmd.ensure_finalized()
+        cmd.easy_install(sdist_script)
+        assert (target / 'mypkg_script').exists()
+
+
+@pytest.mark.filterwarnings('ignore:Unbuilt egg')
+class TestPTHFileWriter:
+    def test_add_from_cwd_site_sets_dirty(self):
+        """a pth file manager should set dirty
+        if a distribution is in site but also the cwd
+        """
+        pth = PthDistributions('does-not_exist', [os.getcwd()])
+        assert not pth.dirty
+        pth.add(PRDistribution(os.getcwd()))
+        assert pth.dirty
+
+    def test_add_from_site_is_ignored(self):
+        location = '/test/location/does-not-have-to-exist'
+        # PthDistributions expects all locations to be normalized
+        location = pkg_resources.normalize_path(location)
+        pth = PthDistributions(
+            'does-not_exist',
+            [
+                location,
+            ],
+        )
+        assert not pth.dirty
+        pth.add(PRDistribution(location))
+        assert not pth.dirty
+
+    def test_many_pth_distributions_merge_together(self, tmpdir):
+        """
+        If the pth file is modified under the hood, then PthDistribution
+        will refresh its content before saving, merging contents when
+        necessary.
+        """
+        # putting the pth file in a dedicated sub-folder,
+        pth_subdir = tmpdir.join("pth_subdir")
+        pth_subdir.mkdir()
+        pth_path = str(pth_subdir.join("file1.pth"))
+        pth1 = PthDistributions(pth_path)
+        pth2 = PthDistributions(pth_path)
+        assert pth1.paths == pth2.paths == [], (
+            "unless there would be some default added at some point"
+        )
+        # and so putting the src_subdir in folder distinct than the pth one,
+        # so to keep it absolute by PthDistributions
+        new_src_path = tmpdir.join("src_subdir")
+        new_src_path.mkdir()  # must exist to be accounted
+        new_src_path_str = str(new_src_path)
+        pth1.paths.append(new_src_path_str)
+        pth1.save()
+        assert pth1.paths, (
+            "the new_src_path added must still be present/valid in pth1 after save"
+        )
+        # now,
+        assert new_src_path_str not in pth2.paths, (
+            "right before we save the entry should still not be present"
+        )
+        pth2.save()
+        assert new_src_path_str in pth2.paths, (
+            "the new_src_path entry should have been added by pth2 with its save() call"
+        )
+        assert pth2.paths[-1] == new_src_path, (
+            "and it should match exactly on the last entry actually "
+            "given we append to it in save()"
+        )
+        # finally,
+        assert PthDistributions(pth_path).paths == pth2.paths, (
+            "and we should have the exact same list at the end "
+            "with a fresh PthDistributions instance"
+        )
+
+
+@pytest.fixture
+def setup_context(tmpdir):
+    with (tmpdir / 'setup.py').open('w', encoding="utf-8") as f:
+        f.write(SETUP_PY)
+    with tmpdir.as_cwd():
+        yield tmpdir
+
+
+@pytest.mark.usefixtures("user_override")
+@pytest.mark.usefixtures("setup_context")
+class TestUserInstallTest:
+    # prevent check that site-packages is writable. easy_install
+    # shouldn't be writing to system site-packages during finalize
+    # options, but while it does, bypass the behavior.
+    prev_sp_write = mock.patch(
+        'setuptools.command.easy_install.easy_install.check_site_dir',
+        mock.Mock(),
+    )
+
+    # simulate setuptools installed in user site packages
+    @mock.patch('setuptools.command.easy_install.__file__', site.USER_SITE)
+    @mock.patch('site.ENABLE_USER_SITE', True)
+    @prev_sp_write
+    def test_user_install_not_implied_user_site_enabled(self):
+        self.assert_not_user_site()
+
+    @mock.patch('site.ENABLE_USER_SITE', False)
+    @prev_sp_write
+    def test_user_install_not_implied_user_site_disabled(self):
+        self.assert_not_user_site()
+
+    @staticmethod
+    def assert_not_user_site():
+        # create a finalized easy_install command
+        dist = Distribution()
+        dist.script_name = 'setup.py'
+        cmd = ei.easy_install(dist)
+        cmd.args = ['py']
+        cmd.ensure_finalized()
+        assert not cmd.user, 'user should not be implied'
+
+    def test_multiproc_atexit(self):
+        pytest.importorskip('multiprocessing')
+
+        log = logging.getLogger('test_easy_install')
+        logging.basicConfig(level=logging.INFO, stream=sys.stderr)
+        log.info('this should not break')
+
+    @pytest.fixture
+    def foo_package(self, tmpdir):
+        egg_file = tmpdir / 'foo-1.0.egg-info'
+        with egg_file.open('w') as f:
+            f.write('Name: foo\n')
+        return str(tmpdir)
+
+    @pytest.fixture
+    def install_target(self, tmpdir):
+        target = str(tmpdir)
+        with mock.patch('sys.path', sys.path + [target]):
+            python_path = os.path.pathsep.join(sys.path)
+            with mock.patch.dict(os.environ, PYTHONPATH=python_path):
+                yield target
+
+    def test_local_index(self, foo_package, install_target):
+        """
+        The local index must be used when easy_install locates installed
+        packages.
+        """
+        dist = Distribution()
+        dist.script_name = 'setup.py'
+        cmd = ei.easy_install(dist)
+        cmd.install_dir = install_target
+        cmd.args = ['foo']
+        cmd.ensure_finalized()
+        cmd.local_index.scan([foo_package])
+        res = cmd.easy_install('foo')
+        actual = os.path.normcase(os.path.realpath(res.location))
+        expected = os.path.normcase(os.path.realpath(foo_package))
+        assert actual == expected
+
+    @contextlib.contextmanager
+    def user_install_setup_context(self, *args, **kwargs):
+        """
+        Wrap sandbox.setup_context to patch easy_install in that context to
+        appear as user-installed.
+        """
+        with self.orig_context(*args, **kwargs):
+            import setuptools.command.easy_install as ei
+
+            ei.__file__ = site.USER_SITE
+            yield
+
+    def patched_setup_context(self):
+        self.orig_context = sandbox.setup_context
+
+        return mock.patch(
+            'setuptools.sandbox.setup_context',
+            self.user_install_setup_context,
+        )
+
+
+@pytest.fixture
+def distutils_package():
+    distutils_setup_py = SETUP_PY.replace(
+        'from setuptools import setup',
+        'from distutils.core import setup',
+    )
+    with contexts.tempdir(cd=os.chdir):
+        with open('setup.py', 'w', encoding="utf-8") as f:
+            f.write(distutils_setup_py)
+        yield
+
+
+@pytest.mark.usefixtures("distutils_package")
+class TestDistutilsPackage:
+    def test_bdist_egg_available_on_distutils_pkg(self):
+        run_setup('setup.py', ['bdist_egg'])
+
+
+@pytest.fixture
+def mock_index():
+    # set up a server which will simulate an alternate package index.
+    p_index = MockServer()
+    if p_index.server_port == 0:
+        # Some platforms (Jython) don't find a port to which to bind,
+        # so skip test for them.
+        pytest.skip("could not find a valid port")
+    p_index.start()
+    return p_index
+
+
+class TestInstallRequires:
+    def test_setup_install_includes_dependencies(self, tmp_path, mock_index):
+        """
+        When ``python setup.py install`` is called directly, it will use easy_install
+        to fetch dependencies.
+        """
+        # TODO: Remove these tests once `setup.py install` is completely removed
+        project_root = tmp_path / "project"
+        project_root.mkdir(exist_ok=True)
+        install_root = tmp_path / "install"
+        install_root.mkdir(exist_ok=True)
+
+        self.create_project(project_root)
+        cmd = [
+            sys.executable,
+            '-c',
+            '__import__("setuptools").setup()',
+            'install',
+            '--install-base',
+            str(install_root),
+            '--install-lib',
+            str(install_root),
+            '--install-headers',
+            str(install_root),
+            '--install-scripts',
+            str(install_root),
+            '--install-data',
+            str(install_root),
+            '--install-purelib',
+            str(install_root),
+            '--install-platlib',
+            str(install_root),
+        ]
+        env = {**os.environ, "__EASYINSTALL_INDEX": mock_index.url}
+        cp = subprocess.run(
+            cmd,
+            cwd=str(project_root),
+            env=env,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.STDOUT,
+            text=True,
+            encoding="utf-8",
+        )
+        assert cp.returncode != 0
+        try:
+            assert '/does-not-exist/' in {r.path for r in mock_index.requests}
+            assert next(
+                line
+                for line in cp.stdout.splitlines()
+                if "not find suitable distribution for" in line
+                and "does-not-exist" in line
+            )
+        except Exception:
+            if "failed to get random numbers" in cp.stdout:
+                pytest.xfail(f"{sys.platform} failure - {cp.stdout}")
+            raise
+
+    def create_project(self, root):
+        config = """
+        [metadata]
+        name = project
+        version = 42
+
+        [options]
+        install_requires = does-not-exist
+        py_modules = mod
+        """
+        (root / 'setup.cfg').write_text(DALS(config), encoding="utf-8")
+        (root / 'mod.py').touch()
+
+
+class TestSetupRequires:
+    def test_setup_requires_honors_fetch_params(self, mock_index, monkeypatch):
+        """
+        When easy_install installs a source distribution which specifies
+        setup_requires, it should honor the fetch parameters (such as
+        index-url, and find-links).
+        """
+        monkeypatch.setenv('PIP_RETRIES', '0')
+        monkeypatch.setenv('PIP_TIMEOUT', '0')
+        monkeypatch.setenv('PIP_NO_INDEX', 'false')
+        with contexts.quiet():
+            # create an sdist that has a build-time dependency.
+            with TestSetupRequires.create_sdist() as dist_file:
+                with contexts.tempdir() as temp_install_dir:
+                    with contexts.environment(PYTHONPATH=temp_install_dir):
+                        cmd = [
+                            sys.executable,
+                            '-c',
+                            '__import__("setuptools").setup()',
+                            'easy_install',
+                            '--index-url',
+                            mock_index.url,
+                            '--exclude-scripts',
+                            '--install-dir',
+                            temp_install_dir,
+                            dist_file,
+                        ]
+                        subprocess.Popen(cmd).wait()
+        # there should have been one requests to the server
+        assert [r.path for r in mock_index.requests] == ['/does-not-exist/']
+
+    @staticmethod
+    @contextlib.contextmanager
+    def create_sdist():
+        """
+        Return an sdist with a setup_requires dependency (of something that
+        doesn't exist)
+        """
+        with contexts.tempdir() as dir:
+            dist_path = os.path.join(dir, 'setuptools-test-fetcher-1.0.tar.gz')
+            make_sdist(
+                dist_path,
+                [
+                    (
+                        'setup.py',
+                        DALS(
+                            """
+                    import setuptools
+                    setuptools.setup(
+                        name="setuptools-test-fetcher",
+                        version="1.0",
+                        setup_requires = ['does-not-exist'],
+                    )
+                """
+                        ),
+                    ),
+                    ('setup.cfg', ''),
+                ],
+            )
+            yield dist_path
+
+    use_setup_cfg = (
+        (),
+        ('dependency_links',),
+        ('setup_requires',),
+        ('dependency_links', 'setup_requires'),
+    )
+
+    @pytest.mark.parametrize('use_setup_cfg', use_setup_cfg)
+    def test_setup_requires_overrides_version_conflict(self, use_setup_cfg):
+        """
+        Regression test for distribution issue 323:
+        https://bitbucket.org/tarek/distribute/issues/323
+
+        Ensures that a distribution's setup_requires requirements can still be
+        installed and used locally even if a conflicting version of that
+        requirement is already on the path.
+        """
+
+        fake_dist = PRDistribution(
+            'does-not-matter', project_name='foobar', version='0.0'
+        )
+        working_set.add(fake_dist)
+
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                test_pkg = create_setup_requires_package(
+                    temp_dir, use_setup_cfg=use_setup_cfg
+                )
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                with contexts.quiet() as (stdout, _stderr):
+                    # Don't even need to install the package, just
+                    # running the setup.py at all is sufficient
+                    run_setup(test_setup_py, ['--name'])
+
+                lines = stdout.readlines()
+                assert len(lines) > 0
+                assert lines[-1].strip() == 'test_pkg'
+
+    @pytest.mark.parametrize('use_setup_cfg', use_setup_cfg)
+    def test_setup_requires_override_nspkg(self, use_setup_cfg):
+        """
+        Like ``test_setup_requires_overrides_version_conflict`` but where the
+        ``setup_requires`` package is part of a namespace package that has
+        *already* been imported.
+        """
+
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                foobar_1_archive = os.path.join(temp_dir, 'foo.bar-0.1.tar.gz')
+                make_nspkg_sdist(foobar_1_archive, 'foo.bar', '0.1')
+                # Now actually go ahead an extract to the temp dir and add the
+                # extracted path to sys.path so foo.bar v0.1 is importable
+                foobar_1_dir = os.path.join(temp_dir, 'foo.bar-0.1')
+                os.mkdir(foobar_1_dir)
+                with tarfile.open(foobar_1_archive) as tf:
+                    tf.extraction_filter = lambda member, path: member
+                    tf.extractall(foobar_1_dir)
+                sys.path.insert(1, foobar_1_dir)
+
+                dist = PRDistribution(
+                    foobar_1_dir, project_name='foo.bar', version='0.1'
+                )
+                working_set.add(dist)
+
+                template = DALS(
+                    """\
+                    import foo  # Even with foo imported first the
+                                # setup_requires package should override
+                    import setuptools
+                    setuptools.setup(**%r)
+
+                    if not (hasattr(foo, '__path__') and
+                            len(foo.__path__) == 2):
+                        print('FAIL')
+
+                    if 'foo.bar-0.2' not in foo.__path__[0]:
+                        print('FAIL')
+                """
+                )
+
+                test_pkg = create_setup_requires_package(
+                    temp_dir,
+                    'foo.bar',
+                    '0.2',
+                    make_nspkg_sdist,
+                    template,
+                    use_setup_cfg=use_setup_cfg,
+                )
+
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+
+                with contexts.quiet() as (stdout, _stderr):
+                    try:
+                        # Don't even need to install the package, just
+                        # running the setup.py at all is sufficient
+                        run_setup(test_setup_py, ['--name'])
+                    except pkg_resources.VersionConflict:
+                        self.fail(
+                            'Installing setup.py requirements caused a VersionConflict'
+                        )
+
+                assert 'FAIL' not in stdout.getvalue()
+                lines = stdout.readlines()
+                assert len(lines) > 0
+                assert lines[-1].strip() == 'test_pkg'
+
+    @pytest.mark.parametrize('use_setup_cfg', use_setup_cfg)
+    def test_setup_requires_with_attr_version(self, use_setup_cfg):
+        def make_dependency_sdist(dist_path, distname, version):
+            files = [
+                (
+                    'setup.py',
+                    DALS(
+                        f"""
+                    import setuptools
+                    setuptools.setup(
+                        name={distname!r},
+                        version={version!r},
+                        py_modules=[{distname!r}],
+                    )
+                    """
+                    ),
+                ),
+                (
+                    distname + '.py',
+                    DALS(
+                        """
+                    version = 42
+                    """
+                    ),
+                ),
+            ]
+            make_sdist(dist_path, files)
+
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                test_pkg = create_setup_requires_package(
+                    temp_dir,
+                    setup_attrs=dict(version='attr: foobar.version'),
+                    make_package=make_dependency_sdist,
+                    use_setup_cfg=use_setup_cfg + ('version',),
+                )
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                with contexts.quiet() as (stdout, _stderr):
+                    run_setup(test_setup_py, ['--version'])
+                lines = stdout.readlines()
+                assert len(lines) > 0
+                assert lines[-1].strip() == '42'
+
+    def test_setup_requires_honors_pip_env(self, mock_index, monkeypatch):
+        monkeypatch.setenv('PIP_RETRIES', '0')
+        monkeypatch.setenv('PIP_TIMEOUT', '0')
+        monkeypatch.setenv('PIP_NO_INDEX', 'false')
+        monkeypatch.setenv('PIP_INDEX_URL', mock_index.url)
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                test_pkg = create_setup_requires_package(
+                    temp_dir,
+                    'python-xlib',
+                    '0.19',
+                    setup_attrs=dict(dependency_links=[]),
+                )
+                test_setup_cfg = os.path.join(test_pkg, 'setup.cfg')
+                with open(test_setup_cfg, 'w', encoding="utf-8") as fp:
+                    fp.write(
+                        DALS(
+                            """
+                        [easy_install]
+                        index_url = https://pypi.org/legacy/
+                        """
+                        )
+                    )
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                with pytest.raises(distutils.errors.DistutilsError):
+                    run_setup(test_setup_py, ['--version'])
+        assert len(mock_index.requests) == 1
+        assert mock_index.requests[0].path == '/python-xlib/'
+
+    def test_setup_requires_with_pep508_url(self, mock_index, monkeypatch):
+        monkeypatch.setenv('PIP_RETRIES', '0')
+        monkeypatch.setenv('PIP_TIMEOUT', '0')
+        monkeypatch.setenv('PIP_INDEX_URL', mock_index.url)
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                dep_sdist = os.path.join(temp_dir, 'dep.tar.gz')
+                make_trivial_sdist(dep_sdist, 'dependency', '42')
+                dep_url = path_to_url(dep_sdist, authority='localhost')
+                test_pkg = create_setup_requires_package(
+                    temp_dir,
+                    # Ignored (overridden by setup_attrs)
+                    'python-xlib',
+                    '0.19',
+                    setup_attrs=dict(setup_requires=f'dependency @ {dep_url}'),
+                )
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                run_setup(test_setup_py, ['--version'])
+        assert len(mock_index.requests) == 0
+
+    def test_setup_requires_with_allow_hosts(self, mock_index):
+        """The `allow-hosts` option in not supported anymore."""
+        files = {
+            'test_pkg': {
+                'setup.py': DALS(
+                    """
+                    from setuptools import setup
+                    setup(setup_requires='python-xlib')
+                    """
+                ),
+                'setup.cfg': DALS(
+                    """
+                    [easy_install]
+                    allow_hosts = *
+                    """
+                ),
+            }
+        }
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                path.build(files, prefix=temp_dir)
+                setup_py = str(pathlib.Path(temp_dir, 'test_pkg', 'setup.py'))
+                with pytest.raises(distutils.errors.DistutilsError):
+                    run_setup(setup_py, ['--version'])
+        assert len(mock_index.requests) == 0
+
+    def test_setup_requires_with_python_requires(self, monkeypatch, tmpdir):
+        """Check `python_requires` is honored."""
+        monkeypatch.setenv('PIP_RETRIES', '0')
+        monkeypatch.setenv('PIP_TIMEOUT', '0')
+        monkeypatch.setenv('PIP_NO_INDEX', '1')
+        monkeypatch.setenv('PIP_VERBOSE', '1')
+        dep_1_0_sdist = 'dep-1.0.tar.gz'
+        dep_1_0_url = path_to_url(str(tmpdir / dep_1_0_sdist))
+        dep_1_0_python_requires = '>=2.7'
+        make_python_requires_sdist(
+            str(tmpdir / dep_1_0_sdist), 'dep', '1.0', dep_1_0_python_requires
+        )
+        dep_2_0_sdist = 'dep-2.0.tar.gz'
+        dep_2_0_url = path_to_url(str(tmpdir / dep_2_0_sdist))
+        dep_2_0_python_requires = (
+            f'!={sys.version_info.major}.{sys.version_info.minor}.*'
+        )
+        make_python_requires_sdist(
+            str(tmpdir / dep_2_0_sdist), 'dep', '2.0', dep_2_0_python_requires
+        )
+        index = tmpdir / 'index.html'
+        index.write_text(
+            DALS(
+                """
+            <!DOCTYPE html>
+            <html><head><title>Links for dep</title></head>
+            <body>
+                <h1>Links for dep</h1>
+                <a href="{dep_1_0_url}"\
+data-requires-python="{dep_1_0_python_requires}">{dep_1_0_sdist}</a><br/>
+                <a href="{dep_2_0_url}"\
+data-requires-python="{dep_2_0_python_requires}">{dep_2_0_sdist}</a><br/>
+            </body>
+            </html>
+            """
+            ).format(
+                dep_1_0_url=dep_1_0_url,
+                dep_1_0_sdist=dep_1_0_sdist,
+                dep_1_0_python_requires=dep_1_0_python_requires,
+                dep_2_0_url=dep_2_0_url,
+                dep_2_0_sdist=dep_2_0_sdist,
+                dep_2_0_python_requires=dep_2_0_python_requires,
+            ),
+            'utf-8',
+        )
+        index_url = path_to_url(str(index))
+        with contexts.save_pkg_resources_state():
+            test_pkg = create_setup_requires_package(
+                str(tmpdir),
+                'python-xlib',
+                '0.19',  # Ignored (overridden by setup_attrs).
+                setup_attrs=dict(setup_requires='dep', dependency_links=[index_url]),
+            )
+            test_setup_py = os.path.join(test_pkg, 'setup.py')
+            run_setup(test_setup_py, ['--version'])
+        eggs = list(
+            map(str, pkg_resources.find_distributions(os.path.join(test_pkg, '.eggs')))
+        )
+        assert eggs == ['dep 1.0']
+
+    @pytest.mark.parametrize('with_dependency_links_in_setup_py', (False, True))
+    def test_setup_requires_with_find_links_in_setup_cfg(
+        self, monkeypatch, with_dependency_links_in_setup_py
+    ):
+        monkeypatch.setenv('PIP_RETRIES', '0')
+        monkeypatch.setenv('PIP_TIMEOUT', '0')
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                make_trivial_sdist(
+                    os.path.join(temp_dir, 'python-xlib-42.tar.gz'), 'python-xlib', '42'
+                )
+                test_pkg = os.path.join(temp_dir, 'test_pkg')
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                test_setup_cfg = os.path.join(test_pkg, 'setup.cfg')
+                os.mkdir(test_pkg)
+                with open(test_setup_py, 'w', encoding="utf-8") as fp:
+                    if with_dependency_links_in_setup_py:
+                        dependency_links = [os.path.join(temp_dir, 'links')]
+                    else:
+                        dependency_links = []
+                    fp.write(
+                        DALS(
+                            """
+                        from setuptools import installer, setup
+                        setup(setup_requires='python-xlib==42',
+                        dependency_links={dependency_links!r})
+                        """
+                        ).format(dependency_links=dependency_links)
+                    )
+                with open(test_setup_cfg, 'w', encoding="utf-8") as fp:
+                    fp.write(
+                        DALS(
+                            """
+                        [easy_install]
+                        index_url = {index_url}
+                        find_links = {find_links}
+                        """
+                        ).format(
+                            index_url=os.path.join(temp_dir, 'index'),
+                            find_links=temp_dir,
+                        )
+                    )
+                run_setup(test_setup_py, ['--version'])
+
+    def test_setup_requires_with_transitive_extra_dependency(self, monkeypatch):
+        """
+        Use case: installing a package with a build dependency on
+        an already installed `dep[extra]`, which in turn depends
+        on `extra_dep` (whose is not already installed).
+        """
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                # Create source distribution for `extra_dep`.
+                make_trivial_sdist(
+                    os.path.join(temp_dir, 'extra_dep-1.0.tar.gz'), 'extra_dep', '1.0'
+                )
+                # Create source tree for `dep`.
+                dep_pkg = os.path.join(temp_dir, 'dep')
+                os.mkdir(dep_pkg)
+                path.build(
+                    {
+                        'setup.py': DALS(
+                            """
+                          import setuptools
+                          setuptools.setup(
+                              name='dep', version='2.0',
+                              extras_require={'extra': ['extra_dep']},
+                          )
+                         """
+                        ),
+                        'setup.cfg': '',
+                    },
+                    prefix=dep_pkg,
+                )
+                # "Install" dep.
+                run_setup(os.path.join(dep_pkg, 'setup.py'), ['dist_info'])
+                working_set.add_entry(dep_pkg)
+                # Create source tree for test package.
+                test_pkg = os.path.join(temp_dir, 'test_pkg')
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                os.mkdir(test_pkg)
+                with open(test_setup_py, 'w', encoding="utf-8") as fp:
+                    fp.write(
+                        DALS(
+                            """
+                        from setuptools import installer, setup
+                        setup(setup_requires='dep[extra]')
+                        """
+                        )
+                    )
+                # Check...
+                monkeypatch.setenv('PIP_FIND_LINKS', str(temp_dir))
+                monkeypatch.setenv('PIP_NO_INDEX', '1')
+                monkeypatch.setenv('PIP_RETRIES', '0')
+                monkeypatch.setenv('PIP_TIMEOUT', '0')
+                run_setup(test_setup_py, ['--version'])
+
+    def test_setup_requires_with_distutils_command_dep(self, monkeypatch):
+        """
+        Use case: ensure build requirements' extras
+        are properly installed and activated.
+        """
+        with contexts.save_pkg_resources_state():
+            with contexts.tempdir() as temp_dir:
+                # Create source distribution for `extra_dep`.
+                make_sdist(
+                    os.path.join(temp_dir, 'extra_dep-1.0.tar.gz'),
+                    [
+                        (
+                            'setup.py',
+                            DALS(
+                                """
+                          import setuptools
+                          setuptools.setup(
+                              name='extra_dep',
+                              version='1.0',
+                              py_modules=['extra_dep'],
+                          )
+                          """
+                            ),
+                        ),
+                        ('setup.cfg', ''),
+                        ('extra_dep.py', ''),
+                    ],
+                )
+                # Create source tree for `epdep`.
+                dep_pkg = os.path.join(temp_dir, 'epdep')
+                os.mkdir(dep_pkg)
+                path.build(
+                    {
+                        'setup.py': DALS(
+                            """
+                          import setuptools
+                          setuptools.setup(
+                              name='dep', version='2.0',
+                              py_modules=['epcmd'],
+                              extras_require={'extra': ['extra_dep']},
+                              entry_points='''
+                                           [distutils.commands]
+                                           epcmd = epcmd:epcmd [extra]
+                                           ''',
+                          )
+                         """
+                        ),
+                        'setup.cfg': '',
+                        'epcmd.py': DALS(
+                            """
+                                     from distutils.command.build_py import build_py
+
+                                     import extra_dep
+
+                                     class epcmd(build_py):
+                                         pass
+                                     """
+                        ),
+                    },
+                    prefix=dep_pkg,
+                )
+                # "Install" dep.
+                run_setup(os.path.join(dep_pkg, 'setup.py'), ['dist_info'])
+                working_set.add_entry(dep_pkg)
+                # Create source tree for test package.
+                test_pkg = os.path.join(temp_dir, 'test_pkg')
+                test_setup_py = os.path.join(test_pkg, 'setup.py')
+                os.mkdir(test_pkg)
+                with open(test_setup_py, 'w', encoding="utf-8") as fp:
+                    fp.write(
+                        DALS(
+                            """
+                        from setuptools import installer, setup
+                        setup(setup_requires='dep[extra]')
+                        """
+                        )
+                    )
+                # Check...
+                monkeypatch.setenv('PIP_FIND_LINKS', str(temp_dir))
+                monkeypatch.setenv('PIP_NO_INDEX', '1')
+                monkeypatch.setenv('PIP_RETRIES', '0')
+                monkeypatch.setenv('PIP_TIMEOUT', '0')
+                run_setup(test_setup_py, ['epcmd'])
+
+
+def make_trivial_sdist(dist_path, distname, version):
+    """
+    Create a simple sdist tarball at dist_path, containing just a simple
+    setup.py.
+    """
+
+    make_sdist(
+        dist_path,
+        [
+            (
+                'setup.py',
+                DALS(
+                    f"""\
+             import setuptools
+             setuptools.setup(
+                 name={distname!r},
+                 version={version!r}
+             )
+         """
+                ),
+            ),
+            ('setup.cfg', ''),
+        ],
+    )
+
+
+def make_nspkg_sdist(dist_path, distname, version):
+    """
+    Make an sdist tarball with distname and version which also contains one
+    package with the same name as distname.  The top-level package is
+    designated a namespace package).
+    """
+
+    parts = distname.split('.')
+    nspackage = parts[0]
+
+    packages = ['.'.join(parts[:idx]) for idx in range(1, len(parts) + 1)]
+
+    setup_py = DALS(
+        f"""\
+        import setuptools
+        setuptools.setup(
+            name={distname!r},
+            version={version!r},
+            packages={packages!r},
+            namespace_packages=[{nspackage!r}]
+        )
+    """
+    )
+
+    init = "__import__('pkg_resources').declare_namespace(__name__)"
+
+    files = [('setup.py', setup_py), (os.path.join(nspackage, '__init__.py'), init)]
+    for package in packages[1:]:
+        filename = os.path.join(*(package.split('.') + ['__init__.py']))
+        files.append((filename, ''))
+
+    make_sdist(dist_path, files)
+
+
+def make_python_requires_sdist(dist_path, distname, version, python_requires):
+    make_sdist(
+        dist_path,
+        [
+            (
+                'setup.py',
+                DALS(
+                    """\
+                import setuptools
+                setuptools.setup(
+                  name={name!r},
+                  version={version!r},
+                  python_requires={python_requires!r},
+                )
+                """
+                ).format(
+                    name=distname, version=version, python_requires=python_requires
+                ),
+            ),
+            ('setup.cfg', ''),
+        ],
+    )
+
+
+def make_sdist(dist_path, files):
+    """
+    Create a simple sdist tarball at dist_path, containing the files
+    listed in ``files`` as ``(filename, content)`` tuples.
+    """
+
+    # Distributions with only one file don't play well with pip.
+    assert len(files) > 1
+    with tarfile.open(dist_path, 'w:gz') as dist:
+        for filename, content in files:
+            file_bytes = io.BytesIO(content.encode('utf-8'))
+            file_info = tarfile.TarInfo(name=filename)
+            file_info.size = len(file_bytes.getvalue())
+            file_info.mtime = int(time.time())
+            dist.addfile(file_info, fileobj=file_bytes)
+
+
+def create_setup_requires_package(
+    path,
+    distname='foobar',
+    version='0.1',
+    make_package=make_trivial_sdist,
+    setup_py_template=None,
+    setup_attrs=None,
+    use_setup_cfg=(),
+):
+    """Creates a source tree under path for a trivial test package that has a
+    single requirement in setup_requires--a tarball for that requirement is
+    also created and added to the dependency_links argument.
+
+    ``distname`` and ``version`` refer to the name/version of the package that
+    the test package requires via ``setup_requires``.  The name of the test
+    package itself is just 'test_pkg'.
+    """
+
+    test_setup_attrs = {
+        'name': 'test_pkg',
+        'version': '0.0',
+        'setup_requires': [f'{distname}=={version}'],
+        'dependency_links': [os.path.abspath(path)],
+    }
+    if setup_attrs:
+        test_setup_attrs.update(setup_attrs)
+
+    test_pkg = os.path.join(path, 'test_pkg')
+    os.mkdir(test_pkg)
+
+    # setup.cfg
+    if use_setup_cfg:
+        options = []
+        metadata = []
+        for name in use_setup_cfg:
+            value = test_setup_attrs.pop(name)
+            if name in 'name version'.split():
+                section = metadata
+            else:
+                section = options
+            if isinstance(value, (tuple, list)):
+                value = ';'.join(value)
+            section.append(f'{name}: {value}')
+        test_setup_cfg_contents = DALS(
+            """
+            [metadata]
+            {metadata}
+            [options]
+            {options}
+            """
+        ).format(
+            options='\n'.join(options),
+            metadata='\n'.join(metadata),
+        )
+    else:
+        test_setup_cfg_contents = ''
+    with open(os.path.join(test_pkg, 'setup.cfg'), 'w', encoding="utf-8") as f:
+        f.write(test_setup_cfg_contents)
+
+    # setup.py
+    if setup_py_template is None:
+        setup_py_template = DALS(
+            """\
+            import setuptools
+            setuptools.setup(**%r)
+        """
+        )
+    with open(os.path.join(test_pkg, 'setup.py'), 'w', encoding="utf-8") as f:
+        f.write(setup_py_template % test_setup_attrs)
+
+    foobar_path = os.path.join(path, f'{distname}-{version}.tar.gz')
+    make_package(foobar_path, distname, version)
+
+    return test_pkg
+
+
+@pytest.mark.skipif(
+    sys.platform.startswith('java') and ei.is_sh(sys.executable),
+    reason="Test cannot run under java when executable is sh",
+)
+class TestScriptHeader:
+    non_ascii_exe = '/Users/José/bin/python'
+    exe_with_spaces = r'C:\Program Files\Python36\python.exe'
+
+    def test_get_script_header(self):
+        expected = f'#!{ei.nt_quote_arg(os.path.normpath(sys.executable))}\n'
+        actual = ei.ScriptWriter.get_header('#!/usr/local/bin/python')
+        assert actual == expected
+
+    def test_get_script_header_args(self):
+        expected = f'#!{ei.nt_quote_arg(os.path.normpath(sys.executable))} -x\n'
+        actual = ei.ScriptWriter.get_header('#!/usr/bin/python -x')
+        assert actual == expected
+
+    def test_get_script_header_non_ascii_exe(self):
+        actual = ei.ScriptWriter.get_header(
+            '#!/usr/bin/python', executable=self.non_ascii_exe
+        )
+        expected = f'#!{self.non_ascii_exe} -x\n'
+        assert actual == expected
+
+    def test_get_script_header_exe_with_spaces(self):
+        actual = ei.ScriptWriter.get_header(
+            '#!/usr/bin/python', executable='"' + self.exe_with_spaces + '"'
+        )
+        expected = f'#!"{self.exe_with_spaces}"\n'
+        assert actual == expected
+
+
+class TestCommandSpec:
+    def test_custom_launch_command(self):
+        """
+        Show how a custom CommandSpec could be used to specify a #! executable
+        which takes parameters.
+        """
+        cmd = ei.CommandSpec(['/usr/bin/env', 'python3'])
+        assert cmd.as_header() == '#!/usr/bin/env python3\n'
+
+    def test_from_param_for_CommandSpec_is_passthrough(self):
+        """
+        from_param should return an instance of a CommandSpec
+        """
+        cmd = ei.CommandSpec(['python'])
+        cmd_new = ei.CommandSpec.from_param(cmd)
+        assert cmd is cmd_new
+
+    @mock.patch('sys.executable', TestScriptHeader.exe_with_spaces)
+    @mock.patch.dict(os.environ)
+    def test_from_environment_with_spaces_in_executable(self):
+        os.environ.pop('__PYVENV_LAUNCHER__', None)
+        cmd = ei.CommandSpec.from_environment()
+        assert len(cmd) == 1
+        assert cmd.as_header().startswith('#!"')
+
+    def test_from_simple_string_uses_shlex(self):
+        """
+        In order to support `executable = /usr/bin/env my-python`, make sure
+        from_param invokes shlex on that input.
+        """
+        cmd = ei.CommandSpec.from_param('/usr/bin/env my-python')
+        assert len(cmd) == 2
+        assert '"' not in cmd.as_header()
+
+    def test_from_param_raises_expected_error(self) -> None:
+        """
+        from_param should raise its own TypeError when the argument's type is unsupported
+        """
+        with pytest.raises(TypeError) as exc_info:
+            ei.CommandSpec.from_param(object())  # type: ignore[arg-type] # We want a type error here
+        assert (
+            str(exc_info.value) == "Argument has an unsupported type <class 'object'>"
+        ), exc_info.value
+
+
+class TestWindowsScriptWriter:
+    def test_header(self):
+        hdr = ei.WindowsScriptWriter.get_header('')
+        assert hdr.startswith('#!')
+        assert hdr.endswith('\n')
+        hdr = hdr.lstrip('#!')
+        hdr = hdr.rstrip('\n')
+        # header should not start with an escaped quote
+        assert not hdr.startswith('\\"')
+
+
+class VersionStub(NamedTuple):
+    major: int
+    minor: int
+    micro: int
+    releaselevel: str
+    serial: int
+
+
+def test_use_correct_python_version_string(tmpdir, tmpdir_cwd, monkeypatch):
+    # In issue #3001, easy_install wrongly uses the `python3.1` directory
+    # when the interpreter is `python3.10` and the `--user` option is given.
+    # See pypa/setuptools#3001.
+    dist = Distribution()
+    cmd = dist.get_command_obj('easy_install')
+    cmd.args = ['ok']
+    cmd.optimize = 0
+    cmd.user = True
+    cmd.install_userbase = str(tmpdir)
+    cmd.install_usersite = None
+    install_cmd = dist.get_command_obj('install')
+    install_cmd.install_userbase = str(tmpdir)
+    install_cmd.install_usersite = None
+
+    with monkeypatch.context() as patch, warnings.catch_warnings():
+        warnings.simplefilter("ignore")
+        version = '3.10.1 (main, Dec 21 2021, 09:17:12) [GCC 10.2.1 20210110]'
+        info = VersionStub(3, 10, 1, "final", 0)
+        patch.setattr('site.ENABLE_USER_SITE', True)
+        patch.setattr('sys.version', version)
+        patch.setattr('sys.version_info', info)
+        patch.setattr(cmd, 'create_home_path', mock.Mock())
+        cmd.finalize_options()
+
+    name = "pypy" if hasattr(sys, 'pypy_version_info') else "python"
+    install_dir = cmd.install_dir.lower()
+
+    # In some platforms (e.g. Windows), install_dir is mostly determined
+    # via `sysconfig`, which define constants eagerly at module creation.
+    # This means that monkeypatching `sys.version` to emulate 3.10 for testing
+    # may have no effect.
+    # The safest test here is to rely on the fact that 3.1 is no longer
+    # supported/tested, and make sure that if 'python3.1' ever appears in the string
+    # it is followed by another digit (e.g. 'python3.10').
+    if re.search(name + r'3\.?1', install_dir):
+        assert re.search(name + r'3\.?1\d', install_dir)
+
+    # The following "variables" are used for interpolation in distutils
+    # installation schemes, so it should be fair to treat them as "semi-public",
+    # or at least public enough so we can have a test to make sure they are correct
+    assert cmd.config_vars['py_version'] == '3.10.1'
+    assert cmd.config_vars['py_version_short'] == '3.10'
+    assert cmd.config_vars['py_version_nodot'] == '310'
+
+
+@pytest.mark.xfail(
+    sys.platform == "darwin",
+    reason="https://github.com/pypa/setuptools/pull/4716#issuecomment-2447624418",
+)
+def test_editable_user_and_build_isolation(setup_context, monkeypatch, tmp_path):
+    """`setup.py develop` should honor `--user` even under build isolation"""
+
+    # == Arrange ==
+    # Pretend that build isolation was enabled
+    # e.g pip sets the environment variable PYTHONNOUSERSITE=1
+    monkeypatch.setattr('site.ENABLE_USER_SITE', False)
+
+    # Patching $HOME for 2 reasons:
+    # 1. setuptools/command/easy_install.py:create_home_path
+    #    tries creating directories in $HOME.
+    #    Given::
+    #        self.config_vars['DESTDIRS'] = (
+    #            "/home/user/.pyenv/versions/3.9.10 "
+    #            "/home/user/.pyenv/versions/3.9.10/lib "
+    #            "/home/user/.pyenv/versions/3.9.10/lib/python3.9 "
+    #            "/home/user/.pyenv/versions/3.9.10/lib/python3.9/lib-dynload")
+    #    `create_home_path` will::
+    #        makedirs(
+    #            "/home/user/.pyenv/versions/3.9.10 "
+    #            "/home/user/.pyenv/versions/3.9.10/lib "
+    #            "/home/user/.pyenv/versions/3.9.10/lib/python3.9 "
+    #            "/home/user/.pyenv/versions/3.9.10/lib/python3.9/lib-dynload")
+    #
+    # 2. We are going to force `site` to update site.USER_BASE and site.USER_SITE
+    #    To point inside our new home
+    monkeypatch.setenv('HOME', str(tmp_path / '.home'))
+    monkeypatch.setenv('USERPROFILE', str(tmp_path / '.home'))
+    monkeypatch.setenv('APPDATA', str(tmp_path / '.home'))
+    monkeypatch.setattr('site.USER_BASE', None)
+    monkeypatch.setattr('site.USER_SITE', None)
+    user_site = Path(site.getusersitepackages())
+    user_site.mkdir(parents=True, exist_ok=True)
+
+    sys_prefix = tmp_path / '.sys_prefix'
+    sys_prefix.mkdir(parents=True, exist_ok=True)
+    monkeypatch.setattr('sys.prefix', str(sys_prefix))
+
+    setup_script = (
+        "__import__('setuptools').setup(name='aproj', version=42, packages=[])\n"
+    )
+    (tmp_path / "setup.py").write_text(setup_script, encoding="utf-8")
+
+    # == Sanity check ==
+    assert list(sys_prefix.glob("*")) == []
+    assert list(user_site.glob("*")) == []
+
+    # == Act ==
+    run_setup('setup.py', ['develop', '--user'])
+
+    # == Assert ==
+    # Should not install to sys.prefix
+    assert list(sys_prefix.glob("*")) == []
+    # Should install to user site
+    installed = {f.name for f in user_site.glob("*")}
+    # sometimes easy-install.pth is created and sometimes not
+    installed = installed - {"easy-install.pth"}
+    assert installed == {'aproj.egg-link'}
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_editable_install.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_editable_install.py
new file mode 100644
index 00000000..038dcadf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_editable_install.py
@@ -0,0 +1,1289 @@
+from __future__ import annotations
+
+import os
+import platform
+import stat
+import subprocess
+import sys
+from copy import deepcopy
+from importlib import import_module
+from importlib.machinery import EXTENSION_SUFFIXES
+from pathlib import Path
+from textwrap import dedent
+from typing import Any
+from unittest.mock import Mock
+from uuid import uuid4
+
+import jaraco.envs
+import jaraco.path
+import pytest
+from path import Path as _Path
+
+from setuptools._importlib import resources as importlib_resources
+from setuptools.command.editable_wheel import (
+    _DebuggingTips,
+    _encode_pth,
+    _find_namespaces,
+    _find_package_roots,
+    _find_virtual_namespaces,
+    _finder_template,
+    _LinkTree,
+    _TopLevelFinder,
+    editable_wheel,
+)
+from setuptools.dist import Distribution
+from setuptools.extension import Extension
+from setuptools.warnings import SetuptoolsDeprecationWarning
+
+from . import contexts, namespaces
+
+from distutils.core import run_setup
+
+
+@pytest.fixture(params=["strict", "lenient"])
+def editable_opts(request):
+    if request.param == "strict":
+        return ["--config-settings", "editable-mode=strict"]
+    return []
+
+
+EXAMPLE = {
+    'pyproject.toml': dedent(
+        """\
+        [build-system]
+        requires = ["setuptools"]
+        build-backend = "setuptools.build_meta"
+
+        [project]
+        name = "mypkg"
+        version = "3.14159"
+        license = {text = "MIT"}
+        description = "This is a Python package"
+        dynamic = ["readme"]
+        classifiers = [
+            "Development Status :: 5 - Production/Stable",
+            "Intended Audience :: Developers"
+        ]
+        urls = {Homepage = "https://github.com"}
+
+        [tool.setuptools]
+        package-dir = {"" = "src"}
+        packages = {find = {where = ["src"]}}
+        license-files = ["LICENSE*"]
+
+        [tool.setuptools.dynamic]
+        readme = {file = "README.rst"}
+
+        [tool.distutils.egg_info]
+        tag-build = ".post0"
+        """
+    ),
+    "MANIFEST.in": dedent(
+        """\
+        global-include *.py *.txt
+        global-exclude *.py[cod]
+        prune dist
+        prune build
+        """
+    ).strip(),
+    "README.rst": "This is a ``README``",
+    "LICENSE.txt": "---- placeholder MIT license ----",
+    "src": {
+        "mypkg": {
+            "__init__.py": dedent(
+                """\
+                import sys
+                from importlib.metadata import PackageNotFoundError, version
+
+                try:
+                    __version__ = version(__name__)
+                except PackageNotFoundError:
+                    __version__ = "unknown"
+                """
+            ),
+            "__main__.py": dedent(
+                """\
+                from importlib.resources import read_text
+                from . import __version__, __name__ as parent
+                from .mod import x
+
+                data = read_text(parent, "data.txt")
+                print(__version__, data, x)
+                """
+            ),
+            "mod.py": "x = ''",
+            "data.txt": "Hello World",
+        }
+    },
+}
+
+
+SETUP_SCRIPT_STUB = "__import__('setuptools').setup()"
+
+
+@pytest.mark.xfail(sys.platform == "darwin", reason="pypa/setuptools#4328")
+@pytest.mark.parametrize(
+    "files",
+    [
+        {**EXAMPLE, "setup.py": SETUP_SCRIPT_STUB},
+        EXAMPLE,  # No setup.py script
+    ],
+)
+def test_editable_with_pyproject(tmp_path, venv, files, editable_opts):
+    project = tmp_path / "mypkg"
+    project.mkdir()
+    jaraco.path.build(files, prefix=project)
+
+    cmd = [
+        "python",
+        "-m",
+        "pip",
+        "install",
+        "--no-build-isolation",  # required to force current version of setuptools
+        "-e",
+        str(project),
+        *editable_opts,
+    ]
+    print(venv.run(cmd))
+
+    cmd = ["python", "-m", "mypkg"]
+    assert venv.run(cmd).strip() == "3.14159.post0 Hello World"
+
+    (project / "src/mypkg/data.txt").write_text("foobar", encoding="utf-8")
+    (project / "src/mypkg/mod.py").write_text("x = 42", encoding="utf-8")
+    assert venv.run(cmd).strip() == "3.14159.post0 foobar 42"
+
+
+def test_editable_with_flat_layout(tmp_path, venv, editable_opts):
+    files = {
+        "mypkg": {
+            "pyproject.toml": dedent(
+                """\
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+
+                [project]
+                name = "mypkg"
+                version = "3.14159"
+
+                [tool.setuptools]
+                packages = ["pkg"]
+                py-modules = ["mod"]
+                """
+            ),
+            "pkg": {"__init__.py": "a = 4"},
+            "mod.py": "b = 2",
+        },
+    }
+    jaraco.path.build(files, prefix=tmp_path)
+    project = tmp_path / "mypkg"
+
+    cmd = [
+        "python",
+        "-m",
+        "pip",
+        "install",
+        "--no-build-isolation",  # required to force current version of setuptools
+        "-e",
+        str(project),
+        *editable_opts,
+    ]
+    print(venv.run(cmd))
+    cmd = ["python", "-c", "import pkg, mod; print(pkg.a, mod.b)"]
+    assert venv.run(cmd).strip() == "4 2"
+
+
+def test_editable_with_single_module(tmp_path, venv, editable_opts):
+    files = {
+        "mypkg": {
+            "pyproject.toml": dedent(
+                """\
+                [build-system]
+                requires = ["setuptools", "wheel"]
+                build-backend = "setuptools.build_meta"
+
+                [project]
+                name = "mod"
+                version = "3.14159"
+
+                [tool.setuptools]
+                py-modules = ["mod"]
+                """
+            ),
+            "mod.py": "b = 2",
+        },
+    }
+    jaraco.path.build(files, prefix=tmp_path)
+    project = tmp_path / "mypkg"
+
+    cmd = [
+        "python",
+        "-m",
+        "pip",
+        "install",
+        "--no-build-isolation",  # required to force current version of setuptools
+        "-e",
+        str(project),
+        *editable_opts,
+    ]
+    print(venv.run(cmd))
+    cmd = ["python", "-c", "import mod; print(mod.b)"]
+    assert venv.run(cmd).strip() == "2"
+
+
+class TestLegacyNamespaces:
+    # legacy => pkg_resources.declare_namespace(...) + setup(namespace_packages=...)
+
+    def test_nspkg_file_is_unique(self, tmp_path, monkeypatch):
+        deprecation = pytest.warns(
+            SetuptoolsDeprecationWarning, match=".*namespace_packages parameter.*"
+        )
+        installation_dir = tmp_path / ".installation_dir"
+        installation_dir.mkdir()
+        examples = (
+            "myns.pkgA",
+            "myns.pkgB",
+            "myns.n.pkgA",
+            "myns.n.pkgB",
+        )
+
+        for name in examples:
+            pkg = namespaces.build_namespace_package(tmp_path, name, version="42")
+            with deprecation, monkeypatch.context() as ctx:
+                ctx.chdir(pkg)
+                dist = run_setup("setup.py", stop_after="config")
+                cmd = editable_wheel(dist)
+                cmd.finalize_options()
+                editable_name = cmd.get_finalized_command("dist_info").name
+                cmd._install_namespaces(installation_dir, editable_name)
+
+        files = list(installation_dir.glob("*-nspkg.pth"))
+        assert len(files) == len(examples)
+
+    @pytest.mark.parametrize(
+        "impl",
+        (
+            "pkg_resources",
+            #  "pkgutil",  => does not work
+        ),
+    )
+    @pytest.mark.parametrize("ns", ("myns.n",))
+    def test_namespace_package_importable(
+        self, venv, tmp_path, ns, impl, editable_opts
+    ):
+        """
+        Installing two packages sharing the same namespace, one installed
+        naturally using pip or `--single-version-externally-managed`
+        and the other installed in editable mode should leave the namespace
+        intact and both packages reachable by import.
+        (Ported from test_develop).
+        """
+        build_system = """\
+        [build-system]
+        requires = ["setuptools"]
+        build-backend = "setuptools.build_meta"
+        """
+        pkg_A = namespaces.build_namespace_package(tmp_path, f"{ns}.pkgA", impl=impl)
+        pkg_B = namespaces.build_namespace_package(tmp_path, f"{ns}.pkgB", impl=impl)
+        (pkg_A / "pyproject.toml").write_text(build_system, encoding="utf-8")
+        (pkg_B / "pyproject.toml").write_text(build_system, encoding="utf-8")
+        # use pip to install to the target directory
+        opts = editable_opts[:]
+        opts.append("--no-build-isolation")  # force current version of setuptools
+        venv.run(["python", "-m", "pip", "install", str(pkg_A), *opts])
+        venv.run(["python", "-m", "pip", "install", "-e", str(pkg_B), *opts])
+        venv.run(["python", "-c", f"import {ns}.pkgA; import {ns}.pkgB"])
+        # additionally ensure that pkg_resources import works
+        venv.run(["python", "-c", "import pkg_resources"])
+
+
+class TestPep420Namespaces:
+    def test_namespace_package_importable(self, venv, tmp_path, editable_opts):
+        """
+        Installing two packages sharing the same namespace, one installed
+        normally using pip and the other installed in editable mode
+        should allow importing both packages.
+        """
+        pkg_A = namespaces.build_pep420_namespace_package(tmp_path, 'myns.n.pkgA')
+        pkg_B = namespaces.build_pep420_namespace_package(tmp_path, 'myns.n.pkgB')
+        # use pip to install to the target directory
+        opts = editable_opts[:]
+        opts.append("--no-build-isolation")  # force current version of setuptools
+        venv.run(["python", "-m", "pip", "install", str(pkg_A), *opts])
+        venv.run(["python", "-m", "pip", "install", "-e", str(pkg_B), *opts])
+        venv.run(["python", "-c", "import myns.n.pkgA; import myns.n.pkgB"])
+
+    def test_namespace_created_via_package_dir(self, venv, tmp_path, editable_opts):
+        """Currently users can create a namespace by tweaking `package_dir`"""
+        files = {
+            "pkgA": {
+                "pyproject.toml": dedent(
+                    """\
+                    [build-system]
+                    requires = ["setuptools", "wheel"]
+                    build-backend = "setuptools.build_meta"
+
+                    [project]
+                    name = "pkgA"
+                    version = "3.14159"
+
+                    [tool.setuptools]
+                    package-dir = {"myns.n.pkgA" = "src"}
+                    """
+                ),
+                "src": {"__init__.py": "a = 1"},
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+        pkg_A = tmp_path / "pkgA"
+        pkg_B = namespaces.build_pep420_namespace_package(tmp_path, 'myns.n.pkgB')
+        pkg_C = namespaces.build_pep420_namespace_package(tmp_path, 'myns.n.pkgC')
+
+        # use pip to install to the target directory
+        opts = editable_opts[:]
+        opts.append("--no-build-isolation")  # force current version of setuptools
+        venv.run(["python", "-m", "pip", "install", str(pkg_A), *opts])
+        venv.run(["python", "-m", "pip", "install", "-e", str(pkg_B), *opts])
+        venv.run(["python", "-m", "pip", "install", "-e", str(pkg_C), *opts])
+        venv.run(["python", "-c", "from myns.n import pkgA, pkgB, pkgC"])
+
+    def test_namespace_accidental_config_in_lenient_mode(self, venv, tmp_path):
+        """Sometimes users might specify an ``include`` pattern that ignores parent
+        packages. In a normal installation this would ignore all modules inside the
+        parent packages, and make them namespaces (reported in issue #3504),
+        so the editable mode should preserve this behaviour.
+        """
+        files = {
+            "pkgA": {
+                "pyproject.toml": dedent(
+                    """\
+                    [build-system]
+                    requires = ["setuptools", "wheel"]
+                    build-backend = "setuptools.build_meta"
+
+                    [project]
+                    name = "pkgA"
+                    version = "3.14159"
+
+                    [tool.setuptools]
+                    packages.find.include = ["mypkg.*"]
+                    """
+                ),
+                "mypkg": {
+                    "__init__.py": "",
+                    "other.py": "b = 1",
+                    "n": {
+                        "__init__.py": "",
+                        "pkgA.py": "a = 1",
+                    },
+                },
+                "MANIFEST.in": EXAMPLE["MANIFEST.in"],
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+        pkg_A = tmp_path / "pkgA"
+
+        # use pip to install to the target directory
+        opts = ["--no-build-isolation"]  # force current version of setuptools
+        venv.run(["python", "-m", "pip", "-v", "install", "-e", str(pkg_A), *opts])
+        out = venv.run(["python", "-c", "from mypkg.n import pkgA; print(pkgA.a)"])
+        assert out.strip() == "1"
+        cmd = """\
+        try:
+            import mypkg.other
+        except ImportError:
+            print("mypkg.other not defined")
+        """
+        out = venv.run(["python", "-c", dedent(cmd)])
+        assert "mypkg.other not defined" in out
+
+
+def test_editable_with_prefix(tmp_path, sample_project, editable_opts):
+    """
+    Editable install to a prefix should be discoverable.
+    """
+    prefix = tmp_path / 'prefix'
+
+    # figure out where pip will likely install the package
+    site_packages_all = [
+        prefix / Path(path).relative_to(sys.prefix)
+        for path in sys.path
+        if 'site-packages' in path and path.startswith(sys.prefix)
+    ]
+
+    for sp in site_packages_all:
+        sp.mkdir(parents=True)
+
+    # install workaround
+    _addsitedirs(site_packages_all)
+
+    env = dict(os.environ, PYTHONPATH=os.pathsep.join(map(str, site_packages_all)))
+    cmd = [
+        sys.executable,
+        '-m',
+        'pip',
+        'install',
+        '--editable',
+        str(sample_project),
+        '--prefix',
+        str(prefix),
+        '--no-build-isolation',
+        *editable_opts,
+    ]
+    subprocess.check_call(cmd, env=env)
+
+    # now run 'sample' with the prefix on the PYTHONPATH
+    bin = 'Scripts' if platform.system() == 'Windows' else 'bin'
+    exe = prefix / bin / 'sample'
+    subprocess.check_call([exe], env=env)
+
+
+class TestFinderTemplate:
+    """This test focus in getting a particular implementation detail right.
+    If at some point in time the implementation is changed for something different,
+    this test can be modified or even excluded.
+    """
+
+    def install_finder(self, finder):
+        loc = {}
+        exec(finder, loc, loc)
+        loc["install"]()
+
+    def test_packages(self, tmp_path):
+        files = {
+            "src1": {
+                "pkg1": {
+                    "__init__.py": "",
+                    "subpkg": {"mod1.py": "a = 42"},
+                },
+            },
+            "src2": {"mod2.py": "a = 43"},
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {
+            "pkg1": str(tmp_path / "src1/pkg1"),
+            "mod2": str(tmp_path / "src2/mod2"),
+        }
+        template = _finder_template(str(uuid4()), mapping, {})
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in ("pkg1", "pkg1.subpkg", "pkg1.subpkg.mod1", "mod2"):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+            mod1 = import_module("pkg1.subpkg.mod1")
+            mod2 = import_module("mod2")
+            subpkg = import_module("pkg1.subpkg")
+
+            assert mod1.a == 42
+            assert mod2.a == 43
+            expected = str((tmp_path / "src1/pkg1/subpkg").resolve())
+            assert_path(subpkg, expected)
+
+    def test_namespace(self, tmp_path):
+        files = {"pkg": {"__init__.py": "a = 13", "text.txt": "abc"}}
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {"ns.othername": str(tmp_path / "pkg")}
+        namespaces = {"ns": []}
+
+        template = _finder_template(str(uuid4()), mapping, namespaces)
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in ("ns", "ns.othername"):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+            pkg = import_module("ns.othername")
+            text = importlib_resources.files(pkg) / "text.txt"
+
+            expected = str((tmp_path / "pkg").resolve())
+            assert_path(pkg, expected)
+            assert pkg.a == 13
+
+            # Make sure resources can also be found
+            assert text.read_text(encoding="utf-8") == "abc"
+
+    def test_combine_namespaces(self, tmp_path):
+        files = {
+            "src1": {"ns": {"pkg1": {"__init__.py": "a = 13"}}},
+            "src2": {"ns": {"mod2.py": "b = 37"}},
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {
+            "ns.pkgA": str(tmp_path / "src1/ns/pkg1"),
+            "ns": str(tmp_path / "src2/ns"),
+        }
+        namespaces_ = {"ns": [str(tmp_path / "src1"), str(tmp_path / "src2")]}
+        template = _finder_template(str(uuid4()), mapping, namespaces_)
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in ("ns", "ns.pkgA", "ns.mod2"):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+            pkgA = import_module("ns.pkgA")
+            mod2 = import_module("ns.mod2")
+
+            expected = str((tmp_path / "src1/ns/pkg1").resolve())
+            assert_path(pkgA, expected)
+            assert pkgA.a == 13
+            assert mod2.b == 37
+
+    def test_combine_namespaces_nested(self, tmp_path):
+        """
+        Users may attempt to combine namespace packages in a nested way via
+        ``package_dir`` as shown in pypa/setuptools#4248.
+        """
+
+        files = {
+            "src": {"my_package": {"my_module.py": "a = 13"}},
+            "src2": {"my_package2": {"my_module2.py": "b = 37"}},
+        }
+
+        stack = jaraco.path.DirectoryStack()
+        with stack.context(tmp_path):
+            jaraco.path.build(files)
+            attrs = {
+                "script_name": "%PEP 517%",
+                "package_dir": {
+                    "different_name": "src/my_package",
+                    "different_name.subpkg": "src2/my_package2",
+                },
+                "packages": ["different_name", "different_name.subpkg"],
+            }
+            dist = Distribution(attrs)
+            finder = _TopLevelFinder(dist, str(uuid4()))
+            code = next(v for k, v in finder.get_implementation() if k.endswith(".py"))
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in attrs["packages"]:
+                sys.modules.pop(mod, None)
+
+            self.install_finder(code)
+            mod1 = import_module("different_name.my_module")
+            mod2 = import_module("different_name.subpkg.my_module2")
+
+            expected = str((tmp_path / "src/my_package/my_module.py").resolve())
+            assert str(Path(mod1.__file__).resolve()) == expected
+
+            expected = str((tmp_path / "src2/my_package2/my_module2.py").resolve())
+            assert str(Path(mod2.__file__).resolve()) == expected
+
+            assert mod1.a == 13
+            assert mod2.b == 37
+
+    def test_dynamic_path_computation(self, tmp_path):
+        # Follows the example in PEP 420
+        files = {
+            "project1": {"parent": {"child": {"one.py": "x = 1"}}},
+            "project2": {"parent": {"child": {"two.py": "x = 2"}}},
+            "project3": {"parent": {"child": {"three.py": "x = 3"}}},
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+        mapping = {}
+        namespaces_ = {"parent": [str(tmp_path / "project1/parent")]}
+        template = _finder_template(str(uuid4()), mapping, namespaces_)
+
+        mods = (f"parent.child.{name}" for name in ("one", "two", "three"))
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in ("parent", "parent.child", "parent.child", *mods):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+
+            one = import_module("parent.child.one")
+            assert one.x == 1
+
+            with pytest.raises(ImportError):
+                import_module("parent.child.two")
+
+            sys.path.append(str(tmp_path / "project2"))
+            two = import_module("parent.child.two")
+            assert two.x == 2
+
+            with pytest.raises(ImportError):
+                import_module("parent.child.three")
+
+            sys.path.append(str(tmp_path / "project3"))
+            three = import_module("parent.child.three")
+            assert three.x == 3
+
+    def test_no_recursion(self, tmp_path):
+        # See issue #3550
+        files = {
+            "pkg": {
+                "__init__.py": "from . import pkg",
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {
+            "pkg": str(tmp_path / "pkg"),
+        }
+        template = _finder_template(str(uuid4()), mapping, {})
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            sys.modules.pop("pkg", None)
+
+            self.install_finder(template)
+            with pytest.raises(ImportError, match="pkg"):
+                import_module("pkg")
+
+    def test_similar_name(self, tmp_path):
+        files = {
+            "foo": {
+                "__init__.py": "",
+                "bar": {
+                    "__init__.py": "",
+                },
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {
+            "foo": str(tmp_path / "foo"),
+        }
+        template = _finder_template(str(uuid4()), mapping, {})
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            sys.modules.pop("foo", None)
+            sys.modules.pop("foo.bar", None)
+
+            self.install_finder(template)
+            with pytest.raises(ImportError, match="foobar"):
+                import_module("foobar")
+
+    def test_case_sensitivity(self, tmp_path):
+        files = {
+            "foo": {
+                "__init__.py": "",
+                "lowercase.py": "x = 1",
+                "bar": {
+                    "__init__.py": "",
+                    "lowercase.py": "x = 2",
+                },
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+        mapping = {
+            "foo": str(tmp_path / "foo"),
+        }
+        template = _finder_template(str(uuid4()), mapping, {})
+        with contexts.save_paths(), contexts.save_sys_modules():
+            sys.modules.pop("foo", None)
+
+            self.install_finder(template)
+            with pytest.raises(ImportError, match="'FOO'"):
+                import_module("FOO")
+
+            with pytest.raises(ImportError, match="'foo\\.LOWERCASE'"):
+                import_module("foo.LOWERCASE")
+
+            with pytest.raises(ImportError, match="'foo\\.bar\\.Lowercase'"):
+                import_module("foo.bar.Lowercase")
+
+            with pytest.raises(ImportError, match="'foo\\.BAR'"):
+                import_module("foo.BAR.lowercase")
+
+            with pytest.raises(ImportError, match="'FOO'"):
+                import_module("FOO.bar.lowercase")
+
+            mod = import_module("foo.lowercase")
+            assert mod.x == 1
+
+            mod = import_module("foo.bar.lowercase")
+            assert mod.x == 2
+
+    def test_namespace_case_sensitivity(self, tmp_path):
+        files = {
+            "pkg": {
+                "__init__.py": "a = 13",
+                "foo": {
+                    "__init__.py": "b = 37",
+                    "bar.py": "c = 42",
+                },
+            },
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {"ns.othername": str(tmp_path / "pkg")}
+        namespaces = {"ns": []}
+
+        template = _finder_template(str(uuid4()), mapping, namespaces)
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in ("ns", "ns.othername"):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+            pkg = import_module("ns.othername")
+            expected = str((tmp_path / "pkg").resolve())
+            assert_path(pkg, expected)
+            assert pkg.a == 13
+
+            foo = import_module("ns.othername.foo")
+            assert foo.b == 37
+
+            bar = import_module("ns.othername.foo.bar")
+            assert bar.c == 42
+
+            with pytest.raises(ImportError, match="'NS'"):
+                import_module("NS.othername.foo")
+
+            with pytest.raises(ImportError, match="'ns\\.othername\\.FOO\\'"):
+                import_module("ns.othername.FOO")
+
+            with pytest.raises(ImportError, match="'ns\\.othername\\.foo\\.BAR\\'"):
+                import_module("ns.othername.foo.BAR")
+
+    def test_intermediate_packages(self, tmp_path):
+        """
+        The finder should not import ``fullname`` if the intermediate segments
+        don't exist (see pypa/setuptools#4019).
+        """
+        files = {
+            "src": {
+                "mypkg": {
+                    "__init__.py": "",
+                    "config.py": "a = 13",
+                    "helloworld.py": "b = 13",
+                    "components": {
+                        "config.py": "a = 37",
+                    },
+                },
+            }
+        }
+        jaraco.path.build(files, prefix=tmp_path)
+
+        mapping = {"mypkg": str(tmp_path / "src/mypkg")}
+        template = _finder_template(str(uuid4()), mapping, {})
+
+        with contexts.save_paths(), contexts.save_sys_modules():
+            for mod in (
+                "mypkg",
+                "mypkg.config",
+                "mypkg.helloworld",
+                "mypkg.components",
+                "mypkg.components.config",
+                "mypkg.components.helloworld",
+            ):
+                sys.modules.pop(mod, None)
+
+            self.install_finder(template)
+
+            config = import_module("mypkg.components.config")
+            assert config.a == 37
+
+            helloworld = import_module("mypkg.helloworld")
+            assert helloworld.b == 13
+
+            with pytest.raises(ImportError):
+                import_module("mypkg.components.helloworld")
+
+
+def test_pkg_roots(tmp_path):
+    """This test focus in getting a particular implementation detail right.
+    If at some point in time the implementation is changed for something different,
+    this test can be modified or even excluded.
+    """
+    files = {
+        "a": {"b": {"__init__.py": "ab = 1"}, "__init__.py": "a = 1"},
+        "d": {"__init__.py": "d = 1", "e": {"__init__.py": "de = 1"}},
+        "f": {"g": {"h": {"__init__.py": "fgh = 1"}}},
+        "other": {"__init__.py": "abc = 1"},
+        "another": {"__init__.py": "abcxyz = 1"},
+        "yet_another": {"__init__.py": "mnopq = 1"},
+    }
+    jaraco.path.build(files, prefix=tmp_path)
+    package_dir = {
+        "a.b.c": "other",
+        "a.b.c.x.y.z": "another",
+        "m.n.o.p.q": "yet_another",
+    }
+    packages = [
+        "a",
+        "a.b",
+        "a.b.c",
+        "a.b.c.x.y",
+        "a.b.c.x.y.z",
+        "d",
+        "d.e",
+        "f",
+        "f.g",
+        "f.g.h",
+        "m.n.o.p.q",
+    ]
+    roots = _find_package_roots(packages, package_dir, tmp_path)
+    assert roots == {
+        "a": str(tmp_path / "a"),
+        "a.b.c": str(tmp_path / "other"),
+        "a.b.c.x.y.z": str(tmp_path / "another"),
+        "d": str(tmp_path / "d"),
+        "f": str(tmp_path / "f"),
+        "m.n.o.p.q": str(tmp_path / "yet_another"),
+    }
+
+    ns = set(dict(_find_namespaces(packages, roots)))
+    assert ns == {"f", "f.g"}
+
+    ns = set(_find_virtual_namespaces(roots))
+    assert ns == {"a.b", "a.b.c.x", "a.b.c.x.y", "m", "m.n", "m.n.o", "m.n.o.p"}
+
+
+class TestOverallBehaviour:
+    PYPROJECT = """\
+        [build-system]
+        requires = ["setuptools"]
+        build-backend = "setuptools.build_meta"
+
+        [project]
+        name = "mypkg"
+        version = "3.14159"
+        """
+
+    # Any: Would need a TypedDict. Keep it simple for tests
+    FLAT_LAYOUT: dict[str, Any] = {
+        "pyproject.toml": dedent(PYPROJECT),
+        "MANIFEST.in": EXAMPLE["MANIFEST.in"],
+        "otherfile.py": "",
+        "mypkg": {
+            "__init__.py": "",
+            "mod1.py": "var = 42",
+            "subpackage": {
+                "__init__.py": "",
+                "mod2.py": "var = 13",
+                "resource_file.txt": "resource 39",
+            },
+        },
+    }
+
+    EXAMPLES = {
+        "flat-layout": FLAT_LAYOUT,
+        "src-layout": {
+            "pyproject.toml": dedent(PYPROJECT),
+            "MANIFEST.in": EXAMPLE["MANIFEST.in"],
+            "otherfile.py": "",
+            "src": {"mypkg": FLAT_LAYOUT["mypkg"]},
+        },
+        "custom-layout": {
+            "pyproject.toml": dedent(PYPROJECT)
+            + dedent(
+                """\
+                [tool.setuptools]
+                packages = ["mypkg", "mypkg.subpackage"]
+
+                [tool.setuptools.package-dir]
+                "mypkg.subpackage" = "other"
+                """
+            ),
+            "MANIFEST.in": EXAMPLE["MANIFEST.in"],
+            "otherfile.py": "",
+            "mypkg": {
+                "__init__.py": "",
+                "mod1.py": FLAT_LAYOUT["mypkg"]["mod1.py"],
+            },
+            "other": FLAT_LAYOUT["mypkg"]["subpackage"],
+        },
+        "namespace": {
+            "pyproject.toml": dedent(PYPROJECT),
+            "MANIFEST.in": EXAMPLE["MANIFEST.in"],
+            "otherfile.py": "",
+            "src": {
+                "mypkg": {
+                    "mod1.py": FLAT_LAYOUT["mypkg"]["mod1.py"],
+                    "subpackage": FLAT_LAYOUT["mypkg"]["subpackage"],
+                },
+            },
+        },
+    }
+
+    @pytest.mark.xfail(sys.platform == "darwin", reason="pypa/setuptools#4328")
+    @pytest.mark.parametrize("layout", EXAMPLES.keys())
+    def test_editable_install(self, tmp_path, venv, layout, editable_opts):
+        project, _ = install_project(
+            "mypkg", venv, tmp_path, self.EXAMPLES[layout], *editable_opts
+        )
+
+        # Ensure stray files are not importable
+        cmd_import_error = """\
+        try:
+            import otherfile
+        except ImportError as ex:
+            print(ex)
+        """
+        out = venv.run(["python", "-c", dedent(cmd_import_error)])
+        assert "No module named 'otherfile'" in out
+
+        # Ensure the modules are importable
+        cmd_get_vars = """\
+        import mypkg, mypkg.mod1, mypkg.subpackage.mod2
+        print(mypkg.mod1.var, mypkg.subpackage.mod2.var)
+        """
+        out = venv.run(["python", "-c", dedent(cmd_get_vars)])
+        assert "42 13" in out
+
+        # Ensure resources are reachable
+        cmd_get_resource = """\
+        import mypkg.subpackage
+        from setuptools._importlib import resources as importlib_resources
+        text = importlib_resources.files(mypkg.subpackage) / "resource_file.txt"
+        print(text.read_text(encoding="utf-8"))
+        """
+        out = venv.run(["python", "-c", dedent(cmd_get_resource)])
+        assert "resource 39" in out
+
+        # Ensure files are editable
+        mod1 = next(project.glob("**/mod1.py"))
+        mod2 = next(project.glob("**/mod2.py"))
+        resource_file = next(project.glob("**/resource_file.txt"))
+
+        mod1.write_text("var = 17", encoding="utf-8")
+        mod2.write_text("var = 781", encoding="utf-8")
+        resource_file.write_text("resource 374", encoding="utf-8")
+
+        out = venv.run(["python", "-c", dedent(cmd_get_vars)])
+        assert "42 13" not in out
+        assert "17 781" in out
+
+        out = venv.run(["python", "-c", dedent(cmd_get_resource)])
+        assert "resource 39" not in out
+        assert "resource 374" in out
+
+
+class TestLinkTree:
+    FILES = deepcopy(TestOverallBehaviour.EXAMPLES["src-layout"])
+    FILES["pyproject.toml"] += dedent(
+        """\
+        [tool.setuptools]
+        # Temporary workaround: both `include-package-data` and `package-data` configs
+        # can be removed after #3260 is fixed.
+        include-package-data = false
+        package-data = {"*" = ["*.txt"]}
+
+        [tool.setuptools.packages.find]
+        where = ["src"]
+        exclude = ["*.subpackage*"]
+        """
+    )
+    FILES["src"]["mypkg"]["resource.not_in_manifest"] = "abc"
+
+    def test_generated_tree(self, tmp_path):
+        jaraco.path.build(self.FILES, prefix=tmp_path)
+
+        with _Path(tmp_path):
+            name = "mypkg-3.14159"
+            dist = Distribution({"script_name": "%PEP 517%"})
+            dist.parse_config_files()
+
+            wheel = Mock()
+            aux = tmp_path / ".aux"
+            build = tmp_path / ".build"
+            aux.mkdir()
+            build.mkdir()
+
+            build_py = dist.get_command_obj("build_py")
+            build_py.editable_mode = True
+            build_py.build_lib = str(build)
+            build_py.ensure_finalized()
+            outputs = build_py.get_outputs()
+            output_mapping = build_py.get_output_mapping()
+
+            make_tree = _LinkTree(dist, name, aux, build)
+            make_tree(wheel, outputs, output_mapping)
+
+            mod1 = next(aux.glob("**/mod1.py"))
+            expected = tmp_path / "src/mypkg/mod1.py"
+            assert_link_to(mod1, expected)
+
+            assert next(aux.glob("**/subpackage"), None) is None
+            assert next(aux.glob("**/mod2.py"), None) is None
+            assert next(aux.glob("**/resource_file.txt"), None) is None
+
+            assert next(aux.glob("**/resource.not_in_manifest"), None) is None
+
+    def test_strict_install(self, tmp_path, venv):
+        opts = ["--config-settings", "editable-mode=strict"]
+        install_project("mypkg", venv, tmp_path, self.FILES, *opts)
+
+        out = venv.run(["python", "-c", "import mypkg.mod1; print(mypkg.mod1.var)"])
+        assert "42" in out
+
+        # Ensure packages excluded from distribution are not importable
+        cmd_import_error = """\
+        try:
+            from mypkg import subpackage
+        except ImportError as ex:
+            print(ex)
+        """
+        out = venv.run(["python", "-c", dedent(cmd_import_error)])
+        assert "cannot import name 'subpackage'" in out
+
+        # Ensure resource files excluded from distribution are not reachable
+        cmd_get_resource = """\
+        import mypkg
+        from setuptools._importlib import resources as importlib_resources
+        try:
+            text = importlib_resources.files(mypkg) / "resource.not_in_manifest"
+            print(text.read_text(encoding="utf-8"))
+        except FileNotFoundError as ex:
+            print(ex)
+        """
+        out = venv.run(["python", "-c", dedent(cmd_get_resource)])
+        assert "No such file or directory" in out
+        assert "resource.not_in_manifest" in out
+
+
+@pytest.mark.filterwarnings("ignore:.*compat.*:setuptools.SetuptoolsDeprecationWarning")
+def test_compat_install(tmp_path, venv):
+    # TODO: Remove `compat` after Dec/2022.
+    opts = ["--config-settings", "editable-mode=compat"]
+    files = TestOverallBehaviour.EXAMPLES["custom-layout"]
+    install_project("mypkg", venv, tmp_path, files, *opts)
+
+    out = venv.run(["python", "-c", "import mypkg.mod1; print(mypkg.mod1.var)"])
+    assert "42" in out
+
+    expected_path = comparable_path(str(tmp_path))
+
+    # Compatible behaviour will make spurious modules and excluded
+    # files importable directly from the original path
+    for cmd in (
+        "import otherfile; print(otherfile)",
+        "import other; print(other)",
+        "import mypkg; print(mypkg)",
+    ):
+        out = comparable_path(venv.run(["python", "-c", cmd]))
+        assert expected_path in out
+
+    # Compatible behaviour will not consider custom mappings
+    cmd = """\
+    try:
+        from mypkg import subpackage;
+    except ImportError as ex:
+        print(ex)
+    """
+    out = venv.run(["python", "-c", dedent(cmd)])
+    assert "cannot import name 'subpackage'" in out
+
+
+def test_pbr_integration(tmp_path, venv, editable_opts):
+    """Ensure editable installs work with pbr, issue #3500"""
+    files = {
+        "pyproject.toml": dedent(
+            """\
+            [build-system]
+            requires = ["setuptools"]
+            build-backend = "setuptools.build_meta"
+            """
+        ),
+        "setup.py": dedent(
+            """\
+            __import__('setuptools').setup(
+                pbr=True,
+                setup_requires=["pbr"],
+            )
+            """
+        ),
+        "setup.cfg": dedent(
+            """\
+            [metadata]
+            name = mypkg
+
+            [files]
+            packages =
+                mypkg
+            """
+        ),
+        "mypkg": {
+            "__init__.py": "",
+            "hello.py": "print('Hello world!')",
+        },
+        "other": {"test.txt": "Another file in here."},
+    }
+    venv.run(["python", "-m", "pip", "install", "pbr"])
+
+    with contexts.environment(PBR_VERSION="0.42"):
+        install_project("mypkg", venv, tmp_path, files, *editable_opts)
+
+    out = venv.run(["python", "-c", "import mypkg.hello"])
+    assert "Hello world!" in out
+
+
+class TestCustomBuildPy:
+    """
+    Issue #3501 indicates that some plugins/customizations might rely on:
+
+    1. ``build_py`` not running
+    2. ``build_py`` always copying files to ``build_lib``
+
+    During the transition period setuptools should prevent potential errors from
+    happening due to those assumptions.
+    """
+
+    # TODO: Remove tests after _run_build_steps is removed.
+
+    FILES = {
+        **TestOverallBehaviour.EXAMPLES["flat-layout"],
+        "setup.py": dedent(
+            """\
+            import pathlib
+            from setuptools import setup
+            from setuptools.command.build_py import build_py as orig
+
+            class my_build_py(orig):
+                def run(self):
+                    super().run()
+                    raise ValueError("TEST_RAISE")
+
+            setup(cmdclass={"build_py": my_build_py})
+            """
+        ),
+    }
+
+    def test_safeguarded_from_errors(self, tmp_path, venv):
+        """Ensure that errors in custom build_py are reported as warnings"""
+        # Warnings should show up
+        _, out = install_project("mypkg", venv, tmp_path, self.FILES)
+        assert "SetuptoolsDeprecationWarning" in out
+        assert "ValueError: TEST_RAISE" in out
+        # but installation should be successful
+        out = venv.run(["python", "-c", "import mypkg.mod1; print(mypkg.mod1.var)"])
+        assert "42" in out
+
+
+class TestCustomBuildWheel:
+    def install_custom_build_wheel(self, dist):
+        bdist_wheel_cls = dist.get_command_class("bdist_wheel")
+
+        class MyBdistWheel(bdist_wheel_cls):
+            def get_tag(self):
+                # In issue #3513, we can see that some extensions may try to access
+                # the `plat_name` property in bdist_wheel
+                if self.plat_name.startswith("macosx-"):
+                    _ = "macOS platform"
+                return super().get_tag()
+
+        dist.cmdclass["bdist_wheel"] = MyBdistWheel
+
+    def test_access_plat_name(self, tmpdir_cwd):
+        # Even when a custom bdist_wheel tries to access plat_name the build should
+        # be successful
+        jaraco.path.build({"module.py": "x = 42"})
+        dist = Distribution()
+        dist.script_name = "setup.py"
+        dist.set_defaults()
+        self.install_custom_build_wheel(dist)
+        cmd = editable_wheel(dist)
+        cmd.ensure_finalized()
+        cmd.run()
+        wheel_file = str(next(Path().glob('dist/*.whl')))
+        assert "editable" in wheel_file
+
+
+class TestCustomBuildExt:
+    def install_custom_build_ext_distutils(self, dist):
+        from distutils.command.build_ext import build_ext as build_ext_cls
+
+        class MyBuildExt(build_ext_cls):
+            pass
+
+        dist.cmdclass["build_ext"] = MyBuildExt
+
+    @pytest.mark.skipif(
+        sys.platform != "linux", reason="compilers may fail without correct setup"
+    )
+    def test_distutils_leave_inplace_files(self, tmpdir_cwd):
+        jaraco.path.build({"module.c": ""})
+        attrs = {
+            "ext_modules": [Extension("module", ["module.c"])],
+        }
+        dist = Distribution(attrs)
+        dist.script_name = "setup.py"
+        dist.set_defaults()
+        self.install_custom_build_ext_distutils(dist)
+        cmd = editable_wheel(dist)
+        cmd.ensure_finalized()
+        cmd.run()
+        wheel_file = str(next(Path().glob('dist/*.whl')))
+        assert "editable" in wheel_file
+        files = [p for p in Path().glob("module.*") if p.suffix != ".c"]
+        assert len(files) == 1
+        name = files[0].name
+        assert any(name.endswith(ext) for ext in EXTENSION_SUFFIXES)
+
+
+def test_debugging_tips(tmpdir_cwd, monkeypatch):
+    """Make sure to display useful debugging tips to the user."""
+    jaraco.path.build({"module.py": "x = 42"})
+    dist = Distribution()
+    dist.script_name = "setup.py"
+    dist.set_defaults()
+    cmd = editable_wheel(dist)
+    cmd.ensure_finalized()
+
+    SimulatedErr = type("SimulatedErr", (Exception,), {})
+    simulated_failure = Mock(side_effect=SimulatedErr())
+    monkeypatch.setattr(cmd, "get_finalized_command", simulated_failure)
+
+    expected_msg = "following steps are recommended to help debug"
+    with pytest.raises(SimulatedErr), pytest.warns(_DebuggingTips, match=expected_msg):
+        cmd.run()
+
+
+@pytest.mark.filterwarnings("error")
+def test_encode_pth():
+    """Ensure _encode_pth function does not produce encoding warnings"""
+    content = _encode_pth("tkmilan_ç_utf8")  # no warnings (would be turned into errors)
+    assert isinstance(content, bytes)
+
+
+def install_project(name, venv, tmp_path, files, *opts):
+    project = tmp_path / name
+    project.mkdir()
+    jaraco.path.build(files, prefix=project)
+    opts = [*opts, "--no-build-isolation"]  # force current version of setuptools
+    out = venv.run(
+        ["python", "-m", "pip", "-v", "install", "-e", str(project), *opts],
+        stderr=subprocess.STDOUT,
+    )
+    return project, out
+
+
+def _addsitedirs(new_dirs):
+    """To use this function, it is necessary to insert new_dir in front of sys.path.
+    The Python process will try to import a ``sitecustomize`` module on startup.
+    If we manipulate sys.path/PYTHONPATH, we can force it to run our code,
+    which invokes ``addsitedir`` and ensure ``.pth`` files are loaded.
+    """
+    content = '\n'.join(
+        ("import site",)
+        + tuple(f"site.addsitedir({os.fspath(new_dir)!r})" for new_dir in new_dirs)
+    )
+    (new_dirs[0] / "sitecustomize.py").write_text(content, encoding="utf-8")
+
+
+# ---- Assertion Helpers ----
+
+
+def assert_path(pkg, expected):
+    # __path__ is not guaranteed to exist, so we have to account for that
+    if pkg.__path__:
+        path = next(iter(pkg.__path__), None)
+        if path:
+            assert str(Path(path).resolve()) == expected
+
+
+def assert_link_to(file: Path, other: Path) -> None:
+    if file.is_symlink():
+        assert str(file.resolve()) == str(other.resolve())
+    else:
+        file_stat = file.stat()
+        other_stat = other.stat()
+        assert file_stat[stat.ST_INO] == other_stat[stat.ST_INO]
+        assert file_stat[stat.ST_DEV] == other_stat[stat.ST_DEV]
+
+
+def comparable_path(str_with_path: str) -> str:
+    return str_with_path.lower().replace(os.sep, "/").replace("//", "/")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_egg_info.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_egg_info.py
new file mode 100644
index 00000000..8233c9b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_egg_info.py
@@ -0,0 +1,1280 @@
+from __future__ import annotations
+
+import ast
+import glob
+import os
+import re
+import stat
+import sys
+import time
+from pathlib import Path
+from unittest import mock
+
+import pytest
+from jaraco import path
+
+from setuptools import errors
+from setuptools.command.egg_info import egg_info, manifest_maker, write_entries
+from setuptools.dist import Distribution
+
+from . import contexts, environment
+from .textwrap import DALS
+
+
+class Environment(str):
+    pass
+
+
+@pytest.fixture
+def env():
+    with contexts.tempdir(prefix='setuptools-test.') as env_dir:
+        env = Environment(env_dir)
+        os.chmod(env_dir, stat.S_IRWXU)
+        subs = 'home', 'lib', 'scripts', 'data', 'egg-base'
+        env.paths = dict((dirname, os.path.join(env_dir, dirname)) for dirname in subs)
+        list(map(os.mkdir, env.paths.values()))
+        path.build({
+            env.paths['home']: {
+                '.pydistutils.cfg': DALS(
+                    """
+                [egg_info]
+                egg-base = {egg-base}
+                """.format(**env.paths)
+                )
+            }
+        })
+        yield env
+
+
+class TestEggInfo:
+    setup_script = DALS(
+        """
+        from setuptools import setup
+
+        setup(
+            name='foo',
+            py_modules=['hello'],
+            entry_points={'console_scripts': ['hi = hello.run']},
+            zip_safe=False,
+        )
+        """
+    )
+
+    def _create_project(self):
+        path.build({
+            'setup.py': self.setup_script,
+            'hello.py': DALS(
+                """
+                def run():
+                    print('hello')
+                """
+            ),
+        })
+
+    @staticmethod
+    def _extract_mv_version(pkg_info_lines: list[str]) -> tuple[int, int]:
+        version_str = pkg_info_lines[0].split(' ')[1]
+        major, minor = map(int, version_str.split('.')[:2])
+        return major, minor
+
+    def test_egg_info_save_version_info_setup_empty(self, tmpdir_cwd, env):
+        """
+        When the egg_info section is empty or not present, running
+        save_version_info should add the settings to the setup.cfg
+        in a deterministic order.
+        """
+        setup_cfg = os.path.join(env.paths['home'], 'setup.cfg')
+        dist = Distribution()
+        ei = egg_info(dist)
+        ei.initialize_options()
+        ei.save_version_info(setup_cfg)
+
+        with open(setup_cfg, 'r', encoding="utf-8") as f:
+            content = f.read()
+
+        assert '[egg_info]' in content
+        assert 'tag_build =' in content
+        assert 'tag_date = 0' in content
+
+        expected_order = (
+            'tag_build',
+            'tag_date',
+        )
+
+        self._validate_content_order(content, expected_order)
+
+    @staticmethod
+    def _validate_content_order(content, expected):
+        """
+        Assert that the strings in expected appear in content
+        in order.
+        """
+        pattern = '.*'.join(expected)
+        flags = re.MULTILINE | re.DOTALL
+        assert re.search(pattern, content, flags)
+
+    def test_egg_info_save_version_info_setup_defaults(self, tmpdir_cwd, env):
+        """
+        When running save_version_info on an existing setup.cfg
+        with the 'default' values present from a previous run,
+        the file should remain unchanged.
+        """
+        setup_cfg = os.path.join(env.paths['home'], 'setup.cfg')
+        path.build({
+            setup_cfg: DALS(
+                """
+            [egg_info]
+            tag_build =
+            tag_date = 0
+            """
+            ),
+        })
+        dist = Distribution()
+        ei = egg_info(dist)
+        ei.initialize_options()
+        ei.save_version_info(setup_cfg)
+
+        with open(setup_cfg, 'r', encoding="utf-8") as f:
+            content = f.read()
+
+        assert '[egg_info]' in content
+        assert 'tag_build =' in content
+        assert 'tag_date = 0' in content
+
+        expected_order = (
+            'tag_build',
+            'tag_date',
+        )
+
+        self._validate_content_order(content, expected_order)
+
+    def test_expected_files_produced(self, tmpdir_cwd, env):
+        self._create_project()
+
+        self._run_egg_info_command(tmpdir_cwd, env)
+        actual = os.listdir('foo.egg-info')
+
+        expected = [
+            'PKG-INFO',
+            'SOURCES.txt',
+            'dependency_links.txt',
+            'entry_points.txt',
+            'not-zip-safe',
+            'top_level.txt',
+        ]
+        assert sorted(actual) == expected
+
+    def test_handling_utime_error(self, tmpdir_cwd, env):
+        dist = Distribution()
+        ei = egg_info(dist)
+        utime_patch = mock.patch('os.utime', side_effect=OSError("TEST"))
+        mkpath_patch = mock.patch(
+            'setuptools.command.egg_info.egg_info.mkpath', return_val=None
+        )
+
+        with utime_patch, mkpath_patch:
+            import distutils.errors
+
+            msg = r"Cannot update time stamp of directory 'None'"
+            with pytest.raises(distutils.errors.DistutilsFileError, match=msg):
+                ei.run()
+
+    def test_license_is_a_string(self, tmpdir_cwd, env):
+        setup_config = DALS(
+            """
+            [metadata]
+            name=foo
+            version=0.0.1
+            license=file:MIT
+            """
+        )
+
+        setup_script = DALS(
+            """
+            from setuptools import setup
+
+            setup()
+            """
+        )
+
+        path.build({
+            'setup.py': setup_script,
+            'setup.cfg': setup_config,
+        })
+
+        # This command should fail with a ValueError, but because it's
+        # currently configured to use a subprocess, the actual traceback
+        # object is lost and we need to parse it from stderr
+        with pytest.raises(AssertionError) as exc:
+            self._run_egg_info_command(tmpdir_cwd, env)
+
+        # The only argument to the assertion error should be a traceback
+        # containing a ValueError
+        assert 'ValueError' in exc.value.args[0]
+
+    def test_rebuilt(self, tmpdir_cwd, env):
+        """Ensure timestamps are updated when the command is re-run."""
+        self._create_project()
+
+        self._run_egg_info_command(tmpdir_cwd, env)
+        timestamp_a = os.path.getmtime('foo.egg-info')
+
+        # arbitrary sleep just to handle *really* fast systems
+        time.sleep(0.001)
+
+        self._run_egg_info_command(tmpdir_cwd, env)
+        timestamp_b = os.path.getmtime('foo.egg-info')
+
+        assert timestamp_a != timestamp_b
+
+    def test_manifest_template_is_read(self, tmpdir_cwd, env):
+        self._create_project()
+        path.build({
+            'MANIFEST.in': DALS(
+                """
+                recursive-include docs *.rst
+            """
+            ),
+            'docs': {
+                'usage.rst': "Run 'hi'",
+            },
+        })
+        self._run_egg_info_command(tmpdir_cwd, env)
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        sources_txt = os.path.join(egg_info_dir, 'SOURCES.txt')
+        with open(sources_txt, encoding="utf-8") as f:
+            assert 'docs/usage.rst' in f.read().split('\n')
+
+    def _setup_script_with_requires(self, requires, use_setup_cfg=False):
+        setup_script = DALS(
+            """
+            from setuptools import setup
+
+            setup(name='foo', zip_safe=False, %s)
+            """
+        ) % ('' if use_setup_cfg else requires)
+        setup_config = requires if use_setup_cfg else ''
+        path.build({
+            'setup.py': setup_script,
+            'setup.cfg': setup_config,
+        })
+
+    mismatch_marker = f"python_version<'{sys.version_info[0]}'"
+    # Alternate equivalent syntax.
+    mismatch_marker_alternate = f'python_version < "{sys.version_info[0]}"'
+    invalid_marker = "<=>++"
+
+    class RequiresTestHelper:
+        @staticmethod
+        def parametrize(*test_list, **format_dict):
+            idlist = []
+            argvalues = []
+            for test in test_list:
+                test_params = test.lstrip().split('\n\n', 3)
+                name_kwargs = test_params.pop(0).split('\n')
+                if len(name_kwargs) > 1:
+                    val = name_kwargs[1].strip()
+                    install_cmd_kwargs = ast.literal_eval(val)
+                else:
+                    install_cmd_kwargs = {}
+                name = name_kwargs[0].strip()
+                setup_py_requires, setup_cfg_requires, expected_requires = [
+                    DALS(a).format(**format_dict) for a in test_params
+                ]
+                for id_, requires, use_cfg in (
+                    (name, setup_py_requires, False),
+                    (name + '_in_setup_cfg', setup_cfg_requires, True),
+                ):
+                    idlist.append(id_)
+                    marks = ()
+                    if requires.startswith('@xfail\n'):
+                        requires = requires[7:]
+                        marks = pytest.mark.xfail
+                    argvalues.append(
+                        pytest.param(
+                            requires,
+                            use_cfg,
+                            expected_requires,
+                            install_cmd_kwargs,
+                            marks=marks,
+                        )
+                    )
+            return pytest.mark.parametrize(
+                'requires,use_setup_cfg,expected_requires,install_cmd_kwargs',
+                argvalues,
+                ids=idlist,
+            )
+
+    @RequiresTestHelper.parametrize(
+        # Format of a test:
+        #
+        # id
+        # install_cmd_kwargs [optional]
+        #
+        # requires block (when used in setup.py)
+        #
+        # requires block (when used in setup.cfg)
+        #
+        # expected contents of requires.txt
+        """
+        install_requires_deterministic
+
+        install_requires=["wheel>=0.5", "pytest"]
+
+        [options]
+        install_requires =
+            wheel>=0.5
+            pytest
+
+        wheel>=0.5
+        pytest
+        """,
+        """
+        install_requires_ordered
+
+        install_requires=["pytest>=3.0.2,!=10.9999"]
+
+        [options]
+        install_requires =
+            pytest>=3.0.2,!=10.9999
+
+        pytest!=10.9999,>=3.0.2
+        """,
+        """
+        install_requires_with_marker
+
+        install_requires=["barbazquux;{mismatch_marker}"],
+
+        [options]
+        install_requires =
+            barbazquux; {mismatch_marker}
+
+        [:{mismatch_marker_alternate}]
+        barbazquux
+        """,
+        """
+        install_requires_with_extra
+        {'cmd': ['egg_info']}
+
+        install_requires=["barbazquux [test]"],
+
+        [options]
+        install_requires =
+            barbazquux [test]
+
+        barbazquux[test]
+        """,
+        """
+        install_requires_with_extra_and_marker
+
+        install_requires=["barbazquux [test]; {mismatch_marker}"],
+
+        [options]
+        install_requires =
+            barbazquux [test]; {mismatch_marker}
+
+        [:{mismatch_marker_alternate}]
+        barbazquux[test]
+        """,
+        """
+        setup_requires_with_markers
+
+        setup_requires=["barbazquux;{mismatch_marker}"],
+
+        [options]
+        setup_requires =
+            barbazquux; {mismatch_marker}
+
+        """,
+        """
+        extras_require_with_extra
+        {'cmd': ['egg_info']}
+
+        extras_require={{"extra": ["barbazquux [test]"]}},
+
+        [options.extras_require]
+        extra = barbazquux [test]
+
+        [extra]
+        barbazquux[test]
+        """,
+        """
+        extras_require_with_extra_and_marker_in_req
+
+        extras_require={{"extra": ["barbazquux [test]; {mismatch_marker}"]}},
+
+        [options.extras_require]
+        extra =
+            barbazquux [test]; {mismatch_marker}
+
+        [extra]
+
+        [extra:{mismatch_marker_alternate}]
+        barbazquux[test]
+        """,
+        # FIXME: ConfigParser does not allow : in key names!
+        """
+        extras_require_with_marker
+
+        extras_require={{":{mismatch_marker}": ["barbazquux"]}},
+
+        @xfail
+        [options.extras_require]
+        :{mismatch_marker} = barbazquux
+
+        [:{mismatch_marker}]
+        barbazquux
+        """,
+        """
+        extras_require_with_marker_in_req
+
+        extras_require={{"extra": ["barbazquux; {mismatch_marker}"]}},
+
+        [options.extras_require]
+        extra =
+            barbazquux; {mismatch_marker}
+
+        [extra]
+
+        [extra:{mismatch_marker_alternate}]
+        barbazquux
+        """,
+        """
+        extras_require_with_empty_section
+
+        extras_require={{"empty": []}},
+
+        [options.extras_require]
+        empty =
+
+        [empty]
+        """,
+        # Format arguments.
+        invalid_marker=invalid_marker,
+        mismatch_marker=mismatch_marker,
+        mismatch_marker_alternate=mismatch_marker_alternate,
+    )
+    def test_requires(
+        self,
+        tmpdir_cwd,
+        env,
+        requires,
+        use_setup_cfg,
+        expected_requires,
+        install_cmd_kwargs,
+    ):
+        self._setup_script_with_requires(requires, use_setup_cfg)
+        self._run_egg_info_command(tmpdir_cwd, env, **install_cmd_kwargs)
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        requires_txt = os.path.join(egg_info_dir, 'requires.txt')
+        if os.path.exists(requires_txt):
+            with open(requires_txt, encoding="utf-8") as fp:
+                install_requires = fp.read()
+        else:
+            install_requires = ''
+        assert install_requires.lstrip() == expected_requires
+        assert glob.glob(os.path.join(env.paths['lib'], 'barbazquux*')) == []
+
+    def test_install_requires_unordered_disallowed(self, tmpdir_cwd, env):
+        """
+        Packages that pass unordered install_requires sequences
+        should be rejected as they produce non-deterministic
+        builds. See #458.
+        """
+        req = 'install_requires={"fake-factory==0.5.2", "pytz"}'
+        self._setup_script_with_requires(req)
+        with pytest.raises(AssertionError):
+            self._run_egg_info_command(tmpdir_cwd, env)
+
+    def test_extras_require_with_invalid_marker(self, tmpdir_cwd, env):
+        tmpl = 'extras_require={{":{marker}": ["barbazquux"]}},'
+        req = tmpl.format(marker=self.invalid_marker)
+        self._setup_script_with_requires(req)
+        with pytest.raises(AssertionError):
+            self._run_egg_info_command(tmpdir_cwd, env)
+        assert glob.glob(os.path.join(env.paths['lib'], 'barbazquux*')) == []
+
+    def test_extras_require_with_invalid_marker_in_req(self, tmpdir_cwd, env):
+        tmpl = 'extras_require={{"extra": ["barbazquux; {marker}"]}},'
+        req = tmpl.format(marker=self.invalid_marker)
+        self._setup_script_with_requires(req)
+        with pytest.raises(AssertionError):
+            self._run_egg_info_command(tmpdir_cwd, env)
+        assert glob.glob(os.path.join(env.paths['lib'], 'barbazquux*')) == []
+
+    def test_provides_extra(self, tmpdir_cwd, env):
+        self._setup_script_with_requires('extras_require={"foobar": ["barbazquux"]},')
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        assert 'Provides-Extra: foobar' in pkg_info_lines
+        assert 'Metadata-Version: 2.2' in pkg_info_lines
+
+    def test_doesnt_provides_extra(self, tmpdir_cwd, env):
+        self._setup_script_with_requires(
+            """install_requires=["spam ; python_version<'3.6'"]"""
+        )
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_text = fp.read()
+        assert 'Provides-Extra:' not in pkg_info_text
+
+    @pytest.mark.parametrize(
+        ('files', 'license_in_sources'),
+        [
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE
+                              """
+                    ),
+                    'LICENSE': "Test license",
+                },
+                True,
+            ),  # with license
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = INVALID_LICENSE
+                              """
+                    ),
+                    'LICENSE': "Test license",
+                },
+                False,
+            ),  # with an invalid license
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              """
+                    ),
+                    'LICENSE': "Test license",
+                },
+                True,
+            ),  # no license_file attribute, LICENSE auto-included
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE
+                              """
+                    ),
+                    'MANIFEST.in': "exclude LICENSE",
+                    'LICENSE': "Test license",
+                },
+                True,
+            ),  # manifest is overwritten by license_file
+            pytest.param(
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICEN[CS]E*
+                              """
+                    ),
+                    'LICENSE': "Test license",
+                },
+                True,
+                id="glob_pattern",
+            ),
+        ],
+    )
+    def test_setup_cfg_license_file(self, tmpdir_cwd, env, files, license_in_sources):
+        self._create_project()
+        path.build(files)
+
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+
+        sources_text = Path(egg_info_dir, "SOURCES.txt").read_text(encoding="utf-8")
+
+        if license_in_sources:
+            assert 'LICENSE' in sources_text
+        else:
+            assert 'LICENSE' not in sources_text
+            # for invalid license test
+            assert 'INVALID_LICENSE' not in sources_text
+
+    @pytest.mark.parametrize(
+        ('files', 'incl_licenses', 'excl_licenses'),
+        [
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE-ABC
+                                  LICENSE-XYZ
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+                [],
+            ),  # with licenses
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files = LICENSE-ABC, LICENSE-XYZ
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+                [],
+            ),  # with commas
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE-ABC
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                ['LICENSE-ABC'],
+                ['LICENSE-XYZ'],
+            ),  # with one license
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                [],
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+            ),  # empty
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files = LICENSE-XYZ
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                ['LICENSE-XYZ'],
+                ['LICENSE-ABC'],
+            ),  # on same line
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE-ABC
+                                  INVALID_LICENSE
+                              """
+                    ),
+                    'LICENSE-ABC': "Test license",
+                },
+                ['LICENSE-ABC'],
+                ['INVALID_LICENSE'],
+            ),  # with an invalid license
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              """
+                    ),
+                    'LICENSE': "Test license",
+                },
+                ['LICENSE'],
+                [],
+            ),  # no license_files attribute, LICENSE auto-included
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files = LICENSE
+                              """
+                    ),
+                    'MANIFEST.in': "exclude LICENSE",
+                    'LICENSE': "Test license",
+                },
+                ['LICENSE'],
+                [],
+            ),  # manifest is overwritten by license_files
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE-ABC
+                                  LICENSE-XYZ
+                              """
+                    ),
+                    'MANIFEST.in': "exclude LICENSE-XYZ",
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                    # manifest is overwritten by license_files
+                },
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+                [],
+            ),
+            pytest.param(
+                {
+                    'setup.cfg': "",
+                    'LICENSE-ABC': "ABC license",
+                    'COPYING-ABC': "ABC copying",
+                    'NOTICE-ABC': "ABC notice",
+                    'AUTHORS-ABC': "ABC authors",
+                    'LICENCE-XYZ': "XYZ license",
+                    'LICENSE': "License",
+                    'INVALID-LICENSE': "Invalid license",
+                },
+                [
+                    'LICENSE-ABC',
+                    'COPYING-ABC',
+                    'NOTICE-ABC',
+                    'AUTHORS-ABC',
+                    'LICENCE-XYZ',
+                    'LICENSE',
+                ],
+                ['INVALID-LICENSE'],
+                # ('LICEN[CS]E*', 'COPYING*', 'NOTICE*', 'AUTHORS*')
+                id="default_glob_patterns",
+            ),
+            pytest.param(
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE*
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'NOTICE-XYZ': "XYZ notice",
+                },
+                ['LICENSE-ABC'],
+                ['NOTICE-XYZ'],
+                id="no_default_glob_patterns",
+            ),
+            pytest.param(
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_files =
+                                  LICENSE-ABC
+                                  LICENSE*
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                },
+                ['LICENSE-ABC'],
+                [],
+                id="files_only_added_once",
+            ),
+        ],
+    )
+    def test_setup_cfg_license_files(
+        self, tmpdir_cwd, env, files, incl_licenses, excl_licenses
+    ):
+        self._create_project()
+        path.build(files)
+
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+
+        sources_text = Path(egg_info_dir, "SOURCES.txt").read_text(encoding="utf-8")
+        sources_lines = [line.strip() for line in sources_text.splitlines()]
+
+        for lf in incl_licenses:
+            assert sources_lines.count(lf) == 1
+
+        for lf in excl_licenses:
+            assert sources_lines.count(lf) == 0
+
+    @pytest.mark.parametrize(
+        ('files', 'incl_licenses', 'excl_licenses'),
+        [
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file =
+                              license_files =
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                [],
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+            ),  # both empty
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file =
+                                  LICENSE-ABC
+                                  LICENSE-XYZ
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-XYZ': "XYZ license",
+                    # license_file is still singular
+                },
+                [],
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+            ),
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE-ABC
+                              license_files =
+                                  LICENSE-XYZ
+                                  LICENSE-PQR
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-PQR': "PQR license",
+                    'LICENSE-XYZ': "XYZ license",
+                },
+                ['LICENSE-ABC', 'LICENSE-PQR', 'LICENSE-XYZ'],
+                [],
+            ),  # combined
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE-ABC
+                              license_files =
+                                  LICENSE-ABC
+                                  LICENSE-XYZ
+                                  LICENSE-PQR
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-PQR': "PQR license",
+                    'LICENSE-XYZ': "XYZ license",
+                    # duplicate license
+                },
+                ['LICENSE-ABC', 'LICENSE-PQR', 'LICENSE-XYZ'],
+                [],
+            ),
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE-ABC
+                              license_files =
+                                  LICENSE-XYZ
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-PQR': "PQR license",
+                    'LICENSE-XYZ': "XYZ license",
+                    # combined subset
+                },
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+                ['LICENSE-PQR'],
+            ),
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE-ABC
+                              license_files =
+                                  LICENSE-XYZ
+                                  LICENSE-PQR
+                              """
+                    ),
+                    'LICENSE-PQR': "Test license",
+                    # with invalid licenses
+                },
+                ['LICENSE-PQR'],
+                ['LICENSE-ABC', 'LICENSE-XYZ'],
+            ),
+            (
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE-ABC
+                              license_files =
+                                LICENSE-PQR
+                                LICENSE-XYZ
+                              """
+                    ),
+                    'MANIFEST.in': "exclude LICENSE-ABC\nexclude LICENSE-PQR",
+                    'LICENSE-ABC': "ABC license",
+                    'LICENSE-PQR': "PQR license",
+                    'LICENSE-XYZ': "XYZ license",
+                    # manifest is overwritten
+                },
+                ['LICENSE-ABC', 'LICENSE-PQR', 'LICENSE-XYZ'],
+                [],
+            ),
+            pytest.param(
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE*
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'NOTICE-XYZ': "XYZ notice",
+                },
+                ['LICENSE-ABC'],
+                ['NOTICE-XYZ'],
+                id="no_default_glob_patterns",
+            ),
+            pytest.param(
+                {
+                    'setup.cfg': DALS(
+                        """
+                              [metadata]
+                              license_file = LICENSE*
+                              license_files =
+                                NOTICE*
+                              """
+                    ),
+                    'LICENSE-ABC': "ABC license",
+                    'NOTICE-ABC': "ABC notice",
+                    'AUTHORS-ABC': "ABC authors",
+                },
+                ['LICENSE-ABC', 'NOTICE-ABC'],
+                ['AUTHORS-ABC'],
+                id="combined_glob_patterrns",
+            ),
+        ],
+    )
+    def test_setup_cfg_license_file_license_files(
+        self, tmpdir_cwd, env, files, incl_licenses, excl_licenses
+    ):
+        self._create_project()
+        path.build(files)
+
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+
+        sources_text = Path(egg_info_dir, "SOURCES.txt").read_text(encoding="utf-8")
+        sources_lines = [line.strip() for line in sources_text.splitlines()]
+
+        for lf in incl_licenses:
+            assert sources_lines.count(lf) == 1
+
+        for lf in excl_licenses:
+            assert sources_lines.count(lf) == 0
+
+    def test_license_file_attr_pkg_info(self, tmpdir_cwd, env):
+        """All matched license files should have a corresponding License-File."""
+        self._create_project()
+        path.build({
+            "setup.cfg": DALS(
+                """
+                              [metadata]
+                              license_files =
+                                  NOTICE*
+                                  LICENSE*
+                              """
+            ),
+            "LICENSE-ABC": "ABC license",
+            "LICENSE-XYZ": "XYZ license",
+            "NOTICE": "included",
+            "IGNORE": "not include",
+        })
+
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        license_file_lines = [
+            line for line in pkg_info_lines if line.startswith('License-File:')
+        ]
+
+        # Only 'NOTICE', LICENSE-ABC', and 'LICENSE-XYZ' should have been matched
+        # Also assert that order from license_files is keeped
+        assert "License-File: NOTICE" == license_file_lines[0]
+        assert "License-File: LICENSE-ABC" in license_file_lines[1:]
+        assert "License-File: LICENSE-XYZ" in license_file_lines[1:]
+
+    def test_metadata_version(self, tmpdir_cwd, env):
+        """Make sure latest metadata version is used by default."""
+        self._setup_script_with_requires("")
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        # Update metadata version if changed
+        assert self._extract_mv_version(pkg_info_lines) == (2, 2)
+
+    def test_long_description_content_type(self, tmpdir_cwd, env):
+        # Test that specifying a `long_description_content_type` keyword arg to
+        # the `setup` function results in writing a `Description-Content-Type`
+        # line to the `PKG-INFO` file in the `<distribution>.egg-info`
+        # directory.
+        # `Description-Content-Type` is described at
+        # https://github.com/pypa/python-packaging-user-guide/pull/258
+
+        self._setup_script_with_requires(
+            """long_description_content_type='text/markdown',"""
+        )
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        expected_line = 'Description-Content-Type: text/markdown'
+        assert expected_line in pkg_info_lines
+        assert 'Metadata-Version: 2.2' in pkg_info_lines
+
+    def test_long_description(self, tmpdir_cwd, env):
+        # Test that specifying `long_description` and `long_description_content_type`
+        # keyword args to the `setup` function results in writing
+        # the description in the message payload of the `PKG-INFO` file
+        # in the `<distribution>.egg-info` directory.
+        self._setup_script_with_requires(
+            "long_description='This is a long description\\nover multiple lines',"
+            "long_description_content_type='text/markdown',"
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        assert 'Metadata-Version: 2.2' in pkg_info_lines
+        assert '' == pkg_info_lines[-1]  # last line should be empty
+        long_desc_lines = pkg_info_lines[pkg_info_lines.index('') :]
+        assert 'This is a long description' in long_desc_lines
+        assert 'over multiple lines' in long_desc_lines
+
+    def test_project_urls(self, tmpdir_cwd, env):
+        # Test that specifying a `project_urls` dict to the `setup`
+        # function results in writing multiple `Project-URL` lines to
+        # the `PKG-INFO` file in the `<distribution>.egg-info`
+        # directory.
+        # `Project-URL` is described at https://packaging.python.org
+        #     /specifications/core-metadata/#project-url-multiple-use
+
+        self._setup_script_with_requires(
+            """project_urls={
+                'Link One': 'https://example.com/one/',
+                'Link Two': 'https://example.com/two/',
+                },"""
+        )
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        expected_line = 'Project-URL: Link One, https://example.com/one/'
+        assert expected_line in pkg_info_lines
+        expected_line = 'Project-URL: Link Two, https://example.com/two/'
+        assert expected_line in pkg_info_lines
+        assert self._extract_mv_version(pkg_info_lines) >= (1, 2)
+
+    def test_license(self, tmpdir_cwd, env):
+        """Test single line license."""
+        self._setup_script_with_requires("license='MIT',")
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        assert 'License: MIT' in pkg_info_lines
+
+    def test_license_escape(self, tmpdir_cwd, env):
+        """Test license is escaped correctly if longer than one line."""
+        self._setup_script_with_requires(
+            "license='This is a long license text \\nover multiple lines',"
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+
+        assert 'License: This is a long license text ' in pkg_info_lines
+        assert '        over multiple lines' in pkg_info_lines
+        assert 'text \n        over multiple' in '\n'.join(pkg_info_lines)
+
+    def test_python_requires_egg_info(self, tmpdir_cwd, env):
+        self._setup_script_with_requires("""python_requires='>=2.7.12',""")
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        environment.run_setup_py(
+            cmd=['egg_info'],
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        assert 'Requires-Python: >=2.7.12' in pkg_info_lines
+        assert self._extract_mv_version(pkg_info_lines) >= (1, 2)
+
+    def test_manifest_maker_warning_suppression(self):
+        fixtures = [
+            "standard file not found: should have one of foo.py, bar.py",
+            "standard file 'setup.py' not found",
+        ]
+
+        for msg in fixtures:
+            assert manifest_maker._should_suppress_warning(msg)
+
+    def test_egg_info_includes_setup_py(self, tmpdir_cwd):
+        self._create_project()
+        dist = Distribution({"name": "foo", "version": "0.0.1"})
+        dist.script_name = "non_setup.py"
+        egg_info_instance = egg_info(dist)
+        egg_info_instance.finalize_options()
+        egg_info_instance.run()
+
+        assert 'setup.py' in egg_info_instance.filelist.files
+
+        with open(egg_info_instance.egg_info + "/SOURCES.txt", encoding="utf-8") as f:
+            sources = f.read().split('\n')
+            assert 'setup.py' in sources
+
+    def _run_egg_info_command(self, tmpdir_cwd, env, cmd=None, output=None):
+        environ = os.environ.copy().update(
+            HOME=env.paths['home'],
+        )
+        if cmd is None:
+            cmd = [
+                'egg_info',
+            ]
+        code, data = environment.run_setup_py(
+            cmd=cmd,
+            pypath=os.pathsep.join([env.paths['lib'], str(tmpdir_cwd)]),
+            data_stream=1,
+            env=environ,
+        )
+        assert not code, data
+
+        if output:
+            assert output in data
+
+    def test_egg_info_tag_only_once(self, tmpdir_cwd, env):
+        self._create_project()
+        path.build({
+            'setup.cfg': DALS(
+                """
+                              [egg_info]
+                              tag_build = dev
+                              tag_date = 0
+                              tag_svn_revision = 0
+                              """
+            ),
+        })
+        self._run_egg_info_command(tmpdir_cwd, env)
+        egg_info_dir = os.path.join('.', 'foo.egg-info')
+        with open(os.path.join(egg_info_dir, 'PKG-INFO'), encoding="utf-8") as fp:
+            pkg_info_lines = fp.read().split('\n')
+        assert 'Version: 0.0.0.dev0' in pkg_info_lines
+
+
+class TestWriteEntries:
+    def test_invalid_entry_point(self, tmpdir_cwd, env):
+        dist = Distribution({"name": "foo", "version": "0.0.1"})
+        dist.entry_points = {"foo": "foo = invalid-identifier:foo"}
+        cmd = dist.get_command_obj("egg_info")
+        expected_msg = r"Problems to parse .*invalid-identifier.*"
+        with pytest.raises(errors.OptionError, match=expected_msg) as ex:
+            write_entries(cmd, "entry_points", "entry_points.txt")
+            assert "ensure entry-point follows the spec" in ex.value.args[0]
+
+    def test_valid_entry_point(self, tmpdir_cwd, env):
+        dist = Distribution({"name": "foo", "version": "0.0.1"})
+        dist.entry_points = {
+            "abc": "foo = bar:baz",
+            "def": ["faa = bor:boz"],
+        }
+        cmd = dist.get_command_obj("egg_info")
+        write_entries(cmd, "entry_points", "entry_points.txt")
+        content = Path("entry_points.txt").read_text(encoding="utf-8")
+        assert "[abc]\nfoo = bar:baz\n" in content
+        assert "[def]\nfaa = bor:boz\n" in content
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_extern.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_extern.py
new file mode 100644
index 00000000..d7eb3c62
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_extern.py
@@ -0,0 +1,15 @@
+import importlib
+import pickle
+
+import packaging
+
+from setuptools import Distribution
+
+
+def test_reimport_extern():
+    packaging2 = importlib.import_module(packaging.__name__)
+    assert packaging is packaging2
+
+
+def test_distribution_picklable():
+    pickle.loads(pickle.dumps(Distribution()))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_packages.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_packages.py
new file mode 100644
index 00000000..9fd9f8f6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_packages.py
@@ -0,0 +1,218 @@
+"""Tests for automatic package discovery"""
+
+import os
+import shutil
+import tempfile
+
+import pytest
+
+from setuptools import find_namespace_packages, find_packages
+from setuptools.discovery import FlatLayoutPackageFinder
+
+from .compat.py39 import os_helper
+
+
+class TestFindPackages:
+    def setup_method(self, method):
+        self.dist_dir = tempfile.mkdtemp()
+        self._make_pkg_structure()
+
+    def teardown_method(self, method):
+        shutil.rmtree(self.dist_dir)
+
+    def _make_pkg_structure(self):
+        """Make basic package structure.
+
+        dist/
+            docs/
+                conf.py
+            pkg/
+                __pycache__/
+                nspkg/
+                    mod.py
+                subpkg/
+                    assets/
+                        asset
+                    __init__.py
+            setup.py
+
+        """
+        self.docs_dir = self._mkdir('docs', self.dist_dir)
+        self._touch('conf.py', self.docs_dir)
+        self.pkg_dir = self._mkdir('pkg', self.dist_dir)
+        self._mkdir('__pycache__', self.pkg_dir)
+        self.ns_pkg_dir = self._mkdir('nspkg', self.pkg_dir)
+        self._touch('mod.py', self.ns_pkg_dir)
+        self.sub_pkg_dir = self._mkdir('subpkg', self.pkg_dir)
+        self.asset_dir = self._mkdir('assets', self.sub_pkg_dir)
+        self._touch('asset', self.asset_dir)
+        self._touch('__init__.py', self.sub_pkg_dir)
+        self._touch('setup.py', self.dist_dir)
+
+    def _mkdir(self, path, parent_dir=None):
+        if parent_dir:
+            path = os.path.join(parent_dir, path)
+        os.mkdir(path)
+        return path
+
+    def _touch(self, path, dir_=None):
+        if dir_:
+            path = os.path.join(dir_, path)
+        open(path, 'wb').close()
+        return path
+
+    def test_regular_package(self):
+        self._touch('__init__.py', self.pkg_dir)
+        packages = find_packages(self.dist_dir)
+        assert packages == ['pkg', 'pkg.subpkg']
+
+    def test_exclude(self):
+        self._touch('__init__.py', self.pkg_dir)
+        packages = find_packages(self.dist_dir, exclude=('pkg.*',))
+        assert packages == ['pkg']
+
+    def test_exclude_recursive(self):
+        """
+        Excluding a parent package should not exclude child packages as well.
+        """
+        self._touch('__init__.py', self.pkg_dir)
+        self._touch('__init__.py', self.sub_pkg_dir)
+        packages = find_packages(self.dist_dir, exclude=('pkg',))
+        assert packages == ['pkg.subpkg']
+
+    def test_include_excludes_other(self):
+        """
+        If include is specified, other packages should be excluded.
+        """
+        self._touch('__init__.py', self.pkg_dir)
+        alt_dir = self._mkdir('other_pkg', self.dist_dir)
+        self._touch('__init__.py', alt_dir)
+        packages = find_packages(self.dist_dir, include=['other_pkg'])
+        assert packages == ['other_pkg']
+
+    def test_dir_with_dot_is_skipped(self):
+        shutil.rmtree(os.path.join(self.dist_dir, 'pkg/subpkg/assets'))
+        data_dir = self._mkdir('some.data', self.pkg_dir)
+        self._touch('__init__.py', data_dir)
+        self._touch('file.dat', data_dir)
+        packages = find_packages(self.dist_dir)
+        assert 'pkg.some.data' not in packages
+
+    def test_dir_with_packages_in_subdir_is_excluded(self):
+        """
+        Ensure that a package in a non-package such as build/pkg/__init__.py
+        is excluded.
+        """
+        build_dir = self._mkdir('build', self.dist_dir)
+        build_pkg_dir = self._mkdir('pkg', build_dir)
+        self._touch('__init__.py', build_pkg_dir)
+        packages = find_packages(self.dist_dir)
+        assert 'build.pkg' not in packages
+
+    @pytest.mark.skipif(not os_helper.can_symlink(), reason='Symlink support required')
+    def test_symlinked_packages_are_included(self):
+        """
+        A symbolically-linked directory should be treated like any other
+        directory when matched as a package.
+
+        Create a link from lpkg -> pkg.
+        """
+        self._touch('__init__.py', self.pkg_dir)
+        linked_pkg = os.path.join(self.dist_dir, 'lpkg')
+        os.symlink('pkg', linked_pkg)
+        assert os.path.isdir(linked_pkg)
+        packages = find_packages(self.dist_dir)
+        assert 'lpkg' in packages
+
+    def _assert_packages(self, actual, expected):
+        assert set(actual) == set(expected)
+
+    def test_pep420_ns_package(self):
+        packages = find_namespace_packages(
+            self.dist_dir, include=['pkg*'], exclude=['pkg.subpkg.assets']
+        )
+        self._assert_packages(packages, ['pkg', 'pkg.nspkg', 'pkg.subpkg'])
+
+    def test_pep420_ns_package_no_includes(self):
+        packages = find_namespace_packages(self.dist_dir, exclude=['pkg.subpkg.assets'])
+        self._assert_packages(packages, ['docs', 'pkg', 'pkg.nspkg', 'pkg.subpkg'])
+
+    def test_pep420_ns_package_no_includes_or_excludes(self):
+        packages = find_namespace_packages(self.dist_dir)
+        expected = ['docs', 'pkg', 'pkg.nspkg', 'pkg.subpkg', 'pkg.subpkg.assets']
+        self._assert_packages(packages, expected)
+
+    def test_regular_package_with_nested_pep420_ns_packages(self):
+        self._touch('__init__.py', self.pkg_dir)
+        packages = find_namespace_packages(
+            self.dist_dir, exclude=['docs', 'pkg.subpkg.assets']
+        )
+        self._assert_packages(packages, ['pkg', 'pkg.nspkg', 'pkg.subpkg'])
+
+    def test_pep420_ns_package_no_non_package_dirs(self):
+        shutil.rmtree(self.docs_dir)
+        shutil.rmtree(os.path.join(self.dist_dir, 'pkg/subpkg/assets'))
+        packages = find_namespace_packages(self.dist_dir)
+        self._assert_packages(packages, ['pkg', 'pkg.nspkg', 'pkg.subpkg'])
+
+
+class TestFlatLayoutPackageFinder:
+    EXAMPLES = {
+        "hidden-folders": (
+            [".pkg/__init__.py", "pkg/__init__.py", "pkg/nested/file.txt"],
+            ["pkg", "pkg.nested"],
+        ),
+        "private-packages": (
+            ["_pkg/__init__.py", "pkg/_private/__init__.py"],
+            ["pkg", "pkg._private"],
+        ),
+        "invalid-name": (
+            ["invalid-pkg/__init__.py", "other.pkg/__init__.py", "yet,another/file.py"],
+            [],
+        ),
+        "docs": (["pkg/__init__.py", "docs/conf.py", "docs/readme.rst"], ["pkg"]),
+        "tests": (
+            ["pkg/__init__.py", "tests/test_pkg.py", "tests/__init__.py"],
+            ["pkg"],
+        ),
+        "examples": (
+            [
+                "pkg/__init__.py",
+                "examples/__init__.py",
+                "examples/file.py",
+                "example/other_file.py",
+                # Sub-packages should always be fine
+                "pkg/example/__init__.py",
+                "pkg/examples/__init__.py",
+            ],
+            ["pkg", "pkg.examples", "pkg.example"],
+        ),
+        "tool-specific": (
+            [
+                "htmlcov/index.html",
+                "pkg/__init__.py",
+                "tasks/__init__.py",
+                "tasks/subpackage/__init__.py",
+                "fabfile/__init__.py",
+                "fabfile/subpackage/__init__.py",
+                # Sub-packages should always be fine
+                "pkg/tasks/__init__.py",
+                "pkg/fabfile/__init__.py",
+            ],
+            ["pkg", "pkg.tasks", "pkg.fabfile"],
+        ),
+    }
+
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_unwanted_directories_not_included(self, tmp_path, example):
+        files, expected_packages = self.EXAMPLES[example]
+        ensure_files(tmp_path, files)
+        found_packages = FlatLayoutPackageFinder.find(str(tmp_path))
+        assert set(found_packages) == set(expected_packages)
+
+
+def ensure_files(root_path, files):
+    for file in files:
+        path = root_path / file
+        path.parent.mkdir(parents=True, exist_ok=True)
+        path.touch()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_py_modules.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_py_modules.py
new file mode 100644
index 00000000..8034b544
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_find_py_modules.py
@@ -0,0 +1,73 @@
+"""Tests for automatic discovery of modules"""
+
+import os
+
+import pytest
+
+from setuptools.discovery import FlatLayoutModuleFinder, ModuleFinder
+
+from .compat.py39 import os_helper
+from .test_find_packages import ensure_files
+
+
+class TestModuleFinder:
+    def find(self, path, *args, **kwargs):
+        return set(ModuleFinder.find(str(path), *args, **kwargs))
+
+    EXAMPLES = {
+        # circumstance: (files, kwargs, expected_modules)
+        "simple_folder": (
+            ["file.py", "other.py"],
+            {},  # kwargs
+            ["file", "other"],
+        ),
+        "exclude": (
+            ["file.py", "other.py"],
+            {"exclude": ["f*"]},
+            ["other"],
+        ),
+        "include": (
+            ["file.py", "fole.py", "other.py"],
+            {"include": ["f*"], "exclude": ["fo*"]},
+            ["file"],
+        ),
+        "invalid-name": (["my-file.py", "other.file.py"], {}, []),
+    }
+
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_finder(self, tmp_path, example):
+        files, kwargs, expected_modules = self.EXAMPLES[example]
+        ensure_files(tmp_path, files)
+        assert self.find(tmp_path, **kwargs) == set(expected_modules)
+
+    @pytest.mark.skipif(not os_helper.can_symlink(), reason='Symlink support required')
+    def test_symlinked_packages_are_included(self, tmp_path):
+        src = "_myfiles/file.py"
+        ensure_files(tmp_path, [src])
+        os.symlink(tmp_path / src, tmp_path / "link.py")
+        assert self.find(tmp_path) == {"link"}
+
+
+class TestFlatLayoutModuleFinder:
+    def find(self, path, *args, **kwargs):
+        return set(FlatLayoutModuleFinder.find(str(path)))
+
+    EXAMPLES = {
+        # circumstance: (files, expected_modules)
+        "hidden-files": ([".module.py"], []),
+        "private-modules": (["_module.py"], []),
+        "common-names": (
+            ["setup.py", "conftest.py", "test.py", "tests.py", "example.py", "mod.py"],
+            ["mod"],
+        ),
+        "tool-specific": (
+            ["tasks.py", "fabfile.py", "noxfile.py", "dodo.py", "manage.py", "mod.py"],
+            ["mod"],
+        ),
+    }
+
+    @pytest.mark.parametrize("example", EXAMPLES.keys())
+    def test_unwanted_files_not_included(self, tmp_path, example):
+        files, expected_modules = self.EXAMPLES[example]
+        ensure_files(tmp_path, files)
+        assert self.find(tmp_path) == set(expected_modules)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_glob.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_glob.py
new file mode 100644
index 00000000..8d225a44
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_glob.py
@@ -0,0 +1,45 @@
+import pytest
+from jaraco import path
+
+from setuptools.glob import glob
+
+
+@pytest.mark.parametrize(
+    ('tree', 'pattern', 'matches'),
+    (
+        ('', b'', []),
+        ('', '', []),
+        (
+            """
+     appveyor.yml
+     CHANGES.rst
+     LICENSE
+     MANIFEST.in
+     pyproject.toml
+     README.rst
+     setup.cfg
+     setup.py
+     """,
+            '*.rst',
+            ('CHANGES.rst', 'README.rst'),
+        ),
+        (
+            """
+     appveyor.yml
+     CHANGES.rst
+     LICENSE
+     MANIFEST.in
+     pyproject.toml
+     README.rst
+     setup.cfg
+     setup.py
+     """,
+            b'*.rst',
+            (b'CHANGES.rst', b'README.rst'),
+        ),
+    ),
+)
+def test_glob(monkeypatch, tmpdir, tree, pattern, matches):
+    monkeypatch.chdir(tmpdir)
+    path.build({name: '' for name in tree.split()})
+    assert list(sorted(glob(pattern))) == list(sorted(matches))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_install_scripts.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_install_scripts.py
new file mode 100644
index 00000000..e62a6b7f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_install_scripts.py
@@ -0,0 +1,89 @@
+"""install_scripts tests"""
+
+import sys
+
+import pytest
+
+from setuptools.command.install_scripts import install_scripts
+from setuptools.dist import Distribution
+
+from . import contexts
+
+
+class TestInstallScripts:
+    settings = dict(
+        name='foo',
+        entry_points={'console_scripts': ['foo=foo:foo']},
+        version='0.0',
+    )
+    unix_exe = '/usr/dummy-test-path/local/bin/python'
+    unix_spaces_exe = '/usr/bin/env dummy-test-python'
+    win32_exe = 'C:\\Dummy Test Path\\Program Files\\Python 3.6\\python.exe'
+
+    def _run_install_scripts(self, install_dir, executable=None):
+        dist = Distribution(self.settings)
+        dist.script_name = 'setup.py'
+        cmd = install_scripts(dist)
+        cmd.install_dir = install_dir
+        if executable is not None:
+            bs = cmd.get_finalized_command('build_scripts')
+            bs.executable = executable
+        cmd.ensure_finalized()
+        with contexts.quiet():
+            cmd.run()
+
+    @pytest.mark.skipif(sys.platform == 'win32', reason='non-Windows only')
+    def test_sys_executable_escaping_unix(self, tmpdir, monkeypatch):
+        """
+        Ensure that shebang is not quoted on Unix when getting the Python exe
+        from sys.executable.
+        """
+        expected = f'#!{self.unix_exe}\n'
+        monkeypatch.setattr('sys.executable', self.unix_exe)
+        with tmpdir.as_cwd():
+            self._run_install_scripts(str(tmpdir))
+            with open(str(tmpdir.join('foo')), 'r', encoding="utf-8") as f:
+                actual = f.readline()
+        assert actual == expected
+
+    @pytest.mark.skipif(sys.platform != 'win32', reason='Windows only')
+    def test_sys_executable_escaping_win32(self, tmpdir, monkeypatch):
+        """
+        Ensure that shebang is quoted on Windows when getting the Python exe
+        from sys.executable and it contains a space.
+        """
+        expected = f'#!"{self.win32_exe}"\n'
+        monkeypatch.setattr('sys.executable', self.win32_exe)
+        with tmpdir.as_cwd():
+            self._run_install_scripts(str(tmpdir))
+            with open(str(tmpdir.join('foo-script.py')), 'r', encoding="utf-8") as f:
+                actual = f.readline()
+        assert actual == expected
+
+    @pytest.mark.skipif(sys.platform == 'win32', reason='non-Windows only')
+    def test_executable_with_spaces_escaping_unix(self, tmpdir):
+        """
+        Ensure that shebang on Unix is not quoted, even when
+        a value with spaces
+        is specified using --executable.
+        """
+        expected = f'#!{self.unix_spaces_exe}\n'
+        with tmpdir.as_cwd():
+            self._run_install_scripts(str(tmpdir), self.unix_spaces_exe)
+            with open(str(tmpdir.join('foo')), 'r', encoding="utf-8") as f:
+                actual = f.readline()
+        assert actual == expected
+
+    @pytest.mark.skipif(sys.platform != 'win32', reason='Windows only')
+    def test_executable_arg_escaping_win32(self, tmpdir):
+        """
+        Ensure that shebang on Windows is quoted when
+        getting a path with spaces
+        from --executable, that is itself properly quoted.
+        """
+        expected = f'#!"{self.win32_exe}"\n'
+        with tmpdir.as_cwd():
+            self._run_install_scripts(str(tmpdir), '"' + self.win32_exe + '"')
+            with open(str(tmpdir.join('foo-script.py')), 'r', encoding="utf-8") as f:
+                actual = f.readline()
+        assert actual == expected
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_logging.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_logging.py
new file mode 100644
index 00000000..ea58001e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_logging.py
@@ -0,0 +1,76 @@
+import functools
+import inspect
+import logging
+import sys
+
+import pytest
+
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+
+setup_py = """\
+from setuptools import setup
+
+setup(
+    name="test_logging",
+    version="0.0"
+)
+"""
+
+
+@pytest.mark.parametrize(
+    ('flag', 'expected_level'), [("--dry-run", "INFO"), ("--verbose", "DEBUG")]
+)
+def test_verbosity_level(tmp_path, monkeypatch, flag, expected_level):
+    """Make sure the correct verbosity level is set (issue #3038)"""
+    import setuptools  # noqa: F401  # import setuptools to monkeypatch distutils
+
+    import distutils  # <- load distutils after all the patches take place
+
+    logger = logging.Logger(__name__)
+    monkeypatch.setattr(logging, "root", logger)
+    unset_log_level = logger.getEffectiveLevel()
+    assert logging.getLevelName(unset_log_level) == "NOTSET"
+
+    setup_script = tmp_path / "setup.py"
+    setup_script.write_text(setup_py, encoding="utf-8")
+    dist = distutils.core.run_setup(setup_script, stop_after="init")
+    dist.script_args = [flag, "sdist"]
+    dist.parse_command_line()  # <- where the log level is set
+    log_level = logger.getEffectiveLevel()
+    log_level_name = logging.getLevelName(log_level)
+    assert log_level_name == expected_level
+
+
+def flaky_on_pypy(func):
+    @functools.wraps(func)
+    def _func():
+        try:
+            func()
+        except AssertionError:  # pragma: no cover
+            if IS_PYPY:
+                msg = "Flaky monkeypatch on PyPy (#4124)"
+                pytest.xfail(f"{msg}. Original discussion in #3707, #3709.")
+            raise
+
+    return _func
+
+
+@flaky_on_pypy
+def test_patching_does_not_cause_problems():
+    # Ensure `dist.log` is only patched if necessary
+
+    import _distutils_hack
+
+    import setuptools.logging
+
+    from distutils import dist
+
+    setuptools.logging.configure()
+
+    if _distutils_hack.enabled():
+        # Modern logging infra, no problematic patching.
+        assert dist.__file__ is None or "setuptools" in dist.__file__
+        assert isinstance(dist.log, logging.Logger)
+    else:
+        assert inspect.ismodule(dist.log)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_manifest.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_manifest.py
new file mode 100644
index 00000000..903a528d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_manifest.py
@@ -0,0 +1,622 @@
+"""sdist tests"""
+
+from __future__ import annotations
+
+import contextlib
+import io
+import itertools
+import logging
+import os
+import shutil
+import sys
+import tempfile
+
+import pytest
+
+from setuptools.command.egg_info import FileList, egg_info, translate_pattern
+from setuptools.dist import Distribution
+from setuptools.tests.textwrap import DALS
+
+from distutils import log
+from distutils.errors import DistutilsTemplateError
+
+IS_PYPY = '__pypy__' in sys.builtin_module_names
+
+
+def make_local_path(s):
+    """Converts '/' in a string to os.sep"""
+    return s.replace('/', os.sep)
+
+
+SETUP_ATTRS = {
+    'name': 'app',
+    'version': '0.0',
+    'packages': ['app'],
+}
+
+SETUP_PY = f"""\
+from setuptools import setup
+
+setup(**{SETUP_ATTRS!r})
+"""
+
+
+@contextlib.contextmanager
+def quiet():
+    old_stdout, old_stderr = sys.stdout, sys.stderr
+    sys.stdout, sys.stderr = io.StringIO(), io.StringIO()
+    try:
+        yield
+    finally:
+        sys.stdout, sys.stderr = old_stdout, old_stderr
+
+
+def touch(filename):
+    open(filename, 'wb').close()
+
+
+# The set of files always in the manifest, including all files in the
+# .egg-info directory
+default_files = frozenset(
+    map(
+        make_local_path,
+        [
+            'README.rst',
+            'MANIFEST.in',
+            'setup.py',
+            'app.egg-info/PKG-INFO',
+            'app.egg-info/SOURCES.txt',
+            'app.egg-info/dependency_links.txt',
+            'app.egg-info/top_level.txt',
+            'app/__init__.py',
+        ],
+    )
+)
+
+
+translate_specs: list[tuple[str, list[str], list[str]]] = [
+    ('foo', ['foo'], ['bar', 'foobar']),
+    ('foo/bar', ['foo/bar'], ['foo/bar/baz', './foo/bar', 'foo']),
+    # Glob matching
+    ('*.txt', ['foo.txt', 'bar.txt'], ['foo/foo.txt']),
+    ('dir/*.txt', ['dir/foo.txt', 'dir/bar.txt', 'dir/.txt'], ['notdir/foo.txt']),
+    ('*/*.py', ['bin/start.py'], []),
+    ('docs/page-?.txt', ['docs/page-9.txt'], ['docs/page-10.txt']),
+    # Globstars change what they mean depending upon where they are
+    (
+        'foo/**/bar',
+        ['foo/bing/bar', 'foo/bing/bang/bar', 'foo/bar'],
+        ['foo/abar'],
+    ),
+    (
+        'foo/**',
+        ['foo/bar/bing.py', 'foo/x'],
+        ['/foo/x'],
+    ),
+    (
+        '**',
+        ['x', 'abc/xyz', '@nything'],
+        [],
+    ),
+    # Character classes
+    (
+        'pre[one]post',
+        ['preopost', 'prenpost', 'preepost'],
+        ['prepost', 'preonepost'],
+    ),
+    (
+        'hello[!one]world',
+        ['helloxworld', 'helloyworld'],
+        ['hellooworld', 'helloworld', 'hellooneworld'],
+    ),
+    (
+        '[]one].txt',
+        ['o.txt', '].txt', 'e.txt'],
+        ['one].txt'],
+    ),
+    (
+        'foo[!]one]bar',
+        ['fooybar'],
+        ['foo]bar', 'fooobar', 'fooebar'],
+    ),
+]
+"""
+A spec of inputs for 'translate_pattern' and matches and mismatches
+for that input.
+"""
+
+match_params = itertools.chain.from_iterable(
+    zip(itertools.repeat(pattern), matches)
+    for pattern, matches, mismatches in translate_specs
+)
+
+
+@pytest.fixture(params=match_params)
+def pattern_match(request):
+    return map(make_local_path, request.param)
+
+
+mismatch_params = itertools.chain.from_iterable(
+    zip(itertools.repeat(pattern), mismatches)
+    for pattern, matches, mismatches in translate_specs
+)
+
+
+@pytest.fixture(params=mismatch_params)
+def pattern_mismatch(request):
+    return map(make_local_path, request.param)
+
+
+def test_translated_pattern_match(pattern_match):
+    pattern, target = pattern_match
+    assert translate_pattern(pattern).match(target)
+
+
+def test_translated_pattern_mismatch(pattern_mismatch):
+    pattern, target = pattern_mismatch
+    assert not translate_pattern(pattern).match(target)
+
+
+class TempDirTestCase:
+    def setup_method(self, method):
+        self.temp_dir = tempfile.mkdtemp()
+        self.old_cwd = os.getcwd()
+        os.chdir(self.temp_dir)
+
+    def teardown_method(self, method):
+        os.chdir(self.old_cwd)
+        shutil.rmtree(self.temp_dir)
+
+
+class TestManifestTest(TempDirTestCase):
+    def setup_method(self, method):
+        super().setup_method(method)
+
+        f = open(os.path.join(self.temp_dir, 'setup.py'), 'w', encoding="utf-8")
+        f.write(SETUP_PY)
+        f.close()
+        """
+        Create a file tree like:
+        - LICENSE
+        - README.rst
+        - testing.rst
+        - .hidden.rst
+        - app/
+            - __init__.py
+            - a.txt
+            - b.txt
+            - c.rst
+            - static/
+                - app.js
+                - app.js.map
+                - app.css
+                - app.css.map
+        """
+
+        for fname in ['README.rst', '.hidden.rst', 'testing.rst', 'LICENSE']:
+            touch(os.path.join(self.temp_dir, fname))
+
+        # Set up the rest of the test package
+        test_pkg = os.path.join(self.temp_dir, 'app')
+        os.mkdir(test_pkg)
+        for fname in ['__init__.py', 'a.txt', 'b.txt', 'c.rst']:
+            touch(os.path.join(test_pkg, fname))
+
+        # Some compiled front-end assets to include
+        static = os.path.join(test_pkg, 'static')
+        os.mkdir(static)
+        for fname in ['app.js', 'app.js.map', 'app.css', 'app.css.map']:
+            touch(os.path.join(static, fname))
+
+    def make_manifest(self, contents):
+        """Write a MANIFEST.in."""
+        manifest = os.path.join(self.temp_dir, 'MANIFEST.in')
+        with open(manifest, 'w', encoding="utf-8") as f:
+            f.write(DALS(contents))
+
+    def get_files(self):
+        """Run egg_info and get all the files to include, as a set"""
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = egg_info(dist)
+        cmd.ensure_finalized()
+
+        cmd.run()
+
+        return set(cmd.filelist.files)
+
+    def test_no_manifest(self):
+        """Check a missing MANIFEST.in includes only the standard files."""
+        assert (default_files - set(['MANIFEST.in'])) == self.get_files()
+
+    def test_empty_files(self):
+        """Check an empty MANIFEST.in includes only the standard files."""
+        self.make_manifest("")
+        assert default_files == self.get_files()
+
+    def test_include(self):
+        """Include extra rst files in the project root."""
+        self.make_manifest("include *.rst")
+        files = default_files | set(['testing.rst', '.hidden.rst'])
+        assert files == self.get_files()
+
+    def test_exclude(self):
+        """Include everything in app/ except the text files"""
+        ml = make_local_path
+        self.make_manifest(
+            """
+            include app/*
+            exclude app/*.txt
+            """
+        )
+        files = default_files | set([ml('app/c.rst')])
+        assert files == self.get_files()
+
+    def test_include_multiple(self):
+        """Include with multiple patterns."""
+        ml = make_local_path
+        self.make_manifest("include app/*.txt app/static/*")
+        files = default_files | set([
+            ml('app/a.txt'),
+            ml('app/b.txt'),
+            ml('app/static/app.js'),
+            ml('app/static/app.js.map'),
+            ml('app/static/app.css'),
+            ml('app/static/app.css.map'),
+        ])
+        assert files == self.get_files()
+
+    def test_graft(self):
+        """Include the whole app/static/ directory."""
+        ml = make_local_path
+        self.make_manifest("graft app/static")
+        files = default_files | set([
+            ml('app/static/app.js'),
+            ml('app/static/app.js.map'),
+            ml('app/static/app.css'),
+            ml('app/static/app.css.map'),
+        ])
+        assert files == self.get_files()
+
+    def test_graft_glob_syntax(self):
+        """Include the whole app/static/ directory."""
+        ml = make_local_path
+        self.make_manifest("graft */static")
+        files = default_files | set([
+            ml('app/static/app.js'),
+            ml('app/static/app.js.map'),
+            ml('app/static/app.css'),
+            ml('app/static/app.css.map'),
+        ])
+        assert files == self.get_files()
+
+    def test_graft_global_exclude(self):
+        """Exclude all *.map files in the project."""
+        ml = make_local_path
+        self.make_manifest(
+            """
+            graft app/static
+            global-exclude *.map
+            """
+        )
+        files = default_files | set([ml('app/static/app.js'), ml('app/static/app.css')])
+        assert files == self.get_files()
+
+    def test_global_include(self):
+        """Include all *.rst, *.js, and *.css files in the whole tree."""
+        ml = make_local_path
+        self.make_manifest(
+            """
+            global-include *.rst *.js *.css
+            """
+        )
+        files = default_files | set([
+            '.hidden.rst',
+            'testing.rst',
+            ml('app/c.rst'),
+            ml('app/static/app.js'),
+            ml('app/static/app.css'),
+        ])
+        assert files == self.get_files()
+
+    def test_graft_prune(self):
+        """Include all files in app/, except for the whole app/static/ dir."""
+        ml = make_local_path
+        self.make_manifest(
+            """
+            graft app
+            prune app/static
+            """
+        )
+        files = default_files | set([ml('app/a.txt'), ml('app/b.txt'), ml('app/c.rst')])
+        assert files == self.get_files()
+
+
+class TestFileListTest(TempDirTestCase):
+    """
+    A copy of the relevant bits of distutils/tests/test_filelist.py,
+    to ensure setuptools' version of FileList keeps parity with distutils.
+    """
+
+    @pytest.fixture(autouse=os.getenv("SETUPTOOLS_USE_DISTUTILS") == "stdlib")
+    def _compat_record_logs(self, monkeypatch, caplog):
+        """Account for stdlib compatibility"""
+
+        def _log(_logger, level, msg, args):
+            exc = sys.exc_info()
+            rec = logging.LogRecord("distutils", level, "", 0, msg, args, exc)
+            caplog.records.append(rec)
+
+        monkeypatch.setattr(log.Log, "_log", _log)
+
+    def get_records(self, caplog, *levels):
+        return [r for r in caplog.records if r.levelno in levels]
+
+    def assertNoWarnings(self, caplog):
+        assert self.get_records(caplog, log.WARN) == []
+        caplog.clear()
+
+    def assertWarnings(self, caplog):
+        if IS_PYPY and not caplog.records:
+            pytest.xfail("caplog checks may not work well in PyPy")
+        else:
+            assert len(self.get_records(caplog, log.WARN)) > 0
+            caplog.clear()
+
+    def make_files(self, files):
+        for file in files:
+            file = os.path.join(self.temp_dir, file)
+            dirname, _basename = os.path.split(file)
+            os.makedirs(dirname, exist_ok=True)
+            touch(file)
+
+    def test_process_template_line(self):
+        # testing  all MANIFEST.in template patterns
+        file_list = FileList()
+        ml = make_local_path
+
+        # simulated file list
+        self.make_files([
+            'foo.tmp',
+            'ok',
+            'xo',
+            'four.txt',
+            'buildout.cfg',
+            # filelist does not filter out VCS directories,
+            # it's sdist that does
+            ml('.hg/last-message.txt'),
+            ml('global/one.txt'),
+            ml('global/two.txt'),
+            ml('global/files.x'),
+            ml('global/here.tmp'),
+            ml('f/o/f.oo'),
+            ml('dir/graft-one'),
+            ml('dir/dir2/graft2'),
+            ml('dir3/ok'),
+            ml('dir3/sub/ok.txt'),
+        ])
+
+        MANIFEST_IN = DALS(
+            """\
+        include ok
+        include xo
+        exclude xo
+        include foo.tmp
+        include buildout.cfg
+        global-include *.x
+        global-include *.txt
+        global-exclude *.tmp
+        recursive-include f *.oo
+        recursive-exclude global *.x
+        graft dir
+        prune dir3
+        """
+        )
+
+        for line in MANIFEST_IN.split('\n'):
+            if not line:
+                continue
+            file_list.process_template_line(line)
+
+        wanted = [
+            'buildout.cfg',
+            'four.txt',
+            'ok',
+            ml('.hg/last-message.txt'),
+            ml('dir/graft-one'),
+            ml('dir/dir2/graft2'),
+            ml('f/o/f.oo'),
+            ml('global/one.txt'),
+            ml('global/two.txt'),
+        ]
+
+        file_list.sort()
+        assert file_list.files == wanted
+
+    def test_exclude_pattern(self):
+        # return False if no match
+        file_list = FileList()
+        assert not file_list.exclude_pattern('*.py')
+
+        # return True if files match
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.py']
+        assert file_list.exclude_pattern('*.py')
+
+        # test excludes
+        file_list = FileList()
+        file_list.files = ['a.py', 'a.txt']
+        file_list.exclude_pattern('*.py')
+        file_list.sort()
+        assert file_list.files == ['a.txt']
+
+    def test_include_pattern(self):
+        # return False if no match
+        file_list = FileList()
+        self.make_files([])
+        assert not file_list.include_pattern('*.py')
+
+        # return True if files match
+        file_list = FileList()
+        self.make_files(['a.py', 'b.txt'])
+        assert file_list.include_pattern('*.py')
+
+        # test * matches all files
+        file_list = FileList()
+        self.make_files(['a.py', 'b.txt'])
+        file_list.include_pattern('*')
+        file_list.sort()
+        assert file_list.files == ['a.py', 'b.txt']
+
+    def test_process_template_line_invalid(self):
+        # invalid lines
+        file_list = FileList()
+        for action in (
+            'include',
+            'exclude',
+            'global-include',
+            'global-exclude',
+            'recursive-include',
+            'recursive-exclude',
+            'graft',
+            'prune',
+            'blarg',
+        ):
+            with pytest.raises(DistutilsTemplateError):
+                file_list.process_template_line(action)
+
+    def test_include(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # include
+        file_list = FileList()
+        self.make_files(['a.py', 'b.txt', ml('d/c.py')])
+
+        file_list.process_template_line('include *.py')
+        file_list.sort()
+        assert file_list.files == ['a.py']
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('include *.rb')
+        file_list.sort()
+        assert file_list.files == ['a.py']
+        self.assertWarnings(caplog)
+
+    def test_exclude(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # exclude
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.txt', ml('d/c.py')]
+
+        file_list.process_template_line('exclude *.py')
+        file_list.sort()
+        assert file_list.files == ['b.txt', ml('d/c.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('exclude *.rb')
+        file_list.sort()
+        assert file_list.files == ['b.txt', ml('d/c.py')]
+        self.assertWarnings(caplog)
+
+    def test_global_include(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # global-include
+        file_list = FileList()
+        self.make_files(['a.py', 'b.txt', ml('d/c.py')])
+
+        file_list.process_template_line('global-include *.py')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('d/c.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('global-include *.rb')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('d/c.py')]
+        self.assertWarnings(caplog)
+
+    def test_global_exclude(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # global-exclude
+        file_list = FileList()
+        file_list.files = ['a.py', 'b.txt', ml('d/c.py')]
+
+        file_list.process_template_line('global-exclude *.py')
+        file_list.sort()
+        assert file_list.files == ['b.txt']
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('global-exclude *.rb')
+        file_list.sort()
+        assert file_list.files == ['b.txt']
+        self.assertWarnings(caplog)
+
+    def test_recursive_include(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # recursive-include
+        file_list = FileList()
+        self.make_files(['a.py', ml('d/b.py'), ml('d/c.txt'), ml('d/d/e.py')])
+
+        file_list.process_template_line('recursive-include d *.py')
+        file_list.sort()
+        assert file_list.files == [ml('d/b.py'), ml('d/d/e.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('recursive-include e *.py')
+        file_list.sort()
+        assert file_list.files == [ml('d/b.py'), ml('d/d/e.py')]
+        self.assertWarnings(caplog)
+
+    def test_recursive_exclude(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # recursive-exclude
+        file_list = FileList()
+        file_list.files = ['a.py', ml('d/b.py'), ml('d/c.txt'), ml('d/d/e.py')]
+
+        file_list.process_template_line('recursive-exclude d *.py')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('d/c.txt')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('recursive-exclude e *.py')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('d/c.txt')]
+        self.assertWarnings(caplog)
+
+    def test_graft(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # graft
+        file_list = FileList()
+        self.make_files(['a.py', ml('d/b.py'), ml('d/d/e.py'), ml('f/f.py')])
+
+        file_list.process_template_line('graft d')
+        file_list.sort()
+        assert file_list.files == [ml('d/b.py'), ml('d/d/e.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('graft e')
+        file_list.sort()
+        assert file_list.files == [ml('d/b.py'), ml('d/d/e.py')]
+        self.assertWarnings(caplog)
+
+    def test_prune(self, caplog):
+        caplog.set_level(logging.DEBUG)
+        ml = make_local_path
+        # prune
+        file_list = FileList()
+        file_list.files = ['a.py', ml('d/b.py'), ml('d/d/e.py'), ml('f/f.py')]
+
+        file_list.process_template_line('prune d')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('f/f.py')]
+        self.assertNoWarnings(caplog)
+
+        file_list.process_template_line('prune e')
+        file_list.sort()
+        assert file_list.files == ['a.py', ml('f/f.py')]
+        self.assertWarnings(caplog)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_namespaces.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_namespaces.py
new file mode 100644
index 00000000..a0f4120b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_namespaces.py
@@ -0,0 +1,138 @@
+import subprocess
+import sys
+
+from setuptools._path import paths_on_pythonpath
+
+from . import namespaces
+
+
+class TestNamespaces:
+    def test_mixed_site_and_non_site(self, tmpdir):
+        """
+        Installing two packages sharing the same namespace, one installed
+        to a site dir and the other installed just to a path on PYTHONPATH
+        should leave the namespace in tact and both packages reachable by
+        import.
+        """
+        pkg_A = namespaces.build_namespace_package(tmpdir, 'myns.pkgA')
+        pkg_B = namespaces.build_namespace_package(tmpdir, 'myns.pkgB')
+        site_packages = tmpdir / 'site-packages'
+        path_packages = tmpdir / 'path-packages'
+        targets = site_packages, path_packages
+        # use pip to install to the target directory
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip.__main__',
+            'install',
+            str(pkg_A),
+            '-t',
+            str(site_packages),
+        ]
+        subprocess.check_call(install_cmd)
+        namespaces.make_site_dir(site_packages)
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip.__main__',
+            'install',
+            str(pkg_B),
+            '-t',
+            str(path_packages),
+        ]
+        subprocess.check_call(install_cmd)
+        try_import = [
+            sys.executable,
+            '-c',
+            'import myns.pkgA; import myns.pkgB',
+        ]
+        with paths_on_pythonpath(map(str, targets)):
+            subprocess.check_call(try_import)
+
+    def test_pkg_resources_import(self, tmpdir):
+        """
+        Ensure that a namespace package doesn't break on import
+        of pkg_resources.
+        """
+        pkg = namespaces.build_namespace_package(tmpdir, 'myns.pkgA')
+        target = tmpdir / 'packages'
+        target.mkdir()
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip',
+            'install',
+            '-t',
+            str(target),
+            str(pkg),
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(install_cmd)
+        namespaces.make_site_dir(target)
+        try_import = [
+            sys.executable,
+            '-c',
+            'import pkg_resources',
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(try_import)
+
+    def test_namespace_package_installed_and_cwd(self, tmpdir):
+        """
+        Installing a namespace packages but also having it in the current
+        working directory, only one version should take precedence.
+        """
+        pkg_A = namespaces.build_namespace_package(tmpdir, 'myns.pkgA')
+        target = tmpdir / 'packages'
+        # use pip to install to the target directory
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip.__main__',
+            'install',
+            str(pkg_A),
+            '-t',
+            str(target),
+        ]
+        subprocess.check_call(install_cmd)
+        namespaces.make_site_dir(target)
+
+        # ensure that package imports and pkg_resources imports
+        pkg_resources_imp = [
+            sys.executable,
+            '-c',
+            'import pkg_resources; import myns.pkgA',
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(pkg_resources_imp, cwd=str(pkg_A))
+
+    def test_packages_in_the_same_namespace_installed_and_cwd(self, tmpdir):
+        """
+        Installing one namespace package and also have another in the same
+        namespace in the current working directory, both of them must be
+        importable.
+        """
+        pkg_A = namespaces.build_namespace_package(tmpdir, 'myns.pkgA')
+        pkg_B = namespaces.build_namespace_package(tmpdir, 'myns.pkgB')
+        target = tmpdir / 'packages'
+        # use pip to install to the target directory
+        install_cmd = [
+            sys.executable,
+            '-m',
+            'pip.__main__',
+            'install',
+            str(pkg_A),
+            '-t',
+            str(target),
+        ]
+        subprocess.check_call(install_cmd)
+        namespaces.make_site_dir(target)
+
+        # ensure that all packages import and pkg_resources imports
+        pkg_resources_imp = [
+            sys.executable,
+            '-c',
+            'import pkg_resources; import myns.pkgA; import myns.pkgB',
+        ]
+        with paths_on_pythonpath([str(target)]):
+            subprocess.check_call(pkg_resources_imp, cwd=str(pkg_B))
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_packageindex.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_packageindex.py
new file mode 100644
index 00000000..2a6e5917
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_packageindex.py
@@ -0,0 +1,267 @@
+import http.client
+import re
+import urllib.error
+import urllib.request
+from inspect import cleandoc
+
+import pytest
+
+import setuptools.package_index
+
+import distutils.errors
+
+
+class TestPackageIndex:
+    def test_regex(self):
+        hash_url = 'http://other_url?:action=show_md5&amp;'
+        hash_url += 'digest=0123456789abcdef0123456789abcdef'
+        doc = """
+            <a href="http://some_url">Name</a>
+            (<a title="MD5 hash"
+            href="{hash_url}">md5</a>)
+        """.lstrip().format(**locals())
+        assert setuptools.package_index.PYPI_MD5.match(doc)
+
+    def test_bad_url_bad_port(self):
+        index = setuptools.package_index.PackageIndex()
+        url = 'http://127.0.0.1:0/nonesuch/test_package_index'
+        with pytest.raises(Exception, match=re.escape(url)):
+            v = index.open_url(url)
+            assert isinstance(v, urllib.error.HTTPError)
+
+    def test_bad_url_typo(self):
+        # issue 16
+        # easy_install inquant.contentmirror.plone breaks because of a typo
+        # in its home URL
+        index = setuptools.package_index.PackageIndex(hosts=('www.example.com',))
+
+        url = 'url:%20https://svn.plone.org/svn/collective/inquant.contentmirror.plone/trunk'
+
+        with pytest.raises(Exception, match=re.escape(url)):
+            v = index.open_url(url)
+            assert isinstance(v, urllib.error.HTTPError)
+
+    def test_bad_url_bad_status_line(self):
+        index = setuptools.package_index.PackageIndex(hosts=('www.example.com',))
+
+        def _urlopen(*args):
+            raise http.client.BadStatusLine('line')
+
+        index.opener = _urlopen
+        url = 'http://example.com'
+        with pytest.raises(Exception, match=r'line'):
+            index.open_url(url)
+
+    def test_bad_url_double_scheme(self):
+        """
+        A bad URL with a double scheme should raise a DistutilsError.
+        """
+        index = setuptools.package_index.PackageIndex(hosts=('www.example.com',))
+
+        # issue 20
+        url = 'http://http://svn.pythonpaste.org/Paste/wphp/trunk'
+        try:
+            index.open_url(url)
+        except distutils.errors.DistutilsError as error:
+            msg = str(error)
+            assert (
+                'nonnumeric port' in msg
+                or 'getaddrinfo failed' in msg
+                or 'Name or service not known' in msg
+            )
+            return
+        raise RuntimeError("Did not raise")
+
+    def test_url_ok(self):
+        index = setuptools.package_index.PackageIndex(hosts=('www.example.com',))
+        url = 'file:///tmp/test_package_index'
+        assert index.url_ok(url, True)
+
+    def test_parse_bdist_wininst(self):
+        parse = setuptools.package_index.parse_bdist_wininst
+
+        actual = parse('reportlab-2.5.win32-py2.4.exe')
+        expected = 'reportlab-2.5', '2.4', 'win32'
+        assert actual == expected
+
+        actual = parse('reportlab-2.5.win32.exe')
+        expected = 'reportlab-2.5', None, 'win32'
+        assert actual == expected
+
+        actual = parse('reportlab-2.5.win-amd64-py2.7.exe')
+        expected = 'reportlab-2.5', '2.7', 'win-amd64'
+        assert actual == expected
+
+        actual = parse('reportlab-2.5.win-amd64.exe')
+        expected = 'reportlab-2.5', None, 'win-amd64'
+        assert actual == expected
+
+    def test__vcs_split_rev_from_url(self):
+        """
+        Test the basic usage of _vcs_split_rev_from_url
+        """
+        vsrfu = setuptools.package_index.PackageIndex._vcs_split_rev_from_url
+        url, rev = vsrfu('https://example.com/bar@2995')
+        assert url == 'https://example.com/bar'
+        assert rev == '2995'
+
+    def test_local_index(self, tmpdir):
+        """
+        local_open should be able to read an index from the file system.
+        """
+        index_file = tmpdir / 'index.html'
+        with index_file.open('w') as f:
+            f.write('<div>content</div>')
+        url = 'file:' + urllib.request.pathname2url(str(tmpdir)) + '/'
+        res = setuptools.package_index.local_open(url)
+        assert 'content' in res.read()
+
+    def test_egg_fragment(self):
+        """
+        EGG fragments must comply to PEP 440
+        """
+        epoch = [
+            '',
+            '1!',
+        ]
+        releases = [
+            '0',
+            '0.0',
+            '0.0.0',
+        ]
+        pre = [
+            'a0',
+            'b0',
+            'rc0',
+        ]
+        post = ['.post0']
+        dev = [
+            '.dev0',
+        ]
+        local = [
+            ('', ''),
+            ('+ubuntu.0', '+ubuntu.0'),
+            ('+ubuntu-0', '+ubuntu.0'),
+            ('+ubuntu_0', '+ubuntu.0'),
+        ]
+        versions = [
+            [''.join([e, r, p, loc]) for loc in locs]
+            for e in epoch
+            for r in releases
+            for p in sum([pre, post, dev], [''])
+            for locs in local
+        ]
+        for v, vc in versions:
+            dists = list(
+                setuptools.package_index.distros_for_url(
+                    'http://example.com/example-foo.zip#egg=example-foo-' + v
+                )
+            )
+            assert dists[0].version == ''
+            assert dists[1].version == vc
+
+    def test_download_git_with_rev(self, tmp_path, fp):
+        url = 'git+https://github.example/group/project@master#egg=foo'
+        index = setuptools.package_index.PackageIndex()
+
+        expected_dir = tmp_path / 'project@master'
+        fp.register([
+            'git',
+            'clone',
+            '--quiet',
+            'https://github.example/group/project',
+            expected_dir,
+        ])
+        fp.register(['git', '-C', expected_dir, 'checkout', '--quiet', 'master'])
+
+        result = index.download(url, tmp_path)
+
+        assert result == str(expected_dir)
+        assert len(fp.calls) == 2
+
+    def test_download_git_no_rev(self, tmp_path, fp):
+        url = 'git+https://github.example/group/project#egg=foo'
+        index = setuptools.package_index.PackageIndex()
+
+        expected_dir = tmp_path / 'project'
+        fp.register([
+            'git',
+            'clone',
+            '--quiet',
+            'https://github.example/group/project',
+            expected_dir,
+        ])
+        index.download(url, tmp_path)
+
+    def test_download_svn(self, tmp_path):
+        url = 'svn+https://svn.example/project#egg=foo'
+        index = setuptools.package_index.PackageIndex()
+
+        msg = r".*SVN download is not supported.*"
+        with pytest.raises(distutils.errors.DistutilsError, match=msg):
+            index.download(url, tmp_path)
+
+
+class TestContentCheckers:
+    def test_md5(self):
+        checker = setuptools.package_index.HashChecker.from_url(
+            'http://foo/bar#md5=f12895fdffbd45007040d2e44df98478'
+        )
+        checker.feed('You should probably not be using MD5'.encode('ascii'))
+        assert checker.hash.hexdigest() == 'f12895fdffbd45007040d2e44df98478'
+        assert checker.is_valid()
+
+    def test_other_fragment(self):
+        "Content checks should succeed silently if no hash is present"
+        checker = setuptools.package_index.HashChecker.from_url(
+            'http://foo/bar#something%20completely%20different'
+        )
+        checker.feed('anything'.encode('ascii'))
+        assert checker.is_valid()
+
+    def test_blank_md5(self):
+        "Content checks should succeed if a hash is empty"
+        checker = setuptools.package_index.HashChecker.from_url('http://foo/bar#md5=')
+        checker.feed('anything'.encode('ascii'))
+        assert checker.is_valid()
+
+    def test_get_hash_name_md5(self):
+        checker = setuptools.package_index.HashChecker.from_url(
+            'http://foo/bar#md5=f12895fdffbd45007040d2e44df98478'
+        )
+        assert checker.hash_name == 'md5'
+
+    def test_report(self):
+        checker = setuptools.package_index.HashChecker.from_url(
+            'http://foo/bar#md5=f12895fdffbd45007040d2e44df98478'
+        )
+        rep = checker.report(lambda x: x, 'My message about %s')
+        assert rep == 'My message about md5'
+
+
+class TestPyPIConfig:
+    def test_percent_in_password(self, tmp_home_dir):
+        pypirc = tmp_home_dir / '.pypirc'
+        pypirc.write_text(
+            cleandoc(
+                """
+                [pypi]
+                repository=https://pypi.org
+                username=jaraco
+                password=pity%
+                """
+            ),
+            encoding="utf-8",
+        )
+        cfg = setuptools.package_index.PyPIConfig()
+        cred = cfg.creds_by_repository['https://pypi.org']
+        assert cred.username == 'jaraco'
+        assert cred.password == 'pity%'
+
+
+@pytest.mark.timeout(1)
+def test_REL_DoS():
+    """
+    REL should not hang on a contrived attack string.
+    """
+    setuptools.package_index.REL.search('< rel=' + ' ' * 2**12)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sandbox.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sandbox.py
new file mode 100644
index 00000000..a476b7c9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sandbox.py
@@ -0,0 +1,134 @@
+"""develop tests"""
+
+import os
+import types
+
+import pytest
+
+import pkg_resources
+import setuptools.sandbox
+
+
+class TestSandbox:
+    def test_devnull(self, tmpdir):
+        with setuptools.sandbox.DirectorySandbox(str(tmpdir)):
+            self._file_writer(os.devnull)
+
+    @staticmethod
+    def _file_writer(path):
+        def do_write():
+            with open(path, 'w', encoding="utf-8") as f:
+                f.write('xxx')
+
+        return do_write
+
+    def test_setup_py_with_BOM(self):
+        """
+        It should be possible to execute a setup.py with a Byte Order Mark
+        """
+        target = pkg_resources.resource_filename(__name__, 'script-with-bom.py')
+        namespace = types.ModuleType('namespace')
+        setuptools.sandbox._execfile(target, vars(namespace))
+        assert namespace.result == 'passed'
+
+    def test_setup_py_with_CRLF(self, tmpdir):
+        setup_py = tmpdir / 'setup.py'
+        with setup_py.open('wb') as stream:
+            stream.write(b'"degenerate script"\r\n')
+        setuptools.sandbox._execfile(str(setup_py), globals())
+
+
+class TestExceptionSaver:
+    def test_exception_trapped(self):
+        with setuptools.sandbox.ExceptionSaver():
+            raise ValueError("details")
+
+    def test_exception_resumed(self):
+        with setuptools.sandbox.ExceptionSaver() as saved_exc:
+            raise ValueError("details")
+
+        with pytest.raises(ValueError) as caught:
+            saved_exc.resume()
+
+        assert isinstance(caught.value, ValueError)
+        assert str(caught.value) == 'details'
+
+    def test_exception_reconstructed(self):
+        orig_exc = ValueError("details")
+
+        with setuptools.sandbox.ExceptionSaver() as saved_exc:
+            raise orig_exc
+
+        with pytest.raises(ValueError) as caught:
+            saved_exc.resume()
+
+        assert isinstance(caught.value, ValueError)
+        assert caught.value is not orig_exc
+
+    def test_no_exception_passes_quietly(self):
+        with setuptools.sandbox.ExceptionSaver() as saved_exc:
+            pass
+
+        saved_exc.resume()
+
+    def test_unpickleable_exception(self):
+        class CantPickleThis(Exception):
+            "This Exception is unpickleable because it's not in globals"
+
+            def __repr__(self) -> str:
+                return f'CantPickleThis{self.args!r}'
+
+        with setuptools.sandbox.ExceptionSaver() as saved_exc:
+            raise CantPickleThis('detail')
+
+        with pytest.raises(setuptools.sandbox.UnpickleableException) as caught:
+            saved_exc.resume()
+
+        assert str(caught.value) == "CantPickleThis('detail',)"
+
+    def test_unpickleable_exception_when_hiding_setuptools(self):
+        """
+        As revealed in #440, an infinite recursion can occur if an unpickleable
+        exception while setuptools is hidden. Ensure this doesn't happen.
+        """
+
+        class ExceptionUnderTest(Exception):
+            """
+            An unpickleable exception (not in globals).
+            """
+
+        with pytest.raises(setuptools.sandbox.UnpickleableException) as caught:
+            with setuptools.sandbox.save_modules():
+                setuptools.sandbox.hide_setuptools()
+                raise ExceptionUnderTest
+
+        (msg,) = caught.value.args
+        assert msg == 'ExceptionUnderTest()'
+
+    def test_sandbox_violation_raised_hiding_setuptools(self, tmpdir):
+        """
+        When in a sandbox with setuptools hidden, a SandboxViolation
+        should reflect a proper exception and not be wrapped in
+        an UnpickleableException.
+        """
+
+        def write_file():
+            "Trigger a SandboxViolation by writing outside the sandbox"
+            with open('/etc/foo', 'w', encoding="utf-8"):
+                pass
+
+        with pytest.raises(setuptools.sandbox.SandboxViolation) as caught:
+            with setuptools.sandbox.save_modules():
+                setuptools.sandbox.hide_setuptools()
+                with setuptools.sandbox.DirectorySandbox(str(tmpdir)):
+                    write_file()
+
+        cmd, args, kwargs = caught.value.args
+        assert cmd == 'open'
+        assert args == ('/etc/foo', 'w')
+        assert kwargs == {"encoding": "utf-8"}
+
+        msg = str(caught.value)
+        assert 'open' in msg
+        assert "('/etc/foo', 'w')" in msg
+        assert "{'encoding': 'utf-8'}" in msg
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sdist.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sdist.py
new file mode 100644
index 00000000..3ee0511b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_sdist.py
@@ -0,0 +1,972 @@
+"""sdist tests"""
+
+import contextlib
+import io
+import logging
+import os
+import pathlib
+import sys
+import tarfile
+import tempfile
+import unicodedata
+from inspect import cleandoc
+from pathlib import Path
+from unittest import mock
+
+import jaraco.path
+import pytest
+
+from setuptools import Command, SetuptoolsDeprecationWarning
+from setuptools._importlib import metadata
+from setuptools.command.egg_info import manifest_maker
+from setuptools.command.sdist import sdist
+from setuptools.dist import Distribution
+from setuptools.extension import Extension
+from setuptools.tests import fail_on_ascii
+
+from .text import Filenames
+
+import distutils
+from distutils.core import run_setup
+
+SETUP_ATTRS = {
+    'name': 'sdist_test',
+    'version': '0.0',
+    'packages': ['sdist_test'],
+    'package_data': {'sdist_test': ['*.txt']},
+    'data_files': [("data", [os.path.join("d", "e.dat")])],
+}
+
+SETUP_PY = f"""\
+from setuptools import setup
+
+setup(**{SETUP_ATTRS!r})
+"""
+
+EXTENSION = Extension(
+    name="sdist_test.f",
+    sources=[os.path.join("sdist_test", "f.c")],
+    depends=[os.path.join("sdist_test", "f.h")],
+)
+EXTENSION_SOURCES = EXTENSION.sources + EXTENSION.depends
+
+
+@contextlib.contextmanager
+def quiet():
+    old_stdout, old_stderr = sys.stdout, sys.stderr
+    sys.stdout, sys.stderr = io.StringIO(), io.StringIO()
+    try:
+        yield
+    finally:
+        sys.stdout, sys.stderr = old_stdout, old_stderr
+
+
+# Convert to POSIX path
+def posix(path):
+    if not isinstance(path, str):
+        return path.replace(os.sep.encode('ascii'), b'/')
+    else:
+        return path.replace(os.sep, '/')
+
+
+# HFS Plus uses decomposed UTF-8
+def decompose(path):
+    if isinstance(path, str):
+        return unicodedata.normalize('NFD', path)
+    try:
+        path = path.decode('utf-8')
+        path = unicodedata.normalize('NFD', path)
+        path = path.encode('utf-8')
+    except UnicodeError:
+        pass  # Not UTF-8
+    return path
+
+
+def read_all_bytes(filename):
+    with open(filename, 'rb') as fp:
+        return fp.read()
+
+
+def latin1_fail():
+    try:
+        desc, filename = tempfile.mkstemp(suffix=Filenames.latin_1)
+        os.close(desc)
+        os.remove(filename)
+    except Exception:
+        return True
+
+
+fail_on_latin1_encoded_filenames = pytest.mark.xfail(
+    latin1_fail(),
+    reason="System does not support latin-1 filenames",
+)
+
+
+skip_under_xdist = pytest.mark.skipif(
+    "os.environ.get('PYTEST_XDIST_WORKER')",
+    reason="pytest-dev/pytest-xdist#843",
+)
+skip_under_stdlib_distutils = pytest.mark.skipif(
+    not distutils.__package__.startswith('setuptools'),
+    reason="the test is not supported with stdlib distutils",
+)
+
+
+def touch(path):
+    open(path, 'wb').close()
+    return path
+
+
+def symlink_or_skip_test(src, dst):
+    try:
+        os.symlink(src, dst)
+    except (OSError, NotImplementedError):
+        pytest.skip("symlink not supported in OS")
+        return None
+    return dst
+
+
+class TestSdistTest:
+    @pytest.fixture(autouse=True)
+    def source_dir(self, tmpdir):
+        tmpdir = tmpdir / "project_root"
+        tmpdir.mkdir()
+
+        (tmpdir / 'setup.py').write_text(SETUP_PY, encoding='utf-8')
+
+        # Set up the rest of the test package
+        test_pkg = tmpdir / 'sdist_test'
+        test_pkg.mkdir()
+        data_folder = tmpdir / 'd'
+        data_folder.mkdir()
+        # *.rst was not included in package_data, so c.rst should not be
+        # automatically added to the manifest when not under version control
+        for fname in ['__init__.py', 'a.txt', 'b.txt', 'c.rst']:
+            touch(test_pkg / fname)
+        touch(data_folder / 'e.dat')
+        # C sources are not included by default, but they will be,
+        # if an extension module uses them as sources or depends
+        for fname in EXTENSION_SOURCES:
+            touch(tmpdir / fname)
+
+        with tmpdir.as_cwd():
+            yield tmpdir
+
+    def assert_package_data_in_manifest(self, cmd):
+        manifest = cmd.filelist.files
+        assert os.path.join('sdist_test', 'a.txt') in manifest
+        assert os.path.join('sdist_test', 'b.txt') in manifest
+        assert os.path.join('sdist_test', 'c.rst') not in manifest
+        assert os.path.join('d', 'e.dat') in manifest
+
+    def setup_with_extension(self):
+        setup_attrs = {**SETUP_ATTRS, 'ext_modules': [EXTENSION]}
+
+        dist = Distribution(setup_attrs)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        return cmd
+
+    def test_package_data_in_sdist(self):
+        """Regression test for pull request #4: ensures that files listed in
+        package_data are included in the manifest even if they're not added to
+        version control.
+        """
+
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        self.assert_package_data_in_manifest(cmd)
+
+    def test_package_data_and_include_package_data_in_sdist(self):
+        """
+        Ensure package_data and include_package_data work
+        together.
+        """
+        setup_attrs = {**SETUP_ATTRS, 'include_package_data': True}
+        assert setup_attrs['package_data']
+
+        dist = Distribution(setup_attrs)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        self.assert_package_data_in_manifest(cmd)
+
+    def test_extension_sources_in_sdist(self):
+        """
+        Ensure that the files listed in Extension.sources and Extension.depends
+        are automatically included in the manifest.
+        """
+        cmd = self.setup_with_extension()
+        self.assert_package_data_in_manifest(cmd)
+        manifest = cmd.filelist.files
+        for path in EXTENSION_SOURCES:
+            assert path in manifest
+
+    def test_missing_extension_sources(self):
+        """
+        Similar to test_extension_sources_in_sdist but the referenced files don't exist.
+        Missing files should not be included in distribution (with no error raised).
+        """
+        for path in EXTENSION_SOURCES:
+            os.remove(path)
+
+        cmd = self.setup_with_extension()
+        self.assert_package_data_in_manifest(cmd)
+        manifest = cmd.filelist.files
+        for path in EXTENSION_SOURCES:
+            assert path not in manifest
+
+    def test_symlinked_extension_sources(self):
+        """
+        Similar to test_extension_sources_in_sdist but the referenced files are
+        instead symbolic links to project-local files. Referenced file paths
+        should be included. Symlink targets themselves should NOT be included.
+        """
+        symlinked = []
+        for path in EXTENSION_SOURCES:
+            base, ext = os.path.splitext(path)
+            target = base + "_target." + ext
+
+            os.rename(path, target)
+            symlink_or_skip_test(os.path.basename(target), path)
+            symlinked.append(target)
+
+        cmd = self.setup_with_extension()
+        self.assert_package_data_in_manifest(cmd)
+        manifest = cmd.filelist.files
+        for path in EXTENSION_SOURCES:
+            assert path in manifest
+        for path in symlinked:
+            assert path not in manifest
+
+    _INVALID_PATHS = {
+        "must be relative": lambda: (
+            os.path.abspath(os.path.join("sdist_test", "f.h"))
+        ),
+        "can't have `..` segments": lambda: (
+            os.path.join("sdist_test", "..", "sdist_test", "f.h")
+        ),
+        "doesn't exist": lambda: (
+            os.path.join("sdist_test", "this_file_does_not_exist.h")
+        ),
+        "must be inside the project root": lambda: (
+            symlink_or_skip_test(
+                touch(os.path.join("..", "outside_of_project_root.h")),
+                "symlink.h",
+            )
+        ),
+    }
+
+    @skip_under_stdlib_distutils
+    @pytest.mark.parametrize("reason", _INVALID_PATHS.keys())
+    def test_invalid_extension_depends(self, reason, caplog):
+        """
+        Due to backwards compatibility reasons, `Extension.depends` should accept
+        invalid/weird paths, but then ignore them when building a sdist.
+
+        This test verifies that the source distribution is still built
+        successfully with such paths, but that instead of adding these paths to
+        the manifest, we emit an informational message, notifying the user that
+        the invalid path won't be automatically included.
+        """
+        invalid_path = self._INVALID_PATHS[reason]()
+        extension = Extension(
+            name="sdist_test.f",
+            sources=[],
+            depends=[invalid_path],
+        )
+        setup_attrs = {**SETUP_ATTRS, 'ext_modules': [extension]}
+
+        dist = Distribution(setup_attrs)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet(), caplog.at_level(logging.INFO):
+            cmd.run()
+
+        self.assert_package_data_in_manifest(cmd)
+        manifest = cmd.filelist.files
+        assert invalid_path not in manifest
+
+        expected_message = [
+            message
+            for (logger, level, message) in caplog.record_tuples
+            if (
+                logger == "root"  #
+                and level == logging.INFO  #
+                and invalid_path in message  #
+            )
+        ]
+        assert len(expected_message) == 1
+        (expected_message,) = expected_message
+        assert reason in expected_message
+
+    def test_custom_build_py(self):
+        """
+        Ensure projects defining custom build_py don't break
+        when creating sdists (issue #2849)
+        """
+        from distutils.command.build_py import build_py as OrigBuildPy
+
+        using_custom_command_guard = mock.Mock()
+
+        class CustomBuildPy(OrigBuildPy):
+            """
+            Some projects have custom commands inheriting from `distutils`
+            """
+
+            def get_data_files(self):
+                using_custom_command_guard()
+                return super().get_data_files()
+
+        setup_attrs = {**SETUP_ATTRS, 'include_package_data': True}
+        assert setup_attrs['package_data']
+
+        dist = Distribution(setup_attrs)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        # Make sure we use the custom command
+        cmd.cmdclass = {'build_py': CustomBuildPy}
+        cmd.distribution.cmdclass = {'build_py': CustomBuildPy}
+        assert cmd.distribution.get_command_class('build_py') == CustomBuildPy
+
+        msg = "setuptools instead of distutils"
+        with quiet(), pytest.warns(SetuptoolsDeprecationWarning, match=msg):
+            cmd.run()
+
+        using_custom_command_guard.assert_called()
+        self.assert_package_data_in_manifest(cmd)
+
+    def test_setup_py_exists(self):
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'foo.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        manifest = cmd.filelist.files
+        assert 'setup.py' in manifest
+
+    def test_setup_py_missing(self):
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'foo.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        if os.path.exists("setup.py"):
+            os.remove("setup.py")
+        with quiet():
+            cmd.run()
+
+        manifest = cmd.filelist.files
+        assert 'setup.py' not in manifest
+
+    def test_setup_py_excluded(self):
+        with open("MANIFEST.in", "w", encoding="utf-8") as manifest_file:
+            manifest_file.write("exclude setup.py")
+
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'foo.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        manifest = cmd.filelist.files
+        assert 'setup.py' not in manifest
+
+    def test_defaults_case_sensitivity(self, source_dir):
+        """
+        Make sure default files (README.*, etc.) are added in a case-sensitive
+        way to avoid problems with packages built on Windows.
+        """
+
+        touch(source_dir / 'readme.rst')
+        touch(source_dir / 'SETUP.cfg')
+
+        dist = Distribution(SETUP_ATTRS)
+        # the extension deliberately capitalized for this test
+        # to make sure the actual filename (not capitalized) gets added
+        # to the manifest
+        dist.script_name = 'setup.PY'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        with quiet():
+            cmd.run()
+
+        # lowercase all names so we can test in a
+        # case-insensitive way to make sure the files
+        # are not included.
+        manifest = map(lambda x: x.lower(), cmd.filelist.files)
+        assert 'readme.rst' not in manifest, manifest
+        assert 'setup.py' not in manifest, manifest
+        assert 'setup.cfg' not in manifest, manifest
+
+    def test_exclude_dev_only_cache_folders(self, source_dir):
+        included = {
+            # Emulate problem in https://github.com/pypa/setuptools/issues/4601
+            "MANIFEST.in": (
+                "global-include LICEN[CS]E* COPYING* NOTICE* AUTHORS*\n"
+                "global-include *.txt\n"
+            ),
+            # For the sake of being conservative and limiting unforeseen side-effects
+            # we just exclude dev-only cache folders at the root of the repository:
+            "test/.venv/lib/python3.9/site-packages/bar-2.dist-info/AUTHORS.rst": "",
+            "src/.nox/py/lib/python3.12/site-packages/bar-2.dist-info/COPYING.txt": "",
+            "doc/.tox/default/lib/python3.11/site-packages/foo-4.dist-info/LICENSE": "",
+            # Let's test against false positives with similarly named files:
+            ".venv-requirements.txt": "",
+            ".tox-coveragerc.txt": "",
+            ".noxy/coveragerc.txt": "",
+        }
+
+        excluded = {
+            # .tox/.nox/.venv are well-know folders present at the root of Python repos
+            # and therefore should be excluded
+            ".tox/release/lib/python3.11/site-packages/foo-4.dist-info/LICENSE": "",
+            ".nox/py/lib/python3.12/site-packages/bar-2.dist-info/COPYING.txt": "",
+            ".venv/lib/python3.9/site-packages/bar-2.dist-info/AUTHORS.rst": "",
+        }
+
+        for file, content in {**excluded, **included}.items():
+            Path(source_dir, file).parent.mkdir(parents=True, exist_ok=True)
+            Path(source_dir, file).write_text(content, encoding="utf-8")
+
+        cmd = self.setup_with_extension()
+        self.assert_package_data_in_manifest(cmd)
+        manifest = {f.replace(os.sep, '/') for f in cmd.filelist.files}
+        for path in excluded:
+            assert os.path.exists(path)
+            assert path not in manifest, (path, manifest)
+        for path in included:
+            assert os.path.exists(path)
+            assert path in manifest, (path, manifest)
+
+    @fail_on_ascii
+    def test_manifest_is_written_with_utf8_encoding(self):
+        # Test for #303.
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        mm = manifest_maker(dist)
+        mm.manifest = os.path.join('sdist_test.egg-info', 'SOURCES.txt')
+        os.mkdir('sdist_test.egg-info')
+
+        # UTF-8 filename
+        filename = os.path.join('sdist_test', 'smörbröd.py')
+
+        # Must create the file or it will get stripped.
+        touch(filename)
+
+        # Add UTF-8 filename and write manifest
+        with quiet():
+            mm.run()
+            mm.filelist.append(filename)
+            mm.write_manifest()
+
+        contents = read_all_bytes(mm.manifest)
+
+        # The manifest should be UTF-8 encoded
+        u_contents = contents.decode('UTF-8')
+
+        # The manifest should contain the UTF-8 filename
+        assert posix(filename) in u_contents
+
+    @fail_on_ascii
+    def test_write_manifest_allows_utf8_filenames(self):
+        # Test for #303.
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        mm = manifest_maker(dist)
+        mm.manifest = os.path.join('sdist_test.egg-info', 'SOURCES.txt')
+        os.mkdir('sdist_test.egg-info')
+
+        filename = os.path.join(b'sdist_test', Filenames.utf_8)
+
+        # Must touch the file or risk removal
+        touch(filename)
+
+        # Add filename and write manifest
+        with quiet():
+            mm.run()
+            u_filename = filename.decode('utf-8')
+            mm.filelist.files.append(u_filename)
+            # Re-write manifest
+            mm.write_manifest()
+
+        contents = read_all_bytes(mm.manifest)
+
+        # The manifest should be UTF-8 encoded
+        contents.decode('UTF-8')
+
+        # The manifest should contain the UTF-8 filename
+        assert posix(filename) in contents
+
+        # The filelist should have been updated as well
+        assert u_filename in mm.filelist.files
+
+    @skip_under_xdist
+    def test_write_manifest_skips_non_utf8_filenames(self):
+        """
+        Files that cannot be encoded to UTF-8 (specifically, those that
+        weren't originally successfully decoded and have surrogate
+        escapes) should be omitted from the manifest.
+        See https://bitbucket.org/tarek/distribute/issue/303 for history.
+        """
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        mm = manifest_maker(dist)
+        mm.manifest = os.path.join('sdist_test.egg-info', 'SOURCES.txt')
+        os.mkdir('sdist_test.egg-info')
+
+        # Latin-1 filename
+        filename = os.path.join(b'sdist_test', Filenames.latin_1)
+
+        # Add filename with surrogates and write manifest
+        with quiet():
+            mm.run()
+            u_filename = filename.decode('utf-8', 'surrogateescape')
+            mm.filelist.append(u_filename)
+            # Re-write manifest
+            mm.write_manifest()
+
+        contents = read_all_bytes(mm.manifest)
+
+        # The manifest should be UTF-8 encoded
+        contents.decode('UTF-8')
+
+        # The Latin-1 filename should have been skipped
+        assert posix(filename) not in contents
+
+        # The filelist should have been updated as well
+        assert u_filename not in mm.filelist.files
+
+    @fail_on_ascii
+    def test_manifest_is_read_with_utf8_encoding(self):
+        # Test for #303.
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        # Create manifest
+        with quiet():
+            cmd.run()
+
+        # Add UTF-8 filename to manifest
+        filename = os.path.join(b'sdist_test', Filenames.utf_8)
+        cmd.manifest = os.path.join('sdist_test.egg-info', 'SOURCES.txt')
+        manifest = open(cmd.manifest, 'ab')
+        manifest.write(b'\n' + filename)
+        manifest.close()
+
+        # The file must exist to be included in the filelist
+        touch(filename)
+
+        # Re-read manifest
+        cmd.filelist.files = []
+        with quiet():
+            cmd.read_manifest()
+
+        # The filelist should contain the UTF-8 filename
+        filename = filename.decode('utf-8')
+        assert filename in cmd.filelist.files
+
+    @fail_on_latin1_encoded_filenames
+    def test_read_manifest_skips_non_utf8_filenames(self):
+        # Test for #303.
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        # Create manifest
+        with quiet():
+            cmd.run()
+
+        # Add Latin-1 filename to manifest
+        filename = os.path.join(b'sdist_test', Filenames.latin_1)
+        cmd.manifest = os.path.join('sdist_test.egg-info', 'SOURCES.txt')
+        manifest = open(cmd.manifest, 'ab')
+        manifest.write(b'\n' + filename)
+        manifest.close()
+
+        # The file must exist to be included in the filelist
+        touch(filename)
+
+        # Re-read manifest
+        cmd.filelist.files = []
+        with quiet():
+            cmd.read_manifest()
+
+        # The Latin-1 filename should have been skipped
+        filename = filename.decode('latin-1')
+        assert filename not in cmd.filelist.files
+
+    @fail_on_ascii
+    @fail_on_latin1_encoded_filenames
+    def test_sdist_with_utf8_encoded_filename(self):
+        # Test for #303.
+        dist = Distribution(self.make_strings(SETUP_ATTRS))
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        filename = os.path.join(b'sdist_test', Filenames.utf_8)
+        touch(filename)
+
+        with quiet():
+            cmd.run()
+
+        if sys.platform == 'darwin':
+            filename = decompose(filename)
+
+        fs_enc = sys.getfilesystemencoding()
+
+        if sys.platform == 'win32':
+            if fs_enc == 'cp1252':
+                # Python mangles the UTF-8 filename
+                filename = filename.decode('cp1252')
+                assert filename in cmd.filelist.files
+            else:
+                filename = filename.decode('mbcs')
+                assert filename in cmd.filelist.files
+        else:
+            filename = filename.decode('utf-8')
+            assert filename in cmd.filelist.files
+
+    @classmethod
+    def make_strings(cls, item):
+        if isinstance(item, dict):
+            return {key: cls.make_strings(value) for key, value in item.items()}
+        if isinstance(item, list):
+            return list(map(cls.make_strings, item))
+        return str(item)
+
+    @fail_on_latin1_encoded_filenames
+    @skip_under_xdist
+    def test_sdist_with_latin1_encoded_filename(self):
+        # Test for #303.
+        dist = Distribution(self.make_strings(SETUP_ATTRS))
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+
+        # Latin-1 filename
+        filename = os.path.join(b'sdist_test', Filenames.latin_1)
+        touch(filename)
+        assert os.path.isfile(filename)
+
+        with quiet():
+            cmd.run()
+
+        # not all windows systems have a default FS encoding of cp1252
+        if sys.platform == 'win32':
+            # Latin-1 is similar to Windows-1252 however
+            # on mbcs filesys it is not in latin-1 encoding
+            fs_enc = sys.getfilesystemencoding()
+            if fs_enc != 'mbcs':
+                fs_enc = 'latin-1'
+            filename = filename.decode(fs_enc)
+
+            assert filename in cmd.filelist.files
+        else:
+            # The Latin-1 filename should have been skipped
+            filename = filename.decode('latin-1')
+            assert filename not in cmd.filelist.files
+
+    _EXAMPLE_DIRECTIVES = {
+        "setup.cfg - long_description and version": """
+            [metadata]
+            name = testing
+            version = file: src/VERSION.txt
+            license_files = DOWHATYOUWANT
+            long_description = file: README.rst, USAGE.rst
+            """,
+        "pyproject.toml - static readme/license files and dynamic version": """
+            [project]
+            name = "testing"
+            readme = "USAGE.rst"
+            license = {file = "DOWHATYOUWANT"}
+            dynamic = ["version"]
+            [tool.setuptools.dynamic]
+            version = {file = ["src/VERSION.txt"]}
+            """,
+        "pyproject.toml - directive with str instead of list": """
+            [project]
+            name = "testing"
+            readme = "USAGE.rst"
+            license = {file = "DOWHATYOUWANT"}
+            dynamic = ["version"]
+            [tool.setuptools.dynamic]
+            version = {file = "src/VERSION.txt"}
+            """,
+    }
+
+    @pytest.mark.parametrize("config", _EXAMPLE_DIRECTIVES.keys())
+    def test_add_files_referenced_by_config_directives(self, source_dir, config):
+        config_file, _, _ = config.partition(" - ")
+        config_text = self._EXAMPLE_DIRECTIVES[config]
+        (source_dir / 'src').mkdir()
+        (source_dir / 'src/VERSION.txt').write_text("0.42", encoding="utf-8")
+        (source_dir / 'README.rst').write_text("hello world!", encoding="utf-8")
+        (source_dir / 'USAGE.rst').write_text("hello world!", encoding="utf-8")
+        (source_dir / 'DOWHATYOUWANT').write_text("hello world!", encoding="utf-8")
+        (source_dir / config_file).write_text(config_text, encoding="utf-8")
+
+        dist = Distribution({"packages": []})
+        dist.script_name = 'setup.py'
+        dist.parse_config_files()
+
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+
+        assert (
+            'src/VERSION.txt' in cmd.filelist.files
+            or 'src\\VERSION.txt' in cmd.filelist.files
+        )
+        assert 'USAGE.rst' in cmd.filelist.files
+        assert 'DOWHATYOUWANT' in cmd.filelist.files
+        assert '/' not in cmd.filelist.files
+        assert '\\' not in cmd.filelist.files
+
+    def test_pyproject_toml_in_sdist(self, source_dir):
+        """
+        Check if pyproject.toml is included in source distribution if present
+        """
+        touch(source_dir / 'pyproject.toml')
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+        manifest = cmd.filelist.files
+        assert 'pyproject.toml' in manifest
+
+    def test_pyproject_toml_excluded(self, source_dir):
+        """
+        Check that pyproject.toml can excluded even if present
+        """
+        touch(source_dir / 'pyproject.toml')
+        with open('MANIFEST.in', 'w', encoding="utf-8") as mts:
+            print('exclude pyproject.toml', file=mts)
+        dist = Distribution(SETUP_ATTRS)
+        dist.script_name = 'setup.py'
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+        manifest = cmd.filelist.files
+        assert 'pyproject.toml' not in manifest
+
+    def test_build_subcommand_source_files(self, source_dir):
+        touch(source_dir / '.myfile~')
+
+        # Sanity check: without custom commands file list should not be affected
+        dist = Distribution({**SETUP_ATTRS, "script_name": "setup.py"})
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+        manifest = cmd.filelist.files
+        assert '.myfile~' not in manifest
+
+        # Test: custom command should be able to augment file list
+        dist = Distribution({**SETUP_ATTRS, "script_name": "setup.py"})
+        build = dist.get_command_obj("build")
+        build.sub_commands = [*build.sub_commands, ("build_custom", None)]
+
+        class build_custom(Command):
+            def initialize_options(self): ...
+
+            def finalize_options(self): ...
+
+            def run(self): ...
+
+            def get_source_files(self):
+                return ['.myfile~']
+
+        dist.cmdclass.update(build_custom=build_custom)
+
+        cmd = sdist(dist)
+        cmd.use_defaults = True
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+        manifest = cmd.filelist.files
+        assert '.myfile~' in manifest
+
+    @pytest.mark.skipif("os.environ.get('SETUPTOOLS_USE_DISTUTILS') == 'stdlib'")
+    def test_build_base_pathlib(self, source_dir):
+        """
+        Ensure if build_base is a pathlib.Path, the build still succeeds.
+        """
+        dist = Distribution({
+            **SETUP_ATTRS,
+            "script_name": "setup.py",
+            "options": {"build": {"build_base": pathlib.Path('build')}},
+        })
+        cmd = sdist(dist)
+        cmd.ensure_finalized()
+        with quiet():
+            cmd.run()
+
+
+def test_default_revctrl():
+    """
+    When _default_revctrl was removed from the `setuptools.command.sdist`
+    module in 10.0, it broke some systems which keep an old install of
+    setuptools (Distribute) around. Those old versions require that the
+    setuptools package continue to implement that interface, so this
+    function provides that interface, stubbed. See #320 for details.
+
+    This interface must be maintained until Ubuntu 12.04 is no longer
+    supported (by Setuptools).
+    """
+    (ep,) = metadata.EntryPoints._from_text(
+        """
+        [setuptools.file_finders]
+        svn_cvs = setuptools.command.sdist:_default_revctrl
+        """
+    )
+    res = ep.load()
+    assert hasattr(res, '__iter__')
+
+
+class TestRegressions:
+    """
+    Can be removed/changed if the project decides to change how it handles symlinks
+    or external files.
+    """
+
+    @staticmethod
+    def files_for_symlink_in_extension_depends(tmp_path, dep_path):
+        return {
+            "external": {
+                "dir": {"file.h": ""},
+            },
+            "project": {
+                "setup.py": cleandoc(
+                    f"""
+                    from setuptools import Extension, setup
+                    setup(
+                        name="myproj",
+                        version="42",
+                        ext_modules=[
+                            Extension(
+                                "hello", sources=["hello.pyx"],
+                                depends=[{dep_path!r}]
+                            )
+                        ],
+                    )
+                    """
+                ),
+                "hello.pyx": "",
+                "MANIFEST.in": "global-include *.h",
+            },
+        }
+
+    @pytest.mark.parametrize(
+        "dep_path", ("myheaders/dir/file.h", "myheaders/dir/../dir/file.h")
+    )
+    def test_symlink_in_extension_depends(self, monkeypatch, tmp_path, dep_path):
+        # Given a project with a symlinked dir and a "depends" targeting that dir
+        files = self.files_for_symlink_in_extension_depends(tmp_path, dep_path)
+        jaraco.path.build(files, prefix=str(tmp_path))
+        symlink_or_skip_test(tmp_path / "external", tmp_path / "project/myheaders")
+
+        # When `sdist` runs, there should be no error
+        members = run_sdist(monkeypatch, tmp_path / "project")
+        # and the sdist should contain the symlinked files
+        for expected in (
+            "myproj-42/hello.pyx",
+            "myproj-42/myheaders/dir/file.h",
+        ):
+            assert expected in members
+
+    @staticmethod
+    def files_for_external_path_in_extension_depends(tmp_path, dep_path):
+        head, _, tail = dep_path.partition("$tmp_path$/")
+        dep_path = tmp_path / tail if tail else head
+
+        return {
+            "external": {
+                "dir": {"file.h": ""},
+            },
+            "project": {
+                "setup.py": cleandoc(
+                    f"""
+                    from setuptools import Extension, setup
+                    setup(
+                        name="myproj",
+                        version="42",
+                        ext_modules=[
+                            Extension(
+                                "hello", sources=["hello.pyx"],
+                                depends=[{str(dep_path)!r}]
+                            )
+                        ],
+                    )
+                    """
+                ),
+                "hello.pyx": "",
+                "MANIFEST.in": "global-include *.h",
+            },
+        }
+
+    @pytest.mark.parametrize(
+        "dep_path", ("$tmp_path$/external/dir/file.h", "../external/dir/file.h")
+    )
+    def test_external_path_in_extension_depends(self, monkeypatch, tmp_path, dep_path):
+        # Given a project with a "depends" targeting an external dir
+        files = self.files_for_external_path_in_extension_depends(tmp_path, dep_path)
+        jaraco.path.build(files, prefix=str(tmp_path))
+        # When `sdist` runs, there should be no error
+        members = run_sdist(monkeypatch, tmp_path / "project")
+        # and the sdist should not contain the external file
+        for name in members:
+            assert "file.h" not in name
+
+
+def run_sdist(monkeypatch, project):
+    """Given a project directory, run the sdist and return its contents"""
+    monkeypatch.chdir(project)
+    with quiet():
+        run_setup("setup.py", ["sdist"])
+
+    archive = next((project / "dist").glob("*.tar.gz"))
+    with tarfile.open(str(archive)) as tar:
+        return set(tar.getnames())
+
+
+def test_sanity_check_setuptools_own_sdist(setuptools_sdist):
+    with tarfile.open(setuptools_sdist) as tar:
+        files = tar.getnames()
+
+    # setuptools sdist should not include the .tox folder
+    tox_files = [name for name in files if ".tox" in name]
+    assert len(tox_files) == 0, f"not empty {tox_files}"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setopt.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setopt.py
new file mode 100644
index 00000000..ccf25618
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setopt.py
@@ -0,0 +1,40 @@
+import configparser
+
+from setuptools.command import setopt
+
+
+class TestEdit:
+    @staticmethod
+    def parse_config(filename):
+        parser = configparser.ConfigParser()
+        with open(filename, encoding='utf-8') as reader:
+            parser.read_file(reader)
+        return parser
+
+    @staticmethod
+    def write_text(file, content):
+        with open(file, 'wb') as strm:
+            strm.write(content.encode('utf-8'))
+
+    def test_utf8_encoding_retained(self, tmpdir):
+        """
+        When editing a file, non-ASCII characters encoded in
+        UTF-8 should be retained.
+        """
+        config = tmpdir.join('setup.cfg')
+        self.write_text(str(config), '[names]\njaraco=джарако')
+        setopt.edit_config(str(config), dict(names=dict(other='yes')))
+        parser = self.parse_config(str(config))
+        assert parser.get('names', 'jaraco') == 'джарако'
+        assert parser.get('names', 'other') == 'yes'
+
+    def test_case_retained(self, tmpdir):
+        """
+        When editing a file, case of keys should be retained.
+        """
+        config = tmpdir.join('setup.cfg')
+        self.write_text(str(config), '[names]\nFoO=bAr')
+        setopt.edit_config(str(config), dict(names=dict(oTher='yes')))
+        actual = config.read_text(encoding='ascii')
+        assert 'FoO' in actual
+        assert 'oTher' in actual
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setuptools.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setuptools.py
new file mode 100644
index 00000000..1d56e1a8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_setuptools.py
@@ -0,0 +1,290 @@
+"""Tests for the 'setuptools' package"""
+
+import os
+import re
+import sys
+from zipfile import ZipFile
+
+import pytest
+from packaging.version import Version
+
+import setuptools
+import setuptools.depends as dep
+import setuptools.dist
+from setuptools.depends import Require
+
+import distutils.cmd
+import distutils.core
+from distutils.core import Extension
+from distutils.errors import DistutilsSetupError
+
+
+@pytest.fixture(autouse=True)
+def isolated_dir(tmpdir_cwd):
+    return
+
+
+def makeSetup(**args):
+    """Return distribution from 'setup(**args)', without executing commands"""
+
+    distutils.core._setup_stop_after = "commandline"
+
+    # Don't let system command line leak into tests!
+    args.setdefault('script_args', ['install'])
+
+    try:
+        return setuptools.setup(**args)
+    finally:
+        distutils.core._setup_stop_after = None
+
+
+needs_bytecode = pytest.mark.skipif(
+    not hasattr(dep, 'get_module_constant'),
+    reason="bytecode support not available",
+)
+
+
+class TestDepends:
+    def testExtractConst(self):
+        if not hasattr(dep, 'extract_constant'):
+            # skip on non-bytecode platforms
+            return
+
+        def f1():
+            global x, y, z
+            x = "test"
+            y = z  # pyright: ignore[reportUnboundVariable] # Explicitly testing for this runtime issue
+
+        fc = f1.__code__
+
+        # unrecognized name
+        assert dep.extract_constant(fc, 'q', -1) is None
+
+        # constant assigned
+        assert dep.extract_constant(fc, 'x', -1) == "test"
+
+        # expression assigned
+        assert dep.extract_constant(fc, 'y', -1) == -1
+
+        # recognized name, not assigned
+        assert dep.extract_constant(fc, 'z', -1) is None
+
+    def testFindModule(self):
+        with pytest.raises(ImportError):
+            dep.find_module('no-such.-thing')
+        with pytest.raises(ImportError):
+            dep.find_module('setuptools.non-existent')
+        f, _p, _i = dep.find_module('setuptools.tests')
+        f.close()
+
+    @needs_bytecode
+    def testModuleExtract(self):
+        from json import __version__
+
+        assert dep.get_module_constant('json', '__version__') == __version__
+        assert dep.get_module_constant('sys', 'version') == sys.version
+        assert (
+            dep.get_module_constant('setuptools.tests.test_setuptools', '__doc__')
+            == __doc__
+        )
+
+    @needs_bytecode
+    def testRequire(self):
+        req = Require('Json', '1.0.3', 'json')
+
+        assert req.name == 'Json'
+        assert req.module == 'json'
+        assert req.requested_version == Version('1.0.3')
+        assert req.attribute == '__version__'
+        assert req.full_name() == 'Json-1.0.3'
+
+        from json import __version__
+
+        assert str(req.get_version()) == __version__
+        assert req.version_ok('1.0.9')
+        assert not req.version_ok('0.9.1')
+        assert not req.version_ok('unknown')
+
+        assert req.is_present()
+        assert req.is_current()
+
+        req = Require('Do-what-I-mean', '1.0', 'd-w-i-m')
+        assert not req.is_present()
+        assert not req.is_current()
+
+    @needs_bytecode
+    def test_require_present(self):
+        # In #1896, this test was failing for months with the only
+        # complaint coming from test runners (not end users).
+        # TODO: Evaluate if this code is needed at all.
+        req = Require('Tests', None, 'tests', homepage="http://example.com")
+        assert req.format is None
+        assert req.attribute is None
+        assert req.requested_version is None
+        assert req.full_name() == 'Tests'
+        assert req.homepage == 'http://example.com'
+
+        from setuptools.tests import __path__
+
+        paths = [os.path.dirname(p) for p in __path__]
+        assert req.is_present(paths)
+        assert req.is_current(paths)
+
+
+class TestDistro:
+    def setup_method(self, method):
+        self.e1 = Extension('bar.ext', ['bar.c'])
+        self.e2 = Extension('c.y', ['y.c'])
+
+        self.dist = makeSetup(
+            packages=['a', 'a.b', 'a.b.c', 'b', 'c'],
+            py_modules=['b.d', 'x'],
+            ext_modules=(self.e1, self.e2),
+            package_dir={},
+        )
+
+    def testDistroType(self):
+        assert isinstance(self.dist, setuptools.dist.Distribution)
+
+    def testExcludePackage(self):
+        self.dist.exclude_package('a')
+        assert self.dist.packages == ['b', 'c']
+
+        self.dist.exclude_package('b')
+        assert self.dist.packages == ['c']
+        assert self.dist.py_modules == ['x']
+        assert self.dist.ext_modules == [self.e1, self.e2]
+
+        self.dist.exclude_package('c')
+        assert self.dist.packages == []
+        assert self.dist.py_modules == ['x']
+        assert self.dist.ext_modules == [self.e1]
+
+        # test removals from unspecified options
+        makeSetup().exclude_package('x')
+
+    def testIncludeExclude(self):
+        # remove an extension
+        self.dist.exclude(ext_modules=[self.e1])
+        assert self.dist.ext_modules == [self.e2]
+
+        # add it back in
+        self.dist.include(ext_modules=[self.e1])
+        assert self.dist.ext_modules == [self.e2, self.e1]
+
+        # should not add duplicate
+        self.dist.include(ext_modules=[self.e1])
+        assert self.dist.ext_modules == [self.e2, self.e1]
+
+    def testExcludePackages(self):
+        self.dist.exclude(packages=['c', 'b', 'a'])
+        assert self.dist.packages == []
+        assert self.dist.py_modules == ['x']
+        assert self.dist.ext_modules == [self.e1]
+
+    def testEmpty(self):
+        dist = makeSetup()
+        dist.include(packages=['a'], py_modules=['b'], ext_modules=[self.e2])
+        dist = makeSetup()
+        dist.exclude(packages=['a'], py_modules=['b'], ext_modules=[self.e2])
+
+    def testContents(self):
+        assert self.dist.has_contents_for('a')
+        self.dist.exclude_package('a')
+        assert not self.dist.has_contents_for('a')
+
+        assert self.dist.has_contents_for('b')
+        self.dist.exclude_package('b')
+        assert not self.dist.has_contents_for('b')
+
+        assert self.dist.has_contents_for('c')
+        self.dist.exclude_package('c')
+        assert not self.dist.has_contents_for('c')
+
+    def testInvalidIncludeExclude(self):
+        with pytest.raises(DistutilsSetupError):
+            self.dist.include(nonexistent_option='x')
+        with pytest.raises(DistutilsSetupError):
+            self.dist.exclude(nonexistent_option='x')
+        with pytest.raises(DistutilsSetupError):
+            self.dist.include(packages={'x': 'y'})
+        with pytest.raises(DistutilsSetupError):
+            self.dist.exclude(packages={'x': 'y'})
+        with pytest.raises(DistutilsSetupError):
+            self.dist.include(ext_modules={'x': 'y'})
+        with pytest.raises(DistutilsSetupError):
+            self.dist.exclude(ext_modules={'x': 'y'})
+
+        with pytest.raises(DistutilsSetupError):
+            self.dist.include(package_dir=['q'])
+        with pytest.raises(DistutilsSetupError):
+            self.dist.exclude(package_dir=['q'])
+
+
+@pytest.fixture
+def example_source(tmpdir):
+    tmpdir.mkdir('foo')
+    (tmpdir / 'foo/bar.py').write('')
+    (tmpdir / 'readme.txt').write('')
+    return tmpdir
+
+
+def test_findall(example_source):
+    found = list(setuptools.findall(str(example_source)))
+    expected = ['readme.txt', 'foo/bar.py']
+    expected = [example_source.join(fn) for fn in expected]
+    assert found == expected
+
+
+def test_findall_curdir(example_source):
+    with example_source.as_cwd():
+        found = list(setuptools.findall())
+    expected = ['readme.txt', os.path.join('foo', 'bar.py')]
+    assert found == expected
+
+
+@pytest.fixture
+def can_symlink(tmpdir):
+    """
+    Skip if cannot create a symbolic link
+    """
+    link_fn = 'link'
+    target_fn = 'target'
+    try:
+        os.symlink(target_fn, link_fn)
+    except (OSError, NotImplementedError, AttributeError):
+        pytest.skip("Cannot create symbolic links")
+    os.remove(link_fn)
+
+
+@pytest.mark.usefixtures("can_symlink")
+def test_findall_missing_symlink(tmpdir):
+    with tmpdir.as_cwd():
+        os.symlink('foo', 'bar')
+        found = list(setuptools.findall())
+        assert found == []
+
+
+@pytest.mark.xfail(reason="unable to exclude tests; #4475 #3260")
+def test_its_own_wheel_does_not_contain_tests(setuptools_wheel):
+    with ZipFile(setuptools_wheel) as zipfile:
+        contents = [f.replace(os.sep, '/') for f in zipfile.namelist()]
+
+    for member in contents:
+        assert '/tests/' not in member
+
+
+def test_wheel_includes_cli_scripts(setuptools_wheel):
+    with ZipFile(setuptools_wheel) as zipfile:
+        contents = [f.replace(os.sep, '/') for f in zipfile.namelist()]
+
+    assert any('cli-64.exe' in member for member in contents)
+
+
+def test_wheel_includes_vendored_metadata(setuptools_wheel):
+    with ZipFile(setuptools_wheel) as zipfile:
+        contents = [f.replace(os.sep, '/') for f in zipfile.namelist()]
+
+    assert any(
+        re.search(r'_vendor/.*\.dist-info/METADATA', member) for member in contents
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_shutil_wrapper.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_shutil_wrapper.py
new file mode 100644
index 00000000..74ff7e9a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_shutil_wrapper.py
@@ -0,0 +1,23 @@
+import stat
+import sys
+from unittest.mock import Mock
+
+from setuptools import _shutil
+
+
+def test_rmtree_readonly(monkeypatch, tmp_path):
+    """Verify onerr works as expected"""
+
+    tmp_dir = tmp_path / "with_readonly"
+    tmp_dir.mkdir()
+    some_file = tmp_dir.joinpath("file.txt")
+    some_file.touch()
+    some_file.chmod(stat.S_IREAD)
+
+    expected_count = 1 if sys.platform.startswith("win") else 0
+    chmod_fn = Mock(wraps=_shutil.attempt_chmod_verbose)
+    monkeypatch.setattr(_shutil, "attempt_chmod_verbose", chmod_fn)
+
+    _shutil.rmtree(tmp_dir)
+    assert chmod_fn.call_count == expected_count
+    assert not tmp_dir.is_dir()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_unicode_utils.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_unicode_utils.py
new file mode 100644
index 00000000..a24a9bd5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_unicode_utils.py
@@ -0,0 +1,10 @@
+from setuptools import unicode_utils
+
+
+def test_filesys_decode_fs_encoding_is_None(monkeypatch):
+    """
+    Test filesys_decode does not raise TypeError when
+    getfilesystemencoding returns None.
+    """
+    monkeypatch.setattr('sys.getfilesystemencoding', lambda: None)
+    unicode_utils.filesys_decode(b'test')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_virtualenv.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_virtualenv.py
new file mode 100644
index 00000000..b02949ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_virtualenv.py
@@ -0,0 +1,113 @@
+import os
+import subprocess
+import sys
+from urllib.error import URLError
+from urllib.request import urlopen
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def pytest_virtualenv_works(venv):
+    """
+    pytest_virtualenv may not work. if it doesn't, skip these
+    tests. See #1284.
+    """
+    venv_prefix = venv.run(["python", "-c", "import sys; print(sys.prefix)"]).strip()
+    if venv_prefix == sys.prefix:
+        pytest.skip("virtualenv is broken (see pypa/setuptools#1284)")
+
+
+def test_clean_env_install(venv_without_setuptools, setuptools_wheel):
+    """
+    Check setuptools can be installed in a clean environment.
+    """
+    cmd = ["python", "-m", "pip", "install", str(setuptools_wheel)]
+    venv_without_setuptools.run(cmd)
+
+
+def access_pypi():
+    # Detect if tests are being run without connectivity
+    if not os.environ.get('NETWORK_REQUIRED', False):  # pragma: nocover
+        try:
+            urlopen('https://pypi.org', timeout=1)
+        except URLError:
+            # No network, disable most of these tests
+            return False
+
+    return True
+
+
+@pytest.mark.skipif(
+    'platform.python_implementation() == "PyPy"',
+    reason="https://github.com/pypa/setuptools/pull/2865#issuecomment-965834995",
+)
+@pytest.mark.skipif(not access_pypi(), reason="no network")
+# ^-- Even when it is not necessary to install a different version of `pip`
+#     the build process will still try to download `wheel`, see #3147 and #2986.
+@pytest.mark.parametrize(
+    'pip_version',
+    [
+        None,
+        pytest.param(
+            'pip<20.1',
+            marks=pytest.mark.xfail(
+                'sys.version_info >= (3, 12)',
+                reason="pip 23.1.2 required for Python 3.12 and later",
+            ),
+        ),
+        pytest.param(
+            'pip<21',
+            marks=pytest.mark.xfail(
+                'sys.version_info >= (3, 12)',
+                reason="pip 23.1.2 required for Python 3.12 and later",
+            ),
+        ),
+        pytest.param(
+            'pip<22',
+            marks=pytest.mark.xfail(
+                'sys.version_info >= (3, 12)',
+                reason="pip 23.1.2 required for Python 3.12 and later",
+            ),
+        ),
+        pytest.param(
+            'pip<23',
+            marks=pytest.mark.xfail(
+                'sys.version_info >= (3, 12)',
+                reason="pip 23.1.2 required for Python 3.12 and later",
+            ),
+        ),
+        pytest.param(
+            'https://github.com/pypa/pip/archive/main.zip',
+            marks=pytest.mark.xfail(reason='#2975'),
+        ),
+    ],
+)
+def test_pip_upgrade_from_source(
+    pip_version, venv_without_setuptools, setuptools_wheel, setuptools_sdist
+):
+    """
+    Check pip can upgrade setuptools from source.
+    """
+    # Install pip/wheel, in a venv without setuptools (as it
+    # should not be needed for bootstrapping from source)
+    venv = venv_without_setuptools
+    venv.run(["pip", "install", "-U", "wheel"])
+    if pip_version is not None:
+        venv.run(["python", "-m", "pip", "install", "-U", pip_version, "--retries=1"])
+    with pytest.raises(subprocess.CalledProcessError):
+        # Meta-test to make sure setuptools is not installed
+        venv.run(["python", "-c", "import setuptools"])
+
+    # Then install from wheel.
+    venv.run(["pip", "install", str(setuptools_wheel)])
+    # And finally try to upgrade from source.
+    venv.run(["pip", "install", "--no-cache-dir", "--upgrade", str(setuptools_sdist)])
+
+
+def test_no_missing_dependencies(bare_venv, request):
+    """
+    Quick and dirty test to ensure all external dependencies are vendored.
+    """
+    setuptools_dir = request.config.rootdir
+    bare_venv.run(['python', 'setup.py', '--help'], cwd=setuptools_dir)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_warnings.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_warnings.py
new file mode 100644
index 00000000..41193d4f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_warnings.py
@@ -0,0 +1,106 @@
+from inspect import cleandoc
+
+import pytest
+
+from setuptools.warnings import SetuptoolsDeprecationWarning, SetuptoolsWarning
+
+_EXAMPLES = {
+    "default": dict(
+        args=("Hello {x}", "\n\t{target} {v:.1f}"),
+        kwargs={"x": 5, "v": 3, "target": "World"},
+        expected="""
+    Hello 5
+    !!
+
+            ********************************************************************************
+            World 3.0
+            ********************************************************************************
+
+    !!
+    """,
+    ),
+    "futue_due_date": dict(
+        args=("Summary", "Lorem ipsum"),
+        kwargs={"due_date": (9999, 11, 22)},
+        expected="""
+    Summary
+    !!
+
+            ********************************************************************************
+            Lorem ipsum
+
+            By 9999-Nov-22, you need to update your project and remove deprecated calls
+            or your builds will no longer be supported.
+            ********************************************************************************
+
+    !!
+    """,
+    ),
+    "past_due_date_with_docs": dict(
+        args=("Summary", "Lorem ipsum"),
+        kwargs={"due_date": (2000, 11, 22), "see_docs": "some_page.html"},
+        expected="""
+    Summary
+    !!
+
+            ********************************************************************************
+            Lorem ipsum
+
+            This deprecation is overdue, please update your project and remove deprecated
+            calls to avoid build errors in the future.
+
+            See https://setuptools.pypa.io/en/latest/some_page.html for details.
+            ********************************************************************************
+
+    !!
+    """,
+    ),
+}
+
+
+@pytest.mark.parametrize("example_name", _EXAMPLES.keys())
+def test_formatting(monkeypatch, example_name):
+    """
+    It should automatically handle indentation, interpolation and things like due date.
+    """
+    args = _EXAMPLES[example_name]["args"]
+    kwargs = _EXAMPLES[example_name]["kwargs"]
+    expected = _EXAMPLES[example_name]["expected"]
+
+    monkeypatch.setenv("SETUPTOOLS_ENFORCE_DEPRECATION", "false")
+    with pytest.warns(SetuptoolsWarning) as warn_info:
+        SetuptoolsWarning.emit(*args, **kwargs)
+    assert _get_message(warn_info) == cleandoc(expected)
+
+
+def test_due_date_enforcement(monkeypatch):
+    class _MyDeprecation(SetuptoolsDeprecationWarning):
+        _SUMMARY = "Summary"
+        _DETAILS = "Lorem ipsum"
+        _DUE_DATE = (2000, 11, 22)
+        _SEE_DOCS = "some_page.html"
+
+    monkeypatch.setenv("SETUPTOOLS_ENFORCE_DEPRECATION", "true")
+    with pytest.raises(SetuptoolsDeprecationWarning) as exc_info:
+        _MyDeprecation.emit()
+
+    expected = """
+    Summary
+    !!
+
+            ********************************************************************************
+            Lorem ipsum
+
+            This deprecation is overdue, please update your project and remove deprecated
+            calls to avoid build errors in the future.
+
+            See https://setuptools.pypa.io/en/latest/some_page.html for details.
+            ********************************************************************************
+
+    !!
+    """
+    assert str(exc_info.value) == cleandoc(expected)
+
+
+def _get_message(warn_info):
+    return next(warn.message.args[0] for warn in warn_info)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_wheel.py
new file mode 100644
index 00000000..70165c60
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_wheel.py
@@ -0,0 +1,714 @@
+"""wheel tests"""
+
+from __future__ import annotations
+
+import contextlib
+import glob
+import inspect
+import os
+import pathlib
+import shutil
+import stat
+import subprocess
+import sys
+import zipfile
+from typing import Any
+
+import pytest
+from jaraco import path
+from packaging.tags import parse_tag
+from packaging.utils import canonicalize_name
+
+from pkg_resources import PY_MAJOR, Distribution, PathMetadata
+from setuptools.wheel import Wheel
+
+from .contexts import tempdir
+from .textwrap import DALS
+
+from distutils.sysconfig import get_config_var
+from distutils.util import get_platform
+
+WHEEL_INFO_TESTS = (
+    ('invalid.whl', ValueError),
+    (
+        'simplewheel-2.0-1-py2.py3-none-any.whl',
+        {
+            'project_name': 'simplewheel',
+            'version': '2.0',
+            'build': '1',
+            'py_version': 'py2.py3',
+            'abi': 'none',
+            'platform': 'any',
+        },
+    ),
+    (
+        'simple.dist-0.1-py2.py3-none-any.whl',
+        {
+            'project_name': 'simple.dist',
+            'version': '0.1',
+            'build': None,
+            'py_version': 'py2.py3',
+            'abi': 'none',
+            'platform': 'any',
+        },
+    ),
+    (
+        'example_pkg_a-1-py3-none-any.whl',
+        {
+            'project_name': 'example_pkg_a',
+            'version': '1',
+            'build': None,
+            'py_version': 'py3',
+            'abi': 'none',
+            'platform': 'any',
+        },
+    ),
+    (
+        'PyQt5-5.9-5.9.1-cp35.cp36.cp37-abi3-manylinux1_x86_64.whl',
+        {
+            'project_name': 'PyQt5',
+            'version': '5.9',
+            'build': '5.9.1',
+            'py_version': 'cp35.cp36.cp37',
+            'abi': 'abi3',
+            'platform': 'manylinux1_x86_64',
+        },
+    ),
+)
+
+
+@pytest.mark.parametrize(
+    ('filename', 'info'), WHEEL_INFO_TESTS, ids=[t[0] for t in WHEEL_INFO_TESTS]
+)
+def test_wheel_info(filename, info):
+    if inspect.isclass(info):
+        with pytest.raises(info):
+            Wheel(filename)
+        return
+    w = Wheel(filename)
+    assert {k: getattr(w, k) for k in info.keys()} == info
+
+
+@contextlib.contextmanager
+def build_wheel(extra_file_defs=None, **kwargs):
+    file_defs = {
+        'setup.py': (
+            DALS(
+                """
+            # -*- coding: utf-8 -*-
+            from setuptools import setup
+            import setuptools
+            setup(**%r)
+            """
+            )
+            % kwargs
+        ).encode('utf-8'),
+    }
+    if extra_file_defs:
+        file_defs.update(extra_file_defs)
+    with tempdir() as source_dir:
+        path.build(file_defs, source_dir)
+        subprocess.check_call(
+            (sys.executable, 'setup.py', '-q', 'bdist_wheel'), cwd=source_dir
+        )
+        yield glob.glob(os.path.join(source_dir, 'dist', '*.whl'))[0]
+
+
+def tree_set(root):
+    contents = set()
+    for dirpath, dirnames, filenames in os.walk(root):
+        for filename in filenames:
+            contents.add(os.path.join(os.path.relpath(dirpath, root), filename))
+    return contents
+
+
+def flatten_tree(tree):
+    """Flatten nested dicts and lists into a full list of paths"""
+    output = set()
+    for node, contents in tree.items():
+        if isinstance(contents, dict):
+            contents = flatten_tree(contents)
+
+        for elem in contents:
+            if isinstance(elem, dict):
+                output |= {os.path.join(node, val) for val in flatten_tree(elem)}
+            else:
+                output.add(os.path.join(node, elem))
+    return output
+
+
+def format_install_tree(tree):
+    return {
+        x.format(
+            py_version=PY_MAJOR,
+            platform=get_platform(),
+            shlib_ext=get_config_var('EXT_SUFFIX') or get_config_var('SO'),
+        )
+        for x in tree
+    }
+
+
+def _check_wheel_install(
+    filename, install_dir, install_tree_includes, project_name, version, requires_txt
+):
+    w = Wheel(filename)
+    egg_path = os.path.join(install_dir, w.egg_name())
+    w.install_as_egg(egg_path)
+    if install_tree_includes is not None:
+        install_tree = format_install_tree(install_tree_includes)
+        exp = tree_set(install_dir)
+        assert install_tree.issubset(exp), install_tree - exp
+
+    metadata = PathMetadata(egg_path, os.path.join(egg_path, 'EGG-INFO'))
+    dist = Distribution.from_filename(egg_path, metadata=metadata)
+    assert dist.project_name == project_name
+    assert dist.version == version
+    if requires_txt is None:
+        assert not dist.has_metadata('requires.txt')
+    else:
+        # Order must match to ensure reproducibility.
+        assert requires_txt == dist.get_metadata('requires.txt').lstrip()
+
+
+class Record:
+    def __init__(self, id, **kwargs):
+        self._id = id
+        self._fields = kwargs
+
+    def __repr__(self) -> str:
+        return f'{self._id}(**{self._fields!r})'
+
+
+# Using Any to avoid possible type union issues later in test
+# making a TypedDict is not worth in a test and anonymous/inline TypedDict are experimental
+# https://github.com/python/mypy/issues/9884
+WHEEL_INSTALL_TESTS: tuple[dict[str, Any], ...] = (
+    dict(
+        id='basic',
+        file_defs={'foo': {'__init__.py': ''}},
+        setup_kwargs=dict(
+            packages=['foo'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': ['PKG-INFO', 'RECORD', 'WHEEL', 'top_level.txt'],
+                'foo': ['__init__.py'],
+            }
+        }),
+    ),
+    dict(
+        id='utf-8',
+        setup_kwargs=dict(
+            description='Description accentuée',
+        ),
+    ),
+    dict(
+        id='data',
+        file_defs={
+            'data.txt': DALS(
+                """
+                Some data...
+                """
+            ),
+        },
+        setup_kwargs=dict(
+            data_files=[('data_dir', ['data.txt'])],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': ['PKG-INFO', 'RECORD', 'WHEEL', 'top_level.txt'],
+                'data_dir': ['data.txt'],
+            }
+        }),
+    ),
+    dict(
+        id='extension',
+        file_defs={
+            'extension.c': DALS(
+                """
+                #include "Python.h"
+
+                #if PY_MAJOR_VERSION >= 3
+
+                static struct PyModuleDef moduledef = {
+                        PyModuleDef_HEAD_INIT,
+                        "extension",
+                        NULL,
+                        0,
+                        NULL,
+                        NULL,
+                        NULL,
+                        NULL,
+                        NULL
+                };
+
+                #define INITERROR return NULL
+
+                PyMODINIT_FUNC PyInit_extension(void)
+
+                #else
+
+                #define INITERROR return
+
+                void initextension(void)
+
+                #endif
+                {
+                #if PY_MAJOR_VERSION >= 3
+                    PyObject *module = PyModule_Create(&moduledef);
+                #else
+                    PyObject *module = Py_InitModule("extension", NULL);
+                #endif
+                    if (module == NULL)
+                        INITERROR;
+                #if PY_MAJOR_VERSION >= 3
+                    return module;
+                #endif
+                }
+                """
+            ),
+        },
+        setup_kwargs=dict(
+            ext_modules=[
+                Record(
+                    'setuptools.Extension', name='extension', sources=['extension.c']
+                )
+            ],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}-{platform}.egg': [
+                'extension{shlib_ext}',
+                {
+                    'EGG-INFO': [
+                        'PKG-INFO',
+                        'RECORD',
+                        'WHEEL',
+                        'top_level.txt',
+                    ]
+                },
+            ]
+        }),
+    ),
+    dict(
+        id='header',
+        file_defs={
+            'header.h': DALS(
+                """
+                """
+            ),
+        },
+        setup_kwargs=dict(
+            headers=['header.h'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': [
+                'header.h',
+                {
+                    'EGG-INFO': [
+                        'PKG-INFO',
+                        'RECORD',
+                        'WHEEL',
+                        'top_level.txt',
+                    ]
+                },
+            ]
+        }),
+    ),
+    dict(
+        id='script',
+        file_defs={
+            'script.py': DALS(
+                """
+                #/usr/bin/python
+                print('hello world!')
+                """
+            ),
+            'script.sh': DALS(
+                """
+                #/bin/sh
+                echo 'hello world!'
+                """
+            ),
+        },
+        setup_kwargs=dict(
+            scripts=['script.py', 'script.sh'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': [
+                    'PKG-INFO',
+                    'RECORD',
+                    'WHEEL',
+                    'top_level.txt',
+                    {'scripts': ['script.py', 'script.sh']},
+                ]
+            }
+        }),
+    ),
+    dict(
+        id='requires1',
+        install_requires='foobar==2.0',
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': [
+                    'PKG-INFO',
+                    'RECORD',
+                    'WHEEL',
+                    'requires.txt',
+                    'top_level.txt',
+                ]
+            }
+        }),
+        requires_txt=DALS(
+            """
+            foobar==2.0
+            """
+        ),
+    ),
+    dict(
+        id='requires2',
+        install_requires=f"""
+        bar
+        foo<=2.0; {sys.platform!r} in sys_platform
+        """,
+        requires_txt=DALS(
+            """
+            bar
+            foo<=2.0
+            """
+        ),
+    ),
+    dict(
+        id='requires3',
+        install_requires=f"""
+        bar; {sys.platform!r} != sys_platform
+        """,
+    ),
+    dict(
+        id='requires4',
+        install_requires="""
+        foo
+        """,
+        extras_require={
+            'extra': 'foobar>3',
+        },
+        requires_txt=DALS(
+            """
+            foo
+
+            [extra]
+            foobar>3
+            """
+        ),
+    ),
+    dict(
+        id='requires5',
+        extras_require={
+            'extra': f'foobar; {sys.platform!r} != sys_platform',
+        },
+        requires_txt=DALS(
+            """
+            [extra]
+            """
+        ),
+    ),
+    dict(
+        id='requires_ensure_order',
+        install_requires="""
+        foo
+        bar
+        baz
+        qux
+        """,
+        extras_require={
+            'extra': """
+            foobar>3
+            barbaz>4
+            bazqux>5
+            quxzap>6
+            """,
+        },
+        requires_txt=DALS(
+            """
+            foo
+            bar
+            baz
+            qux
+
+            [extra]
+            foobar>3
+            barbaz>4
+            bazqux>5
+            quxzap>6
+            """
+        ),
+    ),
+    dict(
+        id='namespace_package',
+        file_defs={
+            'foo': {
+                'bar': {'__init__.py': ''},
+            },
+        },
+        setup_kwargs=dict(
+            namespace_packages=['foo'],
+            packages=['foo.bar'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': [
+                'foo-1.0-py{py_version}-nspkg.pth',
+                {
+                    'EGG-INFO': [
+                        'PKG-INFO',
+                        'RECORD',
+                        'WHEEL',
+                        'namespace_packages.txt',
+                        'top_level.txt',
+                    ]
+                },
+                {
+                    'foo': [
+                        '__init__.py',
+                        {'bar': ['__init__.py']},
+                    ]
+                },
+            ]
+        }),
+    ),
+    dict(
+        id='empty_namespace_package',
+        file_defs={
+            'foobar': {
+                '__init__.py': (
+                    "__import__('pkg_resources').declare_namespace(__name__)"
+                )
+            },
+        },
+        setup_kwargs=dict(
+            namespace_packages=['foobar'],
+            packages=['foobar'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': [
+                'foo-1.0-py{py_version}-nspkg.pth',
+                {
+                    'EGG-INFO': [
+                        'PKG-INFO',
+                        'RECORD',
+                        'WHEEL',
+                        'namespace_packages.txt',
+                        'top_level.txt',
+                    ]
+                },
+                {
+                    'foobar': [
+                        '__init__.py',
+                    ]
+                },
+            ]
+        }),
+    ),
+    dict(
+        id='data_in_package',
+        file_defs={
+            'foo': {
+                '__init__.py': '',
+                'data_dir': {
+                    'data.txt': DALS(
+                        """
+                        Some data...
+                        """
+                    ),
+                },
+            }
+        },
+        setup_kwargs=dict(
+            packages=['foo'],
+            data_files=[('foo/data_dir', ['foo/data_dir/data.txt'])],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': [
+                    'PKG-INFO',
+                    'RECORD',
+                    'WHEEL',
+                    'top_level.txt',
+                ],
+                'foo': [
+                    '__init__.py',
+                    {
+                        'data_dir': [
+                            'data.txt',
+                        ]
+                    },
+                ],
+            }
+        }),
+    ),
+)
+
+
+@pytest.mark.parametrize(
+    'params',
+    WHEEL_INSTALL_TESTS,
+    ids=[params['id'] for params in WHEEL_INSTALL_TESTS],
+)
+def test_wheel_install(params):
+    project_name = params.get('name', 'foo')
+    version = params.get('version', '1.0')
+    install_requires = params.get('install_requires', [])
+    extras_require = params.get('extras_require', {})
+    requires_txt = params.get('requires_txt', None)
+    install_tree = params.get('install_tree')
+    file_defs = params.get('file_defs', {})
+    setup_kwargs = params.get('setup_kwargs', {})
+    with (
+        build_wheel(
+            name=project_name,
+            version=version,
+            install_requires=install_requires,
+            extras_require=extras_require,
+            extra_file_defs=file_defs,
+            **setup_kwargs,
+        ) as filename,
+        tempdir() as install_dir,
+    ):
+        _check_wheel_install(
+            filename, install_dir, install_tree, project_name, version, requires_txt
+        )
+
+
+def test_wheel_install_pep_503():
+    project_name = 'Foo_Bar'  # PEP 503 canonicalized name is "foo-bar"
+    version = '1.0'
+    with (
+        build_wheel(
+            name=project_name,
+            version=version,
+        ) as filename,
+        tempdir() as install_dir,
+    ):
+        new_filename = filename.replace(project_name, canonicalize_name(project_name))
+        shutil.move(filename, new_filename)
+        _check_wheel_install(
+            new_filename,
+            install_dir,
+            None,
+            canonicalize_name(project_name),
+            version,
+            None,
+        )
+
+
+def test_wheel_no_dist_dir():
+    project_name = 'nodistinfo'
+    version = '1.0'
+    wheel_name = f'{project_name}-{version}-py2.py3-none-any.whl'
+    with tempdir() as source_dir:
+        wheel_path = os.path.join(source_dir, wheel_name)
+        # create an empty zip file
+        zipfile.ZipFile(wheel_path, 'w').close()
+        with tempdir() as install_dir:
+            with pytest.raises(ValueError):
+                _check_wheel_install(
+                    wheel_path, install_dir, None, project_name, version, None
+                )
+
+
+def test_wheel_is_compatible(monkeypatch):
+    def sys_tags():
+        return {
+            (t.interpreter, t.abi, t.platform)
+            for t in parse_tag('cp36-cp36m-manylinux1_x86_64')
+        }
+
+    monkeypatch.setattr('setuptools.wheel._get_supported_tags', sys_tags)
+    assert Wheel('onnxruntime-0.1.2-cp36-cp36m-manylinux1_x86_64.whl').is_compatible()
+
+
+def test_wheel_mode():
+    @contextlib.contextmanager
+    def build_wheel(extra_file_defs=None, **kwargs):
+        file_defs = {
+            'setup.py': (
+                DALS(
+                    """
+                # -*- coding: utf-8 -*-
+                from setuptools import setup
+                import setuptools
+                setup(**%r)
+                """
+                )
+                % kwargs
+            ).encode('utf-8'),
+        }
+        if extra_file_defs:
+            file_defs.update(extra_file_defs)
+        with tempdir() as source_dir:
+            path.build(file_defs, source_dir)
+            runsh = pathlib.Path(source_dir) / "script.sh"
+            os.chmod(runsh, 0o777)
+            subprocess.check_call(
+                (sys.executable, 'setup.py', '-q', 'bdist_wheel'), cwd=source_dir
+            )
+            yield glob.glob(os.path.join(source_dir, 'dist', '*.whl'))[0]
+
+    params = dict(
+        id='script',
+        file_defs={
+            'script.py': DALS(
+                """
+                #/usr/bin/python
+                print('hello world!')
+                """
+            ),
+            'script.sh': DALS(
+                """
+                #/bin/sh
+                echo 'hello world!'
+                """
+            ),
+        },
+        setup_kwargs=dict(
+            scripts=['script.py', 'script.sh'],
+        ),
+        install_tree=flatten_tree({
+            'foo-1.0-py{py_version}.egg': {
+                'EGG-INFO': [
+                    'PKG-INFO',
+                    'RECORD',
+                    'WHEEL',
+                    'top_level.txt',
+                    {'scripts': ['script.py', 'script.sh']},
+                ]
+            }
+        }),
+    )
+
+    project_name = params.get('name', 'foo')
+    version = params.get('version', '1.0')
+    install_tree = params.get('install_tree')
+    file_defs = params.get('file_defs', {})
+    setup_kwargs = params.get('setup_kwargs', {})
+
+    with (
+        build_wheel(
+            name=project_name,
+            version=version,
+            install_requires=[],
+            extras_require={},
+            extra_file_defs=file_defs,
+            **setup_kwargs,
+        ) as filename,
+        tempdir() as install_dir,
+    ):
+        _check_wheel_install(
+            filename, install_dir, install_tree, project_name, version, None
+        )
+        w = Wheel(filename)
+        base = pathlib.Path(install_dir) / w.egg_name()
+        script_sh = base / "EGG-INFO" / "scripts" / "script.sh"
+        assert script_sh.exists()
+        if sys.platform != 'win32':
+            # Editable file mode has no effect on Windows
+            assert oct(stat.S_IMODE(script_sh.stat().st_mode)) == "0o777"
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_windows_wrappers.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_windows_wrappers.py
new file mode 100644
index 00000000..f8954853
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/test_windows_wrappers.py
@@ -0,0 +1,259 @@
+"""
+Python Script Wrapper for Windows
+=================================
+
+setuptools includes wrappers for Python scripts that allows them to be
+executed like regular windows programs.  There are 2 wrappers, one
+for command-line programs, cli.exe, and one for graphical programs,
+gui.exe.  These programs are almost identical, function pretty much
+the same way, and are generated from the same source file.  The
+wrapper programs are used by copying them to the directory containing
+the script they are to wrap and with the same name as the script they
+are to wrap.
+"""
+
+import pathlib
+import platform
+import subprocess
+import sys
+import textwrap
+
+import pytest
+
+import pkg_resources
+from setuptools.command.easy_install import nt_quote_arg
+
+pytestmark = pytest.mark.skipif(sys.platform != 'win32', reason="Windows only")
+
+
+class WrapperTester:
+    @classmethod
+    def prep_script(cls, template):
+        python_exe = nt_quote_arg(sys.executable)
+        return template % locals()
+
+    @classmethod
+    def create_script(cls, tmpdir):
+        """
+        Create a simple script, foo-script.py
+
+        Note that the script starts with a Unix-style '#!' line saying which
+        Python executable to run.  The wrapper will use this line to find the
+        correct Python executable.
+        """
+
+        script = cls.prep_script(cls.script_tmpl)
+
+        with (tmpdir / cls.script_name).open('w') as f:
+            f.write(script)
+
+        # also copy cli.exe to the sample directory
+        with (tmpdir / cls.wrapper_name).open('wb') as f:
+            w = pkg_resources.resource_string('setuptools', cls.wrapper_source)
+            f.write(w)
+
+
+def win_launcher_exe(prefix):
+    """A simple routine to select launcher script based on platform."""
+    assert prefix in ('cli', 'gui')
+    if platform.machine() == "ARM64":
+        return f"{prefix}-arm64.exe"
+    else:
+        return f"{prefix}-32.exe"
+
+
+class TestCLI(WrapperTester):
+    script_name = 'foo-script.py'
+    wrapper_name = 'foo.exe'
+    wrapper_source = win_launcher_exe('cli')
+
+    script_tmpl = textwrap.dedent(
+        """
+        #!%(python_exe)s
+        import sys
+        input = repr(sys.stdin.read())
+        print(sys.argv[0][-14:])
+        print(sys.argv[1:])
+        print(input)
+        if __debug__:
+            print('non-optimized')
+        """
+    ).lstrip()
+
+    def test_basic(self, tmpdir):
+        """
+        When the copy of cli.exe, foo.exe in this example, runs, it examines
+        the path name it was run with and computes a Python script path name
+        by removing the '.exe' suffix and adding the '-script.py' suffix. (For
+        GUI programs, the suffix '-script.pyw' is added.)  This is why we
+        named out script the way we did.  Now we can run out script by running
+        the wrapper:
+
+        This example was a little pathological in that it exercised windows
+        (MS C runtime) quoting rules:
+
+        - Strings containing spaces are surrounded by double quotes.
+
+        - Double quotes in strings need to be escaped by preceding them with
+          back slashes.
+
+        - One or more backslashes preceding double quotes need to be escaped
+          by preceding each of them with back slashes.
+        """
+        self.create_script(tmpdir)
+        cmd = [
+            str(tmpdir / 'foo.exe'),
+            'arg1',
+            'arg 2',
+            'arg "2\\"',
+            'arg 4\\',
+            'arg5 a\\\\b',
+        ]
+        proc = subprocess.Popen(
+            cmd,
+            stdout=subprocess.PIPE,
+            stdin=subprocess.PIPE,
+            text=True,
+            encoding="utf-8",
+        )
+        stdout, _stderr = proc.communicate('hello\nworld\n')
+        actual = stdout.replace('\r\n', '\n')
+        expected = textwrap.dedent(
+            r"""
+            \foo-script.py
+            ['arg1', 'arg 2', 'arg "2\\"', 'arg 4\\', 'arg5 a\\\\b']
+            'hello\nworld\n'
+            non-optimized
+            """
+        ).lstrip()
+        assert actual == expected
+
+    def test_symlink(self, tmpdir):
+        """
+        Ensure that symlink for the foo.exe is working correctly.
+        """
+        script_dir = tmpdir / "script_dir"
+        script_dir.mkdir()
+        self.create_script(script_dir)
+        symlink = pathlib.Path(tmpdir / "foo.exe")
+        symlink.symlink_to(script_dir / "foo.exe")
+
+        cmd = [
+            str(tmpdir / 'foo.exe'),
+            'arg1',
+            'arg 2',
+            'arg "2\\"',
+            'arg 4\\',
+            'arg5 a\\\\b',
+        ]
+        proc = subprocess.Popen(
+            cmd,
+            stdout=subprocess.PIPE,
+            stdin=subprocess.PIPE,
+            text=True,
+            encoding="utf-8",
+        )
+        stdout, _stderr = proc.communicate('hello\nworld\n')
+        actual = stdout.replace('\r\n', '\n')
+        expected = textwrap.dedent(
+            r"""
+            \foo-script.py
+            ['arg1', 'arg 2', 'arg "2\\"', 'arg 4\\', 'arg5 a\\\\b']
+            'hello\nworld\n'
+            non-optimized
+            """
+        ).lstrip()
+        assert actual == expected
+
+    def test_with_options(self, tmpdir):
+        """
+        Specifying Python Command-line Options
+        --------------------------------------
+
+        You can specify a single argument on the '#!' line.  This can be used
+        to specify Python options like -O, to run in optimized mode or -i
+        to start the interactive interpreter.  You can combine multiple
+        options as usual. For example, to run in optimized mode and
+        enter the interpreter after running the script, you could use -Oi:
+        """
+        self.create_script(tmpdir)
+        tmpl = textwrap.dedent(
+            """
+            #!%(python_exe)s  -Oi
+            import sys
+            input = repr(sys.stdin.read())
+            print(sys.argv[0][-14:])
+            print(sys.argv[1:])
+            print(input)
+            if __debug__:
+                print('non-optimized')
+            sys.ps1 = '---'
+            """
+        ).lstrip()
+        with (tmpdir / 'foo-script.py').open('w') as f:
+            f.write(self.prep_script(tmpl))
+        cmd = [str(tmpdir / 'foo.exe')]
+        proc = subprocess.Popen(
+            cmd,
+            stdout=subprocess.PIPE,
+            stdin=subprocess.PIPE,
+            stderr=subprocess.STDOUT,
+            text=True,
+            encoding="utf-8",
+        )
+        stdout, _stderr = proc.communicate()
+        actual = stdout.replace('\r\n', '\n')
+        expected = textwrap.dedent(
+            r"""
+            \foo-script.py
+            []
+            ''
+            ---
+            """
+        ).lstrip()
+        assert actual == expected
+
+
+class TestGUI(WrapperTester):
+    """
+    Testing the GUI Version
+    -----------------------
+    """
+
+    script_name = 'bar-script.pyw'
+    wrapper_source = win_launcher_exe('gui')
+    wrapper_name = 'bar.exe'
+
+    script_tmpl = textwrap.dedent(
+        """
+        #!%(python_exe)s
+        import sys
+        f = open(sys.argv[1], 'wb')
+        bytes_written = f.write(repr(sys.argv[2]).encode('utf-8'))
+        f.close()
+        """
+    ).strip()
+
+    def test_basic(self, tmpdir):
+        """Test the GUI version with the simple script, bar-script.py"""
+        self.create_script(tmpdir)
+
+        cmd = [
+            str(tmpdir / 'bar.exe'),
+            str(tmpdir / 'test_output.txt'),
+            'Test Argument',
+        ]
+        proc = subprocess.Popen(
+            cmd,
+            stdout=subprocess.PIPE,
+            stdin=subprocess.PIPE,
+            stderr=subprocess.STDOUT,
+            text=True,
+            encoding="utf-8",
+        )
+        stdout, stderr = proc.communicate()
+        assert not stdout
+        assert not stderr
+        with (tmpdir / 'test_output.txt').open('rb') as f_out:
+            actual = f_out.read().decode('ascii')
+        assert actual == repr('Test Argument')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/text.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/text.py
new file mode 100644
index 00000000..e05cc633
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/text.py
@@ -0,0 +1,4 @@
+class Filenames:
+    unicode = 'smörbröd.py'
+    latin_1 = unicode.encode('latin-1')
+    utf_8 = unicode.encode('utf-8')
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/tests/textwrap.py b/backend/venv/lib/python3.12/site-packages/setuptools/tests/textwrap.py
new file mode 100644
index 00000000..5e39618d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/tests/textwrap.py
@@ -0,0 +1,6 @@
+import textwrap
+
+
+def DALS(s):
+    "dedent and left-strip"
+    return textwrap.dedent(s).lstrip()
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/unicode_utils.py b/backend/venv/lib/python3.12/site-packages/setuptools/unicode_utils.py
new file mode 100644
index 00000000..a6e33f2e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/unicode_utils.py
@@ -0,0 +1,102 @@
+import sys
+import unicodedata
+from configparser import RawConfigParser
+
+from .compat import py39
+from .warnings import SetuptoolsDeprecationWarning
+
+
+# HFS Plus uses decomposed UTF-8
+def decompose(path):
+    if isinstance(path, str):
+        return unicodedata.normalize('NFD', path)
+    try:
+        path = path.decode('utf-8')
+        path = unicodedata.normalize('NFD', path)
+        path = path.encode('utf-8')
+    except UnicodeError:
+        pass  # Not UTF-8
+    return path
+
+
+def filesys_decode(path):
+    """
+    Ensure that the given path is decoded,
+    ``None`` when no expected encoding works
+    """
+
+    if isinstance(path, str):
+        return path
+
+    fs_enc = sys.getfilesystemencoding() or 'utf-8'
+    candidates = fs_enc, 'utf-8'
+
+    for enc in candidates:
+        try:
+            return path.decode(enc)
+        except UnicodeDecodeError:
+            continue
+
+    return None
+
+
+def try_encode(string, enc):
+    "turn unicode encoding into a functional routine"
+    try:
+        return string.encode(enc)
+    except UnicodeEncodeError:
+        return None
+
+
+def _read_utf8_with_fallback(file: str, fallback_encoding=py39.LOCALE_ENCODING) -> str:
+    """
+    First try to read the file with UTF-8, if there is an error fallback to a
+    different encoding ("locale" by default). Returns the content of the file.
+    Also useful when reading files that might have been produced by an older version of
+    setuptools.
+    """
+    try:
+        with open(file, "r", encoding="utf-8") as f:
+            return f.read()
+    except UnicodeDecodeError:  # pragma: no cover
+        _Utf8EncodingNeeded.emit(file=file, fallback_encoding=fallback_encoding)
+        with open(file, "r", encoding=fallback_encoding) as f:
+            return f.read()
+
+
+def _cfg_read_utf8_with_fallback(
+    cfg: RawConfigParser, file: str, fallback_encoding=py39.LOCALE_ENCODING
+) -> None:
+    """Same idea as :func:`_read_utf8_with_fallback`, but for the
+    :meth:`RawConfigParser.read` method.
+
+    This method may call ``cfg.clear()``.
+    """
+    try:
+        cfg.read(file, encoding="utf-8")
+    except UnicodeDecodeError:  # pragma: no cover
+        _Utf8EncodingNeeded.emit(file=file, fallback_encoding=fallback_encoding)
+        cfg.clear()
+        cfg.read(file, encoding=fallback_encoding)
+
+
+class _Utf8EncodingNeeded(SetuptoolsDeprecationWarning):
+    _SUMMARY = """
+    `encoding="utf-8"` fails with {file!r}, trying `encoding={fallback_encoding!r}`.
+    """
+
+    _DETAILS = """
+    Fallback behaviour for UTF-8 is considered **deprecated** and future versions of
+    `setuptools` may not implement it.
+
+    Please encode {file!r} with "utf-8" to ensure future builds will succeed.
+
+    If this file was produced by `setuptools` itself, cleaning up the cached files
+    and re-building/re-installing the package with a newer version of `setuptools`
+    (e.g. by updating `build-system.requires` in its `pyproject.toml`)
+    might solve the problem.
+    """
+    # TODO: Add a deadline?
+    #       Will we be able to remove this?
+    #       The question comes to mind mainly because of sdists that have been produced
+    #       by old versions of setuptools and published to PyPI...
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/version.py b/backend/venv/lib/python3.12/site-packages/setuptools/version.py
new file mode 100644
index 00000000..ec253c41
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/version.py
@@ -0,0 +1,6 @@
+from ._importlib import metadata
+
+try:
+    __version__ = metadata.version('setuptools') or '0.dev0+unknown'
+except Exception:
+    __version__ = '0.dev0+unknown'
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/warnings.py b/backend/venv/lib/python3.12/site-packages/setuptools/warnings.py
new file mode 100644
index 00000000..96467787
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/warnings.py
@@ -0,0 +1,110 @@
+"""Provide basic warnings used by setuptools modules.
+
+Using custom classes (other than ``UserWarning``) allow users to set
+``PYTHONWARNINGS`` filters to run tests and prepare for upcoming changes in
+setuptools.
+"""
+
+from __future__ import annotations
+
+import os
+import warnings
+from datetime import date
+from inspect import cleandoc
+from textwrap import indent
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing_extensions import TypeAlias
+
+_DueDate: TypeAlias = tuple[int, int, int]  # time tuple
+_INDENT = 8 * " "
+_TEMPLATE = f"""{80 * '*'}\n{{details}}\n{80 * '*'}"""
+
+
+class SetuptoolsWarning(UserWarning):
+    """Base class in ``setuptools`` warning hierarchy."""
+
+    @classmethod
+    def emit(
+        cls,
+        summary: str | None = None,
+        details: str | None = None,
+        due_date: _DueDate | None = None,
+        see_docs: str | None = None,
+        see_url: str | None = None,
+        stacklevel: int = 2,
+        **kwargs,
+    ) -> None:
+        """Private: reserved for ``setuptools`` internal use only"""
+        # Default values:
+        summary_ = summary or getattr(cls, "_SUMMARY", None) or ""
+        details_ = details or getattr(cls, "_DETAILS", None) or ""
+        due_date = due_date or getattr(cls, "_DUE_DATE", None)
+        docs_ref = see_docs or getattr(cls, "_SEE_DOCS", None)
+        docs_url = docs_ref and f"https://setuptools.pypa.io/en/latest/{docs_ref}"
+        see_url = see_url or getattr(cls, "_SEE_URL", None)
+        due = date(*due_date) if due_date else None
+
+        text = cls._format(summary_, details_, due, see_url or docs_url, kwargs)
+        if due and due < date.today() and _should_enforce():
+            raise cls(text)
+        warnings.warn(text, cls, stacklevel=stacklevel + 1)
+
+    @classmethod
+    def _format(
+        cls,
+        summary: str,
+        details: str,
+        due_date: date | None = None,
+        see_url: str | None = None,
+        format_args: dict | None = None,
+    ) -> str:
+        """Private: reserved for ``setuptools`` internal use only"""
+        today = date.today()
+        summary = cleandoc(summary).format_map(format_args or {})
+        possible_parts = [
+            cleandoc(details).format_map(format_args or {}),
+            (
+                f"\nBy {due_date:%Y-%b-%d}, you need to update your project and remove "
+                "deprecated calls\nor your builds will no longer be supported."
+                if due_date and due_date > today
+                else None
+            ),
+            (
+                "\nThis deprecation is overdue, please update your project and remove "
+                "deprecated\ncalls to avoid build errors in the future."
+                if due_date and due_date < today
+                else None
+            ),
+            (f"\nSee {see_url} for details." if see_url else None),
+        ]
+        parts = [x for x in possible_parts if x]
+        if parts:
+            body = indent(_TEMPLATE.format(details="\n".join(parts)), _INDENT)
+            return "\n".join([summary, "!!\n", body, "\n!!"])
+        return summary
+
+
+class InformationOnly(SetuptoolsWarning):
+    """Currently there is no clear way of displaying messages to the users
+    that use the setuptools backend directly via ``pip``.
+    The only thing that might work is a warning, although it is not the
+    most appropriate tool for the job...
+
+    See pypa/packaging-problems#558.
+    """
+
+
+class SetuptoolsDeprecationWarning(SetuptoolsWarning):
+    """
+    Base class for warning deprecations in ``setuptools``
+
+    This class is not derived from ``DeprecationWarning``, and as such is
+    visible by default.
+    """
+
+
+def _should_enforce():
+    enforce = os.getenv("SETUPTOOLS_ENFORCE_DEPRECATION", "false").lower()
+    return enforce in ("true", "on", "ok", "1")
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/wheel.py b/backend/venv/lib/python3.12/site-packages/setuptools/wheel.py
new file mode 100644
index 00000000..c7ca43b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/wheel.py
@@ -0,0 +1,236 @@
+"""Wheels support."""
+
+import contextlib
+import email
+import functools
+import itertools
+import os
+import posixpath
+import re
+import zipfile
+
+from packaging.tags import sys_tags
+from packaging.utils import canonicalize_name
+from packaging.version import Version as parse_version
+
+import setuptools
+from setuptools.archive_util import _unpack_zipfile_obj
+from setuptools.command.egg_info import _egg_basename, write_requirements
+
+from .unicode_utils import _read_utf8_with_fallback
+
+from distutils.util import get_platform
+
+WHEEL_NAME = re.compile(
+    r"""^(?P<project_name>.+?)-(?P<version>\d.*?)
+    ((-(?P<build>\d.*?))?-(?P<py_version>.+?)-(?P<abi>.+?)-(?P<platform>.+?)
+    )\.whl$""",
+    re.VERBOSE,
+).match
+
+NAMESPACE_PACKAGE_INIT = "__import__('pkg_resources').declare_namespace(__name__)\n"
+
+
+@functools.cache
+def _get_supported_tags():
+    # We calculate the supported tags only once, otherwise calling
+    # this method on thousands of wheels takes seconds instead of
+    # milliseconds.
+    return {(t.interpreter, t.abi, t.platform) for t in sys_tags()}
+
+
+def unpack(src_dir, dst_dir) -> None:
+    """Move everything under `src_dir` to `dst_dir`, and delete the former."""
+    for dirpath, dirnames, filenames in os.walk(src_dir):
+        subdir = os.path.relpath(dirpath, src_dir)
+        for f in filenames:
+            src = os.path.join(dirpath, f)
+            dst = os.path.join(dst_dir, subdir, f)
+            os.renames(src, dst)
+        for n, d in reversed(list(enumerate(dirnames))):
+            src = os.path.join(dirpath, d)
+            dst = os.path.join(dst_dir, subdir, d)
+            if not os.path.exists(dst):
+                # Directory does not exist in destination,
+                # rename it and prune it from os.walk list.
+                os.renames(src, dst)
+                del dirnames[n]
+    # Cleanup.
+    for dirpath, dirnames, filenames in os.walk(src_dir, topdown=True):
+        assert not filenames
+        os.rmdir(dirpath)
+
+
+@contextlib.contextmanager
+def disable_info_traces():
+    """
+    Temporarily disable info traces.
+    """
+    from distutils import log
+
+    saved = log.set_threshold(log.WARN)
+    try:
+        yield
+    finally:
+        log.set_threshold(saved)
+
+
+class Wheel:
+    def __init__(self, filename) -> None:
+        match = WHEEL_NAME(os.path.basename(filename))
+        if match is None:
+            raise ValueError(f'invalid wheel name: {filename!r}')
+        self.filename = filename
+        for k, v in match.groupdict().items():
+            setattr(self, k, v)
+
+    def tags(self):
+        """List tags (py_version, abi, platform) supported by this wheel."""
+        return itertools.product(
+            self.py_version.split('.'),
+            self.abi.split('.'),
+            self.platform.split('.'),
+        )
+
+    def is_compatible(self):
+        """Is the wheel compatible with the current platform?"""
+        return next((True for t in self.tags() if t in _get_supported_tags()), False)
+
+    def egg_name(self):
+        return (
+            _egg_basename(
+                self.project_name,
+                self.version,
+                platform=(None if self.platform == 'any' else get_platform()),
+            )
+            + ".egg"
+        )
+
+    def get_dist_info(self, zf):
+        # find the correct name of the .dist-info dir in the wheel file
+        for member in zf.namelist():
+            dirname = posixpath.dirname(member)
+            if dirname.endswith('.dist-info') and canonicalize_name(dirname).startswith(
+                canonicalize_name(self.project_name)
+            ):
+                return dirname
+        raise ValueError("unsupported wheel format. .dist-info not found")
+
+    def install_as_egg(self, destination_eggdir) -> None:
+        """Install wheel as an egg directory."""
+        with zipfile.ZipFile(self.filename) as zf:
+            self._install_as_egg(destination_eggdir, zf)
+
+    def _install_as_egg(self, destination_eggdir, zf):
+        dist_basename = f'{self.project_name}-{self.version}'
+        dist_info = self.get_dist_info(zf)
+        dist_data = f'{dist_basename}.data'
+        egg_info = os.path.join(destination_eggdir, 'EGG-INFO')
+
+        self._convert_metadata(zf, destination_eggdir, dist_info, egg_info)
+        self._move_data_entries(destination_eggdir, dist_data)
+        self._fix_namespace_packages(egg_info, destination_eggdir)
+
+    @staticmethod
+    def _convert_metadata(zf, destination_eggdir, dist_info, egg_info):
+        import pkg_resources
+
+        def get_metadata(name):
+            with zf.open(posixpath.join(dist_info, name)) as fp:
+                value = fp.read().decode('utf-8')
+                return email.parser.Parser().parsestr(value)
+
+        wheel_metadata = get_metadata('WHEEL')
+        # Check wheel format version is supported.
+        wheel_version = parse_version(wheel_metadata.get('Wheel-Version'))
+        wheel_v1 = parse_version('1.0') <= wheel_version < parse_version('2.0dev0')
+        if not wheel_v1:
+            raise ValueError(f'unsupported wheel format version: {wheel_version}')
+        # Extract to target directory.
+        _unpack_zipfile_obj(zf, destination_eggdir)
+        # Convert metadata.
+        dist_info = os.path.join(destination_eggdir, dist_info)
+        dist = pkg_resources.Distribution.from_location(
+            destination_eggdir,
+            dist_info,
+            metadata=pkg_resources.PathMetadata(destination_eggdir, dist_info),
+        )
+
+        # Note: Evaluate and strip markers now,
+        # as it's difficult to convert back from the syntax:
+        # foobar; "linux" in sys_platform and extra == 'test'
+        def raw_req(req):
+            req.marker = None
+            return str(req)
+
+        install_requires = list(map(raw_req, dist.requires()))
+        extras_require = {
+            extra: [
+                req
+                for req in map(raw_req, dist.requires((extra,)))
+                if req not in install_requires
+            ]
+            for extra in dist.extras
+        }
+        os.rename(dist_info, egg_info)
+        os.rename(
+            os.path.join(egg_info, 'METADATA'),
+            os.path.join(egg_info, 'PKG-INFO'),
+        )
+        setup_dist = setuptools.Distribution(
+            attrs=dict(
+                install_requires=install_requires,
+                extras_require=extras_require,
+            ),
+        )
+        with disable_info_traces():
+            write_requirements(
+                setup_dist.get_command_obj('egg_info'),
+                None,
+                os.path.join(egg_info, 'requires.txt'),
+            )
+
+    @staticmethod
+    def _move_data_entries(destination_eggdir, dist_data):
+        """Move data entries to their correct location."""
+        dist_data = os.path.join(destination_eggdir, dist_data)
+        dist_data_scripts = os.path.join(dist_data, 'scripts')
+        if os.path.exists(dist_data_scripts):
+            egg_info_scripts = os.path.join(destination_eggdir, 'EGG-INFO', 'scripts')
+            os.mkdir(egg_info_scripts)
+            for entry in os.listdir(dist_data_scripts):
+                # Remove bytecode, as it's not properly handled
+                # during easy_install scripts install phase.
+                if entry.endswith('.pyc'):
+                    os.unlink(os.path.join(dist_data_scripts, entry))
+                else:
+                    os.rename(
+                        os.path.join(dist_data_scripts, entry),
+                        os.path.join(egg_info_scripts, entry),
+                    )
+            os.rmdir(dist_data_scripts)
+        for subdir in filter(
+            os.path.exists,
+            (
+                os.path.join(dist_data, d)
+                for d in ('data', 'headers', 'purelib', 'platlib')
+            ),
+        ):
+            unpack(subdir, destination_eggdir)
+        if os.path.exists(dist_data):
+            os.rmdir(dist_data)
+
+    @staticmethod
+    def _fix_namespace_packages(egg_info, destination_eggdir):
+        namespace_packages = os.path.join(egg_info, 'namespace_packages.txt')
+        if os.path.exists(namespace_packages):
+            namespace_packages = _read_utf8_with_fallback(namespace_packages).split()
+
+            for mod in namespace_packages:
+                mod_dir = os.path.join(destination_eggdir, *mod.split('.'))
+                mod_init = os.path.join(mod_dir, '__init__.py')
+                if not os.path.exists(mod_dir):
+                    os.mkdir(mod_dir)
+                if not os.path.exists(mod_init):
+                    with open(mod_init, 'w', encoding="utf-8") as fp:
+                        fp.write(NAMESPACE_PACKAGE_INIT)
diff --git a/backend/venv/lib/python3.12/site-packages/setuptools/windows_support.py b/backend/venv/lib/python3.12/site-packages/setuptools/windows_support.py
new file mode 100644
index 00000000..7a2b53a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/setuptools/windows_support.py
@@ -0,0 +1,30 @@
+import platform
+
+
+def windows_only(func):
+    if platform.system() != 'Windows':
+        return lambda *args, **kwargs: None
+    return func
+
+
+@windows_only
+def hide_file(path: str) -> None:
+    """
+    Set the hidden attribute on a file or directory.
+
+    From https://stackoverflow.com/questions/19622133/
+
+    `path` must be text.
+    """
+    import ctypes
+    import ctypes.wintypes
+
+    SetFileAttributes = ctypes.windll.kernel32.SetFileAttributesW
+    SetFileAttributes.argtypes = ctypes.wintypes.LPWSTR, ctypes.wintypes.DWORD
+    SetFileAttributes.restype = ctypes.wintypes.BOOL
+
+    FILE_ATTRIBUTE_HIDDEN = 0x02
+
+    ret = SetFileAttributes(path, FILE_ATTRIBUTE_HIDDEN)
+    if not ret:
+        raise ctypes.WinError()
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/LICENSE
new file mode 100644
index 00000000..1cc22a5a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/LICENSE
@@ -0,0 +1,18 @@
+Copyright (c) 2010-2024 Benjamin Peterson
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
+the Software, and to permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/METADATA
new file mode 100644
index 00000000..cfde03c2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/METADATA
@@ -0,0 +1,43 @@
+Metadata-Version: 2.1
+Name: six
+Version: 1.17.0
+Summary: Python 2 and 3 compatibility utilities
+Home-page: https://github.com/benjaminp/six
+Author: Benjamin Peterson
+Author-email: benjamin@python.org
+License: MIT
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Programming Language :: Python :: 2
+Classifier: Programming Language :: Python :: 3
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Utilities
+Requires-Python: >=2.7, !=3.0.*, !=3.1.*, !=3.2.*
+License-File: LICENSE
+
+.. image:: https://img.shields.io/pypi/v/six.svg
+   :target: https://pypi.org/project/six/
+   :alt: six on PyPI
+
+.. image:: https://readthedocs.org/projects/six/badge/?version=latest
+   :target: https://six.readthedocs.io/
+   :alt: six's documentation on Read the Docs
+
+.. image:: https://img.shields.io/badge/license-MIT-green.svg
+   :target: https://github.com/benjaminp/six/blob/master/LICENSE
+   :alt: MIT License badge
+
+Six is a Python 2 and 3 compatibility library.  It provides utility functions
+for smoothing over the differences between the Python versions with the goal of
+writing Python code that is compatible on both Python versions.  See the
+documentation for more information on what is provided.
+
+Six supports Python 2.7 and 3.3+.  It is contained in only one Python
+file, so it can be easily copied into your project. (The copyright and license
+notice must be retained.)
+
+Online documentation is at https://six.readthedocs.io/.
+
+Bugs can be reported to https://github.com/benjaminp/six.  The code can also
+be found there.
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/RECORD
new file mode 100644
index 00000000..bb90c1a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/RECORD
@@ -0,0 +1,8 @@
+__pycache__/six.cpython-312.pyc,,
+six-1.17.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+six-1.17.0.dist-info/LICENSE,sha256=Q3W6IOK5xsTnytKUCmKP2Q6VzD1Q7pKq51VxXYuh-9A,1066
+six-1.17.0.dist-info/METADATA,sha256=ViBCB4wnUlSfbYp8htvF3XCAiKe-bYBnLsewcQC3JGg,1658
+six-1.17.0.dist-info/RECORD,,
+six-1.17.0.dist-info/WHEEL,sha256=pxeNX5JdtCe58PUSYP9upmc7jdRPgvT0Gm9kb1SHlVw,109
+six-1.17.0.dist-info/top_level.txt,sha256=_iVH_iYEtEXnD8nYGQYpYFUvkUW9sEO1GYbkeKSAais,4
+six.py,sha256=xRyR9wPT1LNpbJI8tf7CE-BeddkhU5O--sfy-mo5BN8,34703
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/WHEEL
new file mode 100644
index 00000000..104f3874
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/WHEEL
@@ -0,0 +1,6 @@
+Wheel-Version: 1.0
+Generator: setuptools (75.6.0)
+Root-Is-Purelib: true
+Tag: py2-none-any
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/top_level.txt
new file mode 100644
index 00000000..ffe2fce4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six-1.17.0.dist-info/top_level.txt
@@ -0,0 +1 @@
+six
diff --git a/backend/venv/lib/python3.12/site-packages/six.py b/backend/venv/lib/python3.12/site-packages/six.py
new file mode 100644
index 00000000..3de5969b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/six.py
@@ -0,0 +1,1003 @@
+# Copyright (c) 2010-2024 Benjamin Peterson
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+"""Utilities for writing code that runs on Python 2 and 3"""
+
+from __future__ import absolute_import
+
+import functools
+import itertools
+import operator
+import sys
+import types
+
+__author__ = "Benjamin Peterson <benjamin@python.org>"
+__version__ = "1.17.0"
+
+
+# Useful for very coarse version differentiation.
+PY2 = sys.version_info[0] == 2
+PY3 = sys.version_info[0] == 3
+PY34 = sys.version_info[0:2] >= (3, 4)
+
+if PY3:
+    string_types = str,
+    integer_types = int,
+    class_types = type,
+    text_type = str
+    binary_type = bytes
+
+    MAXSIZE = sys.maxsize
+else:
+    string_types = basestring,
+    integer_types = (int, long)
+    class_types = (type, types.ClassType)
+    text_type = unicode
+    binary_type = str
+
+    if sys.platform.startswith("java"):
+        # Jython always uses 32 bits.
+        MAXSIZE = int((1 << 31) - 1)
+    else:
+        # It's possible to have sizeof(long) != sizeof(Py_ssize_t).
+        class X(object):
+
+            def __len__(self):
+                return 1 << 31
+        try:
+            len(X())
+        except OverflowError:
+            # 32-bit
+            MAXSIZE = int((1 << 31) - 1)
+        else:
+            # 64-bit
+            MAXSIZE = int((1 << 63) - 1)
+        del X
+
+if PY34:
+    from importlib.util import spec_from_loader
+else:
+    spec_from_loader = None
+
+
+def _add_doc(func, doc):
+    """Add documentation to a function."""
+    func.__doc__ = doc
+
+
+def _import_module(name):
+    """Import module, returning the module after the last dot."""
+    __import__(name)
+    return sys.modules[name]
+
+
+class _LazyDescr(object):
+
+    def __init__(self, name):
+        self.name = name
+
+    def __get__(self, obj, tp):
+        result = self._resolve()
+        setattr(obj, self.name, result)  # Invokes __set__.
+        try:
+            # This is a bit ugly, but it avoids running this again by
+            # removing this descriptor.
+            delattr(obj.__class__, self.name)
+        except AttributeError:
+            pass
+        return result
+
+
+class MovedModule(_LazyDescr):
+
+    def __init__(self, name, old, new=None):
+        super(MovedModule, self).__init__(name)
+        if PY3:
+            if new is None:
+                new = name
+            self.mod = new
+        else:
+            self.mod = old
+
+    def _resolve(self):
+        return _import_module(self.mod)
+
+    def __getattr__(self, attr):
+        _module = self._resolve()
+        value = getattr(_module, attr)
+        setattr(self, attr, value)
+        return value
+
+
+class _LazyModule(types.ModuleType):
+
+    def __init__(self, name):
+        super(_LazyModule, self).__init__(name)
+        self.__doc__ = self.__class__.__doc__
+
+    def __dir__(self):
+        attrs = ["__doc__", "__name__"]
+        attrs += [attr.name for attr in self._moved_attributes]
+        return attrs
+
+    # Subclasses should override this
+    _moved_attributes = []
+
+
+class MovedAttribute(_LazyDescr):
+
+    def __init__(self, name, old_mod, new_mod, old_attr=None, new_attr=None):
+        super(MovedAttribute, self).__init__(name)
+        if PY3:
+            if new_mod is None:
+                new_mod = name
+            self.mod = new_mod
+            if new_attr is None:
+                if old_attr is None:
+                    new_attr = name
+                else:
+                    new_attr = old_attr
+            self.attr = new_attr
+        else:
+            self.mod = old_mod
+            if old_attr is None:
+                old_attr = name
+            self.attr = old_attr
+
+    def _resolve(self):
+        module = _import_module(self.mod)
+        return getattr(module, self.attr)
+
+
+class _SixMetaPathImporter(object):
+
+    """
+    A meta path importer to import six.moves and its submodules.
+
+    This class implements a PEP302 finder and loader. It should be compatible
+    with Python 2.5 and all existing versions of Python3
+    """
+
+    def __init__(self, six_module_name):
+        self.name = six_module_name
+        self.known_modules = {}
+
+    def _add_module(self, mod, *fullnames):
+        for fullname in fullnames:
+            self.known_modules[self.name + "." + fullname] = mod
+
+    def _get_module(self, fullname):
+        return self.known_modules[self.name + "." + fullname]
+
+    def find_module(self, fullname, path=None):
+        if fullname in self.known_modules:
+            return self
+        return None
+
+    def find_spec(self, fullname, path, target=None):
+        if fullname in self.known_modules:
+            return spec_from_loader(fullname, self)
+        return None
+
+    def __get_module(self, fullname):
+        try:
+            return self.known_modules[fullname]
+        except KeyError:
+            raise ImportError("This loader does not know module " + fullname)
+
+    def load_module(self, fullname):
+        try:
+            # in case of a reload
+            return sys.modules[fullname]
+        except KeyError:
+            pass
+        mod = self.__get_module(fullname)
+        if isinstance(mod, MovedModule):
+            mod = mod._resolve()
+        else:
+            mod.__loader__ = self
+        sys.modules[fullname] = mod
+        return mod
+
+    def is_package(self, fullname):
+        """
+        Return true, if the named module is a package.
+
+        We need this method to get correct spec objects with
+        Python 3.4 (see PEP451)
+        """
+        return hasattr(self.__get_module(fullname), "__path__")
+
+    def get_code(self, fullname):
+        """Return None
+
+        Required, if is_package is implemented"""
+        self.__get_module(fullname)  # eventually raises ImportError
+        return None
+    get_source = get_code  # same as get_code
+
+    def create_module(self, spec):
+        return self.load_module(spec.name)
+
+    def exec_module(self, module):
+        pass
+
+_importer = _SixMetaPathImporter(__name__)
+
+
+class _MovedItems(_LazyModule):
+
+    """Lazy loading of moved objects"""
+    __path__ = []  # mark as package
+
+
+_moved_attributes = [
+    MovedAttribute("cStringIO", "cStringIO", "io", "StringIO"),
+    MovedAttribute("filter", "itertools", "builtins", "ifilter", "filter"),
+    MovedAttribute("filterfalse", "itertools", "itertools", "ifilterfalse", "filterfalse"),
+    MovedAttribute("input", "__builtin__", "builtins", "raw_input", "input"),
+    MovedAttribute("intern", "__builtin__", "sys"),
+    MovedAttribute("map", "itertools", "builtins", "imap", "map"),
+    MovedAttribute("getcwd", "os", "os", "getcwdu", "getcwd"),
+    MovedAttribute("getcwdb", "os", "os", "getcwd", "getcwdb"),
+    MovedAttribute("getoutput", "commands", "subprocess"),
+    MovedAttribute("range", "__builtin__", "builtins", "xrange", "range"),
+    MovedAttribute("reload_module", "__builtin__", "importlib" if PY34 else "imp", "reload"),
+    MovedAttribute("reduce", "__builtin__", "functools"),
+    MovedAttribute("shlex_quote", "pipes", "shlex", "quote"),
+    MovedAttribute("StringIO", "StringIO", "io"),
+    MovedAttribute("UserDict", "UserDict", "collections", "IterableUserDict", "UserDict"),
+    MovedAttribute("UserList", "UserList", "collections"),
+    MovedAttribute("UserString", "UserString", "collections"),
+    MovedAttribute("xrange", "__builtin__", "builtins", "xrange", "range"),
+    MovedAttribute("zip", "itertools", "builtins", "izip", "zip"),
+    MovedAttribute("zip_longest", "itertools", "itertools", "izip_longest", "zip_longest"),
+    MovedModule("builtins", "__builtin__"),
+    MovedModule("configparser", "ConfigParser"),
+    MovedModule("collections_abc", "collections", "collections.abc" if sys.version_info >= (3, 3) else "collections"),
+    MovedModule("copyreg", "copy_reg"),
+    MovedModule("dbm_gnu", "gdbm", "dbm.gnu"),
+    MovedModule("dbm_ndbm", "dbm", "dbm.ndbm"),
+    MovedModule("_dummy_thread", "dummy_thread", "_dummy_thread" if sys.version_info < (3, 9) else "_thread"),
+    MovedModule("http_cookiejar", "cookielib", "http.cookiejar"),
+    MovedModule("http_cookies", "Cookie", "http.cookies"),
+    MovedModule("html_entities", "htmlentitydefs", "html.entities"),
+    MovedModule("html_parser", "HTMLParser", "html.parser"),
+    MovedModule("http_client", "httplib", "http.client"),
+    MovedModule("email_mime_base", "email.MIMEBase", "email.mime.base"),
+    MovedModule("email_mime_image", "email.MIMEImage", "email.mime.image"),
+    MovedModule("email_mime_multipart", "email.MIMEMultipart", "email.mime.multipart"),
+    MovedModule("email_mime_nonmultipart", "email.MIMENonMultipart", "email.mime.nonmultipart"),
+    MovedModule("email_mime_text", "email.MIMEText", "email.mime.text"),
+    MovedModule("BaseHTTPServer", "BaseHTTPServer", "http.server"),
+    MovedModule("CGIHTTPServer", "CGIHTTPServer", "http.server"),
+    MovedModule("SimpleHTTPServer", "SimpleHTTPServer", "http.server"),
+    MovedModule("cPickle", "cPickle", "pickle"),
+    MovedModule("queue", "Queue"),
+    MovedModule("reprlib", "repr"),
+    MovedModule("socketserver", "SocketServer"),
+    MovedModule("_thread", "thread", "_thread"),
+    MovedModule("tkinter", "Tkinter"),
+    MovedModule("tkinter_dialog", "Dialog", "tkinter.dialog"),
+    MovedModule("tkinter_filedialog", "FileDialog", "tkinter.filedialog"),
+    MovedModule("tkinter_scrolledtext", "ScrolledText", "tkinter.scrolledtext"),
+    MovedModule("tkinter_simpledialog", "SimpleDialog", "tkinter.simpledialog"),
+    MovedModule("tkinter_tix", "Tix", "tkinter.tix"),
+    MovedModule("tkinter_ttk", "ttk", "tkinter.ttk"),
+    MovedModule("tkinter_constants", "Tkconstants", "tkinter.constants"),
+    MovedModule("tkinter_dnd", "Tkdnd", "tkinter.dnd"),
+    MovedModule("tkinter_colorchooser", "tkColorChooser",
+                "tkinter.colorchooser"),
+    MovedModule("tkinter_commondialog", "tkCommonDialog",
+                "tkinter.commondialog"),
+    MovedModule("tkinter_tkfiledialog", "tkFileDialog", "tkinter.filedialog"),
+    MovedModule("tkinter_font", "tkFont", "tkinter.font"),
+    MovedModule("tkinter_messagebox", "tkMessageBox", "tkinter.messagebox"),
+    MovedModule("tkinter_tksimpledialog", "tkSimpleDialog",
+                "tkinter.simpledialog"),
+    MovedModule("urllib_parse", __name__ + ".moves.urllib_parse", "urllib.parse"),
+    MovedModule("urllib_error", __name__ + ".moves.urllib_error", "urllib.error"),
+    MovedModule("urllib", __name__ + ".moves.urllib", __name__ + ".moves.urllib"),
+    MovedModule("urllib_robotparser", "robotparser", "urllib.robotparser"),
+    MovedModule("xmlrpc_client", "xmlrpclib", "xmlrpc.client"),
+    MovedModule("xmlrpc_server", "SimpleXMLRPCServer", "xmlrpc.server"),
+]
+# Add windows specific modules.
+if sys.platform == "win32":
+    _moved_attributes += [
+        MovedModule("winreg", "_winreg"),
+    ]
+
+for attr in _moved_attributes:
+    setattr(_MovedItems, attr.name, attr)
+    if isinstance(attr, MovedModule):
+        _importer._add_module(attr, "moves." + attr.name)
+del attr
+
+_MovedItems._moved_attributes = _moved_attributes
+
+moves = _MovedItems(__name__ + ".moves")
+_importer._add_module(moves, "moves")
+
+
+class Module_six_moves_urllib_parse(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_parse"""
+
+
+_urllib_parse_moved_attributes = [
+    MovedAttribute("ParseResult", "urlparse", "urllib.parse"),
+    MovedAttribute("SplitResult", "urlparse", "urllib.parse"),
+    MovedAttribute("parse_qs", "urlparse", "urllib.parse"),
+    MovedAttribute("parse_qsl", "urlparse", "urllib.parse"),
+    MovedAttribute("urldefrag", "urlparse", "urllib.parse"),
+    MovedAttribute("urljoin", "urlparse", "urllib.parse"),
+    MovedAttribute("urlparse", "urlparse", "urllib.parse"),
+    MovedAttribute("urlsplit", "urlparse", "urllib.parse"),
+    MovedAttribute("urlunparse", "urlparse", "urllib.parse"),
+    MovedAttribute("urlunsplit", "urlparse", "urllib.parse"),
+    MovedAttribute("quote", "urllib", "urllib.parse"),
+    MovedAttribute("quote_plus", "urllib", "urllib.parse"),
+    MovedAttribute("unquote", "urllib", "urllib.parse"),
+    MovedAttribute("unquote_plus", "urllib", "urllib.parse"),
+    MovedAttribute("unquote_to_bytes", "urllib", "urllib.parse", "unquote", "unquote_to_bytes"),
+    MovedAttribute("urlencode", "urllib", "urllib.parse"),
+    MovedAttribute("splitquery", "urllib", "urllib.parse"),
+    MovedAttribute("splittag", "urllib", "urllib.parse"),
+    MovedAttribute("splituser", "urllib", "urllib.parse"),
+    MovedAttribute("splitvalue", "urllib", "urllib.parse"),
+    MovedAttribute("uses_fragment", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_netloc", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_params", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_query", "urlparse", "urllib.parse"),
+    MovedAttribute("uses_relative", "urlparse", "urllib.parse"),
+]
+for attr in _urllib_parse_moved_attributes:
+    setattr(Module_six_moves_urllib_parse, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_parse._moved_attributes = _urllib_parse_moved_attributes
+
+_importer._add_module(Module_six_moves_urllib_parse(__name__ + ".moves.urllib_parse"),
+                      "moves.urllib_parse", "moves.urllib.parse")
+
+
+class Module_six_moves_urllib_error(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_error"""
+
+
+_urllib_error_moved_attributes = [
+    MovedAttribute("URLError", "urllib2", "urllib.error"),
+    MovedAttribute("HTTPError", "urllib2", "urllib.error"),
+    MovedAttribute("ContentTooShortError", "urllib", "urllib.error"),
+]
+for attr in _urllib_error_moved_attributes:
+    setattr(Module_six_moves_urllib_error, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_error._moved_attributes = _urllib_error_moved_attributes
+
+_importer._add_module(Module_six_moves_urllib_error(__name__ + ".moves.urllib.error"),
+                      "moves.urllib_error", "moves.urllib.error")
+
+
+class Module_six_moves_urllib_request(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_request"""
+
+
+_urllib_request_moved_attributes = [
+    MovedAttribute("urlopen", "urllib2", "urllib.request"),
+    MovedAttribute("install_opener", "urllib2", "urllib.request"),
+    MovedAttribute("build_opener", "urllib2", "urllib.request"),
+    MovedAttribute("pathname2url", "urllib", "urllib.request"),
+    MovedAttribute("url2pathname", "urllib", "urllib.request"),
+    MovedAttribute("getproxies", "urllib", "urllib.request"),
+    MovedAttribute("Request", "urllib2", "urllib.request"),
+    MovedAttribute("OpenerDirector", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPDefaultErrorHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPRedirectHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPCookieProcessor", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyHandler", "urllib2", "urllib.request"),
+    MovedAttribute("BaseHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPPasswordMgr", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPPasswordMgrWithDefaultRealm", "urllib2", "urllib.request"),
+    MovedAttribute("AbstractBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyBasicAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("AbstractDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("ProxyDigestAuthHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPSHandler", "urllib2", "urllib.request"),
+    MovedAttribute("FileHandler", "urllib2", "urllib.request"),
+    MovedAttribute("FTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("CacheFTPHandler", "urllib2", "urllib.request"),
+    MovedAttribute("UnknownHandler", "urllib2", "urllib.request"),
+    MovedAttribute("HTTPErrorProcessor", "urllib2", "urllib.request"),
+    MovedAttribute("urlretrieve", "urllib", "urllib.request"),
+    MovedAttribute("urlcleanup", "urllib", "urllib.request"),
+    MovedAttribute("proxy_bypass", "urllib", "urllib.request"),
+    MovedAttribute("parse_http_list", "urllib2", "urllib.request"),
+    MovedAttribute("parse_keqv_list", "urllib2", "urllib.request"),
+]
+if sys.version_info[:2] < (3, 14):
+    _urllib_request_moved_attributes.extend(
+        [
+            MovedAttribute("URLopener", "urllib", "urllib.request"),
+            MovedAttribute("FancyURLopener", "urllib", "urllib.request"),
+        ]
+    )
+for attr in _urllib_request_moved_attributes:
+    setattr(Module_six_moves_urllib_request, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_request._moved_attributes = _urllib_request_moved_attributes
+
+_importer._add_module(Module_six_moves_urllib_request(__name__ + ".moves.urllib.request"),
+                      "moves.urllib_request", "moves.urllib.request")
+
+
+class Module_six_moves_urllib_response(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_response"""
+
+
+_urllib_response_moved_attributes = [
+    MovedAttribute("addbase", "urllib", "urllib.response"),
+    MovedAttribute("addclosehook", "urllib", "urllib.response"),
+    MovedAttribute("addinfo", "urllib", "urllib.response"),
+    MovedAttribute("addinfourl", "urllib", "urllib.response"),
+]
+for attr in _urllib_response_moved_attributes:
+    setattr(Module_six_moves_urllib_response, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_response._moved_attributes = _urllib_response_moved_attributes
+
+_importer._add_module(Module_six_moves_urllib_response(__name__ + ".moves.urllib.response"),
+                      "moves.urllib_response", "moves.urllib.response")
+
+
+class Module_six_moves_urllib_robotparser(_LazyModule):
+
+    """Lazy loading of moved objects in six.moves.urllib_robotparser"""
+
+
+_urllib_robotparser_moved_attributes = [
+    MovedAttribute("RobotFileParser", "robotparser", "urllib.robotparser"),
+]
+for attr in _urllib_robotparser_moved_attributes:
+    setattr(Module_six_moves_urllib_robotparser, attr.name, attr)
+del attr
+
+Module_six_moves_urllib_robotparser._moved_attributes = _urllib_robotparser_moved_attributes
+
+_importer._add_module(Module_six_moves_urllib_robotparser(__name__ + ".moves.urllib.robotparser"),
+                      "moves.urllib_robotparser", "moves.urllib.robotparser")
+
+
+class Module_six_moves_urllib(types.ModuleType):
+
+    """Create a six.moves.urllib namespace that resembles the Python 3 namespace"""
+    __path__ = []  # mark as package
+    parse = _importer._get_module("moves.urllib_parse")
+    error = _importer._get_module("moves.urllib_error")
+    request = _importer._get_module("moves.urllib_request")
+    response = _importer._get_module("moves.urllib_response")
+    robotparser = _importer._get_module("moves.urllib_robotparser")
+
+    def __dir__(self):
+        return ['parse', 'error', 'request', 'response', 'robotparser']
+
+_importer._add_module(Module_six_moves_urllib(__name__ + ".moves.urllib"),
+                      "moves.urllib")
+
+
+def add_move(move):
+    """Add an item to six.moves."""
+    setattr(_MovedItems, move.name, move)
+
+
+def remove_move(name):
+    """Remove item from six.moves."""
+    try:
+        delattr(_MovedItems, name)
+    except AttributeError:
+        try:
+            del moves.__dict__[name]
+        except KeyError:
+            raise AttributeError("no such move, %r" % (name,))
+
+
+if PY3:
+    _meth_func = "__func__"
+    _meth_self = "__self__"
+
+    _func_closure = "__closure__"
+    _func_code = "__code__"
+    _func_defaults = "__defaults__"
+    _func_globals = "__globals__"
+else:
+    _meth_func = "im_func"
+    _meth_self = "im_self"
+
+    _func_closure = "func_closure"
+    _func_code = "func_code"
+    _func_defaults = "func_defaults"
+    _func_globals = "func_globals"
+
+
+try:
+    advance_iterator = next
+except NameError:
+    def advance_iterator(it):
+        return it.next()
+next = advance_iterator
+
+
+try:
+    callable = callable
+except NameError:
+    def callable(obj):
+        return any("__call__" in klass.__dict__ for klass in type(obj).__mro__)
+
+
+if PY3:
+    def get_unbound_function(unbound):
+        return unbound
+
+    create_bound_method = types.MethodType
+
+    def create_unbound_method(func, cls):
+        return func
+
+    Iterator = object
+else:
+    def get_unbound_function(unbound):
+        return unbound.im_func
+
+    def create_bound_method(func, obj):
+        return types.MethodType(func, obj, obj.__class__)
+
+    def create_unbound_method(func, cls):
+        return types.MethodType(func, None, cls)
+
+    class Iterator(object):
+
+        def next(self):
+            return type(self).__next__(self)
+
+    callable = callable
+_add_doc(get_unbound_function,
+         """Get the function out of a possibly unbound function""")
+
+
+get_method_function = operator.attrgetter(_meth_func)
+get_method_self = operator.attrgetter(_meth_self)
+get_function_closure = operator.attrgetter(_func_closure)
+get_function_code = operator.attrgetter(_func_code)
+get_function_defaults = operator.attrgetter(_func_defaults)
+get_function_globals = operator.attrgetter(_func_globals)
+
+
+if PY3:
+    def iterkeys(d, **kw):
+        return iter(d.keys(**kw))
+
+    def itervalues(d, **kw):
+        return iter(d.values(**kw))
+
+    def iteritems(d, **kw):
+        return iter(d.items(**kw))
+
+    def iterlists(d, **kw):
+        return iter(d.lists(**kw))
+
+    viewkeys = operator.methodcaller("keys")
+
+    viewvalues = operator.methodcaller("values")
+
+    viewitems = operator.methodcaller("items")
+else:
+    def iterkeys(d, **kw):
+        return d.iterkeys(**kw)
+
+    def itervalues(d, **kw):
+        return d.itervalues(**kw)
+
+    def iteritems(d, **kw):
+        return d.iteritems(**kw)
+
+    def iterlists(d, **kw):
+        return d.iterlists(**kw)
+
+    viewkeys = operator.methodcaller("viewkeys")
+
+    viewvalues = operator.methodcaller("viewvalues")
+
+    viewitems = operator.methodcaller("viewitems")
+
+_add_doc(iterkeys, "Return an iterator over the keys of a dictionary.")
+_add_doc(itervalues, "Return an iterator over the values of a dictionary.")
+_add_doc(iteritems,
+         "Return an iterator over the (key, value) pairs of a dictionary.")
+_add_doc(iterlists,
+         "Return an iterator over the (key, [values]) pairs of a dictionary.")
+
+
+if PY3:
+    def b(s):
+        return s.encode("latin-1")
+
+    def u(s):
+        return s
+    unichr = chr
+    import struct
+    int2byte = struct.Struct(">B").pack
+    del struct
+    byte2int = operator.itemgetter(0)
+    indexbytes = operator.getitem
+    iterbytes = iter
+    import io
+    StringIO = io.StringIO
+    BytesIO = io.BytesIO
+    del io
+    _assertCountEqual = "assertCountEqual"
+    if sys.version_info[1] <= 1:
+        _assertRaisesRegex = "assertRaisesRegexp"
+        _assertRegex = "assertRegexpMatches"
+        _assertNotRegex = "assertNotRegexpMatches"
+    else:
+        _assertRaisesRegex = "assertRaisesRegex"
+        _assertRegex = "assertRegex"
+        _assertNotRegex = "assertNotRegex"
+else:
+    def b(s):
+        return s
+    # Workaround for standalone backslash
+
+    def u(s):
+        return unicode(s.replace(r'\\', r'\\\\'), "unicode_escape")
+    unichr = unichr
+    int2byte = chr
+
+    def byte2int(bs):
+        return ord(bs[0])
+
+    def indexbytes(buf, i):
+        return ord(buf[i])
+    iterbytes = functools.partial(itertools.imap, ord)
+    import StringIO
+    StringIO = BytesIO = StringIO.StringIO
+    _assertCountEqual = "assertItemsEqual"
+    _assertRaisesRegex = "assertRaisesRegexp"
+    _assertRegex = "assertRegexpMatches"
+    _assertNotRegex = "assertNotRegexpMatches"
+_add_doc(b, """Byte literal""")
+_add_doc(u, """Text literal""")
+
+
+def assertCountEqual(self, *args, **kwargs):
+    return getattr(self, _assertCountEqual)(*args, **kwargs)
+
+
+def assertRaisesRegex(self, *args, **kwargs):
+    return getattr(self, _assertRaisesRegex)(*args, **kwargs)
+
+
+def assertRegex(self, *args, **kwargs):
+    return getattr(self, _assertRegex)(*args, **kwargs)
+
+
+def assertNotRegex(self, *args, **kwargs):
+    return getattr(self, _assertNotRegex)(*args, **kwargs)
+
+
+if PY3:
+    exec_ = getattr(moves.builtins, "exec")
+
+    def reraise(tp, value, tb=None):
+        try:
+            if value is None:
+                value = tp()
+            if value.__traceback__ is not tb:
+                raise value.with_traceback(tb)
+            raise value
+        finally:
+            value = None
+            tb = None
+
+else:
+    def exec_(_code_, _globs_=None, _locs_=None):
+        """Execute code in a namespace."""
+        if _globs_ is None:
+            frame = sys._getframe(1)
+            _globs_ = frame.f_globals
+            if _locs_ is None:
+                _locs_ = frame.f_locals
+            del frame
+        elif _locs_ is None:
+            _locs_ = _globs_
+        exec("""exec _code_ in _globs_, _locs_""")
+
+    exec_("""def reraise(tp, value, tb=None):
+    try:
+        raise tp, value, tb
+    finally:
+        tb = None
+""")
+
+
+if sys.version_info[:2] > (3,):
+    exec_("""def raise_from(value, from_value):
+    try:
+        raise value from from_value
+    finally:
+        value = None
+""")
+else:
+    def raise_from(value, from_value):
+        raise value
+
+
+print_ = getattr(moves.builtins, "print", None)
+if print_ is None:
+    def print_(*args, **kwargs):
+        """The new-style print function for Python 2.4 and 2.5."""
+        fp = kwargs.pop("file", sys.stdout)
+        if fp is None:
+            return
+
+        def write(data):
+            if not isinstance(data, basestring):
+                data = str(data)
+            # If the file has an encoding, encode unicode with it.
+            if (isinstance(fp, file) and
+                    isinstance(data, unicode) and
+                    fp.encoding is not None):
+                errors = getattr(fp, "errors", None)
+                if errors is None:
+                    errors = "strict"
+                data = data.encode(fp.encoding, errors)
+            fp.write(data)
+        want_unicode = False
+        sep = kwargs.pop("sep", None)
+        if sep is not None:
+            if isinstance(sep, unicode):
+                want_unicode = True
+            elif not isinstance(sep, str):
+                raise TypeError("sep must be None or a string")
+        end = kwargs.pop("end", None)
+        if end is not None:
+            if isinstance(end, unicode):
+                want_unicode = True
+            elif not isinstance(end, str):
+                raise TypeError("end must be None or a string")
+        if kwargs:
+            raise TypeError("invalid keyword arguments to print()")
+        if not want_unicode:
+            for arg in args:
+                if isinstance(arg, unicode):
+                    want_unicode = True
+                    break
+        if want_unicode:
+            newline = unicode("\n")
+            space = unicode(" ")
+        else:
+            newline = "\n"
+            space = " "
+        if sep is None:
+            sep = space
+        if end is None:
+            end = newline
+        for i, arg in enumerate(args):
+            if i:
+                write(sep)
+            write(arg)
+        write(end)
+if sys.version_info[:2] < (3, 3):
+    _print = print_
+
+    def print_(*args, **kwargs):
+        fp = kwargs.get("file", sys.stdout)
+        flush = kwargs.pop("flush", False)
+        _print(*args, **kwargs)
+        if flush and fp is not None:
+            fp.flush()
+
+_add_doc(reraise, """Reraise an exception.""")
+
+if sys.version_info[0:2] < (3, 4):
+    # This does exactly the same what the :func:`py3:functools.update_wrapper`
+    # function does on Python versions after 3.2. It sets the ``__wrapped__``
+    # attribute on ``wrapper`` object and it doesn't raise an error if any of
+    # the attributes mentioned in ``assigned`` and ``updated`` are missing on
+    # ``wrapped`` object.
+    def _update_wrapper(wrapper, wrapped,
+                        assigned=functools.WRAPPER_ASSIGNMENTS,
+                        updated=functools.WRAPPER_UPDATES):
+        for attr in assigned:
+            try:
+                value = getattr(wrapped, attr)
+            except AttributeError:
+                continue
+            else:
+                setattr(wrapper, attr, value)
+        for attr in updated:
+            getattr(wrapper, attr).update(getattr(wrapped, attr, {}))
+        wrapper.__wrapped__ = wrapped
+        return wrapper
+    _update_wrapper.__doc__ = functools.update_wrapper.__doc__
+
+    def wraps(wrapped, assigned=functools.WRAPPER_ASSIGNMENTS,
+              updated=functools.WRAPPER_UPDATES):
+        return functools.partial(_update_wrapper, wrapped=wrapped,
+                                 assigned=assigned, updated=updated)
+    wraps.__doc__ = functools.wraps.__doc__
+
+else:
+    wraps = functools.wraps
+
+
+def with_metaclass(meta, *bases):
+    """Create a base class with a metaclass."""
+    # This requires a bit of explanation: the basic idea is to make a dummy
+    # metaclass for one level of class instantiation that replaces itself with
+    # the actual metaclass.
+    class metaclass(type):
+
+        def __new__(cls, name, this_bases, d):
+            if sys.version_info[:2] >= (3, 7):
+                # This version introduced PEP 560 that requires a bit
+                # of extra care (we mimic what is done by __build_class__).
+                resolved_bases = types.resolve_bases(bases)
+                if resolved_bases is not bases:
+                    d['__orig_bases__'] = bases
+            else:
+                resolved_bases = bases
+            return meta(name, resolved_bases, d)
+
+        @classmethod
+        def __prepare__(cls, name, this_bases):
+            return meta.__prepare__(name, bases)
+    return type.__new__(metaclass, 'temporary_class', (), {})
+
+
+def add_metaclass(metaclass):
+    """Class decorator for creating a class with a metaclass."""
+    def wrapper(cls):
+        orig_vars = cls.__dict__.copy()
+        slots = orig_vars.get('__slots__')
+        if slots is not None:
+            if isinstance(slots, str):
+                slots = [slots]
+            for slots_var in slots:
+                orig_vars.pop(slots_var)
+        orig_vars.pop('__dict__', None)
+        orig_vars.pop('__weakref__', None)
+        if hasattr(cls, '__qualname__'):
+            orig_vars['__qualname__'] = cls.__qualname__
+        return metaclass(cls.__name__, cls.__bases__, orig_vars)
+    return wrapper
+
+
+def ensure_binary(s, encoding='utf-8', errors='strict'):
+    """Coerce **s** to six.binary_type.
+
+    For Python 2:
+      - `unicode` -> encoded to `str`
+      - `str` -> `str`
+
+    For Python 3:
+      - `str` -> encoded to `bytes`
+      - `bytes` -> `bytes`
+    """
+    if isinstance(s, binary_type):
+        return s
+    if isinstance(s, text_type):
+        return s.encode(encoding, errors)
+    raise TypeError("not expecting type '%s'" % type(s))
+
+
+def ensure_str(s, encoding='utf-8', errors='strict'):
+    """Coerce *s* to `str`.
+
+    For Python 2:
+      - `unicode` -> encoded to `str`
+      - `str` -> `str`
+
+    For Python 3:
+      - `str` -> `str`
+      - `bytes` -> decoded to `str`
+    """
+    # Optimization: Fast return for the common case.
+    if type(s) is str:
+        return s
+    if PY2 and isinstance(s, text_type):
+        return s.encode(encoding, errors)
+    elif PY3 and isinstance(s, binary_type):
+        return s.decode(encoding, errors)
+    elif not isinstance(s, (text_type, binary_type)):
+        raise TypeError("not expecting type '%s'" % type(s))
+    return s
+
+
+def ensure_text(s, encoding='utf-8', errors='strict'):
+    """Coerce *s* to six.text_type.
+
+    For Python 2:
+      - `unicode` -> `unicode`
+      - `str` -> `unicode`
+
+    For Python 3:
+      - `str` -> `str`
+      - `bytes` -> decoded to `str`
+    """
+    if isinstance(s, binary_type):
+        return s.decode(encoding, errors)
+    elif isinstance(s, text_type):
+        return s
+    else:
+        raise TypeError("not expecting type '%s'" % type(s))
+
+
+def python_2_unicode_compatible(klass):
+    """
+    A class decorator that defines __unicode__ and __str__ methods under Python 2.
+    Under Python 3 it does nothing.
+
+    To support Python 2 and 3 with a single code base, define a __str__ method
+    returning text and apply this decorator to the class.
+    """
+    if PY2:
+        if '__str__' not in klass.__dict__:
+            raise ValueError("@python_2_unicode_compatible cannot be applied "
+                             "to %s because it doesn't define __str__()." %
+                             klass.__name__)
+        klass.__unicode__ = klass.__str__
+        klass.__str__ = lambda self: self.__unicode__().encode('utf-8')
+    return klass
+
+
+# Complete the moves implementation.
+# This code is at the end of this module to speed up module loading.
+# Turn this module into a package.
+__path__ = []  # required for PEP 302 and PEP 451
+__package__ = __name__  # see PEP 366 @ReservedAssignment
+if globals().get("__spec__") is not None:
+    __spec__.submodule_search_locations = []  # PEP 451 @UndefinedVariable
+# Remove other six meta path importers, since they cause problems. This can
+# happen if six is removed from sys.modules and then reloaded. (Setuptools does
+# this for some reason.)
+if sys.meta_path:
+    for i, importer in enumerate(sys.meta_path):
+        # Here's some real nastiness: Another "instance" of the six module might
+        # be floating around. Therefore, we can't use isinstance() to check for
+        # the six meta path importer, since the other six instance will have
+        # inserted an importer with different class.
+        if (type(importer).__name__ == "_SixMetaPathImporter" and
+                importer.name == __name__):
+            del sys.meta_path[i]
+            break
+    del i, importer
+# Finally, add the importer to the meta path import hook.
+sys.meta_path.append(_importer)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__init__.py
new file mode 100644
index 00000000..d67636cf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__init__.py
@@ -0,0 +1,294 @@
+# __init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+
+from . import util as _util
+from .engine import AdaptedConnection as AdaptedConnection
+from .engine import BaseRow as BaseRow
+from .engine import BindTyping as BindTyping
+from .engine import ChunkedIteratorResult as ChunkedIteratorResult
+from .engine import Compiled as Compiled
+from .engine import Connection as Connection
+from .engine import create_engine as create_engine
+from .engine import create_mock_engine as create_mock_engine
+from .engine import create_pool_from_url as create_pool_from_url
+from .engine import CreateEnginePlugin as CreateEnginePlugin
+from .engine import CursorResult as CursorResult
+from .engine import Dialect as Dialect
+from .engine import Engine as Engine
+from .engine import engine_from_config as engine_from_config
+from .engine import ExceptionContext as ExceptionContext
+from .engine import ExecutionContext as ExecutionContext
+from .engine import FrozenResult as FrozenResult
+from .engine import Inspector as Inspector
+from .engine import IteratorResult as IteratorResult
+from .engine import make_url as make_url
+from .engine import MappingResult as MappingResult
+from .engine import MergedResult as MergedResult
+from .engine import NestedTransaction as NestedTransaction
+from .engine import Result as Result
+from .engine import result_tuple as result_tuple
+from .engine import ResultProxy as ResultProxy
+from .engine import RootTransaction as RootTransaction
+from .engine import Row as Row
+from .engine import RowMapping as RowMapping
+from .engine import ScalarResult as ScalarResult
+from .engine import Transaction as Transaction
+from .engine import TwoPhaseTransaction as TwoPhaseTransaction
+from .engine import TypeCompiler as TypeCompiler
+from .engine import URL as URL
+from .inspection import inspect as inspect
+from .pool import AssertionPool as AssertionPool
+from .pool import AsyncAdaptedQueuePool as AsyncAdaptedQueuePool
+from .pool import (
+    FallbackAsyncAdaptedQueuePool as FallbackAsyncAdaptedQueuePool,
+)
+from .pool import NullPool as NullPool
+from .pool import Pool as Pool
+from .pool import PoolProxiedConnection as PoolProxiedConnection
+from .pool import PoolResetState as PoolResetState
+from .pool import QueuePool as QueuePool
+from .pool import SingletonThreadPool as SingletonThreadPool
+from .pool import StaticPool as StaticPool
+from .schema import BaseDDLElement as BaseDDLElement
+from .schema import BLANK_SCHEMA as BLANK_SCHEMA
+from .schema import CheckConstraint as CheckConstraint
+from .schema import Column as Column
+from .schema import ColumnDefault as ColumnDefault
+from .schema import Computed as Computed
+from .schema import Constraint as Constraint
+from .schema import DDL as DDL
+from .schema import DDLElement as DDLElement
+from .schema import DefaultClause as DefaultClause
+from .schema import ExecutableDDLElement as ExecutableDDLElement
+from .schema import FetchedValue as FetchedValue
+from .schema import ForeignKey as ForeignKey
+from .schema import ForeignKeyConstraint as ForeignKeyConstraint
+from .schema import Identity as Identity
+from .schema import Index as Index
+from .schema import insert_sentinel as insert_sentinel
+from .schema import MetaData as MetaData
+from .schema import PrimaryKeyConstraint as PrimaryKeyConstraint
+from .schema import Sequence as Sequence
+from .schema import Table as Table
+from .schema import UniqueConstraint as UniqueConstraint
+from .sql import ColumnExpressionArgument as ColumnExpressionArgument
+from .sql import NotNullable as NotNullable
+from .sql import Nullable as Nullable
+from .sql import SelectLabelStyle as SelectLabelStyle
+from .sql.expression import Alias as Alias
+from .sql.expression import alias as alias
+from .sql.expression import AliasedReturnsRows as AliasedReturnsRows
+from .sql.expression import all_ as all_
+from .sql.expression import and_ as and_
+from .sql.expression import any_ as any_
+from .sql.expression import asc as asc
+from .sql.expression import between as between
+from .sql.expression import BinaryExpression as BinaryExpression
+from .sql.expression import bindparam as bindparam
+from .sql.expression import BindParameter as BindParameter
+from .sql.expression import bitwise_not as bitwise_not
+from .sql.expression import BooleanClauseList as BooleanClauseList
+from .sql.expression import CacheKey as CacheKey
+from .sql.expression import Case as Case
+from .sql.expression import case as case
+from .sql.expression import Cast as Cast
+from .sql.expression import cast as cast
+from .sql.expression import ClauseElement as ClauseElement
+from .sql.expression import ClauseList as ClauseList
+from .sql.expression import collate as collate
+from .sql.expression import CollectionAggregate as CollectionAggregate
+from .sql.expression import column as column
+from .sql.expression import ColumnClause as ColumnClause
+from .sql.expression import ColumnCollection as ColumnCollection
+from .sql.expression import ColumnElement as ColumnElement
+from .sql.expression import ColumnOperators as ColumnOperators
+from .sql.expression import CompoundSelect as CompoundSelect
+from .sql.expression import CTE as CTE
+from .sql.expression import cte as cte
+from .sql.expression import custom_op as custom_op
+from .sql.expression import Delete as Delete
+from .sql.expression import delete as delete
+from .sql.expression import desc as desc
+from .sql.expression import distinct as distinct
+from .sql.expression import except_ as except_
+from .sql.expression import except_all as except_all
+from .sql.expression import Executable as Executable
+from .sql.expression import Exists as Exists
+from .sql.expression import exists as exists
+from .sql.expression import Extract as Extract
+from .sql.expression import extract as extract
+from .sql.expression import false as false
+from .sql.expression import False_ as False_
+from .sql.expression import FromClause as FromClause
+from .sql.expression import FromGrouping as FromGrouping
+from .sql.expression import func as func
+from .sql.expression import funcfilter as funcfilter
+from .sql.expression import Function as Function
+from .sql.expression import FunctionElement as FunctionElement
+from .sql.expression import FunctionFilter as FunctionFilter
+from .sql.expression import GenerativeSelect as GenerativeSelect
+from .sql.expression import Grouping as Grouping
+from .sql.expression import HasCTE as HasCTE
+from .sql.expression import HasPrefixes as HasPrefixes
+from .sql.expression import HasSuffixes as HasSuffixes
+from .sql.expression import Insert as Insert
+from .sql.expression import insert as insert
+from .sql.expression import intersect as intersect
+from .sql.expression import intersect_all as intersect_all
+from .sql.expression import Join as Join
+from .sql.expression import join as join
+from .sql.expression import Label as Label
+from .sql.expression import label as label
+from .sql.expression import LABEL_STYLE_DEFAULT as LABEL_STYLE_DEFAULT
+from .sql.expression import (
+    LABEL_STYLE_DISAMBIGUATE_ONLY as LABEL_STYLE_DISAMBIGUATE_ONLY,
+)
+from .sql.expression import LABEL_STYLE_NONE as LABEL_STYLE_NONE
+from .sql.expression import (
+    LABEL_STYLE_TABLENAME_PLUS_COL as LABEL_STYLE_TABLENAME_PLUS_COL,
+)
+from .sql.expression import lambda_stmt as lambda_stmt
+from .sql.expression import LambdaElement as LambdaElement
+from .sql.expression import Lateral as Lateral
+from .sql.expression import lateral as lateral
+from .sql.expression import literal as literal
+from .sql.expression import literal_column as literal_column
+from .sql.expression import modifier as modifier
+from .sql.expression import not_ as not_
+from .sql.expression import Null as Null
+from .sql.expression import null as null
+from .sql.expression import nulls_first as nulls_first
+from .sql.expression import nulls_last as nulls_last
+from .sql.expression import nullsfirst as nullsfirst
+from .sql.expression import nullslast as nullslast
+from .sql.expression import Operators as Operators
+from .sql.expression import or_ as or_
+from .sql.expression import outerjoin as outerjoin
+from .sql.expression import outparam as outparam
+from .sql.expression import Over as Over
+from .sql.expression import over as over
+from .sql.expression import quoted_name as quoted_name
+from .sql.expression import ReleaseSavepointClause as ReleaseSavepointClause
+from .sql.expression import ReturnsRows as ReturnsRows
+from .sql.expression import (
+    RollbackToSavepointClause as RollbackToSavepointClause,
+)
+from .sql.expression import SavepointClause as SavepointClause
+from .sql.expression import ScalarSelect as ScalarSelect
+from .sql.expression import Select as Select
+from .sql.expression import select as select
+from .sql.expression import Selectable as Selectable
+from .sql.expression import SelectBase as SelectBase
+from .sql.expression import SQLColumnExpression as SQLColumnExpression
+from .sql.expression import StatementLambdaElement as StatementLambdaElement
+from .sql.expression import Subquery as Subquery
+from .sql.expression import table as table
+from .sql.expression import TableClause as TableClause
+from .sql.expression import TableSample as TableSample
+from .sql.expression import tablesample as tablesample
+from .sql.expression import TableValuedAlias as TableValuedAlias
+from .sql.expression import text as text
+from .sql.expression import TextAsFrom as TextAsFrom
+from .sql.expression import TextClause as TextClause
+from .sql.expression import TextualSelect as TextualSelect
+from .sql.expression import true as true
+from .sql.expression import True_ as True_
+from .sql.expression import try_cast as try_cast
+from .sql.expression import TryCast as TryCast
+from .sql.expression import Tuple as Tuple
+from .sql.expression import tuple_ as tuple_
+from .sql.expression import type_coerce as type_coerce
+from .sql.expression import TypeClause as TypeClause
+from .sql.expression import TypeCoerce as TypeCoerce
+from .sql.expression import UnaryExpression as UnaryExpression
+from .sql.expression import union as union
+from .sql.expression import union_all as union_all
+from .sql.expression import Update as Update
+from .sql.expression import update as update
+from .sql.expression import UpdateBase as UpdateBase
+from .sql.expression import Values as Values
+from .sql.expression import values as values
+from .sql.expression import ValuesBase as ValuesBase
+from .sql.expression import Visitable as Visitable
+from .sql.expression import within_group as within_group
+from .sql.expression import WithinGroup as WithinGroup
+from .types import ARRAY as ARRAY
+from .types import BIGINT as BIGINT
+from .types import BigInteger as BigInteger
+from .types import BINARY as BINARY
+from .types import BLOB as BLOB
+from .types import BOOLEAN as BOOLEAN
+from .types import Boolean as Boolean
+from .types import CHAR as CHAR
+from .types import CLOB as CLOB
+from .types import DATE as DATE
+from .types import Date as Date
+from .types import DATETIME as DATETIME
+from .types import DateTime as DateTime
+from .types import DECIMAL as DECIMAL
+from .types import DOUBLE as DOUBLE
+from .types import Double as Double
+from .types import DOUBLE_PRECISION as DOUBLE_PRECISION
+from .types import Enum as Enum
+from .types import FLOAT as FLOAT
+from .types import Float as Float
+from .types import INT as INT
+from .types import INTEGER as INTEGER
+from .types import Integer as Integer
+from .types import Interval as Interval
+from .types import JSON as JSON
+from .types import LargeBinary as LargeBinary
+from .types import NCHAR as NCHAR
+from .types import NUMERIC as NUMERIC
+from .types import Numeric as Numeric
+from .types import NVARCHAR as NVARCHAR
+from .types import PickleType as PickleType
+from .types import REAL as REAL
+from .types import SMALLINT as SMALLINT
+from .types import SmallInteger as SmallInteger
+from .types import String as String
+from .types import TEXT as TEXT
+from .types import Text as Text
+from .types import TIME as TIME
+from .types import Time as Time
+from .types import TIMESTAMP as TIMESTAMP
+from .types import TupleType as TupleType
+from .types import TypeDecorator as TypeDecorator
+from .types import Unicode as Unicode
+from .types import UnicodeText as UnicodeText
+from .types import UUID as UUID
+from .types import Uuid as Uuid
+from .types import VARBINARY as VARBINARY
+from .types import VARCHAR as VARCHAR
+
+__version__ = "2.0.38"
+
+
+def __go(lcls: Any) -> None:
+    _util.preloaded.import_prefix("sqlalchemy")
+
+    from . import exc
+
+    exc._version_token = "".join(__version__.split(".")[0:2])
+
+
+__go(locals())
+
+
+def __getattr__(name: str) -> Any:
+    if name == "SingleonThreadPool":
+        _util.warn_deprecated(
+            "SingleonThreadPool was a typo in the v2 series. "
+            "Please use the correct SingletonThreadPool name.",
+            "2.0.24",
+        )
+        return SingletonThreadPool
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9b088b20
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/events.cpython-312.pyc
new file mode 100644
index 00000000..06c6e405
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/exc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/exc.cpython-312.pyc
new file mode 100644
index 00000000..b6ca8f43
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/exc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/inspection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/inspection.cpython-312.pyc
new file mode 100644
index 00000000..966b074b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/inspection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/log.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/log.cpython-312.pyc
new file mode 100644
index 00000000..d0fe5c52
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/log.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/schema.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/schema.cpython-312.pyc
new file mode 100644
index 00000000..d8c59a5a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/schema.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..8e23e608
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__init__.py
new file mode 100644
index 00000000..43cd1035
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__init__.py
@@ -0,0 +1,18 @@
+# connectors/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+from ..engine.interfaces import Dialect
+
+
+class Connector(Dialect):
+    """Base class for dialect mixins, for DBAPIs that work
+    across entirely different database backends.
+
+    Currently the only such mixin is pyodbc.
+
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..afdc5856
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/aioodbc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/aioodbc.cpython-312.pyc
new file mode 100644
index 00000000..4ff7eba8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/aioodbc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/asyncio.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/asyncio.cpython-312.pyc
new file mode 100644
index 00000000..b467b9f4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/asyncio.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/pyodbc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/pyodbc.cpython-312.pyc
new file mode 100644
index 00000000..a940fce6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/__pycache__/pyodbc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/aioodbc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/aioodbc.py
new file mode 100644
index 00000000..39b2a8a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/aioodbc.py
@@ -0,0 +1,174 @@
+# connectors/aioodbc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from .asyncio import AsyncAdapt_dbapi_connection
+from .asyncio import AsyncAdapt_dbapi_cursor
+from .asyncio import AsyncAdapt_dbapi_ss_cursor
+from .asyncio import AsyncAdaptFallback_dbapi_connection
+from .pyodbc import PyODBCConnector
+from .. import pool
+from .. import util
+from ..util.concurrency import await_fallback
+from ..util.concurrency import await_only
+
+if TYPE_CHECKING:
+    from ..engine.interfaces import ConnectArgsType
+    from ..engine.url import URL
+
+
+class AsyncAdapt_aioodbc_cursor(AsyncAdapt_dbapi_cursor):
+    __slots__ = ()
+
+    def setinputsizes(self, *inputsizes):
+        # see https://github.com/aio-libs/aioodbc/issues/451
+        return self._cursor._impl.setinputsizes(*inputsizes)
+
+        # how it's supposed to work
+        # return self.await_(self._cursor.setinputsizes(*inputsizes))
+
+
+class AsyncAdapt_aioodbc_ss_cursor(
+    AsyncAdapt_aioodbc_cursor, AsyncAdapt_dbapi_ss_cursor
+):
+    __slots__ = ()
+
+
+class AsyncAdapt_aioodbc_connection(AsyncAdapt_dbapi_connection):
+    _cursor_cls = AsyncAdapt_aioodbc_cursor
+    _ss_cursor_cls = AsyncAdapt_aioodbc_ss_cursor
+    __slots__ = ()
+
+    @property
+    def autocommit(self):
+        return self._connection.autocommit
+
+    @autocommit.setter
+    def autocommit(self, value):
+        # https://github.com/aio-libs/aioodbc/issues/448
+        # self._connection.autocommit = value
+
+        self._connection._conn.autocommit = value
+
+    def cursor(self, server_side=False):
+        # aioodbc sets connection=None when closed and just fails with
+        # AttributeError here.  Here we use the same ProgrammingError +
+        # message that pyodbc uses, so it triggers is_disconnect() as well.
+        if self._connection.closed:
+            raise self.dbapi.ProgrammingError(
+                "Attempt to use a closed connection."
+            )
+        return super().cursor(server_side=server_side)
+
+    def rollback(self):
+        # aioodbc sets connection=None when closed and just fails with
+        # AttributeError here.  should be a no-op
+        if not self._connection.closed:
+            super().rollback()
+
+    def commit(self):
+        # aioodbc sets connection=None when closed and just fails with
+        # AttributeError here.  should be a no-op
+        if not self._connection.closed:
+            super().commit()
+
+    def close(self):
+        # aioodbc sets connection=None when closed and just fails with
+        # AttributeError here.  should be a no-op
+        if not self._connection.closed:
+            super().close()
+
+
+class AsyncAdaptFallback_aioodbc_connection(
+    AsyncAdaptFallback_dbapi_connection, AsyncAdapt_aioodbc_connection
+):
+    __slots__ = ()
+
+
+class AsyncAdapt_aioodbc_dbapi:
+    def __init__(self, aioodbc, pyodbc):
+        self.aioodbc = aioodbc
+        self.pyodbc = pyodbc
+        self.paramstyle = pyodbc.paramstyle
+        self._init_dbapi_attributes()
+        self.Cursor = AsyncAdapt_dbapi_cursor
+        self.version = pyodbc.version
+
+    def _init_dbapi_attributes(self):
+        for name in (
+            "Warning",
+            "Error",
+            "InterfaceError",
+            "DataError",
+            "DatabaseError",
+            "OperationalError",
+            "InterfaceError",
+            "IntegrityError",
+            "ProgrammingError",
+            "InternalError",
+            "NotSupportedError",
+            "NUMBER",
+            "STRING",
+            "DATETIME",
+            "BINARY",
+            "Binary",
+            "BinaryNull",
+            "SQL_VARCHAR",
+            "SQL_WVARCHAR",
+        ):
+            setattr(self, name, getattr(self.pyodbc, name))
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop("async_creator_fn", self.aioodbc.connect)
+
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_aioodbc_connection(
+                self,
+                await_fallback(creator_fn(*arg, **kw)),
+            )
+        else:
+            return AsyncAdapt_aioodbc_connection(
+                self,
+                await_only(creator_fn(*arg, **kw)),
+            )
+
+
+class aiodbcConnector(PyODBCConnector):
+    is_async = True
+    supports_statement_cache = True
+
+    supports_server_side_cursors = True
+
+    @classmethod
+    def import_dbapi(cls):
+        return AsyncAdapt_aioodbc_dbapi(
+            __import__("aioodbc"), __import__("pyodbc")
+        )
+
+    def create_connect_args(self, url: URL) -> ConnectArgsType:
+        arg, kw = super().create_connect_args(url)
+        if arg and arg[0]:
+            kw["dsn"] = arg[0]
+
+        return (), kw
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if util.asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def get_driver_connection(self, connection):
+        return connection._connection
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/asyncio.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/asyncio.py
new file mode 100644
index 00000000..c4f0d715
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/asyncio.py
@@ -0,0 +1,213 @@
+# connectors/asyncio.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""generic asyncio-adapted versions of DBAPI connection and cursor"""
+
+from __future__ import annotations
+
+import collections
+
+from ..engine import AdaptedConnection
+from ..util.concurrency import asyncio
+from ..util.concurrency import await_fallback
+from ..util.concurrency import await_only
+
+
+class AsyncAdapt_dbapi_cursor:
+    server_side = False
+    __slots__ = (
+        "_adapt_connection",
+        "_connection",
+        "await_",
+        "_cursor",
+        "_rows",
+    )
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor()
+        self._cursor = self._aenter_cursor(cursor)
+
+        if not self.server_side:
+            self._rows = collections.deque()
+
+    def _aenter_cursor(self, cursor):
+        return self.await_(cursor.__aenter__())
+
+    @property
+    def description(self):
+        return self._cursor.description
+
+    @property
+    def rowcount(self):
+        return self._cursor.rowcount
+
+    @property
+    def arraysize(self):
+        return self._cursor.arraysize
+
+    @arraysize.setter
+    def arraysize(self, value):
+        self._cursor.arraysize = value
+
+    @property
+    def lastrowid(self):
+        return self._cursor.lastrowid
+
+    def close(self):
+        # note we aren't actually closing the cursor here,
+        # we are just letting GC do it.  see notes in aiomysql dialect
+        self._rows.clear()
+
+    def execute(self, operation, parameters=None):
+        return self.await_(self._execute_async(operation, parameters))
+
+    def executemany(self, operation, seq_of_parameters):
+        return self.await_(
+            self._executemany_async(operation, seq_of_parameters)
+        )
+
+    async def _execute_async(self, operation, parameters):
+        async with self._adapt_connection._execute_mutex:
+            result = await self._cursor.execute(operation, parameters or ())
+
+            if self._cursor.description and not self.server_side:
+                self._rows = collections.deque(await self._cursor.fetchall())
+            return result
+
+    async def _executemany_async(self, operation, seq_of_parameters):
+        async with self._adapt_connection._execute_mutex:
+            return await self._cursor.executemany(operation, seq_of_parameters)
+
+    def nextset(self):
+        self.await_(self._cursor.nextset())
+        if self._cursor.description and not self.server_side:
+            self._rows = collections.deque(
+                self.await_(self._cursor.fetchall())
+            )
+
+    def setinputsizes(self, *inputsizes):
+        # NOTE: this is overrridden in aioodbc due to
+        # see https://github.com/aio-libs/aioodbc/issues/451
+        # right now
+
+        return self.await_(self._cursor.setinputsizes(*inputsizes))
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_dbapi_ss_cursor(AsyncAdapt_dbapi_cursor):
+    __slots__ = ()
+    server_side = True
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor()
+
+        self._cursor = self.await_(cursor.__aenter__())
+
+    def close(self):
+        if self._cursor is not None:
+            self.await_(self._cursor.close())
+            self._cursor = None
+
+    def fetchone(self):
+        return self.await_(self._cursor.fetchone())
+
+    def fetchmany(self, size=None):
+        return self.await_(self._cursor.fetchmany(size=size))
+
+    def fetchall(self):
+        return self.await_(self._cursor.fetchall())
+
+    def __iter__(self):
+        iterator = self._cursor.__aiter__()
+        while True:
+            try:
+                yield self.await_(iterator.__anext__())
+            except StopAsyncIteration:
+                break
+
+
+class AsyncAdapt_dbapi_connection(AdaptedConnection):
+    _cursor_cls = AsyncAdapt_dbapi_cursor
+    _ss_cursor_cls = AsyncAdapt_dbapi_ss_cursor
+
+    await_ = staticmethod(await_only)
+    __slots__ = ("dbapi", "_execute_mutex")
+
+    def __init__(self, dbapi, connection):
+        self.dbapi = dbapi
+        self._connection = connection
+        self._execute_mutex = asyncio.Lock()
+
+    def ping(self, reconnect):
+        return self.await_(self._connection.ping(reconnect))
+
+    def add_output_converter(self, *arg, **kw):
+        self._connection.add_output_converter(*arg, **kw)
+
+    def character_set_name(self):
+        return self._connection.character_set_name()
+
+    @property
+    def autocommit(self):
+        return self._connection.autocommit
+
+    @autocommit.setter
+    def autocommit(self, value):
+        # https://github.com/aio-libs/aioodbc/issues/448
+        # self._connection.autocommit = value
+
+        self._connection._conn.autocommit = value
+
+    def cursor(self, server_side=False):
+        if server_side:
+            return self._ss_cursor_cls(self)
+        else:
+            return self._cursor_cls(self)
+
+    def rollback(self):
+        self.await_(self._connection.rollback())
+
+    def commit(self):
+        self.await_(self._connection.commit())
+
+    def close(self):
+        self.await_(self._connection.close())
+
+
+class AsyncAdaptFallback_dbapi_connection(AsyncAdapt_dbapi_connection):
+    __slots__ = ()
+
+    await_ = staticmethod(await_fallback)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/pyodbc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/pyodbc.py
new file mode 100644
index 00000000..92b7c3a4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/connectors/pyodbc.py
@@ -0,0 +1,249 @@
+# connectors/pyodbc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import re
+from types import ModuleType
+import typing
+from typing import Any
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Union
+from urllib.parse import unquote_plus
+
+from . import Connector
+from .. import ExecutionContext
+from .. import pool
+from .. import util
+from ..engine import ConnectArgsType
+from ..engine import Connection
+from ..engine import interfaces
+from ..engine import URL
+from ..sql.type_api import TypeEngine
+
+if typing.TYPE_CHECKING:
+    from ..engine.interfaces import IsolationLevel
+
+
+class PyODBCConnector(Connector):
+    driver = "pyodbc"
+
+    # this is no longer False for pyodbc in general
+    supports_sane_rowcount_returning = True
+    supports_sane_multi_rowcount = False
+
+    supports_native_decimal = True
+    default_paramstyle = "named"
+
+    fast_executemany = False
+
+    # for non-DSN connections, this *may* be used to
+    # hold the desired driver name
+    pyodbc_driver_name: Optional[str] = None
+
+    dbapi: ModuleType
+
+    def __init__(self, use_setinputsizes: bool = False, **kw: Any):
+        super().__init__(**kw)
+        if use_setinputsizes:
+            self.bind_typing = interfaces.BindTyping.SETINPUTSIZES
+
+    @classmethod
+    def import_dbapi(cls) -> ModuleType:
+        return __import__("pyodbc")
+
+    def create_connect_args(self, url: URL) -> ConnectArgsType:
+        opts = url.translate_connect_args(username="user")
+        opts.update(url.query)
+
+        keys = opts
+
+        query = url.query
+
+        connect_args: Dict[str, Any] = {}
+        connectors: List[str]
+
+        for param in ("ansi", "unicode_results", "autocommit"):
+            if param in keys:
+                connect_args[param] = util.asbool(keys.pop(param))
+
+        if "odbc_connect" in keys:
+            connectors = [unquote_plus(keys.pop("odbc_connect"))]
+        else:
+
+            def check_quote(token: str) -> str:
+                if ";" in str(token) or str(token).startswith("{"):
+                    token = "{%s}" % token.replace("}", "}}")
+                return token
+
+            keys = {k: check_quote(v) for k, v in keys.items()}
+
+            dsn_connection = "dsn" in keys or (
+                "host" in keys and "database" not in keys
+            )
+            if dsn_connection:
+                connectors = [
+                    "dsn=%s" % (keys.pop("host", "") or keys.pop("dsn", ""))
+                ]
+            else:
+                port = ""
+                if "port" in keys and "port" not in query:
+                    port = ",%d" % int(keys.pop("port"))
+
+                connectors = []
+                driver = keys.pop("driver", self.pyodbc_driver_name)
+                if driver is None and keys:
+                    # note if keys is empty, this is a totally blank URL
+                    util.warn(
+                        "No driver name specified; "
+                        "this is expected by PyODBC when using "
+                        "DSN-less connections"
+                    )
+                else:
+                    connectors.append("DRIVER={%s}" % driver)
+
+                connectors.extend(
+                    [
+                        "Server=%s%s" % (keys.pop("host", ""), port),
+                        "Database=%s" % keys.pop("database", ""),
+                    ]
+                )
+
+            user = keys.pop("user", None)
+            if user:
+                connectors.append("UID=%s" % user)
+                pwd = keys.pop("password", "")
+                if pwd:
+                    connectors.append("PWD=%s" % pwd)
+            else:
+                authentication = keys.pop("authentication", None)
+                if authentication:
+                    connectors.append("Authentication=%s" % authentication)
+                else:
+                    connectors.append("Trusted_Connection=Yes")
+
+            # if set to 'Yes', the ODBC layer will try to automagically
+            # convert textual data from your database encoding to your
+            # client encoding.  This should obviously be set to 'No' if
+            # you query a cp1253 encoded database from a latin1 client...
+            if "odbc_autotranslate" in keys:
+                connectors.append(
+                    "AutoTranslate=%s" % keys.pop("odbc_autotranslate")
+                )
+
+            connectors.extend(["%s=%s" % (k, v) for k, v in keys.items()])
+
+        return ((";".join(connectors),), connect_args)
+
+    def is_disconnect(
+        self,
+        e: Exception,
+        connection: Optional[
+            Union[pool.PoolProxiedConnection, interfaces.DBAPIConnection]
+        ],
+        cursor: Optional[interfaces.DBAPICursor],
+    ) -> bool:
+        if isinstance(e, self.dbapi.ProgrammingError):
+            return "The cursor's connection has been closed." in str(
+                e
+            ) or "Attempt to use a closed connection." in str(e)
+        else:
+            return False
+
+    def _dbapi_version(self) -> interfaces.VersionInfoType:
+        if not self.dbapi:
+            return ()
+        return self._parse_dbapi_version(self.dbapi.version)
+
+    def _parse_dbapi_version(self, vers: str) -> interfaces.VersionInfoType:
+        m = re.match(r"(?:py.*-)?([\d\.]+)(?:-(\w+))?", vers)
+        if not m:
+            return ()
+        vers_tuple: interfaces.VersionInfoType = tuple(
+            [int(x) for x in m.group(1).split(".")]
+        )
+        if m.group(2):
+            vers_tuple += (m.group(2),)
+        return vers_tuple
+
+    def _get_server_version_info(
+        self, connection: Connection
+    ) -> interfaces.VersionInfoType:
+        # NOTE: this function is not reliable, particularly when
+        # freetds is in use.   Implement database-specific server version
+        # queries.
+        dbapi_con = connection.connection.dbapi_connection
+        version: Tuple[Union[int, str], ...] = ()
+        r = re.compile(r"[.\-]")
+        for n in r.split(dbapi_con.getinfo(self.dbapi.SQL_DBMS_VER)):  # type: ignore[union-attr]  # noqa: E501
+            try:
+                version += (int(n),)
+            except ValueError:
+                pass
+        return tuple(version)
+
+    def do_set_input_sizes(
+        self,
+        cursor: interfaces.DBAPICursor,
+        list_of_tuples: List[Tuple[str, Any, TypeEngine[Any]]],
+        context: ExecutionContext,
+    ) -> None:
+        # the rules for these types seems a little strange, as you can pass
+        # non-tuples as well as tuples, however it seems to assume "0"
+        # for the subsequent values if you don't pass a tuple which fails
+        # for types such as pyodbc.SQL_WLONGVARCHAR, which is the datatype
+        # that ticket #5649 is targeting.
+
+        # NOTE: as of #6058, this won't be called if the use_setinputsizes
+        # parameter were not passed to the dialect, or if no types were
+        # specified in list_of_tuples
+
+        # as of #8177 for 2.0 we assume use_setinputsizes=True and only
+        # omit the setinputsizes calls for .executemany() with
+        # fast_executemany=True
+
+        if (
+            context.execute_style is interfaces.ExecuteStyle.EXECUTEMANY
+            and self.fast_executemany
+        ):
+            return
+
+        cursor.setinputsizes(
+            [
+                (
+                    (dbtype, None, None)
+                    if not isinstance(dbtype, tuple)
+                    else dbtype
+                )
+                for key, dbtype, sqltype in list_of_tuples
+            ]
+        )
+
+    def get_isolation_level_values(
+        self, dbapi_connection: interfaces.DBAPIConnection
+    ) -> List[IsolationLevel]:
+        return super().get_isolation_level_values(dbapi_connection) + [
+            "AUTOCOMMIT"
+        ]
+
+    def set_isolation_level(
+        self,
+        dbapi_connection: interfaces.DBAPIConnection,
+        level: IsolationLevel,
+    ) -> None:
+        # adjust for ConnectionFairy being present
+        # allows attribute set e.g. "connection.autocommit = True"
+        # to work properly
+
+        if level == "AUTOCOMMIT":
+            dbapi_connection.autocommit = True
+        else:
+            dbapi_connection.autocommit = False
+            super().set_isolation_level(dbapi_connection, level)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__init__.py
new file mode 100644
index 00000000..cb8dc2c6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__init__.py
@@ -0,0 +1,6 @@
+# cyextension/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..92777b8d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.cpython-312-darwin.so
new file mode 100755
index 00000000..a091be97
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.pyx b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.pyx
new file mode 100644
index 00000000..86d24852
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/collections.pyx
@@ -0,0 +1,409 @@
+# cyextension/collections.pyx
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+cimport cython
+from cpython.long cimport PyLong_FromLongLong
+from cpython.set cimport PySet_Add
+
+from collections.abc import Collection
+from itertools import filterfalse
+
+cdef bint add_not_present(set seen, object item, hashfunc):
+    hash_value = hashfunc(item)
+    if hash_value not in seen:
+        PySet_Add(seen, hash_value)
+        return True
+    else:
+        return False
+
+cdef list cunique_list(seq, hashfunc=None):
+    cdef set seen = set()
+    if not hashfunc:
+        return [x for x in seq if x not in seen and not PySet_Add(seen, x)]
+    else:
+        return [x for x in seq if add_not_present(seen, x, hashfunc)]
+
+def unique_list(seq, hashfunc=None):
+    return cunique_list(seq, hashfunc)
+
+cdef class OrderedSet(set):
+
+    cdef list _list
+
+    @classmethod
+    def __class_getitem__(cls, key):
+        return cls
+
+    def __init__(self, d=None):
+        set.__init__(self)
+        if d is not None:
+            self._list = cunique_list(d)
+            set.update(self, self._list)
+        else:
+            self._list = []
+
+    cpdef OrderedSet copy(self):
+        cdef OrderedSet cp = OrderedSet.__new__(OrderedSet)
+        cp._list = list(self._list)
+        set.update(cp, cp._list)
+        return cp
+
+    @cython.final
+    cdef OrderedSet _from_list(self, list new_list):
+        cdef OrderedSet new = OrderedSet.__new__(OrderedSet)
+        new._list = new_list
+        set.update(new, new_list)
+        return new
+
+    def add(self, element):
+        if element not in self:
+            self._list.append(element)
+            PySet_Add(self, element)
+
+    def remove(self, element):
+        # set.remove will raise if element is not in self
+        set.remove(self, element)
+        self._list.remove(element)
+
+    def pop(self):
+        try:
+            value = self._list.pop()
+        except IndexError:
+            raise KeyError("pop from an empty set") from None
+        set.remove(self, value)
+        return value
+
+    def insert(self, Py_ssize_t pos, element):
+        if element not in self:
+            self._list.insert(pos, element)
+            PySet_Add(self, element)
+
+    def discard(self, element):
+        if element in self:
+            set.remove(self, element)
+            self._list.remove(element)
+
+    def clear(self):
+        set.clear(self)
+        self._list = []
+
+    def __getitem__(self, key):
+        return self._list[key]
+
+    def __iter__(self):
+        return iter(self._list)
+
+    def __add__(self, other):
+        return self.union(other)
+
+    def __repr__(self):
+        return "%s(%r)" % (self.__class__.__name__, self._list)
+
+    __str__ = __repr__
+
+    def update(self, *iterables):
+        for iterable in iterables:
+            for e in iterable:
+                if e not in self:
+                    self._list.append(e)
+                    set.add(self, e)
+
+    def __ior__(self, iterable):
+        self.update(iterable)
+        return self
+
+    def union(self, *other):
+        result = self.copy()
+        result.update(*other)
+        return result
+
+    def __or__(self, other):
+        return self.union(other)
+
+    def intersection(self, *other):
+        cdef set other_set = set.intersection(self, *other)
+        return self._from_list([a for a in self._list if a in other_set])
+
+    def __and__(self, other):
+        return self.intersection(other)
+
+    def symmetric_difference(self, other):
+        cdef set other_set
+        if isinstance(other, set):
+            other_set = <set> other
+            collection = other_set
+        elif isinstance(other, Collection):
+            collection = other
+            other_set = set(other)
+        else:
+            collection = list(other)
+            other_set = set(collection)
+        result = self._from_list([a for a in self._list if a not in other_set])
+        result.update(a for a in collection if a not in self)
+        return result
+
+    def __xor__(self, other):
+        return self.symmetric_difference(other)
+
+    def difference(self, *other):
+        cdef set other_set = set.difference(self, *other)
+        return self._from_list([a for a in self._list if a in other_set])
+
+    def __sub__(self, other):
+        return self.difference(other)
+
+    def intersection_update(self, *other):
+        set.intersection_update(self, *other)
+        self._list = [a for a in self._list if a in self]
+
+    def __iand__(self, other):
+        self.intersection_update(other)
+        return self
+
+    cpdef symmetric_difference_update(self, other):
+        collection = other if isinstance(other, Collection) else list(other)
+        set.symmetric_difference_update(self, collection)
+        self._list = [a for a in self._list if a in self]
+        self._list += [a for a in collection if a in self]
+
+    def __ixor__(self, other):
+        self.symmetric_difference_update(other)
+        return self
+
+    def difference_update(self, *other):
+        set.difference_update(self, *other)
+        self._list = [a for a in self._list if a in self]
+
+    def __isub__(self, other):
+        self.difference_update(other)
+        return self
+
+cdef object cy_id(object item):
+    return PyLong_FromLongLong(<long long> (<void *>item))
+
+# NOTE: cython 0.x will call __add__, __sub__, etc with the parameter swapped
+# instead of the __rmeth__, so they need to check that also self is of the
+# correct type. This is fixed in cython 3.x. See:
+# https://docs.cython.org/en/latest/src/userguide/special_methods.html#arithmetic-methods
+cdef class IdentitySet:
+    """A set that considers only object id() for uniqueness.
+
+    This strategy has edge cases for builtin types- it's possible to have
+    two 'foo' strings in one of these sets, for example.  Use sparingly.
+
+    """
+
+    cdef dict _members
+
+    def __init__(self, iterable=None):
+        self._members = {}
+        if iterable:
+            self.update(iterable)
+
+    def add(self, value):
+        self._members[cy_id(value)] = value
+
+    def __contains__(self, value):
+        return cy_id(value) in self._members
+
+    cpdef remove(self, value):
+        del self._members[cy_id(value)]
+
+    def discard(self, value):
+        try:
+            self.remove(value)
+        except KeyError:
+            pass
+
+    def pop(self):
+        cdef tuple pair
+        try:
+            pair = self._members.popitem()
+            return pair[1]
+        except KeyError:
+            raise KeyError("pop from an empty set")
+
+    def clear(self):
+        self._members.clear()
+
+    def __eq__(self, other):
+        cdef IdentitySet other_
+        if isinstance(other, IdentitySet):
+            other_ = other
+            return self._members == other_._members
+        else:
+            return False
+
+    def __ne__(self, other):
+        cdef IdentitySet other_
+        if isinstance(other, IdentitySet):
+            other_ = other
+            return self._members != other_._members
+        else:
+            return True
+
+    cpdef issubset(self, iterable):
+        cdef IdentitySet other
+        if isinstance(iterable, self.__class__):
+            other = iterable
+        else:
+            other = self.__class__(iterable)
+
+        if len(self) > len(other):
+            return False
+        for m in filterfalse(other._members.__contains__, self._members):
+            return False
+        return True
+
+    def __le__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.issubset(other)
+
+    def __lt__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return len(self) < len(other) and self.issubset(other)
+
+    cpdef issuperset(self, iterable):
+        cdef IdentitySet other
+        if isinstance(iterable, self.__class__):
+            other = iterable
+        else:
+            other = self.__class__(iterable)
+
+        if len(self) < len(other):
+            return False
+        for m in filterfalse(self._members.__contains__, other._members):
+            return False
+        return True
+
+    def __ge__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.issuperset(other)
+
+    def __gt__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return len(self) > len(other) and self.issuperset(other)
+
+    cpdef IdentitySet union(self, iterable):
+        cdef IdentitySet result = self.__class__()
+        result._members.update(self._members)
+        result.update(iterable)
+        return result
+
+    def __or__(self, other):
+        if not isinstance(other, IdentitySet) or not isinstance(self, IdentitySet):
+            return NotImplemented
+        return self.union(other)
+
+    cpdef update(self, iterable):
+        for obj in iterable:
+            self._members[cy_id(obj)] = obj
+
+    def __ior__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.update(other)
+        return self
+
+    cpdef IdentitySet difference(self, iterable):
+        cdef IdentitySet result = self.__new__(self.__class__)
+        if isinstance(iterable, self.__class__):
+            other = (<IdentitySet>iterable)._members
+        else:
+            other = {cy_id(obj) for obj in iterable}
+        result._members = {k:v for k, v in self._members.items() if k not in other}
+        return result
+
+    def __sub__(self, other):
+        if not isinstance(other, IdentitySet) or not isinstance(self, IdentitySet):
+            return NotImplemented
+        return self.difference(other)
+
+    cpdef difference_update(self, iterable):
+        cdef IdentitySet other = self.difference(iterable)
+        self._members = other._members
+
+    def __isub__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.difference_update(other)
+        return self
+
+    cpdef IdentitySet intersection(self, iterable):
+        cdef IdentitySet result = self.__new__(self.__class__)
+        if isinstance(iterable, self.__class__):
+            other = (<IdentitySet>iterable)._members
+        else:
+            other = {cy_id(obj) for obj in iterable}
+        result._members = {k: v for k, v in self._members.items() if k in other}
+        return result
+
+    def __and__(self, other):
+        if not isinstance(other, IdentitySet) or not isinstance(self, IdentitySet):
+            return NotImplemented
+        return self.intersection(other)
+
+    cpdef intersection_update(self, iterable):
+        cdef IdentitySet other = self.intersection(iterable)
+        self._members = other._members
+
+    def __iand__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.intersection_update(other)
+        return self
+
+    cpdef IdentitySet symmetric_difference(self, iterable):
+        cdef IdentitySet result = self.__new__(self.__class__)
+        cdef dict other
+        if isinstance(iterable, self.__class__):
+            other = (<IdentitySet>iterable)._members
+        else:
+            other = {cy_id(obj): obj for obj in iterable}
+        result._members = {k: v for k, v in self._members.items() if k not in other}
+        result._members.update(
+            [(k, v) for k, v in other.items() if k not in self._members]
+        )
+        return result
+
+    def __xor__(self, other):
+        if not isinstance(other, IdentitySet) or not isinstance(self, IdentitySet):
+            return NotImplemented
+        return self.symmetric_difference(other)
+
+    cpdef symmetric_difference_update(self, iterable):
+        cdef IdentitySet other = self.symmetric_difference(iterable)
+        self._members = other._members
+
+    def __ixor__(self, other):
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.symmetric_difference(other)
+        return self
+
+    cpdef IdentitySet copy(self):
+        cdef IdentitySet cp = self.__new__(self.__class__)
+        cp._members = self._members.copy()
+        return cp
+
+    def __copy__(self):
+        return self.copy()
+
+    def __len__(self):
+        return len(self._members)
+
+    def __iter__(self):
+        return iter(self._members.values())
+
+    def __hash__(self):
+        raise TypeError("set objects are unhashable")
+
+    def __repr__(self):
+        return "%s(%r)" % (type(self).__name__, list(self._members.values()))
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.cpython-312-darwin.so
new file mode 100755
index 00000000..badabfda
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pxd b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pxd
new file mode 100644
index 00000000..76f22893
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pxd
@@ -0,0 +1,8 @@
+# cyextension/immutabledict.pxd
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+cdef class immutabledict(dict):
+    pass
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pyx b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pyx
new file mode 100644
index 00000000..b37eccc4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/immutabledict.pyx
@@ -0,0 +1,133 @@
+# cyextension/immutabledict.pyx
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from cpython.dict cimport PyDict_New, PyDict_Update, PyDict_Size
+
+
+def _readonly_fn(obj):
+    raise TypeError(
+        "%s object is immutable and/or readonly" % obj.__class__.__name__)
+
+
+def _immutable_fn(obj):
+    raise TypeError(
+        "%s object is immutable" % obj.__class__.__name__)
+
+
+class ReadOnlyContainer:
+
+    __slots__ = ()
+
+    def _readonly(self, *a,**kw):
+        _readonly_fn(self)
+
+    __delitem__ = __setitem__ = __setattr__ = _readonly
+
+
+class ImmutableDictBase(dict):
+    def _immutable(self, *a,**kw):
+        _immutable_fn(self)
+
+    @classmethod
+    def __class_getitem__(cls, key):
+        return cls
+
+    __delitem__ = __setitem__ = __setattr__ = _immutable
+    clear = pop = popitem = setdefault = update = _immutable
+
+
+cdef class immutabledict(dict):
+    def __repr__(self):
+        return f"immutabledict({dict.__repr__(self)})"
+
+    @classmethod
+    def __class_getitem__(cls, key):
+        return cls
+
+    def union(self, *args, **kw):
+        cdef dict to_merge = None
+        cdef immutabledict result
+        cdef Py_ssize_t args_len = len(args)
+        if args_len > 1:
+            raise TypeError(
+                f'union expected at most 1 argument, got {args_len}'
+            )
+        if args_len == 1:
+            attribute = args[0]
+            if isinstance(attribute, dict):
+                to_merge = <dict> attribute
+        if to_merge is None:
+            to_merge = dict(*args, **kw)
+
+        if PyDict_Size(to_merge) == 0:
+            return self
+
+        # new + update is faster than immutabledict(self)
+        result = immutabledict()
+        PyDict_Update(result, self)
+        PyDict_Update(result, to_merge)
+        return result
+
+    def merge_with(self, *other):
+        cdef immutabledict result = None
+        cdef object d
+        cdef bint update = False
+        if not other:
+            return self
+        for d in other:
+            if d:
+                if update == False:
+                    update = True
+                    # new + update is faster than immutabledict(self)
+                    result = immutabledict()
+                    PyDict_Update(result, self)
+                PyDict_Update(
+                    result, <dict>(d if isinstance(d, dict) else dict(d))
+                )
+
+        return self if update == False else result
+
+    def copy(self):
+        return self
+
+    def __reduce__(self):
+        return immutabledict, (dict(self), )
+
+    def __delitem__(self, k):
+        _immutable_fn(self)
+
+    def __setitem__(self, k, v):
+        _immutable_fn(self)
+
+    def __setattr__(self, k, v):
+        _immutable_fn(self)
+
+    def clear(self, *args, **kw):
+        _immutable_fn(self)
+
+    def pop(self, *args, **kw):
+        _immutable_fn(self)
+
+    def popitem(self, *args, **kw):
+        _immutable_fn(self)
+
+    def setdefault(self, *args, **kw):
+        _immutable_fn(self)
+
+    def update(self, *args, **kw):
+        _immutable_fn(self)
+
+    # PEP 584
+    def __ior__(self, other):
+        _immutable_fn(self)
+
+    def __or__(self, other):
+        return immutabledict(dict.__or__(self, other))
+
+    def __ror__(self, other):
+        # NOTE: this is used only in cython 3.x;
+        # version 0.x will call __or__ with args inversed
+        return immutabledict(dict.__ror__(self, other))
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.cpython-312-darwin.so
new file mode 100755
index 00000000..5d0eea0d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.pyx b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.pyx
new file mode 100644
index 00000000..3d714569
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/processors.pyx
@@ -0,0 +1,68 @@
+# cyextension/processors.pyx
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+import datetime
+from datetime import datetime as datetime_cls
+from datetime import time as time_cls
+from datetime import date as date_cls
+import re
+
+from cpython.object cimport PyObject_Str
+from cpython.unicode cimport PyUnicode_AsASCIIString, PyUnicode_Check, PyUnicode_Decode
+from libc.stdio cimport sscanf
+
+
+def int_to_boolean(value):
+    if value is None:
+        return None
+    return True if value else False
+
+def to_str(value):
+    return PyObject_Str(value) if value is not None else None
+
+def to_float(value):
+    return float(value) if value is not None else None
+
+cdef inline bytes to_bytes(object value, str type_name):
+    try:
+        return PyUnicode_AsASCIIString(value)
+    except Exception as e:
+        raise ValueError(
+            f"Couldn't parse {type_name} string '{value!r}' "
+            "- value is not a string."
+        ) from e
+
+def str_to_datetime(value):
+    if value is not None:
+        value = datetime_cls.fromisoformat(value)
+    return value
+
+def str_to_time(value):
+    if value is not None:
+        value = time_cls.fromisoformat(value)
+    return value
+
+
+def str_to_date(value):
+    if value is not None:
+        value = date_cls.fromisoformat(value)
+    return value
+
+
+
+cdef class DecimalResultProcessor:
+    cdef object type_
+    cdef str format_
+
+    def __cinit__(self, type_, format_):
+        self.type_ = type_
+        self.format_ = format_
+
+    def process(self, object value):
+        if value is None:
+            return None
+        else:
+            return self.type_(self.format_ % value)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.cpython-312-darwin.so
new file mode 100755
index 00000000..a99f17e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.pyx b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.pyx
new file mode 100644
index 00000000..b6e357a1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/resultproxy.pyx
@@ -0,0 +1,102 @@
+# cyextension/resultproxy.pyx
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+import operator
+
+cdef class BaseRow:
+    cdef readonly object _parent
+    cdef readonly dict _key_to_index
+    cdef readonly tuple _data
+
+    def __init__(self, object parent, object processors, dict key_to_index, object data):
+        """Row objects are constructed by CursorResult objects."""
+
+        self._parent = parent
+
+        self._key_to_index = key_to_index
+
+        if processors:
+            self._data = _apply_processors(processors, data)
+        else:
+            self._data = tuple(data)
+
+    def __reduce__(self):
+        return (
+            rowproxy_reconstructor,
+            (self.__class__, self.__getstate__()),
+        )
+
+    def __getstate__(self):
+        return {"_parent": self._parent, "_data": self._data}
+
+    def __setstate__(self, dict state):
+        parent = state["_parent"]
+        self._parent = parent
+        self._data = state["_data"]
+        self._key_to_index = parent._key_to_index
+
+    def _values_impl(self):
+        return list(self)
+
+    def __iter__(self):
+        return iter(self._data)
+
+    def __len__(self):
+        return len(self._data)
+
+    def __hash__(self):
+        return hash(self._data)
+
+    def __getitem__(self, index):
+        return self._data[index]
+
+    def _get_by_key_impl_mapping(self, key):
+        return self._get_by_key_impl(key, 0)
+
+    cdef _get_by_key_impl(self, object key, int attr_err):
+        index = self._key_to_index.get(key)
+        if index is not None:
+            return self._data[<int>index]
+        self._parent._key_not_found(key, attr_err != 0)
+
+    def __getattr__(self, name):
+        return self._get_by_key_impl(name, 1)
+
+    def _to_tuple_instance(self):
+        return self._data
+
+
+cdef tuple _apply_processors(proc, data):
+    res = []
+    for i in range(len(proc)):
+        p = proc[i]
+        if p is None:
+            res.append(data[i])
+        else:
+            res.append(p(data[i]))
+    return tuple(res)
+
+
+def rowproxy_reconstructor(cls, state):
+    obj = cls.__new__(cls)
+    obj.__setstate__(state)
+    return obj
+
+
+cdef int is_contiguous(tuple indexes):
+    cdef int i
+    for i in range(1, len(indexes)):
+        if indexes[i-1] != indexes[i] -1:
+            return 0
+    return 1
+
+
+def tuplegetter(*indexes):
+    if len(indexes) == 1 or is_contiguous(indexes) != 0:
+        # slice form is faster but returns a list if input is list
+        return operator.itemgetter(slice(indexes[0], indexes[-1] + 1))
+    else:
+        return operator.itemgetter(*indexes)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.cpython-312-darwin.so b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.cpython-312-darwin.so
new file mode 100755
index 00000000..0691d370
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.cpython-312-darwin.so differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.pyx b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.pyx
new file mode 100644
index 00000000..cb17acd6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/cyextension/util.pyx
@@ -0,0 +1,91 @@
+# cyextension/util.pyx
+# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from collections.abc import Mapping
+
+from sqlalchemy import exc
+
+cdef tuple _Empty_Tuple = ()
+
+cdef inline bint _mapping_or_tuple(object value):
+    return isinstance(value, dict) or isinstance(value, tuple) or isinstance(value, Mapping)
+
+cdef inline bint _check_item(object params) except 0:
+    cdef object item
+    cdef bint ret = 1
+    if params:
+        item = params[0]
+        if not _mapping_or_tuple(item):
+            ret = 0
+            raise exc.ArgumentError(
+                "List argument must consist only of tuples or dictionaries"
+            )
+    return ret
+
+def _distill_params_20(object params):
+    if params is None:
+        return _Empty_Tuple
+    elif isinstance(params, list) or isinstance(params, tuple):
+        _check_item(params)
+        return params
+    elif isinstance(params, dict) or isinstance(params, Mapping):
+        return [params]
+    else:
+        raise exc.ArgumentError("mapping or list expected for parameters")
+
+
+def _distill_raw_params(object params):
+    if params is None:
+        return _Empty_Tuple
+    elif isinstance(params, list):
+        _check_item(params)
+        return params
+    elif _mapping_or_tuple(params):
+        return [params]
+    else:
+        raise exc.ArgumentError("mapping or sequence expected for parameters")
+
+cdef class prefix_anon_map(dict):
+    def __missing__(self, str key):
+        cdef str derived
+        cdef int anonymous_counter
+        cdef dict self_dict = self
+
+        derived = key.split(" ", 1)[1]
+
+        anonymous_counter = self_dict.get(derived, 1)
+        self_dict[derived] = anonymous_counter + 1
+        value = f"{derived}_{anonymous_counter}"
+        self_dict[key] = value
+        return value
+
+
+cdef class cache_anon_map(dict):
+    cdef int _index
+
+    def __init__(self):
+        self._index = 0
+
+    def get_anon(self, obj):
+        cdef long long idself
+        cdef str id_
+        cdef dict self_dict = self
+
+        idself = id(obj)
+        if idself in self_dict:
+            return self_dict[idself], True
+        else:
+            id_ = self.__missing__(idself)
+            return id_, False
+
+    def __missing__(self, key):
+        cdef str val
+        cdef dict self_dict = self
+
+        self_dict[key] = val = str(self._index)
+        self._index += 1
+        return val
+
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__init__.py
new file mode 100644
index 00000000..31ce6d64
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__init__.py
@@ -0,0 +1,61 @@
+# dialects/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Callable
+from typing import Optional
+from typing import Type
+from typing import TYPE_CHECKING
+
+from .. import util
+
+if TYPE_CHECKING:
+    from ..engine.interfaces import Dialect
+
+__all__ = ("mssql", "mysql", "oracle", "postgresql", "sqlite")
+
+
+def _auto_fn(name: str) -> Optional[Callable[[], Type[Dialect]]]:
+    """default dialect importer.
+
+    plugs into the :class:`.PluginLoader`
+    as a first-hit system.
+
+    """
+    if "." in name:
+        dialect, driver = name.split(".")
+    else:
+        dialect = name
+        driver = "base"
+
+    try:
+        if dialect == "mariadb":
+            # it's "OK" for us to hardcode here since _auto_fn is already
+            # hardcoded.   if mysql / mariadb etc were third party dialects
+            # they would just publish all the entrypoints, which would actually
+            # look much nicer.
+            module = __import__(
+                "sqlalchemy.dialects.mysql.mariadb"
+            ).dialects.mysql.mariadb
+            return module.loader(driver)  # type: ignore
+        else:
+            module = __import__("sqlalchemy.dialects.%s" % (dialect,)).dialects
+            module = getattr(module, dialect)
+    except ImportError:
+        return None
+
+    if hasattr(module, driver):
+        module = getattr(module, driver)
+        return lambda: module.dialect
+    else:
+        return None
+
+
+registry = util.PluginLoader("sqlalchemy.dialects", auto_fn=_auto_fn)
+
+plugins = util.PluginLoader("sqlalchemy.plugins")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..ba8e4bf1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/_typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/_typing.cpython-312.pyc
new file mode 100644
index 00000000..3e9a4999
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/__pycache__/_typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/_typing.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/_typing.py
new file mode 100644
index 00000000..4dd40d72
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/_typing.py
@@ -0,0 +1,30 @@
+# dialects/_typing.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+from typing import Any
+from typing import Iterable
+from typing import Mapping
+from typing import Optional
+from typing import Union
+
+from ..sql import roles
+from ..sql.base import ColumnCollection
+from ..sql.schema import Column
+from ..sql.schema import ColumnCollectionConstraint
+from ..sql.schema import Index
+
+
+_OnConflictConstraintT = Union[str, ColumnCollectionConstraint, Index, None]
+_OnConflictIndexElementsT = Optional[
+    Iterable[Union[Column[Any], str, roles.DDLConstraintColumnRole]]
+]
+_OnConflictIndexWhereT = Optional[roles.WhereHavingRole]
+_OnConflictSetT = Optional[
+    Union[Mapping[Any, Any], ColumnCollection[Any, Any]]
+]
+_OnConflictWhereT = Optional[roles.WhereHavingRole]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__init__.py
new file mode 100644
index 00000000..20140fdd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__init__.py
@@ -0,0 +1,88 @@
+# dialects/mssql/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from . import aioodbc  # noqa
+from . import base  # noqa
+from . import pymssql  # noqa
+from . import pyodbc  # noqa
+from .base import BIGINT
+from .base import BINARY
+from .base import BIT
+from .base import CHAR
+from .base import DATE
+from .base import DATETIME
+from .base import DATETIME2
+from .base import DATETIMEOFFSET
+from .base import DECIMAL
+from .base import DOUBLE_PRECISION
+from .base import FLOAT
+from .base import IMAGE
+from .base import INTEGER
+from .base import JSON
+from .base import MONEY
+from .base import NCHAR
+from .base import NTEXT
+from .base import NUMERIC
+from .base import NVARCHAR
+from .base import REAL
+from .base import ROWVERSION
+from .base import SMALLDATETIME
+from .base import SMALLINT
+from .base import SMALLMONEY
+from .base import SQL_VARIANT
+from .base import TEXT
+from .base import TIME
+from .base import TIMESTAMP
+from .base import TINYINT
+from .base import UNIQUEIDENTIFIER
+from .base import VARBINARY
+from .base import VARCHAR
+from .base import XML
+from ...sql import try_cast
+
+
+base.dialect = dialect = pyodbc.dialect
+
+
+__all__ = (
+    "JSON",
+    "INTEGER",
+    "BIGINT",
+    "SMALLINT",
+    "TINYINT",
+    "VARCHAR",
+    "NVARCHAR",
+    "CHAR",
+    "NCHAR",
+    "TEXT",
+    "NTEXT",
+    "DECIMAL",
+    "NUMERIC",
+    "FLOAT",
+    "DATETIME",
+    "DATETIME2",
+    "DATETIMEOFFSET",
+    "DATE",
+    "DOUBLE_PRECISION",
+    "TIME",
+    "SMALLDATETIME",
+    "BINARY",
+    "VARBINARY",
+    "BIT",
+    "REAL",
+    "IMAGE",
+    "TIMESTAMP",
+    "ROWVERSION",
+    "MONEY",
+    "SMALLMONEY",
+    "UNIQUEIDENTIFIER",
+    "SQL_VARIANT",
+    "XML",
+    "dialect",
+    "try_cast",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b4f76f63
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/aioodbc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/aioodbc.cpython-312.pyc
new file mode 100644
index 00000000..01062ed2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/aioodbc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..96afea28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/information_schema.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/information_schema.cpython-312.pyc
new file mode 100644
index 00000000..1ff33093
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/information_schema.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/json.cpython-312.pyc
new file mode 100644
index 00000000..ab27f13e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..487e4d07
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pymssql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pymssql.cpython-312.pyc
new file mode 100644
index 00000000..e750db66
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pymssql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pyodbc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pyodbc.cpython-312.pyc
new file mode 100644
index 00000000..633eef54
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/__pycache__/pyodbc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/aioodbc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/aioodbc.py
new file mode 100644
index 00000000..522ad1d6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/aioodbc.py
@@ -0,0 +1,63 @@
+# dialects/mssql/aioodbc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+r"""
+.. dialect:: mssql+aioodbc
+    :name: aioodbc
+    :dbapi: aioodbc
+    :connectstring: mssql+aioodbc://<username>:<password>@<dsnname>
+    :url: https://pypi.org/project/aioodbc/
+
+
+Support for the SQL Server database in asyncio style, using the aioodbc
+driver which itself is a thread-wrapper around pyodbc.
+
+.. versionadded:: 2.0.23  Added the mssql+aioodbc dialect which builds
+   on top of the pyodbc and general aio* dialect architecture.
+
+Using a special asyncio mediation layer, the aioodbc dialect is usable
+as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>`
+extension package.
+
+Most behaviors and caveats for this driver are the same as that of the
+pyodbc dialect used on SQL Server; see :ref:`mssql_pyodbc` for general
+background.
+
+This dialect should normally be used only with the
+:func:`_asyncio.create_async_engine` engine creation function; connection
+styles are otherwise equivalent to those documented in the pyodbc section::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine(
+        "mssql+aioodbc://scott:tiger@mssql2017:1433/test?"
+        "driver=ODBC+Driver+18+for+SQL+Server&TrustServerCertificate=yes"
+    )
+
+"""
+
+from __future__ import annotations
+
+from .pyodbc import MSDialect_pyodbc
+from .pyodbc import MSExecutionContext_pyodbc
+from ...connectors.aioodbc import aiodbcConnector
+
+
+class MSExecutionContext_aioodbc(MSExecutionContext_pyodbc):
+    def create_server_side_cursor(self):
+        return self._dbapi_connection.cursor(server_side=True)
+
+
+class MSDialectAsync_aioodbc(aiodbcConnector, MSDialect_pyodbc):
+    driver = "aioodbc"
+
+    supports_statement_cache = True
+
+    execution_ctx_cls = MSExecutionContext_aioodbc
+
+
+dialect = MSDialectAsync_aioodbc
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/base.py
new file mode 100644
index 00000000..916809e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/base.py
@@ -0,0 +1,4058 @@
+# dialects/mssql/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""
+.. dialect:: mssql
+    :name: Microsoft SQL Server
+    :normal_support: 2012+
+    :best_effort: 2005+
+
+.. _mssql_external_dialects:
+
+External Dialects
+-----------------
+
+In addition to the above DBAPI layers with native SQLAlchemy support, there
+are third-party dialects for other DBAPI layers that are compatible
+with SQL Server. See the "External Dialects" list on the
+:ref:`dialect_toplevel` page.
+
+.. _mssql_identity:
+
+Auto Increment Behavior / IDENTITY Columns
+------------------------------------------
+
+SQL Server provides so-called "auto incrementing" behavior using the
+``IDENTITY`` construct, which can be placed on any single integer column in a
+table. SQLAlchemy considers ``IDENTITY`` within its default "autoincrement"
+behavior for an integer primary key column, described at
+:paramref:`_schema.Column.autoincrement`.  This means that by default,
+the first integer primary key column in a :class:`_schema.Table` will be
+considered to be the identity column - unless it is associated with a
+:class:`.Sequence` - and will generate DDL as such::
+
+    from sqlalchemy import Table, MetaData, Column, Integer
+
+    m = MetaData()
+    t = Table(
+        "t",
+        m,
+        Column("id", Integer, primary_key=True),
+        Column("x", Integer),
+    )
+    m.create_all(engine)
+
+The above example will generate DDL as:
+
+.. sourcecode:: sql
+
+    CREATE TABLE t (
+        id INTEGER NOT NULL IDENTITY,
+        x INTEGER NULL,
+        PRIMARY KEY (id)
+    )
+
+For the case where this default generation of ``IDENTITY`` is not desired,
+specify ``False`` for the :paramref:`_schema.Column.autoincrement` flag,
+on the first integer primary key column::
+
+    m = MetaData()
+    t = Table(
+        "t",
+        m,
+        Column("id", Integer, primary_key=True, autoincrement=False),
+        Column("x", Integer),
+    )
+    m.create_all(engine)
+
+To add the ``IDENTITY`` keyword to a non-primary key column, specify
+``True`` for the :paramref:`_schema.Column.autoincrement` flag on the desired
+:class:`_schema.Column` object, and ensure that
+:paramref:`_schema.Column.autoincrement`
+is set to ``False`` on any integer primary key column::
+
+    m = MetaData()
+    t = Table(
+        "t",
+        m,
+        Column("id", Integer, primary_key=True, autoincrement=False),
+        Column("x", Integer, autoincrement=True),
+    )
+    m.create_all(engine)
+
+.. versionchanged::  1.4   Added :class:`_schema.Identity` construct
+   in a :class:`_schema.Column` to specify the start and increment
+   parameters of an IDENTITY. These replace
+   the use of the :class:`.Sequence` object in order to specify these values.
+
+.. deprecated:: 1.4
+
+   The ``mssql_identity_start`` and ``mssql_identity_increment`` parameters
+   to :class:`_schema.Column` are deprecated and should we replaced by
+   an :class:`_schema.Identity` object. Specifying both ways of configuring
+   an IDENTITY will result in a compile error.
+   These options are also no longer returned as part of the
+   ``dialect_options`` key in :meth:`_reflection.Inspector.get_columns`.
+   Use the information in the ``identity`` key instead.
+
+.. deprecated:: 1.3
+
+   The use of :class:`.Sequence` to specify IDENTITY characteristics is
+   deprecated and will be removed in a future release.   Please use
+   the :class:`_schema.Identity` object parameters
+   :paramref:`_schema.Identity.start` and
+   :paramref:`_schema.Identity.increment`.
+
+.. versionchanged::  1.4   Removed the ability to use a :class:`.Sequence`
+   object to modify IDENTITY characteristics. :class:`.Sequence` objects
+   now only manipulate true T-SQL SEQUENCE types.
+
+.. note::
+
+    There can only be one IDENTITY column on the table.  When using
+    ``autoincrement=True`` to enable the IDENTITY keyword, SQLAlchemy does not
+    guard against multiple columns specifying the option simultaneously.  The
+    SQL Server database will instead reject the ``CREATE TABLE`` statement.
+
+.. note::
+
+    An INSERT statement which attempts to provide a value for a column that is
+    marked with IDENTITY will be rejected by SQL Server.   In order for the
+    value to be accepted, a session-level option "SET IDENTITY_INSERT" must be
+    enabled.   The SQLAlchemy SQL Server dialect will perform this operation
+    automatically when using a core :class:`_expression.Insert`
+    construct; if the
+    execution specifies a value for the IDENTITY column, the "IDENTITY_INSERT"
+    option will be enabled for the span of that statement's invocation.However,
+    this scenario is not high performing and should not be relied upon for
+    normal use.   If a table doesn't actually require IDENTITY behavior in its
+    integer primary key column, the keyword should be disabled when creating
+    the table by ensuring that ``autoincrement=False`` is set.
+
+Controlling "Start" and "Increment"
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Specific control over the "start" and "increment" values for
+the ``IDENTITY`` generator are provided using the
+:paramref:`_schema.Identity.start` and :paramref:`_schema.Identity.increment`
+parameters passed to the :class:`_schema.Identity` object::
+
+    from sqlalchemy import Table, Integer, Column, Identity
+
+    test = Table(
+        "test",
+        metadata,
+        Column(
+            "id", Integer, primary_key=True, Identity(start=100, increment=10)
+        ),
+        Column("name", String(20)),
+    )
+
+The CREATE TABLE for the above :class:`_schema.Table` object would be:
+
+.. sourcecode:: sql
+
+   CREATE TABLE test (
+     id INTEGER NOT NULL IDENTITY(100,10) PRIMARY KEY,
+     name VARCHAR(20) NULL,
+   )
+
+.. note::
+
+   The :class:`_schema.Identity` object supports many other parameter in
+   addition to ``start`` and ``increment``. These are not supported by
+   SQL Server and will be ignored when generating the CREATE TABLE ddl.
+
+.. versionchanged:: 1.3.19  The :class:`_schema.Identity` object is
+   now used to affect the
+   ``IDENTITY`` generator for a :class:`_schema.Column` under  SQL Server.
+   Previously, the :class:`.Sequence` object was used.  As SQL Server now
+   supports real sequences as a separate construct, :class:`.Sequence` will be
+   functional in the normal way starting from SQLAlchemy version 1.4.
+
+
+Using IDENTITY with Non-Integer numeric types
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+SQL Server also allows ``IDENTITY`` to be used with ``NUMERIC`` columns.  To
+implement this pattern smoothly in SQLAlchemy, the primary datatype of the
+column should remain as ``Integer``, however the underlying implementation
+type deployed to the SQL Server database can be specified as ``Numeric`` using
+:meth:`.TypeEngine.with_variant`::
+
+    from sqlalchemy import Column
+    from sqlalchemy import Integer
+    from sqlalchemy import Numeric
+    from sqlalchemy import String
+    from sqlalchemy.ext.declarative import declarative_base
+
+    Base = declarative_base()
+
+
+    class TestTable(Base):
+        __tablename__ = "test"
+        id = Column(
+            Integer().with_variant(Numeric(10, 0), "mssql"),
+            primary_key=True,
+            autoincrement=True,
+        )
+        name = Column(String)
+
+In the above example, ``Integer().with_variant()`` provides clear usage
+information that accurately describes the intent of the code. The general
+restriction that ``autoincrement`` only applies to ``Integer`` is established
+at the metadata level and not at the per-dialect level.
+
+When using the above pattern, the primary key identifier that comes back from
+the insertion of a row, which is also the value that would be assigned to an
+ORM object such as ``TestTable`` above, will be an instance of ``Decimal()``
+and not ``int`` when using SQL Server. The numeric return type of the
+:class:`_types.Numeric` type can be changed to return floats by passing False
+to :paramref:`_types.Numeric.asdecimal`. To normalize the return type of the
+above ``Numeric(10, 0)`` to return Python ints (which also support "long"
+integer values in Python 3), use :class:`_types.TypeDecorator` as follows::
+
+    from sqlalchemy import TypeDecorator
+
+
+    class NumericAsInteger(TypeDecorator):
+        "normalize floating point return values into ints"
+
+        impl = Numeric(10, 0, asdecimal=False)
+        cache_ok = True
+
+        def process_result_value(self, value, dialect):
+            if value is not None:
+                value = int(value)
+            return value
+
+
+    class TestTable(Base):
+        __tablename__ = "test"
+        id = Column(
+            Integer().with_variant(NumericAsInteger, "mssql"),
+            primary_key=True,
+            autoincrement=True,
+        )
+        name = Column(String)
+
+.. _mssql_insert_behavior:
+
+INSERT behavior
+^^^^^^^^^^^^^^^^
+
+Handling of the ``IDENTITY`` column at INSERT time involves two key
+techniques. The most common is being able to fetch the "last inserted value"
+for a given ``IDENTITY`` column, a process which SQLAlchemy performs
+implicitly in many cases, most importantly within the ORM.
+
+The process for fetching this value has several variants:
+
+* In the vast majority of cases, RETURNING is used in conjunction with INSERT
+  statements on SQL Server in order to get newly generated primary key values:
+
+  .. sourcecode:: sql
+
+    INSERT INTO t (x) OUTPUT inserted.id VALUES (?)
+
+  As of SQLAlchemy 2.0, the :ref:`engine_insertmanyvalues` feature is also
+  used by default to optimize many-row INSERT statements; for SQL Server
+  the feature takes place for both RETURNING and-non RETURNING
+  INSERT statements.
+
+  .. versionchanged:: 2.0.10 The :ref:`engine_insertmanyvalues` feature for
+     SQL Server was temporarily disabled for SQLAlchemy version 2.0.9 due to
+     issues with row ordering. As of 2.0.10 the feature is re-enabled, with
+     special case handling for the unit of work's requirement for RETURNING to
+     be ordered.
+
+* When RETURNING is not available or has been disabled via
+  ``implicit_returning=False``, either the ``scope_identity()`` function or
+  the ``@@identity`` variable is used; behavior varies by backend:
+
+  * when using PyODBC, the phrase ``; select scope_identity()`` will be
+    appended to the end of the INSERT statement; a second result set will be
+    fetched in order to receive the value.  Given a table as::
+
+        t = Table(
+            "t",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            implicit_returning=False,
+        )
+
+    an INSERT will look like:
+
+    .. sourcecode:: sql
+
+        INSERT INTO t (x) VALUES (?); select scope_identity()
+
+  * Other dialects such as pymssql will call upon
+    ``SELECT scope_identity() AS lastrowid`` subsequent to an INSERT
+    statement. If the flag ``use_scope_identity=False`` is passed to
+    :func:`_sa.create_engine`,
+    the statement ``SELECT @@identity AS lastrowid``
+    is used instead.
+
+A table that contains an ``IDENTITY`` column will prohibit an INSERT statement
+that refers to the identity column explicitly.  The SQLAlchemy dialect will
+detect when an INSERT construct, created using a core
+:func:`_expression.insert`
+construct (not a plain string SQL), refers to the identity column, and
+in this case will emit ``SET IDENTITY_INSERT ON`` prior to the insert
+statement proceeding, and ``SET IDENTITY_INSERT OFF`` subsequent to the
+execution.  Given this example::
+
+    m = MetaData()
+    t = Table(
+        "t", m, Column("id", Integer, primary_key=True), Column("x", Integer)
+    )
+    m.create_all(engine)
+
+    with engine.begin() as conn:
+        conn.execute(t.insert(), {"id": 1, "x": 1}, {"id": 2, "x": 2})
+
+The above column will be created with IDENTITY, however the INSERT statement
+we emit is specifying explicit values.  In the echo output we can see
+how SQLAlchemy handles this:
+
+.. sourcecode:: sql
+
+    CREATE TABLE t (
+        id INTEGER NOT NULL IDENTITY(1,1),
+        x INTEGER NULL,
+        PRIMARY KEY (id)
+    )
+
+    COMMIT
+    SET IDENTITY_INSERT t ON
+    INSERT INTO t (id, x) VALUES (?, ?)
+    ((1, 1), (2, 2))
+    SET IDENTITY_INSERT t OFF
+    COMMIT
+
+
+
+This is an auxiliary use case suitable for testing and bulk insert scenarios.
+
+SEQUENCE support
+----------------
+
+The :class:`.Sequence` object creates "real" sequences, i.e.,
+``CREATE SEQUENCE``:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import Sequence
+    >>> from sqlalchemy.schema import CreateSequence
+    >>> from sqlalchemy.dialects import mssql
+    >>> print(
+    ...     CreateSequence(Sequence("my_seq", start=1)).compile(
+    ...         dialect=mssql.dialect()
+    ...     )
+    ... )
+    {printsql}CREATE SEQUENCE my_seq START WITH 1
+
+For integer primary key generation, SQL Server's ``IDENTITY`` construct should
+generally be preferred vs. sequence.
+
+.. tip::
+
+    The default start value for T-SQL is ``-2**63`` instead of 1 as
+    in most other SQL databases. Users should explicitly set the
+    :paramref:`.Sequence.start` to 1 if that's the expected default::
+
+        seq = Sequence("my_sequence", start=1)
+
+.. versionadded:: 1.4 added SQL Server support for :class:`.Sequence`
+
+.. versionchanged:: 2.0 The SQL Server dialect will no longer implicitly
+   render "START WITH 1" for ``CREATE SEQUENCE``, which was the behavior
+   first implemented in version 1.4.
+
+MAX on VARCHAR / NVARCHAR
+-------------------------
+
+SQL Server supports the special string "MAX" within the
+:class:`_types.VARCHAR` and :class:`_types.NVARCHAR` datatypes,
+to indicate "maximum length possible".   The dialect currently handles this as
+a length of "None" in the base type, rather than supplying a
+dialect-specific version of these types, so that a base type
+specified such as ``VARCHAR(None)`` can assume "unlengthed" behavior on
+more than one backend without using dialect-specific types.
+
+To build a SQL Server VARCHAR or NVARCHAR with MAX length, use None::
+
+    my_table = Table(
+        "my_table",
+        metadata,
+        Column("my_data", VARCHAR(None)),
+        Column("my_n_data", NVARCHAR(None)),
+    )
+
+Collation Support
+-----------------
+
+Character collations are supported by the base string types,
+specified by the string argument "collation"::
+
+    from sqlalchemy import VARCHAR
+
+    Column("login", VARCHAR(32, collation="Latin1_General_CI_AS"))
+
+When such a column is associated with a :class:`_schema.Table`, the
+CREATE TABLE statement for this column will yield:
+
+.. sourcecode:: sql
+
+    login VARCHAR(32) COLLATE Latin1_General_CI_AS NULL
+
+LIMIT/OFFSET Support
+--------------------
+
+MSSQL has added support for LIMIT / OFFSET as of SQL Server 2012, via the
+"OFFSET n ROWS" and "FETCH NEXT n ROWS" clauses.  SQLAlchemy supports these
+syntaxes automatically if SQL Server 2012 or greater is detected.
+
+.. versionchanged:: 1.4 support added for SQL Server "OFFSET n ROWS" and
+   "FETCH NEXT n ROWS" syntax.
+
+For statements that specify only LIMIT and no OFFSET, all versions of SQL
+Server support the TOP keyword.   This syntax is used for all SQL Server
+versions when no OFFSET clause is present.  A statement such as::
+
+    select(some_table).limit(5)
+
+will render similarly to:
+
+.. sourcecode:: sql
+
+    SELECT TOP 5 col1, col2.. FROM table
+
+For versions of SQL Server prior to SQL Server 2012, a statement that uses
+LIMIT and OFFSET, or just OFFSET alone, will be rendered using the
+``ROW_NUMBER()`` window function.   A statement such as::
+
+    select(some_table).order_by(some_table.c.col3).limit(5).offset(10)
+
+will render similarly to:
+
+.. sourcecode:: sql
+
+    SELECT anon_1.col1, anon_1.col2 FROM (SELECT col1, col2,
+    ROW_NUMBER() OVER (ORDER BY col3) AS
+    mssql_rn FROM table WHERE t.x = :x_1) AS
+    anon_1 WHERE mssql_rn > :param_1 AND mssql_rn <= :param_2 + :param_1
+
+Note that when using LIMIT and/or OFFSET, whether using the older
+or newer SQL Server syntaxes, the statement must have an ORDER BY as well,
+else a :class:`.CompileError` is raised.
+
+.. _mssql_comment_support:
+
+DDL Comment Support
+--------------------
+
+Comment support, which includes DDL rendering for attributes such as
+:paramref:`_schema.Table.comment` and :paramref:`_schema.Column.comment`, as
+well as the ability to reflect these comments, is supported assuming a
+supported version of SQL Server is in use. If a non-supported version such as
+Azure Synapse is detected at first-connect time (based on the presence
+of the ``fn_listextendedproperty`` SQL function), comment support including
+rendering and table-comment reflection is disabled, as both features rely upon
+SQL Server stored procedures and functions that are not available on all
+backend types.
+
+To force comment support to be on or off, bypassing autodetection, set the
+parameter ``supports_comments`` within :func:`_sa.create_engine`::
+
+    e = create_engine("mssql+pyodbc://u:p@dsn", supports_comments=False)
+
+.. versionadded:: 2.0 Added support for table and column comments for
+   the SQL Server dialect, including DDL generation and reflection.
+
+.. _mssql_isolation_level:
+
+Transaction Isolation Level
+---------------------------
+
+All SQL Server dialects support setting of transaction isolation level
+both via a dialect-specific parameter
+:paramref:`_sa.create_engine.isolation_level`
+accepted by :func:`_sa.create_engine`,
+as well as the :paramref:`.Connection.execution_options.isolation_level`
+argument as passed to
+:meth:`_engine.Connection.execution_options`.
+This feature works by issuing the
+command ``SET TRANSACTION ISOLATION LEVEL <level>`` for
+each new connection.
+
+To set isolation level using :func:`_sa.create_engine`::
+
+    engine = create_engine(
+        "mssql+pyodbc://scott:tiger@ms_2008", isolation_level="REPEATABLE READ"
+    )
+
+To set using per-connection execution options::
+
+    connection = engine.connect()
+    connection = connection.execution_options(isolation_level="READ COMMITTED")
+
+Valid values for ``isolation_level`` include:
+
+* ``AUTOCOMMIT`` - pyodbc / pymssql-specific
+* ``READ COMMITTED``
+* ``READ UNCOMMITTED``
+* ``REPEATABLE READ``
+* ``SERIALIZABLE``
+* ``SNAPSHOT`` - specific to SQL Server
+
+There are also more options for isolation level configurations, such as
+"sub-engine" objects linked to a main :class:`_engine.Engine` which each apply
+different isolation level settings.  See the discussion at
+:ref:`dbapi_autocommit` for background.
+
+.. seealso::
+
+    :ref:`dbapi_autocommit`
+
+.. _mssql_reset_on_return:
+
+Temporary Table / Resource Reset for Connection Pooling
+-------------------------------------------------------
+
+The :class:`.QueuePool` connection pool implementation used
+by the SQLAlchemy :class:`.Engine` object includes
+:ref:`reset on return <pool_reset_on_return>` behavior that will invoke
+the DBAPI ``.rollback()`` method when connections are returned to the pool.
+While this rollback will clear out the immediate state used by the previous
+transaction, it does not cover a wider range of session-level state, including
+temporary tables as well as other server state such as prepared statement
+handles and statement caches.   An undocumented SQL Server procedure known
+as ``sp_reset_connection`` is known to be a workaround for this issue which
+will reset most of the session state that builds up on a connection, including
+temporary tables.
+
+To install ``sp_reset_connection`` as the means of performing reset-on-return,
+the :meth:`.PoolEvents.reset` event hook may be used, as demonstrated in the
+example below. The :paramref:`_sa.create_engine.pool_reset_on_return` parameter
+is set to ``None`` so that the custom scheme can replace the default behavior
+completely.   The custom hook implementation calls ``.rollback()`` in any case,
+as it's usually important that the DBAPI's own tracking of commit/rollback
+will remain consistent with the state of the transaction::
+
+    from sqlalchemy import create_engine
+    from sqlalchemy import event
+
+    mssql_engine = create_engine(
+        "mssql+pyodbc://scott:tiger^5HHH@mssql2017:1433/test?driver=ODBC+Driver+17+for+SQL+Server",
+        # disable default reset-on-return scheme
+        pool_reset_on_return=None,
+    )
+
+
+    @event.listens_for(mssql_engine, "reset")
+    def _reset_mssql(dbapi_connection, connection_record, reset_state):
+        if not reset_state.terminate_only:
+            dbapi_connection.execute("{call sys.sp_reset_connection}")
+
+        # so that the DBAPI itself knows that the connection has been
+        # reset
+        dbapi_connection.rollback()
+
+.. versionchanged:: 2.0.0b3  Added additional state arguments to
+   the :meth:`.PoolEvents.reset` event and additionally ensured the event
+   is invoked for all "reset" occurrences, so that it's appropriate
+   as a place for custom "reset" handlers.   Previous schemes which
+   use the :meth:`.PoolEvents.checkin` handler remain usable as well.
+
+.. seealso::
+
+    :ref:`pool_reset_on_return` - in the :ref:`pooling_toplevel` documentation
+
+Nullability
+-----------
+MSSQL has support for three levels of column nullability. The default
+nullability allows nulls and is explicit in the CREATE TABLE
+construct:
+
+.. sourcecode:: sql
+
+    name VARCHAR(20) NULL
+
+If ``nullable=None`` is specified then no specification is made. In
+other words the database's configured default is used. This will
+render:
+
+.. sourcecode:: sql
+
+    name VARCHAR(20)
+
+If ``nullable`` is ``True`` or ``False`` then the column will be
+``NULL`` or ``NOT NULL`` respectively.
+
+Date / Time Handling
+--------------------
+DATE and TIME are supported.   Bind parameters are converted
+to datetime.datetime() objects as required by most MSSQL drivers,
+and results are processed from strings if needed.
+The DATE and TIME types are not available for MSSQL 2005 and
+previous - if a server version below 2008 is detected, DDL
+for these types will be issued as DATETIME.
+
+.. _mssql_large_type_deprecation:
+
+Large Text/Binary Type Deprecation
+----------------------------------
+
+Per
+`SQL Server 2012/2014 Documentation <https://technet.microsoft.com/en-us/library/ms187993.aspx>`_,
+the ``NTEXT``, ``TEXT`` and ``IMAGE`` datatypes are to be removed from SQL
+Server in a future release.   SQLAlchemy normally relates these types to the
+:class:`.UnicodeText`, :class:`_expression.TextClause` and
+:class:`.LargeBinary` datatypes.
+
+In order to accommodate this change, a new flag ``deprecate_large_types``
+is added to the dialect, which will be automatically set based on detection
+of the server version in use, if not otherwise set by the user.  The
+behavior of this flag is as follows:
+
+* When this flag is ``True``, the :class:`.UnicodeText`,
+  :class:`_expression.TextClause` and
+  :class:`.LargeBinary` datatypes, when used to render DDL, will render the
+  types ``NVARCHAR(max)``, ``VARCHAR(max)``, and ``VARBINARY(max)``,
+  respectively.  This is a new behavior as of the addition of this flag.
+
+* When this flag is ``False``, the :class:`.UnicodeText`,
+  :class:`_expression.TextClause` and
+  :class:`.LargeBinary` datatypes, when used to render DDL, will render the
+  types ``NTEXT``, ``TEXT``, and ``IMAGE``,
+  respectively.  This is the long-standing behavior of these types.
+
+* The flag begins with the value ``None``, before a database connection is
+  established.   If the dialect is used to render DDL without the flag being
+  set, it is interpreted the same as ``False``.
+
+* On first connection, the dialect detects if SQL Server version 2012 or
+  greater is in use; if the flag is still at ``None``, it sets it to ``True``
+  or ``False`` based on whether 2012 or greater is detected.
+
+* The flag can be set to either ``True`` or ``False`` when the dialect
+  is created, typically via :func:`_sa.create_engine`::
+
+        eng = create_engine(
+            "mssql+pymssql://user:pass@host/db", deprecate_large_types=True
+        )
+
+* Complete control over whether the "old" or "new" types are rendered is
+  available in all SQLAlchemy versions by using the UPPERCASE type objects
+  instead: :class:`_types.NVARCHAR`, :class:`_types.VARCHAR`,
+  :class:`_types.VARBINARY`, :class:`_types.TEXT`, :class:`_mssql.NTEXT`,
+  :class:`_mssql.IMAGE`
+  will always remain fixed and always output exactly that
+  type.
+
+.. _multipart_schema_names:
+
+Multipart Schema Names
+----------------------
+
+SQL Server schemas sometimes require multiple parts to their "schema"
+qualifier, that is, including the database name and owner name as separate
+tokens, such as ``mydatabase.dbo.some_table``. These multipart names can be set
+at once using the :paramref:`_schema.Table.schema` argument of
+:class:`_schema.Table`::
+
+    Table(
+        "some_table",
+        metadata,
+        Column("q", String(50)),
+        schema="mydatabase.dbo",
+    )
+
+When performing operations such as table or component reflection, a schema
+argument that contains a dot will be split into separate
+"database" and "owner"  components in order to correctly query the SQL
+Server information schema tables, as these two values are stored separately.
+Additionally, when rendering the schema name for DDL or SQL, the two
+components will be quoted separately for case sensitive names and other
+special characters.   Given an argument as below::
+
+    Table(
+        "some_table",
+        metadata,
+        Column("q", String(50)),
+        schema="MyDataBase.dbo",
+    )
+
+The above schema would be rendered as ``[MyDataBase].dbo``, and also in
+reflection, would be reflected using "dbo" as the owner and "MyDataBase"
+as the database name.
+
+To control how the schema name is broken into database / owner,
+specify brackets (which in SQL Server are quoting characters) in the name.
+Below, the "owner" will be considered as ``MyDataBase.dbo`` and the
+"database" will be None::
+
+    Table(
+        "some_table",
+        metadata,
+        Column("q", String(50)),
+        schema="[MyDataBase.dbo]",
+    )
+
+To individually specify both database and owner name with special characters
+or embedded dots, use two sets of brackets::
+
+    Table(
+        "some_table",
+        metadata,
+        Column("q", String(50)),
+        schema="[MyDataBase.Period].[MyOwner.Dot]",
+    )
+
+.. versionchanged:: 1.2 the SQL Server dialect now treats brackets as
+   identifier delimiters splitting the schema into separate database
+   and owner tokens, to allow dots within either name itself.
+
+.. _legacy_schema_rendering:
+
+Legacy Schema Mode
+------------------
+
+Very old versions of the MSSQL dialect introduced the behavior such that a
+schema-qualified table would be auto-aliased when used in a
+SELECT statement; given a table::
+
+    account_table = Table(
+        "account",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("info", String(100)),
+        schema="customer_schema",
+    )
+
+this legacy mode of rendering would assume that "customer_schema.account"
+would not be accepted by all parts of the SQL statement, as illustrated
+below:
+
+.. sourcecode:: pycon+sql
+
+    >>> eng = create_engine("mssql+pymssql://mydsn", legacy_schema_aliasing=True)
+    >>> print(account_table.select().compile(eng))
+    {printsql}SELECT account_1.id, account_1.info
+    FROM customer_schema.account AS account_1
+
+This mode of behavior is now off by default, as it appears to have served
+no purpose; however in the case that legacy applications rely upon it,
+it is available using the ``legacy_schema_aliasing`` argument to
+:func:`_sa.create_engine` as illustrated above.
+
+.. deprecated:: 1.4
+
+   The ``legacy_schema_aliasing`` flag is now
+   deprecated and will be removed in a future release.
+
+.. _mssql_indexes:
+
+Clustered Index Support
+-----------------------
+
+The MSSQL dialect supports clustered indexes (and primary keys) via the
+``mssql_clustered`` option.  This option is available to :class:`.Index`,
+:class:`.UniqueConstraint`. and :class:`.PrimaryKeyConstraint`.
+For indexes this option can be combined with the ``mssql_columnstore`` one
+to create a clustered columnstore index.
+
+To generate a clustered index::
+
+    Index("my_index", table.c.x, mssql_clustered=True)
+
+which renders the index as ``CREATE CLUSTERED INDEX my_index ON table (x)``.
+
+To generate a clustered primary key use::
+
+    Table(
+        "my_table",
+        metadata,
+        Column("x", ...),
+        Column("y", ...),
+        PrimaryKeyConstraint("x", "y", mssql_clustered=True),
+    )
+
+which will render the table, for example, as:
+
+.. sourcecode:: sql
+
+  CREATE TABLE my_table (
+    x INTEGER NOT NULL,
+    y INTEGER NOT NULL,
+    PRIMARY KEY CLUSTERED (x, y)
+  )
+
+Similarly, we can generate a clustered unique constraint using::
+
+    Table(
+        "my_table",
+        metadata,
+        Column("x", ...),
+        Column("y", ...),
+        PrimaryKeyConstraint("x"),
+        UniqueConstraint("y", mssql_clustered=True),
+    )
+
+To explicitly request a non-clustered primary key (for example, when
+a separate clustered index is desired), use::
+
+    Table(
+        "my_table",
+        metadata,
+        Column("x", ...),
+        Column("y", ...),
+        PrimaryKeyConstraint("x", "y", mssql_clustered=False),
+    )
+
+which will render the table, for example, as:
+
+.. sourcecode:: sql
+
+  CREATE TABLE my_table (
+    x INTEGER NOT NULL,
+    y INTEGER NOT NULL,
+    PRIMARY KEY NONCLUSTERED (x, y)
+  )
+
+Columnstore Index Support
+-------------------------
+
+The MSSQL dialect supports columnstore indexes via the ``mssql_columnstore``
+option.  This option is available to :class:`.Index`. It be combined with
+the ``mssql_clustered`` option to create a clustered columnstore index.
+
+To generate a columnstore index::
+
+    Index("my_index", table.c.x, mssql_columnstore=True)
+
+which renders the index as ``CREATE COLUMNSTORE INDEX my_index ON table (x)``.
+
+To generate a clustered columnstore index provide no columns::
+
+    idx = Index("my_index", mssql_clustered=True, mssql_columnstore=True)
+    # required to associate the index with the table
+    table.append_constraint(idx)
+
+the above renders the index as
+``CREATE CLUSTERED COLUMNSTORE INDEX my_index ON table``.
+
+.. versionadded:: 2.0.18
+
+MSSQL-Specific Index Options
+-----------------------------
+
+In addition to clustering, the MSSQL dialect supports other special options
+for :class:`.Index`.
+
+INCLUDE
+^^^^^^^
+
+The ``mssql_include`` option renders INCLUDE(colname) for the given string
+names::
+
+    Index("my_index", table.c.x, mssql_include=["y"])
+
+would render the index as ``CREATE INDEX my_index ON table (x) INCLUDE (y)``
+
+.. _mssql_index_where:
+
+Filtered Indexes
+^^^^^^^^^^^^^^^^
+
+The ``mssql_where`` option renders WHERE(condition) for the given string
+names::
+
+    Index("my_index", table.c.x, mssql_where=table.c.x > 10)
+
+would render the index as ``CREATE INDEX my_index ON table (x) WHERE x > 10``.
+
+.. versionadded:: 1.3.4
+
+Index ordering
+^^^^^^^^^^^^^^
+
+Index ordering is available via functional expressions, such as::
+
+    Index("my_index", table.c.x.desc())
+
+would render the index as ``CREATE INDEX my_index ON table (x DESC)``
+
+.. seealso::
+
+    :ref:`schema_indexes_functional`
+
+Compatibility Levels
+--------------------
+MSSQL supports the notion of setting compatibility levels at the
+database level. This allows, for instance, to run a database that
+is compatible with SQL2000 while running on a SQL2005 database
+server. ``server_version_info`` will always return the database
+server version information (in this case SQL2005) and not the
+compatibility level information. Because of this, if running under
+a backwards compatibility mode SQLAlchemy may attempt to use T-SQL
+statements that are unable to be parsed by the database server.
+
+.. _mssql_triggers:
+
+Triggers
+--------
+
+SQLAlchemy by default uses OUTPUT INSERTED to get at newly
+generated primary key values via IDENTITY columns or other
+server side defaults.   MS-SQL does not
+allow the usage of OUTPUT INSERTED on tables that have triggers.
+To disable the usage of OUTPUT INSERTED on a per-table basis,
+specify ``implicit_returning=False`` for each :class:`_schema.Table`
+which has triggers::
+
+    Table(
+        "mytable",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        # ...,
+        implicit_returning=False,
+    )
+
+Declarative form::
+
+    class MyClass(Base):
+        # ...
+        __table_args__ = {"implicit_returning": False}
+
+.. _mssql_rowcount_versioning:
+
+Rowcount Support / ORM Versioning
+---------------------------------
+
+The SQL Server drivers may have limited ability to return the number
+of rows updated from an UPDATE or DELETE statement.
+
+As of this writing, the PyODBC driver is not able to return a rowcount when
+OUTPUT INSERTED is used.    Previous versions of SQLAlchemy therefore had
+limitations for features such as the "ORM Versioning" feature that relies upon
+accurate rowcounts in order to match version numbers with matched rows.
+
+SQLAlchemy 2.0 now retrieves the "rowcount" manually for these particular use
+cases based on counting the rows that arrived back within RETURNING; so while
+the driver still has this limitation, the ORM Versioning feature is no longer
+impacted by it. As of SQLAlchemy 2.0.5, ORM versioning has been fully
+re-enabled for the pyodbc driver.
+
+.. versionchanged:: 2.0.5  ORM versioning support is restored for the pyodbc
+   driver.  Previously, a warning would be emitted during ORM flush that
+   versioning was not supported.
+
+
+Enabling Snapshot Isolation
+---------------------------
+
+SQL Server has a default transaction
+isolation mode that locks entire tables, and causes even mildly concurrent
+applications to have long held locks and frequent deadlocks.
+Enabling snapshot isolation for the database as a whole is recommended
+for modern levels of concurrency support.  This is accomplished via the
+following ALTER DATABASE commands executed at the SQL prompt:
+
+.. sourcecode:: sql
+
+    ALTER DATABASE MyDatabase SET ALLOW_SNAPSHOT_ISOLATION ON
+
+    ALTER DATABASE MyDatabase SET READ_COMMITTED_SNAPSHOT ON
+
+Background on SQL Server snapshot isolation is available at
+https://msdn.microsoft.com/en-us/library/ms175095.aspx.
+
+"""  # noqa
+
+from __future__ import annotations
+
+import codecs
+import datetime
+import operator
+import re
+from typing import overload
+from typing import TYPE_CHECKING
+from uuid import UUID as _python_UUID
+
+from . import information_schema as ischema
+from .json import JSON
+from .json import JSONIndexType
+from .json import JSONPathType
+from ... import exc
+from ... import Identity
+from ... import schema as sa_schema
+from ... import Sequence
+from ... import sql
+from ... import text
+from ... import util
+from ...engine import cursor as _cursor
+from ...engine import default
+from ...engine import reflection
+from ...engine.reflection import ReflectionDefaults
+from ...sql import coercions
+from ...sql import compiler
+from ...sql import elements
+from ...sql import expression
+from ...sql import func
+from ...sql import quoted_name
+from ...sql import roles
+from ...sql import sqltypes
+from ...sql import try_cast as try_cast  # noqa: F401
+from ...sql import util as sql_util
+from ...sql._typing import is_sql_compiler
+from ...sql.compiler import InsertmanyvaluesSentinelOpts
+from ...sql.elements import TryCast as TryCast  # noqa: F401
+from ...types import BIGINT
+from ...types import BINARY
+from ...types import CHAR
+from ...types import DATE
+from ...types import DATETIME
+from ...types import DECIMAL
+from ...types import FLOAT
+from ...types import INTEGER
+from ...types import NCHAR
+from ...types import NUMERIC
+from ...types import NVARCHAR
+from ...types import SMALLINT
+from ...types import TEXT
+from ...types import VARCHAR
+from ...util import update_wrapper
+from ...util.typing import Literal
+
+if TYPE_CHECKING:
+    from ...sql.dml import DMLState
+    from ...sql.selectable import TableClause
+
+# https://sqlserverbuilds.blogspot.com/
+MS_2017_VERSION = (14,)
+MS_2016_VERSION = (13,)
+MS_2014_VERSION = (12,)
+MS_2012_VERSION = (11,)
+MS_2008_VERSION = (10,)
+MS_2005_VERSION = (9,)
+MS_2000_VERSION = (8,)
+
+RESERVED_WORDS = {
+    "add",
+    "all",
+    "alter",
+    "and",
+    "any",
+    "as",
+    "asc",
+    "authorization",
+    "backup",
+    "begin",
+    "between",
+    "break",
+    "browse",
+    "bulk",
+    "by",
+    "cascade",
+    "case",
+    "check",
+    "checkpoint",
+    "close",
+    "clustered",
+    "coalesce",
+    "collate",
+    "column",
+    "commit",
+    "compute",
+    "constraint",
+    "contains",
+    "containstable",
+    "continue",
+    "convert",
+    "create",
+    "cross",
+    "current",
+    "current_date",
+    "current_time",
+    "current_timestamp",
+    "current_user",
+    "cursor",
+    "database",
+    "dbcc",
+    "deallocate",
+    "declare",
+    "default",
+    "delete",
+    "deny",
+    "desc",
+    "disk",
+    "distinct",
+    "distributed",
+    "double",
+    "drop",
+    "dump",
+    "else",
+    "end",
+    "errlvl",
+    "escape",
+    "except",
+    "exec",
+    "execute",
+    "exists",
+    "exit",
+    "external",
+    "fetch",
+    "file",
+    "fillfactor",
+    "for",
+    "foreign",
+    "freetext",
+    "freetexttable",
+    "from",
+    "full",
+    "function",
+    "goto",
+    "grant",
+    "group",
+    "having",
+    "holdlock",
+    "identity",
+    "identity_insert",
+    "identitycol",
+    "if",
+    "in",
+    "index",
+    "inner",
+    "insert",
+    "intersect",
+    "into",
+    "is",
+    "join",
+    "key",
+    "kill",
+    "left",
+    "like",
+    "lineno",
+    "load",
+    "merge",
+    "national",
+    "nocheck",
+    "nonclustered",
+    "not",
+    "null",
+    "nullif",
+    "of",
+    "off",
+    "offsets",
+    "on",
+    "open",
+    "opendatasource",
+    "openquery",
+    "openrowset",
+    "openxml",
+    "option",
+    "or",
+    "order",
+    "outer",
+    "over",
+    "percent",
+    "pivot",
+    "plan",
+    "precision",
+    "primary",
+    "print",
+    "proc",
+    "procedure",
+    "public",
+    "raiserror",
+    "read",
+    "readtext",
+    "reconfigure",
+    "references",
+    "replication",
+    "restore",
+    "restrict",
+    "return",
+    "revert",
+    "revoke",
+    "right",
+    "rollback",
+    "rowcount",
+    "rowguidcol",
+    "rule",
+    "save",
+    "schema",
+    "securityaudit",
+    "select",
+    "session_user",
+    "set",
+    "setuser",
+    "shutdown",
+    "some",
+    "statistics",
+    "system_user",
+    "table",
+    "tablesample",
+    "textsize",
+    "then",
+    "to",
+    "top",
+    "tran",
+    "transaction",
+    "trigger",
+    "truncate",
+    "tsequal",
+    "union",
+    "unique",
+    "unpivot",
+    "update",
+    "updatetext",
+    "use",
+    "user",
+    "values",
+    "varying",
+    "view",
+    "waitfor",
+    "when",
+    "where",
+    "while",
+    "with",
+    "writetext",
+}
+
+
+class REAL(sqltypes.REAL):
+    """the SQL Server REAL datatype."""
+
+    def __init__(self, **kw):
+        # REAL is a synonym for FLOAT(24) on SQL server.
+        # it is only accepted as the word "REAL" in DDL, the numeric
+        # precision value is not allowed to be present
+        kw.setdefault("precision", 24)
+        super().__init__(**kw)
+
+
+class DOUBLE_PRECISION(sqltypes.DOUBLE_PRECISION):
+    """the SQL Server DOUBLE PRECISION datatype.
+
+    .. versionadded:: 2.0.11
+
+    """
+
+    def __init__(self, **kw):
+        # DOUBLE PRECISION is a synonym for FLOAT(53) on SQL server.
+        # it is only accepted as the word "DOUBLE PRECISION" in DDL,
+        # the numeric precision value is not allowed to be present
+        kw.setdefault("precision", 53)
+        super().__init__(**kw)
+
+
+class TINYINT(sqltypes.Integer):
+    __visit_name__ = "TINYINT"
+
+
+# MSSQL DATE/TIME types have varied behavior, sometimes returning
+# strings.  MSDate/TIME check for everything, and always
+# filter bind parameters into datetime objects (required by pyodbc,
+# not sure about other dialects).
+
+
+class _MSDate(sqltypes.Date):
+    def bind_processor(self, dialect):
+        def process(value):
+            if type(value) == datetime.date:
+                return datetime.datetime(value.year, value.month, value.day)
+            else:
+                return value
+
+        return process
+
+    _reg = re.compile(r"(\d+)-(\d+)-(\d+)")
+
+    def result_processor(self, dialect, coltype):
+        def process(value):
+            if isinstance(value, datetime.datetime):
+                return value.date()
+            elif isinstance(value, str):
+                m = self._reg.match(value)
+                if not m:
+                    raise ValueError(
+                        "could not parse %r as a date value" % (value,)
+                    )
+                return datetime.date(*[int(x or 0) for x in m.groups()])
+            else:
+                return value
+
+        return process
+
+
+class TIME(sqltypes.TIME):
+    def __init__(self, precision=None, **kwargs):
+        self.precision = precision
+        super().__init__()
+
+    __zero_date = datetime.date(1900, 1, 1)
+
+    def bind_processor(self, dialect):
+        def process(value):
+            if isinstance(value, datetime.datetime):
+                value = datetime.datetime.combine(
+                    self.__zero_date, value.time()
+                )
+            elif isinstance(value, datetime.time):
+                """issue #5339
+                per: https://github.com/mkleehammer/pyodbc/wiki/Tips-and-Tricks-by-Database-Platform#time-columns
+                pass TIME value as string
+                """  # noqa
+                value = str(value)
+            return value
+
+        return process
+
+    _reg = re.compile(r"(\d+):(\d+):(\d+)(?:\.(\d{0,6}))?")
+
+    def result_processor(self, dialect, coltype):
+        def process(value):
+            if isinstance(value, datetime.datetime):
+                return value.time()
+            elif isinstance(value, str):
+                m = self._reg.match(value)
+                if not m:
+                    raise ValueError(
+                        "could not parse %r as a time value" % (value,)
+                    )
+                return datetime.time(*[int(x or 0) for x in m.groups()])
+            else:
+                return value
+
+        return process
+
+
+_MSTime = TIME
+
+
+class _BASETIMEIMPL(TIME):
+    __visit_name__ = "_BASETIMEIMPL"
+
+
+class _DateTimeBase:
+    def bind_processor(self, dialect):
+        def process(value):
+            if type(value) == datetime.date:
+                return datetime.datetime(value.year, value.month, value.day)
+            else:
+                return value
+
+        return process
+
+
+class _MSDateTime(_DateTimeBase, sqltypes.DateTime):
+    pass
+
+
+class SMALLDATETIME(_DateTimeBase, sqltypes.DateTime):
+    __visit_name__ = "SMALLDATETIME"
+
+
+class DATETIME2(_DateTimeBase, sqltypes.DateTime):
+    __visit_name__ = "DATETIME2"
+
+    def __init__(self, precision=None, **kw):
+        super().__init__(**kw)
+        self.precision = precision
+
+
+class DATETIMEOFFSET(_DateTimeBase, sqltypes.DateTime):
+    __visit_name__ = "DATETIMEOFFSET"
+
+    def __init__(self, precision=None, **kw):
+        super().__init__(**kw)
+        self.precision = precision
+
+
+class _UnicodeLiteral:
+    def literal_processor(self, dialect):
+        def process(value):
+            value = value.replace("'", "''")
+
+            if dialect.identifier_preparer._double_percents:
+                value = value.replace("%", "%%")
+
+            return "N'%s'" % value
+
+        return process
+
+
+class _MSUnicode(_UnicodeLiteral, sqltypes.Unicode):
+    pass
+
+
+class _MSUnicodeText(_UnicodeLiteral, sqltypes.UnicodeText):
+    pass
+
+
+class TIMESTAMP(sqltypes._Binary):
+    """Implement the SQL Server TIMESTAMP type.
+
+    Note this is **completely different** than the SQL Standard
+    TIMESTAMP type, which is not supported by SQL Server.  It
+    is a read-only datatype that does not support INSERT of values.
+
+    .. versionadded:: 1.2
+
+    .. seealso::
+
+        :class:`_mssql.ROWVERSION`
+
+    """
+
+    __visit_name__ = "TIMESTAMP"
+
+    # expected by _Binary to be present
+    length = None
+
+    def __init__(self, convert_int=False):
+        """Construct a TIMESTAMP or ROWVERSION type.
+
+        :param convert_int: if True, binary integer values will
+         be converted to integers on read.
+
+        .. versionadded:: 1.2
+
+        """
+        self.convert_int = convert_int
+
+    def result_processor(self, dialect, coltype):
+        super_ = super().result_processor(dialect, coltype)
+        if self.convert_int:
+
+            def process(value):
+                if super_:
+                    value = super_(value)
+                if value is not None:
+                    # https://stackoverflow.com/a/30403242/34549
+                    value = int(codecs.encode(value, "hex"), 16)
+                return value
+
+            return process
+        else:
+            return super_
+
+
+class ROWVERSION(TIMESTAMP):
+    """Implement the SQL Server ROWVERSION type.
+
+    The ROWVERSION datatype is a SQL Server synonym for the TIMESTAMP
+    datatype, however current SQL Server documentation suggests using
+    ROWVERSION for new datatypes going forward.
+
+    The ROWVERSION datatype does **not** reflect (e.g. introspect) from the
+    database as itself; the returned datatype will be
+    :class:`_mssql.TIMESTAMP`.
+
+    This is a read-only datatype that does not support INSERT of values.
+
+    .. versionadded:: 1.2
+
+    .. seealso::
+
+        :class:`_mssql.TIMESTAMP`
+
+    """
+
+    __visit_name__ = "ROWVERSION"
+
+
+class NTEXT(sqltypes.UnicodeText):
+    """MSSQL NTEXT type, for variable-length unicode text up to 2^30
+    characters."""
+
+    __visit_name__ = "NTEXT"
+
+
+class VARBINARY(sqltypes.VARBINARY, sqltypes.LargeBinary):
+    """The MSSQL VARBINARY type.
+
+    This type adds additional features to the core :class:`_types.VARBINARY`
+    type, including "deprecate_large_types" mode where
+    either ``VARBINARY(max)`` or IMAGE is rendered, as well as the SQL
+    Server ``FILESTREAM`` option.
+
+    .. seealso::
+
+        :ref:`mssql_large_type_deprecation`
+
+    """
+
+    __visit_name__ = "VARBINARY"
+
+    def __init__(self, length=None, filestream=False):
+        """
+        Construct a VARBINARY type.
+
+        :param length: optional, a length for the column for use in
+          DDL statements, for those binary types that accept a length,
+          such as the MySQL BLOB type.
+
+        :param filestream=False: if True, renders the ``FILESTREAM`` keyword
+          in the table definition. In this case ``length`` must be ``None``
+          or ``'max'``.
+
+          .. versionadded:: 1.4.31
+
+        """
+
+        self.filestream = filestream
+        if self.filestream and length not in (None, "max"):
+            raise ValueError(
+                "length must be None or 'max' when setting filestream"
+            )
+        super().__init__(length=length)
+
+
+class IMAGE(sqltypes.LargeBinary):
+    __visit_name__ = "IMAGE"
+
+
+class XML(sqltypes.Text):
+    """MSSQL XML type.
+
+    This is a placeholder type for reflection purposes that does not include
+    any Python-side datatype support.   It also does not currently support
+    additional arguments, such as "CONTENT", "DOCUMENT",
+    "xml_schema_collection".
+
+    """
+
+    __visit_name__ = "XML"
+
+
+class BIT(sqltypes.Boolean):
+    """MSSQL BIT type.
+
+    Both pyodbc and pymssql return values from BIT columns as
+    Python <class 'bool'> so just subclass Boolean.
+
+    """
+
+    __visit_name__ = "BIT"
+
+
+class MONEY(sqltypes.TypeEngine):
+    __visit_name__ = "MONEY"
+
+
+class SMALLMONEY(sqltypes.TypeEngine):
+    __visit_name__ = "SMALLMONEY"
+
+
+class MSUUid(sqltypes.Uuid):
+    def bind_processor(self, dialect):
+        if self.native_uuid:
+            # this is currently assuming pyodbc; might not work for
+            # some other mssql driver
+            return None
+        else:
+            if self.as_uuid:
+
+                def process(value):
+                    if value is not None:
+                        value = value.hex
+                    return value
+
+                return process
+            else:
+
+                def process(value):
+                    if value is not None:
+                        value = value.replace("-", "").replace("''", "'")
+                    return value
+
+                return process
+
+    def literal_processor(self, dialect):
+        if self.native_uuid:
+
+            def process(value):
+                return f"""'{str(value).replace("''", "'")}'"""
+
+            return process
+        else:
+            if self.as_uuid:
+
+                def process(value):
+                    return f"""'{value.hex}'"""
+
+                return process
+            else:
+
+                def process(value):
+                    return f"""'{
+                        value.replace("-", "").replace("'", "''")
+                    }'"""
+
+                return process
+
+
+class UNIQUEIDENTIFIER(sqltypes.Uuid[sqltypes._UUID_RETURN]):
+    __visit_name__ = "UNIQUEIDENTIFIER"
+
+    @overload
+    def __init__(
+        self: UNIQUEIDENTIFIER[_python_UUID], as_uuid: Literal[True] = ...
+    ): ...
+
+    @overload
+    def __init__(
+        self: UNIQUEIDENTIFIER[str], as_uuid: Literal[False] = ...
+    ): ...
+
+    def __init__(self, as_uuid: bool = True):
+        """Construct a :class:`_mssql.UNIQUEIDENTIFIER` type.
+
+
+        :param as_uuid=True: if True, values will be interpreted
+         as Python uuid objects, converting to/from string via the
+         DBAPI.
+
+         .. versionchanged: 2.0 Added direct "uuid" support to the
+            :class:`_mssql.UNIQUEIDENTIFIER` datatype; uuid interpretation
+            defaults to ``True``.
+
+        """
+        self.as_uuid = as_uuid
+        self.native_uuid = True
+
+
+class SQL_VARIANT(sqltypes.TypeEngine):
+    __visit_name__ = "SQL_VARIANT"
+
+
+# old names.
+MSDateTime = _MSDateTime
+MSDate = _MSDate
+MSReal = REAL
+MSTinyInteger = TINYINT
+MSTime = TIME
+MSSmallDateTime = SMALLDATETIME
+MSDateTime2 = DATETIME2
+MSDateTimeOffset = DATETIMEOFFSET
+MSText = TEXT
+MSNText = NTEXT
+MSString = VARCHAR
+MSNVarchar = NVARCHAR
+MSChar = CHAR
+MSNChar = NCHAR
+MSBinary = BINARY
+MSVarBinary = VARBINARY
+MSImage = IMAGE
+MSBit = BIT
+MSMoney = MONEY
+MSSmallMoney = SMALLMONEY
+MSUniqueIdentifier = UNIQUEIDENTIFIER
+MSVariant = SQL_VARIANT
+
+ischema_names = {
+    "int": INTEGER,
+    "bigint": BIGINT,
+    "smallint": SMALLINT,
+    "tinyint": TINYINT,
+    "varchar": VARCHAR,
+    "nvarchar": NVARCHAR,
+    "char": CHAR,
+    "nchar": NCHAR,
+    "text": TEXT,
+    "ntext": NTEXT,
+    "decimal": DECIMAL,
+    "numeric": NUMERIC,
+    "float": FLOAT,
+    "datetime": DATETIME,
+    "datetime2": DATETIME2,
+    "datetimeoffset": DATETIMEOFFSET,
+    "date": DATE,
+    "time": TIME,
+    "smalldatetime": SMALLDATETIME,
+    "binary": BINARY,
+    "varbinary": VARBINARY,
+    "bit": BIT,
+    "real": REAL,
+    "double precision": DOUBLE_PRECISION,
+    "image": IMAGE,
+    "xml": XML,
+    "timestamp": TIMESTAMP,
+    "money": MONEY,
+    "smallmoney": SMALLMONEY,
+    "uniqueidentifier": UNIQUEIDENTIFIER,
+    "sql_variant": SQL_VARIANT,
+}
+
+
+class MSTypeCompiler(compiler.GenericTypeCompiler):
+    def _extend(self, spec, type_, length=None):
+        """Extend a string-type declaration with standard SQL
+        COLLATE annotations.
+
+        """
+
+        if getattr(type_, "collation", None):
+            collation = "COLLATE %s" % type_.collation
+        else:
+            collation = None
+
+        if not length:
+            length = type_.length
+
+        if length:
+            spec = spec + "(%s)" % length
+
+        return " ".join([c for c in (spec, collation) if c is not None])
+
+    def visit_double(self, type_, **kw):
+        return self.visit_DOUBLE_PRECISION(type_, **kw)
+
+    def visit_FLOAT(self, type_, **kw):
+        precision = getattr(type_, "precision", None)
+        if precision is None:
+            return "FLOAT"
+        else:
+            return "FLOAT(%(precision)s)" % {"precision": precision}
+
+    def visit_TINYINT(self, type_, **kw):
+        return "TINYINT"
+
+    def visit_TIME(self, type_, **kw):
+        precision = getattr(type_, "precision", None)
+        if precision is not None:
+            return "TIME(%s)" % precision
+        else:
+            return "TIME"
+
+    def visit_TIMESTAMP(self, type_, **kw):
+        return "TIMESTAMP"
+
+    def visit_ROWVERSION(self, type_, **kw):
+        return "ROWVERSION"
+
+    def visit_datetime(self, type_, **kw):
+        if type_.timezone:
+            return self.visit_DATETIMEOFFSET(type_, **kw)
+        else:
+            return self.visit_DATETIME(type_, **kw)
+
+    def visit_DATETIMEOFFSET(self, type_, **kw):
+        precision = getattr(type_, "precision", None)
+        if precision is not None:
+            return "DATETIMEOFFSET(%s)" % type_.precision
+        else:
+            return "DATETIMEOFFSET"
+
+    def visit_DATETIME2(self, type_, **kw):
+        precision = getattr(type_, "precision", None)
+        if precision is not None:
+            return "DATETIME2(%s)" % precision
+        else:
+            return "DATETIME2"
+
+    def visit_SMALLDATETIME(self, type_, **kw):
+        return "SMALLDATETIME"
+
+    def visit_unicode(self, type_, **kw):
+        return self.visit_NVARCHAR(type_, **kw)
+
+    def visit_text(self, type_, **kw):
+        if self.dialect.deprecate_large_types:
+            return self.visit_VARCHAR(type_, **kw)
+        else:
+            return self.visit_TEXT(type_, **kw)
+
+    def visit_unicode_text(self, type_, **kw):
+        if self.dialect.deprecate_large_types:
+            return self.visit_NVARCHAR(type_, **kw)
+        else:
+            return self.visit_NTEXT(type_, **kw)
+
+    def visit_NTEXT(self, type_, **kw):
+        return self._extend("NTEXT", type_)
+
+    def visit_TEXT(self, type_, **kw):
+        return self._extend("TEXT", type_)
+
+    def visit_VARCHAR(self, type_, **kw):
+        return self._extend("VARCHAR", type_, length=type_.length or "max")
+
+    def visit_CHAR(self, type_, **kw):
+        return self._extend("CHAR", type_)
+
+    def visit_NCHAR(self, type_, **kw):
+        return self._extend("NCHAR", type_)
+
+    def visit_NVARCHAR(self, type_, **kw):
+        return self._extend("NVARCHAR", type_, length=type_.length or "max")
+
+    def visit_date(self, type_, **kw):
+        if self.dialect.server_version_info < MS_2008_VERSION:
+            return self.visit_DATETIME(type_, **kw)
+        else:
+            return self.visit_DATE(type_, **kw)
+
+    def visit__BASETIMEIMPL(self, type_, **kw):
+        return self.visit_time(type_, **kw)
+
+    def visit_time(self, type_, **kw):
+        if self.dialect.server_version_info < MS_2008_VERSION:
+            return self.visit_DATETIME(type_, **kw)
+        else:
+            return self.visit_TIME(type_, **kw)
+
+    def visit_large_binary(self, type_, **kw):
+        if self.dialect.deprecate_large_types:
+            return self.visit_VARBINARY(type_, **kw)
+        else:
+            return self.visit_IMAGE(type_, **kw)
+
+    def visit_IMAGE(self, type_, **kw):
+        return "IMAGE"
+
+    def visit_XML(self, type_, **kw):
+        return "XML"
+
+    def visit_VARBINARY(self, type_, **kw):
+        text = self._extend("VARBINARY", type_, length=type_.length or "max")
+        if getattr(type_, "filestream", False):
+            text += " FILESTREAM"
+        return text
+
+    def visit_boolean(self, type_, **kw):
+        return self.visit_BIT(type_)
+
+    def visit_BIT(self, type_, **kw):
+        return "BIT"
+
+    def visit_JSON(self, type_, **kw):
+        # this is a bit of a break with SQLAlchemy's convention of
+        # "UPPERCASE name goes to UPPERCASE type name with no modification"
+        return self._extend("NVARCHAR", type_, length="max")
+
+    def visit_MONEY(self, type_, **kw):
+        return "MONEY"
+
+    def visit_SMALLMONEY(self, type_, **kw):
+        return "SMALLMONEY"
+
+    def visit_uuid(self, type_, **kw):
+        if type_.native_uuid:
+            return self.visit_UNIQUEIDENTIFIER(type_, **kw)
+        else:
+            return super().visit_uuid(type_, **kw)
+
+    def visit_UNIQUEIDENTIFIER(self, type_, **kw):
+        return "UNIQUEIDENTIFIER"
+
+    def visit_SQL_VARIANT(self, type_, **kw):
+        return "SQL_VARIANT"
+
+
+class MSExecutionContext(default.DefaultExecutionContext):
+    _enable_identity_insert = False
+    _select_lastrowid = False
+    _lastrowid = None
+
+    dialect: MSDialect
+
+    def _opt_encode(self, statement):
+        if self.compiled and self.compiled.schema_translate_map:
+            rst = self.compiled.preparer._render_schema_translates
+            statement = rst(statement, self.compiled.schema_translate_map)
+
+        return statement
+
+    def pre_exec(self):
+        """Activate IDENTITY_INSERT if needed."""
+
+        if self.isinsert:
+            if TYPE_CHECKING:
+                assert is_sql_compiler(self.compiled)
+                assert isinstance(self.compiled.compile_state, DMLState)
+                assert isinstance(
+                    self.compiled.compile_state.dml_table, TableClause
+                )
+
+            tbl = self.compiled.compile_state.dml_table
+            id_column = tbl._autoincrement_column
+
+            if id_column is not None and (
+                not isinstance(id_column.default, Sequence)
+            ):
+                insert_has_identity = True
+                compile_state = self.compiled.dml_compile_state
+                self._enable_identity_insert = (
+                    id_column.key in self.compiled_parameters[0]
+                ) or (
+                    compile_state._dict_parameters
+                    and (id_column.key in compile_state._insert_col_keys)
+                )
+
+            else:
+                insert_has_identity = False
+                self._enable_identity_insert = False
+
+            self._select_lastrowid = (
+                not self.compiled.inline
+                and insert_has_identity
+                and not self.compiled.effective_returning
+                and not self._enable_identity_insert
+                and not self.executemany
+            )
+
+            if self._enable_identity_insert:
+                self.root_connection._cursor_execute(
+                    self.cursor,
+                    self._opt_encode(
+                        "SET IDENTITY_INSERT %s ON"
+                        % self.identifier_preparer.format_table(tbl)
+                    ),
+                    (),
+                    self,
+                )
+
+    def post_exec(self):
+        """Disable IDENTITY_INSERT if enabled."""
+
+        conn = self.root_connection
+
+        if self.isinsert or self.isupdate or self.isdelete:
+            self._rowcount = self.cursor.rowcount
+
+        if self._select_lastrowid:
+            if self.dialect.use_scope_identity:
+                conn._cursor_execute(
+                    self.cursor,
+                    "SELECT scope_identity() AS lastrowid",
+                    (),
+                    self,
+                )
+            else:
+                conn._cursor_execute(
+                    self.cursor, "SELECT @@identity AS lastrowid", (), self
+                )
+            # fetchall() ensures the cursor is consumed without closing it
+            row = self.cursor.fetchall()[0]
+            self._lastrowid = int(row[0])
+
+            self.cursor_fetch_strategy = _cursor._NO_CURSOR_DML
+        elif (
+            self.compiled is not None
+            and is_sql_compiler(self.compiled)
+            and self.compiled.effective_returning
+        ):
+            self.cursor_fetch_strategy = (
+                _cursor.FullyBufferedCursorFetchStrategy(
+                    self.cursor,
+                    self.cursor.description,
+                    self.cursor.fetchall(),
+                )
+            )
+
+        if self._enable_identity_insert:
+            if TYPE_CHECKING:
+                assert is_sql_compiler(self.compiled)
+                assert isinstance(self.compiled.compile_state, DMLState)
+                assert isinstance(
+                    self.compiled.compile_state.dml_table, TableClause
+                )
+            conn._cursor_execute(
+                self.cursor,
+                self._opt_encode(
+                    "SET IDENTITY_INSERT %s OFF"
+                    % self.identifier_preparer.format_table(
+                        self.compiled.compile_state.dml_table
+                    )
+                ),
+                (),
+                self,
+            )
+
+    def get_lastrowid(self):
+        return self._lastrowid
+
+    def handle_dbapi_exception(self, e):
+        if self._enable_identity_insert:
+            try:
+                self.cursor.execute(
+                    self._opt_encode(
+                        "SET IDENTITY_INSERT %s OFF"
+                        % self.identifier_preparer.format_table(
+                            self.compiled.compile_state.dml_table
+                        )
+                    )
+                )
+            except Exception:
+                pass
+
+    def fire_sequence(self, seq, type_):
+        return self._execute_scalar(
+            (
+                "SELECT NEXT VALUE FOR %s"
+                % self.identifier_preparer.format_sequence(seq)
+            ),
+            type_,
+        )
+
+    def get_insert_default(self, column):
+        if (
+            isinstance(column, sa_schema.Column)
+            and column is column.table._autoincrement_column
+            and isinstance(column.default, sa_schema.Sequence)
+            and column.default.optional
+        ):
+            return None
+        return super().get_insert_default(column)
+
+
+class MSSQLCompiler(compiler.SQLCompiler):
+    returning_precedes_values = True
+
+    extract_map = util.update_copy(
+        compiler.SQLCompiler.extract_map,
+        {
+            "doy": "dayofyear",
+            "dow": "weekday",
+            "milliseconds": "millisecond",
+            "microseconds": "microsecond",
+        },
+    )
+
+    def __init__(self, *args, **kwargs):
+        self.tablealiases = {}
+        super().__init__(*args, **kwargs)
+
+    def _format_frame_clause(self, range_, **kw):
+        kw["literal_execute"] = True
+        return super()._format_frame_clause(range_, **kw)
+
+    def _with_legacy_schema_aliasing(fn):
+        def decorate(self, *arg, **kw):
+            if self.dialect.legacy_schema_aliasing:
+                return fn(self, *arg, **kw)
+            else:
+                super_ = getattr(super(MSSQLCompiler, self), fn.__name__)
+                return super_(*arg, **kw)
+
+        return decorate
+
+    def visit_now_func(self, fn, **kw):
+        return "CURRENT_TIMESTAMP"
+
+    def visit_current_date_func(self, fn, **kw):
+        return "GETDATE()"
+
+    def visit_length_func(self, fn, **kw):
+        return "LEN%s" % self.function_argspec(fn, **kw)
+
+    def visit_char_length_func(self, fn, **kw):
+        return "LEN%s" % self.function_argspec(fn, **kw)
+
+    def visit_aggregate_strings_func(self, fn, **kw):
+        expr = fn.clauses.clauses[0]._compiler_dispatch(self, **kw)
+        kw["literal_execute"] = True
+        delimeter = fn.clauses.clauses[1]._compiler_dispatch(self, **kw)
+        return f"string_agg({expr}, {delimeter})"
+
+    def visit_concat_op_expression_clauselist(
+        self, clauselist, operator, **kw
+    ):
+        return " + ".join(self.process(elem, **kw) for elem in clauselist)
+
+    def visit_concat_op_binary(self, binary, operator, **kw):
+        return "%s + %s" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_true(self, expr, **kw):
+        return "1"
+
+    def visit_false(self, expr, **kw):
+        return "0"
+
+    def visit_match_op_binary(self, binary, operator, **kw):
+        return "CONTAINS (%s, %s)" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def get_select_precolumns(self, select, **kw):
+        """MS-SQL puts TOP, it's version of LIMIT here"""
+
+        s = super().get_select_precolumns(select, **kw)
+
+        if select._has_row_limiting_clause and self._use_top(select):
+            # ODBC drivers and possibly others
+            # don't support bind params in the SELECT clause on SQL Server.
+            # so have to use literal here.
+            kw["literal_execute"] = True
+            s += "TOP %s " % self.process(
+                self._get_limit_or_fetch(select), **kw
+            )
+            if select._fetch_clause is not None:
+                if select._fetch_clause_options["percent"]:
+                    s += "PERCENT "
+                if select._fetch_clause_options["with_ties"]:
+                    s += "WITH TIES "
+
+        return s
+
+    def get_from_hint_text(self, table, text):
+        return text
+
+    def get_crud_hint_text(self, table, text):
+        return text
+
+    def _get_limit_or_fetch(self, select):
+        if select._fetch_clause is None:
+            return select._limit_clause
+        else:
+            return select._fetch_clause
+
+    def _use_top(self, select):
+        return (select._offset_clause is None) and (
+            select._simple_int_clause(select._limit_clause)
+            or (
+                # limit can use TOP with is by itself. fetch only uses TOP
+                # when it needs to because of PERCENT and/or WITH TIES
+                # TODO: Why?  shouldn't we use TOP always ?
+                select._simple_int_clause(select._fetch_clause)
+                and (
+                    select._fetch_clause_options["percent"]
+                    or select._fetch_clause_options["with_ties"]
+                )
+            )
+        )
+
+    def limit_clause(self, cs, **kwargs):
+        return ""
+
+    def _check_can_use_fetch_limit(self, select):
+        # to use ROW_NUMBER(), an ORDER BY is required.
+        # OFFSET are FETCH are options of the ORDER BY clause
+        if not select._order_by_clause.clauses:
+            raise exc.CompileError(
+                "MSSQL requires an order_by when "
+                "using an OFFSET or a non-simple "
+                "LIMIT clause"
+            )
+
+        if select._fetch_clause_options is not None and (
+            select._fetch_clause_options["percent"]
+            or select._fetch_clause_options["with_ties"]
+        ):
+            raise exc.CompileError(
+                "MSSQL needs TOP to use PERCENT and/or WITH TIES. "
+                "Only simple fetch without offset can be used."
+            )
+
+    def _row_limit_clause(self, select, **kw):
+        """MSSQL 2012 supports OFFSET/FETCH operators
+        Use it instead subquery with row_number
+
+        """
+
+        if self.dialect._supports_offset_fetch and not self._use_top(select):
+            self._check_can_use_fetch_limit(select)
+
+            return self.fetch_clause(
+                select,
+                fetch_clause=self._get_limit_or_fetch(select),
+                require_offset=True,
+                **kw,
+            )
+
+        else:
+            return ""
+
+    def visit_try_cast(self, element, **kw):
+        return "TRY_CAST (%s AS %s)" % (
+            self.process(element.clause, **kw),
+            self.process(element.typeclause, **kw),
+        )
+
+    def translate_select_structure(self, select_stmt, **kwargs):
+        """Look for ``LIMIT`` and OFFSET in a select statement, and if
+        so tries to wrap it in a subquery with ``row_number()`` criterion.
+        MSSQL 2012 and above are excluded
+
+        """
+        select = select_stmt
+
+        if (
+            select._has_row_limiting_clause
+            and not self.dialect._supports_offset_fetch
+            and not self._use_top(select)
+            and not getattr(select, "_mssql_visit", None)
+        ):
+            self._check_can_use_fetch_limit(select)
+
+            _order_by_clauses = [
+                sql_util.unwrap_label_reference(elem)
+                for elem in select._order_by_clause.clauses
+            ]
+
+            limit_clause = self._get_limit_or_fetch(select)
+            offset_clause = select._offset_clause
+
+            select = select._generate()
+            select._mssql_visit = True
+            select = (
+                select.add_columns(
+                    sql.func.ROW_NUMBER()
+                    .over(order_by=_order_by_clauses)
+                    .label("mssql_rn")
+                )
+                .order_by(None)
+                .alias()
+            )
+
+            mssql_rn = sql.column("mssql_rn")
+            limitselect = sql.select(
+                *[c for c in select.c if c.key != "mssql_rn"]
+            )
+            if offset_clause is not None:
+                limitselect = limitselect.where(mssql_rn > offset_clause)
+                if limit_clause is not None:
+                    limitselect = limitselect.where(
+                        mssql_rn <= (limit_clause + offset_clause)
+                    )
+            else:
+                limitselect = limitselect.where(mssql_rn <= (limit_clause))
+            return limitselect
+        else:
+            return select
+
+    @_with_legacy_schema_aliasing
+    def visit_table(self, table, mssql_aliased=False, iscrud=False, **kwargs):
+        if mssql_aliased is table or iscrud:
+            return super().visit_table(table, **kwargs)
+
+        # alias schema-qualified tables
+        alias = self._schema_aliased_table(table)
+        if alias is not None:
+            return self.process(alias, mssql_aliased=table, **kwargs)
+        else:
+            return super().visit_table(table, **kwargs)
+
+    @_with_legacy_schema_aliasing
+    def visit_alias(self, alias, **kw):
+        # translate for schema-qualified table aliases
+        kw["mssql_aliased"] = alias.element
+        return super().visit_alias(alias, **kw)
+
+    @_with_legacy_schema_aliasing
+    def visit_column(self, column, add_to_result_map=None, **kw):
+        if (
+            column.table is not None
+            and (not self.isupdate and not self.isdelete)
+            or self.is_subquery()
+        ):
+            # translate for schema-qualified table aliases
+            t = self._schema_aliased_table(column.table)
+            if t is not None:
+                converted = elements._corresponding_column_or_error(t, column)
+                if add_to_result_map is not None:
+                    add_to_result_map(
+                        column.name,
+                        column.name,
+                        (column, column.name, column.key),
+                        column.type,
+                    )
+
+                return super().visit_column(converted, **kw)
+
+        return super().visit_column(
+            column, add_to_result_map=add_to_result_map, **kw
+        )
+
+    def _schema_aliased_table(self, table):
+        if getattr(table, "schema", None) is not None:
+            if table not in self.tablealiases:
+                self.tablealiases[table] = table.alias()
+            return self.tablealiases[table]
+        else:
+            return None
+
+    def visit_extract(self, extract, **kw):
+        field = self.extract_map.get(extract.field, extract.field)
+        return "DATEPART(%s, %s)" % (field, self.process(extract.expr, **kw))
+
+    def visit_savepoint(self, savepoint_stmt, **kw):
+        return "SAVE TRANSACTION %s" % self.preparer.format_savepoint(
+            savepoint_stmt
+        )
+
+    def visit_rollback_to_savepoint(self, savepoint_stmt, **kw):
+        return "ROLLBACK TRANSACTION %s" % self.preparer.format_savepoint(
+            savepoint_stmt
+        )
+
+    def visit_binary(self, binary, **kwargs):
+        """Move bind parameters to the right-hand side of an operator, where
+        possible.
+
+        """
+        if (
+            isinstance(binary.left, expression.BindParameter)
+            and binary.operator == operator.eq
+            and not isinstance(binary.right, expression.BindParameter)
+        ):
+            return self.process(
+                expression.BinaryExpression(
+                    binary.right, binary.left, binary.operator
+                ),
+                **kwargs,
+            )
+        return super().visit_binary(binary, **kwargs)
+
+    def returning_clause(
+        self, stmt, returning_cols, *, populate_result_map, **kw
+    ):
+        # SQL server returning clause requires that the columns refer to
+        # the virtual table names "inserted" or "deleted".   Here, we make
+        # a simple alias of our table with that name, and then adapt the
+        # columns we have from the list of RETURNING columns to that new name
+        # so that they render as "inserted.<colname>" / "deleted.<colname>".
+
+        if stmt.is_insert or stmt.is_update:
+            target = stmt.table.alias("inserted")
+        elif stmt.is_delete:
+            target = stmt.table.alias("deleted")
+        else:
+            assert False, "expected Insert, Update or Delete statement"
+
+        adapter = sql_util.ClauseAdapter(target)
+
+        # adapter.traverse() takes a column from our target table and returns
+        # the one that is linked to the "inserted" / "deleted" tables.  So  in
+        # order to retrieve these values back from the result  (e.g. like
+        # row[column]), tell the compiler to also add the original unadapted
+        # column to the result map.   Before #4877, these were  (unknowingly)
+        # falling back using string name matching in the result set which
+        # necessarily used an expensive KeyError in order to match.
+
+        columns = [
+            self._label_returning_column(
+                stmt,
+                adapter.traverse(column),
+                populate_result_map,
+                {"result_map_targets": (column,)},
+                fallback_label_name=fallback_label_name,
+                column_is_repeated=repeated,
+                name=name,
+                proxy_name=proxy_name,
+                **kw,
+            )
+            for (
+                name,
+                proxy_name,
+                fallback_label_name,
+                column,
+                repeated,
+            ) in stmt._generate_columns_plus_names(
+                True, cols=expression._select_iterables(returning_cols)
+            )
+        ]
+
+        return "OUTPUT " + ", ".join(columns)
+
+    def get_cte_preamble(self, recursive):
+        # SQL Server finds it too inconvenient to accept
+        # an entirely optional, SQL standard specified,
+        # "RECURSIVE" word with their "WITH",
+        # so here we go
+        return "WITH"
+
+    def label_select_column(self, select, column, asfrom):
+        if isinstance(column, expression.Function):
+            return column.label(None)
+        else:
+            return super().label_select_column(select, column, asfrom)
+
+    def for_update_clause(self, select, **kw):
+        # "FOR UPDATE" is only allowed on "DECLARE CURSOR" which
+        # SQLAlchemy doesn't use
+        return ""
+
+    def order_by_clause(self, select, **kw):
+        # MSSQL only allows ORDER BY in subqueries if there is a LIMIT:
+        # "The ORDER BY clause is invalid in views, inline functions,
+        # derived tables, subqueries, and common table expressions,
+        # unless TOP, OFFSET or FOR XML is also specified."
+        if (
+            self.is_subquery()
+            and not self._use_top(select)
+            and (
+                select._offset is None
+                or not self.dialect._supports_offset_fetch
+            )
+        ):
+            # avoid processing the order by clause if we won't end up
+            # using it, because we don't want all the bind params tacked
+            # onto the positional list if that is what the dbapi requires
+            return ""
+
+        order_by = self.process(select._order_by_clause, **kw)
+
+        if order_by:
+            return " ORDER BY " + order_by
+        else:
+            return ""
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Render the UPDATE..FROM clause specific to MSSQL.
+
+        In MSSQL, if the UPDATE statement involves an alias of the table to
+        be updated, then the table itself must be added to the FROM list as
+        well. Otherwise, it is optional. Here, we add it regardless.
+
+        """
+        return "FROM " + ", ".join(
+            t._compiler_dispatch(self, asfrom=True, fromhints=from_hints, **kw)
+            for t in [from_table] + extra_froms
+        )
+
+    def delete_table_clause(self, delete_stmt, from_table, extra_froms, **kw):
+        """If we have extra froms make sure we render any alias as hint."""
+        ashint = False
+        if extra_froms:
+            ashint = True
+        return from_table._compiler_dispatch(
+            self, asfrom=True, iscrud=True, ashint=ashint, **kw
+        )
+
+    def delete_extra_from_clause(
+        self, delete_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Render the DELETE .. FROM clause specific to MSSQL.
+
+        Yes, it has the FROM keyword twice.
+
+        """
+        return "FROM " + ", ".join(
+            t._compiler_dispatch(self, asfrom=True, fromhints=from_hints, **kw)
+            for t in [from_table] + extra_froms
+        )
+
+    def visit_empty_set_expr(self, type_, **kw):
+        return "SELECT 1 WHERE 1!=1"
+
+    def visit_is_distinct_from_binary(self, binary, operator, **kw):
+        return "NOT EXISTS (SELECT %s INTERSECT SELECT %s)" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_is_not_distinct_from_binary(self, binary, operator, **kw):
+        return "EXISTS (SELECT %s INTERSECT SELECT %s)" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def _render_json_extract_from_binary(self, binary, operator, **kw):
+        # note we are intentionally calling upon the process() calls in the
+        # order in which they appear in the SQL String as this is used
+        # by positional parameter rendering
+
+        if binary.type._type_affinity is sqltypes.JSON:
+            return "JSON_QUERY(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+
+        # as with other dialects, start with an explicit test for NULL
+        case_expression = "CASE JSON_VALUE(%s, %s) WHEN NULL THEN NULL" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+        if binary.type._type_affinity is sqltypes.Integer:
+            type_expression = "ELSE CAST(JSON_VALUE(%s, %s) AS INTEGER)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+        elif binary.type._type_affinity is sqltypes.Numeric:
+            type_expression = "ELSE CAST(JSON_VALUE(%s, %s) AS %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+                (
+                    "FLOAT"
+                    if isinstance(binary.type, sqltypes.Float)
+                    else "NUMERIC(%s, %s)"
+                    % (binary.type.precision, binary.type.scale)
+                ),
+            )
+        elif binary.type._type_affinity is sqltypes.Boolean:
+            # the NULL handling is particularly weird with boolean, so
+            # explicitly return numeric (BIT) constants
+            type_expression = (
+                "WHEN 'true' THEN 1 WHEN 'false' THEN 0 ELSE NULL"
+            )
+        elif binary.type._type_affinity is sqltypes.String:
+            # TODO: does this comment (from mysql) apply to here, too?
+            #       this fails with a JSON value that's a four byte unicode
+            #       string.  SQLite has the same problem at the moment
+            type_expression = "ELSE JSON_VALUE(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+        else:
+            # other affinity....this is not expected right now
+            type_expression = "ELSE JSON_QUERY(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+
+        return case_expression + " " + type_expression + " END"
+
+    def visit_json_getitem_op_binary(self, binary, operator, **kw):
+        return self._render_json_extract_from_binary(binary, operator, **kw)
+
+    def visit_json_path_getitem_op_binary(self, binary, operator, **kw):
+        return self._render_json_extract_from_binary(binary, operator, **kw)
+
+    def visit_sequence(self, seq, **kw):
+        return "NEXT VALUE FOR %s" % self.preparer.format_sequence(seq)
+
+
+class MSSQLStrictCompiler(MSSQLCompiler):
+    """A subclass of MSSQLCompiler which disables the usage of bind
+    parameters where not allowed natively by MS-SQL.
+
+    A dialect may use this compiler on a platform where native
+    binds are used.
+
+    """
+
+    ansi_bind_rules = True
+
+    def visit_in_op_binary(self, binary, operator, **kw):
+        kw["literal_execute"] = True
+        return "%s IN %s" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_not_in_op_binary(self, binary, operator, **kw):
+        kw["literal_execute"] = True
+        return "%s NOT IN %s" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def render_literal_value(self, value, type_):
+        """
+        For date and datetime values, convert to a string
+        format acceptable to MSSQL. That seems to be the
+        so-called ODBC canonical date format which looks
+        like this:
+
+            yyyy-mm-dd hh:mi:ss.mmm(24h)
+
+        For other data types, call the base class implementation.
+        """
+        # datetime and date are both subclasses of datetime.date
+        if issubclass(type(value), datetime.date):
+            # SQL Server wants single quotes around the date string.
+            return "'" + str(value) + "'"
+        else:
+            return super().render_literal_value(value, type_)
+
+
+class MSDDLCompiler(compiler.DDLCompiler):
+    def get_column_specification(self, column, **kwargs):
+        colspec = self.preparer.format_column(column)
+
+        # type is not accepted in a computed column
+        if column.computed is not None:
+            colspec += " " + self.process(column.computed)
+        else:
+            colspec += " " + self.dialect.type_compiler_instance.process(
+                column.type, type_expression=column
+            )
+
+        if column.nullable is not None:
+            if (
+                not column.nullable
+                or column.primary_key
+                or isinstance(column.default, sa_schema.Sequence)
+                or column.autoincrement is True
+                or column.identity
+            ):
+                colspec += " NOT NULL"
+            elif column.computed is None:
+                # don't specify "NULL" for computed columns
+                colspec += " NULL"
+
+        if column.table is None:
+            raise exc.CompileError(
+                "mssql requires Table-bound columns "
+                "in order to generate DDL"
+            )
+
+        d_opt = column.dialect_options["mssql"]
+        start = d_opt["identity_start"]
+        increment = d_opt["identity_increment"]
+        if start is not None or increment is not None:
+            if column.identity:
+                raise exc.CompileError(
+                    "Cannot specify options 'mssql_identity_start' and/or "
+                    "'mssql_identity_increment' while also using the "
+                    "'Identity' construct."
+                )
+            util.warn_deprecated(
+                "The dialect options 'mssql_identity_start' and "
+                "'mssql_identity_increment' are deprecated. "
+                "Use the 'Identity' object instead.",
+                "1.4",
+            )
+
+        if column.identity:
+            colspec += self.process(column.identity, **kwargs)
+        elif (
+            column is column.table._autoincrement_column
+            or column.autoincrement is True
+        ) and (
+            not isinstance(column.default, Sequence) or column.default.optional
+        ):
+            colspec += self.process(Identity(start=start, increment=increment))
+        else:
+            default = self.get_column_default_string(column)
+            if default is not None:
+                colspec += " DEFAULT " + default
+
+        return colspec
+
+    def visit_create_index(self, create, include_schema=False, **kw):
+        index = create.element
+        self._verify_index_table(index)
+        preparer = self.preparer
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+
+        # handle clustering option
+        clustered = index.dialect_options["mssql"]["clustered"]
+        if clustered is not None:
+            if clustered:
+                text += "CLUSTERED "
+            else:
+                text += "NONCLUSTERED "
+
+        # handle columnstore option (has no negative value)
+        columnstore = index.dialect_options["mssql"]["columnstore"]
+        if columnstore:
+            text += "COLUMNSTORE "
+
+        text += "INDEX %s ON %s" % (
+            self._prepared_index_name(index, include_schema=include_schema),
+            preparer.format_table(index.table),
+        )
+
+        # in some case mssql allows indexes with no columns defined
+        if len(index.expressions) > 0:
+            text += " (%s)" % ", ".join(
+                self.sql_compiler.process(
+                    expr, include_table=False, literal_binds=True
+                )
+                for expr in index.expressions
+            )
+
+        # handle other included columns
+        if index.dialect_options["mssql"]["include"]:
+            inclusions = [
+                index.table.c[col] if isinstance(col, str) else col
+                for col in index.dialect_options["mssql"]["include"]
+            ]
+
+            text += " INCLUDE (%s)" % ", ".join(
+                [preparer.quote(c.name) for c in inclusions]
+            )
+
+        whereclause = index.dialect_options["mssql"]["where"]
+
+        if whereclause is not None:
+            whereclause = coercions.expect(
+                roles.DDLExpressionRole, whereclause
+            )
+
+            where_compiled = self.sql_compiler.process(
+                whereclause, include_table=False, literal_binds=True
+            )
+            text += " WHERE " + where_compiled
+
+        return text
+
+    def visit_drop_index(self, drop, **kw):
+        return "\nDROP INDEX %s ON %s" % (
+            self._prepared_index_name(drop.element, include_schema=False),
+            self.preparer.format_table(drop.element.table),
+        )
+
+    def visit_primary_key_constraint(self, constraint, **kw):
+        if len(constraint) == 0:
+            return ""
+        text = ""
+        if constraint.name is not None:
+            text += "CONSTRAINT %s " % self.preparer.format_constraint(
+                constraint
+            )
+        text += "PRIMARY KEY "
+
+        clustered = constraint.dialect_options["mssql"]["clustered"]
+        if clustered is not None:
+            if clustered:
+                text += "CLUSTERED "
+            else:
+                text += "NONCLUSTERED "
+
+        text += "(%s)" % ", ".join(
+            self.preparer.quote(c.name) for c in constraint
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def visit_unique_constraint(self, constraint, **kw):
+        if len(constraint) == 0:
+            return ""
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        text += "UNIQUE %s" % self.define_unique_constraint_distinct(
+            constraint, **kw
+        )
+        clustered = constraint.dialect_options["mssql"]["clustered"]
+        if clustered is not None:
+            if clustered:
+                text += "CLUSTERED "
+            else:
+                text += "NONCLUSTERED "
+
+        text += "(%s)" % ", ".join(
+            self.preparer.quote(c.name) for c in constraint
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def visit_computed_column(self, generated, **kw):
+        text = "AS (%s)" % self.sql_compiler.process(
+            generated.sqltext, include_table=False, literal_binds=True
+        )
+        # explicitly check for True|False since None means server default
+        if generated.persisted is True:
+            text += " PERSISTED"
+        return text
+
+    def visit_set_table_comment(self, create, **kw):
+        schema = self.preparer.schema_for_object(create.element)
+        schema_name = schema if schema else self.dialect.default_schema_name
+        return (
+            "execute sp_addextendedproperty 'MS_Description', "
+            "{}, 'schema', {}, 'table', {}".format(
+                self.sql_compiler.render_literal_value(
+                    create.element.comment, sqltypes.NVARCHAR()
+                ),
+                self.preparer.quote_schema(schema_name),
+                self.preparer.format_table(create.element, use_schema=False),
+            )
+        )
+
+    def visit_drop_table_comment(self, drop, **kw):
+        schema = self.preparer.schema_for_object(drop.element)
+        schema_name = schema if schema else self.dialect.default_schema_name
+        return (
+            "execute sp_dropextendedproperty 'MS_Description', 'schema', "
+            "{}, 'table', {}".format(
+                self.preparer.quote_schema(schema_name),
+                self.preparer.format_table(drop.element, use_schema=False),
+            )
+        )
+
+    def visit_set_column_comment(self, create, **kw):
+        schema = self.preparer.schema_for_object(create.element.table)
+        schema_name = schema if schema else self.dialect.default_schema_name
+        return (
+            "execute sp_addextendedproperty 'MS_Description', "
+            "{}, 'schema', {}, 'table', {}, 'column', {}".format(
+                self.sql_compiler.render_literal_value(
+                    create.element.comment, sqltypes.NVARCHAR()
+                ),
+                self.preparer.quote_schema(schema_name),
+                self.preparer.format_table(
+                    create.element.table, use_schema=False
+                ),
+                self.preparer.format_column(create.element),
+            )
+        )
+
+    def visit_drop_column_comment(self, drop, **kw):
+        schema = self.preparer.schema_for_object(drop.element.table)
+        schema_name = schema if schema else self.dialect.default_schema_name
+        return (
+            "execute sp_dropextendedproperty 'MS_Description', 'schema', "
+            "{}, 'table', {}, 'column', {}".format(
+                self.preparer.quote_schema(schema_name),
+                self.preparer.format_table(
+                    drop.element.table, use_schema=False
+                ),
+                self.preparer.format_column(drop.element),
+            )
+        )
+
+    def visit_create_sequence(self, create, **kw):
+        prefix = None
+        if create.element.data_type is not None:
+            data_type = create.element.data_type
+            prefix = " AS %s" % self.type_compiler.process(data_type)
+        return super().visit_create_sequence(create, prefix=prefix, **kw)
+
+    def visit_identity_column(self, identity, **kw):
+        text = " IDENTITY"
+        if identity.start is not None or identity.increment is not None:
+            start = 1 if identity.start is None else identity.start
+            increment = 1 if identity.increment is None else identity.increment
+            text += "(%s,%s)" % (start, increment)
+        return text
+
+
+class MSIdentifierPreparer(compiler.IdentifierPreparer):
+    reserved_words = RESERVED_WORDS
+
+    def __init__(self, dialect):
+        super().__init__(
+            dialect,
+            initial_quote="[",
+            final_quote="]",
+            quote_case_sensitive_collations=False,
+        )
+
+    def _escape_identifier(self, value):
+        return value.replace("]", "]]")
+
+    def _unescape_identifier(self, value):
+        return value.replace("]]", "]")
+
+    def quote_schema(self, schema, force=None):
+        """Prepare a quoted table and schema name."""
+
+        # need to re-implement the deprecation warning entirely
+        if force is not None:
+            # not using the util.deprecated_params() decorator in this
+            # case because of the additional function call overhead on this
+            # very performance-critical spot.
+            util.warn_deprecated(
+                "The IdentifierPreparer.quote_schema.force parameter is "
+                "deprecated and will be removed in a future release.  This "
+                "flag has no effect on the behavior of the "
+                "IdentifierPreparer.quote method; please refer to "
+                "quoted_name().",
+                version="1.3",
+            )
+
+        dbname, owner = _schema_elements(schema)
+        if dbname:
+            result = "%s.%s" % (self.quote(dbname), self.quote(owner))
+        elif owner:
+            result = self.quote(owner)
+        else:
+            result = ""
+        return result
+
+
+def _db_plus_owner_listing(fn):
+    def wrap(dialect, connection, schema=None, **kw):
+        dbname, owner = _owner_plus_db(dialect, schema)
+        return _switch_db(
+            dbname,
+            connection,
+            fn,
+            dialect,
+            connection,
+            dbname,
+            owner,
+            schema,
+            **kw,
+        )
+
+    return update_wrapper(wrap, fn)
+
+
+def _db_plus_owner(fn):
+    def wrap(dialect, connection, tablename, schema=None, **kw):
+        dbname, owner = _owner_plus_db(dialect, schema)
+        return _switch_db(
+            dbname,
+            connection,
+            fn,
+            dialect,
+            connection,
+            tablename,
+            dbname,
+            owner,
+            schema,
+            **kw,
+        )
+
+    return update_wrapper(wrap, fn)
+
+
+def _switch_db(dbname, connection, fn, *arg, **kw):
+    if dbname:
+        current_db = connection.exec_driver_sql("select db_name()").scalar()
+        if current_db != dbname:
+            connection.exec_driver_sql(
+                "use %s" % connection.dialect.identifier_preparer.quote(dbname)
+            )
+    try:
+        return fn(*arg, **kw)
+    finally:
+        if dbname and current_db != dbname:
+            connection.exec_driver_sql(
+                "use %s"
+                % connection.dialect.identifier_preparer.quote(current_db)
+            )
+
+
+def _owner_plus_db(dialect, schema):
+    if not schema:
+        return None, dialect.default_schema_name
+    else:
+        return _schema_elements(schema)
+
+
+_memoized_schema = util.LRUCache()
+
+
+def _schema_elements(schema):
+    if isinstance(schema, quoted_name) and schema.quote:
+        return None, schema
+
+    if schema in _memoized_schema:
+        return _memoized_schema[schema]
+
+    # tests for this function are in:
+    # test/dialect/mssql/test_reflection.py ->
+    #           OwnerPlusDBTest.test_owner_database_pairs
+    # test/dialect/mssql/test_compiler.py -> test_force_schema_*
+    # test/dialect/mssql/test_compiler.py -> test_schema_many_tokens_*
+    #
+
+    if schema.startswith("__[SCHEMA_"):
+        return None, schema
+
+    push = []
+    symbol = ""
+    bracket = False
+    has_brackets = False
+    for token in re.split(r"(\[|\]|\.)", schema):
+        if not token:
+            continue
+        if token == "[":
+            bracket = True
+            has_brackets = True
+        elif token == "]":
+            bracket = False
+        elif not bracket and token == ".":
+            if has_brackets:
+                push.append("[%s]" % symbol)
+            else:
+                push.append(symbol)
+            symbol = ""
+            has_brackets = False
+        else:
+            symbol += token
+    if symbol:
+        push.append(symbol)
+    if len(push) > 1:
+        dbname, owner = ".".join(push[0:-1]), push[-1]
+
+        # test for internal brackets
+        if re.match(r".*\].*\[.*", dbname[1:-1]):
+            dbname = quoted_name(dbname, quote=False)
+        else:
+            dbname = dbname.lstrip("[").rstrip("]")
+
+    elif len(push):
+        dbname, owner = None, push[0]
+    else:
+        dbname, owner = None, None
+
+    _memoized_schema[schema] = dbname, owner
+    return dbname, owner
+
+
+class MSDialect(default.DefaultDialect):
+    # will assume it's at least mssql2005
+    name = "mssql"
+    supports_statement_cache = True
+    supports_default_values = True
+    supports_empty_insert = False
+    favor_returning_over_lastrowid = True
+
+    returns_native_bytes = True
+
+    supports_comments = True
+    supports_default_metavalue = False
+    """dialect supports INSERT... VALUES (DEFAULT) syntax -
+    SQL Server **does** support this, but **not** for the IDENTITY column,
+    so we can't turn this on.
+
+    """
+
+    # supports_native_uuid is partial here, so we implement our
+    # own impl type
+
+    execution_ctx_cls = MSExecutionContext
+    use_scope_identity = True
+    max_identifier_length = 128
+    schema_name = "dbo"
+
+    insert_returning = True
+    update_returning = True
+    delete_returning = True
+    update_returning_multifrom = True
+    delete_returning_multifrom = True
+
+    colspecs = {
+        sqltypes.DateTime: _MSDateTime,
+        sqltypes.Date: _MSDate,
+        sqltypes.JSON: JSON,
+        sqltypes.JSON.JSONIndexType: JSONIndexType,
+        sqltypes.JSON.JSONPathType: JSONPathType,
+        sqltypes.Time: _BASETIMEIMPL,
+        sqltypes.Unicode: _MSUnicode,
+        sqltypes.UnicodeText: _MSUnicodeText,
+        DATETIMEOFFSET: DATETIMEOFFSET,
+        DATETIME2: DATETIME2,
+        SMALLDATETIME: SMALLDATETIME,
+        DATETIME: DATETIME,
+        sqltypes.Uuid: MSUUid,
+    }
+
+    engine_config_types = default.DefaultDialect.engine_config_types.union(
+        {"legacy_schema_aliasing": util.asbool}
+    )
+
+    ischema_names = ischema_names
+
+    supports_sequences = True
+    sequences_optional = True
+    # This is actually used for autoincrement, where itentity is used that
+    # starts with 1.
+    # for sequences T-SQL's actual default is -9223372036854775808
+    default_sequence_base = 1
+
+    supports_native_boolean = False
+    non_native_boolean_check_constraint = False
+    supports_unicode_binds = True
+    postfetch_lastrowid = True
+
+    # may be changed at server inspection time for older SQL server versions
+    supports_multivalues_insert = True
+
+    use_insertmanyvalues = True
+
+    # note pyodbc will set this to False if fast_executemany is set,
+    # as of SQLAlchemy 2.0.9
+    use_insertmanyvalues_wo_returning = True
+
+    insertmanyvalues_implicit_sentinel = (
+        InsertmanyvaluesSentinelOpts.AUTOINCREMENT
+        | InsertmanyvaluesSentinelOpts.IDENTITY
+        | InsertmanyvaluesSentinelOpts.USE_INSERT_FROM_SELECT
+    )
+
+    # "The incoming request has too many parameters. The server supports a "
+    # "maximum of 2100 parameters."
+    # in fact you can have 2099 parameters.
+    insertmanyvalues_max_parameters = 2099
+
+    _supports_offset_fetch = False
+    _supports_nvarchar_max = False
+
+    legacy_schema_aliasing = False
+
+    server_version_info = ()
+
+    statement_compiler = MSSQLCompiler
+    ddl_compiler = MSDDLCompiler
+    type_compiler_cls = MSTypeCompiler
+    preparer = MSIdentifierPreparer
+
+    construct_arguments = [
+        (sa_schema.PrimaryKeyConstraint, {"clustered": None}),
+        (sa_schema.UniqueConstraint, {"clustered": None}),
+        (
+            sa_schema.Index,
+            {
+                "clustered": None,
+                "include": None,
+                "where": None,
+                "columnstore": None,
+            },
+        ),
+        (
+            sa_schema.Column,
+            {"identity_start": None, "identity_increment": None},
+        ),
+    ]
+
+    def __init__(
+        self,
+        query_timeout=None,
+        use_scope_identity=True,
+        schema_name="dbo",
+        deprecate_large_types=None,
+        supports_comments=None,
+        json_serializer=None,
+        json_deserializer=None,
+        legacy_schema_aliasing=None,
+        ignore_no_transaction_on_rollback=False,
+        **opts,
+    ):
+        self.query_timeout = int(query_timeout or 0)
+        self.schema_name = schema_name
+
+        self.use_scope_identity = use_scope_identity
+        self.deprecate_large_types = deprecate_large_types
+        self.ignore_no_transaction_on_rollback = (
+            ignore_no_transaction_on_rollback
+        )
+        self._user_defined_supports_comments = uds = supports_comments
+        if uds is not None:
+            self.supports_comments = uds
+
+        if legacy_schema_aliasing is not None:
+            util.warn_deprecated(
+                "The legacy_schema_aliasing parameter is "
+                "deprecated and will be removed in a future release.",
+                "1.4",
+            )
+            self.legacy_schema_aliasing = legacy_schema_aliasing
+
+        super().__init__(**opts)
+
+        self._json_serializer = json_serializer
+        self._json_deserializer = json_deserializer
+
+    def do_savepoint(self, connection, name):
+        # give the DBAPI a push
+        connection.exec_driver_sql("IF @@TRANCOUNT = 0 BEGIN TRANSACTION")
+        super().do_savepoint(connection, name)
+
+    def do_release_savepoint(self, connection, name):
+        # SQL Server does not support RELEASE SAVEPOINT
+        pass
+
+    def do_rollback(self, dbapi_connection):
+        try:
+            super().do_rollback(dbapi_connection)
+        except self.dbapi.ProgrammingError as e:
+            if self.ignore_no_transaction_on_rollback and re.match(
+                r".*\b111214\b", str(e)
+            ):
+                util.warn(
+                    "ProgrammingError 111214 "
+                    "'No corresponding transaction found.' "
+                    "has been suppressed via "
+                    "ignore_no_transaction_on_rollback=True"
+                )
+            else:
+                raise
+
+    _isolation_lookup = {
+        "SERIALIZABLE",
+        "READ UNCOMMITTED",
+        "READ COMMITTED",
+        "REPEATABLE READ",
+        "SNAPSHOT",
+    }
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return list(self._isolation_lookup)
+
+    def set_isolation_level(self, dbapi_connection, level):
+        cursor = dbapi_connection.cursor()
+        cursor.execute(f"SET TRANSACTION ISOLATION LEVEL {level}")
+        cursor.close()
+        if level == "SNAPSHOT":
+            dbapi_connection.commit()
+
+    def get_isolation_level(self, dbapi_connection):
+        cursor = dbapi_connection.cursor()
+        view_name = "sys.system_views"
+        try:
+            cursor.execute(
+                (
+                    "SELECT name FROM {} WHERE name IN "
+                    "('dm_exec_sessions', 'dm_pdw_nodes_exec_sessions')"
+                ).format(view_name)
+            )
+            row = cursor.fetchone()
+            if not row:
+                raise NotImplementedError(
+                    "Can't fetch isolation level on this particular "
+                    "SQL Server version."
+                )
+
+            view_name = f"sys.{row[0]}"
+
+            cursor.execute(
+                """
+                    SELECT CASE transaction_isolation_level
+                    WHEN 0 THEN NULL
+                    WHEN 1 THEN 'READ UNCOMMITTED'
+                    WHEN 2 THEN 'READ COMMITTED'
+                    WHEN 3 THEN 'REPEATABLE READ'
+                    WHEN 4 THEN 'SERIALIZABLE'
+                    WHEN 5 THEN 'SNAPSHOT' END
+                    AS TRANSACTION_ISOLATION_LEVEL
+                    FROM {}
+                    where session_id = @@SPID
+                """.format(
+                    view_name
+                )
+            )
+        except self.dbapi.Error as err:
+            raise NotImplementedError(
+                "Can't fetch isolation level;  encountered error {} when "
+                'attempting to query the "{}" view.'.format(err, view_name)
+            ) from err
+        else:
+            row = cursor.fetchone()
+            return row[0].upper()
+        finally:
+            cursor.close()
+
+    def initialize(self, connection):
+        super().initialize(connection)
+        self._setup_version_attributes()
+        self._setup_supports_nvarchar_max(connection)
+        self._setup_supports_comments(connection)
+
+    def _setup_version_attributes(self):
+        if self.server_version_info[0] not in list(range(8, 17)):
+            util.warn(
+                "Unrecognized server version info '%s'.  Some SQL Server "
+                "features may not function properly."
+                % ".".join(str(x) for x in self.server_version_info)
+            )
+
+        if self.server_version_info >= MS_2008_VERSION:
+            self.supports_multivalues_insert = True
+        else:
+            self.supports_multivalues_insert = False
+
+        if self.deprecate_large_types is None:
+            self.deprecate_large_types = (
+                self.server_version_info >= MS_2012_VERSION
+            )
+
+        self._supports_offset_fetch = (
+            self.server_version_info and self.server_version_info[0] >= 11
+        )
+
+    def _setup_supports_nvarchar_max(self, connection):
+        try:
+            connection.scalar(
+                sql.text("SELECT CAST('test max support' AS NVARCHAR(max))")
+            )
+        except exc.DBAPIError:
+            self._supports_nvarchar_max = False
+        else:
+            self._supports_nvarchar_max = True
+
+    def _setup_supports_comments(self, connection):
+        if self._user_defined_supports_comments is not None:
+            return
+
+        try:
+            connection.scalar(
+                sql.text(
+                    "SELECT 1 FROM fn_listextendedproperty"
+                    "(default, default, default, default, "
+                    "default, default, default)"
+                )
+            )
+        except exc.DBAPIError:
+            self.supports_comments = False
+        else:
+            self.supports_comments = True
+
+    def _get_default_schema_name(self, connection):
+        query = sql.text("SELECT schema_name()")
+        default_schema_name = connection.scalar(query)
+        if default_schema_name is not None:
+            # guard against the case where the default_schema_name is being
+            # fed back into a table reflection function.
+            return quoted_name(default_schema_name, quote=True)
+        else:
+            return self.schema_name
+
+    @_db_plus_owner
+    def has_table(self, connection, tablename, dbname, owner, schema, **kw):
+        self._ensure_has_table_connection(connection)
+
+        return self._internal_has_table(connection, tablename, owner, **kw)
+
+    @reflection.cache
+    @_db_plus_owner
+    def has_sequence(
+        self, connection, sequencename, dbname, owner, schema, **kw
+    ):
+        sequences = ischema.sequences
+
+        s = sql.select(sequences.c.sequence_name).where(
+            sequences.c.sequence_name == sequencename
+        )
+
+        if owner:
+            s = s.where(sequences.c.sequence_schema == owner)
+
+        c = connection.execute(s)
+
+        return c.first() is not None
+
+    @reflection.cache
+    @_db_plus_owner_listing
+    def get_sequence_names(self, connection, dbname, owner, schema, **kw):
+        sequences = ischema.sequences
+
+        s = sql.select(sequences.c.sequence_name)
+        if owner:
+            s = s.where(sequences.c.sequence_schema == owner)
+
+        c = connection.execute(s)
+
+        return [row[0] for row in c]
+
+    @reflection.cache
+    def get_schema_names(self, connection, **kw):
+        s = sql.select(ischema.schemata.c.schema_name).order_by(
+            ischema.schemata.c.schema_name
+        )
+        schema_names = [r[0] for r in connection.execute(s)]
+        return schema_names
+
+    @reflection.cache
+    @_db_plus_owner_listing
+    def get_table_names(self, connection, dbname, owner, schema, **kw):
+        tables = ischema.tables
+        s = (
+            sql.select(tables.c.table_name)
+            .where(
+                sql.and_(
+                    tables.c.table_schema == owner,
+                    tables.c.table_type == "BASE TABLE",
+                )
+            )
+            .order_by(tables.c.table_name)
+        )
+        table_names = [r[0] for r in connection.execute(s)]
+        return table_names
+
+    @reflection.cache
+    @_db_plus_owner_listing
+    def get_view_names(self, connection, dbname, owner, schema, **kw):
+        tables = ischema.tables
+        s = (
+            sql.select(tables.c.table_name)
+            .where(
+                sql.and_(
+                    tables.c.table_schema == owner,
+                    tables.c.table_type == "VIEW",
+                )
+            )
+            .order_by(tables.c.table_name)
+        )
+        view_names = [r[0] for r in connection.execute(s)]
+        return view_names
+
+    @reflection.cache
+    def _internal_has_table(self, connection, tablename, owner, **kw):
+        if tablename.startswith("#"):  # temporary table
+            # mssql does not support temporary views
+            # SQL Error [4103] [S0001]: "#v": Temporary views are not allowed
+            return bool(
+                connection.scalar(
+                    # U filters on user tables only.
+                    text("SELECT object_id(:table_name, 'U')"),
+                    {"table_name": f"tempdb.dbo.[{tablename}]"},
+                )
+            )
+        else:
+            tables = ischema.tables
+
+            s = sql.select(tables.c.table_name).where(
+                sql.and_(
+                    sql.or_(
+                        tables.c.table_type == "BASE TABLE",
+                        tables.c.table_type == "VIEW",
+                    ),
+                    tables.c.table_name == tablename,
+                )
+            )
+
+            if owner:
+                s = s.where(tables.c.table_schema == owner)
+
+            c = connection.execute(s)
+
+            return c.first() is not None
+
+    def _default_or_error(self, connection, tablename, owner, method, **kw):
+        # TODO: try to avoid having to run a separate query here
+        if self._internal_has_table(connection, tablename, owner, **kw):
+            return method()
+        else:
+            raise exc.NoSuchTableError(f"{owner}.{tablename}")
+
+    @reflection.cache
+    @_db_plus_owner
+    def get_indexes(self, connection, tablename, dbname, owner, schema, **kw):
+        filter_definition = (
+            "ind.filter_definition"
+            if self.server_version_info >= MS_2008_VERSION
+            else "NULL as filter_definition"
+        )
+        rp = connection.execution_options(future_result=True).execute(
+            sql.text(
+                f"""
+select
+    ind.index_id,
+    ind.is_unique,
+    ind.name,
+    ind.type,
+    {filter_definition}
+from
+    sys.indexes as ind
+join sys.tables as tab on
+    ind.object_id = tab.object_id
+join sys.schemas as sch on
+    sch.schema_id = tab.schema_id
+where
+    tab.name = :tabname
+    and sch.name = :schname
+    and ind.is_primary_key = 0
+    and ind.type != 0
+order by
+    ind.name
+                """
+            )
+            .bindparams(
+                sql.bindparam("tabname", tablename, ischema.CoerceUnicode()),
+                sql.bindparam("schname", owner, ischema.CoerceUnicode()),
+            )
+            .columns(name=sqltypes.Unicode())
+        )
+        indexes = {}
+        for row in rp.mappings():
+            indexes[row["index_id"]] = current = {
+                "name": row["name"],
+                "unique": row["is_unique"] == 1,
+                "column_names": [],
+                "include_columns": [],
+                "dialect_options": {},
+            }
+
+            do = current["dialect_options"]
+            index_type = row["type"]
+            if index_type in {1, 2}:
+                do["mssql_clustered"] = index_type == 1
+            if index_type in {5, 6}:
+                do["mssql_clustered"] = index_type == 5
+                do["mssql_columnstore"] = True
+            if row["filter_definition"] is not None:
+                do["mssql_where"] = row["filter_definition"]
+
+        rp = connection.execution_options(future_result=True).execute(
+            sql.text(
+                """
+select
+    ind_col.index_id,
+    col.name,
+    ind_col.is_included_column
+from
+    sys.columns as col
+join sys.tables as tab on
+    tab.object_id = col.object_id
+join sys.index_columns as ind_col on
+    ind_col.column_id = col.column_id
+    and ind_col.object_id = tab.object_id
+join sys.schemas as sch on
+    sch.schema_id = tab.schema_id
+where
+    tab.name = :tabname
+    and sch.name = :schname
+            """
+            )
+            .bindparams(
+                sql.bindparam("tabname", tablename, ischema.CoerceUnicode()),
+                sql.bindparam("schname", owner, ischema.CoerceUnicode()),
+            )
+            .columns(name=sqltypes.Unicode())
+        )
+        for row in rp.mappings():
+            if row["index_id"] not in indexes:
+                continue
+            index_def = indexes[row["index_id"]]
+            is_colstore = index_def["dialect_options"].get("mssql_columnstore")
+            is_clustered = index_def["dialect_options"].get("mssql_clustered")
+            if not (is_colstore and is_clustered):
+                # a clustered columnstore index includes all columns but does
+                # not want them in the index definition
+                if row["is_included_column"] and not is_colstore:
+                    # a noncludsted columnstore index reports that includes
+                    # columns but requires that are listed as normal columns
+                    index_def["include_columns"].append(row["name"])
+                else:
+                    index_def["column_names"].append(row["name"])
+        for index_info in indexes.values():
+            # NOTE: "root level" include_columns is legacy, now part of
+            #       dialect_options (issue #7382)
+            index_info["dialect_options"]["mssql_include"] = index_info[
+                "include_columns"
+            ]
+
+        if indexes:
+            return list(indexes.values())
+        else:
+            return self._default_or_error(
+                connection, tablename, owner, ReflectionDefaults.indexes, **kw
+            )
+
+    @reflection.cache
+    @_db_plus_owner
+    def get_view_definition(
+        self, connection, viewname, dbname, owner, schema, **kw
+    ):
+        view_def = connection.execute(
+            sql.text(
+                "select mod.definition "
+                "from sys.sql_modules as mod "
+                "join sys.views as views on mod.object_id = views.object_id "
+                "join sys.schemas as sch on views.schema_id = sch.schema_id "
+                "where views.name=:viewname and sch.name=:schname"
+            ).bindparams(
+                sql.bindparam("viewname", viewname, ischema.CoerceUnicode()),
+                sql.bindparam("schname", owner, ischema.CoerceUnicode()),
+            )
+        ).scalar()
+        if view_def:
+            return view_def
+        else:
+            raise exc.NoSuchTableError(f"{owner}.{viewname}")
+
+    @reflection.cache
+    def get_table_comment(self, connection, table_name, schema=None, **kw):
+        if not self.supports_comments:
+            raise NotImplementedError(
+                "Can't get table comments on current SQL Server version in use"
+            )
+
+        schema_name = schema if schema else self.default_schema_name
+        COMMENT_SQL = """
+            SELECT cast(com.value as nvarchar(max))
+            FROM fn_listextendedproperty('MS_Description',
+                'schema', :schema, 'table', :table, NULL, NULL
+            ) as com;
+        """
+
+        comment = connection.execute(
+            sql.text(COMMENT_SQL).bindparams(
+                sql.bindparam("schema", schema_name, ischema.CoerceUnicode()),
+                sql.bindparam("table", table_name, ischema.CoerceUnicode()),
+            )
+        ).scalar()
+        if comment:
+            return {"text": comment}
+        else:
+            return self._default_or_error(
+                connection,
+                table_name,
+                None,
+                ReflectionDefaults.table_comment,
+                **kw,
+            )
+
+    def _temp_table_name_like_pattern(self, tablename):
+        # LIKE uses '%' to match zero or more characters and '_' to match any
+        # single character. We want to match literal underscores, so T-SQL
+        # requires that we enclose them in square brackets.
+        return tablename + (
+            ("[_][_][_]%") if not tablename.startswith("##") else ""
+        )
+
+    def _get_internal_temp_table_name(self, connection, tablename):
+        # it's likely that schema is always "dbo", but since we can
+        # get it here, let's get it.
+        # see https://stackoverflow.com/questions/8311959/
+        # specifying-schema-for-temporary-tables
+
+        try:
+            return connection.execute(
+                sql.text(
+                    "select table_schema, table_name "
+                    "from tempdb.information_schema.tables "
+                    "where table_name like :p1"
+                ),
+                {"p1": self._temp_table_name_like_pattern(tablename)},
+            ).one()
+        except exc.MultipleResultsFound as me:
+            raise exc.UnreflectableTableError(
+                "Found more than one temporary table named '%s' in tempdb "
+                "at this time. Cannot reliably resolve that name to its "
+                "internal table name." % tablename
+            ) from me
+        except exc.NoResultFound as ne:
+            raise exc.NoSuchTableError(
+                "Unable to find a temporary table named '%s' in tempdb."
+                % tablename
+            ) from ne
+
+    @reflection.cache
+    @_db_plus_owner
+    def get_columns(self, connection, tablename, dbname, owner, schema, **kw):
+        is_temp_table = tablename.startswith("#")
+        if is_temp_table:
+            owner, tablename = self._get_internal_temp_table_name(
+                connection, tablename
+            )
+
+            columns = ischema.mssql_temp_table_columns
+        else:
+            columns = ischema.columns
+
+        computed_cols = ischema.computed_columns
+        identity_cols = ischema.identity_columns
+        if owner:
+            whereclause = sql.and_(
+                columns.c.table_name == tablename,
+                columns.c.table_schema == owner,
+            )
+            full_name = columns.c.table_schema + "." + columns.c.table_name
+        else:
+            whereclause = columns.c.table_name == tablename
+            full_name = columns.c.table_name
+
+        if self._supports_nvarchar_max:
+            computed_definition = computed_cols.c.definition
+        else:
+            # tds_version 4.2 does not support NVARCHAR(MAX)
+            computed_definition = sql.cast(
+                computed_cols.c.definition, NVARCHAR(4000)
+            )
+
+        object_id = func.object_id(full_name)
+
+        s = (
+            sql.select(
+                columns.c.column_name,
+                columns.c.data_type,
+                columns.c.is_nullable,
+                columns.c.character_maximum_length,
+                columns.c.numeric_precision,
+                columns.c.numeric_scale,
+                columns.c.column_default,
+                columns.c.collation_name,
+                computed_definition,
+                computed_cols.c.is_persisted,
+                identity_cols.c.is_identity,
+                identity_cols.c.seed_value,
+                identity_cols.c.increment_value,
+                ischema.extended_properties.c.value.label("comment"),
+            )
+            .select_from(columns)
+            .outerjoin(
+                computed_cols,
+                onclause=sql.and_(
+                    computed_cols.c.object_id == object_id,
+                    computed_cols.c.name
+                    == columns.c.column_name.collate("DATABASE_DEFAULT"),
+                ),
+            )
+            .outerjoin(
+                identity_cols,
+                onclause=sql.and_(
+                    identity_cols.c.object_id == object_id,
+                    identity_cols.c.name
+                    == columns.c.column_name.collate("DATABASE_DEFAULT"),
+                ),
+            )
+            .outerjoin(
+                ischema.extended_properties,
+                onclause=sql.and_(
+                    ischema.extended_properties.c["class"] == 1,
+                    ischema.extended_properties.c.major_id == object_id,
+                    ischema.extended_properties.c.minor_id
+                    == columns.c.ordinal_position,
+                    ischema.extended_properties.c.name == "MS_Description",
+                ),
+            )
+            .where(whereclause)
+            .order_by(columns.c.ordinal_position)
+        )
+
+        c = connection.execution_options(future_result=True).execute(s)
+
+        cols = []
+        for row in c.mappings():
+            name = row[columns.c.column_name]
+            type_ = row[columns.c.data_type]
+            nullable = row[columns.c.is_nullable] == "YES"
+            charlen = row[columns.c.character_maximum_length]
+            numericprec = row[columns.c.numeric_precision]
+            numericscale = row[columns.c.numeric_scale]
+            default = row[columns.c.column_default]
+            collation = row[columns.c.collation_name]
+            definition = row[computed_definition]
+            is_persisted = row[computed_cols.c.is_persisted]
+            is_identity = row[identity_cols.c.is_identity]
+            identity_start = row[identity_cols.c.seed_value]
+            identity_increment = row[identity_cols.c.increment_value]
+            comment = row[ischema.extended_properties.c.value]
+
+            coltype = self.ischema_names.get(type_, None)
+
+            kwargs = {}
+            if coltype in (
+                MSString,
+                MSChar,
+                MSNVarchar,
+                MSNChar,
+                MSText,
+                MSNText,
+                MSBinary,
+                MSVarBinary,
+                sqltypes.LargeBinary,
+            ):
+                if charlen == -1:
+                    charlen = None
+                kwargs["length"] = charlen
+                if collation:
+                    kwargs["collation"] = collation
+
+            if coltype is None:
+                util.warn(
+                    "Did not recognize type '%s' of column '%s'"
+                    % (type_, name)
+                )
+                coltype = sqltypes.NULLTYPE
+            else:
+                if issubclass(coltype, sqltypes.Numeric):
+                    kwargs["precision"] = numericprec
+
+                    if not issubclass(coltype, sqltypes.Float):
+                        kwargs["scale"] = numericscale
+
+                coltype = coltype(**kwargs)
+            cdict = {
+                "name": name,
+                "type": coltype,
+                "nullable": nullable,
+                "default": default,
+                "autoincrement": is_identity is not None,
+                "comment": comment,
+            }
+
+            if definition is not None and is_persisted is not None:
+                cdict["computed"] = {
+                    "sqltext": definition,
+                    "persisted": is_persisted,
+                }
+
+            if is_identity is not None:
+                # identity_start and identity_increment are Decimal or None
+                if identity_start is None or identity_increment is None:
+                    cdict["identity"] = {}
+                else:
+                    if isinstance(coltype, sqltypes.BigInteger):
+                        start = int(identity_start)
+                        increment = int(identity_increment)
+                    elif isinstance(coltype, sqltypes.Integer):
+                        start = int(identity_start)
+                        increment = int(identity_increment)
+                    else:
+                        start = identity_start
+                        increment = identity_increment
+
+                    cdict["identity"] = {
+                        "start": start,
+                        "increment": increment,
+                    }
+
+            cols.append(cdict)
+
+        if cols:
+            return cols
+        else:
+            return self._default_or_error(
+                connection, tablename, owner, ReflectionDefaults.columns, **kw
+            )
+
+    @reflection.cache
+    @_db_plus_owner
+    def get_pk_constraint(
+        self, connection, tablename, dbname, owner, schema, **kw
+    ):
+        pkeys = []
+        TC = ischema.constraints
+        C = ischema.key_constraints.alias("C")
+
+        # Primary key constraints
+        s = (
+            sql.select(
+                C.c.column_name,
+                TC.c.constraint_type,
+                C.c.constraint_name,
+                func.objectproperty(
+                    func.object_id(
+                        C.c.table_schema + "." + C.c.constraint_name
+                    ),
+                    "CnstIsClustKey",
+                ).label("is_clustered"),
+            )
+            .where(
+                sql.and_(
+                    TC.c.constraint_name == C.c.constraint_name,
+                    TC.c.table_schema == C.c.table_schema,
+                    C.c.table_name == tablename,
+                    C.c.table_schema == owner,
+                ),
+            )
+            .order_by(TC.c.constraint_name, C.c.ordinal_position)
+        )
+        c = connection.execution_options(future_result=True).execute(s)
+        constraint_name = None
+        is_clustered = None
+        for row in c.mappings():
+            if "PRIMARY" in row[TC.c.constraint_type.name]:
+                pkeys.append(row["COLUMN_NAME"])
+                if constraint_name is None:
+                    constraint_name = row[C.c.constraint_name.name]
+                if is_clustered is None:
+                    is_clustered = row["is_clustered"]
+        if pkeys:
+            return {
+                "constrained_columns": pkeys,
+                "name": constraint_name,
+                "dialect_options": {"mssql_clustered": is_clustered},
+            }
+        else:
+            return self._default_or_error(
+                connection,
+                tablename,
+                owner,
+                ReflectionDefaults.pk_constraint,
+                **kw,
+            )
+
+    @reflection.cache
+    @_db_plus_owner
+    def get_foreign_keys(
+        self, connection, tablename, dbname, owner, schema, **kw
+    ):
+        # Foreign key constraints
+        s = (
+            text(
+                """\
+WITH fk_info AS (
+    SELECT
+        ischema_ref_con.constraint_schema,
+        ischema_ref_con.constraint_name,
+        ischema_key_col.ordinal_position,
+        ischema_key_col.table_schema,
+        ischema_key_col.table_name,
+        ischema_ref_con.unique_constraint_schema,
+        ischema_ref_con.unique_constraint_name,
+        ischema_ref_con.match_option,
+        ischema_ref_con.update_rule,
+        ischema_ref_con.delete_rule,
+        ischema_key_col.column_name AS constrained_column
+    FROM
+        INFORMATION_SCHEMA.REFERENTIAL_CONSTRAINTS ischema_ref_con
+        INNER JOIN
+        INFORMATION_SCHEMA.KEY_COLUMN_USAGE ischema_key_col ON
+            ischema_key_col.table_schema = ischema_ref_con.constraint_schema
+            AND ischema_key_col.constraint_name =
+            ischema_ref_con.constraint_name
+    WHERE ischema_key_col.table_name = :tablename
+        AND ischema_key_col.table_schema = :owner
+),
+constraint_info AS (
+    SELECT
+        ischema_key_col.constraint_schema,
+        ischema_key_col.constraint_name,
+        ischema_key_col.ordinal_position,
+        ischema_key_col.table_schema,
+        ischema_key_col.table_name,
+        ischema_key_col.column_name
+    FROM
+        INFORMATION_SCHEMA.KEY_COLUMN_USAGE ischema_key_col
+),
+index_info AS (
+    SELECT
+        sys.schemas.name AS index_schema,
+        sys.indexes.name AS index_name,
+        sys.index_columns.key_ordinal AS ordinal_position,
+        sys.schemas.name AS table_schema,
+        sys.objects.name AS table_name,
+        sys.columns.name AS column_name
+    FROM
+        sys.indexes
+        INNER JOIN
+        sys.objects ON
+            sys.objects.object_id = sys.indexes.object_id
+        INNER JOIN
+        sys.schemas ON
+            sys.schemas.schema_id = sys.objects.schema_id
+        INNER JOIN
+        sys.index_columns ON
+            sys.index_columns.object_id = sys.objects.object_id
+            AND sys.index_columns.index_id = sys.indexes.index_id
+        INNER JOIN
+        sys.columns ON
+            sys.columns.object_id = sys.indexes.object_id
+            AND sys.columns.column_id = sys.index_columns.column_id
+)
+    SELECT
+        fk_info.constraint_schema,
+        fk_info.constraint_name,
+        fk_info.ordinal_position,
+        fk_info.constrained_column,
+        constraint_info.table_schema AS referred_table_schema,
+        constraint_info.table_name AS referred_table_name,
+        constraint_info.column_name AS referred_column,
+        fk_info.match_option,
+        fk_info.update_rule,
+        fk_info.delete_rule
+    FROM
+        fk_info INNER JOIN constraint_info ON
+            constraint_info.constraint_schema =
+                fk_info.unique_constraint_schema
+            AND constraint_info.constraint_name =
+                fk_info.unique_constraint_name
+            AND constraint_info.ordinal_position = fk_info.ordinal_position
+    UNION
+    SELECT
+        fk_info.constraint_schema,
+        fk_info.constraint_name,
+        fk_info.ordinal_position,
+        fk_info.constrained_column,
+        index_info.table_schema AS referred_table_schema,
+        index_info.table_name AS referred_table_name,
+        index_info.column_name AS referred_column,
+        fk_info.match_option,
+        fk_info.update_rule,
+        fk_info.delete_rule
+    FROM
+        fk_info INNER JOIN index_info ON
+            index_info.index_schema = fk_info.unique_constraint_schema
+            AND index_info.index_name = fk_info.unique_constraint_name
+            AND index_info.ordinal_position = fk_info.ordinal_position
+
+    ORDER BY fk_info.constraint_schema, fk_info.constraint_name,
+        fk_info.ordinal_position
+"""
+            )
+            .bindparams(
+                sql.bindparam("tablename", tablename, ischema.CoerceUnicode()),
+                sql.bindparam("owner", owner, ischema.CoerceUnicode()),
+            )
+            .columns(
+                constraint_schema=sqltypes.Unicode(),
+                constraint_name=sqltypes.Unicode(),
+                table_schema=sqltypes.Unicode(),
+                table_name=sqltypes.Unicode(),
+                constrained_column=sqltypes.Unicode(),
+                referred_table_schema=sqltypes.Unicode(),
+                referred_table_name=sqltypes.Unicode(),
+                referred_column=sqltypes.Unicode(),
+            )
+        )
+
+        # group rows by constraint ID, to handle multi-column FKs
+        fkeys = []
+
+        def fkey_rec():
+            return {
+                "name": None,
+                "constrained_columns": [],
+                "referred_schema": None,
+                "referred_table": None,
+                "referred_columns": [],
+                "options": {},
+            }
+
+        fkeys = util.defaultdict(fkey_rec)
+
+        for r in connection.execute(s).all():
+            (
+                _,  # constraint schema
+                rfknm,
+                _,  # ordinal position
+                scol,
+                rschema,
+                rtbl,
+                rcol,
+                # TODO: we support match=<keyword> for foreign keys so
+                # we can support this also, PG has match=FULL for example
+                # but this seems to not be a valid value for SQL Server
+                _,  # match rule
+                fkuprule,
+                fkdelrule,
+            ) = r
+
+            rec = fkeys[rfknm]
+            rec["name"] = rfknm
+
+            if fkuprule != "NO ACTION":
+                rec["options"]["onupdate"] = fkuprule
+
+            if fkdelrule != "NO ACTION":
+                rec["options"]["ondelete"] = fkdelrule
+
+            if not rec["referred_table"]:
+                rec["referred_table"] = rtbl
+                if schema is not None or owner != rschema:
+                    if dbname:
+                        rschema = dbname + "." + rschema
+                    rec["referred_schema"] = rschema
+
+            local_cols, remote_cols = (
+                rec["constrained_columns"],
+                rec["referred_columns"],
+            )
+
+            local_cols.append(scol)
+            remote_cols.append(rcol)
+
+        if fkeys:
+            return list(fkeys.values())
+        else:
+            return self._default_or_error(
+                connection,
+                tablename,
+                owner,
+                ReflectionDefaults.foreign_keys,
+                **kw,
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/information_schema.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/information_schema.py
new file mode 100644
index 00000000..b60bb158
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/information_schema.py
@@ -0,0 +1,254 @@
+# dialects/mssql/information_schema.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import cast
+from ... import Column
+from ... import MetaData
+from ... import Table
+from ...ext.compiler import compiles
+from ...sql import expression
+from ...types import Boolean
+from ...types import Integer
+from ...types import Numeric
+from ...types import NVARCHAR
+from ...types import String
+from ...types import TypeDecorator
+from ...types import Unicode
+
+
+ischema = MetaData()
+
+
+class CoerceUnicode(TypeDecorator):
+    impl = Unicode
+    cache_ok = True
+
+    def bind_expression(self, bindvalue):
+        return _cast_on_2005(bindvalue)
+
+
+class _cast_on_2005(expression.ColumnElement):
+    def __init__(self, bindvalue):
+        self.bindvalue = bindvalue
+
+
+@compiles(_cast_on_2005)
+def _compile(element, compiler, **kw):
+    from . import base
+
+    if (
+        compiler.dialect.server_version_info is None
+        or compiler.dialect.server_version_info < base.MS_2005_VERSION
+    ):
+        return compiler.process(element.bindvalue, **kw)
+    else:
+        return compiler.process(cast(element.bindvalue, Unicode), **kw)
+
+
+schemata = Table(
+    "SCHEMATA",
+    ischema,
+    Column("CATALOG_NAME", CoerceUnicode, key="catalog_name"),
+    Column("SCHEMA_NAME", CoerceUnicode, key="schema_name"),
+    Column("SCHEMA_OWNER", CoerceUnicode, key="schema_owner"),
+    schema="INFORMATION_SCHEMA",
+)
+
+tables = Table(
+    "TABLES",
+    ischema,
+    Column("TABLE_CATALOG", CoerceUnicode, key="table_catalog"),
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("TABLE_TYPE", CoerceUnicode, key="table_type"),
+    schema="INFORMATION_SCHEMA",
+)
+
+columns = Table(
+    "COLUMNS",
+    ischema,
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("COLUMN_NAME", CoerceUnicode, key="column_name"),
+    Column("IS_NULLABLE", Integer, key="is_nullable"),
+    Column("DATA_TYPE", String, key="data_type"),
+    Column("ORDINAL_POSITION", Integer, key="ordinal_position"),
+    Column(
+        "CHARACTER_MAXIMUM_LENGTH", Integer, key="character_maximum_length"
+    ),
+    Column("NUMERIC_PRECISION", Integer, key="numeric_precision"),
+    Column("NUMERIC_SCALE", Integer, key="numeric_scale"),
+    Column("COLUMN_DEFAULT", Integer, key="column_default"),
+    Column("COLLATION_NAME", String, key="collation_name"),
+    schema="INFORMATION_SCHEMA",
+)
+
+mssql_temp_table_columns = Table(
+    "COLUMNS",
+    ischema,
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("COLUMN_NAME", CoerceUnicode, key="column_name"),
+    Column("IS_NULLABLE", Integer, key="is_nullable"),
+    Column("DATA_TYPE", String, key="data_type"),
+    Column("ORDINAL_POSITION", Integer, key="ordinal_position"),
+    Column(
+        "CHARACTER_MAXIMUM_LENGTH", Integer, key="character_maximum_length"
+    ),
+    Column("NUMERIC_PRECISION", Integer, key="numeric_precision"),
+    Column("NUMERIC_SCALE", Integer, key="numeric_scale"),
+    Column("COLUMN_DEFAULT", Integer, key="column_default"),
+    Column("COLLATION_NAME", String, key="collation_name"),
+    schema="tempdb.INFORMATION_SCHEMA",
+)
+
+constraints = Table(
+    "TABLE_CONSTRAINTS",
+    ischema,
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("CONSTRAINT_NAME", CoerceUnicode, key="constraint_name"),
+    Column("CONSTRAINT_TYPE", CoerceUnicode, key="constraint_type"),
+    schema="INFORMATION_SCHEMA",
+)
+
+column_constraints = Table(
+    "CONSTRAINT_COLUMN_USAGE",
+    ischema,
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("COLUMN_NAME", CoerceUnicode, key="column_name"),
+    Column("CONSTRAINT_NAME", CoerceUnicode, key="constraint_name"),
+    schema="INFORMATION_SCHEMA",
+)
+
+key_constraints = Table(
+    "KEY_COLUMN_USAGE",
+    ischema,
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("COLUMN_NAME", CoerceUnicode, key="column_name"),
+    Column("CONSTRAINT_NAME", CoerceUnicode, key="constraint_name"),
+    Column("CONSTRAINT_SCHEMA", CoerceUnicode, key="constraint_schema"),
+    Column("ORDINAL_POSITION", Integer, key="ordinal_position"),
+    schema="INFORMATION_SCHEMA",
+)
+
+ref_constraints = Table(
+    "REFERENTIAL_CONSTRAINTS",
+    ischema,
+    Column("CONSTRAINT_CATALOG", CoerceUnicode, key="constraint_catalog"),
+    Column("CONSTRAINT_SCHEMA", CoerceUnicode, key="constraint_schema"),
+    Column("CONSTRAINT_NAME", CoerceUnicode, key="constraint_name"),
+    # TODO: is CATLOG misspelled ?
+    Column(
+        "UNIQUE_CONSTRAINT_CATLOG",
+        CoerceUnicode,
+        key="unique_constraint_catalog",
+    ),
+    Column(
+        "UNIQUE_CONSTRAINT_SCHEMA",
+        CoerceUnicode,
+        key="unique_constraint_schema",
+    ),
+    Column(
+        "UNIQUE_CONSTRAINT_NAME", CoerceUnicode, key="unique_constraint_name"
+    ),
+    Column("MATCH_OPTION", String, key="match_option"),
+    Column("UPDATE_RULE", String, key="update_rule"),
+    Column("DELETE_RULE", String, key="delete_rule"),
+    schema="INFORMATION_SCHEMA",
+)
+
+views = Table(
+    "VIEWS",
+    ischema,
+    Column("TABLE_CATALOG", CoerceUnicode, key="table_catalog"),
+    Column("TABLE_SCHEMA", CoerceUnicode, key="table_schema"),
+    Column("TABLE_NAME", CoerceUnicode, key="table_name"),
+    Column("VIEW_DEFINITION", CoerceUnicode, key="view_definition"),
+    Column("CHECK_OPTION", String, key="check_option"),
+    Column("IS_UPDATABLE", String, key="is_updatable"),
+    schema="INFORMATION_SCHEMA",
+)
+
+computed_columns = Table(
+    "computed_columns",
+    ischema,
+    Column("object_id", Integer),
+    Column("name", CoerceUnicode),
+    Column("is_computed", Boolean),
+    Column("is_persisted", Boolean),
+    Column("definition", CoerceUnicode),
+    schema="sys",
+)
+
+sequences = Table(
+    "SEQUENCES",
+    ischema,
+    Column("SEQUENCE_CATALOG", CoerceUnicode, key="sequence_catalog"),
+    Column("SEQUENCE_SCHEMA", CoerceUnicode, key="sequence_schema"),
+    Column("SEQUENCE_NAME", CoerceUnicode, key="sequence_name"),
+    schema="INFORMATION_SCHEMA",
+)
+
+
+class NumericSqlVariant(TypeDecorator):
+    r"""This type casts sql_variant columns in the identity_columns view
+    to numeric. This is required because:
+
+    * pyodbc does not support sql_variant
+    * pymssql under python 2 return the byte representation of the number,
+      int 1 is returned as "\x01\x00\x00\x00". On python 3 it returns the
+      correct value as string.
+    """
+
+    impl = Unicode
+    cache_ok = True
+
+    def column_expression(self, colexpr):
+        return cast(colexpr, Numeric(38, 0))
+
+
+identity_columns = Table(
+    "identity_columns",
+    ischema,
+    Column("object_id", Integer),
+    Column("name", CoerceUnicode),
+    Column("is_identity", Boolean),
+    Column("seed_value", NumericSqlVariant),
+    Column("increment_value", NumericSqlVariant),
+    Column("last_value", NumericSqlVariant),
+    Column("is_not_for_replication", Boolean),
+    schema="sys",
+)
+
+
+class NVarcharSqlVariant(TypeDecorator):
+    """This type casts sql_variant columns in the extended_properties view
+    to nvarchar. This is required because pyodbc does not support sql_variant
+    """
+
+    impl = Unicode
+    cache_ok = True
+
+    def column_expression(self, colexpr):
+        return cast(colexpr, NVARCHAR)
+
+
+extended_properties = Table(
+    "extended_properties",
+    ischema,
+    Column("class", Integer),  # TINYINT
+    Column("class_desc", CoerceUnicode),
+    Column("major_id", Integer),
+    Column("minor_id", Integer),
+    Column("name", CoerceUnicode),
+    Column("value", NVarcharSqlVariant),
+    schema="sys",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/json.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/json.py
new file mode 100644
index 00000000..a2d3ce81
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/json.py
@@ -0,0 +1,129 @@
+# dialects/mssql/json.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import types as sqltypes
+
+# technically, all the dialect-specific datatypes that don't have any special
+# behaviors would be private with names like _MSJson. However, we haven't been
+# doing this for mysql.JSON or sqlite.JSON which both have JSON / JSONIndexType
+# / JSONPathType in their json.py files, so keep consistent with that
+# sub-convention for now.  A future change can update them all to be
+# package-private at once.
+
+
+class JSON(sqltypes.JSON):
+    """MSSQL JSON type.
+
+    MSSQL supports JSON-formatted data as of SQL Server 2016.
+
+    The :class:`_mssql.JSON` datatype at the DDL level will represent the
+    datatype as ``NVARCHAR(max)``, but provides for JSON-level comparison
+    functions as well as Python coercion behavior.
+
+    :class:`_mssql.JSON` is used automatically whenever the base
+    :class:`_types.JSON` datatype is used against a SQL Server backend.
+
+    .. seealso::
+
+        :class:`_types.JSON` - main documentation for the generic
+        cross-platform JSON datatype.
+
+    The :class:`_mssql.JSON` type supports persistence of JSON values
+    as well as the core index operations provided by :class:`_types.JSON`
+    datatype, by adapting the operations to render the ``JSON_VALUE``
+    or ``JSON_QUERY`` functions at the database level.
+
+    The SQL Server :class:`_mssql.JSON` type necessarily makes use of the
+    ``JSON_QUERY`` and ``JSON_VALUE`` functions when querying for elements
+    of a JSON object.   These two functions have a major restriction in that
+    they are **mutually exclusive** based on the type of object to be returned.
+    The ``JSON_QUERY`` function **only** returns a JSON dictionary or list,
+    but not an individual string, numeric, or boolean element; the
+    ``JSON_VALUE`` function **only** returns an individual string, numeric,
+    or boolean element.   **both functions either return NULL or raise
+    an error if they are not used against the correct expected value**.
+
+    To handle this awkward requirement, indexed access rules are as follows:
+
+    1. When extracting a sub element from a JSON that is itself a JSON
+       dictionary or list, the :meth:`_types.JSON.Comparator.as_json` accessor
+       should be used::
+
+            stmt = select(data_table.c.data["some key"].as_json()).where(
+                data_table.c.data["some key"].as_json() == {"sub": "structure"}
+            )
+
+    2. When extracting a sub element from a JSON that is a plain boolean,
+       string, integer, or float, use the appropriate method among
+       :meth:`_types.JSON.Comparator.as_boolean`,
+       :meth:`_types.JSON.Comparator.as_string`,
+       :meth:`_types.JSON.Comparator.as_integer`,
+       :meth:`_types.JSON.Comparator.as_float`::
+
+            stmt = select(data_table.c.data["some key"].as_string()).where(
+                data_table.c.data["some key"].as_string() == "some string"
+            )
+
+    .. versionadded:: 1.4
+
+
+    """
+
+    # note there was a result processor here that was looking for "number",
+    # but none of the tests seem to exercise it.
+
+
+# Note: these objects currently match exactly those of MySQL, however since
+# these are not generalizable to all JSON implementations, remain separately
+# implemented for each dialect.
+class _FormatTypeMixin:
+    def _format_value(self, value):
+        raise NotImplementedError()
+
+    def bind_processor(self, dialect):
+        super_proc = self.string_bind_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+    def literal_processor(self, dialect):
+        super_proc = self.string_literal_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+
+class JSONIndexType(_FormatTypeMixin, sqltypes.JSON.JSONIndexType):
+    def _format_value(self, value):
+        if isinstance(value, int):
+            value = "$[%s]" % value
+        else:
+            value = '$."%s"' % value
+        return value
+
+
+class JSONPathType(_FormatTypeMixin, sqltypes.JSON.JSONPathType):
+    def _format_value(self, value):
+        return "$%s" % (
+            "".join(
+                [
+                    "[%s]" % elem if isinstance(elem, int) else '."%s"' % elem
+                    for elem in value
+                ]
+            )
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/provision.py
new file mode 100644
index 00000000..10165856
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/provision.py
@@ -0,0 +1,162 @@
+# dialects/mssql/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from sqlalchemy import inspect
+from sqlalchemy import Integer
+from ... import create_engine
+from ... import exc
+from ...schema import Column
+from ...schema import DropConstraint
+from ...schema import ForeignKeyConstraint
+from ...schema import MetaData
+from ...schema import Table
+from ...testing.provision import create_db
+from ...testing.provision import drop_all_schema_objects_pre_tables
+from ...testing.provision import drop_db
+from ...testing.provision import generate_driver_url
+from ...testing.provision import get_temp_table_name
+from ...testing.provision import log
+from ...testing.provision import normalize_sequence
+from ...testing.provision import post_configure_engine
+from ...testing.provision import run_reap_dbs
+from ...testing.provision import temp_table_keyword_args
+
+
+@post_configure_engine.for_db("mssql")
+def post_configure_engine(url, engine, follower_ident):
+    if engine.driver == "pyodbc":
+        engine.dialect.dbapi.pooling = False
+
+
+@generate_driver_url.for_db("mssql")
+def generate_driver_url(url, driver, query_str):
+    backend = url.get_backend_name()
+
+    new_url = url.set(drivername="%s+%s" % (backend, driver))
+
+    if driver not in ("pyodbc", "aioodbc"):
+        new_url = new_url.set(query="")
+
+    if driver == "aioodbc":
+        new_url = new_url.update_query_dict({"MARS_Connection": "Yes"})
+
+    if query_str:
+        new_url = new_url.update_query_string(query_str)
+
+    try:
+        new_url.get_dialect()
+    except exc.NoSuchModuleError:
+        return None
+    else:
+        return new_url
+
+
+@create_db.for_db("mssql")
+def _mssql_create_db(cfg, eng, ident):
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        conn.exec_driver_sql("create database %s" % ident)
+        conn.exec_driver_sql(
+            "ALTER DATABASE %s SET ALLOW_SNAPSHOT_ISOLATION ON" % ident
+        )
+        conn.exec_driver_sql(
+            "ALTER DATABASE %s SET READ_COMMITTED_SNAPSHOT ON" % ident
+        )
+        conn.exec_driver_sql("use %s" % ident)
+        conn.exec_driver_sql("create schema test_schema")
+        conn.exec_driver_sql("create schema test_schema_2")
+
+
+@drop_db.for_db("mssql")
+def _mssql_drop_db(cfg, eng, ident):
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        _mssql_drop_ignore(conn, ident)
+
+
+def _mssql_drop_ignore(conn, ident):
+    try:
+        # typically when this happens, we can't KILL the session anyway,
+        # so let the cleanup process drop the DBs
+        # for row in conn.exec_driver_sql(
+        #     "select session_id from sys.dm_exec_sessions "
+        #        "where database_id=db_id('%s')" % ident):
+        #    log.info("killing SQL server session %s", row['session_id'])
+        #    conn.exec_driver_sql("kill %s" % row['session_id'])
+        conn.exec_driver_sql("drop database %s" % ident)
+        log.info("Reaped db: %s", ident)
+        return True
+    except exc.DatabaseError as err:
+        log.warning("couldn't drop db: %s", err)
+        return False
+
+
+@run_reap_dbs.for_db("mssql")
+def _reap_mssql_dbs(url, idents):
+    log.info("db reaper connecting to %r", url)
+    eng = create_engine(url)
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        log.info("identifiers in file: %s", ", ".join(idents))
+
+        to_reap = conn.exec_driver_sql(
+            "select d.name from sys.databases as d where name "
+            "like 'TEST_%' and not exists (select session_id "
+            "from sys.dm_exec_sessions "
+            "where database_id=d.database_id)"
+        )
+        all_names = {dbname.lower() for (dbname,) in to_reap}
+        to_drop = set()
+        for name in all_names:
+            if name in idents:
+                to_drop.add(name)
+
+        dropped = total = 0
+        for total, dbname in enumerate(to_drop, 1):
+            if _mssql_drop_ignore(conn, dbname):
+                dropped += 1
+        log.info(
+            "Dropped %d out of %d stale databases detected", dropped, total
+        )
+
+
+@temp_table_keyword_args.for_db("mssql")
+def _mssql_temp_table_keyword_args(cfg, eng):
+    return {}
+
+
+@get_temp_table_name.for_db("mssql")
+def _mssql_get_temp_table_name(cfg, eng, base_name):
+    return "##" + base_name
+
+
+@drop_all_schema_objects_pre_tables.for_db("mssql")
+def drop_all_schema_objects_pre_tables(cfg, eng):
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        inspector = inspect(conn)
+        for schema in (None, "dbo", cfg.test_schema, cfg.test_schema_2):
+            for tname in inspector.get_table_names(schema=schema):
+                tb = Table(
+                    tname,
+                    MetaData(),
+                    Column("x", Integer),
+                    Column("y", Integer),
+                    schema=schema,
+                )
+                for fk in inspect(conn).get_foreign_keys(tname, schema=schema):
+                    conn.execute(
+                        DropConstraint(
+                            ForeignKeyConstraint(
+                                [tb.c.x], [tb.c.y], name=fk["name"]
+                            )
+                        )
+                    )
+
+
+@normalize_sequence.for_db("mssql")
+def normalize_sequence(cfg, sequence):
+    if sequence.start is None:
+        sequence.start = 1
+    return sequence
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pymssql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pymssql.py
new file mode 100644
index 00000000..301a98eb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pymssql.py
@@ -0,0 +1,126 @@
+# dialects/mssql/pymssql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""
+.. dialect:: mssql+pymssql
+    :name: pymssql
+    :dbapi: pymssql
+    :connectstring: mssql+pymssql://<username>:<password>@<freetds_name>/?charset=utf8
+
+pymssql is a Python module that provides a Python DBAPI interface around
+`FreeTDS <https://www.freetds.org/>`_.
+
+.. versionchanged:: 2.0.5
+
+    pymssql was restored to SQLAlchemy's continuous integration testing
+
+
+"""  # noqa
+import re
+
+from .base import MSDialect
+from .base import MSIdentifierPreparer
+from ... import types as sqltypes
+from ... import util
+from ...engine import processors
+
+
+class _MSNumeric_pymssql(sqltypes.Numeric):
+    def result_processor(self, dialect, type_):
+        if not self.asdecimal:
+            return processors.to_float
+        else:
+            return sqltypes.Numeric.result_processor(self, dialect, type_)
+
+
+class MSIdentifierPreparer_pymssql(MSIdentifierPreparer):
+    def __init__(self, dialect):
+        super().__init__(dialect)
+        # pymssql has the very unusual behavior that it uses pyformat
+        # yet does not require that percent signs be doubled
+        self._double_percents = False
+
+
+class MSDialect_pymssql(MSDialect):
+    supports_statement_cache = True
+    supports_native_decimal = True
+    supports_native_uuid = True
+    driver = "pymssql"
+
+    preparer = MSIdentifierPreparer_pymssql
+
+    colspecs = util.update_copy(
+        MSDialect.colspecs,
+        {sqltypes.Numeric: _MSNumeric_pymssql, sqltypes.Float: sqltypes.Float},
+    )
+
+    @classmethod
+    def import_dbapi(cls):
+        module = __import__("pymssql")
+        # pymmsql < 2.1.1 doesn't have a Binary method.  we use string
+        client_ver = tuple(int(x) for x in module.__version__.split("."))
+        if client_ver < (2, 1, 1):
+            # TODO: monkeypatching here is less than ideal
+            module.Binary = lambda x: x if hasattr(x, "decode") else str(x)
+
+        if client_ver < (1,):
+            util.warn(
+                "The pymssql dialect expects at least "
+                "the 1.0 series of the pymssql DBAPI."
+            )
+        return module
+
+    def _get_server_version_info(self, connection):
+        vers = connection.exec_driver_sql("select @@version").scalar()
+        m = re.match(r"Microsoft .*? - (\d+)\.(\d+)\.(\d+)\.(\d+)", vers)
+        if m:
+            return tuple(int(x) for x in m.group(1, 2, 3, 4))
+        else:
+            return None
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args(username="user")
+        opts.update(url.query)
+        port = opts.pop("port", None)
+        if port and "host" in opts:
+            opts["host"] = "%s:%s" % (opts["host"], port)
+        return ([], opts)
+
+    def is_disconnect(self, e, connection, cursor):
+        for msg in (
+            "Adaptive Server connection timed out",
+            "Net-Lib error during Connection reset by peer",
+            "message 20003",  # connection timeout
+            "Error 10054",
+            "Not connected to any MS SQL server",
+            "Connection is closed",
+            "message 20006",  # Write to the server failed
+            "message 20017",  # Unexpected EOF from the server
+            "message 20047",  # DBPROCESS is dead or not enabled
+            "The server failed to resume the transaction",
+        ):
+            if msg in str(e):
+                return True
+        else:
+            return False
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return super().get_isolation_level_values(dbapi_connection) + [
+            "AUTOCOMMIT"
+        ]
+
+    def set_isolation_level(self, dbapi_connection, level):
+        if level == "AUTOCOMMIT":
+            dbapi_connection.autocommit(True)
+        else:
+            dbapi_connection.autocommit(False)
+            super().set_isolation_level(dbapi_connection, level)
+
+
+dialect = MSDialect_pymssql
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pyodbc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pyodbc.py
new file mode 100644
index 00000000..cbf0adbf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mssql/pyodbc.py
@@ -0,0 +1,760 @@
+# dialects/mssql/pyodbc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: mssql+pyodbc
+    :name: PyODBC
+    :dbapi: pyodbc
+    :connectstring: mssql+pyodbc://<username>:<password>@<dsnname>
+    :url: https://pypi.org/project/pyodbc/
+
+Connecting to PyODBC
+--------------------
+
+The URL here is to be translated to PyODBC connection strings, as
+detailed in `ConnectionStrings <https://code.google.com/p/pyodbc/wiki/ConnectionStrings>`_.
+
+DSN Connections
+^^^^^^^^^^^^^^^
+
+A DSN connection in ODBC means that a pre-existing ODBC datasource is
+configured on the client machine.   The application then specifies the name
+of this datasource, which encompasses details such as the specific ODBC driver
+in use as well as the network address of the database.   Assuming a datasource
+is configured on the client, a basic DSN-based connection looks like::
+
+    engine = create_engine("mssql+pyodbc://scott:tiger@some_dsn")
+
+Which above, will pass the following connection string to PyODBC:
+
+.. sourcecode:: text
+
+    DSN=some_dsn;UID=scott;PWD=tiger
+
+If the username and password are omitted, the DSN form will also add
+the ``Trusted_Connection=yes`` directive to the ODBC string.
+
+Hostname Connections
+^^^^^^^^^^^^^^^^^^^^
+
+Hostname-based connections are also supported by pyodbc.  These are often
+easier to use than a DSN and have the additional advantage that the specific
+database name to connect towards may be specified locally in the URL, rather
+than it being fixed as part of a datasource configuration.
+
+When using a hostname connection, the driver name must also be specified in the
+query parameters of the URL.  As these names usually have spaces in them, the
+name must be URL encoded which means using plus signs for spaces::
+
+    engine = create_engine(
+        "mssql+pyodbc://scott:tiger@myhost:port/databasename?driver=ODBC+Driver+17+for+SQL+Server"
+    )
+
+The ``driver`` keyword is significant to the pyodbc dialect and must be
+specified in lowercase.
+
+Any other names passed in the query string are passed through in the pyodbc
+connect string, such as ``authentication``, ``TrustServerCertificate``, etc.
+Multiple keyword arguments must be separated by an ampersand (``&``); these
+will be translated to semicolons when the pyodbc connect string is generated
+internally::
+
+    e = create_engine(
+        "mssql+pyodbc://scott:tiger@mssql2017:1433/test?"
+        "driver=ODBC+Driver+18+for+SQL+Server&TrustServerCertificate=yes"
+        "&authentication=ActiveDirectoryIntegrated"
+    )
+
+The equivalent URL can be constructed using :class:`_sa.engine.URL`::
+
+    from sqlalchemy.engine import URL
+
+    connection_url = URL.create(
+        "mssql+pyodbc",
+        username="scott",
+        password="tiger",
+        host="mssql2017",
+        port=1433,
+        database="test",
+        query={
+            "driver": "ODBC Driver 18 for SQL Server",
+            "TrustServerCertificate": "yes",
+            "authentication": "ActiveDirectoryIntegrated",
+        },
+    )
+
+Pass through exact Pyodbc string
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+A PyODBC connection string can also be sent in pyodbc's format directly, as
+specified in `the PyODBC documentation
+<https://github.com/mkleehammer/pyodbc/wiki/Connecting-to-databases>`_,
+using the parameter ``odbc_connect``.  A :class:`_sa.engine.URL` object
+can help make this easier::
+
+    from sqlalchemy.engine import URL
+
+    connection_string = "DRIVER={SQL Server Native Client 10.0};SERVER=dagger;DATABASE=test;UID=user;PWD=password"
+    connection_url = URL.create(
+        "mssql+pyodbc", query={"odbc_connect": connection_string}
+    )
+
+    engine = create_engine(connection_url)
+
+.. _mssql_pyodbc_access_tokens:
+
+Connecting to databases with access tokens
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Some database servers are set up to only accept access tokens for login. For
+example, SQL Server allows the use of Azure Active Directory tokens to connect
+to databases. This requires creating a credential object using the
+``azure-identity`` library. More information about the authentication step can be
+found in `Microsoft's documentation
+<https://docs.microsoft.com/en-us/azure/developer/python/azure-sdk-authenticate?tabs=bash>`_.
+
+After getting an engine, the credentials need to be sent to ``pyodbc.connect``
+each time a connection is requested. One way to do this is to set up an event
+listener on the engine that adds the credential token to the dialect's connect
+call. This is discussed more generally in :ref:`engines_dynamic_tokens`. For
+SQL Server in particular, this is passed as an ODBC connection attribute with
+a data structure `described by Microsoft
+<https://docs.microsoft.com/en-us/sql/connect/odbc/using-azure-active-directory#authenticating-with-an-access-token>`_.
+
+The following code snippet will create an engine that connects to an Azure SQL
+database using Azure credentials::
+
+    import struct
+    from sqlalchemy import create_engine, event
+    from sqlalchemy.engine.url import URL
+    from azure import identity
+
+    # Connection option for access tokens, as defined in msodbcsql.h
+    SQL_COPT_SS_ACCESS_TOKEN = 1256
+    TOKEN_URL = "https://database.windows.net/"  # The token URL for any Azure SQL database
+
+    connection_string = "mssql+pyodbc://@my-server.database.windows.net/myDb?driver=ODBC+Driver+17+for+SQL+Server"
+
+    engine = create_engine(connection_string)
+
+    azure_credentials = identity.DefaultAzureCredential()
+
+
+    @event.listens_for(engine, "do_connect")
+    def provide_token(dialect, conn_rec, cargs, cparams):
+        # remove the "Trusted_Connection" parameter that SQLAlchemy adds
+        cargs[0] = cargs[0].replace(";Trusted_Connection=Yes", "")
+
+        # create token credential
+        raw_token = azure_credentials.get_token(TOKEN_URL).token.encode(
+            "utf-16-le"
+        )
+        token_struct = struct.pack(
+            f"<I{len(raw_token)}s", len(raw_token), raw_token
+        )
+
+        # apply it to keyword arguments
+        cparams["attrs_before"] = {SQL_COPT_SS_ACCESS_TOKEN: token_struct}
+
+.. tip::
+
+    The ``Trusted_Connection`` token is currently added by the SQLAlchemy
+    pyodbc dialect when no username or password is present.  This needs
+    to be removed per Microsoft's
+    `documentation for Azure access tokens
+    <https://docs.microsoft.com/en-us/sql/connect/odbc/using-azure-active-directory#authenticating-with-an-access-token>`_,
+    stating that a connection string when using an access token must not contain
+    ``UID``, ``PWD``, ``Authentication`` or ``Trusted_Connection`` parameters.
+
+.. _azure_synapse_ignore_no_transaction_on_rollback:
+
+Avoiding transaction-related exceptions on Azure Synapse Analytics
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Azure Synapse Analytics has a significant difference in its transaction
+handling compared to plain SQL Server; in some cases an error within a Synapse
+transaction can cause it to be arbitrarily terminated on the server side, which
+then causes the DBAPI ``.rollback()`` method (as well as ``.commit()``) to
+fail. The issue prevents the usual DBAPI contract of allowing ``.rollback()``
+to pass silently if no transaction is present as the driver does not expect
+this condition. The symptom of this failure is an exception with a message
+resembling 'No corresponding transaction found. (111214)' when attempting to
+emit a ``.rollback()`` after an operation had a failure of some kind.
+
+This specific case can be handled by passing ``ignore_no_transaction_on_rollback=True`` to
+the SQL Server dialect via the :func:`_sa.create_engine` function as follows::
+
+    engine = create_engine(
+        connection_url, ignore_no_transaction_on_rollback=True
+    )
+
+Using the above parameter, the dialect will catch ``ProgrammingError``
+exceptions raised during ``connection.rollback()`` and emit a warning
+if the error message contains code ``111214``, however will not raise
+an exception.
+
+.. versionadded:: 1.4.40  Added the
+   ``ignore_no_transaction_on_rollback=True`` parameter.
+
+Enable autocommit for Azure SQL Data Warehouse (DW) connections
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Azure SQL Data Warehouse does not support transactions,
+and that can cause problems with SQLAlchemy's "autobegin" (and implicit
+commit/rollback) behavior. We can avoid these problems by enabling autocommit
+at both the pyodbc and engine levels::
+
+    connection_url = sa.engine.URL.create(
+        "mssql+pyodbc",
+        username="scott",
+        password="tiger",
+        host="dw.azure.example.com",
+        database="mydb",
+        query={
+            "driver": "ODBC Driver 17 for SQL Server",
+            "autocommit": "True",
+        },
+    )
+
+    engine = create_engine(connection_url).execution_options(
+        isolation_level="AUTOCOMMIT"
+    )
+
+Avoiding sending large string parameters as TEXT/NTEXT
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+By default, for historical reasons, Microsoft's ODBC drivers for SQL Server
+send long string parameters (greater than 4000 SBCS characters or 2000 Unicode
+characters) as TEXT/NTEXT values. TEXT and NTEXT have been deprecated for many
+years and are starting to cause compatibility issues with newer versions of
+SQL_Server/Azure. For example, see `this
+issue <https://github.com/mkleehammer/pyodbc/issues/835>`_.
+
+Starting with ODBC Driver 18 for SQL Server we can override the legacy
+behavior and pass long strings as varchar(max)/nvarchar(max) using the
+``LongAsMax=Yes`` connection string parameter::
+
+    connection_url = sa.engine.URL.create(
+        "mssql+pyodbc",
+        username="scott",
+        password="tiger",
+        host="mssqlserver.example.com",
+        database="mydb",
+        query={
+            "driver": "ODBC Driver 18 for SQL Server",
+            "LongAsMax": "Yes",
+        },
+    )
+
+Pyodbc Pooling / connection close behavior
+------------------------------------------
+
+PyODBC uses internal `pooling
+<https://github.com/mkleehammer/pyodbc/wiki/The-pyodbc-Module#pooling>`_ by
+default, which means connections will be longer lived than they are within
+SQLAlchemy itself.  As SQLAlchemy has its own pooling behavior, it is often
+preferable to disable this behavior.  This behavior can only be disabled
+globally at the PyODBC module level, **before** any connections are made::
+
+    import pyodbc
+
+    pyodbc.pooling = False
+
+    # don't use the engine before pooling is set to False
+    engine = create_engine("mssql+pyodbc://user:pass@dsn")
+
+If this variable is left at its default value of ``True``, **the application
+will continue to maintain active database connections**, even when the
+SQLAlchemy engine itself fully discards a connection or if the engine is
+disposed.
+
+.. seealso::
+
+    `pooling <https://github.com/mkleehammer/pyodbc/wiki/The-pyodbc-Module#pooling>`_ -
+    in the PyODBC documentation.
+
+Driver / Unicode Support
+-------------------------
+
+PyODBC works best with Microsoft ODBC drivers, particularly in the area
+of Unicode support on both Python 2 and Python 3.
+
+Using the FreeTDS ODBC drivers on Linux or OSX with PyODBC is **not**
+recommended; there have been historically many Unicode-related issues
+in this area, including before Microsoft offered ODBC drivers for Linux
+and OSX.   Now that Microsoft offers drivers for all platforms, for
+PyODBC support these are recommended.  FreeTDS remains relevant for
+non-ODBC drivers such as pymssql where it works very well.
+
+
+Rowcount Support
+----------------
+
+Previous limitations with the SQLAlchemy ORM's "versioned rows" feature with
+Pyodbc have been resolved as of SQLAlchemy 2.0.5. See the notes at
+:ref:`mssql_rowcount_versioning`.
+
+.. _mssql_pyodbc_fastexecutemany:
+
+Fast Executemany Mode
+---------------------
+
+The PyODBC driver includes support for a "fast executemany" mode of execution
+which greatly reduces round trips for a DBAPI ``executemany()`` call when using
+Microsoft ODBC drivers, for **limited size batches that fit in memory**.  The
+feature is enabled by setting the attribute ``.fast_executemany`` on the DBAPI
+cursor when an executemany call is to be used.   The SQLAlchemy PyODBC SQL
+Server dialect supports this parameter by passing the
+``fast_executemany`` parameter to
+:func:`_sa.create_engine` , when using the **Microsoft ODBC driver only**::
+
+    engine = create_engine(
+        "mssql+pyodbc://scott:tiger@mssql2017:1433/test?driver=ODBC+Driver+17+for+SQL+Server",
+        fast_executemany=True,
+    )
+
+.. versionchanged:: 2.0.9 - the ``fast_executemany`` parameter now has its
+   intended effect of this PyODBC feature taking effect for all INSERT
+   statements that are executed with multiple parameter sets, which don't
+   include RETURNING.  Previously, SQLAlchemy 2.0's :term:`insertmanyvalues`
+   feature would cause ``fast_executemany`` to not be used in most cases
+   even if specified.
+
+.. versionadded:: 1.3
+
+.. seealso::
+
+    `fast executemany <https://github.com/mkleehammer/pyodbc/wiki/Features-beyond-the-DB-API#fast_executemany>`_
+    - on github
+
+.. _mssql_pyodbc_setinputsizes:
+
+Setinputsizes Support
+-----------------------
+
+As of version 2.0, the pyodbc ``cursor.setinputsizes()`` method is used for
+all statement executions, except for ``cursor.executemany()`` calls when
+fast_executemany=True where it is not supported (assuming
+:ref:`insertmanyvalues <engine_insertmanyvalues>` is kept enabled,
+"fastexecutemany" will not take place for INSERT statements in any case).
+
+The use of ``cursor.setinputsizes()`` can be disabled by passing
+``use_setinputsizes=False`` to :func:`_sa.create_engine`.
+
+When ``use_setinputsizes`` is left at its default of ``True``, the
+specific per-type symbols passed to ``cursor.setinputsizes()`` can be
+programmatically customized using the :meth:`.DialectEvents.do_setinputsizes`
+hook. See that method for usage examples.
+
+.. versionchanged:: 2.0  The mssql+pyodbc dialect now defaults to using
+   ``use_setinputsizes=True`` for all statement executions with the exception of
+   cursor.executemany() calls when fast_executemany=True.  The behavior can
+   be turned off by passing ``use_setinputsizes=False`` to
+   :func:`_sa.create_engine`.
+
+"""  # noqa
+
+
+import datetime
+import decimal
+import re
+import struct
+
+from .base import _MSDateTime
+from .base import _MSUnicode
+from .base import _MSUnicodeText
+from .base import BINARY
+from .base import DATETIMEOFFSET
+from .base import MSDialect
+from .base import MSExecutionContext
+from .base import VARBINARY
+from .json import JSON as _MSJson
+from .json import JSONIndexType as _MSJsonIndexType
+from .json import JSONPathType as _MSJsonPathType
+from ... import exc
+from ... import types as sqltypes
+from ... import util
+from ...connectors.pyodbc import PyODBCConnector
+from ...engine import cursor as _cursor
+
+
+class _ms_numeric_pyodbc:
+    """Turns Decimals with adjusted() < 0 or > 7 into strings.
+
+    The routines here are needed for older pyodbc versions
+    as well as current mxODBC versions.
+
+    """
+
+    def bind_processor(self, dialect):
+        super_process = super().bind_processor(dialect)
+
+        if not dialect._need_decimal_fix:
+            return super_process
+
+        def process(value):
+            if self.asdecimal and isinstance(value, decimal.Decimal):
+                adjusted = value.adjusted()
+                if adjusted < 0:
+                    return self._small_dec_to_string(value)
+                elif adjusted > 7:
+                    return self._large_dec_to_string(value)
+
+            if super_process:
+                return super_process(value)
+            else:
+                return value
+
+        return process
+
+    # these routines needed for older versions of pyodbc.
+    # as of 2.1.8 this logic is integrated.
+
+    def _small_dec_to_string(self, value):
+        return "%s0.%s%s" % (
+            (value < 0 and "-" or ""),
+            "0" * (abs(value.adjusted()) - 1),
+            "".join([str(nint) for nint in value.as_tuple()[1]]),
+        )
+
+    def _large_dec_to_string(self, value):
+        _int = value.as_tuple()[1]
+        if "E" in str(value):
+            result = "%s%s%s" % (
+                (value < 0 and "-" or ""),
+                "".join([str(s) for s in _int]),
+                "0" * (value.adjusted() - (len(_int) - 1)),
+            )
+        else:
+            if (len(_int) - 1) > value.adjusted():
+                result = "%s%s.%s" % (
+                    (value < 0 and "-" or ""),
+                    "".join([str(s) for s in _int][0 : value.adjusted() + 1]),
+                    "".join([str(s) for s in _int][value.adjusted() + 1 :]),
+                )
+            else:
+                result = "%s%s" % (
+                    (value < 0 and "-" or ""),
+                    "".join([str(s) for s in _int][0 : value.adjusted() + 1]),
+                )
+        return result
+
+
+class _MSNumeric_pyodbc(_ms_numeric_pyodbc, sqltypes.Numeric):
+    pass
+
+
+class _MSFloat_pyodbc(_ms_numeric_pyodbc, sqltypes.Float):
+    pass
+
+
+class _ms_binary_pyodbc:
+    """Wraps binary values in dialect-specific Binary wrapper.
+    If the value is null, return a pyodbc-specific BinaryNull
+    object to prevent pyODBC [and FreeTDS] from defaulting binary
+    NULL types to SQLWCHAR and causing implicit conversion errors.
+    """
+
+    def bind_processor(self, dialect):
+        if dialect.dbapi is None:
+            return None
+
+        DBAPIBinary = dialect.dbapi.Binary
+
+        def process(value):
+            if value is not None:
+                return DBAPIBinary(value)
+            else:
+                # pyodbc-specific
+                return dialect.dbapi.BinaryNull
+
+        return process
+
+
+class _ODBCDateTimeBindProcessor:
+    """Add bind processors to handle datetimeoffset behaviors"""
+
+    has_tz = False
+
+    def bind_processor(self, dialect):
+        def process(value):
+            if value is None:
+                return None
+            elif isinstance(value, str):
+                # if a string was passed directly, allow it through
+                return value
+            elif not value.tzinfo or (not self.timezone and not self.has_tz):
+                # for DateTime(timezone=False)
+                return value
+            else:
+                # for DATETIMEOFFSET or DateTime(timezone=True)
+                #
+                # Convert to string format required by T-SQL
+                dto_string = value.strftime("%Y-%m-%d %H:%M:%S.%f %z")
+                # offset needs a colon, e.g., -0700 -> -07:00
+                # "UTC offset in the form (+-)HHMM[SS[.ffffff]]"
+                # backend currently rejects seconds / fractional seconds
+                dto_string = re.sub(
+                    r"([\+\-]\d{2})([\d\.]+)$", r"\1:\2", dto_string
+                )
+                return dto_string
+
+        return process
+
+
+class _ODBCDateTime(_ODBCDateTimeBindProcessor, _MSDateTime):
+    pass
+
+
+class _ODBCDATETIMEOFFSET(_ODBCDateTimeBindProcessor, DATETIMEOFFSET):
+    has_tz = True
+
+
+class _VARBINARY_pyodbc(_ms_binary_pyodbc, VARBINARY):
+    pass
+
+
+class _BINARY_pyodbc(_ms_binary_pyodbc, BINARY):
+    pass
+
+
+class _String_pyodbc(sqltypes.String):
+    def get_dbapi_type(self, dbapi):
+        if self.length in (None, "max") or self.length >= 2000:
+            return (dbapi.SQL_VARCHAR, 0, 0)
+        else:
+            return dbapi.SQL_VARCHAR
+
+
+class _Unicode_pyodbc(_MSUnicode):
+    def get_dbapi_type(self, dbapi):
+        if self.length in (None, "max") or self.length >= 2000:
+            return (dbapi.SQL_WVARCHAR, 0, 0)
+        else:
+            return dbapi.SQL_WVARCHAR
+
+
+class _UnicodeText_pyodbc(_MSUnicodeText):
+    def get_dbapi_type(self, dbapi):
+        if self.length in (None, "max") or self.length >= 2000:
+            return (dbapi.SQL_WVARCHAR, 0, 0)
+        else:
+            return dbapi.SQL_WVARCHAR
+
+
+class _JSON_pyodbc(_MSJson):
+    def get_dbapi_type(self, dbapi):
+        return (dbapi.SQL_WVARCHAR, 0, 0)
+
+
+class _JSONIndexType_pyodbc(_MSJsonIndexType):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.SQL_WVARCHAR
+
+
+class _JSONPathType_pyodbc(_MSJsonPathType):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.SQL_WVARCHAR
+
+
+class MSExecutionContext_pyodbc(MSExecutionContext):
+    _embedded_scope_identity = False
+
+    def pre_exec(self):
+        """where appropriate, issue "select scope_identity()" in the same
+        statement.
+
+        Background on why "scope_identity()" is preferable to "@@identity":
+        https://msdn.microsoft.com/en-us/library/ms190315.aspx
+
+        Background on why we attempt to embed "scope_identity()" into the same
+        statement as the INSERT:
+        https://code.google.com/p/pyodbc/wiki/FAQs#How_do_I_retrieve_autogenerated/identity_values?
+
+        """
+
+        super().pre_exec()
+
+        # don't embed the scope_identity select into an
+        # "INSERT .. DEFAULT VALUES"
+        if (
+            self._select_lastrowid
+            and self.dialect.use_scope_identity
+            and len(self.parameters[0])
+        ):
+            self._embedded_scope_identity = True
+
+            self.statement += "; select scope_identity()"
+
+    def post_exec(self):
+        if self._embedded_scope_identity:
+            # Fetch the last inserted id from the manipulated statement
+            # We may have to skip over a number of result sets with
+            # no data (due to triggers, etc.)
+            while True:
+                try:
+                    # fetchall() ensures the cursor is consumed
+                    # without closing it (FreeTDS particularly)
+                    rows = self.cursor.fetchall()
+                except self.dialect.dbapi.Error:
+                    # no way around this - nextset() consumes the previous set
+                    # so we need to just keep flipping
+                    self.cursor.nextset()
+                else:
+                    if not rows:
+                        # async adapter drivers just return None here
+                        self.cursor.nextset()
+                        continue
+                    row = rows[0]
+                    break
+
+            self._lastrowid = int(row[0])
+
+            self.cursor_fetch_strategy = _cursor._NO_CURSOR_DML
+        else:
+            super().post_exec()
+
+
+class MSDialect_pyodbc(PyODBCConnector, MSDialect):
+    supports_statement_cache = True
+
+    # note this parameter is no longer used by the ORM or default dialect
+    # see #9414
+    supports_sane_rowcount_returning = False
+
+    execution_ctx_cls = MSExecutionContext_pyodbc
+
+    colspecs = util.update_copy(
+        MSDialect.colspecs,
+        {
+            sqltypes.Numeric: _MSNumeric_pyodbc,
+            sqltypes.Float: _MSFloat_pyodbc,
+            BINARY: _BINARY_pyodbc,
+            # support DateTime(timezone=True)
+            sqltypes.DateTime: _ODBCDateTime,
+            DATETIMEOFFSET: _ODBCDATETIMEOFFSET,
+            # SQL Server dialect has a VARBINARY that is just to support
+            # "deprecate_large_types" w/ VARBINARY(max), but also we must
+            # handle the usual SQL standard VARBINARY
+            VARBINARY: _VARBINARY_pyodbc,
+            sqltypes.VARBINARY: _VARBINARY_pyodbc,
+            sqltypes.LargeBinary: _VARBINARY_pyodbc,
+            sqltypes.String: _String_pyodbc,
+            sqltypes.Unicode: _Unicode_pyodbc,
+            sqltypes.UnicodeText: _UnicodeText_pyodbc,
+            sqltypes.JSON: _JSON_pyodbc,
+            sqltypes.JSON.JSONIndexType: _JSONIndexType_pyodbc,
+            sqltypes.JSON.JSONPathType: _JSONPathType_pyodbc,
+            # this excludes Enum from the string/VARCHAR thing for now
+            # it looks like Enum's adaptation doesn't really support the
+            # String type itself having a dialect-level impl
+            sqltypes.Enum: sqltypes.Enum,
+        },
+    )
+
+    def __init__(
+        self,
+        fast_executemany=False,
+        use_setinputsizes=True,
+        **params,
+    ):
+        super().__init__(use_setinputsizes=use_setinputsizes, **params)
+        self.use_scope_identity = (
+            self.use_scope_identity
+            and self.dbapi
+            and hasattr(self.dbapi.Cursor, "nextset")
+        )
+        self._need_decimal_fix = self.dbapi and self._dbapi_version() < (
+            2,
+            1,
+            8,
+        )
+        self.fast_executemany = fast_executemany
+        if fast_executemany:
+            self.use_insertmanyvalues_wo_returning = False
+
+    def _get_server_version_info(self, connection):
+        try:
+            # "Version of the instance of SQL Server, in the form
+            # of 'major.minor.build.revision'"
+            raw = connection.exec_driver_sql(
+                "SELECT CAST(SERVERPROPERTY('ProductVersion') AS VARCHAR)"
+            ).scalar()
+        except exc.DBAPIError:
+            # SQL Server docs indicate this function isn't present prior to
+            # 2008.  Before we had the VARCHAR cast above, pyodbc would also
+            # fail on this query.
+            return super()._get_server_version_info(connection)
+        else:
+            version = []
+            r = re.compile(r"[.\-]")
+            for n in r.split(raw):
+                try:
+                    version.append(int(n))
+                except ValueError:
+                    pass
+            return tuple(version)
+
+    def on_connect(self):
+        super_ = super().on_connect()
+
+        def on_connect(conn):
+            if super_ is not None:
+                super_(conn)
+
+            self._setup_timestampoffset_type(conn)
+
+        return on_connect
+
+    def _setup_timestampoffset_type(self, connection):
+        # output converter function for datetimeoffset
+        def _handle_datetimeoffset(dto_value):
+            tup = struct.unpack("<6hI2h", dto_value)
+            return datetime.datetime(
+                tup[0],
+                tup[1],
+                tup[2],
+                tup[3],
+                tup[4],
+                tup[5],
+                tup[6] // 1000,
+                datetime.timezone(
+                    datetime.timedelta(hours=tup[7], minutes=tup[8])
+                ),
+            )
+
+        odbc_SQL_SS_TIMESTAMPOFFSET = -155  # as defined in SQLNCLI.h
+        connection.add_output_converter(
+            odbc_SQL_SS_TIMESTAMPOFFSET, _handle_datetimeoffset
+        )
+
+    def do_executemany(self, cursor, statement, parameters, context=None):
+        if self.fast_executemany:
+            cursor.fast_executemany = True
+        super().do_executemany(cursor, statement, parameters, context=context)
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(e, self.dbapi.Error):
+            code = e.args[0]
+            if code in {
+                "08S01",
+                "01000",
+                "01002",
+                "08003",
+                "08007",
+                "08S02",
+                "08001",
+                "HYT00",
+                "HY010",
+                "10054",
+            }:
+                return True
+        return super().is_disconnect(e, connection, cursor)
+
+
+dialect = MSDialect_pyodbc
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__init__.py
new file mode 100644
index 00000000..9174c544
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__init__.py
@@ -0,0 +1,104 @@
+# dialects/mysql/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from . import aiomysql  # noqa
+from . import asyncmy  # noqa
+from . import base  # noqa
+from . import cymysql  # noqa
+from . import mariadbconnector  # noqa
+from . import mysqlconnector  # noqa
+from . import mysqldb  # noqa
+from . import pymysql  # noqa
+from . import pyodbc  # noqa
+from .base import BIGINT
+from .base import BINARY
+from .base import BIT
+from .base import BLOB
+from .base import BOOLEAN
+from .base import CHAR
+from .base import DATE
+from .base import DATETIME
+from .base import DECIMAL
+from .base import DOUBLE
+from .base import ENUM
+from .base import FLOAT
+from .base import INTEGER
+from .base import JSON
+from .base import LONGBLOB
+from .base import LONGTEXT
+from .base import MEDIUMBLOB
+from .base import MEDIUMINT
+from .base import MEDIUMTEXT
+from .base import NCHAR
+from .base import NUMERIC
+from .base import NVARCHAR
+from .base import REAL
+from .base import SET
+from .base import SMALLINT
+from .base import TEXT
+from .base import TIME
+from .base import TIMESTAMP
+from .base import TINYBLOB
+from .base import TINYINT
+from .base import TINYTEXT
+from .base import VARBINARY
+from .base import VARCHAR
+from .base import YEAR
+from .dml import Insert
+from .dml import insert
+from .expression import match
+from .mariadb import INET4
+from .mariadb import INET6
+
+# default dialect
+base.dialect = dialect = mysqldb.dialect
+
+__all__ = (
+    "BIGINT",
+    "BINARY",
+    "BIT",
+    "BLOB",
+    "BOOLEAN",
+    "CHAR",
+    "DATE",
+    "DATETIME",
+    "DECIMAL",
+    "DOUBLE",
+    "ENUM",
+    "FLOAT",
+    "INET4",
+    "INET6",
+    "INTEGER",
+    "INTEGER",
+    "JSON",
+    "LONGBLOB",
+    "LONGTEXT",
+    "MEDIUMBLOB",
+    "MEDIUMINT",
+    "MEDIUMTEXT",
+    "NCHAR",
+    "NVARCHAR",
+    "NUMERIC",
+    "SET",
+    "SMALLINT",
+    "REAL",
+    "TEXT",
+    "TIME",
+    "TIMESTAMP",
+    "TINYBLOB",
+    "TINYINT",
+    "TINYTEXT",
+    "VARBINARY",
+    "VARCHAR",
+    "YEAR",
+    "dialect",
+    "insert",
+    "Insert",
+    "match",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..71117205
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/aiomysql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/aiomysql.cpython-312.pyc
new file mode 100644
index 00000000..d865b830
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/aiomysql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/asyncmy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/asyncmy.cpython-312.pyc
new file mode 100644
index 00000000..a0e3699a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/asyncmy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..8bf5965f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/cymysql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/cymysql.cpython-312.pyc
new file mode 100644
index 00000000..958d25b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/cymysql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/dml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/dml.cpython-312.pyc
new file mode 100644
index 00000000..dc909736
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/dml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/enumerated.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/enumerated.cpython-312.pyc
new file mode 100644
index 00000000..f1609291
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/enumerated.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/expression.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/expression.cpython-312.pyc
new file mode 100644
index 00000000..078df76e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/expression.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/json.cpython-312.pyc
new file mode 100644
index 00000000..dcffc2e5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadb.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadb.cpython-312.pyc
new file mode 100644
index 00000000..6a5ad307
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadb.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadbconnector.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadbconnector.cpython-312.pyc
new file mode 100644
index 00000000..15362cb5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mariadbconnector.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqlconnector.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqlconnector.cpython-312.pyc
new file mode 100644
index 00000000..a5032cbc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqlconnector.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqldb.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqldb.cpython-312.pyc
new file mode 100644
index 00000000..7ee996ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/mysqldb.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..12a16072
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pymysql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pymysql.cpython-312.pyc
new file mode 100644
index 00000000..60a2b8b0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pymysql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pyodbc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pyodbc.cpython-312.pyc
new file mode 100644
index 00000000..efc15857
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/pyodbc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reflection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reflection.cpython-312.pyc
new file mode 100644
index 00000000..36e4321b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reflection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reserved_words.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reserved_words.cpython-312.pyc
new file mode 100644
index 00000000..78ad8c2f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/reserved_words.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..8c8207e1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/aiomysql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/aiomysql.py
new file mode 100644
index 00000000..bd5e7de6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/aiomysql.py
@@ -0,0 +1,335 @@
+# dialects/mysql/aiomysql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors <see AUTHORS
+# file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: mysql+aiomysql
+    :name: aiomysql
+    :dbapi: aiomysql
+    :connectstring: mysql+aiomysql://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://github.com/aio-libs/aiomysql
+
+The aiomysql dialect is SQLAlchemy's second Python asyncio dialect.
+
+Using a special asyncio mediation layer, the aiomysql dialect is usable
+as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>`
+extension package.
+
+This dialect should normally be used only with the
+:func:`_asyncio.create_async_engine` engine creation function::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine(
+        "mysql+aiomysql://user:pass@hostname/dbname?charset=utf8mb4"
+    )
+
+"""  # noqa
+from collections import deque
+
+from .pymysql import MySQLDialect_pymysql
+from ... import pool
+from ... import util
+from ...engine import AdaptedConnection
+from ...util.concurrency import asyncio
+from ...util.concurrency import await_fallback
+from ...util.concurrency import await_only
+
+
+class AsyncAdapt_aiomysql_cursor:
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    server_side = False
+    __slots__ = (
+        "_adapt_connection",
+        "_connection",
+        "await_",
+        "_cursor",
+        "_rows",
+    )
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor(adapt_connection.dbapi.Cursor)
+
+        # see https://github.com/aio-libs/aiomysql/issues/543
+        self._cursor = self.await_(cursor.__aenter__())
+        self._rows = deque()
+
+    @property
+    def description(self):
+        return self._cursor.description
+
+    @property
+    def rowcount(self):
+        return self._cursor.rowcount
+
+    @property
+    def arraysize(self):
+        return self._cursor.arraysize
+
+    @arraysize.setter
+    def arraysize(self, value):
+        self._cursor.arraysize = value
+
+    @property
+    def lastrowid(self):
+        return self._cursor.lastrowid
+
+    def close(self):
+        # note we aren't actually closing the cursor here,
+        # we are just letting GC do it.   to allow this to be async
+        # we would need the Result to change how it does "Safe close cursor".
+        # MySQL "cursors" don't actually have state to be "closed" besides
+        # exhausting rows, which we already have done for sync cursor.
+        # another option would be to emulate aiosqlite dialect and assign
+        # cursor only if we are doing server side cursor operation.
+        self._rows.clear()
+
+    def execute(self, operation, parameters=None):
+        return self.await_(self._execute_async(operation, parameters))
+
+    def executemany(self, operation, seq_of_parameters):
+        return self.await_(
+            self._executemany_async(operation, seq_of_parameters)
+        )
+
+    async def _execute_async(self, operation, parameters):
+        async with self._adapt_connection._execute_mutex:
+            result = await self._cursor.execute(operation, parameters)
+
+            if not self.server_side:
+                # aiomysql has a "fake" async result, so we have to pull it out
+                # of that here since our default result is not async.
+                # we could just as easily grab "_rows" here and be done with it
+                # but this is safer.
+                self._rows = deque(await self._cursor.fetchall())
+            return result
+
+    async def _executemany_async(self, operation, seq_of_parameters):
+        async with self._adapt_connection._execute_mutex:
+            return await self._cursor.executemany(operation, seq_of_parameters)
+
+    def setinputsizes(self, *inputsizes):
+        pass
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_aiomysql_ss_cursor(AsyncAdapt_aiomysql_cursor):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    __slots__ = ()
+    server_side = True
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor(adapt_connection.dbapi.SSCursor)
+
+        self._cursor = self.await_(cursor.__aenter__())
+
+    def close(self):
+        if self._cursor is not None:
+            self.await_(self._cursor.close())
+            self._cursor = None
+
+    def fetchone(self):
+        return self.await_(self._cursor.fetchone())
+
+    def fetchmany(self, size=None):
+        return self.await_(self._cursor.fetchmany(size=size))
+
+    def fetchall(self):
+        return self.await_(self._cursor.fetchall())
+
+
+class AsyncAdapt_aiomysql_connection(AdaptedConnection):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    await_ = staticmethod(await_only)
+    __slots__ = ("dbapi", "_execute_mutex")
+
+    def __init__(self, dbapi, connection):
+        self.dbapi = dbapi
+        self._connection = connection
+        self._execute_mutex = asyncio.Lock()
+
+    def ping(self, reconnect):
+        return self.await_(self._connection.ping(reconnect))
+
+    def character_set_name(self):
+        return self._connection.character_set_name()
+
+    def autocommit(self, value):
+        self.await_(self._connection.autocommit(value))
+
+    def cursor(self, server_side=False):
+        if server_side:
+            return AsyncAdapt_aiomysql_ss_cursor(self)
+        else:
+            return AsyncAdapt_aiomysql_cursor(self)
+
+    def rollback(self):
+        self.await_(self._connection.rollback())
+
+    def commit(self):
+        self.await_(self._connection.commit())
+
+    def terminate(self):
+        # it's not awaitable.
+        self._connection.close()
+
+    def close(self) -> None:
+        self.await_(self._connection.ensure_closed())
+
+
+class AsyncAdaptFallback_aiomysql_connection(AsyncAdapt_aiomysql_connection):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    __slots__ = ()
+
+    await_ = staticmethod(await_fallback)
+
+
+class AsyncAdapt_aiomysql_dbapi:
+    def __init__(self, aiomysql, pymysql):
+        self.aiomysql = aiomysql
+        self.pymysql = pymysql
+        self.paramstyle = "format"
+        self._init_dbapi_attributes()
+        self.Cursor, self.SSCursor = self._init_cursors_subclasses()
+
+    def _init_dbapi_attributes(self):
+        for name in (
+            "Warning",
+            "Error",
+            "InterfaceError",
+            "DataError",
+            "DatabaseError",
+            "OperationalError",
+            "InterfaceError",
+            "IntegrityError",
+            "ProgrammingError",
+            "InternalError",
+            "NotSupportedError",
+        ):
+            setattr(self, name, getattr(self.aiomysql, name))
+
+        for name in (
+            "NUMBER",
+            "STRING",
+            "DATETIME",
+            "BINARY",
+            "TIMESTAMP",
+            "Binary",
+        ):
+            setattr(self, name, getattr(self.pymysql, name))
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop("async_creator_fn", self.aiomysql.connect)
+
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_aiomysql_connection(
+                self,
+                await_fallback(creator_fn(*arg, **kw)),
+            )
+        else:
+            return AsyncAdapt_aiomysql_connection(
+                self,
+                await_only(creator_fn(*arg, **kw)),
+            )
+
+    def _init_cursors_subclasses(self):
+        # suppress unconditional warning emitted by aiomysql
+        class Cursor(self.aiomysql.Cursor):
+            async def _show_warnings(self, conn):
+                pass
+
+        class SSCursor(self.aiomysql.SSCursor):
+            async def _show_warnings(self, conn):
+                pass
+
+        return Cursor, SSCursor
+
+
+class MySQLDialect_aiomysql(MySQLDialect_pymysql):
+    driver = "aiomysql"
+    supports_statement_cache = True
+
+    supports_server_side_cursors = True
+    _sscursor = AsyncAdapt_aiomysql_ss_cursor
+
+    is_async = True
+    has_terminate = True
+
+    @classmethod
+    def import_dbapi(cls):
+        return AsyncAdapt_aiomysql_dbapi(
+            __import__("aiomysql"), __import__("pymysql")
+        )
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if util.asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def do_terminate(self, dbapi_connection) -> None:
+        dbapi_connection.terminate()
+
+    def create_connect_args(self, url):
+        return super().create_connect_args(
+            url, _translate_args=dict(username="user", database="db")
+        )
+
+    def is_disconnect(self, e, connection, cursor):
+        if super().is_disconnect(e, connection, cursor):
+            return True
+        else:
+            str_e = str(e).lower()
+            return "not connected" in str_e
+
+    def _found_rows_client_flag(self):
+        from pymysql.constants import CLIENT
+
+        return CLIENT.FOUND_ROWS
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = MySQLDialect_aiomysql
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/asyncmy.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/asyncmy.py
new file mode 100644
index 00000000..9ec54e69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/asyncmy.py
@@ -0,0 +1,339 @@
+# dialects/mysql/asyncmy.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors <see AUTHORS
+# file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: mysql+asyncmy
+    :name: asyncmy
+    :dbapi: asyncmy
+    :connectstring: mysql+asyncmy://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://github.com/long2ice/asyncmy
+
+Using a special asyncio mediation layer, the asyncmy dialect is usable
+as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>`
+extension package.
+
+This dialect should normally be used only with the
+:func:`_asyncio.create_async_engine` engine creation function::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine(
+        "mysql+asyncmy://user:pass@hostname/dbname?charset=utf8mb4"
+    )
+
+"""  # noqa
+from collections import deque
+from contextlib import asynccontextmanager
+
+from .pymysql import MySQLDialect_pymysql
+from ... import pool
+from ... import util
+from ...engine import AdaptedConnection
+from ...util.concurrency import asyncio
+from ...util.concurrency import await_fallback
+from ...util.concurrency import await_only
+
+
+class AsyncAdapt_asyncmy_cursor:
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    server_side = False
+    __slots__ = (
+        "_adapt_connection",
+        "_connection",
+        "await_",
+        "_cursor",
+        "_rows",
+    )
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor()
+
+        self._cursor = self.await_(cursor.__aenter__())
+        self._rows = deque()
+
+    @property
+    def description(self):
+        return self._cursor.description
+
+    @property
+    def rowcount(self):
+        return self._cursor.rowcount
+
+    @property
+    def arraysize(self):
+        return self._cursor.arraysize
+
+    @arraysize.setter
+    def arraysize(self, value):
+        self._cursor.arraysize = value
+
+    @property
+    def lastrowid(self):
+        return self._cursor.lastrowid
+
+    def close(self):
+        # note we aren't actually closing the cursor here,
+        # we are just letting GC do it.   to allow this to be async
+        # we would need the Result to change how it does "Safe close cursor".
+        # MySQL "cursors" don't actually have state to be "closed" besides
+        # exhausting rows, which we already have done for sync cursor.
+        # another option would be to emulate aiosqlite dialect and assign
+        # cursor only if we are doing server side cursor operation.
+        self._rows.clear()
+
+    def execute(self, operation, parameters=None):
+        return self.await_(self._execute_async(operation, parameters))
+
+    def executemany(self, operation, seq_of_parameters):
+        return self.await_(
+            self._executemany_async(operation, seq_of_parameters)
+        )
+
+    async def _execute_async(self, operation, parameters):
+        async with self._adapt_connection._mutex_and_adapt_errors():
+            if parameters is None:
+                result = await self._cursor.execute(operation)
+            else:
+                result = await self._cursor.execute(operation, parameters)
+
+            if not self.server_side:
+                # asyncmy has a "fake" async result, so we have to pull it out
+                # of that here since our default result is not async.
+                # we could just as easily grab "_rows" here and be done with it
+                # but this is safer.
+                self._rows = deque(await self._cursor.fetchall())
+            return result
+
+    async def _executemany_async(self, operation, seq_of_parameters):
+        async with self._adapt_connection._mutex_and_adapt_errors():
+            return await self._cursor.executemany(operation, seq_of_parameters)
+
+    def setinputsizes(self, *inputsizes):
+        pass
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_asyncmy_ss_cursor(AsyncAdapt_asyncmy_cursor):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    __slots__ = ()
+    server_side = True
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+
+        cursor = self._connection.cursor(
+            adapt_connection.dbapi.asyncmy.cursors.SSCursor
+        )
+
+        self._cursor = self.await_(cursor.__aenter__())
+
+    def close(self):
+        if self._cursor is not None:
+            self.await_(self._cursor.close())
+            self._cursor = None
+
+    def fetchone(self):
+        return self.await_(self._cursor.fetchone())
+
+    def fetchmany(self, size=None):
+        return self.await_(self._cursor.fetchmany(size=size))
+
+    def fetchall(self):
+        return self.await_(self._cursor.fetchall())
+
+
+class AsyncAdapt_asyncmy_connection(AdaptedConnection):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    await_ = staticmethod(await_only)
+    __slots__ = ("dbapi", "_execute_mutex")
+
+    def __init__(self, dbapi, connection):
+        self.dbapi = dbapi
+        self._connection = connection
+        self._execute_mutex = asyncio.Lock()
+
+    @asynccontextmanager
+    async def _mutex_and_adapt_errors(self):
+        async with self._execute_mutex:
+            try:
+                yield
+            except AttributeError:
+                raise self.dbapi.InternalError(
+                    "network operation failed due to asyncmy attribute error"
+                )
+
+    def ping(self, reconnect):
+        assert not reconnect
+        return self.await_(self._do_ping())
+
+    async def _do_ping(self):
+        async with self._mutex_and_adapt_errors():
+            return await self._connection.ping(False)
+
+    def character_set_name(self):
+        return self._connection.character_set_name()
+
+    def autocommit(self, value):
+        self.await_(self._connection.autocommit(value))
+
+    def cursor(self, server_side=False):
+        if server_side:
+            return AsyncAdapt_asyncmy_ss_cursor(self)
+        else:
+            return AsyncAdapt_asyncmy_cursor(self)
+
+    def rollback(self):
+        self.await_(self._connection.rollback())
+
+    def commit(self):
+        self.await_(self._connection.commit())
+
+    def terminate(self):
+        # it's not awaitable.
+        self._connection.close()
+
+    def close(self) -> None:
+        self.await_(self._connection.ensure_closed())
+
+
+class AsyncAdaptFallback_asyncmy_connection(AsyncAdapt_asyncmy_connection):
+    __slots__ = ()
+
+    await_ = staticmethod(await_fallback)
+
+
+def _Binary(x):
+    """Return x as a binary type."""
+    return bytes(x)
+
+
+class AsyncAdapt_asyncmy_dbapi:
+    def __init__(self, asyncmy):
+        self.asyncmy = asyncmy
+        self.paramstyle = "format"
+        self._init_dbapi_attributes()
+
+    def _init_dbapi_attributes(self):
+        for name in (
+            "Warning",
+            "Error",
+            "InterfaceError",
+            "DataError",
+            "DatabaseError",
+            "OperationalError",
+            "InterfaceError",
+            "IntegrityError",
+            "ProgrammingError",
+            "InternalError",
+            "NotSupportedError",
+        ):
+            setattr(self, name, getattr(self.asyncmy.errors, name))
+
+    STRING = util.symbol("STRING")
+    NUMBER = util.symbol("NUMBER")
+    BINARY = util.symbol("BINARY")
+    DATETIME = util.symbol("DATETIME")
+    TIMESTAMP = util.symbol("TIMESTAMP")
+    Binary = staticmethod(_Binary)
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop("async_creator_fn", self.asyncmy.connect)
+
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_asyncmy_connection(
+                self,
+                await_fallback(creator_fn(*arg, **kw)),
+            )
+        else:
+            return AsyncAdapt_asyncmy_connection(
+                self,
+                await_only(creator_fn(*arg, **kw)),
+            )
+
+
+class MySQLDialect_asyncmy(MySQLDialect_pymysql):
+    driver = "asyncmy"
+    supports_statement_cache = True
+
+    supports_server_side_cursors = True
+    _sscursor = AsyncAdapt_asyncmy_ss_cursor
+
+    is_async = True
+    has_terminate = True
+
+    @classmethod
+    def import_dbapi(cls):
+        return AsyncAdapt_asyncmy_dbapi(__import__("asyncmy"))
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if util.asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def do_terminate(self, dbapi_connection) -> None:
+        dbapi_connection.terminate()
+
+    def create_connect_args(self, url):
+        return super().create_connect_args(
+            url, _translate_args=dict(username="user", database="db")
+        )
+
+    def is_disconnect(self, e, connection, cursor):
+        if super().is_disconnect(e, connection, cursor):
+            return True
+        else:
+            str_e = str(e).lower()
+            return (
+                "not connected" in str_e or "network operation failed" in str_e
+            )
+
+    def _found_rows_client_flag(self):
+        from asyncmy.constants import CLIENT
+
+        return CLIENT.FOUND_ROWS
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = MySQLDialect_asyncmy
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/base.py
new file mode 100644
index 00000000..1314ee3d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/base.py
@@ -0,0 +1,3550 @@
+# dialects/mysql/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+
+.. dialect:: mysql
+    :name: MySQL / MariaDB
+    :normal_support: 5.6+ / 10+
+    :best_effort: 5.0.2+ / 5.0.2+
+
+Supported Versions and Features
+-------------------------------
+
+SQLAlchemy supports MySQL starting with version 5.0.2 through modern releases,
+as well as all modern versions of MariaDB.   See the official MySQL
+documentation for detailed information about features supported in any given
+server release.
+
+.. versionchanged:: 1.4  minimum MySQL version supported is now 5.0.2.
+
+MariaDB Support
+~~~~~~~~~~~~~~~
+
+The MariaDB variant of MySQL retains fundamental compatibility with MySQL's
+protocols however the development of these two products continues to diverge.
+Within the realm of SQLAlchemy, the two databases have a small number of
+syntactical and behavioral differences that SQLAlchemy accommodates automatically.
+To connect to a MariaDB database, no changes to the database URL are required::
+
+
+    engine = create_engine(
+        "mysql+pymysql://user:pass@some_mariadb/dbname?charset=utf8mb4"
+    )
+
+Upon first connect, the SQLAlchemy dialect employs a
+server version detection scheme that determines if the
+backing database reports as MariaDB.  Based on this flag, the dialect
+can make different choices in those of areas where its behavior
+must be different.
+
+.. _mysql_mariadb_only_mode:
+
+MariaDB-Only Mode
+~~~~~~~~~~~~~~~~~
+
+The dialect also supports an **optional** "MariaDB-only" mode of connection, which may be
+useful for the case where an application makes use of MariaDB-specific features
+and is not compatible with a MySQL database.    To use this mode of operation,
+replace the "mysql" token in the above URL with "mariadb"::
+
+    engine = create_engine(
+        "mariadb+pymysql://user:pass@some_mariadb/dbname?charset=utf8mb4"
+    )
+
+The above engine, upon first connect, will raise an error if the server version
+detection detects that the backing database is not MariaDB.
+
+When using an engine with ``"mariadb"`` as the dialect name, **all mysql-specific options
+that include the name "mysql" in them are now named with "mariadb"**.  This means
+options like ``mysql_engine`` should be named ``mariadb_engine``, etc.  Both
+"mysql" and "mariadb" options can be used simultaneously for applications that
+use URLs with both "mysql" and "mariadb" dialects::
+
+    my_table = Table(
+        "mytable",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("textdata", String(50)),
+        mariadb_engine="InnoDB",
+        mysql_engine="InnoDB",
+    )
+
+    Index(
+        "textdata_ix",
+        my_table.c.textdata,
+        mysql_prefix="FULLTEXT",
+        mariadb_prefix="FULLTEXT",
+    )
+
+Similar behavior will occur when the above structures are reflected, i.e. the
+"mariadb" prefix will be present in the option names when the database URL
+is based on the "mariadb" name.
+
+.. versionadded:: 1.4 Added "mariadb" dialect name supporting "MariaDB-only mode"
+   for the MySQL dialect.
+
+.. _mysql_connection_timeouts:
+
+Connection Timeouts and Disconnects
+-----------------------------------
+
+MySQL / MariaDB feature an automatic connection close behavior, for connections that
+have been idle for a fixed period of time, defaulting to eight hours.
+To circumvent having this issue, use
+the :paramref:`_sa.create_engine.pool_recycle` option which ensures that
+a connection will be discarded and replaced with a new one if it has been
+present in the pool for a fixed number of seconds::
+
+    engine = create_engine("mysql+mysqldb://...", pool_recycle=3600)
+
+For more comprehensive disconnect detection of pooled connections, including
+accommodation of  server restarts and network issues, a pre-ping approach may
+be employed.  See :ref:`pool_disconnects` for current approaches.
+
+.. seealso::
+
+    :ref:`pool_disconnects` - Background on several techniques for dealing
+    with timed out connections as well as database restarts.
+
+.. _mysql_storage_engines:
+
+CREATE TABLE arguments including Storage Engines
+------------------------------------------------
+
+Both MySQL's and MariaDB's CREATE TABLE syntax includes a wide array of special options,
+including ``ENGINE``, ``CHARSET``, ``MAX_ROWS``, ``ROW_FORMAT``,
+``INSERT_METHOD``, and many more.
+To accommodate the rendering of these arguments, specify the form
+``mysql_argument_name="value"``.  For example, to specify a table with
+``ENGINE`` of ``InnoDB``, ``CHARSET`` of ``utf8mb4``, and ``KEY_BLOCK_SIZE``
+of ``1024``::
+
+  Table(
+      "mytable",
+      metadata,
+      Column("data", String(32)),
+      mysql_engine="InnoDB",
+      mysql_charset="utf8mb4",
+      mysql_key_block_size="1024",
+  )
+
+When supporting :ref:`mysql_mariadb_only_mode` mode, similar keys against
+the "mariadb" prefix must be included as well.  The values can of course
+vary independently so that different settings on MySQL vs. MariaDB may
+be maintained::
+
+  # support both "mysql" and "mariadb-only" engine URLs
+
+  Table(
+      "mytable",
+      metadata,
+      Column("data", String(32)),
+      mysql_engine="InnoDB",
+      mariadb_engine="InnoDB",
+      mysql_charset="utf8mb4",
+      mariadb_charset="utf8",
+      mysql_key_block_size="1024",
+      mariadb_key_block_size="1024",
+  )
+
+The MySQL / MariaDB dialects will normally transfer any keyword specified as
+``mysql_keyword_name`` to be rendered as ``KEYWORD_NAME`` in the
+``CREATE TABLE`` statement.  A handful of these names will render with a space
+instead of an underscore; to support this, the MySQL dialect has awareness of
+these particular names, which include ``DATA DIRECTORY``
+(e.g. ``mysql_data_directory``), ``CHARACTER SET`` (e.g.
+``mysql_character_set``) and ``INDEX DIRECTORY`` (e.g.
+``mysql_index_directory``).
+
+The most common argument is ``mysql_engine``, which refers to the storage
+engine for the table.  Historically, MySQL server installations would default
+to ``MyISAM`` for this value, although newer versions may be defaulting
+to ``InnoDB``.  The ``InnoDB`` engine is typically preferred for its support
+of transactions and foreign keys.
+
+A :class:`_schema.Table`
+that is created in a MySQL / MariaDB database with a storage engine
+of ``MyISAM`` will be essentially non-transactional, meaning any
+INSERT/UPDATE/DELETE statement referring to this table will be invoked as
+autocommit.   It also will have no support for foreign key constraints; while
+the ``CREATE TABLE`` statement accepts foreign key options, when using the
+``MyISAM`` storage engine these arguments are discarded.  Reflecting such a
+table will also produce no foreign key constraint information.
+
+For fully atomic transactions as well as support for foreign key
+constraints, all participating ``CREATE TABLE`` statements must specify a
+transactional engine, which in the vast majority of cases is ``InnoDB``.
+
+
+Case Sensitivity and Table Reflection
+-------------------------------------
+
+Both MySQL and MariaDB have inconsistent support for case-sensitive identifier
+names, basing support on specific details of the underlying
+operating system. However, it has been observed that no matter
+what case sensitivity behavior is present, the names of tables in
+foreign key declarations are *always* received from the database
+as all-lower case, making it impossible to accurately reflect a
+schema where inter-related tables use mixed-case identifier names.
+
+Therefore it is strongly advised that table names be declared as
+all lower case both within SQLAlchemy as well as on the MySQL / MariaDB
+database itself, especially if database reflection features are
+to be used.
+
+.. _mysql_isolation_level:
+
+Transaction Isolation Level
+---------------------------
+
+All MySQL / MariaDB dialects support setting of transaction isolation level both via a
+dialect-specific parameter :paramref:`_sa.create_engine.isolation_level`
+accepted
+by :func:`_sa.create_engine`, as well as the
+:paramref:`.Connection.execution_options.isolation_level` argument as passed to
+:meth:`_engine.Connection.execution_options`.
+This feature works by issuing the
+command ``SET SESSION TRANSACTION ISOLATION LEVEL <level>`` for each new
+connection.  For the special AUTOCOMMIT isolation level, DBAPI-specific
+techniques are used.
+
+To set isolation level using :func:`_sa.create_engine`::
+
+    engine = create_engine(
+        "mysql+mysqldb://scott:tiger@localhost/test",
+        isolation_level="READ UNCOMMITTED",
+    )
+
+To set using per-connection execution options::
+
+    connection = engine.connect()
+    connection = connection.execution_options(isolation_level="READ COMMITTED")
+
+Valid values for ``isolation_level`` include:
+
+* ``READ COMMITTED``
+* ``READ UNCOMMITTED``
+* ``REPEATABLE READ``
+* ``SERIALIZABLE``
+* ``AUTOCOMMIT``
+
+The special ``AUTOCOMMIT`` value makes use of the various "autocommit"
+attributes provided by specific DBAPIs, and is currently supported by
+MySQLdb, MySQL-Client, MySQL-Connector Python, and PyMySQL.   Using it,
+the database connection will return true for the value of
+``SELECT @@autocommit;``.
+
+There are also more options for isolation level configurations, such as
+"sub-engine" objects linked to a main :class:`_engine.Engine` which each apply
+different isolation level settings.  See the discussion at
+:ref:`dbapi_autocommit` for background.
+
+.. seealso::
+
+    :ref:`dbapi_autocommit`
+
+AUTO_INCREMENT Behavior
+-----------------------
+
+When creating tables, SQLAlchemy will automatically set ``AUTO_INCREMENT`` on
+the first :class:`.Integer` primary key column which is not marked as a
+foreign key::
+
+  >>> t = Table(
+  ...     "mytable", metadata, Column("mytable_id", Integer, primary_key=True)
+  ... )
+  >>> t.create()
+  CREATE TABLE mytable (
+          id INTEGER NOT NULL AUTO_INCREMENT,
+          PRIMARY KEY (id)
+  )
+
+You can disable this behavior by passing ``False`` to the
+:paramref:`_schema.Column.autoincrement` argument of :class:`_schema.Column`.
+This flag
+can also be used to enable auto-increment on a secondary column in a
+multi-column key for some storage engines::
+
+  Table(
+      "mytable",
+      metadata,
+      Column("gid", Integer, primary_key=True, autoincrement=False),
+      Column("id", Integer, primary_key=True),
+  )
+
+.. _mysql_ss_cursors:
+
+Server Side Cursors
+-------------------
+
+Server-side cursor support is available for the mysqlclient, PyMySQL,
+mariadbconnector dialects and may also be available in others.   This makes use
+of either the "buffered=True/False" flag if available or by using a class such
+as ``MySQLdb.cursors.SSCursor`` or ``pymysql.cursors.SSCursor`` internally.
+
+
+Server side cursors are enabled on a per-statement basis by using the
+:paramref:`.Connection.execution_options.stream_results` connection execution
+option::
+
+    with engine.connect() as conn:
+        result = conn.execution_options(stream_results=True).execute(
+            text("select * from table")
+        )
+
+Note that some kinds of SQL statements may not be supported with
+server side cursors; generally, only SQL statements that return rows should be
+used with this option.
+
+.. deprecated:: 1.4  The dialect-level server_side_cursors flag is deprecated
+   and will be removed in a future release.  Please use the
+   :paramref:`_engine.Connection.stream_results` execution option for
+   unbuffered cursor support.
+
+.. seealso::
+
+    :ref:`engine_stream_results`
+
+.. _mysql_unicode:
+
+Unicode
+-------
+
+Charset Selection
+~~~~~~~~~~~~~~~~~
+
+Most MySQL / MariaDB DBAPIs offer the option to set the client character set for
+a connection.   This is typically delivered using the ``charset`` parameter
+in the URL, such as::
+
+    e = create_engine(
+        "mysql+pymysql://scott:tiger@localhost/test?charset=utf8mb4"
+    )
+
+This charset is the **client character set** for the connection.  Some
+MySQL DBAPIs will default this to a value such as ``latin1``, and some
+will make use of the ``default-character-set`` setting in the ``my.cnf``
+file as well.   Documentation for the DBAPI in use should be consulted
+for specific behavior.
+
+The encoding used for Unicode has traditionally been ``'utf8'``.  However, for
+MySQL versions 5.5.3 and MariaDB 5.5 on forward, a new MySQL-specific encoding
+``'utf8mb4'`` has been introduced, and as of MySQL 8.0 a warning is emitted by
+the server if plain ``utf8`` is specified within any server-side directives,
+replaced with ``utf8mb3``.  The rationale for this new encoding is due to the
+fact that MySQL's legacy utf-8 encoding only supports codepoints up to three
+bytes instead of four.  Therefore, when communicating with a MySQL or MariaDB
+database that includes codepoints more than three bytes in size, this new
+charset is preferred, if supported by both the database as well as the client
+DBAPI, as in::
+
+    e = create_engine(
+        "mysql+pymysql://scott:tiger@localhost/test?charset=utf8mb4"
+    )
+
+All modern DBAPIs should support the ``utf8mb4`` charset.
+
+In order to use ``utf8mb4`` encoding for a schema that was created with  legacy
+``utf8``, changes to the MySQL/MariaDB schema and/or server configuration may be
+required.
+
+.. seealso::
+
+    `The utf8mb4 Character Set \
+    <https://dev.mysql.com/doc/refman/5.5/en/charset-unicode-utf8mb4.html>`_ - \
+    in the MySQL documentation
+
+.. _mysql_binary_introducer:
+
+Dealing with Binary Data Warnings and Unicode
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+MySQL versions 5.6, 5.7 and later (not MariaDB at the time of this writing) now
+emit a warning when attempting to pass binary data to the database, while a
+character set encoding is also in place, when the binary data itself is not
+valid for that encoding:
+
+.. sourcecode:: text
+
+    default.py:509: Warning: (1300, "Invalid utf8mb4 character string:
+    'F9876A'")
+      cursor.execute(statement, parameters)
+
+This warning is due to the fact that the MySQL client library is attempting to
+interpret the binary string as a unicode object even if a datatype such
+as :class:`.LargeBinary` is in use.   To resolve this, the SQL statement requires
+a binary "character set introducer" be present before any non-NULL value
+that renders like this:
+
+.. sourcecode:: sql
+
+    INSERT INTO table (data) VALUES (_binary %s)
+
+These character set introducers are provided by the DBAPI driver, assuming the
+use of mysqlclient or PyMySQL (both of which are recommended).  Add the query
+string parameter ``binary_prefix=true`` to the URL to repair this warning::
+
+    # mysqlclient
+    engine = create_engine(
+        "mysql+mysqldb://scott:tiger@localhost/test?charset=utf8mb4&binary_prefix=true"
+    )
+
+    # PyMySQL
+    engine = create_engine(
+        "mysql+pymysql://scott:tiger@localhost/test?charset=utf8mb4&binary_prefix=true"
+    )
+
+The ``binary_prefix`` flag may or may not be supported by other MySQL drivers.
+
+SQLAlchemy itself cannot render this ``_binary`` prefix reliably, as it does
+not work with the NULL value, which is valid to be sent as a bound parameter.
+As the MySQL driver renders parameters directly into the SQL string, it's the
+most efficient place for this additional keyword to be passed.
+
+.. seealso::
+
+    `Character set introducers <https://dev.mysql.com/doc/refman/5.7/en/charset-introducer.html>`_ - on the MySQL website
+
+
+ANSI Quoting Style
+------------------
+
+MySQL / MariaDB feature two varieties of identifier "quoting style", one using
+backticks and the other using quotes, e.g. ```some_identifier```  vs.
+``"some_identifier"``.   All MySQL dialects detect which version
+is in use by checking the value of :ref:`sql_mode<mysql_sql_mode>` when a connection is first
+established with a particular :class:`_engine.Engine`.
+This quoting style comes
+into play when rendering table and column names as well as when reflecting
+existing database structures.  The detection is entirely automatic and
+no special configuration is needed to use either quoting style.
+
+
+.. _mysql_sql_mode:
+
+Changing the sql_mode
+---------------------
+
+MySQL supports operating in multiple
+`Server SQL Modes <https://dev.mysql.com/doc/refman/8.0/en/sql-mode.html>`_  for
+both Servers and Clients. To change the ``sql_mode`` for a given application, a
+developer can leverage SQLAlchemy's Events system.
+
+In the following example, the event system is used to set the ``sql_mode`` on
+the ``first_connect`` and ``connect`` events::
+
+    from sqlalchemy import create_engine, event
+
+    eng = create_engine(
+        "mysql+mysqldb://scott:tiger@localhost/test", echo="debug"
+    )
+
+
+    # `insert=True` will ensure this is the very first listener to run
+    @event.listens_for(eng, "connect", insert=True)
+    def connect(dbapi_connection, connection_record):
+        cursor = dbapi_connection.cursor()
+        cursor.execute("SET sql_mode = 'STRICT_ALL_TABLES'")
+
+
+    conn = eng.connect()
+
+In the example illustrated above, the "connect" event will invoke the "SET"
+statement on the connection at the moment a particular DBAPI connection is
+first created for a given Pool, before the connection is made available to the
+connection pool.  Additionally, because the function was registered with
+``insert=True``, it will be prepended to the internal list of registered
+functions.
+
+
+MySQL / MariaDB SQL Extensions
+------------------------------
+
+Many of the MySQL / MariaDB SQL extensions are handled through SQLAlchemy's generic
+function and operator support::
+
+  table.select(table.c.password == func.md5("plaintext"))
+  table.select(table.c.username.op("regexp")("^[a-d]"))
+
+And of course any valid SQL statement can be executed as a string as well.
+
+Some limited direct support for MySQL / MariaDB extensions to SQL is currently
+available.
+
+* INSERT..ON DUPLICATE KEY UPDATE:  See
+  :ref:`mysql_insert_on_duplicate_key_update`
+
+* SELECT pragma, use :meth:`_expression.Select.prefix_with` and
+  :meth:`_query.Query.prefix_with`::
+
+    select(...).prefix_with(["HIGH_PRIORITY", "SQL_SMALL_RESULT"])
+
+* UPDATE with LIMIT::
+
+    update(...).with_dialect_options(mysql_limit=10, mariadb_limit=10)
+
+* DELETE
+  with LIMIT::
+
+    delete(...).with_dialect_options(mysql_limit=10, mariadb_limit=10)
+
+  .. versionadded:: 2.0.37 Added delete with limit
+
+* optimizer hints, use :meth:`_expression.Select.prefix_with` and
+  :meth:`_query.Query.prefix_with`::
+
+    select(...).prefix_with("/*+ NO_RANGE_OPTIMIZATION(t4 PRIMARY) */")
+
+* index hints, use :meth:`_expression.Select.with_hint` and
+  :meth:`_query.Query.with_hint`::
+
+    select(...).with_hint(some_table, "USE INDEX xyz")
+
+* MATCH
+  operator support::
+
+        from sqlalchemy.dialects.mysql import match
+
+        select(...).where(match(col1, col2, against="some expr").in_boolean_mode())
+
+  .. seealso::
+
+    :class:`_mysql.match`
+
+INSERT/DELETE...RETURNING
+-------------------------
+
+The MariaDB dialect supports 10.5+'s ``INSERT..RETURNING`` and
+``DELETE..RETURNING`` (10.0+) syntaxes.   ``INSERT..RETURNING`` may be used
+automatically in some cases in order to fetch newly generated identifiers in
+place of the traditional approach of using ``cursor.lastrowid``, however
+``cursor.lastrowid`` is currently still preferred for simple single-statement
+cases for its better performance.
+
+To specify an explicit ``RETURNING`` clause, use the
+:meth:`._UpdateBase.returning` method on a per-statement basis::
+
+    # INSERT..RETURNING
+    result = connection.execute(
+        table.insert().values(name="foo").returning(table.c.col1, table.c.col2)
+    )
+    print(result.all())
+
+    # DELETE..RETURNING
+    result = connection.execute(
+        table.delete()
+        .where(table.c.name == "foo")
+        .returning(table.c.col1, table.c.col2)
+    )
+    print(result.all())
+
+.. versionadded:: 2.0  Added support for MariaDB RETURNING
+
+.. _mysql_insert_on_duplicate_key_update:
+
+INSERT...ON DUPLICATE KEY UPDATE (Upsert)
+------------------------------------------
+
+MySQL / MariaDB allow "upserts" (update or insert)
+of rows into a table via the ``ON DUPLICATE KEY UPDATE`` clause of the
+``INSERT`` statement.  A candidate row will only be inserted if that row does
+not match an existing primary or unique key in the table; otherwise, an UPDATE
+will be performed.   The statement allows for separate specification of the
+values to INSERT versus the values for UPDATE.
+
+SQLAlchemy provides ``ON DUPLICATE KEY UPDATE`` support via the MySQL-specific
+:func:`.mysql.insert()` function, which provides
+the generative method :meth:`~.mysql.Insert.on_duplicate_key_update`:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy.dialects.mysql import insert
+
+    >>> insert_stmt = insert(my_table).values(
+    ...     id="some_existing_id", data="inserted value"
+    ... )
+
+    >>> on_duplicate_key_stmt = insert_stmt.on_duplicate_key_update(
+    ...     data=insert_stmt.inserted.data, status="U"
+    ... )
+    >>> print(on_duplicate_key_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%s, %s)
+    ON DUPLICATE KEY UPDATE data = VALUES(data), status = %s
+
+
+Unlike PostgreSQL's "ON CONFLICT" phrase, the "ON DUPLICATE KEY UPDATE"
+phrase will always match on any primary key or unique key, and will always
+perform an UPDATE if there's a match; there are no options for it to raise
+an error or to skip performing an UPDATE.
+
+``ON DUPLICATE KEY UPDATE`` is used to perform an update of the already
+existing row, using any combination of new values as well as values
+from the proposed insertion.   These values are normally specified using
+keyword arguments passed to the
+:meth:`_mysql.Insert.on_duplicate_key_update`
+given column key values (usually the name of the column, unless it
+specifies :paramref:`_schema.Column.key`
+) as keys and literal or SQL expressions
+as values:
+
+.. sourcecode:: pycon+sql
+
+    >>> insert_stmt = insert(my_table).values(
+    ...     id="some_existing_id", data="inserted value"
+    ... )
+
+    >>> on_duplicate_key_stmt = insert_stmt.on_duplicate_key_update(
+    ...     data="some data",
+    ...     updated_at=func.current_timestamp(),
+    ... )
+
+    >>> print(on_duplicate_key_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%s, %s)
+    ON DUPLICATE KEY UPDATE data = %s, updated_at = CURRENT_TIMESTAMP
+
+In a manner similar to that of :meth:`.UpdateBase.values`, other parameter
+forms are accepted, including a single dictionary:
+
+.. sourcecode:: pycon+sql
+
+    >>> on_duplicate_key_stmt = insert_stmt.on_duplicate_key_update(
+    ...     {"data": "some data", "updated_at": func.current_timestamp()},
+    ... )
+
+as well as a list of 2-tuples, which will automatically provide
+a parameter-ordered UPDATE statement in a manner similar to that described
+at :ref:`tutorial_parameter_ordered_updates`.  Unlike the :class:`_expression.Update`
+object,
+no special flag is needed to specify the intent since the argument form is
+this context is unambiguous:
+
+.. sourcecode:: pycon+sql
+
+    >>> on_duplicate_key_stmt = insert_stmt.on_duplicate_key_update(
+    ...     [
+    ...         ("data", "some data"),
+    ...         ("updated_at", func.current_timestamp()),
+    ...     ]
+    ... )
+
+    >>> print(on_duplicate_key_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%s, %s)
+    ON DUPLICATE KEY UPDATE data = %s, updated_at = CURRENT_TIMESTAMP
+
+.. versionchanged:: 1.3 support for parameter-ordered UPDATE clause within
+   MySQL ON DUPLICATE KEY UPDATE
+
+.. warning::
+
+    The :meth:`_mysql.Insert.on_duplicate_key_update`
+    method does **not** take into
+    account Python-side default UPDATE values or generation functions, e.g.
+    e.g. those specified using :paramref:`_schema.Column.onupdate`.
+    These values will not be exercised for an ON DUPLICATE KEY style of UPDATE,
+    unless they are manually specified explicitly in the parameters.
+
+
+
+In order to refer to the proposed insertion row, the special alias
+:attr:`_mysql.Insert.inserted` is available as an attribute on
+the :class:`_mysql.Insert` object; this object is a
+:class:`_expression.ColumnCollection` which contains all columns of the target
+table:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(
+    ...     id="some_id", data="inserted value", author="jlh"
+    ... )
+
+    >>> do_update_stmt = stmt.on_duplicate_key_update(
+    ...     data="updated value", author=stmt.inserted.author
+    ... )
+
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data, author) VALUES (%s, %s, %s)
+    ON DUPLICATE KEY UPDATE data = %s, author = VALUES(author)
+
+When rendered, the "inserted" namespace will produce the expression
+``VALUES(<columnname>)``.
+
+.. versionadded:: 1.2 Added support for MySQL ON DUPLICATE KEY UPDATE clause
+
+
+
+rowcount Support
+----------------
+
+SQLAlchemy standardizes the DBAPI ``cursor.rowcount`` attribute to be the
+usual definition of "number of rows matched by an UPDATE or DELETE" statement.
+This is in contradiction to the default setting on most MySQL DBAPI drivers,
+which is "number of rows actually modified/deleted".  For this reason, the
+SQLAlchemy MySQL dialects always add the ``constants.CLIENT.FOUND_ROWS``
+flag, or whatever is equivalent for the target dialect, upon connection.
+This setting is currently hardcoded.
+
+.. seealso::
+
+    :attr:`_engine.CursorResult.rowcount`
+
+
+.. _mysql_indexes:
+
+MySQL / MariaDB- Specific Index Options
+-----------------------------------------
+
+MySQL and MariaDB-specific extensions to the :class:`.Index` construct are available.
+
+Index Length
+~~~~~~~~~~~~~
+
+MySQL and MariaDB both provide an option to create index entries with a certain length, where
+"length" refers to the number of characters or bytes in each value which will
+become part of the index. SQLAlchemy provides this feature via the
+``mysql_length`` and/or ``mariadb_length`` parameters::
+
+    Index("my_index", my_table.c.data, mysql_length=10, mariadb_length=10)
+
+    Index("a_b_idx", my_table.c.a, my_table.c.b, mysql_length={"a": 4, "b": 9})
+
+    Index(
+        "a_b_idx", my_table.c.a, my_table.c.b, mariadb_length={"a": 4, "b": 9}
+    )
+
+Prefix lengths are given in characters for nonbinary string types and in bytes
+for binary string types. The value passed to the keyword argument *must* be
+either an integer (and, thus, specify the same prefix length value for all
+columns of the index) or a dict in which keys are column names and values are
+prefix length values for corresponding columns. MySQL and MariaDB only allow a
+length for a column of an index if it is for a CHAR, VARCHAR, TEXT, BINARY,
+VARBINARY and BLOB.
+
+Index Prefixes
+~~~~~~~~~~~~~~
+
+MySQL storage engines permit you to specify an index prefix when creating
+an index. SQLAlchemy provides this feature via the
+``mysql_prefix`` parameter on :class:`.Index`::
+
+    Index("my_index", my_table.c.data, mysql_prefix="FULLTEXT")
+
+The value passed to the keyword argument will be simply passed through to the
+underlying CREATE INDEX, so it *must* be a valid index prefix for your MySQL
+storage engine.
+
+.. seealso::
+
+    `CREATE INDEX <https://dev.mysql.com/doc/refman/5.0/en/create-index.html>`_ - MySQL documentation
+
+Index Types
+~~~~~~~~~~~~~
+
+Some MySQL storage engines permit you to specify an index type when creating
+an index or primary key constraint. SQLAlchemy provides this feature via the
+``mysql_using`` parameter on :class:`.Index`::
+
+    Index(
+        "my_index", my_table.c.data, mysql_using="hash", mariadb_using="hash"
+    )
+
+As well as the ``mysql_using`` parameter on :class:`.PrimaryKeyConstraint`::
+
+    PrimaryKeyConstraint("data", mysql_using="hash", mariadb_using="hash")
+
+The value passed to the keyword argument will be simply passed through to the
+underlying CREATE INDEX or PRIMARY KEY clause, so it *must* be a valid index
+type for your MySQL storage engine.
+
+More information can be found at:
+
+https://dev.mysql.com/doc/refman/5.0/en/create-index.html
+
+https://dev.mysql.com/doc/refman/5.0/en/create-table.html
+
+Index Parsers
+~~~~~~~~~~~~~
+
+CREATE FULLTEXT INDEX in MySQL also supports a "WITH PARSER" option.  This
+is available using the keyword argument ``mysql_with_parser``::
+
+    Index(
+        "my_index",
+        my_table.c.data,
+        mysql_prefix="FULLTEXT",
+        mysql_with_parser="ngram",
+        mariadb_prefix="FULLTEXT",
+        mariadb_with_parser="ngram",
+    )
+
+.. versionadded:: 1.3
+
+
+.. _mysql_foreign_keys:
+
+MySQL / MariaDB Foreign Keys
+-----------------------------
+
+MySQL and MariaDB's behavior regarding foreign keys has some important caveats.
+
+Foreign Key Arguments to Avoid
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Neither MySQL nor MariaDB support the foreign key arguments "DEFERRABLE", "INITIALLY",
+or "MATCH".  Using the ``deferrable`` or ``initially`` keyword argument with
+:class:`_schema.ForeignKeyConstraint` or :class:`_schema.ForeignKey`
+will have the effect of
+these keywords being rendered in a DDL expression, which will then raise an
+error on MySQL or MariaDB.  In order to use these keywords on a foreign key while having
+them ignored on a MySQL / MariaDB backend, use a custom compile rule::
+
+    from sqlalchemy.ext.compiler import compiles
+    from sqlalchemy.schema import ForeignKeyConstraint
+
+
+    @compiles(ForeignKeyConstraint, "mysql", "mariadb")
+    def process(element, compiler, **kw):
+        element.deferrable = element.initially = None
+        return compiler.visit_foreign_key_constraint(element, **kw)
+
+The "MATCH" keyword is in fact more insidious, and is explicitly disallowed
+by SQLAlchemy in conjunction with the MySQL or MariaDB backends.  This argument is
+silently ignored by MySQL / MariaDB, but in addition has the effect of ON UPDATE and ON
+DELETE options also being ignored by the backend.   Therefore MATCH should
+never be used with the MySQL / MariaDB backends; as is the case with DEFERRABLE and
+INITIALLY, custom compilation rules can be used to correct a
+ForeignKeyConstraint at DDL definition time.
+
+Reflection of Foreign Key Constraints
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Not all MySQL / MariaDB storage engines support foreign keys.  When using the
+very common ``MyISAM`` MySQL storage engine, the information loaded by table
+reflection will not include foreign keys.  For these tables, you may supply a
+:class:`~sqlalchemy.ForeignKeyConstraint` at reflection time::
+
+  Table(
+      "mytable",
+      metadata,
+      ForeignKeyConstraint(["other_id"], ["othertable.other_id"]),
+      autoload_with=engine,
+  )
+
+.. seealso::
+
+    :ref:`mysql_storage_engines`
+
+.. _mysql_unique_constraints:
+
+MySQL / MariaDB Unique Constraints and Reflection
+----------------------------------------------------
+
+SQLAlchemy supports both the :class:`.Index` construct with the
+flag ``unique=True``, indicating a UNIQUE index, as well as the
+:class:`.UniqueConstraint` construct, representing a UNIQUE constraint.
+Both objects/syntaxes are supported by MySQL / MariaDB when emitting DDL to create
+these constraints.  However, MySQL / MariaDB does not have a unique constraint
+construct that is separate from a unique index; that is, the "UNIQUE"
+constraint on MySQL / MariaDB is equivalent to creating a "UNIQUE INDEX".
+
+When reflecting these constructs, the
+:meth:`_reflection.Inspector.get_indexes`
+and the :meth:`_reflection.Inspector.get_unique_constraints`
+methods will **both**
+return an entry for a UNIQUE index in MySQL / MariaDB.  However, when performing
+full table reflection using ``Table(..., autoload_with=engine)``,
+the :class:`.UniqueConstraint` construct is
+**not** part of the fully reflected :class:`_schema.Table` construct under any
+circumstances; this construct is always represented by a :class:`.Index`
+with the ``unique=True`` setting present in the :attr:`_schema.Table.indexes`
+collection.
+
+
+TIMESTAMP / DATETIME issues
+---------------------------
+
+.. _mysql_timestamp_onupdate:
+
+Rendering ON UPDATE CURRENT TIMESTAMP for MySQL / MariaDB's explicit_defaults_for_timestamp
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+MySQL / MariaDB have historically expanded the DDL for the :class:`_types.TIMESTAMP`
+datatype into the phrase "TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE
+CURRENT_TIMESTAMP", which includes non-standard SQL that automatically updates
+the column with the current timestamp when an UPDATE occurs, eliminating the
+usual need to use a trigger in such a case where server-side update changes are
+desired.
+
+MySQL 5.6 introduced a new flag `explicit_defaults_for_timestamp
+<https://dev.mysql.com/doc/refman/5.6/en/server-system-variables.html
+#sysvar_explicit_defaults_for_timestamp>`_ which disables the above behavior,
+and in MySQL 8 this flag defaults to true, meaning in order to get a MySQL
+"on update timestamp" without changing this flag, the above DDL must be
+rendered explicitly.   Additionally, the same DDL is valid for use of the
+``DATETIME`` datatype as well.
+
+SQLAlchemy's MySQL dialect does not yet have an option to generate
+MySQL's "ON UPDATE CURRENT_TIMESTAMP" clause, noting that this is not a general
+purpose "ON UPDATE" as there is no such syntax in standard SQL.  SQLAlchemy's
+:paramref:`_schema.Column.server_onupdate` parameter is currently not related
+to this special MySQL behavior.
+
+To generate this DDL, make use of the :paramref:`_schema.Column.server_default`
+parameter and pass a textual clause that also includes the ON UPDATE clause::
+
+    from sqlalchemy import Table, MetaData, Column, Integer, String, TIMESTAMP
+    from sqlalchemy import text
+
+    metadata = MetaData()
+
+    mytable = Table(
+        "mytable",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("data", String(50)),
+        Column(
+            "last_updated",
+            TIMESTAMP,
+            server_default=text(
+                "CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP"
+            ),
+        ),
+    )
+
+The same instructions apply to use of the :class:`_types.DateTime` and
+:class:`_types.DATETIME` datatypes::
+
+    from sqlalchemy import DateTime
+
+    mytable = Table(
+        "mytable",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("data", String(50)),
+        Column(
+            "last_updated",
+            DateTime,
+            server_default=text(
+                "CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP"
+            ),
+        ),
+    )
+
+Even though the :paramref:`_schema.Column.server_onupdate` feature does not
+generate this DDL, it still may be desirable to signal to the ORM that this
+updated value should be fetched.  This syntax looks like the following::
+
+    from sqlalchemy.schema import FetchedValue
+
+
+    class MyClass(Base):
+        __tablename__ = "mytable"
+
+        id = Column(Integer, primary_key=True)
+        data = Column(String(50))
+        last_updated = Column(
+            TIMESTAMP,
+            server_default=text(
+                "CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP"
+            ),
+            server_onupdate=FetchedValue(),
+        )
+
+.. _mysql_timestamp_null:
+
+TIMESTAMP Columns and NULL
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+MySQL historically enforces that a column which specifies the
+TIMESTAMP datatype implicitly includes a default value of
+CURRENT_TIMESTAMP, even though this is not stated, and additionally
+sets the column as NOT NULL, the opposite behavior vs. that of all
+other datatypes:
+
+.. sourcecode:: text
+
+    mysql> CREATE TABLE ts_test (
+        -> a INTEGER,
+        -> b INTEGER NOT NULL,
+        -> c TIMESTAMP,
+        -> d TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+        -> e TIMESTAMP NULL);
+    Query OK, 0 rows affected (0.03 sec)
+
+    mysql> SHOW CREATE TABLE ts_test;
+    +---------+-----------------------------------------------------
+    | Table   | Create Table
+    +---------+-----------------------------------------------------
+    | ts_test | CREATE TABLE `ts_test` (
+      `a` int(11) DEFAULT NULL,
+      `b` int(11) NOT NULL,
+      `c` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
+      `d` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
+      `e` timestamp NULL DEFAULT NULL
+    ) ENGINE=MyISAM DEFAULT CHARSET=latin1
+
+Above, we see that an INTEGER column defaults to NULL, unless it is specified
+with NOT NULL.   But when the column is of type TIMESTAMP, an implicit
+default of CURRENT_TIMESTAMP is generated which also coerces the column
+to be a NOT NULL, even though we did not specify it as such.
+
+This behavior of MySQL can be changed on the MySQL side using the
+`explicit_defaults_for_timestamp
+<https://dev.mysql.com/doc/refman/5.6/en/server-system-variables.html
+#sysvar_explicit_defaults_for_timestamp>`_ configuration flag introduced in
+MySQL 5.6.  With this server setting enabled, TIMESTAMP columns behave like
+any other datatype on the MySQL side with regards to defaults and nullability.
+
+However, to accommodate the vast majority of MySQL databases that do not
+specify this new flag, SQLAlchemy emits the "NULL" specifier explicitly with
+any TIMESTAMP column that does not specify ``nullable=False``.   In order to
+accommodate newer databases that specify ``explicit_defaults_for_timestamp``,
+SQLAlchemy also emits NOT NULL for TIMESTAMP columns that do specify
+``nullable=False``.   The following example illustrates::
+
+    from sqlalchemy import MetaData, Integer, Table, Column, text
+    from sqlalchemy.dialects.mysql import TIMESTAMP
+
+    m = MetaData()
+    t = Table(
+        "ts_test",
+        m,
+        Column("a", Integer),
+        Column("b", Integer, nullable=False),
+        Column("c", TIMESTAMP),
+        Column("d", TIMESTAMP, nullable=False),
+    )
+
+
+    from sqlalchemy import create_engine
+
+    e = create_engine("mysql+mysqldb://scott:tiger@localhost/test", echo=True)
+    m.create_all(e)
+
+output:
+
+.. sourcecode:: sql
+
+    CREATE TABLE ts_test (
+        a INTEGER,
+        b INTEGER NOT NULL,
+        c TIMESTAMP NULL,
+        d TIMESTAMP NOT NULL
+    )
+
+"""  # noqa
+from __future__ import annotations
+
+from array import array as _array
+from collections import defaultdict
+from itertools import compress
+import re
+from typing import cast
+
+from . import reflection as _reflection
+from .enumerated import ENUM
+from .enumerated import SET
+from .json import JSON
+from .json import JSONIndexType
+from .json import JSONPathType
+from .reserved_words import RESERVED_WORDS_MARIADB
+from .reserved_words import RESERVED_WORDS_MYSQL
+from .types import _FloatType
+from .types import _IntegerType
+from .types import _MatchType
+from .types import _NumericType
+from .types import _StringType
+from .types import BIGINT
+from .types import BIT
+from .types import CHAR
+from .types import DATETIME
+from .types import DECIMAL
+from .types import DOUBLE
+from .types import FLOAT
+from .types import INTEGER
+from .types import LONGBLOB
+from .types import LONGTEXT
+from .types import MEDIUMBLOB
+from .types import MEDIUMINT
+from .types import MEDIUMTEXT
+from .types import NCHAR
+from .types import NUMERIC
+from .types import NVARCHAR
+from .types import REAL
+from .types import SMALLINT
+from .types import TEXT
+from .types import TIME
+from .types import TIMESTAMP
+from .types import TINYBLOB
+from .types import TINYINT
+from .types import TINYTEXT
+from .types import VARCHAR
+from .types import YEAR
+from ... import exc
+from ... import literal_column
+from ... import log
+from ... import schema as sa_schema
+from ... import sql
+from ... import util
+from ...engine import cursor as _cursor
+from ...engine import default
+from ...engine import reflection
+from ...engine.reflection import ReflectionDefaults
+from ...sql import coercions
+from ...sql import compiler
+from ...sql import elements
+from ...sql import functions
+from ...sql import operators
+from ...sql import roles
+from ...sql import sqltypes
+from ...sql import util as sql_util
+from ...sql import visitors
+from ...sql.compiler import InsertmanyvaluesSentinelOpts
+from ...sql.compiler import SQLCompiler
+from ...sql.schema import SchemaConst
+from ...types import BINARY
+from ...types import BLOB
+from ...types import BOOLEAN
+from ...types import DATE
+from ...types import UUID
+from ...types import VARBINARY
+from ...util import topological
+
+
+SET_RE = re.compile(
+    r"\s*SET\s+(?:(?:GLOBAL|SESSION)\s+)?\w", re.I | re.UNICODE
+)
+
+# old names
+MSTime = TIME
+MSSet = SET
+MSEnum = ENUM
+MSLongBlob = LONGBLOB
+MSMediumBlob = MEDIUMBLOB
+MSTinyBlob = TINYBLOB
+MSBlob = BLOB
+MSBinary = BINARY
+MSVarBinary = VARBINARY
+MSNChar = NCHAR
+MSNVarChar = NVARCHAR
+MSChar = CHAR
+MSString = VARCHAR
+MSLongText = LONGTEXT
+MSMediumText = MEDIUMTEXT
+MSTinyText = TINYTEXT
+MSText = TEXT
+MSYear = YEAR
+MSTimeStamp = TIMESTAMP
+MSBit = BIT
+MSSmallInteger = SMALLINT
+MSTinyInteger = TINYINT
+MSMediumInteger = MEDIUMINT
+MSBigInteger = BIGINT
+MSNumeric = NUMERIC
+MSDecimal = DECIMAL
+MSDouble = DOUBLE
+MSReal = REAL
+MSFloat = FLOAT
+MSInteger = INTEGER
+
+colspecs = {
+    _IntegerType: _IntegerType,
+    _NumericType: _NumericType,
+    _FloatType: _FloatType,
+    sqltypes.Numeric: NUMERIC,
+    sqltypes.Float: FLOAT,
+    sqltypes.Double: DOUBLE,
+    sqltypes.Time: TIME,
+    sqltypes.Enum: ENUM,
+    sqltypes.MatchType: _MatchType,
+    sqltypes.JSON: JSON,
+    sqltypes.JSON.JSONIndexType: JSONIndexType,
+    sqltypes.JSON.JSONPathType: JSONPathType,
+}
+
+# Everything 3.23 through 5.1 excepting OpenGIS types.
+ischema_names = {
+    "bigint": BIGINT,
+    "binary": BINARY,
+    "bit": BIT,
+    "blob": BLOB,
+    "boolean": BOOLEAN,
+    "char": CHAR,
+    "date": DATE,
+    "datetime": DATETIME,
+    "decimal": DECIMAL,
+    "double": DOUBLE,
+    "enum": ENUM,
+    "fixed": DECIMAL,
+    "float": FLOAT,
+    "int": INTEGER,
+    "integer": INTEGER,
+    "json": JSON,
+    "longblob": LONGBLOB,
+    "longtext": LONGTEXT,
+    "mediumblob": MEDIUMBLOB,
+    "mediumint": MEDIUMINT,
+    "mediumtext": MEDIUMTEXT,
+    "nchar": NCHAR,
+    "nvarchar": NVARCHAR,
+    "numeric": NUMERIC,
+    "set": SET,
+    "smallint": SMALLINT,
+    "text": TEXT,
+    "time": TIME,
+    "timestamp": TIMESTAMP,
+    "tinyblob": TINYBLOB,
+    "tinyint": TINYINT,
+    "tinytext": TINYTEXT,
+    "uuid": UUID,
+    "varbinary": VARBINARY,
+    "varchar": VARCHAR,
+    "year": YEAR,
+}
+
+
+class MySQLExecutionContext(default.DefaultExecutionContext):
+    def post_exec(self):
+        if (
+            self.isdelete
+            and cast(SQLCompiler, self.compiled).effective_returning
+            and not self.cursor.description
+        ):
+            # All MySQL/mariadb drivers appear to not include
+            # cursor.description for DELETE..RETURNING with no rows if the
+            # WHERE criteria is a straight "false" condition such as our EMPTY
+            # IN condition. manufacture an empty result in this case (issue
+            # #10505)
+            #
+            # taken from cx_Oracle implementation
+            self.cursor_fetch_strategy = (
+                _cursor.FullyBufferedCursorFetchStrategy(
+                    self.cursor,
+                    [
+                        (entry.keyname, None)
+                        for entry in cast(
+                            SQLCompiler, self.compiled
+                        )._result_columns
+                    ],
+                    [],
+                )
+            )
+
+    def create_server_side_cursor(self):
+        if self.dialect.supports_server_side_cursors:
+            return self._dbapi_connection.cursor(self.dialect._sscursor)
+        else:
+            raise NotImplementedError()
+
+    def fire_sequence(self, seq, type_):
+        return self._execute_scalar(
+            (
+                "select nextval(%s)"
+                % self.identifier_preparer.format_sequence(seq)
+            ),
+            type_,
+        )
+
+
+class MySQLCompiler(compiler.SQLCompiler):
+    render_table_with_column_in_update_from = True
+    """Overridden from base SQLCompiler value"""
+
+    extract_map = compiler.SQLCompiler.extract_map.copy()
+    extract_map.update({"milliseconds": "millisecond"})
+
+    def default_from(self):
+        """Called when a ``SELECT`` statement has no froms,
+        and no ``FROM`` clause is to be appended.
+
+        """
+        if self.stack:
+            stmt = self.stack[-1]["selectable"]
+            if stmt._where_criteria:
+                return " FROM DUAL"
+
+        return ""
+
+    def visit_random_func(self, fn, **kw):
+        return "rand%s" % self.function_argspec(fn)
+
+    def visit_rollup_func(self, fn, **kw):
+        clause = ", ".join(
+            elem._compiler_dispatch(self, **kw) for elem in fn.clauses
+        )
+        return f"{clause} WITH ROLLUP"
+
+    def visit_aggregate_strings_func(self, fn, **kw):
+        expr, delimeter = (
+            elem._compiler_dispatch(self, **kw) for elem in fn.clauses
+        )
+        return f"group_concat({expr} SEPARATOR {delimeter})"
+
+    def visit_sequence(self, seq, **kw):
+        return "nextval(%s)" % self.preparer.format_sequence(seq)
+
+    def visit_sysdate_func(self, fn, **kw):
+        return "SYSDATE()"
+
+    def _render_json_extract_from_binary(self, binary, operator, **kw):
+        # note we are intentionally calling upon the process() calls in the
+        # order in which they appear in the SQL String as this is used
+        # by positional parameter rendering
+
+        if binary.type._type_affinity is sqltypes.JSON:
+            return "JSON_EXTRACT(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+
+        # for non-JSON, MySQL doesn't handle JSON null at all so it has to
+        # be explicit
+        case_expression = "CASE JSON_EXTRACT(%s, %s) WHEN 'null' THEN NULL" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+        if binary.type._type_affinity is sqltypes.Integer:
+            type_expression = (
+                "ELSE CAST(JSON_EXTRACT(%s, %s) AS SIGNED INTEGER)"
+                % (
+                    self.process(binary.left, **kw),
+                    self.process(binary.right, **kw),
+                )
+            )
+        elif binary.type._type_affinity is sqltypes.Numeric:
+            if (
+                binary.type.scale is not None
+                and binary.type.precision is not None
+            ):
+                # using DECIMAL here because MySQL does not recognize NUMERIC
+                type_expression = (
+                    "ELSE CAST(JSON_EXTRACT(%s, %s) AS DECIMAL(%s, %s))"
+                    % (
+                        self.process(binary.left, **kw),
+                        self.process(binary.right, **kw),
+                        binary.type.precision,
+                        binary.type.scale,
+                    )
+                )
+            else:
+                # FLOAT / REAL not added in MySQL til 8.0.17
+                type_expression = (
+                    "ELSE JSON_EXTRACT(%s, %s)+0.0000000000000000000000"
+                    % (
+                        self.process(binary.left, **kw),
+                        self.process(binary.right, **kw),
+                    )
+                )
+        elif binary.type._type_affinity is sqltypes.Boolean:
+            # the NULL handling is particularly weird with boolean, so
+            # explicitly return true/false constants
+            type_expression = "WHEN true THEN true ELSE false"
+        elif binary.type._type_affinity is sqltypes.String:
+            # (gord): this fails with a JSON value that's a four byte unicode
+            # string.  SQLite has the same problem at the moment
+            # (zzzeek): I'm not really sure.  let's take a look at a test case
+            # that hits each backend and maybe make a requires rule for it?
+            type_expression = "ELSE JSON_UNQUOTE(JSON_EXTRACT(%s, %s))" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+        else:
+            # other affinity....this is not expected right now
+            type_expression = "ELSE JSON_EXTRACT(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+
+        return case_expression + " " + type_expression + " END"
+
+    def visit_json_getitem_op_binary(self, binary, operator, **kw):
+        return self._render_json_extract_from_binary(binary, operator, **kw)
+
+    def visit_json_path_getitem_op_binary(self, binary, operator, **kw):
+        return self._render_json_extract_from_binary(binary, operator, **kw)
+
+    def visit_on_duplicate_key_update(self, on_duplicate, **kw):
+        statement = self.current_executable
+
+        if on_duplicate._parameter_ordering:
+            parameter_ordering = [
+                coercions.expect(roles.DMLColumnRole, key)
+                for key in on_duplicate._parameter_ordering
+            ]
+            ordered_keys = set(parameter_ordering)
+            cols = [
+                statement.table.c[key]
+                for key in parameter_ordering
+                if key in statement.table.c
+            ] + [c for c in statement.table.c if c.key not in ordered_keys]
+        else:
+            cols = statement.table.c
+
+        clauses = []
+
+        requires_mysql8_alias = statement.select is None and (
+            self.dialect._requires_alias_for_on_duplicate_key
+        )
+
+        if requires_mysql8_alias:
+            if statement.table.name.lower() == "new":
+                _on_dup_alias_name = "new_1"
+            else:
+                _on_dup_alias_name = "new"
+
+        on_duplicate_update = {
+            coercions.expect_as_key(roles.DMLColumnRole, key): value
+            for key, value in on_duplicate.update.items()
+        }
+
+        # traverses through all table columns to preserve table column order
+        for column in (col for col in cols if col.key in on_duplicate_update):
+            val = on_duplicate_update[column.key]
+
+            # TODO: this coercion should be up front.  we can't cache
+            # SQL constructs with non-bound literals buried in them
+            if coercions._is_literal(val):
+                val = elements.BindParameter(None, val, type_=column.type)
+                value_text = self.process(val.self_group(), use_schema=False)
+            else:
+
+                def replace(obj):
+                    if (
+                        isinstance(obj, elements.BindParameter)
+                        and obj.type._isnull
+                    ):
+                        obj = obj._clone()
+                        obj.type = column.type
+                        return obj
+                    elif (
+                        isinstance(obj, elements.ColumnClause)
+                        and obj.table is on_duplicate.inserted_alias
+                    ):
+                        if requires_mysql8_alias:
+                            column_literal_clause = (
+                                f"{_on_dup_alias_name}."
+                                f"{self.preparer.quote(obj.name)}"
+                            )
+                        else:
+                            column_literal_clause = (
+                                f"VALUES({self.preparer.quote(obj.name)})"
+                            )
+                        return literal_column(column_literal_clause)
+                    else:
+                        # element is not replaced
+                        return None
+
+                val = visitors.replacement_traverse(val, {}, replace)
+                value_text = self.process(val.self_group(), use_schema=False)
+
+            name_text = self.preparer.quote(column.name)
+            clauses.append("%s = %s" % (name_text, value_text))
+
+        non_matching = set(on_duplicate_update) - {c.key for c in cols}
+        if non_matching:
+            util.warn(
+                "Additional column names not matching "
+                "any column keys in table '%s': %s"
+                % (
+                    self.statement.table.name,
+                    (", ".join("'%s'" % c for c in non_matching)),
+                )
+            )
+
+        if requires_mysql8_alias:
+            return (
+                f"AS {_on_dup_alias_name} "
+                f"ON DUPLICATE KEY UPDATE {', '.join(clauses)}"
+            )
+        else:
+            return f"ON DUPLICATE KEY UPDATE {', '.join(clauses)}"
+
+    def visit_concat_op_expression_clauselist(
+        self, clauselist, operator, **kw
+    ):
+        return "concat(%s)" % (
+            ", ".join(self.process(elem, **kw) for elem in clauselist.clauses)
+        )
+
+    def visit_concat_op_binary(self, binary, operator, **kw):
+        return "concat(%s, %s)" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    _match_valid_flag_combinations = frozenset(
+        (
+            # (boolean_mode, natural_language, query_expansion)
+            (False, False, False),
+            (True, False, False),
+            (False, True, False),
+            (False, False, True),
+            (False, True, True),
+        )
+    )
+
+    _match_flag_expressions = (
+        "IN BOOLEAN MODE",
+        "IN NATURAL LANGUAGE MODE",
+        "WITH QUERY EXPANSION",
+    )
+
+    def visit_mysql_match(self, element, **kw):
+        return self.visit_match_op_binary(element, element.operator, **kw)
+
+    def visit_match_op_binary(self, binary, operator, **kw):
+        """
+        Note that `mysql_boolean_mode` is enabled by default because of
+        backward compatibility
+        """
+
+        modifiers = binary.modifiers
+
+        boolean_mode = modifiers.get("mysql_boolean_mode", True)
+        natural_language = modifiers.get("mysql_natural_language", False)
+        query_expansion = modifiers.get("mysql_query_expansion", False)
+
+        flag_combination = (boolean_mode, natural_language, query_expansion)
+
+        if flag_combination not in self._match_valid_flag_combinations:
+            flags = (
+                "in_boolean_mode=%s" % boolean_mode,
+                "in_natural_language_mode=%s" % natural_language,
+                "with_query_expansion=%s" % query_expansion,
+            )
+
+            flags = ", ".join(flags)
+
+            raise exc.CompileError("Invalid MySQL match flags: %s" % flags)
+
+        match_clause = binary.left
+        match_clause = self.process(match_clause, **kw)
+        against_clause = self.process(binary.right, **kw)
+
+        if any(flag_combination):
+            flag_expressions = compress(
+                self._match_flag_expressions,
+                flag_combination,
+            )
+
+            against_clause = [against_clause]
+            against_clause.extend(flag_expressions)
+
+            against_clause = " ".join(against_clause)
+
+        return "MATCH (%s) AGAINST (%s)" % (match_clause, against_clause)
+
+    def get_from_hint_text(self, table, text):
+        return text
+
+    def visit_typeclause(self, typeclause, type_=None, **kw):
+        if type_ is None:
+            type_ = typeclause.type.dialect_impl(self.dialect)
+        if isinstance(type_, sqltypes.TypeDecorator):
+            return self.visit_typeclause(typeclause, type_.impl, **kw)
+        elif isinstance(type_, sqltypes.Integer):
+            if getattr(type_, "unsigned", False):
+                return "UNSIGNED INTEGER"
+            else:
+                return "SIGNED INTEGER"
+        elif isinstance(type_, sqltypes.TIMESTAMP):
+            return "DATETIME"
+        elif isinstance(
+            type_,
+            (
+                sqltypes.DECIMAL,
+                sqltypes.DateTime,
+                sqltypes.Date,
+                sqltypes.Time,
+            ),
+        ):
+            return self.dialect.type_compiler_instance.process(type_)
+        elif isinstance(type_, sqltypes.String) and not isinstance(
+            type_, (ENUM, SET)
+        ):
+            adapted = CHAR._adapt_string_for_cast(type_)
+            return self.dialect.type_compiler_instance.process(adapted)
+        elif isinstance(type_, sqltypes._Binary):
+            return "BINARY"
+        elif isinstance(type_, sqltypes.JSON):
+            return "JSON"
+        elif isinstance(type_, sqltypes.NUMERIC):
+            return self.dialect.type_compiler_instance.process(type_).replace(
+                "NUMERIC", "DECIMAL"
+            )
+        elif (
+            isinstance(type_, sqltypes.Float)
+            and self.dialect._support_float_cast
+        ):
+            return self.dialect.type_compiler_instance.process(type_)
+        else:
+            return None
+
+    def visit_cast(self, cast, **kw):
+        type_ = self.process(cast.typeclause)
+        if type_ is None:
+            util.warn(
+                "Datatype %s does not support CAST on MySQL/MariaDb; "
+                "the CAST will be skipped."
+                % self.dialect.type_compiler_instance.process(
+                    cast.typeclause.type
+                )
+            )
+            return self.process(cast.clause.self_group(), **kw)
+
+        return "CAST(%s AS %s)" % (self.process(cast.clause, **kw), type_)
+
+    def render_literal_value(self, value, type_):
+        value = super().render_literal_value(value, type_)
+        if self.dialect._backslash_escapes:
+            value = value.replace("\\", "\\\\")
+        return value
+
+    # override native_boolean=False behavior here, as
+    # MySQL still supports native boolean
+    def visit_true(self, element, **kw):
+        return "true"
+
+    def visit_false(self, element, **kw):
+        return "false"
+
+    def get_select_precolumns(self, select, **kw):
+        """Add special MySQL keywords in place of DISTINCT.
+
+        .. deprecated:: 1.4  This usage is deprecated.
+           :meth:`_expression.Select.prefix_with` should be used for special
+           keywords at the start of a SELECT.
+
+        """
+        if isinstance(select._distinct, str):
+            util.warn_deprecated(
+                "Sending string values for 'distinct' is deprecated in the "
+                "MySQL dialect and will be removed in a future release.  "
+                "Please use :meth:`.Select.prefix_with` for special keywords "
+                "at the start of a SELECT statement",
+                version="1.4",
+            )
+            return select._distinct.upper() + " "
+
+        return super().get_select_precolumns(select, **kw)
+
+    def visit_join(self, join, asfrom=False, from_linter=None, **kwargs):
+        if from_linter:
+            from_linter.edges.add((join.left, join.right))
+
+        if join.full:
+            join_type = " FULL OUTER JOIN "
+        elif join.isouter:
+            join_type = " LEFT OUTER JOIN "
+        else:
+            join_type = " INNER JOIN "
+
+        return "".join(
+            (
+                self.process(
+                    join.left, asfrom=True, from_linter=from_linter, **kwargs
+                ),
+                join_type,
+                self.process(
+                    join.right, asfrom=True, from_linter=from_linter, **kwargs
+                ),
+                " ON ",
+                self.process(join.onclause, from_linter=from_linter, **kwargs),
+            )
+        )
+
+    def for_update_clause(self, select, **kw):
+        if select._for_update_arg.read:
+            tmp = " LOCK IN SHARE MODE"
+        else:
+            tmp = " FOR UPDATE"
+
+        if select._for_update_arg.of and self.dialect.supports_for_update_of:
+            tables = util.OrderedSet()
+            for c in select._for_update_arg.of:
+                tables.update(sql_util.surface_selectables_only(c))
+
+            tmp += " OF " + ", ".join(
+                self.process(table, ashint=True, use_schema=False, **kw)
+                for table in tables
+            )
+
+        if select._for_update_arg.nowait:
+            tmp += " NOWAIT"
+
+        if select._for_update_arg.skip_locked:
+            tmp += " SKIP LOCKED"
+
+        return tmp
+
+    def limit_clause(self, select, **kw):
+        # MySQL supports:
+        #   LIMIT <limit>
+        #   LIMIT <offset>, <limit>
+        # and in server versions > 3.3:
+        #   LIMIT <limit> OFFSET <offset>
+        # The latter is more readable for offsets but we're stuck with the
+        # former until we can refine dialects by server revision.
+
+        limit_clause, offset_clause = (
+            select._limit_clause,
+            select._offset_clause,
+        )
+
+        if limit_clause is None and offset_clause is None:
+            return ""
+        elif offset_clause is not None:
+            # As suggested by the MySQL docs, need to apply an
+            # artificial limit if one wasn't provided
+            # https://dev.mysql.com/doc/refman/5.0/en/select.html
+            if limit_clause is None:
+                # TODO: remove ??
+                # hardwire the upper limit.  Currently
+                # needed consistent with the usage of the upper
+                # bound as part of MySQL's "syntax" for OFFSET with
+                # no LIMIT.
+                return " \n LIMIT %s, %s" % (
+                    self.process(offset_clause, **kw),
+                    "18446744073709551615",
+                )
+            else:
+                return " \n LIMIT %s, %s" % (
+                    self.process(offset_clause, **kw),
+                    self.process(limit_clause, **kw),
+                )
+        else:
+            # No offset provided, so just use the limit
+            return " \n LIMIT %s" % (self.process(limit_clause, **kw),)
+
+    def update_limit_clause(self, update_stmt):
+        limit = update_stmt.kwargs.get("%s_limit" % self.dialect.name, None)
+        if limit is not None:
+            return f"LIMIT {int(limit)}"
+        else:
+            return None
+
+    def delete_limit_clause(self, delete_stmt):
+        limit = delete_stmt.kwargs.get("%s_limit" % self.dialect.name, None)
+        if limit is not None:
+            return f"LIMIT {int(limit)}"
+        else:
+            return None
+
+    def update_tables_clause(self, update_stmt, from_table, extra_froms, **kw):
+        kw["asfrom"] = True
+        return ", ".join(
+            t._compiler_dispatch(self, **kw)
+            for t in [from_table] + list(extra_froms)
+        )
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        return None
+
+    def delete_table_clause(self, delete_stmt, from_table, extra_froms, **kw):
+        """If we have extra froms make sure we render any alias as hint."""
+        ashint = False
+        if extra_froms:
+            ashint = True
+        return from_table._compiler_dispatch(
+            self, asfrom=True, iscrud=True, ashint=ashint, **kw
+        )
+
+    def delete_extra_from_clause(
+        self, delete_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Render the DELETE .. USING clause specific to MySQL."""
+        kw["asfrom"] = True
+        return "USING " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in [from_table] + extra_froms
+        )
+
+    def visit_empty_set_expr(self, element_types, **kw):
+        return (
+            "SELECT %(outer)s FROM (SELECT %(inner)s) "
+            "as _empty_set WHERE 1!=1"
+            % {
+                "inner": ", ".join(
+                    "1 AS _in_%s" % idx
+                    for idx, type_ in enumerate(element_types)
+                ),
+                "outer": ", ".join(
+                    "_in_%s" % idx for idx, type_ in enumerate(element_types)
+                ),
+            }
+        )
+
+    def visit_is_distinct_from_binary(self, binary, operator, **kw):
+        return "NOT (%s <=> %s)" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_is_not_distinct_from_binary(self, binary, operator, **kw):
+        return "%s <=> %s" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def _mariadb_regexp_flags(self, flags, pattern, **kw):
+        return "CONCAT('(?', %s, ')', %s)" % (
+            self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            self.process(pattern, **kw),
+        )
+
+    def _regexp_match(self, op_string, binary, operator, **kw):
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return self._generate_generic_binary(binary, op_string, **kw)
+        elif self.dialect.is_mariadb:
+            return "%s%s%s" % (
+                self.process(binary.left, **kw),
+                op_string,
+                self._mariadb_regexp_flags(flags, binary.right),
+            )
+        else:
+            text = "REGEXP_LIKE(%s, %s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+                self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            )
+            if op_string == " NOT REGEXP ":
+                return "NOT %s" % text
+            else:
+                return text
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._regexp_match(" REGEXP ", binary, operator, **kw)
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._regexp_match(" NOT REGEXP ", binary, operator, **kw)
+
+    def visit_regexp_replace_op_binary(self, binary, operator, **kw):
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return "REGEXP_REPLACE(%s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+            )
+        elif self.dialect.is_mariadb:
+            return "REGEXP_REPLACE(%s, %s, %s)" % (
+                self.process(binary.left, **kw),
+                self._mariadb_regexp_flags(flags, binary.right.clauses[0]),
+                self.process(binary.right.clauses[1], **kw),
+            )
+        else:
+            return "REGEXP_REPLACE(%s, %s, %s)" % (
+                self.process(binary.left, **kw),
+                self.process(binary.right, **kw),
+                self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            )
+
+
+class MySQLDDLCompiler(compiler.DDLCompiler):
+    def get_column_specification(self, column, **kw):
+        """Builds column DDL."""
+        if (
+            self.dialect.is_mariadb is True
+            and column.computed is not None
+            and column._user_defined_nullable is SchemaConst.NULL_UNSPECIFIED
+        ):
+            column.nullable = True
+        colspec = [
+            self.preparer.format_column(column),
+            self.dialect.type_compiler_instance.process(
+                column.type, type_expression=column
+            ),
+        ]
+
+        if column.computed is not None:
+            colspec.append(self.process(column.computed))
+
+        is_timestamp = isinstance(
+            column.type._unwrapped_dialect_impl(self.dialect),
+            sqltypes.TIMESTAMP,
+        )
+
+        if not column.nullable:
+            colspec.append("NOT NULL")
+
+        # see: https://docs.sqlalchemy.org/en/latest/dialects/mysql.html#mysql_timestamp_null  # noqa
+        elif column.nullable and is_timestamp:
+            colspec.append("NULL")
+
+        comment = column.comment
+        if comment is not None:
+            literal = self.sql_compiler.render_literal_value(
+                comment, sqltypes.String()
+            )
+            colspec.append("COMMENT " + literal)
+
+        if (
+            column.table is not None
+            and column is column.table._autoincrement_column
+            and (
+                column.server_default is None
+                or isinstance(column.server_default, sa_schema.Identity)
+            )
+            and not (
+                self.dialect.supports_sequences
+                and isinstance(column.default, sa_schema.Sequence)
+                and not column.default.optional
+            )
+        ):
+            colspec.append("AUTO_INCREMENT")
+        else:
+            default = self.get_column_default_string(column)
+            if default is not None:
+                if (
+                    isinstance(
+                        column.server_default.arg, functions.FunctionElement
+                    )
+                    and self.dialect._support_default_function
+                ):
+                    colspec.append(f"DEFAULT ({default})")
+                else:
+                    colspec.append("DEFAULT " + default)
+        return " ".join(colspec)
+
+    def post_create_table(self, table):
+        """Build table-level CREATE options like ENGINE and COLLATE."""
+
+        table_opts = []
+
+        opts = {
+            k[len(self.dialect.name) + 1 :].upper(): v
+            for k, v in table.kwargs.items()
+            if k.startswith("%s_" % self.dialect.name)
+        }
+
+        if table.comment is not None:
+            opts["COMMENT"] = table.comment
+
+        partition_options = [
+            "PARTITION_BY",
+            "PARTITIONS",
+            "SUBPARTITIONS",
+            "SUBPARTITION_BY",
+        ]
+
+        nonpart_options = set(opts).difference(partition_options)
+        part_options = set(opts).intersection(partition_options)
+
+        for opt in topological.sort(
+            [
+                ("DEFAULT_CHARSET", "COLLATE"),
+                ("DEFAULT_CHARACTER_SET", "COLLATE"),
+                ("CHARSET", "COLLATE"),
+                ("CHARACTER_SET", "COLLATE"),
+            ],
+            nonpart_options,
+        ):
+            arg = opts[opt]
+            if opt in _reflection._options_of_type_string:
+                arg = self.sql_compiler.render_literal_value(
+                    arg, sqltypes.String()
+                )
+
+            if opt in (
+                "DATA_DIRECTORY",
+                "INDEX_DIRECTORY",
+                "DEFAULT_CHARACTER_SET",
+                "CHARACTER_SET",
+                "DEFAULT_CHARSET",
+                "DEFAULT_COLLATE",
+            ):
+                opt = opt.replace("_", " ")
+
+            joiner = "="
+            if opt in (
+                "TABLESPACE",
+                "DEFAULT CHARACTER SET",
+                "CHARACTER SET",
+                "COLLATE",
+            ):
+                joiner = " "
+
+            table_opts.append(joiner.join((opt, arg)))
+
+        for opt in topological.sort(
+            [
+                ("PARTITION_BY", "PARTITIONS"),
+                ("PARTITION_BY", "SUBPARTITION_BY"),
+                ("PARTITION_BY", "SUBPARTITIONS"),
+                ("PARTITIONS", "SUBPARTITIONS"),
+                ("PARTITIONS", "SUBPARTITION_BY"),
+                ("SUBPARTITION_BY", "SUBPARTITIONS"),
+            ],
+            part_options,
+        ):
+            arg = opts[opt]
+            if opt in _reflection._options_of_type_string:
+                arg = self.sql_compiler.render_literal_value(
+                    arg, sqltypes.String()
+                )
+
+            opt = opt.replace("_", " ")
+            joiner = " "
+
+            table_opts.append(joiner.join((opt, arg)))
+
+        return " ".join(table_opts)
+
+    def visit_create_index(self, create, **kw):
+        index = create.element
+        self._verify_index_table(index)
+        preparer = self.preparer
+        table = preparer.format_table(index.table)
+
+        columns = [
+            self.sql_compiler.process(
+                (
+                    elements.Grouping(expr)
+                    if (
+                        isinstance(expr, elements.BinaryExpression)
+                        or (
+                            isinstance(expr, elements.UnaryExpression)
+                            and expr.modifier
+                            not in (operators.desc_op, operators.asc_op)
+                        )
+                        or isinstance(expr, functions.FunctionElement)
+                    )
+                    else expr
+                ),
+                include_table=False,
+                literal_binds=True,
+            )
+            for expr in index.expressions
+        ]
+
+        name = self._prepared_index_name(index)
+
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+
+        index_prefix = index.kwargs.get("%s_prefix" % self.dialect.name, None)
+        if index_prefix:
+            text += index_prefix + " "
+
+        text += "INDEX "
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+        text += "%s ON %s " % (name, table)
+
+        length = index.dialect_options[self.dialect.name]["length"]
+        if length is not None:
+            if isinstance(length, dict):
+                # length value can be a (column_name --> integer value)
+                # mapping specifying the prefix length for each column of the
+                # index
+                columns = ", ".join(
+                    (
+                        "%s(%d)" % (expr, length[col.name])
+                        if col.name in length
+                        else (
+                            "%s(%d)" % (expr, length[expr])
+                            if expr in length
+                            else "%s" % expr
+                        )
+                    )
+                    for col, expr in zip(index.expressions, columns)
+                )
+            else:
+                # or can be an integer value specifying the same
+                # prefix length for all columns of the index
+                columns = ", ".join(
+                    "%s(%d)" % (col, length) for col in columns
+                )
+        else:
+            columns = ", ".join(columns)
+        text += "(%s)" % columns
+
+        parser = index.dialect_options["mysql"]["with_parser"]
+        if parser is not None:
+            text += " WITH PARSER %s" % (parser,)
+
+        using = index.dialect_options["mysql"]["using"]
+        if using is not None:
+            text += " USING %s" % (preparer.quote(using))
+
+        return text
+
+    def visit_primary_key_constraint(self, constraint, **kw):
+        text = super().visit_primary_key_constraint(constraint)
+        using = constraint.dialect_options["mysql"]["using"]
+        if using:
+            text += " USING %s" % (self.preparer.quote(using))
+        return text
+
+    def visit_drop_index(self, drop, **kw):
+        index = drop.element
+        text = "\nDROP INDEX "
+        if drop.if_exists:
+            text += "IF EXISTS "
+
+        return text + "%s ON %s" % (
+            self._prepared_index_name(index, include_schema=False),
+            self.preparer.format_table(index.table),
+        )
+
+    def visit_drop_constraint(self, drop, **kw):
+        constraint = drop.element
+        if isinstance(constraint, sa_schema.ForeignKeyConstraint):
+            qual = "FOREIGN KEY "
+            const = self.preparer.format_constraint(constraint)
+        elif isinstance(constraint, sa_schema.PrimaryKeyConstraint):
+            qual = "PRIMARY KEY "
+            const = ""
+        elif isinstance(constraint, sa_schema.UniqueConstraint):
+            qual = "INDEX "
+            const = self.preparer.format_constraint(constraint)
+        elif isinstance(constraint, sa_schema.CheckConstraint):
+            if self.dialect.is_mariadb:
+                qual = "CONSTRAINT "
+            else:
+                qual = "CHECK "
+            const = self.preparer.format_constraint(constraint)
+        else:
+            qual = ""
+            const = self.preparer.format_constraint(constraint)
+        return "ALTER TABLE %s DROP %s%s" % (
+            self.preparer.format_table(constraint.table),
+            qual,
+            const,
+        )
+
+    def define_constraint_match(self, constraint):
+        if constraint.match is not None:
+            raise exc.CompileError(
+                "MySQL ignores the 'MATCH' keyword while at the same time "
+                "causes ON UPDATE/ON DELETE clauses to be ignored."
+            )
+        return ""
+
+    def visit_set_table_comment(self, create, **kw):
+        return "ALTER TABLE %s COMMENT %s" % (
+            self.preparer.format_table(create.element),
+            self.sql_compiler.render_literal_value(
+                create.element.comment, sqltypes.String()
+            ),
+        )
+
+    def visit_drop_table_comment(self, create, **kw):
+        return "ALTER TABLE %s COMMENT ''" % (
+            self.preparer.format_table(create.element)
+        )
+
+    def visit_set_column_comment(self, create, **kw):
+        return "ALTER TABLE %s CHANGE %s %s" % (
+            self.preparer.format_table(create.element.table),
+            self.preparer.format_column(create.element),
+            self.get_column_specification(create.element),
+        )
+
+
+class MySQLTypeCompiler(compiler.GenericTypeCompiler):
+    def _extend_numeric(self, type_, spec):
+        "Extend a numeric-type declaration with MySQL specific extensions."
+
+        if not self._mysql_type(type_):
+            return spec
+
+        if type_.unsigned:
+            spec += " UNSIGNED"
+        if type_.zerofill:
+            spec += " ZEROFILL"
+        return spec
+
+    def _extend_string(self, type_, defaults, spec):
+        """Extend a string-type declaration with standard SQL CHARACTER SET /
+        COLLATE annotations and MySQL specific extensions.
+
+        """
+
+        def attr(name):
+            return getattr(type_, name, defaults.get(name))
+
+        if attr("charset"):
+            charset = "CHARACTER SET %s" % attr("charset")
+        elif attr("ascii"):
+            charset = "ASCII"
+        elif attr("unicode"):
+            charset = "UNICODE"
+        else:
+            charset = None
+
+        if attr("collation"):
+            collation = "COLLATE %s" % type_.collation
+        elif attr("binary"):
+            collation = "BINARY"
+        else:
+            collation = None
+
+        if attr("national"):
+            # NATIONAL (aka NCHAR/NVARCHAR) trumps charsets.
+            return " ".join(
+                [c for c in ("NATIONAL", spec, collation) if c is not None]
+            )
+        return " ".join(
+            [c for c in (spec, charset, collation) if c is not None]
+        )
+
+    def _mysql_type(self, type_):
+        return isinstance(type_, (_StringType, _NumericType))
+
+    def visit_NUMERIC(self, type_, **kw):
+        if type_.precision is None:
+            return self._extend_numeric(type_, "NUMERIC")
+        elif type_.scale is None:
+            return self._extend_numeric(
+                type_,
+                "NUMERIC(%(precision)s)" % {"precision": type_.precision},
+            )
+        else:
+            return self._extend_numeric(
+                type_,
+                "NUMERIC(%(precision)s, %(scale)s)"
+                % {"precision": type_.precision, "scale": type_.scale},
+            )
+
+    def visit_DECIMAL(self, type_, **kw):
+        if type_.precision is None:
+            return self._extend_numeric(type_, "DECIMAL")
+        elif type_.scale is None:
+            return self._extend_numeric(
+                type_,
+                "DECIMAL(%(precision)s)" % {"precision": type_.precision},
+            )
+        else:
+            return self._extend_numeric(
+                type_,
+                "DECIMAL(%(precision)s, %(scale)s)"
+                % {"precision": type_.precision, "scale": type_.scale},
+            )
+
+    def visit_DOUBLE(self, type_, **kw):
+        if type_.precision is not None and type_.scale is not None:
+            return self._extend_numeric(
+                type_,
+                "DOUBLE(%(precision)s, %(scale)s)"
+                % {"precision": type_.precision, "scale": type_.scale},
+            )
+        else:
+            return self._extend_numeric(type_, "DOUBLE")
+
+    def visit_REAL(self, type_, **kw):
+        if type_.precision is not None and type_.scale is not None:
+            return self._extend_numeric(
+                type_,
+                "REAL(%(precision)s, %(scale)s)"
+                % {"precision": type_.precision, "scale": type_.scale},
+            )
+        else:
+            return self._extend_numeric(type_, "REAL")
+
+    def visit_FLOAT(self, type_, **kw):
+        if (
+            self._mysql_type(type_)
+            and type_.scale is not None
+            and type_.precision is not None
+        ):
+            return self._extend_numeric(
+                type_, "FLOAT(%s, %s)" % (type_.precision, type_.scale)
+            )
+        elif type_.precision is not None:
+            return self._extend_numeric(
+                type_, "FLOAT(%s)" % (type_.precision,)
+            )
+        else:
+            return self._extend_numeric(type_, "FLOAT")
+
+    def visit_INTEGER(self, type_, **kw):
+        if self._mysql_type(type_) and type_.display_width is not None:
+            return self._extend_numeric(
+                type_,
+                "INTEGER(%(display_width)s)"
+                % {"display_width": type_.display_width},
+            )
+        else:
+            return self._extend_numeric(type_, "INTEGER")
+
+    def visit_BIGINT(self, type_, **kw):
+        if self._mysql_type(type_) and type_.display_width is not None:
+            return self._extend_numeric(
+                type_,
+                "BIGINT(%(display_width)s)"
+                % {"display_width": type_.display_width},
+            )
+        else:
+            return self._extend_numeric(type_, "BIGINT")
+
+    def visit_MEDIUMINT(self, type_, **kw):
+        if self._mysql_type(type_) and type_.display_width is not None:
+            return self._extend_numeric(
+                type_,
+                "MEDIUMINT(%(display_width)s)"
+                % {"display_width": type_.display_width},
+            )
+        else:
+            return self._extend_numeric(type_, "MEDIUMINT")
+
+    def visit_TINYINT(self, type_, **kw):
+        if self._mysql_type(type_) and type_.display_width is not None:
+            return self._extend_numeric(
+                type_, "TINYINT(%s)" % type_.display_width
+            )
+        else:
+            return self._extend_numeric(type_, "TINYINT")
+
+    def visit_SMALLINT(self, type_, **kw):
+        if self._mysql_type(type_) and type_.display_width is not None:
+            return self._extend_numeric(
+                type_,
+                "SMALLINT(%(display_width)s)"
+                % {"display_width": type_.display_width},
+            )
+        else:
+            return self._extend_numeric(type_, "SMALLINT")
+
+    def visit_BIT(self, type_, **kw):
+        if type_.length is not None:
+            return "BIT(%s)" % type_.length
+        else:
+            return "BIT"
+
+    def visit_DATETIME(self, type_, **kw):
+        if getattr(type_, "fsp", None):
+            return "DATETIME(%d)" % type_.fsp
+        else:
+            return "DATETIME"
+
+    def visit_DATE(self, type_, **kw):
+        return "DATE"
+
+    def visit_TIME(self, type_, **kw):
+        if getattr(type_, "fsp", None):
+            return "TIME(%d)" % type_.fsp
+        else:
+            return "TIME"
+
+    def visit_TIMESTAMP(self, type_, **kw):
+        if getattr(type_, "fsp", None):
+            return "TIMESTAMP(%d)" % type_.fsp
+        else:
+            return "TIMESTAMP"
+
+    def visit_YEAR(self, type_, **kw):
+        if type_.display_width is None:
+            return "YEAR"
+        else:
+            return "YEAR(%s)" % type_.display_width
+
+    def visit_TEXT(self, type_, **kw):
+        if type_.length is not None:
+            return self._extend_string(type_, {}, "TEXT(%d)" % type_.length)
+        else:
+            return self._extend_string(type_, {}, "TEXT")
+
+    def visit_TINYTEXT(self, type_, **kw):
+        return self._extend_string(type_, {}, "TINYTEXT")
+
+    def visit_MEDIUMTEXT(self, type_, **kw):
+        return self._extend_string(type_, {}, "MEDIUMTEXT")
+
+    def visit_LONGTEXT(self, type_, **kw):
+        return self._extend_string(type_, {}, "LONGTEXT")
+
+    def visit_VARCHAR(self, type_, **kw):
+        if type_.length is not None:
+            return self._extend_string(type_, {}, "VARCHAR(%d)" % type_.length)
+        else:
+            raise exc.CompileError(
+                "VARCHAR requires a length on dialect %s" % self.dialect.name
+            )
+
+    def visit_CHAR(self, type_, **kw):
+        if type_.length is not None:
+            return self._extend_string(
+                type_, {}, "CHAR(%(length)s)" % {"length": type_.length}
+            )
+        else:
+            return self._extend_string(type_, {}, "CHAR")
+
+    def visit_NVARCHAR(self, type_, **kw):
+        # We'll actually generate the equiv. "NATIONAL VARCHAR" instead
+        # of "NVARCHAR".
+        if type_.length is not None:
+            return self._extend_string(
+                type_,
+                {"national": True},
+                "VARCHAR(%(length)s)" % {"length": type_.length},
+            )
+        else:
+            raise exc.CompileError(
+                "NVARCHAR requires a length on dialect %s" % self.dialect.name
+            )
+
+    def visit_NCHAR(self, type_, **kw):
+        # We'll actually generate the equiv.
+        # "NATIONAL CHAR" instead of "NCHAR".
+        if type_.length is not None:
+            return self._extend_string(
+                type_,
+                {"national": True},
+                "CHAR(%(length)s)" % {"length": type_.length},
+            )
+        else:
+            return self._extend_string(type_, {"national": True}, "CHAR")
+
+    def visit_UUID(self, type_, **kw):
+        return "UUID"
+
+    def visit_VARBINARY(self, type_, **kw):
+        return "VARBINARY(%d)" % type_.length
+
+    def visit_JSON(self, type_, **kw):
+        return "JSON"
+
+    def visit_large_binary(self, type_, **kw):
+        return self.visit_BLOB(type_)
+
+    def visit_enum(self, type_, **kw):
+        if not type_.native_enum:
+            return super().visit_enum(type_)
+        else:
+            return self._visit_enumerated_values("ENUM", type_, type_.enums)
+
+    def visit_BLOB(self, type_, **kw):
+        if type_.length is not None:
+            return "BLOB(%d)" % type_.length
+        else:
+            return "BLOB"
+
+    def visit_TINYBLOB(self, type_, **kw):
+        return "TINYBLOB"
+
+    def visit_MEDIUMBLOB(self, type_, **kw):
+        return "MEDIUMBLOB"
+
+    def visit_LONGBLOB(self, type_, **kw):
+        return "LONGBLOB"
+
+    def _visit_enumerated_values(self, name, type_, enumerated_values):
+        quoted_enums = []
+        for e in enumerated_values:
+            if self.dialect.identifier_preparer._double_percents:
+                e = e.replace("%", "%%")
+            quoted_enums.append("'%s'" % e.replace("'", "''"))
+        return self._extend_string(
+            type_, {}, "%s(%s)" % (name, ",".join(quoted_enums))
+        )
+
+    def visit_ENUM(self, type_, **kw):
+        return self._visit_enumerated_values("ENUM", type_, type_.enums)
+
+    def visit_SET(self, type_, **kw):
+        return self._visit_enumerated_values("SET", type_, type_.values)
+
+    def visit_BOOLEAN(self, type_, **kw):
+        return "BOOL"
+
+
+class MySQLIdentifierPreparer(compiler.IdentifierPreparer):
+    reserved_words = RESERVED_WORDS_MYSQL
+
+    def __init__(self, dialect, server_ansiquotes=False, **kw):
+        if not server_ansiquotes:
+            quote = "`"
+        else:
+            quote = '"'
+
+        super().__init__(dialect, initial_quote=quote, escape_quote=quote)
+
+    def _quote_free_identifiers(self, *ids):
+        """Unilaterally identifier-quote any number of strings."""
+
+        return tuple([self.quote_identifier(i) for i in ids if i is not None])
+
+
+class MariaDBIdentifierPreparer(MySQLIdentifierPreparer):
+    reserved_words = RESERVED_WORDS_MARIADB
+
+
+@log.class_logger
+class MySQLDialect(default.DefaultDialect):
+    """Details of the MySQL dialect.
+    Not used directly in application code.
+    """
+
+    name = "mysql"
+    supports_statement_cache = True
+
+    supports_alter = True
+
+    # MySQL has no true "boolean" type; we
+    # allow for the "true" and "false" keywords, however
+    supports_native_boolean = False
+
+    # identifiers are 64, however aliases can be 255...
+    max_identifier_length = 255
+    max_index_name_length = 64
+    max_constraint_name_length = 64
+
+    div_is_floordiv = False
+
+    supports_native_enum = True
+
+    returns_native_bytes = True
+
+    supports_sequences = False  # default for MySQL ...
+    # ... may be updated to True for MariaDB 10.3+ in initialize()
+
+    sequences_optional = False
+
+    supports_for_update_of = False  # default for MySQL ...
+    # ... may be updated to True for MySQL 8+ in initialize()
+
+    _requires_alias_for_on_duplicate_key = False  # Only available ...
+    # ... in MySQL 8+
+
+    # MySQL doesn't support "DEFAULT VALUES" but *does* support
+    # "VALUES (DEFAULT)"
+    supports_default_values = False
+    supports_default_metavalue = True
+
+    use_insertmanyvalues: bool = True
+    insertmanyvalues_implicit_sentinel = (
+        InsertmanyvaluesSentinelOpts.ANY_AUTOINCREMENT
+    )
+
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = False
+    supports_multivalues_insert = True
+    insert_null_pk_still_autoincrements = True
+
+    supports_comments = True
+    inline_comments = True
+    default_paramstyle = "format"
+    colspecs = colspecs
+
+    cte_follows_insert = True
+
+    statement_compiler = MySQLCompiler
+    ddl_compiler = MySQLDDLCompiler
+    type_compiler_cls = MySQLTypeCompiler
+    ischema_names = ischema_names
+    preparer = MySQLIdentifierPreparer
+
+    is_mariadb = False
+    _mariadb_normalized_version_info = None
+
+    # default SQL compilation settings -
+    # these are modified upon initialize(),
+    # i.e. first connect
+    _backslash_escapes = True
+    _server_ansiquotes = False
+
+    construct_arguments = [
+        (sa_schema.Table, {"*": None}),
+        (sql.Update, {"limit": None}),
+        (sql.Delete, {"limit": None}),
+        (sa_schema.PrimaryKeyConstraint, {"using": None}),
+        (
+            sa_schema.Index,
+            {
+                "using": None,
+                "length": None,
+                "prefix": None,
+                "with_parser": None,
+            },
+        ),
+    ]
+
+    def __init__(
+        self,
+        json_serializer=None,
+        json_deserializer=None,
+        is_mariadb=None,
+        **kwargs,
+    ):
+        kwargs.pop("use_ansiquotes", None)  # legacy
+        default.DefaultDialect.__init__(self, **kwargs)
+        self._json_serializer = json_serializer
+        self._json_deserializer = json_deserializer
+        self._set_mariadb(is_mariadb, None)
+
+    def get_isolation_level_values(self, dbapi_conn):
+        return (
+            "SERIALIZABLE",
+            "READ UNCOMMITTED",
+            "READ COMMITTED",
+            "REPEATABLE READ",
+        )
+
+    def set_isolation_level(self, dbapi_connection, level):
+        cursor = dbapi_connection.cursor()
+        cursor.execute(f"SET SESSION TRANSACTION ISOLATION LEVEL {level}")
+        cursor.execute("COMMIT")
+        cursor.close()
+
+    def get_isolation_level(self, dbapi_connection):
+        cursor = dbapi_connection.cursor()
+        if self._is_mysql and self.server_version_info >= (5, 7, 20):
+            cursor.execute("SELECT @@transaction_isolation")
+        else:
+            cursor.execute("SELECT @@tx_isolation")
+        row = cursor.fetchone()
+        if row is None:
+            util.warn(
+                "Could not retrieve transaction isolation level for MySQL "
+                "connection."
+            )
+            raise NotImplementedError()
+        val = row[0]
+        cursor.close()
+        if isinstance(val, bytes):
+            val = val.decode()
+        return val.upper().replace("-", " ")
+
+    @classmethod
+    def _is_mariadb_from_url(cls, url):
+        dbapi = cls.import_dbapi()
+        dialect = cls(dbapi=dbapi)
+
+        cargs, cparams = dialect.create_connect_args(url)
+        conn = dialect.connect(*cargs, **cparams)
+        try:
+            cursor = conn.cursor()
+            cursor.execute("SELECT VERSION() LIKE '%MariaDB%'")
+            val = cursor.fetchone()[0]
+        except:
+            raise
+        else:
+            return bool(val)
+        finally:
+            conn.close()
+
+    def _get_server_version_info(self, connection):
+        # get database server version info explicitly over the wire
+        # to avoid proxy servers like MaxScale getting in the
+        # way with their own values, see #4205
+        dbapi_con = connection.connection
+        cursor = dbapi_con.cursor()
+        cursor.execute("SELECT VERSION()")
+        val = cursor.fetchone()[0]
+        cursor.close()
+        if isinstance(val, bytes):
+            val = val.decode()
+
+        return self._parse_server_version(val)
+
+    def _parse_server_version(self, val):
+        version = []
+        is_mariadb = False
+
+        r = re.compile(r"[.\-+]")
+        tokens = r.split(val)
+        for token in tokens:
+            parsed_token = re.match(
+                r"^(?:(\d+)(?:a|b|c)?|(MariaDB\w*))$", token
+            )
+            if not parsed_token:
+                continue
+            elif parsed_token.group(2):
+                self._mariadb_normalized_version_info = tuple(version[-3:])
+                is_mariadb = True
+            else:
+                digit = int(parsed_token.group(1))
+                version.append(digit)
+
+        server_version_info = tuple(version)
+
+        self._set_mariadb(
+            server_version_info and is_mariadb, server_version_info
+        )
+
+        if not is_mariadb:
+            self._mariadb_normalized_version_info = server_version_info
+
+        if server_version_info < (5, 0, 2):
+            raise NotImplementedError(
+                "the MySQL/MariaDB dialect supports server "
+                "version info 5.0.2 and above."
+            )
+
+        # setting it here to help w the test suite
+        self.server_version_info = server_version_info
+        return server_version_info
+
+    def _set_mariadb(self, is_mariadb, server_version_info):
+        if is_mariadb is None:
+            return
+
+        if not is_mariadb and self.is_mariadb:
+            raise exc.InvalidRequestError(
+                "MySQL version %s is not a MariaDB variant."
+                % (".".join(map(str, server_version_info)),)
+            )
+        if is_mariadb:
+            self.preparer = MariaDBIdentifierPreparer
+            # this would have been set by the default dialect already,
+            # so set it again
+            self.identifier_preparer = self.preparer(self)
+
+            # this will be updated on first connect in initialize()
+            # if using older mariadb version
+            self.delete_returning = True
+            self.insert_returning = True
+
+        self.is_mariadb = is_mariadb
+
+    def do_begin_twophase(self, connection, xid):
+        connection.execute(sql.text("XA BEGIN :xid"), dict(xid=xid))
+
+    def do_prepare_twophase(self, connection, xid):
+        connection.execute(sql.text("XA END :xid"), dict(xid=xid))
+        connection.execute(sql.text("XA PREPARE :xid"), dict(xid=xid))
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if not is_prepared:
+            connection.execute(sql.text("XA END :xid"), dict(xid=xid))
+        connection.execute(sql.text("XA ROLLBACK :xid"), dict(xid=xid))
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if not is_prepared:
+            self.do_prepare_twophase(connection, xid)
+        connection.execute(sql.text("XA COMMIT :xid"), dict(xid=xid))
+
+    def do_recover_twophase(self, connection):
+        resultset = connection.exec_driver_sql("XA RECOVER")
+        return [row["data"][0 : row["gtrid_length"]] for row in resultset]
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(
+            e,
+            (
+                self.dbapi.OperationalError,
+                self.dbapi.ProgrammingError,
+                self.dbapi.InterfaceError,
+            ),
+        ) and self._extract_error_code(e) in (
+            1927,
+            2006,
+            2013,
+            2014,
+            2045,
+            2055,
+            4031,
+        ):
+            return True
+        elif isinstance(
+            e, (self.dbapi.InterfaceError, self.dbapi.InternalError)
+        ):
+            # if underlying connection is closed,
+            # this is the error you get
+            return "(0, '')" in str(e)
+        else:
+            return False
+
+    def _compat_fetchall(self, rp, charset=None):
+        """Proxy result rows to smooth over MySQL-Python driver
+        inconsistencies."""
+
+        return [_DecodingRow(row, charset) for row in rp.fetchall()]
+
+    def _compat_fetchone(self, rp, charset=None):
+        """Proxy a result row to smooth over MySQL-Python driver
+        inconsistencies."""
+
+        row = rp.fetchone()
+        if row:
+            return _DecodingRow(row, charset)
+        else:
+            return None
+
+    def _compat_first(self, rp, charset=None):
+        """Proxy a result row to smooth over MySQL-Python driver
+        inconsistencies."""
+
+        row = rp.first()
+        if row:
+            return _DecodingRow(row, charset)
+        else:
+            return None
+
+    def _extract_error_code(self, exception):
+        raise NotImplementedError()
+
+    def _get_default_schema_name(self, connection):
+        return connection.exec_driver_sql("SELECT DATABASE()").scalar()
+
+    @reflection.cache
+    def has_table(self, connection, table_name, schema=None, **kw):
+        self._ensure_has_table_connection(connection)
+
+        if schema is None:
+            schema = self.default_schema_name
+
+        assert schema is not None
+
+        full_name = ".".join(
+            self.identifier_preparer._quote_free_identifiers(
+                schema, table_name
+            )
+        )
+
+        # DESCRIBE *must* be used because there is no information schema
+        # table that returns information on temp tables that is consistently
+        # available on MariaDB / MySQL / engine-agnostic etc.
+        # therefore we have no choice but to use DESCRIBE and an error catch
+        # to detect "False".  See issue #9058
+
+        try:
+            with connection.exec_driver_sql(
+                f"DESCRIBE {full_name}",
+                execution_options={"skip_user_error_events": True},
+            ) as rs:
+                return rs.fetchone() is not None
+        except exc.DBAPIError as e:
+            # https://dev.mysql.com/doc/mysql-errors/8.0/en/server-error-reference.html  # noqa: E501
+            # there are a lot of codes that *may* pop up here at some point
+            # but we continue to be fairly conservative.  We include:
+            # 1146: Table '%s.%s' doesn't exist - what every MySQL has emitted
+            # for decades
+            #
+            # mysql 8 suddenly started emitting:
+            # 1049: Unknown database '%s'  - for nonexistent schema
+            #
+            # also added:
+            # 1051: Unknown table '%s' - not known to emit
+            #
+            # there's more "doesn't exist" kinds of messages but they are
+            # less clear if mysql 8 would suddenly start using one of those
+            if self._extract_error_code(e.orig) in (1146, 1049, 1051):
+                return False
+            raise
+
+    @reflection.cache
+    def has_sequence(self, connection, sequence_name, schema=None, **kw):
+        if not self.supports_sequences:
+            self._sequences_not_supported()
+        if not schema:
+            schema = self.default_schema_name
+        # MariaDB implements sequences as a special type of table
+        #
+        cursor = connection.execute(
+            sql.text(
+                "SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES "
+                "WHERE TABLE_TYPE='SEQUENCE' and TABLE_NAME=:name AND "
+                "TABLE_SCHEMA=:schema_name"
+            ),
+            dict(
+                name=str(sequence_name),
+                schema_name=str(schema),
+            ),
+        )
+        return cursor.first() is not None
+
+    def _sequences_not_supported(self):
+        raise NotImplementedError(
+            "Sequences are supported only by the "
+            "MariaDB series 10.3 or greater"
+        )
+
+    @reflection.cache
+    def get_sequence_names(self, connection, schema=None, **kw):
+        if not self.supports_sequences:
+            self._sequences_not_supported()
+        if not schema:
+            schema = self.default_schema_name
+        # MariaDB implements sequences as a special type of table
+        cursor = connection.execute(
+            sql.text(
+                "SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES "
+                "WHERE TABLE_TYPE='SEQUENCE' and TABLE_SCHEMA=:schema_name"
+            ),
+            dict(schema_name=schema),
+        )
+        return [
+            row[0]
+            for row in self._compat_fetchall(
+                cursor, charset=self._connection_charset
+            )
+        ]
+
+    def initialize(self, connection):
+        # this is driver-based, does not need server version info
+        # and is fairly critical for even basic SQL operations
+        self._connection_charset = self._detect_charset(connection)
+
+        # call super().initialize() because we need to have
+        # server_version_info set up.  in 1.4 under python 2 only this does the
+        # "check unicode returns" thing, which is the one area that some
+        # SQL gets compiled within initialize() currently
+        default.DefaultDialect.initialize(self, connection)
+
+        self._detect_sql_mode(connection)
+        self._detect_ansiquotes(connection)  # depends on sql mode
+        self._detect_casing(connection)
+        if self._server_ansiquotes:
+            # if ansiquotes == True, build a new IdentifierPreparer
+            # with the new setting
+            self.identifier_preparer = self.preparer(
+                self, server_ansiquotes=self._server_ansiquotes
+            )
+
+        self.supports_sequences = (
+            self.is_mariadb and self.server_version_info >= (10, 3)
+        )
+
+        self.supports_for_update_of = (
+            self._is_mysql and self.server_version_info >= (8,)
+        )
+
+        self._needs_correct_for_88718_96365 = (
+            not self.is_mariadb and self.server_version_info >= (8,)
+        )
+
+        self.delete_returning = (
+            self.is_mariadb and self.server_version_info >= (10, 0, 5)
+        )
+
+        self.insert_returning = (
+            self.is_mariadb and self.server_version_info >= (10, 5)
+        )
+
+        self._requires_alias_for_on_duplicate_key = (
+            self._is_mysql and self.server_version_info >= (8, 0, 20)
+        )
+
+        self._warn_for_known_db_issues()
+
+    def _warn_for_known_db_issues(self):
+        if self.is_mariadb:
+            mdb_version = self._mariadb_normalized_version_info
+            if mdb_version > (10, 2) and mdb_version < (10, 2, 9):
+                util.warn(
+                    "MariaDB %r before 10.2.9 has known issues regarding "
+                    "CHECK constraints, which impact handling of NULL values "
+                    "with SQLAlchemy's boolean datatype (MDEV-13596). An "
+                    "additional issue prevents proper migrations of columns "
+                    "with CHECK constraints (MDEV-11114).  Please upgrade to "
+                    "MariaDB 10.2.9 or greater, or use the MariaDB 10.1 "
+                    "series, to avoid these issues." % (mdb_version,)
+                )
+
+    @property
+    def _support_float_cast(self):
+        if not self.server_version_info:
+            return False
+        elif self.is_mariadb:
+            # ref https://mariadb.com/kb/en/mariadb-1045-release-notes/
+            return self.server_version_info >= (10, 4, 5)
+        else:
+            # ref https://dev.mysql.com/doc/relnotes/mysql/8.0/en/news-8-0-17.html#mysqld-8-0-17-feature  # noqa
+            return self.server_version_info >= (8, 0, 17)
+
+    @property
+    def _support_default_function(self):
+        if not self.server_version_info:
+            return False
+        elif self.is_mariadb:
+            # ref https://mariadb.com/kb/en/mariadb-1021-release-notes/
+            return self.server_version_info >= (10, 2, 1)
+        else:
+            # ref https://dev.mysql.com/doc/refman/8.0/en/data-type-defaults.html # noqa
+            return self.server_version_info >= (8, 0, 13)
+
+    @property
+    def _is_mariadb(self):
+        return self.is_mariadb
+
+    @property
+    def _is_mysql(self):
+        return not self.is_mariadb
+
+    @property
+    def _is_mariadb_102(self):
+        return self.is_mariadb and self._mariadb_normalized_version_info > (
+            10,
+            2,
+        )
+
+    @reflection.cache
+    def get_schema_names(self, connection, **kw):
+        rp = connection.exec_driver_sql("SHOW schemas")
+        return [r[0] for r in rp]
+
+    @reflection.cache
+    def get_table_names(self, connection, schema=None, **kw):
+        """Return a Unicode SHOW TABLES from a given schema."""
+        if schema is not None:
+            current_schema = schema
+        else:
+            current_schema = self.default_schema_name
+
+        charset = self._connection_charset
+
+        rp = connection.exec_driver_sql(
+            "SHOW FULL TABLES FROM %s"
+            % self.identifier_preparer.quote_identifier(current_schema)
+        )
+
+        return [
+            row[0]
+            for row in self._compat_fetchall(rp, charset=charset)
+            if row[1] == "BASE TABLE"
+        ]
+
+    @reflection.cache
+    def get_view_names(self, connection, schema=None, **kw):
+        if schema is None:
+            schema = self.default_schema_name
+        charset = self._connection_charset
+        rp = connection.exec_driver_sql(
+            "SHOW FULL TABLES FROM %s"
+            % self.identifier_preparer.quote_identifier(schema)
+        )
+        return [
+            row[0]
+            for row in self._compat_fetchall(rp, charset=charset)
+            if row[1] in ("VIEW", "SYSTEM VIEW")
+        ]
+
+    @reflection.cache
+    def get_table_options(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+        if parsed_state.table_options:
+            return parsed_state.table_options
+        else:
+            return ReflectionDefaults.table_options()
+
+    @reflection.cache
+    def get_columns(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+        if parsed_state.columns:
+            return parsed_state.columns
+        else:
+            return ReflectionDefaults.columns()
+
+    @reflection.cache
+    def get_pk_constraint(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+        for key in parsed_state.keys:
+            if key["type"] == "PRIMARY":
+                # There can be only one.
+                cols = [s[0] for s in key["columns"]]
+                return {"constrained_columns": cols, "name": None}
+        return ReflectionDefaults.pk_constraint()
+
+    @reflection.cache
+    def get_foreign_keys(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+        default_schema = None
+
+        fkeys = []
+
+        for spec in parsed_state.fk_constraints:
+            ref_name = spec["table"][-1]
+            ref_schema = len(spec["table"]) > 1 and spec["table"][-2] or schema
+
+            if not ref_schema:
+                if default_schema is None:
+                    default_schema = connection.dialect.default_schema_name
+                if schema == default_schema:
+                    ref_schema = schema
+
+            loc_names = spec["local"]
+            ref_names = spec["foreign"]
+
+            con_kw = {}
+            for opt in ("onupdate", "ondelete"):
+                if spec.get(opt, False) not in ("NO ACTION", None):
+                    con_kw[opt] = spec[opt]
+
+            fkey_d = {
+                "name": spec["name"],
+                "constrained_columns": loc_names,
+                "referred_schema": ref_schema,
+                "referred_table": ref_name,
+                "referred_columns": ref_names,
+                "options": con_kw,
+            }
+            fkeys.append(fkey_d)
+
+        if self._needs_correct_for_88718_96365:
+            self._correct_for_mysql_bugs_88718_96365(fkeys, connection)
+
+        return fkeys if fkeys else ReflectionDefaults.foreign_keys()
+
+    def _correct_for_mysql_bugs_88718_96365(self, fkeys, connection):
+        # Foreign key is always in lower case (MySQL 8.0)
+        # https://bugs.mysql.com/bug.php?id=88718
+        # issue #4344 for SQLAlchemy
+
+        # table name also for MySQL 8.0
+        # https://bugs.mysql.com/bug.php?id=96365
+        # issue #4751 for SQLAlchemy
+
+        # for lower_case_table_names=2, information_schema.columns
+        # preserves the original table/schema casing, but SHOW CREATE
+        # TABLE does not.   this problem is not in lower_case_table_names=1,
+        # but use case-insensitive matching for these two modes in any case.
+
+        if self._casing in (1, 2):
+
+            def lower(s):
+                return s.lower()
+
+        else:
+            # if on case sensitive, there can be two tables referenced
+            # with the same name different casing, so we need to use
+            # case-sensitive matching.
+            def lower(s):
+                return s
+
+        default_schema_name = connection.dialect.default_schema_name
+
+        # NOTE: using (table_schema, table_name, lower(column_name)) in (...)
+        # is very slow since mysql does not seem able to properly use indexse.
+        # Unpack the where condition instead.
+        schema_by_table_by_column = defaultdict(lambda: defaultdict(list))
+        for rec in fkeys:
+            sch = lower(rec["referred_schema"] or default_schema_name)
+            tbl = lower(rec["referred_table"])
+            for col_name in rec["referred_columns"]:
+                schema_by_table_by_column[sch][tbl].append(col_name)
+
+        if schema_by_table_by_column:
+
+            condition = sql.or_(
+                *(
+                    sql.and_(
+                        _info_columns.c.table_schema == schema,
+                        sql.or_(
+                            *(
+                                sql.and_(
+                                    _info_columns.c.table_name == table,
+                                    sql.func.lower(
+                                        _info_columns.c.column_name
+                                    ).in_(columns),
+                                )
+                                for table, columns in tables.items()
+                            )
+                        ),
+                    )
+                    for schema, tables in schema_by_table_by_column.items()
+                )
+            )
+
+            select = sql.select(
+                _info_columns.c.table_schema,
+                _info_columns.c.table_name,
+                _info_columns.c.column_name,
+            ).where(condition)
+
+            correct_for_wrong_fk_case = connection.execute(select)
+
+            # in casing=0, table name and schema name come back in their
+            # exact case.
+            # in casing=1, table name and schema name come back in lower
+            # case.
+            # in casing=2, table name and schema name come back from the
+            # information_schema.columns view in the case
+            # that was used in CREATE DATABASE and CREATE TABLE, but
+            # SHOW CREATE TABLE converts them to *lower case*, therefore
+            # not matching.  So for this case, case-insensitive lookup
+            # is necessary
+            d = defaultdict(dict)
+            for schema, tname, cname in correct_for_wrong_fk_case:
+                d[(lower(schema), lower(tname))]["SCHEMANAME"] = schema
+                d[(lower(schema), lower(tname))]["TABLENAME"] = tname
+                d[(lower(schema), lower(tname))][cname.lower()] = cname
+
+            for fkey in fkeys:
+                rec = d[
+                    (
+                        lower(fkey["referred_schema"] or default_schema_name),
+                        lower(fkey["referred_table"]),
+                    )
+                ]
+
+                fkey["referred_table"] = rec["TABLENAME"]
+                if fkey["referred_schema"] is not None:
+                    fkey["referred_schema"] = rec["SCHEMANAME"]
+
+                fkey["referred_columns"] = [
+                    rec[col.lower()] for col in fkey["referred_columns"]
+                ]
+
+    @reflection.cache
+    def get_check_constraints(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+
+        cks = [
+            {"name": spec["name"], "sqltext": spec["sqltext"]}
+            for spec in parsed_state.ck_constraints
+        ]
+        cks.sort(key=lambda d: d["name"] or "~")  # sort None as last
+        return cks if cks else ReflectionDefaults.check_constraints()
+
+    @reflection.cache
+    def get_table_comment(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+        comment = parsed_state.table_options.get(f"{self.name}_comment", None)
+        if comment is not None:
+            return {"text": comment}
+        else:
+            return ReflectionDefaults.table_comment()
+
+    @reflection.cache
+    def get_indexes(self, connection, table_name, schema=None, **kw):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+
+        indexes = []
+
+        for spec in parsed_state.keys:
+            dialect_options = {}
+            unique = False
+            flavor = spec["type"]
+            if flavor == "PRIMARY":
+                continue
+            if flavor == "UNIQUE":
+                unique = True
+            elif flavor in ("FULLTEXT", "SPATIAL"):
+                dialect_options["%s_prefix" % self.name] = flavor
+            elif flavor is None:
+                pass
+            else:
+                self.logger.info(
+                    "Converting unknown KEY type %s to a plain KEY", flavor
+                )
+                pass
+
+            if spec["parser"]:
+                dialect_options["%s_with_parser" % (self.name)] = spec[
+                    "parser"
+                ]
+
+            index_d = {}
+
+            index_d["name"] = spec["name"]
+            index_d["column_names"] = [s[0] for s in spec["columns"]]
+            mysql_length = {
+                s[0]: s[1] for s in spec["columns"] if s[1] is not None
+            }
+            if mysql_length:
+                dialect_options["%s_length" % self.name] = mysql_length
+
+            index_d["unique"] = unique
+            if flavor:
+                index_d["type"] = flavor
+
+            if dialect_options:
+                index_d["dialect_options"] = dialect_options
+
+            indexes.append(index_d)
+        indexes.sort(key=lambda d: d["name"] or "~")  # sort None as last
+        return indexes if indexes else ReflectionDefaults.indexes()
+
+    @reflection.cache
+    def get_unique_constraints(
+        self, connection, table_name, schema=None, **kw
+    ):
+        parsed_state = self._parsed_state_or_create(
+            connection, table_name, schema, **kw
+        )
+
+        ucs = [
+            {
+                "name": key["name"],
+                "column_names": [col[0] for col in key["columns"]],
+                "duplicates_index": key["name"],
+            }
+            for key in parsed_state.keys
+            if key["type"] == "UNIQUE"
+        ]
+        ucs.sort(key=lambda d: d["name"] or "~")  # sort None as last
+        if ucs:
+            return ucs
+        else:
+            return ReflectionDefaults.unique_constraints()
+
+    @reflection.cache
+    def get_view_definition(self, connection, view_name, schema=None, **kw):
+        charset = self._connection_charset
+        full_name = ".".join(
+            self.identifier_preparer._quote_free_identifiers(schema, view_name)
+        )
+        sql = self._show_create_table(
+            connection, None, charset, full_name=full_name
+        )
+        if sql.upper().startswith("CREATE TABLE"):
+            # it's a table, not a view
+            raise exc.NoSuchTableError(full_name)
+        return sql
+
+    def _parsed_state_or_create(
+        self, connection, table_name, schema=None, **kw
+    ):
+        return self._setup_parser(
+            connection,
+            table_name,
+            schema,
+            info_cache=kw.get("info_cache", None),
+        )
+
+    @util.memoized_property
+    def _tabledef_parser(self):
+        """return the MySQLTableDefinitionParser, generate if needed.
+
+        The deferred creation ensures that the dialect has
+        retrieved server version information first.
+
+        """
+        preparer = self.identifier_preparer
+        return _reflection.MySQLTableDefinitionParser(self, preparer)
+
+    @reflection.cache
+    def _setup_parser(self, connection, table_name, schema=None, **kw):
+        charset = self._connection_charset
+        parser = self._tabledef_parser
+        full_name = ".".join(
+            self.identifier_preparer._quote_free_identifiers(
+                schema, table_name
+            )
+        )
+        sql = self._show_create_table(
+            connection, None, charset, full_name=full_name
+        )
+        if parser._check_view(sql):
+            # Adapt views to something table-like.
+            columns = self._describe_table(
+                connection, None, charset, full_name=full_name
+            )
+            sql = parser._describe_to_create(table_name, columns)
+        return parser.parse(sql, charset)
+
+    def _fetch_setting(self, connection, setting_name):
+        charset = self._connection_charset
+
+        if self.server_version_info and self.server_version_info < (5, 6):
+            sql = "SHOW VARIABLES LIKE '%s'" % setting_name
+            fetch_col = 1
+        else:
+            sql = "SELECT @@%s" % setting_name
+            fetch_col = 0
+
+        show_var = connection.exec_driver_sql(sql)
+        row = self._compat_first(show_var, charset=charset)
+        if not row:
+            return None
+        else:
+            return row[fetch_col]
+
+    def _detect_charset(self, connection):
+        raise NotImplementedError()
+
+    def _detect_casing(self, connection):
+        """Sniff out identifier case sensitivity.
+
+        Cached per-connection. This value can not change without a server
+        restart.
+
+        """
+        # https://dev.mysql.com/doc/refman/en/identifier-case-sensitivity.html
+
+        setting = self._fetch_setting(connection, "lower_case_table_names")
+        if setting is None:
+            cs = 0
+        else:
+            # 4.0.15 returns OFF or ON according to [ticket:489]
+            # 3.23 doesn't, 4.0.27 doesn't..
+            if setting == "OFF":
+                cs = 0
+            elif setting == "ON":
+                cs = 1
+            else:
+                cs = int(setting)
+        self._casing = cs
+        return cs
+
+    def _detect_collations(self, connection):
+        """Pull the active COLLATIONS list from the server.
+
+        Cached per-connection.
+        """
+
+        collations = {}
+        charset = self._connection_charset
+        rs = connection.exec_driver_sql("SHOW COLLATION")
+        for row in self._compat_fetchall(rs, charset):
+            collations[row[0]] = row[1]
+        return collations
+
+    def _detect_sql_mode(self, connection):
+        setting = self._fetch_setting(connection, "sql_mode")
+
+        if setting is None:
+            util.warn(
+                "Could not retrieve SQL_MODE; please ensure the "
+                "MySQL user has permissions to SHOW VARIABLES"
+            )
+            self._sql_mode = ""
+        else:
+            self._sql_mode = setting or ""
+
+    def _detect_ansiquotes(self, connection):
+        """Detect and adjust for the ANSI_QUOTES sql mode."""
+
+        mode = self._sql_mode
+        if not mode:
+            mode = ""
+        elif mode.isdigit():
+            mode_no = int(mode)
+            mode = (mode_no | 4 == mode_no) and "ANSI_QUOTES" or ""
+
+        self._server_ansiquotes = "ANSI_QUOTES" in mode
+
+        # as of MySQL 5.0.1
+        self._backslash_escapes = "NO_BACKSLASH_ESCAPES" not in mode
+
+    def _show_create_table(
+        self, connection, table, charset=None, full_name=None
+    ):
+        """Run SHOW CREATE TABLE for a ``Table``."""
+
+        if full_name is None:
+            full_name = self.identifier_preparer.format_table(table)
+        st = "SHOW CREATE TABLE %s" % full_name
+
+        rp = None
+        try:
+            rp = connection.execution_options(
+                skip_user_error_events=True
+            ).exec_driver_sql(st)
+        except exc.DBAPIError as e:
+            if self._extract_error_code(e.orig) == 1146:
+                raise exc.NoSuchTableError(full_name) from e
+            else:
+                raise
+        row = self._compat_first(rp, charset=charset)
+        if not row:
+            raise exc.NoSuchTableError(full_name)
+        return row[1].strip()
+
+    def _describe_table(self, connection, table, charset=None, full_name=None):
+        """Run DESCRIBE for a ``Table`` and return processed rows."""
+
+        if full_name is None:
+            full_name = self.identifier_preparer.format_table(table)
+        st = "DESCRIBE %s" % full_name
+
+        rp, rows = None, None
+        try:
+            try:
+                rp = connection.execution_options(
+                    skip_user_error_events=True
+                ).exec_driver_sql(st)
+            except exc.DBAPIError as e:
+                code = self._extract_error_code(e.orig)
+                if code == 1146:
+                    raise exc.NoSuchTableError(full_name) from e
+
+                elif code == 1356:
+                    raise exc.UnreflectableTableError(
+                        "Table or view named %s could not be "
+                        "reflected: %s" % (full_name, e)
+                    ) from e
+
+                else:
+                    raise
+            rows = self._compat_fetchall(rp, charset=charset)
+        finally:
+            if rp:
+                rp.close()
+        return rows
+
+
+class _DecodingRow:
+    """Return unicode-decoded values based on type inspection.
+
+    Smooth over data type issues (esp. with alpha driver versions) and
+    normalize strings as Unicode regardless of user-configured driver
+    encoding settings.
+
+    """
+
+    # Some MySQL-python versions can return some columns as
+    # sets.Set(['value']) (seriously) but thankfully that doesn't
+    # seem to come up in DDL queries.
+
+    _encoding_compat = {
+        "koi8r": "koi8_r",
+        "koi8u": "koi8_u",
+        "utf16": "utf-16-be",  # MySQL's uft16 is always bigendian
+        "utf8mb4": "utf8",  # real utf8
+        "utf8mb3": "utf8",  # real utf8; saw this happen on CI but I cannot
+        # reproduce, possibly mariadb10.6 related
+        "eucjpms": "ujis",
+    }
+
+    def __init__(self, rowproxy, charset):
+        self.rowproxy = rowproxy
+        self.charset = self._encoding_compat.get(charset, charset)
+
+    def __getitem__(self, index):
+        item = self.rowproxy[index]
+        if isinstance(item, _array):
+            item = item.tostring()
+
+        if self.charset and isinstance(item, bytes):
+            return item.decode(self.charset)
+        else:
+            return item
+
+    def __getattr__(self, attr):
+        item = getattr(self.rowproxy, attr)
+        if isinstance(item, _array):
+            item = item.tostring()
+        if self.charset and isinstance(item, bytes):
+            return item.decode(self.charset)
+        else:
+            return item
+
+
+_info_columns = sql.table(
+    "columns",
+    sql.column("table_schema", VARCHAR(64)),
+    sql.column("table_name", VARCHAR(64)),
+    sql.column("column_name", VARCHAR(64)),
+    schema="information_schema",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/cymysql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/cymysql.py
new file mode 100644
index 00000000..5c00ada9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/cymysql.py
@@ -0,0 +1,84 @@
+# dialects/mysql/cymysql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+
+.. dialect:: mysql+cymysql
+    :name: CyMySQL
+    :dbapi: cymysql
+    :connectstring: mysql+cymysql://<username>:<password>@<host>/<dbname>[?<options>]
+    :url: https://github.com/nakagami/CyMySQL
+
+.. note::
+
+    The CyMySQL dialect is **not tested as part of SQLAlchemy's continuous
+    integration** and may have unresolved issues.  The recommended MySQL
+    dialects are mysqlclient and PyMySQL.
+
+"""  # noqa
+
+from .base import BIT
+from .base import MySQLDialect
+from .mysqldb import MySQLDialect_mysqldb
+from ... import util
+
+
+class _cymysqlBIT(BIT):
+    def result_processor(self, dialect, coltype):
+        """Convert MySQL's 64 bit, variable length binary string to a long."""
+
+        def process(value):
+            if value is not None:
+                v = 0
+                for i in iter(value):
+                    v = v << 8 | i
+                return v
+            return value
+
+        return process
+
+
+class MySQLDialect_cymysql(MySQLDialect_mysqldb):
+    driver = "cymysql"
+    supports_statement_cache = True
+
+    description_encoding = None
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = False
+    supports_unicode_statements = True
+
+    colspecs = util.update_copy(MySQLDialect.colspecs, {BIT: _cymysqlBIT})
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("cymysql")
+
+    def _detect_charset(self, connection):
+        return connection.connection.charset
+
+    def _extract_error_code(self, exception):
+        return exception.errno
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(e, self.dbapi.OperationalError):
+            return self._extract_error_code(e) in (
+                2006,
+                2013,
+                2014,
+                2045,
+                2055,
+            )
+        elif isinstance(e, self.dbapi.InterfaceError):
+            # if underlying connection is closed,
+            # this is the error you get
+            return True
+        else:
+            return False
+
+
+dialect = MySQLDialect_cymysql
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/dml.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/dml.py
new file mode 100644
index 00000000..cceb0818
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/dml.py
@@ -0,0 +1,225 @@
+# dialects/mysql/dml.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+from typing import Any
+from typing import Dict
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import Tuple
+from typing import Union
+
+from ... import exc
+from ... import util
+from ...sql._typing import _DMLTableArgument
+from ...sql.base import _exclusive_against
+from ...sql.base import _generative
+from ...sql.base import ColumnCollection
+from ...sql.base import ReadOnlyColumnCollection
+from ...sql.dml import Insert as StandardInsert
+from ...sql.elements import ClauseElement
+from ...sql.elements import KeyedColumnElement
+from ...sql.expression import alias
+from ...sql.selectable import NamedFromClause
+from ...util.typing import Self
+
+
+__all__ = ("Insert", "insert")
+
+
+def insert(table: _DMLTableArgument) -> Insert:
+    """Construct a MySQL/MariaDB-specific variant :class:`_mysql.Insert`
+    construct.
+
+    .. container:: inherited_member
+
+        The :func:`sqlalchemy.dialects.mysql.insert` function creates
+        a :class:`sqlalchemy.dialects.mysql.Insert`.  This class is based
+        on the dialect-agnostic :class:`_sql.Insert` construct which may
+        be constructed using the :func:`_sql.insert` function in
+        SQLAlchemy Core.
+
+    The :class:`_mysql.Insert` construct includes additional methods
+    :meth:`_mysql.Insert.on_duplicate_key_update`.
+
+    """
+    return Insert(table)
+
+
+class Insert(StandardInsert):
+    """MySQL-specific implementation of INSERT.
+
+    Adds methods for MySQL-specific syntaxes such as ON DUPLICATE KEY UPDATE.
+
+    The :class:`~.mysql.Insert` object is created using the
+    :func:`sqlalchemy.dialects.mysql.insert` function.
+
+    .. versionadded:: 1.2
+
+    """
+
+    stringify_dialect = "mysql"
+    inherit_cache = False
+
+    @property
+    def inserted(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """Provide the "inserted" namespace for an ON DUPLICATE KEY UPDATE
+        statement
+
+        MySQL's ON DUPLICATE KEY UPDATE clause allows reference to the row
+        that would be inserted, via a special function called ``VALUES()``.
+        This attribute provides all columns in this row to be referenceable
+        such that they will render within a ``VALUES()`` function inside the
+        ON DUPLICATE KEY UPDATE clause.    The attribute is named ``.inserted``
+        so as not to conflict with the existing
+        :meth:`_expression.Insert.values` method.
+
+        .. tip::  The :attr:`_mysql.Insert.inserted` attribute is an instance
+            of :class:`_expression.ColumnCollection`, which provides an
+            interface the same as that of the :attr:`_schema.Table.c`
+            collection described at :ref:`metadata_tables_and_columns`.
+            With this collection, ordinary names are accessible like attributes
+            (e.g. ``stmt.inserted.some_column``), but special names and
+            dictionary method names should be accessed using indexed access,
+            such as ``stmt.inserted["column name"]`` or
+            ``stmt.inserted["values"]``.  See the docstring for
+            :class:`_expression.ColumnCollection` for further examples.
+
+        .. seealso::
+
+            :ref:`mysql_insert_on_duplicate_key_update` - example of how
+            to use :attr:`_expression.Insert.inserted`
+
+        """
+        return self.inserted_alias.columns
+
+    @util.memoized_property
+    def inserted_alias(self) -> NamedFromClause:
+        return alias(self.table, name="inserted")
+
+    @_generative
+    @_exclusive_against(
+        "_post_values_clause",
+        msgs={
+            "_post_values_clause": "This Insert construct already "
+            "has an ON DUPLICATE KEY clause present"
+        },
+    )
+    def on_duplicate_key_update(self, *args: _UpdateArg, **kw: Any) -> Self:
+        r"""
+        Specifies the ON DUPLICATE KEY UPDATE clause.
+
+        :param \**kw:  Column keys linked to UPDATE values.  The
+         values may be any SQL expression or supported literal Python
+         values.
+
+        .. warning:: This dictionary does **not** take into account
+           Python-specified default UPDATE values or generation functions,
+           e.g. those specified using :paramref:`_schema.Column.onupdate`.
+           These values will not be exercised for an ON DUPLICATE KEY UPDATE
+           style of UPDATE, unless values are manually specified here.
+
+        :param \*args: As an alternative to passing key/value parameters,
+         a dictionary or list of 2-tuples can be passed as a single positional
+         argument.
+
+         Passing a single dictionary is equivalent to the keyword argument
+         form::
+
+            insert().on_duplicate_key_update({"name": "some name"})
+
+         Passing a list of 2-tuples indicates that the parameter assignments
+         in the UPDATE clause should be ordered as sent, in a manner similar
+         to that described for the :class:`_expression.Update`
+         construct overall
+         in :ref:`tutorial_parameter_ordered_updates`::
+
+            insert().on_duplicate_key_update(
+                [
+                    ("name", "some name"),
+                    ("value", "some value"),
+                ]
+            )
+
+         .. versionchanged:: 1.3 parameters can be specified as a dictionary
+            or list of 2-tuples; the latter form provides for parameter
+            ordering.
+
+
+        .. versionadded:: 1.2
+
+        .. seealso::
+
+            :ref:`mysql_insert_on_duplicate_key_update`
+
+        """
+        if args and kw:
+            raise exc.ArgumentError(
+                "Can't pass kwargs and positional arguments simultaneously"
+            )
+
+        if args:
+            if len(args) > 1:
+                raise exc.ArgumentError(
+                    "Only a single dictionary or list of tuples "
+                    "is accepted positionally."
+                )
+            values = args[0]
+        else:
+            values = kw
+
+        self._post_values_clause = OnDuplicateClause(
+            self.inserted_alias, values
+        )
+        return self
+
+
+class OnDuplicateClause(ClauseElement):
+    __visit_name__ = "on_duplicate_key_update"
+
+    _parameter_ordering: Optional[List[str]] = None
+
+    update: Dict[str, Any]
+    stringify_dialect = "mysql"
+
+    def __init__(
+        self, inserted_alias: NamedFromClause, update: _UpdateArg
+    ) -> None:
+        self.inserted_alias = inserted_alias
+
+        # auto-detect that parameters should be ordered.   This is copied from
+        # Update._proces_colparams(), however we don't look for a special flag
+        # in this case since we are not disambiguating from other use cases as
+        # we are in Update.values().
+        if isinstance(update, list) and (
+            update and isinstance(update[0], tuple)
+        ):
+            self._parameter_ordering = [key for key, value in update]
+            update = dict(update)
+
+        if isinstance(update, dict):
+            if not update:
+                raise ValueError(
+                    "update parameter dictionary must not be empty"
+                )
+        elif isinstance(update, ColumnCollection):
+            update = dict(update)
+        else:
+            raise ValueError(
+                "update parameter must be a non-empty dictionary "
+                "or a ColumnCollection such as the `.c.` collection "
+                "of a Table object"
+            )
+        self.update = update
+
+
+_UpdateArg = Union[
+    Mapping[Any, Any], List[Tuple[str, Any]], ColumnCollection[Any, Any]
+]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/enumerated.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/enumerated.py
new file mode 100644
index 00000000..6745cae5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/enumerated.py
@@ -0,0 +1,243 @@
+# dialects/mysql/enumerated.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import re
+
+from .types import _StringType
+from ... import exc
+from ... import sql
+from ... import util
+from ...sql import sqltypes
+
+
+class ENUM(sqltypes.NativeForEmulated, sqltypes.Enum, _StringType):
+    """MySQL ENUM type."""
+
+    __visit_name__ = "ENUM"
+
+    native_enum = True
+
+    def __init__(self, *enums, **kw):
+        """Construct an ENUM.
+
+        E.g.::
+
+          Column("myenum", ENUM("foo", "bar", "baz"))
+
+        :param enums: The range of valid values for this ENUM.  Values in
+          enums are not quoted, they will be escaped and surrounded by single
+          quotes when generating the schema.  This object may also be a
+          PEP-435-compliant enumerated type.
+
+          .. versionadded: 1.1 added support for PEP-435-compliant enumerated
+             types.
+
+        :param strict: This flag has no effect.
+
+         .. versionchanged:: The MySQL ENUM type as well as the base Enum
+            type now validates all Python data values.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        kw.pop("strict", None)
+        self._enum_init(enums, kw)
+        _StringType.__init__(self, length=self.length, **kw)
+
+    @classmethod
+    def adapt_emulated_to_native(cls, impl, **kw):
+        """Produce a MySQL native :class:`.mysql.ENUM` from plain
+        :class:`.Enum`.
+
+        """
+        kw.setdefault("validate_strings", impl.validate_strings)
+        kw.setdefault("values_callable", impl.values_callable)
+        kw.setdefault("omit_aliases", impl._omit_aliases)
+        return cls(**kw)
+
+    def _object_value_for_elem(self, elem):
+        # mysql sends back a blank string for any value that
+        # was persisted that was not in the enums; that is, it does no
+        # validation on the incoming data, it "truncates" it to be
+        # the blank string.  Return it straight.
+        if elem == "":
+            return elem
+        else:
+            return super()._object_value_for_elem(elem)
+
+    def __repr__(self):
+        return util.generic_repr(
+            self, to_inspect=[ENUM, _StringType, sqltypes.Enum]
+        )
+
+
+class SET(_StringType):
+    """MySQL SET type."""
+
+    __visit_name__ = "SET"
+
+    def __init__(self, *values, **kw):
+        """Construct a SET.
+
+        E.g.::
+
+          Column("myset", SET("foo", "bar", "baz"))
+
+        The list of potential values is required in the case that this
+        set will be used to generate DDL for a table, or if the
+        :paramref:`.SET.retrieve_as_bitwise` flag is set to True.
+
+        :param values: The range of valid values for this SET. The values
+          are not quoted, they will be escaped and surrounded by single
+          quotes when generating the schema.
+
+        :param convert_unicode: Same flag as that of
+         :paramref:`.String.convert_unicode`.
+
+        :param collation: same as that of :paramref:`.String.collation`
+
+        :param charset: same as that of :paramref:`.VARCHAR.charset`.
+
+        :param ascii: same as that of :paramref:`.VARCHAR.ascii`.
+
+        :param unicode: same as that of :paramref:`.VARCHAR.unicode`.
+
+        :param binary: same as that of :paramref:`.VARCHAR.binary`.
+
+        :param retrieve_as_bitwise: if True, the data for the set type will be
+          persisted and selected using an integer value, where a set is coerced
+          into a bitwise mask for persistence.  MySQL allows this mode which
+          has the advantage of being able to store values unambiguously,
+          such as the blank string ``''``.   The datatype will appear
+          as the expression ``col + 0`` in a SELECT statement, so that the
+          value is coerced into an integer value in result sets.
+          This flag is required if one wishes
+          to persist a set that can store the blank string ``''`` as a value.
+
+          .. warning::
+
+            When using :paramref:`.mysql.SET.retrieve_as_bitwise`, it is
+            essential that the list of set values is expressed in the
+            **exact same order** as exists on the MySQL database.
+
+        """
+        self.retrieve_as_bitwise = kw.pop("retrieve_as_bitwise", False)
+        self.values = tuple(values)
+        if not self.retrieve_as_bitwise and "" in values:
+            raise exc.ArgumentError(
+                "Can't use the blank value '' in a SET without "
+                "setting retrieve_as_bitwise=True"
+            )
+        if self.retrieve_as_bitwise:
+            self._bitmap = {
+                value: 2**idx for idx, value in enumerate(self.values)
+            }
+            self._bitmap.update(
+                (2**idx, value) for idx, value in enumerate(self.values)
+            )
+        length = max([len(v) for v in values] + [0])
+        kw.setdefault("length", length)
+        super().__init__(**kw)
+
+    def column_expression(self, colexpr):
+        if self.retrieve_as_bitwise:
+            return sql.type_coerce(
+                sql.type_coerce(colexpr, sqltypes.Integer) + 0, self
+            )
+        else:
+            return colexpr
+
+    def result_processor(self, dialect, coltype):
+        if self.retrieve_as_bitwise:
+
+            def process(value):
+                if value is not None:
+                    value = int(value)
+
+                    return set(util.map_bits(self._bitmap.__getitem__, value))
+                else:
+                    return None
+
+        else:
+            super_convert = super().result_processor(dialect, coltype)
+
+            def process(value):
+                if isinstance(value, str):
+                    # MySQLdb returns a string, let's parse
+                    if super_convert:
+                        value = super_convert(value)
+                    return set(re.findall(r"[^,]+", value))
+                else:
+                    # mysql-connector-python does a naive
+                    # split(",") which throws in an empty string
+                    if value is not None:
+                        value.discard("")
+                    return value
+
+        return process
+
+    def bind_processor(self, dialect):
+        super_convert = super().bind_processor(dialect)
+        if self.retrieve_as_bitwise:
+
+            def process(value):
+                if value is None:
+                    return None
+                elif isinstance(value, (int, str)):
+                    if super_convert:
+                        return super_convert(value)
+                    else:
+                        return value
+                else:
+                    int_value = 0
+                    for v in value:
+                        int_value |= self._bitmap[v]
+                    return int_value
+
+        else:
+
+            def process(value):
+                # accept strings and int (actually bitflag) values directly
+                if value is not None and not isinstance(value, (int, str)):
+                    value = ",".join(value)
+
+                if super_convert:
+                    return super_convert(value)
+                else:
+                    return value
+
+        return process
+
+    def adapt(self, impltype, **kw):
+        kw["retrieve_as_bitwise"] = self.retrieve_as_bitwise
+        return util.constructor_copy(self, impltype, *self.values, **kw)
+
+    def __repr__(self):
+        return util.generic_repr(
+            self,
+            to_inspect=[SET, _StringType],
+            additional_kw=[
+                ("retrieve_as_bitwise", False),
+            ],
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/expression.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/expression.py
new file mode 100644
index 00000000..b60a0888
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/expression.py
@@ -0,0 +1,143 @@
+# dialects/mysql/expression.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from ... import exc
+from ... import util
+from ...sql import coercions
+from ...sql import elements
+from ...sql import operators
+from ...sql import roles
+from ...sql.base import _generative
+from ...sql.base import Generative
+from ...util.typing import Self
+
+
+class match(Generative, elements.BinaryExpression):
+    """Produce a ``MATCH (X, Y) AGAINST ('TEXT')`` clause.
+
+    E.g.::
+
+        from sqlalchemy import desc
+        from sqlalchemy.dialects.mysql import match
+
+        match_expr = match(
+            users_table.c.firstname,
+            users_table.c.lastname,
+            against="Firstname Lastname",
+        )
+
+        stmt = (
+            select(users_table)
+            .where(match_expr.in_boolean_mode())
+            .order_by(desc(match_expr))
+        )
+
+    Would produce SQL resembling:
+
+    .. sourcecode:: sql
+
+        SELECT id, firstname, lastname
+        FROM user
+        WHERE MATCH(firstname, lastname) AGAINST (:param_1 IN BOOLEAN MODE)
+        ORDER BY MATCH(firstname, lastname) AGAINST (:param_2) DESC
+
+    The :func:`_mysql.match` function is a standalone version of the
+    :meth:`_sql.ColumnElement.match` method available on all
+    SQL expressions, as when :meth:`_expression.ColumnElement.match` is
+    used, but allows to pass multiple columns
+
+    :param cols: column expressions to match against
+
+    :param against: expression to be compared towards
+
+    :param in_boolean_mode: boolean, set "boolean mode" to true
+
+    :param in_natural_language_mode: boolean , set "natural language" to true
+
+    :param with_query_expansion: boolean, set "query expansion" to true
+
+    .. versionadded:: 1.4.19
+
+    .. seealso::
+
+        :meth:`_expression.ColumnElement.match`
+
+    """
+
+    __visit_name__ = "mysql_match"
+
+    inherit_cache = True
+
+    def __init__(self, *cols, **kw):
+        if not cols:
+            raise exc.ArgumentError("columns are required")
+
+        against = kw.pop("against", None)
+
+        if against is None:
+            raise exc.ArgumentError("against is required")
+        against = coercions.expect(
+            roles.ExpressionElementRole,
+            against,
+        )
+
+        left = elements.BooleanClauseList._construct_raw(
+            operators.comma_op,
+            clauses=cols,
+        )
+        left.group = False
+
+        flags = util.immutabledict(
+            {
+                "mysql_boolean_mode": kw.pop("in_boolean_mode", False),
+                "mysql_natural_language": kw.pop(
+                    "in_natural_language_mode", False
+                ),
+                "mysql_query_expansion": kw.pop("with_query_expansion", False),
+            }
+        )
+
+        if kw:
+            raise exc.ArgumentError("unknown arguments: %s" % (", ".join(kw)))
+
+        super().__init__(left, against, operators.match_op, modifiers=flags)
+
+    @_generative
+    def in_boolean_mode(self) -> Self:
+        """Apply the "IN BOOLEAN MODE" modifier to the MATCH expression.
+
+        :return: a new :class:`_mysql.match` instance with modifications
+         applied.
+        """
+
+        self.modifiers = self.modifiers.union({"mysql_boolean_mode": True})
+        return self
+
+    @_generative
+    def in_natural_language_mode(self) -> Self:
+        """Apply the "IN NATURAL LANGUAGE MODE" modifier to the MATCH
+        expression.
+
+        :return: a new :class:`_mysql.match` instance with modifications
+         applied.
+        """
+
+        self.modifiers = self.modifiers.union({"mysql_natural_language": True})
+        return self
+
+    @_generative
+    def with_query_expansion(self) -> Self:
+        """Apply the "WITH QUERY EXPANSION" modifier to the MATCH expression.
+
+        :return: a new :class:`_mysql.match` instance with modifications
+         applied.
+        """
+
+        self.modifiers = self.modifiers.union({"mysql_query_expansion": True})
+        return self
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/json.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/json.py
new file mode 100644
index 00000000..8912af36
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/json.py
@@ -0,0 +1,81 @@
+# dialects/mysql/json.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import types as sqltypes
+
+
+class JSON(sqltypes.JSON):
+    """MySQL JSON type.
+
+    MySQL supports JSON as of version 5.7.
+    MariaDB supports JSON (as an alias for LONGTEXT) as of version 10.2.
+
+    :class:`_mysql.JSON` is used automatically whenever the base
+    :class:`_types.JSON` datatype is used against a MySQL or MariaDB backend.
+
+    .. seealso::
+
+        :class:`_types.JSON` - main documentation for the generic
+        cross-platform JSON datatype.
+
+    The :class:`.mysql.JSON` type supports persistence of JSON values
+    as well as the core index operations provided by :class:`_types.JSON`
+    datatype, by adapting the operations to render the ``JSON_EXTRACT``
+    function at the database level.
+
+    """
+
+    pass
+
+
+class _FormatTypeMixin:
+    def _format_value(self, value):
+        raise NotImplementedError()
+
+    def bind_processor(self, dialect):
+        super_proc = self.string_bind_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+    def literal_processor(self, dialect):
+        super_proc = self.string_literal_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+
+class JSONIndexType(_FormatTypeMixin, sqltypes.JSON.JSONIndexType):
+    def _format_value(self, value):
+        if isinstance(value, int):
+            value = "$[%s]" % value
+        else:
+            value = '$."%s"' % value
+        return value
+
+
+class JSONPathType(_FormatTypeMixin, sqltypes.JSON.JSONPathType):
+    def _format_value(self, value):
+        return "$%s" % (
+            "".join(
+                [
+                    "[%s]" % elem if isinstance(elem, int) else '."%s"' % elem
+                    for elem in value
+                ]
+            )
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadb.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadb.py
new file mode 100644
index 00000000..ac2cfbd1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadb.py
@@ -0,0 +1,61 @@
+# dialects/mysql/mariadb.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from .base import MariaDBIdentifierPreparer
+from .base import MySQLDialect
+from .base import MySQLTypeCompiler
+from ...sql import sqltypes
+
+
+class INET4(sqltypes.TypeEngine[str]):
+    """INET4 column type for MariaDB
+
+    .. versionadded:: 2.0.37
+    """
+
+    __visit_name__ = "INET4"
+
+
+class INET6(sqltypes.TypeEngine[str]):
+    """INET6 column type for MariaDB
+
+    .. versionadded:: 2.0.37
+    """
+
+    __visit_name__ = "INET6"
+
+
+class MariaDBTypeCompiler(MySQLTypeCompiler):
+    def visit_INET4(self, type_, **kwargs) -> str:
+        return "INET4"
+
+    def visit_INET6(self, type_, **kwargs) -> str:
+        return "INET6"
+
+
+class MariaDBDialect(MySQLDialect):
+    is_mariadb = True
+    supports_statement_cache = True
+    name = "mariadb"
+    preparer = MariaDBIdentifierPreparer
+    type_compiler_cls = MariaDBTypeCompiler
+
+
+def loader(driver):
+    driver_mod = __import__(
+        "sqlalchemy.dialects.mysql.%s" % driver
+    ).dialects.mysql
+    driver_cls = getattr(driver_mod, driver).dialect
+
+    return type(
+        "MariaDBDialect_%s" % driver,
+        (
+            MariaDBDialect,
+            driver_cls,
+        ),
+        {"supports_statement_cache": True},
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadbconnector.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadbconnector.py
new file mode 100644
index 00000000..2d2ad199
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mariadbconnector.py
@@ -0,0 +1,277 @@
+# dialects/mysql/mariadbconnector.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""
+
+.. dialect:: mysql+mariadbconnector
+    :name: MariaDB Connector/Python
+    :dbapi: mariadb
+    :connectstring: mariadb+mariadbconnector://<user>:<password>@<host>[:<port>]/<dbname>
+    :url: https://pypi.org/project/mariadb/
+
+Driver Status
+-------------
+
+MariaDB Connector/Python enables Python programs to access MariaDB and MySQL
+databases using an API which is compliant with the Python DB API 2.0 (PEP-249).
+It is written in C and uses MariaDB Connector/C client library for client server
+communication.
+
+Note that the default driver for a ``mariadb://`` connection URI continues to
+be ``mysqldb``. ``mariadb+mariadbconnector://`` is required to use this driver.
+
+.. mariadb: https://github.com/mariadb-corporation/mariadb-connector-python
+
+"""  # noqa
+import re
+from uuid import UUID as _python_UUID
+
+from .base import MySQLCompiler
+from .base import MySQLDialect
+from .base import MySQLExecutionContext
+from ... import sql
+from ... import util
+from ...sql import sqltypes
+
+
+mariadb_cpy_minimum_version = (1, 0, 1)
+
+
+class _MariaDBUUID(sqltypes.UUID[sqltypes._UUID_RETURN]):
+    # work around JIRA issue
+    # https://jira.mariadb.org/browse/CONPY-270.  When that issue is fixed,
+    # this type can be removed.
+    def result_processor(self, dialect, coltype):
+        if self.as_uuid:
+
+            def process(value):
+                if value is not None:
+                    if hasattr(value, "decode"):
+                        value = value.decode("ascii")
+                    value = _python_UUID(value)
+                return value
+
+            return process
+        else:
+
+            def process(value):
+                if value is not None:
+                    if hasattr(value, "decode"):
+                        value = value.decode("ascii")
+                    value = str(_python_UUID(value))
+                return value
+
+            return process
+
+
+class MySQLExecutionContext_mariadbconnector(MySQLExecutionContext):
+    _lastrowid = None
+
+    def create_server_side_cursor(self):
+        return self._dbapi_connection.cursor(buffered=False)
+
+    def create_default_cursor(self):
+        return self._dbapi_connection.cursor(buffered=True)
+
+    def post_exec(self):
+        super().post_exec()
+
+        self._rowcount = self.cursor.rowcount
+
+        if self.isinsert and self.compiled.postfetch_lastrowid:
+            self._lastrowid = self.cursor.lastrowid
+
+    def get_lastrowid(self):
+        return self._lastrowid
+
+
+class MySQLCompiler_mariadbconnector(MySQLCompiler):
+    pass
+
+
+class MySQLDialect_mariadbconnector(MySQLDialect):
+    driver = "mariadbconnector"
+    supports_statement_cache = True
+
+    # set this to True at the module level to prevent the driver from running
+    # against a backend that server detects as MySQL. currently this appears to
+    # be unnecessary as MariaDB client libraries have always worked against
+    # MySQL databases.   However, if this changes at some point, this can be
+    # adjusted, but PLEASE ADD A TEST in test/dialect/mysql/test_dialect.py if
+    # this change is made at some point to ensure the correct exception
+    # is raised at the correct point when running the driver against
+    # a MySQL backend.
+    # is_mariadb = True
+
+    supports_unicode_statements = True
+    encoding = "utf8mb4"
+    convert_unicode = True
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = True
+    supports_native_decimal = True
+    default_paramstyle = "qmark"
+    execution_ctx_cls = MySQLExecutionContext_mariadbconnector
+    statement_compiler = MySQLCompiler_mariadbconnector
+
+    supports_server_side_cursors = True
+
+    colspecs = util.update_copy(
+        MySQLDialect.colspecs, {sqltypes.Uuid: _MariaDBUUID}
+    )
+
+    @util.memoized_property
+    def _dbapi_version(self):
+        if self.dbapi and hasattr(self.dbapi, "__version__"):
+            return tuple(
+                [
+                    int(x)
+                    for x in re.findall(
+                        r"(\d+)(?:[-\.]?|$)", self.dbapi.__version__
+                    )
+                ]
+            )
+        else:
+            return (99, 99, 99)
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self.paramstyle = "qmark"
+        if self.dbapi is not None:
+            if self._dbapi_version < mariadb_cpy_minimum_version:
+                raise NotImplementedError(
+                    "The minimum required version for MariaDB "
+                    "Connector/Python is %s"
+                    % ".".join(str(x) for x in mariadb_cpy_minimum_version)
+                )
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("mariadb")
+
+    def is_disconnect(self, e, connection, cursor):
+        if super().is_disconnect(e, connection, cursor):
+            return True
+        elif isinstance(e, self.dbapi.Error):
+            str_e = str(e).lower()
+            return "not connected" in str_e or "isn't valid" in str_e
+        else:
+            return False
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args()
+        opts.update(url.query)
+
+        int_params = [
+            "connect_timeout",
+            "read_timeout",
+            "write_timeout",
+            "client_flag",
+            "port",
+            "pool_size",
+        ]
+        bool_params = [
+            "local_infile",
+            "ssl_verify_cert",
+            "ssl",
+            "pool_reset_connection",
+            "compress",
+        ]
+
+        for key in int_params:
+            util.coerce_kw_type(opts, key, int)
+        for key in bool_params:
+            util.coerce_kw_type(opts, key, bool)
+
+        # FOUND_ROWS must be set in CLIENT_FLAGS to enable
+        # supports_sane_rowcount.
+        client_flag = opts.get("client_flag", 0)
+        if self.dbapi is not None:
+            try:
+                CLIENT_FLAGS = __import__(
+                    self.dbapi.__name__ + ".constants.CLIENT"
+                ).constants.CLIENT
+                client_flag |= CLIENT_FLAGS.FOUND_ROWS
+            except (AttributeError, ImportError):
+                self.supports_sane_rowcount = False
+            opts["client_flag"] = client_flag
+        return [[], opts]
+
+    def _extract_error_code(self, exception):
+        try:
+            rc = exception.errno
+        except:
+            rc = -1
+        return rc
+
+    def _detect_charset(self, connection):
+        return "utf8mb4"
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return (
+            "SERIALIZABLE",
+            "READ UNCOMMITTED",
+            "READ COMMITTED",
+            "REPEATABLE READ",
+            "AUTOCOMMIT",
+        )
+
+    def set_isolation_level(self, connection, level):
+        if level == "AUTOCOMMIT":
+            connection.autocommit = True
+        else:
+            connection.autocommit = False
+            super().set_isolation_level(connection, level)
+
+    def do_begin_twophase(self, connection, xid):
+        connection.execute(
+            sql.text("XA BEGIN :xid").bindparams(
+                sql.bindparam("xid", xid, literal_execute=True)
+            )
+        )
+
+    def do_prepare_twophase(self, connection, xid):
+        connection.execute(
+            sql.text("XA END :xid").bindparams(
+                sql.bindparam("xid", xid, literal_execute=True)
+            )
+        )
+        connection.execute(
+            sql.text("XA PREPARE :xid").bindparams(
+                sql.bindparam("xid", xid, literal_execute=True)
+            )
+        )
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if not is_prepared:
+            connection.execute(
+                sql.text("XA END :xid").bindparams(
+                    sql.bindparam("xid", xid, literal_execute=True)
+                )
+            )
+        connection.execute(
+            sql.text("XA ROLLBACK :xid").bindparams(
+                sql.bindparam("xid", xid, literal_execute=True)
+            )
+        )
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if not is_prepared:
+            self.do_prepare_twophase(connection, xid)
+        connection.execute(
+            sql.text("XA COMMIT :xid").bindparams(
+                sql.bindparam("xid", xid, literal_execute=True)
+            )
+        )
+
+
+dialect = MySQLDialect_mariadbconnector
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqlconnector.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqlconnector.py
new file mode 100644
index 00000000..e88f8fd7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqlconnector.py
@@ -0,0 +1,180 @@
+# dialects/mysql/mysqlconnector.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+.. dialect:: mysql+mysqlconnector
+    :name: MySQL Connector/Python
+    :dbapi: myconnpy
+    :connectstring: mysql+mysqlconnector://<user>:<password>@<host>[:<port>]/<dbname>
+    :url: https://pypi.org/project/mysql-connector-python/
+
+.. note::
+
+    The MySQL Connector/Python DBAPI has had many issues since its release,
+    some of which may remain unresolved, and the mysqlconnector dialect is
+    **not tested as part of SQLAlchemy's continuous integration**.
+    The recommended MySQL dialects are mysqlclient and PyMySQL.
+
+"""  # noqa
+
+import re
+
+from .base import BIT
+from .base import MySQLCompiler
+from .base import MySQLDialect
+from .base import MySQLIdentifierPreparer
+from ... import util
+
+
+class MySQLCompiler_mysqlconnector(MySQLCompiler):
+    def visit_mod_binary(self, binary, operator, **kw):
+        return (
+            self.process(binary.left, **kw)
+            + " % "
+            + self.process(binary.right, **kw)
+        )
+
+
+class MySQLIdentifierPreparer_mysqlconnector(MySQLIdentifierPreparer):
+    @property
+    def _double_percents(self):
+        return False
+
+    @_double_percents.setter
+    def _double_percents(self, value):
+        pass
+
+    def _escape_identifier(self, value):
+        value = value.replace(self.escape_quote, self.escape_to_quote)
+        return value
+
+
+class _myconnpyBIT(BIT):
+    def result_processor(self, dialect, coltype):
+        """MySQL-connector already converts mysql bits, so."""
+
+        return None
+
+
+class MySQLDialect_mysqlconnector(MySQLDialect):
+    driver = "mysqlconnector"
+    supports_statement_cache = True
+
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = True
+
+    supports_native_decimal = True
+
+    default_paramstyle = "format"
+    statement_compiler = MySQLCompiler_mysqlconnector
+
+    preparer = MySQLIdentifierPreparer_mysqlconnector
+
+    colspecs = util.update_copy(MySQLDialect.colspecs, {BIT: _myconnpyBIT})
+
+    @classmethod
+    def import_dbapi(cls):
+        from mysql import connector
+
+        return connector
+
+    def do_ping(self, dbapi_connection):
+        dbapi_connection.ping(False)
+        return True
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args(username="user")
+
+        opts.update(url.query)
+
+        util.coerce_kw_type(opts, "allow_local_infile", bool)
+        util.coerce_kw_type(opts, "autocommit", bool)
+        util.coerce_kw_type(opts, "buffered", bool)
+        util.coerce_kw_type(opts, "client_flag", int)
+        util.coerce_kw_type(opts, "compress", bool)
+        util.coerce_kw_type(opts, "connection_timeout", int)
+        util.coerce_kw_type(opts, "connect_timeout", int)
+        util.coerce_kw_type(opts, "consume_results", bool)
+        util.coerce_kw_type(opts, "force_ipv6", bool)
+        util.coerce_kw_type(opts, "get_warnings", bool)
+        util.coerce_kw_type(opts, "pool_reset_session", bool)
+        util.coerce_kw_type(opts, "pool_size", int)
+        util.coerce_kw_type(opts, "raise_on_warnings", bool)
+        util.coerce_kw_type(opts, "raw", bool)
+        util.coerce_kw_type(opts, "ssl_verify_cert", bool)
+        util.coerce_kw_type(opts, "use_pure", bool)
+        util.coerce_kw_type(opts, "use_unicode", bool)
+
+        # unfortunately, MySQL/connector python refuses to release a
+        # cursor without reading fully, so non-buffered isn't an option
+        opts.setdefault("buffered", True)
+
+        # FOUND_ROWS must be set in ClientFlag to enable
+        # supports_sane_rowcount.
+        if self.dbapi is not None:
+            try:
+                from mysql.connector.constants import ClientFlag
+
+                client_flags = opts.get(
+                    "client_flags", ClientFlag.get_default()
+                )
+                client_flags |= ClientFlag.FOUND_ROWS
+                opts["client_flags"] = client_flags
+            except Exception:
+                pass
+        return [[], opts]
+
+    @util.memoized_property
+    def _mysqlconnector_version_info(self):
+        if self.dbapi and hasattr(self.dbapi, "__version__"):
+            m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", self.dbapi.__version__)
+            if m:
+                return tuple(int(x) for x in m.group(1, 2, 3) if x is not None)
+
+    def _detect_charset(self, connection):
+        return connection.connection.charset
+
+    def _extract_error_code(self, exception):
+        return exception.errno
+
+    def is_disconnect(self, e, connection, cursor):
+        errnos = (2006, 2013, 2014, 2045, 2055, 2048)
+        exceptions = (self.dbapi.OperationalError, self.dbapi.InterfaceError)
+        if isinstance(e, exceptions):
+            return (
+                e.errno in errnos
+                or "MySQL Connection not available." in str(e)
+                or "Connection to MySQL is not available" in str(e)
+            )
+        else:
+            return False
+
+    def _compat_fetchall(self, rp, charset=None):
+        return rp.fetchall()
+
+    def _compat_fetchone(self, rp, charset=None):
+        return rp.fetchone()
+
+    _isolation_lookup = {
+        "SERIALIZABLE",
+        "READ UNCOMMITTED",
+        "READ COMMITTED",
+        "REPEATABLE READ",
+        "AUTOCOMMIT",
+    }
+
+    def _set_isolation_level(self, connection, level):
+        if level == "AUTOCOMMIT":
+            connection.autocommit = True
+        else:
+            connection.autocommit = False
+            super()._set_isolation_level(connection, level)
+
+
+dialect = MySQLDialect_mysqlconnector
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqldb.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqldb.py
new file mode 100644
index 00000000..3cf56c1f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/mysqldb.py
@@ -0,0 +1,305 @@
+# dialects/mysql/mysqldb.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""
+
+.. dialect:: mysql+mysqldb
+    :name: mysqlclient (maintained fork of MySQL-Python)
+    :dbapi: mysqldb
+    :connectstring: mysql+mysqldb://<user>:<password>@<host>[:<port>]/<dbname>
+    :url: https://pypi.org/project/mysqlclient/
+
+Driver Status
+-------------
+
+The mysqlclient DBAPI is a maintained fork of the
+`MySQL-Python <https://sourceforge.net/projects/mysql-python>`_ DBAPI
+that is no longer maintained.  `mysqlclient`_ supports Python 2 and Python 3
+and is very stable.
+
+.. _mysqlclient: https://github.com/PyMySQL/mysqlclient-python
+
+.. _mysqldb_unicode:
+
+Unicode
+-------
+
+Please see :ref:`mysql_unicode` for current recommendations on unicode
+handling.
+
+.. _mysqldb_ssl:
+
+SSL Connections
+----------------
+
+The mysqlclient and PyMySQL DBAPIs accept an additional dictionary under the
+key "ssl", which may be specified using the
+:paramref:`_sa.create_engine.connect_args` dictionary::
+
+    engine = create_engine(
+        "mysql+mysqldb://scott:tiger@192.168.0.134/test",
+        connect_args={
+            "ssl": {
+                "ca": "/home/gord/client-ssl/ca.pem",
+                "cert": "/home/gord/client-ssl/client-cert.pem",
+                "key": "/home/gord/client-ssl/client-key.pem",
+            }
+        },
+    )
+
+For convenience, the following keys may also be specified inline within the URL
+where they will be interpreted into the "ssl" dictionary automatically:
+"ssl_ca", "ssl_cert", "ssl_key", "ssl_capath", "ssl_cipher",
+"ssl_check_hostname". An example is as follows::
+
+    connection_uri = (
+        "mysql+mysqldb://scott:tiger@192.168.0.134/test"
+        "?ssl_ca=/home/gord/client-ssl/ca.pem"
+        "&ssl_cert=/home/gord/client-ssl/client-cert.pem"
+        "&ssl_key=/home/gord/client-ssl/client-key.pem"
+    )
+
+.. seealso::
+
+    :ref:`pymysql_ssl` in the PyMySQL dialect
+
+
+Using MySQLdb with Google Cloud SQL
+-----------------------------------
+
+Google Cloud SQL now recommends use of the MySQLdb dialect.  Connect
+using a URL like the following:
+
+.. sourcecode:: text
+
+    mysql+mysqldb://root@/<dbname>?unix_socket=/cloudsql/<projectid>:<instancename>
+
+Server Side Cursors
+-------------------
+
+The mysqldb dialect supports server-side cursors. See :ref:`mysql_ss_cursors`.
+
+"""
+
+import re
+
+from .base import MySQLCompiler
+from .base import MySQLDialect
+from .base import MySQLExecutionContext
+from .base import MySQLIdentifierPreparer
+from .base import TEXT
+from ... import sql
+from ... import util
+
+
+class MySQLExecutionContext_mysqldb(MySQLExecutionContext):
+    pass
+
+
+class MySQLCompiler_mysqldb(MySQLCompiler):
+    pass
+
+
+class MySQLDialect_mysqldb(MySQLDialect):
+    driver = "mysqldb"
+    supports_statement_cache = True
+    supports_unicode_statements = True
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = True
+
+    supports_native_decimal = True
+
+    default_paramstyle = "format"
+    execution_ctx_cls = MySQLExecutionContext_mysqldb
+    statement_compiler = MySQLCompiler_mysqldb
+    preparer = MySQLIdentifierPreparer
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self._mysql_dbapi_version = (
+            self._parse_dbapi_version(self.dbapi.__version__)
+            if self.dbapi is not None and hasattr(self.dbapi, "__version__")
+            else (0, 0, 0)
+        )
+
+    def _parse_dbapi_version(self, version):
+        m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", version)
+        if m:
+            return tuple(int(x) for x in m.group(1, 2, 3) if x is not None)
+        else:
+            return (0, 0, 0)
+
+    @util.langhelpers.memoized_property
+    def supports_server_side_cursors(self):
+        try:
+            cursors = __import__("MySQLdb.cursors").cursors
+            self._sscursor = cursors.SSCursor
+            return True
+        except (ImportError, AttributeError):
+            return False
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("MySQLdb")
+
+    def on_connect(self):
+        super_ = super().on_connect()
+
+        def on_connect(conn):
+            if super_ is not None:
+                super_(conn)
+
+            charset_name = conn.character_set_name()
+
+            if charset_name is not None:
+                cursor = conn.cursor()
+                cursor.execute("SET NAMES %s" % charset_name)
+                cursor.close()
+
+        return on_connect
+
+    def do_ping(self, dbapi_connection):
+        dbapi_connection.ping()
+        return True
+
+    def do_executemany(self, cursor, statement, parameters, context=None):
+        rowcount = cursor.executemany(statement, parameters)
+        if context is not None:
+            context._rowcount = rowcount
+
+    def _check_unicode_returns(self, connection):
+        # work around issue fixed in
+        # https://github.com/farcepest/MySQLdb1/commit/cd44524fef63bd3fcb71947392326e9742d520e8
+        # specific issue w/ the utf8mb4_bin collation and unicode returns
+
+        collation = connection.exec_driver_sql(
+            "show collation where %s = 'utf8mb4' and %s = 'utf8mb4_bin'"
+            % (
+                self.identifier_preparer.quote("Charset"),
+                self.identifier_preparer.quote("Collation"),
+            )
+        ).scalar()
+        has_utf8mb4_bin = self.server_version_info > (5,) and collation
+        if has_utf8mb4_bin:
+            additional_tests = [
+                sql.collate(
+                    sql.cast(
+                        sql.literal_column("'test collated returns'"),
+                        TEXT(charset="utf8mb4"),
+                    ),
+                    "utf8mb4_bin",
+                )
+            ]
+        else:
+            additional_tests = []
+        return super()._check_unicode_returns(connection, additional_tests)
+
+    def create_connect_args(self, url, _translate_args=None):
+        if _translate_args is None:
+            _translate_args = dict(
+                database="db", username="user", password="passwd"
+            )
+
+        opts = url.translate_connect_args(**_translate_args)
+        opts.update(url.query)
+
+        util.coerce_kw_type(opts, "compress", bool)
+        util.coerce_kw_type(opts, "connect_timeout", int)
+        util.coerce_kw_type(opts, "read_timeout", int)
+        util.coerce_kw_type(opts, "write_timeout", int)
+        util.coerce_kw_type(opts, "client_flag", int)
+        util.coerce_kw_type(opts, "local_infile", bool)
+        # Note: using either of the below will cause all strings to be
+        # returned as Unicode, both in raw SQL operations and with column
+        # types like String and MSString.
+        util.coerce_kw_type(opts, "use_unicode", bool)
+        util.coerce_kw_type(opts, "charset", str)
+
+        # Rich values 'cursorclass' and 'conv' are not supported via
+        # query string.
+
+        ssl = {}
+        keys = [
+            ("ssl_ca", str),
+            ("ssl_key", str),
+            ("ssl_cert", str),
+            ("ssl_capath", str),
+            ("ssl_cipher", str),
+            ("ssl_check_hostname", bool),
+        ]
+        for key, kw_type in keys:
+            if key in opts:
+                ssl[key[4:]] = opts[key]
+                util.coerce_kw_type(ssl, key[4:], kw_type)
+                del opts[key]
+        if ssl:
+            opts["ssl"] = ssl
+
+        # FOUND_ROWS must be set in CLIENT_FLAGS to enable
+        # supports_sane_rowcount.
+        client_flag = opts.get("client_flag", 0)
+
+        client_flag_found_rows = self._found_rows_client_flag()
+        if client_flag_found_rows is not None:
+            client_flag |= client_flag_found_rows
+            opts["client_flag"] = client_flag
+        return [[], opts]
+
+    def _found_rows_client_flag(self):
+        if self.dbapi is not None:
+            try:
+                CLIENT_FLAGS = __import__(
+                    self.dbapi.__name__ + ".constants.CLIENT"
+                ).constants.CLIENT
+            except (AttributeError, ImportError):
+                return None
+            else:
+                return CLIENT_FLAGS.FOUND_ROWS
+        else:
+            return None
+
+    def _extract_error_code(self, exception):
+        return exception.args[0]
+
+    def _detect_charset(self, connection):
+        """Sniff out the character set in use for connection results."""
+
+        try:
+            # note: the SQL here would be
+            # "SHOW VARIABLES LIKE 'character_set%%'"
+            cset_name = connection.connection.character_set_name
+        except AttributeError:
+            util.warn(
+                "No 'character_set_name' can be detected with "
+                "this MySQL-Python version; "
+                "please upgrade to a recent version of MySQL-Python.  "
+                "Assuming latin1."
+            )
+            return "latin1"
+        else:
+            return cset_name()
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return (
+            "SERIALIZABLE",
+            "READ UNCOMMITTED",
+            "READ COMMITTED",
+            "REPEATABLE READ",
+            "AUTOCOMMIT",
+        )
+
+    def set_isolation_level(self, dbapi_connection, level):
+        if level == "AUTOCOMMIT":
+            dbapi_connection.autocommit(True)
+        else:
+            dbapi_connection.autocommit(False)
+            super().set_isolation_level(dbapi_connection, level)
+
+
+dialect = MySQLDialect_mysqldb
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/provision.py
new file mode 100644
index 00000000..7807af40
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/provision.py
@@ -0,0 +1,110 @@
+# dialects/mysql/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import exc
+from ...testing.provision import configure_follower
+from ...testing.provision import create_db
+from ...testing.provision import drop_db
+from ...testing.provision import generate_driver_url
+from ...testing.provision import temp_table_keyword_args
+from ...testing.provision import upsert
+
+
+@generate_driver_url.for_db("mysql", "mariadb")
+def generate_driver_url(url, driver, query_str):
+    backend = url.get_backend_name()
+
+    # NOTE: at the moment, tests are running mariadbconnector
+    # against both mariadb and mysql backends.   if we want this to be
+    # limited, do the decision making here to reject a "mysql+mariadbconnector"
+    # URL.  Optionally also re-enable the module level
+    # MySQLDialect_mariadbconnector.is_mysql flag as well, which must include
+    # a unit and/or functional test.
+
+    # all the Jenkins tests have been running mysqlclient Python library
+    # built against mariadb client drivers for years against all MySQL /
+    # MariaDB versions going back to MySQL 5.6, currently they can talk
+    # to MySQL databases without problems.
+
+    if backend == "mysql":
+        dialect_cls = url.get_dialect()
+        if dialect_cls._is_mariadb_from_url(url):
+            backend = "mariadb"
+
+    new_url = url.set(
+        drivername="%s+%s" % (backend, driver)
+    ).update_query_string(query_str)
+
+    if driver == "mariadbconnector":
+        new_url = new_url.difference_update_query(["charset"])
+
+    try:
+        new_url.get_dialect()
+    except exc.NoSuchModuleError:
+        return None
+    else:
+        return new_url
+
+
+@create_db.for_db("mysql", "mariadb")
+def _mysql_create_db(cfg, eng, ident):
+    with eng.begin() as conn:
+        try:
+            _mysql_drop_db(cfg, conn, ident)
+        except Exception:
+            pass
+
+    with eng.begin() as conn:
+        conn.exec_driver_sql(
+            "CREATE DATABASE %s CHARACTER SET utf8mb4" % ident
+        )
+        conn.exec_driver_sql(
+            "CREATE DATABASE %s_test_schema CHARACTER SET utf8mb4" % ident
+        )
+        conn.exec_driver_sql(
+            "CREATE DATABASE %s_test_schema_2 CHARACTER SET utf8mb4" % ident
+        )
+
+
+@configure_follower.for_db("mysql", "mariadb")
+def _mysql_configure_follower(config, ident):
+    config.test_schema = "%s_test_schema" % ident
+    config.test_schema_2 = "%s_test_schema_2" % ident
+
+
+@drop_db.for_db("mysql", "mariadb")
+def _mysql_drop_db(cfg, eng, ident):
+    with eng.begin() as conn:
+        conn.exec_driver_sql("DROP DATABASE %s_test_schema" % ident)
+        conn.exec_driver_sql("DROP DATABASE %s_test_schema_2" % ident)
+        conn.exec_driver_sql("DROP DATABASE %s" % ident)
+
+
+@temp_table_keyword_args.for_db("mysql", "mariadb")
+def _mysql_temp_table_keyword_args(cfg, eng):
+    return {"prefixes": ["TEMPORARY"]}
+
+
+@upsert.for_db("mariadb")
+def _upsert(
+    cfg, table, returning, *, set_lambda=None, sort_by_parameter_order=False
+):
+    from sqlalchemy.dialects.mysql import insert
+
+    stmt = insert(table)
+
+    if set_lambda:
+        stmt = stmt.on_duplicate_key_update(**set_lambda(stmt.inserted))
+    else:
+        pk1 = table.primary_key.c[0]
+        stmt = stmt.on_duplicate_key_update({pk1.key: pk1})
+
+    stmt = stmt.returning(
+        *returning, sort_by_parameter_order=sort_by_parameter_order
+    )
+    return stmt
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pymysql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pymysql.py
new file mode 100644
index 00000000..67cb4cdd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pymysql.py
@@ -0,0 +1,136 @@
+# dialects/mysql/pymysql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+
+.. dialect:: mysql+pymysql
+    :name: PyMySQL
+    :dbapi: pymysql
+    :connectstring: mysql+pymysql://<username>:<password>@<host>/<dbname>[?<options>]
+    :url: https://pymysql.readthedocs.io/
+
+Unicode
+-------
+
+Please see :ref:`mysql_unicode` for current recommendations on unicode
+handling.
+
+.. _pymysql_ssl:
+
+SSL Connections
+------------------
+
+The PyMySQL DBAPI accepts the same SSL arguments as that of MySQLdb,
+described at :ref:`mysqldb_ssl`.   See that section for additional examples.
+
+If the server uses an automatically-generated certificate that is self-signed
+or does not match the host name (as seen from the client), it may also be
+necessary to indicate ``ssl_check_hostname=false`` in PyMySQL::
+
+    connection_uri = (
+        "mysql+pymysql://scott:tiger@192.168.0.134/test"
+        "?ssl_ca=/home/gord/client-ssl/ca.pem"
+        "&ssl_cert=/home/gord/client-ssl/client-cert.pem"
+        "&ssl_key=/home/gord/client-ssl/client-key.pem"
+        "&ssl_check_hostname=false"
+    )
+
+MySQL-Python Compatibility
+--------------------------
+
+The pymysql DBAPI is a pure Python port of the MySQL-python (MySQLdb) driver,
+and targets 100% compatibility.   Most behavioral notes for MySQL-python apply
+to the pymysql driver as well.
+
+"""  # noqa
+
+from .mysqldb import MySQLDialect_mysqldb
+from ...util import langhelpers
+
+
+class MySQLDialect_pymysql(MySQLDialect_mysqldb):
+    driver = "pymysql"
+    supports_statement_cache = True
+
+    description_encoding = None
+
+    @langhelpers.memoized_property
+    def supports_server_side_cursors(self):
+        try:
+            cursors = __import__("pymysql.cursors").cursors
+            self._sscursor = cursors.SSCursor
+            return True
+        except (ImportError, AttributeError):
+            return False
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("pymysql")
+
+    @langhelpers.memoized_property
+    def _send_false_to_ping(self):
+        """determine if pymysql has deprecated, changed the default of,
+        or removed the 'reconnect' argument of connection.ping().
+
+        See #10492 and
+        https://github.com/PyMySQL/mysqlclient/discussions/651#discussioncomment-7308971
+        for background.
+
+        """  # noqa: E501
+
+        try:
+            Connection = __import__(
+                "pymysql.connections"
+            ).connections.Connection
+        except (ImportError, AttributeError):
+            return True
+        else:
+            insp = langhelpers.get_callable_argspec(Connection.ping)
+            try:
+                reconnect_arg = insp.args[1]
+            except IndexError:
+                return False
+            else:
+                return reconnect_arg == "reconnect" and (
+                    not insp.defaults or insp.defaults[0] is not False
+                )
+
+    def do_ping(self, dbapi_connection):
+        if self._send_false_to_ping:
+            dbapi_connection.ping(False)
+        else:
+            dbapi_connection.ping()
+
+        return True
+
+    def create_connect_args(self, url, _translate_args=None):
+        if _translate_args is None:
+            _translate_args = dict(username="user")
+        return super().create_connect_args(
+            url, _translate_args=_translate_args
+        )
+
+    def is_disconnect(self, e, connection, cursor):
+        if super().is_disconnect(e, connection, cursor):
+            return True
+        elif isinstance(e, self.dbapi.Error):
+            str_e = str(e).lower()
+            return (
+                "already closed" in str_e or "connection was killed" in str_e
+            )
+        else:
+            return False
+
+    def _extract_error_code(self, exception):
+        if isinstance(exception.args[0], Exception):
+            exception = exception.args[0]
+        return exception.args[0]
+
+
+dialect = MySQLDialect_pymysql
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pyodbc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pyodbc.py
new file mode 100644
index 00000000..6d44bd38
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/pyodbc.py
@@ -0,0 +1,139 @@
+# dialects/mysql/pyodbc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+
+
+.. dialect:: mysql+pyodbc
+    :name: PyODBC
+    :dbapi: pyodbc
+    :connectstring: mysql+pyodbc://<username>:<password>@<dsnname>
+    :url: https://pypi.org/project/pyodbc/
+
+.. note::
+
+    The PyODBC for MySQL dialect is **not tested as part of
+    SQLAlchemy's continuous integration**.
+    The recommended MySQL dialects are mysqlclient and PyMySQL.
+    However, if you want to use the mysql+pyodbc dialect and require
+    full support for ``utf8mb4`` characters (including supplementary
+    characters like emoji) be sure to use a current release of
+    MySQL Connector/ODBC and specify the "ANSI" (**not** "Unicode")
+    version of the driver in your DSN or connection string.
+
+Pass through exact pyodbc connection string::
+
+    import urllib
+
+    connection_string = (
+        "DRIVER=MySQL ODBC 8.0 ANSI Driver;"
+        "SERVER=localhost;"
+        "PORT=3307;"
+        "DATABASE=mydb;"
+        "UID=root;"
+        "PWD=(whatever);"
+        "charset=utf8mb4;"
+    )
+    params = urllib.parse.quote_plus(connection_string)
+    connection_uri = "mysql+pyodbc:///?odbc_connect=%s" % params
+
+"""  # noqa
+
+import re
+
+from .base import MySQLDialect
+from .base import MySQLExecutionContext
+from .types import TIME
+from ... import exc
+from ... import util
+from ...connectors.pyodbc import PyODBCConnector
+from ...sql.sqltypes import Time
+
+
+class _pyodbcTIME(TIME):
+    def result_processor(self, dialect, coltype):
+        def process(value):
+            # pyodbc returns a datetime.time object; no need to convert
+            return value
+
+        return process
+
+
+class MySQLExecutionContext_pyodbc(MySQLExecutionContext):
+    def get_lastrowid(self):
+        cursor = self.create_cursor()
+        cursor.execute("SELECT LAST_INSERT_ID()")
+        lastrowid = cursor.fetchone()[0]
+        cursor.close()
+        return lastrowid
+
+
+class MySQLDialect_pyodbc(PyODBCConnector, MySQLDialect):
+    supports_statement_cache = True
+    colspecs = util.update_copy(MySQLDialect.colspecs, {Time: _pyodbcTIME})
+    supports_unicode_statements = True
+    execution_ctx_cls = MySQLExecutionContext_pyodbc
+
+    pyodbc_driver_name = "MySQL"
+
+    def _detect_charset(self, connection):
+        """Sniff out the character set in use for connection results."""
+
+        # Prefer 'character_set_results' for the current connection over the
+        # value in the driver.  SET NAMES or individual variable SETs will
+        # change the charset without updating the driver's view of the world.
+        #
+        # If it's decided that issuing that sort of SQL leaves you SOL, then
+        # this can prefer the driver value.
+
+        # set this to None as _fetch_setting attempts to use it (None is OK)
+        self._connection_charset = None
+        try:
+            value = self._fetch_setting(connection, "character_set_client")
+            if value:
+                return value
+        except exc.DBAPIError:
+            pass
+
+        util.warn(
+            "Could not detect the connection character set.  "
+            "Assuming latin1."
+        )
+        return "latin1"
+
+    def _get_server_version_info(self, connection):
+        return MySQLDialect._get_server_version_info(self, connection)
+
+    def _extract_error_code(self, exception):
+        m = re.compile(r"\((\d+)\)").search(str(exception.args))
+        c = m.group(1)
+        if c:
+            return int(c)
+        else:
+            return None
+
+    def on_connect(self):
+        super_ = super().on_connect()
+
+        def on_connect(conn):
+            if super_ is not None:
+                super_(conn)
+
+            # declare Unicode encoding for pyodbc as per
+            #   https://github.com/mkleehammer/pyodbc/wiki/Unicode
+            pyodbc_SQL_CHAR = 1  # pyodbc.SQL_CHAR
+            pyodbc_SQL_WCHAR = -8  # pyodbc.SQL_WCHAR
+            conn.setdecoding(pyodbc_SQL_CHAR, encoding="utf-8")
+            conn.setdecoding(pyodbc_SQL_WCHAR, encoding="utf-8")
+            conn.setencoding(encoding="utf-8")
+
+        return on_connect
+
+
+dialect = MySQLDialect_pyodbc
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reflection.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reflection.py
new file mode 100644
index 00000000..3998be97
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reflection.py
@@ -0,0 +1,677 @@
+# dialects/mysql/reflection.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import re
+
+from .enumerated import ENUM
+from .enumerated import SET
+from .types import DATETIME
+from .types import TIME
+from .types import TIMESTAMP
+from ... import log
+from ... import types as sqltypes
+from ... import util
+
+
+class ReflectedState:
+    """Stores raw information about a SHOW CREATE TABLE statement."""
+
+    def __init__(self):
+        self.columns = []
+        self.table_options = {}
+        self.table_name = None
+        self.keys = []
+        self.fk_constraints = []
+        self.ck_constraints = []
+
+
+@log.class_logger
+class MySQLTableDefinitionParser:
+    """Parses the results of a SHOW CREATE TABLE statement."""
+
+    def __init__(self, dialect, preparer):
+        self.dialect = dialect
+        self.preparer = preparer
+        self._prep_regexes()
+
+    def parse(self, show_create, charset):
+        state = ReflectedState()
+        state.charset = charset
+        for line in re.split(r"\r?\n", show_create):
+            if line.startswith("  " + self.preparer.initial_quote):
+                self._parse_column(line, state)
+            # a regular table options line
+            elif line.startswith(") "):
+                self._parse_table_options(line, state)
+            # an ANSI-mode table options line
+            elif line == ")":
+                pass
+            elif line.startswith("CREATE "):
+                self._parse_table_name(line, state)
+            elif "PARTITION" in line:
+                self._parse_partition_options(line, state)
+            # Not present in real reflection, but may be if
+            # loading from a file.
+            elif not line:
+                pass
+            else:
+                type_, spec = self._parse_constraints(line)
+                if type_ is None:
+                    util.warn("Unknown schema content: %r" % line)
+                elif type_ == "key":
+                    state.keys.append(spec)
+                elif type_ == "fk_constraint":
+                    state.fk_constraints.append(spec)
+                elif type_ == "ck_constraint":
+                    state.ck_constraints.append(spec)
+                else:
+                    pass
+        return state
+
+    def _check_view(self, sql: str) -> bool:
+        return bool(self._re_is_view.match(sql))
+
+    def _parse_constraints(self, line):
+        """Parse a KEY or CONSTRAINT line.
+
+        :param line: A line of SHOW CREATE TABLE output
+        """
+
+        # KEY
+        m = self._re_key.match(line)
+        if m:
+            spec = m.groupdict()
+            # convert columns into name, length pairs
+            # NOTE: we may want to consider SHOW INDEX as the
+            # format of indexes in MySQL becomes more complex
+            spec["columns"] = self._parse_keyexprs(spec["columns"])
+            if spec["version_sql"]:
+                m2 = self._re_key_version_sql.match(spec["version_sql"])
+                if m2 and m2.groupdict()["parser"]:
+                    spec["parser"] = m2.groupdict()["parser"]
+            if spec["parser"]:
+                spec["parser"] = self.preparer.unformat_identifiers(
+                    spec["parser"]
+                )[0]
+            return "key", spec
+
+        # FOREIGN KEY CONSTRAINT
+        m = self._re_fk_constraint.match(line)
+        if m:
+            spec = m.groupdict()
+            spec["table"] = self.preparer.unformat_identifiers(spec["table"])
+            spec["local"] = [c[0] for c in self._parse_keyexprs(spec["local"])]
+            spec["foreign"] = [
+                c[0] for c in self._parse_keyexprs(spec["foreign"])
+            ]
+            return "fk_constraint", spec
+
+        # CHECK constraint
+        m = self._re_ck_constraint.match(line)
+        if m:
+            spec = m.groupdict()
+            return "ck_constraint", spec
+
+        # PARTITION and SUBPARTITION
+        m = self._re_partition.match(line)
+        if m:
+            # Punt!
+            return "partition", line
+
+        # No match.
+        return (None, line)
+
+    def _parse_table_name(self, line, state):
+        """Extract the table name.
+
+        :param line: The first line of SHOW CREATE TABLE
+        """
+
+        regex, cleanup = self._pr_name
+        m = regex.match(line)
+        if m:
+            state.table_name = cleanup(m.group("name"))
+
+    def _parse_table_options(self, line, state):
+        """Build a dictionary of all reflected table-level options.
+
+        :param line: The final line of SHOW CREATE TABLE output.
+        """
+
+        options = {}
+
+        if line and line != ")":
+            rest_of_line = line
+            for regex, cleanup in self._pr_options:
+                m = regex.search(rest_of_line)
+                if not m:
+                    continue
+                directive, value = m.group("directive"), m.group("val")
+                if cleanup:
+                    value = cleanup(value)
+                options[directive.lower()] = value
+                rest_of_line = regex.sub("", rest_of_line)
+
+        for nope in ("auto_increment", "data directory", "index directory"):
+            options.pop(nope, None)
+
+        for opt, val in options.items():
+            state.table_options["%s_%s" % (self.dialect.name, opt)] = val
+
+    def _parse_partition_options(self, line, state):
+        options = {}
+        new_line = line[:]
+
+        while new_line.startswith("(") or new_line.startswith(" "):
+            new_line = new_line[1:]
+
+        for regex, cleanup in self._pr_options:
+            m = regex.search(new_line)
+            if not m or "PARTITION" not in regex.pattern:
+                continue
+
+            directive = m.group("directive")
+            directive = directive.lower()
+            is_subpartition = directive == "subpartition"
+
+            if directive == "partition" or is_subpartition:
+                new_line = new_line.replace(") */", "")
+                new_line = new_line.replace(",", "")
+                if is_subpartition and new_line.endswith(")"):
+                    new_line = new_line[:-1]
+                if self.dialect.name == "mariadb" and new_line.endswith(")"):
+                    if (
+                        "MAXVALUE" in new_line
+                        or "MINVALUE" in new_line
+                        or "ENGINE" in new_line
+                    ):
+                        # final line of MariaDB partition endswith ")"
+                        new_line = new_line[:-1]
+
+                defs = "%s_%s_definitions" % (self.dialect.name, directive)
+                options[defs] = new_line
+
+            else:
+                directive = directive.replace(" ", "_")
+                value = m.group("val")
+                if cleanup:
+                    value = cleanup(value)
+                options[directive] = value
+            break
+
+        for opt, val in options.items():
+            part_def = "%s_partition_definitions" % (self.dialect.name)
+            subpart_def = "%s_subpartition_definitions" % (self.dialect.name)
+            if opt == part_def or opt == subpart_def:
+                # builds a string of definitions
+                if opt not in state.table_options:
+                    state.table_options[opt] = val
+                else:
+                    state.table_options[opt] = "%s, %s" % (
+                        state.table_options[opt],
+                        val,
+                    )
+            else:
+                state.table_options["%s_%s" % (self.dialect.name, opt)] = val
+
+    def _parse_column(self, line, state):
+        """Extract column details.
+
+        Falls back to a 'minimal support' variant if full parse fails.
+
+        :param line: Any column-bearing line from SHOW CREATE TABLE
+        """
+
+        spec = None
+        m = self._re_column.match(line)
+        if m:
+            spec = m.groupdict()
+            spec["full"] = True
+        else:
+            m = self._re_column_loose.match(line)
+            if m:
+                spec = m.groupdict()
+                spec["full"] = False
+        if not spec:
+            util.warn("Unknown column definition %r" % line)
+            return
+        if not spec["full"]:
+            util.warn("Incomplete reflection of column definition %r" % line)
+
+        name, type_, args = spec["name"], spec["coltype"], spec["arg"]
+
+        try:
+            col_type = self.dialect.ischema_names[type_]
+        except KeyError:
+            util.warn(
+                "Did not recognize type '%s' of column '%s'" % (type_, name)
+            )
+            col_type = sqltypes.NullType
+
+        # Column type positional arguments eg. varchar(32)
+        if args is None or args == "":
+            type_args = []
+        elif args[0] == "'" and args[-1] == "'":
+            type_args = self._re_csv_str.findall(args)
+        else:
+            type_args = [int(v) for v in self._re_csv_int.findall(args)]
+
+        # Column type keyword options
+        type_kw = {}
+
+        if issubclass(col_type, (DATETIME, TIME, TIMESTAMP)):
+            if type_args:
+                type_kw["fsp"] = type_args.pop(0)
+
+        for kw in ("unsigned", "zerofill"):
+            if spec.get(kw, False):
+                type_kw[kw] = True
+        for kw in ("charset", "collate"):
+            if spec.get(kw, False):
+                type_kw[kw] = spec[kw]
+        if issubclass(col_type, (ENUM, SET)):
+            type_args = _strip_values(type_args)
+
+            if issubclass(col_type, SET) and "" in type_args:
+                type_kw["retrieve_as_bitwise"] = True
+
+        type_instance = col_type(*type_args, **type_kw)
+
+        col_kw = {}
+
+        # NOT NULL
+        col_kw["nullable"] = True
+        # this can be "NULL" in the case of TIMESTAMP
+        if spec.get("notnull", False) == "NOT NULL":
+            col_kw["nullable"] = False
+        # For generated columns, the nullability is marked in a different place
+        if spec.get("notnull_generated", False) == "NOT NULL":
+            col_kw["nullable"] = False
+
+        # AUTO_INCREMENT
+        if spec.get("autoincr", False):
+            col_kw["autoincrement"] = True
+        elif issubclass(col_type, sqltypes.Integer):
+            col_kw["autoincrement"] = False
+
+        # DEFAULT
+        default = spec.get("default", None)
+
+        if default == "NULL":
+            # eliminates the need to deal with this later.
+            default = None
+
+        comment = spec.get("comment", None)
+
+        if comment is not None:
+            comment = cleanup_text(comment)
+
+        sqltext = spec.get("generated")
+        if sqltext is not None:
+            computed = dict(sqltext=sqltext)
+            persisted = spec.get("persistence")
+            if persisted is not None:
+                computed["persisted"] = persisted == "STORED"
+            col_kw["computed"] = computed
+
+        col_d = dict(
+            name=name, type=type_instance, default=default, comment=comment
+        )
+        col_d.update(col_kw)
+        state.columns.append(col_d)
+
+    def _describe_to_create(self, table_name, columns):
+        """Re-format DESCRIBE output as a SHOW CREATE TABLE string.
+
+        DESCRIBE is a much simpler reflection and is sufficient for
+        reflecting views for runtime use.  This method formats DDL
+        for columns only- keys are omitted.
+
+        :param columns: A sequence of DESCRIBE or SHOW COLUMNS 6-tuples.
+          SHOW FULL COLUMNS FROM rows must be rearranged for use with
+          this function.
+        """
+
+        buffer = []
+        for row in columns:
+            (name, col_type, nullable, default, extra) = (
+                row[i] for i in (0, 1, 2, 4, 5)
+            )
+
+            line = [" "]
+            line.append(self.preparer.quote_identifier(name))
+            line.append(col_type)
+            if not nullable:
+                line.append("NOT NULL")
+            if default:
+                if "auto_increment" in default:
+                    pass
+                elif col_type.startswith("timestamp") and default.startswith(
+                    "C"
+                ):
+                    line.append("DEFAULT")
+                    line.append(default)
+                elif default == "NULL":
+                    line.append("DEFAULT")
+                    line.append(default)
+                else:
+                    line.append("DEFAULT")
+                    line.append("'%s'" % default.replace("'", "''"))
+            if extra:
+                line.append(extra)
+
+            buffer.append(" ".join(line))
+
+        return "".join(
+            [
+                (
+                    "CREATE TABLE %s (\n"
+                    % self.preparer.quote_identifier(table_name)
+                ),
+                ",\n".join(buffer),
+                "\n) ",
+            ]
+        )
+
+    def _parse_keyexprs(self, identifiers):
+        """Unpack '"col"(2),"col" ASC'-ish strings into components."""
+
+        return [
+            (colname, int(length) if length else None, modifiers)
+            for colname, length, modifiers in self._re_keyexprs.findall(
+                identifiers
+            )
+        ]
+
+    def _prep_regexes(self):
+        """Pre-compile regular expressions."""
+
+        self._re_columns = []
+        self._pr_options = []
+
+        _final = self.preparer.final_quote
+
+        quotes = dict(
+            zip(
+                ("iq", "fq", "esc_fq"),
+                [
+                    re.escape(s)
+                    for s in (
+                        self.preparer.initial_quote,
+                        _final,
+                        self.preparer._escape_identifier(_final),
+                    )
+                ],
+            )
+        )
+
+        self._pr_name = _pr_compile(
+            r"^CREATE (?:\w+ +)?TABLE +"
+            r"%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s +\($" % quotes,
+            self.preparer._unescape_identifier,
+        )
+
+        self._re_is_view = _re_compile(r"^CREATE(?! TABLE)(\s.*)?\sVIEW")
+
+        # `col`,`col2`(32),`col3`(15) DESC
+        #
+        self._re_keyexprs = _re_compile(
+            r"(?:"
+            r"(?:%(iq)s((?:%(esc_fq)s|[^%(fq)s])+)%(fq)s)"
+            r"(?:\((\d+)\))?(?: +(ASC|DESC))?(?=\,|$))+" % quotes
+        )
+
+        # 'foo' or 'foo','bar' or 'fo,o','ba''a''r'
+        self._re_csv_str = _re_compile(r"\x27(?:\x27\x27|[^\x27])*\x27")
+
+        # 123 or 123,456
+        self._re_csv_int = _re_compile(r"\d+")
+
+        # `colname` <type> [type opts]
+        #  (NOT NULL | NULL)
+        #   DEFAULT ('value' | CURRENT_TIMESTAMP...)
+        #   COMMENT 'comment'
+        #  COLUMN_FORMAT (FIXED|DYNAMIC|DEFAULT)
+        #  STORAGE (DISK|MEMORY)
+        self._re_column = _re_compile(
+            r"  "
+            r"%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s +"
+            r"(?P<coltype>\w+)"
+            r"(?:\((?P<arg>(?:\d+|\d+,\d+|"
+            r"(?:'(?:''|[^'])*',?)+))\))?"
+            r"(?: +(?P<unsigned>UNSIGNED))?"
+            r"(?: +(?P<zerofill>ZEROFILL))?"
+            r"(?: +CHARACTER SET +(?P<charset>[\w_]+))?"
+            r"(?: +COLLATE +(?P<collate>[\w_]+))?"
+            r"(?: +(?P<notnull>(?:NOT )?NULL))?"
+            r"(?: +DEFAULT +(?P<default>"
+            r"(?:NULL|'(?:''|[^'])*'|[\-\w\.\(\)]+"
+            r"(?: +ON UPDATE [\-\w\.\(\)]+)?)"
+            r"))?"
+            r"(?: +(?:GENERATED ALWAYS)? ?AS +(?P<generated>\("
+            r".*\))? ?(?P<persistence>VIRTUAL|STORED)?"
+            r"(?: +(?P<notnull_generated>(?:NOT )?NULL))?"
+            r")?"
+            r"(?: +(?P<autoincr>AUTO_INCREMENT))?"
+            r"(?: +COMMENT +'(?P<comment>(?:''|[^'])*)')?"
+            r"(?: +COLUMN_FORMAT +(?P<colfmt>\w+))?"
+            r"(?: +STORAGE +(?P<storage>\w+))?"
+            r"(?: +(?P<extra>.*))?"
+            r",?$" % quotes
+        )
+
+        # Fallback, try to parse as little as possible
+        self._re_column_loose = _re_compile(
+            r"  "
+            r"%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s +"
+            r"(?P<coltype>\w+)"
+            r"(?:\((?P<arg>(?:\d+|\d+,\d+|\x27(?:\x27\x27|[^\x27])+\x27))\))?"
+            r".*?(?P<notnull>(?:NOT )NULL)?" % quotes
+        )
+
+        # (PRIMARY|UNIQUE|FULLTEXT|SPATIAL) INDEX `name` (USING (BTREE|HASH))?
+        # (`col` (ASC|DESC)?, `col` (ASC|DESC)?)
+        # KEY_BLOCK_SIZE size | WITH PARSER name  /*!50100 WITH PARSER name */
+        self._re_key = _re_compile(
+            r"  "
+            r"(?:(?P<type>\S+) )?KEY"
+            r"(?: +%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s)?"
+            r"(?: +USING +(?P<using_pre>\S+))?"
+            r" +\((?P<columns>.+?)\)"
+            r"(?: +USING +(?P<using_post>\S+))?"
+            r"(?: +KEY_BLOCK_SIZE *[ =]? *(?P<keyblock>\S+))?"
+            r"(?: +WITH PARSER +(?P<parser>\S+))?"
+            r"(?: +COMMENT +(?P<comment>(\x27\x27|\x27([^\x27])*?\x27)+))?"
+            r"(?: +/\*(?P<version_sql>.+)\*/ *)?"
+            r",?$" % quotes
+        )
+
+        # https://forums.mysql.com/read.php?20,567102,567111#msg-567111
+        # It means if the MySQL version >= \d+, execute what's in the comment
+        self._re_key_version_sql = _re_compile(
+            r"\!\d+ " r"(?: *WITH PARSER +(?P<parser>\S+) *)?"
+        )
+
+        # CONSTRAINT `name` FOREIGN KEY (`local_col`)
+        # REFERENCES `remote` (`remote_col`)
+        # MATCH FULL | MATCH PARTIAL | MATCH SIMPLE
+        # ON DELETE CASCADE ON UPDATE RESTRICT
+        #
+        # unique constraints come back as KEYs
+        kw = quotes.copy()
+        kw["on"] = "RESTRICT|CASCADE|SET NULL|NO ACTION|SET DEFAULT"
+        self._re_fk_constraint = _re_compile(
+            r"  "
+            r"CONSTRAINT +"
+            r"%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s +"
+            r"FOREIGN KEY +"
+            r"\((?P<local>[^\)]+?)\) REFERENCES +"
+            r"(?P<table>%(iq)s[^%(fq)s]+%(fq)s"
+            r"(?:\.%(iq)s[^%(fq)s]+%(fq)s)?) +"
+            r"\((?P<foreign>(?:%(iq)s[^%(fq)s]+%(fq)s(?: *, *)?)+)\)"
+            r"(?: +(?P<match>MATCH \w+))?"
+            r"(?: +ON DELETE (?P<ondelete>%(on)s))?"
+            r"(?: +ON UPDATE (?P<onupdate>%(on)s))?" % kw
+        )
+
+        # CONSTRAINT `CONSTRAINT_1` CHECK (`x` > 5)'
+        # testing on MariaDB 10.2 shows that the CHECK constraint
+        # is returned on a line by itself, so to match without worrying
+        # about parenthesis in the expression we go to the end of the line
+        self._re_ck_constraint = _re_compile(
+            r"  "
+            r"CONSTRAINT +"
+            r"%(iq)s(?P<name>(?:%(esc_fq)s|[^%(fq)s])+)%(fq)s +"
+            r"CHECK +"
+            r"\((?P<sqltext>.+)\),?" % kw
+        )
+
+        # PARTITION
+        #
+        # punt!
+        self._re_partition = _re_compile(r"(?:.*)(?:SUB)?PARTITION(?:.*)")
+
+        # Table-level options (COLLATE, ENGINE, etc.)
+        # Do the string options first, since they have quoted
+        # strings we need to get rid of.
+        for option in _options_of_type_string:
+            self._add_option_string(option)
+
+        for option in (
+            "ENGINE",
+            "TYPE",
+            "AUTO_INCREMENT",
+            "AVG_ROW_LENGTH",
+            "CHARACTER SET",
+            "DEFAULT CHARSET",
+            "CHECKSUM",
+            "COLLATE",
+            "DELAY_KEY_WRITE",
+            "INSERT_METHOD",
+            "MAX_ROWS",
+            "MIN_ROWS",
+            "PACK_KEYS",
+            "ROW_FORMAT",
+            "KEY_BLOCK_SIZE",
+            "STATS_SAMPLE_PAGES",
+        ):
+            self._add_option_word(option)
+
+        for option in (
+            "PARTITION BY",
+            "SUBPARTITION BY",
+            "PARTITIONS",
+            "SUBPARTITIONS",
+            "PARTITION",
+            "SUBPARTITION",
+        ):
+            self._add_partition_option_word(option)
+
+        self._add_option_regex("UNION", r"\([^\)]+\)")
+        self._add_option_regex("TABLESPACE", r".*? STORAGE DISK")
+        self._add_option_regex(
+            "RAID_TYPE",
+            r"\w+\s+RAID_CHUNKS\s*\=\s*\w+RAID_CHUNKSIZE\s*=\s*\w+",
+        )
+
+    _optional_equals = r"(?:\s*(?:=\s*)|\s+)"
+
+    def _add_option_string(self, directive):
+        regex = r"(?P<directive>%s)%s" r"'(?P<val>(?:[^']|'')*?)'(?!')" % (
+            re.escape(directive),
+            self._optional_equals,
+        )
+        self._pr_options.append(_pr_compile(regex, cleanup_text))
+
+    def _add_option_word(self, directive):
+        regex = r"(?P<directive>%s)%s" r"(?P<val>\w+)" % (
+            re.escape(directive),
+            self._optional_equals,
+        )
+        self._pr_options.append(_pr_compile(regex))
+
+    def _add_partition_option_word(self, directive):
+        if directive == "PARTITION BY" or directive == "SUBPARTITION BY":
+            regex = r"(?<!\S)(?P<directive>%s)%s" r"(?P<val>\w+.*)" % (
+                re.escape(directive),
+                self._optional_equals,
+            )
+        elif directive == "SUBPARTITIONS" or directive == "PARTITIONS":
+            regex = r"(?<!\S)(?P<directive>%s)%s" r"(?P<val>\d+)" % (
+                re.escape(directive),
+                self._optional_equals,
+            )
+        else:
+            regex = r"(?<!\S)(?P<directive>%s)(?!\S)" % (re.escape(directive),)
+        self._pr_options.append(_pr_compile(regex))
+
+    def _add_option_regex(self, directive, regex):
+        regex = r"(?P<directive>%s)%s" r"(?P<val>%s)" % (
+            re.escape(directive),
+            self._optional_equals,
+            regex,
+        )
+        self._pr_options.append(_pr_compile(regex))
+
+
+_options_of_type_string = (
+    "COMMENT",
+    "DATA DIRECTORY",
+    "INDEX DIRECTORY",
+    "PASSWORD",
+    "CONNECTION",
+)
+
+
+def _pr_compile(regex, cleanup=None):
+    """Prepare a 2-tuple of compiled regex and callable."""
+
+    return (_re_compile(regex), cleanup)
+
+
+def _re_compile(regex):
+    """Compile a string to regex, I and UNICODE."""
+
+    return re.compile(regex, re.I | re.UNICODE)
+
+
+def _strip_values(values):
+    "Strip reflected values quotes"
+    strip_values = []
+    for a in values:
+        if a[0:1] == '"' or a[0:1] == "'":
+            # strip enclosing quotes and unquote interior
+            a = a[1:-1].replace(a[0] * 2, a[0])
+        strip_values.append(a)
+    return strip_values
+
+
+def cleanup_text(raw_text: str) -> str:
+    if "\\" in raw_text:
+        raw_text = re.sub(
+            _control_char_regexp, lambda s: _control_char_map[s[0]], raw_text
+        )
+    return raw_text.replace("''", "'")
+
+
+_control_char_map = {
+    "\\\\": "\\",
+    "\\0": "\0",
+    "\\a": "\a",
+    "\\b": "\b",
+    "\\t": "\t",
+    "\\n": "\n",
+    "\\v": "\v",
+    "\\f": "\f",
+    "\\r": "\r",
+    # '\\e':'\e',
+}
+_control_char_regexp = re.compile(
+    "|".join(re.escape(k) for k in _control_char_map)
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reserved_words.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reserved_words.py
new file mode 100644
index 00000000..34fecf42
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/reserved_words.py
@@ -0,0 +1,571 @@
+# dialects/mysql/reserved_words.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+# generated using:
+# https://gist.github.com/kkirsche/4f31f2153ed7a3248be1ec44ca6ddbc9
+#
+# https://mariadb.com/kb/en/reserved-words/
+# includes: Reserved Words, Oracle Mode (separate set unioned)
+# excludes: Exceptions, Function Names
+# mypy: ignore-errors
+
+RESERVED_WORDS_MARIADB = {
+    "accessible",
+    "add",
+    "all",
+    "alter",
+    "analyze",
+    "and",
+    "as",
+    "asc",
+    "asensitive",
+    "before",
+    "between",
+    "bigint",
+    "binary",
+    "blob",
+    "both",
+    "by",
+    "call",
+    "cascade",
+    "case",
+    "change",
+    "char",
+    "character",
+    "check",
+    "collate",
+    "column",
+    "condition",
+    "constraint",
+    "continue",
+    "convert",
+    "create",
+    "cross",
+    "current_date",
+    "current_role",
+    "current_time",
+    "current_timestamp",
+    "current_user",
+    "cursor",
+    "database",
+    "databases",
+    "day_hour",
+    "day_microsecond",
+    "day_minute",
+    "day_second",
+    "dec",
+    "decimal",
+    "declare",
+    "default",
+    "delayed",
+    "delete",
+    "desc",
+    "describe",
+    "deterministic",
+    "distinct",
+    "distinctrow",
+    "div",
+    "do_domain_ids",
+    "double",
+    "drop",
+    "dual",
+    "each",
+    "else",
+    "elseif",
+    "enclosed",
+    "escaped",
+    "except",
+    "exists",
+    "exit",
+    "explain",
+    "false",
+    "fetch",
+    "float",
+    "float4",
+    "float8",
+    "for",
+    "force",
+    "foreign",
+    "from",
+    "fulltext",
+    "general",
+    "grant",
+    "group",
+    "having",
+    "high_priority",
+    "hour_microsecond",
+    "hour_minute",
+    "hour_second",
+    "if",
+    "ignore",
+    "ignore_domain_ids",
+    "ignore_server_ids",
+    "in",
+    "index",
+    "infile",
+    "inner",
+    "inout",
+    "insensitive",
+    "insert",
+    "int",
+    "int1",
+    "int2",
+    "int3",
+    "int4",
+    "int8",
+    "integer",
+    "intersect",
+    "interval",
+    "into",
+    "is",
+    "iterate",
+    "join",
+    "key",
+    "keys",
+    "kill",
+    "leading",
+    "leave",
+    "left",
+    "like",
+    "limit",
+    "linear",
+    "lines",
+    "load",
+    "localtime",
+    "localtimestamp",
+    "lock",
+    "long",
+    "longblob",
+    "longtext",
+    "loop",
+    "low_priority",
+    "master_heartbeat_period",
+    "master_ssl_verify_server_cert",
+    "match",
+    "maxvalue",
+    "mediumblob",
+    "mediumint",
+    "mediumtext",
+    "middleint",
+    "minute_microsecond",
+    "minute_second",
+    "mod",
+    "modifies",
+    "natural",
+    "no_write_to_binlog",
+    "not",
+    "null",
+    "numeric",
+    "offset",
+    "on",
+    "optimize",
+    "option",
+    "optionally",
+    "or",
+    "order",
+    "out",
+    "outer",
+    "outfile",
+    "over",
+    "page_checksum",
+    "parse_vcol_expr",
+    "partition",
+    "position",
+    "precision",
+    "primary",
+    "procedure",
+    "purge",
+    "range",
+    "read",
+    "read_write",
+    "reads",
+    "real",
+    "recursive",
+    "ref_system_id",
+    "references",
+    "regexp",
+    "release",
+    "rename",
+    "repeat",
+    "replace",
+    "require",
+    "resignal",
+    "restrict",
+    "return",
+    "returning",
+    "revoke",
+    "right",
+    "rlike",
+    "rows",
+    "row_number",
+    "schema",
+    "schemas",
+    "second_microsecond",
+    "select",
+    "sensitive",
+    "separator",
+    "set",
+    "show",
+    "signal",
+    "slow",
+    "smallint",
+    "spatial",
+    "specific",
+    "sql",
+    "sql_big_result",
+    "sql_calc_found_rows",
+    "sql_small_result",
+    "sqlexception",
+    "sqlstate",
+    "sqlwarning",
+    "ssl",
+    "starting",
+    "stats_auto_recalc",
+    "stats_persistent",
+    "stats_sample_pages",
+    "straight_join",
+    "table",
+    "terminated",
+    "then",
+    "tinyblob",
+    "tinyint",
+    "tinytext",
+    "to",
+    "trailing",
+    "trigger",
+    "true",
+    "undo",
+    "union",
+    "unique",
+    "unlock",
+    "unsigned",
+    "update",
+    "usage",
+    "use",
+    "using",
+    "utc_date",
+    "utc_time",
+    "utc_timestamp",
+    "values",
+    "varbinary",
+    "varchar",
+    "varcharacter",
+    "varying",
+    "when",
+    "where",
+    "while",
+    "window",
+    "with",
+    "write",
+    "xor",
+    "year_month",
+    "zerofill",
+}.union(
+    {
+        "body",
+        "elsif",
+        "goto",
+        "history",
+        "others",
+        "package",
+        "period",
+        "raise",
+        "rowtype",
+        "system",
+        "system_time",
+        "versioning",
+        "without",
+    }
+)
+
+# https://dev.mysql.com/doc/refman/8.3/en/keywords.html
+# https://dev.mysql.com/doc/refman/8.0/en/keywords.html
+# https://dev.mysql.com/doc/refman/5.7/en/keywords.html
+# https://dev.mysql.com/doc/refman/5.6/en/keywords.html
+# includes: MySQL x.0 Keywords and Reserved Words
+# excludes: MySQL x.0 New Keywords and Reserved Words,
+#       MySQL x.0 Removed Keywords and Reserved Words
+RESERVED_WORDS_MYSQL = {
+    "accessible",
+    "add",
+    "admin",
+    "all",
+    "alter",
+    "analyze",
+    "and",
+    "array",
+    "as",
+    "asc",
+    "asensitive",
+    "before",
+    "between",
+    "bigint",
+    "binary",
+    "blob",
+    "both",
+    "by",
+    "call",
+    "cascade",
+    "case",
+    "change",
+    "char",
+    "character",
+    "check",
+    "collate",
+    "column",
+    "condition",
+    "constraint",
+    "continue",
+    "convert",
+    "create",
+    "cross",
+    "cube",
+    "cume_dist",
+    "current_date",
+    "current_time",
+    "current_timestamp",
+    "current_user",
+    "cursor",
+    "database",
+    "databases",
+    "day_hour",
+    "day_microsecond",
+    "day_minute",
+    "day_second",
+    "dec",
+    "decimal",
+    "declare",
+    "default",
+    "delayed",
+    "delete",
+    "dense_rank",
+    "desc",
+    "describe",
+    "deterministic",
+    "distinct",
+    "distinctrow",
+    "div",
+    "double",
+    "drop",
+    "dual",
+    "each",
+    "else",
+    "elseif",
+    "empty",
+    "enclosed",
+    "escaped",
+    "except",
+    "exists",
+    "exit",
+    "explain",
+    "false",
+    "fetch",
+    "first_value",
+    "float",
+    "float4",
+    "float8",
+    "for",
+    "force",
+    "foreign",
+    "from",
+    "fulltext",
+    "function",
+    "general",
+    "generated",
+    "get",
+    "get_master_public_key",
+    "grant",
+    "group",
+    "grouping",
+    "groups",
+    "having",
+    "high_priority",
+    "hour_microsecond",
+    "hour_minute",
+    "hour_second",
+    "if",
+    "ignore",
+    "ignore_server_ids",
+    "in",
+    "index",
+    "infile",
+    "inner",
+    "inout",
+    "insensitive",
+    "insert",
+    "int",
+    "int1",
+    "int2",
+    "int3",
+    "int4",
+    "int8",
+    "integer",
+    "intersect",
+    "interval",
+    "into",
+    "io_after_gtids",
+    "io_before_gtids",
+    "is",
+    "iterate",
+    "join",
+    "json_table",
+    "key",
+    "keys",
+    "kill",
+    "lag",
+    "last_value",
+    "lateral",
+    "lead",
+    "leading",
+    "leave",
+    "left",
+    "like",
+    "limit",
+    "linear",
+    "lines",
+    "load",
+    "localtime",
+    "localtimestamp",
+    "lock",
+    "long",
+    "longblob",
+    "longtext",
+    "loop",
+    "low_priority",
+    "master_bind",
+    "master_heartbeat_period",
+    "master_ssl_verify_server_cert",
+    "match",
+    "maxvalue",
+    "mediumblob",
+    "mediumint",
+    "mediumtext",
+    "member",
+    "middleint",
+    "minute_microsecond",
+    "minute_second",
+    "mod",
+    "modifies",
+    "natural",
+    "no_write_to_binlog",
+    "not",
+    "nth_value",
+    "ntile",
+    "null",
+    "numeric",
+    "of",
+    "on",
+    "optimize",
+    "optimizer_costs",
+    "option",
+    "optionally",
+    "or",
+    "order",
+    "out",
+    "outer",
+    "outfile",
+    "over",
+    "parse_gcol_expr",
+    "parallel",
+    "partition",
+    "percent_rank",
+    "persist",
+    "persist_only",
+    "precision",
+    "primary",
+    "procedure",
+    "purge",
+    "qualify",
+    "range",
+    "rank",
+    "read",
+    "read_write",
+    "reads",
+    "real",
+    "recursive",
+    "references",
+    "regexp",
+    "release",
+    "rename",
+    "repeat",
+    "replace",
+    "require",
+    "resignal",
+    "restrict",
+    "return",
+    "revoke",
+    "right",
+    "rlike",
+    "role",
+    "row",
+    "row_number",
+    "rows",
+    "schema",
+    "schemas",
+    "second_microsecond",
+    "select",
+    "sensitive",
+    "separator",
+    "set",
+    "show",
+    "signal",
+    "slow",
+    "smallint",
+    "spatial",
+    "specific",
+    "sql",
+    "sql_after_gtids",
+    "sql_before_gtids",
+    "sql_big_result",
+    "sql_calc_found_rows",
+    "sql_small_result",
+    "sqlexception",
+    "sqlstate",
+    "sqlwarning",
+    "ssl",
+    "starting",
+    "stored",
+    "straight_join",
+    "system",
+    "table",
+    "terminated",
+    "then",
+    "tinyblob",
+    "tinyint",
+    "tinytext",
+    "to",
+    "trailing",
+    "trigger",
+    "true",
+    "undo",
+    "union",
+    "unique",
+    "unlock",
+    "unsigned",
+    "update",
+    "usage",
+    "use",
+    "using",
+    "utc_date",
+    "utc_time",
+    "utc_timestamp",
+    "values",
+    "varbinary",
+    "varchar",
+    "varcharacter",
+    "varying",
+    "virtual",
+    "when",
+    "where",
+    "while",
+    "window",
+    "with",
+    "write",
+    "xor",
+    "year_month",
+    "zerofill",
+}
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/types.py
new file mode 100644
index 00000000..0c05aacb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/mysql/types.py
@@ -0,0 +1,774 @@
+# dialects/mysql/types.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import datetime
+
+from ... import exc
+from ... import util
+from ...sql import sqltypes
+
+
+class _NumericType:
+    """Base for MySQL numeric types.
+
+    This is the base both for NUMERIC as well as INTEGER, hence
+    it's a mixin.
+
+    """
+
+    def __init__(self, unsigned=False, zerofill=False, **kw):
+        self.unsigned = unsigned
+        self.zerofill = zerofill
+        super().__init__(**kw)
+
+    def __repr__(self):
+        return util.generic_repr(
+            self, to_inspect=[_NumericType, sqltypes.Numeric]
+        )
+
+
+class _FloatType(_NumericType, sqltypes.Float):
+    def __init__(self, precision=None, scale=None, asdecimal=True, **kw):
+        if isinstance(self, (REAL, DOUBLE)) and (
+            (precision is None and scale is not None)
+            or (precision is not None and scale is None)
+        ):
+            raise exc.ArgumentError(
+                "You must specify both precision and scale or omit "
+                "both altogether."
+            )
+        super().__init__(precision=precision, asdecimal=asdecimal, **kw)
+        self.scale = scale
+
+    def __repr__(self):
+        return util.generic_repr(
+            self, to_inspect=[_FloatType, _NumericType, sqltypes.Float]
+        )
+
+
+class _IntegerType(_NumericType, sqltypes.Integer):
+    def __init__(self, display_width=None, **kw):
+        self.display_width = display_width
+        super().__init__(**kw)
+
+    def __repr__(self):
+        return util.generic_repr(
+            self, to_inspect=[_IntegerType, _NumericType, sqltypes.Integer]
+        )
+
+
+class _StringType(sqltypes.String):
+    """Base for MySQL string types."""
+
+    def __init__(
+        self,
+        charset=None,
+        collation=None,
+        ascii=False,  # noqa
+        binary=False,
+        unicode=False,
+        national=False,
+        **kw,
+    ):
+        self.charset = charset
+
+        # allow collate= or collation=
+        kw.setdefault("collation", kw.pop("collate", collation))
+
+        self.ascii = ascii
+        self.unicode = unicode
+        self.binary = binary
+        self.national = national
+        super().__init__(**kw)
+
+    def __repr__(self):
+        return util.generic_repr(
+            self, to_inspect=[_StringType, sqltypes.String]
+        )
+
+
+class _MatchType(sqltypes.Float, sqltypes.MatchType):
+    def __init__(self, **kw):
+        # TODO: float arguments?
+        sqltypes.Float.__init__(self)
+        sqltypes.MatchType.__init__(self)
+
+
+class NUMERIC(_NumericType, sqltypes.NUMERIC):
+    """MySQL NUMERIC type."""
+
+    __visit_name__ = "NUMERIC"
+
+    def __init__(self, precision=None, scale=None, asdecimal=True, **kw):
+        """Construct a NUMERIC.
+
+        :param precision: Total digits in this number.  If scale and precision
+          are both None, values are stored to limits allowed by the server.
+
+        :param scale: The number of digits after the decimal point.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(
+            precision=precision, scale=scale, asdecimal=asdecimal, **kw
+        )
+
+
+class DECIMAL(_NumericType, sqltypes.DECIMAL):
+    """MySQL DECIMAL type."""
+
+    __visit_name__ = "DECIMAL"
+
+    def __init__(self, precision=None, scale=None, asdecimal=True, **kw):
+        """Construct a DECIMAL.
+
+        :param precision: Total digits in this number.  If scale and precision
+          are both None, values are stored to limits allowed by the server.
+
+        :param scale: The number of digits after the decimal point.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(
+            precision=precision, scale=scale, asdecimal=asdecimal, **kw
+        )
+
+
+class DOUBLE(_FloatType, sqltypes.DOUBLE):
+    """MySQL DOUBLE type."""
+
+    __visit_name__ = "DOUBLE"
+
+    def __init__(self, precision=None, scale=None, asdecimal=True, **kw):
+        """Construct a DOUBLE.
+
+        .. note::
+
+            The :class:`.DOUBLE` type by default converts from float
+            to Decimal, using a truncation that defaults to 10 digits.
+            Specify either ``scale=n`` or ``decimal_return_scale=n`` in order
+            to change this scale, or ``asdecimal=False`` to return values
+            directly as Python floating points.
+
+        :param precision: Total digits in this number.  If scale and precision
+          are both None, values are stored to limits allowed by the server.
+
+        :param scale: The number of digits after the decimal point.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(
+            precision=precision, scale=scale, asdecimal=asdecimal, **kw
+        )
+
+
+class REAL(_FloatType, sqltypes.REAL):
+    """MySQL REAL type."""
+
+    __visit_name__ = "REAL"
+
+    def __init__(self, precision=None, scale=None, asdecimal=True, **kw):
+        """Construct a REAL.
+
+        .. note::
+
+            The :class:`.REAL` type by default converts from float
+            to Decimal, using a truncation that defaults to 10 digits.
+            Specify either ``scale=n`` or ``decimal_return_scale=n`` in order
+            to change this scale, or ``asdecimal=False`` to return values
+            directly as Python floating points.
+
+        :param precision: Total digits in this number.  If scale and precision
+          are both None, values are stored to limits allowed by the server.
+
+        :param scale: The number of digits after the decimal point.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(
+            precision=precision, scale=scale, asdecimal=asdecimal, **kw
+        )
+
+
+class FLOAT(_FloatType, sqltypes.FLOAT):
+    """MySQL FLOAT type."""
+
+    __visit_name__ = "FLOAT"
+
+    def __init__(self, precision=None, scale=None, asdecimal=False, **kw):
+        """Construct a FLOAT.
+
+        :param precision: Total digits in this number.  If scale and precision
+          are both None, values are stored to limits allowed by the server.
+
+        :param scale: The number of digits after the decimal point.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(
+            precision=precision, scale=scale, asdecimal=asdecimal, **kw
+        )
+
+    def bind_processor(self, dialect):
+        return None
+
+
+class INTEGER(_IntegerType, sqltypes.INTEGER):
+    """MySQL INTEGER type."""
+
+    __visit_name__ = "INTEGER"
+
+    def __init__(self, display_width=None, **kw):
+        """Construct an INTEGER.
+
+        :param display_width: Optional, maximum display width for this number.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(display_width=display_width, **kw)
+
+
+class BIGINT(_IntegerType, sqltypes.BIGINT):
+    """MySQL BIGINTEGER type."""
+
+    __visit_name__ = "BIGINT"
+
+    def __init__(self, display_width=None, **kw):
+        """Construct a BIGINTEGER.
+
+        :param display_width: Optional, maximum display width for this number.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(display_width=display_width, **kw)
+
+
+class MEDIUMINT(_IntegerType):
+    """MySQL MEDIUMINTEGER type."""
+
+    __visit_name__ = "MEDIUMINT"
+
+    def __init__(self, display_width=None, **kw):
+        """Construct a MEDIUMINTEGER
+
+        :param display_width: Optional, maximum display width for this number.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(display_width=display_width, **kw)
+
+
+class TINYINT(_IntegerType):
+    """MySQL TINYINT type."""
+
+    __visit_name__ = "TINYINT"
+
+    def __init__(self, display_width=None, **kw):
+        """Construct a TINYINT.
+
+        :param display_width: Optional, maximum display width for this number.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(display_width=display_width, **kw)
+
+
+class SMALLINT(_IntegerType, sqltypes.SMALLINT):
+    """MySQL SMALLINTEGER type."""
+
+    __visit_name__ = "SMALLINT"
+
+    def __init__(self, display_width=None, **kw):
+        """Construct a SMALLINTEGER.
+
+        :param display_width: Optional, maximum display width for this number.
+
+        :param unsigned: a boolean, optional.
+
+        :param zerofill: Optional. If true, values will be stored as strings
+          left-padded with zeros. Note that this does not effect the values
+          returned by the underlying database API, which continue to be
+          numeric.
+
+        """
+        super().__init__(display_width=display_width, **kw)
+
+
+class BIT(sqltypes.TypeEngine):
+    """MySQL BIT type.
+
+    This type is for MySQL 5.0.3 or greater for MyISAM, and 5.0.5 or greater
+    for MyISAM, MEMORY, InnoDB and BDB.  For older versions, use a
+    MSTinyInteger() type.
+
+    """
+
+    __visit_name__ = "BIT"
+
+    def __init__(self, length=None):
+        """Construct a BIT.
+
+        :param length: Optional, number of bits.
+
+        """
+        self.length = length
+
+    def result_processor(self, dialect, coltype):
+        """Convert a MySQL's 64 bit, variable length binary string to a long.
+
+        TODO: this is MySQL-db, pyodbc specific.  OurSQL and mysqlconnector
+        already do this, so this logic should be moved to those dialects.
+
+        """
+
+        def process(value):
+            if value is not None:
+                v = 0
+                for i in value:
+                    if not isinstance(i, int):
+                        i = ord(i)  # convert byte to int on Python 2
+                    v = v << 8 | i
+                return v
+            return value
+
+        return process
+
+
+class TIME(sqltypes.TIME):
+    """MySQL TIME type."""
+
+    __visit_name__ = "TIME"
+
+    def __init__(self, timezone=False, fsp=None):
+        """Construct a MySQL TIME type.
+
+        :param timezone: not used by the MySQL dialect.
+        :param fsp: fractional seconds precision value.
+         MySQL 5.6 supports storage of fractional seconds;
+         this parameter will be used when emitting DDL
+         for the TIME type.
+
+         .. note::
+
+            DBAPI driver support for fractional seconds may
+            be limited; current support includes
+            MySQL Connector/Python.
+
+        """
+        super().__init__(timezone=timezone)
+        self.fsp = fsp
+
+    def result_processor(self, dialect, coltype):
+        time = datetime.time
+
+        def process(value):
+            # convert from a timedelta value
+            if value is not None:
+                microseconds = value.microseconds
+                seconds = value.seconds
+                minutes = seconds // 60
+                return time(
+                    minutes // 60,
+                    minutes % 60,
+                    seconds - minutes * 60,
+                    microsecond=microseconds,
+                )
+            else:
+                return None
+
+        return process
+
+
+class TIMESTAMP(sqltypes.TIMESTAMP):
+    """MySQL TIMESTAMP type."""
+
+    __visit_name__ = "TIMESTAMP"
+
+    def __init__(self, timezone=False, fsp=None):
+        """Construct a MySQL TIMESTAMP type.
+
+        :param timezone: not used by the MySQL dialect.
+        :param fsp: fractional seconds precision value.
+         MySQL 5.6.4 supports storage of fractional seconds;
+         this parameter will be used when emitting DDL
+         for the TIMESTAMP type.
+
+         .. note::
+
+            DBAPI driver support for fractional seconds may
+            be limited; current support includes
+            MySQL Connector/Python.
+
+        """
+        super().__init__(timezone=timezone)
+        self.fsp = fsp
+
+
+class DATETIME(sqltypes.DATETIME):
+    """MySQL DATETIME type."""
+
+    __visit_name__ = "DATETIME"
+
+    def __init__(self, timezone=False, fsp=None):
+        """Construct a MySQL DATETIME type.
+
+        :param timezone: not used by the MySQL dialect.
+        :param fsp: fractional seconds precision value.
+         MySQL 5.6.4 supports storage of fractional seconds;
+         this parameter will be used when emitting DDL
+         for the DATETIME type.
+
+         .. note::
+
+            DBAPI driver support for fractional seconds may
+            be limited; current support includes
+            MySQL Connector/Python.
+
+        """
+        super().__init__(timezone=timezone)
+        self.fsp = fsp
+
+
+class YEAR(sqltypes.TypeEngine):
+    """MySQL YEAR type, for single byte storage of years 1901-2155."""
+
+    __visit_name__ = "YEAR"
+
+    def __init__(self, display_width=None):
+        self.display_width = display_width
+
+
+class TEXT(_StringType, sqltypes.TEXT):
+    """MySQL TEXT type, for character storage encoded up to 2^16 bytes."""
+
+    __visit_name__ = "TEXT"
+
+    def __init__(self, length=None, **kw):
+        """Construct a TEXT.
+
+        :param length: Optional, if provided the server may optimize storage
+          by substituting the smallest TEXT type sufficient to store
+          ``length`` bytes of characters.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param national: Optional. If true, use the server's configured
+          national character set.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        super().__init__(length=length, **kw)
+
+
+class TINYTEXT(_StringType):
+    """MySQL TINYTEXT type, for character storage encoded up to 2^8 bytes."""
+
+    __visit_name__ = "TINYTEXT"
+
+    def __init__(self, **kwargs):
+        """Construct a TINYTEXT.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param national: Optional. If true, use the server's configured
+          national character set.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        super().__init__(**kwargs)
+
+
+class MEDIUMTEXT(_StringType):
+    """MySQL MEDIUMTEXT type, for character storage encoded up
+    to 2^24 bytes."""
+
+    __visit_name__ = "MEDIUMTEXT"
+
+    def __init__(self, **kwargs):
+        """Construct a MEDIUMTEXT.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param national: Optional. If true, use the server's configured
+          national character set.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        super().__init__(**kwargs)
+
+
+class LONGTEXT(_StringType):
+    """MySQL LONGTEXT type, for character storage encoded up to 2^32 bytes."""
+
+    __visit_name__ = "LONGTEXT"
+
+    def __init__(self, **kwargs):
+        """Construct a LONGTEXT.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param national: Optional. If true, use the server's configured
+          national character set.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        super().__init__(**kwargs)
+
+
+class VARCHAR(_StringType, sqltypes.VARCHAR):
+    """MySQL VARCHAR type, for variable-length character data."""
+
+    __visit_name__ = "VARCHAR"
+
+    def __init__(self, length=None, **kwargs):
+        """Construct a VARCHAR.
+
+        :param charset: Optional, a column-level character set for this string
+          value.  Takes precedence to 'ascii' or 'unicode' short-hand.
+
+        :param collation: Optional, a column-level collation for this string
+          value.  Takes precedence to 'binary' short-hand.
+
+        :param ascii: Defaults to False: short-hand for the ``latin1``
+          character set, generates ASCII in schema.
+
+        :param unicode: Defaults to False: short-hand for the ``ucs2``
+          character set, generates UNICODE in schema.
+
+        :param national: Optional. If true, use the server's configured
+          national character set.
+
+        :param binary: Defaults to False: short-hand, pick the binary
+          collation type that matches the column's character set.  Generates
+          BINARY in schema.  This does not affect the type of data stored,
+          only the collation of character data.
+
+        """
+        super().__init__(length=length, **kwargs)
+
+
+class CHAR(_StringType, sqltypes.CHAR):
+    """MySQL CHAR type, for fixed-length character data."""
+
+    __visit_name__ = "CHAR"
+
+    def __init__(self, length=None, **kwargs):
+        """Construct a CHAR.
+
+        :param length: Maximum data length, in characters.
+
+        :param binary: Optional, use the default binary collation for the
+          national character set.  This does not affect the type of data
+          stored, use a BINARY type for binary data.
+
+        :param collation: Optional, request a particular collation.  Must be
+          compatible with the national character set.
+
+        """
+        super().__init__(length=length, **kwargs)
+
+    @classmethod
+    def _adapt_string_for_cast(cls, type_):
+        # copy the given string type into a CHAR
+        # for the purposes of rendering a CAST expression
+        type_ = sqltypes.to_instance(type_)
+        if isinstance(type_, sqltypes.CHAR):
+            return type_
+        elif isinstance(type_, _StringType):
+            return CHAR(
+                length=type_.length,
+                charset=type_.charset,
+                collation=type_.collation,
+                ascii=type_.ascii,
+                binary=type_.binary,
+                unicode=type_.unicode,
+                national=False,  # not supported in CAST
+            )
+        else:
+            return CHAR(length=type_.length)
+
+
+class NVARCHAR(_StringType, sqltypes.NVARCHAR):
+    """MySQL NVARCHAR type.
+
+    For variable-length character data in the server's configured national
+    character set.
+    """
+
+    __visit_name__ = "NVARCHAR"
+
+    def __init__(self, length=None, **kwargs):
+        """Construct an NVARCHAR.
+
+        :param length: Maximum data length, in characters.
+
+        :param binary: Optional, use the default binary collation for the
+          national character set.  This does not affect the type of data
+          stored, use a BINARY type for binary data.
+
+        :param collation: Optional, request a particular collation.  Must be
+          compatible with the national character set.
+
+        """
+        kwargs["national"] = True
+        super().__init__(length=length, **kwargs)
+
+
+class NCHAR(_StringType, sqltypes.NCHAR):
+    """MySQL NCHAR type.
+
+    For fixed-length character data in the server's configured national
+    character set.
+    """
+
+    __visit_name__ = "NCHAR"
+
+    def __init__(self, length=None, **kwargs):
+        """Construct an NCHAR.
+
+        :param length: Maximum data length, in characters.
+
+        :param binary: Optional, use the default binary collation for the
+          national character set.  This does not affect the type of data
+          stored, use a BINARY type for binary data.
+
+        :param collation: Optional, request a particular collation.  Must be
+          compatible with the national character set.
+
+        """
+        kwargs["national"] = True
+        super().__init__(length=length, **kwargs)
+
+
+class TINYBLOB(sqltypes._Binary):
+    """MySQL TINYBLOB type, for binary data up to 2^8 bytes."""
+
+    __visit_name__ = "TINYBLOB"
+
+
+class MEDIUMBLOB(sqltypes._Binary):
+    """MySQL MEDIUMBLOB type, for binary data up to 2^24 bytes."""
+
+    __visit_name__ = "MEDIUMBLOB"
+
+
+class LONGBLOB(sqltypes._Binary):
+    """MySQL LONGBLOB type, for binary data up to 2^32 bytes."""
+
+    __visit_name__ = "LONGBLOB"
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__init__.py
new file mode 100644
index 00000000..7ceb743d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__init__.py
@@ -0,0 +1,67 @@
+# dialects/oracle/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from types import ModuleType
+
+from . import base  # noqa
+from . import cx_oracle  # noqa
+from . import oracledb  # noqa
+from .base import BFILE
+from .base import BINARY_DOUBLE
+from .base import BINARY_FLOAT
+from .base import BLOB
+from .base import CHAR
+from .base import CLOB
+from .base import DATE
+from .base import DOUBLE_PRECISION
+from .base import FLOAT
+from .base import INTERVAL
+from .base import LONG
+from .base import NCHAR
+from .base import NCLOB
+from .base import NUMBER
+from .base import NVARCHAR
+from .base import NVARCHAR2
+from .base import RAW
+from .base import REAL
+from .base import ROWID
+from .base import TIMESTAMP
+from .base import VARCHAR
+from .base import VARCHAR2
+
+# Alias oracledb also as oracledb_async
+oracledb_async = type(
+    "oracledb_async", (ModuleType,), {"dialect": oracledb.dialect_async}
+)
+
+base.dialect = dialect = cx_oracle.dialect
+
+__all__ = (
+    "VARCHAR",
+    "NVARCHAR",
+    "CHAR",
+    "NCHAR",
+    "DATE",
+    "NUMBER",
+    "BLOB",
+    "BFILE",
+    "CLOB",
+    "NCLOB",
+    "TIMESTAMP",
+    "RAW",
+    "FLOAT",
+    "DOUBLE_PRECISION",
+    "BINARY_DOUBLE",
+    "BINARY_FLOAT",
+    "LONG",
+    "dialect",
+    "INTERVAL",
+    "VARCHAR2",
+    "NVARCHAR2",
+    "ROWID",
+    "REAL",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..2b0cce8f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..162bb4ba
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/cx_oracle.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/cx_oracle.cpython-312.pyc
new file mode 100644
index 00000000..43ea25fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/cx_oracle.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/dictionary.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/dictionary.cpython-312.pyc
new file mode 100644
index 00000000..d1bfe815
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/dictionary.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/oracledb.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/oracledb.cpython-312.pyc
new file mode 100644
index 00000000..34f335f9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/oracledb.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..313080c2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..81ee42cd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/base.py
new file mode 100644
index 00000000..02aa4d53
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/base.py
@@ -0,0 +1,3484 @@
+# dialects/oracle/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+.. dialect:: oracle
+    :name: Oracle Database
+    :normal_support: 11+
+    :best_effort: 9+
+
+
+Auto Increment Behavior
+-----------------------
+
+SQLAlchemy Table objects which include integer primary keys are usually assumed
+to have "autoincrementing" behavior, meaning they can generate their own
+primary key values upon INSERT. For use within Oracle Database, two options are
+available, which are the use of IDENTITY columns (Oracle Database 12 and above
+only) or the association of a SEQUENCE with the column.
+
+Specifying GENERATED AS IDENTITY (Oracle Database 12 and above)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Starting from version 12, Oracle Database can make use of identity columns
+using the :class:`_sql.Identity` to specify the autoincrementing behavior::
+
+    t = Table(
+        "mytable",
+        metadata,
+        Column("id", Integer, Identity(start=3), primary_key=True),
+        Column(...),
+        ...,
+    )
+
+The CREATE TABLE for the above :class:`_schema.Table` object would be:
+
+.. sourcecode:: sql
+
+    CREATE TABLE mytable (
+        id INTEGER GENERATED BY DEFAULT AS IDENTITY (START WITH 3),
+        ...,
+        PRIMARY KEY (id)
+    )
+
+The :class:`_schema.Identity` object support many options to control the
+"autoincrementing" behavior of the column, like the starting value, the
+incrementing value, etc.  In addition to the standard options, Oracle Database
+supports setting :paramref:`_schema.Identity.always` to ``None`` to use the
+default generated mode, rendering GENERATED AS IDENTITY in the DDL. It also supports
+setting :paramref:`_schema.Identity.on_null` to ``True`` to specify ON NULL
+in conjunction with a 'BY DEFAULT' identity column.
+
+Using a SEQUENCE (all Oracle Database versions)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Older version of Oracle Database had no "autoincrement" feature: SQLAlchemy
+relies upon sequences to produce these values.  With the older Oracle Database
+versions, *a sequence must always be explicitly specified to enable
+autoincrement*.  This is divergent with the majority of documentation examples
+which assume the usage of an autoincrement-capable database.  To specify
+sequences, use the sqlalchemy.schema.Sequence object which is passed to a
+Column construct::
+
+  t = Table(
+      "mytable",
+      metadata,
+      Column("id", Integer, Sequence("id_seq", start=1), primary_key=True),
+      Column(...),
+      ...,
+  )
+
+This step is also required when using table reflection, i.e. autoload_with=engine::
+
+  t = Table(
+      "mytable",
+      metadata,
+      Column("id", Integer, Sequence("id_seq", start=1), primary_key=True),
+      autoload_with=engine,
+  )
+
+.. versionchanged::  1.4   Added :class:`_schema.Identity` construct
+   in a :class:`_schema.Column` to specify the option of an autoincrementing
+   column.
+
+.. _oracle_isolation_level:
+
+Transaction Isolation Level / Autocommit
+----------------------------------------
+
+Oracle Database supports "READ COMMITTED" and "SERIALIZABLE" modes of
+isolation. The AUTOCOMMIT isolation level is also supported by the
+python-oracledb and cx_Oracle dialects.
+
+To set using per-connection execution options::
+
+    connection = engine.connect()
+    connection = connection.execution_options(isolation_level="AUTOCOMMIT")
+
+For ``READ COMMITTED`` and ``SERIALIZABLE``, the Oracle Database dialects sets
+the level at the session level using ``ALTER SESSION``, which is reverted back
+to its default setting when the connection is returned to the connection pool.
+
+Valid values for ``isolation_level`` include:
+
+* ``READ COMMITTED``
+* ``AUTOCOMMIT``
+* ``SERIALIZABLE``
+
+.. note:: The implementation for the
+   :meth:`_engine.Connection.get_isolation_level` method as implemented by the
+   Oracle Database dialects necessarily force the start of a transaction using the
+   Oracle Database DBMS_TRANSACTION.LOCAL_TRANSACTION_ID function; otherwise no
+   level is normally readable.
+
+   Additionally, the :meth:`_engine.Connection.get_isolation_level` method will
+   raise an exception if the ``v$transaction`` view is not available due to
+   permissions or other reasons, which is a common occurrence in Oracle Database
+   installations.
+
+   The python-oracledb and cx_Oracle dialects attempt to call the
+   :meth:`_engine.Connection.get_isolation_level` method when the dialect makes
+   its first connection to the database in order to acquire the
+   "default"isolation level.  This default level is necessary so that the level
+   can be reset on a connection after it has been temporarily modified using
+   :meth:`_engine.Connection.execution_options` method.  In the common event
+   that the :meth:`_engine.Connection.get_isolation_level` method raises an
+   exception due to ``v$transaction`` not being readable as well as any other
+   database-related failure, the level is assumed to be "READ COMMITTED".  No
+   warning is emitted for this initial first-connect condition as it is
+   expected to be a common restriction on Oracle databases.
+
+.. versionadded:: 1.3.16 added support for AUTOCOMMIT to the cx_Oracle dialect
+   as well as the notion of a default isolation level
+
+.. versionadded:: 1.3.21 Added support for SERIALIZABLE as well as live
+   reading of the isolation level.
+
+.. versionchanged:: 1.3.22 In the event that the default isolation
+   level cannot be read due to permissions on the v$transaction view as
+   is common in Oracle installations, the default isolation level is hardcoded
+   to "READ COMMITTED" which was the behavior prior to 1.3.21.
+
+.. seealso::
+
+    :ref:`dbapi_autocommit`
+
+Identifier Casing
+-----------------
+
+In Oracle Database, the data dictionary represents all case insensitive
+identifier names using UPPERCASE text.  This is in contradiction to the
+expectations of SQLAlchemy, which assume a case insensitive name is represented
+as lowercase text.
+
+As an example of case insensitive identifier names, consider the following table:
+
+.. sourcecode:: sql
+
+    CREATE TABLE MyTable (Identifier INTEGER PRIMARY KEY)
+
+If you were to ask Oracle Database for information about this table, the
+table name would be reported as ``MYTABLE`` and the column name would
+be reported as ``IDENTIFIER``.    Compare to most other databases such as
+PostgreSQL and MySQL which would report these names as ``mytable`` and
+``identifier``.   The names are **not quoted, therefore are case insensitive**.
+The special casing of ``MyTable`` and ``Identifier`` would only be maintained
+if they were quoted in the table definition:
+
+.. sourcecode:: sql
+
+    CREATE TABLE "MyTable" ("Identifier" INTEGER PRIMARY KEY)
+
+When constructing a SQLAlchemy :class:`.Table` object, **an all lowercase name
+is considered to be case insensitive**.   So the following table assumes
+case insensitive names::
+
+    Table("mytable", metadata, Column("identifier", Integer, primary_key=True))
+
+Whereas when mixed case or UPPERCASE names are used, case sensitivity is
+assumed::
+
+    Table("MyTable", metadata, Column("Identifier", Integer, primary_key=True))
+
+A similar situation occurs at the database driver level when emitting a
+textual SQL SELECT statement and looking at column names in the DBAPI
+``cursor.description`` attribute.  A database like PostgreSQL will normalize
+case insensitive names to be lowercase::
+
+    >>> pg_engine = create_engine("postgresql://scott:tiger@localhost/test")
+    >>> pg_connection = pg_engine.connect()
+    >>> result = pg_connection.exec_driver_sql("SELECT 1 AS SomeName")
+    >>> result.cursor.description
+    (Column(name='somename', type_code=23),)
+
+Whereas Oracle normalizes them to UPPERCASE::
+
+    >>> oracle_engine = create_engine("oracle+oracledb://scott:tiger@oracle18c/xe")
+    >>> oracle_connection = oracle_engine.connect()
+    >>> result = oracle_connection.exec_driver_sql(
+    ...     "SELECT 1 AS SomeName FROM DUAL"
+    ... )
+    >>> result.cursor.description
+    [('SOMENAME', <DbType DB_TYPE_NUMBER>, 127, None, 0, -127, True)]
+
+In order to achieve cross-database parity for the two cases of a. table
+reflection and b. textual-only SQL statement round trips, SQLAlchemy performs a step
+called **name normalization** when using the Oracle dialect.  This process may
+also apply to other third party dialects that have similar UPPERCASE handling
+of case insensitive names.
+
+When using name normalization, SQLAlchemy attempts to detect if a name is
+case insensitive by checking if all characters are UPPERCASE letters only;
+if so, then it assumes this is a case insensitive name and is delivered as
+a lowercase name.
+
+For table reflection, a tablename that is seen represented as all UPPERCASE
+in Oracle Database's catalog tables will be assumed to have a case insensitive
+name.  This is what allows the ``Table`` definition to use lower case names
+and be equally compatible from a reflection point of view on Oracle Database
+and all other databases such as PostgreSQL and MySQL::
+
+    # matches a table created with CREATE TABLE mytable
+    Table("mytable", metadata, autoload_with=some_engine)
+
+Above, the all lowercase name ``"mytable"`` is case insensitive; it will match
+a table reported by PostgreSQL as ``"mytable"`` and a table reported by
+Oracle as ``"MYTABLE"``.  If name normalization were not present, it would
+not be possible for the above :class:`.Table` definition to be introspectable
+in a cross-database way, since we are dealing with a case insensitive name
+that is not reported by each database in the same way.
+
+Case sensitivity can be forced on in this case, such as if we wanted to represent
+the quoted tablename ``"MYTABLE"`` with that exact casing, most simply by using
+that casing directly, which will be seen as a case sensitive name::
+
+    # matches a table created with CREATE TABLE "MYTABLE"
+    Table("MYTABLE", metadata, autoload_with=some_engine)
+
+For the unusual case of a quoted all-lowercase name, the :class:`.quoted_name`
+construct may be used::
+
+    from sqlalchemy import quoted_name
+
+    # matches a table created with CREATE TABLE "mytable"
+    Table(
+        quoted_name("mytable", quote=True), metadata, autoload_with=some_engine
+    )
+
+Name normalization also takes place when handling result sets from **purely
+textual SQL strings**, that have no other :class:`.Table` or :class:`.Column`
+metadata associated with them. This includes SQL strings executed using
+:meth:`.Connection.exec_driver_sql` and SQL strings executed using the
+:func:`.text` construct which do not include :class:`.Column` metadata.
+
+Returning to the Oracle Database SELECT statement, we see that even though
+``cursor.description`` reports the column name as ``SOMENAME``, SQLAlchemy
+name normalizes this to ``somename``::
+
+    >>> oracle_engine = create_engine("oracle+oracledb://scott:tiger@oracle18c/xe")
+    >>> oracle_connection = oracle_engine.connect()
+    >>> result = oracle_connection.exec_driver_sql(
+    ...     "SELECT 1 AS SomeName FROM DUAL"
+    ... )
+    >>> result.cursor.description
+    [('SOMENAME', <DbType DB_TYPE_NUMBER>, 127, None, 0, -127, True)]
+    >>> result.keys()
+    RMKeyView(['somename'])
+
+The single scenario where the above behavior produces inaccurate results
+is when using an all-uppercase, quoted name.  SQLAlchemy has no way to determine
+that a particular name in ``cursor.description`` was quoted, and is therefore
+case sensitive, or was not quoted, and should be name normalized::
+
+    >>> result = oracle_connection.exec_driver_sql(
+    ...     'SELECT 1 AS "SOMENAME" FROM DUAL'
+    ... )
+    >>> result.cursor.description
+    [('SOMENAME', <DbType DB_TYPE_NUMBER>, 127, None, 0, -127, True)]
+    >>> result.keys()
+    RMKeyView(['somename'])
+
+For this case, a new feature will be available in SQLAlchemy 2.1 to disable
+the name normalization behavior in specific cases.
+
+
+.. _oracle_max_identifier_lengths:
+
+Maximum Identifier Lengths
+--------------------------
+
+SQLAlchemy is sensitive to the maximum identifier length supported by Oracle
+Database. This affects generated SQL label names as well as the generation of
+constraint names, particularly in the case where the constraint naming
+convention feature described at :ref:`constraint_naming_conventions` is being
+used.
+
+Oracle Database 12.2 increased the default maximum identifier length from 30 to
+128. As of SQLAlchemy 1.4, the default maximum identifier length for the Oracle
+dialects is 128 characters.  Upon first connection, the maximum length actually
+supported by the database is obtained. In all cases, setting the
+:paramref:`_sa.create_engine.max_identifier_length` parameter will bypass this
+change and the value given will be used as is::
+
+    engine = create_engine(
+        "oracle+oracledb://scott:tiger@localhost:1521?service_name=freepdb1",
+        max_identifier_length=30,
+    )
+
+If :paramref:`_sa.create_engine.max_identifier_length` is not set, the oracledb
+dialect internally uses the ``max_identifier_length`` attribute available on
+driver connections since python-oracledb version 2.5. When using an older
+driver version, or using the cx_Oracle dialect, SQLAlchemy will instead attempt
+to use the query ``SELECT value FROM v$parameter WHERE name = 'compatible'``
+upon first connect in order to determine the effective compatibility version of
+the database. The "compatibility" version is a version number that is
+independent of the actual database version. It is used to assist database
+migration. It is configured by an Oracle Database initialization parameter. The
+compatibility version then determines the maximum allowed identifier length for
+the database. If the V$ view is not available, the database version information
+is used instead.
+
+The maximum identifier length comes into play both when generating anonymized
+SQL labels in SELECT statements, but more crucially when generating constraint
+names from a naming convention.  It is this area that has created the need for
+SQLAlchemy to change this default conservatively.  For example, the following
+naming convention produces two very different constraint names based on the
+identifier length::
+
+    from sqlalchemy import Column
+    from sqlalchemy import Index
+    from sqlalchemy import Integer
+    from sqlalchemy import MetaData
+    from sqlalchemy import Table
+    from sqlalchemy.dialects import oracle
+    from sqlalchemy.schema import CreateIndex
+
+    m = MetaData(naming_convention={"ix": "ix_%(column_0N_name)s"})
+
+    t = Table(
+        "t",
+        m,
+        Column("some_column_name_1", Integer),
+        Column("some_column_name_2", Integer),
+        Column("some_column_name_3", Integer),
+    )
+
+    ix = Index(
+        None,
+        t.c.some_column_name_1,
+        t.c.some_column_name_2,
+        t.c.some_column_name_3,
+    )
+
+    oracle_dialect = oracle.dialect(max_identifier_length=30)
+    print(CreateIndex(ix).compile(dialect=oracle_dialect))
+
+With an identifier length of 30, the above CREATE INDEX looks like:
+
+.. sourcecode:: sql
+
+    CREATE INDEX ix_some_column_name_1s_70cd ON t
+    (some_column_name_1, some_column_name_2, some_column_name_3)
+
+However with length of 128, it becomes::
+
+.. sourcecode:: sql
+
+    CREATE INDEX ix_some_column_name_1some_column_name_2some_column_name_3 ON t
+    (some_column_name_1, some_column_name_2, some_column_name_3)
+
+Applications which have run versions of SQLAlchemy prior to 1.4 on Oracle
+Database version 12.2 or greater are therefore subject to the scenario of a
+database migration that wishes to "DROP CONSTRAINT" on a name that was
+previously generated with the shorter length.  This migration will fail when
+the identifier length is changed without the name of the index or constraint
+first being adjusted.  Such applications are strongly advised to make use of
+:paramref:`_sa.create_engine.max_identifier_length` in order to maintain
+control of the generation of truncated names, and to fully review and test all
+database migrations in a staging environment when changing this value to ensure
+that the impact of this change has been mitigated.
+
+.. versionchanged:: 1.4 the default max_identifier_length for Oracle Database
+   is 128 characters, which is adjusted down to 30 upon first connect if the
+   Oracle Database, or its compatibility setting, are lower than version 12.2.
+
+
+LIMIT/OFFSET/FETCH Support
+--------------------------
+
+Methods like :meth:`_sql.Select.limit` and :meth:`_sql.Select.offset` make use
+of ``FETCH FIRST N ROW / OFFSET N ROWS`` syntax assuming Oracle Database 12c or
+above, and assuming the SELECT statement is not embedded within a compound
+statement like UNION.  This syntax is also available directly by using the
+:meth:`_sql.Select.fetch` method.
+
+.. versionchanged:: 2.0 the Oracle Database dialects now use ``FETCH FIRST N
+   ROW / OFFSET N ROWS`` for all :meth:`_sql.Select.limit` and
+   :meth:`_sql.Select.offset` usage including within the ORM and legacy
+   :class:`_orm.Query`.  To force the legacy behavior using window functions,
+   specify the ``enable_offset_fetch=False`` dialect parameter to
+   :func:`_sa.create_engine`.
+
+The use of ``FETCH FIRST / OFFSET`` may be disabled on any Oracle Database
+version by passing ``enable_offset_fetch=False`` to :func:`_sa.create_engine`,
+which will force the use of "legacy" mode that makes use of window functions.
+This mode is also selected automatically when using a version of Oracle
+Database prior to 12c.
+
+When using legacy mode, or when a :class:`.Select` statement with limit/offset
+is embedded in a compound statement, an emulated approach for LIMIT / OFFSET
+based on window functions is used, which involves creation of a subquery using
+``ROW_NUMBER`` that is prone to performance issues as well as SQL construction
+issues for complex statements. However, this approach is supported by all
+Oracle Database versions. See notes below.
+
+Notes on LIMIT / OFFSET emulation (when fetch() method cannot be used)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+If using :meth:`_sql.Select.limit` and :meth:`_sql.Select.offset`, or with the
+ORM the :meth:`_orm.Query.limit` and :meth:`_orm.Query.offset` methods on an
+Oracle Database version prior to 12c, the following notes apply:
+
+* SQLAlchemy currently makes use of ROWNUM to achieve
+  LIMIT/OFFSET; the exact methodology is taken from
+  https://blogs.oracle.com/oraclemagazine/on-rownum-and-limiting-results .
+
+* the "FIRST_ROWS()" optimization keyword is not used by default.  To enable
+  the usage of this optimization directive, specify ``optimize_limits=True``
+  to :func:`_sa.create_engine`.
+
+  .. versionchanged:: 1.4
+
+      The Oracle Database dialect renders limit/offset integer values using a
+      "post compile" scheme which renders the integer directly before passing
+      the statement to the cursor for execution.  The ``use_binds_for_limits``
+      flag no longer has an effect.
+
+      .. seealso::
+
+          :ref:`change_4808`.
+
+.. _oracle_returning:
+
+RETURNING Support
+-----------------
+
+Oracle Database supports RETURNING fully for INSERT, UPDATE and DELETE
+statements that are invoked with a single collection of bound parameters (that
+is, a ``cursor.execute()`` style statement; SQLAlchemy does not generally
+support RETURNING with :term:`executemany` statements).  Multiple rows may be
+returned as well.
+
+.. versionchanged:: 2.0 the Oracle Database backend has full support for
+   RETURNING on parity with other backends.
+
+
+ON UPDATE CASCADE
+-----------------
+
+Oracle Database doesn't have native ON UPDATE CASCADE functionality.  A trigger
+based solution is available at
+https://web.archive.org/web/20090317041251/https://asktom.oracle.com/tkyte/update_cascade/index.html
+
+When using the SQLAlchemy ORM, the ORM has limited ability to manually issue
+cascading updates - specify ForeignKey objects using the
+"deferrable=True, initially='deferred'" keyword arguments,
+and specify "passive_updates=False" on each relationship().
+
+Oracle Database 8 Compatibility
+-------------------------------
+
+.. warning:: The status of Oracle Database 8 compatibility is not known for
+   SQLAlchemy 2.0.
+
+When Oracle Database 8 is detected, the dialect internally configures itself to
+the following behaviors:
+
+* the use_ansi flag is set to False.  This has the effect of converting all
+  JOIN phrases into the WHERE clause, and in the case of LEFT OUTER JOIN
+  makes use of Oracle's (+) operator.
+
+* the NVARCHAR2 and NCLOB datatypes are no longer generated as DDL when
+  the :class:`~sqlalchemy.types.Unicode` is used - VARCHAR2 and CLOB are issued
+  instead. This because these types don't seem to work correctly on Oracle 8
+  even though they are available. The :class:`~sqlalchemy.types.NVARCHAR` and
+  :class:`~sqlalchemy.dialects.oracle.NCLOB` types will always generate
+  NVARCHAR2 and NCLOB.
+
+
+Synonym/DBLINK Reflection
+-------------------------
+
+When using reflection with Table objects, the dialect can optionally search
+for tables indicated by synonyms, either in local or remote schemas or
+accessed over DBLINK, by passing the flag ``oracle_resolve_synonyms=True`` as
+a keyword argument to the :class:`_schema.Table` construct::
+
+    some_table = Table(
+        "some_table", autoload_with=some_engine, oracle_resolve_synonyms=True
+    )
+
+When this flag is set, the given name (such as ``some_table`` above) will be
+searched not just in the ``ALL_TABLES`` view, but also within the
+``ALL_SYNONYMS`` view to see if this name is actually a synonym to another
+name.  If the synonym is located and refers to a DBLINK, the Oracle Database
+dialects know how to locate the table's information using DBLINK syntax(e.g.
+``@dblink``).
+
+``oracle_resolve_synonyms`` is accepted wherever reflection arguments are
+accepted, including methods such as :meth:`_schema.MetaData.reflect` and
+:meth:`_reflection.Inspector.get_columns`.
+
+If synonyms are not in use, this flag should be left disabled.
+
+.. _oracle_constraint_reflection:
+
+Constraint Reflection
+---------------------
+
+The Oracle Database dialects can return information about foreign key, unique,
+and CHECK constraints, as well as indexes on tables.
+
+Raw information regarding these constraints can be acquired using
+:meth:`_reflection.Inspector.get_foreign_keys`,
+:meth:`_reflection.Inspector.get_unique_constraints`,
+:meth:`_reflection.Inspector.get_check_constraints`, and
+:meth:`_reflection.Inspector.get_indexes`.
+
+.. versionchanged:: 1.2 The Oracle Database dialect can now reflect UNIQUE and
+   CHECK constraints.
+
+When using reflection at the :class:`_schema.Table` level, the
+:class:`_schema.Table`
+will also include these constraints.
+
+Note the following caveats:
+
+* When using the :meth:`_reflection.Inspector.get_check_constraints` method,
+  Oracle Database builds a special "IS NOT NULL" constraint for columns that
+  specify "NOT NULL".  This constraint is **not** returned by default; to
+  include the "IS NOT NULL" constraints, pass the flag ``include_all=True``::
+
+      from sqlalchemy import create_engine, inspect
+
+      engine = create_engine(
+          "oracle+oracledb://scott:tiger@localhost:1521?service_name=freepdb1"
+      )
+      inspector = inspect(engine)
+      all_check_constraints = inspector.get_check_constraints(
+          "some_table", include_all=True
+      )
+
+* in most cases, when reflecting a :class:`_schema.Table`, a UNIQUE constraint
+  will **not** be available as a :class:`.UniqueConstraint` object, as Oracle
+  Database mirrors unique constraints with a UNIQUE index in most cases (the
+  exception seems to be when two or more unique constraints represent the same
+  columns); the :class:`_schema.Table` will instead represent these using
+  :class:`.Index` with the ``unique=True`` flag set.
+
+* Oracle Database creates an implicit index for the primary key of a table;
+  this index is **excluded** from all index results.
+
+* the list of columns reflected for an index will not include column names
+  that start with SYS_NC.
+
+Table names with SYSTEM/SYSAUX tablespaces
+-------------------------------------------
+
+The :meth:`_reflection.Inspector.get_table_names` and
+:meth:`_reflection.Inspector.get_temp_table_names`
+methods each return a list of table names for the current engine. These methods
+are also part of the reflection which occurs within an operation such as
+:meth:`_schema.MetaData.reflect`.  By default,
+these operations exclude the ``SYSTEM``
+and ``SYSAUX`` tablespaces from the operation.   In order to change this, the
+default list of tablespaces excluded can be changed at the engine level using
+the ``exclude_tablespaces`` parameter::
+
+    # exclude SYSAUX and SOME_TABLESPACE, but not SYSTEM
+    e = create_engine(
+        "oracle+oracledb://scott:tiger@localhost:1521/?service_name=freepdb1",
+        exclude_tablespaces=["SYSAUX", "SOME_TABLESPACE"],
+    )
+
+.. _oracle_float_support:
+
+FLOAT / DOUBLE Support and Behaviors
+------------------------------------
+
+The SQLAlchemy :class:`.Float` and :class:`.Double` datatypes are generic
+datatypes that resolve to the "least surprising" datatype for a given backend.
+For Oracle Database, this means they resolve to the ``FLOAT`` and ``DOUBLE``
+types::
+
+    >>> from sqlalchemy import cast, literal, Float
+    >>> from sqlalchemy.dialects import oracle
+    >>> float_datatype = Float()
+    >>> print(cast(literal(5.0), float_datatype).compile(dialect=oracle.dialect()))
+    CAST(:param_1 AS FLOAT)
+
+Oracle's ``FLOAT`` / ``DOUBLE`` datatypes are aliases for ``NUMBER``.   Oracle
+Database stores ``NUMBER`` values with full precision, not floating point
+precision, which means that ``FLOAT`` / ``DOUBLE`` do not actually behave like
+native FP values. Oracle Database instead offers special datatypes
+``BINARY_FLOAT`` and ``BINARY_DOUBLE`` to deliver real 4- and 8- byte FP
+values.
+
+SQLAlchemy supports these datatypes directly using :class:`.BINARY_FLOAT` and
+:class:`.BINARY_DOUBLE`.   To use the :class:`.Float` or :class:`.Double`
+datatypes in a database agnostic way, while allowing Oracle backends to utilize
+one of these types, use the :meth:`.TypeEngine.with_variant` method to set up a
+variant::
+
+    >>> from sqlalchemy import cast, literal, Float
+    >>> from sqlalchemy.dialects import oracle
+    >>> float_datatype = Float().with_variant(oracle.BINARY_FLOAT(), "oracle")
+    >>> print(cast(literal(5.0), float_datatype).compile(dialect=oracle.dialect()))
+    CAST(:param_1 AS BINARY_FLOAT)
+
+E.g. to use this datatype in a :class:`.Table` definition::
+
+    my_table = Table(
+        "my_table",
+        metadata,
+        Column(
+            "fp_data", Float().with_variant(oracle.BINARY_FLOAT(), "oracle")
+        ),
+    )
+
+DateTime Compatibility
+----------------------
+
+Oracle Database has no datatype known as ``DATETIME``, it instead has only
+``DATE``, which can actually store a date and time value.  For this reason, the
+Oracle Database dialects provide a type :class:`_oracle.DATE` which is a
+subclass of :class:`.DateTime`.  This type has no special behavior, and is only
+present as a "marker" for this type; additionally, when a database column is
+reflected and the type is reported as ``DATE``, the time-supporting
+:class:`_oracle.DATE` type is used.
+
+.. _oracle_table_options:
+
+Oracle Database Table Options
+-----------------------------
+
+The CREATE TABLE phrase supports the following options with Oracle Database
+dialects in conjunction with the :class:`_schema.Table` construct:
+
+
+* ``ON COMMIT``::
+
+    Table(
+        "some_table",
+        metadata,
+        ...,
+        prefixes=["GLOBAL TEMPORARY"],
+        oracle_on_commit="PRESERVE ROWS",
+    )
+
+*
+  ``COMPRESS``::
+
+     Table(
+         "mytable", metadata, Column("data", String(32)), oracle_compress=True
+     )
+
+     Table("mytable", metadata, Column("data", String(32)), oracle_compress=6)
+
+  The ``oracle_compress`` parameter accepts either an integer compression
+  level, or ``True`` to use the default compression level.
+
+*
+  ``TABLESPACE``::
+
+     Table("mytable", metadata, ..., oracle_tablespace="EXAMPLE_TABLESPACE")
+
+  The ``oracle_tablespace`` parameter specifies the tablespace in which the
+  table is to be created. This is useful when you want to create a table in a
+  tablespace other than the default tablespace of the user.
+
+  .. versionadded:: 2.0.37
+
+.. _oracle_index_options:
+
+Oracle Database Specific Index Options
+--------------------------------------
+
+Bitmap Indexes
+~~~~~~~~~~~~~~
+
+You can specify the ``oracle_bitmap`` parameter to create a bitmap index
+instead of a B-tree index::
+
+    Index("my_index", my_table.c.data, oracle_bitmap=True)
+
+Bitmap indexes cannot be unique and cannot be compressed. SQLAlchemy will not
+check for such limitations, only the database will.
+
+Index compression
+~~~~~~~~~~~~~~~~~
+
+Oracle Database has a more efficient storage mode for indexes containing lots
+of repeated values. Use the ``oracle_compress`` parameter to turn on key
+compression::
+
+    Index("my_index", my_table.c.data, oracle_compress=True)
+
+    Index(
+        "my_index",
+        my_table.c.data1,
+        my_table.c.data2,
+        unique=True,
+        oracle_compress=1,
+    )
+
+The ``oracle_compress`` parameter accepts either an integer specifying the
+number of prefix columns to compress, or ``True`` to use the default (all
+columns for non-unique indexes, all but the last column for unique indexes).
+
+"""  # noqa
+
+from __future__ import annotations
+
+from collections import defaultdict
+from functools import lru_cache
+from functools import wraps
+import re
+
+from . import dictionary
+from .types import _OracleBoolean
+from .types import _OracleDate
+from .types import BFILE
+from .types import BINARY_DOUBLE
+from .types import BINARY_FLOAT
+from .types import DATE
+from .types import FLOAT
+from .types import INTERVAL
+from .types import LONG
+from .types import NCLOB
+from .types import NUMBER
+from .types import NVARCHAR2  # noqa
+from .types import OracleRaw  # noqa
+from .types import RAW
+from .types import ROWID  # noqa
+from .types import TIMESTAMP
+from .types import VARCHAR2  # noqa
+from ... import Computed
+from ... import exc
+from ... import schema as sa_schema
+from ... import sql
+from ... import util
+from ...engine import default
+from ...engine import ObjectKind
+from ...engine import ObjectScope
+from ...engine import reflection
+from ...engine.reflection import ReflectionDefaults
+from ...sql import and_
+from ...sql import bindparam
+from ...sql import compiler
+from ...sql import expression
+from ...sql import func
+from ...sql import null
+from ...sql import or_
+from ...sql import select
+from ...sql import sqltypes
+from ...sql import util as sql_util
+from ...sql import visitors
+from ...sql.visitors import InternalTraversal
+from ...types import BLOB
+from ...types import CHAR
+from ...types import CLOB
+from ...types import DOUBLE_PRECISION
+from ...types import INTEGER
+from ...types import NCHAR
+from ...types import NVARCHAR
+from ...types import REAL
+from ...types import VARCHAR
+
+RESERVED_WORDS = set(
+    "SHARE RAW DROP BETWEEN FROM DESC OPTION PRIOR LONG THEN "
+    "DEFAULT ALTER IS INTO MINUS INTEGER NUMBER GRANT IDENTIFIED "
+    "ALL TO ORDER ON FLOAT DATE HAVING CLUSTER NOWAIT RESOURCE "
+    "ANY TABLE INDEX FOR UPDATE WHERE CHECK SMALLINT WITH DELETE "
+    "BY ASC REVOKE LIKE SIZE RENAME NOCOMPRESS NULL GROUP VALUES "
+    "AS IN VIEW EXCLUSIVE COMPRESS SYNONYM SELECT INSERT EXISTS "
+    "NOT TRIGGER ELSE CREATE INTERSECT PCTFREE DISTINCT USER "
+    "CONNECT SET MODE OF UNIQUE VARCHAR2 VARCHAR LOCK OR CHAR "
+    "DECIMAL UNION PUBLIC AND START UID COMMENT CURRENT LEVEL".split()
+)
+
+NO_ARG_FNS = set(
+    "UID CURRENT_DATE SYSDATE USER CURRENT_TIME CURRENT_TIMESTAMP".split()
+)
+
+
+colspecs = {
+    sqltypes.Boolean: _OracleBoolean,
+    sqltypes.Interval: INTERVAL,
+    sqltypes.DateTime: DATE,
+    sqltypes.Date: _OracleDate,
+}
+
+ischema_names = {
+    "VARCHAR2": VARCHAR,
+    "NVARCHAR2": NVARCHAR,
+    "CHAR": CHAR,
+    "NCHAR": NCHAR,
+    "DATE": DATE,
+    "NUMBER": NUMBER,
+    "BLOB": BLOB,
+    "BFILE": BFILE,
+    "CLOB": CLOB,
+    "NCLOB": NCLOB,
+    "TIMESTAMP": TIMESTAMP,
+    "TIMESTAMP WITH TIME ZONE": TIMESTAMP,
+    "TIMESTAMP WITH LOCAL TIME ZONE": TIMESTAMP,
+    "INTERVAL DAY TO SECOND": INTERVAL,
+    "RAW": RAW,
+    "FLOAT": FLOAT,
+    "DOUBLE PRECISION": DOUBLE_PRECISION,
+    "REAL": REAL,
+    "LONG": LONG,
+    "BINARY_DOUBLE": BINARY_DOUBLE,
+    "BINARY_FLOAT": BINARY_FLOAT,
+    "ROWID": ROWID,
+}
+
+
+class OracleTypeCompiler(compiler.GenericTypeCompiler):
+    # Note:
+    # Oracle DATE == DATETIME
+    # Oracle does not allow milliseconds in DATE
+    # Oracle does not support TIME columns
+
+    def visit_datetime(self, type_, **kw):
+        return self.visit_DATE(type_, **kw)
+
+    def visit_float(self, type_, **kw):
+        return self.visit_FLOAT(type_, **kw)
+
+    def visit_double(self, type_, **kw):
+        return self.visit_DOUBLE_PRECISION(type_, **kw)
+
+    def visit_unicode(self, type_, **kw):
+        if self.dialect._use_nchar_for_unicode:
+            return self.visit_NVARCHAR2(type_, **kw)
+        else:
+            return self.visit_VARCHAR2(type_, **kw)
+
+    def visit_INTERVAL(self, type_, **kw):
+        return "INTERVAL DAY%s TO SECOND%s" % (
+            type_.day_precision is not None
+            and "(%d)" % type_.day_precision
+            or "",
+            type_.second_precision is not None
+            and "(%d)" % type_.second_precision
+            or "",
+        )
+
+    def visit_LONG(self, type_, **kw):
+        return "LONG"
+
+    def visit_TIMESTAMP(self, type_, **kw):
+        if getattr(type_, "local_timezone", False):
+            return "TIMESTAMP WITH LOCAL TIME ZONE"
+        elif type_.timezone:
+            return "TIMESTAMP WITH TIME ZONE"
+        else:
+            return "TIMESTAMP"
+
+    def visit_DOUBLE_PRECISION(self, type_, **kw):
+        return self._generate_numeric(type_, "DOUBLE PRECISION", **kw)
+
+    def visit_BINARY_DOUBLE(self, type_, **kw):
+        return self._generate_numeric(type_, "BINARY_DOUBLE", **kw)
+
+    def visit_BINARY_FLOAT(self, type_, **kw):
+        return self._generate_numeric(type_, "BINARY_FLOAT", **kw)
+
+    def visit_FLOAT(self, type_, **kw):
+        kw["_requires_binary_precision"] = True
+        return self._generate_numeric(type_, "FLOAT", **kw)
+
+    def visit_NUMBER(self, type_, **kw):
+        return self._generate_numeric(type_, "NUMBER", **kw)
+
+    def _generate_numeric(
+        self,
+        type_,
+        name,
+        precision=None,
+        scale=None,
+        _requires_binary_precision=False,
+        **kw,
+    ):
+        if precision is None:
+            precision = getattr(type_, "precision", None)
+
+        if _requires_binary_precision:
+            binary_precision = getattr(type_, "binary_precision", None)
+
+            if precision and binary_precision is None:
+                # https://www.oracletutorial.com/oracle-basics/oracle-float/
+                estimated_binary_precision = int(precision / 0.30103)
+                raise exc.ArgumentError(
+                    "Oracle Database FLOAT types use 'binary precision', "
+                    "which does not convert cleanly from decimal "
+                    "'precision'.  Please specify "
+                    "this type with a separate Oracle Database variant, such "
+                    f"as {type_.__class__.__name__}(precision={precision})."
+                    f"with_variant(oracle.FLOAT"
+                    f"(binary_precision="
+                    f"{estimated_binary_precision}), 'oracle'), so that the "
+                    "Oracle Database specific 'binary_precision' may be "
+                    "specified accurately."
+                )
+            else:
+                precision = binary_precision
+
+        if scale is None:
+            scale = getattr(type_, "scale", None)
+
+        if precision is None:
+            return name
+        elif scale is None:
+            n = "%(name)s(%(precision)s)"
+            return n % {"name": name, "precision": precision}
+        else:
+            n = "%(name)s(%(precision)s, %(scale)s)"
+            return n % {"name": name, "precision": precision, "scale": scale}
+
+    def visit_string(self, type_, **kw):
+        return self.visit_VARCHAR2(type_, **kw)
+
+    def visit_VARCHAR2(self, type_, **kw):
+        return self._visit_varchar(type_, "", "2")
+
+    def visit_NVARCHAR2(self, type_, **kw):
+        return self._visit_varchar(type_, "N", "2")
+
+    visit_NVARCHAR = visit_NVARCHAR2
+
+    def visit_VARCHAR(self, type_, **kw):
+        return self._visit_varchar(type_, "", "")
+
+    def _visit_varchar(self, type_, n, num):
+        if not type_.length:
+            return "%(n)sVARCHAR%(two)s" % {"two": num, "n": n}
+        elif not n and self.dialect._supports_char_length:
+            varchar = "VARCHAR%(two)s(%(length)s CHAR)"
+            return varchar % {"length": type_.length, "two": num}
+        else:
+            varchar = "%(n)sVARCHAR%(two)s(%(length)s)"
+            return varchar % {"length": type_.length, "two": num, "n": n}
+
+    def visit_text(self, type_, **kw):
+        return self.visit_CLOB(type_, **kw)
+
+    def visit_unicode_text(self, type_, **kw):
+        if self.dialect._use_nchar_for_unicode:
+            return self.visit_NCLOB(type_, **kw)
+        else:
+            return self.visit_CLOB(type_, **kw)
+
+    def visit_large_binary(self, type_, **kw):
+        return self.visit_BLOB(type_, **kw)
+
+    def visit_big_integer(self, type_, **kw):
+        return self.visit_NUMBER(type_, precision=19, **kw)
+
+    def visit_boolean(self, type_, **kw):
+        return self.visit_SMALLINT(type_, **kw)
+
+    def visit_RAW(self, type_, **kw):
+        if type_.length:
+            return "RAW(%(length)s)" % {"length": type_.length}
+        else:
+            return "RAW"
+
+    def visit_ROWID(self, type_, **kw):
+        return "ROWID"
+
+
+class OracleCompiler(compiler.SQLCompiler):
+    """Oracle compiler modifies the lexical structure of Select
+    statements to work under non-ANSI configured Oracle databases, if
+    the use_ansi flag is False.
+    """
+
+    compound_keywords = util.update_copy(
+        compiler.SQLCompiler.compound_keywords,
+        {expression.CompoundSelect.EXCEPT: "MINUS"},
+    )
+
+    def __init__(self, *args, **kwargs):
+        self.__wheres = {}
+        super().__init__(*args, **kwargs)
+
+    def visit_mod_binary(self, binary, operator, **kw):
+        return "mod(%s, %s)" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_now_func(self, fn, **kw):
+        return "CURRENT_TIMESTAMP"
+
+    def visit_char_length_func(self, fn, **kw):
+        return "LENGTH" + self.function_argspec(fn, **kw)
+
+    def visit_match_op_binary(self, binary, operator, **kw):
+        return "CONTAINS (%s, %s)" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_true(self, expr, **kw):
+        return "1"
+
+    def visit_false(self, expr, **kw):
+        return "0"
+
+    def get_cte_preamble(self, recursive):
+        return "WITH"
+
+    def get_select_hint_text(self, byfroms):
+        return " ".join("/*+ %s */" % text for table, text in byfroms.items())
+
+    def function_argspec(self, fn, **kw):
+        if len(fn.clauses) > 0 or fn.name.upper() not in NO_ARG_FNS:
+            return compiler.SQLCompiler.function_argspec(self, fn, **kw)
+        else:
+            return ""
+
+    def visit_function(self, func, **kw):
+        text = super().visit_function(func, **kw)
+        if kw.get("asfrom", False) and func.name.lower() != "table":
+            text = "TABLE (%s)" % text
+        return text
+
+    def visit_table_valued_column(self, element, **kw):
+        text = super().visit_table_valued_column(element, **kw)
+        text = text + ".COLUMN_VALUE"
+        return text
+
+    def default_from(self):
+        """Called when a ``SELECT`` statement has no froms,
+        and no ``FROM`` clause is to be appended.
+
+        The Oracle compiler tacks a "FROM DUAL" to the statement.
+        """
+
+        return " FROM DUAL"
+
+    def visit_join(self, join, from_linter=None, **kwargs):
+        if self.dialect.use_ansi:
+            return compiler.SQLCompiler.visit_join(
+                self, join, from_linter=from_linter, **kwargs
+            )
+        else:
+            if from_linter:
+                from_linter.edges.add((join.left, join.right))
+
+            kwargs["asfrom"] = True
+            if isinstance(join.right, expression.FromGrouping):
+                right = join.right.element
+            else:
+                right = join.right
+            return (
+                self.process(join.left, from_linter=from_linter, **kwargs)
+                + ", "
+                + self.process(right, from_linter=from_linter, **kwargs)
+            )
+
+    def _get_nonansi_join_whereclause(self, froms):
+        clauses = []
+
+        def visit_join(join):
+            if join.isouter:
+                # https://docs.oracle.com/database/121/SQLRF/queries006.htm#SQLRF52354
+                # "apply the outer join operator (+) to all columns of B in
+                # the join condition in the WHERE clause" - that is,
+                # unconditionally regardless of operator or the other side
+                def visit_binary(binary):
+                    if isinstance(
+                        binary.left, expression.ColumnClause
+                    ) and join.right.is_derived_from(binary.left.table):
+                        binary.left = _OuterJoinColumn(binary.left)
+                    elif isinstance(
+                        binary.right, expression.ColumnClause
+                    ) and join.right.is_derived_from(binary.right.table):
+                        binary.right = _OuterJoinColumn(binary.right)
+
+                clauses.append(
+                    visitors.cloned_traverse(
+                        join.onclause, {}, {"binary": visit_binary}
+                    )
+                )
+            else:
+                clauses.append(join.onclause)
+
+            for j in join.left, join.right:
+                if isinstance(j, expression.Join):
+                    visit_join(j)
+                elif isinstance(j, expression.FromGrouping):
+                    visit_join(j.element)
+
+        for f in froms:
+            if isinstance(f, expression.Join):
+                visit_join(f)
+
+        if not clauses:
+            return None
+        else:
+            return sql.and_(*clauses)
+
+    def visit_outer_join_column(self, vc, **kw):
+        return self.process(vc.column, **kw) + "(+)"
+
+    def visit_sequence(self, seq, **kw):
+        return self.preparer.format_sequence(seq) + ".nextval"
+
+    def get_render_as_alias_suffix(self, alias_name_text):
+        """Oracle doesn't like ``FROM table AS alias``"""
+
+        return " " + alias_name_text
+
+    def returning_clause(
+        self, stmt, returning_cols, *, populate_result_map, **kw
+    ):
+        columns = []
+        binds = []
+
+        for i, column in enumerate(
+            expression._select_iterables(returning_cols)
+        ):
+            if (
+                self.isupdate
+                and isinstance(column, sa_schema.Column)
+                and isinstance(column.server_default, Computed)
+                and not self.dialect._supports_update_returning_computed_cols
+            ):
+                util.warn(
+                    "Computed columns don't work with Oracle Database UPDATE "
+                    "statements that use RETURNING; the value of the column "
+                    "*before* the UPDATE takes place is returned.   It is "
+                    "advised to not use RETURNING with an Oracle Database "
+                    "computed column.  Consider setting implicit_returning "
+                    "to False on the Table object in order to avoid implicit "
+                    "RETURNING clauses from being generated for this Table."
+                )
+            if column.type._has_column_expression:
+                col_expr = column.type.column_expression(column)
+            else:
+                col_expr = column
+
+            outparam = sql.outparam("ret_%d" % i, type_=column.type)
+            self.binds[outparam.key] = outparam
+            binds.append(
+                self.bindparam_string(self._truncate_bindparam(outparam))
+            )
+
+            # has_out_parameters would in a normal case be set to True
+            # as a result of the compiler visiting an outparam() object.
+            # in this case, the above outparam() objects are not being
+            # visited.   Ensure the statement itself didn't have other
+            # outparam() objects independently.
+            # technically, this could be supported, but as it would be
+            # a very strange use case without a clear rationale, disallow it
+            if self.has_out_parameters:
+                raise exc.InvalidRequestError(
+                    "Using explicit outparam() objects with "
+                    "UpdateBase.returning() in the same Core DML statement "
+                    "is not supported in the Oracle Database dialects."
+                )
+
+            self._oracle_returning = True
+
+            columns.append(self.process(col_expr, within_columns_clause=False))
+            if populate_result_map:
+                self._add_to_result_map(
+                    getattr(col_expr, "name", col_expr._anon_name_label),
+                    getattr(col_expr, "name", col_expr._anon_name_label),
+                    (
+                        column,
+                        getattr(column, "name", None),
+                        getattr(column, "key", None),
+                    ),
+                    column.type,
+                )
+
+        return "RETURNING " + ", ".join(columns) + " INTO " + ", ".join(binds)
+
+    def _row_limit_clause(self, select, **kw):
+        """Oracle Database 12c supports OFFSET/FETCH operators
+        Use it instead subquery with row_number
+
+        """
+
+        if (
+            select._fetch_clause is not None
+            or not self.dialect._supports_offset_fetch
+        ):
+            return super()._row_limit_clause(
+                select, use_literal_execute_for_simple_int=True, **kw
+            )
+        else:
+            return self.fetch_clause(
+                select,
+                fetch_clause=self._get_limit_or_fetch(select),
+                use_literal_execute_for_simple_int=True,
+                **kw,
+            )
+
+    def _get_limit_or_fetch(self, select):
+        if select._fetch_clause is None:
+            return select._limit_clause
+        else:
+            return select._fetch_clause
+
+    def translate_select_structure(self, select_stmt, **kwargs):
+        select = select_stmt
+
+        if not getattr(select, "_oracle_visit", None):
+            if not self.dialect.use_ansi:
+                froms = self._display_froms_for_select(
+                    select, kwargs.get("asfrom", False)
+                )
+                whereclause = self._get_nonansi_join_whereclause(froms)
+                if whereclause is not None:
+                    select = select.where(whereclause)
+                    select._oracle_visit = True
+
+            # if fetch is used this is not needed
+            if (
+                select._has_row_limiting_clause
+                and not self.dialect._supports_offset_fetch
+                and select._fetch_clause is None
+            ):
+                limit_clause = select._limit_clause
+                offset_clause = select._offset_clause
+
+                if select._simple_int_clause(limit_clause):
+                    limit_clause = limit_clause.render_literal_execute()
+
+                if select._simple_int_clause(offset_clause):
+                    offset_clause = offset_clause.render_literal_execute()
+
+                # currently using form at:
+                # https://blogs.oracle.com/oraclemagazine/\
+                # on-rownum-and-limiting-results
+
+                orig_select = select
+                select = select._generate()
+                select._oracle_visit = True
+
+                # add expressions to accommodate FOR UPDATE OF
+                for_update = select._for_update_arg
+                if for_update is not None and for_update.of:
+                    for_update = for_update._clone()
+                    for_update._copy_internals()
+
+                    for elem in for_update.of:
+                        if not select.selected_columns.contains_column(elem):
+                            select = select.add_columns(elem)
+
+                # Wrap the middle select and add the hint
+                inner_subquery = select.alias()
+                limitselect = sql.select(
+                    *[
+                        c
+                        for c in inner_subquery.c
+                        if orig_select.selected_columns.corresponding_column(c)
+                        is not None
+                    ]
+                )
+
+                if (
+                    limit_clause is not None
+                    and self.dialect.optimize_limits
+                    and select._simple_int_clause(limit_clause)
+                ):
+                    limitselect = limitselect.prefix_with(
+                        expression.text(
+                            "/*+ FIRST_ROWS(%s) */"
+                            % self.process(limit_clause, **kwargs)
+                        )
+                    )
+
+                limitselect._oracle_visit = True
+                limitselect._is_wrapper = True
+
+                # add expressions to accommodate FOR UPDATE OF
+                if for_update is not None and for_update.of:
+                    adapter = sql_util.ClauseAdapter(inner_subquery)
+                    for_update.of = [
+                        adapter.traverse(elem) for elem in for_update.of
+                    ]
+
+                # If needed, add the limiting clause
+                if limit_clause is not None:
+                    if select._simple_int_clause(limit_clause) and (
+                        offset_clause is None
+                        or select._simple_int_clause(offset_clause)
+                    ):
+                        max_row = limit_clause
+
+                        if offset_clause is not None:
+                            max_row = max_row + offset_clause
+
+                    else:
+                        max_row = limit_clause
+
+                        if offset_clause is not None:
+                            max_row = max_row + offset_clause
+                    limitselect = limitselect.where(
+                        sql.literal_column("ROWNUM") <= max_row
+                    )
+
+                # If needed, add the ora_rn, and wrap again with offset.
+                if offset_clause is None:
+                    limitselect._for_update_arg = for_update
+                    select = limitselect
+                else:
+                    limitselect = limitselect.add_columns(
+                        sql.literal_column("ROWNUM").label("ora_rn")
+                    )
+                    limitselect._oracle_visit = True
+                    limitselect._is_wrapper = True
+
+                    if for_update is not None and for_update.of:
+                        limitselect_cols = limitselect.selected_columns
+                        for elem in for_update.of:
+                            if (
+                                limitselect_cols.corresponding_column(elem)
+                                is None
+                            ):
+                                limitselect = limitselect.add_columns(elem)
+
+                    limit_subquery = limitselect.alias()
+                    origselect_cols = orig_select.selected_columns
+                    offsetselect = sql.select(
+                        *[
+                            c
+                            for c in limit_subquery.c
+                            if origselect_cols.corresponding_column(c)
+                            is not None
+                        ]
+                    )
+
+                    offsetselect._oracle_visit = True
+                    offsetselect._is_wrapper = True
+
+                    if for_update is not None and for_update.of:
+                        adapter = sql_util.ClauseAdapter(limit_subquery)
+                        for_update.of = [
+                            adapter.traverse(elem) for elem in for_update.of
+                        ]
+
+                    offsetselect = offsetselect.where(
+                        sql.literal_column("ora_rn") > offset_clause
+                    )
+
+                    offsetselect._for_update_arg = for_update
+                    select = offsetselect
+
+        return select
+
+    def limit_clause(self, select, **kw):
+        return ""
+
+    def visit_empty_set_expr(self, type_, **kw):
+        return "SELECT 1 FROM DUAL WHERE 1!=1"
+
+    def for_update_clause(self, select, **kw):
+        if self.is_subquery():
+            return ""
+
+        tmp = " FOR UPDATE"
+
+        if select._for_update_arg.of:
+            tmp += " OF " + ", ".join(
+                self.process(elem, **kw) for elem in select._for_update_arg.of
+            )
+
+        if select._for_update_arg.nowait:
+            tmp += " NOWAIT"
+        if select._for_update_arg.skip_locked:
+            tmp += " SKIP LOCKED"
+
+        return tmp
+
+    def visit_is_distinct_from_binary(self, binary, operator, **kw):
+        return "DECODE(%s, %s, 0, 1) = 1" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_is_not_distinct_from_binary(self, binary, operator, **kw):
+        return "DECODE(%s, %s, 0, 1) = 0" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        string = self.process(binary.left, **kw)
+        pattern = self.process(binary.right, **kw)
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return "REGEXP_LIKE(%s, %s)" % (string, pattern)
+        else:
+            return "REGEXP_LIKE(%s, %s, %s)" % (
+                string,
+                pattern,
+                self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            )
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        return "NOT %s" % self.visit_regexp_match_op_binary(
+            binary, operator, **kw
+        )
+
+    def visit_regexp_replace_op_binary(self, binary, operator, **kw):
+        string = self.process(binary.left, **kw)
+        pattern_replace = self.process(binary.right, **kw)
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return "REGEXP_REPLACE(%s, %s)" % (
+                string,
+                pattern_replace,
+            )
+        else:
+            return "REGEXP_REPLACE(%s, %s, %s)" % (
+                string,
+                pattern_replace,
+                self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            )
+
+    def visit_aggregate_strings_func(self, fn, **kw):
+        return "LISTAGG%s" % self.function_argspec(fn, **kw)
+
+    def _visit_bitwise(self, binary, fn_name, custom_right=None, **kw):
+        left = self.process(binary.left, **kw)
+        right = self.process(
+            custom_right if custom_right is not None else binary.right, **kw
+        )
+        return f"{fn_name}({left}, {right})"
+
+    def visit_bitwise_xor_op_binary(self, binary, operator, **kw):
+        return self._visit_bitwise(binary, "BITXOR", **kw)
+
+    def visit_bitwise_or_op_binary(self, binary, operator, **kw):
+        return self._visit_bitwise(binary, "BITOR", **kw)
+
+    def visit_bitwise_and_op_binary(self, binary, operator, **kw):
+        return self._visit_bitwise(binary, "BITAND", **kw)
+
+    def visit_bitwise_rshift_op_binary(self, binary, operator, **kw):
+        raise exc.CompileError("Cannot compile bitwise_rshift in oracle")
+
+    def visit_bitwise_lshift_op_binary(self, binary, operator, **kw):
+        raise exc.CompileError("Cannot compile bitwise_lshift in oracle")
+
+    def visit_bitwise_not_op_unary_operator(self, element, operator, **kw):
+        raise exc.CompileError("Cannot compile bitwise_not in oracle")
+
+
+class OracleDDLCompiler(compiler.DDLCompiler):
+    def define_constraint_cascades(self, constraint):
+        text = ""
+        if constraint.ondelete is not None:
+            text += " ON DELETE %s" % constraint.ondelete
+
+        # oracle has no ON UPDATE CASCADE -
+        # its only available via triggers
+        # https://web.archive.org/web/20090317041251/https://asktom.oracle.com/tkyte/update_cascade/index.html
+        if constraint.onupdate is not None:
+            util.warn(
+                "Oracle Database does not contain native UPDATE CASCADE "
+                "functionality - onupdates will not be rendered for foreign "
+                "keys.  Consider using deferrable=True, initially='deferred' "
+                "or triggers."
+            )
+
+        return text
+
+    def visit_drop_table_comment(self, drop, **kw):
+        return "COMMENT ON TABLE %s IS ''" % self.preparer.format_table(
+            drop.element
+        )
+
+    def visit_create_index(self, create, **kw):
+        index = create.element
+        self._verify_index_table(index)
+        preparer = self.preparer
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+        if index.dialect_options["oracle"]["bitmap"]:
+            text += "BITMAP "
+        text += "INDEX %s ON %s (%s)" % (
+            self._prepared_index_name(index, include_schema=True),
+            preparer.format_table(index.table, use_schema=True),
+            ", ".join(
+                self.sql_compiler.process(
+                    expr, include_table=False, literal_binds=True
+                )
+                for expr in index.expressions
+            ),
+        )
+        if index.dialect_options["oracle"]["compress"] is not False:
+            if index.dialect_options["oracle"]["compress"] is True:
+                text += " COMPRESS"
+            else:
+                text += " COMPRESS %d" % (
+                    index.dialect_options["oracle"]["compress"]
+                )
+        return text
+
+    def post_create_table(self, table):
+        table_opts = []
+        opts = table.dialect_options["oracle"]
+
+        if opts["on_commit"]:
+            on_commit_options = opts["on_commit"].replace("_", " ").upper()
+            table_opts.append("\n ON COMMIT %s" % on_commit_options)
+
+        if opts["compress"]:
+            if opts["compress"] is True:
+                table_opts.append("\n COMPRESS")
+            else:
+                table_opts.append("\n COMPRESS FOR %s" % (opts["compress"]))
+        if opts["tablespace"]:
+            table_opts.append(
+                "\n TABLESPACE %s" % self.preparer.quote(opts["tablespace"])
+            )
+        return "".join(table_opts)
+
+    def get_identity_options(self, identity_options):
+        text = super().get_identity_options(identity_options)
+        text = text.replace("NO MINVALUE", "NOMINVALUE")
+        text = text.replace("NO MAXVALUE", "NOMAXVALUE")
+        text = text.replace("NO CYCLE", "NOCYCLE")
+        if identity_options.order is not None:
+            text += " ORDER" if identity_options.order else " NOORDER"
+        return text.strip()
+
+    def visit_computed_column(self, generated, **kw):
+        text = "GENERATED ALWAYS AS (%s)" % self.sql_compiler.process(
+            generated.sqltext, include_table=False, literal_binds=True
+        )
+        if generated.persisted is True:
+            raise exc.CompileError(
+                "Oracle Database computed columns do not support 'stored' "
+                "persistence; set the 'persisted' flag to None or False for "
+                "Oracle Database support."
+            )
+        elif generated.persisted is False:
+            text += " VIRTUAL"
+        return text
+
+    def visit_identity_column(self, identity, **kw):
+        if identity.always is None:
+            kind = ""
+        else:
+            kind = "ALWAYS" if identity.always else "BY DEFAULT"
+        text = "GENERATED %s" % kind
+        if identity.on_null:
+            text += " ON NULL"
+        text += " AS IDENTITY"
+        options = self.get_identity_options(identity)
+        if options:
+            text += " (%s)" % options
+        return text
+
+
+class OracleIdentifierPreparer(compiler.IdentifierPreparer):
+    reserved_words = {x.lower() for x in RESERVED_WORDS}
+    illegal_initial_characters = {str(dig) for dig in range(0, 10)}.union(
+        ["_", "$"]
+    )
+
+    def _bindparam_requires_quotes(self, value):
+        """Return True if the given identifier requires quoting."""
+        lc_value = value.lower()
+        return (
+            lc_value in self.reserved_words
+            or value[0] in self.illegal_initial_characters
+            or not self.legal_characters.match(str(value))
+        )
+
+    def format_savepoint(self, savepoint):
+        name = savepoint.ident.lstrip("_")
+        return super().format_savepoint(savepoint, name)
+
+
+class OracleExecutionContext(default.DefaultExecutionContext):
+    def fire_sequence(self, seq, type_):
+        return self._execute_scalar(
+            "SELECT "
+            + self.identifier_preparer.format_sequence(seq)
+            + ".nextval FROM DUAL",
+            type_,
+        )
+
+    def pre_exec(self):
+        if self.statement and "_oracle_dblink" in self.execution_options:
+            self.statement = self.statement.replace(
+                dictionary.DB_LINK_PLACEHOLDER,
+                self.execution_options["_oracle_dblink"],
+            )
+
+
+class OracleDialect(default.DefaultDialect):
+    name = "oracle"
+    supports_statement_cache = True
+    supports_alter = True
+    max_identifier_length = 128
+
+    _supports_offset_fetch = True
+
+    insert_returning = True
+    update_returning = True
+    delete_returning = True
+
+    div_is_floordiv = False
+
+    supports_simple_order_by_label = False
+    cte_follows_insert = True
+    returns_native_bytes = True
+
+    supports_sequences = True
+    sequences_optional = False
+    postfetch_lastrowid = False
+
+    default_paramstyle = "named"
+    colspecs = colspecs
+    ischema_names = ischema_names
+    requires_name_normalize = True
+
+    supports_comments = True
+
+    supports_default_values = False
+    supports_default_metavalue = True
+    supports_empty_insert = False
+    supports_identity_columns = True
+
+    statement_compiler = OracleCompiler
+    ddl_compiler = OracleDDLCompiler
+    type_compiler_cls = OracleTypeCompiler
+    preparer = OracleIdentifierPreparer
+    execution_ctx_cls = OracleExecutionContext
+
+    reflection_options = ("oracle_resolve_synonyms",)
+
+    _use_nchar_for_unicode = False
+
+    construct_arguments = [
+        (
+            sa_schema.Table,
+            {
+                "resolve_synonyms": False,
+                "on_commit": None,
+                "compress": False,
+                "tablespace": None,
+            },
+        ),
+        (sa_schema.Index, {"bitmap": False, "compress": False}),
+    ]
+
+    @util.deprecated_params(
+        use_binds_for_limits=(
+            "1.4",
+            "The ``use_binds_for_limits`` Oracle Database dialect parameter "
+            "is deprecated. The dialect now renders LIMIT / OFFSET integers "
+            "inline in all cases using a post-compilation hook, so that the "
+            "value is still represented by a 'bound parameter' on the Core "
+            "Expression side.",
+        )
+    )
+    def __init__(
+        self,
+        use_ansi=True,
+        optimize_limits=False,
+        use_binds_for_limits=None,
+        use_nchar_for_unicode=False,
+        exclude_tablespaces=("SYSTEM", "SYSAUX"),
+        enable_offset_fetch=True,
+        **kwargs,
+    ):
+        default.DefaultDialect.__init__(self, **kwargs)
+        self._use_nchar_for_unicode = use_nchar_for_unicode
+        self.use_ansi = use_ansi
+        self.optimize_limits = optimize_limits
+        self.exclude_tablespaces = exclude_tablespaces
+        self.enable_offset_fetch = self._supports_offset_fetch = (
+            enable_offset_fetch
+        )
+
+    def initialize(self, connection):
+        super().initialize(connection)
+
+        # Oracle 8i has RETURNING:
+        # https://docs.oracle.com/cd/A87860_01/doc/index.htm
+
+        # so does Oracle8:
+        # https://docs.oracle.com/cd/A64702_01/doc/index.htm
+
+        if self._is_oracle_8:
+            self.colspecs = self.colspecs.copy()
+            self.colspecs.pop(sqltypes.Interval)
+            self.use_ansi = False
+
+        self.supports_identity_columns = self.server_version_info >= (12,)
+        self._supports_offset_fetch = (
+            self.enable_offset_fetch and self.server_version_info >= (12,)
+        )
+
+    def _get_effective_compat_server_version_info(self, connection):
+        # dialect does not need compat levels below 12.2, so don't query
+        # in those cases
+
+        if self.server_version_info < (12, 2):
+            return self.server_version_info
+        try:
+            compat = connection.exec_driver_sql(
+                "SELECT value FROM v$parameter WHERE name = 'compatible'"
+            ).scalar()
+        except exc.DBAPIError:
+            compat = None
+
+        if compat:
+            try:
+                return tuple(int(x) for x in compat.split("."))
+            except:
+                return self.server_version_info
+        else:
+            return self.server_version_info
+
+    @property
+    def _is_oracle_8(self):
+        return self.server_version_info and self.server_version_info < (9,)
+
+    @property
+    def _supports_table_compression(self):
+        return self.server_version_info and self.server_version_info >= (10, 1)
+
+    @property
+    def _supports_table_compress_for(self):
+        return self.server_version_info and self.server_version_info >= (11,)
+
+    @property
+    def _supports_char_length(self):
+        return not self._is_oracle_8
+
+    @property
+    def _supports_update_returning_computed_cols(self):
+        # on version 18 this error is no longet present while it happens on 11
+        # it may work also on versions before the 18
+        return self.server_version_info and self.server_version_info >= (18,)
+
+    @property
+    def _supports_except_all(self):
+        return self.server_version_info and self.server_version_info >= (21,)
+
+    def do_release_savepoint(self, connection, name):
+        # Oracle does not support RELEASE SAVEPOINT
+        pass
+
+    def _check_max_identifier_length(self, connection):
+        if self._get_effective_compat_server_version_info(connection) < (
+            12,
+            2,
+        ):
+            return 30
+        else:
+            # use the default
+            return None
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return ["READ COMMITTED", "SERIALIZABLE"]
+
+    def get_default_isolation_level(self, dbapi_conn):
+        try:
+            return self.get_isolation_level(dbapi_conn)
+        except NotImplementedError:
+            raise
+        except:
+            return "READ COMMITTED"
+
+    def _execute_reflection(
+        self, connection, query, dblink, returns_long, params=None
+    ):
+        if dblink and not dblink.startswith("@"):
+            dblink = f"@{dblink}"
+        execution_options = {
+            # handle db links
+            "_oracle_dblink": dblink or "",
+            # override any schema translate map
+            "schema_translate_map": None,
+        }
+
+        if dblink and returns_long:
+            # Oracle seems to error with
+            # "ORA-00997: illegal use of LONG datatype" when returning
+            # LONG columns via a dblink in a query with bind params
+            # This type seems to be very hard to cast into something else
+            # so it seems easier to just use bind param in this case
+            def visit_bindparam(bindparam):
+                bindparam.literal_execute = True
+
+            query = visitors.cloned_traverse(
+                query, {}, {"bindparam": visit_bindparam}
+            )
+        return connection.execute(
+            query, params, execution_options=execution_options
+        )
+
+    @util.memoized_property
+    def _has_table_query(self):
+        # materialized views are returned by all_tables
+        tables = (
+            select(
+                dictionary.all_tables.c.table_name,
+                dictionary.all_tables.c.owner,
+            )
+            .union_all(
+                select(
+                    dictionary.all_views.c.view_name.label("table_name"),
+                    dictionary.all_views.c.owner,
+                )
+            )
+            .subquery("tables_and_views")
+        )
+
+        query = select(tables.c.table_name).where(
+            tables.c.table_name == bindparam("table_name"),
+            tables.c.owner == bindparam("owner"),
+        )
+        return query
+
+    @reflection.cache
+    def has_table(
+        self, connection, table_name, schema=None, dblink=None, **kw
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        self._ensure_has_table_connection(connection)
+
+        if not schema:
+            schema = self.default_schema_name
+
+        params = {
+            "table_name": self.denormalize_name(table_name),
+            "owner": self.denormalize_schema_name(schema),
+        }
+        cursor = self._execute_reflection(
+            connection,
+            self._has_table_query,
+            dblink,
+            returns_long=False,
+            params=params,
+        )
+        return bool(cursor.scalar())
+
+    @reflection.cache
+    def has_sequence(
+        self, connection, sequence_name, schema=None, dblink=None, **kw
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        if not schema:
+            schema = self.default_schema_name
+
+        query = select(dictionary.all_sequences.c.sequence_name).where(
+            dictionary.all_sequences.c.sequence_name
+            == self.denormalize_schema_name(sequence_name),
+            dictionary.all_sequences.c.sequence_owner
+            == self.denormalize_schema_name(schema),
+        )
+
+        cursor = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        )
+        return bool(cursor.scalar())
+
+    def _get_default_schema_name(self, connection):
+        return self.normalize_name(
+            connection.exec_driver_sql(
+                "select sys_context( 'userenv', 'current_schema' ) from dual"
+            ).scalar()
+        )
+
+    def denormalize_schema_name(self, name):
+        # look for quoted_name
+        force = getattr(name, "quote", None)
+        if force is None and name == "public":
+            # look for case insensitive, no quoting specified, "public"
+            return "PUBLIC"
+        return super().denormalize_name(name)
+
+    @reflection.flexi_cache(
+        ("schema", InternalTraversal.dp_string),
+        ("filter_names", InternalTraversal.dp_string_list),
+        ("dblink", InternalTraversal.dp_string),
+    )
+    def _get_synonyms(self, connection, schema, filter_names, dblink, **kw):
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = select(
+            dictionary.all_synonyms.c.synonym_name,
+            dictionary.all_synonyms.c.table_name,
+            dictionary.all_synonyms.c.table_owner,
+            dictionary.all_synonyms.c.db_link,
+        ).where(dictionary.all_synonyms.c.owner == owner)
+        if has_filter_names:
+            query = query.where(
+                dictionary.all_synonyms.c.synonym_name.in_(
+                    params["filter_names"]
+                )
+            )
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).mappings()
+        return result.all()
+
+    @lru_cache()
+    def _all_objects_query(
+        self, owner, scope, kind, has_filter_names, has_mat_views
+    ):
+        query = (
+            select(dictionary.all_objects.c.object_name)
+            .select_from(dictionary.all_objects)
+            .where(dictionary.all_objects.c.owner == owner)
+        )
+
+        # NOTE: materialized views are listed in all_objects twice;
+        # once as MATERIALIZE VIEW and once as TABLE
+        if kind is ObjectKind.ANY:
+            # materilaized view are listed also as tables so there is no
+            # need to add them to the in_.
+            query = query.where(
+                dictionary.all_objects.c.object_type.in_(("TABLE", "VIEW"))
+            )
+        else:
+            object_type = []
+            if ObjectKind.VIEW in kind:
+                object_type.append("VIEW")
+            if (
+                ObjectKind.MATERIALIZED_VIEW in kind
+                and ObjectKind.TABLE not in kind
+            ):
+                # materilaized view are listed also as tables so there is no
+                # need to add them to the in_ if also selecting tables.
+                object_type.append("MATERIALIZED VIEW")
+            if ObjectKind.TABLE in kind:
+                object_type.append("TABLE")
+                if has_mat_views and ObjectKind.MATERIALIZED_VIEW not in kind:
+                    # materialized view are listed also as tables,
+                    # so they need to be filtered out
+                    # EXCEPT ALL / MINUS profiles as faster than using
+                    # NOT EXISTS or NOT IN with a subquery, but it's in
+                    # general faster to get the mat view names and exclude
+                    # them only when needed
+                    query = query.where(
+                        dictionary.all_objects.c.object_name.not_in(
+                            bindparam("mat_views")
+                        )
+                    )
+            query = query.where(
+                dictionary.all_objects.c.object_type.in_(object_type)
+            )
+
+        # handles scope
+        if scope is ObjectScope.DEFAULT:
+            query = query.where(dictionary.all_objects.c.temporary == "N")
+        elif scope is ObjectScope.TEMPORARY:
+            query = query.where(dictionary.all_objects.c.temporary == "Y")
+
+        if has_filter_names:
+            query = query.where(
+                dictionary.all_objects.c.object_name.in_(
+                    bindparam("filter_names")
+                )
+            )
+        return query
+
+    @reflection.flexi_cache(
+        ("schema", InternalTraversal.dp_string),
+        ("scope", InternalTraversal.dp_plain_obj),
+        ("kind", InternalTraversal.dp_plain_obj),
+        ("filter_names", InternalTraversal.dp_string_list),
+        ("dblink", InternalTraversal.dp_string),
+    )
+    def _get_all_objects(
+        self, connection, schema, scope, kind, filter_names, dblink, **kw
+    ):
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        has_mat_views = False
+        if (
+            ObjectKind.TABLE in kind
+            and ObjectKind.MATERIALIZED_VIEW not in kind
+        ):
+            # see note in _all_objects_query
+            mat_views = self.get_materialized_view_names(
+                connection, schema, dblink, _normalize=False, **kw
+            )
+            if mat_views:
+                params["mat_views"] = mat_views
+                has_mat_views = True
+
+        query = self._all_objects_query(
+            owner, scope, kind, has_filter_names, has_mat_views
+        )
+
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False, params=params
+        ).scalars()
+
+        return result.all()
+
+    def _handle_synonyms_decorator(fn):
+        @wraps(fn)
+        def wrapper(self, *args, **kwargs):
+            return self._handle_synonyms(fn, *args, **kwargs)
+
+        return wrapper
+
+    def _handle_synonyms(self, fn, connection, *args, **kwargs):
+        if not kwargs.get("oracle_resolve_synonyms", False):
+            return fn(self, connection, *args, **kwargs)
+
+        original_kw = kwargs.copy()
+        schema = kwargs.pop("schema", None)
+        result = self._get_synonyms(
+            connection,
+            schema=schema,
+            filter_names=kwargs.pop("filter_names", None),
+            dblink=kwargs.pop("dblink", None),
+            info_cache=kwargs.get("info_cache", None),
+        )
+
+        dblinks_owners = defaultdict(dict)
+        for row in result:
+            key = row["db_link"], row["table_owner"]
+            tn = self.normalize_name(row["table_name"])
+            dblinks_owners[key][tn] = row["synonym_name"]
+
+        if not dblinks_owners:
+            # No synonym, do the plain thing
+            return fn(self, connection, *args, **original_kw)
+
+        data = {}
+        for (dblink, table_owner), mapping in dblinks_owners.items():
+            call_kw = {
+                **original_kw,
+                "schema": table_owner,
+                "dblink": self.normalize_name(dblink),
+                "filter_names": mapping.keys(),
+            }
+            call_result = fn(self, connection, *args, **call_kw)
+            for (_, tn), value in call_result:
+                synonym_name = self.normalize_name(mapping[tn])
+                data[(schema, synonym_name)] = value
+        return data.items()
+
+    @reflection.cache
+    def get_schema_names(self, connection, dblink=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        query = select(dictionary.all_users.c.username).order_by(
+            dictionary.all_users.c.username
+        )
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(row) for row in result]
+
+    @reflection.cache
+    def get_table_names(self, connection, schema=None, dblink=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        # note that table_names() isn't loading DBLINKed or synonym'ed tables
+        if schema is None:
+            schema = self.default_schema_name
+
+        den_schema = self.denormalize_schema_name(schema)
+        if kw.get("oracle_resolve_synonyms", False):
+            tables = (
+                select(
+                    dictionary.all_tables.c.table_name,
+                    dictionary.all_tables.c.owner,
+                    dictionary.all_tables.c.iot_name,
+                    dictionary.all_tables.c.duration,
+                    dictionary.all_tables.c.tablespace_name,
+                )
+                .union_all(
+                    select(
+                        dictionary.all_synonyms.c.synonym_name.label(
+                            "table_name"
+                        ),
+                        dictionary.all_synonyms.c.owner,
+                        dictionary.all_tables.c.iot_name,
+                        dictionary.all_tables.c.duration,
+                        dictionary.all_tables.c.tablespace_name,
+                    )
+                    .select_from(dictionary.all_tables)
+                    .join(
+                        dictionary.all_synonyms,
+                        and_(
+                            dictionary.all_tables.c.table_name
+                            == dictionary.all_synonyms.c.table_name,
+                            dictionary.all_tables.c.owner
+                            == func.coalesce(
+                                dictionary.all_synonyms.c.table_owner,
+                                dictionary.all_synonyms.c.owner,
+                            ),
+                        ),
+                    )
+                )
+                .subquery("available_tables")
+            )
+        else:
+            tables = dictionary.all_tables
+
+        query = select(tables.c.table_name)
+        if self.exclude_tablespaces:
+            query = query.where(
+                func.coalesce(
+                    tables.c.tablespace_name, "no tablespace"
+                ).not_in(self.exclude_tablespaces)
+            )
+        query = query.where(
+            tables.c.owner == den_schema,
+            tables.c.iot_name.is_(null()),
+            tables.c.duration.is_(null()),
+        )
+
+        # remove materialized views
+        mat_query = select(
+            dictionary.all_mviews.c.mview_name.label("table_name")
+        ).where(dictionary.all_mviews.c.owner == den_schema)
+
+        query = (
+            query.except_all(mat_query)
+            if self._supports_except_all
+            else query.except_(mat_query)
+        )
+
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(row) for row in result]
+
+    @reflection.cache
+    def get_temp_table_names(self, connection, dblink=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        schema = self.denormalize_schema_name(self.default_schema_name)
+
+        query = select(dictionary.all_tables.c.table_name)
+        if self.exclude_tablespaces:
+            query = query.where(
+                func.coalesce(
+                    dictionary.all_tables.c.tablespace_name, "no tablespace"
+                ).not_in(self.exclude_tablespaces)
+            )
+        query = query.where(
+            dictionary.all_tables.c.owner == schema,
+            dictionary.all_tables.c.iot_name.is_(null()),
+            dictionary.all_tables.c.duration.is_not(null()),
+        )
+
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(row) for row in result]
+
+    @reflection.cache
+    def get_materialized_view_names(
+        self, connection, schema=None, dblink=None, _normalize=True, **kw
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        if not schema:
+            schema = self.default_schema_name
+
+        query = select(dictionary.all_mviews.c.mview_name).where(
+            dictionary.all_mviews.c.owner
+            == self.denormalize_schema_name(schema)
+        )
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        if _normalize:
+            return [self.normalize_name(row) for row in result]
+        else:
+            return result.all()
+
+    @reflection.cache
+    def get_view_names(self, connection, schema=None, dblink=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        if not schema:
+            schema = self.default_schema_name
+
+        query = select(dictionary.all_views.c.view_name).where(
+            dictionary.all_views.c.owner
+            == self.denormalize_schema_name(schema)
+        )
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(row) for row in result]
+
+    @reflection.cache
+    def get_sequence_names(self, connection, schema=None, dblink=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link."""
+        if not schema:
+            schema = self.default_schema_name
+        query = select(dictionary.all_sequences.c.sequence_name).where(
+            dictionary.all_sequences.c.sequence_owner
+            == self.denormalize_schema_name(schema)
+        )
+
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(row) for row in result]
+
+    def _value_or_raise(self, data, table, schema):
+        table = self.normalize_name(str(table))
+        try:
+            return dict(data)[(schema, table)]
+        except KeyError:
+            raise exc.NoSuchTableError(
+                f"{schema}.{table}" if schema else table
+            ) from None
+
+    def _prepare_filter_names(self, filter_names):
+        if filter_names:
+            fn = [self.denormalize_name(name) for name in filter_names]
+            return True, {"filter_names": fn}
+        else:
+            return False, {}
+
+    @reflection.cache
+    def get_table_options(self, connection, table_name, schema=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_table_options(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _table_options_query(
+        self, owner, scope, kind, has_filter_names, has_mat_views
+    ):
+        query = select(
+            dictionary.all_tables.c.table_name,
+            (
+                dictionary.all_tables.c.compression
+                if self._supports_table_compression
+                else sql.null().label("compression")
+            ),
+            (
+                dictionary.all_tables.c.compress_for
+                if self._supports_table_compress_for
+                else sql.null().label("compress_for")
+            ),
+            dictionary.all_tables.c.tablespace_name,
+        ).where(dictionary.all_tables.c.owner == owner)
+        if has_filter_names:
+            query = query.where(
+                dictionary.all_tables.c.table_name.in_(
+                    bindparam("filter_names")
+                )
+            )
+        if scope is ObjectScope.DEFAULT:
+            query = query.where(dictionary.all_tables.c.duration.is_(null()))
+        elif scope is ObjectScope.TEMPORARY:
+            query = query.where(
+                dictionary.all_tables.c.duration.is_not(null())
+            )
+
+        if (
+            has_mat_views
+            and ObjectKind.TABLE in kind
+            and ObjectKind.MATERIALIZED_VIEW not in kind
+        ):
+            # cant use EXCEPT ALL / MINUS here because we don't have an
+            # excludable row vs. the query above
+            # outerjoin + where null works better on oracle 21 but 11 does
+            # not like it at all. this is the next best thing
+
+            query = query.where(
+                dictionary.all_tables.c.table_name.not_in(
+                    bindparam("mat_views")
+                )
+            )
+        elif (
+            ObjectKind.TABLE not in kind
+            and ObjectKind.MATERIALIZED_VIEW in kind
+        ):
+            query = query.where(
+                dictionary.all_tables.c.table_name.in_(bindparam("mat_views"))
+            )
+        return query
+
+    @_handle_synonyms_decorator
+    def get_multi_table_options(
+        self,
+        connection,
+        *,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        has_mat_views = False
+
+        if (
+            ObjectKind.TABLE in kind
+            and ObjectKind.MATERIALIZED_VIEW not in kind
+        ):
+            # see note in _table_options_query
+            mat_views = self.get_materialized_view_names(
+                connection, schema, dblink, _normalize=False, **kw
+            )
+            if mat_views:
+                params["mat_views"] = mat_views
+                has_mat_views = True
+        elif (
+            ObjectKind.TABLE not in kind
+            and ObjectKind.MATERIALIZED_VIEW in kind
+        ):
+            mat_views = self.get_materialized_view_names(
+                connection, schema, dblink, _normalize=False, **kw
+            )
+            params["mat_views"] = mat_views
+
+        options = {}
+        default = ReflectionDefaults.table_options
+
+        if ObjectKind.TABLE in kind or ObjectKind.MATERIALIZED_VIEW in kind:
+            query = self._table_options_query(
+                owner, scope, kind, has_filter_names, has_mat_views
+            )
+            result = self._execute_reflection(
+                connection, query, dblink, returns_long=False, params=params
+            )
+
+            for table, compression, compress_for, tablespace in result:
+                data = default()
+                if compression == "ENABLED":
+                    data["oracle_compress"] = compress_for
+                if tablespace:
+                    data["oracle_tablespace"] = tablespace
+                options[(schema, self.normalize_name(table))] = data
+        if ObjectKind.VIEW in kind and ObjectScope.DEFAULT in scope:
+            # add the views (no temporary views)
+            for view in self.get_view_names(connection, schema, dblink, **kw):
+                if not filter_names or view in filter_names:
+                    options[(schema, view)] = default()
+
+        return options.items()
+
+    @reflection.cache
+    def get_columns(self, connection, table_name, schema=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+
+        data = self.get_multi_columns(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    def _run_batches(
+        self, connection, query, dblink, returns_long, mappings, all_objects
+    ):
+        each_batch = 500
+        batches = list(all_objects)
+        while batches:
+            batch = batches[0:each_batch]
+            batches[0:each_batch] = []
+
+            result = self._execute_reflection(
+                connection,
+                query,
+                dblink,
+                returns_long=returns_long,
+                params={"all_objects": batch},
+            )
+            if mappings:
+                yield from result.mappings()
+            else:
+                yield from result
+
+    @lru_cache()
+    def _column_query(self, owner):
+        all_cols = dictionary.all_tab_cols
+        all_comments = dictionary.all_col_comments
+        all_ids = dictionary.all_tab_identity_cols
+
+        if self.server_version_info >= (12,):
+            add_cols = (
+                all_cols.c.default_on_null,
+                sql.case(
+                    (all_ids.c.table_name.is_(None), sql.null()),
+                    else_=all_ids.c.generation_type
+                    + ","
+                    + all_ids.c.identity_options,
+                ).label("identity_options"),
+            )
+            join_identity_cols = True
+        else:
+            add_cols = (
+                sql.null().label("default_on_null"),
+                sql.null().label("identity_options"),
+            )
+            join_identity_cols = False
+
+        # NOTE: on oracle cannot create tables/views without columns and
+        # a table cannot have all column hidden:
+        # ORA-54039: table must have at least one column that is not invisible
+        # all_tab_cols returns data for tables/views/mat-views.
+        # all_tab_cols does not return recycled tables
+
+        query = (
+            select(
+                all_cols.c.table_name,
+                all_cols.c.column_name,
+                all_cols.c.data_type,
+                all_cols.c.char_length,
+                all_cols.c.data_precision,
+                all_cols.c.data_scale,
+                all_cols.c.nullable,
+                all_cols.c.data_default,
+                all_comments.c.comments,
+                all_cols.c.virtual_column,
+                *add_cols,
+            ).select_from(all_cols)
+            # NOTE: all_col_comments has a row for each column even if no
+            # comment is present, so a join could be performed, but there
+            # seems to be no difference compared to an outer join
+            .outerjoin(
+                all_comments,
+                and_(
+                    all_cols.c.table_name == all_comments.c.table_name,
+                    all_cols.c.column_name == all_comments.c.column_name,
+                    all_cols.c.owner == all_comments.c.owner,
+                ),
+            )
+        )
+        if join_identity_cols:
+            query = query.outerjoin(
+                all_ids,
+                and_(
+                    all_cols.c.table_name == all_ids.c.table_name,
+                    all_cols.c.column_name == all_ids.c.column_name,
+                    all_cols.c.owner == all_ids.c.owner,
+                ),
+            )
+
+        query = query.where(
+            all_cols.c.table_name.in_(bindparam("all_objects")),
+            all_cols.c.hidden_column == "NO",
+            all_cols.c.owner == owner,
+        ).order_by(all_cols.c.table_name, all_cols.c.column_id)
+        return query
+
+    @_handle_synonyms_decorator
+    def get_multi_columns(
+        self,
+        connection,
+        *,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+        query = self._column_query(owner)
+
+        if (
+            filter_names
+            and kind is ObjectKind.ANY
+            and scope is ObjectScope.ANY
+        ):
+            all_objects = [self.denormalize_name(n) for n in filter_names]
+        else:
+            all_objects = self._get_all_objects(
+                connection, schema, scope, kind, filter_names, dblink, **kw
+            )
+
+        columns = defaultdict(list)
+
+        # all_tab_cols.data_default is LONG
+        result = self._run_batches(
+            connection,
+            query,
+            dblink,
+            returns_long=True,
+            mappings=True,
+            all_objects=all_objects,
+        )
+
+        def maybe_int(value):
+            if isinstance(value, float) and value.is_integer():
+                return int(value)
+            else:
+                return value
+
+        remove_size = re.compile(r"\(\d+\)")
+
+        for row_dict in result:
+            table_name = self.normalize_name(row_dict["table_name"])
+            orig_colname = row_dict["column_name"]
+            colname = self.normalize_name(orig_colname)
+            coltype = row_dict["data_type"]
+            precision = maybe_int(row_dict["data_precision"])
+
+            if coltype == "NUMBER":
+                scale = maybe_int(row_dict["data_scale"])
+                if precision is None and scale == 0:
+                    coltype = INTEGER()
+                else:
+                    coltype = NUMBER(precision, scale)
+            elif coltype == "FLOAT":
+                # https://docs.oracle.com/cd/B14117_01/server.101/b10758/sqlqr06.htm
+                if precision == 126:
+                    # The DOUBLE PRECISION datatype is a floating-point
+                    # number with binary precision 126.
+                    coltype = DOUBLE_PRECISION()
+                elif precision == 63:
+                    # The REAL datatype is a floating-point number with a
+                    # binary precision of 63, or 18 decimal.
+                    coltype = REAL()
+                else:
+                    # non standard precision
+                    coltype = FLOAT(binary_precision=precision)
+
+            elif coltype in ("VARCHAR2", "NVARCHAR2", "CHAR", "NCHAR"):
+                char_length = maybe_int(row_dict["char_length"])
+                coltype = self.ischema_names.get(coltype)(char_length)
+            elif "WITH TIME ZONE" in coltype:
+                coltype = TIMESTAMP(timezone=True)
+            elif "WITH LOCAL TIME ZONE" in coltype:
+                coltype = TIMESTAMP(local_timezone=True)
+            else:
+                coltype = re.sub(remove_size, "", coltype)
+                try:
+                    coltype = self.ischema_names[coltype]
+                except KeyError:
+                    util.warn(
+                        "Did not recognize type '%s' of column '%s'"
+                        % (coltype, colname)
+                    )
+                    coltype = sqltypes.NULLTYPE
+
+            default = row_dict["data_default"]
+            if row_dict["virtual_column"] == "YES":
+                computed = dict(sqltext=default)
+                default = None
+            else:
+                computed = None
+
+            identity_options = row_dict["identity_options"]
+            if identity_options is not None:
+                identity = self._parse_identity_options(
+                    identity_options, row_dict["default_on_null"]
+                )
+                default = None
+            else:
+                identity = None
+
+            cdict = {
+                "name": colname,
+                "type": coltype,
+                "nullable": row_dict["nullable"] == "Y",
+                "default": default,
+                "comment": row_dict["comments"],
+            }
+            if orig_colname.lower() == orig_colname:
+                cdict["quote"] = True
+            if computed is not None:
+                cdict["computed"] = computed
+            if identity is not None:
+                cdict["identity"] = identity
+
+            columns[(schema, table_name)].append(cdict)
+
+        # NOTE: default not needed since all tables have columns
+        # default = ReflectionDefaults.columns
+        # return (
+        #     (key, value if value else default())
+        #     for key, value in columns.items()
+        # )
+        return columns.items()
+
+    def _parse_identity_options(self, identity_options, default_on_null):
+        # identity_options is a string that starts with 'ALWAYS,' or
+        # 'BY DEFAULT,' and continues with
+        # START WITH: 1, INCREMENT BY: 1, MAX_VALUE: 123, MIN_VALUE: 1,
+        # CYCLE_FLAG: N, CACHE_SIZE: 1, ORDER_FLAG: N, SCALE_FLAG: N,
+        # EXTEND_FLAG: N, SESSION_FLAG: N, KEEP_VALUE: N
+        parts = [p.strip() for p in identity_options.split(",")]
+        identity = {
+            "always": parts[0] == "ALWAYS",
+            "on_null": default_on_null == "YES",
+        }
+
+        for part in parts[1:]:
+            option, value = part.split(":")
+            value = value.strip()
+
+            if "START WITH" in option:
+                identity["start"] = int(value)
+            elif "INCREMENT BY" in option:
+                identity["increment"] = int(value)
+            elif "MAX_VALUE" in option:
+                identity["maxvalue"] = int(value)
+            elif "MIN_VALUE" in option:
+                identity["minvalue"] = int(value)
+            elif "CYCLE_FLAG" in option:
+                identity["cycle"] = value == "Y"
+            elif "CACHE_SIZE" in option:
+                identity["cache"] = int(value)
+            elif "ORDER_FLAG" in option:
+                identity["order"] = value == "Y"
+        return identity
+
+    @reflection.cache
+    def get_table_comment(self, connection, table_name, schema=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_table_comment(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _comment_query(self, owner, scope, kind, has_filter_names):
+        # NOTE: all_tab_comments / all_mview_comments have a row for all
+        # object even if they don't have comments
+        queries = []
+        if ObjectKind.TABLE in kind or ObjectKind.VIEW in kind:
+            # all_tab_comments returns also plain views
+            tbl_view = select(
+                dictionary.all_tab_comments.c.table_name,
+                dictionary.all_tab_comments.c.comments,
+            ).where(
+                dictionary.all_tab_comments.c.owner == owner,
+                dictionary.all_tab_comments.c.table_name.not_like("BIN$%"),
+            )
+            if ObjectKind.VIEW not in kind:
+                tbl_view = tbl_view.where(
+                    dictionary.all_tab_comments.c.table_type == "TABLE"
+                )
+            elif ObjectKind.TABLE not in kind:
+                tbl_view = tbl_view.where(
+                    dictionary.all_tab_comments.c.table_type == "VIEW"
+                )
+            queries.append(tbl_view)
+        if ObjectKind.MATERIALIZED_VIEW in kind:
+            mat_view = select(
+                dictionary.all_mview_comments.c.mview_name.label("table_name"),
+                dictionary.all_mview_comments.c.comments,
+            ).where(
+                dictionary.all_mview_comments.c.owner == owner,
+                dictionary.all_mview_comments.c.mview_name.not_like("BIN$%"),
+            )
+            queries.append(mat_view)
+        if len(queries) == 1:
+            query = queries[0]
+        else:
+            union = sql.union_all(*queries).subquery("tables_and_views")
+            query = select(union.c.table_name, union.c.comments)
+
+        name_col = query.selected_columns.table_name
+
+        if scope in (ObjectScope.DEFAULT, ObjectScope.TEMPORARY):
+            temp = "Y" if scope is ObjectScope.TEMPORARY else "N"
+            # need distinct since materialized view are listed also
+            # as tables in all_objects
+            query = query.distinct().join(
+                dictionary.all_objects,
+                and_(
+                    dictionary.all_objects.c.owner == owner,
+                    dictionary.all_objects.c.object_name == name_col,
+                    dictionary.all_objects.c.temporary == temp,
+                ),
+            )
+        if has_filter_names:
+            query = query.where(name_col.in_(bindparam("filter_names")))
+        return query
+
+    @_handle_synonyms_decorator
+    def get_multi_table_comment(
+        self,
+        connection,
+        *,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = self._comment_query(owner, scope, kind, has_filter_names)
+
+        result = self._execute_reflection(
+            connection, query, dblink, returns_long=False, params=params
+        )
+        default = ReflectionDefaults.table_comment
+        # materialized views by default seem to have a comment like
+        # "snapshot table for snapshot owner.mat_view_name"
+        ignore_mat_view = "snapshot table for snapshot "
+        return (
+            (
+                (schema, self.normalize_name(table)),
+                (
+                    {"text": comment}
+                    if comment is not None
+                    and not comment.startswith(ignore_mat_view)
+                    else default()
+                ),
+            )
+            for table, comment in result
+        )
+
+    @reflection.cache
+    def get_indexes(self, connection, table_name, schema=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_indexes(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _index_query(self, owner):
+        return (
+            select(
+                dictionary.all_ind_columns.c.table_name,
+                dictionary.all_ind_columns.c.index_name,
+                dictionary.all_ind_columns.c.column_name,
+                dictionary.all_indexes.c.index_type,
+                dictionary.all_indexes.c.uniqueness,
+                dictionary.all_indexes.c.compression,
+                dictionary.all_indexes.c.prefix_length,
+                dictionary.all_ind_columns.c.descend,
+                dictionary.all_ind_expressions.c.column_expression,
+            )
+            .select_from(dictionary.all_ind_columns)
+            .join(
+                dictionary.all_indexes,
+                sql.and_(
+                    dictionary.all_ind_columns.c.index_name
+                    == dictionary.all_indexes.c.index_name,
+                    dictionary.all_ind_columns.c.index_owner
+                    == dictionary.all_indexes.c.owner,
+                ),
+            )
+            .outerjoin(
+                # NOTE: this adds about 20% to the query time. Using a
+                # case expression with a scalar subquery only when needed
+                # with the assumption that most indexes are not expression
+                # would be faster but oracle does not like that with
+                # LONG datatype. It errors with:
+                # ORA-00997: illegal use of LONG datatype
+                dictionary.all_ind_expressions,
+                sql.and_(
+                    dictionary.all_ind_expressions.c.index_name
+                    == dictionary.all_ind_columns.c.index_name,
+                    dictionary.all_ind_expressions.c.index_owner
+                    == dictionary.all_ind_columns.c.index_owner,
+                    dictionary.all_ind_expressions.c.column_position
+                    == dictionary.all_ind_columns.c.column_position,
+                ),
+            )
+            .where(
+                dictionary.all_indexes.c.table_owner == owner,
+                dictionary.all_indexes.c.table_name.in_(
+                    bindparam("all_objects")
+                ),
+            )
+            .order_by(
+                dictionary.all_ind_columns.c.index_name,
+                dictionary.all_ind_columns.c.column_position,
+            )
+        )
+
+    @reflection.flexi_cache(
+        ("schema", InternalTraversal.dp_string),
+        ("dblink", InternalTraversal.dp_string),
+        ("all_objects", InternalTraversal.dp_string_list),
+    )
+    def _get_indexes_rows(self, connection, schema, dblink, all_objects, **kw):
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+
+        query = self._index_query(owner)
+
+        pks = {
+            row_dict["constraint_name"]
+            for row_dict in self._get_all_constraint_rows(
+                connection, schema, dblink, all_objects, **kw
+            )
+            if row_dict["constraint_type"] == "P"
+        }
+
+        # all_ind_expressions.column_expression is LONG
+        result = self._run_batches(
+            connection,
+            query,
+            dblink,
+            returns_long=True,
+            mappings=True,
+            all_objects=all_objects,
+        )
+
+        return [
+            row_dict
+            for row_dict in result
+            if row_dict["index_name"] not in pks
+        ]
+
+    @_handle_synonyms_decorator
+    def get_multi_indexes(
+        self,
+        connection,
+        *,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        all_objects = self._get_all_objects(
+            connection, schema, scope, kind, filter_names, dblink, **kw
+        )
+
+        uniqueness = {"NONUNIQUE": False, "UNIQUE": True}
+        enabled = {"DISABLED": False, "ENABLED": True}
+        is_bitmap = {"BITMAP", "FUNCTION-BASED BITMAP"}
+
+        indexes = defaultdict(dict)
+
+        for row_dict in self._get_indexes_rows(
+            connection, schema, dblink, all_objects, **kw
+        ):
+            index_name = self.normalize_name(row_dict["index_name"])
+            table_name = self.normalize_name(row_dict["table_name"])
+            table_indexes = indexes[(schema, table_name)]
+
+            if index_name not in table_indexes:
+                table_indexes[index_name] = index_dict = {
+                    "name": index_name,
+                    "column_names": [],
+                    "dialect_options": {},
+                    "unique": uniqueness.get(row_dict["uniqueness"], False),
+                }
+                do = index_dict["dialect_options"]
+                if row_dict["index_type"] in is_bitmap:
+                    do["oracle_bitmap"] = True
+                if enabled.get(row_dict["compression"], False):
+                    do["oracle_compress"] = row_dict["prefix_length"]
+
+            else:
+                index_dict = table_indexes[index_name]
+
+            expr = row_dict["column_expression"]
+            if expr is not None:
+                index_dict["column_names"].append(None)
+                if "expressions" in index_dict:
+                    index_dict["expressions"].append(expr)
+                else:
+                    index_dict["expressions"] = index_dict["column_names"][:-1]
+                    index_dict["expressions"].append(expr)
+
+                if row_dict["descend"].lower() != "asc":
+                    assert row_dict["descend"].lower() == "desc"
+                    cs = index_dict.setdefault("column_sorting", {})
+                    cs[expr] = ("desc",)
+            else:
+                assert row_dict["descend"].lower() == "asc"
+                cn = self.normalize_name(row_dict["column_name"])
+                index_dict["column_names"].append(cn)
+                if "expressions" in index_dict:
+                    index_dict["expressions"].append(cn)
+
+        default = ReflectionDefaults.indexes
+
+        return (
+            (key, list(indexes[key].values()) if key in indexes else default())
+            for key in (
+                (schema, self.normalize_name(obj_name))
+                for obj_name in all_objects
+            )
+        )
+
+    @reflection.cache
+    def get_pk_constraint(self, connection, table_name, schema=None, **kw):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_pk_constraint(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _constraint_query(self, owner):
+        local = dictionary.all_cons_columns.alias("local")
+        remote = dictionary.all_cons_columns.alias("remote")
+        return (
+            select(
+                dictionary.all_constraints.c.table_name,
+                dictionary.all_constraints.c.constraint_type,
+                dictionary.all_constraints.c.constraint_name,
+                local.c.column_name.label("local_column"),
+                remote.c.table_name.label("remote_table"),
+                remote.c.column_name.label("remote_column"),
+                remote.c.owner.label("remote_owner"),
+                dictionary.all_constraints.c.search_condition,
+                dictionary.all_constraints.c.delete_rule,
+            )
+            .select_from(dictionary.all_constraints)
+            .join(
+                local,
+                and_(
+                    local.c.owner == dictionary.all_constraints.c.owner,
+                    dictionary.all_constraints.c.constraint_name
+                    == local.c.constraint_name,
+                ),
+            )
+            .outerjoin(
+                remote,
+                and_(
+                    dictionary.all_constraints.c.r_owner == remote.c.owner,
+                    dictionary.all_constraints.c.r_constraint_name
+                    == remote.c.constraint_name,
+                    or_(
+                        remote.c.position.is_(sql.null()),
+                        local.c.position == remote.c.position,
+                    ),
+                ),
+            )
+            .where(
+                dictionary.all_constraints.c.owner == owner,
+                dictionary.all_constraints.c.table_name.in_(
+                    bindparam("all_objects")
+                ),
+                dictionary.all_constraints.c.constraint_type.in_(
+                    ("R", "P", "U", "C")
+                ),
+            )
+            .order_by(
+                dictionary.all_constraints.c.constraint_name, local.c.position
+            )
+        )
+
+    @reflection.flexi_cache(
+        ("schema", InternalTraversal.dp_string),
+        ("dblink", InternalTraversal.dp_string),
+        ("all_objects", InternalTraversal.dp_string_list),
+    )
+    def _get_all_constraint_rows(
+        self, connection, schema, dblink, all_objects, **kw
+    ):
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+        query = self._constraint_query(owner)
+
+        # since the result is cached a list must be created
+        values = list(
+            self._run_batches(
+                connection,
+                query,
+                dblink,
+                returns_long=False,
+                mappings=True,
+                all_objects=all_objects,
+            )
+        )
+        return values
+
+    @_handle_synonyms_decorator
+    def get_multi_pk_constraint(
+        self,
+        connection,
+        *,
+        scope,
+        schema,
+        filter_names,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        all_objects = self._get_all_objects(
+            connection, schema, scope, kind, filter_names, dblink, **kw
+        )
+
+        primary_keys = defaultdict(dict)
+        default = ReflectionDefaults.pk_constraint
+
+        for row_dict in self._get_all_constraint_rows(
+            connection, schema, dblink, all_objects, **kw
+        ):
+            if row_dict["constraint_type"] != "P":
+                continue
+            table_name = self.normalize_name(row_dict["table_name"])
+            constraint_name = self.normalize_name(row_dict["constraint_name"])
+            column_name = self.normalize_name(row_dict["local_column"])
+
+            table_pk = primary_keys[(schema, table_name)]
+            if not table_pk:
+                table_pk["name"] = constraint_name
+                table_pk["constrained_columns"] = [column_name]
+            else:
+                table_pk["constrained_columns"].append(column_name)
+
+        return (
+            (key, primary_keys[key] if key in primary_keys else default())
+            for key in (
+                (schema, self.normalize_name(obj_name))
+                for obj_name in all_objects
+            )
+        )
+
+    @reflection.cache
+    def get_foreign_keys(
+        self,
+        connection,
+        table_name,
+        schema=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_foreign_keys(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @_handle_synonyms_decorator
+    def get_multi_foreign_keys(
+        self,
+        connection,
+        *,
+        scope,
+        schema,
+        filter_names,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        all_objects = self._get_all_objects(
+            connection, schema, scope, kind, filter_names, dblink, **kw
+        )
+
+        resolve_synonyms = kw.get("oracle_resolve_synonyms", False)
+
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+
+        all_remote_owners = set()
+        fkeys = defaultdict(dict)
+
+        for row_dict in self._get_all_constraint_rows(
+            connection, schema, dblink, all_objects, **kw
+        ):
+            if row_dict["constraint_type"] != "R":
+                continue
+
+            table_name = self.normalize_name(row_dict["table_name"])
+            constraint_name = self.normalize_name(row_dict["constraint_name"])
+            table_fkey = fkeys[(schema, table_name)]
+
+            assert constraint_name is not None
+
+            local_column = self.normalize_name(row_dict["local_column"])
+            remote_table = self.normalize_name(row_dict["remote_table"])
+            remote_column = self.normalize_name(row_dict["remote_column"])
+            remote_owner_orig = row_dict["remote_owner"]
+            remote_owner = self.normalize_name(remote_owner_orig)
+            if remote_owner_orig is not None:
+                all_remote_owners.add(remote_owner_orig)
+
+            if remote_table is None:
+                # ticket 363
+                if dblink and not dblink.startswith("@"):
+                    dblink = f"@{dblink}"
+                util.warn(
+                    "Got 'None' querying 'table_name' from "
+                    f"all_cons_columns{dblink or ''} - does the user have "
+                    "proper rights to the table?"
+                )
+                continue
+
+            if constraint_name not in table_fkey:
+                table_fkey[constraint_name] = fkey = {
+                    "name": constraint_name,
+                    "constrained_columns": [],
+                    "referred_schema": None,
+                    "referred_table": remote_table,
+                    "referred_columns": [],
+                    "options": {},
+                }
+
+                if resolve_synonyms:
+                    # will be removed below
+                    fkey["_ref_schema"] = remote_owner
+
+                if schema is not None or remote_owner_orig != owner:
+                    fkey["referred_schema"] = remote_owner
+
+                delete_rule = row_dict["delete_rule"]
+                if delete_rule != "NO ACTION":
+                    fkey["options"]["ondelete"] = delete_rule
+
+            else:
+                fkey = table_fkey[constraint_name]
+
+            fkey["constrained_columns"].append(local_column)
+            fkey["referred_columns"].append(remote_column)
+
+        if resolve_synonyms and all_remote_owners:
+            query = select(
+                dictionary.all_synonyms.c.owner,
+                dictionary.all_synonyms.c.table_name,
+                dictionary.all_synonyms.c.table_owner,
+                dictionary.all_synonyms.c.synonym_name,
+            ).where(dictionary.all_synonyms.c.owner.in_(all_remote_owners))
+
+            result = self._execute_reflection(
+                connection, query, dblink, returns_long=False
+            ).mappings()
+
+            remote_owners_lut = {}
+            for row in result:
+                synonym_owner = self.normalize_name(row["owner"])
+                table_name = self.normalize_name(row["table_name"])
+
+                remote_owners_lut[(synonym_owner, table_name)] = (
+                    row["table_owner"],
+                    row["synonym_name"],
+                )
+
+            empty = (None, None)
+            for table_fkeys in fkeys.values():
+                for table_fkey in table_fkeys.values():
+                    key = (
+                        table_fkey.pop("_ref_schema"),
+                        table_fkey["referred_table"],
+                    )
+                    remote_owner, syn_name = remote_owners_lut.get(key, empty)
+                    if syn_name:
+                        sn = self.normalize_name(syn_name)
+                        table_fkey["referred_table"] = sn
+                        if schema is not None or remote_owner != owner:
+                            ro = self.normalize_name(remote_owner)
+                            table_fkey["referred_schema"] = ro
+                        else:
+                            table_fkey["referred_schema"] = None
+        default = ReflectionDefaults.foreign_keys
+
+        return (
+            (key, list(fkeys[key].values()) if key in fkeys else default())
+            for key in (
+                (schema, self.normalize_name(obj_name))
+                for obj_name in all_objects
+            )
+        )
+
+    @reflection.cache
+    def get_unique_constraints(
+        self, connection, table_name, schema=None, **kw
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_unique_constraints(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @_handle_synonyms_decorator
+    def get_multi_unique_constraints(
+        self,
+        connection,
+        *,
+        scope,
+        schema,
+        filter_names,
+        kind,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        all_objects = self._get_all_objects(
+            connection, schema, scope, kind, filter_names, dblink, **kw
+        )
+
+        unique_cons = defaultdict(dict)
+
+        index_names = {
+            row_dict["index_name"]
+            for row_dict in self._get_indexes_rows(
+                connection, schema, dblink, all_objects, **kw
+            )
+        }
+
+        for row_dict in self._get_all_constraint_rows(
+            connection, schema, dblink, all_objects, **kw
+        ):
+            if row_dict["constraint_type"] != "U":
+                continue
+            table_name = self.normalize_name(row_dict["table_name"])
+            constraint_name_orig = row_dict["constraint_name"]
+            constraint_name = self.normalize_name(constraint_name_orig)
+            column_name = self.normalize_name(row_dict["local_column"])
+            table_uc = unique_cons[(schema, table_name)]
+
+            assert constraint_name is not None
+
+            if constraint_name not in table_uc:
+                table_uc[constraint_name] = uc = {
+                    "name": constraint_name,
+                    "column_names": [],
+                    "duplicates_index": (
+                        constraint_name
+                        if constraint_name_orig in index_names
+                        else None
+                    ),
+                }
+            else:
+                uc = table_uc[constraint_name]
+
+            uc["column_names"].append(column_name)
+
+        default = ReflectionDefaults.unique_constraints
+
+        return (
+            (
+                key,
+                (
+                    list(unique_cons[key].values())
+                    if key in unique_cons
+                    else default()
+                ),
+            )
+            for key in (
+                (schema, self.normalize_name(obj_name))
+                for obj_name in all_objects
+            )
+        )
+
+    @reflection.cache
+    def get_view_definition(
+        self,
+        connection,
+        view_name,
+        schema=None,
+        dblink=None,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        if kw.get("oracle_resolve_synonyms", False):
+            synonyms = self._get_synonyms(
+                connection, schema, filter_names=[view_name], dblink=dblink
+            )
+            if synonyms:
+                assert len(synonyms) == 1
+                row_dict = synonyms[0]
+                dblink = self.normalize_name(row_dict["db_link"])
+                schema = row_dict["table_owner"]
+                view_name = row_dict["table_name"]
+
+        name = self.denormalize_name(view_name)
+        owner = self.denormalize_schema_name(
+            schema or self.default_schema_name
+        )
+        query = (
+            select(dictionary.all_views.c.text)
+            .where(
+                dictionary.all_views.c.view_name == name,
+                dictionary.all_views.c.owner == owner,
+            )
+            .union_all(
+                select(dictionary.all_mviews.c.query).where(
+                    dictionary.all_mviews.c.mview_name == name,
+                    dictionary.all_mviews.c.owner == owner,
+                )
+            )
+        )
+
+        rp = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalar()
+        if rp is None:
+            raise exc.NoSuchTableError(
+                f"{schema}.{view_name}" if schema else view_name
+            )
+        else:
+            return rp
+
+    @reflection.cache
+    def get_check_constraints(
+        self, connection, table_name, schema=None, include_all=False, **kw
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        data = self.get_multi_check_constraints(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            include_all=include_all,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @_handle_synonyms_decorator
+    def get_multi_check_constraints(
+        self,
+        connection,
+        *,
+        schema,
+        filter_names,
+        dblink=None,
+        scope,
+        kind,
+        include_all=False,
+        **kw,
+    ):
+        """Supported kw arguments are: ``dblink`` to reflect via a db link;
+        ``oracle_resolve_synonyms`` to resolve names to synonyms
+        """
+        all_objects = self._get_all_objects(
+            connection, schema, scope, kind, filter_names, dblink, **kw
+        )
+
+        not_null = re.compile(r"..+?. IS NOT NULL$")
+
+        check_constraints = defaultdict(list)
+
+        for row_dict in self._get_all_constraint_rows(
+            connection, schema, dblink, all_objects, **kw
+        ):
+            if row_dict["constraint_type"] != "C":
+                continue
+            table_name = self.normalize_name(row_dict["table_name"])
+            constraint_name = self.normalize_name(row_dict["constraint_name"])
+            search_condition = row_dict["search_condition"]
+
+            table_checks = check_constraints[(schema, table_name)]
+            if constraint_name is not None and (
+                include_all or not not_null.match(search_condition)
+            ):
+                table_checks.append(
+                    {"name": constraint_name, "sqltext": search_condition}
+                )
+
+        default = ReflectionDefaults.check_constraints
+
+        return (
+            (
+                key,
+                (
+                    check_constraints[key]
+                    if key in check_constraints
+                    else default()
+                ),
+            )
+            for key in (
+                (schema, self.normalize_name(obj_name))
+                for obj_name in all_objects
+            )
+        )
+
+    def _list_dblinks(self, connection, dblink=None):
+        query = select(dictionary.all_db_links.c.db_link)
+        links = self._execute_reflection(
+            connection, query, dblink, returns_long=False
+        ).scalars()
+        return [self.normalize_name(link) for link in links]
+
+
+class _OuterJoinColumn(sql.ClauseElement):
+    __visit_name__ = "outer_join_column"
+
+    def __init__(self, column):
+        self.column = column
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/cx_oracle.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/cx_oracle.py
new file mode 100644
index 00000000..0514ebbc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/cx_oracle.py
@@ -0,0 +1,1552 @@
+# dialects/oracle/cx_oracle.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r""".. dialect:: oracle+cx_oracle
+    :name: cx-Oracle
+    :dbapi: cx_oracle
+    :connectstring: oracle+cx_oracle://user:pass@hostname:port[/dbname][?service_name=<service>[&key=value&key=value...]]
+    :url: https://oracle.github.io/python-cx_Oracle/
+
+Description
+-----------
+
+cx_Oracle was the original driver for Oracle Database. It was superseded by
+python-oracledb which should be used instead.
+
+DSN vs. Hostname connections
+-----------------------------
+
+cx_Oracle provides several methods of indicating the target database.  The
+dialect translates from a series of different URL forms.
+
+Hostname Connections with Easy Connect Syntax
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Given a hostname, port and service name of the target database, for example
+from Oracle Database's Easy Connect syntax then connect in SQLAlchemy using the
+``service_name`` query string parameter::
+
+    engine = create_engine(
+        "oracle+cx_oracle://scott:tiger@hostname:port?service_name=myservice&encoding=UTF-8&nencoding=UTF-8"
+    )
+
+Note that the default driver value for encoding and nencoding was changed to
+“UTF-8” in cx_Oracle 8.0 so these parameters can be omitted when using that
+version, or later.
+
+To use a full Easy Connect string, pass it as the ``dsn`` key value in a
+:paramref:`_sa.create_engine.connect_args` dictionary::
+
+    import cx_Oracle
+
+    e = create_engine(
+        "oracle+cx_oracle://@",
+        connect_args={
+            "user": "scott",
+            "password": "tiger",
+            "dsn": "hostname:port/myservice?transport_connect_timeout=30&expire_time=60",
+        },
+    )
+
+Connections with tnsnames.ora or to Oracle Autonomous Database
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Alternatively, if no port, database name, or service name is provided, the
+dialect will use an Oracle Database DSN "connection string".  This takes the
+"hostname" portion of the URL as the data source name.  For example, if the
+``tnsnames.ora`` file contains a TNS Alias of ``myalias`` as below:
+
+.. sourcecode:: text
+
+    myalias =
+      (DESCRIPTION =
+        (ADDRESS = (PROTOCOL = TCP)(HOST = mymachine.example.com)(PORT = 1521))
+        (CONNECT_DATA =
+          (SERVER = DEDICATED)
+          (SERVICE_NAME = orclpdb1)
+        )
+      )
+
+The cx_Oracle dialect connects to this database service when ``myalias`` is the
+hostname portion of the URL, without specifying a port, database name or
+``service_name``::
+
+    engine = create_engine("oracle+cx_oracle://scott:tiger@myalias")
+
+Users of Oracle Autonomous Database should use this syntax. If the database is
+configured for mutural TLS ("mTLS"), then you must also configure the cloud
+wallet as shown in cx_Oracle documentation `Connecting to Autononmous Databases
+<https://cx-oracle.readthedocs.io/en/latest/user_guide/connection_handling.html#autonomousdb>`_.
+
+SID Connections
+^^^^^^^^^^^^^^^
+
+To use Oracle Database's obsolete System Identifier connection syntax, the SID
+can be passed in a "database name" portion of the URL::
+
+    engine = create_engine(
+        "oracle+cx_oracle://scott:tiger@hostname:port/dbname"
+    )
+
+Above, the DSN passed to cx_Oracle is created by ``cx_Oracle.makedsn()`` as
+follows::
+
+    >>> import cx_Oracle
+    >>> cx_Oracle.makedsn("hostname", 1521, sid="dbname")
+    '(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=hostname)(PORT=1521))(CONNECT_DATA=(SID=dbname)))'
+
+Note that although the SQLAlchemy syntax ``hostname:port/dbname`` looks like
+Oracle's Easy Connect syntax it is different. It uses a SID in place of the
+service name required by Easy Connect.  The Easy Connect syntax does not
+support SIDs.
+
+Passing cx_Oracle connect arguments
+-----------------------------------
+
+Additional connection arguments can usually be passed via the URL query string;
+particular symbols like ``SYSDBA`` are intercepted and converted to the correct
+symbol::
+
+    e = create_engine(
+        "oracle+cx_oracle://user:pass@dsn?encoding=UTF-8&nencoding=UTF-8&mode=SYSDBA&events=true"
+    )
+
+.. versionchanged:: 1.3 the cx_Oracle dialect now accepts all argument names
+   within the URL string itself, to be passed to the cx_Oracle DBAPI.   As
+   was the case earlier but not correctly documented, the
+   :paramref:`_sa.create_engine.connect_args` parameter also accepts all
+   cx_Oracle DBAPI connect arguments.
+
+To pass arguments directly to ``.connect()`` without using the query
+string, use the :paramref:`_sa.create_engine.connect_args` dictionary.
+Any cx_Oracle parameter value and/or constant may be passed, such as::
+
+    import cx_Oracle
+
+    e = create_engine(
+        "oracle+cx_oracle://user:pass@dsn",
+        connect_args={
+            "encoding": "UTF-8",
+            "nencoding": "UTF-8",
+            "mode": cx_Oracle.SYSDBA,
+            "events": True,
+        },
+    )
+
+Note that the default driver value for ``encoding`` and ``nencoding`` was
+changed to "UTF-8" in cx_Oracle 8.0 so these parameters can be omitted when
+using that version, or later.
+
+Options consumed by the SQLAlchemy cx_Oracle dialect outside of the driver
+--------------------------------------------------------------------------
+
+There are also options that are consumed by the SQLAlchemy cx_oracle dialect
+itself.  These options are always passed directly to :func:`_sa.create_engine`
+, such as::
+
+    e = create_engine(
+        "oracle+cx_oracle://user:pass@dsn", coerce_to_decimal=False
+    )
+
+The parameters accepted by the cx_oracle dialect are as follows:
+
+* ``arraysize`` - set the cx_oracle.arraysize value on cursors; defaults
+  to ``None``, indicating that the driver default should be used (typically
+  the value is 100).  This setting controls how many rows are buffered when
+  fetching rows, and can have a significant effect on performance when
+  modified.
+
+  .. versionchanged:: 2.0.26 - changed the default value from 50 to None,
+    to use the default value of the driver itself.
+
+* ``auto_convert_lobs`` - defaults to True; See :ref:`cx_oracle_lob`.
+
+* ``coerce_to_decimal`` - see :ref:`cx_oracle_numeric` for detail.
+
+* ``encoding_errors`` - see :ref:`cx_oracle_unicode_encoding_errors` for detail.
+
+.. _cx_oracle_sessionpool:
+
+Using cx_Oracle SessionPool
+---------------------------
+
+The cx_Oracle driver provides its own connection pool implementation that may
+be used in place of SQLAlchemy's pooling functionality. The driver pool
+supports Oracle Database features such dead connection detection, connection
+draining for planned database downtime, support for Oracle Application
+Continuity and Transparent Application Continuity, and gives support for
+Database Resident Connection Pooling (DRCP).
+
+Using the driver pool can be achieved by using the
+:paramref:`_sa.create_engine.creator` parameter to provide a function that
+returns a new connection, along with setting
+:paramref:`_sa.create_engine.pool_class` to ``NullPool`` to disable
+SQLAlchemy's pooling::
+
+    import cx_Oracle
+    from sqlalchemy import create_engine
+    from sqlalchemy.pool import NullPool
+
+    pool = cx_Oracle.SessionPool(
+        user="scott",
+        password="tiger",
+        dsn="orclpdb",
+        min=1,
+        max=4,
+        increment=1,
+        threaded=True,
+        encoding="UTF-8",
+        nencoding="UTF-8",
+    )
+
+    engine = create_engine(
+        "oracle+cx_oracle://", creator=pool.acquire, poolclass=NullPool
+    )
+
+The above engine may then be used normally where cx_Oracle's pool handles
+connection pooling::
+
+    with engine.connect() as conn:
+        print(conn.scalar("select 1 from dual"))
+
+As well as providing a scalable solution for multi-user applications, the
+cx_Oracle session pool supports some Oracle features such as DRCP and
+`Application Continuity
+<https://cx-oracle.readthedocs.io/en/latest/user_guide/ha.html#application-continuity-ac>`_.
+
+Note that the pool creation parameters ``threaded``, ``encoding`` and
+``nencoding`` were deprecated in later cx_Oracle releases.
+
+Using Oracle Database Resident Connection Pooling (DRCP)
+--------------------------------------------------------
+
+When using Oracle Database's DRCP, the best practice is to pass a connection
+class and "purity" when acquiring a connection from the SessionPool.  Refer to
+the `cx_Oracle DRCP documentation
+<https://cx-oracle.readthedocs.io/en/latest/user_guide/connection_handling.html#database-resident-connection-pooling-drcp>`_.
+
+This can be achieved by wrapping ``pool.acquire()``::
+
+    import cx_Oracle
+    from sqlalchemy import create_engine
+    from sqlalchemy.pool import NullPool
+
+    pool = cx_Oracle.SessionPool(
+        user="scott",
+        password="tiger",
+        dsn="orclpdb",
+        min=2,
+        max=5,
+        increment=1,
+        threaded=True,
+        encoding="UTF-8",
+        nencoding="UTF-8",
+    )
+
+
+    def creator():
+        return pool.acquire(
+            cclass="MYCLASS", purity=cx_Oracle.ATTR_PURITY_SELF
+        )
+
+
+    engine = create_engine(
+        "oracle+cx_oracle://", creator=creator, poolclass=NullPool
+    )
+
+The above engine may then be used normally where cx_Oracle handles session
+pooling and Oracle Database additionally uses DRCP::
+
+    with engine.connect() as conn:
+        print(conn.scalar("select 1 from dual"))
+
+.. _cx_oracle_unicode:
+
+Unicode
+-------
+
+As is the case for all DBAPIs under Python 3, all strings are inherently
+Unicode strings. In all cases however, the driver requires an explicit
+encoding configuration.
+
+Ensuring the Correct Client Encoding
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The long accepted standard for establishing client encoding for nearly all
+Oracle Database related software is via the `NLS_LANG
+<https://www.oracle.com/database/technologies/faq-nls-lang.html>`_ environment
+variable.  Older versions of cx_Oracle use this environment variable as the
+source of its encoding configuration.  The format of this variable is
+Territory_Country.CharacterSet; a typical value would be
+``AMERICAN_AMERICA.AL32UTF8``.  cx_Oracle version 8 and later use the character
+set "UTF-8" by default, and ignore the character set component of NLS_LANG.
+
+The cx_Oracle driver also supported a programmatic alternative which is to pass
+the ``encoding`` and ``nencoding`` parameters directly to its ``.connect()``
+function.  These can be present in the URL as follows::
+
+    engine = create_engine(
+        "oracle+cx_oracle://scott:tiger@tnsalias?encoding=UTF-8&nencoding=UTF-8"
+    )
+
+For the meaning of the ``encoding`` and ``nencoding`` parameters, please
+consult
+`Characters Sets and National Language Support (NLS) <https://cx-oracle.readthedocs.io/en/latest/user_guide/globalization.html#globalization>`_.
+
+.. seealso::
+
+    `Characters Sets and National Language Support (NLS) <https://cx-oracle.readthedocs.io/en/latest/user_guide/globalization.html#globalization>`_
+    - in the cx_Oracle documentation.
+
+
+Unicode-specific Column datatypes
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The Core expression language handles unicode data by use of the
+:class:`.Unicode` and :class:`.UnicodeText` datatypes.  These types correspond
+to the VARCHAR2 and CLOB Oracle Database datatypes by default.  When using
+these datatypes with Unicode data, it is expected that the database is
+configured with a Unicode-aware character set, as well as that the ``NLS_LANG``
+environment variable is set appropriately (this applies to older versions of
+cx_Oracle), so that the VARCHAR2 and CLOB datatypes can accommodate the data.
+
+In the case that Oracle Database is not configured with a Unicode character
+set, the two options are to use the :class:`_types.NCHAR` and
+:class:`_oracle.NCLOB` datatypes explicitly, or to pass the flag
+``use_nchar_for_unicode=True`` to :func:`_sa.create_engine`, which will cause
+the SQLAlchemy dialect to use NCHAR/NCLOB for the :class:`.Unicode` /
+:class:`.UnicodeText` datatypes instead of VARCHAR/CLOB.
+
+.. versionchanged:: 1.3 The :class:`.Unicode` and :class:`.UnicodeText`
+   datatypes now correspond to the ``VARCHAR2`` and ``CLOB`` Oracle Database
+   datatypes unless the ``use_nchar_for_unicode=True`` is passed to the dialect
+   when :func:`_sa.create_engine` is called.
+
+
+.. _cx_oracle_unicode_encoding_errors:
+
+Encoding Errors
+^^^^^^^^^^^^^^^
+
+For the unusual case that data in Oracle Database is present with a broken
+encoding, the dialect accepts a parameter ``encoding_errors`` which will be
+passed to Unicode decoding functions in order to affect how decoding errors are
+handled.  The value is ultimately consumed by the Python `decode
+<https://docs.python.org/3/library/stdtypes.html#bytes.decode>`_ function, and
+is passed both via cx_Oracle's ``encodingErrors`` parameter consumed by
+``Cursor.var()``, as well as SQLAlchemy's own decoding function, as the
+cx_Oracle dialect makes use of both under different circumstances.
+
+.. versionadded:: 1.3.11
+
+
+.. _cx_oracle_setinputsizes:
+
+Fine grained control over cx_Oracle data binding performance with setinputsizes
+-------------------------------------------------------------------------------
+
+The cx_Oracle DBAPI has a deep and fundamental reliance upon the usage of the
+DBAPI ``setinputsizes()`` call.  The purpose of this call is to establish the
+datatypes that are bound to a SQL statement for Python values being passed as
+parameters.  While virtually no other DBAPI assigns any use to the
+``setinputsizes()`` call, the cx_Oracle DBAPI relies upon it heavily in its
+interactions with the Oracle Database client interface, and in some scenarios
+it is not possible for SQLAlchemy to know exactly how data should be bound, as
+some settings can cause profoundly different performance characteristics, while
+altering the type coercion behavior at the same time.
+
+Users of the cx_Oracle dialect are **strongly encouraged** to read through
+cx_Oracle's list of built-in datatype symbols at
+https://cx-oracle.readthedocs.io/en/latest/api_manual/module.html#database-types.
+Note that in some cases, significant performance degradation can occur when
+using these types vs. not, in particular when specifying ``cx_Oracle.CLOB``.
+
+On the SQLAlchemy side, the :meth:`.DialectEvents.do_setinputsizes` event can
+be used both for runtime visibility (e.g. logging) of the setinputsizes step as
+well as to fully control how ``setinputsizes()`` is used on a per-statement
+basis.
+
+.. versionadded:: 1.2.9 Added :meth:`.DialectEvents.setinputsizes`
+
+
+Example 1 - logging all setinputsizes calls
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The following example illustrates how to log the intermediary values from a
+SQLAlchemy perspective before they are converted to the raw ``setinputsizes()``
+parameter dictionary.  The keys of the dictionary are :class:`.BindParameter`
+objects which have a ``.key`` and a ``.type`` attribute::
+
+    from sqlalchemy import create_engine, event
+
+    engine = create_engine("oracle+cx_oracle://scott:tiger@host/xe")
+
+
+    @event.listens_for(engine, "do_setinputsizes")
+    def _log_setinputsizes(inputsizes, cursor, statement, parameters, context):
+        for bindparam, dbapitype in inputsizes.items():
+            log.info(
+                "Bound parameter name: %s  SQLAlchemy type: %r DBAPI object: %s",
+                bindparam.key,
+                bindparam.type,
+                dbapitype,
+            )
+
+Example 2 - remove all bindings to CLOB
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``CLOB`` datatype in cx_Oracle incurs a significant performance overhead,
+however is set by default for the ``Text`` type within the SQLAlchemy 1.2
+series.   This setting can be modified as follows::
+
+    from sqlalchemy import create_engine, event
+    from cx_Oracle import CLOB
+
+    engine = create_engine("oracle+cx_oracle://scott:tiger@host/xe")
+
+
+    @event.listens_for(engine, "do_setinputsizes")
+    def _remove_clob(inputsizes, cursor, statement, parameters, context):
+        for bindparam, dbapitype in list(inputsizes.items()):
+            if dbapitype is CLOB:
+                del inputsizes[bindparam]
+
+.. _cx_oracle_lob:
+
+LOB Datatypes
+--------------
+
+LOB datatypes refer to the "large object" datatypes such as CLOB, NCLOB and
+BLOB. Modern versions of cx_Oracle is optimized for these datatypes to be
+delivered as a single buffer. As such, SQLAlchemy makes use of these newer type
+handlers by default.
+
+To disable the use of newer type handlers and deliver LOB objects as classic
+buffered objects with a ``read()`` method, the parameter
+``auto_convert_lobs=False`` may be passed to :func:`_sa.create_engine`,
+which takes place only engine-wide.
+
+.. _cx_oracle_returning:
+
+RETURNING Support
+-----------------
+
+The cx_Oracle dialect implements RETURNING using OUT parameters.
+The dialect supports RETURNING fully.
+
+Two Phase Transactions Not Supported
+------------------------------------
+
+Two phase transactions are **not supported** under cx_Oracle due to poor driver
+support. The newer :ref:`oracledb` dialect however **does** support two phase
+transactions.
+
+.. _cx_oracle_numeric:
+
+Precision Numerics
+------------------
+
+SQLAlchemy's numeric types can handle receiving and returning values as Python
+``Decimal`` objects or float objects.  When a :class:`.Numeric` object, or a
+subclass such as :class:`.Float`, :class:`_oracle.DOUBLE_PRECISION` etc. is in
+use, the :paramref:`.Numeric.asdecimal` flag determines if values should be
+coerced to ``Decimal`` upon return, or returned as float objects.  To make
+matters more complicated under Oracle Database, the ``NUMBER`` type can also
+represent integer values if the "scale" is zero, so the Oracle
+Database-specific :class:`_oracle.NUMBER` type takes this into account as well.
+
+The cx_Oracle dialect makes extensive use of connection- and cursor-level
+"outputtypehandler" callables in order to coerce numeric values as requested.
+These callables are specific to the specific flavor of :class:`.Numeric` in
+use, as well as if no SQLAlchemy typing objects are present.  There are
+observed scenarios where Oracle Database may send incomplete or ambiguous
+information about the numeric types being returned, such as a query where the
+numeric types are buried under multiple levels of subquery.  The type handlers
+do their best to make the right decision in all cases, deferring to the
+underlying cx_Oracle DBAPI for all those cases where the driver can make the
+best decision.
+
+When no typing objects are present, as when executing plain SQL strings, a
+default "outputtypehandler" is present which will generally return numeric
+values which specify precision and scale as Python ``Decimal`` objects.  To
+disable this coercion to decimal for performance reasons, pass the flag
+``coerce_to_decimal=False`` to :func:`_sa.create_engine`::
+
+    engine = create_engine("oracle+cx_oracle://dsn", coerce_to_decimal=False)
+
+The ``coerce_to_decimal`` flag only impacts the results of plain string
+SQL statements that are not otherwise associated with a :class:`.Numeric`
+SQLAlchemy type (or a subclass of such).
+
+.. versionchanged:: 1.2  The numeric handling system for cx_Oracle has been
+   reworked to take advantage of newer cx_Oracle features as well
+   as better integration of outputtypehandlers.
+
+"""  # noqa
+from __future__ import annotations
+
+import decimal
+import random
+import re
+
+from . import base as oracle
+from .base import OracleCompiler
+from .base import OracleDialect
+from .base import OracleExecutionContext
+from .types import _OracleDateLiteralRender
+from ... import exc
+from ... import util
+from ...engine import cursor as _cursor
+from ...engine import interfaces
+from ...engine import processors
+from ...sql import sqltypes
+from ...sql._typing import is_sql_compiler
+
+# source:
+# https://github.com/oracle/python-cx_Oracle/issues/596#issuecomment-999243649
+_CX_ORACLE_MAGIC_LOB_SIZE = 131072
+
+
+class _OracleInteger(sqltypes.Integer):
+    def get_dbapi_type(self, dbapi):
+        # see https://github.com/oracle/python-cx_Oracle/issues/
+        # 208#issuecomment-409715955
+        return int
+
+    def _cx_oracle_var(self, dialect, cursor, arraysize=None):
+        cx_Oracle = dialect.dbapi
+        return cursor.var(
+            cx_Oracle.STRING,
+            255,
+            arraysize=arraysize if arraysize is not None else cursor.arraysize,
+            outconverter=int,
+        )
+
+    def _cx_oracle_outputtypehandler(self, dialect):
+        def handler(cursor, name, default_type, size, precision, scale):
+            return self._cx_oracle_var(dialect, cursor)
+
+        return handler
+
+
+class _OracleNumeric(sqltypes.Numeric):
+    is_number = False
+
+    def bind_processor(self, dialect):
+        if self.scale == 0:
+            return None
+        elif self.asdecimal:
+            processor = processors.to_decimal_processor_factory(
+                decimal.Decimal, self._effective_decimal_return_scale
+            )
+
+            def process(value):
+                if isinstance(value, (int, float)):
+                    return processor(value)
+                elif value is not None and value.is_infinite():
+                    return float(value)
+                else:
+                    return value
+
+            return process
+        else:
+            return processors.to_float
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+    def _cx_oracle_outputtypehandler(self, dialect):
+        cx_Oracle = dialect.dbapi
+
+        def handler(cursor, name, default_type, size, precision, scale):
+            outconverter = None
+
+            if precision:
+                if self.asdecimal:
+                    if default_type == cx_Oracle.NATIVE_FLOAT:
+                        # receiving float and doing Decimal after the fact
+                        # allows for float("inf") to be handled
+                        type_ = default_type
+                        outconverter = decimal.Decimal
+                    else:
+                        type_ = decimal.Decimal
+                else:
+                    if self.is_number and scale == 0:
+                        # integer. cx_Oracle is observed to handle the widest
+                        # variety of ints when no directives are passed,
+                        # from 5.2 to 7.0.  See [ticket:4457]
+                        return None
+                    else:
+                        type_ = cx_Oracle.NATIVE_FLOAT
+
+            else:
+                if self.asdecimal:
+                    if default_type == cx_Oracle.NATIVE_FLOAT:
+                        type_ = default_type
+                        outconverter = decimal.Decimal
+                    else:
+                        type_ = decimal.Decimal
+                else:
+                    if self.is_number and scale == 0:
+                        # integer. cx_Oracle is observed to handle the widest
+                        # variety of ints when no directives are passed,
+                        # from 5.2 to 7.0.  See [ticket:4457]
+                        return None
+                    else:
+                        type_ = cx_Oracle.NATIVE_FLOAT
+
+            return cursor.var(
+                type_,
+                255,
+                arraysize=cursor.arraysize,
+                outconverter=outconverter,
+            )
+
+        return handler
+
+
+class _OracleUUID(sqltypes.Uuid):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.STRING
+
+
+class _OracleBinaryFloat(_OracleNumeric):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.NATIVE_FLOAT
+
+
+class _OracleBINARY_FLOAT(_OracleBinaryFloat, oracle.BINARY_FLOAT):
+    pass
+
+
+class _OracleBINARY_DOUBLE(_OracleBinaryFloat, oracle.BINARY_DOUBLE):
+    pass
+
+
+class _OracleNUMBER(_OracleNumeric):
+    is_number = True
+
+
+class _CXOracleDate(oracle._OracleDate):
+    def bind_processor(self, dialect):
+        return None
+
+    def result_processor(self, dialect, coltype):
+        def process(value):
+            if value is not None:
+                return value.date()
+            else:
+                return value
+
+        return process
+
+
+class _CXOracleTIMESTAMP(_OracleDateLiteralRender, sqltypes.TIMESTAMP):
+    def literal_processor(self, dialect):
+        return self._literal_processor_datetime(dialect)
+
+
+class _LOBDataType:
+    pass
+
+
+# TODO: the names used across CHAR / VARCHAR / NCHAR / NVARCHAR
+# here are inconsistent and not very good
+class _OracleChar(sqltypes.CHAR):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.FIXED_CHAR
+
+
+class _OracleNChar(sqltypes.NCHAR):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.FIXED_NCHAR
+
+
+class _OracleUnicodeStringNCHAR(oracle.NVARCHAR2):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.NCHAR
+
+
+class _OracleUnicodeStringCHAR(sqltypes.Unicode):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.LONG_STRING
+
+
+class _OracleUnicodeTextNCLOB(_LOBDataType, oracle.NCLOB):
+    def get_dbapi_type(self, dbapi):
+        # previously, this was dbapi.NCLOB.
+        # DB_TYPE_NVARCHAR will instead be passed to setinputsizes()
+        # when this datatype is used.
+        return dbapi.DB_TYPE_NVARCHAR
+
+
+class _OracleUnicodeTextCLOB(_LOBDataType, sqltypes.UnicodeText):
+    def get_dbapi_type(self, dbapi):
+        # previously, this was dbapi.CLOB.
+        # DB_TYPE_NVARCHAR will instead be passed to setinputsizes()
+        # when this datatype is used.
+        return dbapi.DB_TYPE_NVARCHAR
+
+
+class _OracleText(_LOBDataType, sqltypes.Text):
+    def get_dbapi_type(self, dbapi):
+        # previously, this was dbapi.CLOB.
+        # DB_TYPE_NVARCHAR will instead be passed to setinputsizes()
+        # when this datatype is used.
+        return dbapi.DB_TYPE_NVARCHAR
+
+
+class _OracleLong(_LOBDataType, oracle.LONG):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.LONG_STRING
+
+
+class _OracleString(sqltypes.String):
+    pass
+
+
+class _OracleEnum(sqltypes.Enum):
+    def bind_processor(self, dialect):
+        enum_proc = sqltypes.Enum.bind_processor(self, dialect)
+
+        def process(value):
+            raw_str = enum_proc(value)
+            return raw_str
+
+        return process
+
+
+class _OracleBinary(_LOBDataType, sqltypes.LargeBinary):
+    def get_dbapi_type(self, dbapi):
+        # previously, this was dbapi.BLOB.
+        # DB_TYPE_RAW will instead be passed to setinputsizes()
+        # when this datatype is used.
+        return dbapi.DB_TYPE_RAW
+
+    def bind_processor(self, dialect):
+        return None
+
+    def result_processor(self, dialect, coltype):
+        if not dialect.auto_convert_lobs:
+            return None
+        else:
+            return super().result_processor(dialect, coltype)
+
+
+class _OracleInterval(oracle.INTERVAL):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.INTERVAL
+
+
+class _OracleRaw(oracle.RAW):
+    pass
+
+
+class _OracleRowid(oracle.ROWID):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.ROWID
+
+
+class OracleCompiler_cx_oracle(OracleCompiler):
+    _oracle_cx_sql_compiler = True
+
+    _oracle_returning = False
+
+    # Oracle bind names can't start with digits or underscores.
+    # currently we rely upon Oracle-specific quoting of bind names in most
+    # cases.  however for expanding params, the escape chars are used.
+    # see #8708
+    bindname_escape_characters = util.immutabledict(
+        {
+            "%": "P",
+            "(": "A",
+            ")": "Z",
+            ":": "C",
+            ".": "C",
+            "[": "C",
+            "]": "C",
+            " ": "C",
+            "\\": "C",
+            "/": "C",
+            "?": "C",
+        }
+    )
+
+    def bindparam_string(self, name, **kw):
+        quote = getattr(name, "quote", None)
+        if (
+            quote is True
+            or quote is not False
+            and self.preparer._bindparam_requires_quotes(name)
+            # bind param quoting for Oracle doesn't work with post_compile
+            # params.  For those, the default bindparam_string will escape
+            # special chars, and the appending of a number "_1" etc. will
+            # take care of reserved words
+            and not kw.get("post_compile", False)
+        ):
+            # interesting to note about expanding parameters - since the
+            # new parameters take the form <paramname>_<int>, at least if
+            # they are originally formed from reserved words, they no longer
+            # need quoting :).    names that include illegal characters
+            # won't work however.
+            quoted_name = '"%s"' % name
+            kw["escaped_from"] = name
+            name = quoted_name
+            return OracleCompiler.bindparam_string(self, name, **kw)
+
+        # TODO: we could likely do away with quoting altogether for
+        # Oracle parameters and use the custom escaping here
+        escaped_from = kw.get("escaped_from", None)
+        if not escaped_from:
+            if self._bind_translate_re.search(name):
+                # not quite the translate use case as we want to
+                # also get a quick boolean if we even found
+                # unusual characters in the name
+                new_name = self._bind_translate_re.sub(
+                    lambda m: self._bind_translate_chars[m.group(0)],
+                    name,
+                )
+                if new_name[0].isdigit() or new_name[0] == "_":
+                    new_name = "D" + new_name
+                kw["escaped_from"] = name
+                name = new_name
+            elif name[0].isdigit() or name[0] == "_":
+                new_name = "D" + name
+                kw["escaped_from"] = name
+                name = new_name
+
+        return OracleCompiler.bindparam_string(self, name, **kw)
+
+
+class OracleExecutionContext_cx_oracle(OracleExecutionContext):
+    out_parameters = None
+
+    def _generate_out_parameter_vars(self):
+        # check for has_out_parameters or RETURNING, create cx_Oracle.var
+        # objects if so
+        if self.compiled.has_out_parameters or self.compiled._oracle_returning:
+            out_parameters = self.out_parameters
+            assert out_parameters is not None
+
+            len_params = len(self.parameters)
+
+            quoted_bind_names = self.compiled.escaped_bind_names
+            for bindparam in self.compiled.binds.values():
+                if bindparam.isoutparam:
+                    name = self.compiled.bind_names[bindparam]
+                    type_impl = bindparam.type.dialect_impl(self.dialect)
+
+                    if hasattr(type_impl, "_cx_oracle_var"):
+                        out_parameters[name] = type_impl._cx_oracle_var(
+                            self.dialect, self.cursor, arraysize=len_params
+                        )
+                    else:
+                        dbtype = type_impl.get_dbapi_type(self.dialect.dbapi)
+
+                        cx_Oracle = self.dialect.dbapi
+
+                        assert cx_Oracle is not None
+
+                        if dbtype is None:
+                            raise exc.InvalidRequestError(
+                                "Cannot create out parameter for "
+                                "parameter "
+                                "%r - its type %r is not supported by"
+                                " cx_oracle" % (bindparam.key, bindparam.type)
+                            )
+
+                        # note this is an OUT parameter.   Using
+                        # non-LOB datavalues with large unicode-holding
+                        # values causes the failure (both cx_Oracle and
+                        # oracledb):
+                        # ORA-22835: Buffer too small for CLOB to CHAR or
+                        # BLOB to RAW conversion (actual: 16507,
+                        # maximum: 4000)
+                        # [SQL: INSERT INTO long_text (x, y, z) VALUES
+                        # (:x, :y, :z) RETURNING long_text.x, long_text.y,
+                        # long_text.z INTO :ret_0, :ret_1, :ret_2]
+                        # so even for DB_TYPE_NVARCHAR we convert to a LOB
+
+                        if isinstance(type_impl, _LOBDataType):
+                            if dbtype == cx_Oracle.DB_TYPE_NVARCHAR:
+                                dbtype = cx_Oracle.NCLOB
+                            elif dbtype == cx_Oracle.DB_TYPE_RAW:
+                                dbtype = cx_Oracle.BLOB
+                            # other LOB types go in directly
+
+                            out_parameters[name] = self.cursor.var(
+                                dbtype,
+                                # this is fine also in oracledb_async since
+                                # the driver will await the read coroutine
+                                outconverter=lambda value: value.read(),
+                                arraysize=len_params,
+                            )
+                        elif (
+                            isinstance(type_impl, _OracleNumeric)
+                            and type_impl.asdecimal
+                        ):
+                            out_parameters[name] = self.cursor.var(
+                                decimal.Decimal,
+                                arraysize=len_params,
+                            )
+
+                        else:
+                            out_parameters[name] = self.cursor.var(
+                                dbtype, arraysize=len_params
+                            )
+
+                    for param in self.parameters:
+                        param[quoted_bind_names.get(name, name)] = (
+                            out_parameters[name]
+                        )
+
+    def _generate_cursor_outputtype_handler(self):
+        output_handlers = {}
+
+        for keyname, name, objects, type_ in self.compiled._result_columns:
+            handler = type_._cached_custom_processor(
+                self.dialect,
+                "cx_oracle_outputtypehandler",
+                self._get_cx_oracle_type_handler,
+            )
+
+            if handler:
+                denormalized_name = self.dialect.denormalize_name(keyname)
+                output_handlers[denormalized_name] = handler
+
+        if output_handlers:
+            default_handler = self._dbapi_connection.outputtypehandler
+
+            def output_type_handler(
+                cursor, name, default_type, size, precision, scale
+            ):
+                if name in output_handlers:
+                    return output_handlers[name](
+                        cursor, name, default_type, size, precision, scale
+                    )
+                else:
+                    return default_handler(
+                        cursor, name, default_type, size, precision, scale
+                    )
+
+            self.cursor.outputtypehandler = output_type_handler
+
+    def _get_cx_oracle_type_handler(self, impl):
+        if hasattr(impl, "_cx_oracle_outputtypehandler"):
+            return impl._cx_oracle_outputtypehandler(self.dialect)
+        else:
+            return None
+
+    def pre_exec(self):
+        super().pre_exec()
+        if not getattr(self.compiled, "_oracle_cx_sql_compiler", False):
+            return
+
+        self.out_parameters = {}
+
+        self._generate_out_parameter_vars()
+
+        self._generate_cursor_outputtype_handler()
+
+    def post_exec(self):
+        if (
+            self.compiled
+            and is_sql_compiler(self.compiled)
+            and self.compiled._oracle_returning
+        ):
+            initial_buffer = self.fetchall_for_returning(
+                self.cursor, _internal=True
+            )
+
+            fetch_strategy = _cursor.FullyBufferedCursorFetchStrategy(
+                self.cursor,
+                [
+                    (entry.keyname, None)
+                    for entry in self.compiled._result_columns
+                ],
+                initial_buffer=initial_buffer,
+            )
+
+            self.cursor_fetch_strategy = fetch_strategy
+
+    def create_cursor(self):
+        c = self._dbapi_connection.cursor()
+        if self.dialect.arraysize:
+            c.arraysize = self.dialect.arraysize
+
+        return c
+
+    def fetchall_for_returning(self, cursor, *, _internal=False):
+        compiled = self.compiled
+        if (
+            not _internal
+            and compiled is None
+            or not is_sql_compiler(compiled)
+            or not compiled._oracle_returning
+        ):
+            raise NotImplementedError(
+                "execution context was not prepared for Oracle RETURNING"
+            )
+
+        # create a fake cursor result from the out parameters. unlike
+        # get_out_parameter_values(), the result-row handlers here will be
+        # applied at the Result level
+
+        numcols = len(self.out_parameters)
+
+        # [stmt_result for stmt_result in outparam.values] == each
+        # statement in executemany
+        # [val for val in stmt_result] == each row for a particular
+        # statement
+        return list(
+            zip(
+                *[
+                    [
+                        val
+                        for stmt_result in self.out_parameters[
+                            f"ret_{j}"
+                        ].values
+                        for val in (stmt_result or ())
+                    ]
+                    for j in range(numcols)
+                ]
+            )
+        )
+
+    def get_out_parameter_values(self, out_param_names):
+        # this method should not be called when the compiler has
+        # RETURNING as we've turned the has_out_parameters flag set to
+        # False.
+        assert not self.compiled.returning
+
+        return [
+            self.dialect._paramval(self.out_parameters[name])
+            for name in out_param_names
+        ]
+
+
+class OracleDialect_cx_oracle(OracleDialect):
+    supports_statement_cache = True
+    execution_ctx_cls = OracleExecutionContext_cx_oracle
+    statement_compiler = OracleCompiler_cx_oracle
+
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = True
+
+    insert_executemany_returning = True
+    insert_executemany_returning_sort_by_parameter_order = True
+    update_executemany_returning = True
+    delete_executemany_returning = True
+
+    bind_typing = interfaces.BindTyping.SETINPUTSIZES
+
+    driver = "cx_oracle"
+
+    colspecs = util.update_copy(
+        OracleDialect.colspecs,
+        {
+            sqltypes.TIMESTAMP: _CXOracleTIMESTAMP,
+            sqltypes.Numeric: _OracleNumeric,
+            sqltypes.Float: _OracleNumeric,
+            oracle.BINARY_FLOAT: _OracleBINARY_FLOAT,
+            oracle.BINARY_DOUBLE: _OracleBINARY_DOUBLE,
+            sqltypes.Integer: _OracleInteger,
+            oracle.NUMBER: _OracleNUMBER,
+            sqltypes.Date: _CXOracleDate,
+            sqltypes.LargeBinary: _OracleBinary,
+            sqltypes.Boolean: oracle._OracleBoolean,
+            sqltypes.Interval: _OracleInterval,
+            oracle.INTERVAL: _OracleInterval,
+            sqltypes.Text: _OracleText,
+            sqltypes.String: _OracleString,
+            sqltypes.UnicodeText: _OracleUnicodeTextCLOB,
+            sqltypes.CHAR: _OracleChar,
+            sqltypes.NCHAR: _OracleNChar,
+            sqltypes.Enum: _OracleEnum,
+            oracle.LONG: _OracleLong,
+            oracle.RAW: _OracleRaw,
+            sqltypes.Unicode: _OracleUnicodeStringCHAR,
+            sqltypes.NVARCHAR: _OracleUnicodeStringNCHAR,
+            sqltypes.Uuid: _OracleUUID,
+            oracle.NCLOB: _OracleUnicodeTextNCLOB,
+            oracle.ROWID: _OracleRowid,
+        },
+    )
+
+    execute_sequence_format = list
+
+    _cx_oracle_threaded = None
+
+    _cursor_var_unicode_kwargs = util.immutabledict()
+
+    @util.deprecated_params(
+        threaded=(
+            "1.3",
+            "The 'threaded' parameter to the cx_oracle/oracledb dialect "
+            "is deprecated as a dialect-level argument, and will be removed "
+            "in a future release.  As of version 1.3, it defaults to False "
+            "rather than True.  The 'threaded' option can be passed to "
+            "cx_Oracle directly in the URL query string passed to "
+            ":func:`_sa.create_engine`.",
+        )
+    )
+    def __init__(
+        self,
+        auto_convert_lobs=True,
+        coerce_to_decimal=True,
+        arraysize=None,
+        encoding_errors=None,
+        threaded=None,
+        **kwargs,
+    ):
+        OracleDialect.__init__(self, **kwargs)
+        self.arraysize = arraysize
+        self.encoding_errors = encoding_errors
+        if encoding_errors:
+            self._cursor_var_unicode_kwargs = {
+                "encodingErrors": encoding_errors
+            }
+        if threaded is not None:
+            self._cx_oracle_threaded = threaded
+        self.auto_convert_lobs = auto_convert_lobs
+        self.coerce_to_decimal = coerce_to_decimal
+        if self._use_nchar_for_unicode:
+            self.colspecs = self.colspecs.copy()
+            self.colspecs[sqltypes.Unicode] = _OracleUnicodeStringNCHAR
+            self.colspecs[sqltypes.UnicodeText] = _OracleUnicodeTextNCLOB
+
+        dbapi_module = self.dbapi
+        self._load_version(dbapi_module)
+
+        if dbapi_module is not None:
+            # these constants will first be seen in SQLAlchemy datatypes
+            # coming from the get_dbapi_type() method.   We then
+            # will place the following types into setinputsizes() calls
+            # on each statement.  Oracle constants that are not in this
+            # list will not be put into setinputsizes().
+            self.include_set_input_sizes = {
+                dbapi_module.DATETIME,
+                dbapi_module.DB_TYPE_NVARCHAR,  # used for CLOB, NCLOB
+                dbapi_module.DB_TYPE_RAW,  # used for BLOB
+                dbapi_module.NCLOB,  # not currently used except for OUT param
+                dbapi_module.CLOB,  # not currently used except for OUT param
+                dbapi_module.LOB,  # not currently used
+                dbapi_module.BLOB,  # not currently used except for OUT param
+                dbapi_module.NCHAR,
+                dbapi_module.FIXED_NCHAR,
+                dbapi_module.FIXED_CHAR,
+                dbapi_module.TIMESTAMP,
+                int,  # _OracleInteger,
+                # _OracleBINARY_FLOAT, _OracleBINARY_DOUBLE,
+                dbapi_module.NATIVE_FLOAT,
+            }
+
+            self._paramval = lambda value: value.getvalue()
+
+    def _load_version(self, dbapi_module):
+        version = (0, 0, 0)
+        if dbapi_module is not None:
+            m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", dbapi_module.version)
+            if m:
+                version = tuple(
+                    int(x) for x in m.group(1, 2, 3) if x is not None
+                )
+        self.cx_oracle_ver = version
+        if self.cx_oracle_ver < (8,) and self.cx_oracle_ver > (0, 0, 0):
+            raise exc.InvalidRequestError(
+                "cx_Oracle version 8 and above are supported"
+            )
+
+    @classmethod
+    def import_dbapi(cls):
+        import cx_Oracle
+
+        return cx_Oracle
+
+    def initialize(self, connection):
+        super().initialize(connection)
+        self._detect_decimal_char(connection)
+
+    def get_isolation_level(self, dbapi_connection):
+        # sources:
+
+        # general idea of transaction id, have to start one, etc.
+        # https://stackoverflow.com/questions/10711204/how-to-check-isoloation-level
+
+        # how to decode xid cols from v$transaction to match
+        # https://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:9532779900346079444
+
+        # Oracle tuple comparison without using IN:
+        # https://www.sql-workbench.eu/comparison/tuple_comparison.html
+
+        with dbapi_connection.cursor() as cursor:
+            # this is the only way to ensure a transaction is started without
+            # actually running DML.   There's no way to see the configured
+            # isolation level without getting it from v$transaction which
+            # means transaction has to be started.
+            outval = cursor.var(str)
+            cursor.execute(
+                """
+                begin
+                   :trans_id := dbms_transaction.local_transaction_id( TRUE );
+                end;
+                """,
+                {"trans_id": outval},
+            )
+            trans_id = outval.getvalue()
+            xidusn, xidslot, xidsqn = trans_id.split(".", 2)
+
+            cursor.execute(
+                "SELECT CASE BITAND(t.flag, POWER(2, 28)) "
+                "WHEN 0 THEN 'READ COMMITTED' "
+                "ELSE 'SERIALIZABLE' END AS isolation_level "
+                "FROM v$transaction t WHERE "
+                "(t.xidusn, t.xidslot, t.xidsqn) = "
+                "((:xidusn, :xidslot, :xidsqn))",
+                {"xidusn": xidusn, "xidslot": xidslot, "xidsqn": xidsqn},
+            )
+            row = cursor.fetchone()
+            if row is None:
+                raise exc.InvalidRequestError(
+                    "could not retrieve isolation level"
+                )
+            result = row[0]
+
+        return result
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return super().get_isolation_level_values(dbapi_connection) + [
+            "AUTOCOMMIT"
+        ]
+
+    def set_isolation_level(self, dbapi_connection, level):
+        if level == "AUTOCOMMIT":
+            dbapi_connection.autocommit = True
+        else:
+            dbapi_connection.autocommit = False
+            dbapi_connection.rollback()
+            with dbapi_connection.cursor() as cursor:
+                cursor.execute(f"ALTER SESSION SET ISOLATION_LEVEL={level}")
+
+    def _detect_decimal_char(self, connection):
+        # we have the option to change this setting upon connect,
+        # or just look at what it is upon connect and convert.
+        # to minimize the chance of interference with changes to
+        # NLS_TERRITORY or formatting behavior of the DB, we opt
+        # to just look at it
+
+        dbapi_connection = connection.connection
+
+        with dbapi_connection.cursor() as cursor:
+            # issue #8744
+            # nls_session_parameters is not available in some Oracle
+            # modes like "mount mode".  But then, v$nls_parameters is not
+            # available if the connection doesn't have SYSDBA priv.
+            #
+            # simplify the whole thing and just use the method that we were
+            # doing in the test suite already, selecting a number
+
+            def output_type_handler(
+                cursor, name, defaultType, size, precision, scale
+            ):
+                return cursor.var(
+                    self.dbapi.STRING, 255, arraysize=cursor.arraysize
+                )
+
+            cursor.outputtypehandler = output_type_handler
+            cursor.execute("SELECT 1.1 FROM DUAL")
+            value = cursor.fetchone()[0]
+
+            decimal_char = value.lstrip("0")[1]
+            assert not decimal_char[0].isdigit()
+
+        self._decimal_char = decimal_char
+
+        if self._decimal_char != ".":
+            _detect_decimal = self._detect_decimal
+            _to_decimal = self._to_decimal
+
+            self._detect_decimal = lambda value: _detect_decimal(
+                value.replace(self._decimal_char, ".")
+            )
+            self._to_decimal = lambda value: _to_decimal(
+                value.replace(self._decimal_char, ".")
+            )
+
+    def _detect_decimal(self, value):
+        if "." in value:
+            return self._to_decimal(value)
+        else:
+            return int(value)
+
+    _to_decimal = decimal.Decimal
+
+    def _generate_connection_outputtype_handler(self):
+        """establish the default outputtypehandler established at the
+        connection level.
+
+        """
+
+        dialect = self
+        cx_Oracle = dialect.dbapi
+
+        number_handler = _OracleNUMBER(
+            asdecimal=True
+        )._cx_oracle_outputtypehandler(dialect)
+        float_handler = _OracleNUMBER(
+            asdecimal=False
+        )._cx_oracle_outputtypehandler(dialect)
+
+        def output_type_handler(
+            cursor, name, default_type, size, precision, scale
+        ):
+            if (
+                default_type == cx_Oracle.NUMBER
+                and default_type is not cx_Oracle.NATIVE_FLOAT
+            ):
+                if not dialect.coerce_to_decimal:
+                    return None
+                elif precision == 0 and scale in (0, -127):
+                    # ambiguous type, this occurs when selecting
+                    # numbers from deep subqueries
+                    return cursor.var(
+                        cx_Oracle.STRING,
+                        255,
+                        outconverter=dialect._detect_decimal,
+                        arraysize=cursor.arraysize,
+                    )
+                elif precision and scale > 0:
+                    return number_handler(
+                        cursor, name, default_type, size, precision, scale
+                    )
+                else:
+                    return float_handler(
+                        cursor, name, default_type, size, precision, scale
+                    )
+
+            # if unicode options were specified, add a decoder, otherwise
+            # cx_Oracle should return Unicode
+            elif (
+                dialect._cursor_var_unicode_kwargs
+                and default_type
+                in (
+                    cx_Oracle.STRING,
+                    cx_Oracle.FIXED_CHAR,
+                )
+                and default_type is not cx_Oracle.CLOB
+                and default_type is not cx_Oracle.NCLOB
+            ):
+                return cursor.var(
+                    str,
+                    size,
+                    cursor.arraysize,
+                    **dialect._cursor_var_unicode_kwargs,
+                )
+
+            elif dialect.auto_convert_lobs and default_type in (
+                cx_Oracle.CLOB,
+                cx_Oracle.NCLOB,
+            ):
+                typ = (
+                    cx_Oracle.DB_TYPE_VARCHAR
+                    if default_type is cx_Oracle.CLOB
+                    else cx_Oracle.DB_TYPE_NVARCHAR
+                )
+                return cursor.var(
+                    typ,
+                    _CX_ORACLE_MAGIC_LOB_SIZE,
+                    cursor.arraysize,
+                    **dialect._cursor_var_unicode_kwargs,
+                )
+
+            elif dialect.auto_convert_lobs and default_type in (
+                cx_Oracle.BLOB,
+            ):
+                return cursor.var(
+                    cx_Oracle.DB_TYPE_RAW,
+                    _CX_ORACLE_MAGIC_LOB_SIZE,
+                    cursor.arraysize,
+                )
+
+        return output_type_handler
+
+    def on_connect(self):
+        output_type_handler = self._generate_connection_outputtype_handler()
+
+        def on_connect(conn):
+            conn.outputtypehandler = output_type_handler
+
+        return on_connect
+
+    def create_connect_args(self, url):
+        opts = dict(url.query)
+
+        for opt in ("use_ansi", "auto_convert_lobs"):
+            if opt in opts:
+                util.warn_deprecated(
+                    f"{self.driver} dialect option {opt!r} should only be "
+                    "passed to create_engine directly, not within the URL "
+                    "string",
+                    version="1.3",
+                )
+                util.coerce_kw_type(opts, opt, bool)
+                setattr(self, opt, opts.pop(opt))
+
+        database = url.database
+        service_name = opts.pop("service_name", None)
+        if database or service_name:
+            # if we have a database, then we have a remote host
+            port = url.port
+            if port:
+                port = int(port)
+            else:
+                port = 1521
+
+            if database and service_name:
+                raise exc.InvalidRequestError(
+                    '"service_name" option shouldn\'t '
+                    'be used with a "database" part of the url'
+                )
+            if database:
+                makedsn_kwargs = {"sid": database}
+            if service_name:
+                makedsn_kwargs = {"service_name": service_name}
+
+            dsn = self.dbapi.makedsn(url.host, port, **makedsn_kwargs)
+        else:
+            # we have a local tnsname
+            dsn = url.host
+
+        if dsn is not None:
+            opts["dsn"] = dsn
+        if url.password is not None:
+            opts["password"] = url.password
+        if url.username is not None:
+            opts["user"] = url.username
+
+        if self._cx_oracle_threaded is not None:
+            opts.setdefault("threaded", self._cx_oracle_threaded)
+
+        def convert_cx_oracle_constant(value):
+            if isinstance(value, str):
+                try:
+                    int_val = int(value)
+                except ValueError:
+                    value = value.upper()
+                    return getattr(self.dbapi, value)
+                else:
+                    return int_val
+            else:
+                return value
+
+        util.coerce_kw_type(opts, "mode", convert_cx_oracle_constant)
+        util.coerce_kw_type(opts, "threaded", bool)
+        util.coerce_kw_type(opts, "events", bool)
+        util.coerce_kw_type(opts, "purity", convert_cx_oracle_constant)
+        return ([], opts)
+
+    def _get_server_version_info(self, connection):
+        return tuple(int(x) for x in connection.connection.version.split("."))
+
+    def is_disconnect(self, e, connection, cursor):
+        (error,) = e.args
+        if isinstance(
+            e, (self.dbapi.InterfaceError, self.dbapi.DatabaseError)
+        ) and "not connected" in str(e):
+            return True
+
+        if hasattr(error, "code") and error.code in {
+            28,
+            3114,
+            3113,
+            3135,
+            1033,
+            2396,
+        }:
+            # ORA-00028: your session has been killed
+            # ORA-03114: not connected to ORACLE
+            # ORA-03113: end-of-file on communication channel
+            # ORA-03135: connection lost contact
+            # ORA-01033: ORACLE initialization or shutdown in progress
+            # ORA-02396: exceeded maximum idle time, please connect again
+            # TODO: Others ?
+            return True
+
+        if re.match(r"^(?:DPI-1010|DPI-1080|DPY-1001|DPY-4011)", str(e)):
+            # DPI-1010: not connected
+            # DPI-1080: connection was closed by ORA-3113
+            # python-oracledb's DPY-1001: not connected to database
+            # python-oracledb's DPY-4011: the database or network closed the
+            # connection
+            # TODO: others?
+            return True
+
+        return False
+
+    def create_xid(self):
+        id_ = random.randint(0, 2**128)
+        return (0x1234, "%032x" % id_, "%032x" % 9)
+
+    def do_executemany(self, cursor, statement, parameters, context=None):
+        if isinstance(parameters, tuple):
+            parameters = list(parameters)
+        cursor.executemany(statement, parameters)
+
+    def do_begin_twophase(self, connection, xid):
+        connection.connection.begin(*xid)
+        connection.connection.info["cx_oracle_xid"] = xid
+
+    def do_prepare_twophase(self, connection, xid):
+        result = connection.connection.prepare()
+        connection.info["cx_oracle_prepared"] = result
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        self.do_rollback(connection.connection)
+        # TODO: need to end XA state here
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if not is_prepared:
+            self.do_commit(connection.connection)
+        else:
+            if recover:
+                raise NotImplementedError(
+                    "2pc recovery not implemented for cx_Oracle"
+                )
+            oci_prepared = connection.info["cx_oracle_prepared"]
+            if oci_prepared:
+                self.do_commit(connection.connection)
+        # TODO: need to end XA state here
+
+    def do_set_input_sizes(self, cursor, list_of_tuples, context):
+        if self.positional:
+            # not usually used, here to support if someone is modifying
+            # the dialect to use positional style
+            cursor.setinputsizes(
+                *[dbtype for key, dbtype, sqltype in list_of_tuples]
+            )
+        else:
+            collection = (
+                (key, dbtype)
+                for key, dbtype, sqltype in list_of_tuples
+                if dbtype
+            )
+
+            cursor.setinputsizes(**{key: dbtype for key, dbtype in collection})
+
+    def do_recover_twophase(self, connection):
+        raise NotImplementedError(
+            "recover two phase query for cx_Oracle not implemented"
+        )
+
+
+dialect = OracleDialect_cx_oracle
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/dictionary.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/dictionary.py
new file mode 100644
index 00000000..f785a66e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/dictionary.py
@@ -0,0 +1,507 @@
+# dialects/oracle/dictionary.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .types import DATE
+from .types import LONG
+from .types import NUMBER
+from .types import RAW
+from .types import VARCHAR2
+from ... import Column
+from ... import MetaData
+from ... import Table
+from ... import table
+from ...sql.sqltypes import CHAR
+
+# constants
+DB_LINK_PLACEHOLDER = "__$sa_dblink$__"
+# tables
+dual = table("dual")
+dictionary_meta = MetaData()
+
+# NOTE: all the dictionary_meta are aliases because oracle does not like
+# using the full table@dblink for every column in query, and complains with
+# ORA-00960: ambiguous column naming in select list
+all_tables = Table(
+    "all_tables" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("tablespace_name", VARCHAR2(30)),
+    Column("cluster_name", VARCHAR2(128)),
+    Column("iot_name", VARCHAR2(128)),
+    Column("status", VARCHAR2(8)),
+    Column("pct_free", NUMBER),
+    Column("pct_used", NUMBER),
+    Column("ini_trans", NUMBER),
+    Column("max_trans", NUMBER),
+    Column("initial_extent", NUMBER),
+    Column("next_extent", NUMBER),
+    Column("min_extents", NUMBER),
+    Column("max_extents", NUMBER),
+    Column("pct_increase", NUMBER),
+    Column("freelists", NUMBER),
+    Column("freelist_groups", NUMBER),
+    Column("logging", VARCHAR2(3)),
+    Column("backed_up", VARCHAR2(1)),
+    Column("num_rows", NUMBER),
+    Column("blocks", NUMBER),
+    Column("empty_blocks", NUMBER),
+    Column("avg_space", NUMBER),
+    Column("chain_cnt", NUMBER),
+    Column("avg_row_len", NUMBER),
+    Column("avg_space_freelist_blocks", NUMBER),
+    Column("num_freelist_blocks", NUMBER),
+    Column("degree", VARCHAR2(10)),
+    Column("instances", VARCHAR2(10)),
+    Column("cache", VARCHAR2(5)),
+    Column("table_lock", VARCHAR2(8)),
+    Column("sample_size", NUMBER),
+    Column("last_analyzed", DATE),
+    Column("partitioned", VARCHAR2(3)),
+    Column("iot_type", VARCHAR2(12)),
+    Column("temporary", VARCHAR2(1)),
+    Column("secondary", VARCHAR2(1)),
+    Column("nested", VARCHAR2(3)),
+    Column("buffer_pool", VARCHAR2(7)),
+    Column("flash_cache", VARCHAR2(7)),
+    Column("cell_flash_cache", VARCHAR2(7)),
+    Column("row_movement", VARCHAR2(8)),
+    Column("global_stats", VARCHAR2(3)),
+    Column("user_stats", VARCHAR2(3)),
+    Column("duration", VARCHAR2(15)),
+    Column("skip_corrupt", VARCHAR2(8)),
+    Column("monitoring", VARCHAR2(3)),
+    Column("cluster_owner", VARCHAR2(128)),
+    Column("dependencies", VARCHAR2(8)),
+    Column("compression", VARCHAR2(8)),
+    Column("compress_for", VARCHAR2(30)),
+    Column("dropped", VARCHAR2(3)),
+    Column("read_only", VARCHAR2(3)),
+    Column("segment_created", VARCHAR2(3)),
+    Column("result_cache", VARCHAR2(7)),
+    Column("clustering", VARCHAR2(3)),
+    Column("activity_tracking", VARCHAR2(23)),
+    Column("dml_timestamp", VARCHAR2(25)),
+    Column("has_identity", VARCHAR2(3)),
+    Column("container_data", VARCHAR2(3)),
+    Column("inmemory", VARCHAR2(8)),
+    Column("inmemory_priority", VARCHAR2(8)),
+    Column("inmemory_distribute", VARCHAR2(15)),
+    Column("inmemory_compression", VARCHAR2(17)),
+    Column("inmemory_duplicate", VARCHAR2(13)),
+    Column("default_collation", VARCHAR2(100)),
+    Column("duplicated", VARCHAR2(1)),
+    Column("sharded", VARCHAR2(1)),
+    Column("externally_sharded", VARCHAR2(1)),
+    Column("externally_duplicated", VARCHAR2(1)),
+    Column("external", VARCHAR2(3)),
+    Column("hybrid", VARCHAR2(3)),
+    Column("cellmemory", VARCHAR2(24)),
+    Column("containers_default", VARCHAR2(3)),
+    Column("container_map", VARCHAR2(3)),
+    Column("extended_data_link", VARCHAR2(3)),
+    Column("extended_data_link_map", VARCHAR2(3)),
+    Column("inmemory_service", VARCHAR2(12)),
+    Column("inmemory_service_name", VARCHAR2(1000)),
+    Column("container_map_object", VARCHAR2(3)),
+    Column("memoptimize_read", VARCHAR2(8)),
+    Column("memoptimize_write", VARCHAR2(8)),
+    Column("has_sensitive_column", VARCHAR2(3)),
+    Column("admit_null", VARCHAR2(3)),
+    Column("data_link_dml_enabled", VARCHAR2(3)),
+    Column("logical_replication", VARCHAR2(8)),
+).alias("a_tables")
+
+all_views = Table(
+    "all_views" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("view_name", VARCHAR2(128), nullable=False),
+    Column("text_length", NUMBER),
+    Column("text", LONG),
+    Column("text_vc", VARCHAR2(4000)),
+    Column("type_text_length", NUMBER),
+    Column("type_text", VARCHAR2(4000)),
+    Column("oid_text_length", NUMBER),
+    Column("oid_text", VARCHAR2(4000)),
+    Column("view_type_owner", VARCHAR2(128)),
+    Column("view_type", VARCHAR2(128)),
+    Column("superview_name", VARCHAR2(128)),
+    Column("editioning_view", VARCHAR2(1)),
+    Column("read_only", VARCHAR2(1)),
+    Column("container_data", VARCHAR2(1)),
+    Column("bequeath", VARCHAR2(12)),
+    Column("origin_con_id", VARCHAR2(256)),
+    Column("default_collation", VARCHAR2(100)),
+    Column("containers_default", VARCHAR2(3)),
+    Column("container_map", VARCHAR2(3)),
+    Column("extended_data_link", VARCHAR2(3)),
+    Column("extended_data_link_map", VARCHAR2(3)),
+    Column("has_sensitive_column", VARCHAR2(3)),
+    Column("admit_null", VARCHAR2(3)),
+    Column("pdb_local_only", VARCHAR2(3)),
+).alias("a_views")
+
+all_sequences = Table(
+    "all_sequences" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("sequence_owner", VARCHAR2(128), nullable=False),
+    Column("sequence_name", VARCHAR2(128), nullable=False),
+    Column("min_value", NUMBER),
+    Column("max_value", NUMBER),
+    Column("increment_by", NUMBER, nullable=False),
+    Column("cycle_flag", VARCHAR2(1)),
+    Column("order_flag", VARCHAR2(1)),
+    Column("cache_size", NUMBER, nullable=False),
+    Column("last_number", NUMBER, nullable=False),
+    Column("scale_flag", VARCHAR2(1)),
+    Column("extend_flag", VARCHAR2(1)),
+    Column("sharded_flag", VARCHAR2(1)),
+    Column("session_flag", VARCHAR2(1)),
+    Column("keep_value", VARCHAR2(1)),
+).alias("a_sequences")
+
+all_users = Table(
+    "all_users" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("username", VARCHAR2(128), nullable=False),
+    Column("user_id", NUMBER, nullable=False),
+    Column("created", DATE, nullable=False),
+    Column("common", VARCHAR2(3)),
+    Column("oracle_maintained", VARCHAR2(1)),
+    Column("inherited", VARCHAR2(3)),
+    Column("default_collation", VARCHAR2(100)),
+    Column("implicit", VARCHAR2(3)),
+    Column("all_shard", VARCHAR2(3)),
+    Column("external_shard", VARCHAR2(3)),
+).alias("a_users")
+
+all_mviews = Table(
+    "all_mviews" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("mview_name", VARCHAR2(128), nullable=False),
+    Column("container_name", VARCHAR2(128), nullable=False),
+    Column("query", LONG),
+    Column("query_len", NUMBER(38)),
+    Column("updatable", VARCHAR2(1)),
+    Column("update_log", VARCHAR2(128)),
+    Column("master_rollback_seg", VARCHAR2(128)),
+    Column("master_link", VARCHAR2(128)),
+    Column("rewrite_enabled", VARCHAR2(1)),
+    Column("rewrite_capability", VARCHAR2(9)),
+    Column("refresh_mode", VARCHAR2(6)),
+    Column("refresh_method", VARCHAR2(8)),
+    Column("build_mode", VARCHAR2(9)),
+    Column("fast_refreshable", VARCHAR2(18)),
+    Column("last_refresh_type", VARCHAR2(8)),
+    Column("last_refresh_date", DATE),
+    Column("last_refresh_end_time", DATE),
+    Column("staleness", VARCHAR2(19)),
+    Column("after_fast_refresh", VARCHAR2(19)),
+    Column("unknown_prebuilt", VARCHAR2(1)),
+    Column("unknown_plsql_func", VARCHAR2(1)),
+    Column("unknown_external_table", VARCHAR2(1)),
+    Column("unknown_consider_fresh", VARCHAR2(1)),
+    Column("unknown_import", VARCHAR2(1)),
+    Column("unknown_trusted_fd", VARCHAR2(1)),
+    Column("compile_state", VARCHAR2(19)),
+    Column("use_no_index", VARCHAR2(1)),
+    Column("stale_since", DATE),
+    Column("num_pct_tables", NUMBER),
+    Column("num_fresh_pct_regions", NUMBER),
+    Column("num_stale_pct_regions", NUMBER),
+    Column("segment_created", VARCHAR2(3)),
+    Column("evaluation_edition", VARCHAR2(128)),
+    Column("unusable_before", VARCHAR2(128)),
+    Column("unusable_beginning", VARCHAR2(128)),
+    Column("default_collation", VARCHAR2(100)),
+    Column("on_query_computation", VARCHAR2(1)),
+    Column("auto", VARCHAR2(3)),
+).alias("a_mviews")
+
+all_tab_identity_cols = Table(
+    "all_tab_identity_cols" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_name", VARCHAR2(128), nullable=False),
+    Column("generation_type", VARCHAR2(10)),
+    Column("sequence_name", VARCHAR2(128), nullable=False),
+    Column("identity_options", VARCHAR2(298)),
+).alias("a_tab_identity_cols")
+
+all_tab_cols = Table(
+    "all_tab_cols" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_name", VARCHAR2(128), nullable=False),
+    Column("data_type", VARCHAR2(128)),
+    Column("data_type_mod", VARCHAR2(3)),
+    Column("data_type_owner", VARCHAR2(128)),
+    Column("data_length", NUMBER, nullable=False),
+    Column("data_precision", NUMBER),
+    Column("data_scale", NUMBER),
+    Column("nullable", VARCHAR2(1)),
+    Column("column_id", NUMBER),
+    Column("default_length", NUMBER),
+    Column("data_default", LONG),
+    Column("num_distinct", NUMBER),
+    Column("low_value", RAW(1000)),
+    Column("high_value", RAW(1000)),
+    Column("density", NUMBER),
+    Column("num_nulls", NUMBER),
+    Column("num_buckets", NUMBER),
+    Column("last_analyzed", DATE),
+    Column("sample_size", NUMBER),
+    Column("character_set_name", VARCHAR2(44)),
+    Column("char_col_decl_length", NUMBER),
+    Column("global_stats", VARCHAR2(3)),
+    Column("user_stats", VARCHAR2(3)),
+    Column("avg_col_len", NUMBER),
+    Column("char_length", NUMBER),
+    Column("char_used", VARCHAR2(1)),
+    Column("v80_fmt_image", VARCHAR2(3)),
+    Column("data_upgraded", VARCHAR2(3)),
+    Column("hidden_column", VARCHAR2(3)),
+    Column("virtual_column", VARCHAR2(3)),
+    Column("segment_column_id", NUMBER),
+    Column("internal_column_id", NUMBER, nullable=False),
+    Column("histogram", VARCHAR2(15)),
+    Column("qualified_col_name", VARCHAR2(4000)),
+    Column("user_generated", VARCHAR2(3)),
+    Column("default_on_null", VARCHAR2(3)),
+    Column("identity_column", VARCHAR2(3)),
+    Column("evaluation_edition", VARCHAR2(128)),
+    Column("unusable_before", VARCHAR2(128)),
+    Column("unusable_beginning", VARCHAR2(128)),
+    Column("collation", VARCHAR2(100)),
+    Column("collated_column_id", NUMBER),
+).alias("a_tab_cols")
+
+all_tab_comments = Table(
+    "all_tab_comments" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("table_type", VARCHAR2(11)),
+    Column("comments", VARCHAR2(4000)),
+    Column("origin_con_id", NUMBER),
+).alias("a_tab_comments")
+
+all_col_comments = Table(
+    "all_col_comments" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_name", VARCHAR2(128), nullable=False),
+    Column("comments", VARCHAR2(4000)),
+    Column("origin_con_id", NUMBER),
+).alias("a_col_comments")
+
+all_mview_comments = Table(
+    "all_mview_comments" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("mview_name", VARCHAR2(128), nullable=False),
+    Column("comments", VARCHAR2(4000)),
+).alias("a_mview_comments")
+
+all_ind_columns = Table(
+    "all_ind_columns" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("index_owner", VARCHAR2(128), nullable=False),
+    Column("index_name", VARCHAR2(128), nullable=False),
+    Column("table_owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_name", VARCHAR2(4000)),
+    Column("column_position", NUMBER, nullable=False),
+    Column("column_length", NUMBER, nullable=False),
+    Column("char_length", NUMBER),
+    Column("descend", VARCHAR2(4)),
+    Column("collated_column_id", NUMBER),
+).alias("a_ind_columns")
+
+all_indexes = Table(
+    "all_indexes" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("index_name", VARCHAR2(128), nullable=False),
+    Column("index_type", VARCHAR2(27)),
+    Column("table_owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("table_type", CHAR(11)),
+    Column("uniqueness", VARCHAR2(9)),
+    Column("compression", VARCHAR2(13)),
+    Column("prefix_length", NUMBER),
+    Column("tablespace_name", VARCHAR2(30)),
+    Column("ini_trans", NUMBER),
+    Column("max_trans", NUMBER),
+    Column("initial_extent", NUMBER),
+    Column("next_extent", NUMBER),
+    Column("min_extents", NUMBER),
+    Column("max_extents", NUMBER),
+    Column("pct_increase", NUMBER),
+    Column("pct_threshold", NUMBER),
+    Column("include_column", NUMBER),
+    Column("freelists", NUMBER),
+    Column("freelist_groups", NUMBER),
+    Column("pct_free", NUMBER),
+    Column("logging", VARCHAR2(3)),
+    Column("blevel", NUMBER),
+    Column("leaf_blocks", NUMBER),
+    Column("distinct_keys", NUMBER),
+    Column("avg_leaf_blocks_per_key", NUMBER),
+    Column("avg_data_blocks_per_key", NUMBER),
+    Column("clustering_factor", NUMBER),
+    Column("status", VARCHAR2(8)),
+    Column("num_rows", NUMBER),
+    Column("sample_size", NUMBER),
+    Column("last_analyzed", DATE),
+    Column("degree", VARCHAR2(40)),
+    Column("instances", VARCHAR2(40)),
+    Column("partitioned", VARCHAR2(3)),
+    Column("temporary", VARCHAR2(1)),
+    Column("generated", VARCHAR2(1)),
+    Column("secondary", VARCHAR2(1)),
+    Column("buffer_pool", VARCHAR2(7)),
+    Column("flash_cache", VARCHAR2(7)),
+    Column("cell_flash_cache", VARCHAR2(7)),
+    Column("user_stats", VARCHAR2(3)),
+    Column("duration", VARCHAR2(15)),
+    Column("pct_direct_access", NUMBER),
+    Column("ityp_owner", VARCHAR2(128)),
+    Column("ityp_name", VARCHAR2(128)),
+    Column("parameters", VARCHAR2(1000)),
+    Column("global_stats", VARCHAR2(3)),
+    Column("domidx_status", VARCHAR2(12)),
+    Column("domidx_opstatus", VARCHAR2(6)),
+    Column("funcidx_status", VARCHAR2(8)),
+    Column("join_index", VARCHAR2(3)),
+    Column("iot_redundant_pkey_elim", VARCHAR2(3)),
+    Column("dropped", VARCHAR2(3)),
+    Column("visibility", VARCHAR2(9)),
+    Column("domidx_management", VARCHAR2(14)),
+    Column("segment_created", VARCHAR2(3)),
+    Column("orphaned_entries", VARCHAR2(3)),
+    Column("indexing", VARCHAR2(7)),
+    Column("auto", VARCHAR2(3)),
+).alias("a_indexes")
+
+all_ind_expressions = Table(
+    "all_ind_expressions" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("index_owner", VARCHAR2(128), nullable=False),
+    Column("index_name", VARCHAR2(128), nullable=False),
+    Column("table_owner", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_expression", LONG),
+    Column("column_position", NUMBER, nullable=False),
+).alias("a_ind_expressions")
+
+all_constraints = Table(
+    "all_constraints" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128)),
+    Column("constraint_name", VARCHAR2(128)),
+    Column("constraint_type", VARCHAR2(1)),
+    Column("table_name", VARCHAR2(128)),
+    Column("search_condition", LONG),
+    Column("search_condition_vc", VARCHAR2(4000)),
+    Column("r_owner", VARCHAR2(128)),
+    Column("r_constraint_name", VARCHAR2(128)),
+    Column("delete_rule", VARCHAR2(9)),
+    Column("status", VARCHAR2(8)),
+    Column("deferrable", VARCHAR2(14)),
+    Column("deferred", VARCHAR2(9)),
+    Column("validated", VARCHAR2(13)),
+    Column("generated", VARCHAR2(14)),
+    Column("bad", VARCHAR2(3)),
+    Column("rely", VARCHAR2(4)),
+    Column("last_change", DATE),
+    Column("index_owner", VARCHAR2(128)),
+    Column("index_name", VARCHAR2(128)),
+    Column("invalid", VARCHAR2(7)),
+    Column("view_related", VARCHAR2(14)),
+    Column("origin_con_id", VARCHAR2(256)),
+).alias("a_constraints")
+
+all_cons_columns = Table(
+    "all_cons_columns" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("constraint_name", VARCHAR2(128), nullable=False),
+    Column("table_name", VARCHAR2(128), nullable=False),
+    Column("column_name", VARCHAR2(4000)),
+    Column("position", NUMBER),
+).alias("a_cons_columns")
+
+# TODO figure out if it's still relevant, since there is no mention from here
+# https://docs.oracle.com/en/database/oracle/oracle-database/21/refrn/ALL_DB_LINKS.html
+# original note:
+# using user_db_links here since all_db_links appears
+# to have more restricted permissions.
+# https://docs.oracle.com/cd/B28359_01/server.111/b28310/ds_admin005.htm
+# will need to hear from more users if we are doing
+# the right thing here.  See [ticket:2619]
+all_db_links = Table(
+    "all_db_links" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("db_link", VARCHAR2(128), nullable=False),
+    Column("username", VARCHAR2(128)),
+    Column("host", VARCHAR2(2000)),
+    Column("created", DATE, nullable=False),
+    Column("hidden", VARCHAR2(3)),
+    Column("shard_internal", VARCHAR2(3)),
+    Column("valid", VARCHAR2(3)),
+    Column("intra_cdb", VARCHAR2(3)),
+).alias("a_db_links")
+
+all_synonyms = Table(
+    "all_synonyms" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128)),
+    Column("synonym_name", VARCHAR2(128)),
+    Column("table_owner", VARCHAR2(128)),
+    Column("table_name", VARCHAR2(128)),
+    Column("db_link", VARCHAR2(128)),
+    Column("origin_con_id", VARCHAR2(256)),
+).alias("a_synonyms")
+
+all_objects = Table(
+    "all_objects" + DB_LINK_PLACEHOLDER,
+    dictionary_meta,
+    Column("owner", VARCHAR2(128), nullable=False),
+    Column("object_name", VARCHAR2(128), nullable=False),
+    Column("subobject_name", VARCHAR2(128)),
+    Column("object_id", NUMBER, nullable=False),
+    Column("data_object_id", NUMBER),
+    Column("object_type", VARCHAR2(23)),
+    Column("created", DATE, nullable=False),
+    Column("last_ddl_time", DATE, nullable=False),
+    Column("timestamp", VARCHAR2(19)),
+    Column("status", VARCHAR2(7)),
+    Column("temporary", VARCHAR2(1)),
+    Column("generated", VARCHAR2(1)),
+    Column("secondary", VARCHAR2(1)),
+    Column("namespace", NUMBER, nullable=False),
+    Column("edition_name", VARCHAR2(128)),
+    Column("sharing", VARCHAR2(13)),
+    Column("editionable", VARCHAR2(1)),
+    Column("oracle_maintained", VARCHAR2(1)),
+    Column("application", VARCHAR2(1)),
+    Column("default_collation", VARCHAR2(100)),
+    Column("duplicated", VARCHAR2(1)),
+    Column("sharded", VARCHAR2(1)),
+    Column("created_appid", NUMBER),
+    Column("created_vsnid", NUMBER),
+    Column("modified_appid", NUMBER),
+    Column("modified_vsnid", NUMBER),
+).alias("a_objects")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/oracledb.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/oracledb.py
new file mode 100644
index 00000000..c09d2bae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/oracledb.py
@@ -0,0 +1,947 @@
+# dialects/oracle/oracledb.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r""".. dialect:: oracle+oracledb
+    :name: python-oracledb
+    :dbapi: oracledb
+    :connectstring: oracle+oracledb://user:pass@hostname:port[/dbname][?service_name=<service>[&key=value&key=value...]]
+    :url: https://oracle.github.io/python-oracledb/
+
+Description
+-----------
+
+Python-oracledb is the Oracle Database driver for Python. It features a default
+"thin" client mode that requires no dependencies, and an optional "thick" mode
+that uses Oracle Client libraries.  It supports SQLAlchemy features including
+two phase transactions and Asyncio.
+
+Python-oracle is the renamed, updated cx_Oracle driver. Oracle is no longer
+doing any releases in the cx_Oracle namespace.
+
+The SQLAlchemy ``oracledb`` dialect provides both a sync and an async
+implementation under the same dialect name. The proper version is
+selected depending on how the engine is created:
+
+* calling :func:`_sa.create_engine` with ``oracle+oracledb://...`` will
+  automatically select the sync version::
+
+    from sqlalchemy import create_engine
+
+    sync_engine = create_engine(
+        "oracle+oracledb://scott:tiger@localhost?service_name=FREEPDB1"
+    )
+
+* calling :func:`_asyncio.create_async_engine` with ``oracle+oracledb://...``
+  will automatically select the async version::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    asyncio_engine = create_async_engine(
+        "oracle+oracledb://scott:tiger@localhost?service_name=FREEPDB1"
+    )
+
+  The asyncio version of the dialect may also be specified explicitly using the
+  ``oracledb_async`` suffix::
+
+      from sqlalchemy.ext.asyncio import create_async_engine
+
+      asyncio_engine = create_async_engine(
+          "oracle+oracledb_async://scott:tiger@localhost?service_name=FREEPDB1"
+      )
+
+.. versionadded:: 2.0.25 added support for the async version of oracledb.
+
+Thick mode support
+------------------
+
+By default, the python-oracledb driver runs in a "thin" mode that does not
+require Oracle Client libraries to be installed. The driver also supports a
+"thick" mode that uses Oracle Client libraries to get functionality such as
+Oracle Application Continuity.
+
+To enable thick mode, call `oracledb.init_oracle_client()
+<https://python-oracledb.readthedocs.io/en/latest/api_manual/module.html#oracledb.init_oracle_client>`_
+explicitly, or pass the parameter ``thick_mode=True`` to
+:func:`_sa.create_engine`. To pass custom arguments to
+``init_oracle_client()``, like the ``lib_dir`` path, a dict may be passed, for
+example::
+
+    engine = sa.create_engine(
+        "oracle+oracledb://...",
+        thick_mode={
+            "lib_dir": "/path/to/oracle/client/lib",
+            "config_dir": "/path/to/network_config_file_directory",
+            "driver_name": "my-app : 1.0.0",
+        },
+    )
+
+Note that passing a ``lib_dir`` path should only be done on macOS or
+Windows. On Linux it does not behave as you might expect.
+
+.. seealso::
+
+    python-oracledb documentation `Enabling python-oracledb Thick mode
+    <https://python-oracledb.readthedocs.io/en/latest/user_guide/initialization.html#enabling-python-oracledb-thick-mode>`_
+
+Connecting to Oracle Database
+-----------------------------
+
+python-oracledb provides several methods of indicating the target database.
+The dialect translates from a series of different URL forms.
+
+Given the hostname, port and service name of the target database, you can
+connect in SQLAlchemy using the ``service_name`` query string parameter::
+
+    engine = create_engine(
+        "oracle+oracledb://scott:tiger@hostname:port?service_name=myservice"
+    )
+
+Connecting with Easy Connect strings
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+You can pass any valid python-oracledb connection string as the ``dsn`` key
+value in a :paramref:`_sa.create_engine.connect_args` dictionary.  See
+python-oracledb documentation `Oracle Net Services Connection Strings
+<https://python-oracledb.readthedocs.io/en/latest/user_guide/connection_handling.html#oracle-net-services-connection-strings>`_.
+
+For example to use an `Easy Connect string
+<https://download.oracle.com/ocomdocs/global/Oracle-Net-Easy-Connect-Plus.pdf>`_
+with a timeout to prevent connection establishment from hanging if the network
+transport to the database cannot be establishd in 30 seconds, and also setting
+a keep-alive time of 60 seconds to stop idle network connections from being
+terminated by a firewall::
+
+    e = create_engine(
+        "oracle+oracledb://@",
+        connect_args={
+            "user": "scott",
+            "password": "tiger",
+            "dsn": "hostname:port/myservice?transport_connect_timeout=30&expire_time=60",
+        },
+    )
+
+The Easy Connect syntax has been enhanced during the life of Oracle Database.
+Review the documentation for your database version.  The current documentation
+is at `Understanding the Easy Connect Naming Method
+<https://www.oracle.com/pls/topic/lookup?ctx=dblatest&id=GUID-B0437826-43C1-49EC-A94D-B650B6A4A6EE>`_.
+
+The general syntax is similar to:
+
+.. sourcecode:: text
+
+    [[protocol:]//]host[:port][/[service_name]][?parameter_name=value{&parameter_name=value}]
+
+Note that although the SQLAlchemy URL syntax ``hostname:port/dbname`` looks
+like Oracle's Easy Connect syntax, it is different. SQLAlchemy's URL requires a
+system identifier (SID) for the ``dbname`` component::
+
+    engine = create_engine("oracle+oracledb://scott:tiger@hostname:port/sid")
+
+Easy Connect syntax does not support SIDs. It uses services names, which are
+the preferred choice for connecting to Oracle Database.
+
+Passing python-oracledb connect arguments
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Other python-oracledb driver `connection options
+<https://python-oracledb.readthedocs.io/en/latest/api_manual/module.html#oracledb.connect>`_
+can be passed in ``connect_args``.  For example::
+
+    e = create_engine(
+        "oracle+oracledb://@",
+        connect_args={
+            "user": "scott",
+            "password": "tiger",
+            "dsn": "hostname:port/myservice",
+            "events": True,
+            "mode": oracledb.AUTH_MODE_SYSDBA,
+        },
+    )
+
+Connecting with tnsnames.ora TNS aliases
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+If no port, database name, or service name is provided, the dialect will use an
+Oracle Database DSN "connection string".  This takes the "hostname" portion of
+the URL as the data source name.  For example, if the ``tnsnames.ora`` file
+contains a `TNS Alias
+<https://python-oracledb.readthedocs.io/en/latest/user_guide/connection_handling.html#tns-aliases-for-connection-strings>`_
+of ``myalias`` as below:
+
+.. sourcecode:: text
+
+    myalias =
+      (DESCRIPTION =
+        (ADDRESS = (PROTOCOL = TCP)(HOST = mymachine.example.com)(PORT = 1521))
+        (CONNECT_DATA =
+          (SERVER = DEDICATED)
+          (SERVICE_NAME = orclpdb1)
+        )
+      )
+
+The python-oracledb dialect connects to this database service when ``myalias`` is the
+hostname portion of the URL, without specifying a port, database name or
+``service_name``::
+
+    engine = create_engine("oracle+oracledb://scott:tiger@myalias")
+
+Connecting to Oracle Autonomous Database
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Users of Oracle Autonomous Database should use either use the TNS Alias URL
+shown above, or pass the TNS Alias as the ``dsn`` key value in a
+:paramref:`_sa.create_engine.connect_args` dictionary.
+
+If Oracle Autonomous Database is configured for mutual TLS ("mTLS")
+connections, then additional configuration is required as shown in `Connecting
+to Oracle Cloud Autonomous Databases
+<https://python-oracledb.readthedocs.io/en/latest/user_guide/connection_handling.html#connecting-to-oracle-cloud-autonomous-databases>`_. In
+summary, Thick mode users should configure file locations and set the wallet
+path in ``sqlnet.ora`` appropriately::
+
+    e = create_engine(
+        "oracle+oracledb://@",
+        thick_mode={
+            # directory containing tnsnames.ora and cwallet.so
+            "config_dir": "/opt/oracle/wallet_dir",
+        },
+        connect_args={
+            "user": "scott",
+            "password": "tiger",
+            "dsn": "mydb_high",
+        },
+    )
+
+Thin mode users of mTLS should pass the appropriate directories and PEM wallet
+password when creating the engine, similar to::
+
+    e = create_engine(
+        "oracle+oracledb://@",
+        connect_args={
+            "user": "scott",
+            "password": "tiger",
+            "dsn": "mydb_high",
+            "config_dir": "/opt/oracle/wallet_dir",  # directory containing tnsnames.ora
+            "wallet_location": "/opt/oracle/wallet_dir",  # directory containing ewallet.pem
+            "wallet_password": "top secret",  # password for the PEM file
+        },
+    )
+
+Typically ``config_dir`` and ``wallet_location`` are the same directory, which
+is where the Oracle Autonomous Database wallet zip file was extracted.  Note
+this directory should be protected.
+
+Connection Pooling
+------------------
+
+Applications with multiple concurrent users should use connection pooling. A
+minimal sized connection pool is also beneficial for long-running, single-user
+applications that do not frequently use a connection.
+
+The python-oracledb driver provides its own connection pool implementation that
+may be used in place of SQLAlchemy's pooling functionality.  The driver pool
+gives support for high availability features such as dead connection detection,
+connection draining for planned database downtime, support for Oracle
+Application Continuity and Transparent Application Continuity, and gives
+support for `Database Resident Connection Pooling (DRCP)
+<https://python-oracledb.readthedocs.io/en/latest/user_guide/connection_handling.html#database-resident-connection-pooling-drcp>`_.
+
+To take advantage of python-oracledb's pool, use the
+:paramref:`_sa.create_engine.creator` parameter to provide a function that
+returns a new connection, along with setting
+:paramref:`_sa.create_engine.pool_class` to ``NullPool`` to disable
+SQLAlchemy's pooling::
+
+    import oracledb
+    from sqlalchemy import create_engine
+    from sqlalchemy import text
+    from sqlalchemy.pool import NullPool
+
+    # Uncomment to use the optional python-oracledb Thick mode.
+    # Review the python-oracledb doc for the appropriate parameters
+    # oracledb.init_oracle_client(<your parameters>)
+
+    pool = oracledb.create_pool(
+        user="scott",
+        password="tiger",
+        dsn="localhost:1521/freepdb1",
+        min=1,
+        max=4,
+        increment=1,
+    )
+    engine = create_engine(
+        "oracle+oracledb://", creator=pool.acquire, poolclass=NullPool
+    )
+
+The above engine may then be used normally. Internally, python-oracledb handles
+connection pooling::
+
+    with engine.connect() as conn:
+        print(conn.scalar(text("select 1 from dual")))
+
+Refer to the python-oracledb documentation for `oracledb.create_pool()
+<https://python-oracledb.readthedocs.io/en/latest/api_manual/module.html#oracledb.create_pool>`_
+for the arguments that can be used when creating a connection pool.
+
+.. _drcp:
+
+Using Oracle Database Resident Connection Pooling (DRCP)
+--------------------------------------------------------
+
+When using Oracle Database's Database Resident Connection Pooling (DRCP), the
+best practice is to specify a connection class and "purity". Refer to the
+`python-oracledb documentation on DRCP
+<https://python-oracledb.readthedocs.io/en/latest/user_guide/connection_handling.html#database-resident-connection-pooling-drcp>`_.
+For example::
+
+    import oracledb
+    from sqlalchemy import create_engine
+    from sqlalchemy import text
+    from sqlalchemy.pool import NullPool
+
+    # Uncomment to use the optional python-oracledb Thick mode.
+    # Review the python-oracledb doc for the appropriate parameters
+    # oracledb.init_oracle_client(<your parameters>)
+
+    pool = oracledb.create_pool(
+        user="scott",
+        password="tiger",
+        dsn="localhost:1521/freepdb1",
+        min=1,
+        max=4,
+        increment=1,
+        cclass="MYCLASS",
+        purity=oracledb.PURITY_SELF,
+    )
+    engine = create_engine(
+        "oracle+oracledb://", creator=pool.acquire, poolclass=NullPool
+    )
+
+The above engine may then be used normally where python-oracledb handles
+application connection pooling and Oracle Database additionally uses DRCP::
+
+    with engine.connect() as conn:
+        print(conn.scalar(text("select 1 from dual")))
+
+If you wish to use different connection classes or purities for different
+connections, then wrap ``pool.acquire()``::
+
+    import oracledb
+    from sqlalchemy import create_engine
+    from sqlalchemy import text
+    from sqlalchemy.pool import NullPool
+
+    # Uncomment to use python-oracledb Thick mode.
+    # Review the python-oracledb doc for the appropriate parameters
+    # oracledb.init_oracle_client(<your parameters>)
+
+    pool = oracledb.create_pool(
+        user="scott",
+        password="tiger",
+        dsn="localhost:1521/freepdb1",
+        min=1,
+        max=4,
+        increment=1,
+        cclass="MYCLASS",
+        purity=oracledb.PURITY_SELF,
+    )
+
+
+    def creator():
+        return pool.acquire(cclass="MYOTHERCLASS", purity=oracledb.PURITY_NEW)
+
+
+    engine = create_engine(
+        "oracle+oracledb://", creator=creator, poolclass=NullPool
+    )
+
+Engine Options consumed by the SQLAlchemy oracledb dialect outside of the driver
+--------------------------------------------------------------------------------
+
+There are also options that are consumed by the SQLAlchemy oracledb dialect
+itself.  These options are always passed directly to :func:`_sa.create_engine`,
+such as::
+
+    e = create_engine("oracle+oracledb://user:pass@tnsalias", arraysize=500)
+
+The parameters accepted by the oracledb dialect are as follows:
+
+* ``arraysize`` - set the driver cursor.arraysize value. It defaults to
+  ``None``, indicating that the driver default value of 100 should be used.
+  This setting controls how many rows are buffered when fetching rows, and can
+  have a significant effect on performance if increased for queries that return
+  large numbers of rows.
+
+  .. versionchanged:: 2.0.26 - changed the default value from 50 to None,
+    to use the default value of the driver itself.
+
+* ``auto_convert_lobs`` - defaults to True; See :ref:`oracledb_lob`.
+
+* ``coerce_to_decimal`` - see :ref:`oracledb_numeric` for detail.
+
+* ``encoding_errors`` - see :ref:`oracledb_unicode_encoding_errors` for detail.
+
+.. _oracledb_unicode:
+
+Unicode
+-------
+
+As is the case for all DBAPIs under Python 3, all strings are inherently
+Unicode strings.
+
+Ensuring the Correct Client Encoding
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+In python-oracledb, the encoding used for all character data is "UTF-8".
+
+Unicode-specific Column datatypes
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The Core expression language handles unicode data by use of the
+:class:`.Unicode` and :class:`.UnicodeText` datatypes.  These types correspond
+to the VARCHAR2 and CLOB Oracle Database datatypes by default.  When using
+these datatypes with Unicode data, it is expected that the database is
+configured with a Unicode-aware character set so that the VARCHAR2 and CLOB
+datatypes can accommodate the data.
+
+In the case that Oracle Database is not configured with a Unicode character
+set, the two options are to use the :class:`_types.NCHAR` and
+:class:`_oracle.NCLOB` datatypes explicitly, or to pass the flag
+``use_nchar_for_unicode=True`` to :func:`_sa.create_engine`, which will cause
+the SQLAlchemy dialect to use NCHAR/NCLOB for the :class:`.Unicode` /
+:class:`.UnicodeText` datatypes instead of VARCHAR/CLOB.
+
+.. versionchanged:: 1.3 The :class:`.Unicode` and :class:`.UnicodeText`
+   datatypes now correspond to the ``VARCHAR2`` and ``CLOB`` Oracle Database
+   datatypes unless the ``use_nchar_for_unicode=True`` is passed to the dialect
+   when :func:`_sa.create_engine` is called.
+
+
+.. _oracledb_unicode_encoding_errors:
+
+Encoding Errors
+^^^^^^^^^^^^^^^
+
+For the unusual case that data in Oracle Database is present with a broken
+encoding, the dialect accepts a parameter ``encoding_errors`` which will be
+passed to Unicode decoding functions in order to affect how decoding errors are
+handled.  The value is ultimately consumed by the Python `decode
+<https://docs.python.org/3/library/stdtypes.html#bytes.decode>`_ function, and
+is passed both via python-oracledb's ``encodingErrors`` parameter consumed by
+``Cursor.var()``, as well as SQLAlchemy's own decoding function, as the
+python-oracledb dialect makes use of both under different circumstances.
+
+.. versionadded:: 1.3.11
+
+
+.. _oracledb_setinputsizes:
+
+Fine grained control over python-oracledb data binding with setinputsizes
+-------------------------------------------------------------------------
+
+The python-oracle DBAPI has a deep and fundamental reliance upon the usage of
+the DBAPI ``setinputsizes()`` call.  The purpose of this call is to establish
+the datatypes that are bound to a SQL statement for Python values being passed
+as parameters.  While virtually no other DBAPI assigns any use to the
+``setinputsizes()`` call, the python-oracledb DBAPI relies upon it heavily in
+its interactions with the Oracle Database, and in some scenarios it is not
+possible for SQLAlchemy to know exactly how data should be bound, as some
+settings can cause profoundly different performance characteristics, while
+altering the type coercion behavior at the same time.
+
+Users of the oracledb dialect are **strongly encouraged** to read through
+python-oracledb's list of built-in datatype symbols at `Database Types
+<https://python-oracledb.readthedocs.io/en/latest/api_manual/module.html#database-types>`_
+Note that in some cases, significant performance degradation can occur when
+using these types vs. not.
+
+On the SQLAlchemy side, the :meth:`.DialectEvents.do_setinputsizes` event can
+be used both for runtime visibility (e.g. logging) of the setinputsizes step as
+well as to fully control how ``setinputsizes()`` is used on a per-statement
+basis.
+
+.. versionadded:: 1.2.9 Added :meth:`.DialectEvents.setinputsizes`
+
+
+Example 1 - logging all setinputsizes calls
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The following example illustrates how to log the intermediary values from a
+SQLAlchemy perspective before they are converted to the raw ``setinputsizes()``
+parameter dictionary.  The keys of the dictionary are :class:`.BindParameter`
+objects which have a ``.key`` and a ``.type`` attribute::
+
+    from sqlalchemy import create_engine, event
+
+    engine = create_engine(
+        "oracle+oracledb://scott:tiger@localhost:1521?service_name=freepdb1"
+    )
+
+
+    @event.listens_for(engine, "do_setinputsizes")
+    def _log_setinputsizes(inputsizes, cursor, statement, parameters, context):
+        for bindparam, dbapitype in inputsizes.items():
+            log.info(
+                "Bound parameter name: %s  SQLAlchemy type: %r DBAPI object: %s",
+                bindparam.key,
+                bindparam.type,
+                dbapitype,
+            )
+
+Example 2 - remove all bindings to CLOB
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+For performance, fetching LOB datatypes from Oracle Database is set by default
+for the ``Text`` type within SQLAlchemy.  This setting can be modified as
+follows::
+
+
+    from sqlalchemy import create_engine, event
+    from oracledb import CLOB
+
+    engine = create_engine(
+        "oracle+oracledb://scott:tiger@localhost:1521?service_name=freepdb1"
+    )
+
+
+    @event.listens_for(engine, "do_setinputsizes")
+    def _remove_clob(inputsizes, cursor, statement, parameters, context):
+        for bindparam, dbapitype in list(inputsizes.items()):
+            if dbapitype is CLOB:
+                del inputsizes[bindparam]
+
+.. _oracledb_lob:
+
+LOB Datatypes
+--------------
+
+LOB datatypes refer to the "large object" datatypes such as CLOB, NCLOB and
+BLOB. Oracle Database can efficiently return these datatypes as a single
+buffer. SQLAlchemy makes use of type handlers to do this by default.
+
+To disable the use of the type handlers and deliver LOB objects as classic
+buffered objects with a ``read()`` method, the parameter
+``auto_convert_lobs=False`` may be passed to :func:`_sa.create_engine`.
+
+.. _oracledb_returning:
+
+RETURNING Support
+-----------------
+
+The oracledb dialect implements RETURNING using OUT parameters.  The dialect
+supports RETURNING fully.
+
+Two Phase Transaction Support
+-----------------------------
+
+Two phase transactions are fully supported with python-oracledb. (Thin mode
+requires python-oracledb 2.3).  APIs for two phase transactions are provided at
+the Core level via :meth:`_engine.Connection.begin_twophase` and
+:paramref:`_orm.Session.twophase` for transparent ORM use.
+
+.. versionchanged:: 2.0.32 added support for two phase transactions
+
+.. _oracledb_numeric:
+
+Precision Numerics
+------------------
+
+SQLAlchemy's numeric types can handle receiving and returning values as Python
+``Decimal`` objects or float objects.  When a :class:`.Numeric` object, or a
+subclass such as :class:`.Float`, :class:`_oracle.DOUBLE_PRECISION` etc. is in
+use, the :paramref:`.Numeric.asdecimal` flag determines if values should be
+coerced to ``Decimal`` upon return, or returned as float objects.  To make
+matters more complicated under Oracle Database, the ``NUMBER`` type can also
+represent integer values if the "scale" is zero, so the Oracle
+Database-specific :class:`_oracle.NUMBER` type takes this into account as well.
+
+The oracledb dialect makes extensive use of connection- and cursor-level
+"outputtypehandler" callables in order to coerce numeric values as requested.
+These callables are specific to the specific flavor of :class:`.Numeric` in
+use, as well as if no SQLAlchemy typing objects are present.  There are
+observed scenarios where Oracle Database may send incomplete or ambiguous
+information about the numeric types being returned, such as a query where the
+numeric types are buried under multiple levels of subquery.  The type handlers
+do their best to make the right decision in all cases, deferring to the
+underlying python-oracledb DBAPI for all those cases where the driver can make
+the best decision.
+
+When no typing objects are present, as when executing plain SQL strings, a
+default "outputtypehandler" is present which will generally return numeric
+values which specify precision and scale as Python ``Decimal`` objects.  To
+disable this coercion to decimal for performance reasons, pass the flag
+``coerce_to_decimal=False`` to :func:`_sa.create_engine`::
+
+    engine = create_engine(
+        "oracle+oracledb://scott:tiger@tnsalias", coerce_to_decimal=False
+    )
+
+The ``coerce_to_decimal`` flag only impacts the results of plain string
+SQL statements that are not otherwise associated with a :class:`.Numeric`
+SQLAlchemy type (or a subclass of such).
+
+.. versionchanged:: 1.2 The numeric handling system for the oracle dialects has
+   been reworked to take advantage of newer driver features as well as better
+   integration of outputtypehandlers.
+
+.. versionadded:: 2.0.0 added support for the python-oracledb driver.
+
+"""  # noqa
+from __future__ import annotations
+
+import collections
+import re
+from typing import Any
+from typing import TYPE_CHECKING
+
+from . import cx_oracle as _cx_oracle
+from ... import exc
+from ... import pool
+from ...connectors.asyncio import AsyncAdapt_dbapi_connection
+from ...connectors.asyncio import AsyncAdapt_dbapi_cursor
+from ...connectors.asyncio import AsyncAdapt_dbapi_ss_cursor
+from ...connectors.asyncio import AsyncAdaptFallback_dbapi_connection
+from ...engine import default
+from ...util import asbool
+from ...util import await_fallback
+from ...util import await_only
+
+if TYPE_CHECKING:
+    from oracledb import AsyncConnection
+    from oracledb import AsyncCursor
+
+
+class OracleExecutionContext_oracledb(
+    _cx_oracle.OracleExecutionContext_cx_oracle
+):
+    pass
+
+
+class OracleDialect_oracledb(_cx_oracle.OracleDialect_cx_oracle):
+    supports_statement_cache = True
+    execution_ctx_cls = OracleExecutionContext_oracledb
+
+    driver = "oracledb"
+    _min_version = (1,)
+
+    def __init__(
+        self,
+        auto_convert_lobs=True,
+        coerce_to_decimal=True,
+        arraysize=None,
+        encoding_errors=None,
+        thick_mode=None,
+        **kwargs,
+    ):
+        super().__init__(
+            auto_convert_lobs,
+            coerce_to_decimal,
+            arraysize,
+            encoding_errors,
+            **kwargs,
+        )
+
+        if self.dbapi is not None and (
+            thick_mode or isinstance(thick_mode, dict)
+        ):
+            kw = thick_mode if isinstance(thick_mode, dict) else {}
+            self.dbapi.init_oracle_client(**kw)
+
+    @classmethod
+    def import_dbapi(cls):
+        import oracledb
+
+        return oracledb
+
+    @classmethod
+    def is_thin_mode(cls, connection):
+        return connection.connection.dbapi_connection.thin
+
+    @classmethod
+    def get_async_dialect_cls(cls, url):
+        return OracleDialectAsync_oracledb
+
+    def _load_version(self, dbapi_module):
+        version = (0, 0, 0)
+        if dbapi_module is not None:
+            m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", dbapi_module.version)
+            if m:
+                version = tuple(
+                    int(x) for x in m.group(1, 2, 3) if x is not None
+                )
+        self.oracledb_ver = version
+        if (
+            self.oracledb_ver > (0, 0, 0)
+            and self.oracledb_ver < self._min_version
+        ):
+            raise exc.InvalidRequestError(
+                f"oracledb version {self._min_version} and above are supported"
+            )
+
+    def do_begin_twophase(self, connection, xid):
+        conn_xis = connection.connection.xid(*xid)
+        connection.connection.tpc_begin(conn_xis)
+        connection.connection.info["oracledb_xid"] = conn_xis
+
+    def do_prepare_twophase(self, connection, xid):
+        should_commit = connection.connection.tpc_prepare()
+        connection.info["oracledb_should_commit"] = should_commit
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if recover:
+            conn_xid = connection.connection.xid(*xid)
+        else:
+            conn_xid = None
+        connection.connection.tpc_rollback(conn_xid)
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        conn_xid = None
+        if not is_prepared:
+            should_commit = connection.connection.tpc_prepare()
+        elif recover:
+            conn_xid = connection.connection.xid(*xid)
+            should_commit = True
+        else:
+            should_commit = connection.info["oracledb_should_commit"]
+        if should_commit:
+            connection.connection.tpc_commit(conn_xid)
+
+    def do_recover_twophase(self, connection):
+        return [
+            # oracledb seems to return bytes
+            (
+                fi,
+                gti.decode() if isinstance(gti, bytes) else gti,
+                bq.decode() if isinstance(bq, bytes) else bq,
+            )
+            for fi, gti, bq in connection.connection.tpc_recover()
+        ]
+
+    def _check_max_identifier_length(self, connection):
+        if self.oracledb_ver >= (2, 5):
+            max_len = connection.connection.max_identifier_length
+            if max_len is not None:
+                return max_len
+        return super()._check_max_identifier_length(connection)
+
+
+class AsyncAdapt_oracledb_cursor(AsyncAdapt_dbapi_cursor):
+    _cursor: AsyncCursor
+    __slots__ = ()
+
+    @property
+    def outputtypehandler(self):
+        return self._cursor.outputtypehandler
+
+    @outputtypehandler.setter
+    def outputtypehandler(self, value):
+        self._cursor.outputtypehandler = value
+
+    def var(self, *args, **kwargs):
+        return self._cursor.var(*args, **kwargs)
+
+    def close(self):
+        self._rows.clear()
+        self._cursor.close()
+
+    def setinputsizes(self, *args: Any, **kwargs: Any) -> Any:
+        return self._cursor.setinputsizes(*args, **kwargs)
+
+    def _aenter_cursor(self, cursor: AsyncCursor) -> AsyncCursor:
+        try:
+            return cursor.__enter__()
+        except Exception as error:
+            self._adapt_connection._handle_exception(error)
+
+    async def _execute_async(self, operation, parameters):
+        # override to not use mutex, oracledb already has a mutex
+
+        if parameters is None:
+            result = await self._cursor.execute(operation)
+        else:
+            result = await self._cursor.execute(operation, parameters)
+
+        if self._cursor.description and not self.server_side:
+            self._rows = collections.deque(await self._cursor.fetchall())
+        return result
+
+    async def _executemany_async(
+        self,
+        operation,
+        seq_of_parameters,
+    ):
+        # override to not use mutex, oracledb already has a mutex
+        return await self._cursor.executemany(operation, seq_of_parameters)
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        self.close()
+
+
+class AsyncAdapt_oracledb_ss_cursor(
+    AsyncAdapt_dbapi_ss_cursor, AsyncAdapt_oracledb_cursor
+):
+    __slots__ = ()
+
+    def close(self) -> None:
+        if self._cursor is not None:
+            self._cursor.close()
+            self._cursor = None  # type: ignore
+
+
+class AsyncAdapt_oracledb_connection(AsyncAdapt_dbapi_connection):
+    _connection: AsyncConnection
+    __slots__ = ()
+
+    thin = True
+
+    _cursor_cls = AsyncAdapt_oracledb_cursor
+    _ss_cursor_cls = None
+
+    @property
+    def autocommit(self):
+        return self._connection.autocommit
+
+    @autocommit.setter
+    def autocommit(self, value):
+        self._connection.autocommit = value
+
+    @property
+    def outputtypehandler(self):
+        return self._connection.outputtypehandler
+
+    @outputtypehandler.setter
+    def outputtypehandler(self, value):
+        self._connection.outputtypehandler = value
+
+    @property
+    def version(self):
+        return self._connection.version
+
+    @property
+    def stmtcachesize(self):
+        return self._connection.stmtcachesize
+
+    @stmtcachesize.setter
+    def stmtcachesize(self, value):
+        self._connection.stmtcachesize = value
+
+    @property
+    def max_identifier_length(self):
+        return self._connection.max_identifier_length
+
+    def cursor(self):
+        return AsyncAdapt_oracledb_cursor(self)
+
+    def ss_cursor(self):
+        return AsyncAdapt_oracledb_ss_cursor(self)
+
+    def xid(self, *args: Any, **kwargs: Any) -> Any:
+        return self._connection.xid(*args, **kwargs)
+
+    def tpc_begin(self, *args: Any, **kwargs: Any) -> Any:
+        return self.await_(self._connection.tpc_begin(*args, **kwargs))
+
+    def tpc_commit(self, *args: Any, **kwargs: Any) -> Any:
+        return self.await_(self._connection.tpc_commit(*args, **kwargs))
+
+    def tpc_prepare(self, *args: Any, **kwargs: Any) -> Any:
+        return self.await_(self._connection.tpc_prepare(*args, **kwargs))
+
+    def tpc_recover(self, *args: Any, **kwargs: Any) -> Any:
+        return self.await_(self._connection.tpc_recover(*args, **kwargs))
+
+    def tpc_rollback(self, *args: Any, **kwargs: Any) -> Any:
+        return self.await_(self._connection.tpc_rollback(*args, **kwargs))
+
+
+class AsyncAdaptFallback_oracledb_connection(
+    AsyncAdaptFallback_dbapi_connection, AsyncAdapt_oracledb_connection
+):
+    __slots__ = ()
+
+
+class OracledbAdaptDBAPI:
+    def __init__(self, oracledb) -> None:
+        self.oracledb = oracledb
+
+        for k, v in self.oracledb.__dict__.items():
+            if k != "connect":
+                self.__dict__[k] = v
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop("async_creator_fn", self.oracledb.connect_async)
+
+        if asbool(async_fallback):
+            return AsyncAdaptFallback_oracledb_connection(
+                self, await_fallback(creator_fn(*arg, **kw))
+            )
+
+        else:
+            return AsyncAdapt_oracledb_connection(
+                self, await_only(creator_fn(*arg, **kw))
+            )
+
+
+class OracleExecutionContextAsync_oracledb(OracleExecutionContext_oracledb):
+    # restore default create cursor
+    create_cursor = default.DefaultExecutionContext.create_cursor
+
+    def create_default_cursor(self):
+        # copy of OracleExecutionContext_cx_oracle.create_cursor
+        c = self._dbapi_connection.cursor()
+        if self.dialect.arraysize:
+            c.arraysize = self.dialect.arraysize
+
+        return c
+
+    def create_server_side_cursor(self):
+        c = self._dbapi_connection.ss_cursor()
+        if self.dialect.arraysize:
+            c.arraysize = self.dialect.arraysize
+
+        return c
+
+
+class OracleDialectAsync_oracledb(OracleDialect_oracledb):
+    is_async = True
+    supports_server_side_cursors = True
+    supports_statement_cache = True
+    execution_ctx_cls = OracleExecutionContextAsync_oracledb
+
+    _min_version = (2,)
+
+    # thick_mode mode is not supported by asyncio, oracledb will raise
+    @classmethod
+    def import_dbapi(cls):
+        import oracledb
+
+        return OracledbAdaptDBAPI(oracledb)
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = OracleDialect_oracledb
+dialect_async = OracleDialectAsync_oracledb
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/provision.py
new file mode 100644
index 00000000..3587de9d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/provision.py
@@ -0,0 +1,220 @@
+# dialects/oracle/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import create_engine
+from ... import exc
+from ... import inspect
+from ...engine import url as sa_url
+from ...testing.provision import configure_follower
+from ...testing.provision import create_db
+from ...testing.provision import drop_all_schema_objects_post_tables
+from ...testing.provision import drop_all_schema_objects_pre_tables
+from ...testing.provision import drop_db
+from ...testing.provision import follower_url_from_main
+from ...testing.provision import log
+from ...testing.provision import post_configure_engine
+from ...testing.provision import run_reap_dbs
+from ...testing.provision import set_default_schema_on_connection
+from ...testing.provision import stop_test_class_outside_fixtures
+from ...testing.provision import temp_table_keyword_args
+from ...testing.provision import update_db_opts
+
+
+@create_db.for_db("oracle")
+def _oracle_create_db(cfg, eng, ident):
+    # NOTE: make sure you've run "ALTER DATABASE default tablespace users" or
+    # similar, so that the default tablespace is not "system"; reflection will
+    # fail otherwise
+    with eng.begin() as conn:
+        conn.exec_driver_sql("create user %s identified by xe" % ident)
+        conn.exec_driver_sql("create user %s_ts1 identified by xe" % ident)
+        conn.exec_driver_sql("create user %s_ts2 identified by xe" % ident)
+        conn.exec_driver_sql("grant dba to %s" % (ident,))
+        conn.exec_driver_sql("grant unlimited tablespace to %s" % ident)
+        conn.exec_driver_sql("grant unlimited tablespace to %s_ts1" % ident)
+        conn.exec_driver_sql("grant unlimited tablespace to %s_ts2" % ident)
+        # these are needed to create materialized views
+        conn.exec_driver_sql("grant create table to %s" % ident)
+        conn.exec_driver_sql("grant create table to %s_ts1" % ident)
+        conn.exec_driver_sql("grant create table to %s_ts2" % ident)
+
+
+@configure_follower.for_db("oracle")
+def _oracle_configure_follower(config, ident):
+    config.test_schema = "%s_ts1" % ident
+    config.test_schema_2 = "%s_ts2" % ident
+
+
+def _ora_drop_ignore(conn, dbname):
+    try:
+        conn.exec_driver_sql("drop user %s cascade" % dbname)
+        log.info("Reaped db: %s", dbname)
+        return True
+    except exc.DatabaseError as err:
+        log.warning("couldn't drop db: %s", err)
+        return False
+
+
+@drop_all_schema_objects_pre_tables.for_db("oracle")
+def _ora_drop_all_schema_objects_pre_tables(cfg, eng):
+    _purge_recyclebin(eng)
+    _purge_recyclebin(eng, cfg.test_schema)
+
+
+@drop_all_schema_objects_post_tables.for_db("oracle")
+def _ora_drop_all_schema_objects_post_tables(cfg, eng):
+    with eng.begin() as conn:
+        for syn in conn.dialect._get_synonyms(conn, None, None, None):
+            conn.exec_driver_sql(f"drop synonym {syn['synonym_name']}")
+
+        for syn in conn.dialect._get_synonyms(
+            conn, cfg.test_schema, None, None
+        ):
+            conn.exec_driver_sql(
+                f"drop synonym {cfg.test_schema}.{syn['synonym_name']}"
+            )
+
+        for tmp_table in inspect(conn).get_temp_table_names():
+            conn.exec_driver_sql(f"drop table {tmp_table}")
+
+
+@drop_db.for_db("oracle")
+def _oracle_drop_db(cfg, eng, ident):
+    with eng.begin() as conn:
+        # cx_Oracle seems to occasionally leak open connections when a large
+        # suite it run, even if we confirm we have zero references to
+        # connection objects.
+        # while there is a "kill session" command in Oracle Database,
+        # it unfortunately does not release the connection sufficiently.
+        _ora_drop_ignore(conn, ident)
+        _ora_drop_ignore(conn, "%s_ts1" % ident)
+        _ora_drop_ignore(conn, "%s_ts2" % ident)
+
+
+@stop_test_class_outside_fixtures.for_db("oracle")
+def _ora_stop_test_class_outside_fixtures(config, db, cls):
+    try:
+        _purge_recyclebin(db)
+    except exc.DatabaseError as err:
+        log.warning("purge recyclebin command failed: %s", err)
+
+    # clear statement cache on all connections that were used
+    # https://github.com/oracle/python-cx_Oracle/issues/519
+
+    for cx_oracle_conn in _all_conns:
+        try:
+            sc = cx_oracle_conn.stmtcachesize
+        except db.dialect.dbapi.InterfaceError:
+            # connection closed
+            pass
+        else:
+            cx_oracle_conn.stmtcachesize = 0
+            cx_oracle_conn.stmtcachesize = sc
+    _all_conns.clear()
+
+
+def _purge_recyclebin(eng, schema=None):
+    with eng.begin() as conn:
+        if schema is None:
+            # run magic command to get rid of identity sequences
+            # https://floo.bar/2019/11/29/drop-the-underlying-sequence-of-an-identity-column/  # noqa: E501
+            conn.exec_driver_sql("purge recyclebin")
+        else:
+            # per user: https://community.oracle.com/tech/developers/discussion/2255402/how-to-clear-dba-recyclebin-for-a-particular-user  # noqa: E501
+            for owner, object_name, type_ in conn.exec_driver_sql(
+                "select owner, object_name,type from "
+                "dba_recyclebin where owner=:schema and type='TABLE'",
+                {"schema": conn.dialect.denormalize_name(schema)},
+            ).all():
+                conn.exec_driver_sql(f'purge {type_} {owner}."{object_name}"')
+
+
+_all_conns = set()
+
+
+@post_configure_engine.for_db("oracle")
+def _oracle_post_configure_engine(url, engine, follower_ident):
+    from sqlalchemy import event
+
+    @event.listens_for(engine, "checkout")
+    def checkout(dbapi_con, con_record, con_proxy):
+        _all_conns.add(dbapi_con)
+
+    @event.listens_for(engine, "checkin")
+    def checkin(dbapi_connection, connection_record):
+        # work around cx_Oracle issue:
+        # https://github.com/oracle/python-cx_Oracle/issues/530
+        # invalidate oracle connections that had 2pc set up
+        if "cx_oracle_xid" in connection_record.info:
+            connection_record.invalidate()
+
+
+@run_reap_dbs.for_db("oracle")
+def _reap_oracle_dbs(url, idents):
+    log.info("db reaper connecting to %r", url)
+    eng = create_engine(url)
+    with eng.begin() as conn:
+        log.info("identifiers in file: %s", ", ".join(idents))
+
+        to_reap = conn.exec_driver_sql(
+            "select u.username from all_users u where username "
+            "like 'TEST_%' and not exists (select username "
+            "from v$session where username=u.username)"
+        )
+        all_names = {username.lower() for (username,) in to_reap}
+        to_drop = set()
+        for name in all_names:
+            if name.endswith("_ts1") or name.endswith("_ts2"):
+                continue
+            elif name in idents:
+                to_drop.add(name)
+                if "%s_ts1" % name in all_names:
+                    to_drop.add("%s_ts1" % name)
+                if "%s_ts2" % name in all_names:
+                    to_drop.add("%s_ts2" % name)
+
+        dropped = total = 0
+        for total, username in enumerate(to_drop, 1):
+            if _ora_drop_ignore(conn, username):
+                dropped += 1
+        log.info(
+            "Dropped %d out of %d stale databases detected", dropped, total
+        )
+
+
+@follower_url_from_main.for_db("oracle")
+def _oracle_follower_url_from_main(url, ident):
+    url = sa_url.make_url(url)
+    return url.set(username=ident, password="xe")
+
+
+@temp_table_keyword_args.for_db("oracle")
+def _oracle_temp_table_keyword_args(cfg, eng):
+    return {
+        "prefixes": ["GLOBAL TEMPORARY"],
+        "oracle_on_commit": "PRESERVE ROWS",
+    }
+
+
+@set_default_schema_on_connection.for_db("oracle")
+def _oracle_set_default_schema_on_connection(
+    cfg, dbapi_connection, schema_name
+):
+    cursor = dbapi_connection.cursor()
+    cursor.execute("ALTER SESSION SET CURRENT_SCHEMA=%s" % schema_name)
+    cursor.close()
+
+
+@update_db_opts.for_db("oracle")
+def _update_db_opts(db_url, db_opts, options):
+    """Set database options (db_opts) for a test database that we created."""
+    if (
+        options.oracledb_thick_mode
+        and sa_url.make_url(db_url).get_driver_name() == "oracledb"
+    ):
+        db_opts["thick_mode"] = True
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/types.py
new file mode 100644
index 00000000..06aeaace
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/oracle/types.py
@@ -0,0 +1,316 @@
+# dialects/oracle/types.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+import datetime as dt
+from typing import Optional
+from typing import Type
+from typing import TYPE_CHECKING
+
+from ... import exc
+from ...sql import sqltypes
+from ...types import NVARCHAR
+from ...types import VARCHAR
+
+if TYPE_CHECKING:
+    from ...engine.interfaces import Dialect
+    from ...sql.type_api import _LiteralProcessorType
+
+
+class RAW(sqltypes._Binary):
+    __visit_name__ = "RAW"
+
+
+OracleRaw = RAW
+
+
+class NCLOB(sqltypes.Text):
+    __visit_name__ = "NCLOB"
+
+
+class VARCHAR2(VARCHAR):
+    __visit_name__ = "VARCHAR2"
+
+
+NVARCHAR2 = NVARCHAR
+
+
+class NUMBER(sqltypes.Numeric, sqltypes.Integer):
+    __visit_name__ = "NUMBER"
+
+    def __init__(self, precision=None, scale=None, asdecimal=None):
+        if asdecimal is None:
+            asdecimal = bool(scale and scale > 0)
+
+        super().__init__(precision=precision, scale=scale, asdecimal=asdecimal)
+
+    def adapt(self, impltype):
+        ret = super().adapt(impltype)
+        # leave a hint for the DBAPI handler
+        ret._is_oracle_number = True
+        return ret
+
+    @property
+    def _type_affinity(self):
+        if bool(self.scale and self.scale > 0):
+            return sqltypes.Numeric
+        else:
+            return sqltypes.Integer
+
+
+class FLOAT(sqltypes.FLOAT):
+    """Oracle Database FLOAT.
+
+    This is the same as :class:`_sqltypes.FLOAT` except that
+    an Oracle Database -specific :paramref:`_oracle.FLOAT.binary_precision`
+    parameter is accepted, and
+    the :paramref:`_sqltypes.Float.precision` parameter is not accepted.
+
+    Oracle Database FLOAT types indicate precision in terms of "binary
+    precision", which defaults to 126. For a REAL type, the value is 63. This
+    parameter does not cleanly map to a specific number of decimal places but
+    is roughly equivalent to the desired number of decimal places divided by
+    0.3103.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __visit_name__ = "FLOAT"
+
+    def __init__(
+        self,
+        binary_precision=None,
+        asdecimal=False,
+        decimal_return_scale=None,
+    ):
+        r"""
+        Construct a FLOAT
+
+        :param binary_precision: Oracle Database binary precision value to be
+         rendered in DDL. This may be approximated to the number of decimal
+         characters using the formula "decimal precision = 0.30103 * binary
+         precision".  The default value used by Oracle Database for FLOAT /
+         DOUBLE PRECISION is 126.
+
+        :param asdecimal: See :paramref:`_sqltypes.Float.asdecimal`
+
+        :param decimal_return_scale: See
+         :paramref:`_sqltypes.Float.decimal_return_scale`
+
+        """
+        super().__init__(
+            asdecimal=asdecimal, decimal_return_scale=decimal_return_scale
+        )
+        self.binary_precision = binary_precision
+
+
+class BINARY_DOUBLE(sqltypes.Double):
+    """Implement the Oracle ``BINARY_DOUBLE`` datatype.
+
+    This datatype differs from the Oracle ``DOUBLE`` datatype in that it
+    delivers a true 8-byte FP value.   The datatype may be combined with a
+    generic :class:`.Double` datatype using :meth:`.TypeEngine.with_variant`.
+
+    .. seealso::
+
+        :ref:`oracle_float_support`
+
+
+    """
+
+    __visit_name__ = "BINARY_DOUBLE"
+
+
+class BINARY_FLOAT(sqltypes.Float):
+    """Implement the Oracle ``BINARY_FLOAT`` datatype.
+
+    This datatype differs from the Oracle ``FLOAT`` datatype in that it
+    delivers a true 4-byte FP value.   The datatype may be combined with a
+    generic :class:`.Float` datatype using :meth:`.TypeEngine.with_variant`.
+
+    .. seealso::
+
+        :ref:`oracle_float_support`
+
+
+    """
+
+    __visit_name__ = "BINARY_FLOAT"
+
+
+class BFILE(sqltypes.LargeBinary):
+    __visit_name__ = "BFILE"
+
+
+class LONG(sqltypes.Text):
+    __visit_name__ = "LONG"
+
+
+class _OracleDateLiteralRender:
+    def _literal_processor_datetime(self, dialect):
+        def process(value):
+            if getattr(value, "microsecond", None):
+                value = (
+                    f"""TO_TIMESTAMP"""
+                    f"""('{value.isoformat().replace("T", " ")}', """
+                    """'YYYY-MM-DD HH24:MI:SS.FF')"""
+                )
+            else:
+                value = (
+                    f"""TO_DATE"""
+                    f"""('{value.isoformat().replace("T", " ")}', """
+                    """'YYYY-MM-DD HH24:MI:SS')"""
+                )
+            return value
+
+        return process
+
+    def _literal_processor_date(self, dialect):
+        def process(value):
+            if getattr(value, "microsecond", None):
+                value = (
+                    f"""TO_TIMESTAMP"""
+                    f"""('{value.isoformat().split("T")[0]}', """
+                    """'YYYY-MM-DD')"""
+                )
+            else:
+                value = (
+                    f"""TO_DATE"""
+                    f"""('{value.isoformat().split("T")[0]}', """
+                    """'YYYY-MM-DD')"""
+                )
+            return value
+
+        return process
+
+
+class DATE(_OracleDateLiteralRender, sqltypes.DateTime):
+    """Provide the Oracle Database DATE type.
+
+    This type has no special Python behavior, except that it subclasses
+    :class:`_types.DateTime`; this is to suit the fact that the Oracle Database
+    ``DATE`` type supports a time value.
+
+    """
+
+    __visit_name__ = "DATE"
+
+    def literal_processor(self, dialect):
+        return self._literal_processor_datetime(dialect)
+
+    def _compare_type_affinity(self, other):
+        return other._type_affinity in (sqltypes.DateTime, sqltypes.Date)
+
+
+class _OracleDate(_OracleDateLiteralRender, sqltypes.Date):
+    def literal_processor(self, dialect):
+        return self._literal_processor_date(dialect)
+
+
+class INTERVAL(sqltypes.NativeForEmulated, sqltypes._AbstractInterval):
+    __visit_name__ = "INTERVAL"
+
+    def __init__(self, day_precision=None, second_precision=None):
+        """Construct an INTERVAL.
+
+        Note that only DAY TO SECOND intervals are currently supported.
+        This is due to a lack of support for YEAR TO MONTH intervals
+        within available DBAPIs.
+
+        :param day_precision: the day precision value.  this is the number of
+          digits to store for the day field.  Defaults to "2"
+        :param second_precision: the second precision value.  this is the
+          number of digits to store for the fractional seconds field.
+          Defaults to "6".
+
+        """
+        self.day_precision = day_precision
+        self.second_precision = second_precision
+
+    @classmethod
+    def _adapt_from_generic_interval(cls, interval):
+        return INTERVAL(
+            day_precision=interval.day_precision,
+            second_precision=interval.second_precision,
+        )
+
+    @classmethod
+    def adapt_emulated_to_native(
+        cls, interval: sqltypes.Interval, **kw  # type: ignore[override]
+    ):
+        return INTERVAL(
+            day_precision=interval.day_precision,
+            second_precision=interval.second_precision,
+        )
+
+    @property
+    def _type_affinity(self):
+        return sqltypes.Interval
+
+    def as_generic(self, allow_nulltype=False):
+        return sqltypes.Interval(
+            native=True,
+            second_precision=self.second_precision,
+            day_precision=self.day_precision,
+        )
+
+    @property
+    def python_type(self) -> Type[dt.timedelta]:
+        return dt.timedelta
+
+    def literal_processor(
+        self, dialect: Dialect
+    ) -> Optional[_LiteralProcessorType[dt.timedelta]]:
+        def process(value: dt.timedelta) -> str:
+            return f"NUMTODSINTERVAL({value.total_seconds()}, 'SECOND')"
+
+        return process
+
+
+class TIMESTAMP(sqltypes.TIMESTAMP):
+    """Oracle Database implementation of ``TIMESTAMP``, which supports
+    additional Oracle Database-specific modes
+
+    .. versionadded:: 2.0
+
+    """
+
+    def __init__(self, timezone: bool = False, local_timezone: bool = False):
+        """Construct a new :class:`_oracle.TIMESTAMP`.
+
+        :param timezone: boolean.  Indicates that the TIMESTAMP type should
+         use Oracle Database's ``TIMESTAMP WITH TIME ZONE`` datatype.
+
+        :param local_timezone: boolean.  Indicates that the TIMESTAMP type
+         should use Oracle Database's ``TIMESTAMP WITH LOCAL TIME ZONE``
+         datatype.
+
+
+        """
+        if timezone and local_timezone:
+            raise exc.ArgumentError(
+                "timezone and local_timezone are mutually exclusive"
+            )
+        super().__init__(timezone=timezone)
+        self.local_timezone = local_timezone
+
+
+class ROWID(sqltypes.TypeEngine):
+    """Oracle Database ROWID type.
+
+    When used in a cast() or similar, generates ROWID.
+
+    """
+
+    __visit_name__ = "ROWID"
+
+
+class _OracleBoolean(sqltypes.Boolean):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.NUMBER
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__init__.py
new file mode 100644
index 00000000..88935e20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__init__.py
@@ -0,0 +1,167 @@
+# dialects/postgresql/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from types import ModuleType
+
+from . import array as arraylib  # noqa # keep above base and other dialects
+from . import asyncpg  # noqa
+from . import base
+from . import pg8000  # noqa
+from . import psycopg  # noqa
+from . import psycopg2  # noqa
+from . import psycopg2cffi  # noqa
+from .array import All
+from .array import Any
+from .array import ARRAY
+from .array import array
+from .base import BIGINT
+from .base import BOOLEAN
+from .base import CHAR
+from .base import DATE
+from .base import DOMAIN
+from .base import DOUBLE_PRECISION
+from .base import FLOAT
+from .base import INTEGER
+from .base import NUMERIC
+from .base import REAL
+from .base import SMALLINT
+from .base import TEXT
+from .base import UUID
+from .base import VARCHAR
+from .dml import Insert
+from .dml import insert
+from .ext import aggregate_order_by
+from .ext import array_agg
+from .ext import ExcludeConstraint
+from .ext import phraseto_tsquery
+from .ext import plainto_tsquery
+from .ext import to_tsquery
+from .ext import to_tsvector
+from .ext import ts_headline
+from .ext import websearch_to_tsquery
+from .hstore import HSTORE
+from .hstore import hstore
+from .json import JSON
+from .json import JSONB
+from .json import JSONPATH
+from .named_types import CreateDomainType
+from .named_types import CreateEnumType
+from .named_types import DropDomainType
+from .named_types import DropEnumType
+from .named_types import ENUM
+from .named_types import NamedType
+from .ranges import AbstractMultiRange
+from .ranges import AbstractRange
+from .ranges import AbstractSingleRange
+from .ranges import DATEMULTIRANGE
+from .ranges import DATERANGE
+from .ranges import INT4MULTIRANGE
+from .ranges import INT4RANGE
+from .ranges import INT8MULTIRANGE
+from .ranges import INT8RANGE
+from .ranges import MultiRange
+from .ranges import NUMMULTIRANGE
+from .ranges import NUMRANGE
+from .ranges import Range
+from .ranges import TSMULTIRANGE
+from .ranges import TSRANGE
+from .ranges import TSTZMULTIRANGE
+from .ranges import TSTZRANGE
+from .types import BIT
+from .types import BYTEA
+from .types import CIDR
+from .types import CITEXT
+from .types import INET
+from .types import INTERVAL
+from .types import MACADDR
+from .types import MACADDR8
+from .types import MONEY
+from .types import OID
+from .types import REGCLASS
+from .types import REGCONFIG
+from .types import TIME
+from .types import TIMESTAMP
+from .types import TSQUERY
+from .types import TSVECTOR
+
+
+# Alias psycopg also as psycopg_async
+psycopg_async = type(
+    "psycopg_async", (ModuleType,), {"dialect": psycopg.dialect_async}
+)
+
+base.dialect = dialect = psycopg2.dialect
+
+
+__all__ = (
+    "INTEGER",
+    "BIGINT",
+    "SMALLINT",
+    "VARCHAR",
+    "CHAR",
+    "TEXT",
+    "NUMERIC",
+    "FLOAT",
+    "REAL",
+    "INET",
+    "CIDR",
+    "CITEXT",
+    "UUID",
+    "BIT",
+    "MACADDR",
+    "MACADDR8",
+    "MONEY",
+    "OID",
+    "REGCLASS",
+    "REGCONFIG",
+    "TSQUERY",
+    "TSVECTOR",
+    "DOUBLE_PRECISION",
+    "TIMESTAMP",
+    "TIME",
+    "DATE",
+    "BYTEA",
+    "BOOLEAN",
+    "INTERVAL",
+    "ARRAY",
+    "ENUM",
+    "DOMAIN",
+    "dialect",
+    "array",
+    "HSTORE",
+    "hstore",
+    "INT4RANGE",
+    "INT8RANGE",
+    "NUMRANGE",
+    "DATERANGE",
+    "INT4MULTIRANGE",
+    "INT8MULTIRANGE",
+    "NUMMULTIRANGE",
+    "DATEMULTIRANGE",
+    "TSVECTOR",
+    "TSRANGE",
+    "TSTZRANGE",
+    "TSMULTIRANGE",
+    "TSTZMULTIRANGE",
+    "JSON",
+    "JSONB",
+    "JSONPATH",
+    "Any",
+    "All",
+    "DropEnumType",
+    "DropDomainType",
+    "CreateDomainType",
+    "NamedType",
+    "CreateEnumType",
+    "ExcludeConstraint",
+    "Range",
+    "aggregate_order_by",
+    "array_agg",
+    "insert",
+    "Insert",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..f039f92a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-312.pyc
new file mode 100644
index 00000000..7fbebc1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-312.pyc
new file mode 100644
index 00000000..644fb8a4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-312.pyc
new file mode 100644
index 00000000..d2382c1e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..1e9e81dd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-312.pyc
new file mode 100644
index 00000000..3c817454
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-312.pyc
new file mode 100644
index 00000000..0a2aef4f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-312.pyc
new file mode 100644
index 00000000..3af98de7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-312.pyc
new file mode 100644
index 00000000..a41b63fc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-312.pyc
new file mode 100644
index 00000000..007078cd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-312.pyc
new file mode 100644
index 00000000..9ea3e256
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-312.pyc
new file mode 100644
index 00000000..59558f0e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-312.pyc
new file mode 100644
index 00000000..c1541fc3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..5691f031
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-312.pyc
new file mode 100644
index 00000000..cdafef0c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-312.pyc
new file mode 100644
index 00000000..35d69064
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-312.pyc
new file mode 100644
index 00000000..de6aa1d6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-312.pyc
new file mode 100644
index 00000000..047d03a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..1d8f6a99
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py
new file mode 100644
index 00000000..d827e054
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py
@@ -0,0 +1,187 @@
+# dialects/postgresql/_psycopg_common.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+import decimal
+
+from .array import ARRAY as PGARRAY
+from .base import _DECIMAL_TYPES
+from .base import _FLOAT_TYPES
+from .base import _INT_TYPES
+from .base import PGDialect
+from .base import PGExecutionContext
+from .hstore import HSTORE
+from .pg_catalog import _SpaceVector
+from .pg_catalog import INT2VECTOR
+from .pg_catalog import OIDVECTOR
+from ... import exc
+from ... import types as sqltypes
+from ... import util
+from ...engine import processors
+
+_server_side_id = util.counter()
+
+
+class _PsycopgNumeric(sqltypes.Numeric):
+    def bind_processor(self, dialect):
+        return None
+
+    def result_processor(self, dialect, coltype):
+        if self.asdecimal:
+            if coltype in _FLOAT_TYPES:
+                return processors.to_decimal_processor_factory(
+                    decimal.Decimal, self._effective_decimal_return_scale
+                )
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                # psycopg returns Decimal natively for 1700
+                return None
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+        else:
+            if coltype in _FLOAT_TYPES:
+                # psycopg returns float natively for 701
+                return None
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                return processors.to_float
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+
+
+class _PsycopgFloat(_PsycopgNumeric):
+    __visit_name__ = "float"
+
+
+class _PsycopgHStore(HSTORE):
+    def bind_processor(self, dialect):
+        if dialect._has_native_hstore:
+            return None
+        else:
+            return super().bind_processor(dialect)
+
+    def result_processor(self, dialect, coltype):
+        if dialect._has_native_hstore:
+            return None
+        else:
+            return super().result_processor(dialect, coltype)
+
+
+class _PsycopgARRAY(PGARRAY):
+    render_bind_cast = True
+
+
+class _PsycopgINT2VECTOR(_SpaceVector, INT2VECTOR):
+    pass
+
+
+class _PsycopgOIDVECTOR(_SpaceVector, OIDVECTOR):
+    pass
+
+
+class _PGExecutionContext_common_psycopg(PGExecutionContext):
+    def create_server_side_cursor(self):
+        # use server-side cursors:
+        # psycopg
+        # https://www.psycopg.org/psycopg3/docs/advanced/cursors.html#server-side-cursors
+        # psycopg2
+        # https://www.psycopg.org/docs/usage.html#server-side-cursors
+        ident = "c_%s_%s" % (hex(id(self))[2:], hex(_server_side_id())[2:])
+        return self._dbapi_connection.cursor(ident)
+
+
+class _PGDialect_common_psycopg(PGDialect):
+    supports_statement_cache = True
+    supports_server_side_cursors = True
+
+    default_paramstyle = "pyformat"
+
+    _has_native_hstore = True
+
+    colspecs = util.update_copy(
+        PGDialect.colspecs,
+        {
+            sqltypes.Numeric: _PsycopgNumeric,
+            sqltypes.Float: _PsycopgFloat,
+            HSTORE: _PsycopgHStore,
+            sqltypes.ARRAY: _PsycopgARRAY,
+            INT2VECTOR: _PsycopgINT2VECTOR,
+            OIDVECTOR: _PsycopgOIDVECTOR,
+        },
+    )
+
+    def __init__(
+        self,
+        client_encoding=None,
+        use_native_hstore=True,
+        **kwargs,
+    ):
+        PGDialect.__init__(self, **kwargs)
+        if not use_native_hstore:
+            self._has_native_hstore = False
+        self.use_native_hstore = use_native_hstore
+        self.client_encoding = client_encoding
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args(username="user", database="dbname")
+
+        multihosts, multiports = self._split_multihost_from_url(url)
+
+        if opts or url.query:
+            if not opts:
+                opts = {}
+            if "port" in opts:
+                opts["port"] = int(opts["port"])
+            opts.update(url.query)
+
+            if multihosts:
+                opts["host"] = ",".join(multihosts)
+                comma_ports = ",".join(str(p) if p else "" for p in multiports)
+                if comma_ports:
+                    opts["port"] = comma_ports
+            return ([], opts)
+        else:
+            # no connection arguments whatsoever; psycopg2.connect()
+            # requires that "dsn" be present as a blank string.
+            return ([""], opts)
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return (
+            "AUTOCOMMIT",
+            "READ COMMITTED",
+            "READ UNCOMMITTED",
+            "REPEATABLE READ",
+            "SERIALIZABLE",
+        )
+
+    def set_deferrable(self, connection, value):
+        connection.deferrable = value
+
+    def get_deferrable(self, connection):
+        return connection.deferrable
+
+    def _do_autocommit(self, connection, value):
+        connection.autocommit = value
+
+    def do_ping(self, dbapi_connection):
+        cursor = None
+        before_autocommit = dbapi_connection.autocommit
+
+        if not before_autocommit:
+            dbapi_connection.autocommit = True
+        cursor = dbapi_connection.cursor()
+        try:
+            cursor.execute(self._dialect_specific_select_one)
+        finally:
+            cursor.close()
+            if not before_autocommit and not dbapi_connection.closed:
+                dbapi_connection.autocommit = before_autocommit
+
+        return True
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/array.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/array.py
new file mode 100644
index 00000000..7708769c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/array.py
@@ -0,0 +1,435 @@
+# dialects/postgresql/array.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import re
+from typing import Any
+from typing import Optional
+from typing import TypeVar
+
+from .operators import CONTAINED_BY
+from .operators import CONTAINS
+from .operators import OVERLAP
+from ... import types as sqltypes
+from ... import util
+from ...sql import expression
+from ...sql import operators
+from ...sql._typing import _TypeEngineArgument
+
+
+_T = TypeVar("_T", bound=Any)
+
+
+def Any(other, arrexpr, operator=operators.eq):
+    """A synonym for the ARRAY-level :meth:`.ARRAY.Comparator.any` method.
+    See that method for details.
+
+    """
+
+    return arrexpr.any(other, operator)
+
+
+def All(other, arrexpr, operator=operators.eq):
+    """A synonym for the ARRAY-level :meth:`.ARRAY.Comparator.all` method.
+    See that method for details.
+
+    """
+
+    return arrexpr.all(other, operator)
+
+
+class array(expression.ExpressionClauseList[_T]):
+    """A PostgreSQL ARRAY literal.
+
+    This is used to produce ARRAY literals in SQL expressions, e.g.::
+
+        from sqlalchemy.dialects.postgresql import array
+        from sqlalchemy.dialects import postgresql
+        from sqlalchemy import select, func
+
+        stmt = select(array([1, 2]) + array([3, 4, 5]))
+
+        print(stmt.compile(dialect=postgresql.dialect()))
+
+    Produces the SQL:
+
+    .. sourcecode:: sql
+
+        SELECT ARRAY[%(param_1)s, %(param_2)s] ||
+            ARRAY[%(param_3)s, %(param_4)s, %(param_5)s]) AS anon_1
+
+    An instance of :class:`.array` will always have the datatype
+    :class:`_types.ARRAY`.  The "inner" type of the array is inferred from
+    the values present, unless the ``type_`` keyword argument is passed::
+
+        array(["foo", "bar"], type_=CHAR)
+
+    Multidimensional arrays are produced by nesting :class:`.array` constructs.
+    The dimensionality of the final :class:`_types.ARRAY`
+    type is calculated by
+    recursively adding the dimensions of the inner :class:`_types.ARRAY`
+    type::
+
+        stmt = select(
+            array(
+                [array([1, 2]), array([3, 4]), array([column("q"), column("x")])]
+            )
+        )
+        print(stmt.compile(dialect=postgresql.dialect()))
+
+    Produces:
+
+    .. sourcecode:: sql
+
+        SELECT ARRAY[
+            ARRAY[%(param_1)s, %(param_2)s],
+            ARRAY[%(param_3)s, %(param_4)s],
+            ARRAY[q, x]
+        ] AS anon_1
+
+    .. versionadded:: 1.3.6 added support for multidimensional array literals
+
+    .. seealso::
+
+        :class:`_postgresql.ARRAY`
+
+    """  # noqa: E501
+
+    __visit_name__ = "array"
+
+    stringify_dialect = "postgresql"
+    inherit_cache = True
+
+    def __init__(self, clauses, **kw):
+        type_arg = kw.pop("type_", None)
+        super().__init__(operators.comma_op, *clauses, **kw)
+
+        self._type_tuple = [arg.type for arg in self.clauses]
+
+        main_type = (
+            type_arg
+            if type_arg is not None
+            else self._type_tuple[0] if self._type_tuple else sqltypes.NULLTYPE
+        )
+
+        if isinstance(main_type, ARRAY):
+            self.type = ARRAY(
+                main_type.item_type,
+                dimensions=(
+                    main_type.dimensions + 1
+                    if main_type.dimensions is not None
+                    else 2
+                ),
+            )
+        else:
+            self.type = ARRAY(main_type)
+
+    @property
+    def _select_iterable(self):
+        return (self,)
+
+    def _bind_param(self, operator, obj, _assume_scalar=False, type_=None):
+        if _assume_scalar or operator is operators.getitem:
+            return expression.BindParameter(
+                None,
+                obj,
+                _compared_to_operator=operator,
+                type_=type_,
+                _compared_to_type=self.type,
+                unique=True,
+            )
+
+        else:
+            return array(
+                [
+                    self._bind_param(
+                        operator, o, _assume_scalar=True, type_=type_
+                    )
+                    for o in obj
+                ]
+            )
+
+    def self_group(self, against=None):
+        if against in (operators.any_op, operators.all_op, operators.getitem):
+            return expression.Grouping(self)
+        else:
+            return self
+
+
+class ARRAY(sqltypes.ARRAY):
+    """PostgreSQL ARRAY type.
+
+    The :class:`_postgresql.ARRAY` type is constructed in the same way
+    as the core :class:`_types.ARRAY` type; a member type is required, and a
+    number of dimensions is recommended if the type is to be used for more
+    than one dimension::
+
+        from sqlalchemy.dialects import postgresql
+
+        mytable = Table(
+            "mytable",
+            metadata,
+            Column("data", postgresql.ARRAY(Integer, dimensions=2)),
+        )
+
+    The :class:`_postgresql.ARRAY` type provides all operations defined on the
+    core :class:`_types.ARRAY` type, including support for "dimensions",
+    indexed access, and simple matching such as
+    :meth:`.types.ARRAY.Comparator.any` and
+    :meth:`.types.ARRAY.Comparator.all`.  :class:`_postgresql.ARRAY`
+    class also
+    provides PostgreSQL-specific methods for containment operations, including
+    :meth:`.postgresql.ARRAY.Comparator.contains`
+    :meth:`.postgresql.ARRAY.Comparator.contained_by`, and
+    :meth:`.postgresql.ARRAY.Comparator.overlap`, e.g.::
+
+        mytable.c.data.contains([1, 2])
+
+    Indexed access is one-based by default, to match that of PostgreSQL;
+    for zero-based indexed access, set
+    :paramref:`_postgresql.ARRAY.zero_indexes`.
+
+    Additionally, the :class:`_postgresql.ARRAY`
+    type does not work directly in
+    conjunction with the :class:`.ENUM` type.  For a workaround, see the
+    special type at :ref:`postgresql_array_of_enum`.
+
+    .. container:: topic
+
+        **Detecting Changes in ARRAY columns when using the ORM**
+
+        The :class:`_postgresql.ARRAY` type, when used with the SQLAlchemy ORM,
+        does not detect in-place mutations to the array. In order to detect
+        these, the :mod:`sqlalchemy.ext.mutable` extension must be used, using
+        the :class:`.MutableList` class::
+
+            from sqlalchemy.dialects.postgresql import ARRAY
+            from sqlalchemy.ext.mutable import MutableList
+
+
+            class SomeOrmClass(Base):
+                # ...
+
+                data = Column(MutableList.as_mutable(ARRAY(Integer)))
+
+        This extension will allow "in-place" changes such to the array
+        such as ``.append()`` to produce events which will be detected by the
+        unit of work.  Note that changes to elements **inside** the array,
+        including subarrays that are mutated in place, are **not** detected.
+
+        Alternatively, assigning a new array value to an ORM element that
+        replaces the old one will always trigger a change event.
+
+    .. seealso::
+
+        :class:`_types.ARRAY` - base array type
+
+        :class:`_postgresql.array` - produces a literal array value.
+
+    """
+
+    def __init__(
+        self,
+        item_type: _TypeEngineArgument[Any],
+        as_tuple: bool = False,
+        dimensions: Optional[int] = None,
+        zero_indexes: bool = False,
+    ):
+        """Construct an ARRAY.
+
+        E.g.::
+
+          Column("myarray", ARRAY(Integer))
+
+        Arguments are:
+
+        :param item_type: The data type of items of this array. Note that
+          dimensionality is irrelevant here, so multi-dimensional arrays like
+          ``INTEGER[][]``, are constructed as ``ARRAY(Integer)``, not as
+          ``ARRAY(ARRAY(Integer))`` or such.
+
+        :param as_tuple=False: Specify whether return results
+          should be converted to tuples from lists. DBAPIs such
+          as psycopg2 return lists by default. When tuples are
+          returned, the results are hashable.
+
+        :param dimensions: if non-None, the ARRAY will assume a fixed
+         number of dimensions.  This will cause the DDL emitted for this
+         ARRAY to include the exact number of bracket clauses ``[]``,
+         and will also optimize the performance of the type overall.
+         Note that PG arrays are always implicitly "non-dimensioned",
+         meaning they can store any number of dimensions no matter how
+         they were declared.
+
+        :param zero_indexes=False: when True, index values will be converted
+         between Python zero-based and PostgreSQL one-based indexes, e.g.
+         a value of one will be added to all index values before passing
+         to the database.
+
+        """
+        if isinstance(item_type, ARRAY):
+            raise ValueError(
+                "Do not nest ARRAY types; ARRAY(basetype) "
+                "handles multi-dimensional arrays of basetype"
+            )
+        if isinstance(item_type, type):
+            item_type = item_type()
+        self.item_type = item_type
+        self.as_tuple = as_tuple
+        self.dimensions = dimensions
+        self.zero_indexes = zero_indexes
+
+    class Comparator(sqltypes.ARRAY.Comparator):
+        """Define comparison operations for :class:`_types.ARRAY`.
+
+        Note that these operations are in addition to those provided
+        by the base :class:`.types.ARRAY.Comparator` class, including
+        :meth:`.types.ARRAY.Comparator.any` and
+        :meth:`.types.ARRAY.Comparator.all`.
+
+        """
+
+        def contains(self, other, **kwargs):
+            """Boolean expression.  Test if elements are a superset of the
+            elements of the argument array expression.
+
+            kwargs may be ignored by this operator but are required for API
+            conformance.
+            """
+            return self.operate(CONTAINS, other, result_type=sqltypes.Boolean)
+
+        def contained_by(self, other):
+            """Boolean expression.  Test if elements are a proper subset of the
+            elements of the argument array expression.
+            """
+            return self.operate(
+                CONTAINED_BY, other, result_type=sqltypes.Boolean
+            )
+
+        def overlap(self, other):
+            """Boolean expression.  Test if array has elements in common with
+            an argument array expression.
+            """
+            return self.operate(OVERLAP, other, result_type=sqltypes.Boolean)
+
+    comparator_factory = Comparator
+
+    @property
+    def hashable(self):
+        return self.as_tuple
+
+    @property
+    def python_type(self):
+        return list
+
+    def compare_values(self, x, y):
+        return x == y
+
+    @util.memoized_property
+    def _against_native_enum(self):
+        return (
+            isinstance(self.item_type, sqltypes.Enum)
+            and self.item_type.native_enum
+        )
+
+    def literal_processor(self, dialect):
+        item_proc = self.item_type.dialect_impl(dialect).literal_processor(
+            dialect
+        )
+        if item_proc is None:
+            return None
+
+        def to_str(elements):
+            return f"ARRAY[{', '.join(elements)}]"
+
+        def process(value):
+            inner = self._apply_item_processor(
+                value, item_proc, self.dimensions, to_str
+            )
+            return inner
+
+        return process
+
+    def bind_processor(self, dialect):
+        item_proc = self.item_type.dialect_impl(dialect).bind_processor(
+            dialect
+        )
+
+        def process(value):
+            if value is None:
+                return value
+            else:
+                return self._apply_item_processor(
+                    value, item_proc, self.dimensions, list
+                )
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        item_proc = self.item_type.dialect_impl(dialect).result_processor(
+            dialect, coltype
+        )
+
+        def process(value):
+            if value is None:
+                return value
+            else:
+                return self._apply_item_processor(
+                    value,
+                    item_proc,
+                    self.dimensions,
+                    tuple if self.as_tuple else list,
+                )
+
+        if self._against_native_enum:
+            super_rp = process
+            pattern = re.compile(r"^{(.*)}$")
+
+            def handle_raw_string(value):
+                inner = pattern.match(value).group(1)
+                return _split_enum_values(inner)
+
+            def process(value):
+                if value is None:
+                    return value
+                # isinstance(value, str) is required to handle
+                # the case where a TypeDecorator for and Array of Enum is
+                # used like was required in sa < 1.3.17
+                return super_rp(
+                    handle_raw_string(value)
+                    if isinstance(value, str)
+                    else value
+                )
+
+        return process
+
+
+def _split_enum_values(array_string):
+    if '"' not in array_string:
+        # no escape char is present so it can just split on the comma
+        return array_string.split(",") if array_string else []
+
+    # handles quoted strings from:
+    # r'abc,"quoted","also\\\\quoted", "quoted, comma", "esc \" quot", qpr'
+    # returns
+    # ['abc', 'quoted', 'also\\quoted', 'quoted, comma', 'esc " quot', 'qpr']
+    text = array_string.replace(r"\"", "_$ESC_QUOTE$_")
+    text = text.replace(r"\\", "\\")
+    result = []
+    on_quotes = re.split(r'(")', text)
+    in_quotes = False
+    for tok in on_quotes:
+        if tok == '"':
+            in_quotes = not in_quotes
+        elif in_quotes:
+            result.append(tok.replace("_$ESC_QUOTE$_", '"'))
+        else:
+            result.extend(re.findall(r"([^\s,]+),?", tok))
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py
new file mode 100644
index 00000000..09689212
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py
@@ -0,0 +1,1287 @@
+# dialects/postgresql/asyncpg.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors <see AUTHORS
+# file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql+asyncpg
+    :name: asyncpg
+    :dbapi: asyncpg
+    :connectstring: postgresql+asyncpg://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://magicstack.github.io/asyncpg/
+
+The asyncpg dialect is SQLAlchemy's first Python asyncio dialect.
+
+Using a special asyncio mediation layer, the asyncpg dialect is usable
+as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>`
+extension package.
+
+This dialect should normally be used only with the
+:func:`_asyncio.create_async_engine` engine creation function::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:pass@hostname/dbname"
+    )
+
+.. versionadded:: 1.4
+
+.. note::
+
+    By default asyncpg does not decode the ``json`` and ``jsonb`` types and
+    returns them as strings. SQLAlchemy sets default type decoder for ``json``
+    and ``jsonb`` types using the python builtin ``json.loads`` function.
+    The json implementation used can be changed by setting the attribute
+    ``json_deserializer`` when creating the engine with
+    :func:`create_engine` or :func:`create_async_engine`.
+
+.. _asyncpg_multihost:
+
+Multihost Connections
+--------------------------
+
+The asyncpg dialect features support for multiple fallback hosts in the
+same way as that of the psycopg2 and psycopg dialects.  The
+syntax is the same,
+using ``host=<host>:<port>`` combinations as additional query string arguments;
+however, there is no default port, so all hosts must have a complete port number
+present, otherwise an exception is raised::
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:password@/dbname?host=HostA:5432&host=HostB:5432&host=HostC:5432"
+    )
+
+For complete background on this syntax, see :ref:`psycopg2_multi_host`.
+
+.. versionadded:: 2.0.18
+
+.. seealso::
+
+    :ref:`psycopg2_multi_host`
+
+.. _asyncpg_prepared_statement_cache:
+
+Prepared Statement Cache
+--------------------------
+
+The asyncpg SQLAlchemy dialect makes use of ``asyncpg.connection.prepare()``
+for all statements.   The prepared statement objects are cached after
+construction which appears to grant a 10% or more performance improvement for
+statement invocation.   The cache is on a per-DBAPI connection basis, which
+means that the primary storage for prepared statements is within DBAPI
+connections pooled within the connection pool.   The size of this cache
+defaults to 100 statements per DBAPI connection and may be adjusted using the
+``prepared_statement_cache_size`` DBAPI argument (note that while this argument
+is implemented by SQLAlchemy, it is part of the DBAPI emulation portion of the
+asyncpg dialect, therefore is handled as a DBAPI argument, not a dialect
+argument)::
+
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:pass@hostname/dbname?prepared_statement_cache_size=500"
+    )
+
+To disable the prepared statement cache, use a value of zero::
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:pass@hostname/dbname?prepared_statement_cache_size=0"
+    )
+
+.. versionadded:: 1.4.0b2 Added ``prepared_statement_cache_size`` for asyncpg.
+
+
+.. warning::  The ``asyncpg`` database driver necessarily uses caches for
+   PostgreSQL type OIDs, which become stale when custom PostgreSQL datatypes
+   such as ``ENUM`` objects are changed via DDL operations.   Additionally,
+   prepared statements themselves which are optionally cached by SQLAlchemy's
+   driver as described above may also become "stale" when DDL has been emitted
+   to the PostgreSQL database which modifies the tables or other objects
+   involved in a particular prepared statement.
+
+   The SQLAlchemy asyncpg dialect will invalidate these caches within its local
+   process when statements that represent DDL are emitted on a local
+   connection, but this is only controllable within a single Python process /
+   database engine.     If DDL changes are made from other database engines
+   and/or processes, a running application may encounter asyncpg exceptions
+   ``InvalidCachedStatementError`` and/or ``InternalServerError("cache lookup
+   failed for type <oid>")`` if it refers to pooled database connections which
+   operated upon the previous structures. The SQLAlchemy asyncpg dialect will
+   recover from these error cases when the driver raises these exceptions by
+   clearing its internal caches as well as those of the asyncpg driver in
+   response to them, but cannot prevent them from being raised in the first
+   place if the cached prepared statement or asyncpg type caches have gone
+   stale, nor can it retry the statement as the PostgreSQL transaction is
+   invalidated when these errors occur.
+
+.. _asyncpg_prepared_statement_name:
+
+Prepared Statement Name with PGBouncer
+--------------------------------------
+
+By default, asyncpg enumerates prepared statements in numeric order, which
+can lead to errors if a name has already been taken for another prepared
+statement. This issue can arise if your application uses database proxies
+such as PgBouncer to handle connections. One possible workaround is to
+use dynamic prepared statement names, which asyncpg now supports through
+an optional ``name`` value for the statement name. This allows you to
+generate your own unique names that won't conflict with existing ones.
+To achieve this, you can provide a function that will be called every time
+a prepared statement is prepared::
+
+    from uuid import uuid4
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:pass@somepgbouncer/dbname",
+        poolclass=NullPool,
+        connect_args={
+            "prepared_statement_name_func": lambda: f"__asyncpg_{uuid4()}__",
+        },
+    )
+
+.. seealso::
+
+   https://github.com/MagicStack/asyncpg/issues/837
+
+   https://github.com/sqlalchemy/sqlalchemy/issues/6467
+
+.. warning:: When using PGBouncer, to prevent a buildup of useless prepared statements in
+   your application, it's important to use the :class:`.NullPool` pool
+   class, and to configure PgBouncer to use `DISCARD <https://www.postgresql.org/docs/current/sql-discard.html>`_
+   when returning connections.  The DISCARD command is used to release resources held by the db connection,
+   including prepared statements. Without proper setup, prepared statements can
+   accumulate quickly and cause performance issues.
+
+Disabling the PostgreSQL JIT to improve ENUM datatype handling
+---------------------------------------------------------------
+
+Asyncpg has an `issue <https://github.com/MagicStack/asyncpg/issues/727>`_ when
+using PostgreSQL ENUM datatypes, where upon the creation of new database
+connections, an expensive query may be emitted in order to retrieve metadata
+regarding custom types which has been shown to negatively affect performance.
+To mitigate this issue, the PostgreSQL "jit" setting may be disabled from the
+client using this setting passed to :func:`_asyncio.create_async_engine`::
+
+    engine = create_async_engine(
+        "postgresql+asyncpg://user:password@localhost/tmp",
+        connect_args={"server_settings": {"jit": "off"}},
+    )
+
+.. seealso::
+
+    https://github.com/MagicStack/asyncpg/issues/727
+
+"""  # noqa
+
+from __future__ import annotations
+
+from collections import deque
+import decimal
+import json as _py_json
+import re
+import time
+
+from . import json
+from . import ranges
+from .array import ARRAY as PGARRAY
+from .base import _DECIMAL_TYPES
+from .base import _FLOAT_TYPES
+from .base import _INT_TYPES
+from .base import ENUM
+from .base import INTERVAL
+from .base import OID
+from .base import PGCompiler
+from .base import PGDialect
+from .base import PGExecutionContext
+from .base import PGIdentifierPreparer
+from .base import REGCLASS
+from .base import REGCONFIG
+from .types import BIT
+from .types import BYTEA
+from .types import CITEXT
+from ... import exc
+from ... import pool
+from ... import util
+from ...engine import AdaptedConnection
+from ...engine import processors
+from ...sql import sqltypes
+from ...util.concurrency import asyncio
+from ...util.concurrency import await_fallback
+from ...util.concurrency import await_only
+
+
+class AsyncpgARRAY(PGARRAY):
+    render_bind_cast = True
+
+
+class AsyncpgString(sqltypes.String):
+    render_bind_cast = True
+
+
+class AsyncpgREGCONFIG(REGCONFIG):
+    render_bind_cast = True
+
+
+class AsyncpgTime(sqltypes.Time):
+    render_bind_cast = True
+
+
+class AsyncpgBit(BIT):
+    render_bind_cast = True
+
+
+class AsyncpgByteA(BYTEA):
+    render_bind_cast = True
+
+
+class AsyncpgDate(sqltypes.Date):
+    render_bind_cast = True
+
+
+class AsyncpgDateTime(sqltypes.DateTime):
+    render_bind_cast = True
+
+
+class AsyncpgBoolean(sqltypes.Boolean):
+    render_bind_cast = True
+
+
+class AsyncPgInterval(INTERVAL):
+    render_bind_cast = True
+
+    @classmethod
+    def adapt_emulated_to_native(cls, interval, **kw):
+        return AsyncPgInterval(precision=interval.second_precision)
+
+
+class AsyncPgEnum(ENUM):
+    render_bind_cast = True
+
+
+class AsyncpgInteger(sqltypes.Integer):
+    render_bind_cast = True
+
+
+class AsyncpgSmallInteger(sqltypes.SmallInteger):
+    render_bind_cast = True
+
+
+class AsyncpgBigInteger(sqltypes.BigInteger):
+    render_bind_cast = True
+
+
+class AsyncpgJSON(json.JSON):
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class AsyncpgJSONB(json.JSONB):
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class AsyncpgJSONIndexType(sqltypes.JSON.JSONIndexType):
+    pass
+
+
+class AsyncpgJSONIntIndexType(sqltypes.JSON.JSONIntIndexType):
+    __visit_name__ = "json_int_index"
+
+    render_bind_cast = True
+
+
+class AsyncpgJSONStrIndexType(sqltypes.JSON.JSONStrIndexType):
+    __visit_name__ = "json_str_index"
+
+    render_bind_cast = True
+
+
+class AsyncpgJSONPathType(json.JSONPathType):
+    def bind_processor(self, dialect):
+        def process(value):
+            if isinstance(value, str):
+                # If it's already a string assume that it's in json path
+                # format. This allows using cast with json paths literals
+                return value
+            elif value:
+                tokens = [str(elem) for elem in value]
+                return tokens
+            else:
+                return []
+
+        return process
+
+
+class AsyncpgNumeric(sqltypes.Numeric):
+    render_bind_cast = True
+
+    def bind_processor(self, dialect):
+        return None
+
+    def result_processor(self, dialect, coltype):
+        if self.asdecimal:
+            if coltype in _FLOAT_TYPES:
+                return processors.to_decimal_processor_factory(
+                    decimal.Decimal, self._effective_decimal_return_scale
+                )
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                # pg8000 returns Decimal natively for 1700
+                return None
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+        else:
+            if coltype in _FLOAT_TYPES:
+                # pg8000 returns float natively for 701
+                return None
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                return processors.to_float
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+
+
+class AsyncpgFloat(AsyncpgNumeric, sqltypes.Float):
+    __visit_name__ = "float"
+    render_bind_cast = True
+
+
+class AsyncpgREGCLASS(REGCLASS):
+    render_bind_cast = True
+
+
+class AsyncpgOID(OID):
+    render_bind_cast = True
+
+
+class AsyncpgCHAR(sqltypes.CHAR):
+    render_bind_cast = True
+
+
+class _AsyncpgRange(ranges.AbstractSingleRangeImpl):
+    def bind_processor(self, dialect):
+        asyncpg_Range = dialect.dbapi.asyncpg.Range
+
+        def to_range(value):
+            if isinstance(value, ranges.Range):
+                value = asyncpg_Range(
+                    value.lower,
+                    value.upper,
+                    lower_inc=value.bounds[0] == "[",
+                    upper_inc=value.bounds[1] == "]",
+                    empty=value.empty,
+                )
+            return value
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range(value):
+            if value is not None:
+                empty = value.isempty
+                value = ranges.Range(
+                    value.lower,
+                    value.upper,
+                    bounds=f"{'[' if empty or value.lower_inc else '('}"  # type: ignore  # noqa: E501
+                    f"{']' if not empty and value.upper_inc else ')'}",
+                    empty=empty,
+                )
+            return value
+
+        return to_range
+
+
+class _AsyncpgMultiRange(ranges.AbstractMultiRangeImpl):
+    def bind_processor(self, dialect):
+        asyncpg_Range = dialect.dbapi.asyncpg.Range
+
+        NoneType = type(None)
+
+        def to_range(value):
+            if isinstance(value, (str, NoneType)):
+                return value
+
+            def to_range(value):
+                if isinstance(value, ranges.Range):
+                    value = asyncpg_Range(
+                        value.lower,
+                        value.upper,
+                        lower_inc=value.bounds[0] == "[",
+                        upper_inc=value.bounds[1] == "]",
+                        empty=value.empty,
+                    )
+                return value
+
+            return [to_range(element) for element in value]
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range_array(value):
+            def to_range(rvalue):
+                if rvalue is not None:
+                    empty = rvalue.isempty
+                    rvalue = ranges.Range(
+                        rvalue.lower,
+                        rvalue.upper,
+                        bounds=f"{'[' if empty or rvalue.lower_inc else '('}"  # type: ignore  # noqa: E501
+                        f"{']' if not empty and rvalue.upper_inc else ')'}",
+                        empty=empty,
+                    )
+                return rvalue
+
+            if value is not None:
+                value = ranges.MultiRange(to_range(elem) for elem in value)
+
+            return value
+
+        return to_range_array
+
+
+class PGExecutionContext_asyncpg(PGExecutionContext):
+    def handle_dbapi_exception(self, e):
+        if isinstance(
+            e,
+            (
+                self.dialect.dbapi.InvalidCachedStatementError,
+                self.dialect.dbapi.InternalServerError,
+            ),
+        ):
+            self.dialect._invalidate_schema_cache()
+
+    def pre_exec(self):
+        if self.isddl:
+            self.dialect._invalidate_schema_cache()
+
+        self.cursor._invalidate_schema_cache_asof = (
+            self.dialect._invalidate_schema_cache_asof
+        )
+
+        if not self.compiled:
+            return
+
+    def create_server_side_cursor(self):
+        return self._dbapi_connection.cursor(server_side=True)
+
+
+class PGCompiler_asyncpg(PGCompiler):
+    pass
+
+
+class PGIdentifierPreparer_asyncpg(PGIdentifierPreparer):
+    pass
+
+
+class AsyncAdapt_asyncpg_cursor:
+    __slots__ = (
+        "_adapt_connection",
+        "_connection",
+        "_rows",
+        "description",
+        "arraysize",
+        "rowcount",
+        "_cursor",
+        "_invalidate_schema_cache_asof",
+    )
+
+    server_side = False
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self._rows = deque()
+        self._cursor = None
+        self.description = None
+        self.arraysize = 1
+        self.rowcount = -1
+        self._invalidate_schema_cache_asof = 0
+
+    def close(self):
+        self._rows.clear()
+
+    def _handle_exception(self, error):
+        self._adapt_connection._handle_exception(error)
+
+    async def _prepare_and_execute(self, operation, parameters):
+        adapt_connection = self._adapt_connection
+
+        async with adapt_connection._execute_mutex:
+            if not adapt_connection._started:
+                await adapt_connection._start_transaction()
+
+            if parameters is None:
+                parameters = ()
+
+            try:
+                prepared_stmt, attributes = await adapt_connection._prepare(
+                    operation, self._invalidate_schema_cache_asof
+                )
+
+                if attributes:
+                    self.description = [
+                        (
+                            attr.name,
+                            attr.type.oid,
+                            None,
+                            None,
+                            None,
+                            None,
+                            None,
+                        )
+                        for attr in attributes
+                    ]
+                else:
+                    self.description = None
+
+                if self.server_side:
+                    self._cursor = await prepared_stmt.cursor(*parameters)
+                    self.rowcount = -1
+                else:
+                    self._rows = deque(await prepared_stmt.fetch(*parameters))
+                    status = prepared_stmt.get_statusmsg()
+
+                    reg = re.match(
+                        r"(?:SELECT|UPDATE|DELETE|INSERT \d+) (\d+)",
+                        status or "",
+                    )
+                    if reg:
+                        self.rowcount = int(reg.group(1))
+                    else:
+                        self.rowcount = -1
+
+            except Exception as error:
+                self._handle_exception(error)
+
+    async def _executemany(self, operation, seq_of_parameters):
+        adapt_connection = self._adapt_connection
+
+        self.description = None
+        async with adapt_connection._execute_mutex:
+            await adapt_connection._check_type_cache_invalidation(
+                self._invalidate_schema_cache_asof
+            )
+
+            if not adapt_connection._started:
+                await adapt_connection._start_transaction()
+
+            try:
+                return await self._connection.executemany(
+                    operation, seq_of_parameters
+                )
+            except Exception as error:
+                self._handle_exception(error)
+
+    def execute(self, operation, parameters=None):
+        self._adapt_connection.await_(
+            self._prepare_and_execute(operation, parameters)
+        )
+
+    def executemany(self, operation, seq_of_parameters):
+        return self._adapt_connection.await_(
+            self._executemany(operation, seq_of_parameters)
+        )
+
+    def setinputsizes(self, *inputsizes):
+        raise NotImplementedError()
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_asyncpg_ss_cursor(AsyncAdapt_asyncpg_cursor):
+    server_side = True
+    __slots__ = ("_rowbuffer",)
+
+    def __init__(self, adapt_connection):
+        super().__init__(adapt_connection)
+        self._rowbuffer = deque()
+
+    def close(self):
+        self._cursor = None
+        self._rowbuffer.clear()
+
+    def _buffer_rows(self):
+        assert self._cursor is not None
+        new_rows = self._adapt_connection.await_(self._cursor.fetch(50))
+        self._rowbuffer.extend(new_rows)
+
+    def __aiter__(self):
+        return self
+
+    async def __anext__(self):
+        while True:
+            while self._rowbuffer:
+                yield self._rowbuffer.popleft()
+
+            self._buffer_rows()
+            if not self._rowbuffer:
+                break
+
+    def fetchone(self):
+        if not self._rowbuffer:
+            self._buffer_rows()
+            if not self._rowbuffer:
+                return None
+        return self._rowbuffer.popleft()
+
+    def fetchmany(self, size=None):
+        if size is None:
+            return self.fetchall()
+
+        if not self._rowbuffer:
+            self._buffer_rows()
+
+        assert self._cursor is not None
+        rb = self._rowbuffer
+        lb = len(rb)
+        if size > lb:
+            rb.extend(
+                self._adapt_connection.await_(self._cursor.fetch(size - lb))
+            )
+
+        return [rb.popleft() for _ in range(min(size, len(rb)))]
+
+    def fetchall(self):
+        ret = list(self._rowbuffer)
+        ret.extend(self._adapt_connection.await_(self._all()))
+        self._rowbuffer.clear()
+        return ret
+
+    async def _all(self):
+        rows = []
+
+        # TODO: looks like we have to hand-roll some kind of batching here.
+        # hardcoding for the moment but this should be improved.
+        while True:
+            batch = await self._cursor.fetch(1000)
+            if batch:
+                rows.extend(batch)
+                continue
+            else:
+                break
+        return rows
+
+    def executemany(self, operation, seq_of_parameters):
+        raise NotImplementedError(
+            "server side cursor doesn't support executemany yet"
+        )
+
+
+class AsyncAdapt_asyncpg_connection(AdaptedConnection):
+    __slots__ = (
+        "dbapi",
+        "isolation_level",
+        "_isolation_setting",
+        "readonly",
+        "deferrable",
+        "_transaction",
+        "_started",
+        "_prepared_statement_cache",
+        "_prepared_statement_name_func",
+        "_invalidate_schema_cache_asof",
+        "_execute_mutex",
+    )
+
+    await_ = staticmethod(await_only)
+
+    def __init__(
+        self,
+        dbapi,
+        connection,
+        prepared_statement_cache_size=100,
+        prepared_statement_name_func=None,
+    ):
+        self.dbapi = dbapi
+        self._connection = connection
+        self.isolation_level = self._isolation_setting = None
+        self.readonly = False
+        self.deferrable = False
+        self._transaction = None
+        self._started = False
+        self._invalidate_schema_cache_asof = time.time()
+        self._execute_mutex = asyncio.Lock()
+
+        if prepared_statement_cache_size:
+            self._prepared_statement_cache = util.LRUCache(
+                prepared_statement_cache_size
+            )
+        else:
+            self._prepared_statement_cache = None
+
+        if prepared_statement_name_func:
+            self._prepared_statement_name_func = prepared_statement_name_func
+        else:
+            self._prepared_statement_name_func = self._default_name_func
+
+    async def _check_type_cache_invalidation(self, invalidate_timestamp):
+        if invalidate_timestamp > self._invalidate_schema_cache_asof:
+            await self._connection.reload_schema_state()
+            self._invalidate_schema_cache_asof = invalidate_timestamp
+
+    async def _prepare(self, operation, invalidate_timestamp):
+        await self._check_type_cache_invalidation(invalidate_timestamp)
+
+        cache = self._prepared_statement_cache
+        if cache is None:
+            prepared_stmt = await self._connection.prepare(
+                operation, name=self._prepared_statement_name_func()
+            )
+            attributes = prepared_stmt.get_attributes()
+            return prepared_stmt, attributes
+
+        # asyncpg uses a type cache for the "attributes" which seems to go
+        # stale independently of the PreparedStatement itself, so place that
+        # collection in the cache as well.
+        if operation in cache:
+            prepared_stmt, attributes, cached_timestamp = cache[operation]
+
+            # preparedstatements themselves also go stale for certain DDL
+            # changes such as size of a VARCHAR changing, so there is also
+            # a cross-connection invalidation timestamp
+            if cached_timestamp > invalidate_timestamp:
+                return prepared_stmt, attributes
+
+        prepared_stmt = await self._connection.prepare(
+            operation, name=self._prepared_statement_name_func()
+        )
+        attributes = prepared_stmt.get_attributes()
+        cache[operation] = (prepared_stmt, attributes, time.time())
+
+        return prepared_stmt, attributes
+
+    def _handle_exception(self, error):
+        if self._connection.is_closed():
+            self._transaction = None
+            self._started = False
+
+        if not isinstance(error, AsyncAdapt_asyncpg_dbapi.Error):
+            exception_mapping = self.dbapi._asyncpg_error_translate
+
+            for super_ in type(error).__mro__:
+                if super_ in exception_mapping:
+                    translated_error = exception_mapping[super_](
+                        "%s: %s" % (type(error), error)
+                    )
+                    translated_error.pgcode = translated_error.sqlstate = (
+                        getattr(error, "sqlstate", None)
+                    )
+                    raise translated_error from error
+            else:
+                raise error
+        else:
+            raise error
+
+    @property
+    def autocommit(self):
+        return self.isolation_level == "autocommit"
+
+    @autocommit.setter
+    def autocommit(self, value):
+        if value:
+            self.isolation_level = "autocommit"
+        else:
+            self.isolation_level = self._isolation_setting
+
+    def ping(self):
+        try:
+            _ = self.await_(self._async_ping())
+        except Exception as error:
+            self._handle_exception(error)
+
+    async def _async_ping(self):
+        if self._transaction is None and self.isolation_level != "autocommit":
+            # create a tranasction explicitly to support pgbouncer
+            # transaction mode.   See #10226
+            tr = self._connection.transaction()
+            await tr.start()
+            try:
+                await self._connection.fetchrow(";")
+            finally:
+                await tr.rollback()
+        else:
+            await self._connection.fetchrow(";")
+
+    def set_isolation_level(self, level):
+        if self._started:
+            self.rollback()
+        self.isolation_level = self._isolation_setting = level
+
+    async def _start_transaction(self):
+        if self.isolation_level == "autocommit":
+            return
+
+        try:
+            self._transaction = self._connection.transaction(
+                isolation=self.isolation_level,
+                readonly=self.readonly,
+                deferrable=self.deferrable,
+            )
+            await self._transaction.start()
+        except Exception as error:
+            self._handle_exception(error)
+        else:
+            self._started = True
+
+    def cursor(self, server_side=False):
+        if server_side:
+            return AsyncAdapt_asyncpg_ss_cursor(self)
+        else:
+            return AsyncAdapt_asyncpg_cursor(self)
+
+    async def _rollback_and_discard(self):
+        try:
+            await self._transaction.rollback()
+        finally:
+            # if asyncpg .rollback() was actually called, then whether or
+            # not it raised or succeeded, the transation is done, discard it
+            self._transaction = None
+            self._started = False
+
+    async def _commit_and_discard(self):
+        try:
+            await self._transaction.commit()
+        finally:
+            # if asyncpg .commit() was actually called, then whether or
+            # not it raised or succeeded, the transation is done, discard it
+            self._transaction = None
+            self._started = False
+
+    def rollback(self):
+        if self._started:
+            try:
+                self.await_(self._rollback_and_discard())
+                self._transaction = None
+                self._started = False
+            except Exception as error:
+                # don't dereference asyncpg transaction if we didn't
+                # actually try to call rollback() on it
+                self._handle_exception(error)
+
+    def commit(self):
+        if self._started:
+            try:
+                self.await_(self._commit_and_discard())
+                self._transaction = None
+                self._started = False
+            except Exception as error:
+                # don't dereference asyncpg transaction if we didn't
+                # actually try to call commit() on it
+                self._handle_exception(error)
+
+    def close(self):
+        self.rollback()
+
+        self.await_(self._connection.close())
+
+    def terminate(self):
+        if util.concurrency.in_greenlet():
+            # in a greenlet; this is the connection was invalidated
+            # case.
+            try:
+                # try to gracefully close; see #10717
+                # timeout added in asyncpg 0.14.0 December 2017
+                self.await_(asyncio.shield(self._connection.close(timeout=2)))
+            except (
+                asyncio.TimeoutError,
+                asyncio.CancelledError,
+                OSError,
+                self.dbapi.asyncpg.PostgresError,
+            ):
+                # in the case where we are recycling an old connection
+                # that may have already been disconnected, close() will
+                # fail with the above timeout.  in this case, terminate
+                # the connection without any further waiting.
+                # see issue #8419
+                self._connection.terminate()
+        else:
+            # not in a greenlet; this is the gc cleanup case
+            self._connection.terminate()
+        self._started = False
+
+    @staticmethod
+    def _default_name_func():
+        return None
+
+
+class AsyncAdaptFallback_asyncpg_connection(AsyncAdapt_asyncpg_connection):
+    __slots__ = ()
+
+    await_ = staticmethod(await_fallback)
+
+
+class AsyncAdapt_asyncpg_dbapi:
+    def __init__(self, asyncpg):
+        self.asyncpg = asyncpg
+        self.paramstyle = "numeric_dollar"
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop("async_creator_fn", self.asyncpg.connect)
+        prepared_statement_cache_size = kw.pop(
+            "prepared_statement_cache_size", 100
+        )
+        prepared_statement_name_func = kw.pop(
+            "prepared_statement_name_func", None
+        )
+
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_asyncpg_connection(
+                self,
+                await_fallback(creator_fn(*arg, **kw)),
+                prepared_statement_cache_size=prepared_statement_cache_size,
+                prepared_statement_name_func=prepared_statement_name_func,
+            )
+        else:
+            return AsyncAdapt_asyncpg_connection(
+                self,
+                await_only(creator_fn(*arg, **kw)),
+                prepared_statement_cache_size=prepared_statement_cache_size,
+                prepared_statement_name_func=prepared_statement_name_func,
+            )
+
+    class Error(Exception):
+        pass
+
+    class Warning(Exception):  # noqa
+        pass
+
+    class InterfaceError(Error):
+        pass
+
+    class DatabaseError(Error):
+        pass
+
+    class InternalError(DatabaseError):
+        pass
+
+    class OperationalError(DatabaseError):
+        pass
+
+    class ProgrammingError(DatabaseError):
+        pass
+
+    class IntegrityError(DatabaseError):
+        pass
+
+    class DataError(DatabaseError):
+        pass
+
+    class NotSupportedError(DatabaseError):
+        pass
+
+    class InternalServerError(InternalError):
+        pass
+
+    class InvalidCachedStatementError(NotSupportedError):
+        def __init__(self, message):
+            super().__init__(
+                message + " (SQLAlchemy asyncpg dialect will now invalidate "
+                "all prepared caches in response to this exception)",
+            )
+
+    # pep-249 datatype placeholders.  As of SQLAlchemy 2.0 these aren't
+    # used, however the test suite looks for these in a few cases.
+    STRING = util.symbol("STRING")
+    NUMBER = util.symbol("NUMBER")
+    DATETIME = util.symbol("DATETIME")
+
+    @util.memoized_property
+    def _asyncpg_error_translate(self):
+        import asyncpg
+
+        return {
+            asyncpg.exceptions.IntegrityConstraintViolationError: self.IntegrityError,  # noqa: E501
+            asyncpg.exceptions.PostgresError: self.Error,
+            asyncpg.exceptions.SyntaxOrAccessError: self.ProgrammingError,
+            asyncpg.exceptions.InterfaceError: self.InterfaceError,
+            asyncpg.exceptions.InvalidCachedStatementError: self.InvalidCachedStatementError,  # noqa: E501
+            asyncpg.exceptions.InternalServerError: self.InternalServerError,
+        }
+
+    def Binary(self, value):
+        return value
+
+
+class PGDialect_asyncpg(PGDialect):
+    driver = "asyncpg"
+    supports_statement_cache = True
+
+    supports_server_side_cursors = True
+
+    render_bind_cast = True
+    has_terminate = True
+
+    default_paramstyle = "numeric_dollar"
+    supports_sane_multi_rowcount = False
+    execution_ctx_cls = PGExecutionContext_asyncpg
+    statement_compiler = PGCompiler_asyncpg
+    preparer = PGIdentifierPreparer_asyncpg
+
+    colspecs = util.update_copy(
+        PGDialect.colspecs,
+        {
+            sqltypes.String: AsyncpgString,
+            sqltypes.ARRAY: AsyncpgARRAY,
+            BIT: AsyncpgBit,
+            CITEXT: CITEXT,
+            REGCONFIG: AsyncpgREGCONFIG,
+            sqltypes.Time: AsyncpgTime,
+            sqltypes.Date: AsyncpgDate,
+            sqltypes.DateTime: AsyncpgDateTime,
+            sqltypes.Interval: AsyncPgInterval,
+            INTERVAL: AsyncPgInterval,
+            sqltypes.Boolean: AsyncpgBoolean,
+            sqltypes.Integer: AsyncpgInteger,
+            sqltypes.SmallInteger: AsyncpgSmallInteger,
+            sqltypes.BigInteger: AsyncpgBigInteger,
+            sqltypes.Numeric: AsyncpgNumeric,
+            sqltypes.Float: AsyncpgFloat,
+            sqltypes.JSON: AsyncpgJSON,
+            sqltypes.LargeBinary: AsyncpgByteA,
+            json.JSONB: AsyncpgJSONB,
+            sqltypes.JSON.JSONPathType: AsyncpgJSONPathType,
+            sqltypes.JSON.JSONIndexType: AsyncpgJSONIndexType,
+            sqltypes.JSON.JSONIntIndexType: AsyncpgJSONIntIndexType,
+            sqltypes.JSON.JSONStrIndexType: AsyncpgJSONStrIndexType,
+            sqltypes.Enum: AsyncPgEnum,
+            OID: AsyncpgOID,
+            REGCLASS: AsyncpgREGCLASS,
+            sqltypes.CHAR: AsyncpgCHAR,
+            ranges.AbstractSingleRange: _AsyncpgRange,
+            ranges.AbstractMultiRange: _AsyncpgMultiRange,
+        },
+    )
+    is_async = True
+    _invalidate_schema_cache_asof = 0
+
+    def _invalidate_schema_cache(self):
+        self._invalidate_schema_cache_asof = time.time()
+
+    @util.memoized_property
+    def _dbapi_version(self):
+        if self.dbapi and hasattr(self.dbapi, "__version__"):
+            return tuple(
+                [
+                    int(x)
+                    for x in re.findall(
+                        r"(\d+)(?:[-\.]?|$)", self.dbapi.__version__
+                    )
+                ]
+            )
+        else:
+            return (99, 99, 99)
+
+    @classmethod
+    def import_dbapi(cls):
+        return AsyncAdapt_asyncpg_dbapi(__import__("asyncpg"))
+
+    @util.memoized_property
+    def _isolation_lookup(self):
+        return {
+            "AUTOCOMMIT": "autocommit",
+            "READ COMMITTED": "read_committed",
+            "REPEATABLE READ": "repeatable_read",
+            "SERIALIZABLE": "serializable",
+        }
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return list(self._isolation_lookup)
+
+    def set_isolation_level(self, dbapi_connection, level):
+        dbapi_connection.set_isolation_level(self._isolation_lookup[level])
+
+    def set_readonly(self, connection, value):
+        connection.readonly = value
+
+    def get_readonly(self, connection):
+        return connection.readonly
+
+    def set_deferrable(self, connection, value):
+        connection.deferrable = value
+
+    def get_deferrable(self, connection):
+        return connection.deferrable
+
+    def do_terminate(self, dbapi_connection) -> None:
+        dbapi_connection.terminate()
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args(username="user")
+        multihosts, multiports = self._split_multihost_from_url(url)
+
+        opts.update(url.query)
+
+        if multihosts:
+            assert multiports
+            if len(multihosts) == 1:
+                opts["host"] = multihosts[0]
+                if multiports[0] is not None:
+                    opts["port"] = multiports[0]
+            elif not all(multihosts):
+                raise exc.ArgumentError(
+                    "All hosts are required to be present"
+                    " for asyncpg multiple host URL"
+                )
+            elif not all(multiports):
+                raise exc.ArgumentError(
+                    "All ports are required to be present"
+                    " for asyncpg multiple host URL"
+                )
+            else:
+                opts["host"] = list(multihosts)
+                opts["port"] = list(multiports)
+        else:
+            util.coerce_kw_type(opts, "port", int)
+        util.coerce_kw_type(opts, "prepared_statement_cache_size", int)
+        return ([], opts)
+
+    def do_ping(self, dbapi_connection):
+        dbapi_connection.ping()
+        return True
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if util.asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def is_disconnect(self, e, connection, cursor):
+        if connection:
+            return connection._connection.is_closed()
+        else:
+            return isinstance(
+                e, self.dbapi.InterfaceError
+            ) and "connection is closed" in str(e)
+
+    async def setup_asyncpg_json_codec(self, conn):
+        """set up JSON codec for asyncpg.
+
+        This occurs for all new connections and
+        can be overridden by third party dialects.
+
+        .. versionadded:: 1.4.27
+
+        """
+
+        asyncpg_connection = conn._connection
+        deserializer = self._json_deserializer or _py_json.loads
+
+        def _json_decoder(bin_value):
+            return deserializer(bin_value.decode())
+
+        await asyncpg_connection.set_type_codec(
+            "json",
+            encoder=str.encode,
+            decoder=_json_decoder,
+            schema="pg_catalog",
+            format="binary",
+        )
+
+    async def setup_asyncpg_jsonb_codec(self, conn):
+        """set up JSONB codec for asyncpg.
+
+        This occurs for all new connections and
+        can be overridden by third party dialects.
+
+        .. versionadded:: 1.4.27
+
+        """
+
+        asyncpg_connection = conn._connection
+        deserializer = self._json_deserializer or _py_json.loads
+
+        def _jsonb_encoder(str_value):
+            # \x01 is the prefix for jsonb used by PostgreSQL.
+            # asyncpg requires it when format='binary'
+            return b"\x01" + str_value.encode()
+
+        deserializer = self._json_deserializer or _py_json.loads
+
+        def _jsonb_decoder(bin_value):
+            # the byte is the \x01 prefix for jsonb used by PostgreSQL.
+            # asyncpg returns it when format='binary'
+            return deserializer(bin_value[1:].decode())
+
+        await asyncpg_connection.set_type_codec(
+            "jsonb",
+            encoder=_jsonb_encoder,
+            decoder=_jsonb_decoder,
+            schema="pg_catalog",
+            format="binary",
+        )
+
+    async def _disable_asyncpg_inet_codecs(self, conn):
+        asyncpg_connection = conn._connection
+
+        await asyncpg_connection.set_type_codec(
+            "inet",
+            encoder=lambda s: s,
+            decoder=lambda s: s,
+            schema="pg_catalog",
+            format="text",
+        )
+
+        await asyncpg_connection.set_type_codec(
+            "cidr",
+            encoder=lambda s: s,
+            decoder=lambda s: s,
+            schema="pg_catalog",
+            format="text",
+        )
+
+    def on_connect(self):
+        """on_connect for asyncpg
+
+        A major component of this for asyncpg is to set up type decoders at the
+        asyncpg level.
+
+        See https://github.com/MagicStack/asyncpg/issues/623 for
+        notes on JSON/JSONB implementation.
+
+        """
+
+        super_connect = super().on_connect()
+
+        def connect(conn):
+            conn.await_(self.setup_asyncpg_json_codec(conn))
+            conn.await_(self.setup_asyncpg_jsonb_codec(conn))
+
+            if self._native_inet_types is False:
+                conn.await_(self._disable_asyncpg_inet_codecs(conn))
+            if super_connect is not None:
+                super_connect(conn)
+
+        return connect
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = PGDialect_asyncpg
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/base.py
new file mode 100644
index 00000000..9bfaa277
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/base.py
@@ -0,0 +1,5038 @@
+# dialects/postgresql/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql
+    :name: PostgreSQL
+    :normal_support: 9.6+
+    :best_effort: 9+
+
+.. _postgresql_sequences:
+
+Sequences/SERIAL/IDENTITY
+-------------------------
+
+PostgreSQL supports sequences, and SQLAlchemy uses these as the default means
+of creating new primary key values for integer-based primary key columns. When
+creating tables, SQLAlchemy will issue the ``SERIAL`` datatype for
+integer-based primary key columns, which generates a sequence and server side
+default corresponding to the column.
+
+To specify a specific named sequence to be used for primary key generation,
+use the :func:`~sqlalchemy.schema.Sequence` construct::
+
+    Table(
+        "sometable",
+        metadata,
+        Column(
+            "id", Integer, Sequence("some_id_seq", start=1), primary_key=True
+        ),
+    )
+
+When SQLAlchemy issues a single INSERT statement, to fulfill the contract of
+having the "last insert identifier" available, a RETURNING clause is added to
+the INSERT statement which specifies the primary key columns should be
+returned after the statement completes. The RETURNING functionality only takes
+place if PostgreSQL 8.2 or later is in use. As a fallback approach, the
+sequence, whether specified explicitly or implicitly via ``SERIAL``, is
+executed independently beforehand, the returned value to be used in the
+subsequent insert. Note that when an
+:func:`~sqlalchemy.sql.expression.insert()` construct is executed using
+"executemany" semantics, the "last inserted identifier" functionality does not
+apply; no RETURNING clause is emitted nor is the sequence pre-executed in this
+case.
+
+
+PostgreSQL 10 and above IDENTITY columns
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+PostgreSQL 10 and above have a new IDENTITY feature that supersedes the use
+of SERIAL. The :class:`_schema.Identity` construct in a
+:class:`_schema.Column` can be used to control its behavior::
+
+    from sqlalchemy import Table, Column, MetaData, Integer, Computed
+
+    metadata = MetaData()
+
+    data = Table(
+        "data",
+        metadata,
+        Column(
+            "id", Integer, Identity(start=42, cycle=True), primary_key=True
+        ),
+        Column("data", String),
+    )
+
+The CREATE TABLE for the above :class:`_schema.Table` object would be:
+
+.. sourcecode:: sql
+
+    CREATE TABLE data (
+        id INTEGER GENERATED BY DEFAULT AS IDENTITY (START WITH 42 CYCLE),
+        data VARCHAR,
+        PRIMARY KEY (id)
+    )
+
+.. versionchanged::  1.4   Added :class:`_schema.Identity` construct
+   in a :class:`_schema.Column` to specify the option of an autoincrementing
+   column.
+
+.. note::
+
+   Previous versions of SQLAlchemy did not have built-in support for rendering
+   of IDENTITY, and could use the following compilation hook to replace
+   occurrences of SERIAL with IDENTITY::
+
+       from sqlalchemy.schema import CreateColumn
+       from sqlalchemy.ext.compiler import compiles
+
+
+       @compiles(CreateColumn, "postgresql")
+       def use_identity(element, compiler, **kw):
+           text = compiler.visit_create_column(element, **kw)
+           text = text.replace("SERIAL", "INT GENERATED BY DEFAULT AS IDENTITY")
+           return text
+
+   Using the above, a table such as::
+
+       t = Table(
+           "t", m, Column("id", Integer, primary_key=True), Column("data", String)
+       )
+
+   Will generate on the backing database as:
+
+   .. sourcecode:: sql
+
+       CREATE TABLE t (
+           id INT GENERATED BY DEFAULT AS IDENTITY,
+           data VARCHAR,
+           PRIMARY KEY (id)
+       )
+
+.. _postgresql_ss_cursors:
+
+Server Side Cursors
+-------------------
+
+Server-side cursor support is available for the psycopg2, asyncpg
+dialects and may also be available in others.
+
+Server side cursors are enabled on a per-statement basis by using the
+:paramref:`.Connection.execution_options.stream_results` connection execution
+option::
+
+    with engine.connect() as conn:
+        result = conn.execution_options(stream_results=True).execute(
+            text("select * from table")
+        )
+
+Note that some kinds of SQL statements may not be supported with
+server side cursors; generally, only SQL statements that return rows should be
+used with this option.
+
+.. deprecated:: 1.4  The dialect-level server_side_cursors flag is deprecated
+   and will be removed in a future release.  Please use the
+   :paramref:`_engine.Connection.stream_results` execution option for
+   unbuffered cursor support.
+
+.. seealso::
+
+    :ref:`engine_stream_results`
+
+.. _postgresql_isolation_level:
+
+Transaction Isolation Level
+---------------------------
+
+Most SQLAlchemy dialects support setting of transaction isolation level
+using the :paramref:`_sa.create_engine.isolation_level` parameter
+at the :func:`_sa.create_engine` level, and at the :class:`_engine.Connection`
+level via the :paramref:`.Connection.execution_options.isolation_level`
+parameter.
+
+For PostgreSQL dialects, this feature works either by making use of the
+DBAPI-specific features, such as psycopg2's isolation level flags which will
+embed the isolation level setting inline with the ``"BEGIN"`` statement, or for
+DBAPIs with no direct support by emitting ``SET SESSION CHARACTERISTICS AS
+TRANSACTION ISOLATION LEVEL <level>`` ahead of the ``"BEGIN"`` statement
+emitted by the DBAPI.   For the special AUTOCOMMIT isolation level,
+DBAPI-specific techniques are used which is typically an ``.autocommit``
+flag on the DBAPI connection object.
+
+To set isolation level using :func:`_sa.create_engine`::
+
+    engine = create_engine(
+        "postgresql+pg8000://scott:tiger@localhost/test",
+        isolation_level="REPEATABLE READ",
+    )
+
+To set using per-connection execution options::
+
+    with engine.connect() as conn:
+        conn = conn.execution_options(isolation_level="REPEATABLE READ")
+        with conn.begin():
+            ...  # work with transaction
+
+There are also more options for isolation level configurations, such as
+"sub-engine" objects linked to a main :class:`_engine.Engine` which each apply
+different isolation level settings.  See the discussion at
+:ref:`dbapi_autocommit` for background.
+
+Valid values for ``isolation_level`` on most PostgreSQL dialects include:
+
+* ``READ COMMITTED``
+* ``READ UNCOMMITTED``
+* ``REPEATABLE READ``
+* ``SERIALIZABLE``
+* ``AUTOCOMMIT``
+
+.. seealso::
+
+    :ref:`dbapi_autocommit`
+
+    :ref:`postgresql_readonly_deferrable`
+
+    :ref:`psycopg2_isolation_level`
+
+    :ref:`pg8000_isolation_level`
+
+.. _postgresql_readonly_deferrable:
+
+Setting READ ONLY / DEFERRABLE
+------------------------------
+
+Most PostgreSQL dialects support setting the "READ ONLY" and "DEFERRABLE"
+characteristics of the transaction, which is in addition to the isolation level
+setting. These two attributes can be established either in conjunction with or
+independently of the isolation level by passing the ``postgresql_readonly`` and
+``postgresql_deferrable`` flags with
+:meth:`_engine.Connection.execution_options`.  The example below illustrates
+passing the ``"SERIALIZABLE"`` isolation level at the same time as setting
+"READ ONLY" and "DEFERRABLE"::
+
+    with engine.connect() as conn:
+        conn = conn.execution_options(
+            isolation_level="SERIALIZABLE",
+            postgresql_readonly=True,
+            postgresql_deferrable=True,
+        )
+        with conn.begin():
+            ...  # work with transaction
+
+Note that some DBAPIs such as asyncpg only support "readonly" with
+SERIALIZABLE isolation.
+
+.. versionadded:: 1.4 added support for the ``postgresql_readonly``
+   and ``postgresql_deferrable`` execution options.
+
+.. _postgresql_reset_on_return:
+
+Temporary Table / Resource Reset for Connection Pooling
+-------------------------------------------------------
+
+The :class:`.QueuePool` connection pool implementation used
+by the SQLAlchemy :class:`.Engine` object includes
+:ref:`reset on return <pool_reset_on_return>` behavior that will invoke
+the DBAPI ``.rollback()`` method when connections are returned to the pool.
+While this rollback will clear out the immediate state used by the previous
+transaction, it does not cover a wider range of session-level state, including
+temporary tables as well as other server state such as prepared statement
+handles and statement caches.   The PostgreSQL database includes a variety
+of commands which may be used to reset this state, including
+``DISCARD``, ``RESET``, ``DEALLOCATE``, and ``UNLISTEN``.
+
+
+To install
+one or more of these commands as the means of performing reset-on-return,
+the :meth:`.PoolEvents.reset` event hook may be used, as demonstrated
+in the example below. The implementation
+will end transactions in progress as well as discard temporary tables
+using the ``CLOSE``, ``RESET`` and ``DISCARD`` commands; see the PostgreSQL
+documentation for background on what each of these statements do.
+
+The :paramref:`_sa.create_engine.pool_reset_on_return` parameter
+is set to ``None`` so that the custom scheme can replace the default behavior
+completely.   The custom hook implementation calls ``.rollback()`` in any case,
+as it's usually important that the DBAPI's own tracking of commit/rollback
+will remain consistent with the state of the transaction::
+
+
+    from sqlalchemy import create_engine
+    from sqlalchemy import event
+
+    postgresql_engine = create_engine(
+        "postgresql+pyscopg2://scott:tiger@hostname/dbname",
+        # disable default reset-on-return scheme
+        pool_reset_on_return=None,
+    )
+
+
+    @event.listens_for(postgresql_engine, "reset")
+    def _reset_postgresql(dbapi_connection, connection_record, reset_state):
+        if not reset_state.terminate_only:
+            dbapi_connection.execute("CLOSE ALL")
+            dbapi_connection.execute("RESET ALL")
+            dbapi_connection.execute("DISCARD TEMP")
+
+        # so that the DBAPI itself knows that the connection has been
+        # reset
+        dbapi_connection.rollback()
+
+.. versionchanged:: 2.0.0b3  Added additional state arguments to
+   the :meth:`.PoolEvents.reset` event and additionally ensured the event
+   is invoked for all "reset" occurrences, so that it's appropriate
+   as a place for custom "reset" handlers.   Previous schemes which
+   use the :meth:`.PoolEvents.checkin` handler remain usable as well.
+
+.. seealso::
+
+    :ref:`pool_reset_on_return` - in the :ref:`pooling_toplevel` documentation
+
+.. _postgresql_alternate_search_path:
+
+Setting Alternate Search Paths on Connect
+------------------------------------------
+
+The PostgreSQL ``search_path`` variable refers to the list of schema names
+that will be implicitly referenced when a particular table or other
+object is referenced in a SQL statement.  As detailed in the next section
+:ref:`postgresql_schema_reflection`, SQLAlchemy is generally organized around
+the concept of keeping this variable at its default value of ``public``,
+however, in order to have it set to any arbitrary name or names when connections
+are used automatically, the "SET SESSION search_path" command may be invoked
+for all connections in a pool using the following event handler, as discussed
+at :ref:`schema_set_default_connections`::
+
+    from sqlalchemy import event
+    from sqlalchemy import create_engine
+
+    engine = create_engine("postgresql+psycopg2://scott:tiger@host/dbname")
+
+
+    @event.listens_for(engine, "connect", insert=True)
+    def set_search_path(dbapi_connection, connection_record):
+        existing_autocommit = dbapi_connection.autocommit
+        dbapi_connection.autocommit = True
+        cursor = dbapi_connection.cursor()
+        cursor.execute("SET SESSION search_path='%s'" % schema_name)
+        cursor.close()
+        dbapi_connection.autocommit = existing_autocommit
+
+The reason the recipe is complicated by use of the ``.autocommit`` DBAPI
+attribute is so that when the ``SET SESSION search_path`` directive is invoked,
+it is invoked outside of the scope of any transaction and therefore will not
+be reverted when the DBAPI connection has a rollback.
+
+.. seealso::
+
+  :ref:`schema_set_default_connections` - in the :ref:`metadata_toplevel` documentation
+
+.. _postgresql_schema_reflection:
+
+Remote-Schema Table Introspection and PostgreSQL search_path
+------------------------------------------------------------
+
+.. admonition:: Section Best Practices Summarized
+
+    keep the ``search_path`` variable set to its default of ``public``, without
+    any other schema names. Ensure the username used to connect **does not**
+    match remote schemas, or ensure the ``"$user"`` token is **removed** from
+    ``search_path``.  For other schema names, name these explicitly
+    within :class:`_schema.Table` definitions. Alternatively, the
+    ``postgresql_ignore_search_path`` option will cause all reflected
+    :class:`_schema.Table` objects to have a :attr:`_schema.Table.schema`
+    attribute set up.
+
+The PostgreSQL dialect can reflect tables from any schema, as outlined in
+:ref:`metadata_reflection_schemas`.
+
+In all cases, the first thing SQLAlchemy does when reflecting tables is
+to **determine the default schema for the current database connection**.
+It does this using the PostgreSQL ``current_schema()``
+function, illustated below using a PostgreSQL client session (i.e. using
+the ``psql`` tool):
+
+.. sourcecode:: sql
+
+    test=> select current_schema();
+    current_schema
+    ----------------
+    public
+    (1 row)
+
+Above we see that on a plain install of PostgreSQL, the default schema name
+is the name ``public``.
+
+However, if your database username **matches the name of a schema**, PostgreSQL's
+default is to then **use that name as the default schema**.  Below, we log in
+using the username ``scott``.  When we create a schema named ``scott``, **it
+implicitly changes the default schema**:
+
+.. sourcecode:: sql
+
+    test=> select current_schema();
+    current_schema
+    ----------------
+    public
+    (1 row)
+
+    test=> create schema scott;
+    CREATE SCHEMA
+    test=> select current_schema();
+    current_schema
+    ----------------
+    scott
+    (1 row)
+
+The behavior of ``current_schema()`` is derived from the
+`PostgreSQL search path
+<https://www.postgresql.org/docs/current/static/ddl-schemas.html#DDL-SCHEMAS-PATH>`_
+variable ``search_path``, which in modern PostgreSQL versions defaults to this:
+
+.. sourcecode:: sql
+
+    test=> show search_path;
+    search_path
+    -----------------
+    "$user", public
+    (1 row)
+
+Where above, the ``"$user"`` variable will inject the current username as the
+default schema, if one exists.   Otherwise, ``public`` is used.
+
+When a :class:`_schema.Table` object is reflected, if it is present in the
+schema indicated by the ``current_schema()`` function, **the schema name assigned
+to the ".schema" attribute of the Table is the Python "None" value**.  Otherwise, the
+".schema" attribute will be assigned the string name of that schema.
+
+With regards to tables which these :class:`_schema.Table`
+objects refer to via foreign key constraint, a decision must be made as to how
+the ``.schema`` is represented in those remote tables, in the case where that
+remote schema name is also a member of the current ``search_path``.
+
+By default, the PostgreSQL dialect mimics the behavior encouraged by
+PostgreSQL's own ``pg_get_constraintdef()`` builtin procedure.  This function
+returns a sample definition for a particular foreign key constraint,
+omitting the referenced schema name from that definition when the name is
+also in the PostgreSQL schema search path.  The interaction below
+illustrates this behavior:
+
+.. sourcecode:: sql
+
+    test=> CREATE TABLE test_schema.referred(id INTEGER PRIMARY KEY);
+    CREATE TABLE
+    test=> CREATE TABLE referring(
+    test(>         id INTEGER PRIMARY KEY,
+    test(>         referred_id INTEGER REFERENCES test_schema.referred(id));
+    CREATE TABLE
+    test=> SET search_path TO public, test_schema;
+    test=> SELECT pg_catalog.pg_get_constraintdef(r.oid, true) FROM
+    test-> pg_catalog.pg_class c JOIN pg_catalog.pg_namespace n
+    test-> ON n.oid = c.relnamespace
+    test-> JOIN pg_catalog.pg_constraint r  ON c.oid = r.conrelid
+    test-> WHERE c.relname='referring' AND r.contype = 'f'
+    test-> ;
+                   pg_get_constraintdef
+    ---------------------------------------------------
+     FOREIGN KEY (referred_id) REFERENCES referred(id)
+    (1 row)
+
+Above, we created a table ``referred`` as a member of the remote schema
+``test_schema``, however when we added ``test_schema`` to the
+PG ``search_path`` and then asked ``pg_get_constraintdef()`` for the
+``FOREIGN KEY`` syntax, ``test_schema`` was not included in the output of
+the function.
+
+On the other hand, if we set the search path back to the typical default
+of ``public``:
+
+.. sourcecode:: sql
+
+    test=> SET search_path TO public;
+    SET
+
+The same query against ``pg_get_constraintdef()`` now returns the fully
+schema-qualified name for us:
+
+.. sourcecode:: sql
+
+    test=> SELECT pg_catalog.pg_get_constraintdef(r.oid, true) FROM
+    test-> pg_catalog.pg_class c JOIN pg_catalog.pg_namespace n
+    test-> ON n.oid = c.relnamespace
+    test-> JOIN pg_catalog.pg_constraint r  ON c.oid = r.conrelid
+    test-> WHERE c.relname='referring' AND r.contype = 'f';
+                         pg_get_constraintdef
+    ---------------------------------------------------------------
+     FOREIGN KEY (referred_id) REFERENCES test_schema.referred(id)
+    (1 row)
+
+SQLAlchemy will by default use the return value of ``pg_get_constraintdef()``
+in order to determine the remote schema name.  That is, if our ``search_path``
+were set to include ``test_schema``, and we invoked a table
+reflection process as follows::
+
+    >>> from sqlalchemy import Table, MetaData, create_engine, text
+    >>> engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test")
+    >>> with engine.connect() as conn:
+    ...     conn.execute(text("SET search_path TO test_schema, public"))
+    ...     metadata_obj = MetaData()
+    ...     referring = Table("referring", metadata_obj, autoload_with=conn)
+    <sqlalchemy.engine.result.CursorResult object at 0x101612ed0>
+
+The above process would deliver to the :attr:`_schema.MetaData.tables`
+collection
+``referred`` table named **without** the schema::
+
+    >>> metadata_obj.tables["referred"].schema is None
+    True
+
+To alter the behavior of reflection such that the referred schema is
+maintained regardless of the ``search_path`` setting, use the
+``postgresql_ignore_search_path`` option, which can be specified as a
+dialect-specific argument to both :class:`_schema.Table` as well as
+:meth:`_schema.MetaData.reflect`::
+
+    >>> with engine.connect() as conn:
+    ...     conn.execute(text("SET search_path TO test_schema, public"))
+    ...     metadata_obj = MetaData()
+    ...     referring = Table(
+    ...         "referring",
+    ...         metadata_obj,
+    ...         autoload_with=conn,
+    ...         postgresql_ignore_search_path=True,
+    ...     )
+    <sqlalchemy.engine.result.CursorResult object at 0x1016126d0>
+
+We will now have ``test_schema.referred`` stored as schema-qualified::
+
+    >>> metadata_obj.tables["test_schema.referred"].schema
+    'test_schema'
+
+.. sidebar:: Best Practices for PostgreSQL Schema reflection
+
+    The description of PostgreSQL schema reflection behavior is complex, and
+    is the product of many years of dealing with widely varied use cases and
+    user preferences. But in fact, there's no need to understand any of it if
+    you just stick to the simplest use pattern: leave the ``search_path`` set
+    to its default of ``public`` only, never refer to the name ``public`` as
+    an explicit schema name otherwise, and refer to all other schema names
+    explicitly when building up a :class:`_schema.Table` object.  The options
+    described here are only for those users who can't, or prefer not to, stay
+    within these guidelines.
+
+.. seealso::
+
+    :ref:`reflection_schema_qualified_interaction` - discussion of the issue
+    from a backend-agnostic perspective
+
+    `The Schema Search Path
+    <https://www.postgresql.org/docs/current/static/ddl-schemas.html#DDL-SCHEMAS-PATH>`_
+    - on the PostgreSQL website.
+
+INSERT/UPDATE...RETURNING
+-------------------------
+
+The dialect supports PG 8.2's ``INSERT..RETURNING``, ``UPDATE..RETURNING`` and
+``DELETE..RETURNING`` syntaxes.   ``INSERT..RETURNING`` is used by default
+for single-row INSERT statements in order to fetch newly generated
+primary key identifiers.   To specify an explicit ``RETURNING`` clause,
+use the :meth:`._UpdateBase.returning` method on a per-statement basis::
+
+    # INSERT..RETURNING
+    result = (
+        table.insert().returning(table.c.col1, table.c.col2).values(name="foo")
+    )
+    print(result.fetchall())
+
+    # UPDATE..RETURNING
+    result = (
+        table.update()
+        .returning(table.c.col1, table.c.col2)
+        .where(table.c.name == "foo")
+        .values(name="bar")
+    )
+    print(result.fetchall())
+
+    # DELETE..RETURNING
+    result = (
+        table.delete()
+        .returning(table.c.col1, table.c.col2)
+        .where(table.c.name == "foo")
+    )
+    print(result.fetchall())
+
+.. _postgresql_insert_on_conflict:
+
+INSERT...ON CONFLICT (Upsert)
+------------------------------
+
+Starting with version 9.5, PostgreSQL allows "upserts" (update or insert) of
+rows into a table via the ``ON CONFLICT`` clause of the ``INSERT`` statement. A
+candidate row will only be inserted if that row does not violate any unique
+constraints.  In the case of a unique constraint violation, a secondary action
+can occur which can be either "DO UPDATE", indicating that the data in the
+target row should be updated, or "DO NOTHING", which indicates to silently skip
+this row.
+
+Conflicts are determined using existing unique constraints and indexes.  These
+constraints may be identified either using their name as stated in DDL,
+or they may be inferred by stating the columns and conditions that comprise
+the indexes.
+
+SQLAlchemy provides ``ON CONFLICT`` support via the PostgreSQL-specific
+:func:`_postgresql.insert()` function, which provides
+the generative methods :meth:`_postgresql.Insert.on_conflict_do_update`
+and :meth:`~.postgresql.Insert.on_conflict_do_nothing`:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy.dialects.postgresql import insert
+    >>> insert_stmt = insert(my_table).values(
+    ...     id="some_existing_id", data="inserted value"
+    ... )
+    >>> do_nothing_stmt = insert_stmt.on_conflict_do_nothing(index_elements=["id"])
+    >>> print(do_nothing_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO NOTHING
+    {stop}
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     constraint="pk_my_table", set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT ON CONSTRAINT pk_my_table DO UPDATE SET data = %(param_1)s
+
+.. seealso::
+
+    `INSERT .. ON CONFLICT
+    <https://www.postgresql.org/docs/current/static/sql-insert.html#SQL-ON-CONFLICT>`_
+    - in the PostgreSQL documentation.
+
+Specifying the Target
+^^^^^^^^^^^^^^^^^^^^^
+
+Both methods supply the "target" of the conflict using either the
+named constraint or by column inference:
+
+* The :paramref:`_postgresql.Insert.on_conflict_do_update.index_elements` argument
+  specifies a sequence containing string column names, :class:`_schema.Column`
+  objects, and/or SQL expression elements, which would identify a unique
+  index:
+
+  .. sourcecode:: pycon+sql
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     index_elements=["id"], set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s
+    {stop}
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     index_elements=[my_table.c.id], set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s
+
+* When using :paramref:`_postgresql.Insert.on_conflict_do_update.index_elements` to
+  infer an index, a partial index can be inferred by also specifying the
+  use the :paramref:`_postgresql.Insert.on_conflict_do_update.index_where` parameter:
+
+  .. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(user_email="a@b.com", data="inserted data")
+    >>> stmt = stmt.on_conflict_do_update(
+    ...     index_elements=[my_table.c.user_email],
+    ...     index_where=my_table.c.user_email.like("%@gmail.com"),
+    ...     set_=dict(data=stmt.excluded.data),
+    ... )
+    >>> print(stmt)
+    {printsql}INSERT INTO my_table (data, user_email)
+    VALUES (%(data)s, %(user_email)s) ON CONFLICT (user_email)
+    WHERE user_email LIKE %(user_email_1)s DO UPDATE SET data = excluded.data
+
+* The :paramref:`_postgresql.Insert.on_conflict_do_update.constraint` argument is
+  used to specify an index directly rather than inferring it.  This can be
+  the name of a UNIQUE constraint, a PRIMARY KEY constraint, or an INDEX:
+
+  .. sourcecode:: pycon+sql
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     constraint="my_table_idx_1", set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT ON CONSTRAINT my_table_idx_1 DO UPDATE SET data = %(param_1)s
+    {stop}
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     constraint="my_table_pk", set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT ON CONSTRAINT my_table_pk DO UPDATE SET data = %(param_1)s
+    {stop}
+
+* The :paramref:`_postgresql.Insert.on_conflict_do_update.constraint` argument may
+  also refer to a SQLAlchemy construct representing a constraint,
+  e.g. :class:`.UniqueConstraint`, :class:`.PrimaryKeyConstraint`,
+  :class:`.Index`, or :class:`.ExcludeConstraint`.   In this use,
+  if the constraint has a name, it is used directly.  Otherwise, if the
+  constraint is unnamed, then inference will be used, where the expressions
+  and optional WHERE clause of the constraint will be spelled out in the
+  construct.  This use is especially convenient
+  to refer to the named or unnamed primary key of a :class:`_schema.Table`
+  using the
+  :attr:`_schema.Table.primary_key` attribute:
+
+  .. sourcecode:: pycon+sql
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     constraint=my_table.primary_key, set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s
+
+The SET Clause
+^^^^^^^^^^^^^^^
+
+``ON CONFLICT...DO UPDATE`` is used to perform an update of the already
+existing row, using any combination of new values as well as values
+from the proposed insertion.   These values are specified using the
+:paramref:`_postgresql.Insert.on_conflict_do_update.set_` parameter.  This
+parameter accepts a dictionary which consists of direct values
+for UPDATE:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+    >>> do_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"], set_=dict(data="updated value")
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s
+
+.. warning::
+
+    The :meth:`_expression.Insert.on_conflict_do_update`
+    method does **not** take into
+    account Python-side default UPDATE values or generation functions, e.g.
+    those specified using :paramref:`_schema.Column.onupdate`.
+    These values will not be exercised for an ON CONFLICT style of UPDATE,
+    unless they are manually specified in the
+    :paramref:`_postgresql.Insert.on_conflict_do_update.set_` dictionary.
+
+Updating using the Excluded INSERT Values
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+In order to refer to the proposed insertion row, the special alias
+:attr:`~.postgresql.Insert.excluded` is available as an attribute on
+the :class:`_postgresql.Insert` object; this object is a
+:class:`_expression.ColumnCollection`
+which alias contains all columns of the target
+table:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(
+    ...     id="some_id", data="inserted value", author="jlh"
+    ... )
+    >>> do_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"],
+    ...     set_=dict(data="updated value", author=stmt.excluded.author),
+    ... )
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data, author)
+    VALUES (%(id)s, %(data)s, %(author)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s, author = excluded.author
+
+Additional WHERE Criteria
+^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The :meth:`_expression.Insert.on_conflict_do_update` method also accepts
+a WHERE clause using the :paramref:`_postgresql.Insert.on_conflict_do_update.where`
+parameter, which will limit those rows which receive an UPDATE:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(
+    ...     id="some_id", data="inserted value", author="jlh"
+    ... )
+    >>> on_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"],
+    ...     set_=dict(data="updated value", author=stmt.excluded.author),
+    ...     where=(my_table.c.status == 2),
+    ... )
+    >>> print(on_update_stmt)
+    {printsql}INSERT INTO my_table (id, data, author)
+    VALUES (%(id)s, %(data)s, %(author)s)
+    ON CONFLICT (id) DO UPDATE SET data = %(param_1)s, author = excluded.author
+    WHERE my_table.status = %(status_1)s
+
+Skipping Rows with DO NOTHING
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+``ON CONFLICT`` may be used to skip inserting a row entirely
+if any conflict with a unique or exclusion constraint occurs; below
+this is illustrated using the
+:meth:`~.postgresql.Insert.on_conflict_do_nothing` method:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+    >>> stmt = stmt.on_conflict_do_nothing(index_elements=["id"])
+    >>> print(stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT (id) DO NOTHING
+
+If ``DO NOTHING`` is used without specifying any columns or constraint,
+it has the effect of skipping the INSERT for any unique or exclusion
+constraint violation which occurs:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+    >>> stmt = stmt.on_conflict_do_nothing()
+    >>> print(stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s)
+    ON CONFLICT DO NOTHING
+
+.. _postgresql_match:
+
+Full Text Search
+----------------
+
+PostgreSQL's full text search system is available through the use of the
+:data:`.func` namespace, combined with the use of custom operators
+via the :meth:`.Operators.bool_op` method.    For simple cases with some
+degree of cross-backend compatibility, the :meth:`.Operators.match` operator
+may also be used.
+
+.. _postgresql_simple_match:
+
+Simple plain text matching with ``match()``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The :meth:`.Operators.match` operator provides for cross-compatible simple
+text matching.   For the PostgreSQL backend, it's hardcoded to generate
+an expression using the ``@@`` operator in conjunction with the
+``plainto_tsquery()`` PostgreSQL function.
+
+On the PostgreSQL dialect, an expression like the following::
+
+    select(sometable.c.text.match("search string"))
+
+would emit to the database:
+
+.. sourcecode:: sql
+
+    SELECT text @@ plainto_tsquery('search string') FROM table
+
+Above, passing a plain string to :meth:`.Operators.match` will automatically
+make use of ``plainto_tsquery()`` to specify the type of tsquery.  This
+establishes basic database cross-compatibility for :meth:`.Operators.match`
+with other backends.
+
+.. versionchanged:: 2.0 The default tsquery generation function used by the
+   PostgreSQL dialect with :meth:`.Operators.match` is ``plainto_tsquery()``.
+
+   To render exactly what was rendered in 1.4, use the following form::
+
+        from sqlalchemy import func
+
+        select(sometable.c.text.bool_op("@@")(func.to_tsquery("search string")))
+
+   Which would emit:
+
+   .. sourcecode:: sql
+
+        SELECT text @@ to_tsquery('search string') FROM table
+
+Using PostgreSQL full text functions and operators directly
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Text search operations beyond the simple use of :meth:`.Operators.match`
+may make use of the :data:`.func` namespace to generate PostgreSQL full-text
+functions, in combination with :meth:`.Operators.bool_op` to generate
+any boolean operator.
+
+For example, the query::
+
+    select(func.to_tsquery("cat").bool_op("@>")(func.to_tsquery("cat & rat")))
+
+would generate:
+
+.. sourcecode:: sql
+
+    SELECT to_tsquery('cat') @> to_tsquery('cat & rat')
+
+
+The :class:`_postgresql.TSVECTOR` type can provide for explicit CAST::
+
+    from sqlalchemy.dialects.postgresql import TSVECTOR
+    from sqlalchemy import select, cast
+
+    select(cast("some text", TSVECTOR))
+
+produces a statement equivalent to:
+
+.. sourcecode:: sql
+
+    SELECT CAST('some text' AS TSVECTOR) AS anon_1
+
+The ``func`` namespace is augmented by the PostgreSQL dialect to set up
+correct argument and return types for most full text search functions.
+These functions are used automatically by the :attr:`_sql.func` namespace
+assuming the ``sqlalchemy.dialects.postgresql`` package has been imported,
+or :func:`_sa.create_engine` has been invoked using a ``postgresql``
+dialect.  These functions are documented at:
+
+* :class:`_postgresql.to_tsvector`
+* :class:`_postgresql.to_tsquery`
+* :class:`_postgresql.plainto_tsquery`
+* :class:`_postgresql.phraseto_tsquery`
+* :class:`_postgresql.websearch_to_tsquery`
+* :class:`_postgresql.ts_headline`
+
+Specifying the "regconfig" with ``match()`` or custom operators
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+PostgreSQL's ``plainto_tsquery()`` function accepts an optional
+"regconfig" argument that is used to instruct PostgreSQL to use a
+particular pre-computed GIN or GiST index in order to perform the search.
+When using :meth:`.Operators.match`, this additional parameter may be
+specified using the ``postgresql_regconfig`` parameter, such as::
+
+    select(mytable.c.id).where(
+        mytable.c.title.match("somestring", postgresql_regconfig="english")
+    )
+
+Which would emit:
+
+.. sourcecode:: sql
+
+    SELECT mytable.id FROM mytable
+    WHERE mytable.title @@ plainto_tsquery('english', 'somestring')
+
+When using other PostgreSQL search functions with :data:`.func`, the
+"regconfig" parameter may be passed directly as the initial argument::
+
+    select(mytable.c.id).where(
+        func.to_tsvector("english", mytable.c.title).bool_op("@@")(
+            func.to_tsquery("english", "somestring")
+        )
+    )
+
+produces a statement equivalent to:
+
+.. sourcecode:: sql
+
+    SELECT mytable.id FROM mytable
+    WHERE to_tsvector('english', mytable.title) @@
+        to_tsquery('english', 'somestring')
+
+It is recommended that you use the ``EXPLAIN ANALYZE...`` tool from
+PostgreSQL to ensure that you are generating queries with SQLAlchemy that
+take full advantage of any indexes you may have created for full text search.
+
+.. seealso::
+
+    `Full Text Search <https://www.postgresql.org/docs/current/textsearch-controls.html>`_ - in the PostgreSQL documentation
+
+
+FROM ONLY ...
+-------------
+
+The dialect supports PostgreSQL's ONLY keyword for targeting only a particular
+table in an inheritance hierarchy. This can be used to produce the
+``SELECT ... FROM ONLY``, ``UPDATE ONLY ...``, and ``DELETE FROM ONLY ...``
+syntaxes. It uses SQLAlchemy's hints mechanism::
+
+    # SELECT ... FROM ONLY ...
+    result = table.select().with_hint(table, "ONLY", "postgresql")
+    print(result.fetchall())
+
+    # UPDATE ONLY ...
+    table.update(values=dict(foo="bar")).with_hint(
+        "ONLY", dialect_name="postgresql"
+    )
+
+    # DELETE FROM ONLY ...
+    table.delete().with_hint("ONLY", dialect_name="postgresql")
+
+.. _postgresql_indexes:
+
+PostgreSQL-Specific Index Options
+---------------------------------
+
+Several extensions to the :class:`.Index` construct are available, specific
+to the PostgreSQL dialect.
+
+Covering Indexes
+^^^^^^^^^^^^^^^^
+
+The ``postgresql_include`` option renders INCLUDE(colname) for the given
+string names::
+
+    Index("my_index", table.c.x, postgresql_include=["y"])
+
+would render the index as ``CREATE INDEX my_index ON table (x) INCLUDE (y)``
+
+Note that this feature requires PostgreSQL 11 or later.
+
+.. versionadded:: 1.4
+
+.. _postgresql_partial_indexes:
+
+Partial Indexes
+^^^^^^^^^^^^^^^
+
+Partial indexes add criterion to the index definition so that the index is
+applied to a subset of rows.   These can be specified on :class:`.Index`
+using the ``postgresql_where`` keyword argument::
+
+  Index("my_index", my_table.c.id, postgresql_where=my_table.c.value > 10)
+
+.. _postgresql_operator_classes:
+
+Operator Classes
+^^^^^^^^^^^^^^^^
+
+PostgreSQL allows the specification of an *operator class* for each column of
+an index (see
+https://www.postgresql.org/docs/current/interactive/indexes-opclass.html).
+The :class:`.Index` construct allows these to be specified via the
+``postgresql_ops`` keyword argument::
+
+    Index(
+        "my_index",
+        my_table.c.id,
+        my_table.c.data,
+        postgresql_ops={"data": "text_pattern_ops", "id": "int4_ops"},
+    )
+
+Note that the keys in the ``postgresql_ops`` dictionaries are the
+"key" name of the :class:`_schema.Column`, i.e. the name used to access it from
+the ``.c`` collection of :class:`_schema.Table`, which can be configured to be
+different than the actual name of the column as expressed in the database.
+
+If ``postgresql_ops`` is to be used against a complex SQL expression such
+as a function call, then to apply to the column it must be given a label
+that is identified in the dictionary by name, e.g.::
+
+    Index(
+        "my_index",
+        my_table.c.id,
+        func.lower(my_table.c.data).label("data_lower"),
+        postgresql_ops={"data_lower": "text_pattern_ops", "id": "int4_ops"},
+    )
+
+Operator classes are also supported by the
+:class:`_postgresql.ExcludeConstraint` construct using the
+:paramref:`_postgresql.ExcludeConstraint.ops` parameter. See that parameter for
+details.
+
+.. versionadded:: 1.3.21 added support for operator classes with
+   :class:`_postgresql.ExcludeConstraint`.
+
+
+Index Types
+^^^^^^^^^^^
+
+PostgreSQL provides several index types: B-Tree, Hash, GiST, and GIN, as well
+as the ability for users to create their own (see
+https://www.postgresql.org/docs/current/static/indexes-types.html). These can be
+specified on :class:`.Index` using the ``postgresql_using`` keyword argument::
+
+    Index("my_index", my_table.c.data, postgresql_using="gin")
+
+The value passed to the keyword argument will be simply passed through to the
+underlying CREATE INDEX command, so it *must* be a valid index type for your
+version of PostgreSQL.
+
+.. _postgresql_index_storage:
+
+Index Storage Parameters
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+PostgreSQL allows storage parameters to be set on indexes. The storage
+parameters available depend on the index method used by the index. Storage
+parameters can be specified on :class:`.Index` using the ``postgresql_with``
+keyword argument::
+
+    Index("my_index", my_table.c.data, postgresql_with={"fillfactor": 50})
+
+PostgreSQL allows to define the tablespace in which to create the index.
+The tablespace can be specified on :class:`.Index` using the
+``postgresql_tablespace`` keyword argument::
+
+    Index("my_index", my_table.c.data, postgresql_tablespace="my_tablespace")
+
+Note that the same option is available on :class:`_schema.Table` as well.
+
+.. _postgresql_index_concurrently:
+
+Indexes with CONCURRENTLY
+^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The PostgreSQL index option CONCURRENTLY is supported by passing the
+flag ``postgresql_concurrently`` to the :class:`.Index` construct::
+
+    tbl = Table("testtbl", m, Column("data", Integer))
+
+    idx1 = Index("test_idx1", tbl.c.data, postgresql_concurrently=True)
+
+The above index construct will render DDL for CREATE INDEX, assuming
+PostgreSQL 8.2 or higher is detected or for a connection-less dialect, as:
+
+.. sourcecode:: sql
+
+    CREATE INDEX CONCURRENTLY test_idx1 ON testtbl (data)
+
+For DROP INDEX, assuming PostgreSQL 9.2 or higher is detected or for
+a connection-less dialect, it will emit:
+
+.. sourcecode:: sql
+
+    DROP INDEX CONCURRENTLY test_idx1
+
+When using CONCURRENTLY, the PostgreSQL database requires that the statement
+be invoked outside of a transaction block.   The Python DBAPI enforces that
+even for a single statement, a transaction is present, so to use this
+construct, the DBAPI's "autocommit" mode must be used::
+
+    metadata = MetaData()
+    table = Table("foo", metadata, Column("id", String))
+    index = Index("foo_idx", table.c.id, postgresql_concurrently=True)
+
+    with engine.connect() as conn:
+        with conn.execution_options(isolation_level="AUTOCOMMIT"):
+            table.create(conn)
+
+.. seealso::
+
+    :ref:`postgresql_isolation_level`
+
+.. _postgresql_index_reflection:
+
+PostgreSQL Index Reflection
+---------------------------
+
+The PostgreSQL database creates a UNIQUE INDEX implicitly whenever the
+UNIQUE CONSTRAINT construct is used.   When inspecting a table using
+:class:`_reflection.Inspector`, the :meth:`_reflection.Inspector.get_indexes`
+and the :meth:`_reflection.Inspector.get_unique_constraints`
+will report on these
+two constructs distinctly; in the case of the index, the key
+``duplicates_constraint`` will be present in the index entry if it is
+detected as mirroring a constraint.   When performing reflection using
+``Table(..., autoload_with=engine)``, the UNIQUE INDEX is **not** returned
+in :attr:`_schema.Table.indexes` when it is detected as mirroring a
+:class:`.UniqueConstraint` in the :attr:`_schema.Table.constraints` collection
+.
+
+Special Reflection Options
+--------------------------
+
+The :class:`_reflection.Inspector`
+used for the PostgreSQL backend is an instance
+of :class:`.PGInspector`, which offers additional methods::
+
+    from sqlalchemy import create_engine, inspect
+
+    engine = create_engine("postgresql+psycopg2://localhost/test")
+    insp = inspect(engine)  # will be a PGInspector
+
+    print(insp.get_enums())
+
+.. autoclass:: PGInspector
+    :members:
+
+.. _postgresql_table_options:
+
+PostgreSQL Table Options
+------------------------
+
+Several options for CREATE TABLE are supported directly by the PostgreSQL
+dialect in conjunction with the :class:`_schema.Table` construct:
+
+* ``INHERITS``::
+
+    Table("some_table", metadata, ..., postgresql_inherits="some_supertable")
+
+    Table("some_table", metadata, ..., postgresql_inherits=("t1", "t2", ...))
+
+* ``ON COMMIT``::
+
+    Table("some_table", metadata, ..., postgresql_on_commit="PRESERVE ROWS")
+
+*
+  ``PARTITION BY``::
+
+    Table(
+        "some_table",
+        metadata,
+        ...,
+        postgresql_partition_by="LIST (part_column)",
+    )
+
+  .. versionadded:: 1.2.6
+
+*
+  ``TABLESPACE``::
+
+    Table("some_table", metadata, ..., postgresql_tablespace="some_tablespace")
+
+  The above option is also available on the :class:`.Index` construct.
+
+*
+  ``USING``::
+
+    Table("some_table", metadata, ..., postgresql_using="heap")
+
+  .. versionadded:: 2.0.26
+
+* ``WITH OIDS``::
+
+    Table("some_table", metadata, ..., postgresql_with_oids=True)
+
+* ``WITHOUT OIDS``::
+
+    Table("some_table", metadata, ..., postgresql_with_oids=False)
+
+.. seealso::
+
+    `PostgreSQL CREATE TABLE options
+    <https://www.postgresql.org/docs/current/static/sql-createtable.html>`_ -
+    in the PostgreSQL documentation.
+
+.. _postgresql_constraint_options:
+
+PostgreSQL Constraint Options
+-----------------------------
+
+The following option(s) are supported by the PostgreSQL dialect in conjunction
+with selected constraint constructs:
+
+* ``NOT VALID``:  This option applies towards CHECK and FOREIGN KEY constraints
+  when the constraint is being added to an existing table via ALTER TABLE,
+  and has the effect that existing rows are not scanned during the ALTER
+  operation against the constraint being added.
+
+  When using a SQL migration tool such as `Alembic <https://alembic.sqlalchemy.org>`_
+  that renders ALTER TABLE constructs, the ``postgresql_not_valid`` argument
+  may be specified as an additional keyword argument within the operation
+  that creates the constraint, as in the following Alembic example::
+
+        def update():
+            op.create_foreign_key(
+                "fk_user_address",
+                "address",
+                "user",
+                ["user_id"],
+                ["id"],
+                postgresql_not_valid=True,
+            )
+
+  The keyword is ultimately accepted directly by the
+  :class:`_schema.CheckConstraint`, :class:`_schema.ForeignKeyConstraint`
+  and :class:`_schema.ForeignKey` constructs; when using a tool like
+  Alembic, dialect-specific keyword arguments are passed through to
+  these constructs from the migration operation directives::
+
+       CheckConstraint("some_field IS NOT NULL", postgresql_not_valid=True)
+
+       ForeignKeyConstraint(
+           ["some_id"], ["some_table.some_id"], postgresql_not_valid=True
+       )
+
+  .. versionadded:: 1.4.32
+
+  .. seealso::
+
+      `PostgreSQL ALTER TABLE options
+      <https://www.postgresql.org/docs/current/static/sql-altertable.html>`_ -
+      in the PostgreSQL documentation.
+
+.. _postgresql_table_valued_overview:
+
+Table values, Table and Column valued functions, Row and Tuple objects
+-----------------------------------------------------------------------
+
+PostgreSQL makes great use of modern SQL forms such as table-valued functions,
+tables and rows as values.   These constructs are commonly used as part
+of PostgreSQL's support for complex datatypes such as JSON, ARRAY, and other
+datatypes.  SQLAlchemy's SQL expression language has native support for
+most table-valued and row-valued forms.
+
+.. _postgresql_table_valued:
+
+Table-Valued Functions
+^^^^^^^^^^^^^^^^^^^^^^^
+
+Many PostgreSQL built-in functions are intended to be used in the FROM clause
+of a SELECT statement, and are capable of returning table rows or sets of table
+rows. A large portion of PostgreSQL's JSON functions for example such as
+``json_array_elements()``, ``json_object_keys()``, ``json_each_text()``,
+``json_each()``, ``json_to_record()``, ``json_populate_recordset()`` use such
+forms. These classes of SQL function calling forms in SQLAlchemy are available
+using the :meth:`_functions.FunctionElement.table_valued` method in conjunction
+with :class:`_functions.Function` objects generated from the :data:`_sql.func`
+namespace.
+
+Examples from PostgreSQL's reference documentation follow below:
+
+* ``json_each()``:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select, func
+    >>> stmt = select(
+    ...     func.json_each('{"a":"foo", "b":"bar"}').table_valued("key", "value")
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT anon_1.key, anon_1.value
+    FROM json_each(:json_each_1) AS anon_1
+
+* ``json_populate_record()``:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select, func, literal_column
+    >>> stmt = select(
+    ...     func.json_populate_record(
+    ...         literal_column("null::myrowtype"), '{"a":1,"b":2}'
+    ...     ).table_valued("a", "b", name="x")
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT x.a, x.b
+    FROM json_populate_record(null::myrowtype, :json_populate_record_1) AS x
+
+* ``json_to_record()`` - this form uses a PostgreSQL specific form of derived
+  columns in the alias, where we may make use of :func:`_sql.column` elements with
+  types to produce them.  The :meth:`_functions.FunctionElement.table_valued`
+  method produces  a :class:`_sql.TableValuedAlias` construct, and the method
+  :meth:`_sql.TableValuedAlias.render_derived` method sets up the derived
+  columns specification:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select, func, column, Integer, Text
+    >>> stmt = select(
+    ...     func.json_to_record('{"a":1,"b":[1,2,3],"c":"bar"}')
+    ...     .table_valued(
+    ...         column("a", Integer),
+    ...         column("b", Text),
+    ...         column("d", Text),
+    ...     )
+    ...     .render_derived(name="x", with_types=True)
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT x.a, x.b, x.d
+    FROM json_to_record(:json_to_record_1) AS x(a INTEGER, b TEXT, d TEXT)
+
+* ``WITH ORDINALITY`` - part of the SQL standard, ``WITH ORDINALITY`` adds an
+  ordinal counter to the output of a function and is accepted by a limited set
+  of PostgreSQL functions including ``unnest()`` and ``generate_series()``. The
+  :meth:`_functions.FunctionElement.table_valued` method accepts a keyword
+  parameter ``with_ordinality`` for this purpose, which accepts the string name
+  that will be applied to the "ordinality" column:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select, func
+    >>> stmt = select(
+    ...     func.generate_series(4, 1, -1)
+    ...     .table_valued("value", with_ordinality="ordinality")
+    ...     .render_derived()
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT anon_1.value, anon_1.ordinality
+    FROM generate_series(:generate_series_1, :generate_series_2, :generate_series_3)
+    WITH ORDINALITY AS anon_1(value, ordinality)
+
+.. versionadded:: 1.4.0b2
+
+.. seealso::
+
+    :ref:`tutorial_functions_table_valued` - in the :ref:`unified_tutorial`
+
+.. _postgresql_column_valued:
+
+Column Valued Functions
+^^^^^^^^^^^^^^^^^^^^^^^
+
+Similar to the table valued function, a column valued function is present
+in the FROM clause, but delivers itself to the columns clause as a single
+scalar value.  PostgreSQL functions such as ``json_array_elements()``,
+``unnest()`` and ``generate_series()`` may use this form. Column valued functions are available using the
+:meth:`_functions.FunctionElement.column_valued` method of :class:`_functions.FunctionElement`:
+
+* ``json_array_elements()``:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select, func
+    >>> stmt = select(
+    ...     func.json_array_elements('["one", "two"]').column_valued("x")
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT x
+    FROM json_array_elements(:json_array_elements_1) AS x
+
+* ``unnest()`` - in order to generate a PostgreSQL ARRAY literal, the
+  :func:`_postgresql.array` construct may be used:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy.dialects.postgresql import array
+    >>> from sqlalchemy import select, func
+    >>> stmt = select(func.unnest(array([1, 2])).column_valued())
+    >>> print(stmt)
+    {printsql}SELECT anon_1
+    FROM unnest(ARRAY[%(param_1)s, %(param_2)s]) AS anon_1
+
+  The function can of course be used against an existing table-bound column
+  that's of type :class:`_types.ARRAY`:
+
+  .. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import table, column, ARRAY, Integer
+    >>> from sqlalchemy import select, func
+    >>> t = table("t", column("value", ARRAY(Integer)))
+    >>> stmt = select(func.unnest(t.c.value).column_valued("unnested_value"))
+    >>> print(stmt)
+    {printsql}SELECT unnested_value
+    FROM unnest(t.value) AS unnested_value
+
+.. seealso::
+
+    :ref:`tutorial_functions_column_valued` - in the :ref:`unified_tutorial`
+
+
+Row Types
+^^^^^^^^^
+
+Built-in support for rendering a ``ROW`` may be approximated using
+``func.ROW`` with the :attr:`_sa.func` namespace, or by using the
+:func:`_sql.tuple_` construct:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import table, column, func, tuple_
+    >>> t = table("t", column("id"), column("fk"))
+    >>> stmt = (
+    ...     t.select()
+    ...     .where(tuple_(t.c.id, t.c.fk) > (1, 2))
+    ...     .where(func.ROW(t.c.id, t.c.fk) < func.ROW(3, 7))
+    ... )
+    >>> print(stmt)
+    {printsql}SELECT t.id, t.fk
+    FROM t
+    WHERE (t.id, t.fk) > (:param_1, :param_2) AND ROW(t.id, t.fk) < ROW(:ROW_1, :ROW_2)
+
+.. seealso::
+
+    `PostgreSQL Row Constructors
+    <https://www.postgresql.org/docs/current/sql-expressions.html#SQL-SYNTAX-ROW-CONSTRUCTORS>`_
+
+    `PostgreSQL Row Constructor Comparison
+    <https://www.postgresql.org/docs/current/functions-comparisons.html#ROW-WISE-COMPARISON>`_
+
+Table Types passed to Functions
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+PostgreSQL supports passing a table as an argument to a function, which is
+known as a "record" type. SQLAlchemy :class:`_sql.FromClause` objects
+such as :class:`_schema.Table` support this special form using the
+:meth:`_sql.FromClause.table_valued` method, which is comparable to the
+:meth:`_functions.FunctionElement.table_valued` method except that the collection
+of columns is already established by that of the :class:`_sql.FromClause`
+itself:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import table, column, func, select
+    >>> a = table("a", column("id"), column("x"), column("y"))
+    >>> stmt = select(func.row_to_json(a.table_valued()))
+    >>> print(stmt)
+    {printsql}SELECT row_to_json(a) AS row_to_json_1
+    FROM a
+
+.. versionadded:: 1.4.0b2
+
+
+
+"""  # noqa: E501
+
+from __future__ import annotations
+
+from collections import defaultdict
+from functools import lru_cache
+import re
+from typing import Any
+from typing import cast
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import arraylib as _array
+from . import json as _json
+from . import pg_catalog
+from . import ranges as _ranges
+from .ext import _regconfig_fn
+from .ext import aggregate_order_by
+from .hstore import HSTORE
+from .named_types import CreateDomainType as CreateDomainType  # noqa: F401
+from .named_types import CreateEnumType as CreateEnumType  # noqa: F401
+from .named_types import DOMAIN as DOMAIN  # noqa: F401
+from .named_types import DropDomainType as DropDomainType  # noqa: F401
+from .named_types import DropEnumType as DropEnumType  # noqa: F401
+from .named_types import ENUM as ENUM  # noqa: F401
+from .named_types import NamedType as NamedType  # noqa: F401
+from .types import _DECIMAL_TYPES  # noqa: F401
+from .types import _FLOAT_TYPES  # noqa: F401
+from .types import _INT_TYPES  # noqa: F401
+from .types import BIT as BIT
+from .types import BYTEA as BYTEA
+from .types import CIDR as CIDR
+from .types import CITEXT as CITEXT
+from .types import INET as INET
+from .types import INTERVAL as INTERVAL
+from .types import MACADDR as MACADDR
+from .types import MACADDR8 as MACADDR8
+from .types import MONEY as MONEY
+from .types import OID as OID
+from .types import PGBit as PGBit  # noqa: F401
+from .types import PGCidr as PGCidr  # noqa: F401
+from .types import PGInet as PGInet  # noqa: F401
+from .types import PGInterval as PGInterval  # noqa: F401
+from .types import PGMacAddr as PGMacAddr  # noqa: F401
+from .types import PGMacAddr8 as PGMacAddr8  # noqa: F401
+from .types import PGUuid as PGUuid
+from .types import REGCLASS as REGCLASS
+from .types import REGCONFIG as REGCONFIG  # noqa: F401
+from .types import TIME as TIME
+from .types import TIMESTAMP as TIMESTAMP
+from .types import TSVECTOR as TSVECTOR
+from ... import exc
+from ... import schema
+from ... import select
+from ... import sql
+from ... import util
+from ...engine import characteristics
+from ...engine import default
+from ...engine import interfaces
+from ...engine import ObjectKind
+from ...engine import ObjectScope
+from ...engine import reflection
+from ...engine import URL
+from ...engine.reflection import ReflectionDefaults
+from ...sql import bindparam
+from ...sql import coercions
+from ...sql import compiler
+from ...sql import elements
+from ...sql import expression
+from ...sql import roles
+from ...sql import sqltypes
+from ...sql import util as sql_util
+from ...sql.compiler import InsertmanyvaluesSentinelOpts
+from ...sql.visitors import InternalTraversal
+from ...types import BIGINT
+from ...types import BOOLEAN
+from ...types import CHAR
+from ...types import DATE
+from ...types import DOUBLE_PRECISION
+from ...types import FLOAT
+from ...types import INTEGER
+from ...types import NUMERIC
+from ...types import REAL
+from ...types import SMALLINT
+from ...types import TEXT
+from ...types import UUID as UUID
+from ...types import VARCHAR
+from ...util.typing import TypedDict
+
+IDX_USING = re.compile(r"^(?:btree|hash|gist|gin|[\w_]+)$", re.I)
+
+RESERVED_WORDS = {
+    "all",
+    "analyse",
+    "analyze",
+    "and",
+    "any",
+    "array",
+    "as",
+    "asc",
+    "asymmetric",
+    "both",
+    "case",
+    "cast",
+    "check",
+    "collate",
+    "column",
+    "constraint",
+    "create",
+    "current_catalog",
+    "current_date",
+    "current_role",
+    "current_time",
+    "current_timestamp",
+    "current_user",
+    "default",
+    "deferrable",
+    "desc",
+    "distinct",
+    "do",
+    "else",
+    "end",
+    "except",
+    "false",
+    "fetch",
+    "for",
+    "foreign",
+    "from",
+    "grant",
+    "group",
+    "having",
+    "in",
+    "initially",
+    "intersect",
+    "into",
+    "leading",
+    "limit",
+    "localtime",
+    "localtimestamp",
+    "new",
+    "not",
+    "null",
+    "of",
+    "off",
+    "offset",
+    "old",
+    "on",
+    "only",
+    "or",
+    "order",
+    "placing",
+    "primary",
+    "references",
+    "returning",
+    "select",
+    "session_user",
+    "some",
+    "symmetric",
+    "table",
+    "then",
+    "to",
+    "trailing",
+    "true",
+    "union",
+    "unique",
+    "user",
+    "using",
+    "variadic",
+    "when",
+    "where",
+    "window",
+    "with",
+    "authorization",
+    "between",
+    "binary",
+    "cross",
+    "current_schema",
+    "freeze",
+    "full",
+    "ilike",
+    "inner",
+    "is",
+    "isnull",
+    "join",
+    "left",
+    "like",
+    "natural",
+    "notnull",
+    "outer",
+    "over",
+    "overlaps",
+    "right",
+    "similar",
+    "verbose",
+}
+
+colspecs = {
+    sqltypes.ARRAY: _array.ARRAY,
+    sqltypes.Interval: INTERVAL,
+    sqltypes.Enum: ENUM,
+    sqltypes.JSON.JSONPathType: _json.JSONPATH,
+    sqltypes.JSON: _json.JSON,
+    sqltypes.Uuid: PGUuid,
+}
+
+
+ischema_names = {
+    "_array": _array.ARRAY,
+    "hstore": HSTORE,
+    "json": _json.JSON,
+    "jsonb": _json.JSONB,
+    "int4range": _ranges.INT4RANGE,
+    "int8range": _ranges.INT8RANGE,
+    "numrange": _ranges.NUMRANGE,
+    "daterange": _ranges.DATERANGE,
+    "tsrange": _ranges.TSRANGE,
+    "tstzrange": _ranges.TSTZRANGE,
+    "int4multirange": _ranges.INT4MULTIRANGE,
+    "int8multirange": _ranges.INT8MULTIRANGE,
+    "nummultirange": _ranges.NUMMULTIRANGE,
+    "datemultirange": _ranges.DATEMULTIRANGE,
+    "tsmultirange": _ranges.TSMULTIRANGE,
+    "tstzmultirange": _ranges.TSTZMULTIRANGE,
+    "integer": INTEGER,
+    "bigint": BIGINT,
+    "smallint": SMALLINT,
+    "character varying": VARCHAR,
+    "character": CHAR,
+    '"char"': sqltypes.String,
+    "name": sqltypes.String,
+    "text": TEXT,
+    "numeric": NUMERIC,
+    "float": FLOAT,
+    "real": REAL,
+    "inet": INET,
+    "cidr": CIDR,
+    "citext": CITEXT,
+    "uuid": UUID,
+    "bit": BIT,
+    "bit varying": BIT,
+    "macaddr": MACADDR,
+    "macaddr8": MACADDR8,
+    "money": MONEY,
+    "oid": OID,
+    "regclass": REGCLASS,
+    "double precision": DOUBLE_PRECISION,
+    "timestamp": TIMESTAMP,
+    "timestamp with time zone": TIMESTAMP,
+    "timestamp without time zone": TIMESTAMP,
+    "time with time zone": TIME,
+    "time without time zone": TIME,
+    "date": DATE,
+    "time": TIME,
+    "bytea": BYTEA,
+    "boolean": BOOLEAN,
+    "interval": INTERVAL,
+    "tsvector": TSVECTOR,
+}
+
+
+class PGCompiler(compiler.SQLCompiler):
+    def visit_to_tsvector_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def visit_to_tsquery_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def visit_plainto_tsquery_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def visit_phraseto_tsquery_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def visit_websearch_to_tsquery_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def visit_ts_headline_func(self, element, **kw):
+        return self._assert_pg_ts_ext(element, **kw)
+
+    def _assert_pg_ts_ext(self, element, **kw):
+        if not isinstance(element, _regconfig_fn):
+            # other options here include trying to rewrite the function
+            # with the correct types.  however, that means we have to
+            # "un-SQL-ize" the first argument, which can't work in a
+            # generalized way. Also, parent compiler class has already added
+            # the incorrect return type to the result map.   So let's just
+            # make sure the function we want is used up front.
+
+            raise exc.CompileError(
+                f'Can\'t compile "{element.name}()" full text search '
+                f"function construct that does not originate from the "
+                f'"sqlalchemy.dialects.postgresql" package.  '
+                f'Please ensure "import sqlalchemy.dialects.postgresql" is '
+                f"called before constructing "
+                f'"sqlalchemy.func.{element.name}()" to ensure registration '
+                f"of the correct argument and return types."
+            )
+
+        return f"{element.name}{self.function_argspec(element, **kw)}"
+
+    def render_bind_cast(self, type_, dbapi_type, sqltext):
+        if dbapi_type._type_affinity is sqltypes.String and dbapi_type.length:
+            # use VARCHAR with no length for VARCHAR cast.
+            # see #9511
+            dbapi_type = sqltypes.STRINGTYPE
+        return f"""{sqltext}::{
+            self.dialect.type_compiler_instance.process(
+                dbapi_type, identifier_preparer=self.preparer
+            )
+        }"""
+
+    def visit_array(self, element, **kw):
+        return "ARRAY[%s]" % self.visit_clauselist(element, **kw)
+
+    def visit_slice(self, element, **kw):
+        return "%s:%s" % (
+            self.process(element.start, **kw),
+            self.process(element.stop, **kw),
+        )
+
+    def visit_bitwise_xor_op_binary(self, binary, operator, **kw):
+        return self._generate_generic_binary(binary, " # ", **kw)
+
+    def visit_json_getitem_op_binary(
+        self, binary, operator, _cast_applied=False, **kw
+    ):
+        if (
+            not _cast_applied
+            and binary.type._type_affinity is not sqltypes.JSON
+        ):
+            kw["_cast_applied"] = True
+            return self.process(sql.cast(binary, binary.type), **kw)
+
+        kw["eager_grouping"] = True
+
+        return self._generate_generic_binary(
+            binary, " -> " if not _cast_applied else " ->> ", **kw
+        )
+
+    def visit_json_path_getitem_op_binary(
+        self, binary, operator, _cast_applied=False, **kw
+    ):
+        if (
+            not _cast_applied
+            and binary.type._type_affinity is not sqltypes.JSON
+        ):
+            kw["_cast_applied"] = True
+            return self.process(sql.cast(binary, binary.type), **kw)
+
+        kw["eager_grouping"] = True
+        return self._generate_generic_binary(
+            binary, " #> " if not _cast_applied else " #>> ", **kw
+        )
+
+    def visit_getitem_binary(self, binary, operator, **kw):
+        return "%s[%s]" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_aggregate_order_by(self, element, **kw):
+        return "%s ORDER BY %s" % (
+            self.process(element.target, **kw),
+            self.process(element.order_by, **kw),
+        )
+
+    def visit_match_op_binary(self, binary, operator, **kw):
+        if "postgresql_regconfig" in binary.modifiers:
+            regconfig = self.render_literal_value(
+                binary.modifiers["postgresql_regconfig"], sqltypes.STRINGTYPE
+            )
+            if regconfig:
+                return "%s @@ plainto_tsquery(%s, %s)" % (
+                    self.process(binary.left, **kw),
+                    regconfig,
+                    self.process(binary.right, **kw),
+                )
+        return "%s @@ plainto_tsquery(%s)" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_ilike_case_insensitive_operand(self, element, **kw):
+        return element.element._compiler_dispatch(self, **kw)
+
+    def visit_ilike_op_binary(self, binary, operator, **kw):
+        escape = binary.modifiers.get("escape", None)
+
+        return "%s ILIKE %s" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        ) + (
+            " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE)
+            if escape is not None
+            else ""
+        )
+
+    def visit_not_ilike_op_binary(self, binary, operator, **kw):
+        escape = binary.modifiers.get("escape", None)
+        return "%s NOT ILIKE %s" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        ) + (
+            " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE)
+            if escape is not None
+            else ""
+        )
+
+    def _regexp_match(self, base_op, binary, operator, kw):
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return self._generate_generic_binary(
+                binary, " %s " % base_op, **kw
+            )
+        if flags == "i":
+            return self._generate_generic_binary(
+                binary, " %s* " % base_op, **kw
+            )
+        return "%s %s CONCAT('(?', %s, ')', %s)" % (
+            self.process(binary.left, **kw),
+            base_op,
+            self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._regexp_match("~", binary, operator, kw)
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._regexp_match("!~", binary, operator, kw)
+
+    def visit_regexp_replace_op_binary(self, binary, operator, **kw):
+        string = self.process(binary.left, **kw)
+        pattern_replace = self.process(binary.right, **kw)
+        flags = binary.modifiers["flags"]
+        if flags is None:
+            return "REGEXP_REPLACE(%s, %s)" % (
+                string,
+                pattern_replace,
+            )
+        else:
+            return "REGEXP_REPLACE(%s, %s, %s)" % (
+                string,
+                pattern_replace,
+                self.render_literal_value(flags, sqltypes.STRINGTYPE),
+            )
+
+    def visit_empty_set_expr(self, element_types, **kw):
+        # cast the empty set to the type we are comparing against.  if
+        # we are comparing against the null type, pick an arbitrary
+        # datatype for the empty set
+        return "SELECT %s WHERE 1!=1" % (
+            ", ".join(
+                "CAST(NULL AS %s)"
+                % self.dialect.type_compiler_instance.process(
+                    INTEGER() if type_._isnull else type_
+                )
+                for type_ in element_types or [INTEGER()]
+            ),
+        )
+
+    def render_literal_value(self, value, type_):
+        value = super().render_literal_value(value, type_)
+
+        if self.dialect._backslash_escapes:
+            value = value.replace("\\", "\\\\")
+        return value
+
+    def visit_aggregate_strings_func(self, fn, **kw):
+        return "string_agg%s" % self.function_argspec(fn)
+
+    def visit_sequence(self, seq, **kw):
+        return "nextval('%s')" % self.preparer.format_sequence(seq)
+
+    def limit_clause(self, select, **kw):
+        text = ""
+        if select._limit_clause is not None:
+            text += " \n LIMIT " + self.process(select._limit_clause, **kw)
+        if select._offset_clause is not None:
+            if select._limit_clause is None:
+                text += "\n LIMIT ALL"
+            text += " OFFSET " + self.process(select._offset_clause, **kw)
+        return text
+
+    def format_from_hint_text(self, sqltext, table, hint, iscrud):
+        if hint.upper() != "ONLY":
+            raise exc.CompileError("Unrecognized hint: %r" % hint)
+        return "ONLY " + sqltext
+
+    def get_select_precolumns(self, select, **kw):
+        # Do not call super().get_select_precolumns because
+        # it will warn/raise when distinct on is present
+        if select._distinct or select._distinct_on:
+            if select._distinct_on:
+                return (
+                    "DISTINCT ON ("
+                    + ", ".join(
+                        [
+                            self.process(col, **kw)
+                            for col in select._distinct_on
+                        ]
+                    )
+                    + ") "
+                )
+            else:
+                return "DISTINCT "
+        else:
+            return ""
+
+    def for_update_clause(self, select, **kw):
+        if select._for_update_arg.read:
+            if select._for_update_arg.key_share:
+                tmp = " FOR KEY SHARE"
+            else:
+                tmp = " FOR SHARE"
+        elif select._for_update_arg.key_share:
+            tmp = " FOR NO KEY UPDATE"
+        else:
+            tmp = " FOR UPDATE"
+
+        if select._for_update_arg.of:
+            tables = util.OrderedSet()
+            for c in select._for_update_arg.of:
+                tables.update(sql_util.surface_selectables_only(c))
+
+            tmp += " OF " + ", ".join(
+                self.process(table, ashint=True, use_schema=False, **kw)
+                for table in tables
+            )
+
+        if select._for_update_arg.nowait:
+            tmp += " NOWAIT"
+        if select._for_update_arg.skip_locked:
+            tmp += " SKIP LOCKED"
+
+        return tmp
+
+    def visit_substring_func(self, func, **kw):
+        s = self.process(func.clauses.clauses[0], **kw)
+        start = self.process(func.clauses.clauses[1], **kw)
+        if len(func.clauses.clauses) > 2:
+            length = self.process(func.clauses.clauses[2], **kw)
+            return "SUBSTRING(%s FROM %s FOR %s)" % (s, start, length)
+        else:
+            return "SUBSTRING(%s FROM %s)" % (s, start)
+
+    def _on_conflict_target(self, clause, **kw):
+        if clause.constraint_target is not None:
+            # target may be a name of an Index, UniqueConstraint or
+            # ExcludeConstraint.  While there is a separate
+            # "max_identifier_length" for indexes, PostgreSQL uses the same
+            # length for all objects so we can use
+            # truncate_and_render_constraint_name
+            target_text = (
+                "ON CONSTRAINT %s"
+                % self.preparer.truncate_and_render_constraint_name(
+                    clause.constraint_target
+                )
+            )
+        elif clause.inferred_target_elements is not None:
+            target_text = "(%s)" % ", ".join(
+                (
+                    self.preparer.quote(c)
+                    if isinstance(c, str)
+                    else self.process(c, include_table=False, use_schema=False)
+                )
+                for c in clause.inferred_target_elements
+            )
+            if clause.inferred_target_whereclause is not None:
+                target_text += " WHERE %s" % self.process(
+                    clause.inferred_target_whereclause,
+                    include_table=False,
+                    use_schema=False,
+                )
+        else:
+            target_text = ""
+
+        return target_text
+
+    def visit_on_conflict_do_nothing(self, on_conflict, **kw):
+        target_text = self._on_conflict_target(on_conflict, **kw)
+
+        if target_text:
+            return "ON CONFLICT %s DO NOTHING" % target_text
+        else:
+            return "ON CONFLICT DO NOTHING"
+
+    def visit_on_conflict_do_update(self, on_conflict, **kw):
+        clause = on_conflict
+
+        target_text = self._on_conflict_target(on_conflict, **kw)
+
+        action_set_ops = []
+
+        set_parameters = dict(clause.update_values_to_set)
+        # create a list of column assignment clauses as tuples
+
+        insert_statement = self.stack[-1]["selectable"]
+        cols = insert_statement.table.c
+        for c in cols:
+            col_key = c.key
+
+            if col_key in set_parameters:
+                value = set_parameters.pop(col_key)
+            elif c in set_parameters:
+                value = set_parameters.pop(c)
+            else:
+                continue
+
+            # TODO: this coercion should be up front.  we can't cache
+            # SQL constructs with non-bound literals buried in them
+            if coercions._is_literal(value):
+                value = elements.BindParameter(None, value, type_=c.type)
+
+            else:
+                if (
+                    isinstance(value, elements.BindParameter)
+                    and value.type._isnull
+                ):
+                    value = value._clone()
+                    value.type = c.type
+            value_text = self.process(value.self_group(), use_schema=False)
+
+            key_text = self.preparer.quote(c.name)
+            action_set_ops.append("%s = %s" % (key_text, value_text))
+
+        # check for names that don't match columns
+        if set_parameters:
+            util.warn(
+                "Additional column names not matching "
+                "any column keys in table '%s': %s"
+                % (
+                    self.current_executable.table.name,
+                    (", ".join("'%s'" % c for c in set_parameters)),
+                )
+            )
+            for k, v in set_parameters.items():
+                key_text = (
+                    self.preparer.quote(k)
+                    if isinstance(k, str)
+                    else self.process(k, use_schema=False)
+                )
+                value_text = self.process(
+                    coercions.expect(roles.ExpressionElementRole, v),
+                    use_schema=False,
+                )
+                action_set_ops.append("%s = %s" % (key_text, value_text))
+
+        action_text = ", ".join(action_set_ops)
+        if clause.update_whereclause is not None:
+            action_text += " WHERE %s" % self.process(
+                clause.update_whereclause, include_table=True, use_schema=False
+            )
+
+        return "ON CONFLICT %s DO UPDATE SET %s" % (target_text, action_text)
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        kw["asfrom"] = True
+        return "FROM " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in extra_froms
+        )
+
+    def delete_extra_from_clause(
+        self, delete_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Render the DELETE .. USING clause specific to PostgreSQL."""
+        kw["asfrom"] = True
+        return "USING " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in extra_froms
+        )
+
+    def fetch_clause(self, select, **kw):
+        # pg requires parens for non literal clauses. It's also required for
+        # bind parameters if a ::type casts is used by the driver (asyncpg),
+        # so it's easiest to just always add it
+        text = ""
+        if select._offset_clause is not None:
+            text += "\n OFFSET (%s) ROWS" % self.process(
+                select._offset_clause, **kw
+            )
+        if select._fetch_clause is not None:
+            text += "\n FETCH FIRST (%s)%s ROWS %s" % (
+                self.process(select._fetch_clause, **kw),
+                " PERCENT" if select._fetch_clause_options["percent"] else "",
+                (
+                    "WITH TIES"
+                    if select._fetch_clause_options["with_ties"]
+                    else "ONLY"
+                ),
+            )
+        return text
+
+
+class PGDDLCompiler(compiler.DDLCompiler):
+    def get_column_specification(self, column, **kwargs):
+        colspec = self.preparer.format_column(column)
+        impl_type = column.type.dialect_impl(self.dialect)
+        if isinstance(impl_type, sqltypes.TypeDecorator):
+            impl_type = impl_type.impl
+
+        has_identity = (
+            column.identity is not None
+            and self.dialect.supports_identity_columns
+        )
+
+        if (
+            column.primary_key
+            and column is column.table._autoincrement_column
+            and (
+                self.dialect.supports_smallserial
+                or not isinstance(impl_type, sqltypes.SmallInteger)
+            )
+            and not has_identity
+            and (
+                column.default is None
+                or (
+                    isinstance(column.default, schema.Sequence)
+                    and column.default.optional
+                )
+            )
+        ):
+            if isinstance(impl_type, sqltypes.BigInteger):
+                colspec += " BIGSERIAL"
+            elif isinstance(impl_type, sqltypes.SmallInteger):
+                colspec += " SMALLSERIAL"
+            else:
+                colspec += " SERIAL"
+        else:
+            colspec += " " + self.dialect.type_compiler_instance.process(
+                column.type,
+                type_expression=column,
+                identifier_preparer=self.preparer,
+            )
+            default = self.get_column_default_string(column)
+            if default is not None:
+                colspec += " DEFAULT " + default
+
+        if column.computed is not None:
+            colspec += " " + self.process(column.computed)
+        if has_identity:
+            colspec += " " + self.process(column.identity)
+
+        if not column.nullable and not has_identity:
+            colspec += " NOT NULL"
+        elif column.nullable and has_identity:
+            colspec += " NULL"
+        return colspec
+
+    def _define_constraint_validity(self, constraint):
+        not_valid = constraint.dialect_options["postgresql"]["not_valid"]
+        return " NOT VALID" if not_valid else ""
+
+    def visit_check_constraint(self, constraint, **kw):
+        if constraint._type_bound:
+            typ = list(constraint.columns)[0].type
+            if (
+                isinstance(typ, sqltypes.ARRAY)
+                and isinstance(typ.item_type, sqltypes.Enum)
+                and not typ.item_type.native_enum
+            ):
+                raise exc.CompileError(
+                    "PostgreSQL dialect cannot produce the CHECK constraint "
+                    "for ARRAY of non-native ENUM; please specify "
+                    "create_constraint=False on this Enum datatype."
+                )
+
+        text = super().visit_check_constraint(constraint)
+        text += self._define_constraint_validity(constraint)
+        return text
+
+    def visit_foreign_key_constraint(self, constraint, **kw):
+        text = super().visit_foreign_key_constraint(constraint)
+        text += self._define_constraint_validity(constraint)
+        return text
+
+    def visit_create_enum_type(self, create, **kw):
+        type_ = create.element
+
+        return "CREATE TYPE %s AS ENUM (%s)" % (
+            self.preparer.format_type(type_),
+            ", ".join(
+                self.sql_compiler.process(sql.literal(e), literal_binds=True)
+                for e in type_.enums
+            ),
+        )
+
+    def visit_drop_enum_type(self, drop, **kw):
+        type_ = drop.element
+
+        return "DROP TYPE %s" % (self.preparer.format_type(type_))
+
+    def visit_create_domain_type(self, create, **kw):
+        domain: DOMAIN = create.element
+
+        options = []
+        if domain.collation is not None:
+            options.append(f"COLLATE {self.preparer.quote(domain.collation)}")
+        if domain.default is not None:
+            default = self.render_default_string(domain.default)
+            options.append(f"DEFAULT {default}")
+        if domain.constraint_name is not None:
+            name = self.preparer.truncate_and_render_constraint_name(
+                domain.constraint_name
+            )
+            options.append(f"CONSTRAINT {name}")
+        if domain.not_null:
+            options.append("NOT NULL")
+        if domain.check is not None:
+            check = self.sql_compiler.process(
+                domain.check, include_table=False, literal_binds=True
+            )
+            options.append(f"CHECK ({check})")
+
+        return (
+            f"CREATE DOMAIN {self.preparer.format_type(domain)} AS "
+            f"{self.type_compiler.process(domain.data_type)} "
+            f"{' '.join(options)}"
+        )
+
+    def visit_drop_domain_type(self, drop, **kw):
+        domain = drop.element
+        return f"DROP DOMAIN {self.preparer.format_type(domain)}"
+
+    def visit_create_index(self, create, **kw):
+        preparer = self.preparer
+        index = create.element
+        self._verify_index_table(index)
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+
+        text += "INDEX "
+
+        if self.dialect._supports_create_index_concurrently:
+            concurrently = index.dialect_options["postgresql"]["concurrently"]
+            if concurrently:
+                text += "CONCURRENTLY "
+
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+
+        text += "%s ON %s " % (
+            self._prepared_index_name(index, include_schema=False),
+            preparer.format_table(index.table),
+        )
+
+        using = index.dialect_options["postgresql"]["using"]
+        if using:
+            text += (
+                "USING %s "
+                % self.preparer.validate_sql_phrase(using, IDX_USING).lower()
+            )
+
+        ops = index.dialect_options["postgresql"]["ops"]
+        text += "(%s)" % (
+            ", ".join(
+                [
+                    self.sql_compiler.process(
+                        (
+                            expr.self_group()
+                            if not isinstance(expr, expression.ColumnClause)
+                            else expr
+                        ),
+                        include_table=False,
+                        literal_binds=True,
+                    )
+                    + (
+                        (" " + ops[expr.key])
+                        if hasattr(expr, "key") and expr.key in ops
+                        else ""
+                    )
+                    for expr in index.expressions
+                ]
+            )
+        )
+
+        includeclause = index.dialect_options["postgresql"]["include"]
+        if includeclause:
+            inclusions = [
+                index.table.c[col] if isinstance(col, str) else col
+                for col in includeclause
+            ]
+            text += " INCLUDE (%s)" % ", ".join(
+                [preparer.quote(c.name) for c in inclusions]
+            )
+
+        nulls_not_distinct = index.dialect_options["postgresql"][
+            "nulls_not_distinct"
+        ]
+        if nulls_not_distinct is True:
+            text += " NULLS NOT DISTINCT"
+        elif nulls_not_distinct is False:
+            text += " NULLS DISTINCT"
+
+        withclause = index.dialect_options["postgresql"]["with"]
+        if withclause:
+            text += " WITH (%s)" % (
+                ", ".join(
+                    [
+                        "%s = %s" % storage_parameter
+                        for storage_parameter in withclause.items()
+                    ]
+                )
+            )
+
+        tablespace_name = index.dialect_options["postgresql"]["tablespace"]
+        if tablespace_name:
+            text += " TABLESPACE %s" % preparer.quote(tablespace_name)
+
+        whereclause = index.dialect_options["postgresql"]["where"]
+        if whereclause is not None:
+            whereclause = coercions.expect(
+                roles.DDLExpressionRole, whereclause
+            )
+
+            where_compiled = self.sql_compiler.process(
+                whereclause, include_table=False, literal_binds=True
+            )
+            text += " WHERE " + where_compiled
+
+        return text
+
+    def define_unique_constraint_distinct(self, constraint, **kw):
+        nulls_not_distinct = constraint.dialect_options["postgresql"][
+            "nulls_not_distinct"
+        ]
+        if nulls_not_distinct is True:
+            nulls_not_distinct_param = "NULLS NOT DISTINCT "
+        elif nulls_not_distinct is False:
+            nulls_not_distinct_param = "NULLS DISTINCT "
+        else:
+            nulls_not_distinct_param = ""
+        return nulls_not_distinct_param
+
+    def visit_drop_index(self, drop, **kw):
+        index = drop.element
+
+        text = "\nDROP INDEX "
+
+        if self.dialect._supports_drop_index_concurrently:
+            concurrently = index.dialect_options["postgresql"]["concurrently"]
+            if concurrently:
+                text += "CONCURRENTLY "
+
+        if drop.if_exists:
+            text += "IF EXISTS "
+
+        text += self._prepared_index_name(index, include_schema=True)
+        return text
+
+    def visit_exclude_constraint(self, constraint, **kw):
+        text = ""
+        if constraint.name is not None:
+            text += "CONSTRAINT %s " % self.preparer.format_constraint(
+                constraint
+            )
+        elements = []
+        kw["include_table"] = False
+        kw["literal_binds"] = True
+        for expr, name, op in constraint._render_exprs:
+            exclude_element = self.sql_compiler.process(expr, **kw) + (
+                (" " + constraint.ops[expr.key])
+                if hasattr(expr, "key") and expr.key in constraint.ops
+                else ""
+            )
+
+            elements.append("%s WITH %s" % (exclude_element, op))
+        text += "EXCLUDE USING %s (%s)" % (
+            self.preparer.validate_sql_phrase(
+                constraint.using, IDX_USING
+            ).lower(),
+            ", ".join(elements),
+        )
+        if constraint.where is not None:
+            text += " WHERE (%s)" % self.sql_compiler.process(
+                constraint.where, literal_binds=True
+            )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def post_create_table(self, table):
+        table_opts = []
+        pg_opts = table.dialect_options["postgresql"]
+
+        inherits = pg_opts.get("inherits")
+        if inherits is not None:
+            if not isinstance(inherits, (list, tuple)):
+                inherits = (inherits,)
+            table_opts.append(
+                "\n INHERITS ( "
+                + ", ".join(self.preparer.quote(name) for name in inherits)
+                + " )"
+            )
+
+        if pg_opts["partition_by"]:
+            table_opts.append("\n PARTITION BY %s" % pg_opts["partition_by"])
+
+        if pg_opts["using"]:
+            table_opts.append("\n USING %s" % pg_opts["using"])
+
+        if pg_opts["with_oids"] is True:
+            table_opts.append("\n WITH OIDS")
+        elif pg_opts["with_oids"] is False:
+            table_opts.append("\n WITHOUT OIDS")
+
+        if pg_opts["on_commit"]:
+            on_commit_options = pg_opts["on_commit"].replace("_", " ").upper()
+            table_opts.append("\n ON COMMIT %s" % on_commit_options)
+
+        if pg_opts["tablespace"]:
+            tablespace_name = pg_opts["tablespace"]
+            table_opts.append(
+                "\n TABLESPACE %s" % self.preparer.quote(tablespace_name)
+            )
+
+        return "".join(table_opts)
+
+    def visit_computed_column(self, generated, **kw):
+        if generated.persisted is False:
+            raise exc.CompileError(
+                "PostrgreSQL computed columns do not support 'virtual' "
+                "persistence; set the 'persisted' flag to None or True for "
+                "PostgreSQL support."
+            )
+
+        return "GENERATED ALWAYS AS (%s) STORED" % self.sql_compiler.process(
+            generated.sqltext, include_table=False, literal_binds=True
+        )
+
+    def visit_create_sequence(self, create, **kw):
+        prefix = None
+        if create.element.data_type is not None:
+            prefix = " AS %s" % self.type_compiler.process(
+                create.element.data_type
+            )
+
+        return super().visit_create_sequence(create, prefix=prefix, **kw)
+
+    def _can_comment_on_constraint(self, ddl_instance):
+        constraint = ddl_instance.element
+        if constraint.name is None:
+            raise exc.CompileError(
+                f"Can't emit COMMENT ON for constraint {constraint!r}: "
+                "it has no name"
+            )
+        if constraint.table is None:
+            raise exc.CompileError(
+                f"Can't emit COMMENT ON for constraint {constraint!r}: "
+                "it has no associated table"
+            )
+
+    def visit_set_constraint_comment(self, create, **kw):
+        self._can_comment_on_constraint(create)
+        return "COMMENT ON CONSTRAINT %s ON %s IS %s" % (
+            self.preparer.format_constraint(create.element),
+            self.preparer.format_table(create.element.table),
+            self.sql_compiler.render_literal_value(
+                create.element.comment, sqltypes.String()
+            ),
+        )
+
+    def visit_drop_constraint_comment(self, drop, **kw):
+        self._can_comment_on_constraint(drop)
+        return "COMMENT ON CONSTRAINT %s ON %s IS NULL" % (
+            self.preparer.format_constraint(drop.element),
+            self.preparer.format_table(drop.element.table),
+        )
+
+
+class PGTypeCompiler(compiler.GenericTypeCompiler):
+    def visit_TSVECTOR(self, type_, **kw):
+        return "TSVECTOR"
+
+    def visit_TSQUERY(self, type_, **kw):
+        return "TSQUERY"
+
+    def visit_INET(self, type_, **kw):
+        return "INET"
+
+    def visit_CIDR(self, type_, **kw):
+        return "CIDR"
+
+    def visit_CITEXT(self, type_, **kw):
+        return "CITEXT"
+
+    def visit_MACADDR(self, type_, **kw):
+        return "MACADDR"
+
+    def visit_MACADDR8(self, type_, **kw):
+        return "MACADDR8"
+
+    def visit_MONEY(self, type_, **kw):
+        return "MONEY"
+
+    def visit_OID(self, type_, **kw):
+        return "OID"
+
+    def visit_REGCONFIG(self, type_, **kw):
+        return "REGCONFIG"
+
+    def visit_REGCLASS(self, type_, **kw):
+        return "REGCLASS"
+
+    def visit_FLOAT(self, type_, **kw):
+        if not type_.precision:
+            return "FLOAT"
+        else:
+            return "FLOAT(%(precision)s)" % {"precision": type_.precision}
+
+    def visit_double(self, type_, **kw):
+        return self.visit_DOUBLE_PRECISION(type, **kw)
+
+    def visit_BIGINT(self, type_, **kw):
+        return "BIGINT"
+
+    def visit_HSTORE(self, type_, **kw):
+        return "HSTORE"
+
+    def visit_JSON(self, type_, **kw):
+        return "JSON"
+
+    def visit_JSONB(self, type_, **kw):
+        return "JSONB"
+
+    def visit_INT4MULTIRANGE(self, type_, **kw):
+        return "INT4MULTIRANGE"
+
+    def visit_INT8MULTIRANGE(self, type_, **kw):
+        return "INT8MULTIRANGE"
+
+    def visit_NUMMULTIRANGE(self, type_, **kw):
+        return "NUMMULTIRANGE"
+
+    def visit_DATEMULTIRANGE(self, type_, **kw):
+        return "DATEMULTIRANGE"
+
+    def visit_TSMULTIRANGE(self, type_, **kw):
+        return "TSMULTIRANGE"
+
+    def visit_TSTZMULTIRANGE(self, type_, **kw):
+        return "TSTZMULTIRANGE"
+
+    def visit_INT4RANGE(self, type_, **kw):
+        return "INT4RANGE"
+
+    def visit_INT8RANGE(self, type_, **kw):
+        return "INT8RANGE"
+
+    def visit_NUMRANGE(self, type_, **kw):
+        return "NUMRANGE"
+
+    def visit_DATERANGE(self, type_, **kw):
+        return "DATERANGE"
+
+    def visit_TSRANGE(self, type_, **kw):
+        return "TSRANGE"
+
+    def visit_TSTZRANGE(self, type_, **kw):
+        return "TSTZRANGE"
+
+    def visit_json_int_index(self, type_, **kw):
+        return "INT"
+
+    def visit_json_str_index(self, type_, **kw):
+        return "TEXT"
+
+    def visit_datetime(self, type_, **kw):
+        return self.visit_TIMESTAMP(type_, **kw)
+
+    def visit_enum(self, type_, **kw):
+        if not type_.native_enum or not self.dialect.supports_native_enum:
+            return super().visit_enum(type_, **kw)
+        else:
+            return self.visit_ENUM(type_, **kw)
+
+    def visit_ENUM(self, type_, identifier_preparer=None, **kw):
+        if identifier_preparer is None:
+            identifier_preparer = self.dialect.identifier_preparer
+        return identifier_preparer.format_type(type_)
+
+    def visit_DOMAIN(self, type_, identifier_preparer=None, **kw):
+        if identifier_preparer is None:
+            identifier_preparer = self.dialect.identifier_preparer
+        return identifier_preparer.format_type(type_)
+
+    def visit_TIMESTAMP(self, type_, **kw):
+        return "TIMESTAMP%s %s" % (
+            (
+                "(%d)" % type_.precision
+                if getattr(type_, "precision", None) is not None
+                else ""
+            ),
+            (type_.timezone and "WITH" or "WITHOUT") + " TIME ZONE",
+        )
+
+    def visit_TIME(self, type_, **kw):
+        return "TIME%s %s" % (
+            (
+                "(%d)" % type_.precision
+                if getattr(type_, "precision", None) is not None
+                else ""
+            ),
+            (type_.timezone and "WITH" or "WITHOUT") + " TIME ZONE",
+        )
+
+    def visit_INTERVAL(self, type_, **kw):
+        text = "INTERVAL"
+        if type_.fields is not None:
+            text += " " + type_.fields
+        if type_.precision is not None:
+            text += " (%d)" % type_.precision
+        return text
+
+    def visit_BIT(self, type_, **kw):
+        if type_.varying:
+            compiled = "BIT VARYING"
+            if type_.length is not None:
+                compiled += "(%d)" % type_.length
+        else:
+            compiled = "BIT(%d)" % type_.length
+        return compiled
+
+    def visit_uuid(self, type_, **kw):
+        if type_.native_uuid:
+            return self.visit_UUID(type_, **kw)
+        else:
+            return super().visit_uuid(type_, **kw)
+
+    def visit_UUID(self, type_, **kw):
+        return "UUID"
+
+    def visit_large_binary(self, type_, **kw):
+        return self.visit_BYTEA(type_, **kw)
+
+    def visit_BYTEA(self, type_, **kw):
+        return "BYTEA"
+
+    def visit_ARRAY(self, type_, **kw):
+        inner = self.process(type_.item_type, **kw)
+        return re.sub(
+            r"((?: COLLATE.*)?)$",
+            (
+                r"%s\1"
+                % (
+                    "[]"
+                    * (type_.dimensions if type_.dimensions is not None else 1)
+                )
+            ),
+            inner,
+            count=1,
+        )
+
+    def visit_json_path(self, type_, **kw):
+        return self.visit_JSONPATH(type_, **kw)
+
+    def visit_JSONPATH(self, type_, **kw):
+        return "JSONPATH"
+
+
+class PGIdentifierPreparer(compiler.IdentifierPreparer):
+    reserved_words = RESERVED_WORDS
+
+    def _unquote_identifier(self, value):
+        if value[0] == self.initial_quote:
+            value = value[1:-1].replace(
+                self.escape_to_quote, self.escape_quote
+            )
+        return value
+
+    def format_type(self, type_, use_schema=True):
+        if not type_.name:
+            raise exc.CompileError(
+                f"PostgreSQL {type_.__class__.__name__} type requires a name."
+            )
+
+        name = self.quote(type_.name)
+        effective_schema = self.schema_for_object(type_)
+
+        if (
+            not self.omit_schema
+            and use_schema
+            and effective_schema is not None
+        ):
+            name = f"{self.quote_schema(effective_schema)}.{name}"
+        return name
+
+
+class ReflectedNamedType(TypedDict):
+    """Represents a reflected named type."""
+
+    name: str
+    """Name of the type."""
+    schema: str
+    """The schema of the type."""
+    visible: bool
+    """Indicates if this type is in the current search path."""
+
+
+class ReflectedDomainConstraint(TypedDict):
+    """Represents a reflect check constraint of a domain."""
+
+    name: str
+    """Name of the constraint."""
+    check: str
+    """The check constraint text."""
+
+
+class ReflectedDomain(ReflectedNamedType):
+    """Represents a reflected enum."""
+
+    type: str
+    """The string name of the underlying data type of the domain."""
+    nullable: bool
+    """Indicates if the domain allows null or not."""
+    default: Optional[str]
+    """The string representation of the default value of this domain
+    or ``None`` if none present.
+    """
+    constraints: List[ReflectedDomainConstraint]
+    """The constraints defined in the domain, if any.
+    The constraint are in order of evaluation by postgresql.
+    """
+    collation: Optional[str]
+    """The collation for the domain."""
+
+
+class ReflectedEnum(ReflectedNamedType):
+    """Represents a reflected enum."""
+
+    labels: List[str]
+    """The labels that compose the enum."""
+
+
+class PGInspector(reflection.Inspector):
+    dialect: PGDialect
+
+    def get_table_oid(
+        self, table_name: str, schema: Optional[str] = None
+    ) -> int:
+        """Return the OID for the given table name.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_table_oid(
+                conn, table_name, schema, info_cache=self.info_cache
+            )
+
+    def get_domains(
+        self, schema: Optional[str] = None
+    ) -> List[ReflectedDomain]:
+        """Return a list of DOMAIN objects.
+
+        Each member is a dictionary containing these fields:
+
+            * name - name of the domain
+            * schema - the schema name for the domain.
+            * visible - boolean, whether or not this domain is visible
+              in the default search path.
+            * type - the type defined by this domain.
+            * nullable - Indicates if this domain can be ``NULL``.
+            * default - The default value of the domain or ``None`` if the
+              domain has no default.
+            * constraints - A list of dict wit the constraint defined by this
+              domain. Each element constaints two keys: ``name`` of the
+              constraint and ``check`` with the constraint text.
+
+        :param schema: schema name.  If None, the default schema
+         (typically 'public') is used.  May also be set to ``'*'`` to
+         indicate load domains for all schemas.
+
+        .. versionadded:: 2.0
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect._load_domains(
+                conn, schema, info_cache=self.info_cache
+            )
+
+    def get_enums(self, schema: Optional[str] = None) -> List[ReflectedEnum]:
+        """Return a list of ENUM objects.
+
+        Each member is a dictionary containing these fields:
+
+            * name - name of the enum
+            * schema - the schema name for the enum.
+            * visible - boolean, whether or not this enum is visible
+              in the default search path.
+            * labels - a list of string labels that apply to the enum.
+
+        :param schema: schema name.  If None, the default schema
+         (typically 'public') is used.  May also be set to ``'*'`` to
+         indicate load enums for all schemas.
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect._load_enums(
+                conn, schema, info_cache=self.info_cache
+            )
+
+    def get_foreign_table_names(
+        self, schema: Optional[str] = None
+    ) -> List[str]:
+        """Return a list of FOREIGN TABLE names.
+
+        Behavior is similar to that of
+        :meth:`_reflection.Inspector.get_table_names`,
+        except that the list is limited to those tables that report a
+        ``relkind`` value of ``f``.
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect._get_foreign_table_names(
+                conn, schema, info_cache=self.info_cache
+            )
+
+    def has_type(
+        self, type_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> bool:
+        """Return if the database has the specified type in the provided
+        schema.
+
+        :param type_name: the type to check.
+        :param schema: schema name.  If None, the default schema
+         (typically 'public') is used.  May also be set to ``'*'`` to
+         check in all schemas.
+
+        .. versionadded:: 2.0
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.has_type(
+                conn, type_name, schema, info_cache=self.info_cache
+            )
+
+
+class PGExecutionContext(default.DefaultExecutionContext):
+    def fire_sequence(self, seq, type_):
+        return self._execute_scalar(
+            (
+                "select nextval('%s')"
+                % self.identifier_preparer.format_sequence(seq)
+            ),
+            type_,
+        )
+
+    def get_insert_default(self, column):
+        if column.primary_key and column is column.table._autoincrement_column:
+            if column.server_default and column.server_default.has_argument:
+                # pre-execute passive defaults on primary key columns
+                return self._execute_scalar(
+                    "select %s" % column.server_default.arg, column.type
+                )
+
+            elif column.default is None or (
+                column.default.is_sequence and column.default.optional
+            ):
+                # execute the sequence associated with a SERIAL primary
+                # key column. for non-primary-key SERIAL, the ID just
+                # generates server side.
+
+                try:
+                    seq_name = column._postgresql_seq_name
+                except AttributeError:
+                    tab = column.table.name
+                    col = column.name
+                    tab = tab[0 : 29 + max(0, (29 - len(col)))]
+                    col = col[0 : 29 + max(0, (29 - len(tab)))]
+                    name = "%s_%s_seq" % (tab, col)
+                    column._postgresql_seq_name = seq_name = name
+
+                if column.table is not None:
+                    effective_schema = self.connection.schema_for_object(
+                        column.table
+                    )
+                else:
+                    effective_schema = None
+
+                if effective_schema is not None:
+                    exc = 'select nextval(\'"%s"."%s"\')' % (
+                        effective_schema,
+                        seq_name,
+                    )
+                else:
+                    exc = "select nextval('\"%s\"')" % (seq_name,)
+
+                return self._execute_scalar(exc, column.type)
+
+        return super().get_insert_default(column)
+
+
+class PGReadOnlyConnectionCharacteristic(
+    characteristics.ConnectionCharacteristic
+):
+    transactional = True
+
+    def reset_characteristic(self, dialect, dbapi_conn):
+        dialect.set_readonly(dbapi_conn, False)
+
+    def set_characteristic(self, dialect, dbapi_conn, value):
+        dialect.set_readonly(dbapi_conn, value)
+
+    def get_characteristic(self, dialect, dbapi_conn):
+        return dialect.get_readonly(dbapi_conn)
+
+
+class PGDeferrableConnectionCharacteristic(
+    characteristics.ConnectionCharacteristic
+):
+    transactional = True
+
+    def reset_characteristic(self, dialect, dbapi_conn):
+        dialect.set_deferrable(dbapi_conn, False)
+
+    def set_characteristic(self, dialect, dbapi_conn, value):
+        dialect.set_deferrable(dbapi_conn, value)
+
+    def get_characteristic(self, dialect, dbapi_conn):
+        return dialect.get_deferrable(dbapi_conn)
+
+
+class PGDialect(default.DefaultDialect):
+    name = "postgresql"
+    supports_statement_cache = True
+    supports_alter = True
+    max_identifier_length = 63
+    supports_sane_rowcount = True
+
+    bind_typing = interfaces.BindTyping.RENDER_CASTS
+
+    supports_native_enum = True
+    supports_native_boolean = True
+    supports_native_uuid = True
+    supports_smallserial = True
+
+    supports_sequences = True
+    sequences_optional = True
+    preexecute_autoincrement_sequences = True
+    postfetch_lastrowid = False
+    use_insertmanyvalues = True
+
+    returns_native_bytes = True
+
+    insertmanyvalues_implicit_sentinel = (
+        InsertmanyvaluesSentinelOpts.ANY_AUTOINCREMENT
+        | InsertmanyvaluesSentinelOpts.USE_INSERT_FROM_SELECT
+        | InsertmanyvaluesSentinelOpts.RENDER_SELECT_COL_CASTS
+    )
+
+    supports_comments = True
+    supports_constraint_comments = True
+    supports_default_values = True
+
+    supports_default_metavalue = True
+
+    supports_empty_insert = False
+    supports_multivalues_insert = True
+
+    supports_identity_columns = True
+
+    default_paramstyle = "pyformat"
+    ischema_names = ischema_names
+    colspecs = colspecs
+
+    statement_compiler = PGCompiler
+    ddl_compiler = PGDDLCompiler
+    type_compiler_cls = PGTypeCompiler
+    preparer = PGIdentifierPreparer
+    execution_ctx_cls = PGExecutionContext
+    inspector = PGInspector
+
+    update_returning = True
+    delete_returning = True
+    insert_returning = True
+    update_returning_multifrom = True
+    delete_returning_multifrom = True
+
+    connection_characteristics = (
+        default.DefaultDialect.connection_characteristics
+    )
+    connection_characteristics = connection_characteristics.union(
+        {
+            "postgresql_readonly": PGReadOnlyConnectionCharacteristic(),
+            "postgresql_deferrable": PGDeferrableConnectionCharacteristic(),
+        }
+    )
+
+    construct_arguments = [
+        (
+            schema.Index,
+            {
+                "using": False,
+                "include": None,
+                "where": None,
+                "ops": {},
+                "concurrently": False,
+                "with": {},
+                "tablespace": None,
+                "nulls_not_distinct": None,
+            },
+        ),
+        (
+            schema.Table,
+            {
+                "ignore_search_path": False,
+                "tablespace": None,
+                "partition_by": None,
+                "with_oids": None,
+                "on_commit": None,
+                "inherits": None,
+                "using": None,
+            },
+        ),
+        (
+            schema.CheckConstraint,
+            {
+                "not_valid": False,
+            },
+        ),
+        (
+            schema.ForeignKeyConstraint,
+            {
+                "not_valid": False,
+            },
+        ),
+        (
+            schema.UniqueConstraint,
+            {"nulls_not_distinct": None},
+        ),
+    ]
+
+    reflection_options = ("postgresql_ignore_search_path",)
+
+    _backslash_escapes = True
+    _supports_create_index_concurrently = True
+    _supports_drop_index_concurrently = True
+
+    def __init__(
+        self,
+        native_inet_types=None,
+        json_serializer=None,
+        json_deserializer=None,
+        **kwargs,
+    ):
+        default.DefaultDialect.__init__(self, **kwargs)
+
+        self._native_inet_types = native_inet_types
+        self._json_deserializer = json_deserializer
+        self._json_serializer = json_serializer
+
+    def initialize(self, connection):
+        super().initialize(connection)
+
+        # https://www.postgresql.org/docs/9.3/static/release-9-2.html#AEN116689
+        self.supports_smallserial = self.server_version_info >= (9, 2)
+
+        self._set_backslash_escapes(connection)
+
+        self._supports_drop_index_concurrently = self.server_version_info >= (
+            9,
+            2,
+        )
+        self.supports_identity_columns = self.server_version_info >= (10,)
+
+    def get_isolation_level_values(self, dbapi_conn):
+        # note the generic dialect doesn't have AUTOCOMMIT, however
+        # all postgresql dialects should include AUTOCOMMIT.
+        return (
+            "SERIALIZABLE",
+            "READ UNCOMMITTED",
+            "READ COMMITTED",
+            "REPEATABLE READ",
+        )
+
+    def set_isolation_level(self, dbapi_connection, level):
+        cursor = dbapi_connection.cursor()
+        cursor.execute(
+            "SET SESSION CHARACTERISTICS AS TRANSACTION "
+            f"ISOLATION LEVEL {level}"
+        )
+        cursor.execute("COMMIT")
+        cursor.close()
+
+    def get_isolation_level(self, dbapi_connection):
+        cursor = dbapi_connection.cursor()
+        cursor.execute("show transaction isolation level")
+        val = cursor.fetchone()[0]
+        cursor.close()
+        return val.upper()
+
+    def set_readonly(self, connection, value):
+        raise NotImplementedError()
+
+    def get_readonly(self, connection):
+        raise NotImplementedError()
+
+    def set_deferrable(self, connection, value):
+        raise NotImplementedError()
+
+    def get_deferrable(self, connection):
+        raise NotImplementedError()
+
+    def _split_multihost_from_url(self, url: URL) -> Union[
+        Tuple[None, None],
+        Tuple[Tuple[Optional[str], ...], Tuple[Optional[int], ...]],
+    ]:
+        hosts: Optional[Tuple[Optional[str], ...]] = None
+        ports_str: Union[str, Tuple[Optional[str], ...], None] = None
+
+        integrated_multihost = False
+
+        if "host" in url.query:
+            if isinstance(url.query["host"], (list, tuple)):
+                integrated_multihost = True
+                hosts, ports_str = zip(
+                    *[
+                        token.split(":") if ":" in token else (token, None)
+                        for token in url.query["host"]
+                    ]
+                )
+
+            elif isinstance(url.query["host"], str):
+                hosts = tuple(url.query["host"].split(","))
+
+                if (
+                    "port" not in url.query
+                    and len(hosts) == 1
+                    and ":" in hosts[0]
+                ):
+                    # internet host is alphanumeric plus dots or hyphens.
+                    # this is essentially rfc1123, which refers to rfc952.
+                    # https://stackoverflow.com/questions/3523028/
+                    # valid-characters-of-a-hostname
+                    host_port_match = re.match(
+                        r"^([a-zA-Z0-9\-\.]*)(?:\:(\d*))?$", hosts[0]
+                    )
+                    if host_port_match:
+                        integrated_multihost = True
+                        h, p = host_port_match.group(1, 2)
+                        if TYPE_CHECKING:
+                            assert isinstance(h, str)
+                            assert isinstance(p, str)
+                        hosts = (h,)
+                        ports_str = cast(
+                            "Tuple[Optional[str], ...]", (p,) if p else (None,)
+                        )
+
+        if "port" in url.query:
+            if integrated_multihost:
+                raise exc.ArgumentError(
+                    "Can't mix 'multihost' formats together; use "
+                    '"host=h1,h2,h3&port=p1,p2,p3" or '
+                    '"host=h1:p1&host=h2:p2&host=h3:p3" separately'
+                )
+            if isinstance(url.query["port"], (list, tuple)):
+                ports_str = url.query["port"]
+            elif isinstance(url.query["port"], str):
+                ports_str = tuple(url.query["port"].split(","))
+
+        ports: Optional[Tuple[Optional[int], ...]] = None
+
+        if ports_str:
+            try:
+                ports = tuple(int(x) if x else None for x in ports_str)
+            except ValueError:
+                raise exc.ArgumentError(
+                    f"Received non-integer port arguments: {ports_str}"
+                ) from None
+
+        if ports and (
+            (not hosts and len(ports) > 1)
+            or (
+                hosts
+                and ports
+                and len(hosts) != len(ports)
+                and (len(hosts) > 1 or len(ports) > 1)
+            )
+        ):
+            raise exc.ArgumentError("number of hosts and ports don't match")
+
+        if hosts is not None:
+            if ports is None:
+                ports = tuple(None for _ in hosts)
+
+        return hosts, ports  # type: ignore
+
+    def do_begin_twophase(self, connection, xid):
+        self.do_begin(connection.connection)
+
+    def do_prepare_twophase(self, connection, xid):
+        connection.exec_driver_sql("PREPARE TRANSACTION '%s'" % xid)
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if is_prepared:
+            if recover:
+                # FIXME: ugly hack to get out of transaction
+                # context when committing recoverable transactions
+                # Must find out a way how to make the dbapi not
+                # open a transaction.
+                connection.exec_driver_sql("ROLLBACK")
+            connection.exec_driver_sql("ROLLBACK PREPARED '%s'" % xid)
+            connection.exec_driver_sql("BEGIN")
+            self.do_rollback(connection.connection)
+        else:
+            self.do_rollback(connection.connection)
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if is_prepared:
+            if recover:
+                connection.exec_driver_sql("ROLLBACK")
+            connection.exec_driver_sql("COMMIT PREPARED '%s'" % xid)
+            connection.exec_driver_sql("BEGIN")
+            self.do_rollback(connection.connection)
+        else:
+            self.do_commit(connection.connection)
+
+    def do_recover_twophase(self, connection):
+        return connection.scalars(
+            sql.text("SELECT gid FROM pg_prepared_xacts")
+        ).all()
+
+    def _get_default_schema_name(self, connection):
+        return connection.exec_driver_sql("select current_schema()").scalar()
+
+    @reflection.cache
+    def has_schema(self, connection, schema, **kw):
+        query = select(pg_catalog.pg_namespace.c.nspname).where(
+            pg_catalog.pg_namespace.c.nspname == schema
+        )
+        return bool(connection.scalar(query))
+
+    def _pg_class_filter_scope_schema(
+        self, query, schema, scope, pg_class_table=None
+    ):
+        if pg_class_table is None:
+            pg_class_table = pg_catalog.pg_class
+        query = query.join(
+            pg_catalog.pg_namespace,
+            pg_catalog.pg_namespace.c.oid == pg_class_table.c.relnamespace,
+        )
+
+        if scope is ObjectScope.DEFAULT:
+            query = query.where(pg_class_table.c.relpersistence != "t")
+        elif scope is ObjectScope.TEMPORARY:
+            query = query.where(pg_class_table.c.relpersistence == "t")
+
+        if schema is None:
+            query = query.where(
+                pg_catalog.pg_table_is_visible(pg_class_table.c.oid),
+                # ignore pg_catalog schema
+                pg_catalog.pg_namespace.c.nspname != "pg_catalog",
+            )
+        else:
+            query = query.where(pg_catalog.pg_namespace.c.nspname == schema)
+        return query
+
+    def _pg_class_relkind_condition(self, relkinds, pg_class_table=None):
+        if pg_class_table is None:
+            pg_class_table = pg_catalog.pg_class
+        # uses the any form instead of in otherwise postgresql complaings
+        # that 'IN could not convert type character to "char"'
+        return pg_class_table.c.relkind == sql.any_(_array.array(relkinds))
+
+    @lru_cache()
+    def _has_table_query(self, schema):
+        query = select(pg_catalog.pg_class.c.relname).where(
+            pg_catalog.pg_class.c.relname == bindparam("table_name"),
+            self._pg_class_relkind_condition(
+                pg_catalog.RELKINDS_ALL_TABLE_LIKE
+            ),
+        )
+        return self._pg_class_filter_scope_schema(
+            query, schema, scope=ObjectScope.ANY
+        )
+
+    @reflection.cache
+    def has_table(self, connection, table_name, schema=None, **kw):
+        self._ensure_has_table_connection(connection)
+        query = self._has_table_query(schema)
+        return bool(connection.scalar(query, {"table_name": table_name}))
+
+    @reflection.cache
+    def has_sequence(self, connection, sequence_name, schema=None, **kw):
+        query = select(pg_catalog.pg_class.c.relname).where(
+            pg_catalog.pg_class.c.relkind == "S",
+            pg_catalog.pg_class.c.relname == sequence_name,
+        )
+        query = self._pg_class_filter_scope_schema(
+            query, schema, scope=ObjectScope.ANY
+        )
+        return bool(connection.scalar(query))
+
+    @reflection.cache
+    def has_type(self, connection, type_name, schema=None, **kw):
+        query = (
+            select(pg_catalog.pg_type.c.typname)
+            .join(
+                pg_catalog.pg_namespace,
+                pg_catalog.pg_namespace.c.oid
+                == pg_catalog.pg_type.c.typnamespace,
+            )
+            .where(pg_catalog.pg_type.c.typname == type_name)
+        )
+        if schema is None:
+            query = query.where(
+                pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid),
+                # ignore pg_catalog schema
+                pg_catalog.pg_namespace.c.nspname != "pg_catalog",
+            )
+        elif schema != "*":
+            query = query.where(pg_catalog.pg_namespace.c.nspname == schema)
+
+        return bool(connection.scalar(query))
+
+    def _get_server_version_info(self, connection):
+        v = connection.exec_driver_sql("select pg_catalog.version()").scalar()
+        m = re.match(
+            r".*(?:PostgreSQL|EnterpriseDB) "
+            r"(\d+)\.?(\d+)?(?:\.(\d+))?(?:\.\d+)?(?:devel|beta)?",
+            v,
+        )
+        if not m:
+            raise AssertionError(
+                "Could not determine version from string '%s'" % v
+            )
+        return tuple([int(x) for x in m.group(1, 2, 3) if x is not None])
+
+    @reflection.cache
+    def get_table_oid(self, connection, table_name, schema=None, **kw):
+        """Fetch the oid for schema.table_name."""
+        query = select(pg_catalog.pg_class.c.oid).where(
+            pg_catalog.pg_class.c.relname == table_name,
+            self._pg_class_relkind_condition(
+                pg_catalog.RELKINDS_ALL_TABLE_LIKE
+            ),
+        )
+        query = self._pg_class_filter_scope_schema(
+            query, schema, scope=ObjectScope.ANY
+        )
+        table_oid = connection.scalar(query)
+        if table_oid is None:
+            raise exc.NoSuchTableError(
+                f"{schema}.{table_name}" if schema else table_name
+            )
+        return table_oid
+
+    @reflection.cache
+    def get_schema_names(self, connection, **kw):
+        query = (
+            select(pg_catalog.pg_namespace.c.nspname)
+            .where(pg_catalog.pg_namespace.c.nspname.not_like("pg_%"))
+            .order_by(pg_catalog.pg_namespace.c.nspname)
+        )
+        return connection.scalars(query).all()
+
+    def _get_relnames_for_relkinds(self, connection, schema, relkinds, scope):
+        query = select(pg_catalog.pg_class.c.relname).where(
+            self._pg_class_relkind_condition(relkinds)
+        )
+        query = self._pg_class_filter_scope_schema(query, schema, scope=scope)
+        return connection.scalars(query).all()
+
+    @reflection.cache
+    def get_table_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection,
+            schema,
+            pg_catalog.RELKINDS_TABLE_NO_FOREIGN,
+            scope=ObjectScope.DEFAULT,
+        )
+
+    @reflection.cache
+    def get_temp_table_names(self, connection, **kw):
+        return self._get_relnames_for_relkinds(
+            connection,
+            schema=None,
+            relkinds=pg_catalog.RELKINDS_TABLE_NO_FOREIGN,
+            scope=ObjectScope.TEMPORARY,
+        )
+
+    @reflection.cache
+    def _get_foreign_table_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection, schema, relkinds=("f",), scope=ObjectScope.ANY
+        )
+
+    @reflection.cache
+    def get_view_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection,
+            schema,
+            pg_catalog.RELKINDS_VIEW,
+            scope=ObjectScope.DEFAULT,
+        )
+
+    @reflection.cache
+    def get_materialized_view_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection,
+            schema,
+            pg_catalog.RELKINDS_MAT_VIEW,
+            scope=ObjectScope.DEFAULT,
+        )
+
+    @reflection.cache
+    def get_temp_view_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection,
+            schema,
+            # NOTE: do not include temp materialzied views (that do not
+            # seem to be a thing at least up to version 14)
+            pg_catalog.RELKINDS_VIEW,
+            scope=ObjectScope.TEMPORARY,
+        )
+
+    @reflection.cache
+    def get_sequence_names(self, connection, schema=None, **kw):
+        return self._get_relnames_for_relkinds(
+            connection, schema, relkinds=("S",), scope=ObjectScope.ANY
+        )
+
+    @reflection.cache
+    def get_view_definition(self, connection, view_name, schema=None, **kw):
+        query = (
+            select(pg_catalog.pg_get_viewdef(pg_catalog.pg_class.c.oid))
+            .select_from(pg_catalog.pg_class)
+            .where(
+                pg_catalog.pg_class.c.relname == view_name,
+                self._pg_class_relkind_condition(
+                    pg_catalog.RELKINDS_VIEW + pg_catalog.RELKINDS_MAT_VIEW
+                ),
+            )
+        )
+        query = self._pg_class_filter_scope_schema(
+            query, schema, scope=ObjectScope.ANY
+        )
+        res = connection.scalar(query)
+        if res is None:
+            raise exc.NoSuchTableError(
+                f"{schema}.{view_name}" if schema else view_name
+            )
+        else:
+            return res
+
+    def _value_or_raise(self, data, table, schema):
+        try:
+            return dict(data)[(schema, table)]
+        except KeyError:
+            raise exc.NoSuchTableError(
+                f"{schema}.{table}" if schema else table
+            ) from None
+
+    def _prepare_filter_names(self, filter_names):
+        if filter_names:
+            return True, {"filter_names": filter_names}
+        else:
+            return False, {}
+
+    def _kind_to_relkinds(self, kind: ObjectKind) -> Tuple[str, ...]:
+        if kind is ObjectKind.ANY:
+            return pg_catalog.RELKINDS_ALL_TABLE_LIKE
+        relkinds = ()
+        if ObjectKind.TABLE in kind:
+            relkinds += pg_catalog.RELKINDS_TABLE
+        if ObjectKind.VIEW in kind:
+            relkinds += pg_catalog.RELKINDS_VIEW
+        if ObjectKind.MATERIALIZED_VIEW in kind:
+            relkinds += pg_catalog.RELKINDS_MAT_VIEW
+        return relkinds
+
+    @reflection.cache
+    def get_columns(self, connection, table_name, schema=None, **kw):
+        data = self.get_multi_columns(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _columns_query(self, schema, has_filter_names, scope, kind):
+        # NOTE: the query with the default and identity options scalar
+        # subquery is faster than trying to use outer joins for them
+        generated = (
+            pg_catalog.pg_attribute.c.attgenerated.label("generated")
+            if self.server_version_info >= (12,)
+            else sql.null().label("generated")
+        )
+        if self.server_version_info >= (10,):
+            # join lateral performs worse (~2x slower) than a scalar_subquery
+            identity = (
+                select(
+                    sql.func.json_build_object(
+                        "always",
+                        pg_catalog.pg_attribute.c.attidentity == "a",
+                        "start",
+                        pg_catalog.pg_sequence.c.seqstart,
+                        "increment",
+                        pg_catalog.pg_sequence.c.seqincrement,
+                        "minvalue",
+                        pg_catalog.pg_sequence.c.seqmin,
+                        "maxvalue",
+                        pg_catalog.pg_sequence.c.seqmax,
+                        "cache",
+                        pg_catalog.pg_sequence.c.seqcache,
+                        "cycle",
+                        pg_catalog.pg_sequence.c.seqcycle,
+                    )
+                )
+                .select_from(pg_catalog.pg_sequence)
+                .where(
+                    # attidentity != '' is required or it will reflect also
+                    # serial columns as identity.
+                    pg_catalog.pg_attribute.c.attidentity != "",
+                    pg_catalog.pg_sequence.c.seqrelid
+                    == sql.cast(
+                        sql.cast(
+                            pg_catalog.pg_get_serial_sequence(
+                                sql.cast(
+                                    sql.cast(
+                                        pg_catalog.pg_attribute.c.attrelid,
+                                        REGCLASS,
+                                    ),
+                                    TEXT,
+                                ),
+                                pg_catalog.pg_attribute.c.attname,
+                            ),
+                            REGCLASS,
+                        ),
+                        OID,
+                    ),
+                )
+                .correlate(pg_catalog.pg_attribute)
+                .scalar_subquery()
+                .label("identity_options")
+            )
+        else:
+            identity = sql.null().label("identity_options")
+
+        # join lateral performs the same as scalar_subquery here
+        default = (
+            select(
+                pg_catalog.pg_get_expr(
+                    pg_catalog.pg_attrdef.c.adbin,
+                    pg_catalog.pg_attrdef.c.adrelid,
+                )
+            )
+            .select_from(pg_catalog.pg_attrdef)
+            .where(
+                pg_catalog.pg_attrdef.c.adrelid
+                == pg_catalog.pg_attribute.c.attrelid,
+                pg_catalog.pg_attrdef.c.adnum
+                == pg_catalog.pg_attribute.c.attnum,
+                pg_catalog.pg_attribute.c.atthasdef,
+            )
+            .correlate(pg_catalog.pg_attribute)
+            .scalar_subquery()
+            .label("default")
+        )
+        relkinds = self._kind_to_relkinds(kind)
+        query = (
+            select(
+                pg_catalog.pg_attribute.c.attname.label("name"),
+                pg_catalog.format_type(
+                    pg_catalog.pg_attribute.c.atttypid,
+                    pg_catalog.pg_attribute.c.atttypmod,
+                ).label("format_type"),
+                default,
+                pg_catalog.pg_attribute.c.attnotnull.label("not_null"),
+                pg_catalog.pg_class.c.relname.label("table_name"),
+                pg_catalog.pg_description.c.description.label("comment"),
+                generated,
+                identity,
+            )
+            .select_from(pg_catalog.pg_class)
+            # NOTE: postgresql support table with no user column, meaning
+            # there is no row with pg_attribute.attnum > 0. use a left outer
+            # join to avoid filtering these tables.
+            .outerjoin(
+                pg_catalog.pg_attribute,
+                sql.and_(
+                    pg_catalog.pg_class.c.oid
+                    == pg_catalog.pg_attribute.c.attrelid,
+                    pg_catalog.pg_attribute.c.attnum > 0,
+                    ~pg_catalog.pg_attribute.c.attisdropped,
+                ),
+            )
+            .outerjoin(
+                pg_catalog.pg_description,
+                sql.and_(
+                    pg_catalog.pg_description.c.objoid
+                    == pg_catalog.pg_attribute.c.attrelid,
+                    pg_catalog.pg_description.c.objsubid
+                    == pg_catalog.pg_attribute.c.attnum,
+                ),
+            )
+            .where(self._pg_class_relkind_condition(relkinds))
+            .order_by(
+                pg_catalog.pg_class.c.relname, pg_catalog.pg_attribute.c.attnum
+            )
+        )
+        query = self._pg_class_filter_scope_schema(query, schema, scope=scope)
+        if has_filter_names:
+            query = query.where(
+                pg_catalog.pg_class.c.relname.in_(bindparam("filter_names"))
+            )
+        return query
+
+    def get_multi_columns(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = self._columns_query(schema, has_filter_names, scope, kind)
+        rows = connection.execute(query, params).mappings()
+
+        # dictionary with (name, ) if default search path or (schema, name)
+        # as keys
+        domains = {
+            ((d["schema"], d["name"]) if not d["visible"] else (d["name"],)): d
+            for d in self._load_domains(
+                connection, schema="*", info_cache=kw.get("info_cache")
+            )
+        }
+
+        # dictionary with (name, ) if default search path or (schema, name)
+        # as keys
+        enums = dict(
+            (
+                ((rec["name"],), rec)
+                if rec["visible"]
+                else ((rec["schema"], rec["name"]), rec)
+            )
+            for rec in self._load_enums(
+                connection, schema="*", info_cache=kw.get("info_cache")
+            )
+        )
+
+        columns = self._get_columns_info(rows, domains, enums, schema)
+
+        return columns.items()
+
+    _format_type_args_pattern = re.compile(r"\((.*)\)")
+    _format_type_args_delim = re.compile(r"\s*,\s*")
+    _format_array_spec_pattern = re.compile(r"((?:\[\])*)$")
+
+    def _reflect_type(
+        self,
+        format_type: Optional[str],
+        domains: dict[str, ReflectedDomain],
+        enums: dict[str, ReflectedEnum],
+        type_description: str,
+    ) -> sqltypes.TypeEngine[Any]:
+        """
+        Attempts to reconstruct a column type defined in ischema_names based
+        on the information available in the format_type.
+
+        If the `format_type` cannot be associated with a known `ischema_names`,
+        it is treated as a reference to a known PostgreSQL named `ENUM` or
+        `DOMAIN` type.
+        """
+        type_description = type_description or "unknown type"
+        if format_type is None:
+            util.warn(
+                "PostgreSQL format_type() returned NULL for %s"
+                % type_description
+            )
+            return sqltypes.NULLTYPE
+
+        attype_args_match = self._format_type_args_pattern.search(format_type)
+        if attype_args_match and attype_args_match.group(1):
+            attype_args = self._format_type_args_delim.split(
+                attype_args_match.group(1)
+            )
+        else:
+            attype_args = ()
+
+        match_array_dim = self._format_array_spec_pattern.search(format_type)
+        # Each "[]" in array specs corresponds to an array dimension
+        array_dim = len(match_array_dim.group(1) or "") // 2
+
+        # Remove all parameters and array specs from format_type to obtain an
+        # ischema_name candidate
+        attype = self._format_type_args_pattern.sub("", format_type)
+        attype = self._format_array_spec_pattern.sub("", attype)
+
+        schema_type = self.ischema_names.get(attype.lower(), None)
+        args, kwargs = (), {}
+
+        if attype == "numeric":
+            if len(attype_args) == 2:
+                precision, scale = map(int, attype_args)
+                args = (precision, scale)
+
+        elif attype == "double precision":
+            args = (53,)
+
+        elif attype == "integer":
+            args = ()
+
+        elif attype in ("timestamp with time zone", "time with time zone"):
+            kwargs["timezone"] = True
+            if len(attype_args) == 1:
+                kwargs["precision"] = int(attype_args[0])
+
+        elif attype in (
+            "timestamp without time zone",
+            "time without time zone",
+            "time",
+        ):
+            kwargs["timezone"] = False
+            if len(attype_args) == 1:
+                kwargs["precision"] = int(attype_args[0])
+
+        elif attype == "bit varying":
+            kwargs["varying"] = True
+            if len(attype_args) == 1:
+                charlen = int(attype_args[0])
+                args = (charlen,)
+
+        elif attype.startswith("interval"):
+            schema_type = INTERVAL
+
+            field_match = re.match(r"interval (.+)", attype)
+            if field_match:
+                kwargs["fields"] = field_match.group(1)
+
+            if len(attype_args) == 1:
+                kwargs["precision"] = int(attype_args[0])
+
+        else:
+            enum_or_domain_key = tuple(util.quoted_token_parser(attype))
+
+            if enum_or_domain_key in enums:
+                schema_type = ENUM
+                enum = enums[enum_or_domain_key]
+
+                args = tuple(enum["labels"])
+                kwargs["name"] = enum["name"]
+
+                if not enum["visible"]:
+                    kwargs["schema"] = enum["schema"]
+                args = tuple(enum["labels"])
+            elif enum_or_domain_key in domains:
+                schema_type = DOMAIN
+                domain = domains[enum_or_domain_key]
+
+                data_type = self._reflect_type(
+                    domain["type"],
+                    domains,
+                    enums,
+                    type_description="DOMAIN '%s'" % domain["name"],
+                )
+                args = (domain["name"], data_type)
+
+                kwargs["collation"] = domain["collation"]
+                kwargs["default"] = domain["default"]
+                kwargs["not_null"] = not domain["nullable"]
+                kwargs["create_type"] = False
+
+                if domain["constraints"]:
+                    # We only support a single constraint
+                    check_constraint = domain["constraints"][0]
+
+                    kwargs["constraint_name"] = check_constraint["name"]
+                    kwargs["check"] = check_constraint["check"]
+
+                if not domain["visible"]:
+                    kwargs["schema"] = domain["schema"]
+
+            else:
+                try:
+                    charlen = int(attype_args[0])
+                    args = (charlen, *attype_args[1:])
+                except (ValueError, IndexError):
+                    args = attype_args
+
+        if not schema_type:
+            util.warn(
+                "Did not recognize type '%s' of %s"
+                % (attype, type_description)
+            )
+            return sqltypes.NULLTYPE
+
+        data_type = schema_type(*args, **kwargs)
+        if array_dim >= 1:
+            # postgres does not preserve dimensionality or size of array types.
+            data_type = _array.ARRAY(data_type)
+
+        return data_type
+
+    def _get_columns_info(self, rows, domains, enums, schema):
+        columns = defaultdict(list)
+        for row_dict in rows:
+            # ensure that each table has an entry, even if it has no columns
+            if row_dict["name"] is None:
+                columns[(schema, row_dict["table_name"])] = (
+                    ReflectionDefaults.columns()
+                )
+                continue
+            table_cols = columns[(schema, row_dict["table_name"])]
+
+            coltype = self._reflect_type(
+                row_dict["format_type"],
+                domains,
+                enums,
+                type_description="column '%s'" % row_dict["name"],
+            )
+
+            default = row_dict["default"]
+            name = row_dict["name"]
+            generated = row_dict["generated"]
+            nullable = not row_dict["not_null"]
+
+            if isinstance(coltype, DOMAIN):
+                if not default:
+                    # domain can override the default value but
+                    # cant set it to None
+                    if coltype.default is not None:
+                        default = coltype.default
+
+                nullable = nullable and not coltype.not_null
+
+            identity = row_dict["identity_options"]
+
+            # If a zero byte or blank string depending on driver (is also
+            # absent for older PG versions), then not a generated column.
+            # Otherwise, s = stored. (Other values might be added in the
+            # future.)
+            if generated not in (None, "", b"\x00"):
+                computed = dict(
+                    sqltext=default, persisted=generated in ("s", b"s")
+                )
+                default = None
+            else:
+                computed = None
+
+            # adjust the default value
+            autoincrement = False
+            if default is not None:
+                match = re.search(r"""(nextval\(')([^']+)('.*$)""", default)
+                if match is not None:
+                    if issubclass(coltype._type_affinity, sqltypes.Integer):
+                        autoincrement = True
+                    # the default is related to a Sequence
+                    if "." not in match.group(2) and schema is not None:
+                        # unconditionally quote the schema name.  this could
+                        # later be enhanced to obey quoting rules /
+                        # "quote schema"
+                        default = (
+                            match.group(1)
+                            + ('"%s"' % schema)
+                            + "."
+                            + match.group(2)
+                            + match.group(3)
+                        )
+
+            column_info = {
+                "name": name,
+                "type": coltype,
+                "nullable": nullable,
+                "default": default,
+                "autoincrement": autoincrement or identity is not None,
+                "comment": row_dict["comment"],
+            }
+            if computed is not None:
+                column_info["computed"] = computed
+            if identity is not None:
+                column_info["identity"] = identity
+
+            table_cols.append(column_info)
+
+        return columns
+
+    @lru_cache()
+    def _table_oids_query(self, schema, has_filter_names, scope, kind):
+        relkinds = self._kind_to_relkinds(kind)
+        oid_q = select(
+            pg_catalog.pg_class.c.oid, pg_catalog.pg_class.c.relname
+        ).where(self._pg_class_relkind_condition(relkinds))
+        oid_q = self._pg_class_filter_scope_schema(oid_q, schema, scope=scope)
+
+        if has_filter_names:
+            oid_q = oid_q.where(
+                pg_catalog.pg_class.c.relname.in_(bindparam("filter_names"))
+            )
+        return oid_q
+
+    @reflection.flexi_cache(
+        ("schema", InternalTraversal.dp_string),
+        ("filter_names", InternalTraversal.dp_string_list),
+        ("kind", InternalTraversal.dp_plain_obj),
+        ("scope", InternalTraversal.dp_plain_obj),
+    )
+    def _get_table_oids(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        oid_q = self._table_oids_query(schema, has_filter_names, scope, kind)
+        result = connection.execute(oid_q, params)
+        return result.all()
+
+    @lru_cache()
+    def _constraint_query(self, is_unique):
+        con_sq = (
+            select(
+                pg_catalog.pg_constraint.c.conrelid,
+                pg_catalog.pg_constraint.c.conname,
+                pg_catalog.pg_constraint.c.conindid,
+                sql.func.unnest(pg_catalog.pg_constraint.c.conkey).label(
+                    "attnum"
+                ),
+                sql.func.generate_subscripts(
+                    pg_catalog.pg_constraint.c.conkey, 1
+                ).label("ord"),
+                pg_catalog.pg_description.c.description,
+            )
+            .outerjoin(
+                pg_catalog.pg_description,
+                pg_catalog.pg_description.c.objoid
+                == pg_catalog.pg_constraint.c.oid,
+            )
+            .where(
+                pg_catalog.pg_constraint.c.contype == bindparam("contype"),
+                pg_catalog.pg_constraint.c.conrelid.in_(bindparam("oids")),
+            )
+            .subquery("con")
+        )
+
+        attr_sq = (
+            select(
+                con_sq.c.conrelid,
+                con_sq.c.conname,
+                con_sq.c.conindid,
+                con_sq.c.description,
+                con_sq.c.ord,
+                pg_catalog.pg_attribute.c.attname,
+            )
+            .select_from(pg_catalog.pg_attribute)
+            .join(
+                con_sq,
+                sql.and_(
+                    pg_catalog.pg_attribute.c.attnum == con_sq.c.attnum,
+                    pg_catalog.pg_attribute.c.attrelid == con_sq.c.conrelid,
+                ),
+            )
+            .where(
+                # NOTE: restate the condition here, since pg15 otherwise
+                # seems to get confused on pscopg2 sometimes, doing
+                # a sequential scan of pg_attribute.
+                # The condition in the con_sq subquery is not actually needed
+                # in pg15, but it may be needed in older versions. Keeping it
+                # does not seems to have any inpact in any case.
+                con_sq.c.conrelid.in_(bindparam("oids"))
+            )
+            .subquery("attr")
+        )
+
+        constraint_query = (
+            select(
+                attr_sq.c.conrelid,
+                sql.func.array_agg(
+                    # NOTE: cast since some postgresql derivatives may
+                    # not support array_agg on the name type
+                    aggregate_order_by(
+                        attr_sq.c.attname.cast(TEXT), attr_sq.c.ord
+                    )
+                ).label("cols"),
+                attr_sq.c.conname,
+                sql.func.min(attr_sq.c.description).label("description"),
+            )
+            .group_by(attr_sq.c.conrelid, attr_sq.c.conname)
+            .order_by(attr_sq.c.conrelid, attr_sq.c.conname)
+        )
+
+        if is_unique:
+            if self.server_version_info >= (15,):
+                constraint_query = constraint_query.join(
+                    pg_catalog.pg_index,
+                    attr_sq.c.conindid == pg_catalog.pg_index.c.indexrelid,
+                ).add_columns(
+                    sql.func.bool_and(
+                        pg_catalog.pg_index.c.indnullsnotdistinct
+                    ).label("indnullsnotdistinct")
+                )
+            else:
+                constraint_query = constraint_query.add_columns(
+                    sql.false().label("indnullsnotdistinct")
+                )
+        else:
+            constraint_query = constraint_query.add_columns(
+                sql.null().label("extra")
+            )
+        return constraint_query
+
+    def _reflect_constraint(
+        self, connection, contype, schema, filter_names, scope, kind, **kw
+    ):
+        # used to reflect primary and unique constraint
+        table_oids = self._get_table_oids(
+            connection, schema, filter_names, scope, kind, **kw
+        )
+        batches = list(table_oids)
+        is_unique = contype == "u"
+
+        while batches:
+            batch = batches[0:3000]
+            batches[0:3000] = []
+
+            result = connection.execute(
+                self._constraint_query(is_unique),
+                {"oids": [r[0] for r in batch], "contype": contype},
+            )
+
+            result_by_oid = defaultdict(list)
+            for oid, cols, constraint_name, comment, extra in result:
+                result_by_oid[oid].append(
+                    (cols, constraint_name, comment, extra)
+                )
+
+            for oid, tablename in batch:
+                for_oid = result_by_oid.get(oid, ())
+                if for_oid:
+                    for cols, constraint, comment, extra in for_oid:
+                        if is_unique:
+                            yield tablename, cols, constraint, comment, {
+                                "nullsnotdistinct": extra
+                            }
+                        else:
+                            yield tablename, cols, constraint, comment, None
+                else:
+                    yield tablename, None, None, None, None
+
+    @reflection.cache
+    def get_pk_constraint(self, connection, table_name, schema=None, **kw):
+        data = self.get_multi_pk_constraint(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    def get_multi_pk_constraint(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        result = self._reflect_constraint(
+            connection, "p", schema, filter_names, scope, kind, **kw
+        )
+
+        # only a single pk can be present for each table. Return an entry
+        # even if a table has no primary key
+        default = ReflectionDefaults.pk_constraint
+        return (
+            (
+                (schema, table_name),
+                (
+                    {
+                        "constrained_columns": [] if cols is None else cols,
+                        "name": pk_name,
+                        "comment": comment,
+                    }
+                    if pk_name is not None
+                    else default()
+                ),
+            )
+            for table_name, cols, pk_name, comment, _ in result
+        )
+
+    @reflection.cache
+    def get_foreign_keys(
+        self,
+        connection,
+        table_name,
+        schema=None,
+        postgresql_ignore_search_path=False,
+        **kw,
+    ):
+        data = self.get_multi_foreign_keys(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            postgresql_ignore_search_path=postgresql_ignore_search_path,
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _foreing_key_query(self, schema, has_filter_names, scope, kind):
+        pg_class_ref = pg_catalog.pg_class.alias("cls_ref")
+        pg_namespace_ref = pg_catalog.pg_namespace.alias("nsp_ref")
+        relkinds = self._kind_to_relkinds(kind)
+        query = (
+            select(
+                pg_catalog.pg_class.c.relname,
+                pg_catalog.pg_constraint.c.conname,
+                # NOTE: avoid calling pg_get_constraintdef when not needed
+                # to speed up the query
+                sql.case(
+                    (
+                        pg_catalog.pg_constraint.c.oid.is_not(None),
+                        pg_catalog.pg_get_constraintdef(
+                            pg_catalog.pg_constraint.c.oid, True
+                        ),
+                    ),
+                    else_=None,
+                ),
+                pg_namespace_ref.c.nspname,
+                pg_catalog.pg_description.c.description,
+            )
+            .select_from(pg_catalog.pg_class)
+            .outerjoin(
+                pg_catalog.pg_constraint,
+                sql.and_(
+                    pg_catalog.pg_class.c.oid
+                    == pg_catalog.pg_constraint.c.conrelid,
+                    pg_catalog.pg_constraint.c.contype == "f",
+                ),
+            )
+            .outerjoin(
+                pg_class_ref,
+                pg_class_ref.c.oid == pg_catalog.pg_constraint.c.confrelid,
+            )
+            .outerjoin(
+                pg_namespace_ref,
+                pg_class_ref.c.relnamespace == pg_namespace_ref.c.oid,
+            )
+            .outerjoin(
+                pg_catalog.pg_description,
+                pg_catalog.pg_description.c.objoid
+                == pg_catalog.pg_constraint.c.oid,
+            )
+            .order_by(
+                pg_catalog.pg_class.c.relname,
+                pg_catalog.pg_constraint.c.conname,
+            )
+            .where(self._pg_class_relkind_condition(relkinds))
+        )
+        query = self._pg_class_filter_scope_schema(query, schema, scope)
+        if has_filter_names:
+            query = query.where(
+                pg_catalog.pg_class.c.relname.in_(bindparam("filter_names"))
+            )
+        return query
+
+    @util.memoized_property
+    def _fk_regex_pattern(self):
+        # optionally quoted token
+        qtoken = '(?:"[^"]+"|[A-Za-z0-9_]+?)'
+
+        # https://www.postgresql.org/docs/current/static/sql-createtable.html
+        return re.compile(
+            r"FOREIGN KEY \((.*?)\) "
+            rf"REFERENCES (?:({qtoken})\.)?({qtoken})\(((?:{qtoken}(?: *, *)?)+)\)"  # noqa: E501
+            r"[\s]?(MATCH (FULL|PARTIAL|SIMPLE)+)?"
+            r"[\s]?(ON UPDATE "
+            r"(CASCADE|RESTRICT|NO ACTION|SET NULL|SET DEFAULT)+)?"
+            r"[\s]?(ON DELETE "
+            r"(CASCADE|RESTRICT|NO ACTION|SET NULL|SET DEFAULT)+)?"
+            r"[\s]?(DEFERRABLE|NOT DEFERRABLE)?"
+            r"[\s]?(INITIALLY (DEFERRED|IMMEDIATE)+)?"
+        )
+
+    def get_multi_foreign_keys(
+        self,
+        connection,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        postgresql_ignore_search_path=False,
+        **kw,
+    ):
+        preparer = self.identifier_preparer
+
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = self._foreing_key_query(schema, has_filter_names, scope, kind)
+        result = connection.execute(query, params)
+
+        FK_REGEX = self._fk_regex_pattern
+
+        fkeys = defaultdict(list)
+        default = ReflectionDefaults.foreign_keys
+        for table_name, conname, condef, conschema, comment in result:
+            # ensure that each table has an entry, even if it has
+            # no foreign keys
+            if conname is None:
+                fkeys[(schema, table_name)] = default()
+                continue
+            table_fks = fkeys[(schema, table_name)]
+            m = re.search(FK_REGEX, condef).groups()
+
+            (
+                constrained_columns,
+                referred_schema,
+                referred_table,
+                referred_columns,
+                _,
+                match,
+                _,
+                onupdate,
+                _,
+                ondelete,
+                deferrable,
+                _,
+                initially,
+            ) = m
+
+            if deferrable is not None:
+                deferrable = True if deferrable == "DEFERRABLE" else False
+            constrained_columns = [
+                preparer._unquote_identifier(x)
+                for x in re.split(r"\s*,\s*", constrained_columns)
+            ]
+
+            if postgresql_ignore_search_path:
+                # when ignoring search path, we use the actual schema
+                # provided it isn't the "default" schema
+                if conschema != self.default_schema_name:
+                    referred_schema = conschema
+                else:
+                    referred_schema = schema
+            elif referred_schema:
+                # referred_schema is the schema that we regexp'ed from
+                # pg_get_constraintdef().  If the schema is in the search
+                # path, pg_get_constraintdef() will give us None.
+                referred_schema = preparer._unquote_identifier(referred_schema)
+            elif schema is not None and schema == conschema:
+                # If the actual schema matches the schema of the table
+                # we're reflecting, then we will use that.
+                referred_schema = schema
+
+            referred_table = preparer._unquote_identifier(referred_table)
+            referred_columns = [
+                preparer._unquote_identifier(x)
+                for x in re.split(r"\s*,\s", referred_columns)
+            ]
+            options = {
+                k: v
+                for k, v in [
+                    ("onupdate", onupdate),
+                    ("ondelete", ondelete),
+                    ("initially", initially),
+                    ("deferrable", deferrable),
+                    ("match", match),
+                ]
+                if v is not None and v != "NO ACTION"
+            }
+            fkey_d = {
+                "name": conname,
+                "constrained_columns": constrained_columns,
+                "referred_schema": referred_schema,
+                "referred_table": referred_table,
+                "referred_columns": referred_columns,
+                "options": options,
+                "comment": comment,
+            }
+            table_fks.append(fkey_d)
+        return fkeys.items()
+
+    @reflection.cache
+    def get_indexes(self, connection, table_name, schema=None, **kw):
+        data = self.get_multi_indexes(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @util.memoized_property
+    def _index_query(self):
+        pg_class_index = pg_catalog.pg_class.alias("cls_idx")
+        # NOTE: repeating oids clause improve query performance
+
+        # subquery to get the columns
+        idx_sq = (
+            select(
+                pg_catalog.pg_index.c.indexrelid,
+                pg_catalog.pg_index.c.indrelid,
+                sql.func.unnest(pg_catalog.pg_index.c.indkey).label("attnum"),
+                sql.func.generate_subscripts(
+                    pg_catalog.pg_index.c.indkey, 1
+                ).label("ord"),
+            )
+            .where(
+                ~pg_catalog.pg_index.c.indisprimary,
+                pg_catalog.pg_index.c.indrelid.in_(bindparam("oids")),
+            )
+            .subquery("idx")
+        )
+
+        attr_sq = (
+            select(
+                idx_sq.c.indexrelid,
+                idx_sq.c.indrelid,
+                idx_sq.c.ord,
+                # NOTE: always using pg_get_indexdef is too slow so just
+                # invoke when the element is an expression
+                sql.case(
+                    (
+                        idx_sq.c.attnum == 0,
+                        pg_catalog.pg_get_indexdef(
+                            idx_sq.c.indexrelid, idx_sq.c.ord + 1, True
+                        ),
+                    ),
+                    # NOTE: need to cast this since attname is of type "name"
+                    # that's limited to 63 bytes, while pg_get_indexdef
+                    # returns "text" so its output may get cut
+                    else_=pg_catalog.pg_attribute.c.attname.cast(TEXT),
+                ).label("element"),
+                (idx_sq.c.attnum == 0).label("is_expr"),
+            )
+            .select_from(idx_sq)
+            .outerjoin(
+                # do not remove rows where idx_sq.c.attnum is 0
+                pg_catalog.pg_attribute,
+                sql.and_(
+                    pg_catalog.pg_attribute.c.attnum == idx_sq.c.attnum,
+                    pg_catalog.pg_attribute.c.attrelid == idx_sq.c.indrelid,
+                ),
+            )
+            .where(idx_sq.c.indrelid.in_(bindparam("oids")))
+            .subquery("idx_attr")
+        )
+
+        cols_sq = (
+            select(
+                attr_sq.c.indexrelid,
+                sql.func.min(attr_sq.c.indrelid),
+                sql.func.array_agg(
+                    aggregate_order_by(attr_sq.c.element, attr_sq.c.ord)
+                ).label("elements"),
+                sql.func.array_agg(
+                    aggregate_order_by(attr_sq.c.is_expr, attr_sq.c.ord)
+                ).label("elements_is_expr"),
+            )
+            .group_by(attr_sq.c.indexrelid)
+            .subquery("idx_cols")
+        )
+
+        if self.server_version_info >= (11, 0):
+            indnkeyatts = pg_catalog.pg_index.c.indnkeyatts
+        else:
+            indnkeyatts = sql.null().label("indnkeyatts")
+
+        if self.server_version_info >= (15,):
+            nulls_not_distinct = pg_catalog.pg_index.c.indnullsnotdistinct
+        else:
+            nulls_not_distinct = sql.false().label("indnullsnotdistinct")
+
+        return (
+            select(
+                pg_catalog.pg_index.c.indrelid,
+                pg_class_index.c.relname.label("relname_index"),
+                pg_catalog.pg_index.c.indisunique,
+                pg_catalog.pg_constraint.c.conrelid.is_not(None).label(
+                    "has_constraint"
+                ),
+                pg_catalog.pg_index.c.indoption,
+                pg_class_index.c.reloptions,
+                pg_catalog.pg_am.c.amname,
+                # NOTE: pg_get_expr is very fast so this case has almost no
+                # performance impact
+                sql.case(
+                    (
+                        pg_catalog.pg_index.c.indpred.is_not(None),
+                        pg_catalog.pg_get_expr(
+                            pg_catalog.pg_index.c.indpred,
+                            pg_catalog.pg_index.c.indrelid,
+                        ),
+                    ),
+                    else_=None,
+                ).label("filter_definition"),
+                indnkeyatts,
+                nulls_not_distinct,
+                cols_sq.c.elements,
+                cols_sq.c.elements_is_expr,
+            )
+            .select_from(pg_catalog.pg_index)
+            .where(
+                pg_catalog.pg_index.c.indrelid.in_(bindparam("oids")),
+                ~pg_catalog.pg_index.c.indisprimary,
+            )
+            .join(
+                pg_class_index,
+                pg_catalog.pg_index.c.indexrelid == pg_class_index.c.oid,
+            )
+            .join(
+                pg_catalog.pg_am,
+                pg_class_index.c.relam == pg_catalog.pg_am.c.oid,
+            )
+            .outerjoin(
+                cols_sq,
+                pg_catalog.pg_index.c.indexrelid == cols_sq.c.indexrelid,
+            )
+            .outerjoin(
+                pg_catalog.pg_constraint,
+                sql.and_(
+                    pg_catalog.pg_index.c.indrelid
+                    == pg_catalog.pg_constraint.c.conrelid,
+                    pg_catalog.pg_index.c.indexrelid
+                    == pg_catalog.pg_constraint.c.conindid,
+                    pg_catalog.pg_constraint.c.contype
+                    == sql.any_(_array.array(("p", "u", "x"))),
+                ),
+            )
+            .order_by(pg_catalog.pg_index.c.indrelid, pg_class_index.c.relname)
+        )
+
+    def get_multi_indexes(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        table_oids = self._get_table_oids(
+            connection, schema, filter_names, scope, kind, **kw
+        )
+
+        indexes = defaultdict(list)
+        default = ReflectionDefaults.indexes
+
+        batches = list(table_oids)
+
+        while batches:
+            batch = batches[0:3000]
+            batches[0:3000] = []
+
+            result = connection.execute(
+                self._index_query, {"oids": [r[0] for r in batch]}
+            ).mappings()
+
+            result_by_oid = defaultdict(list)
+            for row_dict in result:
+                result_by_oid[row_dict["indrelid"]].append(row_dict)
+
+            for oid, table_name in batch:
+                if oid not in result_by_oid:
+                    # ensure that each table has an entry, even if reflection
+                    # is skipped because not supported
+                    indexes[(schema, table_name)] = default()
+                    continue
+
+                for row in result_by_oid[oid]:
+                    index_name = row["relname_index"]
+
+                    table_indexes = indexes[(schema, table_name)]
+
+                    all_elements = row["elements"]
+                    all_elements_is_expr = row["elements_is_expr"]
+                    indnkeyatts = row["indnkeyatts"]
+                    # "The number of key columns in the index, not counting any
+                    # included columns, which are merely stored and do not
+                    # participate in the index semantics"
+                    if indnkeyatts and len(all_elements) > indnkeyatts:
+                        # this is a "covering index" which has INCLUDE columns
+                        # as well as regular index columns
+                        inc_cols = all_elements[indnkeyatts:]
+                        idx_elements = all_elements[:indnkeyatts]
+                        idx_elements_is_expr = all_elements_is_expr[
+                            :indnkeyatts
+                        ]
+                        # postgresql does not support expression on included
+                        # columns as of v14: "ERROR: expressions are not
+                        # supported in included columns".
+                        assert all(
+                            not is_expr
+                            for is_expr in all_elements_is_expr[indnkeyatts:]
+                        )
+                    else:
+                        idx_elements = all_elements
+                        idx_elements_is_expr = all_elements_is_expr
+                        inc_cols = []
+
+                    index = {"name": index_name, "unique": row["indisunique"]}
+                    if any(idx_elements_is_expr):
+                        index["column_names"] = [
+                            None if is_expr else expr
+                            for expr, is_expr in zip(
+                                idx_elements, idx_elements_is_expr
+                            )
+                        ]
+                        index["expressions"] = idx_elements
+                    else:
+                        index["column_names"] = idx_elements
+
+                    sorting = {}
+                    for col_index, col_flags in enumerate(row["indoption"]):
+                        col_sorting = ()
+                        # try to set flags only if they differ from PG
+                        # defaults...
+                        if col_flags & 0x01:
+                            col_sorting += ("desc",)
+                            if not (col_flags & 0x02):
+                                col_sorting += ("nulls_last",)
+                        else:
+                            if col_flags & 0x02:
+                                col_sorting += ("nulls_first",)
+                        if col_sorting:
+                            sorting[idx_elements[col_index]] = col_sorting
+                    if sorting:
+                        index["column_sorting"] = sorting
+                    if row["has_constraint"]:
+                        index["duplicates_constraint"] = index_name
+
+                    dialect_options = {}
+                    if row["reloptions"]:
+                        dialect_options["postgresql_with"] = dict(
+                            [
+                                option.split("=", 1)
+                                for option in row["reloptions"]
+                            ]
+                        )
+                    # it *might* be nice to include that this is 'btree' in the
+                    # reflection info.  But we don't want an Index object
+                    # to have a ``postgresql_using`` in it that is just the
+                    # default, so for the moment leaving this out.
+                    amname = row["amname"]
+                    if amname != "btree":
+                        dialect_options["postgresql_using"] = row["amname"]
+                    if row["filter_definition"]:
+                        dialect_options["postgresql_where"] = row[
+                            "filter_definition"
+                        ]
+                    if self.server_version_info >= (11,):
+                        # NOTE: this is legacy, this is part of
+                        # dialect_options now as of #7382
+                        index["include_columns"] = inc_cols
+                        dialect_options["postgresql_include"] = inc_cols
+                    if row["indnullsnotdistinct"]:
+                        # the default is False, so ignore it.
+                        dialect_options["postgresql_nulls_not_distinct"] = row[
+                            "indnullsnotdistinct"
+                        ]
+
+                    if dialect_options:
+                        index["dialect_options"] = dialect_options
+
+                    table_indexes.append(index)
+        return indexes.items()
+
+    @reflection.cache
+    def get_unique_constraints(
+        self, connection, table_name, schema=None, **kw
+    ):
+        data = self.get_multi_unique_constraints(
+            connection,
+            schema=schema,
+            filter_names=[table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    def get_multi_unique_constraints(
+        self,
+        connection,
+        schema,
+        filter_names,
+        scope,
+        kind,
+        **kw,
+    ):
+        result = self._reflect_constraint(
+            connection, "u", schema, filter_names, scope, kind, **kw
+        )
+
+        # each table can have multiple unique constraints
+        uniques = defaultdict(list)
+        default = ReflectionDefaults.unique_constraints
+        for table_name, cols, con_name, comment, options in result:
+            # ensure a list is created for each table. leave it empty if
+            # the table has no unique cosntraint
+            if con_name is None:
+                uniques[(schema, table_name)] = default()
+                continue
+
+            uc_dict = {
+                "column_names": cols,
+                "name": con_name,
+                "comment": comment,
+            }
+            if options:
+                if options["nullsnotdistinct"]:
+                    uc_dict["dialect_options"] = {
+                        "postgresql_nulls_not_distinct": options[
+                            "nullsnotdistinct"
+                        ]
+                    }
+
+            uniques[(schema, table_name)].append(uc_dict)
+        return uniques.items()
+
+    @reflection.cache
+    def get_table_comment(self, connection, table_name, schema=None, **kw):
+        data = self.get_multi_table_comment(
+            connection,
+            schema,
+            [table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _comment_query(self, schema, has_filter_names, scope, kind):
+        relkinds = self._kind_to_relkinds(kind)
+        query = (
+            select(
+                pg_catalog.pg_class.c.relname,
+                pg_catalog.pg_description.c.description,
+            )
+            .select_from(pg_catalog.pg_class)
+            .outerjoin(
+                pg_catalog.pg_description,
+                sql.and_(
+                    pg_catalog.pg_class.c.oid
+                    == pg_catalog.pg_description.c.objoid,
+                    pg_catalog.pg_description.c.objsubid == 0,
+                    pg_catalog.pg_description.c.classoid
+                    == sql.func.cast("pg_catalog.pg_class", REGCLASS),
+                ),
+            )
+            .where(self._pg_class_relkind_condition(relkinds))
+        )
+        query = self._pg_class_filter_scope_schema(query, schema, scope)
+        if has_filter_names:
+            query = query.where(
+                pg_catalog.pg_class.c.relname.in_(bindparam("filter_names"))
+            )
+        return query
+
+    def get_multi_table_comment(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = self._comment_query(schema, has_filter_names, scope, kind)
+        result = connection.execute(query, params)
+
+        default = ReflectionDefaults.table_comment
+        return (
+            (
+                (schema, table),
+                {"text": comment} if comment is not None else default(),
+            )
+            for table, comment in result
+        )
+
+    @reflection.cache
+    def get_check_constraints(self, connection, table_name, schema=None, **kw):
+        data = self.get_multi_check_constraints(
+            connection,
+            schema,
+            [table_name],
+            scope=ObjectScope.ANY,
+            kind=ObjectKind.ANY,
+            **kw,
+        )
+        return self._value_or_raise(data, table_name, schema)
+
+    @lru_cache()
+    def _check_constraint_query(self, schema, has_filter_names, scope, kind):
+        relkinds = self._kind_to_relkinds(kind)
+        query = (
+            select(
+                pg_catalog.pg_class.c.relname,
+                pg_catalog.pg_constraint.c.conname,
+                # NOTE: avoid calling pg_get_constraintdef when not needed
+                # to speed up the query
+                sql.case(
+                    (
+                        pg_catalog.pg_constraint.c.oid.is_not(None),
+                        pg_catalog.pg_get_constraintdef(
+                            pg_catalog.pg_constraint.c.oid, True
+                        ),
+                    ),
+                    else_=None,
+                ),
+                pg_catalog.pg_description.c.description,
+            )
+            .select_from(pg_catalog.pg_class)
+            .outerjoin(
+                pg_catalog.pg_constraint,
+                sql.and_(
+                    pg_catalog.pg_class.c.oid
+                    == pg_catalog.pg_constraint.c.conrelid,
+                    pg_catalog.pg_constraint.c.contype == "c",
+                ),
+            )
+            .outerjoin(
+                pg_catalog.pg_description,
+                pg_catalog.pg_description.c.objoid
+                == pg_catalog.pg_constraint.c.oid,
+            )
+            .order_by(
+                pg_catalog.pg_class.c.relname,
+                pg_catalog.pg_constraint.c.conname,
+            )
+            .where(self._pg_class_relkind_condition(relkinds))
+        )
+        query = self._pg_class_filter_scope_schema(query, schema, scope)
+        if has_filter_names:
+            query = query.where(
+                pg_catalog.pg_class.c.relname.in_(bindparam("filter_names"))
+            )
+        return query
+
+    def get_multi_check_constraints(
+        self, connection, schema, filter_names, scope, kind, **kw
+    ):
+        has_filter_names, params = self._prepare_filter_names(filter_names)
+        query = self._check_constraint_query(
+            schema, has_filter_names, scope, kind
+        )
+        result = connection.execute(query, params)
+
+        check_constraints = defaultdict(list)
+        default = ReflectionDefaults.check_constraints
+        for table_name, check_name, src, comment in result:
+            # only two cases for check_name and src: both null or both defined
+            if check_name is None and src is None:
+                check_constraints[(schema, table_name)] = default()
+                continue
+            # samples:
+            # "CHECK (((a > 1) AND (a < 5)))"
+            # "CHECK (((a = 1) OR ((a > 2) AND (a < 5))))"
+            # "CHECK (((a > 1) AND (a < 5))) NOT VALID"
+            # "CHECK (some_boolean_function(a))"
+            # "CHECK (((a\n < 1)\n OR\n (a\n >= 5))\n)"
+            # "CHECK (a NOT NULL) NO INHERIT"
+            # "CHECK (a NOT NULL) NO INHERIT NOT VALID"
+
+            m = re.match(
+                r"^CHECK *\((.+)\)( NO INHERIT)?( NOT VALID)?$",
+                src,
+                flags=re.DOTALL,
+            )
+            if not m:
+                util.warn("Could not parse CHECK constraint text: %r" % src)
+                sqltext = ""
+            else:
+                sqltext = re.compile(
+                    r"^[\s\n]*\((.+)\)[\s\n]*$", flags=re.DOTALL
+                ).sub(r"\1", m.group(1))
+            entry = {
+                "name": check_name,
+                "sqltext": sqltext,
+                "comment": comment,
+            }
+            if m:
+                do = {}
+                if " NOT VALID" in m.groups():
+                    do["not_valid"] = True
+                if " NO INHERIT" in m.groups():
+                    do["no_inherit"] = True
+                if do:
+                    entry["dialect_options"] = do
+
+            check_constraints[(schema, table_name)].append(entry)
+        return check_constraints.items()
+
+    def _pg_type_filter_schema(self, query, schema):
+        if schema is None:
+            query = query.where(
+                pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid),
+                # ignore pg_catalog schema
+                pg_catalog.pg_namespace.c.nspname != "pg_catalog",
+            )
+        elif schema != "*":
+            query = query.where(pg_catalog.pg_namespace.c.nspname == schema)
+        return query
+
+    @lru_cache()
+    def _enum_query(self, schema):
+        lbl_agg_sq = (
+            select(
+                pg_catalog.pg_enum.c.enumtypid,
+                sql.func.array_agg(
+                    aggregate_order_by(
+                        # NOTE: cast since some postgresql derivatives may
+                        # not support array_agg on the name type
+                        pg_catalog.pg_enum.c.enumlabel.cast(TEXT),
+                        pg_catalog.pg_enum.c.enumsortorder,
+                    )
+                ).label("labels"),
+            )
+            .group_by(pg_catalog.pg_enum.c.enumtypid)
+            .subquery("lbl_agg")
+        )
+
+        query = (
+            select(
+                pg_catalog.pg_type.c.typname.label("name"),
+                pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid).label(
+                    "visible"
+                ),
+                pg_catalog.pg_namespace.c.nspname.label("schema"),
+                lbl_agg_sq.c.labels.label("labels"),
+            )
+            .join(
+                pg_catalog.pg_namespace,
+                pg_catalog.pg_namespace.c.oid
+                == pg_catalog.pg_type.c.typnamespace,
+            )
+            .outerjoin(
+                lbl_agg_sq, pg_catalog.pg_type.c.oid == lbl_agg_sq.c.enumtypid
+            )
+            .where(pg_catalog.pg_type.c.typtype == "e")
+            .order_by(
+                pg_catalog.pg_namespace.c.nspname, pg_catalog.pg_type.c.typname
+            )
+        )
+
+        return self._pg_type_filter_schema(query, schema)
+
+    @reflection.cache
+    def _load_enums(self, connection, schema=None, **kw):
+        if not self.supports_native_enum:
+            return []
+
+        result = connection.execute(self._enum_query(schema))
+
+        enums = []
+        for name, visible, schema, labels in result:
+            enums.append(
+                {
+                    "name": name,
+                    "schema": schema,
+                    "visible": visible,
+                    "labels": [] if labels is None else labels,
+                }
+            )
+        return enums
+
+    @lru_cache()
+    def _domain_query(self, schema):
+        con_sq = (
+            select(
+                pg_catalog.pg_constraint.c.contypid,
+                sql.func.array_agg(
+                    pg_catalog.pg_get_constraintdef(
+                        pg_catalog.pg_constraint.c.oid, True
+                    )
+                ).label("condefs"),
+                sql.func.array_agg(
+                    # NOTE: cast since some postgresql derivatives may
+                    # not support array_agg on the name type
+                    pg_catalog.pg_constraint.c.conname.cast(TEXT)
+                ).label("connames"),
+            )
+            # The domain this constraint is on; zero if not a domain constraint
+            .where(pg_catalog.pg_constraint.c.contypid != 0)
+            .group_by(pg_catalog.pg_constraint.c.contypid)
+            .subquery("domain_constraints")
+        )
+
+        query = (
+            select(
+                pg_catalog.pg_type.c.typname.label("name"),
+                pg_catalog.format_type(
+                    pg_catalog.pg_type.c.typbasetype,
+                    pg_catalog.pg_type.c.typtypmod,
+                ).label("attype"),
+                (~pg_catalog.pg_type.c.typnotnull).label("nullable"),
+                pg_catalog.pg_type.c.typdefault.label("default"),
+                pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid).label(
+                    "visible"
+                ),
+                pg_catalog.pg_namespace.c.nspname.label("schema"),
+                con_sq.c.condefs,
+                con_sq.c.connames,
+                pg_catalog.pg_collation.c.collname,
+            )
+            .join(
+                pg_catalog.pg_namespace,
+                pg_catalog.pg_namespace.c.oid
+                == pg_catalog.pg_type.c.typnamespace,
+            )
+            .outerjoin(
+                pg_catalog.pg_collation,
+                pg_catalog.pg_type.c.typcollation
+                == pg_catalog.pg_collation.c.oid,
+            )
+            .outerjoin(
+                con_sq,
+                pg_catalog.pg_type.c.oid == con_sq.c.contypid,
+            )
+            .where(pg_catalog.pg_type.c.typtype == "d")
+            .order_by(
+                pg_catalog.pg_namespace.c.nspname, pg_catalog.pg_type.c.typname
+            )
+        )
+        return self._pg_type_filter_schema(query, schema)
+
+    @reflection.cache
+    def _load_domains(self, connection, schema=None, **kw):
+        result = connection.execute(self._domain_query(schema))
+
+        domains: List[ReflectedDomain] = []
+        for domain in result.mappings():
+            # strip (30) from character varying(30)
+            attype = re.search(r"([^\(]+)", domain["attype"]).group(1)
+            constraints: List[ReflectedDomainConstraint] = []
+            if domain["connames"]:
+                # When a domain has multiple CHECK constraints, they will
+                # be tested in alphabetical order by name.
+                sorted_constraints = sorted(
+                    zip(domain["connames"], domain["condefs"]),
+                    key=lambda t: t[0],
+                )
+                for name, def_ in sorted_constraints:
+                    # constraint is in the form "CHECK (expression)".
+                    # remove "CHECK (" and the tailing ")".
+                    check = def_[7:-1]
+                    constraints.append({"name": name, "check": check})
+
+            domain_rec: ReflectedDomain = {
+                "name": domain["name"],
+                "schema": domain["schema"],
+                "visible": domain["visible"],
+                "type": attype,
+                "nullable": domain["nullable"],
+                "default": domain["default"],
+                "constraints": constraints,
+                "collation": domain["collname"],
+            }
+            domains.append(domain_rec)
+
+        return domains
+
+    def _set_backslash_escapes(self, connection):
+        # this method is provided as an override hook for descendant
+        # dialects (e.g. Redshift), so removing it may break them
+        std_string = connection.exec_driver_sql(
+            "show standard_conforming_strings"
+        ).scalar()
+        self._backslash_escapes = std_string == "off"
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/dml.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/dml.py
new file mode 100644
index 00000000..1187b6bf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/dml.py
@@ -0,0 +1,339 @@
+# dialects/postgresql/dml.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+from typing import Any
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Union
+
+from . import ext
+from .._typing import _OnConflictConstraintT
+from .._typing import _OnConflictIndexElementsT
+from .._typing import _OnConflictIndexWhereT
+from .._typing import _OnConflictSetT
+from .._typing import _OnConflictWhereT
+from ... import util
+from ...sql import coercions
+from ...sql import roles
+from ...sql import schema
+from ...sql._typing import _DMLTableArgument
+from ...sql.base import _exclusive_against
+from ...sql.base import _generative
+from ...sql.base import ColumnCollection
+from ...sql.base import ReadOnlyColumnCollection
+from ...sql.dml import Insert as StandardInsert
+from ...sql.elements import ClauseElement
+from ...sql.elements import ColumnElement
+from ...sql.elements import KeyedColumnElement
+from ...sql.elements import TextClause
+from ...sql.expression import alias
+from ...util.typing import Self
+
+
+__all__ = ("Insert", "insert")
+
+
+def insert(table: _DMLTableArgument) -> Insert:
+    """Construct a PostgreSQL-specific variant :class:`_postgresql.Insert`
+    construct.
+
+    .. container:: inherited_member
+
+        The :func:`sqlalchemy.dialects.postgresql.insert` function creates
+        a :class:`sqlalchemy.dialects.postgresql.Insert`.  This class is based
+        on the dialect-agnostic :class:`_sql.Insert` construct which may
+        be constructed using the :func:`_sql.insert` function in
+        SQLAlchemy Core.
+
+    The :class:`_postgresql.Insert` construct includes additional methods
+    :meth:`_postgresql.Insert.on_conflict_do_update`,
+    :meth:`_postgresql.Insert.on_conflict_do_nothing`.
+
+    """
+    return Insert(table)
+
+
+class Insert(StandardInsert):
+    """PostgreSQL-specific implementation of INSERT.
+
+    Adds methods for PG-specific syntaxes such as ON CONFLICT.
+
+    The :class:`_postgresql.Insert` object is created using the
+    :func:`sqlalchemy.dialects.postgresql.insert` function.
+
+    """
+
+    stringify_dialect = "postgresql"
+    inherit_cache = False
+
+    @util.memoized_property
+    def excluded(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """Provide the ``excluded`` namespace for an ON CONFLICT statement
+
+        PG's ON CONFLICT clause allows reference to the row that would
+        be inserted, known as ``excluded``.  This attribute provides
+        all columns in this row to be referenceable.
+
+        .. tip::  The :attr:`_postgresql.Insert.excluded` attribute is an
+            instance of :class:`_expression.ColumnCollection`, which provides
+            an interface the same as that of the :attr:`_schema.Table.c`
+            collection described at :ref:`metadata_tables_and_columns`.
+            With this collection, ordinary names are accessible like attributes
+            (e.g. ``stmt.excluded.some_column``), but special names and
+            dictionary method names should be accessed using indexed access,
+            such as ``stmt.excluded["column name"]`` or
+            ``stmt.excluded["values"]``.   See the docstring for
+            :class:`_expression.ColumnCollection` for further examples.
+
+        .. seealso::
+
+            :ref:`postgresql_insert_on_conflict` - example of how
+            to use :attr:`_expression.Insert.excluded`
+
+        """
+        return alias(self.table, name="excluded").columns
+
+    _on_conflict_exclusive = _exclusive_against(
+        "_post_values_clause",
+        msgs={
+            "_post_values_clause": "This Insert construct already has "
+            "an ON CONFLICT clause established"
+        },
+    )
+
+    @_generative
+    @_on_conflict_exclusive
+    def on_conflict_do_update(
+        self,
+        constraint: _OnConflictConstraintT = None,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+        set_: _OnConflictSetT = None,
+        where: _OnConflictWhereT = None,
+    ) -> Self:
+        r"""
+        Specifies a DO UPDATE SET action for ON CONFLICT clause.
+
+        Either the ``constraint`` or ``index_elements`` argument is
+        required, but only one of these can be specified.
+
+        :param constraint:
+         The name of a unique or exclusion constraint on the table,
+         or the constraint object itself if it has a .name attribute.
+
+        :param index_elements:
+         A sequence consisting of string column names, :class:`_schema.Column`
+         objects, or other column expression objects that will be used
+         to infer a target index.
+
+        :param index_where:
+         Additional WHERE criterion that can be used to infer a
+         conditional target index.
+
+        :param set\_:
+         A dictionary or other mapping object
+         where the keys are either names of columns in the target table,
+         or :class:`_schema.Column` objects or other ORM-mapped columns
+         matching that of the target table, and expressions or literals
+         as values, specifying the ``SET`` actions to take.
+
+         .. versionadded:: 1.4 The
+            :paramref:`_postgresql.Insert.on_conflict_do_update.set_`
+            parameter supports :class:`_schema.Column` objects from the target
+            :class:`_schema.Table` as keys.
+
+         .. warning:: This dictionary does **not** take into account
+            Python-specified default UPDATE values or generation functions,
+            e.g. those specified using :paramref:`_schema.Column.onupdate`.
+            These values will not be exercised for an ON CONFLICT style of
+            UPDATE, unless they are manually specified in the
+            :paramref:`.Insert.on_conflict_do_update.set_` dictionary.
+
+        :param where:
+         Optional argument. An expression object representing a ``WHERE``
+         clause that restricts the rows affected by ``DO UPDATE SET``. Rows not
+         meeting the ``WHERE`` condition will not be updated (effectively a
+         ``DO NOTHING`` for those rows).
+
+
+        .. seealso::
+
+            :ref:`postgresql_insert_on_conflict`
+
+        """
+        self._post_values_clause = OnConflictDoUpdate(
+            constraint, index_elements, index_where, set_, where
+        )
+        return self
+
+    @_generative
+    @_on_conflict_exclusive
+    def on_conflict_do_nothing(
+        self,
+        constraint: _OnConflictConstraintT = None,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+    ) -> Self:
+        """
+        Specifies a DO NOTHING action for ON CONFLICT clause.
+
+        The ``constraint`` and ``index_elements`` arguments
+        are optional, but only one of these can be specified.
+
+        :param constraint:
+         The name of a unique or exclusion constraint on the table,
+         or the constraint object itself if it has a .name attribute.
+
+        :param index_elements:
+         A sequence consisting of string column names, :class:`_schema.Column`
+         objects, or other column expression objects that will be used
+         to infer a target index.
+
+        :param index_where:
+         Additional WHERE criterion that can be used to infer a
+         conditional target index.
+
+        .. seealso::
+
+            :ref:`postgresql_insert_on_conflict`
+
+        """
+        self._post_values_clause = OnConflictDoNothing(
+            constraint, index_elements, index_where
+        )
+        return self
+
+
+class OnConflictClause(ClauseElement):
+    stringify_dialect = "postgresql"
+
+    constraint_target: Optional[str]
+    inferred_target_elements: Optional[List[Union[str, schema.Column[Any]]]]
+    inferred_target_whereclause: Optional[
+        Union[ColumnElement[Any], TextClause]
+    ]
+
+    def __init__(
+        self,
+        constraint: _OnConflictConstraintT = None,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+    ):
+        if constraint is not None:
+            if not isinstance(constraint, str) and isinstance(
+                constraint,
+                (schema.Constraint, ext.ExcludeConstraint),
+            ):
+                constraint = getattr(constraint, "name") or constraint
+
+        if constraint is not None:
+            if index_elements is not None:
+                raise ValueError(
+                    "'constraint' and 'index_elements' are mutually exclusive"
+                )
+
+            if isinstance(constraint, str):
+                self.constraint_target = constraint
+                self.inferred_target_elements = None
+                self.inferred_target_whereclause = None
+            elif isinstance(constraint, schema.Index):
+                index_elements = constraint.expressions
+                index_where = constraint.dialect_options["postgresql"].get(
+                    "where"
+                )
+            elif isinstance(constraint, ext.ExcludeConstraint):
+                index_elements = constraint.columns
+                index_where = constraint.where
+            else:
+                index_elements = constraint.columns
+                index_where = constraint.dialect_options["postgresql"].get(
+                    "where"
+                )
+
+        if index_elements is not None:
+            self.constraint_target = None
+            self.inferred_target_elements = [
+                coercions.expect(roles.DDLConstraintColumnRole, column)
+                for column in index_elements
+            ]
+
+            self.inferred_target_whereclause = (
+                coercions.expect(
+                    (
+                        roles.StatementOptionRole
+                        if isinstance(constraint, ext.ExcludeConstraint)
+                        else roles.WhereHavingRole
+                    ),
+                    index_where,
+                )
+                if index_where is not None
+                else None
+            )
+
+        elif constraint is None:
+            self.constraint_target = self.inferred_target_elements = (
+                self.inferred_target_whereclause
+            ) = None
+
+
+class OnConflictDoNothing(OnConflictClause):
+    __visit_name__ = "on_conflict_do_nothing"
+
+
+class OnConflictDoUpdate(OnConflictClause):
+    __visit_name__ = "on_conflict_do_update"
+
+    update_values_to_set: List[Tuple[Union[schema.Column[Any], str], Any]]
+    update_whereclause: Optional[ColumnElement[Any]]
+
+    def __init__(
+        self,
+        constraint: _OnConflictConstraintT = None,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+        set_: _OnConflictSetT = None,
+        where: _OnConflictWhereT = None,
+    ):
+        super().__init__(
+            constraint=constraint,
+            index_elements=index_elements,
+            index_where=index_where,
+        )
+
+        if (
+            self.inferred_target_elements is None
+            and self.constraint_target is None
+        ):
+            raise ValueError(
+                "Either constraint or index_elements, "
+                "but not both, must be specified unless DO NOTHING"
+            )
+
+        if isinstance(set_, dict):
+            if not set_:
+                raise ValueError("set parameter dictionary must not be empty")
+        elif isinstance(set_, ColumnCollection):
+            set_ = dict(set_)
+        else:
+            raise ValueError(
+                "set parameter must be a non-empty dictionary "
+                "or a ColumnCollection such as the `.c.` collection "
+                "of a Table object"
+            )
+        self.update_values_to_set = [
+            (coercions.expect(roles.DMLColumnRole, key), value)
+            for key, value in set_.items()
+        ]
+        self.update_whereclause = (
+            coercions.expect(roles.WhereHavingRole, where)
+            if where is not None
+            else None
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ext.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ext.py
new file mode 100644
index 00000000..94466ae0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ext.py
@@ -0,0 +1,501 @@
+# dialects/postgresql/ext.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+from typing import Any
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from . import types
+from .array import ARRAY
+from ...sql import coercions
+from ...sql import elements
+from ...sql import expression
+from ...sql import functions
+from ...sql import roles
+from ...sql import schema
+from ...sql.schema import ColumnCollectionConstraint
+from ...sql.sqltypes import TEXT
+from ...sql.visitors import InternalTraversal
+
+_T = TypeVar("_T", bound=Any)
+
+if TYPE_CHECKING:
+    from ...sql.visitors import _TraverseInternalsType
+
+
+class aggregate_order_by(expression.ColumnElement):
+    """Represent a PostgreSQL aggregate order by expression.
+
+    E.g.::
+
+        from sqlalchemy.dialects.postgresql import aggregate_order_by
+
+        expr = func.array_agg(aggregate_order_by(table.c.a, table.c.b.desc()))
+        stmt = select(expr)
+
+    would represent the expression:
+
+    .. sourcecode:: sql
+
+        SELECT array_agg(a ORDER BY b DESC) FROM table;
+
+    Similarly::
+
+        expr = func.string_agg(
+            table.c.a, aggregate_order_by(literal_column("','"), table.c.a)
+        )
+        stmt = select(expr)
+
+    Would represent:
+
+    .. sourcecode:: sql
+
+        SELECT string_agg(a, ',' ORDER BY a) FROM table;
+
+    .. versionchanged:: 1.2.13 - the ORDER BY argument may be multiple terms
+
+    .. seealso::
+
+        :class:`_functions.array_agg`
+
+    """
+
+    __visit_name__ = "aggregate_order_by"
+
+    stringify_dialect = "postgresql"
+    _traverse_internals: _TraverseInternalsType = [
+        ("target", InternalTraversal.dp_clauseelement),
+        ("type", InternalTraversal.dp_type),
+        ("order_by", InternalTraversal.dp_clauseelement),
+    ]
+
+    def __init__(self, target, *order_by):
+        self.target = coercions.expect(roles.ExpressionElementRole, target)
+        self.type = self.target.type
+
+        _lob = len(order_by)
+        if _lob == 0:
+            raise TypeError("at least one ORDER BY element is required")
+        elif _lob == 1:
+            self.order_by = coercions.expect(
+                roles.ExpressionElementRole, order_by[0]
+            )
+        else:
+            self.order_by = elements.ClauseList(
+                *order_by, _literal_as_text_role=roles.ExpressionElementRole
+            )
+
+    def self_group(self, against=None):
+        return self
+
+    def get_children(self, **kwargs):
+        return self.target, self.order_by
+
+    def _copy_internals(self, clone=elements._clone, **kw):
+        self.target = clone(self.target, **kw)
+        self.order_by = clone(self.order_by, **kw)
+
+    @property
+    def _from_objects(self):
+        return self.target._from_objects + self.order_by._from_objects
+
+
+class ExcludeConstraint(ColumnCollectionConstraint):
+    """A table-level EXCLUDE constraint.
+
+    Defines an EXCLUDE constraint as described in the `PostgreSQL
+    documentation`__.
+
+    __ https://www.postgresql.org/docs/current/static/sql-createtable.html#SQL-CREATETABLE-EXCLUDE
+
+    """  # noqa
+
+    __visit_name__ = "exclude_constraint"
+
+    where = None
+    inherit_cache = False
+
+    create_drop_stringify_dialect = "postgresql"
+
+    @elements._document_text_coercion(
+        "where",
+        ":class:`.ExcludeConstraint`",
+        ":paramref:`.ExcludeConstraint.where`",
+    )
+    def __init__(self, *elements, **kw):
+        r"""
+        Create an :class:`.ExcludeConstraint` object.
+
+        E.g.::
+
+            const = ExcludeConstraint(
+                (Column("period"), "&&"),
+                (Column("group"), "="),
+                where=(Column("group") != "some group"),
+                ops={"group": "my_operator_class"},
+            )
+
+        The constraint is normally embedded into the :class:`_schema.Table`
+        construct
+        directly, or added later using :meth:`.append_constraint`::
+
+            some_table = Table(
+                "some_table",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("period", TSRANGE()),
+                Column("group", String),
+            )
+
+            some_table.append_constraint(
+                ExcludeConstraint(
+                    (some_table.c.period, "&&"),
+                    (some_table.c.group, "="),
+                    where=some_table.c.group != "some group",
+                    name="some_table_excl_const",
+                    ops={"group": "my_operator_class"},
+                )
+            )
+
+        The exclude constraint defined in this example requires the
+        ``btree_gist`` extension, that can be created using the
+        command ``CREATE EXTENSION btree_gist;``.
+
+        :param \*elements:
+
+          A sequence of two tuples of the form ``(column, operator)`` where
+          "column" is either a :class:`_schema.Column` object, or a SQL
+          expression element (e.g. ``func.int8range(table.from, table.to)``)
+          or the name of a column as string, and "operator" is a string
+          containing the operator to use (e.g. `"&&"` or `"="`).
+
+          In order to specify a column name when a :class:`_schema.Column`
+          object is not available, while ensuring
+          that any necessary quoting rules take effect, an ad-hoc
+          :class:`_schema.Column` or :func:`_expression.column`
+          object should be used.
+          The ``column`` may also be a string SQL expression when
+          passed as :func:`_expression.literal_column` or
+          :func:`_expression.text`
+
+        :param name:
+          Optional, the in-database name of this constraint.
+
+        :param deferrable:
+          Optional bool.  If set, emit DEFERRABLE or NOT DEFERRABLE when
+          issuing DDL for this constraint.
+
+        :param initially:
+          Optional string.  If set, emit INITIALLY <value> when issuing DDL
+          for this constraint.
+
+        :param using:
+          Optional string.  If set, emit USING <index_method> when issuing DDL
+          for this constraint. Defaults to 'gist'.
+
+        :param where:
+          Optional SQL expression construct or literal SQL string.
+          If set, emit WHERE <predicate> when issuing DDL
+          for this constraint.
+
+        :param ops:
+          Optional dictionary.  Used to define operator classes for the
+          elements; works the same way as that of the
+          :ref:`postgresql_ops <postgresql_operator_classes>`
+          parameter specified to the :class:`_schema.Index` construct.
+
+          .. versionadded:: 1.3.21
+
+          .. seealso::
+
+            :ref:`postgresql_operator_classes` - general description of how
+            PostgreSQL operator classes are specified.
+
+        """
+        columns = []
+        render_exprs = []
+        self.operators = {}
+
+        expressions, operators = zip(*elements)
+
+        for (expr, column, strname, add_element), operator in zip(
+            coercions.expect_col_expression_collection(
+                roles.DDLConstraintColumnRole, expressions
+            ),
+            operators,
+        ):
+            if add_element is not None:
+                columns.append(add_element)
+
+            name = column.name if column is not None else strname
+
+            if name is not None:
+                # backwards compat
+                self.operators[name] = operator
+
+            render_exprs.append((expr, name, operator))
+
+        self._render_exprs = render_exprs
+
+        ColumnCollectionConstraint.__init__(
+            self,
+            *columns,
+            name=kw.get("name"),
+            deferrable=kw.get("deferrable"),
+            initially=kw.get("initially"),
+        )
+        self.using = kw.get("using", "gist")
+        where = kw.get("where")
+        if where is not None:
+            self.where = coercions.expect(roles.StatementOptionRole, where)
+
+        self.ops = kw.get("ops", {})
+
+    def _set_parent(self, table, **kw):
+        super()._set_parent(table)
+
+        self._render_exprs = [
+            (
+                expr if not isinstance(expr, str) else table.c[expr],
+                name,
+                operator,
+            )
+            for expr, name, operator in (self._render_exprs)
+        ]
+
+    def _copy(self, target_table=None, **kw):
+        elements = [
+            (
+                schema._copy_expression(expr, self.parent, target_table),
+                operator,
+            )
+            for expr, _, operator in self._render_exprs
+        ]
+        c = self.__class__(
+            *elements,
+            name=self.name,
+            deferrable=self.deferrable,
+            initially=self.initially,
+            where=self.where,
+            using=self.using,
+        )
+        c.dispatch._update(self.dispatch)
+        return c
+
+
+def array_agg(*arg, **kw):
+    """PostgreSQL-specific form of :class:`_functions.array_agg`, ensures
+    return type is :class:`_postgresql.ARRAY` and not
+    the plain :class:`_types.ARRAY`, unless an explicit ``type_``
+    is passed.
+
+    """
+    kw["_default_array_type"] = ARRAY
+    return functions.func.array_agg(*arg, **kw)
+
+
+class _regconfig_fn(functions.GenericFunction[_T]):
+    inherit_cache = True
+
+    def __init__(self, *args, **kwargs):
+        args = list(args)
+        if len(args) > 1:
+            initial_arg = coercions.expect(
+                roles.ExpressionElementRole,
+                args.pop(0),
+                name=getattr(self, "name", None),
+                apply_propagate_attrs=self,
+                type_=types.REGCONFIG,
+            )
+            initial_arg = [initial_arg]
+        else:
+            initial_arg = []
+
+        addtl_args = [
+            coercions.expect(
+                roles.ExpressionElementRole,
+                c,
+                name=getattr(self, "name", None),
+                apply_propagate_attrs=self,
+            )
+            for c in args
+        ]
+        super().__init__(*(initial_arg + addtl_args), **kwargs)
+
+
+class to_tsvector(_regconfig_fn):
+    """The PostgreSQL ``to_tsvector`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_postgresql.TSVECTOR`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.to_tsvector` will be used automatically when invoking
+    ``sqlalchemy.func.to_tsvector()``, ensuring the correct argument and return
+    type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = types.TSVECTOR
+
+
+class to_tsquery(_regconfig_fn):
+    """The PostgreSQL ``to_tsquery`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_postgresql.TSQUERY`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.to_tsquery` will be used automatically when invoking
+    ``sqlalchemy.func.to_tsquery()``, ensuring the correct argument and return
+    type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = types.TSQUERY
+
+
+class plainto_tsquery(_regconfig_fn):
+    """The PostgreSQL ``plainto_tsquery`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_postgresql.TSQUERY`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.plainto_tsquery` will be used automatically when
+    invoking ``sqlalchemy.func.plainto_tsquery()``, ensuring the correct
+    argument and return type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = types.TSQUERY
+
+
+class phraseto_tsquery(_regconfig_fn):
+    """The PostgreSQL ``phraseto_tsquery`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_postgresql.TSQUERY`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.phraseto_tsquery` will be used automatically when
+    invoking ``sqlalchemy.func.phraseto_tsquery()``, ensuring the correct
+    argument and return type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = types.TSQUERY
+
+
+class websearch_to_tsquery(_regconfig_fn):
+    """The PostgreSQL ``websearch_to_tsquery`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_postgresql.TSQUERY`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.websearch_to_tsquery` will be used automatically when
+    invoking ``sqlalchemy.func.websearch_to_tsquery()``, ensuring the correct
+    argument and return type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = types.TSQUERY
+
+
+class ts_headline(_regconfig_fn):
+    """The PostgreSQL ``ts_headline`` SQL function.
+
+    This function applies automatic casting of the REGCONFIG argument
+    to use the :class:`_postgresql.REGCONFIG` datatype automatically,
+    and applies a return type of :class:`_types.TEXT`.
+
+    Assuming the PostgreSQL dialect has been imported, either by invoking
+    ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL
+    engine using ``create_engine("postgresql...")``,
+    :class:`_postgresql.ts_headline` will be used automatically when invoking
+    ``sqlalchemy.func.ts_headline()``, ensuring the correct argument and return
+    type handlers are used at compile and execution time.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    inherit_cache = True
+    type = TEXT
+
+    def __init__(self, *args, **kwargs):
+        args = list(args)
+
+        # parse types according to
+        # https://www.postgresql.org/docs/current/textsearch-controls.html#TEXTSEARCH-HEADLINE
+        if len(args) < 2:
+            # invalid args; don't do anything
+            has_regconfig = False
+        elif (
+            isinstance(args[1], elements.ColumnElement)
+            and args[1].type._type_affinity is types.TSQUERY
+        ):
+            # tsquery is second argument, no regconfig argument
+            has_regconfig = False
+        else:
+            has_regconfig = True
+
+        if has_regconfig:
+            initial_arg = coercions.expect(
+                roles.ExpressionElementRole,
+                args.pop(0),
+                apply_propagate_attrs=self,
+                name=getattr(self, "name", None),
+                type_=types.REGCONFIG,
+            )
+            initial_arg = [initial_arg]
+        else:
+            initial_arg = []
+
+        addtl_args = [
+            coercions.expect(
+                roles.ExpressionElementRole,
+                c,
+                name=getattr(self, "name", None),
+                apply_propagate_attrs=self,
+            )
+            for c in args
+        ]
+        super().__init__(*(initial_arg + addtl_args), **kwargs)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/hstore.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/hstore.py
new file mode 100644
index 00000000..0a915b17
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/hstore.py
@@ -0,0 +1,406 @@
+# dialects/postgresql/hstore.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import re
+
+from .array import ARRAY
+from .operators import CONTAINED_BY
+from .operators import CONTAINS
+from .operators import GETITEM
+from .operators import HAS_ALL
+from .operators import HAS_ANY
+from .operators import HAS_KEY
+from ... import types as sqltypes
+from ...sql import functions as sqlfunc
+
+
+__all__ = ("HSTORE", "hstore")
+
+
+class HSTORE(sqltypes.Indexable, sqltypes.Concatenable, sqltypes.TypeEngine):
+    """Represent the PostgreSQL HSTORE type.
+
+    The :class:`.HSTORE` type stores dictionaries containing strings, e.g.::
+
+        data_table = Table(
+            "data_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", HSTORE),
+        )
+
+        with engine.connect() as conn:
+            conn.execute(
+                data_table.insert(), data={"key1": "value1", "key2": "value2"}
+            )
+
+    :class:`.HSTORE` provides for a wide range of operations, including:
+
+    * Index operations::
+
+        data_table.c.data["some key"] == "some value"
+
+    * Containment operations::
+
+        data_table.c.data.has_key("some key")
+
+        data_table.c.data.has_all(["one", "two", "three"])
+
+    * Concatenation::
+
+        data_table.c.data + {"k1": "v1"}
+
+    For a full list of special methods see
+    :class:`.HSTORE.comparator_factory`.
+
+    .. container:: topic
+
+        **Detecting Changes in HSTORE columns when using the ORM**
+
+        For usage with the SQLAlchemy ORM, it may be desirable to combine the
+        usage of :class:`.HSTORE` with :class:`.MutableDict` dictionary now
+        part of the :mod:`sqlalchemy.ext.mutable` extension. This extension
+        will allow "in-place" changes to the dictionary, e.g. addition of new
+        keys or replacement/removal of existing keys to/from the current
+        dictionary, to produce events which will be detected by the unit of
+        work::
+
+            from sqlalchemy.ext.mutable import MutableDict
+
+
+            class MyClass(Base):
+                __tablename__ = "data_table"
+
+                id = Column(Integer, primary_key=True)
+                data = Column(MutableDict.as_mutable(HSTORE))
+
+
+            my_object = session.query(MyClass).one()
+
+            # in-place mutation, requires Mutable extension
+            # in order for the ORM to detect
+            my_object.data["some_key"] = "some value"
+
+            session.commit()
+
+        When the :mod:`sqlalchemy.ext.mutable` extension is not used, the ORM
+        will not be alerted to any changes to the contents of an existing
+        dictionary, unless that dictionary value is re-assigned to the
+        HSTORE-attribute itself, thus generating a change event.
+
+    .. seealso::
+
+        :class:`.hstore` - render the PostgreSQL ``hstore()`` function.
+
+
+    """  # noqa: E501
+
+    __visit_name__ = "HSTORE"
+    hashable = False
+    text_type = sqltypes.Text()
+
+    def __init__(self, text_type=None):
+        """Construct a new :class:`.HSTORE`.
+
+        :param text_type: the type that should be used for indexed values.
+         Defaults to :class:`_types.Text`.
+
+        """
+        if text_type is not None:
+            self.text_type = text_type
+
+    class Comparator(
+        sqltypes.Indexable.Comparator, sqltypes.Concatenable.Comparator
+    ):
+        """Define comparison operations for :class:`.HSTORE`."""
+
+        def has_key(self, other):
+            """Boolean expression.  Test for presence of a key.  Note that the
+            key may be a SQLA expression.
+            """
+            return self.operate(HAS_KEY, other, result_type=sqltypes.Boolean)
+
+        def has_all(self, other):
+            """Boolean expression.  Test for presence of all keys in jsonb"""
+            return self.operate(HAS_ALL, other, result_type=sqltypes.Boolean)
+
+        def has_any(self, other):
+            """Boolean expression.  Test for presence of any key in jsonb"""
+            return self.operate(HAS_ANY, other, result_type=sqltypes.Boolean)
+
+        def contains(self, other, **kwargs):
+            """Boolean expression.  Test if keys (or array) are a superset
+            of/contained the keys of the argument jsonb expression.
+
+            kwargs may be ignored by this operator but are required for API
+            conformance.
+            """
+            return self.operate(CONTAINS, other, result_type=sqltypes.Boolean)
+
+        def contained_by(self, other):
+            """Boolean expression.  Test if keys are a proper subset of the
+            keys of the argument jsonb expression.
+            """
+            return self.operate(
+                CONTAINED_BY, other, result_type=sqltypes.Boolean
+            )
+
+        def _setup_getitem(self, index):
+            return GETITEM, index, self.type.text_type
+
+        def defined(self, key):
+            """Boolean expression.  Test for presence of a non-NULL value for
+            the key.  Note that the key may be a SQLA expression.
+            """
+            return _HStoreDefinedFunction(self.expr, key)
+
+        def delete(self, key):
+            """HStore expression.  Returns the contents of this hstore with the
+            given key deleted.  Note that the key may be a SQLA expression.
+            """
+            if isinstance(key, dict):
+                key = _serialize_hstore(key)
+            return _HStoreDeleteFunction(self.expr, key)
+
+        def slice(self, array):
+            """HStore expression.  Returns a subset of an hstore defined by
+            array of keys.
+            """
+            return _HStoreSliceFunction(self.expr, array)
+
+        def keys(self):
+            """Text array expression.  Returns array of keys."""
+            return _HStoreKeysFunction(self.expr)
+
+        def vals(self):
+            """Text array expression.  Returns array of values."""
+            return _HStoreValsFunction(self.expr)
+
+        def array(self):
+            """Text array expression.  Returns array of alternating keys and
+            values.
+            """
+            return _HStoreArrayFunction(self.expr)
+
+        def matrix(self):
+            """Text array expression.  Returns array of [key, value] pairs."""
+            return _HStoreMatrixFunction(self.expr)
+
+    comparator_factory = Comparator
+
+    def bind_processor(self, dialect):
+        # note that dialect-specific types like that of psycopg and
+        # psycopg2 will override this method to allow driver-level conversion
+        # instead, see _PsycopgHStore
+        def process(value):
+            if isinstance(value, dict):
+                return _serialize_hstore(value)
+            else:
+                return value
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        # note that dialect-specific types like that of psycopg and
+        # psycopg2 will override this method to allow driver-level conversion
+        # instead, see _PsycopgHStore
+        def process(value):
+            if value is not None:
+                return _parse_hstore(value)
+            else:
+                return value
+
+        return process
+
+
+class hstore(sqlfunc.GenericFunction):
+    """Construct an hstore value within a SQL expression using the
+    PostgreSQL ``hstore()`` function.
+
+    The :class:`.hstore` function accepts one or two arguments as described
+    in the PostgreSQL documentation.
+
+    E.g.::
+
+        from sqlalchemy.dialects.postgresql import array, hstore
+
+        select(hstore("key1", "value1"))
+
+        select(
+            hstore(
+                array(["key1", "key2", "key3"]),
+                array(["value1", "value2", "value3"]),
+            )
+        )
+
+    .. seealso::
+
+        :class:`.HSTORE` - the PostgreSQL ``HSTORE`` datatype.
+
+    """
+
+    type = HSTORE
+    name = "hstore"
+    inherit_cache = True
+
+
+class _HStoreDefinedFunction(sqlfunc.GenericFunction):
+    type = sqltypes.Boolean
+    name = "defined"
+    inherit_cache = True
+
+
+class _HStoreDeleteFunction(sqlfunc.GenericFunction):
+    type = HSTORE
+    name = "delete"
+    inherit_cache = True
+
+
+class _HStoreSliceFunction(sqlfunc.GenericFunction):
+    type = HSTORE
+    name = "slice"
+    inherit_cache = True
+
+
+class _HStoreKeysFunction(sqlfunc.GenericFunction):
+    type = ARRAY(sqltypes.Text)
+    name = "akeys"
+    inherit_cache = True
+
+
+class _HStoreValsFunction(sqlfunc.GenericFunction):
+    type = ARRAY(sqltypes.Text)
+    name = "avals"
+    inherit_cache = True
+
+
+class _HStoreArrayFunction(sqlfunc.GenericFunction):
+    type = ARRAY(sqltypes.Text)
+    name = "hstore_to_array"
+    inherit_cache = True
+
+
+class _HStoreMatrixFunction(sqlfunc.GenericFunction):
+    type = ARRAY(sqltypes.Text)
+    name = "hstore_to_matrix"
+    inherit_cache = True
+
+
+#
+# parsing.  note that none of this is used with the psycopg2 backend,
+# which provides its own native extensions.
+#
+
+# My best guess at the parsing rules of hstore literals, since no formal
+# grammar is given.  This is mostly reverse engineered from PG's input parser
+# behavior.
+HSTORE_PAIR_RE = re.compile(
+    r"""
+(
+  "(?P<key> (\\ . | [^"])* )"       # Quoted key
+)
+[ ]* => [ ]*    # Pair operator, optional adjoining whitespace
+(
+    (?P<value_null> NULL )          # NULL value
+  | "(?P<value> (\\ . | [^"])* )"   # Quoted value
+)
+""",
+    re.VERBOSE,
+)
+
+HSTORE_DELIMITER_RE = re.compile(
+    r"""
+[ ]* , [ ]*
+""",
+    re.VERBOSE,
+)
+
+
+def _parse_error(hstore_str, pos):
+    """format an unmarshalling error."""
+
+    ctx = 20
+    hslen = len(hstore_str)
+
+    parsed_tail = hstore_str[max(pos - ctx - 1, 0) : min(pos, hslen)]
+    residual = hstore_str[min(pos, hslen) : min(pos + ctx + 1, hslen)]
+
+    if len(parsed_tail) > ctx:
+        parsed_tail = "[...]" + parsed_tail[1:]
+    if len(residual) > ctx:
+        residual = residual[:-1] + "[...]"
+
+    return "After %r, could not parse residual at position %d: %r" % (
+        parsed_tail,
+        pos,
+        residual,
+    )
+
+
+def _parse_hstore(hstore_str):
+    """Parse an hstore from its literal string representation.
+
+    Attempts to approximate PG's hstore input parsing rules as closely as
+    possible. Although currently this is not strictly necessary, since the
+    current implementation of hstore's output syntax is stricter than what it
+    accepts as input, the documentation makes no guarantees that will always
+    be the case.
+
+
+
+    """
+    result = {}
+    pos = 0
+    pair_match = HSTORE_PAIR_RE.match(hstore_str)
+
+    while pair_match is not None:
+        key = pair_match.group("key").replace(r"\"", '"').replace("\\\\", "\\")
+        if pair_match.group("value_null"):
+            value = None
+        else:
+            value = (
+                pair_match.group("value")
+                .replace(r"\"", '"')
+                .replace("\\\\", "\\")
+            )
+        result[key] = value
+
+        pos += pair_match.end()
+
+        delim_match = HSTORE_DELIMITER_RE.match(hstore_str[pos:])
+        if delim_match is not None:
+            pos += delim_match.end()
+
+        pair_match = HSTORE_PAIR_RE.match(hstore_str[pos:])
+
+    if pos != len(hstore_str):
+        raise ValueError(_parse_error(hstore_str, pos))
+
+    return result
+
+
+def _serialize_hstore(val):
+    """Serialize a dictionary into an hstore literal.  Keys and values must
+    both be strings (except None for values).
+
+    """
+
+    def esc(s, position):
+        if position == "value" and s is None:
+            return "NULL"
+        elif isinstance(s, str):
+            return '"%s"' % s.replace("\\", "\\\\").replace('"', r"\"")
+        else:
+            raise ValueError(
+                "%r in %s position is not a string." % (s, position)
+            )
+
+    return ", ".join(
+        "%s=>%s" % (esc(k, "key"), esc(v, "value")) for k, v in val.items()
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/json.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/json.py
new file mode 100644
index 00000000..2f26b39e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/json.py
@@ -0,0 +1,338 @@
+# dialects/postgresql/json.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from .array import ARRAY
+from .array import array as _pg_array
+from .operators import ASTEXT
+from .operators import CONTAINED_BY
+from .operators import CONTAINS
+from .operators import DELETE_PATH
+from .operators import HAS_ALL
+from .operators import HAS_ANY
+from .operators import HAS_KEY
+from .operators import JSONPATH_ASTEXT
+from .operators import PATH_EXISTS
+from .operators import PATH_MATCH
+from ... import types as sqltypes
+from ...sql import cast
+
+__all__ = ("JSON", "JSONB")
+
+
+class JSONPathType(sqltypes.JSON.JSONPathType):
+    def _processor(self, dialect, super_proc):
+        def process(value):
+            if isinstance(value, str):
+                # If it's already a string assume that it's in json path
+                # format. This allows using cast with json paths literals
+                return value
+            elif value:
+                # If it's already a string assume that it's in json path
+                # format. This allows using cast with json paths literals
+                value = "{%s}" % (", ".join(map(str, value)))
+            else:
+                value = "{}"
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+    def bind_processor(self, dialect):
+        return self._processor(dialect, self.string_bind_processor(dialect))
+
+    def literal_processor(self, dialect):
+        return self._processor(dialect, self.string_literal_processor(dialect))
+
+
+class JSONPATH(JSONPathType):
+    """JSON Path Type.
+
+    This is usually required to cast literal values to json path when using
+    json search like function, such as ``jsonb_path_query_array`` or
+    ``jsonb_path_exists``::
+
+        stmt = sa.select(
+            sa.func.jsonb_path_query_array(
+                table.c.jsonb_col, cast("$.address.id", JSONPATH)
+            )
+        )
+
+    """
+
+    __visit_name__ = "JSONPATH"
+
+
+class JSON(sqltypes.JSON):
+    """Represent the PostgreSQL JSON type.
+
+    :class:`_postgresql.JSON` is used automatically whenever the base
+    :class:`_types.JSON` datatype is used against a PostgreSQL backend,
+    however base :class:`_types.JSON` datatype does not provide Python
+    accessors for PostgreSQL-specific comparison methods such as
+    :meth:`_postgresql.JSON.Comparator.astext`; additionally, to use
+    PostgreSQL ``JSONB``, the :class:`_postgresql.JSONB` datatype should
+    be used explicitly.
+
+    .. seealso::
+
+        :class:`_types.JSON` - main documentation for the generic
+        cross-platform JSON datatype.
+
+    The operators provided by the PostgreSQL version of :class:`_types.JSON`
+    include:
+
+    * Index operations (the ``->`` operator)::
+
+        data_table.c.data["some key"]
+
+        data_table.c.data[5]
+
+    * Index operations returning text
+      (the ``->>`` operator)::
+
+        data_table.c.data["some key"].astext == "some value"
+
+      Note that equivalent functionality is available via the
+      :attr:`.JSON.Comparator.as_string` accessor.
+
+    * Index operations with CAST
+      (equivalent to ``CAST(col ->> ['some key'] AS <type>)``)::
+
+        data_table.c.data["some key"].astext.cast(Integer) == 5
+
+      Note that equivalent functionality is available via the
+      :attr:`.JSON.Comparator.as_integer` and similar accessors.
+
+    * Path index operations (the ``#>`` operator)::
+
+        data_table.c.data[("key_1", "key_2", 5, ..., "key_n")]
+
+    * Path index operations returning text (the ``#>>`` operator)::
+
+        data_table.c.data[
+            ("key_1", "key_2", 5, ..., "key_n")
+        ].astext == "some value"
+
+    Index operations return an expression object whose type defaults to
+    :class:`_types.JSON` by default,
+    so that further JSON-oriented instructions
+    may be called upon the result type.
+
+    Custom serializers and deserializers are specified at the dialect level,
+    that is using :func:`_sa.create_engine`.  The reason for this is that when
+    using psycopg2, the DBAPI only allows serializers at the per-cursor
+    or per-connection level.   E.g.::
+
+        engine = create_engine(
+            "postgresql+psycopg2://scott:tiger@localhost/test",
+            json_serializer=my_serialize_fn,
+            json_deserializer=my_deserialize_fn,
+        )
+
+    When using the psycopg2 dialect, the json_deserializer is registered
+    against the database using ``psycopg2.extras.register_default_json``.
+
+    .. seealso::
+
+        :class:`_types.JSON` - Core level JSON type
+
+        :class:`_postgresql.JSONB`
+
+    """  # noqa
+
+    render_bind_cast = True
+    astext_type = sqltypes.Text()
+
+    def __init__(self, none_as_null=False, astext_type=None):
+        """Construct a :class:`_types.JSON` type.
+
+        :param none_as_null: if True, persist the value ``None`` as a
+         SQL NULL value, not the JSON encoding of ``null``.   Note that
+         when this flag is False, the :func:`.null` construct can still
+         be used to persist a NULL value::
+
+             from sqlalchemy import null
+
+             conn.execute(table.insert(), {"data": null()})
+
+         .. seealso::
+
+              :attr:`_types.JSON.NULL`
+
+        :param astext_type: the type to use for the
+         :attr:`.JSON.Comparator.astext`
+         accessor on indexed attributes.  Defaults to :class:`_types.Text`.
+
+        """
+        super().__init__(none_as_null=none_as_null)
+        if astext_type is not None:
+            self.astext_type = astext_type
+
+    class Comparator(sqltypes.JSON.Comparator):
+        """Define comparison operations for :class:`_types.JSON`."""
+
+        @property
+        def astext(self):
+            """On an indexed expression, use the "astext" (e.g. "->>")
+            conversion when rendered in SQL.
+
+            E.g.::
+
+                select(data_table.c.data["some key"].astext)
+
+            .. seealso::
+
+                :meth:`_expression.ColumnElement.cast`
+
+            """
+            if isinstance(self.expr.right.type, sqltypes.JSON.JSONPathType):
+                return self.expr.left.operate(
+                    JSONPATH_ASTEXT,
+                    self.expr.right,
+                    result_type=self.type.astext_type,
+                )
+            else:
+                return self.expr.left.operate(
+                    ASTEXT, self.expr.right, result_type=self.type.astext_type
+                )
+
+    comparator_factory = Comparator
+
+
+class JSONB(JSON):
+    """Represent the PostgreSQL JSONB type.
+
+    The :class:`_postgresql.JSONB` type stores arbitrary JSONB format data,
+    e.g.::
+
+        data_table = Table(
+            "data_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", JSONB),
+        )
+
+        with engine.connect() as conn:
+            conn.execute(
+                data_table.insert(), data={"key1": "value1", "key2": "value2"}
+            )
+
+    The :class:`_postgresql.JSONB` type includes all operations provided by
+    :class:`_types.JSON`, including the same behaviors for indexing
+    operations.
+    It also adds additional operators specific to JSONB, including
+    :meth:`.JSONB.Comparator.has_key`, :meth:`.JSONB.Comparator.has_all`,
+    :meth:`.JSONB.Comparator.has_any`, :meth:`.JSONB.Comparator.contains`,
+    :meth:`.JSONB.Comparator.contained_by`,
+    :meth:`.JSONB.Comparator.delete_path`,
+    :meth:`.JSONB.Comparator.path_exists` and
+    :meth:`.JSONB.Comparator.path_match`.
+
+    Like the :class:`_types.JSON` type, the :class:`_postgresql.JSONB`
+    type does not detect
+    in-place changes when used with the ORM, unless the
+    :mod:`sqlalchemy.ext.mutable` extension is used.
+
+    Custom serializers and deserializers
+    are shared with the :class:`_types.JSON` class,
+    using the ``json_serializer``
+    and ``json_deserializer`` keyword arguments.  These must be specified
+    at the dialect level using :func:`_sa.create_engine`.  When using
+    psycopg2, the serializers are associated with the jsonb type using
+    ``psycopg2.extras.register_default_jsonb`` on a per-connection basis,
+    in the same way that ``psycopg2.extras.register_default_json`` is used
+    to register these handlers with the json type.
+
+    .. seealso::
+
+        :class:`_types.JSON`
+
+    """
+
+    __visit_name__ = "JSONB"
+
+    class Comparator(JSON.Comparator):
+        """Define comparison operations for :class:`_types.JSON`."""
+
+        def has_key(self, other):
+            """Boolean expression.  Test for presence of a key (equivalent of
+            the ``?`` operator).  Note that the key may be a SQLA expression.
+            """
+            return self.operate(HAS_KEY, other, result_type=sqltypes.Boolean)
+
+        def has_all(self, other):
+            """Boolean expression.  Test for presence of all keys in jsonb
+            (equivalent of the ``?&`` operator)
+            """
+            return self.operate(HAS_ALL, other, result_type=sqltypes.Boolean)
+
+        def has_any(self, other):
+            """Boolean expression.  Test for presence of any key in jsonb
+            (equivalent of the ``?|`` operator)
+            """
+            return self.operate(HAS_ANY, other, result_type=sqltypes.Boolean)
+
+        def contains(self, other, **kwargs):
+            """Boolean expression.  Test if keys (or array) are a superset
+            of/contained the keys of the argument jsonb expression
+            (equivalent of the ``@>`` operator).
+
+            kwargs may be ignored by this operator but are required for API
+            conformance.
+            """
+            return self.operate(CONTAINS, other, result_type=sqltypes.Boolean)
+
+        def contained_by(self, other):
+            """Boolean expression.  Test if keys are a proper subset of the
+            keys of the argument jsonb expression
+            (equivalent of the ``<@`` operator).
+            """
+            return self.operate(
+                CONTAINED_BY, other, result_type=sqltypes.Boolean
+            )
+
+        def delete_path(self, array):
+            """JSONB expression. Deletes field or array element specified in
+            the argument array (equivalent of the ``#-`` operator).
+
+            The input may be a list of strings that will be coerced to an
+            ``ARRAY`` or an instance of :meth:`_postgres.array`.
+
+            .. versionadded:: 2.0
+            """
+            if not isinstance(array, _pg_array):
+                array = _pg_array(array)
+            right_side = cast(array, ARRAY(sqltypes.TEXT))
+            return self.operate(DELETE_PATH, right_side, result_type=JSONB)
+
+        def path_exists(self, other):
+            """Boolean expression. Test for presence of item given by the
+            argument JSONPath expression (equivalent of the ``@?`` operator).
+
+            .. versionadded:: 2.0
+            """
+            return self.operate(
+                PATH_EXISTS, other, result_type=sqltypes.Boolean
+            )
+
+        def path_match(self, other):
+            """Boolean expression. Test if JSONPath predicate given by the
+            argument JSONPath expression matches
+            (equivalent of the ``@@`` operator).
+
+            Only the first item of the result is taken into account.
+
+            .. versionadded:: 2.0
+            """
+            return self.operate(
+                PATH_MATCH, other, result_type=sqltypes.Boolean
+            )
+
+    comparator_factory = Comparator
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/named_types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/named_types.py
new file mode 100644
index 00000000..e1b8e84c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/named_types.py
@@ -0,0 +1,505 @@
+# dialects/postgresql/named_types.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+from typing import Any
+from typing import Optional
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+
+from ... import schema
+from ... import util
+from ...sql import coercions
+from ...sql import elements
+from ...sql import roles
+from ...sql import sqltypes
+from ...sql import type_api
+from ...sql.base import _NoArg
+from ...sql.ddl import InvokeCreateDDLBase
+from ...sql.ddl import InvokeDropDDLBase
+
+if TYPE_CHECKING:
+    from ...sql._typing import _TypeEngineArgument
+
+
+class NamedType(sqltypes.TypeEngine):
+    """Base for named types."""
+
+    __abstract__ = True
+    DDLGenerator: Type[NamedTypeGenerator]
+    DDLDropper: Type[NamedTypeDropper]
+    create_type: bool
+
+    def create(self, bind, checkfirst=True, **kw):
+        """Emit ``CREATE`` DDL for this type.
+
+        :param bind: a connectable :class:`_engine.Engine`,
+         :class:`_engine.Connection`, or similar object to emit
+         SQL.
+        :param checkfirst: if ``True``, a query against
+         the PG catalog will be first performed to see
+         if the type does not exist already before
+         creating.
+
+        """
+        bind._run_ddl_visitor(self.DDLGenerator, self, checkfirst=checkfirst)
+
+    def drop(self, bind, checkfirst=True, **kw):
+        """Emit ``DROP`` DDL for this type.
+
+        :param bind: a connectable :class:`_engine.Engine`,
+         :class:`_engine.Connection`, or similar object to emit
+         SQL.
+        :param checkfirst: if ``True``, a query against
+         the PG catalog will be first performed to see
+         if the type actually exists before dropping.
+
+        """
+        bind._run_ddl_visitor(self.DDLDropper, self, checkfirst=checkfirst)
+
+    def _check_for_name_in_memos(self, checkfirst, kw):
+        """Look in the 'ddl runner' for 'memos', then
+        note our name in that collection.
+
+        This to ensure a particular named type is operated
+        upon only once within any kind of create/drop
+        sequence without relying upon "checkfirst".
+
+        """
+        if not self.create_type:
+            return True
+        if "_ddl_runner" in kw:
+            ddl_runner = kw["_ddl_runner"]
+            type_name = f"pg_{self.__visit_name__}"
+            if type_name in ddl_runner.memo:
+                existing = ddl_runner.memo[type_name]
+            else:
+                existing = ddl_runner.memo[type_name] = set()
+            present = (self.schema, self.name) in existing
+            existing.add((self.schema, self.name))
+            return present
+        else:
+            return False
+
+    def _on_table_create(self, target, bind, checkfirst=False, **kw):
+        if (
+            checkfirst
+            or (
+                not self.metadata
+                and not kw.get("_is_metadata_operation", False)
+            )
+        ) and not self._check_for_name_in_memos(checkfirst, kw):
+            self.create(bind=bind, checkfirst=checkfirst)
+
+    def _on_table_drop(self, target, bind, checkfirst=False, **kw):
+        if (
+            not self.metadata
+            and not kw.get("_is_metadata_operation", False)
+            and not self._check_for_name_in_memos(checkfirst, kw)
+        ):
+            self.drop(bind=bind, checkfirst=checkfirst)
+
+    def _on_metadata_create(self, target, bind, checkfirst=False, **kw):
+        if not self._check_for_name_in_memos(checkfirst, kw):
+            self.create(bind=bind, checkfirst=checkfirst)
+
+    def _on_metadata_drop(self, target, bind, checkfirst=False, **kw):
+        if not self._check_for_name_in_memos(checkfirst, kw):
+            self.drop(bind=bind, checkfirst=checkfirst)
+
+
+class NamedTypeGenerator(InvokeCreateDDLBase):
+    def __init__(self, dialect, connection, checkfirst=False, **kwargs):
+        super().__init__(connection, **kwargs)
+        self.checkfirst = checkfirst
+
+    def _can_create_type(self, type_):
+        if not self.checkfirst:
+            return True
+
+        effective_schema = self.connection.schema_for_object(type_)
+        return not self.connection.dialect.has_type(
+            self.connection, type_.name, schema=effective_schema
+        )
+
+
+class NamedTypeDropper(InvokeDropDDLBase):
+    def __init__(self, dialect, connection, checkfirst=False, **kwargs):
+        super().__init__(connection, **kwargs)
+        self.checkfirst = checkfirst
+
+    def _can_drop_type(self, type_):
+        if not self.checkfirst:
+            return True
+
+        effective_schema = self.connection.schema_for_object(type_)
+        return self.connection.dialect.has_type(
+            self.connection, type_.name, schema=effective_schema
+        )
+
+
+class EnumGenerator(NamedTypeGenerator):
+    def visit_enum(self, enum):
+        if not self._can_create_type(enum):
+            return
+
+        with self.with_ddl_events(enum):
+            self.connection.execute(CreateEnumType(enum))
+
+
+class EnumDropper(NamedTypeDropper):
+    def visit_enum(self, enum):
+        if not self._can_drop_type(enum):
+            return
+
+        with self.with_ddl_events(enum):
+            self.connection.execute(DropEnumType(enum))
+
+
+class ENUM(NamedType, type_api.NativeForEmulated, sqltypes.Enum):
+    """PostgreSQL ENUM type.
+
+    This is a subclass of :class:`_types.Enum` which includes
+    support for PG's ``CREATE TYPE`` and ``DROP TYPE``.
+
+    When the builtin type :class:`_types.Enum` is used and the
+    :paramref:`.Enum.native_enum` flag is left at its default of
+    True, the PostgreSQL backend will use a :class:`_postgresql.ENUM`
+    type as the implementation, so the special create/drop rules
+    will be used.
+
+    The create/drop behavior of ENUM is necessarily intricate, due to the
+    awkward relationship the ENUM type has in relationship to the
+    parent table, in that it may be "owned" by just a single table, or
+    may be shared among many tables.
+
+    When using :class:`_types.Enum` or :class:`_postgresql.ENUM`
+    in an "inline" fashion, the ``CREATE TYPE`` and ``DROP TYPE`` is emitted
+    corresponding to when the :meth:`_schema.Table.create` and
+    :meth:`_schema.Table.drop`
+    methods are called::
+
+        table = Table(
+            "sometable",
+            metadata,
+            Column("some_enum", ENUM("a", "b", "c", name="myenum")),
+        )
+
+        table.create(engine)  # will emit CREATE ENUM and CREATE TABLE
+        table.drop(engine)  # will emit DROP TABLE and DROP ENUM
+
+    To use a common enumerated type between multiple tables, the best
+    practice is to declare the :class:`_types.Enum` or
+    :class:`_postgresql.ENUM` independently, and associate it with the
+    :class:`_schema.MetaData` object itself::
+
+        my_enum = ENUM("a", "b", "c", name="myenum", metadata=metadata)
+
+        t1 = Table("sometable_one", metadata, Column("some_enum", myenum))
+
+        t2 = Table("sometable_two", metadata, Column("some_enum", myenum))
+
+    When this pattern is used, care must still be taken at the level
+    of individual table creates.  Emitting CREATE TABLE without also
+    specifying ``checkfirst=True`` will still cause issues::
+
+        t1.create(engine)  # will fail: no such type 'myenum'
+
+    If we specify ``checkfirst=True``, the individual table-level create
+    operation will check for the ``ENUM`` and create if not exists::
+
+        # will check if enum exists, and emit CREATE TYPE if not
+        t1.create(engine, checkfirst=True)
+
+    When using a metadata-level ENUM type, the type will always be created
+    and dropped if either the metadata-wide create/drop is called::
+
+        metadata.create_all(engine)  # will emit CREATE TYPE
+        metadata.drop_all(engine)  # will emit DROP TYPE
+
+    The type can also be created and dropped directly::
+
+        my_enum.create(engine)
+        my_enum.drop(engine)
+
+    """
+
+    native_enum = True
+    DDLGenerator = EnumGenerator
+    DDLDropper = EnumDropper
+
+    def __init__(
+        self,
+        *enums,
+        name: Union[str, _NoArg, None] = _NoArg.NO_ARG,
+        create_type: bool = True,
+        **kw,
+    ):
+        """Construct an :class:`_postgresql.ENUM`.
+
+        Arguments are the same as that of
+        :class:`_types.Enum`, but also including
+        the following parameters.
+
+        :param create_type: Defaults to True.
+         Indicates that ``CREATE TYPE`` should be
+         emitted, after optionally checking for the
+         presence of the type, when the parent
+         table is being created; and additionally
+         that ``DROP TYPE`` is called when the table
+         is dropped.    When ``False``, no check
+         will be performed and no ``CREATE TYPE``
+         or ``DROP TYPE`` is emitted, unless
+         :meth:`~.postgresql.ENUM.create`
+         or :meth:`~.postgresql.ENUM.drop`
+         are called directly.
+         Setting to ``False`` is helpful
+         when invoking a creation scheme to a SQL file
+         without access to the actual database -
+         the :meth:`~.postgresql.ENUM.create` and
+         :meth:`~.postgresql.ENUM.drop` methods can
+         be used to emit SQL to a target bind.
+
+        """
+        native_enum = kw.pop("native_enum", None)
+        if native_enum is False:
+            util.warn(
+                "the native_enum flag does not apply to the "
+                "sqlalchemy.dialects.postgresql.ENUM datatype; this type "
+                "always refers to ENUM.   Use sqlalchemy.types.Enum for "
+                "non-native enum."
+            )
+        self.create_type = create_type
+        if name is not _NoArg.NO_ARG:
+            kw["name"] = name
+        super().__init__(*enums, **kw)
+
+    def coerce_compared_value(self, op, value):
+        super_coerced_type = super().coerce_compared_value(op, value)
+        if (
+            super_coerced_type._type_affinity
+            is type_api.STRINGTYPE._type_affinity
+        ):
+            return self
+        else:
+            return super_coerced_type
+
+    @classmethod
+    def __test_init__(cls):
+        return cls(name="name")
+
+    @classmethod
+    def adapt_emulated_to_native(cls, impl, **kw):
+        """Produce a PostgreSQL native :class:`_postgresql.ENUM` from plain
+        :class:`.Enum`.
+
+        """
+        kw.setdefault("validate_strings", impl.validate_strings)
+        kw.setdefault("name", impl.name)
+        kw.setdefault("schema", impl.schema)
+        kw.setdefault("inherit_schema", impl.inherit_schema)
+        kw.setdefault("metadata", impl.metadata)
+        kw.setdefault("_create_events", False)
+        kw.setdefault("values_callable", impl.values_callable)
+        kw.setdefault("omit_aliases", impl._omit_aliases)
+        kw.setdefault("_adapted_from", impl)
+        if type_api._is_native_for_emulated(impl.__class__):
+            kw.setdefault("create_type", impl.create_type)
+
+        return cls(**kw)
+
+    def create(self, bind=None, checkfirst=True):
+        """Emit ``CREATE TYPE`` for this
+        :class:`_postgresql.ENUM`.
+
+        If the underlying dialect does not support
+        PostgreSQL CREATE TYPE, no action is taken.
+
+        :param bind: a connectable :class:`_engine.Engine`,
+         :class:`_engine.Connection`, or similar object to emit
+         SQL.
+        :param checkfirst: if ``True``, a query against
+         the PG catalog will be first performed to see
+         if the type does not exist already before
+         creating.
+
+        """
+        if not bind.dialect.supports_native_enum:
+            return
+
+        super().create(bind, checkfirst=checkfirst)
+
+    def drop(self, bind=None, checkfirst=True):
+        """Emit ``DROP TYPE`` for this
+        :class:`_postgresql.ENUM`.
+
+        If the underlying dialect does not support
+        PostgreSQL DROP TYPE, no action is taken.
+
+        :param bind: a connectable :class:`_engine.Engine`,
+         :class:`_engine.Connection`, or similar object to emit
+         SQL.
+        :param checkfirst: if ``True``, a query against
+         the PG catalog will be first performed to see
+         if the type actually exists before dropping.
+
+        """
+        if not bind.dialect.supports_native_enum:
+            return
+
+        super().drop(bind, checkfirst=checkfirst)
+
+    def get_dbapi_type(self, dbapi):
+        """dont return dbapi.STRING for ENUM in PostgreSQL, since that's
+        a different type"""
+
+        return None
+
+
+class DomainGenerator(NamedTypeGenerator):
+    def visit_DOMAIN(self, domain):
+        if not self._can_create_type(domain):
+            return
+        with self.with_ddl_events(domain):
+            self.connection.execute(CreateDomainType(domain))
+
+
+class DomainDropper(NamedTypeDropper):
+    def visit_DOMAIN(self, domain):
+        if not self._can_drop_type(domain):
+            return
+
+        with self.with_ddl_events(domain):
+            self.connection.execute(DropDomainType(domain))
+
+
+class DOMAIN(NamedType, sqltypes.SchemaType):
+    r"""Represent the DOMAIN PostgreSQL type.
+
+    A domain is essentially a data type with optional constraints
+    that restrict the allowed set of values. E.g.::
+
+        PositiveInt = DOMAIN("pos_int", Integer, check="VALUE > 0", not_null=True)
+
+        UsPostalCode = DOMAIN(
+            "us_postal_code",
+            Text,
+            check="VALUE ~ '^\d{5}$' OR VALUE ~ '^\d{5}-\d{4}$'",
+        )
+
+    See the `PostgreSQL documentation`__ for additional details
+
+    __ https://www.postgresql.org/docs/current/sql-createdomain.html
+
+    .. versionadded:: 2.0
+
+    """  # noqa: E501
+
+    DDLGenerator = DomainGenerator
+    DDLDropper = DomainDropper
+
+    __visit_name__ = "DOMAIN"
+
+    def __init__(
+        self,
+        name: str,
+        data_type: _TypeEngineArgument[Any],
+        *,
+        collation: Optional[str] = None,
+        default: Union[elements.TextClause, str, None] = None,
+        constraint_name: Optional[str] = None,
+        not_null: Optional[bool] = None,
+        check: Union[elements.TextClause, str, None] = None,
+        create_type: bool = True,
+        **kw: Any,
+    ):
+        """
+        Construct a DOMAIN.
+
+        :param name: the name of the domain
+        :param data_type: The underlying data type of the domain.
+          This can include array specifiers.
+        :param collation: An optional collation for the domain.
+          If no collation is specified, the underlying data type's default
+          collation is used. The underlying type must be collatable if
+          ``collation`` is specified.
+        :param default: The DEFAULT clause specifies a default value for
+          columns of the domain data type. The default should be a string
+          or a :func:`_expression.text` value.
+          If no default value is specified, then the default value is
+          the null value.
+        :param constraint_name: An optional name for a constraint.
+          If not specified, the backend generates a name.
+        :param not_null: Values of this domain are prevented from being null.
+          By default domain are allowed to be null. If not specified
+          no nullability clause will be emitted.
+        :param check: CHECK clause specify integrity constraint or test
+          which values of the domain must satisfy. A constraint must be
+          an expression producing a Boolean result that can use the key
+          word VALUE to refer to the value being tested.
+          Differently from PostgreSQL, only a single check clause is
+          currently allowed in SQLAlchemy.
+        :param schema: optional schema name
+        :param metadata: optional :class:`_schema.MetaData` object which
+         this :class:`_postgresql.DOMAIN` will be directly associated
+        :param create_type: Defaults to True.
+         Indicates that ``CREATE TYPE`` should be emitted, after optionally
+         checking for the presence of the type, when the parent table is
+         being created; and additionally that ``DROP TYPE`` is called
+         when the table is dropped.
+
+        """
+        self.data_type = type_api.to_instance(data_type)
+        self.default = default
+        self.collation = collation
+        self.constraint_name = constraint_name
+        self.not_null = bool(not_null)
+        if check is not None:
+            check = coercions.expect(roles.DDLExpressionRole, check)
+        self.check = check
+        self.create_type = create_type
+        super().__init__(name=name, **kw)
+
+    @classmethod
+    def __test_init__(cls):
+        return cls("name", sqltypes.Integer)
+
+    def adapt(self, impl, **kw):
+        if self.default:
+            kw["default"] = self.default
+        if self.constraint_name is not None:
+            kw["constraint_name"] = self.constraint_name
+        if self.not_null:
+            kw["not_null"] = self.not_null
+        if self.check is not None:
+            kw["check"] = str(self.check)
+        if self.create_type:
+            kw["create_type"] = self.create_type
+
+        return super().adapt(impl, **kw)
+
+
+class CreateEnumType(schema._CreateDropBase):
+    __visit_name__ = "create_enum_type"
+
+
+class DropEnumType(schema._CreateDropBase):
+    __visit_name__ = "drop_enum_type"
+
+
+class CreateDomainType(schema._CreateDropBase):
+    """Represent a CREATE DOMAIN statement."""
+
+    __visit_name__ = "create_domain_type"
+
+
+class DropDomainType(schema._CreateDropBase):
+    """Represent a DROP DOMAIN statement."""
+
+    __visit_name__ = "drop_domain_type"
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/operators.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/operators.py
new file mode 100644
index 00000000..ebcafcba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/operators.py
@@ -0,0 +1,129 @@
+# dialects/postgresql/operators.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from ...sql import operators
+
+
+_getitem_precedence = operators._PRECEDENCE[operators.json_getitem_op]
+_eq_precedence = operators._PRECEDENCE[operators.eq]
+
+# JSON + JSONB
+ASTEXT = operators.custom_op(
+    "->>",
+    precedence=_getitem_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+)
+
+JSONPATH_ASTEXT = operators.custom_op(
+    "#>>",
+    precedence=_getitem_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+)
+
+# JSONB + HSTORE
+HAS_KEY = operators.custom_op(
+    "?",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+HAS_ALL = operators.custom_op(
+    "?&",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+HAS_ANY = operators.custom_op(
+    "?|",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+# JSONB
+DELETE_PATH = operators.custom_op(
+    "#-",
+    precedence=_getitem_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+)
+
+PATH_EXISTS = operators.custom_op(
+    "@?",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+PATH_MATCH = operators.custom_op(
+    "@@",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+# JSONB + ARRAY + HSTORE + RANGE
+CONTAINS = operators.custom_op(
+    "@>",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+CONTAINED_BY = operators.custom_op(
+    "<@",
+    precedence=_eq_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+    is_comparison=True,
+)
+
+# ARRAY + RANGE
+OVERLAP = operators.custom_op(
+    "&&",
+    precedence=_eq_precedence,
+    is_comparison=True,
+)
+
+# RANGE
+STRICTLY_LEFT_OF = operators.custom_op(
+    "<<", precedence=_eq_precedence, is_comparison=True
+)
+
+STRICTLY_RIGHT_OF = operators.custom_op(
+    ">>", precedence=_eq_precedence, is_comparison=True
+)
+
+NOT_EXTEND_RIGHT_OF = operators.custom_op(
+    "&<", precedence=_eq_precedence, is_comparison=True
+)
+
+NOT_EXTEND_LEFT_OF = operators.custom_op(
+    "&>", precedence=_eq_precedence, is_comparison=True
+)
+
+ADJACENT_TO = operators.custom_op(
+    "-|-", precedence=_eq_precedence, is_comparison=True
+)
+
+# HSTORE
+GETITEM = operators.custom_op(
+    "->",
+    precedence=_getitem_precedence,
+    natural_self_precedent=True,
+    eager_grouping=True,
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg8000.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg8000.py
new file mode 100644
index 00000000..bf113230
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg8000.py
@@ -0,0 +1,666 @@
+# dialects/postgresql/pg8000.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors <see AUTHORS
+# file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql+pg8000
+    :name: pg8000
+    :dbapi: pg8000
+    :connectstring: postgresql+pg8000://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://pypi.org/project/pg8000/
+
+.. versionchanged:: 1.4  The pg8000 dialect has been updated for version
+   1.16.6 and higher, and is again part of SQLAlchemy's continuous integration
+   with full feature support.
+
+.. _pg8000_unicode:
+
+Unicode
+-------
+
+pg8000 will encode / decode string values between it and the server using the
+PostgreSQL ``client_encoding`` parameter; by default this is the value in
+the ``postgresql.conf`` file, which often defaults to ``SQL_ASCII``.
+Typically, this can be changed to ``utf-8``, as a more useful default::
+
+    # client_encoding = sql_ascii # actually, defaults to database encoding
+    client_encoding = utf8
+
+The ``client_encoding`` can be overridden for a session by executing the SQL:
+
+.. sourcecode:: sql
+
+    SET CLIENT_ENCODING TO 'utf8';
+
+SQLAlchemy will execute this SQL on all new connections based on the value
+passed to :func:`_sa.create_engine` using the ``client_encoding`` parameter::
+
+    engine = create_engine(
+        "postgresql+pg8000://user:pass@host/dbname", client_encoding="utf8"
+    )
+
+.. _pg8000_ssl:
+
+SSL Connections
+---------------
+
+pg8000 accepts a Python ``SSLContext`` object which may be specified using the
+:paramref:`_sa.create_engine.connect_args` dictionary::
+
+    import ssl
+
+    ssl_context = ssl.create_default_context()
+    engine = sa.create_engine(
+        "postgresql+pg8000://scott:tiger@192.168.0.199/test",
+        connect_args={"ssl_context": ssl_context},
+    )
+
+If the server uses an automatically-generated certificate that is self-signed
+or does not match the host name (as seen from the client), it may also be
+necessary to disable hostname checking::
+
+    import ssl
+
+    ssl_context = ssl.create_default_context()
+    ssl_context.check_hostname = False
+    ssl_context.verify_mode = ssl.CERT_NONE
+    engine = sa.create_engine(
+        "postgresql+pg8000://scott:tiger@192.168.0.199/test",
+        connect_args={"ssl_context": ssl_context},
+    )
+
+.. _pg8000_isolation_level:
+
+pg8000 Transaction Isolation Level
+-------------------------------------
+
+The pg8000 dialect offers the same isolation level settings as that
+of the :ref:`psycopg2 <psycopg2_isolation_level>` dialect:
+
+* ``READ COMMITTED``
+* ``READ UNCOMMITTED``
+* ``REPEATABLE READ``
+* ``SERIALIZABLE``
+* ``AUTOCOMMIT``
+
+.. seealso::
+
+    :ref:`postgresql_isolation_level`
+
+    :ref:`psycopg2_isolation_level`
+
+
+"""  # noqa
+import decimal
+import re
+
+from . import ranges
+from .array import ARRAY as PGARRAY
+from .base import _DECIMAL_TYPES
+from .base import _FLOAT_TYPES
+from .base import _INT_TYPES
+from .base import ENUM
+from .base import INTERVAL
+from .base import PGCompiler
+from .base import PGDialect
+from .base import PGExecutionContext
+from .base import PGIdentifierPreparer
+from .json import JSON
+from .json import JSONB
+from .json import JSONPathType
+from .pg_catalog import _SpaceVector
+from .pg_catalog import OIDVECTOR
+from .types import CITEXT
+from ... import exc
+from ... import util
+from ...engine import processors
+from ...sql import sqltypes
+from ...sql.elements import quoted_name
+
+
+class _PGString(sqltypes.String):
+    render_bind_cast = True
+
+
+class _PGNumeric(sqltypes.Numeric):
+    render_bind_cast = True
+
+    def result_processor(self, dialect, coltype):
+        if self.asdecimal:
+            if coltype in _FLOAT_TYPES:
+                return processors.to_decimal_processor_factory(
+                    decimal.Decimal, self._effective_decimal_return_scale
+                )
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                # pg8000 returns Decimal natively for 1700
+                return None
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+        else:
+            if coltype in _FLOAT_TYPES:
+                # pg8000 returns float natively for 701
+                return None
+            elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES:
+                return processors.to_float
+            else:
+                raise exc.InvalidRequestError(
+                    "Unknown PG numeric type: %d" % coltype
+                )
+
+
+class _PGFloat(_PGNumeric, sqltypes.Float):
+    __visit_name__ = "float"
+    render_bind_cast = True
+
+
+class _PGNumericNoBind(_PGNumeric):
+    def bind_processor(self, dialect):
+        return None
+
+
+class _PGJSON(JSON):
+    render_bind_cast = True
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _PGJSONB(JSONB):
+    render_bind_cast = True
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _PGJSONIndexType(sqltypes.JSON.JSONIndexType):
+    def get_dbapi_type(self, dbapi):
+        raise NotImplementedError("should not be here")
+
+
+class _PGJSONIntIndexType(sqltypes.JSON.JSONIntIndexType):
+    __visit_name__ = "json_int_index"
+
+    render_bind_cast = True
+
+
+class _PGJSONStrIndexType(sqltypes.JSON.JSONStrIndexType):
+    __visit_name__ = "json_str_index"
+
+    render_bind_cast = True
+
+
+class _PGJSONPathType(JSONPathType):
+    pass
+
+    # DBAPI type 1009
+
+
+class _PGEnum(ENUM):
+    def get_dbapi_type(self, dbapi):
+        return dbapi.UNKNOWN
+
+
+class _PGInterval(INTERVAL):
+    render_bind_cast = True
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.INTERVAL
+
+    @classmethod
+    def adapt_emulated_to_native(cls, interval, **kw):
+        return _PGInterval(precision=interval.second_precision)
+
+
+class _PGTimeStamp(sqltypes.DateTime):
+    render_bind_cast = True
+
+
+class _PGDate(sqltypes.Date):
+    render_bind_cast = True
+
+
+class _PGTime(sqltypes.Time):
+    render_bind_cast = True
+
+
+class _PGInteger(sqltypes.Integer):
+    render_bind_cast = True
+
+
+class _PGSmallInteger(sqltypes.SmallInteger):
+    render_bind_cast = True
+
+
+class _PGNullType(sqltypes.NullType):
+    pass
+
+
+class _PGBigInteger(sqltypes.BigInteger):
+    render_bind_cast = True
+
+
+class _PGBoolean(sqltypes.Boolean):
+    render_bind_cast = True
+
+
+class _PGARRAY(PGARRAY):
+    render_bind_cast = True
+
+
+class _PGOIDVECTOR(_SpaceVector, OIDVECTOR):
+    pass
+
+
+class _Pg8000Range(ranges.AbstractSingleRangeImpl):
+    def bind_processor(self, dialect):
+        pg8000_Range = dialect.dbapi.Range
+
+        def to_range(value):
+            if isinstance(value, ranges.Range):
+                value = pg8000_Range(
+                    value.lower, value.upper, value.bounds, value.empty
+                )
+            return value
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range(value):
+            if value is not None:
+                value = ranges.Range(
+                    value.lower,
+                    value.upper,
+                    bounds=value.bounds,
+                    empty=value.is_empty,
+                )
+            return value
+
+        return to_range
+
+
+class _Pg8000MultiRange(ranges.AbstractMultiRangeImpl):
+    def bind_processor(self, dialect):
+        pg8000_Range = dialect.dbapi.Range
+
+        def to_multirange(value):
+            if isinstance(value, list):
+                mr = []
+                for v in value:
+                    if isinstance(v, ranges.Range):
+                        mr.append(
+                            pg8000_Range(v.lower, v.upper, v.bounds, v.empty)
+                        )
+                    else:
+                        mr.append(v)
+                return mr
+            else:
+                return value
+
+        return to_multirange
+
+    def result_processor(self, dialect, coltype):
+        def to_multirange(value):
+            if value is None:
+                return None
+            else:
+                return ranges.MultiRange(
+                    ranges.Range(
+                        v.lower, v.upper, bounds=v.bounds, empty=v.is_empty
+                    )
+                    for v in value
+                )
+
+        return to_multirange
+
+
+_server_side_id = util.counter()
+
+
+class PGExecutionContext_pg8000(PGExecutionContext):
+    def create_server_side_cursor(self):
+        ident = "c_%s_%s" % (hex(id(self))[2:], hex(_server_side_id())[2:])
+        return ServerSideCursor(self._dbapi_connection.cursor(), ident)
+
+    def pre_exec(self):
+        if not self.compiled:
+            return
+
+
+class ServerSideCursor:
+    server_side = True
+
+    def __init__(self, cursor, ident):
+        self.ident = ident
+        self.cursor = cursor
+
+    @property
+    def connection(self):
+        return self.cursor.connection
+
+    @property
+    def rowcount(self):
+        return self.cursor.rowcount
+
+    @property
+    def description(self):
+        return self.cursor.description
+
+    def execute(self, operation, args=(), stream=None):
+        op = "DECLARE " + self.ident + " NO SCROLL CURSOR FOR " + operation
+        self.cursor.execute(op, args, stream=stream)
+        return self
+
+    def executemany(self, operation, param_sets):
+        self.cursor.executemany(operation, param_sets)
+        return self
+
+    def fetchone(self):
+        self.cursor.execute("FETCH FORWARD 1 FROM " + self.ident)
+        return self.cursor.fetchone()
+
+    def fetchmany(self, num=None):
+        if num is None:
+            return self.fetchall()
+        else:
+            self.cursor.execute(
+                "FETCH FORWARD " + str(int(num)) + " FROM " + self.ident
+            )
+            return self.cursor.fetchall()
+
+    def fetchall(self):
+        self.cursor.execute("FETCH FORWARD ALL FROM " + self.ident)
+        return self.cursor.fetchall()
+
+    def close(self):
+        self.cursor.execute("CLOSE " + self.ident)
+        self.cursor.close()
+
+    def setinputsizes(self, *sizes):
+        self.cursor.setinputsizes(*sizes)
+
+    def setoutputsize(self, size, column=None):
+        pass
+
+
+class PGCompiler_pg8000(PGCompiler):
+    def visit_mod_binary(self, binary, operator, **kw):
+        return (
+            self.process(binary.left, **kw)
+            + " %% "
+            + self.process(binary.right, **kw)
+        )
+
+
+class PGIdentifierPreparer_pg8000(PGIdentifierPreparer):
+    def __init__(self, *args, **kwargs):
+        PGIdentifierPreparer.__init__(self, *args, **kwargs)
+        self._double_percents = False
+
+
+class PGDialect_pg8000(PGDialect):
+    driver = "pg8000"
+    supports_statement_cache = True
+
+    supports_unicode_statements = True
+
+    supports_unicode_binds = True
+
+    default_paramstyle = "format"
+    supports_sane_multi_rowcount = True
+    execution_ctx_cls = PGExecutionContext_pg8000
+    statement_compiler = PGCompiler_pg8000
+    preparer = PGIdentifierPreparer_pg8000
+    supports_server_side_cursors = True
+
+    render_bind_cast = True
+
+    # reversed as of pg8000 1.16.6.  1.16.5 and lower
+    # are no longer compatible
+    description_encoding = None
+    # description_encoding = "use_encoding"
+
+    colspecs = util.update_copy(
+        PGDialect.colspecs,
+        {
+            sqltypes.String: _PGString,
+            sqltypes.Numeric: _PGNumericNoBind,
+            sqltypes.Float: _PGFloat,
+            sqltypes.JSON: _PGJSON,
+            sqltypes.Boolean: _PGBoolean,
+            sqltypes.NullType: _PGNullType,
+            JSONB: _PGJSONB,
+            CITEXT: CITEXT,
+            sqltypes.JSON.JSONPathType: _PGJSONPathType,
+            sqltypes.JSON.JSONIndexType: _PGJSONIndexType,
+            sqltypes.JSON.JSONIntIndexType: _PGJSONIntIndexType,
+            sqltypes.JSON.JSONStrIndexType: _PGJSONStrIndexType,
+            sqltypes.Interval: _PGInterval,
+            INTERVAL: _PGInterval,
+            sqltypes.DateTime: _PGTimeStamp,
+            sqltypes.DateTime: _PGTimeStamp,
+            sqltypes.Date: _PGDate,
+            sqltypes.Time: _PGTime,
+            sqltypes.Integer: _PGInteger,
+            sqltypes.SmallInteger: _PGSmallInteger,
+            sqltypes.BigInteger: _PGBigInteger,
+            sqltypes.Enum: _PGEnum,
+            sqltypes.ARRAY: _PGARRAY,
+            OIDVECTOR: _PGOIDVECTOR,
+            ranges.INT4RANGE: _Pg8000Range,
+            ranges.INT8RANGE: _Pg8000Range,
+            ranges.NUMRANGE: _Pg8000Range,
+            ranges.DATERANGE: _Pg8000Range,
+            ranges.TSRANGE: _Pg8000Range,
+            ranges.TSTZRANGE: _Pg8000Range,
+            ranges.INT4MULTIRANGE: _Pg8000MultiRange,
+            ranges.INT8MULTIRANGE: _Pg8000MultiRange,
+            ranges.NUMMULTIRANGE: _Pg8000MultiRange,
+            ranges.DATEMULTIRANGE: _Pg8000MultiRange,
+            ranges.TSMULTIRANGE: _Pg8000MultiRange,
+            ranges.TSTZMULTIRANGE: _Pg8000MultiRange,
+        },
+    )
+
+    def __init__(self, client_encoding=None, **kwargs):
+        PGDialect.__init__(self, **kwargs)
+        self.client_encoding = client_encoding
+
+        if self._dbapi_version < (1, 16, 6):
+            raise NotImplementedError("pg8000 1.16.6 or greater is required")
+
+        if self._native_inet_types:
+            raise NotImplementedError(
+                "The pg8000 dialect does not fully implement "
+                "ipaddress type handling; INET is supported by default, "
+                "CIDR is not"
+            )
+
+    @util.memoized_property
+    def _dbapi_version(self):
+        if self.dbapi and hasattr(self.dbapi, "__version__"):
+            return tuple(
+                [
+                    int(x)
+                    for x in re.findall(
+                        r"(\d+)(?:[-\.]?|$)", self.dbapi.__version__
+                    )
+                ]
+            )
+        else:
+            return (99, 99, 99)
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("pg8000")
+
+    def create_connect_args(self, url):
+        opts = url.translate_connect_args(username="user")
+        if "port" in opts:
+            opts["port"] = int(opts["port"])
+        opts.update(url.query)
+        return ([], opts)
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(e, self.dbapi.InterfaceError) and "network error" in str(
+            e
+        ):
+            # new as of pg8000 1.19.0 for broken connections
+            return True
+
+        # connection was closed normally
+        return "connection is closed" in str(e)
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return (
+            "AUTOCOMMIT",
+            "READ COMMITTED",
+            "READ UNCOMMITTED",
+            "REPEATABLE READ",
+            "SERIALIZABLE",
+        )
+
+    def set_isolation_level(self, dbapi_connection, level):
+        level = level.replace("_", " ")
+
+        if level == "AUTOCOMMIT":
+            dbapi_connection.autocommit = True
+        else:
+            dbapi_connection.autocommit = False
+            cursor = dbapi_connection.cursor()
+            cursor.execute(
+                "SET SESSION CHARACTERISTICS AS TRANSACTION "
+                f"ISOLATION LEVEL {level}"
+            )
+            cursor.execute("COMMIT")
+            cursor.close()
+
+    def set_readonly(self, connection, value):
+        cursor = connection.cursor()
+        try:
+            cursor.execute(
+                "SET SESSION CHARACTERISTICS AS TRANSACTION %s"
+                % ("READ ONLY" if value else "READ WRITE")
+            )
+            cursor.execute("COMMIT")
+        finally:
+            cursor.close()
+
+    def get_readonly(self, connection):
+        cursor = connection.cursor()
+        try:
+            cursor.execute("show transaction_read_only")
+            val = cursor.fetchone()[0]
+        finally:
+            cursor.close()
+
+        return val == "on"
+
+    def set_deferrable(self, connection, value):
+        cursor = connection.cursor()
+        try:
+            cursor.execute(
+                "SET SESSION CHARACTERISTICS AS TRANSACTION %s"
+                % ("DEFERRABLE" if value else "NOT DEFERRABLE")
+            )
+            cursor.execute("COMMIT")
+        finally:
+            cursor.close()
+
+    def get_deferrable(self, connection):
+        cursor = connection.cursor()
+        try:
+            cursor.execute("show transaction_deferrable")
+            val = cursor.fetchone()[0]
+        finally:
+            cursor.close()
+
+        return val == "on"
+
+    def _set_client_encoding(self, dbapi_connection, client_encoding):
+        cursor = dbapi_connection.cursor()
+        cursor.execute(
+            f"""SET CLIENT_ENCODING TO '{
+                client_encoding.replace("'", "''")
+            }'"""
+        )
+        cursor.execute("COMMIT")
+        cursor.close()
+
+    def do_begin_twophase(self, connection, xid):
+        connection.connection.tpc_begin((0, xid, ""))
+
+    def do_prepare_twophase(self, connection, xid):
+        connection.connection.tpc_prepare()
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        connection.connection.tpc_rollback((0, xid, ""))
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        connection.connection.tpc_commit((0, xid, ""))
+
+    def do_recover_twophase(self, connection):
+        return [row[1] for row in connection.connection.tpc_recover()]
+
+    def on_connect(self):
+        fns = []
+
+        def on_connect(conn):
+            conn.py_types[quoted_name] = conn.py_types[str]
+
+        fns.append(on_connect)
+
+        if self.client_encoding is not None:
+
+            def on_connect(conn):
+                self._set_client_encoding(conn, self.client_encoding)
+
+            fns.append(on_connect)
+
+        if self._native_inet_types is False:
+
+            def on_connect(conn):
+                # inet
+                conn.register_in_adapter(869, lambda s: s)
+
+                # cidr
+                conn.register_in_adapter(650, lambda s: s)
+
+            fns.append(on_connect)
+
+        if self._json_deserializer:
+
+            def on_connect(conn):
+                # json
+                conn.register_in_adapter(114, self._json_deserializer)
+
+                # jsonb
+                conn.register_in_adapter(3802, self._json_deserializer)
+
+            fns.append(on_connect)
+
+        if len(fns) > 0:
+
+            def on_connect(conn):
+                for fn in fns:
+                    fn(conn)
+
+            return on_connect
+        else:
+            return None
+
+    @util.memoized_property
+    def _dialect_specific_select_one(self):
+        return ";"
+
+
+dialect = PGDialect_pg8000
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py
new file mode 100644
index 00000000..78f390a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py
@@ -0,0 +1,300 @@
+# dialects/postgresql/pg_catalog.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .array import ARRAY
+from .types import OID
+from .types import REGCLASS
+from ... import Column
+from ... import func
+from ... import MetaData
+from ... import Table
+from ...types import BigInteger
+from ...types import Boolean
+from ...types import CHAR
+from ...types import Float
+from ...types import Integer
+from ...types import SmallInteger
+from ...types import String
+from ...types import Text
+from ...types import TypeDecorator
+
+
+# types
+class NAME(TypeDecorator):
+    impl = String(64, collation="C")
+    cache_ok = True
+
+
+class PG_NODE_TREE(TypeDecorator):
+    impl = Text(collation="C")
+    cache_ok = True
+
+
+class INT2VECTOR(TypeDecorator):
+    impl = ARRAY(SmallInteger)
+    cache_ok = True
+
+
+class OIDVECTOR(TypeDecorator):
+    impl = ARRAY(OID)
+    cache_ok = True
+
+
+class _SpaceVector:
+    def result_processor(self, dialect, coltype):
+        def process(value):
+            if value is None:
+                return value
+            return [int(p) for p in value.split(" ")]
+
+        return process
+
+
+REGPROC = REGCLASS  # seems an alias
+
+# functions
+_pg_cat = func.pg_catalog
+quote_ident = _pg_cat.quote_ident
+pg_table_is_visible = _pg_cat.pg_table_is_visible
+pg_type_is_visible = _pg_cat.pg_type_is_visible
+pg_get_viewdef = _pg_cat.pg_get_viewdef
+pg_get_serial_sequence = _pg_cat.pg_get_serial_sequence
+format_type = _pg_cat.format_type
+pg_get_expr = _pg_cat.pg_get_expr
+pg_get_constraintdef = _pg_cat.pg_get_constraintdef
+pg_get_indexdef = _pg_cat.pg_get_indexdef
+
+# constants
+RELKINDS_TABLE_NO_FOREIGN = ("r", "p")
+RELKINDS_TABLE = RELKINDS_TABLE_NO_FOREIGN + ("f",)
+RELKINDS_VIEW = ("v",)
+RELKINDS_MAT_VIEW = ("m",)
+RELKINDS_ALL_TABLE_LIKE = RELKINDS_TABLE + RELKINDS_VIEW + RELKINDS_MAT_VIEW
+
+# tables
+pg_catalog_meta = MetaData(schema="pg_catalog")
+
+pg_namespace = Table(
+    "pg_namespace",
+    pg_catalog_meta,
+    Column("oid", OID),
+    Column("nspname", NAME),
+    Column("nspowner", OID),
+)
+
+pg_class = Table(
+    "pg_class",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("relname", NAME),
+    Column("relnamespace", OID),
+    Column("reltype", OID),
+    Column("reloftype", OID),
+    Column("relowner", OID),
+    Column("relam", OID),
+    Column("relfilenode", OID),
+    Column("reltablespace", OID),
+    Column("relpages", Integer),
+    Column("reltuples", Float),
+    Column("relallvisible", Integer, info={"server_version": (9, 2)}),
+    Column("reltoastrelid", OID),
+    Column("relhasindex", Boolean),
+    Column("relisshared", Boolean),
+    Column("relpersistence", CHAR, info={"server_version": (9, 1)}),
+    Column("relkind", CHAR),
+    Column("relnatts", SmallInteger),
+    Column("relchecks", SmallInteger),
+    Column("relhasrules", Boolean),
+    Column("relhastriggers", Boolean),
+    Column("relhassubclass", Boolean),
+    Column("relrowsecurity", Boolean),
+    Column("relforcerowsecurity", Boolean, info={"server_version": (9, 5)}),
+    Column("relispopulated", Boolean, info={"server_version": (9, 3)}),
+    Column("relreplident", CHAR, info={"server_version": (9, 4)}),
+    Column("relispartition", Boolean, info={"server_version": (10,)}),
+    Column("relrewrite", OID, info={"server_version": (11,)}),
+    Column("reloptions", ARRAY(Text)),
+)
+
+pg_type = Table(
+    "pg_type",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("typname", NAME),
+    Column("typnamespace", OID),
+    Column("typowner", OID),
+    Column("typlen", SmallInteger),
+    Column("typbyval", Boolean),
+    Column("typtype", CHAR),
+    Column("typcategory", CHAR),
+    Column("typispreferred", Boolean),
+    Column("typisdefined", Boolean),
+    Column("typdelim", CHAR),
+    Column("typrelid", OID),
+    Column("typelem", OID),
+    Column("typarray", OID),
+    Column("typinput", REGPROC),
+    Column("typoutput", REGPROC),
+    Column("typreceive", REGPROC),
+    Column("typsend", REGPROC),
+    Column("typmodin", REGPROC),
+    Column("typmodout", REGPROC),
+    Column("typanalyze", REGPROC),
+    Column("typalign", CHAR),
+    Column("typstorage", CHAR),
+    Column("typnotnull", Boolean),
+    Column("typbasetype", OID),
+    Column("typtypmod", Integer),
+    Column("typndims", Integer),
+    Column("typcollation", OID, info={"server_version": (9, 1)}),
+    Column("typdefault", Text),
+)
+
+pg_index = Table(
+    "pg_index",
+    pg_catalog_meta,
+    Column("indexrelid", OID),
+    Column("indrelid", OID),
+    Column("indnatts", SmallInteger),
+    Column("indnkeyatts", SmallInteger, info={"server_version": (11,)}),
+    Column("indisunique", Boolean),
+    Column("indnullsnotdistinct", Boolean, info={"server_version": (15,)}),
+    Column("indisprimary", Boolean),
+    Column("indisexclusion", Boolean, info={"server_version": (9, 1)}),
+    Column("indimmediate", Boolean),
+    Column("indisclustered", Boolean),
+    Column("indisvalid", Boolean),
+    Column("indcheckxmin", Boolean),
+    Column("indisready", Boolean),
+    Column("indislive", Boolean, info={"server_version": (9, 3)}),  # 9.3
+    Column("indisreplident", Boolean),
+    Column("indkey", INT2VECTOR),
+    Column("indcollation", OIDVECTOR, info={"server_version": (9, 1)}),  # 9.1
+    Column("indclass", OIDVECTOR),
+    Column("indoption", INT2VECTOR),
+    Column("indexprs", PG_NODE_TREE),
+    Column("indpred", PG_NODE_TREE),
+)
+
+pg_attribute = Table(
+    "pg_attribute",
+    pg_catalog_meta,
+    Column("attrelid", OID),
+    Column("attname", NAME),
+    Column("atttypid", OID),
+    Column("attstattarget", Integer),
+    Column("attlen", SmallInteger),
+    Column("attnum", SmallInteger),
+    Column("attndims", Integer),
+    Column("attcacheoff", Integer),
+    Column("atttypmod", Integer),
+    Column("attbyval", Boolean),
+    Column("attstorage", CHAR),
+    Column("attalign", CHAR),
+    Column("attnotnull", Boolean),
+    Column("atthasdef", Boolean),
+    Column("atthasmissing", Boolean, info={"server_version": (11,)}),
+    Column("attidentity", CHAR, info={"server_version": (10,)}),
+    Column("attgenerated", CHAR, info={"server_version": (12,)}),
+    Column("attisdropped", Boolean),
+    Column("attislocal", Boolean),
+    Column("attinhcount", Integer),
+    Column("attcollation", OID, info={"server_version": (9, 1)}),
+)
+
+pg_constraint = Table(
+    "pg_constraint",
+    pg_catalog_meta,
+    Column("oid", OID),  # 9.3
+    Column("conname", NAME),
+    Column("connamespace", OID),
+    Column("contype", CHAR),
+    Column("condeferrable", Boolean),
+    Column("condeferred", Boolean),
+    Column("convalidated", Boolean, info={"server_version": (9, 1)}),
+    Column("conrelid", OID),
+    Column("contypid", OID),
+    Column("conindid", OID),
+    Column("conparentid", OID, info={"server_version": (11,)}),
+    Column("confrelid", OID),
+    Column("confupdtype", CHAR),
+    Column("confdeltype", CHAR),
+    Column("confmatchtype", CHAR),
+    Column("conislocal", Boolean),
+    Column("coninhcount", Integer),
+    Column("connoinherit", Boolean, info={"server_version": (9, 2)}),
+    Column("conkey", ARRAY(SmallInteger)),
+    Column("confkey", ARRAY(SmallInteger)),
+)
+
+pg_sequence = Table(
+    "pg_sequence",
+    pg_catalog_meta,
+    Column("seqrelid", OID),
+    Column("seqtypid", OID),
+    Column("seqstart", BigInteger),
+    Column("seqincrement", BigInteger),
+    Column("seqmax", BigInteger),
+    Column("seqmin", BigInteger),
+    Column("seqcache", BigInteger),
+    Column("seqcycle", Boolean),
+    info={"server_version": (10,)},
+)
+
+pg_attrdef = Table(
+    "pg_attrdef",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("adrelid", OID),
+    Column("adnum", SmallInteger),
+    Column("adbin", PG_NODE_TREE),
+)
+
+pg_description = Table(
+    "pg_description",
+    pg_catalog_meta,
+    Column("objoid", OID),
+    Column("classoid", OID),
+    Column("objsubid", Integer),
+    Column("description", Text(collation="C")),
+)
+
+pg_enum = Table(
+    "pg_enum",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("enumtypid", OID),
+    Column("enumsortorder", Float(), info={"server_version": (9, 1)}),
+    Column("enumlabel", NAME),
+)
+
+pg_am = Table(
+    "pg_am",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("amname", NAME),
+    Column("amhandler", REGPROC, info={"server_version": (9, 6)}),
+    Column("amtype", CHAR, info={"server_version": (9, 6)}),
+)
+
+pg_collation = Table(
+    "pg_collation",
+    pg_catalog_meta,
+    Column("oid", OID, info={"server_version": (9, 3)}),
+    Column("collname", NAME),
+    Column("collnamespace", OID),
+    Column("collowner", OID),
+    Column("collprovider", CHAR, info={"server_version": (10,)}),
+    Column("collisdeterministic", Boolean, info={"server_version": (12,)}),
+    Column("collencoding", Integer),
+    Column("collcollate", Text),
+    Column("collctype", Text),
+    Column("colliculocale", Text),
+    Column("collicurules", Text, info={"server_version": (16,)}),
+    Column("collversion", Text, info={"server_version": (10,)}),
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/provision.py
new file mode 100644
index 00000000..c76f5f51
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/provision.py
@@ -0,0 +1,175 @@
+# dialects/postgresql/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import time
+
+from ... import exc
+from ... import inspect
+from ... import text
+from ...testing import warn_test_suite
+from ...testing.provision import create_db
+from ...testing.provision import drop_all_schema_objects_post_tables
+from ...testing.provision import drop_all_schema_objects_pre_tables
+from ...testing.provision import drop_db
+from ...testing.provision import log
+from ...testing.provision import post_configure_engine
+from ...testing.provision import prepare_for_drop_tables
+from ...testing.provision import set_default_schema_on_connection
+from ...testing.provision import temp_table_keyword_args
+from ...testing.provision import upsert
+
+
+@create_db.for_db("postgresql")
+def _pg_create_db(cfg, eng, ident):
+    template_db = cfg.options.postgresql_templatedb
+
+    with eng.execution_options(isolation_level="AUTOCOMMIT").begin() as conn:
+        if not template_db:
+            template_db = conn.exec_driver_sql(
+                "select current_database()"
+            ).scalar()
+
+        attempt = 0
+        while True:
+            try:
+                conn.exec_driver_sql(
+                    "CREATE DATABASE %s TEMPLATE %s" % (ident, template_db)
+                )
+            except exc.OperationalError as err:
+                attempt += 1
+                if attempt >= 3:
+                    raise
+                if "accessed by other users" in str(err):
+                    log.info(
+                        "Waiting to create %s, URI %r, "
+                        "template DB %s is in use sleeping for .5",
+                        ident,
+                        eng.url,
+                        template_db,
+                    )
+                    time.sleep(0.5)
+            except:
+                raise
+            else:
+                break
+
+
+@drop_db.for_db("postgresql")
+def _pg_drop_db(cfg, eng, ident):
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        with conn.begin():
+            conn.execute(
+                text(
+                    "select pg_terminate_backend(pid) from pg_stat_activity "
+                    "where usename=current_user and pid != pg_backend_pid() "
+                    "and datname=:dname"
+                ),
+                dict(dname=ident),
+            )
+            conn.exec_driver_sql("DROP DATABASE %s" % ident)
+
+
+@temp_table_keyword_args.for_db("postgresql")
+def _postgresql_temp_table_keyword_args(cfg, eng):
+    return {"prefixes": ["TEMPORARY"]}
+
+
+@set_default_schema_on_connection.for_db("postgresql")
+def _postgresql_set_default_schema_on_connection(
+    cfg, dbapi_connection, schema_name
+):
+    existing_autocommit = dbapi_connection.autocommit
+    dbapi_connection.autocommit = True
+    cursor = dbapi_connection.cursor()
+    cursor.execute("SET SESSION search_path='%s'" % schema_name)
+    cursor.close()
+    dbapi_connection.autocommit = existing_autocommit
+
+
+@drop_all_schema_objects_pre_tables.for_db("postgresql")
+def drop_all_schema_objects_pre_tables(cfg, eng):
+    with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn:
+        for xid in conn.exec_driver_sql(
+            "select gid from pg_prepared_xacts"
+        ).scalars():
+            conn.exec_driver_sql("ROLLBACK PREPARED '%s'" % xid)
+
+
+@drop_all_schema_objects_post_tables.for_db("postgresql")
+def drop_all_schema_objects_post_tables(cfg, eng):
+    from sqlalchemy.dialects import postgresql
+
+    inspector = inspect(eng)
+    with eng.begin() as conn:
+        for enum in inspector.get_enums("*"):
+            conn.execute(
+                postgresql.DropEnumType(
+                    postgresql.ENUM(name=enum["name"], schema=enum["schema"])
+                )
+            )
+
+
+@prepare_for_drop_tables.for_db("postgresql")
+def prepare_for_drop_tables(config, connection):
+    """Ensure there are no locks on the current username/database."""
+
+    result = connection.exec_driver_sql(
+        "select pid, state, wait_event_type, query "
+        # "select pg_terminate_backend(pid), state, wait_event_type "
+        "from pg_stat_activity where "
+        "usename=current_user "
+        "and datname=current_database() and state='idle in transaction' "
+        "and pid != pg_backend_pid()"
+    )
+    rows = result.all()  # noqa
+    if rows:
+        warn_test_suite(
+            "PostgreSQL may not be able to DROP tables due to "
+            "idle in transaction: %s"
+            % ("; ".join(row._mapping["query"] for row in rows))
+        )
+
+
+@upsert.for_db("postgresql")
+def _upsert(
+    cfg, table, returning, *, set_lambda=None, sort_by_parameter_order=False
+):
+    from sqlalchemy.dialects.postgresql import insert
+
+    stmt = insert(table)
+
+    table_pk = inspect(table).selectable
+
+    if set_lambda:
+        stmt = stmt.on_conflict_do_update(
+            index_elements=table_pk.primary_key, set_=set_lambda(stmt.excluded)
+        )
+    else:
+        stmt = stmt.on_conflict_do_nothing()
+
+    stmt = stmt.returning(
+        *returning, sort_by_parameter_order=sort_by_parameter_order
+    )
+    return stmt
+
+
+_extensions = [
+    ("citext", (13,)),
+    ("hstore", (13,)),
+]
+
+
+@post_configure_engine.for_db("postgresql")
+def _create_citext_extension(url, engine, follower_ident):
+    with engine.connect() as conn:
+        for extension, min_version in _extensions:
+            if conn.dialect.server_version_info >= min_version:
+                conn.execute(
+                    text(f"CREATE EXTENSION IF NOT EXISTS {extension}")
+                )
+                conn.commit()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg.py
new file mode 100644
index 00000000..0554048c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg.py
@@ -0,0 +1,783 @@
+# dialects/postgresql/psycopg.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql+psycopg
+    :name: psycopg (a.k.a. psycopg 3)
+    :dbapi: psycopg
+    :connectstring: postgresql+psycopg://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://pypi.org/project/psycopg/
+
+``psycopg`` is the package and module name for version 3 of the ``psycopg``
+database driver, formerly known as ``psycopg2``.  This driver is different
+enough from its ``psycopg2`` predecessor that SQLAlchemy supports it
+via a totally separate dialect; support for ``psycopg2`` is expected to remain
+for as long as that package continues to function for modern Python versions,
+and also remains the default dialect for the ``postgresql://`` dialect
+series.
+
+The SQLAlchemy ``psycopg`` dialect provides both a sync and an async
+implementation under the same dialect name. The proper version is
+selected depending on how the engine is created:
+
+* calling :func:`_sa.create_engine` with ``postgresql+psycopg://...`` will
+  automatically select the sync version, e.g.::
+
+    from sqlalchemy import create_engine
+
+    sync_engine = create_engine(
+        "postgresql+psycopg://scott:tiger@localhost/test"
+    )
+
+* calling :func:`_asyncio.create_async_engine` with
+  ``postgresql+psycopg://...`` will automatically select the async version,
+  e.g.::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    asyncio_engine = create_async_engine(
+        "postgresql+psycopg://scott:tiger@localhost/test"
+    )
+
+The asyncio version of the dialect may also be specified explicitly using the
+``psycopg_async`` suffix, as::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    asyncio_engine = create_async_engine(
+        "postgresql+psycopg_async://scott:tiger@localhost/test"
+    )
+
+.. seealso::
+
+    :ref:`postgresql_psycopg2` - The SQLAlchemy ``psycopg``
+    dialect shares most of its behavior with the ``psycopg2`` dialect.
+    Further documentation is available there.
+
+Using a different Cursor class
+------------------------------
+
+One of the differences between ``psycopg`` and the older ``psycopg2``
+is how bound parameters are handled: ``psycopg2`` would bind them
+client side, while ``psycopg`` by default will bind them server side.
+
+It's possible to configure ``psycopg`` to do client side binding by
+specifying the ``cursor_factory`` to be ``ClientCursor`` when creating
+the engine::
+
+    from psycopg import ClientCursor
+
+    client_side_engine = create_engine(
+        "postgresql+psycopg://...",
+        connect_args={"cursor_factory": ClientCursor},
+    )
+
+Similarly when using an async engine the ``AsyncClientCursor`` can be
+specified::
+
+    from psycopg import AsyncClientCursor
+
+    client_side_engine = create_async_engine(
+        "postgresql+psycopg://...",
+        connect_args={"cursor_factory": AsyncClientCursor},
+    )
+
+.. seealso::
+
+    `Client-side-binding cursors <https://www.psycopg.org/psycopg3/docs/advanced/cursors.html#client-side-binding-cursors>`_
+
+"""  # noqa
+from __future__ import annotations
+
+from collections import deque
+import logging
+import re
+from typing import cast
+from typing import TYPE_CHECKING
+
+from . import ranges
+from ._psycopg_common import _PGDialect_common_psycopg
+from ._psycopg_common import _PGExecutionContext_common_psycopg
+from .base import INTERVAL
+from .base import PGCompiler
+from .base import PGIdentifierPreparer
+from .base import REGCONFIG
+from .json import JSON
+from .json import JSONB
+from .json import JSONPathType
+from .types import CITEXT
+from ... import pool
+from ... import util
+from ...engine import AdaptedConnection
+from ...sql import sqltypes
+from ...util.concurrency import await_fallback
+from ...util.concurrency import await_only
+
+if TYPE_CHECKING:
+    from typing import Iterable
+
+    from psycopg import AsyncConnection
+
+logger = logging.getLogger("sqlalchemy.dialects.postgresql")
+
+
+class _PGString(sqltypes.String):
+    render_bind_cast = True
+
+
+class _PGREGCONFIG(REGCONFIG):
+    render_bind_cast = True
+
+
+class _PGJSON(JSON):
+    def bind_processor(self, dialect):
+        return self._make_bind_processor(None, dialect._psycopg_Json)
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _PGJSONB(JSONB):
+    def bind_processor(self, dialect):
+        return self._make_bind_processor(None, dialect._psycopg_Jsonb)
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _PGJSONIntIndexType(sqltypes.JSON.JSONIntIndexType):
+    __visit_name__ = "json_int_index"
+
+    render_bind_cast = True
+
+
+class _PGJSONStrIndexType(sqltypes.JSON.JSONStrIndexType):
+    __visit_name__ = "json_str_index"
+
+    render_bind_cast = True
+
+
+class _PGJSONPathType(JSONPathType):
+    pass
+
+
+class _PGInterval(INTERVAL):
+    render_bind_cast = True
+
+
+class _PGTimeStamp(sqltypes.DateTime):
+    render_bind_cast = True
+
+
+class _PGDate(sqltypes.Date):
+    render_bind_cast = True
+
+
+class _PGTime(sqltypes.Time):
+    render_bind_cast = True
+
+
+class _PGInteger(sqltypes.Integer):
+    render_bind_cast = True
+
+
+class _PGSmallInteger(sqltypes.SmallInteger):
+    render_bind_cast = True
+
+
+class _PGNullType(sqltypes.NullType):
+    render_bind_cast = True
+
+
+class _PGBigInteger(sqltypes.BigInteger):
+    render_bind_cast = True
+
+
+class _PGBoolean(sqltypes.Boolean):
+    render_bind_cast = True
+
+
+class _PsycopgRange(ranges.AbstractSingleRangeImpl):
+    def bind_processor(self, dialect):
+        psycopg_Range = cast(PGDialect_psycopg, dialect)._psycopg_Range
+
+        def to_range(value):
+            if isinstance(value, ranges.Range):
+                value = psycopg_Range(
+                    value.lower, value.upper, value.bounds, value.empty
+                )
+            return value
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range(value):
+            if value is not None:
+                value = ranges.Range(
+                    value._lower,
+                    value._upper,
+                    bounds=value._bounds if value._bounds else "[)",
+                    empty=not value._bounds,
+                )
+            return value
+
+        return to_range
+
+
+class _PsycopgMultiRange(ranges.AbstractMultiRangeImpl):
+    def bind_processor(self, dialect):
+        psycopg_Range = cast(PGDialect_psycopg, dialect)._psycopg_Range
+        psycopg_Multirange = cast(
+            PGDialect_psycopg, dialect
+        )._psycopg_Multirange
+
+        NoneType = type(None)
+
+        def to_range(value):
+            if isinstance(value, (str, NoneType, psycopg_Multirange)):
+                return value
+
+            return psycopg_Multirange(
+                [
+                    psycopg_Range(
+                        element.lower,
+                        element.upper,
+                        element.bounds,
+                        element.empty,
+                    )
+                    for element in cast("Iterable[ranges.Range]", value)
+                ]
+            )
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range(value):
+            if value is None:
+                return None
+            else:
+                return ranges.MultiRange(
+                    ranges.Range(
+                        elem._lower,
+                        elem._upper,
+                        bounds=elem._bounds if elem._bounds else "[)",
+                        empty=not elem._bounds,
+                    )
+                    for elem in value
+                )
+
+        return to_range
+
+
+class PGExecutionContext_psycopg(_PGExecutionContext_common_psycopg):
+    pass
+
+
+class PGCompiler_psycopg(PGCompiler):
+    pass
+
+
+class PGIdentifierPreparer_psycopg(PGIdentifierPreparer):
+    pass
+
+
+def _log_notices(diagnostic):
+    logger.info("%s: %s", diagnostic.severity, diagnostic.message_primary)
+
+
+class PGDialect_psycopg(_PGDialect_common_psycopg):
+    driver = "psycopg"
+
+    supports_statement_cache = True
+    supports_server_side_cursors = True
+    default_paramstyle = "pyformat"
+    supports_sane_multi_rowcount = True
+
+    execution_ctx_cls = PGExecutionContext_psycopg
+    statement_compiler = PGCompiler_psycopg
+    preparer = PGIdentifierPreparer_psycopg
+    psycopg_version = (0, 0)
+
+    _has_native_hstore = True
+    _psycopg_adapters_map = None
+
+    colspecs = util.update_copy(
+        _PGDialect_common_psycopg.colspecs,
+        {
+            sqltypes.String: _PGString,
+            REGCONFIG: _PGREGCONFIG,
+            JSON: _PGJSON,
+            CITEXT: CITEXT,
+            sqltypes.JSON: _PGJSON,
+            JSONB: _PGJSONB,
+            sqltypes.JSON.JSONPathType: _PGJSONPathType,
+            sqltypes.JSON.JSONIntIndexType: _PGJSONIntIndexType,
+            sqltypes.JSON.JSONStrIndexType: _PGJSONStrIndexType,
+            sqltypes.Interval: _PGInterval,
+            INTERVAL: _PGInterval,
+            sqltypes.Date: _PGDate,
+            sqltypes.DateTime: _PGTimeStamp,
+            sqltypes.Time: _PGTime,
+            sqltypes.Integer: _PGInteger,
+            sqltypes.SmallInteger: _PGSmallInteger,
+            sqltypes.BigInteger: _PGBigInteger,
+            ranges.AbstractSingleRange: _PsycopgRange,
+            ranges.AbstractMultiRange: _PsycopgMultiRange,
+        },
+    )
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+
+        if self.dbapi:
+            m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", self.dbapi.__version__)
+            if m:
+                self.psycopg_version = tuple(
+                    int(x) for x in m.group(1, 2, 3) if x is not None
+                )
+
+            if self.psycopg_version < (3, 0, 2):
+                raise ImportError(
+                    "psycopg version 3.0.2 or higher is required."
+                )
+
+            from psycopg.adapt import AdaptersMap
+
+            self._psycopg_adapters_map = adapters_map = AdaptersMap(
+                self.dbapi.adapters
+            )
+
+            if self._native_inet_types is False:
+                import psycopg.types.string
+
+                adapters_map.register_loader(
+                    "inet", psycopg.types.string.TextLoader
+                )
+                adapters_map.register_loader(
+                    "cidr", psycopg.types.string.TextLoader
+                )
+
+            if self._json_deserializer:
+                from psycopg.types.json import set_json_loads
+
+                set_json_loads(self._json_deserializer, adapters_map)
+
+            if self._json_serializer:
+                from psycopg.types.json import set_json_dumps
+
+                set_json_dumps(self._json_serializer, adapters_map)
+
+    def create_connect_args(self, url):
+        # see https://github.com/psycopg/psycopg/issues/83
+        cargs, cparams = super().create_connect_args(url)
+
+        if self._psycopg_adapters_map:
+            cparams["context"] = self._psycopg_adapters_map
+        if self.client_encoding is not None:
+            cparams["client_encoding"] = self.client_encoding
+        return cargs, cparams
+
+    def _type_info_fetch(self, connection, name):
+        from psycopg.types import TypeInfo
+
+        return TypeInfo.fetch(connection.connection.driver_connection, name)
+
+    def initialize(self, connection):
+        super().initialize(connection)
+
+        # PGDialect.initialize() checks server version for <= 8.2 and sets
+        # this flag to False if so
+        if not self.insert_returning:
+            self.insert_executemany_returning = False
+
+        # HSTORE can't be registered until we have a connection so that
+        # we can look up its OID, so we set up this adapter in
+        # initialize()
+        if self.use_native_hstore:
+            info = self._type_info_fetch(connection, "hstore")
+            self._has_native_hstore = info is not None
+            if self._has_native_hstore:
+                from psycopg.types.hstore import register_hstore
+
+                # register the adapter for connections made subsequent to
+                # this one
+                assert self._psycopg_adapters_map
+                register_hstore(info, self._psycopg_adapters_map)
+
+                # register the adapter for this connection
+                assert connection.connection
+                register_hstore(info, connection.connection.driver_connection)
+
+    @classmethod
+    def import_dbapi(cls):
+        import psycopg
+
+        return psycopg
+
+    @classmethod
+    def get_async_dialect_cls(cls, url):
+        return PGDialectAsync_psycopg
+
+    @util.memoized_property
+    def _isolation_lookup(self):
+        return {
+            "READ COMMITTED": self.dbapi.IsolationLevel.READ_COMMITTED,
+            "READ UNCOMMITTED": self.dbapi.IsolationLevel.READ_UNCOMMITTED,
+            "REPEATABLE READ": self.dbapi.IsolationLevel.REPEATABLE_READ,
+            "SERIALIZABLE": self.dbapi.IsolationLevel.SERIALIZABLE,
+        }
+
+    @util.memoized_property
+    def _psycopg_Json(self):
+        from psycopg.types import json
+
+        return json.Json
+
+    @util.memoized_property
+    def _psycopg_Jsonb(self):
+        from psycopg.types import json
+
+        return json.Jsonb
+
+    @util.memoized_property
+    def _psycopg_TransactionStatus(self):
+        from psycopg.pq import TransactionStatus
+
+        return TransactionStatus
+
+    @util.memoized_property
+    def _psycopg_Range(self):
+        from psycopg.types.range import Range
+
+        return Range
+
+    @util.memoized_property
+    def _psycopg_Multirange(self):
+        from psycopg.types.multirange import Multirange
+
+        return Multirange
+
+    def _do_isolation_level(self, connection, autocommit, isolation_level):
+        connection.autocommit = autocommit
+        connection.isolation_level = isolation_level
+
+    def get_isolation_level(self, dbapi_connection):
+        status_before = dbapi_connection.info.transaction_status
+        value = super().get_isolation_level(dbapi_connection)
+
+        # don't rely on psycopg providing enum symbols, compare with
+        # eq/ne
+        if status_before == self._psycopg_TransactionStatus.IDLE:
+            dbapi_connection.rollback()
+        return value
+
+    def set_isolation_level(self, dbapi_connection, level):
+        if level == "AUTOCOMMIT":
+            self._do_isolation_level(
+                dbapi_connection, autocommit=True, isolation_level=None
+            )
+        else:
+            self._do_isolation_level(
+                dbapi_connection,
+                autocommit=False,
+                isolation_level=self._isolation_lookup[level],
+            )
+
+    def set_readonly(self, connection, value):
+        connection.read_only = value
+
+    def get_readonly(self, connection):
+        return connection.read_only
+
+    def on_connect(self):
+        def notices(conn):
+            conn.add_notice_handler(_log_notices)
+
+        fns = [notices]
+
+        if self.isolation_level is not None:
+
+            def on_connect(conn):
+                self.set_isolation_level(conn, self.isolation_level)
+
+            fns.append(on_connect)
+
+        # fns always has the notices function
+        def on_connect(conn):
+            for fn in fns:
+                fn(conn)
+
+        return on_connect
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(e, self.dbapi.Error) and connection is not None:
+            if connection.closed or connection.broken:
+                return True
+        return False
+
+    def _do_prepared_twophase(self, connection, command, recover=False):
+        dbapi_conn = connection.connection.dbapi_connection
+        if (
+            recover
+            # don't rely on psycopg providing enum symbols, compare with
+            # eq/ne
+            or dbapi_conn.info.transaction_status
+            != self._psycopg_TransactionStatus.IDLE
+        ):
+            dbapi_conn.rollback()
+        before_autocommit = dbapi_conn.autocommit
+        try:
+            if not before_autocommit:
+                self._do_autocommit(dbapi_conn, True)
+            dbapi_conn.execute(command)
+        finally:
+            if not before_autocommit:
+                self._do_autocommit(dbapi_conn, before_autocommit)
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if is_prepared:
+            self._do_prepared_twophase(
+                connection, f"ROLLBACK PREPARED '{xid}'", recover=recover
+            )
+        else:
+            self.do_rollback(connection.connection)
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        if is_prepared:
+            self._do_prepared_twophase(
+                connection, f"COMMIT PREPARED '{xid}'", recover=recover
+            )
+        else:
+            self.do_commit(connection.connection)
+
+    @util.memoized_property
+    def _dialect_specific_select_one(self):
+        return ";"
+
+
+class AsyncAdapt_psycopg_cursor:
+    __slots__ = ("_cursor", "await_", "_rows")
+
+    _psycopg_ExecStatus = None
+
+    def __init__(self, cursor, await_) -> None:
+        self._cursor = cursor
+        self.await_ = await_
+        self._rows = deque()
+
+    def __getattr__(self, name):
+        return getattr(self._cursor, name)
+
+    @property
+    def arraysize(self):
+        return self._cursor.arraysize
+
+    @arraysize.setter
+    def arraysize(self, value):
+        self._cursor.arraysize = value
+
+    def close(self):
+        self._rows.clear()
+        # Normal cursor just call _close() in a non-sync way.
+        self._cursor._close()
+
+    def execute(self, query, params=None, **kw):
+        result = self.await_(self._cursor.execute(query, params, **kw))
+        # sqlalchemy result is not async, so need to pull all rows here
+        res = self._cursor.pgresult
+
+        # don't rely on psycopg providing enum symbols, compare with
+        # eq/ne
+        if res and res.status == self._psycopg_ExecStatus.TUPLES_OK:
+            rows = self.await_(self._cursor.fetchall())
+            self._rows = deque(rows)
+        return result
+
+    def executemany(self, query, params_seq):
+        return self.await_(self._cursor.executemany(query, params_seq))
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self._cursor.arraysize
+
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_psycopg_ss_cursor(AsyncAdapt_psycopg_cursor):
+    def execute(self, query, params=None, **kw):
+        self.await_(self._cursor.execute(query, params, **kw))
+        return self
+
+    def close(self):
+        self.await_(self._cursor.close())
+
+    def fetchone(self):
+        return self.await_(self._cursor.fetchone())
+
+    def fetchmany(self, size=0):
+        return self.await_(self._cursor.fetchmany(size))
+
+    def fetchall(self):
+        return self.await_(self._cursor.fetchall())
+
+    def __iter__(self):
+        iterator = self._cursor.__aiter__()
+        while True:
+            try:
+                yield self.await_(iterator.__anext__())
+            except StopAsyncIteration:
+                break
+
+
+class AsyncAdapt_psycopg_connection(AdaptedConnection):
+    _connection: AsyncConnection
+    __slots__ = ()
+    await_ = staticmethod(await_only)
+
+    def __init__(self, connection) -> None:
+        self._connection = connection
+
+    def __getattr__(self, name):
+        return getattr(self._connection, name)
+
+    def execute(self, query, params=None, **kw):
+        cursor = self.await_(self._connection.execute(query, params, **kw))
+        return AsyncAdapt_psycopg_cursor(cursor, self.await_)
+
+    def cursor(self, *args, **kw):
+        cursor = self._connection.cursor(*args, **kw)
+        if hasattr(cursor, "name"):
+            return AsyncAdapt_psycopg_ss_cursor(cursor, self.await_)
+        else:
+            return AsyncAdapt_psycopg_cursor(cursor, self.await_)
+
+    def commit(self):
+        self.await_(self._connection.commit())
+
+    def rollback(self):
+        self.await_(self._connection.rollback())
+
+    def close(self):
+        self.await_(self._connection.close())
+
+    @property
+    def autocommit(self):
+        return self._connection.autocommit
+
+    @autocommit.setter
+    def autocommit(self, value):
+        self.set_autocommit(value)
+
+    def set_autocommit(self, value):
+        self.await_(self._connection.set_autocommit(value))
+
+    def set_isolation_level(self, value):
+        self.await_(self._connection.set_isolation_level(value))
+
+    def set_read_only(self, value):
+        self.await_(self._connection.set_read_only(value))
+
+    def set_deferrable(self, value):
+        self.await_(self._connection.set_deferrable(value))
+
+
+class AsyncAdaptFallback_psycopg_connection(AsyncAdapt_psycopg_connection):
+    __slots__ = ()
+    await_ = staticmethod(await_fallback)
+
+
+class PsycopgAdaptDBAPI:
+    def __init__(self, psycopg) -> None:
+        self.psycopg = psycopg
+
+        for k, v in self.psycopg.__dict__.items():
+            if k != "connect":
+                self.__dict__[k] = v
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+        creator_fn = kw.pop(
+            "async_creator_fn", self.psycopg.AsyncConnection.connect
+        )
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_psycopg_connection(
+                await_fallback(creator_fn(*arg, **kw))
+            )
+        else:
+            return AsyncAdapt_psycopg_connection(
+                await_only(creator_fn(*arg, **kw))
+            )
+
+
+class PGDialectAsync_psycopg(PGDialect_psycopg):
+    is_async = True
+    supports_statement_cache = True
+
+    @classmethod
+    def import_dbapi(cls):
+        import psycopg
+        from psycopg.pq import ExecStatus
+
+        AsyncAdapt_psycopg_cursor._psycopg_ExecStatus = ExecStatus
+
+        return PsycopgAdaptDBAPI(psycopg)
+
+    @classmethod
+    def get_pool_class(cls, url):
+        async_fallback = url.query.get("async_fallback", False)
+
+        if util.asbool(async_fallback):
+            return pool.FallbackAsyncAdaptedQueuePool
+        else:
+            return pool.AsyncAdaptedQueuePool
+
+    def _type_info_fetch(self, connection, name):
+        from psycopg.types import TypeInfo
+
+        adapted = connection.connection
+        return adapted.await_(TypeInfo.fetch(adapted.driver_connection, name))
+
+    def _do_isolation_level(self, connection, autocommit, isolation_level):
+        connection.set_autocommit(autocommit)
+        connection.set_isolation_level(isolation_level)
+
+    def _do_autocommit(self, connection, value):
+        connection.set_autocommit(value)
+
+    def set_readonly(self, connection, value):
+        connection.set_read_only(value)
+
+    def set_deferrable(self, connection, value):
+        connection.set_deferrable(value)
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = PGDialect_psycopg
+dialect_async = PGDialectAsync_psycopg
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py
new file mode 100644
index 00000000..eeb7604f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py
@@ -0,0 +1,892 @@
+# dialects/postgresql/psycopg2.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql+psycopg2
+    :name: psycopg2
+    :dbapi: psycopg2
+    :connectstring: postgresql+psycopg2://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://pypi.org/project/psycopg2/
+
+.. _psycopg2_toplevel:
+
+psycopg2 Connect Arguments
+--------------------------
+
+Keyword arguments that are specific to the SQLAlchemy psycopg2 dialect
+may be passed to :func:`_sa.create_engine()`, and include the following:
+
+
+* ``isolation_level``: This option, available for all PostgreSQL dialects,
+  includes the ``AUTOCOMMIT`` isolation level when using the psycopg2
+  dialect.   This option sets the **default** isolation level for the
+  connection that is set immediately upon connection to the database before
+  the connection is pooled.  This option is generally superseded by the more
+  modern :paramref:`_engine.Connection.execution_options.isolation_level`
+  execution option, detailed at :ref:`dbapi_autocommit`.
+
+  .. seealso::
+
+    :ref:`psycopg2_isolation_level`
+
+    :ref:`dbapi_autocommit`
+
+
+* ``client_encoding``: sets the client encoding in a libpq-agnostic way,
+  using psycopg2's ``set_client_encoding()`` method.
+
+  .. seealso::
+
+    :ref:`psycopg2_unicode`
+
+
+* ``executemany_mode``, ``executemany_batch_page_size``,
+  ``executemany_values_page_size``: Allows use of psycopg2
+  extensions for optimizing "executemany"-style queries.  See the referenced
+  section below for details.
+
+  .. seealso::
+
+    :ref:`psycopg2_executemany_mode`
+
+.. tip::
+
+    The above keyword arguments are **dialect** keyword arguments, meaning
+    that they are passed as explicit keyword arguments to :func:`_sa.create_engine()`::
+
+        engine = create_engine(
+            "postgresql+psycopg2://scott:tiger@localhost/test",
+            isolation_level="SERIALIZABLE",
+        )
+
+    These should not be confused with **DBAPI** connect arguments, which
+    are passed as part of the :paramref:`_sa.create_engine.connect_args`
+    dictionary and/or are passed in the URL query string, as detailed in
+    the section :ref:`custom_dbapi_args`.
+
+.. _psycopg2_ssl:
+
+SSL Connections
+---------------
+
+The psycopg2 module has a connection argument named ``sslmode`` for
+controlling its behavior regarding secure (SSL) connections. The default is
+``sslmode=prefer``; it will attempt an SSL connection and if that fails it
+will fall back to an unencrypted connection. ``sslmode=require`` may be used
+to ensure that only secure connections are established.  Consult the
+psycopg2 / libpq documentation for further options that are available.
+
+Note that ``sslmode`` is specific to psycopg2 so it is included in the
+connection URI::
+
+    engine = sa.create_engine(
+        "postgresql+psycopg2://scott:tiger@192.168.0.199:5432/test?sslmode=require"
+    )
+
+Unix Domain Connections
+------------------------
+
+psycopg2 supports connecting via Unix domain connections.   When the ``host``
+portion of the URL is omitted, SQLAlchemy passes ``None`` to psycopg2,
+which specifies Unix-domain communication rather than TCP/IP communication::
+
+    create_engine("postgresql+psycopg2://user:password@/dbname")
+
+By default, the socket file used is to connect to a Unix-domain socket
+in ``/tmp``, or whatever socket directory was specified when PostgreSQL
+was built.  This value can be overridden by passing a pathname to psycopg2,
+using ``host`` as an additional keyword argument::
+
+    create_engine(
+        "postgresql+psycopg2://user:password@/dbname?host=/var/lib/postgresql"
+    )
+
+.. warning::  The format accepted here allows for a hostname in the main URL
+   in addition to the "host" query string argument.  **When using this URL
+   format, the initial host is silently ignored**.  That is, this URL::
+
+        engine = create_engine(
+            "postgresql+psycopg2://user:password@myhost1/dbname?host=myhost2"
+        )
+
+   Above, the hostname ``myhost1`` is **silently ignored and discarded.**  The
+   host which is connected is the ``myhost2`` host.
+
+   This is to maintain some degree of compatibility with PostgreSQL's own URL
+   format which has been tested to behave the same way and for which tools like
+   PifPaf hardcode two hostnames.
+
+.. seealso::
+
+    `PQconnectdbParams \
+    <https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PQCONNECTDBPARAMS>`_
+
+.. _psycopg2_multi_host:
+
+Specifying multiple fallback hosts
+-----------------------------------
+
+psycopg2 supports multiple connection points in the connection string.
+When the ``host`` parameter is used multiple times in the query section of
+the URL, SQLAlchemy will create a single string of the host and port
+information provided to make the connections.  Tokens may consist of
+``host::port`` or just ``host``; in the latter case, the default port
+is selected by libpq.  In the example below, three host connections
+are specified, for ``HostA::PortA``, ``HostB`` connecting to the default port,
+and ``HostC::PortC``::
+
+    create_engine(
+        "postgresql+psycopg2://user:password@/dbname?host=HostA:PortA&host=HostB&host=HostC:PortC"
+    )
+
+As an alternative, libpq query string format also may be used; this specifies
+``host`` and ``port`` as single query string arguments with comma-separated
+lists - the default port can be chosen by indicating an empty value
+in the comma separated list::
+
+    create_engine(
+        "postgresql+psycopg2://user:password@/dbname?host=HostA,HostB,HostC&port=PortA,,PortC"
+    )
+
+With either URL style, connections to each host is attempted based on a
+configurable strategy, which may be configured using the libpq
+``target_session_attrs`` parameter.  Per libpq this defaults to ``any``
+which indicates a connection to each host is then attempted until a connection is successful.
+Other strategies include ``primary``, ``prefer-standby``, etc.  The complete
+list is documented by PostgreSQL at
+`libpq connection strings <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNSTRING>`_.
+
+For example, to indicate two hosts using the ``primary`` strategy::
+
+    create_engine(
+        "postgresql+psycopg2://user:password@/dbname?host=HostA:PortA&host=HostB&host=HostC:PortC&target_session_attrs=primary"
+    )
+
+.. versionchanged:: 1.4.40 Port specification in psycopg2 multiple host format
+   is repaired, previously ports were not correctly interpreted in this context.
+   libpq comma-separated format is also now supported.
+
+.. versionadded:: 1.3.20 Support for multiple hosts in PostgreSQL connection
+   string.
+
+.. seealso::
+
+    `libpq connection strings <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNSTRING>`_ - please refer
+    to this section in the libpq documentation for complete background on multiple host support.
+
+
+Empty DSN Connections / Environment Variable Connections
+---------------------------------------------------------
+
+The psycopg2 DBAPI can connect to PostgreSQL by passing an empty DSN to the
+libpq client library, which by default indicates to connect to a localhost
+PostgreSQL database that is open for "trust" connections.  This behavior can be
+further tailored using a particular set of environment variables which are
+prefixed with ``PG_...``, which are  consumed by ``libpq`` to take the place of
+any or all elements of the connection string.
+
+For this form, the URL can be passed without any elements other than the
+initial scheme::
+
+    engine = create_engine("postgresql+psycopg2://")
+
+In the above form, a blank "dsn" string is passed to the ``psycopg2.connect()``
+function which in turn represents an empty DSN passed to libpq.
+
+.. versionadded:: 1.3.2 support for parameter-less connections with psycopg2.
+
+.. seealso::
+
+    `Environment Variables\
+    <https://www.postgresql.org/docs/current/libpq-envars.html>`_ -
+    PostgreSQL documentation on how to use ``PG_...``
+    environment variables for connections.
+
+.. _psycopg2_execution_options:
+
+Per-Statement/Connection Execution Options
+-------------------------------------------
+
+The following DBAPI-specific options are respected when used with
+:meth:`_engine.Connection.execution_options`,
+:meth:`.Executable.execution_options`,
+:meth:`_query.Query.execution_options`,
+in addition to those not specific to DBAPIs:
+
+* ``isolation_level`` - Set the transaction isolation level for the lifespan
+  of a :class:`_engine.Connection` (can only be set on a connection,
+  not a statement
+  or query).   See :ref:`psycopg2_isolation_level`.
+
+* ``stream_results`` - Enable or disable usage of psycopg2 server side
+  cursors - this feature makes use of "named" cursors in combination with
+  special result handling methods so that result rows are not fully buffered.
+  Defaults to False, meaning cursors are buffered by default.
+
+* ``max_row_buffer`` - when using ``stream_results``, an integer value that
+  specifies the maximum number of rows to buffer at a time.  This is
+  interpreted by the :class:`.BufferedRowCursorResult`, and if omitted the
+  buffer will grow to ultimately store 1000 rows at a time.
+
+  .. versionchanged:: 1.4  The ``max_row_buffer`` size can now be greater than
+     1000, and the buffer will grow to that size.
+
+.. _psycopg2_batch_mode:
+
+.. _psycopg2_executemany_mode:
+
+Psycopg2 Fast Execution Helpers
+-------------------------------
+
+Modern versions of psycopg2 include a feature known as
+`Fast Execution Helpers \
+<https://www.psycopg.org/docs/extras.html#fast-execution-helpers>`_, which
+have been shown in benchmarking to improve psycopg2's executemany()
+performance, primarily with INSERT statements, by at least
+an order of magnitude.
+
+SQLAlchemy implements a native form of the "insert many values"
+handler that will rewrite a single-row INSERT statement to accommodate for
+many values at once within an extended VALUES clause; this handler is
+equivalent to psycopg2's ``execute_values()`` handler; an overview of this
+feature and its configuration are at :ref:`engine_insertmanyvalues`.
+
+.. versionadded:: 2.0 Replaced psycopg2's ``execute_values()`` fast execution
+   helper with a native SQLAlchemy mechanism known as
+   :ref:`insertmanyvalues <engine_insertmanyvalues>`.
+
+The psycopg2 dialect retains the ability to use the psycopg2-specific
+``execute_batch()`` feature, although it is not expected that this is a widely
+used feature.  The use of this extension may be enabled using the
+``executemany_mode`` flag which may be passed to :func:`_sa.create_engine`::
+
+    engine = create_engine(
+        "postgresql+psycopg2://scott:tiger@host/dbname",
+        executemany_mode="values_plus_batch",
+    )
+
+Possible options for ``executemany_mode`` include:
+
+* ``values_only`` - this is the default value.  SQLAlchemy's native
+  :ref:`insertmanyvalues <engine_insertmanyvalues>` handler is used for qualifying
+  INSERT statements, assuming
+  :paramref:`_sa.create_engine.use_insertmanyvalues` is left at
+  its default value of ``True``.  This handler rewrites simple
+  INSERT statements to include multiple VALUES clauses so that many
+  parameter sets can be inserted with one statement.
+
+* ``'values_plus_batch'``- SQLAlchemy's native
+  :ref:`insertmanyvalues <engine_insertmanyvalues>` handler is used for qualifying
+  INSERT statements, assuming
+  :paramref:`_sa.create_engine.use_insertmanyvalues` is left at its default
+  value of ``True``. Then, psycopg2's ``execute_batch()`` handler is used for
+  qualifying UPDATE and DELETE statements when executed with multiple parameter
+  sets. When using this mode, the :attr:`_engine.CursorResult.rowcount`
+  attribute will not contain a value for executemany-style executions against
+  UPDATE and DELETE statements.
+
+.. versionchanged:: 2.0 Removed the ``'batch'`` and ``'None'`` options
+   from psycopg2 ``executemany_mode``.  Control over batching for INSERT
+   statements is now configured via the
+   :paramref:`_sa.create_engine.use_insertmanyvalues` engine-level parameter.
+
+The term "qualifying statements" refers to the statement being executed
+being a Core :func:`_expression.insert`, :func:`_expression.update`
+or :func:`_expression.delete` construct, and **not** a plain textual SQL
+string or one constructed using :func:`_expression.text`.  It also may **not** be
+a special "extension" statement such as an "ON CONFLICT" "upsert" statement.
+When using the ORM, all insert/update/delete statements used by the ORM flush process
+are qualifying.
+
+The "page size" for the psycopg2 "batch" strategy can be affected
+by using the ``executemany_batch_page_size`` parameter, which defaults to
+100.
+
+For the "insertmanyvalues" feature, the page size can be controlled using the
+:paramref:`_sa.create_engine.insertmanyvalues_page_size` parameter,
+which defaults to 1000.  An example of modifying both parameters
+is below::
+
+    engine = create_engine(
+        "postgresql+psycopg2://scott:tiger@host/dbname",
+        executemany_mode="values_plus_batch",
+        insertmanyvalues_page_size=5000,
+        executemany_batch_page_size=500,
+    )
+
+.. seealso::
+
+    :ref:`engine_insertmanyvalues` - background on "insertmanyvalues"
+
+    :ref:`tutorial_multiple_parameters` - General information on using the
+    :class:`_engine.Connection`
+    object to execute statements in such a way as to make
+    use of the DBAPI ``.executemany()`` method.
+
+
+.. _psycopg2_unicode:
+
+Unicode with Psycopg2
+----------------------
+
+The psycopg2 DBAPI driver supports Unicode data transparently.
+
+The client character encoding can be controlled for the psycopg2 dialect
+in the following ways:
+
+* For PostgreSQL 9.1 and above, the ``client_encoding`` parameter may be
+  passed in the database URL; this parameter is consumed by the underlying
+  ``libpq`` PostgreSQL client library::
+
+    engine = create_engine(
+        "postgresql+psycopg2://user:pass@host/dbname?client_encoding=utf8"
+    )
+
+  Alternatively, the above ``client_encoding`` value may be passed using
+  :paramref:`_sa.create_engine.connect_args` for programmatic establishment with
+  ``libpq``::
+
+    engine = create_engine(
+        "postgresql+psycopg2://user:pass@host/dbname",
+        connect_args={"client_encoding": "utf8"},
+    )
+
+* For all PostgreSQL versions, psycopg2 supports a client-side encoding
+  value that will be passed to database connections when they are first
+  established.  The SQLAlchemy psycopg2 dialect supports this using the
+  ``client_encoding`` parameter passed to :func:`_sa.create_engine`::
+
+      engine = create_engine(
+          "postgresql+psycopg2://user:pass@host/dbname", client_encoding="utf8"
+      )
+
+  .. tip:: The above ``client_encoding`` parameter admittedly is very similar
+      in appearance to usage of the parameter within the
+      :paramref:`_sa.create_engine.connect_args` dictionary; the difference
+      above is that the parameter is consumed by psycopg2 and is
+      passed to the database connection using ``SET client_encoding TO
+      'utf8'``; in the previously mentioned style, the parameter is instead
+      passed through psycopg2 and consumed by the ``libpq`` library.
+
+* A common way to set up client encoding with PostgreSQL databases is to
+  ensure it is configured within the server-side postgresql.conf file;
+  this is the recommended way to set encoding for a server that is
+  consistently of one encoding in all databases::
+
+    # postgresql.conf file
+
+    # client_encoding = sql_ascii # actually, defaults to database
+    # encoding
+    client_encoding = utf8
+
+Transactions
+------------
+
+The psycopg2 dialect fully supports SAVEPOINT and two-phase commit operations.
+
+.. _psycopg2_isolation_level:
+
+Psycopg2 Transaction Isolation Level
+-------------------------------------
+
+As discussed in :ref:`postgresql_isolation_level`,
+all PostgreSQL dialects support setting of transaction isolation level
+both via the ``isolation_level`` parameter passed to :func:`_sa.create_engine`
+,
+as well as the ``isolation_level`` argument used by
+:meth:`_engine.Connection.execution_options`.  When using the psycopg2 dialect
+, these
+options make use of psycopg2's ``set_isolation_level()`` connection method,
+rather than emitting a PostgreSQL directive; this is because psycopg2's
+API-level setting is always emitted at the start of each transaction in any
+case.
+
+The psycopg2 dialect supports these constants for isolation level:
+
+* ``READ COMMITTED``
+* ``READ UNCOMMITTED``
+* ``REPEATABLE READ``
+* ``SERIALIZABLE``
+* ``AUTOCOMMIT``
+
+.. seealso::
+
+    :ref:`postgresql_isolation_level`
+
+    :ref:`pg8000_isolation_level`
+
+
+NOTICE logging
+---------------
+
+The psycopg2 dialect will log PostgreSQL NOTICE messages
+via the ``sqlalchemy.dialects.postgresql`` logger.  When this logger
+is set to the ``logging.INFO`` level, notice messages will be logged::
+
+    import logging
+
+    logging.getLogger("sqlalchemy.dialects.postgresql").setLevel(logging.INFO)
+
+Above, it is assumed that logging is configured externally.  If this is not
+the case, configuration such as ``logging.basicConfig()`` must be utilized::
+
+    import logging
+
+    logging.basicConfig()  # log messages to stdout
+    logging.getLogger("sqlalchemy.dialects.postgresql").setLevel(logging.INFO)
+
+.. seealso::
+
+    `Logging HOWTO <https://docs.python.org/3/howto/logging.html>`_ - on the python.org website
+
+.. _psycopg2_hstore:
+
+HSTORE type
+------------
+
+The ``psycopg2`` DBAPI includes an extension to natively handle marshalling of
+the HSTORE type.   The SQLAlchemy psycopg2 dialect will enable this extension
+by default when psycopg2 version 2.4 or greater is used, and
+it is detected that the target database has the HSTORE type set up for use.
+In other words, when the dialect makes the first
+connection, a sequence like the following is performed:
+
+1. Request the available HSTORE oids using
+   ``psycopg2.extras.HstoreAdapter.get_oids()``.
+   If this function returns a list of HSTORE identifiers, we then determine
+   that the ``HSTORE`` extension is present.
+   This function is **skipped** if the version of psycopg2 installed is
+   less than version 2.4.
+
+2. If the ``use_native_hstore`` flag is at its default of ``True``, and
+   we've detected that ``HSTORE`` oids are available, the
+   ``psycopg2.extensions.register_hstore()`` extension is invoked for all
+   connections.
+
+The ``register_hstore()`` extension has the effect of **all Python
+dictionaries being accepted as parameters regardless of the type of target
+column in SQL**. The dictionaries are converted by this extension into a
+textual HSTORE expression.  If this behavior is not desired, disable the
+use of the hstore extension by setting ``use_native_hstore`` to ``False`` as
+follows::
+
+    engine = create_engine(
+        "postgresql+psycopg2://scott:tiger@localhost/test",
+        use_native_hstore=False,
+    )
+
+The ``HSTORE`` type is **still supported** when the
+``psycopg2.extensions.register_hstore()`` extension is not used.  It merely
+means that the coercion between Python dictionaries and the HSTORE
+string format, on both the parameter side and the result side, will take
+place within SQLAlchemy's own marshalling logic, and not that of ``psycopg2``
+which may be more performant.
+
+"""  # noqa
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import logging
+import re
+from typing import cast
+
+from . import ranges
+from ._psycopg_common import _PGDialect_common_psycopg
+from ._psycopg_common import _PGExecutionContext_common_psycopg
+from .base import PGIdentifierPreparer
+from .json import JSON
+from .json import JSONB
+from ... import types as sqltypes
+from ... import util
+from ...util import FastIntFlag
+from ...util import parse_user_argument_for_enum
+
+logger = logging.getLogger("sqlalchemy.dialects.postgresql")
+
+
+class _PGJSON(JSON):
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _PGJSONB(JSONB):
+    def result_processor(self, dialect, coltype):
+        return None
+
+
+class _Psycopg2Range(ranges.AbstractSingleRangeImpl):
+    _psycopg2_range_cls = "none"
+
+    def bind_processor(self, dialect):
+        psycopg2_Range = getattr(
+            cast(PGDialect_psycopg2, dialect)._psycopg2_extras,
+            self._psycopg2_range_cls,
+        )
+
+        def to_range(value):
+            if isinstance(value, ranges.Range):
+                value = psycopg2_Range(
+                    value.lower, value.upper, value.bounds, value.empty
+                )
+            return value
+
+        return to_range
+
+    def result_processor(self, dialect, coltype):
+        def to_range(value):
+            if value is not None:
+                value = ranges.Range(
+                    value._lower,
+                    value._upper,
+                    bounds=value._bounds if value._bounds else "[)",
+                    empty=not value._bounds,
+                )
+            return value
+
+        return to_range
+
+
+class _Psycopg2NumericRange(_Psycopg2Range):
+    _psycopg2_range_cls = "NumericRange"
+
+
+class _Psycopg2DateRange(_Psycopg2Range):
+    _psycopg2_range_cls = "DateRange"
+
+
+class _Psycopg2DateTimeRange(_Psycopg2Range):
+    _psycopg2_range_cls = "DateTimeRange"
+
+
+class _Psycopg2DateTimeTZRange(_Psycopg2Range):
+    _psycopg2_range_cls = "DateTimeTZRange"
+
+
+class PGExecutionContext_psycopg2(_PGExecutionContext_common_psycopg):
+    _psycopg2_fetched_rows = None
+
+    def post_exec(self):
+        self._log_notices(self.cursor)
+
+    def _log_notices(self, cursor):
+        # check also that notices is an iterable, after it's already
+        # established that we will be iterating through it.  This is to get
+        # around test suites such as SQLAlchemy's using a Mock object for
+        # cursor
+        if not cursor.connection.notices or not isinstance(
+            cursor.connection.notices, collections_abc.Iterable
+        ):
+            return
+
+        for notice in cursor.connection.notices:
+            # NOTICE messages have a
+            # newline character at the end
+            logger.info(notice.rstrip())
+
+        cursor.connection.notices[:] = []
+
+
+class PGIdentifierPreparer_psycopg2(PGIdentifierPreparer):
+    pass
+
+
+class ExecutemanyMode(FastIntFlag):
+    EXECUTEMANY_VALUES = 0
+    EXECUTEMANY_VALUES_PLUS_BATCH = 1
+
+
+(
+    EXECUTEMANY_VALUES,
+    EXECUTEMANY_VALUES_PLUS_BATCH,
+) = ExecutemanyMode.__members__.values()
+
+
+class PGDialect_psycopg2(_PGDialect_common_psycopg):
+    driver = "psycopg2"
+
+    supports_statement_cache = True
+    supports_server_side_cursors = True
+
+    default_paramstyle = "pyformat"
+    # set to true based on psycopg2 version
+    supports_sane_multi_rowcount = False
+    execution_ctx_cls = PGExecutionContext_psycopg2
+    preparer = PGIdentifierPreparer_psycopg2
+    psycopg2_version = (0, 0)
+    use_insertmanyvalues_wo_returning = True
+
+    returns_native_bytes = False
+
+    _has_native_hstore = True
+
+    colspecs = util.update_copy(
+        _PGDialect_common_psycopg.colspecs,
+        {
+            JSON: _PGJSON,
+            sqltypes.JSON: _PGJSON,
+            JSONB: _PGJSONB,
+            ranges.INT4RANGE: _Psycopg2NumericRange,
+            ranges.INT8RANGE: _Psycopg2NumericRange,
+            ranges.NUMRANGE: _Psycopg2NumericRange,
+            ranges.DATERANGE: _Psycopg2DateRange,
+            ranges.TSRANGE: _Psycopg2DateTimeRange,
+            ranges.TSTZRANGE: _Psycopg2DateTimeTZRange,
+        },
+    )
+
+    def __init__(
+        self,
+        executemany_mode="values_only",
+        executemany_batch_page_size=100,
+        **kwargs,
+    ):
+        _PGDialect_common_psycopg.__init__(self, **kwargs)
+
+        if self._native_inet_types:
+            raise NotImplementedError(
+                "The psycopg2 dialect does not implement "
+                "ipaddress type handling; native_inet_types cannot be set "
+                "to ``True`` when using this dialect."
+            )
+
+        # Parse executemany_mode argument, allowing it to be only one of the
+        # symbol names
+        self.executemany_mode = parse_user_argument_for_enum(
+            executemany_mode,
+            {
+                EXECUTEMANY_VALUES: ["values_only"],
+                EXECUTEMANY_VALUES_PLUS_BATCH: ["values_plus_batch"],
+            },
+            "executemany_mode",
+        )
+
+        self.executemany_batch_page_size = executemany_batch_page_size
+
+        if self.dbapi and hasattr(self.dbapi, "__version__"):
+            m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", self.dbapi.__version__)
+            if m:
+                self.psycopg2_version = tuple(
+                    int(x) for x in m.group(1, 2, 3) if x is not None
+                )
+
+            if self.psycopg2_version < (2, 7):
+                raise ImportError(
+                    "psycopg2 version 2.7 or higher is required."
+                )
+
+    def initialize(self, connection):
+        super().initialize(connection)
+        self._has_native_hstore = (
+            self.use_native_hstore
+            and self._hstore_oids(connection.connection.dbapi_connection)
+            is not None
+        )
+
+        self.supports_sane_multi_rowcount = (
+            self.executemany_mode is not EXECUTEMANY_VALUES_PLUS_BATCH
+        )
+
+    @classmethod
+    def import_dbapi(cls):
+        import psycopg2
+
+        return psycopg2
+
+    @util.memoized_property
+    def _psycopg2_extensions(cls):
+        from psycopg2 import extensions
+
+        return extensions
+
+    @util.memoized_property
+    def _psycopg2_extras(cls):
+        from psycopg2 import extras
+
+        return extras
+
+    @util.memoized_property
+    def _isolation_lookup(self):
+        extensions = self._psycopg2_extensions
+        return {
+            "AUTOCOMMIT": extensions.ISOLATION_LEVEL_AUTOCOMMIT,
+            "READ COMMITTED": extensions.ISOLATION_LEVEL_READ_COMMITTED,
+            "READ UNCOMMITTED": extensions.ISOLATION_LEVEL_READ_UNCOMMITTED,
+            "REPEATABLE READ": extensions.ISOLATION_LEVEL_REPEATABLE_READ,
+            "SERIALIZABLE": extensions.ISOLATION_LEVEL_SERIALIZABLE,
+        }
+
+    def set_isolation_level(self, dbapi_connection, level):
+        dbapi_connection.set_isolation_level(self._isolation_lookup[level])
+
+    def set_readonly(self, connection, value):
+        connection.readonly = value
+
+    def get_readonly(self, connection):
+        return connection.readonly
+
+    def set_deferrable(self, connection, value):
+        connection.deferrable = value
+
+    def get_deferrable(self, connection):
+        return connection.deferrable
+
+    def on_connect(self):
+        extras = self._psycopg2_extras
+
+        fns = []
+        if self.client_encoding is not None:
+
+            def on_connect(dbapi_conn):
+                dbapi_conn.set_client_encoding(self.client_encoding)
+
+            fns.append(on_connect)
+
+        if self.dbapi:
+
+            def on_connect(dbapi_conn):
+                extras.register_uuid(None, dbapi_conn)
+
+            fns.append(on_connect)
+
+        if self.dbapi and self.use_native_hstore:
+
+            def on_connect(dbapi_conn):
+                hstore_oids = self._hstore_oids(dbapi_conn)
+                if hstore_oids is not None:
+                    oid, array_oid = hstore_oids
+                    kw = {"oid": oid}
+                    kw["array_oid"] = array_oid
+                    extras.register_hstore(dbapi_conn, **kw)
+
+            fns.append(on_connect)
+
+        if self.dbapi and self._json_deserializer:
+
+            def on_connect(dbapi_conn):
+                extras.register_default_json(
+                    dbapi_conn, loads=self._json_deserializer
+                )
+                extras.register_default_jsonb(
+                    dbapi_conn, loads=self._json_deserializer
+                )
+
+            fns.append(on_connect)
+
+        if fns:
+
+            def on_connect(dbapi_conn):
+                for fn in fns:
+                    fn(dbapi_conn)
+
+            return on_connect
+        else:
+            return None
+
+    def do_executemany(self, cursor, statement, parameters, context=None):
+        if self.executemany_mode is EXECUTEMANY_VALUES_PLUS_BATCH:
+            if self.executemany_batch_page_size:
+                kwargs = {"page_size": self.executemany_batch_page_size}
+            else:
+                kwargs = {}
+            self._psycopg2_extras.execute_batch(
+                cursor, statement, parameters, **kwargs
+            )
+        else:
+            cursor.executemany(statement, parameters)
+
+    def do_begin_twophase(self, connection, xid):
+        connection.connection.tpc_begin(xid)
+
+    def do_prepare_twophase(self, connection, xid):
+        connection.connection.tpc_prepare()
+
+    def _do_twophase(self, dbapi_conn, operation, xid, recover=False):
+        if recover:
+            if dbapi_conn.status != self._psycopg2_extensions.STATUS_READY:
+                dbapi_conn.rollback()
+            operation(xid)
+        else:
+            operation()
+
+    def do_rollback_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        dbapi_conn = connection.connection.dbapi_connection
+        self._do_twophase(
+            dbapi_conn, dbapi_conn.tpc_rollback, xid, recover=recover
+        )
+
+    def do_commit_twophase(
+        self, connection, xid, is_prepared=True, recover=False
+    ):
+        dbapi_conn = connection.connection.dbapi_connection
+        self._do_twophase(
+            dbapi_conn, dbapi_conn.tpc_commit, xid, recover=recover
+        )
+
+    @util.memoized_instancemethod
+    def _hstore_oids(self, dbapi_connection):
+        extras = self._psycopg2_extras
+        oids = extras.HstoreAdapter.get_oids(dbapi_connection)
+        if oids is not None and oids[0]:
+            return oids[0:2]
+        else:
+            return None
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(e, self.dbapi.Error):
+            # check the "closed" flag.  this might not be
+            # present on old psycopg2 versions.   Also,
+            # this flag doesn't actually help in a lot of disconnect
+            # situations, so don't rely on it.
+            if getattr(connection, "closed", False):
+                return True
+
+            # checks based on strings.  in the case that .closed
+            # didn't cut it, fall back onto these.
+            str_e = str(e).partition("\n")[0]
+            for msg in self._is_disconnect_messages:
+                idx = str_e.find(msg)
+                if idx >= 0 and '"' not in str_e[:idx]:
+                    return True
+        return False
+
+    @util.memoized_property
+    def _is_disconnect_messages(self):
+        return (
+            # these error messages from libpq: interfaces/libpq/fe-misc.c
+            # and interfaces/libpq/fe-secure.c.
+            "terminating connection",
+            "closed the connection",
+            "connection not open",
+            "could not receive data from server",
+            "could not send data to server",
+            # psycopg2 client errors, psycopg2/connection.h,
+            # psycopg2/cursor.h
+            "connection already closed",
+            "cursor already closed",
+            # not sure where this path is originally from, it may
+            # be obsolete.   It really says "losed", not "closed".
+            "losed the connection unexpectedly",
+            # these can occur in newer SSL
+            "connection has been closed unexpectedly",
+            "SSL error: decryption failed or bad record mac",
+            "SSL SYSCALL error: Bad file descriptor",
+            "SSL SYSCALL error: EOF detected",
+            "SSL SYSCALL error: Operation timed out",
+            "SSL SYSCALL error: Bad address",
+            # This can occur in OpenSSL 1 when an unexpected EOF occurs.
+            # https://www.openssl.org/docs/man1.1.1/man3/SSL_get_error.html#BUGS
+            # It may also occur in newer OpenSSL for a non-recoverable I/O
+            # error as a result of a system call that does not set 'errno'
+            # in libc.
+            "SSL SYSCALL error: Success",
+        )
+
+
+dialect = PGDialect_psycopg2
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py
new file mode 100644
index 00000000..55e17607
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py
@@ -0,0 +1,61 @@
+# dialects/postgresql/psycopg2cffi.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+r"""
+.. dialect:: postgresql+psycopg2cffi
+    :name: psycopg2cffi
+    :dbapi: psycopg2cffi
+    :connectstring: postgresql+psycopg2cffi://user:password@host:port/dbname[?key=value&key=value...]
+    :url: https://pypi.org/project/psycopg2cffi/
+
+``psycopg2cffi`` is an adaptation of ``psycopg2``, using CFFI for the C
+layer. This makes it suitable for use in e.g. PyPy. Documentation
+is as per ``psycopg2``.
+
+.. seealso::
+
+    :mod:`sqlalchemy.dialects.postgresql.psycopg2`
+
+"""  # noqa
+from .psycopg2 import PGDialect_psycopg2
+from ... import util
+
+
+class PGDialect_psycopg2cffi(PGDialect_psycopg2):
+    driver = "psycopg2cffi"
+    supports_unicode_statements = True
+    supports_statement_cache = True
+
+    # psycopg2cffi's first release is 2.5.0, but reports
+    # __version__ as 2.4.4.  Subsequent releases seem to have
+    # fixed this.
+
+    FEATURE_VERSION_MAP = dict(
+        native_json=(2, 4, 4),
+        native_jsonb=(2, 7, 1),
+        sane_multi_rowcount=(2, 4, 4),
+        array_oid=(2, 4, 4),
+        hstore_adapter=(2, 4, 4),
+    )
+
+    @classmethod
+    def import_dbapi(cls):
+        return __import__("psycopg2cffi")
+
+    @util.memoized_property
+    def _psycopg2_extensions(cls):
+        root = __import__("psycopg2cffi", fromlist=["extensions"])
+        return root.extensions
+
+    @util.memoized_property
+    def _psycopg2_extras(cls):
+        root = __import__("psycopg2cffi", fromlist=["extras"])
+        return root.extras
+
+
+dialect = PGDialect_psycopg2cffi
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ranges.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ranges.py
new file mode 100644
index 00000000..93253570
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/ranges.py
@@ -0,0 +1,1031 @@
+# dialects/postgresql/ranges.py
+# Copyright (C) 2013-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import dataclasses
+from datetime import date
+from datetime import datetime
+from datetime import timedelta
+from decimal import Decimal
+from typing import Any
+from typing import cast
+from typing import Generic
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .operators import ADJACENT_TO
+from .operators import CONTAINED_BY
+from .operators import CONTAINS
+from .operators import NOT_EXTEND_LEFT_OF
+from .operators import NOT_EXTEND_RIGHT_OF
+from .operators import OVERLAP
+from .operators import STRICTLY_LEFT_OF
+from .operators import STRICTLY_RIGHT_OF
+from ... import types as sqltypes
+from ...sql import operators
+from ...sql.type_api import TypeEngine
+from ...util import py310
+from ...util.typing import Literal
+
+if TYPE_CHECKING:
+    from ...sql.elements import ColumnElement
+    from ...sql.type_api import _TE
+    from ...sql.type_api import TypeEngineMixin
+
+_T = TypeVar("_T", bound=Any)
+
+_BoundsType = Literal["()", "[)", "(]", "[]"]
+
+if py310:
+    dc_slots = {"slots": True}
+    dc_kwonly = {"kw_only": True}
+else:
+    dc_slots = {}
+    dc_kwonly = {}
+
+
+@dataclasses.dataclass(frozen=True, **dc_slots)
+class Range(Generic[_T]):
+    """Represent a PostgreSQL range.
+
+    E.g.::
+
+        r = Range(10, 50, bounds="()")
+
+    The calling style is similar to that of psycopg and psycopg2, in part
+    to allow easier migration from previous SQLAlchemy versions that used
+    these objects directly.
+
+    :param lower: Lower bound value, or None
+    :param upper: Upper bound value, or None
+    :param bounds: keyword-only, optional string value that is one of
+     ``"()"``, ``"[)"``, ``"(]"``, ``"[]"``.  Defaults to ``"[)"``.
+    :param empty: keyword-only, optional bool indicating this is an "empty"
+     range
+
+    .. versionadded:: 2.0
+
+    """
+
+    lower: Optional[_T] = None
+    """the lower bound"""
+
+    upper: Optional[_T] = None
+    """the upper bound"""
+
+    if TYPE_CHECKING:
+        bounds: _BoundsType = dataclasses.field(default="[)")
+        empty: bool = dataclasses.field(default=False)
+    else:
+        bounds: _BoundsType = dataclasses.field(default="[)", **dc_kwonly)
+        empty: bool = dataclasses.field(default=False, **dc_kwonly)
+
+    if not py310:
+
+        def __init__(
+            self,
+            lower: Optional[_T] = None,
+            upper: Optional[_T] = None,
+            *,
+            bounds: _BoundsType = "[)",
+            empty: bool = False,
+        ):
+            # no __slots__ either so we can update dict
+            self.__dict__.update(
+                {
+                    "lower": lower,
+                    "upper": upper,
+                    "bounds": bounds,
+                    "empty": empty,
+                }
+            )
+
+    def __bool__(self) -> bool:
+        return not self.empty
+
+    @property
+    def isempty(self) -> bool:
+        "A synonym for the 'empty' attribute."
+
+        return self.empty
+
+    @property
+    def is_empty(self) -> bool:
+        "A synonym for the 'empty' attribute."
+
+        return self.empty
+
+    @property
+    def lower_inc(self) -> bool:
+        """Return True if the lower bound is inclusive."""
+
+        return self.bounds[0] == "["
+
+    @property
+    def lower_inf(self) -> bool:
+        """Return True if this range is non-empty and lower bound is
+        infinite."""
+
+        return not self.empty and self.lower is None
+
+    @property
+    def upper_inc(self) -> bool:
+        """Return True if the upper bound is inclusive."""
+
+        return self.bounds[1] == "]"
+
+    @property
+    def upper_inf(self) -> bool:
+        """Return True if this range is non-empty and the upper bound is
+        infinite."""
+
+        return not self.empty and self.upper is None
+
+    @property
+    def __sa_type_engine__(self) -> AbstractSingleRange[_T]:
+        return AbstractSingleRange()
+
+    def _contains_value(self, value: _T) -> bool:
+        """Return True if this range contains the given value."""
+
+        if self.empty:
+            return False
+
+        if self.lower is None:
+            return self.upper is None or (
+                value < self.upper
+                if self.bounds[1] == ")"
+                else value <= self.upper
+            )
+
+        if self.upper is None:
+            return (  # type: ignore
+                value > self.lower
+                if self.bounds[0] == "("
+                else value >= self.lower
+            )
+
+        return (  # type: ignore
+            value > self.lower
+            if self.bounds[0] == "("
+            else value >= self.lower
+        ) and (
+            value < self.upper
+            if self.bounds[1] == ")"
+            else value <= self.upper
+        )
+
+    def _get_discrete_step(self) -> Any:
+        "Determine the “step” for this range, if it is a discrete one."
+
+        # See
+        # https://www.postgresql.org/docs/current/rangetypes.html#RANGETYPES-DISCRETE
+        # for the rationale
+
+        if isinstance(self.lower, int) or isinstance(self.upper, int):
+            return 1
+        elif isinstance(self.lower, datetime) or isinstance(
+            self.upper, datetime
+        ):
+            # This is required, because a `isinstance(datetime.now(), date)`
+            # is True
+            return None
+        elif isinstance(self.lower, date) or isinstance(self.upper, date):
+            return timedelta(days=1)
+        else:
+            return None
+
+    def _compare_edges(
+        self,
+        value1: Optional[_T],
+        bound1: str,
+        value2: Optional[_T],
+        bound2: str,
+        only_values: bool = False,
+    ) -> int:
+        """Compare two range bounds.
+
+        Return -1, 0 or 1 respectively when `value1` is less than,
+        equal to or greater than `value2`.
+
+        When `only_value` is ``True``, do not consider the *inclusivity*
+        of the edges, just their values.
+        """
+
+        value1_is_lower_bound = bound1 in {"[", "("}
+        value2_is_lower_bound = bound2 in {"[", "("}
+
+        # Infinite edges are equal when they are on the same side,
+        # otherwise a lower edge is considered less than the upper end
+        if value1 is value2 is None:
+            if value1_is_lower_bound == value2_is_lower_bound:
+                return 0
+            else:
+                return -1 if value1_is_lower_bound else 1
+        elif value1 is None:
+            return -1 if value1_is_lower_bound else 1
+        elif value2 is None:
+            return 1 if value2_is_lower_bound else -1
+
+        # Short path for trivial case
+        if bound1 == bound2 and value1 == value2:
+            return 0
+
+        value1_inc = bound1 in {"[", "]"}
+        value2_inc = bound2 in {"[", "]"}
+        step = self._get_discrete_step()
+
+        if step is not None:
+            # "Normalize" the two edges as '[)', to simplify successive
+            # logic when the range is discrete: otherwise we would need
+            # to handle the comparison between ``(0`` and ``[1`` that
+            # are equal when dealing with integers while for floats the
+            # former is lesser than the latter
+
+            if value1_is_lower_bound:
+                if not value1_inc:
+                    value1 += step
+                    value1_inc = True
+            else:
+                if value1_inc:
+                    value1 += step
+                    value1_inc = False
+            if value2_is_lower_bound:
+                if not value2_inc:
+                    value2 += step
+                    value2_inc = True
+            else:
+                if value2_inc:
+                    value2 += step
+                    value2_inc = False
+
+        if value1 < value2:  # type: ignore
+            return -1
+        elif value1 > value2:  # type: ignore
+            return 1
+        elif only_values:
+            return 0
+        else:
+            # Neither one is infinite but are equal, so we
+            # need to consider the respective inclusive/exclusive
+            # flag
+
+            if value1_inc and value2_inc:
+                return 0
+            elif not value1_inc and not value2_inc:
+                if value1_is_lower_bound == value2_is_lower_bound:
+                    return 0
+                else:
+                    return 1 if value1_is_lower_bound else -1
+            elif not value1_inc:
+                return 1 if value1_is_lower_bound else -1
+            elif not value2_inc:
+                return -1 if value2_is_lower_bound else 1
+            else:
+                return 0
+
+    def __eq__(self, other: Any) -> bool:
+        """Compare this range to the `other` taking into account
+        bounds inclusivity, returning ``True`` if they are equal.
+        """
+
+        if not isinstance(other, Range):
+            return NotImplemented
+
+        if self.empty and other.empty:
+            return True
+        elif self.empty != other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        return (
+            self._compare_edges(slower, slower_b, olower, olower_b) == 0
+            and self._compare_edges(supper, supper_b, oupper, oupper_b) == 0
+        )
+
+    def contained_by(self, other: Range[_T]) -> bool:
+        "Determine whether this range is a contained by `other`."
+
+        # Any range contains the empty one
+        if self.empty:
+            return True
+
+        # An empty range does not contain any range except the empty one
+        if other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        olower = other.lower
+        olower_b = other.bounds[0]
+
+        if self._compare_edges(slower, slower_b, olower, olower_b) < 0:
+            return False
+
+        supper = self.upper
+        supper_b = self.bounds[1]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0:
+            return False
+
+        return True
+
+    def contains(self, value: Union[_T, Range[_T]]) -> bool:
+        "Determine whether this range contains `value`."
+
+        if isinstance(value, Range):
+            return value.contained_by(self)
+        else:
+            return self._contains_value(value)
+
+    __contains__ = contains
+
+    def overlaps(self, other: Range[_T]) -> bool:
+        "Determine whether this range overlaps with `other`."
+
+        # Empty ranges never overlap with any other range
+        if self.empty or other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        # Check whether this lower bound is contained in the other range
+        if (
+            self._compare_edges(slower, slower_b, olower, olower_b) >= 0
+            and self._compare_edges(slower, slower_b, oupper, oupper_b) <= 0
+        ):
+            return True
+
+        # Check whether other lower bound is contained in this range
+        if (
+            self._compare_edges(olower, olower_b, slower, slower_b) >= 0
+            and self._compare_edges(olower, olower_b, supper, supper_b) <= 0
+        ):
+            return True
+
+        return False
+
+    def strictly_left_of(self, other: Range[_T]) -> bool:
+        "Determine whether this range is completely to the left of `other`."
+
+        # Empty ranges are neither to left nor to the right of any other range
+        if self.empty or other.empty:
+            return False
+
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+
+        # Check whether this upper edge is less than other's lower end
+        return self._compare_edges(supper, supper_b, olower, olower_b) < 0
+
+    __lshift__ = strictly_left_of
+
+    def strictly_right_of(self, other: Range[_T]) -> bool:
+        "Determine whether this range is completely to the right of `other`."
+
+        # Empty ranges are neither to left nor to the right of any other range
+        if self.empty or other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        # Check whether this lower edge is greater than other's upper end
+        return self._compare_edges(slower, slower_b, oupper, oupper_b) > 0
+
+    __rshift__ = strictly_right_of
+
+    def not_extend_left_of(self, other: Range[_T]) -> bool:
+        "Determine whether this does not extend to the left of `other`."
+
+        # Empty ranges are neither to left nor to the right of any other range
+        if self.empty or other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        olower = other.lower
+        olower_b = other.bounds[0]
+
+        # Check whether this lower edge is not less than other's lower end
+        return self._compare_edges(slower, slower_b, olower, olower_b) >= 0
+
+    def not_extend_right_of(self, other: Range[_T]) -> bool:
+        "Determine whether this does not extend to the right of `other`."
+
+        # Empty ranges are neither to left nor to the right of any other range
+        if self.empty or other.empty:
+            return False
+
+        supper = self.upper
+        supper_b = self.bounds[1]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        # Check whether this upper edge is not greater than other's upper end
+        return self._compare_edges(supper, supper_b, oupper, oupper_b) <= 0
+
+    def _upper_edge_adjacent_to_lower(
+        self,
+        value1: Optional[_T],
+        bound1: str,
+        value2: Optional[_T],
+        bound2: str,
+    ) -> bool:
+        """Determine whether an upper bound is immediately successive to a
+        lower bound."""
+
+        # Since we need a peculiar way to handle the bounds inclusivity,
+        # just do a comparison by value here
+        res = self._compare_edges(value1, bound1, value2, bound2, True)
+        if res == -1:
+            step = self._get_discrete_step()
+            if step is None:
+                return False
+            if bound1 == "]":
+                if bound2 == "[":
+                    return value1 == value2 - step  # type: ignore
+                else:
+                    return value1 == value2
+            else:
+                if bound2 == "[":
+                    return value1 == value2
+                else:
+                    return value1 == value2 - step  # type: ignore
+        elif res == 0:
+            # Cover cases like [0,0] -|- [1,] and [0,2) -|- (1,3]
+            if (
+                bound1 == "]"
+                and bound2 == "["
+                or bound1 == ")"
+                and bound2 == "("
+            ):
+                step = self._get_discrete_step()
+                if step is not None:
+                    return True
+            return (
+                bound1 == ")"
+                and bound2 == "["
+                or bound1 == "]"
+                and bound2 == "("
+            )
+        else:
+            return False
+
+    def adjacent_to(self, other: Range[_T]) -> bool:
+        "Determine whether this range is adjacent to the `other`."
+
+        # Empty ranges are not adjacent to any other range
+        if self.empty or other.empty:
+            return False
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        return self._upper_edge_adjacent_to_lower(
+            supper, supper_b, olower, olower_b
+        ) or self._upper_edge_adjacent_to_lower(
+            oupper, oupper_b, slower, slower_b
+        )
+
+    def union(self, other: Range[_T]) -> Range[_T]:
+        """Compute the union of this range with the `other`.
+
+        This raises a ``ValueError`` exception if the two ranges are
+        "disjunct", that is neither adjacent nor overlapping.
+        """
+
+        # Empty ranges are "additive identities"
+        if self.empty:
+            return other
+        if other.empty:
+            return self
+
+        if not self.overlaps(other) and not self.adjacent_to(other):
+            raise ValueError(
+                "Adding non-overlapping and non-adjacent"
+                " ranges is not implemented"
+            )
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        if self._compare_edges(slower, slower_b, olower, olower_b) < 0:
+            rlower = slower
+            rlower_b = slower_b
+        else:
+            rlower = olower
+            rlower_b = olower_b
+
+        if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0:
+            rupper = supper
+            rupper_b = supper_b
+        else:
+            rupper = oupper
+            rupper_b = oupper_b
+
+        return Range(
+            rlower, rupper, bounds=cast(_BoundsType, rlower_b + rupper_b)
+        )
+
+    def __add__(self, other: Range[_T]) -> Range[_T]:
+        return self.union(other)
+
+    def difference(self, other: Range[_T]) -> Range[_T]:
+        """Compute the difference between this range and the `other`.
+
+        This raises a ``ValueError`` exception if the two ranges are
+        "disjunct", that is neither adjacent nor overlapping.
+        """
+
+        # Subtracting an empty range is a no-op
+        if self.empty or other.empty:
+            return self
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        sl_vs_ol = self._compare_edges(slower, slower_b, olower, olower_b)
+        su_vs_ou = self._compare_edges(supper, supper_b, oupper, oupper_b)
+        if sl_vs_ol < 0 and su_vs_ou > 0:
+            raise ValueError(
+                "Subtracting a strictly inner range is not implemented"
+            )
+
+        sl_vs_ou = self._compare_edges(slower, slower_b, oupper, oupper_b)
+        su_vs_ol = self._compare_edges(supper, supper_b, olower, olower_b)
+
+        # If the ranges do not overlap, result is simply the first
+        if sl_vs_ou > 0 or su_vs_ol < 0:
+            return self
+
+        # If this range is completely contained by the other, result is empty
+        if sl_vs_ol >= 0 and su_vs_ou <= 0:
+            return Range(None, None, empty=True)
+
+        # If this range extends to the left of the other and ends in its
+        # middle
+        if sl_vs_ol <= 0 and su_vs_ol >= 0 and su_vs_ou <= 0:
+            rupper_b = ")" if olower_b == "[" else "]"
+            if (
+                slower_b != "["
+                and rupper_b != "]"
+                and self._compare_edges(slower, slower_b, olower, rupper_b)
+                == 0
+            ):
+                return Range(None, None, empty=True)
+            else:
+                return Range(
+                    slower,
+                    olower,
+                    bounds=cast(_BoundsType, slower_b + rupper_b),
+                )
+
+        # If this range starts in the middle of the other and extends to its
+        # right
+        if sl_vs_ol >= 0 and su_vs_ou >= 0 and sl_vs_ou <= 0:
+            rlower_b = "(" if oupper_b == "]" else "["
+            if (
+                rlower_b != "["
+                and supper_b != "]"
+                and self._compare_edges(oupper, rlower_b, supper, supper_b)
+                == 0
+            ):
+                return Range(None, None, empty=True)
+            else:
+                return Range(
+                    oupper,
+                    supper,
+                    bounds=cast(_BoundsType, rlower_b + supper_b),
+                )
+
+        assert False, f"Unhandled case computing {self} - {other}"
+
+    def __sub__(self, other: Range[_T]) -> Range[_T]:
+        return self.difference(other)
+
+    def intersection(self, other: Range[_T]) -> Range[_T]:
+        """Compute the intersection of this range with the `other`.
+
+        .. versionadded:: 2.0.10
+
+        """
+        if self.empty or other.empty or not self.overlaps(other):
+            return Range(None, None, empty=True)
+
+        slower = self.lower
+        slower_b = self.bounds[0]
+        supper = self.upper
+        supper_b = self.bounds[1]
+        olower = other.lower
+        olower_b = other.bounds[0]
+        oupper = other.upper
+        oupper_b = other.bounds[1]
+
+        if self._compare_edges(slower, slower_b, olower, olower_b) < 0:
+            rlower = olower
+            rlower_b = olower_b
+        else:
+            rlower = slower
+            rlower_b = slower_b
+
+        if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0:
+            rupper = oupper
+            rupper_b = oupper_b
+        else:
+            rupper = supper
+            rupper_b = supper_b
+
+        return Range(
+            rlower,
+            rupper,
+            bounds=cast(_BoundsType, rlower_b + rupper_b),
+        )
+
+    def __mul__(self, other: Range[_T]) -> Range[_T]:
+        return self.intersection(other)
+
+    def __str__(self) -> str:
+        return self._stringify()
+
+    def _stringify(self) -> str:
+        if self.empty:
+            return "empty"
+
+        l, r = self.lower, self.upper
+        l = "" if l is None else l  # type: ignore
+        r = "" if r is None else r  # type: ignore
+
+        b0, b1 = cast("Tuple[str, str]", self.bounds)
+
+        return f"{b0}{l},{r}{b1}"
+
+
+class MultiRange(List[Range[_T]]):
+    """Represents a multirange sequence.
+
+    This list subclass is an utility to allow automatic type inference of
+    the proper multi-range SQL type depending on the single range values.
+    This is useful when operating on literal multi-ranges::
+
+        import sqlalchemy as sa
+        from sqlalchemy.dialects.postgresql import MultiRange, Range
+
+        value = literal(MultiRange([Range(2, 4)]))
+
+        select(tbl).where(tbl.c.value.op("@")(MultiRange([Range(-3, 7)])))
+
+    .. versionadded:: 2.0.26
+
+    .. seealso::
+
+        - :ref:`postgresql_multirange_list_use`.
+    """
+
+    @property
+    def __sa_type_engine__(self) -> AbstractMultiRange[_T]:
+        return AbstractMultiRange()
+
+
+class AbstractRange(sqltypes.TypeEngine[_T]):
+    """Base class for single and multi Range SQL types."""
+
+    render_bind_cast = True
+
+    __abstract__ = True
+
+    @overload
+    def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ...
+
+    @overload
+    def adapt(
+        self, cls: Type[TypeEngineMixin], **kw: Any
+    ) -> TypeEngine[Any]: ...
+
+    def adapt(
+        self,
+        cls: Type[Union[TypeEngine[Any], TypeEngineMixin]],
+        **kw: Any,
+    ) -> TypeEngine[Any]:
+        """Dynamically adapt a range type to an abstract impl.
+
+        For example ``INT4RANGE().adapt(_Psycopg2NumericRange)`` should
+        produce a type that will have ``_Psycopg2NumericRange`` behaviors
+        and also render as ``INT4RANGE`` in SQL and DDL.
+
+        """
+        if (
+            issubclass(cls, (AbstractSingleRangeImpl, AbstractMultiRangeImpl))
+            and cls is not self.__class__
+        ):
+            # two ways to do this are:  1. create a new type on the fly
+            # or 2. have AbstractRangeImpl(visit_name) constructor and a
+            # visit_abstract_range_impl() method in the PG compiler.
+            # I'm choosing #1 as the resulting type object
+            # will then make use of the same mechanics
+            # as if we had made all these sub-types explicitly, and will
+            # also look more obvious under pdb etc.
+            # The adapt() operation here is cached per type-class-per-dialect,
+            # so is not much of a performance concern
+            visit_name = self.__visit_name__
+            return type(  # type: ignore
+                f"{visit_name}RangeImpl",
+                (cls, self.__class__),
+                {"__visit_name__": visit_name},
+            )()
+        else:
+            return super().adapt(cls)
+
+    class comparator_factory(TypeEngine.Comparator[Range[Any]]):
+        """Define comparison operations for range types."""
+
+        def contains(self, other: Any, **kw: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the right hand operand,
+            which can be an element or a range, is contained within the
+            column.
+
+            kwargs may be ignored by this operator but are required for API
+            conformance.
+            """
+            return self.expr.operate(CONTAINS, other)
+
+        def contained_by(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the column is contained
+            within the right hand operand.
+            """
+            return self.expr.operate(CONTAINED_BY, other)
+
+        def overlaps(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the column overlaps
+            (has points in common with) the right hand operand.
+            """
+            return self.expr.operate(OVERLAP, other)
+
+        def strictly_left_of(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the column is strictly
+            left of the right hand operand.
+            """
+            return self.expr.operate(STRICTLY_LEFT_OF, other)
+
+        __lshift__ = strictly_left_of
+
+        def strictly_right_of(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the column is strictly
+            right of the right hand operand.
+            """
+            return self.expr.operate(STRICTLY_RIGHT_OF, other)
+
+        __rshift__ = strictly_right_of
+
+        def not_extend_right_of(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the range in the column
+            does not extend right of the range in the operand.
+            """
+            return self.expr.operate(NOT_EXTEND_RIGHT_OF, other)
+
+        def not_extend_left_of(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the range in the column
+            does not extend left of the range in the operand.
+            """
+            return self.expr.operate(NOT_EXTEND_LEFT_OF, other)
+
+        def adjacent_to(self, other: Any) -> ColumnElement[bool]:
+            """Boolean expression. Returns true if the range in the column
+            is adjacent to the range in the operand.
+            """
+            return self.expr.operate(ADJACENT_TO, other)
+
+        def union(self, other: Any) -> ColumnElement[bool]:
+            """Range expression. Returns the union of the two ranges.
+            Will raise an exception if the resulting range is not
+            contiguous.
+            """
+            return self.expr.operate(operators.add, other)
+
+        def difference(self, other: Any) -> ColumnElement[bool]:
+            """Range expression. Returns the union of the two ranges.
+            Will raise an exception if the resulting range is not
+            contiguous.
+            """
+            return self.expr.operate(operators.sub, other)
+
+        def intersection(self, other: Any) -> ColumnElement[Range[_T]]:
+            """Range expression. Returns the intersection of the two ranges.
+            Will raise an exception if the resulting range is not
+            contiguous.
+            """
+            return self.expr.operate(operators.mul, other)
+
+
+class AbstractSingleRange(AbstractRange[Range[_T]]):
+    """Base for PostgreSQL RANGE types.
+
+    These are types that return a single :class:`_postgresql.Range` object.
+
+    .. seealso::
+
+        `PostgreSQL range functions <https://www.postgresql.org/docs/current/static/functions-range.html>`_
+
+    """  # noqa: E501
+
+    __abstract__ = True
+
+    def _resolve_for_literal(self, value: Range[Any]) -> Any:
+        spec = value.lower if value.lower is not None else value.upper
+
+        if isinstance(spec, int):
+            # pg is unreasonably picky here: the query
+            # "select 1::INTEGER <@ '[1, 4)'::INT8RANGE" raises
+            # "operator does not exist: integer <@ int8range" as of pg 16
+            if _is_int32(value):
+                return INT4RANGE()
+            else:
+                return INT8RANGE()
+        elif isinstance(spec, (Decimal, float)):
+            return NUMRANGE()
+        elif isinstance(spec, datetime):
+            return TSRANGE() if not spec.tzinfo else TSTZRANGE()
+        elif isinstance(spec, date):
+            return DATERANGE()
+        else:
+            # empty Range, SQL datatype can't be determined here
+            return sqltypes.NULLTYPE
+
+
+class AbstractSingleRangeImpl(AbstractSingleRange[_T]):
+    """Marker for AbstractSingleRange that will apply a subclass-specific
+    adaptation"""
+
+
+class AbstractMultiRange(AbstractRange[Sequence[Range[_T]]]):
+    """Base for PostgreSQL MULTIRANGE types.
+
+    these are types that return a sequence of :class:`_postgresql.Range`
+    objects.
+
+    """
+
+    __abstract__ = True
+
+    def _resolve_for_literal(self, value: Sequence[Range[Any]]) -> Any:
+        if not value:
+            # empty MultiRange, SQL datatype can't be determined here
+            return sqltypes.NULLTYPE
+        first = value[0]
+        spec = first.lower if first.lower is not None else first.upper
+
+        if isinstance(spec, int):
+            # pg is unreasonably picky here: the query
+            # "select 1::INTEGER <@ '{[1, 4),[6,19)}'::INT8MULTIRANGE" raises
+            # "operator does not exist: integer <@ int8multirange" as of pg 16
+            if all(_is_int32(r) for r in value):
+                return INT4MULTIRANGE()
+            else:
+                return INT8MULTIRANGE()
+        elif isinstance(spec, (Decimal, float)):
+            return NUMMULTIRANGE()
+        elif isinstance(spec, datetime):
+            return TSMULTIRANGE() if not spec.tzinfo else TSTZMULTIRANGE()
+        elif isinstance(spec, date):
+            return DATEMULTIRANGE()
+        else:
+            # empty Range, SQL datatype can't be determined here
+            return sqltypes.NULLTYPE
+
+
+class AbstractMultiRangeImpl(AbstractMultiRange[_T]):
+    """Marker for AbstractMultiRange that will apply a subclass-specific
+    adaptation"""
+
+
+class INT4RANGE(AbstractSingleRange[int]):
+    """Represent the PostgreSQL INT4RANGE type."""
+
+    __visit_name__ = "INT4RANGE"
+
+
+class INT8RANGE(AbstractSingleRange[int]):
+    """Represent the PostgreSQL INT8RANGE type."""
+
+    __visit_name__ = "INT8RANGE"
+
+
+class NUMRANGE(AbstractSingleRange[Decimal]):
+    """Represent the PostgreSQL NUMRANGE type."""
+
+    __visit_name__ = "NUMRANGE"
+
+
+class DATERANGE(AbstractSingleRange[date]):
+    """Represent the PostgreSQL DATERANGE type."""
+
+    __visit_name__ = "DATERANGE"
+
+
+class TSRANGE(AbstractSingleRange[datetime]):
+    """Represent the PostgreSQL TSRANGE type."""
+
+    __visit_name__ = "TSRANGE"
+
+
+class TSTZRANGE(AbstractSingleRange[datetime]):
+    """Represent the PostgreSQL TSTZRANGE type."""
+
+    __visit_name__ = "TSTZRANGE"
+
+
+class INT4MULTIRANGE(AbstractMultiRange[int]):
+    """Represent the PostgreSQL INT4MULTIRANGE type."""
+
+    __visit_name__ = "INT4MULTIRANGE"
+
+
+class INT8MULTIRANGE(AbstractMultiRange[int]):
+    """Represent the PostgreSQL INT8MULTIRANGE type."""
+
+    __visit_name__ = "INT8MULTIRANGE"
+
+
+class NUMMULTIRANGE(AbstractMultiRange[Decimal]):
+    """Represent the PostgreSQL NUMMULTIRANGE type."""
+
+    __visit_name__ = "NUMMULTIRANGE"
+
+
+class DATEMULTIRANGE(AbstractMultiRange[date]):
+    """Represent the PostgreSQL DATEMULTIRANGE type."""
+
+    __visit_name__ = "DATEMULTIRANGE"
+
+
+class TSMULTIRANGE(AbstractMultiRange[datetime]):
+    """Represent the PostgreSQL TSRANGE type."""
+
+    __visit_name__ = "TSMULTIRANGE"
+
+
+class TSTZMULTIRANGE(AbstractMultiRange[datetime]):
+    """Represent the PostgreSQL TSTZRANGE type."""
+
+    __visit_name__ = "TSTZMULTIRANGE"
+
+
+_max_int_32 = 2**31 - 1
+_min_int_32 = -(2**31)
+
+
+def _is_int32(r: Range[int]) -> bool:
+    return (r.lower is None or _min_int_32 <= r.lower <= _max_int_32) and (
+        r.upper is None or _min_int_32 <= r.upper <= _max_int_32
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/types.py
new file mode 100644
index 00000000..6fe4f576
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/postgresql/types.py
@@ -0,0 +1,303 @@
+# dialects/postgresql/types.py
+# Copyright (C) 2013-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import datetime as dt
+from typing import Any
+from typing import Optional
+from typing import overload
+from typing import Type
+from typing import TYPE_CHECKING
+from uuid import UUID as _python_UUID
+
+from ...sql import sqltypes
+from ...sql import type_api
+from ...util.typing import Literal
+
+if TYPE_CHECKING:
+    from ...engine.interfaces import Dialect
+    from ...sql.operators import OperatorType
+    from ...sql.type_api import _LiteralProcessorType
+    from ...sql.type_api import TypeEngine
+
+_DECIMAL_TYPES = (1231, 1700)
+_FLOAT_TYPES = (700, 701, 1021, 1022)
+_INT_TYPES = (20, 21, 23, 26, 1005, 1007, 1016)
+
+
+class PGUuid(sqltypes.UUID[sqltypes._UUID_RETURN]):
+    render_bind_cast = True
+    render_literal_cast = True
+
+    if TYPE_CHECKING:
+
+        @overload
+        def __init__(
+            self: PGUuid[_python_UUID], as_uuid: Literal[True] = ...
+        ) -> None: ...
+
+        @overload
+        def __init__(
+            self: PGUuid[str], as_uuid: Literal[False] = ...
+        ) -> None: ...
+
+        def __init__(self, as_uuid: bool = True) -> None: ...
+
+
+class BYTEA(sqltypes.LargeBinary):
+    __visit_name__ = "BYTEA"
+
+
+class INET(sqltypes.TypeEngine[str]):
+    __visit_name__ = "INET"
+
+
+PGInet = INET
+
+
+class CIDR(sqltypes.TypeEngine[str]):
+    __visit_name__ = "CIDR"
+
+
+PGCidr = CIDR
+
+
+class MACADDR(sqltypes.TypeEngine[str]):
+    __visit_name__ = "MACADDR"
+
+
+PGMacAddr = MACADDR
+
+
+class MACADDR8(sqltypes.TypeEngine[str]):
+    __visit_name__ = "MACADDR8"
+
+
+PGMacAddr8 = MACADDR8
+
+
+class MONEY(sqltypes.TypeEngine[str]):
+    r"""Provide the PostgreSQL MONEY type.
+
+    Depending on driver, result rows using this type may return a
+    string value which includes currency symbols.
+
+    For this reason, it may be preferable to provide conversion to a
+    numerically-based currency datatype using :class:`_types.TypeDecorator`::
+
+        import re
+        import decimal
+        from sqlalchemy import Dialect
+        from sqlalchemy import TypeDecorator
+
+
+        class NumericMoney(TypeDecorator):
+            impl = MONEY
+
+            def process_result_value(self, value: Any, dialect: Dialect) -> None:
+                if value is not None:
+                    # adjust this for the currency and numeric
+                    m = re.match(r"\$([\d.]+)", value)
+                    if m:
+                        value = decimal.Decimal(m.group(1))
+                return value
+
+    Alternatively, the conversion may be applied as a CAST using
+    the :meth:`_types.TypeDecorator.column_expression` method as follows::
+
+        import decimal
+        from sqlalchemy import cast
+        from sqlalchemy import TypeDecorator
+
+
+        class NumericMoney(TypeDecorator):
+            impl = MONEY
+
+            def column_expression(self, column: Any):
+                return cast(column, Numeric())
+
+    .. versionadded:: 1.2
+
+    """  # noqa: E501
+
+    __visit_name__ = "MONEY"
+
+
+class OID(sqltypes.TypeEngine[int]):
+    """Provide the PostgreSQL OID type."""
+
+    __visit_name__ = "OID"
+
+
+class REGCONFIG(sqltypes.TypeEngine[str]):
+    """Provide the PostgreSQL REGCONFIG type.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    __visit_name__ = "REGCONFIG"
+
+
+class TSQUERY(sqltypes.TypeEngine[str]):
+    """Provide the PostgreSQL TSQUERY type.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    __visit_name__ = "TSQUERY"
+
+
+class REGCLASS(sqltypes.TypeEngine[str]):
+    """Provide the PostgreSQL REGCLASS type.
+
+    .. versionadded:: 1.2.7
+
+    """
+
+    __visit_name__ = "REGCLASS"
+
+
+class TIMESTAMP(sqltypes.TIMESTAMP):
+    """Provide the PostgreSQL TIMESTAMP type."""
+
+    __visit_name__ = "TIMESTAMP"
+
+    def __init__(
+        self, timezone: bool = False, precision: Optional[int] = None
+    ) -> None:
+        """Construct a TIMESTAMP.
+
+        :param timezone: boolean value if timezone present, default False
+        :param precision: optional integer precision value
+
+         .. versionadded:: 1.4
+
+        """
+        super().__init__(timezone=timezone)
+        self.precision = precision
+
+
+class TIME(sqltypes.TIME):
+    """PostgreSQL TIME type."""
+
+    __visit_name__ = "TIME"
+
+    def __init__(
+        self, timezone: bool = False, precision: Optional[int] = None
+    ) -> None:
+        """Construct a TIME.
+
+        :param timezone: boolean value if timezone present, default False
+        :param precision: optional integer precision value
+
+         .. versionadded:: 1.4
+
+        """
+        super().__init__(timezone=timezone)
+        self.precision = precision
+
+
+class INTERVAL(type_api.NativeForEmulated, sqltypes._AbstractInterval):
+    """PostgreSQL INTERVAL type."""
+
+    __visit_name__ = "INTERVAL"
+    native = True
+
+    def __init__(
+        self, precision: Optional[int] = None, fields: Optional[str] = None
+    ) -> None:
+        """Construct an INTERVAL.
+
+        :param precision: optional integer precision value
+        :param fields: string fields specifier.  allows storage of fields
+         to be limited, such as ``"YEAR"``, ``"MONTH"``, ``"DAY TO HOUR"``,
+         etc.
+
+         .. versionadded:: 1.2
+
+        """
+        self.precision = precision
+        self.fields = fields
+
+    @classmethod
+    def adapt_emulated_to_native(
+        cls, interval: sqltypes.Interval, **kw: Any  # type: ignore[override]
+    ) -> INTERVAL:
+        return INTERVAL(precision=interval.second_precision)
+
+    @property
+    def _type_affinity(self) -> Type[sqltypes.Interval]:
+        return sqltypes.Interval
+
+    def as_generic(self, allow_nulltype: bool = False) -> sqltypes.Interval:
+        return sqltypes.Interval(native=True, second_precision=self.precision)
+
+    @property
+    def python_type(self) -> Type[dt.timedelta]:
+        return dt.timedelta
+
+    def literal_processor(
+        self, dialect: Dialect
+    ) -> Optional[_LiteralProcessorType[dt.timedelta]]:
+        def process(value: dt.timedelta) -> str:
+            return f"make_interval(secs=>{value.total_seconds()})"
+
+        return process
+
+
+PGInterval = INTERVAL
+
+
+class BIT(sqltypes.TypeEngine[int]):
+    __visit_name__ = "BIT"
+
+    def __init__(
+        self, length: Optional[int] = None, varying: bool = False
+    ) -> None:
+        if varying:
+            # BIT VARYING can be unlimited-length, so no default
+            self.length = length
+        else:
+            # BIT without VARYING defaults to length 1
+            self.length = length or 1
+        self.varying = varying
+
+
+PGBit = BIT
+
+
+class TSVECTOR(sqltypes.TypeEngine[str]):
+    """The :class:`_postgresql.TSVECTOR` type implements the PostgreSQL
+    text search type TSVECTOR.
+
+    It can be used to do full text queries on natural language
+    documents.
+
+    .. seealso::
+
+        :ref:`postgresql_match`
+
+    """
+
+    __visit_name__ = "TSVECTOR"
+
+
+class CITEXT(sqltypes.TEXT):
+    """Provide the PostgreSQL CITEXT type.
+
+    .. versionadded:: 2.0.7
+
+    """
+
+    __visit_name__ = "CITEXT"
+
+    def coerce_compared_value(
+        self, op: Optional[OperatorType], value: Any
+    ) -> TypeEngine[Any]:
+        return self
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__init__.py
new file mode 100644
index 00000000..7b381fa6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__init__.py
@@ -0,0 +1,57 @@
+# dialects/sqlite/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from . import aiosqlite  # noqa
+from . import base  # noqa
+from . import pysqlcipher  # noqa
+from . import pysqlite  # noqa
+from .base import BLOB
+from .base import BOOLEAN
+from .base import CHAR
+from .base import DATE
+from .base import DATETIME
+from .base import DECIMAL
+from .base import FLOAT
+from .base import INTEGER
+from .base import JSON
+from .base import NUMERIC
+from .base import REAL
+from .base import SMALLINT
+from .base import TEXT
+from .base import TIME
+from .base import TIMESTAMP
+from .base import VARCHAR
+from .dml import Insert
+from .dml import insert
+
+# default dialect
+base.dialect = dialect = pysqlite.dialect
+
+
+__all__ = (
+    "BLOB",
+    "BOOLEAN",
+    "CHAR",
+    "DATE",
+    "DATETIME",
+    "DECIMAL",
+    "FLOAT",
+    "INTEGER",
+    "JSON",
+    "NUMERIC",
+    "SMALLINT",
+    "TEXT",
+    "TIME",
+    "TIMESTAMP",
+    "VARCHAR",
+    "REAL",
+    "Insert",
+    "insert",
+    "dialect",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..380a8984
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/aiosqlite.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/aiosqlite.cpython-312.pyc
new file mode 100644
index 00000000..243088d9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/aiosqlite.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..ae47291a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/dml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/dml.cpython-312.pyc
new file mode 100644
index 00000000..3f523f41
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/dml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/json.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/json.cpython-312.pyc
new file mode 100644
index 00000000..d027d8b4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/json.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..54a7a966
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlcipher.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlcipher.cpython-312.pyc
new file mode 100644
index 00000000..cafa62c7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlcipher.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlite.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlite.cpython-312.pyc
new file mode 100644
index 00000000..933c086d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/__pycache__/pysqlite.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/aiosqlite.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/aiosqlite.py
new file mode 100644
index 00000000..82802245
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/aiosqlite.py
@@ -0,0 +1,421 @@
+# dialects/sqlite/aiosqlite.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+
+.. dialect:: sqlite+aiosqlite
+    :name: aiosqlite
+    :dbapi: aiosqlite
+    :connectstring: sqlite+aiosqlite:///file_path
+    :url: https://pypi.org/project/aiosqlite/
+
+The aiosqlite dialect provides support for the SQLAlchemy asyncio interface
+running on top of pysqlite.
+
+aiosqlite is a wrapper around pysqlite that uses a background thread for
+each connection.   It does not actually use non-blocking IO, as SQLite
+databases are not socket-based.  However it does provide a working asyncio
+interface that's useful for testing and prototyping purposes.
+
+Using a special asyncio mediation layer, the aiosqlite dialect is usable
+as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>`
+extension package.
+
+This dialect should normally be used only with the
+:func:`_asyncio.create_async_engine` engine creation function::
+
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine("sqlite+aiosqlite:///filename")
+
+The URL passes through all arguments to the ``pysqlite`` driver, so all
+connection arguments are the same as they are for that of :ref:`pysqlite`.
+
+.. _aiosqlite_udfs:
+
+User-Defined Functions
+----------------------
+
+aiosqlite extends pysqlite to support async, so we can create our own user-defined functions (UDFs)
+in Python and use them directly in SQLite queries as described here: :ref:`pysqlite_udfs`.
+
+.. _aiosqlite_serializable:
+
+Serializable isolation / Savepoints / Transactional DDL (asyncio version)
+-------------------------------------------------------------------------
+
+Similarly to pysqlite, aiosqlite does not support SAVEPOINT feature.
+
+The solution is similar to :ref:`pysqlite_serializable`. This is achieved by the event listeners in async::
+
+    from sqlalchemy import create_engine, event
+    from sqlalchemy.ext.asyncio import create_async_engine
+
+    engine = create_async_engine("sqlite+aiosqlite:///myfile.db")
+
+
+    @event.listens_for(engine.sync_engine, "connect")
+    def do_connect(dbapi_connection, connection_record):
+        # disable aiosqlite's emitting of the BEGIN statement entirely.
+        # also stops it from emitting COMMIT before any DDL.
+        dbapi_connection.isolation_level = None
+
+
+    @event.listens_for(engine.sync_engine, "begin")
+    def do_begin(conn):
+        # emit our own BEGIN
+        conn.exec_driver_sql("BEGIN")
+
+.. warning:: When using the above recipe, it is advised to not use the
+   :paramref:`.Connection.execution_options.isolation_level` setting on
+   :class:`_engine.Connection` and :func:`_sa.create_engine`
+   with the SQLite driver,
+   as this function necessarily will also alter the ".isolation_level" setting.
+
+.. _aiosqlite_pooling:
+
+Pooling Behavior
+----------------
+
+The SQLAlchemy ``aiosqlite`` DBAPI establishes the connection pool differently
+based on the kind of SQLite database that's requested:
+
+* When a ``:memory:`` SQLite database is specified, the dialect by default
+  will use :class:`.StaticPool`. This pool maintains a single
+  connection, so that all access to the engine
+  use the same ``:memory:`` database.
+* When a file-based database is specified, the dialect will use
+  :class:`.AsyncAdaptedQueuePool` as the source of connections.
+
+  .. versionchanged:: 2.0.38
+
+    SQLite file database engines now use :class:`.AsyncAdaptedQueuePool` by default.
+    Previously, :class:`.NullPool` were used.  The :class:`.NullPool` class
+    may be used by specifying it via the
+    :paramref:`_sa.create_engine.poolclass` parameter.
+
+"""  # noqa
+
+import asyncio
+from collections import deque
+from functools import partial
+
+from .base import SQLiteExecutionContext
+from .pysqlite import SQLiteDialect_pysqlite
+from ... import pool
+from ... import util
+from ...engine import AdaptedConnection
+from ...util.concurrency import await_fallback
+from ...util.concurrency import await_only
+
+
+class AsyncAdapt_aiosqlite_cursor:
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+
+    __slots__ = (
+        "_adapt_connection",
+        "_connection",
+        "description",
+        "await_",
+        "_rows",
+        "arraysize",
+        "rowcount",
+        "lastrowid",
+    )
+
+    server_side = False
+
+    def __init__(self, adapt_connection):
+        self._adapt_connection = adapt_connection
+        self._connection = adapt_connection._connection
+        self.await_ = adapt_connection.await_
+        self.arraysize = 1
+        self.rowcount = -1
+        self.description = None
+        self._rows = deque()
+
+    def close(self):
+        self._rows.clear()
+
+    def execute(self, operation, parameters=None):
+        try:
+            _cursor = self.await_(self._connection.cursor())
+
+            if parameters is None:
+                self.await_(_cursor.execute(operation))
+            else:
+                self.await_(_cursor.execute(operation, parameters))
+
+            if _cursor.description:
+                self.description = _cursor.description
+                self.lastrowid = self.rowcount = -1
+
+                if not self.server_side:
+                    self._rows = deque(self.await_(_cursor.fetchall()))
+            else:
+                self.description = None
+                self.lastrowid = _cursor.lastrowid
+                self.rowcount = _cursor.rowcount
+
+            if not self.server_side:
+                self.await_(_cursor.close())
+            else:
+                self._cursor = _cursor
+        except Exception as error:
+            self._adapt_connection._handle_exception(error)
+
+    def executemany(self, operation, seq_of_parameters):
+        try:
+            _cursor = self.await_(self._connection.cursor())
+            self.await_(_cursor.executemany(operation, seq_of_parameters))
+            self.description = None
+            self.lastrowid = _cursor.lastrowid
+            self.rowcount = _cursor.rowcount
+            self.await_(_cursor.close())
+        except Exception as error:
+            self._adapt_connection._handle_exception(error)
+
+    def setinputsizes(self, *inputsizes):
+        pass
+
+    def __iter__(self):
+        while self._rows:
+            yield self._rows.popleft()
+
+    def fetchone(self):
+        if self._rows:
+            return self._rows.popleft()
+        else:
+            return None
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+
+        rr = self._rows
+        return [rr.popleft() for _ in range(min(size, len(rr)))]
+
+    def fetchall(self):
+        retval = list(self._rows)
+        self._rows.clear()
+        return retval
+
+
+class AsyncAdapt_aiosqlite_ss_cursor(AsyncAdapt_aiosqlite_cursor):
+    # TODO: base on connectors/asyncio.py
+    # see #10415
+    __slots__ = "_cursor"
+
+    server_side = True
+
+    def __init__(self, *arg, **kw):
+        super().__init__(*arg, **kw)
+        self._cursor = None
+
+    def close(self):
+        if self._cursor is not None:
+            self.await_(self._cursor.close())
+            self._cursor = None
+
+    def fetchone(self):
+        return self.await_(self._cursor.fetchone())
+
+    def fetchmany(self, size=None):
+        if size is None:
+            size = self.arraysize
+        return self.await_(self._cursor.fetchmany(size=size))
+
+    def fetchall(self):
+        return self.await_(self._cursor.fetchall())
+
+
+class AsyncAdapt_aiosqlite_connection(AdaptedConnection):
+    await_ = staticmethod(await_only)
+    __slots__ = ("dbapi",)
+
+    def __init__(self, dbapi, connection):
+        self.dbapi = dbapi
+        self._connection = connection
+
+    @property
+    def isolation_level(self):
+        return self._connection.isolation_level
+
+    @isolation_level.setter
+    def isolation_level(self, value):
+        # aiosqlite's isolation_level setter works outside the Thread
+        # that it's supposed to, necessitating setting check_same_thread=False.
+        # for improved stability, we instead invent our own awaitable version
+        # using aiosqlite's async queue directly.
+
+        def set_iso(connection, value):
+            connection.isolation_level = value
+
+        function = partial(set_iso, self._connection._conn, value)
+        future = asyncio.get_event_loop().create_future()
+
+        self._connection._tx.put_nowait((future, function))
+
+        try:
+            return self.await_(future)
+        except Exception as error:
+            self._handle_exception(error)
+
+    def create_function(self, *args, **kw):
+        try:
+            self.await_(self._connection.create_function(*args, **kw))
+        except Exception as error:
+            self._handle_exception(error)
+
+    def cursor(self, server_side=False):
+        if server_side:
+            return AsyncAdapt_aiosqlite_ss_cursor(self)
+        else:
+            return AsyncAdapt_aiosqlite_cursor(self)
+
+    def execute(self, *args, **kw):
+        return self.await_(self._connection.execute(*args, **kw))
+
+    def rollback(self):
+        try:
+            self.await_(self._connection.rollback())
+        except Exception as error:
+            self._handle_exception(error)
+
+    def commit(self):
+        try:
+            self.await_(self._connection.commit())
+        except Exception as error:
+            self._handle_exception(error)
+
+    def close(self):
+        try:
+            self.await_(self._connection.close())
+        except ValueError:
+            # this is undocumented for aiosqlite, that ValueError
+            # was raised if .close() was called more than once, which is
+            # both not customary for DBAPI and is also not a DBAPI.Error
+            # exception. This is now fixed in aiosqlite via my PR
+            # https://github.com/omnilib/aiosqlite/pull/238, so we can be
+            # assured this will not become some other kind of exception,
+            # since it doesn't raise anymore.
+
+            pass
+        except Exception as error:
+            self._handle_exception(error)
+
+    def _handle_exception(self, error):
+        if (
+            isinstance(error, ValueError)
+            and error.args[0] == "no active connection"
+        ):
+            raise self.dbapi.sqlite.OperationalError(
+                "no active connection"
+            ) from error
+        else:
+            raise error
+
+
+class AsyncAdaptFallback_aiosqlite_connection(AsyncAdapt_aiosqlite_connection):
+    __slots__ = ()
+
+    await_ = staticmethod(await_fallback)
+
+
+class AsyncAdapt_aiosqlite_dbapi:
+    def __init__(self, aiosqlite, sqlite):
+        self.aiosqlite = aiosqlite
+        self.sqlite = sqlite
+        self.paramstyle = "qmark"
+        self._init_dbapi_attributes()
+
+    def _init_dbapi_attributes(self):
+        for name in (
+            "DatabaseError",
+            "Error",
+            "IntegrityError",
+            "NotSupportedError",
+            "OperationalError",
+            "ProgrammingError",
+            "sqlite_version",
+            "sqlite_version_info",
+        ):
+            setattr(self, name, getattr(self.aiosqlite, name))
+
+        for name in ("PARSE_COLNAMES", "PARSE_DECLTYPES"):
+            setattr(self, name, getattr(self.sqlite, name))
+
+        for name in ("Binary",):
+            setattr(self, name, getattr(self.sqlite, name))
+
+    def connect(self, *arg, **kw):
+        async_fallback = kw.pop("async_fallback", False)
+
+        creator_fn = kw.pop("async_creator_fn", None)
+        if creator_fn:
+            connection = creator_fn(*arg, **kw)
+        else:
+            connection = self.aiosqlite.connect(*arg, **kw)
+            # it's a Thread.   you'll thank us later
+            connection.daemon = True
+
+        if util.asbool(async_fallback):
+            return AsyncAdaptFallback_aiosqlite_connection(
+                self,
+                await_fallback(connection),
+            )
+        else:
+            return AsyncAdapt_aiosqlite_connection(
+                self,
+                await_only(connection),
+            )
+
+
+class SQLiteExecutionContext_aiosqlite(SQLiteExecutionContext):
+    def create_server_side_cursor(self):
+        return self._dbapi_connection.cursor(server_side=True)
+
+
+class SQLiteDialect_aiosqlite(SQLiteDialect_pysqlite):
+    driver = "aiosqlite"
+    supports_statement_cache = True
+
+    is_async = True
+
+    supports_server_side_cursors = True
+
+    execution_ctx_cls = SQLiteExecutionContext_aiosqlite
+
+    @classmethod
+    def import_dbapi(cls):
+        return AsyncAdapt_aiosqlite_dbapi(
+            __import__("aiosqlite"), __import__("sqlite3")
+        )
+
+    @classmethod
+    def get_pool_class(cls, url):
+        if cls._is_url_file_db(url):
+            return pool.AsyncAdaptedQueuePool
+        else:
+            return pool.StaticPool
+
+    def is_disconnect(self, e, connection, cursor):
+        if isinstance(
+            e, self.dbapi.OperationalError
+        ) and "no active connection" in str(e):
+            return True
+
+        return super().is_disconnect(e, connection, cursor)
+
+    def get_driver_connection(self, connection):
+        return connection._connection
+
+
+dialect = SQLiteDialect_aiosqlite
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/base.py
new file mode 100644
index 00000000..e0c0f6e8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/base.py
@@ -0,0 +1,2860 @@
+# dialects/sqlite/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r'''
+.. dialect:: sqlite
+    :name: SQLite
+    :normal_support: 3.12+
+    :best_effort: 3.7.16+
+
+.. _sqlite_datetime:
+
+Date and Time Types
+-------------------
+
+SQLite does not have built-in DATE, TIME, or DATETIME types, and pysqlite does
+not provide out of the box functionality for translating values between Python
+`datetime` objects and a SQLite-supported format. SQLAlchemy's own
+:class:`~sqlalchemy.types.DateTime` and related types provide date formatting
+and parsing functionality when SQLite is used. The implementation classes are
+:class:`_sqlite.DATETIME`, :class:`_sqlite.DATE` and :class:`_sqlite.TIME`.
+These types represent dates and times as ISO formatted strings, which also
+nicely support ordering. There's no reliance on typical "libc" internals for
+these functions so historical dates are fully supported.
+
+Ensuring Text affinity
+^^^^^^^^^^^^^^^^^^^^^^
+
+The DDL rendered for these types is the standard ``DATE``, ``TIME``
+and ``DATETIME`` indicators.    However, custom storage formats can also be
+applied to these types.   When the
+storage format is detected as containing no alpha characters, the DDL for
+these types is rendered as ``DATE_CHAR``, ``TIME_CHAR``, and ``DATETIME_CHAR``,
+so that the column continues to have textual affinity.
+
+.. seealso::
+
+    `Type Affinity <https://www.sqlite.org/datatype3.html#affinity>`_ -
+    in the SQLite documentation
+
+.. _sqlite_autoincrement:
+
+SQLite Auto Incrementing Behavior
+----------------------------------
+
+Background on SQLite's autoincrement is at: https://sqlite.org/autoinc.html
+
+Key concepts:
+
+* SQLite has an implicit "auto increment" feature that takes place for any
+  non-composite primary-key column that is specifically created using
+  "INTEGER PRIMARY KEY" for the type + primary key.
+
+* SQLite also has an explicit "AUTOINCREMENT" keyword, that is **not**
+  equivalent to the implicit autoincrement feature; this keyword is not
+  recommended for general use.  SQLAlchemy does not render this keyword
+  unless a special SQLite-specific directive is used (see below).  However,
+  it still requires that the column's type is named "INTEGER".
+
+Using the AUTOINCREMENT Keyword
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+To specifically render the AUTOINCREMENT keyword on the primary key column
+when rendering DDL, add the flag ``sqlite_autoincrement=True`` to the Table
+construct::
+
+    Table(
+        "sometable",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        sqlite_autoincrement=True,
+    )
+
+Allowing autoincrement behavior SQLAlchemy types other than Integer/INTEGER
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+SQLite's typing model is based on naming conventions.  Among other things, this
+means that any type name which contains the substring ``"INT"`` will be
+determined to be of "integer affinity".  A type named ``"BIGINT"``,
+``"SPECIAL_INT"`` or even ``"XYZINTQPR"``, will be considered by SQLite to be
+of "integer" affinity.  However, **the SQLite autoincrement feature, whether
+implicitly or explicitly enabled, requires that the name of the column's type
+is exactly the string "INTEGER"**.  Therefore, if an application uses a type
+like :class:`.BigInteger` for a primary key, on SQLite this type will need to
+be rendered as the name ``"INTEGER"`` when emitting the initial ``CREATE
+TABLE`` statement in order for the autoincrement behavior to be available.
+
+One approach to achieve this is to use :class:`.Integer` on SQLite
+only using :meth:`.TypeEngine.with_variant`::
+
+    table = Table(
+        "my_table",
+        metadata,
+        Column(
+            "id",
+            BigInteger().with_variant(Integer, "sqlite"),
+            primary_key=True,
+        ),
+    )
+
+Another is to use a subclass of :class:`.BigInteger` that overrides its DDL
+name to be ``INTEGER`` when compiled against SQLite::
+
+    from sqlalchemy import BigInteger
+    from sqlalchemy.ext.compiler import compiles
+
+
+    class SLBigInteger(BigInteger):
+        pass
+
+
+    @compiles(SLBigInteger, "sqlite")
+    def bi_c(element, compiler, **kw):
+        return "INTEGER"
+
+
+    @compiles(SLBigInteger)
+    def bi_c(element, compiler, **kw):
+        return compiler.visit_BIGINT(element, **kw)
+
+
+    table = Table(
+        "my_table", metadata, Column("id", SLBigInteger(), primary_key=True)
+    )
+
+.. seealso::
+
+    :meth:`.TypeEngine.with_variant`
+
+    :ref:`sqlalchemy.ext.compiler_toplevel`
+
+    `Datatypes In SQLite Version 3 <https://sqlite.org/datatype3.html>`_
+
+.. _sqlite_concurrency:
+
+Database Locking Behavior / Concurrency
+---------------------------------------
+
+SQLite is not designed for a high level of write concurrency. The database
+itself, being a file, is locked completely during write operations within
+transactions, meaning exactly one "connection" (in reality a file handle)
+has exclusive access to the database during this period - all other
+"connections" will be blocked during this time.
+
+The Python DBAPI specification also calls for a connection model that is
+always in a transaction; there is no ``connection.begin()`` method,
+only ``connection.commit()`` and ``connection.rollback()``, upon which a
+new transaction is to be begun immediately.  This may seem to imply
+that the SQLite driver would in theory allow only a single filehandle on a
+particular database file at any time; however, there are several
+factors both within SQLite itself as well as within the pysqlite driver
+which loosen this restriction significantly.
+
+However, no matter what locking modes are used, SQLite will still always
+lock the database file once a transaction is started and DML (e.g. INSERT,
+UPDATE, DELETE) has at least been emitted, and this will block
+other transactions at least at the point that they also attempt to emit DML.
+By default, the length of time on this block is very short before it times out
+with an error.
+
+This behavior becomes more critical when used in conjunction with the
+SQLAlchemy ORM.  SQLAlchemy's :class:`.Session` object by default runs
+within a transaction, and with its autoflush model, may emit DML preceding
+any SELECT statement.   This may lead to a SQLite database that locks
+more quickly than is expected.   The locking mode of SQLite and the pysqlite
+driver can be manipulated to some degree, however it should be noted that
+achieving a high degree of write-concurrency with SQLite is a losing battle.
+
+For more information on SQLite's lack of write concurrency by design, please
+see
+`Situations Where Another RDBMS May Work Better - High Concurrency
+<https://www.sqlite.org/whentouse.html>`_ near the bottom of the page.
+
+The following subsections introduce areas that are impacted by SQLite's
+file-based architecture and additionally will usually require workarounds to
+work when using the pysqlite driver.
+
+.. _sqlite_isolation_level:
+
+Transaction Isolation Level / Autocommit
+----------------------------------------
+
+SQLite supports "transaction isolation" in a non-standard way, along two
+axes.  One is that of the
+`PRAGMA read_uncommitted <https://www.sqlite.org/pragma.html#pragma_read_uncommitted>`_
+instruction.   This setting can essentially switch SQLite between its
+default mode of ``SERIALIZABLE`` isolation, and a "dirty read" isolation
+mode normally referred to as ``READ UNCOMMITTED``.
+
+SQLAlchemy ties into this PRAGMA statement using the
+:paramref:`_sa.create_engine.isolation_level` parameter of
+:func:`_sa.create_engine`.
+Valid values for this parameter when used with SQLite are ``"SERIALIZABLE"``
+and ``"READ UNCOMMITTED"`` corresponding to a value of 0 and 1, respectively.
+SQLite defaults to ``SERIALIZABLE``, however its behavior is impacted by
+the pysqlite driver's default behavior.
+
+When using the pysqlite driver, the ``"AUTOCOMMIT"`` isolation level is also
+available, which will alter the pysqlite connection using the ``.isolation_level``
+attribute on the DBAPI connection and set it to None for the duration
+of the setting.
+
+.. versionadded:: 1.3.16 added support for SQLite AUTOCOMMIT isolation level
+   when using the pysqlite / sqlite3 SQLite driver.
+
+
+The other axis along which SQLite's transactional locking is impacted is
+via the nature of the ``BEGIN`` statement used.   The three varieties
+are "deferred", "immediate", and "exclusive", as described at
+`BEGIN TRANSACTION <https://sqlite.org/lang_transaction.html>`_.   A straight
+``BEGIN`` statement uses the "deferred" mode, where the database file is
+not locked until the first read or write operation, and read access remains
+open to other transactions until the first write operation.  But again,
+it is critical to note that the pysqlite driver interferes with this behavior
+by *not even emitting BEGIN* until the first write operation.
+
+.. warning::
+
+    SQLite's transactional scope is impacted by unresolved
+    issues in the pysqlite driver, which defers BEGIN statements to a greater
+    degree than is often feasible. See the section :ref:`pysqlite_serializable`
+    or :ref:`aiosqlite_serializable` for techniques to work around this behavior.
+
+.. seealso::
+
+    :ref:`dbapi_autocommit`
+
+INSERT/UPDATE/DELETE...RETURNING
+---------------------------------
+
+The SQLite dialect supports SQLite 3.35's  ``INSERT|UPDATE|DELETE..RETURNING``
+syntax.   ``INSERT..RETURNING`` may be used
+automatically in some cases in order to fetch newly generated identifiers in
+place of the traditional approach of using ``cursor.lastrowid``, however
+``cursor.lastrowid`` is currently still preferred for simple single-statement
+cases for its better performance.
+
+To specify an explicit ``RETURNING`` clause, use the
+:meth:`._UpdateBase.returning` method on a per-statement basis::
+
+    # INSERT..RETURNING
+    result = connection.execute(
+        table.insert().values(name="foo").returning(table.c.col1, table.c.col2)
+    )
+    print(result.all())
+
+    # UPDATE..RETURNING
+    result = connection.execute(
+        table.update()
+        .where(table.c.name == "foo")
+        .values(name="bar")
+        .returning(table.c.col1, table.c.col2)
+    )
+    print(result.all())
+
+    # DELETE..RETURNING
+    result = connection.execute(
+        table.delete()
+        .where(table.c.name == "foo")
+        .returning(table.c.col1, table.c.col2)
+    )
+    print(result.all())
+
+.. versionadded:: 2.0  Added support for SQLite RETURNING
+
+SAVEPOINT Support
+----------------------------
+
+SQLite supports SAVEPOINTs, which only function once a transaction is
+begun.   SQLAlchemy's SAVEPOINT support is available using the
+:meth:`_engine.Connection.begin_nested` method at the Core level, and
+:meth:`.Session.begin_nested` at the ORM level.   However, SAVEPOINTs
+won't work at all with pysqlite unless workarounds are taken.
+
+.. warning::
+
+    SQLite's SAVEPOINT feature is impacted by unresolved
+    issues in the pysqlite and aiosqlite drivers, which defer BEGIN statements
+    to a greater degree than is often feasible. See the sections
+    :ref:`pysqlite_serializable` and :ref:`aiosqlite_serializable`
+    for techniques to work around this behavior.
+
+Transactional DDL
+----------------------------
+
+The SQLite database supports transactional :term:`DDL` as well.
+In this case, the pysqlite driver is not only failing to start transactions,
+it also is ending any existing transaction when DDL is detected, so again,
+workarounds are required.
+
+.. warning::
+
+    SQLite's transactional DDL is impacted by unresolved issues
+    in the pysqlite driver, which fails to emit BEGIN and additionally
+    forces a COMMIT to cancel any transaction when DDL is encountered.
+    See the section :ref:`pysqlite_serializable`
+    for techniques to work around this behavior.
+
+.. _sqlite_foreign_keys:
+
+Foreign Key Support
+-------------------
+
+SQLite supports FOREIGN KEY syntax when emitting CREATE statements for tables,
+however by default these constraints have no effect on the operation of the
+table.
+
+Constraint checking on SQLite has three prerequisites:
+
+* At least version 3.6.19 of SQLite must be in use
+* The SQLite library must be compiled *without* the SQLITE_OMIT_FOREIGN_KEY
+  or SQLITE_OMIT_TRIGGER symbols enabled.
+* The ``PRAGMA foreign_keys = ON`` statement must be emitted on all
+  connections before use -- including the initial call to
+  :meth:`sqlalchemy.schema.MetaData.create_all`.
+
+SQLAlchemy allows for the ``PRAGMA`` statement to be emitted automatically for
+new connections through the usage of events::
+
+    from sqlalchemy.engine import Engine
+    from sqlalchemy import event
+
+
+    @event.listens_for(Engine, "connect")
+    def set_sqlite_pragma(dbapi_connection, connection_record):
+        cursor = dbapi_connection.cursor()
+        cursor.execute("PRAGMA foreign_keys=ON")
+        cursor.close()
+
+.. warning::
+
+    When SQLite foreign keys are enabled, it is **not possible**
+    to emit CREATE or DROP statements for tables that contain
+    mutually-dependent foreign key constraints;
+    to emit the DDL for these tables requires that ALTER TABLE be used to
+    create or drop these constraints separately, for which SQLite has
+    no support.
+
+.. seealso::
+
+    `SQLite Foreign Key Support <https://www.sqlite.org/foreignkeys.html>`_
+    - on the SQLite web site.
+
+    :ref:`event_toplevel` - SQLAlchemy event API.
+
+    :ref:`use_alter` - more information on SQLAlchemy's facilities for handling
+     mutually-dependent foreign key constraints.
+
+.. _sqlite_on_conflict_ddl:
+
+ON CONFLICT support for constraints
+-----------------------------------
+
+.. seealso:: This section describes the :term:`DDL` version of "ON CONFLICT" for
+   SQLite, which occurs within a CREATE TABLE statement.  For "ON CONFLICT" as
+   applied to an INSERT statement, see :ref:`sqlite_on_conflict_insert`.
+
+SQLite supports a non-standard DDL clause known as ON CONFLICT which can be applied
+to primary key, unique, check, and not null constraints.   In DDL, it is
+rendered either within the "CONSTRAINT" clause or within the column definition
+itself depending on the location of the target constraint.    To render this
+clause within DDL, the extension parameter ``sqlite_on_conflict`` can be
+specified with a string conflict resolution algorithm within the
+:class:`.PrimaryKeyConstraint`, :class:`.UniqueConstraint`,
+:class:`.CheckConstraint` objects.  Within the :class:`_schema.Column` object,
+there
+are individual parameters ``sqlite_on_conflict_not_null``,
+``sqlite_on_conflict_primary_key``, ``sqlite_on_conflict_unique`` which each
+correspond to the three types of relevant constraint types that can be
+indicated from a :class:`_schema.Column` object.
+
+.. seealso::
+
+    `ON CONFLICT <https://www.sqlite.org/lang_conflict.html>`_ - in the SQLite
+    documentation
+
+.. versionadded:: 1.3
+
+
+The ``sqlite_on_conflict`` parameters accept a  string argument which is just
+the resolution name to be chosen, which on SQLite can be one of ROLLBACK,
+ABORT, FAIL, IGNORE, and REPLACE.   For example, to add a UNIQUE constraint
+that specifies the IGNORE algorithm::
+
+    some_table = Table(
+        "some_table",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("data", Integer),
+        UniqueConstraint("id", "data", sqlite_on_conflict="IGNORE"),
+    )
+
+The above renders CREATE TABLE DDL as:
+
+.. sourcecode:: sql
+
+    CREATE TABLE some_table (
+        id INTEGER NOT NULL,
+        data INTEGER,
+        PRIMARY KEY (id),
+        UNIQUE (id, data) ON CONFLICT IGNORE
+    )
+
+
+When using the :paramref:`_schema.Column.unique`
+flag to add a UNIQUE constraint
+to a single column, the ``sqlite_on_conflict_unique`` parameter can
+be added to the :class:`_schema.Column` as well, which will be added to the
+UNIQUE constraint in the DDL::
+
+    some_table = Table(
+        "some_table",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column(
+            "data", Integer, unique=True, sqlite_on_conflict_unique="IGNORE"
+        ),
+    )
+
+rendering:
+
+.. sourcecode:: sql
+
+    CREATE TABLE some_table (
+        id INTEGER NOT NULL,
+        data INTEGER,
+        PRIMARY KEY (id),
+        UNIQUE (data) ON CONFLICT IGNORE
+    )
+
+To apply the FAIL algorithm for a NOT NULL constraint,
+``sqlite_on_conflict_not_null`` is used::
+
+    some_table = Table(
+        "some_table",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column(
+            "data", Integer, nullable=False, sqlite_on_conflict_not_null="FAIL"
+        ),
+    )
+
+this renders the column inline ON CONFLICT phrase:
+
+.. sourcecode:: sql
+
+    CREATE TABLE some_table (
+        id INTEGER NOT NULL,
+        data INTEGER NOT NULL ON CONFLICT FAIL,
+        PRIMARY KEY (id)
+    )
+
+
+Similarly, for an inline primary key, use ``sqlite_on_conflict_primary_key``::
+
+    some_table = Table(
+        "some_table",
+        metadata,
+        Column(
+            "id",
+            Integer,
+            primary_key=True,
+            sqlite_on_conflict_primary_key="FAIL",
+        ),
+    )
+
+SQLAlchemy renders the PRIMARY KEY constraint separately, so the conflict
+resolution algorithm is applied to the constraint itself:
+
+.. sourcecode:: sql
+
+    CREATE TABLE some_table (
+        id INTEGER NOT NULL,
+        PRIMARY KEY (id) ON CONFLICT FAIL
+    )
+
+.. _sqlite_on_conflict_insert:
+
+INSERT...ON CONFLICT (Upsert)
+-----------------------------
+
+.. seealso:: This section describes the :term:`DML` version of "ON CONFLICT" for
+   SQLite, which occurs within an INSERT statement.  For "ON CONFLICT" as
+   applied to a CREATE TABLE statement, see :ref:`sqlite_on_conflict_ddl`.
+
+From version 3.24.0 onwards, SQLite supports "upserts" (update or insert)
+of rows into a table via the ``ON CONFLICT`` clause of the ``INSERT``
+statement. A candidate row will only be inserted if that row does not violate
+any unique or primary key constraints. In the case of a unique constraint violation, a
+secondary action can occur which can be either "DO UPDATE", indicating that
+the data in the target row should be updated, or "DO NOTHING", which indicates
+to silently skip this row.
+
+Conflicts are determined using columns that are part of existing unique
+constraints and indexes.  These constraints are identified by stating the
+columns and conditions that comprise the indexes.
+
+SQLAlchemy provides ``ON CONFLICT`` support via the SQLite-specific
+:func:`_sqlite.insert()` function, which provides
+the generative methods :meth:`_sqlite.Insert.on_conflict_do_update`
+and :meth:`_sqlite.Insert.on_conflict_do_nothing`:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy.dialects.sqlite import insert
+
+    >>> insert_stmt = insert(my_table).values(
+    ...     id="some_existing_id", data="inserted value"
+    ... )
+
+    >>> do_update_stmt = insert_stmt.on_conflict_do_update(
+    ...     index_elements=["id"], set_=dict(data="updated value")
+    ... )
+
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (?, ?)
+    ON CONFLICT (id) DO UPDATE SET data = ?{stop}
+
+    >>> do_nothing_stmt = insert_stmt.on_conflict_do_nothing(index_elements=["id"])
+
+    >>> print(do_nothing_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (?, ?)
+    ON CONFLICT (id) DO NOTHING
+
+.. versionadded:: 1.4
+
+.. seealso::
+
+    `Upsert
+    <https://sqlite.org/lang_UPSERT.html>`_
+    - in the SQLite documentation.
+
+
+Specifying the Target
+^^^^^^^^^^^^^^^^^^^^^
+
+Both methods supply the "target" of the conflict using column inference:
+
+* The :paramref:`_sqlite.Insert.on_conflict_do_update.index_elements` argument
+  specifies a sequence containing string column names, :class:`_schema.Column`
+  objects, and/or SQL expression elements, which would identify a unique index
+  or unique constraint.
+
+* When using :paramref:`_sqlite.Insert.on_conflict_do_update.index_elements`
+  to infer an index, a partial index can be inferred by also specifying the
+  :paramref:`_sqlite.Insert.on_conflict_do_update.index_where` parameter:
+
+  .. sourcecode:: pycon+sql
+
+        >>> stmt = insert(my_table).values(user_email="a@b.com", data="inserted data")
+
+        >>> do_update_stmt = stmt.on_conflict_do_update(
+        ...     index_elements=[my_table.c.user_email],
+        ...     index_where=my_table.c.user_email.like("%@gmail.com"),
+        ...     set_=dict(data=stmt.excluded.data),
+        ... )
+
+        >>> print(do_update_stmt)
+        {printsql}INSERT INTO my_table (data, user_email) VALUES (?, ?)
+        ON CONFLICT (user_email)
+        WHERE user_email LIKE '%@gmail.com'
+        DO UPDATE SET data = excluded.data
+
+The SET Clause
+^^^^^^^^^^^^^^^
+
+``ON CONFLICT...DO UPDATE`` is used to perform an update of the already
+existing row, using any combination of new values as well as values
+from the proposed insertion. These values are specified using the
+:paramref:`_sqlite.Insert.on_conflict_do_update.set_` parameter.  This
+parameter accepts a dictionary which consists of direct values
+for UPDATE:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+
+    >>> do_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"], set_=dict(data="updated value")
+    ... )
+
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (?, ?)
+    ON CONFLICT (id) DO UPDATE SET data = ?
+
+.. warning::
+
+    The :meth:`_sqlite.Insert.on_conflict_do_update` method does **not** take
+    into account Python-side default UPDATE values or generation functions,
+    e.g. those specified using :paramref:`_schema.Column.onupdate`. These
+    values will not be exercised for an ON CONFLICT style of UPDATE, unless
+    they are manually specified in the
+    :paramref:`_sqlite.Insert.on_conflict_do_update.set_` dictionary.
+
+Updating using the Excluded INSERT Values
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+In order to refer to the proposed insertion row, the special alias
+:attr:`~.sqlite.Insert.excluded` is available as an attribute on
+the :class:`_sqlite.Insert` object; this object creates an "excluded." prefix
+on a column, that informs the DO UPDATE to update the row with the value that
+would have been inserted had the constraint not failed:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(
+    ...     id="some_id", data="inserted value", author="jlh"
+    ... )
+
+    >>> do_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"],
+    ...     set_=dict(data="updated value", author=stmt.excluded.author),
+    ... )
+
+    >>> print(do_update_stmt)
+    {printsql}INSERT INTO my_table (id, data, author) VALUES (?, ?, ?)
+    ON CONFLICT (id) DO UPDATE SET data = ?, author = excluded.author
+
+Additional WHERE Criteria
+^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The :meth:`_sqlite.Insert.on_conflict_do_update` method also accepts
+a WHERE clause using the :paramref:`_sqlite.Insert.on_conflict_do_update.where`
+parameter, which will limit those rows which receive an UPDATE:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(
+    ...     id="some_id", data="inserted value", author="jlh"
+    ... )
+
+    >>> on_update_stmt = stmt.on_conflict_do_update(
+    ...     index_elements=["id"],
+    ...     set_=dict(data="updated value", author=stmt.excluded.author),
+    ...     where=(my_table.c.status == 2),
+    ... )
+    >>> print(on_update_stmt)
+    {printsql}INSERT INTO my_table (id, data, author) VALUES (?, ?, ?)
+    ON CONFLICT (id) DO UPDATE SET data = ?, author = excluded.author
+    WHERE my_table.status = ?
+
+
+Skipping Rows with DO NOTHING
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+``ON CONFLICT`` may be used to skip inserting a row entirely
+if any conflict with a unique constraint occurs; below this is illustrated
+using the :meth:`_sqlite.Insert.on_conflict_do_nothing` method:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+    >>> stmt = stmt.on_conflict_do_nothing(index_elements=["id"])
+    >>> print(stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (?, ?) ON CONFLICT (id) DO NOTHING
+
+
+If ``DO NOTHING`` is used without specifying any columns or constraint,
+it has the effect of skipping the INSERT for any unique violation which
+occurs:
+
+.. sourcecode:: pycon+sql
+
+    >>> stmt = insert(my_table).values(id="some_id", data="inserted value")
+    >>> stmt = stmt.on_conflict_do_nothing()
+    >>> print(stmt)
+    {printsql}INSERT INTO my_table (id, data) VALUES (?, ?) ON CONFLICT DO NOTHING
+
+.. _sqlite_type_reflection:
+
+Type Reflection
+---------------
+
+SQLite types are unlike those of most other database backends, in that
+the string name of the type usually does not correspond to a "type" in a
+one-to-one fashion.  Instead, SQLite links per-column typing behavior
+to one of five so-called "type affinities" based on a string matching
+pattern for the type.
+
+SQLAlchemy's reflection process, when inspecting types, uses a simple
+lookup table to link the keywords returned to provided SQLAlchemy types.
+This lookup table is present within the SQLite dialect as it is for all
+other dialects.  However, the SQLite dialect has a different "fallback"
+routine for when a particular type name is not located in the lookup map;
+it instead implements the SQLite "type affinity" scheme located at
+https://www.sqlite.org/datatype3.html section 2.1.
+
+The provided typemap will make direct associations from an exact string
+name match for the following types:
+
+:class:`_types.BIGINT`, :class:`_types.BLOB`,
+:class:`_types.BOOLEAN`, :class:`_types.BOOLEAN`,
+:class:`_types.CHAR`, :class:`_types.DATE`,
+:class:`_types.DATETIME`, :class:`_types.FLOAT`,
+:class:`_types.DECIMAL`, :class:`_types.FLOAT`,
+:class:`_types.INTEGER`, :class:`_types.INTEGER`,
+:class:`_types.NUMERIC`, :class:`_types.REAL`,
+:class:`_types.SMALLINT`, :class:`_types.TEXT`,
+:class:`_types.TIME`, :class:`_types.TIMESTAMP`,
+:class:`_types.VARCHAR`, :class:`_types.NVARCHAR`,
+:class:`_types.NCHAR`
+
+When a type name does not match one of the above types, the "type affinity"
+lookup is used instead:
+
+* :class:`_types.INTEGER` is returned if the type name includes the
+  string ``INT``
+* :class:`_types.TEXT` is returned if the type name includes the
+  string ``CHAR``, ``CLOB`` or ``TEXT``
+* :class:`_types.NullType` is returned if the type name includes the
+  string ``BLOB``
+* :class:`_types.REAL` is returned if the type name includes the string
+  ``REAL``, ``FLOA`` or ``DOUB``.
+* Otherwise, the :class:`_types.NUMERIC` type is used.
+
+.. _sqlite_partial_index:
+
+Partial Indexes
+---------------
+
+A partial index, e.g. one which uses a WHERE clause, can be specified
+with the DDL system using the argument ``sqlite_where``::
+
+    tbl = Table("testtbl", m, Column("data", Integer))
+    idx = Index(
+        "test_idx1",
+        tbl.c.data,
+        sqlite_where=and_(tbl.c.data > 5, tbl.c.data < 10),
+    )
+
+The index will be rendered at create time as:
+
+.. sourcecode:: sql
+
+    CREATE INDEX test_idx1 ON testtbl (data)
+    WHERE data > 5 AND data < 10
+
+.. _sqlite_dotted_column_names:
+
+Dotted Column Names
+-------------------
+
+Using table or column names that explicitly have periods in them is
+**not recommended**.   While this is generally a bad idea for relational
+databases in general, as the dot is a syntactically significant character,
+the SQLite driver up until version **3.10.0** of SQLite has a bug which
+requires that SQLAlchemy filter out these dots in result sets.
+
+The bug, entirely outside of SQLAlchemy, can be illustrated thusly::
+
+    import sqlite3
+
+    assert sqlite3.sqlite_version_info < (
+        3,
+        10,
+        0,
+    ), "bug is fixed in this version"
+
+    conn = sqlite3.connect(":memory:")
+    cursor = conn.cursor()
+
+    cursor.execute("create table x (a integer, b integer)")
+    cursor.execute("insert into x (a, b) values (1, 1)")
+    cursor.execute("insert into x (a, b) values (2, 2)")
+
+    cursor.execute("select x.a, x.b from x")
+    assert [c[0] for c in cursor.description] == ["a", "b"]
+
+    cursor.execute(
+        """
+        select x.a, x.b from x where a=1
+        union
+        select x.a, x.b from x where a=2
+        """
+    )
+    assert [c[0] for c in cursor.description] == ["a", "b"], [
+        c[0] for c in cursor.description
+    ]
+
+The second assertion fails:
+
+.. sourcecode:: text
+
+    Traceback (most recent call last):
+      File "test.py", line 19, in <module>
+        [c[0] for c in cursor.description]
+    AssertionError: ['x.a', 'x.b']
+
+Where above, the driver incorrectly reports the names of the columns
+including the name of the table, which is entirely inconsistent vs.
+when the UNION is not present.
+
+SQLAlchemy relies upon column names being predictable in how they match
+to the original statement, so the SQLAlchemy dialect has no choice but
+to filter these out::
+
+
+    from sqlalchemy import create_engine
+
+    eng = create_engine("sqlite://")
+    conn = eng.connect()
+
+    conn.exec_driver_sql("create table x (a integer, b integer)")
+    conn.exec_driver_sql("insert into x (a, b) values (1, 1)")
+    conn.exec_driver_sql("insert into x (a, b) values (2, 2)")
+
+    result = conn.exec_driver_sql("select x.a, x.b from x")
+    assert result.keys() == ["a", "b"]
+
+    result = conn.exec_driver_sql(
+        """
+        select x.a, x.b from x where a=1
+        union
+        select x.a, x.b from x where a=2
+        """
+    )
+    assert result.keys() == ["a", "b"]
+
+Note that above, even though SQLAlchemy filters out the dots, *both
+names are still addressable*::
+
+    >>> row = result.first()
+    >>> row["a"]
+    1
+    >>> row["x.a"]
+    1
+    >>> row["b"]
+    1
+    >>> row["x.b"]
+    1
+
+Therefore, the workaround applied by SQLAlchemy only impacts
+:meth:`_engine.CursorResult.keys` and :meth:`.Row.keys()` in the public API. In
+the very specific case where an application is forced to use column names that
+contain dots, and the functionality of :meth:`_engine.CursorResult.keys` and
+:meth:`.Row.keys()` is required to return these dotted names unmodified,
+the ``sqlite_raw_colnames`` execution option may be provided, either on a
+per-:class:`_engine.Connection` basis::
+
+    result = conn.execution_options(sqlite_raw_colnames=True).exec_driver_sql(
+        """
+        select x.a, x.b from x where a=1
+        union
+        select x.a, x.b from x where a=2
+        """
+    )
+    assert result.keys() == ["x.a", "x.b"]
+
+or on a per-:class:`_engine.Engine` basis::
+
+    engine = create_engine(
+        "sqlite://", execution_options={"sqlite_raw_colnames": True}
+    )
+
+When using the per-:class:`_engine.Engine` execution option, note that
+**Core and ORM queries that use UNION may not function properly**.
+
+SQLite-specific table options
+-----------------------------
+
+One option for CREATE TABLE is supported directly by the SQLite
+dialect in conjunction with the :class:`_schema.Table` construct:
+
+* ``WITHOUT ROWID``::
+
+    Table("some_table", metadata, ..., sqlite_with_rowid=False)
+
+*
+  ``STRICT``::
+
+    Table("some_table", metadata, ..., sqlite_strict=True)
+
+  .. versionadded:: 2.0.37
+
+.. seealso::
+
+    `SQLite CREATE TABLE options
+    <https://www.sqlite.org/lang_createtable.html>`_
+
+.. _sqlite_include_internal:
+
+Reflecting internal schema tables
+----------------------------------
+
+Reflection methods that return lists of tables will omit so-called
+"SQLite internal schema object" names, which are considered by SQLite
+as any object name that is prefixed with ``sqlite_``.  An example of
+such an object is the ``sqlite_sequence`` table that's generated when
+the ``AUTOINCREMENT`` column parameter is used.   In order to return
+these objects, the parameter ``sqlite_include_internal=True`` may be
+passed to methods such as :meth:`_schema.MetaData.reflect` or
+:meth:`.Inspector.get_table_names`.
+
+.. versionadded:: 2.0  Added the ``sqlite_include_internal=True`` parameter.
+   Previously, these tables were not ignored by SQLAlchemy reflection
+   methods.
+
+.. note::
+
+    The ``sqlite_include_internal`` parameter does not refer to the
+    "system" tables that are present in schemas such as ``sqlite_master``.
+
+.. seealso::
+
+    `SQLite Internal Schema Objects <https://www.sqlite.org/fileformat2.html#intschema>`_ - in the SQLite
+    documentation.
+
+'''  # noqa
+from __future__ import annotations
+
+import datetime
+import numbers
+import re
+from typing import Optional
+
+from .json import JSON
+from .json import JSONIndexType
+from .json import JSONPathType
+from ... import exc
+from ... import schema as sa_schema
+from ... import sql
+from ... import text
+from ... import types as sqltypes
+from ... import util
+from ...engine import default
+from ...engine import processors
+from ...engine import reflection
+from ...engine.reflection import ReflectionDefaults
+from ...sql import coercions
+from ...sql import ColumnElement
+from ...sql import compiler
+from ...sql import elements
+from ...sql import roles
+from ...sql import schema
+from ...types import BLOB  # noqa
+from ...types import BOOLEAN  # noqa
+from ...types import CHAR  # noqa
+from ...types import DECIMAL  # noqa
+from ...types import FLOAT  # noqa
+from ...types import INTEGER  # noqa
+from ...types import NUMERIC  # noqa
+from ...types import REAL  # noqa
+from ...types import SMALLINT  # noqa
+from ...types import TEXT  # noqa
+from ...types import TIMESTAMP  # noqa
+from ...types import VARCHAR  # noqa
+
+
+class _SQliteJson(JSON):
+    def result_processor(self, dialect, coltype):
+        default_processor = super().result_processor(dialect, coltype)
+
+        def process(value):
+            try:
+                return default_processor(value)
+            except TypeError:
+                if isinstance(value, numbers.Number):
+                    return value
+                else:
+                    raise
+
+        return process
+
+
+class _DateTimeMixin:
+    _reg = None
+    _storage_format = None
+
+    def __init__(self, storage_format=None, regexp=None, **kw):
+        super().__init__(**kw)
+        if regexp is not None:
+            self._reg = re.compile(regexp)
+        if storage_format is not None:
+            self._storage_format = storage_format
+
+    @property
+    def format_is_text_affinity(self):
+        """return True if the storage format will automatically imply
+        a TEXT affinity.
+
+        If the storage format contains no non-numeric characters,
+        it will imply a NUMERIC storage format on SQLite; in this case,
+        the type will generate its DDL as DATE_CHAR, DATETIME_CHAR,
+        TIME_CHAR.
+
+        """
+        spec = self._storage_format % {
+            "year": 0,
+            "month": 0,
+            "day": 0,
+            "hour": 0,
+            "minute": 0,
+            "second": 0,
+            "microsecond": 0,
+        }
+        return bool(re.search(r"[^0-9]", spec))
+
+    def adapt(self, cls, **kw):
+        if issubclass(cls, _DateTimeMixin):
+            if self._storage_format:
+                kw["storage_format"] = self._storage_format
+            if self._reg:
+                kw["regexp"] = self._reg
+        return super().adapt(cls, **kw)
+
+    def literal_processor(self, dialect):
+        bp = self.bind_processor(dialect)
+
+        def process(value):
+            return "'%s'" % bp(value)
+
+        return process
+
+
+class DATETIME(_DateTimeMixin, sqltypes.DateTime):
+    r"""Represent a Python datetime object in SQLite using a string.
+
+    The default string storage format is::
+
+        "%(year)04d-%(month)02d-%(day)02d %(hour)02d:%(minute)02d:%(second)02d.%(microsecond)06d"
+
+    e.g.:
+
+    .. sourcecode:: text
+
+        2021-03-15 12:05:57.105542
+
+    The incoming storage format is by default parsed using the
+    Python ``datetime.fromisoformat()`` function.
+
+    .. versionchanged:: 2.0  ``datetime.fromisoformat()`` is used for default
+       datetime string parsing.
+
+    The storage format can be customized to some degree using the
+    ``storage_format`` and ``regexp`` parameters, such as::
+
+        import re
+        from sqlalchemy.dialects.sqlite import DATETIME
+
+        dt = DATETIME(
+            storage_format=(
+                "%(year)04d/%(month)02d/%(day)02d %(hour)02d:%(minute)02d:%(second)02d"
+            ),
+            regexp=r"(\d+)/(\d+)/(\d+) (\d+)-(\d+)-(\d+)",
+        )
+
+    :param storage_format: format string which will be applied to the dict
+     with keys year, month, day, hour, minute, second, and microsecond.
+
+    :param regexp: regular expression which will be applied to incoming result
+     rows, replacing the use of ``datetime.fromisoformat()`` to parse incoming
+     strings. If the regexp contains named groups, the resulting match dict is
+     applied to the Python datetime() constructor as keyword arguments.
+     Otherwise, if positional groups are used, the datetime() constructor
+     is called with positional arguments via
+     ``*map(int, match_obj.groups(0))``.
+
+    """  # noqa
+
+    _storage_format = (
+        "%(year)04d-%(month)02d-%(day)02d "
+        "%(hour)02d:%(minute)02d:%(second)02d.%(microsecond)06d"
+    )
+
+    def __init__(self, *args, **kwargs):
+        truncate_microseconds = kwargs.pop("truncate_microseconds", False)
+        super().__init__(*args, **kwargs)
+        if truncate_microseconds:
+            assert "storage_format" not in kwargs, (
+                "You can specify only "
+                "one of truncate_microseconds or storage_format."
+            )
+            assert "regexp" not in kwargs, (
+                "You can specify only one of "
+                "truncate_microseconds or regexp."
+            )
+            self._storage_format = (
+                "%(year)04d-%(month)02d-%(day)02d "
+                "%(hour)02d:%(minute)02d:%(second)02d"
+            )
+
+    def bind_processor(self, dialect):
+        datetime_datetime = datetime.datetime
+        datetime_date = datetime.date
+        format_ = self._storage_format
+
+        def process(value):
+            if value is None:
+                return None
+            elif isinstance(value, datetime_datetime):
+                return format_ % {
+                    "year": value.year,
+                    "month": value.month,
+                    "day": value.day,
+                    "hour": value.hour,
+                    "minute": value.minute,
+                    "second": value.second,
+                    "microsecond": value.microsecond,
+                }
+            elif isinstance(value, datetime_date):
+                return format_ % {
+                    "year": value.year,
+                    "month": value.month,
+                    "day": value.day,
+                    "hour": 0,
+                    "minute": 0,
+                    "second": 0,
+                    "microsecond": 0,
+                }
+            else:
+                raise TypeError(
+                    "SQLite DateTime type only accepts Python "
+                    "datetime and date objects as input."
+                )
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        if self._reg:
+            return processors.str_to_datetime_processor_factory(
+                self._reg, datetime.datetime
+            )
+        else:
+            return processors.str_to_datetime
+
+
+class DATE(_DateTimeMixin, sqltypes.Date):
+    r"""Represent a Python date object in SQLite using a string.
+
+    The default string storage format is::
+
+        "%(year)04d-%(month)02d-%(day)02d"
+
+    e.g.:
+
+    .. sourcecode:: text
+
+        2011-03-15
+
+    The incoming storage format is by default parsed using the
+    Python ``date.fromisoformat()`` function.
+
+    .. versionchanged:: 2.0  ``date.fromisoformat()`` is used for default
+       date string parsing.
+
+
+    The storage format can be customized to some degree using the
+    ``storage_format`` and ``regexp`` parameters, such as::
+
+        import re
+        from sqlalchemy.dialects.sqlite import DATE
+
+        d = DATE(
+            storage_format="%(month)02d/%(day)02d/%(year)04d",
+            regexp=re.compile("(?P<month>\d+)/(?P<day>\d+)/(?P<year>\d+)"),
+        )
+
+    :param storage_format: format string which will be applied to the
+     dict with keys year, month, and day.
+
+    :param regexp: regular expression which will be applied to
+     incoming result rows, replacing the use of ``date.fromisoformat()`` to
+     parse incoming strings. If the regexp contains named groups, the resulting
+     match dict is applied to the Python date() constructor as keyword
+     arguments. Otherwise, if positional groups are used, the date()
+     constructor is called with positional arguments via
+     ``*map(int, match_obj.groups(0))``.
+
+    """
+
+    _storage_format = "%(year)04d-%(month)02d-%(day)02d"
+
+    def bind_processor(self, dialect):
+        datetime_date = datetime.date
+        format_ = self._storage_format
+
+        def process(value):
+            if value is None:
+                return None
+            elif isinstance(value, datetime_date):
+                return format_ % {
+                    "year": value.year,
+                    "month": value.month,
+                    "day": value.day,
+                }
+            else:
+                raise TypeError(
+                    "SQLite Date type only accepts Python "
+                    "date objects as input."
+                )
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        if self._reg:
+            return processors.str_to_datetime_processor_factory(
+                self._reg, datetime.date
+            )
+        else:
+            return processors.str_to_date
+
+
+class TIME(_DateTimeMixin, sqltypes.Time):
+    r"""Represent a Python time object in SQLite using a string.
+
+    The default string storage format is::
+
+        "%(hour)02d:%(minute)02d:%(second)02d.%(microsecond)06d"
+
+    e.g.:
+
+    .. sourcecode:: text
+
+        12:05:57.10558
+
+    The incoming storage format is by default parsed using the
+    Python ``time.fromisoformat()`` function.
+
+    .. versionchanged:: 2.0  ``time.fromisoformat()`` is used for default
+       time string parsing.
+
+    The storage format can be customized to some degree using the
+    ``storage_format`` and ``regexp`` parameters, such as::
+
+        import re
+        from sqlalchemy.dialects.sqlite import TIME
+
+        t = TIME(
+            storage_format="%(hour)02d-%(minute)02d-%(second)02d-%(microsecond)06d",
+            regexp=re.compile("(\d+)-(\d+)-(\d+)-(?:-(\d+))?"),
+        )
+
+    :param storage_format: format string which will be applied to the dict
+     with keys hour, minute, second, and microsecond.
+
+    :param regexp: regular expression which will be applied to incoming result
+     rows, replacing the use of ``datetime.fromisoformat()`` to parse incoming
+     strings. If the regexp contains named groups, the resulting match dict is
+     applied to the Python time() constructor as keyword arguments. Otherwise,
+     if positional groups are used, the time() constructor is called with
+     positional arguments via ``*map(int, match_obj.groups(0))``.
+
+    """
+
+    _storage_format = "%(hour)02d:%(minute)02d:%(second)02d.%(microsecond)06d"
+
+    def __init__(self, *args, **kwargs):
+        truncate_microseconds = kwargs.pop("truncate_microseconds", False)
+        super().__init__(*args, **kwargs)
+        if truncate_microseconds:
+            assert "storage_format" not in kwargs, (
+                "You can specify only "
+                "one of truncate_microseconds or storage_format."
+            )
+            assert "regexp" not in kwargs, (
+                "You can specify only one of "
+                "truncate_microseconds or regexp."
+            )
+            self._storage_format = "%(hour)02d:%(minute)02d:%(second)02d"
+
+    def bind_processor(self, dialect):
+        datetime_time = datetime.time
+        format_ = self._storage_format
+
+        def process(value):
+            if value is None:
+                return None
+            elif isinstance(value, datetime_time):
+                return format_ % {
+                    "hour": value.hour,
+                    "minute": value.minute,
+                    "second": value.second,
+                    "microsecond": value.microsecond,
+                }
+            else:
+                raise TypeError(
+                    "SQLite Time type only accepts Python "
+                    "time objects as input."
+                )
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        if self._reg:
+            return processors.str_to_datetime_processor_factory(
+                self._reg, datetime.time
+            )
+        else:
+            return processors.str_to_time
+
+
+colspecs = {
+    sqltypes.Date: DATE,
+    sqltypes.DateTime: DATETIME,
+    sqltypes.JSON: _SQliteJson,
+    sqltypes.JSON.JSONIndexType: JSONIndexType,
+    sqltypes.JSON.JSONPathType: JSONPathType,
+    sqltypes.Time: TIME,
+}
+
+ischema_names = {
+    "BIGINT": sqltypes.BIGINT,
+    "BLOB": sqltypes.BLOB,
+    "BOOL": sqltypes.BOOLEAN,
+    "BOOLEAN": sqltypes.BOOLEAN,
+    "CHAR": sqltypes.CHAR,
+    "DATE": sqltypes.DATE,
+    "DATE_CHAR": sqltypes.DATE,
+    "DATETIME": sqltypes.DATETIME,
+    "DATETIME_CHAR": sqltypes.DATETIME,
+    "DOUBLE": sqltypes.DOUBLE,
+    "DECIMAL": sqltypes.DECIMAL,
+    "FLOAT": sqltypes.FLOAT,
+    "INT": sqltypes.INTEGER,
+    "INTEGER": sqltypes.INTEGER,
+    "JSON": JSON,
+    "NUMERIC": sqltypes.NUMERIC,
+    "REAL": sqltypes.REAL,
+    "SMALLINT": sqltypes.SMALLINT,
+    "TEXT": sqltypes.TEXT,
+    "TIME": sqltypes.TIME,
+    "TIME_CHAR": sqltypes.TIME,
+    "TIMESTAMP": sqltypes.TIMESTAMP,
+    "VARCHAR": sqltypes.VARCHAR,
+    "NVARCHAR": sqltypes.NVARCHAR,
+    "NCHAR": sqltypes.NCHAR,
+}
+
+
+class SQLiteCompiler(compiler.SQLCompiler):
+    extract_map = util.update_copy(
+        compiler.SQLCompiler.extract_map,
+        {
+            "month": "%m",
+            "day": "%d",
+            "year": "%Y",
+            "second": "%S",
+            "hour": "%H",
+            "doy": "%j",
+            "minute": "%M",
+            "epoch": "%s",
+            "dow": "%w",
+            "week": "%W",
+        },
+    )
+
+    def visit_truediv_binary(self, binary, operator, **kw):
+        return (
+            self.process(binary.left, **kw)
+            + " / "
+            + "(%s + 0.0)" % self.process(binary.right, **kw)
+        )
+
+    def visit_now_func(self, fn, **kw):
+        return "CURRENT_TIMESTAMP"
+
+    def visit_localtimestamp_func(self, func, **kw):
+        return 'DATETIME(CURRENT_TIMESTAMP, "localtime")'
+
+    def visit_true(self, expr, **kw):
+        return "1"
+
+    def visit_false(self, expr, **kw):
+        return "0"
+
+    def visit_char_length_func(self, fn, **kw):
+        return "length%s" % self.function_argspec(fn)
+
+    def visit_aggregate_strings_func(self, fn, **kw):
+        return "group_concat%s" % self.function_argspec(fn)
+
+    def visit_cast(self, cast, **kwargs):
+        if self.dialect.supports_cast:
+            return super().visit_cast(cast, **kwargs)
+        else:
+            return self.process(cast.clause, **kwargs)
+
+    def visit_extract(self, extract, **kw):
+        try:
+            return "CAST(STRFTIME('%s', %s) AS INTEGER)" % (
+                self.extract_map[extract.field],
+                self.process(extract.expr, **kw),
+            )
+        except KeyError as err:
+            raise exc.CompileError(
+                "%s is not a valid extract argument." % extract.field
+            ) from err
+
+    def returning_clause(
+        self,
+        stmt,
+        returning_cols,
+        *,
+        populate_result_map,
+        **kw,
+    ):
+        kw["include_table"] = False
+        return super().returning_clause(
+            stmt, returning_cols, populate_result_map=populate_result_map, **kw
+        )
+
+    def limit_clause(self, select, **kw):
+        text = ""
+        if select._limit_clause is not None:
+            text += "\n LIMIT " + self.process(select._limit_clause, **kw)
+        if select._offset_clause is not None:
+            if select._limit_clause is None:
+                text += "\n LIMIT " + self.process(sql.literal(-1))
+            text += " OFFSET " + self.process(select._offset_clause, **kw)
+        else:
+            text += " OFFSET " + self.process(sql.literal(0), **kw)
+        return text
+
+    def for_update_clause(self, select, **kw):
+        # sqlite has no "FOR UPDATE" AFAICT
+        return ""
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        kw["asfrom"] = True
+        return "FROM " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in extra_froms
+        )
+
+    def visit_is_distinct_from_binary(self, binary, operator, **kw):
+        return "%s IS NOT %s" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_is_not_distinct_from_binary(self, binary, operator, **kw):
+        return "%s IS %s" % (
+            self.process(binary.left),
+            self.process(binary.right),
+        )
+
+    def visit_json_getitem_op_binary(self, binary, operator, **kw):
+        if binary.type._type_affinity is sqltypes.JSON:
+            expr = "JSON_QUOTE(JSON_EXTRACT(%s, %s))"
+        else:
+            expr = "JSON_EXTRACT(%s, %s)"
+
+        return expr % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_json_path_getitem_op_binary(self, binary, operator, **kw):
+        if binary.type._type_affinity is sqltypes.JSON:
+            expr = "JSON_QUOTE(JSON_EXTRACT(%s, %s))"
+        else:
+            expr = "JSON_EXTRACT(%s, %s)"
+
+        return expr % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_empty_set_op_expr(self, type_, expand_op, **kw):
+        # slightly old SQLite versions don't seem to be able to handle
+        # the empty set impl
+        return self.visit_empty_set_expr(type_)
+
+    def visit_empty_set_expr(self, element_types, **kw):
+        return "SELECT %s FROM (SELECT %s) WHERE 1!=1" % (
+            ", ".join("1" for type_ in element_types or [INTEGER()]),
+            ", ".join("1" for type_ in element_types or [INTEGER()]),
+        )
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._generate_generic_binary(binary, " REGEXP ", **kw)
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._generate_generic_binary(binary, " NOT REGEXP ", **kw)
+
+    def _on_conflict_target(self, clause, **kw):
+        if clause.inferred_target_elements is not None:
+            target_text = "(%s)" % ", ".join(
+                (
+                    self.preparer.quote(c)
+                    if isinstance(c, str)
+                    else self.process(c, include_table=False, use_schema=False)
+                )
+                for c in clause.inferred_target_elements
+            )
+            if clause.inferred_target_whereclause is not None:
+                target_text += " WHERE %s" % self.process(
+                    clause.inferred_target_whereclause,
+                    include_table=False,
+                    use_schema=False,
+                    literal_execute=True,
+                )
+
+        else:
+            target_text = ""
+
+        return target_text
+
+    def visit_on_conflict_do_nothing(self, on_conflict, **kw):
+        target_text = self._on_conflict_target(on_conflict, **kw)
+
+        if target_text:
+            return "ON CONFLICT %s DO NOTHING" % target_text
+        else:
+            return "ON CONFLICT DO NOTHING"
+
+    def visit_on_conflict_do_update(self, on_conflict, **kw):
+        clause = on_conflict
+
+        target_text = self._on_conflict_target(on_conflict, **kw)
+
+        action_set_ops = []
+
+        set_parameters = dict(clause.update_values_to_set)
+        # create a list of column assignment clauses as tuples
+
+        insert_statement = self.stack[-1]["selectable"]
+        cols = insert_statement.table.c
+        for c in cols:
+            col_key = c.key
+
+            if col_key in set_parameters:
+                value = set_parameters.pop(col_key)
+            elif c in set_parameters:
+                value = set_parameters.pop(c)
+            else:
+                continue
+
+            if coercions._is_literal(value):
+                value = elements.BindParameter(None, value, type_=c.type)
+
+            else:
+                if (
+                    isinstance(value, elements.BindParameter)
+                    and value.type._isnull
+                ):
+                    value = value._clone()
+                    value.type = c.type
+            value_text = self.process(value.self_group(), use_schema=False)
+
+            key_text = self.preparer.quote(c.name)
+            action_set_ops.append("%s = %s" % (key_text, value_text))
+
+        # check for names that don't match columns
+        if set_parameters:
+            util.warn(
+                "Additional column names not matching "
+                "any column keys in table '%s': %s"
+                % (
+                    self.current_executable.table.name,
+                    (", ".join("'%s'" % c for c in set_parameters)),
+                )
+            )
+            for k, v in set_parameters.items():
+                key_text = (
+                    self.preparer.quote(k)
+                    if isinstance(k, str)
+                    else self.process(k, use_schema=False)
+                )
+                value_text = self.process(
+                    coercions.expect(roles.ExpressionElementRole, v),
+                    use_schema=False,
+                )
+                action_set_ops.append("%s = %s" % (key_text, value_text))
+
+        action_text = ", ".join(action_set_ops)
+        if clause.update_whereclause is not None:
+            action_text += " WHERE %s" % self.process(
+                clause.update_whereclause, include_table=True, use_schema=False
+            )
+
+        return "ON CONFLICT %s DO UPDATE SET %s" % (target_text, action_text)
+
+    def visit_bitwise_xor_op_binary(self, binary, operator, **kw):
+        # sqlite has no xor. Use "a XOR b" = "(a | b) - (a & b)".
+        kw["eager_grouping"] = True
+        or_ = self._generate_generic_binary(binary, " | ", **kw)
+        and_ = self._generate_generic_binary(binary, " & ", **kw)
+        return f"({or_} - {and_})"
+
+
+class SQLiteDDLCompiler(compiler.DDLCompiler):
+    def get_column_specification(self, column, **kwargs):
+        coltype = self.dialect.type_compiler_instance.process(
+            column.type, type_expression=column
+        )
+        colspec = self.preparer.format_column(column) + " " + coltype
+        default = self.get_column_default_string(column)
+        if default is not None:
+            if isinstance(column.server_default.arg, ColumnElement):
+                default = "(" + default + ")"
+            colspec += " DEFAULT " + default
+
+        if not column.nullable:
+            colspec += " NOT NULL"
+
+            on_conflict_clause = column.dialect_options["sqlite"][
+                "on_conflict_not_null"
+            ]
+            if on_conflict_clause is not None:
+                colspec += " ON CONFLICT " + on_conflict_clause
+
+        if column.primary_key:
+            if (
+                column.autoincrement is True
+                and len(column.table.primary_key.columns) != 1
+            ):
+                raise exc.CompileError(
+                    "SQLite does not support autoincrement for "
+                    "composite primary keys"
+                )
+
+            if (
+                column.table.dialect_options["sqlite"]["autoincrement"]
+                and len(column.table.primary_key.columns) == 1
+                and issubclass(column.type._type_affinity, sqltypes.Integer)
+                and not column.foreign_keys
+            ):
+                colspec += " PRIMARY KEY"
+
+                on_conflict_clause = column.dialect_options["sqlite"][
+                    "on_conflict_primary_key"
+                ]
+                if on_conflict_clause is not None:
+                    colspec += " ON CONFLICT " + on_conflict_clause
+
+                colspec += " AUTOINCREMENT"
+
+        if column.computed is not None:
+            colspec += " " + self.process(column.computed)
+
+        return colspec
+
+    def visit_primary_key_constraint(self, constraint, **kw):
+        # for columns with sqlite_autoincrement=True,
+        # the PRIMARY KEY constraint can only be inline
+        # with the column itself.
+        if len(constraint.columns) == 1:
+            c = list(constraint)[0]
+            if (
+                c.primary_key
+                and c.table.dialect_options["sqlite"]["autoincrement"]
+                and issubclass(c.type._type_affinity, sqltypes.Integer)
+                and not c.foreign_keys
+            ):
+                return None
+
+        text = super().visit_primary_key_constraint(constraint)
+
+        on_conflict_clause = constraint.dialect_options["sqlite"][
+            "on_conflict"
+        ]
+        if on_conflict_clause is None and len(constraint.columns) == 1:
+            on_conflict_clause = list(constraint)[0].dialect_options["sqlite"][
+                "on_conflict_primary_key"
+            ]
+
+        if on_conflict_clause is not None:
+            text += " ON CONFLICT " + on_conflict_clause
+
+        return text
+
+    def visit_unique_constraint(self, constraint, **kw):
+        text = super().visit_unique_constraint(constraint)
+
+        on_conflict_clause = constraint.dialect_options["sqlite"][
+            "on_conflict"
+        ]
+        if on_conflict_clause is None and len(constraint.columns) == 1:
+            col1 = list(constraint)[0]
+            if isinstance(col1, schema.SchemaItem):
+                on_conflict_clause = list(constraint)[0].dialect_options[
+                    "sqlite"
+                ]["on_conflict_unique"]
+
+        if on_conflict_clause is not None:
+            text += " ON CONFLICT " + on_conflict_clause
+
+        return text
+
+    def visit_check_constraint(self, constraint, **kw):
+        text = super().visit_check_constraint(constraint)
+
+        on_conflict_clause = constraint.dialect_options["sqlite"][
+            "on_conflict"
+        ]
+
+        if on_conflict_clause is not None:
+            text += " ON CONFLICT " + on_conflict_clause
+
+        return text
+
+    def visit_column_check_constraint(self, constraint, **kw):
+        text = super().visit_column_check_constraint(constraint)
+
+        if constraint.dialect_options["sqlite"]["on_conflict"] is not None:
+            raise exc.CompileError(
+                "SQLite does not support on conflict clause for "
+                "column check constraint"
+            )
+
+        return text
+
+    def visit_foreign_key_constraint(self, constraint, **kw):
+        local_table = constraint.elements[0].parent.table
+        remote_table = constraint.elements[0].column.table
+
+        if local_table.schema != remote_table.schema:
+            return None
+        else:
+            return super().visit_foreign_key_constraint(constraint)
+
+    def define_constraint_remote_table(self, constraint, table, preparer):
+        """Format the remote table clause of a CREATE CONSTRAINT clause."""
+
+        return preparer.format_table(table, use_schema=False)
+
+    def visit_create_index(
+        self, create, include_schema=False, include_table_schema=True, **kw
+    ):
+        index = create.element
+        self._verify_index_table(index)
+        preparer = self.preparer
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+
+        text += "INDEX "
+
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+
+        text += "%s ON %s (%s)" % (
+            self._prepared_index_name(index, include_schema=True),
+            preparer.format_table(index.table, use_schema=False),
+            ", ".join(
+                self.sql_compiler.process(
+                    expr, include_table=False, literal_binds=True
+                )
+                for expr in index.expressions
+            ),
+        )
+
+        whereclause = index.dialect_options["sqlite"]["where"]
+        if whereclause is not None:
+            where_compiled = self.sql_compiler.process(
+                whereclause, include_table=False, literal_binds=True
+            )
+            text += " WHERE " + where_compiled
+
+        return text
+
+    def post_create_table(self, table):
+        text = ""
+        if table.dialect_options["sqlite"]["with_rowid"] is False:
+            text += "\n WITHOUT ROWID"
+        if table.dialect_options["sqlite"]["strict"] is True:
+            text += "\n STRICT"
+        return text
+
+
+class SQLiteTypeCompiler(compiler.GenericTypeCompiler):
+    def visit_large_binary(self, type_, **kw):
+        return self.visit_BLOB(type_)
+
+    def visit_DATETIME(self, type_, **kw):
+        if (
+            not isinstance(type_, _DateTimeMixin)
+            or type_.format_is_text_affinity
+        ):
+            return super().visit_DATETIME(type_)
+        else:
+            return "DATETIME_CHAR"
+
+    def visit_DATE(self, type_, **kw):
+        if (
+            not isinstance(type_, _DateTimeMixin)
+            or type_.format_is_text_affinity
+        ):
+            return super().visit_DATE(type_)
+        else:
+            return "DATE_CHAR"
+
+    def visit_TIME(self, type_, **kw):
+        if (
+            not isinstance(type_, _DateTimeMixin)
+            or type_.format_is_text_affinity
+        ):
+            return super().visit_TIME(type_)
+        else:
+            return "TIME_CHAR"
+
+    def visit_JSON(self, type_, **kw):
+        # note this name provides NUMERIC affinity, not TEXT.
+        # should not be an issue unless the JSON value consists of a single
+        # numeric value.   JSONTEXT can be used if this case is required.
+        return "JSON"
+
+
+class SQLiteIdentifierPreparer(compiler.IdentifierPreparer):
+    reserved_words = {
+        "add",
+        "after",
+        "all",
+        "alter",
+        "analyze",
+        "and",
+        "as",
+        "asc",
+        "attach",
+        "autoincrement",
+        "before",
+        "begin",
+        "between",
+        "by",
+        "cascade",
+        "case",
+        "cast",
+        "check",
+        "collate",
+        "column",
+        "commit",
+        "conflict",
+        "constraint",
+        "create",
+        "cross",
+        "current_date",
+        "current_time",
+        "current_timestamp",
+        "database",
+        "default",
+        "deferrable",
+        "deferred",
+        "delete",
+        "desc",
+        "detach",
+        "distinct",
+        "drop",
+        "each",
+        "else",
+        "end",
+        "escape",
+        "except",
+        "exclusive",
+        "exists",
+        "explain",
+        "false",
+        "fail",
+        "for",
+        "foreign",
+        "from",
+        "full",
+        "glob",
+        "group",
+        "having",
+        "if",
+        "ignore",
+        "immediate",
+        "in",
+        "index",
+        "indexed",
+        "initially",
+        "inner",
+        "insert",
+        "instead",
+        "intersect",
+        "into",
+        "is",
+        "isnull",
+        "join",
+        "key",
+        "left",
+        "like",
+        "limit",
+        "match",
+        "natural",
+        "not",
+        "notnull",
+        "null",
+        "of",
+        "offset",
+        "on",
+        "or",
+        "order",
+        "outer",
+        "plan",
+        "pragma",
+        "primary",
+        "query",
+        "raise",
+        "references",
+        "reindex",
+        "rename",
+        "replace",
+        "restrict",
+        "right",
+        "rollback",
+        "row",
+        "select",
+        "set",
+        "table",
+        "temp",
+        "temporary",
+        "then",
+        "to",
+        "transaction",
+        "trigger",
+        "true",
+        "union",
+        "unique",
+        "update",
+        "using",
+        "vacuum",
+        "values",
+        "view",
+        "virtual",
+        "when",
+        "where",
+    }
+
+
+class SQLiteExecutionContext(default.DefaultExecutionContext):
+    @util.memoized_property
+    def _preserve_raw_colnames(self):
+        return (
+            not self.dialect._broken_dotted_colnames
+            or self.execution_options.get("sqlite_raw_colnames", False)
+        )
+
+    def _translate_colname(self, colname):
+        # TODO: detect SQLite version 3.10.0 or greater;
+        # see [ticket:3633]
+
+        # adjust for dotted column names.  SQLite
+        # in the case of UNION may store col names as
+        # "tablename.colname", or if using an attached database,
+        # "database.tablename.colname", in cursor.description
+        if not self._preserve_raw_colnames and "." in colname:
+            return colname.split(".")[-1], colname
+        else:
+            return colname, None
+
+
+class SQLiteDialect(default.DefaultDialect):
+    name = "sqlite"
+    supports_alter = False
+
+    # SQlite supports "DEFAULT VALUES" but *does not* support
+    # "VALUES (DEFAULT)"
+    supports_default_values = True
+    supports_default_metavalue = False
+
+    # sqlite issue:
+    # https://github.com/python/cpython/issues/93421
+    # note this parameter is no longer used by the ORM or default dialect
+    # see #9414
+    supports_sane_rowcount_returning = False
+
+    supports_empty_insert = False
+    supports_cast = True
+    supports_multivalues_insert = True
+    use_insertmanyvalues = True
+    tuple_in_values = True
+    supports_statement_cache = True
+    insert_null_pk_still_autoincrements = True
+    insert_returning = True
+    update_returning = True
+    update_returning_multifrom = True
+    delete_returning = True
+    update_returning_multifrom = True
+
+    supports_default_metavalue = True
+    """dialect supports INSERT... VALUES (DEFAULT) syntax"""
+
+    default_metavalue_token = "NULL"
+    """for INSERT... VALUES (DEFAULT) syntax, the token to put in the
+    parenthesis."""
+
+    default_paramstyle = "qmark"
+    execution_ctx_cls = SQLiteExecutionContext
+    statement_compiler = SQLiteCompiler
+    ddl_compiler = SQLiteDDLCompiler
+    type_compiler_cls = SQLiteTypeCompiler
+    preparer = SQLiteIdentifierPreparer
+    ischema_names = ischema_names
+    colspecs = colspecs
+
+    construct_arguments = [
+        (
+            sa_schema.Table,
+            {
+                "autoincrement": False,
+                "with_rowid": True,
+                "strict": False,
+            },
+        ),
+        (sa_schema.Index, {"where": None}),
+        (
+            sa_schema.Column,
+            {
+                "on_conflict_primary_key": None,
+                "on_conflict_not_null": None,
+                "on_conflict_unique": None,
+            },
+        ),
+        (sa_schema.Constraint, {"on_conflict": None}),
+    ]
+
+    _broken_fk_pragma_quotes = False
+    _broken_dotted_colnames = False
+
+    @util.deprecated_params(
+        _json_serializer=(
+            "1.3.7",
+            "The _json_serializer argument to the SQLite dialect has "
+            "been renamed to the correct name of json_serializer.  The old "
+            "argument name will be removed in a future release.",
+        ),
+        _json_deserializer=(
+            "1.3.7",
+            "The _json_deserializer argument to the SQLite dialect has "
+            "been renamed to the correct name of json_deserializer.  The old "
+            "argument name will be removed in a future release.",
+        ),
+    )
+    def __init__(
+        self,
+        native_datetime=False,
+        json_serializer=None,
+        json_deserializer=None,
+        _json_serializer=None,
+        _json_deserializer=None,
+        **kwargs,
+    ):
+        default.DefaultDialect.__init__(self, **kwargs)
+
+        if _json_serializer:
+            json_serializer = _json_serializer
+        if _json_deserializer:
+            json_deserializer = _json_deserializer
+        self._json_serializer = json_serializer
+        self._json_deserializer = json_deserializer
+
+        # this flag used by pysqlite dialect, and perhaps others in the
+        # future, to indicate the driver is handling date/timestamp
+        # conversions (and perhaps datetime/time as well on some hypothetical
+        # driver ?)
+        self.native_datetime = native_datetime
+
+        if self.dbapi is not None:
+            if self.dbapi.sqlite_version_info < (3, 7, 16):
+                util.warn(
+                    "SQLite version %s is older than 3.7.16, and will not "
+                    "support right nested joins, as are sometimes used in "
+                    "more complex ORM scenarios.  SQLAlchemy 1.4 and above "
+                    "no longer tries to rewrite these joins."
+                    % (self.dbapi.sqlite_version_info,)
+                )
+
+            # NOTE: python 3.7 on fedora for me has SQLite 3.34.1.  These
+            # version checks are getting very stale.
+            self._broken_dotted_colnames = self.dbapi.sqlite_version_info < (
+                3,
+                10,
+                0,
+            )
+            self.supports_default_values = self.dbapi.sqlite_version_info >= (
+                3,
+                3,
+                8,
+            )
+            self.supports_cast = self.dbapi.sqlite_version_info >= (3, 2, 3)
+            self.supports_multivalues_insert = (
+                # https://www.sqlite.org/releaselog/3_7_11.html
+                self.dbapi.sqlite_version_info
+                >= (3, 7, 11)
+            )
+            # see https://www.sqlalchemy.org/trac/ticket/2568
+            # as well as https://www.sqlite.org/src/info/600482d161
+            self._broken_fk_pragma_quotes = self.dbapi.sqlite_version_info < (
+                3,
+                6,
+                14,
+            )
+
+            if self.dbapi.sqlite_version_info < (3, 35) or util.pypy:
+                self.update_returning = self.delete_returning = (
+                    self.insert_returning
+                ) = False
+
+            if self.dbapi.sqlite_version_info < (3, 32, 0):
+                # https://www.sqlite.org/limits.html
+                self.insertmanyvalues_max_parameters = 999
+
+    _isolation_lookup = util.immutabledict(
+        {"READ UNCOMMITTED": 1, "SERIALIZABLE": 0}
+    )
+
+    def get_isolation_level_values(self, dbapi_connection):
+        return list(self._isolation_lookup)
+
+    def set_isolation_level(self, dbapi_connection, level):
+        isolation_level = self._isolation_lookup[level]
+
+        cursor = dbapi_connection.cursor()
+        cursor.execute(f"PRAGMA read_uncommitted = {isolation_level}")
+        cursor.close()
+
+    def get_isolation_level(self, dbapi_connection):
+        cursor = dbapi_connection.cursor()
+        cursor.execute("PRAGMA read_uncommitted")
+        res = cursor.fetchone()
+        if res:
+            value = res[0]
+        else:
+            # https://www.sqlite.org/changes.html#version_3_3_3
+            # "Optional READ UNCOMMITTED isolation (instead of the
+            # default isolation level of SERIALIZABLE) and
+            # table level locking when database connections
+            # share a common cache.""
+            # pre-SQLite 3.3.0 default to 0
+            value = 0
+        cursor.close()
+        if value == 0:
+            return "SERIALIZABLE"
+        elif value == 1:
+            return "READ UNCOMMITTED"
+        else:
+            assert False, "Unknown isolation level %s" % value
+
+    @reflection.cache
+    def get_schema_names(self, connection, **kw):
+        s = "PRAGMA database_list"
+        dl = connection.exec_driver_sql(s)
+
+        return [db[1] for db in dl if db[1] != "temp"]
+
+    def _format_schema(self, schema, table_name):
+        if schema is not None:
+            qschema = self.identifier_preparer.quote_identifier(schema)
+            name = f"{qschema}.{table_name}"
+        else:
+            name = table_name
+        return name
+
+    def _sqlite_main_query(
+        self,
+        table: str,
+        type_: str,
+        schema: Optional[str],
+        sqlite_include_internal: bool,
+    ):
+        main = self._format_schema(schema, table)
+        if not sqlite_include_internal:
+            filter_table = " AND name NOT LIKE 'sqlite~_%' ESCAPE '~'"
+        else:
+            filter_table = ""
+        query = (
+            f"SELECT name FROM {main} "
+            f"WHERE type='{type_}'{filter_table} "
+            "ORDER BY name"
+        )
+        return query
+
+    @reflection.cache
+    def get_table_names(
+        self, connection, schema=None, sqlite_include_internal=False, **kw
+    ):
+        query = self._sqlite_main_query(
+            "sqlite_master", "table", schema, sqlite_include_internal
+        )
+        names = connection.exec_driver_sql(query).scalars().all()
+        return names
+
+    @reflection.cache
+    def get_temp_table_names(
+        self, connection, sqlite_include_internal=False, **kw
+    ):
+        query = self._sqlite_main_query(
+            "sqlite_temp_master", "table", None, sqlite_include_internal
+        )
+        names = connection.exec_driver_sql(query).scalars().all()
+        return names
+
+    @reflection.cache
+    def get_temp_view_names(
+        self, connection, sqlite_include_internal=False, **kw
+    ):
+        query = self._sqlite_main_query(
+            "sqlite_temp_master", "view", None, sqlite_include_internal
+        )
+        names = connection.exec_driver_sql(query).scalars().all()
+        return names
+
+    @reflection.cache
+    def has_table(self, connection, table_name, schema=None, **kw):
+        self._ensure_has_table_connection(connection)
+
+        if schema is not None and schema not in self.get_schema_names(
+            connection, **kw
+        ):
+            return False
+
+        info = self._get_table_pragma(
+            connection, "table_info", table_name, schema=schema
+        )
+        return bool(info)
+
+    def _get_default_schema_name(self, connection):
+        return "main"
+
+    @reflection.cache
+    def get_view_names(
+        self, connection, schema=None, sqlite_include_internal=False, **kw
+    ):
+        query = self._sqlite_main_query(
+            "sqlite_master", "view", schema, sqlite_include_internal
+        )
+        names = connection.exec_driver_sql(query).scalars().all()
+        return names
+
+    @reflection.cache
+    def get_view_definition(self, connection, view_name, schema=None, **kw):
+        if schema is not None:
+            qschema = self.identifier_preparer.quote_identifier(schema)
+            master = f"{qschema}.sqlite_master"
+            s = ("SELECT sql FROM %s WHERE name = ? AND type='view'") % (
+                master,
+            )
+            rs = connection.exec_driver_sql(s, (view_name,))
+        else:
+            try:
+                s = (
+                    "SELECT sql FROM "
+                    " (SELECT * FROM sqlite_master UNION ALL "
+                    "  SELECT * FROM sqlite_temp_master) "
+                    "WHERE name = ? "
+                    "AND type='view'"
+                )
+                rs = connection.exec_driver_sql(s, (view_name,))
+            except exc.DBAPIError:
+                s = (
+                    "SELECT sql FROM sqlite_master WHERE name = ? "
+                    "AND type='view'"
+                )
+                rs = connection.exec_driver_sql(s, (view_name,))
+
+        result = rs.fetchall()
+        if result:
+            return result[0].sql
+        else:
+            raise exc.NoSuchTableError(
+                f"{schema}.{view_name}" if schema else view_name
+            )
+
+    @reflection.cache
+    def get_columns(self, connection, table_name, schema=None, **kw):
+        pragma = "table_info"
+        # computed columns are threaded as hidden, they require table_xinfo
+        if self.server_version_info >= (3, 31):
+            pragma = "table_xinfo"
+        info = self._get_table_pragma(
+            connection, pragma, table_name, schema=schema
+        )
+        columns = []
+        tablesql = None
+        for row in info:
+            name = row[1]
+            type_ = row[2].upper()
+            nullable = not row[3]
+            default = row[4]
+            primary_key = row[5]
+            hidden = row[6] if pragma == "table_xinfo" else 0
+
+            # hidden has value 0 for normal columns, 1 for hidden columns,
+            # 2 for computed virtual columns and 3 for computed stored columns
+            # https://www.sqlite.org/src/info/069351b85f9a706f60d3e98fbc8aaf40c374356b967c0464aede30ead3d9d18b
+            if hidden == 1:
+                continue
+
+            generated = bool(hidden)
+            persisted = hidden == 3
+
+            if tablesql is None and generated:
+                tablesql = self._get_table_sql(
+                    connection, table_name, schema, **kw
+                )
+                # remove create table
+                match = re.match(
+                    r"create table .*?\((.*)\)$",
+                    tablesql.strip(),
+                    re.DOTALL | re.IGNORECASE,
+                )
+                assert match, f"create table not found in {tablesql}"
+                tablesql = match.group(1).strip()
+
+            columns.append(
+                self._get_column_info(
+                    name,
+                    type_,
+                    nullable,
+                    default,
+                    primary_key,
+                    generated,
+                    persisted,
+                    tablesql,
+                )
+            )
+        if columns:
+            return columns
+        elif not self.has_table(connection, table_name, schema):
+            raise exc.NoSuchTableError(
+                f"{schema}.{table_name}" if schema else table_name
+            )
+        else:
+            return ReflectionDefaults.columns()
+
+    def _get_column_info(
+        self,
+        name,
+        type_,
+        nullable,
+        default,
+        primary_key,
+        generated,
+        persisted,
+        tablesql,
+    ):
+        if generated:
+            # the type of a column "cc INTEGER GENERATED ALWAYS AS (1 + 42)"
+            # somehow is "INTEGER GENERATED ALWAYS"
+            type_ = re.sub("generated", "", type_, flags=re.IGNORECASE)
+            type_ = re.sub("always", "", type_, flags=re.IGNORECASE).strip()
+
+        coltype = self._resolve_type_affinity(type_)
+
+        if default is not None:
+            default = str(default)
+
+        colspec = {
+            "name": name,
+            "type": coltype,
+            "nullable": nullable,
+            "default": default,
+            "primary_key": primary_key,
+        }
+        if generated:
+            sqltext = ""
+            if tablesql:
+                pattern = (
+                    r"[^,]*\s+GENERATED\s+ALWAYS\s+AS"
+                    r"\s+\((.*)\)\s*(?:virtual|stored)?"
+                )
+                match = re.search(
+                    re.escape(name) + pattern, tablesql, re.IGNORECASE
+                )
+                if match:
+                    sqltext = match.group(1)
+            colspec["computed"] = {"sqltext": sqltext, "persisted": persisted}
+        return colspec
+
+    def _resolve_type_affinity(self, type_):
+        """Return a data type from a reflected column, using affinity rules.
+
+        SQLite's goal for universal compatibility introduces some complexity
+        during reflection, as a column's defined type might not actually be a
+        type that SQLite understands - or indeed, my not be defined *at all*.
+        Internally, SQLite handles this with a 'data type affinity' for each
+        column definition, mapping to one of 'TEXT', 'NUMERIC', 'INTEGER',
+        'REAL', or 'NONE' (raw bits). The algorithm that determines this is
+        listed in https://www.sqlite.org/datatype3.html section 2.1.
+
+        This method allows SQLAlchemy to support that algorithm, while still
+        providing access to smarter reflection utilities by recognizing
+        column definitions that SQLite only supports through affinity (like
+        DATE and DOUBLE).
+
+        """
+        match = re.match(r"([\w ]+)(\(.*?\))?", type_)
+        if match:
+            coltype = match.group(1)
+            args = match.group(2)
+        else:
+            coltype = ""
+            args = ""
+
+        if coltype in self.ischema_names:
+            coltype = self.ischema_names[coltype]
+        elif "INT" in coltype:
+            coltype = sqltypes.INTEGER
+        elif "CHAR" in coltype or "CLOB" in coltype or "TEXT" in coltype:
+            coltype = sqltypes.TEXT
+        elif "BLOB" in coltype or not coltype:
+            coltype = sqltypes.NullType
+        elif "REAL" in coltype or "FLOA" in coltype or "DOUB" in coltype:
+            coltype = sqltypes.REAL
+        else:
+            coltype = sqltypes.NUMERIC
+
+        if args is not None:
+            args = re.findall(r"(\d+)", args)
+            try:
+                coltype = coltype(*[int(a) for a in args])
+            except TypeError:
+                util.warn(
+                    "Could not instantiate type %s with "
+                    "reflected arguments %s; using no arguments."
+                    % (coltype, args)
+                )
+                coltype = coltype()
+        else:
+            coltype = coltype()
+
+        return coltype
+
+    @reflection.cache
+    def get_pk_constraint(self, connection, table_name, schema=None, **kw):
+        constraint_name = None
+        table_data = self._get_table_sql(connection, table_name, schema=schema)
+        if table_data:
+            PK_PATTERN = r"CONSTRAINT (\w+) PRIMARY KEY"
+            result = re.search(PK_PATTERN, table_data, re.I)
+            constraint_name = result.group(1) if result else None
+
+        cols = self.get_columns(connection, table_name, schema, **kw)
+        # consider only pk columns. This also avoids sorting the cached
+        # value returned by get_columns
+        cols = [col for col in cols if col.get("primary_key", 0) > 0]
+        cols.sort(key=lambda col: col.get("primary_key"))
+        pkeys = [col["name"] for col in cols]
+
+        if pkeys:
+            return {"constrained_columns": pkeys, "name": constraint_name}
+        else:
+            return ReflectionDefaults.pk_constraint()
+
+    @reflection.cache
+    def get_foreign_keys(self, connection, table_name, schema=None, **kw):
+        # sqlite makes this *extremely difficult*.
+        # First, use the pragma to get the actual FKs.
+        pragma_fks = self._get_table_pragma(
+            connection, "foreign_key_list", table_name, schema=schema
+        )
+
+        fks = {}
+
+        for row in pragma_fks:
+            (numerical_id, rtbl, lcol, rcol) = (row[0], row[2], row[3], row[4])
+
+            if not rcol:
+                # no referred column, which means it was not named in the
+                # original DDL.  The referred columns of the foreign key
+                # constraint are therefore the primary key of the referred
+                # table.
+                try:
+                    referred_pk = self.get_pk_constraint(
+                        connection, rtbl, schema=schema, **kw
+                    )
+                    referred_columns = referred_pk["constrained_columns"]
+                except exc.NoSuchTableError:
+                    # ignore not existing parents
+                    referred_columns = []
+            else:
+                # note we use this list only if this is the first column
+                # in the constraint.  for subsequent columns we ignore the
+                # list and append "rcol" if present.
+                referred_columns = []
+
+            if self._broken_fk_pragma_quotes:
+                rtbl = re.sub(r"^[\"\[`\']|[\"\]`\']$", "", rtbl)
+
+            if numerical_id in fks:
+                fk = fks[numerical_id]
+            else:
+                fk = fks[numerical_id] = {
+                    "name": None,
+                    "constrained_columns": [],
+                    "referred_schema": schema,
+                    "referred_table": rtbl,
+                    "referred_columns": referred_columns,
+                    "options": {},
+                }
+                fks[numerical_id] = fk
+
+            fk["constrained_columns"].append(lcol)
+
+            if rcol:
+                fk["referred_columns"].append(rcol)
+
+        def fk_sig(constrained_columns, referred_table, referred_columns):
+            return (
+                tuple(constrained_columns)
+                + (referred_table,)
+                + tuple(referred_columns)
+            )
+
+        # then, parse the actual SQL and attempt to find DDL that matches
+        # the names as well.   SQLite saves the DDL in whatever format
+        # it was typed in as, so need to be liberal here.
+
+        keys_by_signature = {
+            fk_sig(
+                fk["constrained_columns"],
+                fk["referred_table"],
+                fk["referred_columns"],
+            ): fk
+            for fk in fks.values()
+        }
+
+        table_data = self._get_table_sql(connection, table_name, schema=schema)
+
+        def parse_fks():
+            if table_data is None:
+                # system tables, etc.
+                return
+
+            # note that we already have the FKs from PRAGMA above.  This whole
+            # regexp thing is trying to locate additional detail about the
+            # FKs, namely the name of the constraint and other options.
+            # so parsing the columns is really about matching it up to what
+            # we already have.
+            FK_PATTERN = (
+                r"(?:CONSTRAINT (\w+) +)?"
+                r"FOREIGN KEY *\( *(.+?) *\) +"
+                r'REFERENCES +(?:(?:"(.+?)")|([a-z0-9_]+)) *\( *((?:(?:"[^"]+"|[a-z0-9_]+) *(?:, *)?)+)\) *'  # noqa: E501
+                r"((?:ON (?:DELETE|UPDATE) "
+                r"(?:SET NULL|SET DEFAULT|CASCADE|RESTRICT|NO ACTION) *)*)"
+                r"((?:NOT +)?DEFERRABLE)?"
+                r"(?: +INITIALLY +(DEFERRED|IMMEDIATE))?"
+            )
+            for match in re.finditer(FK_PATTERN, table_data, re.I):
+                (
+                    constraint_name,
+                    constrained_columns,
+                    referred_quoted_name,
+                    referred_name,
+                    referred_columns,
+                    onupdatedelete,
+                    deferrable,
+                    initially,
+                ) = match.group(1, 2, 3, 4, 5, 6, 7, 8)
+                constrained_columns = list(
+                    self._find_cols_in_sig(constrained_columns)
+                )
+                if not referred_columns:
+                    referred_columns = constrained_columns
+                else:
+                    referred_columns = list(
+                        self._find_cols_in_sig(referred_columns)
+                    )
+                referred_name = referred_quoted_name or referred_name
+                options = {}
+
+                for token in re.split(r" *\bON\b *", onupdatedelete.upper()):
+                    if token.startswith("DELETE"):
+                        ondelete = token[6:].strip()
+                        if ondelete and ondelete != "NO ACTION":
+                            options["ondelete"] = ondelete
+                    elif token.startswith("UPDATE"):
+                        onupdate = token[6:].strip()
+                        if onupdate and onupdate != "NO ACTION":
+                            options["onupdate"] = onupdate
+
+                if deferrable:
+                    options["deferrable"] = "NOT" not in deferrable.upper()
+                if initially:
+                    options["initially"] = initially.upper()
+
+                yield (
+                    constraint_name,
+                    constrained_columns,
+                    referred_name,
+                    referred_columns,
+                    options,
+                )
+
+        fkeys = []
+
+        for (
+            constraint_name,
+            constrained_columns,
+            referred_name,
+            referred_columns,
+            options,
+        ) in parse_fks():
+            sig = fk_sig(constrained_columns, referred_name, referred_columns)
+            if sig not in keys_by_signature:
+                util.warn(
+                    "WARNING: SQL-parsed foreign key constraint "
+                    "'%s' could not be located in PRAGMA "
+                    "foreign_keys for table %s" % (sig, table_name)
+                )
+                continue
+            key = keys_by_signature.pop(sig)
+            key["name"] = constraint_name
+            key["options"] = options
+            fkeys.append(key)
+        # assume the remainders are the unnamed, inline constraints, just
+        # use them as is as it's extremely difficult to parse inline
+        # constraints
+        fkeys.extend(keys_by_signature.values())
+        if fkeys:
+            return fkeys
+        else:
+            return ReflectionDefaults.foreign_keys()
+
+    def _find_cols_in_sig(self, sig):
+        for match in re.finditer(r'(?:"(.+?)")|([a-z0-9_]+)', sig, re.I):
+            yield match.group(1) or match.group(2)
+
+    @reflection.cache
+    def get_unique_constraints(
+        self, connection, table_name, schema=None, **kw
+    ):
+        auto_index_by_sig = {}
+        for idx in self.get_indexes(
+            connection,
+            table_name,
+            schema=schema,
+            include_auto_indexes=True,
+            **kw,
+        ):
+            if not idx["name"].startswith("sqlite_autoindex"):
+                continue
+            sig = tuple(idx["column_names"])
+            auto_index_by_sig[sig] = idx
+
+        table_data = self._get_table_sql(
+            connection, table_name, schema=schema, **kw
+        )
+        unique_constraints = []
+
+        def parse_uqs():
+            if table_data is None:
+                return
+            UNIQUE_PATTERN = r'(?:CONSTRAINT "?(.+?)"? +)?UNIQUE *\((.+?)\)'
+            INLINE_UNIQUE_PATTERN = (
+                r'(?:(".+?")|(?:[\[`])?([a-z0-9_]+)(?:[\]`])?)[\t ]'
+                r"+[a-z0-9_ ]+?[\t ]+UNIQUE"
+            )
+
+            for match in re.finditer(UNIQUE_PATTERN, table_data, re.I):
+                name, cols = match.group(1, 2)
+                yield name, list(self._find_cols_in_sig(cols))
+
+            # we need to match inlines as well, as we seek to differentiate
+            # a UNIQUE constraint from a UNIQUE INDEX, even though these
+            # are kind of the same thing :)
+            for match in re.finditer(INLINE_UNIQUE_PATTERN, table_data, re.I):
+                cols = list(
+                    self._find_cols_in_sig(match.group(1) or match.group(2))
+                )
+                yield None, cols
+
+        for name, cols in parse_uqs():
+            sig = tuple(cols)
+            if sig in auto_index_by_sig:
+                auto_index_by_sig.pop(sig)
+                parsed_constraint = {"name": name, "column_names": cols}
+                unique_constraints.append(parsed_constraint)
+        # NOTE: auto_index_by_sig might not be empty here,
+        # the PRIMARY KEY may have an entry.
+        if unique_constraints:
+            return unique_constraints
+        else:
+            return ReflectionDefaults.unique_constraints()
+
+    @reflection.cache
+    def get_check_constraints(self, connection, table_name, schema=None, **kw):
+        table_data = self._get_table_sql(
+            connection, table_name, schema=schema, **kw
+        )
+
+        # NOTE NOTE NOTE
+        # DO NOT CHANGE THIS REGULAR EXPRESSION.   There is no known way
+        # to parse CHECK constraints that contain newlines themselves using
+        # regular expressions, and the approach here relies upon each
+        # individual
+        # CHECK constraint being on a single line by itself.   This
+        # necessarily makes assumptions as to how the CREATE TABLE
+        # was emitted.   A more comprehensive DDL parsing solution would be
+        # needed to improve upon the current situation. See #11840 for
+        # background
+        CHECK_PATTERN = r"(?:CONSTRAINT (.+) +)?CHECK *\( *(.+) *\),? *"
+        cks = []
+
+        for match in re.finditer(CHECK_PATTERN, table_data or "", re.I):
+
+            name = match.group(1)
+
+            if name:
+                name = re.sub(r'^"|"$', "", name)
+
+            cks.append({"sqltext": match.group(2), "name": name})
+        cks.sort(key=lambda d: d["name"] or "~")  # sort None as last
+        if cks:
+            return cks
+        else:
+            return ReflectionDefaults.check_constraints()
+
+    @reflection.cache
+    def get_indexes(self, connection, table_name, schema=None, **kw):
+        pragma_indexes = self._get_table_pragma(
+            connection, "index_list", table_name, schema=schema
+        )
+        indexes = []
+
+        # regular expression to extract the filter predicate of a partial
+        # index. this could fail to extract the predicate correctly on
+        # indexes created like
+        #   CREATE INDEX i ON t (col || ') where') WHERE col <> ''
+        # but as this function does not support expression-based indexes
+        # this case does not occur.
+        partial_pred_re = re.compile(r"\)\s+where\s+(.+)", re.IGNORECASE)
+
+        if schema:
+            schema_expr = "%s." % self.identifier_preparer.quote_identifier(
+                schema
+            )
+        else:
+            schema_expr = ""
+
+        include_auto_indexes = kw.pop("include_auto_indexes", False)
+        for row in pragma_indexes:
+            # ignore implicit primary key index.
+            # https://www.mail-archive.com/sqlite-users@sqlite.org/msg30517.html
+            if not include_auto_indexes and row[1].startswith(
+                "sqlite_autoindex"
+            ):
+                continue
+            indexes.append(
+                dict(
+                    name=row[1],
+                    column_names=[],
+                    unique=row[2],
+                    dialect_options={},
+                )
+            )
+
+            # check partial indexes
+            if len(row) >= 5 and row[4]:
+                s = (
+                    "SELECT sql FROM %(schema)ssqlite_master "
+                    "WHERE name = ? "
+                    "AND type = 'index'" % {"schema": schema_expr}
+                )
+                rs = connection.exec_driver_sql(s, (row[1],))
+                index_sql = rs.scalar()
+                predicate_match = partial_pred_re.search(index_sql)
+                if predicate_match is None:
+                    # unless the regex is broken this case shouldn't happen
+                    # because we know this is a partial index, so the
+                    # definition sql should match the regex
+                    util.warn(
+                        "Failed to look up filter predicate of "
+                        "partial index %s" % row[1]
+                    )
+                else:
+                    predicate = predicate_match.group(1)
+                    indexes[-1]["dialect_options"]["sqlite_where"] = text(
+                        predicate
+                    )
+
+        # loop thru unique indexes to get the column names.
+        for idx in list(indexes):
+            pragma_index = self._get_table_pragma(
+                connection, "index_info", idx["name"], schema=schema
+            )
+
+            for row in pragma_index:
+                if row[2] is None:
+                    util.warn(
+                        "Skipped unsupported reflection of "
+                        "expression-based index %s" % idx["name"]
+                    )
+                    indexes.remove(idx)
+                    break
+                else:
+                    idx["column_names"].append(row[2])
+
+        indexes.sort(key=lambda d: d["name"] or "~")  # sort None as last
+        if indexes:
+            return indexes
+        elif not self.has_table(connection, table_name, schema):
+            raise exc.NoSuchTableError(
+                f"{schema}.{table_name}" if schema else table_name
+            )
+        else:
+            return ReflectionDefaults.indexes()
+
+    def _is_sys_table(self, table_name):
+        return table_name in {
+            "sqlite_schema",
+            "sqlite_master",
+            "sqlite_temp_schema",
+            "sqlite_temp_master",
+        }
+
+    @reflection.cache
+    def _get_table_sql(self, connection, table_name, schema=None, **kw):
+        if schema:
+            schema_expr = "%s." % (
+                self.identifier_preparer.quote_identifier(schema)
+            )
+        else:
+            schema_expr = ""
+        try:
+            s = (
+                "SELECT sql FROM "
+                " (SELECT * FROM %(schema)ssqlite_master UNION ALL "
+                "  SELECT * FROM %(schema)ssqlite_temp_master) "
+                "WHERE name = ? "
+                "AND type in ('table', 'view')" % {"schema": schema_expr}
+            )
+            rs = connection.exec_driver_sql(s, (table_name,))
+        except exc.DBAPIError:
+            s = (
+                "SELECT sql FROM %(schema)ssqlite_master "
+                "WHERE name = ? "
+                "AND type in ('table', 'view')" % {"schema": schema_expr}
+            )
+            rs = connection.exec_driver_sql(s, (table_name,))
+        value = rs.scalar()
+        if value is None and not self._is_sys_table(table_name):
+            raise exc.NoSuchTableError(f"{schema_expr}{table_name}")
+        return value
+
+    def _get_table_pragma(self, connection, pragma, table_name, schema=None):
+        quote = self.identifier_preparer.quote_identifier
+        if schema is not None:
+            statements = [f"PRAGMA {quote(schema)}."]
+        else:
+            # because PRAGMA looks in all attached databases if no schema
+            # given, need to specify "main" schema, however since we want
+            # 'temp' tables in the same namespace as 'main', need to run
+            # the PRAGMA twice
+            statements = ["PRAGMA main.", "PRAGMA temp."]
+
+        qtable = quote(table_name)
+        for statement in statements:
+            statement = f"{statement}{pragma}({qtable})"
+            cursor = connection.exec_driver_sql(statement)
+            if not cursor._soft_closed:
+                # work around SQLite issue whereby cursor.description
+                # is blank when PRAGMA returns no rows:
+                # https://www.sqlite.org/cvstrac/tktview?tn=1884
+                result = cursor.fetchall()
+            else:
+                result = []
+            if result:
+                return result
+        else:
+            return []
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/dml.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/dml.py
new file mode 100644
index 00000000..84cdb8be
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/dml.py
@@ -0,0 +1,263 @@
+# dialects/sqlite/dml.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+from typing import Any
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Union
+
+from .._typing import _OnConflictIndexElementsT
+from .._typing import _OnConflictIndexWhereT
+from .._typing import _OnConflictSetT
+from .._typing import _OnConflictWhereT
+from ... import util
+from ...sql import coercions
+from ...sql import roles
+from ...sql import schema
+from ...sql._typing import _DMLTableArgument
+from ...sql.base import _exclusive_against
+from ...sql.base import _generative
+from ...sql.base import ColumnCollection
+from ...sql.base import ReadOnlyColumnCollection
+from ...sql.dml import Insert as StandardInsert
+from ...sql.elements import ClauseElement
+from ...sql.elements import ColumnElement
+from ...sql.elements import KeyedColumnElement
+from ...sql.elements import TextClause
+from ...sql.expression import alias
+from ...util.typing import Self
+
+__all__ = ("Insert", "insert")
+
+
+def insert(table: _DMLTableArgument) -> Insert:
+    """Construct a sqlite-specific variant :class:`_sqlite.Insert`
+    construct.
+
+    .. container:: inherited_member
+
+        The :func:`sqlalchemy.dialects.sqlite.insert` function creates
+        a :class:`sqlalchemy.dialects.sqlite.Insert`.  This class is based
+        on the dialect-agnostic :class:`_sql.Insert` construct which may
+        be constructed using the :func:`_sql.insert` function in
+        SQLAlchemy Core.
+
+    The :class:`_sqlite.Insert` construct includes additional methods
+    :meth:`_sqlite.Insert.on_conflict_do_update`,
+    :meth:`_sqlite.Insert.on_conflict_do_nothing`.
+
+    """
+    return Insert(table)
+
+
+class Insert(StandardInsert):
+    """SQLite-specific implementation of INSERT.
+
+    Adds methods for SQLite-specific syntaxes such as ON CONFLICT.
+
+    The :class:`_sqlite.Insert` object is created using the
+    :func:`sqlalchemy.dialects.sqlite.insert` function.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`sqlite_on_conflict_insert`
+
+    """
+
+    stringify_dialect = "sqlite"
+    inherit_cache = False
+
+    @util.memoized_property
+    def excluded(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """Provide the ``excluded`` namespace for an ON CONFLICT statement
+
+        SQLite's ON CONFLICT clause allows reference to the row that would
+        be inserted, known as ``excluded``.  This attribute provides
+        all columns in this row to be referenceable.
+
+        .. tip::  The :attr:`_sqlite.Insert.excluded` attribute is an instance
+            of :class:`_expression.ColumnCollection`, which provides an
+            interface the same as that of the :attr:`_schema.Table.c`
+            collection described at :ref:`metadata_tables_and_columns`.
+            With this collection, ordinary names are accessible like attributes
+            (e.g. ``stmt.excluded.some_column``), but special names and
+            dictionary method names should be accessed using indexed access,
+            such as ``stmt.excluded["column name"]`` or
+            ``stmt.excluded["values"]``.  See the docstring for
+            :class:`_expression.ColumnCollection` for further examples.
+
+        """
+        return alias(self.table, name="excluded").columns
+
+    _on_conflict_exclusive = _exclusive_against(
+        "_post_values_clause",
+        msgs={
+            "_post_values_clause": "This Insert construct already has "
+            "an ON CONFLICT clause established"
+        },
+    )
+
+    @_generative
+    @_on_conflict_exclusive
+    def on_conflict_do_update(
+        self,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+        set_: _OnConflictSetT = None,
+        where: _OnConflictWhereT = None,
+    ) -> Self:
+        r"""
+        Specifies a DO UPDATE SET action for ON CONFLICT clause.
+
+        :param index_elements:
+         A sequence consisting of string column names, :class:`_schema.Column`
+         objects, or other column expression objects that will be used
+         to infer a target index or unique constraint.
+
+        :param index_where:
+         Additional WHERE criterion that can be used to infer a
+         conditional target index.
+
+        :param set\_:
+         A dictionary or other mapping object
+         where the keys are either names of columns in the target table,
+         or :class:`_schema.Column` objects or other ORM-mapped columns
+         matching that of the target table, and expressions or literals
+         as values, specifying the ``SET`` actions to take.
+
+         .. versionadded:: 1.4 The
+            :paramref:`_sqlite.Insert.on_conflict_do_update.set_`
+            parameter supports :class:`_schema.Column` objects from the target
+            :class:`_schema.Table` as keys.
+
+         .. warning:: This dictionary does **not** take into account
+            Python-specified default UPDATE values or generation functions,
+            e.g. those specified using :paramref:`_schema.Column.onupdate`.
+            These values will not be exercised for an ON CONFLICT style of
+            UPDATE, unless they are manually specified in the
+            :paramref:`.Insert.on_conflict_do_update.set_` dictionary.
+
+        :param where:
+         Optional argument. An expression object representing a ``WHERE``
+         clause that restricts the rows affected by ``DO UPDATE SET``. Rows not
+         meeting the ``WHERE`` condition will not be updated (effectively a
+         ``DO NOTHING`` for those rows).
+
+        """
+
+        self._post_values_clause = OnConflictDoUpdate(
+            index_elements, index_where, set_, where
+        )
+        return self
+
+    @_generative
+    @_on_conflict_exclusive
+    def on_conflict_do_nothing(
+        self,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+    ) -> Self:
+        """
+        Specifies a DO NOTHING action for ON CONFLICT clause.
+
+        :param index_elements:
+         A sequence consisting of string column names, :class:`_schema.Column`
+         objects, or other column expression objects that will be used
+         to infer a target index or unique constraint.
+
+        :param index_where:
+         Additional WHERE criterion that can be used to infer a
+         conditional target index.
+
+        """
+
+        self._post_values_clause = OnConflictDoNothing(
+            index_elements, index_where
+        )
+        return self
+
+
+class OnConflictClause(ClauseElement):
+    stringify_dialect = "sqlite"
+
+    inferred_target_elements: Optional[List[Union[str, schema.Column[Any]]]]
+    inferred_target_whereclause: Optional[
+        Union[ColumnElement[Any], TextClause]
+    ]
+
+    def __init__(
+        self,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+    ):
+        if index_elements is not None:
+            self.inferred_target_elements = [
+                coercions.expect(roles.DDLConstraintColumnRole, column)
+                for column in index_elements
+            ]
+            self.inferred_target_whereclause = (
+                coercions.expect(
+                    roles.WhereHavingRole,
+                    index_where,
+                )
+                if index_where is not None
+                else None
+            )
+        else:
+            self.inferred_target_elements = (
+                self.inferred_target_whereclause
+            ) = None
+
+
+class OnConflictDoNothing(OnConflictClause):
+    __visit_name__ = "on_conflict_do_nothing"
+
+
+class OnConflictDoUpdate(OnConflictClause):
+    __visit_name__ = "on_conflict_do_update"
+
+    update_values_to_set: List[Tuple[Union[schema.Column[Any], str], Any]]
+    update_whereclause: Optional[ColumnElement[Any]]
+
+    def __init__(
+        self,
+        index_elements: _OnConflictIndexElementsT = None,
+        index_where: _OnConflictIndexWhereT = None,
+        set_: _OnConflictSetT = None,
+        where: _OnConflictWhereT = None,
+    ):
+        super().__init__(
+            index_elements=index_elements,
+            index_where=index_where,
+        )
+
+        if isinstance(set_, dict):
+            if not set_:
+                raise ValueError("set parameter dictionary must not be empty")
+        elif isinstance(set_, ColumnCollection):
+            set_ = dict(set_)
+        else:
+            raise ValueError(
+                "set parameter must be a non-empty dictionary "
+                "or a ColumnCollection such as the `.c.` collection "
+                "of a Table object"
+            )
+        self.update_values_to_set = [
+            (coercions.expect(roles.DMLColumnRole, key), value)
+            for key, value in set_.items()
+        ]
+        self.update_whereclause = (
+            coercions.expect(roles.WhereHavingRole, where)
+            if where is not None
+            else None
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/json.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/json.py
new file mode 100644
index 00000000..02f4ea4c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/json.py
@@ -0,0 +1,92 @@
+# dialects/sqlite/json.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from ... import types as sqltypes
+
+
+class JSON(sqltypes.JSON):
+    """SQLite JSON type.
+
+    SQLite supports JSON as of version 3.9 through its JSON1_ extension. Note
+    that JSON1_ is a
+    `loadable extension <https://www.sqlite.org/loadext.html>`_ and as such
+    may not be available, or may require run-time loading.
+
+    :class:`_sqlite.JSON` is used automatically whenever the base
+    :class:`_types.JSON` datatype is used against a SQLite backend.
+
+    .. seealso::
+
+        :class:`_types.JSON` - main documentation for the generic
+        cross-platform JSON datatype.
+
+    The :class:`_sqlite.JSON` type supports persistence of JSON values
+    as well as the core index operations provided by :class:`_types.JSON`
+    datatype, by adapting the operations to render the ``JSON_EXTRACT``
+    function wrapped in the ``JSON_QUOTE`` function at the database level.
+    Extracted values are quoted in order to ensure that the results are
+    always JSON string values.
+
+
+    .. versionadded:: 1.3
+
+
+    .. _JSON1: https://www.sqlite.org/json1.html
+
+    """
+
+
+# Note: these objects currently match exactly those of MySQL, however since
+# these are not generalizable to all JSON implementations, remain separately
+# implemented for each dialect.
+class _FormatTypeMixin:
+    def _format_value(self, value):
+        raise NotImplementedError()
+
+    def bind_processor(self, dialect):
+        super_proc = self.string_bind_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+    def literal_processor(self, dialect):
+        super_proc = self.string_literal_processor(dialect)
+
+        def process(value):
+            value = self._format_value(value)
+            if super_proc:
+                value = super_proc(value)
+            return value
+
+        return process
+
+
+class JSONIndexType(_FormatTypeMixin, sqltypes.JSON.JSONIndexType):
+    def _format_value(self, value):
+        if isinstance(value, int):
+            value = "$[%s]" % value
+        else:
+            value = '$."%s"' % value
+        return value
+
+
+class JSONPathType(_FormatTypeMixin, sqltypes.JSON.JSONPathType):
+    def _format_value(self, value):
+        return "$%s" % (
+            "".join(
+                [
+                    "[%s]" % elem if isinstance(elem, int) else '."%s"' % elem
+                    for elem in value
+                ]
+            )
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/provision.py
new file mode 100644
index 00000000..97f882e7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/provision.py
@@ -0,0 +1,198 @@
+# dialects/sqlite/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import os
+import re
+
+from ... import exc
+from ...engine import url as sa_url
+from ...testing.provision import create_db
+from ...testing.provision import drop_db
+from ...testing.provision import follower_url_from_main
+from ...testing.provision import generate_driver_url
+from ...testing.provision import log
+from ...testing.provision import post_configure_engine
+from ...testing.provision import run_reap_dbs
+from ...testing.provision import stop_test_class_outside_fixtures
+from ...testing.provision import temp_table_keyword_args
+from ...testing.provision import upsert
+
+
+# TODO: I can't get this to build dynamically with pytest-xdist procs
+_drivernames = {
+    "pysqlite",
+    "aiosqlite",
+    "pysqlcipher",
+    "pysqlite_numeric",
+    "pysqlite_dollar",
+}
+
+
+def _format_url(url, driver, ident):
+    """given a sqlite url + desired driver + ident, make a canonical
+    URL out of it
+
+    """
+    url = sa_url.make_url(url)
+
+    if driver is None:
+        driver = url.get_driver_name()
+
+    filename = url.database
+
+    needs_enc = driver == "pysqlcipher"
+    name_token = None
+
+    if filename and filename != ":memory:":
+        assert "test_schema" not in filename
+        tokens = re.split(r"[_\.]", filename)
+
+        new_filename = f"{driver}"
+
+        for token in tokens:
+            if token in _drivernames:
+                if driver is None:
+                    driver = token
+                continue
+            elif token in ("db", "enc"):
+                continue
+            elif name_token is None:
+                name_token = token.strip("_")
+
+        assert name_token, f"sqlite filename has no name token: {url.database}"
+
+        new_filename = f"{name_token}_{driver}"
+        if ident:
+            new_filename += f"_{ident}"
+        new_filename += ".db"
+        if needs_enc:
+            new_filename += ".enc"
+        url = url.set(database=new_filename)
+
+    if needs_enc:
+        url = url.set(password="test")
+
+    url = url.set(drivername="sqlite+%s" % (driver,))
+
+    return url
+
+
+@generate_driver_url.for_db("sqlite")
+def generate_driver_url(url, driver, query_str):
+    url = _format_url(url, driver, None)
+
+    try:
+        url.get_dialect()
+    except exc.NoSuchModuleError:
+        return None
+    else:
+        return url
+
+
+@follower_url_from_main.for_db("sqlite")
+def _sqlite_follower_url_from_main(url, ident):
+    return _format_url(url, None, ident)
+
+
+@post_configure_engine.for_db("sqlite")
+def _sqlite_post_configure_engine(url, engine, follower_ident):
+    from sqlalchemy import event
+
+    if follower_ident:
+        attach_path = f"{follower_ident}_{engine.driver}_test_schema.db"
+    else:
+        attach_path = f"{engine.driver}_test_schema.db"
+
+    @event.listens_for(engine, "connect")
+    def connect(dbapi_connection, connection_record):
+        # use file DBs in all cases, memory acts kind of strangely
+        # as an attached
+
+        # NOTE!  this has to be done *per connection*.  New sqlite connection,
+        # as we get with say, QueuePool, the attaches are gone.
+        # so schemes to delete those attached files have to be done at the
+        # filesystem level and not rely upon what attachments are in a
+        # particular SQLite connection
+        dbapi_connection.execute(
+            f'ATTACH DATABASE "{attach_path}" AS test_schema'
+        )
+
+    @event.listens_for(engine, "engine_disposed")
+    def dispose(engine):
+        """most databases should be dropped using
+        stop_test_class_outside_fixtures
+
+        however a few tests like AttachedDBTest might not get triggered on
+        that main hook
+
+        """
+
+        if os.path.exists(attach_path):
+            os.remove(attach_path)
+
+        filename = engine.url.database
+
+        if filename and filename != ":memory:" and os.path.exists(filename):
+            os.remove(filename)
+
+
+@create_db.for_db("sqlite")
+def _sqlite_create_db(cfg, eng, ident):
+    pass
+
+
+@drop_db.for_db("sqlite")
+def _sqlite_drop_db(cfg, eng, ident):
+    _drop_dbs_w_ident(eng.url.database, eng.driver, ident)
+
+
+def _drop_dbs_w_ident(databasename, driver, ident):
+    for path in os.listdir("."):
+        fname, ext = os.path.split(path)
+        if ident in fname and ext in [".db", ".db.enc"]:
+            log.info("deleting SQLite database file: %s", path)
+            os.remove(path)
+
+
+@stop_test_class_outside_fixtures.for_db("sqlite")
+def stop_test_class_outside_fixtures(config, db, cls):
+    db.dispose()
+
+
+@temp_table_keyword_args.for_db("sqlite")
+def _sqlite_temp_table_keyword_args(cfg, eng):
+    return {"prefixes": ["TEMPORARY"]}
+
+
+@run_reap_dbs.for_db("sqlite")
+def _reap_sqlite_dbs(url, idents):
+    log.info("db reaper connecting to %r", url)
+    log.info("identifiers in file: %s", ", ".join(idents))
+    url = sa_url.make_url(url)
+    for ident in idents:
+        for drivername in _drivernames:
+            _drop_dbs_w_ident(url.database, drivername, ident)
+
+
+@upsert.for_db("sqlite")
+def _upsert(
+    cfg, table, returning, *, set_lambda=None, sort_by_parameter_order=False
+):
+    from sqlalchemy.dialects.sqlite import insert
+
+    stmt = insert(table)
+
+    if set_lambda:
+        stmt = stmt.on_conflict_do_update(set_=set_lambda(stmt.excluded))
+    else:
+        stmt = stmt.on_conflict_do_nothing()
+
+    stmt = stmt.returning(
+        *returning, sort_by_parameter_order=sort_by_parameter_order
+    )
+    return stmt
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlcipher.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlcipher.py
new file mode 100644
index 00000000..7a3dc1ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlcipher.py
@@ -0,0 +1,157 @@
+# dialects/sqlite/pysqlcipher.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""
+.. dialect:: sqlite+pysqlcipher
+    :name: pysqlcipher
+    :dbapi: sqlcipher 3 or pysqlcipher
+    :connectstring: sqlite+pysqlcipher://:passphrase@/file_path[?kdf_iter=<iter>]
+
+    Dialect for support of DBAPIs that make use of the
+    `SQLCipher <https://www.zetetic.net/sqlcipher>`_ backend.
+
+
+Driver
+------
+
+Current dialect selection logic is:
+
+* If the :paramref:`_sa.create_engine.module` parameter supplies a DBAPI module,
+  that module is used.
+* Otherwise for Python 3, choose https://pypi.org/project/sqlcipher3/
+* If not available, fall back to https://pypi.org/project/pysqlcipher3/
+* For Python 2, https://pypi.org/project/pysqlcipher/ is used.
+
+.. warning:: The ``pysqlcipher3`` and ``pysqlcipher`` DBAPI drivers are no
+   longer maintained; the ``sqlcipher3`` driver as of this writing appears
+   to be current.  For future compatibility, any pysqlcipher-compatible DBAPI
+   may be used as follows::
+
+        import sqlcipher_compatible_driver
+
+        from sqlalchemy import create_engine
+
+        e = create_engine(
+            "sqlite+pysqlcipher://:password@/dbname.db",
+            module=sqlcipher_compatible_driver,
+        )
+
+These drivers make use of the SQLCipher engine. This system essentially
+introduces new PRAGMA commands to SQLite which allows the setting of a
+passphrase and other encryption parameters, allowing the database file to be
+encrypted.
+
+
+Connect Strings
+---------------
+
+The format of the connect string is in every way the same as that
+of the :mod:`~sqlalchemy.dialects.sqlite.pysqlite` driver, except that the
+"password" field is now accepted, which should contain a passphrase::
+
+    e = create_engine("sqlite+pysqlcipher://:testing@/foo.db")
+
+For an absolute file path, two leading slashes should be used for the
+database name::
+
+    e = create_engine("sqlite+pysqlcipher://:testing@//path/to/foo.db")
+
+A selection of additional encryption-related pragmas supported by SQLCipher
+as documented at https://www.zetetic.net/sqlcipher/sqlcipher-api/ can be passed
+in the query string, and will result in that PRAGMA being called for each
+new connection.  Currently, ``cipher``, ``kdf_iter``
+``cipher_page_size`` and ``cipher_use_hmac`` are supported::
+
+    e = create_engine(
+        "sqlite+pysqlcipher://:testing@/foo.db?cipher=aes-256-cfb&kdf_iter=64000"
+    )
+
+.. warning:: Previous versions of sqlalchemy did not take into consideration
+   the encryption-related pragmas passed in the url string, that were silently
+   ignored. This may cause errors when opening files saved by a
+   previous sqlalchemy version if the encryption options do not match.
+
+
+Pooling Behavior
+----------------
+
+The driver makes a change to the default pool behavior of pysqlite
+as described in :ref:`pysqlite_threading_pooling`.   The pysqlcipher driver
+has been observed to be significantly slower on connection than the
+pysqlite driver, most likely due to the encryption overhead, so the
+dialect here defaults to using the :class:`.SingletonThreadPool`
+implementation,
+instead of the :class:`.NullPool` pool used by pysqlite.  As always, the pool
+implementation is entirely configurable using the
+:paramref:`_sa.create_engine.poolclass` parameter; the :class:`.
+StaticPool` may
+be more feasible for single-threaded use, or :class:`.NullPool` may be used
+to prevent unencrypted connections from being held open for long periods of
+time, at the expense of slower startup time for new connections.
+
+
+"""  # noqa
+
+from .pysqlite import SQLiteDialect_pysqlite
+from ... import pool
+
+
+class SQLiteDialect_pysqlcipher(SQLiteDialect_pysqlite):
+    driver = "pysqlcipher"
+    supports_statement_cache = True
+
+    pragmas = ("kdf_iter", "cipher", "cipher_page_size", "cipher_use_hmac")
+
+    @classmethod
+    def import_dbapi(cls):
+        try:
+            import sqlcipher3 as sqlcipher
+        except ImportError:
+            pass
+        else:
+            return sqlcipher
+
+        from pysqlcipher3 import dbapi2 as sqlcipher
+
+        return sqlcipher
+
+    @classmethod
+    def get_pool_class(cls, url):
+        return pool.SingletonThreadPool
+
+    def on_connect_url(self, url):
+        super_on_connect = super().on_connect_url(url)
+
+        # pull the info we need from the URL early.  Even though URL
+        # is immutable, we don't want any in-place changes to the URL
+        # to affect things
+        passphrase = url.password or ""
+        url_query = dict(url.query)
+
+        def on_connect(conn):
+            cursor = conn.cursor()
+            cursor.execute('pragma key="%s"' % passphrase)
+            for prag in self.pragmas:
+                value = url_query.get(prag, None)
+                if value is not None:
+                    cursor.execute('pragma %s="%s"' % (prag, value))
+            cursor.close()
+
+            if super_on_connect:
+                super_on_connect(conn)
+
+        return on_connect
+
+    def create_connect_args(self, url):
+        plain_url = url._replace(password=None)
+        plain_url = plain_url.difference_update_query(self.pragmas)
+        return super().create_connect_args(plain_url)
+
+
+dialect = SQLiteDialect_pysqlcipher
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlite.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlite.py
new file mode 100644
index 00000000..9dafda6d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/sqlite/pysqlite.py
@@ -0,0 +1,771 @@
+# dialects/sqlite/pysqlite.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+r"""
+.. dialect:: sqlite+pysqlite
+    :name: pysqlite
+    :dbapi: sqlite3
+    :connectstring: sqlite+pysqlite:///file_path
+    :url: https://docs.python.org/library/sqlite3.html
+
+    Note that ``pysqlite`` is the same driver as the ``sqlite3``
+    module included with the Python distribution.
+
+Driver
+------
+
+The ``sqlite3`` Python DBAPI is standard on all modern Python versions;
+for cPython and Pypy, no additional installation is necessary.
+
+
+Connect Strings
+---------------
+
+The file specification for the SQLite database is taken as the "database"
+portion of the URL.  Note that the format of a SQLAlchemy url is:
+
+.. sourcecode:: text
+
+    driver://user:pass@host/database
+
+This means that the actual filename to be used starts with the characters to
+the **right** of the third slash.   So connecting to a relative filepath
+looks like::
+
+    # relative path
+    e = create_engine("sqlite:///path/to/database.db")
+
+An absolute path, which is denoted by starting with a slash, means you
+need **four** slashes::
+
+    # absolute path
+    e = create_engine("sqlite:////path/to/database.db")
+
+To use a Windows path, regular drive specifications and backslashes can be
+used. Double backslashes are probably needed::
+
+    # absolute path on Windows
+    e = create_engine("sqlite:///C:\\path\\to\\database.db")
+
+To use sqlite ``:memory:`` database specify it as the filename using
+``sqlite:///:memory:``. It's also the default if no filepath is
+present, specifying only ``sqlite://`` and nothing else::
+
+    # in-memory database (note three slashes)
+    e = create_engine("sqlite:///:memory:")
+    # also in-memory database
+    e2 = create_engine("sqlite://")
+
+.. _pysqlite_uri_connections:
+
+URI Connections
+^^^^^^^^^^^^^^^
+
+Modern versions of SQLite support an alternative system of connecting using a
+`driver level URI <https://www.sqlite.org/uri.html>`_, which has the  advantage
+that additional driver-level arguments can be passed including options such as
+"read only".   The Python sqlite3 driver supports this mode under modern Python
+3 versions.   The SQLAlchemy pysqlite driver supports this mode of use by
+specifying "uri=true" in the URL query string.  The SQLite-level "URI" is kept
+as the "database" portion of the SQLAlchemy url (that is, following a slash)::
+
+    e = create_engine("sqlite:///file:path/to/database?mode=ro&uri=true")
+
+.. note::  The "uri=true" parameter must appear in the **query string**
+   of the URL.  It will not currently work as expected if it is only
+   present in the :paramref:`_sa.create_engine.connect_args`
+   parameter dictionary.
+
+The logic reconciles the simultaneous presence of SQLAlchemy's query string and
+SQLite's query string by separating out the parameters that belong to the
+Python sqlite3 driver vs. those that belong to the SQLite URI.  This is
+achieved through the use of a fixed list of parameters known to be accepted by
+the Python side of the driver.  For example, to include a URL that indicates
+the Python sqlite3 "timeout" and "check_same_thread" parameters, along with the
+SQLite "mode" and "nolock" parameters, they can all be passed together on the
+query string::
+
+    e = create_engine(
+        "sqlite:///file:path/to/database?"
+        "check_same_thread=true&timeout=10&mode=ro&nolock=1&uri=true"
+    )
+
+Above, the pysqlite / sqlite3 DBAPI would be passed arguments as::
+
+    sqlite3.connect(
+        "file:path/to/database?mode=ro&nolock=1",
+        check_same_thread=True,
+        timeout=10,
+        uri=True,
+    )
+
+Regarding future parameters added to either the Python or native drivers. new
+parameter names added to the SQLite URI scheme should be automatically
+accommodated by this scheme.  New parameter names added to the Python driver
+side can be accommodated by specifying them in the
+:paramref:`_sa.create_engine.connect_args` dictionary,
+until dialect support is
+added by SQLAlchemy.   For the less likely case that the native SQLite driver
+adds a new parameter name that overlaps with one of the existing, known Python
+driver parameters (such as "timeout" perhaps), SQLAlchemy's dialect would
+require adjustment for the URL scheme to continue to support this.
+
+As is always the case for all SQLAlchemy dialects, the entire "URL" process
+can be bypassed in :func:`_sa.create_engine` through the use of the
+:paramref:`_sa.create_engine.creator`
+parameter which allows for a custom callable
+that creates a Python sqlite3 driver level connection directly.
+
+.. versionadded:: 1.3.9
+
+.. seealso::
+
+    `Uniform Resource Identifiers <https://www.sqlite.org/uri.html>`_ - in
+    the SQLite documentation
+
+.. _pysqlite_regexp:
+
+Regular Expression Support
+---------------------------
+
+.. versionadded:: 1.4
+
+Support for the :meth:`_sql.ColumnOperators.regexp_match` operator is provided
+using Python's re.search_ function.  SQLite itself does not include a working
+regular expression operator; instead, it includes a non-implemented placeholder
+operator ``REGEXP`` that calls a user-defined function that must be provided.
+
+SQLAlchemy's implementation makes use of the pysqlite create_function_ hook
+as follows::
+
+
+    def regexp(a, b):
+        return re.search(a, b) is not None
+
+
+    sqlite_connection.create_function(
+        "regexp",
+        2,
+        regexp,
+    )
+
+There is currently no support for regular expression flags as a separate
+argument, as these are not supported by SQLite's REGEXP operator, however these
+may be included inline within the regular expression string.  See `Python regular expressions`_ for
+details.
+
+.. seealso::
+
+    `Python regular expressions`_: Documentation for Python's regular expression syntax.
+
+.. _create_function: https://docs.python.org/3/library/sqlite3.html#sqlite3.Connection.create_function
+
+.. _re.search: https://docs.python.org/3/library/re.html#re.search
+
+.. _Python regular expressions: https://docs.python.org/3/library/re.html#re.search
+
+
+
+Compatibility with sqlite3 "native" date and datetime types
+-----------------------------------------------------------
+
+The pysqlite driver includes the sqlite3.PARSE_DECLTYPES and
+sqlite3.PARSE_COLNAMES options, which have the effect of any column
+or expression explicitly cast as "date" or "timestamp" will be converted
+to a Python date or datetime object.  The date and datetime types provided
+with the pysqlite dialect are not currently compatible with these options,
+since they render the ISO date/datetime including microseconds, which
+pysqlite's driver does not.   Additionally, SQLAlchemy does not at
+this time automatically render the "cast" syntax required for the
+freestanding functions "current_timestamp" and "current_date" to return
+datetime/date types natively.   Unfortunately, pysqlite
+does not provide the standard DBAPI types in ``cursor.description``,
+leaving SQLAlchemy with no way to detect these types on the fly
+without expensive per-row type checks.
+
+Keeping in mind that pysqlite's parsing option is not recommended,
+nor should be necessary, for use with SQLAlchemy, usage of PARSE_DECLTYPES
+can be forced if one configures "native_datetime=True" on create_engine()::
+
+    engine = create_engine(
+        "sqlite://",
+        connect_args={
+            "detect_types": sqlite3.PARSE_DECLTYPES | sqlite3.PARSE_COLNAMES
+        },
+        native_datetime=True,
+    )
+
+With this flag enabled, the DATE and TIMESTAMP types (but note - not the
+DATETIME or TIME types...confused yet ?) will not perform any bind parameter
+or result processing. Execution of "func.current_date()" will return a string.
+"func.current_timestamp()" is registered as returning a DATETIME type in
+SQLAlchemy, so this function still receives SQLAlchemy-level result
+processing.
+
+.. _pysqlite_threading_pooling:
+
+Threading/Pooling Behavior
+---------------------------
+
+The ``sqlite3`` DBAPI by default prohibits the use of a particular connection
+in a thread which is not the one in which it was created.  As SQLite has
+matured, it's behavior under multiple threads has improved, and even includes
+options for memory only databases to be used in multiple threads.
+
+The thread prohibition is known as "check same thread" and may be controlled
+using the ``sqlite3`` parameter ``check_same_thread``, which will disable or
+enable this check. SQLAlchemy's default behavior here is to set
+``check_same_thread`` to ``False`` automatically whenever a file-based database
+is in use, to establish compatibility with the default pool class
+:class:`.QueuePool`.
+
+The SQLAlchemy ``pysqlite`` DBAPI establishes the connection pool differently
+based on the kind of SQLite database that's requested:
+
+* When a ``:memory:`` SQLite database is specified, the dialect by default
+  will use :class:`.SingletonThreadPool`. This pool maintains a single
+  connection per thread, so that all access to the engine within the current
+  thread use the same ``:memory:`` database - other threads would access a
+  different ``:memory:`` database.  The ``check_same_thread`` parameter
+  defaults to ``True``.
+* When a file-based database is specified, the dialect will use
+  :class:`.QueuePool` as the source of connections.   at the same time,
+  the ``check_same_thread`` flag is set to False by default unless overridden.
+
+  .. versionchanged:: 2.0
+
+    SQLite file database engines now use :class:`.QueuePool` by default.
+    Previously, :class:`.NullPool` were used.  The :class:`.NullPool` class
+    may be used by specifying it via the
+    :paramref:`_sa.create_engine.poolclass` parameter.
+
+Disabling Connection Pooling for File Databases
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Pooling may be disabled for a file based database by specifying the
+:class:`.NullPool` implementation for the :func:`_sa.create_engine.poolclass`
+parameter::
+
+    from sqlalchemy import NullPool
+
+    engine = create_engine("sqlite:///myfile.db", poolclass=NullPool)
+
+It's been observed that the :class:`.NullPool` implementation incurs an
+extremely small performance overhead for repeated checkouts due to the lack of
+connection re-use implemented by :class:`.QueuePool`.  However, it still
+may be beneficial to use this class if the application is experiencing
+issues with files being locked.
+
+Using a Memory Database in Multiple Threads
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+To use a ``:memory:`` database in a multithreaded scenario, the same
+connection object must be shared among threads, since the database exists
+only within the scope of that connection.   The
+:class:`.StaticPool` implementation will maintain a single connection
+globally, and the ``check_same_thread`` flag can be passed to Pysqlite
+as ``False``::
+
+    from sqlalchemy.pool import StaticPool
+
+    engine = create_engine(
+        "sqlite://",
+        connect_args={"check_same_thread": False},
+        poolclass=StaticPool,
+    )
+
+Note that using a ``:memory:`` database in multiple threads requires a recent
+version of SQLite.
+
+Using Temporary Tables with SQLite
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Due to the way SQLite deals with temporary tables, if you wish to use a
+temporary table in a file-based SQLite database across multiple checkouts
+from the connection pool, such as when using an ORM :class:`.Session` where
+the temporary table should continue to remain after :meth:`.Session.commit` or
+:meth:`.Session.rollback` is called, a pool which maintains a single
+connection must be used.   Use :class:`.SingletonThreadPool` if the scope is
+only needed within the current thread, or :class:`.StaticPool` is scope is
+needed within multiple threads for this case::
+
+    # maintain the same connection per thread
+    from sqlalchemy.pool import SingletonThreadPool
+
+    engine = create_engine("sqlite:///mydb.db", poolclass=SingletonThreadPool)
+
+
+    # maintain the same connection across all threads
+    from sqlalchemy.pool import StaticPool
+
+    engine = create_engine("sqlite:///mydb.db", poolclass=StaticPool)
+
+Note that :class:`.SingletonThreadPool` should be configured for the number
+of threads that are to be used; beyond that number, connections will be
+closed out in a non deterministic way.
+
+
+Dealing with Mixed String / Binary Columns
+------------------------------------------------------
+
+The SQLite database is weakly typed, and as such it is possible when using
+binary values, which in Python are represented as ``b'some string'``, that a
+particular SQLite database can have data values within different rows where
+some of them will be returned as a ``b''`` value by the Pysqlite driver, and
+others will be returned as Python strings, e.g. ``''`` values.   This situation
+is not known to occur if the SQLAlchemy :class:`.LargeBinary` datatype is used
+consistently, however if a particular SQLite database has data that was
+inserted using the Pysqlite driver directly, or when using the SQLAlchemy
+:class:`.String` type which was later changed to :class:`.LargeBinary`, the
+table will not be consistently readable because SQLAlchemy's
+:class:`.LargeBinary` datatype does not handle strings so it has no way of
+"encoding" a value that is in string format.
+
+To deal with a SQLite table that has mixed string / binary data in the
+same column, use a custom type that will check each row individually::
+
+    from sqlalchemy import String
+    from sqlalchemy import TypeDecorator
+
+
+    class MixedBinary(TypeDecorator):
+        impl = String
+        cache_ok = True
+
+        def process_result_value(self, value, dialect):
+            if isinstance(value, str):
+                value = bytes(value, "utf-8")
+            elif value is not None:
+                value = bytes(value)
+
+            return value
+
+Then use the above ``MixedBinary`` datatype in the place where
+:class:`.LargeBinary` would normally be used.
+
+.. _pysqlite_serializable:
+
+Serializable isolation / Savepoints / Transactional DDL
+-------------------------------------------------------
+
+In the section :ref:`sqlite_concurrency`, we refer to the pysqlite
+driver's assortment of issues that prevent several features of SQLite
+from working correctly.  The pysqlite DBAPI driver has several
+long-standing bugs which impact the correctness of its transactional
+behavior.   In its default mode of operation, SQLite features such as
+SERIALIZABLE isolation, transactional DDL, and SAVEPOINT support are
+non-functional, and in order to use these features, workarounds must
+be taken.
+
+The issue is essentially that the driver attempts to second-guess the user's
+intent, failing to start transactions and sometimes ending them prematurely, in
+an effort to minimize the SQLite databases's file locking behavior, even
+though SQLite itself uses "shared" locks for read-only activities.
+
+SQLAlchemy chooses to not alter this behavior by default, as it is the
+long-expected behavior of the pysqlite driver; if and when the pysqlite
+driver attempts to repair these issues, that will be more of a driver towards
+defaults for SQLAlchemy.
+
+The good news is that with a few events, we can implement transactional
+support fully, by disabling pysqlite's feature entirely and emitting BEGIN
+ourselves. This is achieved using two event listeners::
+
+    from sqlalchemy import create_engine, event
+
+    engine = create_engine("sqlite:///myfile.db")
+
+
+    @event.listens_for(engine, "connect")
+    def do_connect(dbapi_connection, connection_record):
+        # disable pysqlite's emitting of the BEGIN statement entirely.
+        # also stops it from emitting COMMIT before any DDL.
+        dbapi_connection.isolation_level = None
+
+
+    @event.listens_for(engine, "begin")
+    def do_begin(conn):
+        # emit our own BEGIN
+        conn.exec_driver_sql("BEGIN")
+
+.. warning:: When using the above recipe, it is advised to not use the
+   :paramref:`.Connection.execution_options.isolation_level` setting on
+   :class:`_engine.Connection` and :func:`_sa.create_engine`
+   with the SQLite driver,
+   as this function necessarily will also alter the ".isolation_level" setting.
+
+
+Above, we intercept a new pysqlite connection and disable any transactional
+integration.   Then, at the point at which SQLAlchemy knows that transaction
+scope is to begin, we emit ``"BEGIN"`` ourselves.
+
+When we take control of ``"BEGIN"``, we can also control directly SQLite's
+locking modes, introduced at
+`BEGIN TRANSACTION <https://sqlite.org/lang_transaction.html>`_,
+by adding the desired locking mode to our ``"BEGIN"``::
+
+    @event.listens_for(engine, "begin")
+    def do_begin(conn):
+        conn.exec_driver_sql("BEGIN EXCLUSIVE")
+
+.. seealso::
+
+    `BEGIN TRANSACTION <https://sqlite.org/lang_transaction.html>`_ -
+    on the SQLite site
+
+    `sqlite3 SELECT does not BEGIN a transaction <https://bugs.python.org/issue9924>`_ -
+    on the Python bug tracker
+
+    `sqlite3 module breaks transactions and potentially corrupts data <https://bugs.python.org/issue10740>`_ -
+    on the Python bug tracker
+
+.. _pysqlite_udfs:
+
+User-Defined Functions
+----------------------
+
+pysqlite supports a `create_function() <https://docs.python.org/3/library/sqlite3.html#sqlite3.Connection.create_function>`_
+method that allows us to create our own user-defined functions (UDFs) in Python and use them directly in SQLite queries.
+These functions are registered with a specific DBAPI Connection.
+
+SQLAlchemy uses connection pooling with file-based SQLite databases, so we need to ensure that the UDF is attached to the
+connection when it is created. That is accomplished with an event listener::
+
+    from sqlalchemy import create_engine
+    from sqlalchemy import event
+    from sqlalchemy import text
+
+
+    def udf():
+        return "udf-ok"
+
+
+    engine = create_engine("sqlite:///./db_file")
+
+
+    @event.listens_for(engine, "connect")
+    def connect(conn, rec):
+        conn.create_function("udf", 0, udf)
+
+
+    for i in range(5):
+        with engine.connect() as conn:
+            print(conn.scalar(text("SELECT UDF()")))
+
+"""  # noqa
+
+import math
+import os
+import re
+
+from .base import DATE
+from .base import DATETIME
+from .base import SQLiteDialect
+from ... import exc
+from ... import pool
+from ... import types as sqltypes
+from ... import util
+
+
+class _SQLite_pysqliteTimeStamp(DATETIME):
+    def bind_processor(self, dialect):
+        if dialect.native_datetime:
+            return None
+        else:
+            return DATETIME.bind_processor(self, dialect)
+
+    def result_processor(self, dialect, coltype):
+        if dialect.native_datetime:
+            return None
+        else:
+            return DATETIME.result_processor(self, dialect, coltype)
+
+
+class _SQLite_pysqliteDate(DATE):
+    def bind_processor(self, dialect):
+        if dialect.native_datetime:
+            return None
+        else:
+            return DATE.bind_processor(self, dialect)
+
+    def result_processor(self, dialect, coltype):
+        if dialect.native_datetime:
+            return None
+        else:
+            return DATE.result_processor(self, dialect, coltype)
+
+
+class SQLiteDialect_pysqlite(SQLiteDialect):
+    default_paramstyle = "qmark"
+    supports_statement_cache = True
+    returns_native_bytes = True
+
+    colspecs = util.update_copy(
+        SQLiteDialect.colspecs,
+        {
+            sqltypes.Date: _SQLite_pysqliteDate,
+            sqltypes.TIMESTAMP: _SQLite_pysqliteTimeStamp,
+        },
+    )
+
+    description_encoding = None
+
+    driver = "pysqlite"
+
+    @classmethod
+    def import_dbapi(cls):
+        from sqlite3 import dbapi2 as sqlite
+
+        return sqlite
+
+    @classmethod
+    def _is_url_file_db(cls, url):
+        if (url.database and url.database != ":memory:") and (
+            url.query.get("mode", None) != "memory"
+        ):
+            return True
+        else:
+            return False
+
+    @classmethod
+    def get_pool_class(cls, url):
+        if cls._is_url_file_db(url):
+            return pool.QueuePool
+        else:
+            return pool.SingletonThreadPool
+
+    def _get_server_version_info(self, connection):
+        return self.dbapi.sqlite_version_info
+
+    _isolation_lookup = SQLiteDialect._isolation_lookup.union(
+        {
+            "AUTOCOMMIT": None,
+        }
+    )
+
+    def set_isolation_level(self, dbapi_connection, level):
+        if level == "AUTOCOMMIT":
+            dbapi_connection.isolation_level = None
+        else:
+            dbapi_connection.isolation_level = ""
+            return super().set_isolation_level(dbapi_connection, level)
+
+    def on_connect(self):
+        def regexp(a, b):
+            if b is None:
+                return None
+            return re.search(a, b) is not None
+
+        if util.py38 and self._get_server_version_info(None) >= (3, 9):
+            # sqlite must be greater than 3.8.3 for deterministic=True
+            # https://docs.python.org/3/library/sqlite3.html#sqlite3.Connection.create_function
+            # the check is more conservative since there were still issues
+            # with following 3.8 sqlite versions
+            create_func_kw = {"deterministic": True}
+        else:
+            create_func_kw = {}
+
+        def set_regexp(dbapi_connection):
+            dbapi_connection.create_function(
+                "regexp", 2, regexp, **create_func_kw
+            )
+
+        def floor_func(dbapi_connection):
+            # NOTE: floor is optionally present in sqlite 3.35+ , however
+            # as it is normally non-present we deliver floor() unconditionally
+            # for now.
+            # https://www.sqlite.org/lang_mathfunc.html
+            dbapi_connection.create_function(
+                "floor", 1, math.floor, **create_func_kw
+            )
+
+        fns = [set_regexp, floor_func]
+
+        def connect(conn):
+            for fn in fns:
+                fn(conn)
+
+        return connect
+
+    def create_connect_args(self, url):
+        if url.username or url.password or url.host or url.port:
+            raise exc.ArgumentError(
+                "Invalid SQLite URL: %s\n"
+                "Valid SQLite URL forms are:\n"
+                " sqlite:///:memory: (or, sqlite://)\n"
+                " sqlite:///relative/path/to/file.db\n"
+                " sqlite:////absolute/path/to/file.db" % (url,)
+            )
+
+        # theoretically, this list can be augmented, at least as far as
+        # parameter names accepted by sqlite3/pysqlite, using
+        # inspect.getfullargspec().  for the moment this seems like overkill
+        # as these parameters don't change very often, and as always,
+        # parameters passed to connect_args will always go to the
+        # sqlite3/pysqlite driver.
+        pysqlite_args = [
+            ("uri", bool),
+            ("timeout", float),
+            ("isolation_level", str),
+            ("detect_types", int),
+            ("check_same_thread", bool),
+            ("cached_statements", int),
+        ]
+        opts = url.query
+        pysqlite_opts = {}
+        for key, type_ in pysqlite_args:
+            util.coerce_kw_type(opts, key, type_, dest=pysqlite_opts)
+
+        if pysqlite_opts.get("uri", False):
+            uri_opts = dict(opts)
+            # here, we are actually separating the parameters that go to
+            # sqlite3/pysqlite vs. those that go the SQLite URI.  What if
+            # two names conflict?  again, this seems to be not the case right
+            # now, and in the case that new names are added to
+            # either side which overlap, again the sqlite3/pysqlite parameters
+            # can be passed through connect_args instead of in the URL.
+            # If SQLite native URIs add a parameter like "timeout" that
+            # we already have listed here for the python driver, then we need
+            # to adjust for that here.
+            for key, type_ in pysqlite_args:
+                uri_opts.pop(key, None)
+            filename = url.database
+            if uri_opts:
+                # sorting of keys is for unit test support
+                filename += "?" + (
+                    "&".join(
+                        "%s=%s" % (key, uri_opts[key])
+                        for key in sorted(uri_opts)
+                    )
+                )
+        else:
+            filename = url.database or ":memory:"
+            if filename != ":memory:":
+                filename = os.path.abspath(filename)
+
+        pysqlite_opts.setdefault(
+            "check_same_thread", not self._is_url_file_db(url)
+        )
+
+        return ([filename], pysqlite_opts)
+
+    def is_disconnect(self, e, connection, cursor):
+        return isinstance(
+            e, self.dbapi.ProgrammingError
+        ) and "Cannot operate on a closed database." in str(e)
+
+
+dialect = SQLiteDialect_pysqlite
+
+
+class _SQLiteDialect_pysqlite_numeric(SQLiteDialect_pysqlite):
+    """numeric dialect for testing only
+
+    internal use only.  This dialect is **NOT** supported by SQLAlchemy
+    and may change at any time.
+
+    """
+
+    supports_statement_cache = True
+    default_paramstyle = "numeric"
+    driver = "pysqlite_numeric"
+
+    _first_bind = ":1"
+    _not_in_statement_regexp = None
+
+    def __init__(self, *arg, **kw):
+        kw.setdefault("paramstyle", "numeric")
+        super().__init__(*arg, **kw)
+
+    def create_connect_args(self, url):
+        arg, opts = super().create_connect_args(url)
+        opts["factory"] = self._fix_sqlite_issue_99953()
+        return arg, opts
+
+    def _fix_sqlite_issue_99953(self):
+        import sqlite3
+
+        first_bind = self._first_bind
+        if self._not_in_statement_regexp:
+            nis = self._not_in_statement_regexp
+
+            def _test_sql(sql):
+                m = nis.search(sql)
+                assert not m, f"Found {nis.pattern!r} in {sql!r}"
+
+        else:
+
+            def _test_sql(sql):
+                pass
+
+        def _numeric_param_as_dict(parameters):
+            if parameters:
+                assert isinstance(parameters, tuple)
+                return {
+                    str(idx): value for idx, value in enumerate(parameters, 1)
+                }
+            else:
+                return ()
+
+        class SQLiteFix99953Cursor(sqlite3.Cursor):
+            def execute(self, sql, parameters=()):
+                _test_sql(sql)
+                if first_bind in sql:
+                    parameters = _numeric_param_as_dict(parameters)
+                return super().execute(sql, parameters)
+
+            def executemany(self, sql, parameters):
+                _test_sql(sql)
+                if first_bind in sql:
+                    parameters = [
+                        _numeric_param_as_dict(p) for p in parameters
+                    ]
+                return super().executemany(sql, parameters)
+
+        class SQLiteFix99953Connection(sqlite3.Connection):
+            def cursor(self, factory=None):
+                if factory is None:
+                    factory = SQLiteFix99953Cursor
+                return super().cursor(factory=factory)
+
+            def execute(self, sql, parameters=()):
+                _test_sql(sql)
+                if first_bind in sql:
+                    parameters = _numeric_param_as_dict(parameters)
+                return super().execute(sql, parameters)
+
+            def executemany(self, sql, parameters):
+                _test_sql(sql)
+                if first_bind in sql:
+                    parameters = [
+                        _numeric_param_as_dict(p) for p in parameters
+                    ]
+                return super().executemany(sql, parameters)
+
+        return SQLiteFix99953Connection
+
+
+class _SQLiteDialect_pysqlite_dollar(_SQLiteDialect_pysqlite_numeric):
+    """numeric dialect that uses $ for testing only
+
+    internal use only.  This dialect is **NOT** supported by SQLAlchemy
+    and may change at any time.
+
+    """
+
+    supports_statement_cache = True
+    default_paramstyle = "numeric_dollar"
+    driver = "pysqlite_dollar"
+
+    _first_bind = "$1"
+    _not_in_statement_regexp = re.compile(r"[^\d]:\d+")
+
+    def __init__(self, *arg, **kw):
+        kw.setdefault("paramstyle", "numeric_dollar")
+        super().__init__(*arg, **kw)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/type_migration_guidelines.txt b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/type_migration_guidelines.txt
new file mode 100644
index 00000000..e6be2056
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/dialects/type_migration_guidelines.txt
@@ -0,0 +1,145 @@
+Rules for Migrating TypeEngine classes to 0.6
+---------------------------------------------
+
+1. the TypeEngine classes are used for:
+
+    a. Specifying behavior which needs to occur for bind parameters
+    or result row columns.
+
+    b. Specifying types that are entirely specific to the database
+    in use and have no analogue in the sqlalchemy.types package.
+
+    c. Specifying types where there is an analogue in sqlalchemy.types,
+    but the database in use takes vendor-specific flags for those
+    types.
+
+    d. If a TypeEngine class doesn't provide any of this, it should be
+    *removed* from the dialect.
+
+2. the TypeEngine classes are *no longer* used for generating DDL.  Dialects
+now have a TypeCompiler subclass which uses the same visit_XXX model as
+other compilers.
+
+3. the "ischema_names" and "colspecs" dictionaries are now required members on
+the Dialect class.
+
+4. The names of types within dialects are now important.   If a dialect-specific type
+is a subclass of an existing generic type and is only provided for bind/result behavior, 
+the current mixed case naming can remain, i.e. _PGNumeric for Numeric - in this case, 
+end users would never need to use _PGNumeric directly.   However, if a dialect-specific 
+type is specifying a type *or* arguments that are not present generically, it should
+match the real name of the type on that backend, in uppercase.  E.g. postgresql.INET,
+mysql.ENUM, postgresql.ARRAY.
+
+Or follow this handy flowchart:
+
+    is the type meant to provide bind/result                  is the type the same name as an
+    behavior to a generic type (i.e. MixedCase)  ---- no ---> UPPERCASE type in types.py ?
+    type in types.py ?                                          |                     |
+                    |                                           no                    yes
+                   yes                                          |                     |
+                    |                                           |             does your type need special
+                    |                                           +<--- yes --- behavior or arguments ?
+                    |                                           |                               |
+                    |                                           |                              no
+           name the type using                                  |                               |
+           _MixedCase, i.e.                                     v                               V
+           _OracleBoolean. it                          name the type                        don't make a
+           stays private to the dialect                identically as that                  type, make sure the dialect's
+           and is invoked *only* via                   within the DB,                       base.py imports the types.py
+           the colspecs dict.                          using UPPERCASE                      UPPERCASE name into its namespace
+                    |                                  (i.e. BIT, NCHAR, INTERVAL).
+                    |                                  Users can import it.
+                    |                                       |
+                    v                                       v
+           subclass the closest                        is the name of this type
+           MixedCase type types.py,                    identical to an UPPERCASE
+           i.e.                        <--- no ------- name in types.py ?
+           class _DateTime(types.DateTime),
+           class DATETIME2(types.DateTime),                   |
+           class BIT(types.TypeEngine).                      yes
+                                                              |
+                                                              v
+                                                        the type should
+                                                        subclass the
+                                                        UPPERCASE
+                                                        type in types.py
+                                                        (i.e. class BLOB(types.BLOB))
+
+
+Example 1.   pysqlite needs bind/result processing for the DateTime type in types.py, 
+which applies to all DateTimes and subclasses.   It's named _SLDateTime and 
+subclasses types.DateTime.
+
+Example 2.  MS-SQL has a TIME type which takes a non-standard "precision" argument
+that is rendered within DDL.   So it's named TIME in the MS-SQL dialect's base.py, 
+and subclasses types.TIME.  Users can then say mssql.TIME(precision=10).
+
+Example 3.  MS-SQL dialects also need special bind/result processing for date 
+But its DATE type doesn't render DDL differently than that of a plain 
+DATE, i.e. it takes no special arguments.  Therefore we are just adding behavior
+to types.Date, so it's named _MSDate in the MS-SQL dialect's base.py, and subclasses
+types.Date.
+
+Example 4.  MySQL has a SET type, there's no analogue for this in types.py. So
+MySQL names it SET in the dialect's base.py, and it subclasses types.String, since 
+it ultimately deals with strings.
+
+Example 5.  PostgreSQL has a DATETIME type.  The DBAPIs handle dates correctly,
+and no special arguments are used in PG's DDL beyond what types.py provides.
+PostgreSQL dialect therefore imports types.DATETIME into its base.py.
+
+Ideally one should be able to specify a schema using names imported completely from a 
+dialect, all matching the real name on that backend:
+
+   from sqlalchemy.dialects.postgresql import base as pg
+
+   t = Table('mytable', metadata,
+              Column('id', pg.INTEGER, primary_key=True),
+              Column('name', pg.VARCHAR(300)),
+              Column('inetaddr', pg.INET)
+   )
+
+where above, the INTEGER and VARCHAR types are ultimately from sqlalchemy.types, 
+but the PG dialect makes them available in its own namespace.
+
+5. "colspecs" now is a dictionary of generic or uppercased types from sqlalchemy.types
+linked to types specified in the dialect.   Again, if a type in the dialect does not
+specify any special behavior for bind_processor() or result_processor() and does not
+indicate a special type only available in this database, it must be *removed* from the 
+module and from this dictionary.
+
+6. "ischema_names" indicates string descriptions of types as returned from the database
+linked to TypeEngine classes.
+
+    a. The string name should be matched to the most specific type possible within
+    sqlalchemy.types, unless there is no matching type within sqlalchemy.types in which
+    case it points to a dialect type.   *It doesn't matter* if the dialect has its 
+    own subclass of that type with special bind/result behavior - reflect to the types.py
+    UPPERCASE type as much as possible.   With very few exceptions, all types
+    should reflect to an UPPERCASE type.
+
+    b. If the dialect contains a matching dialect-specific type that takes extra arguments 
+    which the generic one does not, then point to the dialect-specific type.  E.g.
+    mssql.VARCHAR takes a "collation" parameter which should be preserved.
+
+5. DDL, or what was formerly issued by "get_col_spec()", is now handled exclusively by
+a subclass of compiler.GenericTypeCompiler.
+
+    a. your TypeCompiler class will receive generic and uppercase types from 
+    sqlalchemy.types.  Do not assume the presence of dialect-specific attributes on
+    these types. 
+
+    b. the visit_UPPERCASE methods on GenericTypeCompiler should *not* be overridden with
+    methods that produce a different DDL name.   Uppercase types don't do any kind of 
+    "guessing" - if visit_TIMESTAMP is called, the DDL should render as TIMESTAMP in
+    all cases, regardless of whether or not that type is legal on the backend database.
+
+    c. the visit_UPPERCASE methods *should* be overridden with methods that add additional
+    arguments and flags to those types.
+
+    d. the visit_lowercase methods are overridden to provide an interpretation of a generic 
+    type.  E.g.  visit_large_binary() might be overridden to say "return self.visit_BIT(type_)".
+
+    e. visit_lowercase methods should *never* render strings directly - it should always
+    be via calling a visit_UPPERCASE() method.
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__init__.py
new file mode 100644
index 00000000..f4205d89
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__init__.py
@@ -0,0 +1,62 @@
+# engine/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""SQL connections, SQL execution and high-level DB-API interface.
+
+The engine package defines the basic components used to interface
+DB-API modules with higher-level statement construction,
+connection-management, execution and result contexts.  The primary
+"entry point" class into this package is the Engine and its public
+constructor ``create_engine()``.
+
+"""
+
+from . import events as events
+from . import util as util
+from .base import Connection as Connection
+from .base import Engine as Engine
+from .base import NestedTransaction as NestedTransaction
+from .base import RootTransaction as RootTransaction
+from .base import Transaction as Transaction
+from .base import TwoPhaseTransaction as TwoPhaseTransaction
+from .create import create_engine as create_engine
+from .create import create_pool_from_url as create_pool_from_url
+from .create import engine_from_config as engine_from_config
+from .cursor import CursorResult as CursorResult
+from .cursor import ResultProxy as ResultProxy
+from .interfaces import AdaptedConnection as AdaptedConnection
+from .interfaces import BindTyping as BindTyping
+from .interfaces import Compiled as Compiled
+from .interfaces import Connectable as Connectable
+from .interfaces import ConnectArgsType as ConnectArgsType
+from .interfaces import ConnectionEventsTarget as ConnectionEventsTarget
+from .interfaces import CreateEnginePlugin as CreateEnginePlugin
+from .interfaces import Dialect as Dialect
+from .interfaces import ExceptionContext as ExceptionContext
+from .interfaces import ExecutionContext as ExecutionContext
+from .interfaces import TypeCompiler as TypeCompiler
+from .mock import create_mock_engine as create_mock_engine
+from .reflection import Inspector as Inspector
+from .reflection import ObjectKind as ObjectKind
+from .reflection import ObjectScope as ObjectScope
+from .result import ChunkedIteratorResult as ChunkedIteratorResult
+from .result import FilterResult as FilterResult
+from .result import FrozenResult as FrozenResult
+from .result import IteratorResult as IteratorResult
+from .result import MappingResult as MappingResult
+from .result import MergedResult as MergedResult
+from .result import Result as Result
+from .result import result_tuple as result_tuple
+from .result import ScalarResult as ScalarResult
+from .result import TupleResult as TupleResult
+from .row import BaseRow as BaseRow
+from .row import Row as Row
+from .row import RowMapping as RowMapping
+from .url import make_url as make_url
+from .url import URL as URL
+from .util import connection_memoize as connection_memoize
+from ..sql import ddl as ddl
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..47ace703
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_processors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_processors.cpython-312.pyc
new file mode 100644
index 00000000..32a4a4ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_processors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_row.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_row.cpython-312.pyc
new file mode 100644
index 00000000..afb6f3bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_row.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_util.cpython-312.pyc
new file mode 100644
index 00000000..5a7178ad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/_py_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..ea19b57e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/characteristics.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/characteristics.cpython-312.pyc
new file mode 100644
index 00000000..46822712
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/characteristics.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/create.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/create.cpython-312.pyc
new file mode 100644
index 00000000..dce8213a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/create.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/cursor.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/cursor.cpython-312.pyc
new file mode 100644
index 00000000..44da1a67
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/cursor.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/default.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/default.cpython-312.pyc
new file mode 100644
index 00000000..0d3cc1bf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/default.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/events.cpython-312.pyc
new file mode 100644
index 00000000..a3b9e155
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/interfaces.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/interfaces.cpython-312.pyc
new file mode 100644
index 00000000..48cc1b53
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/interfaces.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/mock.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/mock.cpython-312.pyc
new file mode 100644
index 00000000..ea6d707b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/mock.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/processors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/processors.cpython-312.pyc
new file mode 100644
index 00000000..93167e1b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/processors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/reflection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/reflection.cpython-312.pyc
new file mode 100644
index 00000000..86ceb768
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/reflection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/result.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/result.cpython-312.pyc
new file mode 100644
index 00000000..c9be7a8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/result.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/row.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/row.cpython-312.pyc
new file mode 100644
index 00000000..32cf1cfb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/row.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/strategies.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/strategies.cpython-312.pyc
new file mode 100644
index 00000000..a24fb763
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/strategies.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/url.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/url.cpython-312.pyc
new file mode 100644
index 00000000..82c966a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/url.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..502310aa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_processors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_processors.py
new file mode 100644
index 00000000..8536d53d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_processors.py
@@ -0,0 +1,136 @@
+# engine/_py_processors.py
+# Copyright (C) 2010-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+# Copyright (C) 2010 Gaetan de Menten gdementen@gmail.com
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""defines generic type conversion functions, as used in bind and result
+processors.
+
+They all share one common characteristic: None is passed through unchanged.
+
+"""
+
+from __future__ import annotations
+
+import datetime
+from datetime import date as date_cls
+from datetime import datetime as datetime_cls
+from datetime import time as time_cls
+from decimal import Decimal
+import typing
+from typing import Any
+from typing import Callable
+from typing import Optional
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+
+_DT = TypeVar(
+    "_DT", bound=Union[datetime.datetime, datetime.time, datetime.date]
+)
+
+
+def str_to_datetime_processor_factory(
+    regexp: typing.Pattern[str], type_: Callable[..., _DT]
+) -> Callable[[Optional[str]], Optional[_DT]]:
+    rmatch = regexp.match
+    # Even on python2.6 datetime.strptime is both slower than this code
+    # and it does not support microseconds.
+    has_named_groups = bool(regexp.groupindex)
+
+    def process(value: Optional[str]) -> Optional[_DT]:
+        if value is None:
+            return None
+        else:
+            try:
+                m = rmatch(value)
+            except TypeError as err:
+                raise ValueError(
+                    "Couldn't parse %s string '%r' "
+                    "- value is not a string." % (type_.__name__, value)
+                ) from err
+
+            if m is None:
+                raise ValueError(
+                    "Couldn't parse %s string: "
+                    "'%s'" % (type_.__name__, value)
+                )
+            if has_named_groups:
+                groups = m.groupdict(0)
+                return type_(
+                    **dict(
+                        list(
+                            zip(
+                                iter(groups.keys()),
+                                list(map(int, iter(groups.values()))),
+                            )
+                        )
+                    )
+                )
+            else:
+                return type_(*list(map(int, m.groups(0))))
+
+    return process
+
+
+def to_decimal_processor_factory(
+    target_class: Type[Decimal], scale: int
+) -> Callable[[Optional[float]], Optional[Decimal]]:
+    fstring = "%%.%df" % scale
+
+    def process(value: Optional[float]) -> Optional[Decimal]:
+        if value is None:
+            return None
+        else:
+            return target_class(fstring % value)
+
+    return process
+
+
+def to_float(value: Optional[Union[int, float]]) -> Optional[float]:
+    if value is None:
+        return None
+    else:
+        return float(value)
+
+
+def to_str(value: Optional[Any]) -> Optional[str]:
+    if value is None:
+        return None
+    else:
+        return str(value)
+
+
+def int_to_boolean(value: Optional[int]) -> Optional[bool]:
+    if value is None:
+        return None
+    else:
+        return bool(value)
+
+
+def str_to_datetime(value: Optional[str]) -> Optional[datetime.datetime]:
+    if value is not None:
+        dt_value = datetime_cls.fromisoformat(value)
+    else:
+        dt_value = None
+    return dt_value
+
+
+def str_to_time(value: Optional[str]) -> Optional[datetime.time]:
+    if value is not None:
+        dt_value = time_cls.fromisoformat(value)
+    else:
+        dt_value = None
+    return dt_value
+
+
+def str_to_date(value: Optional[str]) -> Optional[datetime.date]:
+    if value is not None:
+        dt_value = date_cls.fromisoformat(value)
+    else:
+        dt_value = None
+    return dt_value
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_row.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_row.py
new file mode 100644
index 00000000..38c60fcd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_row.py
@@ -0,0 +1,128 @@
+# engine/_py_row.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import Tuple
+from typing import Type
+
+if typing.TYPE_CHECKING:
+    from .result import _KeyType
+    from .result import _ProcessorsType
+    from .result import _RawRowType
+    from .result import _TupleGetterType
+    from .result import ResultMetaData
+
+MD_INDEX = 0  # integer index in cursor.description
+
+
+class BaseRow:
+    __slots__ = ("_parent", "_data", "_key_to_index")
+
+    _parent: ResultMetaData
+    _key_to_index: Mapping[_KeyType, int]
+    _data: _RawRowType
+
+    def __init__(
+        self,
+        parent: ResultMetaData,
+        processors: Optional[_ProcessorsType],
+        key_to_index: Mapping[_KeyType, int],
+        data: _RawRowType,
+    ):
+        """Row objects are constructed by CursorResult objects."""
+        object.__setattr__(self, "_parent", parent)
+
+        object.__setattr__(self, "_key_to_index", key_to_index)
+
+        if processors:
+            object.__setattr__(
+                self,
+                "_data",
+                tuple(
+                    [
+                        proc(value) if proc else value
+                        for proc, value in zip(processors, data)
+                    ]
+                ),
+            )
+        else:
+            object.__setattr__(self, "_data", tuple(data))
+
+    def __reduce__(self) -> Tuple[Callable[..., BaseRow], Tuple[Any, ...]]:
+        return (
+            rowproxy_reconstructor,
+            (self.__class__, self.__getstate__()),
+        )
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {"_parent": self._parent, "_data": self._data}
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        parent = state["_parent"]
+        object.__setattr__(self, "_parent", parent)
+        object.__setattr__(self, "_data", state["_data"])
+        object.__setattr__(self, "_key_to_index", parent._key_to_index)
+
+    def _values_impl(self) -> List[Any]:
+        return list(self)
+
+    def __iter__(self) -> Iterator[Any]:
+        return iter(self._data)
+
+    def __len__(self) -> int:
+        return len(self._data)
+
+    def __hash__(self) -> int:
+        return hash(self._data)
+
+    def __getitem__(self, key: Any) -> Any:
+        return self._data[key]
+
+    def _get_by_key_impl_mapping(self, key: str) -> Any:
+        try:
+            return self._data[self._key_to_index[key]]
+        except KeyError:
+            pass
+        self._parent._key_not_found(key, False)
+
+    def __getattr__(self, name: str) -> Any:
+        try:
+            return self._data[self._key_to_index[name]]
+        except KeyError:
+            pass
+        self._parent._key_not_found(name, True)
+
+    def _to_tuple_instance(self) -> Tuple[Any, ...]:
+        return self._data
+
+
+# This reconstructor is necessary so that pickles with the Cy extension or
+# without use the same Binary format.
+def rowproxy_reconstructor(
+    cls: Type[BaseRow], state: Dict[str, Any]
+) -> BaseRow:
+    obj = cls.__new__(cls)
+    obj.__setstate__(state)
+    return obj
+
+
+def tuplegetter(*indexes: int) -> _TupleGetterType:
+    if len(indexes) != 1:
+        for i in range(1, len(indexes)):
+            if indexes[i - 1] != indexes[i] - 1:
+                return operator.itemgetter(*indexes)
+    # slice form is faster but returns a list if input is list
+    return operator.itemgetter(slice(indexes[0], indexes[-1] + 1))
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_util.py
new file mode 100644
index 00000000..50badea2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/_py_util.py
@@ -0,0 +1,74 @@
+# engine/_py_util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Mapping
+from typing import Optional
+from typing import Tuple
+
+from .. import exc
+
+if typing.TYPE_CHECKING:
+    from .interfaces import _CoreAnyExecuteParams
+    from .interfaces import _CoreMultiExecuteParams
+    from .interfaces import _DBAPIAnyExecuteParams
+    from .interfaces import _DBAPIMultiExecuteParams
+
+
+_no_tuple: Tuple[Any, ...] = ()
+
+
+def _distill_params_20(
+    params: Optional[_CoreAnyExecuteParams],
+) -> _CoreMultiExecuteParams:
+    if params is None:
+        return _no_tuple
+    # Assume list is more likely than tuple
+    elif isinstance(params, list) or isinstance(params, tuple):
+        # collections_abc.MutableSequence): # avoid abc.__instancecheck__
+        if params and not isinstance(params[0], (tuple, Mapping)):
+            raise exc.ArgumentError(
+                "List argument must consist only of tuples or dictionaries"
+            )
+
+        return params
+    elif isinstance(params, dict) or isinstance(
+        # only do immutabledict or abc.__instancecheck__ for Mapping after
+        # we've checked for plain dictionaries and would otherwise raise
+        params,
+        Mapping,
+    ):
+        return [params]
+    else:
+        raise exc.ArgumentError("mapping or list expected for parameters")
+
+
+def _distill_raw_params(
+    params: Optional[_DBAPIAnyExecuteParams],
+) -> _DBAPIMultiExecuteParams:
+    if params is None:
+        return _no_tuple
+    elif isinstance(params, list):
+        # collections_abc.MutableSequence): # avoid abc.__instancecheck__
+        if params and not isinstance(params[0], (tuple, Mapping)):
+            raise exc.ArgumentError(
+                "List argument must consist only of tuples or dictionaries"
+            )
+
+        return params
+    elif isinstance(params, (tuple, dict)) or isinstance(
+        # only do abc.__instancecheck__ for Mapping after we've checked
+        # for plain dictionaries and would otherwise raise
+        params,
+        Mapping,
+    ):
+        # cast("Union[List[Mapping[str, Any]], Tuple[Any, ...]]", [params])
+        return [params]  # type: ignore
+    else:
+        raise exc.ArgumentError("mapping or sequence expected for parameters")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/base.py
new file mode 100644
index 00000000..cbf11acf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/base.py
@@ -0,0 +1,3371 @@
+# engine/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+"""Defines :class:`_engine.Connection` and :class:`_engine.Engine`.
+
+"""
+from __future__ import annotations
+
+import contextlib
+import sys
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+from .interfaces import BindTyping
+from .interfaces import ConnectionEventsTarget
+from .interfaces import DBAPICursor
+from .interfaces import ExceptionContext
+from .interfaces import ExecuteStyle
+from .interfaces import ExecutionContext
+from .interfaces import IsolationLevel
+from .util import _distill_params_20
+from .util import _distill_raw_params
+from .util import TransactionalContext
+from .. import exc
+from .. import inspection
+from .. import log
+from .. import util
+from ..sql import compiler
+from ..sql import util as sql_util
+
+if typing.TYPE_CHECKING:
+    from . import CursorResult
+    from . import ScalarResult
+    from .interfaces import _AnyExecuteParams
+    from .interfaces import _AnyMultiExecuteParams
+    from .interfaces import _CoreAnyExecuteParams
+    from .interfaces import _CoreMultiExecuteParams
+    from .interfaces import _CoreSingleExecuteParams
+    from .interfaces import _DBAPIAnyExecuteParams
+    from .interfaces import _DBAPISingleExecuteParams
+    from .interfaces import _ExecuteOptions
+    from .interfaces import CompiledCacheType
+    from .interfaces import CoreExecuteOptionsParameter
+    from .interfaces import Dialect
+    from .interfaces import SchemaTranslateMapType
+    from .reflection import Inspector  # noqa
+    from .url import URL
+    from ..event import dispatcher
+    from ..log import _EchoFlagType
+    from ..pool import _ConnectionFairy
+    from ..pool import Pool
+    from ..pool import PoolProxiedConnection
+    from ..sql import Executable
+    from ..sql._typing import _InfoType
+    from ..sql.compiler import Compiled
+    from ..sql.ddl import ExecutableDDLElement
+    from ..sql.ddl import SchemaDropper
+    from ..sql.ddl import SchemaGenerator
+    from ..sql.functions import FunctionElement
+    from ..sql.schema import DefaultGenerator
+    from ..sql.schema import HasSchemaAttr
+    from ..sql.schema import SchemaItem
+    from ..sql.selectable import TypedReturnsRows
+
+
+_T = TypeVar("_T", bound=Any)
+_EMPTY_EXECUTION_OPTS: _ExecuteOptions = util.EMPTY_DICT
+NO_OPTIONS: Mapping[str, Any] = util.EMPTY_DICT
+
+
+class Connection(ConnectionEventsTarget, inspection.Inspectable["Inspector"]):
+    """Provides high-level functionality for a wrapped DB-API connection.
+
+    The :class:`_engine.Connection` object is procured by calling the
+    :meth:`_engine.Engine.connect` method of the :class:`_engine.Engine`
+    object, and provides services for execution of SQL statements as well
+    as transaction control.
+
+    The Connection object is **not** thread-safe. While a Connection can be
+    shared among threads using properly synchronized access, it is still
+    possible that the underlying DBAPI connection may not support shared
+    access between threads. Check the DBAPI documentation for details.
+
+    The Connection object represents a single DBAPI connection checked out
+    from the connection pool. In this state, the connection pool has no
+    affect upon the connection, including its expiration or timeout state.
+    For the connection pool to properly manage connections, connections
+    should be returned to the connection pool (i.e. ``connection.close()``)
+    whenever the connection is not in use.
+
+    .. index::
+      single: thread safety; Connection
+
+    """
+
+    dialect: Dialect
+    dispatch: dispatcher[ConnectionEventsTarget]
+
+    _sqla_logger_namespace = "sqlalchemy.engine.Connection"
+
+    # used by sqlalchemy.engine.util.TransactionalContext
+    _trans_context_manager: Optional[TransactionalContext] = None
+
+    # legacy as of 2.0, should be eventually deprecated and
+    # removed.  was used in the "pre_ping" recipe that's been in the docs
+    # a long time
+    should_close_with_result = False
+
+    _dbapi_connection: Optional[PoolProxiedConnection]
+
+    _execution_options: _ExecuteOptions
+
+    _transaction: Optional[RootTransaction]
+    _nested_transaction: Optional[NestedTransaction]
+
+    def __init__(
+        self,
+        engine: Engine,
+        connection: Optional[PoolProxiedConnection] = None,
+        _has_events: Optional[bool] = None,
+        _allow_revalidate: bool = True,
+        _allow_autobegin: bool = True,
+    ):
+        """Construct a new Connection."""
+        self.engine = engine
+        self.dialect = dialect = engine.dialect
+
+        if connection is None:
+            try:
+                self._dbapi_connection = engine.raw_connection()
+            except dialect.loaded_dbapi.Error as err:
+                Connection._handle_dbapi_exception_noconnection(
+                    err, dialect, engine
+                )
+                raise
+        else:
+            self._dbapi_connection = connection
+
+        self._transaction = self._nested_transaction = None
+        self.__savepoint_seq = 0
+        self.__in_begin = False
+
+        self.__can_reconnect = _allow_revalidate
+        self._allow_autobegin = _allow_autobegin
+        self._echo = self.engine._should_log_info()
+
+        if _has_events is None:
+            # if _has_events is sent explicitly as False,
+            # then don't join the dispatch of the engine; we don't
+            # want to handle any of the engine's events in that case.
+            self.dispatch = self.dispatch._join(engine.dispatch)
+        self._has_events = _has_events or (
+            _has_events is None and engine._has_events
+        )
+
+        self._execution_options = engine._execution_options
+
+        if self._has_events or self.engine._has_events:
+            self.dispatch.engine_connect(self)
+
+    # this can be assigned differently via
+    # characteristics.LoggingTokenCharacteristic
+    _message_formatter: Any = None
+
+    def _log_info(self, message: str, *arg: Any, **kw: Any) -> None:
+        fmt = self._message_formatter
+
+        if fmt:
+            message = fmt(message)
+
+        if log.STACKLEVEL:
+            kw["stacklevel"] = 1 + log.STACKLEVEL_OFFSET
+
+        self.engine.logger.info(message, *arg, **kw)
+
+    def _log_debug(self, message: str, *arg: Any, **kw: Any) -> None:
+        fmt = self._message_formatter
+
+        if fmt:
+            message = fmt(message)
+
+        if log.STACKLEVEL:
+            kw["stacklevel"] = 1 + log.STACKLEVEL_OFFSET
+
+        self.engine.logger.debug(message, *arg, **kw)
+
+    @property
+    def _schema_translate_map(self) -> Optional[SchemaTranslateMapType]:
+        schema_translate_map: Optional[SchemaTranslateMapType] = (
+            self._execution_options.get("schema_translate_map", None)
+        )
+
+        return schema_translate_map
+
+    def schema_for_object(self, obj: HasSchemaAttr) -> Optional[str]:
+        """Return the schema name for the given schema item taking into
+        account current schema translate map.
+
+        """
+
+        name = obj.schema
+        schema_translate_map: Optional[SchemaTranslateMapType] = (
+            self._execution_options.get("schema_translate_map", None)
+        )
+
+        if (
+            schema_translate_map
+            and name in schema_translate_map
+            and obj._use_schema_map
+        ):
+            return schema_translate_map[name]
+        else:
+            return name
+
+    def __enter__(self) -> Connection:
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        self.close()
+
+    @overload
+    def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        no_parameters: bool = False,
+        stream_results: bool = False,
+        max_row_buffer: int = ...,
+        yield_per: int = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        preserve_rowcount: bool = False,
+        **opt: Any,
+    ) -> Connection: ...
+
+    @overload
+    def execution_options(self, **opt: Any) -> Connection: ...
+
+    def execution_options(self, **opt: Any) -> Connection:
+        r"""Set non-SQL options for the connection which take effect
+        during execution.
+
+        This method modifies this :class:`_engine.Connection` **in-place**;
+        the return value is the same :class:`_engine.Connection` object
+        upon which the method is called.   Note that this is in contrast
+        to the behavior of the ``execution_options`` methods on other
+        objects such as :meth:`_engine.Engine.execution_options` and
+        :meth:`_sql.Executable.execution_options`.  The rationale is that many
+        such execution options necessarily modify the state of the base
+        DBAPI connection in any case so there is no feasible means of
+        keeping the effect of such an option localized to a "sub" connection.
+
+        .. versionchanged:: 2.0  The :meth:`_engine.Connection.execution_options`
+           method, in contrast to other objects with this method, modifies
+           the connection in-place without creating copy of it.
+
+        As discussed elsewhere, the :meth:`_engine.Connection.execution_options`
+        method accepts any arbitrary parameters including user defined names.
+        All parameters given are consumable in a number of ways including
+        by using the :meth:`_engine.Connection.get_execution_options` method.
+        See the examples at :meth:`_sql.Executable.execution_options`
+        and :meth:`_engine.Engine.execution_options`.
+
+        The keywords that are currently recognized by SQLAlchemy itself
+        include all those listed under :meth:`.Executable.execution_options`,
+        as well as others that are specific to :class:`_engine.Connection`.
+
+        :param compiled_cache: Available on: :class:`_engine.Connection`,
+          :class:`_engine.Engine`.
+
+          A dictionary where :class:`.Compiled` objects
+          will be cached when the :class:`_engine.Connection`
+          compiles a clause
+          expression into a :class:`.Compiled` object.  This dictionary will
+          supersede the statement cache that may be configured on the
+          :class:`_engine.Engine` itself.   If set to None, caching
+          is disabled, even if the engine has a configured cache size.
+
+          Note that the ORM makes use of its own "compiled" caches for
+          some operations, including flush operations.  The caching
+          used by the ORM internally supersedes a cache dictionary
+          specified here.
+
+        :param logging_token: Available on: :class:`_engine.Connection`,
+          :class:`_engine.Engine`, :class:`_sql.Executable`.
+
+          Adds the specified string token surrounded by brackets in log
+          messages logged by the connection, i.e. the logging that's enabled
+          either via the :paramref:`_sa.create_engine.echo` flag or via the
+          ``logging.getLogger("sqlalchemy.engine")`` logger. This allows a
+          per-connection or per-sub-engine token to be available which is
+          useful for debugging concurrent connection scenarios.
+
+          .. versionadded:: 1.4.0b2
+
+          .. seealso::
+
+            :ref:`dbengine_logging_tokens` - usage example
+
+            :paramref:`_sa.create_engine.logging_name` - adds a name to the
+            name used by the Python logger object itself.
+
+        :param isolation_level: Available on: :class:`_engine.Connection`,
+          :class:`_engine.Engine`.
+
+          Set the transaction isolation level for the lifespan of this
+          :class:`_engine.Connection` object.
+          Valid values include those string
+          values accepted by the :paramref:`_sa.create_engine.isolation_level`
+          parameter passed to :func:`_sa.create_engine`.  These levels are
+          semi-database specific; see individual dialect documentation for
+          valid levels.
+
+          The isolation level option applies the isolation level by emitting
+          statements on the DBAPI connection, and **necessarily affects the
+          original Connection object overall**. The isolation level will remain
+          at the given setting until explicitly changed, or when the DBAPI
+          connection itself is :term:`released` to the connection pool, i.e. the
+          :meth:`_engine.Connection.close` method is called, at which time an
+          event handler will emit additional statements on the DBAPI connection
+          in order to revert the isolation level change.
+
+          .. note:: The ``isolation_level`` execution option may only be
+             established before the :meth:`_engine.Connection.begin` method is
+             called, as well as before any SQL statements are emitted which
+             would otherwise trigger "autobegin", or directly after a call to
+             :meth:`_engine.Connection.commit` or
+             :meth:`_engine.Connection.rollback`. A database cannot change the
+             isolation level on a transaction in progress.
+
+          .. note:: The ``isolation_level`` execution option is implicitly
+             reset if the :class:`_engine.Connection` is invalidated, e.g. via
+             the :meth:`_engine.Connection.invalidate` method, or if a
+             disconnection error occurs. The new connection produced after the
+             invalidation will **not** have the selected isolation level
+             re-applied to it automatically.
+
+          .. seealso::
+
+                :ref:`dbapi_autocommit`
+
+                :meth:`_engine.Connection.get_isolation_level`
+                - view current actual level
+
+        :param no_parameters: Available on: :class:`_engine.Connection`,
+          :class:`_sql.Executable`.
+
+          When ``True``, if the final parameter
+          list or dictionary is totally empty, will invoke the
+          statement on the cursor as ``cursor.execute(statement)``,
+          not passing the parameter collection at all.
+          Some DBAPIs such as psycopg2 and mysql-python consider
+          percent signs as significant only when parameters are
+          present; this option allows code to generate SQL
+          containing percent signs (and possibly other characters)
+          that is neutral regarding whether it's executed by the DBAPI
+          or piped into a script that's later invoked by
+          command line tools.
+
+        :param stream_results: Available on: :class:`_engine.Connection`,
+          :class:`_sql.Executable`.
+
+          Indicate to the dialect that results should be "streamed" and not
+          pre-buffered, if possible.  For backends such as PostgreSQL, MySQL
+          and MariaDB, this indicates the use of a "server side cursor" as
+          opposed to a client side cursor.  Other backends such as that of
+          Oracle Database may already use server side cursors by default.
+
+          The usage of
+          :paramref:`_engine.Connection.execution_options.stream_results` is
+          usually combined with setting a fixed number of rows to to be fetched
+          in batches, to allow for efficient iteration of database rows while
+          at the same time not loading all result rows into memory at once;
+          this can be configured on a :class:`_engine.Result` object using the
+          :meth:`_engine.Result.yield_per` method, after execution has
+          returned a new :class:`_engine.Result`.   If
+          :meth:`_engine.Result.yield_per` is not used,
+          the :paramref:`_engine.Connection.execution_options.stream_results`
+          mode of operation will instead use a dynamically sized buffer
+          which buffers sets of rows at a time, growing on each batch
+          based on a fixed growth size up until a limit which may
+          be configured using the
+          :paramref:`_engine.Connection.execution_options.max_row_buffer`
+          parameter.
+
+          When using the ORM to fetch ORM mapped objects from a result,
+          :meth:`_engine.Result.yield_per` should always be used with
+          :paramref:`_engine.Connection.execution_options.stream_results`,
+          so that the ORM does not fetch all rows into new ORM objects at once.
+
+          For typical use, the
+          :paramref:`_engine.Connection.execution_options.yield_per` execution
+          option should be preferred, which sets up both
+          :paramref:`_engine.Connection.execution_options.stream_results` and
+          :meth:`_engine.Result.yield_per` at once. This option is supported
+          both at a core level by :class:`_engine.Connection` as well as by the
+          ORM :class:`_engine.Session`; the latter is described at
+          :ref:`orm_queryguide_yield_per`.
+
+          .. seealso::
+
+            :ref:`engine_stream_results` - background on
+            :paramref:`_engine.Connection.execution_options.stream_results`
+
+            :paramref:`_engine.Connection.execution_options.max_row_buffer`
+
+            :paramref:`_engine.Connection.execution_options.yield_per`
+
+            :ref:`orm_queryguide_yield_per` - in the :ref:`queryguide_toplevel`
+            describing the ORM version of ``yield_per``
+
+        :param max_row_buffer: Available on: :class:`_engine.Connection`,
+          :class:`_sql.Executable`.  Sets a maximum
+          buffer size to use when the
+          :paramref:`_engine.Connection.execution_options.stream_results`
+          execution option is used on a backend that supports server side
+          cursors.  The default value if not specified is 1000.
+
+          .. seealso::
+
+            :paramref:`_engine.Connection.execution_options.stream_results`
+
+            :ref:`engine_stream_results`
+
+
+        :param yield_per: Available on: :class:`_engine.Connection`,
+          :class:`_sql.Executable`.  Integer value applied which will
+          set the :paramref:`_engine.Connection.execution_options.stream_results`
+          execution option and invoke :meth:`_engine.Result.yield_per`
+          automatically at once.  Allows equivalent functionality as
+          is present when using this parameter with the ORM.
+
+          .. versionadded:: 1.4.40
+
+          .. seealso::
+
+            :ref:`engine_stream_results` - background and examples
+            on using server side cursors with Core.
+
+            :ref:`orm_queryguide_yield_per` - in the :ref:`queryguide_toplevel`
+            describing the ORM version of ``yield_per``
+
+        :param insertmanyvalues_page_size: Available on: :class:`_engine.Connection`,
+            :class:`_engine.Engine`. Number of rows to format into an
+            INSERT statement when the statement uses "insertmanyvalues" mode,
+            which is a paged form of bulk insert that is used for many backends
+            when using :term:`executemany` execution typically in conjunction
+            with RETURNING. Defaults to 1000. May also be modified on a
+            per-engine basis using the
+            :paramref:`_sa.create_engine.insertmanyvalues_page_size` parameter.
+
+            .. versionadded:: 2.0
+
+            .. seealso::
+
+                :ref:`engine_insertmanyvalues`
+
+        :param schema_translate_map: Available on: :class:`_engine.Connection`,
+          :class:`_engine.Engine`, :class:`_sql.Executable`.
+
+          A dictionary mapping schema names to schema names, that will be
+          applied to the :paramref:`_schema.Table.schema` element of each
+          :class:`_schema.Table`
+          encountered when SQL or DDL expression elements
+          are compiled into strings; the resulting schema name will be
+          converted based on presence in the map of the original name.
+
+          .. seealso::
+
+            :ref:`schema_translating`
+
+        :param preserve_rowcount: Boolean; when True, the ``cursor.rowcount``
+          attribute will be unconditionally memoized within the result and
+          made available via the :attr:`.CursorResult.rowcount` attribute.
+          Normally, this attribute is only preserved for UPDATE and DELETE
+          statements.  Using this option, the DBAPIs rowcount value can
+          be accessed for other kinds of statements such as INSERT and SELECT,
+          to the degree that the DBAPI supports these statements.  See
+          :attr:`.CursorResult.rowcount` for notes regarding the behavior
+          of this attribute.
+
+          .. versionadded:: 2.0.28
+
+        .. seealso::
+
+            :meth:`_engine.Engine.execution_options`
+
+            :meth:`.Executable.execution_options`
+
+            :meth:`_engine.Connection.get_execution_options`
+
+            :ref:`orm_queryguide_execution_options` - documentation on all
+            ORM-specific execution options
+
+        """  # noqa
+        if self._has_events or self.engine._has_events:
+            self.dispatch.set_connection_execution_options(self, opt)
+        self._execution_options = self._execution_options.union(opt)
+        self.dialect.set_connection_execution_options(self, opt)
+        return self
+
+    def get_execution_options(self) -> _ExecuteOptions:
+        """Get the non-SQL options which will take effect during execution.
+
+        .. versionadded:: 1.3
+
+        .. seealso::
+
+            :meth:`_engine.Connection.execution_options`
+        """
+        return self._execution_options
+
+    @property
+    def _still_open_and_dbapi_connection_is_valid(self) -> bool:
+        pool_proxied_connection = self._dbapi_connection
+        return (
+            pool_proxied_connection is not None
+            and pool_proxied_connection.is_valid
+        )
+
+    @property
+    def closed(self) -> bool:
+        """Return True if this connection is closed."""
+
+        return self._dbapi_connection is None and not self.__can_reconnect
+
+    @property
+    def invalidated(self) -> bool:
+        """Return True if this connection was invalidated.
+
+        This does not indicate whether or not the connection was
+        invalidated at the pool level, however
+
+        """
+
+        # prior to 1.4, "invalid" was stored as a state independent of
+        # "closed", meaning an invalidated connection could be "closed",
+        # the _dbapi_connection would be None and closed=True, yet the
+        # "invalid" flag would stay True.  This meant that there were
+        # three separate states (open/valid, closed/valid, closed/invalid)
+        # when there is really no reason for that; a connection that's
+        # "closed" does not need to be "invalid".  So the state is now
+        # represented by the two facts alone.
+
+        pool_proxied_connection = self._dbapi_connection
+        return pool_proxied_connection is None and self.__can_reconnect
+
+    @property
+    def connection(self) -> PoolProxiedConnection:
+        """The underlying DB-API connection managed by this Connection.
+
+        This is a SQLAlchemy connection-pool proxied connection
+        which then has the attribute
+        :attr:`_pool._ConnectionFairy.dbapi_connection` that refers to the
+        actual driver connection.
+
+        .. seealso::
+
+
+            :ref:`dbapi_connections`
+
+        """
+
+        if self._dbapi_connection is None:
+            try:
+                return self._revalidate_connection()
+            except (exc.PendingRollbackError, exc.ResourceClosedError):
+                raise
+            except BaseException as e:
+                self._handle_dbapi_exception(e, None, None, None, None)
+        else:
+            return self._dbapi_connection
+
+    def get_isolation_level(self) -> IsolationLevel:
+        """Return the current **actual** isolation level that's present on
+        the database within the scope of this connection.
+
+        This attribute will perform a live SQL operation against the database
+        in order to procure the current isolation level, so the value returned
+        is the actual level on the underlying DBAPI connection regardless of
+        how this state was set. This will be one of the four actual isolation
+        modes ``READ UNCOMMITTED``, ``READ COMMITTED``, ``REPEATABLE READ``,
+        ``SERIALIZABLE``. It will **not** include the ``AUTOCOMMIT`` isolation
+        level setting. Third party dialects may also feature additional
+        isolation level settings.
+
+        .. note::  This method **will not report** on the ``AUTOCOMMIT``
+          isolation level, which is a separate :term:`dbapi` setting that's
+          independent of **actual** isolation level.  When ``AUTOCOMMIT`` is
+          in use, the database connection still has a "traditional" isolation
+          mode in effect, that is typically one of the four values
+          ``READ UNCOMMITTED``, ``READ COMMITTED``, ``REPEATABLE READ``,
+          ``SERIALIZABLE``.
+
+        Compare to the :attr:`_engine.Connection.default_isolation_level`
+        accessor which returns the isolation level that is present on the
+        database at initial connection time.
+
+        .. seealso::
+
+            :attr:`_engine.Connection.default_isolation_level`
+            - view default level
+
+            :paramref:`_sa.create_engine.isolation_level`
+            - set per :class:`_engine.Engine` isolation level
+
+            :paramref:`.Connection.execution_options.isolation_level`
+            - set per :class:`_engine.Connection` isolation level
+
+        """
+        dbapi_connection = self.connection.dbapi_connection
+        assert dbapi_connection is not None
+        try:
+            return self.dialect.get_isolation_level(dbapi_connection)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+
+    @property
+    def default_isolation_level(self) -> Optional[IsolationLevel]:
+        """The initial-connection time isolation level associated with the
+        :class:`_engine.Dialect` in use.
+
+        This value is independent of the
+        :paramref:`.Connection.execution_options.isolation_level` and
+        :paramref:`.Engine.execution_options.isolation_level` execution
+        options, and is determined by the :class:`_engine.Dialect` when the
+        first connection is created, by performing a SQL query against the
+        database for the current isolation level before any additional commands
+        have been emitted.
+
+        Calling this accessor does not invoke any new SQL queries.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.get_isolation_level`
+            - view current actual isolation level
+
+            :paramref:`_sa.create_engine.isolation_level`
+            - set per :class:`_engine.Engine` isolation level
+
+            :paramref:`.Connection.execution_options.isolation_level`
+            - set per :class:`_engine.Connection` isolation level
+
+        """
+        return self.dialect.default_isolation_level
+
+    def _invalid_transaction(self) -> NoReturn:
+        raise exc.PendingRollbackError(
+            "Can't reconnect until invalid %stransaction is rolled "
+            "back.  Please rollback() fully before proceeding"
+            % ("savepoint " if self._nested_transaction is not None else ""),
+            code="8s2b",
+        )
+
+    def _revalidate_connection(self) -> PoolProxiedConnection:
+        if self.__can_reconnect and self.invalidated:
+            if self._transaction is not None:
+                self._invalid_transaction()
+            self._dbapi_connection = self.engine.raw_connection()
+            return self._dbapi_connection
+        raise exc.ResourceClosedError("This Connection is closed")
+
+    @property
+    def info(self) -> _InfoType:
+        """Info dictionary associated with the underlying DBAPI connection
+        referred to by this :class:`_engine.Connection`, allowing user-defined
+        data to be associated with the connection.
+
+        The data here will follow along with the DBAPI connection including
+        after it is returned to the connection pool and used again
+        in subsequent instances of :class:`_engine.Connection`.
+
+        """
+
+        return self.connection.info
+
+    def invalidate(self, exception: Optional[BaseException] = None) -> None:
+        """Invalidate the underlying DBAPI connection associated with
+        this :class:`_engine.Connection`.
+
+        An attempt will be made to close the underlying DBAPI connection
+        immediately; however if this operation fails, the error is logged
+        but not raised.  The connection is then discarded whether or not
+        close() succeeded.
+
+        Upon the next use (where "use" typically means using the
+        :meth:`_engine.Connection.execute` method or similar),
+        this :class:`_engine.Connection` will attempt to
+        procure a new DBAPI connection using the services of the
+        :class:`_pool.Pool` as a source of connectivity (e.g.
+        a "reconnection").
+
+        If a transaction was in progress (e.g. the
+        :meth:`_engine.Connection.begin` method has been called) when
+        :meth:`_engine.Connection.invalidate` method is called, at the DBAPI
+        level all state associated with this transaction is lost, as
+        the DBAPI connection is closed.  The :class:`_engine.Connection`
+        will not allow a reconnection to proceed until the
+        :class:`.Transaction` object is ended, by calling the
+        :meth:`.Transaction.rollback` method; until that point, any attempt at
+        continuing to use the :class:`_engine.Connection` will raise an
+        :class:`~sqlalchemy.exc.InvalidRequestError`.
+        This is to prevent applications from accidentally
+        continuing an ongoing transactional operations despite the
+        fact that the transaction has been lost due to an
+        invalidation.
+
+        The :meth:`_engine.Connection.invalidate` method,
+        just like auto-invalidation,
+        will at the connection pool level invoke the
+        :meth:`_events.PoolEvents.invalidate` event.
+
+        :param exception: an optional ``Exception`` instance that's the
+         reason for the invalidation.  is passed along to event handlers
+         and logging functions.
+
+        .. seealso::
+
+            :ref:`pool_connection_invalidation`
+
+        """
+
+        if self.invalidated:
+            return
+
+        if self.closed:
+            raise exc.ResourceClosedError("This Connection is closed")
+
+        if self._still_open_and_dbapi_connection_is_valid:
+            pool_proxied_connection = self._dbapi_connection
+            assert pool_proxied_connection is not None
+            pool_proxied_connection.invalidate(exception)
+
+        self._dbapi_connection = None
+
+    def detach(self) -> None:
+        """Detach the underlying DB-API connection from its connection pool.
+
+        E.g.::
+
+            with engine.connect() as conn:
+                conn.detach()
+                conn.execute(text("SET search_path TO schema1, schema2"))
+
+                # work with connection
+
+            # connection is fully closed (since we used "with:", can
+            # also call .close())
+
+        This :class:`_engine.Connection` instance will remain usable.
+        When closed
+        (or exited from a context manager context as above),
+        the DB-API connection will be literally closed and not
+        returned to its originating pool.
+
+        This method can be used to insulate the rest of an application
+        from a modified state on a connection (such as a transaction
+        isolation level or similar).
+
+        """
+
+        if self.closed:
+            raise exc.ResourceClosedError("This Connection is closed")
+
+        pool_proxied_connection = self._dbapi_connection
+        if pool_proxied_connection is None:
+            raise exc.InvalidRequestError(
+                "Can't detach an invalidated Connection"
+            )
+        pool_proxied_connection.detach()
+
+    def _autobegin(self) -> None:
+        if self._allow_autobegin and not self.__in_begin:
+            self.begin()
+
+    def begin(self) -> RootTransaction:
+        """Begin a transaction prior to autobegin occurring.
+
+        E.g.::
+
+            with engine.connect() as conn:
+                with conn.begin() as trans:
+                    conn.execute(table.insert(), {"username": "sandy"})
+
+        The returned object is an instance of :class:`_engine.RootTransaction`.
+        This object represents the "scope" of the transaction,
+        which completes when either the :meth:`_engine.Transaction.rollback`
+        or :meth:`_engine.Transaction.commit` method is called; the object
+        also works as a context manager as illustrated above.
+
+        The :meth:`_engine.Connection.begin` method begins a
+        transaction that normally will be begun in any case when the connection
+        is first used to execute a statement.  The reason this method might be
+        used would be to invoke the :meth:`_events.ConnectionEvents.begin`
+        event at a specific time, or to organize code within the scope of a
+        connection checkout in terms of context managed blocks, such as::
+
+            with engine.connect() as conn:
+                with conn.begin():
+                    conn.execute(...)
+                    conn.execute(...)
+
+                with conn.begin():
+                    conn.execute(...)
+                    conn.execute(...)
+
+        The above code is not  fundamentally any different in its behavior than
+        the following code  which does not use
+        :meth:`_engine.Connection.begin`; the below style is known
+        as "commit as you go" style::
+
+            with engine.connect() as conn:
+                conn.execute(...)
+                conn.execute(...)
+                conn.commit()
+
+                conn.execute(...)
+                conn.execute(...)
+                conn.commit()
+
+        From a database point of view, the :meth:`_engine.Connection.begin`
+        method does not emit any SQL or change the state of the underlying
+        DBAPI connection in any way; the Python DBAPI does not have any
+        concept of explicit transaction begin.
+
+        .. seealso::
+
+            :ref:`tutorial_working_with_transactions` - in the
+            :ref:`unified_tutorial`
+
+            :meth:`_engine.Connection.begin_nested` - use a SAVEPOINT
+
+            :meth:`_engine.Connection.begin_twophase` -
+            use a two phase /XID transaction
+
+            :meth:`_engine.Engine.begin` - context manager available from
+            :class:`_engine.Engine`
+
+        """
+        if self._transaction is None:
+            self._transaction = RootTransaction(self)
+            return self._transaction
+        else:
+            raise exc.InvalidRequestError(
+                "This connection has already initialized a SQLAlchemy "
+                "Transaction() object via begin() or autobegin; can't "
+                "call begin() here unless rollback() or commit() "
+                "is called first."
+            )
+
+    def begin_nested(self) -> NestedTransaction:
+        """Begin a nested transaction (i.e. SAVEPOINT) and return a transaction
+        handle that controls the scope of the SAVEPOINT.
+
+        E.g.::
+
+            with engine.begin() as connection:
+                with connection.begin_nested():
+                    connection.execute(table.insert(), {"username": "sandy"})
+
+        The returned object is an instance of
+        :class:`_engine.NestedTransaction`, which includes transactional
+        methods :meth:`_engine.NestedTransaction.commit` and
+        :meth:`_engine.NestedTransaction.rollback`; for a nested transaction,
+        these methods correspond to the operations "RELEASE SAVEPOINT <name>"
+        and "ROLLBACK TO SAVEPOINT <name>". The name of the savepoint is local
+        to the :class:`_engine.NestedTransaction` object and is generated
+        automatically. Like any other :class:`_engine.Transaction`, the
+        :class:`_engine.NestedTransaction` may be used as a context manager as
+        illustrated above which will "release" or "rollback" corresponding to
+        if the operation within the block were successful or raised an
+        exception.
+
+        Nested transactions require SAVEPOINT support in the underlying
+        database, else the behavior is undefined. SAVEPOINT is commonly used to
+        run operations within a transaction that may fail, while continuing the
+        outer transaction. E.g.::
+
+            from sqlalchemy import exc
+
+            with engine.begin() as connection:
+                trans = connection.begin_nested()
+                try:
+                    connection.execute(table.insert(), {"username": "sandy"})
+                    trans.commit()
+                except exc.IntegrityError:  # catch for duplicate username
+                    trans.rollback()  # rollback to savepoint
+
+                # outer transaction continues
+                connection.execute(...)
+
+        If :meth:`_engine.Connection.begin_nested` is called without first
+        calling :meth:`_engine.Connection.begin` or
+        :meth:`_engine.Engine.begin`, the :class:`_engine.Connection` object
+        will "autobegin" the outer transaction first. This outer transaction
+        may be committed using "commit-as-you-go" style, e.g.::
+
+            with engine.connect() as connection:  # begin() wasn't called
+
+                with connection.begin_nested():  # will auto-"begin()" first
+                    connection.execute(...)
+                # savepoint is released
+
+                connection.execute(...)
+
+                # explicitly commit outer transaction
+                connection.commit()
+
+                # can continue working with connection here
+
+        .. versionchanged:: 2.0
+
+            :meth:`_engine.Connection.begin_nested` will now participate
+            in the connection "autobegin" behavior that is new as of
+            2.0 / "future" style connections in 1.4.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.begin`
+
+            :ref:`session_begin_nested` - ORM support for SAVEPOINT
+
+        """
+        if self._transaction is None:
+            self._autobegin()
+
+        return NestedTransaction(self)
+
+    def begin_twophase(self, xid: Optional[Any] = None) -> TwoPhaseTransaction:
+        """Begin a two-phase or XA transaction and return a transaction
+        handle.
+
+        The returned object is an instance of :class:`.TwoPhaseTransaction`,
+        which in addition to the methods provided by
+        :class:`.Transaction`, also provides a
+        :meth:`~.TwoPhaseTransaction.prepare` method.
+
+        :param xid: the two phase transaction id.  If not supplied, a
+          random id will be generated.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.begin`
+
+            :meth:`_engine.Connection.begin_twophase`
+
+        """
+
+        if self._transaction is not None:
+            raise exc.InvalidRequestError(
+                "Cannot start a two phase transaction when a transaction "
+                "is already in progress."
+            )
+        if xid is None:
+            xid = self.engine.dialect.create_xid()
+        return TwoPhaseTransaction(self, xid)
+
+    def commit(self) -> None:
+        """Commit the transaction that is currently in progress.
+
+        This method commits the current transaction if one has been started.
+        If no transaction was started, the method has no effect, assuming
+        the connection is in a non-invalidated state.
+
+        A transaction is begun on a :class:`_engine.Connection` automatically
+        whenever a statement is first executed, or when the
+        :meth:`_engine.Connection.begin` method is called.
+
+        .. note:: The :meth:`_engine.Connection.commit` method only acts upon
+          the primary database transaction that is linked to the
+          :class:`_engine.Connection` object.  It does not operate upon a
+          SAVEPOINT that would have been invoked from the
+          :meth:`_engine.Connection.begin_nested` method; for control of a
+          SAVEPOINT, call :meth:`_engine.NestedTransaction.commit` on the
+          :class:`_engine.NestedTransaction` that is returned by the
+          :meth:`_engine.Connection.begin_nested` method itself.
+
+
+        """
+        if self._transaction:
+            self._transaction.commit()
+
+    def rollback(self) -> None:
+        """Roll back the transaction that is currently in progress.
+
+        This method rolls back the current transaction if one has been started.
+        If no transaction was started, the method has no effect.  If a
+        transaction was started and the connection is in an invalidated state,
+        the transaction is cleared using this method.
+
+        A transaction is begun on a :class:`_engine.Connection` automatically
+        whenever a statement is first executed, or when the
+        :meth:`_engine.Connection.begin` method is called.
+
+        .. note:: The :meth:`_engine.Connection.rollback` method only acts
+          upon the primary database transaction that is linked to the
+          :class:`_engine.Connection` object.  It does not operate upon a
+          SAVEPOINT that would have been invoked from the
+          :meth:`_engine.Connection.begin_nested` method; for control of a
+          SAVEPOINT, call :meth:`_engine.NestedTransaction.rollback` on the
+          :class:`_engine.NestedTransaction` that is returned by the
+          :meth:`_engine.Connection.begin_nested` method itself.
+
+
+        """
+        if self._transaction:
+            self._transaction.rollback()
+
+    def recover_twophase(self) -> List[Any]:
+        return self.engine.dialect.do_recover_twophase(self)
+
+    def rollback_prepared(self, xid: Any, recover: bool = False) -> None:
+        self.engine.dialect.do_rollback_twophase(self, xid, recover=recover)
+
+    def commit_prepared(self, xid: Any, recover: bool = False) -> None:
+        self.engine.dialect.do_commit_twophase(self, xid, recover=recover)
+
+    def in_transaction(self) -> bool:
+        """Return True if a transaction is in progress."""
+        return self._transaction is not None and self._transaction.is_active
+
+    def in_nested_transaction(self) -> bool:
+        """Return True if a transaction is in progress."""
+        return (
+            self._nested_transaction is not None
+            and self._nested_transaction.is_active
+        )
+
+    def _is_autocommit_isolation(self) -> bool:
+        opt_iso = self._execution_options.get("isolation_level", None)
+        return bool(
+            opt_iso == "AUTOCOMMIT"
+            or (
+                opt_iso is None
+                and self.engine.dialect._on_connect_isolation_level
+                == "AUTOCOMMIT"
+            )
+        )
+
+    def _get_required_transaction(self) -> RootTransaction:
+        trans = self._transaction
+        if trans is None:
+            raise exc.InvalidRequestError("connection is not in a transaction")
+        return trans
+
+    def _get_required_nested_transaction(self) -> NestedTransaction:
+        trans = self._nested_transaction
+        if trans is None:
+            raise exc.InvalidRequestError(
+                "connection is not in a nested transaction"
+            )
+        return trans
+
+    def get_transaction(self) -> Optional[RootTransaction]:
+        """Return the current root transaction in progress, if any.
+
+        .. versionadded:: 1.4
+
+        """
+
+        return self._transaction
+
+    def get_nested_transaction(self) -> Optional[NestedTransaction]:
+        """Return the current nested transaction in progress, if any.
+
+        .. versionadded:: 1.4
+
+        """
+        return self._nested_transaction
+
+    def _begin_impl(self, transaction: RootTransaction) -> None:
+        if self._echo:
+            if self._is_autocommit_isolation():
+                self._log_info(
+                    "BEGIN (implicit; DBAPI should not BEGIN due to "
+                    "autocommit mode)"
+                )
+            else:
+                self._log_info("BEGIN (implicit)")
+
+        self.__in_begin = True
+
+        if self._has_events or self.engine._has_events:
+            self.dispatch.begin(self)
+
+        try:
+            self.engine.dialect.do_begin(self.connection)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+        finally:
+            self.__in_begin = False
+
+    def _rollback_impl(self) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.rollback(self)
+
+        if self._still_open_and_dbapi_connection_is_valid:
+            if self._echo:
+                if self._is_autocommit_isolation():
+                    self._log_info(
+                        "ROLLBACK using DBAPI connection.rollback(), "
+                        "DBAPI should ignore due to autocommit mode"
+                    )
+                else:
+                    self._log_info("ROLLBACK")
+            try:
+                self.engine.dialect.do_rollback(self.connection)
+            except BaseException as e:
+                self._handle_dbapi_exception(e, None, None, None, None)
+
+    def _commit_impl(self) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.commit(self)
+
+        if self._echo:
+            if self._is_autocommit_isolation():
+                self._log_info(
+                    "COMMIT using DBAPI connection.commit(), "
+                    "DBAPI should ignore due to autocommit mode"
+                )
+            else:
+                self._log_info("COMMIT")
+        try:
+            self.engine.dialect.do_commit(self.connection)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+
+    def _savepoint_impl(self, name: Optional[str] = None) -> str:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.savepoint(self, name)
+
+        if name is None:
+            self.__savepoint_seq += 1
+            name = "sa_savepoint_%s" % self.__savepoint_seq
+        self.engine.dialect.do_savepoint(self, name)
+        return name
+
+    def _rollback_to_savepoint_impl(self, name: str) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.rollback_savepoint(self, name, None)
+
+        if self._still_open_and_dbapi_connection_is_valid:
+            self.engine.dialect.do_rollback_to_savepoint(self, name)
+
+    def _release_savepoint_impl(self, name: str) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.release_savepoint(self, name, None)
+
+        self.engine.dialect.do_release_savepoint(self, name)
+
+    def _begin_twophase_impl(self, transaction: TwoPhaseTransaction) -> None:
+        if self._echo:
+            self._log_info("BEGIN TWOPHASE (implicit)")
+        if self._has_events or self.engine._has_events:
+            self.dispatch.begin_twophase(self, transaction.xid)
+
+        self.__in_begin = True
+        try:
+            self.engine.dialect.do_begin_twophase(self, transaction.xid)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+        finally:
+            self.__in_begin = False
+
+    def _prepare_twophase_impl(self, xid: Any) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.prepare_twophase(self, xid)
+
+        assert isinstance(self._transaction, TwoPhaseTransaction)
+        try:
+            self.engine.dialect.do_prepare_twophase(self, xid)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+
+    def _rollback_twophase_impl(self, xid: Any, is_prepared: bool) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.rollback_twophase(self, xid, is_prepared)
+
+        if self._still_open_and_dbapi_connection_is_valid:
+            assert isinstance(self._transaction, TwoPhaseTransaction)
+            try:
+                self.engine.dialect.do_rollback_twophase(
+                    self, xid, is_prepared
+                )
+            except BaseException as e:
+                self._handle_dbapi_exception(e, None, None, None, None)
+
+    def _commit_twophase_impl(self, xid: Any, is_prepared: bool) -> None:
+        if self._has_events or self.engine._has_events:
+            self.dispatch.commit_twophase(self, xid, is_prepared)
+
+        assert isinstance(self._transaction, TwoPhaseTransaction)
+        try:
+            self.engine.dialect.do_commit_twophase(self, xid, is_prepared)
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+
+    def close(self) -> None:
+        """Close this :class:`_engine.Connection`.
+
+        This results in a release of the underlying database
+        resources, that is, the DBAPI connection referenced
+        internally. The DBAPI connection is typically restored
+        back to the connection-holding :class:`_pool.Pool` referenced
+        by the :class:`_engine.Engine` that produced this
+        :class:`_engine.Connection`. Any transactional state present on
+        the DBAPI connection is also unconditionally released via
+        the DBAPI connection's ``rollback()`` method, regardless
+        of any :class:`.Transaction` object that may be
+        outstanding with regards to this :class:`_engine.Connection`.
+
+        This has the effect of also calling :meth:`_engine.Connection.rollback`
+        if any transaction is in place.
+
+        After :meth:`_engine.Connection.close` is called, the
+        :class:`_engine.Connection` is permanently in a closed state,
+        and will allow no further operations.
+
+        """
+
+        if self._transaction:
+            self._transaction.close()
+            skip_reset = True
+        else:
+            skip_reset = False
+
+        if self._dbapi_connection is not None:
+            conn = self._dbapi_connection
+
+            # as we just closed the transaction, close the connection
+            # pool connection without doing an additional reset
+            if skip_reset:
+                cast("_ConnectionFairy", conn)._close_special(
+                    transaction_reset=True
+                )
+            else:
+                conn.close()
+
+            # There is a slight chance that conn.close() may have
+            # triggered an invalidation here in which case
+            # _dbapi_connection would already be None, however usually
+            # it will be non-None here and in a "closed" state.
+            self._dbapi_connection = None
+        self.__can_reconnect = False
+
+    @overload
+    def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Optional[_T]: ...
+
+    @overload
+    def scalar(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Any: ...
+
+    def scalar(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Any:
+        r"""Executes a SQL statement construct and returns a scalar object.
+
+        This method is shorthand for invoking the
+        :meth:`_engine.Result.scalar` method after invoking the
+        :meth:`_engine.Connection.execute` method.  Parameters are equivalent.
+
+        :return: a scalar Python value representing the first column of the
+         first row returned.
+
+        """
+        distilled_parameters = _distill_params_20(parameters)
+        try:
+            meth = statement._execute_on_scalar
+        except AttributeError as err:
+            raise exc.ObjectNotExecutableError(statement) from err
+        else:
+            return meth(
+                self,
+                distilled_parameters,
+                execution_options or NO_OPTIONS,
+            )
+
+    @overload
+    def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    def scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[Any]: ...
+
+    def scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[Any]:
+        """Executes and returns a scalar result set, which yields scalar values
+        from the first column of each row.
+
+        This method is equivalent to calling :meth:`_engine.Connection.execute`
+        to receive a :class:`_result.Result` object, then invoking the
+        :meth:`_result.Result.scalars` method to produce a
+        :class:`_result.ScalarResult` instance.
+
+        :return: a :class:`_result.ScalarResult`
+
+        .. versionadded:: 1.4.24
+
+        """
+
+        return self.execute(
+            statement, parameters, execution_options=execution_options
+        ).scalars()
+
+    @overload
+    def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[_T]: ...
+
+    @overload
+    def execute(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]: ...
+
+    def execute(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]:
+        r"""Executes a SQL statement construct and returns a
+        :class:`_engine.CursorResult`.
+
+        :param statement: The statement to be executed.  This is always
+         an object that is in both the :class:`_expression.ClauseElement` and
+         :class:`_expression.Executable` hierarchies, including:
+
+         * :class:`_expression.Select`
+         * :class:`_expression.Insert`, :class:`_expression.Update`,
+           :class:`_expression.Delete`
+         * :class:`_expression.TextClause` and
+           :class:`_expression.TextualSelect`
+         * :class:`_schema.DDL` and objects which inherit from
+           :class:`_schema.ExecutableDDLElement`
+
+        :param parameters: parameters which will be bound into the statement.
+         This may be either a dictionary of parameter names to values,
+         or a mutable sequence (e.g. a list) of dictionaries.  When a
+         list of dictionaries is passed, the underlying statement execution
+         will make use of the DBAPI ``cursor.executemany()`` method.
+         When a single dictionary is passed, the DBAPI ``cursor.execute()``
+         method will be used.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the statement execution.  This
+         dictionary can provide a subset of the options that are accepted
+         by :meth:`_engine.Connection.execution_options`.
+
+        :return: a :class:`_engine.Result` object.
+
+        """
+        distilled_parameters = _distill_params_20(parameters)
+        try:
+            meth = statement._execute_on_connection
+        except AttributeError as err:
+            raise exc.ObjectNotExecutableError(statement) from err
+        else:
+            return meth(
+                self,
+                distilled_parameters,
+                execution_options or NO_OPTIONS,
+            )
+
+    def _execute_function(
+        self,
+        func: FunctionElement[Any],
+        distilled_parameters: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> CursorResult[Any]:
+        """Execute a sql.FunctionElement object."""
+
+        return self._execute_clauseelement(
+            func.select(), distilled_parameters, execution_options
+        )
+
+    def _execute_default(
+        self,
+        default: DefaultGenerator,
+        distilled_parameters: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> Any:
+        """Execute a schema.ColumnDefault object."""
+
+        execution_options = self._execution_options.merge_with(
+            execution_options
+        )
+
+        event_multiparams: Optional[_CoreMultiExecuteParams]
+        event_params: Optional[_CoreAnyExecuteParams]
+
+        # note for event handlers, the "distilled parameters" which is always
+        # a list of dicts is broken out into separate "multiparams" and
+        # "params" collections, which allows the handler to distinguish
+        # between an executemany and execute style set of parameters.
+        if self._has_events or self.engine._has_events:
+            (
+                default,
+                distilled_parameters,
+                event_multiparams,
+                event_params,
+            ) = self._invoke_before_exec_event(
+                default, distilled_parameters, execution_options
+            )
+        else:
+            event_multiparams = event_params = None
+
+        try:
+            conn = self._dbapi_connection
+            if conn is None:
+                conn = self._revalidate_connection()
+
+            dialect = self.dialect
+            ctx = dialect.execution_ctx_cls._init_default(
+                dialect, self, conn, execution_options
+            )
+        except (exc.PendingRollbackError, exc.ResourceClosedError):
+            raise
+        except BaseException as e:
+            self._handle_dbapi_exception(e, None, None, None, None)
+
+        ret = ctx._exec_default(None, default, None)
+
+        if self._has_events or self.engine._has_events:
+            self.dispatch.after_execute(
+                self,
+                default,
+                event_multiparams,
+                event_params,
+                execution_options,
+                ret,
+            )
+
+        return ret
+
+    def _execute_ddl(
+        self,
+        ddl: ExecutableDDLElement,
+        distilled_parameters: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> CursorResult[Any]:
+        """Execute a schema.DDL object."""
+
+        exec_opts = ddl._execution_options.merge_with(
+            self._execution_options, execution_options
+        )
+
+        event_multiparams: Optional[_CoreMultiExecuteParams]
+        event_params: Optional[_CoreSingleExecuteParams]
+
+        if self._has_events or self.engine._has_events:
+            (
+                ddl,
+                distilled_parameters,
+                event_multiparams,
+                event_params,
+            ) = self._invoke_before_exec_event(
+                ddl, distilled_parameters, exec_opts
+            )
+        else:
+            event_multiparams = event_params = None
+
+        schema_translate_map = exec_opts.get("schema_translate_map", None)
+
+        dialect = self.dialect
+
+        compiled = ddl.compile(
+            dialect=dialect, schema_translate_map=schema_translate_map
+        )
+        ret = self._execute_context(
+            dialect,
+            dialect.execution_ctx_cls._init_ddl,
+            compiled,
+            None,
+            exec_opts,
+            compiled,
+        )
+        if self._has_events or self.engine._has_events:
+            self.dispatch.after_execute(
+                self,
+                ddl,
+                event_multiparams,
+                event_params,
+                exec_opts,
+                ret,
+            )
+        return ret
+
+    def _invoke_before_exec_event(
+        self,
+        elem: Any,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: _ExecuteOptions,
+    ) -> Tuple[
+        Any,
+        _CoreMultiExecuteParams,
+        _CoreMultiExecuteParams,
+        _CoreSingleExecuteParams,
+    ]:
+        event_multiparams: _CoreMultiExecuteParams
+        event_params: _CoreSingleExecuteParams
+
+        if len(distilled_params) == 1:
+            event_multiparams, event_params = [], distilled_params[0]
+        else:
+            event_multiparams, event_params = distilled_params, {}
+
+        for fn in self.dispatch.before_execute:
+            elem, event_multiparams, event_params = fn(
+                self,
+                elem,
+                event_multiparams,
+                event_params,
+                execution_options,
+            )
+
+        if event_multiparams:
+            distilled_params = list(event_multiparams)
+            if event_params:
+                raise exc.InvalidRequestError(
+                    "Event handler can't return non-empty multiparams "
+                    "and params at the same time"
+                )
+        elif event_params:
+            distilled_params = [event_params]
+        else:
+            distilled_params = []
+
+        return elem, distilled_params, event_multiparams, event_params
+
+    def _execute_clauseelement(
+        self,
+        elem: Executable,
+        distilled_parameters: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> CursorResult[Any]:
+        """Execute a sql.ClauseElement object."""
+
+        execution_options = elem._execution_options.merge_with(
+            self._execution_options, execution_options
+        )
+
+        has_events = self._has_events or self.engine._has_events
+        if has_events:
+            (
+                elem,
+                distilled_parameters,
+                event_multiparams,
+                event_params,
+            ) = self._invoke_before_exec_event(
+                elem, distilled_parameters, execution_options
+            )
+
+        if distilled_parameters:
+            # ensure we don't retain a link to the view object for keys()
+            # which links to the values, which we don't want to cache
+            keys = sorted(distilled_parameters[0])
+            for_executemany = len(distilled_parameters) > 1
+        else:
+            keys = []
+            for_executemany = False
+
+        dialect = self.dialect
+
+        schema_translate_map = execution_options.get(
+            "schema_translate_map", None
+        )
+
+        compiled_cache: Optional[CompiledCacheType] = execution_options.get(
+            "compiled_cache", self.engine._compiled_cache
+        )
+
+        compiled_sql, extracted_params, cache_hit = elem._compile_w_cache(
+            dialect=dialect,
+            compiled_cache=compiled_cache,
+            column_keys=keys,
+            for_executemany=for_executemany,
+            schema_translate_map=schema_translate_map,
+            linting=self.dialect.compiler_linting | compiler.WARN_LINTING,
+        )
+        ret = self._execute_context(
+            dialect,
+            dialect.execution_ctx_cls._init_compiled,
+            compiled_sql,
+            distilled_parameters,
+            execution_options,
+            compiled_sql,
+            distilled_parameters,
+            elem,
+            extracted_params,
+            cache_hit=cache_hit,
+        )
+        if has_events:
+            self.dispatch.after_execute(
+                self,
+                elem,
+                event_multiparams,
+                event_params,
+                execution_options,
+                ret,
+            )
+        return ret
+
+    def _execute_compiled(
+        self,
+        compiled: Compiled,
+        distilled_parameters: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter = _EMPTY_EXECUTION_OPTS,
+    ) -> CursorResult[Any]:
+        """Execute a sql.Compiled object.
+
+        TODO: why do we have this?   likely deprecate or remove
+
+        """
+
+        execution_options = compiled.execution_options.merge_with(
+            self._execution_options, execution_options
+        )
+
+        if self._has_events or self.engine._has_events:
+            (
+                compiled,
+                distilled_parameters,
+                event_multiparams,
+                event_params,
+            ) = self._invoke_before_exec_event(
+                compiled, distilled_parameters, execution_options
+            )
+
+        dialect = self.dialect
+
+        ret = self._execute_context(
+            dialect,
+            dialect.execution_ctx_cls._init_compiled,
+            compiled,
+            distilled_parameters,
+            execution_options,
+            compiled,
+            distilled_parameters,
+            None,
+            None,
+        )
+        if self._has_events or self.engine._has_events:
+            self.dispatch.after_execute(
+                self,
+                compiled,
+                event_multiparams,
+                event_params,
+                execution_options,
+                ret,
+            )
+        return ret
+
+    def exec_driver_sql(
+        self,
+        statement: str,
+        parameters: Optional[_DBAPIAnyExecuteParams] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]:
+        r"""Executes a string SQL statement on the DBAPI cursor directly,
+        without any SQL compilation steps.
+
+        This can be used to pass any string directly to the
+        ``cursor.execute()`` method of the DBAPI in use.
+
+        :param statement: The statement str to be executed.   Bound parameters
+         must use the underlying DBAPI's paramstyle, such as "qmark",
+         "pyformat", "format", etc.
+
+        :param parameters: represent bound parameter values to be used in the
+         execution.  The format is one of:   a dictionary of named parameters,
+         a tuple of positional parameters, or a list containing either
+         dictionaries or tuples for multiple-execute support.
+
+        :return: a :class:`_engine.CursorResult`.
+
+         E.g. multiple dictionaries::
+
+
+             conn.exec_driver_sql(
+                 "INSERT INTO table (id, value) VALUES (%(id)s, %(value)s)",
+                 [{"id": 1, "value": "v1"}, {"id": 2, "value": "v2"}],
+             )
+
+         Single dictionary::
+
+             conn.exec_driver_sql(
+                 "INSERT INTO table (id, value) VALUES (%(id)s, %(value)s)",
+                 dict(id=1, value="v1"),
+             )
+
+         Single tuple::
+
+             conn.exec_driver_sql(
+                 "INSERT INTO table (id, value) VALUES (?, ?)", (1, "v1")
+             )
+
+         .. note:: The :meth:`_engine.Connection.exec_driver_sql` method does
+             not participate in the
+             :meth:`_events.ConnectionEvents.before_execute` and
+             :meth:`_events.ConnectionEvents.after_execute` events.   To
+             intercept calls to :meth:`_engine.Connection.exec_driver_sql`, use
+             :meth:`_events.ConnectionEvents.before_cursor_execute` and
+             :meth:`_events.ConnectionEvents.after_cursor_execute`.
+
+         .. seealso::
+
+            :pep:`249`
+
+        """
+
+        distilled_parameters = _distill_raw_params(parameters)
+
+        execution_options = self._execution_options.merge_with(
+            execution_options
+        )
+
+        dialect = self.dialect
+        ret = self._execute_context(
+            dialect,
+            dialect.execution_ctx_cls._init_statement,
+            statement,
+            None,
+            execution_options,
+            statement,
+            distilled_parameters,
+        )
+
+        return ret
+
+    def _execute_context(
+        self,
+        dialect: Dialect,
+        constructor: Callable[..., ExecutionContext],
+        statement: Union[str, Compiled],
+        parameters: Optional[_AnyMultiExecuteParams],
+        execution_options: _ExecuteOptions,
+        *args: Any,
+        **kw: Any,
+    ) -> CursorResult[Any]:
+        """Create an :class:`.ExecutionContext` and execute, returning
+        a :class:`_engine.CursorResult`."""
+
+        if execution_options:
+            yp = execution_options.get("yield_per", None)
+            if yp:
+                execution_options = execution_options.union(
+                    {"stream_results": True, "max_row_buffer": yp}
+                )
+        try:
+            conn = self._dbapi_connection
+            if conn is None:
+                conn = self._revalidate_connection()
+
+            context = constructor(
+                dialect, self, conn, execution_options, *args, **kw
+            )
+        except (exc.PendingRollbackError, exc.ResourceClosedError):
+            raise
+        except BaseException as e:
+            self._handle_dbapi_exception(
+                e, str(statement), parameters, None, None
+            )
+
+        if (
+            self._transaction
+            and not self._transaction.is_active
+            or (
+                self._nested_transaction
+                and not self._nested_transaction.is_active
+            )
+        ):
+            self._invalid_transaction()
+
+        elif self._trans_context_manager:
+            TransactionalContext._trans_ctx_check(self)
+
+        if self._transaction is None:
+            self._autobegin()
+
+        context.pre_exec()
+
+        if context.execute_style is ExecuteStyle.INSERTMANYVALUES:
+            return self._exec_insertmany_context(dialect, context)
+        else:
+            return self._exec_single_context(
+                dialect, context, statement, parameters
+            )
+
+    def _exec_single_context(
+        self,
+        dialect: Dialect,
+        context: ExecutionContext,
+        statement: Union[str, Compiled],
+        parameters: Optional[_AnyMultiExecuteParams],
+    ) -> CursorResult[Any]:
+        """continue the _execute_context() method for a single DBAPI
+        cursor.execute() or cursor.executemany() call.
+
+        """
+        if dialect.bind_typing is BindTyping.SETINPUTSIZES:
+            generic_setinputsizes = context._prepare_set_input_sizes()
+
+            if generic_setinputsizes:
+                try:
+                    dialect.do_set_input_sizes(
+                        context.cursor, generic_setinputsizes, context
+                    )
+                except BaseException as e:
+                    self._handle_dbapi_exception(
+                        e, str(statement), parameters, None, context
+                    )
+
+        cursor, str_statement, parameters = (
+            context.cursor,
+            context.statement,
+            context.parameters,
+        )
+
+        effective_parameters: Optional[_AnyExecuteParams]
+
+        if not context.executemany:
+            effective_parameters = parameters[0]
+        else:
+            effective_parameters = parameters
+
+        if self._has_events or self.engine._has_events:
+            for fn in self.dispatch.before_cursor_execute:
+                str_statement, effective_parameters = fn(
+                    self,
+                    cursor,
+                    str_statement,
+                    effective_parameters,
+                    context,
+                    context.executemany,
+                )
+
+        if self._echo:
+            self._log_info(str_statement)
+
+            stats = context._get_cache_stats()
+
+            if not self.engine.hide_parameters:
+                self._log_info(
+                    "[%s] %r",
+                    stats,
+                    sql_util._repr_params(
+                        effective_parameters,
+                        batches=10,
+                        ismulti=context.executemany,
+                    ),
+                )
+            else:
+                self._log_info(
+                    "[%s] [SQL parameters hidden due to hide_parameters=True]",
+                    stats,
+                )
+
+        evt_handled: bool = False
+        try:
+            if context.execute_style is ExecuteStyle.EXECUTEMANY:
+                effective_parameters = cast(
+                    "_CoreMultiExecuteParams", effective_parameters
+                )
+                if self.dialect._has_events:
+                    for fn in self.dialect.dispatch.do_executemany:
+                        if fn(
+                            cursor,
+                            str_statement,
+                            effective_parameters,
+                            context,
+                        ):
+                            evt_handled = True
+                            break
+                if not evt_handled:
+                    self.dialect.do_executemany(
+                        cursor,
+                        str_statement,
+                        effective_parameters,
+                        context,
+                    )
+            elif not effective_parameters and context.no_parameters:
+                if self.dialect._has_events:
+                    for fn in self.dialect.dispatch.do_execute_no_params:
+                        if fn(cursor, str_statement, context):
+                            evt_handled = True
+                            break
+                if not evt_handled:
+                    self.dialect.do_execute_no_params(
+                        cursor, str_statement, context
+                    )
+            else:
+                effective_parameters = cast(
+                    "_CoreSingleExecuteParams", effective_parameters
+                )
+                if self.dialect._has_events:
+                    for fn in self.dialect.dispatch.do_execute:
+                        if fn(
+                            cursor,
+                            str_statement,
+                            effective_parameters,
+                            context,
+                        ):
+                            evt_handled = True
+                            break
+                if not evt_handled:
+                    self.dialect.do_execute(
+                        cursor, str_statement, effective_parameters, context
+                    )
+
+            if self._has_events or self.engine._has_events:
+                self.dispatch.after_cursor_execute(
+                    self,
+                    cursor,
+                    str_statement,
+                    effective_parameters,
+                    context,
+                    context.executemany,
+                )
+
+            context.post_exec()
+
+            result = context._setup_result_proxy()
+
+        except BaseException as e:
+            self._handle_dbapi_exception(
+                e, str_statement, effective_parameters, cursor, context
+            )
+
+        return result
+
+    def _exec_insertmany_context(
+        self,
+        dialect: Dialect,
+        context: ExecutionContext,
+    ) -> CursorResult[Any]:
+        """continue the _execute_context() method for an "insertmanyvalues"
+        operation, which will invoke DBAPI
+        cursor.execute() one or more times with individual log and
+        event hook calls.
+
+        """
+
+        if dialect.bind_typing is BindTyping.SETINPUTSIZES:
+            generic_setinputsizes = context._prepare_set_input_sizes()
+        else:
+            generic_setinputsizes = None
+
+        cursor, str_statement, parameters = (
+            context.cursor,
+            context.statement,
+            context.parameters,
+        )
+
+        effective_parameters = parameters
+
+        engine_events = self._has_events or self.engine._has_events
+        if self.dialect._has_events:
+            do_execute_dispatch: Iterable[Any] = (
+                self.dialect.dispatch.do_execute
+            )
+        else:
+            do_execute_dispatch = ()
+
+        if self._echo:
+            stats = context._get_cache_stats() + " (insertmanyvalues)"
+
+        preserve_rowcount = context.execution_options.get(
+            "preserve_rowcount", False
+        )
+        rowcount = 0
+
+        for imv_batch in dialect._deliver_insertmanyvalues_batches(
+            self,
+            cursor,
+            str_statement,
+            effective_parameters,
+            generic_setinputsizes,
+            context,
+        ):
+            if imv_batch.processed_setinputsizes:
+                try:
+                    dialect.do_set_input_sizes(
+                        context.cursor,
+                        imv_batch.processed_setinputsizes,
+                        context,
+                    )
+                except BaseException as e:
+                    self._handle_dbapi_exception(
+                        e,
+                        sql_util._long_statement(imv_batch.replaced_statement),
+                        imv_batch.replaced_parameters,
+                        None,
+                        context,
+                        is_sub_exec=True,
+                    )
+
+            sub_stmt = imv_batch.replaced_statement
+            sub_params = imv_batch.replaced_parameters
+
+            if engine_events:
+                for fn in self.dispatch.before_cursor_execute:
+                    sub_stmt, sub_params = fn(
+                        self,
+                        cursor,
+                        sub_stmt,
+                        sub_params,
+                        context,
+                        True,
+                    )
+
+            if self._echo:
+                self._log_info(sql_util._long_statement(sub_stmt))
+
+                imv_stats = f""" {imv_batch.batchnum}/{
+                            imv_batch.total_batches
+                } ({
+                    'ordered'
+                    if imv_batch.rows_sorted else 'unordered'
+                }{
+                    '; batch not supported'
+                    if imv_batch.is_downgraded
+                    else ''
+                })"""
+
+                if imv_batch.batchnum == 1:
+                    stats += imv_stats
+                else:
+                    stats = f"insertmanyvalues{imv_stats}"
+
+                if not self.engine.hide_parameters:
+                    self._log_info(
+                        "[%s] %r",
+                        stats,
+                        sql_util._repr_params(
+                            sub_params,
+                            batches=10,
+                            ismulti=False,
+                        ),
+                    )
+                else:
+                    self._log_info(
+                        "[%s] [SQL parameters hidden due to "
+                        "hide_parameters=True]",
+                        stats,
+                    )
+
+            try:
+                for fn in do_execute_dispatch:
+                    if fn(
+                        cursor,
+                        sub_stmt,
+                        sub_params,
+                        context,
+                    ):
+                        break
+                else:
+                    dialect.do_execute(
+                        cursor,
+                        sub_stmt,
+                        sub_params,
+                        context,
+                    )
+
+            except BaseException as e:
+                self._handle_dbapi_exception(
+                    e,
+                    sql_util._long_statement(sub_stmt),
+                    sub_params,
+                    cursor,
+                    context,
+                    is_sub_exec=True,
+                )
+
+            if engine_events:
+                self.dispatch.after_cursor_execute(
+                    self,
+                    cursor,
+                    str_statement,
+                    effective_parameters,
+                    context,
+                    context.executemany,
+                )
+
+            if preserve_rowcount:
+                rowcount += imv_batch.current_batch_size
+
+        try:
+            context.post_exec()
+
+            if preserve_rowcount:
+                context._rowcount = rowcount  # type: ignore[attr-defined]
+
+            result = context._setup_result_proxy()
+
+        except BaseException as e:
+            self._handle_dbapi_exception(
+                e, str_statement, effective_parameters, cursor, context
+            )
+
+        return result
+
+    def _cursor_execute(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPISingleExecuteParams,
+        context: Optional[ExecutionContext] = None,
+    ) -> None:
+        """Execute a statement + params on the given cursor.
+
+        Adds appropriate logging and exception handling.
+
+        This method is used by DefaultDialect for special-case
+        executions, such as for sequences and column defaults.
+        The path of statement execution in the majority of cases
+        terminates at _execute_context().
+
+        """
+        if self._has_events or self.engine._has_events:
+            for fn in self.dispatch.before_cursor_execute:
+                statement, parameters = fn(
+                    self, cursor, statement, parameters, context, False
+                )
+
+        if self._echo:
+            self._log_info(statement)
+            self._log_info("[raw sql] %r", parameters)
+        try:
+            for fn in (
+                ()
+                if not self.dialect._has_events
+                else self.dialect.dispatch.do_execute
+            ):
+                if fn(cursor, statement, parameters, context):
+                    break
+            else:
+                self.dialect.do_execute(cursor, statement, parameters, context)
+        except BaseException as e:
+            self._handle_dbapi_exception(
+                e, statement, parameters, cursor, context
+            )
+
+        if self._has_events or self.engine._has_events:
+            self.dispatch.after_cursor_execute(
+                self, cursor, statement, parameters, context, False
+            )
+
+    def _safe_close_cursor(self, cursor: DBAPICursor) -> None:
+        """Close the given cursor, catching exceptions
+        and turning into log warnings.
+
+        """
+        try:
+            cursor.close()
+        except Exception:
+            # log the error through the connection pool's logger.
+            self.engine.pool.logger.error(
+                "Error closing cursor", exc_info=True
+            )
+
+    _reentrant_error = False
+    _is_disconnect = False
+
+    def _handle_dbapi_exception(
+        self,
+        e: BaseException,
+        statement: Optional[str],
+        parameters: Optional[_AnyExecuteParams],
+        cursor: Optional[DBAPICursor],
+        context: Optional[ExecutionContext],
+        is_sub_exec: bool = False,
+    ) -> NoReturn:
+        exc_info = sys.exc_info()
+
+        is_exit_exception = util.is_exit_exception(e)
+
+        if not self._is_disconnect:
+            self._is_disconnect = (
+                isinstance(e, self.dialect.loaded_dbapi.Error)
+                and not self.closed
+                and self.dialect.is_disconnect(
+                    e,
+                    self._dbapi_connection if not self.invalidated else None,
+                    cursor,
+                )
+            ) or (is_exit_exception and not self.closed)
+
+        invalidate_pool_on_disconnect = not is_exit_exception
+
+        ismulti: bool = (
+            not is_sub_exec and context.executemany
+            if context is not None
+            else False
+        )
+        if self._reentrant_error:
+            raise exc.DBAPIError.instance(
+                statement,
+                parameters,
+                e,
+                self.dialect.loaded_dbapi.Error,
+                hide_parameters=self.engine.hide_parameters,
+                dialect=self.dialect,
+                ismulti=ismulti,
+            ).with_traceback(exc_info[2]) from e
+        self._reentrant_error = True
+        try:
+            # non-DBAPI error - if we already got a context,
+            # or there's no string statement, don't wrap it
+            should_wrap = isinstance(e, self.dialect.loaded_dbapi.Error) or (
+                statement is not None
+                and context is None
+                and not is_exit_exception
+            )
+
+            if should_wrap:
+                sqlalchemy_exception = exc.DBAPIError.instance(
+                    statement,
+                    parameters,
+                    cast(Exception, e),
+                    self.dialect.loaded_dbapi.Error,
+                    hide_parameters=self.engine.hide_parameters,
+                    connection_invalidated=self._is_disconnect,
+                    dialect=self.dialect,
+                    ismulti=ismulti,
+                )
+            else:
+                sqlalchemy_exception = None
+
+            newraise = None
+
+            if (self.dialect._has_events) and not self._execution_options.get(
+                "skip_user_error_events", False
+            ):
+                ctx = ExceptionContextImpl(
+                    e,
+                    sqlalchemy_exception,
+                    self.engine,
+                    self.dialect,
+                    self,
+                    cursor,
+                    statement,
+                    parameters,
+                    context,
+                    self._is_disconnect,
+                    invalidate_pool_on_disconnect,
+                    False,
+                )
+
+                for fn in self.dialect.dispatch.handle_error:
+                    try:
+                        # handler returns an exception;
+                        # call next handler in a chain
+                        per_fn = fn(ctx)
+                        if per_fn is not None:
+                            ctx.chained_exception = newraise = per_fn
+                    except Exception as _raised:
+                        # handler raises an exception - stop processing
+                        newraise = _raised
+                        break
+
+                if self._is_disconnect != ctx.is_disconnect:
+                    self._is_disconnect = ctx.is_disconnect
+                    if sqlalchemy_exception:
+                        sqlalchemy_exception.connection_invalidated = (
+                            ctx.is_disconnect
+                        )
+
+                # set up potentially user-defined value for
+                # invalidate pool.
+                invalidate_pool_on_disconnect = (
+                    ctx.invalidate_pool_on_disconnect
+                )
+
+            if should_wrap and context:
+                context.handle_dbapi_exception(e)
+
+            if not self._is_disconnect:
+                if cursor:
+                    self._safe_close_cursor(cursor)
+                # "autorollback" was mostly relevant in 1.x series.
+                # It's very unlikely to reach here, as the connection
+                # does autobegin so when we are here, we are usually
+                # in an explicit / semi-explicit transaction.
+                # however we have a test which manufactures this
+                # scenario in any case using an event handler.
+                # test/engine/test_execute.py-> test_actual_autorollback
+                if not self.in_transaction():
+                    self._rollback_impl()
+
+            if newraise:
+                raise newraise.with_traceback(exc_info[2]) from e
+            elif should_wrap:
+                assert sqlalchemy_exception is not None
+                raise sqlalchemy_exception.with_traceback(exc_info[2]) from e
+            else:
+                assert exc_info[1] is not None
+                raise exc_info[1].with_traceback(exc_info[2])
+        finally:
+            del self._reentrant_error
+            if self._is_disconnect:
+                del self._is_disconnect
+                if not self.invalidated:
+                    dbapi_conn_wrapper = self._dbapi_connection
+                    assert dbapi_conn_wrapper is not None
+                    if invalidate_pool_on_disconnect:
+                        self.engine.pool._invalidate(dbapi_conn_wrapper, e)
+                    self.invalidate(e)
+
+    @classmethod
+    def _handle_dbapi_exception_noconnection(
+        cls,
+        e: BaseException,
+        dialect: Dialect,
+        engine: Optional[Engine] = None,
+        is_disconnect: Optional[bool] = None,
+        invalidate_pool_on_disconnect: bool = True,
+        is_pre_ping: bool = False,
+    ) -> NoReturn:
+        exc_info = sys.exc_info()
+
+        if is_disconnect is None:
+            is_disconnect = isinstance(
+                e, dialect.loaded_dbapi.Error
+            ) and dialect.is_disconnect(e, None, None)
+
+        should_wrap = isinstance(e, dialect.loaded_dbapi.Error)
+
+        if should_wrap:
+            sqlalchemy_exception = exc.DBAPIError.instance(
+                None,
+                None,
+                cast(Exception, e),
+                dialect.loaded_dbapi.Error,
+                hide_parameters=(
+                    engine.hide_parameters if engine is not None else False
+                ),
+                connection_invalidated=is_disconnect,
+                dialect=dialect,
+            )
+        else:
+            sqlalchemy_exception = None
+
+        newraise = None
+
+        if dialect._has_events:
+            ctx = ExceptionContextImpl(
+                e,
+                sqlalchemy_exception,
+                engine,
+                dialect,
+                None,
+                None,
+                None,
+                None,
+                None,
+                is_disconnect,
+                invalidate_pool_on_disconnect,
+                is_pre_ping,
+            )
+            for fn in dialect.dispatch.handle_error:
+                try:
+                    # handler returns an exception;
+                    # call next handler in a chain
+                    per_fn = fn(ctx)
+                    if per_fn is not None:
+                        ctx.chained_exception = newraise = per_fn
+                except Exception as _raised:
+                    # handler raises an exception - stop processing
+                    newraise = _raised
+                    break
+
+            if sqlalchemy_exception and is_disconnect != ctx.is_disconnect:
+                sqlalchemy_exception.connection_invalidated = is_disconnect = (
+                    ctx.is_disconnect
+                )
+
+        if newraise:
+            raise newraise.with_traceback(exc_info[2]) from e
+        elif should_wrap:
+            assert sqlalchemy_exception is not None
+            raise sqlalchemy_exception.with_traceback(exc_info[2]) from e
+        else:
+            assert exc_info[1] is not None
+            raise exc_info[1].with_traceback(exc_info[2])
+
+    def _run_ddl_visitor(
+        self,
+        visitorcallable: Type[Union[SchemaGenerator, SchemaDropper]],
+        element: SchemaItem,
+        **kwargs: Any,
+    ) -> None:
+        """run a DDL visitor.
+
+        This method is only here so that the MockConnection can change the
+        options given to the visitor so that "checkfirst" is skipped.
+
+        """
+        visitorcallable(self.dialect, self, **kwargs).traverse_single(element)
+
+
+class ExceptionContextImpl(ExceptionContext):
+    """Implement the :class:`.ExceptionContext` interface."""
+
+    __slots__ = (
+        "connection",
+        "engine",
+        "dialect",
+        "cursor",
+        "statement",
+        "parameters",
+        "original_exception",
+        "sqlalchemy_exception",
+        "chained_exception",
+        "execution_context",
+        "is_disconnect",
+        "invalidate_pool_on_disconnect",
+        "is_pre_ping",
+    )
+
+    def __init__(
+        self,
+        exception: BaseException,
+        sqlalchemy_exception: Optional[exc.StatementError],
+        engine: Optional[Engine],
+        dialect: Dialect,
+        connection: Optional[Connection],
+        cursor: Optional[DBAPICursor],
+        statement: Optional[str],
+        parameters: Optional[_DBAPIAnyExecuteParams],
+        context: Optional[ExecutionContext],
+        is_disconnect: bool,
+        invalidate_pool_on_disconnect: bool,
+        is_pre_ping: bool,
+    ):
+        self.engine = engine
+        self.dialect = dialect
+        self.connection = connection
+        self.sqlalchemy_exception = sqlalchemy_exception
+        self.original_exception = exception
+        self.execution_context = context
+        self.statement = statement
+        self.parameters = parameters
+        self.is_disconnect = is_disconnect
+        self.invalidate_pool_on_disconnect = invalidate_pool_on_disconnect
+        self.is_pre_ping = is_pre_ping
+
+
+class Transaction(TransactionalContext):
+    """Represent a database transaction in progress.
+
+    The :class:`.Transaction` object is procured by
+    calling the :meth:`_engine.Connection.begin` method of
+    :class:`_engine.Connection`::
+
+        from sqlalchemy import create_engine
+
+        engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test")
+        connection = engine.connect()
+        trans = connection.begin()
+        connection.execute(text("insert into x (a, b) values (1, 2)"))
+        trans.commit()
+
+    The object provides :meth:`.rollback` and :meth:`.commit`
+    methods in order to control transaction boundaries.  It
+    also implements a context manager interface so that
+    the Python ``with`` statement can be used with the
+    :meth:`_engine.Connection.begin` method::
+
+        with connection.begin():
+            connection.execute(text("insert into x (a, b) values (1, 2)"))
+
+    The Transaction object is **not** threadsafe.
+
+    .. seealso::
+
+        :meth:`_engine.Connection.begin`
+
+        :meth:`_engine.Connection.begin_twophase`
+
+        :meth:`_engine.Connection.begin_nested`
+
+    .. index::
+      single: thread safety; Transaction
+    """  # noqa
+
+    __slots__ = ()
+
+    _is_root: bool = False
+    is_active: bool
+    connection: Connection
+
+    def __init__(self, connection: Connection):
+        raise NotImplementedError()
+
+    @property
+    def _deactivated_from_connection(self) -> bool:
+        """True if this transaction is totally deactivated from the connection
+        and therefore can no longer affect its state.
+
+        """
+        raise NotImplementedError()
+
+    def _do_close(self) -> None:
+        raise NotImplementedError()
+
+    def _do_rollback(self) -> None:
+        raise NotImplementedError()
+
+    def _do_commit(self) -> None:
+        raise NotImplementedError()
+
+    @property
+    def is_valid(self) -> bool:
+        return self.is_active and not self.connection.invalidated
+
+    def close(self) -> None:
+        """Close this :class:`.Transaction`.
+
+        If this transaction is the base transaction in a begin/commit
+        nesting, the transaction will rollback().  Otherwise, the
+        method returns.
+
+        This is used to cancel a Transaction without affecting the scope of
+        an enclosing transaction.
+
+        """
+        try:
+            self._do_close()
+        finally:
+            assert not self.is_active
+
+    def rollback(self) -> None:
+        """Roll back this :class:`.Transaction`.
+
+        The implementation of this may vary based on the type of transaction in
+        use:
+
+        * For a simple database transaction (e.g. :class:`.RootTransaction`),
+          it corresponds to a ROLLBACK.
+
+        * For a :class:`.NestedTransaction`, it corresponds to a
+          "ROLLBACK TO SAVEPOINT" operation.
+
+        * For a :class:`.TwoPhaseTransaction`, DBAPI-specific methods for two
+          phase transactions may be used.
+
+
+        """
+        try:
+            self._do_rollback()
+        finally:
+            assert not self.is_active
+
+    def commit(self) -> None:
+        """Commit this :class:`.Transaction`.
+
+        The implementation of this may vary based on the type of transaction in
+        use:
+
+        * For a simple database transaction (e.g. :class:`.RootTransaction`),
+          it corresponds to a COMMIT.
+
+        * For a :class:`.NestedTransaction`, it corresponds to a
+          "RELEASE SAVEPOINT" operation.
+
+        * For a :class:`.TwoPhaseTransaction`, DBAPI-specific methods for two
+          phase transactions may be used.
+
+        """
+        try:
+            self._do_commit()
+        finally:
+            assert not self.is_active
+
+    def _get_subject(self) -> Connection:
+        return self.connection
+
+    def _transaction_is_active(self) -> bool:
+        return self.is_active
+
+    def _transaction_is_closed(self) -> bool:
+        return not self._deactivated_from_connection
+
+    def _rollback_can_be_called(self) -> bool:
+        # for RootTransaction / NestedTransaction, it's safe to call
+        # rollback() even if the transaction is deactive and no warnings
+        # will be emitted.  tested in
+        # test_transaction.py -> test_no_rollback_in_deactive(?:_savepoint)?
+        return True
+
+
+class RootTransaction(Transaction):
+    """Represent the "root" transaction on a :class:`_engine.Connection`.
+
+    This corresponds to the current "BEGIN/COMMIT/ROLLBACK" that's occurring
+    for the :class:`_engine.Connection`. The :class:`_engine.RootTransaction`
+    is created by calling upon the :meth:`_engine.Connection.begin` method, and
+    remains associated with the :class:`_engine.Connection` throughout its
+    active span. The current :class:`_engine.RootTransaction` in use is
+    accessible via the :attr:`_engine.Connection.get_transaction` method of
+    :class:`_engine.Connection`.
+
+    In :term:`2.0 style` use, the :class:`_engine.Connection` also employs
+    "autobegin" behavior that will create a new
+    :class:`_engine.RootTransaction` whenever a connection in a
+    non-transactional state is used to emit commands on the DBAPI connection.
+    The scope of the :class:`_engine.RootTransaction` in 2.0 style
+    use can be controlled using the :meth:`_engine.Connection.commit` and
+    :meth:`_engine.Connection.rollback` methods.
+
+
+    """
+
+    _is_root = True
+
+    __slots__ = ("connection", "is_active")
+
+    def __init__(self, connection: Connection):
+        assert connection._transaction is None
+        if connection._trans_context_manager:
+            TransactionalContext._trans_ctx_check(connection)
+        self.connection = connection
+        self._connection_begin_impl()
+        connection._transaction = self
+
+        self.is_active = True
+
+    def _deactivate_from_connection(self) -> None:
+        if self.is_active:
+            assert self.connection._transaction is self
+            self.is_active = False
+
+        elif self.connection._transaction is not self:
+            util.warn("transaction already deassociated from connection")
+
+    @property
+    def _deactivated_from_connection(self) -> bool:
+        return self.connection._transaction is not self
+
+    def _connection_begin_impl(self) -> None:
+        self.connection._begin_impl(self)
+
+    def _connection_rollback_impl(self) -> None:
+        self.connection._rollback_impl()
+
+    def _connection_commit_impl(self) -> None:
+        self.connection._commit_impl()
+
+    def _close_impl(self, try_deactivate: bool = False) -> None:
+        try:
+            if self.is_active:
+                self._connection_rollback_impl()
+
+            if self.connection._nested_transaction:
+                self.connection._nested_transaction._cancel()
+        finally:
+            if self.is_active or try_deactivate:
+                self._deactivate_from_connection()
+            if self.connection._transaction is self:
+                self.connection._transaction = None
+
+        assert not self.is_active
+        assert self.connection._transaction is not self
+
+    def _do_close(self) -> None:
+        self._close_impl()
+
+    def _do_rollback(self) -> None:
+        self._close_impl(try_deactivate=True)
+
+    def _do_commit(self) -> None:
+        if self.is_active:
+            assert self.connection._transaction is self
+
+            try:
+                self._connection_commit_impl()
+            finally:
+                # whether or not commit succeeds, cancel any
+                # nested transactions, make this transaction "inactive"
+                # and remove it as a reset agent
+                if self.connection._nested_transaction:
+                    self.connection._nested_transaction._cancel()
+
+                self._deactivate_from_connection()
+
+            # ...however only remove as the connection's current transaction
+            # if commit succeeded.  otherwise it stays on so that a rollback
+            # needs to occur.
+            self.connection._transaction = None
+        else:
+            if self.connection._transaction is self:
+                self.connection._invalid_transaction()
+            else:
+                raise exc.InvalidRequestError("This transaction is inactive")
+
+        assert not self.is_active
+        assert self.connection._transaction is not self
+
+
+class NestedTransaction(Transaction):
+    """Represent a 'nested', or SAVEPOINT transaction.
+
+    The :class:`.NestedTransaction` object is created by calling the
+    :meth:`_engine.Connection.begin_nested` method of
+    :class:`_engine.Connection`.
+
+    When using :class:`.NestedTransaction`, the semantics of "begin" /
+    "commit" / "rollback" are as follows:
+
+    * the "begin" operation corresponds to the "BEGIN SAVEPOINT" command, where
+      the savepoint is given an explicit name that is part of the state
+      of this object.
+
+    * The :meth:`.NestedTransaction.commit` method corresponds to a
+      "RELEASE SAVEPOINT" operation, using the savepoint identifier associated
+      with this :class:`.NestedTransaction`.
+
+    * The :meth:`.NestedTransaction.rollback` method corresponds to a
+      "ROLLBACK TO SAVEPOINT" operation, using the savepoint identifier
+      associated with this :class:`.NestedTransaction`.
+
+    The rationale for mimicking the semantics of an outer transaction in
+    terms of savepoints so that code may deal with a "savepoint" transaction
+    and an "outer" transaction in an agnostic way.
+
+    .. seealso::
+
+        :ref:`session_begin_nested` - ORM version of the SAVEPOINT API.
+
+    """
+
+    __slots__ = ("connection", "is_active", "_savepoint", "_previous_nested")
+
+    _savepoint: str
+
+    def __init__(self, connection: Connection):
+        assert connection._transaction is not None
+        if connection._trans_context_manager:
+            TransactionalContext._trans_ctx_check(connection)
+        self.connection = connection
+        self._savepoint = self.connection._savepoint_impl()
+        self.is_active = True
+        self._previous_nested = connection._nested_transaction
+        connection._nested_transaction = self
+
+    def _deactivate_from_connection(self, warn: bool = True) -> None:
+        if self.connection._nested_transaction is self:
+            self.connection._nested_transaction = self._previous_nested
+        elif warn:
+            util.warn(
+                "nested transaction already deassociated from connection"
+            )
+
+    @property
+    def _deactivated_from_connection(self) -> bool:
+        return self.connection._nested_transaction is not self
+
+    def _cancel(self) -> None:
+        # called by RootTransaction when the outer transaction is
+        # committed, rolled back, or closed to cancel all savepoints
+        # without any action being taken
+        self.is_active = False
+        self._deactivate_from_connection()
+        if self._previous_nested:
+            self._previous_nested._cancel()
+
+    def _close_impl(
+        self, deactivate_from_connection: bool, warn_already_deactive: bool
+    ) -> None:
+        try:
+            if (
+                self.is_active
+                and self.connection._transaction
+                and self.connection._transaction.is_active
+            ):
+                self.connection._rollback_to_savepoint_impl(self._savepoint)
+        finally:
+            self.is_active = False
+
+            if deactivate_from_connection:
+                self._deactivate_from_connection(warn=warn_already_deactive)
+
+        assert not self.is_active
+        if deactivate_from_connection:
+            assert self.connection._nested_transaction is not self
+
+    def _do_close(self) -> None:
+        self._close_impl(True, False)
+
+    def _do_rollback(self) -> None:
+        self._close_impl(True, True)
+
+    def _do_commit(self) -> None:
+        if self.is_active:
+            try:
+                self.connection._release_savepoint_impl(self._savepoint)
+            finally:
+                # nested trans becomes inactive on failed release
+                # unconditionally.  this prevents it from trying to
+                # emit SQL when it rolls back.
+                self.is_active = False
+
+            # but only de-associate from connection if it succeeded
+            self._deactivate_from_connection()
+        else:
+            if self.connection._nested_transaction is self:
+                self.connection._invalid_transaction()
+            else:
+                raise exc.InvalidRequestError(
+                    "This nested transaction is inactive"
+                )
+
+
+class TwoPhaseTransaction(RootTransaction):
+    """Represent a two-phase transaction.
+
+    A new :class:`.TwoPhaseTransaction` object may be procured
+    using the :meth:`_engine.Connection.begin_twophase` method.
+
+    The interface is the same as that of :class:`.Transaction`
+    with the addition of the :meth:`prepare` method.
+
+    """
+
+    __slots__ = ("xid", "_is_prepared")
+
+    xid: Any
+
+    def __init__(self, connection: Connection, xid: Any):
+        self._is_prepared = False
+        self.xid = xid
+        super().__init__(connection)
+
+    def prepare(self) -> None:
+        """Prepare this :class:`.TwoPhaseTransaction`.
+
+        After a PREPARE, the transaction can be committed.
+
+        """
+        if not self.is_active:
+            raise exc.InvalidRequestError("This transaction is inactive")
+        self.connection._prepare_twophase_impl(self.xid)
+        self._is_prepared = True
+
+    def _connection_begin_impl(self) -> None:
+        self.connection._begin_twophase_impl(self)
+
+    def _connection_rollback_impl(self) -> None:
+        self.connection._rollback_twophase_impl(self.xid, self._is_prepared)
+
+    def _connection_commit_impl(self) -> None:
+        self.connection._commit_twophase_impl(self.xid, self._is_prepared)
+
+
+class Engine(
+    ConnectionEventsTarget, log.Identified, inspection.Inspectable["Inspector"]
+):
+    """
+    Connects a :class:`~sqlalchemy.pool.Pool` and
+    :class:`~sqlalchemy.engine.interfaces.Dialect` together to provide a
+    source of database connectivity and behavior.
+
+    An :class:`_engine.Engine` object is instantiated publicly using the
+    :func:`~sqlalchemy.create_engine` function.
+
+    .. seealso::
+
+        :doc:`/core/engines`
+
+        :ref:`connections_toplevel`
+
+    """
+
+    dispatch: dispatcher[ConnectionEventsTarget]
+
+    _compiled_cache: Optional[CompiledCacheType]
+
+    _execution_options: _ExecuteOptions = _EMPTY_EXECUTION_OPTS
+    _has_events: bool = False
+    _connection_cls: Type[Connection] = Connection
+    _sqla_logger_namespace: str = "sqlalchemy.engine.Engine"
+    _is_future: bool = False
+
+    _schema_translate_map: Optional[SchemaTranslateMapType] = None
+    _option_cls: Type[OptionEngine]
+
+    dialect: Dialect
+    pool: Pool
+    url: URL
+    hide_parameters: bool
+
+    def __init__(
+        self,
+        pool: Pool,
+        dialect: Dialect,
+        url: URL,
+        logging_name: Optional[str] = None,
+        echo: Optional[_EchoFlagType] = None,
+        query_cache_size: int = 500,
+        execution_options: Optional[Mapping[str, Any]] = None,
+        hide_parameters: bool = False,
+    ):
+        self.pool = pool
+        self.url = url
+        self.dialect = dialect
+        if logging_name:
+            self.logging_name = logging_name
+        self.echo = echo
+        self.hide_parameters = hide_parameters
+        if query_cache_size != 0:
+            self._compiled_cache = util.LRUCache(
+                query_cache_size, size_alert=self._lru_size_alert
+            )
+        else:
+            self._compiled_cache = None
+        log.instance_logger(self, echoflag=echo)
+        if execution_options:
+            self.update_execution_options(**execution_options)
+
+    def _lru_size_alert(self, cache: util.LRUCache[Any, Any]) -> None:
+        if self._should_log_info():
+            self.logger.info(
+                "Compiled cache size pruning from %d items to %d.  "
+                "Increase cache size to reduce the frequency of pruning.",
+                len(cache),
+                cache.capacity,
+            )
+
+    @property
+    def engine(self) -> Engine:
+        """Returns this :class:`.Engine`.
+
+        Used for legacy schemes that accept :class:`.Connection` /
+        :class:`.Engine` objects within the same variable.
+
+        """
+        return self
+
+    def clear_compiled_cache(self) -> None:
+        """Clear the compiled cache associated with the dialect.
+
+        This applies **only** to the built-in cache that is established
+        via the :paramref:`_engine.create_engine.query_cache_size` parameter.
+        It will not impact any dictionary caches that were passed via the
+        :paramref:`.Connection.execution_options.compiled_cache` parameter.
+
+        .. versionadded:: 1.4
+
+        """
+        if self._compiled_cache:
+            self._compiled_cache.clear()
+
+    def update_execution_options(self, **opt: Any) -> None:
+        r"""Update the default execution_options dictionary
+        of this :class:`_engine.Engine`.
+
+        The given keys/values in \**opt are added to the
+        default execution options that will be used for
+        all connections.  The initial contents of this dictionary
+        can be sent via the ``execution_options`` parameter
+        to :func:`_sa.create_engine`.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.execution_options`
+
+            :meth:`_engine.Engine.execution_options`
+
+        """
+        self.dispatch.set_engine_execution_options(self, opt)
+        self._execution_options = self._execution_options.union(opt)
+        self.dialect.set_engine_execution_options(self, opt)
+
+    @overload
+    def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        **opt: Any,
+    ) -> OptionEngine: ...
+
+    @overload
+    def execution_options(self, **opt: Any) -> OptionEngine: ...
+
+    def execution_options(self, **opt: Any) -> OptionEngine:
+        """Return a new :class:`_engine.Engine` that will provide
+        :class:`_engine.Connection` objects with the given execution options.
+
+        The returned :class:`_engine.Engine` remains related to the original
+        :class:`_engine.Engine` in that it shares the same connection pool and
+        other state:
+
+        * The :class:`_pool.Pool` used by the new :class:`_engine.Engine`
+          is the
+          same instance.  The :meth:`_engine.Engine.dispose`
+          method will replace
+          the connection pool instance for the parent engine as well
+          as this one.
+        * Event listeners are "cascaded" - meaning, the new
+          :class:`_engine.Engine`
+          inherits the events of the parent, and new events can be associated
+          with the new :class:`_engine.Engine` individually.
+        * The logging configuration and logging_name is copied from the parent
+          :class:`_engine.Engine`.
+
+        The intent of the :meth:`_engine.Engine.execution_options` method is
+        to implement schemes where multiple :class:`_engine.Engine`
+        objects refer to the same connection pool, but are differentiated
+        by options that affect some execution-level behavior for each
+        engine.    One such example is breaking into separate "reader" and
+        "writer" :class:`_engine.Engine` instances, where one
+        :class:`_engine.Engine`
+        has a lower :term:`isolation level` setting configured or is even
+        transaction-disabled using "autocommit".  An example of this
+        configuration is at :ref:`dbapi_autocommit_multiple`.
+
+        Another example is one that
+        uses a custom option ``shard_id`` which is consumed by an event
+        to change the current schema on a database connection::
+
+            from sqlalchemy import event
+            from sqlalchemy.engine import Engine
+
+            primary_engine = create_engine("mysql+mysqldb://")
+            shard1 = primary_engine.execution_options(shard_id="shard1")
+            shard2 = primary_engine.execution_options(shard_id="shard2")
+
+            shards = {"default": "base", "shard_1": "db1", "shard_2": "db2"}
+
+
+            @event.listens_for(Engine, "before_cursor_execute")
+            def _switch_shard(conn, cursor, stmt, params, context, executemany):
+                shard_id = conn.get_execution_options().get("shard_id", "default")
+                current_shard = conn.info.get("current_shard", None)
+
+                if current_shard != shard_id:
+                    cursor.execute("use %s" % shards[shard_id])
+                    conn.info["current_shard"] = shard_id
+
+        The above recipe illustrates two :class:`_engine.Engine` objects that
+        will each serve as factories for :class:`_engine.Connection` objects
+        that have pre-established "shard_id" execution options present. A
+        :meth:`_events.ConnectionEvents.before_cursor_execute` event handler
+        then interprets this execution option to emit a MySQL ``use`` statement
+        to switch databases before a statement execution, while at the same
+        time keeping track of which database we've established using the
+        :attr:`_engine.Connection.info` dictionary.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.execution_options`
+            - update execution options
+            on a :class:`_engine.Connection` object.
+
+            :meth:`_engine.Engine.update_execution_options`
+            - update the execution
+            options for a given :class:`_engine.Engine` in place.
+
+            :meth:`_engine.Engine.get_execution_options`
+
+
+        """  # noqa: E501
+        return self._option_cls(self, opt)
+
+    def get_execution_options(self) -> _ExecuteOptions:
+        """Get the non-SQL options which will take effect during execution.
+
+        .. versionadded: 1.3
+
+        .. seealso::
+
+            :meth:`_engine.Engine.execution_options`
+        """
+        return self._execution_options
+
+    @property
+    def name(self) -> str:
+        """String name of the :class:`~sqlalchemy.engine.interfaces.Dialect`
+        in use by this :class:`Engine`.
+
+        """
+
+        return self.dialect.name
+
+    @property
+    def driver(self) -> str:
+        """Driver name of the :class:`~sqlalchemy.engine.interfaces.Dialect`
+        in use by this :class:`Engine`.
+
+        """
+
+        return self.dialect.driver
+
+    echo = log.echo_property()
+
+    def __repr__(self) -> str:
+        return "Engine(%r)" % (self.url,)
+
+    def dispose(self, close: bool = True) -> None:
+        """Dispose of the connection pool used by this
+        :class:`_engine.Engine`.
+
+        A new connection pool is created immediately after the old one has been
+        disposed. The previous connection pool is disposed either actively, by
+        closing out all currently checked-in connections in that pool, or
+        passively, by losing references to it but otherwise not closing any
+        connections. The latter strategy is more appropriate for an initializer
+        in a forked Python process.
+
+        :param close: if left at its default of ``True``, has the
+         effect of fully closing all **currently checked in**
+         database connections.  Connections that are still checked out
+         will **not** be closed, however they will no longer be associated
+         with this :class:`_engine.Engine`,
+         so when they are closed individually, eventually the
+         :class:`_pool.Pool` which they are associated with will
+         be garbage collected and they will be closed out fully, if
+         not already closed on checkin.
+
+         If set to ``False``, the previous connection pool is de-referenced,
+         and otherwise not touched in any way.
+
+        .. versionadded:: 1.4.33  Added the :paramref:`.Engine.dispose.close`
+            parameter to allow the replacement of a connection pool in a child
+            process without interfering with the connections used by the parent
+            process.
+
+
+        .. seealso::
+
+            :ref:`engine_disposal`
+
+            :ref:`pooling_multiprocessing`
+
+        """
+        if close:
+            self.pool.dispose()
+        self.pool = self.pool.recreate()
+        self.dispatch.engine_disposed(self)
+
+    @contextlib.contextmanager
+    def _optional_conn_ctx_manager(
+        self, connection: Optional[Connection] = None
+    ) -> Iterator[Connection]:
+        if connection is None:
+            with self.connect() as conn:
+                yield conn
+        else:
+            yield connection
+
+    @contextlib.contextmanager
+    def begin(self) -> Iterator[Connection]:
+        """Return a context manager delivering a :class:`_engine.Connection`
+        with a :class:`.Transaction` established.
+
+        E.g.::
+
+            with engine.begin() as conn:
+                conn.execute(text("insert into table (x, y, z) values (1, 2, 3)"))
+                conn.execute(text("my_special_procedure(5)"))
+
+        Upon successful operation, the :class:`.Transaction`
+        is committed.  If an error is raised, the :class:`.Transaction`
+        is rolled back.
+
+        .. seealso::
+
+            :meth:`_engine.Engine.connect` - procure a
+            :class:`_engine.Connection` from
+            an :class:`_engine.Engine`.
+
+            :meth:`_engine.Connection.begin` - start a :class:`.Transaction`
+            for a particular :class:`_engine.Connection`.
+
+        """  # noqa: E501
+        with self.connect() as conn:
+            with conn.begin():
+                yield conn
+
+    def _run_ddl_visitor(
+        self,
+        visitorcallable: Type[Union[SchemaGenerator, SchemaDropper]],
+        element: SchemaItem,
+        **kwargs: Any,
+    ) -> None:
+        with self.begin() as conn:
+            conn._run_ddl_visitor(visitorcallable, element, **kwargs)
+
+    def connect(self) -> Connection:
+        """Return a new :class:`_engine.Connection` object.
+
+        The :class:`_engine.Connection` acts as a Python context manager, so
+        the typical use of this method looks like::
+
+            with engine.connect() as connection:
+                connection.execute(text("insert into table values ('foo')"))
+                connection.commit()
+
+        Where above, after the block is completed, the connection is "closed"
+        and its underlying DBAPI resources are returned to the connection pool.
+        This also has the effect of rolling back any transaction that
+        was explicitly begun or was begun via autobegin, and will
+        emit the :meth:`_events.ConnectionEvents.rollback` event if one was
+        started and is still in progress.
+
+        .. seealso::
+
+            :meth:`_engine.Engine.begin`
+
+        """
+
+        return self._connection_cls(self)
+
+    def raw_connection(self) -> PoolProxiedConnection:
+        """Return a "raw" DBAPI connection from the connection pool.
+
+        The returned object is a proxied version of the DBAPI
+        connection object used by the underlying driver in use.
+        The object will have all the same behavior as the real DBAPI
+        connection, except that its ``close()`` method will result in the
+        connection being returned to the pool, rather than being closed
+        for real.
+
+        This method provides direct DBAPI connection access for
+        special situations when the API provided by
+        :class:`_engine.Connection`
+        is not needed.   When a :class:`_engine.Connection` object is already
+        present, the DBAPI connection is available using
+        the :attr:`_engine.Connection.connection` accessor.
+
+        .. seealso::
+
+            :ref:`dbapi_connections`
+
+        """
+        return self.pool.connect()
+
+
+class OptionEngineMixin(log.Identified):
+    _sa_propagate_class_events = False
+
+    dispatch: dispatcher[ConnectionEventsTarget]
+    _compiled_cache: Optional[CompiledCacheType]
+    dialect: Dialect
+    pool: Pool
+    url: URL
+    hide_parameters: bool
+    echo: log.echo_property
+
+    def __init__(
+        self, proxied: Engine, execution_options: CoreExecuteOptionsParameter
+    ):
+        self._proxied = proxied
+        self.url = proxied.url
+        self.dialect = proxied.dialect
+        self.logging_name = proxied.logging_name
+        self.echo = proxied.echo
+        self._compiled_cache = proxied._compiled_cache
+        self.hide_parameters = proxied.hide_parameters
+        log.instance_logger(self, echoflag=self.echo)
+
+        # note: this will propagate events that are assigned to the parent
+        # engine after this OptionEngine is created.   Since we share
+        # the events of the parent we also disallow class-level events
+        # to apply to the OptionEngine class directly.
+        #
+        # the other way this can work would be to transfer existing
+        # events only, using:
+        # self.dispatch._update(proxied.dispatch)
+        #
+        # that might be more appropriate however it would be a behavioral
+        # change for logic that assigns events to the parent engine and
+        # would like it to take effect for the already-created sub-engine.
+        self.dispatch = self.dispatch._join(proxied.dispatch)
+
+        self._execution_options = proxied._execution_options
+        self.update_execution_options(**execution_options)
+
+    def update_execution_options(self, **opt: Any) -> None:
+        raise NotImplementedError()
+
+    if not typing.TYPE_CHECKING:
+        # https://github.com/python/typing/discussions/1095
+
+        @property
+        def pool(self) -> Pool:
+            return self._proxied.pool
+
+        @pool.setter
+        def pool(self, pool: Pool) -> None:
+            self._proxied.pool = pool
+
+        @property
+        def _has_events(self) -> bool:
+            return self._proxied._has_events or self.__dict__.get(
+                "_has_events", False
+            )
+
+        @_has_events.setter
+        def _has_events(self, value: bool) -> None:
+            self.__dict__["_has_events"] = value
+
+
+class OptionEngine(OptionEngineMixin, Engine):
+    def update_execution_options(self, **opt: Any) -> None:
+        Engine.update_execution_options(self, **opt)
+
+
+Engine._option_cls = OptionEngine
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/characteristics.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/characteristics.py
new file mode 100644
index 00000000..322c28b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/characteristics.py
@@ -0,0 +1,155 @@
+# engine/characteristics.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import abc
+import typing
+from typing import Any
+from typing import ClassVar
+
+if typing.TYPE_CHECKING:
+    from .base import Connection
+    from .interfaces import DBAPIConnection
+    from .interfaces import Dialect
+
+
+class ConnectionCharacteristic(abc.ABC):
+    """An abstract base for an object that can set, get and reset a
+    per-connection characteristic, typically one that gets reset when the
+    connection is returned to the connection pool.
+
+    transaction isolation is the canonical example, and the
+    ``IsolationLevelCharacteristic`` implementation provides this for the
+    ``DefaultDialect``.
+
+    The ``ConnectionCharacteristic`` class should call upon the ``Dialect`` for
+    the implementation of each method.   The object exists strictly to serve as
+    a dialect visitor that can be placed into the
+    ``DefaultDialect.connection_characteristics`` dictionary where it will take
+    effect for calls to :meth:`_engine.Connection.execution_options` and
+    related APIs.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    transactional: ClassVar[bool] = False
+
+    @abc.abstractmethod
+    def reset_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> None:
+        """Reset the characteristic on the DBAPI connection to its default
+        value."""
+
+    @abc.abstractmethod
+    def set_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection, value: Any
+    ) -> None:
+        """set characteristic on the DBAPI connection to a given value."""
+
+    def set_connection_characteristic(
+        self,
+        dialect: Dialect,
+        conn: Connection,
+        dbapi_conn: DBAPIConnection,
+        value: Any,
+    ) -> None:
+        """set characteristic on the :class:`_engine.Connection` to a given
+        value.
+
+        .. versionadded:: 2.0.30 - added to support elements that are local
+           to the :class:`_engine.Connection` itself.
+
+        """
+        self.set_characteristic(dialect, dbapi_conn, value)
+
+    @abc.abstractmethod
+    def get_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> Any:
+        """Given a DBAPI connection, get the current value of the
+        characteristic.
+
+        """
+
+    def get_connection_characteristic(
+        self, dialect: Dialect, conn: Connection, dbapi_conn: DBAPIConnection
+    ) -> Any:
+        """Given a :class:`_engine.Connection`, get the current value of the
+        characteristic.
+
+        .. versionadded:: 2.0.30 - added to support elements that are local
+           to the :class:`_engine.Connection` itself.
+
+        """
+        return self.get_characteristic(dialect, dbapi_conn)
+
+
+class IsolationLevelCharacteristic(ConnectionCharacteristic):
+    """Manage the isolation level on a DBAPI connection"""
+
+    transactional: ClassVar[bool] = True
+
+    def reset_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> None:
+        dialect.reset_isolation_level(dbapi_conn)
+
+    def set_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection, value: Any
+    ) -> None:
+        dialect._assert_and_set_isolation_level(dbapi_conn, value)
+
+    def get_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> Any:
+        return dialect.get_isolation_level(dbapi_conn)
+
+
+class LoggingTokenCharacteristic(ConnectionCharacteristic):
+    """Manage the 'logging_token' option of a :class:`_engine.Connection`.
+
+    .. versionadded:: 2.0.30
+
+    """
+
+    transactional: ClassVar[bool] = False
+
+    def reset_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> None:
+        pass
+
+    def set_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection, value: Any
+    ) -> None:
+        raise NotImplementedError()
+
+    def set_connection_characteristic(
+        self,
+        dialect: Dialect,
+        conn: Connection,
+        dbapi_conn: DBAPIConnection,
+        value: Any,
+    ) -> None:
+        if value:
+            conn._message_formatter = lambda msg: "[%s] %s" % (value, msg)
+        else:
+            del conn._message_formatter
+
+    def get_characteristic(
+        self, dialect: Dialect, dbapi_conn: DBAPIConnection
+    ) -> Any:
+        raise NotImplementedError()
+
+    def get_connection_characteristic(
+        self, dialect: Dialect, conn: Connection, dbapi_conn: DBAPIConnection
+    ) -> Any:
+        return conn._execution_options.get("logging_token", None)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/create.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/create.py
new file mode 100644
index 00000000..920f620b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/create.py
@@ -0,0 +1,878 @@
+# engine/create.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import inspect
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Type
+from typing import Union
+
+from . import base
+from . import url as _url
+from .interfaces import DBAPIConnection
+from .mock import create_mock_engine
+from .. import event
+from .. import exc
+from .. import util
+from ..pool import _AdhocProxiedConnection
+from ..pool import ConnectionPoolEntry
+from ..sql import compiler
+from ..util import immutabledict
+
+if typing.TYPE_CHECKING:
+    from .base import Engine
+    from .interfaces import _ExecuteOptions
+    from .interfaces import _ParamStyle
+    from .interfaces import IsolationLevel
+    from .url import URL
+    from ..log import _EchoFlagType
+    from ..pool import _CreatorFnType
+    from ..pool import _CreatorWRecFnType
+    from ..pool import _ResetStyleArgType
+    from ..pool import Pool
+    from ..util.typing import Literal
+
+
+@overload
+def create_engine(
+    url: Union[str, URL],
+    *,
+    connect_args: Dict[Any, Any] = ...,
+    convert_unicode: bool = ...,
+    creator: Union[_CreatorFnType, _CreatorWRecFnType] = ...,
+    echo: _EchoFlagType = ...,
+    echo_pool: _EchoFlagType = ...,
+    enable_from_linting: bool = ...,
+    execution_options: _ExecuteOptions = ...,
+    future: Literal[True],
+    hide_parameters: bool = ...,
+    implicit_returning: Literal[True] = ...,
+    insertmanyvalues_page_size: int = ...,
+    isolation_level: IsolationLevel = ...,
+    json_deserializer: Callable[..., Any] = ...,
+    json_serializer: Callable[..., Any] = ...,
+    label_length: Optional[int] = ...,
+    logging_name: str = ...,
+    max_identifier_length: Optional[int] = ...,
+    max_overflow: int = ...,
+    module: Optional[Any] = ...,
+    paramstyle: Optional[_ParamStyle] = ...,
+    pool: Optional[Pool] = ...,
+    poolclass: Optional[Type[Pool]] = ...,
+    pool_logging_name: str = ...,
+    pool_pre_ping: bool = ...,
+    pool_size: int = ...,
+    pool_recycle: int = ...,
+    pool_reset_on_return: Optional[_ResetStyleArgType] = ...,
+    pool_timeout: float = ...,
+    pool_use_lifo: bool = ...,
+    plugins: List[str] = ...,
+    query_cache_size: int = ...,
+    use_insertmanyvalues: bool = ...,
+    **kwargs: Any,
+) -> Engine: ...
+
+
+@overload
+def create_engine(url: Union[str, URL], **kwargs: Any) -> Engine: ...
+
+
+@util.deprecated_params(
+    strategy=(
+        "1.4",
+        "The :paramref:`_sa.create_engine.strategy` keyword is deprecated, "
+        "and the only argument accepted is 'mock'; please use "
+        ":func:`.create_mock_engine` going forward.  For general "
+        "customization of create_engine which may have been accomplished "
+        "using strategies, see :class:`.CreateEnginePlugin`.",
+    ),
+    empty_in_strategy=(
+        "1.4",
+        "The :paramref:`_sa.create_engine.empty_in_strategy` keyword is "
+        "deprecated, and no longer has any effect.  All IN expressions "
+        "are now rendered using "
+        'the "expanding parameter" strategy which renders a set of bound'
+        'expressions, or an "empty set" SELECT, at statement execution'
+        "time.",
+    ),
+    implicit_returning=(
+        "2.0",
+        "The :paramref:`_sa.create_engine.implicit_returning` parameter "
+        "is deprecated and will be removed in a future release. ",
+    ),
+)
+def create_engine(url: Union[str, _url.URL], **kwargs: Any) -> Engine:
+    """Create a new :class:`_engine.Engine` instance.
+
+    The standard calling form is to send the :ref:`URL <database_urls>` as the
+    first positional argument, usually a string
+    that indicates database dialect and connection arguments::
+
+        engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test")
+
+    .. note::
+
+        Please review :ref:`database_urls` for general guidelines in composing
+        URL strings.  In particular, special characters, such as those often
+        part of passwords, must be URL encoded to be properly parsed.
+
+    Additional keyword arguments may then follow it which
+    establish various options on the resulting :class:`_engine.Engine`
+    and its underlying :class:`.Dialect` and :class:`_pool.Pool`
+    constructs::
+
+        engine = create_engine(
+            "mysql+mysqldb://scott:tiger@hostname/dbname",
+            pool_recycle=3600,
+            echo=True,
+        )
+
+    The string form of the URL is
+    ``dialect[+driver]://user:password@host/dbname[?key=value..]``, where
+    ``dialect`` is a database name such as ``mysql``, ``oracle``,
+    ``postgresql``, etc., and ``driver`` the name of a DBAPI, such as
+    ``psycopg2``, ``pyodbc``, ``cx_oracle``, etc.  Alternatively,
+    the URL can be an instance of :class:`~sqlalchemy.engine.url.URL`.
+
+    ``**kwargs`` takes a wide variety of options which are routed
+    towards their appropriate components.  Arguments may be specific to
+    the :class:`_engine.Engine`, the underlying :class:`.Dialect`,
+    as well as the
+    :class:`_pool.Pool`.  Specific dialects also accept keyword arguments that
+    are unique to that dialect.   Here, we describe the parameters
+    that are common to most :func:`_sa.create_engine()` usage.
+
+    Once established, the newly resulting :class:`_engine.Engine` will
+    request a connection from the underlying :class:`_pool.Pool` once
+    :meth:`_engine.Engine.connect` is called, or a method which depends on it
+    such as :meth:`_engine.Engine.execute` is invoked.   The
+    :class:`_pool.Pool` in turn
+    will establish the first actual DBAPI connection when this request
+    is received.   The :func:`_sa.create_engine` call itself does **not**
+    establish any actual DBAPI connections directly.
+
+    .. seealso::
+
+        :doc:`/core/engines`
+
+        :doc:`/dialects/index`
+
+        :ref:`connections_toplevel`
+
+    :param connect_args: a dictionary of options which will be
+        passed directly to the DBAPI's ``connect()`` method as
+        additional keyword arguments.  See the example
+        at :ref:`custom_dbapi_args`.
+
+    :param creator: a callable which returns a DBAPI connection.
+        This creation function will be passed to the underlying
+        connection pool and will be used to create all new database
+        connections. Usage of this function causes connection
+        parameters specified in the URL argument to be bypassed.
+
+        This hook is not as flexible as the newer
+        :meth:`_events.DialectEvents.do_connect` hook which allows complete
+        control over how a connection is made to the database, given the full
+        set of URL arguments and state beforehand.
+
+        .. seealso::
+
+            :meth:`_events.DialectEvents.do_connect` - event hook that allows
+            full control over DBAPI connection mechanics.
+
+            :ref:`custom_dbapi_args`
+
+    :param echo=False: if True, the Engine will log all statements
+        as well as a ``repr()`` of their parameter lists to the default log
+        handler, which defaults to ``sys.stdout`` for output.   If set to the
+        string ``"debug"``, result rows will be printed to the standard output
+        as well. The ``echo`` attribute of ``Engine`` can be modified at any
+        time to turn logging on and off; direct control of logging is also
+        available using the standard Python ``logging`` module.
+
+        .. seealso::
+
+            :ref:`dbengine_logging` - further detail on how to configure
+            logging.
+
+
+    :param echo_pool=False: if True, the connection pool will log
+        informational output such as when connections are invalidated
+        as well as when connections are recycled to the default log handler,
+        which defaults to ``sys.stdout`` for output.   If set to the string
+        ``"debug"``, the logging will include pool checkouts and checkins.
+        Direct control of logging is also available using the standard Python
+        ``logging`` module.
+
+        .. seealso::
+
+            :ref:`dbengine_logging` - further detail on how to configure
+            logging.
+
+
+    :param empty_in_strategy:   No longer used; SQLAlchemy now uses
+        "empty set" behavior for IN in all cases.
+
+    :param enable_from_linting: defaults to True.  Will emit a warning
+        if a given SELECT statement is found to have un-linked FROM elements
+        which would cause a cartesian product.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :ref:`change_4737`
+
+    :param execution_options: Dictionary execution options which will
+        be applied to all connections.  See
+        :meth:`~sqlalchemy.engine.Connection.execution_options`
+
+    :param future: Use the 2.0 style :class:`_engine.Engine` and
+        :class:`_engine.Connection` API.
+
+        As of SQLAlchemy 2.0, this parameter is present for backwards
+        compatibility only and must remain at its default value of ``True``.
+
+        The :paramref:`_sa.create_engine.future` parameter will be
+        deprecated in a subsequent 2.x release and eventually removed.
+
+        .. versionadded:: 1.4
+
+        .. versionchanged:: 2.0 All :class:`_engine.Engine` objects are
+           "future" style engines and there is no longer a ``future=False``
+           mode of operation.
+
+        .. seealso::
+
+            :ref:`migration_20_toplevel`
+
+    :param hide_parameters: Boolean, when set to True, SQL statement parameters
+        will not be displayed in INFO logging nor will they be formatted into
+        the string representation of :class:`.StatementError` objects.
+
+        .. versionadded:: 1.3.8
+
+        .. seealso::
+
+            :ref:`dbengine_logging` - further detail on how to configure
+            logging.
+
+    :param implicit_returning=True:  Legacy parameter that may only be set
+        to True. In SQLAlchemy 2.0, this parameter does nothing. In order to
+        disable "implicit returning" for statements invoked by the ORM,
+        configure this on a per-table basis using the
+        :paramref:`.Table.implicit_returning` parameter.
+
+
+    :param insertmanyvalues_page_size: number of rows to format into an
+     INSERT statement when the statement uses "insertmanyvalues" mode, which is
+     a paged form of bulk insert that is used for many backends when using
+     :term:`executemany` execution typically in conjunction with RETURNING.
+     Defaults to 1000, but may also be subject to dialect-specific limiting
+     factors which may override this value on a per-statement basis.
+
+     .. versionadded:: 2.0
+
+     .. seealso::
+
+        :ref:`engine_insertmanyvalues`
+
+        :ref:`engine_insertmanyvalues_page_size`
+
+        :paramref:`_engine.Connection.execution_options.insertmanyvalues_page_size`
+
+    :param isolation_level: optional string name of an isolation level
+        which will be set on all new connections unconditionally.
+        Isolation levels are typically some subset of the string names
+        ``"SERIALIZABLE"``, ``"REPEATABLE READ"``,
+        ``"READ COMMITTED"``, ``"READ UNCOMMITTED"`` and ``"AUTOCOMMIT"``
+        based on backend.
+
+        The :paramref:`_sa.create_engine.isolation_level` parameter is
+        in contrast to the
+        :paramref:`.Connection.execution_options.isolation_level`
+        execution option, which may be set on an individual
+        :class:`.Connection`, as well as the same parameter passed to
+        :meth:`.Engine.execution_options`, where it may be used to create
+        multiple engines with different isolation levels that share a common
+        connection pool and dialect.
+
+        .. versionchanged:: 2.0 The
+           :paramref:`_sa.create_engine.isolation_level`
+           parameter has been generalized to work on all dialects which support
+           the concept of isolation level, and is provided as a more succinct,
+           up front configuration switch in contrast to the execution option
+           which is more of an ad-hoc programmatic option.
+
+        .. seealso::
+
+            :ref:`dbapi_autocommit`
+
+    :param json_deserializer: for dialects that support the
+        :class:`_types.JSON`
+        datatype, this is a Python callable that will convert a JSON string
+        to a Python object.  By default, the Python ``json.loads`` function is
+        used.
+
+        .. versionchanged:: 1.3.7  The SQLite dialect renamed this from
+           ``_json_deserializer``.
+
+    :param json_serializer: for dialects that support the :class:`_types.JSON`
+        datatype, this is a Python callable that will render a given object
+        as JSON.   By default, the Python ``json.dumps`` function is used.
+
+        .. versionchanged:: 1.3.7  The SQLite dialect renamed this from
+           ``_json_serializer``.
+
+
+    :param label_length=None: optional integer value which limits
+        the size of dynamically generated column labels to that many
+        characters. If less than 6, labels are generated as
+        "_(counter)". If ``None``, the value of
+        ``dialect.max_identifier_length``, which may be affected via the
+        :paramref:`_sa.create_engine.max_identifier_length` parameter,
+        is used instead.   The value of
+        :paramref:`_sa.create_engine.label_length`
+        may not be larger than that of
+        :paramref:`_sa.create_engine.max_identfier_length`.
+
+        .. seealso::
+
+            :paramref:`_sa.create_engine.max_identifier_length`
+
+    :param logging_name:  String identifier which will be used within
+        the "name" field of logging records generated within the
+        "sqlalchemy.engine" logger. Defaults to a hexstring of the
+        object's id.
+
+        .. seealso::
+
+            :ref:`dbengine_logging` - further detail on how to configure
+            logging.
+
+            :paramref:`_engine.Connection.execution_options.logging_token`
+
+    :param max_identifier_length: integer; override the max_identifier_length
+        determined by the dialect.  if ``None`` or zero, has no effect.  This
+        is the database's configured maximum number of characters that may be
+        used in a SQL identifier such as a table name, column name, or label
+        name. All dialects determine this value automatically, however in the
+        case of a new database version for which this value has changed but
+        SQLAlchemy's dialect has not been adjusted, the value may be passed
+        here.
+
+        .. versionadded:: 1.3.9
+
+        .. seealso::
+
+            :paramref:`_sa.create_engine.label_length`
+
+    :param max_overflow=10: the number of connections to allow in
+        connection pool "overflow", that is connections that can be
+        opened above and beyond the pool_size setting, which defaults
+        to five. this is only used with :class:`~sqlalchemy.pool.QueuePool`.
+
+    :param module=None: reference to a Python module object (the module
+        itself, not its string name).  Specifies an alternate DBAPI module to
+        be used by the engine's dialect.  Each sub-dialect references a
+        specific DBAPI which will be imported before first connect.  This
+        parameter causes the import to be bypassed, and the given module to
+        be used instead. Can be used for testing of DBAPIs as well as to
+        inject "mock" DBAPI implementations into the :class:`_engine.Engine`.
+
+    :param paramstyle=None: The `paramstyle <https://legacy.python.org/dev/peps/pep-0249/#paramstyle>`_
+        to use when rendering bound parameters.  This style defaults to the
+        one recommended by the DBAPI itself, which is retrieved from the
+        ``.paramstyle`` attribute of the DBAPI.  However, most DBAPIs accept
+        more than one paramstyle, and in particular it may be desirable
+        to change a "named" paramstyle into a "positional" one, or vice versa.
+        When this attribute is passed, it should be one of the values
+        ``"qmark"``, ``"numeric"``, ``"named"``, ``"format"`` or
+        ``"pyformat"``, and should correspond to a parameter style known
+        to be supported by the DBAPI in use.
+
+    :param pool=None: an already-constructed instance of
+        :class:`~sqlalchemy.pool.Pool`, such as a
+        :class:`~sqlalchemy.pool.QueuePool` instance. If non-None, this
+        pool will be used directly as the underlying connection pool
+        for the engine, bypassing whatever connection parameters are
+        present in the URL argument. For information on constructing
+        connection pools manually, see :ref:`pooling_toplevel`.
+
+    :param poolclass=None: a :class:`~sqlalchemy.pool.Pool`
+        subclass, which will be used to create a connection pool
+        instance using the connection parameters given in the URL. Note
+        this differs from ``pool`` in that you don't actually
+        instantiate the pool in this case, you just indicate what type
+        of pool to be used.
+
+    :param pool_logging_name:  String identifier which will be used within
+       the "name" field of logging records generated within the
+       "sqlalchemy.pool" logger. Defaults to a hexstring of the object's
+       id.
+
+       .. seealso::
+
+            :ref:`dbengine_logging` - further detail on how to configure
+            logging.
+
+    :param pool_pre_ping: boolean, if True will enable the connection pool
+        "pre-ping" feature that tests connections for liveness upon
+        each checkout.
+
+        .. versionadded:: 1.2
+
+        .. seealso::
+
+            :ref:`pool_disconnects_pessimistic`
+
+    :param pool_size=5: the number of connections to keep open
+        inside the connection pool. This used with
+        :class:`~sqlalchemy.pool.QueuePool` as
+        well as :class:`~sqlalchemy.pool.SingletonThreadPool`.  With
+        :class:`~sqlalchemy.pool.QueuePool`, a ``pool_size`` setting
+        of 0 indicates no limit; to disable pooling, set ``poolclass`` to
+        :class:`~sqlalchemy.pool.NullPool` instead.
+
+    :param pool_recycle=-1: this setting causes the pool to recycle
+        connections after the given number of seconds has passed. It
+        defaults to -1, or no timeout. For example, setting to 3600
+        means connections will be recycled after one hour. Note that
+        MySQL in particular will disconnect automatically if no
+        activity is detected on a connection for eight hours (although
+        this is configurable with the MySQLDB connection itself and the
+        server configuration as well).
+
+        .. seealso::
+
+            :ref:`pool_setting_recycle`
+
+    :param pool_reset_on_return='rollback': set the
+        :paramref:`_pool.Pool.reset_on_return` parameter of the underlying
+        :class:`_pool.Pool` object, which can be set to the values
+        ``"rollback"``, ``"commit"``, or ``None``.
+
+        .. seealso::
+
+            :ref:`pool_reset_on_return`
+
+    :param pool_timeout=30: number of seconds to wait before giving
+        up on getting a connection from the pool. This is only used
+        with :class:`~sqlalchemy.pool.QueuePool`. This can be a float but is
+        subject to the limitations of Python time functions which may not be
+        reliable in the tens of milliseconds.
+
+        .. note: don't use 30.0 above, it seems to break with the :param tag
+
+    :param pool_use_lifo=False: use LIFO (last-in-first-out) when retrieving
+        connections from :class:`.QueuePool` instead of FIFO
+        (first-in-first-out). Using LIFO, a server-side timeout scheme can
+        reduce the number of connections used during non- peak   periods of
+        use.   When planning for server-side timeouts, ensure that a recycle or
+        pre-ping strategy is in use to gracefully   handle stale connections.
+
+          .. versionadded:: 1.3
+
+          .. seealso::
+
+            :ref:`pool_use_lifo`
+
+            :ref:`pool_disconnects`
+
+    :param plugins: string list of plugin names to load.  See
+        :class:`.CreateEnginePlugin` for background.
+
+        .. versionadded:: 1.2.3
+
+    :param query_cache_size: size of the cache used to cache the SQL string
+     form of queries.  Set to zero to disable caching.
+
+     The cache is pruned of its least recently used items when its size reaches
+     N * 1.5.  Defaults to 500, meaning the cache will always store at least
+     500 SQL statements when filled, and will grow up to 750 items at which
+     point it is pruned back down to 500 by removing the 250 least recently
+     used items.
+
+     Caching is accomplished on a per-statement basis by generating a
+     cache key that represents the statement's structure, then generating
+     string SQL for the current dialect only if that key is not present
+     in the cache.   All statements support caching, however some features
+     such as an INSERT with a large set of parameters will intentionally
+     bypass the cache.   SQL logging will indicate statistics for each
+     statement whether or not it were pull from the cache.
+
+     .. note:: some ORM functions related to unit-of-work persistence as well
+        as some attribute loading strategies will make use of individual
+        per-mapper caches outside of the main cache.
+
+
+     .. seealso::
+
+        :ref:`sql_caching`
+
+     .. versionadded:: 1.4
+
+    :param use_insertmanyvalues: True by default, use the "insertmanyvalues"
+     execution style for INSERT..RETURNING statements by default.
+
+     .. versionadded:: 2.0
+
+     .. seealso::
+
+        :ref:`engine_insertmanyvalues`
+
+    """  # noqa
+
+    if "strategy" in kwargs:
+        strat = kwargs.pop("strategy")
+        if strat == "mock":
+            # this case is deprecated
+            return create_mock_engine(url, **kwargs)  # type: ignore
+        else:
+            raise exc.ArgumentError("unknown strategy: %r" % strat)
+
+    kwargs.pop("empty_in_strategy", None)
+
+    # create url.URL object
+    u = _url.make_url(url)
+
+    u, plugins, kwargs = u._instantiate_plugins(kwargs)
+
+    entrypoint = u._get_entrypoint()
+    _is_async = kwargs.pop("_is_async", False)
+    if _is_async:
+        dialect_cls = entrypoint.get_async_dialect_cls(u)
+    else:
+        dialect_cls = entrypoint.get_dialect_cls(u)
+
+    if kwargs.pop("_coerce_config", False):
+
+        def pop_kwarg(key: str, default: Optional[Any] = None) -> Any:
+            value = kwargs.pop(key, default)
+            if key in dialect_cls.engine_config_types:
+                value = dialect_cls.engine_config_types[key](value)
+            return value
+
+    else:
+        pop_kwarg = kwargs.pop  # type: ignore
+
+    dialect_args = {}
+    # consume dialect arguments from kwargs
+    for k in util.get_cls_kwargs(dialect_cls):
+        if k in kwargs:
+            dialect_args[k] = pop_kwarg(k)
+
+    dbapi = kwargs.pop("module", None)
+    if dbapi is None:
+        dbapi_args = {}
+
+        if "import_dbapi" in dialect_cls.__dict__:
+            dbapi_meth = dialect_cls.import_dbapi
+
+        elif hasattr(dialect_cls, "dbapi") and inspect.ismethod(
+            dialect_cls.dbapi
+        ):
+            util.warn_deprecated(
+                "The dbapi() classmethod on dialect classes has been "
+                "renamed to import_dbapi().  Implement an import_dbapi() "
+                f"classmethod directly on class {dialect_cls} to remove this "
+                "warning; the old .dbapi() classmethod may be maintained for "
+                "backwards compatibility.",
+                "2.0",
+            )
+            dbapi_meth = dialect_cls.dbapi
+        else:
+            dbapi_meth = dialect_cls.import_dbapi
+
+        for k in util.get_func_kwargs(dbapi_meth):
+            if k in kwargs:
+                dbapi_args[k] = pop_kwarg(k)
+        dbapi = dbapi_meth(**dbapi_args)
+
+    dialect_args["dbapi"] = dbapi
+
+    dialect_args.setdefault("compiler_linting", compiler.NO_LINTING)
+    enable_from_linting = kwargs.pop("enable_from_linting", True)
+    if enable_from_linting:
+        dialect_args["compiler_linting"] ^= compiler.COLLECT_CARTESIAN_PRODUCTS
+
+    for plugin in plugins:
+        plugin.handle_dialect_kwargs(dialect_cls, dialect_args)
+
+    # create dialect
+    dialect = dialect_cls(**dialect_args)
+
+    # assemble connection arguments
+    (cargs_tup, cparams) = dialect.create_connect_args(u)
+    cparams.update(pop_kwarg("connect_args", {}))
+
+    if "async_fallback" in cparams and util.asbool(cparams["async_fallback"]):
+        util.warn_deprecated(
+            "The async_fallback dialect argument is deprecated and will be "
+            "removed in SQLAlchemy 2.1.",
+            "2.0",
+        )
+
+    cargs = list(cargs_tup)  # allow mutability
+
+    # look for existing pool or create
+    pool = pop_kwarg("pool", None)
+    if pool is None:
+
+        def connect(
+            connection_record: Optional[ConnectionPoolEntry] = None,
+        ) -> DBAPIConnection:
+            if dialect._has_events:
+                for fn in dialect.dispatch.do_connect:
+                    connection = cast(
+                        DBAPIConnection,
+                        fn(dialect, connection_record, cargs, cparams),
+                    )
+                    if connection is not None:
+                        return connection
+
+            return dialect.connect(*cargs, **cparams)
+
+        creator = pop_kwarg("creator", connect)
+
+        poolclass = pop_kwarg("poolclass", None)
+        if poolclass is None:
+            poolclass = dialect.get_dialect_pool_class(u)
+        pool_args = {"dialect": dialect}
+
+        # consume pool arguments from kwargs, translating a few of
+        # the arguments
+        for k in util.get_cls_kwargs(poolclass):
+            tk = _pool_translate_kwargs.get(k, k)
+            if tk in kwargs:
+                pool_args[k] = pop_kwarg(tk)
+
+        for plugin in plugins:
+            plugin.handle_pool_kwargs(poolclass, pool_args)
+
+        pool = poolclass(creator, **pool_args)
+    else:
+        pool._dialect = dialect
+
+    if (
+        hasattr(pool, "_is_asyncio")
+        and pool._is_asyncio is not dialect.is_async
+    ):
+        raise exc.ArgumentError(
+            f"Pool class {pool.__class__.__name__} cannot be "
+            f"used with {'non-' if not dialect.is_async else ''}"
+            "asyncio engine",
+            code="pcls",
+        )
+
+    # create engine.
+    if not pop_kwarg("future", True):
+        raise exc.ArgumentError(
+            "The 'future' parameter passed to "
+            "create_engine() may only be set to True."
+        )
+
+    engineclass = base.Engine
+
+    engine_args = {}
+    for k in util.get_cls_kwargs(engineclass):
+        if k in kwargs:
+            engine_args[k] = pop_kwarg(k)
+
+    # internal flags used by the test suite for instrumenting / proxying
+    # engines with mocks etc.
+    _initialize = kwargs.pop("_initialize", True)
+
+    # all kwargs should be consumed
+    if kwargs:
+        raise TypeError(
+            "Invalid argument(s) %s sent to create_engine(), "
+            "using configuration %s/%s/%s.  Please check that the "
+            "keyword arguments are appropriate for this combination "
+            "of components."
+            % (
+                ",".join("'%s'" % k for k in kwargs),
+                dialect.__class__.__name__,
+                pool.__class__.__name__,
+                engineclass.__name__,
+            )
+        )
+
+    engine = engineclass(pool, dialect, u, **engine_args)
+
+    if _initialize:
+        do_on_connect = dialect.on_connect_url(u)
+        if do_on_connect:
+
+            def on_connect(
+                dbapi_connection: DBAPIConnection,
+                connection_record: ConnectionPoolEntry,
+            ) -> None:
+                assert do_on_connect is not None
+                do_on_connect(dbapi_connection)
+
+            event.listen(pool, "connect", on_connect)
+
+        builtin_on_connect = dialect._builtin_onconnect()
+        if builtin_on_connect:
+            event.listen(pool, "connect", builtin_on_connect)
+
+        def first_connect(
+            dbapi_connection: DBAPIConnection,
+            connection_record: ConnectionPoolEntry,
+        ) -> None:
+            c = base.Connection(
+                engine,
+                connection=_AdhocProxiedConnection(
+                    dbapi_connection, connection_record
+                ),
+                _has_events=False,
+                # reconnecting will be a reentrant condition, so if the
+                # connection goes away, Connection is then closed
+                _allow_revalidate=False,
+                # dont trigger the autobegin sequence
+                # within the up front dialect checks
+                _allow_autobegin=False,
+            )
+            c._execution_options = util.EMPTY_DICT
+
+            try:
+                dialect.initialize(c)
+            finally:
+                # note that "invalidated" and "closed" are mutually
+                # exclusive in 1.4 Connection.
+                if not c.invalidated and not c.closed:
+                    # transaction is rolled back otherwise, tested by
+                    # test/dialect/postgresql/test_dialect.py
+                    # ::MiscBackendTest::test_initial_transaction_state
+                    dialect.do_rollback(c.connection)
+
+        # previously, the "first_connect" event was used here, which was then
+        # scaled back if the "on_connect" handler were present.  now,
+        # since "on_connect" is virtually always present, just use
+        # "connect" event with once_unless_exception in all cases so that
+        # the connection event flow is consistent in all cases.
+        event.listen(
+            pool, "connect", first_connect, _once_unless_exception=True
+        )
+
+    dialect_cls.engine_created(engine)
+    if entrypoint is not dialect_cls:
+        entrypoint.engine_created(engine)
+
+    for plugin in plugins:
+        plugin.engine_created(engine)
+
+    return engine
+
+
+def engine_from_config(
+    configuration: Dict[str, Any], prefix: str = "sqlalchemy.", **kwargs: Any
+) -> Engine:
+    """Create a new Engine instance using a configuration dictionary.
+
+    The dictionary is typically produced from a config file.
+
+    The keys of interest to ``engine_from_config()`` should be prefixed, e.g.
+    ``sqlalchemy.url``, ``sqlalchemy.echo``, etc.  The 'prefix' argument
+    indicates the prefix to be searched for.  Each matching key (after the
+    prefix is stripped) is treated as though it were the corresponding keyword
+    argument to a :func:`_sa.create_engine` call.
+
+    The only required key is (assuming the default prefix) ``sqlalchemy.url``,
+    which provides the :ref:`database URL <database_urls>`.
+
+    A select set of keyword arguments will be "coerced" to their
+    expected type based on string values.    The set of arguments
+    is extensible per-dialect using the ``engine_config_types`` accessor.
+
+    :param configuration: A dictionary (typically produced from a config file,
+        but this is not a requirement).  Items whose keys start with the value
+        of 'prefix' will have that prefix stripped, and will then be passed to
+        :func:`_sa.create_engine`.
+
+    :param prefix: Prefix to match and then strip from keys
+        in 'configuration'.
+
+    :param kwargs: Each keyword argument to ``engine_from_config()`` itself
+        overrides the corresponding item taken from the 'configuration'
+        dictionary.  Keyword arguments should *not* be prefixed.
+
+    """
+
+    options = {
+        key[len(prefix) :]: configuration[key]
+        for key in configuration
+        if key.startswith(prefix)
+    }
+    options["_coerce_config"] = True
+    options.update(kwargs)
+    url = options.pop("url")
+    return create_engine(url, **options)
+
+
+@overload
+def create_pool_from_url(
+    url: Union[str, URL],
+    *,
+    poolclass: Optional[Type[Pool]] = ...,
+    logging_name: str = ...,
+    pre_ping: bool = ...,
+    size: int = ...,
+    recycle: int = ...,
+    reset_on_return: Optional[_ResetStyleArgType] = ...,
+    timeout: float = ...,
+    use_lifo: bool = ...,
+    **kwargs: Any,
+) -> Pool: ...
+
+
+@overload
+def create_pool_from_url(url: Union[str, URL], **kwargs: Any) -> Pool: ...
+
+
+def create_pool_from_url(url: Union[str, URL], **kwargs: Any) -> Pool:
+    """Create a pool instance from the given url.
+
+    If ``poolclass`` is not provided the pool class used
+    is selected using the dialect specified in the URL.
+
+    The arguments passed to :func:`_sa.create_pool_from_url` are
+    identical to the pool argument passed to the :func:`_sa.create_engine`
+    function.
+
+    .. versionadded:: 2.0.10
+    """
+
+    for key in _pool_translate_kwargs:
+        if key in kwargs:
+            kwargs[_pool_translate_kwargs[key]] = kwargs.pop(key)
+
+    engine = create_engine(url, **kwargs, _initialize=False)
+    return engine.pool
+
+
+_pool_translate_kwargs = immutabledict(
+    {
+        "logging_name": "pool_logging_name",
+        "echo": "echo_pool",
+        "timeout": "pool_timeout",
+        "recycle": "pool_recycle",
+        "events": "pool_events",  # deprecated
+        "reset_on_return": "pool_reset_on_return",
+        "pre_ping": "pool_pre_ping",
+        "use_lifo": "pool_use_lifo",
+    }
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/cursor.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/cursor.py
new file mode 100644
index 00000000..ff14ad8e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/cursor.py
@@ -0,0 +1,2176 @@
+# engine/cursor.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Define cursor-specific result set constructs including
+:class:`.CursorResult`."""
+
+
+from __future__ import annotations
+
+import collections
+import functools
+import operator
+import typing
+from typing import Any
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .result import IteratorResult
+from .result import MergedResult
+from .result import Result
+from .result import ResultMetaData
+from .result import SimpleResultMetaData
+from .result import tuplegetter
+from .row import Row
+from .. import exc
+from .. import util
+from ..sql import elements
+from ..sql import sqltypes
+from ..sql import util as sql_util
+from ..sql.base import _generative
+from ..sql.compiler import ResultColumnsEntry
+from ..sql.compiler import RM_NAME
+from ..sql.compiler import RM_OBJECTS
+from ..sql.compiler import RM_RENDERED_NAME
+from ..sql.compiler import RM_TYPE
+from ..sql.type_api import TypeEngine
+from ..util import compat
+from ..util.typing import Literal
+from ..util.typing import Self
+
+
+if typing.TYPE_CHECKING:
+    from .base import Connection
+    from .default import DefaultExecutionContext
+    from .interfaces import _DBAPICursorDescription
+    from .interfaces import DBAPICursor
+    from .interfaces import Dialect
+    from .interfaces import ExecutionContext
+    from .result import _KeyIndexType
+    from .result import _KeyMapRecType
+    from .result import _KeyMapType
+    from .result import _KeyType
+    from .result import _ProcessorsType
+    from .result import _TupleGetterType
+    from ..sql.type_api import _ResultProcessorType
+
+
+_T = TypeVar("_T", bound=Any)
+
+
+# metadata entry tuple indexes.
+# using raw tuple is faster than namedtuple.
+# these match up to the positions in
+# _CursorKeyMapRecType
+MD_INDEX: Literal[0] = 0
+"""integer index in cursor.description
+
+"""
+
+MD_RESULT_MAP_INDEX: Literal[1] = 1
+"""integer index in compiled._result_columns"""
+
+MD_OBJECTS: Literal[2] = 2
+"""other string keys and ColumnElement obj that can match.
+
+This comes from compiler.RM_OBJECTS / compiler.ResultColumnsEntry.objects
+
+"""
+
+MD_LOOKUP_KEY: Literal[3] = 3
+"""string key we usually expect for key-based lookup
+
+this comes from compiler.RM_NAME / compiler.ResultColumnsEntry.name
+"""
+
+
+MD_RENDERED_NAME: Literal[4] = 4
+"""name that is usually in cursor.description
+
+this comes from compiler.RENDERED_NAME / compiler.ResultColumnsEntry.keyname
+"""
+
+
+MD_PROCESSOR: Literal[5] = 5
+"""callable to process a result value into a row"""
+
+MD_UNTRANSLATED: Literal[6] = 6
+"""raw name from cursor.description"""
+
+
+_CursorKeyMapRecType = Tuple[
+    Optional[int],  # MD_INDEX, None means the record is ambiguously named
+    int,  # MD_RESULT_MAP_INDEX
+    List[Any],  # MD_OBJECTS
+    str,  # MD_LOOKUP_KEY
+    str,  # MD_RENDERED_NAME
+    Optional["_ResultProcessorType[Any]"],  # MD_PROCESSOR
+    Optional[str],  # MD_UNTRANSLATED
+]
+
+_CursorKeyMapType = Mapping["_KeyType", _CursorKeyMapRecType]
+
+# same as _CursorKeyMapRecType except the MD_INDEX value is definitely
+# not None
+_NonAmbigCursorKeyMapRecType = Tuple[
+    int,
+    int,
+    List[Any],
+    str,
+    str,
+    Optional["_ResultProcessorType[Any]"],
+    str,
+]
+
+
+class CursorResultMetaData(ResultMetaData):
+    """Result metadata for DBAPI cursors."""
+
+    __slots__ = (
+        "_keymap",
+        "_processors",
+        "_keys",
+        "_keymap_by_result_column_idx",
+        "_tuplefilter",
+        "_translated_indexes",
+        "_safe_for_cache",
+        "_unpickled",
+        "_key_to_index",
+        # don't need _unique_filters support here for now.  Can be added
+        # if a need arises.
+    )
+
+    _keymap: _CursorKeyMapType
+    _processors: _ProcessorsType
+    _keymap_by_result_column_idx: Optional[Dict[int, _KeyMapRecType]]
+    _unpickled: bool
+    _safe_for_cache: bool
+    _translated_indexes: Optional[List[int]]
+
+    returns_rows: ClassVar[bool] = True
+
+    def _has_key(self, key: Any) -> bool:
+        return key in self._keymap
+
+    def _for_freeze(self) -> ResultMetaData:
+        return SimpleResultMetaData(
+            self._keys,
+            extra=[self._keymap[key][MD_OBJECTS] for key in self._keys],
+        )
+
+    def _make_new_metadata(
+        self,
+        *,
+        unpickled: bool,
+        processors: _ProcessorsType,
+        keys: Sequence[str],
+        keymap: _KeyMapType,
+        tuplefilter: Optional[_TupleGetterType],
+        translated_indexes: Optional[List[int]],
+        safe_for_cache: bool,
+        keymap_by_result_column_idx: Any,
+    ) -> CursorResultMetaData:
+        new_obj = self.__class__.__new__(self.__class__)
+        new_obj._unpickled = unpickled
+        new_obj._processors = processors
+        new_obj._keys = keys
+        new_obj._keymap = keymap
+        new_obj._tuplefilter = tuplefilter
+        new_obj._translated_indexes = translated_indexes
+        new_obj._safe_for_cache = safe_for_cache
+        new_obj._keymap_by_result_column_idx = keymap_by_result_column_idx
+        new_obj._key_to_index = self._make_key_to_index(keymap, MD_INDEX)
+        return new_obj
+
+    def _remove_processors(self) -> CursorResultMetaData:
+        assert not self._tuplefilter
+        return self._make_new_metadata(
+            unpickled=self._unpickled,
+            processors=[None] * len(self._processors),
+            tuplefilter=None,
+            translated_indexes=None,
+            keymap={
+                key: value[0:5] + (None,) + value[6:]
+                for key, value in self._keymap.items()
+            },
+            keys=self._keys,
+            safe_for_cache=self._safe_for_cache,
+            keymap_by_result_column_idx=self._keymap_by_result_column_idx,
+        )
+
+    def _splice_horizontally(
+        self, other: CursorResultMetaData
+    ) -> CursorResultMetaData:
+        assert not self._tuplefilter
+
+        keymap = dict(self._keymap)
+        offset = len(self._keys)
+        keymap.update(
+            {
+                key: (
+                    # int index should be None for ambiguous key
+                    (
+                        value[0] + offset
+                        if value[0] is not None and key not in keymap
+                        else None
+                    ),
+                    value[1] + offset,
+                    *value[2:],
+                )
+                for key, value in other._keymap.items()
+            }
+        )
+        return self._make_new_metadata(
+            unpickled=self._unpickled,
+            processors=self._processors + other._processors,  # type: ignore
+            tuplefilter=None,
+            translated_indexes=None,
+            keys=self._keys + other._keys,  # type: ignore
+            keymap=keymap,
+            safe_for_cache=self._safe_for_cache,
+            keymap_by_result_column_idx={
+                metadata_entry[MD_RESULT_MAP_INDEX]: metadata_entry
+                for metadata_entry in keymap.values()
+            },
+        )
+
+    def _reduce(self, keys: Sequence[_KeyIndexType]) -> ResultMetaData:
+        recs = list(self._metadata_for_keys(keys))
+
+        indexes = [rec[MD_INDEX] for rec in recs]
+        new_keys: List[str] = [rec[MD_LOOKUP_KEY] for rec in recs]
+
+        if self._translated_indexes:
+            indexes = [self._translated_indexes[idx] for idx in indexes]
+        tup = tuplegetter(*indexes)
+        new_recs = [(index,) + rec[1:] for index, rec in enumerate(recs)]
+
+        keymap = {rec[MD_LOOKUP_KEY]: rec for rec in new_recs}
+        # TODO: need unit test for:
+        # result = connection.execute("raw sql, no columns").scalars()
+        # without the "or ()" it's failing because MD_OBJECTS is None
+        keymap.update(
+            (e, new_rec)
+            for new_rec in new_recs
+            for e in new_rec[MD_OBJECTS] or ()
+        )
+
+        return self._make_new_metadata(
+            unpickled=self._unpickled,
+            processors=self._processors,
+            keys=new_keys,
+            tuplefilter=tup,
+            translated_indexes=indexes,
+            keymap=keymap,  # type: ignore[arg-type]
+            safe_for_cache=self._safe_for_cache,
+            keymap_by_result_column_idx=self._keymap_by_result_column_idx,
+        )
+
+    def _adapt_to_context(self, context: ExecutionContext) -> ResultMetaData:
+        """When using a cached Compiled construct that has a _result_map,
+        for a new statement that used the cached Compiled, we need to ensure
+        the keymap has the Column objects from our new statement as keys.
+        So here we rewrite keymap with new entries for the new columns
+        as matched to those of the cached statement.
+
+        """
+
+        if not context.compiled or not context.compiled._result_columns:
+            return self
+
+        compiled_statement = context.compiled.statement
+        invoked_statement = context.invoked_statement
+
+        if TYPE_CHECKING:
+            assert isinstance(invoked_statement, elements.ClauseElement)
+
+        if compiled_statement is invoked_statement:
+            return self
+
+        assert invoked_statement is not None
+
+        # this is the most common path for Core statements when
+        # caching is used.  In ORM use, this codepath is not really used
+        # as the _result_disable_adapt_to_context execution option is
+        # set by the ORM.
+
+        # make a copy and add the columns from the invoked statement
+        # to the result map.
+
+        keymap_by_position = self._keymap_by_result_column_idx
+
+        if keymap_by_position is None:
+            # first retrival from cache, this map will not be set up yet,
+            # initialize lazily
+            keymap_by_position = self._keymap_by_result_column_idx = {
+                metadata_entry[MD_RESULT_MAP_INDEX]: metadata_entry
+                for metadata_entry in self._keymap.values()
+            }
+
+        assert not self._tuplefilter
+        return self._make_new_metadata(
+            keymap=compat.dict_union(
+                self._keymap,
+                {
+                    new: keymap_by_position[idx]
+                    for idx, new in enumerate(
+                        invoked_statement._all_selected_columns
+                    )
+                    if idx in keymap_by_position
+                },
+            ),
+            unpickled=self._unpickled,
+            processors=self._processors,
+            tuplefilter=None,
+            translated_indexes=None,
+            keys=self._keys,
+            safe_for_cache=self._safe_for_cache,
+            keymap_by_result_column_idx=self._keymap_by_result_column_idx,
+        )
+
+    def __init__(
+        self,
+        parent: CursorResult[Any],
+        cursor_description: _DBAPICursorDescription,
+    ):
+        context = parent.context
+        self._tuplefilter = None
+        self._translated_indexes = None
+        self._safe_for_cache = self._unpickled = False
+
+        if context.result_column_struct:
+            (
+                result_columns,
+                cols_are_ordered,
+                textual_ordered,
+                ad_hoc_textual,
+                loose_column_name_matching,
+            ) = context.result_column_struct
+            num_ctx_cols = len(result_columns)
+        else:
+            result_columns = cols_are_ordered = (  # type: ignore
+                num_ctx_cols
+            ) = ad_hoc_textual = loose_column_name_matching = (
+                textual_ordered
+            ) = False
+
+        # merge cursor.description with the column info
+        # present in the compiled structure, if any
+        raw = self._merge_cursor_description(
+            context,
+            cursor_description,
+            result_columns,
+            num_ctx_cols,
+            cols_are_ordered,
+            textual_ordered,
+            ad_hoc_textual,
+            loose_column_name_matching,
+        )
+
+        # processors in key order which are used when building up
+        # a row
+        self._processors = [
+            metadata_entry[MD_PROCESSOR] for metadata_entry in raw
+        ]
+
+        # this is used when using this ResultMetaData in a Core-only cache
+        # retrieval context.  it's initialized on first cache retrieval
+        # when the _result_disable_adapt_to_context execution option
+        # (which the ORM generally sets) is not set.
+        self._keymap_by_result_column_idx = None
+
+        # for compiled SQL constructs, copy additional lookup keys into
+        # the key lookup map, such as Column objects, labels,
+        # column keys and other names
+        if num_ctx_cols:
+            # keymap by primary string...
+            by_key = {
+                metadata_entry[MD_LOOKUP_KEY]: metadata_entry
+                for metadata_entry in raw
+            }
+
+            if len(by_key) != num_ctx_cols:
+                # if by-primary-string dictionary smaller than
+                # number of columns, assume we have dupes; (this check
+                # is also in place if string dictionary is bigger, as
+                # can occur when '*' was used as one of the compiled columns,
+                # which may or may not be suggestive of dupes), rewrite
+                # dupe records with "None" for index which results in
+                # ambiguous column exception when accessed.
+                #
+                # this is considered to be the less common case as it is not
+                # common to have dupe column keys in a SELECT statement.
+                #
+                # new in 1.4: get the complete set of all possible keys,
+                # strings, objects, whatever, that are dupes across two
+                # different records, first.
+                index_by_key: Dict[Any, Any] = {}
+                dupes = set()
+                for metadata_entry in raw:
+                    for key in (metadata_entry[MD_RENDERED_NAME],) + (
+                        metadata_entry[MD_OBJECTS] or ()
+                    ):
+                        idx = metadata_entry[MD_INDEX]
+                        # if this key has been associated with more than one
+                        # positional index, it's a dupe
+                        if index_by_key.setdefault(key, idx) != idx:
+                            dupes.add(key)
+
+                # then put everything we have into the keymap excluding only
+                # those keys that are dupes.
+                self._keymap = {
+                    obj_elem: metadata_entry
+                    for metadata_entry in raw
+                    if metadata_entry[MD_OBJECTS]
+                    for obj_elem in metadata_entry[MD_OBJECTS]
+                    if obj_elem not in dupes
+                }
+
+                # then for the dupe keys, put the "ambiguous column"
+                # record into by_key.
+                by_key.update(
+                    {
+                        key: (None, None, [], key, key, None, None)
+                        for key in dupes
+                    }
+                )
+
+            else:
+                # no dupes - copy secondary elements from compiled
+                # columns into self._keymap.  this is the most common
+                # codepath for Core / ORM statement executions before the
+                # result metadata is cached
+                self._keymap = {
+                    obj_elem: metadata_entry
+                    for metadata_entry in raw
+                    if metadata_entry[MD_OBJECTS]
+                    for obj_elem in metadata_entry[MD_OBJECTS]
+                }
+            # update keymap with primary string names taking
+            # precedence
+            self._keymap.update(by_key)
+        else:
+            # no compiled objects to map, just create keymap by primary string
+            self._keymap = {
+                metadata_entry[MD_LOOKUP_KEY]: metadata_entry
+                for metadata_entry in raw
+            }
+
+        # update keymap with "translated" names.  In SQLAlchemy this is a
+        # sqlite only thing, and in fact impacting only extremely old SQLite
+        # versions unlikely to be present in modern Python versions.
+        # however, the pyhive third party dialect is
+        # also using this hook, which means others still might use it as well.
+        # I dislike having this awkward hook here but as long as we need
+        # to use names in cursor.description in some cases we need to have
+        # some hook to accomplish this.
+        if not num_ctx_cols and context._translate_colname:
+            self._keymap.update(
+                {
+                    metadata_entry[MD_UNTRANSLATED]: self._keymap[
+                        metadata_entry[MD_LOOKUP_KEY]
+                    ]
+                    for metadata_entry in raw
+                    if metadata_entry[MD_UNTRANSLATED]
+                }
+            )
+
+        self._key_to_index = self._make_key_to_index(self._keymap, MD_INDEX)
+
+    def _merge_cursor_description(
+        self,
+        context,
+        cursor_description,
+        result_columns,
+        num_ctx_cols,
+        cols_are_ordered,
+        textual_ordered,
+        ad_hoc_textual,
+        loose_column_name_matching,
+    ):
+        """Merge a cursor.description with compiled result column information.
+
+        There are at least four separate strategies used here, selected
+        depending on the type of SQL construct used to start with.
+
+        The most common case is that of the compiled SQL expression construct,
+        which generated the column names present in the raw SQL string and
+        which has the identical number of columns as were reported by
+        cursor.description.  In this case, we assume a 1-1 positional mapping
+        between the entries in cursor.description and the compiled object.
+        This is also the most performant case as we disregard extracting /
+        decoding the column names present in cursor.description since we
+        already have the desired name we generated in the compiled SQL
+        construct.
+
+        The next common case is that of the completely raw string SQL,
+        such as passed to connection.execute().  In this case we have no
+        compiled construct to work with, so we extract and decode the
+        names from cursor.description and index those as the primary
+        result row target keys.
+
+        The remaining fairly common case is that of the textual SQL
+        that includes at least partial column information; this is when
+        we use a :class:`_expression.TextualSelect` construct.
+        This construct may have
+        unordered or ordered column information.  In the ordered case, we
+        merge the cursor.description and the compiled construct's information
+        positionally, and warn if there are additional description names
+        present, however we still decode the names in cursor.description
+        as we don't have a guarantee that the names in the columns match
+        on these.   In the unordered case, we match names in cursor.description
+        to that of the compiled construct based on name matching.
+        In both of these cases, the cursor.description names and the column
+        expression objects and names are indexed as result row target keys.
+
+        The final case is much less common, where we have a compiled
+        non-textual SQL expression construct, but the number of columns
+        in cursor.description doesn't match what's in the compiled
+        construct.  We make the guess here that there might be textual
+        column expressions in the compiled construct that themselves include
+        a comma in them causing them to split.  We do the same name-matching
+        as with textual non-ordered columns.
+
+        The name-matched system of merging is the same as that used by
+        SQLAlchemy for all cases up through the 0.9 series.   Positional
+        matching for compiled SQL expressions was introduced in 1.0 as a
+        major performance feature, and positional matching for textual
+        :class:`_expression.TextualSelect` objects in 1.1.
+        As name matching is no longer
+        a common case, it was acceptable to factor it into smaller generator-
+        oriented methods that are easier to understand, but incur slightly
+        more performance overhead.
+
+        """
+
+        if (
+            num_ctx_cols
+            and cols_are_ordered
+            and not textual_ordered
+            and num_ctx_cols == len(cursor_description)
+        ):
+            self._keys = [elem[0] for elem in result_columns]
+            # pure positional 1-1 case; doesn't need to read
+            # the names from cursor.description
+
+            # most common case for Core and ORM
+
+            # this metadata is safe to cache because we are guaranteed
+            # to have the columns in the same order for new executions
+            self._safe_for_cache = True
+            return [
+                (
+                    idx,
+                    idx,
+                    rmap_entry[RM_OBJECTS],
+                    rmap_entry[RM_NAME],
+                    rmap_entry[RM_RENDERED_NAME],
+                    context.get_result_processor(
+                        rmap_entry[RM_TYPE],
+                        rmap_entry[RM_RENDERED_NAME],
+                        cursor_description[idx][1],
+                    ),
+                    None,
+                )
+                for idx, rmap_entry in enumerate(result_columns)
+            ]
+        else:
+            # name-based or text-positional cases, where we need
+            # to read cursor.description names
+
+            if textual_ordered or (
+                ad_hoc_textual and len(cursor_description) == num_ctx_cols
+            ):
+                self._safe_for_cache = True
+                # textual positional case
+                raw_iterator = self._merge_textual_cols_by_position(
+                    context, cursor_description, result_columns
+                )
+            elif num_ctx_cols:
+                # compiled SQL with a mismatch of description cols
+                # vs. compiled cols, or textual w/ unordered columns
+                # the order of columns can change if the query is
+                # against a "select *", so not safe to cache
+                self._safe_for_cache = False
+                raw_iterator = self._merge_cols_by_name(
+                    context,
+                    cursor_description,
+                    result_columns,
+                    loose_column_name_matching,
+                )
+            else:
+                # no compiled SQL, just a raw string, order of columns
+                # can change for "select *"
+                self._safe_for_cache = False
+                raw_iterator = self._merge_cols_by_none(
+                    context, cursor_description
+                )
+
+            return [
+                (
+                    idx,
+                    ridx,
+                    obj,
+                    cursor_colname,
+                    cursor_colname,
+                    context.get_result_processor(
+                        mapped_type, cursor_colname, coltype
+                    ),
+                    untranslated,
+                )
+                for (
+                    idx,
+                    ridx,
+                    cursor_colname,
+                    mapped_type,
+                    coltype,
+                    obj,
+                    untranslated,
+                ) in raw_iterator
+            ]
+
+    def _colnames_from_description(self, context, cursor_description):
+        """Extract column names and data types from a cursor.description.
+
+        Applies unicode decoding, column translation, "normalization",
+        and case sensitivity rules to the names based on the dialect.
+
+        """
+
+        dialect = context.dialect
+        translate_colname = context._translate_colname
+        normalize_name = (
+            dialect.normalize_name if dialect.requires_name_normalize else None
+        )
+        untranslated = None
+
+        self._keys = []
+
+        for idx, rec in enumerate(cursor_description):
+            colname = rec[0]
+            coltype = rec[1]
+
+            if translate_colname:
+                colname, untranslated = translate_colname(colname)
+
+            if normalize_name:
+                colname = normalize_name(colname)
+
+            self._keys.append(colname)
+
+            yield idx, colname, untranslated, coltype
+
+    def _merge_textual_cols_by_position(
+        self, context, cursor_description, result_columns
+    ):
+        num_ctx_cols = len(result_columns)
+
+        if num_ctx_cols > len(cursor_description):
+            util.warn(
+                "Number of columns in textual SQL (%d) is "
+                "smaller than number of columns requested (%d)"
+                % (num_ctx_cols, len(cursor_description))
+            )
+        seen = set()
+
+        for (
+            idx,
+            colname,
+            untranslated,
+            coltype,
+        ) in self._colnames_from_description(context, cursor_description):
+            if idx < num_ctx_cols:
+                ctx_rec = result_columns[idx]
+                obj = ctx_rec[RM_OBJECTS]
+                ridx = idx
+                mapped_type = ctx_rec[RM_TYPE]
+                if obj[0] in seen:
+                    raise exc.InvalidRequestError(
+                        "Duplicate column expression requested "
+                        "in textual SQL: %r" % obj[0]
+                    )
+                seen.add(obj[0])
+            else:
+                mapped_type = sqltypes.NULLTYPE
+                obj = None
+                ridx = None
+            yield idx, ridx, colname, mapped_type, coltype, obj, untranslated
+
+    def _merge_cols_by_name(
+        self,
+        context,
+        cursor_description,
+        result_columns,
+        loose_column_name_matching,
+    ):
+        match_map = self._create_description_match_map(
+            result_columns, loose_column_name_matching
+        )
+        mapped_type: TypeEngine[Any]
+
+        for (
+            idx,
+            colname,
+            untranslated,
+            coltype,
+        ) in self._colnames_from_description(context, cursor_description):
+            try:
+                ctx_rec = match_map[colname]
+            except KeyError:
+                mapped_type = sqltypes.NULLTYPE
+                obj = None
+                result_columns_idx = None
+            else:
+                obj = ctx_rec[1]
+                mapped_type = ctx_rec[2]
+                result_columns_idx = ctx_rec[3]
+            yield (
+                idx,
+                result_columns_idx,
+                colname,
+                mapped_type,
+                coltype,
+                obj,
+                untranslated,
+            )
+
+    @classmethod
+    def _create_description_match_map(
+        cls,
+        result_columns: List[ResultColumnsEntry],
+        loose_column_name_matching: bool = False,
+    ) -> Dict[
+        Union[str, object], Tuple[str, Tuple[Any, ...], TypeEngine[Any], int]
+    ]:
+        """when matching cursor.description to a set of names that are present
+        in a Compiled object, as is the case with TextualSelect, get all the
+        names we expect might match those in cursor.description.
+        """
+
+        d: Dict[
+            Union[str, object],
+            Tuple[str, Tuple[Any, ...], TypeEngine[Any], int],
+        ] = {}
+        for ridx, elem in enumerate(result_columns):
+            key = elem[RM_RENDERED_NAME]
+            if key in d:
+                # conflicting keyname - just add the column-linked objects
+                # to the existing record.  if there is a duplicate column
+                # name in the cursor description, this will allow all of those
+                # objects to raise an ambiguous column error
+                e_name, e_obj, e_type, e_ridx = d[key]
+                d[key] = e_name, e_obj + elem[RM_OBJECTS], e_type, ridx
+            else:
+                d[key] = (elem[RM_NAME], elem[RM_OBJECTS], elem[RM_TYPE], ridx)
+
+            if loose_column_name_matching:
+                # when using a textual statement with an unordered set
+                # of columns that line up, we are expecting the user
+                # to be using label names in the SQL that match to the column
+                # expressions.  Enable more liberal matching for this case;
+                # duplicate keys that are ambiguous will be fixed later.
+                for r_key in elem[RM_OBJECTS]:
+                    d.setdefault(
+                        r_key,
+                        (elem[RM_NAME], elem[RM_OBJECTS], elem[RM_TYPE], ridx),
+                    )
+        return d
+
+    def _merge_cols_by_none(self, context, cursor_description):
+        for (
+            idx,
+            colname,
+            untranslated,
+            coltype,
+        ) in self._colnames_from_description(context, cursor_description):
+            yield (
+                idx,
+                None,
+                colname,
+                sqltypes.NULLTYPE,
+                coltype,
+                None,
+                untranslated,
+            )
+
+    if not TYPE_CHECKING:
+
+        def _key_fallback(
+            self, key: Any, err: Optional[Exception], raiseerr: bool = True
+        ) -> Optional[NoReturn]:
+            if raiseerr:
+                if self._unpickled and isinstance(key, elements.ColumnElement):
+                    raise exc.NoSuchColumnError(
+                        "Row was unpickled; lookup by ColumnElement "
+                        "is unsupported"
+                    ) from err
+                else:
+                    raise exc.NoSuchColumnError(
+                        "Could not locate column in row for column '%s'"
+                        % util.string_or_unprintable(key)
+                    ) from err
+            else:
+                return None
+
+    def _raise_for_ambiguous_column_name(self, rec):
+        raise exc.InvalidRequestError(
+            "Ambiguous column name '%s' in "
+            "result set column descriptions" % rec[MD_LOOKUP_KEY]
+        )
+
+    def _index_for_key(self, key: Any, raiseerr: bool = True) -> Optional[int]:
+        # TODO: can consider pre-loading ints and negative ints
+        # into _keymap - also no coverage here
+        if isinstance(key, int):
+            key = self._keys[key]
+
+        try:
+            rec = self._keymap[key]
+        except KeyError as ke:
+            x = self._key_fallback(key, ke, raiseerr)
+            assert x is None
+            return None
+
+        index = rec[0]
+
+        if index is None:
+            self._raise_for_ambiguous_column_name(rec)
+        return index
+
+    def _indexes_for_keys(self, keys):
+        try:
+            return [self._keymap[key][0] for key in keys]
+        except KeyError as ke:
+            # ensure it raises
+            CursorResultMetaData._key_fallback(self, ke.args[0], ke)
+
+    def _metadata_for_keys(
+        self, keys: Sequence[Any]
+    ) -> Iterator[_NonAmbigCursorKeyMapRecType]:
+        for key in keys:
+            if int in key.__class__.__mro__:
+                key = self._keys[key]
+
+            try:
+                rec = self._keymap[key]
+            except KeyError as ke:
+                # ensure it raises
+                CursorResultMetaData._key_fallback(self, ke.args[0], ke)
+
+            index = rec[MD_INDEX]
+
+            if index is None:
+                self._raise_for_ambiguous_column_name(rec)
+
+            yield cast(_NonAmbigCursorKeyMapRecType, rec)
+
+    def __getstate__(self):
+        # TODO: consider serializing this as SimpleResultMetaData
+        return {
+            "_keymap": {
+                key: (
+                    rec[MD_INDEX],
+                    rec[MD_RESULT_MAP_INDEX],
+                    [],
+                    key,
+                    rec[MD_RENDERED_NAME],
+                    None,
+                    None,
+                )
+                for key, rec in self._keymap.items()
+                if isinstance(key, (str, int))
+            },
+            "_keys": self._keys,
+            "_translated_indexes": self._translated_indexes,
+        }
+
+    def __setstate__(self, state):
+        self._processors = [None for _ in range(len(state["_keys"]))]
+        self._keymap = state["_keymap"]
+        self._keymap_by_result_column_idx = None
+        self._key_to_index = self._make_key_to_index(self._keymap, MD_INDEX)
+        self._keys = state["_keys"]
+        self._unpickled = True
+        if state["_translated_indexes"]:
+            self._translated_indexes = cast(
+                "List[int]", state["_translated_indexes"]
+            )
+            self._tuplefilter = tuplegetter(*self._translated_indexes)
+        else:
+            self._translated_indexes = self._tuplefilter = None
+
+
+class ResultFetchStrategy:
+    """Define a fetching strategy for a result object.
+
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    alternate_cursor_description: Optional[_DBAPICursorDescription] = None
+
+    def soft_close(
+        self, result: CursorResult[Any], dbapi_cursor: Optional[DBAPICursor]
+    ) -> None:
+        raise NotImplementedError()
+
+    def hard_close(
+        self, result: CursorResult[Any], dbapi_cursor: Optional[DBAPICursor]
+    ) -> None:
+        raise NotImplementedError()
+
+    def yield_per(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: Optional[DBAPICursor],
+        num: int,
+    ) -> None:
+        return
+
+    def fetchone(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+        hard_close: bool = False,
+    ) -> Any:
+        raise NotImplementedError()
+
+    def fetchmany(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+        size: Optional[int] = None,
+    ) -> Any:
+        raise NotImplementedError()
+
+    def fetchall(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+    ) -> Any:
+        raise NotImplementedError()
+
+    def handle_exception(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: Optional[DBAPICursor],
+        err: BaseException,
+    ) -> NoReturn:
+        raise err
+
+
+class NoCursorFetchStrategy(ResultFetchStrategy):
+    """Cursor strategy for a result that has no open cursor.
+
+    There are two varieties of this strategy, one for DQL and one for
+    DML (and also DDL), each of which represent a result that had a cursor
+    but no longer has one.
+
+    """
+
+    __slots__ = ()
+
+    def soft_close(self, result, dbapi_cursor):
+        pass
+
+    def hard_close(self, result, dbapi_cursor):
+        pass
+
+    def fetchone(self, result, dbapi_cursor, hard_close=False):
+        return self._non_result(result, None)
+
+    def fetchmany(self, result, dbapi_cursor, size=None):
+        return self._non_result(result, [])
+
+    def fetchall(self, result, dbapi_cursor):
+        return self._non_result(result, [])
+
+    def _non_result(self, result, default, err=None):
+        raise NotImplementedError()
+
+
+class NoCursorDQLFetchStrategy(NoCursorFetchStrategy):
+    """Cursor strategy for a DQL result that has no open cursor.
+
+    This is a result set that can return rows, i.e. for a SELECT, or for an
+    INSERT, UPDATE, DELETE that includes RETURNING. However it is in the state
+    where the cursor is closed and no rows remain available.  The owning result
+    object may or may not be "hard closed", which determines if the fetch
+    methods send empty results or raise for closed result.
+
+    """
+
+    __slots__ = ()
+
+    def _non_result(self, result, default, err=None):
+        if result.closed:
+            raise exc.ResourceClosedError(
+                "This result object is closed."
+            ) from err
+        else:
+            return default
+
+
+_NO_CURSOR_DQL = NoCursorDQLFetchStrategy()
+
+
+class NoCursorDMLFetchStrategy(NoCursorFetchStrategy):
+    """Cursor strategy for a DML result that has no open cursor.
+
+    This is a result set that does not return rows, i.e. for an INSERT,
+    UPDATE, DELETE that does not include RETURNING.
+
+    """
+
+    __slots__ = ()
+
+    def _non_result(self, result, default, err=None):
+        # we only expect to have a _NoResultMetaData() here right now.
+        assert not result._metadata.returns_rows
+        result._metadata._we_dont_return_rows(err)
+
+
+_NO_CURSOR_DML = NoCursorDMLFetchStrategy()
+
+
+class CursorFetchStrategy(ResultFetchStrategy):
+    """Call fetch methods from a DBAPI cursor.
+
+    Alternate versions of this class may instead buffer the rows from
+    cursors or not use cursors at all.
+
+    """
+
+    __slots__ = ()
+
+    def soft_close(
+        self, result: CursorResult[Any], dbapi_cursor: Optional[DBAPICursor]
+    ) -> None:
+        result.cursor_strategy = _NO_CURSOR_DQL
+
+    def hard_close(
+        self, result: CursorResult[Any], dbapi_cursor: Optional[DBAPICursor]
+    ) -> None:
+        result.cursor_strategy = _NO_CURSOR_DQL
+
+    def handle_exception(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: Optional[DBAPICursor],
+        err: BaseException,
+    ) -> NoReturn:
+        result.connection._handle_dbapi_exception(
+            err, None, None, dbapi_cursor, result.context
+        )
+
+    def yield_per(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: Optional[DBAPICursor],
+        num: int,
+    ) -> None:
+        result.cursor_strategy = BufferedRowCursorFetchStrategy(
+            dbapi_cursor,
+            {"max_row_buffer": num},
+            initial_buffer=collections.deque(),
+            growth_factor=0,
+        )
+
+    def fetchone(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+        hard_close: bool = False,
+    ) -> Any:
+        try:
+            row = dbapi_cursor.fetchone()
+            if row is None:
+                result._soft_close(hard=hard_close)
+            return row
+        except BaseException as e:
+            self.handle_exception(result, dbapi_cursor, e)
+
+    def fetchmany(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+        size: Optional[int] = None,
+    ) -> Any:
+        try:
+            if size is None:
+                l = dbapi_cursor.fetchmany()
+            else:
+                l = dbapi_cursor.fetchmany(size)
+
+            if not l:
+                result._soft_close()
+            return l
+        except BaseException as e:
+            self.handle_exception(result, dbapi_cursor, e)
+
+    def fetchall(
+        self,
+        result: CursorResult[Any],
+        dbapi_cursor: DBAPICursor,
+    ) -> Any:
+        try:
+            rows = dbapi_cursor.fetchall()
+            result._soft_close()
+            return rows
+        except BaseException as e:
+            self.handle_exception(result, dbapi_cursor, e)
+
+
+_DEFAULT_FETCH = CursorFetchStrategy()
+
+
+class BufferedRowCursorFetchStrategy(CursorFetchStrategy):
+    """A cursor fetch strategy with row buffering behavior.
+
+    This strategy buffers the contents of a selection of rows
+    before ``fetchone()`` is called.  This is to allow the results of
+    ``cursor.description`` to be available immediately, when
+    interfacing with a DB-API that requires rows to be consumed before
+    this information is available (currently psycopg2, when used with
+    server-side cursors).
+
+    The pre-fetching behavior fetches only one row initially, and then
+    grows its buffer size by a fixed amount with each successive need
+    for additional rows up the ``max_row_buffer`` size, which defaults
+    to 1000::
+
+        with psycopg2_engine.connect() as conn:
+
+            result = conn.execution_options(
+                stream_results=True, max_row_buffer=50
+            ).execute(text("select * from table"))
+
+    .. versionadded:: 1.4 ``max_row_buffer`` may now exceed 1000 rows.
+
+    .. seealso::
+
+        :ref:`psycopg2_execution_options`
+    """
+
+    __slots__ = ("_max_row_buffer", "_rowbuffer", "_bufsize", "_growth_factor")
+
+    def __init__(
+        self,
+        dbapi_cursor,
+        execution_options,
+        growth_factor=5,
+        initial_buffer=None,
+    ):
+        self._max_row_buffer = execution_options.get("max_row_buffer", 1000)
+
+        if initial_buffer is not None:
+            self._rowbuffer = initial_buffer
+        else:
+            self._rowbuffer = collections.deque(dbapi_cursor.fetchmany(1))
+        self._growth_factor = growth_factor
+
+        if growth_factor:
+            self._bufsize = min(self._max_row_buffer, self._growth_factor)
+        else:
+            self._bufsize = self._max_row_buffer
+
+    @classmethod
+    def create(cls, result):
+        return BufferedRowCursorFetchStrategy(
+            result.cursor,
+            result.context.execution_options,
+        )
+
+    def _buffer_rows(self, result, dbapi_cursor):
+        """this is currently used only by fetchone()."""
+
+        size = self._bufsize
+        try:
+            if size < 1:
+                new_rows = dbapi_cursor.fetchall()
+            else:
+                new_rows = dbapi_cursor.fetchmany(size)
+        except BaseException as e:
+            self.handle_exception(result, dbapi_cursor, e)
+
+        if not new_rows:
+            return
+        self._rowbuffer = collections.deque(new_rows)
+        if self._growth_factor and size < self._max_row_buffer:
+            self._bufsize = min(
+                self._max_row_buffer, size * self._growth_factor
+            )
+
+    def yield_per(self, result, dbapi_cursor, num):
+        self._growth_factor = 0
+        self._max_row_buffer = self._bufsize = num
+
+    def soft_close(self, result, dbapi_cursor):
+        self._rowbuffer.clear()
+        super().soft_close(result, dbapi_cursor)
+
+    def hard_close(self, result, dbapi_cursor):
+        self._rowbuffer.clear()
+        super().hard_close(result, dbapi_cursor)
+
+    def fetchone(self, result, dbapi_cursor, hard_close=False):
+        if not self._rowbuffer:
+            self._buffer_rows(result, dbapi_cursor)
+            if not self._rowbuffer:
+                try:
+                    result._soft_close(hard=hard_close)
+                except BaseException as e:
+                    self.handle_exception(result, dbapi_cursor, e)
+                return None
+        return self._rowbuffer.popleft()
+
+    def fetchmany(self, result, dbapi_cursor, size=None):
+        if size is None:
+            return self.fetchall(result, dbapi_cursor)
+
+        rb = self._rowbuffer
+        lb = len(rb)
+        close = False
+        if size > lb:
+            try:
+                new = dbapi_cursor.fetchmany(size - lb)
+            except BaseException as e:
+                self.handle_exception(result, dbapi_cursor, e)
+            else:
+                if not new:
+                    # defer closing since it may clear the row buffer
+                    close = True
+                else:
+                    rb.extend(new)
+
+        res = [rb.popleft() for _ in range(min(size, len(rb)))]
+        if close:
+            result._soft_close()
+        return res
+
+    def fetchall(self, result, dbapi_cursor):
+        try:
+            ret = list(self._rowbuffer) + list(dbapi_cursor.fetchall())
+            self._rowbuffer.clear()
+            result._soft_close()
+            return ret
+        except BaseException as e:
+            self.handle_exception(result, dbapi_cursor, e)
+
+
+class FullyBufferedCursorFetchStrategy(CursorFetchStrategy):
+    """A cursor strategy that buffers rows fully upon creation.
+
+    Used for operations where a result is to be delivered
+    after the database conversation can not be continued,
+    such as MSSQL INSERT...OUTPUT after an autocommit.
+
+    """
+
+    __slots__ = ("_rowbuffer", "alternate_cursor_description")
+
+    def __init__(
+        self, dbapi_cursor, alternate_description=None, initial_buffer=None
+    ):
+        self.alternate_cursor_description = alternate_description
+        if initial_buffer is not None:
+            self._rowbuffer = collections.deque(initial_buffer)
+        else:
+            self._rowbuffer = collections.deque(dbapi_cursor.fetchall())
+
+    def yield_per(self, result, dbapi_cursor, num):
+        pass
+
+    def soft_close(self, result, dbapi_cursor):
+        self._rowbuffer.clear()
+        super().soft_close(result, dbapi_cursor)
+
+    def hard_close(self, result, dbapi_cursor):
+        self._rowbuffer.clear()
+        super().hard_close(result, dbapi_cursor)
+
+    def fetchone(self, result, dbapi_cursor, hard_close=False):
+        if self._rowbuffer:
+            return self._rowbuffer.popleft()
+        else:
+            result._soft_close(hard=hard_close)
+            return None
+
+    def fetchmany(self, result, dbapi_cursor, size=None):
+        if size is None:
+            return self.fetchall(result, dbapi_cursor)
+
+        rb = self._rowbuffer
+        rows = [rb.popleft() for _ in range(min(size, len(rb)))]
+        if not rows:
+            result._soft_close()
+        return rows
+
+    def fetchall(self, result, dbapi_cursor):
+        ret = self._rowbuffer
+        self._rowbuffer = collections.deque()
+        result._soft_close()
+        return ret
+
+
+class _NoResultMetaData(ResultMetaData):
+    __slots__ = ()
+
+    returns_rows = False
+
+    def _we_dont_return_rows(self, err=None):
+        raise exc.ResourceClosedError(
+            "This result object does not return rows. "
+            "It has been closed automatically."
+        ) from err
+
+    def _index_for_key(self, keys, raiseerr):
+        self._we_dont_return_rows()
+
+    def _metadata_for_keys(self, key):
+        self._we_dont_return_rows()
+
+    def _reduce(self, keys):
+        self._we_dont_return_rows()
+
+    @property
+    def _keymap(self):
+        self._we_dont_return_rows()
+
+    @property
+    def _key_to_index(self):
+        self._we_dont_return_rows()
+
+    @property
+    def _processors(self):
+        self._we_dont_return_rows()
+
+    @property
+    def keys(self):
+        self._we_dont_return_rows()
+
+
+_NO_RESULT_METADATA = _NoResultMetaData()
+
+
+def null_dml_result() -> IteratorResult[Any]:
+    it: IteratorResult[Any] = IteratorResult(_NoResultMetaData(), iter([]))
+    it._soft_close()
+    return it
+
+
+class CursorResult(Result[_T]):
+    """A Result that is representing state from a DBAPI cursor.
+
+    .. versionchanged:: 1.4  The :class:`.CursorResult``
+       class replaces the previous :class:`.ResultProxy` interface.
+       This classes are based on the :class:`.Result` calling API
+       which provides an updated usage model and calling facade for
+       SQLAlchemy Core and SQLAlchemy ORM.
+
+    Returns database rows via the :class:`.Row` class, which provides
+    additional API features and behaviors on top of the raw data returned by
+    the DBAPI.   Through the use of filters such as the :meth:`.Result.scalars`
+    method, other kinds of objects may also be returned.
+
+    .. seealso::
+
+        :ref:`tutorial_selecting_data` - introductory material for accessing
+        :class:`_engine.CursorResult` and :class:`.Row` objects.
+
+    """
+
+    __slots__ = (
+        "context",
+        "dialect",
+        "cursor",
+        "cursor_strategy",
+        "_echo",
+        "connection",
+    )
+
+    _metadata: Union[CursorResultMetaData, _NoResultMetaData]
+    _no_result_metadata = _NO_RESULT_METADATA
+    _soft_closed: bool = False
+    closed: bool = False
+    _is_cursor = True
+
+    context: DefaultExecutionContext
+    dialect: Dialect
+    cursor_strategy: ResultFetchStrategy
+    connection: Connection
+
+    def __init__(
+        self,
+        context: DefaultExecutionContext,
+        cursor_strategy: ResultFetchStrategy,
+        cursor_description: Optional[_DBAPICursorDescription],
+    ):
+        self.context = context
+        self.dialect = context.dialect
+        self.cursor = context.cursor
+        self.cursor_strategy = cursor_strategy
+        self.connection = context.root_connection
+        self._echo = echo = (
+            self.connection._echo and context.engine._should_log_debug()
+        )
+
+        if cursor_description is not None:
+            # inline of Result._row_getter(), set up an initial row
+            # getter assuming no transformations will be called as this
+            # is the most common case
+
+            metadata = self._init_metadata(context, cursor_description)
+
+            _make_row: Any
+            _make_row = functools.partial(
+                Row,
+                metadata,
+                metadata._effective_processors,
+                metadata._key_to_index,
+            )
+
+            if context._num_sentinel_cols:
+                sentinel_filter = operator.itemgetter(
+                    slice(-context._num_sentinel_cols)
+                )
+
+                def _sliced_row(raw_data):
+                    return _make_row(sentinel_filter(raw_data))
+
+                sliced_row = _sliced_row
+            else:
+                sliced_row = _make_row
+
+            if echo:
+                log = self.context.connection._log_debug
+
+                def _log_row(row):
+                    log("Row %r", sql_util._repr_row(row))
+                    return row
+
+                self._row_logging_fn = _log_row
+
+                def _make_row_2(row):
+                    return _log_row(sliced_row(row))
+
+                make_row = _make_row_2
+            else:
+                make_row = sliced_row
+            self._set_memoized_attribute("_row_getter", make_row)
+
+        else:
+            assert context._num_sentinel_cols == 0
+            self._metadata = self._no_result_metadata
+
+    def _init_metadata(self, context, cursor_description):
+        if context.compiled:
+            compiled = context.compiled
+
+            if compiled._cached_metadata:
+                metadata = compiled._cached_metadata
+            else:
+                metadata = CursorResultMetaData(self, cursor_description)
+                if metadata._safe_for_cache:
+                    compiled._cached_metadata = metadata
+
+            # result rewrite/ adapt step.  this is to suit the case
+            # when we are invoked against a cached Compiled object, we want
+            # to rewrite the ResultMetaData to reflect the Column objects
+            # that are in our current SQL statement object, not the one
+            # that is associated with the cached Compiled object.
+            # the Compiled object may also tell us to not
+            # actually do this step; this is to support the ORM where
+            # it is to produce a new Result object in any case, and will
+            # be using the cached Column objects against this database result
+            # so we don't want to rewrite them.
+            #
+            # Basically this step suits the use case where the end user
+            # is using Core SQL expressions and is accessing columns in the
+            # result row using row._mapping[table.c.column].
+            if (
+                not context.execution_options.get(
+                    "_result_disable_adapt_to_context", False
+                )
+                and compiled._result_columns
+                and context.cache_hit is context.dialect.CACHE_HIT
+                and compiled.statement is not context.invoked_statement
+            ):
+                metadata = metadata._adapt_to_context(context)
+
+            self._metadata = metadata
+
+        else:
+            self._metadata = metadata = CursorResultMetaData(
+                self, cursor_description
+            )
+        if self._echo:
+            context.connection._log_debug(
+                "Col %r", tuple(x[0] for x in cursor_description)
+            )
+        return metadata
+
+    def _soft_close(self, hard=False):
+        """Soft close this :class:`_engine.CursorResult`.
+
+        This releases all DBAPI cursor resources, but leaves the
+        CursorResult "open" from a semantic perspective, meaning the
+        fetchXXX() methods will continue to return empty results.
+
+        This method is called automatically when:
+
+        * all result rows are exhausted using the fetchXXX() methods.
+        * cursor.description is None.
+
+        This method is **not public**, but is documented in order to clarify
+        the "autoclose" process used.
+
+        .. seealso::
+
+            :meth:`_engine.CursorResult.close`
+
+
+        """
+
+        if (not hard and self._soft_closed) or (hard and self.closed):
+            return
+
+        if hard:
+            self.closed = True
+            self.cursor_strategy.hard_close(self, self.cursor)
+        else:
+            self.cursor_strategy.soft_close(self, self.cursor)
+
+        if not self._soft_closed:
+            cursor = self.cursor
+            self.cursor = None  # type: ignore
+            self.connection._safe_close_cursor(cursor)
+            self._soft_closed = True
+
+    @property
+    def inserted_primary_key_rows(self):
+        """Return the value of
+        :attr:`_engine.CursorResult.inserted_primary_key`
+        as a row contained within a list; some dialects may support a
+        multiple row form as well.
+
+        .. note:: As indicated below, in current SQLAlchemy versions this
+           accessor is only useful beyond what's already supplied by
+           :attr:`_engine.CursorResult.inserted_primary_key` when using the
+           :ref:`postgresql_psycopg2` dialect.   Future versions hope to
+           generalize this feature to more dialects.
+
+        This accessor is added to support dialects that offer the feature
+        that is currently implemented by the :ref:`psycopg2_executemany_mode`
+        feature, currently **only the psycopg2 dialect**, which provides
+        for many rows to be INSERTed at once while still retaining the
+        behavior of being able to return server-generated primary key values.
+
+        * **When using the psycopg2 dialect, or other dialects that may support
+          "fast executemany" style inserts in upcoming releases** : When
+          invoking an INSERT statement while passing a list of rows as the
+          second argument to :meth:`_engine.Connection.execute`, this accessor
+          will then provide a list of rows, where each row contains the primary
+          key value for each row that was INSERTed.
+
+        * **When using all other dialects / backends that don't yet support
+          this feature**: This accessor is only useful for **single row INSERT
+          statements**, and returns the same information as that of the
+          :attr:`_engine.CursorResult.inserted_primary_key` within a
+          single-element list. When an INSERT statement is executed in
+          conjunction with a list of rows to be INSERTed, the list will contain
+          one row per row inserted in the statement, however it will contain
+          ``None`` for any server-generated values.
+
+        Future releases of SQLAlchemy will further generalize the
+        "fast execution helper" feature of psycopg2 to suit other dialects,
+        thus allowing this accessor to be of more general use.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_engine.CursorResult.inserted_primary_key`
+
+        """
+        if not self.context.compiled:
+            raise exc.InvalidRequestError(
+                "Statement is not a compiled expression construct."
+            )
+        elif not self.context.isinsert:
+            raise exc.InvalidRequestError(
+                "Statement is not an insert() expression construct."
+            )
+        elif self.context._is_explicit_returning:
+            raise exc.InvalidRequestError(
+                "Can't call inserted_primary_key "
+                "when returning() "
+                "is used."
+            )
+        return self.context.inserted_primary_key_rows
+
+    @property
+    def inserted_primary_key(self):
+        """Return the primary key for the row just inserted.
+
+        The return value is a :class:`_result.Row` object representing
+        a named tuple of primary key values in the order in which the
+        primary key columns are configured in the source
+        :class:`_schema.Table`.
+
+        .. versionchanged:: 1.4.8 - the
+           :attr:`_engine.CursorResult.inserted_primary_key`
+           value is now a named tuple via the :class:`_result.Row` class,
+           rather than a plain tuple.
+
+        This accessor only applies to single row :func:`_expression.insert`
+        constructs which did not explicitly specify
+        :meth:`_expression.Insert.returning`.    Support for multirow inserts,
+        while not yet available for most backends, would be accessed using
+        the :attr:`_engine.CursorResult.inserted_primary_key_rows` accessor.
+
+        Note that primary key columns which specify a server_default clause, or
+        otherwise do not qualify as "autoincrement" columns (see the notes at
+        :class:`_schema.Column`), and were generated using the database-side
+        default, will appear in this list as ``None`` unless the backend
+        supports "returning" and the insert statement executed with the
+        "implicit returning" enabled.
+
+        Raises :class:`~sqlalchemy.exc.InvalidRequestError` if the executed
+        statement is not a compiled expression construct
+        or is not an insert() construct.
+
+        """
+
+        if self.context.executemany:
+            raise exc.InvalidRequestError(
+                "This statement was an executemany call; if primary key "
+                "returning is supported, please "
+                "use .inserted_primary_key_rows."
+            )
+
+        ikp = self.inserted_primary_key_rows
+        if ikp:
+            return ikp[0]
+        else:
+            return None
+
+    def last_updated_params(self):
+        """Return the collection of updated parameters from this
+        execution.
+
+        Raises :class:`~sqlalchemy.exc.InvalidRequestError` if the executed
+        statement is not a compiled expression construct
+        or is not an update() construct.
+
+        """
+        if not self.context.compiled:
+            raise exc.InvalidRequestError(
+                "Statement is not a compiled expression construct."
+            )
+        elif not self.context.isupdate:
+            raise exc.InvalidRequestError(
+                "Statement is not an update() expression construct."
+            )
+        elif self.context.executemany:
+            return self.context.compiled_parameters
+        else:
+            return self.context.compiled_parameters[0]
+
+    def last_inserted_params(self):
+        """Return the collection of inserted parameters from this
+        execution.
+
+        Raises :class:`~sqlalchemy.exc.InvalidRequestError` if the executed
+        statement is not a compiled expression construct
+        or is not an insert() construct.
+
+        """
+        if not self.context.compiled:
+            raise exc.InvalidRequestError(
+                "Statement is not a compiled expression construct."
+            )
+        elif not self.context.isinsert:
+            raise exc.InvalidRequestError(
+                "Statement is not an insert() expression construct."
+            )
+        elif self.context.executemany:
+            return self.context.compiled_parameters
+        else:
+            return self.context.compiled_parameters[0]
+
+    @property
+    def returned_defaults_rows(self):
+        """Return a list of rows each containing the values of default
+        columns that were fetched using
+        the :meth:`.ValuesBase.return_defaults` feature.
+
+        The return value is a list of :class:`.Row` objects.
+
+        .. versionadded:: 1.4
+
+        """
+        return self.context.returned_default_rows
+
+    def splice_horizontally(self, other):
+        """Return a new :class:`.CursorResult` that "horizontally splices"
+        together the rows of this :class:`.CursorResult` with that of another
+        :class:`.CursorResult`.
+
+        .. tip::  This method is for the benefit of the SQLAlchemy ORM and is
+           not intended for general use.
+
+        "horizontally splices" means that for each row in the first and second
+        result sets, a new row that concatenates the two rows together is
+        produced, which then becomes the new row.  The incoming
+        :class:`.CursorResult` must have the identical number of rows.  It is
+        typically expected that the two result sets come from the same sort
+        order as well, as the result rows are spliced together based on their
+        position in the result.
+
+        The expected use case here is so that multiple INSERT..RETURNING
+        statements (which definitely need to be sorted) against different
+        tables can produce a single result that looks like a JOIN of those two
+        tables.
+
+        E.g.::
+
+            r1 = connection.execute(
+                users.insert().returning(
+                    users.c.user_name, users.c.user_id, sort_by_parameter_order=True
+                ),
+                user_values,
+            )
+
+            r2 = connection.execute(
+                addresses.insert().returning(
+                    addresses.c.address_id,
+                    addresses.c.address,
+                    addresses.c.user_id,
+                    sort_by_parameter_order=True,
+                ),
+                address_values,
+            )
+
+            rows = r1.splice_horizontally(r2).all()
+            assert rows == [
+                ("john", 1, 1, "foo@bar.com", 1),
+                ("jack", 2, 2, "bar@bat.com", 2),
+            ]
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :meth:`.CursorResult.splice_vertically`
+
+
+        """  # noqa: E501
+
+        clone = self._generate()
+        total_rows = [
+            tuple(r1) + tuple(r2)
+            for r1, r2 in zip(
+                list(self._raw_row_iterator()),
+                list(other._raw_row_iterator()),
+            )
+        ]
+
+        clone._metadata = clone._metadata._splice_horizontally(other._metadata)
+
+        clone.cursor_strategy = FullyBufferedCursorFetchStrategy(
+            None,
+            initial_buffer=total_rows,
+        )
+        clone._reset_memoizations()
+        return clone
+
+    def splice_vertically(self, other):
+        """Return a new :class:`.CursorResult` that "vertically splices",
+        i.e. "extends", the rows of this :class:`.CursorResult` with that of
+        another :class:`.CursorResult`.
+
+        .. tip::  This method is for the benefit of the SQLAlchemy ORM and is
+           not intended for general use.
+
+        "vertically splices" means the rows of the given result are appended to
+        the rows of this cursor result. The incoming :class:`.CursorResult`
+        must have rows that represent the identical list of columns in the
+        identical order as they are in this :class:`.CursorResult`.
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :meth:`.CursorResult.splice_horizontally`
+
+        """
+        clone = self._generate()
+        total_rows = list(self._raw_row_iterator()) + list(
+            other._raw_row_iterator()
+        )
+
+        clone.cursor_strategy = FullyBufferedCursorFetchStrategy(
+            None,
+            initial_buffer=total_rows,
+        )
+        clone._reset_memoizations()
+        return clone
+
+    def _rewind(self, rows):
+        """rewind this result back to the given rowset.
+
+        this is used internally for the case where an :class:`.Insert`
+        construct combines the use of
+        :meth:`.Insert.return_defaults` along with the
+        "supplemental columns" feature.
+
+        """
+
+        if self._echo:
+            self.context.connection._log_debug(
+                "CursorResult rewound %d row(s)", len(rows)
+            )
+
+        # the rows given are expected to be Row objects, so we
+        # have to clear out processors which have already run on these
+        # rows
+        self._metadata = cast(
+            CursorResultMetaData, self._metadata
+        )._remove_processors()
+
+        self.cursor_strategy = FullyBufferedCursorFetchStrategy(
+            None,
+            # TODO: if these are Row objects, can we save on not having to
+            # re-make new Row objects out of them a second time?  is that
+            # what's actually happening right now?  maybe look into this
+            initial_buffer=rows,
+        )
+        self._reset_memoizations()
+        return self
+
+    @property
+    def returned_defaults(self):
+        """Return the values of default columns that were fetched using
+        the :meth:`.ValuesBase.return_defaults` feature.
+
+        The value is an instance of :class:`.Row`, or ``None``
+        if :meth:`.ValuesBase.return_defaults` was not used or if the
+        backend does not support RETURNING.
+
+        .. seealso::
+
+            :meth:`.ValuesBase.return_defaults`
+
+        """
+
+        if self.context.executemany:
+            raise exc.InvalidRequestError(
+                "This statement was an executemany call; if return defaults "
+                "is supported, please use .returned_defaults_rows."
+            )
+
+        rows = self.context.returned_default_rows
+        if rows:
+            return rows[0]
+        else:
+            return None
+
+    def lastrow_has_defaults(self):
+        """Return ``lastrow_has_defaults()`` from the underlying
+        :class:`.ExecutionContext`.
+
+        See :class:`.ExecutionContext` for details.
+
+        """
+
+        return self.context.lastrow_has_defaults()
+
+    def postfetch_cols(self):
+        """Return ``postfetch_cols()`` from the underlying
+        :class:`.ExecutionContext`.
+
+        See :class:`.ExecutionContext` for details.
+
+        Raises :class:`~sqlalchemy.exc.InvalidRequestError` if the executed
+        statement is not a compiled expression construct
+        or is not an insert() or update() construct.
+
+        """
+
+        if not self.context.compiled:
+            raise exc.InvalidRequestError(
+                "Statement is not a compiled expression construct."
+            )
+        elif not self.context.isinsert and not self.context.isupdate:
+            raise exc.InvalidRequestError(
+                "Statement is not an insert() or update() "
+                "expression construct."
+            )
+        return self.context.postfetch_cols
+
+    def prefetch_cols(self):
+        """Return ``prefetch_cols()`` from the underlying
+        :class:`.ExecutionContext`.
+
+        See :class:`.ExecutionContext` for details.
+
+        Raises :class:`~sqlalchemy.exc.InvalidRequestError` if the executed
+        statement is not a compiled expression construct
+        or is not an insert() or update() construct.
+
+        """
+
+        if not self.context.compiled:
+            raise exc.InvalidRequestError(
+                "Statement is not a compiled expression construct."
+            )
+        elif not self.context.isinsert and not self.context.isupdate:
+            raise exc.InvalidRequestError(
+                "Statement is not an insert() or update() "
+                "expression construct."
+            )
+        return self.context.prefetch_cols
+
+    def supports_sane_rowcount(self):
+        """Return ``supports_sane_rowcount`` from the dialect.
+
+        See :attr:`_engine.CursorResult.rowcount` for background.
+
+        """
+
+        return self.dialect.supports_sane_rowcount
+
+    def supports_sane_multi_rowcount(self):
+        """Return ``supports_sane_multi_rowcount`` from the dialect.
+
+        See :attr:`_engine.CursorResult.rowcount` for background.
+
+        """
+
+        return self.dialect.supports_sane_multi_rowcount
+
+    @util.memoized_property
+    def rowcount(self) -> int:
+        """Return the 'rowcount' for this result.
+
+        The primary purpose of 'rowcount' is to report the number of rows
+        matched by the WHERE criterion of an UPDATE or DELETE statement
+        executed once (i.e. for a single parameter set), which may then be
+        compared to the number of rows expected to be updated or deleted as a
+        means of asserting data integrity.
+
+        This attribute is transferred from the ``cursor.rowcount`` attribute
+        of the DBAPI before the cursor is closed, to support DBAPIs that
+        don't make this value available after cursor close.   Some DBAPIs may
+        offer meaningful values for other kinds of statements, such as INSERT
+        and SELECT statements as well.  In order to retrieve ``cursor.rowcount``
+        for these statements, set the
+        :paramref:`.Connection.execution_options.preserve_rowcount`
+        execution option to True, which will cause the ``cursor.rowcount``
+        value to be unconditionally memoized before any results are returned
+        or the cursor is closed, regardless of statement type.
+
+        For cases where the DBAPI does not support rowcount for a particular
+        kind of statement and/or execution, the returned value will be ``-1``,
+        which is delivered directly from the DBAPI and is part of :pep:`249`.
+        All DBAPIs should support rowcount for single-parameter-set
+        UPDATE and DELETE statements, however.
+
+        .. note::
+
+           Notes regarding :attr:`_engine.CursorResult.rowcount`:
+
+
+           * This attribute returns the number of rows *matched*,
+             which is not necessarily the same as the number of rows
+             that were actually *modified*. For example, an UPDATE statement
+             may have no net change on a given row if the SET values
+             given are the same as those present in the row already.
+             Such a row would be matched but not modified.
+             On backends that feature both styles, such as MySQL,
+             rowcount is configured to return the match
+             count in all cases.
+
+           * :attr:`_engine.CursorResult.rowcount` in the default case is
+             *only* useful in conjunction with an UPDATE or DELETE statement,
+             and only with a single set of parameters. For other kinds of
+             statements, SQLAlchemy will not attempt to pre-memoize the value
+             unless the
+             :paramref:`.Connection.execution_options.preserve_rowcount`
+             execution option is used.  Note that contrary to :pep:`249`, many
+             DBAPIs do not support rowcount values for statements that are not
+             UPDATE or DELETE, particularly when rows are being returned which
+             are not fully pre-buffered.   DBAPIs that dont support rowcount
+             for a particular kind of statement should return the value ``-1``
+             for such statements.
+
+           * :attr:`_engine.CursorResult.rowcount` may not be meaningful
+             when executing a single statement with multiple parameter sets
+             (i.e. an :term:`executemany`). Most DBAPIs do not sum "rowcount"
+             values across multiple parameter sets and will return ``-1``
+             when accessed.
+
+           * SQLAlchemy's :ref:`engine_insertmanyvalues` feature does support
+             a correct population of :attr:`_engine.CursorResult.rowcount`
+             when the :paramref:`.Connection.execution_options.preserve_rowcount`
+             execution option is set to True.
+
+           * Statements that use RETURNING may not support rowcount, returning
+             a ``-1`` value instead.
+
+        .. seealso::
+
+            :ref:`tutorial_update_delete_rowcount` - in the :ref:`unified_tutorial`
+
+            :paramref:`.Connection.execution_options.preserve_rowcount`
+
+        """  # noqa: E501
+        try:
+            return self.context.rowcount
+        except BaseException as e:
+            self.cursor_strategy.handle_exception(self, self.cursor, e)
+            raise  # not called
+
+    @property
+    def lastrowid(self):
+        """Return the 'lastrowid' accessor on the DBAPI cursor.
+
+        This is a DBAPI specific method and is only functional
+        for those backends which support it, for statements
+        where it is appropriate.  It's behavior is not
+        consistent across backends.
+
+        Usage of this method is normally unnecessary when
+        using insert() expression constructs; the
+        :attr:`~CursorResult.inserted_primary_key` attribute provides a
+        tuple of primary key values for a newly inserted row,
+        regardless of database backend.
+
+        """
+        try:
+            return self.context.get_lastrowid()
+        except BaseException as e:
+            self.cursor_strategy.handle_exception(self, self.cursor, e)
+
+    @property
+    def returns_rows(self):
+        """True if this :class:`_engine.CursorResult` returns zero or more
+        rows.
+
+        I.e. if it is legal to call the methods
+        :meth:`_engine.CursorResult.fetchone`,
+        :meth:`_engine.CursorResult.fetchmany`
+        :meth:`_engine.CursorResult.fetchall`.
+
+        Overall, the value of :attr:`_engine.CursorResult.returns_rows` should
+        always be synonymous with whether or not the DBAPI cursor had a
+        ``.description`` attribute, indicating the presence of result columns,
+        noting that a cursor that returns zero rows still has a
+        ``.description`` if a row-returning statement was emitted.
+
+        This attribute should be True for all results that are against
+        SELECT statements, as well as for DML statements INSERT/UPDATE/DELETE
+        that use RETURNING.   For INSERT/UPDATE/DELETE statements that were
+        not using RETURNING, the value will usually be False, however
+        there are some dialect-specific exceptions to this, such as when
+        using the MSSQL / pyodbc dialect a SELECT is emitted inline in
+        order to retrieve an inserted primary key value.
+
+
+        """
+        return self._metadata.returns_rows
+
+    @property
+    def is_insert(self):
+        """True if this :class:`_engine.CursorResult` is the result
+        of a executing an expression language compiled
+        :func:`_expression.insert` construct.
+
+        When True, this implies that the
+        :attr:`inserted_primary_key` attribute is accessible,
+        assuming the statement did not include
+        a user defined "returning" construct.
+
+        """
+        return self.context.isinsert
+
+    def _fetchiter_impl(self):
+        fetchone = self.cursor_strategy.fetchone
+
+        while True:
+            row = fetchone(self, self.cursor)
+            if row is None:
+                break
+            yield row
+
+    def _fetchone_impl(self, hard_close=False):
+        return self.cursor_strategy.fetchone(self, self.cursor, hard_close)
+
+    def _fetchall_impl(self):
+        return self.cursor_strategy.fetchall(self, self.cursor)
+
+    def _fetchmany_impl(self, size=None):
+        return self.cursor_strategy.fetchmany(self, self.cursor, size)
+
+    def _raw_row_iterator(self):
+        return self._fetchiter_impl()
+
+    def merge(self, *others: Result[Any]) -> MergedResult[Any]:
+        merged_result = super().merge(*others)
+        if self.context._has_rowcount:
+            merged_result.rowcount = sum(
+                cast("CursorResult[Any]", result).rowcount
+                for result in (self,) + others
+            )
+        return merged_result
+
+    def close(self) -> Any:
+        """Close this :class:`_engine.CursorResult`.
+
+        This closes out the underlying DBAPI cursor corresponding to the
+        statement execution, if one is still present.  Note that the DBAPI
+        cursor is automatically released when the :class:`_engine.CursorResult`
+        exhausts all available rows.  :meth:`_engine.CursorResult.close` is
+        generally an optional method except in the case when discarding a
+        :class:`_engine.CursorResult` that still has additional rows pending
+        for fetch.
+
+        After this method is called, it is no longer valid to call upon
+        the fetch methods, which will raise a :class:`.ResourceClosedError`
+        on subsequent use.
+
+        .. seealso::
+
+            :ref:`connections_toplevel`
+
+        """
+        self._soft_close(hard=True)
+
+    @_generative
+    def yield_per(self, num: int) -> Self:
+        self._yield_per = num
+        self.cursor_strategy.yield_per(self, self.cursor, num)
+        return self
+
+
+ResultProxy = CursorResult
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/default.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/default.py
new file mode 100644
index 00000000..dd4250ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/default.py
@@ -0,0 +1,2367 @@
+# engine/default.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Default implementations of per-dialect sqlalchemy.engine classes.
+
+These are semi-private implementation classes which are only of importance
+to database dialect authors; dialects will usually use the classes here
+as the base class for their own corresponding classes.
+
+"""
+
+from __future__ import annotations
+
+import functools
+import operator
+import random
+import re
+from time import perf_counter
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import MutableSequence
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+import weakref
+
+from . import characteristics
+from . import cursor as _cursor
+from . import interfaces
+from .base import Connection
+from .interfaces import CacheStats
+from .interfaces import DBAPICursor
+from .interfaces import Dialect
+from .interfaces import ExecuteStyle
+from .interfaces import ExecutionContext
+from .reflection import ObjectKind
+from .reflection import ObjectScope
+from .. import event
+from .. import exc
+from .. import pool
+from .. import util
+from ..sql import compiler
+from ..sql import dml
+from ..sql import expression
+from ..sql import type_api
+from ..sql import util as sql_util
+from ..sql._typing import is_tuple_type
+from ..sql.base import _NoArg
+from ..sql.compiler import DDLCompiler
+from ..sql.compiler import InsertmanyvaluesSentinelOpts
+from ..sql.compiler import SQLCompiler
+from ..sql.elements import quoted_name
+from ..util.typing import Final
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from types import ModuleType
+
+    from .base import Engine
+    from .cursor import ResultFetchStrategy
+    from .interfaces import _CoreMultiExecuteParams
+    from .interfaces import _CoreSingleExecuteParams
+    from .interfaces import _DBAPICursorDescription
+    from .interfaces import _DBAPIMultiExecuteParams
+    from .interfaces import _ExecuteOptions
+    from .interfaces import _MutableCoreSingleExecuteParams
+    from .interfaces import _ParamStyle
+    from .interfaces import DBAPIConnection
+    from .interfaces import IsolationLevel
+    from .row import Row
+    from .url import URL
+    from ..event import _ListenerFnType
+    from ..pool import Pool
+    from ..pool import PoolProxiedConnection
+    from ..sql import Executable
+    from ..sql.compiler import Compiled
+    from ..sql.compiler import Linting
+    from ..sql.compiler import ResultColumnsEntry
+    from ..sql.dml import DMLState
+    from ..sql.dml import UpdateBase
+    from ..sql.elements import BindParameter
+    from ..sql.schema import Column
+    from ..sql.type_api import _BindProcessorType
+    from ..sql.type_api import _ResultProcessorType
+    from ..sql.type_api import TypeEngine
+
+# When we're handed literal SQL, ensure it's a SELECT query
+SERVER_SIDE_CURSOR_RE = re.compile(r"\s*SELECT", re.I | re.UNICODE)
+
+
+(
+    CACHE_HIT,
+    CACHE_MISS,
+    CACHING_DISABLED,
+    NO_CACHE_KEY,
+    NO_DIALECT_SUPPORT,
+) = list(CacheStats)
+
+
+class DefaultDialect(Dialect):
+    """Default implementation of Dialect"""
+
+    statement_compiler = compiler.SQLCompiler
+    ddl_compiler = compiler.DDLCompiler
+    type_compiler_cls = compiler.GenericTypeCompiler
+
+    preparer = compiler.IdentifierPreparer
+    supports_alter = True
+    supports_comments = False
+    supports_constraint_comments = False
+    inline_comments = False
+    supports_statement_cache = True
+
+    div_is_floordiv = True
+
+    bind_typing = interfaces.BindTyping.NONE
+
+    include_set_input_sizes: Optional[Set[Any]] = None
+    exclude_set_input_sizes: Optional[Set[Any]] = None
+
+    # the first value we'd get for an autoincrement column.
+    default_sequence_base = 1
+
+    # most DBAPIs happy with this for execute().
+    # not cx_oracle.
+    execute_sequence_format = tuple
+
+    supports_schemas = True
+    supports_views = True
+    supports_sequences = False
+    sequences_optional = False
+    preexecute_autoincrement_sequences = False
+    supports_identity_columns = False
+    postfetch_lastrowid = True
+    favor_returning_over_lastrowid = False
+    insert_null_pk_still_autoincrements = False
+    update_returning = False
+    delete_returning = False
+    update_returning_multifrom = False
+    delete_returning_multifrom = False
+    insert_returning = False
+
+    cte_follows_insert = False
+
+    supports_native_enum = False
+    supports_native_boolean = False
+    supports_native_uuid = False
+    returns_native_bytes = False
+
+    non_native_boolean_check_constraint = True
+
+    supports_simple_order_by_label = True
+
+    tuple_in_values = False
+
+    connection_characteristics = util.immutabledict(
+        {
+            "isolation_level": characteristics.IsolationLevelCharacteristic(),
+            "logging_token": characteristics.LoggingTokenCharacteristic(),
+        }
+    )
+
+    engine_config_types: Mapping[str, Any] = util.immutabledict(
+        {
+            "pool_timeout": util.asint,
+            "echo": util.bool_or_str("debug"),
+            "echo_pool": util.bool_or_str("debug"),
+            "pool_recycle": util.asint,
+            "pool_size": util.asint,
+            "max_overflow": util.asint,
+            "future": util.asbool,
+        }
+    )
+
+    # if the NUMERIC type
+    # returns decimal.Decimal.
+    # *not* the FLOAT type however.
+    supports_native_decimal = False
+
+    name = "default"
+
+    # length at which to truncate
+    # any identifier.
+    max_identifier_length = 9999
+    _user_defined_max_identifier_length: Optional[int] = None
+
+    isolation_level: Optional[str] = None
+
+    # sub-categories of max_identifier_length.
+    # currently these accommodate for MySQL which allows alias names
+    # of 255 but DDL names only of 64.
+    max_index_name_length: Optional[int] = None
+    max_constraint_name_length: Optional[int] = None
+
+    supports_sane_rowcount = True
+    supports_sane_multi_rowcount = True
+    colspecs: MutableMapping[Type[TypeEngine[Any]], Type[TypeEngine[Any]]] = {}
+    default_paramstyle = "named"
+
+    supports_default_values = False
+    """dialect supports INSERT... DEFAULT VALUES syntax"""
+
+    supports_default_metavalue = False
+    """dialect supports INSERT... VALUES (DEFAULT) syntax"""
+
+    default_metavalue_token = "DEFAULT"
+    """for INSERT... VALUES (DEFAULT) syntax, the token to put in the
+    parenthesis."""
+
+    # not sure if this is a real thing but the compiler will deliver it
+    # if this is the only flag enabled.
+    supports_empty_insert = True
+    """dialect supports INSERT () VALUES ()"""
+
+    supports_multivalues_insert = False
+
+    use_insertmanyvalues: bool = False
+
+    use_insertmanyvalues_wo_returning: bool = False
+
+    insertmanyvalues_implicit_sentinel: InsertmanyvaluesSentinelOpts = (
+        InsertmanyvaluesSentinelOpts.NOT_SUPPORTED
+    )
+
+    insertmanyvalues_page_size: int = 1000
+    insertmanyvalues_max_parameters = 32700
+
+    supports_is_distinct_from = True
+
+    supports_server_side_cursors = False
+
+    server_side_cursors = False
+
+    # extra record-level locking features (#4860)
+    supports_for_update_of = False
+
+    server_version_info = None
+
+    default_schema_name: Optional[str] = None
+
+    # indicates symbol names are
+    # UPPERCASED if they are case insensitive
+    # within the database.
+    # if this is True, the methods normalize_name()
+    # and denormalize_name() must be provided.
+    requires_name_normalize = False
+
+    is_async = False
+
+    has_terminate = False
+
+    # TODO: this is not to be part of 2.0.  implement rudimentary binary
+    # literals for SQLite, PostgreSQL, MySQL only within
+    # _Binary.literal_processor
+    _legacy_binary_type_literal_encoding = "utf-8"
+
+    @util.deprecated_params(
+        empty_in_strategy=(
+            "1.4",
+            "The :paramref:`_sa.create_engine.empty_in_strategy` keyword is "
+            "deprecated, and no longer has any effect.  All IN expressions "
+            "are now rendered using "
+            'the "expanding parameter" strategy which renders a set of bound'
+            'expressions, or an "empty set" SELECT, at statement execution'
+            "time.",
+        ),
+        server_side_cursors=(
+            "1.4",
+            "The :paramref:`_sa.create_engine.server_side_cursors` parameter "
+            "is deprecated and will be removed in a future release.  Please "
+            "use the "
+            ":paramref:`_engine.Connection.execution_options.stream_results` "
+            "parameter.",
+        ),
+    )
+    def __init__(
+        self,
+        paramstyle: Optional[_ParamStyle] = None,
+        isolation_level: Optional[IsolationLevel] = None,
+        dbapi: Optional[ModuleType] = None,
+        implicit_returning: Literal[True] = True,
+        supports_native_boolean: Optional[bool] = None,
+        max_identifier_length: Optional[int] = None,
+        label_length: Optional[int] = None,
+        insertmanyvalues_page_size: Union[_NoArg, int] = _NoArg.NO_ARG,
+        use_insertmanyvalues: Optional[bool] = None,
+        # util.deprecated_params decorator cannot render the
+        # Linting.NO_LINTING constant
+        compiler_linting: Linting = int(compiler.NO_LINTING),  # type: ignore
+        server_side_cursors: bool = False,
+        **kwargs: Any,
+    ):
+        if server_side_cursors:
+            if not self.supports_server_side_cursors:
+                raise exc.ArgumentError(
+                    "Dialect %s does not support server side cursors" % self
+                )
+            else:
+                self.server_side_cursors = True
+
+        if getattr(self, "use_setinputsizes", False):
+            util.warn_deprecated(
+                "The dialect-level use_setinputsizes attribute is "
+                "deprecated.  Please use "
+                "bind_typing = BindTyping.SETINPUTSIZES",
+                "2.0",
+            )
+            self.bind_typing = interfaces.BindTyping.SETINPUTSIZES
+
+        self.positional = False
+        self._ischema = None
+
+        self.dbapi = dbapi
+
+        if paramstyle is not None:
+            self.paramstyle = paramstyle
+        elif self.dbapi is not None:
+            self.paramstyle = self.dbapi.paramstyle
+        else:
+            self.paramstyle = self.default_paramstyle
+        self.positional = self.paramstyle in (
+            "qmark",
+            "format",
+            "numeric",
+            "numeric_dollar",
+        )
+        self.identifier_preparer = self.preparer(self)
+        self._on_connect_isolation_level = isolation_level
+
+        legacy_tt_callable = getattr(self, "type_compiler", None)
+        if legacy_tt_callable is not None:
+            tt_callable = cast(
+                Type[compiler.GenericTypeCompiler],
+                self.type_compiler,
+            )
+        else:
+            tt_callable = self.type_compiler_cls
+
+        self.type_compiler_instance = self.type_compiler = tt_callable(self)
+
+        if supports_native_boolean is not None:
+            self.supports_native_boolean = supports_native_boolean
+
+        self._user_defined_max_identifier_length = max_identifier_length
+        if self._user_defined_max_identifier_length:
+            self.max_identifier_length = (
+                self._user_defined_max_identifier_length
+            )
+        self.label_length = label_length
+        self.compiler_linting = compiler_linting
+
+        if use_insertmanyvalues is not None:
+            self.use_insertmanyvalues = use_insertmanyvalues
+
+        if insertmanyvalues_page_size is not _NoArg.NO_ARG:
+            self.insertmanyvalues_page_size = insertmanyvalues_page_size
+
+    @property
+    @util.deprecated(
+        "2.0",
+        "full_returning is deprecated, please use insert_returning, "
+        "update_returning, delete_returning",
+    )
+    def full_returning(self):
+        return (
+            self.insert_returning
+            and self.update_returning
+            and self.delete_returning
+        )
+
+    @util.memoized_property
+    def insert_executemany_returning(self):
+        """Default implementation for insert_executemany_returning, if not
+        otherwise overridden by the specific dialect.
+
+        The default dialect determines "insert_executemany_returning" is
+        available if the dialect in use has opted into using the
+        "use_insertmanyvalues" feature. If they haven't opted into that, then
+        this attribute is False, unless the dialect in question overrides this
+        and provides some other implementation (such as the Oracle Database
+        dialects).
+
+        """
+        return self.insert_returning and self.use_insertmanyvalues
+
+    @util.memoized_property
+    def insert_executemany_returning_sort_by_parameter_order(self):
+        """Default implementation for
+        insert_executemany_returning_deterministic_order, if not otherwise
+        overridden by the specific dialect.
+
+        The default dialect determines "insert_executemany_returning" can have
+        deterministic order only if the dialect in use has opted into using the
+        "use_insertmanyvalues" feature, which implements deterministic ordering
+        using client side sentinel columns only by default.  The
+        "insertmanyvalues" feature also features alternate forms that can
+        use server-generated PK values as "sentinels", but those are only
+        used if the :attr:`.Dialect.insertmanyvalues_implicit_sentinel`
+        bitflag enables those alternate SQL forms, which are disabled
+        by default.
+
+        If the dialect in use hasn't opted into that, then this attribute is
+        False, unless the dialect in question overrides this and provides some
+        other implementation (such as the Oracle Database dialects).
+
+        """
+        return self.insert_returning and self.use_insertmanyvalues
+
+    update_executemany_returning = False
+    delete_executemany_returning = False
+
+    @util.memoized_property
+    def loaded_dbapi(self) -> ModuleType:
+        if self.dbapi is None:
+            raise exc.InvalidRequestError(
+                f"Dialect {self} does not have a Python DBAPI established "
+                "and cannot be used for actual database interaction"
+            )
+        return self.dbapi
+
+    @util.memoized_property
+    def _bind_typing_render_casts(self):
+        return self.bind_typing is interfaces.BindTyping.RENDER_CASTS
+
+    def _ensure_has_table_connection(self, arg):
+        if not isinstance(arg, Connection):
+            raise exc.ArgumentError(
+                "The argument passed to Dialect.has_table() should be a "
+                "%s, got %s. "
+                "Additionally, the Dialect.has_table() method is for "
+                "internal dialect "
+                "use only; please use "
+                "``inspect(some_engine).has_table(<tablename>>)`` "
+                "for public API use." % (Connection, type(arg))
+            )
+
+    @util.memoized_property
+    def _supports_statement_cache(self):
+        ssc = self.__class__.__dict__.get("supports_statement_cache", None)
+        if ssc is None:
+            util.warn(
+                "Dialect %s:%s will not make use of SQL compilation caching "
+                "as it does not set the 'supports_statement_cache' attribute "
+                "to ``True``.  This can have "
+                "significant performance implications including some "
+                "performance degradations in comparison to prior SQLAlchemy "
+                "versions.  Dialect maintainers should seek to set this "
+                "attribute to True after appropriate development and testing "
+                "for SQLAlchemy 1.4 caching support.   Alternatively, this "
+                "attribute may be set to False which will disable this "
+                "warning." % (self.name, self.driver),
+                code="cprf",
+            )
+
+        return bool(ssc)
+
+    @util.memoized_property
+    def _type_memos(self):
+        return weakref.WeakKeyDictionary()
+
+    @property
+    def dialect_description(self):
+        return self.name + "+" + self.driver
+
+    @property
+    def supports_sane_rowcount_returning(self):
+        """True if this dialect supports sane rowcount even if RETURNING is
+        in use.
+
+        For dialects that don't support RETURNING, this is synonymous with
+        ``supports_sane_rowcount``.
+
+        """
+        return self.supports_sane_rowcount
+
+    @classmethod
+    def get_pool_class(cls, url: URL) -> Type[Pool]:
+        return getattr(cls, "poolclass", pool.QueuePool)
+
+    def get_dialect_pool_class(self, url: URL) -> Type[Pool]:
+        return self.get_pool_class(url)
+
+    @classmethod
+    def load_provisioning(cls):
+        package = ".".join(cls.__module__.split(".")[0:-1])
+        try:
+            __import__(package + ".provision")
+        except ImportError:
+            pass
+
+    def _builtin_onconnect(self) -> Optional[_ListenerFnType]:
+        if self._on_connect_isolation_level is not None:
+
+            def builtin_connect(dbapi_conn, conn_rec):
+                self._assert_and_set_isolation_level(
+                    dbapi_conn, self._on_connect_isolation_level
+                )
+
+            return builtin_connect
+        else:
+            return None
+
+    def initialize(self, connection):
+        try:
+            self.server_version_info = self._get_server_version_info(
+                connection
+            )
+        except NotImplementedError:
+            self.server_version_info = None
+        try:
+            self.default_schema_name = self._get_default_schema_name(
+                connection
+            )
+        except NotImplementedError:
+            self.default_schema_name = None
+
+        try:
+            self.default_isolation_level = self.get_default_isolation_level(
+                connection.connection.dbapi_connection
+            )
+        except NotImplementedError:
+            self.default_isolation_level = None
+
+        if not self._user_defined_max_identifier_length:
+            max_ident_length = self._check_max_identifier_length(connection)
+            if max_ident_length:
+                self.max_identifier_length = max_ident_length
+
+        if (
+            self.label_length
+            and self.label_length > self.max_identifier_length
+        ):
+            raise exc.ArgumentError(
+                "Label length of %d is greater than this dialect's"
+                " maximum identifier length of %d"
+                % (self.label_length, self.max_identifier_length)
+            )
+
+    def on_connect(self):
+        # inherits the docstring from interfaces.Dialect.on_connect
+        return None
+
+    def _check_max_identifier_length(self, connection):
+        """Perform a connection / server version specific check to determine
+        the max_identifier_length.
+
+        If the dialect's class level max_identifier_length should be used,
+        can return None.
+
+        .. versionadded:: 1.3.9
+
+        """
+        return None
+
+    def get_default_isolation_level(self, dbapi_conn):
+        """Given a DBAPI connection, return its isolation level, or
+        a default isolation level if one cannot be retrieved.
+
+        May be overridden by subclasses in order to provide a
+        "fallback" isolation level for databases that cannot reliably
+        retrieve the actual isolation level.
+
+        By default, calls the :meth:`_engine.Interfaces.get_isolation_level`
+        method, propagating any exceptions raised.
+
+        .. versionadded:: 1.3.22
+
+        """
+        return self.get_isolation_level(dbapi_conn)
+
+    def type_descriptor(self, typeobj):
+        """Provide a database-specific :class:`.TypeEngine` object, given
+        the generic object which comes from the types module.
+
+        This method looks for a dictionary called
+        ``colspecs`` as a class or instance-level variable,
+        and passes on to :func:`_types.adapt_type`.
+
+        """
+        return type_api.adapt_type(typeobj, self.colspecs)
+
+    def has_index(self, connection, table_name, index_name, schema=None, **kw):
+        if not self.has_table(connection, table_name, schema=schema, **kw):
+            return False
+        for idx in self.get_indexes(
+            connection, table_name, schema=schema, **kw
+        ):
+            if idx["name"] == index_name:
+                return True
+        else:
+            return False
+
+    def has_schema(
+        self, connection: Connection, schema_name: str, **kw: Any
+    ) -> bool:
+        return schema_name in self.get_schema_names(connection, **kw)
+
+    def validate_identifier(self, ident):
+        if len(ident) > self.max_identifier_length:
+            raise exc.IdentifierError(
+                "Identifier '%s' exceeds maximum length of %d characters"
+                % (ident, self.max_identifier_length)
+            )
+
+    def connect(self, *cargs, **cparams):
+        # inherits the docstring from interfaces.Dialect.connect
+        return self.loaded_dbapi.connect(*cargs, **cparams)
+
+    def create_connect_args(self, url):
+        # inherits the docstring from interfaces.Dialect.create_connect_args
+        opts = url.translate_connect_args()
+        opts.update(url.query)
+        return ([], opts)
+
+    def set_engine_execution_options(
+        self, engine: Engine, opts: Mapping[str, Any]
+    ) -> None:
+        supported_names = set(self.connection_characteristics).intersection(
+            opts
+        )
+        if supported_names:
+            characteristics: Mapping[str, Any] = util.immutabledict(
+                (name, opts[name]) for name in supported_names
+            )
+
+            @event.listens_for(engine, "engine_connect")
+            def set_connection_characteristics(connection):
+                self._set_connection_characteristics(
+                    connection, characteristics
+                )
+
+    def set_connection_execution_options(
+        self, connection: Connection, opts: Mapping[str, Any]
+    ) -> None:
+        supported_names = set(self.connection_characteristics).intersection(
+            opts
+        )
+        if supported_names:
+            characteristics: Mapping[str, Any] = util.immutabledict(
+                (name, opts[name]) for name in supported_names
+            )
+            self._set_connection_characteristics(connection, characteristics)
+
+    def _set_connection_characteristics(self, connection, characteristics):
+        characteristic_values = [
+            (name, self.connection_characteristics[name], value)
+            for name, value in characteristics.items()
+        ]
+
+        if connection.in_transaction():
+            trans_objs = [
+                (name, obj)
+                for name, obj, _ in characteristic_values
+                if obj.transactional
+            ]
+            if trans_objs:
+                raise exc.InvalidRequestError(
+                    "This connection has already initialized a SQLAlchemy "
+                    "Transaction() object via begin() or autobegin; "
+                    "%s may not be altered unless rollback() or commit() "
+                    "is called first."
+                    % (", ".join(name for name, obj in trans_objs))
+                )
+
+        dbapi_connection = connection.connection.dbapi_connection
+        for _, characteristic, value in characteristic_values:
+            characteristic.set_connection_characteristic(
+                self, connection, dbapi_connection, value
+            )
+        connection.connection._connection_record.finalize_callback.append(
+            functools.partial(self._reset_characteristics, characteristics)
+        )
+
+    def _reset_characteristics(self, characteristics, dbapi_connection):
+        for characteristic_name in characteristics:
+            characteristic = self.connection_characteristics[
+                characteristic_name
+            ]
+            characteristic.reset_characteristic(self, dbapi_connection)
+
+    def do_begin(self, dbapi_connection):
+        pass
+
+    def do_rollback(self, dbapi_connection):
+        dbapi_connection.rollback()
+
+    def do_commit(self, dbapi_connection):
+        dbapi_connection.commit()
+
+    def do_terminate(self, dbapi_connection):
+        self.do_close(dbapi_connection)
+
+    def do_close(self, dbapi_connection):
+        dbapi_connection.close()
+
+    @util.memoized_property
+    def _dialect_specific_select_one(self):
+        return str(expression.select(1).compile(dialect=self))
+
+    def _do_ping_w_event(self, dbapi_connection: DBAPIConnection) -> bool:
+        try:
+            return self.do_ping(dbapi_connection)
+        except self.loaded_dbapi.Error as err:
+            is_disconnect = self.is_disconnect(err, dbapi_connection, None)
+
+            if self._has_events:
+                try:
+                    Connection._handle_dbapi_exception_noconnection(
+                        err,
+                        self,
+                        is_disconnect=is_disconnect,
+                        invalidate_pool_on_disconnect=False,
+                        is_pre_ping=True,
+                    )
+                except exc.StatementError as new_err:
+                    is_disconnect = new_err.connection_invalidated
+
+            if is_disconnect:
+                return False
+            else:
+                raise
+
+    def do_ping(self, dbapi_connection: DBAPIConnection) -> bool:
+        cursor = None
+
+        cursor = dbapi_connection.cursor()
+        try:
+            cursor.execute(self._dialect_specific_select_one)
+        finally:
+            cursor.close()
+        return True
+
+    def create_xid(self):
+        """Create a random two-phase transaction ID.
+
+        This id will be passed to do_begin_twophase(), do_rollback_twophase(),
+        do_commit_twophase().  Its format is unspecified.
+        """
+
+        return "_sa_%032x" % random.randint(0, 2**128)
+
+    def do_savepoint(self, connection, name):
+        connection.execute(expression.SavepointClause(name))
+
+    def do_rollback_to_savepoint(self, connection, name):
+        connection.execute(expression.RollbackToSavepointClause(name))
+
+    def do_release_savepoint(self, connection, name):
+        connection.execute(expression.ReleaseSavepointClause(name))
+
+    def _deliver_insertmanyvalues_batches(
+        self,
+        connection,
+        cursor,
+        statement,
+        parameters,
+        generic_setinputsizes,
+        context,
+    ):
+        context = cast(DefaultExecutionContext, context)
+        compiled = cast(SQLCompiler, context.compiled)
+
+        _composite_sentinel_proc: Sequence[
+            Optional[_ResultProcessorType[Any]]
+        ] = ()
+        _scalar_sentinel_proc: Optional[_ResultProcessorType[Any]] = None
+        _sentinel_proc_initialized: bool = False
+
+        compiled_parameters = context.compiled_parameters
+
+        imv = compiled._insertmanyvalues
+        assert imv is not None
+
+        is_returning: Final[bool] = bool(compiled.effective_returning)
+        batch_size = context.execution_options.get(
+            "insertmanyvalues_page_size", self.insertmanyvalues_page_size
+        )
+
+        if compiled.schema_translate_map:
+            schema_translate_map = context.execution_options.get(
+                "schema_translate_map", {}
+            )
+        else:
+            schema_translate_map = None
+
+        if is_returning:
+            result: Optional[List[Any]] = []
+            context._insertmanyvalues_rows = result
+
+            sort_by_parameter_order = imv.sort_by_parameter_order
+
+        else:
+            sort_by_parameter_order = False
+            result = None
+
+        for imv_batch in compiled._deliver_insertmanyvalues_batches(
+            statement,
+            parameters,
+            compiled_parameters,
+            generic_setinputsizes,
+            batch_size,
+            sort_by_parameter_order,
+            schema_translate_map,
+        ):
+            yield imv_batch
+
+            if is_returning:
+
+                try:
+                    rows = context.fetchall_for_returning(cursor)
+                except BaseException as be:
+                    connection._handle_dbapi_exception(
+                        be,
+                        sql_util._long_statement(imv_batch.replaced_statement),
+                        imv_batch.replaced_parameters,
+                        None,
+                        context,
+                        is_sub_exec=True,
+                    )
+
+                # I would have thought "is_returning: Final[bool]"
+                # would have assured this but pylance thinks not
+                assert result is not None
+
+                if imv.num_sentinel_columns and not imv_batch.is_downgraded:
+                    composite_sentinel = imv.num_sentinel_columns > 1
+                    if imv.implicit_sentinel:
+                        # for implicit sentinel, which is currently single-col
+                        # integer autoincrement, do a simple sort.
+                        assert not composite_sentinel
+                        result.extend(
+                            sorted(rows, key=operator.itemgetter(-1))
+                        )
+                        continue
+
+                    # otherwise, create dictionaries to match up batches
+                    # with parameters
+                    assert imv.sentinel_param_keys
+                    assert imv.sentinel_columns
+
+                    _nsc = imv.num_sentinel_columns
+
+                    if not _sentinel_proc_initialized:
+                        if composite_sentinel:
+                            _composite_sentinel_proc = [
+                                col.type._cached_result_processor(
+                                    self, cursor_desc[1]
+                                )
+                                for col, cursor_desc in zip(
+                                    imv.sentinel_columns,
+                                    cursor.description[-_nsc:],
+                                )
+                            ]
+                        else:
+                            _scalar_sentinel_proc = (
+                                imv.sentinel_columns[0]
+                            ).type._cached_result_processor(
+                                self, cursor.description[-1][1]
+                            )
+                        _sentinel_proc_initialized = True
+
+                    rows_by_sentinel: Union[
+                        Dict[Tuple[Any, ...], Any],
+                        Dict[Any, Any],
+                    ]
+                    if composite_sentinel:
+                        rows_by_sentinel = {
+                            tuple(
+                                (proc(val) if proc else val)
+                                for val, proc in zip(
+                                    row[-_nsc:], _composite_sentinel_proc
+                                )
+                            ): row
+                            for row in rows
+                        }
+                    elif _scalar_sentinel_proc:
+                        rows_by_sentinel = {
+                            _scalar_sentinel_proc(row[-1]): row for row in rows
+                        }
+                    else:
+                        rows_by_sentinel = {row[-1]: row for row in rows}
+
+                    if len(rows_by_sentinel) != len(imv_batch.batch):
+                        # see test_insert_exec.py::
+                        # IMVSentinelTest::test_sentinel_incorrect_rowcount
+                        # for coverage / demonstration
+                        raise exc.InvalidRequestError(
+                            f"Sentinel-keyed result set did not produce "
+                            f"correct number of rows {len(imv_batch.batch)}; "
+                            "produced "
+                            f"{len(rows_by_sentinel)}.  Please ensure the "
+                            "sentinel column is fully unique and populated in "
+                            "all cases."
+                        )
+
+                    try:
+                        ordered_rows = [
+                            rows_by_sentinel[sentinel_keys]
+                            for sentinel_keys in imv_batch.sentinel_values
+                        ]
+                    except KeyError as ke:
+                        # see test_insert_exec.py::
+                        # IMVSentinelTest::test_sentinel_cant_match_keys
+                        # for coverage / demonstration
+                        raise exc.InvalidRequestError(
+                            f"Can't match sentinel values in result set to "
+                            f"parameter sets; key {ke.args[0]!r} was not "
+                            "found. "
+                            "There may be a mismatch between the datatype "
+                            "passed to the DBAPI driver vs. that which it "
+                            "returns in a result row.  Ensure the given "
+                            "Python value matches the expected result type "
+                            "*exactly*, taking care to not rely upon implicit "
+                            "conversions which may occur such as when using "
+                            "strings in place of UUID or integer values, etc. "
+                        ) from ke
+
+                    result.extend(ordered_rows)
+
+                else:
+                    result.extend(rows)
+
+    def do_executemany(self, cursor, statement, parameters, context=None):
+        cursor.executemany(statement, parameters)
+
+    def do_execute(self, cursor, statement, parameters, context=None):
+        cursor.execute(statement, parameters)
+
+    def do_execute_no_params(self, cursor, statement, context=None):
+        cursor.execute(statement)
+
+    def is_disconnect(self, e, connection, cursor):
+        return False
+
+    @util.memoized_instancemethod
+    def _gen_allowed_isolation_levels(self, dbapi_conn):
+        try:
+            raw_levels = list(self.get_isolation_level_values(dbapi_conn))
+        except NotImplementedError:
+            return None
+        else:
+            normalized_levels = [
+                level.replace("_", " ").upper() for level in raw_levels
+            ]
+            if raw_levels != normalized_levels:
+                raise ValueError(
+                    f"Dialect {self.name!r} get_isolation_level_values() "
+                    f"method should return names as UPPERCASE using spaces, "
+                    f"not underscores; got "
+                    f"{sorted(set(raw_levels).difference(normalized_levels))}"
+                )
+            return tuple(normalized_levels)
+
+    def _assert_and_set_isolation_level(self, dbapi_conn, level):
+        level = level.replace("_", " ").upper()
+
+        _allowed_isolation_levels = self._gen_allowed_isolation_levels(
+            dbapi_conn
+        )
+        if (
+            _allowed_isolation_levels
+            and level not in _allowed_isolation_levels
+        ):
+            raise exc.ArgumentError(
+                f"Invalid value {level!r} for isolation_level. "
+                f"Valid isolation levels for {self.name!r} are "
+                f"{', '.join(_allowed_isolation_levels)}"
+            )
+
+        self.set_isolation_level(dbapi_conn, level)
+
+    def reset_isolation_level(self, dbapi_conn):
+        if self._on_connect_isolation_level is not None:
+            assert (
+                self._on_connect_isolation_level == "AUTOCOMMIT"
+                or self._on_connect_isolation_level
+                == self.default_isolation_level
+            )
+            self._assert_and_set_isolation_level(
+                dbapi_conn, self._on_connect_isolation_level
+            )
+        else:
+            assert self.default_isolation_level is not None
+            self._assert_and_set_isolation_level(
+                dbapi_conn,
+                self.default_isolation_level,
+            )
+
+    def normalize_name(self, name):
+        if name is None:
+            return None
+
+        name_lower = name.lower()
+        name_upper = name.upper()
+
+        if name_upper == name_lower:
+            # name has no upper/lower conversion, e.g. non-european characters.
+            # return unchanged
+            return name
+        elif name_upper == name and not (
+            self.identifier_preparer._requires_quotes
+        )(name_lower):
+            # name is all uppercase and doesn't require quoting; normalize
+            # to all lower case
+            return name_lower
+        elif name_lower == name:
+            # name is all lower case, which if denormalized means we need to
+            # force quoting on it
+            return quoted_name(name, quote=True)
+        else:
+            # name is mixed case, means it will be quoted in SQL when used
+            # later, no normalizes
+            return name
+
+    def denormalize_name(self, name):
+        if name is None:
+            return None
+
+        name_lower = name.lower()
+        name_upper = name.upper()
+
+        if name_upper == name_lower:
+            # name has no upper/lower conversion, e.g. non-european characters.
+            # return unchanged
+            return name
+        elif name_lower == name and not (
+            self.identifier_preparer._requires_quotes
+        )(name_lower):
+            name = name_upper
+        return name
+
+    def get_driver_connection(self, connection):
+        return connection
+
+    def _overrides_default(self, method):
+        return (
+            getattr(type(self), method).__code__
+            is not getattr(DefaultDialect, method).__code__
+        )
+
+    def _default_multi_reflect(
+        self,
+        single_tbl_method,
+        connection,
+        kind,
+        schema,
+        filter_names,
+        scope,
+        **kw,
+    ):
+        names_fns = []
+        temp_names_fns = []
+        if ObjectKind.TABLE in kind:
+            names_fns.append(self.get_table_names)
+            temp_names_fns.append(self.get_temp_table_names)
+        if ObjectKind.VIEW in kind:
+            names_fns.append(self.get_view_names)
+            temp_names_fns.append(self.get_temp_view_names)
+        if ObjectKind.MATERIALIZED_VIEW in kind:
+            names_fns.append(self.get_materialized_view_names)
+            # no temp materialized view at the moment
+            # temp_names_fns.append(self.get_temp_materialized_view_names)
+
+        unreflectable = kw.pop("unreflectable", {})
+
+        if (
+            filter_names
+            and scope is ObjectScope.ANY
+            and kind is ObjectKind.ANY
+        ):
+            # if names are given and no qualification on type of table
+            # (i.e. the Table(..., autoload) case), take the names as given,
+            # don't run names queries. If a table does not exit
+            # NoSuchTableError is raised and it's skipped
+
+            # this also suits the case for mssql where we can reflect
+            # individual temp tables but there's no temp_names_fn
+            names = filter_names
+        else:
+            names = []
+            name_kw = {"schema": schema, **kw}
+            fns = []
+            if ObjectScope.DEFAULT in scope:
+                fns.extend(names_fns)
+            if ObjectScope.TEMPORARY in scope:
+                fns.extend(temp_names_fns)
+
+            for fn in fns:
+                try:
+                    names.extend(fn(connection, **name_kw))
+                except NotImplementedError:
+                    pass
+
+        if filter_names:
+            filter_names = set(filter_names)
+
+        # iterate over all the tables/views and call the single table method
+        for table in names:
+            if not filter_names or table in filter_names:
+                key = (schema, table)
+                try:
+                    yield (
+                        key,
+                        single_tbl_method(
+                            connection, table, schema=schema, **kw
+                        ),
+                    )
+                except exc.UnreflectableTableError as err:
+                    if key not in unreflectable:
+                        unreflectable[key] = err
+                except exc.NoSuchTableError:
+                    pass
+
+    def get_multi_table_options(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_table_options, connection, **kw
+        )
+
+    def get_multi_columns(self, connection, **kw):
+        return self._default_multi_reflect(self.get_columns, connection, **kw)
+
+    def get_multi_pk_constraint(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_pk_constraint, connection, **kw
+        )
+
+    def get_multi_foreign_keys(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_foreign_keys, connection, **kw
+        )
+
+    def get_multi_indexes(self, connection, **kw):
+        return self._default_multi_reflect(self.get_indexes, connection, **kw)
+
+    def get_multi_unique_constraints(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_unique_constraints, connection, **kw
+        )
+
+    def get_multi_check_constraints(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_check_constraints, connection, **kw
+        )
+
+    def get_multi_table_comment(self, connection, **kw):
+        return self._default_multi_reflect(
+            self.get_table_comment, connection, **kw
+        )
+
+
+class StrCompileDialect(DefaultDialect):
+    statement_compiler = compiler.StrSQLCompiler
+    ddl_compiler = compiler.DDLCompiler
+    type_compiler_cls = compiler.StrSQLTypeCompiler
+    preparer = compiler.IdentifierPreparer
+
+    insert_returning = True
+    update_returning = True
+    delete_returning = True
+
+    supports_statement_cache = True
+
+    supports_identity_columns = True
+
+    supports_sequences = True
+    sequences_optional = True
+    preexecute_autoincrement_sequences = False
+
+    supports_native_boolean = True
+
+    supports_multivalues_insert = True
+    supports_simple_order_by_label = True
+
+
+class DefaultExecutionContext(ExecutionContext):
+    isinsert = False
+    isupdate = False
+    isdelete = False
+    is_crud = False
+    is_text = False
+    isddl = False
+
+    execute_style: ExecuteStyle = ExecuteStyle.EXECUTE
+
+    compiled: Optional[Compiled] = None
+    result_column_struct: Optional[
+        Tuple[List[ResultColumnsEntry], bool, bool, bool, bool]
+    ] = None
+    returned_default_rows: Optional[Sequence[Row[Any]]] = None
+
+    execution_options: _ExecuteOptions = util.EMPTY_DICT
+
+    cursor_fetch_strategy = _cursor._DEFAULT_FETCH
+
+    invoked_statement: Optional[Executable] = None
+
+    _is_implicit_returning = False
+    _is_explicit_returning = False
+    _is_supplemental_returning = False
+    _is_server_side = False
+
+    _soft_closed = False
+
+    _rowcount: Optional[int] = None
+
+    # a hook for SQLite's translation of
+    # result column names
+    # NOTE: pyhive is using this hook, can't remove it :(
+    _translate_colname: Optional[Callable[[str], str]] = None
+
+    _expanded_parameters: Mapping[str, List[str]] = util.immutabledict()
+    """used by set_input_sizes().
+
+    This collection comes from ``ExpandedState.parameter_expansion``.
+
+    """
+
+    cache_hit = NO_CACHE_KEY
+
+    root_connection: Connection
+    _dbapi_connection: PoolProxiedConnection
+    dialect: Dialect
+    unicode_statement: str
+    cursor: DBAPICursor
+    compiled_parameters: List[_MutableCoreSingleExecuteParams]
+    parameters: _DBAPIMultiExecuteParams
+    extracted_parameters: Optional[Sequence[BindParameter[Any]]]
+
+    _empty_dict_params = cast("Mapping[str, Any]", util.EMPTY_DICT)
+
+    _insertmanyvalues_rows: Optional[List[Tuple[Any, ...]]] = None
+    _num_sentinel_cols: int = 0
+
+    @classmethod
+    def _init_ddl(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        compiled_ddl: DDLCompiler,
+    ) -> ExecutionContext:
+        """Initialize execution context for an ExecutableDDLElement
+        construct."""
+
+        self = cls.__new__(cls)
+        self.root_connection = connection
+        self._dbapi_connection = dbapi_connection
+        self.dialect = connection.dialect
+
+        self.compiled = compiled = compiled_ddl
+        self.isddl = True
+
+        self.execution_options = execution_options
+
+        self.unicode_statement = str(compiled)
+        if compiled.schema_translate_map:
+            schema_translate_map = self.execution_options.get(
+                "schema_translate_map", {}
+            )
+
+            rst = compiled.preparer._render_schema_translates
+            self.unicode_statement = rst(
+                self.unicode_statement, schema_translate_map
+            )
+
+        self.statement = self.unicode_statement
+
+        self.cursor = self.create_cursor()
+        self.compiled_parameters = []
+
+        if dialect.positional:
+            self.parameters = [dialect.execute_sequence_format()]
+        else:
+            self.parameters = [self._empty_dict_params]
+
+        return self
+
+    @classmethod
+    def _init_compiled(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        compiled: SQLCompiler,
+        parameters: _CoreMultiExecuteParams,
+        invoked_statement: Executable,
+        extracted_parameters: Optional[Sequence[BindParameter[Any]]],
+        cache_hit: CacheStats = CacheStats.CACHING_DISABLED,
+    ) -> ExecutionContext:
+        """Initialize execution context for a Compiled construct."""
+
+        self = cls.__new__(cls)
+        self.root_connection = connection
+        self._dbapi_connection = dbapi_connection
+        self.dialect = connection.dialect
+        self.extracted_parameters = extracted_parameters
+        self.invoked_statement = invoked_statement
+        self.compiled = compiled
+        self.cache_hit = cache_hit
+
+        self.execution_options = execution_options
+
+        self.result_column_struct = (
+            compiled._result_columns,
+            compiled._ordered_columns,
+            compiled._textual_ordered_columns,
+            compiled._ad_hoc_textual,
+            compiled._loose_column_name_matching,
+        )
+
+        self.isinsert = ii = compiled.isinsert
+        self.isupdate = iu = compiled.isupdate
+        self.isdelete = id_ = compiled.isdelete
+        self.is_text = compiled.isplaintext
+
+        if ii or iu or id_:
+            dml_statement = compiled.compile_state.statement  # type: ignore
+            if TYPE_CHECKING:
+                assert isinstance(dml_statement, UpdateBase)
+            self.is_crud = True
+            self._is_explicit_returning = ier = bool(dml_statement._returning)
+            self._is_implicit_returning = iir = bool(
+                compiled.implicit_returning
+            )
+            if iir and dml_statement._supplemental_returning:
+                self._is_supplemental_returning = True
+
+            # dont mix implicit and explicit returning
+            assert not (iir and ier)
+
+            if (ier or iir) and compiled.for_executemany:
+                if ii and not self.dialect.insert_executemany_returning:
+                    raise exc.InvalidRequestError(
+                        f"Dialect {self.dialect.dialect_description} with "
+                        f"current server capabilities does not support "
+                        "INSERT..RETURNING when executemany is used"
+                    )
+                elif (
+                    ii
+                    and dml_statement._sort_by_parameter_order
+                    and not self.dialect.insert_executemany_returning_sort_by_parameter_order  # noqa: E501
+                ):
+                    raise exc.InvalidRequestError(
+                        f"Dialect {self.dialect.dialect_description} with "
+                        f"current server capabilities does not support "
+                        "INSERT..RETURNING with deterministic row ordering "
+                        "when executemany is used"
+                    )
+                elif (
+                    ii
+                    and self.dialect.use_insertmanyvalues
+                    and not compiled._insertmanyvalues
+                ):
+                    raise exc.InvalidRequestError(
+                        'Statement does not have "insertmanyvalues" '
+                        "enabled, can't use INSERT..RETURNING with "
+                        "executemany in this case."
+                    )
+                elif iu and not self.dialect.update_executemany_returning:
+                    raise exc.InvalidRequestError(
+                        f"Dialect {self.dialect.dialect_description} with "
+                        f"current server capabilities does not support "
+                        "UPDATE..RETURNING when executemany is used"
+                    )
+                elif id_ and not self.dialect.delete_executemany_returning:
+                    raise exc.InvalidRequestError(
+                        f"Dialect {self.dialect.dialect_description} with "
+                        f"current server capabilities does not support "
+                        "DELETE..RETURNING when executemany is used"
+                    )
+
+        if not parameters:
+            self.compiled_parameters = [
+                compiled.construct_params(
+                    extracted_parameters=extracted_parameters,
+                    escape_names=False,
+                )
+            ]
+        else:
+            self.compiled_parameters = [
+                compiled.construct_params(
+                    m,
+                    escape_names=False,
+                    _group_number=grp,
+                    extracted_parameters=extracted_parameters,
+                )
+                for grp, m in enumerate(parameters)
+            ]
+
+            if len(parameters) > 1:
+                if self.isinsert and compiled._insertmanyvalues:
+                    self.execute_style = ExecuteStyle.INSERTMANYVALUES
+
+                    imv = compiled._insertmanyvalues
+                    if imv.sentinel_columns is not None:
+                        self._num_sentinel_cols = imv.num_sentinel_columns
+                else:
+                    self.execute_style = ExecuteStyle.EXECUTEMANY
+
+        self.unicode_statement = compiled.string
+
+        self.cursor = self.create_cursor()
+
+        if self.compiled.insert_prefetch or self.compiled.update_prefetch:
+            self._process_execute_defaults()
+
+        processors = compiled._bind_processors
+
+        flattened_processors: Mapping[
+            str, _BindProcessorType[Any]
+        ] = processors  # type: ignore[assignment]
+
+        if compiled.literal_execute_params or compiled.post_compile_params:
+            if self.executemany:
+                raise exc.InvalidRequestError(
+                    "'literal_execute' or 'expanding' parameters can't be "
+                    "used with executemany()"
+                )
+
+            expanded_state = compiled._process_parameters_for_postcompile(
+                self.compiled_parameters[0]
+            )
+
+            # re-assign self.unicode_statement
+            self.unicode_statement = expanded_state.statement
+
+            self._expanded_parameters = expanded_state.parameter_expansion
+
+            flattened_processors = dict(processors)  # type: ignore
+            flattened_processors.update(expanded_state.processors)
+            positiontup = expanded_state.positiontup
+        elif compiled.positional:
+            positiontup = self.compiled.positiontup
+        else:
+            positiontup = None
+
+        if compiled.schema_translate_map:
+            schema_translate_map = self.execution_options.get(
+                "schema_translate_map", {}
+            )
+            rst = compiled.preparer._render_schema_translates
+            self.unicode_statement = rst(
+                self.unicode_statement, schema_translate_map
+            )
+
+        # final self.unicode_statement is now assigned, encode if needed
+        # by dialect
+        self.statement = self.unicode_statement
+
+        # Convert the dictionary of bind parameter values
+        # into a dict or list to be sent to the DBAPI's
+        # execute() or executemany() method.
+
+        if compiled.positional:
+            core_positional_parameters: MutableSequence[Sequence[Any]] = []
+            assert positiontup is not None
+            for compiled_params in self.compiled_parameters:
+                l_param: List[Any] = [
+                    (
+                        flattened_processors[key](compiled_params[key])
+                        if key in flattened_processors
+                        else compiled_params[key]
+                    )
+                    for key in positiontup
+                ]
+                core_positional_parameters.append(
+                    dialect.execute_sequence_format(l_param)
+                )
+
+            self.parameters = core_positional_parameters
+        else:
+            core_dict_parameters: MutableSequence[Dict[str, Any]] = []
+            escaped_names = compiled.escaped_bind_names
+
+            # note that currently, "expanded" parameters will be present
+            # in self.compiled_parameters in their quoted form.   This is
+            # slightly inconsistent with the approach taken as of
+            # #8056 where self.compiled_parameters is meant to contain unquoted
+            # param names.
+            d_param: Dict[str, Any]
+            for compiled_params in self.compiled_parameters:
+                if escaped_names:
+                    d_param = {
+                        escaped_names.get(key, key): (
+                            flattened_processors[key](compiled_params[key])
+                            if key in flattened_processors
+                            else compiled_params[key]
+                        )
+                        for key in compiled_params
+                    }
+                else:
+                    d_param = {
+                        key: (
+                            flattened_processors[key](compiled_params[key])
+                            if key in flattened_processors
+                            else compiled_params[key]
+                        )
+                        for key in compiled_params
+                    }
+
+                core_dict_parameters.append(d_param)
+
+            self.parameters = core_dict_parameters
+
+        return self
+
+    @classmethod
+    def _init_statement(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+    ) -> ExecutionContext:
+        """Initialize execution context for a string SQL statement."""
+
+        self = cls.__new__(cls)
+        self.root_connection = connection
+        self._dbapi_connection = dbapi_connection
+        self.dialect = connection.dialect
+        self.is_text = True
+
+        self.execution_options = execution_options
+
+        if not parameters:
+            if self.dialect.positional:
+                self.parameters = [dialect.execute_sequence_format()]
+            else:
+                self.parameters = [self._empty_dict_params]
+        elif isinstance(parameters[0], dialect.execute_sequence_format):
+            self.parameters = parameters
+        elif isinstance(parameters[0], dict):
+            self.parameters = parameters
+        else:
+            self.parameters = [
+                dialect.execute_sequence_format(p) for p in parameters
+            ]
+
+        if len(parameters) > 1:
+            self.execute_style = ExecuteStyle.EXECUTEMANY
+
+        self.statement = self.unicode_statement = statement
+
+        self.cursor = self.create_cursor()
+        return self
+
+    @classmethod
+    def _init_default(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+    ) -> ExecutionContext:
+        """Initialize execution context for a ColumnDefault construct."""
+
+        self = cls.__new__(cls)
+        self.root_connection = connection
+        self._dbapi_connection = dbapi_connection
+        self.dialect = connection.dialect
+
+        self.execution_options = execution_options
+
+        self.cursor = self.create_cursor()
+        return self
+
+    def _get_cache_stats(self) -> str:
+        if self.compiled is None:
+            return "raw sql"
+
+        now = perf_counter()
+
+        ch = self.cache_hit
+
+        gen_time = self.compiled._gen_time
+        assert gen_time is not None
+
+        if ch is NO_CACHE_KEY:
+            return "no key %.5fs" % (now - gen_time,)
+        elif ch is CACHE_HIT:
+            return "cached since %.4gs ago" % (now - gen_time,)
+        elif ch is CACHE_MISS:
+            return "generated in %.5fs" % (now - gen_time,)
+        elif ch is CACHING_DISABLED:
+            if "_cache_disable_reason" in self.execution_options:
+                return "caching disabled (%s) %.5fs " % (
+                    self.execution_options["_cache_disable_reason"],
+                    now - gen_time,
+                )
+            else:
+                return "caching disabled %.5fs" % (now - gen_time,)
+        elif ch is NO_DIALECT_SUPPORT:
+            return "dialect %s+%s does not support caching %.5fs" % (
+                self.dialect.name,
+                self.dialect.driver,
+                now - gen_time,
+            )
+        else:
+            return "unknown"
+
+    @property
+    def executemany(self):
+        return self.execute_style in (
+            ExecuteStyle.EXECUTEMANY,
+            ExecuteStyle.INSERTMANYVALUES,
+        )
+
+    @util.memoized_property
+    def identifier_preparer(self):
+        if self.compiled:
+            return self.compiled.preparer
+        elif "schema_translate_map" in self.execution_options:
+            return self.dialect.identifier_preparer._with_schema_translate(
+                self.execution_options["schema_translate_map"]
+            )
+        else:
+            return self.dialect.identifier_preparer
+
+    @util.memoized_property
+    def engine(self):
+        return self.root_connection.engine
+
+    @util.memoized_property
+    def postfetch_cols(self) -> Optional[Sequence[Column[Any]]]:
+        if TYPE_CHECKING:
+            assert isinstance(self.compiled, SQLCompiler)
+        return self.compiled.postfetch
+
+    @util.memoized_property
+    def prefetch_cols(self) -> Optional[Sequence[Column[Any]]]:
+        if TYPE_CHECKING:
+            assert isinstance(self.compiled, SQLCompiler)
+        if self.isinsert:
+            return self.compiled.insert_prefetch
+        elif self.isupdate:
+            return self.compiled.update_prefetch
+        else:
+            return ()
+
+    @util.memoized_property
+    def no_parameters(self):
+        return self.execution_options.get("no_parameters", False)
+
+    def _execute_scalar(self, stmt, type_, parameters=None):
+        """Execute a string statement on the current cursor, returning a
+        scalar result.
+
+        Used to fire off sequences, default phrases, and "select lastrowid"
+        types of statements individually or in the context of a parent INSERT
+        or UPDATE statement.
+
+        """
+
+        conn = self.root_connection
+
+        if "schema_translate_map" in self.execution_options:
+            schema_translate_map = self.execution_options.get(
+                "schema_translate_map", {}
+            )
+
+            rst = self.identifier_preparer._render_schema_translates
+            stmt = rst(stmt, schema_translate_map)
+
+        if not parameters:
+            if self.dialect.positional:
+                parameters = self.dialect.execute_sequence_format()
+            else:
+                parameters = {}
+
+        conn._cursor_execute(self.cursor, stmt, parameters, context=self)
+        row = self.cursor.fetchone()
+        if row is not None:
+            r = row[0]
+        else:
+            r = None
+        if type_ is not None:
+            # apply type post processors to the result
+            proc = type_._cached_result_processor(
+                self.dialect, self.cursor.description[0][1]
+            )
+            if proc:
+                return proc(r)
+        return r
+
+    @util.memoized_property
+    def connection(self):
+        return self.root_connection
+
+    def _use_server_side_cursor(self):
+        if not self.dialect.supports_server_side_cursors:
+            return False
+
+        if self.dialect.server_side_cursors:
+            # this is deprecated
+            use_server_side = self.execution_options.get(
+                "stream_results", True
+            ) and (
+                self.compiled
+                and isinstance(self.compiled.statement, expression.Selectable)
+                or (
+                    (
+                        not self.compiled
+                        or isinstance(
+                            self.compiled.statement, expression.TextClause
+                        )
+                    )
+                    and self.unicode_statement
+                    and SERVER_SIDE_CURSOR_RE.match(self.unicode_statement)
+                )
+            )
+        else:
+            use_server_side = self.execution_options.get(
+                "stream_results", False
+            )
+
+        return use_server_side
+
+    def create_cursor(self):
+        if (
+            # inlining initial preference checks for SS cursors
+            self.dialect.supports_server_side_cursors
+            and (
+                self.execution_options.get("stream_results", False)
+                or (
+                    self.dialect.server_side_cursors
+                    and self._use_server_side_cursor()
+                )
+            )
+        ):
+            self._is_server_side = True
+            return self.create_server_side_cursor()
+        else:
+            self._is_server_side = False
+            return self.create_default_cursor()
+
+    def fetchall_for_returning(self, cursor):
+        return cursor.fetchall()
+
+    def create_default_cursor(self):
+        return self._dbapi_connection.cursor()
+
+    def create_server_side_cursor(self):
+        raise NotImplementedError()
+
+    def pre_exec(self):
+        pass
+
+    def get_out_parameter_values(self, names):
+        raise NotImplementedError(
+            "This dialect does not support OUT parameters"
+        )
+
+    def post_exec(self):
+        pass
+
+    def get_result_processor(self, type_, colname, coltype):
+        """Return a 'result processor' for a given type as present in
+        cursor.description.
+
+        This has a default implementation that dialects can override
+        for context-sensitive result type handling.
+
+        """
+        return type_._cached_result_processor(self.dialect, coltype)
+
+    def get_lastrowid(self):
+        """return self.cursor.lastrowid, or equivalent, after an INSERT.
+
+        This may involve calling special cursor functions, issuing a new SELECT
+        on the cursor (or a new one), or returning a stored value that was
+        calculated within post_exec().
+
+        This function will only be called for dialects which support "implicit"
+        primary key generation, keep preexecute_autoincrement_sequences set to
+        False, and when no explicit id value was bound to the statement.
+
+        The function is called once for an INSERT statement that would need to
+        return the last inserted primary key for those dialects that make use
+        of the lastrowid concept.  In these cases, it is called directly after
+        :meth:`.ExecutionContext.post_exec`.
+
+        """
+        return self.cursor.lastrowid
+
+    def handle_dbapi_exception(self, e):
+        pass
+
+    @util.non_memoized_property
+    def rowcount(self) -> int:
+        if self._rowcount is not None:
+            return self._rowcount
+        else:
+            return self.cursor.rowcount
+
+    @property
+    def _has_rowcount(self):
+        return self._rowcount is not None
+
+    def supports_sane_rowcount(self):
+        return self.dialect.supports_sane_rowcount
+
+    def supports_sane_multi_rowcount(self):
+        return self.dialect.supports_sane_multi_rowcount
+
+    def _setup_result_proxy(self):
+        exec_opt = self.execution_options
+
+        if self._rowcount is None and exec_opt.get("preserve_rowcount", False):
+            self._rowcount = self.cursor.rowcount
+
+        if self.is_crud or self.is_text:
+            result = self._setup_dml_or_text_result()
+            yp = sr = False
+        else:
+            yp = exec_opt.get("yield_per", None)
+            sr = self._is_server_side or exec_opt.get("stream_results", False)
+            strategy = self.cursor_fetch_strategy
+            if sr and strategy is _cursor._DEFAULT_FETCH:
+                strategy = _cursor.BufferedRowCursorFetchStrategy(
+                    self.cursor, self.execution_options
+                )
+            cursor_description: _DBAPICursorDescription = (
+                strategy.alternate_cursor_description
+                or self.cursor.description
+            )
+            if cursor_description is None:
+                strategy = _cursor._NO_CURSOR_DQL
+
+            result = _cursor.CursorResult(self, strategy, cursor_description)
+
+        compiled = self.compiled
+
+        if (
+            compiled
+            and not self.isddl
+            and cast(SQLCompiler, compiled).has_out_parameters
+        ):
+            self._setup_out_parameters(result)
+
+        self._soft_closed = result._soft_closed
+
+        if yp:
+            result = result.yield_per(yp)
+
+        return result
+
+    def _setup_out_parameters(self, result):
+        compiled = cast(SQLCompiler, self.compiled)
+
+        out_bindparams = [
+            (param, name)
+            for param, name in compiled.bind_names.items()
+            if param.isoutparam
+        ]
+        out_parameters = {}
+
+        for bindparam, raw_value in zip(
+            [param for param, name in out_bindparams],
+            self.get_out_parameter_values(
+                [name for param, name in out_bindparams]
+            ),
+        ):
+            type_ = bindparam.type
+            impl_type = type_.dialect_impl(self.dialect)
+            dbapi_type = impl_type.get_dbapi_type(self.dialect.loaded_dbapi)
+            result_processor = impl_type.result_processor(
+                self.dialect, dbapi_type
+            )
+            if result_processor is not None:
+                raw_value = result_processor(raw_value)
+            out_parameters[bindparam.key] = raw_value
+
+        result.out_parameters = out_parameters
+
+    def _setup_dml_or_text_result(self):
+        compiled = cast(SQLCompiler, self.compiled)
+
+        strategy: ResultFetchStrategy = self.cursor_fetch_strategy
+
+        if self.isinsert:
+            if (
+                self.execute_style is ExecuteStyle.INSERTMANYVALUES
+                and compiled.effective_returning
+            ):
+                strategy = _cursor.FullyBufferedCursorFetchStrategy(
+                    self.cursor,
+                    initial_buffer=self._insertmanyvalues_rows,
+                    # maintain alt cursor description if set by the
+                    # dialect, e.g. mssql preserves it
+                    alternate_description=(
+                        strategy.alternate_cursor_description
+                    ),
+                )
+
+            if compiled.postfetch_lastrowid:
+                self.inserted_primary_key_rows = (
+                    self._setup_ins_pk_from_lastrowid()
+                )
+            # else if not self._is_implicit_returning,
+            # the default inserted_primary_key_rows accessor will
+            # return an "empty" primary key collection when accessed.
+
+        if self._is_server_side and strategy is _cursor._DEFAULT_FETCH:
+            strategy = _cursor.BufferedRowCursorFetchStrategy(
+                self.cursor, self.execution_options
+            )
+
+        if strategy is _cursor._NO_CURSOR_DML:
+            cursor_description = None
+        else:
+            cursor_description = (
+                strategy.alternate_cursor_description
+                or self.cursor.description
+            )
+
+        if cursor_description is None:
+            strategy = _cursor._NO_CURSOR_DML
+        elif self._num_sentinel_cols:
+            assert self.execute_style is ExecuteStyle.INSERTMANYVALUES
+            # strip out the sentinel columns from cursor description
+            # a similar logic is done to the rows only in CursorResult
+            cursor_description = cursor_description[
+                0 : -self._num_sentinel_cols
+            ]
+
+        result: _cursor.CursorResult[Any] = _cursor.CursorResult(
+            self, strategy, cursor_description
+        )
+
+        if self.isinsert:
+            if self._is_implicit_returning:
+                rows = result.all()
+
+                self.returned_default_rows = rows
+
+                self.inserted_primary_key_rows = (
+                    self._setup_ins_pk_from_implicit_returning(result, rows)
+                )
+
+                # test that it has a cursor metadata that is accurate. the
+                # first row will have been fetched and current assumptions
+                # are that the result has only one row, until executemany()
+                # support is added here.
+                assert result._metadata.returns_rows
+
+                # Insert statement has both return_defaults() and
+                # returning().  rewind the result on the list of rows
+                # we just used.
+                if self._is_supplemental_returning:
+                    result._rewind(rows)
+                else:
+                    result._soft_close()
+            elif not self._is_explicit_returning:
+                result._soft_close()
+
+                # we assume here the result does not return any rows.
+                # *usually*, this will be true.  However, some dialects
+                # such as that of MSSQL/pyodbc need to SELECT a post fetch
+                # function so this is not necessarily true.
+                # assert not result.returns_rows
+
+        elif self._is_implicit_returning:
+            rows = result.all()
+
+            if rows:
+                self.returned_default_rows = rows
+            self._rowcount = len(rows)
+
+            if self._is_supplemental_returning:
+                result._rewind(rows)
+            else:
+                result._soft_close()
+
+            # test that it has a cursor metadata that is accurate.
+            # the rows have all been fetched however.
+            assert result._metadata.returns_rows
+
+        elif not result._metadata.returns_rows:
+            # no results, get rowcount
+            # (which requires open cursor on some drivers)
+            if self._rowcount is None:
+                self._rowcount = self.cursor.rowcount
+            result._soft_close()
+        elif self.isupdate or self.isdelete:
+            if self._rowcount is None:
+                self._rowcount = self.cursor.rowcount
+        return result
+
+    @util.memoized_property
+    def inserted_primary_key_rows(self):
+        # if no specific "get primary key" strategy was set up
+        # during execution, return a "default" primary key based
+        # on what's in the compiled_parameters and nothing else.
+        return self._setup_ins_pk_from_empty()
+
+    def _setup_ins_pk_from_lastrowid(self):
+        getter = cast(
+            SQLCompiler, self.compiled
+        )._inserted_primary_key_from_lastrowid_getter
+        lastrowid = self.get_lastrowid()
+        return [getter(lastrowid, self.compiled_parameters[0])]
+
+    def _setup_ins_pk_from_empty(self):
+        getter = cast(
+            SQLCompiler, self.compiled
+        )._inserted_primary_key_from_lastrowid_getter
+        return [getter(None, param) for param in self.compiled_parameters]
+
+    def _setup_ins_pk_from_implicit_returning(self, result, rows):
+        if not rows:
+            return []
+
+        getter = cast(
+            SQLCompiler, self.compiled
+        )._inserted_primary_key_from_returning_getter
+        compiled_params = self.compiled_parameters
+
+        return [
+            getter(row, param) for row, param in zip(rows, compiled_params)
+        ]
+
+    def lastrow_has_defaults(self):
+        return (self.isinsert or self.isupdate) and bool(
+            cast(SQLCompiler, self.compiled).postfetch
+        )
+
+    def _prepare_set_input_sizes(
+        self,
+    ) -> Optional[List[Tuple[str, Any, TypeEngine[Any]]]]:
+        """Given a cursor and ClauseParameters, prepare arguments
+        in order to call the appropriate
+        style of ``setinputsizes()`` on the cursor, using DB-API types
+        from the bind parameter's ``TypeEngine`` objects.
+
+        This method only called by those dialects which set the
+        :attr:`.Dialect.bind_typing` attribute to
+        :attr:`.BindTyping.SETINPUTSIZES`.  Python-oracledb and cx_Oracle are
+        the only DBAPIs that requires setinputsizes(); pyodbc offers it as an
+        option.
+
+        Prior to SQLAlchemy 2.0, the setinputsizes() approach was also used
+        for pg8000 and asyncpg, which has been changed to inline rendering
+        of casts.
+
+        """
+        if self.isddl or self.is_text:
+            return None
+
+        compiled = cast(SQLCompiler, self.compiled)
+
+        inputsizes = compiled._get_set_input_sizes_lookup()
+
+        if inputsizes is None:
+            return None
+
+        dialect = self.dialect
+
+        # all of the rest of this... cython?
+
+        if dialect._has_events:
+            inputsizes = dict(inputsizes)
+            dialect.dispatch.do_setinputsizes(
+                inputsizes, self.cursor, self.statement, self.parameters, self
+            )
+
+        if compiled.escaped_bind_names:
+            escaped_bind_names = compiled.escaped_bind_names
+        else:
+            escaped_bind_names = None
+
+        if dialect.positional:
+            items = [
+                (key, compiled.binds[key])
+                for key in compiled.positiontup or ()
+            ]
+        else:
+            items = [
+                (key, bindparam)
+                for bindparam, key in compiled.bind_names.items()
+            ]
+
+        generic_inputsizes: List[Tuple[str, Any, TypeEngine[Any]]] = []
+        for key, bindparam in items:
+            if bindparam in compiled.literal_execute_params:
+                continue
+
+            if key in self._expanded_parameters:
+                if is_tuple_type(bindparam.type):
+                    num = len(bindparam.type.types)
+                    dbtypes = inputsizes[bindparam]
+                    generic_inputsizes.extend(
+                        (
+                            (
+                                escaped_bind_names.get(paramname, paramname)
+                                if escaped_bind_names is not None
+                                else paramname
+                            ),
+                            dbtypes[idx % num],
+                            bindparam.type.types[idx % num],
+                        )
+                        for idx, paramname in enumerate(
+                            self._expanded_parameters[key]
+                        )
+                    )
+                else:
+                    dbtype = inputsizes.get(bindparam, None)
+                    generic_inputsizes.extend(
+                        (
+                            (
+                                escaped_bind_names.get(paramname, paramname)
+                                if escaped_bind_names is not None
+                                else paramname
+                            ),
+                            dbtype,
+                            bindparam.type,
+                        )
+                        for paramname in self._expanded_parameters[key]
+                    )
+            else:
+                dbtype = inputsizes.get(bindparam, None)
+
+                escaped_name = (
+                    escaped_bind_names.get(key, key)
+                    if escaped_bind_names is not None
+                    else key
+                )
+
+                generic_inputsizes.append(
+                    (escaped_name, dbtype, bindparam.type)
+                )
+
+        return generic_inputsizes
+
+    def _exec_default(self, column, default, type_):
+        if default.is_sequence:
+            return self.fire_sequence(default, type_)
+        elif default.is_callable:
+            # this codepath is not normally used as it's inlined
+            # into _process_execute_defaults
+            self.current_column = column
+            return default.arg(self)
+        elif default.is_clause_element:
+            return self._exec_default_clause_element(column, default, type_)
+        else:
+            # this codepath is not normally used as it's inlined
+            # into _process_execute_defaults
+            return default.arg
+
+    def _exec_default_clause_element(self, column, default, type_):
+        # execute a default that's a complete clause element.  Here, we have
+        # to re-implement a miniature version of the compile->parameters->
+        # cursor.execute() sequence, since we don't want to modify the state
+        # of the connection  / result in progress or create new connection/
+        # result objects etc.
+        # .. versionchanged:: 1.4
+
+        if not default._arg_is_typed:
+            default_arg = expression.type_coerce(default.arg, type_)
+        else:
+            default_arg = default.arg
+        compiled = expression.select(default_arg).compile(dialect=self.dialect)
+        compiled_params = compiled.construct_params()
+        processors = compiled._bind_processors
+        if compiled.positional:
+            parameters = self.dialect.execute_sequence_format(
+                [
+                    (
+                        processors[key](compiled_params[key])  # type: ignore
+                        if key in processors
+                        else compiled_params[key]
+                    )
+                    for key in compiled.positiontup or ()
+                ]
+            )
+        else:
+            parameters = {
+                key: (
+                    processors[key](compiled_params[key])  # type: ignore
+                    if key in processors
+                    else compiled_params[key]
+                )
+                for key in compiled_params
+            }
+        return self._execute_scalar(
+            str(compiled), type_, parameters=parameters
+        )
+
+    current_parameters: Optional[_CoreSingleExecuteParams] = None
+    """A dictionary of parameters applied to the current row.
+
+    This attribute is only available in the context of a user-defined default
+    generation function, e.g. as described at :ref:`context_default_functions`.
+    It consists of a dictionary which includes entries for each column/value
+    pair that is to be part of the INSERT or UPDATE statement. The keys of the
+    dictionary will be the key value of each :class:`_schema.Column`,
+    which is usually
+    synonymous with the name.
+
+    Note that the :attr:`.DefaultExecutionContext.current_parameters` attribute
+    does not accommodate for the "multi-values" feature of the
+    :meth:`_expression.Insert.values` method.  The
+    :meth:`.DefaultExecutionContext.get_current_parameters` method should be
+    preferred.
+
+    .. seealso::
+
+        :meth:`.DefaultExecutionContext.get_current_parameters`
+
+        :ref:`context_default_functions`
+
+    """
+
+    def get_current_parameters(self, isolate_multiinsert_groups=True):
+        """Return a dictionary of parameters applied to the current row.
+
+        This method can only be used in the context of a user-defined default
+        generation function, e.g. as described at
+        :ref:`context_default_functions`. When invoked, a dictionary is
+        returned which includes entries for each column/value pair that is part
+        of the INSERT or UPDATE statement. The keys of the dictionary will be
+        the key value of each :class:`_schema.Column`,
+        which is usually synonymous
+        with the name.
+
+        :param isolate_multiinsert_groups=True: indicates that multi-valued
+         INSERT constructs created using :meth:`_expression.Insert.values`
+         should be
+         handled by returning only the subset of parameters that are local
+         to the current column default invocation.   When ``False``, the
+         raw parameters of the statement are returned including the
+         naming convention used in the case of multi-valued INSERT.
+
+        .. versionadded:: 1.2  added
+           :meth:`.DefaultExecutionContext.get_current_parameters`
+           which provides more functionality over the existing
+           :attr:`.DefaultExecutionContext.current_parameters`
+           attribute.
+
+        .. seealso::
+
+            :attr:`.DefaultExecutionContext.current_parameters`
+
+            :ref:`context_default_functions`
+
+        """
+        try:
+            parameters = self.current_parameters
+            column = self.current_column
+        except AttributeError:
+            raise exc.InvalidRequestError(
+                "get_current_parameters() can only be invoked in the "
+                "context of a Python side column default function"
+            )
+        else:
+            assert column is not None
+            assert parameters is not None
+        compile_state = cast(
+            "DMLState", cast(SQLCompiler, self.compiled).compile_state
+        )
+        assert compile_state is not None
+        if (
+            isolate_multiinsert_groups
+            and dml.isinsert(compile_state)
+            and compile_state._has_multi_parameters
+        ):
+            if column._is_multiparam_column:
+                index = column.index + 1
+                d = {column.original.key: parameters[column.key]}
+            else:
+                d = {column.key: parameters[column.key]}
+                index = 0
+            assert compile_state._dict_parameters is not None
+            keys = compile_state._dict_parameters.keys()
+            d.update(
+                (key, parameters["%s_m%d" % (key, index)]) for key in keys
+            )
+            return d
+        else:
+            return parameters
+
+    def get_insert_default(self, column):
+        if column.default is None:
+            return None
+        else:
+            return self._exec_default(column, column.default, column.type)
+
+    def get_update_default(self, column):
+        if column.onupdate is None:
+            return None
+        else:
+            return self._exec_default(column, column.onupdate, column.type)
+
+    def _process_execute_defaults(self):
+        compiled = cast(SQLCompiler, self.compiled)
+
+        key_getter = compiled._within_exec_param_key_getter
+
+        sentinel_counter = 0
+
+        if compiled.insert_prefetch:
+            prefetch_recs = [
+                (
+                    c,
+                    key_getter(c),
+                    c._default_description_tuple,
+                    self.get_insert_default,
+                )
+                for c in compiled.insert_prefetch
+            ]
+        elif compiled.update_prefetch:
+            prefetch_recs = [
+                (
+                    c,
+                    key_getter(c),
+                    c._onupdate_description_tuple,
+                    self.get_update_default,
+                )
+                for c in compiled.update_prefetch
+            ]
+        else:
+            prefetch_recs = []
+
+        for param in self.compiled_parameters:
+            self.current_parameters = param
+
+            for (
+                c,
+                param_key,
+                (arg, is_scalar, is_callable, is_sentinel),
+                fallback,
+            ) in prefetch_recs:
+                if is_sentinel:
+                    param[param_key] = sentinel_counter
+                    sentinel_counter += 1
+                elif is_scalar:
+                    param[param_key] = arg
+                elif is_callable:
+                    self.current_column = c
+                    param[param_key] = arg(self)
+                else:
+                    val = fallback(c)
+                    if val is not None:
+                        param[param_key] = val
+
+        del self.current_parameters
+
+
+DefaultDialect.execution_ctx_cls = DefaultExecutionContext
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/events.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/events.py
new file mode 100644
index 00000000..b759382c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/events.py
@@ -0,0 +1,965 @@
+# engine/events.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Dict
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import Union
+
+from .base import Connection
+from .base import Engine
+from .interfaces import ConnectionEventsTarget
+from .interfaces import DBAPIConnection
+from .interfaces import DBAPICursor
+from .interfaces import Dialect
+from .. import event
+from .. import exc
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from .interfaces import _CoreMultiExecuteParams
+    from .interfaces import _CoreSingleExecuteParams
+    from .interfaces import _DBAPIAnyExecuteParams
+    from .interfaces import _DBAPIMultiExecuteParams
+    from .interfaces import _DBAPISingleExecuteParams
+    from .interfaces import _ExecuteOptions
+    from .interfaces import ExceptionContext
+    from .interfaces import ExecutionContext
+    from .result import Result
+    from ..pool import ConnectionPoolEntry
+    from ..sql import Executable
+    from ..sql.elements import BindParameter
+
+
+class ConnectionEvents(event.Events[ConnectionEventsTarget]):
+    """Available events for
+    :class:`_engine.Connection` and :class:`_engine.Engine`.
+
+    The methods here define the name of an event as well as the names of
+    members that are passed to listener functions.
+
+    An event listener can be associated with any
+    :class:`_engine.Connection` or :class:`_engine.Engine`
+    class or instance, such as an :class:`_engine.Engine`, e.g.::
+
+        from sqlalchemy import event, create_engine
+
+
+        def before_cursor_execute(
+            conn, cursor, statement, parameters, context, executemany
+        ):
+            log.info("Received statement: %s", statement)
+
+
+        engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test")
+        event.listen(engine, "before_cursor_execute", before_cursor_execute)
+
+    or with a specific :class:`_engine.Connection`::
+
+        with engine.begin() as conn:
+
+            @event.listens_for(conn, "before_cursor_execute")
+            def before_cursor_execute(
+                conn, cursor, statement, parameters, context, executemany
+            ):
+                log.info("Received statement: %s", statement)
+
+    When the methods are called with a `statement` parameter, such as in
+    :meth:`.after_cursor_execute` or :meth:`.before_cursor_execute`,
+    the statement is the exact SQL string that was prepared for transmission
+    to the DBAPI ``cursor`` in the connection's :class:`.Dialect`.
+
+    The :meth:`.before_execute` and :meth:`.before_cursor_execute`
+    events can also be established with the ``retval=True`` flag, which
+    allows modification of the statement and parameters to be sent
+    to the database.  The :meth:`.before_cursor_execute` event is
+    particularly useful here to add ad-hoc string transformations, such
+    as comments, to all executions::
+
+        from sqlalchemy.engine import Engine
+        from sqlalchemy import event
+
+
+        @event.listens_for(Engine, "before_cursor_execute", retval=True)
+        def comment_sql_calls(
+            conn, cursor, statement, parameters, context, executemany
+        ):
+            statement = statement + " -- some comment"
+            return statement, parameters
+
+    .. note:: :class:`_events.ConnectionEvents` can be established on any
+       combination of :class:`_engine.Engine`, :class:`_engine.Connection`,
+       as well
+       as instances of each of those classes.  Events across all
+       four scopes will fire off for a given instance of
+       :class:`_engine.Connection`.  However, for performance reasons, the
+       :class:`_engine.Connection` object determines at instantiation time
+       whether or not its parent :class:`_engine.Engine` has event listeners
+       established.   Event listeners added to the :class:`_engine.Engine`
+       class or to an instance of :class:`_engine.Engine`
+       *after* the instantiation
+       of a dependent :class:`_engine.Connection` instance will usually
+       *not* be available on that :class:`_engine.Connection` instance.
+       The newly
+       added listeners will instead take effect for
+       :class:`_engine.Connection`
+       instances created subsequent to those event listeners being
+       established on the parent :class:`_engine.Engine` class or instance.
+
+    :param retval=False: Applies to the :meth:`.before_execute` and
+      :meth:`.before_cursor_execute` events only.  When True, the
+      user-defined event function must have a return value, which
+      is a tuple of parameters that replace the given statement
+      and parameters.  See those methods for a description of
+      specific return arguments.
+
+    """  # noqa
+
+    _target_class_doc = "SomeEngine"
+    _dispatch_target = ConnectionEventsTarget
+
+    @classmethod
+    def _accept_with(
+        cls,
+        target: Union[ConnectionEventsTarget, Type[ConnectionEventsTarget]],
+        identifier: str,
+    ) -> Optional[Union[ConnectionEventsTarget, Type[ConnectionEventsTarget]]]:
+        default_dispatch = super()._accept_with(target, identifier)
+        if default_dispatch is None and hasattr(
+            target, "_no_async_engine_events"
+        ):
+            target._no_async_engine_events()
+
+        return default_dispatch
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: event._EventKey[ConnectionEventsTarget],
+        *,
+        retval: bool = False,
+        **kw: Any,
+    ) -> None:
+        target, identifier, fn = (
+            event_key.dispatch_target,
+            event_key.identifier,
+            event_key._listen_fn,
+        )
+        target._has_events = True
+
+        if not retval:
+            if identifier == "before_execute":
+                orig_fn = fn
+
+                def wrap_before_execute(  # type: ignore
+                    conn, clauseelement, multiparams, params, execution_options
+                ):
+                    orig_fn(
+                        conn,
+                        clauseelement,
+                        multiparams,
+                        params,
+                        execution_options,
+                    )
+                    return clauseelement, multiparams, params
+
+                fn = wrap_before_execute
+            elif identifier == "before_cursor_execute":
+                orig_fn = fn
+
+                def wrap_before_cursor_execute(  # type: ignore
+                    conn, cursor, statement, parameters, context, executemany
+                ):
+                    orig_fn(
+                        conn,
+                        cursor,
+                        statement,
+                        parameters,
+                        context,
+                        executemany,
+                    )
+                    return statement, parameters
+
+                fn = wrap_before_cursor_execute
+        elif retval and identifier not in (
+            "before_execute",
+            "before_cursor_execute",
+        ):
+            raise exc.ArgumentError(
+                "Only the 'before_execute', "
+                "'before_cursor_execute' and 'handle_error' engine "
+                "event listeners accept the 'retval=True' "
+                "argument."
+            )
+        event_key.with_wrapper(fn).base_listen()
+
+    @event._legacy_signature(
+        "1.4",
+        ["conn", "clauseelement", "multiparams", "params"],
+        lambda conn, clauseelement, multiparams, params, execution_options: (
+            conn,
+            clauseelement,
+            multiparams,
+            params,
+        ),
+    )
+    def before_execute(
+        self,
+        conn: Connection,
+        clauseelement: Executable,
+        multiparams: _CoreMultiExecuteParams,
+        params: _CoreSingleExecuteParams,
+        execution_options: _ExecuteOptions,
+    ) -> Optional[
+        Tuple[Executable, _CoreMultiExecuteParams, _CoreSingleExecuteParams]
+    ]:
+        """Intercept high level execute() events, receiving uncompiled
+        SQL constructs and other objects prior to rendering into SQL.
+
+        This event is good for debugging SQL compilation issues as well
+        as early manipulation of the parameters being sent to the database,
+        as the parameter lists will be in a consistent format here.
+
+        This event can be optionally established with the ``retval=True``
+        flag.  The ``clauseelement``, ``multiparams``, and ``params``
+        arguments should be returned as a three-tuple in this case::
+
+            @event.listens_for(Engine, "before_execute", retval=True)
+            def before_execute(conn, clauseelement, multiparams, params):
+                # do something with clauseelement, multiparams, params
+                return clauseelement, multiparams, params
+
+        :param conn: :class:`_engine.Connection` object
+        :param clauseelement: SQL expression construct, :class:`.Compiled`
+         instance, or string statement passed to
+         :meth:`_engine.Connection.execute`.
+        :param multiparams: Multiple parameter sets, a list of dictionaries.
+        :param params: Single parameter set, a single dictionary.
+        :param execution_options: dictionary of execution
+         options passed along with the statement, if any.  This is a merge
+         of all options that will be used, including those of the statement,
+         the connection, and those passed in to the method itself for
+         the 2.0 style of execution.
+
+         .. versionadded: 1.4
+
+        .. seealso::
+
+            :meth:`.before_cursor_execute`
+
+        """
+
+    @event._legacy_signature(
+        "1.4",
+        ["conn", "clauseelement", "multiparams", "params", "result"],
+        lambda conn, clauseelement, multiparams, params, execution_options, result: (  # noqa
+            conn,
+            clauseelement,
+            multiparams,
+            params,
+            result,
+        ),
+    )
+    def after_execute(
+        self,
+        conn: Connection,
+        clauseelement: Executable,
+        multiparams: _CoreMultiExecuteParams,
+        params: _CoreSingleExecuteParams,
+        execution_options: _ExecuteOptions,
+        result: Result[Any],
+    ) -> None:
+        """Intercept high level execute() events after execute.
+
+
+        :param conn: :class:`_engine.Connection` object
+        :param clauseelement: SQL expression construct, :class:`.Compiled`
+         instance, or string statement passed to
+         :meth:`_engine.Connection.execute`.
+        :param multiparams: Multiple parameter sets, a list of dictionaries.
+        :param params: Single parameter set, a single dictionary.
+        :param execution_options: dictionary of execution
+         options passed along with the statement, if any.  This is a merge
+         of all options that will be used, including those of the statement,
+         the connection, and those passed in to the method itself for
+         the 2.0 style of execution.
+
+         .. versionadded: 1.4
+
+        :param result: :class:`_engine.CursorResult` generated by the
+         execution.
+
+        """
+
+    def before_cursor_execute(
+        self,
+        conn: Connection,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIAnyExecuteParams,
+        context: Optional[ExecutionContext],
+        executemany: bool,
+    ) -> Optional[Tuple[str, _DBAPIAnyExecuteParams]]:
+        """Intercept low-level cursor execute() events before execution,
+        receiving the string SQL statement and DBAPI-specific parameter list to
+        be invoked against a cursor.
+
+        This event is a good choice for logging as well as late modifications
+        to the SQL string.  It's less ideal for parameter modifications except
+        for those which are specific to a target backend.
+
+        This event can be optionally established with the ``retval=True``
+        flag.  The ``statement`` and ``parameters`` arguments should be
+        returned as a two-tuple in this case::
+
+            @event.listens_for(Engine, "before_cursor_execute", retval=True)
+            def before_cursor_execute(
+                conn, cursor, statement, parameters, context, executemany
+            ):
+                # do something with statement, parameters
+                return statement, parameters
+
+        See the example at :class:`_events.ConnectionEvents`.
+
+        :param conn: :class:`_engine.Connection` object
+        :param cursor: DBAPI cursor object
+        :param statement: string SQL statement, as to be passed to the DBAPI
+        :param parameters: Dictionary, tuple, or list of parameters being
+         passed to the ``execute()`` or ``executemany()`` method of the
+         DBAPI ``cursor``.  In some cases may be ``None``.
+        :param context: :class:`.ExecutionContext` object in use.  May
+         be ``None``.
+        :param executemany: boolean, if ``True``, this is an ``executemany()``
+         call, if ``False``, this is an ``execute()`` call.
+
+        .. seealso::
+
+            :meth:`.before_execute`
+
+            :meth:`.after_cursor_execute`
+
+        """
+
+    def after_cursor_execute(
+        self,
+        conn: Connection,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIAnyExecuteParams,
+        context: Optional[ExecutionContext],
+        executemany: bool,
+    ) -> None:
+        """Intercept low-level cursor execute() events after execution.
+
+        :param conn: :class:`_engine.Connection` object
+        :param cursor: DBAPI cursor object.  Will have results pending
+         if the statement was a SELECT, but these should not be consumed
+         as they will be needed by the :class:`_engine.CursorResult`.
+        :param statement: string SQL statement, as passed to the DBAPI
+        :param parameters: Dictionary, tuple, or list of parameters being
+         passed to the ``execute()`` or ``executemany()`` method of the
+         DBAPI ``cursor``.  In some cases may be ``None``.
+        :param context: :class:`.ExecutionContext` object in use.  May
+         be ``None``.
+        :param executemany: boolean, if ``True``, this is an ``executemany()``
+         call, if ``False``, this is an ``execute()`` call.
+
+        """
+
+    @event._legacy_signature(
+        "2.0", ["conn", "branch"], converter=lambda conn: (conn, False)
+    )
+    def engine_connect(self, conn: Connection) -> None:
+        """Intercept the creation of a new :class:`_engine.Connection`.
+
+        This event is called typically as the direct result of calling
+        the :meth:`_engine.Engine.connect` method.
+
+        It differs from the :meth:`_events.PoolEvents.connect` method, which
+        refers to the actual connection to a database at the DBAPI level;
+        a DBAPI connection may be pooled and reused for many operations.
+        In contrast, this event refers only to the production of a higher level
+        :class:`_engine.Connection` wrapper around such a DBAPI connection.
+
+        It also differs from the :meth:`_events.PoolEvents.checkout` event
+        in that it is specific to the :class:`_engine.Connection` object,
+        not the
+        DBAPI connection that :meth:`_events.PoolEvents.checkout` deals with,
+        although
+        this DBAPI connection is available here via the
+        :attr:`_engine.Connection.connection` attribute.
+        But note there can in fact
+        be multiple :meth:`_events.PoolEvents.checkout`
+        events within the lifespan
+        of a single :class:`_engine.Connection` object, if that
+        :class:`_engine.Connection`
+        is invalidated and re-established.
+
+        :param conn: :class:`_engine.Connection` object.
+
+        .. seealso::
+
+            :meth:`_events.PoolEvents.checkout`
+            the lower-level pool checkout event
+            for an individual DBAPI connection
+
+        """
+
+    def set_connection_execution_options(
+        self, conn: Connection, opts: Dict[str, Any]
+    ) -> None:
+        """Intercept when the :meth:`_engine.Connection.execution_options`
+        method is called.
+
+        This method is called after the new :class:`_engine.Connection`
+        has been
+        produced, with the newly updated execution options collection, but
+        before the :class:`.Dialect` has acted upon any of those new options.
+
+        Note that this method is not called when a new
+        :class:`_engine.Connection`
+        is produced which is inheriting execution options from its parent
+        :class:`_engine.Engine`; to intercept this condition, use the
+        :meth:`_events.ConnectionEvents.engine_connect` event.
+
+        :param conn: The newly copied :class:`_engine.Connection` object
+
+        :param opts: dictionary of options that were passed to the
+         :meth:`_engine.Connection.execution_options` method.
+         This dictionary may be modified in place to affect the ultimate
+         options which take effect.
+
+         .. versionadded:: 2.0 the ``opts`` dictionary may be modified
+            in place.
+
+
+        .. seealso::
+
+            :meth:`_events.ConnectionEvents.set_engine_execution_options`
+            - event
+            which is called when :meth:`_engine.Engine.execution_options`
+            is called.
+
+
+        """
+
+    def set_engine_execution_options(
+        self, engine: Engine, opts: Dict[str, Any]
+    ) -> None:
+        """Intercept when the :meth:`_engine.Engine.execution_options`
+        method is called.
+
+        The :meth:`_engine.Engine.execution_options` method produces a shallow
+        copy of the :class:`_engine.Engine` which stores the new options.
+        That new
+        :class:`_engine.Engine` is passed here.
+        A particular application of this
+        method is to add a :meth:`_events.ConnectionEvents.engine_connect`
+        event
+        handler to the given :class:`_engine.Engine`
+        which will perform some per-
+        :class:`_engine.Connection` task specific to these execution options.
+
+        :param conn: The newly copied :class:`_engine.Engine` object
+
+        :param opts: dictionary of options that were passed to the
+         :meth:`_engine.Connection.execution_options` method.
+         This dictionary may be modified in place to affect the ultimate
+         options which take effect.
+
+         .. versionadded:: 2.0 the ``opts`` dictionary may be modified
+            in place.
+
+        .. seealso::
+
+            :meth:`_events.ConnectionEvents.set_connection_execution_options`
+            - event
+            which is called when :meth:`_engine.Connection.execution_options`
+            is
+            called.
+
+        """
+
+    def engine_disposed(self, engine: Engine) -> None:
+        """Intercept when the :meth:`_engine.Engine.dispose` method is called.
+
+        The :meth:`_engine.Engine.dispose` method instructs the engine to
+        "dispose" of it's connection pool (e.g. :class:`_pool.Pool`), and
+        replaces it with a new one.  Disposing of the old pool has the
+        effect that existing checked-in connections are closed.  The new
+        pool does not establish any new connections until it is first used.
+
+        This event can be used to indicate that resources related to the
+        :class:`_engine.Engine` should also be cleaned up,
+        keeping in mind that the
+        :class:`_engine.Engine`
+        can still be used for new requests in which case
+        it re-acquires connection resources.
+
+        """
+
+    def begin(self, conn: Connection) -> None:
+        """Intercept begin() events.
+
+        :param conn: :class:`_engine.Connection` object
+
+        """
+
+    def rollback(self, conn: Connection) -> None:
+        """Intercept rollback() events, as initiated by a
+        :class:`.Transaction`.
+
+        Note that the :class:`_pool.Pool` also "auto-rolls back"
+        a DBAPI connection upon checkin, if the ``reset_on_return``
+        flag is set to its default value of ``'rollback'``.
+        To intercept this
+        rollback, use the :meth:`_events.PoolEvents.reset` hook.
+
+        :param conn: :class:`_engine.Connection` object
+
+        .. seealso::
+
+            :meth:`_events.PoolEvents.reset`
+
+        """
+
+    def commit(self, conn: Connection) -> None:
+        """Intercept commit() events, as initiated by a
+        :class:`.Transaction`.
+
+        Note that the :class:`_pool.Pool` may also "auto-commit"
+        a DBAPI connection upon checkin, if the ``reset_on_return``
+        flag is set to the value ``'commit'``.  To intercept this
+        commit, use the :meth:`_events.PoolEvents.reset` hook.
+
+        :param conn: :class:`_engine.Connection` object
+        """
+
+    def savepoint(self, conn: Connection, name: str) -> None:
+        """Intercept savepoint() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param name: specified name used for the savepoint.
+
+        """
+
+    def rollback_savepoint(
+        self, conn: Connection, name: str, context: None
+    ) -> None:
+        """Intercept rollback_savepoint() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param name: specified name used for the savepoint.
+        :param context: not used
+
+        """
+        # TODO: deprecate "context"
+
+    def release_savepoint(
+        self, conn: Connection, name: str, context: None
+    ) -> None:
+        """Intercept release_savepoint() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param name: specified name used for the savepoint.
+        :param context: not used
+
+        """
+        # TODO: deprecate "context"
+
+    def begin_twophase(self, conn: Connection, xid: Any) -> None:
+        """Intercept begin_twophase() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param xid: two-phase XID identifier
+
+        """
+
+    def prepare_twophase(self, conn: Connection, xid: Any) -> None:
+        """Intercept prepare_twophase() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param xid: two-phase XID identifier
+        """
+
+    def rollback_twophase(
+        self, conn: Connection, xid: Any, is_prepared: bool
+    ) -> None:
+        """Intercept rollback_twophase() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param xid: two-phase XID identifier
+        :param is_prepared: boolean, indicates if
+         :meth:`.TwoPhaseTransaction.prepare` was called.
+
+        """
+
+    def commit_twophase(
+        self, conn: Connection, xid: Any, is_prepared: bool
+    ) -> None:
+        """Intercept commit_twophase() events.
+
+        :param conn: :class:`_engine.Connection` object
+        :param xid: two-phase XID identifier
+        :param is_prepared: boolean, indicates if
+         :meth:`.TwoPhaseTransaction.prepare` was called.
+
+        """
+
+
+class DialectEvents(event.Events[Dialect]):
+    """event interface for execution-replacement functions.
+
+    These events allow direct instrumentation and replacement
+    of key dialect functions which interact with the DBAPI.
+
+    .. note::
+
+        :class:`.DialectEvents` hooks should be considered **semi-public**
+        and experimental.
+        These hooks are not for general use and are only for those situations
+        where intricate re-statement of DBAPI mechanics must be injected onto
+        an existing dialect.  For general-use statement-interception events,
+        please use the :class:`_events.ConnectionEvents` interface.
+
+    .. seealso::
+
+        :meth:`_events.ConnectionEvents.before_cursor_execute`
+
+        :meth:`_events.ConnectionEvents.before_execute`
+
+        :meth:`_events.ConnectionEvents.after_cursor_execute`
+
+        :meth:`_events.ConnectionEvents.after_execute`
+
+    """
+
+    _target_class_doc = "SomeEngine"
+    _dispatch_target = Dialect
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: event._EventKey[Dialect],
+        *,
+        retval: bool = False,
+        **kw: Any,
+    ) -> None:
+        target = event_key.dispatch_target
+
+        target._has_events = True
+        event_key.base_listen()
+
+    @classmethod
+    def _accept_with(
+        cls,
+        target: Union[Engine, Type[Engine], Dialect, Type[Dialect]],
+        identifier: str,
+    ) -> Optional[Union[Dialect, Type[Dialect]]]:
+        if isinstance(target, type):
+            if issubclass(target, Engine):
+                return Dialect
+            elif issubclass(target, Dialect):
+                return target
+        elif isinstance(target, Engine):
+            return target.dialect
+        elif isinstance(target, Dialect):
+            return target
+        elif isinstance(target, Connection) and identifier == "handle_error":
+            raise exc.InvalidRequestError(
+                "The handle_error() event hook as of SQLAlchemy 2.0 is "
+                "established on the Dialect, and may only be applied to the "
+                "Engine as a whole or to a specific Dialect as a whole, "
+                "not on a per-Connection basis."
+            )
+        elif hasattr(target, "_no_async_engine_events"):
+            target._no_async_engine_events()
+        else:
+            return None
+
+    def handle_error(
+        self, exception_context: ExceptionContext
+    ) -> Optional[BaseException]:
+        r"""Intercept all exceptions processed by the
+        :class:`_engine.Dialect`, typically but not limited to those
+        emitted within the scope of a :class:`_engine.Connection`.
+
+        .. versionchanged:: 2.0 the :meth:`.DialectEvents.handle_error` event
+           is moved to the :class:`.DialectEvents` class, moved from the
+           :class:`.ConnectionEvents` class, so that it may also participate in
+           the "pre ping" operation configured with the
+           :paramref:`_sa.create_engine.pool_pre_ping` parameter. The event
+           remains registered by using the :class:`_engine.Engine` as the event
+           target, however note that using the :class:`_engine.Connection` as
+           an event target for :meth:`.DialectEvents.handle_error` is no longer
+           supported.
+
+        This includes all exceptions emitted by the DBAPI as well as
+        within SQLAlchemy's statement invocation process, including
+        encoding errors and other statement validation errors.  Other areas
+        in which the event is invoked include transaction begin and end,
+        result row fetching, cursor creation.
+
+        Note that :meth:`.handle_error` may support new kinds of exceptions
+        and new calling scenarios at *any time*.  Code which uses this
+        event must expect new calling patterns to be present in minor
+        releases.
+
+        To support the wide variety of members that correspond to an exception,
+        as well as to allow extensibility of the event without backwards
+        incompatibility, the sole argument received is an instance of
+        :class:`.ExceptionContext`.   This object contains data members
+        representing detail about the exception.
+
+        Use cases supported by this hook include:
+
+        * read-only, low-level exception handling for logging and
+          debugging purposes
+        * Establishing whether a DBAPI connection error message indicates
+          that the database connection needs to be reconnected, including
+          for the "pre_ping" handler used by **some** dialects
+        * Establishing or disabling whether a connection or the owning
+          connection pool is invalidated or expired in response to a
+          specific exception
+        * exception re-writing
+
+        The hook is called while the cursor from the failed operation
+        (if any) is still open and accessible.   Special cleanup operations
+        can be called on this cursor; SQLAlchemy will attempt to close
+        this cursor subsequent to this hook being invoked.
+
+        As of SQLAlchemy 2.0, the "pre_ping" handler enabled using the
+        :paramref:`_sa.create_engine.pool_pre_ping` parameter will also
+        participate in the :meth:`.handle_error` process, **for those dialects
+        that rely upon disconnect codes to detect database liveness**. Note
+        that some dialects such as psycopg, psycopg2, and most MySQL dialects
+        make use of a native ``ping()`` method supplied by the DBAPI which does
+        not make use of disconnect codes.
+
+        .. versionchanged:: 2.0.0 The :meth:`.DialectEvents.handle_error`
+           event hook participates in connection pool "pre-ping" operations.
+           Within this usage, the :attr:`.ExceptionContext.engine` attribute
+           will be ``None``, however the :class:`.Dialect` in use is always
+           available via the :attr:`.ExceptionContext.dialect` attribute.
+
+        .. versionchanged:: 2.0.5 Added :attr:`.ExceptionContext.is_pre_ping`
+           attribute which will be set to ``True`` when the
+           :meth:`.DialectEvents.handle_error` event hook is triggered within
+           a connection pool pre-ping operation.
+
+        .. versionchanged:: 2.0.5 An issue was repaired that allows for the
+           PostgreSQL ``psycopg`` and ``psycopg2`` drivers, as well as all
+           MySQL drivers, to properly participate in the
+           :meth:`.DialectEvents.handle_error` event hook during
+           connection pool "pre-ping" operations; previously, the
+           implementation was non-working for these drivers.
+
+
+        A handler function has two options for replacing
+        the SQLAlchemy-constructed exception into one that is user
+        defined.   It can either raise this new exception directly, in
+        which case all further event listeners are bypassed and the
+        exception will be raised, after appropriate cleanup as taken
+        place::
+
+            @event.listens_for(Engine, "handle_error")
+            def handle_exception(context):
+                if isinstance(
+                    context.original_exception, psycopg2.OperationalError
+                ) and "failed" in str(context.original_exception):
+                    raise MySpecialException("failed operation")
+
+        .. warning::  Because the
+           :meth:`_events.DialectEvents.handle_error`
+           event specifically provides for exceptions to be re-thrown as
+           the ultimate exception raised by the failed statement,
+           **stack traces will be misleading** if the user-defined event
+           handler itself fails and throws an unexpected exception;
+           the stack trace may not illustrate the actual code line that
+           failed!  It is advised to code carefully here and use
+           logging and/or inline debugging if unexpected exceptions are
+           occurring.
+
+        Alternatively, a "chained" style of event handling can be
+        used, by configuring the handler with the ``retval=True``
+        modifier and returning the new exception instance from the
+        function.  In this case, event handling will continue onto the
+        next handler.   The "chained" exception is available using
+        :attr:`.ExceptionContext.chained_exception`::
+
+            @event.listens_for(Engine, "handle_error", retval=True)
+            def handle_exception(context):
+                if (
+                    context.chained_exception is not None
+                    and "special" in context.chained_exception.message
+                ):
+                    return MySpecialException(
+                        "failed", cause=context.chained_exception
+                    )
+
+        Handlers that return ``None`` may be used within the chain; when
+        a handler returns ``None``, the previous exception instance,
+        if any, is maintained as the current exception that is passed onto the
+        next handler.
+
+        When a custom exception is raised or returned, SQLAlchemy raises
+        this new exception as-is, it is not wrapped by any SQLAlchemy
+        object.  If the exception is not a subclass of
+        :class:`sqlalchemy.exc.StatementError`,
+        certain features may not be available; currently this includes
+        the ORM's feature of adding a detail hint about "autoflush" to
+        exceptions raised within the autoflush process.
+
+        :param context: an :class:`.ExceptionContext` object.  See this
+         class for details on all available members.
+
+
+        .. seealso::
+
+            :ref:`pool_new_disconnect_codes`
+
+        """
+
+    def do_connect(
+        self,
+        dialect: Dialect,
+        conn_rec: ConnectionPoolEntry,
+        cargs: Tuple[Any, ...],
+        cparams: Dict[str, Any],
+    ) -> Optional[DBAPIConnection]:
+        """Receive connection arguments before a connection is made.
+
+        This event is useful in that it allows the handler to manipulate the
+        cargs and/or cparams collections that control how the DBAPI
+        ``connect()`` function will be called. ``cargs`` will always be a
+        Python list that can be mutated in-place, and ``cparams`` a Python
+        dictionary that may also be mutated::
+
+            e = create_engine("postgresql+psycopg2://user@host/dbname")
+
+
+            @event.listens_for(e, "do_connect")
+            def receive_do_connect(dialect, conn_rec, cargs, cparams):
+                cparams["password"] = "some_password"
+
+        The event hook may also be used to override the call to ``connect()``
+        entirely, by returning a non-``None`` DBAPI connection object::
+
+            e = create_engine("postgresql+psycopg2://user@host/dbname")
+
+
+            @event.listens_for(e, "do_connect")
+            def receive_do_connect(dialect, conn_rec, cargs, cparams):
+                return psycopg2.connect(*cargs, **cparams)
+
+        .. seealso::
+
+            :ref:`custom_dbapi_args`
+
+        """
+
+    def do_executemany(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+        context: ExecutionContext,
+    ) -> Optional[Literal[True]]:
+        """Receive a cursor to have executemany() called.
+
+        Return the value True to halt further events from invoking,
+        and to indicate that the cursor execution has already taken
+        place within the event handler.
+
+        """
+
+    def do_execute_no_params(
+        self, cursor: DBAPICursor, statement: str, context: ExecutionContext
+    ) -> Optional[Literal[True]]:
+        """Receive a cursor to have execute() with no parameters called.
+
+        Return the value True to halt further events from invoking,
+        and to indicate that the cursor execution has already taken
+        place within the event handler.
+
+        """
+
+    def do_execute(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPISingleExecuteParams,
+        context: ExecutionContext,
+    ) -> Optional[Literal[True]]:
+        """Receive a cursor to have execute() called.
+
+        Return the value True to halt further events from invoking,
+        and to indicate that the cursor execution has already taken
+        place within the event handler.
+
+        """
+
+    def do_setinputsizes(
+        self,
+        inputsizes: Dict[BindParameter[Any], Any],
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIAnyExecuteParams,
+        context: ExecutionContext,
+    ) -> None:
+        """Receive the setinputsizes dictionary for possible modification.
+
+        This event is emitted in the case where the dialect makes use of the
+        DBAPI ``cursor.setinputsizes()`` method which passes information about
+        parameter binding for a particular statement.   The given
+        ``inputsizes`` dictionary will contain :class:`.BindParameter` objects
+        as keys, linked to DBAPI-specific type objects as values; for
+        parameters that are not bound, they are added to the dictionary with
+        ``None`` as the value, which means the parameter will not be included
+        in the ultimate setinputsizes call.   The event may be used to inspect
+        and/or log the datatypes that are being bound, as well as to modify the
+        dictionary in place.  Parameters can be added, modified, or removed
+        from this dictionary.   Callers will typically want to inspect the
+        :attr:`.BindParameter.type` attribute of the given bind objects in
+        order to make decisions about the DBAPI object.
+
+        After the event, the ``inputsizes`` dictionary is converted into
+        an appropriate datastructure to be passed to ``cursor.setinputsizes``;
+        either a list for a positional bound parameter execution style,
+        or a dictionary of string parameter keys to DBAPI type objects for
+        a named bound parameter execution style.
+
+        The setinputsizes hook overall is only used for dialects which include
+        the flag ``use_setinputsizes=True``.  Dialects which use this
+        include python-oracledb, cx_Oracle, pg8000, asyncpg, and pyodbc
+        dialects.
+
+        .. note::
+
+            For use with pyodbc, the ``use_setinputsizes`` flag
+            must be passed to the dialect, e.g.::
+
+                create_engine("mssql+pyodbc://...", use_setinputsizes=True)
+
+            .. seealso::
+
+                  :ref:`mssql_pyodbc_setinputsizes`
+
+        .. versionadded:: 1.2.9
+
+        .. seealso::
+
+            :ref:`cx_oracle_setinputsizes`
+
+        """
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/interfaces.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/interfaces.py
new file mode 100644
index 00000000..9fb39db7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/interfaces.py
@@ -0,0 +1,3406 @@
+# engine/interfaces.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Define core interfaces used by the engine system."""
+
+from __future__ import annotations
+
+from enum import Enum
+from types import ModuleType
+from typing import Any
+from typing import Awaitable
+from typing import Callable
+from typing import ClassVar
+from typing import Collection
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import util
+from ..event import EventTarget
+from ..pool import Pool
+from ..pool import PoolProxiedConnection
+from ..sql.compiler import Compiled as Compiled
+from ..sql.compiler import Compiled  # noqa
+from ..sql.compiler import TypeCompiler as TypeCompiler
+from ..sql.compiler import TypeCompiler  # noqa
+from ..util import immutabledict
+from ..util.concurrency import await_only
+from ..util.typing import Literal
+from ..util.typing import NotRequired
+from ..util.typing import Protocol
+from ..util.typing import TypedDict
+
+if TYPE_CHECKING:
+    from .base import Connection
+    from .base import Engine
+    from .cursor import CursorResult
+    from .url import URL
+    from ..event import _ListenerFnType
+    from ..event import dispatcher
+    from ..exc import StatementError
+    from ..sql import Executable
+    from ..sql.compiler import _InsertManyValuesBatch
+    from ..sql.compiler import DDLCompiler
+    from ..sql.compiler import IdentifierPreparer
+    from ..sql.compiler import InsertmanyvaluesSentinelOpts
+    from ..sql.compiler import Linting
+    from ..sql.compiler import SQLCompiler
+    from ..sql.elements import BindParameter
+    from ..sql.elements import ClauseElement
+    from ..sql.schema import Column
+    from ..sql.schema import DefaultGenerator
+    from ..sql.schema import SchemaItem
+    from ..sql.schema import Sequence as Sequence_SchemaItem
+    from ..sql.sqltypes import Integer
+    from ..sql.type_api import _TypeMemoDict
+    from ..sql.type_api import TypeEngine
+
+ConnectArgsType = Tuple[Sequence[str], MutableMapping[str, Any]]
+
+_T = TypeVar("_T", bound="Any")
+
+
+class CacheStats(Enum):
+    CACHE_HIT = 0
+    CACHE_MISS = 1
+    CACHING_DISABLED = 2
+    NO_CACHE_KEY = 3
+    NO_DIALECT_SUPPORT = 4
+
+
+class ExecuteStyle(Enum):
+    """indicates the :term:`DBAPI` cursor method that will be used to invoke
+    a statement."""
+
+    EXECUTE = 0
+    """indicates cursor.execute() will be used"""
+
+    EXECUTEMANY = 1
+    """indicates cursor.executemany() will be used."""
+
+    INSERTMANYVALUES = 2
+    """indicates cursor.execute() will be used with an INSERT where the
+    VALUES expression will be expanded to accommodate for multiple
+    parameter sets
+
+    .. seealso::
+
+        :ref:`engine_insertmanyvalues`
+
+    """
+
+
+class DBAPIConnection(Protocol):
+    """protocol representing a :pep:`249` database connection.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        `Connection Objects <https://www.python.org/dev/peps/pep-0249/#connection-objects>`_
+        - in :pep:`249`
+
+    """  # noqa: E501
+
+    def close(self) -> None: ...
+
+    def commit(self) -> None: ...
+
+    def cursor(self) -> DBAPICursor: ...
+
+    def rollback(self) -> None: ...
+
+    autocommit: bool
+
+
+class DBAPIType(Protocol):
+    """protocol representing a :pep:`249` database type.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        `Type Objects <https://www.python.org/dev/peps/pep-0249/#type-objects>`_
+        - in :pep:`249`
+
+    """  # noqa: E501
+
+
+class DBAPICursor(Protocol):
+    """protocol representing a :pep:`249` database cursor.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        `Cursor Objects <https://www.python.org/dev/peps/pep-0249/#cursor-objects>`_
+        - in :pep:`249`
+
+    """  # noqa: E501
+
+    @property
+    def description(
+        self,
+    ) -> _DBAPICursorDescription:
+        """The description attribute of the Cursor.
+
+        .. seealso::
+
+            `cursor.description <https://www.python.org/dev/peps/pep-0249/#description>`_
+            - in :pep:`249`
+
+
+        """  # noqa: E501
+        ...
+
+    @property
+    def rowcount(self) -> int: ...
+
+    arraysize: int
+
+    lastrowid: int
+
+    def close(self) -> None: ...
+
+    def execute(
+        self,
+        operation: Any,
+        parameters: Optional[_DBAPISingleExecuteParams] = None,
+    ) -> Any: ...
+
+    def executemany(
+        self,
+        operation: Any,
+        parameters: _DBAPIMultiExecuteParams,
+    ) -> Any: ...
+
+    def fetchone(self) -> Optional[Any]: ...
+
+    def fetchmany(self, size: int = ...) -> Sequence[Any]: ...
+
+    def fetchall(self) -> Sequence[Any]: ...
+
+    def setinputsizes(self, sizes: Sequence[Any]) -> None: ...
+
+    def setoutputsize(self, size: Any, column: Any) -> None: ...
+
+    def callproc(
+        self, procname: str, parameters: Sequence[Any] = ...
+    ) -> Any: ...
+
+    def nextset(self) -> Optional[bool]: ...
+
+    def __getattr__(self, key: str) -> Any: ...
+
+
+_CoreSingleExecuteParams = Mapping[str, Any]
+_MutableCoreSingleExecuteParams = MutableMapping[str, Any]
+_CoreMultiExecuteParams = Sequence[_CoreSingleExecuteParams]
+_CoreAnyExecuteParams = Union[
+    _CoreMultiExecuteParams, _CoreSingleExecuteParams
+]
+
+_DBAPISingleExecuteParams = Union[Sequence[Any], _CoreSingleExecuteParams]
+
+_DBAPIMultiExecuteParams = Union[
+    Sequence[Sequence[Any]], _CoreMultiExecuteParams
+]
+_DBAPIAnyExecuteParams = Union[
+    _DBAPIMultiExecuteParams, _DBAPISingleExecuteParams
+]
+_DBAPICursorDescription = Sequence[
+    Tuple[
+        str,
+        "DBAPIType",
+        Optional[int],
+        Optional[int],
+        Optional[int],
+        Optional[int],
+        Optional[bool],
+    ]
+]
+
+_AnySingleExecuteParams = _DBAPISingleExecuteParams
+_AnyMultiExecuteParams = _DBAPIMultiExecuteParams
+_AnyExecuteParams = _DBAPIAnyExecuteParams
+
+CompiledCacheType = MutableMapping[Any, "Compiled"]
+SchemaTranslateMapType = Mapping[Optional[str], Optional[str]]
+
+_ImmutableExecuteOptions = immutabledict[str, Any]
+
+_ParamStyle = Literal[
+    "qmark", "numeric", "named", "format", "pyformat", "numeric_dollar"
+]
+
+_GenericSetInputSizesType = List[Tuple[str, Any, "TypeEngine[Any]"]]
+
+IsolationLevel = Literal[
+    "SERIALIZABLE",
+    "REPEATABLE READ",
+    "READ COMMITTED",
+    "READ UNCOMMITTED",
+    "AUTOCOMMIT",
+]
+
+
+class _CoreKnownExecutionOptions(TypedDict, total=False):
+    compiled_cache: Optional[CompiledCacheType]
+    logging_token: str
+    isolation_level: IsolationLevel
+    no_parameters: bool
+    stream_results: bool
+    max_row_buffer: int
+    yield_per: int
+    insertmanyvalues_page_size: int
+    schema_translate_map: Optional[SchemaTranslateMapType]
+    preserve_rowcount: bool
+
+
+_ExecuteOptions = immutabledict[str, Any]
+CoreExecuteOptionsParameter = Union[
+    _CoreKnownExecutionOptions, Mapping[str, Any]
+]
+
+
+class ReflectedIdentity(TypedDict):
+    """represent the reflected IDENTITY structure of a column, corresponding
+    to the :class:`_schema.Identity` construct.
+
+    The :class:`.ReflectedIdentity` structure is part of the
+    :class:`.ReflectedColumn` structure, which is returned by the
+    :meth:`.Inspector.get_columns` method.
+
+    """
+
+    always: bool
+    """type of identity column"""
+
+    on_null: bool
+    """indicates ON NULL"""
+
+    start: int
+    """starting index of the sequence"""
+
+    increment: int
+    """increment value of the sequence"""
+
+    minvalue: int
+    """the minimum value of the sequence."""
+
+    maxvalue: int
+    """the maximum value of the sequence."""
+
+    nominvalue: bool
+    """no minimum value of the sequence."""
+
+    nomaxvalue: bool
+    """no maximum value of the sequence."""
+
+    cycle: bool
+    """allows the sequence to wrap around when the maxvalue
+    or minvalue has been reached."""
+
+    cache: Optional[int]
+    """number of future values in the
+    sequence which are calculated in advance."""
+
+    order: bool
+    """if true, renders the ORDER keyword."""
+
+
+class ReflectedComputed(TypedDict):
+    """Represent the reflected elements of a computed column, corresponding
+    to the :class:`_schema.Computed` construct.
+
+    The :class:`.ReflectedComputed` structure is part of the
+    :class:`.ReflectedColumn` structure, which is returned by the
+    :meth:`.Inspector.get_columns` method.
+
+    """
+
+    sqltext: str
+    """the expression used to generate this column returned
+    as a string SQL expression"""
+
+    persisted: NotRequired[bool]
+    """indicates if the value is stored in the table or computed on demand"""
+
+
+class ReflectedColumn(TypedDict):
+    """Dictionary representing the reflected elements corresponding to
+    a :class:`_schema.Column` object.
+
+    The :class:`.ReflectedColumn` structure is returned by the
+    :class:`.Inspector.get_columns` method.
+
+    """
+
+    name: str
+    """column name"""
+
+    type: TypeEngine[Any]
+    """column type represented as a :class:`.TypeEngine` instance."""
+
+    nullable: bool
+    """boolean flag if the column is NULL or NOT NULL"""
+
+    default: Optional[str]
+    """column default expression as a SQL string"""
+
+    autoincrement: NotRequired[bool]
+    """database-dependent autoincrement flag.
+
+    This flag indicates if the column has a database-side "autoincrement"
+    flag of some kind.   Within SQLAlchemy, other kinds of columns may
+    also act as an "autoincrement" column without necessarily having
+    such a flag on them.
+
+    See :paramref:`_schema.Column.autoincrement` for more background on
+    "autoincrement".
+
+    """
+
+    comment: NotRequired[Optional[str]]
+    """comment for the column, if present.
+    Only some dialects return this key
+    """
+
+    computed: NotRequired[ReflectedComputed]
+    """indicates that this column is computed by the database.
+    Only some dialects return this key.
+
+    .. versionadded:: 1.3.16 - added support for computed reflection.
+    """
+
+    identity: NotRequired[ReflectedIdentity]
+    """indicates this column is an IDENTITY column.
+    Only some dialects return this key.
+
+    .. versionadded:: 1.4 - added support for identity column reflection.
+    """
+
+    dialect_options: NotRequired[Dict[str, Any]]
+    """Additional dialect-specific options detected for this reflected
+    object"""
+
+
+class ReflectedConstraint(TypedDict):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.Constraint`
+
+    A base class for all constraints
+    """
+
+    name: Optional[str]
+    """constraint name"""
+
+    comment: NotRequired[Optional[str]]
+    """comment for the constraint, if present"""
+
+
+class ReflectedCheckConstraint(ReflectedConstraint):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.CheckConstraint`.
+
+    The :class:`.ReflectedCheckConstraint` structure is returned by the
+    :meth:`.Inspector.get_check_constraints` method.
+
+    """
+
+    sqltext: str
+    """the check constraint's SQL expression"""
+
+    dialect_options: NotRequired[Dict[str, Any]]
+    """Additional dialect-specific options detected for this check constraint
+
+    .. versionadded:: 1.3.8
+    """
+
+
+class ReflectedUniqueConstraint(ReflectedConstraint):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.UniqueConstraint`.
+
+    The :class:`.ReflectedUniqueConstraint` structure is returned by the
+    :meth:`.Inspector.get_unique_constraints` method.
+
+    """
+
+    column_names: List[str]
+    """column names which comprise the unique constraint"""
+
+    duplicates_index: NotRequired[Optional[str]]
+    "Indicates if this unique constraint duplicates an index with this name"
+
+    dialect_options: NotRequired[Dict[str, Any]]
+    """Additional dialect-specific options detected for this unique
+    constraint"""
+
+
+class ReflectedPrimaryKeyConstraint(ReflectedConstraint):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.PrimaryKeyConstraint`.
+
+    The :class:`.ReflectedPrimaryKeyConstraint` structure is returned by the
+    :meth:`.Inspector.get_pk_constraint` method.
+
+    """
+
+    constrained_columns: List[str]
+    """column names which comprise the primary key"""
+
+    dialect_options: NotRequired[Dict[str, Any]]
+    """Additional dialect-specific options detected for this primary key"""
+
+
+class ReflectedForeignKeyConstraint(ReflectedConstraint):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.ForeignKeyConstraint`.
+
+    The :class:`.ReflectedForeignKeyConstraint` structure is returned by
+    the :meth:`.Inspector.get_foreign_keys` method.
+
+    """
+
+    constrained_columns: List[str]
+    """local column names which comprise the foreign key"""
+
+    referred_schema: Optional[str]
+    """schema name of the table being referred"""
+
+    referred_table: str
+    """name of the table being referred"""
+
+    referred_columns: List[str]
+    """referred column names that correspond to ``constrained_columns``"""
+
+    options: NotRequired[Dict[str, Any]]
+    """Additional options detected for this foreign key constraint"""
+
+
+class ReflectedIndex(TypedDict):
+    """Dictionary representing the reflected elements corresponding to
+    :class:`.Index`.
+
+    The :class:`.ReflectedIndex` structure is returned by the
+    :meth:`.Inspector.get_indexes` method.
+
+    """
+
+    name: Optional[str]
+    """index name"""
+
+    column_names: List[Optional[str]]
+    """column names which the index references.
+    An element of this list is ``None`` if it's an expression and is
+    returned in the ``expressions`` list.
+    """
+
+    expressions: NotRequired[List[str]]
+    """Expressions that compose the index. This list, when present, contains
+    both plain column names (that are also in ``column_names``) and
+    expressions (that are ``None`` in ``column_names``).
+    """
+
+    unique: bool
+    """whether or not the index has a unique flag"""
+
+    duplicates_constraint: NotRequired[Optional[str]]
+    "Indicates if this index mirrors a constraint with this name"
+
+    include_columns: NotRequired[List[str]]
+    """columns to include in the INCLUDE clause for supporting databases.
+
+    .. deprecated:: 2.0
+
+        Legacy value, will be replaced with
+        ``index_dict["dialect_options"]["<dialect name>_include"]``
+
+    """
+
+    column_sorting: NotRequired[Dict[str, Tuple[str]]]
+    """optional dict mapping column names or expressions to tuple of sort
+    keywords, which may include ``asc``, ``desc``, ``nulls_first``,
+    ``nulls_last``.
+
+    .. versionadded:: 1.3.5
+    """
+
+    dialect_options: NotRequired[Dict[str, Any]]
+    """Additional dialect-specific options detected for this index"""
+
+
+class ReflectedTableComment(TypedDict):
+    """Dictionary representing the reflected comment corresponding to
+    the :attr:`_schema.Table.comment` attribute.
+
+    The :class:`.ReflectedTableComment` structure is returned by the
+    :meth:`.Inspector.get_table_comment` method.
+
+    """
+
+    text: Optional[str]
+    """text of the comment"""
+
+
+class BindTyping(Enum):
+    """Define different methods of passing typing information for
+    bound parameters in a statement to the database driver.
+
+    .. versionadded:: 2.0
+
+    """
+
+    NONE = 1
+    """No steps are taken to pass typing information to the database driver.
+
+    This is the default behavior for databases such as SQLite, MySQL / MariaDB,
+    SQL Server.
+
+    """
+
+    SETINPUTSIZES = 2
+    """Use the pep-249 setinputsizes method.
+
+    This is only implemented for DBAPIs that support this method and for which
+    the SQLAlchemy dialect has the appropriate infrastructure for that dialect
+    set up.  Current dialects include python-oracledb, cx_Oracle as well as
+    optional support for SQL Server using pyodbc.
+
+    When using setinputsizes, dialects also have a means of only using the
+    method for certain datatypes using include/exclude lists.
+
+    When SETINPUTSIZES is used, the :meth:`.Dialect.do_set_input_sizes` method
+    is called for each statement executed which has bound parameters.
+
+    """
+
+    RENDER_CASTS = 3
+    """Render casts or other directives in the SQL string.
+
+    This method is used for all PostgreSQL dialects, including asyncpg,
+    pg8000, psycopg, psycopg2.   Dialects which implement this can choose
+    which kinds of datatypes are explicitly cast in SQL statements and which
+    aren't.
+
+    When RENDER_CASTS is used, the compiler will invoke the
+    :meth:`.SQLCompiler.render_bind_cast` method for the rendered
+    string representation of each :class:`.BindParameter` object whose
+    dialect-level type sets the :attr:`.TypeEngine.render_bind_cast` attribute.
+
+    The :meth:`.SQLCompiler.render_bind_cast` is also used to render casts
+    for one form of "insertmanyvalues" query, when both
+    :attr:`.InsertmanyvaluesSentinelOpts.USE_INSERT_FROM_SELECT` and
+    :attr:`.InsertmanyvaluesSentinelOpts.RENDER_SELECT_COL_CASTS` are set,
+    where the casts are applied to the intermediary columns e.g.
+    "INSERT INTO t (a, b, c) SELECT p0::TYP, p1::TYP, p2::TYP "
+    "FROM (VALUES (?, ?), (?, ?), ...)".
+
+    .. versionadded:: 2.0.10 - :meth:`.SQLCompiler.render_bind_cast` is now
+       used within some elements of the "insertmanyvalues" implementation.
+
+
+    """
+
+
+VersionInfoType = Tuple[Union[int, str], ...]
+TableKey = Tuple[Optional[str], str]
+
+
+class Dialect(EventTarget):
+    """Define the behavior of a specific database and DB-API combination.
+
+    Any aspect of metadata definition, SQL query generation,
+    execution, result-set handling, or anything else which varies
+    between databases is defined under the general category of the
+    Dialect.  The Dialect acts as a factory for other
+    database-specific object implementations including
+    ExecutionContext, Compiled, DefaultGenerator, and TypeEngine.
+
+    .. note:: Third party dialects should not subclass :class:`.Dialect`
+       directly.  Instead, subclass :class:`.default.DefaultDialect` or
+       descendant class.
+
+    """
+
+    CACHE_HIT = CacheStats.CACHE_HIT
+    CACHE_MISS = CacheStats.CACHE_MISS
+    CACHING_DISABLED = CacheStats.CACHING_DISABLED
+    NO_CACHE_KEY = CacheStats.NO_CACHE_KEY
+    NO_DIALECT_SUPPORT = CacheStats.NO_DIALECT_SUPPORT
+
+    dispatch: dispatcher[Dialect]
+
+    name: str
+    """identifying name for the dialect from a DBAPI-neutral point of view
+      (i.e. 'sqlite')
+    """
+
+    driver: str
+    """identifying name for the dialect's DBAPI"""
+
+    dialect_description: str
+
+    dbapi: Optional[ModuleType]
+    """A reference to the DBAPI module object itself.
+
+    SQLAlchemy dialects import DBAPI modules using the classmethod
+    :meth:`.Dialect.import_dbapi`. The rationale is so that any dialect
+    module can be imported and used to generate SQL statements without the
+    need for the actual DBAPI driver to be installed.  Only when an
+    :class:`.Engine` is constructed using :func:`.create_engine` does the
+    DBAPI get imported; at that point, the creation process will assign
+    the DBAPI module to this attribute.
+
+    Dialects should therefore implement :meth:`.Dialect.import_dbapi`
+    which will import the necessary module and return it, and then refer
+    to ``self.dbapi`` in dialect code in order to refer to the DBAPI module
+    contents.
+
+    .. versionchanged:: The :attr:`.Dialect.dbapi` attribute is exclusively
+       used as the per-:class:`.Dialect`-instance reference to the DBAPI
+       module.   The previous not-fully-documented ``.Dialect.dbapi()``
+       classmethod is deprecated and replaced by :meth:`.Dialect.import_dbapi`.
+
+    """
+
+    @util.non_memoized_property
+    def loaded_dbapi(self) -> ModuleType:
+        """same as .dbapi, but is never None; will raise an error if no
+        DBAPI was set up.
+
+        .. versionadded:: 2.0
+
+        """
+        raise NotImplementedError()
+
+    positional: bool
+    """True if the paramstyle for this Dialect is positional."""
+
+    paramstyle: str
+    """the paramstyle to be used (some DB-APIs support multiple
+      paramstyles).
+    """
+
+    compiler_linting: Linting
+
+    statement_compiler: Type[SQLCompiler]
+    """a :class:`.Compiled` class used to compile SQL statements"""
+
+    ddl_compiler: Type[DDLCompiler]
+    """a :class:`.Compiled` class used to compile DDL statements"""
+
+    type_compiler_cls: ClassVar[Type[TypeCompiler]]
+    """a :class:`.Compiled` class used to compile SQL type objects
+
+    .. versionadded:: 2.0
+
+    """
+
+    type_compiler_instance: TypeCompiler
+    """instance of a :class:`.Compiled` class used to compile SQL type
+    objects
+
+    .. versionadded:: 2.0
+
+    """
+
+    type_compiler: Any
+    """legacy; this is a TypeCompiler class at the class level, a
+    TypeCompiler instance at the instance level.
+
+    Refer to type_compiler_instance instead.
+
+    """
+
+    preparer: Type[IdentifierPreparer]
+    """a :class:`.IdentifierPreparer` class used to
+    quote identifiers.
+    """
+
+    identifier_preparer: IdentifierPreparer
+    """This element will refer to an instance of :class:`.IdentifierPreparer`
+    once a :class:`.DefaultDialect` has been constructed.
+
+    """
+
+    server_version_info: Optional[Tuple[Any, ...]]
+    """a tuple containing a version number for the DB backend in use.
+
+    This value is only available for supporting dialects, and is
+    typically populated during the initial connection to the database.
+    """
+
+    default_schema_name: Optional[str]
+    """the name of the default schema.  This value is only available for
+    supporting dialects, and is typically populated during the
+    initial connection to the database.
+
+    """
+
+    # NOTE: this does not take into effect engine-level isolation level.
+    # not clear if this should be changed, seems like it should
+    default_isolation_level: Optional[IsolationLevel]
+    """the isolation that is implicitly present on new connections"""
+
+    # create_engine()  -> isolation_level  currently goes here
+    _on_connect_isolation_level: Optional[IsolationLevel]
+
+    execution_ctx_cls: Type[ExecutionContext]
+    """a :class:`.ExecutionContext` class used to handle statement execution"""
+
+    execute_sequence_format: Union[
+        Type[Tuple[Any, ...]], Type[Tuple[List[Any]]]
+    ]
+    """either the 'tuple' or 'list' type, depending on what cursor.execute()
+    accepts for the second argument (they vary)."""
+
+    supports_alter: bool
+    """``True`` if the database supports ``ALTER TABLE`` - used only for
+    generating foreign key constraints in certain circumstances
+    """
+
+    max_identifier_length: int
+    """The maximum length of identifier names."""
+
+    supports_server_side_cursors: bool
+    """indicates if the dialect supports server side cursors"""
+
+    server_side_cursors: bool
+    """deprecated; indicates if the dialect should attempt to use server
+    side cursors by default"""
+
+    supports_sane_rowcount: bool
+    """Indicate whether the dialect properly implements rowcount for
+      ``UPDATE`` and ``DELETE`` statements.
+    """
+
+    supports_sane_multi_rowcount: bool
+    """Indicate whether the dialect properly implements rowcount for
+      ``UPDATE`` and ``DELETE`` statements when executed via
+      executemany.
+    """
+
+    supports_empty_insert: bool
+    """dialect supports INSERT () VALUES (), i.e. a plain INSERT with no
+    columns in it.
+
+    This is not usually supported; an "empty" insert is typically
+    suited using either "INSERT..DEFAULT VALUES" or
+    "INSERT ... (col) VALUES (DEFAULT)".
+
+    """
+
+    supports_default_values: bool
+    """dialect supports INSERT... DEFAULT VALUES syntax"""
+
+    supports_default_metavalue: bool
+    """dialect supports INSERT...(col) VALUES (DEFAULT) syntax.
+
+    Most databases support this in some way, e.g. SQLite supports it using
+    ``VALUES (NULL)``.    MS SQL Server supports the syntax also however
+    is the only included dialect where we have this disabled, as
+    MSSQL does not support the field for the IDENTITY column, which is
+    usually where we like to make use of the feature.
+
+    """
+
+    default_metavalue_token: str = "DEFAULT"
+    """for INSERT... VALUES (DEFAULT) syntax, the token to put in the
+    parenthesis.
+
+    E.g. for SQLite this is the keyword "NULL".
+
+    """
+
+    supports_multivalues_insert: bool
+    """Target database supports INSERT...VALUES with multiple value
+    sets, i.e. INSERT INTO table (cols) VALUES (...), (...), (...), ...
+
+    """
+
+    insert_executemany_returning: bool
+    """dialect / driver / database supports some means of providing
+    INSERT...RETURNING support when dialect.do_executemany() is used.
+
+    """
+
+    insert_executemany_returning_sort_by_parameter_order: bool
+    """dialect / driver / database supports some means of providing
+    INSERT...RETURNING support when dialect.do_executemany() is used
+    along with the :paramref:`_dml.Insert.returning.sort_by_parameter_order`
+    parameter being set.
+
+    """
+
+    update_executemany_returning: bool
+    """dialect supports UPDATE..RETURNING with executemany."""
+
+    delete_executemany_returning: bool
+    """dialect supports DELETE..RETURNING with executemany."""
+
+    use_insertmanyvalues: bool
+    """if True, indicates "insertmanyvalues" functionality should be used
+    to allow for ``insert_executemany_returning`` behavior, if possible.
+
+    In practice, setting this to True means:
+
+    if ``supports_multivalues_insert``, ``insert_returning`` and
+    ``use_insertmanyvalues`` are all True, the SQL compiler will produce
+    an INSERT that will be interpreted by the :class:`.DefaultDialect`
+    as an :attr:`.ExecuteStyle.INSERTMANYVALUES` execution that allows
+    for INSERT of many rows with RETURNING by rewriting a single-row
+    INSERT statement to have multiple VALUES clauses, also executing
+    the statement multiple times for a series of batches when large numbers
+    of rows are given.
+
+    The parameter is False for the default dialect, and is set to True for
+    SQLAlchemy internal dialects SQLite, MySQL/MariaDB, PostgreSQL, SQL Server.
+    It remains at False for Oracle Database, which provides native "executemany
+    with RETURNING" support and also does not support
+    ``supports_multivalues_insert``.  For MySQL/MariaDB, those MySQL dialects
+    that don't support RETURNING will not report
+    ``insert_executemany_returning`` as True.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`engine_insertmanyvalues`
+
+    """
+
+    use_insertmanyvalues_wo_returning: bool
+    """if True, and use_insertmanyvalues is also True, INSERT statements
+    that don't include RETURNING will also use "insertmanyvalues".
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`engine_insertmanyvalues`
+
+    """
+
+    insertmanyvalues_implicit_sentinel: InsertmanyvaluesSentinelOpts
+    """Options indicating the database supports a form of bulk INSERT where
+    the autoincrement integer primary key can be reliably used as an ordering
+    for INSERTed rows.
+
+    .. versionadded:: 2.0.10
+
+    .. seealso::
+
+        :ref:`engine_insertmanyvalues_returning_order`
+
+    """
+
+    insertmanyvalues_page_size: int
+    """Number of rows to render into an individual INSERT..VALUES() statement
+    for :attr:`.ExecuteStyle.INSERTMANYVALUES` executions.
+
+    The default dialect defaults this to 1000.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :paramref:`_engine.Connection.execution_options.insertmanyvalues_page_size` -
+        execution option available on :class:`_engine.Connection`, statements
+
+    """  # noqa: E501
+
+    insertmanyvalues_max_parameters: int
+    """Alternate to insertmanyvalues_page_size, will additionally limit
+    page size based on number of parameters total in the statement.
+
+
+    """
+
+    preexecute_autoincrement_sequences: bool
+    """True if 'implicit' primary key functions must be executed separately
+      in order to get their value, if RETURNING is not used.
+
+      This is currently oriented towards PostgreSQL when the
+      ``implicit_returning=False`` parameter is used on a :class:`.Table`
+      object.
+
+    """
+
+    insert_returning: bool
+    """if the dialect supports RETURNING with INSERT
+
+    .. versionadded:: 2.0
+
+    """
+
+    update_returning: bool
+    """if the dialect supports RETURNING with UPDATE
+
+    .. versionadded:: 2.0
+
+    """
+
+    update_returning_multifrom: bool
+    """if the dialect supports RETURNING with UPDATE..FROM
+
+    .. versionadded:: 2.0
+
+    """
+
+    delete_returning: bool
+    """if the dialect supports RETURNING with DELETE
+
+    .. versionadded:: 2.0
+
+    """
+
+    delete_returning_multifrom: bool
+    """if the dialect supports RETURNING with DELETE..FROM
+
+    .. versionadded:: 2.0
+
+    """
+
+    favor_returning_over_lastrowid: bool
+    """for backends that support both a lastrowid and a RETURNING insert
+    strategy, favor RETURNING for simple single-int pk inserts.
+
+    cursor.lastrowid tends to be more performant on most backends.
+
+    """
+
+    supports_identity_columns: bool
+    """target database supports IDENTITY"""
+
+    cte_follows_insert: bool
+    """target database, when given a CTE with an INSERT statement, needs
+    the CTE to be below the INSERT"""
+
+    colspecs: MutableMapping[Type[TypeEngine[Any]], Type[TypeEngine[Any]]]
+    """A dictionary of TypeEngine classes from sqlalchemy.types mapped
+      to subclasses that are specific to the dialect class.  This
+      dictionary is class-level only and is not accessed from the
+      dialect instance itself.
+    """
+
+    supports_sequences: bool
+    """Indicates if the dialect supports CREATE SEQUENCE or similar."""
+
+    sequences_optional: bool
+    """If True, indicates if the :paramref:`_schema.Sequence.optional`
+      parameter on the :class:`_schema.Sequence` construct
+      should signal to not generate a CREATE SEQUENCE. Applies only to
+      dialects that support sequences. Currently used only to allow PostgreSQL
+      SERIAL to be used on a column that specifies Sequence() for usage on
+      other backends.
+    """
+
+    default_sequence_base: int
+    """the default value that will be rendered as the "START WITH" portion of
+    a CREATE SEQUENCE DDL statement.
+
+    """
+
+    supports_native_enum: bool
+    """Indicates if the dialect supports a native ENUM construct.
+      This will prevent :class:`_types.Enum` from generating a CHECK
+      constraint when that type is used in "native" mode.
+    """
+
+    supports_native_boolean: bool
+    """Indicates if the dialect supports a native boolean construct.
+      This will prevent :class:`_types.Boolean` from generating a CHECK
+      constraint when that type is used.
+    """
+
+    supports_native_decimal: bool
+    """indicates if Decimal objects are handled and returned for precision
+    numeric types, or if floats are returned"""
+
+    supports_native_uuid: bool
+    """indicates if Python UUID() objects are handled natively by the
+    driver for SQL UUID datatypes.
+
+    .. versionadded:: 2.0
+
+    """
+
+    returns_native_bytes: bool
+    """indicates if Python bytes() objects are returned natively by the
+    driver for SQL "binary" datatypes.
+
+    .. versionadded:: 2.0.11
+
+    """
+
+    construct_arguments: Optional[
+        List[Tuple[Type[Union[SchemaItem, ClauseElement]], Mapping[str, Any]]]
+    ] = None
+    """Optional set of argument specifiers for various SQLAlchemy
+    constructs, typically schema items.
+
+    To implement, establish as a series of tuples, as in::
+
+        construct_arguments = [
+            (schema.Index, {"using": False, "where": None, "ops": None}),
+        ]
+
+    If the above construct is established on the PostgreSQL dialect,
+    the :class:`.Index` construct will now accept the keyword arguments
+    ``postgresql_using``, ``postgresql_where``, nad ``postgresql_ops``.
+    Any other argument specified to the constructor of :class:`.Index`
+    which is prefixed with ``postgresql_`` will raise :class:`.ArgumentError`.
+
+    A dialect which does not include a ``construct_arguments`` member will
+    not participate in the argument validation system.  For such a dialect,
+    any argument name is accepted by all participating constructs, within
+    the namespace of arguments prefixed with that dialect name.  The rationale
+    here is so that third-party dialects that haven't yet implemented this
+    feature continue to function in the old way.
+
+    .. seealso::
+
+        :class:`.DialectKWArgs` - implementing base class which consumes
+        :attr:`.DefaultDialect.construct_arguments`
+
+
+    """
+
+    reflection_options: Sequence[str] = ()
+    """Sequence of string names indicating keyword arguments that can be
+    established on a :class:`.Table` object which will be passed as
+    "reflection options" when using :paramref:`.Table.autoload_with`.
+
+    Current example is "oracle_resolve_synonyms" in the Oracle Database
+    dialects.
+
+    """
+
+    dbapi_exception_translation_map: Mapping[str, str] = util.EMPTY_DICT
+    """A dictionary of names that will contain as values the names of
+       pep-249 exceptions ("IntegrityError", "OperationalError", etc)
+       keyed to alternate class names, to support the case where a
+       DBAPI has exception classes that aren't named as they are
+       referred to (e.g. IntegrityError = MyException).   In the vast
+       majority of cases this dictionary is empty.
+    """
+
+    supports_comments: bool
+    """Indicates the dialect supports comment DDL on tables and columns."""
+
+    inline_comments: bool
+    """Indicates the dialect supports comment DDL that's inline with the
+    definition of a Table or Column.  If False, this implies that ALTER must
+    be used to set table and column comments."""
+
+    supports_constraint_comments: bool
+    """Indicates if the dialect supports comment DDL on constraints.
+
+    .. versionadded:: 2.0
+    """
+
+    _has_events = False
+
+    supports_statement_cache: bool = True
+    """indicates if this dialect supports caching.
+
+    All dialects that are compatible with statement caching should set this
+    flag to True directly on each dialect class and subclass that supports
+    it.  SQLAlchemy tests that this flag is locally present on each dialect
+    subclass before it will use statement caching.  This is to provide
+    safety for legacy or new dialects that are not yet fully tested to be
+    compliant with SQL statement caching.
+
+    .. versionadded:: 1.4.5
+
+    .. seealso::
+
+        :ref:`engine_thirdparty_caching`
+
+    """
+
+    _supports_statement_cache: bool
+    """internal evaluation for supports_statement_cache"""
+
+    bind_typing = BindTyping.NONE
+    """define a means of passing typing information to the database and/or
+    driver for bound parameters.
+
+    See :class:`.BindTyping` for values.
+
+    .. versionadded:: 2.0
+
+    """
+
+    is_async: bool
+    """Whether or not this dialect is intended for asyncio use."""
+
+    has_terminate: bool
+    """Whether or not this dialect has a separate "terminate" implementation
+    that does not block or require awaiting."""
+
+    engine_config_types: Mapping[str, Any]
+    """a mapping of string keys that can be in an engine config linked to
+    type conversion functions.
+
+    """
+
+    label_length: Optional[int]
+    """optional user-defined max length for SQL labels"""
+
+    include_set_input_sizes: Optional[Set[Any]]
+    """set of DBAPI type objects that should be included in
+    automatic cursor.setinputsizes() calls.
+
+    This is only used if bind_typing is BindTyping.SET_INPUT_SIZES
+
+    """
+
+    exclude_set_input_sizes: Optional[Set[Any]]
+    """set of DBAPI type objects that should be excluded in
+    automatic cursor.setinputsizes() calls.
+
+    This is only used if bind_typing is BindTyping.SET_INPUT_SIZES
+
+    """
+
+    supports_simple_order_by_label: bool
+    """target database supports ORDER BY <labelname>, where <labelname>
+    refers to a label in the columns clause of the SELECT"""
+
+    div_is_floordiv: bool
+    """target database treats the / division operator as "floor division" """
+
+    tuple_in_values: bool
+    """target database supports tuple IN, i.e. (x, y) IN ((q, p), (r, z))"""
+
+    _bind_typing_render_casts: bool
+
+    _type_memos: MutableMapping[TypeEngine[Any], _TypeMemoDict]
+
+    def _builtin_onconnect(self) -> Optional[_ListenerFnType]:
+        raise NotImplementedError()
+
+    def create_connect_args(self, url: URL) -> ConnectArgsType:
+        """Build DB-API compatible connection arguments.
+
+        Given a :class:`.URL` object, returns a tuple
+        consisting of a ``(*args, **kwargs)`` suitable to send directly
+        to the dbapi's connect function.   The arguments are sent to the
+        :meth:`.Dialect.connect` method which then runs the DBAPI-level
+        ``connect()`` function.
+
+        The method typically makes use of the
+        :meth:`.URL.translate_connect_args`
+        method in order to generate a dictionary of options.
+
+        The default implementation is::
+
+            def create_connect_args(self, url):
+                opts = url.translate_connect_args()
+                opts.update(url.query)
+                return ([], opts)
+
+        :param url: a :class:`.URL` object
+
+        :return: a tuple of ``(*args, **kwargs)`` which will be passed to the
+         :meth:`.Dialect.connect` method.
+
+        .. seealso::
+
+            :meth:`.URL.translate_connect_args`
+
+        """
+
+        raise NotImplementedError()
+
+    @classmethod
+    def import_dbapi(cls) -> ModuleType:
+        """Import the DBAPI module that is used by this dialect.
+
+        The Python module object returned here will be assigned as an
+        instance variable to a constructed dialect under the name
+        ``.dbapi``.
+
+        .. versionchanged:: 2.0  The :meth:`.Dialect.import_dbapi` class
+           method is renamed from the previous method ``.Dialect.dbapi()``,
+           which would be replaced at dialect instantiation time by the
+           DBAPI module itself, thus using the same name in two different ways.
+           If a ``.Dialect.dbapi()`` classmethod is present on a third-party
+           dialect, it will be used and a deprecation warning will be emitted.
+
+        """
+        raise NotImplementedError()
+
+    def type_descriptor(self, typeobj: TypeEngine[_T]) -> TypeEngine[_T]:
+        """Transform a generic type to a dialect-specific type.
+
+        Dialect classes will usually use the
+        :func:`_types.adapt_type` function in the types module to
+        accomplish this.
+
+        The returned result is cached *per dialect class* so can
+        contain no dialect-instance state.
+
+        """
+
+        raise NotImplementedError()
+
+    def initialize(self, connection: Connection) -> None:
+        """Called during strategized creation of the dialect with a
+        connection.
+
+        Allows dialects to configure options based on server version info or
+        other properties.
+
+        The connection passed here is a SQLAlchemy Connection object,
+        with full capabilities.
+
+        The initialize() method of the base dialect should be called via
+        super().
+
+        .. note:: as of SQLAlchemy 1.4, this method is called **before**
+           any :meth:`_engine.Dialect.on_connect` hooks are called.
+
+        """
+
+        pass
+
+    if TYPE_CHECKING:
+
+        def _overrides_default(self, method_name: str) -> bool: ...
+
+    def get_columns(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[ReflectedColumn]:
+        """Return information about columns in ``table_name``.
+
+        Given a :class:`_engine.Connection`, a string
+        ``table_name``, and an optional string ``schema``, return column
+        information as a list of dictionaries
+        corresponding to the :class:`.ReflectedColumn` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_columns`.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_columns(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, List[ReflectedColumn]]]:
+        """Return information about columns in all tables in the
+        given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_multi_columns`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_pk_constraint(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> ReflectedPrimaryKeyConstraint:
+        """Return information about the primary key constraint on
+        table_name`.
+
+        Given a :class:`_engine.Connection`, a string
+        ``table_name``, and an optional string ``schema``, return primary
+        key information as a dictionary corresponding to the
+        :class:`.ReflectedPrimaryKeyConstraint` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_pk_constraint`.
+
+        """
+        raise NotImplementedError()
+
+    def get_multi_pk_constraint(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, ReflectedPrimaryKeyConstraint]]:
+        """Return information about primary key constraints in
+        all tables in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_multi_pk_constraint`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+        raise NotImplementedError()
+
+    def get_foreign_keys(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[ReflectedForeignKeyConstraint]:
+        """Return information about foreign_keys in ``table_name``.
+
+        Given a :class:`_engine.Connection`, a string
+        ``table_name``, and an optional string ``schema``, return foreign
+        key information as a list of dicts corresponding to the
+        :class:`.ReflectedForeignKeyConstraint` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_foreign_keys`.
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_foreign_keys(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, List[ReflectedForeignKeyConstraint]]]:
+        """Return information about foreign_keys in all tables
+        in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_multi_foreign_keys`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_table_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of table names for ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_table_names`.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_temp_table_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of temporary table names on the given connection,
+        if supported by the underlying backend.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_temp_table_names`.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_view_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of all non-materialized view names available in the
+        database.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_view_names`.
+
+        :param schema: schema name to query, if not the default schema.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_materialized_view_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of all materialized view names available in the
+        database.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_materialized_view_names`.
+
+        :param schema: schema name to query, if not the default schema.
+
+         .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_sequence_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of all sequence names available in the database.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_sequence_names`.
+
+        :param schema: schema name to query, if not the default schema.
+
+        .. versionadded:: 1.4
+        """
+
+        raise NotImplementedError()
+
+    def get_temp_view_names(
+        self, connection: Connection, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        """Return a list of temporary view names on the given connection,
+        if supported by the underlying backend.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_temp_view_names`.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_schema_names(self, connection: Connection, **kw: Any) -> List[str]:
+        """Return a list of all schema names available in the database.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_schema_names`.
+        """
+        raise NotImplementedError()
+
+    def get_view_definition(
+        self,
+        connection: Connection,
+        view_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> str:
+        """Return plain or materialized view definition.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_view_definition`.
+
+        Given a :class:`_engine.Connection`, a string
+        ``view_name``, and an optional string ``schema``, return the view
+        definition.
+        """
+
+        raise NotImplementedError()
+
+    def get_indexes(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[ReflectedIndex]:
+        """Return information about indexes in ``table_name``.
+
+        Given a :class:`_engine.Connection`, a string
+        ``table_name`` and an optional string ``schema``, return index
+        information as a list of dictionaries corresponding to the
+        :class:`.ReflectedIndex` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_indexes`.
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_indexes(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, List[ReflectedIndex]]]:
+        """Return information about indexes in in all tables
+        in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_multi_indexes`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_unique_constraints(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[ReflectedUniqueConstraint]:
+        r"""Return information about unique constraints in ``table_name``.
+
+        Given a string ``table_name`` and an optional string ``schema``, return
+        unique constraint information as a list of dicts corresponding
+        to the :class:`.ReflectedUniqueConstraint` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_unique_constraints`.
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_unique_constraints(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, List[ReflectedUniqueConstraint]]]:
+        """Return information about unique constraints in all tables
+        in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_multi_unique_constraints`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_check_constraints(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[ReflectedCheckConstraint]:
+        r"""Return information about check constraints in ``table_name``.
+
+        Given a string ``table_name`` and an optional string ``schema``, return
+        check constraint information as a list of dicts corresponding
+        to the :class:`.ReflectedCheckConstraint` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_check_constraints`.
+
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_check_constraints(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, List[ReflectedCheckConstraint]]]:
+        """Return information about check constraints in all tables
+        in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_multi_check_constraints`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def get_table_options(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> Dict[str, Any]:
+        """Return a dictionary of options specified when ``table_name``
+        was created.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_table_options`.
+        """
+        raise NotImplementedError()
+
+    def get_multi_table_options(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, Dict[str, Any]]]:
+        """Return a dictionary of options specified when the tables in the
+        given schema were created.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_multi_table_options`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+        raise NotImplementedError()
+
+    def get_table_comment(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> ReflectedTableComment:
+        r"""Return the "comment" for the table identified by ``table_name``.
+
+        Given a string ``table_name`` and an optional string ``schema``, return
+        table comment information as a dictionary corresponding to the
+        :class:`.ReflectedTableComment` dictionary.
+
+        This is an internal dialect method. Applications should use
+        :meth:`.Inspector.get_table_comment`.
+
+        :raise: ``NotImplementedError`` for dialects that don't support
+         comments.
+
+        .. versionadded:: 1.2
+
+        """
+
+        raise NotImplementedError()
+
+    def get_multi_table_comment(
+        self,
+        connection: Connection,
+        *,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        **kw: Any,
+    ) -> Iterable[Tuple[TableKey, ReflectedTableComment]]:
+        """Return information about the table comment in all tables
+        in the given ``schema``.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.get_multi_table_comment`.
+
+        .. note:: The :class:`_engine.DefaultDialect` provides a default
+          implementation that will call the single table method for
+          each object returned by :meth:`Dialect.get_table_names`,
+          :meth:`Dialect.get_view_names` or
+          :meth:`Dialect.get_materialized_view_names` depending on the
+          provided ``kind``. Dialects that want to support a faster
+          implementation should implement this method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def normalize_name(self, name: str) -> str:
+        """convert the given name to lowercase if it is detected as
+        case insensitive.
+
+        This method is only used if the dialect defines
+        requires_name_normalize=True.
+
+        """
+        raise NotImplementedError()
+
+    def denormalize_name(self, name: str) -> str:
+        """convert the given name to a case insensitive identifier
+        for the backend if it is an all-lowercase name.
+
+        This method is only used if the dialect defines
+        requires_name_normalize=True.
+
+        """
+        raise NotImplementedError()
+
+    def has_table(
+        self,
+        connection: Connection,
+        table_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> bool:
+        """For internal dialect use, check the existence of a particular table
+        or view in the database.
+
+        Given a :class:`_engine.Connection` object, a string table_name and
+        optional schema name, return True if the given table exists in the
+        database, False otherwise.
+
+        This method serves as the underlying implementation of the
+        public facing :meth:`.Inspector.has_table` method, and is also used
+        internally to implement the "checkfirst" behavior for methods like
+        :meth:`_schema.Table.create` and :meth:`_schema.MetaData.create_all`.
+
+        .. note:: This method is used internally by SQLAlchemy, and is
+           published so that third-party dialects may provide an
+           implementation. It is **not** the public API for checking for table
+           presence. Please use the :meth:`.Inspector.has_table` method.
+
+        .. versionchanged:: 2.0:: :meth:`_engine.Dialect.has_table` now
+           formally supports checking for additional table-like objects:
+
+           * any type of views (plain or materialized)
+           * temporary tables of any kind
+
+           Previously, these two checks were not formally specified and
+           different dialects would vary in their behavior.   The dialect
+           testing suite now includes tests for all of these object types,
+           and dialects to the degree that the backing database supports views
+           or temporary tables should seek to support locating these objects
+           for full compliance.
+
+        """
+
+        raise NotImplementedError()
+
+    def has_index(
+        self,
+        connection: Connection,
+        table_name: str,
+        index_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> bool:
+        """Check the existence of a particular index name in the database.
+
+        Given a :class:`_engine.Connection` object, a string
+        ``table_name`` and string index name, return ``True`` if an index of
+        the given name on the given table exists, ``False`` otherwise.
+
+        The :class:`.DefaultDialect` implements this in terms of the
+        :meth:`.Dialect.has_table` and :meth:`.Dialect.get_indexes` methods,
+        however dialects can implement a more performant version.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.has_index`.
+
+        .. versionadded:: 1.4
+
+        """
+
+        raise NotImplementedError()
+
+    def has_sequence(
+        self,
+        connection: Connection,
+        sequence_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> bool:
+        """Check the existence of a particular sequence in the database.
+
+        Given a :class:`_engine.Connection` object and a string
+        `sequence_name`, return ``True`` if the given sequence exists in
+        the database, ``False`` otherwise.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.has_sequence`.
+        """
+
+        raise NotImplementedError()
+
+    def has_schema(
+        self, connection: Connection, schema_name: str, **kw: Any
+    ) -> bool:
+        """Check the existence of a particular schema name in the database.
+
+        Given a :class:`_engine.Connection` object, a string
+        ``schema_name``, return ``True`` if a schema of the
+        given exists, ``False`` otherwise.
+
+        The :class:`.DefaultDialect` implements this by checking
+        the presence of ``schema_name`` among the schemas returned by
+        :meth:`.Dialect.get_schema_names`,
+        however dialects can implement a more performant version.
+
+        This is an internal dialect method. Applications should use
+        :meth:`_engine.Inspector.has_schema`.
+
+        .. versionadded:: 2.0
+
+        """
+
+        raise NotImplementedError()
+
+    def _get_server_version_info(self, connection: Connection) -> Any:
+        """Retrieve the server version info from the given connection.
+
+        This is used by the default implementation to populate the
+        "server_version_info" attribute and is called exactly
+        once upon first connect.
+
+        """
+
+        raise NotImplementedError()
+
+    def _get_default_schema_name(self, connection: Connection) -> str:
+        """Return the string name of the currently selected schema from
+        the given connection.
+
+        This is used by the default implementation to populate the
+        "default_schema_name" attribute and is called exactly
+        once upon first connect.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_begin(self, dbapi_connection: PoolProxiedConnection) -> None:
+        """Provide an implementation of ``connection.begin()``, given a
+        DB-API connection.
+
+        The DBAPI has no dedicated "begin" method and it is expected
+        that transactions are implicit.  This hook is provided for those
+        DBAPIs that might need additional help in this area.
+
+        :param dbapi_connection: a DBAPI connection, typically
+         proxied within a :class:`.ConnectionFairy`.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_rollback(self, dbapi_connection: PoolProxiedConnection) -> None:
+        """Provide an implementation of ``connection.rollback()``, given
+        a DB-API connection.
+
+        :param dbapi_connection: a DBAPI connection, typically
+         proxied within a :class:`.ConnectionFairy`.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_commit(self, dbapi_connection: PoolProxiedConnection) -> None:
+        """Provide an implementation of ``connection.commit()``, given a
+        DB-API connection.
+
+        :param dbapi_connection: a DBAPI connection, typically
+         proxied within a :class:`.ConnectionFairy`.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_terminate(self, dbapi_connection: DBAPIConnection) -> None:
+        """Provide an implementation of ``connection.close()`` that tries as
+        much as possible to not block, given a DBAPI
+        connection.
+
+        In the vast majority of cases this just calls .close(), however
+        for some asyncio dialects may call upon different API features.
+
+        This hook is called by the :class:`_pool.Pool`
+        when a connection is being recycled or has been invalidated.
+
+        .. versionadded:: 1.4.41
+
+        """
+
+        raise NotImplementedError()
+
+    def do_close(self, dbapi_connection: DBAPIConnection) -> None:
+        """Provide an implementation of ``connection.close()``, given a DBAPI
+        connection.
+
+        This hook is called by the :class:`_pool.Pool`
+        when a connection has been
+        detached from the pool, or is being returned beyond the normal
+        capacity of the pool.
+
+        """
+
+        raise NotImplementedError()
+
+    def _do_ping_w_event(self, dbapi_connection: DBAPIConnection) -> bool:
+        raise NotImplementedError()
+
+    def do_ping(self, dbapi_connection: DBAPIConnection) -> bool:
+        """ping the DBAPI connection and return True if the connection is
+        usable."""
+        raise NotImplementedError()
+
+    def do_set_input_sizes(
+        self,
+        cursor: DBAPICursor,
+        list_of_tuples: _GenericSetInputSizesType,
+        context: ExecutionContext,
+    ) -> Any:
+        """invoke the cursor.setinputsizes() method with appropriate arguments
+
+        This hook is called if the :attr:`.Dialect.bind_typing` attribute is
+        set to the
+        :attr:`.BindTyping.SETINPUTSIZES` value.
+        Parameter data is passed in a list of tuples (paramname, dbtype,
+        sqltype), where ``paramname`` is the key of the parameter in the
+        statement, ``dbtype`` is the DBAPI datatype and ``sqltype`` is the
+        SQLAlchemy type. The order of tuples is in the correct parameter order.
+
+        .. versionadded:: 1.4
+
+        .. versionchanged:: 2.0  - setinputsizes mode is now enabled by
+           setting :attr:`.Dialect.bind_typing` to
+           :attr:`.BindTyping.SETINPUTSIZES`.  Dialects which accept
+           a ``use_setinputsizes`` parameter should set this value
+           appropriately.
+
+
+        """
+        raise NotImplementedError()
+
+    def create_xid(self) -> Any:
+        """Create a two-phase transaction ID.
+
+        This id will be passed to do_begin_twophase(),
+        do_rollback_twophase(), do_commit_twophase().  Its format is
+        unspecified.
+        """
+
+        raise NotImplementedError()
+
+    def do_savepoint(self, connection: Connection, name: str) -> None:
+        """Create a savepoint with the given name.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param name: savepoint name.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_rollback_to_savepoint(
+        self, connection: Connection, name: str
+    ) -> None:
+        """Rollback a connection to the named savepoint.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param name: savepoint name.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_release_savepoint(self, connection: Connection, name: str) -> None:
+        """Release the named savepoint on a connection.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param name: savepoint name.
+        """
+
+        raise NotImplementedError()
+
+    def do_begin_twophase(self, connection: Connection, xid: Any) -> None:
+        """Begin a two phase transaction on the given connection.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param xid: xid
+
+        """
+
+        raise NotImplementedError()
+
+    def do_prepare_twophase(self, connection: Connection, xid: Any) -> None:
+        """Prepare a two phase transaction on the given connection.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param xid: xid
+
+        """
+
+        raise NotImplementedError()
+
+    def do_rollback_twophase(
+        self,
+        connection: Connection,
+        xid: Any,
+        is_prepared: bool = True,
+        recover: bool = False,
+    ) -> None:
+        """Rollback a two phase transaction on the given connection.
+
+        :param connection: a :class:`_engine.Connection`.
+        :param xid: xid
+        :param is_prepared: whether or not
+         :meth:`.TwoPhaseTransaction.prepare` was called.
+        :param recover: if the recover flag was passed.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_commit_twophase(
+        self,
+        connection: Connection,
+        xid: Any,
+        is_prepared: bool = True,
+        recover: bool = False,
+    ) -> None:
+        """Commit a two phase transaction on the given connection.
+
+
+        :param connection: a :class:`_engine.Connection`.
+        :param xid: xid
+        :param is_prepared: whether or not
+         :meth:`.TwoPhaseTransaction.prepare` was called.
+        :param recover: if the recover flag was passed.
+
+        """
+
+        raise NotImplementedError()
+
+    def do_recover_twophase(self, connection: Connection) -> List[Any]:
+        """Recover list of uncommitted prepared two phase transaction
+        identifiers on the given connection.
+
+        :param connection: a :class:`_engine.Connection`.
+
+        """
+
+        raise NotImplementedError()
+
+    def _deliver_insertmanyvalues_batches(
+        self,
+        connection: Connection,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+        generic_setinputsizes: Optional[_GenericSetInputSizesType],
+        context: ExecutionContext,
+    ) -> Iterator[_InsertManyValuesBatch]:
+        """convert executemany parameters for an INSERT into an iterator
+        of statement/single execute values, used by the insertmanyvalues
+        feature.
+
+        """
+        raise NotImplementedError()
+
+    def do_executemany(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+        context: Optional[ExecutionContext] = None,
+    ) -> None:
+        """Provide an implementation of ``cursor.executemany(statement,
+        parameters)``."""
+
+        raise NotImplementedError()
+
+    def do_execute(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        parameters: Optional[_DBAPISingleExecuteParams],
+        context: Optional[ExecutionContext] = None,
+    ) -> None:
+        """Provide an implementation of ``cursor.execute(statement,
+        parameters)``."""
+
+        raise NotImplementedError()
+
+    def do_execute_no_params(
+        self,
+        cursor: DBAPICursor,
+        statement: str,
+        context: Optional[ExecutionContext] = None,
+    ) -> None:
+        """Provide an implementation of ``cursor.execute(statement)``.
+
+        The parameter collection should not be sent.
+
+        """
+
+        raise NotImplementedError()
+
+    def is_disconnect(
+        self,
+        e: Exception,
+        connection: Optional[Union[PoolProxiedConnection, DBAPIConnection]],
+        cursor: Optional[DBAPICursor],
+    ) -> bool:
+        """Return True if the given DB-API error indicates an invalid
+        connection"""
+
+        raise NotImplementedError()
+
+    def connect(self, *cargs: Any, **cparams: Any) -> DBAPIConnection:
+        r"""Establish a connection using this dialect's DBAPI.
+
+        The default implementation of this method is::
+
+            def connect(self, *cargs, **cparams):
+                return self.dbapi.connect(*cargs, **cparams)
+
+        The ``*cargs, **cparams`` parameters are generated directly
+        from this dialect's :meth:`.Dialect.create_connect_args` method.
+
+        This method may be used for dialects that need to perform programmatic
+        per-connection steps when a new connection is procured from the
+        DBAPI.
+
+
+        :param \*cargs: positional parameters returned from the
+         :meth:`.Dialect.create_connect_args` method
+
+        :param \*\*cparams: keyword parameters returned from the
+         :meth:`.Dialect.create_connect_args` method.
+
+        :return: a DBAPI connection, typically from the :pep:`249` module
+         level ``.connect()`` function.
+
+        .. seealso::
+
+            :meth:`.Dialect.create_connect_args`
+
+            :meth:`.Dialect.on_connect`
+
+        """
+        raise NotImplementedError()
+
+    def on_connect_url(self, url: URL) -> Optional[Callable[[Any], Any]]:
+        """return a callable which sets up a newly created DBAPI connection.
+
+        This method is a new hook that supersedes the
+        :meth:`_engine.Dialect.on_connect` method when implemented by a
+        dialect.   When not implemented by a dialect, it invokes the
+        :meth:`_engine.Dialect.on_connect` method directly to maintain
+        compatibility with existing dialects.   There is no deprecation
+        for :meth:`_engine.Dialect.on_connect` expected.
+
+        The callable should accept a single argument "conn" which is the
+        DBAPI connection itself.  The inner callable has no
+        return value.
+
+        E.g.::
+
+            class MyDialect(default.DefaultDialect):
+                # ...
+
+                def on_connect_url(self, url):
+                    def do_on_connect(connection):
+                        connection.execute("SET SPECIAL FLAGS etc")
+
+                    return do_on_connect
+
+        This is used to set dialect-wide per-connection options such as
+        isolation modes, Unicode modes, etc.
+
+        This method differs from :meth:`_engine.Dialect.on_connect` in that
+        it is passed the :class:`_engine.URL` object that's relevant to the
+        connect args.  Normally the only way to get this is from the
+        :meth:`_engine.Dialect.on_connect` hook is to look on the
+        :class:`_engine.Engine` itself, however this URL object may have been
+        replaced by plugins.
+
+        .. note::
+
+            The default implementation of
+            :meth:`_engine.Dialect.on_connect_url` is to invoke the
+            :meth:`_engine.Dialect.on_connect` method. Therefore if a dialect
+            implements this method, the :meth:`_engine.Dialect.on_connect`
+            method **will not be called** unless the overriding dialect calls
+            it directly from here.
+
+        .. versionadded:: 1.4.3 added :meth:`_engine.Dialect.on_connect_url`
+           which normally calls into :meth:`_engine.Dialect.on_connect`.
+
+        :param url: a :class:`_engine.URL` object representing the
+         :class:`_engine.URL` that was passed to the
+         :meth:`_engine.Dialect.create_connect_args` method.
+
+        :return: a callable that accepts a single DBAPI connection as an
+         argument, or None.
+
+        .. seealso::
+
+            :meth:`_engine.Dialect.on_connect`
+
+        """
+        return self.on_connect()
+
+    def on_connect(self) -> Optional[Callable[[Any], Any]]:
+        """return a callable which sets up a newly created DBAPI connection.
+
+        The callable should accept a single argument "conn" which is the
+        DBAPI connection itself.  The inner callable has no
+        return value.
+
+        E.g.::
+
+            class MyDialect(default.DefaultDialect):
+                # ...
+
+                def on_connect(self):
+                    def do_on_connect(connection):
+                        connection.execute("SET SPECIAL FLAGS etc")
+
+                    return do_on_connect
+
+        This is used to set dialect-wide per-connection options such as
+        isolation modes, Unicode modes, etc.
+
+        The "do_on_connect" callable is invoked by using the
+        :meth:`_events.PoolEvents.connect` event
+        hook, then unwrapping the DBAPI connection and passing it into the
+        callable.
+
+        .. versionchanged:: 1.4 the on_connect hook is no longer called twice
+           for the first connection of a dialect.  The on_connect hook is still
+           called before the :meth:`_engine.Dialect.initialize` method however.
+
+        .. versionchanged:: 1.4.3 the on_connect hook is invoked from a new
+           method on_connect_url that passes the URL that was used to create
+           the connect args.   Dialects can implement on_connect_url instead
+           of on_connect if they need the URL object that was used for the
+           connection in order to get additional context.
+
+        If None is returned, no event listener is generated.
+
+        :return: a callable that accepts a single DBAPI connection as an
+         argument, or None.
+
+        .. seealso::
+
+            :meth:`.Dialect.connect` - allows the DBAPI ``connect()`` sequence
+            itself to be controlled.
+
+            :meth:`.Dialect.on_connect_url` - supersedes
+            :meth:`.Dialect.on_connect` to also receive the
+            :class:`_engine.URL` object in context.
+
+        """
+        return None
+
+    def reset_isolation_level(self, dbapi_connection: DBAPIConnection) -> None:
+        """Given a DBAPI connection, revert its isolation to the default.
+
+        Note that this is a dialect-level method which is used as part
+        of the implementation of the :class:`_engine.Connection` and
+        :class:`_engine.Engine`
+        isolation level facilities; these APIs should be preferred for
+        most typical use cases.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.get_isolation_level`
+            - view current level
+
+            :attr:`_engine.Connection.default_isolation_level`
+            - view default level
+
+            :paramref:`.Connection.execution_options.isolation_level` -
+            set per :class:`_engine.Connection` isolation level
+
+            :paramref:`_sa.create_engine.isolation_level` -
+            set per :class:`_engine.Engine` isolation level
+
+        """
+
+        raise NotImplementedError()
+
+    def set_isolation_level(
+        self, dbapi_connection: DBAPIConnection, level: IsolationLevel
+    ) -> None:
+        """Given a DBAPI connection, set its isolation level.
+
+        Note that this is a dialect-level method which is used as part
+        of the implementation of the :class:`_engine.Connection` and
+        :class:`_engine.Engine`
+        isolation level facilities; these APIs should be preferred for
+        most typical use cases.
+
+        If the dialect also implements the
+        :meth:`.Dialect.get_isolation_level_values` method, then the given
+        level is guaranteed to be one of the string names within that sequence,
+        and the method will not need to anticipate a lookup failure.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.get_isolation_level`
+            - view current level
+
+            :attr:`_engine.Connection.default_isolation_level`
+            - view default level
+
+            :paramref:`.Connection.execution_options.isolation_level` -
+            set per :class:`_engine.Connection` isolation level
+
+            :paramref:`_sa.create_engine.isolation_level` -
+            set per :class:`_engine.Engine` isolation level
+
+        """
+
+        raise NotImplementedError()
+
+    def get_isolation_level(
+        self, dbapi_connection: DBAPIConnection
+    ) -> IsolationLevel:
+        """Given a DBAPI connection, return its isolation level.
+
+        When working with a :class:`_engine.Connection` object,
+        the corresponding
+        DBAPI connection may be procured using the
+        :attr:`_engine.Connection.connection` accessor.
+
+        Note that this is a dialect-level method which is used as part
+        of the implementation of the :class:`_engine.Connection` and
+        :class:`_engine.Engine` isolation level facilities;
+        these APIs should be preferred for most typical use cases.
+
+
+        .. seealso::
+
+            :meth:`_engine.Connection.get_isolation_level`
+            - view current level
+
+            :attr:`_engine.Connection.default_isolation_level`
+            - view default level
+
+            :paramref:`.Connection.execution_options.isolation_level` -
+            set per :class:`_engine.Connection` isolation level
+
+            :paramref:`_sa.create_engine.isolation_level` -
+            set per :class:`_engine.Engine` isolation level
+
+
+        """
+
+        raise NotImplementedError()
+
+    def get_default_isolation_level(
+        self, dbapi_conn: DBAPIConnection
+    ) -> IsolationLevel:
+        """Given a DBAPI connection, return its isolation level, or
+        a default isolation level if one cannot be retrieved.
+
+        This method may only raise NotImplementedError and
+        **must not raise any other exception**, as it is used implicitly upon
+        first connect.
+
+        The method **must return a value** for a dialect that supports
+        isolation level settings, as this level is what will be reverted
+        towards when a per-connection isolation level change is made.
+
+        The method defaults to using the :meth:`.Dialect.get_isolation_level`
+        method unless overridden by a dialect.
+
+        .. versionadded:: 1.3.22
+
+        """
+        raise NotImplementedError()
+
+    def get_isolation_level_values(
+        self, dbapi_conn: DBAPIConnection
+    ) -> List[IsolationLevel]:
+        """return a sequence of string isolation level names that are accepted
+        by this dialect.
+
+        The available names should use the following conventions:
+
+        * use UPPERCASE names.   isolation level methods will accept lowercase
+          names but these are normalized into UPPERCASE before being passed
+          along to the dialect.
+        * separate words should be separated by spaces, not underscores, e.g.
+          ``REPEATABLE READ``.  isolation level names will have underscores
+          converted to spaces before being passed along to the dialect.
+        * The names for the four standard isolation names to the extent that
+          they are supported by the backend should be ``READ UNCOMMITTED``,
+          ``READ COMMITTED``, ``REPEATABLE READ``, ``SERIALIZABLE``
+        * if the dialect supports an autocommit option it should be provided
+          using the isolation level name ``AUTOCOMMIT``.
+        * Other isolation modes may also be present, provided that they
+          are named in UPPERCASE and use spaces not underscores.
+
+        This function is used so that the default dialect can check that
+        a given isolation level parameter is valid, else raises an
+        :class:`_exc.ArgumentError`.
+
+        A DBAPI connection is passed to the method, in the unlikely event that
+        the dialect needs to interrogate the connection itself to determine
+        this list, however it is expected that most backends will return
+        a hardcoded list of values.  If the dialect supports "AUTOCOMMIT",
+        that value should also be present in the sequence returned.
+
+        The method raises ``NotImplementedError`` by default.  If a dialect
+        does not implement this method, then the default dialect will not
+        perform any checking on a given isolation level value before passing
+        it onto the :meth:`.Dialect.set_isolation_level` method.  This is
+        to allow backwards-compatibility with third party dialects that may
+        not yet be implementing this method.
+
+        .. versionadded:: 2.0
+
+        """
+        raise NotImplementedError()
+
+    def _assert_and_set_isolation_level(
+        self, dbapi_conn: DBAPIConnection, level: IsolationLevel
+    ) -> None:
+        raise NotImplementedError()
+
+    @classmethod
+    def get_dialect_cls(cls, url: URL) -> Type[Dialect]:
+        """Given a URL, return the :class:`.Dialect` that will be used.
+
+        This is a hook that allows an external plugin to provide functionality
+        around an existing dialect, by allowing the plugin to be loaded
+        from the url based on an entrypoint, and then the plugin returns
+        the actual dialect to be used.
+
+        By default this just returns the cls.
+
+        """
+        return cls
+
+    @classmethod
+    def get_async_dialect_cls(cls, url: URL) -> Type[Dialect]:
+        """Given a URL, return the :class:`.Dialect` that will be used by
+        an async engine.
+
+        By default this is an alias of :meth:`.Dialect.get_dialect_cls` and
+        just returns the cls. It may be used if a dialect provides
+        both a sync and async version under the same name, like the
+        ``psycopg`` driver.
+
+        .. versionadded:: 2
+
+        .. seealso::
+
+            :meth:`.Dialect.get_dialect_cls`
+
+        """
+        return cls.get_dialect_cls(url)
+
+    @classmethod
+    def load_provisioning(cls) -> None:
+        """set up the provision.py module for this dialect.
+
+        For dialects that include a provision.py module that sets up
+        provisioning followers, this method should initiate that process.
+
+        A typical implementation would be::
+
+            @classmethod
+            def load_provisioning(cls):
+                __import__("mydialect.provision")
+
+        The default method assumes a module named ``provision.py`` inside
+        the owning package of the current dialect, based on the ``__module__``
+        attribute::
+
+            @classmethod
+            def load_provisioning(cls):
+                package = ".".join(cls.__module__.split(".")[0:-1])
+                try:
+                    __import__(package + ".provision")
+                except ImportError:
+                    pass
+
+        .. versionadded:: 1.3.14
+
+        """
+
+    @classmethod
+    def engine_created(cls, engine: Engine) -> None:
+        """A convenience hook called before returning the final
+        :class:`_engine.Engine`.
+
+        If the dialect returned a different class from the
+        :meth:`.get_dialect_cls`
+        method, then the hook is called on both classes, first on
+        the dialect class returned by the :meth:`.get_dialect_cls` method and
+        then on the class on which the method was called.
+
+        The hook should be used by dialects and/or wrappers to apply special
+        events to the engine or its components.   In particular, it allows
+        a dialect-wrapping class to apply dialect-level events.
+
+        """
+
+    def get_driver_connection(self, connection: DBAPIConnection) -> Any:
+        """Returns the connection object as returned by the external driver
+        package.
+
+        For normal dialects that use a DBAPI compliant driver this call
+        will just return the ``connection`` passed as argument.
+        For dialects that instead adapt a non DBAPI compliant driver, like
+        when adapting an asyncio driver, this call will return the
+        connection-like object as returned by the driver.
+
+        .. versionadded:: 1.4.24
+
+        """
+        raise NotImplementedError()
+
+    def set_engine_execution_options(
+        self, engine: Engine, opts: CoreExecuteOptionsParameter
+    ) -> None:
+        """Establish execution options for a given engine.
+
+        This is implemented by :class:`.DefaultDialect` to establish
+        event hooks for new :class:`.Connection` instances created
+        by the given :class:`.Engine` which will then invoke the
+        :meth:`.Dialect.set_connection_execution_options` method for that
+        connection.
+
+        """
+        raise NotImplementedError()
+
+    def set_connection_execution_options(
+        self, connection: Connection, opts: CoreExecuteOptionsParameter
+    ) -> None:
+        """Establish execution options for a given connection.
+
+        This is implemented by :class:`.DefaultDialect` in order to implement
+        the :paramref:`_engine.Connection.execution_options.isolation_level`
+        execution option.  Dialects can intercept various execution options
+        which may need to modify state on a particular DBAPI connection.
+
+        .. versionadded:: 1.4
+
+        """
+        raise NotImplementedError()
+
+    def get_dialect_pool_class(self, url: URL) -> Type[Pool]:
+        """return a Pool class to use for a given URL"""
+        raise NotImplementedError()
+
+
+class CreateEnginePlugin:
+    """A set of hooks intended to augment the construction of an
+    :class:`_engine.Engine` object based on entrypoint names in a URL.
+
+    The purpose of :class:`_engine.CreateEnginePlugin` is to allow third-party
+    systems to apply engine, pool and dialect level event listeners without
+    the need for the target application to be modified; instead, the plugin
+    names can be added to the database URL.  Target applications for
+    :class:`_engine.CreateEnginePlugin` include:
+
+    * connection and SQL performance tools, e.g. which use events to track
+      number of checkouts and/or time spent with statements
+
+    * connectivity plugins such as proxies
+
+    A rudimentary :class:`_engine.CreateEnginePlugin` that attaches a logger
+    to an :class:`_engine.Engine` object might look like::
+
+
+        import logging
+
+        from sqlalchemy.engine import CreateEnginePlugin
+        from sqlalchemy import event
+
+
+        class LogCursorEventsPlugin(CreateEnginePlugin):
+            def __init__(self, url, kwargs):
+                # consume the parameter "log_cursor_logging_name" from the
+                # URL query
+                logging_name = url.query.get(
+                    "log_cursor_logging_name", "log_cursor"
+                )
+
+                self.log = logging.getLogger(logging_name)
+
+            def update_url(self, url):
+                "update the URL to one that no longer includes our parameters"
+                return url.difference_update_query(["log_cursor_logging_name"])
+
+            def engine_created(self, engine):
+                "attach an event listener after the new Engine is constructed"
+                event.listen(engine, "before_cursor_execute", self._log_event)
+
+            def _log_event(
+                self,
+                conn,
+                cursor,
+                statement,
+                parameters,
+                context,
+                executemany,
+            ):
+
+                self.log.info("Plugin logged cursor event: %s", statement)
+
+    Plugins are registered using entry points in a similar way as that
+    of dialects::
+
+        entry_points = {
+            "sqlalchemy.plugins": [
+                "log_cursor_plugin = myapp.plugins:LogCursorEventsPlugin"
+            ]
+        }
+
+    A plugin that uses the above names would be invoked from a database
+    URL as in::
+
+        from sqlalchemy import create_engine
+
+        engine = create_engine(
+            "mysql+pymysql://scott:tiger@localhost/test?"
+            "plugin=log_cursor_plugin&log_cursor_logging_name=mylogger"
+        )
+
+    The ``plugin`` URL parameter supports multiple instances, so that a URL
+    may specify multiple plugins; they are loaded in the order stated
+    in the URL::
+
+        engine = create_engine(
+            "mysql+pymysql://scott:tiger@localhost/test?"
+            "plugin=plugin_one&plugin=plugin_twp&plugin=plugin_three"
+        )
+
+    The plugin names may also be passed directly to :func:`_sa.create_engine`
+    using the :paramref:`_sa.create_engine.plugins` argument::
+
+        engine = create_engine(
+            "mysql+pymysql://scott:tiger@localhost/test", plugins=["myplugin"]
+        )
+
+    .. versionadded:: 1.2.3  plugin names can also be specified
+       to :func:`_sa.create_engine` as a list
+
+    A plugin may consume plugin-specific arguments from the
+    :class:`_engine.URL` object as well as the ``kwargs`` dictionary, which is
+    the dictionary of arguments passed to the :func:`_sa.create_engine`
+    call.  "Consuming" these arguments includes that they must be removed
+    when the plugin initializes, so that the arguments are not passed along
+    to the :class:`_engine.Dialect` constructor, where they will raise an
+    :class:`_exc.ArgumentError` because they are not known by the dialect.
+
+    As of version 1.4 of SQLAlchemy, arguments should continue to be consumed
+    from the ``kwargs`` dictionary directly, by removing the values with a
+    method such as ``dict.pop``. Arguments from the :class:`_engine.URL` object
+    should be consumed by implementing the
+    :meth:`_engine.CreateEnginePlugin.update_url` method, returning a new copy
+    of the :class:`_engine.URL` with plugin-specific parameters removed::
+
+        class MyPlugin(CreateEnginePlugin):
+            def __init__(self, url, kwargs):
+                self.my_argument_one = url.query["my_argument_one"]
+                self.my_argument_two = url.query["my_argument_two"]
+                self.my_argument_three = kwargs.pop("my_argument_three", None)
+
+            def update_url(self, url):
+                return url.difference_update_query(
+                    ["my_argument_one", "my_argument_two"]
+                )
+
+    Arguments like those illustrated above would be consumed from a
+    :func:`_sa.create_engine` call such as::
+
+        from sqlalchemy import create_engine
+
+        engine = create_engine(
+            "mysql+pymysql://scott:tiger@localhost/test?"
+            "plugin=myplugin&my_argument_one=foo&my_argument_two=bar",
+            my_argument_three="bat",
+        )
+
+    .. versionchanged:: 1.4
+
+        The :class:`_engine.URL` object is now immutable; a
+        :class:`_engine.CreateEnginePlugin` that needs to alter the
+        :class:`_engine.URL` should implement the newly added
+        :meth:`_engine.CreateEnginePlugin.update_url` method, which
+        is invoked after the plugin is constructed.
+
+        For migration, construct the plugin in the following way, checking
+        for the existence of the :meth:`_engine.CreateEnginePlugin.update_url`
+        method to detect which version is running::
+
+            class MyPlugin(CreateEnginePlugin):
+                def __init__(self, url, kwargs):
+                    if hasattr(CreateEnginePlugin, "update_url"):
+                        # detect the 1.4 API
+                        self.my_argument_one = url.query["my_argument_one"]
+                        self.my_argument_two = url.query["my_argument_two"]
+                    else:
+                        # detect the 1.3 and earlier API - mutate the
+                        # URL directly
+                        self.my_argument_one = url.query.pop("my_argument_one")
+                        self.my_argument_two = url.query.pop("my_argument_two")
+
+                    self.my_argument_three = kwargs.pop("my_argument_three", None)
+
+                def update_url(self, url):
+                    # this method is only called in the 1.4 version
+                    return url.difference_update_query(
+                        ["my_argument_one", "my_argument_two"]
+                    )
+
+        .. seealso::
+
+            :ref:`change_5526` - overview of the :class:`_engine.URL` change which
+            also includes notes regarding :class:`_engine.CreateEnginePlugin`.
+
+
+    When the engine creation process completes and produces the
+    :class:`_engine.Engine` object, it is again passed to the plugin via the
+    :meth:`_engine.CreateEnginePlugin.engine_created` hook.  In this hook, additional
+    changes can be made to the engine, most typically involving setup of
+    events (e.g. those defined in :ref:`core_event_toplevel`).
+
+    """  # noqa: E501
+
+    def __init__(self, url: URL, kwargs: Dict[str, Any]):
+        """Construct a new :class:`.CreateEnginePlugin`.
+
+        The plugin object is instantiated individually for each call
+        to :func:`_sa.create_engine`.  A single :class:`_engine.
+        Engine` will be
+        passed to the :meth:`.CreateEnginePlugin.engine_created` method
+        corresponding to this URL.
+
+        :param url: the :class:`_engine.URL` object.  The plugin may inspect
+         the :class:`_engine.URL` for arguments.  Arguments used by the
+         plugin should be removed, by returning an updated :class:`_engine.URL`
+         from the :meth:`_engine.CreateEnginePlugin.update_url` method.
+
+         .. versionchanged::  1.4
+
+            The :class:`_engine.URL` object is now immutable, so a
+            :class:`_engine.CreateEnginePlugin` that needs to alter the
+            :class:`_engine.URL` object should implement the
+            :meth:`_engine.CreateEnginePlugin.update_url` method.
+
+        :param kwargs: The keyword arguments passed to
+         :func:`_sa.create_engine`.
+
+        """
+        self.url = url
+
+    def update_url(self, url: URL) -> URL:
+        """Update the :class:`_engine.URL`.
+
+        A new :class:`_engine.URL` should be returned.   This method is
+        typically used to consume configuration arguments from the
+        :class:`_engine.URL` which must be removed, as they will not be
+        recognized by the dialect.  The
+        :meth:`_engine.URL.difference_update_query` method is available
+        to remove these arguments.   See the docstring at
+        :class:`_engine.CreateEnginePlugin` for an example.
+
+
+        .. versionadded:: 1.4
+
+        """
+        raise NotImplementedError()
+
+    def handle_dialect_kwargs(
+        self, dialect_cls: Type[Dialect], dialect_args: Dict[str, Any]
+    ) -> None:
+        """parse and modify dialect kwargs"""
+
+    def handle_pool_kwargs(
+        self, pool_cls: Type[Pool], pool_args: Dict[str, Any]
+    ) -> None:
+        """parse and modify pool kwargs"""
+
+    def engine_created(self, engine: Engine) -> None:
+        """Receive the :class:`_engine.Engine`
+        object when it is fully constructed.
+
+        The plugin may make additional changes to the engine, such as
+        registering engine or connection pool events.
+
+        """
+
+
+class ExecutionContext:
+    """A messenger object for a Dialect that corresponds to a single
+    execution.
+
+    """
+
+    engine: Engine
+    """engine which the Connection is associated with"""
+
+    connection: Connection
+    """Connection object which can be freely used by default value
+      generators to execute SQL.  This Connection should reference the
+      same underlying connection/transactional resources of
+      root_connection."""
+
+    root_connection: Connection
+    """Connection object which is the source of this ExecutionContext."""
+
+    dialect: Dialect
+    """dialect which created this ExecutionContext."""
+
+    cursor: DBAPICursor
+    """DB-API cursor procured from the connection"""
+
+    compiled: Optional[Compiled]
+    """if passed to constructor, sqlalchemy.engine.base.Compiled object
+      being executed"""
+
+    statement: str
+    """string version of the statement to be executed.  Is either
+      passed to the constructor, or must be created from the
+      sql.Compiled object by the time pre_exec() has completed."""
+
+    invoked_statement: Optional[Executable]
+    """The Executable statement object that was given in the first place.
+
+    This should be structurally equivalent to compiled.statement, but not
+    necessarily the same object as in a caching scenario the compiled form
+    will have been extracted from the cache.
+
+    """
+
+    parameters: _AnyMultiExecuteParams
+    """bind parameters passed to the execute() or exec_driver_sql() methods.
+
+    These are always stored as a list of parameter entries.  A single-element
+    list corresponds to a ``cursor.execute()`` call and a multiple-element
+    list corresponds to ``cursor.executemany()``, except in the case
+    of :attr:`.ExecuteStyle.INSERTMANYVALUES` which will use
+    ``cursor.execute()`` one or more times.
+
+    """
+
+    no_parameters: bool
+    """True if the execution style does not use parameters"""
+
+    isinsert: bool
+    """True if the statement is an INSERT."""
+
+    isupdate: bool
+    """True if the statement is an UPDATE."""
+
+    execute_style: ExecuteStyle
+    """the style of DBAPI cursor method that will be used to execute
+    a statement.
+
+    .. versionadded:: 2.0
+
+    """
+
+    executemany: bool
+    """True if the context has a list of more than one parameter set.
+
+    Historically this attribute links to whether ``cursor.execute()`` or
+    ``cursor.executemany()`` will be used.  It also can now mean that
+    "insertmanyvalues" may be used which indicates one or more
+    ``cursor.execute()`` calls.
+
+    """
+
+    prefetch_cols: util.generic_fn_descriptor[Optional[Sequence[Column[Any]]]]
+    """a list of Column objects for which a client-side default
+      was fired off.  Applies to inserts and updates."""
+
+    postfetch_cols: util.generic_fn_descriptor[Optional[Sequence[Column[Any]]]]
+    """a list of Column objects for which a server-side default or
+      inline SQL expression value was fired off.  Applies to inserts
+      and updates."""
+
+    execution_options: _ExecuteOptions
+    """Execution options associated with the current statement execution"""
+
+    @classmethod
+    def _init_ddl(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        compiled_ddl: DDLCompiler,
+    ) -> ExecutionContext:
+        raise NotImplementedError()
+
+    @classmethod
+    def _init_compiled(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        compiled: SQLCompiler,
+        parameters: _CoreMultiExecuteParams,
+        invoked_statement: Executable,
+        extracted_parameters: Optional[Sequence[BindParameter[Any]]],
+        cache_hit: CacheStats = CacheStats.CACHING_DISABLED,
+    ) -> ExecutionContext:
+        raise NotImplementedError()
+
+    @classmethod
+    def _init_statement(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+    ) -> ExecutionContext:
+        raise NotImplementedError()
+
+    @classmethod
+    def _init_default(
+        cls,
+        dialect: Dialect,
+        connection: Connection,
+        dbapi_connection: PoolProxiedConnection,
+        execution_options: _ExecuteOptions,
+    ) -> ExecutionContext:
+        raise NotImplementedError()
+
+    def _exec_default(
+        self,
+        column: Optional[Column[Any]],
+        default: DefaultGenerator,
+        type_: Optional[TypeEngine[Any]],
+    ) -> Any:
+        raise NotImplementedError()
+
+    def _prepare_set_input_sizes(
+        self,
+    ) -> Optional[List[Tuple[str, Any, TypeEngine[Any]]]]:
+        raise NotImplementedError()
+
+    def _get_cache_stats(self) -> str:
+        raise NotImplementedError()
+
+    def _setup_result_proxy(self) -> CursorResult[Any]:
+        raise NotImplementedError()
+
+    def fire_sequence(self, seq: Sequence_SchemaItem, type_: Integer) -> int:
+        """given a :class:`.Sequence`, invoke it and return the next int
+        value"""
+        raise NotImplementedError()
+
+    def create_cursor(self) -> DBAPICursor:
+        """Return a new cursor generated from this ExecutionContext's
+        connection.
+
+        Some dialects may wish to change the behavior of
+        connection.cursor(), such as postgresql which may return a PG
+        "server side" cursor.
+        """
+
+        raise NotImplementedError()
+
+    def pre_exec(self) -> None:
+        """Called before an execution of a compiled statement.
+
+        If a compiled statement was passed to this ExecutionContext,
+        the `statement` and `parameters` datamembers must be
+        initialized after this statement is complete.
+        """
+
+        raise NotImplementedError()
+
+    def get_out_parameter_values(
+        self, out_param_names: Sequence[str]
+    ) -> Sequence[Any]:
+        """Return a sequence of OUT parameter values from a cursor.
+
+        For dialects that support OUT parameters, this method will be called
+        when there is a :class:`.SQLCompiler` object which has the
+        :attr:`.SQLCompiler.has_out_parameters` flag set.  This flag in turn
+        will be set to True if the statement itself has :class:`.BindParameter`
+        objects that have the ``.isoutparam`` flag set which are consumed by
+        the :meth:`.SQLCompiler.visit_bindparam` method.  If the dialect
+        compiler produces :class:`.BindParameter` objects with ``.isoutparam``
+        set which are not handled by :meth:`.SQLCompiler.visit_bindparam`, it
+        should set this flag explicitly.
+
+        The list of names that were rendered for each bound parameter
+        is passed to the method.  The method should then return a sequence of
+        values corresponding to the list of parameter objects. Unlike in
+        previous SQLAlchemy versions, the values can be the **raw values** from
+        the DBAPI; the execution context will apply the appropriate type
+        handler based on what's present in self.compiled.binds and update the
+        values.  The processed dictionary will then be made available via the
+        ``.out_parameters`` collection on the result object.  Note that
+        SQLAlchemy 1.4 has multiple kinds of result object as part of the 2.0
+        transition.
+
+        .. versionadded:: 1.4 - added
+           :meth:`.ExecutionContext.get_out_parameter_values`, which is invoked
+           automatically by the :class:`.DefaultExecutionContext` when there
+           are :class:`.BindParameter` objects with the ``.isoutparam`` flag
+           set.  This replaces the practice of setting out parameters within
+           the now-removed ``get_result_proxy()`` method.
+
+        """
+        raise NotImplementedError()
+
+    def post_exec(self) -> None:
+        """Called after the execution of a compiled statement.
+
+        If a compiled statement was passed to this ExecutionContext,
+        the `last_insert_ids`, `last_inserted_params`, etc.
+        datamembers should be available after this method completes.
+        """
+
+        raise NotImplementedError()
+
+    def handle_dbapi_exception(self, e: BaseException) -> None:
+        """Receive a DBAPI exception which occurred upon execute, result
+        fetch, etc."""
+
+        raise NotImplementedError()
+
+    def lastrow_has_defaults(self) -> bool:
+        """Return True if the last INSERT or UPDATE row contained
+        inlined or database-side defaults.
+        """
+
+        raise NotImplementedError()
+
+    def get_rowcount(self) -> Optional[int]:
+        """Return the DBAPI ``cursor.rowcount`` value, or in some
+        cases an interpreted value.
+
+        See :attr:`_engine.CursorResult.rowcount` for details on this.
+
+        """
+
+        raise NotImplementedError()
+
+    def fetchall_for_returning(self, cursor: DBAPICursor) -> Sequence[Any]:
+        """For a RETURNING result, deliver cursor.fetchall() from the
+        DBAPI cursor.
+
+        This is a dialect-specific hook for dialects that have special
+        considerations when calling upon the rows delivered for a
+        "RETURNING" statement.   Default implementation is
+        ``cursor.fetchall()``.
+
+        This hook is currently used only by the :term:`insertmanyvalues`
+        feature.   Dialects that don't set ``use_insertmanyvalues=True``
+        don't need to consider this hook.
+
+        .. versionadded:: 2.0.10
+
+        """
+        raise NotImplementedError()
+
+
+class ConnectionEventsTarget(EventTarget):
+    """An object which can accept events from :class:`.ConnectionEvents`.
+
+    Includes :class:`_engine.Connection` and :class:`_engine.Engine`.
+
+    .. versionadded:: 2.0
+
+    """
+
+    dispatch: dispatcher[ConnectionEventsTarget]
+
+
+Connectable = ConnectionEventsTarget
+
+
+class ExceptionContext:
+    """Encapsulate information about an error condition in progress.
+
+    This object exists solely to be passed to the
+    :meth:`_events.DialectEvents.handle_error` event,
+    supporting an interface that
+    can be extended without backwards-incompatibility.
+
+
+    """
+
+    __slots__ = ()
+
+    dialect: Dialect
+    """The :class:`_engine.Dialect` in use.
+
+    This member is present for all invocations of the event hook.
+
+    .. versionadded:: 2.0
+
+    """
+
+    connection: Optional[Connection]
+    """The :class:`_engine.Connection` in use during the exception.
+
+    This member is present, except in the case of a failure when
+    first connecting.
+
+    .. seealso::
+
+        :attr:`.ExceptionContext.engine`
+
+
+    """
+
+    engine: Optional[Engine]
+    """The :class:`_engine.Engine` in use during the exception.
+
+    This member is present in all cases except for when handling an error
+    within the connection pool "pre-ping" process.
+
+    """
+
+    cursor: Optional[DBAPICursor]
+    """The DBAPI cursor object.
+
+    May be None.
+
+    """
+
+    statement: Optional[str]
+    """String SQL statement that was emitted directly to the DBAPI.
+
+    May be None.
+
+    """
+
+    parameters: Optional[_DBAPIAnyExecuteParams]
+    """Parameter collection that was emitted directly to the DBAPI.
+
+    May be None.
+
+    """
+
+    original_exception: BaseException
+    """The exception object which was caught.
+
+    This member is always present.
+
+    """
+
+    sqlalchemy_exception: Optional[StatementError]
+    """The :class:`sqlalchemy.exc.StatementError` which wraps the original,
+    and will be raised if exception handling is not circumvented by the event.
+
+    May be None, as not all exception types are wrapped by SQLAlchemy.
+    For DBAPI-level exceptions that subclass the dbapi's Error class, this
+    field will always be present.
+
+    """
+
+    chained_exception: Optional[BaseException]
+    """The exception that was returned by the previous handler in the
+    exception chain, if any.
+
+    If present, this exception will be the one ultimately raised by
+    SQLAlchemy unless a subsequent handler replaces it.
+
+    May be None.
+
+    """
+
+    execution_context: Optional[ExecutionContext]
+    """The :class:`.ExecutionContext` corresponding to the execution
+    operation in progress.
+
+    This is present for statement execution operations, but not for
+    operations such as transaction begin/end.  It also is not present when
+    the exception was raised before the :class:`.ExecutionContext`
+    could be constructed.
+
+    Note that the :attr:`.ExceptionContext.statement` and
+    :attr:`.ExceptionContext.parameters` members may represent a
+    different value than that of the :class:`.ExecutionContext`,
+    potentially in the case where a
+    :meth:`_events.ConnectionEvents.before_cursor_execute` event or similar
+    modified the statement/parameters to be sent.
+
+    May be None.
+
+    """
+
+    is_disconnect: bool
+    """Represent whether the exception as occurred represents a "disconnect"
+    condition.
+
+    This flag will always be True or False within the scope of the
+    :meth:`_events.DialectEvents.handle_error` handler.
+
+    SQLAlchemy will defer to this flag in order to determine whether or not
+    the connection should be invalidated subsequently.    That is, by
+    assigning to this flag, a "disconnect" event which then results in
+    a connection and pool invalidation can be invoked or prevented by
+    changing this flag.
+
+
+    .. note:: The pool "pre_ping" handler enabled using the
+        :paramref:`_sa.create_engine.pool_pre_ping` parameter does **not**
+        consult this event before deciding if the "ping" returned false,
+        as opposed to receiving an unhandled error.   For this use case, the
+        :ref:`legacy recipe based on engine_connect() may be used
+        <pool_disconnects_pessimistic_custom>`.  A future API allow more
+        comprehensive customization of the "disconnect" detection mechanism
+        across all functions.
+
+    """
+
+    invalidate_pool_on_disconnect: bool
+    """Represent whether all connections in the pool should be invalidated
+    when a "disconnect" condition is in effect.
+
+    Setting this flag to False within the scope of the
+    :meth:`_events.DialectEvents.handle_error`
+    event will have the effect such
+    that the full collection of connections in the pool will not be
+    invalidated during a disconnect; only the current connection that is the
+    subject of the error will actually be invalidated.
+
+    The purpose of this flag is for custom disconnect-handling schemes where
+    the invalidation of other connections in the pool is to be performed
+    based on other conditions, or even on a per-connection basis.
+
+    """
+
+    is_pre_ping: bool
+    """Indicates if this error is occurring within the "pre-ping" step
+    performed when :paramref:`_sa.create_engine.pool_pre_ping` is set to
+    ``True``.  In this mode, the :attr:`.ExceptionContext.engine` attribute
+    will be ``None``.  The dialect in use is accessible via the
+    :attr:`.ExceptionContext.dialect` attribute.
+
+    .. versionadded:: 2.0.5
+
+    """
+
+
+class AdaptedConnection:
+    """Interface of an adapted connection object to support the DBAPI protocol.
+
+    Used by asyncio dialects to provide a sync-style pep-249 facade on top
+    of the asyncio connection/cursor API provided by the driver.
+
+    .. versionadded:: 1.4.24
+
+    """
+
+    __slots__ = ("_connection",)
+
+    _connection: Any
+
+    @property
+    def driver_connection(self) -> Any:
+        """The connection object as returned by the driver after a connect."""
+        return self._connection
+
+    def run_async(self, fn: Callable[[Any], Awaitable[_T]]) -> _T:
+        """Run the awaitable returned by the given function, which is passed
+        the raw asyncio driver connection.
+
+        This is used to invoke awaitable-only methods on the driver connection
+        within the context of a "synchronous" method, like a connection
+        pool event handler.
+
+        E.g.::
+
+            engine = create_async_engine(...)
+
+
+            @event.listens_for(engine.sync_engine, "connect")
+            def register_custom_types(
+                dbapi_connection,  # ...
+            ):
+                dbapi_connection.run_async(
+                    lambda connection: connection.set_type_codec(
+                        "MyCustomType", encoder, decoder, ...
+                    )
+                )
+
+        .. versionadded:: 1.4.30
+
+        .. seealso::
+
+            :ref:`asyncio_events_run_async`
+
+        """
+        return await_only(fn(self._connection))
+
+    def __repr__(self) -> str:
+        return "<AdaptedConnection %s>" % self._connection
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/mock.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/mock.py
new file mode 100644
index 00000000..08dba5a6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/mock.py
@@ -0,0 +1,133 @@
+# engine/mock.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from operator import attrgetter
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Optional
+from typing import Type
+from typing import Union
+
+from . import url as _url
+from .. import util
+
+
+if typing.TYPE_CHECKING:
+    from .base import Engine
+    from .interfaces import _CoreAnyExecuteParams
+    from .interfaces import CoreExecuteOptionsParameter
+    from .interfaces import Dialect
+    from .url import URL
+    from ..sql.base import Executable
+    from ..sql.ddl import SchemaDropper
+    from ..sql.ddl import SchemaGenerator
+    from ..sql.schema import HasSchemaAttr
+    from ..sql.schema import SchemaItem
+
+
+class MockConnection:
+    def __init__(self, dialect: Dialect, execute: Callable[..., Any]):
+        self._dialect = dialect
+        self._execute_impl = execute
+
+    engine: Engine = cast(Any, property(lambda s: s))
+    dialect: Dialect = cast(Any, property(attrgetter("_dialect")))
+    name: str = cast(Any, property(lambda s: s._dialect.name))
+
+    def connect(self, **kwargs: Any) -> MockConnection:
+        return self
+
+    def schema_for_object(self, obj: HasSchemaAttr) -> Optional[str]:
+        return obj.schema
+
+    def execution_options(self, **kw: Any) -> MockConnection:
+        return self
+
+    def _run_ddl_visitor(
+        self,
+        visitorcallable: Type[Union[SchemaGenerator, SchemaDropper]],
+        element: SchemaItem,
+        **kwargs: Any,
+    ) -> None:
+        kwargs["checkfirst"] = False
+        visitorcallable(self.dialect, self, **kwargs).traverse_single(element)
+
+    def execute(
+        self,
+        obj: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Any:
+        return self._execute_impl(obj, parameters)
+
+
+def create_mock_engine(
+    url: Union[str, URL], executor: Any, **kw: Any
+) -> MockConnection:
+    """Create a "mock" engine used for echoing DDL.
+
+    This is a utility function used for debugging or storing the output of DDL
+    sequences as generated by :meth:`_schema.MetaData.create_all`
+    and related methods.
+
+    The function accepts a URL which is used only to determine the kind of
+    dialect to be used, as well as an "executor" callable function which
+    will receive a SQL expression object and parameters, which can then be
+    echoed or otherwise printed.   The executor's return value is not handled,
+    nor does the engine allow regular string statements to be invoked, and
+    is therefore only useful for DDL that is sent to the database without
+    receiving any results.
+
+    E.g.::
+
+        from sqlalchemy import create_mock_engine
+
+
+        def dump(sql, *multiparams, **params):
+            print(sql.compile(dialect=engine.dialect))
+
+
+        engine = create_mock_engine("postgresql+psycopg2://", dump)
+        metadata.create_all(engine, checkfirst=False)
+
+    :param url: A string URL which typically needs to contain only the
+     database backend name.
+
+    :param executor: a callable which receives the arguments ``sql``,
+     ``*multiparams`` and ``**params``.  The ``sql`` parameter is typically
+     an instance of :class:`.ExecutableDDLElement`, which can then be compiled
+     into a string using :meth:`.ExecutableDDLElement.compile`.
+
+    .. versionadded:: 1.4 - the :func:`.create_mock_engine` function replaces
+       the previous "mock" engine strategy used with
+       :func:`_sa.create_engine`.
+
+    .. seealso::
+
+        :ref:`faq_ddl_as_string`
+
+    """
+
+    # create url.URL object
+    u = _url.make_url(url)
+
+    dialect_cls = u.get_dialect()
+
+    dialect_args = {}
+    # consume dialect arguments from kwargs
+    for k in util.get_cls_kwargs(dialect_cls):
+        if k in kw:
+            dialect_args[k] = kw.pop(k)
+
+    # create dialect
+    dialect = dialect_cls(**dialect_args)
+
+    return MockConnection(dialect, executor)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/processors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/processors.py
new file mode 100644
index 00000000..b3f93308
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/processors.py
@@ -0,0 +1,61 @@
+# engine/processors.py
+# Copyright (C) 2010-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+# Copyright (C) 2010 Gaetan de Menten gdementen@gmail.com
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""defines generic type conversion functions, as used in bind and result
+processors.
+
+They all share one common characteristic: None is passed through unchanged.
+
+"""
+from __future__ import annotations
+
+import typing
+
+from ._py_processors import str_to_datetime_processor_factory  # noqa
+from ..util._has_cy import HAS_CYEXTENSION
+
+if typing.TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_processors import int_to_boolean as int_to_boolean
+    from ._py_processors import str_to_date as str_to_date
+    from ._py_processors import str_to_datetime as str_to_datetime
+    from ._py_processors import str_to_time as str_to_time
+    from ._py_processors import (
+        to_decimal_processor_factory as to_decimal_processor_factory,
+    )
+    from ._py_processors import to_float as to_float
+    from ._py_processors import to_str as to_str
+else:
+    from sqlalchemy.cyextension.processors import (
+        DecimalResultProcessor,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401
+        int_to_boolean as int_to_boolean,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401,E501
+        str_to_date as str_to_date,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401
+        str_to_datetime as str_to_datetime,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401,E501
+        str_to_time as str_to_time,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401,E501
+        to_float as to_float,
+    )
+    from sqlalchemy.cyextension.processors import (  # noqa: F401,E501
+        to_str as to_str,
+    )
+
+    def to_decimal_processor_factory(target_class, scale):
+        # Note that the scale argument is not taken into account for integer
+        # values in the C implementation while it is in the Python one.
+        # For example, the Python implementation might return
+        # Decimal('5.00000') whereas the C implementation will
+        # return Decimal('5'). These are equivalent of course.
+        return DecimalResultProcessor(target_class, "%%.%df" % scale).process
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/reflection.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/reflection.py
new file mode 100644
index 00000000..5d754c67
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/reflection.py
@@ -0,0 +1,2099 @@
+# engine/reflection.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Provides an abstraction for obtaining database schema information.
+
+Usage Notes:
+
+Here are some general conventions when accessing the low level inspector
+methods such as get_table_names, get_columns, etc.
+
+1. Inspector methods return lists of dicts in most cases for the following
+   reasons:
+
+   * They're both standard types that can be serialized.
+   * Using a dict instead of a tuple allows easy expansion of attributes.
+   * Using a list for the outer structure maintains order and is easy to work
+     with (e.g. list comprehension [d['name'] for d in cols]).
+
+2. Records that contain a name, such as the column name in a column record
+   use the key 'name'. So for most return values, each record will have a
+   'name' attribute..
+"""
+from __future__ import annotations
+
+import contextlib
+from dataclasses import dataclass
+from enum import auto
+from enum import Flag
+from enum import unique
+from typing import Any
+from typing import Callable
+from typing import Collection
+from typing import Dict
+from typing import Generator
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .base import Connection
+from .base import Engine
+from .. import exc
+from .. import inspection
+from .. import sql
+from .. import util
+from ..sql import operators
+from ..sql import schema as sa_schema
+from ..sql.cache_key import _ad_hoc_cache_key_from_args
+from ..sql.elements import quoted_name
+from ..sql.elements import TextClause
+from ..sql.type_api import TypeEngine
+from ..sql.visitors import InternalTraversal
+from ..util import topological
+from ..util.typing import final
+
+if TYPE_CHECKING:
+    from .interfaces import Dialect
+    from .interfaces import ReflectedCheckConstraint
+    from .interfaces import ReflectedColumn
+    from .interfaces import ReflectedForeignKeyConstraint
+    from .interfaces import ReflectedIndex
+    from .interfaces import ReflectedPrimaryKeyConstraint
+    from .interfaces import ReflectedTableComment
+    from .interfaces import ReflectedUniqueConstraint
+    from .interfaces import TableKey
+
+_R = TypeVar("_R")
+
+
+@util.decorator
+def cache(
+    fn: Callable[..., _R],
+    self: Dialect,
+    con: Connection,
+    *args: Any,
+    **kw: Any,
+) -> _R:
+    info_cache = kw.get("info_cache", None)
+    if info_cache is None:
+        return fn(self, con, *args, **kw)
+    exclude = {"info_cache", "unreflectable"}
+    key = (
+        fn.__name__,
+        tuple(
+            (str(a), a.quote) if isinstance(a, quoted_name) else a
+            for a in args
+            if isinstance(a, str)
+        ),
+        tuple(
+            (k, (str(v), v.quote) if isinstance(v, quoted_name) else v)
+            for k, v in kw.items()
+            if k not in exclude
+        ),
+    )
+    ret: _R = info_cache.get(key)
+    if ret is None:
+        ret = fn(self, con, *args, **kw)
+        info_cache[key] = ret
+    return ret
+
+
+def flexi_cache(
+    *traverse_args: Tuple[str, InternalTraversal]
+) -> Callable[[Callable[..., _R]], Callable[..., _R]]:
+    @util.decorator
+    def go(
+        fn: Callable[..., _R],
+        self: Dialect,
+        con: Connection,
+        *args: Any,
+        **kw: Any,
+    ) -> _R:
+        info_cache = kw.get("info_cache", None)
+        if info_cache is None:
+            return fn(self, con, *args, **kw)
+        key = _ad_hoc_cache_key_from_args((fn.__name__,), traverse_args, args)
+        ret: _R = info_cache.get(key)
+        if ret is None:
+            ret = fn(self, con, *args, **kw)
+            info_cache[key] = ret
+        return ret
+
+    return go
+
+
+@unique
+class ObjectKind(Flag):
+    """Enumerator that indicates which kind of object to return when calling
+    the ``get_multi`` methods.
+
+    This is a Flag enum, so custom combinations can be passed. For example,
+    to reflect tables and plain views ``ObjectKind.TABLE | ObjectKind.VIEW``
+    may be used.
+
+    .. note::
+      Not all dialect may support all kind of object. If a dialect does
+      not support a particular object an empty dict is returned.
+      In case a dialect supports an object, but the requested method
+      is not applicable for the specified kind the default value
+      will be returned for each reflected object. For example reflecting
+      check constraints of view return a dict with all the views with
+      empty lists as values.
+    """
+
+    TABLE = auto()
+    "Reflect table objects"
+    VIEW = auto()
+    "Reflect plain view objects"
+    MATERIALIZED_VIEW = auto()
+    "Reflect materialized view object"
+
+    ANY_VIEW = VIEW | MATERIALIZED_VIEW
+    "Reflect any kind of view objects"
+    ANY = TABLE | VIEW | MATERIALIZED_VIEW
+    "Reflect all type of objects"
+
+
+@unique
+class ObjectScope(Flag):
+    """Enumerator that indicates which scope to use when calling
+    the ``get_multi`` methods.
+    """
+
+    DEFAULT = auto()
+    "Include default scope"
+    TEMPORARY = auto()
+    "Include only temp scope"
+    ANY = DEFAULT | TEMPORARY
+    "Include both default and temp scope"
+
+
+@inspection._self_inspects
+class Inspector(inspection.Inspectable["Inspector"]):
+    """Performs database schema inspection.
+
+    The Inspector acts as a proxy to the reflection methods of the
+    :class:`~sqlalchemy.engine.interfaces.Dialect`, providing a
+    consistent interface as well as caching support for previously
+    fetched metadata.
+
+    A :class:`_reflection.Inspector` object is usually created via the
+    :func:`_sa.inspect` function, which may be passed an
+    :class:`_engine.Engine`
+    or a :class:`_engine.Connection`::
+
+        from sqlalchemy import inspect, create_engine
+
+        engine = create_engine("...")
+        insp = inspect(engine)
+
+    Where above, the :class:`~sqlalchemy.engine.interfaces.Dialect` associated
+    with the engine may opt to return an :class:`_reflection.Inspector`
+    subclass that
+    provides additional methods specific to the dialect's target database.
+
+    """
+
+    bind: Union[Engine, Connection]
+    engine: Engine
+    _op_context_requires_connect: bool
+    dialect: Dialect
+    info_cache: Dict[Any, Any]
+
+    @util.deprecated(
+        "1.4",
+        "The __init__() method on :class:`_reflection.Inspector` "
+        "is deprecated and "
+        "will be removed in a future release.  Please use the "
+        ":func:`.sqlalchemy.inspect` "
+        "function on an :class:`_engine.Engine` or "
+        ":class:`_engine.Connection` "
+        "in order to "
+        "acquire an :class:`_reflection.Inspector`.",
+    )
+    def __init__(self, bind: Union[Engine, Connection]):
+        """Initialize a new :class:`_reflection.Inspector`.
+
+        :param bind: a :class:`~sqlalchemy.engine.Connection`,
+          which is typically an instance of
+          :class:`~sqlalchemy.engine.Engine` or
+          :class:`~sqlalchemy.engine.Connection`.
+
+        For a dialect-specific instance of :class:`_reflection.Inspector`, see
+        :meth:`_reflection.Inspector.from_engine`
+
+        """
+        self._init_legacy(bind)
+
+    @classmethod
+    def _construct(
+        cls, init: Callable[..., Any], bind: Union[Engine, Connection]
+    ) -> Inspector:
+        if hasattr(bind.dialect, "inspector"):
+            cls = bind.dialect.inspector
+
+        self = cls.__new__(cls)
+        init(self, bind)
+        return self
+
+    def _init_legacy(self, bind: Union[Engine, Connection]) -> None:
+        if hasattr(bind, "exec_driver_sql"):
+            self._init_connection(bind)  # type: ignore[arg-type]
+        else:
+            self._init_engine(bind)
+
+    def _init_engine(self, engine: Engine) -> None:
+        self.bind = self.engine = engine
+        engine.connect().close()
+        self._op_context_requires_connect = True
+        self.dialect = self.engine.dialect
+        self.info_cache = {}
+
+    def _init_connection(self, connection: Connection) -> None:
+        self.bind = connection
+        self.engine = connection.engine
+        self._op_context_requires_connect = False
+        self.dialect = self.engine.dialect
+        self.info_cache = {}
+
+    def clear_cache(self) -> None:
+        """reset the cache for this :class:`.Inspector`.
+
+        Inspection methods that have data cached will emit SQL queries
+        when next called to get new data.
+
+        .. versionadded:: 2.0
+
+        """
+        self.info_cache.clear()
+
+    @classmethod
+    @util.deprecated(
+        "1.4",
+        "The from_engine() method on :class:`_reflection.Inspector` "
+        "is deprecated and "
+        "will be removed in a future release.  Please use the "
+        ":func:`.sqlalchemy.inspect` "
+        "function on an :class:`_engine.Engine` or "
+        ":class:`_engine.Connection` "
+        "in order to "
+        "acquire an :class:`_reflection.Inspector`.",
+    )
+    def from_engine(cls, bind: Engine) -> Inspector:
+        """Construct a new dialect-specific Inspector object from the given
+        engine or connection.
+
+        :param bind: a :class:`~sqlalchemy.engine.Connection`
+         or :class:`~sqlalchemy.engine.Engine`.
+
+        This method differs from direct a direct constructor call of
+        :class:`_reflection.Inspector` in that the
+        :class:`~sqlalchemy.engine.interfaces.Dialect` is given a chance to
+        provide a dialect-specific :class:`_reflection.Inspector` instance,
+        which may
+        provide additional methods.
+
+        See the example at :class:`_reflection.Inspector`.
+
+        """
+        return cls._construct(cls._init_legacy, bind)
+
+    @inspection._inspects(Engine)
+    def _engine_insp(bind: Engine) -> Inspector:  # type: ignore[misc]
+        return Inspector._construct(Inspector._init_engine, bind)
+
+    @inspection._inspects(Connection)
+    def _connection_insp(bind: Connection) -> Inspector:  # type: ignore[misc]
+        return Inspector._construct(Inspector._init_connection, bind)
+
+    @contextlib.contextmanager
+    def _operation_context(self) -> Generator[Connection, None, None]:
+        """Return a context that optimizes for multiple operations on a single
+        transaction.
+
+        This essentially allows connect()/close() to be called if we detected
+        that we're against an :class:`_engine.Engine` and not a
+        :class:`_engine.Connection`.
+
+        """
+        conn: Connection
+        if self._op_context_requires_connect:
+            conn = self.bind.connect()  # type: ignore[union-attr]
+        else:
+            conn = self.bind  # type: ignore[assignment]
+        try:
+            yield conn
+        finally:
+            if self._op_context_requires_connect:
+                conn.close()
+
+    @contextlib.contextmanager
+    def _inspection_context(self) -> Generator[Inspector, None, None]:
+        """Return an :class:`_reflection.Inspector`
+        from this one that will run all
+        operations on a single connection.
+
+        """
+
+        with self._operation_context() as conn:
+            sub_insp = self._construct(self.__class__._init_connection, conn)
+            sub_insp.info_cache = self.info_cache
+            yield sub_insp
+
+    @property
+    def default_schema_name(self) -> Optional[str]:
+        """Return the default schema name presented by the dialect
+        for the current engine's database user.
+
+        E.g. this is typically ``public`` for PostgreSQL and ``dbo``
+        for SQL Server.
+
+        """
+        return self.dialect.default_schema_name
+
+    def get_schema_names(self, **kw: Any) -> List[str]:
+        r"""Return all schema names.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_schema_names(
+                conn, info_cache=self.info_cache, **kw
+            )
+
+    def get_table_names(
+        self, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        r"""Return all table names within a particular schema.
+
+        The names are expected to be real tables only, not views.
+        Views are instead returned using the
+        :meth:`_reflection.Inspector.get_view_names` and/or
+        :meth:`_reflection.Inspector.get_materialized_view_names`
+        methods.
+
+        :param schema: Schema name. If ``schema`` is left at ``None``, the
+         database's default schema is
+         used, else the named schema is searched.  If the database does not
+         support named schemas, behavior is undefined if ``schema`` is not
+         passed as ``None``.  For special quoting, use :class:`.quoted_name`.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. seealso::
+
+            :meth:`_reflection.Inspector.get_sorted_table_and_fkc_names`
+
+            :attr:`_schema.MetaData.sorted_tables`
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_table_names(
+                conn, schema, info_cache=self.info_cache, **kw
+            )
+
+    def has_table(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> bool:
+        r"""Return True if the backend has a table, view, or temporary
+        table of the given name.
+
+        :param table_name: name of the table to check
+        :param schema: schema name to query, if not the default schema.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. versionadded:: 1.4 - the :meth:`.Inspector.has_table` method
+           replaces the :meth:`_engine.Engine.has_table` method.
+
+        .. versionchanged:: 2.0:: :meth:`.Inspector.has_table` now formally
+           supports checking for additional table-like objects:
+
+           * any type of views (plain or materialized)
+           * temporary tables of any kind
+
+           Previously, these two checks were not formally specified and
+           different dialects would vary in their behavior.   The dialect
+           testing suite now includes tests for all of these object types
+           and should be supported by all SQLAlchemy-included dialects.
+           Support among third party dialects may be lagging, however.
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.has_table(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def has_sequence(
+        self, sequence_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> bool:
+        r"""Return True if the backend has a sequence with the given name.
+
+        :param sequence_name: name of the sequence to check
+        :param schema: schema name to query, if not the default schema.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. versionadded:: 1.4
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.has_sequence(
+                conn, sequence_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def has_index(
+        self,
+        table_name: str,
+        index_name: str,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> bool:
+        r"""Check the existence of a particular index name in the database.
+
+        :param table_name: the name of the table the index belongs to
+        :param index_name: the name of the index to check
+        :param schema: schema name to query, if not the default schema.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. versionadded:: 2.0
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.has_index(
+                conn,
+                table_name,
+                index_name,
+                schema,
+                info_cache=self.info_cache,
+                **kw,
+            )
+
+    def has_schema(self, schema_name: str, **kw: Any) -> bool:
+        r"""Return True if the backend has a schema with the given name.
+
+        :param schema_name: name of the schema to check
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. versionadded:: 2.0
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.has_schema(
+                conn, schema_name, info_cache=self.info_cache, **kw
+            )
+
+    def get_sorted_table_and_fkc_names(
+        self,
+        schema: Optional[str] = None,
+        **kw: Any,
+    ) -> List[Tuple[Optional[str], List[Tuple[str, Optional[str]]]]]:
+        r"""Return dependency-sorted table and foreign key constraint names in
+        referred to within a particular schema.
+
+        This will yield 2-tuples of
+        ``(tablename, [(tname, fkname), (tname, fkname), ...])``
+        consisting of table names in CREATE order grouped with the foreign key
+        constraint names that are not detected as belonging to a cycle.
+        The final element
+        will be ``(None, [(tname, fkname), (tname, fkname), ..])``
+        which will consist of remaining
+        foreign key constraint names that would require a separate CREATE
+        step after-the-fact, based on dependencies between tables.
+
+        :param schema: schema name to query, if not the default schema.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. seealso::
+
+            :meth:`_reflection.Inspector.get_table_names`
+
+            :func:`.sort_tables_and_constraints` - similar method which works
+            with an already-given :class:`_schema.MetaData`.
+
+        """
+
+        return [
+            (
+                table_key[1] if table_key else None,
+                [(tname, fks) for (_, tname), fks in fk_collection],
+            )
+            for (
+                table_key,
+                fk_collection,
+            ) in self.sort_tables_on_foreign_key_dependency(
+                consider_schemas=(schema,)
+            )
+        ]
+
+    def sort_tables_on_foreign_key_dependency(
+        self,
+        consider_schemas: Collection[Optional[str]] = (None,),
+        **kw: Any,
+    ) -> List[
+        Tuple[
+            Optional[Tuple[Optional[str], str]],
+            List[Tuple[Tuple[Optional[str], str], Optional[str]]],
+        ]
+    ]:
+        r"""Return dependency-sorted table and foreign key constraint names
+        referred to within multiple schemas.
+
+        This method may be compared to
+        :meth:`.Inspector.get_sorted_table_and_fkc_names`, which
+        works on one schema at a time; here, the method is a generalization
+        that will consider multiple schemas at once including that it will
+        resolve for cross-schema foreign keys.
+
+        .. versionadded:: 2.0
+
+        """
+        SchemaTab = Tuple[Optional[str], str]
+
+        tuples: Set[Tuple[SchemaTab, SchemaTab]] = set()
+        remaining_fkcs: Set[Tuple[SchemaTab, Optional[str]]] = set()
+        fknames_for_table: Dict[SchemaTab, Set[Optional[str]]] = {}
+        tnames: List[SchemaTab] = []
+
+        for schname in consider_schemas:
+            schema_fkeys = self.get_multi_foreign_keys(schname, **kw)
+            tnames.extend(schema_fkeys)
+            for (_, tname), fkeys in schema_fkeys.items():
+                fknames_for_table[(schname, tname)] = {
+                    fk["name"] for fk in fkeys
+                }
+                for fkey in fkeys:
+                    if (
+                        tname != fkey["referred_table"]
+                        or schname != fkey["referred_schema"]
+                    ):
+                        tuples.add(
+                            (
+                                (
+                                    fkey["referred_schema"],
+                                    fkey["referred_table"],
+                                ),
+                                (schname, tname),
+                            )
+                        )
+        try:
+            candidate_sort = list(topological.sort(tuples, tnames))
+        except exc.CircularDependencyError as err:
+            edge: Tuple[SchemaTab, SchemaTab]
+            for edge in err.edges:
+                tuples.remove(edge)
+                remaining_fkcs.update(
+                    (edge[1], fkc) for fkc in fknames_for_table[edge[1]]
+                )
+
+            candidate_sort = list(topological.sort(tuples, tnames))
+        ret: List[
+            Tuple[Optional[SchemaTab], List[Tuple[SchemaTab, Optional[str]]]]
+        ]
+        ret = [
+            (
+                (schname, tname),
+                [
+                    ((schname, tname), fk)
+                    for fk in fknames_for_table[(schname, tname)].difference(
+                        name for _, name in remaining_fkcs
+                    )
+                ],
+            )
+            for (schname, tname) in candidate_sort
+        ]
+        return ret + [(None, list(remaining_fkcs))]
+
+    def get_temp_table_names(self, **kw: Any) -> List[str]:
+        r"""Return a list of temporary table names for the current bind.
+
+        This method is unsupported by most dialects; currently
+        only Oracle Database, PostgreSQL and SQLite implements it.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_temp_table_names(
+                conn, info_cache=self.info_cache, **kw
+            )
+
+    def get_temp_view_names(self, **kw: Any) -> List[str]:
+        r"""Return a list of temporary view names for the current bind.
+
+        This method is unsupported by most dialects; currently
+        only PostgreSQL and SQLite implements it.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.get_temp_view_names(
+                conn, info_cache=self.info_cache, **kw
+            )
+
+    def get_table_options(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> Dict[str, Any]:
+        r"""Return a dictionary of options specified when the table of the
+        given name was created.
+
+        This currently includes some options that apply to MySQL and Oracle
+        Database tables.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dict with the table options. The returned keys depend on the
+         dialect in use. Each one is prefixed with the dialect name.
+
+        .. seealso:: :meth:`Inspector.get_multi_table_options`
+
+        """
+        with self._operation_context() as conn:
+            return self.dialect.get_table_options(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_table_options(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, Dict[str, Any]]:
+        r"""Return a dictionary of options specified when the tables in the
+        given schema were created.
+
+        The tables can be filtered by passing the names to use to
+        ``filter_names``.
+
+        This currently includes some options that apply to MySQL and Oracle
+        tables.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if options of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are dictionaries with the table options.
+         The returned keys in each dict depend on the
+         dialect in use. Each one is prefixed with the dialect name.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_table_options`
+        """
+        with self._operation_context() as conn:
+            res = self.dialect.get_multi_table_options(
+                conn,
+                schema=schema,
+                filter_names=filter_names,
+                kind=kind,
+                scope=scope,
+                info_cache=self.info_cache,
+                **kw,
+            )
+            return dict(res)
+
+    def get_view_names(
+        self, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        r"""Return all non-materialized view names in `schema`.
+
+        :param schema: Optional, retrieve names from a non-default schema.
+         For special quoting, use :class:`.quoted_name`.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+
+        .. versionchanged:: 2.0  For those dialects that previously included
+           the names of materialized views in this list (currently PostgreSQL),
+           this method no longer returns the names of materialized views.
+           the :meth:`.Inspector.get_materialized_view_names` method should
+           be used instead.
+
+        .. seealso::
+
+            :meth:`.Inspector.get_materialized_view_names`
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_view_names(
+                conn, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_materialized_view_names(
+        self, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        r"""Return all materialized view names in `schema`.
+
+        :param schema: Optional, retrieve names from a non-default schema.
+         For special quoting, use :class:`.quoted_name`.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :meth:`.Inspector.get_view_names`
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_materialized_view_names(
+                conn, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_sequence_names(
+        self, schema: Optional[str] = None, **kw: Any
+    ) -> List[str]:
+        r"""Return all sequence names in `schema`.
+
+        :param schema: Optional, retrieve names from a non-default schema.
+         For special quoting, use :class:`.quoted_name`.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_sequence_names(
+                conn, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_view_definition(
+        self, view_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> str:
+        r"""Return definition for the plain or materialized view called
+        ``view_name``.
+
+        :param view_name: Name of the view.
+        :param schema: Optional, retrieve names from a non-default schema.
+         For special quoting, use :class:`.quoted_name`.
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_view_definition(
+                conn, view_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_columns(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> List[ReflectedColumn]:
+        r"""Return information about columns in ``table_name``.
+
+        Given a string ``table_name`` and an optional string ``schema``,
+        return column information as a list of :class:`.ReflectedColumn`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: list of dictionaries, each representing the definition of
+         a database column.
+
+        .. seealso:: :meth:`Inspector.get_multi_columns`.
+
+        """
+
+        with self._operation_context() as conn:
+            col_defs = self.dialect.get_columns(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+        if col_defs:
+            self._instantiate_types([col_defs])
+        return col_defs
+
+    def _instantiate_types(
+        self, data: Iterable[List[ReflectedColumn]]
+    ) -> None:
+        # make this easy and only return instances for coltype
+        for col_defs in data:
+            for col_def in col_defs:
+                coltype = col_def["type"]
+                if not isinstance(coltype, TypeEngine):
+                    col_def["type"] = coltype()
+
+    def get_multi_columns(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, List[ReflectedColumn]]:
+        r"""Return information about columns in all objects in the given
+        schema.
+
+        The objects can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a list of :class:`.ReflectedColumn`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if columns of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are list of dictionaries, each representing the
+         definition of a database column.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_columns`
+        """
+
+        with self._operation_context() as conn:
+            table_col_defs = dict(
+                self.dialect.get_multi_columns(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+        self._instantiate_types(table_col_defs.values())
+        return table_col_defs
+
+    def get_pk_constraint(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> ReflectedPrimaryKeyConstraint:
+        r"""Return information about primary key constraint in ``table_name``.
+
+        Given a string ``table_name``, and an optional string `schema`, return
+        primary key information as a :class:`.ReflectedPrimaryKeyConstraint`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary representing the definition of
+         a primary key constraint.
+
+        .. seealso:: :meth:`Inspector.get_multi_pk_constraint`
+        """
+        with self._operation_context() as conn:
+            return self.dialect.get_pk_constraint(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_pk_constraint(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, ReflectedPrimaryKeyConstraint]:
+        r"""Return information about primary key constraints in
+        all tables in the given schema.
+
+        The tables can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a :class:`.ReflectedPrimaryKeyConstraint`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if primary keys of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are dictionaries, each representing the
+         definition of a primary key constraint.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_pk_constraint`
+        """
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_pk_constraint(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def get_foreign_keys(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> List[ReflectedForeignKeyConstraint]:
+        r"""Return information about foreign_keys in ``table_name``.
+
+        Given a string ``table_name``, and an optional string `schema`, return
+        foreign key information as a list of
+        :class:`.ReflectedForeignKeyConstraint`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a list of dictionaries, each representing the
+         a foreign key definition.
+
+        .. seealso:: :meth:`Inspector.get_multi_foreign_keys`
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_foreign_keys(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_foreign_keys(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, List[ReflectedForeignKeyConstraint]]:
+        r"""Return information about foreign_keys in all tables
+        in the given schema.
+
+        The tables can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a list of
+        :class:`.ReflectedForeignKeyConstraint`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if foreign keys of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are list of dictionaries, each representing
+         a foreign key definition.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_foreign_keys`
+        """
+
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_foreign_keys(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def get_indexes(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> List[ReflectedIndex]:
+        r"""Return information about indexes in ``table_name``.
+
+        Given a string ``table_name`` and an optional string `schema`, return
+        index information as a list of :class:`.ReflectedIndex`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a list of dictionaries, each representing the
+         definition of an index.
+
+        .. seealso:: :meth:`Inspector.get_multi_indexes`
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_indexes(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_indexes(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, List[ReflectedIndex]]:
+        r"""Return information about indexes in in all objects
+        in the given schema.
+
+        The objects can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a list of :class:`.ReflectedIndex`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if indexes of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are list of dictionaries, each representing the
+         definition of an index.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_indexes`
+        """
+
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_indexes(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def get_unique_constraints(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> List[ReflectedUniqueConstraint]:
+        r"""Return information about unique constraints in ``table_name``.
+
+        Given a string ``table_name`` and an optional string `schema`, return
+        unique constraint information as a list of
+        :class:`.ReflectedUniqueConstraint`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a list of dictionaries, each representing the
+         definition of an unique constraint.
+
+        .. seealso:: :meth:`Inspector.get_multi_unique_constraints`
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_unique_constraints(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_unique_constraints(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, List[ReflectedUniqueConstraint]]:
+        r"""Return information about unique constraints in all tables
+        in the given schema.
+
+        The tables can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a list of
+        :class:`.ReflectedUniqueConstraint`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if constraints of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are list of dictionaries, each representing the
+         definition of an unique constraint.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_unique_constraints`
+        """
+
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_unique_constraints(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def get_table_comment(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> ReflectedTableComment:
+        r"""Return information about the table comment for ``table_name``.
+
+        Given a string ``table_name`` and an optional string ``schema``,
+        return table comment information as a :class:`.ReflectedTableComment`.
+
+        Raises ``NotImplementedError`` for a dialect that does not support
+        comments.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary, with the table comment.
+
+        .. versionadded:: 1.2
+
+        .. seealso:: :meth:`Inspector.get_multi_table_comment`
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_table_comment(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_table_comment(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, ReflectedTableComment]:
+        r"""Return information about the table comment in all objects
+        in the given schema.
+
+        The objects can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a :class:`.ReflectedTableComment`.
+
+        Raises ``NotImplementedError`` for a dialect that does not support
+        comments.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if comments of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are dictionaries, representing the
+         table comments.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_table_comment`
+        """
+
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_table_comment(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def get_check_constraints(
+        self, table_name: str, schema: Optional[str] = None, **kw: Any
+    ) -> List[ReflectedCheckConstraint]:
+        r"""Return information about check constraints in ``table_name``.
+
+        Given a string ``table_name`` and an optional string `schema`, return
+        check constraint information as a list of
+        :class:`.ReflectedCheckConstraint`.
+
+        :param table_name: string name of the table.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a list of dictionaries, each representing the
+         definition of a check constraints.
+
+        .. seealso:: :meth:`Inspector.get_multi_check_constraints`
+        """
+
+        with self._operation_context() as conn:
+            return self.dialect.get_check_constraints(
+                conn, table_name, schema, info_cache=self.info_cache, **kw
+            )
+
+    def get_multi_check_constraints(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Sequence[str]] = None,
+        kind: ObjectKind = ObjectKind.TABLE,
+        scope: ObjectScope = ObjectScope.DEFAULT,
+        **kw: Any,
+    ) -> Dict[TableKey, List[ReflectedCheckConstraint]]:
+        r"""Return information about check constraints in all tables
+        in the given schema.
+
+        The tables can be filtered by passing the names to use to
+        ``filter_names``.
+
+        For each table the value is a list of
+        :class:`.ReflectedCheckConstraint`.
+
+        :param schema: string schema name; if omitted, uses the default schema
+         of the database connection.  For special quoting,
+         use :class:`.quoted_name`.
+
+        :param filter_names: optionally return information only for the
+         objects listed here.
+
+        :param kind: a :class:`.ObjectKind` that specifies the type of objects
+         to reflect. Defaults to ``ObjectKind.TABLE``.
+
+        :param scope: a :class:`.ObjectScope` that specifies if constraints of
+         default, temporary or any tables should be reflected.
+         Defaults to ``ObjectScope.DEFAULT``.
+
+        :param \**kw: Additional keyword argument to pass to the dialect
+         specific implementation. See the documentation of the dialect
+         in use for more information.
+
+        :return: a dictionary where the keys are two-tuple schema,table-name
+         and the values are list of dictionaries, each representing the
+         definition of a check constraints.
+         The schema is ``None`` if no schema is provided.
+
+        .. versionadded:: 2.0
+
+        .. seealso:: :meth:`Inspector.get_check_constraints`
+        """
+
+        with self._operation_context() as conn:
+            return dict(
+                self.dialect.get_multi_check_constraints(
+                    conn,
+                    schema=schema,
+                    filter_names=filter_names,
+                    kind=kind,
+                    scope=scope,
+                    info_cache=self.info_cache,
+                    **kw,
+                )
+            )
+
+    def reflect_table(
+        self,
+        table: sa_schema.Table,
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str] = (),
+        resolve_fks: bool = True,
+        _extend_on: Optional[Set[sa_schema.Table]] = None,
+        _reflect_info: Optional[_ReflectionInfo] = None,
+    ) -> None:
+        """Given a :class:`_schema.Table` object, load its internal
+        constructs based on introspection.
+
+        This is the underlying method used by most dialects to produce
+        table reflection.  Direct usage is like::
+
+            from sqlalchemy import create_engine, MetaData, Table
+            from sqlalchemy import inspect
+
+            engine = create_engine("...")
+            meta = MetaData()
+            user_table = Table("user", meta)
+            insp = inspect(engine)
+            insp.reflect_table(user_table, None)
+
+        .. versionchanged:: 1.4 Renamed from ``reflecttable`` to
+           ``reflect_table``
+
+        :param table: a :class:`~sqlalchemy.schema.Table` instance.
+        :param include_columns: a list of string column names to include
+          in the reflection process.  If ``None``, all columns are reflected.
+
+        """
+
+        if _extend_on is not None:
+            if table in _extend_on:
+                return
+            else:
+                _extend_on.add(table)
+
+        dialect = self.bind.dialect
+
+        with self._operation_context() as conn:
+            schema = conn.schema_for_object(table)
+
+        table_name = table.name
+
+        # get table-level arguments that are specifically
+        # intended for reflection, e.g. oracle_resolve_synonyms.
+        # these are unconditionally passed to related Table
+        # objects
+        reflection_options = {
+            k: table.dialect_kwargs.get(k)
+            for k in dialect.reflection_options
+            if k in table.dialect_kwargs
+        }
+
+        table_key = (schema, table_name)
+        if _reflect_info is None or table_key not in _reflect_info.columns:
+            _reflect_info = self._get_reflection_info(
+                schema,
+                filter_names=[table_name],
+                kind=ObjectKind.ANY,
+                scope=ObjectScope.ANY,
+                _reflect_info=_reflect_info,
+                **table.dialect_kwargs,
+            )
+        if table_key in _reflect_info.unreflectable:
+            raise _reflect_info.unreflectable[table_key]
+
+        if table_key not in _reflect_info.columns:
+            raise exc.NoSuchTableError(table_name)
+
+        # reflect table options, like mysql_engine
+        if _reflect_info.table_options:
+            tbl_opts = _reflect_info.table_options.get(table_key)
+            if tbl_opts:
+                # add additional kwargs to the Table if the dialect
+                # returned them
+                table._validate_dialect_kwargs(tbl_opts)
+
+        found_table = False
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]] = {}
+
+        for col_d in _reflect_info.columns[table_key]:
+            found_table = True
+
+            self._reflect_column(
+                table,
+                col_d,
+                include_columns,
+                exclude_columns,
+                cols_by_orig_name,
+            )
+
+        # NOTE: support tables/views with no columns
+        if not found_table and not self.has_table(table_name, schema):
+            raise exc.NoSuchTableError(table_name)
+
+        self._reflect_pk(
+            _reflect_info, table_key, table, cols_by_orig_name, exclude_columns
+        )
+
+        self._reflect_fk(
+            _reflect_info,
+            table_key,
+            table,
+            cols_by_orig_name,
+            include_columns,
+            exclude_columns,
+            resolve_fks,
+            _extend_on,
+            reflection_options,
+        )
+
+        self._reflect_indexes(
+            _reflect_info,
+            table_key,
+            table,
+            cols_by_orig_name,
+            include_columns,
+            exclude_columns,
+            reflection_options,
+        )
+
+        self._reflect_unique_constraints(
+            _reflect_info,
+            table_key,
+            table,
+            cols_by_orig_name,
+            include_columns,
+            exclude_columns,
+            reflection_options,
+        )
+
+        self._reflect_check_constraints(
+            _reflect_info,
+            table_key,
+            table,
+            cols_by_orig_name,
+            include_columns,
+            exclude_columns,
+            reflection_options,
+        )
+
+        self._reflect_table_comment(
+            _reflect_info,
+            table_key,
+            table,
+            reflection_options,
+        )
+
+    def _reflect_column(
+        self,
+        table: sa_schema.Table,
+        col_d: ReflectedColumn,
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str],
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+    ) -> None:
+        orig_name = col_d["name"]
+
+        table.metadata.dispatch.column_reflect(self, table, col_d)
+        table.dispatch.column_reflect(self, table, col_d)
+
+        # fetch name again as column_reflect is allowed to
+        # change it
+        name = col_d["name"]
+        if (include_columns and name not in include_columns) or (
+            exclude_columns and name in exclude_columns
+        ):
+            return
+
+        coltype = col_d["type"]
+
+        col_kw = {
+            k: col_d[k]  # type: ignore[literal-required]
+            for k in [
+                "nullable",
+                "autoincrement",
+                "quote",
+                "info",
+                "key",
+                "comment",
+            ]
+            if k in col_d
+        }
+
+        if "dialect_options" in col_d:
+            col_kw.update(col_d["dialect_options"])
+
+        colargs = []
+        default: Any
+        if col_d.get("default") is not None:
+            default_text = col_d["default"]
+            assert default_text is not None
+            if isinstance(default_text, TextClause):
+                default = sa_schema.DefaultClause(
+                    default_text, _reflected=True
+                )
+            elif not isinstance(default_text, sa_schema.FetchedValue):
+                default = sa_schema.DefaultClause(
+                    sql.text(default_text), _reflected=True
+                )
+            else:
+                default = default_text
+            colargs.append(default)
+
+        if "computed" in col_d:
+            computed = sa_schema.Computed(**col_d["computed"])
+            colargs.append(computed)
+
+        if "identity" in col_d:
+            identity = sa_schema.Identity(**col_d["identity"])
+            colargs.append(identity)
+
+        cols_by_orig_name[orig_name] = col = sa_schema.Column(
+            name, coltype, *colargs, **col_kw
+        )
+
+        if col.key in table.primary_key:
+            col.primary_key = True
+        table.append_column(col, replace_existing=True)
+
+    def _reflect_pk(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+        exclude_columns: Collection[str],
+    ) -> None:
+        pk_cons = _reflect_info.pk_constraint.get(table_key)
+        if pk_cons:
+            pk_cols = [
+                cols_by_orig_name[pk]
+                for pk in pk_cons["constrained_columns"]
+                if pk in cols_by_orig_name and pk not in exclude_columns
+            ]
+
+            # update pk constraint name and comment
+            table.primary_key.name = pk_cons.get("name")
+            table.primary_key.comment = pk_cons.get("comment", None)
+
+            # tell the PKConstraint to re-initialize
+            # its column collection
+            table.primary_key._reload(pk_cols)
+
+    def _reflect_fk(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str],
+        resolve_fks: bool,
+        _extend_on: Optional[Set[sa_schema.Table]],
+        reflection_options: Dict[str, Any],
+    ) -> None:
+        fkeys = _reflect_info.foreign_keys.get(table_key, [])
+        for fkey_d in fkeys:
+            conname = fkey_d["name"]
+            # look for columns by orig name in cols_by_orig_name,
+            # but support columns that are in-Python only as fallback
+            constrained_columns = [
+                cols_by_orig_name[c].key if c in cols_by_orig_name else c
+                for c in fkey_d["constrained_columns"]
+            ]
+
+            if (
+                exclude_columns
+                and set(constrained_columns).intersection(exclude_columns)
+                or (
+                    include_columns
+                    and set(constrained_columns).difference(include_columns)
+                )
+            ):
+                continue
+
+            referred_schema = fkey_d["referred_schema"]
+            referred_table = fkey_d["referred_table"]
+            referred_columns = fkey_d["referred_columns"]
+            refspec = []
+            if referred_schema is not None:
+                if resolve_fks:
+                    sa_schema.Table(
+                        referred_table,
+                        table.metadata,
+                        schema=referred_schema,
+                        autoload_with=self.bind,
+                        _extend_on=_extend_on,
+                        _reflect_info=_reflect_info,
+                        **reflection_options,
+                    )
+                for column in referred_columns:
+                    refspec.append(
+                        ".".join([referred_schema, referred_table, column])
+                    )
+            else:
+                if resolve_fks:
+                    sa_schema.Table(
+                        referred_table,
+                        table.metadata,
+                        autoload_with=self.bind,
+                        schema=sa_schema.BLANK_SCHEMA,
+                        _extend_on=_extend_on,
+                        _reflect_info=_reflect_info,
+                        **reflection_options,
+                    )
+                for column in referred_columns:
+                    refspec.append(".".join([referred_table, column]))
+            if "options" in fkey_d:
+                options = fkey_d["options"]
+            else:
+                options = {}
+
+            try:
+                table.append_constraint(
+                    sa_schema.ForeignKeyConstraint(
+                        constrained_columns,
+                        refspec,
+                        conname,
+                        link_to_name=True,
+                        comment=fkey_d.get("comment"),
+                        **options,
+                    )
+                )
+            except exc.ConstraintColumnNotFoundError:
+                util.warn(
+                    f"On reflected table {table.name}, skipping reflection of "
+                    "foreign key constraint "
+                    f"{conname}; one or more subject columns within "
+                    f"name(s) {', '.join(constrained_columns)} are not "
+                    "present in the table"
+                )
+
+    _index_sort_exprs = {
+        "asc": operators.asc_op,
+        "desc": operators.desc_op,
+        "nulls_first": operators.nulls_first_op,
+        "nulls_last": operators.nulls_last_op,
+    }
+
+    def _reflect_indexes(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str],
+        reflection_options: Dict[str, Any],
+    ) -> None:
+        # Indexes
+        indexes = _reflect_info.indexes.get(table_key, [])
+        for index_d in indexes:
+            name = index_d["name"]
+            columns = index_d["column_names"]
+            expressions = index_d.get("expressions")
+            column_sorting = index_d.get("column_sorting", {})
+            unique = index_d["unique"]
+            flavor = index_d.get("type", "index")
+            dialect_options = index_d.get("dialect_options", {})
+
+            duplicates = index_d.get("duplicates_constraint")
+            if include_columns and not set(columns).issubset(include_columns):
+                continue
+            if duplicates:
+                continue
+            # look for columns by orig name in cols_by_orig_name,
+            # but support columns that are in-Python only as fallback
+            idx_element: Any
+            idx_elements = []
+            for index, c in enumerate(columns):
+                if c is None:
+                    if not expressions:
+                        util.warn(
+                            f"Skipping {flavor} {name!r} because key "
+                            f"{index + 1} reflected as None but no "
+                            "'expressions' were returned"
+                        )
+                        break
+                    idx_element = sql.text(expressions[index])
+                else:
+                    try:
+                        if c in cols_by_orig_name:
+                            idx_element = cols_by_orig_name[c]
+                        else:
+                            idx_element = table.c[c]
+                    except KeyError:
+                        util.warn(
+                            f"{flavor} key {c!r} was not located in "
+                            f"columns for table {table.name!r}"
+                        )
+                        continue
+                    for option in column_sorting.get(c, ()):
+                        if option in self._index_sort_exprs:
+                            op = self._index_sort_exprs[option]
+                            idx_element = op(idx_element)
+                idx_elements.append(idx_element)
+            else:
+                sa_schema.Index(
+                    name,
+                    *idx_elements,
+                    _table=table,
+                    unique=unique,
+                    **dialect_options,
+                )
+
+    def _reflect_unique_constraints(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str],
+        reflection_options: Dict[str, Any],
+    ) -> None:
+        constraints = _reflect_info.unique_constraints.get(table_key, [])
+        # Unique Constraints
+        for const_d in constraints:
+            conname = const_d["name"]
+            columns = const_d["column_names"]
+            comment = const_d.get("comment")
+            duplicates = const_d.get("duplicates_index")
+            dialect_options = const_d.get("dialect_options", {})
+            if include_columns and not set(columns).issubset(include_columns):
+                continue
+            if duplicates:
+                continue
+            # look for columns by orig name in cols_by_orig_name,
+            # but support columns that are in-Python only as fallback
+            constrained_cols = []
+            for c in columns:
+                try:
+                    constrained_col = (
+                        cols_by_orig_name[c]
+                        if c in cols_by_orig_name
+                        else table.c[c]
+                    )
+                except KeyError:
+                    util.warn(
+                        "unique constraint key '%s' was not located in "
+                        "columns for table '%s'" % (c, table.name)
+                    )
+                else:
+                    constrained_cols.append(constrained_col)
+            table.append_constraint(
+                sa_schema.UniqueConstraint(
+                    *constrained_cols,
+                    name=conname,
+                    comment=comment,
+                    **dialect_options,
+                )
+            )
+
+    def _reflect_check_constraints(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        cols_by_orig_name: Dict[str, sa_schema.Column[Any]],
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str],
+        reflection_options: Dict[str, Any],
+    ) -> None:
+        constraints = _reflect_info.check_constraints.get(table_key, [])
+        for const_d in constraints:
+            table.append_constraint(sa_schema.CheckConstraint(**const_d))
+
+    def _reflect_table_comment(
+        self,
+        _reflect_info: _ReflectionInfo,
+        table_key: TableKey,
+        table: sa_schema.Table,
+        reflection_options: Dict[str, Any],
+    ) -> None:
+        comment_dict = _reflect_info.table_comment.get(table_key)
+        if comment_dict:
+            table.comment = comment_dict["text"]
+
+    def _get_reflection_info(
+        self,
+        schema: Optional[str] = None,
+        filter_names: Optional[Collection[str]] = None,
+        available: Optional[Collection[str]] = None,
+        _reflect_info: Optional[_ReflectionInfo] = None,
+        **kw: Any,
+    ) -> _ReflectionInfo:
+        kw["schema"] = schema
+
+        if filter_names and available and len(filter_names) > 100:
+            fraction = len(filter_names) / len(available)
+        else:
+            fraction = None
+
+        unreflectable: Dict[TableKey, exc.UnreflectableTableError]
+        kw["unreflectable"] = unreflectable = {}
+
+        has_result: bool = True
+
+        def run(
+            meth: Any,
+            *,
+            optional: bool = False,
+            check_filter_names_from_meth: bool = False,
+        ) -> Any:
+            nonlocal has_result
+            # simple heuristic to improve reflection performance if a
+            # dialect implements multi_reflection:
+            # if more than 50% of the tables in the db are in filter_names
+            # load all the tables, since it's most likely faster to avoid
+            # a filter on that many tables.
+            if (
+                fraction is None
+                or fraction <= 0.5
+                or not self.dialect._overrides_default(meth.__name__)
+            ):
+                _fn = filter_names
+            else:
+                _fn = None
+            try:
+                if has_result:
+                    res = meth(filter_names=_fn, **kw)
+                    if check_filter_names_from_meth and not res:
+                        # method returned no result data.
+                        # skip any future call methods
+                        has_result = False
+                else:
+                    res = {}
+            except NotImplementedError:
+                if not optional:
+                    raise
+                res = {}
+            return res
+
+        info = _ReflectionInfo(
+            columns=run(
+                self.get_multi_columns, check_filter_names_from_meth=True
+            ),
+            pk_constraint=run(self.get_multi_pk_constraint),
+            foreign_keys=run(self.get_multi_foreign_keys),
+            indexes=run(self.get_multi_indexes),
+            unique_constraints=run(
+                self.get_multi_unique_constraints, optional=True
+            ),
+            table_comment=run(self.get_multi_table_comment, optional=True),
+            check_constraints=run(
+                self.get_multi_check_constraints, optional=True
+            ),
+            table_options=run(self.get_multi_table_options, optional=True),
+            unreflectable=unreflectable,
+        )
+        if _reflect_info:
+            _reflect_info.update(info)
+            return _reflect_info
+        else:
+            return info
+
+
+@final
+class ReflectionDefaults:
+    """provides blank default values for reflection methods."""
+
+    @classmethod
+    def columns(cls) -> List[ReflectedColumn]:
+        return []
+
+    @classmethod
+    def pk_constraint(cls) -> ReflectedPrimaryKeyConstraint:
+        return {
+            "name": None,
+            "constrained_columns": [],
+        }
+
+    @classmethod
+    def foreign_keys(cls) -> List[ReflectedForeignKeyConstraint]:
+        return []
+
+    @classmethod
+    def indexes(cls) -> List[ReflectedIndex]:
+        return []
+
+    @classmethod
+    def unique_constraints(cls) -> List[ReflectedUniqueConstraint]:
+        return []
+
+    @classmethod
+    def check_constraints(cls) -> List[ReflectedCheckConstraint]:
+        return []
+
+    @classmethod
+    def table_options(cls) -> Dict[str, Any]:
+        return {}
+
+    @classmethod
+    def table_comment(cls) -> ReflectedTableComment:
+        return {"text": None}
+
+
+@dataclass
+class _ReflectionInfo:
+    columns: Dict[TableKey, List[ReflectedColumn]]
+    pk_constraint: Dict[TableKey, Optional[ReflectedPrimaryKeyConstraint]]
+    foreign_keys: Dict[TableKey, List[ReflectedForeignKeyConstraint]]
+    indexes: Dict[TableKey, List[ReflectedIndex]]
+    # optionals
+    unique_constraints: Dict[TableKey, List[ReflectedUniqueConstraint]]
+    table_comment: Dict[TableKey, Optional[ReflectedTableComment]]
+    check_constraints: Dict[TableKey, List[ReflectedCheckConstraint]]
+    table_options: Dict[TableKey, Dict[str, Any]]
+    unreflectable: Dict[TableKey, exc.UnreflectableTableError]
+
+    def update(self, other: _ReflectionInfo) -> None:
+        for k, v in self.__dict__.items():
+            ov = getattr(other, k)
+            if ov is not None:
+                if v is None:
+                    setattr(self, k, ov)
+                else:
+                    v.update(ov)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/result.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/result.py
new file mode 100644
index 00000000..7411fd74
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/result.py
@@ -0,0 +1,2380 @@
+# engine/result.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Define generic result set constructs."""
+
+from __future__ import annotations
+
+from enum import Enum
+import functools
+import itertools
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .row import Row
+from .row import RowMapping
+from .. import exc
+from .. import util
+from ..sql.base import _generative
+from ..sql.base import HasMemoized
+from ..sql.base import InPlaceGenerative
+from ..util import HasMemoized_ro_memoized_attribute
+from ..util import NONE_SET
+from ..util._has_cy import HAS_CYEXTENSION
+from ..util.typing import Literal
+from ..util.typing import Self
+
+if typing.TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_row import tuplegetter as tuplegetter
+else:
+    from sqlalchemy.cyextension.resultproxy import tuplegetter as tuplegetter
+
+if typing.TYPE_CHECKING:
+    from ..sql.schema import Column
+    from ..sql.type_api import _ResultProcessorType
+
+_KeyType = Union[str, "Column[Any]"]
+_KeyIndexType = Union[str, "Column[Any]", int]
+
+# is overridden in cursor using _CursorKeyMapRecType
+_KeyMapRecType = Any
+
+_KeyMapType = Mapping[_KeyType, _KeyMapRecType]
+
+
+_RowData = Union[Row[Any], RowMapping, Any]
+"""A generic form of "row" that accommodates for the different kinds of
+"rows" that different result objects return, including row, row mapping, and
+scalar values"""
+
+_RawRowType = Tuple[Any, ...]
+"""represents the kind of row we get from a DBAPI cursor"""
+
+_R = TypeVar("_R", bound=_RowData)
+_T = TypeVar("_T", bound=Any)
+_TP = TypeVar("_TP", bound=Tuple[Any, ...])
+
+_InterimRowType = Union[_R, _RawRowType]
+"""a catchall "anything" kind of return type that can be applied
+across all the result types
+
+"""
+
+_InterimSupportsScalarsRowType = Union[Row[Any], Any]
+
+_ProcessorsType = Sequence[Optional["_ResultProcessorType[Any]"]]
+_TupleGetterType = Callable[[Sequence[Any]], Sequence[Any]]
+_UniqueFilterType = Callable[[Any], Any]
+_UniqueFilterStateType = Tuple[Set[Any], Optional[_UniqueFilterType]]
+
+
+class ResultMetaData:
+    """Base for metadata about result rows."""
+
+    __slots__ = ()
+
+    _tuplefilter: Optional[_TupleGetterType] = None
+    _translated_indexes: Optional[Sequence[int]] = None
+    _unique_filters: Optional[Sequence[Callable[[Any], Any]]] = None
+    _keymap: _KeyMapType
+    _keys: Sequence[str]
+    _processors: Optional[_ProcessorsType]
+    _key_to_index: Mapping[_KeyType, int]
+
+    @property
+    def keys(self) -> RMKeyView:
+        return RMKeyView(self)
+
+    def _has_key(self, key: object) -> bool:
+        raise NotImplementedError()
+
+    def _for_freeze(self) -> ResultMetaData:
+        raise NotImplementedError()
+
+    @overload
+    def _key_fallback(
+        self, key: Any, err: Optional[Exception], raiseerr: Literal[True] = ...
+    ) -> NoReturn: ...
+
+    @overload
+    def _key_fallback(
+        self,
+        key: Any,
+        err: Optional[Exception],
+        raiseerr: Literal[False] = ...,
+    ) -> None: ...
+
+    @overload
+    def _key_fallback(
+        self, key: Any, err: Optional[Exception], raiseerr: bool = ...
+    ) -> Optional[NoReturn]: ...
+
+    def _key_fallback(
+        self, key: Any, err: Optional[Exception], raiseerr: bool = True
+    ) -> Optional[NoReturn]:
+        assert raiseerr
+        raise KeyError(key) from err
+
+    def _raise_for_ambiguous_column_name(
+        self, rec: _KeyMapRecType
+    ) -> NoReturn:
+        raise NotImplementedError(
+            "ambiguous column name logic is implemented for "
+            "CursorResultMetaData"
+        )
+
+    def _index_for_key(
+        self, key: _KeyIndexType, raiseerr: bool
+    ) -> Optional[int]:
+        raise NotImplementedError()
+
+    def _indexes_for_keys(
+        self, keys: Sequence[_KeyIndexType]
+    ) -> Sequence[int]:
+        raise NotImplementedError()
+
+    def _metadata_for_keys(
+        self, keys: Sequence[_KeyIndexType]
+    ) -> Iterator[_KeyMapRecType]:
+        raise NotImplementedError()
+
+    def _reduce(self, keys: Sequence[_KeyIndexType]) -> ResultMetaData:
+        raise NotImplementedError()
+
+    def _getter(
+        self, key: Any, raiseerr: bool = True
+    ) -> Optional[Callable[[Row[Any]], Any]]:
+        index = self._index_for_key(key, raiseerr)
+
+        if index is not None:
+            return operator.itemgetter(index)
+        else:
+            return None
+
+    def _row_as_tuple_getter(
+        self, keys: Sequence[_KeyIndexType]
+    ) -> _TupleGetterType:
+        indexes = self._indexes_for_keys(keys)
+        return tuplegetter(*indexes)
+
+    def _make_key_to_index(
+        self, keymap: Mapping[_KeyType, Sequence[Any]], index: int
+    ) -> Mapping[_KeyType, int]:
+        return {
+            key: rec[index]
+            for key, rec in keymap.items()
+            if rec[index] is not None
+        }
+
+    def _key_not_found(self, key: Any, attr_error: bool) -> NoReturn:
+        if key in self._keymap:
+            # the index must be none in this case
+            self._raise_for_ambiguous_column_name(self._keymap[key])
+        else:
+            # unknown key
+            if attr_error:
+                try:
+                    self._key_fallback(key, None)
+                except KeyError as ke:
+                    raise AttributeError(ke.args[0]) from ke
+            else:
+                self._key_fallback(key, None)
+
+    @property
+    def _effective_processors(self) -> Optional[_ProcessorsType]:
+        if not self._processors or NONE_SET.issuperset(self._processors):
+            return None
+        else:
+            return self._processors
+
+
+class RMKeyView(typing.KeysView[Any]):
+    __slots__ = ("_parent", "_keys")
+
+    _parent: ResultMetaData
+    _keys: Sequence[str]
+
+    def __init__(self, parent: ResultMetaData):
+        self._parent = parent
+        self._keys = [k for k in parent._keys if k is not None]
+
+    def __len__(self) -> int:
+        return len(self._keys)
+
+    def __repr__(self) -> str:
+        return "{0.__class__.__name__}({0._keys!r})".format(self)
+
+    def __iter__(self) -> Iterator[str]:
+        return iter(self._keys)
+
+    def __contains__(self, item: Any) -> bool:
+        if isinstance(item, int):
+            return False
+
+        # note this also includes special key fallback behaviors
+        # which also don't seem to be tested in test_resultset right now
+        return self._parent._has_key(item)
+
+    def __eq__(self, other: Any) -> bool:
+        return list(other) == list(self)
+
+    def __ne__(self, other: Any) -> bool:
+        return list(other) != list(self)
+
+
+class SimpleResultMetaData(ResultMetaData):
+    """result metadata for in-memory collections."""
+
+    __slots__ = (
+        "_keys",
+        "_keymap",
+        "_processors",
+        "_tuplefilter",
+        "_translated_indexes",
+        "_unique_filters",
+        "_key_to_index",
+    )
+
+    _keys: Sequence[str]
+
+    def __init__(
+        self,
+        keys: Sequence[str],
+        extra: Optional[Sequence[Any]] = None,
+        _processors: Optional[_ProcessorsType] = None,
+        _tuplefilter: Optional[_TupleGetterType] = None,
+        _translated_indexes: Optional[Sequence[int]] = None,
+        _unique_filters: Optional[Sequence[Callable[[Any], Any]]] = None,
+    ):
+        self._keys = list(keys)
+        self._tuplefilter = _tuplefilter
+        self._translated_indexes = _translated_indexes
+        self._unique_filters = _unique_filters
+        if extra:
+            recs_names = [
+                (
+                    (name,) + (extras if extras else ()),
+                    (index, name, extras),
+                )
+                for index, (name, extras) in enumerate(zip(self._keys, extra))
+            ]
+        else:
+            recs_names = [
+                ((name,), (index, name, ()))
+                for index, name in enumerate(self._keys)
+            ]
+
+        self._keymap = {key: rec for keys, rec in recs_names for key in keys}
+
+        self._processors = _processors
+
+        self._key_to_index = self._make_key_to_index(self._keymap, 0)
+
+    def _has_key(self, key: object) -> bool:
+        return key in self._keymap
+
+    def _for_freeze(self) -> ResultMetaData:
+        unique_filters = self._unique_filters
+        if unique_filters and self._tuplefilter:
+            unique_filters = self._tuplefilter(unique_filters)
+
+        # TODO: are we freezing the result with or without uniqueness
+        # applied?
+        return SimpleResultMetaData(
+            self._keys,
+            extra=[self._keymap[key][2] for key in self._keys],
+            _unique_filters=unique_filters,
+        )
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {
+            "_keys": self._keys,
+            "_translated_indexes": self._translated_indexes,
+        }
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        if state["_translated_indexes"]:
+            _translated_indexes = state["_translated_indexes"]
+            _tuplefilter = tuplegetter(*_translated_indexes)
+        else:
+            _translated_indexes = _tuplefilter = None
+        self.__init__(  # type: ignore
+            state["_keys"],
+            _translated_indexes=_translated_indexes,
+            _tuplefilter=_tuplefilter,
+        )
+
+    def _index_for_key(self, key: Any, raiseerr: bool = True) -> int:
+        if int in key.__class__.__mro__:
+            key = self._keys[key]
+        try:
+            rec = self._keymap[key]
+        except KeyError as ke:
+            rec = self._key_fallback(key, ke, raiseerr)
+
+        return rec[0]  # type: ignore[no-any-return]
+
+    def _indexes_for_keys(self, keys: Sequence[Any]) -> Sequence[int]:
+        return [self._keymap[key][0] for key in keys]
+
+    def _metadata_for_keys(
+        self, keys: Sequence[Any]
+    ) -> Iterator[_KeyMapRecType]:
+        for key in keys:
+            if int in key.__class__.__mro__:
+                key = self._keys[key]
+
+            try:
+                rec = self._keymap[key]
+            except KeyError as ke:
+                rec = self._key_fallback(key, ke, True)
+
+            yield rec
+
+    def _reduce(self, keys: Sequence[Any]) -> ResultMetaData:
+        try:
+            metadata_for_keys = [
+                self._keymap[
+                    self._keys[key] if int in key.__class__.__mro__ else key
+                ]
+                for key in keys
+            ]
+        except KeyError as ke:
+            self._key_fallback(ke.args[0], ke, True)
+
+        indexes: Sequence[int]
+        new_keys: Sequence[str]
+        extra: Sequence[Any]
+        indexes, new_keys, extra = zip(*metadata_for_keys)
+
+        if self._translated_indexes:
+            indexes = [self._translated_indexes[idx] for idx in indexes]
+
+        tup = tuplegetter(*indexes)
+
+        new_metadata = SimpleResultMetaData(
+            new_keys,
+            extra=extra,
+            _tuplefilter=tup,
+            _translated_indexes=indexes,
+            _processors=self._processors,
+            _unique_filters=self._unique_filters,
+        )
+
+        return new_metadata
+
+
+def result_tuple(
+    fields: Sequence[str], extra: Optional[Any] = None
+) -> Callable[[Iterable[Any]], Row[Any]]:
+    parent = SimpleResultMetaData(fields, extra)
+    return functools.partial(
+        Row, parent, parent._effective_processors, parent._key_to_index
+    )
+
+
+# a symbol that indicates to internal Result methods that
+# "no row is returned".  We can't use None for those cases where a scalar
+# filter is applied to rows.
+class _NoRow(Enum):
+    _NO_ROW = 0
+
+
+_NO_ROW = _NoRow._NO_ROW
+
+
+class ResultInternal(InPlaceGenerative, Generic[_R]):
+    __slots__ = ()
+
+    _real_result: Optional[Result[Any]] = None
+    _generate_rows: bool = True
+    _row_logging_fn: Optional[Callable[[Any], Any]]
+
+    _unique_filter_state: Optional[_UniqueFilterStateType] = None
+    _post_creational_filter: Optional[Callable[[Any], Any]] = None
+    _is_cursor = False
+
+    _metadata: ResultMetaData
+
+    _source_supports_scalars: bool
+
+    def _fetchiter_impl(self) -> Iterator[_InterimRowType[Row[Any]]]:
+        raise NotImplementedError()
+
+    def _fetchone_impl(
+        self, hard_close: bool = False
+    ) -> Optional[_InterimRowType[Row[Any]]]:
+        raise NotImplementedError()
+
+    def _fetchmany_impl(
+        self, size: Optional[int] = None
+    ) -> List[_InterimRowType[Row[Any]]]:
+        raise NotImplementedError()
+
+    def _fetchall_impl(self) -> List[_InterimRowType[Row[Any]]]:
+        raise NotImplementedError()
+
+    def _soft_close(self, hard: bool = False) -> None:
+        raise NotImplementedError()
+
+    @HasMemoized_ro_memoized_attribute
+    def _row_getter(self) -> Optional[Callable[..., _R]]:
+        real_result: Result[Any] = (
+            self._real_result
+            if self._real_result
+            else cast("Result[Any]", self)
+        )
+
+        if real_result._source_supports_scalars:
+            if not self._generate_rows:
+                return None
+            else:
+                _proc = Row
+
+                def process_row(
+                    metadata: ResultMetaData,
+                    processors: Optional[_ProcessorsType],
+                    key_to_index: Mapping[_KeyType, int],
+                    scalar_obj: Any,
+                ) -> Row[Any]:
+                    return _proc(
+                        metadata, processors, key_to_index, (scalar_obj,)
+                    )
+
+        else:
+            process_row = Row  # type: ignore
+
+        metadata = self._metadata
+
+        key_to_index = metadata._key_to_index
+        processors = metadata._effective_processors
+        tf = metadata._tuplefilter
+
+        if tf and not real_result._source_supports_scalars:
+            if processors:
+                processors = tf(processors)
+
+            _make_row_orig: Callable[..., _R] = functools.partial(  # type: ignore  # noqa E501
+                process_row, metadata, processors, key_to_index
+            )
+
+            fixed_tf = tf
+
+            def make_row(row: _InterimRowType[Row[Any]]) -> _R:
+                return _make_row_orig(fixed_tf(row))
+
+        else:
+            make_row = functools.partial(  # type: ignore
+                process_row, metadata, processors, key_to_index
+            )
+
+        if real_result._row_logging_fn:
+            _log_row = real_result._row_logging_fn
+            _make_row = make_row
+
+            def make_row(row: _InterimRowType[Row[Any]]) -> _R:
+                return _log_row(_make_row(row))  # type: ignore
+
+        return make_row
+
+    @HasMemoized_ro_memoized_attribute
+    def _iterator_getter(self) -> Callable[..., Iterator[_R]]:
+        make_row = self._row_getter
+
+        post_creational_filter = self._post_creational_filter
+
+        if self._unique_filter_state:
+            uniques, strategy = self._unique_strategy
+
+            def iterrows(self: Result[Any]) -> Iterator[_R]:
+                for raw_row in self._fetchiter_impl():
+                    obj: _InterimRowType[Any] = (
+                        make_row(raw_row) if make_row else raw_row
+                    )
+                    hashed = strategy(obj) if strategy else obj
+                    if hashed in uniques:
+                        continue
+                    uniques.add(hashed)
+                    if post_creational_filter:
+                        obj = post_creational_filter(obj)
+                    yield obj  # type: ignore
+
+        else:
+
+            def iterrows(self: Result[Any]) -> Iterator[_R]:
+                for raw_row in self._fetchiter_impl():
+                    row: _InterimRowType[Any] = (
+                        make_row(raw_row) if make_row else raw_row
+                    )
+                    if post_creational_filter:
+                        row = post_creational_filter(row)
+                    yield row  # type: ignore
+
+        return iterrows
+
+    def _raw_all_rows(self) -> List[_R]:
+        make_row = self._row_getter
+        assert make_row is not None
+        rows = self._fetchall_impl()
+        return [make_row(row) for row in rows]
+
+    def _allrows(self) -> List[_R]:
+        post_creational_filter = self._post_creational_filter
+
+        make_row = self._row_getter
+
+        rows = self._fetchall_impl()
+        made_rows: List[_InterimRowType[_R]]
+        if make_row:
+            made_rows = [make_row(row) for row in rows]
+        else:
+            made_rows = rows  # type: ignore
+
+        interim_rows: List[_R]
+
+        if self._unique_filter_state:
+            uniques, strategy = self._unique_strategy
+
+            interim_rows = [
+                made_row  # type: ignore
+                for made_row, sig_row in [
+                    (
+                        made_row,
+                        strategy(made_row) if strategy else made_row,
+                    )
+                    for made_row in made_rows
+                ]
+                if sig_row not in uniques and not uniques.add(sig_row)  # type: ignore # noqa: E501
+            ]
+        else:
+            interim_rows = made_rows  # type: ignore
+
+        if post_creational_filter:
+            interim_rows = [
+                post_creational_filter(row) for row in interim_rows
+            ]
+        return interim_rows
+
+    @HasMemoized_ro_memoized_attribute
+    def _onerow_getter(
+        self,
+    ) -> Callable[..., Union[Literal[_NoRow._NO_ROW], _R]]:
+        make_row = self._row_getter
+
+        post_creational_filter = self._post_creational_filter
+
+        if self._unique_filter_state:
+            uniques, strategy = self._unique_strategy
+
+            def onerow(self: Result[Any]) -> Union[_NoRow, _R]:
+                _onerow = self._fetchone_impl
+                while True:
+                    row = _onerow()
+                    if row is None:
+                        return _NO_ROW
+                    else:
+                        obj: _InterimRowType[Any] = (
+                            make_row(row) if make_row else row
+                        )
+                        hashed = strategy(obj) if strategy else obj
+                        if hashed in uniques:
+                            continue
+                        else:
+                            uniques.add(hashed)
+                        if post_creational_filter:
+                            obj = post_creational_filter(obj)
+                        return obj  # type: ignore
+
+        else:
+
+            def onerow(self: Result[Any]) -> Union[_NoRow, _R]:
+                row = self._fetchone_impl()
+                if row is None:
+                    return _NO_ROW
+                else:
+                    interim_row: _InterimRowType[Any] = (
+                        make_row(row) if make_row else row
+                    )
+                    if post_creational_filter:
+                        interim_row = post_creational_filter(interim_row)
+                    return interim_row  # type: ignore
+
+        return onerow
+
+    @HasMemoized_ro_memoized_attribute
+    def _manyrow_getter(self) -> Callable[..., List[_R]]:
+        make_row = self._row_getter
+
+        post_creational_filter = self._post_creational_filter
+
+        if self._unique_filter_state:
+            uniques, strategy = self._unique_strategy
+
+            def filterrows(
+                make_row: Optional[Callable[..., _R]],
+                rows: List[Any],
+                strategy: Optional[Callable[[List[Any]], Any]],
+                uniques: Set[Any],
+            ) -> List[_R]:
+                if make_row:
+                    rows = [make_row(row) for row in rows]
+
+                if strategy:
+                    made_rows = (
+                        (made_row, strategy(made_row)) for made_row in rows
+                    )
+                else:
+                    made_rows = ((made_row, made_row) for made_row in rows)
+                return [
+                    made_row
+                    for made_row, sig_row in made_rows
+                    if sig_row not in uniques and not uniques.add(sig_row)  # type: ignore  # noqa: E501
+                ]
+
+            def manyrows(
+                self: ResultInternal[_R], num: Optional[int]
+            ) -> List[_R]:
+                collect: List[_R] = []
+
+                _manyrows = self._fetchmany_impl
+
+                if num is None:
+                    # if None is passed, we don't know the default
+                    # manyrows number, DBAPI has this as cursor.arraysize
+                    # different DBAPIs / fetch strategies may be different.
+                    # do a fetch to find what the number is.  if there are
+                    # only fewer rows left, then it doesn't matter.
+                    real_result = (
+                        self._real_result
+                        if self._real_result
+                        else cast("Result[Any]", self)
+                    )
+                    if real_result._yield_per:
+                        num_required = num = real_result._yield_per
+                    else:
+                        rows = _manyrows(num)
+                        num = len(rows)
+                        assert make_row is not None
+                        collect.extend(
+                            filterrows(make_row, rows, strategy, uniques)
+                        )
+                        num_required = num - len(collect)
+                else:
+                    num_required = num
+
+                assert num is not None
+
+                while num_required:
+                    rows = _manyrows(num_required)
+                    if not rows:
+                        break
+
+                    collect.extend(
+                        filterrows(make_row, rows, strategy, uniques)
+                    )
+                    num_required = num - len(collect)
+
+                if post_creational_filter:
+                    collect = [post_creational_filter(row) for row in collect]
+                return collect
+
+        else:
+
+            def manyrows(
+                self: ResultInternal[_R], num: Optional[int]
+            ) -> List[_R]:
+                if num is None:
+                    real_result = (
+                        self._real_result
+                        if self._real_result
+                        else cast("Result[Any]", self)
+                    )
+                    num = real_result._yield_per
+
+                rows: List[_InterimRowType[Any]] = self._fetchmany_impl(num)
+                if make_row:
+                    rows = [make_row(row) for row in rows]
+                if post_creational_filter:
+                    rows = [post_creational_filter(row) for row in rows]
+                return rows  # type: ignore
+
+        return manyrows
+
+    @overload
+    def _only_one_row(
+        self,
+        raise_for_second_row: bool,
+        raise_for_none: Literal[True],
+        scalar: bool,
+    ) -> _R: ...
+
+    @overload
+    def _only_one_row(
+        self,
+        raise_for_second_row: bool,
+        raise_for_none: bool,
+        scalar: bool,
+    ) -> Optional[_R]: ...
+
+    def _only_one_row(
+        self,
+        raise_for_second_row: bool,
+        raise_for_none: bool,
+        scalar: bool,
+    ) -> Optional[_R]:
+        onerow = self._fetchone_impl
+
+        row: Optional[_InterimRowType[Any]] = onerow(hard_close=True)
+        if row is None:
+            if raise_for_none:
+                raise exc.NoResultFound(
+                    "No row was found when one was required"
+                )
+            else:
+                return None
+
+        if scalar and self._source_supports_scalars:
+            self._generate_rows = False
+            make_row = None
+        else:
+            make_row = self._row_getter
+
+        try:
+            row = make_row(row) if make_row else row
+        except:
+            self._soft_close(hard=True)
+            raise
+
+        if raise_for_second_row:
+            if self._unique_filter_state:
+                # for no second row but uniqueness, need to essentially
+                # consume the entire result :(
+                uniques, strategy = self._unique_strategy
+
+                existing_row_hash = strategy(row) if strategy else row
+
+                while True:
+                    next_row: Any = onerow(hard_close=True)
+                    if next_row is None:
+                        next_row = _NO_ROW
+                        break
+
+                    try:
+                        next_row = make_row(next_row) if make_row else next_row
+
+                        if strategy:
+                            assert next_row is not _NO_ROW
+                            if existing_row_hash == strategy(next_row):
+                                continue
+                        elif row == next_row:
+                            continue
+                        # here, we have a row and it's different
+                        break
+                    except:
+                        self._soft_close(hard=True)
+                        raise
+            else:
+                next_row = onerow(hard_close=True)
+                if next_row is None:
+                    next_row = _NO_ROW
+
+            if next_row is not _NO_ROW:
+                self._soft_close(hard=True)
+                raise exc.MultipleResultsFound(
+                    "Multiple rows were found when exactly one was required"
+                    if raise_for_none
+                    else "Multiple rows were found when one or none "
+                    "was required"
+                )
+        else:
+            next_row = _NO_ROW
+            # if we checked for second row then that would have
+            # closed us :)
+            self._soft_close(hard=True)
+
+        if not scalar:
+            post_creational_filter = self._post_creational_filter
+            if post_creational_filter:
+                row = post_creational_filter(row)
+
+        if scalar and make_row:
+            return row[0]  # type: ignore
+        else:
+            return row  # type: ignore
+
+    def _iter_impl(self) -> Iterator[_R]:
+        return self._iterator_getter(self)
+
+    def _next_impl(self) -> _R:
+        row = self._onerow_getter(self)
+        if row is _NO_ROW:
+            raise StopIteration()
+        else:
+            return row
+
+    @_generative
+    def _column_slices(self, indexes: Sequence[_KeyIndexType]) -> Self:
+        real_result = (
+            self._real_result
+            if self._real_result
+            else cast("Result[Any]", self)
+        )
+
+        if not real_result._source_supports_scalars or len(indexes) != 1:
+            self._metadata = self._metadata._reduce(indexes)
+
+        assert self._generate_rows
+
+        return self
+
+    @HasMemoized.memoized_attribute
+    def _unique_strategy(self) -> _UniqueFilterStateType:
+        assert self._unique_filter_state is not None
+        uniques, strategy = self._unique_filter_state
+
+        real_result = (
+            self._real_result
+            if self._real_result is not None
+            else cast("Result[Any]", self)
+        )
+
+        if not strategy and self._metadata._unique_filters:
+            if (
+                real_result._source_supports_scalars
+                and not self._generate_rows
+            ):
+                strategy = self._metadata._unique_filters[0]
+            else:
+                filters = self._metadata._unique_filters
+                if self._metadata._tuplefilter:
+                    filters = self._metadata._tuplefilter(filters)
+
+                strategy = operator.methodcaller("_filter_on_values", filters)
+        return uniques, strategy
+
+
+class _WithKeys:
+    __slots__ = ()
+
+    _metadata: ResultMetaData
+
+    # used mainly to share documentation on the keys method.
+    def keys(self) -> RMKeyView:
+        """Return an iterable view which yields the string keys that would
+        be represented by each :class:`_engine.Row`.
+
+        The keys can represent the labels of the columns returned by a core
+        statement or the names of the orm classes returned by an orm
+        execution.
+
+        The view also can be tested for key containment using the Python
+        ``in`` operator, which will test both for the string keys represented
+        in the view, as well as for alternate keys such as column objects.
+
+        .. versionchanged:: 1.4 a key view object is returned rather than a
+           plain list.
+
+
+        """
+        return self._metadata.keys
+
+
+class Result(_WithKeys, ResultInternal[Row[_TP]]):
+    """Represent a set of database results.
+
+    .. versionadded:: 1.4  The :class:`_engine.Result` object provides a
+       completely updated usage model and calling facade for SQLAlchemy
+       Core and SQLAlchemy ORM.   In Core, it forms the basis of the
+       :class:`_engine.CursorResult` object which replaces the previous
+       :class:`_engine.ResultProxy` interface.   When using the ORM, a
+       higher level object called :class:`_engine.ChunkedIteratorResult`
+       is normally used.
+
+    .. note:: In SQLAlchemy 1.4 and above, this object is
+       used for ORM results returned by :meth:`_orm.Session.execute`, which can
+       yield instances of ORM mapped objects either individually or within
+       tuple-like rows. Note that the :class:`_engine.Result` object does not
+       deduplicate instances or rows automatically as is the case with the
+       legacy :class:`_orm.Query` object. For in-Python de-duplication of
+       instances or rows, use the :meth:`_engine.Result.unique` modifier
+       method.
+
+    .. seealso::
+
+        :ref:`tutorial_fetching_rows` - in the :doc:`/tutorial/index`
+
+    """
+
+    __slots__ = ("_metadata", "__dict__")
+
+    _row_logging_fn: Optional[Callable[[Row[Any]], Row[Any]]] = None
+
+    _source_supports_scalars: bool = False
+
+    _yield_per: Optional[int] = None
+
+    _attributes: util.immutabledict[Any, Any] = util.immutabledict()
+
+    def __init__(self, cursor_metadata: ResultMetaData):
+        self._metadata = cursor_metadata
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        self.close()
+
+    def close(self) -> None:
+        """close this :class:`_engine.Result`.
+
+        The behavior of this method is implementation specific, and is
+        not implemented by default.    The method should generally end
+        the resources in use by the result object and also cause any
+        subsequent iteration or row fetching to raise
+        :class:`.ResourceClosedError`.
+
+        .. versionadded:: 1.4.27 - ``.close()`` was previously not generally
+           available for all :class:`_engine.Result` classes, instead only
+           being available on the :class:`_engine.CursorResult` returned for
+           Core statement executions. As most other result objects, namely the
+           ones used by the ORM, are proxying a :class:`_engine.CursorResult`
+           in any case, this allows the underlying cursor result to be closed
+           from the outside facade for the case when the ORM query is using
+           the ``yield_per`` execution option where it does not immediately
+           exhaust and autoclose the database cursor.
+
+        """
+        self._soft_close(hard=True)
+
+    @property
+    def _soft_closed(self) -> bool:
+        raise NotImplementedError()
+
+    @property
+    def closed(self) -> bool:
+        """return ``True`` if this :class:`_engine.Result` reports .closed
+
+        .. versionadded:: 1.4.43
+
+        """
+        raise NotImplementedError()
+
+    @_generative
+    def yield_per(self, num: int) -> Self:
+        """Configure the row-fetching strategy to fetch ``num`` rows at a time.
+
+        This impacts the underlying behavior of the result when iterating over
+        the result object, or otherwise making use of  methods such as
+        :meth:`_engine.Result.fetchone` that return one row at a time.   Data
+        from the underlying cursor or other data source will be buffered up to
+        this many rows in memory, and the buffered collection will then be
+        yielded out one row at a time or as many rows are requested. Each time
+        the buffer clears, it will be refreshed to this many rows or as many
+        rows remain if fewer remain.
+
+        The :meth:`_engine.Result.yield_per` method is generally used in
+        conjunction with the
+        :paramref:`_engine.Connection.execution_options.stream_results`
+        execution option, which will allow the database dialect in use to make
+        use of a server side cursor, if the DBAPI supports a specific "server
+        side cursor" mode separate from its default mode of operation.
+
+        .. tip::
+
+            Consider using the
+            :paramref:`_engine.Connection.execution_options.yield_per`
+            execution option, which will simultaneously set
+            :paramref:`_engine.Connection.execution_options.stream_results`
+            to ensure the use of server side cursors, as well as automatically
+            invoke the :meth:`_engine.Result.yield_per` method to establish
+            a fixed row buffer size at once.
+
+            The :paramref:`_engine.Connection.execution_options.yield_per`
+            execution option is available for ORM operations, with
+            :class:`_orm.Session`-oriented use described at
+            :ref:`orm_queryguide_yield_per`. The Core-only version which works
+            with :class:`_engine.Connection` is new as of SQLAlchemy 1.4.40.
+
+        .. versionadded:: 1.4
+
+        :param num: number of rows to fetch each time the buffer is refilled.
+         If set to a value below 1, fetches all rows for the next buffer.
+
+        .. seealso::
+
+            :ref:`engine_stream_results` - describes Core behavior for
+            :meth:`_engine.Result.yield_per`
+
+            :ref:`orm_queryguide_yield_per` - in the :ref:`queryguide_toplevel`
+
+        """
+        self._yield_per = num
+        return self
+
+    @_generative
+    def unique(self, strategy: Optional[_UniqueFilterType] = None) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_engine.Result`.
+
+        When this filter is applied with no arguments, the rows or objects
+        returned will filtered such that each row is returned uniquely. The
+        algorithm used to determine this uniqueness is by default the Python
+        hashing identity of the whole tuple.   In some cases a specialized
+        per-entity hashing scheme may be used, such as when using the ORM, a
+        scheme is applied which  works against the primary key identity of
+        returned objects.
+
+        The unique filter is applied **after all other filters**, which means
+        if the columns returned have been refined using a method such as the
+        :meth:`_engine.Result.columns` or :meth:`_engine.Result.scalars`
+        method, the uniquing is applied to **only the column or columns
+        returned**.   This occurs regardless of the order in which these
+        methods have been called upon the :class:`_engine.Result` object.
+
+        The unique filter also changes the calculus used for methods like
+        :meth:`_engine.Result.fetchmany` and :meth:`_engine.Result.partitions`.
+        When using :meth:`_engine.Result.unique`, these methods will continue
+        to yield the number of rows or objects requested, after uniquing
+        has been applied.  However, this necessarily impacts the buffering
+        behavior of the underlying cursor or datasource, such that multiple
+        underlying calls to ``cursor.fetchmany()`` may be necessary in order
+        to accumulate enough objects in order to provide a unique collection
+        of the requested size.
+
+        :param strategy: a callable that will be applied to rows or objects
+         being iterated, which should return an object that represents the
+         unique value of the row.   A Python ``set()`` is used to store
+         these identities.   If not passed, a default uniqueness strategy
+         is used which may have been assembled by the source of this
+         :class:`_engine.Result` object.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    def columns(self, *col_expressions: _KeyIndexType) -> Self:
+        r"""Establish the columns that should be returned in each row.
+
+        This method may be used to limit the columns returned as well
+        as to reorder them.   The given list of expressions are normally
+        a series of integers or string key names.   They may also be
+        appropriate :class:`.ColumnElement` objects which correspond to
+        a given statement construct.
+
+        .. versionchanged:: 2.0  Due to a bug in 1.4, the
+           :meth:`_engine.Result.columns` method had an incorrect behavior
+           where calling upon the method with just one index would cause the
+           :class:`_engine.Result` object to yield scalar values rather than
+           :class:`_engine.Row` objects.   In version 2.0, this behavior
+           has been corrected such that calling upon
+           :meth:`_engine.Result.columns` with a single index will
+           produce a :class:`_engine.Result` object that continues
+           to yield :class:`_engine.Row` objects, which include
+           only a single column.
+
+        E.g.::
+
+            statement = select(table.c.x, table.c.y, table.c.z)
+            result = connection.execute(statement)
+
+            for z, y in result.columns("z", "y"):
+                ...
+
+        Example of using the column objects from the statement itself::
+
+            for z, y in result.columns(
+                statement.selected_columns.c.z, statement.selected_columns.c.y
+            ):
+                ...
+
+        .. versionadded:: 1.4
+
+        :param \*col_expressions: indicates columns to be returned.  Elements
+         may be integer row indexes, string column names, or appropriate
+         :class:`.ColumnElement` objects corresponding to a select construct.
+
+        :return: this :class:`_engine.Result` object with the modifications
+         given.
+
+        """
+        return self._column_slices(col_expressions)
+
+    @overload
+    def scalars(self: Result[Tuple[_T]]) -> ScalarResult[_T]: ...
+
+    @overload
+    def scalars(
+        self: Result[Tuple[_T]], index: Literal[0]
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    def scalars(self, index: _KeyIndexType = 0) -> ScalarResult[Any]: ...
+
+    def scalars(self, index: _KeyIndexType = 0) -> ScalarResult[Any]:
+        """Return a :class:`_engine.ScalarResult` filtering object which
+        will return single elements rather than :class:`_row.Row` objects.
+
+        E.g.::
+
+            >>> result = conn.execute(text("select int_id from table"))
+            >>> result.scalars().all()
+            [1, 2, 3]
+
+        When results are fetched from the :class:`_engine.ScalarResult`
+        filtering object, the single column-row that would be returned by the
+        :class:`_engine.Result` is instead returned as the column's value.
+
+        .. versionadded:: 1.4
+
+        :param index: integer or row key indicating the column to be fetched
+         from each row, defaults to ``0`` indicating the first column.
+
+        :return: a new :class:`_engine.ScalarResult` filtering object referring
+         to this :class:`_engine.Result` object.
+
+        """
+        return ScalarResult(self, index)
+
+    def _getter(
+        self, key: _KeyIndexType, raiseerr: bool = True
+    ) -> Optional[Callable[[Row[Any]], Any]]:
+        """return a callable that will retrieve the given key from a
+        :class:`_engine.Row`.
+
+        """
+        if self._source_supports_scalars:
+            raise NotImplementedError(
+                "can't use this function in 'only scalars' mode"
+            )
+        return self._metadata._getter(key, raiseerr)
+
+    def _tuple_getter(self, keys: Sequence[_KeyIndexType]) -> _TupleGetterType:
+        """return a callable that will retrieve the given keys from a
+        :class:`_engine.Row`.
+
+        """
+        if self._source_supports_scalars:
+            raise NotImplementedError(
+                "can't use this function in 'only scalars' mode"
+            )
+        return self._metadata._row_as_tuple_getter(keys)
+
+    def mappings(self) -> MappingResult:
+        """Apply a mappings filter to returned rows, returning an instance of
+        :class:`_engine.MappingResult`.
+
+        When this filter is applied, fetching rows will return
+        :class:`_engine.RowMapping` objects instead of :class:`_engine.Row`
+        objects.
+
+        .. versionadded:: 1.4
+
+        :return: a new :class:`_engine.MappingResult` filtering object
+         referring to this :class:`_engine.Result` object.
+
+        """
+
+        return MappingResult(self)
+
+    @property
+    def t(self) -> TupleResult[_TP]:
+        """Apply a "typed tuple" typing filter to returned rows.
+
+        The :attr:`_engine.Result.t` attribute is a synonym for
+        calling the :meth:`_engine.Result.tuples` method.
+
+        .. versionadded:: 2.0
+
+        """
+        return self  # type: ignore
+
+    def tuples(self) -> TupleResult[_TP]:
+        """Apply a "typed tuple" typing filter to returned rows.
+
+        This method returns the same :class:`_engine.Result` object
+        at runtime,
+        however annotates as returning a :class:`_engine.TupleResult` object
+        that will indicate to :pep:`484` typing tools that plain typed
+        ``Tuple`` instances are returned rather than rows.  This allows
+        tuple unpacking and ``__getitem__`` access of :class:`_engine.Row`
+        objects to by typed, for those cases where the statement invoked
+        itself included typing information.
+
+        .. versionadded:: 2.0
+
+        :return: the :class:`_engine.TupleResult` type at typing time.
+
+        .. seealso::
+
+            :attr:`_engine.Result.t` - shorter synonym
+
+            :attr:`_engine.Row._t` - :class:`_engine.Row` version
+
+        """
+
+        return self  # type: ignore
+
+    def _raw_row_iterator(self) -> Iterator[_RowData]:
+        """Return a safe iterator that yields raw row data.
+
+        This is used by the :meth:`_engine.Result.merge` method
+        to merge multiple compatible results together.
+
+        """
+        raise NotImplementedError()
+
+    def __iter__(self) -> Iterator[Row[_TP]]:
+        return self._iter_impl()
+
+    def __next__(self) -> Row[_TP]:
+        return self._next_impl()
+
+    def partitions(
+        self, size: Optional[int] = None
+    ) -> Iterator[Sequence[Row[_TP]]]:
+        """Iterate through sub-lists of rows of the size given.
+
+        Each list will be of the size given, excluding the last list to
+        be yielded, which may have a small number of rows.  No empty
+        lists will be yielded.
+
+        The result object is automatically closed when the iterator
+        is fully consumed.
+
+        Note that the backend driver will usually buffer the entire result
+        ahead of time unless the
+        :paramref:`.Connection.execution_options.stream_results` execution
+        option is used indicating that the driver should not pre-buffer
+        results, if possible.   Not all drivers support this option and
+        the option is silently ignored for those who do not.
+
+        When using the ORM, the :meth:`_engine.Result.partitions` method
+        is typically more effective from a memory perspective when it is
+        combined with use of the
+        :ref:`yield_per execution option <orm_queryguide_yield_per>`,
+        which instructs both the DBAPI driver to use server side cursors,
+        if available, as well as instructs the ORM loading internals to only
+        build a certain amount of ORM objects from a result at a time before
+        yielding them out.
+
+        .. versionadded:: 1.4
+
+        :param size: indicate the maximum number of rows to be present
+         in each list yielded.  If None, makes use of the value set by
+         the :meth:`_engine.Result.yield_per`, method, if it were called,
+         or the :paramref:`_engine.Connection.execution_options.yield_per`
+         execution option, which is equivalent in this regard.  If
+         yield_per weren't set, it makes use of the
+         :meth:`_engine.Result.fetchmany` default, which may be backend
+         specific and not well defined.
+
+        :return: iterator of lists
+
+        .. seealso::
+
+            :ref:`engine_stream_results`
+
+            :ref:`orm_queryguide_yield_per` - in the :ref:`queryguide_toplevel`
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = getter(self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    def fetchall(self) -> Sequence[Row[_TP]]:
+        """A synonym for the :meth:`_engine.Result.all` method."""
+
+        return self._allrows()
+
+    def fetchone(self) -> Optional[Row[_TP]]:
+        """Fetch one row.
+
+        When all rows are exhausted, returns None.
+
+        This method is provided for backwards compatibility with
+        SQLAlchemy 1.x.x.
+
+        To fetch the first row of a result only, use the
+        :meth:`_engine.Result.first` method.  To iterate through all
+        rows, iterate the :class:`_engine.Result` object directly.
+
+        :return: a :class:`_engine.Row` object if no filters are applied,
+         or ``None`` if no rows remain.
+
+        """
+        row = self._onerow_getter(self)
+        if row is _NO_ROW:
+            return None
+        else:
+            return row
+
+    def fetchmany(self, size: Optional[int] = None) -> Sequence[Row[_TP]]:
+        """Fetch many rows.
+
+        When all rows are exhausted, returns an empty sequence.
+
+        This method is provided for backwards compatibility with
+        SQLAlchemy 1.x.x.
+
+        To fetch rows in groups, use the :meth:`_engine.Result.partitions`
+        method.
+
+        :return: a sequence of :class:`_engine.Row` objects.
+
+        .. seealso::
+
+            :meth:`_engine.Result.partitions`
+
+        """
+
+        return self._manyrow_getter(self, size)
+
+    def all(self) -> Sequence[Row[_TP]]:
+        """Return all rows in a sequence.
+
+        Closes the result set after invocation.   Subsequent invocations
+        will return an empty sequence.
+
+        .. versionadded:: 1.4
+
+        :return: a sequence of :class:`_engine.Row` objects.
+
+        .. seealso::
+
+            :ref:`engine_stream_results` - How to stream a large result set
+            without loading it completely in python.
+
+        """
+
+        return self._allrows()
+
+    def first(self) -> Optional[Row[_TP]]:
+        """Fetch the first row or ``None`` if no row is present.
+
+        Closes the result set and discards remaining rows.
+
+        .. note::  This method returns one **row**, e.g. tuple, by default.
+           To return exactly one single scalar value, that is, the first
+           column of the first row, use the
+           :meth:`_engine.Result.scalar` method,
+           or combine :meth:`_engine.Result.scalars` and
+           :meth:`_engine.Result.first`.
+
+           Additionally, in contrast to the behavior of the legacy  ORM
+           :meth:`_orm.Query.first` method, **no limit is applied** to the
+           SQL query which was invoked to produce this
+           :class:`_engine.Result`;
+           for a DBAPI driver that buffers results in memory before yielding
+           rows, all rows will be sent to the Python process and all but
+           the first row will be discarded.
+
+           .. seealso::
+
+                :ref:`migration_20_unify_select`
+
+        :return: a :class:`_engine.Row` object, or None
+         if no rows remain.
+
+        .. seealso::
+
+            :meth:`_engine.Result.scalar`
+
+            :meth:`_engine.Result.one`
+
+        """
+
+        return self._only_one_row(
+            raise_for_second_row=False, raise_for_none=False, scalar=False
+        )
+
+    def one_or_none(self) -> Optional[Row[_TP]]:
+        """Return at most one result or raise an exception.
+
+        Returns ``None`` if the result has no rows.
+        Raises :class:`.MultipleResultsFound`
+        if multiple rows are returned.
+
+        .. versionadded:: 1.4
+
+        :return: The first :class:`_engine.Row` or ``None`` if no row
+         is available.
+
+        :raises: :class:`.MultipleResultsFound`
+
+        .. seealso::
+
+            :meth:`_engine.Result.first`
+
+            :meth:`_engine.Result.one`
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=False, scalar=False
+        )
+
+    @overload
+    def scalar_one(self: Result[Tuple[_T]]) -> _T: ...
+
+    @overload
+    def scalar_one(self) -> Any: ...
+
+    def scalar_one(self) -> Any:
+        """Return exactly one scalar result or raise an exception.
+
+        This is equivalent to calling :meth:`_engine.Result.scalars` and
+        then :meth:`_engine.ScalarResult.one`.
+
+        .. seealso::
+
+            :meth:`_engine.ScalarResult.one`
+
+            :meth:`_engine.Result.scalars`
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=True, scalar=True
+        )
+
+    @overload
+    def scalar_one_or_none(self: Result[Tuple[_T]]) -> Optional[_T]: ...
+
+    @overload
+    def scalar_one_or_none(self) -> Optional[Any]: ...
+
+    def scalar_one_or_none(self) -> Optional[Any]:
+        """Return exactly one scalar result or ``None``.
+
+        This is equivalent to calling :meth:`_engine.Result.scalars` and
+        then :meth:`_engine.ScalarResult.one_or_none`.
+
+        .. seealso::
+
+            :meth:`_engine.ScalarResult.one_or_none`
+
+            :meth:`_engine.Result.scalars`
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=False, scalar=True
+        )
+
+    def one(self) -> Row[_TP]:
+        """Return exactly one row or raise an exception.
+
+        Raises :class:`.NoResultFound` if the result returns no
+        rows, or :class:`.MultipleResultsFound` if multiple rows
+        would be returned.
+
+        .. note::  This method returns one **row**, e.g. tuple, by default.
+           To return exactly one single scalar value, that is, the first
+           column of the first row, use the
+           :meth:`_engine.Result.scalar_one` method, or combine
+           :meth:`_engine.Result.scalars` and
+           :meth:`_engine.Result.one`.
+
+        .. versionadded:: 1.4
+
+        :return: The first :class:`_engine.Row`.
+
+        :raises: :class:`.MultipleResultsFound`, :class:`.NoResultFound`
+
+        .. seealso::
+
+            :meth:`_engine.Result.first`
+
+            :meth:`_engine.Result.one_or_none`
+
+            :meth:`_engine.Result.scalar_one`
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=True, scalar=False
+        )
+
+    @overload
+    def scalar(self: Result[Tuple[_T]]) -> Optional[_T]: ...
+
+    @overload
+    def scalar(self) -> Any: ...
+
+    def scalar(self) -> Any:
+        """Fetch the first column of the first row, and close the result set.
+
+        Returns ``None`` if there are no rows to fetch.
+
+        No validation is performed to test if additional rows remain.
+
+        After calling this method, the object is fully closed,
+        e.g. the :meth:`_engine.CursorResult.close`
+        method will have been called.
+
+        :return: a Python scalar value, or ``None`` if no rows remain.
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=False, raise_for_none=False, scalar=True
+        )
+
+    def freeze(self) -> FrozenResult[_TP]:
+        """Return a callable object that will produce copies of this
+        :class:`_engine.Result` when invoked.
+
+        The callable object returned is an instance of
+        :class:`_engine.FrozenResult`.
+
+        This is used for result set caching.  The method must be called
+        on the result when it has been unconsumed, and calling the method
+        will consume the result fully.   When the :class:`_engine.FrozenResult`
+        is retrieved from a cache, it can be called any number of times where
+        it will produce a new :class:`_engine.Result` object each time
+        against its stored set of rows.
+
+        .. seealso::
+
+            :ref:`do_orm_execute_re_executing` - example usage within the
+            ORM to implement a result-set cache.
+
+        """
+
+        return FrozenResult(self)
+
+    def merge(self, *others: Result[Any]) -> MergedResult[_TP]:
+        """Merge this :class:`_engine.Result` with other compatible result
+        objects.
+
+        The object returned is an instance of :class:`_engine.MergedResult`,
+        which will be composed of iterators from the given result
+        objects.
+
+        The new result will use the metadata from this result object.
+        The subsequent result objects must be against an identical
+        set of result / cursor metadata, otherwise the behavior is
+        undefined.
+
+        """
+        return MergedResult(self._metadata, (self,) + others)
+
+
+class FilterResult(ResultInternal[_R]):
+    """A wrapper for a :class:`_engine.Result` that returns objects other than
+    :class:`_engine.Row` objects, such as dictionaries or scalar objects.
+
+    :class:`_engine.FilterResult` is the common base for additional result
+    APIs including :class:`_engine.MappingResult`,
+    :class:`_engine.ScalarResult` and :class:`_engine.AsyncResult`.
+
+    """
+
+    __slots__ = (
+        "_real_result",
+        "_post_creational_filter",
+        "_metadata",
+        "_unique_filter_state",
+        "__dict__",
+    )
+
+    _post_creational_filter: Optional[Callable[[Any], Any]]
+
+    _real_result: Result[Any]
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        self._real_result.__exit__(type_, value, traceback)
+
+    @_generative
+    def yield_per(self, num: int) -> Self:
+        """Configure the row-fetching strategy to fetch ``num`` rows at a time.
+
+        The :meth:`_engine.FilterResult.yield_per` method is a pass through
+        to the :meth:`_engine.Result.yield_per` method.  See that method's
+        documentation for usage notes.
+
+        .. versionadded:: 1.4.40 - added :meth:`_engine.FilterResult.yield_per`
+           so that the method is available on all result set implementations
+
+        .. seealso::
+
+            :ref:`engine_stream_results` - describes Core behavior for
+            :meth:`_engine.Result.yield_per`
+
+            :ref:`orm_queryguide_yield_per` - in the :ref:`queryguide_toplevel`
+
+        """
+        self._real_result = self._real_result.yield_per(num)
+        return self
+
+    def _soft_close(self, hard: bool = False) -> None:
+        self._real_result._soft_close(hard=hard)
+
+    @property
+    def _soft_closed(self) -> bool:
+        return self._real_result._soft_closed
+
+    @property
+    def closed(self) -> bool:
+        """Return ``True`` if the underlying :class:`_engine.Result` reports
+        closed
+
+        .. versionadded:: 1.4.43
+
+        """
+        return self._real_result.closed
+
+    def close(self) -> None:
+        """Close this :class:`_engine.FilterResult`.
+
+        .. versionadded:: 1.4.43
+
+        """
+        self._real_result.close()
+
+    @property
+    def _attributes(self) -> Dict[Any, Any]:
+        return self._real_result._attributes
+
+    def _fetchiter_impl(self) -> Iterator[_InterimRowType[Row[Any]]]:
+        return self._real_result._fetchiter_impl()
+
+    def _fetchone_impl(
+        self, hard_close: bool = False
+    ) -> Optional[_InterimRowType[Row[Any]]]:
+        return self._real_result._fetchone_impl(hard_close=hard_close)
+
+    def _fetchall_impl(self) -> List[_InterimRowType[Row[Any]]]:
+        return self._real_result._fetchall_impl()
+
+    def _fetchmany_impl(
+        self, size: Optional[int] = None
+    ) -> List[_InterimRowType[Row[Any]]]:
+        return self._real_result._fetchmany_impl(size=size)
+
+
+class ScalarResult(FilterResult[_R]):
+    """A wrapper for a :class:`_engine.Result` that returns scalar values
+    rather than :class:`_row.Row` values.
+
+    The :class:`_engine.ScalarResult` object is acquired by calling the
+    :meth:`_engine.Result.scalars` method.
+
+    A special limitation of :class:`_engine.ScalarResult` is that it has
+    no ``fetchone()`` method; since the semantics of ``fetchone()`` are that
+    the ``None`` value indicates no more results, this is not compatible
+    with :class:`_engine.ScalarResult` since there is no way to distinguish
+    between ``None`` as a row value versus ``None`` as an indicator.  Use
+    ``next(result)`` to receive values individually.
+
+    """
+
+    __slots__ = ()
+
+    _generate_rows = False
+
+    _post_creational_filter: Optional[Callable[[Any], Any]]
+
+    def __init__(self, real_result: Result[Any], index: _KeyIndexType):
+        self._real_result = real_result
+
+        if real_result._source_supports_scalars:
+            self._metadata = real_result._metadata
+            self._post_creational_filter = None
+        else:
+            self._metadata = real_result._metadata._reduce([index])
+            self._post_creational_filter = operator.itemgetter(0)
+
+        self._unique_filter_state = real_result._unique_filter_state
+
+    def unique(self, strategy: Optional[_UniqueFilterType] = None) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_engine.ScalarResult`.
+
+        See :meth:`_engine.Result.unique` for usage details.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    def partitions(self, size: Optional[int] = None) -> Iterator[Sequence[_R]]:
+        """Iterate through sub-lists of elements of the size given.
+
+        Equivalent to :meth:`_engine.Result.partitions` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = getter(self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    def fetchall(self) -> Sequence[_R]:
+        """A synonym for the :meth:`_engine.ScalarResult.all` method."""
+
+        return self._allrows()
+
+    def fetchmany(self, size: Optional[int] = None) -> Sequence[_R]:
+        """Fetch many objects.
+
+        Equivalent to :meth:`_engine.Result.fetchmany` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return self._manyrow_getter(self, size)
+
+    def all(self) -> Sequence[_R]:
+        """Return all scalar values in a sequence.
+
+        Equivalent to :meth:`_engine.Result.all` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return self._allrows()
+
+    def __iter__(self) -> Iterator[_R]:
+        return self._iter_impl()
+
+    def __next__(self) -> _R:
+        return self._next_impl()
+
+    def first(self) -> Optional[_R]:
+        """Fetch the first object or ``None`` if no object is present.
+
+        Equivalent to :meth:`_engine.Result.first` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=False, raise_for_none=False, scalar=False
+        )
+
+    def one_or_none(self) -> Optional[_R]:
+        """Return at most one object or raise an exception.
+
+        Equivalent to :meth:`_engine.Result.one_or_none` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=False, scalar=False
+        )
+
+    def one(self) -> _R:
+        """Return exactly one object or raise an exception.
+
+        Equivalent to :meth:`_engine.Result.one` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=True, scalar=False
+        )
+
+
+class TupleResult(FilterResult[_R], util.TypingOnly):
+    """A :class:`_engine.Result` that's typed as returning plain
+    Python tuples instead of rows.
+
+    Since :class:`_engine.Row` acts like a tuple in every way already,
+    this class is a typing only class, regular :class:`_engine.Result` is
+    still used at runtime.
+
+    """
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        def partitions(
+            self, size: Optional[int] = None
+        ) -> Iterator[Sequence[_R]]:
+            """Iterate through sub-lists of elements of the size given.
+
+            Equivalent to :meth:`_engine.Result.partitions` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        def fetchone(self) -> Optional[_R]:
+            """Fetch one tuple.
+
+            Equivalent to :meth:`_engine.Result.fetchone` except that
+            tuple values, rather than :class:`_engine.Row`
+            objects, are returned.
+
+            """
+            ...
+
+        def fetchall(self) -> Sequence[_R]:
+            """A synonym for the :meth:`_engine.ScalarResult.all` method."""
+            ...
+
+        def fetchmany(self, size: Optional[int] = None) -> Sequence[_R]:
+            """Fetch many objects.
+
+            Equivalent to :meth:`_engine.Result.fetchmany` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        def all(self) -> Sequence[_R]:  # noqa: A001
+            """Return all scalar values in a sequence.
+
+            Equivalent to :meth:`_engine.Result.all` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        def __iter__(self) -> Iterator[_R]: ...
+
+        def __next__(self) -> _R: ...
+
+        def first(self) -> Optional[_R]:
+            """Fetch the first object or ``None`` if no object is present.
+
+            Equivalent to :meth:`_engine.Result.first` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+
+            """
+            ...
+
+        def one_or_none(self) -> Optional[_R]:
+            """Return at most one object or raise an exception.
+
+            Equivalent to :meth:`_engine.Result.one_or_none` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        def one(self) -> _R:
+            """Return exactly one object or raise an exception.
+
+            Equivalent to :meth:`_engine.Result.one` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        @overload
+        def scalar_one(self: TupleResult[Tuple[_T]]) -> _T: ...
+
+        @overload
+        def scalar_one(self) -> Any: ...
+
+        def scalar_one(self) -> Any:
+            """Return exactly one scalar result or raise an exception.
+
+            This is equivalent to calling :meth:`_engine.Result.scalars`
+            and then :meth:`_engine.ScalarResult.one`.
+
+            .. seealso::
+
+                :meth:`_engine.ScalarResult.one`
+
+                :meth:`_engine.Result.scalars`
+
+            """
+            ...
+
+        @overload
+        def scalar_one_or_none(
+            self: TupleResult[Tuple[_T]],
+        ) -> Optional[_T]: ...
+
+        @overload
+        def scalar_one_or_none(self) -> Optional[Any]: ...
+
+        def scalar_one_or_none(self) -> Optional[Any]:
+            """Return exactly one or no scalar result.
+
+            This is equivalent to calling :meth:`_engine.Result.scalars`
+            and then :meth:`_engine.ScalarResult.one_or_none`.
+
+            .. seealso::
+
+                :meth:`_engine.ScalarResult.one_or_none`
+
+                :meth:`_engine.Result.scalars`
+
+            """
+            ...
+
+        @overload
+        def scalar(self: TupleResult[Tuple[_T]]) -> Optional[_T]: ...
+
+        @overload
+        def scalar(self) -> Any: ...
+
+        def scalar(self) -> Any:
+            """Fetch the first column of the first row, and close the result
+            set.
+
+            Returns ``None`` if there are no rows to fetch.
+
+            No validation is performed to test if additional rows remain.
+
+            After calling this method, the object is fully closed,
+            e.g. the :meth:`_engine.CursorResult.close`
+            method will have been called.
+
+            :return: a Python scalar value , or ``None`` if no rows remain.
+
+            """
+            ...
+
+
+class MappingResult(_WithKeys, FilterResult[RowMapping]):
+    """A wrapper for a :class:`_engine.Result` that returns dictionary values
+    rather than :class:`_engine.Row` values.
+
+    The :class:`_engine.MappingResult` object is acquired by calling the
+    :meth:`_engine.Result.mappings` method.
+
+    """
+
+    __slots__ = ()
+
+    _generate_rows = True
+
+    _post_creational_filter = operator.attrgetter("_mapping")
+
+    def __init__(self, result: Result[Any]):
+        self._real_result = result
+        self._unique_filter_state = result._unique_filter_state
+        self._metadata = result._metadata
+        if result._source_supports_scalars:
+            self._metadata = self._metadata._reduce([0])
+
+    def unique(self, strategy: Optional[_UniqueFilterType] = None) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_engine.MappingResult`.
+
+        See :meth:`_engine.Result.unique` for usage details.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    def columns(self, *col_expressions: _KeyIndexType) -> Self:
+        r"""Establish the columns that should be returned in each row."""
+        return self._column_slices(col_expressions)
+
+    def partitions(
+        self, size: Optional[int] = None
+    ) -> Iterator[Sequence[RowMapping]]:
+        """Iterate through sub-lists of elements of the size given.
+
+        Equivalent to :meth:`_engine.Result.partitions` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = getter(self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    def fetchall(self) -> Sequence[RowMapping]:
+        """A synonym for the :meth:`_engine.MappingResult.all` method."""
+
+        return self._allrows()
+
+    def fetchone(self) -> Optional[RowMapping]:
+        """Fetch one object.
+
+        Equivalent to :meth:`_engine.Result.fetchone` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        row = self._onerow_getter(self)
+        if row is _NO_ROW:
+            return None
+        else:
+            return row
+
+    def fetchmany(self, size: Optional[int] = None) -> Sequence[RowMapping]:
+        """Fetch many objects.
+
+        Equivalent to :meth:`_engine.Result.fetchmany` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        return self._manyrow_getter(self, size)
+
+    def all(self) -> Sequence[RowMapping]:
+        """Return all scalar values in a sequence.
+
+        Equivalent to :meth:`_engine.Result.all` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        return self._allrows()
+
+    def __iter__(self) -> Iterator[RowMapping]:
+        return self._iter_impl()
+
+    def __next__(self) -> RowMapping:
+        return self._next_impl()
+
+    def first(self) -> Optional[RowMapping]:
+        """Fetch the first object or ``None`` if no object is present.
+
+        Equivalent to :meth:`_engine.Result.first` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=False, raise_for_none=False, scalar=False
+        )
+
+    def one_or_none(self) -> Optional[RowMapping]:
+        """Return at most one object or raise an exception.
+
+        Equivalent to :meth:`_engine.Result.one_or_none` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=False, scalar=False
+        )
+
+    def one(self) -> RowMapping:
+        """Return exactly one object or raise an exception.
+
+        Equivalent to :meth:`_engine.Result.one` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+        return self._only_one_row(
+            raise_for_second_row=True, raise_for_none=True, scalar=False
+        )
+
+
+class FrozenResult(Generic[_TP]):
+    """Represents a :class:`_engine.Result` object in a "frozen" state suitable
+    for caching.
+
+    The :class:`_engine.FrozenResult` object is returned from the
+    :meth:`_engine.Result.freeze` method of any :class:`_engine.Result`
+    object.
+
+    A new iterable :class:`_engine.Result` object is generated from a fixed
+    set of data each time the :class:`_engine.FrozenResult` is invoked as
+    a callable::
+
+
+        result = connection.execute(query)
+
+        frozen = result.freeze()
+
+        unfrozen_result_one = frozen()
+
+        for row in unfrozen_result_one:
+            print(row)
+
+        unfrozen_result_two = frozen()
+        rows = unfrozen_result_two.all()
+
+        # ... etc
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`do_orm_execute_re_executing` - example usage within the
+        ORM to implement a result-set cache.
+
+        :func:`_orm.loading.merge_frozen_result` - ORM function to merge
+        a frozen result back into a :class:`_orm.Session`.
+
+    """
+
+    data: Sequence[Any]
+
+    def __init__(self, result: Result[_TP]):
+        self.metadata = result._metadata._for_freeze()
+        self._source_supports_scalars = result._source_supports_scalars
+        self._attributes = result._attributes
+
+        if self._source_supports_scalars:
+            self.data = list(result._raw_row_iterator())
+        else:
+            self.data = result.fetchall()
+
+    def rewrite_rows(self) -> Sequence[Sequence[Any]]:
+        if self._source_supports_scalars:
+            return [[elem] for elem in self.data]
+        else:
+            return [list(row) for row in self.data]
+
+    def with_new_rows(
+        self, tuple_data: Sequence[Row[_TP]]
+    ) -> FrozenResult[_TP]:
+        fr = FrozenResult.__new__(FrozenResult)
+        fr.metadata = self.metadata
+        fr._attributes = self._attributes
+        fr._source_supports_scalars = self._source_supports_scalars
+
+        if self._source_supports_scalars:
+            fr.data = [d[0] for d in tuple_data]
+        else:
+            fr.data = tuple_data
+        return fr
+
+    def __call__(self) -> Result[_TP]:
+        result: IteratorResult[_TP] = IteratorResult(
+            self.metadata, iter(self.data)
+        )
+        result._attributes = self._attributes
+        result._source_supports_scalars = self._source_supports_scalars
+        return result
+
+
+class IteratorResult(Result[_TP]):
+    """A :class:`_engine.Result` that gets data from a Python iterator of
+    :class:`_engine.Row` objects or similar row-like data.
+
+    .. versionadded:: 1.4
+
+    """
+
+    _hard_closed = False
+    _soft_closed = False
+
+    def __init__(
+        self,
+        cursor_metadata: ResultMetaData,
+        iterator: Iterator[_InterimSupportsScalarsRowType],
+        raw: Optional[Result[Any]] = None,
+        _source_supports_scalars: bool = False,
+    ):
+        self._metadata = cursor_metadata
+        self.iterator = iterator
+        self.raw = raw
+        self._source_supports_scalars = _source_supports_scalars
+
+    @property
+    def closed(self) -> bool:
+        """Return ``True`` if this :class:`_engine.IteratorResult` has
+        been closed
+
+        .. versionadded:: 1.4.43
+
+        """
+        return self._hard_closed
+
+    def _soft_close(self, hard: bool = False, **kw: Any) -> None:
+        if hard:
+            self._hard_closed = True
+        if self.raw is not None:
+            self.raw._soft_close(hard=hard, **kw)
+        self.iterator = iter([])
+        self._reset_memoizations()
+        self._soft_closed = True
+
+    def _raise_hard_closed(self) -> NoReturn:
+        raise exc.ResourceClosedError("This result object is closed.")
+
+    def _raw_row_iterator(self) -> Iterator[_RowData]:
+        return self.iterator
+
+    def _fetchiter_impl(self) -> Iterator[_InterimSupportsScalarsRowType]:
+        if self._hard_closed:
+            self._raise_hard_closed()
+        return self.iterator
+
+    def _fetchone_impl(
+        self, hard_close: bool = False
+    ) -> Optional[_InterimRowType[Row[Any]]]:
+        if self._hard_closed:
+            self._raise_hard_closed()
+
+        row = next(self.iterator, _NO_ROW)
+        if row is _NO_ROW:
+            self._soft_close(hard=hard_close)
+            return None
+        else:
+            return row
+
+    def _fetchall_impl(self) -> List[_InterimRowType[Row[Any]]]:
+        if self._hard_closed:
+            self._raise_hard_closed()
+        try:
+            return list(self.iterator)
+        finally:
+            self._soft_close()
+
+    def _fetchmany_impl(
+        self, size: Optional[int] = None
+    ) -> List[_InterimRowType[Row[Any]]]:
+        if self._hard_closed:
+            self._raise_hard_closed()
+
+        return list(itertools.islice(self.iterator, 0, size))
+
+
+def null_result() -> IteratorResult[Any]:
+    return IteratorResult(SimpleResultMetaData([]), iter([]))
+
+
+class ChunkedIteratorResult(IteratorResult[_TP]):
+    """An :class:`_engine.IteratorResult` that works from an
+    iterator-producing callable.
+
+    The given ``chunks`` argument is a function that is given a number of rows
+    to return in each chunk, or ``None`` for all rows.  The function should
+    then return an un-consumed iterator of lists, each list of the requested
+    size.
+
+    The function can be called at any time again, in which case it should
+    continue from the same result set but adjust the chunk size as given.
+
+    .. versionadded:: 1.4
+
+    """
+
+    def __init__(
+        self,
+        cursor_metadata: ResultMetaData,
+        chunks: Callable[
+            [Optional[int]], Iterator[Sequence[_InterimRowType[_R]]]
+        ],
+        source_supports_scalars: bool = False,
+        raw: Optional[Result[Any]] = None,
+        dynamic_yield_per: bool = False,
+    ):
+        self._metadata = cursor_metadata
+        self.chunks = chunks
+        self._source_supports_scalars = source_supports_scalars
+        self.raw = raw
+        self.iterator = itertools.chain.from_iterable(self.chunks(None))
+        self.dynamic_yield_per = dynamic_yield_per
+
+    @_generative
+    def yield_per(self, num: int) -> Self:
+        # TODO: this throws away the iterator which may be holding
+        # onto a chunk.   the yield_per cannot be changed once any
+        # rows have been fetched.   either find a way to enforce this,
+        # or we can't use itertools.chain and will instead have to
+        # keep track.
+
+        self._yield_per = num
+        self.iterator = itertools.chain.from_iterable(self.chunks(num))
+        return self
+
+    def _soft_close(self, hard: bool = False, **kw: Any) -> None:
+        super()._soft_close(hard=hard, **kw)
+        self.chunks = lambda size: []  # type: ignore
+
+    def _fetchmany_impl(
+        self, size: Optional[int] = None
+    ) -> List[_InterimRowType[Row[Any]]]:
+        if self.dynamic_yield_per:
+            self.iterator = itertools.chain.from_iterable(self.chunks(size))
+        return super()._fetchmany_impl(size=size)
+
+
+class MergedResult(IteratorResult[_TP]):
+    """A :class:`_engine.Result` that is merged from any number of
+    :class:`_engine.Result` objects.
+
+    Returned by the :meth:`_engine.Result.merge` method.
+
+    .. versionadded:: 1.4
+
+    """
+
+    closed = False
+    rowcount: Optional[int]
+
+    def __init__(
+        self, cursor_metadata: ResultMetaData, results: Sequence[Result[_TP]]
+    ):
+        self._results = results
+        super().__init__(
+            cursor_metadata,
+            itertools.chain.from_iterable(
+                r._raw_row_iterator() for r in results
+            ),
+        )
+
+        self._unique_filter_state = results[0]._unique_filter_state
+        self._yield_per = results[0]._yield_per
+
+        # going to try something w/ this in next rev
+        self._source_supports_scalars = results[0]._source_supports_scalars
+
+        self._attributes = self._attributes.merge_with(
+            *[r._attributes for r in results]
+        )
+
+    def _soft_close(self, hard: bool = False, **kw: Any) -> None:
+        for r in self._results:
+            r._soft_close(hard=hard, **kw)
+        if hard:
+            self.closed = True
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/row.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/row.py
new file mode 100644
index 00000000..da7ae9af
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/row.py
@@ -0,0 +1,400 @@
+# engine/row.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Define row constructs including :class:`.Row`."""
+
+from __future__ import annotations
+
+from abc import ABC
+import collections.abc as collections_abc
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from ..sql import util as sql_util
+from ..util import deprecated
+from ..util._has_cy import HAS_CYEXTENSION
+
+if TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_row import BaseRow as BaseRow
+else:
+    from sqlalchemy.cyextension.resultproxy import BaseRow as BaseRow
+
+if TYPE_CHECKING:
+    from .result import _KeyType
+    from .result import _ProcessorsType
+    from .result import RMKeyView
+
+_T = TypeVar("_T", bound=Any)
+_TP = TypeVar("_TP", bound=Tuple[Any, ...])
+
+
+class Row(BaseRow, Sequence[Any], Generic[_TP]):
+    """Represent a single result row.
+
+    The :class:`.Row` object represents a row of a database result.  It is
+    typically associated in the 1.x series of SQLAlchemy with the
+    :class:`_engine.CursorResult` object, however is also used by the ORM for
+    tuple-like results as of SQLAlchemy 1.4.
+
+    The :class:`.Row` object seeks to act as much like a Python named
+    tuple as possible.   For mapping (i.e. dictionary) behavior on a row,
+    such as testing for containment of keys, refer to the :attr:`.Row._mapping`
+    attribute.
+
+    .. seealso::
+
+        :ref:`tutorial_selecting_data` - includes examples of selecting
+        rows from SELECT statements.
+
+    .. versionchanged:: 1.4
+
+        Renamed ``RowProxy`` to :class:`.Row`. :class:`.Row` is no longer a
+        "proxy" object in that it contains the final form of data within it,
+        and now acts mostly like a named tuple. Mapping-like functionality is
+        moved to the :attr:`.Row._mapping` attribute. See
+        :ref:`change_4710_core` for background on this change.
+
+    """
+
+    __slots__ = ()
+
+    def __setattr__(self, name: str, value: Any) -> NoReturn:
+        raise AttributeError("can't set attribute")
+
+    def __delattr__(self, name: str) -> NoReturn:
+        raise AttributeError("can't delete attribute")
+
+    def _tuple(self) -> _TP:
+        """Return a 'tuple' form of this :class:`.Row`.
+
+        At runtime, this method returns "self"; the :class:`.Row` object is
+        already a named tuple. However, at the typing level, if this
+        :class:`.Row` is typed, the "tuple" return type will be a :pep:`484`
+        ``Tuple`` datatype that contains typing information about individual
+        elements, supporting typed unpacking and attribute access.
+
+        .. versionadded:: 2.0.19 - The :meth:`.Row._tuple` method supersedes
+           the previous :meth:`.Row.tuple` method, which is now underscored
+           to avoid name conflicts with column names in the same way as other
+           named-tuple methods on :class:`.Row`.
+
+        .. seealso::
+
+            :attr:`.Row._t` - shorthand attribute notation
+
+            :meth:`.Result.tuples`
+
+
+        """
+        return self  # type: ignore
+
+    @deprecated(
+        "2.0.19",
+        "The :meth:`.Row.tuple` method is deprecated in favor of "
+        ":meth:`.Row._tuple`; all :class:`.Row` "
+        "methods and library-level attributes are intended to be underscored "
+        "to avoid name conflicts.  Please use :meth:`Row._tuple`.",
+    )
+    def tuple(self) -> _TP:
+        """Return a 'tuple' form of this :class:`.Row`.
+
+        .. versionadded:: 2.0
+
+        """
+        return self._tuple()
+
+    @property
+    def _t(self) -> _TP:
+        """A synonym for :meth:`.Row._tuple`.
+
+        .. versionadded:: 2.0.19 - The :attr:`.Row._t` attribute supersedes
+           the previous :attr:`.Row.t` attribute, which is now underscored
+           to avoid name conflicts with column names in the same way as other
+           named-tuple methods on :class:`.Row`.
+
+        .. seealso::
+
+            :attr:`.Result.t`
+        """
+        return self  # type: ignore
+
+    @property
+    @deprecated(
+        "2.0.19",
+        "The :attr:`.Row.t` attribute is deprecated in favor of "
+        ":attr:`.Row._t`; all :class:`.Row` "
+        "methods and library-level attributes are intended to be underscored "
+        "to avoid name conflicts.  Please use :attr:`Row._t`.",
+    )
+    def t(self) -> _TP:
+        """A synonym for :meth:`.Row._tuple`.
+
+        .. versionadded:: 2.0
+
+        """
+        return self._t
+
+    @property
+    def _mapping(self) -> RowMapping:
+        """Return a :class:`.RowMapping` for this :class:`.Row`.
+
+        This object provides a consistent Python mapping (i.e. dictionary)
+        interface for the data contained within the row.   The :class:`.Row`
+        by itself behaves like a named tuple.
+
+        .. seealso::
+
+            :attr:`.Row._fields`
+
+        .. versionadded:: 1.4
+
+        """
+        return RowMapping(self._parent, None, self._key_to_index, self._data)
+
+    def _filter_on_values(
+        self, processor: Optional[_ProcessorsType]
+    ) -> Row[Any]:
+        return Row(self._parent, processor, self._key_to_index, self._data)
+
+    if not TYPE_CHECKING:
+
+        def _special_name_accessor(name: str) -> Any:
+            """Handle ambiguous names such as "count" and "index" """
+
+            @property
+            def go(self: Row) -> Any:
+                if self._parent._has_key(name):
+                    return self.__getattr__(name)
+                else:
+
+                    def meth(*arg: Any, **kw: Any) -> Any:
+                        return getattr(collections_abc.Sequence, name)(
+                            self, *arg, **kw
+                        )
+
+                    return meth
+
+            return go
+
+        count = _special_name_accessor("count")
+        index = _special_name_accessor("index")
+
+    def __contains__(self, key: Any) -> bool:
+        return key in self._data
+
+    def _op(self, other: Any, op: Callable[[Any, Any], bool]) -> bool:
+        return (
+            op(self._to_tuple_instance(), other._to_tuple_instance())
+            if isinstance(other, Row)
+            else op(self._to_tuple_instance(), other)
+        )
+
+    __hash__ = BaseRow.__hash__
+
+    if TYPE_CHECKING:
+
+        @overload
+        def __getitem__(self, index: int) -> Any: ...
+
+        @overload
+        def __getitem__(self, index: slice) -> Sequence[Any]: ...
+
+        def __getitem__(self, index: Union[int, slice]) -> Any: ...
+
+    def __lt__(self, other: Any) -> bool:
+        return self._op(other, operator.lt)
+
+    def __le__(self, other: Any) -> bool:
+        return self._op(other, operator.le)
+
+    def __ge__(self, other: Any) -> bool:
+        return self._op(other, operator.ge)
+
+    def __gt__(self, other: Any) -> bool:
+        return self._op(other, operator.gt)
+
+    def __eq__(self, other: Any) -> bool:
+        return self._op(other, operator.eq)
+
+    def __ne__(self, other: Any) -> bool:
+        return self._op(other, operator.ne)
+
+    def __repr__(self) -> str:
+        return repr(sql_util._repr_row(self))
+
+    @property
+    def _fields(self) -> Tuple[str, ...]:
+        """Return a tuple of string keys as represented by this
+        :class:`.Row`.
+
+        The keys can represent the labels of the columns returned by a core
+        statement or the names of the orm classes returned by an orm
+        execution.
+
+        This attribute is analogous to the Python named tuple ``._fields``
+        attribute.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`.Row._mapping`
+
+        """
+        return tuple([k for k in self._parent.keys if k is not None])
+
+    def _asdict(self) -> Dict[str, Any]:
+        """Return a new dict which maps field names to their corresponding
+        values.
+
+        This method is analogous to the Python named tuple ``._asdict()``
+        method, and works by applying the ``dict()`` constructor to the
+        :attr:`.Row._mapping` attribute.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`.Row._mapping`
+
+        """
+        return dict(self._mapping)
+
+
+BaseRowProxy = BaseRow
+RowProxy = Row
+
+
+class ROMappingView(ABC):
+    __slots__ = ()
+
+    _items: Sequence[Any]
+    _mapping: Mapping["_KeyType", Any]
+
+    def __init__(
+        self, mapping: Mapping["_KeyType", Any], items: Sequence[Any]
+    ):
+        self._mapping = mapping  # type: ignore[misc]
+        self._items = items  # type: ignore[misc]
+
+    def __len__(self) -> int:
+        return len(self._items)
+
+    def __repr__(self) -> str:
+        return "{0.__class__.__name__}({0._mapping!r})".format(self)
+
+    def __iter__(self) -> Iterator[Any]:
+        return iter(self._items)
+
+    def __contains__(self, item: Any) -> bool:
+        return item in self._items
+
+    def __eq__(self, other: Any) -> bool:
+        return list(other) == list(self)
+
+    def __ne__(self, other: Any) -> bool:
+        return list(other) != list(self)
+
+
+class ROMappingKeysValuesView(
+    ROMappingView, typing.KeysView["_KeyType"], typing.ValuesView[Any]
+):
+    __slots__ = ("_items",)  # mapping slot is provided by KeysView
+
+
+class ROMappingItemsView(ROMappingView, typing.ItemsView["_KeyType", Any]):
+    __slots__ = ("_items",)  # mapping slot is provided by ItemsView
+
+
+class RowMapping(BaseRow, typing.Mapping["_KeyType", Any]):
+    """A ``Mapping`` that maps column names and objects to :class:`.Row`
+    values.
+
+    The :class:`.RowMapping` is available from a :class:`.Row` via the
+    :attr:`.Row._mapping` attribute, as well as from the iterable interface
+    provided by the :class:`.MappingResult` object returned by the
+    :meth:`_engine.Result.mappings` method.
+
+    :class:`.RowMapping` supplies Python mapping (i.e. dictionary) access to
+    the  contents of the row.   This includes support for testing of
+    containment of specific keys (string column names or objects), as well
+    as iteration of keys, values, and items::
+
+        for row in result:
+            if "a" in row._mapping:
+                print("Column 'a': %s" % row._mapping["a"])
+
+            print("Column b: %s" % row._mapping[table.c.b])
+
+    .. versionadded:: 1.4 The :class:`.RowMapping` object replaces the
+       mapping-like access previously provided by a database result row,
+       which now seeks to behave mostly like a named tuple.
+
+    """
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        def __getitem__(self, key: _KeyType) -> Any: ...
+
+    else:
+        __getitem__ = BaseRow._get_by_key_impl_mapping
+
+    def _values_impl(self) -> List[Any]:
+        return list(self._data)
+
+    def __iter__(self) -> Iterator[str]:
+        return (k for k in self._parent.keys if k is not None)
+
+    def __len__(self) -> int:
+        return len(self._data)
+
+    def __contains__(self, key: object) -> bool:
+        return self._parent._has_key(key)
+
+    def __repr__(self) -> str:
+        return repr(dict(self))
+
+    def items(self) -> ROMappingItemsView:
+        """Return a view of key/value tuples for the elements in the
+        underlying :class:`.Row`.
+
+        """
+        return ROMappingItemsView(
+            self, [(key, self[key]) for key in self.keys()]
+        )
+
+    def keys(self) -> RMKeyView:
+        """Return a view of 'keys' for string column names represented
+        by the underlying :class:`.Row`.
+
+        """
+
+        return self._parent.keys
+
+    def values(self) -> ROMappingKeysValuesView:
+        """Return a view of values for the values represented in the
+        underlying :class:`.Row`.
+
+        """
+        return ROMappingKeysValuesView(self, self._values_impl())
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/strategies.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/strategies.py
new file mode 100644
index 00000000..5dd7bca9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/strategies.py
@@ -0,0 +1,19 @@
+# engine/strategies.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Deprecated mock engine strategy used by Alembic.
+
+
+"""
+
+from __future__ import annotations
+
+from .mock import MockConnection  # noqa
+
+
+class MockEngineStrategy:
+    MockConnection = MockConnection
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/url.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/url.py
new file mode 100644
index 00000000..bb004f11
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/url.py
@@ -0,0 +1,924 @@
+# engine/url.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Provides the :class:`~sqlalchemy.engine.url.URL` class which encapsulates
+information about a database connection specification.
+
+The URL object is created automatically when
+:func:`~sqlalchemy.engine.create_engine` is called with a string
+argument; alternatively, the URL is a public-facing construct which can
+be used directly and is also accepted directly by ``create_engine()``.
+"""
+
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import re
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Mapping
+from typing import NamedTuple
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import Union
+from urllib.parse import parse_qsl
+from urllib.parse import quote
+from urllib.parse import quote_plus
+from urllib.parse import unquote
+
+from .interfaces import Dialect
+from .. import exc
+from .. import util
+from ..dialects import plugins
+from ..dialects import registry
+
+
+class URL(NamedTuple):
+    """
+    Represent the components of a URL used to connect to a database.
+
+    URLs are typically constructed from a fully formatted URL string, where the
+    :func:`.make_url` function is used internally by the
+    :func:`_sa.create_engine` function in order to parse the URL string into
+    its individual components, which are then used to construct a new
+    :class:`.URL` object. When parsing from a formatted URL string, the parsing
+    format generally follows
+    `RFC-1738 <https://www.ietf.org/rfc/rfc1738.txt>`_, with some exceptions.
+
+    A :class:`_engine.URL` object may also be produced directly, either by
+    using the :func:`.make_url` function with a fully formed URL string, or
+    by using the :meth:`_engine.URL.create` constructor in order
+    to construct a :class:`_engine.URL` programmatically given individual
+    fields. The resulting :class:`.URL` object may be passed directly to
+    :func:`_sa.create_engine` in place of a string argument, which will bypass
+    the usage of :func:`.make_url` within the engine's creation process.
+
+    .. versionchanged:: 1.4
+
+        The :class:`_engine.URL` object is now an immutable object.  To
+        create a URL, use the :func:`_engine.make_url` or
+        :meth:`_engine.URL.create` function / method.  To modify
+        a :class:`_engine.URL`, use methods like
+        :meth:`_engine.URL.set` and
+        :meth:`_engine.URL.update_query_dict` to return a new
+        :class:`_engine.URL` object with modifications.   See notes for this
+        change at :ref:`change_5526`.
+
+    .. seealso::
+
+        :ref:`database_urls`
+
+    :class:`_engine.URL` contains the following attributes:
+
+    * :attr:`_engine.URL.drivername`: database backend and driver name, such as
+      ``postgresql+psycopg2``
+    * :attr:`_engine.URL.username`: username string
+    * :attr:`_engine.URL.password`: password string
+    * :attr:`_engine.URL.host`: string hostname
+    * :attr:`_engine.URL.port`: integer port number
+    * :attr:`_engine.URL.database`: string database name
+    * :attr:`_engine.URL.query`: an immutable mapping representing the query
+      string.  contains strings for keys and either strings or tuples of
+      strings for values.
+
+
+    """
+
+    drivername: str
+    """database backend and driver name, such as
+    ``postgresql+psycopg2``
+
+    """
+
+    username: Optional[str]
+    "username string"
+
+    password: Optional[str]
+    """password, which is normally a string but may also be any
+    object that has a ``__str__()`` method."""
+
+    host: Optional[str]
+    """hostname or IP number.  May also be a data source name for some
+    drivers."""
+
+    port: Optional[int]
+    """integer port number"""
+
+    database: Optional[str]
+    """database name"""
+
+    query: util.immutabledict[str, Union[Tuple[str, ...], str]]
+    """an immutable mapping representing the query string.  contains strings
+       for keys and either strings or tuples of strings for values, e.g.::
+
+            >>> from sqlalchemy.engine import make_url
+            >>> url = make_url(
+            ...     "postgresql+psycopg2://user:pass@host/dbname?alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt"
+            ... )
+            >>> url.query
+            immutabledict({'alt_host': ('host1', 'host2'), 'ssl_cipher': '/path/to/crt'})
+
+         To create a mutable copy of this mapping, use the ``dict`` constructor::
+
+            mutable_query_opts = dict(url.query)
+
+       .. seealso::
+
+          :attr:`_engine.URL.normalized_query` - normalizes all values into sequences
+          for consistent processing
+
+          Methods for altering the contents of :attr:`_engine.URL.query`:
+
+          :meth:`_engine.URL.update_query_dict`
+
+          :meth:`_engine.URL.update_query_string`
+
+          :meth:`_engine.URL.update_query_pairs`
+
+          :meth:`_engine.URL.difference_update_query`
+
+    """  # noqa: E501
+
+    @classmethod
+    def create(
+        cls,
+        drivername: str,
+        username: Optional[str] = None,
+        password: Optional[str] = None,
+        host: Optional[str] = None,
+        port: Optional[int] = None,
+        database: Optional[str] = None,
+        query: Mapping[str, Union[Sequence[str], str]] = util.EMPTY_DICT,
+    ) -> URL:
+        """Create a new :class:`_engine.URL` object.
+
+        .. seealso::
+
+            :ref:`database_urls`
+
+        :param drivername: the name of the database backend. This name will
+          correspond to a module in sqlalchemy/databases or a third party
+          plug-in.
+        :param username: The user name.
+        :param password: database password.  Is typically a string, but may
+          also be an object that can be stringified with ``str()``.
+
+          .. note:: The password string should **not** be URL encoded when
+             passed as an argument to :meth:`_engine.URL.create`; the string
+             should contain the password characters exactly as they would be
+             typed.
+
+          .. note::  A password-producing object will be stringified only
+             **once** per :class:`_engine.Engine` object.  For dynamic password
+             generation per connect, see :ref:`engines_dynamic_tokens`.
+
+        :param host: The name of the host.
+        :param port: The port number.
+        :param database: The database name.
+        :param query: A dictionary of string keys to string values to be passed
+          to the dialect and/or the DBAPI upon connect.   To specify non-string
+          parameters to a Python DBAPI directly, use the
+          :paramref:`_sa.create_engine.connect_args` parameter to
+          :func:`_sa.create_engine`.   See also
+          :attr:`_engine.URL.normalized_query` for a dictionary that is
+          consistently string->list of string.
+        :return: new :class:`_engine.URL` object.
+
+        .. versionadded:: 1.4
+
+            The :class:`_engine.URL` object is now an **immutable named
+            tuple**.  In addition, the ``query`` dictionary is also immutable.
+            To create a URL, use the :func:`_engine.url.make_url` or
+            :meth:`_engine.URL.create` function/ method.  To modify a
+            :class:`_engine.URL`, use the :meth:`_engine.URL.set` and
+            :meth:`_engine.URL.update_query` methods.
+
+        """
+
+        return cls(
+            cls._assert_str(drivername, "drivername"),
+            cls._assert_none_str(username, "username"),
+            password,
+            cls._assert_none_str(host, "host"),
+            cls._assert_port(port),
+            cls._assert_none_str(database, "database"),
+            cls._str_dict(query),
+        )
+
+    @classmethod
+    def _assert_port(cls, port: Optional[int]) -> Optional[int]:
+        if port is None:
+            return None
+        try:
+            return int(port)
+        except TypeError:
+            raise TypeError("Port argument must be an integer or None")
+
+    @classmethod
+    def _assert_str(cls, v: str, paramname: str) -> str:
+        if not isinstance(v, str):
+            raise TypeError("%s must be a string" % paramname)
+        return v
+
+    @classmethod
+    def _assert_none_str(
+        cls, v: Optional[str], paramname: str
+    ) -> Optional[str]:
+        if v is None:
+            return v
+
+        return cls._assert_str(v, paramname)
+
+    @classmethod
+    def _str_dict(
+        cls,
+        dict_: Optional[
+            Union[
+                Sequence[Tuple[str, Union[Sequence[str], str]]],
+                Mapping[str, Union[Sequence[str], str]],
+            ]
+        ],
+    ) -> util.immutabledict[str, Union[Tuple[str, ...], str]]:
+        if dict_ is None:
+            return util.EMPTY_DICT
+
+        @overload
+        def _assert_value(
+            val: str,
+        ) -> str: ...
+
+        @overload
+        def _assert_value(
+            val: Sequence[str],
+        ) -> Union[str, Tuple[str, ...]]: ...
+
+        def _assert_value(
+            val: Union[str, Sequence[str]],
+        ) -> Union[str, Tuple[str, ...]]:
+            if isinstance(val, str):
+                return val
+            elif isinstance(val, collections_abc.Sequence):
+                return tuple(_assert_value(elem) for elem in val)
+            else:
+                raise TypeError(
+                    "Query dictionary values must be strings or "
+                    "sequences of strings"
+                )
+
+        def _assert_str(v: str) -> str:
+            if not isinstance(v, str):
+                raise TypeError("Query dictionary keys must be strings")
+            return v
+
+        dict_items: Iterable[Tuple[str, Union[Sequence[str], str]]]
+        if isinstance(dict_, collections_abc.Sequence):
+            dict_items = dict_
+        else:
+            dict_items = dict_.items()
+
+        return util.immutabledict(
+            {
+                _assert_str(key): _assert_value(
+                    value,
+                )
+                for key, value in dict_items
+            }
+        )
+
+    def set(
+        self,
+        drivername: Optional[str] = None,
+        username: Optional[str] = None,
+        password: Optional[str] = None,
+        host: Optional[str] = None,
+        port: Optional[int] = None,
+        database: Optional[str] = None,
+        query: Optional[Mapping[str, Union[Sequence[str], str]]] = None,
+    ) -> URL:
+        """return a new :class:`_engine.URL` object with modifications.
+
+        Values are used if they are non-None.  To set a value to ``None``
+        explicitly, use the :meth:`_engine.URL._replace` method adapted
+        from ``namedtuple``.
+
+        :param drivername: new drivername
+        :param username: new username
+        :param password: new password
+        :param host: new hostname
+        :param port: new port
+        :param query: new query parameters, passed a dict of string keys
+         referring to string or sequence of string values.  Fully
+         replaces the previous list of arguments.
+
+        :return: new :class:`_engine.URL` object.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :meth:`_engine.URL.update_query_dict`
+
+        """
+
+        kw: Dict[str, Any] = {}
+        if drivername is not None:
+            kw["drivername"] = drivername
+        if username is not None:
+            kw["username"] = username
+        if password is not None:
+            kw["password"] = password
+        if host is not None:
+            kw["host"] = host
+        if port is not None:
+            kw["port"] = port
+        if database is not None:
+            kw["database"] = database
+        if query is not None:
+            kw["query"] = query
+
+        return self._assert_replace(**kw)
+
+    def _assert_replace(self, **kw: Any) -> URL:
+        """argument checks before calling _replace()"""
+
+        if "drivername" in kw:
+            self._assert_str(kw["drivername"], "drivername")
+        for name in "username", "host", "database":
+            if name in kw:
+                self._assert_none_str(kw[name], name)
+        if "port" in kw:
+            self._assert_port(kw["port"])
+        if "query" in kw:
+            kw["query"] = self._str_dict(kw["query"])
+
+        return self._replace(**kw)
+
+    def update_query_string(
+        self, query_string: str, append: bool = False
+    ) -> URL:
+        """Return a new :class:`_engine.URL` object with the :attr:`_engine.URL.query`
+        parameter dictionary updated by the given query string.
+
+        E.g.::
+
+            >>> from sqlalchemy.engine import make_url
+            >>> url = make_url("postgresql+psycopg2://user:pass@host/dbname")
+            >>> url = url.update_query_string(
+            ...     "alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt"
+            ... )
+            >>> str(url)
+            'postgresql+psycopg2://user:pass@host/dbname?alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt'
+
+        :param query_string: a URL escaped query string, not including the
+         question mark.
+
+        :param append: if True, parameters in the existing query string will
+         not be removed; new parameters will be in addition to those present.
+         If left at its default of False, keys present in the given query
+         parameters will replace those of the existing query string.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_engine.URL.query`
+
+            :meth:`_engine.URL.update_query_dict`
+
+        """  # noqa: E501
+        return self.update_query_pairs(parse_qsl(query_string), append=append)
+
+    def update_query_pairs(
+        self,
+        key_value_pairs: Iterable[Tuple[str, Union[str, List[str]]]],
+        append: bool = False,
+    ) -> URL:
+        """Return a new :class:`_engine.URL` object with the
+        :attr:`_engine.URL.query`
+        parameter dictionary updated by the given sequence of key/value pairs
+
+        E.g.::
+
+            >>> from sqlalchemy.engine import make_url
+            >>> url = make_url("postgresql+psycopg2://user:pass@host/dbname")
+            >>> url = url.update_query_pairs(
+            ...     [
+            ...         ("alt_host", "host1"),
+            ...         ("alt_host", "host2"),
+            ...         ("ssl_cipher", "/path/to/crt"),
+            ...     ]
+            ... )
+            >>> str(url)
+            'postgresql+psycopg2://user:pass@host/dbname?alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt'
+
+        :param key_value_pairs: A sequence of tuples containing two strings
+         each.
+
+        :param append: if True, parameters in the existing query string will
+         not be removed; new parameters will be in addition to those present.
+         If left at its default of False, keys present in the given query
+         parameters will replace those of the existing query string.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_engine.URL.query`
+
+            :meth:`_engine.URL.difference_update_query`
+
+            :meth:`_engine.URL.set`
+
+        """  # noqa: E501
+
+        existing_query = self.query
+        new_keys: Dict[str, Union[str, List[str]]] = {}
+
+        for key, value in key_value_pairs:
+            if key in new_keys:
+                new_keys[key] = util.to_list(new_keys[key])
+                cast("List[str]", new_keys[key]).append(cast(str, value))
+            else:
+                new_keys[key] = (
+                    list(value) if isinstance(value, (list, tuple)) else value
+                )
+
+        new_query: Mapping[str, Union[str, Sequence[str]]]
+        if append:
+            new_query = {}
+
+            for k in new_keys:
+                if k in existing_query:
+                    new_query[k] = tuple(
+                        util.to_list(existing_query[k])
+                        + util.to_list(new_keys[k])
+                    )
+                else:
+                    new_query[k] = new_keys[k]
+
+            new_query.update(
+                {
+                    k: existing_query[k]
+                    for k in set(existing_query).difference(new_keys)
+                }
+            )
+        else:
+            new_query = self.query.union(
+                {
+                    k: tuple(v) if isinstance(v, list) else v
+                    for k, v in new_keys.items()
+                }
+            )
+        return self.set(query=new_query)
+
+    def update_query_dict(
+        self,
+        query_parameters: Mapping[str, Union[str, List[str]]],
+        append: bool = False,
+    ) -> URL:
+        """Return a new :class:`_engine.URL` object with the
+        :attr:`_engine.URL.query` parameter dictionary updated by the given
+        dictionary.
+
+        The dictionary typically contains string keys and string values.
+        In order to represent a query parameter that is expressed multiple
+        times, pass a sequence of string values.
+
+        E.g.::
+
+
+            >>> from sqlalchemy.engine import make_url
+            >>> url = make_url("postgresql+psycopg2://user:pass@host/dbname")
+            >>> url = url.update_query_dict(
+            ...     {"alt_host": ["host1", "host2"], "ssl_cipher": "/path/to/crt"}
+            ... )
+            >>> str(url)
+            'postgresql+psycopg2://user:pass@host/dbname?alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt'
+
+
+        :param query_parameters: A dictionary with string keys and values
+         that are either strings, or sequences of strings.
+
+        :param append: if True, parameters in the existing query string will
+         not be removed; new parameters will be in addition to those present.
+         If left at its default of False, keys present in the given query
+         parameters will replace those of the existing query string.
+
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_engine.URL.query`
+
+            :meth:`_engine.URL.update_query_string`
+
+            :meth:`_engine.URL.update_query_pairs`
+
+            :meth:`_engine.URL.difference_update_query`
+
+            :meth:`_engine.URL.set`
+
+        """  # noqa: E501
+        return self.update_query_pairs(query_parameters.items(), append=append)
+
+    def difference_update_query(self, names: Iterable[str]) -> URL:
+        """
+        Remove the given names from the :attr:`_engine.URL.query` dictionary,
+        returning the new :class:`_engine.URL`.
+
+        E.g.::
+
+            url = url.difference_update_query(["foo", "bar"])
+
+        Equivalent to using :meth:`_engine.URL.set` as follows::
+
+            url = url.set(
+                query={
+                    key: url.query[key]
+                    for key in set(url.query).difference(["foo", "bar"])
+                }
+            )
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_engine.URL.query`
+
+            :meth:`_engine.URL.update_query_dict`
+
+            :meth:`_engine.URL.set`
+
+        """
+
+        if not set(names).intersection(self.query):
+            return self
+
+        return URL(
+            self.drivername,
+            self.username,
+            self.password,
+            self.host,
+            self.port,
+            self.database,
+            util.immutabledict(
+                {
+                    key: self.query[key]
+                    for key in set(self.query).difference(names)
+                }
+            ),
+        )
+
+    @property
+    def normalized_query(self) -> Mapping[str, Sequence[str]]:
+        """Return the :attr:`_engine.URL.query` dictionary with values normalized
+        into sequences.
+
+        As the :attr:`_engine.URL.query` dictionary may contain either
+        string values or sequences of string values to differentiate between
+        parameters that are specified multiple times in the query string,
+        code that needs to handle multiple parameters generically will wish
+        to use this attribute so that all parameters present are presented
+        as sequences.   Inspiration is from Python's ``urllib.parse.parse_qs``
+        function.  E.g.::
+
+
+            >>> from sqlalchemy.engine import make_url
+            >>> url = make_url(
+            ...     "postgresql+psycopg2://user:pass@host/dbname?alt_host=host1&alt_host=host2&ssl_cipher=%2Fpath%2Fto%2Fcrt"
+            ... )
+            >>> url.query
+            immutabledict({'alt_host': ('host1', 'host2'), 'ssl_cipher': '/path/to/crt'})
+            >>> url.normalized_query
+            immutabledict({'alt_host': ('host1', 'host2'), 'ssl_cipher': ('/path/to/crt',)})
+
+        """  # noqa: E501
+
+        return util.immutabledict(
+            {
+                k: (v,) if not isinstance(v, tuple) else v
+                for k, v in self.query.items()
+            }
+        )
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_engine.URL.__to_string__ method is deprecated and will "
+        "be removed in a future release.  Please use the "
+        ":meth:`_engine.URL.render_as_string` method.",
+    )
+    def __to_string__(self, hide_password: bool = True) -> str:
+        """Render this :class:`_engine.URL` object as a string.
+
+        :param hide_password: Defaults to True.   The password is not shown
+         in the string unless this is set to False.
+
+        """
+        return self.render_as_string(hide_password=hide_password)
+
+    def render_as_string(self, hide_password: bool = True) -> str:
+        """Render this :class:`_engine.URL` object as a string.
+
+        This method is used when the ``__str__()`` or ``__repr__()``
+        methods are used.   The method directly includes additional options.
+
+        :param hide_password: Defaults to True.   The password is not shown
+         in the string unless this is set to False.
+
+        """
+        s = self.drivername + "://"
+        if self.username is not None:
+            s += quote(self.username, safe=" +")
+            if self.password is not None:
+                s += ":" + (
+                    "***"
+                    if hide_password
+                    else quote(str(self.password), safe=" +")
+                )
+            s += "@"
+        if self.host is not None:
+            if ":" in self.host:
+                s += f"[{self.host}]"
+            else:
+                s += self.host
+        if self.port is not None:
+            s += ":" + str(self.port)
+        if self.database is not None:
+            s += "/" + self.database
+        if self.query:
+            keys = list(self.query)
+            keys.sort()
+            s += "?" + "&".join(
+                f"{quote_plus(k)}={quote_plus(element)}"
+                for k in keys
+                for element in util.to_list(self.query[k])
+            )
+        return s
+
+    def __repr__(self) -> str:
+        return self.render_as_string()
+
+    def __copy__(self) -> URL:
+        return self.__class__.create(
+            self.drivername,
+            self.username,
+            self.password,
+            self.host,
+            self.port,
+            self.database,
+            # note this is an immutabledict of str-> str / tuple of str,
+            # also fully immutable.  does not require deepcopy
+            self.query,
+        )
+
+    def __deepcopy__(self, memo: Any) -> URL:
+        return self.__copy__()
+
+    def __hash__(self) -> int:
+        return hash(str(self))
+
+    def __eq__(self, other: Any) -> bool:
+        return (
+            isinstance(other, URL)
+            and self.drivername == other.drivername
+            and self.username == other.username
+            and self.password == other.password
+            and self.host == other.host
+            and self.database == other.database
+            and self.query == other.query
+            and self.port == other.port
+        )
+
+    def __ne__(self, other: Any) -> bool:
+        return not self == other
+
+    def get_backend_name(self) -> str:
+        """Return the backend name.
+
+        This is the name that corresponds to the database backend in
+        use, and is the portion of the :attr:`_engine.URL.drivername`
+        that is to the left of the plus sign.
+
+        """
+        if "+" not in self.drivername:
+            return self.drivername
+        else:
+            return self.drivername.split("+")[0]
+
+    def get_driver_name(self) -> str:
+        """Return the backend name.
+
+        This is the name that corresponds to the DBAPI driver in
+        use, and is the portion of the :attr:`_engine.URL.drivername`
+        that is to the right of the plus sign.
+
+        If the :attr:`_engine.URL.drivername` does not include a plus sign,
+        then the default :class:`_engine.Dialect` for this :class:`_engine.URL`
+        is imported in order to get the driver name.
+
+        """
+
+        if "+" not in self.drivername:
+            return self.get_dialect().driver
+        else:
+            return self.drivername.split("+")[1]
+
+    def _instantiate_plugins(
+        self, kwargs: Mapping[str, Any]
+    ) -> Tuple[URL, List[Any], Dict[str, Any]]:
+        plugin_names = util.to_list(self.query.get("plugin", ()))
+        plugin_names += kwargs.get("plugins", [])
+
+        kwargs = dict(kwargs)
+
+        loaded_plugins = [
+            plugins.load(plugin_name)(self, kwargs)
+            for plugin_name in plugin_names
+        ]
+
+        u = self.difference_update_query(["plugin", "plugins"])
+
+        for plugin in loaded_plugins:
+            new_u = plugin.update_url(u)
+            if new_u is not None:
+                u = new_u
+
+        kwargs.pop("plugins", None)
+
+        return u, loaded_plugins, kwargs
+
+    def _get_entrypoint(self) -> Type[Dialect]:
+        """Return the "entry point" dialect class.
+
+        This is normally the dialect itself except in the case when the
+        returned class implements the get_dialect_cls() method.
+
+        """
+        if "+" not in self.drivername:
+            name = self.drivername
+        else:
+            name = self.drivername.replace("+", ".")
+        cls = registry.load(name)
+        # check for legacy dialects that
+        # would return a module with 'dialect' as the
+        # actual class
+        if (
+            hasattr(cls, "dialect")
+            and isinstance(cls.dialect, type)
+            and issubclass(cls.dialect, Dialect)
+        ):
+            return cls.dialect
+        else:
+            return cast("Type[Dialect]", cls)
+
+    def get_dialect(self, _is_async: bool = False) -> Type[Dialect]:
+        """Return the SQLAlchemy :class:`_engine.Dialect` class corresponding
+        to this URL's driver name.
+
+        """
+        entrypoint = self._get_entrypoint()
+        if _is_async:
+            dialect_cls = entrypoint.get_async_dialect_cls(self)
+        else:
+            dialect_cls = entrypoint.get_dialect_cls(self)
+        return dialect_cls
+
+    def translate_connect_args(
+        self, names: Optional[List[str]] = None, **kw: Any
+    ) -> Dict[str, Any]:
+        r"""Translate url attributes into a dictionary of connection arguments.
+
+        Returns attributes of this url (`host`, `database`, `username`,
+        `password`, `port`) as a plain dictionary.  The attribute names are
+        used as the keys by default.  Unset or false attributes are omitted
+        from the final dictionary.
+
+        :param \**kw: Optional, alternate key names for url attributes.
+
+        :param names: Deprecated.  Same purpose as the keyword-based alternate
+            names, but correlates the name to the original positionally.
+        """
+
+        if names is not None:
+            util.warn_deprecated(
+                "The `URL.translate_connect_args.name`s parameter is "
+                "deprecated. Please pass the "
+                "alternate names as kw arguments.",
+                "1.4",
+            )
+
+        translated = {}
+        attribute_names = ["host", "database", "username", "password", "port"]
+        for sname in attribute_names:
+            if names:
+                name = names.pop(0)
+            elif sname in kw:
+                name = kw[sname]
+            else:
+                name = sname
+            if name is not None and getattr(self, sname, False):
+                if sname == "password":
+                    translated[name] = str(getattr(self, sname))
+                else:
+                    translated[name] = getattr(self, sname)
+
+        return translated
+
+
+def make_url(name_or_url: Union[str, URL]) -> URL:
+    """Given a string, produce a new URL instance.
+
+    The format of the URL generally follows `RFC-1738
+    <https://www.ietf.org/rfc/rfc1738.txt>`_, with some exceptions, including
+    that underscores, and not dashes or periods, are accepted within the
+    "scheme" portion.
+
+    If a :class:`.URL` object is passed, it is returned as is.
+
+    .. seealso::
+
+        :ref:`database_urls`
+
+    """
+
+    if isinstance(name_or_url, str):
+        return _parse_url(name_or_url)
+    elif not isinstance(name_or_url, URL) and not hasattr(
+        name_or_url, "_sqla_is_testing_if_this_is_a_mock_object"
+    ):
+        raise exc.ArgumentError(
+            f"Expected string or URL object, got {name_or_url!r}"
+        )
+    else:
+        return name_or_url
+
+
+def _parse_url(name: str) -> URL:
+    pattern = re.compile(
+        r"""
+            (?P<name>[\w\+]+)://
+            (?:
+                (?P<username>[^:/]*)
+                (?::(?P<password>[^@]*))?
+            @)?
+            (?:
+                (?:
+                    \[(?P<ipv6host>[^/\?]+)\] |
+                    (?P<ipv4host>[^/:\?]+)
+                )?
+                (?::(?P<port>[^/\?]*))?
+            )?
+            (?:/(?P<database>[^\?]*))?
+            (?:\?(?P<query>.*))?
+            """,
+        re.X,
+    )
+
+    m = pattern.match(name)
+    if m is not None:
+        components = m.groupdict()
+        query: Optional[Dict[str, Union[str, List[str]]]]
+        if components["query"] is not None:
+            query = {}
+
+            for key, value in parse_qsl(components["query"]):
+                if key in query:
+                    query[key] = util.to_list(query[key])
+                    cast("List[str]", query[key]).append(value)
+                else:
+                    query[key] = value
+        else:
+            query = None
+        components["query"] = query
+
+        if components["username"] is not None:
+            components["username"] = unquote(components["username"])
+
+        if components["password"] is not None:
+            components["password"] = unquote(components["password"])
+
+        ipv4host = components.pop("ipv4host")
+        ipv6host = components.pop("ipv6host")
+        components["host"] = ipv4host or ipv6host
+        name = components.pop("name")
+
+        if components["port"]:
+            components["port"] = int(components["port"])
+
+        return URL.create(name, **components)  # type: ignore
+
+    else:
+        raise exc.ArgumentError(
+            "Could not parse SQLAlchemy URL from string '%s'" % name
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/util.py
new file mode 100644
index 00000000..e499efa9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/engine/util.py
@@ -0,0 +1,167 @@
+# engine/util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import Optional
+from typing import TypeVar
+
+from .. import exc
+from .. import util
+from ..util._has_cy import HAS_CYEXTENSION
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+if typing.TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_util import _distill_params_20 as _distill_params_20
+    from ._py_util import _distill_raw_params as _distill_raw_params
+else:
+    from sqlalchemy.cyextension.util import (  # noqa: F401
+        _distill_params_20 as _distill_params_20,
+    )
+    from sqlalchemy.cyextension.util import (  # noqa: F401
+        _distill_raw_params as _distill_raw_params,
+    )
+
+_C = TypeVar("_C", bound=Callable[[], Any])
+
+
+def connection_memoize(key: str) -> Callable[[_C], _C]:
+    """Decorator, memoize a function in a connection.info stash.
+
+    Only applicable to functions which take no arguments other than a
+    connection.  The memo will be stored in ``connection.info[key]``.
+    """
+
+    @util.decorator
+    def decorated(fn, self, connection):  # type: ignore
+        connection = connection.connect()
+        try:
+            return connection.info[key]
+        except KeyError:
+            connection.info[key] = val = fn(self, connection)
+            return val
+
+    return decorated
+
+
+class _TConsSubject(Protocol):
+    _trans_context_manager: Optional[TransactionalContext]
+
+
+class TransactionalContext:
+    """Apply Python context manager behavior to transaction objects.
+
+    Performs validation to ensure the subject of the transaction is not
+    used if the transaction were ended prematurely.
+
+    """
+
+    __slots__ = ("_outer_trans_ctx", "_trans_subject", "__weakref__")
+
+    _trans_subject: Optional[_TConsSubject]
+
+    def _transaction_is_active(self) -> bool:
+        raise NotImplementedError()
+
+    def _transaction_is_closed(self) -> bool:
+        raise NotImplementedError()
+
+    def _rollback_can_be_called(self) -> bool:
+        """indicates the object is in a state that is known to be acceptable
+        for rollback() to be called.
+
+        This does not necessarily mean rollback() will succeed or not raise
+        an error, just that there is currently no state detected that indicates
+        rollback() would fail or emit warnings.
+
+        It also does not mean that there's a transaction in progress, as
+        it is usually safe to call rollback() even if no transaction is
+        present.
+
+        .. versionadded:: 1.4.28
+
+        """
+        raise NotImplementedError()
+
+    def _get_subject(self) -> _TConsSubject:
+        raise NotImplementedError()
+
+    def commit(self) -> None:
+        raise NotImplementedError()
+
+    def rollback(self) -> None:
+        raise NotImplementedError()
+
+    def close(self) -> None:
+        raise NotImplementedError()
+
+    @classmethod
+    def _trans_ctx_check(cls, subject: _TConsSubject) -> None:
+        trans_context = subject._trans_context_manager
+        if trans_context:
+            if not trans_context._transaction_is_active():
+                raise exc.InvalidRequestError(
+                    "Can't operate on closed transaction inside context "
+                    "manager.  Please complete the context manager "
+                    "before emitting further commands."
+                )
+
+    def __enter__(self) -> Self:
+        subject = self._get_subject()
+
+        # none for outer transaction, may be non-None for nested
+        # savepoint, legacy nesting cases
+        trans_context = subject._trans_context_manager
+        self._outer_trans_ctx = trans_context
+
+        self._trans_subject = subject
+        subject._trans_context_manager = self
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        subject = getattr(self, "_trans_subject", None)
+
+        # simplistically we could assume that
+        # "subject._trans_context_manager is self".  However, any calling
+        # code that is manipulating __exit__ directly would break this
+        # assumption.  alembic context manager
+        # is an example of partial use that just calls __exit__ and
+        # not __enter__ at the moment.  it's safe to assume this is being done
+        # in the wild also
+        out_of_band_exit = (
+            subject is None or subject._trans_context_manager is not self
+        )
+
+        if type_ is None and self._transaction_is_active():
+            try:
+                self.commit()
+            except:
+                with util.safe_reraise():
+                    if self._rollback_can_be_called():
+                        self.rollback()
+            finally:
+                if not out_of_band_exit:
+                    assert subject is not None
+                    subject._trans_context_manager = self._outer_trans_ctx
+                self._trans_subject = self._outer_trans_ctx = None
+        else:
+            try:
+                if not self._transaction_is_active():
+                    if not self._transaction_is_closed():
+                        self.close()
+                else:
+                    if self._rollback_can_be_called():
+                        self.rollback()
+            finally:
+                if not out_of_band_exit:
+                    assert subject is not None
+                    subject._trans_context_manager = self._outer_trans_ctx
+                self._trans_subject = self._outer_trans_ctx = None
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__init__.py
new file mode 100644
index 00000000..309b7bd3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__init__.py
@@ -0,0 +1,25 @@
+# event/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from .api import CANCEL as CANCEL
+from .api import contains as contains
+from .api import listen as listen
+from .api import listens_for as listens_for
+from .api import NO_RETVAL as NO_RETVAL
+from .api import remove as remove
+from .attr import _InstanceLevelDispatch as _InstanceLevelDispatch
+from .attr import RefCollection as RefCollection
+from .base import _Dispatch as _Dispatch
+from .base import _DispatchCommon as _DispatchCommon
+from .base import dispatcher as dispatcher
+from .base import Events as Events
+from .legacy import _legacy_signature as _legacy_signature
+from .registry import _EventKey as _EventKey
+from .registry import _ListenerFnType as _ListenerFnType
+from .registry import EventTarget as EventTarget
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..5a0e057e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/api.cpython-312.pyc
new file mode 100644
index 00000000..b0d2ea0e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/attr.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/attr.cpython-312.pyc
new file mode 100644
index 00000000..0a4fd2b7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/attr.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..c8af7620
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/legacy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/legacy.cpython-312.pyc
new file mode 100644
index 00000000..efc3d18c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/legacy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/registry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/registry.cpython-312.pyc
new file mode 100644
index 00000000..025e1e2f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/__pycache__/registry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/api.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/api.py
new file mode 100644
index 00000000..b6ec8f6d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/api.py
@@ -0,0 +1,222 @@
+# event/api.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Public API functions for the event system.
+
+"""
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+
+from .base import _registrars
+from .registry import _ET
+from .registry import _EventKey
+from .registry import _ListenerFnType
+from .. import exc
+from .. import util
+
+
+CANCEL = util.symbol("CANCEL")
+NO_RETVAL = util.symbol("NO_RETVAL")
+
+
+def _event_key(
+    target: _ET, identifier: str, fn: _ListenerFnType
+) -> _EventKey[_ET]:
+    for evt_cls in _registrars[identifier]:
+        tgt = evt_cls._accept_with(target, identifier)
+        if tgt is not None:
+            return _EventKey(target, identifier, fn, tgt)
+    else:
+        raise exc.InvalidRequestError(
+            "No such event '%s' for target '%s'" % (identifier, target)
+        )
+
+
+def listen(
+    target: Any, identifier: str, fn: Callable[..., Any], *args: Any, **kw: Any
+) -> None:
+    """Register a listener function for the given target.
+
+    The :func:`.listen` function is part of the primary interface for the
+    SQLAlchemy event system, documented at :ref:`event_toplevel`.
+
+    e.g.::
+
+        from sqlalchemy import event
+        from sqlalchemy.schema import UniqueConstraint
+
+
+        def unique_constraint_name(const, table):
+            const.name = "uq_%s_%s" % (table.name, list(const.columns)[0].name)
+
+
+        event.listen(
+            UniqueConstraint, "after_parent_attach", unique_constraint_name
+        )
+
+    :param bool insert: The default behavior for event handlers is to append
+      the decorated user defined function to an internal list of registered
+      event listeners upon discovery. If a user registers a function with
+      ``insert=True``, SQLAlchemy will insert (prepend) the function to the
+      internal list upon discovery. This feature is not typically used or
+      recommended by the SQLAlchemy maintainers, but is provided to ensure
+      certain user defined functions can run before others, such as when
+      :ref:`Changing the sql_mode in MySQL <mysql_sql_mode>`.
+
+    :param bool named: When using named argument passing, the names listed in
+      the function argument specification will be used as keys in the
+      dictionary.
+      See :ref:`event_named_argument_styles`.
+
+    :param bool once: Private/Internal API usage. Deprecated.  This parameter
+      would provide that an event function would run only once per given
+      target. It does not however imply automatic de-registration of the
+      listener function; associating an arbitrarily high number of listeners
+      without explicitly removing them will cause memory to grow unbounded even
+      if ``once=True`` is specified.
+
+    :param bool propagate: The ``propagate`` kwarg is available when working
+      with ORM instrumentation and mapping events.
+      See :class:`_ormevent.MapperEvents` and
+      :meth:`_ormevent.MapperEvents.before_mapper_configured` for examples.
+
+    :param bool retval: This flag applies only to specific event listeners,
+      each of which includes documentation explaining when it should be used.
+      By default, no listener ever requires a return value.
+      However, some listeners do support special behaviors for return values,
+      and include in their documentation that the ``retval=True`` flag is
+      necessary for a return value to be processed.
+
+      Event listener suites that make use of :paramref:`_event.listen.retval`
+      include :class:`_events.ConnectionEvents` and
+      :class:`_ormevent.AttributeEvents`.
+
+    .. note::
+
+        The :func:`.listen` function cannot be called at the same time
+        that the target event is being run.   This has implications
+        for thread safety, and also means an event cannot be added
+        from inside the listener function for itself.  The list of
+        events to be run are present inside of a mutable collection
+        that can't be changed during iteration.
+
+        Event registration and removal is not intended to be a "high
+        velocity" operation; it is a configurational operation.  For
+        systems that need to quickly associate and deassociate with
+        events at high scale, use a mutable structure that is handled
+        from inside of a single listener.
+
+    .. seealso::
+
+        :func:`.listens_for`
+
+        :func:`.remove`
+
+    """
+
+    _event_key(target, identifier, fn).listen(*args, **kw)
+
+
+def listens_for(
+    target: Any, identifier: str, *args: Any, **kw: Any
+) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
+    """Decorate a function as a listener for the given target + identifier.
+
+    The :func:`.listens_for` decorator is part of the primary interface for the
+    SQLAlchemy event system, documented at :ref:`event_toplevel`.
+
+    This function generally shares the same kwargs as :func:`.listen`.
+
+    e.g.::
+
+        from sqlalchemy import event
+        from sqlalchemy.schema import UniqueConstraint
+
+
+        @event.listens_for(UniqueConstraint, "after_parent_attach")
+        def unique_constraint_name(const, table):
+            const.name = "uq_%s_%s" % (table.name, list(const.columns)[0].name)
+
+    A given function can also be invoked for only the first invocation
+    of the event using the ``once`` argument::
+
+        @event.listens_for(Mapper, "before_configure", once=True)
+        def on_config():
+            do_config()
+
+    .. warning:: The ``once`` argument does not imply automatic de-registration
+       of the listener function after it has been invoked a first time; a
+       listener entry will remain associated with the target object.
+       Associating an arbitrarily high number of listeners without explicitly
+       removing them will cause memory to grow unbounded even if ``once=True``
+       is specified.
+
+    .. seealso::
+
+        :func:`.listen` - general description of event listening
+
+    """
+
+    def decorate(fn: Callable[..., Any]) -> Callable[..., Any]:
+        listen(target, identifier, fn, *args, **kw)
+        return fn
+
+    return decorate
+
+
+def remove(target: Any, identifier: str, fn: Callable[..., Any]) -> None:
+    """Remove an event listener.
+
+    The arguments here should match exactly those which were sent to
+    :func:`.listen`; all the event registration which proceeded as a result
+    of this call will be reverted by calling :func:`.remove` with the same
+    arguments.
+
+    e.g.::
+
+        # if a function was registered like this...
+        @event.listens_for(SomeMappedClass, "before_insert", propagate=True)
+        def my_listener_function(*arg):
+            pass
+
+
+        # ... it's removed like this
+        event.remove(SomeMappedClass, "before_insert", my_listener_function)
+
+    Above, the listener function associated with ``SomeMappedClass`` was also
+    propagated to subclasses of ``SomeMappedClass``; the :func:`.remove`
+    function will revert all of these operations.
+
+    .. note::
+
+        The :func:`.remove` function cannot be called at the same time
+        that the target event is being run.   This has implications
+        for thread safety, and also means an event cannot be removed
+        from inside the listener function for itself.  The list of
+        events to be run are present inside of a mutable collection
+        that can't be changed during iteration.
+
+        Event registration and removal is not intended to be a "high
+        velocity" operation; it is a configurational operation.  For
+        systems that need to quickly associate and deassociate with
+        events at high scale, use a mutable structure that is handled
+        from inside of a single listener.
+
+    .. seealso::
+
+        :func:`.listen`
+
+    """
+    _event_key(target, identifier, fn).remove()
+
+
+def contains(target: Any, identifier: str, fn: Callable[..., Any]) -> bool:
+    """Return True if the given target/ident/fn is set up to listen."""
+
+    return _event_key(target, identifier, fn).contains()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/attr.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/attr.py
new file mode 100644
index 00000000..ec5d5822
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/attr.py
@@ -0,0 +1,655 @@
+# event/attr.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Attribute implementation for _Dispatch classes.
+
+The various listener targets for a particular event class are represented
+as attributes, which refer to collections of listeners to be fired off.
+These collections can exist at the class level as well as at the instance
+level.  An event is fired off using code like this::
+
+    some_object.dispatch.first_connect(arg1, arg2)
+
+Above, ``some_object.dispatch`` would be an instance of ``_Dispatch`` and
+``first_connect`` is typically an instance of ``_ListenerCollection``
+if event listeners are present, or ``_EmptyListener`` if none are present.
+
+The attribute mechanics here spend effort trying to ensure listener functions
+are available with a minimum of function call overhead, that unnecessary
+objects aren't created (i.e. many empty per-instance listener collections),
+as well as that everything is garbage collectable when owning references are
+lost.  Other features such as "propagation" of listener functions across
+many ``_Dispatch`` instances, "joining" of multiple ``_Dispatch`` instances,
+as well as support for subclass propagation (e.g. events assigned to
+``Pool`` vs. ``QueuePool``) are all implemented here.
+
+"""
+from __future__ import annotations
+
+import collections
+from itertools import chain
+import threading
+from types import TracebackType
+import typing
+from typing import Any
+from typing import cast
+from typing import Collection
+from typing import Deque
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterator
+from typing import MutableMapping
+from typing import MutableSequence
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import legacy
+from . import registry
+from .registry import _ET
+from .registry import _EventKey
+from .registry import _ListenerFnType
+from .. import exc
+from .. import util
+from ..util.concurrency import AsyncAdaptedLock
+from ..util.typing import Protocol
+
+_T = TypeVar("_T", bound=Any)
+
+if typing.TYPE_CHECKING:
+    from .base import _Dispatch
+    from .base import _DispatchCommon
+    from .base import _HasEventsDispatch
+
+
+class RefCollection(util.MemoizedSlots, Generic[_ET]):
+    __slots__ = ("ref",)
+
+    ref: weakref.ref[RefCollection[_ET]]
+
+    def _memoized_attr_ref(self) -> weakref.ref[RefCollection[_ET]]:
+        return weakref.ref(self, registry._collection_gced)
+
+
+class _empty_collection(Collection[_T]):
+    def append(self, element: _T) -> None:
+        pass
+
+    def appendleft(self, element: _T) -> None:
+        pass
+
+    def extend(self, other: Sequence[_T]) -> None:
+        pass
+
+    def remove(self, element: _T) -> None:
+        pass
+
+    def __contains__(self, element: Any) -> bool:
+        return False
+
+    def __iter__(self) -> Iterator[_T]:
+        return iter([])
+
+    def clear(self) -> None:
+        pass
+
+    def __len__(self) -> int:
+        return 0
+
+
+_ListenerFnSequenceType = Union[Deque[_T], _empty_collection[_T]]
+
+
+class _ClsLevelDispatch(RefCollection[_ET]):
+    """Class-level events on :class:`._Dispatch` classes."""
+
+    __slots__ = (
+        "clsname",
+        "name",
+        "arg_names",
+        "has_kw",
+        "legacy_signatures",
+        "_clslevel",
+        "__weakref__",
+    )
+
+    clsname: str
+    name: str
+    arg_names: Sequence[str]
+    has_kw: bool
+    legacy_signatures: MutableSequence[legacy._LegacySignatureType]
+    _clslevel: MutableMapping[
+        Type[_ET], _ListenerFnSequenceType[_ListenerFnType]
+    ]
+
+    def __init__(
+        self,
+        parent_dispatch_cls: Type[_HasEventsDispatch[_ET]],
+        fn: _ListenerFnType,
+    ):
+        self.name = fn.__name__
+        self.clsname = parent_dispatch_cls.__name__
+        argspec = util.inspect_getfullargspec(fn)
+        self.arg_names = argspec.args[1:]
+        self.has_kw = bool(argspec.varkw)
+        self.legacy_signatures = list(
+            reversed(
+                sorted(
+                    getattr(fn, "_legacy_signatures", []), key=lambda s: s[0]
+                )
+            )
+        )
+        fn.__doc__ = legacy._augment_fn_docs(self, parent_dispatch_cls, fn)
+
+        self._clslevel = weakref.WeakKeyDictionary()
+
+    def _adjust_fn_spec(
+        self, fn: _ListenerFnType, named: bool
+    ) -> _ListenerFnType:
+        if named:
+            fn = self._wrap_fn_for_kw(fn)
+        if self.legacy_signatures:
+            try:
+                argspec = util.get_callable_argspec(fn, no_self=True)
+            except TypeError:
+                pass
+            else:
+                fn = legacy._wrap_fn_for_legacy(self, fn, argspec)
+        return fn
+
+    def _wrap_fn_for_kw(self, fn: _ListenerFnType) -> _ListenerFnType:
+        def wrap_kw(*args: Any, **kw: Any) -> Any:
+            argdict = dict(zip(self.arg_names, args))
+            argdict.update(kw)
+            return fn(**argdict)
+
+        return wrap_kw
+
+    def _do_insert_or_append(
+        self, event_key: _EventKey[_ET], is_append: bool
+    ) -> None:
+        target = event_key.dispatch_target
+        assert isinstance(
+            target, type
+        ), "Class-level Event targets must be classes."
+        if not getattr(target, "_sa_propagate_class_events", True):
+            raise exc.InvalidRequestError(
+                f"Can't assign an event directly to the {target} class"
+            )
+
+        cls: Type[_ET]
+
+        for cls in util.walk_subclasses(target):
+            if cls is not target and cls not in self._clslevel:
+                self.update_subclass(cls)
+            else:
+                if cls not in self._clslevel:
+                    self.update_subclass(cls)
+                if is_append:
+                    self._clslevel[cls].append(event_key._listen_fn)
+                else:
+                    self._clslevel[cls].appendleft(event_key._listen_fn)
+        registry._stored_in_collection(event_key, self)
+
+    def insert(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        self._do_insert_or_append(event_key, is_append=False)
+
+    def append(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        self._do_insert_or_append(event_key, is_append=True)
+
+    def update_subclass(self, target: Type[_ET]) -> None:
+        if target not in self._clslevel:
+            if getattr(target, "_sa_propagate_class_events", True):
+                self._clslevel[target] = collections.deque()
+            else:
+                self._clslevel[target] = _empty_collection()
+
+        clslevel = self._clslevel[target]
+        cls: Type[_ET]
+        for cls in target.__mro__[1:]:
+            if cls in self._clslevel:
+                clslevel.extend(
+                    [fn for fn in self._clslevel[cls] if fn not in clslevel]
+                )
+
+    def remove(self, event_key: _EventKey[_ET]) -> None:
+        target = event_key.dispatch_target
+        cls: Type[_ET]
+        for cls in util.walk_subclasses(target):
+            if cls in self._clslevel:
+                self._clslevel[cls].remove(event_key._listen_fn)
+        registry._removed_from_collection(event_key, self)
+
+    def clear(self) -> None:
+        """Clear all class level listeners"""
+
+        to_clear: Set[_ListenerFnType] = set()
+        for dispatcher in self._clslevel.values():
+            to_clear.update(dispatcher)
+            dispatcher.clear()
+        registry._clear(self, to_clear)
+
+    def for_modify(self, obj: _Dispatch[_ET]) -> _ClsLevelDispatch[_ET]:
+        """Return an event collection which can be modified.
+
+        For _ClsLevelDispatch at the class level of
+        a dispatcher, this returns self.
+
+        """
+        return self
+
+
+class _InstanceLevelDispatch(RefCollection[_ET], Collection[_ListenerFnType]):
+    __slots__ = ()
+
+    parent: _ClsLevelDispatch[_ET]
+
+    def _adjust_fn_spec(
+        self, fn: _ListenerFnType, named: bool
+    ) -> _ListenerFnType:
+        return self.parent._adjust_fn_spec(fn, named)
+
+    def __contains__(self, item: Any) -> bool:
+        raise NotImplementedError()
+
+    def __len__(self) -> int:
+        raise NotImplementedError()
+
+    def __iter__(self) -> Iterator[_ListenerFnType]:
+        raise NotImplementedError()
+
+    def __bool__(self) -> bool:
+        raise NotImplementedError()
+
+    def exec_once(self, *args: Any, **kw: Any) -> None:
+        raise NotImplementedError()
+
+    def exec_once_unless_exception(self, *args: Any, **kw: Any) -> None:
+        raise NotImplementedError()
+
+    def _exec_w_sync_on_first_run(self, *args: Any, **kw: Any) -> None:
+        raise NotImplementedError()
+
+    def __call__(self, *args: Any, **kw: Any) -> None:
+        raise NotImplementedError()
+
+    def insert(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        raise NotImplementedError()
+
+    def append(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        raise NotImplementedError()
+
+    def remove(self, event_key: _EventKey[_ET]) -> None:
+        raise NotImplementedError()
+
+    def for_modify(
+        self, obj: _DispatchCommon[_ET]
+    ) -> _InstanceLevelDispatch[_ET]:
+        """Return an event collection which can be modified.
+
+        For _ClsLevelDispatch at the class level of
+        a dispatcher, this returns self.
+
+        """
+        return self
+
+
+class _EmptyListener(_InstanceLevelDispatch[_ET]):
+    """Serves as a proxy interface to the events
+    served by a _ClsLevelDispatch, when there are no
+    instance-level events present.
+
+    Is replaced by _ListenerCollection when instance-level
+    events are added.
+
+    """
+
+    __slots__ = "parent", "parent_listeners", "name"
+
+    propagate: FrozenSet[_ListenerFnType] = frozenset()
+    listeners: Tuple[()] = ()
+    parent: _ClsLevelDispatch[_ET]
+    parent_listeners: _ListenerFnSequenceType[_ListenerFnType]
+    name: str
+
+    def __init__(self, parent: _ClsLevelDispatch[_ET], target_cls: Type[_ET]):
+        if target_cls not in parent._clslevel:
+            parent.update_subclass(target_cls)
+        self.parent = parent
+        self.parent_listeners = parent._clslevel[target_cls]
+        self.name = parent.name
+
+    def for_modify(
+        self, obj: _DispatchCommon[_ET]
+    ) -> _ListenerCollection[_ET]:
+        """Return an event collection which can be modified.
+
+        For _EmptyListener at the instance level of
+        a dispatcher, this generates a new
+        _ListenerCollection, applies it to the instance,
+        and returns it.
+
+        """
+        obj = cast("_Dispatch[_ET]", obj)
+
+        assert obj._instance_cls is not None
+        result = _ListenerCollection(self.parent, obj._instance_cls)
+        if getattr(obj, self.name) is self:
+            setattr(obj, self.name, result)
+        else:
+            assert isinstance(getattr(obj, self.name), _JoinedListener)
+        return result
+
+    def _needs_modify(self, *args: Any, **kw: Any) -> NoReturn:
+        raise NotImplementedError("need to call for_modify()")
+
+    def exec_once(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def exec_once_unless_exception(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def insert(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def append(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def remove(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def clear(self, *args: Any, **kw: Any) -> NoReturn:
+        self._needs_modify(*args, **kw)
+
+    def __call__(self, *args: Any, **kw: Any) -> None:
+        """Execute this event."""
+
+        for fn in self.parent_listeners:
+            fn(*args, **kw)
+
+    def __contains__(self, item: Any) -> bool:
+        return item in self.parent_listeners
+
+    def __len__(self) -> int:
+        return len(self.parent_listeners)
+
+    def __iter__(self) -> Iterator[_ListenerFnType]:
+        return iter(self.parent_listeners)
+
+    def __bool__(self) -> bool:
+        return bool(self.parent_listeners)
+
+
+class _MutexProtocol(Protocol):
+    def __enter__(self) -> bool: ...
+
+    def __exit__(
+        self,
+        exc_type: Optional[Type[BaseException]],
+        exc_val: Optional[BaseException],
+        exc_tb: Optional[TracebackType],
+    ) -> Optional[bool]: ...
+
+
+class _CompoundListener(_InstanceLevelDispatch[_ET]):
+    __slots__ = (
+        "_exec_once_mutex",
+        "_exec_once",
+        "_exec_w_sync_once",
+        "_is_asyncio",
+    )
+
+    _exec_once_mutex: _MutexProtocol
+    parent_listeners: Collection[_ListenerFnType]
+    listeners: Collection[_ListenerFnType]
+    _exec_once: bool
+    _exec_w_sync_once: bool
+
+    def __init__(self, *arg: Any, **kw: Any):
+        super().__init__(*arg, **kw)
+        self._is_asyncio = False
+
+    def _set_asyncio(self) -> None:
+        self._is_asyncio = True
+
+    def _memoized_attr__exec_once_mutex(self) -> _MutexProtocol:
+        if self._is_asyncio:
+            return AsyncAdaptedLock()
+        else:
+            return threading.Lock()
+
+    def _exec_once_impl(
+        self, retry_on_exception: bool, *args: Any, **kw: Any
+    ) -> None:
+        with self._exec_once_mutex:
+            if not self._exec_once:
+                try:
+                    self(*args, **kw)
+                    exception = False
+                except:
+                    exception = True
+                    raise
+                finally:
+                    if not exception or not retry_on_exception:
+                        self._exec_once = True
+
+    def exec_once(self, *args: Any, **kw: Any) -> None:
+        """Execute this event, but only if it has not been
+        executed already for this collection."""
+
+        if not self._exec_once:
+            self._exec_once_impl(False, *args, **kw)
+
+    def exec_once_unless_exception(self, *args: Any, **kw: Any) -> None:
+        """Execute this event, but only if it has not been
+        executed already for this collection, or was called
+        by a previous exec_once_unless_exception call and
+        raised an exception.
+
+        If exec_once was already called, then this method will never run
+        the callable regardless of whether it raised or not.
+
+        .. versionadded:: 1.3.8
+
+        """
+        if not self._exec_once:
+            self._exec_once_impl(True, *args, **kw)
+
+    def _exec_w_sync_on_first_run(self, *args: Any, **kw: Any) -> None:
+        """Execute this event, and use a mutex if it has not been
+        executed already for this collection, or was called
+        by a previous _exec_w_sync_on_first_run call and
+        raised an exception.
+
+        If _exec_w_sync_on_first_run was already called and didn't raise an
+        exception, then a mutex is not used.
+
+        .. versionadded:: 1.4.11
+
+        """
+        if not self._exec_w_sync_once:
+            with self._exec_once_mutex:
+                try:
+                    self(*args, **kw)
+                except:
+                    raise
+                else:
+                    self._exec_w_sync_once = True
+        else:
+            self(*args, **kw)
+
+    def __call__(self, *args: Any, **kw: Any) -> None:
+        """Execute this event."""
+
+        for fn in self.parent_listeners:
+            fn(*args, **kw)
+        for fn in self.listeners:
+            fn(*args, **kw)
+
+    def __contains__(self, item: Any) -> bool:
+        return item in self.parent_listeners or item in self.listeners
+
+    def __len__(self) -> int:
+        return len(self.parent_listeners) + len(self.listeners)
+
+    def __iter__(self) -> Iterator[_ListenerFnType]:
+        return chain(self.parent_listeners, self.listeners)
+
+    def __bool__(self) -> bool:
+        return bool(self.listeners or self.parent_listeners)
+
+
+class _ListenerCollection(_CompoundListener[_ET]):
+    """Instance-level attributes on instances of :class:`._Dispatch`.
+
+    Represents a collection of listeners.
+
+    As of 0.7.9, _ListenerCollection is only first
+    created via the _EmptyListener.for_modify() method.
+
+    """
+
+    __slots__ = (
+        "parent_listeners",
+        "parent",
+        "name",
+        "listeners",
+        "propagate",
+        "__weakref__",
+    )
+
+    parent_listeners: Collection[_ListenerFnType]
+    parent: _ClsLevelDispatch[_ET]
+    name: str
+    listeners: Deque[_ListenerFnType]
+    propagate: Set[_ListenerFnType]
+
+    def __init__(self, parent: _ClsLevelDispatch[_ET], target_cls: Type[_ET]):
+        super().__init__()
+        if target_cls not in parent._clslevel:
+            parent.update_subclass(target_cls)
+        self._exec_once = False
+        self._exec_w_sync_once = False
+        self.parent_listeners = parent._clslevel[target_cls]
+        self.parent = parent
+        self.name = parent.name
+        self.listeners = collections.deque()
+        self.propagate = set()
+
+    def for_modify(
+        self, obj: _DispatchCommon[_ET]
+    ) -> _ListenerCollection[_ET]:
+        """Return an event collection which can be modified.
+
+        For _ListenerCollection at the instance level of
+        a dispatcher, this returns self.
+
+        """
+        return self
+
+    def _update(
+        self, other: _ListenerCollection[_ET], only_propagate: bool = True
+    ) -> None:
+        """Populate from the listeners in another :class:`_Dispatch`
+        object."""
+        existing_listeners = self.listeners
+        existing_listener_set = set(existing_listeners)
+        self.propagate.update(other.propagate)
+        other_listeners = [
+            l
+            for l in other.listeners
+            if l not in existing_listener_set
+            and not only_propagate
+            or l in self.propagate
+        ]
+
+        existing_listeners.extend(other_listeners)
+
+        if other._is_asyncio:
+            self._set_asyncio()
+
+        to_associate = other.propagate.union(other_listeners)
+        registry._stored_in_collection_multi(self, other, to_associate)
+
+    def insert(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        if event_key.prepend_to_list(self, self.listeners):
+            if propagate:
+                self.propagate.add(event_key._listen_fn)
+
+    def append(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        if event_key.append_to_list(self, self.listeners):
+            if propagate:
+                self.propagate.add(event_key._listen_fn)
+
+    def remove(self, event_key: _EventKey[_ET]) -> None:
+        self.listeners.remove(event_key._listen_fn)
+        self.propagate.discard(event_key._listen_fn)
+        registry._removed_from_collection(event_key, self)
+
+    def clear(self) -> None:
+        registry._clear(self, self.listeners)
+        self.propagate.clear()
+        self.listeners.clear()
+
+
+class _JoinedListener(_CompoundListener[_ET]):
+    __slots__ = "parent_dispatch", "name", "local", "parent_listeners"
+
+    parent_dispatch: _DispatchCommon[_ET]
+    name: str
+    local: _InstanceLevelDispatch[_ET]
+    parent_listeners: Collection[_ListenerFnType]
+
+    def __init__(
+        self,
+        parent_dispatch: _DispatchCommon[_ET],
+        name: str,
+        local: _EmptyListener[_ET],
+    ):
+        self._exec_once = False
+        self.parent_dispatch = parent_dispatch
+        self.name = name
+        self.local = local
+        self.parent_listeners = self.local
+
+    if not typing.TYPE_CHECKING:
+        # first error, I don't really understand:
+        # Signature of "listeners" incompatible with
+        # supertype "_CompoundListener"  [override]
+        # the name / return type are exactly the same
+        # second error is getattr_isn't typed, the cast() here
+        # adds too much method overhead
+        @property
+        def listeners(self) -> Collection[_ListenerFnType]:
+            return getattr(self.parent_dispatch, self.name)
+
+    def _adjust_fn_spec(
+        self, fn: _ListenerFnType, named: bool
+    ) -> _ListenerFnType:
+        return self.local._adjust_fn_spec(fn, named)
+
+    def for_modify(self, obj: _DispatchCommon[_ET]) -> _JoinedListener[_ET]:
+        self.local = self.parent_listeners = self.local.for_modify(obj)
+        return self
+
+    def insert(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        self.local.insert(event_key, propagate)
+
+    def append(self, event_key: _EventKey[_ET], propagate: bool) -> None:
+        self.local.append(event_key, propagate)
+
+    def remove(self, event_key: _EventKey[_ET]) -> None:
+        self.local.remove(event_key)
+
+    def clear(self) -> None:
+        raise NotImplementedError()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/base.py
new file mode 100644
index 00000000..66dc1299
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/base.py
@@ -0,0 +1,472 @@
+# event/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Base implementation classes.
+
+The public-facing ``Events`` serves as the base class for an event interface;
+its public attributes represent different kinds of events.   These attributes
+are mirrored onto a ``_Dispatch`` class, which serves as a container for
+collections of listener functions.   These collections are represented both
+at the class level of a particular ``_Dispatch`` class as well as within
+instances of ``_Dispatch``.
+
+"""
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import Union
+import weakref
+
+from .attr import _ClsLevelDispatch
+from .attr import _EmptyListener
+from .attr import _InstanceLevelDispatch
+from .attr import _JoinedListener
+from .registry import _ET
+from .registry import _EventKey
+from .. import util
+from ..util.typing import Literal
+
+_registrars: MutableMapping[str, List[Type[_HasEventsDispatch[Any]]]] = (
+    util.defaultdict(list)
+)
+
+
+def _is_event_name(name: str) -> bool:
+    # _sa_event prefix is special to support internal-only event names.
+    # most event names are just plain method names that aren't
+    # underscored.
+
+    return (
+        not name.startswith("_") and name != "dispatch"
+    ) or name.startswith("_sa_event")
+
+
+class _UnpickleDispatch:
+    """Serializable callable that re-generates an instance of
+    :class:`_Dispatch` given a particular :class:`.Events` subclass.
+
+    """
+
+    def __call__(self, _instance_cls: Type[_ET]) -> _Dispatch[_ET]:
+        for cls in _instance_cls.__mro__:
+            if "dispatch" in cls.__dict__:
+                return cast(
+                    "_Dispatch[_ET]", cls.__dict__["dispatch"].dispatch
+                )._for_class(_instance_cls)
+        else:
+            raise AttributeError("No class with a 'dispatch' member present.")
+
+
+class _DispatchCommon(Generic[_ET]):
+    __slots__ = ()
+
+    _instance_cls: Optional[Type[_ET]]
+
+    def _join(self, other: _DispatchCommon[_ET]) -> _JoinedDispatcher[_ET]:
+        raise NotImplementedError()
+
+    def __getattr__(self, name: str) -> _InstanceLevelDispatch[_ET]:
+        raise NotImplementedError()
+
+    @property
+    def _events(self) -> Type[_HasEventsDispatch[_ET]]:
+        raise NotImplementedError()
+
+
+class _Dispatch(_DispatchCommon[_ET]):
+    """Mirror the event listening definitions of an Events class with
+    listener collections.
+
+    Classes which define a "dispatch" member will return a
+    non-instantiated :class:`._Dispatch` subclass when the member
+    is accessed at the class level.  When the "dispatch" member is
+    accessed at the instance level of its owner, an instance
+    of the :class:`._Dispatch` class is returned.
+
+    A :class:`._Dispatch` class is generated for each :class:`.Events`
+    class defined, by the :meth:`._HasEventsDispatch._create_dispatcher_class`
+    method.  The original :class:`.Events` classes remain untouched.
+    This decouples the construction of :class:`.Events` subclasses from
+    the implementation used by the event internals, and allows
+    inspecting tools like Sphinx to work in an unsurprising
+    way against the public API.
+
+    """
+
+    # "active_history" is an ORM case we add here.   ideally a better
+    # system would be in place for ad-hoc attributes.
+    __slots__ = "_parent", "_instance_cls", "__dict__", "_empty_listeners"
+
+    _active_history: bool
+
+    _empty_listener_reg: MutableMapping[
+        Type[_ET], Dict[str, _EmptyListener[_ET]]
+    ] = weakref.WeakKeyDictionary()
+
+    _empty_listeners: Dict[str, _EmptyListener[_ET]]
+
+    _event_names: List[str]
+
+    _instance_cls: Optional[Type[_ET]]
+
+    _joined_dispatch_cls: Type[_JoinedDispatcher[_ET]]
+
+    _events: Type[_HasEventsDispatch[_ET]]
+    """reference back to the Events class.
+
+    Bidirectional against _HasEventsDispatch.dispatch
+
+    """
+
+    def __init__(
+        self,
+        parent: Optional[_Dispatch[_ET]],
+        instance_cls: Optional[Type[_ET]] = None,
+    ):
+        self._parent = parent
+        self._instance_cls = instance_cls
+
+        if instance_cls:
+            assert parent is not None
+            try:
+                self._empty_listeners = self._empty_listener_reg[instance_cls]
+            except KeyError:
+                self._empty_listeners = self._empty_listener_reg[
+                    instance_cls
+                ] = {
+                    ls.name: _EmptyListener(ls, instance_cls)
+                    for ls in parent._event_descriptors
+                }
+        else:
+            self._empty_listeners = {}
+
+    def __getattr__(self, name: str) -> _InstanceLevelDispatch[_ET]:
+        # Assign EmptyListeners as attributes on demand
+        # to reduce startup time for new dispatch objects.
+        try:
+            ls = self._empty_listeners[name]
+        except KeyError:
+            raise AttributeError(name)
+        else:
+            setattr(self, ls.name, ls)
+            return ls
+
+    @property
+    def _event_descriptors(self) -> Iterator[_ClsLevelDispatch[_ET]]:
+        for k in self._event_names:
+            # Yield _ClsLevelDispatch related
+            # to relevant event name.
+            yield getattr(self, k)
+
+    def _listen(self, event_key: _EventKey[_ET], **kw: Any) -> None:
+        return self._events._listen(event_key, **kw)
+
+    def _for_class(self, instance_cls: Type[_ET]) -> _Dispatch[_ET]:
+        return self.__class__(self, instance_cls)
+
+    def _for_instance(self, instance: _ET) -> _Dispatch[_ET]:
+        instance_cls = instance.__class__
+        return self._for_class(instance_cls)
+
+    def _join(self, other: _DispatchCommon[_ET]) -> _JoinedDispatcher[_ET]:
+        """Create a 'join' of this :class:`._Dispatch` and another.
+
+        This new dispatcher will dispatch events to both
+        :class:`._Dispatch` objects.
+
+        """
+        assert "_joined_dispatch_cls" in self.__class__.__dict__
+
+        return self._joined_dispatch_cls(self, other)
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return _UnpickleDispatch(), (self._instance_cls,)
+
+    def _update(
+        self, other: _Dispatch[_ET], only_propagate: bool = True
+    ) -> None:
+        """Populate from the listeners in another :class:`_Dispatch`
+        object."""
+        for ls in other._event_descriptors:
+            if isinstance(ls, _EmptyListener):
+                continue
+            getattr(self, ls.name).for_modify(self)._update(
+                ls, only_propagate=only_propagate
+            )
+
+    def _clear(self) -> None:
+        for ls in self._event_descriptors:
+            ls.for_modify(self).clear()
+
+
+def _remove_dispatcher(cls: Type[_HasEventsDispatch[_ET]]) -> None:
+    for k in cls.dispatch._event_names:
+        _registrars[k].remove(cls)
+        if not _registrars[k]:
+            del _registrars[k]
+
+
+class _HasEventsDispatch(Generic[_ET]):
+    _dispatch_target: Optional[Type[_ET]]
+    """class which will receive the .dispatch collection"""
+
+    dispatch: _Dispatch[_ET]
+    """reference back to the _Dispatch class.
+
+    Bidirectional against _Dispatch._events
+
+    """
+
+    if typing.TYPE_CHECKING:
+
+        def __getattr__(self, name: str) -> _InstanceLevelDispatch[_ET]: ...
+
+    def __init_subclass__(cls) -> None:
+        """Intercept new Event subclasses and create associated _Dispatch
+        classes."""
+
+        cls._create_dispatcher_class(cls.__name__, cls.__bases__, cls.__dict__)
+
+    @classmethod
+    def _accept_with(
+        cls, target: Union[_ET, Type[_ET]], identifier: str
+    ) -> Optional[Union[_ET, Type[_ET]]]:
+        raise NotImplementedError()
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: _EventKey[_ET],
+        *,
+        propagate: bool = False,
+        insert: bool = False,
+        named: bool = False,
+        asyncio: bool = False,
+    ) -> None:
+        raise NotImplementedError()
+
+    @staticmethod
+    def _set_dispatch(
+        klass: Type[_HasEventsDispatch[_ET]],
+        dispatch_cls: Type[_Dispatch[_ET]],
+    ) -> _Dispatch[_ET]:
+        # This allows an Events subclass to define additional utility
+        # methods made available to the target via
+        # "self.dispatch._events.<utilitymethod>"
+        # @staticmethod to allow easy "super" calls while in a metaclass
+        # constructor.
+        klass.dispatch = dispatch_cls(None)
+        dispatch_cls._events = klass
+        return klass.dispatch
+
+    @classmethod
+    def _create_dispatcher_class(
+        cls, classname: str, bases: Tuple[type, ...], dict_: Mapping[str, Any]
+    ) -> None:
+        """Create a :class:`._Dispatch` class corresponding to an
+        :class:`.Events` class."""
+
+        # there's all kinds of ways to do this,
+        # i.e. make a Dispatch class that shares the '_listen' method
+        # of the Event class, this is the straight monkeypatch.
+        if hasattr(cls, "dispatch"):
+            dispatch_base = cls.dispatch.__class__
+        else:
+            dispatch_base = _Dispatch
+
+        event_names = [k for k in dict_ if _is_event_name(k)]
+        dispatch_cls = cast(
+            "Type[_Dispatch[_ET]]",
+            type(
+                "%sDispatch" % classname,
+                (dispatch_base,),
+                {"__slots__": event_names},
+            ),
+        )
+
+        dispatch_cls._event_names = event_names
+        dispatch_inst = cls._set_dispatch(cls, dispatch_cls)
+        for k in dispatch_cls._event_names:
+            setattr(dispatch_inst, k, _ClsLevelDispatch(cls, dict_[k]))
+            _registrars[k].append(cls)
+
+        for super_ in dispatch_cls.__bases__:
+            if issubclass(super_, _Dispatch) and super_ is not _Dispatch:
+                for ls in super_._events.dispatch._event_descriptors:
+                    setattr(dispatch_inst, ls.name, ls)
+                    dispatch_cls._event_names.append(ls.name)
+
+        if getattr(cls, "_dispatch_target", None):
+            dispatch_target_cls = cls._dispatch_target
+            assert dispatch_target_cls is not None
+            if (
+                hasattr(dispatch_target_cls, "__slots__")
+                and "_slots_dispatch" in dispatch_target_cls.__slots__
+            ):
+                dispatch_target_cls.dispatch = slots_dispatcher(cls)
+            else:
+                dispatch_target_cls.dispatch = dispatcher(cls)
+
+        klass = type(
+            "Joined%s" % dispatch_cls.__name__,
+            (_JoinedDispatcher,),
+            {"__slots__": event_names},
+        )
+        dispatch_cls._joined_dispatch_cls = klass
+
+        # establish pickle capability by adding it to this module
+        globals()[klass.__name__] = klass
+
+
+class _JoinedDispatcher(_DispatchCommon[_ET]):
+    """Represent a connection between two _Dispatch objects."""
+
+    __slots__ = "local", "parent", "_instance_cls"
+
+    local: _DispatchCommon[_ET]
+    parent: _DispatchCommon[_ET]
+    _instance_cls: Optional[Type[_ET]]
+
+    def __init__(
+        self, local: _DispatchCommon[_ET], parent: _DispatchCommon[_ET]
+    ):
+        self.local = local
+        self.parent = parent
+        self._instance_cls = self.local._instance_cls
+
+    def __reduce__(self) -> Any:
+        return (self.__class__, (self.local, self.parent))
+
+    def __getattr__(self, name: str) -> _JoinedListener[_ET]:
+        # Assign _JoinedListeners as attributes on demand
+        # to reduce startup time for new dispatch objects.
+        ls = getattr(self.local, name)
+        jl = _JoinedListener(self.parent, ls.name, ls)
+        setattr(self, ls.name, jl)
+        return jl
+
+    def _listen(self, event_key: _EventKey[_ET], **kw: Any) -> None:
+        return self.parent._listen(event_key, **kw)
+
+    @property
+    def _events(self) -> Type[_HasEventsDispatch[_ET]]:
+        return self.parent._events
+
+
+class Events(_HasEventsDispatch[_ET]):
+    """Define event listening functions for a particular target type."""
+
+    @classmethod
+    def _accept_with(
+        cls, target: Union[_ET, Type[_ET]], identifier: str
+    ) -> Optional[Union[_ET, Type[_ET]]]:
+        def dispatch_is(*types: Type[Any]) -> bool:
+            return all(isinstance(target.dispatch, t) for t in types)
+
+        def dispatch_parent_is(t: Type[Any]) -> bool:
+            parent = cast("_JoinedDispatcher[_ET]", target.dispatch).parent
+            while isinstance(parent, _JoinedDispatcher):
+                parent = cast("_JoinedDispatcher[_ET]", parent).parent
+
+            return isinstance(parent, t)
+
+        # Mapper, ClassManager, Session override this to
+        # also accept classes, scoped_sessions, sessionmakers, etc.
+        if hasattr(target, "dispatch"):
+            if (
+                dispatch_is(cls.dispatch.__class__)
+                or dispatch_is(type, cls.dispatch.__class__)
+                or (
+                    dispatch_is(_JoinedDispatcher)
+                    and dispatch_parent_is(cls.dispatch.__class__)
+                )
+            ):
+                return target
+
+        return None
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: _EventKey[_ET],
+        *,
+        propagate: bool = False,
+        insert: bool = False,
+        named: bool = False,
+        asyncio: bool = False,
+    ) -> None:
+        event_key.base_listen(
+            propagate=propagate, insert=insert, named=named, asyncio=asyncio
+        )
+
+    @classmethod
+    def _remove(cls, event_key: _EventKey[_ET]) -> None:
+        event_key.remove()
+
+    @classmethod
+    def _clear(cls) -> None:
+        cls.dispatch._clear()
+
+
+class dispatcher(Generic[_ET]):
+    """Descriptor used by target classes to
+    deliver the _Dispatch class at the class level
+    and produce new _Dispatch instances for target
+    instances.
+
+    """
+
+    def __init__(self, events: Type[_HasEventsDispatch[_ET]]):
+        self.dispatch = events.dispatch
+        self.events = events
+
+    @overload
+    def __get__(
+        self, obj: Literal[None], cls: Type[Any]
+    ) -> Type[_Dispatch[_ET]]: ...
+
+    @overload
+    def __get__(self, obj: Any, cls: Type[Any]) -> _DispatchCommon[_ET]: ...
+
+    def __get__(self, obj: Any, cls: Type[Any]) -> Any:
+        if obj is None:
+            return self.dispatch
+
+        disp = self.dispatch._for_instance(obj)
+        try:
+            obj.__dict__["dispatch"] = disp
+        except AttributeError as ae:
+            raise TypeError(
+                "target %r doesn't have __dict__, should it be "
+                "defining _slots_dispatch?" % (obj,)
+            ) from ae
+        return disp
+
+
+class slots_dispatcher(dispatcher[_ET]):
+    def __get__(self, obj: Any, cls: Type[Any]) -> Any:
+        if obj is None:
+            return self.dispatch
+
+        if hasattr(obj, "_slots_dispatch"):
+            return obj._slots_dispatch
+
+        disp = self.dispatch._for_instance(obj)
+        obj._slots_dispatch = disp
+        return disp
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/legacy.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/legacy.py
new file mode 100644
index 00000000..e60fd9a5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/legacy.py
@@ -0,0 +1,246 @@
+# event/legacy.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Routines to handle adaption of legacy call signatures,
+generation of deprecation notes and docstrings.
+
+"""
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Type
+
+from .registry import _ET
+from .registry import _ListenerFnType
+from .. import util
+from ..util.compat import FullArgSpec
+
+if typing.TYPE_CHECKING:
+    from .attr import _ClsLevelDispatch
+    from .base import _HasEventsDispatch
+
+
+_LegacySignatureType = Tuple[str, List[str], Optional[Callable[..., Any]]]
+
+
+def _legacy_signature(
+    since: str,
+    argnames: List[str],
+    converter: Optional[Callable[..., Any]] = None,
+) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
+    """legacy sig decorator
+
+
+    :param since: string version for deprecation warning
+    :param argnames: list of strings, which is *all* arguments that the legacy
+     version accepted, including arguments that are still there
+    :param converter: lambda that will accept tuple of this full arg signature
+     and return tuple of new arg signature.
+
+    """
+
+    def leg(fn: Callable[..., Any]) -> Callable[..., Any]:
+        if not hasattr(fn, "_legacy_signatures"):
+            fn._legacy_signatures = []  # type: ignore[attr-defined]
+        fn._legacy_signatures.append((since, argnames, converter))  # type: ignore[attr-defined] # noqa: E501
+        return fn
+
+    return leg
+
+
+def _wrap_fn_for_legacy(
+    dispatch_collection: _ClsLevelDispatch[_ET],
+    fn: _ListenerFnType,
+    argspec: FullArgSpec,
+) -> _ListenerFnType:
+    for since, argnames, conv in dispatch_collection.legacy_signatures:
+        if argnames[-1] == "**kw":
+            has_kw = True
+            argnames = argnames[0:-1]
+        else:
+            has_kw = False
+
+        if len(argnames) == len(argspec.args) and has_kw is bool(
+            argspec.varkw
+        ):
+            formatted_def = "def %s(%s%s)" % (
+                dispatch_collection.name,
+                ", ".join(dispatch_collection.arg_names),
+                ", **kw" if has_kw else "",
+            )
+            warning_txt = (
+                'The argument signature for the "%s.%s" event listener '
+                "has changed as of version %s, and conversion for "
+                "the old argument signature will be removed in a "
+                'future release.  The new signature is "%s"'
+                % (
+                    dispatch_collection.clsname,
+                    dispatch_collection.name,
+                    since,
+                    formatted_def,
+                )
+            )
+
+            if conv is not None:
+                assert not has_kw
+
+                def wrap_leg(*args: Any, **kw: Any) -> Any:
+                    util.warn_deprecated(warning_txt, version=since)
+                    assert conv is not None
+                    return fn(*conv(*args))
+
+            else:
+
+                def wrap_leg(*args: Any, **kw: Any) -> Any:
+                    util.warn_deprecated(warning_txt, version=since)
+                    argdict = dict(zip(dispatch_collection.arg_names, args))
+                    args_from_dict = [argdict[name] for name in argnames]
+                    if has_kw:
+                        return fn(*args_from_dict, **kw)
+                    else:
+                        return fn(*args_from_dict)
+
+            return wrap_leg
+    else:
+        return fn
+
+
+def _indent(text: str, indent: str) -> str:
+    return "\n".join(indent + line for line in text.split("\n"))
+
+
+def _standard_listen_example(
+    dispatch_collection: _ClsLevelDispatch[_ET],
+    sample_target: Any,
+    fn: _ListenerFnType,
+) -> str:
+    example_kw_arg = _indent(
+        "\n".join(
+            "%(arg)s = kw['%(arg)s']" % {"arg": arg}
+            for arg in dispatch_collection.arg_names[0:2]
+        ),
+        "    ",
+    )
+    if dispatch_collection.legacy_signatures:
+        current_since = max(
+            since
+            for since, args, conv in dispatch_collection.legacy_signatures
+        )
+    else:
+        current_since = None
+    text = (
+        "from sqlalchemy import event\n\n\n"
+        "@event.listens_for(%(sample_target)s, '%(event_name)s')\n"
+        "def receive_%(event_name)s("
+        "%(named_event_arguments)s%(has_kw_arguments)s):\n"
+        "    \"listen for the '%(event_name)s' event\"\n"
+        "\n    # ... (event handling logic) ...\n"
+    )
+
+    text %= {
+        "current_since": (
+            " (arguments as of %s)" % current_since if current_since else ""
+        ),
+        "event_name": fn.__name__,
+        "has_kw_arguments": ", **kw" if dispatch_collection.has_kw else "",
+        "named_event_arguments": ", ".join(dispatch_collection.arg_names),
+        "example_kw_arg": example_kw_arg,
+        "sample_target": sample_target,
+    }
+    return text
+
+
+def _legacy_listen_examples(
+    dispatch_collection: _ClsLevelDispatch[_ET],
+    sample_target: str,
+    fn: _ListenerFnType,
+) -> str:
+    text = ""
+    for since, args, conv in dispatch_collection.legacy_signatures:
+        text += (
+            "\n# DEPRECATED calling style (pre-%(since)s, "
+            "will be removed in a future release)\n"
+            "@event.listens_for(%(sample_target)s, '%(event_name)s')\n"
+            "def receive_%(event_name)s("
+            "%(named_event_arguments)s%(has_kw_arguments)s):\n"
+            "    \"listen for the '%(event_name)s' event\"\n"
+            "\n    # ... (event handling logic) ...\n"
+            % {
+                "since": since,
+                "event_name": fn.__name__,
+                "has_kw_arguments": (
+                    " **kw" if dispatch_collection.has_kw else ""
+                ),
+                "named_event_arguments": ", ".join(args),
+                "sample_target": sample_target,
+            }
+        )
+    return text
+
+
+def _version_signature_changes(
+    parent_dispatch_cls: Type[_HasEventsDispatch[_ET]],
+    dispatch_collection: _ClsLevelDispatch[_ET],
+) -> str:
+    since, args, conv = dispatch_collection.legacy_signatures[0]
+    return (
+        "\n.. versionchanged:: %(since)s\n"
+        "    The :meth:`.%(clsname)s.%(event_name)s` event now accepts the \n"
+        "    arguments %(named_event_arguments)s%(has_kw_arguments)s.\n"
+        "    Support for listener functions which accept the previous \n"
+        '    argument signature(s) listed above as "deprecated" will be \n'
+        "    removed in a future release."
+        % {
+            "since": since,
+            "clsname": parent_dispatch_cls.__name__,
+            "event_name": dispatch_collection.name,
+            "named_event_arguments": ", ".join(
+                ":paramref:`.%(clsname)s.%(event_name)s.%(param_name)s`"
+                % {
+                    "clsname": parent_dispatch_cls.__name__,
+                    "event_name": dispatch_collection.name,
+                    "param_name": param_name,
+                }
+                for param_name in dispatch_collection.arg_names
+            ),
+            "has_kw_arguments": ", **kw" if dispatch_collection.has_kw else "",
+        }
+    )
+
+
+def _augment_fn_docs(
+    dispatch_collection: _ClsLevelDispatch[_ET],
+    parent_dispatch_cls: Type[_HasEventsDispatch[_ET]],
+    fn: _ListenerFnType,
+) -> str:
+    header = (
+        ".. container:: event_signatures\n\n"
+        "     Example argument forms::\n"
+        "\n"
+    )
+
+    sample_target = getattr(parent_dispatch_cls, "_target_class_doc", "obj")
+    text = header + _indent(
+        _standard_listen_example(dispatch_collection, sample_target, fn),
+        " " * 8,
+    )
+    if dispatch_collection.legacy_signatures:
+        text += _indent(
+            _legacy_listen_examples(dispatch_collection, sample_target, fn),
+            " " * 8,
+        )
+
+        text += _version_signature_changes(
+            parent_dispatch_cls, dispatch_collection
+        )
+
+    return util.inject_docstring_text(fn.__doc__, text, 1)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/registry.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/registry.py
new file mode 100644
index 00000000..d7e4b321
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/event/registry.py
@@ -0,0 +1,390 @@
+# event/registry.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Provides managed registration services on behalf of :func:`.listen`
+arguments.
+
+By "managed registration", we mean that event listening functions and
+other objects can be added to various collections in such a way that their
+membership in all those collections can be revoked at once, based on
+an equivalent :class:`._EventKey`.
+
+"""
+from __future__ import annotations
+
+import collections
+import types
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Deque
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Optional
+from typing import Tuple
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from .. import exc
+from .. import util
+
+if typing.TYPE_CHECKING:
+    from .attr import RefCollection
+    from .base import dispatcher
+
+_ListenerFnType = Callable[..., Any]
+_ListenerFnKeyType = Union[int, Tuple[int, int]]
+_EventKeyTupleType = Tuple[int, str, _ListenerFnKeyType]
+
+
+_ET = TypeVar("_ET", bound="EventTarget")
+
+
+class EventTarget:
+    """represents an event target, that is, something we can listen on
+    either with that target as a class or as an instance.
+
+    Examples include:  Connection, Mapper, Table, Session,
+    InstrumentedAttribute, Engine, Pool, Dialect.
+
+    """
+
+    __slots__ = ()
+
+    dispatch: dispatcher[Any]
+
+
+_RefCollectionToListenerType = Dict[
+    "weakref.ref[RefCollection[Any]]",
+    "weakref.ref[_ListenerFnType]",
+]
+
+_key_to_collection: Dict[_EventKeyTupleType, _RefCollectionToListenerType] = (
+    collections.defaultdict(dict)
+)
+"""
+Given an original listen() argument, can locate all
+listener collections and the listener fn contained
+
+(target, identifier, fn) -> {
+                            ref(listenercollection) -> ref(listener_fn)
+                            ref(listenercollection) -> ref(listener_fn)
+                            ref(listenercollection) -> ref(listener_fn)
+                        }
+"""
+
+_ListenerToEventKeyType = Dict[
+    "weakref.ref[_ListenerFnType]",
+    _EventKeyTupleType,
+]
+_collection_to_key: Dict[
+    weakref.ref[RefCollection[Any]],
+    _ListenerToEventKeyType,
+] = collections.defaultdict(dict)
+"""
+Given a _ListenerCollection or _ClsLevelListener, can locate
+all the original listen() arguments and the listener fn contained
+
+ref(listenercollection) -> {
+                            ref(listener_fn) -> (target, identifier, fn),
+                            ref(listener_fn) -> (target, identifier, fn),
+                            ref(listener_fn) -> (target, identifier, fn),
+                        }
+"""
+
+
+def _collection_gced(ref: weakref.ref[Any]) -> None:
+    # defaultdict, so can't get a KeyError
+    if not _collection_to_key or ref not in _collection_to_key:
+        return
+
+    ref = cast("weakref.ref[RefCollection[EventTarget]]", ref)
+
+    listener_to_key = _collection_to_key.pop(ref)
+    for key in listener_to_key.values():
+        if key in _key_to_collection:
+            # defaultdict, so can't get a KeyError
+            dispatch_reg = _key_to_collection[key]
+            dispatch_reg.pop(ref)
+            if not dispatch_reg:
+                _key_to_collection.pop(key)
+
+
+def _stored_in_collection(
+    event_key: _EventKey[_ET], owner: RefCollection[_ET]
+) -> bool:
+    key = event_key._key
+
+    dispatch_reg = _key_to_collection[key]
+
+    owner_ref = owner.ref
+    listen_ref = weakref.ref(event_key._listen_fn)
+
+    if owner_ref in dispatch_reg:
+        return False
+
+    dispatch_reg[owner_ref] = listen_ref
+
+    listener_to_key = _collection_to_key[owner_ref]
+    listener_to_key[listen_ref] = key
+
+    return True
+
+
+def _removed_from_collection(
+    event_key: _EventKey[_ET], owner: RefCollection[_ET]
+) -> None:
+    key = event_key._key
+
+    dispatch_reg = _key_to_collection[key]
+
+    listen_ref = weakref.ref(event_key._listen_fn)
+
+    owner_ref = owner.ref
+    dispatch_reg.pop(owner_ref, None)
+    if not dispatch_reg:
+        del _key_to_collection[key]
+
+    if owner_ref in _collection_to_key:
+        listener_to_key = _collection_to_key[owner_ref]
+        # see #12216 - this guards against a removal that already occurred
+        # here. however, I cannot come up with a test that shows any negative
+        # side effects occurring from this removal happening, even though an
+        # event key may still be referenced from a clsleveldispatch here
+        listener_to_key.pop(listen_ref, None)
+
+
+def _stored_in_collection_multi(
+    newowner: RefCollection[_ET],
+    oldowner: RefCollection[_ET],
+    elements: Iterable[_ListenerFnType],
+) -> None:
+    if not elements:
+        return
+
+    oldowner_ref = oldowner.ref
+    newowner_ref = newowner.ref
+
+    old_listener_to_key = _collection_to_key[oldowner_ref]
+    new_listener_to_key = _collection_to_key[newowner_ref]
+
+    for listen_fn in elements:
+        listen_ref = weakref.ref(listen_fn)
+        try:
+            key = old_listener_to_key[listen_ref]
+        except KeyError:
+            # can occur during interpreter shutdown.
+            # see #6740
+            continue
+
+        try:
+            dispatch_reg = _key_to_collection[key]
+        except KeyError:
+            continue
+
+        if newowner_ref in dispatch_reg:
+            assert dispatch_reg[newowner_ref] == listen_ref
+        else:
+            dispatch_reg[newowner_ref] = listen_ref
+
+        new_listener_to_key[listen_ref] = key
+
+
+def _clear(
+    owner: RefCollection[_ET],
+    elements: Iterable[_ListenerFnType],
+) -> None:
+    if not elements:
+        return
+
+    owner_ref = owner.ref
+    listener_to_key = _collection_to_key[owner_ref]
+    for listen_fn in elements:
+        listen_ref = weakref.ref(listen_fn)
+        key = listener_to_key[listen_ref]
+        dispatch_reg = _key_to_collection[key]
+        dispatch_reg.pop(owner_ref, None)
+
+        if not dispatch_reg:
+            del _key_to_collection[key]
+
+
+class _EventKey(Generic[_ET]):
+    """Represent :func:`.listen` arguments."""
+
+    __slots__ = (
+        "target",
+        "identifier",
+        "fn",
+        "fn_key",
+        "fn_wrap",
+        "dispatch_target",
+    )
+
+    target: _ET
+    identifier: str
+    fn: _ListenerFnType
+    fn_key: _ListenerFnKeyType
+    dispatch_target: Any
+    _fn_wrap: Optional[_ListenerFnType]
+
+    def __init__(
+        self,
+        target: _ET,
+        identifier: str,
+        fn: _ListenerFnType,
+        dispatch_target: Any,
+        _fn_wrap: Optional[_ListenerFnType] = None,
+    ):
+        self.target = target
+        self.identifier = identifier
+        self.fn = fn
+        if isinstance(fn, types.MethodType):
+            self.fn_key = id(fn.__func__), id(fn.__self__)
+        else:
+            self.fn_key = id(fn)
+        self.fn_wrap = _fn_wrap
+        self.dispatch_target = dispatch_target
+
+    @property
+    def _key(self) -> _EventKeyTupleType:
+        return (id(self.target), self.identifier, self.fn_key)
+
+    def with_wrapper(self, fn_wrap: _ListenerFnType) -> _EventKey[_ET]:
+        if fn_wrap is self._listen_fn:
+            return self
+        else:
+            return _EventKey(
+                self.target,
+                self.identifier,
+                self.fn,
+                self.dispatch_target,
+                _fn_wrap=fn_wrap,
+            )
+
+    def with_dispatch_target(self, dispatch_target: Any) -> _EventKey[_ET]:
+        if dispatch_target is self.dispatch_target:
+            return self
+        else:
+            return _EventKey(
+                self.target,
+                self.identifier,
+                self.fn,
+                dispatch_target,
+                _fn_wrap=self.fn_wrap,
+            )
+
+    def listen(self, *args: Any, **kw: Any) -> None:
+        once = kw.pop("once", False)
+        once_unless_exception = kw.pop("_once_unless_exception", False)
+        named = kw.pop("named", False)
+
+        target, identifier, fn = (
+            self.dispatch_target,
+            self.identifier,
+            self._listen_fn,
+        )
+
+        dispatch_collection = getattr(target.dispatch, identifier)
+
+        adjusted_fn = dispatch_collection._adjust_fn_spec(fn, named)
+
+        self = self.with_wrapper(adjusted_fn)
+
+        stub_function = getattr(
+            self.dispatch_target.dispatch._events, self.identifier
+        )
+        if hasattr(stub_function, "_sa_warn"):
+            stub_function._sa_warn()
+
+        if once or once_unless_exception:
+            self.with_wrapper(
+                util.only_once(
+                    self._listen_fn, retry_on_exception=once_unless_exception
+                )
+            ).listen(*args, **kw)
+        else:
+            self.dispatch_target.dispatch._listen(self, *args, **kw)
+
+    def remove(self) -> None:
+        key = self._key
+
+        if key not in _key_to_collection:
+            raise exc.InvalidRequestError(
+                "No listeners found for event %s / %r / %s "
+                % (self.target, self.identifier, self.fn)
+            )
+
+        dispatch_reg = _key_to_collection.pop(key)
+
+        for collection_ref, listener_ref in dispatch_reg.items():
+            collection = collection_ref()
+            listener_fn = listener_ref()
+            if collection is not None and listener_fn is not None:
+                collection.remove(self.with_wrapper(listener_fn))
+
+    def contains(self) -> bool:
+        """Return True if this event key is registered to listen."""
+        return self._key in _key_to_collection
+
+    def base_listen(
+        self,
+        propagate: bool = False,
+        insert: bool = False,
+        named: bool = False,
+        retval: Optional[bool] = None,
+        asyncio: bool = False,
+    ) -> None:
+        target, identifier = self.dispatch_target, self.identifier
+
+        dispatch_collection = getattr(target.dispatch, identifier)
+
+        for_modify = dispatch_collection.for_modify(target.dispatch)
+        if asyncio:
+            for_modify._set_asyncio()
+
+        if insert:
+            for_modify.insert(self, propagate)
+        else:
+            for_modify.append(self, propagate)
+
+    @property
+    def _listen_fn(self) -> _ListenerFnType:
+        return self.fn_wrap or self.fn
+
+    def append_to_list(
+        self,
+        owner: RefCollection[_ET],
+        list_: Deque[_ListenerFnType],
+    ) -> bool:
+        if _stored_in_collection(self, owner):
+            list_.append(self._listen_fn)
+            return True
+        else:
+            return False
+
+    def remove_from_list(
+        self,
+        owner: RefCollection[_ET],
+        list_: Deque[_ListenerFnType],
+    ) -> None:
+        _removed_from_collection(self, owner)
+        list_.remove(self._listen_fn)
+
+    def prepend_to_list(
+        self,
+        owner: RefCollection[_ET],
+        list_: Deque[_ListenerFnType],
+    ) -> bool:
+        if _stored_in_collection(self, owner):
+            list_.appendleft(self._listen_fn)
+            return True
+        else:
+            return False
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/events.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/events.py
new file mode 100644
index 00000000..ce832439
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/events.py
@@ -0,0 +1,17 @@
+# events.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Core event interfaces."""
+
+from __future__ import annotations
+
+from .engine.events import ConnectionEvents
+from .engine.events import DialectEvents
+from .pool import PoolResetState
+from .pool.events import PoolEvents
+from .sql.base import SchemaEventTarget
+from .sql.events import DDLEvents
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/exc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/exc.py
new file mode 100644
index 00000000..71e5dd81
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/exc.py
@@ -0,0 +1,832 @@
+# exc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Exceptions used with SQLAlchemy.
+
+The base exception class is :exc:`.SQLAlchemyError`.  Exceptions which are
+raised as a result of DBAPI exceptions are all subclasses of
+:exc:`.DBAPIError`.
+
+"""
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import Union
+
+from .util import compat
+from .util import preloaded as _preloaded
+
+if typing.TYPE_CHECKING:
+    from .engine.interfaces import _AnyExecuteParams
+    from .engine.interfaces import Dialect
+    from .sql.compiler import Compiled
+    from .sql.compiler import TypeCompiler
+    from .sql.elements import ClauseElement
+
+if typing.TYPE_CHECKING:
+    _version_token: str
+else:
+    # set by __init__.py
+    _version_token = None
+
+
+class HasDescriptionCode:
+    """helper which adds 'code' as an attribute and '_code_str' as a method"""
+
+    code: Optional[str] = None
+
+    def __init__(self, *arg: Any, **kw: Any):
+        code = kw.pop("code", None)
+        if code is not None:
+            self.code = code
+        super().__init__(*arg, **kw)
+
+    _what_are_we = "error"
+
+    def _code_str(self) -> str:
+        if not self.code:
+            return ""
+        else:
+            return (
+                f"(Background on this {self._what_are_we} at: "
+                f"https://sqlalche.me/e/{_version_token}/{self.code})"
+            )
+
+    def __str__(self) -> str:
+        message = super().__str__()
+        if self.code:
+            message = "%s %s" % (message, self._code_str())
+        return message
+
+
+class SQLAlchemyError(HasDescriptionCode, Exception):
+    """Generic error class."""
+
+    def _message(self) -> str:
+        # rules:
+        #
+        # 1. single arg string will usually be a unicode
+        # object, but since __str__() must return unicode, check for
+        # bytestring just in case
+        #
+        # 2. for multiple self.args, this is not a case in current
+        # SQLAlchemy though this is happening in at least one known external
+        # library, call str() which does a repr().
+        #
+        text: str
+
+        if len(self.args) == 1:
+            arg_text = self.args[0]
+
+            if isinstance(arg_text, bytes):
+                text = compat.decode_backslashreplace(arg_text, "utf-8")
+            # This is for when the argument is not a string of any sort.
+            # Otherwise, converting this exception to string would fail for
+            # non-string arguments.
+            else:
+                text = str(arg_text)
+
+            return text
+        else:
+            # this is not a normal case within SQLAlchemy but is here for
+            # compatibility with Exception.args - the str() comes out as
+            # a repr() of the tuple
+            return str(self.args)
+
+    def _sql_message(self) -> str:
+        message = self._message()
+
+        if self.code:
+            message = "%s %s" % (message, self._code_str())
+
+        return message
+
+    def __str__(self) -> str:
+        return self._sql_message()
+
+
+class ArgumentError(SQLAlchemyError):
+    """Raised when an invalid or conflicting function argument is supplied.
+
+    This error generally corresponds to construction time state errors.
+
+    """
+
+
+class DuplicateColumnError(ArgumentError):
+    """a Column is being added to a Table that would replace another
+    Column, without appropriate parameters to allow this in place.
+
+    .. versionadded:: 2.0.0b4
+
+    """
+
+
+class ObjectNotExecutableError(ArgumentError):
+    """Raised when an object is passed to .execute() that can't be
+    executed as SQL.
+
+    """
+
+    def __init__(self, target: Any):
+        super().__init__("Not an executable object: %r" % target)
+        self.target = target
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return self.__class__, (self.target,)
+
+
+class NoSuchModuleError(ArgumentError):
+    """Raised when a dynamically-loaded module (usually a database dialect)
+    of a particular name cannot be located."""
+
+
+class NoForeignKeysError(ArgumentError):
+    """Raised when no foreign keys can be located between two selectables
+    during a join."""
+
+
+class AmbiguousForeignKeysError(ArgumentError):
+    """Raised when more than one foreign key matching can be located
+    between two selectables during a join."""
+
+
+class ConstraintColumnNotFoundError(ArgumentError):
+    """raised when a constraint refers to a string column name that
+    is not present in the table being constrained.
+
+    .. versionadded:: 2.0
+
+    """
+
+
+class CircularDependencyError(SQLAlchemyError):
+    """Raised by topological sorts when a circular dependency is detected.
+
+    There are two scenarios where this error occurs:
+
+    * In a Session flush operation, if two objects are mutually dependent
+      on each other, they can not be inserted or deleted via INSERT or
+      DELETE statements alone; an UPDATE will be needed to post-associate
+      or pre-deassociate one of the foreign key constrained values.
+      The ``post_update`` flag described at :ref:`post_update` can resolve
+      this cycle.
+    * In a :attr:`_schema.MetaData.sorted_tables` operation, two
+      :class:`_schema.ForeignKey`
+      or :class:`_schema.ForeignKeyConstraint` objects mutually refer to each
+      other.  Apply the ``use_alter=True`` flag to one or both,
+      see :ref:`use_alter`.
+
+    """
+
+    def __init__(
+        self,
+        message: str,
+        cycles: Any,
+        edges: Any,
+        msg: Optional[str] = None,
+        code: Optional[str] = None,
+    ):
+        if msg is None:
+            message += " (%s)" % ", ".join(repr(s) for s in cycles)
+        else:
+            message = msg
+        SQLAlchemyError.__init__(self, message, code=code)
+        self.cycles = cycles
+        self.edges = edges
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return (
+            self.__class__,
+            (None, self.cycles, self.edges, self.args[0]),
+            {"code": self.code} if self.code is not None else {},
+        )
+
+
+class CompileError(SQLAlchemyError):
+    """Raised when an error occurs during SQL compilation"""
+
+
+class UnsupportedCompilationError(CompileError):
+    """Raised when an operation is not supported by the given compiler.
+
+    .. seealso::
+
+        :ref:`faq_sql_expression_string`
+
+        :ref:`error_l7de`
+    """
+
+    code = "l7de"
+
+    def __init__(
+        self,
+        compiler: Union[Compiled, TypeCompiler],
+        element_type: Type[ClauseElement],
+        message: Optional[str] = None,
+    ):
+        super().__init__(
+            "Compiler %r can't render element of type %s%s"
+            % (compiler, element_type, ": %s" % message if message else "")
+        )
+        self.compiler = compiler
+        self.element_type = element_type
+        self.message = message
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return self.__class__, (self.compiler, self.element_type, self.message)
+
+
+class IdentifierError(SQLAlchemyError):
+    """Raised when a schema name is beyond the max character limit"""
+
+
+class DisconnectionError(SQLAlchemyError):
+    """A disconnect is detected on a raw DB-API connection.
+
+    This error is raised and consumed internally by a connection pool.  It can
+    be raised by the :meth:`_events.PoolEvents.checkout`
+    event so that the host pool
+    forces a retry; the exception will be caught three times in a row before
+    the pool gives up and raises :class:`~sqlalchemy.exc.InvalidRequestError`
+    regarding the connection attempt.
+
+    """
+
+    invalidate_pool: bool = False
+
+
+class InvalidatePoolError(DisconnectionError):
+    """Raised when the connection pool should invalidate all stale connections.
+
+    A subclass of :class:`_exc.DisconnectionError` that indicates that the
+    disconnect situation encountered on the connection probably means the
+    entire pool should be invalidated, as the database has been restarted.
+
+    This exception will be handled otherwise the same way as
+    :class:`_exc.DisconnectionError`, allowing three attempts to reconnect
+    before giving up.
+
+    .. versionadded:: 1.2
+
+    """
+
+    invalidate_pool: bool = True
+
+
+class TimeoutError(SQLAlchemyError):  # noqa
+    """Raised when a connection pool times out on getting a connection."""
+
+
+class InvalidRequestError(SQLAlchemyError):
+    """SQLAlchemy was asked to do something it can't do.
+
+    This error generally corresponds to runtime state errors.
+
+    """
+
+
+class IllegalStateChangeError(InvalidRequestError):
+    """An object that tracks state encountered an illegal state change
+    of some kind.
+
+    .. versionadded:: 2.0
+
+    """
+
+
+class NoInspectionAvailable(InvalidRequestError):
+    """A subject passed to :func:`sqlalchemy.inspection.inspect` produced
+    no context for inspection."""
+
+
+class PendingRollbackError(InvalidRequestError):
+    """A transaction has failed and needs to be rolled back before
+    continuing.
+
+    .. versionadded:: 1.4
+
+    """
+
+
+class ResourceClosedError(InvalidRequestError):
+    """An operation was requested from a connection, cursor, or other
+    object that's in a closed state."""
+
+
+class NoSuchColumnError(InvalidRequestError, KeyError):
+    """A nonexistent column is requested from a ``Row``."""
+
+
+class NoResultFound(InvalidRequestError):
+    """A database result was required but none was found.
+
+
+    .. versionchanged:: 1.4  This exception is now part of the
+       ``sqlalchemy.exc`` module in Core, moved from the ORM.  The symbol
+       remains importable from ``sqlalchemy.orm.exc``.
+
+
+    """
+
+
+class MultipleResultsFound(InvalidRequestError):
+    """A single database result was required but more than one were found.
+
+    .. versionchanged:: 1.4  This exception is now part of the
+       ``sqlalchemy.exc`` module in Core, moved from the ORM.  The symbol
+       remains importable from ``sqlalchemy.orm.exc``.
+
+
+    """
+
+
+class NoReferenceError(InvalidRequestError):
+    """Raised by ``ForeignKey`` to indicate a reference cannot be resolved."""
+
+    table_name: str
+
+
+class AwaitRequired(InvalidRequestError):
+    """Error raised by the async greenlet spawn if no async operation
+    was awaited when it required one.
+
+    """
+
+    code = "xd1r"
+
+
+class MissingGreenlet(InvalidRequestError):
+    r"""Error raised by the async greenlet await\_ if called while not inside
+    the greenlet spawn context.
+
+    """
+
+    code = "xd2s"
+
+
+class NoReferencedTableError(NoReferenceError):
+    """Raised by ``ForeignKey`` when the referred ``Table`` cannot be
+    located.
+
+    """
+
+    def __init__(self, message: str, tname: str):
+        NoReferenceError.__init__(self, message)
+        self.table_name = tname
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return self.__class__, (self.args[0], self.table_name)
+
+
+class NoReferencedColumnError(NoReferenceError):
+    """Raised by ``ForeignKey`` when the referred ``Column`` cannot be
+    located.
+
+    """
+
+    def __init__(self, message: str, tname: str, cname: str):
+        NoReferenceError.__init__(self, message)
+        self.table_name = tname
+        self.column_name = cname
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return (
+            self.__class__,
+            (self.args[0], self.table_name, self.column_name),
+        )
+
+
+class NoSuchTableError(InvalidRequestError):
+    """Table does not exist or is not visible to a connection."""
+
+
+class UnreflectableTableError(InvalidRequestError):
+    """Table exists but can't be reflected for some reason.
+
+    .. versionadded:: 1.2
+
+    """
+
+
+class UnboundExecutionError(InvalidRequestError):
+    """SQL was attempted without a database connection to execute it on."""
+
+
+class DontWrapMixin:
+    """A mixin class which, when applied to a user-defined Exception class,
+    will not be wrapped inside of :exc:`.StatementError` if the error is
+    emitted within the process of executing a statement.
+
+    E.g.::
+
+        from sqlalchemy.exc import DontWrapMixin
+
+
+        class MyCustomException(Exception, DontWrapMixin):
+            pass
+
+
+        class MySpecialType(TypeDecorator):
+            impl = String
+
+            def process_bind_param(self, value, dialect):
+                if value == "invalid":
+                    raise MyCustomException("invalid!")
+
+    """
+
+
+class StatementError(SQLAlchemyError):
+    """An error occurred during execution of a SQL statement.
+
+    :class:`StatementError` wraps the exception raised
+    during execution, and features :attr:`.statement`
+    and :attr:`.params` attributes which supply context regarding
+    the specifics of the statement which had an issue.
+
+    The wrapped exception object is available in
+    the :attr:`.orig` attribute.
+
+    """
+
+    statement: Optional[str] = None
+    """The string SQL statement being invoked when this exception occurred."""
+
+    params: Optional[_AnyExecuteParams] = None
+    """The parameter list being used when this exception occurred."""
+
+    orig: Optional[BaseException] = None
+    """The original exception that was thrown.
+
+    """
+
+    ismulti: Optional[bool] = None
+    """multi parameter passed to repr_params().  None is meaningful."""
+
+    connection_invalidated: bool = False
+
+    def __init__(
+        self,
+        message: str,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: Optional[BaseException],
+        hide_parameters: bool = False,
+        code: Optional[str] = None,
+        ismulti: Optional[bool] = None,
+    ):
+        SQLAlchemyError.__init__(self, message, code=code)
+        self.statement = statement
+        self.params = params
+        self.orig = orig
+        self.ismulti = ismulti
+        self.hide_parameters = hide_parameters
+        self.detail: List[str] = []
+
+    def add_detail(self, msg: str) -> None:
+        self.detail.append(msg)
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return (
+            self.__class__,
+            (
+                self.args[0],
+                self.statement,
+                self.params,
+                self.orig,
+                self.hide_parameters,
+                self.__dict__.get("code"),
+                self.ismulti,
+            ),
+            {"detail": self.detail},
+        )
+
+    @_preloaded.preload_module("sqlalchemy.sql.util")
+    def _sql_message(self) -> str:
+        util = _preloaded.sql_util
+
+        details = [self._message()]
+        if self.statement:
+            stmt_detail = "[SQL: %s]" % self.statement
+            details.append(stmt_detail)
+            if self.params:
+                if self.hide_parameters:
+                    details.append(
+                        "[SQL parameters hidden due to hide_parameters=True]"
+                    )
+                else:
+                    params_repr = util._repr_params(
+                        self.params, 10, ismulti=self.ismulti
+                    )
+                    details.append("[parameters: %r]" % params_repr)
+        code_str = self._code_str()
+        if code_str:
+            details.append(code_str)
+        return "\n".join(["(%s)" % det for det in self.detail] + details)
+
+
+class DBAPIError(StatementError):
+    """Raised when the execution of a database operation fails.
+
+    Wraps exceptions raised by the DB-API underlying the
+    database operation.  Driver-specific implementations of the standard
+    DB-API exception types are wrapped by matching sub-types of SQLAlchemy's
+    :class:`DBAPIError` when possible.  DB-API's ``Error`` type maps to
+    :class:`DBAPIError` in SQLAlchemy, otherwise the names are identical.  Note
+    that there is no guarantee that different DB-API implementations will
+    raise the same exception type for any given error condition.
+
+    :class:`DBAPIError` features :attr:`~.StatementError.statement`
+    and :attr:`~.StatementError.params` attributes which supply context
+    regarding the specifics of the statement which had an issue, for the
+    typical case when the error was raised within the context of
+    emitting a SQL statement.
+
+    The wrapped exception object is available in the
+    :attr:`~.StatementError.orig` attribute. Its type and properties are
+    DB-API implementation specific.
+
+    """
+
+    code = "dbapi"
+
+    @overload
+    @classmethod
+    def instance(
+        cls,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: Exception,
+        dbapi_base_err: Type[Exception],
+        hide_parameters: bool = False,
+        connection_invalidated: bool = False,
+        dialect: Optional[Dialect] = None,
+        ismulti: Optional[bool] = None,
+    ) -> StatementError: ...
+
+    @overload
+    @classmethod
+    def instance(
+        cls,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: DontWrapMixin,
+        dbapi_base_err: Type[Exception],
+        hide_parameters: bool = False,
+        connection_invalidated: bool = False,
+        dialect: Optional[Dialect] = None,
+        ismulti: Optional[bool] = None,
+    ) -> DontWrapMixin: ...
+
+    @overload
+    @classmethod
+    def instance(
+        cls,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: BaseException,
+        dbapi_base_err: Type[Exception],
+        hide_parameters: bool = False,
+        connection_invalidated: bool = False,
+        dialect: Optional[Dialect] = None,
+        ismulti: Optional[bool] = None,
+    ) -> BaseException: ...
+
+    @classmethod
+    def instance(
+        cls,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: Union[BaseException, DontWrapMixin],
+        dbapi_base_err: Type[Exception],
+        hide_parameters: bool = False,
+        connection_invalidated: bool = False,
+        dialect: Optional[Dialect] = None,
+        ismulti: Optional[bool] = None,
+    ) -> Union[BaseException, DontWrapMixin]:
+        # Don't ever wrap these, just return them directly as if
+        # DBAPIError didn't exist.
+        if (
+            isinstance(orig, BaseException) and not isinstance(orig, Exception)
+        ) or isinstance(orig, DontWrapMixin):
+            return orig
+
+        if orig is not None:
+            # not a DBAPI error, statement is present.
+            # raise a StatementError
+            if isinstance(orig, SQLAlchemyError) and statement:
+                return StatementError(
+                    "(%s.%s) %s"
+                    % (
+                        orig.__class__.__module__,
+                        orig.__class__.__name__,
+                        orig.args[0],
+                    ),
+                    statement,
+                    params,
+                    orig,
+                    hide_parameters=hide_parameters,
+                    code=orig.code,
+                    ismulti=ismulti,
+                )
+            elif not isinstance(orig, dbapi_base_err) and statement:
+                return StatementError(
+                    "(%s.%s) %s"
+                    % (
+                        orig.__class__.__module__,
+                        orig.__class__.__name__,
+                        orig,
+                    ),
+                    statement,
+                    params,
+                    orig,
+                    hide_parameters=hide_parameters,
+                    ismulti=ismulti,
+                )
+
+            glob = globals()
+            for super_ in orig.__class__.__mro__:
+                name = super_.__name__
+                if dialect:
+                    name = dialect.dbapi_exception_translation_map.get(
+                        name, name
+                    )
+                if name in glob and issubclass(glob[name], DBAPIError):
+                    cls = glob[name]
+                    break
+
+        return cls(
+            statement,
+            params,
+            orig,
+            connection_invalidated=connection_invalidated,
+            hide_parameters=hide_parameters,
+            code=cls.code,
+            ismulti=ismulti,
+        )
+
+    def __reduce__(self) -> Union[str, Tuple[Any, ...]]:
+        return (
+            self.__class__,
+            (
+                self.statement,
+                self.params,
+                self.orig,
+                self.hide_parameters,
+                self.connection_invalidated,
+                self.__dict__.get("code"),
+                self.ismulti,
+            ),
+            {"detail": self.detail},
+        )
+
+    def __init__(
+        self,
+        statement: Optional[str],
+        params: Optional[_AnyExecuteParams],
+        orig: BaseException,
+        hide_parameters: bool = False,
+        connection_invalidated: bool = False,
+        code: Optional[str] = None,
+        ismulti: Optional[bool] = None,
+    ):
+        try:
+            text = str(orig)
+        except Exception as e:
+            text = "Error in str() of DB-API-generated exception: " + str(e)
+        StatementError.__init__(
+            self,
+            "(%s.%s) %s"
+            % (orig.__class__.__module__, orig.__class__.__name__, text),
+            statement,
+            params,
+            orig,
+            hide_parameters,
+            code=code,
+            ismulti=ismulti,
+        )
+        self.connection_invalidated = connection_invalidated
+
+
+class InterfaceError(DBAPIError):
+    """Wraps a DB-API InterfaceError."""
+
+    code = "rvf5"
+
+
+class DatabaseError(DBAPIError):
+    """Wraps a DB-API DatabaseError."""
+
+    code = "4xp6"
+
+
+class DataError(DatabaseError):
+    """Wraps a DB-API DataError."""
+
+    code = "9h9h"
+
+
+class OperationalError(DatabaseError):
+    """Wraps a DB-API OperationalError."""
+
+    code = "e3q8"
+
+
+class IntegrityError(DatabaseError):
+    """Wraps a DB-API IntegrityError."""
+
+    code = "gkpj"
+
+
+class InternalError(DatabaseError):
+    """Wraps a DB-API InternalError."""
+
+    code = "2j85"
+
+
+class ProgrammingError(DatabaseError):
+    """Wraps a DB-API ProgrammingError."""
+
+    code = "f405"
+
+
+class NotSupportedError(DatabaseError):
+    """Wraps a DB-API NotSupportedError."""
+
+    code = "tw8g"
+
+
+# Warnings
+
+
+class SATestSuiteWarning(Warning):
+    """warning for a condition detected during tests that is non-fatal
+
+    Currently outside of SAWarning so that we can work around tools like
+    Alembic doing the wrong thing with warnings.
+
+    """
+
+
+class SADeprecationWarning(HasDescriptionCode, DeprecationWarning):
+    """Issued for usage of deprecated APIs."""
+
+    deprecated_since: Optional[str] = None
+    "Indicates the version that started raising this deprecation warning"
+
+
+class Base20DeprecationWarning(SADeprecationWarning):
+    """Issued for usage of APIs specifically deprecated or legacy in
+    SQLAlchemy 2.0.
+
+    .. seealso::
+
+        :ref:`error_b8d9`.
+
+        :ref:`deprecation_20_mode`
+
+    """
+
+    deprecated_since: Optional[str] = "1.4"
+    "Indicates the version that started raising this deprecation warning"
+
+    def __str__(self) -> str:
+        return (
+            super().__str__()
+            + " (Background on SQLAlchemy 2.0 at: https://sqlalche.me/e/b8d9)"
+        )
+
+
+class LegacyAPIWarning(Base20DeprecationWarning):
+    """indicates an API that is in 'legacy' status, a long term deprecation."""
+
+
+class MovedIn20Warning(Base20DeprecationWarning):
+    """Subtype of RemovedIn20Warning to indicate an API that moved only."""
+
+
+class SAPendingDeprecationWarning(PendingDeprecationWarning):
+    """A similar warning as :class:`_exc.SADeprecationWarning`, this warning
+    is not used in modern versions of SQLAlchemy.
+
+    """
+
+    deprecated_since: Optional[str] = None
+    "Indicates the version that started raising this deprecation warning"
+
+
+class SAWarning(HasDescriptionCode, RuntimeWarning):
+    """Issued at runtime."""
+
+    _what_are_we = "warning"
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__init__.py
new file mode 100644
index 00000000..2751bcf9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__init__.py
@@ -0,0 +1,11 @@
+# ext/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from .. import util as _sa_util
+
+
+_sa_util.preloaded.import_prefix("sqlalchemy.ext")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..75baf25e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/associationproxy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/associationproxy.cpython-312.pyc
new file mode 100644
index 00000000..3ca13124
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/associationproxy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/automap.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/automap.cpython-312.pyc
new file mode 100644
index 00000000..d604a283
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/automap.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/baked.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/baked.cpython-312.pyc
new file mode 100644
index 00000000..6ad467d7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/baked.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/compiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/compiler.cpython-312.pyc
new file mode 100644
index 00000000..9a32aef4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/compiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/horizontal_shard.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/horizontal_shard.cpython-312.pyc
new file mode 100644
index 00000000..d0d33b51
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/horizontal_shard.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/hybrid.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/hybrid.cpython-312.pyc
new file mode 100644
index 00000000..c6696a1b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/hybrid.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/indexable.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/indexable.cpython-312.pyc
new file mode 100644
index 00000000..26b5ac0e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/indexable.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/instrumentation.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/instrumentation.cpython-312.pyc
new file mode 100644
index 00000000..5d1962d4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/instrumentation.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/mutable.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/mutable.cpython-312.pyc
new file mode 100644
index 00000000..6da4030e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/mutable.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/orderinglist.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/orderinglist.cpython-312.pyc
new file mode 100644
index 00000000..7ebe2a04
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/orderinglist.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/serializer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/serializer.cpython-312.pyc
new file mode 100644
index 00000000..4de670cf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/__pycache__/serializer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/associationproxy.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/associationproxy.py
new file mode 100644
index 00000000..8f2c19b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/associationproxy.py
@@ -0,0 +1,2013 @@
+# ext/associationproxy.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Contain the ``AssociationProxy`` class.
+
+The ``AssociationProxy`` is a Python property object which provides
+transparent proxied access to the endpoint of an association object.
+
+See the example ``examples/association/proxied_association.py``.
+
+"""
+from __future__ import annotations
+
+import operator
+import typing
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Generic
+from typing import ItemsView
+from typing import Iterable
+from typing import Iterator
+from typing import KeysView
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import MutableSequence
+from typing import MutableSet
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+from typing import ValuesView
+
+from .. import ColumnElement
+from .. import exc
+from .. import inspect
+from .. import orm
+from .. import util
+from ..orm import collections
+from ..orm import InspectionAttrExtensionType
+from ..orm import interfaces
+from ..orm import ORMDescriptor
+from ..orm.base import SQLORMOperations
+from ..orm.interfaces import _AttributeOptions
+from ..orm.interfaces import _DCAttributeOptions
+from ..orm.interfaces import _DEFAULT_ATTRIBUTE_OPTIONS
+from ..sql import operators
+from ..sql import or_
+from ..sql.base import _NoArg
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import Self
+from ..util.typing import SupportsIndex
+from ..util.typing import SupportsKeysAndGetItem
+
+if typing.TYPE_CHECKING:
+    from ..orm.interfaces import MapperProperty
+    from ..orm.interfaces import PropComparator
+    from ..orm.mapper import Mapper
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _InfoType
+
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+_T_con = TypeVar("_T_con", bound=Any, contravariant=True)
+_S = TypeVar("_S", bound=Any)
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+
+
+def association_proxy(
+    target_collection: str,
+    attr: str,
+    *,
+    creator: Optional[_CreatorProtocol] = None,
+    getset_factory: Optional[_GetSetFactoryProtocol] = None,
+    proxy_factory: Optional[_ProxyFactoryProtocol] = None,
+    proxy_bulk_set: Optional[_ProxyBulkSetProtocol] = None,
+    info: Optional[_InfoType] = None,
+    cascade_scalar_deletes: bool = False,
+    create_on_none_assignment: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+) -> AssociationProxy[Any]:
+    r"""Return a Python property implementing a view of a target
+    attribute which references an attribute on members of the
+    target.
+
+    The returned value is an instance of :class:`.AssociationProxy`.
+
+    Implements a Python property representing a relationship as a collection
+    of simpler values, or a scalar value.  The proxied property will mimic
+    the collection type of the target (list, dict or set), or, in the case of
+    a one to one relationship, a simple scalar value.
+
+    :param target_collection: Name of the attribute that is the immediate
+      target.  This attribute is typically mapped by
+      :func:`~sqlalchemy.orm.relationship` to link to a target collection, but
+      can also be a many-to-one or non-scalar relationship.
+
+    :param attr: Attribute on the associated instance or instances that
+      are available on instances of the target object.
+
+    :param creator: optional.
+
+      Defines custom behavior when new items are added to the proxied
+      collection.
+
+      By default, adding new items to the collection will trigger a
+      construction of an instance of the target object, passing the given
+      item as a positional argument to the target constructor.  For cases
+      where this isn't sufficient, :paramref:`.association_proxy.creator`
+      can supply a callable that will construct the object in the
+      appropriate way, given the item that was passed.
+
+      For list- and set- oriented collections, a single argument is
+      passed to the callable. For dictionary oriented collections, two
+      arguments are passed, corresponding to the key and value.
+
+      The :paramref:`.association_proxy.creator` callable is also invoked
+      for scalar (i.e. many-to-one, one-to-one) relationships. If the
+      current value of the target relationship attribute is ``None``, the
+      callable is used to construct a new object.  If an object value already
+      exists, the given attribute value is populated onto that object.
+
+      .. seealso::
+
+        :ref:`associationproxy_creator`
+
+    :param cascade_scalar_deletes: when True, indicates that setting
+        the proxied value to ``None``, or deleting it via ``del``, should
+        also remove the source object.  Only applies to scalar attributes.
+        Normally, removing the proxied target will not remove the proxy
+        source, as this object may have other state that is still to be
+        kept.
+
+        .. versionadded:: 1.3
+
+        .. seealso::
+
+            :ref:`cascade_scalar_deletes` - complete usage example
+
+    :param create_on_none_assignment: when True, indicates that setting
+      the proxied value to ``None`` should **create** the source object
+      if it does not exist, using the creator.  Only applies to scalar
+      attributes.  This is mutually exclusive
+      vs. the :paramref:`.assocation_proxy.cascade_scalar_deletes`.
+
+      .. versionadded:: 2.0.18
+
+    :param init: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__init__()``
+     method as generated by the dataclass process.
+
+     .. versionadded:: 2.0.0b4
+
+    :param repr: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the attribute established by this :class:`.AssociationProxy`
+     should be part of the ``__repr__()`` method as generated by the dataclass
+     process.
+
+     .. versionadded:: 2.0.0b4
+
+    :param default_factory: Specific to
+     :ref:`orm_declarative_native_dataclasses`, specifies a default-value
+     generation function that will take place as part of the ``__init__()``
+     method as generated by the dataclass process.
+
+     .. versionadded:: 2.0.0b4
+
+    :param compare: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be included in comparison operations when generating the
+     ``__eq__()`` and ``__ne__()`` methods for the mapped class.
+
+     .. versionadded:: 2.0.0b4
+
+    :param kw_only: Specific to :ref:`orm_declarative_native_dataclasses`,
+     indicates if this field should be marked as keyword-only when generating
+     the ``__init__()`` method as generated by the dataclass process.
+
+     .. versionadded:: 2.0.0b4
+
+    :param hash: Specific to
+     :ref:`orm_declarative_native_dataclasses`, controls if this field
+     is included when generating the ``__hash__()`` method for the mapped
+     class.
+
+     .. versionadded:: 2.0.36
+
+    :param info: optional, will be assigned to
+     :attr:`.AssociationProxy.info` if present.
+
+
+    The following additional parameters involve injection of custom behaviors
+    within the :class:`.AssociationProxy` object and are for advanced use
+    only:
+
+    :param getset_factory: Optional.  Proxied attribute access is
+        automatically handled by routines that get and set values based on
+        the `attr` argument for this proxy.
+
+        If you would like to customize this behavior, you may supply a
+        `getset_factory` callable that produces a tuple of `getter` and
+        `setter` functions.  The factory is called with two arguments, the
+        abstract type of the underlying collection and this proxy instance.
+
+    :param proxy_factory: Optional.  The type of collection to emulate is
+        determined by sniffing the target collection.  If your collection
+        type can't be determined by duck typing or you'd like to use a
+        different collection implementation, you may supply a factory
+        function to produce those collections.  Only applicable to
+        non-scalar relationships.
+
+    :param proxy_bulk_set: Optional, use with proxy_factory.
+
+
+    """
+    return AssociationProxy(
+        target_collection,
+        attr,
+        creator=creator,
+        getset_factory=getset_factory,
+        proxy_factory=proxy_factory,
+        proxy_bulk_set=proxy_bulk_set,
+        info=info,
+        cascade_scalar_deletes=cascade_scalar_deletes,
+        create_on_none_assignment=create_on_none_assignment,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+    )
+
+
+class AssociationProxyExtensionType(InspectionAttrExtensionType):
+    ASSOCIATION_PROXY = "ASSOCIATION_PROXY"
+    """Symbol indicating an :class:`.InspectionAttr` that's
+    of type :class:`.AssociationProxy`.
+
+    Is assigned to the :attr:`.InspectionAttr.extension_type`
+    attribute.
+
+    """
+
+
+class _GetterProtocol(Protocol[_T_co]):
+    def __call__(self, instance: Any) -> _T_co: ...
+
+
+# mypy 0.990 we are no longer allowed to make this Protocol[_T_con]
+class _SetterProtocol(Protocol): ...
+
+
+class _PlainSetterProtocol(_SetterProtocol, Protocol[_T_con]):
+    def __call__(self, instance: Any, value: _T_con) -> None: ...
+
+
+class _DictSetterProtocol(_SetterProtocol, Protocol[_T_con]):
+    def __call__(self, instance: Any, key: Any, value: _T_con) -> None: ...
+
+
+# mypy 0.990 we are no longer allowed to make this Protocol[_T_con]
+class _CreatorProtocol(Protocol): ...
+
+
+class _PlainCreatorProtocol(_CreatorProtocol, Protocol[_T_con]):
+    def __call__(self, value: _T_con) -> Any: ...
+
+
+class _KeyCreatorProtocol(_CreatorProtocol, Protocol[_T_con]):
+    def __call__(self, key: Any, value: Optional[_T_con]) -> Any: ...
+
+
+class _LazyCollectionProtocol(Protocol[_T]):
+    def __call__(
+        self,
+    ) -> Union[
+        MutableSet[_T], MutableMapping[Any, _T], MutableSequence[_T]
+    ]: ...
+
+
+class _GetSetFactoryProtocol(Protocol):
+    def __call__(
+        self,
+        collection_class: Optional[Type[Any]],
+        assoc_instance: AssociationProxyInstance[Any],
+    ) -> Tuple[_GetterProtocol[Any], _SetterProtocol]: ...
+
+
+class _ProxyFactoryProtocol(Protocol):
+    def __call__(
+        self,
+        lazy_collection: _LazyCollectionProtocol[Any],
+        creator: _CreatorProtocol,
+        value_attr: str,
+        parent: AssociationProxyInstance[Any],
+    ) -> Any: ...
+
+
+class _ProxyBulkSetProtocol(Protocol):
+    def __call__(
+        self, proxy: _AssociationCollection[Any], collection: Iterable[Any]
+    ) -> None: ...
+
+
+class _AssociationProxyProtocol(Protocol[_T]):
+    """describes the interface of :class:`.AssociationProxy`
+    without including descriptor methods in the interface."""
+
+    creator: Optional[_CreatorProtocol]
+    key: str
+    target_collection: str
+    value_attr: str
+    cascade_scalar_deletes: bool
+    create_on_none_assignment: bool
+    getset_factory: Optional[_GetSetFactoryProtocol]
+    proxy_factory: Optional[_ProxyFactoryProtocol]
+    proxy_bulk_set: Optional[_ProxyBulkSetProtocol]
+
+    @util.ro_memoized_property
+    def info(self) -> _InfoType: ...
+
+    def for_class(
+        self, class_: Type[Any], obj: Optional[object] = None
+    ) -> AssociationProxyInstance[_T]: ...
+
+    def _default_getset(
+        self, collection_class: Any
+    ) -> Tuple[_GetterProtocol[Any], _SetterProtocol]: ...
+
+
+class AssociationProxy(
+    interfaces.InspectionAttrInfo,
+    ORMDescriptor[_T],
+    _DCAttributeOptions,
+    _AssociationProxyProtocol[_T],
+):
+    """A descriptor that presents a read/write view of an object attribute."""
+
+    is_attribute = True
+    extension_type = AssociationProxyExtensionType.ASSOCIATION_PROXY
+
+    def __init__(
+        self,
+        target_collection: str,
+        attr: str,
+        *,
+        creator: Optional[_CreatorProtocol] = None,
+        getset_factory: Optional[_GetSetFactoryProtocol] = None,
+        proxy_factory: Optional[_ProxyFactoryProtocol] = None,
+        proxy_bulk_set: Optional[_ProxyBulkSetProtocol] = None,
+        info: Optional[_InfoType] = None,
+        cascade_scalar_deletes: bool = False,
+        create_on_none_assignment: bool = False,
+        attribute_options: Optional[_AttributeOptions] = None,
+    ):
+        """Construct a new :class:`.AssociationProxy`.
+
+        The :class:`.AssociationProxy` object is typically constructed using
+        the :func:`.association_proxy` constructor function. See the
+        description of :func:`.association_proxy` for a description of all
+        parameters.
+
+
+        """
+        self.target_collection = target_collection
+        self.value_attr = attr
+        self.creator = creator
+        self.getset_factory = getset_factory
+        self.proxy_factory = proxy_factory
+        self.proxy_bulk_set = proxy_bulk_set
+
+        if cascade_scalar_deletes and create_on_none_assignment:
+            raise exc.ArgumentError(
+                "The cascade_scalar_deletes and create_on_none_assignment "
+                "parameters are mutually exclusive."
+            )
+        self.cascade_scalar_deletes = cascade_scalar_deletes
+        self.create_on_none_assignment = create_on_none_assignment
+
+        self.key = "_%s_%s_%s" % (
+            type(self).__name__,
+            target_collection,
+            id(self),
+        )
+        if info:
+            self.info = info  # type: ignore
+
+        if (
+            attribute_options
+            and attribute_options != _DEFAULT_ATTRIBUTE_OPTIONS
+        ):
+            self._has_dataclass_arguments = True
+            self._attribute_options = attribute_options
+        else:
+            self._has_dataclass_arguments = False
+            self._attribute_options = _DEFAULT_ATTRIBUTE_OPTIONS
+
+    @overload
+    def __get__(
+        self, instance: Literal[None], owner: Literal[None]
+    ) -> Self: ...
+
+    @overload
+    def __get__(
+        self, instance: Literal[None], owner: Any
+    ) -> AssociationProxyInstance[_T]: ...
+
+    @overload
+    def __get__(self, instance: object, owner: Any) -> _T: ...
+
+    def __get__(
+        self, instance: object, owner: Any
+    ) -> Union[AssociationProxyInstance[_T], _T, AssociationProxy[_T]]:
+        if owner is None:
+            return self
+        inst = self._as_instance(owner, instance)
+        if inst:
+            return inst.get(instance)
+
+        assert instance is None
+
+        return self
+
+    def __set__(self, instance: object, values: _T) -> None:
+        class_ = type(instance)
+        self._as_instance(class_, instance).set(instance, values)
+
+    def __delete__(self, instance: object) -> None:
+        class_ = type(instance)
+        self._as_instance(class_, instance).delete(instance)
+
+    def for_class(
+        self, class_: Type[Any], obj: Optional[object] = None
+    ) -> AssociationProxyInstance[_T]:
+        r"""Return the internal state local to a specific mapped class.
+
+        E.g., given a class ``User``::
+
+            class User(Base):
+                # ...
+
+                keywords = association_proxy("kws", "keyword")
+
+        If we access this :class:`.AssociationProxy` from
+        :attr:`_orm.Mapper.all_orm_descriptors`, and we want to view the
+        target class for this proxy as mapped by ``User``::
+
+            inspect(User).all_orm_descriptors["keywords"].for_class(User).target_class
+
+        This returns an instance of :class:`.AssociationProxyInstance` that
+        is specific to the ``User`` class.   The :class:`.AssociationProxy`
+        object remains agnostic of its parent class.
+
+        :param class\_: the class that we are returning state for.
+
+        :param obj: optional, an instance of the class that is required
+         if the attribute refers to a polymorphic target, e.g. where we have
+         to look at the type of the actual destination object to get the
+         complete path.
+
+        .. versionadded:: 1.3 - :class:`.AssociationProxy` no longer stores
+           any state specific to a particular parent class; the state is now
+           stored in per-class :class:`.AssociationProxyInstance` objects.
+
+
+        """
+        return self._as_instance(class_, obj)
+
+    def _as_instance(
+        self, class_: Any, obj: Any
+    ) -> AssociationProxyInstance[_T]:
+        try:
+            inst = class_.__dict__[self.key + "_inst"]
+        except KeyError:
+            inst = None
+
+        # avoid exception context
+        if inst is None:
+            owner = self._calc_owner(class_)
+            if owner is not None:
+                inst = AssociationProxyInstance.for_proxy(self, owner, obj)
+                setattr(class_, self.key + "_inst", inst)
+            else:
+                inst = None
+
+        if inst is not None and not inst._is_canonical:
+            # the AssociationProxyInstance can't be generalized
+            # since the proxied attribute is not on the targeted
+            # class, only on subclasses of it, which might be
+            # different.  only return for the specific
+            # object's current value
+            return inst._non_canonical_get_for_object(obj)  # type: ignore
+        else:
+            return inst  # type: ignore  # TODO
+
+    def _calc_owner(self, target_cls: Any) -> Any:
+        # we might be getting invoked for a subclass
+        # that is not mapped yet, in some declarative situations.
+        # save until we are mapped
+        try:
+            insp = inspect(target_cls)
+        except exc.NoInspectionAvailable:
+            # can't find a mapper, don't set owner. if we are a not-yet-mapped
+            # subclass, we can also scan through __mro__ to find a mapped
+            # class, but instead just wait for us to be called again against a
+            # mapped class normally.
+            return None
+        else:
+            return insp.mapper.class_manager.class_
+
+    def _default_getset(
+        self, collection_class: Any
+    ) -> Tuple[_GetterProtocol[Any], _SetterProtocol]:
+        attr = self.value_attr
+        _getter = operator.attrgetter(attr)
+
+        def getter(instance: Any) -> Optional[Any]:
+            return _getter(instance) if instance is not None else None
+
+        if collection_class is dict:
+
+            def dict_setter(instance: Any, k: Any, value: Any) -> None:
+                setattr(instance, attr, value)
+
+            return getter, dict_setter
+
+        else:
+
+            def plain_setter(o: Any, v: Any) -> None:
+                setattr(o, attr, v)
+
+            return getter, plain_setter
+
+    def __repr__(self) -> str:
+        return "AssociationProxy(%r, %r)" % (
+            self.target_collection,
+            self.value_attr,
+        )
+
+
+# the pep-673 Self type does not work in Mypy for a "hybrid"
+# style method that returns type or Self, so for one specific case
+# we still need to use the pre-pep-673 workaround.
+_Self = TypeVar("_Self", bound="AssociationProxyInstance[Any]")
+
+
+class AssociationProxyInstance(SQLORMOperations[_T]):
+    """A per-class object that serves class- and object-specific results.
+
+    This is used by :class:`.AssociationProxy` when it is invoked
+    in terms of a specific class or instance of a class, i.e. when it is
+    used as a regular Python descriptor.
+
+    When referring to the :class:`.AssociationProxy` as a normal Python
+    descriptor, the :class:`.AssociationProxyInstance` is the object that
+    actually serves the information.   Under normal circumstances, its presence
+    is transparent::
+
+        >>> User.keywords.scalar
+        False
+
+    In the special case that the :class:`.AssociationProxy` object is being
+    accessed directly, in order to get an explicit handle to the
+    :class:`.AssociationProxyInstance`, use the
+    :meth:`.AssociationProxy.for_class` method::
+
+        proxy_state = inspect(User).all_orm_descriptors["keywords"].for_class(User)
+
+        # view if proxy object is scalar or not
+        >>> proxy_state.scalar
+        False
+
+    .. versionadded:: 1.3
+
+    """  # noqa
+
+    collection_class: Optional[Type[Any]]
+    parent: _AssociationProxyProtocol[_T]
+
+    def __init__(
+        self,
+        parent: _AssociationProxyProtocol[_T],
+        owning_class: Type[Any],
+        target_class: Type[Any],
+        value_attr: str,
+    ):
+        self.parent = parent
+        self.key = parent.key
+        self.owning_class = owning_class
+        self.target_collection = parent.target_collection
+        self.collection_class = None
+        self.target_class = target_class
+        self.value_attr = value_attr
+
+    target_class: Type[Any]
+    """The intermediary class handled by this
+    :class:`.AssociationProxyInstance`.
+
+    Intercepted append/set/assignment events will result
+    in the generation of new instances of this class.
+
+    """
+
+    @classmethod
+    def for_proxy(
+        cls,
+        parent: AssociationProxy[_T],
+        owning_class: Type[Any],
+        parent_instance: Any,
+    ) -> AssociationProxyInstance[_T]:
+        target_collection = parent.target_collection
+        value_attr = parent.value_attr
+        prop = cast(
+            "orm.RelationshipProperty[_T]",
+            orm.class_mapper(owning_class).get_property(target_collection),
+        )
+
+        # this was never asserted before but this should be made clear.
+        if not isinstance(prop, orm.RelationshipProperty):
+            raise NotImplementedError(
+                "association proxy to a non-relationship "
+                "intermediary is not supported"
+            ) from None
+
+        target_class = prop.mapper.class_
+
+        try:
+            target_assoc = cast(
+                "AssociationProxyInstance[_T]",
+                cls._cls_unwrap_target_assoc_proxy(target_class, value_attr),
+            )
+        except AttributeError:
+            # the proxied attribute doesn't exist on the target class;
+            # return an "ambiguous" instance that will work on a per-object
+            # basis
+            return AmbiguousAssociationProxyInstance(
+                parent, owning_class, target_class, value_attr
+            )
+        except Exception as err:
+            raise exc.InvalidRequestError(
+                f"Association proxy received an unexpected error when "
+                f"trying to retreive attribute "
+                f'"{target_class.__name__}.{parent.value_attr}" from '
+                f'class "{target_class.__name__}": {err}'
+            ) from err
+        else:
+            return cls._construct_for_assoc(
+                target_assoc, parent, owning_class, target_class, value_attr
+            )
+
+    @classmethod
+    def _construct_for_assoc(
+        cls,
+        target_assoc: Optional[AssociationProxyInstance[_T]],
+        parent: _AssociationProxyProtocol[_T],
+        owning_class: Type[Any],
+        target_class: Type[Any],
+        value_attr: str,
+    ) -> AssociationProxyInstance[_T]:
+        if target_assoc is not None:
+            return ObjectAssociationProxyInstance(
+                parent, owning_class, target_class, value_attr
+            )
+
+        attr = getattr(target_class, value_attr)
+        if not hasattr(attr, "_is_internal_proxy"):
+            return AmbiguousAssociationProxyInstance(
+                parent, owning_class, target_class, value_attr
+            )
+        is_object = attr._impl_uses_objects
+        if is_object:
+            return ObjectAssociationProxyInstance(
+                parent, owning_class, target_class, value_attr
+            )
+        else:
+            return ColumnAssociationProxyInstance(
+                parent, owning_class, target_class, value_attr
+            )
+
+    def _get_property(self) -> MapperProperty[Any]:
+        return orm.class_mapper(self.owning_class).get_property(
+            self.target_collection
+        )
+
+    @property
+    def _comparator(self) -> PropComparator[Any]:
+        return getattr(  # type: ignore
+            self.owning_class, self.target_collection
+        ).comparator
+
+    def __clause_element__(self) -> NoReturn:
+        raise NotImplementedError(
+            "The association proxy can't be used as a plain column "
+            "expression; it only works inside of a comparison expression"
+        )
+
+    @classmethod
+    def _cls_unwrap_target_assoc_proxy(
+        cls, target_class: Any, value_attr: str
+    ) -> Optional[AssociationProxyInstance[_T]]:
+        attr = getattr(target_class, value_attr)
+        assert not isinstance(attr, AssociationProxy)
+        if isinstance(attr, AssociationProxyInstance):
+            return attr
+        return None
+
+    @util.memoized_property
+    def _unwrap_target_assoc_proxy(
+        self,
+    ) -> Optional[AssociationProxyInstance[_T]]:
+        return self._cls_unwrap_target_assoc_proxy(
+            self.target_class, self.value_attr
+        )
+
+    @property
+    def remote_attr(self) -> SQLORMOperations[_T]:
+        """The 'remote' class attribute referenced by this
+        :class:`.AssociationProxyInstance`.
+
+        .. seealso::
+
+            :attr:`.AssociationProxyInstance.attr`
+
+            :attr:`.AssociationProxyInstance.local_attr`
+
+        """
+        return cast(
+            "SQLORMOperations[_T]", getattr(self.target_class, self.value_attr)
+        )
+
+    @property
+    def local_attr(self) -> SQLORMOperations[Any]:
+        """The 'local' class attribute referenced by this
+        :class:`.AssociationProxyInstance`.
+
+        .. seealso::
+
+            :attr:`.AssociationProxyInstance.attr`
+
+            :attr:`.AssociationProxyInstance.remote_attr`
+
+        """
+        return cast(
+            "SQLORMOperations[Any]",
+            getattr(self.owning_class, self.target_collection),
+        )
+
+    @property
+    def attr(self) -> Tuple[SQLORMOperations[Any], SQLORMOperations[_T]]:
+        """Return a tuple of ``(local_attr, remote_attr)``.
+
+        This attribute was originally intended to facilitate using the
+        :meth:`_query.Query.join` method to join across the two relationships
+        at once, however this makes use of a deprecated calling style.
+
+        To use :meth:`_sql.select.join` or :meth:`_orm.Query.join` with
+        an association proxy, the current method is to make use of the
+        :attr:`.AssociationProxyInstance.local_attr` and
+        :attr:`.AssociationProxyInstance.remote_attr` attributes separately::
+
+            stmt = (
+                select(Parent)
+                .join(Parent.proxied.local_attr)
+                .join(Parent.proxied.remote_attr)
+            )
+
+        A future release may seek to provide a more succinct join pattern
+        for association proxy attributes.
+
+        .. seealso::
+
+            :attr:`.AssociationProxyInstance.local_attr`
+
+            :attr:`.AssociationProxyInstance.remote_attr`
+
+        """
+        return (self.local_attr, self.remote_attr)
+
+    @util.memoized_property
+    def scalar(self) -> bool:
+        """Return ``True`` if this :class:`.AssociationProxyInstance`
+        proxies a scalar relationship on the local side."""
+
+        scalar = not self._get_property().uselist
+        if scalar:
+            self._initialize_scalar_accessors()
+        return scalar
+
+    @util.memoized_property
+    def _value_is_scalar(self) -> bool:
+        return (
+            not self._get_property()
+            .mapper.get_property(self.value_attr)
+            .uselist
+        )
+
+    @property
+    def _target_is_object(self) -> bool:
+        raise NotImplementedError()
+
+    _scalar_get: _GetterProtocol[_T]
+    _scalar_set: _PlainSetterProtocol[_T]
+
+    def _initialize_scalar_accessors(self) -> None:
+        if self.parent.getset_factory:
+            get, set_ = self.parent.getset_factory(None, self)
+        else:
+            get, set_ = self.parent._default_getset(None)
+        self._scalar_get, self._scalar_set = get, cast(
+            "_PlainSetterProtocol[_T]", set_
+        )
+
+    def _default_getset(
+        self, collection_class: Any
+    ) -> Tuple[_GetterProtocol[Any], _SetterProtocol]:
+        attr = self.value_attr
+        _getter = operator.attrgetter(attr)
+
+        def getter(instance: Any) -> Optional[_T]:
+            return _getter(instance) if instance is not None else None
+
+        if collection_class is dict:
+
+            def dict_setter(instance: Any, k: Any, value: _T) -> None:
+                setattr(instance, attr, value)
+
+            return getter, dict_setter
+        else:
+
+            def plain_setter(o: Any, v: _T) -> None:
+                setattr(o, attr, v)
+
+            return getter, plain_setter
+
+    @util.ro_non_memoized_property
+    def info(self) -> _InfoType:
+        return self.parent.info
+
+    @overload
+    def get(self: _Self, obj: Literal[None]) -> _Self: ...
+
+    @overload
+    def get(self, obj: Any) -> _T: ...
+
+    def get(
+        self, obj: Any
+    ) -> Union[Optional[_T], AssociationProxyInstance[_T]]:
+        if obj is None:
+            return self
+
+        proxy: _T
+
+        if self.scalar:
+            target = getattr(obj, self.target_collection)
+            return self._scalar_get(target)
+        else:
+            try:
+                # If the owning instance is reborn (orm session resurrect,
+                # etc.), refresh the proxy cache.
+                creator_id, self_id, proxy = cast(
+                    "Tuple[int, int, _T]", getattr(obj, self.key)
+                )
+            except AttributeError:
+                pass
+            else:
+                if id(obj) == creator_id and id(self) == self_id:
+                    assert self.collection_class is not None
+                    return proxy
+
+            self.collection_class, proxy = self._new(
+                _lazy_collection(obj, self.target_collection)
+            )
+            setattr(obj, self.key, (id(obj), id(self), proxy))
+            return proxy
+
+    def set(self, obj: Any, values: _T) -> None:
+        if self.scalar:
+            creator = cast(
+                "_PlainCreatorProtocol[_T]",
+                (
+                    self.parent.creator
+                    if self.parent.creator
+                    else self.target_class
+                ),
+            )
+            target = getattr(obj, self.target_collection)
+            if target is None:
+                if (
+                    values is None
+                    and not self.parent.create_on_none_assignment
+                ):
+                    return
+                setattr(obj, self.target_collection, creator(values))
+            else:
+                self._scalar_set(target, values)
+                if values is None and self.parent.cascade_scalar_deletes:
+                    setattr(obj, self.target_collection, None)
+        else:
+            proxy = self.get(obj)
+            assert self.collection_class is not None
+            if proxy is not values:
+                proxy._bulk_replace(self, values)
+
+    def delete(self, obj: Any) -> None:
+        if self.owning_class is None:
+            self._calc_owner(obj, None)
+
+        if self.scalar:
+            target = getattr(obj, self.target_collection)
+            if target is not None:
+                delattr(target, self.value_attr)
+        delattr(obj, self.target_collection)
+
+    def _new(
+        self, lazy_collection: _LazyCollectionProtocol[_T]
+    ) -> Tuple[Type[Any], _T]:
+        creator = (
+            self.parent.creator
+            if self.parent.creator is not None
+            else cast("_CreatorProtocol", self.target_class)
+        )
+        collection_class = util.duck_type_collection(lazy_collection())
+
+        if collection_class is None:
+            raise exc.InvalidRequestError(
+                f"lazy collection factory did not return a "
+                f"valid collection type, got {collection_class}"
+            )
+        if self.parent.proxy_factory:
+            return (
+                collection_class,
+                self.parent.proxy_factory(
+                    lazy_collection, creator, self.value_attr, self
+                ),
+            )
+
+        if self.parent.getset_factory:
+            getter, setter = self.parent.getset_factory(collection_class, self)
+        else:
+            getter, setter = self.parent._default_getset(collection_class)
+
+        if collection_class is list:
+            return (
+                collection_class,
+                cast(
+                    _T,
+                    _AssociationList(
+                        lazy_collection, creator, getter, setter, self
+                    ),
+                ),
+            )
+        elif collection_class is dict:
+            return (
+                collection_class,
+                cast(
+                    _T,
+                    _AssociationDict(
+                        lazy_collection, creator, getter, setter, self
+                    ),
+                ),
+            )
+        elif collection_class is set:
+            return (
+                collection_class,
+                cast(
+                    _T,
+                    _AssociationSet(
+                        lazy_collection, creator, getter, setter, self
+                    ),
+                ),
+            )
+        else:
+            raise exc.ArgumentError(
+                "could not guess which interface to use for "
+                'collection_class "%s" backing "%s"; specify a '
+                "proxy_factory and proxy_bulk_set manually"
+                % (self.collection_class, self.target_collection)
+            )
+
+    def _set(
+        self, proxy: _AssociationCollection[Any], values: Iterable[Any]
+    ) -> None:
+        if self.parent.proxy_bulk_set:
+            self.parent.proxy_bulk_set(proxy, values)
+        elif self.collection_class is list:
+            cast("_AssociationList[Any]", proxy).extend(values)
+        elif self.collection_class is dict:
+            cast("_AssociationDict[Any, Any]", proxy).update(values)
+        elif self.collection_class is set:
+            cast("_AssociationSet[Any]", proxy).update(values)
+        else:
+            raise exc.ArgumentError(
+                "no proxy_bulk_set supplied for custom "
+                "collection_class implementation"
+            )
+
+    def _inflate(self, proxy: _AssociationCollection[Any]) -> None:
+        creator = (
+            self.parent.creator
+            and self.parent.creator
+            or cast(_CreatorProtocol, self.target_class)
+        )
+
+        if self.parent.getset_factory:
+            getter, setter = self.parent.getset_factory(
+                self.collection_class, self
+            )
+        else:
+            getter, setter = self.parent._default_getset(self.collection_class)
+
+        proxy.creator = creator
+        proxy.getter = getter
+        proxy.setter = setter
+
+    def _criterion_exists(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[bool]:
+        is_has = kwargs.pop("is_has", None)
+
+        target_assoc = self._unwrap_target_assoc_proxy
+        if target_assoc is not None:
+            inner = target_assoc._criterion_exists(
+                criterion=criterion, **kwargs
+            )
+            return self._comparator._criterion_exists(inner)
+
+        if self._target_is_object:
+            attr = getattr(self.target_class, self.value_attr)
+            value_expr = attr.comparator._criterion_exists(criterion, **kwargs)
+        else:
+            if kwargs:
+                raise exc.ArgumentError(
+                    "Can't apply keyword arguments to column-targeted "
+                    "association proxy; use =="
+                )
+            elif is_has and criterion is not None:
+                raise exc.ArgumentError(
+                    "Non-empty has() not allowed for "
+                    "column-targeted association proxy; use =="
+                )
+
+            value_expr = criterion
+
+        return self._comparator._criterion_exists(value_expr)
+
+    def any(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[bool]:
+        """Produce a proxied 'any' expression using EXISTS.
+
+        This expression will be a composed product
+        using the :meth:`.Relationship.Comparator.any`
+        and/or :meth:`.Relationship.Comparator.has`
+        operators of the underlying proxied attributes.
+
+        """
+        if self._unwrap_target_assoc_proxy is None and (
+            self.scalar
+            and (not self._target_is_object or self._value_is_scalar)
+        ):
+            raise exc.InvalidRequestError(
+                "'any()' not implemented for scalar attributes. Use has()."
+            )
+        return self._criterion_exists(
+            criterion=criterion, is_has=False, **kwargs
+        )
+
+    def has(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[bool]:
+        """Produce a proxied 'has' expression using EXISTS.
+
+        This expression will be a composed product
+        using the :meth:`.Relationship.Comparator.any`
+        and/or :meth:`.Relationship.Comparator.has`
+        operators of the underlying proxied attributes.
+
+        """
+        if self._unwrap_target_assoc_proxy is None and (
+            not self.scalar
+            or (self._target_is_object and not self._value_is_scalar)
+        ):
+            raise exc.InvalidRequestError(
+                "'has()' not implemented for collections. Use any()."
+            )
+        return self._criterion_exists(
+            criterion=criterion, is_has=True, **kwargs
+        )
+
+    def __repr__(self) -> str:
+        return "%s(%r)" % (self.__class__.__name__, self.parent)
+
+
+class AmbiguousAssociationProxyInstance(AssociationProxyInstance[_T]):
+    """an :class:`.AssociationProxyInstance` where we cannot determine
+    the type of target object.
+    """
+
+    _is_canonical = False
+
+    def _ambiguous(self) -> NoReturn:
+        raise AttributeError(
+            "Association proxy %s.%s refers to an attribute '%s' that is not "
+            "directly mapped on class %s; therefore this operation cannot "
+            "proceed since we don't know what type of object is referred "
+            "towards"
+            % (
+                self.owning_class.__name__,
+                self.target_collection,
+                self.value_attr,
+                self.target_class,
+            )
+        )
+
+    def get(self, obj: Any) -> Any:
+        if obj is None:
+            return self
+        else:
+            return super().get(obj)
+
+    def __eq__(self, obj: object) -> NoReturn:
+        self._ambiguous()
+
+    def __ne__(self, obj: object) -> NoReturn:
+        self._ambiguous()
+
+    def any(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> NoReturn:
+        self._ambiguous()
+
+    def has(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> NoReturn:
+        self._ambiguous()
+
+    @util.memoized_property
+    def _lookup_cache(self) -> Dict[Type[Any], AssociationProxyInstance[_T]]:
+        # mapping of <subclass>->AssociationProxyInstance.
+        # e.g. proxy is A-> A.b -> B -> B.b_attr, but B.b_attr doesn't exist;
+        # only B1(B) and B2(B) have "b_attr", keys in here would be B1, B2
+        return {}
+
+    def _non_canonical_get_for_object(
+        self, parent_instance: Any
+    ) -> AssociationProxyInstance[_T]:
+        if parent_instance is not None:
+            actual_obj = getattr(parent_instance, self.target_collection)
+            if actual_obj is not None:
+                try:
+                    insp = inspect(actual_obj)
+                except exc.NoInspectionAvailable:
+                    pass
+                else:
+                    mapper = insp.mapper
+                    instance_class = mapper.class_
+                    if instance_class not in self._lookup_cache:
+                        self._populate_cache(instance_class, mapper)
+
+                    try:
+                        return self._lookup_cache[instance_class]
+                    except KeyError:
+                        pass
+
+        # no object or ambiguous object given, so return "self", which
+        # is a proxy with generally only instance-level functionality
+        return self
+
+    def _populate_cache(
+        self, instance_class: Any, mapper: Mapper[Any]
+    ) -> None:
+        prop = orm.class_mapper(self.owning_class).get_property(
+            self.target_collection
+        )
+
+        if mapper.isa(prop.mapper):
+            target_class = instance_class
+            try:
+                target_assoc = self._cls_unwrap_target_assoc_proxy(
+                    target_class, self.value_attr
+                )
+            except AttributeError:
+                pass
+            else:
+                self._lookup_cache[instance_class] = self._construct_for_assoc(
+                    cast("AssociationProxyInstance[_T]", target_assoc),
+                    self.parent,
+                    self.owning_class,
+                    target_class,
+                    self.value_attr,
+                )
+
+
+class ObjectAssociationProxyInstance(AssociationProxyInstance[_T]):
+    """an :class:`.AssociationProxyInstance` that has an object as a target."""
+
+    _target_is_object: bool = True
+    _is_canonical = True
+
+    def contains(self, other: Any, **kw: Any) -> ColumnElement[bool]:
+        """Produce a proxied 'contains' expression using EXISTS.
+
+        This expression will be a composed product
+        using the :meth:`.Relationship.Comparator.any`,
+        :meth:`.Relationship.Comparator.has`,
+        and/or :meth:`.Relationship.Comparator.contains`
+        operators of the underlying proxied attributes.
+        """
+
+        target_assoc = self._unwrap_target_assoc_proxy
+        if target_assoc is not None:
+            return self._comparator._criterion_exists(
+                target_assoc.contains(other)
+                if not target_assoc.scalar
+                else target_assoc == other
+            )
+        elif (
+            self._target_is_object
+            and self.scalar
+            and not self._value_is_scalar
+        ):
+            return self._comparator.has(
+                getattr(self.target_class, self.value_attr).contains(other)
+            )
+        elif self._target_is_object and self.scalar and self._value_is_scalar:
+            raise exc.InvalidRequestError(
+                "contains() doesn't apply to a scalar object endpoint; use =="
+            )
+        else:
+            return self._comparator._criterion_exists(
+                **{self.value_attr: other}
+            )
+
+    def __eq__(self, obj: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+        # note the has() here will fail for collections; eq_()
+        # is only allowed with a scalar.
+        if obj is None:
+            return or_(
+                self._comparator.has(**{self.value_attr: obj}),
+                self._comparator == None,
+            )
+        else:
+            return self._comparator.has(**{self.value_attr: obj})
+
+    def __ne__(self, obj: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+        # note the has() here will fail for collections; eq_()
+        # is only allowed with a scalar.
+        return self._comparator.has(
+            getattr(self.target_class, self.value_attr) != obj
+        )
+
+
+class ColumnAssociationProxyInstance(AssociationProxyInstance[_T]):
+    """an :class:`.AssociationProxyInstance` that has a database column as a
+    target.
+    """
+
+    _target_is_object: bool = False
+    _is_canonical = True
+
+    def __eq__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+        # special case "is None" to check for no related row as well
+        expr = self._criterion_exists(
+            self.remote_attr.operate(operators.eq, other)
+        )
+        if other is None:
+            return or_(expr, self._comparator == None)
+        else:
+            return expr
+
+    def operate(
+        self, op: operators.OperatorType, *other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return self._criterion_exists(
+            self.remote_attr.operate(op, *other, **kwargs)
+        )
+
+
+class _lazy_collection(_LazyCollectionProtocol[_T]):
+    def __init__(self, obj: Any, target: str):
+        self.parent = obj
+        self.target = target
+
+    def __call__(
+        self,
+    ) -> Union[MutableSet[_T], MutableMapping[Any, _T], MutableSequence[_T]]:
+        return getattr(self.parent, self.target)  # type: ignore[no-any-return]
+
+    def __getstate__(self) -> Any:
+        return {"obj": self.parent, "target": self.target}
+
+    def __setstate__(self, state: Any) -> None:
+        self.parent = state["obj"]
+        self.target = state["target"]
+
+
+_IT = TypeVar("_IT", bound="Any")
+"""instance type - this is the type of object inside a collection.
+
+this is not the same as the _T of AssociationProxy and
+AssociationProxyInstance itself, which will often refer to the
+collection[_IT] type.
+
+"""
+
+
+class _AssociationCollection(Generic[_IT]):
+    getter: _GetterProtocol[_IT]
+    """A function.  Given an associated object, return the 'value'."""
+
+    creator: _CreatorProtocol
+    """
+    A function that creates new target entities.  Given one parameter:
+    value.  This assertion is assumed::
+
+    obj = creator(somevalue)
+    assert getter(obj) == somevalue
+    """
+
+    parent: AssociationProxyInstance[_IT]
+    setter: _SetterProtocol
+    """A function.  Given an associated object and a value, store that
+        value on the object.
+    """
+
+    lazy_collection: _LazyCollectionProtocol[_IT]
+    """A callable returning a list-based collection of entities (usually an
+          object attribute managed by a SQLAlchemy relationship())"""
+
+    def __init__(
+        self,
+        lazy_collection: _LazyCollectionProtocol[_IT],
+        creator: _CreatorProtocol,
+        getter: _GetterProtocol[_IT],
+        setter: _SetterProtocol,
+        parent: AssociationProxyInstance[_IT],
+    ):
+        """Constructs an _AssociationCollection.
+
+        This will always be a subclass of either _AssociationList,
+        _AssociationSet, or _AssociationDict.
+
+        """
+        self.lazy_collection = lazy_collection
+        self.creator = creator
+        self.getter = getter
+        self.setter = setter
+        self.parent = parent
+
+    if typing.TYPE_CHECKING:
+        col: Collection[_IT]
+    else:
+        col = property(lambda self: self.lazy_collection())
+
+    def __len__(self) -> int:
+        return len(self.col)
+
+    def __bool__(self) -> bool:
+        return bool(self.col)
+
+    def __getstate__(self) -> Any:
+        return {"parent": self.parent, "lazy_collection": self.lazy_collection}
+
+    def __setstate__(self, state: Any) -> None:
+        self.parent = state["parent"]
+        self.lazy_collection = state["lazy_collection"]
+        self.parent._inflate(self)
+
+    def clear(self) -> None:
+        raise NotImplementedError()
+
+
+class _AssociationSingleItem(_AssociationCollection[_T]):
+    setter: _PlainSetterProtocol[_T]
+    creator: _PlainCreatorProtocol[_T]
+
+    def _create(self, value: _T) -> Any:
+        return self.creator(value)
+
+    def _get(self, object_: Any) -> _T:
+        return self.getter(object_)
+
+    def _bulk_replace(
+        self, assoc_proxy: AssociationProxyInstance[Any], values: Iterable[_IT]
+    ) -> None:
+        self.clear()
+        assoc_proxy._set(self, values)
+
+
+class _AssociationList(_AssociationSingleItem[_T], MutableSequence[_T]):
+    """Generic, converting, list-to-list proxy."""
+
+    col: MutableSequence[_T]
+
+    def _set(self, object_: Any, value: _T) -> None:
+        self.setter(object_, value)
+
+    @overload
+    def __getitem__(self, index: int) -> _T: ...
+
+    @overload
+    def __getitem__(self, index: slice) -> MutableSequence[_T]: ...
+
+    def __getitem__(
+        self, index: Union[int, slice]
+    ) -> Union[_T, MutableSequence[_T]]:
+        if not isinstance(index, slice):
+            return self._get(self.col[index])
+        else:
+            return [self._get(member) for member in self.col[index]]
+
+    @overload
+    def __setitem__(self, index: int, value: _T) -> None: ...
+
+    @overload
+    def __setitem__(self, index: slice, value: Iterable[_T]) -> None: ...
+
+    def __setitem__(
+        self, index: Union[int, slice], value: Union[_T, Iterable[_T]]
+    ) -> None:
+        if not isinstance(index, slice):
+            self._set(self.col[index], cast("_T", value))
+        else:
+            if index.stop is None:
+                stop = len(self)
+            elif index.stop < 0:
+                stop = len(self) + index.stop
+            else:
+                stop = index.stop
+            step = index.step or 1
+
+            start = index.start or 0
+            rng = list(range(index.start or 0, stop, step))
+
+            sized_value = list(value)
+
+            if step == 1:
+                for i in rng:
+                    del self[start]
+                i = start
+                for item in sized_value:
+                    self.insert(i, item)
+                    i += 1
+            else:
+                if len(sized_value) != len(rng):
+                    raise ValueError(
+                        "attempt to assign sequence of size %s to "
+                        "extended slice of size %s"
+                        % (len(sized_value), len(rng))
+                    )
+                for i, item in zip(rng, value):
+                    self._set(self.col[i], item)
+
+    @overload
+    def __delitem__(self, index: int) -> None: ...
+
+    @overload
+    def __delitem__(self, index: slice) -> None: ...
+
+    def __delitem__(self, index: Union[slice, int]) -> None:
+        del self.col[index]
+
+    def __contains__(self, value: object) -> bool:
+        for member in self.col:
+            # testlib.pragma exempt:__eq__
+            if self._get(member) == value:
+                return True
+        return False
+
+    def __iter__(self) -> Iterator[_T]:
+        """Iterate over proxied values.
+
+        For the actual domain objects, iterate over .col instead or
+        just use the underlying collection directly from its property
+        on the parent.
+        """
+
+        for member in self.col:
+            yield self._get(member)
+        return
+
+    def append(self, value: _T) -> None:
+        col = self.col
+        item = self._create(value)
+        col.append(item)
+
+    def count(self, value: Any) -> int:
+        count = 0
+        for v in self:
+            if v == value:
+                count += 1
+        return count
+
+    def extend(self, values: Iterable[_T]) -> None:
+        for v in values:
+            self.append(v)
+
+    def insert(self, index: int, value: _T) -> None:
+        self.col[index:index] = [self._create(value)]
+
+    def pop(self, index: int = -1) -> _T:
+        return self.getter(self.col.pop(index))
+
+    def remove(self, value: _T) -> None:
+        for i, val in enumerate(self):
+            if val == value:
+                del self.col[i]
+                return
+        raise ValueError("value not in list")
+
+    def reverse(self) -> NoReturn:
+        """Not supported, use reversed(mylist)"""
+
+        raise NotImplementedError()
+
+    def sort(self) -> NoReturn:
+        """Not supported, use sorted(mylist)"""
+
+        raise NotImplementedError()
+
+    def clear(self) -> None:
+        del self.col[0 : len(self.col)]
+
+    def __eq__(self, other: object) -> bool:
+        return list(self) == other
+
+    def __ne__(self, other: object) -> bool:
+        return list(self) != other
+
+    def __lt__(self, other: List[_T]) -> bool:
+        return list(self) < other
+
+    def __le__(self, other: List[_T]) -> bool:
+        return list(self) <= other
+
+    def __gt__(self, other: List[_T]) -> bool:
+        return list(self) > other
+
+    def __ge__(self, other: List[_T]) -> bool:
+        return list(self) >= other
+
+    def __add__(self, other: List[_T]) -> List[_T]:
+        try:
+            other = list(other)
+        except TypeError:
+            return NotImplemented
+        return list(self) + other
+
+    def __radd__(self, other: List[_T]) -> List[_T]:
+        try:
+            other = list(other)
+        except TypeError:
+            return NotImplemented
+        return other + list(self)
+
+    def __mul__(self, n: SupportsIndex) -> List[_T]:
+        if not isinstance(n, int):
+            return NotImplemented
+        return list(self) * n
+
+    def __rmul__(self, n: SupportsIndex) -> List[_T]:
+        if not isinstance(n, int):
+            return NotImplemented
+        return n * list(self)
+
+    def __iadd__(self, iterable: Iterable[_T]) -> Self:
+        self.extend(iterable)
+        return self
+
+    def __imul__(self, n: SupportsIndex) -> Self:
+        # unlike a regular list *=, proxied __imul__ will generate unique
+        # backing objects for each copy.  *= on proxied lists is a bit of
+        # a stretch anyhow, and this interpretation of the __imul__ contract
+        # is more plausibly useful than copying the backing objects.
+        if not isinstance(n, int):
+            raise NotImplementedError()
+        if n == 0:
+            self.clear()
+        elif n > 1:
+            self.extend(list(self) * (n - 1))
+        return self
+
+    if typing.TYPE_CHECKING:
+        # TODO: no idea how to do this without separate "stub"
+        def index(
+            self, value: Any, start: int = ..., stop: int = ...
+        ) -> int: ...
+
+    else:
+
+        def index(self, value: Any, *arg) -> int:
+            ls = list(self)
+            return ls.index(value, *arg)
+
+    def copy(self) -> List[_T]:
+        return list(self)
+
+    def __repr__(self) -> str:
+        return repr(list(self))
+
+    def __hash__(self) -> NoReturn:
+        raise TypeError("%s objects are unhashable" % type(self).__name__)
+
+    if not typing.TYPE_CHECKING:
+        for func_name, func in list(locals().items()):
+            if (
+                callable(func)
+                and func.__name__ == func_name
+                and not func.__doc__
+                and hasattr(list, func_name)
+            ):
+                func.__doc__ = getattr(list, func_name).__doc__
+        del func_name, func
+
+
+class _AssociationDict(_AssociationCollection[_VT], MutableMapping[_KT, _VT]):
+    """Generic, converting, dict-to-dict proxy."""
+
+    setter: _DictSetterProtocol[_VT]
+    creator: _KeyCreatorProtocol[_VT]
+    col: MutableMapping[_KT, Optional[_VT]]
+
+    def _create(self, key: _KT, value: Optional[_VT]) -> Any:
+        return self.creator(key, value)
+
+    def _get(self, object_: Any) -> _VT:
+        return self.getter(object_)
+
+    def _set(self, object_: Any, key: _KT, value: _VT) -> None:
+        return self.setter(object_, key, value)
+
+    def __getitem__(self, key: _KT) -> _VT:
+        return self._get(self.col[key])
+
+    def __setitem__(self, key: _KT, value: _VT) -> None:
+        if key in self.col:
+            self._set(self.col[key], key, value)
+        else:
+            self.col[key] = self._create(key, value)
+
+    def __delitem__(self, key: _KT) -> None:
+        del self.col[key]
+
+    def __contains__(self, key: object) -> bool:
+        return key in self.col
+
+    def __iter__(self) -> Iterator[_KT]:
+        return iter(self.col.keys())
+
+    def clear(self) -> None:
+        self.col.clear()
+
+    def __eq__(self, other: object) -> bool:
+        return dict(self) == other
+
+    def __ne__(self, other: object) -> bool:
+        return dict(self) != other
+
+    def __repr__(self) -> str:
+        return repr(dict(self))
+
+    @overload
+    def get(self, __key: _KT) -> Optional[_VT]: ...
+
+    @overload
+    def get(self, __key: _KT, default: Union[_VT, _T]) -> Union[_VT, _T]: ...
+
+    def get(
+        self, key: _KT, default: Optional[Union[_VT, _T]] = None
+    ) -> Union[_VT, _T, None]:
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def setdefault(self, key: _KT, default: Optional[_VT] = None) -> _VT:
+        # TODO: again, no idea how to create an actual MutableMapping.
+        # default must allow None, return type can't include None,
+        # the stub explicitly allows for default of None with a cryptic message
+        # "This overload should be allowed only if the value type is
+        # compatible with None.".
+        if key not in self.col:
+            self.col[key] = self._create(key, default)
+            return default  # type: ignore
+        else:
+            return self[key]
+
+    def keys(self) -> KeysView[_KT]:
+        return self.col.keys()
+
+    def items(self) -> ItemsView[_KT, _VT]:
+        return ItemsView(self)
+
+    def values(self) -> ValuesView[_VT]:
+        return ValuesView(self)
+
+    @overload
+    def pop(self, __key: _KT) -> _VT: ...
+
+    @overload
+    def pop(
+        self, __key: _KT, default: Union[_VT, _T] = ...
+    ) -> Union[_VT, _T]: ...
+
+    def pop(self, __key: _KT, *arg: Any, **kw: Any) -> Union[_VT, _T]:
+        member = self.col.pop(__key, *arg, **kw)
+        return self._get(member)
+
+    def popitem(self) -> Tuple[_KT, _VT]:
+        item = self.col.popitem()
+        return (item[0], self._get(item[1]))
+
+    @overload
+    def update(
+        self, __m: SupportsKeysAndGetItem[_KT, _VT], **kwargs: _VT
+    ) -> None: ...
+
+    @overload
+    def update(
+        self, __m: Iterable[tuple[_KT, _VT]], **kwargs: _VT
+    ) -> None: ...
+
+    @overload
+    def update(self, **kwargs: _VT) -> None: ...
+
+    def update(self, *a: Any, **kw: Any) -> None:
+        up: Dict[_KT, _VT] = {}
+        up.update(*a, **kw)
+
+        for key, value in up.items():
+            self[key] = value
+
+    def _bulk_replace(
+        self,
+        assoc_proxy: AssociationProxyInstance[Any],
+        values: Mapping[_KT, _VT],
+    ) -> None:
+        existing = set(self)
+        constants = existing.intersection(values or ())
+        additions = set(values or ()).difference(constants)
+        removals = existing.difference(constants)
+
+        for key, member in values.items() or ():
+            if key in additions:
+                self[key] = member
+            elif key in constants:
+                self[key] = member
+
+        for key in removals:
+            del self[key]
+
+    def copy(self) -> Dict[_KT, _VT]:
+        return dict(self.items())
+
+    def __hash__(self) -> NoReturn:
+        raise TypeError("%s objects are unhashable" % type(self).__name__)
+
+    if not typing.TYPE_CHECKING:
+        for func_name, func in list(locals().items()):
+            if (
+                callable(func)
+                and func.__name__ == func_name
+                and not func.__doc__
+                and hasattr(dict, func_name)
+            ):
+                func.__doc__ = getattr(dict, func_name).__doc__
+        del func_name, func
+
+
+class _AssociationSet(_AssociationSingleItem[_T], MutableSet[_T]):
+    """Generic, converting, set-to-set proxy."""
+
+    col: MutableSet[_T]
+
+    def __len__(self) -> int:
+        return len(self.col)
+
+    def __bool__(self) -> bool:
+        if self.col:
+            return True
+        else:
+            return False
+
+    def __contains__(self, __o: object) -> bool:
+        for member in self.col:
+            if self._get(member) == __o:
+                return True
+        return False
+
+    def __iter__(self) -> Iterator[_T]:
+        """Iterate over proxied values.
+
+        For the actual domain objects, iterate over .col instead or just use
+        the underlying collection directly from its property on the parent.
+
+        """
+        for member in self.col:
+            yield self._get(member)
+        return
+
+    def add(self, __element: _T) -> None:
+        if __element not in self:
+            self.col.add(self._create(__element))
+
+    # for discard and remove, choosing a more expensive check strategy rather
+    # than call self.creator()
+    def discard(self, __element: _T) -> None:
+        for member in self.col:
+            if self._get(member) == __element:
+                self.col.discard(member)
+                break
+
+    def remove(self, __element: _T) -> None:
+        for member in self.col:
+            if self._get(member) == __element:
+                self.col.discard(member)
+                return
+        raise KeyError(__element)
+
+    def pop(self) -> _T:
+        if not self.col:
+            raise KeyError("pop from an empty set")
+        member = self.col.pop()
+        return self._get(member)
+
+    def update(self, *s: Iterable[_T]) -> None:
+        for iterable in s:
+            for value in iterable:
+                self.add(value)
+
+    def _bulk_replace(self, assoc_proxy: Any, values: Iterable[_T]) -> None:
+        existing = set(self)
+        constants = existing.intersection(values or ())
+        additions = set(values or ()).difference(constants)
+        removals = existing.difference(constants)
+
+        appender = self.add
+        remover = self.remove
+
+        for member in values or ():
+            if member in additions:
+                appender(member)
+            elif member in constants:
+                appender(member)
+
+        for member in removals:
+            remover(member)
+
+    def __ior__(  # type: ignore
+        self, other: AbstractSet[_S]
+    ) -> MutableSet[Union[_T, _S]]:
+        if not collections._set_binops_check_strict(self, other):
+            raise NotImplementedError()
+        for value in other:
+            self.add(value)
+        return self
+
+    def _set(self) -> Set[_T]:
+        return set(iter(self))
+
+    def union(self, *s: Iterable[_S]) -> MutableSet[Union[_T, _S]]:
+        return set(self).union(*s)
+
+    def __or__(self, __s: AbstractSet[_S]) -> MutableSet[Union[_T, _S]]:
+        return self.union(__s)
+
+    def difference(self, *s: Iterable[Any]) -> MutableSet[_T]:
+        return set(self).difference(*s)
+
+    def __sub__(self, s: AbstractSet[Any]) -> MutableSet[_T]:
+        return self.difference(s)
+
+    def difference_update(self, *s: Iterable[Any]) -> None:
+        for other in s:
+            for value in other:
+                self.discard(value)
+
+    def __isub__(self, s: AbstractSet[Any]) -> Self:
+        if not collections._set_binops_check_strict(self, s):
+            raise NotImplementedError()
+        for value in s:
+            self.discard(value)
+        return self
+
+    def intersection(self, *s: Iterable[Any]) -> MutableSet[_T]:
+        return set(self).intersection(*s)
+
+    def __and__(self, s: AbstractSet[Any]) -> MutableSet[_T]:
+        return self.intersection(s)
+
+    def intersection_update(self, *s: Iterable[Any]) -> None:
+        for other in s:
+            want, have = self.intersection(other), set(self)
+
+            remove, add = have - want, want - have
+
+            for value in remove:
+                self.remove(value)
+            for value in add:
+                self.add(value)
+
+    def __iand__(self, s: AbstractSet[Any]) -> Self:
+        if not collections._set_binops_check_strict(self, s):
+            raise NotImplementedError()
+        want = self.intersection(s)
+        have: Set[_T] = set(self)
+
+        remove, add = have - want, want - have
+
+        for value in remove:
+            self.remove(value)
+        for value in add:
+            self.add(value)
+        return self
+
+    def symmetric_difference(self, __s: Iterable[_T]) -> MutableSet[_T]:
+        return set(self).symmetric_difference(__s)
+
+    def __xor__(self, s: AbstractSet[_S]) -> MutableSet[Union[_T, _S]]:
+        return self.symmetric_difference(s)
+
+    def symmetric_difference_update(self, other: Iterable[Any]) -> None:
+        want, have = self.symmetric_difference(other), set(self)
+
+        remove, add = have - want, want - have
+
+        for value in remove:
+            self.remove(value)
+        for value in add:
+            self.add(value)
+
+    def __ixor__(self, other: AbstractSet[_S]) -> MutableSet[Union[_T, _S]]:  # type: ignore  # noqa: E501
+        if not collections._set_binops_check_strict(self, other):
+            raise NotImplementedError()
+
+        self.symmetric_difference_update(other)
+        return self
+
+    def issubset(self, __s: Iterable[Any]) -> bool:
+        return set(self).issubset(__s)
+
+    def issuperset(self, __s: Iterable[Any]) -> bool:
+        return set(self).issuperset(__s)
+
+    def clear(self) -> None:
+        self.col.clear()
+
+    def copy(self) -> AbstractSet[_T]:
+        return set(self)
+
+    def __eq__(self, other: object) -> bool:
+        return set(self) == other
+
+    def __ne__(self, other: object) -> bool:
+        return set(self) != other
+
+    def __lt__(self, other: AbstractSet[Any]) -> bool:
+        return set(self) < other
+
+    def __le__(self, other: AbstractSet[Any]) -> bool:
+        return set(self) <= other
+
+    def __gt__(self, other: AbstractSet[Any]) -> bool:
+        return set(self) > other
+
+    def __ge__(self, other: AbstractSet[Any]) -> bool:
+        return set(self) >= other
+
+    def __repr__(self) -> str:
+        return repr(set(self))
+
+    def __hash__(self) -> NoReturn:
+        raise TypeError("%s objects are unhashable" % type(self).__name__)
+
+    if not typing.TYPE_CHECKING:
+        for func_name, func in list(locals().items()):
+            if (
+                callable(func)
+                and func.__name__ == func_name
+                and not func.__doc__
+                and hasattr(set, func_name)
+            ):
+                func.__doc__ = getattr(set, func_name).__doc__
+        del func_name, func
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__init__.py
new file mode 100644
index 00000000..7d8a04bd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__init__.py
@@ -0,0 +1,25 @@
+# ext/asyncio/__init__.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from .engine import async_engine_from_config as async_engine_from_config
+from .engine import AsyncConnection as AsyncConnection
+from .engine import AsyncEngine as AsyncEngine
+from .engine import AsyncTransaction as AsyncTransaction
+from .engine import create_async_engine as create_async_engine
+from .engine import create_async_pool_from_url as create_async_pool_from_url
+from .result import AsyncMappingResult as AsyncMappingResult
+from .result import AsyncResult as AsyncResult
+from .result import AsyncScalarResult as AsyncScalarResult
+from .result import AsyncTupleResult as AsyncTupleResult
+from .scoping import async_scoped_session as async_scoped_session
+from .session import async_object_session as async_object_session
+from .session import async_session as async_session
+from .session import async_sessionmaker as async_sessionmaker
+from .session import AsyncAttrs as AsyncAttrs
+from .session import AsyncSession as AsyncSession
+from .session import AsyncSessionTransaction as AsyncSessionTransaction
+from .session import close_all_sessions as close_all_sessions
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..096a9c0e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..8c4de161
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/engine.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/engine.cpython-312.pyc
new file mode 100644
index 00000000..8fc0354e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/engine.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/exc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/exc.cpython-312.pyc
new file mode 100644
index 00000000..57ad2e7a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/exc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/result.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/result.cpython-312.pyc
new file mode 100644
index 00000000..42848ab8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/result.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/scoping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/scoping.cpython-312.pyc
new file mode 100644
index 00000000..77381a28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/scoping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/session.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/session.cpython-312.pyc
new file mode 100644
index 00000000..c9754e07
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/__pycache__/session.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/base.py
new file mode 100644
index 00000000..b53d53b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/base.py
@@ -0,0 +1,281 @@
+# ext/asyncio/base.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import abc
+import functools
+from typing import Any
+from typing import AsyncGenerator
+from typing import AsyncIterator
+from typing import Awaitable
+from typing import Callable
+from typing import ClassVar
+from typing import Dict
+from typing import Generator
+from typing import Generic
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import TypeVar
+import weakref
+
+from . import exc as async_exc
+from ... import util
+from ...util.typing import Literal
+from ...util.typing import Self
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+
+_PT = TypeVar("_PT", bound=Any)
+
+
+class ReversibleProxy(Generic[_PT]):
+    _proxy_objects: ClassVar[
+        Dict[weakref.ref[Any], weakref.ref[ReversibleProxy[Any]]]
+    ] = {}
+    __slots__ = ("__weakref__",)
+
+    @overload
+    def _assign_proxied(self, target: _PT) -> _PT: ...
+
+    @overload
+    def _assign_proxied(self, target: None) -> None: ...
+
+    def _assign_proxied(self, target: Optional[_PT]) -> Optional[_PT]:
+        if target is not None:
+            target_ref: weakref.ref[_PT] = weakref.ref(
+                target, ReversibleProxy._target_gced
+            )
+            proxy_ref = weakref.ref(
+                self,
+                functools.partial(ReversibleProxy._target_gced, target_ref),
+            )
+            ReversibleProxy._proxy_objects[target_ref] = proxy_ref
+
+        return target
+
+    @classmethod
+    def _target_gced(
+        cls,
+        ref: weakref.ref[_PT],
+        proxy_ref: Optional[weakref.ref[Self]] = None,  # noqa: U100
+    ) -> None:
+        cls._proxy_objects.pop(ref, None)
+
+    @classmethod
+    def _regenerate_proxy_for_target(cls, target: _PT) -> Self:
+        raise NotImplementedError()
+
+    @overload
+    @classmethod
+    def _retrieve_proxy_for_target(
+        cls,
+        target: _PT,
+        regenerate: Literal[True] = ...,
+    ) -> Self: ...
+
+    @overload
+    @classmethod
+    def _retrieve_proxy_for_target(
+        cls, target: _PT, regenerate: bool = True
+    ) -> Optional[Self]: ...
+
+    @classmethod
+    def _retrieve_proxy_for_target(
+        cls, target: _PT, regenerate: bool = True
+    ) -> Optional[Self]:
+        try:
+            proxy_ref = cls._proxy_objects[weakref.ref(target)]
+        except KeyError:
+            pass
+        else:
+            proxy = proxy_ref()
+            if proxy is not None:
+                return proxy  # type: ignore
+
+        if regenerate:
+            return cls._regenerate_proxy_for_target(target)
+        else:
+            return None
+
+
+class StartableContext(Awaitable[_T_co], abc.ABC):
+    __slots__ = ()
+
+    @abc.abstractmethod
+    async def start(self, is_ctxmanager: bool = False) -> _T_co:
+        raise NotImplementedError()
+
+    def __await__(self) -> Generator[Any, Any, _T_co]:
+        return self.start().__await__()
+
+    async def __aenter__(self) -> _T_co:
+        return await self.start(is_ctxmanager=True)
+
+    @abc.abstractmethod
+    async def __aexit__(
+        self, type_: Any, value: Any, traceback: Any
+    ) -> Optional[bool]:
+        pass
+
+    def _raise_for_not_started(self) -> NoReturn:
+        raise async_exc.AsyncContextNotStarted(
+            "%s context has not been started and object has not been awaited."
+            % (self.__class__.__name__)
+        )
+
+
+class GeneratorStartableContext(StartableContext[_T_co]):
+    __slots__ = ("gen",)
+
+    gen: AsyncGenerator[_T_co, Any]
+
+    def __init__(
+        self,
+        func: Callable[..., AsyncIterator[_T_co]],
+        args: Tuple[Any, ...],
+        kwds: Dict[str, Any],
+    ):
+        self.gen = func(*args, **kwds)  # type: ignore
+
+    async def start(self, is_ctxmanager: bool = False) -> _T_co:
+        try:
+            start_value = await util.anext_(self.gen)
+        except StopAsyncIteration:
+            raise RuntimeError("generator didn't yield") from None
+
+        # if not a context manager, then interrupt the generator, don't
+        # let it complete.   this step is technically not needed, as the
+        # generator will close in any case at gc time.  not clear if having
+        # this here is a good idea or not (though it helps for clarity IMO)
+        if not is_ctxmanager:
+            await self.gen.aclose()
+
+        return start_value
+
+    async def __aexit__(
+        self, typ: Any, value: Any, traceback: Any
+    ) -> Optional[bool]:
+        # vendored from contextlib.py
+        if typ is None:
+            try:
+                await util.anext_(self.gen)
+            except StopAsyncIteration:
+                return False
+            else:
+                raise RuntimeError("generator didn't stop")
+        else:
+            if value is None:
+                # Need to force instantiation so we can reliably
+                # tell if we get the same exception back
+                value = typ()
+            try:
+                await self.gen.athrow(value)
+            except StopAsyncIteration as exc:
+                # Suppress StopIteration *unless* it's the same exception that
+                # was passed to throw().  This prevents a StopIteration
+                # raised inside the "with" statement from being suppressed.
+                return exc is not value
+            except RuntimeError as exc:
+                # Don't re-raise the passed in exception. (issue27122)
+                if exc is value:
+                    return False
+                # Avoid suppressing if a Stop(Async)Iteration exception
+                # was passed to athrow() and later wrapped into a RuntimeError
+                # (see PEP 479 for sync generators; async generators also
+                # have this behavior). But do this only if the exception
+                # wrapped
+                # by the RuntimeError is actully Stop(Async)Iteration (see
+                # issue29692).
+                if (
+                    isinstance(value, (StopIteration, StopAsyncIteration))
+                    and exc.__cause__ is value
+                ):
+                    return False
+                raise
+            except BaseException as exc:
+                # only re-raise if it's *not* the exception that was
+                # passed to throw(), because __exit__() must not raise
+                # an exception unless __exit__() itself failed.  But throw()
+                # has to raise the exception to signal propagation, so this
+                # fixes the impedance mismatch between the throw() protocol
+                # and the __exit__() protocol.
+                if exc is not value:
+                    raise
+                return False
+            raise RuntimeError("generator didn't stop after athrow()")
+
+
+def asyncstartablecontext(
+    func: Callable[..., AsyncIterator[_T_co]]
+) -> Callable[..., GeneratorStartableContext[_T_co]]:
+    """@asyncstartablecontext decorator.
+
+    the decorated function can be called either as ``async with fn()``, **or**
+    ``await fn()``.   This is decidedly different from what
+    ``@contextlib.asynccontextmanager`` supports, and the usage pattern
+    is different as well.
+
+    Typical usage:
+
+    .. sourcecode:: text
+
+        @asyncstartablecontext
+        async def some_async_generator(<arguments>):
+            <setup>
+            try:
+                yield <value>
+            except GeneratorExit:
+                # return value was awaited, no context manager is present
+                # and caller will .close() the resource explicitly
+                pass
+            else:
+                <context manager cleanup>
+
+
+    Above, ``GeneratorExit`` is caught if the function were used as an
+    ``await``.  In this case, it's essential that the cleanup does **not**
+    occur, so there should not be a ``finally`` block.
+
+    If ``GeneratorExit`` is not invoked, this means we're in ``__aexit__``
+    and we were invoked as a context manager, and cleanup should proceed.
+
+
+    """
+
+    @functools.wraps(func)
+    def helper(*args: Any, **kwds: Any) -> GeneratorStartableContext[_T_co]:
+        return GeneratorStartableContext(func, args, kwds)
+
+    return helper
+
+
+class ProxyComparable(ReversibleProxy[_PT]):
+    __slots__ = ()
+
+    @util.ro_non_memoized_property
+    def _proxied(self) -> _PT:
+        raise NotImplementedError()
+
+    def __hash__(self) -> int:
+        return id(self)
+
+    def __eq__(self, other: Any) -> bool:
+        return (
+            isinstance(other, self.__class__)
+            and self._proxied == other._proxied
+        )
+
+    def __ne__(self, other: Any) -> bool:
+        return (
+            not isinstance(other, self.__class__)
+            or self._proxied != other._proxied
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/engine.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/engine.py
new file mode 100644
index 00000000..2c9b499f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/engine.py
@@ -0,0 +1,1467 @@
+# ext/asyncio/engine.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import asyncio
+import contextlib
+from typing import Any
+from typing import AsyncIterator
+from typing import Callable
+from typing import Dict
+from typing import Generator
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import exc as async_exc
+from .base import asyncstartablecontext
+from .base import GeneratorStartableContext
+from .base import ProxyComparable
+from .base import StartableContext
+from .result import _ensure_sync_result
+from .result import AsyncResult
+from .result import AsyncScalarResult
+from ... import exc
+from ... import inspection
+from ... import util
+from ...engine import Connection
+from ...engine import create_engine as _create_engine
+from ...engine import create_pool_from_url as _create_pool_from_url
+from ...engine import Engine
+from ...engine.base import NestedTransaction
+from ...engine.base import Transaction
+from ...exc import ArgumentError
+from ...util.concurrency import greenlet_spawn
+from ...util.typing import Concatenate
+from ...util.typing import ParamSpec
+
+if TYPE_CHECKING:
+    from ...engine.cursor import CursorResult
+    from ...engine.interfaces import _CoreAnyExecuteParams
+    from ...engine.interfaces import _CoreSingleExecuteParams
+    from ...engine.interfaces import _DBAPIAnyExecuteParams
+    from ...engine.interfaces import _ExecuteOptions
+    from ...engine.interfaces import CompiledCacheType
+    from ...engine.interfaces import CoreExecuteOptionsParameter
+    from ...engine.interfaces import Dialect
+    from ...engine.interfaces import IsolationLevel
+    from ...engine.interfaces import SchemaTranslateMapType
+    from ...engine.result import ScalarResult
+    from ...engine.url import URL
+    from ...pool import Pool
+    from ...pool import PoolProxiedConnection
+    from ...sql._typing import _InfoType
+    from ...sql.base import Executable
+    from ...sql.selectable import TypedReturnsRows
+
+_P = ParamSpec("_P")
+_T = TypeVar("_T", bound=Any)
+
+
+def create_async_engine(url: Union[str, URL], **kw: Any) -> AsyncEngine:
+    """Create a new async engine instance.
+
+    Arguments passed to :func:`_asyncio.create_async_engine` are mostly
+    identical to those passed to the :func:`_sa.create_engine` function.
+    The specified dialect must be an asyncio-compatible dialect
+    such as :ref:`dialect-postgresql-asyncpg`.
+
+    .. versionadded:: 1.4
+
+    :param async_creator: an async callable which returns a driver-level
+        asyncio connection. If given, the function should take no arguments,
+        and return a new asyncio connection from the underlying asyncio
+        database driver; the connection will be wrapped in the appropriate
+        structures to be used with the :class:`.AsyncEngine`.   Note that the
+        parameters specified in the URL are not applied here, and the creator
+        function should use its own connection parameters.
+
+        This parameter is the asyncio equivalent of the
+        :paramref:`_sa.create_engine.creator` parameter of the
+        :func:`_sa.create_engine` function.
+
+        .. versionadded:: 2.0.16
+
+    """
+
+    if kw.get("server_side_cursors", False):
+        raise async_exc.AsyncMethodRequired(
+            "Can't set server_side_cursors for async engine globally; "
+            "use the connection.stream() method for an async "
+            "streaming result set"
+        )
+    kw["_is_async"] = True
+    async_creator = kw.pop("async_creator", None)
+    if async_creator:
+        if kw.get("creator", None):
+            raise ArgumentError(
+                "Can only specify one of 'async_creator' or 'creator', "
+                "not both."
+            )
+
+        def creator() -> Any:
+            # note that to send adapted arguments like
+            # prepared_statement_cache_size, user would use
+            # "creator" and emulate this form here
+            return sync_engine.dialect.dbapi.connect(  # type: ignore
+                async_creator_fn=async_creator
+            )
+
+        kw["creator"] = creator
+    sync_engine = _create_engine(url, **kw)
+    return AsyncEngine(sync_engine)
+
+
+def async_engine_from_config(
+    configuration: Dict[str, Any], prefix: str = "sqlalchemy.", **kwargs: Any
+) -> AsyncEngine:
+    """Create a new AsyncEngine instance using a configuration dictionary.
+
+    This function is analogous to the :func:`_sa.engine_from_config` function
+    in SQLAlchemy Core, except that the requested dialect must be an
+    asyncio-compatible dialect such as :ref:`dialect-postgresql-asyncpg`.
+    The argument signature of the function is identical to that
+    of :func:`_sa.engine_from_config`.
+
+    .. versionadded:: 1.4.29
+
+    """
+    options = {
+        key[len(prefix) :]: value
+        for key, value in configuration.items()
+        if key.startswith(prefix)
+    }
+    options["_coerce_config"] = True
+    options.update(kwargs)
+    url = options.pop("url")
+    return create_async_engine(url, **options)
+
+
+def create_async_pool_from_url(url: Union[str, URL], **kwargs: Any) -> Pool:
+    """Create a new async engine instance.
+
+    Arguments passed to :func:`_asyncio.create_async_pool_from_url` are mostly
+    identical to those passed to the :func:`_sa.create_pool_from_url` function.
+    The specified dialect must be an asyncio-compatible dialect
+    such as :ref:`dialect-postgresql-asyncpg`.
+
+    .. versionadded:: 2.0.10
+
+    """
+    kwargs["_is_async"] = True
+    return _create_pool_from_url(url, **kwargs)
+
+
+class AsyncConnectable:
+    __slots__ = "_slots_dispatch", "__weakref__"
+
+    @classmethod
+    def _no_async_engine_events(cls) -> NoReturn:
+        raise NotImplementedError(
+            "asynchronous events are not implemented at this time.  Apply "
+            "synchronous listeners to the AsyncEngine.sync_engine or "
+            "AsyncConnection.sync_connection attributes."
+        )
+
+
+@util.create_proxy_methods(
+    Connection,
+    ":class:`_engine.Connection`",
+    ":class:`_asyncio.AsyncConnection`",
+    classmethods=[],
+    methods=[],
+    attributes=[
+        "closed",
+        "invalidated",
+        "dialect",
+        "default_isolation_level",
+    ],
+)
+class AsyncConnection(
+    ProxyComparable[Connection],
+    StartableContext["AsyncConnection"],
+    AsyncConnectable,
+):
+    """An asyncio proxy for a :class:`_engine.Connection`.
+
+    :class:`_asyncio.AsyncConnection` is acquired using the
+    :meth:`_asyncio.AsyncEngine.connect`
+    method of :class:`_asyncio.AsyncEngine`::
+
+        from sqlalchemy.ext.asyncio import create_async_engine
+
+        engine = create_async_engine("postgresql+asyncpg://user:pass@host/dbname")
+
+        async with engine.connect() as conn:
+            result = await conn.execute(select(table))
+
+    .. versionadded:: 1.4
+
+    """  # noqa
+
+    # AsyncConnection is a thin proxy; no state should be added here
+    # that is not retrievable from the "sync" engine / connection, e.g.
+    # current transaction, info, etc.   It should be possible to
+    # create a new AsyncConnection that matches this one given only the
+    # "sync" elements.
+    __slots__ = (
+        "engine",
+        "sync_engine",
+        "sync_connection",
+    )
+
+    def __init__(
+        self,
+        async_engine: AsyncEngine,
+        sync_connection: Optional[Connection] = None,
+    ):
+        self.engine = async_engine
+        self.sync_engine = async_engine.sync_engine
+        self.sync_connection = self._assign_proxied(sync_connection)
+
+    sync_connection: Optional[Connection]
+    """Reference to the sync-style :class:`_engine.Connection` this
+    :class:`_asyncio.AsyncConnection` proxies requests towards.
+
+    This instance can be used as an event target.
+
+    .. seealso::
+
+        :ref:`asyncio_events`
+
+    """
+
+    sync_engine: Engine
+    """Reference to the sync-style :class:`_engine.Engine` this
+    :class:`_asyncio.AsyncConnection` is associated with via its underlying
+    :class:`_engine.Connection`.
+
+    This instance can be used as an event target.
+
+    .. seealso::
+
+        :ref:`asyncio_events`
+
+    """
+
+    @classmethod
+    def _regenerate_proxy_for_target(
+        cls, target: Connection
+    ) -> AsyncConnection:
+        return AsyncConnection(
+            AsyncEngine._retrieve_proxy_for_target(target.engine), target
+        )
+
+    async def start(
+        self, is_ctxmanager: bool = False  # noqa: U100
+    ) -> AsyncConnection:
+        """Start this :class:`_asyncio.AsyncConnection` object's context
+        outside of using a Python ``with:`` block.
+
+        """
+        if self.sync_connection:
+            raise exc.InvalidRequestError("connection is already started")
+        self.sync_connection = self._assign_proxied(
+            await greenlet_spawn(self.sync_engine.connect)
+        )
+        return self
+
+    @property
+    def connection(self) -> NoReturn:
+        """Not implemented for async; call
+        :meth:`_asyncio.AsyncConnection.get_raw_connection`.
+        """
+        raise exc.InvalidRequestError(
+            "AsyncConnection.connection accessor is not implemented as the "
+            "attribute may need to reconnect on an invalidated connection.  "
+            "Use the get_raw_connection() method."
+        )
+
+    async def get_raw_connection(self) -> PoolProxiedConnection:
+        """Return the pooled DBAPI-level connection in use by this
+        :class:`_asyncio.AsyncConnection`.
+
+        This is a SQLAlchemy connection-pool proxied connection
+        which then has the attribute
+        :attr:`_pool._ConnectionFairy.driver_connection` that refers to the
+        actual driver connection. Its
+        :attr:`_pool._ConnectionFairy.dbapi_connection` refers instead
+        to an :class:`_engine.AdaptedConnection` instance that
+        adapts the driver connection to the DBAPI protocol.
+
+        """
+
+        return await greenlet_spawn(getattr, self._proxied, "connection")
+
+    @util.ro_non_memoized_property
+    def info(self) -> _InfoType:
+        """Return the :attr:`_engine.Connection.info` dictionary of the
+        underlying :class:`_engine.Connection`.
+
+        This dictionary is freely writable for user-defined state to be
+        associated with the database connection.
+
+        This attribute is only available if the :class:`.AsyncConnection` is
+        currently connected.   If the :attr:`.AsyncConnection.closed` attribute
+        is ``True``, then accessing this attribute will raise
+        :class:`.ResourceClosedError`.
+
+        .. versionadded:: 1.4.0b2
+
+        """
+        return self._proxied.info
+
+    @util.ro_non_memoized_property
+    def _proxied(self) -> Connection:
+        if not self.sync_connection:
+            self._raise_for_not_started()
+        return self.sync_connection
+
+    def begin(self) -> AsyncTransaction:
+        """Begin a transaction prior to autobegin occurring."""
+        assert self._proxied
+        return AsyncTransaction(self)
+
+    def begin_nested(self) -> AsyncTransaction:
+        """Begin a nested transaction and return a transaction handle."""
+        assert self._proxied
+        return AsyncTransaction(self, nested=True)
+
+    async def invalidate(
+        self, exception: Optional[BaseException] = None
+    ) -> None:
+        """Invalidate the underlying DBAPI connection associated with
+        this :class:`_engine.Connection`.
+
+        See the method :meth:`_engine.Connection.invalidate` for full
+        detail on this method.
+
+        """
+
+        return await greenlet_spawn(
+            self._proxied.invalidate, exception=exception
+        )
+
+    async def get_isolation_level(self) -> IsolationLevel:
+        return await greenlet_spawn(self._proxied.get_isolation_level)
+
+    def in_transaction(self) -> bool:
+        """Return True if a transaction is in progress."""
+
+        return self._proxied.in_transaction()
+
+    def in_nested_transaction(self) -> bool:
+        """Return True if a transaction is in progress.
+
+        .. versionadded:: 1.4.0b2
+
+        """
+        return self._proxied.in_nested_transaction()
+
+    def get_transaction(self) -> Optional[AsyncTransaction]:
+        """Return an :class:`.AsyncTransaction` representing the current
+        transaction, if any.
+
+        This makes use of the underlying synchronous connection's
+        :meth:`_engine.Connection.get_transaction` method to get the current
+        :class:`_engine.Transaction`, which is then proxied in a new
+        :class:`.AsyncTransaction` object.
+
+        .. versionadded:: 1.4.0b2
+
+        """
+
+        trans = self._proxied.get_transaction()
+        if trans is not None:
+            return AsyncTransaction._retrieve_proxy_for_target(trans)
+        else:
+            return None
+
+    def get_nested_transaction(self) -> Optional[AsyncTransaction]:
+        """Return an :class:`.AsyncTransaction` representing the current
+        nested (savepoint) transaction, if any.
+
+        This makes use of the underlying synchronous connection's
+        :meth:`_engine.Connection.get_nested_transaction` method to get the
+        current :class:`_engine.Transaction`, which is then proxied in a new
+        :class:`.AsyncTransaction` object.
+
+        .. versionadded:: 1.4.0b2
+
+        """
+
+        trans = self._proxied.get_nested_transaction()
+        if trans is not None:
+            return AsyncTransaction._retrieve_proxy_for_target(trans)
+        else:
+            return None
+
+    @overload
+    async def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        no_parameters: bool = False,
+        stream_results: bool = False,
+        max_row_buffer: int = ...,
+        yield_per: int = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        preserve_rowcount: bool = False,
+        **opt: Any,
+    ) -> AsyncConnection: ...
+
+    @overload
+    async def execution_options(self, **opt: Any) -> AsyncConnection: ...
+
+    async def execution_options(self, **opt: Any) -> AsyncConnection:
+        r"""Set non-SQL options for the connection which take effect
+        during execution.
+
+        This returns this :class:`_asyncio.AsyncConnection` object with
+        the new options added.
+
+        See :meth:`_engine.Connection.execution_options` for full details
+        on this method.
+
+        """
+
+        conn = self._proxied
+        c2 = await greenlet_spawn(conn.execution_options, **opt)
+        assert c2 is conn
+        return self
+
+    async def commit(self) -> None:
+        """Commit the transaction that is currently in progress.
+
+        This method commits the current transaction if one has been started.
+        If no transaction was started, the method has no effect, assuming
+        the connection is in a non-invalidated state.
+
+        A transaction is begun on a :class:`_engine.Connection` automatically
+        whenever a statement is first executed, or when the
+        :meth:`_engine.Connection.begin` method is called.
+
+        """
+        await greenlet_spawn(self._proxied.commit)
+
+    async def rollback(self) -> None:
+        """Roll back the transaction that is currently in progress.
+
+        This method rolls back the current transaction if one has been started.
+        If no transaction was started, the method has no effect.  If a
+        transaction was started and the connection is in an invalidated state,
+        the transaction is cleared using this method.
+
+        A transaction is begun on a :class:`_engine.Connection` automatically
+        whenever a statement is first executed, or when the
+        :meth:`_engine.Connection.begin` method is called.
+
+
+        """
+        await greenlet_spawn(self._proxied.rollback)
+
+    async def close(self) -> None:
+        """Close this :class:`_asyncio.AsyncConnection`.
+
+        This has the effect of also rolling back the transaction if one
+        is in place.
+
+        """
+        await greenlet_spawn(self._proxied.close)
+
+    async def aclose(self) -> None:
+        """A synonym for :meth:`_asyncio.AsyncConnection.close`.
+
+        The :meth:`_asyncio.AsyncConnection.aclose` name is specifically
+        to support the Python standard library ``@contextlib.aclosing``
+        context manager function.
+
+        .. versionadded:: 2.0.20
+
+        """
+        await self.close()
+
+    async def exec_driver_sql(
+        self,
+        statement: str,
+        parameters: Optional[_DBAPIAnyExecuteParams] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]:
+        r"""Executes a driver-level SQL string and return buffered
+        :class:`_engine.Result`.
+
+        """
+
+        result = await greenlet_spawn(
+            self._proxied.exec_driver_sql,
+            statement,
+            parameters,
+            execution_options,
+            _require_await=True,
+        )
+
+        return await _ensure_sync_result(result, self.exec_driver_sql)
+
+    @overload
+    def stream(
+        self,
+        statement: TypedReturnsRows[_T],
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> GeneratorStartableContext[AsyncResult[_T]]: ...
+
+    @overload
+    def stream(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> GeneratorStartableContext[AsyncResult[Any]]: ...
+
+    @asyncstartablecontext
+    async def stream(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> AsyncIterator[AsyncResult[Any]]:
+        """Execute a statement and return an awaitable yielding a
+        :class:`_asyncio.AsyncResult` object.
+
+        E.g.::
+
+            result = await conn.stream(stmt)
+            async for row in result:
+                print(f"{row}")
+
+        The :meth:`.AsyncConnection.stream`
+        method supports optional context manager use against the
+        :class:`.AsyncResult` object, as in::
+
+            async with conn.stream(stmt) as result:
+                async for row in result:
+                    print(f"{row}")
+
+        In the above pattern, the :meth:`.AsyncResult.close` method is
+        invoked unconditionally, even if the iterator is interrupted by an
+        exception throw.   Context manager use remains optional, however,
+        and the function may be called in either an ``async with fn():`` or
+        ``await fn()`` style.
+
+        .. versionadded:: 2.0.0b3 added context manager support
+
+
+        :return: an awaitable object that will yield an
+         :class:`_asyncio.AsyncResult` object.
+
+        .. seealso::
+
+            :meth:`.AsyncConnection.stream_scalars`
+
+        """
+        if not self.dialect.supports_server_side_cursors:
+            raise exc.InvalidRequestError(
+                "Cant use `stream` or `stream_scalars` with the current "
+                "dialect since it does not support server side cursors."
+            )
+
+        result = await greenlet_spawn(
+            self._proxied.execute,
+            statement,
+            parameters,
+            execution_options=util.EMPTY_DICT.merge_with(
+                execution_options, {"stream_results": True}
+            ),
+            _require_await=True,
+        )
+        assert result.context._is_server_side
+        ar = AsyncResult(result)
+        try:
+            yield ar
+        except GeneratorExit:
+            pass
+        else:
+            task = asyncio.create_task(ar.close())
+            await asyncio.shield(task)
+
+    @overload
+    async def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[_T]: ...
+
+    @overload
+    async def execute(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]: ...
+
+    async def execute(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> CursorResult[Any]:
+        r"""Executes a SQL statement construct and return a buffered
+        :class:`_engine.Result`.
+
+        :param object: The statement to be executed.  This is always
+         an object that is in both the :class:`_expression.ClauseElement` and
+         :class:`_expression.Executable` hierarchies, including:
+
+         * :class:`_expression.Select`
+         * :class:`_expression.Insert`, :class:`_expression.Update`,
+           :class:`_expression.Delete`
+         * :class:`_expression.TextClause` and
+           :class:`_expression.TextualSelect`
+         * :class:`_schema.DDL` and objects which inherit from
+           :class:`_schema.ExecutableDDLElement`
+
+        :param parameters: parameters which will be bound into the statement.
+         This may be either a dictionary of parameter names to values,
+         or a mutable sequence (e.g. a list) of dictionaries.  When a
+         list of dictionaries is passed, the underlying statement execution
+         will make use of the DBAPI ``cursor.executemany()`` method.
+         When a single dictionary is passed, the DBAPI ``cursor.execute()``
+         method will be used.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the statement execution.  This
+         dictionary can provide a subset of the options that are accepted
+         by :meth:`_engine.Connection.execution_options`.
+
+        :return: a :class:`_engine.Result` object.
+
+        """
+        result = await greenlet_spawn(
+            self._proxied.execute,
+            statement,
+            parameters,
+            execution_options=execution_options,
+            _require_await=True,
+        )
+        return await _ensure_sync_result(result, self.execute)
+
+    @overload
+    async def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Optional[_T]: ...
+
+    @overload
+    async def scalar(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Any: ...
+
+    async def scalar(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Any:
+        r"""Executes a SQL statement construct and returns a scalar object.
+
+        This method is shorthand for invoking the
+        :meth:`_engine.Result.scalar` method after invoking the
+        :meth:`_engine.Connection.execute` method.  Parameters are equivalent.
+
+        :return: a scalar Python value representing the first column of the
+         first row returned.
+
+        """
+        result = await self.execute(
+            statement, parameters, execution_options=execution_options
+        )
+        return result.scalar()
+
+    @overload
+    async def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    async def scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[Any]: ...
+
+    async def scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> ScalarResult[Any]:
+        r"""Executes a SQL statement construct and returns a scalar objects.
+
+        This method is shorthand for invoking the
+        :meth:`_engine.Result.scalars` method after invoking the
+        :meth:`_engine.Connection.execute` method.  Parameters are equivalent.
+
+        :return: a :class:`_engine.ScalarResult` object.
+
+        .. versionadded:: 1.4.24
+
+        """
+        result = await self.execute(
+            statement, parameters, execution_options=execution_options
+        )
+        return result.scalars()
+
+    @overload
+    def stream_scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> GeneratorStartableContext[AsyncScalarResult[_T]]: ...
+
+    @overload
+    def stream_scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> GeneratorStartableContext[AsyncScalarResult[Any]]: ...
+
+    @asyncstartablecontext
+    async def stream_scalars(
+        self,
+        statement: Executable,
+        parameters: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> AsyncIterator[AsyncScalarResult[Any]]:
+        r"""Execute a statement and return an awaitable yielding a
+        :class:`_asyncio.AsyncScalarResult` object.
+
+        E.g.::
+
+            result = await conn.stream_scalars(stmt)
+            async for scalar in result:
+                print(f"{scalar}")
+
+        This method is shorthand for invoking the
+        :meth:`_engine.AsyncResult.scalars` method after invoking the
+        :meth:`_engine.Connection.stream` method.  Parameters are equivalent.
+
+        The :meth:`.AsyncConnection.stream_scalars`
+        method supports optional context manager use against the
+        :class:`.AsyncScalarResult` object, as in::
+
+            async with conn.stream_scalars(stmt) as result:
+                async for scalar in result:
+                    print(f"{scalar}")
+
+        In the above pattern, the :meth:`.AsyncScalarResult.close` method is
+        invoked unconditionally, even if the iterator is interrupted by an
+        exception throw.  Context manager use remains optional, however,
+        and the function may be called in either an ``async with fn():`` or
+        ``await fn()`` style.
+
+        .. versionadded:: 2.0.0b3 added context manager support
+
+        :return: an awaitable object that will yield an
+         :class:`_asyncio.AsyncScalarResult` object.
+
+        .. versionadded:: 1.4.24
+
+        .. seealso::
+
+            :meth:`.AsyncConnection.stream`
+
+        """
+
+        async with self.stream(
+            statement, parameters, execution_options=execution_options
+        ) as result:
+            yield result.scalars()
+
+    async def run_sync(
+        self,
+        fn: Callable[Concatenate[Connection, _P], _T],
+        *arg: _P.args,
+        **kw: _P.kwargs,
+    ) -> _T:
+        '''Invoke the given synchronous (i.e. not async) callable,
+        passing a synchronous-style :class:`_engine.Connection` as the first
+        argument.
+
+        This method allows traditional synchronous SQLAlchemy functions to
+        run within the context of an asyncio application.
+
+        E.g.::
+
+            def do_something_with_core(conn: Connection, arg1: int, arg2: str) -> str:
+                """A synchronous function that does not require awaiting
+
+                :param conn: a Core SQLAlchemy Connection, used synchronously
+
+                :return: an optional return value is supported
+
+                """
+                conn.execute(some_table.insert().values(int_col=arg1, str_col=arg2))
+                return "success"
+
+
+            async def do_something_async(async_engine: AsyncEngine) -> None:
+                """an async function that uses awaiting"""
+
+                async with async_engine.begin() as async_conn:
+                    # run do_something_with_core() with a sync-style
+                    # Connection, proxied into an awaitable
+                    return_code = await async_conn.run_sync(
+                        do_something_with_core, 5, "strval"
+                    )
+                    print(return_code)
+
+        This method maintains the asyncio event loop all the way through
+        to the database connection by running the given callable in a
+        specially instrumented greenlet.
+
+        The most rudimentary use of :meth:`.AsyncConnection.run_sync` is to
+        invoke methods such as :meth:`_schema.MetaData.create_all`, given
+        an :class:`.AsyncConnection` that needs to be provided to
+        :meth:`_schema.MetaData.create_all` as a :class:`_engine.Connection`
+        object::
+
+            # run metadata.create_all(conn) with a sync-style Connection,
+            # proxied into an awaitable
+            with async_engine.begin() as conn:
+                await conn.run_sync(metadata.create_all)
+
+        .. note::
+
+            The provided callable is invoked inline within the asyncio event
+            loop, and will block on traditional IO calls.  IO within this
+            callable should only call into SQLAlchemy's asyncio database
+            APIs which will be properly adapted to the greenlet context.
+
+        .. seealso::
+
+            :meth:`.AsyncSession.run_sync`
+
+            :ref:`session_run_sync`
+
+        '''  # noqa: E501
+
+        return await greenlet_spawn(
+            fn, self._proxied, *arg, _require_await=False, **kw
+        )
+
+    def __await__(self) -> Generator[Any, None, AsyncConnection]:
+        return self.start().__await__()
+
+    async def __aexit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        task = asyncio.create_task(self.close())
+        await asyncio.shield(task)
+
+    # START PROXY METHODS AsyncConnection
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_proxy_methods.py
+
+    @property
+    def closed(self) -> Any:
+        r"""Return True if this connection is closed.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Connection` class
+            on behalf of the :class:`_asyncio.AsyncConnection` class.
+
+        """  # noqa: E501
+
+        return self._proxied.closed
+
+    @property
+    def invalidated(self) -> Any:
+        r"""Return True if this connection was invalidated.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Connection` class
+            on behalf of the :class:`_asyncio.AsyncConnection` class.
+
+        This does not indicate whether or not the connection was
+        invalidated at the pool level, however
+
+
+        """  # noqa: E501
+
+        return self._proxied.invalidated
+
+    @property
+    def dialect(self) -> Dialect:
+        r"""Proxy for the :attr:`_engine.Connection.dialect` attribute
+        on behalf of the :class:`_asyncio.AsyncConnection` class.
+
+        """  # noqa: E501
+
+        return self._proxied.dialect
+
+    @dialect.setter
+    def dialect(self, attr: Dialect) -> None:
+        self._proxied.dialect = attr
+
+    @property
+    def default_isolation_level(self) -> Any:
+        r"""The initial-connection time isolation level associated with the
+        :class:`_engine.Dialect` in use.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Connection` class
+            on behalf of the :class:`_asyncio.AsyncConnection` class.
+
+        This value is independent of the
+        :paramref:`.Connection.execution_options.isolation_level` and
+        :paramref:`.Engine.execution_options.isolation_level` execution
+        options, and is determined by the :class:`_engine.Dialect` when the
+        first connection is created, by performing a SQL query against the
+        database for the current isolation level before any additional commands
+        have been emitted.
+
+        Calling this accessor does not invoke any new SQL queries.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.get_isolation_level`
+            - view current actual isolation level
+
+            :paramref:`_sa.create_engine.isolation_level`
+            - set per :class:`_engine.Engine` isolation level
+
+            :paramref:`.Connection.execution_options.isolation_level`
+            - set per :class:`_engine.Connection` isolation level
+
+
+        """  # noqa: E501
+
+        return self._proxied.default_isolation_level
+
+    # END PROXY METHODS AsyncConnection
+
+
+@util.create_proxy_methods(
+    Engine,
+    ":class:`_engine.Engine`",
+    ":class:`_asyncio.AsyncEngine`",
+    classmethods=[],
+    methods=[
+        "clear_compiled_cache",
+        "update_execution_options",
+        "get_execution_options",
+    ],
+    attributes=["url", "pool", "dialect", "engine", "name", "driver", "echo"],
+)
+class AsyncEngine(ProxyComparable[Engine], AsyncConnectable):
+    """An asyncio proxy for a :class:`_engine.Engine`.
+
+    :class:`_asyncio.AsyncEngine` is acquired using the
+    :func:`_asyncio.create_async_engine` function::
+
+        from sqlalchemy.ext.asyncio import create_async_engine
+
+        engine = create_async_engine("postgresql+asyncpg://user:pass@host/dbname")
+
+    .. versionadded:: 1.4
+
+    """  # noqa
+
+    # AsyncEngine is a thin proxy; no state should be added here
+    # that is not retrievable from the "sync" engine / connection, e.g.
+    # current transaction, info, etc.   It should be possible to
+    # create a new AsyncEngine that matches this one given only the
+    # "sync" elements.
+    __slots__ = "sync_engine"
+
+    _connection_cls: Type[AsyncConnection] = AsyncConnection
+
+    sync_engine: Engine
+    """Reference to the sync-style :class:`_engine.Engine` this
+    :class:`_asyncio.AsyncEngine` proxies requests towards.
+
+    This instance can be used as an event target.
+
+    .. seealso::
+
+        :ref:`asyncio_events`
+    """
+
+    def __init__(self, sync_engine: Engine):
+        if not sync_engine.dialect.is_async:
+            raise exc.InvalidRequestError(
+                "The asyncio extension requires an async driver to be used. "
+                f"The loaded {sync_engine.dialect.driver!r} is not async."
+            )
+        self.sync_engine = self._assign_proxied(sync_engine)
+
+    @util.ro_non_memoized_property
+    def _proxied(self) -> Engine:
+        return self.sync_engine
+
+    @classmethod
+    def _regenerate_proxy_for_target(cls, target: Engine) -> AsyncEngine:
+        return AsyncEngine(target)
+
+    @contextlib.asynccontextmanager
+    async def begin(self) -> AsyncIterator[AsyncConnection]:
+        """Return a context manager which when entered will deliver an
+        :class:`_asyncio.AsyncConnection` with an
+        :class:`_asyncio.AsyncTransaction` established.
+
+        E.g.::
+
+            async with async_engine.begin() as conn:
+                await conn.execute(
+                    text("insert into table (x, y, z) values (1, 2, 3)")
+                )
+                await conn.execute(text("my_special_procedure(5)"))
+
+        """
+        conn = self.connect()
+
+        async with conn:
+            async with conn.begin():
+                yield conn
+
+    def connect(self) -> AsyncConnection:
+        """Return an :class:`_asyncio.AsyncConnection` object.
+
+        The :class:`_asyncio.AsyncConnection` will procure a database
+        connection from the underlying connection pool when it is entered
+        as an async context manager::
+
+            async with async_engine.connect() as conn:
+                result = await conn.execute(select(user_table))
+
+        The :class:`_asyncio.AsyncConnection` may also be started outside of a
+        context manager by invoking its :meth:`_asyncio.AsyncConnection.start`
+        method.
+
+        """
+
+        return self._connection_cls(self)
+
+    async def raw_connection(self) -> PoolProxiedConnection:
+        """Return a "raw" DBAPI connection from the connection pool.
+
+        .. seealso::
+
+            :ref:`dbapi_connections`
+
+        """
+        return await greenlet_spawn(self.sync_engine.raw_connection)
+
+    @overload
+    def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        **opt: Any,
+    ) -> AsyncEngine: ...
+
+    @overload
+    def execution_options(self, **opt: Any) -> AsyncEngine: ...
+
+    def execution_options(self, **opt: Any) -> AsyncEngine:
+        """Return a new :class:`_asyncio.AsyncEngine` that will provide
+        :class:`_asyncio.AsyncConnection` objects with the given execution
+        options.
+
+        Proxied from :meth:`_engine.Engine.execution_options`.  See that
+        method for details.
+
+        """
+
+        return AsyncEngine(self.sync_engine.execution_options(**opt))
+
+    async def dispose(self, close: bool = True) -> None:
+        """Dispose of the connection pool used by this
+        :class:`_asyncio.AsyncEngine`.
+
+        :param close: if left at its default of ``True``, has the
+         effect of fully closing all **currently checked in**
+         database connections.  Connections that are still checked out
+         will **not** be closed, however they will no longer be associated
+         with this :class:`_engine.Engine`,
+         so when they are closed individually, eventually the
+         :class:`_pool.Pool` which they are associated with will
+         be garbage collected and they will be closed out fully, if
+         not already closed on checkin.
+
+         If set to ``False``, the previous connection pool is de-referenced,
+         and otherwise not touched in any way.
+
+        .. seealso::
+
+            :meth:`_engine.Engine.dispose`
+
+        """
+
+        await greenlet_spawn(self.sync_engine.dispose, close=close)
+
+    # START PROXY METHODS AsyncEngine
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_proxy_methods.py
+
+    def clear_compiled_cache(self) -> None:
+        r"""Clear the compiled cache associated with the dialect.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class on
+            behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        This applies **only** to the built-in cache that is established
+        via the :paramref:`_engine.create_engine.query_cache_size` parameter.
+        It will not impact any dictionary caches that were passed via the
+        :paramref:`.Connection.execution_options.compiled_cache` parameter.
+
+        .. versionadded:: 1.4
+
+
+        """  # noqa: E501
+
+        return self._proxied.clear_compiled_cache()
+
+    def update_execution_options(self, **opt: Any) -> None:
+        r"""Update the default execution_options dictionary
+        of this :class:`_engine.Engine`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class on
+            behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        The given keys/values in \**opt are added to the
+        default execution options that will be used for
+        all connections.  The initial contents of this dictionary
+        can be sent via the ``execution_options`` parameter
+        to :func:`_sa.create_engine`.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.execution_options`
+
+            :meth:`_engine.Engine.execution_options`
+
+
+        """  # noqa: E501
+
+        return self._proxied.update_execution_options(**opt)
+
+    def get_execution_options(self) -> _ExecuteOptions:
+        r"""Get the non-SQL options which will take effect during execution.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class on
+            behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        .. versionadded: 1.3
+
+        .. seealso::
+
+            :meth:`_engine.Engine.execution_options`
+
+        """  # noqa: E501
+
+        return self._proxied.get_execution_options()
+
+    @property
+    def url(self) -> URL:
+        r"""Proxy for the :attr:`_engine.Engine.url` attribute
+        on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        """  # noqa: E501
+
+        return self._proxied.url
+
+    @url.setter
+    def url(self, attr: URL) -> None:
+        self._proxied.url = attr
+
+    @property
+    def pool(self) -> Pool:
+        r"""Proxy for the :attr:`_engine.Engine.pool` attribute
+        on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        """  # noqa: E501
+
+        return self._proxied.pool
+
+    @pool.setter
+    def pool(self, attr: Pool) -> None:
+        self._proxied.pool = attr
+
+    @property
+    def dialect(self) -> Dialect:
+        r"""Proxy for the :attr:`_engine.Engine.dialect` attribute
+        on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        """  # noqa: E501
+
+        return self._proxied.dialect
+
+    @dialect.setter
+    def dialect(self, attr: Dialect) -> None:
+        self._proxied.dialect = attr
+
+    @property
+    def engine(self) -> Any:
+        r"""Returns this :class:`.Engine`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class
+            on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        Used for legacy schemes that accept :class:`.Connection` /
+        :class:`.Engine` objects within the same variable.
+
+
+        """  # noqa: E501
+
+        return self._proxied.engine
+
+    @property
+    def name(self) -> Any:
+        r"""String name of the :class:`~sqlalchemy.engine.interfaces.Dialect`
+        in use by this :class:`Engine`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class
+            on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.name
+
+    @property
+    def driver(self) -> Any:
+        r"""Driver name of the :class:`~sqlalchemy.engine.interfaces.Dialect`
+        in use by this :class:`Engine`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class
+            on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.driver
+
+    @property
+    def echo(self) -> Any:
+        r"""When ``True``, enable log output for this element.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_engine.Engine` class
+            on behalf of the :class:`_asyncio.AsyncEngine` class.
+
+        This has the effect of setting the Python logging level for the namespace
+        of this element's class and object reference.  A value of boolean ``True``
+        indicates that the loglevel ``logging.INFO`` will be set for the logger,
+        whereas the string value ``debug`` will set the loglevel to
+        ``logging.DEBUG``.
+
+        """  # noqa: E501
+
+        return self._proxied.echo
+
+    @echo.setter
+    def echo(self, attr: Any) -> None:
+        self._proxied.echo = attr
+
+    # END PROXY METHODS AsyncEngine
+
+
+class AsyncTransaction(
+    ProxyComparable[Transaction], StartableContext["AsyncTransaction"]
+):
+    """An asyncio proxy for a :class:`_engine.Transaction`."""
+
+    __slots__ = ("connection", "sync_transaction", "nested")
+
+    sync_transaction: Optional[Transaction]
+    connection: AsyncConnection
+    nested: bool
+
+    def __init__(self, connection: AsyncConnection, nested: bool = False):
+        self.connection = connection
+        self.sync_transaction = None
+        self.nested = nested
+
+    @classmethod
+    def _regenerate_proxy_for_target(
+        cls, target: Transaction
+    ) -> AsyncTransaction:
+        sync_connection = target.connection
+        sync_transaction = target
+        nested = isinstance(target, NestedTransaction)
+
+        async_connection = AsyncConnection._retrieve_proxy_for_target(
+            sync_connection
+        )
+        assert async_connection is not None
+
+        obj = cls.__new__(cls)
+        obj.connection = async_connection
+        obj.sync_transaction = obj._assign_proxied(sync_transaction)
+        obj.nested = nested
+        return obj
+
+    @util.ro_non_memoized_property
+    def _proxied(self) -> Transaction:
+        if not self.sync_transaction:
+            self._raise_for_not_started()
+        return self.sync_transaction
+
+    @property
+    def is_valid(self) -> bool:
+        return self._proxied.is_valid
+
+    @property
+    def is_active(self) -> bool:
+        return self._proxied.is_active
+
+    async def close(self) -> None:
+        """Close this :class:`.AsyncTransaction`.
+
+        If this transaction is the base transaction in a begin/commit
+        nesting, the transaction will rollback().  Otherwise, the
+        method returns.
+
+        This is used to cancel a Transaction without affecting the scope of
+        an enclosing transaction.
+
+        """
+        await greenlet_spawn(self._proxied.close)
+
+    async def rollback(self) -> None:
+        """Roll back this :class:`.AsyncTransaction`."""
+        await greenlet_spawn(self._proxied.rollback)
+
+    async def commit(self) -> None:
+        """Commit this :class:`.AsyncTransaction`."""
+
+        await greenlet_spawn(self._proxied.commit)
+
+    async def start(self, is_ctxmanager: bool = False) -> AsyncTransaction:
+        """Start this :class:`_asyncio.AsyncTransaction` object's context
+        outside of using a Python ``with:`` block.
+
+        """
+
+        self.sync_transaction = self._assign_proxied(
+            await greenlet_spawn(
+                self.connection._proxied.begin_nested
+                if self.nested
+                else self.connection._proxied.begin
+            )
+        )
+        if is_ctxmanager:
+            self.sync_transaction.__enter__()
+        return self
+
+    async def __aexit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        await greenlet_spawn(self._proxied.__exit__, type_, value, traceback)
+
+
+@overload
+def _get_sync_engine_or_connection(async_engine: AsyncEngine) -> Engine: ...
+
+
+@overload
+def _get_sync_engine_or_connection(
+    async_engine: AsyncConnection,
+) -> Connection: ...
+
+
+def _get_sync_engine_or_connection(
+    async_engine: Union[AsyncEngine, AsyncConnection]
+) -> Union[Engine, Connection]:
+    if isinstance(async_engine, AsyncConnection):
+        return async_engine._proxied
+
+    try:
+        return async_engine.sync_engine
+    except AttributeError as e:
+        raise exc.ArgumentError(
+            "AsyncEngine expected, got %r" % async_engine
+        ) from e
+
+
+@inspection._inspects(AsyncConnection)
+def _no_insp_for_async_conn_yet(
+    subject: AsyncConnection,  # noqa: U100
+) -> NoReturn:
+    raise exc.NoInspectionAvailable(
+        "Inspection on an AsyncConnection is currently not supported. "
+        "Please use ``run_sync`` to pass a callable where it's possible "
+        "to call ``inspect`` on the passed connection.",
+        code="xd3s",
+    )
+
+
+@inspection._inspects(AsyncEngine)
+def _no_insp_for_async_engine_xyet(
+    subject: AsyncEngine,  # noqa: U100
+) -> NoReturn:
+    raise exc.NoInspectionAvailable(
+        "Inspection on an AsyncEngine is currently not supported. "
+        "Please obtain a connection then use ``conn.run_sync`` to pass a "
+        "callable where it's possible to call ``inspect`` on the "
+        "passed connection.",
+        code="xd3s",
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/exc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/exc.py
new file mode 100644
index 00000000..558187c0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/exc.py
@@ -0,0 +1,21 @@
+# ext/asyncio/exc.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from ... import exc
+
+
+class AsyncMethodRequired(exc.InvalidRequestError):
+    """an API can't be used because its result would not be
+    compatible with async"""
+
+
+class AsyncContextNotStarted(exc.InvalidRequestError):
+    """a startable context manager has not been started."""
+
+
+class AsyncContextAlreadyStarted(exc.InvalidRequestError):
+    """a startable context manager is already started."""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/result.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/result.py
new file mode 100644
index 00000000..c51e166d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/result.py
@@ -0,0 +1,961 @@
+# ext/asyncio/result.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import operator
+from typing import Any
+from typing import AsyncIterator
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from . import exc as async_exc
+from ... import util
+from ...engine import Result
+from ...engine.result import _NO_ROW
+from ...engine.result import _R
+from ...engine.result import _WithKeys
+from ...engine.result import FilterResult
+from ...engine.result import FrozenResult
+from ...engine.result import ResultMetaData
+from ...engine.row import Row
+from ...engine.row import RowMapping
+from ...sql.base import _generative
+from ...util.concurrency import greenlet_spawn
+from ...util.typing import Literal
+from ...util.typing import Self
+
+if TYPE_CHECKING:
+    from ...engine import CursorResult
+    from ...engine.result import _KeyIndexType
+    from ...engine.result import _UniqueFilterType
+
+_T = TypeVar("_T", bound=Any)
+_TP = TypeVar("_TP", bound=Tuple[Any, ...])
+
+
+class AsyncCommon(FilterResult[_R]):
+    __slots__ = ()
+
+    _real_result: Result[Any]
+    _metadata: ResultMetaData
+
+    async def close(self) -> None:  # type: ignore[override]
+        """Close this result."""
+
+        await greenlet_spawn(self._real_result.close)
+
+    @property
+    def closed(self) -> bool:
+        """proxies the .closed attribute of the underlying result object,
+        if any, else raises ``AttributeError``.
+
+        .. versionadded:: 2.0.0b3
+
+        """
+        return self._real_result.closed
+
+
+class AsyncResult(_WithKeys, AsyncCommon[Row[_TP]]):
+    """An asyncio wrapper around a :class:`_result.Result` object.
+
+    The :class:`_asyncio.AsyncResult` only applies to statement executions that
+    use a server-side cursor.  It is returned only from the
+    :meth:`_asyncio.AsyncConnection.stream` and
+    :meth:`_asyncio.AsyncSession.stream` methods.
+
+    .. note:: As is the case with :class:`_engine.Result`, this object is
+       used for ORM results returned by :meth:`_asyncio.AsyncSession.execute`,
+       which can yield instances of ORM mapped objects either individually or
+       within tuple-like rows.  Note that these result objects do not
+       deduplicate instances or rows automatically as is the case with the
+       legacy :class:`_orm.Query` object. For in-Python de-duplication of
+       instances or rows, use the :meth:`_asyncio.AsyncResult.unique` modifier
+       method.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    _real_result: Result[_TP]
+
+    def __init__(self, real_result: Result[_TP]):
+        self._real_result = real_result
+
+        self._metadata = real_result._metadata
+        self._unique_filter_state = real_result._unique_filter_state
+        self._post_creational_filter = None
+
+        # BaseCursorResult pre-generates the "_row_getter".  Use that
+        # if available rather than building a second one
+        if "_row_getter" in real_result.__dict__:
+            self._set_memoized_attribute(
+                "_row_getter", real_result.__dict__["_row_getter"]
+            )
+
+    @property
+    def t(self) -> AsyncTupleResult[_TP]:
+        """Apply a "typed tuple" typing filter to returned rows.
+
+        The :attr:`_asyncio.AsyncResult.t` attribute is a synonym for
+        calling the :meth:`_asyncio.AsyncResult.tuples` method.
+
+        .. versionadded:: 2.0
+
+        """
+        return self  # type: ignore
+
+    def tuples(self) -> AsyncTupleResult[_TP]:
+        """Apply a "typed tuple" typing filter to returned rows.
+
+        This method returns the same :class:`_asyncio.AsyncResult` object
+        at runtime,
+        however annotates as returning a :class:`_asyncio.AsyncTupleResult`
+        object that will indicate to :pep:`484` typing tools that plain typed
+        ``Tuple`` instances are returned rather than rows.  This allows
+        tuple unpacking and ``__getitem__`` access of :class:`_engine.Row`
+        objects to by typed, for those cases where the statement invoked
+        itself included typing information.
+
+        .. versionadded:: 2.0
+
+        :return: the :class:`_result.AsyncTupleResult` type at typing time.
+
+        .. seealso::
+
+            :attr:`_asyncio.AsyncResult.t` - shorter synonym
+
+            :attr:`_engine.Row.t` - :class:`_engine.Row` version
+
+        """
+
+        return self  # type: ignore
+
+    @_generative
+    def unique(self, strategy: Optional[_UniqueFilterType] = None) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_asyncio.AsyncResult`.
+
+        Refer to :meth:`_engine.Result.unique` in the synchronous
+        SQLAlchemy API for a complete behavioral description.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    def columns(self, *col_expressions: _KeyIndexType) -> Self:
+        r"""Establish the columns that should be returned in each row.
+
+        Refer to :meth:`_engine.Result.columns` in the synchronous
+        SQLAlchemy API for a complete behavioral description.
+
+        """
+        return self._column_slices(col_expressions)
+
+    async def partitions(
+        self, size: Optional[int] = None
+    ) -> AsyncIterator[Sequence[Row[_TP]]]:
+        """Iterate through sub-lists of rows of the size given.
+
+        An async iterator is returned::
+
+            async def scroll_results(connection):
+                result = await connection.stream(select(users_table))
+
+                async for partition in result.partitions(100):
+                    print("list of rows: %s" % partition)
+
+        Refer to :meth:`_engine.Result.partitions` in the synchronous
+        SQLAlchemy API for a complete behavioral description.
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = await greenlet_spawn(getter, self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    async def fetchall(self) -> Sequence[Row[_TP]]:
+        """A synonym for the :meth:`_asyncio.AsyncResult.all` method.
+
+        .. versionadded:: 2.0
+
+        """
+
+        return await greenlet_spawn(self._allrows)
+
+    async def fetchone(self) -> Optional[Row[_TP]]:
+        """Fetch one row.
+
+        When all rows are exhausted, returns None.
+
+        This method is provided for backwards compatibility with
+        SQLAlchemy 1.x.x.
+
+        To fetch the first row of a result only, use the
+        :meth:`_asyncio.AsyncResult.first` method.  To iterate through all
+        rows, iterate the :class:`_asyncio.AsyncResult` object directly.
+
+        :return: a :class:`_engine.Row` object if no filters are applied,
+         or ``None`` if no rows remain.
+
+        """
+        row = await greenlet_spawn(self._onerow_getter, self)
+        if row is _NO_ROW:
+            return None
+        else:
+            return row
+
+    async def fetchmany(
+        self, size: Optional[int] = None
+    ) -> Sequence[Row[_TP]]:
+        """Fetch many rows.
+
+        When all rows are exhausted, returns an empty list.
+
+        This method is provided for backwards compatibility with
+        SQLAlchemy 1.x.x.
+
+        To fetch rows in groups, use the
+        :meth:`._asyncio.AsyncResult.partitions` method.
+
+        :return: a list of :class:`_engine.Row` objects.
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncResult.partitions`
+
+        """
+
+        return await greenlet_spawn(self._manyrow_getter, self, size)
+
+    async def all(self) -> Sequence[Row[_TP]]:
+        """Return all rows in a list.
+
+        Closes the result set after invocation.   Subsequent invocations
+        will return an empty list.
+
+        :return: a list of :class:`_engine.Row` objects.
+
+        """
+
+        return await greenlet_spawn(self._allrows)
+
+    def __aiter__(self) -> AsyncResult[_TP]:
+        return self
+
+    async def __anext__(self) -> Row[_TP]:
+        row = await greenlet_spawn(self._onerow_getter, self)
+        if row is _NO_ROW:
+            raise StopAsyncIteration()
+        else:
+            return row
+
+    async def first(self) -> Optional[Row[_TP]]:
+        """Fetch the first row or ``None`` if no row is present.
+
+        Closes the result set and discards remaining rows.
+
+        .. note::  This method returns one **row**, e.g. tuple, by default.
+           To return exactly one single scalar value, that is, the first
+           column of the first row, use the
+           :meth:`_asyncio.AsyncResult.scalar` method,
+           or combine :meth:`_asyncio.AsyncResult.scalars` and
+           :meth:`_asyncio.AsyncResult.first`.
+
+           Additionally, in contrast to the behavior of the legacy  ORM
+           :meth:`_orm.Query.first` method, **no limit is applied** to the
+           SQL query which was invoked to produce this
+           :class:`_asyncio.AsyncResult`;
+           for a DBAPI driver that buffers results in memory before yielding
+           rows, all rows will be sent to the Python process and all but
+           the first row will be discarded.
+
+           .. seealso::
+
+                :ref:`migration_20_unify_select`
+
+        :return: a :class:`_engine.Row` object, or None
+         if no rows remain.
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncResult.scalar`
+
+            :meth:`_asyncio.AsyncResult.one`
+
+        """
+        return await greenlet_spawn(self._only_one_row, False, False, False)
+
+    async def one_or_none(self) -> Optional[Row[_TP]]:
+        """Return at most one result or raise an exception.
+
+        Returns ``None`` if the result has no rows.
+        Raises :class:`.MultipleResultsFound`
+        if multiple rows are returned.
+
+        .. versionadded:: 1.4
+
+        :return: The first :class:`_engine.Row` or ``None`` if no row
+         is available.
+
+        :raises: :class:`.MultipleResultsFound`
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncResult.first`
+
+            :meth:`_asyncio.AsyncResult.one`
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, False, False)
+
+    @overload
+    async def scalar_one(self: AsyncResult[Tuple[_T]]) -> _T: ...
+
+    @overload
+    async def scalar_one(self) -> Any: ...
+
+    async def scalar_one(self) -> Any:
+        """Return exactly one scalar result or raise an exception.
+
+        This is equivalent to calling :meth:`_asyncio.AsyncResult.scalars` and
+        then :meth:`_asyncio.AsyncScalarResult.one`.
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncScalarResult.one`
+
+            :meth:`_asyncio.AsyncResult.scalars`
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, True, True)
+
+    @overload
+    async def scalar_one_or_none(
+        self: AsyncResult[Tuple[_T]],
+    ) -> Optional[_T]: ...
+
+    @overload
+    async def scalar_one_or_none(self) -> Optional[Any]: ...
+
+    async def scalar_one_or_none(self) -> Optional[Any]:
+        """Return exactly one scalar result or ``None``.
+
+        This is equivalent to calling :meth:`_asyncio.AsyncResult.scalars` and
+        then :meth:`_asyncio.AsyncScalarResult.one_or_none`.
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncScalarResult.one_or_none`
+
+            :meth:`_asyncio.AsyncResult.scalars`
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, False, True)
+
+    async def one(self) -> Row[_TP]:
+        """Return exactly one row or raise an exception.
+
+        Raises :class:`.NoResultFound` if the result returns no
+        rows, or :class:`.MultipleResultsFound` if multiple rows
+        would be returned.
+
+        .. note::  This method returns one **row**, e.g. tuple, by default.
+           To return exactly one single scalar value, that is, the first
+           column of the first row, use the
+           :meth:`_asyncio.AsyncResult.scalar_one` method, or combine
+           :meth:`_asyncio.AsyncResult.scalars` and
+           :meth:`_asyncio.AsyncResult.one`.
+
+        .. versionadded:: 1.4
+
+        :return: The first :class:`_engine.Row`.
+
+        :raises: :class:`.MultipleResultsFound`, :class:`.NoResultFound`
+
+        .. seealso::
+
+            :meth:`_asyncio.AsyncResult.first`
+
+            :meth:`_asyncio.AsyncResult.one_or_none`
+
+            :meth:`_asyncio.AsyncResult.scalar_one`
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, True, False)
+
+    @overload
+    async def scalar(self: AsyncResult[Tuple[_T]]) -> Optional[_T]: ...
+
+    @overload
+    async def scalar(self) -> Any: ...
+
+    async def scalar(self) -> Any:
+        """Fetch the first column of the first row, and close the result set.
+
+        Returns ``None`` if there are no rows to fetch.
+
+        No validation is performed to test if additional rows remain.
+
+        After calling this method, the object is fully closed,
+        e.g. the :meth:`_engine.CursorResult.close`
+        method will have been called.
+
+        :return: a Python scalar value, or ``None`` if no rows remain.
+
+        """
+        return await greenlet_spawn(self._only_one_row, False, False, True)
+
+    async def freeze(self) -> FrozenResult[_TP]:
+        """Return a callable object that will produce copies of this
+        :class:`_asyncio.AsyncResult` when invoked.
+
+        The callable object returned is an instance of
+        :class:`_engine.FrozenResult`.
+
+        This is used for result set caching.  The method must be called
+        on the result when it has been unconsumed, and calling the method
+        will consume the result fully.   When the :class:`_engine.FrozenResult`
+        is retrieved from a cache, it can be called any number of times where
+        it will produce a new :class:`_engine.Result` object each time
+        against its stored set of rows.
+
+        .. seealso::
+
+            :ref:`do_orm_execute_re_executing` - example usage within the
+            ORM to implement a result-set cache.
+
+        """
+
+        return await greenlet_spawn(FrozenResult, self)
+
+    @overload
+    def scalars(
+        self: AsyncResult[Tuple[_T]], index: Literal[0]
+    ) -> AsyncScalarResult[_T]: ...
+
+    @overload
+    def scalars(self: AsyncResult[Tuple[_T]]) -> AsyncScalarResult[_T]: ...
+
+    @overload
+    def scalars(self, index: _KeyIndexType = 0) -> AsyncScalarResult[Any]: ...
+
+    def scalars(self, index: _KeyIndexType = 0) -> AsyncScalarResult[Any]:
+        """Return an :class:`_asyncio.AsyncScalarResult` filtering object which
+        will return single elements rather than :class:`_row.Row` objects.
+
+        Refer to :meth:`_result.Result.scalars` in the synchronous
+        SQLAlchemy API for a complete behavioral description.
+
+        :param index: integer or row key indicating the column to be fetched
+         from each row, defaults to ``0`` indicating the first column.
+
+        :return: a new :class:`_asyncio.AsyncScalarResult` filtering object
+         referring to this :class:`_asyncio.AsyncResult` object.
+
+        """
+        return AsyncScalarResult(self._real_result, index)
+
+    def mappings(self) -> AsyncMappingResult:
+        """Apply a mappings filter to returned rows, returning an instance of
+        :class:`_asyncio.AsyncMappingResult`.
+
+        When this filter is applied, fetching rows will return
+        :class:`_engine.RowMapping` objects instead of :class:`_engine.Row`
+        objects.
+
+        :return: a new :class:`_asyncio.AsyncMappingResult` filtering object
+         referring to the underlying :class:`_result.Result` object.
+
+        """
+
+        return AsyncMappingResult(self._real_result)
+
+
+class AsyncScalarResult(AsyncCommon[_R]):
+    """A wrapper for a :class:`_asyncio.AsyncResult` that returns scalar values
+    rather than :class:`_row.Row` values.
+
+    The :class:`_asyncio.AsyncScalarResult` object is acquired by calling the
+    :meth:`_asyncio.AsyncResult.scalars` method.
+
+    Refer to the :class:`_result.ScalarResult` object in the synchronous
+    SQLAlchemy API for a complete behavioral description.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    _generate_rows = False
+
+    def __init__(self, real_result: Result[Any], index: _KeyIndexType):
+        self._real_result = real_result
+
+        if real_result._source_supports_scalars:
+            self._metadata = real_result._metadata
+            self._post_creational_filter = None
+        else:
+            self._metadata = real_result._metadata._reduce([index])
+            self._post_creational_filter = operator.itemgetter(0)
+
+        self._unique_filter_state = real_result._unique_filter_state
+
+    def unique(
+        self,
+        strategy: Optional[_UniqueFilterType] = None,
+    ) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_asyncio.AsyncScalarResult`.
+
+        See :meth:`_asyncio.AsyncResult.unique` for usage details.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    async def partitions(
+        self, size: Optional[int] = None
+    ) -> AsyncIterator[Sequence[_R]]:
+        """Iterate through sub-lists of elements of the size given.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.partitions` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = await greenlet_spawn(getter, self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    async def fetchall(self) -> Sequence[_R]:
+        """A synonym for the :meth:`_asyncio.AsyncScalarResult.all` method."""
+
+        return await greenlet_spawn(self._allrows)
+
+    async def fetchmany(self, size: Optional[int] = None) -> Sequence[_R]:
+        """Fetch many objects.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.fetchmany` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return await greenlet_spawn(self._manyrow_getter, self, size)
+
+    async def all(self) -> Sequence[_R]:
+        """Return all scalar values in a list.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.all` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return await greenlet_spawn(self._allrows)
+
+    def __aiter__(self) -> AsyncScalarResult[_R]:
+        return self
+
+    async def __anext__(self) -> _R:
+        row = await greenlet_spawn(self._onerow_getter, self)
+        if row is _NO_ROW:
+            raise StopAsyncIteration()
+        else:
+            return row
+
+    async def first(self) -> Optional[_R]:
+        """Fetch the first object or ``None`` if no object is present.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.first` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, False, False, False)
+
+    async def one_or_none(self) -> Optional[_R]:
+        """Return at most one object or raise an exception.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.one_or_none` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, False, False)
+
+    async def one(self) -> _R:
+        """Return exactly one object or raise an exception.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.one` except that
+        scalar values, rather than :class:`_engine.Row` objects,
+        are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, True, False)
+
+
+class AsyncMappingResult(_WithKeys, AsyncCommon[RowMapping]):
+    """A wrapper for a :class:`_asyncio.AsyncResult` that returns dictionary
+    values rather than :class:`_engine.Row` values.
+
+    The :class:`_asyncio.AsyncMappingResult` object is acquired by calling the
+    :meth:`_asyncio.AsyncResult.mappings` method.
+
+    Refer to the :class:`_result.MappingResult` object in the synchronous
+    SQLAlchemy API for a complete behavioral description.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    _generate_rows = True
+
+    _post_creational_filter = operator.attrgetter("_mapping")
+
+    def __init__(self, result: Result[Any]):
+        self._real_result = result
+        self._unique_filter_state = result._unique_filter_state
+        self._metadata = result._metadata
+        if result._source_supports_scalars:
+            self._metadata = self._metadata._reduce([0])
+
+    def unique(
+        self,
+        strategy: Optional[_UniqueFilterType] = None,
+    ) -> Self:
+        """Apply unique filtering to the objects returned by this
+        :class:`_asyncio.AsyncMappingResult`.
+
+        See :meth:`_asyncio.AsyncResult.unique` for usage details.
+
+        """
+        self._unique_filter_state = (set(), strategy)
+        return self
+
+    def columns(self, *col_expressions: _KeyIndexType) -> Self:
+        r"""Establish the columns that should be returned in each row."""
+        return self._column_slices(col_expressions)
+
+    async def partitions(
+        self, size: Optional[int] = None
+    ) -> AsyncIterator[Sequence[RowMapping]]:
+        """Iterate through sub-lists of elements of the size given.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.partitions` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        getter = self._manyrow_getter
+
+        while True:
+            partition = await greenlet_spawn(getter, self, size)
+            if partition:
+                yield partition
+            else:
+                break
+
+    async def fetchall(self) -> Sequence[RowMapping]:
+        """A synonym for the :meth:`_asyncio.AsyncMappingResult.all` method."""
+
+        return await greenlet_spawn(self._allrows)
+
+    async def fetchone(self) -> Optional[RowMapping]:
+        """Fetch one object.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.fetchone` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        row = await greenlet_spawn(self._onerow_getter, self)
+        if row is _NO_ROW:
+            return None
+        else:
+            return row
+
+    async def fetchmany(
+        self, size: Optional[int] = None
+    ) -> Sequence[RowMapping]:
+        """Fetch many rows.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.fetchmany` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        return await greenlet_spawn(self._manyrow_getter, self, size)
+
+    async def all(self) -> Sequence[RowMapping]:
+        """Return all rows in a list.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.all` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+
+        return await greenlet_spawn(self._allrows)
+
+    def __aiter__(self) -> AsyncMappingResult:
+        return self
+
+    async def __anext__(self) -> RowMapping:
+        row = await greenlet_spawn(self._onerow_getter, self)
+        if row is _NO_ROW:
+            raise StopAsyncIteration()
+        else:
+            return row
+
+    async def first(self) -> Optional[RowMapping]:
+        """Fetch the first object or ``None`` if no object is present.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.first` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, False, False, False)
+
+    async def one_or_none(self) -> Optional[RowMapping]:
+        """Return at most one object or raise an exception.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.one_or_none` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, False, False)
+
+    async def one(self) -> RowMapping:
+        """Return exactly one object or raise an exception.
+
+        Equivalent to :meth:`_asyncio.AsyncResult.one` except that
+        :class:`_engine.RowMapping` values, rather than :class:`_engine.Row`
+        objects, are returned.
+
+        """
+        return await greenlet_spawn(self._only_one_row, True, True, False)
+
+
+class AsyncTupleResult(AsyncCommon[_R], util.TypingOnly):
+    """A :class:`_asyncio.AsyncResult` that's typed as returning plain
+    Python tuples instead of rows.
+
+    Since :class:`_engine.Row` acts like a tuple in every way already,
+    this class is a typing only class, regular :class:`_asyncio.AsyncResult` is
+    still used at runtime.
+
+    """
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        async def partitions(
+            self, size: Optional[int] = None
+        ) -> AsyncIterator[Sequence[_R]]:
+            """Iterate through sub-lists of elements of the size given.
+
+            Equivalent to :meth:`_result.Result.partitions` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        async def fetchone(self) -> Optional[_R]:
+            """Fetch one tuple.
+
+            Equivalent to :meth:`_result.Result.fetchone` except that
+            tuple values, rather than :class:`_engine.Row`
+            objects, are returned.
+
+            """
+            ...
+
+        async def fetchall(self) -> Sequence[_R]:
+            """A synonym for the :meth:`_engine.ScalarResult.all` method."""
+            ...
+
+        async def fetchmany(self, size: Optional[int] = None) -> Sequence[_R]:
+            """Fetch many objects.
+
+            Equivalent to :meth:`_result.Result.fetchmany` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        async def all(self) -> Sequence[_R]:  # noqa: A001
+            """Return all scalar values in a list.
+
+            Equivalent to :meth:`_result.Result.all` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        async def __aiter__(self) -> AsyncIterator[_R]: ...
+
+        async def __anext__(self) -> _R: ...
+
+        async def first(self) -> Optional[_R]:
+            """Fetch the first object or ``None`` if no object is present.
+
+            Equivalent to :meth:`_result.Result.first` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+
+            """
+            ...
+
+        async def one_or_none(self) -> Optional[_R]:
+            """Return at most one object or raise an exception.
+
+            Equivalent to :meth:`_result.Result.one_or_none` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        async def one(self) -> _R:
+            """Return exactly one object or raise an exception.
+
+            Equivalent to :meth:`_result.Result.one` except that
+            tuple values, rather than :class:`_engine.Row` objects,
+            are returned.
+
+            """
+            ...
+
+        @overload
+        async def scalar_one(self: AsyncTupleResult[Tuple[_T]]) -> _T: ...
+
+        @overload
+        async def scalar_one(self) -> Any: ...
+
+        async def scalar_one(self) -> Any:
+            """Return exactly one scalar result or raise an exception.
+
+            This is equivalent to calling :meth:`_engine.Result.scalars`
+            and then :meth:`_engine.AsyncScalarResult.one`.
+
+            .. seealso::
+
+                :meth:`_engine.AsyncScalarResult.one`
+
+                :meth:`_engine.Result.scalars`
+
+            """
+            ...
+
+        @overload
+        async def scalar_one_or_none(
+            self: AsyncTupleResult[Tuple[_T]],
+        ) -> Optional[_T]: ...
+
+        @overload
+        async def scalar_one_or_none(self) -> Optional[Any]: ...
+
+        async def scalar_one_or_none(self) -> Optional[Any]:
+            """Return exactly one or no scalar result.
+
+            This is equivalent to calling :meth:`_engine.Result.scalars`
+            and then :meth:`_engine.AsyncScalarResult.one_or_none`.
+
+            .. seealso::
+
+                :meth:`_engine.AsyncScalarResult.one_or_none`
+
+                :meth:`_engine.Result.scalars`
+
+            """
+            ...
+
+        @overload
+        async def scalar(
+            self: AsyncTupleResult[Tuple[_T]],
+        ) -> Optional[_T]: ...
+
+        @overload
+        async def scalar(self) -> Any: ...
+
+        async def scalar(self) -> Any:
+            """Fetch the first column of the first row, and close the result
+            set.
+
+            Returns ``None`` if there are no rows to fetch.
+
+            No validation is performed to test if additional rows remain.
+
+            After calling this method, the object is fully closed,
+            e.g. the :meth:`_engine.CursorResult.close`
+            method will have been called.
+
+            :return: a Python scalar value , or ``None`` if no rows remain.
+
+            """
+            ...
+
+
+_RT = TypeVar("_RT", bound="Result[Any]")
+
+
+async def _ensure_sync_result(result: _RT, calling_method: Any) -> _RT:
+    cursor_result: CursorResult[Any]
+
+    try:
+        is_cursor = result._is_cursor
+    except AttributeError:
+        # legacy execute(DefaultGenerator) case
+        return result
+
+    if not is_cursor:
+        cursor_result = getattr(result, "raw", None)  # type: ignore
+    else:
+        cursor_result = result  # type: ignore
+    if cursor_result and cursor_result.context._is_server_side:
+        await greenlet_spawn(cursor_result.close)
+        raise async_exc.AsyncMethodRequired(
+            "Can't use the %s.%s() method with a "
+            "server-side cursor. "
+            "Use the %s.stream() method for an async "
+            "streaming result set."
+            % (
+                calling_method.__self__.__class__.__name__,
+                calling_method.__name__,
+                calling_method.__self__.__class__.__name__,
+            )
+        )
+    return result
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/scoping.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/scoping.py
new file mode 100644
index 00000000..7ecab37b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/scoping.py
@@ -0,0 +1,1614 @@
+# ext/asyncio/scoping.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .session import _AS
+from .session import async_sessionmaker
+from .session import AsyncSession
+from ... import exc as sa_exc
+from ... import util
+from ...orm.session import Session
+from ...util import create_proxy_methods
+from ...util import ScopedRegistry
+from ...util import warn
+from ...util import warn_deprecated
+
+if TYPE_CHECKING:
+    from .engine import AsyncConnection
+    from .result import AsyncResult
+    from .result import AsyncScalarResult
+    from .session import AsyncSessionTransaction
+    from ...engine import Connection
+    from ...engine import CursorResult
+    from ...engine import Engine
+    from ...engine import Result
+    from ...engine import Row
+    from ...engine import RowMapping
+    from ...engine.interfaces import _CoreAnyExecuteParams
+    from ...engine.interfaces import CoreExecuteOptionsParameter
+    from ...engine.result import ScalarResult
+    from ...orm._typing import _IdentityKeyType
+    from ...orm._typing import _O
+    from ...orm._typing import OrmExecuteOptionsParameter
+    from ...orm.interfaces import ORMOption
+    from ...orm.session import _BindArguments
+    from ...orm.session import _EntityBindKey
+    from ...orm.session import _PKIdentityArgument
+    from ...orm.session import _SessionBind
+    from ...sql.base import Executable
+    from ...sql.dml import UpdateBase
+    from ...sql.elements import ClauseElement
+    from ...sql.selectable import ForUpdateParameter
+    from ...sql.selectable import TypedReturnsRows
+
+_T = TypeVar("_T", bound=Any)
+
+
+@create_proxy_methods(
+    AsyncSession,
+    ":class:`_asyncio.AsyncSession`",
+    ":class:`_asyncio.scoping.async_scoped_session`",
+    classmethods=["close_all", "object_session", "identity_key"],
+    methods=[
+        "__contains__",
+        "__iter__",
+        "aclose",
+        "add",
+        "add_all",
+        "begin",
+        "begin_nested",
+        "close",
+        "reset",
+        "commit",
+        "connection",
+        "delete",
+        "execute",
+        "expire",
+        "expire_all",
+        "expunge",
+        "expunge_all",
+        "flush",
+        "get_bind",
+        "is_modified",
+        "invalidate",
+        "merge",
+        "refresh",
+        "rollback",
+        "scalar",
+        "scalars",
+        "get",
+        "get_one",
+        "stream",
+        "stream_scalars",
+    ],
+    attributes=[
+        "bind",
+        "dirty",
+        "deleted",
+        "new",
+        "identity_map",
+        "is_active",
+        "autoflush",
+        "no_autoflush",
+        "info",
+    ],
+    use_intermediate_variable=["get"],
+)
+class async_scoped_session(Generic[_AS]):
+    """Provides scoped management of :class:`.AsyncSession` objects.
+
+    See the section :ref:`asyncio_scoped_session` for usage details.
+
+    .. versionadded:: 1.4.19
+
+
+    """
+
+    _support_async = True
+
+    session_factory: async_sessionmaker[_AS]
+    """The `session_factory` provided to `__init__` is stored in this
+    attribute and may be accessed at a later time.  This can be useful when
+    a new non-scoped :class:`.AsyncSession` is needed."""
+
+    registry: ScopedRegistry[_AS]
+
+    def __init__(
+        self,
+        session_factory: async_sessionmaker[_AS],
+        scopefunc: Callable[[], Any],
+    ):
+        """Construct a new :class:`_asyncio.async_scoped_session`.
+
+        :param session_factory: a factory to create new :class:`_asyncio.AsyncSession`
+         instances. This is usually, but not necessarily, an instance
+         of :class:`_asyncio.async_sessionmaker`.
+
+        :param scopefunc: function which defines
+         the current scope.   A function such as ``asyncio.current_task``
+         may be useful here.
+
+        """  # noqa: E501
+
+        self.session_factory = session_factory
+        self.registry = ScopedRegistry(session_factory, scopefunc)
+
+    @property
+    def _proxied(self) -> _AS:
+        return self.registry()
+
+    def __call__(self, **kw: Any) -> _AS:
+        r"""Return the current :class:`.AsyncSession`, creating it
+        using the :attr:`.scoped_session.session_factory` if not present.
+
+        :param \**kw: Keyword arguments will be passed to the
+         :attr:`.scoped_session.session_factory` callable, if an existing
+         :class:`.AsyncSession` is not present.  If the
+         :class:`.AsyncSession` is present
+         and keyword arguments have been passed,
+         :exc:`~sqlalchemy.exc.InvalidRequestError` is raised.
+
+        """
+        if kw:
+            if self.registry.has():
+                raise sa_exc.InvalidRequestError(
+                    "Scoped session is already present; "
+                    "no new arguments may be specified."
+                )
+            else:
+                sess = self.session_factory(**kw)
+                self.registry.set(sess)
+        else:
+            sess = self.registry()
+        if not self._support_async and sess._is_asyncio:
+            warn_deprecated(
+                "Using `scoped_session` with asyncio is deprecated and "
+                "will raise an error in a future version. "
+                "Please use `async_scoped_session` instead.",
+                "1.4.23",
+            )
+        return sess
+
+    def configure(self, **kwargs: Any) -> None:
+        """reconfigure the :class:`.sessionmaker` used by this
+        :class:`.scoped_session`.
+
+        See :meth:`.sessionmaker.configure`.
+
+        """
+
+        if self.registry.has():
+            warn(
+                "At least one scoped session is already present. "
+                " configure() can not affect sessions that have "
+                "already been created."
+            )
+
+        self.session_factory.configure(**kwargs)
+
+    async def remove(self) -> None:
+        """Dispose of the current :class:`.AsyncSession`, if present.
+
+        Different from scoped_session's remove method, this method would use
+        await to wait for the close method of AsyncSession.
+
+        """
+
+        if self.registry.has():
+            await self.registry().close()
+        self.registry.clear()
+
+    # START PROXY METHODS async_scoped_session
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_proxy_methods.py
+
+    def __contains__(self, instance: object) -> bool:
+        r"""Return True if the instance is associated with this session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        The instance may be pending or persistent within the Session for a
+        result of True.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.__contains__(instance)
+
+    def __iter__(self) -> Iterator[object]:
+        r"""Iterate over all pending or persistent instances within this
+        Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.__iter__()
+
+    async def aclose(self) -> None:
+        r"""A synonym for :meth:`_asyncio.AsyncSession.close`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        The :meth:`_asyncio.AsyncSession.aclose` name is specifically
+        to support the Python standard library ``@contextlib.aclosing``
+        context manager function.
+
+        .. versionadded:: 2.0.20
+
+
+        """  # noqa: E501
+
+        return await self._proxied.aclose()
+
+    def add(self, instance: object, _warn: bool = True) -> None:
+        r"""Place an object into this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        Objects that are in the :term:`transient` state when passed to the
+        :meth:`_orm.Session.add` method will move to the
+        :term:`pending` state, until the next flush, at which point they
+        will move to the :term:`persistent` state.
+
+        Objects that are in the :term:`detached` state when passed to the
+        :meth:`_orm.Session.add` method will move to the :term:`persistent`
+        state directly.
+
+        If the transaction used by the :class:`_orm.Session` is rolled back,
+        objects which were transient when they were passed to
+        :meth:`_orm.Session.add` will be moved back to the
+        :term:`transient` state, and will no longer be present within this
+        :class:`_orm.Session`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add_all`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.add(instance, _warn=_warn)
+
+    def add_all(self, instances: Iterable[object]) -> None:
+        r"""Add the given collection of instances to this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        See the documentation for :meth:`_orm.Session.add` for a general
+        behavioral description.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.add_all(instances)
+
+    def begin(self) -> AsyncSessionTransaction:
+        r"""Return an :class:`_asyncio.AsyncSessionTransaction` object.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        The underlying :class:`_orm.Session` will perform the
+        "begin" action when the :class:`_asyncio.AsyncSessionTransaction`
+        object is entered::
+
+            async with async_session.begin():
+                ...  # ORM transaction is begun
+
+        Note that database IO will not normally occur when the session-level
+        transaction is begun, as database transactions begin on an
+        on-demand basis.  However, the begin block is async to accommodate
+        for a :meth:`_orm.SessionEvents.after_transaction_create`
+        event hook that may perform IO.
+
+        For a general description of ORM begin, see
+        :meth:`_orm.Session.begin`.
+
+
+        """  # noqa: E501
+
+        return self._proxied.begin()
+
+    def begin_nested(self) -> AsyncSessionTransaction:
+        r"""Return an :class:`_asyncio.AsyncSessionTransaction` object
+        which will begin a "nested" transaction, e.g. SAVEPOINT.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        Behavior is the same as that of :meth:`_asyncio.AsyncSession.begin`.
+
+        For a general description of ORM begin nested, see
+        :meth:`_orm.Session.begin_nested`.
+
+        .. seealso::
+
+            :ref:`aiosqlite_serializable` - special workarounds required
+            with the SQLite asyncio driver in order for SAVEPOINT to work
+            correctly.
+
+
+        """  # noqa: E501
+
+        return self._proxied.begin_nested()
+
+    async def close(self) -> None:
+        r"""Close out the transactional resources and ORM objects used by this
+        :class:`_asyncio.AsyncSession`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.close` - main documentation for
+            "close"
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_asyncio.AsyncSession.close` and
+            :meth:`_asyncio.AsyncSession.reset`.
+
+
+        """  # noqa: E501
+
+        return await self._proxied.close()
+
+    async def reset(self) -> None:
+        r"""Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`, resetting the session to its initial state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. versionadded:: 2.0.22
+
+        .. seealso::
+
+            :meth:`_orm.Session.reset` - main documentation for
+            "reset"
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_asyncio.AsyncSession.close` and
+            :meth:`_asyncio.AsyncSession.reset`.
+
+
+        """  # noqa: E501
+
+        return await self._proxied.reset()
+
+    async def commit(self) -> None:
+        r"""Commit the current transaction in progress.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.commit` - main documentation for
+            "commit"
+
+        """  # noqa: E501
+
+        return await self._proxied.commit()
+
+    async def connection(
+        self,
+        bind_arguments: Optional[_BindArguments] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+        **kw: Any,
+    ) -> AsyncConnection:
+        r"""Return a :class:`_asyncio.AsyncConnection` object corresponding to
+        this :class:`.Session` object's transactional state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        This method may also be used to establish execution options for the
+        database connection used by the current transaction.
+
+        .. versionadded:: 1.4.24  Added \**kw arguments which are passed
+           through to the underlying :meth:`_orm.Session.connection` method.
+
+        .. seealso::
+
+            :meth:`_orm.Session.connection` - main documentation for
+            "connection"
+
+
+        """  # noqa: E501
+
+        return await self._proxied.connection(
+            bind_arguments=bind_arguments,
+            execution_options=execution_options,
+            **kw,
+        )
+
+    async def delete(self, instance: object) -> None:
+        r"""Mark an instance as deleted.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        The database delete operation occurs upon ``flush()``.
+
+        As this operation may need to cascade along unloaded relationships,
+        it is awaitable to allow for those queries to take place.
+
+        .. seealso::
+
+            :meth:`_orm.Session.delete` - main documentation for delete
+
+
+        """  # noqa: E501
+
+        return await self._proxied.delete(instance)
+
+    @overload
+    async def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[_T]: ...
+
+    @overload
+    async def execute(
+        self,
+        statement: UpdateBase,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> CursorResult[Any]: ...
+
+    @overload
+    async def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]: ...
+
+    async def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Result[Any]:
+        r"""Execute a statement and return a buffered
+        :class:`_engine.Result` object.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.execute` - main documentation for execute
+
+
+        """  # noqa: E501
+
+        return await self._proxied.execute(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    def expire(
+        self, instance: object, attribute_names: Optional[Iterable[str]] = None
+    ) -> None:
+        r"""Expire the attributes on an instance.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        Marks the attributes of an instance as out of date. When an expired
+        attribute is next accessed, a query will be issued to the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire all objects in the :class:`.Session` simultaneously,
+        use :meth:`Session.expire_all`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire` only makes sense for the specific
+        case that a non-ORM SQL statement was emitted in the current
+        transaction.
+
+        :param instance: The instance to be refreshed.
+        :param attribute_names: optional list of string attribute names
+          indicating a subset of attributes to be expired.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire(instance, attribute_names=attribute_names)
+
+    def expire_all(self) -> None:
+        r"""Expires all persistent instances within this Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        When any attributes on a persistent instance is next accessed,
+        a query will be issued using the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire individual objects and individual attributes
+        on those objects, use :meth:`Session.expire`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire_all` is not usually needed,
+        assuming the transaction is isolated.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire_all()
+
+    def expunge(self, instance: object) -> None:
+        r"""Remove the `instance` from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This will free all internal references to the instance.  Cascading
+        will be applied according to the *expunge* cascade rule.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge(instance)
+
+    def expunge_all(self) -> None:
+        r"""Remove all object instances from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is equivalent to calling ``expunge(obj)`` on all objects in this
+        ``Session``.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge_all()
+
+    async def flush(self, objects: Optional[Sequence[Any]] = None) -> None:
+        r"""Flush all the object changes to the database.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.flush` - main documentation for flush
+
+
+        """  # noqa: E501
+
+        return await self._proxied.flush(objects=objects)
+
+    def get_bind(
+        self,
+        mapper: Optional[_EntityBindKey[_O]] = None,
+        clause: Optional[ClauseElement] = None,
+        bind: Optional[_SessionBind] = None,
+        **kw: Any,
+    ) -> Union[Engine, Connection]:
+        r"""Return a "bind" to which the synchronous proxied :class:`_orm.Session`
+        is bound.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        Unlike the :meth:`_orm.Session.get_bind` method, this method is
+        currently **not** used by this :class:`.AsyncSession` in any way
+        in order to resolve engines for requests.
+
+        .. note::
+
+            This method proxies directly to the :meth:`_orm.Session.get_bind`
+            method, however is currently **not** useful as an override target,
+            in contrast to that of the :meth:`_orm.Session.get_bind` method.
+            The example below illustrates how to implement custom
+            :meth:`_orm.Session.get_bind` schemes that work with
+            :class:`.AsyncSession` and :class:`.AsyncEngine`.
+
+        The pattern introduced at :ref:`session_custom_partitioning`
+        illustrates how to apply a custom bind-lookup scheme to a
+        :class:`_orm.Session` given a set of :class:`_engine.Engine` objects.
+        To apply a corresponding :meth:`_orm.Session.get_bind` implementation
+        for use with a :class:`.AsyncSession` and :class:`.AsyncEngine`
+        objects, continue to subclass :class:`_orm.Session` and apply it to
+        :class:`.AsyncSession` using
+        :paramref:`.AsyncSession.sync_session_class`. The inner method must
+        continue to return :class:`_engine.Engine` instances, which can be
+        acquired from a :class:`_asyncio.AsyncEngine` using the
+        :attr:`_asyncio.AsyncEngine.sync_engine` attribute::
+
+            # using example from "Custom Vertical Partitioning"
+
+
+            import random
+
+            from sqlalchemy.ext.asyncio import AsyncSession
+            from sqlalchemy.ext.asyncio import create_async_engine
+            from sqlalchemy.ext.asyncio import async_sessionmaker
+            from sqlalchemy.orm import Session
+
+            # construct async engines w/ async drivers
+            engines = {
+                "leader": create_async_engine("sqlite+aiosqlite:///leader.db"),
+                "other": create_async_engine("sqlite+aiosqlite:///other.db"),
+                "follower1": create_async_engine("sqlite+aiosqlite:///follower1.db"),
+                "follower2": create_async_engine("sqlite+aiosqlite:///follower2.db"),
+            }
+
+
+            class RoutingSession(Session):
+                def get_bind(self, mapper=None, clause=None, **kw):
+                    # within get_bind(), return sync engines
+                    if mapper and issubclass(mapper.class_, MyOtherClass):
+                        return engines["other"].sync_engine
+                    elif self._flushing or isinstance(clause, (Update, Delete)):
+                        return engines["leader"].sync_engine
+                    else:
+                        return engines[
+                            random.choice(["follower1", "follower2"])
+                        ].sync_engine
+
+
+            # apply to AsyncSession using sync_session_class
+            AsyncSessionMaker = async_sessionmaker(sync_session_class=RoutingSession)
+
+        The :meth:`_orm.Session.get_bind` method is called in a non-asyncio,
+        implicitly non-blocking context in the same manner as ORM event hooks
+        and functions that are invoked via :meth:`.AsyncSession.run_sync`, so
+        routines that wish to run SQL commands inside of
+        :meth:`_orm.Session.get_bind` can continue to do so using
+        blocking-style code, which will be translated to implicitly async calls
+        at the point of invoking IO on the database drivers.
+
+
+        """  # noqa: E501
+
+        return self._proxied.get_bind(
+            mapper=mapper, clause=clause, bind=bind, **kw
+        )
+
+    def is_modified(
+        self, instance: object, include_collections: bool = True
+    ) -> bool:
+        r"""Return ``True`` if the given instance has locally
+        modified attributes.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This method retrieves the history for each instrumented
+        attribute on the instance and performs a comparison of the current
+        value to its previously flushed or committed value, if any.
+
+        It is in effect a more expensive and accurate
+        version of checking for the given instance in the
+        :attr:`.Session.dirty` collection; a full test for
+        each attribute's net "dirty" status is performed.
+
+        E.g.::
+
+            return session.is_modified(someobject)
+
+        A few caveats to this method apply:
+
+        * Instances present in the :attr:`.Session.dirty` collection may
+          report ``False`` when tested with this method.  This is because
+          the object may have received change events via attribute mutation,
+          thus placing it in :attr:`.Session.dirty`, but ultimately the state
+          is the same as that loaded from the database, resulting in no net
+          change here.
+        * Scalar attributes may not have recorded the previously set
+          value when a new value was applied, if the attribute was not loaded,
+          or was expired, at the time the new value was received - in these
+          cases, the attribute is assumed to have a change, even if there is
+          ultimately no net change against its database value. SQLAlchemy in
+          most cases does not need the "old" value when a set event occurs, so
+          it skips the expense of a SQL call if the old value isn't present,
+          based on the assumption that an UPDATE of the scalar value is
+          usually needed, and in those few cases where it isn't, is less
+          expensive on average than issuing a defensive SELECT.
+
+          The "old" value is fetched unconditionally upon set only if the
+          attribute container has the ``active_history`` flag set to ``True``.
+          This flag is set typically for primary key attributes and scalar
+          object references that are not a simple many-to-one.  To set this
+          flag for any arbitrary mapped column, use the ``active_history``
+          argument with :func:`.column_property`.
+
+        :param instance: mapped instance to be tested for pending changes.
+        :param include_collections: Indicates if multivalued collections
+         should be included in the operation.  Setting this to ``False`` is a
+         way to detect only local-column based properties (i.e. scalar columns
+         or many-to-one foreign keys) that would result in an UPDATE for this
+         instance upon flush.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_modified(
+            instance, include_collections=include_collections
+        )
+
+    async def invalidate(self) -> None:
+        r"""Close this Session, using connection invalidation.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        For a complete description, see :meth:`_orm.Session.invalidate`.
+
+        """  # noqa: E501
+
+        return await self._proxied.invalidate()
+
+    async def merge(
+        self,
+        instance: _O,
+        *,
+        load: bool = True,
+        options: Optional[Sequence[ORMOption]] = None,
+    ) -> _O:
+        r"""Copy the state of a given instance into a corresponding instance
+        within this :class:`_asyncio.AsyncSession`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.merge` - main documentation for merge
+
+
+        """  # noqa: E501
+
+        return await self._proxied.merge(instance, load=load, options=options)
+
+    async def refresh(
+        self,
+        instance: object,
+        attribute_names: Optional[Iterable[str]] = None,
+        with_for_update: ForUpdateParameter = None,
+    ) -> None:
+        r"""Expire and refresh the attributes on the given instance.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        A query will be issued to the database and all attributes will be
+        refreshed with their current database value.
+
+        This is the async version of the :meth:`_orm.Session.refresh` method.
+        See that method for a complete description of all options.
+
+        .. seealso::
+
+            :meth:`_orm.Session.refresh` - main documentation for refresh
+
+
+        """  # noqa: E501
+
+        return await self._proxied.refresh(
+            instance,
+            attribute_names=attribute_names,
+            with_for_update=with_for_update,
+        )
+
+    async def rollback(self) -> None:
+        r"""Rollback the current transaction in progress.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.rollback` - main documentation for
+            "rollback"
+
+        """  # noqa: E501
+
+        return await self._proxied.rollback()
+
+    @overload
+    async def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Optional[_T]: ...
+
+    @overload
+    async def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any: ...
+
+    async def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any:
+        r"""Execute a statement and return a scalar result.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalar` - main documentation for scalar
+
+
+        """  # noqa: E501
+
+        return await self._proxied.scalar(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @overload
+    async def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    async def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]: ...
+
+    async def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]:
+        r"""Execute a statement and return scalar results.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        :return: a :class:`_result.ScalarResult` object
+
+        .. versionadded:: 1.4.24 Added :meth:`_asyncio.AsyncSession.scalars`
+
+        .. versionadded:: 1.4.26 Added
+           :meth:`_asyncio.async_scoped_session.scalars`
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalars` - main documentation for scalars
+
+            :meth:`_asyncio.AsyncSession.stream_scalars` - streaming version
+
+
+        """  # noqa: E501
+
+        return await self._proxied.scalars(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    async def get(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+    ) -> Union[_O, None]:
+        r"""Return an instance based on the given primary key identifier,
+        or ``None`` if not found.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. seealso::
+
+            :meth:`_orm.Session.get` - main documentation for get
+
+
+
+        """  # noqa: E501
+
+        result = await self._proxied.get(
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+        )
+        return result
+
+    async def get_one(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+    ) -> _O:
+        r"""Return an instance based on the given primary key identifier,
+        or raise an exception if not found.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        Raises ``sqlalchemy.orm.exc.NoResultFound`` if the query selects
+        no rows.
+
+        ..versionadded: 2.0.22
+
+        .. seealso::
+
+            :meth:`_orm.Session.get_one` - main documentation for get_one
+
+
+        """  # noqa: E501
+
+        return await self._proxied.get_one(
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+        )
+
+    @overload
+    async def stream(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[_T]: ...
+
+    @overload
+    async def stream(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[Any]: ...
+
+    async def stream(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[Any]:
+        r"""Execute a statement and return a streaming
+        :class:`_asyncio.AsyncResult` object.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+
+        """  # noqa: E501
+
+        return await self._proxied.stream(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @overload
+    async def stream_scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[_T]: ...
+
+    @overload
+    async def stream_scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[Any]: ...
+
+    async def stream_scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[Any]:
+        r"""Execute a statement and return a stream of scalar results.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        :return: an :class:`_asyncio.AsyncScalarResult` object
+
+        .. versionadded:: 1.4.24
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalars` - main documentation for scalars
+
+            :meth:`_asyncio.AsyncSession.scalars` - non streaming version
+
+
+        """  # noqa: E501
+
+        return await self._proxied.stream_scalars(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @property
+    def bind(self) -> Any:
+        r"""Proxy for the :attr:`_asyncio.AsyncSession.bind` attribute
+        on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.bind
+
+    @bind.setter
+    def bind(self, attr: Any) -> None:
+        self._proxied.bind = attr
+
+    @property
+    def dirty(self) -> Any:
+        r"""The set of all persistent instances considered dirty.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        E.g.::
+
+            some_mapped_object in session.dirty
+
+        Instances are considered dirty when they were modified but not
+        deleted.
+
+        Note that this 'dirty' calculation is 'optimistic'; most
+        attribute-setting or collection modification operations will
+        mark an instance as 'dirty' and place it in this set, even if
+        there is no net change to the attribute's value.  At flush
+        time, the value of each attribute is compared to its
+        previously saved value, and if there's no net change, no SQL
+        operation will occur (this is a more expensive operation so
+        it's only done at flush time).
+
+        To check if an instance has actionable net changes to its
+        attributes, use the :meth:`.Session.is_modified` method.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.dirty
+
+    @property
+    def deleted(self) -> Any:
+        r"""The set of all instances marked as 'deleted' within this ``Session``
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.deleted
+
+    @property
+    def new(self) -> Any:
+        r"""The set of all instances marked as 'new' within this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.new
+
+    @property
+    def identity_map(self) -> Any:
+        r"""Proxy for the :attr:`_orm.Session.identity_map` attribute
+        on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.identity_map
+
+    @identity_map.setter
+    def identity_map(self, attr: Any) -> None:
+        self._proxied.identity_map = attr
+
+    @property
+    def is_active(self) -> Any:
+        r"""True if this :class:`.Session` not in "partial rollback" state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. versionchanged:: 1.4 The :class:`_orm.Session` no longer begins
+           a new transaction immediately, so this attribute will be False
+           when the :class:`_orm.Session` is first instantiated.
+
+        "partial rollback" state typically indicates that the flush process
+        of the :class:`_orm.Session` has failed, and that the
+        :meth:`_orm.Session.rollback` method must be emitted in order to
+        fully roll back the transaction.
+
+        If this :class:`_orm.Session` is not in a transaction at all, the
+        :class:`_orm.Session` will autobegin when it is first used, so in this
+        case :attr:`_orm.Session.is_active` will return True.
+
+        Otherwise, if this :class:`_orm.Session` is within a transaction,
+        and that transaction has not been rolled back internally, the
+        :attr:`_orm.Session.is_active` will also return True.
+
+        .. seealso::
+
+            :ref:`faq_session_rollback`
+
+            :meth:`_orm.Session.in_transaction`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_active
+
+    @property
+    def autoflush(self) -> Any:
+        r"""Proxy for the :attr:`_orm.Session.autoflush` attribute
+        on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.autoflush
+
+    @autoflush.setter
+    def autoflush(self, attr: Any) -> None:
+        self._proxied.autoflush = attr
+
+    @property
+    def no_autoflush(self) -> Any:
+        r"""Return a context manager that disables autoflush.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        e.g.::
+
+            with session.no_autoflush:
+
+                some_object = SomeClass()
+                session.add(some_object)
+                # won't autoflush
+                some_object.related_thing = session.query(SomeRelated).first()
+
+        Operations that proceed within the ``with:`` block
+        will not be subject to flushes occurring upon query
+        access.  This is useful when initializing a series
+        of objects which involve existing database queries,
+        where the uncompleted object should not yet be flushed.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.no_autoflush
+
+    @property
+    def info(self) -> Any:
+        r"""A user-modifiable dictionary.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class
+            on behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        The initial value of this dictionary can be populated using the
+        ``info`` argument to the :class:`.Session` constructor or
+        :class:`.sessionmaker` constructor or factory methods.  The dictionary
+        here is always local to this :class:`.Session` and can be modified
+        independently of all other :class:`.Session` objects.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.info
+
+    @classmethod
+    async def close_all(cls) -> None:
+        r"""Close all :class:`_asyncio.AsyncSession` sessions.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. deprecated:: 2.0 The :meth:`.AsyncSession.close_all` method is deprecated and will be removed in a future release.  Please refer to :func:`_asyncio.close_all_sessions`.
+
+        """  # noqa: E501
+
+        return await AsyncSession.close_all()
+
+    @classmethod
+    def object_session(cls, instance: object) -> Optional[Session]:
+        r"""Return the :class:`.Session` to which an object belongs.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is an alias of :func:`.object_session`.
+
+
+
+        """  # noqa: E501
+
+        return AsyncSession.object_session(instance)
+
+    @classmethod
+    def identity_key(
+        cls,
+        class_: Optional[Type[Any]] = None,
+        ident: Union[Any, Tuple[Any, ...]] = None,
+        *,
+        instance: Optional[Any] = None,
+        row: Optional[Union[Row[Any], RowMapping]] = None,
+        identity_token: Optional[Any] = None,
+    ) -> _IdentityKeyType[Any]:
+        r"""Return an identity key.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_asyncio.AsyncSession` class on
+            behalf of the :class:`_asyncio.scoping.async_scoped_session` class.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is an alias of :func:`.util.identity_key`.
+
+
+
+        """  # noqa: E501
+
+        return AsyncSession.identity_key(
+            class_=class_,
+            ident=ident,
+            instance=instance,
+            row=row,
+            identity_token=identity_token,
+        )
+
+    # END PROXY METHODS async_scoped_session
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/session.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/session.py
new file mode 100644
index 00000000..bb276943
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/asyncio/session.py
@@ -0,0 +1,1943 @@
+# ext/asyncio/session.py
+# Copyright (C) 2020-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import asyncio
+from typing import Any
+from typing import Awaitable
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import engine
+from .base import ReversibleProxy
+from .base import StartableContext
+from .result import _ensure_sync_result
+from .result import AsyncResult
+from .result import AsyncScalarResult
+from ... import util
+from ...orm import close_all_sessions as _sync_close_all_sessions
+from ...orm import object_session
+from ...orm import Session
+from ...orm import SessionTransaction
+from ...orm import state as _instance_state
+from ...util.concurrency import greenlet_spawn
+from ...util.typing import Concatenate
+from ...util.typing import ParamSpec
+
+
+if TYPE_CHECKING:
+    from .engine import AsyncConnection
+    from .engine import AsyncEngine
+    from ...engine import Connection
+    from ...engine import CursorResult
+    from ...engine import Engine
+    from ...engine import Result
+    from ...engine import Row
+    from ...engine import RowMapping
+    from ...engine import ScalarResult
+    from ...engine.interfaces import _CoreAnyExecuteParams
+    from ...engine.interfaces import CoreExecuteOptionsParameter
+    from ...event import dispatcher
+    from ...orm._typing import _IdentityKeyType
+    from ...orm._typing import _O
+    from ...orm._typing import OrmExecuteOptionsParameter
+    from ...orm.identity import IdentityMap
+    from ...orm.interfaces import ORMOption
+    from ...orm.session import _BindArguments
+    from ...orm.session import _EntityBindKey
+    from ...orm.session import _PKIdentityArgument
+    from ...orm.session import _SessionBind
+    from ...orm.session import _SessionBindKey
+    from ...sql._typing import _InfoType
+    from ...sql.base import Executable
+    from ...sql.dml import UpdateBase
+    from ...sql.elements import ClauseElement
+    from ...sql.selectable import ForUpdateParameter
+    from ...sql.selectable import TypedReturnsRows
+
+_AsyncSessionBind = Union["AsyncEngine", "AsyncConnection"]
+
+_P = ParamSpec("_P")
+_T = TypeVar("_T", bound=Any)
+
+
+_EXECUTE_OPTIONS = util.immutabledict({"prebuffer_rows": True})
+_STREAM_OPTIONS = util.immutabledict({"stream_results": True})
+
+
+class AsyncAttrs:
+    """Mixin class which provides an awaitable accessor for all attributes.
+
+    E.g.::
+
+        from __future__ import annotations
+
+        from typing import List
+
+        from sqlalchemy import ForeignKey
+        from sqlalchemy import func
+        from sqlalchemy.ext.asyncio import AsyncAttrs
+        from sqlalchemy.orm import DeclarativeBase
+        from sqlalchemy.orm import Mapped
+        from sqlalchemy.orm import mapped_column
+        from sqlalchemy.orm import relationship
+
+
+        class Base(AsyncAttrs, DeclarativeBase):
+            pass
+
+
+        class A(Base):
+            __tablename__ = "a"
+
+            id: Mapped[int] = mapped_column(primary_key=True)
+            data: Mapped[str]
+            bs: Mapped[List[B]] = relationship()
+
+
+        class B(Base):
+            __tablename__ = "b"
+            id: Mapped[int] = mapped_column(primary_key=True)
+            a_id: Mapped[int] = mapped_column(ForeignKey("a.id"))
+            data: Mapped[str]
+
+    In the above example, the :class:`_asyncio.AsyncAttrs` mixin is applied to
+    the declarative ``Base`` class where it takes effect for all subclasses.
+    This mixin adds a single new attribute
+    :attr:`_asyncio.AsyncAttrs.awaitable_attrs` to all classes, which will
+    yield the value of any attribute as an awaitable. This allows attributes
+    which may be subject to lazy loading or deferred / unexpiry loading to be
+    accessed such that IO can still be emitted::
+
+        a1 = (await async_session.scalars(select(A).where(A.id == 5))).one()
+
+        # use the lazy loader on ``a1.bs`` via the ``.awaitable_attrs``
+        # interface, so that it may be awaited
+        for b1 in await a1.awaitable_attrs.bs:
+            print(b1)
+
+    The :attr:`_asyncio.AsyncAttrs.awaitable_attrs` performs a call against the
+    attribute that is approximately equivalent to using the
+    :meth:`_asyncio.AsyncSession.run_sync` method, e.g.::
+
+        for b1 in await async_session.run_sync(lambda sess: a1.bs):
+            print(b1)
+
+    .. versionadded:: 2.0.13
+
+    .. seealso::
+
+        :ref:`asyncio_orm_avoid_lazyloads`
+
+    """
+
+    class _AsyncAttrGetitem:
+        __slots__ = "_instance"
+
+        def __init__(self, _instance: Any):
+            self._instance = _instance
+
+        def __getattr__(self, name: str) -> Awaitable[Any]:
+            return greenlet_spawn(getattr, self._instance, name)
+
+    @property
+    def awaitable_attrs(self) -> AsyncAttrs._AsyncAttrGetitem:
+        """provide a namespace of all attributes on this object wrapped
+        as awaitables.
+
+        e.g.::
+
+
+            a1 = (await async_session.scalars(select(A).where(A.id == 5))).one()
+
+            some_attribute = await a1.awaitable_attrs.some_deferred_attribute
+            some_collection = await a1.awaitable_attrs.some_collection
+
+        """  # noqa: E501
+
+        return AsyncAttrs._AsyncAttrGetitem(self)
+
+
+@util.create_proxy_methods(
+    Session,
+    ":class:`_orm.Session`",
+    ":class:`_asyncio.AsyncSession`",
+    classmethods=["object_session", "identity_key"],
+    methods=[
+        "__contains__",
+        "__iter__",
+        "add",
+        "add_all",
+        "expire",
+        "expire_all",
+        "expunge",
+        "expunge_all",
+        "is_modified",
+        "in_transaction",
+        "in_nested_transaction",
+    ],
+    attributes=[
+        "dirty",
+        "deleted",
+        "new",
+        "identity_map",
+        "is_active",
+        "autoflush",
+        "no_autoflush",
+        "info",
+    ],
+)
+class AsyncSession(ReversibleProxy[Session]):
+    """Asyncio version of :class:`_orm.Session`.
+
+    The :class:`_asyncio.AsyncSession` is a proxy for a traditional
+    :class:`_orm.Session` instance.
+
+    The :class:`_asyncio.AsyncSession` is **not safe for use in concurrent
+    tasks.**.  See :ref:`session_faq_threadsafe` for background.
+
+    .. versionadded:: 1.4
+
+    To use an :class:`_asyncio.AsyncSession` with custom :class:`_orm.Session`
+    implementations, see the
+    :paramref:`_asyncio.AsyncSession.sync_session_class` parameter.
+
+
+    """
+
+    _is_asyncio = True
+
+    dispatch: dispatcher[Session]
+
+    def __init__(
+        self,
+        bind: Optional[_AsyncSessionBind] = None,
+        *,
+        binds: Optional[Dict[_SessionBindKey, _AsyncSessionBind]] = None,
+        sync_session_class: Optional[Type[Session]] = None,
+        **kw: Any,
+    ):
+        r"""Construct a new :class:`_asyncio.AsyncSession`.
+
+        All parameters other than ``sync_session_class`` are passed to the
+        ``sync_session_class`` callable directly to instantiate a new
+        :class:`_orm.Session`. Refer to :meth:`_orm.Session.__init__` for
+        parameter documentation.
+
+        :param sync_session_class:
+          A :class:`_orm.Session` subclass or other callable which will be used
+          to construct the :class:`_orm.Session` which will be proxied. This
+          parameter may be used to provide custom :class:`_orm.Session`
+          subclasses. Defaults to the
+          :attr:`_asyncio.AsyncSession.sync_session_class` class-level
+          attribute.
+
+          .. versionadded:: 1.4.24
+
+        """
+        sync_bind = sync_binds = None
+
+        if bind:
+            self.bind = bind
+            sync_bind = engine._get_sync_engine_or_connection(bind)
+
+        if binds:
+            self.binds = binds
+            sync_binds = {
+                key: engine._get_sync_engine_or_connection(b)
+                for key, b in binds.items()
+            }
+
+        if sync_session_class:
+            self.sync_session_class = sync_session_class
+
+        self.sync_session = self._proxied = self._assign_proxied(
+            self.sync_session_class(bind=sync_bind, binds=sync_binds, **kw)
+        )
+
+    sync_session_class: Type[Session] = Session
+    """The class or callable that provides the
+    underlying :class:`_orm.Session` instance for a particular
+    :class:`_asyncio.AsyncSession`.
+
+    At the class level, this attribute is the default value for the
+    :paramref:`_asyncio.AsyncSession.sync_session_class` parameter. Custom
+    subclasses of :class:`_asyncio.AsyncSession` can override this.
+
+    At the instance level, this attribute indicates the current class or
+    callable that was used to provide the :class:`_orm.Session` instance for
+    this :class:`_asyncio.AsyncSession` instance.
+
+    .. versionadded:: 1.4.24
+
+    """
+
+    sync_session: Session
+    """Reference to the underlying :class:`_orm.Session` this
+    :class:`_asyncio.AsyncSession` proxies requests towards.
+
+    This instance can be used as an event target.
+
+    .. seealso::
+
+        :ref:`asyncio_events`
+
+    """
+
+    @classmethod
+    def _no_async_engine_events(cls) -> NoReturn:
+        raise NotImplementedError(
+            "asynchronous events are not implemented at this time.  Apply "
+            "synchronous listeners to the AsyncSession.sync_session."
+        )
+
+    async def refresh(
+        self,
+        instance: object,
+        attribute_names: Optional[Iterable[str]] = None,
+        with_for_update: ForUpdateParameter = None,
+    ) -> None:
+        """Expire and refresh the attributes on the given instance.
+
+        A query will be issued to the database and all attributes will be
+        refreshed with their current database value.
+
+        This is the async version of the :meth:`_orm.Session.refresh` method.
+        See that method for a complete description of all options.
+
+        .. seealso::
+
+            :meth:`_orm.Session.refresh` - main documentation for refresh
+
+        """
+
+        await greenlet_spawn(
+            self.sync_session.refresh,
+            instance,
+            attribute_names=attribute_names,
+            with_for_update=with_for_update,
+        )
+
+    async def run_sync(
+        self,
+        fn: Callable[Concatenate[Session, _P], _T],
+        *arg: _P.args,
+        **kw: _P.kwargs,
+    ) -> _T:
+        '''Invoke the given synchronous (i.e. not async) callable,
+        passing a synchronous-style :class:`_orm.Session` as the first
+        argument.
+
+        This method allows traditional synchronous SQLAlchemy functions to
+        run within the context of an asyncio application.
+
+        E.g.::
+
+            def some_business_method(session: Session, param: str) -> str:
+                """A synchronous function that does not require awaiting
+
+                :param session: a SQLAlchemy Session, used synchronously
+
+                :return: an optional return value is supported
+
+                """
+                session.add(MyObject(param=param))
+                session.flush()
+                return "success"
+
+
+            async def do_something_async(async_engine: AsyncEngine) -> None:
+                """an async function that uses awaiting"""
+
+                with AsyncSession(async_engine) as async_session:
+                    # run some_business_method() with a sync-style
+                    # Session, proxied into an awaitable
+                    return_code = await async_session.run_sync(
+                        some_business_method, param="param1"
+                    )
+                    print(return_code)
+
+        This method maintains the asyncio event loop all the way through
+        to the database connection by running the given callable in a
+        specially instrumented greenlet.
+
+        .. tip::
+
+            The provided callable is invoked inline within the asyncio event
+            loop, and will block on traditional IO calls.  IO within this
+            callable should only call into SQLAlchemy's asyncio database
+            APIs which will be properly adapted to the greenlet context.
+
+        .. seealso::
+
+            :class:`.AsyncAttrs`  - a mixin for ORM mapped classes that provides
+            a similar feature more succinctly on a per-attribute basis
+
+            :meth:`.AsyncConnection.run_sync`
+
+            :ref:`session_run_sync`
+        '''  # noqa: E501
+
+        return await greenlet_spawn(
+            fn, self.sync_session, *arg, _require_await=False, **kw
+        )
+
+    @overload
+    async def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[_T]: ...
+
+    @overload
+    async def execute(
+        self,
+        statement: UpdateBase,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> CursorResult[Any]: ...
+
+    @overload
+    async def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]: ...
+
+    async def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Result[Any]:
+        """Execute a statement and return a buffered
+        :class:`_engine.Result` object.
+
+        .. seealso::
+
+            :meth:`_orm.Session.execute` - main documentation for execute
+
+        """
+
+        if execution_options:
+            execution_options = util.immutabledict(execution_options).union(
+                _EXECUTE_OPTIONS
+            )
+        else:
+            execution_options = _EXECUTE_OPTIONS
+
+        result = await greenlet_spawn(
+            self.sync_session.execute,
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+        return await _ensure_sync_result(result, self.execute)
+
+    @overload
+    async def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Optional[_T]: ...
+
+    @overload
+    async def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any: ...
+
+    async def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any:
+        """Execute a statement and return a scalar result.
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalar` - main documentation for scalar
+
+        """
+
+        if execution_options:
+            execution_options = util.immutabledict(execution_options).union(
+                _EXECUTE_OPTIONS
+            )
+        else:
+            execution_options = _EXECUTE_OPTIONS
+
+        return await greenlet_spawn(
+            self.sync_session.scalar,
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @overload
+    async def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    async def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]: ...
+
+    async def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]:
+        """Execute a statement and return scalar results.
+
+        :return: a :class:`_result.ScalarResult` object
+
+        .. versionadded:: 1.4.24 Added :meth:`_asyncio.AsyncSession.scalars`
+
+        .. versionadded:: 1.4.26 Added
+           :meth:`_asyncio.async_scoped_session.scalars`
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalars` - main documentation for scalars
+
+            :meth:`_asyncio.AsyncSession.stream_scalars` - streaming version
+
+        """
+
+        result = await self.execute(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+        return result.scalars()
+
+    async def get(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+    ) -> Union[_O, None]:
+        """Return an instance based on the given primary key identifier,
+        or ``None`` if not found.
+
+        .. seealso::
+
+            :meth:`_orm.Session.get` - main documentation for get
+
+
+        """
+
+        return await greenlet_spawn(
+            cast("Callable[..., _O]", self.sync_session.get),
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+        )
+
+    async def get_one(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+    ) -> _O:
+        """Return an instance based on the given primary key identifier,
+        or raise an exception if not found.
+
+        Raises ``sqlalchemy.orm.exc.NoResultFound`` if the query selects
+        no rows.
+
+        ..versionadded: 2.0.22
+
+        .. seealso::
+
+            :meth:`_orm.Session.get_one` - main documentation for get_one
+
+        """
+
+        return await greenlet_spawn(
+            cast("Callable[..., _O]", self.sync_session.get_one),
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+        )
+
+    @overload
+    async def stream(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[_T]: ...
+
+    @overload
+    async def stream(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[Any]: ...
+
+    async def stream(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncResult[Any]:
+        """Execute a statement and return a streaming
+        :class:`_asyncio.AsyncResult` object.
+
+        """
+
+        if execution_options:
+            execution_options = util.immutabledict(execution_options).union(
+                _STREAM_OPTIONS
+            )
+        else:
+            execution_options = _STREAM_OPTIONS
+
+        result = await greenlet_spawn(
+            self.sync_session.execute,
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+        return AsyncResult(result)
+
+    @overload
+    async def stream_scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[_T]: ...
+
+    @overload
+    async def stream_scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[Any]: ...
+
+    async def stream_scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> AsyncScalarResult[Any]:
+        """Execute a statement and return a stream of scalar results.
+
+        :return: an :class:`_asyncio.AsyncScalarResult` object
+
+        .. versionadded:: 1.4.24
+
+        .. seealso::
+
+            :meth:`_orm.Session.scalars` - main documentation for scalars
+
+            :meth:`_asyncio.AsyncSession.scalars` - non streaming version
+
+        """
+
+        result = await self.stream(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+        return result.scalars()
+
+    async def delete(self, instance: object) -> None:
+        """Mark an instance as deleted.
+
+        The database delete operation occurs upon ``flush()``.
+
+        As this operation may need to cascade along unloaded relationships,
+        it is awaitable to allow for those queries to take place.
+
+        .. seealso::
+
+            :meth:`_orm.Session.delete` - main documentation for delete
+
+        """
+        await greenlet_spawn(self.sync_session.delete, instance)
+
+    async def merge(
+        self,
+        instance: _O,
+        *,
+        load: bool = True,
+        options: Optional[Sequence[ORMOption]] = None,
+    ) -> _O:
+        """Copy the state of a given instance into a corresponding instance
+        within this :class:`_asyncio.AsyncSession`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.merge` - main documentation for merge
+
+        """
+        return await greenlet_spawn(
+            self.sync_session.merge, instance, load=load, options=options
+        )
+
+    async def flush(self, objects: Optional[Sequence[Any]] = None) -> None:
+        """Flush all the object changes to the database.
+
+        .. seealso::
+
+            :meth:`_orm.Session.flush` - main documentation for flush
+
+        """
+        await greenlet_spawn(self.sync_session.flush, objects=objects)
+
+    def get_transaction(self) -> Optional[AsyncSessionTransaction]:
+        """Return the current root transaction in progress, if any.
+
+        :return: an :class:`_asyncio.AsyncSessionTransaction` object, or
+         ``None``.
+
+        .. versionadded:: 1.4.18
+
+        """
+        trans = self.sync_session.get_transaction()
+        if trans is not None:
+            return AsyncSessionTransaction._retrieve_proxy_for_target(trans)
+        else:
+            return None
+
+    def get_nested_transaction(self) -> Optional[AsyncSessionTransaction]:
+        """Return the current nested transaction in progress, if any.
+
+        :return: an :class:`_asyncio.AsyncSessionTransaction` object, or
+         ``None``.
+
+        .. versionadded:: 1.4.18
+
+        """
+
+        trans = self.sync_session.get_nested_transaction()
+        if trans is not None:
+            return AsyncSessionTransaction._retrieve_proxy_for_target(trans)
+        else:
+            return None
+
+    def get_bind(
+        self,
+        mapper: Optional[_EntityBindKey[_O]] = None,
+        clause: Optional[ClauseElement] = None,
+        bind: Optional[_SessionBind] = None,
+        **kw: Any,
+    ) -> Union[Engine, Connection]:
+        """Return a "bind" to which the synchronous proxied :class:`_orm.Session`
+        is bound.
+
+        Unlike the :meth:`_orm.Session.get_bind` method, this method is
+        currently **not** used by this :class:`.AsyncSession` in any way
+        in order to resolve engines for requests.
+
+        .. note::
+
+            This method proxies directly to the :meth:`_orm.Session.get_bind`
+            method, however is currently **not** useful as an override target,
+            in contrast to that of the :meth:`_orm.Session.get_bind` method.
+            The example below illustrates how to implement custom
+            :meth:`_orm.Session.get_bind` schemes that work with
+            :class:`.AsyncSession` and :class:`.AsyncEngine`.
+
+        The pattern introduced at :ref:`session_custom_partitioning`
+        illustrates how to apply a custom bind-lookup scheme to a
+        :class:`_orm.Session` given a set of :class:`_engine.Engine` objects.
+        To apply a corresponding :meth:`_orm.Session.get_bind` implementation
+        for use with a :class:`.AsyncSession` and :class:`.AsyncEngine`
+        objects, continue to subclass :class:`_orm.Session` and apply it to
+        :class:`.AsyncSession` using
+        :paramref:`.AsyncSession.sync_session_class`. The inner method must
+        continue to return :class:`_engine.Engine` instances, which can be
+        acquired from a :class:`_asyncio.AsyncEngine` using the
+        :attr:`_asyncio.AsyncEngine.sync_engine` attribute::
+
+            # using example from "Custom Vertical Partitioning"
+
+
+            import random
+
+            from sqlalchemy.ext.asyncio import AsyncSession
+            from sqlalchemy.ext.asyncio import create_async_engine
+            from sqlalchemy.ext.asyncio import async_sessionmaker
+            from sqlalchemy.orm import Session
+
+            # construct async engines w/ async drivers
+            engines = {
+                "leader": create_async_engine("sqlite+aiosqlite:///leader.db"),
+                "other": create_async_engine("sqlite+aiosqlite:///other.db"),
+                "follower1": create_async_engine("sqlite+aiosqlite:///follower1.db"),
+                "follower2": create_async_engine("sqlite+aiosqlite:///follower2.db"),
+            }
+
+
+            class RoutingSession(Session):
+                def get_bind(self, mapper=None, clause=None, **kw):
+                    # within get_bind(), return sync engines
+                    if mapper and issubclass(mapper.class_, MyOtherClass):
+                        return engines["other"].sync_engine
+                    elif self._flushing or isinstance(clause, (Update, Delete)):
+                        return engines["leader"].sync_engine
+                    else:
+                        return engines[
+                            random.choice(["follower1", "follower2"])
+                        ].sync_engine
+
+
+            # apply to AsyncSession using sync_session_class
+            AsyncSessionMaker = async_sessionmaker(sync_session_class=RoutingSession)
+
+        The :meth:`_orm.Session.get_bind` method is called in a non-asyncio,
+        implicitly non-blocking context in the same manner as ORM event hooks
+        and functions that are invoked via :meth:`.AsyncSession.run_sync`, so
+        routines that wish to run SQL commands inside of
+        :meth:`_orm.Session.get_bind` can continue to do so using
+        blocking-style code, which will be translated to implicitly async calls
+        at the point of invoking IO on the database drivers.
+
+        """  # noqa: E501
+
+        return self.sync_session.get_bind(
+            mapper=mapper, clause=clause, bind=bind, **kw
+        )
+
+    async def connection(
+        self,
+        bind_arguments: Optional[_BindArguments] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+        **kw: Any,
+    ) -> AsyncConnection:
+        r"""Return a :class:`_asyncio.AsyncConnection` object corresponding to
+        this :class:`.Session` object's transactional state.
+
+        This method may also be used to establish execution options for the
+        database connection used by the current transaction.
+
+        .. versionadded:: 1.4.24  Added \**kw arguments which are passed
+           through to the underlying :meth:`_orm.Session.connection` method.
+
+        .. seealso::
+
+            :meth:`_orm.Session.connection` - main documentation for
+            "connection"
+
+        """
+
+        sync_connection = await greenlet_spawn(
+            self.sync_session.connection,
+            bind_arguments=bind_arguments,
+            execution_options=execution_options,
+            **kw,
+        )
+        return engine.AsyncConnection._retrieve_proxy_for_target(
+            sync_connection
+        )
+
+    def begin(self) -> AsyncSessionTransaction:
+        """Return an :class:`_asyncio.AsyncSessionTransaction` object.
+
+        The underlying :class:`_orm.Session` will perform the
+        "begin" action when the :class:`_asyncio.AsyncSessionTransaction`
+        object is entered::
+
+            async with async_session.begin():
+                ...  # ORM transaction is begun
+
+        Note that database IO will not normally occur when the session-level
+        transaction is begun, as database transactions begin on an
+        on-demand basis.  However, the begin block is async to accommodate
+        for a :meth:`_orm.SessionEvents.after_transaction_create`
+        event hook that may perform IO.
+
+        For a general description of ORM begin, see
+        :meth:`_orm.Session.begin`.
+
+        """
+
+        return AsyncSessionTransaction(self)
+
+    def begin_nested(self) -> AsyncSessionTransaction:
+        """Return an :class:`_asyncio.AsyncSessionTransaction` object
+        which will begin a "nested" transaction, e.g. SAVEPOINT.
+
+        Behavior is the same as that of :meth:`_asyncio.AsyncSession.begin`.
+
+        For a general description of ORM begin nested, see
+        :meth:`_orm.Session.begin_nested`.
+
+        .. seealso::
+
+            :ref:`aiosqlite_serializable` - special workarounds required
+            with the SQLite asyncio driver in order for SAVEPOINT to work
+            correctly.
+
+        """
+
+        return AsyncSessionTransaction(self, nested=True)
+
+    async def rollback(self) -> None:
+        """Rollback the current transaction in progress.
+
+        .. seealso::
+
+            :meth:`_orm.Session.rollback` - main documentation for
+            "rollback"
+        """
+        await greenlet_spawn(self.sync_session.rollback)
+
+    async def commit(self) -> None:
+        """Commit the current transaction in progress.
+
+        .. seealso::
+
+            :meth:`_orm.Session.commit` - main documentation for
+            "commit"
+        """
+        await greenlet_spawn(self.sync_session.commit)
+
+    async def close(self) -> None:
+        """Close out the transactional resources and ORM objects used by this
+        :class:`_asyncio.AsyncSession`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.close` - main documentation for
+            "close"
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_asyncio.AsyncSession.close` and
+            :meth:`_asyncio.AsyncSession.reset`.
+
+        """
+        await greenlet_spawn(self.sync_session.close)
+
+    async def reset(self) -> None:
+        """Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`, resetting the session to its initial state.
+
+        .. versionadded:: 2.0.22
+
+        .. seealso::
+
+            :meth:`_orm.Session.reset` - main documentation for
+            "reset"
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_asyncio.AsyncSession.close` and
+            :meth:`_asyncio.AsyncSession.reset`.
+
+        """
+        await greenlet_spawn(self.sync_session.reset)
+
+    async def aclose(self) -> None:
+        """A synonym for :meth:`_asyncio.AsyncSession.close`.
+
+        The :meth:`_asyncio.AsyncSession.aclose` name is specifically
+        to support the Python standard library ``@contextlib.aclosing``
+        context manager function.
+
+        .. versionadded:: 2.0.20
+
+        """
+        await self.close()
+
+    async def invalidate(self) -> None:
+        """Close this Session, using connection invalidation.
+
+        For a complete description, see :meth:`_orm.Session.invalidate`.
+        """
+        await greenlet_spawn(self.sync_session.invalidate)
+
+    @classmethod
+    @util.deprecated(
+        "2.0",
+        "The :meth:`.AsyncSession.close_all` method is deprecated and will be "
+        "removed in a future release.  Please refer to "
+        ":func:`_asyncio.close_all_sessions`.",
+    )
+    async def close_all(cls) -> None:
+        """Close all :class:`_asyncio.AsyncSession` sessions."""
+        await close_all_sessions()
+
+    async def __aenter__(self: _AS) -> _AS:
+        return self
+
+    async def __aexit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        task = asyncio.create_task(self.close())
+        await asyncio.shield(task)
+
+    def _maker_context_manager(self: _AS) -> _AsyncSessionContextManager[_AS]:
+        return _AsyncSessionContextManager(self)
+
+    # START PROXY METHODS AsyncSession
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_proxy_methods.py
+
+    def __contains__(self, instance: object) -> bool:
+        r"""Return True if the instance is associated with this session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        The instance may be pending or persistent within the Session for a
+        result of True.
+
+
+        """  # noqa: E501
+
+        return self._proxied.__contains__(instance)
+
+    def __iter__(self) -> Iterator[object]:
+        r"""Iterate over all pending or persistent instances within this
+        Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.__iter__()
+
+    def add(self, instance: object, _warn: bool = True) -> None:
+        r"""Place an object into this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        Objects that are in the :term:`transient` state when passed to the
+        :meth:`_orm.Session.add` method will move to the
+        :term:`pending` state, until the next flush, at which point they
+        will move to the :term:`persistent` state.
+
+        Objects that are in the :term:`detached` state when passed to the
+        :meth:`_orm.Session.add` method will move to the :term:`persistent`
+        state directly.
+
+        If the transaction used by the :class:`_orm.Session` is rolled back,
+        objects which were transient when they were passed to
+        :meth:`_orm.Session.add` will be moved back to the
+        :term:`transient` state, and will no longer be present within this
+        :class:`_orm.Session`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add_all`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+        """  # noqa: E501
+
+        return self._proxied.add(instance, _warn=_warn)
+
+    def add_all(self, instances: Iterable[object]) -> None:
+        r"""Add the given collection of instances to this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        See the documentation for :meth:`_orm.Session.add` for a general
+        behavioral description.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+        """  # noqa: E501
+
+        return self._proxied.add_all(instances)
+
+    def expire(
+        self, instance: object, attribute_names: Optional[Iterable[str]] = None
+    ) -> None:
+        r"""Expire the attributes on an instance.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        Marks the attributes of an instance as out of date. When an expired
+        attribute is next accessed, a query will be issued to the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire all objects in the :class:`.Session` simultaneously,
+        use :meth:`Session.expire_all`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire` only makes sense for the specific
+        case that a non-ORM SQL statement was emitted in the current
+        transaction.
+
+        :param instance: The instance to be refreshed.
+        :param attribute_names: optional list of string attribute names
+          indicating a subset of attributes to be expired.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire(instance, attribute_names=attribute_names)
+
+    def expire_all(self) -> None:
+        r"""Expires all persistent instances within this Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        When any attributes on a persistent instance is next accessed,
+        a query will be issued using the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire individual objects and individual attributes
+        on those objects, use :meth:`Session.expire`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire_all` is not usually needed,
+        assuming the transaction is isolated.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire_all()
+
+    def expunge(self, instance: object) -> None:
+        r"""Remove the `instance` from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This will free all internal references to the instance.  Cascading
+        will be applied according to the *expunge* cascade rule.
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge(instance)
+
+    def expunge_all(self) -> None:
+        r"""Remove all object instances from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is equivalent to calling ``expunge(obj)`` on all objects in this
+        ``Session``.
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge_all()
+
+    def is_modified(
+        self, instance: object, include_collections: bool = True
+    ) -> bool:
+        r"""Return ``True`` if the given instance has locally
+        modified attributes.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This method retrieves the history for each instrumented
+        attribute on the instance and performs a comparison of the current
+        value to its previously flushed or committed value, if any.
+
+        It is in effect a more expensive and accurate
+        version of checking for the given instance in the
+        :attr:`.Session.dirty` collection; a full test for
+        each attribute's net "dirty" status is performed.
+
+        E.g.::
+
+            return session.is_modified(someobject)
+
+        A few caveats to this method apply:
+
+        * Instances present in the :attr:`.Session.dirty` collection may
+          report ``False`` when tested with this method.  This is because
+          the object may have received change events via attribute mutation,
+          thus placing it in :attr:`.Session.dirty`, but ultimately the state
+          is the same as that loaded from the database, resulting in no net
+          change here.
+        * Scalar attributes may not have recorded the previously set
+          value when a new value was applied, if the attribute was not loaded,
+          or was expired, at the time the new value was received - in these
+          cases, the attribute is assumed to have a change, even if there is
+          ultimately no net change against its database value. SQLAlchemy in
+          most cases does not need the "old" value when a set event occurs, so
+          it skips the expense of a SQL call if the old value isn't present,
+          based on the assumption that an UPDATE of the scalar value is
+          usually needed, and in those few cases where it isn't, is less
+          expensive on average than issuing a defensive SELECT.
+
+          The "old" value is fetched unconditionally upon set only if the
+          attribute container has the ``active_history`` flag set to ``True``.
+          This flag is set typically for primary key attributes and scalar
+          object references that are not a simple many-to-one.  To set this
+          flag for any arbitrary mapped column, use the ``active_history``
+          argument with :func:`.column_property`.
+
+        :param instance: mapped instance to be tested for pending changes.
+        :param include_collections: Indicates if multivalued collections
+         should be included in the operation.  Setting this to ``False`` is a
+         way to detect only local-column based properties (i.e. scalar columns
+         or many-to-one foreign keys) that would result in an UPDATE for this
+         instance upon flush.
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_modified(
+            instance, include_collections=include_collections
+        )
+
+    def in_transaction(self) -> bool:
+        r"""Return True if this :class:`_orm.Session` has begun a transaction.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_orm.Session.is_active`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.in_transaction()
+
+    def in_nested_transaction(self) -> bool:
+        r"""Return True if this :class:`_orm.Session` has begun a nested
+        transaction, e.g. SAVEPOINT.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. versionadded:: 1.4
+
+
+        """  # noqa: E501
+
+        return self._proxied.in_nested_transaction()
+
+    @property
+    def dirty(self) -> Any:
+        r"""The set of all persistent instances considered dirty.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        E.g.::
+
+            some_mapped_object in session.dirty
+
+        Instances are considered dirty when they were modified but not
+        deleted.
+
+        Note that this 'dirty' calculation is 'optimistic'; most
+        attribute-setting or collection modification operations will
+        mark an instance as 'dirty' and place it in this set, even if
+        there is no net change to the attribute's value.  At flush
+        time, the value of each attribute is compared to its
+        previously saved value, and if there's no net change, no SQL
+        operation will occur (this is a more expensive operation so
+        it's only done at flush time).
+
+        To check if an instance has actionable net changes to its
+        attributes, use the :meth:`.Session.is_modified` method.
+
+
+        """  # noqa: E501
+
+        return self._proxied.dirty
+
+    @property
+    def deleted(self) -> Any:
+        r"""The set of all instances marked as 'deleted' within this ``Session``
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        """  # noqa: E501
+
+        return self._proxied.deleted
+
+    @property
+    def new(self) -> Any:
+        r"""The set of all instances marked as 'new' within this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        """  # noqa: E501
+
+        return self._proxied.new
+
+    @property
+    def identity_map(self) -> IdentityMap:
+        r"""Proxy for the :attr:`_orm.Session.identity_map` attribute
+        on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        """  # noqa: E501
+
+        return self._proxied.identity_map
+
+    @identity_map.setter
+    def identity_map(self, attr: IdentityMap) -> None:
+        self._proxied.identity_map = attr
+
+    @property
+    def is_active(self) -> Any:
+        r"""True if this :class:`.Session` not in "partial rollback" state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        .. versionchanged:: 1.4 The :class:`_orm.Session` no longer begins
+           a new transaction immediately, so this attribute will be False
+           when the :class:`_orm.Session` is first instantiated.
+
+        "partial rollback" state typically indicates that the flush process
+        of the :class:`_orm.Session` has failed, and that the
+        :meth:`_orm.Session.rollback` method must be emitted in order to
+        fully roll back the transaction.
+
+        If this :class:`_orm.Session` is not in a transaction at all, the
+        :class:`_orm.Session` will autobegin when it is first used, so in this
+        case :attr:`_orm.Session.is_active` will return True.
+
+        Otherwise, if this :class:`_orm.Session` is within a transaction,
+        and that transaction has not been rolled back internally, the
+        :attr:`_orm.Session.is_active` will also return True.
+
+        .. seealso::
+
+            :ref:`faq_session_rollback`
+
+            :meth:`_orm.Session.in_transaction`
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_active
+
+    @property
+    def autoflush(self) -> bool:
+        r"""Proxy for the :attr:`_orm.Session.autoflush` attribute
+        on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        """  # noqa: E501
+
+        return self._proxied.autoflush
+
+    @autoflush.setter
+    def autoflush(self, attr: bool) -> None:
+        self._proxied.autoflush = attr
+
+    @property
+    def no_autoflush(self) -> Any:
+        r"""Return a context manager that disables autoflush.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        e.g.::
+
+            with session.no_autoflush:
+
+                some_object = SomeClass()
+                session.add(some_object)
+                # won't autoflush
+                some_object.related_thing = session.query(SomeRelated).first()
+
+        Operations that proceed within the ``with:`` block
+        will not be subject to flushes occurring upon query
+        access.  This is useful when initializing a series
+        of objects which involve existing database queries,
+        where the uncompleted object should not yet be flushed.
+
+
+        """  # noqa: E501
+
+        return self._proxied.no_autoflush
+
+    @property
+    def info(self) -> Any:
+        r"""A user-modifiable dictionary.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_asyncio.AsyncSession` class.
+
+        The initial value of this dictionary can be populated using the
+        ``info`` argument to the :class:`.Session` constructor or
+        :class:`.sessionmaker` constructor or factory methods.  The dictionary
+        here is always local to this :class:`.Session` and can be modified
+        independently of all other :class:`.Session` objects.
+
+
+        """  # noqa: E501
+
+        return self._proxied.info
+
+    @classmethod
+    def object_session(cls, instance: object) -> Optional[Session]:
+        r"""Return the :class:`.Session` to which an object belongs.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is an alias of :func:`.object_session`.
+
+
+        """  # noqa: E501
+
+        return Session.object_session(instance)
+
+    @classmethod
+    def identity_key(
+        cls,
+        class_: Optional[Type[Any]] = None,
+        ident: Union[Any, Tuple[Any, ...]] = None,
+        *,
+        instance: Optional[Any] = None,
+        row: Optional[Union[Row[Any], RowMapping]] = None,
+        identity_token: Optional[Any] = None,
+    ) -> _IdentityKeyType[Any]:
+        r"""Return an identity key.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_asyncio.AsyncSession` class.
+
+        This is an alias of :func:`.util.identity_key`.
+
+
+        """  # noqa: E501
+
+        return Session.identity_key(
+            class_=class_,
+            ident=ident,
+            instance=instance,
+            row=row,
+            identity_token=identity_token,
+        )
+
+    # END PROXY METHODS AsyncSession
+
+
+_AS = TypeVar("_AS", bound="AsyncSession")
+
+
+class async_sessionmaker(Generic[_AS]):
+    """A configurable :class:`.AsyncSession` factory.
+
+    The :class:`.async_sessionmaker` factory works in the same way as the
+    :class:`.sessionmaker` factory, to generate new :class:`.AsyncSession`
+    objects when called, creating them given
+    the configurational arguments established here.
+
+    e.g.::
+
+        from sqlalchemy.ext.asyncio import create_async_engine
+        from sqlalchemy.ext.asyncio import AsyncSession
+        from sqlalchemy.ext.asyncio import async_sessionmaker
+
+
+        async def run_some_sql(
+            async_session: async_sessionmaker[AsyncSession],
+        ) -> None:
+            async with async_session() as session:
+                session.add(SomeObject(data="object"))
+                session.add(SomeOtherObject(name="other object"))
+                await session.commit()
+
+
+        async def main() -> None:
+            # an AsyncEngine, which the AsyncSession will use for connection
+            # resources
+            engine = create_async_engine(
+                "postgresql+asyncpg://scott:tiger@localhost/"
+            )
+
+            # create a reusable factory for new AsyncSession instances
+            async_session = async_sessionmaker(engine)
+
+            await run_some_sql(async_session)
+
+            await engine.dispose()
+
+    The :class:`.async_sessionmaker` is useful so that different parts
+    of a program can create new :class:`.AsyncSession` objects with a
+    fixed configuration established up front.  Note that :class:`.AsyncSession`
+    objects may also be instantiated directly when not using
+    :class:`.async_sessionmaker`.
+
+    .. versionadded:: 2.0  :class:`.async_sessionmaker` provides a
+       :class:`.sessionmaker` class that's dedicated to the
+       :class:`.AsyncSession` object, including pep-484 typing support.
+
+    .. seealso::
+
+        :ref:`asyncio_orm` - shows example use
+
+        :class:`.sessionmaker`  - general overview of the
+         :class:`.sessionmaker` architecture
+
+
+        :ref:`session_getting` - introductory text on creating
+        sessions using :class:`.sessionmaker`.
+
+    """  # noqa E501
+
+    class_: Type[_AS]
+
+    @overload
+    def __init__(
+        self,
+        bind: Optional[_AsyncSessionBind] = ...,
+        *,
+        class_: Type[_AS],
+        autoflush: bool = ...,
+        expire_on_commit: bool = ...,
+        info: Optional[_InfoType] = ...,
+        **kw: Any,
+    ): ...
+
+    @overload
+    def __init__(
+        self: "async_sessionmaker[AsyncSession]",
+        bind: Optional[_AsyncSessionBind] = ...,
+        *,
+        autoflush: bool = ...,
+        expire_on_commit: bool = ...,
+        info: Optional[_InfoType] = ...,
+        **kw: Any,
+    ): ...
+
+    def __init__(
+        self,
+        bind: Optional[_AsyncSessionBind] = None,
+        *,
+        class_: Type[_AS] = AsyncSession,  # type: ignore
+        autoflush: bool = True,
+        expire_on_commit: bool = True,
+        info: Optional[_InfoType] = None,
+        **kw: Any,
+    ):
+        r"""Construct a new :class:`.async_sessionmaker`.
+
+        All arguments here except for ``class_`` correspond to arguments
+        accepted by :class:`.Session` directly. See the
+        :meth:`.AsyncSession.__init__` docstring for more details on
+        parameters.
+
+
+        """
+        kw["bind"] = bind
+        kw["autoflush"] = autoflush
+        kw["expire_on_commit"] = expire_on_commit
+        if info is not None:
+            kw["info"] = info
+        self.kw = kw
+        self.class_ = class_
+
+    def begin(self) -> _AsyncSessionContextManager[_AS]:
+        """Produce a context manager that both provides a new
+        :class:`_orm.AsyncSession` as well as a transaction that commits.
+
+
+        e.g.::
+
+            async def main():
+                Session = async_sessionmaker(some_engine)
+
+                async with Session.begin() as session:
+                    session.add(some_object)
+
+                # commits transaction, closes session
+
+        """
+
+        session = self()
+        return session._maker_context_manager()
+
+    def __call__(self, **local_kw: Any) -> _AS:
+        """Produce a new :class:`.AsyncSession` object using the configuration
+        established in this :class:`.async_sessionmaker`.
+
+        In Python, the ``__call__`` method is invoked on an object when
+        it is "called" in the same way as a function::
+
+            AsyncSession = async_sessionmaker(async_engine, expire_on_commit=False)
+            session = AsyncSession()  # invokes sessionmaker.__call__()
+
+        """  # noqa E501
+        for k, v in self.kw.items():
+            if k == "info" and "info" in local_kw:
+                d = v.copy()
+                d.update(local_kw["info"])
+                local_kw["info"] = d
+            else:
+                local_kw.setdefault(k, v)
+        return self.class_(**local_kw)
+
+    def configure(self, **new_kw: Any) -> None:
+        """(Re)configure the arguments for this async_sessionmaker.
+
+        e.g.::
+
+            AsyncSession = async_sessionmaker(some_engine)
+
+            AsyncSession.configure(bind=create_async_engine("sqlite+aiosqlite://"))
+        """  # noqa E501
+
+        self.kw.update(new_kw)
+
+    def __repr__(self) -> str:
+        return "%s(class_=%r, %s)" % (
+            self.__class__.__name__,
+            self.class_.__name__,
+            ", ".join("%s=%r" % (k, v) for k, v in self.kw.items()),
+        )
+
+
+class _AsyncSessionContextManager(Generic[_AS]):
+    __slots__ = ("async_session", "trans")
+
+    async_session: _AS
+    trans: AsyncSessionTransaction
+
+    def __init__(self, async_session: _AS):
+        self.async_session = async_session
+
+    async def __aenter__(self) -> _AS:
+        self.trans = self.async_session.begin()
+        await self.trans.__aenter__()
+        return self.async_session
+
+    async def __aexit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        async def go() -> None:
+            await self.trans.__aexit__(type_, value, traceback)
+            await self.async_session.__aexit__(type_, value, traceback)
+
+        task = asyncio.create_task(go())
+        await asyncio.shield(task)
+
+
+class AsyncSessionTransaction(
+    ReversibleProxy[SessionTransaction],
+    StartableContext["AsyncSessionTransaction"],
+):
+    """A wrapper for the ORM :class:`_orm.SessionTransaction` object.
+
+    This object is provided so that a transaction-holding object
+    for the :meth:`_asyncio.AsyncSession.begin` may be returned.
+
+    The object supports both explicit calls to
+    :meth:`_asyncio.AsyncSessionTransaction.commit` and
+    :meth:`_asyncio.AsyncSessionTransaction.rollback`, as well as use as an
+    async context manager.
+
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ("session", "sync_transaction", "nested")
+
+    session: AsyncSession
+    sync_transaction: Optional[SessionTransaction]
+
+    def __init__(self, session: AsyncSession, nested: bool = False):
+        self.session = session
+        self.nested = nested
+        self.sync_transaction = None
+
+    @property
+    def is_active(self) -> bool:
+        return (
+            self._sync_transaction() is not None
+            and self._sync_transaction().is_active
+        )
+
+    def _sync_transaction(self) -> SessionTransaction:
+        if not self.sync_transaction:
+            self._raise_for_not_started()
+        return self.sync_transaction
+
+    async def rollback(self) -> None:
+        """Roll back this :class:`_asyncio.AsyncTransaction`."""
+        await greenlet_spawn(self._sync_transaction().rollback)
+
+    async def commit(self) -> None:
+        """Commit this :class:`_asyncio.AsyncTransaction`."""
+
+        await greenlet_spawn(self._sync_transaction().commit)
+
+    async def start(
+        self, is_ctxmanager: bool = False
+    ) -> AsyncSessionTransaction:
+        self.sync_transaction = self._assign_proxied(
+            await greenlet_spawn(
+                self.session.sync_session.begin_nested
+                if self.nested
+                else self.session.sync_session.begin
+            )
+        )
+        if is_ctxmanager:
+            self.sync_transaction.__enter__()
+        return self
+
+    async def __aexit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        await greenlet_spawn(
+            self._sync_transaction().__exit__, type_, value, traceback
+        )
+
+
+def async_object_session(instance: object) -> Optional[AsyncSession]:
+    """Return the :class:`_asyncio.AsyncSession` to which the given instance
+    belongs.
+
+    This function makes use of the sync-API function
+    :class:`_orm.object_session` to retrieve the :class:`_orm.Session` which
+    refers to the given instance, and from there links it to the original
+    :class:`_asyncio.AsyncSession`.
+
+    If the :class:`_asyncio.AsyncSession` has been garbage collected, the
+    return value is ``None``.
+
+    This functionality is also available from the
+    :attr:`_orm.InstanceState.async_session` accessor.
+
+    :param instance: an ORM mapped instance
+    :return: an :class:`_asyncio.AsyncSession` object, or ``None``.
+
+    .. versionadded:: 1.4.18
+
+    """
+
+    session = object_session(instance)
+    if session is not None:
+        return async_session(session)
+    else:
+        return None
+
+
+def async_session(session: Session) -> Optional[AsyncSession]:
+    """Return the :class:`_asyncio.AsyncSession` which is proxying the given
+    :class:`_orm.Session` object, if any.
+
+    :param session: a :class:`_orm.Session` instance.
+    :return: a :class:`_asyncio.AsyncSession` instance, or ``None``.
+
+    .. versionadded:: 1.4.18
+
+    """
+    return AsyncSession._retrieve_proxy_for_target(session, regenerate=False)
+
+
+async def close_all_sessions() -> None:
+    """Close all :class:`_asyncio.AsyncSession` sessions.
+
+    .. versionadded:: 2.0.23
+
+    .. seealso::
+
+        :func:`.session.close_all_sessions`
+
+    """
+    await greenlet_spawn(_sync_close_all_sessions)
+
+
+_instance_state._async_provider = async_session  # type: ignore
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/automap.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/automap.py
new file mode 100644
index 00000000..817f91d2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/automap.py
@@ -0,0 +1,1701 @@
+# ext/automap.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+r"""Define an extension to the :mod:`sqlalchemy.ext.declarative` system
+which automatically generates mapped classes and relationships from a database
+schema, typically though not necessarily one which is reflected.
+
+It is hoped that the :class:`.AutomapBase` system provides a quick
+and modernized solution to the problem that the very famous
+`SQLSoup <https://pypi.org/project/sqlsoup/>`_
+also tries to solve, that of generating a quick and rudimentary object
+model from an existing database on the fly.  By addressing the issue strictly
+at the mapper configuration level, and integrating fully with existing
+Declarative class techniques, :class:`.AutomapBase` seeks to provide
+a well-integrated approach to the issue of expediently auto-generating ad-hoc
+mappings.
+
+.. tip:: The :ref:`automap_toplevel` extension is geared towards a
+   "zero declaration" approach, where a complete ORM model including classes
+   and pre-named relationships can be generated on the fly from a database
+   schema. For applications that still want to use explicit class declarations
+   including explicit relationship definitions in conjunction with reflection
+   of tables, the :class:`.DeferredReflection` class, described at
+   :ref:`orm_declarative_reflected_deferred_reflection`, is a better choice.
+
+.. _automap_basic_use:
+
+Basic Use
+=========
+
+The simplest usage is to reflect an existing database into a new model.
+We create a new :class:`.AutomapBase` class in a similar manner as to how
+we create a declarative base class, using :func:`.automap_base`.
+We then call :meth:`.AutomapBase.prepare` on the resulting base class,
+asking it to reflect the schema and produce mappings::
+
+    from sqlalchemy.ext.automap import automap_base
+    from sqlalchemy.orm import Session
+    from sqlalchemy import create_engine
+
+    Base = automap_base()
+
+    # engine, suppose it has two tables 'user' and 'address' set up
+    engine = create_engine("sqlite:///mydatabase.db")
+
+    # reflect the tables
+    Base.prepare(autoload_with=engine)
+
+    # mapped classes are now created with names by default
+    # matching that of the table name.
+    User = Base.classes.user
+    Address = Base.classes.address
+
+    session = Session(engine)
+
+    # rudimentary relationships are produced
+    session.add(Address(email_address="foo@bar.com", user=User(name="foo")))
+    session.commit()
+
+    # collection-based relationships are by default named
+    # "<classname>_collection"
+    u1 = session.query(User).first()
+    print(u1.address_collection)
+
+Above, calling :meth:`.AutomapBase.prepare` while passing along the
+:paramref:`.AutomapBase.prepare.reflect` parameter indicates that the
+:meth:`_schema.MetaData.reflect`
+method will be called on this declarative base
+classes' :class:`_schema.MetaData` collection; then, each **viable**
+:class:`_schema.Table` within the :class:`_schema.MetaData`
+will get a new mapped class
+generated automatically.  The :class:`_schema.ForeignKeyConstraint`
+objects which
+link the various tables together will be used to produce new, bidirectional
+:func:`_orm.relationship` objects between classes.
+The classes and relationships
+follow along a default naming scheme that we can customize.  At this point,
+our basic mapping consisting of related ``User`` and ``Address`` classes is
+ready to use in the traditional way.
+
+.. note:: By **viable**, we mean that for a table to be mapped, it must
+   specify a primary key.  Additionally, if the table is detected as being
+   a pure association table between two other tables, it will not be directly
+   mapped and will instead be configured as a many-to-many table between
+   the mappings for the two referring tables.
+
+Generating Mappings from an Existing MetaData
+=============================================
+
+We can pass a pre-declared :class:`_schema.MetaData` object to
+:func:`.automap_base`.
+This object can be constructed in any way, including programmatically, from
+a serialized file, or from itself being reflected using
+:meth:`_schema.MetaData.reflect`.
+Below we illustrate a combination of reflection and
+explicit table declaration::
+
+    from sqlalchemy import create_engine, MetaData, Table, Column, ForeignKey
+    from sqlalchemy.ext.automap import automap_base
+
+    engine = create_engine("sqlite:///mydatabase.db")
+
+    # produce our own MetaData object
+    metadata = MetaData()
+
+    # we can reflect it ourselves from a database, using options
+    # such as 'only' to limit what tables we look at...
+    metadata.reflect(engine, only=["user", "address"])
+
+    # ... or just define our own Table objects with it (or combine both)
+    Table(
+        "user_order",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("user_id", ForeignKey("user.id")),
+    )
+
+    # we can then produce a set of mappings from this MetaData.
+    Base = automap_base(metadata=metadata)
+
+    # calling prepare() just sets up mapped classes and relationships.
+    Base.prepare()
+
+    # mapped classes are ready
+    User = Base.classes.user
+    Address = Base.classes.address
+    Order = Base.classes.user_order
+
+.. _automap_by_module:
+
+Generating Mappings from Multiple Schemas
+=========================================
+
+The :meth:`.AutomapBase.prepare` method when used with reflection may reflect
+tables from one schema at a time at most, using the
+:paramref:`.AutomapBase.prepare.schema` parameter to indicate the name of a
+schema to be reflected from. In order to populate the :class:`.AutomapBase`
+with tables from multiple schemas, :meth:`.AutomapBase.prepare` may be invoked
+multiple times, each time passing a different name to the
+:paramref:`.AutomapBase.prepare.schema` parameter. The
+:meth:`.AutomapBase.prepare` method keeps an internal list of
+:class:`_schema.Table` objects that have already been mapped, and will add new
+mappings only for those :class:`_schema.Table` objects that are new since the
+last time :meth:`.AutomapBase.prepare` was run::
+
+    e = create_engine("postgresql://scott:tiger@localhost/test")
+
+    Base.metadata.create_all(e)
+
+    Base = automap_base()
+
+    Base.prepare(e)
+    Base.prepare(e, schema="test_schema")
+    Base.prepare(e, schema="test_schema_2")
+
+.. versionadded:: 2.0  The :meth:`.AutomapBase.prepare` method may be called
+   any number of times; only newly added tables will be mapped
+   on each run.   Previously in version 1.4 and earlier, multiple calls would
+   cause errors as it would attempt to re-map an already mapped class.
+   The previous workaround approach of invoking
+   :meth:`_schema.MetaData.reflect` directly remains available as well.
+
+Automapping same-named tables across multiple schemas
+-----------------------------------------------------
+
+For the common case where multiple schemas may have same-named tables and
+therefore would generate same-named classes, conflicts can be resolved either
+through use of the :paramref:`.AutomapBase.prepare.classname_for_table` hook to
+apply different classnames on a per-schema basis, or by using the
+:paramref:`.AutomapBase.prepare.modulename_for_table` hook, which allows
+disambiguation of same-named classes by changing their effective ``__module__``
+attribute. In the example below, this hook is used to create a ``__module__``
+attribute for all classes that is of the form ``mymodule.<schemaname>``, where
+the schema name ``default`` is used if no schema is present::
+
+    e = create_engine("postgresql://scott:tiger@localhost/test")
+
+    Base.metadata.create_all(e)
+
+
+    def module_name_for_table(cls, tablename, table):
+        if table.schema is not None:
+            return f"mymodule.{table.schema}"
+        else:
+            return f"mymodule.default"
+
+
+    Base = automap_base()
+
+    Base.prepare(e, modulename_for_table=module_name_for_table)
+    Base.prepare(
+        e, schema="test_schema", modulename_for_table=module_name_for_table
+    )
+    Base.prepare(
+        e, schema="test_schema_2", modulename_for_table=module_name_for_table
+    )
+
+The same named-classes are organized into a hierarchical collection available
+at :attr:`.AutomapBase.by_module`.  This collection is traversed using the
+dot-separated name of a particular package/module down into the desired
+class name.
+
+.. note:: When using the :paramref:`.AutomapBase.prepare.modulename_for_table`
+   hook to return a new ``__module__`` that is not ``None``, the class is
+   **not** placed into the :attr:`.AutomapBase.classes` collection; only
+   classes that were not given an explicit modulename are placed here, as the
+   collection cannot represent same-named classes individually.
+
+In the example above, if the database contained a table named ``accounts`` in
+all three of the default schema, the ``test_schema`` schema, and the
+``test_schema_2`` schema, three separate classes will be available as::
+
+    Base.by_module.mymodule.default.accounts
+    Base.by_module.mymodule.test_schema.accounts
+    Base.by_module.mymodule.test_schema_2.accounts
+
+The default module namespace generated for all :class:`.AutomapBase` classes is
+``sqlalchemy.ext.automap``. If no
+:paramref:`.AutomapBase.prepare.modulename_for_table` hook is used, the
+contents of :attr:`.AutomapBase.by_module` will be entirely within the
+``sqlalchemy.ext.automap`` namespace (e.g.
+``MyBase.by_module.sqlalchemy.ext.automap.<classname>``), which would contain
+the same series of classes as what would be seen in
+:attr:`.AutomapBase.classes`. Therefore it's generally only necessary to use
+:attr:`.AutomapBase.by_module` when explicit ``__module__`` conventions are
+present.
+
+.. versionadded: 2.0
+
+    Added the :attr:`.AutomapBase.by_module` collection, which stores
+    classes within a named hierarchy based on dot-separated module names,
+    as well as the :paramref:`.Automap.prepare.modulename_for_table` parameter
+    which allows for custom ``__module__`` schemes for automapped
+    classes.
+
+
+
+Specifying Classes Explicitly
+=============================
+
+.. tip:: If explicit classes are expected to be prominent in an application,
+   consider using :class:`.DeferredReflection` instead.
+
+The :mod:`.sqlalchemy.ext.automap` extension allows classes to be defined
+explicitly, in a way similar to that of the :class:`.DeferredReflection` class.
+Classes that extend from :class:`.AutomapBase` act like regular declarative
+classes, but are not immediately mapped after their construction, and are
+instead mapped when we call :meth:`.AutomapBase.prepare`.  The
+:meth:`.AutomapBase.prepare` method will make use of the classes we've
+established based on the table name we use.  If our schema contains tables
+``user`` and ``address``, we can define one or both of the classes to be used::
+
+    from sqlalchemy.ext.automap import automap_base
+    from sqlalchemy import create_engine
+
+    # automap base
+    Base = automap_base()
+
+
+    # pre-declare User for the 'user' table
+    class User(Base):
+        __tablename__ = "user"
+
+        # override schema elements like Columns
+        user_name = Column("name", String)
+
+        # override relationships too, if desired.
+        # we must use the same name that automap would use for the
+        # relationship, and also must refer to the class name that automap will
+        # generate for "address"
+        address_collection = relationship("address", collection_class=set)
+
+
+    # reflect
+    engine = create_engine("sqlite:///mydatabase.db")
+    Base.prepare(autoload_with=engine)
+
+    # we still have Address generated from the tablename "address",
+    # but User is the same as Base.classes.User now
+
+    Address = Base.classes.address
+
+    u1 = session.query(User).first()
+    print(u1.address_collection)
+
+    # the backref is still there:
+    a1 = session.query(Address).first()
+    print(a1.user)
+
+Above, one of the more intricate details is that we illustrated overriding
+one of the :func:`_orm.relationship` objects that automap would have created.
+To do this, we needed to make sure the names match up with what automap
+would normally generate, in that the relationship name would be
+``User.address_collection`` and the name of the class referred to, from
+automap's perspective, is called ``address``, even though we are referring to
+it as ``Address`` within our usage of this class.
+
+Overriding Naming Schemes
+=========================
+
+:mod:`.sqlalchemy.ext.automap` is tasked with producing mapped classes and
+relationship names based on a schema, which means it has decision points in how
+these names are determined.  These three decision points are provided using
+functions which can be passed to the :meth:`.AutomapBase.prepare` method, and
+are known as :func:`.classname_for_table`,
+:func:`.name_for_scalar_relationship`,
+and :func:`.name_for_collection_relationship`.  Any or all of these
+functions are provided as in the example below, where we use a "camel case"
+scheme for class names and a "pluralizer" for collection names using the
+`Inflect <https://pypi.org/project/inflect>`_ package::
+
+    import re
+    import inflect
+
+
+    def camelize_classname(base, tablename, table):
+        "Produce a 'camelized' class name, e.g."
+        "'words_and_underscores' -> 'WordsAndUnderscores'"
+
+        return str(
+            tablename[0].upper()
+            + re.sub(
+                r"_([a-z])",
+                lambda m: m.group(1).upper(),
+                tablename[1:],
+            )
+        )
+
+
+    _pluralizer = inflect.engine()
+
+
+    def pluralize_collection(base, local_cls, referred_cls, constraint):
+        "Produce an 'uncamelized', 'pluralized' class name, e.g."
+        "'SomeTerm' -> 'some_terms'"
+
+        referred_name = referred_cls.__name__
+        uncamelized = re.sub(
+            r"[A-Z]",
+            lambda m: "_%s" % m.group(0).lower(),
+            referred_name,
+        )[1:]
+        pluralized = _pluralizer.plural(uncamelized)
+        return pluralized
+
+
+    from sqlalchemy.ext.automap import automap_base
+
+    Base = automap_base()
+
+    engine = create_engine("sqlite:///mydatabase.db")
+
+    Base.prepare(
+        autoload_with=engine,
+        classname_for_table=camelize_classname,
+        name_for_collection_relationship=pluralize_collection,
+    )
+
+From the above mapping, we would now have classes ``User`` and ``Address``,
+where the collection from ``User`` to ``Address`` is called
+``User.addresses``::
+
+    User, Address = Base.classes.User, Base.classes.Address
+
+    u1 = User(addresses=[Address(email="foo@bar.com")])
+
+Relationship Detection
+======================
+
+The vast majority of what automap accomplishes is the generation of
+:func:`_orm.relationship` structures based on foreign keys.  The mechanism
+by which this works for many-to-one and one-to-many relationships is as
+follows:
+
+1. A given :class:`_schema.Table`, known to be mapped to a particular class,
+   is examined for :class:`_schema.ForeignKeyConstraint` objects.
+
+2. From each :class:`_schema.ForeignKeyConstraint`, the remote
+   :class:`_schema.Table`
+   object present is matched up to the class to which it is to be mapped,
+   if any, else it is skipped.
+
+3. As the :class:`_schema.ForeignKeyConstraint`
+   we are examining corresponds to a
+   reference from the immediate mapped class,  the relationship will be set up
+   as a many-to-one referring to the referred class; a corresponding
+   one-to-many backref will be created on the referred class referring
+   to this class.
+
+4. If any of the columns that are part of the
+   :class:`_schema.ForeignKeyConstraint`
+   are not nullable (e.g. ``nullable=False``), a
+   :paramref:`_orm.relationship.cascade` keyword argument
+   of ``all, delete-orphan`` will be added to the keyword arguments to
+   be passed to the relationship or backref.  If the
+   :class:`_schema.ForeignKeyConstraint` reports that
+   :paramref:`_schema.ForeignKeyConstraint.ondelete`
+   is set to ``CASCADE`` for a not null or ``SET NULL`` for a nullable
+   set of columns, the option :paramref:`_orm.relationship.passive_deletes`
+   flag is set to ``True`` in the set of relationship keyword arguments.
+   Note that not all backends support reflection of ON DELETE.
+
+5. The names of the relationships are determined using the
+   :paramref:`.AutomapBase.prepare.name_for_scalar_relationship` and
+   :paramref:`.AutomapBase.prepare.name_for_collection_relationship`
+   callable functions.  It is important to note that the default relationship
+   naming derives the name from the **the actual class name**.  If you've
+   given a particular class an explicit name by declaring it, or specified an
+   alternate class naming scheme, that's the name from which the relationship
+   name will be derived.
+
+6. The classes are inspected for an existing mapped property matching these
+   names.  If one is detected on one side, but none on the other side,
+   :class:`.AutomapBase` attempts to create a relationship on the missing side,
+   then uses the :paramref:`_orm.relationship.back_populates`
+   parameter in order to
+   point the new relationship to the other side.
+
+7. In the usual case where no relationship is on either side,
+   :meth:`.AutomapBase.prepare` produces a :func:`_orm.relationship` on the
+   "many-to-one" side and matches it to the other using the
+   :paramref:`_orm.relationship.backref` parameter.
+
+8. Production of the :func:`_orm.relationship` and optionally the
+   :func:`.backref`
+   is handed off to the :paramref:`.AutomapBase.prepare.generate_relationship`
+   function, which can be supplied by the end-user in order to augment
+   the arguments passed to :func:`_orm.relationship` or :func:`.backref` or to
+   make use of custom implementations of these functions.
+
+Custom Relationship Arguments
+-----------------------------
+
+The :paramref:`.AutomapBase.prepare.generate_relationship` hook can be used
+to add parameters to relationships.  For most cases, we can make use of the
+existing :func:`.automap.generate_relationship` function to return
+the object, after augmenting the given keyword dictionary with our own
+arguments.
+
+Below is an illustration of how to send
+:paramref:`_orm.relationship.cascade` and
+:paramref:`_orm.relationship.passive_deletes`
+options along to all one-to-many relationships::
+
+    from sqlalchemy.ext.automap import generate_relationship
+    from sqlalchemy.orm import interfaces
+
+
+    def _gen_relationship(
+        base, direction, return_fn, attrname, local_cls, referred_cls, **kw
+    ):
+        if direction is interfaces.ONETOMANY:
+            kw["cascade"] = "all, delete-orphan"
+            kw["passive_deletes"] = True
+        # make use of the built-in function to actually return
+        # the result.
+        return generate_relationship(
+            base, direction, return_fn, attrname, local_cls, referred_cls, **kw
+        )
+
+
+    from sqlalchemy.ext.automap import automap_base
+    from sqlalchemy import create_engine
+
+    # automap base
+    Base = automap_base()
+
+    engine = create_engine("sqlite:///mydatabase.db")
+    Base.prepare(autoload_with=engine, generate_relationship=_gen_relationship)
+
+Many-to-Many relationships
+--------------------------
+
+:mod:`.sqlalchemy.ext.automap` will generate many-to-many relationships, e.g.
+those which contain a ``secondary`` argument.  The process for producing these
+is as follows:
+
+1. A given :class:`_schema.Table` is examined for
+   :class:`_schema.ForeignKeyConstraint`
+   objects, before any mapped class has been assigned to it.
+
+2. If the table contains two and exactly two
+   :class:`_schema.ForeignKeyConstraint`
+   objects, and all columns within this table are members of these two
+   :class:`_schema.ForeignKeyConstraint` objects, the table is assumed to be a
+   "secondary" table, and will **not be mapped directly**.
+
+3. The two (or one, for self-referential) external tables to which the
+   :class:`_schema.Table`
+   refers to are matched to the classes to which they will be
+   mapped, if any.
+
+4. If mapped classes for both sides are located, a many-to-many bi-directional
+   :func:`_orm.relationship` / :func:`.backref`
+   pair is created between the two
+   classes.
+
+5. The override logic for many-to-many works the same as that of one-to-many/
+   many-to-one; the :func:`.generate_relationship` function is called upon
+   to generate the structures and existing attributes will be maintained.
+
+Relationships with Inheritance
+------------------------------
+
+:mod:`.sqlalchemy.ext.automap` will not generate any relationships between
+two classes that are in an inheritance relationship.   That is, with two
+classes given as follows::
+
+    class Employee(Base):
+        __tablename__ = "employee"
+        id = Column(Integer, primary_key=True)
+        type = Column(String(50))
+        __mapper_args__ = {
+            "polymorphic_identity": "employee",
+            "polymorphic_on": type,
+        }
+
+
+    class Engineer(Employee):
+        __tablename__ = "engineer"
+        id = Column(Integer, ForeignKey("employee.id"), primary_key=True)
+        __mapper_args__ = {
+            "polymorphic_identity": "engineer",
+        }
+
+The foreign key from ``Engineer`` to ``Employee`` is used not for a
+relationship, but to establish joined inheritance between the two classes.
+
+Note that this means automap will not generate *any* relationships
+for foreign keys that link from a subclass to a superclass.  If a mapping
+has actual relationships from subclass to superclass as well, those
+need to be explicit.  Below, as we have two separate foreign keys
+from ``Engineer`` to ``Employee``, we need to set up both the relationship
+we want as well as the ``inherit_condition``, as these are not things
+SQLAlchemy can guess::
+
+    class Employee(Base):
+        __tablename__ = "employee"
+        id = Column(Integer, primary_key=True)
+        type = Column(String(50))
+
+        __mapper_args__ = {
+            "polymorphic_identity": "employee",
+            "polymorphic_on": type,
+        }
+
+
+    class Engineer(Employee):
+        __tablename__ = "engineer"
+        id = Column(Integer, ForeignKey("employee.id"), primary_key=True)
+        favorite_employee_id = Column(Integer, ForeignKey("employee.id"))
+
+        favorite_employee = relationship(
+            Employee, foreign_keys=favorite_employee_id
+        )
+
+        __mapper_args__ = {
+            "polymorphic_identity": "engineer",
+            "inherit_condition": id == Employee.id,
+        }
+
+Handling Simple Naming Conflicts
+--------------------------------
+
+In the case of naming conflicts during mapping, override any of
+:func:`.classname_for_table`, :func:`.name_for_scalar_relationship`,
+and :func:`.name_for_collection_relationship` as needed.  For example, if
+automap is attempting to name a many-to-one relationship the same as an
+existing column, an alternate convention can be conditionally selected.  Given
+a schema:
+
+.. sourcecode:: sql
+
+    CREATE TABLE table_a (
+        id INTEGER PRIMARY KEY
+    );
+
+    CREATE TABLE table_b (
+        id INTEGER PRIMARY KEY,
+        table_a INTEGER,
+        FOREIGN KEY(table_a) REFERENCES table_a(id)
+    );
+
+The above schema will first automap the ``table_a`` table as a class named
+``table_a``; it will then automap a relationship onto the class for ``table_b``
+with the same name as this related class, e.g. ``table_a``.  This
+relationship name conflicts with the mapping column ``table_b.table_a``,
+and will emit an error on mapping.
+
+We can resolve this conflict by using an underscore as follows::
+
+    def name_for_scalar_relationship(
+        base, local_cls, referred_cls, constraint
+    ):
+        name = referred_cls.__name__.lower()
+        local_table = local_cls.__table__
+        if name in local_table.columns:
+            newname = name + "_"
+            warnings.warn(
+                "Already detected name %s present.  using %s" % (name, newname)
+            )
+            return newname
+        return name
+
+
+    Base.prepare(
+        autoload_with=engine,
+        name_for_scalar_relationship=name_for_scalar_relationship,
+    )
+
+Alternatively, we can change the name on the column side.   The columns
+that are mapped can be modified using the technique described at
+:ref:`mapper_column_distinct_names`, by assigning the column explicitly
+to a new name::
+
+    Base = automap_base()
+
+
+    class TableB(Base):
+        __tablename__ = "table_b"
+        _table_a = Column("table_a", ForeignKey("table_a.id"))
+
+
+    Base.prepare(autoload_with=engine)
+
+Using Automap with Explicit Declarations
+========================================
+
+As noted previously, automap has no dependency on reflection, and can make
+use of any collection of :class:`_schema.Table` objects within a
+:class:`_schema.MetaData`
+collection.  From this, it follows that automap can also be used
+generate missing relationships given an otherwise complete model that fully
+defines table metadata::
+
+    from sqlalchemy.ext.automap import automap_base
+    from sqlalchemy import Column, Integer, String, ForeignKey
+
+    Base = automap_base()
+
+
+    class User(Base):
+        __tablename__ = "user"
+
+        id = Column(Integer, primary_key=True)
+        name = Column(String)
+
+
+    class Address(Base):
+        __tablename__ = "address"
+
+        id = Column(Integer, primary_key=True)
+        email = Column(String)
+        user_id = Column(ForeignKey("user.id"))
+
+
+    # produce relationships
+    Base.prepare()
+
+    # mapping is complete, with "address_collection" and
+    # "user" relationships
+    a1 = Address(email="u1")
+    a2 = Address(email="u2")
+    u1 = User(address_collection=[a1, a2])
+    assert a1.user is u1
+
+Above, given mostly complete ``User`` and ``Address`` mappings, the
+:class:`_schema.ForeignKey` which we defined on ``Address.user_id`` allowed a
+bidirectional relationship pair ``Address.user`` and
+``User.address_collection`` to be generated on the mapped classes.
+
+Note that when subclassing :class:`.AutomapBase`,
+the :meth:`.AutomapBase.prepare` method is required; if not called, the classes
+we've declared are in an un-mapped state.
+
+
+.. _automap_intercepting_columns:
+
+Intercepting Column Definitions
+===============================
+
+The :class:`_schema.MetaData` and :class:`_schema.Table` objects support an
+event hook :meth:`_events.DDLEvents.column_reflect` that may be used to intercept
+the information reflected about a database column before the :class:`_schema.Column`
+object is constructed.   For example if we wanted to map columns using a
+naming convention such as ``"attr_<columnname>"``, the event could
+be applied as::
+
+    @event.listens_for(Base.metadata, "column_reflect")
+    def column_reflect(inspector, table, column_info):
+        # set column.key = "attr_<lower_case_name>"
+        column_info["key"] = "attr_%s" % column_info["name"].lower()
+
+
+    # run reflection
+    Base.prepare(autoload_with=engine)
+
+.. versionadded:: 1.4.0b2 the :meth:`_events.DDLEvents.column_reflect` event
+   may be applied to a :class:`_schema.MetaData` object.
+
+.. seealso::
+
+      :meth:`_events.DDLEvents.column_reflect`
+
+      :ref:`mapper_automated_reflection_schemes` - in the ORM mapping documentation
+
+
+"""  # noqa
+from __future__ import annotations
+
+import dataclasses
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import util
+from ..orm import backref
+from ..orm import declarative_base as _declarative_base
+from ..orm import exc as orm_exc
+from ..orm import interfaces
+from ..orm import relationship
+from ..orm.decl_base import _DeferredMapperConfig
+from ..orm.mapper import _CONFIGURE_MUTEX
+from ..schema import ForeignKeyConstraint
+from ..sql import and_
+from ..util import Properties
+from ..util.typing import Protocol
+
+if TYPE_CHECKING:
+    from ..engine.base import Engine
+    from ..orm.base import RelationshipDirection
+    from ..orm.relationships import ORMBackrefArgument
+    from ..orm.relationships import Relationship
+    from ..sql.schema import Column
+    from ..sql.schema import MetaData
+    from ..sql.schema import Table
+    from ..util import immutabledict
+
+
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+
+
+class PythonNameForTableType(Protocol):
+    def __call__(
+        self, base: Type[Any], tablename: str, table: Table
+    ) -> str: ...
+
+
+def classname_for_table(
+    base: Type[Any],
+    tablename: str,
+    table: Table,
+) -> str:
+    """Return the class name that should be used, given the name
+    of a table.
+
+    The default implementation is::
+
+        return str(tablename)
+
+    Alternate implementations can be specified using the
+    :paramref:`.AutomapBase.prepare.classname_for_table`
+    parameter.
+
+    :param base: the :class:`.AutomapBase` class doing the prepare.
+
+    :param tablename: string name of the :class:`_schema.Table`.
+
+    :param table: the :class:`_schema.Table` object itself.
+
+    :return: a string class name.
+
+     .. note::
+
+        In Python 2, the string used for the class name **must** be a
+        non-Unicode object, e.g. a ``str()`` object.  The ``.name`` attribute
+        of :class:`_schema.Table` is typically a Python unicode subclass,
+        so the
+        ``str()`` function should be applied to this name, after accounting for
+        any non-ASCII characters.
+
+    """
+    return str(tablename)
+
+
+class NameForScalarRelationshipType(Protocol):
+    def __call__(
+        self,
+        base: Type[Any],
+        local_cls: Type[Any],
+        referred_cls: Type[Any],
+        constraint: ForeignKeyConstraint,
+    ) -> str: ...
+
+
+def name_for_scalar_relationship(
+    base: Type[Any],
+    local_cls: Type[Any],
+    referred_cls: Type[Any],
+    constraint: ForeignKeyConstraint,
+) -> str:
+    """Return the attribute name that should be used to refer from one
+    class to another, for a scalar object reference.
+
+    The default implementation is::
+
+        return referred_cls.__name__.lower()
+
+    Alternate implementations can be specified using the
+    :paramref:`.AutomapBase.prepare.name_for_scalar_relationship`
+    parameter.
+
+    :param base: the :class:`.AutomapBase` class doing the prepare.
+
+    :param local_cls: the class to be mapped on the local side.
+
+    :param referred_cls: the class to be mapped on the referring side.
+
+    :param constraint: the :class:`_schema.ForeignKeyConstraint` that is being
+     inspected to produce this relationship.
+
+    """
+    return referred_cls.__name__.lower()
+
+
+class NameForCollectionRelationshipType(Protocol):
+    def __call__(
+        self,
+        base: Type[Any],
+        local_cls: Type[Any],
+        referred_cls: Type[Any],
+        constraint: ForeignKeyConstraint,
+    ) -> str: ...
+
+
+def name_for_collection_relationship(
+    base: Type[Any],
+    local_cls: Type[Any],
+    referred_cls: Type[Any],
+    constraint: ForeignKeyConstraint,
+) -> str:
+    """Return the attribute name that should be used to refer from one
+    class to another, for a collection reference.
+
+    The default implementation is::
+
+        return referred_cls.__name__.lower() + "_collection"
+
+    Alternate implementations
+    can be specified using the
+    :paramref:`.AutomapBase.prepare.name_for_collection_relationship`
+    parameter.
+
+    :param base: the :class:`.AutomapBase` class doing the prepare.
+
+    :param local_cls: the class to be mapped on the local side.
+
+    :param referred_cls: the class to be mapped on the referring side.
+
+    :param constraint: the :class:`_schema.ForeignKeyConstraint` that is being
+     inspected to produce this relationship.
+
+    """
+    return referred_cls.__name__.lower() + "_collection"
+
+
+class GenerateRelationshipType(Protocol):
+    @overload
+    def __call__(
+        self,
+        base: Type[Any],
+        direction: RelationshipDirection,
+        return_fn: Callable[..., Relationship[Any]],
+        attrname: str,
+        local_cls: Type[Any],
+        referred_cls: Type[Any],
+        **kw: Any,
+    ) -> Relationship[Any]: ...
+
+    @overload
+    def __call__(
+        self,
+        base: Type[Any],
+        direction: RelationshipDirection,
+        return_fn: Callable[..., ORMBackrefArgument],
+        attrname: str,
+        local_cls: Type[Any],
+        referred_cls: Type[Any],
+        **kw: Any,
+    ) -> ORMBackrefArgument: ...
+
+    def __call__(
+        self,
+        base: Type[Any],
+        direction: RelationshipDirection,
+        return_fn: Union[
+            Callable[..., Relationship[Any]], Callable[..., ORMBackrefArgument]
+        ],
+        attrname: str,
+        local_cls: Type[Any],
+        referred_cls: Type[Any],
+        **kw: Any,
+    ) -> Union[ORMBackrefArgument, Relationship[Any]]: ...
+
+
+@overload
+def generate_relationship(
+    base: Type[Any],
+    direction: RelationshipDirection,
+    return_fn: Callable[..., Relationship[Any]],
+    attrname: str,
+    local_cls: Type[Any],
+    referred_cls: Type[Any],
+    **kw: Any,
+) -> Relationship[Any]: ...
+
+
+@overload
+def generate_relationship(
+    base: Type[Any],
+    direction: RelationshipDirection,
+    return_fn: Callable[..., ORMBackrefArgument],
+    attrname: str,
+    local_cls: Type[Any],
+    referred_cls: Type[Any],
+    **kw: Any,
+) -> ORMBackrefArgument: ...
+
+
+def generate_relationship(
+    base: Type[Any],
+    direction: RelationshipDirection,
+    return_fn: Union[
+        Callable[..., Relationship[Any]], Callable[..., ORMBackrefArgument]
+    ],
+    attrname: str,
+    local_cls: Type[Any],
+    referred_cls: Type[Any],
+    **kw: Any,
+) -> Union[Relationship[Any], ORMBackrefArgument]:
+    r"""Generate a :func:`_orm.relationship` or :func:`.backref`
+    on behalf of two
+    mapped classes.
+
+    An alternate implementation of this function can be specified using the
+    :paramref:`.AutomapBase.prepare.generate_relationship` parameter.
+
+    The default implementation of this function is as follows::
+
+        if return_fn is backref:
+            return return_fn(attrname, **kw)
+        elif return_fn is relationship:
+            return return_fn(referred_cls, **kw)
+        else:
+            raise TypeError("Unknown relationship function: %s" % return_fn)
+
+    :param base: the :class:`.AutomapBase` class doing the prepare.
+
+    :param direction: indicate the "direction" of the relationship; this will
+     be one of :data:`.ONETOMANY`, :data:`.MANYTOONE`, :data:`.MANYTOMANY`.
+
+    :param return_fn: the function that is used by default to create the
+     relationship.  This will be either :func:`_orm.relationship` or
+     :func:`.backref`.  The :func:`.backref` function's result will be used to
+     produce a new :func:`_orm.relationship` in a second step,
+     so it is critical
+     that user-defined implementations correctly differentiate between the two
+     functions, if a custom relationship function is being used.
+
+    :param attrname: the attribute name to which this relationship is being
+     assigned. If the value of :paramref:`.generate_relationship.return_fn` is
+     the :func:`.backref` function, then this name is the name that is being
+     assigned to the backref.
+
+    :param local_cls: the "local" class to which this relationship or backref
+     will be locally present.
+
+    :param referred_cls: the "referred" class to which the relationship or
+     backref refers to.
+
+    :param \**kw: all additional keyword arguments are passed along to the
+     function.
+
+    :return: a :func:`_orm.relationship` or :func:`.backref` construct,
+     as dictated
+     by the :paramref:`.generate_relationship.return_fn` parameter.
+
+    """
+
+    if return_fn is backref:
+        return return_fn(attrname, **kw)
+    elif return_fn is relationship:
+        return return_fn(referred_cls, **kw)
+    else:
+        raise TypeError("Unknown relationship function: %s" % return_fn)
+
+
+ByModuleProperties = Properties[Union["ByModuleProperties", Type[Any]]]
+
+
+class AutomapBase:
+    """Base class for an "automap" schema.
+
+    The :class:`.AutomapBase` class can be compared to the "declarative base"
+    class that is produced by the :func:`.declarative.declarative_base`
+    function.  In practice, the :class:`.AutomapBase` class is always used
+    as a mixin along with an actual declarative base.
+
+    A new subclassable :class:`.AutomapBase` is typically instantiated
+    using the :func:`.automap_base` function.
+
+    .. seealso::
+
+        :ref:`automap_toplevel`
+
+    """
+
+    __abstract__ = True
+
+    classes: ClassVar[Properties[Type[Any]]]
+    """An instance of :class:`.util.Properties` containing classes.
+
+    This object behaves much like the ``.c`` collection on a table.  Classes
+    are present under the name they were given, e.g.::
+
+        Base = automap_base()
+        Base.prepare(autoload_with=some_engine)
+
+        User, Address = Base.classes.User, Base.classes.Address
+
+    For class names that overlap with a method name of
+    :class:`.util.Properties`, such as ``items()``, the getitem form
+    is also supported::
+
+        Item = Base.classes["items"]
+
+    """
+
+    by_module: ClassVar[ByModuleProperties]
+    """An instance of :class:`.util.Properties` containing a hierarchal
+    structure of dot-separated module names linked to classes.
+
+    This collection is an alternative to the :attr:`.AutomapBase.classes`
+    collection that is useful when making use of the
+    :paramref:`.AutomapBase.prepare.modulename_for_table` parameter, which will
+    apply distinct ``__module__`` attributes to generated classes.
+
+    The default ``__module__`` an automap-generated class is
+    ``sqlalchemy.ext.automap``; to access this namespace using
+    :attr:`.AutomapBase.by_module` looks like::
+
+        User = Base.by_module.sqlalchemy.ext.automap.User
+
+    If a class had a ``__module__`` of ``mymodule.account``, accessing
+    this namespace looks like::
+
+        MyClass = Base.by_module.mymodule.account.MyClass
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`automap_by_module`
+
+    """
+
+    metadata: ClassVar[MetaData]
+    """Refers to the :class:`_schema.MetaData` collection that will be used
+    for new :class:`_schema.Table` objects.
+
+    .. seealso::
+
+        :ref:`orm_declarative_metadata`
+
+    """
+
+    _sa_automapbase_bookkeeping: ClassVar[_Bookkeeping]
+
+    @classmethod
+    @util.deprecated_params(
+        engine=(
+            "2.0",
+            "The :paramref:`_automap.AutomapBase.prepare.engine` parameter "
+            "is deprecated and will be removed in a future release.  "
+            "Please use the "
+            ":paramref:`_automap.AutomapBase.prepare.autoload_with` "
+            "parameter.",
+        ),
+        reflect=(
+            "2.0",
+            "The :paramref:`_automap.AutomapBase.prepare.reflect` "
+            "parameter is deprecated and will be removed in a future "
+            "release.  Reflection is enabled when "
+            ":paramref:`_automap.AutomapBase.prepare.autoload_with` "
+            "is passed.",
+        ),
+    )
+    def prepare(
+        cls: Type[AutomapBase],
+        autoload_with: Optional[Engine] = None,
+        engine: Optional[Any] = None,
+        reflect: bool = False,
+        schema: Optional[str] = None,
+        classname_for_table: Optional[PythonNameForTableType] = None,
+        modulename_for_table: Optional[PythonNameForTableType] = None,
+        collection_class: Optional[Any] = None,
+        name_for_scalar_relationship: Optional[
+            NameForScalarRelationshipType
+        ] = None,
+        name_for_collection_relationship: Optional[
+            NameForCollectionRelationshipType
+        ] = None,
+        generate_relationship: Optional[GenerateRelationshipType] = None,
+        reflection_options: Union[
+            Dict[_KT, _VT], immutabledict[_KT, _VT]
+        ] = util.EMPTY_DICT,
+    ) -> None:
+        """Extract mapped classes and relationships from the
+        :class:`_schema.MetaData` and perform mappings.
+
+        For full documentation and examples see
+        :ref:`automap_basic_use`.
+
+        :param autoload_with: an :class:`_engine.Engine` or
+         :class:`_engine.Connection` with which
+         to perform schema reflection; when specified, the
+         :meth:`_schema.MetaData.reflect` method will be invoked within
+         the scope of this method.
+
+        :param engine: legacy; use :paramref:`.AutomapBase.autoload_with`.
+         Used to indicate the :class:`_engine.Engine` or
+         :class:`_engine.Connection` with which to reflect tables with,
+         if :paramref:`.AutomapBase.reflect` is True.
+
+        :param reflect: legacy; use :paramref:`.AutomapBase.autoload_with`.
+         Indicates that :meth:`_schema.MetaData.reflect` should be invoked.
+
+        :param classname_for_table: callable function which will be used to
+         produce new class names, given a table name.  Defaults to
+         :func:`.classname_for_table`.
+
+        :param modulename_for_table: callable function which will be used to
+         produce the effective ``__module__`` for an internally generated
+         class, to allow for multiple classes of the same name in a single
+         automap base which would be in different "modules".
+
+         Defaults to ``None``, which will indicate that ``__module__`` will not
+         be set explicitly; the Python runtime will use the value
+         ``sqlalchemy.ext.automap`` for these classes.
+
+         When assigning ``__module__`` to generated classes, they can be
+         accessed based on dot-separated module names using the
+         :attr:`.AutomapBase.by_module` collection.   Classes that have
+         an explicit ``__module_`` assigned using this hook do **not** get
+         placed into the :attr:`.AutomapBase.classes` collection, only
+         into :attr:`.AutomapBase.by_module`.
+
+         .. versionadded:: 2.0
+
+         .. seealso::
+
+            :ref:`automap_by_module`
+
+        :param name_for_scalar_relationship: callable function which will be
+         used to produce relationship names for scalar relationships.  Defaults
+         to :func:`.name_for_scalar_relationship`.
+
+        :param name_for_collection_relationship: callable function which will
+         be used to produce relationship names for collection-oriented
+         relationships.  Defaults to :func:`.name_for_collection_relationship`.
+
+        :param generate_relationship: callable function which will be used to
+         actually generate :func:`_orm.relationship` and :func:`.backref`
+         constructs.  Defaults to :func:`.generate_relationship`.
+
+        :param collection_class: the Python collection class that will be used
+         when a new :func:`_orm.relationship`
+         object is created that represents a
+         collection.  Defaults to ``list``.
+
+        :param schema: Schema name to reflect when reflecting tables using
+         the :paramref:`.AutomapBase.prepare.autoload_with` parameter. The name
+         is passed to the :paramref:`_schema.MetaData.reflect.schema` parameter
+         of :meth:`_schema.MetaData.reflect`. When omitted, the default schema
+         in use by the database connection is used.
+
+         .. note:: The :paramref:`.AutomapBase.prepare.schema`
+            parameter supports reflection of a single schema at a time.
+            In order to include tables from many schemas, use
+            multiple calls to :meth:`.AutomapBase.prepare`.
+
+            For an overview of multiple-schema automap including the use
+            of additional naming conventions to resolve table name
+            conflicts, see the section :ref:`automap_by_module`.
+
+            .. versionadded:: 2.0 :meth:`.AutomapBase.prepare` supports being
+               directly invoked any number of times, keeping track of tables
+               that have already been processed to avoid processing them
+               a second time.
+
+        :param reflection_options: When present, this dictionary of options
+         will be passed to :meth:`_schema.MetaData.reflect`
+         to supply general reflection-specific options like ``only`` and/or
+         dialect-specific options like ``oracle_resolve_synonyms``.
+
+         .. versionadded:: 1.4
+
+        """
+
+        for mr in cls.__mro__:
+            if "_sa_automapbase_bookkeeping" in mr.__dict__:
+                automap_base = cast("Type[AutomapBase]", mr)
+                break
+        else:
+            assert False, "Can't locate automap base in class hierarchy"
+
+        glbls = globals()
+        if classname_for_table is None:
+            classname_for_table = glbls["classname_for_table"]
+        if name_for_scalar_relationship is None:
+            name_for_scalar_relationship = glbls[
+                "name_for_scalar_relationship"
+            ]
+        if name_for_collection_relationship is None:
+            name_for_collection_relationship = glbls[
+                "name_for_collection_relationship"
+            ]
+        if generate_relationship is None:
+            generate_relationship = glbls["generate_relationship"]
+        if collection_class is None:
+            collection_class = list
+
+        if autoload_with:
+            reflect = True
+
+        if engine:
+            autoload_with = engine
+
+        if reflect:
+            assert autoload_with
+            opts = dict(
+                schema=schema,
+                extend_existing=True,
+                autoload_replace=False,
+            )
+            if reflection_options:
+                opts.update(reflection_options)
+            cls.metadata.reflect(autoload_with, **opts)  # type: ignore[arg-type]  # noqa: E501
+
+        with _CONFIGURE_MUTEX:
+            table_to_map_config: Union[
+                Dict[Optional[Table], _DeferredMapperConfig],
+                Dict[Table, _DeferredMapperConfig],
+            ] = {
+                cast("Table", m.local_table): m
+                for m in _DeferredMapperConfig.classes_for_base(
+                    cls, sort=False
+                )
+            }
+
+            many_to_many: List[
+                Tuple[Table, Table, List[ForeignKeyConstraint], Table]
+            ]
+            many_to_many = []
+
+            bookkeeping = automap_base._sa_automapbase_bookkeeping
+            metadata_tables = cls.metadata.tables
+
+            for table_key in set(metadata_tables).difference(
+                bookkeeping.table_keys
+            ):
+                table = metadata_tables[table_key]
+                bookkeeping.table_keys.add(table_key)
+
+                lcl_m2m, rem_m2m, m2m_const = _is_many_to_many(cls, table)
+                if lcl_m2m is not None:
+                    assert rem_m2m is not None
+                    assert m2m_const is not None
+                    many_to_many.append((lcl_m2m, rem_m2m, m2m_const, table))
+                elif not table.primary_key:
+                    continue
+                elif table not in table_to_map_config:
+                    clsdict: Dict[str, Any] = {"__table__": table}
+                    if modulename_for_table is not None:
+                        new_module = modulename_for_table(
+                            cls, table.name, table
+                        )
+                        if new_module is not None:
+                            clsdict["__module__"] = new_module
+                    else:
+                        new_module = None
+
+                    newname = classname_for_table(cls, table.name, table)
+                    if new_module is None and newname in cls.classes:
+                        util.warn(
+                            "Ignoring duplicate class name "
+                            f"'{newname}' "
+                            "received in automap base for table "
+                            f"{table.key} without "
+                            "``__module__`` being set; consider using the "
+                            "``modulename_for_table`` hook"
+                        )
+                        continue
+
+                    mapped_cls = type(
+                        newname,
+                        (automap_base,),
+                        clsdict,
+                    )
+                    map_config = _DeferredMapperConfig.config_for_cls(
+                        mapped_cls
+                    )
+                    assert map_config.cls.__name__ == newname
+                    if new_module is None:
+                        cls.classes[newname] = mapped_cls
+
+                    by_module_properties: ByModuleProperties = cls.by_module
+                    for token in map_config.cls.__module__.split("."):
+                        if token not in by_module_properties:
+                            by_module_properties[token] = util.Properties({})
+
+                        props = by_module_properties[token]
+
+                        # we can assert this because the clsregistry
+                        # module would have raised if there was a mismatch
+                        # between modules/classes already.
+                        # see test_cls_schema_name_conflict
+                        assert isinstance(props, Properties)
+                        by_module_properties = props
+
+                    by_module_properties[map_config.cls.__name__] = mapped_cls
+
+                    table_to_map_config[table] = map_config
+
+            for map_config in table_to_map_config.values():
+                _relationships_for_fks(
+                    automap_base,
+                    map_config,
+                    table_to_map_config,
+                    collection_class,
+                    name_for_scalar_relationship,
+                    name_for_collection_relationship,
+                    generate_relationship,
+                )
+
+            for lcl_m2m, rem_m2m, m2m_const, table in many_to_many:
+                _m2m_relationship(
+                    automap_base,
+                    lcl_m2m,
+                    rem_m2m,
+                    m2m_const,
+                    table,
+                    table_to_map_config,
+                    collection_class,
+                    name_for_scalar_relationship,
+                    name_for_collection_relationship,
+                    generate_relationship,
+                )
+
+            for map_config in _DeferredMapperConfig.classes_for_base(
+                automap_base
+            ):
+                map_config.map()
+
+    _sa_decl_prepare = True
+    """Indicate that the mapping of classes should be deferred.
+
+    The presence of this attribute name indicates to declarative
+    that the call to mapper() should not occur immediately; instead,
+    information about the table and attributes to be mapped are gathered
+    into an internal structure called _DeferredMapperConfig.  These
+    objects can be collected later using classes_for_base(), additional
+    mapping decisions can be made, and then the map() method will actually
+    apply the mapping.
+
+    The only real reason this deferral of the whole
+    thing is needed is to support primary key columns that aren't reflected
+    yet when the class is declared; everything else can theoretically be
+    added to the mapper later.  However, the _DeferredMapperConfig is a
+    nice interface in any case which exists at that not usually exposed point
+    at which declarative has the class and the Table but hasn't called
+    mapper() yet.
+
+    """
+
+    @classmethod
+    def _sa_raise_deferred_config(cls) -> NoReturn:
+        raise orm_exc.UnmappedClassError(
+            cls,
+            msg="Class %s is a subclass of AutomapBase.  "
+            "Mappings are not produced until the .prepare() "
+            "method is called on the class hierarchy."
+            % orm_exc._safe_cls_name(cls),
+        )
+
+
+@dataclasses.dataclass
+class _Bookkeeping:
+    __slots__ = ("table_keys",)
+
+    table_keys: Set[str]
+
+
+def automap_base(
+    declarative_base: Optional[Type[Any]] = None, **kw: Any
+) -> Any:
+    r"""Produce a declarative automap base.
+
+    This function produces a new base class that is a product of the
+    :class:`.AutomapBase` class as well a declarative base produced by
+    :func:`.declarative.declarative_base`.
+
+    All parameters other than ``declarative_base`` are keyword arguments
+    that are passed directly to the :func:`.declarative.declarative_base`
+    function.
+
+    :param declarative_base: an existing class produced by
+     :func:`.declarative.declarative_base`.  When this is passed, the function
+     no longer invokes :func:`.declarative.declarative_base` itself, and all
+     other keyword arguments are ignored.
+
+    :param \**kw: keyword arguments are passed along to
+     :func:`.declarative.declarative_base`.
+
+    """
+    if declarative_base is None:
+        Base = _declarative_base(**kw)
+    else:
+        Base = declarative_base
+
+    return type(
+        Base.__name__,
+        (AutomapBase, Base),
+        {
+            "__abstract__": True,
+            "classes": util.Properties({}),
+            "by_module": util.Properties({}),
+            "_sa_automapbase_bookkeeping": _Bookkeeping(set()),
+        },
+    )
+
+
+def _is_many_to_many(
+    automap_base: Type[Any], table: Table
+) -> Tuple[
+    Optional[Table], Optional[Table], Optional[list[ForeignKeyConstraint]]
+]:
+    fk_constraints = [
+        const
+        for const in table.constraints
+        if isinstance(const, ForeignKeyConstraint)
+    ]
+    if len(fk_constraints) != 2:
+        return None, None, None
+
+    cols: List[Column[Any]] = sum(
+        [
+            [fk.parent for fk in fk_constraint.elements]
+            for fk_constraint in fk_constraints
+        ],
+        [],
+    )
+
+    if set(cols) != set(table.c):
+        return None, None, None
+
+    return (
+        fk_constraints[0].elements[0].column.table,
+        fk_constraints[1].elements[0].column.table,
+        fk_constraints,
+    )
+
+
+def _relationships_for_fks(
+    automap_base: Type[Any],
+    map_config: _DeferredMapperConfig,
+    table_to_map_config: Union[
+        Dict[Optional[Table], _DeferredMapperConfig],
+        Dict[Table, _DeferredMapperConfig],
+    ],
+    collection_class: type,
+    name_for_scalar_relationship: NameForScalarRelationshipType,
+    name_for_collection_relationship: NameForCollectionRelationshipType,
+    generate_relationship: GenerateRelationshipType,
+) -> None:
+    local_table = cast("Optional[Table]", map_config.local_table)
+    local_cls = cast(
+        "Optional[Type[Any]]", map_config.cls
+    )  # derived from a weakref, may be None
+
+    if local_table is None or local_cls is None:
+        return
+    for constraint in local_table.constraints:
+        if isinstance(constraint, ForeignKeyConstraint):
+            fks = constraint.elements
+            referred_table = fks[0].column.table
+            referred_cfg = table_to_map_config.get(referred_table, None)
+            if referred_cfg is None:
+                continue
+            referred_cls = referred_cfg.cls
+
+            if local_cls is not referred_cls and issubclass(
+                local_cls, referred_cls
+            ):
+                continue
+
+            relationship_name = name_for_scalar_relationship(
+                automap_base, local_cls, referred_cls, constraint
+            )
+            backref_name = name_for_collection_relationship(
+                automap_base, referred_cls, local_cls, constraint
+            )
+
+            o2m_kws: Dict[str, Union[str, bool]] = {}
+            nullable = False not in {fk.parent.nullable for fk in fks}
+            if not nullable:
+                o2m_kws["cascade"] = "all, delete-orphan"
+
+                if (
+                    constraint.ondelete
+                    and constraint.ondelete.lower() == "cascade"
+                ):
+                    o2m_kws["passive_deletes"] = True
+            else:
+                if (
+                    constraint.ondelete
+                    and constraint.ondelete.lower() == "set null"
+                ):
+                    o2m_kws["passive_deletes"] = True
+
+            create_backref = backref_name not in referred_cfg.properties
+
+            if relationship_name not in map_config.properties:
+                if create_backref:
+                    backref_obj = generate_relationship(
+                        automap_base,
+                        interfaces.ONETOMANY,
+                        backref,
+                        backref_name,
+                        referred_cls,
+                        local_cls,
+                        collection_class=collection_class,
+                        **o2m_kws,
+                    )
+                else:
+                    backref_obj = None
+                rel = generate_relationship(
+                    automap_base,
+                    interfaces.MANYTOONE,
+                    relationship,
+                    relationship_name,
+                    local_cls,
+                    referred_cls,
+                    foreign_keys=[fk.parent for fk in constraint.elements],
+                    backref=backref_obj,
+                    remote_side=[fk.column for fk in constraint.elements],
+                )
+                if rel is not None:
+                    map_config.properties[relationship_name] = rel
+                    if not create_backref:
+                        referred_cfg.properties[
+                            backref_name
+                        ].back_populates = relationship_name  # type: ignore[union-attr] # noqa: E501
+            elif create_backref:
+                rel = generate_relationship(
+                    automap_base,
+                    interfaces.ONETOMANY,
+                    relationship,
+                    backref_name,
+                    referred_cls,
+                    local_cls,
+                    foreign_keys=[fk.parent for fk in constraint.elements],
+                    back_populates=relationship_name,
+                    collection_class=collection_class,
+                    **o2m_kws,
+                )
+                if rel is not None:
+                    referred_cfg.properties[backref_name] = rel
+                    map_config.properties[
+                        relationship_name
+                    ].back_populates = backref_name  # type: ignore[union-attr]
+
+
+def _m2m_relationship(
+    automap_base: Type[Any],
+    lcl_m2m: Table,
+    rem_m2m: Table,
+    m2m_const: List[ForeignKeyConstraint],
+    table: Table,
+    table_to_map_config: Union[
+        Dict[Optional[Table], _DeferredMapperConfig],
+        Dict[Table, _DeferredMapperConfig],
+    ],
+    collection_class: type,
+    name_for_scalar_relationship: NameForCollectionRelationshipType,
+    name_for_collection_relationship: NameForCollectionRelationshipType,
+    generate_relationship: GenerateRelationshipType,
+) -> None:
+    map_config = table_to_map_config.get(lcl_m2m, None)
+    referred_cfg = table_to_map_config.get(rem_m2m, None)
+    if map_config is None or referred_cfg is None:
+        return
+
+    local_cls = map_config.cls
+    referred_cls = referred_cfg.cls
+
+    relationship_name = name_for_collection_relationship(
+        automap_base, local_cls, referred_cls, m2m_const[0]
+    )
+    backref_name = name_for_collection_relationship(
+        automap_base, referred_cls, local_cls, m2m_const[1]
+    )
+
+    create_backref = backref_name not in referred_cfg.properties
+
+    if table in table_to_map_config:
+        overlaps = "__*"
+    else:
+        overlaps = None
+
+    if relationship_name not in map_config.properties:
+        if create_backref:
+            backref_obj = generate_relationship(
+                automap_base,
+                interfaces.MANYTOMANY,
+                backref,
+                backref_name,
+                referred_cls,
+                local_cls,
+                collection_class=collection_class,
+                overlaps=overlaps,
+            )
+        else:
+            backref_obj = None
+
+        rel = generate_relationship(
+            automap_base,
+            interfaces.MANYTOMANY,
+            relationship,
+            relationship_name,
+            local_cls,
+            referred_cls,
+            overlaps=overlaps,
+            secondary=table,
+            primaryjoin=and_(
+                fk.column == fk.parent for fk in m2m_const[0].elements
+            ),  # type: ignore [arg-type]
+            secondaryjoin=and_(
+                fk.column == fk.parent for fk in m2m_const[1].elements
+            ),  # type: ignore [arg-type]
+            backref=backref_obj,
+            collection_class=collection_class,
+        )
+        if rel is not None:
+            map_config.properties[relationship_name] = rel
+
+            if not create_backref:
+                referred_cfg.properties[
+                    backref_name
+                ].back_populates = relationship_name  # type: ignore[union-attr] # noqa: E501
+    elif create_backref:
+        rel = generate_relationship(
+            automap_base,
+            interfaces.MANYTOMANY,
+            relationship,
+            backref_name,
+            referred_cls,
+            local_cls,
+            overlaps=overlaps,
+            secondary=table,
+            primaryjoin=and_(
+                fk.column == fk.parent for fk in m2m_const[1].elements
+            ),  # type: ignore [arg-type]
+            secondaryjoin=and_(
+                fk.column == fk.parent for fk in m2m_const[0].elements
+            ),  # type: ignore [arg-type]
+            back_populates=relationship_name,
+            collection_class=collection_class,
+        )
+        if rel is not None:
+            referred_cfg.properties[backref_name] = rel
+            map_config.properties[
+                relationship_name
+            ].back_populates = backref_name  # type: ignore[union-attr]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/baked.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/baked.py
new file mode 100644
index 00000000..cd3e0879
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/baked.py
@@ -0,0 +1,570 @@
+# ext/baked.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Baked query extension.
+
+Provides a creational pattern for the :class:`.query.Query` object which
+allows the fully constructed object, Core select statement, and string
+compiled result to be fully cached.
+
+
+"""
+
+import collections.abc as collections_abc
+import logging
+
+from .. import exc as sa_exc
+from .. import util
+from ..orm import exc as orm_exc
+from ..orm.query import Query
+from ..orm.session import Session
+from ..sql import func
+from ..sql import literal_column
+from ..sql import util as sql_util
+
+
+log = logging.getLogger(__name__)
+
+
+class Bakery:
+    """Callable which returns a :class:`.BakedQuery`.
+
+    This object is returned by the class method
+    :meth:`.BakedQuery.bakery`.  It exists as an object
+    so that the "cache" can be easily inspected.
+
+    .. versionadded:: 1.2
+
+
+    """
+
+    __slots__ = "cls", "cache"
+
+    def __init__(self, cls_, cache):
+        self.cls = cls_
+        self.cache = cache
+
+    def __call__(self, initial_fn, *args):
+        return self.cls(self.cache, initial_fn, args)
+
+
+class BakedQuery:
+    """A builder object for :class:`.query.Query` objects."""
+
+    __slots__ = "steps", "_bakery", "_cache_key", "_spoiled"
+
+    def __init__(self, bakery, initial_fn, args=()):
+        self._cache_key = ()
+        self._update_cache_key(initial_fn, args)
+        self.steps = [initial_fn]
+        self._spoiled = False
+        self._bakery = bakery
+
+    @classmethod
+    def bakery(cls, size=200, _size_alert=None):
+        """Construct a new bakery.
+
+        :return: an instance of :class:`.Bakery`
+
+        """
+
+        return Bakery(cls, util.LRUCache(size, size_alert=_size_alert))
+
+    def _clone(self):
+        b1 = BakedQuery.__new__(BakedQuery)
+        b1._cache_key = self._cache_key
+        b1.steps = list(self.steps)
+        b1._bakery = self._bakery
+        b1._spoiled = self._spoiled
+        return b1
+
+    def _update_cache_key(self, fn, args=()):
+        self._cache_key += (fn.__code__,) + args
+
+    def __iadd__(self, other):
+        if isinstance(other, tuple):
+            self.add_criteria(*other)
+        else:
+            self.add_criteria(other)
+        return self
+
+    def __add__(self, other):
+        if isinstance(other, tuple):
+            return self.with_criteria(*other)
+        else:
+            return self.with_criteria(other)
+
+    def add_criteria(self, fn, *args):
+        """Add a criteria function to this :class:`.BakedQuery`.
+
+        This is equivalent to using the ``+=`` operator to
+        modify a :class:`.BakedQuery` in-place.
+
+        """
+        self._update_cache_key(fn, args)
+        self.steps.append(fn)
+        return self
+
+    def with_criteria(self, fn, *args):
+        """Add a criteria function to a :class:`.BakedQuery` cloned from this
+        one.
+
+        This is equivalent to using the ``+`` operator to
+        produce a new :class:`.BakedQuery` with modifications.
+
+        """
+        return self._clone().add_criteria(fn, *args)
+
+    def for_session(self, session):
+        """Return a :class:`_baked.Result` object for this
+        :class:`.BakedQuery`.
+
+        This is equivalent to calling the :class:`.BakedQuery` as a
+        Python callable, e.g. ``result = my_baked_query(session)``.
+
+        """
+        return Result(self, session)
+
+    def __call__(self, session):
+        return self.for_session(session)
+
+    def spoil(self, full=False):
+        """Cancel any query caching that will occur on this BakedQuery object.
+
+        The BakedQuery can continue to be used normally, however additional
+        creational functions will not be cached; they will be called
+        on every invocation.
+
+        This is to support the case where a particular step in constructing
+        a baked query disqualifies the query from being cacheable, such
+        as a variant that relies upon some uncacheable value.
+
+        :param full: if False, only functions added to this
+         :class:`.BakedQuery` object subsequent to the spoil step will be
+         non-cached; the state of the :class:`.BakedQuery` up until
+         this point will be pulled from the cache.   If True, then the
+         entire :class:`_query.Query` object is built from scratch each
+         time, with all creational functions being called on each
+         invocation.
+
+        """
+        if not full and not self._spoiled:
+            _spoil_point = self._clone()
+            _spoil_point._cache_key += ("_query_only",)
+            self.steps = [_spoil_point._retrieve_baked_query]
+        self._spoiled = True
+        return self
+
+    def _effective_key(self, session):
+        """Return the key that actually goes into the cache dictionary for
+        this :class:`.BakedQuery`, taking into account the given
+        :class:`.Session`.
+
+        This basically means we also will include the session's query_class,
+        as the actual :class:`_query.Query` object is part of what's cached
+        and needs to match the type of :class:`_query.Query` that a later
+        session will want to use.
+
+        """
+        return self._cache_key + (session._query_cls,)
+
+    def _with_lazyload_options(self, options, effective_path, cache_path=None):
+        """Cloning version of _add_lazyload_options."""
+        q = self._clone()
+        q._add_lazyload_options(options, effective_path, cache_path=cache_path)
+        return q
+
+    def _add_lazyload_options(self, options, effective_path, cache_path=None):
+        """Used by per-state lazy loaders to add options to the
+        "lazy load" query from a parent query.
+
+        Creates a cache key based on given load path and query options;
+        if a repeatable cache key cannot be generated, the query is
+        "spoiled" so that it won't use caching.
+
+        """
+
+        key = ()
+
+        if not cache_path:
+            cache_path = effective_path
+
+        for opt in options:
+            if opt._is_legacy_option or opt._is_compile_state:
+                ck = opt._generate_cache_key()
+                if ck is None:
+                    self.spoil(full=True)
+                else:
+                    assert not ck[1], (
+                        "loader options with variable bound parameters "
+                        "not supported with baked queries.  Please "
+                        "use new-style select() statements for cached "
+                        "ORM queries."
+                    )
+                    key += ck[0]
+
+        self.add_criteria(
+            lambda q: q._with_current_path(effective_path).options(*options),
+            cache_path.path,
+            key,
+        )
+
+    def _retrieve_baked_query(self, session):
+        query = self._bakery.get(self._effective_key(session), None)
+        if query is None:
+            query = self._as_query(session)
+            self._bakery[self._effective_key(session)] = query.with_session(
+                None
+            )
+        return query.with_session(session)
+
+    def _bake(self, session):
+        query = self._as_query(session)
+        query.session = None
+
+        # in 1.4, this is where before_compile() event is
+        # invoked
+        statement = query._statement_20()
+
+        # if the query is not safe to cache, we still do everything as though
+        # we did cache it, since the receiver of _bake() assumes subqueryload
+        # context was set up, etc.
+        #
+        # note also we want to cache the statement itself because this
+        # allows the statement itself to hold onto its cache key that is
+        # used by the Connection, which in itself is more expensive to
+        # generate than what BakedQuery was able to provide in 1.3 and prior
+
+        if statement._compile_options._bake_ok:
+            self._bakery[self._effective_key(session)] = (
+                query,
+                statement,
+            )
+
+        return query, statement
+
+    def to_query(self, query_or_session):
+        """Return the :class:`_query.Query` object for use as a subquery.
+
+        This method should be used within the lambda callable being used
+        to generate a step of an enclosing :class:`.BakedQuery`.   The
+        parameter should normally be the :class:`_query.Query` object that
+        is passed to the lambda::
+
+            sub_bq = self.bakery(lambda s: s.query(User.name))
+            sub_bq += lambda q: q.filter(User.id == Address.user_id).correlate(Address)
+
+            main_bq = self.bakery(lambda s: s.query(Address))
+            main_bq += lambda q: q.filter(sub_bq.to_query(q).exists())
+
+        In the case where the subquery is used in the first callable against
+        a :class:`.Session`, the :class:`.Session` is also accepted::
+
+            sub_bq = self.bakery(lambda s: s.query(User.name))
+            sub_bq += lambda q: q.filter(User.id == Address.user_id).correlate(Address)
+
+            main_bq = self.bakery(
+                lambda s: s.query(Address.id, sub_bq.to_query(q).scalar_subquery())
+            )
+
+        :param query_or_session: a :class:`_query.Query` object or a class
+         :class:`.Session` object, that is assumed to be within the context
+         of an enclosing :class:`.BakedQuery` callable.
+
+
+         .. versionadded:: 1.3
+
+
+        """  # noqa: E501
+
+        if isinstance(query_or_session, Session):
+            session = query_or_session
+        elif isinstance(query_or_session, Query):
+            session = query_or_session.session
+            if session is None:
+                raise sa_exc.ArgumentError(
+                    "Given Query needs to be associated with a Session"
+                )
+        else:
+            raise TypeError(
+                "Query or Session object expected, got %r."
+                % type(query_or_session)
+            )
+        return self._as_query(session)
+
+    def _as_query(self, session):
+        query = self.steps[0](session)
+
+        for step in self.steps[1:]:
+            query = step(query)
+
+        return query
+
+
+class Result:
+    """Invokes a :class:`.BakedQuery` against a :class:`.Session`.
+
+    The :class:`_baked.Result` object is where the actual :class:`.query.Query`
+    object gets created, or retrieved from the cache,
+    against a target :class:`.Session`, and is then invoked for results.
+
+    """
+
+    __slots__ = "bq", "session", "_params", "_post_criteria"
+
+    def __init__(self, bq, session):
+        self.bq = bq
+        self.session = session
+        self._params = {}
+        self._post_criteria = []
+
+    def params(self, *args, **kw):
+        """Specify parameters to be replaced into the string SQL statement."""
+
+        if len(args) == 1:
+            kw.update(args[0])
+        elif len(args) > 0:
+            raise sa_exc.ArgumentError(
+                "params() takes zero or one positional argument, "
+                "which is a dictionary."
+            )
+        self._params.update(kw)
+        return self
+
+    def _using_post_criteria(self, fns):
+        if fns:
+            self._post_criteria.extend(fns)
+        return self
+
+    def with_post_criteria(self, fn):
+        """Add a criteria function that will be applied post-cache.
+
+        This adds a function that will be run against the
+        :class:`_query.Query` object after it is retrieved from the
+        cache.    This currently includes **only** the
+        :meth:`_query.Query.params` and :meth:`_query.Query.execution_options`
+        methods.
+
+        .. warning::  :meth:`_baked.Result.with_post_criteria`
+           functions are applied
+           to the :class:`_query.Query`
+           object **after** the query's SQL statement
+           object has been retrieved from the cache.   Only
+           :meth:`_query.Query.params` and
+           :meth:`_query.Query.execution_options`
+           methods should be used.
+
+
+        .. versionadded:: 1.2
+
+
+        """
+        return self._using_post_criteria([fn])
+
+    def _as_query(self):
+        q = self.bq._as_query(self.session).params(self._params)
+        for fn in self._post_criteria:
+            q = fn(q)
+        return q
+
+    def __str__(self):
+        return str(self._as_query())
+
+    def __iter__(self):
+        return self._iter().__iter__()
+
+    def _iter(self):
+        bq = self.bq
+
+        if not self.session.enable_baked_queries or bq._spoiled:
+            return self._as_query()._iter()
+
+        query, statement = bq._bakery.get(
+            bq._effective_key(self.session), (None, None)
+        )
+        if query is None:
+            query, statement = bq._bake(self.session)
+
+        if self._params:
+            q = query.params(self._params)
+        else:
+            q = query
+        for fn in self._post_criteria:
+            q = fn(q)
+
+        params = q._params
+        execution_options = dict(q._execution_options)
+        execution_options.update(
+            {
+                "_sa_orm_load_options": q.load_options,
+                "compiled_cache": bq._bakery,
+            }
+        )
+
+        result = self.session.execute(
+            statement, params, execution_options=execution_options
+        )
+        if result._attributes.get("is_single_entity", False):
+            result = result.scalars()
+
+        if result._attributes.get("filtered", False):
+            result = result.unique()
+
+        return result
+
+    def count(self):
+        """return the 'count'.
+
+        Equivalent to :meth:`_query.Query.count`.
+
+        Note this uses a subquery to ensure an accurate count regardless
+        of the structure of the original statement.
+
+        """
+
+        col = func.count(literal_column("*"))
+        bq = self.bq.with_criteria(lambda q: q._legacy_from_self(col))
+        return bq.for_session(self.session).params(self._params).scalar()
+
+    def scalar(self):
+        """Return the first element of the first result or None
+        if no rows present.  If multiple rows are returned,
+        raises MultipleResultsFound.
+
+        Equivalent to :meth:`_query.Query.scalar`.
+
+        """
+        try:
+            ret = self.one()
+            if not isinstance(ret, collections_abc.Sequence):
+                return ret
+            return ret[0]
+        except orm_exc.NoResultFound:
+            return None
+
+    def first(self):
+        """Return the first row.
+
+        Equivalent to :meth:`_query.Query.first`.
+
+        """
+
+        bq = self.bq.with_criteria(lambda q: q.slice(0, 1))
+        return (
+            bq.for_session(self.session)
+            .params(self._params)
+            ._using_post_criteria(self._post_criteria)
+            ._iter()
+            .first()
+        )
+
+    def one(self):
+        """Return exactly one result or raise an exception.
+
+        Equivalent to :meth:`_query.Query.one`.
+
+        """
+        return self._iter().one()
+
+    def one_or_none(self):
+        """Return one or zero results, or raise an exception for multiple
+        rows.
+
+        Equivalent to :meth:`_query.Query.one_or_none`.
+
+        """
+        return self._iter().one_or_none()
+
+    def all(self):
+        """Return all rows.
+
+        Equivalent to :meth:`_query.Query.all`.
+
+        """
+        return self._iter().all()
+
+    def get(self, ident):
+        """Retrieve an object based on identity.
+
+        Equivalent to :meth:`_query.Query.get`.
+
+        """
+
+        query = self.bq.steps[0](self.session)
+        return query._get_impl(ident, self._load_on_pk_identity)
+
+    def _load_on_pk_identity(self, session, query, primary_key_identity, **kw):
+        """Load the given primary key identity from the database."""
+
+        mapper = query._raw_columns[0]._annotations["parententity"]
+
+        _get_clause, _get_params = mapper._get_clause
+
+        def setup(query):
+            _lcl_get_clause = _get_clause
+            q = query._clone()
+            q._get_condition()
+            q._order_by = None
+
+            # None present in ident - turn those comparisons
+            # into "IS NULL"
+            if None in primary_key_identity:
+                nones = {
+                    _get_params[col].key
+                    for col, value in zip(
+                        mapper.primary_key, primary_key_identity
+                    )
+                    if value is None
+                }
+                _lcl_get_clause = sql_util.adapt_criterion_to_null(
+                    _lcl_get_clause, nones
+                )
+
+            # TODO: can mapper._get_clause be pre-adapted?
+            q._where_criteria = (
+                sql_util._deep_annotate(_lcl_get_clause, {"_orm_adapt": True}),
+            )
+
+            for fn in self._post_criteria:
+                q = fn(q)
+            return q
+
+        # cache the query against a key that includes
+        # which positions in the primary key are NULL
+        # (remember, we can map to an OUTER JOIN)
+        bq = self.bq
+
+        # add the clause we got from mapper._get_clause to the cache
+        # key so that if a race causes multiple calls to _get_clause,
+        # we've cached on ours
+        bq = bq._clone()
+        bq._cache_key += (_get_clause,)
+
+        bq = bq.with_criteria(
+            setup, tuple(elem is None for elem in primary_key_identity)
+        )
+
+        params = {
+            _get_params[primary_key].key: id_val
+            for id_val, primary_key in zip(
+                primary_key_identity, mapper.primary_key
+            )
+        }
+
+        result = list(bq.for_session(self.session).params(**params))
+        l = len(result)
+        if l > 1:
+            raise orm_exc.MultipleResultsFound()
+        elif l:
+            return result[0]
+        else:
+            return None
+
+
+bakery = BakedQuery.bakery
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/compiler.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/compiler.py
new file mode 100644
index 00000000..cc64477e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/compiler.py
@@ -0,0 +1,600 @@
+# ext/compiler.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+r"""Provides an API for creation of custom ClauseElements and compilers.
+
+Synopsis
+========
+
+Usage involves the creation of one or more
+:class:`~sqlalchemy.sql.expression.ClauseElement` subclasses and one or
+more callables defining its compilation::
+
+    from sqlalchemy.ext.compiler import compiles
+    from sqlalchemy.sql.expression import ColumnClause
+
+
+    class MyColumn(ColumnClause):
+        inherit_cache = True
+
+
+    @compiles(MyColumn)
+    def compile_mycolumn(element, compiler, **kw):
+        return "[%s]" % element.name
+
+Above, ``MyColumn`` extends :class:`~sqlalchemy.sql.expression.ColumnClause`,
+the base expression element for named column objects. The ``compiles``
+decorator registers itself with the ``MyColumn`` class so that it is invoked
+when the object is compiled to a string::
+
+    from sqlalchemy import select
+
+    s = select(MyColumn("x"), MyColumn("y"))
+    print(str(s))
+
+Produces:
+
+.. sourcecode:: sql
+
+    SELECT [x], [y]
+
+Dialect-specific compilation rules
+==================================
+
+Compilers can also be made dialect-specific. The appropriate compiler will be
+invoked for the dialect in use::
+
+    from sqlalchemy.schema import DDLElement
+
+
+    class AlterColumn(DDLElement):
+        inherit_cache = False
+
+        def __init__(self, column, cmd):
+            self.column = column
+            self.cmd = cmd
+
+
+    @compiles(AlterColumn)
+    def visit_alter_column(element, compiler, **kw):
+        return "ALTER COLUMN %s ..." % element.column.name
+
+
+    @compiles(AlterColumn, "postgresql")
+    def visit_alter_column(element, compiler, **kw):
+        return "ALTER TABLE %s ALTER COLUMN %s ..." % (
+            element.table.name,
+            element.column.name,
+        )
+
+The second ``visit_alter_table`` will be invoked when any ``postgresql``
+dialect is used.
+
+.. _compilerext_compiling_subelements:
+
+Compiling sub-elements of a custom expression construct
+=======================================================
+
+The ``compiler`` argument is the
+:class:`~sqlalchemy.engine.interfaces.Compiled` object in use. This object
+can be inspected for any information about the in-progress compilation,
+including ``compiler.dialect``, ``compiler.statement`` etc. The
+:class:`~sqlalchemy.sql.compiler.SQLCompiler` and
+:class:`~sqlalchemy.sql.compiler.DDLCompiler` both include a ``process()``
+method which can be used for compilation of embedded attributes::
+
+    from sqlalchemy.sql.expression import Executable, ClauseElement
+
+
+    class InsertFromSelect(Executable, ClauseElement):
+        inherit_cache = False
+
+        def __init__(self, table, select):
+            self.table = table
+            self.select = select
+
+
+    @compiles(InsertFromSelect)
+    def visit_insert_from_select(element, compiler, **kw):
+        return "INSERT INTO %s (%s)" % (
+            compiler.process(element.table, asfrom=True, **kw),
+            compiler.process(element.select, **kw),
+        )
+
+
+    insert = InsertFromSelect(t1, select(t1).where(t1.c.x > 5))
+    print(insert)
+
+Produces (formatted for readability):
+
+.. sourcecode:: sql
+
+    INSERT INTO mytable (
+        SELECT mytable.x, mytable.y, mytable.z
+        FROM mytable
+        WHERE mytable.x > :x_1
+    )
+
+.. note::
+
+    The above ``InsertFromSelect`` construct is only an example, this actual
+    functionality is already available using the
+    :meth:`_expression.Insert.from_select` method.
+
+
+Cross Compiling between SQL and DDL compilers
+---------------------------------------------
+
+SQL and DDL constructs are each compiled using different base compilers -
+``SQLCompiler`` and ``DDLCompiler``.   A common need is to access the
+compilation rules of SQL expressions from within a DDL expression. The
+``DDLCompiler`` includes an accessor ``sql_compiler`` for this reason, such as
+below where we generate a CHECK constraint that embeds a SQL expression::
+
+    @compiles(MyConstraint)
+    def compile_my_constraint(constraint, ddlcompiler, **kw):
+        kw["literal_binds"] = True
+        return "CONSTRAINT %s CHECK (%s)" % (
+            constraint.name,
+            ddlcompiler.sql_compiler.process(constraint.expression, **kw),
+        )
+
+Above, we add an additional flag to the process step as called by
+:meth:`.SQLCompiler.process`, which is the ``literal_binds`` flag.  This
+indicates that any SQL expression which refers to a :class:`.BindParameter`
+object or other "literal" object such as those which refer to strings or
+integers should be rendered **in-place**, rather than being referred to as
+a bound parameter;  when emitting DDL, bound parameters are typically not
+supported.
+
+
+Changing the default compilation of existing constructs
+=======================================================
+
+The compiler extension applies just as well to the existing constructs.  When
+overriding the compilation of a built in SQL construct, the @compiles
+decorator is invoked upon the appropriate class (be sure to use the class,
+i.e. ``Insert`` or ``Select``, instead of the creation function such
+as ``insert()`` or ``select()``).
+
+Within the new compilation function, to get at the "original" compilation
+routine, use the appropriate visit_XXX method - this
+because compiler.process() will call upon the overriding routine and cause
+an endless loop.   Such as, to add "prefix" to all insert statements::
+
+    from sqlalchemy.sql.expression import Insert
+
+
+    @compiles(Insert)
+    def prefix_inserts(insert, compiler, **kw):
+        return compiler.visit_insert(insert.prefix_with("some prefix"), **kw)
+
+The above compiler will prefix all INSERT statements with "some prefix" when
+compiled.
+
+.. _type_compilation_extension:
+
+Changing Compilation of Types
+=============================
+
+``compiler`` works for types, too, such as below where we implement the
+MS-SQL specific 'max' keyword for ``String``/``VARCHAR``::
+
+    @compiles(String, "mssql")
+    @compiles(VARCHAR, "mssql")
+    def compile_varchar(element, compiler, **kw):
+        if element.length == "max":
+            return "VARCHAR('max')"
+        else:
+            return compiler.visit_VARCHAR(element, **kw)
+
+
+    foo = Table("foo", metadata, Column("data", VARCHAR("max")))
+
+Subclassing Guidelines
+======================
+
+A big part of using the compiler extension is subclassing SQLAlchemy
+expression constructs. To make this easier, the expression and
+schema packages feature a set of "bases" intended for common tasks.
+A synopsis is as follows:
+
+* :class:`~sqlalchemy.sql.expression.ClauseElement` - This is the root
+  expression class. Any SQL expression can be derived from this base, and is
+  probably the best choice for longer constructs such as specialized INSERT
+  statements.
+
+* :class:`~sqlalchemy.sql.expression.ColumnElement` - The root of all
+  "column-like" elements. Anything that you'd place in the "columns" clause of
+  a SELECT statement (as well as order by and group by) can derive from this -
+  the object will automatically have Python "comparison" behavior.
+
+  :class:`~sqlalchemy.sql.expression.ColumnElement` classes want to have a
+  ``type`` member which is expression's return type.  This can be established
+  at the instance level in the constructor, or at the class level if its
+  generally constant::
+
+      class timestamp(ColumnElement):
+          type = TIMESTAMP()
+          inherit_cache = True
+
+* :class:`~sqlalchemy.sql.functions.FunctionElement` - This is a hybrid of a
+  ``ColumnElement`` and a "from clause" like object, and represents a SQL
+  function or stored procedure type of call. Since most databases support
+  statements along the line of "SELECT FROM <some function>"
+  ``FunctionElement`` adds in the ability to be used in the FROM clause of a
+  ``select()`` construct::
+
+      from sqlalchemy.sql.expression import FunctionElement
+
+
+      class coalesce(FunctionElement):
+          name = "coalesce"
+          inherit_cache = True
+
+
+      @compiles(coalesce)
+      def compile(element, compiler, **kw):
+          return "coalesce(%s)" % compiler.process(element.clauses, **kw)
+
+
+      @compiles(coalesce, "oracle")
+      def compile(element, compiler, **kw):
+          if len(element.clauses) > 2:
+              raise TypeError(
+                  "coalesce only supports two arguments on " "Oracle Database"
+              )
+          return "nvl(%s)" % compiler.process(element.clauses, **kw)
+
+* :class:`.ExecutableDDLElement` - The root of all DDL expressions,
+  like CREATE TABLE, ALTER TABLE, etc. Compilation of
+  :class:`.ExecutableDDLElement` subclasses is issued by a
+  :class:`.DDLCompiler` instead of a :class:`.SQLCompiler`.
+  :class:`.ExecutableDDLElement` can also be used as an event hook in
+  conjunction with event hooks like :meth:`.DDLEvents.before_create` and
+  :meth:`.DDLEvents.after_create`, allowing the construct to be invoked
+  automatically during CREATE TABLE and DROP TABLE sequences.
+
+  .. seealso::
+
+    :ref:`metadata_ddl_toplevel` - contains examples of associating
+    :class:`.DDL` objects (which are themselves :class:`.ExecutableDDLElement`
+    instances) with :class:`.DDLEvents` event hooks.
+
+* :class:`~sqlalchemy.sql.expression.Executable` - This is a mixin which
+  should be used with any expression class that represents a "standalone"
+  SQL statement that can be passed directly to an ``execute()`` method.  It
+  is already implicit within ``DDLElement`` and ``FunctionElement``.
+
+Most of the above constructs also respond to SQL statement caching.   A
+subclassed construct will want to define the caching behavior for the object,
+which usually means setting the flag ``inherit_cache`` to the value of
+``False`` or ``True``.  See the next section :ref:`compilerext_caching`
+for background.
+
+
+.. _compilerext_caching:
+
+Enabling Caching Support for Custom Constructs
+==============================================
+
+SQLAlchemy as of version 1.4 includes a
+:ref:`SQL compilation caching facility <sql_caching>` which will allow
+equivalent SQL constructs to cache their stringified form, along with other
+structural information used to fetch results from the statement.
+
+For reasons discussed at :ref:`caching_caveats`, the implementation of this
+caching system takes a conservative approach towards including custom SQL
+constructs and/or subclasses within the caching system.   This includes that
+any user-defined SQL constructs, including all the examples for this
+extension, will not participate in caching by default unless they positively
+assert that they are able to do so.  The :attr:`.HasCacheKey.inherit_cache`
+attribute when set to ``True`` at the class level of a specific subclass
+will indicate that instances of this class may be safely cached, using the
+cache key generation scheme of the immediate superclass.  This applies
+for example to the "synopsis" example indicated previously::
+
+    class MyColumn(ColumnClause):
+        inherit_cache = True
+
+
+    @compiles(MyColumn)
+    def compile_mycolumn(element, compiler, **kw):
+        return "[%s]" % element.name
+
+Above, the ``MyColumn`` class does not include any new state that
+affects its SQL compilation; the cache key of ``MyColumn`` instances will
+make use of that of the ``ColumnClause`` superclass, meaning it will take
+into account the class of the object (``MyColumn``), the string name and
+datatype of the object::
+
+    >>> MyColumn("some_name", String())._generate_cache_key()
+    CacheKey(
+        key=('0', <class '__main__.MyColumn'>,
+        'name', 'some_name',
+        'type', (<class 'sqlalchemy.sql.sqltypes.String'>,
+                 ('length', None), ('collation', None))
+    ), bindparams=[])
+
+For objects that are likely to be **used liberally as components within many
+larger statements**, such as :class:`_schema.Column` subclasses and custom SQL
+datatypes, it's important that **caching be enabled as much as possible**, as
+this may otherwise negatively affect performance.
+
+An example of an object that **does** contain state which affects its SQL
+compilation is the one illustrated at :ref:`compilerext_compiling_subelements`;
+this is an "INSERT FROM SELECT" construct that combines together a
+:class:`_schema.Table` as well as a :class:`_sql.Select` construct, each of
+which independently affect the SQL string generation of the construct.  For
+this class, the example illustrates that it simply does not participate in
+caching::
+
+    class InsertFromSelect(Executable, ClauseElement):
+        inherit_cache = False
+
+        def __init__(self, table, select):
+            self.table = table
+            self.select = select
+
+
+    @compiles(InsertFromSelect)
+    def visit_insert_from_select(element, compiler, **kw):
+        return "INSERT INTO %s (%s)" % (
+            compiler.process(element.table, asfrom=True, **kw),
+            compiler.process(element.select, **kw),
+        )
+
+While it is also possible that the above ``InsertFromSelect`` could be made to
+produce a cache key that is composed of that of the :class:`_schema.Table` and
+:class:`_sql.Select` components together, the API for this is not at the moment
+fully public. However, for an "INSERT FROM SELECT" construct, which is only
+used by itself for specific operations, caching is not as critical as in the
+previous example.
+
+For objects that are **used in relative isolation and are generally
+standalone**, such as custom :term:`DML` constructs like an "INSERT FROM
+SELECT", **caching is generally less critical** as the lack of caching for such
+a construct will have only localized implications for that specific operation.
+
+
+Further Examples
+================
+
+"UTC timestamp" function
+-------------------------
+
+A function that works like "CURRENT_TIMESTAMP" except applies the
+appropriate conversions so that the time is in UTC time.   Timestamps are best
+stored in relational databases as UTC, without time zones.   UTC so that your
+database doesn't think time has gone backwards in the hour when daylight
+savings ends, without timezones because timezones are like character
+encodings - they're best applied only at the endpoints of an application
+(i.e. convert to UTC upon user input, re-apply desired timezone upon display).
+
+For PostgreSQL and Microsoft SQL Server::
+
+    from sqlalchemy.sql import expression
+    from sqlalchemy.ext.compiler import compiles
+    from sqlalchemy.types import DateTime
+
+
+    class utcnow(expression.FunctionElement):
+        type = DateTime()
+        inherit_cache = True
+
+
+    @compiles(utcnow, "postgresql")
+    def pg_utcnow(element, compiler, **kw):
+        return "TIMEZONE('utc', CURRENT_TIMESTAMP)"
+
+
+    @compiles(utcnow, "mssql")
+    def ms_utcnow(element, compiler, **kw):
+        return "GETUTCDATE()"
+
+Example usage::
+
+    from sqlalchemy import Table, Column, Integer, String, DateTime, MetaData
+
+    metadata = MetaData()
+    event = Table(
+        "event",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("description", String(50), nullable=False),
+        Column("timestamp", DateTime, server_default=utcnow()),
+    )
+
+"GREATEST" function
+-------------------
+
+The "GREATEST" function is given any number of arguments and returns the one
+that is of the highest value - its equivalent to Python's ``max``
+function.  A SQL standard version versus a CASE based version which only
+accommodates two arguments::
+
+    from sqlalchemy.sql import expression, case
+    from sqlalchemy.ext.compiler import compiles
+    from sqlalchemy.types import Numeric
+
+
+    class greatest(expression.FunctionElement):
+        type = Numeric()
+        name = "greatest"
+        inherit_cache = True
+
+
+    @compiles(greatest)
+    def default_greatest(element, compiler, **kw):
+        return compiler.visit_function(element)
+
+
+    @compiles(greatest, "sqlite")
+    @compiles(greatest, "mssql")
+    @compiles(greatest, "oracle")
+    def case_greatest(element, compiler, **kw):
+        arg1, arg2 = list(element.clauses)
+        return compiler.process(case((arg1 > arg2, arg1), else_=arg2), **kw)
+
+Example usage::
+
+    Session.query(Account).filter(
+        greatest(Account.checking_balance, Account.savings_balance) > 10000
+    )
+
+"false" expression
+------------------
+
+Render a "false" constant expression, rendering as "0" on platforms that
+don't have a "false" constant::
+
+    from sqlalchemy.sql import expression
+    from sqlalchemy.ext.compiler import compiles
+
+
+    class sql_false(expression.ColumnElement):
+        inherit_cache = True
+
+
+    @compiles(sql_false)
+    def default_false(element, compiler, **kw):
+        return "false"
+
+
+    @compiles(sql_false, "mssql")
+    @compiles(sql_false, "mysql")
+    @compiles(sql_false, "oracle")
+    def int_false(element, compiler, **kw):
+        return "0"
+
+Example usage::
+
+    from sqlalchemy import select, union_all
+
+    exp = union_all(
+        select(users.c.name, sql_false().label("enrolled")),
+        select(customers.c.name, customers.c.enrolled),
+    )
+
+"""
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from .. import exc
+from ..sql import sqltypes
+
+if TYPE_CHECKING:
+    from ..sql.compiler import SQLCompiler
+
+_F = TypeVar("_F", bound=Callable[..., Any])
+
+
+def compiles(class_: Type[Any], *specs: str) -> Callable[[_F], _F]:
+    """Register a function as a compiler for a
+    given :class:`_expression.ClauseElement` type."""
+
+    def decorate(fn: _F) -> _F:
+        # get an existing @compiles handler
+        existing = class_.__dict__.get("_compiler_dispatcher", None)
+
+        # get the original handler.  All ClauseElement classes have one
+        # of these, but some TypeEngine classes will not.
+        existing_dispatch = getattr(class_, "_compiler_dispatch", None)
+
+        if not existing:
+            existing = _dispatcher()
+
+            if existing_dispatch:
+
+                def _wrap_existing_dispatch(
+                    element: Any, compiler: SQLCompiler, **kw: Any
+                ) -> Any:
+                    try:
+                        return existing_dispatch(element, compiler, **kw)
+                    except exc.UnsupportedCompilationError as uce:
+                        raise exc.UnsupportedCompilationError(
+                            compiler,
+                            type(element),
+                            message="%s construct has no default "
+                            "compilation handler." % type(element),
+                        ) from uce
+
+                existing.specs["default"] = _wrap_existing_dispatch
+
+            # TODO: why is the lambda needed ?
+            setattr(
+                class_,
+                "_compiler_dispatch",
+                lambda *arg, **kw: existing(*arg, **kw),
+            )
+            setattr(class_, "_compiler_dispatcher", existing)
+
+        if specs:
+            for s in specs:
+                existing.specs[s] = fn
+
+        else:
+            existing.specs["default"] = fn
+        return fn
+
+    return decorate
+
+
+def deregister(class_: Type[Any]) -> None:
+    """Remove all custom compilers associated with a given
+    :class:`_expression.ClauseElement` type.
+
+    """
+
+    if hasattr(class_, "_compiler_dispatcher"):
+        class_._compiler_dispatch = class_._original_compiler_dispatch
+        del class_._compiler_dispatcher
+
+
+class _dispatcher:
+    def __init__(self) -> None:
+        self.specs: Dict[str, Callable[..., Any]] = {}
+
+    def __call__(self, element: Any, compiler: SQLCompiler, **kw: Any) -> Any:
+        # TODO: yes, this could also switch off of DBAPI in use.
+        fn = self.specs.get(compiler.dialect.name, None)
+        if not fn:
+            try:
+                fn = self.specs["default"]
+            except KeyError as ke:
+                raise exc.UnsupportedCompilationError(
+                    compiler,
+                    type(element),
+                    message="%s construct has no default "
+                    "compilation handler." % type(element),
+                ) from ke
+
+        # if compilation includes add_to_result_map, collect add_to_result_map
+        # arguments from the user-defined callable, which are probably none
+        # because this is not public API.  if it wasn't called, then call it
+        # ourselves.
+        arm = kw.get("add_to_result_map", None)
+        if arm:
+            arm_collection = []
+            kw["add_to_result_map"] = lambda *args: arm_collection.append(args)
+
+        expr = fn(element, compiler, **kw)
+
+        if arm:
+            if not arm_collection:
+                arm_collection.append(
+                    (None, None, (element,), sqltypes.NULLTYPE)
+                )
+            for tup in arm_collection:
+                arm(*tup)
+        return expr
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__init__.py
new file mode 100644
index 00000000..0383f9d3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__init__.py
@@ -0,0 +1,65 @@
+# ext/declarative/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from .extensions import AbstractConcreteBase
+from .extensions import ConcreteBase
+from .extensions import DeferredReflection
+from ... import util
+from ...orm.decl_api import as_declarative as _as_declarative
+from ...orm.decl_api import declarative_base as _declarative_base
+from ...orm.decl_api import DeclarativeMeta
+from ...orm.decl_api import declared_attr
+from ...orm.decl_api import has_inherited_table as _has_inherited_table
+from ...orm.decl_api import synonym_for as _synonym_for
+
+
+@util.moved_20(
+    "The ``declarative_base()`` function is now available as "
+    ":func:`sqlalchemy.orm.declarative_base`."
+)
+def declarative_base(*arg, **kw):
+    return _declarative_base(*arg, **kw)
+
+
+@util.moved_20(
+    "The ``as_declarative()`` function is now available as "
+    ":func:`sqlalchemy.orm.as_declarative`"
+)
+def as_declarative(*arg, **kw):
+    return _as_declarative(*arg, **kw)
+
+
+@util.moved_20(
+    "The ``has_inherited_table()`` function is now available as "
+    ":func:`sqlalchemy.orm.has_inherited_table`."
+)
+def has_inherited_table(*arg, **kw):
+    return _has_inherited_table(*arg, **kw)
+
+
+@util.moved_20(
+    "The ``synonym_for()`` function is now available as "
+    ":func:`sqlalchemy.orm.synonym_for`"
+)
+def synonym_for(*arg, **kw):
+    return _synonym_for(*arg, **kw)
+
+
+__all__ = [
+    "declarative_base",
+    "synonym_for",
+    "has_inherited_table",
+    "instrument_declarative",
+    "declared_attr",
+    "as_declarative",
+    "ConcreteBase",
+    "AbstractConcreteBase",
+    "DeclarativeMeta",
+    "DeferredReflection",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e1b5ce73
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/extensions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/extensions.cpython-312.pyc
new file mode 100644
index 00000000..eb3b7abe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/__pycache__/extensions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/extensions.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/extensions.py
new file mode 100644
index 00000000..3dc6bf69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/declarative/extensions.py
@@ -0,0 +1,564 @@
+# ext/declarative/extensions.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Public API functions and helpers for declarative."""
+from __future__ import annotations
+
+import collections
+import contextlib
+from typing import Any
+from typing import Callable
+from typing import TYPE_CHECKING
+from typing import Union
+
+from ... import exc as sa_exc
+from ...engine import Connection
+from ...engine import Engine
+from ...orm import exc as orm_exc
+from ...orm import relationships
+from ...orm.base import _mapper_or_none
+from ...orm.clsregistry import _resolver
+from ...orm.decl_base import _DeferredMapperConfig
+from ...orm.util import polymorphic_union
+from ...schema import Table
+from ...util import OrderedDict
+
+if TYPE_CHECKING:
+    from ...sql.schema import MetaData
+
+
+class ConcreteBase:
+    """A helper class for 'concrete' declarative mappings.
+
+    :class:`.ConcreteBase` will use the :func:`.polymorphic_union`
+    function automatically, against all tables mapped as a subclass
+    to this class.   The function is called via the
+    ``__declare_last__()`` function, which is essentially
+    a hook for the :meth:`.after_configured` event.
+
+    :class:`.ConcreteBase` produces a mapped
+    table for the class itself.  Compare to :class:`.AbstractConcreteBase`,
+    which does not.
+
+    Example::
+
+        from sqlalchemy.ext.declarative import ConcreteBase
+
+
+        class Employee(ConcreteBase, Base):
+            __tablename__ = "employee"
+            employee_id = Column(Integer, primary_key=True)
+            name = Column(String(50))
+            __mapper_args__ = {
+                "polymorphic_identity": "employee",
+                "concrete": True,
+            }
+
+
+        class Manager(Employee):
+            __tablename__ = "manager"
+            employee_id = Column(Integer, primary_key=True)
+            name = Column(String(50))
+            manager_data = Column(String(40))
+            __mapper_args__ = {
+                "polymorphic_identity": "manager",
+                "concrete": True,
+            }
+
+    The name of the discriminator column used by :func:`.polymorphic_union`
+    defaults to the name ``type``.  To suit the use case of a mapping where an
+    actual column in a mapped table is already named ``type``, the
+    discriminator name can be configured by setting the
+    ``_concrete_discriminator_name`` attribute::
+
+        class Employee(ConcreteBase, Base):
+            _concrete_discriminator_name = "_concrete_discriminator"
+
+    .. versionadded:: 1.3.19 Added the ``_concrete_discriminator_name``
+       attribute to :class:`_declarative.ConcreteBase` so that the
+       virtual discriminator column name can be customized.
+
+    .. versionchanged:: 1.4.2 The ``_concrete_discriminator_name`` attribute
+       need only be placed on the basemost class to take correct effect for
+       all subclasses.   An explicit error message is now raised if the
+       mapped column names conflict with the discriminator name, whereas
+       in the 1.3.x series there would be some warnings and then a non-useful
+       query would be generated.
+
+    .. seealso::
+
+        :class:`.AbstractConcreteBase`
+
+        :ref:`concrete_inheritance`
+
+
+    """
+
+    @classmethod
+    def _create_polymorphic_union(cls, mappers, discriminator_name):
+        return polymorphic_union(
+            OrderedDict(
+                (mp.polymorphic_identity, mp.local_table) for mp in mappers
+            ),
+            discriminator_name,
+            "pjoin",
+        )
+
+    @classmethod
+    def __declare_first__(cls):
+        m = cls.__mapper__
+        if m.with_polymorphic:
+            return
+
+        discriminator_name = (
+            getattr(cls, "_concrete_discriminator_name", None) or "type"
+        )
+
+        mappers = list(m.self_and_descendants)
+        pjoin = cls._create_polymorphic_union(mappers, discriminator_name)
+        m._set_with_polymorphic(("*", pjoin))
+        m._set_polymorphic_on(pjoin.c[discriminator_name])
+
+
+class AbstractConcreteBase(ConcreteBase):
+    """A helper class for 'concrete' declarative mappings.
+
+    :class:`.AbstractConcreteBase` will use the :func:`.polymorphic_union`
+    function automatically, against all tables mapped as a subclass
+    to this class.   The function is called via the
+    ``__declare_first__()`` function, which is essentially
+    a hook for the :meth:`.before_configured` event.
+
+    :class:`.AbstractConcreteBase` applies :class:`_orm.Mapper` for its
+    immediately inheriting class, as would occur for any other
+    declarative mapped class. However, the :class:`_orm.Mapper` is not
+    mapped to any particular :class:`.Table` object.  Instead, it's
+    mapped directly to the "polymorphic" selectable produced by
+    :func:`.polymorphic_union`, and performs no persistence operations on its
+    own.  Compare to :class:`.ConcreteBase`, which maps its
+    immediately inheriting class to an actual
+    :class:`.Table` that stores rows directly.
+
+    .. note::
+
+        The :class:`.AbstractConcreteBase` delays the mapper creation of the
+        base class until all the subclasses have been defined,
+        as it needs to create a mapping against a selectable that will include
+        all subclass tables.  In order to achieve this, it waits for the
+        **mapper configuration event** to occur, at which point it scans
+        through all the configured subclasses and sets up a mapping that will
+        query against all subclasses at once.
+
+        While this event is normally invoked automatically, in the case of
+        :class:`.AbstractConcreteBase`, it may be necessary to invoke it
+        explicitly after **all** subclass mappings are defined, if the first
+        operation is to be a query against this base class. To do so, once all
+        the desired classes have been configured, the
+        :meth:`_orm.registry.configure` method on the :class:`_orm.registry`
+        in use can be invoked, which is available in relation to a particular
+        declarative base class::
+
+            Base.registry.configure()
+
+    Example::
+
+        from sqlalchemy.orm import DeclarativeBase
+        from sqlalchemy.ext.declarative import AbstractConcreteBase
+
+
+        class Base(DeclarativeBase):
+            pass
+
+
+        class Employee(AbstractConcreteBase, Base):
+            pass
+
+
+        class Manager(Employee):
+            __tablename__ = "manager"
+            employee_id = Column(Integer, primary_key=True)
+            name = Column(String(50))
+            manager_data = Column(String(40))
+
+            __mapper_args__ = {
+                "polymorphic_identity": "manager",
+                "concrete": True,
+            }
+
+
+        Base.registry.configure()
+
+    The abstract base class is handled by declarative in a special way;
+    at class configuration time, it behaves like a declarative mixin
+    or an ``__abstract__`` base class.   Once classes are configured
+    and mappings are produced, it then gets mapped itself, but
+    after all of its descendants.  This is a very unique system of mapping
+    not found in any other SQLAlchemy API feature.
+
+    Using this approach, we can specify columns and properties
+    that will take place on mapped subclasses, in the way that
+    we normally do as in :ref:`declarative_mixins`::
+
+        from sqlalchemy.ext.declarative import AbstractConcreteBase
+
+
+        class Company(Base):
+            __tablename__ = "company"
+            id = Column(Integer, primary_key=True)
+
+
+        class Employee(AbstractConcreteBase, Base):
+            strict_attrs = True
+
+            employee_id = Column(Integer, primary_key=True)
+
+            @declared_attr
+            def company_id(cls):
+                return Column(ForeignKey("company.id"))
+
+            @declared_attr
+            def company(cls):
+                return relationship("Company")
+
+
+        class Manager(Employee):
+            __tablename__ = "manager"
+
+            name = Column(String(50))
+            manager_data = Column(String(40))
+
+            __mapper_args__ = {
+                "polymorphic_identity": "manager",
+                "concrete": True,
+            }
+
+
+        Base.registry.configure()
+
+    When we make use of our mappings however, both ``Manager`` and
+    ``Employee`` will have an independently usable ``.company`` attribute::
+
+        session.execute(select(Employee).filter(Employee.company.has(id=5)))
+
+    :param strict_attrs: when specified on the base class, "strict" attribute
+     mode is enabled which attempts to limit ORM mapped attributes on the
+     base class to only those that are immediately present, while still
+     preserving "polymorphic" loading behavior.
+
+     .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`.ConcreteBase`
+
+        :ref:`concrete_inheritance`
+
+        :ref:`abstract_concrete_base`
+
+    """
+
+    __no_table__ = True
+
+    @classmethod
+    def __declare_first__(cls):
+        cls._sa_decl_prepare_nocascade()
+
+    @classmethod
+    def _sa_decl_prepare_nocascade(cls):
+        if getattr(cls, "__mapper__", None):
+            return
+
+        to_map = _DeferredMapperConfig.config_for_cls(cls)
+
+        # can't rely on 'self_and_descendants' here
+        # since technically an immediate subclass
+        # might not be mapped, but a subclass
+        # may be.
+        mappers = []
+        stack = list(cls.__subclasses__())
+        while stack:
+            klass = stack.pop()
+            stack.extend(klass.__subclasses__())
+            mn = _mapper_or_none(klass)
+            if mn is not None:
+                mappers.append(mn)
+
+        discriminator_name = (
+            getattr(cls, "_concrete_discriminator_name", None) or "type"
+        )
+        pjoin = cls._create_polymorphic_union(mappers, discriminator_name)
+
+        # For columns that were declared on the class, these
+        # are normally ignored with the "__no_table__" mapping,
+        # unless they have a different attribute key vs. col name
+        # and are in the properties argument.
+        # In that case, ensure we update the properties entry
+        # to the correct column from the pjoin target table.
+        declared_cols = set(to_map.declared_columns)
+        declared_col_keys = {c.key for c in declared_cols}
+        for k, v in list(to_map.properties.items()):
+            if v in declared_cols:
+                to_map.properties[k] = pjoin.c[v.key]
+                declared_col_keys.remove(v.key)
+
+        to_map.local_table = pjoin
+
+        strict_attrs = cls.__dict__.get("strict_attrs", False)
+
+        m_args = to_map.mapper_args_fn or dict
+
+        def mapper_args():
+            args = m_args()
+            args["polymorphic_on"] = pjoin.c[discriminator_name]
+            args["polymorphic_abstract"] = True
+            if strict_attrs:
+                args["include_properties"] = (
+                    set(pjoin.primary_key)
+                    | declared_col_keys
+                    | {discriminator_name}
+                )
+                args["with_polymorphic"] = ("*", pjoin)
+            return args
+
+        to_map.mapper_args_fn = mapper_args
+
+        to_map.map()
+
+        stack = [cls]
+        while stack:
+            scls = stack.pop(0)
+            stack.extend(scls.__subclasses__())
+            sm = _mapper_or_none(scls)
+            if sm and sm.concrete and sm.inherits is None:
+                for sup_ in scls.__mro__[1:]:
+                    sup_sm = _mapper_or_none(sup_)
+                    if sup_sm:
+                        sm._set_concrete_base(sup_sm)
+                        break
+
+    @classmethod
+    def _sa_raise_deferred_config(cls):
+        raise orm_exc.UnmappedClassError(
+            cls,
+            msg="Class %s is a subclass of AbstractConcreteBase and "
+            "has a mapping pending until all subclasses are defined. "
+            "Call the sqlalchemy.orm.configure_mappers() function after "
+            "all subclasses have been defined to "
+            "complete the mapping of this class."
+            % orm_exc._safe_cls_name(cls),
+        )
+
+
+class DeferredReflection:
+    """A helper class for construction of mappings based on
+    a deferred reflection step.
+
+    Normally, declarative can be used with reflection by
+    setting a :class:`_schema.Table` object using autoload_with=engine
+    as the ``__table__`` attribute on a declarative class.
+    The caveat is that the :class:`_schema.Table` must be fully
+    reflected, or at the very least have a primary key column,
+    at the point at which a normal declarative mapping is
+    constructed, meaning the :class:`_engine.Engine` must be available
+    at class declaration time.
+
+    The :class:`.DeferredReflection` mixin moves the construction
+    of mappers to be at a later point, after a specific
+    method is called which first reflects all :class:`_schema.Table`
+    objects created so far.   Classes can define it as such::
+
+        from sqlalchemy.ext.declarative import declarative_base
+        from sqlalchemy.ext.declarative import DeferredReflection
+
+        Base = declarative_base()
+
+
+        class MyClass(DeferredReflection, Base):
+            __tablename__ = "mytable"
+
+    Above, ``MyClass`` is not yet mapped.   After a series of
+    classes have been defined in the above fashion, all tables
+    can be reflected and mappings created using
+    :meth:`.prepare`::
+
+        engine = create_engine("someengine://...")
+        DeferredReflection.prepare(engine)
+
+    The :class:`.DeferredReflection` mixin can be applied to individual
+    classes, used as the base for the declarative base itself,
+    or used in a custom abstract class.   Using an abstract base
+    allows that only a subset of classes to be prepared for a
+    particular prepare step, which is necessary for applications
+    that use more than one engine.  For example, if an application
+    has two engines, you might use two bases, and prepare each
+    separately, e.g.::
+
+        class ReflectedOne(DeferredReflection, Base):
+            __abstract__ = True
+
+
+        class ReflectedTwo(DeferredReflection, Base):
+            __abstract__ = True
+
+
+        class MyClass(ReflectedOne):
+            __tablename__ = "mytable"
+
+
+        class MyOtherClass(ReflectedOne):
+            __tablename__ = "myothertable"
+
+
+        class YetAnotherClass(ReflectedTwo):
+            __tablename__ = "yetanothertable"
+
+
+        # ... etc.
+
+    Above, the class hierarchies for ``ReflectedOne`` and
+    ``ReflectedTwo`` can be configured separately::
+
+        ReflectedOne.prepare(engine_one)
+        ReflectedTwo.prepare(engine_two)
+
+    .. seealso::
+
+        :ref:`orm_declarative_reflected_deferred_reflection` - in the
+        :ref:`orm_declarative_table_config_toplevel` section.
+
+    """
+
+    @classmethod
+    def prepare(
+        cls, bind: Union[Engine, Connection], **reflect_kw: Any
+    ) -> None:
+        r"""Reflect all :class:`_schema.Table` objects for all current
+        :class:`.DeferredReflection` subclasses
+
+        :param bind: :class:`_engine.Engine` or :class:`_engine.Connection`
+         instance
+
+         ..versionchanged:: 2.0.16 a :class:`_engine.Connection` is also
+         accepted.
+
+        :param \**reflect_kw: additional keyword arguments passed to
+         :meth:`_schema.MetaData.reflect`, such as
+         :paramref:`_schema.MetaData.reflect.views`.
+
+         .. versionadded:: 2.0.16
+
+        """
+
+        to_map = _DeferredMapperConfig.classes_for_base(cls)
+
+        metadata_to_table = collections.defaultdict(set)
+
+        # first collect the primary __table__ for each class into a
+        # collection of metadata/schemaname -> table names
+        for thingy in to_map:
+            if thingy.local_table is not None:
+                metadata_to_table[
+                    (thingy.local_table.metadata, thingy.local_table.schema)
+                ].add(thingy.local_table.name)
+
+        # then reflect all those tables into their metadatas
+
+        if isinstance(bind, Connection):
+            conn = bind
+            ctx = contextlib.nullcontext(enter_result=conn)
+        elif isinstance(bind, Engine):
+            ctx = bind.connect()
+        else:
+            raise sa_exc.ArgumentError(
+                f"Expected Engine or Connection, got {bind!r}"
+            )
+
+        with ctx as conn:
+            for (metadata, schema), table_names in metadata_to_table.items():
+                metadata.reflect(
+                    conn,
+                    only=table_names,
+                    schema=schema,
+                    extend_existing=True,
+                    autoload_replace=False,
+                    **reflect_kw,
+                )
+
+            metadata_to_table.clear()
+
+            # .map() each class, then go through relationships and look
+            # for secondary
+            for thingy in to_map:
+                thingy.map()
+
+                mapper = thingy.cls.__mapper__
+                metadata = mapper.class_.metadata
+
+                for rel in mapper._props.values():
+                    if (
+                        isinstance(rel, relationships.RelationshipProperty)
+                        and rel._init_args.secondary._is_populated()
+                    ):
+                        secondary_arg = rel._init_args.secondary
+
+                        if isinstance(secondary_arg.argument, Table):
+                            secondary_table = secondary_arg.argument
+                            metadata_to_table[
+                                (
+                                    secondary_table.metadata,
+                                    secondary_table.schema,
+                                )
+                            ].add(secondary_table.name)
+                        elif isinstance(secondary_arg.argument, str):
+                            _, resolve_arg = _resolver(rel.parent.class_, rel)
+
+                            resolver = resolve_arg(
+                                secondary_arg.argument, True
+                            )
+                            metadata_to_table[
+                                (metadata, thingy.local_table.schema)
+                            ].add(secondary_arg.argument)
+
+                            resolver._resolvers += (
+                                cls._sa_deferred_table_resolver(metadata),
+                            )
+
+                            secondary_arg.argument = resolver()
+
+            for (metadata, schema), table_names in metadata_to_table.items():
+                metadata.reflect(
+                    conn,
+                    only=table_names,
+                    schema=schema,
+                    extend_existing=True,
+                    autoload_replace=False,
+                )
+
+    @classmethod
+    def _sa_deferred_table_resolver(
+        cls, metadata: MetaData
+    ) -> Callable[[str], Table]:
+        def _resolve(key: str) -> Table:
+            # reflection has already occurred so this Table would have
+            # its contents already
+            return Table(key, metadata)
+
+        return _resolve
+
+    _sa_decl_prepare = True
+
+    @classmethod
+    def _sa_raise_deferred_config(cls):
+        raise orm_exc.UnmappedClassError(
+            cls,
+            msg="Class %s is a subclass of DeferredReflection.  "
+            "Mappings are not produced until the .prepare() "
+            "method is called on the class hierarchy."
+            % orm_exc._safe_cls_name(cls),
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/horizontal_shard.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/horizontal_shard.py
new file mode 100644
index 00000000..3ea3304e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/horizontal_shard.py
@@ -0,0 +1,478 @@
+# ext/horizontal_shard.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Horizontal sharding support.
+
+Defines a rudimental 'horizontal sharding' system which allows a Session to
+distribute queries and persistence operations across multiple databases.
+
+For a usage example, see the :ref:`examples_sharding` example included in
+the source distribution.
+
+.. deepalchemy:: The horizontal sharding extension is an advanced feature,
+   involving a complex statement -> database interaction as well as
+   use of semi-public APIs for non-trivial cases.   Simpler approaches to
+   refering to multiple database "shards", most commonly using a distinct
+   :class:`_orm.Session` per "shard", should always be considered first
+   before using this more complex and less-production-tested system.
+
+
+
+"""
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Iterable
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import event
+from .. import exc
+from .. import inspect
+from .. import util
+from ..orm import PassiveFlag
+from ..orm._typing import OrmExecuteOptionsParameter
+from ..orm.interfaces import ORMOption
+from ..orm.mapper import Mapper
+from ..orm.query import Query
+from ..orm.session import _BindArguments
+from ..orm.session import _PKIdentityArgument
+from ..orm.session import Session
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from ..engine.base import Connection
+    from ..engine.base import Engine
+    from ..engine.base import OptionEngine
+    from ..engine.result import IteratorResult
+    from ..engine.result import Result
+    from ..orm import LoaderCallableStatus
+    from ..orm._typing import _O
+    from ..orm.bulk_persistence import BulkUDCompileState
+    from ..orm.context import QueryContext
+    from ..orm.session import _EntityBindKey
+    from ..orm.session import _SessionBind
+    from ..orm.session import ORMExecuteState
+    from ..orm.state import InstanceState
+    from ..sql import Executable
+    from ..sql._typing import _TP
+    from ..sql.elements import ClauseElement
+
+__all__ = ["ShardedSession", "ShardedQuery"]
+
+_T = TypeVar("_T", bound=Any)
+
+
+ShardIdentifier = str
+
+
+class ShardChooser(Protocol):
+    def __call__(
+        self,
+        mapper: Optional[Mapper[_T]],
+        instance: Any,
+        clause: Optional[ClauseElement],
+    ) -> Any: ...
+
+
+class IdentityChooser(Protocol):
+    def __call__(
+        self,
+        mapper: Mapper[_T],
+        primary_key: _PKIdentityArgument,
+        *,
+        lazy_loaded_from: Optional[InstanceState[Any]],
+        execution_options: OrmExecuteOptionsParameter,
+        bind_arguments: _BindArguments,
+        **kw: Any,
+    ) -> Any: ...
+
+
+class ShardedQuery(Query[_T]):
+    """Query class used with :class:`.ShardedSession`.
+
+    .. legacy:: The :class:`.ShardedQuery` is a subclass of the legacy
+       :class:`.Query` class.   The :class:`.ShardedSession` now supports
+       2.0 style execution via the :meth:`.ShardedSession.execute` method.
+
+    """
+
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        super().__init__(*args, **kwargs)
+        assert isinstance(self.session, ShardedSession)
+
+        self.identity_chooser = self.session.identity_chooser
+        self.execute_chooser = self.session.execute_chooser
+        self._shard_id = None
+
+    def set_shard(self, shard_id: ShardIdentifier) -> Self:
+        """Return a new query, limited to a single shard ID.
+
+        All subsequent operations with the returned query will
+        be against the single shard regardless of other state.
+
+        The shard_id can be passed for a 2.0 style execution to the
+        bind_arguments dictionary of :meth:`.Session.execute`::
+
+            results = session.execute(stmt, bind_arguments={"shard_id": "my_shard"})
+
+        """  # noqa: E501
+        return self.execution_options(_sa_shard_id=shard_id)
+
+
+class ShardedSession(Session):
+    shard_chooser: ShardChooser
+    identity_chooser: IdentityChooser
+    execute_chooser: Callable[[ORMExecuteState], Iterable[Any]]
+
+    def __init__(
+        self,
+        shard_chooser: ShardChooser,
+        identity_chooser: Optional[IdentityChooser] = None,
+        execute_chooser: Optional[
+            Callable[[ORMExecuteState], Iterable[Any]]
+        ] = None,
+        shards: Optional[Dict[str, Any]] = None,
+        query_cls: Type[Query[_T]] = ShardedQuery,
+        *,
+        id_chooser: Optional[
+            Callable[[Query[_T], Iterable[_T]], Iterable[Any]]
+        ] = None,
+        query_chooser: Optional[Callable[[Executable], Iterable[Any]]] = None,
+        **kwargs: Any,
+    ) -> None:
+        """Construct a ShardedSession.
+
+        :param shard_chooser: A callable which, passed a Mapper, a mapped
+          instance, and possibly a SQL clause, returns a shard ID.  This id
+          may be based off of the attributes present within the object, or on
+          some round-robin scheme. If the scheme is based on a selection, it
+          should set whatever state on the instance to mark it in the future as
+          participating in that shard.
+
+        :param identity_chooser: A callable, passed a Mapper and primary key
+         argument, which should return a list of shard ids where this
+         primary key might reside.
+
+          .. versionchanged:: 2.0  The ``identity_chooser`` parameter
+             supersedes the ``id_chooser`` parameter.
+
+        :param execute_chooser: For a given :class:`.ORMExecuteState`,
+          returns the list of shard_ids
+          where the query should be issued.  Results from all shards returned
+          will be combined together into a single listing.
+
+          .. versionchanged:: 1.4  The ``execute_chooser`` parameter
+             supersedes the ``query_chooser`` parameter.
+
+        :param shards: A dictionary of string shard names
+          to :class:`~sqlalchemy.engine.Engine` objects.
+
+        """
+        super().__init__(query_cls=query_cls, **kwargs)
+
+        event.listen(
+            self, "do_orm_execute", execute_and_instances, retval=True
+        )
+        self.shard_chooser = shard_chooser
+
+        if id_chooser:
+            _id_chooser = id_chooser
+            util.warn_deprecated(
+                "The ``id_chooser`` parameter is deprecated; "
+                "please use ``identity_chooser``.",
+                "2.0",
+            )
+
+            def _legacy_identity_chooser(
+                mapper: Mapper[_T],
+                primary_key: _PKIdentityArgument,
+                *,
+                lazy_loaded_from: Optional[InstanceState[Any]],
+                execution_options: OrmExecuteOptionsParameter,
+                bind_arguments: _BindArguments,
+                **kw: Any,
+            ) -> Any:
+                q = self.query(mapper)
+                if lazy_loaded_from:
+                    q = q._set_lazyload_from(lazy_loaded_from)
+                return _id_chooser(q, primary_key)
+
+            self.identity_chooser = _legacy_identity_chooser
+        elif identity_chooser:
+            self.identity_chooser = identity_chooser
+        else:
+            raise exc.ArgumentError(
+                "identity_chooser or id_chooser is required"
+            )
+
+        if query_chooser:
+            _query_chooser = query_chooser
+            util.warn_deprecated(
+                "The ``query_chooser`` parameter is deprecated; "
+                "please use ``execute_chooser``.",
+                "1.4",
+            )
+            if execute_chooser:
+                raise exc.ArgumentError(
+                    "Can't pass query_chooser and execute_chooser "
+                    "at the same time."
+                )
+
+            def _default_execute_chooser(
+                orm_context: ORMExecuteState,
+            ) -> Iterable[Any]:
+                return _query_chooser(orm_context.statement)
+
+            if execute_chooser is None:
+                execute_chooser = _default_execute_chooser
+
+        if execute_chooser is None:
+            raise exc.ArgumentError(
+                "execute_chooser or query_chooser is required"
+            )
+        self.execute_chooser = execute_chooser
+        self.__shards: Dict[ShardIdentifier, _SessionBind] = {}
+        if shards is not None:
+            for k in shards:
+                self.bind_shard(k, shards[k])
+
+    def _identity_lookup(
+        self,
+        mapper: Mapper[_O],
+        primary_key_identity: Union[Any, Tuple[Any, ...]],
+        identity_token: Optional[Any] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+        lazy_loaded_from: Optional[InstanceState[Any]] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Union[Optional[_O], LoaderCallableStatus]:
+        """override the default :meth:`.Session._identity_lookup` method so
+        that we search for a given non-token primary key identity across all
+        possible identity tokens (e.g. shard ids).
+
+        .. versionchanged:: 1.4  Moved :meth:`.Session._identity_lookup` from
+           the :class:`_query.Query` object to the :class:`.Session`.
+
+        """
+
+        if identity_token is not None:
+            obj = super()._identity_lookup(
+                mapper,
+                primary_key_identity,
+                identity_token=identity_token,
+                **kw,
+            )
+
+            return obj
+        else:
+            for shard_id in self.identity_chooser(
+                mapper,
+                primary_key_identity,
+                lazy_loaded_from=lazy_loaded_from,
+                execution_options=execution_options,
+                bind_arguments=dict(bind_arguments) if bind_arguments else {},
+            ):
+                obj2 = super()._identity_lookup(
+                    mapper,
+                    primary_key_identity,
+                    identity_token=shard_id,
+                    lazy_loaded_from=lazy_loaded_from,
+                    **kw,
+                )
+                if obj2 is not None:
+                    return obj2
+
+            return None
+
+    def _choose_shard_and_assign(
+        self,
+        mapper: Optional[_EntityBindKey[_O]],
+        instance: Any,
+        **kw: Any,
+    ) -> Any:
+        if instance is not None:
+            state = inspect(instance)
+            if state.key:
+                token = state.key[2]
+                assert token is not None
+                return token
+            elif state.identity_token:
+                return state.identity_token
+
+        assert isinstance(mapper, Mapper)
+        shard_id = self.shard_chooser(mapper, instance, **kw)
+        if instance is not None:
+            state.identity_token = shard_id
+        return shard_id
+
+    def connection_callable(
+        self,
+        mapper: Optional[Mapper[_T]] = None,
+        instance: Optional[Any] = None,
+        shard_id: Optional[ShardIdentifier] = None,
+        **kw: Any,
+    ) -> Connection:
+        """Provide a :class:`_engine.Connection` to use in the unit of work
+        flush process.
+
+        """
+
+        if shard_id is None:
+            shard_id = self._choose_shard_and_assign(mapper, instance)
+
+        if self.in_transaction():
+            trans = self.get_transaction()
+            assert trans is not None
+            return trans.connection(mapper, shard_id=shard_id)
+        else:
+            bind = self.get_bind(
+                mapper=mapper, shard_id=shard_id, instance=instance
+            )
+
+            if isinstance(bind, Engine):
+                return bind.connect(**kw)
+            else:
+                assert isinstance(bind, Connection)
+                return bind
+
+    def get_bind(
+        self,
+        mapper: Optional[_EntityBindKey[_O]] = None,
+        *,
+        shard_id: Optional[ShardIdentifier] = None,
+        instance: Optional[Any] = None,
+        clause: Optional[ClauseElement] = None,
+        **kw: Any,
+    ) -> _SessionBind:
+        if shard_id is None:
+            shard_id = self._choose_shard_and_assign(
+                mapper, instance=instance, clause=clause
+            )
+            assert shard_id is not None
+        return self.__shards[shard_id]
+
+    def bind_shard(
+        self, shard_id: ShardIdentifier, bind: Union[Engine, OptionEngine]
+    ) -> None:
+        self.__shards[shard_id] = bind
+
+
+class set_shard_id(ORMOption):
+    """a loader option for statements to apply a specific shard id to the
+    primary query as well as for additional relationship and column
+    loaders.
+
+    The :class:`_horizontal.set_shard_id` option may be applied using
+    the :meth:`_sql.Executable.options` method of any executable statement::
+
+        stmt = (
+            select(MyObject)
+            .where(MyObject.name == "some name")
+            .options(set_shard_id("shard1"))
+        )
+
+    Above, the statement when invoked will limit to the "shard1" shard
+    identifier for the primary query as well as for all relationship and
+    column loading strategies, including eager loaders such as
+    :func:`_orm.selectinload`, deferred column loaders like :func:`_orm.defer`,
+    and the lazy relationship loader :func:`_orm.lazyload`.
+
+    In this way, the :class:`_horizontal.set_shard_id` option has much wider
+    scope than using the "shard_id" argument within the
+    :paramref:`_orm.Session.execute.bind_arguments` dictionary.
+
+
+    .. versionadded:: 2.0.0
+
+    """
+
+    __slots__ = ("shard_id", "propagate_to_loaders")
+
+    def __init__(
+        self, shard_id: ShardIdentifier, propagate_to_loaders: bool = True
+    ):
+        """Construct a :class:`_horizontal.set_shard_id` option.
+
+        :param shard_id: shard identifier
+        :param propagate_to_loaders: if left at its default of ``True``, the
+         shard option will take place for lazy loaders such as
+         :func:`_orm.lazyload` and :func:`_orm.defer`; if False, the option
+         will not be propagated to loaded objects. Note that :func:`_orm.defer`
+         always limits to the shard_id of the parent row in any case, so the
+         parameter only has a net effect on the behavior of the
+         :func:`_orm.lazyload` strategy.
+
+        """
+        self.shard_id = shard_id
+        self.propagate_to_loaders = propagate_to_loaders
+
+
+def execute_and_instances(
+    orm_context: ORMExecuteState,
+) -> Union[Result[_T], IteratorResult[_TP]]:
+    active_options: Union[
+        None,
+        QueryContext.default_load_options,
+        Type[QueryContext.default_load_options],
+        BulkUDCompileState.default_update_options,
+        Type[BulkUDCompileState.default_update_options],
+    ]
+
+    if orm_context.is_select:
+        active_options = orm_context.load_options
+
+    elif orm_context.is_update or orm_context.is_delete:
+        active_options = orm_context.update_delete_options
+    else:
+        active_options = None
+
+    session = orm_context.session
+    assert isinstance(session, ShardedSession)
+
+    def iter_for_shard(
+        shard_id: ShardIdentifier,
+    ) -> Union[Result[_T], IteratorResult[_TP]]:
+        bind_arguments = dict(orm_context.bind_arguments)
+        bind_arguments["shard_id"] = shard_id
+
+        orm_context.update_execution_options(identity_token=shard_id)
+        return orm_context.invoke_statement(bind_arguments=bind_arguments)
+
+    for orm_opt in orm_context._non_compile_orm_options:
+        # TODO: if we had an ORMOption that gets applied at ORM statement
+        # execution time, that would allow this to be more generalized.
+        # for now just iterate and look for our options
+        if isinstance(orm_opt, set_shard_id):
+            shard_id = orm_opt.shard_id
+            break
+    else:
+        if active_options and active_options._identity_token is not None:
+            shard_id = active_options._identity_token
+        elif "_sa_shard_id" in orm_context.execution_options:
+            shard_id = orm_context.execution_options["_sa_shard_id"]
+        elif "shard_id" in orm_context.bind_arguments:
+            shard_id = orm_context.bind_arguments["shard_id"]
+        else:
+            shard_id = None
+
+    if shard_id is not None:
+        return iter_for_shard(shard_id)
+    else:
+        partial = []
+        for shard_id in session.execute_chooser(orm_context):
+            result_ = iter_for_shard(shard_id)
+            partial.append(result_)
+        return partial[0].merge(*partial[1:])
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/hybrid.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/hybrid.py
new file mode 100644
index 00000000..c1c46e7c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/hybrid.py
@@ -0,0 +1,1533 @@
+# ext/hybrid.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+r"""Define attributes on ORM-mapped classes that have "hybrid" behavior.
+
+"hybrid" means the attribute has distinct behaviors defined at the
+class level and at the instance level.
+
+The :mod:`~sqlalchemy.ext.hybrid` extension provides a special form of
+method decorator and has minimal dependencies on the rest of SQLAlchemy.
+Its basic theory of operation can work with any descriptor-based expression
+system.
+
+Consider a mapping ``Interval``, representing integer ``start`` and ``end``
+values. We can define higher level functions on mapped classes that produce SQL
+expressions at the class level, and Python expression evaluation at the
+instance level.  Below, each function decorated with :class:`.hybrid_method` or
+:class:`.hybrid_property` may receive ``self`` as an instance of the class, or
+may receive the class directly, depending on context::
+
+    from __future__ import annotations
+
+    from sqlalchemy.ext.hybrid import hybrid_method
+    from sqlalchemy.ext.hybrid import hybrid_property
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class Interval(Base):
+        __tablename__ = "interval"
+
+        id: Mapped[int] = mapped_column(primary_key=True)
+        start: Mapped[int]
+        end: Mapped[int]
+
+        def __init__(self, start: int, end: int):
+            self.start = start
+            self.end = end
+
+        @hybrid_property
+        def length(self) -> int:
+            return self.end - self.start
+
+        @hybrid_method
+        def contains(self, point: int) -> bool:
+            return (self.start <= point) & (point <= self.end)
+
+        @hybrid_method
+        def intersects(self, other: Interval) -> bool:
+            return self.contains(other.start) | self.contains(other.end)
+
+Above, the ``length`` property returns the difference between the
+``end`` and ``start`` attributes.  With an instance of ``Interval``,
+this subtraction occurs in Python, using normal Python descriptor
+mechanics::
+
+    >>> i1 = Interval(5, 10)
+    >>> i1.length
+    5
+
+When dealing with the ``Interval`` class itself, the :class:`.hybrid_property`
+descriptor evaluates the function body given the ``Interval`` class as
+the argument, which when evaluated with SQLAlchemy expression mechanics
+returns a new SQL expression:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> print(select(Interval.length))
+    {printsql}SELECT interval."end" - interval.start AS length
+    FROM interval{stop}
+
+
+    >>> print(select(Interval).filter(Interval.length > 10))
+    {printsql}SELECT interval.id, interval.start, interval."end"
+    FROM interval
+    WHERE interval."end" - interval.start > :param_1
+
+Filtering methods such as :meth:`.Select.filter_by` are supported
+with hybrid attributes as well:
+
+.. sourcecode:: pycon+sql
+
+    >>> print(select(Interval).filter_by(length=5))
+    {printsql}SELECT interval.id, interval.start, interval."end"
+    FROM interval
+    WHERE interval."end" - interval.start = :param_1
+
+The ``Interval`` class example also illustrates two methods,
+``contains()`` and ``intersects()``, decorated with
+:class:`.hybrid_method`. This decorator applies the same idea to
+methods that :class:`.hybrid_property` applies to attributes.   The
+methods return boolean values, and take advantage of the Python ``|``
+and ``&`` bitwise operators to produce equivalent instance-level and
+SQL expression-level boolean behavior:
+
+.. sourcecode:: pycon+sql
+
+    >>> i1.contains(6)
+    True
+    >>> i1.contains(15)
+    False
+    >>> i1.intersects(Interval(7, 18))
+    True
+    >>> i1.intersects(Interval(25, 29))
+    False
+
+    >>> print(select(Interval).filter(Interval.contains(15)))
+    {printsql}SELECT interval.id, interval.start, interval."end"
+    FROM interval
+    WHERE interval.start <= :start_1 AND interval."end" > :end_1{stop}
+
+    >>> ia = aliased(Interval)
+    >>> print(select(Interval, ia).filter(Interval.intersects(ia)))
+    {printsql}SELECT interval.id, interval.start,
+    interval."end", interval_1.id AS interval_1_id,
+    interval_1.start AS interval_1_start, interval_1."end" AS interval_1_end
+    FROM interval, interval AS interval_1
+    WHERE interval.start <= interval_1.start
+        AND interval."end" > interval_1.start
+        OR interval.start <= interval_1."end"
+        AND interval."end" > interval_1."end"{stop}
+
+.. _hybrid_distinct_expression:
+
+Defining Expression Behavior Distinct from Attribute Behavior
+--------------------------------------------------------------
+
+In the previous section, our usage of the ``&`` and ``|`` bitwise operators
+within the ``Interval.contains`` and ``Interval.intersects`` methods was
+fortunate, considering our functions operated on two boolean values to return a
+new one. In many cases, the construction of an in-Python function and a
+SQLAlchemy SQL expression have enough differences that two separate Python
+expressions should be defined. The :mod:`~sqlalchemy.ext.hybrid` decorator
+defines a **modifier** :meth:`.hybrid_property.expression` for this purpose. As an
+example we'll define the radius of the interval, which requires the usage of
+the absolute value function::
+
+    from sqlalchemy import ColumnElement
+    from sqlalchemy import Float
+    from sqlalchemy import func
+    from sqlalchemy import type_coerce
+
+
+    class Interval(Base):
+        # ...
+
+        @hybrid_property
+        def radius(self) -> float:
+            return abs(self.length) / 2
+
+        @radius.inplace.expression
+        @classmethod
+        def _radius_expression(cls) -> ColumnElement[float]:
+            return type_coerce(func.abs(cls.length) / 2, Float)
+
+In the above example, the :class:`.hybrid_property` first assigned to the
+name ``Interval.radius`` is amended by a subsequent method called
+``Interval._radius_expression``, using the decorator
+``@radius.inplace.expression``, which chains together two modifiers
+:attr:`.hybrid_property.inplace` and :attr:`.hybrid_property.expression`.
+The use of :attr:`.hybrid_property.inplace` indicates that the
+:meth:`.hybrid_property.expression` modifier should mutate the
+existing hybrid object at ``Interval.radius`` in place, without creating a
+new object.   Notes on this modifier and its
+rationale are discussed in the next section :ref:`hybrid_pep484_naming`.
+The use of ``@classmethod`` is optional, and is strictly to give typing
+tools a hint that ``cls`` in this case is expected to be the ``Interval``
+class, and not an instance of ``Interval``.
+
+.. note:: :attr:`.hybrid_property.inplace` as well as the use of ``@classmethod``
+   for proper typing support are available as of SQLAlchemy 2.0.4, and will
+   not work in earlier versions.
+
+With ``Interval.radius`` now including an expression element, the SQL
+function ``ABS()`` is returned when accessing ``Interval.radius``
+at the class level:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> print(select(Interval).filter(Interval.radius > 5))
+    {printsql}SELECT interval.id, interval.start, interval."end"
+    FROM interval
+    WHERE abs(interval."end" - interval.start) / :abs_1 > :param_1
+
+
+.. _hybrid_pep484_naming:
+
+Using ``inplace`` to create pep-484 compliant hybrid properties
+---------------------------------------------------------------
+
+In the previous section, a :class:`.hybrid_property` decorator is illustrated
+which includes two separate method-level functions being decorated, both
+to produce a single object attribute referenced as ``Interval.radius``.
+There are actually several different modifiers we can use for
+:class:`.hybrid_property` including :meth:`.hybrid_property.expression`,
+:meth:`.hybrid_property.setter` and :meth:`.hybrid_property.update_expression`.
+
+SQLAlchemy's :class:`.hybrid_property` decorator intends that adding on these
+methods may be done in the identical manner as Python's built-in
+``@property`` decorator, where idiomatic use is to continue to redefine the
+attribute repeatedly, using the **same attribute name** each time, as in the
+example below that illustrates the use of :meth:`.hybrid_property.setter` and
+:meth:`.hybrid_property.expression` for the ``Interval.radius`` descriptor::
+
+    # correct use, however is not accepted by pep-484 tooling
+
+
+    class Interval(Base):
+        # ...
+
+        @hybrid_property
+        def radius(self):
+            return abs(self.length) / 2
+
+        @radius.setter
+        def radius(self, value):
+            self.length = value * 2
+
+        @radius.expression
+        def radius(cls):
+            return type_coerce(func.abs(cls.length) / 2, Float)
+
+Above, there are three ``Interval.radius`` methods, but as each are decorated,
+first by the :class:`.hybrid_property` decorator and then by the
+``@radius`` name itself, the end effect is that ``Interval.radius`` is
+a single attribute with three different functions contained within it.
+This style of use is taken from `Python's documented use of @property
+<https://docs.python.org/3/library/functions.html#property>`_.
+It is important to note that the way both ``@property`` as well as
+:class:`.hybrid_property` work, a **copy of the descriptor is made each time**.
+That is, each call to ``@radius.expression``, ``@radius.setter`` etc.
+make a new object entirely.  This allows the attribute to be re-defined in
+subclasses without issue (see :ref:`hybrid_reuse_subclass` later in this
+section for how this is used).
+
+However, the above approach is not compatible with typing tools such as
+mypy and pyright.  Python's own ``@property`` decorator does not have this
+limitation only because
+`these tools hardcode the behavior of @property
+<https://github.com/python/typing/discussions/1102>`_, meaning this syntax
+is not available to SQLAlchemy under :pep:`484` compliance.
+
+In order to produce a reasonable syntax while remaining typing compliant,
+the :attr:`.hybrid_property.inplace` decorator allows the same
+decorator to be re-used with different method names, while still producing
+a single decorator under one name::
+
+    # correct use which is also accepted by pep-484 tooling
+
+
+    class Interval(Base):
+        # ...
+
+        @hybrid_property
+        def radius(self) -> float:
+            return abs(self.length) / 2
+
+        @radius.inplace.setter
+        def _radius_setter(self, value: float) -> None:
+            # for example only
+            self.length = value * 2
+
+        @radius.inplace.expression
+        @classmethod
+        def _radius_expression(cls) -> ColumnElement[float]:
+            return type_coerce(func.abs(cls.length) / 2, Float)
+
+Using :attr:`.hybrid_property.inplace` further qualifies the use of the
+decorator that a new copy should not be made, thereby maintaining the
+``Interval.radius`` name while allowing additional methods
+``Interval._radius_setter`` and ``Interval._radius_expression`` to be
+differently named.
+
+
+.. versionadded:: 2.0.4 Added :attr:`.hybrid_property.inplace` to allow
+   less verbose construction of composite :class:`.hybrid_property` objects
+   while not having to use repeated method names.   Additionally allowed the
+   use of ``@classmethod`` within :attr:`.hybrid_property.expression`,
+   :attr:`.hybrid_property.update_expression`, and
+   :attr:`.hybrid_property.comparator` to allow typing tools to identify
+   ``cls`` as a class and not an instance in the method signature.
+
+
+Defining Setters
+----------------
+
+The :meth:`.hybrid_property.setter` modifier allows the construction of a
+custom setter method, that can modify values on the object::
+
+    class Interval(Base):
+        # ...
+
+        @hybrid_property
+        def length(self) -> int:
+            return self.end - self.start
+
+        @length.inplace.setter
+        def _length_setter(self, value: int) -> None:
+            self.end = self.start + value
+
+The ``length(self, value)`` method is now called upon set::
+
+    >>> i1 = Interval(5, 10)
+    >>> i1.length
+    5
+    >>> i1.length = 12
+    >>> i1.end
+    17
+
+.. _hybrid_bulk_update:
+
+Allowing Bulk ORM Update
+------------------------
+
+A hybrid can define a custom "UPDATE" handler for when using
+ORM-enabled updates, allowing the hybrid to be used in the
+SET clause of the update.
+
+Normally, when using a hybrid with :func:`_sql.update`, the SQL
+expression is used as the column that's the target of the SET.  If our
+``Interval`` class had a hybrid ``start_point`` that linked to
+``Interval.start``, this could be substituted directly::
+
+    from sqlalchemy import update
+
+    stmt = update(Interval).values({Interval.start_point: 10})
+
+However, when using a composite hybrid like ``Interval.length``, this
+hybrid represents more than one column.   We can set up a handler that will
+accommodate a value passed in the VALUES expression which can affect
+this, using the :meth:`.hybrid_property.update_expression` decorator.
+A handler that works similarly to our setter would be::
+
+    from typing import List, Tuple, Any
+
+
+    class Interval(Base):
+        # ...
+
+        @hybrid_property
+        def length(self) -> int:
+            return self.end - self.start
+
+        @length.inplace.setter
+        def _length_setter(self, value: int) -> None:
+            self.end = self.start + value
+
+        @length.inplace.update_expression
+        def _length_update_expression(
+            cls, value: Any
+        ) -> List[Tuple[Any, Any]]:
+            return [(cls.end, cls.start + value)]
+
+Above, if we use ``Interval.length`` in an UPDATE expression, we get
+a hybrid SET expression:
+
+.. sourcecode:: pycon+sql
+
+
+    >>> from sqlalchemy import update
+    >>> print(update(Interval).values({Interval.length: 25}))
+    {printsql}UPDATE interval SET "end"=(interval.start + :start_1)
+
+This SET expression is accommodated by the ORM automatically.
+
+.. seealso::
+
+    :ref:`orm_expression_update_delete` - includes background on ORM-enabled
+    UPDATE statements
+
+
+Working with Relationships
+--------------------------
+
+There's no essential difference when creating hybrids that work with
+related objects as opposed to column-based data. The need for distinct
+expressions tends to be greater.  The two variants we'll illustrate
+are the "join-dependent" hybrid, and the "correlated subquery" hybrid.
+
+Join-Dependent Relationship Hybrid
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Consider the following declarative
+mapping which relates a ``User`` to a ``SavingsAccount``::
+
+    from __future__ import annotations
+
+    from decimal import Decimal
+    from typing import cast
+    from typing import List
+    from typing import Optional
+
+    from sqlalchemy import ForeignKey
+    from sqlalchemy import Numeric
+    from sqlalchemy import String
+    from sqlalchemy import SQLColumnExpression
+    from sqlalchemy.ext.hybrid import hybrid_property
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+    from sqlalchemy.orm import relationship
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class SavingsAccount(Base):
+        __tablename__ = "account"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        user_id: Mapped[int] = mapped_column(ForeignKey("user.id"))
+        balance: Mapped[Decimal] = mapped_column(Numeric(15, 5))
+
+        owner: Mapped[User] = relationship(back_populates="accounts")
+
+
+    class User(Base):
+        __tablename__ = "user"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        name: Mapped[str] = mapped_column(String(100))
+
+        accounts: Mapped[List[SavingsAccount]] = relationship(
+            back_populates="owner", lazy="selectin"
+        )
+
+        @hybrid_property
+        def balance(self) -> Optional[Decimal]:
+            if self.accounts:
+                return self.accounts[0].balance
+            else:
+                return None
+
+        @balance.inplace.setter
+        def _balance_setter(self, value: Optional[Decimal]) -> None:
+            assert value is not None
+
+            if not self.accounts:
+                account = SavingsAccount(owner=self)
+            else:
+                account = self.accounts[0]
+            account.balance = value
+
+        @balance.inplace.expression
+        @classmethod
+        def _balance_expression(cls) -> SQLColumnExpression[Optional[Decimal]]:
+            return cast(
+                "SQLColumnExpression[Optional[Decimal]]",
+                SavingsAccount.balance,
+            )
+
+The above hybrid property ``balance`` works with the first
+``SavingsAccount`` entry in the list of accounts for this user.   The
+in-Python getter/setter methods can treat ``accounts`` as a Python
+list available on ``self``.
+
+.. tip:: The ``User.balance`` getter in the above example accesses the
+   ``self.acccounts`` collection, which will normally be loaded via the
+   :func:`.selectinload` loader strategy configured on the ``User.balance``
+   :func:`_orm.relationship`. The default loader strategy when not otherwise
+   stated on :func:`_orm.relationship` is :func:`.lazyload`, which emits SQL on
+   demand. When using asyncio, on-demand loaders such as :func:`.lazyload` are
+   not supported, so care should be taken to ensure the ``self.accounts``
+   collection is accessible to this hybrid accessor when using asyncio.
+
+At the expression level, it's expected that the ``User`` class will
+be used in an appropriate context such that an appropriate join to
+``SavingsAccount`` will be present:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> print(
+    ...     select(User, User.balance)
+    ...     .join(User.accounts)
+    ...     .filter(User.balance > 5000)
+    ... )
+    {printsql}SELECT "user".id AS user_id, "user".name AS user_name,
+    account.balance AS account_balance
+    FROM "user" JOIN account ON "user".id = account.user_id
+    WHERE account.balance > :balance_1
+
+Note however, that while the instance level accessors need to worry
+about whether ``self.accounts`` is even present, this issue expresses
+itself differently at the SQL expression level, where we basically
+would use an outer join:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> from sqlalchemy import or_
+    >>> print(
+    ...     select(User, User.balance)
+    ...     .outerjoin(User.accounts)
+    ...     .filter(or_(User.balance < 5000, User.balance == None))
+    ... )
+    {printsql}SELECT "user".id AS user_id, "user".name AS user_name,
+    account.balance AS account_balance
+    FROM "user" LEFT OUTER JOIN account ON "user".id = account.user_id
+    WHERE account.balance <  :balance_1 OR account.balance IS NULL
+
+Correlated Subquery Relationship Hybrid
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+We can, of course, forego being dependent on the enclosing query's usage
+of joins in favor of the correlated subquery, which can portably be packed
+into a single column expression. A correlated subquery is more portable, but
+often performs more poorly at the SQL level. Using the same technique
+illustrated at :ref:`mapper_column_property_sql_expressions`,
+we can adjust our ``SavingsAccount`` example to aggregate the balances for
+*all* accounts, and use a correlated subquery for the column expression::
+
+    from __future__ import annotations
+
+    from decimal import Decimal
+    from typing import List
+
+    from sqlalchemy import ForeignKey
+    from sqlalchemy import func
+    from sqlalchemy import Numeric
+    from sqlalchemy import select
+    from sqlalchemy import SQLColumnExpression
+    from sqlalchemy import String
+    from sqlalchemy.ext.hybrid import hybrid_property
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+    from sqlalchemy.orm import relationship
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class SavingsAccount(Base):
+        __tablename__ = "account"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        user_id: Mapped[int] = mapped_column(ForeignKey("user.id"))
+        balance: Mapped[Decimal] = mapped_column(Numeric(15, 5))
+
+        owner: Mapped[User] = relationship(back_populates="accounts")
+
+
+    class User(Base):
+        __tablename__ = "user"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        name: Mapped[str] = mapped_column(String(100))
+
+        accounts: Mapped[List[SavingsAccount]] = relationship(
+            back_populates="owner", lazy="selectin"
+        )
+
+        @hybrid_property
+        def balance(self) -> Decimal:
+            return sum(
+                (acc.balance for acc in self.accounts), start=Decimal("0")
+            )
+
+        @balance.inplace.expression
+        @classmethod
+        def _balance_expression(cls) -> SQLColumnExpression[Decimal]:
+            return (
+                select(func.sum(SavingsAccount.balance))
+                .where(SavingsAccount.user_id == cls.id)
+                .label("total_balance")
+            )
+
+The above recipe will give us the ``balance`` column which renders
+a correlated SELECT:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> print(select(User).filter(User.balance > 400))
+    {printsql}SELECT "user".id, "user".name
+    FROM "user"
+    WHERE (
+        SELECT sum(account.balance) AS sum_1 FROM account
+        WHERE account.user_id = "user".id
+    ) > :param_1
+
+
+.. _hybrid_custom_comparators:
+
+Building Custom Comparators
+---------------------------
+
+The hybrid property also includes a helper that allows construction of
+custom comparators. A comparator object allows one to customize the
+behavior of each SQLAlchemy expression operator individually.  They
+are useful when creating custom types that have some highly
+idiosyncratic behavior on the SQL side.
+
+.. note::  The :meth:`.hybrid_property.comparator` decorator introduced
+   in this section **replaces** the use of the
+   :meth:`.hybrid_property.expression` decorator.
+   They cannot be used together.
+
+The example class below allows case-insensitive comparisons on the attribute
+named ``word_insensitive``::
+
+    from __future__ import annotations
+
+    from typing import Any
+
+    from sqlalchemy import ColumnElement
+    from sqlalchemy import func
+    from sqlalchemy.ext.hybrid import Comparator
+    from sqlalchemy.ext.hybrid import hybrid_property
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class CaseInsensitiveComparator(Comparator[str]):
+        def __eq__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            return func.lower(self.__clause_element__()) == func.lower(other)
+
+
+    class SearchWord(Base):
+        __tablename__ = "searchword"
+
+        id: Mapped[int] = mapped_column(primary_key=True)
+        word: Mapped[str]
+
+        @hybrid_property
+        def word_insensitive(self) -> str:
+            return self.word.lower()
+
+        @word_insensitive.inplace.comparator
+        @classmethod
+        def _word_insensitive_comparator(cls) -> CaseInsensitiveComparator:
+            return CaseInsensitiveComparator(cls.word)
+
+Above, SQL expressions against ``word_insensitive`` will apply the ``LOWER()``
+SQL function to both sides:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy import select
+    >>> print(select(SearchWord).filter_by(word_insensitive="Trucks"))
+    {printsql}SELECT searchword.id, searchword.word
+    FROM searchword
+    WHERE lower(searchword.word) = lower(:lower_1)
+
+
+The ``CaseInsensitiveComparator`` above implements part of the
+:class:`.ColumnOperators` interface.   A "coercion" operation like
+lowercasing can be applied to all comparison operations (i.e. ``eq``,
+``lt``, ``gt``, etc.) using :meth:`.Operators.operate`::
+
+    class CaseInsensitiveComparator(Comparator):
+        def operate(self, op, other, **kwargs):
+            return op(
+                func.lower(self.__clause_element__()),
+                func.lower(other),
+                **kwargs,
+            )
+
+.. _hybrid_reuse_subclass:
+
+Reusing Hybrid Properties across Subclasses
+-------------------------------------------
+
+A hybrid can be referred to from a superclass, to allow modifying
+methods like :meth:`.hybrid_property.getter`, :meth:`.hybrid_property.setter`
+to be used to redefine those methods on a subclass.  This is similar to
+how the standard Python ``@property`` object works::
+
+    class FirstNameOnly(Base):
+        # ...
+
+        first_name: Mapped[str]
+
+        @hybrid_property
+        def name(self) -> str:
+            return self.first_name
+
+        @name.inplace.setter
+        def _name_setter(self, value: str) -> None:
+            self.first_name = value
+
+
+    class FirstNameLastName(FirstNameOnly):
+        # ...
+
+        last_name: Mapped[str]
+
+        # 'inplace' is not used here; calling getter creates a copy
+        # of FirstNameOnly.name that is local to FirstNameLastName
+        @FirstNameOnly.name.getter
+        def name(self) -> str:
+            return self.first_name + " " + self.last_name
+
+        @name.inplace.setter
+        def _name_setter(self, value: str) -> None:
+            self.first_name, self.last_name = value.split(" ", 1)
+
+Above, the ``FirstNameLastName`` class refers to the hybrid from
+``FirstNameOnly.name`` to repurpose its getter and setter for the subclass.
+
+When overriding :meth:`.hybrid_property.expression` and
+:meth:`.hybrid_property.comparator` alone as the first reference to the
+superclass, these names conflict with the same-named accessors on the class-
+level :class:`.QueryableAttribute` object returned at the class level.  To
+override these methods when referring directly to the parent class descriptor,
+add the special qualifier :attr:`.hybrid_property.overrides`, which will de-
+reference the instrumented attribute back to the hybrid object::
+
+    class FirstNameLastName(FirstNameOnly):
+        # ...
+
+        last_name: Mapped[str]
+
+        @FirstNameOnly.name.overrides.expression
+        @classmethod
+        def name(cls):
+            return func.concat(cls.first_name, " ", cls.last_name)
+
+Hybrid Value Objects
+--------------------
+
+Note in our previous example, if we were to compare the ``word_insensitive``
+attribute of a ``SearchWord`` instance to a plain Python string, the plain
+Python string would not be coerced to lower case - the
+``CaseInsensitiveComparator`` we built, being returned by
+``@word_insensitive.comparator``, only applies to the SQL side.
+
+A more comprehensive form of the custom comparator is to construct a *Hybrid
+Value Object*. This technique applies the target value or expression to a value
+object which is then returned by the accessor in all cases.   The value object
+allows control of all operations upon the value as well as how compared values
+are treated, both on the SQL expression side as well as the Python value side.
+Replacing the previous ``CaseInsensitiveComparator`` class with a new
+``CaseInsensitiveWord`` class::
+
+    class CaseInsensitiveWord(Comparator):
+        "Hybrid value representing a lower case representation of a word."
+
+        def __init__(self, word):
+            if isinstance(word, basestring):
+                self.word = word.lower()
+            elif isinstance(word, CaseInsensitiveWord):
+                self.word = word.word
+            else:
+                self.word = func.lower(word)
+
+        def operate(self, op, other, **kwargs):
+            if not isinstance(other, CaseInsensitiveWord):
+                other = CaseInsensitiveWord(other)
+            return op(self.word, other.word, **kwargs)
+
+        def __clause_element__(self):
+            return self.word
+
+        def __str__(self):
+            return self.word
+
+        key = "word"
+        "Label to apply to Query tuple results"
+
+Above, the ``CaseInsensitiveWord`` object represents ``self.word``, which may
+be a SQL function, or may be a Python native.   By overriding ``operate()`` and
+``__clause_element__()`` to work in terms of ``self.word``, all comparison
+operations will work against the "converted" form of ``word``, whether it be
+SQL side or Python side. Our ``SearchWord`` class can now deliver the
+``CaseInsensitiveWord`` object unconditionally from a single hybrid call::
+
+    class SearchWord(Base):
+        __tablename__ = "searchword"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        word: Mapped[str]
+
+        @hybrid_property
+        def word_insensitive(self) -> CaseInsensitiveWord:
+            return CaseInsensitiveWord(self.word)
+
+The ``word_insensitive`` attribute now has case-insensitive comparison behavior
+universally, including SQL expression vs. Python expression (note the Python
+value is converted to lower case on the Python side here):
+
+.. sourcecode:: pycon+sql
+
+    >>> print(select(SearchWord).filter_by(word_insensitive="Trucks"))
+    {printsql}SELECT searchword.id AS searchword_id, searchword.word AS searchword_word
+    FROM searchword
+    WHERE lower(searchword.word) = :lower_1
+
+SQL expression versus SQL expression:
+
+.. sourcecode:: pycon+sql
+
+    >>> from sqlalchemy.orm import aliased
+    >>> sw1 = aliased(SearchWord)
+    >>> sw2 = aliased(SearchWord)
+    >>> print(
+    ...     select(sw1.word_insensitive, sw2.word_insensitive).filter(
+    ...         sw1.word_insensitive > sw2.word_insensitive
+    ...     )
+    ... )
+    {printsql}SELECT lower(searchword_1.word) AS lower_1,
+    lower(searchword_2.word) AS lower_2
+    FROM searchword AS searchword_1, searchword AS searchword_2
+    WHERE lower(searchword_1.word) > lower(searchword_2.word)
+
+Python only expression::
+
+    >>> ws1 = SearchWord(word="SomeWord")
+    >>> ws1.word_insensitive == "sOmEwOrD"
+    True
+    >>> ws1.word_insensitive == "XOmEwOrX"
+    False
+    >>> print(ws1.word_insensitive)
+    someword
+
+The Hybrid Value pattern is very useful for any kind of value that may have
+multiple representations, such as timestamps, time deltas, units of
+measurement, currencies and encrypted passwords.
+
+.. seealso::
+
+    `Hybrids and Value Agnostic Types
+    <https://techspot.zzzeek.org/2011/10/21/hybrids-and-value-agnostic-types/>`_
+    - on the techspot.zzzeek.org blog
+
+    `Value Agnostic Types, Part II
+    <https://techspot.zzzeek.org/2011/10/29/value-agnostic-types-part-ii/>`_ -
+    on the techspot.zzzeek.org blog
+
+
+"""  # noqa
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Generic
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import util
+from ..orm import attributes
+from ..orm import InspectionAttrExtensionType
+from ..orm import interfaces
+from ..orm import ORMDescriptor
+from ..orm.attributes import QueryableAttribute
+from ..sql import roles
+from ..sql._typing import is_has_clause_element
+from ..sql.elements import ColumnElement
+from ..sql.elements import SQLCoreOperations
+from ..util.typing import Concatenate
+from ..util.typing import Literal
+from ..util.typing import ParamSpec
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from ..orm.interfaces import MapperProperty
+    from ..orm.util import AliasedInsp
+    from ..sql import SQLColumnExpression
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _DMLColumnArgument
+    from ..sql._typing import _HasClauseElement
+    from ..sql._typing import _InfoType
+    from ..sql.operators import OperatorType
+
+_P = ParamSpec("_P")
+_R = TypeVar("_R")
+_T = TypeVar("_T", bound=Any)
+_TE = TypeVar("_TE", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+_T_con = TypeVar("_T_con", bound=Any, contravariant=True)
+
+
+class HybridExtensionType(InspectionAttrExtensionType):
+    HYBRID_METHOD = "HYBRID_METHOD"
+    """Symbol indicating an :class:`InspectionAttr` that's
+    of type :class:`.hybrid_method`.
+
+    Is assigned to the :attr:`.InspectionAttr.extension_type`
+    attribute.
+
+    .. seealso::
+
+        :attr:`_orm.Mapper.all_orm_attributes`
+
+    """
+
+    HYBRID_PROPERTY = "HYBRID_PROPERTY"
+    """Symbol indicating an :class:`InspectionAttr` that's
+        of type :class:`.hybrid_method`.
+
+    Is assigned to the :attr:`.InspectionAttr.extension_type`
+    attribute.
+
+    .. seealso::
+
+        :attr:`_orm.Mapper.all_orm_attributes`
+
+    """
+
+
+class _HybridGetterType(Protocol[_T_co]):
+    def __call__(s, self: Any) -> _T_co: ...
+
+
+class _HybridSetterType(Protocol[_T_con]):
+    def __call__(s, self: Any, value: _T_con) -> None: ...
+
+
+class _HybridUpdaterType(Protocol[_T_con]):
+    def __call__(
+        s,
+        cls: Any,
+        value: Union[_T_con, _ColumnExpressionArgument[_T_con]],
+    ) -> List[Tuple[_DMLColumnArgument, Any]]: ...
+
+
+class _HybridDeleterType(Protocol[_T_co]):
+    def __call__(s, self: Any) -> None: ...
+
+
+class _HybridExprCallableType(Protocol[_T_co]):
+    def __call__(
+        s, cls: Any
+    ) -> Union[_HasClauseElement[_T_co], SQLColumnExpression[_T_co]]: ...
+
+
+class _HybridComparatorCallableType(Protocol[_T]):
+    def __call__(self, cls: Any) -> Comparator[_T]: ...
+
+
+class _HybridClassLevelAccessor(QueryableAttribute[_T]):
+    """Describe the object returned by a hybrid_property() when
+    called as a class-level descriptor.
+
+    """
+
+    if TYPE_CHECKING:
+
+        def getter(
+            self, fget: _HybridGetterType[_T]
+        ) -> hybrid_property[_T]: ...
+
+        def setter(
+            self, fset: _HybridSetterType[_T]
+        ) -> hybrid_property[_T]: ...
+
+        def deleter(
+            self, fdel: _HybridDeleterType[_T]
+        ) -> hybrid_property[_T]: ...
+
+        @property
+        def overrides(self) -> hybrid_property[_T]: ...
+
+        def update_expression(
+            self, meth: _HybridUpdaterType[_T]
+        ) -> hybrid_property[_T]: ...
+
+
+class hybrid_method(interfaces.InspectionAttrInfo, Generic[_P, _R]):
+    """A decorator which allows definition of a Python object method with both
+    instance-level and class-level behavior.
+
+    """
+
+    is_attribute = True
+    extension_type = HybridExtensionType.HYBRID_METHOD
+
+    def __init__(
+        self,
+        func: Callable[Concatenate[Any, _P], _R],
+        expr: Optional[
+            Callable[Concatenate[Any, _P], SQLCoreOperations[_R]]
+        ] = None,
+    ):
+        """Create a new :class:`.hybrid_method`.
+
+        Usage is typically via decorator::
+
+            from sqlalchemy.ext.hybrid import hybrid_method
+
+
+            class SomeClass:
+                @hybrid_method
+                def value(self, x, y):
+                    return self._value + x + y
+
+                @value.expression
+                @classmethod
+                def value(cls, x, y):
+                    return func.some_function(cls._value, x, y)
+
+        """
+        self.func = func
+        if expr is not None:
+            self.expression(expr)
+        else:
+            self.expression(func)  # type: ignore
+
+    @property
+    def inplace(self) -> Self:
+        """Return the inplace mutator for this :class:`.hybrid_method`.
+
+        The :class:`.hybrid_method` class already performs "in place" mutation
+        when the :meth:`.hybrid_method.expression` decorator is called,
+        so this attribute returns Self.
+
+        .. versionadded:: 2.0.4
+
+        .. seealso::
+
+            :ref:`hybrid_pep484_naming`
+
+        """
+        return self
+
+    @overload
+    def __get__(
+        self, instance: Literal[None], owner: Type[object]
+    ) -> Callable[_P, SQLCoreOperations[_R]]: ...
+
+    @overload
+    def __get__(
+        self, instance: object, owner: Type[object]
+    ) -> Callable[_P, _R]: ...
+
+    def __get__(
+        self, instance: Optional[object], owner: Type[object]
+    ) -> Union[Callable[_P, _R], Callable[_P, SQLCoreOperations[_R]]]:
+        if instance is None:
+            return self.expr.__get__(owner, owner)  # type: ignore
+        else:
+            return self.func.__get__(instance, owner)  # type: ignore
+
+    def expression(
+        self, expr: Callable[Concatenate[Any, _P], SQLCoreOperations[_R]]
+    ) -> hybrid_method[_P, _R]:
+        """Provide a modifying decorator that defines a
+        SQL-expression producing method."""
+
+        self.expr = expr
+        if not self.expr.__doc__:
+            self.expr.__doc__ = self.func.__doc__
+        return self
+
+
+def _unwrap_classmethod(meth: _T) -> _T:
+    if isinstance(meth, classmethod):
+        return meth.__func__  # type: ignore
+    else:
+        return meth
+
+
+class hybrid_property(interfaces.InspectionAttrInfo, ORMDescriptor[_T]):
+    """A decorator which allows definition of a Python descriptor with both
+    instance-level and class-level behavior.
+
+    """
+
+    is_attribute = True
+    extension_type = HybridExtensionType.HYBRID_PROPERTY
+
+    __name__: str
+
+    def __init__(
+        self,
+        fget: _HybridGetterType[_T],
+        fset: Optional[_HybridSetterType[_T]] = None,
+        fdel: Optional[_HybridDeleterType[_T]] = None,
+        expr: Optional[_HybridExprCallableType[_T]] = None,
+        custom_comparator: Optional[Comparator[_T]] = None,
+        update_expr: Optional[_HybridUpdaterType[_T]] = None,
+    ):
+        """Create a new :class:`.hybrid_property`.
+
+        Usage is typically via decorator::
+
+            from sqlalchemy.ext.hybrid import hybrid_property
+
+
+            class SomeClass:
+                @hybrid_property
+                def value(self):
+                    return self._value
+
+                @value.setter
+                def value(self, value):
+                    self._value = value
+
+        """
+        self.fget = fget
+        self.fset = fset
+        self.fdel = fdel
+        self.expr = _unwrap_classmethod(expr)
+        self.custom_comparator = _unwrap_classmethod(custom_comparator)
+        self.update_expr = _unwrap_classmethod(update_expr)
+        util.update_wrapper(self, fget)  # type: ignore[arg-type]
+
+    @overload
+    def __get__(self, instance: Any, owner: Literal[None]) -> Self: ...
+
+    @overload
+    def __get__(
+        self, instance: Literal[None], owner: Type[object]
+    ) -> _HybridClassLevelAccessor[_T]: ...
+
+    @overload
+    def __get__(self, instance: object, owner: Type[object]) -> _T: ...
+
+    def __get__(
+        self, instance: Optional[object], owner: Optional[Type[object]]
+    ) -> Union[hybrid_property[_T], _HybridClassLevelAccessor[_T], _T]:
+        if owner is None:
+            return self
+        elif instance is None:
+            return self._expr_comparator(owner)
+        else:
+            return self.fget(instance)
+
+    def __set__(self, instance: object, value: Any) -> None:
+        if self.fset is None:
+            raise AttributeError("can't set attribute")
+        self.fset(instance, value)
+
+    def __delete__(self, instance: object) -> None:
+        if self.fdel is None:
+            raise AttributeError("can't delete attribute")
+        self.fdel(instance)
+
+    def _copy(self, **kw: Any) -> hybrid_property[_T]:
+        defaults = {
+            key: value
+            for key, value in self.__dict__.items()
+            if not key.startswith("_")
+        }
+        defaults.update(**kw)
+        return type(self)(**defaults)
+
+    @property
+    def overrides(self) -> Self:
+        """Prefix for a method that is overriding an existing attribute.
+
+        The :attr:`.hybrid_property.overrides` accessor just returns
+        this hybrid object, which when called at the class level from
+        a parent class, will de-reference the "instrumented attribute"
+        normally returned at this level, and allow modifying decorators
+        like :meth:`.hybrid_property.expression` and
+        :meth:`.hybrid_property.comparator`
+        to be used without conflicting with the same-named attributes
+        normally present on the :class:`.QueryableAttribute`::
+
+            class SuperClass:
+                # ...
+
+                @hybrid_property
+                def foobar(self):
+                    return self._foobar
+
+
+            class SubClass(SuperClass):
+                # ...
+
+                @SuperClass.foobar.overrides.expression
+                def foobar(cls):
+                    return func.subfoobar(self._foobar)
+
+        .. versionadded:: 1.2
+
+        .. seealso::
+
+            :ref:`hybrid_reuse_subclass`
+
+        """
+        return self
+
+    class _InPlace(Generic[_TE]):
+        """A builder helper for .hybrid_property.
+
+        .. versionadded:: 2.0.4
+
+        """
+
+        __slots__ = ("attr",)
+
+        def __init__(self, attr: hybrid_property[_TE]):
+            self.attr = attr
+
+        def _set(self, **kw: Any) -> hybrid_property[_TE]:
+            for k, v in kw.items():
+                setattr(self.attr, k, _unwrap_classmethod(v))
+            return self.attr
+
+        def getter(self, fget: _HybridGetterType[_TE]) -> hybrid_property[_TE]:
+            return self._set(fget=fget)
+
+        def setter(self, fset: _HybridSetterType[_TE]) -> hybrid_property[_TE]:
+            return self._set(fset=fset)
+
+        def deleter(
+            self, fdel: _HybridDeleterType[_TE]
+        ) -> hybrid_property[_TE]:
+            return self._set(fdel=fdel)
+
+        def expression(
+            self, expr: _HybridExprCallableType[_TE]
+        ) -> hybrid_property[_TE]:
+            return self._set(expr=expr)
+
+        def comparator(
+            self, comparator: _HybridComparatorCallableType[_TE]
+        ) -> hybrid_property[_TE]:
+            return self._set(custom_comparator=comparator)
+
+        def update_expression(
+            self, meth: _HybridUpdaterType[_TE]
+        ) -> hybrid_property[_TE]:
+            return self._set(update_expr=meth)
+
+    @property
+    def inplace(self) -> _InPlace[_T]:
+        """Return the inplace mutator for this :class:`.hybrid_property`.
+
+        This is to allow in-place mutation of the hybrid, allowing the first
+        hybrid method of a certain name to be re-used in order to add
+        more methods without having to name those methods the same, e.g.::
+
+            class Interval(Base):
+                # ...
+
+                @hybrid_property
+                def radius(self) -> float:
+                    return abs(self.length) / 2
+
+                @radius.inplace.setter
+                def _radius_setter(self, value: float) -> None:
+                    self.length = value * 2
+
+                @radius.inplace.expression
+                def _radius_expression(cls) -> ColumnElement[float]:
+                    return type_coerce(func.abs(cls.length) / 2, Float)
+
+        .. versionadded:: 2.0.4
+
+        .. seealso::
+
+            :ref:`hybrid_pep484_naming`
+
+        """
+        return hybrid_property._InPlace(self)
+
+    def getter(self, fget: _HybridGetterType[_T]) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines a getter method.
+
+        .. versionadded:: 1.2
+
+        """
+
+        return self._copy(fget=fget)
+
+    def setter(self, fset: _HybridSetterType[_T]) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines a setter method."""
+
+        return self._copy(fset=fset)
+
+    def deleter(self, fdel: _HybridDeleterType[_T]) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines a deletion method."""
+
+        return self._copy(fdel=fdel)
+
+    def expression(
+        self, expr: _HybridExprCallableType[_T]
+    ) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines a SQL-expression
+        producing method.
+
+        When a hybrid is invoked at the class level, the SQL expression given
+        here is wrapped inside of a specialized :class:`.QueryableAttribute`,
+        which is the same kind of object used by the ORM to represent other
+        mapped attributes.   The reason for this is so that other class-level
+        attributes such as docstrings and a reference to the hybrid itself may
+        be maintained within the structure that's returned, without any
+        modifications to the original SQL expression passed in.
+
+        .. note::
+
+           When referring to a hybrid property  from an owning class (e.g.
+           ``SomeClass.some_hybrid``), an instance of
+           :class:`.QueryableAttribute` is returned, representing the
+           expression or comparator object as well as this  hybrid object.
+           However, that object itself has accessors called ``expression`` and
+           ``comparator``; so when attempting to override these decorators on a
+           subclass, it may be necessary to qualify it using the
+           :attr:`.hybrid_property.overrides` modifier first.  See that
+           modifier for details.
+
+        .. seealso::
+
+            :ref:`hybrid_distinct_expression`
+
+        """
+
+        return self._copy(expr=expr)
+
+    def comparator(
+        self, comparator: _HybridComparatorCallableType[_T]
+    ) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines a custom
+        comparator producing method.
+
+        The return value of the decorated method should be an instance of
+        :class:`~.hybrid.Comparator`.
+
+        .. note::  The :meth:`.hybrid_property.comparator` decorator
+           **replaces** the use of the :meth:`.hybrid_property.expression`
+           decorator.  They cannot be used together.
+
+        When a hybrid is invoked at the class level, the
+        :class:`~.hybrid.Comparator` object given here is wrapped inside of a
+        specialized :class:`.QueryableAttribute`, which is the same kind of
+        object used by the ORM to represent other mapped attributes.   The
+        reason for this is so that other class-level attributes such as
+        docstrings and a reference to the hybrid itself may be maintained
+        within the structure that's returned, without any modifications to the
+        original comparator object passed in.
+
+        .. note::
+
+           When referring to a hybrid property  from an owning class (e.g.
+           ``SomeClass.some_hybrid``), an instance of
+           :class:`.QueryableAttribute` is returned, representing the
+           expression or comparator object as this  hybrid object.  However,
+           that object itself has accessors called ``expression`` and
+           ``comparator``; so when attempting to override these decorators on a
+           subclass, it may be necessary to qualify it using the
+           :attr:`.hybrid_property.overrides` modifier first.  See that
+           modifier for details.
+
+        """
+        return self._copy(custom_comparator=comparator)
+
+    def update_expression(
+        self, meth: _HybridUpdaterType[_T]
+    ) -> hybrid_property[_T]:
+        """Provide a modifying decorator that defines an UPDATE tuple
+        producing method.
+
+        The method accepts a single value, which is the value to be
+        rendered into the SET clause of an UPDATE statement.  The method
+        should then process this value into individual column expressions
+        that fit into the ultimate SET clause, and return them as a
+        sequence of 2-tuples.  Each tuple
+        contains a column expression as the key and a value to be rendered.
+
+        E.g.::
+
+            class Person(Base):
+                # ...
+
+                first_name = Column(String)
+                last_name = Column(String)
+
+                @hybrid_property
+                def fullname(self):
+                    return first_name + " " + last_name
+
+                @fullname.update_expression
+                def fullname(cls, value):
+                    fname, lname = value.split(" ", 1)
+                    return [(cls.first_name, fname), (cls.last_name, lname)]
+
+        .. versionadded:: 1.2
+
+        """
+        return self._copy(update_expr=meth)
+
+    @util.memoized_property
+    def _expr_comparator(
+        self,
+    ) -> Callable[[Any], _HybridClassLevelAccessor[_T]]:
+        if self.custom_comparator is not None:
+            return self._get_comparator(self.custom_comparator)
+        elif self.expr is not None:
+            return self._get_expr(self.expr)
+        else:
+            return self._get_expr(cast(_HybridExprCallableType[_T], self.fget))
+
+    def _get_expr(
+        self, expr: _HybridExprCallableType[_T]
+    ) -> Callable[[Any], _HybridClassLevelAccessor[_T]]:
+        def _expr(cls: Any) -> ExprComparator[_T]:
+            return ExprComparator(cls, expr(cls), self)
+
+        util.update_wrapper(_expr, expr)
+
+        return self._get_comparator(_expr)
+
+    def _get_comparator(
+        self, comparator: Any
+    ) -> Callable[[Any], _HybridClassLevelAccessor[_T]]:
+        proxy_attr = attributes.create_proxied_attribute(self)
+
+        def expr_comparator(
+            owner: Type[object],
+        ) -> _HybridClassLevelAccessor[_T]:
+            # because this is the descriptor protocol, we don't really know
+            # what our attribute name is.  so search for it through the
+            # MRO.
+            for lookup in owner.__mro__:
+                if self.__name__ in lookup.__dict__:
+                    if lookup.__dict__[self.__name__] is self:
+                        name = self.__name__
+                        break
+            else:
+                name = attributes._UNKNOWN_ATTR_KEY  # type: ignore[assignment]
+
+            return cast(
+                "_HybridClassLevelAccessor[_T]",
+                proxy_attr(
+                    owner,
+                    name,
+                    self,
+                    comparator(owner),
+                    doc=comparator.__doc__ or self.__doc__,
+                ),
+            )
+
+        return expr_comparator
+
+
+class Comparator(interfaces.PropComparator[_T]):
+    """A helper class that allows easy construction of custom
+    :class:`~.orm.interfaces.PropComparator`
+    classes for usage with hybrids."""
+
+    def __init__(
+        self, expression: Union[_HasClauseElement[_T], SQLColumnExpression[_T]]
+    ):
+        self.expression = expression
+
+    def __clause_element__(self) -> roles.ColumnsClauseRole:
+        expr = self.expression
+        if is_has_clause_element(expr):
+            ret_expr = expr.__clause_element__()
+        else:
+            if TYPE_CHECKING:
+                assert isinstance(expr, ColumnElement)
+            ret_expr = expr
+
+        if TYPE_CHECKING:
+            # see test_hybrid->test_expression_isnt_clause_element
+            # that exercises the usual place this is caught if not
+            # true
+            assert isinstance(ret_expr, ColumnElement)
+        return ret_expr
+
+    @util.non_memoized_property
+    def property(self) -> interfaces.MapperProperty[_T]:
+        raise NotImplementedError()
+
+    def adapt_to_entity(
+        self, adapt_to_entity: AliasedInsp[Any]
+    ) -> Comparator[_T]:
+        # interesting....
+        return self
+
+
+class ExprComparator(Comparator[_T]):
+    def __init__(
+        self,
+        cls: Type[Any],
+        expression: Union[_HasClauseElement[_T], SQLColumnExpression[_T]],
+        hybrid: hybrid_property[_T],
+    ):
+        self.cls = cls
+        self.expression = expression
+        self.hybrid = hybrid
+
+    def __getattr__(self, key: str) -> Any:
+        return getattr(self.expression, key)
+
+    @util.ro_non_memoized_property
+    def info(self) -> _InfoType:
+        return self.hybrid.info
+
+    def _bulk_update_tuples(
+        self, value: Any
+    ) -> Sequence[Tuple[_DMLColumnArgument, Any]]:
+        if isinstance(self.expression, attributes.QueryableAttribute):
+            return self.expression._bulk_update_tuples(value)
+        elif self.hybrid.update_expr is not None:
+            return self.hybrid.update_expr(self.cls, value)
+        else:
+            return [(self.expression, value)]
+
+    @util.non_memoized_property
+    def property(self) -> MapperProperty[_T]:
+        # this accessor is not normally used, however is accessed by things
+        # like ORM synonyms if the hybrid is used in this context; the
+        # .property attribute is not necessarily accessible
+        return self.expression.property  # type: ignore
+
+    def operate(
+        self, op: OperatorType, *other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(self.expression, *other, **kwargs)
+
+    def reverse_operate(
+        self, op: OperatorType, other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(other, self.expression, **kwargs)  # type: ignore
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/indexable.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/indexable.py
new file mode 100644
index 00000000..886069ce
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/indexable.py
@@ -0,0 +1,345 @@
+# ext/indexable.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""Define attributes on ORM-mapped classes that have "index" attributes for
+columns with :class:`_types.Indexable` types.
+
+"index" means the attribute is associated with an element of an
+:class:`_types.Indexable` column with the predefined index to access it.
+The :class:`_types.Indexable` types include types such as
+:class:`_types.ARRAY`, :class:`_types.JSON` and
+:class:`_postgresql.HSTORE`.
+
+
+
+The :mod:`~sqlalchemy.ext.indexable` extension provides
+:class:`_schema.Column`-like interface for any element of an
+:class:`_types.Indexable` typed column. In simple cases, it can be
+treated as a :class:`_schema.Column` - mapped attribute.
+
+Synopsis
+========
+
+Given ``Person`` as a model with a primary key and JSON data field.
+While this field may have any number of elements encoded within it,
+we would like to refer to the element called ``name`` individually
+as a dedicated attribute which behaves like a standalone column::
+
+    from sqlalchemy import Column, JSON, Integer
+    from sqlalchemy.ext.declarative import declarative_base
+    from sqlalchemy.ext.indexable import index_property
+
+    Base = declarative_base()
+
+
+    class Person(Base):
+        __tablename__ = "person"
+
+        id = Column(Integer, primary_key=True)
+        data = Column(JSON)
+
+        name = index_property("data", "name")
+
+Above, the ``name`` attribute now behaves like a mapped column.   We
+can compose a new ``Person`` and set the value of ``name``::
+
+    >>> person = Person(name="Alchemist")
+
+The value is now accessible::
+
+    >>> person.name
+    'Alchemist'
+
+Behind the scenes, the JSON field was initialized to a new blank dictionary
+and the field was set::
+
+    >>> person.data
+    {'name': 'Alchemist'}
+
+The field is mutable in place::
+
+    >>> person.name = "Renamed"
+    >>> person.name
+    'Renamed'
+    >>> person.data
+    {'name': 'Renamed'}
+
+When using :class:`.index_property`, the change that we make to the indexable
+structure is also automatically tracked as history; we no longer need
+to use :class:`~.mutable.MutableDict` in order to track this change
+for the unit of work.
+
+Deletions work normally as well::
+
+    >>> del person.name
+    >>> person.data
+    {}
+
+Above, deletion of ``person.name`` deletes the value from the dictionary,
+but not the dictionary itself.
+
+A missing key will produce ``AttributeError``::
+
+    >>> person = Person()
+    >>> person.name
+    AttributeError: 'name'
+
+Unless you set a default value::
+
+    >>> class Person(Base):
+    ...     __tablename__ = "person"
+    ...
+    ...     id = Column(Integer, primary_key=True)
+    ...     data = Column(JSON)
+    ...
+    ...     name = index_property("data", "name", default=None)  # See default
+
+    >>> person = Person()
+    >>> print(person.name)
+    None
+
+
+The attributes are also accessible at the class level.
+Below, we illustrate ``Person.name`` used to generate
+an indexed SQL criteria::
+
+    >>> from sqlalchemy.orm import Session
+    >>> session = Session()
+    >>> query = session.query(Person).filter(Person.name == "Alchemist")
+
+The above query is equivalent to::
+
+    >>> query = session.query(Person).filter(Person.data["name"] == "Alchemist")
+
+Multiple :class:`.index_property` objects can be chained to produce
+multiple levels of indexing::
+
+    from sqlalchemy import Column, JSON, Integer
+    from sqlalchemy.ext.declarative import declarative_base
+    from sqlalchemy.ext.indexable import index_property
+
+    Base = declarative_base()
+
+
+    class Person(Base):
+        __tablename__ = "person"
+
+        id = Column(Integer, primary_key=True)
+        data = Column(JSON)
+
+        birthday = index_property("data", "birthday")
+        year = index_property("birthday", "year")
+        month = index_property("birthday", "month")
+        day = index_property("birthday", "day")
+
+Above, a query such as::
+
+    q = session.query(Person).filter(Person.year == "1980")
+
+On a PostgreSQL backend, the above query will render as:
+
+.. sourcecode:: sql
+
+    SELECT person.id, person.data
+    FROM person
+    WHERE person.data -> %(data_1)s -> %(param_1)s = %(param_2)s
+
+Default Values
+==============
+
+:class:`.index_property` includes special behaviors for when the indexed
+data structure does not exist, and a set operation is called:
+
+* For an :class:`.index_property` that is given an integer index value,
+  the default data structure will be a Python list of ``None`` values,
+  at least as long as the index value; the value is then set at its
+  place in the list.  This means for an index value of zero, the list
+  will be initialized to ``[None]`` before setting the given value,
+  and for an index value of five, the list will be initialized to
+  ``[None, None, None, None, None]`` before setting the fifth element
+  to the given value.   Note that an existing list is **not** extended
+  in place to receive a value.
+
+* for an :class:`.index_property` that is given any other kind of index
+  value (e.g. strings usually), a Python dictionary is used as the
+  default data structure.
+
+* The default data structure can be set to any Python callable using the
+  :paramref:`.index_property.datatype` parameter, overriding the previous
+  rules.
+
+
+Subclassing
+===========
+
+:class:`.index_property` can be subclassed, in particular for the common
+use case of providing coercion of values or SQL expressions as they are
+accessed.  Below is a common recipe for use with a PostgreSQL JSON type,
+where we want to also include automatic casting plus ``astext()``::
+
+    class pg_json_property(index_property):
+        def __init__(self, attr_name, index, cast_type):
+            super(pg_json_property, self).__init__(attr_name, index)
+            self.cast_type = cast_type
+
+        def expr(self, model):
+            expr = super(pg_json_property, self).expr(model)
+            return expr.astext.cast(self.cast_type)
+
+The above subclass can be used with the PostgreSQL-specific
+version of :class:`_postgresql.JSON`::
+
+    from sqlalchemy import Column, Integer
+    from sqlalchemy.ext.declarative import declarative_base
+    from sqlalchemy.dialects.postgresql import JSON
+
+    Base = declarative_base()
+
+
+    class Person(Base):
+        __tablename__ = "person"
+
+        id = Column(Integer, primary_key=True)
+        data = Column(JSON)
+
+        age = pg_json_property("data", "age", Integer)
+
+The ``age`` attribute at the instance level works as before; however
+when rendering SQL, PostgreSQL's ``->>`` operator will be used
+for indexed access, instead of the usual index operator of ``->``::
+
+    >>> query = session.query(Person).filter(Person.age < 20)
+
+The above query will render:
+.. sourcecode:: sql
+
+    SELECT person.id, person.data
+    FROM person
+    WHERE CAST(person.data ->> %(data_1)s AS INTEGER) < %(param_1)s
+
+"""  # noqa
+from .. import inspect
+from ..ext.hybrid import hybrid_property
+from ..orm.attributes import flag_modified
+
+
+__all__ = ["index_property"]
+
+
+class index_property(hybrid_property):  # noqa
+    """A property generator. The generated property describes an object
+    attribute that corresponds to an :class:`_types.Indexable`
+    column.
+
+    .. seealso::
+
+        :mod:`sqlalchemy.ext.indexable`
+
+    """
+
+    _NO_DEFAULT_ARGUMENT = object()
+
+    def __init__(
+        self,
+        attr_name,
+        index,
+        default=_NO_DEFAULT_ARGUMENT,
+        datatype=None,
+        mutable=True,
+        onebased=True,
+    ):
+        """Create a new :class:`.index_property`.
+
+        :param attr_name:
+            An attribute name of an `Indexable` typed column, or other
+            attribute that returns an indexable structure.
+        :param index:
+            The index to be used for getting and setting this value.  This
+            should be the Python-side index value for integers.
+        :param default:
+            A value which will be returned instead of `AttributeError`
+            when there is not a value at given index.
+        :param datatype: default datatype to use when the field is empty.
+            By default, this is derived from the type of index used; a
+            Python list for an integer index, or a Python dictionary for
+            any other style of index.   For a list, the list will be
+            initialized to a list of None values that is at least
+            ``index`` elements long.
+        :param mutable: if False, writes and deletes to the attribute will
+            be disallowed.
+        :param onebased: assume the SQL representation of this value is
+            one-based; that is, the first index in SQL is 1, not zero.
+        """
+
+        if mutable:
+            super().__init__(self.fget, self.fset, self.fdel, self.expr)
+        else:
+            super().__init__(self.fget, None, None, self.expr)
+        self.attr_name = attr_name
+        self.index = index
+        self.default = default
+        is_numeric = isinstance(index, int)
+        onebased = is_numeric and onebased
+
+        if datatype is not None:
+            self.datatype = datatype
+        else:
+            if is_numeric:
+                self.datatype = lambda: [None for x in range(index + 1)]
+            else:
+                self.datatype = dict
+        self.onebased = onebased
+
+    def _fget_default(self, err=None):
+        if self.default == self._NO_DEFAULT_ARGUMENT:
+            raise AttributeError(self.attr_name) from err
+        else:
+            return self.default
+
+    def fget(self, instance):
+        attr_name = self.attr_name
+        column_value = getattr(instance, attr_name)
+        if column_value is None:
+            return self._fget_default()
+        try:
+            value = column_value[self.index]
+        except (KeyError, IndexError) as err:
+            return self._fget_default(err)
+        else:
+            return value
+
+    def fset(self, instance, value):
+        attr_name = self.attr_name
+        column_value = getattr(instance, attr_name, None)
+        if column_value is None:
+            column_value = self.datatype()
+            setattr(instance, attr_name, column_value)
+        column_value[self.index] = value
+        setattr(instance, attr_name, column_value)
+        if attr_name in inspect(instance).mapper.attrs:
+            flag_modified(instance, attr_name)
+
+    def fdel(self, instance):
+        attr_name = self.attr_name
+        column_value = getattr(instance, attr_name)
+        if column_value is None:
+            raise AttributeError(self.attr_name)
+        try:
+            del column_value[self.index]
+        except KeyError as err:
+            raise AttributeError(self.attr_name) from err
+        else:
+            setattr(instance, attr_name, column_value)
+            flag_modified(instance, attr_name)
+
+    def expr(self, model):
+        column = getattr(model, self.attr_name)
+        index = self.index
+        if self.onebased:
+            index += 1
+        return column[index]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/instrumentation.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/instrumentation.py
new file mode 100644
index 00000000..8bb01985
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/instrumentation.py
@@ -0,0 +1,450 @@
+# ext/instrumentation.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""Extensible class instrumentation.
+
+The :mod:`sqlalchemy.ext.instrumentation` package provides for alternate
+systems of class instrumentation within the ORM.  Class instrumentation
+refers to how the ORM places attributes on the class which maintain
+data and track changes to that data, as well as event hooks installed
+on the class.
+
+.. note::
+    The extension package is provided for the benefit of integration
+    with other object management packages, which already perform
+    their own instrumentation.  It is not intended for general use.
+
+For examples of how the instrumentation extension is used,
+see the example :ref:`examples_instrumentation`.
+
+"""
+import weakref
+
+from .. import util
+from ..orm import attributes
+from ..orm import base as orm_base
+from ..orm import collections
+from ..orm import exc as orm_exc
+from ..orm import instrumentation as orm_instrumentation
+from ..orm import util as orm_util
+from ..orm.instrumentation import _default_dict_getter
+from ..orm.instrumentation import _default_manager_getter
+from ..orm.instrumentation import _default_opt_manager_getter
+from ..orm.instrumentation import _default_state_getter
+from ..orm.instrumentation import ClassManager
+from ..orm.instrumentation import InstrumentationFactory
+
+
+INSTRUMENTATION_MANAGER = "__sa_instrumentation_manager__"
+"""Attribute, elects custom instrumentation when present on a mapped class.
+
+Allows a class to specify a slightly or wildly different technique for
+tracking changes made to mapped attributes and collections.
+
+Only one instrumentation implementation is allowed in a given object
+inheritance hierarchy.
+
+The value of this attribute must be a callable and will be passed a class
+object.  The callable must return one of:
+
+  - An instance of an :class:`.InstrumentationManager` or subclass
+  - An object implementing all or some of InstrumentationManager (TODO)
+  - A dictionary of callables, implementing all or some of the above (TODO)
+  - An instance of a :class:`.ClassManager` or subclass
+
+This attribute is consulted by SQLAlchemy instrumentation
+resolution, once the :mod:`sqlalchemy.ext.instrumentation` module
+has been imported.  If custom finders are installed in the global
+instrumentation_finders list, they may or may not choose to honor this
+attribute.
+
+"""
+
+
+def find_native_user_instrumentation_hook(cls):
+    """Find user-specified instrumentation management for a class."""
+    return getattr(cls, INSTRUMENTATION_MANAGER, None)
+
+
+instrumentation_finders = [find_native_user_instrumentation_hook]
+"""An extensible sequence of callables which return instrumentation
+implementations
+
+When a class is registered, each callable will be passed a class object.
+If None is returned, the
+next finder in the sequence is consulted.  Otherwise the return must be an
+instrumentation factory that follows the same guidelines as
+sqlalchemy.ext.instrumentation.INSTRUMENTATION_MANAGER.
+
+By default, the only finder is find_native_user_instrumentation_hook, which
+searches for INSTRUMENTATION_MANAGER.  If all finders return None, standard
+ClassManager instrumentation is used.
+
+"""
+
+
+class ExtendedInstrumentationRegistry(InstrumentationFactory):
+    """Extends :class:`.InstrumentationFactory` with additional
+    bookkeeping, to accommodate multiple types of
+    class managers.
+
+    """
+
+    _manager_finders = weakref.WeakKeyDictionary()
+    _state_finders = weakref.WeakKeyDictionary()
+    _dict_finders = weakref.WeakKeyDictionary()
+    _extended = False
+
+    def _locate_extended_factory(self, class_):
+        for finder in instrumentation_finders:
+            factory = finder(class_)
+            if factory is not None:
+                manager = self._extended_class_manager(class_, factory)
+                return manager, factory
+        else:
+            return None, None
+
+    def _check_conflicts(self, class_, factory):
+        existing_factories = self._collect_management_factories_for(
+            class_
+        ).difference([factory])
+        if existing_factories:
+            raise TypeError(
+                "multiple instrumentation implementations specified "
+                "in %s inheritance hierarchy: %r"
+                % (class_.__name__, list(existing_factories))
+            )
+
+    def _extended_class_manager(self, class_, factory):
+        manager = factory(class_)
+        if not isinstance(manager, ClassManager):
+            manager = _ClassInstrumentationAdapter(class_, manager)
+
+        if factory != ClassManager and not self._extended:
+            # somebody invoked a custom ClassManager.
+            # reinstall global "getter" functions with the more
+            # expensive ones.
+            self._extended = True
+            _install_instrumented_lookups()
+
+        self._manager_finders[class_] = manager.manager_getter()
+        self._state_finders[class_] = manager.state_getter()
+        self._dict_finders[class_] = manager.dict_getter()
+        return manager
+
+    def _collect_management_factories_for(self, cls):
+        """Return a collection of factories in play or specified for a
+        hierarchy.
+
+        Traverses the entire inheritance graph of a cls and returns a
+        collection of instrumentation factories for those classes. Factories
+        are extracted from active ClassManagers, if available, otherwise
+        instrumentation_finders is consulted.
+
+        """
+        hierarchy = util.class_hierarchy(cls)
+        factories = set()
+        for member in hierarchy:
+            manager = self.opt_manager_of_class(member)
+            if manager is not None:
+                factories.add(manager.factory)
+            else:
+                for finder in instrumentation_finders:
+                    factory = finder(member)
+                    if factory is not None:
+                        break
+                else:
+                    factory = None
+                factories.add(factory)
+        factories.discard(None)
+        return factories
+
+    def unregister(self, class_):
+        super().unregister(class_)
+        if class_ in self._manager_finders:
+            del self._manager_finders[class_]
+            del self._state_finders[class_]
+            del self._dict_finders[class_]
+
+    def opt_manager_of_class(self, cls):
+        try:
+            finder = self._manager_finders.get(
+                cls, _default_opt_manager_getter
+            )
+        except TypeError:
+            # due to weakref lookup on invalid object
+            return None
+        else:
+            return finder(cls)
+
+    def manager_of_class(self, cls):
+        try:
+            finder = self._manager_finders.get(cls, _default_manager_getter)
+        except TypeError:
+            # due to weakref lookup on invalid object
+            raise orm_exc.UnmappedClassError(
+                cls, f"Can't locate an instrumentation manager for class {cls}"
+            )
+        else:
+            manager = finder(cls)
+            if manager is None:
+                raise orm_exc.UnmappedClassError(
+                    cls,
+                    f"Can't locate an instrumentation manager for class {cls}",
+                )
+            return manager
+
+    def state_of(self, instance):
+        if instance is None:
+            raise AttributeError("None has no persistent state.")
+        return self._state_finders.get(
+            instance.__class__, _default_state_getter
+        )(instance)
+
+    def dict_of(self, instance):
+        if instance is None:
+            raise AttributeError("None has no persistent state.")
+        return self._dict_finders.get(
+            instance.__class__, _default_dict_getter
+        )(instance)
+
+
+orm_instrumentation._instrumentation_factory = _instrumentation_factory = (
+    ExtendedInstrumentationRegistry()
+)
+orm_instrumentation.instrumentation_finders = instrumentation_finders
+
+
+class InstrumentationManager:
+    """User-defined class instrumentation extension.
+
+    :class:`.InstrumentationManager` can be subclassed in order
+    to change
+    how class instrumentation proceeds. This class exists for
+    the purposes of integration with other object management
+    frameworks which would like to entirely modify the
+    instrumentation methodology of the ORM, and is not intended
+    for regular usage.  For interception of class instrumentation
+    events, see :class:`.InstrumentationEvents`.
+
+    The API for this class should be considered as semi-stable,
+    and may change slightly with new releases.
+
+    """
+
+    # r4361 added a mandatory (cls) constructor to this interface.
+    # given that, perhaps class_ should be dropped from all of these
+    # signatures.
+
+    def __init__(self, class_):
+        pass
+
+    def manage(self, class_, manager):
+        setattr(class_, "_default_class_manager", manager)
+
+    def unregister(self, class_, manager):
+        delattr(class_, "_default_class_manager")
+
+    def manager_getter(self, class_):
+        def get(cls):
+            return cls._default_class_manager
+
+        return get
+
+    def instrument_attribute(self, class_, key, inst):
+        pass
+
+    def post_configure_attribute(self, class_, key, inst):
+        pass
+
+    def install_descriptor(self, class_, key, inst):
+        setattr(class_, key, inst)
+
+    def uninstall_descriptor(self, class_, key):
+        delattr(class_, key)
+
+    def install_member(self, class_, key, implementation):
+        setattr(class_, key, implementation)
+
+    def uninstall_member(self, class_, key):
+        delattr(class_, key)
+
+    def instrument_collection_class(self, class_, key, collection_class):
+        return collections.prepare_instrumentation(collection_class)
+
+    def get_instance_dict(self, class_, instance):
+        return instance.__dict__
+
+    def initialize_instance_dict(self, class_, instance):
+        pass
+
+    def install_state(self, class_, instance, state):
+        setattr(instance, "_default_state", state)
+
+    def remove_state(self, class_, instance):
+        delattr(instance, "_default_state")
+
+    def state_getter(self, class_):
+        return lambda instance: getattr(instance, "_default_state")
+
+    def dict_getter(self, class_):
+        return lambda inst: self.get_instance_dict(class_, inst)
+
+
+class _ClassInstrumentationAdapter(ClassManager):
+    """Adapts a user-defined InstrumentationManager to a ClassManager."""
+
+    def __init__(self, class_, override):
+        self._adapted = override
+        self._get_state = self._adapted.state_getter(class_)
+        self._get_dict = self._adapted.dict_getter(class_)
+
+        ClassManager.__init__(self, class_)
+
+    def manage(self):
+        self._adapted.manage(self.class_, self)
+
+    def unregister(self):
+        self._adapted.unregister(self.class_, self)
+
+    def manager_getter(self):
+        return self._adapted.manager_getter(self.class_)
+
+    def instrument_attribute(self, key, inst, propagated=False):
+        ClassManager.instrument_attribute(self, key, inst, propagated)
+        if not propagated:
+            self._adapted.instrument_attribute(self.class_, key, inst)
+
+    def post_configure_attribute(self, key):
+        super().post_configure_attribute(key)
+        self._adapted.post_configure_attribute(self.class_, key, self[key])
+
+    def install_descriptor(self, key, inst):
+        self._adapted.install_descriptor(self.class_, key, inst)
+
+    def uninstall_descriptor(self, key):
+        self._adapted.uninstall_descriptor(self.class_, key)
+
+    def install_member(self, key, implementation):
+        self._adapted.install_member(self.class_, key, implementation)
+
+    def uninstall_member(self, key):
+        self._adapted.uninstall_member(self.class_, key)
+
+    def instrument_collection_class(self, key, collection_class):
+        return self._adapted.instrument_collection_class(
+            self.class_, key, collection_class
+        )
+
+    def initialize_collection(self, key, state, factory):
+        delegate = getattr(self._adapted, "initialize_collection", None)
+        if delegate:
+            return delegate(key, state, factory)
+        else:
+            return ClassManager.initialize_collection(
+                self, key, state, factory
+            )
+
+    def new_instance(self, state=None):
+        instance = self.class_.__new__(self.class_)
+        self.setup_instance(instance, state)
+        return instance
+
+    def _new_state_if_none(self, instance):
+        """Install a default InstanceState if none is present.
+
+        A private convenience method used by the __init__ decorator.
+        """
+        if self.has_state(instance):
+            return False
+        else:
+            return self.setup_instance(instance)
+
+    def setup_instance(self, instance, state=None):
+        self._adapted.initialize_instance_dict(self.class_, instance)
+
+        if state is None:
+            state = self._state_constructor(instance, self)
+
+        # the given instance is assumed to have no state
+        self._adapted.install_state(self.class_, instance, state)
+        return state
+
+    def teardown_instance(self, instance):
+        self._adapted.remove_state(self.class_, instance)
+
+    def has_state(self, instance):
+        try:
+            self._get_state(instance)
+        except orm_exc.NO_STATE:
+            return False
+        else:
+            return True
+
+    def state_getter(self):
+        return self._get_state
+
+    def dict_getter(self):
+        return self._get_dict
+
+
+def _install_instrumented_lookups():
+    """Replace global class/object management functions
+    with ExtendedInstrumentationRegistry implementations, which
+    allow multiple types of class managers to be present,
+    at the cost of performance.
+
+    This function is called only by ExtendedInstrumentationRegistry
+    and unit tests specific to this behavior.
+
+    The _reinstall_default_lookups() function can be called
+    after this one to re-establish the default functions.
+
+    """
+    _install_lookups(
+        dict(
+            instance_state=_instrumentation_factory.state_of,
+            instance_dict=_instrumentation_factory.dict_of,
+            manager_of_class=_instrumentation_factory.manager_of_class,
+            opt_manager_of_class=_instrumentation_factory.opt_manager_of_class,
+        )
+    )
+
+
+def _reinstall_default_lookups():
+    """Restore simplified lookups."""
+    _install_lookups(
+        dict(
+            instance_state=_default_state_getter,
+            instance_dict=_default_dict_getter,
+            manager_of_class=_default_manager_getter,
+            opt_manager_of_class=_default_opt_manager_getter,
+        )
+    )
+    _instrumentation_factory._extended = False
+
+
+def _install_lookups(lookups):
+    global instance_state, instance_dict
+    global manager_of_class, opt_manager_of_class
+    instance_state = lookups["instance_state"]
+    instance_dict = lookups["instance_dict"]
+    manager_of_class = lookups["manager_of_class"]
+    opt_manager_of_class = lookups["opt_manager_of_class"]
+    orm_base.instance_state = attributes.instance_state = (
+        orm_instrumentation.instance_state
+    ) = instance_state
+    orm_base.instance_dict = attributes.instance_dict = (
+        orm_instrumentation.instance_dict
+    ) = instance_dict
+    orm_base.manager_of_class = attributes.manager_of_class = (
+        orm_instrumentation.manager_of_class
+    ) = manager_of_class
+    orm_base.opt_manager_of_class = orm_util.opt_manager_of_class = (
+        attributes.opt_manager_of_class
+    ) = orm_instrumentation.opt_manager_of_class = opt_manager_of_class
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mutable.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mutable.py
new file mode 100644
index 00000000..ed618c38
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mutable.py
@@ -0,0 +1,1095 @@
+# ext/mutable.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+r"""Provide support for tracking of in-place changes to scalar values,
+which are propagated into ORM change events on owning parent objects.
+
+.. _mutable_scalars:
+
+Establishing Mutability on Scalar Column Values
+===============================================
+
+A typical example of a "mutable" structure is a Python dictionary.
+Following the example introduced in :ref:`types_toplevel`, we
+begin with a custom type that marshals Python dictionaries into
+JSON strings before being persisted::
+
+    from sqlalchemy.types import TypeDecorator, VARCHAR
+    import json
+
+
+    class JSONEncodedDict(TypeDecorator):
+        "Represents an immutable structure as a json-encoded string."
+
+        impl = VARCHAR
+
+        def process_bind_param(self, value, dialect):
+            if value is not None:
+                value = json.dumps(value)
+            return value
+
+        def process_result_value(self, value, dialect):
+            if value is not None:
+                value = json.loads(value)
+            return value
+
+The usage of ``json`` is only for the purposes of example. The
+:mod:`sqlalchemy.ext.mutable` extension can be used
+with any type whose target Python type may be mutable, including
+:class:`.PickleType`, :class:`_postgresql.ARRAY`, etc.
+
+When using the :mod:`sqlalchemy.ext.mutable` extension, the value itself
+tracks all parents which reference it.  Below, we illustrate a simple
+version of the :class:`.MutableDict` dictionary object, which applies
+the :class:`.Mutable` mixin to a plain Python dictionary::
+
+    from sqlalchemy.ext.mutable import Mutable
+
+
+    class MutableDict(Mutable, dict):
+        @classmethod
+        def coerce(cls, key, value):
+            "Convert plain dictionaries to MutableDict."
+
+            if not isinstance(value, MutableDict):
+                if isinstance(value, dict):
+                    return MutableDict(value)
+
+                # this call will raise ValueError
+                return Mutable.coerce(key, value)
+            else:
+                return value
+
+        def __setitem__(self, key, value):
+            "Detect dictionary set events and emit change events."
+
+            dict.__setitem__(self, key, value)
+            self.changed()
+
+        def __delitem__(self, key):
+            "Detect dictionary del events and emit change events."
+
+            dict.__delitem__(self, key)
+            self.changed()
+
+The above dictionary class takes the approach of subclassing the Python
+built-in ``dict`` to produce a dict
+subclass which routes all mutation events through ``__setitem__``.  There are
+variants on this approach, such as subclassing ``UserDict.UserDict`` or
+``collections.MutableMapping``; the part that's important to this example is
+that the :meth:`.Mutable.changed` method is called whenever an in-place
+change to the datastructure takes place.
+
+We also redefine the :meth:`.Mutable.coerce` method which will be used to
+convert any values that are not instances of ``MutableDict``, such
+as the plain dictionaries returned by the ``json`` module, into the
+appropriate type.  Defining this method is optional; we could just as well
+created our ``JSONEncodedDict`` such that it always returns an instance
+of ``MutableDict``, and additionally ensured that all calling code
+uses ``MutableDict`` explicitly.  When :meth:`.Mutable.coerce` is not
+overridden, any values applied to a parent object which are not instances
+of the mutable type will raise a ``ValueError``.
+
+Our new ``MutableDict`` type offers a class method
+:meth:`~.Mutable.as_mutable` which we can use within column metadata
+to associate with types. This method grabs the given type object or
+class and associates a listener that will detect all future mappings
+of this type, applying event listening instrumentation to the mapped
+attribute. Such as, with classical table metadata::
+
+    from sqlalchemy import Table, Column, Integer
+
+    my_data = Table(
+        "my_data",
+        metadata,
+        Column("id", Integer, primary_key=True),
+        Column("data", MutableDict.as_mutable(JSONEncodedDict)),
+    )
+
+Above, :meth:`~.Mutable.as_mutable` returns an instance of ``JSONEncodedDict``
+(if the type object was not an instance already), which will intercept any
+attributes which are mapped against this type.  Below we establish a simple
+mapping against the ``my_data`` table::
+
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class MyDataClass(Base):
+        __tablename__ = "my_data"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        data: Mapped[dict[str, str]] = mapped_column(
+            MutableDict.as_mutable(JSONEncodedDict)
+        )
+
+The ``MyDataClass.data`` member will now be notified of in place changes
+to its value.
+
+Any in-place changes to the ``MyDataClass.data`` member
+will flag the attribute as "dirty" on the parent object::
+
+    >>> from sqlalchemy.orm import Session
+
+    >>> sess = Session(some_engine)
+    >>> m1 = MyDataClass(data={"value1": "foo"})
+    >>> sess.add(m1)
+    >>> sess.commit()
+
+    >>> m1.data["value1"] = "bar"
+    >>> assert m1 in sess.dirty
+    True
+
+The ``MutableDict`` can be associated with all future instances
+of ``JSONEncodedDict`` in one step, using
+:meth:`~.Mutable.associate_with`.  This is similar to
+:meth:`~.Mutable.as_mutable` except it will intercept all occurrences
+of ``MutableDict`` in all mappings unconditionally, without
+the need to declare it individually::
+
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+
+    MutableDict.associate_with(JSONEncodedDict)
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class MyDataClass(Base):
+        __tablename__ = "my_data"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        data: Mapped[dict[str, str]] = mapped_column(JSONEncodedDict)
+
+Supporting Pickling
+--------------------
+
+The key to the :mod:`sqlalchemy.ext.mutable` extension relies upon the
+placement of a ``weakref.WeakKeyDictionary`` upon the value object, which
+stores a mapping of parent mapped objects keyed to the attribute name under
+which they are associated with this value. ``WeakKeyDictionary`` objects are
+not picklable, due to the fact that they contain weakrefs and function
+callbacks. In our case, this is a good thing, since if this dictionary were
+picklable, it could lead to an excessively large pickle size for our value
+objects that are pickled by themselves outside of the context of the parent.
+The developer responsibility here is only to provide a ``__getstate__`` method
+that excludes the :meth:`~MutableBase._parents` collection from the pickle
+stream::
+
+    class MyMutableType(Mutable):
+        def __getstate__(self):
+            d = self.__dict__.copy()
+            d.pop("_parents", None)
+            return d
+
+With our dictionary example, we need to return the contents of the dict itself
+(and also restore them on __setstate__)::
+
+    class MutableDict(Mutable, dict):
+        # ....
+
+        def __getstate__(self):
+            return dict(self)
+
+        def __setstate__(self, state):
+            self.update(state)
+
+In the case that our mutable value object is pickled as it is attached to one
+or more parent objects that are also part of the pickle, the :class:`.Mutable`
+mixin will re-establish the :attr:`.Mutable._parents` collection on each value
+object as the owning parents themselves are unpickled.
+
+Receiving Events
+----------------
+
+The :meth:`.AttributeEvents.modified` event handler may be used to receive
+an event when a mutable scalar emits a change event.  This event handler
+is called when the :func:`.attributes.flag_modified` function is called
+from within the mutable extension::
+
+    from sqlalchemy.orm import DeclarativeBase
+    from sqlalchemy.orm import Mapped
+    from sqlalchemy.orm import mapped_column
+    from sqlalchemy import event
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class MyDataClass(Base):
+        __tablename__ = "my_data"
+        id: Mapped[int] = mapped_column(primary_key=True)
+        data: Mapped[dict[str, str]] = mapped_column(
+            MutableDict.as_mutable(JSONEncodedDict)
+        )
+
+
+    @event.listens_for(MyDataClass.data, "modified")
+    def modified_json(instance, initiator):
+        print("json value modified:", instance.data)
+
+.. _mutable_composites:
+
+Establishing Mutability on Composites
+=====================================
+
+Composites are a special ORM feature which allow a single scalar attribute to
+be assigned an object value which represents information "composed" from one
+or more columns from the underlying mapped table. The usual example is that of
+a geometric "point", and is introduced in :ref:`mapper_composite`.
+
+As is the case with :class:`.Mutable`, the user-defined composite class
+subclasses :class:`.MutableComposite` as a mixin, and detects and delivers
+change events to its parents via the :meth:`.MutableComposite.changed` method.
+In the case of a composite class, the detection is usually via the usage of the
+special Python method ``__setattr__()``. In the example below, we expand upon the ``Point``
+class introduced in :ref:`mapper_composite` to include
+:class:`.MutableComposite` in its bases and to route attribute set events via
+``__setattr__`` to the :meth:`.MutableComposite.changed` method::
+
+    import dataclasses
+    from sqlalchemy.ext.mutable import MutableComposite
+
+
+    @dataclasses.dataclass
+    class Point(MutableComposite):
+        x: int
+        y: int
+
+        def __setattr__(self, key, value):
+            "Intercept set events"
+
+            # set the attribute
+            object.__setattr__(self, key, value)
+
+            # alert all parents to the change
+            self.changed()
+
+The :class:`.MutableComposite` class makes use of class mapping events to
+automatically establish listeners for any usage of :func:`_orm.composite` that
+specifies our ``Point`` type. Below, when ``Point`` is mapped to the ``Vertex``
+class, listeners are established which will route change events from ``Point``
+objects to each of the ``Vertex.start`` and ``Vertex.end`` attributes::
+
+    from sqlalchemy.orm import DeclarativeBase, Mapped
+    from sqlalchemy.orm import composite, mapped_column
+
+
+    class Base(DeclarativeBase):
+        pass
+
+
+    class Vertex(Base):
+        __tablename__ = "vertices"
+
+        id: Mapped[int] = mapped_column(primary_key=True)
+
+        start: Mapped[Point] = composite(
+            mapped_column("x1"), mapped_column("y1")
+        )
+        end: Mapped[Point] = composite(
+            mapped_column("x2"), mapped_column("y2")
+        )
+
+        def __repr__(self):
+            return f"Vertex(start={self.start}, end={self.end})"
+
+Any in-place changes to the ``Vertex.start`` or ``Vertex.end`` members
+will flag the attribute as "dirty" on the parent object:
+
+.. sourcecode:: python+sql
+
+    >>> from sqlalchemy.orm import Session
+    >>> sess = Session(engine)
+    >>> v1 = Vertex(start=Point(3, 4), end=Point(12, 15))
+    >>> sess.add(v1)
+    {sql}>>> sess.flush()
+    BEGIN (implicit)
+    INSERT INTO vertices (x1, y1, x2, y2) VALUES (?, ?, ?, ?)
+    [...] (3, 4, 12, 15)
+
+    {stop}>>> v1.end.x = 8
+    >>> assert v1 in sess.dirty
+    True
+    {sql}>>> sess.commit()
+    UPDATE vertices SET x2=? WHERE vertices.id = ?
+    [...] (8, 1)
+    COMMIT
+
+Coercing Mutable Composites
+---------------------------
+
+The :meth:`.MutableBase.coerce` method is also supported on composite types.
+In the case of :class:`.MutableComposite`, the :meth:`.MutableBase.coerce`
+method is only called for attribute set operations, not load operations.
+Overriding the :meth:`.MutableBase.coerce` method is essentially equivalent
+to using a :func:`.validates` validation routine for all attributes which
+make use of the custom composite type::
+
+    @dataclasses.dataclass
+    class Point(MutableComposite):
+        # other Point methods
+        # ...
+
+        def coerce(cls, key, value):
+            if isinstance(value, tuple):
+                value = Point(*value)
+            elif not isinstance(value, Point):
+                raise ValueError("tuple or Point expected")
+            return value
+
+Supporting Pickling
+--------------------
+
+As is the case with :class:`.Mutable`, the :class:`.MutableComposite` helper
+class uses a ``weakref.WeakKeyDictionary`` available via the
+:meth:`MutableBase._parents` attribute which isn't picklable. If we need to
+pickle instances of ``Point`` or its owning class ``Vertex``, we at least need
+to define a ``__getstate__`` that doesn't include the ``_parents`` dictionary.
+Below we define both a ``__getstate__`` and a ``__setstate__`` that package up
+the minimal form of our ``Point`` class::
+
+    @dataclasses.dataclass
+    class Point(MutableComposite):
+        # ...
+
+        def __getstate__(self):
+            return self.x, self.y
+
+        def __setstate__(self, state):
+            self.x, self.y = state
+
+As with :class:`.Mutable`, the :class:`.MutableComposite` augments the
+pickling process of the parent's object-relational state so that the
+:meth:`MutableBase._parents` collection is restored to all ``Point`` objects.
+
+"""  # noqa: E501
+
+from __future__ import annotations
+
+from collections import defaultdict
+from typing import AbstractSet
+from typing import Any
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+from weakref import WeakKeyDictionary
+
+from .. import event
+from .. import inspect
+from .. import types
+from .. import util
+from ..orm import Mapper
+from ..orm._typing import _ExternalEntityType
+from ..orm._typing import _O
+from ..orm._typing import _T
+from ..orm.attributes import AttributeEventToken
+from ..orm.attributes import flag_modified
+from ..orm.attributes import InstrumentedAttribute
+from ..orm.attributes import QueryableAttribute
+from ..orm.context import QueryContext
+from ..orm.decl_api import DeclarativeAttributeIntercept
+from ..orm.state import InstanceState
+from ..orm.unitofwork import UOWTransaction
+from ..sql._typing import _TypeEngineArgument
+from ..sql.base import SchemaEventTarget
+from ..sql.schema import Column
+from ..sql.type_api import TypeEngine
+from ..util import memoized_property
+from ..util.typing import SupportsIndex
+from ..util.typing import TypeGuard
+
+_KT = TypeVar("_KT")  # Key type.
+_VT = TypeVar("_VT")  # Value type.
+
+
+class MutableBase:
+    """Common base class to :class:`.Mutable`
+    and :class:`.MutableComposite`.
+
+    """
+
+    @memoized_property
+    def _parents(self) -> WeakKeyDictionary[Any, Any]:
+        """Dictionary of parent object's :class:`.InstanceState`->attribute
+        name on the parent.
+
+        This attribute is a so-called "memoized" property.  It initializes
+        itself with a new ``weakref.WeakKeyDictionary`` the first time
+        it is accessed, returning the same object upon subsequent access.
+
+        .. versionchanged:: 1.4 the :class:`.InstanceState` is now used
+           as the key in the weak dictionary rather than the instance
+           itself.
+
+        """
+
+        return weakref.WeakKeyDictionary()
+
+    @classmethod
+    def coerce(cls, key: str, value: Any) -> Optional[Any]:
+        """Given a value, coerce it into the target type.
+
+        Can be overridden by custom subclasses to coerce incoming
+        data into a particular type.
+
+        By default, raises ``ValueError``.
+
+        This method is called in different scenarios depending on if
+        the parent class is of type :class:`.Mutable` or of type
+        :class:`.MutableComposite`.  In the case of the former, it is called
+        for both attribute-set operations as well as during ORM loading
+        operations.  For the latter, it is only called during attribute-set
+        operations; the mechanics of the :func:`.composite` construct
+        handle coercion during load operations.
+
+
+        :param key: string name of the ORM-mapped attribute being set.
+        :param value: the incoming value.
+        :return: the method should return the coerced value, or raise
+         ``ValueError`` if the coercion cannot be completed.
+
+        """
+        if value is None:
+            return None
+        msg = "Attribute '%s' does not accept objects of type %s"
+        raise ValueError(msg % (key, type(value)))
+
+    @classmethod
+    def _get_listen_keys(cls, attribute: QueryableAttribute[Any]) -> Set[str]:
+        """Given a descriptor attribute, return a ``set()`` of the attribute
+        keys which indicate a change in the state of this attribute.
+
+        This is normally just ``set([attribute.key])``, but can be overridden
+        to provide for additional keys.  E.g. a :class:`.MutableComposite`
+        augments this set with the attribute keys associated with the columns
+        that comprise the composite value.
+
+        This collection is consulted in the case of intercepting the
+        :meth:`.InstanceEvents.refresh` and
+        :meth:`.InstanceEvents.refresh_flush` events, which pass along a list
+        of attribute names that have been refreshed; the list is compared
+        against this set to determine if action needs to be taken.
+
+        """
+        return {attribute.key}
+
+    @classmethod
+    def _listen_on_attribute(
+        cls,
+        attribute: QueryableAttribute[Any],
+        coerce: bool,
+        parent_cls: _ExternalEntityType[Any],
+    ) -> None:
+        """Establish this type as a mutation listener for the given
+        mapped descriptor.
+
+        """
+        key = attribute.key
+        if parent_cls is not attribute.class_:
+            return
+
+        # rely on "propagate" here
+        parent_cls = attribute.class_
+
+        listen_keys = cls._get_listen_keys(attribute)
+
+        def load(state: InstanceState[_O], *args: Any) -> None:
+            """Listen for objects loaded or refreshed.
+
+            Wrap the target data member's value with
+            ``Mutable``.
+
+            """
+            val = state.dict.get(key, None)
+            if val is not None:
+                if coerce:
+                    val = cls.coerce(key, val)
+                    state.dict[key] = val
+                val._parents[state] = key
+
+        def load_attrs(
+            state: InstanceState[_O],
+            ctx: Union[object, QueryContext, UOWTransaction],
+            attrs: Iterable[Any],
+        ) -> None:
+            if not attrs or listen_keys.intersection(attrs):
+                load(state)
+
+        def set_(
+            target: InstanceState[_O],
+            value: MutableBase | None,
+            oldvalue: MutableBase | None,
+            initiator: AttributeEventToken,
+        ) -> MutableBase | None:
+            """Listen for set/replace events on the target
+            data member.
+
+            Establish a weak reference to the parent object
+            on the incoming value, remove it for the one
+            outgoing.
+
+            """
+            if value is oldvalue:
+                return value
+
+            if not isinstance(value, cls):
+                value = cls.coerce(key, value)
+            if value is not None:
+                value._parents[target] = key
+            if isinstance(oldvalue, cls):
+                oldvalue._parents.pop(inspect(target), None)
+            return value
+
+        def pickle(
+            state: InstanceState[_O], state_dict: Dict[str, Any]
+        ) -> None:
+            val = state.dict.get(key, None)
+            if val is not None:
+                if "ext.mutable.values" not in state_dict:
+                    state_dict["ext.mutable.values"] = defaultdict(list)
+                state_dict["ext.mutable.values"][key].append(val)
+
+        def unpickle(
+            state: InstanceState[_O], state_dict: Dict[str, Any]
+        ) -> None:
+            if "ext.mutable.values" in state_dict:
+                collection = state_dict["ext.mutable.values"]
+                if isinstance(collection, list):
+                    # legacy format
+                    for val in collection:
+                        val._parents[state] = key
+                else:
+                    for val in state_dict["ext.mutable.values"][key]:
+                        val._parents[state] = key
+
+        event.listen(
+            parent_cls,
+            "_sa_event_merge_wo_load",
+            load,
+            raw=True,
+            propagate=True,
+        )
+
+        event.listen(parent_cls, "load", load, raw=True, propagate=True)
+        event.listen(
+            parent_cls, "refresh", load_attrs, raw=True, propagate=True
+        )
+        event.listen(
+            parent_cls, "refresh_flush", load_attrs, raw=True, propagate=True
+        )
+        event.listen(
+            attribute, "set", set_, raw=True, retval=True, propagate=True
+        )
+        event.listen(parent_cls, "pickle", pickle, raw=True, propagate=True)
+        event.listen(
+            parent_cls, "unpickle", unpickle, raw=True, propagate=True
+        )
+
+
+class Mutable(MutableBase):
+    """Mixin that defines transparent propagation of change
+    events to a parent object.
+
+    See the example in :ref:`mutable_scalars` for usage information.
+
+    """
+
+    def changed(self) -> None:
+        """Subclasses should call this method whenever change events occur."""
+
+        for parent, key in self._parents.items():
+            flag_modified(parent.obj(), key)
+
+    @classmethod
+    def associate_with_attribute(
+        cls, attribute: InstrumentedAttribute[_O]
+    ) -> None:
+        """Establish this type as a mutation listener for the given
+        mapped descriptor.
+
+        """
+        cls._listen_on_attribute(attribute, True, attribute.class_)
+
+    @classmethod
+    def associate_with(cls, sqltype: type) -> None:
+        """Associate this wrapper with all future mapped columns
+        of the given type.
+
+        This is a convenience method that calls
+        ``associate_with_attribute`` automatically.
+
+        .. warning::
+
+           The listeners established by this method are *global*
+           to all mappers, and are *not* garbage collected.   Only use
+           :meth:`.associate_with` for types that are permanent to an
+           application, not with ad-hoc types else this will cause unbounded
+           growth in memory usage.
+
+        """
+
+        def listen_for_type(mapper: Mapper[_O], class_: type) -> None:
+            if mapper.non_primary:
+                return
+            for prop in mapper.column_attrs:
+                if isinstance(prop.columns[0].type, sqltype):
+                    cls.associate_with_attribute(getattr(class_, prop.key))
+
+        event.listen(Mapper, "mapper_configured", listen_for_type)
+
+    @classmethod
+    def as_mutable(cls, sqltype: _TypeEngineArgument[_T]) -> TypeEngine[_T]:
+        """Associate a SQL type with this mutable Python type.
+
+        This establishes listeners that will detect ORM mappings against
+        the given type, adding mutation event trackers to those mappings.
+
+        The type is returned, unconditionally as an instance, so that
+        :meth:`.as_mutable` can be used inline::
+
+            Table(
+                "mytable",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("data", MyMutableType.as_mutable(PickleType)),
+            )
+
+        Note that the returned type is always an instance, even if a class
+        is given, and that only columns which are declared specifically with
+        that type instance receive additional instrumentation.
+
+        To associate a particular mutable type with all occurrences of a
+        particular type, use the :meth:`.Mutable.associate_with` classmethod
+        of the particular :class:`.Mutable` subclass to establish a global
+        association.
+
+        .. warning::
+
+           The listeners established by this method are *global*
+           to all mappers, and are *not* garbage collected.   Only use
+           :meth:`.as_mutable` for types that are permanent to an application,
+           not with ad-hoc types else this will cause unbounded growth
+           in memory usage.
+
+        """
+        sqltype = types.to_instance(sqltype)
+
+        # a SchemaType will be copied when the Column is copied,
+        # and we'll lose our ability to link that type back to the original.
+        # so track our original type w/ columns
+        if isinstance(sqltype, SchemaEventTarget):
+
+            @event.listens_for(sqltype, "before_parent_attach")
+            def _add_column_memo(
+                sqltyp: TypeEngine[Any],
+                parent: Column[_T],
+            ) -> None:
+                parent.info["_ext_mutable_orig_type"] = sqltyp
+
+            schema_event_check = True
+        else:
+            schema_event_check = False
+
+        def listen_for_type(
+            mapper: Mapper[_T],
+            class_: Union[DeclarativeAttributeIntercept, type],
+        ) -> None:
+            if mapper.non_primary:
+                return
+            _APPLIED_KEY = "_ext_mutable_listener_applied"
+
+            for prop in mapper.column_attrs:
+                if (
+                    # all Mutable types refer to a Column that's mapped,
+                    # since this is the only kind of Core target the ORM can
+                    # "mutate"
+                    isinstance(prop.expression, Column)
+                    and (
+                        (
+                            schema_event_check
+                            and prop.expression.info.get(
+                                "_ext_mutable_orig_type"
+                            )
+                            is sqltype
+                        )
+                        or prop.expression.type is sqltype
+                    )
+                ):
+                    if not prop.expression.info.get(_APPLIED_KEY, False):
+                        prop.expression.info[_APPLIED_KEY] = True
+                        cls.associate_with_attribute(getattr(class_, prop.key))
+
+        event.listen(Mapper, "mapper_configured", listen_for_type)
+
+        return sqltype
+
+
+class MutableComposite(MutableBase):
+    """Mixin that defines transparent propagation of change
+    events on a SQLAlchemy "composite" object to its
+    owning parent or parents.
+
+    See the example in :ref:`mutable_composites` for usage information.
+
+    """
+
+    @classmethod
+    def _get_listen_keys(cls, attribute: QueryableAttribute[_O]) -> Set[str]:
+        return {attribute.key}.union(attribute.property._attribute_keys)
+
+    def changed(self) -> None:
+        """Subclasses should call this method whenever change events occur."""
+
+        for parent, key in self._parents.items():
+            prop = parent.mapper.get_property(key)
+            for value, attr_name in zip(
+                prop._composite_values_from_instance(self),
+                prop._attribute_keys,
+            ):
+                setattr(parent.obj(), attr_name, value)
+
+
+def _setup_composite_listener() -> None:
+    def _listen_for_type(mapper: Mapper[_T], class_: type) -> None:
+        for prop in mapper.iterate_properties:
+            if (
+                hasattr(prop, "composite_class")
+                and isinstance(prop.composite_class, type)
+                and issubclass(prop.composite_class, MutableComposite)
+            ):
+                prop.composite_class._listen_on_attribute(
+                    getattr(class_, prop.key), False, class_
+                )
+
+    if not event.contains(Mapper, "mapper_configured", _listen_for_type):
+        event.listen(Mapper, "mapper_configured", _listen_for_type)
+
+
+_setup_composite_listener()
+
+
+class MutableDict(Mutable, Dict[_KT, _VT]):
+    """A dictionary type that implements :class:`.Mutable`.
+
+    The :class:`.MutableDict` object implements a dictionary that will
+    emit change events to the underlying mapping when the contents of
+    the dictionary are altered, including when values are added or removed.
+
+    Note that :class:`.MutableDict` does **not** apply mutable tracking to  the
+    *values themselves* inside the dictionary. Therefore it is not a sufficient
+    solution for the use case of tracking deep changes to a *recursive*
+    dictionary structure, such as a JSON structure.  To support this use case,
+    build a subclass of  :class:`.MutableDict` that provides appropriate
+    coercion to the values placed in the dictionary so that they too are
+    "mutable", and emit events up to their parent structure.
+
+    .. seealso::
+
+        :class:`.MutableList`
+
+        :class:`.MutableSet`
+
+    """
+
+    def __setitem__(self, key: _KT, value: _VT) -> None:
+        """Detect dictionary set events and emit change events."""
+        super().__setitem__(key, value)
+        self.changed()
+
+    if TYPE_CHECKING:
+        # from https://github.com/python/mypy/issues/14858
+
+        @overload
+        def setdefault(
+            self: MutableDict[_KT, Optional[_T]], key: _KT, value: None = None
+        ) -> Optional[_T]: ...
+
+        @overload
+        def setdefault(self, key: _KT, value: _VT) -> _VT: ...
+
+        def setdefault(self, key: _KT, value: object = None) -> object: ...
+
+    else:
+
+        def setdefault(self, *arg):  # noqa: F811
+            result = super().setdefault(*arg)
+            self.changed()
+            return result
+
+    def __delitem__(self, key: _KT) -> None:
+        """Detect dictionary del events and emit change events."""
+        super().__delitem__(key)
+        self.changed()
+
+    def update(self, *a: Any, **kw: _VT) -> None:
+        super().update(*a, **kw)
+        self.changed()
+
+    if TYPE_CHECKING:
+
+        @overload
+        def pop(self, __key: _KT) -> _VT: ...
+
+        @overload
+        def pop(self, __key: _KT, __default: _VT | _T) -> _VT | _T: ...
+
+        def pop(
+            self, __key: _KT, __default: _VT | _T | None = None
+        ) -> _VT | _T: ...
+
+    else:
+
+        def pop(self, *arg):  # noqa: F811
+            result = super().pop(*arg)
+            self.changed()
+            return result
+
+    def popitem(self) -> Tuple[_KT, _VT]:
+        result = super().popitem()
+        self.changed()
+        return result
+
+    def clear(self) -> None:
+        super().clear()
+        self.changed()
+
+    @classmethod
+    def coerce(cls, key: str, value: Any) -> MutableDict[_KT, _VT] | None:
+        """Convert plain dictionary to instance of this class."""
+        if not isinstance(value, cls):
+            if isinstance(value, dict):
+                return cls(value)
+            return Mutable.coerce(key, value)
+        else:
+            return value
+
+    def __getstate__(self) -> Dict[_KT, _VT]:
+        return dict(self)
+
+    def __setstate__(
+        self, state: Union[Dict[str, int], Dict[str, str]]
+    ) -> None:
+        self.update(state)
+
+
+class MutableList(Mutable, List[_T]):
+    """A list type that implements :class:`.Mutable`.
+
+    The :class:`.MutableList` object implements a list that will
+    emit change events to the underlying mapping when the contents of
+    the list are altered, including when values are added or removed.
+
+    Note that :class:`.MutableList` does **not** apply mutable tracking to  the
+    *values themselves* inside the list. Therefore it is not a sufficient
+    solution for the use case of tracking deep changes to a *recursive*
+    mutable structure, such as a JSON structure.  To support this use case,
+    build a subclass of  :class:`.MutableList` that provides appropriate
+    coercion to the values placed in the dictionary so that they too are
+    "mutable", and emit events up to their parent structure.
+
+    .. seealso::
+
+        :class:`.MutableDict`
+
+        :class:`.MutableSet`
+
+    """
+
+    def __reduce_ex__(
+        self, proto: SupportsIndex
+    ) -> Tuple[type, Tuple[List[int]]]:
+        return (self.__class__, (list(self),))
+
+    # needed for backwards compatibility with
+    # older pickles
+    def __setstate__(self, state: Iterable[_T]) -> None:
+        self[:] = state
+
+    def is_scalar(self, value: _T | Iterable[_T]) -> TypeGuard[_T]:
+        return not util.is_non_string_iterable(value)
+
+    def is_iterable(self, value: _T | Iterable[_T]) -> TypeGuard[Iterable[_T]]:
+        return util.is_non_string_iterable(value)
+
+    def __setitem__(
+        self, index: SupportsIndex | slice, value: _T | Iterable[_T]
+    ) -> None:
+        """Detect list set events and emit change events."""
+        if isinstance(index, SupportsIndex) and self.is_scalar(value):
+            super().__setitem__(index, value)
+        elif isinstance(index, slice) and self.is_iterable(value):
+            super().__setitem__(index, value)
+        self.changed()
+
+    def __delitem__(self, index: SupportsIndex | slice) -> None:
+        """Detect list del events and emit change events."""
+        super().__delitem__(index)
+        self.changed()
+
+    def pop(self, *arg: SupportsIndex) -> _T:
+        result = super().pop(*arg)
+        self.changed()
+        return result
+
+    def append(self, x: _T) -> None:
+        super().append(x)
+        self.changed()
+
+    def extend(self, x: Iterable[_T]) -> None:
+        super().extend(x)
+        self.changed()
+
+    def __iadd__(self, x: Iterable[_T]) -> MutableList[_T]:  # type: ignore[override,misc] # noqa: E501
+        self.extend(x)
+        return self
+
+    def insert(self, i: SupportsIndex, x: _T) -> None:
+        super().insert(i, x)
+        self.changed()
+
+    def remove(self, i: _T) -> None:
+        super().remove(i)
+        self.changed()
+
+    def clear(self) -> None:
+        super().clear()
+        self.changed()
+
+    def sort(self, **kw: Any) -> None:
+        super().sort(**kw)
+        self.changed()
+
+    def reverse(self) -> None:
+        super().reverse()
+        self.changed()
+
+    @classmethod
+    def coerce(
+        cls, key: str, value: MutableList[_T] | _T
+    ) -> Optional[MutableList[_T]]:
+        """Convert plain list to instance of this class."""
+        if not isinstance(value, cls):
+            if isinstance(value, list):
+                return cls(value)
+            return Mutable.coerce(key, value)
+        else:
+            return value
+
+
+class MutableSet(Mutable, Set[_T]):
+    """A set type that implements :class:`.Mutable`.
+
+    The :class:`.MutableSet` object implements a set that will
+    emit change events to the underlying mapping when the contents of
+    the set are altered, including when values are added or removed.
+
+    Note that :class:`.MutableSet` does **not** apply mutable tracking to  the
+    *values themselves* inside the set. Therefore it is not a sufficient
+    solution for the use case of tracking deep changes to a *recursive*
+    mutable structure.  To support this use case,
+    build a subclass of  :class:`.MutableSet` that provides appropriate
+    coercion to the values placed in the dictionary so that they too are
+    "mutable", and emit events up to their parent structure.
+
+    .. seealso::
+
+        :class:`.MutableDict`
+
+        :class:`.MutableList`
+
+
+    """
+
+    def update(self, *arg: Iterable[_T]) -> None:
+        super().update(*arg)
+        self.changed()
+
+    def intersection_update(self, *arg: Iterable[Any]) -> None:
+        super().intersection_update(*arg)
+        self.changed()
+
+    def difference_update(self, *arg: Iterable[Any]) -> None:
+        super().difference_update(*arg)
+        self.changed()
+
+    def symmetric_difference_update(self, *arg: Iterable[_T]) -> None:
+        super().symmetric_difference_update(*arg)
+        self.changed()
+
+    def __ior__(self, other: AbstractSet[_T]) -> MutableSet[_T]:  # type: ignore[override,misc] # noqa: E501
+        self.update(other)
+        return self
+
+    def __iand__(self, other: AbstractSet[object]) -> MutableSet[_T]:
+        self.intersection_update(other)
+        return self
+
+    def __ixor__(self, other: AbstractSet[_T]) -> MutableSet[_T]:  # type: ignore[override,misc] # noqa: E501
+        self.symmetric_difference_update(other)
+        return self
+
+    def __isub__(self, other: AbstractSet[object]) -> MutableSet[_T]:  # type: ignore[misc] # noqa: E501
+        self.difference_update(other)
+        return self
+
+    def add(self, elem: _T) -> None:
+        super().add(elem)
+        self.changed()
+
+    def remove(self, elem: _T) -> None:
+        super().remove(elem)
+        self.changed()
+
+    def discard(self, elem: _T) -> None:
+        super().discard(elem)
+        self.changed()
+
+    def pop(self, *arg: Any) -> _T:
+        result = super().pop(*arg)
+        self.changed()
+        return result
+
+    def clear(self) -> None:
+        super().clear()
+        self.changed()
+
+    @classmethod
+    def coerce(cls, index: str, value: Any) -> Optional[MutableSet[_T]]:
+        """Convert plain set to instance of this class."""
+        if not isinstance(value, cls):
+            if isinstance(value, set):
+                return cls(value)
+            return Mutable.coerce(index, value)
+        else:
+            return value
+
+    def __getstate__(self) -> Set[_T]:
+        return set(self)
+
+    def __setstate__(self, state: Iterable[_T]) -> None:
+        self.update(state)
+
+    def __reduce_ex__(
+        self, proto: SupportsIndex
+    ) -> Tuple[type, Tuple[List[int]]]:
+        return (self.__class__, (list(self),))
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__init__.py
new file mode 100644
index 00000000..b5827cb8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__init__.py
@@ -0,0 +1,6 @@
+# ext/mypy/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0e3d7f89
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/apply.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/apply.cpython-312.pyc
new file mode 100644
index 00000000..8f364a23
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/apply.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/decl_class.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/decl_class.cpython-312.pyc
new file mode 100644
index 00000000..bf959a1a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/decl_class.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/infer.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/infer.cpython-312.pyc
new file mode 100644
index 00000000..6d0eb9de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/infer.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/names.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/names.cpython-312.pyc
new file mode 100644
index 00000000..38cca867
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/names.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/plugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/plugin.cpython-312.pyc
new file mode 100644
index 00000000..de2cb92f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/plugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..dd793e67
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/apply.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/apply.py
new file mode 100644
index 00000000..02908cc1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/apply.py
@@ -0,0 +1,324 @@
+# ext/mypy/apply.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import List
+from typing import Optional
+from typing import Union
+
+from mypy.nodes import ARG_NAMED_OPT
+from mypy.nodes import Argument
+from mypy.nodes import AssignmentStmt
+from mypy.nodes import CallExpr
+from mypy.nodes import ClassDef
+from mypy.nodes import MDEF
+from mypy.nodes import MemberExpr
+from mypy.nodes import NameExpr
+from mypy.nodes import RefExpr
+from mypy.nodes import StrExpr
+from mypy.nodes import SymbolTableNode
+from mypy.nodes import TempNode
+from mypy.nodes import TypeInfo
+from mypy.nodes import Var
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.plugins.common import add_method_to_class
+from mypy.types import AnyType
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import NoneTyp
+from mypy.types import ProperType
+from mypy.types import TypeOfAny
+from mypy.types import UnboundType
+from mypy.types import UnionType
+
+from . import infer
+from . import util
+from .names import expr_to_mapped_constructor
+from .names import NAMED_TYPE_SQLA_MAPPED
+
+
+def apply_mypy_mapped_attr(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    item: Union[NameExpr, StrExpr],
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    if isinstance(item, NameExpr):
+        name = item.name
+    elif isinstance(item, StrExpr):
+        name = item.value
+    else:
+        return None
+
+    for stmt in cls.defs.body:
+        if (
+            isinstance(stmt, AssignmentStmt)
+            and isinstance(stmt.lvalues[0], NameExpr)
+            and stmt.lvalues[0].name == name
+        ):
+            break
+    else:
+        util.fail(api, f"Can't find mapped attribute {name}", cls)
+        return None
+
+    if stmt.type is None:
+        util.fail(
+            api,
+            "Statement linked from _mypy_mapped_attrs has no "
+            "typing information",
+            stmt,
+        )
+        return None
+
+    left_hand_explicit_type = get_proper_type(stmt.type)
+    assert isinstance(
+        left_hand_explicit_type, (Instance, UnionType, UnboundType)
+    )
+
+    attributes.append(
+        util.SQLAlchemyAttribute(
+            name=name,
+            line=item.line,
+            column=item.column,
+            typ=left_hand_explicit_type,
+            info=cls.info,
+        )
+    )
+
+    apply_type_to_mapped_statement(
+        api, stmt, stmt.lvalues[0], left_hand_explicit_type, None
+    )
+
+
+def re_apply_declarative_assignments(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    """For multiple class passes, re-apply our left-hand side types as mypy
+    seems to reset them in place.
+
+    """
+    mapped_attr_lookup = {attr.name: attr for attr in attributes}
+    update_cls_metadata = False
+
+    for stmt in cls.defs.body:
+        # for a re-apply, all of our statements are AssignmentStmt;
+        # @declared_attr calls will have been converted and this
+        # currently seems to be preserved by mypy (but who knows if this
+        # will change).
+        if (
+            isinstance(stmt, AssignmentStmt)
+            and isinstance(stmt.lvalues[0], NameExpr)
+            and stmt.lvalues[0].name in mapped_attr_lookup
+            and isinstance(stmt.lvalues[0].node, Var)
+        ):
+            left_node = stmt.lvalues[0].node
+
+            python_type_for_type = mapped_attr_lookup[
+                stmt.lvalues[0].name
+            ].type
+
+            left_node_proper_type = get_proper_type(left_node.type)
+
+            # if we have scanned an UnboundType and now there's a more
+            # specific type than UnboundType, call the re-scan so we
+            # can get that set up correctly
+            if (
+                isinstance(python_type_for_type, UnboundType)
+                and not isinstance(left_node_proper_type, UnboundType)
+                and (
+                    isinstance(stmt.rvalue, CallExpr)
+                    and isinstance(stmt.rvalue.callee, MemberExpr)
+                    and isinstance(stmt.rvalue.callee.expr, NameExpr)
+                    and stmt.rvalue.callee.expr.node is not None
+                    and stmt.rvalue.callee.expr.node.fullname
+                    == NAMED_TYPE_SQLA_MAPPED
+                    and stmt.rvalue.callee.name == "_empty_constructor"
+                    and isinstance(stmt.rvalue.args[0], CallExpr)
+                    and isinstance(stmt.rvalue.args[0].callee, RefExpr)
+                )
+            ):
+                new_python_type_for_type = (
+                    infer.infer_type_from_right_hand_nameexpr(
+                        api,
+                        stmt,
+                        left_node,
+                        left_node_proper_type,
+                        stmt.rvalue.args[0].callee,
+                    )
+                )
+
+                if new_python_type_for_type is not None and not isinstance(
+                    new_python_type_for_type, UnboundType
+                ):
+                    python_type_for_type = new_python_type_for_type
+
+                    # update the SQLAlchemyAttribute with the better
+                    # information
+                    mapped_attr_lookup[stmt.lvalues[0].name].type = (
+                        python_type_for_type
+                    )
+
+                    update_cls_metadata = True
+
+            if (
+                not isinstance(left_node.type, Instance)
+                or left_node.type.type.fullname != NAMED_TYPE_SQLA_MAPPED
+            ):
+                assert python_type_for_type is not None
+                left_node.type = api.named_type(
+                    NAMED_TYPE_SQLA_MAPPED, [python_type_for_type]
+                )
+
+    if update_cls_metadata:
+        util.set_mapped_attributes(cls.info, attributes)
+
+
+def apply_type_to_mapped_statement(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    lvalue: NameExpr,
+    left_hand_explicit_type: Optional[ProperType],
+    python_type_for_type: Optional[ProperType],
+) -> None:
+    """Apply the Mapped[<type>] annotation and right hand object to a
+    declarative assignment statement.
+
+    This converts a Python declarative class statement such as::
+
+        class User(Base):
+            # ...
+
+            attrname = Column(Integer)
+
+    To one that describes the final Python behavior to Mypy::
+
+    ... format: off
+
+        class User(Base):
+            # ...
+
+            attrname : Mapped[Optional[int]] = <meaningless temp node>
+
+    ... format: on
+
+    """
+    left_node = lvalue.node
+    assert isinstance(left_node, Var)
+
+    # to be completely honest I have no idea what the difference between
+    # left_node.type and stmt.type is, what it means if these are different
+    # vs. the same, why in order to get tests to pass I have to assign
+    # to stmt.type for the second case and not the first.  this is complete
+    # trying every combination until it works stuff.
+
+    if left_hand_explicit_type is not None:
+        lvalue.is_inferred_def = False
+        left_node.type = api.named_type(
+            NAMED_TYPE_SQLA_MAPPED, [left_hand_explicit_type]
+        )
+    else:
+        lvalue.is_inferred_def = False
+        left_node.type = api.named_type(
+            NAMED_TYPE_SQLA_MAPPED,
+            (
+                [AnyType(TypeOfAny.special_form)]
+                if python_type_for_type is None
+                else [python_type_for_type]
+            ),
+        )
+
+    # so to have it skip the right side totally, we can do this:
+    # stmt.rvalue = TempNode(AnyType(TypeOfAny.special_form))
+
+    # however, if we instead manufacture a new node that uses the old
+    # one, then we can still get type checking for the call itself,
+    # e.g. the Column, relationship() call, etc.
+
+    # rewrite the node as:
+    # <attr> : Mapped[<typ>] =
+    # _sa_Mapped._empty_constructor(<original CallExpr from rvalue>)
+    # the original right-hand side is maintained so it gets type checked
+    # internally
+    stmt.rvalue = expr_to_mapped_constructor(stmt.rvalue)
+
+    if stmt.type is not None and python_type_for_type is not None:
+        stmt.type = python_type_for_type
+
+
+def add_additional_orm_attributes(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    """Apply __init__, __table__ and other attributes to the mapped class."""
+
+    info = util.info_for_cls(cls, api)
+
+    if info is None:
+        return
+
+    is_base = util.get_is_base(info)
+
+    if "__init__" not in info.names and not is_base:
+        mapped_attr_names = {attr.name: attr.type for attr in attributes}
+
+        for base in info.mro[1:-1]:
+            if "sqlalchemy" not in info.metadata:
+                continue
+
+            base_cls_attributes = util.get_mapped_attributes(base, api)
+            if base_cls_attributes is None:
+                continue
+
+            for attr in base_cls_attributes:
+                mapped_attr_names.setdefault(attr.name, attr.type)
+
+        arguments = []
+        for name, typ in mapped_attr_names.items():
+            if typ is None:
+                typ = AnyType(TypeOfAny.special_form)
+            arguments.append(
+                Argument(
+                    variable=Var(name, typ),
+                    type_annotation=typ,
+                    initializer=TempNode(typ),
+                    kind=ARG_NAMED_OPT,
+                )
+            )
+
+        add_method_to_class(api, cls, "__init__", arguments, NoneTyp())
+
+    if "__table__" not in info.names and util.get_has_table(info):
+        _apply_placeholder_attr_to_class(
+            api, cls, "sqlalchemy.sql.schema.Table", "__table__"
+        )
+    if not is_base:
+        _apply_placeholder_attr_to_class(
+            api, cls, "sqlalchemy.orm.mapper.Mapper", "__mapper__"
+        )
+
+
+def _apply_placeholder_attr_to_class(
+    api: SemanticAnalyzerPluginInterface,
+    cls: ClassDef,
+    qualified_name: str,
+    attrname: str,
+) -> None:
+    sym = api.lookup_fully_qualified_or_none(qualified_name)
+    if sym:
+        assert isinstance(sym.node, TypeInfo)
+        type_: ProperType = Instance(sym.node, [])
+    else:
+        type_ = AnyType(TypeOfAny.special_form)
+    var = Var(attrname)
+    var._fullname = cls.fullname + "." + attrname
+    var.info = cls.info
+    var.type = type_
+    cls.info.names[attrname] = SymbolTableNode(MDEF, var)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/decl_class.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/decl_class.py
new file mode 100644
index 00000000..2ce7ad56
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/decl_class.py
@@ -0,0 +1,515 @@
+# ext/mypy/decl_class.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import List
+from typing import Optional
+from typing import Union
+
+from mypy.nodes import AssignmentStmt
+from mypy.nodes import CallExpr
+from mypy.nodes import ClassDef
+from mypy.nodes import Decorator
+from mypy.nodes import LambdaExpr
+from mypy.nodes import ListExpr
+from mypy.nodes import MemberExpr
+from mypy.nodes import NameExpr
+from mypy.nodes import PlaceholderNode
+from mypy.nodes import RefExpr
+from mypy.nodes import StrExpr
+from mypy.nodes import SymbolNode
+from mypy.nodes import SymbolTableNode
+from mypy.nodes import TempNode
+from mypy.nodes import TypeInfo
+from mypy.nodes import Var
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.types import AnyType
+from mypy.types import CallableType
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import NoneType
+from mypy.types import ProperType
+from mypy.types import Type
+from mypy.types import TypeOfAny
+from mypy.types import UnboundType
+from mypy.types import UnionType
+
+from . import apply
+from . import infer
+from . import names
+from . import util
+
+
+def scan_declarative_assignments_and_apply_types(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    is_mixin_scan: bool = False,
+) -> Optional[List[util.SQLAlchemyAttribute]]:
+    info = util.info_for_cls(cls, api)
+
+    if info is None:
+        # this can occur during cached passes
+        return None
+    elif cls.fullname.startswith("builtins"):
+        return None
+
+    mapped_attributes: Optional[List[util.SQLAlchemyAttribute]] = (
+        util.get_mapped_attributes(info, api)
+    )
+
+    # used by assign.add_additional_orm_attributes among others
+    util.establish_as_sqlalchemy(info)
+
+    if mapped_attributes is not None:
+        # ensure that a class that's mapped is always picked up by
+        # its mapped() decorator or declarative metaclass before
+        # it would be detected as an unmapped mixin class
+
+        if not is_mixin_scan:
+            # mypy can call us more than once.  it then *may* have reset the
+            # left hand side of everything, but not the right that we removed,
+            # removing our ability to re-scan.   but we have the types
+            # here, so lets re-apply them, or if we have an UnboundType,
+            # we can re-scan
+
+            apply.re_apply_declarative_assignments(cls, api, mapped_attributes)
+
+        return mapped_attributes
+
+    mapped_attributes = []
+
+    if not cls.defs.body:
+        # when we get a mixin class from another file, the body is
+        # empty (!) but the names are in the symbol table.  so use that.
+
+        for sym_name, sym in info.names.items():
+            _scan_symbol_table_entry(
+                cls, api, sym_name, sym, mapped_attributes
+            )
+    else:
+        for stmt in util.flatten_typechecking(cls.defs.body):
+            if isinstance(stmt, AssignmentStmt):
+                _scan_declarative_assignment_stmt(
+                    cls, api, stmt, mapped_attributes
+                )
+            elif isinstance(stmt, Decorator):
+                _scan_declarative_decorator_stmt(
+                    cls, api, stmt, mapped_attributes
+                )
+    _scan_for_mapped_bases(cls, api)
+
+    if not is_mixin_scan:
+        apply.add_additional_orm_attributes(cls, api, mapped_attributes)
+
+    util.set_mapped_attributes(info, mapped_attributes)
+
+    return mapped_attributes
+
+
+def _scan_symbol_table_entry(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    name: str,
+    value: SymbolTableNode,
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    """Extract mapping information from a SymbolTableNode that's in the
+    type.names dictionary.
+
+    """
+    value_type = get_proper_type(value.type)
+    if not isinstance(value_type, Instance):
+        return
+
+    left_hand_explicit_type = None
+    type_id = names.type_id_for_named_node(value_type.type)
+    # type_id = names._type_id_for_unbound_type(value.type.type, cls, api)
+
+    err = False
+
+    # TODO: this is nearly the same logic as that of
+    # _scan_declarative_decorator_stmt, likely can be merged
+    if type_id in {
+        names.MAPPED,
+        names.RELATIONSHIP,
+        names.COMPOSITE_PROPERTY,
+        names.MAPPER_PROPERTY,
+        names.SYNONYM_PROPERTY,
+        names.COLUMN_PROPERTY,
+    }:
+        if value_type.args:
+            left_hand_explicit_type = get_proper_type(value_type.args[0])
+        else:
+            err = True
+    elif type_id is names.COLUMN:
+        if not value_type.args:
+            err = True
+        else:
+            typeengine_arg: Union[ProperType, TypeInfo] = get_proper_type(
+                value_type.args[0]
+            )
+            if isinstance(typeengine_arg, Instance):
+                typeengine_arg = typeengine_arg.type
+
+            if isinstance(typeengine_arg, (UnboundType, TypeInfo)):
+                sym = api.lookup_qualified(typeengine_arg.name, typeengine_arg)
+                if sym is not None and isinstance(sym.node, TypeInfo):
+                    if names.has_base_type_id(sym.node, names.TYPEENGINE):
+                        left_hand_explicit_type = UnionType(
+                            [
+                                infer.extract_python_type_from_typeengine(
+                                    api, sym.node, []
+                                ),
+                                NoneType(),
+                            ]
+                        )
+                    else:
+                        util.fail(
+                            api,
+                            "Column type should be a TypeEngine "
+                            "subclass not '{}'".format(sym.node.fullname),
+                            value_type,
+                        )
+
+    if err:
+        msg = (
+            "Can't infer type from attribute {} on class {}. "
+            "please specify a return type from this function that is "
+            "one of: Mapped[<python type>], relationship[<target class>], "
+            "Column[<TypeEngine>], MapperProperty[<python type>]"
+        )
+        util.fail(api, msg.format(name, cls.name), cls)
+
+        left_hand_explicit_type = AnyType(TypeOfAny.special_form)
+
+    if left_hand_explicit_type is not None:
+        assert value.node is not None
+        attributes.append(
+            util.SQLAlchemyAttribute(
+                name=name,
+                line=value.node.line,
+                column=value.node.column,
+                typ=left_hand_explicit_type,
+                info=cls.info,
+            )
+        )
+
+
+def _scan_declarative_decorator_stmt(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    stmt: Decorator,
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    """Extract mapping information from a @declared_attr in a declarative
+    class.
+
+    E.g.::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            @declared_attr
+            def updated_at(cls) -> Column[DateTime]:
+                return Column(DateTime)
+
+    Will resolve in mypy as::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            updated_at: Mapped[Optional[datetime.datetime]]
+
+    """
+    for dec in stmt.decorators:
+        if (
+            isinstance(dec, (NameExpr, MemberExpr, SymbolNode))
+            and names.type_id_for_named_node(dec) is names.DECLARED_ATTR
+        ):
+            break
+    else:
+        return
+
+    dec_index = cls.defs.body.index(stmt)
+
+    left_hand_explicit_type: Optional[ProperType] = None
+
+    if util.name_is_dunder(stmt.name):
+        # for dunder names like __table_args__, __tablename__,
+        # __mapper_args__ etc., rewrite these as simple assignment
+        # statements; otherwise mypy doesn't like if the decorated
+        # function has an annotation like ``cls: Type[Foo]`` because
+        # it isn't @classmethod
+        any_ = AnyType(TypeOfAny.special_form)
+        left_node = NameExpr(stmt.var.name)
+        left_node.node = stmt.var
+        new_stmt = AssignmentStmt([left_node], TempNode(any_))
+        new_stmt.type = left_node.node.type
+        cls.defs.body[dec_index] = new_stmt
+        return
+    elif isinstance(stmt.func.type, CallableType):
+        func_type = stmt.func.type.ret_type
+        if isinstance(func_type, UnboundType):
+            type_id = names.type_id_for_unbound_type(func_type, cls, api)
+        else:
+            # this does not seem to occur unless the type argument is
+            # incorrect
+            return
+
+        if (
+            type_id
+            in {
+                names.MAPPED,
+                names.RELATIONSHIP,
+                names.COMPOSITE_PROPERTY,
+                names.MAPPER_PROPERTY,
+                names.SYNONYM_PROPERTY,
+                names.COLUMN_PROPERTY,
+            }
+            and func_type.args
+        ):
+            left_hand_explicit_type = get_proper_type(func_type.args[0])
+        elif type_id is names.COLUMN and func_type.args:
+            typeengine_arg = func_type.args[0]
+            if isinstance(typeengine_arg, UnboundType):
+                sym = api.lookup_qualified(typeengine_arg.name, typeengine_arg)
+                if sym is not None and isinstance(sym.node, TypeInfo):
+                    if names.has_base_type_id(sym.node, names.TYPEENGINE):
+                        left_hand_explicit_type = UnionType(
+                            [
+                                infer.extract_python_type_from_typeengine(
+                                    api, sym.node, []
+                                ),
+                                NoneType(),
+                            ]
+                        )
+                    else:
+                        util.fail(
+                            api,
+                            "Column type should be a TypeEngine "
+                            "subclass not '{}'".format(sym.node.fullname),
+                            func_type,
+                        )
+
+    if left_hand_explicit_type is None:
+        # no type on the decorated function.  our option here is to
+        # dig into the function body and get the return type, but they
+        # should just have an annotation.
+        msg = (
+            "Can't infer type from @declared_attr on function '{}';  "
+            "please specify a return type from this function that is "
+            "one of: Mapped[<python type>], relationship[<target class>], "
+            "Column[<TypeEngine>], MapperProperty[<python type>]"
+        )
+        util.fail(api, msg.format(stmt.var.name), stmt)
+
+        left_hand_explicit_type = AnyType(TypeOfAny.special_form)
+
+    left_node = NameExpr(stmt.var.name)
+    left_node.node = stmt.var
+
+    # totally feeling around in the dark here as I don't totally understand
+    # the significance of UnboundType.  It seems to be something that is
+    # not going to do what's expected when it is applied as the type of
+    # an AssignmentStatement.  So do a feeling-around-in-the-dark version
+    # of converting it to the regular Instance/TypeInfo/UnionType structures
+    # we see everywhere else.
+    if isinstance(left_hand_explicit_type, UnboundType):
+        left_hand_explicit_type = get_proper_type(
+            util.unbound_to_instance(api, left_hand_explicit_type)
+        )
+
+    left_node.node.type = api.named_type(
+        names.NAMED_TYPE_SQLA_MAPPED, [left_hand_explicit_type]
+    )
+
+    # this will ignore the rvalue entirely
+    # rvalue = TempNode(AnyType(TypeOfAny.special_form))
+
+    # rewrite the node as:
+    # <attr> : Mapped[<typ>] =
+    # _sa_Mapped._empty_constructor(lambda: <function body>)
+    # the function body is maintained so it gets type checked internally
+    rvalue = names.expr_to_mapped_constructor(
+        LambdaExpr(stmt.func.arguments, stmt.func.body)
+    )
+
+    new_stmt = AssignmentStmt([left_node], rvalue)
+    new_stmt.type = left_node.node.type
+
+    attributes.append(
+        util.SQLAlchemyAttribute(
+            name=left_node.name,
+            line=stmt.line,
+            column=stmt.column,
+            typ=left_hand_explicit_type,
+            info=cls.info,
+        )
+    )
+    cls.defs.body[dec_index] = new_stmt
+
+
+def _scan_declarative_assignment_stmt(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    attributes: List[util.SQLAlchemyAttribute],
+) -> None:
+    """Extract mapping information from an assignment statement in a
+    declarative class.
+
+    """
+    lvalue = stmt.lvalues[0]
+    if not isinstance(lvalue, NameExpr):
+        return
+
+    sym = cls.info.names.get(lvalue.name)
+
+    # this establishes that semantic analysis has taken place, which
+    # means the nodes are populated and we are called from an appropriate
+    # hook.
+    assert sym is not None
+    node = sym.node
+
+    if isinstance(node, PlaceholderNode):
+        return
+
+    assert node is lvalue.node
+    assert isinstance(node, Var)
+
+    if node.name == "__abstract__":
+        if api.parse_bool(stmt.rvalue) is True:
+            util.set_is_base(cls.info)
+        return
+    elif node.name == "__tablename__":
+        util.set_has_table(cls.info)
+    elif node.name.startswith("__"):
+        return
+    elif node.name == "_mypy_mapped_attrs":
+        if not isinstance(stmt.rvalue, ListExpr):
+            util.fail(api, "_mypy_mapped_attrs is expected to be a list", stmt)
+        else:
+            for item in stmt.rvalue.items:
+                if isinstance(item, (NameExpr, StrExpr)):
+                    apply.apply_mypy_mapped_attr(cls, api, item, attributes)
+
+    left_hand_mapped_type: Optional[Type] = None
+    left_hand_explicit_type: Optional[ProperType] = None
+
+    if node.is_inferred or node.type is None:
+        if isinstance(stmt.type, UnboundType):
+            # look for an explicit Mapped[] type annotation on the left
+            # side with nothing on the right
+
+            # print(stmt.type)
+            # Mapped?[Optional?[A?]]
+
+            left_hand_explicit_type = stmt.type
+
+            if stmt.type.name == "Mapped":
+                mapped_sym = api.lookup_qualified("Mapped", cls)
+                if (
+                    mapped_sym is not None
+                    and mapped_sym.node is not None
+                    and names.type_id_for_named_node(mapped_sym.node)
+                    is names.MAPPED
+                ):
+                    left_hand_explicit_type = get_proper_type(
+                        stmt.type.args[0]
+                    )
+                    left_hand_mapped_type = stmt.type
+
+            # TODO: do we need to convert from unbound for this case?
+            # left_hand_explicit_type = util._unbound_to_instance(
+            #     api, left_hand_explicit_type
+            # )
+    else:
+        node_type = get_proper_type(node.type)
+        if (
+            isinstance(node_type, Instance)
+            and names.type_id_for_named_node(node_type.type) is names.MAPPED
+        ):
+            # print(node.type)
+            # sqlalchemy.orm.attributes.Mapped[<python type>]
+            left_hand_explicit_type = get_proper_type(node_type.args[0])
+            left_hand_mapped_type = node_type
+        else:
+            # print(node.type)
+            # <python type>
+            left_hand_explicit_type = node_type
+            left_hand_mapped_type = None
+
+    if isinstance(stmt.rvalue, TempNode) and left_hand_mapped_type is not None:
+        # annotation without assignment and Mapped is present
+        # as type annotation
+        # equivalent to using _infer_type_from_left_hand_type_only.
+
+        python_type_for_type = left_hand_explicit_type
+    elif isinstance(stmt.rvalue, CallExpr) and isinstance(
+        stmt.rvalue.callee, RefExpr
+    ):
+        python_type_for_type = infer.infer_type_from_right_hand_nameexpr(
+            api, stmt, node, left_hand_explicit_type, stmt.rvalue.callee
+        )
+
+        if python_type_for_type is None:
+            return
+
+    else:
+        return
+
+    assert python_type_for_type is not None
+
+    attributes.append(
+        util.SQLAlchemyAttribute(
+            name=node.name,
+            line=stmt.line,
+            column=stmt.column,
+            typ=python_type_for_type,
+            info=cls.info,
+        )
+    )
+
+    apply.apply_type_to_mapped_statement(
+        api,
+        stmt,
+        lvalue,
+        left_hand_explicit_type,
+        python_type_for_type,
+    )
+
+
+def _scan_for_mapped_bases(
+    cls: ClassDef,
+    api: SemanticAnalyzerPluginInterface,
+) -> None:
+    """Given a class, iterate through its superclass hierarchy to find
+    all other classes that are considered as ORM-significant.
+
+    Locates non-mapped mixins and scans them for mapped attributes to be
+    applied to subclasses.
+
+    """
+
+    info = util.info_for_cls(cls, api)
+
+    if info is None:
+        return
+
+    for base_info in info.mro[1:-1]:
+        if base_info.fullname.startswith("builtins"):
+            continue
+
+        # scan each base for mapped attributes.  if they are not already
+        # scanned (but have all their type info), that means they are unmapped
+        # mixins
+        scan_declarative_assignments_and_apply_types(
+            base_info.defn, api, is_mixin_scan=True
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/infer.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/infer.py
new file mode 100644
index 00000000..26a83cca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/infer.py
@@ -0,0 +1,590 @@
+# ext/mypy/infer.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Optional
+from typing import Sequence
+
+from mypy.maptype import map_instance_to_supertype
+from mypy.nodes import AssignmentStmt
+from mypy.nodes import CallExpr
+from mypy.nodes import Expression
+from mypy.nodes import FuncDef
+from mypy.nodes import LambdaExpr
+from mypy.nodes import MemberExpr
+from mypy.nodes import NameExpr
+from mypy.nodes import RefExpr
+from mypy.nodes import StrExpr
+from mypy.nodes import TypeInfo
+from mypy.nodes import Var
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.subtypes import is_subtype
+from mypy.types import AnyType
+from mypy.types import CallableType
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import NoneType
+from mypy.types import ProperType
+from mypy.types import TypeOfAny
+from mypy.types import UnionType
+
+from . import names
+from . import util
+
+
+def infer_type_from_right_hand_nameexpr(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+    infer_from_right_side: RefExpr,
+) -> Optional[ProperType]:
+    type_id = names.type_id_for_callee(infer_from_right_side)
+    if type_id is None:
+        return None
+    elif type_id is names.MAPPED:
+        python_type_for_type = _infer_type_from_mapped(
+            api, stmt, node, left_hand_explicit_type, infer_from_right_side
+        )
+    elif type_id is names.COLUMN:
+        python_type_for_type = _infer_type_from_decl_column(
+            api, stmt, node, left_hand_explicit_type
+        )
+    elif type_id is names.RELATIONSHIP:
+        python_type_for_type = _infer_type_from_relationship(
+            api, stmt, node, left_hand_explicit_type
+        )
+    elif type_id is names.COLUMN_PROPERTY:
+        python_type_for_type = _infer_type_from_decl_column_property(
+            api, stmt, node, left_hand_explicit_type
+        )
+    elif type_id is names.SYNONYM_PROPERTY:
+        python_type_for_type = infer_type_from_left_hand_type_only(
+            api, node, left_hand_explicit_type
+        )
+    elif type_id is names.COMPOSITE_PROPERTY:
+        python_type_for_type = _infer_type_from_decl_composite_property(
+            api, stmt, node, left_hand_explicit_type
+        )
+    else:
+        return None
+
+    return python_type_for_type
+
+
+def _infer_type_from_relationship(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+) -> Optional[ProperType]:
+    """Infer the type of mapping from a relationship.
+
+    E.g.::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            addresses = relationship(Address, uselist=True)
+
+            order: Mapped["Order"] = relationship("Order")
+
+    Will resolve in mypy as::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            addresses: Mapped[List[Address]]
+
+            order: Mapped["Order"]
+
+    """
+
+    assert isinstance(stmt.rvalue, CallExpr)
+    target_cls_arg = stmt.rvalue.args[0]
+    python_type_for_type: Optional[ProperType] = None
+
+    if isinstance(target_cls_arg, NameExpr) and isinstance(
+        target_cls_arg.node, TypeInfo
+    ):
+        # type
+        related_object_type = target_cls_arg.node
+        python_type_for_type = Instance(related_object_type, [])
+
+    # other cases not covered - an error message directs the user
+    # to set an explicit type annotation
+    #
+    # node.type == str, it's a string
+    # if isinstance(target_cls_arg, NameExpr) and isinstance(
+    #     target_cls_arg.node, Var
+    # )
+    # points to a type
+    # isinstance(target_cls_arg, NameExpr) and isinstance(
+    #     target_cls_arg.node, TypeAlias
+    # )
+    # string expression
+    # isinstance(target_cls_arg, StrExpr)
+
+    uselist_arg = util.get_callexpr_kwarg(stmt.rvalue, "uselist")
+    collection_cls_arg: Optional[Expression] = util.get_callexpr_kwarg(
+        stmt.rvalue, "collection_class"
+    )
+    type_is_a_collection = False
+
+    # this can be used to determine Optional for a many-to-one
+    # in the same way nullable=False could be used, if we start supporting
+    # that.
+    # innerjoin_arg = util.get_callexpr_kwarg(stmt.rvalue, "innerjoin")
+
+    if (
+        uselist_arg is not None
+        and api.parse_bool(uselist_arg) is True
+        and collection_cls_arg is None
+    ):
+        type_is_a_collection = True
+        if python_type_for_type is not None:
+            python_type_for_type = api.named_type(
+                names.NAMED_TYPE_BUILTINS_LIST, [python_type_for_type]
+            )
+    elif (
+        uselist_arg is None or api.parse_bool(uselist_arg) is True
+    ) and collection_cls_arg is not None:
+        type_is_a_collection = True
+        if isinstance(collection_cls_arg, CallExpr):
+            collection_cls_arg = collection_cls_arg.callee
+
+        if isinstance(collection_cls_arg, NameExpr) and isinstance(
+            collection_cls_arg.node, TypeInfo
+        ):
+            if python_type_for_type is not None:
+                # this can still be overridden by the left hand side
+                # within _infer_Type_from_left_and_inferred_right
+                python_type_for_type = Instance(
+                    collection_cls_arg.node, [python_type_for_type]
+                )
+        elif (
+            isinstance(collection_cls_arg, NameExpr)
+            and isinstance(collection_cls_arg.node, FuncDef)
+            and collection_cls_arg.node.type is not None
+        ):
+            if python_type_for_type is not None:
+                # this can still be overridden by the left hand side
+                # within _infer_Type_from_left_and_inferred_right
+
+                # TODO: handle mypy.types.Overloaded
+                if isinstance(collection_cls_arg.node.type, CallableType):
+                    rt = get_proper_type(collection_cls_arg.node.type.ret_type)
+
+                    if isinstance(rt, CallableType):
+                        callable_ret_type = get_proper_type(rt.ret_type)
+                        if isinstance(callable_ret_type, Instance):
+                            python_type_for_type = Instance(
+                                callable_ret_type.type,
+                                [python_type_for_type],
+                            )
+        else:
+            util.fail(
+                api,
+                "Expected Python collection type for "
+                "collection_class parameter",
+                stmt.rvalue,
+            )
+            python_type_for_type = None
+    elif uselist_arg is not None and api.parse_bool(uselist_arg) is False:
+        if collection_cls_arg is not None:
+            util.fail(
+                api,
+                "Sending uselist=False and collection_class at the same time "
+                "does not make sense",
+                stmt.rvalue,
+            )
+        if python_type_for_type is not None:
+            python_type_for_type = UnionType(
+                [python_type_for_type, NoneType()]
+            )
+
+    else:
+        if left_hand_explicit_type is None:
+            msg = (
+                "Can't infer scalar or collection for ORM mapped expression "
+                "assigned to attribute '{}' if both 'uselist' and "
+                "'collection_class' arguments are absent from the "
+                "relationship(); please specify a "
+                "type annotation on the left hand side."
+            )
+            util.fail(api, msg.format(node.name), node)
+
+    if python_type_for_type is None:
+        return infer_type_from_left_hand_type_only(
+            api, node, left_hand_explicit_type
+        )
+    elif left_hand_explicit_type is not None:
+        if type_is_a_collection:
+            assert isinstance(left_hand_explicit_type, Instance)
+            assert isinstance(python_type_for_type, Instance)
+            return _infer_collection_type_from_left_and_inferred_right(
+                api, node, left_hand_explicit_type, python_type_for_type
+            )
+        else:
+            return _infer_type_from_left_and_inferred_right(
+                api,
+                node,
+                left_hand_explicit_type,
+                python_type_for_type,
+            )
+    else:
+        return python_type_for_type
+
+
+def _infer_type_from_decl_composite_property(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+) -> Optional[ProperType]:
+    """Infer the type of mapping from a Composite."""
+
+    assert isinstance(stmt.rvalue, CallExpr)
+    target_cls_arg = stmt.rvalue.args[0]
+    python_type_for_type = None
+
+    if isinstance(target_cls_arg, NameExpr) and isinstance(
+        target_cls_arg.node, TypeInfo
+    ):
+        related_object_type = target_cls_arg.node
+        python_type_for_type = Instance(related_object_type, [])
+    else:
+        python_type_for_type = None
+
+    if python_type_for_type is None:
+        return infer_type_from_left_hand_type_only(
+            api, node, left_hand_explicit_type
+        )
+    elif left_hand_explicit_type is not None:
+        return _infer_type_from_left_and_inferred_right(
+            api, node, left_hand_explicit_type, python_type_for_type
+        )
+    else:
+        return python_type_for_type
+
+
+def _infer_type_from_mapped(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+    infer_from_right_side: RefExpr,
+) -> Optional[ProperType]:
+    """Infer the type of mapping from a right side expression
+    that returns Mapped.
+
+
+    """
+    assert isinstance(stmt.rvalue, CallExpr)
+
+    # (Pdb) print(stmt.rvalue.callee)
+    # NameExpr(query_expression [sqlalchemy.orm._orm_constructors.query_expression])  # noqa: E501
+    # (Pdb) stmt.rvalue.callee.node
+    # <mypy.nodes.FuncDef object at 0x7f8d92fb5940>
+    # (Pdb) stmt.rvalue.callee.node.type
+    # def [_T] (default_expr: sqlalchemy.sql.elements.ColumnElement[_T`-1] =) -> sqlalchemy.orm.base.Mapped[_T`-1]  # noqa: E501
+    # sqlalchemy.orm.base.Mapped[_T`-1]
+    # the_mapped_type = stmt.rvalue.callee.node.type.ret_type
+
+    # TODO: look at generic ref and either use that,
+    # or reconcile w/ what's present, etc.
+    the_mapped_type = util.type_for_callee(infer_from_right_side)  # noqa
+
+    return infer_type_from_left_hand_type_only(
+        api, node, left_hand_explicit_type
+    )
+
+
+def _infer_type_from_decl_column_property(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+) -> Optional[ProperType]:
+    """Infer the type of mapping from a ColumnProperty.
+
+    This includes mappings against ``column_property()`` as well as the
+    ``deferred()`` function.
+
+    """
+    assert isinstance(stmt.rvalue, CallExpr)
+
+    if stmt.rvalue.args:
+        first_prop_arg = stmt.rvalue.args[0]
+
+        if isinstance(first_prop_arg, CallExpr):
+            type_id = names.type_id_for_callee(first_prop_arg.callee)
+
+            # look for column_property() / deferred() etc with Column as first
+            # argument
+            if type_id is names.COLUMN:
+                return _infer_type_from_decl_column(
+                    api,
+                    stmt,
+                    node,
+                    left_hand_explicit_type,
+                    right_hand_expression=first_prop_arg,
+                )
+
+    if isinstance(stmt.rvalue, CallExpr):
+        type_id = names.type_id_for_callee(stmt.rvalue.callee)
+        # this is probably not strictly necessary as we have to use the left
+        # hand type for query expression in any case.  any other no-arg
+        # column prop objects would go here also
+        if type_id is names.QUERY_EXPRESSION:
+            return _infer_type_from_decl_column(
+                api,
+                stmt,
+                node,
+                left_hand_explicit_type,
+            )
+
+    return infer_type_from_left_hand_type_only(
+        api, node, left_hand_explicit_type
+    )
+
+
+def _infer_type_from_decl_column(
+    api: SemanticAnalyzerPluginInterface,
+    stmt: AssignmentStmt,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+    right_hand_expression: Optional[CallExpr] = None,
+) -> Optional[ProperType]:
+    """Infer the type of mapping from a Column.
+
+    E.g.::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            a = Column(Integer)
+
+            b = Column("b", String)
+
+            c: Mapped[int] = Column(Integer)
+
+            d: bool = Column(Boolean)
+
+    Will resolve in MyPy as::
+
+        @reg.mapped
+        class MyClass:
+            # ...
+
+            a: Mapped[int]
+
+            b: Mapped[str]
+
+            c: Mapped[int]
+
+            d: Mapped[bool]
+
+    """
+    assert isinstance(node, Var)
+
+    callee = None
+
+    if right_hand_expression is None:
+        if not isinstance(stmt.rvalue, CallExpr):
+            return None
+
+        right_hand_expression = stmt.rvalue
+
+    for column_arg in right_hand_expression.args[0:2]:
+        if isinstance(column_arg, CallExpr):
+            if isinstance(column_arg.callee, RefExpr):
+                # x = Column(String(50))
+                callee = column_arg.callee
+                type_args: Sequence[Expression] = column_arg.args
+                break
+        elif isinstance(column_arg, (NameExpr, MemberExpr)):
+            if isinstance(column_arg.node, TypeInfo):
+                # x = Column(String)
+                callee = column_arg
+                type_args = ()
+                break
+            else:
+                # x = Column(some_name, String), go to next argument
+                continue
+        elif isinstance(column_arg, (StrExpr,)):
+            # x = Column("name", String), go to next argument
+            continue
+        elif isinstance(column_arg, (LambdaExpr,)):
+            # x = Column("name", String, default=lambda: uuid.uuid4())
+            # go to next argument
+            continue
+        else:
+            assert False
+
+    if callee is None:
+        return None
+
+    if isinstance(callee.node, TypeInfo) and names.mro_has_id(
+        callee.node.mro, names.TYPEENGINE
+    ):
+        python_type_for_type = extract_python_type_from_typeengine(
+            api, callee.node, type_args
+        )
+
+        if left_hand_explicit_type is not None:
+            return _infer_type_from_left_and_inferred_right(
+                api, node, left_hand_explicit_type, python_type_for_type
+            )
+
+        else:
+            return UnionType([python_type_for_type, NoneType()])
+    else:
+        # it's not TypeEngine, it's typically implicitly typed
+        # like ForeignKey.  we can't infer from the right side.
+        return infer_type_from_left_hand_type_only(
+            api, node, left_hand_explicit_type
+        )
+
+
+def _infer_type_from_left_and_inferred_right(
+    api: SemanticAnalyzerPluginInterface,
+    node: Var,
+    left_hand_explicit_type: ProperType,
+    python_type_for_type: ProperType,
+    orig_left_hand_type: Optional[ProperType] = None,
+    orig_python_type_for_type: Optional[ProperType] = None,
+) -> Optional[ProperType]:
+    """Validate type when a left hand annotation is present and we also
+    could infer the right hand side::
+
+        attrname: SomeType = Column(SomeDBType)
+
+    """
+
+    if orig_left_hand_type is None:
+        orig_left_hand_type = left_hand_explicit_type
+    if orig_python_type_for_type is None:
+        orig_python_type_for_type = python_type_for_type
+
+    if not is_subtype(left_hand_explicit_type, python_type_for_type):
+        effective_type = api.named_type(
+            names.NAMED_TYPE_SQLA_MAPPED, [orig_python_type_for_type]
+        )
+
+        msg = (
+            "Left hand assignment '{}: {}' not compatible "
+            "with ORM mapped expression of type {}"
+        )
+        util.fail(
+            api,
+            msg.format(
+                node.name,
+                util.format_type(orig_left_hand_type, api.options),
+                util.format_type(effective_type, api.options),
+            ),
+            node,
+        )
+
+    return orig_left_hand_type
+
+
+def _infer_collection_type_from_left_and_inferred_right(
+    api: SemanticAnalyzerPluginInterface,
+    node: Var,
+    left_hand_explicit_type: Instance,
+    python_type_for_type: Instance,
+) -> Optional[ProperType]:
+    orig_left_hand_type = left_hand_explicit_type
+    orig_python_type_for_type = python_type_for_type
+
+    if left_hand_explicit_type.args:
+        left_hand_arg = get_proper_type(left_hand_explicit_type.args[0])
+        python_type_arg = get_proper_type(python_type_for_type.args[0])
+    else:
+        left_hand_arg = left_hand_explicit_type
+        python_type_arg = python_type_for_type
+
+    assert isinstance(left_hand_arg, (Instance, UnionType))
+    assert isinstance(python_type_arg, (Instance, UnionType))
+
+    return _infer_type_from_left_and_inferred_right(
+        api,
+        node,
+        left_hand_arg,
+        python_type_arg,
+        orig_left_hand_type=orig_left_hand_type,
+        orig_python_type_for_type=orig_python_type_for_type,
+    )
+
+
+def infer_type_from_left_hand_type_only(
+    api: SemanticAnalyzerPluginInterface,
+    node: Var,
+    left_hand_explicit_type: Optional[ProperType],
+) -> Optional[ProperType]:
+    """Determine the type based on explicit annotation only.
+
+    if no annotation were present, note that we need one there to know
+    the type.
+
+    """
+    if left_hand_explicit_type is None:
+        msg = (
+            "Can't infer type from ORM mapped expression "
+            "assigned to attribute '{}'; please specify a "
+            "Python type or "
+            "Mapped[<python type>] on the left hand side."
+        )
+        util.fail(api, msg.format(node.name), node)
+
+        return api.named_type(
+            names.NAMED_TYPE_SQLA_MAPPED, [AnyType(TypeOfAny.special_form)]
+        )
+
+    else:
+        # use type from the left hand side
+        return left_hand_explicit_type
+
+
+def extract_python_type_from_typeengine(
+    api: SemanticAnalyzerPluginInterface,
+    node: TypeInfo,
+    type_args: Sequence[Expression],
+) -> ProperType:
+    if node.fullname == "sqlalchemy.sql.sqltypes.Enum" and type_args:
+        first_arg = type_args[0]
+        if isinstance(first_arg, RefExpr) and isinstance(
+            first_arg.node, TypeInfo
+        ):
+            for base_ in first_arg.node.mro:
+                if base_.fullname == "enum.Enum":
+                    return Instance(first_arg.node, [])
+            # TODO: support other pep-435 types here
+        else:
+            return api.named_type(names.NAMED_TYPE_BUILTINS_STR, [])
+
+    assert node.has_base("sqlalchemy.sql.type_api.TypeEngine"), (
+        "could not extract Python type from node: %s" % node
+    )
+
+    type_engine_sym = api.lookup_fully_qualified_or_none(
+        "sqlalchemy.sql.type_api.TypeEngine"
+    )
+
+    assert type_engine_sym is not None and isinstance(
+        type_engine_sym.node, TypeInfo
+    )
+    type_engine = map_instance_to_supertype(
+        Instance(node, []),
+        type_engine_sym.node,
+    )
+    return get_proper_type(type_engine.args[-1])
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/names.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/names.py
new file mode 100644
index 00000000..31978628
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/names.py
@@ -0,0 +1,335 @@
+# ext/mypy/names.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Union
+
+from mypy.nodes import ARG_POS
+from mypy.nodes import CallExpr
+from mypy.nodes import ClassDef
+from mypy.nodes import Decorator
+from mypy.nodes import Expression
+from mypy.nodes import FuncDef
+from mypy.nodes import MemberExpr
+from mypy.nodes import NameExpr
+from mypy.nodes import OverloadedFuncDef
+from mypy.nodes import SymbolNode
+from mypy.nodes import TypeAlias
+from mypy.nodes import TypeInfo
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.types import CallableType
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import UnboundType
+
+from ... import util
+
+COLUMN: int = util.symbol("COLUMN")
+RELATIONSHIP: int = util.symbol("RELATIONSHIP")
+REGISTRY: int = util.symbol("REGISTRY")
+COLUMN_PROPERTY: int = util.symbol("COLUMN_PROPERTY")
+TYPEENGINE: int = util.symbol("TYPEENGNE")
+MAPPED: int = util.symbol("MAPPED")
+DECLARATIVE_BASE: int = util.symbol("DECLARATIVE_BASE")
+DECLARATIVE_META: int = util.symbol("DECLARATIVE_META")
+MAPPED_DECORATOR: int = util.symbol("MAPPED_DECORATOR")
+SYNONYM_PROPERTY: int = util.symbol("SYNONYM_PROPERTY")
+COMPOSITE_PROPERTY: int = util.symbol("COMPOSITE_PROPERTY")
+DECLARED_ATTR: int = util.symbol("DECLARED_ATTR")
+MAPPER_PROPERTY: int = util.symbol("MAPPER_PROPERTY")
+AS_DECLARATIVE: int = util.symbol("AS_DECLARATIVE")
+AS_DECLARATIVE_BASE: int = util.symbol("AS_DECLARATIVE_BASE")
+DECLARATIVE_MIXIN: int = util.symbol("DECLARATIVE_MIXIN")
+QUERY_EXPRESSION: int = util.symbol("QUERY_EXPRESSION")
+
+# names that must succeed with mypy.api.named_type
+NAMED_TYPE_BUILTINS_OBJECT = "builtins.object"
+NAMED_TYPE_BUILTINS_STR = "builtins.str"
+NAMED_TYPE_BUILTINS_LIST = "builtins.list"
+NAMED_TYPE_SQLA_MAPPED = "sqlalchemy.orm.base.Mapped"
+
+_RelFullNames = {
+    "sqlalchemy.orm.relationships.Relationship",
+    "sqlalchemy.orm.relationships.RelationshipProperty",
+    "sqlalchemy.orm.relationships._RelationshipDeclared",
+    "sqlalchemy.orm.Relationship",
+    "sqlalchemy.orm.RelationshipProperty",
+}
+
+_lookup: Dict[str, Tuple[int, Set[str]]] = {
+    "Column": (
+        COLUMN,
+        {
+            "sqlalchemy.sql.schema.Column",
+            "sqlalchemy.sql.Column",
+        },
+    ),
+    "Relationship": (RELATIONSHIP, _RelFullNames),
+    "RelationshipProperty": (RELATIONSHIP, _RelFullNames),
+    "_RelationshipDeclared": (RELATIONSHIP, _RelFullNames),
+    "registry": (
+        REGISTRY,
+        {
+            "sqlalchemy.orm.decl_api.registry",
+            "sqlalchemy.orm.registry",
+        },
+    ),
+    "ColumnProperty": (
+        COLUMN_PROPERTY,
+        {
+            "sqlalchemy.orm.properties.MappedSQLExpression",
+            "sqlalchemy.orm.MappedSQLExpression",
+            "sqlalchemy.orm.properties.ColumnProperty",
+            "sqlalchemy.orm.ColumnProperty",
+        },
+    ),
+    "MappedSQLExpression": (
+        COLUMN_PROPERTY,
+        {
+            "sqlalchemy.orm.properties.MappedSQLExpression",
+            "sqlalchemy.orm.MappedSQLExpression",
+            "sqlalchemy.orm.properties.ColumnProperty",
+            "sqlalchemy.orm.ColumnProperty",
+        },
+    ),
+    "Synonym": (
+        SYNONYM_PROPERTY,
+        {
+            "sqlalchemy.orm.descriptor_props.Synonym",
+            "sqlalchemy.orm.Synonym",
+            "sqlalchemy.orm.descriptor_props.SynonymProperty",
+            "sqlalchemy.orm.SynonymProperty",
+        },
+    ),
+    "SynonymProperty": (
+        SYNONYM_PROPERTY,
+        {
+            "sqlalchemy.orm.descriptor_props.Synonym",
+            "sqlalchemy.orm.Synonym",
+            "sqlalchemy.orm.descriptor_props.SynonymProperty",
+            "sqlalchemy.orm.SynonymProperty",
+        },
+    ),
+    "Composite": (
+        COMPOSITE_PROPERTY,
+        {
+            "sqlalchemy.orm.descriptor_props.Composite",
+            "sqlalchemy.orm.Composite",
+            "sqlalchemy.orm.descriptor_props.CompositeProperty",
+            "sqlalchemy.orm.CompositeProperty",
+        },
+    ),
+    "CompositeProperty": (
+        COMPOSITE_PROPERTY,
+        {
+            "sqlalchemy.orm.descriptor_props.Composite",
+            "sqlalchemy.orm.Composite",
+            "sqlalchemy.orm.descriptor_props.CompositeProperty",
+            "sqlalchemy.orm.CompositeProperty",
+        },
+    ),
+    "MapperProperty": (
+        MAPPER_PROPERTY,
+        {
+            "sqlalchemy.orm.interfaces.MapperProperty",
+            "sqlalchemy.orm.MapperProperty",
+        },
+    ),
+    "TypeEngine": (TYPEENGINE, {"sqlalchemy.sql.type_api.TypeEngine"}),
+    "Mapped": (MAPPED, {NAMED_TYPE_SQLA_MAPPED}),
+    "declarative_base": (
+        DECLARATIVE_BASE,
+        {
+            "sqlalchemy.ext.declarative.declarative_base",
+            "sqlalchemy.orm.declarative_base",
+            "sqlalchemy.orm.decl_api.declarative_base",
+        },
+    ),
+    "DeclarativeMeta": (
+        DECLARATIVE_META,
+        {
+            "sqlalchemy.ext.declarative.DeclarativeMeta",
+            "sqlalchemy.orm.DeclarativeMeta",
+            "sqlalchemy.orm.decl_api.DeclarativeMeta",
+        },
+    ),
+    "mapped": (
+        MAPPED_DECORATOR,
+        {
+            "sqlalchemy.orm.decl_api.registry.mapped",
+            "sqlalchemy.orm.registry.mapped",
+        },
+    ),
+    "as_declarative": (
+        AS_DECLARATIVE,
+        {
+            "sqlalchemy.ext.declarative.as_declarative",
+            "sqlalchemy.orm.decl_api.as_declarative",
+            "sqlalchemy.orm.as_declarative",
+        },
+    ),
+    "as_declarative_base": (
+        AS_DECLARATIVE_BASE,
+        {
+            "sqlalchemy.orm.decl_api.registry.as_declarative_base",
+            "sqlalchemy.orm.registry.as_declarative_base",
+        },
+    ),
+    "declared_attr": (
+        DECLARED_ATTR,
+        {
+            "sqlalchemy.orm.decl_api.declared_attr",
+            "sqlalchemy.orm.declared_attr",
+        },
+    ),
+    "declarative_mixin": (
+        DECLARATIVE_MIXIN,
+        {
+            "sqlalchemy.orm.decl_api.declarative_mixin",
+            "sqlalchemy.orm.declarative_mixin",
+        },
+    ),
+    "query_expression": (
+        QUERY_EXPRESSION,
+        {
+            "sqlalchemy.orm.query_expression",
+            "sqlalchemy.orm._orm_constructors.query_expression",
+        },
+    ),
+}
+
+
+def has_base_type_id(info: TypeInfo, type_id: int) -> bool:
+    for mr in info.mro:
+        check_type_id, fullnames = _lookup.get(mr.name, (None, None))
+        if check_type_id == type_id:
+            break
+    else:
+        return False
+
+    if fullnames is None:
+        return False
+
+    return mr.fullname in fullnames
+
+
+def mro_has_id(mro: List[TypeInfo], type_id: int) -> bool:
+    for mr in mro:
+        check_type_id, fullnames = _lookup.get(mr.name, (None, None))
+        if check_type_id == type_id:
+            break
+    else:
+        return False
+
+    if fullnames is None:
+        return False
+
+    return mr.fullname in fullnames
+
+
+def type_id_for_unbound_type(
+    type_: UnboundType, cls: ClassDef, api: SemanticAnalyzerPluginInterface
+) -> Optional[int]:
+    sym = api.lookup_qualified(type_.name, type_)
+    if sym is not None:
+        if isinstance(sym.node, TypeAlias):
+            target_type = get_proper_type(sym.node.target)
+            if isinstance(target_type, Instance):
+                return type_id_for_named_node(target_type.type)
+        elif isinstance(sym.node, TypeInfo):
+            return type_id_for_named_node(sym.node)
+
+    return None
+
+
+def type_id_for_callee(callee: Expression) -> Optional[int]:
+    if isinstance(callee, (MemberExpr, NameExpr)):
+        if isinstance(callee.node, Decorator) and isinstance(
+            callee.node.func, FuncDef
+        ):
+            if callee.node.func.type and isinstance(
+                callee.node.func.type, CallableType
+            ):
+                ret_type = get_proper_type(callee.node.func.type.ret_type)
+
+                if isinstance(ret_type, Instance):
+                    return type_id_for_fullname(ret_type.type.fullname)
+
+            return None
+
+        elif isinstance(callee.node, OverloadedFuncDef):
+            if (
+                callee.node.impl
+                and callee.node.impl.type
+                and isinstance(callee.node.impl.type, CallableType)
+            ):
+                ret_type = get_proper_type(callee.node.impl.type.ret_type)
+
+                if isinstance(ret_type, Instance):
+                    return type_id_for_fullname(ret_type.type.fullname)
+
+            return None
+        elif isinstance(callee.node, FuncDef):
+            if callee.node.type and isinstance(callee.node.type, CallableType):
+                ret_type = get_proper_type(callee.node.type.ret_type)
+
+                if isinstance(ret_type, Instance):
+                    return type_id_for_fullname(ret_type.type.fullname)
+
+            return None
+        elif isinstance(callee.node, TypeAlias):
+            target_type = get_proper_type(callee.node.target)
+            if isinstance(target_type, Instance):
+                return type_id_for_fullname(target_type.type.fullname)
+        elif isinstance(callee.node, TypeInfo):
+            return type_id_for_named_node(callee)
+    return None
+
+
+def type_id_for_named_node(
+    node: Union[NameExpr, MemberExpr, SymbolNode]
+) -> Optional[int]:
+    type_id, fullnames = _lookup.get(node.name, (None, None))
+
+    if type_id is None or fullnames is None:
+        return None
+    elif node.fullname in fullnames:
+        return type_id
+    else:
+        return None
+
+
+def type_id_for_fullname(fullname: str) -> Optional[int]:
+    tokens = fullname.split(".")
+    immediate = tokens[-1]
+
+    type_id, fullnames = _lookup.get(immediate, (None, None))
+
+    if type_id is None or fullnames is None:
+        return None
+    elif fullname in fullnames:
+        return type_id
+    else:
+        return None
+
+
+def expr_to_mapped_constructor(expr: Expression) -> CallExpr:
+    column_descriptor = NameExpr("__sa_Mapped")
+    column_descriptor.fullname = NAMED_TYPE_SQLA_MAPPED
+    member_expr = MemberExpr(column_descriptor, "_empty_constructor")
+    return CallExpr(
+        member_expr,
+        [expr],
+        [ARG_POS],
+        ["arg1"],
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/plugin.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/plugin.py
new file mode 100644
index 00000000..1ec2c02b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/plugin.py
@@ -0,0 +1,303 @@
+# ext/mypy/plugin.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""
+Mypy plugin for SQLAlchemy ORM.
+
+"""
+from __future__ import annotations
+
+from typing import Callable
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Type as TypingType
+from typing import Union
+
+from mypy import nodes
+from mypy.mro import calculate_mro
+from mypy.mro import MroError
+from mypy.nodes import Block
+from mypy.nodes import ClassDef
+from mypy.nodes import GDEF
+from mypy.nodes import MypyFile
+from mypy.nodes import NameExpr
+from mypy.nodes import SymbolTable
+from mypy.nodes import SymbolTableNode
+from mypy.nodes import TypeInfo
+from mypy.plugin import AttributeContext
+from mypy.plugin import ClassDefContext
+from mypy.plugin import DynamicClassDefContext
+from mypy.plugin import Plugin
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import Type
+
+from . import decl_class
+from . import names
+from . import util
+
+try:
+    __import__("sqlalchemy-stubs")
+except ImportError:
+    pass
+else:
+    raise ImportError(
+        "The SQLAlchemy mypy plugin in SQLAlchemy "
+        "2.0 does not work with sqlalchemy-stubs or "
+        "sqlalchemy2-stubs installed, as well as with any other third party "
+        "SQLAlchemy stubs.  Please uninstall all SQLAlchemy stubs "
+        "packages."
+    )
+
+
+class SQLAlchemyPlugin(Plugin):
+    def get_dynamic_class_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[DynamicClassDefContext], None]]:
+        if names.type_id_for_fullname(fullname) is names.DECLARATIVE_BASE:
+            return _dynamic_class_hook
+        return None
+
+    def get_customize_class_mro_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[ClassDefContext], None]]:
+        return _fill_in_decorators
+
+    def get_class_decorator_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[ClassDefContext], None]]:
+        sym = self.lookup_fully_qualified(fullname)
+
+        if sym is not None and sym.node is not None:
+            type_id = names.type_id_for_named_node(sym.node)
+            if type_id is names.MAPPED_DECORATOR:
+                return _cls_decorator_hook
+            elif type_id in (
+                names.AS_DECLARATIVE,
+                names.AS_DECLARATIVE_BASE,
+            ):
+                return _base_cls_decorator_hook
+            elif type_id is names.DECLARATIVE_MIXIN:
+                return _declarative_mixin_hook
+
+        return None
+
+    def get_metaclass_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[ClassDefContext], None]]:
+        if names.type_id_for_fullname(fullname) is names.DECLARATIVE_META:
+            # Set any classes that explicitly have metaclass=DeclarativeMeta
+            # as declarative so the check in `get_base_class_hook()` works
+            return _metaclass_cls_hook
+
+        return None
+
+    def get_base_class_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[ClassDefContext], None]]:
+        sym = self.lookup_fully_qualified(fullname)
+
+        if (
+            sym
+            and isinstance(sym.node, TypeInfo)
+            and util.has_declarative_base(sym.node)
+        ):
+            return _base_cls_hook
+
+        return None
+
+    def get_attribute_hook(
+        self, fullname: str
+    ) -> Optional[Callable[[AttributeContext], Type]]:
+        if fullname.startswith(
+            "sqlalchemy.orm.attributes.QueryableAttribute."
+        ):
+            return _queryable_getattr_hook
+
+        return None
+
+    def get_additional_deps(
+        self, file: MypyFile
+    ) -> List[Tuple[int, str, int]]:
+        return [
+            #
+            (10, "sqlalchemy.orm", -1),
+            (10, "sqlalchemy.orm.attributes", -1),
+            (10, "sqlalchemy.orm.decl_api", -1),
+        ]
+
+
+def plugin(version: str) -> TypingType[SQLAlchemyPlugin]:
+    return SQLAlchemyPlugin
+
+
+def _dynamic_class_hook(ctx: DynamicClassDefContext) -> None:
+    """Generate a declarative Base class when the declarative_base() function
+    is encountered."""
+
+    _add_globals(ctx)
+
+    cls = ClassDef(ctx.name, Block([]))
+    cls.fullname = ctx.api.qualified_name(ctx.name)
+
+    info = TypeInfo(SymbolTable(), cls, ctx.api.cur_mod_id)
+    cls.info = info
+    _set_declarative_metaclass(ctx.api, cls)
+
+    cls_arg = util.get_callexpr_kwarg(ctx.call, "cls", expr_types=(NameExpr,))
+    if cls_arg is not None and isinstance(cls_arg.node, TypeInfo):
+        util.set_is_base(cls_arg.node)
+        decl_class.scan_declarative_assignments_and_apply_types(
+            cls_arg.node.defn, ctx.api, is_mixin_scan=True
+        )
+        info.bases = [Instance(cls_arg.node, [])]
+    else:
+        obj = ctx.api.named_type(names.NAMED_TYPE_BUILTINS_OBJECT)
+
+        info.bases = [obj]
+
+    try:
+        calculate_mro(info)
+    except MroError:
+        util.fail(
+            ctx.api, "Not able to calculate MRO for declarative base", ctx.call
+        )
+        obj = ctx.api.named_type(names.NAMED_TYPE_BUILTINS_OBJECT)
+        info.bases = [obj]
+        info.fallback_to_any = True
+
+    ctx.api.add_symbol_table_node(ctx.name, SymbolTableNode(GDEF, info))
+    util.set_is_base(info)
+
+
+def _fill_in_decorators(ctx: ClassDefContext) -> None:
+    for decorator in ctx.cls.decorators:
+        # set the ".fullname" attribute of a class decorator
+        # that is a MemberExpr.   This causes the logic in
+        # semanal.py->apply_class_plugin_hooks to invoke the
+        # get_class_decorator_hook for our "registry.map_class()"
+        # and "registry.as_declarative_base()" methods.
+        # this seems like a bug in mypy that these decorators are otherwise
+        # skipped.
+
+        if (
+            isinstance(decorator, nodes.CallExpr)
+            and isinstance(decorator.callee, nodes.MemberExpr)
+            and decorator.callee.name == "as_declarative_base"
+        ):
+            target = decorator.callee
+        elif (
+            isinstance(decorator, nodes.MemberExpr)
+            and decorator.name == "mapped"
+        ):
+            target = decorator
+        else:
+            continue
+
+        if isinstance(target.expr, NameExpr):
+            sym = ctx.api.lookup_qualified(
+                target.expr.name, target, suppress_errors=True
+            )
+        else:
+            continue
+
+        if sym and sym.node:
+            sym_type = get_proper_type(sym.type)
+            if isinstance(sym_type, Instance):
+                target.fullname = f"{sym_type.type.fullname}.{target.name}"
+            else:
+                # if the registry is in the same file as where the
+                # decorator is used, it might not have semantic
+                # symbols applied and we can't get a fully qualified
+                # name or an inferred type, so we are actually going to
+                # flag an error in this case that they need to annotate
+                # it.  The "registry" is declared just
+                # once (or few times), so they have to just not use
+                # type inference for its assignment in this one case.
+                util.fail(
+                    ctx.api,
+                    "Class decorator called %s(), but we can't "
+                    "tell if it's from an ORM registry.  Please "
+                    "annotate the registry assignment, e.g. "
+                    "my_registry: registry = registry()" % target.name,
+                    sym.node,
+                )
+
+
+def _cls_decorator_hook(ctx: ClassDefContext) -> None:
+    _add_globals(ctx)
+    assert isinstance(ctx.reason, nodes.MemberExpr)
+    expr = ctx.reason.expr
+
+    assert isinstance(expr, nodes.RefExpr) and isinstance(expr.node, nodes.Var)
+
+    node_type = get_proper_type(expr.node.type)
+
+    assert (
+        isinstance(node_type, Instance)
+        and names.type_id_for_named_node(node_type.type) is names.REGISTRY
+    )
+
+    decl_class.scan_declarative_assignments_and_apply_types(ctx.cls, ctx.api)
+
+
+def _base_cls_decorator_hook(ctx: ClassDefContext) -> None:
+    _add_globals(ctx)
+
+    cls = ctx.cls
+
+    _set_declarative_metaclass(ctx.api, cls)
+
+    util.set_is_base(ctx.cls.info)
+    decl_class.scan_declarative_assignments_and_apply_types(
+        cls, ctx.api, is_mixin_scan=True
+    )
+
+
+def _declarative_mixin_hook(ctx: ClassDefContext) -> None:
+    _add_globals(ctx)
+    util.set_is_base(ctx.cls.info)
+    decl_class.scan_declarative_assignments_and_apply_types(
+        ctx.cls, ctx.api, is_mixin_scan=True
+    )
+
+
+def _metaclass_cls_hook(ctx: ClassDefContext) -> None:
+    util.set_is_base(ctx.cls.info)
+
+
+def _base_cls_hook(ctx: ClassDefContext) -> None:
+    _add_globals(ctx)
+    decl_class.scan_declarative_assignments_and_apply_types(ctx.cls, ctx.api)
+
+
+def _queryable_getattr_hook(ctx: AttributeContext) -> Type:
+    # how do I....tell it it has no attribute of a certain name?
+    # can't find any Type that seems to match that
+    return ctx.default_attr_type
+
+
+def _add_globals(ctx: Union[ClassDefContext, DynamicClassDefContext]) -> None:
+    """Add __sa_DeclarativeMeta and __sa_Mapped symbol to the global space
+    for all class defs
+
+    """
+
+    util.add_global(ctx, "sqlalchemy.orm", "Mapped", "__sa_Mapped")
+
+
+def _set_declarative_metaclass(
+    api: SemanticAnalyzerPluginInterface, target_cls: ClassDef
+) -> None:
+    info = target_cls.info
+    sym = api.lookup_fully_qualified_or_none(
+        "sqlalchemy.orm.decl_api.DeclarativeMeta"
+    )
+    assert sym is not None and isinstance(sym.node, TypeInfo)
+    info.declared_metaclass = info.metaclass_type = Instance(sym.node, [])
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/util.py
new file mode 100644
index 00000000..16761b9a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/mypy/util.py
@@ -0,0 +1,357 @@
+# ext/mypy/util.py
+# Copyright (C) 2021-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import re
+from typing import Any
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type as TypingType
+from typing import TypeVar
+from typing import Union
+
+from mypy import version
+from mypy.messages import format_type as _mypy_format_type
+from mypy.nodes import CallExpr
+from mypy.nodes import ClassDef
+from mypy.nodes import CLASSDEF_NO_INFO
+from mypy.nodes import Context
+from mypy.nodes import Expression
+from mypy.nodes import FuncDef
+from mypy.nodes import IfStmt
+from mypy.nodes import JsonDict
+from mypy.nodes import MemberExpr
+from mypy.nodes import NameExpr
+from mypy.nodes import Statement
+from mypy.nodes import SymbolTableNode
+from mypy.nodes import TypeAlias
+from mypy.nodes import TypeInfo
+from mypy.options import Options
+from mypy.plugin import ClassDefContext
+from mypy.plugin import DynamicClassDefContext
+from mypy.plugin import SemanticAnalyzerPluginInterface
+from mypy.plugins.common import deserialize_and_fixup_type
+from mypy.typeops import map_type_from_supertype
+from mypy.types import CallableType
+from mypy.types import get_proper_type
+from mypy.types import Instance
+from mypy.types import NoneType
+from mypy.types import Type
+from mypy.types import TypeVarType
+from mypy.types import UnboundType
+from mypy.types import UnionType
+
+_vers = tuple(
+    [int(x) for x in version.__version__.split(".") if re.match(r"^\d+$", x)]
+)
+mypy_14 = _vers >= (1, 4)
+
+
+_TArgType = TypeVar("_TArgType", bound=Union[CallExpr, NameExpr])
+
+
+class SQLAlchemyAttribute:
+    def __init__(
+        self,
+        name: str,
+        line: int,
+        column: int,
+        typ: Optional[Type],
+        info: TypeInfo,
+    ) -> None:
+        self.name = name
+        self.line = line
+        self.column = column
+        self.type = typ
+        self.info = info
+
+    def serialize(self) -> JsonDict:
+        assert self.type
+        return {
+            "name": self.name,
+            "line": self.line,
+            "column": self.column,
+            "type": serialize_type(self.type),
+        }
+
+    def expand_typevar_from_subtype(self, sub_type: TypeInfo) -> None:
+        """Expands type vars in the context of a subtype when an attribute is
+        inherited from a generic super type.
+        """
+        if not isinstance(self.type, TypeVarType):
+            return
+
+        self.type = map_type_from_supertype(self.type, sub_type, self.info)
+
+    @classmethod
+    def deserialize(
+        cls,
+        info: TypeInfo,
+        data: JsonDict,
+        api: SemanticAnalyzerPluginInterface,
+    ) -> SQLAlchemyAttribute:
+        data = data.copy()
+        typ = deserialize_and_fixup_type(data.pop("type"), api)
+        return cls(typ=typ, info=info, **data)
+
+
+def name_is_dunder(name: str) -> bool:
+    return bool(re.match(r"^__.+?__$", name))
+
+
+def _set_info_metadata(info: TypeInfo, key: str, data: Any) -> None:
+    info.metadata.setdefault("sqlalchemy", {})[key] = data
+
+
+def _get_info_metadata(info: TypeInfo, key: str) -> Optional[Any]:
+    return info.metadata.get("sqlalchemy", {}).get(key, None)
+
+
+def _get_info_mro_metadata(info: TypeInfo, key: str) -> Optional[Any]:
+    if info.mro:
+        for base in info.mro:
+            metadata = _get_info_metadata(base, key)
+            if metadata is not None:
+                return metadata
+    return None
+
+
+def establish_as_sqlalchemy(info: TypeInfo) -> None:
+    info.metadata.setdefault("sqlalchemy", {})
+
+
+def set_is_base(info: TypeInfo) -> None:
+    _set_info_metadata(info, "is_base", True)
+
+
+def get_is_base(info: TypeInfo) -> bool:
+    is_base = _get_info_metadata(info, "is_base")
+    return is_base is True
+
+
+def has_declarative_base(info: TypeInfo) -> bool:
+    is_base = _get_info_mro_metadata(info, "is_base")
+    return is_base is True
+
+
+def set_has_table(info: TypeInfo) -> None:
+    _set_info_metadata(info, "has_table", True)
+
+
+def get_has_table(info: TypeInfo) -> bool:
+    is_base = _get_info_metadata(info, "has_table")
+    return is_base is True
+
+
+def get_mapped_attributes(
+    info: TypeInfo, api: SemanticAnalyzerPluginInterface
+) -> Optional[List[SQLAlchemyAttribute]]:
+    mapped_attributes: Optional[List[JsonDict]] = _get_info_metadata(
+        info, "mapped_attributes"
+    )
+    if mapped_attributes is None:
+        return None
+
+    attributes: List[SQLAlchemyAttribute] = []
+
+    for data in mapped_attributes:
+        attr = SQLAlchemyAttribute.deserialize(info, data, api)
+        attr.expand_typevar_from_subtype(info)
+        attributes.append(attr)
+
+    return attributes
+
+
+def format_type(typ_: Type, options: Options) -> str:
+    if mypy_14:
+        return _mypy_format_type(typ_, options)
+    else:
+        return _mypy_format_type(typ_)  # type: ignore
+
+
+def set_mapped_attributes(
+    info: TypeInfo, attributes: List[SQLAlchemyAttribute]
+) -> None:
+    _set_info_metadata(
+        info,
+        "mapped_attributes",
+        [attribute.serialize() for attribute in attributes],
+    )
+
+
+def fail(api: SemanticAnalyzerPluginInterface, msg: str, ctx: Context) -> None:
+    msg = "[SQLAlchemy Mypy plugin] %s" % msg
+    return api.fail(msg, ctx)
+
+
+def add_global(
+    ctx: Union[ClassDefContext, DynamicClassDefContext],
+    module: str,
+    symbol_name: str,
+    asname: str,
+) -> None:
+    module_globals = ctx.api.modules[ctx.api.cur_mod_id].names
+
+    if asname not in module_globals:
+        lookup_sym: SymbolTableNode = ctx.api.modules[module].names[
+            symbol_name
+        ]
+
+        module_globals[asname] = lookup_sym
+
+
+@overload
+def get_callexpr_kwarg(
+    callexpr: CallExpr, name: str, *, expr_types: None = ...
+) -> Optional[Union[CallExpr, NameExpr]]: ...
+
+
+@overload
+def get_callexpr_kwarg(
+    callexpr: CallExpr,
+    name: str,
+    *,
+    expr_types: Tuple[TypingType[_TArgType], ...],
+) -> Optional[_TArgType]: ...
+
+
+def get_callexpr_kwarg(
+    callexpr: CallExpr,
+    name: str,
+    *,
+    expr_types: Optional[Tuple[TypingType[Any], ...]] = None,
+) -> Optional[Any]:
+    try:
+        arg_idx = callexpr.arg_names.index(name)
+    except ValueError:
+        return None
+
+    kwarg = callexpr.args[arg_idx]
+    if isinstance(
+        kwarg, expr_types if expr_types is not None else (NameExpr, CallExpr)
+    ):
+        return kwarg
+
+    return None
+
+
+def flatten_typechecking(stmts: Iterable[Statement]) -> Iterator[Statement]:
+    for stmt in stmts:
+        if (
+            isinstance(stmt, IfStmt)
+            and isinstance(stmt.expr[0], NameExpr)
+            and stmt.expr[0].fullname == "typing.TYPE_CHECKING"
+        ):
+            yield from stmt.body[0].body
+        else:
+            yield stmt
+
+
+def type_for_callee(callee: Expression) -> Optional[Union[Instance, TypeInfo]]:
+    if isinstance(callee, (MemberExpr, NameExpr)):
+        if isinstance(callee.node, FuncDef):
+            if callee.node.type and isinstance(callee.node.type, CallableType):
+                ret_type = get_proper_type(callee.node.type.ret_type)
+
+                if isinstance(ret_type, Instance):
+                    return ret_type
+
+            return None
+        elif isinstance(callee.node, TypeAlias):
+            target_type = get_proper_type(callee.node.target)
+            if isinstance(target_type, Instance):
+                return target_type
+        elif isinstance(callee.node, TypeInfo):
+            return callee.node
+    return None
+
+
+def unbound_to_instance(
+    api: SemanticAnalyzerPluginInterface, typ: Type
+) -> Type:
+    """Take the UnboundType that we seem to get as the ret_type from a FuncDef
+    and convert it into an Instance/TypeInfo kind of structure that seems
+    to work as the left-hand type of an AssignmentStatement.
+
+    """
+
+    if not isinstance(typ, UnboundType):
+        return typ
+
+    # TODO: figure out a more robust way to check this.  The node is some
+    # kind of _SpecialForm, there's a typing.Optional that's _SpecialForm,
+    # but I can't figure out how to get them to match up
+    if typ.name == "Optional":
+        # convert from "Optional?" to the more familiar
+        # UnionType[..., NoneType()]
+        return unbound_to_instance(
+            api,
+            UnionType(
+                [unbound_to_instance(api, typ_arg) for typ_arg in typ.args]
+                + [NoneType()]
+            ),
+        )
+
+    node = api.lookup_qualified(typ.name, typ)
+
+    if (
+        node is not None
+        and isinstance(node, SymbolTableNode)
+        and isinstance(node.node, TypeInfo)
+    ):
+        bound_type = node.node
+
+        return Instance(
+            bound_type,
+            [
+                (
+                    unbound_to_instance(api, arg)
+                    if isinstance(arg, UnboundType)
+                    else arg
+                )
+                for arg in typ.args
+            ],
+        )
+    else:
+        return typ
+
+
+def info_for_cls(
+    cls: ClassDef, api: SemanticAnalyzerPluginInterface
+) -> Optional[TypeInfo]:
+    if cls.info is CLASSDEF_NO_INFO:
+        sym = api.lookup_qualified(cls.name, cls)
+        if sym is None:
+            return None
+        assert sym and isinstance(sym.node, TypeInfo)
+        return sym.node
+
+    return cls.info
+
+
+def serialize_type(typ: Type) -> Union[str, JsonDict]:
+    try:
+        return typ.serialize()
+    except Exception:
+        pass
+    if hasattr(typ, "args"):
+        typ.args = tuple(
+            (
+                a.resolve_string_annotation()
+                if hasattr(a, "resolve_string_annotation")
+                else a
+            )
+            for a in typ.args
+        )
+    elif hasattr(typ, "resolve_string_annotation"):
+        typ = typ.resolve_string_annotation()
+    return typ.serialize()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/orderinglist.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/orderinglist.py
new file mode 100644
index 00000000..3cc67b18
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/orderinglist.py
@@ -0,0 +1,427 @@
+# ext/orderinglist.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""A custom list that manages index/position information for contained
+elements.
+
+:author: Jason Kirtland
+
+``orderinglist`` is a helper for mutable ordered relationships.  It will
+intercept list operations performed on a :func:`_orm.relationship`-managed
+collection and
+automatically synchronize changes in list position onto a target scalar
+attribute.
+
+Example: A ``slide`` table, where each row refers to zero or more entries
+in a related ``bullet`` table.   The bullets within a slide are
+displayed in order based on the value of the ``position`` column in the
+``bullet`` table.   As entries are reordered in memory, the value of the
+``position`` attribute should be updated to reflect the new sort order::
+
+
+    Base = declarative_base()
+
+
+    class Slide(Base):
+        __tablename__ = "slide"
+
+        id = Column(Integer, primary_key=True)
+        name = Column(String)
+
+        bullets = relationship("Bullet", order_by="Bullet.position")
+
+
+    class Bullet(Base):
+        __tablename__ = "bullet"
+        id = Column(Integer, primary_key=True)
+        slide_id = Column(Integer, ForeignKey("slide.id"))
+        position = Column(Integer)
+        text = Column(String)
+
+The standard relationship mapping will produce a list-like attribute on each
+``Slide`` containing all related ``Bullet`` objects,
+but coping with changes in ordering is not handled automatically.
+When appending a ``Bullet`` into ``Slide.bullets``, the ``Bullet.position``
+attribute will remain unset until manually assigned.   When the ``Bullet``
+is inserted into the middle of the list, the following ``Bullet`` objects
+will also need to be renumbered.
+
+The :class:`.OrderingList` object automates this task, managing the
+``position`` attribute on all ``Bullet`` objects in the collection.  It is
+constructed using the :func:`.ordering_list` factory::
+
+    from sqlalchemy.ext.orderinglist import ordering_list
+
+    Base = declarative_base()
+
+
+    class Slide(Base):
+        __tablename__ = "slide"
+
+        id = Column(Integer, primary_key=True)
+        name = Column(String)
+
+        bullets = relationship(
+            "Bullet",
+            order_by="Bullet.position",
+            collection_class=ordering_list("position"),
+        )
+
+
+    class Bullet(Base):
+        __tablename__ = "bullet"
+        id = Column(Integer, primary_key=True)
+        slide_id = Column(Integer, ForeignKey("slide.id"))
+        position = Column(Integer)
+        text = Column(String)
+
+With the above mapping the ``Bullet.position`` attribute is managed::
+
+    s = Slide()
+    s.bullets.append(Bullet())
+    s.bullets.append(Bullet())
+    s.bullets[1].position
+    >>> 1
+    s.bullets.insert(1, Bullet())
+    s.bullets[2].position
+    >>> 2
+
+The :class:`.OrderingList` construct only works with **changes** to a
+collection, and not the initial load from the database, and requires that the
+list be sorted when loaded.  Therefore, be sure to specify ``order_by`` on the
+:func:`_orm.relationship` against the target ordering attribute, so that the
+ordering is correct when first loaded.
+
+.. warning::
+
+  :class:`.OrderingList` only provides limited functionality when a primary
+  key column or unique column is the target of the sort.  Operations
+  that are unsupported or are problematic include:
+
+    * two entries must trade values.  This is not supported directly in the
+      case of a primary key or unique constraint because it means at least
+      one row would need to be temporarily removed first, or changed to
+      a third, neutral value while the switch occurs.
+
+    * an entry must be deleted in order to make room for a new entry.
+      SQLAlchemy's unit of work performs all INSERTs before DELETEs within a
+      single flush.  In the case of a primary key, it will trade
+      an INSERT/DELETE of the same primary key for an UPDATE statement in order
+      to lessen the impact of this limitation, however this does not take place
+      for a UNIQUE column.
+      A future feature will allow the "DELETE before INSERT" behavior to be
+      possible, alleviating this limitation, though this feature will require
+      explicit configuration at the mapper level for sets of columns that
+      are to be handled in this way.
+
+:func:`.ordering_list` takes the name of the related object's ordering
+attribute as an argument.  By default, the zero-based integer index of the
+object's position in the :func:`.ordering_list` is synchronized with the
+ordering attribute: index 0 will get position 0, index 1 position 1, etc.  To
+start numbering at 1 or some other integer, provide ``count_from=1``.
+
+
+"""
+from __future__ import annotations
+
+from typing import Callable
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import TypeVar
+
+from ..orm.collections import collection
+from ..orm.collections import collection_adapter
+
+_T = TypeVar("_T")
+OrderingFunc = Callable[[int, Sequence[_T]], int]
+
+
+__all__ = ["ordering_list"]
+
+
+def ordering_list(
+    attr: str,
+    count_from: Optional[int] = None,
+    ordering_func: Optional[OrderingFunc] = None,
+    reorder_on_append: bool = False,
+) -> Callable[[], OrderingList]:
+    """Prepares an :class:`OrderingList` factory for use in mapper definitions.
+
+    Returns an object suitable for use as an argument to a Mapper
+    relationship's ``collection_class`` option.  e.g.::
+
+        from sqlalchemy.ext.orderinglist import ordering_list
+
+
+        class Slide(Base):
+            __tablename__ = "slide"
+
+            id = Column(Integer, primary_key=True)
+            name = Column(String)
+
+            bullets = relationship(
+                "Bullet",
+                order_by="Bullet.position",
+                collection_class=ordering_list("position"),
+            )
+
+    :param attr:
+      Name of the mapped attribute to use for storage and retrieval of
+      ordering information
+
+    :param count_from:
+      Set up an integer-based ordering, starting at ``count_from``.  For
+      example, ``ordering_list('pos', count_from=1)`` would create a 1-based
+      list in SQL, storing the value in the 'pos' column.  Ignored if
+      ``ordering_func`` is supplied.
+
+    Additional arguments are passed to the :class:`.OrderingList` constructor.
+
+    """
+
+    kw = _unsugar_count_from(
+        count_from=count_from,
+        ordering_func=ordering_func,
+        reorder_on_append=reorder_on_append,
+    )
+    return lambda: OrderingList(attr, **kw)
+
+
+# Ordering utility functions
+
+
+def count_from_0(index, collection):
+    """Numbering function: consecutive integers starting at 0."""
+
+    return index
+
+
+def count_from_1(index, collection):
+    """Numbering function: consecutive integers starting at 1."""
+
+    return index + 1
+
+
+def count_from_n_factory(start):
+    """Numbering function: consecutive integers starting at arbitrary start."""
+
+    def f(index, collection):
+        return index + start
+
+    try:
+        f.__name__ = "count_from_%i" % start
+    except TypeError:
+        pass
+    return f
+
+
+def _unsugar_count_from(**kw):
+    """Builds counting functions from keyword arguments.
+
+    Keyword argument filter, prepares a simple ``ordering_func`` from a
+    ``count_from`` argument, otherwise passes ``ordering_func`` on unchanged.
+    """
+
+    count_from = kw.pop("count_from", None)
+    if kw.get("ordering_func", None) is None and count_from is not None:
+        if count_from == 0:
+            kw["ordering_func"] = count_from_0
+        elif count_from == 1:
+            kw["ordering_func"] = count_from_1
+        else:
+            kw["ordering_func"] = count_from_n_factory(count_from)
+    return kw
+
+
+class OrderingList(List[_T]):
+    """A custom list that manages position information for its children.
+
+    The :class:`.OrderingList` object is normally set up using the
+    :func:`.ordering_list` factory function, used in conjunction with
+    the :func:`_orm.relationship` function.
+
+    """
+
+    ordering_attr: str
+    ordering_func: OrderingFunc
+    reorder_on_append: bool
+
+    def __init__(
+        self,
+        ordering_attr: Optional[str] = None,
+        ordering_func: Optional[OrderingFunc] = None,
+        reorder_on_append: bool = False,
+    ):
+        """A custom list that manages position information for its children.
+
+        ``OrderingList`` is a ``collection_class`` list implementation that
+        syncs position in a Python list with a position attribute on the
+        mapped objects.
+
+        This implementation relies on the list starting in the proper order,
+        so be **sure** to put an ``order_by`` on your relationship.
+
+        :param ordering_attr:
+          Name of the attribute that stores the object's order in the
+          relationship.
+
+        :param ordering_func: Optional.  A function that maps the position in
+          the Python list to a value to store in the
+          ``ordering_attr``.  Values returned are usually (but need not be!)
+          integers.
+
+          An ``ordering_func`` is called with two positional parameters: the
+          index of the element in the list, and the list itself.
+
+          If omitted, Python list indexes are used for the attribute values.
+          Two basic pre-built numbering functions are provided in this module:
+          ``count_from_0`` and ``count_from_1``.  For more exotic examples
+          like stepped numbering, alphabetical and Fibonacci numbering, see
+          the unit tests.
+
+        :param reorder_on_append:
+          Default False.  When appending an object with an existing (non-None)
+          ordering value, that value will be left untouched unless
+          ``reorder_on_append`` is true.  This is an optimization to avoid a
+          variety of dangerous unexpected database writes.
+
+          SQLAlchemy will add instances to the list via append() when your
+          object loads.  If for some reason the result set from the database
+          skips a step in the ordering (say, row '1' is missing but you get
+          '2', '3', and '4'), reorder_on_append=True would immediately
+          renumber the items to '1', '2', '3'.  If you have multiple sessions
+          making changes, any of whom happen to load this collection even in
+          passing, all of the sessions would try to "clean up" the numbering
+          in their commits, possibly causing all but one to fail with a
+          concurrent modification error.
+
+          Recommend leaving this with the default of False, and just call
+          ``reorder()`` if you're doing ``append()`` operations with
+          previously ordered instances or when doing some housekeeping after
+          manual sql operations.
+
+        """
+        self.ordering_attr = ordering_attr
+        if ordering_func is None:
+            ordering_func = count_from_0
+        self.ordering_func = ordering_func
+        self.reorder_on_append = reorder_on_append
+
+    # More complex serialization schemes (multi column, e.g.) are possible by
+    # subclassing and reimplementing these two methods.
+    def _get_order_value(self, entity):
+        return getattr(entity, self.ordering_attr)
+
+    def _set_order_value(self, entity, value):
+        setattr(entity, self.ordering_attr, value)
+
+    def reorder(self) -> None:
+        """Synchronize ordering for the entire collection.
+
+        Sweeps through the list and ensures that each object has accurate
+        ordering information set.
+
+        """
+        for index, entity in enumerate(self):
+            self._order_entity(index, entity, True)
+
+    # As of 0.5, _reorder is no longer semi-private
+    _reorder = reorder
+
+    def _order_entity(self, index, entity, reorder=True):
+        have = self._get_order_value(entity)
+
+        # Don't disturb existing ordering if reorder is False
+        if have is not None and not reorder:
+            return
+
+        should_be = self.ordering_func(index, self)
+        if have != should_be:
+            self._set_order_value(entity, should_be)
+
+    def append(self, entity):
+        super().append(entity)
+        self._order_entity(len(self) - 1, entity, self.reorder_on_append)
+
+    def _raw_append(self, entity):
+        """Append without any ordering behavior."""
+
+        super().append(entity)
+
+    _raw_append = collection.adds(1)(_raw_append)
+
+    def insert(self, index, entity):
+        super().insert(index, entity)
+        self._reorder()
+
+    def remove(self, entity):
+        super().remove(entity)
+
+        adapter = collection_adapter(self)
+        if adapter and adapter._referenced_by_owner:
+            self._reorder()
+
+    def pop(self, index=-1):
+        entity = super().pop(index)
+        self._reorder()
+        return entity
+
+    def __setitem__(self, index, entity):
+        if isinstance(index, slice):
+            step = index.step or 1
+            start = index.start or 0
+            if start < 0:
+                start += len(self)
+            stop = index.stop or len(self)
+            if stop < 0:
+                stop += len(self)
+
+            for i in range(start, stop, step):
+                self.__setitem__(i, entity[i])
+        else:
+            self._order_entity(index, entity, True)
+            super().__setitem__(index, entity)
+
+    def __delitem__(self, index):
+        super().__delitem__(index)
+        self._reorder()
+
+    def __setslice__(self, start, end, values):
+        super().__setslice__(start, end, values)
+        self._reorder()
+
+    def __delslice__(self, start, end):
+        super().__delslice__(start, end)
+        self._reorder()
+
+    def __reduce__(self):
+        return _reconstitute, (self.__class__, self.__dict__, list(self))
+
+    for func_name, func in list(locals().items()):
+        if (
+            callable(func)
+            and func.__name__ == func_name
+            and not func.__doc__
+            and hasattr(list, func_name)
+        ):
+            func.__doc__ = getattr(list, func_name).__doc__
+    del func_name, func
+
+
+def _reconstitute(cls, dict_, items):
+    """Reconstitute an :class:`.OrderingList`.
+
+    This is the adjoint to :meth:`.OrderingList.__reduce__`.  It is used for
+    unpickling :class:`.OrderingList` objects.
+
+    """
+    obj = cls.__new__(cls)
+    obj.__dict__.update(dict_)
+    list.extend(obj, items)
+    return obj
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/serializer.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/serializer.py
new file mode 100644
index 00000000..b7032b65
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/ext/serializer.py
@@ -0,0 +1,185 @@
+# ext/serializer.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""Serializer/Deserializer objects for usage with SQLAlchemy query structures,
+allowing "contextual" deserialization.
+
+.. legacy::
+
+    The serializer extension is **legacy** and should not be used for
+    new development.
+
+Any SQLAlchemy query structure, either based on sqlalchemy.sql.*
+or sqlalchemy.orm.* can be used.  The mappers, Tables, Columns, Session
+etc. which are referenced by the structure are not persisted in serialized
+form, but are instead re-associated with the query structure
+when it is deserialized.
+
+.. warning:: The serializer extension uses pickle to serialize and
+   deserialize objects, so the same security consideration mentioned
+   in the `python documentation
+   <https://docs.python.org/3/library/pickle.html>`_ apply.
+
+Usage is nearly the same as that of the standard Python pickle module::
+
+    from sqlalchemy.ext.serializer import loads, dumps
+
+    metadata = MetaData(bind=some_engine)
+    Session = scoped_session(sessionmaker())
+
+    # ... define mappers
+
+    query = (
+        Session.query(MyClass)
+        .filter(MyClass.somedata == "foo")
+        .order_by(MyClass.sortkey)
+    )
+
+    # pickle the query
+    serialized = dumps(query)
+
+    # unpickle.  Pass in metadata + scoped_session
+    query2 = loads(serialized, metadata, Session)
+
+    print(query2.all())
+
+Similar restrictions as when using raw pickle apply; mapped classes must be
+themselves be pickleable, meaning they are importable from a module-level
+namespace.
+
+The serializer module is only appropriate for query structures.  It is not
+needed for:
+
+* instances of user-defined classes.   These contain no references to engines,
+  sessions or expression constructs in the typical case and can be serialized
+  directly.
+
+* Table metadata that is to be loaded entirely from the serialized structure
+  (i.e. is not already declared in the application).   Regular
+  pickle.loads()/dumps() can be used to fully dump any ``MetaData`` object,
+  typically one which was reflected from an existing database at some previous
+  point in time.  The serializer module is specifically for the opposite case,
+  where the Table metadata is already present in memory.
+
+"""
+
+from io import BytesIO
+import pickle
+import re
+
+from .. import Column
+from .. import Table
+from ..engine import Engine
+from ..orm import class_mapper
+from ..orm.interfaces import MapperProperty
+from ..orm.mapper import Mapper
+from ..orm.session import Session
+from ..util import b64decode
+from ..util import b64encode
+
+
+__all__ = ["Serializer", "Deserializer", "dumps", "loads"]
+
+
+class Serializer(pickle.Pickler):
+
+    def persistent_id(self, obj):
+        # print "serializing:", repr(obj)
+        if isinstance(obj, Mapper) and not obj.non_primary:
+            id_ = "mapper:" + b64encode(pickle.dumps(obj.class_))
+        elif isinstance(obj, MapperProperty) and not obj.parent.non_primary:
+            id_ = (
+                "mapperprop:"
+                + b64encode(pickle.dumps(obj.parent.class_))
+                + ":"
+                + obj.key
+            )
+        elif isinstance(obj, Table):
+            if "parententity" in obj._annotations:
+                id_ = "mapper_selectable:" + b64encode(
+                    pickle.dumps(obj._annotations["parententity"].class_)
+                )
+            else:
+                id_ = f"table:{obj.key}"
+        elif isinstance(obj, Column) and isinstance(obj.table, Table):
+            id_ = f"column:{obj.table.key}:{obj.key}"
+        elif isinstance(obj, Session):
+            id_ = "session:"
+        elif isinstance(obj, Engine):
+            id_ = "engine:"
+        else:
+            return None
+        return id_
+
+
+our_ids = re.compile(
+    r"(mapperprop|mapper|mapper_selectable|table|column|"
+    r"session|attribute|engine):(.*)"
+)
+
+
+class Deserializer(pickle.Unpickler):
+
+    def __init__(self, file, metadata=None, scoped_session=None, engine=None):
+        super().__init__(file)
+        self.metadata = metadata
+        self.scoped_session = scoped_session
+        self.engine = engine
+
+    def get_engine(self):
+        if self.engine:
+            return self.engine
+        elif self.scoped_session and self.scoped_session().bind:
+            return self.scoped_session().bind
+        else:
+            return None
+
+    def persistent_load(self, id_):
+        m = our_ids.match(str(id_))
+        if not m:
+            return None
+        else:
+            type_, args = m.group(1, 2)
+            if type_ == "attribute":
+                key, clsarg = args.split(":")
+                cls = pickle.loads(b64decode(clsarg))
+                return getattr(cls, key)
+            elif type_ == "mapper":
+                cls = pickle.loads(b64decode(args))
+                return class_mapper(cls)
+            elif type_ == "mapper_selectable":
+                cls = pickle.loads(b64decode(args))
+                return class_mapper(cls).__clause_element__()
+            elif type_ == "mapperprop":
+                mapper, keyname = args.split(":")
+                cls = pickle.loads(b64decode(mapper))
+                return class_mapper(cls).attrs[keyname]
+            elif type_ == "table":
+                return self.metadata.tables[args]
+            elif type_ == "column":
+                table, colname = args.split(":")
+                return self.metadata.tables[table].c[colname]
+            elif type_ == "session":
+                return self.scoped_session()
+            elif type_ == "engine":
+                return self.get_engine()
+            else:
+                raise Exception("Unknown token: %s" % type_)
+
+
+def dumps(obj, protocol=pickle.HIGHEST_PROTOCOL):
+    buf = BytesIO()
+    pickler = Serializer(buf, protocol)
+    pickler.dump(obj)
+    return buf.getvalue()
+
+
+def loads(data, metadata=None, scoped_session=None, engine=None):
+    buf = BytesIO(data)
+    unpickler = Deserializer(buf, metadata, scoped_session, engine)
+    return unpickler.load()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__init__.py
new file mode 100644
index 00000000..ef9afb1a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__init__.py
@@ -0,0 +1,16 @@
+# future/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""2.0 API features.
+
+this module is legacy as 2.0 APIs are now standard.
+
+"""
+from .engine import Connection as Connection
+from .engine import create_engine as create_engine
+from .engine import Engine as Engine
+from ..sql._selectable_constructors import select as select
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..75551899
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/engine.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/engine.cpython-312.pyc
new file mode 100644
index 00000000..e6c731a7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/__pycache__/engine.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/engine.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/engine.py
new file mode 100644
index 00000000..0449c3d9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/future/engine.py
@@ -0,0 +1,15 @@
+# future/engine.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+"""2.0 API features.
+
+this module is legacy as 2.0 APIs are now standard.
+
+"""
+
+from ..engine import Connection as Connection  # noqa: F401
+from ..engine import create_engine as create_engine  # noqa: F401
+from ..engine import Engine as Engine  # noqa: F401
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/inspection.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/inspection.py
new file mode 100644
index 00000000..2e5b2201
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/inspection.py
@@ -0,0 +1,174 @@
+# inspection.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""The inspection module provides the :func:`_sa.inspect` function,
+which delivers runtime information about a wide variety
+of SQLAlchemy objects, both within the Core as well as the
+ORM.
+
+The :func:`_sa.inspect` function is the entry point to SQLAlchemy's
+public API for viewing the configuration and construction
+of in-memory objects.   Depending on the type of object
+passed to :func:`_sa.inspect`, the return value will either be
+a related object which provides a known interface, or in many
+cases it will return the object itself.
+
+The rationale for :func:`_sa.inspect` is twofold.  One is that
+it replaces the need to be aware of a large variety of "information
+getting" functions in SQLAlchemy, such as
+:meth:`_reflection.Inspector.from_engine` (deprecated in 1.4),
+:func:`.orm.attributes.instance_state`, :func:`_orm.class_mapper`,
+and others.    The other is that the return value of :func:`_sa.inspect`
+is guaranteed to obey a documented API, thus allowing third party
+tools which build on top of SQLAlchemy configurations to be constructed
+in a forwards-compatible way.
+
+"""
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import Optional
+from typing import overload
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+from . import exc
+from .util.typing import Literal
+from .util.typing import Protocol
+
+_T = TypeVar("_T", bound=Any)
+_TCov = TypeVar("_TCov", bound=Any, covariant=True)
+_F = TypeVar("_F", bound=Callable[..., Any])
+
+_IN = TypeVar("_IN", bound=Any)
+
+_registrars: Dict[type, Union[Literal[True], Callable[[Any], Any]]] = {}
+
+
+class Inspectable(Generic[_T]):
+    """define a class as inspectable.
+
+    This allows typing to set up a linkage between an object that
+    can be inspected and the type of inspection it returns.
+
+    Unfortunately we cannot at the moment get all classes that are
+    returned by inspection to suit this interface as we get into
+    MRO issues.
+
+    """
+
+    __slots__ = ()
+
+
+class _InspectableTypeProtocol(Protocol[_TCov]):
+    """a protocol defining a method that's used when a type (ie the class
+    itself) is passed to inspect().
+
+    """
+
+    def _sa_inspect_type(self) -> _TCov: ...
+
+
+class _InspectableProtocol(Protocol[_TCov]):
+    """a protocol defining a method that's used when an instance is
+    passed to inspect().
+
+    """
+
+    def _sa_inspect_instance(self) -> _TCov: ...
+
+
+@overload
+def inspect(
+    subject: Type[_InspectableTypeProtocol[_IN]], raiseerr: bool = True
+) -> _IN: ...
+
+
+@overload
+def inspect(
+    subject: _InspectableProtocol[_IN], raiseerr: bool = True
+) -> _IN: ...
+
+
+@overload
+def inspect(subject: Inspectable[_IN], raiseerr: bool = True) -> _IN: ...
+
+
+@overload
+def inspect(subject: Any, raiseerr: Literal[False] = ...) -> Optional[Any]: ...
+
+
+@overload
+def inspect(subject: Any, raiseerr: bool = True) -> Any: ...
+
+
+def inspect(subject: Any, raiseerr: bool = True) -> Any:
+    """Produce an inspection object for the given target.
+
+    The returned value in some cases may be the
+    same object as the one given, such as if a
+    :class:`_orm.Mapper` object is passed.   In other
+    cases, it will be an instance of the registered
+    inspection type for the given object, such as
+    if an :class:`_engine.Engine` is passed, an
+    :class:`_reflection.Inspector` object is returned.
+
+    :param subject: the subject to be inspected.
+    :param raiseerr: When ``True``, if the given subject
+     does not
+     correspond to a known SQLAlchemy inspected type,
+     :class:`sqlalchemy.exc.NoInspectionAvailable`
+     is raised.  If ``False``, ``None`` is returned.
+
+    """
+    type_ = type(subject)
+    for cls in type_.__mro__:
+        if cls in _registrars:
+            reg = _registrars.get(cls, None)
+            if reg is None:
+                continue
+            elif reg is True:
+                return subject
+            ret = reg(subject)
+            if ret is not None:
+                return ret
+    else:
+        reg = ret = None
+
+    if raiseerr and (reg is None or ret is None):
+        raise exc.NoInspectionAvailable(
+            "No inspection system is "
+            "available for object of type %s" % type_
+        )
+    return ret
+
+
+def _inspects(
+    *types: Type[Any],
+) -> Callable[[_F], _F]:
+    def decorate(fn_or_cls: _F) -> _F:
+        for type_ in types:
+            if type_ in _registrars:
+                raise AssertionError("Type %s is already registered" % type_)
+            _registrars[type_] = fn_or_cls
+        return fn_or_cls
+
+    return decorate
+
+
+_TT = TypeVar("_TT", bound="Type[Any]")
+
+
+def _self_inspects(cls: _TT) -> _TT:
+    if cls in _registrars:
+        raise AssertionError("Type %s is already registered" % cls)
+    _registrars[cls] = True
+    return cls
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/log.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/log.py
new file mode 100644
index 00000000..849a0bfa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/log.py
@@ -0,0 +1,288 @@
+# log.py
+# Copyright (C) 2006-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+# Includes alterations by Vinay Sajip vinay_sajip@yahoo.co.uk
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Logging control and utilities.
+
+Control of logging for SA can be performed from the regular python logging
+module.  The regular dotted module namespace is used, starting at
+'sqlalchemy'.  For class-level logging, the class name is appended.
+
+The "echo" keyword parameter, available on SQLA :class:`_engine.Engine`
+and :class:`_pool.Pool` objects, corresponds to a logger specific to that
+instance only.
+
+"""
+from __future__ import annotations
+
+import logging
+import sys
+from typing import Any
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+from .util import py311
+from .util import py38
+from .util.typing import Literal
+
+
+if py38:
+    STACKLEVEL = True
+    # needed as of py3.11.0b1
+    # #8019
+    STACKLEVEL_OFFSET = 2 if py311 else 1
+else:
+    STACKLEVEL = False
+    STACKLEVEL_OFFSET = 0
+
+_IT = TypeVar("_IT", bound="Identified")
+
+_EchoFlagType = Union[None, bool, Literal["debug"]]
+
+# set initial level to WARN.  This so that
+# log statements don't occur in the absence of explicit
+# logging being enabled for 'sqlalchemy'.
+rootlogger = logging.getLogger("sqlalchemy")
+if rootlogger.level == logging.NOTSET:
+    rootlogger.setLevel(logging.WARN)
+
+
+def _add_default_handler(logger: logging.Logger) -> None:
+    handler = logging.StreamHandler(sys.stdout)
+    handler.setFormatter(
+        logging.Formatter("%(asctime)s %(levelname)s %(name)s %(message)s")
+    )
+    logger.addHandler(handler)
+
+
+_logged_classes: Set[Type[Identified]] = set()
+
+
+def _qual_logger_name_for_cls(cls: Type[Identified]) -> str:
+    return (
+        getattr(cls, "_sqla_logger_namespace", None)
+        or cls.__module__ + "." + cls.__name__
+    )
+
+
+def class_logger(cls: Type[_IT]) -> Type[_IT]:
+    logger = logging.getLogger(_qual_logger_name_for_cls(cls))
+    cls._should_log_debug = lambda self: logger.isEnabledFor(  # type: ignore[method-assign]  # noqa: E501
+        logging.DEBUG
+    )
+    cls._should_log_info = lambda self: logger.isEnabledFor(  # type: ignore[method-assign]  # noqa: E501
+        logging.INFO
+    )
+    cls.logger = logger
+    _logged_classes.add(cls)
+    return cls
+
+
+_IdentifiedLoggerType = Union[logging.Logger, "InstanceLogger"]
+
+
+class Identified:
+    __slots__ = ()
+
+    logging_name: Optional[str] = None
+
+    logger: _IdentifiedLoggerType
+
+    _echo: _EchoFlagType
+
+    def _should_log_debug(self) -> bool:
+        return self.logger.isEnabledFor(logging.DEBUG)
+
+    def _should_log_info(self) -> bool:
+        return self.logger.isEnabledFor(logging.INFO)
+
+
+class InstanceLogger:
+    """A logger adapter (wrapper) for :class:`.Identified` subclasses.
+
+    This allows multiple instances (e.g. Engine or Pool instances)
+    to share a logger, but have its verbosity controlled on a
+    per-instance basis.
+
+    The basic functionality is to return a logging level
+    which is based on an instance's echo setting.
+
+    Default implementation is:
+
+    'debug' -> logging.DEBUG
+    True    -> logging.INFO
+    False   -> Effective level of underlying logger (
+    logging.WARNING by default)
+    None    -> same as False
+    """
+
+    # Map echo settings to logger levels
+    _echo_map = {
+        None: logging.NOTSET,
+        False: logging.NOTSET,
+        True: logging.INFO,
+        "debug": logging.DEBUG,
+    }
+
+    _echo: _EchoFlagType
+
+    __slots__ = ("echo", "logger")
+
+    def __init__(self, echo: _EchoFlagType, name: str):
+        self.echo = echo
+        self.logger = logging.getLogger(name)
+
+        # if echo flag is enabled and no handlers,
+        # add a handler to the list
+        if self._echo_map[echo] <= logging.INFO and not self.logger.handlers:
+            _add_default_handler(self.logger)
+
+    #
+    # Boilerplate convenience methods
+    #
+    def debug(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate a debug call to the underlying logger."""
+
+        self.log(logging.DEBUG, msg, *args, **kwargs)
+
+    def info(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate an info call to the underlying logger."""
+
+        self.log(logging.INFO, msg, *args, **kwargs)
+
+    def warning(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate a warning call to the underlying logger."""
+
+        self.log(logging.WARNING, msg, *args, **kwargs)
+
+    warn = warning
+
+    def error(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """
+        Delegate an error call to the underlying logger.
+        """
+        self.log(logging.ERROR, msg, *args, **kwargs)
+
+    def exception(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate an exception call to the underlying logger."""
+
+        kwargs["exc_info"] = 1
+        self.log(logging.ERROR, msg, *args, **kwargs)
+
+    def critical(self, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate a critical call to the underlying logger."""
+
+        self.log(logging.CRITICAL, msg, *args, **kwargs)
+
+    def log(self, level: int, msg: str, *args: Any, **kwargs: Any) -> None:
+        """Delegate a log call to the underlying logger.
+
+        The level here is determined by the echo
+        flag as well as that of the underlying logger, and
+        logger._log() is called directly.
+
+        """
+
+        # inline the logic from isEnabledFor(),
+        # getEffectiveLevel(), to avoid overhead.
+
+        if self.logger.manager.disable >= level:
+            return
+
+        selected_level = self._echo_map[self.echo]
+        if selected_level == logging.NOTSET:
+            selected_level = self.logger.getEffectiveLevel()
+
+        if level >= selected_level:
+            if STACKLEVEL:
+                kwargs["stacklevel"] = (
+                    kwargs.get("stacklevel", 1) + STACKLEVEL_OFFSET
+                )
+
+            self.logger._log(level, msg, args, **kwargs)
+
+    def isEnabledFor(self, level: int) -> bool:
+        """Is this logger enabled for level 'level'?"""
+
+        if self.logger.manager.disable >= level:
+            return False
+        return level >= self.getEffectiveLevel()
+
+    def getEffectiveLevel(self) -> int:
+        """What's the effective level for this logger?"""
+
+        level = self._echo_map[self.echo]
+        if level == logging.NOTSET:
+            level = self.logger.getEffectiveLevel()
+        return level
+
+
+def instance_logger(
+    instance: Identified, echoflag: _EchoFlagType = None
+) -> None:
+    """create a logger for an instance that implements :class:`.Identified`."""
+
+    if instance.logging_name:
+        name = "%s.%s" % (
+            _qual_logger_name_for_cls(instance.__class__),
+            instance.logging_name,
+        )
+    else:
+        name = _qual_logger_name_for_cls(instance.__class__)
+
+    instance._echo = echoflag  # type: ignore
+
+    logger: Union[logging.Logger, InstanceLogger]
+
+    if echoflag in (False, None):
+        # if no echo setting or False, return a Logger directly,
+        # avoiding overhead of filtering
+        logger = logging.getLogger(name)
+    else:
+        # if a specified echo flag, return an EchoLogger,
+        # which checks the flag, overrides normal log
+        # levels by calling logger._log()
+        logger = InstanceLogger(echoflag, name)
+
+    instance.logger = logger  # type: ignore
+
+
+class echo_property:
+    __doc__ = """\
+    When ``True``, enable log output for this element.
+
+    This has the effect of setting the Python logging level for the namespace
+    of this element's class and object reference.  A value of boolean ``True``
+    indicates that the loglevel ``logging.INFO`` will be set for the logger,
+    whereas the string value ``debug`` will set the loglevel to
+    ``logging.DEBUG``.
+    """
+
+    @overload
+    def __get__(
+        self, instance: Literal[None], owner: Type[Identified]
+    ) -> echo_property: ...
+
+    @overload
+    def __get__(
+        self, instance: Identified, owner: Type[Identified]
+    ) -> _EchoFlagType: ...
+
+    def __get__(
+        self, instance: Optional[Identified], owner: Type[Identified]
+    ) -> Union[echo_property, _EchoFlagType]:
+        if instance is None:
+            return self
+        else:
+            return instance._echo
+
+    def __set__(self, instance: Identified, value: _EchoFlagType) -> None:
+        instance_logger(instance, echoflag=value)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__init__.py
new file mode 100644
index 00000000..7771de47
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__init__.py
@@ -0,0 +1,170 @@
+# orm/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""
+Functional constructs for ORM configuration.
+
+See the SQLAlchemy object relational tutorial and mapper configuration
+documentation for an overview of how this module is used.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from . import exc as exc
+from . import mapper as mapperlib
+from . import strategy_options as strategy_options
+from ._orm_constructors import _mapper_fn as mapper
+from ._orm_constructors import aliased as aliased
+from ._orm_constructors import backref as backref
+from ._orm_constructors import clear_mappers as clear_mappers
+from ._orm_constructors import column_property as column_property
+from ._orm_constructors import composite as composite
+from ._orm_constructors import contains_alias as contains_alias
+from ._orm_constructors import create_session as create_session
+from ._orm_constructors import deferred as deferred
+from ._orm_constructors import dynamic_loader as dynamic_loader
+from ._orm_constructors import join as join
+from ._orm_constructors import mapped_column as mapped_column
+from ._orm_constructors import orm_insert_sentinel as orm_insert_sentinel
+from ._orm_constructors import outerjoin as outerjoin
+from ._orm_constructors import query_expression as query_expression
+from ._orm_constructors import relationship as relationship
+from ._orm_constructors import synonym as synonym
+from ._orm_constructors import with_loader_criteria as with_loader_criteria
+from ._orm_constructors import with_polymorphic as with_polymorphic
+from .attributes import AttributeEventToken as AttributeEventToken
+from .attributes import InstrumentedAttribute as InstrumentedAttribute
+from .attributes import QueryableAttribute as QueryableAttribute
+from .base import class_mapper as class_mapper
+from .base import DynamicMapped as DynamicMapped
+from .base import InspectionAttrExtensionType as InspectionAttrExtensionType
+from .base import LoaderCallableStatus as LoaderCallableStatus
+from .base import Mapped as Mapped
+from .base import NotExtension as NotExtension
+from .base import ORMDescriptor as ORMDescriptor
+from .base import PassiveFlag as PassiveFlag
+from .base import SQLORMExpression as SQLORMExpression
+from .base import WriteOnlyMapped as WriteOnlyMapped
+from .context import FromStatement as FromStatement
+from .context import QueryContext as QueryContext
+from .decl_api import add_mapped_attribute as add_mapped_attribute
+from .decl_api import as_declarative as as_declarative
+from .decl_api import declarative_base as declarative_base
+from .decl_api import declarative_mixin as declarative_mixin
+from .decl_api import DeclarativeBase as DeclarativeBase
+from .decl_api import DeclarativeBaseNoMeta as DeclarativeBaseNoMeta
+from .decl_api import DeclarativeMeta as DeclarativeMeta
+from .decl_api import declared_attr as declared_attr
+from .decl_api import has_inherited_table as has_inherited_table
+from .decl_api import MappedAsDataclass as MappedAsDataclass
+from .decl_api import registry as registry
+from .decl_api import synonym_for as synonym_for
+from .decl_base import MappedClassProtocol as MappedClassProtocol
+from .descriptor_props import Composite as Composite
+from .descriptor_props import CompositeProperty as CompositeProperty
+from .descriptor_props import Synonym as Synonym
+from .descriptor_props import SynonymProperty as SynonymProperty
+from .dynamic import AppenderQuery as AppenderQuery
+from .events import AttributeEvents as AttributeEvents
+from .events import InstanceEvents as InstanceEvents
+from .events import InstrumentationEvents as InstrumentationEvents
+from .events import MapperEvents as MapperEvents
+from .events import QueryEvents as QueryEvents
+from .events import SessionEvents as SessionEvents
+from .identity import IdentityMap as IdentityMap
+from .instrumentation import ClassManager as ClassManager
+from .interfaces import EXT_CONTINUE as EXT_CONTINUE
+from .interfaces import EXT_SKIP as EXT_SKIP
+from .interfaces import EXT_STOP as EXT_STOP
+from .interfaces import InspectionAttr as InspectionAttr
+from .interfaces import InspectionAttrInfo as InspectionAttrInfo
+from .interfaces import MANYTOMANY as MANYTOMANY
+from .interfaces import MANYTOONE as MANYTOONE
+from .interfaces import MapperProperty as MapperProperty
+from .interfaces import NO_KEY as NO_KEY
+from .interfaces import NO_VALUE as NO_VALUE
+from .interfaces import ONETOMANY as ONETOMANY
+from .interfaces import PropComparator as PropComparator
+from .interfaces import RelationshipDirection as RelationshipDirection
+from .interfaces import UserDefinedOption as UserDefinedOption
+from .loading import merge_frozen_result as merge_frozen_result
+from .loading import merge_result as merge_result
+from .mapped_collection import attribute_keyed_dict as attribute_keyed_dict
+from .mapped_collection import (
+    attribute_mapped_collection as attribute_mapped_collection,
+)
+from .mapped_collection import column_keyed_dict as column_keyed_dict
+from .mapped_collection import (
+    column_mapped_collection as column_mapped_collection,
+)
+from .mapped_collection import keyfunc_mapping as keyfunc_mapping
+from .mapped_collection import KeyFuncDict as KeyFuncDict
+from .mapped_collection import mapped_collection as mapped_collection
+from .mapped_collection import MappedCollection as MappedCollection
+from .mapper import configure_mappers as configure_mappers
+from .mapper import Mapper as Mapper
+from .mapper import reconstructor as reconstructor
+from .mapper import validates as validates
+from .properties import ColumnProperty as ColumnProperty
+from .properties import MappedColumn as MappedColumn
+from .properties import MappedSQLExpression as MappedSQLExpression
+from .query import AliasOption as AliasOption
+from .query import Query as Query
+from .relationships import foreign as foreign
+from .relationships import Relationship as Relationship
+from .relationships import RelationshipProperty as RelationshipProperty
+from .relationships import remote as remote
+from .scoping import QueryPropertyDescriptor as QueryPropertyDescriptor
+from .scoping import scoped_session as scoped_session
+from .session import close_all_sessions as close_all_sessions
+from .session import make_transient as make_transient
+from .session import make_transient_to_detached as make_transient_to_detached
+from .session import object_session as object_session
+from .session import ORMExecuteState as ORMExecuteState
+from .session import Session as Session
+from .session import sessionmaker as sessionmaker
+from .session import SessionTransaction as SessionTransaction
+from .session import SessionTransactionOrigin as SessionTransactionOrigin
+from .state import AttributeState as AttributeState
+from .state import InstanceState as InstanceState
+from .strategy_options import contains_eager as contains_eager
+from .strategy_options import defaultload as defaultload
+from .strategy_options import defer as defer
+from .strategy_options import immediateload as immediateload
+from .strategy_options import joinedload as joinedload
+from .strategy_options import lazyload as lazyload
+from .strategy_options import Load as Load
+from .strategy_options import load_only as load_only
+from .strategy_options import noload as noload
+from .strategy_options import raiseload as raiseload
+from .strategy_options import selectin_polymorphic as selectin_polymorphic
+from .strategy_options import selectinload as selectinload
+from .strategy_options import subqueryload as subqueryload
+from .strategy_options import undefer as undefer
+from .strategy_options import undefer_group as undefer_group
+from .strategy_options import with_expression as with_expression
+from .unitofwork import UOWTransaction as UOWTransaction
+from .util import Bundle as Bundle
+from .util import CascadeOptions as CascadeOptions
+from .util import LoaderCriteriaOption as LoaderCriteriaOption
+from .util import object_mapper as object_mapper
+from .util import polymorphic_union as polymorphic_union
+from .util import was_deleted as was_deleted
+from .util import with_parent as with_parent
+from .writeonly import WriteOnlyCollection as WriteOnlyCollection
+from .. import util as _sa_util
+
+
+def __go(lcls: Any) -> None:
+    _sa_util.preloaded.import_prefix("sqlalchemy.orm")
+    _sa_util.preloaded.import_prefix("sqlalchemy.ext")
+
+
+__go(locals())
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..28a533c8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_orm_constructors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_orm_constructors.cpython-312.pyc
new file mode 100644
index 00000000..1ca9fab2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_orm_constructors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_typing.cpython-312.pyc
new file mode 100644
index 00000000..6858ad59
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/_typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/attributes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/attributes.cpython-312.pyc
new file mode 100644
index 00000000..03358733
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/attributes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..15353ca6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/bulk_persistence.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/bulk_persistence.cpython-312.pyc
new file mode 100644
index 00000000..484dc563
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/bulk_persistence.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/clsregistry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/clsregistry.cpython-312.pyc
new file mode 100644
index 00000000..689e1ef4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/clsregistry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/collections.cpython-312.pyc
new file mode 100644
index 00000000..7763e6cc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/context.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/context.cpython-312.pyc
new file mode 100644
index 00000000..983b8706
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/context.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_api.cpython-312.pyc
new file mode 100644
index 00000000..3e50bb77
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_base.cpython-312.pyc
new file mode 100644
index 00000000..53be5e5e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/decl_base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dependency.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dependency.cpython-312.pyc
new file mode 100644
index 00000000..aff99dcf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dependency.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/descriptor_props.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/descriptor_props.cpython-312.pyc
new file mode 100644
index 00000000..875fcc63
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/descriptor_props.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dynamic.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dynamic.cpython-312.pyc
new file mode 100644
index 00000000..01dd5177
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/dynamic.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/evaluator.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/evaluator.cpython-312.pyc
new file mode 100644
index 00000000..19b7d84b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/evaluator.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/events.cpython-312.pyc
new file mode 100644
index 00000000..2d061417
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/exc.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/exc.cpython-312.pyc
new file mode 100644
index 00000000..184de6fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/exc.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/identity.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/identity.cpython-312.pyc
new file mode 100644
index 00000000..3ad622a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/identity.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/instrumentation.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/instrumentation.cpython-312.pyc
new file mode 100644
index 00000000..f1956378
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/instrumentation.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/interfaces.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/interfaces.cpython-312.pyc
new file mode 100644
index 00000000..c799e047
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/interfaces.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/loading.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/loading.cpython-312.pyc
new file mode 100644
index 00000000..f84cdb1f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/loading.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapped_collection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapped_collection.cpython-312.pyc
new file mode 100644
index 00000000..82ce6849
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapped_collection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapper.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapper.cpython-312.pyc
new file mode 100644
index 00000000..0d543c09
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/mapper.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/path_registry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/path_registry.cpython-312.pyc
new file mode 100644
index 00000000..3e4f3bd1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/path_registry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/persistence.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/persistence.cpython-312.pyc
new file mode 100644
index 00000000..7b321a44
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/persistence.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/properties.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/properties.cpython-312.pyc
new file mode 100644
index 00000000..b1d33d73
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/properties.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/query.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/query.cpython-312.pyc
new file mode 100644
index 00000000..46e81b17
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/query.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/relationships.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/relationships.cpython-312.pyc
new file mode 100644
index 00000000..fc031ea4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/relationships.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/scoping.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/scoping.cpython-312.pyc
new file mode 100644
index 00000000..cf1f1625
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/scoping.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/session.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/session.cpython-312.pyc
new file mode 100644
index 00000000..b9788f7e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/session.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state.cpython-312.pyc
new file mode 100644
index 00000000..5301b49d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state_changes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state_changes.cpython-312.pyc
new file mode 100644
index 00000000..37ac4066
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/state_changes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategies.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategies.cpython-312.pyc
new file mode 100644
index 00000000..3af799d8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategies.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategy_options.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategy_options.cpython-312.pyc
new file mode 100644
index 00000000..5cdaf88f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/strategy_options.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/sync.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/sync.cpython-312.pyc
new file mode 100644
index 00000000..57d68574
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/sync.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/unitofwork.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/unitofwork.cpython-312.pyc
new file mode 100644
index 00000000..958215d4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/unitofwork.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..7d40eabd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/writeonly.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/writeonly.cpython-312.pyc
new file mode 100644
index 00000000..83571c9f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/__pycache__/writeonly.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_orm_constructors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_orm_constructors.py
new file mode 100644
index 00000000..d9e3ec37
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_orm_constructors.py
@@ -0,0 +1,2590 @@
+# orm/_orm_constructors.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import Collection
+from typing import Iterable
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import mapperlib as mapperlib
+from ._typing import _O
+from .descriptor_props import Composite
+from .descriptor_props import Synonym
+from .interfaces import _AttributeOptions
+from .properties import MappedColumn
+from .properties import MappedSQLExpression
+from .query import AliasOption
+from .relationships import _RelationshipArgumentType
+from .relationships import _RelationshipDeclared
+from .relationships import _RelationshipSecondaryArgument
+from .relationships import RelationshipProperty
+from .session import Session
+from .util import _ORMJoin
+from .util import AliasedClass
+from .util import AliasedInsp
+from .util import LoaderCriteriaOption
+from .. import sql
+from .. import util
+from ..exc import InvalidRequestError
+from ..sql._typing import _no_kw
+from ..sql.base import _NoArg
+from ..sql.base import SchemaEventTarget
+from ..sql.schema import _InsertSentinelColumnDefault
+from ..sql.schema import SchemaConst
+from ..sql.selectable import FromClause
+from ..util.typing import Annotated
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _ORMColumnExprArgument
+    from .descriptor_props import _CC
+    from .descriptor_props import _CompositeAttrType
+    from .interfaces import PropComparator
+    from .mapper import Mapper
+    from .query import Query
+    from .relationships import _LazyLoadArgumentType
+    from .relationships import _ORMColCollectionArgument
+    from .relationships import _ORMOrderByArgument
+    from .relationships import _RelationshipJoinConditionArgument
+    from .relationships import ORMBackrefArgument
+    from .session import _SessionBind
+    from ..sql._typing import _AutoIncrementType
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _FromClauseArgument
+    from ..sql._typing import _InfoType
+    from ..sql._typing import _OnClauseArgument
+    from ..sql._typing import _TypeEngineArgument
+    from ..sql.elements import ColumnElement
+    from ..sql.schema import _ServerDefaultArgument
+    from ..sql.schema import _ServerOnUpdateArgument
+    from ..sql.selectable import Alias
+    from ..sql.selectable import Subquery
+
+
+_T = typing.TypeVar("_T")
+
+
+@util.deprecated(
+    "1.4",
+    "The :class:`.AliasOption` object is not necessary "
+    "for entities to be matched up to a query that is established "
+    "via :meth:`.Query.from_statement` and now does nothing.",
+    enable_warnings=False,  # AliasOption itself warns
+)
+def contains_alias(alias: Union[Alias, Subquery]) -> AliasOption:
+    r"""Return a :class:`.MapperOption` that will indicate to the
+    :class:`_query.Query`
+    that the main table has been aliased.
+
+    """
+    return AliasOption(alias)
+
+
+def mapped_column(
+    __name_pos: Optional[
+        Union[str, _TypeEngineArgument[Any], SchemaEventTarget]
+    ] = None,
+    __type_pos: Optional[
+        Union[_TypeEngineArgument[Any], SchemaEventTarget]
+    ] = None,
+    *args: SchemaEventTarget,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    nullable: Optional[
+        Union[bool, Literal[SchemaConst.NULL_UNSPECIFIED]]
+    ] = SchemaConst.NULL_UNSPECIFIED,
+    primary_key: Optional[bool] = False,
+    deferred: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    deferred_group: Optional[str] = None,
+    deferred_raiseload: Optional[bool] = None,
+    use_existing_column: bool = False,
+    name: Optional[str] = None,
+    type_: Optional[_TypeEngineArgument[Any]] = None,
+    autoincrement: _AutoIncrementType = "auto",
+    doc: Optional[str] = None,
+    key: Optional[str] = None,
+    index: Optional[bool] = None,
+    unique: Optional[bool] = None,
+    info: Optional[_InfoType] = None,
+    onupdate: Optional[Any] = None,
+    insert_default: Optional[Any] = _NoArg.NO_ARG,
+    server_default: Optional[_ServerDefaultArgument] = None,
+    server_onupdate: Optional[_ServerOnUpdateArgument] = None,
+    active_history: bool = False,
+    quote: Optional[bool] = None,
+    system: bool = False,
+    comment: Optional[str] = None,
+    sort_order: Union[_NoArg, int] = _NoArg.NO_ARG,
+    **kw: Any,
+) -> MappedColumn[Any]:
+    r"""declare a new ORM-mapped :class:`_schema.Column` construct
+    for use within :ref:`Declarative Table <orm_declarative_table>`
+    configuration.
+
+    The :func:`_orm.mapped_column` function provides an ORM-aware and
+    Python-typing-compatible construct which is used with
+    :ref:`declarative <orm_declarative_mapping>` mappings to indicate an
+    attribute that's mapped to a Core :class:`_schema.Column` object.  It
+    provides the equivalent feature as mapping an attribute to a
+    :class:`_schema.Column` object directly when using Declarative,
+    specifically when using :ref:`Declarative Table <orm_declarative_table>`
+    configuration.
+
+    .. versionadded:: 2.0
+
+    :func:`_orm.mapped_column` is normally used with explicit typing along with
+    the :class:`_orm.Mapped` annotation type, where it can derive the SQL
+    type and nullability for the column based on what's present within the
+    :class:`_orm.Mapped` annotation.   It also may be used without annotations
+    as a drop-in replacement for how :class:`_schema.Column` is used in
+    Declarative mappings in SQLAlchemy 1.x style.
+
+    For usage examples of :func:`_orm.mapped_column`, see the documentation
+    at :ref:`orm_declarative_table`.
+
+    .. seealso::
+
+        :ref:`orm_declarative_table` - complete documentation
+
+        :ref:`whatsnew_20_orm_declarative_typing` - migration notes for
+        Declarative mappings using 1.x style mappings
+
+    :param __name: String name to give to the :class:`_schema.Column`.  This
+     is an optional, positional only argument that if present must be the
+     first positional argument passed.  If omitted, the attribute name to
+     which the :func:`_orm.mapped_column`  is mapped will be used as the SQL
+     column name.
+    :param __type: :class:`_types.TypeEngine` type or instance which will
+     indicate the datatype to be associated with the :class:`_schema.Column`.
+     This is an optional, positional-only argument that if present must
+     immediately follow the ``__name`` parameter if present also, or otherwise
+     be the first positional parameter.  If omitted, the ultimate type for
+     the column may be derived either from the annotated type, or if a
+     :class:`_schema.ForeignKey` is present, from the datatype of the
+     referenced column.
+    :param \*args: Additional positional arguments include constructs such
+     as :class:`_schema.ForeignKey`, :class:`_schema.CheckConstraint`,
+     and :class:`_schema.Identity`, which are passed through to the constructed
+     :class:`_schema.Column`.
+    :param nullable: Optional bool, whether the column should be "NULL" or
+     "NOT NULL". If omitted, the nullability is derived from the type
+     annotation based on whether or not ``typing.Optional`` is present.
+     ``nullable`` defaults to ``True`` otherwise for non-primary key columns,
+     and ``False`` for primary key columns.
+    :param primary_key: optional bool, indicates the :class:`_schema.Column`
+     would be part of the table's primary key or not.
+    :param deferred: Optional bool - this keyword argument is consumed by the
+     ORM declarative process, and is not part of the :class:`_schema.Column`
+     itself; instead, it indicates that this column should be "deferred" for
+     loading as though mapped by :func:`_orm.deferred`.
+
+     .. seealso::
+
+        :ref:`orm_queryguide_deferred_declarative`
+
+    :param deferred_group: Implies :paramref:`_orm.mapped_column.deferred`
+     to ``True``, and set the :paramref:`_orm.deferred.group` parameter.
+
+     .. seealso::
+
+        :ref:`orm_queryguide_deferred_group`
+
+    :param deferred_raiseload: Implies :paramref:`_orm.mapped_column.deferred`
+     to ``True``, and set the :paramref:`_orm.deferred.raiseload` parameter.
+
+     .. seealso::
+
+        :ref:`orm_queryguide_deferred_raiseload`
+
+    :param use_existing_column: if True, will attempt to locate the given
+     column name on an inherited superclass (typically single inheriting
+     superclass), and if present, will not produce a new column, mapping
+     to the superclass column as though it were omitted from this class.
+     This is used for mixins that add new columns to an inherited superclass.
+
+     .. seealso::
+
+        :ref:`orm_inheritance_column_conflicts`
+
+     .. versionadded:: 2.0.0b4
+
+    :param default: Passed directly to the
+     :paramref:`_schema.Column.default` parameter if the
+     :paramref:`_orm.mapped_column.insert_default` parameter is not present.
+     Additionally, when used with :ref:`orm_declarative_native_dataclasses`,
+     indicates a default Python value that should be applied to the keyword
+     constructor within the generated ``__init__()`` method.
+
+     Note that in the case of dataclass generation when
+     :paramref:`_orm.mapped_column.insert_default` is not present, this means
+     the :paramref:`_orm.mapped_column.default` value is used in **two**
+     places, both the ``__init__()`` method as well as the
+     :paramref:`_schema.Column.default` parameter. While this behavior may
+     change in a future release, for the moment this tends to "work out"; a
+     default of ``None`` will mean that the :class:`_schema.Column` gets no
+     default generator, whereas a default that refers to a non-``None`` Python
+     or SQL expression value will be assigned up front on the object when
+     ``__init__()`` is called, which is the same value that the Core
+     :class:`_sql.Insert` construct would use in any case, leading to the same
+     end result.
+
+     .. note:: When using Core level column defaults that are callables to
+        be interpreted by the underlying :class:`_schema.Column` in conjunction
+        with :ref:`ORM-mapped dataclasses
+        <orm_declarative_native_dataclasses>`, especially those that are
+        :ref:`context-aware default functions <context_default_functions>`,
+        **the** :paramref:`_orm.mapped_column.insert_default` **parameter must
+        be used instead**.  This is necessary to disambiguate the callable from
+        being interpreted as a dataclass level default.
+
+     .. seealso::
+
+        :ref:`defaults_default_factory_insert_default`
+
+        :paramref:`_orm.mapped_column.insert_default`
+
+        :paramref:`_orm.mapped_column.default_factory`
+
+    :param insert_default: Passed directly to the
+     :paramref:`_schema.Column.default` parameter; will supersede the value
+     of :paramref:`_orm.mapped_column.default` when present, however
+     :paramref:`_orm.mapped_column.default` will always apply to the
+     constructor default for a dataclasses mapping.
+
+     .. seealso::
+
+        :ref:`defaults_default_factory_insert_default`
+
+        :paramref:`_orm.mapped_column.default`
+
+        :paramref:`_orm.mapped_column.default_factory`
+
+    :param sort_order: An integer that indicates how this mapped column
+     should be sorted compared to the others when the ORM is creating a
+     :class:`_schema.Table`. Among mapped columns that have the same
+     value the default ordering is used, placing first the mapped columns
+     defined in the main class, then the ones in the super classes.
+     Defaults to 0. The sort is ascending.
+
+     .. versionadded:: 2.0.4
+
+    :param active_history=False:
+
+        When ``True``, indicates that the "previous" value for a
+        scalar attribute should be loaded when replaced, if not
+        already loaded. Normally, history tracking logic for
+        simple non-primary-key scalar values only needs to be
+        aware of the "new" value in order to perform a flush. This
+        flag is available for applications that make use of
+        :func:`.attributes.get_history` or :meth:`.Session.is_modified`
+        which also need to know the "previous" value of the attribute.
+
+        .. versionadded:: 2.0.10
+
+
+    :param init: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__init__()``
+     method as generated by the dataclass process.
+    :param repr: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__repr__()``
+     method as generated by the dataclass process.
+    :param default_factory: Specific to
+     :ref:`orm_declarative_native_dataclasses`,
+     specifies a default-value generation function that will take place
+     as part of the ``__init__()``
+     method as generated by the dataclass process.
+
+     .. seealso::
+
+        :ref:`defaults_default_factory_insert_default`
+
+        :paramref:`_orm.mapped_column.default`
+
+        :paramref:`_orm.mapped_column.insert_default`
+
+    :param compare: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be included in comparison operations when generating the
+     ``__eq__()`` and ``__ne__()`` methods for the mapped class.
+
+     .. versionadded:: 2.0.0b4
+
+    :param kw_only: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be marked as keyword-only when generating the ``__init__()``.
+
+    :param hash: Specific to
+     :ref:`orm_declarative_native_dataclasses`, controls if this field
+     is included when generating the ``__hash__()`` method for the mapped
+     class.
+
+     .. versionadded:: 2.0.36
+
+    :param \**kw: All remaining keyword arguments are passed through to the
+     constructor for the :class:`_schema.Column`.
+
+    """
+
+    return MappedColumn(
+        __name_pos,
+        __type_pos,
+        *args,
+        name=name,
+        type_=type_,
+        autoincrement=autoincrement,
+        insert_default=insert_default,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+        doc=doc,
+        key=key,
+        index=index,
+        unique=unique,
+        info=info,
+        active_history=active_history,
+        nullable=nullable,
+        onupdate=onupdate,
+        primary_key=primary_key,
+        server_default=server_default,
+        server_onupdate=server_onupdate,
+        use_existing_column=use_existing_column,
+        quote=quote,
+        comment=comment,
+        system=system,
+        deferred=deferred,
+        deferred_group=deferred_group,
+        deferred_raiseload=deferred_raiseload,
+        sort_order=sort_order,
+        **kw,
+    )
+
+
+def orm_insert_sentinel(
+    name: Optional[str] = None,
+    type_: Optional[_TypeEngineArgument[Any]] = None,
+    *,
+    default: Optional[Any] = None,
+    omit_from_statements: bool = True,
+) -> MappedColumn[Any]:
+    """Provides a surrogate :func:`_orm.mapped_column` that generates
+    a so-called :term:`sentinel` column, allowing efficient bulk
+    inserts with deterministic RETURNING sorting for tables that don't
+    otherwise have qualifying primary key configurations.
+
+    Use of :func:`_orm.orm_insert_sentinel` is analogous to the use of the
+    :func:`_schema.insert_sentinel` construct within a Core
+    :class:`_schema.Table` construct.
+
+    Guidelines for adding this construct to a Declarative mapped class
+    are the same as that of the :func:`_schema.insert_sentinel` construct;
+    the database table itself also needs to have a column with this name
+    present.
+
+    For background on how this object is used, see the section
+    :ref:`engine_insertmanyvalues_sentinel_columns` as part of the
+    section :ref:`engine_insertmanyvalues`.
+
+    .. seealso::
+
+        :func:`_schema.insert_sentinel`
+
+        :ref:`engine_insertmanyvalues`
+
+        :ref:`engine_insertmanyvalues_sentinel_columns`
+
+
+    .. versionadded:: 2.0.10
+
+    """
+
+    return mapped_column(
+        name=name,
+        default=(
+            default if default is not None else _InsertSentinelColumnDefault()
+        ),
+        _omit_from_statements=omit_from_statements,
+        insert_sentinel=True,
+        use_existing_column=True,
+        nullable=True,
+    )
+
+
+@util.deprecated_params(
+    **{
+        arg: (
+            "2.0",
+            f"The :paramref:`_orm.column_property.{arg}` parameter is "
+            "deprecated for :func:`_orm.column_property`.  This parameter "
+            "applies to a writeable-attribute in a Declarative Dataclasses "
+            "configuration only, and :func:`_orm.column_property` is treated "
+            "as a read-only attribute in this context.",
+        )
+        for arg in ("init", "kw_only", "default", "default_factory")
+    }
+)
+def column_property(
+    column: _ORMColumnExprArgument[_T],
+    *additional_columns: _ORMColumnExprArgument[Any],
+    group: Optional[str] = None,
+    deferred: bool = False,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[PropComparator[_T]]] = None,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    active_history: bool = False,
+    expire_on_flush: bool = True,
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+) -> MappedSQLExpression[_T]:
+    r"""Provide a column-level property for use with a mapping.
+
+    With Declarative mappings, :func:`_orm.column_property` is used to
+    map read-only SQL expressions to a mapped class.
+
+    When using Imperative mappings, :func:`_orm.column_property` also
+    takes on the role of mapping table columns with additional features.
+    When using fully Declarative mappings, the :func:`_orm.mapped_column`
+    construct should be used for this purpose.
+
+    With Declarative Dataclass mappings, :func:`_orm.column_property`
+    is considered to be **read only**, and will not be included in the
+    Dataclass ``__init__()`` constructor.
+
+    The :func:`_orm.column_property` function returns an instance of
+    :class:`.ColumnProperty`.
+
+    .. seealso::
+
+        :ref:`mapper_column_property_sql_expressions` - general use of
+        :func:`_orm.column_property` to map SQL expressions
+
+        :ref:`orm_imperative_table_column_options` - usage of
+        :func:`_orm.column_property` with Imperative Table mappings to apply
+        additional options to a plain :class:`_schema.Column` object
+
+    :param \*cols:
+        list of Column objects to be mapped.
+
+    :param active_history=False:
+
+        Used only for Imperative Table mappings, or legacy-style Declarative
+        mappings (i.e. which have not been upgraded to
+        :func:`_orm.mapped_column`), for column-based attributes that are
+        expected to be writeable; use :func:`_orm.mapped_column` with
+        :paramref:`_orm.mapped_column.active_history` for Declarative mappings.
+        See that parameter for functional details.
+
+    :param comparator_factory: a class which extends
+        :class:`.ColumnProperty.Comparator` which provides custom SQL
+        clause generation for comparison operations.
+
+    :param group:
+        a group name for this property when marked as deferred.
+
+    :param deferred:
+        when True, the column property is "deferred", meaning that
+        it does not load immediately, and is instead loaded when the
+        attribute is first accessed on an instance.  See also
+        :func:`~sqlalchemy.orm.deferred`.
+
+    :param doc:
+        optional string that will be applied as the doc on the
+        class-bound descriptor.
+
+    :param expire_on_flush=True:
+        Disable expiry on flush.   A column_property() which refers
+        to a SQL expression (and not a single table-bound column)
+        is considered to be a "read only" property; populating it
+        has no effect on the state of data, and it can only return
+        database state.   For this reason a column_property()'s value
+        is expired whenever the parent object is involved in a
+        flush, that is, has any kind of "dirty" state within a flush.
+        Setting this parameter to ``False`` will have the effect of
+        leaving any existing value present after the flush proceeds.
+        Note that the :class:`.Session` with default expiration
+        settings still expires
+        all attributes after a :meth:`.Session.commit` call, however.
+
+    :param info: Optional data dictionary which will be populated into the
+        :attr:`.MapperProperty.info` attribute of this object.
+
+    :param raiseload: if True, indicates the column should raise an error
+        when undeferred, rather than loading the value.  This can be
+        altered at query time by using the :func:`.deferred` option with
+        raiseload=False.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :ref:`orm_queryguide_deferred_raiseload`
+
+    :param init: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__init__()``
+     method as generated by the dataclass process.
+    :param repr: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__repr__()``
+     method as generated by the dataclass process.
+    :param default_factory: Specific to
+     :ref:`orm_declarative_native_dataclasses`,
+     specifies a default-value generation function that will take place
+     as part of the ``__init__()``
+     method as generated by the dataclass process.
+
+     .. seealso::
+
+        :ref:`defaults_default_factory_insert_default`
+
+        :paramref:`_orm.mapped_column.default`
+
+        :paramref:`_orm.mapped_column.insert_default`
+
+    :param compare: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be included in comparison operations when generating the
+     ``__eq__()`` and ``__ne__()`` methods for the mapped class.
+
+     .. versionadded:: 2.0.0b4
+
+    :param kw_only: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be marked as keyword-only when generating the ``__init__()``.
+
+    :param hash: Specific to
+     :ref:`orm_declarative_native_dataclasses`, controls if this field
+     is included when generating the ``__hash__()`` method for the mapped
+     class.
+
+     .. versionadded:: 2.0.36
+
+    """
+    return MappedSQLExpression(
+        column,
+        *additional_columns,
+        attribute_options=_AttributeOptions(
+            False if init is _NoArg.NO_ARG else init,
+            repr,
+            default,
+            default_factory,
+            compare,
+            kw_only,
+            hash,
+        ),
+        group=group,
+        deferred=deferred,
+        raiseload=raiseload,
+        comparator_factory=comparator_factory,
+        active_history=active_history,
+        expire_on_flush=expire_on_flush,
+        info=info,
+        doc=doc,
+        _assume_readonly_dc_attributes=True,
+    )
+
+
+@overload
+def composite(
+    _class_or_attr: _CompositeAttrType[Any],
+    *attrs: _CompositeAttrType[Any],
+    group: Optional[str] = None,
+    deferred: bool = False,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[Composite.Comparator[_T]]] = None,
+    active_history: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+    **__kw: Any,
+) -> Composite[Any]: ...
+
+
+@overload
+def composite(
+    _class_or_attr: Type[_CC],
+    *attrs: _CompositeAttrType[Any],
+    group: Optional[str] = None,
+    deferred: bool = False,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[Composite.Comparator[_T]]] = None,
+    active_history: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+    **__kw: Any,
+) -> Composite[_CC]: ...
+
+
+@overload
+def composite(
+    _class_or_attr: Callable[..., _CC],
+    *attrs: _CompositeAttrType[Any],
+    group: Optional[str] = None,
+    deferred: bool = False,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[Composite.Comparator[_T]]] = None,
+    active_history: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+    **__kw: Any,
+) -> Composite[_CC]: ...
+
+
+def composite(
+    _class_or_attr: Union[
+        None, Type[_CC], Callable[..., _CC], _CompositeAttrType[Any]
+    ] = None,
+    *attrs: _CompositeAttrType[Any],
+    group: Optional[str] = None,
+    deferred: bool = False,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[Composite.Comparator[_T]]] = None,
+    active_history: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+    **__kw: Any,
+) -> Composite[Any]:
+    r"""Return a composite column-based property for use with a Mapper.
+
+    See the mapping documentation section :ref:`mapper_composite` for a
+    full usage example.
+
+    The :class:`.MapperProperty` returned by :func:`.composite`
+    is the :class:`.Composite`.
+
+    :param class\_:
+      The "composite type" class, or any classmethod or callable which
+      will produce a new instance of the composite object given the
+      column values in order.
+
+    :param \*attrs:
+      List of elements to be mapped, which may include:
+
+      * :class:`_schema.Column` objects
+      * :func:`_orm.mapped_column` constructs
+      * string names of other attributes on the mapped class, which may be
+        any other SQL or object-mapped attribute.  This can for
+        example allow a composite that refers to a many-to-one relationship
+
+    :param active_history=False:
+      When ``True``, indicates that the "previous" value for a
+      scalar attribute should be loaded when replaced, if not
+      already loaded.  See the same flag on :func:`.column_property`.
+
+    :param group:
+      A group name for this property when marked as deferred.
+
+    :param deferred:
+      When True, the column property is "deferred", meaning that it does
+      not load immediately, and is instead loaded when the attribute is
+      first accessed on an instance.  See also
+      :func:`~sqlalchemy.orm.deferred`.
+
+    :param comparator_factory:  a class which extends
+      :class:`.Composite.Comparator` which provides custom SQL
+      clause generation for comparison operations.
+
+    :param doc:
+      optional string that will be applied as the doc on the
+      class-bound descriptor.
+
+    :param info: Optional data dictionary which will be populated into the
+        :attr:`.MapperProperty.info` attribute of this object.
+
+    :param init: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__init__()``
+     method as generated by the dataclass process.
+    :param repr: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__repr__()``
+     method as generated by the dataclass process.
+    :param default_factory: Specific to
+     :ref:`orm_declarative_native_dataclasses`,
+     specifies a default-value generation function that will take place
+     as part of the ``__init__()``
+     method as generated by the dataclass process.
+
+    :param compare: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be included in comparison operations when generating the
+     ``__eq__()`` and ``__ne__()`` methods for the mapped class.
+
+     .. versionadded:: 2.0.0b4
+
+    :param kw_only: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be marked as keyword-only when generating the ``__init__()``.
+
+    :param hash: Specific to
+     :ref:`orm_declarative_native_dataclasses`, controls if this field
+     is included when generating the ``__hash__()`` method for the mapped
+     class.
+
+     .. versionadded:: 2.0.36
+    """
+    if __kw:
+        raise _no_kw()
+
+    return Composite(
+        _class_or_attr,
+        *attrs,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+        group=group,
+        deferred=deferred,
+        raiseload=raiseload,
+        comparator_factory=comparator_factory,
+        active_history=active_history,
+        info=info,
+        doc=doc,
+    )
+
+
+def with_loader_criteria(
+    entity_or_base: _EntityType[Any],
+    where_criteria: Union[
+        _ColumnExpressionArgument[bool],
+        Callable[[Any], _ColumnExpressionArgument[bool]],
+    ],
+    loader_only: bool = False,
+    include_aliases: bool = False,
+    propagate_to_loaders: bool = True,
+    track_closure_variables: bool = True,
+) -> LoaderCriteriaOption:
+    """Add additional WHERE criteria to the load for all occurrences of
+    a particular entity.
+
+    .. versionadded:: 1.4
+
+    The :func:`_orm.with_loader_criteria` option is intended to add
+    limiting criteria to a particular kind of entity in a query,
+    **globally**, meaning it will apply to the entity as it appears
+    in the SELECT query as well as within any subqueries, join
+    conditions, and relationship loads, including both eager and lazy
+    loaders, without the need for it to be specified in any particular
+    part of the query.    The rendering logic uses the same system used by
+    single table inheritance to ensure a certain discriminator is applied
+    to a table.
+
+    E.g., using :term:`2.0-style` queries, we can limit the way the
+    ``User.addresses`` collection is loaded, regardless of the kind
+    of loading used::
+
+        from sqlalchemy.orm import with_loader_criteria
+
+        stmt = select(User).options(
+            selectinload(User.addresses),
+            with_loader_criteria(Address, Address.email_address != "foo"),
+        )
+
+    Above, the "selectinload" for ``User.addresses`` will apply the
+    given filtering criteria to the WHERE clause.
+
+    Another example, where the filtering will be applied to the
+    ON clause of the join, in this example using :term:`1.x style`
+    queries::
+
+        q = (
+            session.query(User)
+            .outerjoin(User.addresses)
+            .options(with_loader_criteria(Address, Address.email_address != "foo"))
+        )
+
+    The primary purpose of :func:`_orm.with_loader_criteria` is to use
+    it in the :meth:`_orm.SessionEvents.do_orm_execute` event handler
+    to ensure that all occurrences of a particular entity are filtered
+    in a certain way, such as filtering for access control roles.    It
+    also can be used to apply criteria to relationship loads.  In the
+    example below, we can apply a certain set of rules to all queries
+    emitted by a particular :class:`_orm.Session`::
+
+        session = Session(bind=engine)
+
+
+        @event.listens_for("do_orm_execute", session)
+        def _add_filtering_criteria(execute_state):
+
+            if (
+                execute_state.is_select
+                and not execute_state.is_column_load
+                and not execute_state.is_relationship_load
+            ):
+                execute_state.statement = execute_state.statement.options(
+                    with_loader_criteria(
+                        SecurityRole,
+                        lambda cls: cls.role.in_(["some_role"]),
+                        include_aliases=True,
+                    )
+                )
+
+    In the above example, the :meth:`_orm.SessionEvents.do_orm_execute`
+    event will intercept all queries emitted using the
+    :class:`_orm.Session`. For those queries which are SELECT statements
+    and are not attribute or relationship loads a custom
+    :func:`_orm.with_loader_criteria` option is added to the query.    The
+    :func:`_orm.with_loader_criteria` option will be used in the given
+    statement and will also be automatically propagated to all relationship
+    loads that descend from this query.
+
+    The criteria argument given is a ``lambda`` that accepts a ``cls``
+    argument.  The given class will expand to include all mapped subclass
+    and need not itself be a mapped class.
+
+    .. tip::
+
+       When using :func:`_orm.with_loader_criteria` option in
+       conjunction with the :func:`_orm.contains_eager` loader option,
+       it's important to note that :func:`_orm.with_loader_criteria` only
+       affects the part of the query that determines what SQL is rendered
+       in terms of the WHERE and FROM clauses. The
+       :func:`_orm.contains_eager` option does not affect the rendering of
+       the SELECT statement outside of the columns clause, so does not have
+       any interaction with the :func:`_orm.with_loader_criteria` option.
+       However, the way things "work" is that :func:`_orm.contains_eager`
+       is meant to be used with a query that is already selecting from the
+       additional entities in some way, where
+       :func:`_orm.with_loader_criteria` can apply it's additional
+       criteria.
+
+       In the example below, assuming a mapping relationship as
+       ``A -> A.bs -> B``, the given :func:`_orm.with_loader_criteria`
+       option will affect the way in which the JOIN is rendered::
+
+            stmt = (
+                select(A)
+                .join(A.bs)
+                .options(contains_eager(A.bs), with_loader_criteria(B, B.flag == 1))
+            )
+
+       Above, the given :func:`_orm.with_loader_criteria` option will
+       affect the ON clause of the JOIN that is specified by
+       ``.join(A.bs)``, so is applied as expected. The
+       :func:`_orm.contains_eager` option has the effect that columns from
+       ``B`` are added to the columns clause:
+
+       .. sourcecode:: sql
+
+            SELECT
+                b.id, b.a_id, b.data, b.flag,
+                a.id AS id_1,
+                a.data AS data_1
+            FROM a JOIN b ON a.id = b.a_id AND b.flag = :flag_1
+
+
+       The use of the :func:`_orm.contains_eager` option within the above
+       statement has no effect on the behavior of the
+       :func:`_orm.with_loader_criteria` option. If the
+       :func:`_orm.contains_eager` option were omitted, the SQL would be
+       the same as regards the FROM and WHERE clauses, where
+       :func:`_orm.with_loader_criteria` continues to add its criteria to
+       the ON clause of the JOIN. The addition of
+       :func:`_orm.contains_eager` only affects the columns clause, in that
+       additional columns against ``b`` are added which are then consumed
+       by the ORM to produce ``B`` instances.
+
+    .. warning:: The use of a lambda inside of the call to
+      :func:`_orm.with_loader_criteria` is only invoked **once per unique
+      class**. Custom functions should not be invoked within this lambda.
+      See :ref:`engine_lambda_caching` for an overview of the "lambda SQL"
+      feature, which is for advanced use only.
+
+    :param entity_or_base: a mapped class, or a class that is a super
+     class of a particular set of mapped classes, to which the rule
+     will apply.
+
+    :param where_criteria: a Core SQL expression that applies limiting
+     criteria.   This may also be a "lambda:" or Python function that
+     accepts a target class as an argument, when the given class is
+     a base with many different mapped subclasses.
+
+     .. note:: To support pickling, use a module-level Python function to
+        produce the SQL expression instead of a lambda or a fixed SQL
+        expression, which tend to not be picklable.
+
+    :param include_aliases: if True, apply the rule to :func:`_orm.aliased`
+     constructs as well.
+
+    :param propagate_to_loaders: defaults to True, apply to relationship
+     loaders such as lazy loaders.   This indicates that the
+     option object itself including SQL expression is carried along with
+     each loaded instance.  Set to ``False`` to prevent the object from
+     being assigned to individual instances.
+
+
+     .. seealso::
+
+        :ref:`examples_session_orm_events` - includes examples of using
+        :func:`_orm.with_loader_criteria`.
+
+        :ref:`do_orm_execute_global_criteria` - basic example on how to
+        combine :func:`_orm.with_loader_criteria` with the
+        :meth:`_orm.SessionEvents.do_orm_execute` event.
+
+    :param track_closure_variables: when False, closure variables inside
+     of a lambda expression will not be used as part of
+     any cache key.    This allows more complex expressions to be used
+     inside of a lambda expression but requires that the lambda ensures
+     it returns the identical SQL every time given a particular class.
+
+     .. versionadded:: 1.4.0b2
+
+    """  # noqa: E501
+    return LoaderCriteriaOption(
+        entity_or_base,
+        where_criteria,
+        loader_only,
+        include_aliases,
+        propagate_to_loaders,
+        track_closure_variables,
+    )
+
+
+def relationship(
+    argument: Optional[_RelationshipArgumentType[Any]] = None,
+    secondary: Optional[_RelationshipSecondaryArgument] = None,
+    *,
+    uselist: Optional[bool] = None,
+    collection_class: Optional[
+        Union[Type[Collection[Any]], Callable[[], Collection[Any]]]
+    ] = None,
+    primaryjoin: Optional[_RelationshipJoinConditionArgument] = None,
+    secondaryjoin: Optional[_RelationshipJoinConditionArgument] = None,
+    back_populates: Optional[str] = None,
+    order_by: _ORMOrderByArgument = False,
+    backref: Optional[ORMBackrefArgument] = None,
+    overlaps: Optional[str] = None,
+    post_update: bool = False,
+    cascade: str = "save-update, merge",
+    viewonly: bool = False,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Union[_NoArg, _T] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    lazy: _LazyLoadArgumentType = "select",
+    passive_deletes: Union[Literal["all"], bool] = False,
+    passive_updates: bool = True,
+    active_history: bool = False,
+    enable_typechecks: bool = True,
+    foreign_keys: Optional[_ORMColCollectionArgument] = None,
+    remote_side: Optional[_ORMColCollectionArgument] = None,
+    join_depth: Optional[int] = None,
+    comparator_factory: Optional[
+        Type[RelationshipProperty.Comparator[Any]]
+    ] = None,
+    single_parent: bool = False,
+    innerjoin: bool = False,
+    distinct_target_key: Optional[bool] = None,
+    load_on_pending: bool = False,
+    query_class: Optional[Type[Query[Any]]] = None,
+    info: Optional[_InfoType] = None,
+    omit_join: Literal[None, False] = None,
+    sync_backref: Optional[bool] = None,
+    **kw: Any,
+) -> _RelationshipDeclared[Any]:
+    """Provide a relationship between two mapped classes.
+
+    This corresponds to a parent-child or associative table relationship.
+    The constructed class is an instance of :class:`.Relationship`.
+
+    .. seealso::
+
+        :ref:`tutorial_orm_related_objects` - tutorial introduction
+        to :func:`_orm.relationship` in the :ref:`unified_tutorial`
+
+        :ref:`relationship_config_toplevel` - narrative documentation
+
+    :param argument:
+      This parameter refers to the class that is to be related.   It
+      accepts several forms, including a direct reference to the target
+      class itself, the :class:`_orm.Mapper` instance for the target class,
+      a Python callable / lambda that will return a reference to the
+      class or :class:`_orm.Mapper` when called, and finally a string
+      name for the class, which will be resolved from the
+      :class:`_orm.registry` in use in order to locate the class, e.g.::
+
+            class SomeClass(Base):
+                # ...
+
+                related = relationship("RelatedClass")
+
+      The :paramref:`_orm.relationship.argument` may also be omitted from the
+      :func:`_orm.relationship` construct entirely, and instead placed inside
+      a :class:`_orm.Mapped` annotation on the left side, which should
+      include a Python collection type if the relationship is expected
+      to be a collection, such as::
+
+            class SomeClass(Base):
+                # ...
+
+                related_items: Mapped[List["RelatedItem"]] = relationship()
+
+      Or for a many-to-one or one-to-one relationship::
+
+            class SomeClass(Base):
+                # ...
+
+                related_item: Mapped["RelatedItem"] = relationship()
+
+      .. seealso::
+
+        :ref:`orm_declarative_properties` - further detail
+        on relationship configuration when using Declarative.
+
+    :param secondary:
+      For a many-to-many relationship, specifies the intermediary
+      table, and is typically an instance of :class:`_schema.Table`.
+      In less common circumstances, the argument may also be specified
+      as an :class:`_expression.Alias` construct, or even a
+      :class:`_expression.Join` construct.
+
+      :paramref:`_orm.relationship.secondary` may
+      also be passed as a callable function which is evaluated at
+      mapper initialization time.  When using Declarative, it may also
+      be a string argument noting the name of a :class:`_schema.Table`
+      that is
+      present in the :class:`_schema.MetaData`
+      collection associated with the
+      parent-mapped :class:`_schema.Table`.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+      The :paramref:`_orm.relationship.secondary` keyword argument is
+      typically applied in the case where the intermediary
+      :class:`_schema.Table`
+      is not otherwise expressed in any direct class mapping. If the
+      "secondary" table is also explicitly mapped elsewhere (e.g. as in
+      :ref:`association_pattern`), one should consider applying the
+      :paramref:`_orm.relationship.viewonly` flag so that this
+      :func:`_orm.relationship`
+      is not used for persistence operations which
+      may conflict with those of the association object pattern.
+
+      .. seealso::
+
+          :ref:`relationships_many_to_many` - Reference example of "many
+          to many".
+
+          :ref:`self_referential_many_to_many` - Specifics on using
+          many-to-many in a self-referential case.
+
+          :ref:`declarative_many_to_many` - Additional options when using
+          Declarative.
+
+          :ref:`association_pattern` - an alternative to
+          :paramref:`_orm.relationship.secondary`
+          when composing association
+          table relationships, allowing additional attributes to be
+          specified on the association table.
+
+          :ref:`composite_secondary_join` - a lesser-used pattern which
+          in some cases can enable complex :func:`_orm.relationship` SQL
+          conditions to be used.
+
+    :param active_history=False:
+      When ``True``, indicates that the "previous" value for a
+      many-to-one reference should be loaded when replaced, if
+      not already loaded. Normally, history tracking logic for
+      simple many-to-ones only needs to be aware of the "new"
+      value in order to perform a flush. This flag is available
+      for applications that make use of
+      :func:`.attributes.get_history` which also need to know
+      the "previous" value of the attribute.
+
+    :param backref:
+      A reference to a string relationship name, or a :func:`_orm.backref`
+      construct, which will be used to automatically generate a new
+      :func:`_orm.relationship` on the related class, which then refers to this
+      one using a bi-directional :paramref:`_orm.relationship.back_populates`
+      configuration.
+
+      In modern Python, explicit use of :func:`_orm.relationship`
+      with :paramref:`_orm.relationship.back_populates` should be preferred,
+      as it is more robust in terms of mapper configuration as well as
+      more conceptually straightforward.  It also integrates with
+      new :pep:`484` typing features introduced in SQLAlchemy 2.0 which
+      is not possible with dynamically generated attributes.
+
+      .. seealso::
+
+        :ref:`relationships_backref` - notes on using
+        :paramref:`_orm.relationship.backref`
+
+        :ref:`tutorial_orm_related_objects` - in the :ref:`unified_tutorial`,
+        presents an overview of bi-directional relationship configuration
+        and behaviors using :paramref:`_orm.relationship.back_populates`
+
+        :func:`.backref` - allows control over :func:`_orm.relationship`
+        configuration when using :paramref:`_orm.relationship.backref`.
+
+
+    :param back_populates:
+      Indicates the name of a :func:`_orm.relationship` on the related
+      class that will be synchronized with this one.   It is usually
+      expected that the :func:`_orm.relationship` on the related class
+      also refer to this one.  This allows objects on both sides of
+      each :func:`_orm.relationship` to synchronize in-Python state
+      changes and also provides directives to the :term:`unit of work`
+      flush process how changes along these relationships should
+      be persisted.
+
+      .. seealso::
+
+        :ref:`tutorial_orm_related_objects` - in the :ref:`unified_tutorial`,
+        presents an overview of bi-directional relationship configuration
+        and behaviors.
+
+        :ref:`relationship_patterns` - includes many examples of
+        :paramref:`_orm.relationship.back_populates`.
+
+        :paramref:`_orm.relationship.backref` - legacy form which allows
+        more succinct configuration, but does not support explicit typing
+
+    :param overlaps:
+       A string name or comma-delimited set of names of other relationships
+       on either this mapper, a descendant mapper, or a target mapper with
+       which this relationship may write to the same foreign keys upon
+       persistence.   The only effect this has is to eliminate the
+       warning that this relationship will conflict with another upon
+       persistence.   This is used for such relationships that are truly
+       capable of conflicting with each other on write, but the application
+       will ensure that no such conflicts occur.
+
+       .. versionadded:: 1.4
+
+       .. seealso::
+
+            :ref:`error_qzyx` - usage example
+
+    :param cascade:
+      A comma-separated list of cascade rules which determines how
+      Session operations should be "cascaded" from parent to child.
+      This defaults to ``False``, which means the default cascade
+      should be used - this default cascade is ``"save-update, merge"``.
+
+      The available cascades are ``save-update``, ``merge``,
+      ``expunge``, ``delete``, ``delete-orphan``, and ``refresh-expire``.
+      An additional option, ``all`` indicates shorthand for
+      ``"save-update, merge, refresh-expire,
+      expunge, delete"``, and is often used as in ``"all, delete-orphan"``
+      to indicate that related objects should follow along with the
+      parent object in all cases, and be deleted when de-associated.
+
+      .. seealso::
+
+        :ref:`unitofwork_cascades` - Full detail on each of the available
+        cascade options.
+
+    :param cascade_backrefs=False:
+      Legacy; this flag is always False.
+
+      .. versionchanged:: 2.0 "cascade_backrefs" functionality has been
+         removed.
+
+    :param collection_class:
+      A class or callable that returns a new list-holding object. will
+      be used in place of a plain list for storing elements.
+
+      .. seealso::
+
+        :ref:`custom_collections` - Introductory documentation and
+        examples.
+
+    :param comparator_factory:
+      A class which extends :class:`.Relationship.Comparator`
+      which provides custom SQL clause generation for comparison
+      operations.
+
+      .. seealso::
+
+        :class:`.PropComparator` - some detail on redefining comparators
+        at this level.
+
+        :ref:`custom_comparators` - Brief intro to this feature.
+
+
+    :param distinct_target_key=None:
+      Indicate if a "subquery" eager load should apply the DISTINCT
+      keyword to the innermost SELECT statement.  When left as ``None``,
+      the DISTINCT keyword will be applied in those cases when the target
+      columns do not comprise the full primary key of the target table.
+      When set to ``True``, the DISTINCT keyword is applied to the
+      innermost SELECT unconditionally.
+
+      It may be desirable to set this flag to False when the DISTINCT is
+      reducing performance of the innermost subquery beyond that of what
+      duplicate innermost rows may be causing.
+
+      .. seealso::
+
+        :ref:`loading_toplevel` - includes an introduction to subquery
+        eager loading.
+
+    :param doc:
+      Docstring which will be applied to the resulting descriptor.
+
+    :param foreign_keys:
+
+      A list of columns which are to be used as "foreign key"
+      columns, or columns which refer to the value in a remote
+      column, within the context of this :func:`_orm.relationship`
+      object's :paramref:`_orm.relationship.primaryjoin` condition.
+      That is, if the :paramref:`_orm.relationship.primaryjoin`
+      condition of this :func:`_orm.relationship` is ``a.id ==
+      b.a_id``, and the values in ``b.a_id`` are required to be
+      present in ``a.id``, then the "foreign key" column of this
+      :func:`_orm.relationship` is ``b.a_id``.
+
+      In normal cases, the :paramref:`_orm.relationship.foreign_keys`
+      parameter is **not required.** :func:`_orm.relationship` will
+      automatically determine which columns in the
+      :paramref:`_orm.relationship.primaryjoin` condition are to be
+      considered "foreign key" columns based on those
+      :class:`_schema.Column` objects that specify
+      :class:`_schema.ForeignKey`,
+      or are otherwise listed as referencing columns in a
+      :class:`_schema.ForeignKeyConstraint` construct.
+      :paramref:`_orm.relationship.foreign_keys` is only needed when:
+
+        1. There is more than one way to construct a join from the local
+           table to the remote table, as there are multiple foreign key
+           references present.  Setting ``foreign_keys`` will limit the
+           :func:`_orm.relationship`
+           to consider just those columns specified
+           here as "foreign".
+
+        2. The :class:`_schema.Table` being mapped does not actually have
+           :class:`_schema.ForeignKey` or
+           :class:`_schema.ForeignKeyConstraint`
+           constructs present, often because the table
+           was reflected from a database that does not support foreign key
+           reflection (MySQL MyISAM).
+
+        3. The :paramref:`_orm.relationship.primaryjoin`
+           argument is used to
+           construct a non-standard join condition, which makes use of
+           columns or expressions that do not normally refer to their
+           "parent" column, such as a join condition expressed by a
+           complex comparison using a SQL function.
+
+      The :func:`_orm.relationship` construct will raise informative
+      error messages that suggest the use of the
+      :paramref:`_orm.relationship.foreign_keys` parameter when
+      presented with an ambiguous condition.   In typical cases,
+      if :func:`_orm.relationship` doesn't raise any exceptions, the
+      :paramref:`_orm.relationship.foreign_keys` parameter is usually
+      not needed.
+
+      :paramref:`_orm.relationship.foreign_keys` may also be passed as a
+      callable function which is evaluated at mapper initialization time,
+      and may be passed as a Python-evaluable string when using
+      Declarative.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+      .. seealso::
+
+        :ref:`relationship_foreign_keys`
+
+        :ref:`relationship_custom_foreign`
+
+        :func:`.foreign` - allows direct annotation of the "foreign"
+        columns within a :paramref:`_orm.relationship.primaryjoin`
+        condition.
+
+    :param info: Optional data dictionary which will be populated into the
+        :attr:`.MapperProperty.info` attribute of this object.
+
+    :param innerjoin=False:
+      When ``True``, joined eager loads will use an inner join to join
+      against related tables instead of an outer join.  The purpose
+      of this option is generally one of performance, as inner joins
+      generally perform better than outer joins.
+
+      This flag can be set to ``True`` when the relationship references an
+      object via many-to-one using local foreign keys that are not
+      nullable, or when the reference is one-to-one or a collection that
+      is guaranteed to have one or at least one entry.
+
+      The option supports the same "nested" and "unnested" options as
+      that of :paramref:`_orm.joinedload.innerjoin`.  See that flag
+      for details on nested / unnested behaviors.
+
+      .. seealso::
+
+        :paramref:`_orm.joinedload.innerjoin` - the option as specified by
+        loader option, including detail on nesting behavior.
+
+        :ref:`what_kind_of_loading` - Discussion of some details of
+        various loader options.
+
+
+    :param join_depth:
+      When non-``None``, an integer value indicating how many levels
+      deep "eager" loaders should join on a self-referring or cyclical
+      relationship.  The number counts how many times the same Mapper
+      shall be present in the loading condition along a particular join
+      branch.  When left at its default of ``None``, eager loaders
+      will stop chaining when they encounter a the same target mapper
+      which is already higher up in the chain.  This option applies
+      both to joined- and subquery- eager loaders.
+
+      .. seealso::
+
+        :ref:`self_referential_eager_loading` - Introductory documentation
+        and examples.
+
+    :param lazy='select': specifies
+      How the related items should be loaded.  Default value is
+      ``select``.  Values include:
+
+      * ``select`` - items should be loaded lazily when the property is
+        first accessed, using a separate SELECT statement, or identity map
+        fetch for simple many-to-one references.
+
+      * ``immediate`` - items should be loaded as the parents are loaded,
+        using a separate SELECT statement, or identity map fetch for
+        simple many-to-one references.
+
+      * ``joined`` - items should be loaded "eagerly" in the same query as
+        that of the parent, using a JOIN or LEFT OUTER JOIN.  Whether
+        the join is "outer" or not is determined by the
+        :paramref:`_orm.relationship.innerjoin` parameter.
+
+      * ``subquery`` - items should be loaded "eagerly" as the parents are
+        loaded, using one additional SQL statement, which issues a JOIN to
+        a subquery of the original statement, for each collection
+        requested.
+
+      * ``selectin`` - items should be loaded "eagerly" as the parents
+        are loaded, using one or more additional SQL statements, which
+        issues a JOIN to the immediate parent object, specifying primary
+        key identifiers using an IN clause.
+
+      * ``noload`` - no loading should occur at any time.  The related
+        collection will remain empty.   The ``noload`` strategy is not
+        recommended for general use.  For a general use "never load"
+        approach, see :ref:`write_only_relationship`
+
+      * ``raise`` - lazy loading is disallowed; accessing
+        the attribute, if its value were not already loaded via eager
+        loading, will raise an :exc:`~sqlalchemy.exc.InvalidRequestError`.
+        This strategy can be used when objects are to be detached from
+        their attached :class:`.Session` after they are loaded.
+
+      * ``raise_on_sql`` - lazy loading that emits SQL is disallowed;
+        accessing the attribute, if its value were not already loaded via
+        eager loading, will raise an
+        :exc:`~sqlalchemy.exc.InvalidRequestError`, **if the lazy load
+        needs to emit SQL**.  If the lazy load can pull the related value
+        from the identity map or determine that it should be None, the
+        value is loaded.  This strategy can be used when objects will
+        remain associated with the attached :class:`.Session`, however
+        additional SELECT statements should be blocked.
+
+      * ``write_only`` - the attribute will be configured with a special
+        "virtual collection" that may receive
+        :meth:`_orm.WriteOnlyCollection.add` and
+        :meth:`_orm.WriteOnlyCollection.remove` commands to add or remove
+        individual objects, but will not under any circumstances load or
+        iterate the full set of objects from the database directly. Instead,
+        methods such as :meth:`_orm.WriteOnlyCollection.select`,
+        :meth:`_orm.WriteOnlyCollection.insert`,
+        :meth:`_orm.WriteOnlyCollection.update` and
+        :meth:`_orm.WriteOnlyCollection.delete` are provided which generate SQL
+        constructs that may be used to load and modify rows in bulk. Used for
+        large collections that are never appropriate to load at once into
+        memory.
+
+        The ``write_only`` loader style is configured automatically when
+        the :class:`_orm.WriteOnlyMapped` annotation is provided on the
+        left hand side within a Declarative mapping.  See the section
+        :ref:`write_only_relationship` for examples.
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :ref:`write_only_relationship` - in the :ref:`queryguide_toplevel`
+
+      * ``dynamic`` - the attribute will return a pre-configured
+        :class:`_query.Query` object for all read
+        operations, onto which further filtering operations can be
+        applied before iterating the results.
+
+        The ``dynamic`` loader style is configured automatically when
+        the :class:`_orm.DynamicMapped` annotation is provided on the
+        left hand side within a Declarative mapping.  See the section
+        :ref:`dynamic_relationship` for examples.
+
+        .. legacy::  The "dynamic" lazy loader strategy is the legacy form of
+           what is now the "write_only" strategy described in the section
+           :ref:`write_only_relationship`.
+
+        .. seealso::
+
+            :ref:`dynamic_relationship` - in the :ref:`queryguide_toplevel`
+
+            :ref:`write_only_relationship` - more generally useful approach
+            for large collections that should not fully load into memory
+
+      * True - a synonym for 'select'
+
+      * False - a synonym for 'joined'
+
+      * None - a synonym for 'noload'
+
+      .. seealso::
+
+        :ref:`orm_queryguide_relationship_loaders` - Full documentation on
+        relationship loader configuration in the :ref:`queryguide_toplevel`.
+
+
+    :param load_on_pending=False:
+      Indicates loading behavior for transient or pending parent objects.
+
+      When set to ``True``, causes the lazy-loader to
+      issue a query for a parent object that is not persistent, meaning it
+      has never been flushed.  This may take effect for a pending object
+      when autoflush is disabled, or for a transient object that has been
+      "attached" to a :class:`.Session` but is not part of its pending
+      collection.
+
+      The :paramref:`_orm.relationship.load_on_pending`
+      flag does not improve
+      behavior when the ORM is used normally - object references should be
+      constructed at the object level, not at the foreign key level, so
+      that they are present in an ordinary way before a flush proceeds.
+      This flag is not not intended for general use.
+
+      .. seealso::
+
+          :meth:`.Session.enable_relationship_loading` - this method
+          establishes "load on pending" behavior for the whole object, and
+          also allows loading on objects that remain transient or
+          detached.
+
+    :param order_by:
+      Indicates the ordering that should be applied when loading these
+      items.  :paramref:`_orm.relationship.order_by`
+      is expected to refer to
+      one of the :class:`_schema.Column`
+      objects to which the target class is
+      mapped, or the attribute itself bound to the target class which
+      refers to the column.
+
+      :paramref:`_orm.relationship.order_by`
+      may also be passed as a callable
+      function which is evaluated at mapper initialization time, and may
+      be passed as a Python-evaluable string when using Declarative.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+    :param passive_deletes=False:
+       Indicates loading behavior during delete operations.
+
+       A value of True indicates that unloaded child items should not
+       be loaded during a delete operation on the parent.  Normally,
+       when a parent item is deleted, all child items are loaded so
+       that they can either be marked as deleted, or have their
+       foreign key to the parent set to NULL.  Marking this flag as
+       True usually implies an ON DELETE <CASCADE|SET NULL> rule is in
+       place which will handle updating/deleting child rows on the
+       database side.
+
+       Additionally, setting the flag to the string value 'all' will
+       disable the "nulling out" of the child foreign keys, when the parent
+       object is deleted and there is no delete or delete-orphan cascade
+       enabled.  This is typically used when a triggering or error raise
+       scenario is in place on the database side.  Note that the foreign
+       key attributes on in-session child objects will not be changed after
+       a flush occurs so this is a very special use-case setting.
+       Additionally, the "nulling out" will still occur if the child
+       object is de-associated with the parent.
+
+       .. seealso::
+
+            :ref:`passive_deletes` - Introductory documentation
+            and examples.
+
+    :param passive_updates=True:
+      Indicates the persistence behavior to take when a referenced
+      primary key value changes in place, indicating that the referencing
+      foreign key columns will also need their value changed.
+
+      When True, it is assumed that ``ON UPDATE CASCADE`` is configured on
+      the foreign key in the database, and that the database will
+      handle propagation of an UPDATE from a source column to
+      dependent rows.  When False, the SQLAlchemy
+      :func:`_orm.relationship`
+      construct will attempt to emit its own UPDATE statements to
+      modify related targets.  However note that SQLAlchemy **cannot**
+      emit an UPDATE for more than one level of cascade.  Also,
+      setting this flag to False is not compatible in the case where
+      the database is in fact enforcing referential integrity, unless
+      those constraints are explicitly "deferred", if the target backend
+      supports it.
+
+      It is highly advised that an application which is employing
+      mutable primary keys keeps ``passive_updates`` set to True,
+      and instead uses the referential integrity features of the database
+      itself in order to handle the change efficiently and fully.
+
+      .. seealso::
+
+          :ref:`passive_updates` - Introductory documentation and
+          examples.
+
+          :paramref:`.mapper.passive_updates` - a similar flag which
+          takes effect for joined-table inheritance mappings.
+
+    :param post_update:
+      This indicates that the relationship should be handled by a
+      second UPDATE statement after an INSERT or before a
+      DELETE. This flag is used to handle saving bi-directional
+      dependencies between two individual rows (i.e. each row
+      references the other), where it would otherwise be impossible to
+      INSERT or DELETE both rows fully since one row exists before the
+      other. Use this flag when a particular mapping arrangement will
+      incur two rows that are dependent on each other, such as a table
+      that has a one-to-many relationship to a set of child rows, and
+      also has a column that references a single child row within that
+      list (i.e. both tables contain a foreign key to each other). If
+      a flush operation returns an error that a "cyclical
+      dependency" was detected, this is a cue that you might want to
+      use :paramref:`_orm.relationship.post_update` to "break" the cycle.
+
+      .. seealso::
+
+          :ref:`post_update` - Introductory documentation and examples.
+
+    :param primaryjoin:
+      A SQL expression that will be used as the primary
+      join of the child object against the parent object, or in a
+      many-to-many relationship the join of the parent object to the
+      association table. By default, this value is computed based on the
+      foreign key relationships of the parent and child tables (or
+      association table).
+
+      :paramref:`_orm.relationship.primaryjoin` may also be passed as a
+      callable function which is evaluated at mapper initialization time,
+      and may be passed as a Python-evaluable string when using
+      Declarative.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+      .. seealso::
+
+          :ref:`relationship_primaryjoin`
+
+    :param remote_side:
+      Used for self-referential relationships, indicates the column or
+      list of columns that form the "remote side" of the relationship.
+
+      :paramref:`_orm.relationship.remote_side` may also be passed as a
+      callable function which is evaluated at mapper initialization time,
+      and may be passed as a Python-evaluable string when using
+      Declarative.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+      .. seealso::
+
+        :ref:`self_referential` - in-depth explanation of how
+        :paramref:`_orm.relationship.remote_side`
+        is used to configure self-referential relationships.
+
+        :func:`.remote` - an annotation function that accomplishes the
+        same purpose as :paramref:`_orm.relationship.remote_side`,
+        typically
+        when a custom :paramref:`_orm.relationship.primaryjoin` condition
+        is used.
+
+    :param query_class:
+      A :class:`_query.Query`
+      subclass that will be used internally by the
+      ``AppenderQuery`` returned by a "dynamic" relationship, that
+      is, a relationship that specifies ``lazy="dynamic"`` or was
+      otherwise constructed using the :func:`_orm.dynamic_loader`
+      function.
+
+      .. seealso::
+
+        :ref:`dynamic_relationship` - Introduction to "dynamic"
+        relationship loaders.
+
+    :param secondaryjoin:
+      A SQL expression that will be used as the join of
+      an association table to the child object. By default, this value is
+      computed based on the foreign key relationships of the association
+      and child tables.
+
+      :paramref:`_orm.relationship.secondaryjoin` may also be passed as a
+      callable function which is evaluated at mapper initialization time,
+      and may be passed as a Python-evaluable string when using
+      Declarative.
+
+      .. warning:: When passed as a Python-evaluable string, the
+         argument is interpreted using Python's ``eval()`` function.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+         See :ref:`declarative_relationship_eval` for details on
+         declarative evaluation of :func:`_orm.relationship` arguments.
+
+      .. seealso::
+
+          :ref:`relationship_primaryjoin`
+
+    :param single_parent:
+      When True, installs a validator which will prevent objects
+      from being associated with more than one parent at a time.
+      This is used for many-to-one or many-to-many relationships that
+      should be treated either as one-to-one or one-to-many.  Its usage
+      is optional, except for :func:`_orm.relationship` constructs which
+      are many-to-one or many-to-many and also
+      specify the ``delete-orphan`` cascade option.  The
+      :func:`_orm.relationship` construct itself will raise an error
+      instructing when this option is required.
+
+      .. seealso::
+
+        :ref:`unitofwork_cascades` - includes detail on when the
+        :paramref:`_orm.relationship.single_parent`
+        flag may be appropriate.
+
+    :param uselist:
+      A boolean that indicates if this property should be loaded as a
+      list or a scalar. In most cases, this value is determined
+      automatically by :func:`_orm.relationship` at mapper configuration
+      time.  When using explicit :class:`_orm.Mapped` annotations,
+      :paramref:`_orm.relationship.uselist` may be derived from the
+      whether or not the annotation within :class:`_orm.Mapped` contains
+      a collection class.
+      Otherwise, :paramref:`_orm.relationship.uselist` may be derived from
+      the type and direction
+      of the relationship - one to many forms a list, many to one
+      forms a scalar, many to many is a list. If a scalar is desired
+      where normally a list would be present, such as a bi-directional
+      one-to-one relationship, use an appropriate :class:`_orm.Mapped`
+      annotation or set :paramref:`_orm.relationship.uselist` to False.
+
+      The :paramref:`_orm.relationship.uselist`
+      flag is also available on an
+      existing :func:`_orm.relationship`
+      construct as a read-only attribute,
+      which can be used to determine if this :func:`_orm.relationship`
+      deals
+      with collections or scalar attributes::
+
+          >>> User.addresses.property.uselist
+          True
+
+      .. seealso::
+
+          :ref:`relationships_one_to_one` - Introduction to the "one to
+          one" relationship pattern, which is typically when an alternate
+          setting for :paramref:`_orm.relationship.uselist` is involved.
+
+    :param viewonly=False:
+      When set to ``True``, the relationship is used only for loading
+      objects, and not for any persistence operation.  A
+      :func:`_orm.relationship` which specifies
+      :paramref:`_orm.relationship.viewonly` can work
+      with a wider range of SQL operations within the
+      :paramref:`_orm.relationship.primaryjoin` condition, including
+      operations that feature the use of a variety of comparison operators
+      as well as SQL functions such as :func:`_expression.cast`.  The
+      :paramref:`_orm.relationship.viewonly`
+      flag is also of general use when defining any kind of
+      :func:`_orm.relationship` that doesn't represent
+      the full set of related objects, to prevent modifications of the
+      collection from resulting in persistence operations.
+
+      .. seealso::
+
+        :ref:`relationship_viewonly_notes` - more details on best practices
+        when using :paramref:`_orm.relationship.viewonly`.
+
+    :param sync_backref:
+      A boolean that enables the events used to synchronize the in-Python
+      attributes when this relationship is target of either
+      :paramref:`_orm.relationship.backref` or
+      :paramref:`_orm.relationship.back_populates`.
+
+      Defaults to ``None``, which indicates that an automatic value should
+      be selected based on the value of the
+      :paramref:`_orm.relationship.viewonly` flag.  When left at its
+      default, changes in state will be back-populated only if neither
+      sides of a relationship is viewonly.
+
+      .. versionadded:: 1.3.17
+
+      .. versionchanged:: 1.4 - A relationship that specifies
+         :paramref:`_orm.relationship.viewonly` automatically implies
+         that :paramref:`_orm.relationship.sync_backref` is ``False``.
+
+      .. seealso::
+
+        :paramref:`_orm.relationship.viewonly`
+
+    :param omit_join:
+      Allows manual control over the "selectin" automatic join
+      optimization.  Set to ``False`` to disable the "omit join" feature
+      added in SQLAlchemy 1.3; or leave as ``None`` to leave automatic
+      optimization in place.
+
+      .. note:: This flag may only be set to ``False``.   It is not
+         necessary to set it to ``True`` as the "omit_join" optimization is
+         automatically detected; if it is not detected, then the
+         optimization is not supported.
+
+         .. versionchanged:: 1.3.11  setting ``omit_join`` to True will now
+            emit a warning as this was not the intended use of this flag.
+
+      .. versionadded:: 1.3
+
+    :param init: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__init__()``
+     method as generated by the dataclass process.
+    :param repr: Specific to :ref:`orm_declarative_native_dataclasses`,
+     specifies if the mapped attribute should be part of the ``__repr__()``
+     method as generated by the dataclass process.
+    :param default_factory: Specific to
+     :ref:`orm_declarative_native_dataclasses`,
+     specifies a default-value generation function that will take place
+     as part of the ``__init__()``
+     method as generated by the dataclass process.
+    :param compare: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be included in comparison operations when generating the
+     ``__eq__()`` and ``__ne__()`` methods for the mapped class.
+
+     .. versionadded:: 2.0.0b4
+
+    :param kw_only: Specific to
+     :ref:`orm_declarative_native_dataclasses`, indicates if this field
+     should be marked as keyword-only when generating the ``__init__()``.
+
+    :param hash: Specific to
+     :ref:`orm_declarative_native_dataclasses`, controls if this field
+     is included when generating the ``__hash__()`` method for the mapped
+     class.
+
+     .. versionadded:: 2.0.36
+    """
+
+    return _RelationshipDeclared(
+        argument,
+        secondary=secondary,
+        uselist=uselist,
+        collection_class=collection_class,
+        primaryjoin=primaryjoin,
+        secondaryjoin=secondaryjoin,
+        back_populates=back_populates,
+        order_by=order_by,
+        backref=backref,
+        overlaps=overlaps,
+        post_update=post_update,
+        cascade=cascade,
+        viewonly=viewonly,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+        lazy=lazy,
+        passive_deletes=passive_deletes,
+        passive_updates=passive_updates,
+        active_history=active_history,
+        enable_typechecks=enable_typechecks,
+        foreign_keys=foreign_keys,
+        remote_side=remote_side,
+        join_depth=join_depth,
+        comparator_factory=comparator_factory,
+        single_parent=single_parent,
+        innerjoin=innerjoin,
+        distinct_target_key=distinct_target_key,
+        load_on_pending=load_on_pending,
+        query_class=query_class,
+        info=info,
+        omit_join=omit_join,
+        sync_backref=sync_backref,
+        **kw,
+    )
+
+
+def synonym(
+    name: str,
+    *,
+    map_column: Optional[bool] = None,
+    descriptor: Optional[Any] = None,
+    comparator_factory: Optional[Type[PropComparator[_T]]] = None,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Union[_NoArg, _T] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+) -> Synonym[Any]:
+    """Denote an attribute name as a synonym to a mapped property,
+    in that the attribute will mirror the value and expression behavior
+    of another attribute.
+
+    e.g.::
+
+        class MyClass(Base):
+            __tablename__ = "my_table"
+
+            id = Column(Integer, primary_key=True)
+            job_status = Column(String(50))
+
+            status = synonym("job_status")
+
+    :param name: the name of the existing mapped property.  This
+      can refer to the string name ORM-mapped attribute
+      configured on the class, including column-bound attributes
+      and relationships.
+
+    :param descriptor: a Python :term:`descriptor` that will be used
+      as a getter (and potentially a setter) when this attribute is
+      accessed at the instance level.
+
+    :param map_column: **For classical mappings and mappings against
+      an existing Table object only**.  if ``True``, the :func:`.synonym`
+      construct will locate the :class:`_schema.Column`
+      object upon the mapped
+      table that would normally be associated with the attribute name of
+      this synonym, and produce a new :class:`.ColumnProperty` that instead
+      maps this :class:`_schema.Column`
+      to the alternate name given as the "name"
+      argument of the synonym; in this way, the usual step of redefining
+      the mapping of the :class:`_schema.Column`
+      to be under a different name is
+      unnecessary. This is usually intended to be used when a
+      :class:`_schema.Column`
+      is to be replaced with an attribute that also uses a
+      descriptor, that is, in conjunction with the
+      :paramref:`.synonym.descriptor` parameter::
+
+        my_table = Table(
+            "my_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("job_status", String(50)),
+        )
+
+
+        class MyClass:
+            @property
+            def _job_status_descriptor(self):
+                return "Status: %s" % self._job_status
+
+
+        mapper(
+            MyClass,
+            my_table,
+            properties={
+                "job_status": synonym(
+                    "_job_status",
+                    map_column=True,
+                    descriptor=MyClass._job_status_descriptor,
+                )
+            },
+        )
+
+      Above, the attribute named ``_job_status`` is automatically
+      mapped to the ``job_status`` column::
+
+        >>> j1 = MyClass()
+        >>> j1._job_status = "employed"
+        >>> j1.job_status
+        Status: employed
+
+      When using Declarative, in order to provide a descriptor in
+      conjunction with a synonym, use the
+      :func:`sqlalchemy.ext.declarative.synonym_for` helper.  However,
+      note that the :ref:`hybrid properties <mapper_hybrids>` feature
+      should usually be preferred, particularly when redefining attribute
+      behavior.
+
+    :param info: Optional data dictionary which will be populated into the
+        :attr:`.InspectionAttr.info` attribute of this object.
+
+    :param comparator_factory: A subclass of :class:`.PropComparator`
+      that will provide custom comparison behavior at the SQL expression
+      level.
+
+      .. note::
+
+        For the use case of providing an attribute which redefines both
+        Python-level and SQL-expression level behavior of an attribute,
+        please refer to the Hybrid attribute introduced at
+        :ref:`mapper_hybrids` for a more effective technique.
+
+    .. seealso::
+
+        :ref:`synonyms` - Overview of synonyms
+
+        :func:`.synonym_for` - a helper oriented towards Declarative
+
+        :ref:`mapper_hybrids` - The Hybrid Attribute extension provides an
+        updated approach to augmenting attribute behavior more flexibly
+        than can be achieved with synonyms.
+
+    """
+    return Synonym(
+        name,
+        map_column=map_column,
+        descriptor=descriptor,
+        comparator_factory=comparator_factory,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+        doc=doc,
+        info=info,
+    )
+
+
+def create_session(
+    bind: Optional[_SessionBind] = None, **kwargs: Any
+) -> Session:
+    r"""Create a new :class:`.Session`
+    with no automation enabled by default.
+
+    This function is used primarily for testing.   The usual
+    route to :class:`.Session` creation is via its constructor
+    or the :func:`.sessionmaker` function.
+
+    :param bind: optional, a single Connectable to use for all
+      database access in the created
+      :class:`~sqlalchemy.orm.session.Session`.
+
+    :param \*\*kwargs: optional, passed through to the
+      :class:`.Session` constructor.
+
+    :returns: an :class:`~sqlalchemy.orm.session.Session` instance
+
+    The defaults of create_session() are the opposite of that of
+    :func:`sessionmaker`; ``autoflush`` and ``expire_on_commit`` are
+    False.
+
+    Usage::
+
+      >>> from sqlalchemy.orm import create_session
+      >>> session = create_session()
+
+    It is recommended to use :func:`sessionmaker` instead of
+    create_session().
+
+    """
+
+    kwargs.setdefault("autoflush", False)
+    kwargs.setdefault("expire_on_commit", False)
+    return Session(bind=bind, **kwargs)
+
+
+def _mapper_fn(*arg: Any, **kw: Any) -> NoReturn:
+    """Placeholder for the now-removed ``mapper()`` function.
+
+    Classical mappings should be performed using the
+    :meth:`_orm.registry.map_imperatively` method.
+
+    This symbol remains in SQLAlchemy 2.0 to suit the deprecated use case
+    of using the ``mapper()`` function as a target for ORM event listeners,
+    which failed to be marked as deprecated in the 1.4 series.
+
+    Global ORM mapper listeners should instead use the :class:`_orm.Mapper`
+    class as the target.
+
+    .. versionchanged:: 2.0  The ``mapper()`` function was removed; the
+       symbol remains temporarily as a placeholder for the event listening
+       use case.
+
+    """
+    raise InvalidRequestError(
+        "The 'sqlalchemy.orm.mapper()' function is removed as of "
+        "SQLAlchemy 2.0.  Use the "
+        "'sqlalchemy.orm.registry.map_imperatively()` "
+        "method of the ``sqlalchemy.orm.registry`` class to perform "
+        "classical mapping."
+    )
+
+
+def dynamic_loader(
+    argument: Optional[_RelationshipArgumentType[Any]] = None, **kw: Any
+) -> RelationshipProperty[Any]:
+    """Construct a dynamically-loading mapper property.
+
+    This is essentially the same as
+    using the ``lazy='dynamic'`` argument with :func:`relationship`::
+
+        dynamic_loader(SomeClass)
+
+        # is the same as
+
+        relationship(SomeClass, lazy="dynamic")
+
+    See the section :ref:`dynamic_relationship` for more details
+    on dynamic loading.
+
+    """
+    kw["lazy"] = "dynamic"
+    return relationship(argument, **kw)
+
+
+def backref(name: str, **kwargs: Any) -> ORMBackrefArgument:
+    """When using the :paramref:`_orm.relationship.backref` parameter,
+    provides specific parameters to be used when the new
+    :func:`_orm.relationship` is generated.
+
+    E.g.::
+
+        "items": relationship(SomeItem, backref=backref("parent", lazy="subquery"))
+
+    The :paramref:`_orm.relationship.backref` parameter is generally
+    considered to be legacy; for modern applications, using
+    explicit :func:`_orm.relationship` constructs linked together using
+    the :paramref:`_orm.relationship.back_populates` parameter should be
+    preferred.
+
+    .. seealso::
+
+        :ref:`relationships_backref` - background on backrefs
+
+    """  # noqa: E501
+
+    return (name, kwargs)
+
+
+def deferred(
+    column: _ORMColumnExprArgument[_T],
+    *additional_columns: _ORMColumnExprArgument[Any],
+    group: Optional[str] = None,
+    raiseload: bool = False,
+    comparator_factory: Optional[Type[PropComparator[_T]]] = None,
+    init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    default: Optional[Any] = _NoArg.NO_ARG,
+    default_factory: Union[_NoArg, Callable[[], _T]] = _NoArg.NO_ARG,
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+    hash: Union[_NoArg, bool, None] = _NoArg.NO_ARG,  # noqa: A002
+    active_history: bool = False,
+    expire_on_flush: bool = True,
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+) -> MappedSQLExpression[_T]:
+    r"""Indicate a column-based mapped attribute that by default will
+    not load unless accessed.
+
+    When using :func:`_orm.mapped_column`, the same functionality as
+    that of :func:`_orm.deferred` construct is provided by using the
+    :paramref:`_orm.mapped_column.deferred` parameter.
+
+    :param \*columns: columns to be mapped.  This is typically a single
+     :class:`_schema.Column` object,
+     however a collection is supported in order
+     to support multiple columns mapped under the same attribute.
+
+    :param raiseload: boolean, if True, indicates an exception should be raised
+     if the load operation is to take place.
+
+     .. versionadded:: 1.4
+
+
+    Additional arguments are the same as that of :func:`_orm.column_property`.
+
+    .. seealso::
+
+        :ref:`orm_queryguide_deferred_imperative`
+
+    """
+    return MappedSQLExpression(
+        column,
+        *additional_columns,
+        attribute_options=_AttributeOptions(
+            init, repr, default, default_factory, compare, kw_only, hash
+        ),
+        group=group,
+        deferred=True,
+        raiseload=raiseload,
+        comparator_factory=comparator_factory,
+        active_history=active_history,
+        expire_on_flush=expire_on_flush,
+        info=info,
+        doc=doc,
+    )
+
+
+def query_expression(
+    default_expr: _ORMColumnExprArgument[_T] = sql.null(),
+    *,
+    repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    compare: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+    expire_on_flush: bool = True,
+    info: Optional[_InfoType] = None,
+    doc: Optional[str] = None,
+) -> MappedSQLExpression[_T]:
+    """Indicate an attribute that populates from a query-time SQL expression.
+
+    :param default_expr: Optional SQL expression object that will be used in
+        all cases if not assigned later with :func:`_orm.with_expression`.
+
+    .. versionadded:: 1.2
+
+    .. seealso::
+
+        :ref:`orm_queryguide_with_expression` - background and usage examples
+
+    """
+    prop = MappedSQLExpression(
+        default_expr,
+        attribute_options=_AttributeOptions(
+            False,
+            repr,
+            _NoArg.NO_ARG,
+            _NoArg.NO_ARG,
+            compare,
+            _NoArg.NO_ARG,
+            _NoArg.NO_ARG,
+        ),
+        expire_on_flush=expire_on_flush,
+        info=info,
+        doc=doc,
+        _assume_readonly_dc_attributes=True,
+    )
+
+    prop.strategy_key = (("query_expression", True),)
+    return prop
+
+
+def clear_mappers() -> None:
+    """Remove all mappers from all classes.
+
+    .. versionchanged:: 1.4  This function now locates all
+       :class:`_orm.registry` objects and calls upon the
+       :meth:`_orm.registry.dispose` method of each.
+
+    This function removes all instrumentation from classes and disposes
+    of their associated mappers.  Once called, the classes are unmapped
+    and can be later re-mapped with new mappers.
+
+    :func:`.clear_mappers` is *not* for normal use, as there is literally no
+    valid usage for it outside of very specific testing scenarios. Normally,
+    mappers are permanent structural components of user-defined classes, and
+    are never discarded independently of their class.  If a mapped class
+    itself is garbage collected, its mapper is automatically disposed of as
+    well. As such, :func:`.clear_mappers` is only for usage in test suites
+    that re-use the same classes with different mappings, which is itself an
+    extremely rare use case - the only such use case is in fact SQLAlchemy's
+    own test suite, and possibly the test suites of other ORM extension
+    libraries which intend to test various combinations of mapper construction
+    upon a fixed set of classes.
+
+    """
+
+    mapperlib._dispose_registries(mapperlib._all_registries(), False)
+
+
+# I would really like a way to get the Type[] here that shows up
+# in a different way in typing tools, however there is no current method
+# that is accepted by mypy (subclass of Type[_O] works in pylance, rejected
+# by mypy).
+AliasedType = Annotated[Type[_O], "aliased"]
+
+
+@overload
+def aliased(
+    element: Type[_O],
+    alias: Optional[FromClause] = None,
+    name: Optional[str] = None,
+    flat: bool = False,
+    adapt_on_names: bool = False,
+) -> AliasedType[_O]: ...
+
+
+@overload
+def aliased(
+    element: Union[AliasedClass[_O], Mapper[_O], AliasedInsp[_O]],
+    alias: Optional[FromClause] = None,
+    name: Optional[str] = None,
+    flat: bool = False,
+    adapt_on_names: bool = False,
+) -> AliasedClass[_O]: ...
+
+
+@overload
+def aliased(
+    element: FromClause,
+    alias: None = None,
+    name: Optional[str] = None,
+    flat: bool = False,
+    adapt_on_names: bool = False,
+) -> FromClause: ...
+
+
+def aliased(
+    element: Union[_EntityType[_O], FromClause],
+    alias: Optional[FromClause] = None,
+    name: Optional[str] = None,
+    flat: bool = False,
+    adapt_on_names: bool = False,
+) -> Union[AliasedClass[_O], FromClause, AliasedType[_O]]:
+    """Produce an alias of the given element, usually an :class:`.AliasedClass`
+    instance.
+
+    E.g.::
+
+        my_alias = aliased(MyClass)
+
+        stmt = select(MyClass, my_alias).filter(MyClass.id > my_alias.id)
+        result = session.execute(stmt)
+
+    The :func:`.aliased` function is used to create an ad-hoc mapping of a
+    mapped class to a new selectable.  By default, a selectable is generated
+    from the normally mapped selectable (typically a :class:`_schema.Table`
+    ) using the
+    :meth:`_expression.FromClause.alias` method. However, :func:`.aliased`
+    can also be
+    used to link the class to a new :func:`_expression.select` statement.
+    Also, the :func:`.with_polymorphic` function is a variant of
+    :func:`.aliased` that is intended to specify a so-called "polymorphic
+    selectable", that corresponds to the union of several joined-inheritance
+    subclasses at once.
+
+    For convenience, the :func:`.aliased` function also accepts plain
+    :class:`_expression.FromClause` constructs, such as a
+    :class:`_schema.Table` or
+    :func:`_expression.select` construct.   In those cases, the
+    :meth:`_expression.FromClause.alias`
+    method is called on the object and the new
+    :class:`_expression.Alias` object returned.  The returned
+    :class:`_expression.Alias` is not
+    ORM-mapped in this case.
+
+    .. seealso::
+
+        :ref:`tutorial_orm_entity_aliases` - in the :ref:`unified_tutorial`
+
+        :ref:`orm_queryguide_orm_aliases` - in the :ref:`queryguide_toplevel`
+
+    :param element: element to be aliased.  Is normally a mapped class,
+     but for convenience can also be a :class:`_expression.FromClause`
+     element.
+
+    :param alias: Optional selectable unit to map the element to.  This is
+     usually used to link the object to a subquery, and should be an aliased
+     select construct as one would produce from the
+     :meth:`_query.Query.subquery` method or
+     the :meth:`_expression.Select.subquery` or
+     :meth:`_expression.Select.alias` methods of the :func:`_expression.select`
+     construct.
+
+    :param name: optional string name to use for the alias, if not specified
+     by the ``alias`` parameter.  The name, among other things, forms the
+     attribute name that will be accessible via tuples returned by a
+     :class:`_query.Query` object.  Not supported when creating aliases
+     of :class:`_sql.Join` objects.
+
+    :param flat: Boolean, will be passed through to the
+     :meth:`_expression.FromClause.alias` call so that aliases of
+     :class:`_expression.Join` objects will alias the individual tables
+     inside the join, rather than creating a subquery.  This is generally
+     supported by all modern databases with regards to right-nested joins
+     and generally produces more efficient queries.
+
+     When :paramref:`_orm.aliased.flat` is combined with
+     :paramref:`_orm.aliased.name`, the resulting joins will alias individual
+     tables using a naming scheme similar to ``<prefix>_<tablename>``.  This
+     naming scheme is for visibility / debugging purposes only and the
+     specific scheme is subject to change without notice.
+
+     .. versionadded:: 2.0.32 added support for combining
+        :paramref:`_orm.aliased.name` with :paramref:`_orm.aliased.flat`.
+        Previously, this would raise ``NotImplementedError``.
+
+    :param adapt_on_names: if True, more liberal "matching" will be used when
+     mapping the mapped columns of the ORM entity to those of the
+     given selectable - a name-based match will be performed if the
+     given selectable doesn't otherwise have a column that corresponds
+     to one on the entity.  The use case for this is when associating
+     an entity with some derived selectable such as one that uses
+     aggregate functions::
+
+        class UnitPrice(Base):
+            __tablename__ = "unit_price"
+            ...
+            unit_id = Column(Integer)
+            price = Column(Numeric)
+
+
+        aggregated_unit_price = (
+            Session.query(func.sum(UnitPrice.price).label("price"))
+            .group_by(UnitPrice.unit_id)
+            .subquery()
+        )
+
+        aggregated_unit_price = aliased(
+            UnitPrice, alias=aggregated_unit_price, adapt_on_names=True
+        )
+
+     Above, functions on ``aggregated_unit_price`` which refer to
+     ``.price`` will return the
+     ``func.sum(UnitPrice.price).label('price')`` column, as it is
+     matched on the name "price".  Ordinarily, the "price" function
+     wouldn't have any "column correspondence" to the actual
+     ``UnitPrice.price`` column as it is not a proxy of the original.
+
+    """
+    return AliasedInsp._alias_factory(
+        element,
+        alias=alias,
+        name=name,
+        flat=flat,
+        adapt_on_names=adapt_on_names,
+    )
+
+
+def with_polymorphic(
+    base: Union[Type[_O], Mapper[_O]],
+    classes: Union[Literal["*"], Iterable[Type[Any]]],
+    selectable: Union[Literal[False, None], FromClause] = False,
+    flat: bool = False,
+    polymorphic_on: Optional[ColumnElement[Any]] = None,
+    aliased: bool = False,
+    innerjoin: bool = False,
+    adapt_on_names: bool = False,
+    name: Optional[str] = None,
+    _use_mapper_path: bool = False,
+) -> AliasedClass[_O]:
+    """Produce an :class:`.AliasedClass` construct which specifies
+    columns for descendant mappers of the given base.
+
+    Using this method will ensure that each descendant mapper's
+    tables are included in the FROM clause, and will allow filter()
+    criterion to be used against those tables.  The resulting
+    instances will also have those columns already loaded so that
+    no "post fetch" of those columns will be required.
+
+    .. seealso::
+
+        :ref:`with_polymorphic` - full discussion of
+        :func:`_orm.with_polymorphic`.
+
+    :param base: Base class to be aliased.
+
+    :param classes: a single class or mapper, or list of
+        class/mappers, which inherit from the base class.
+        Alternatively, it may also be the string ``'*'``, in which case
+        all descending mapped classes will be added to the FROM clause.
+
+    :param aliased: when True, the selectable will be aliased.   For a
+        JOIN, this means the JOIN will be SELECTed from inside of a subquery
+        unless the :paramref:`_orm.with_polymorphic.flat` flag is set to
+        True, which is recommended for simpler use cases.
+
+    :param flat: Boolean, will be passed through to the
+     :meth:`_expression.FromClause.alias` call so that aliases of
+     :class:`_expression.Join` objects will alias the individual tables
+     inside the join, rather than creating a subquery.  This is generally
+     supported by all modern databases with regards to right-nested joins
+     and generally produces more efficient queries.  Setting this flag is
+     recommended as long as the resulting SQL is functional.
+
+    :param selectable: a table or subquery that will
+        be used in place of the generated FROM clause. This argument is
+        required if any of the desired classes use concrete table
+        inheritance, since SQLAlchemy currently cannot generate UNIONs
+        among tables automatically. If used, the ``selectable`` argument
+        must represent the full set of tables and columns mapped by every
+        mapped class. Otherwise, the unaccounted mapped columns will
+        result in their table being appended directly to the FROM clause
+        which will usually lead to incorrect results.
+
+        When left at its default value of ``False``, the polymorphic
+        selectable assigned to the base mapper is used for selecting rows.
+        However, it may also be passed as ``None``, which will bypass the
+        configured polymorphic selectable and instead construct an ad-hoc
+        selectable for the target classes given; for joined table inheritance
+        this will be a join that includes all target mappers and their
+        subclasses.
+
+    :param polymorphic_on: a column to be used as the "discriminator"
+        column for the given selectable. If not given, the polymorphic_on
+        attribute of the base classes' mapper will be used, if any. This
+        is useful for mappings that don't have polymorphic loading
+        behavior by default.
+
+    :param innerjoin: if True, an INNER JOIN will be used.  This should
+       only be specified if querying for one specific subtype only
+
+    :param adapt_on_names: Passes through the
+      :paramref:`_orm.aliased.adapt_on_names`
+      parameter to the aliased object.  This may be useful in situations where
+      the given selectable is not directly related to the existing mapped
+      selectable.
+
+      .. versionadded:: 1.4.33
+
+    :param name: Name given to the generated :class:`.AliasedClass`.
+
+      .. versionadded:: 2.0.31
+
+    """
+    return AliasedInsp._with_polymorphic_factory(
+        base,
+        classes,
+        selectable=selectable,
+        flat=flat,
+        polymorphic_on=polymorphic_on,
+        adapt_on_names=adapt_on_names,
+        aliased=aliased,
+        innerjoin=innerjoin,
+        name=name,
+        _use_mapper_path=_use_mapper_path,
+    )
+
+
+def join(
+    left: _FromClauseArgument,
+    right: _FromClauseArgument,
+    onclause: Optional[_OnClauseArgument] = None,
+    isouter: bool = False,
+    full: bool = False,
+) -> _ORMJoin:
+    r"""Produce an inner join between left and right clauses.
+
+    :func:`_orm.join` is an extension to the core join interface
+    provided by :func:`_expression.join()`, where the
+    left and right selectable may be not only core selectable
+    objects such as :class:`_schema.Table`, but also mapped classes or
+    :class:`.AliasedClass` instances.   The "on" clause can
+    be a SQL expression or an ORM mapped attribute
+    referencing a configured :func:`_orm.relationship`.
+
+    :func:`_orm.join` is not commonly needed in modern usage,
+    as its functionality is encapsulated within that of the
+    :meth:`_sql.Select.join` and :meth:`_query.Query.join`
+    methods. which feature a
+    significant amount of automation beyond :func:`_orm.join`
+    by itself.  Explicit use of :func:`_orm.join`
+    with ORM-enabled SELECT statements involves use of the
+    :meth:`_sql.Select.select_from` method, as in::
+
+        from sqlalchemy.orm import join
+
+        stmt = (
+            select(User)
+            .select_from(join(User, Address, User.addresses))
+            .filter(Address.email_address == "foo@bar.com")
+        )
+
+    In modern SQLAlchemy the above join can be written more
+    succinctly as::
+
+        stmt = (
+            select(User)
+            .join(User.addresses)
+            .filter(Address.email_address == "foo@bar.com")
+        )
+
+    .. warning:: using :func:`_orm.join` directly may not work properly
+       with modern ORM options such as :func:`_orm.with_loader_criteria`.
+       It is strongly recommended to use the idiomatic join patterns
+       provided by methods such as :meth:`.Select.join` and
+       :meth:`.Select.join_from` when creating ORM joins.
+
+    .. seealso::
+
+        :ref:`orm_queryguide_joins` - in the :ref:`queryguide_toplevel` for
+        background on idiomatic ORM join patterns
+
+    """
+    return _ORMJoin(left, right, onclause, isouter, full)
+
+
+def outerjoin(
+    left: _FromClauseArgument,
+    right: _FromClauseArgument,
+    onclause: Optional[_OnClauseArgument] = None,
+    full: bool = False,
+) -> _ORMJoin:
+    """Produce a left outer join between left and right clauses.
+
+    This is the "outer join" version of the :func:`_orm.join` function,
+    featuring the same behavior except that an OUTER JOIN is generated.
+    See that function's documentation for other usage details.
+
+    """
+    return _ORMJoin(left, right, onclause, True, full)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_typing.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_typing.py
new file mode 100644
index 00000000..ccb8413b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/_typing.py
@@ -0,0 +1,179 @@
+# orm/_typing.py
+# Copyright (C) 2022-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import operator
+from typing import Any
+from typing import Dict
+from typing import Mapping
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from ..engine.interfaces import _CoreKnownExecutionOptions
+from ..sql import roles
+from ..sql._orm_types import DMLStrategyArgument as DMLStrategyArgument
+from ..sql._orm_types import (
+    SynchronizeSessionArgument as SynchronizeSessionArgument,
+)
+from ..sql._typing import _HasClauseElement
+from ..sql.elements import ColumnElement
+from ..util.typing import Protocol
+from ..util.typing import TypeGuard
+
+if TYPE_CHECKING:
+    from .attributes import AttributeImpl
+    from .attributes import CollectionAttributeImpl
+    from .attributes import HasCollectionAdapter
+    from .attributes import QueryableAttribute
+    from .base import PassiveFlag
+    from .decl_api import registry as _registry_type
+    from .interfaces import InspectionAttr
+    from .interfaces import MapperProperty
+    from .interfaces import ORMOption
+    from .interfaces import UserDefinedOption
+    from .mapper import Mapper
+    from .relationships import RelationshipProperty
+    from .state import InstanceState
+    from .util import AliasedClass
+    from .util import AliasedInsp
+    from ..sql._typing import _CE
+    from ..sql.base import ExecutableOption
+
+_T = TypeVar("_T", bound=Any)
+
+
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+_O = TypeVar("_O", bound=object)
+"""The 'ORM mapped object' type.
+
+"""
+
+
+if TYPE_CHECKING:
+    _RegistryType = _registry_type
+
+_InternalEntityType = Union["Mapper[_T]", "AliasedInsp[_T]"]
+
+_ExternalEntityType = Union[Type[_T], "AliasedClass[_T]"]
+
+_EntityType = Union[
+    Type[_T], "AliasedClass[_T]", "Mapper[_T]", "AliasedInsp[_T]"
+]
+
+
+_ClassDict = Mapping[str, Any]
+_InstanceDict = Dict[str, Any]
+
+_IdentityKeyType = Tuple[Type[_T], Tuple[Any, ...], Optional[Any]]
+
+_ORMColumnExprArgument = Union[
+    ColumnElement[_T],
+    _HasClauseElement[_T],
+    roles.ExpressionElementRole[_T],
+]
+
+
+_ORMCOLEXPR = TypeVar("_ORMCOLEXPR", bound=ColumnElement[Any])
+
+
+class _OrmKnownExecutionOptions(_CoreKnownExecutionOptions, total=False):
+    populate_existing: bool
+    autoflush: bool
+    synchronize_session: SynchronizeSessionArgument
+    dml_strategy: DMLStrategyArgument
+    is_delete_using: bool
+    is_update_from: bool
+    render_nulls: bool
+
+
+OrmExecuteOptionsParameter = Union[
+    _OrmKnownExecutionOptions, Mapping[str, Any]
+]
+
+
+class _ORMAdapterProto(Protocol):
+    """protocol for the :class:`.AliasedInsp._orm_adapt_element` method
+    which is a synonym for :class:`.AliasedInsp._adapt_element`.
+
+
+    """
+
+    def __call__(self, obj: _CE, key: Optional[str] = None) -> _CE: ...
+
+
+class _LoaderCallable(Protocol):
+    def __call__(
+        self, state: InstanceState[Any], passive: PassiveFlag
+    ) -> Any: ...
+
+
+def is_orm_option(
+    opt: ExecutableOption,
+) -> TypeGuard[ORMOption]:
+    return not opt._is_core
+
+
+def is_user_defined_option(
+    opt: ExecutableOption,
+) -> TypeGuard[UserDefinedOption]:
+    return not opt._is_core and opt._is_user_defined  # type: ignore
+
+
+def is_composite_class(obj: Any) -> bool:
+    # inlining is_dataclass(obj)
+    return hasattr(obj, "__composite_values__") or hasattr(
+        obj, "__dataclass_fields__"
+    )
+
+
+if TYPE_CHECKING:
+
+    def insp_is_mapper_property(
+        obj: Any,
+    ) -> TypeGuard[MapperProperty[Any]]: ...
+
+    def insp_is_mapper(obj: Any) -> TypeGuard[Mapper[Any]]: ...
+
+    def insp_is_aliased_class(obj: Any) -> TypeGuard[AliasedInsp[Any]]: ...
+
+    def insp_is_attribute(
+        obj: InspectionAttr,
+    ) -> TypeGuard[QueryableAttribute[Any]]: ...
+
+    def attr_is_internal_proxy(
+        obj: InspectionAttr,
+    ) -> TypeGuard[QueryableAttribute[Any]]: ...
+
+    def prop_is_relationship(
+        prop: MapperProperty[Any],
+    ) -> TypeGuard[RelationshipProperty[Any]]: ...
+
+    def is_collection_impl(
+        impl: AttributeImpl,
+    ) -> TypeGuard[CollectionAttributeImpl]: ...
+
+    def is_has_collection_adapter(
+        impl: AttributeImpl,
+    ) -> TypeGuard[HasCollectionAdapter]: ...
+
+else:
+    insp_is_mapper_property = operator.attrgetter("is_property")
+    insp_is_mapper = operator.attrgetter("is_mapper")
+    insp_is_aliased_class = operator.attrgetter("is_aliased_class")
+    insp_is_attribute = operator.attrgetter("is_attribute")
+    attr_is_internal_proxy = operator.attrgetter("_is_internal_proxy")
+    is_collection_impl = operator.attrgetter("collection")
+    prop_is_relationship = operator.attrgetter("_is_relationship")
+    is_has_collection_adapter = operator.attrgetter(
+        "_is_has_collection_adapter"
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/attributes.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/attributes.py
new file mode 100644
index 00000000..8207b4ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/attributes.py
@@ -0,0 +1,2835 @@
+# orm/attributes.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Defines instrumentation for class attributes and their interaction
+with instances.
+
+This module is usually not directly visible to user applications, but
+defines a large part of the ORM's interactivity.
+
+
+"""
+
+from __future__ import annotations
+
+import dataclasses
+import operator
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import NamedTuple
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import collections
+from . import exc as orm_exc
+from . import interfaces
+from ._typing import insp_is_aliased_class
+from .base import _DeclarativeMapped
+from .base import ATTR_EMPTY
+from .base import ATTR_WAS_SET
+from .base import CALLABLES_OK
+from .base import DEFERRED_HISTORY_LOAD
+from .base import INCLUDE_PENDING_MUTATIONS  # noqa
+from .base import INIT_OK
+from .base import instance_dict as instance_dict
+from .base import instance_state as instance_state
+from .base import instance_str
+from .base import LOAD_AGAINST_COMMITTED
+from .base import LoaderCallableStatus
+from .base import manager_of_class as manager_of_class
+from .base import Mapped as Mapped  # noqa
+from .base import NEVER_SET  # noqa
+from .base import NO_AUTOFLUSH
+from .base import NO_CHANGE  # noqa
+from .base import NO_KEY
+from .base import NO_RAISE
+from .base import NO_VALUE
+from .base import NON_PERSISTENT_OK  # noqa
+from .base import opt_manager_of_class as opt_manager_of_class
+from .base import PASSIVE_CLASS_MISMATCH  # noqa
+from .base import PASSIVE_NO_FETCH
+from .base import PASSIVE_NO_FETCH_RELATED  # noqa
+from .base import PASSIVE_NO_INITIALIZE
+from .base import PASSIVE_NO_RESULT
+from .base import PASSIVE_OFF
+from .base import PASSIVE_ONLY_PERSISTENT
+from .base import PASSIVE_RETURN_NO_VALUE
+from .base import PassiveFlag
+from .base import RELATED_OBJECT_OK  # noqa
+from .base import SQL_OK  # noqa
+from .base import SQLORMExpression
+from .base import state_str
+from .. import event
+from .. import exc
+from .. import inspection
+from .. import util
+from ..event import dispatcher
+from ..event import EventTarget
+from ..sql import base as sql_base
+from ..sql import cache_key
+from ..sql import coercions
+from ..sql import roles
+from ..sql import visitors
+from ..sql.cache_key import HasCacheKey
+from ..sql.visitors import _TraverseInternalsType
+from ..sql.visitors import InternalTraversal
+from ..util.typing import Literal
+from ..util.typing import Self
+from ..util.typing import TypeGuard
+
+if TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _ExternalEntityType
+    from ._typing import _InstanceDict
+    from ._typing import _InternalEntityType
+    from ._typing import _LoaderCallable
+    from ._typing import _O
+    from .collections import _AdaptedCollectionProtocol
+    from .collections import CollectionAdapter
+    from .interfaces import MapperProperty
+    from .relationships import RelationshipProperty
+    from .state import InstanceState
+    from .util import AliasedInsp
+    from .writeonly import WriteOnlyAttributeImpl
+    from ..event.base import _Dispatch
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _DMLColumnArgument
+    from ..sql._typing import _InfoType
+    from ..sql._typing import _PropagateAttrsType
+    from ..sql.annotation import _AnnotationDict
+    from ..sql.elements import ColumnElement
+    from ..sql.elements import Label
+    from ..sql.operators import OperatorType
+    from ..sql.selectable import FromClause
+
+
+_T = TypeVar("_T")
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+
+_AllPendingType = Sequence[
+    Tuple[Optional["InstanceState[Any]"], Optional[object]]
+]
+
+
+_UNKNOWN_ATTR_KEY = object()
+
+
+@inspection._self_inspects
+class QueryableAttribute(
+    _DeclarativeMapped[_T_co],
+    SQLORMExpression[_T_co],
+    interfaces.InspectionAttr,
+    interfaces.PropComparator[_T_co],
+    roles.JoinTargetRole,
+    roles.OnClauseRole,
+    sql_base.Immutable,
+    cache_key.SlotsMemoizedHasCacheKey,
+    util.MemoizedSlots,
+    EventTarget,
+):
+    """Base class for :term:`descriptor` objects that intercept
+    attribute events on behalf of a :class:`.MapperProperty`
+    object.  The actual :class:`.MapperProperty` is accessible
+    via the :attr:`.QueryableAttribute.property`
+    attribute.
+
+
+    .. seealso::
+
+        :class:`.InstrumentedAttribute`
+
+        :class:`.MapperProperty`
+
+        :attr:`_orm.Mapper.all_orm_descriptors`
+
+        :attr:`_orm.Mapper.attrs`
+    """
+
+    __slots__ = (
+        "class_",
+        "key",
+        "impl",
+        "comparator",
+        "property",
+        "parent",
+        "expression",
+        "_of_type",
+        "_extra_criteria",
+        "_slots_dispatch",
+        "_propagate_attrs",
+        "_doc",
+    )
+
+    is_attribute = True
+
+    dispatch: dispatcher[QueryableAttribute[_T_co]]
+
+    class_: _ExternalEntityType[Any]
+    key: str
+    parententity: _InternalEntityType[Any]
+    impl: AttributeImpl
+    comparator: interfaces.PropComparator[_T_co]
+    _of_type: Optional[_InternalEntityType[Any]]
+    _extra_criteria: Tuple[ColumnElement[bool], ...]
+    _doc: Optional[str]
+
+    # PropComparator has a __visit_name__ to participate within
+    # traversals.   Disambiguate the attribute vs. a comparator.
+    __visit_name__ = "orm_instrumented_attribute"
+
+    def __init__(
+        self,
+        class_: _ExternalEntityType[_O],
+        key: str,
+        parententity: _InternalEntityType[_O],
+        comparator: interfaces.PropComparator[_T_co],
+        impl: Optional[AttributeImpl] = None,
+        of_type: Optional[_InternalEntityType[Any]] = None,
+        extra_criteria: Tuple[ColumnElement[bool], ...] = (),
+    ):
+        self.class_ = class_
+        self.key = key
+
+        self._parententity = self.parent = parententity
+
+        # this attribute is non-None after mappers are set up, however in the
+        # interim class manager setup, there's a check for None to see if it
+        # needs to be populated, so we assign None here leaving the attribute
+        # in a temporarily not-type-correct state
+        self.impl = impl  # type: ignore
+
+        assert comparator is not None
+        self.comparator = comparator
+        self._of_type = of_type
+        self._extra_criteria = extra_criteria
+        self._doc = None
+
+        manager = opt_manager_of_class(class_)
+        # manager is None in the case of AliasedClass
+        if manager:
+            # propagate existing event listeners from
+            # immediate superclass
+            for base in manager._bases:
+                if key in base:
+                    self.dispatch._update(base[key].dispatch)
+                    if base[key].dispatch._active_history:
+                        self.dispatch._active_history = True  # type: ignore
+
+    _cache_key_traversal = [
+        ("key", visitors.ExtendedInternalTraversal.dp_string),
+        ("_parententity", visitors.ExtendedInternalTraversal.dp_multi),
+        ("_of_type", visitors.ExtendedInternalTraversal.dp_multi),
+        ("_extra_criteria", visitors.InternalTraversal.dp_clauseelement_list),
+    ]
+
+    def __reduce__(self) -> Any:
+        # this method is only used in terms of the
+        # sqlalchemy.ext.serializer extension
+        return (
+            _queryable_attribute_unreduce,
+            (
+                self.key,
+                self._parententity.mapper.class_,
+                self._parententity,
+                self._parententity.entity,
+            ),
+        )
+
+    @property
+    def _impl_uses_objects(self) -> bool:
+        return self.impl.uses_objects
+
+    def get_history(
+        self, instance: Any, passive: PassiveFlag = PASSIVE_OFF
+    ) -> History:
+        return self.impl.get_history(
+            instance_state(instance), instance_dict(instance), passive
+        )
+
+    @property
+    def info(self) -> _InfoType:
+        """Return the 'info' dictionary for the underlying SQL element.
+
+        The behavior here is as follows:
+
+        * If the attribute is a column-mapped property, i.e.
+          :class:`.ColumnProperty`, which is mapped directly
+          to a schema-level :class:`_schema.Column` object, this attribute
+          will return the :attr:`.SchemaItem.info` dictionary associated
+          with the core-level :class:`_schema.Column` object.
+
+        * If the attribute is a :class:`.ColumnProperty` but is mapped to
+          any other kind of SQL expression other than a
+          :class:`_schema.Column`,
+          the attribute will refer to the :attr:`.MapperProperty.info`
+          dictionary associated directly with the :class:`.ColumnProperty`,
+          assuming the SQL expression itself does not have its own ``.info``
+          attribute (which should be the case, unless a user-defined SQL
+          construct has defined one).
+
+        * If the attribute refers to any other kind of
+          :class:`.MapperProperty`, including :class:`.Relationship`,
+          the attribute will refer to the :attr:`.MapperProperty.info`
+          dictionary associated with that :class:`.MapperProperty`.
+
+        * To access the :attr:`.MapperProperty.info` dictionary of the
+          :class:`.MapperProperty` unconditionally, including for a
+          :class:`.ColumnProperty` that's associated directly with a
+          :class:`_schema.Column`, the attribute can be referred to using
+          :attr:`.QueryableAttribute.property` attribute, as
+          ``MyClass.someattribute.property.info``.
+
+        .. seealso::
+
+            :attr:`.SchemaItem.info`
+
+            :attr:`.MapperProperty.info`
+
+        """
+        return self.comparator.info
+
+    parent: _InternalEntityType[Any]
+    """Return an inspection instance representing the parent.
+
+    This will be either an instance of :class:`_orm.Mapper`
+    or :class:`.AliasedInsp`, depending upon the nature
+    of the parent entity which this attribute is associated
+    with.
+
+    """
+
+    expression: ColumnElement[_T_co]
+    """The SQL expression object represented by this
+    :class:`.QueryableAttribute`.
+
+    This will typically be an instance of a :class:`_sql.ColumnElement`
+    subclass representing a column expression.
+
+    """
+
+    def _memoized_attr_expression(self) -> ColumnElement[_T]:
+        annotations: _AnnotationDict
+
+        # applies only to Proxy() as used by hybrid.
+        # currently is an exception to typing rather than feeding through
+        # non-string keys.
+        # ideally Proxy() would have a separate set of methods to deal
+        # with this case.
+        entity_namespace = self._entity_namespace
+        assert isinstance(entity_namespace, HasCacheKey)
+
+        if self.key is _UNKNOWN_ATTR_KEY:
+            annotations = {"entity_namespace": entity_namespace}
+        else:
+            annotations = {
+                "proxy_key": self.key,
+                "proxy_owner": self._parententity,
+                "entity_namespace": entity_namespace,
+            }
+
+        ce = self.comparator.__clause_element__()
+        try:
+            if TYPE_CHECKING:
+                assert isinstance(ce, ColumnElement)
+            anno = ce._annotate
+        except AttributeError as ae:
+            raise exc.InvalidRequestError(
+                'When interpreting attribute "%s" as a SQL expression, '
+                "expected __clause_element__() to return "
+                "a ClauseElement object, got: %r" % (self, ce)
+            ) from ae
+        else:
+            return anno(annotations)
+
+    def _memoized_attr__propagate_attrs(self) -> _PropagateAttrsType:
+        # this suits the case in coercions where we don't actually
+        # call ``__clause_element__()`` but still need to get
+        # resolved._propagate_attrs.  See #6558.
+        return util.immutabledict(
+            {
+                "compile_state_plugin": "orm",
+                "plugin_subject": self._parentmapper,
+            }
+        )
+
+    @property
+    def _entity_namespace(self) -> _InternalEntityType[Any]:
+        return self._parententity
+
+    @property
+    def _annotations(self) -> _AnnotationDict:
+        return self.__clause_element__()._annotations
+
+    def __clause_element__(self) -> ColumnElement[_T_co]:
+        return self.expression
+
+    @property
+    def _from_objects(self) -> List[FromClause]:
+        return self.expression._from_objects
+
+    def _bulk_update_tuples(
+        self, value: Any
+    ) -> Sequence[Tuple[_DMLColumnArgument, Any]]:
+        """Return setter tuples for a bulk UPDATE."""
+
+        return self.comparator._bulk_update_tuples(value)
+
+    def adapt_to_entity(self, adapt_to_entity: AliasedInsp[Any]) -> Self:
+        assert not self._of_type
+        return self.__class__(
+            adapt_to_entity.entity,
+            self.key,
+            impl=self.impl,
+            comparator=self.comparator.adapt_to_entity(adapt_to_entity),
+            parententity=adapt_to_entity,
+        )
+
+    def of_type(self, entity: _EntityType[_T]) -> QueryableAttribute[_T]:
+        return QueryableAttribute(
+            self.class_,
+            self.key,
+            self._parententity,
+            impl=self.impl,
+            comparator=self.comparator.of_type(entity),
+            of_type=inspection.inspect(entity),
+            extra_criteria=self._extra_criteria,
+        )
+
+    def and_(
+        self, *clauses: _ColumnExpressionArgument[bool]
+    ) -> QueryableAttribute[bool]:
+        if TYPE_CHECKING:
+            assert isinstance(self.comparator, RelationshipProperty.Comparator)
+
+        exprs = tuple(
+            coercions.expect(roles.WhereHavingRole, clause)
+            for clause in util.coerce_generator_arg(clauses)
+        )
+
+        return QueryableAttribute(
+            self.class_,
+            self.key,
+            self._parententity,
+            impl=self.impl,
+            comparator=self.comparator.and_(*exprs),
+            of_type=self._of_type,
+            extra_criteria=self._extra_criteria + exprs,
+        )
+
+    def _clone(self, **kw: Any) -> QueryableAttribute[_T]:
+        return QueryableAttribute(
+            self.class_,
+            self.key,
+            self._parententity,
+            impl=self.impl,
+            comparator=self.comparator,
+            of_type=self._of_type,
+            extra_criteria=self._extra_criteria,
+        )
+
+    def label(self, name: Optional[str]) -> Label[_T_co]:
+        return self.__clause_element__().label(name)
+
+    def operate(
+        self, op: OperatorType, *other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(self.comparator, *other, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def reverse_operate(
+        self, op: OperatorType, other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(other, self.comparator, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def hasparent(
+        self, state: InstanceState[Any], optimistic: bool = False
+    ) -> bool:
+        return self.impl.hasparent(state, optimistic=optimistic) is not False
+
+    def __getattr__(self, key: str) -> Any:
+        try:
+            return util.MemoizedSlots.__getattr__(self, key)
+        except AttributeError:
+            pass
+
+        try:
+            return getattr(self.comparator, key)
+        except AttributeError as err:
+            raise AttributeError(
+                "Neither %r object nor %r object associated with %s "
+                "has an attribute %r"
+                % (
+                    type(self).__name__,
+                    type(self.comparator).__name__,
+                    self,
+                    key,
+                )
+            ) from err
+
+    def __str__(self) -> str:
+        return f"{self.class_.__name__}.{self.key}"
+
+    def _memoized_attr_property(self) -> Optional[MapperProperty[Any]]:
+        return self.comparator.property
+
+
+def _queryable_attribute_unreduce(
+    key: str,
+    mapped_class: Type[_O],
+    parententity: _InternalEntityType[_O],
+    entity: _ExternalEntityType[Any],
+) -> Any:
+    # this method is only used in terms of the
+    # sqlalchemy.ext.serializer extension
+    if insp_is_aliased_class(parententity):
+        return entity._get_from_serialized(key, mapped_class, parententity)
+    else:
+        return getattr(entity, key)
+
+
+class InstrumentedAttribute(QueryableAttribute[_T_co]):
+    """Class bound instrumented attribute which adds basic
+    :term:`descriptor` methods.
+
+    See :class:`.QueryableAttribute` for a description of most features.
+
+
+    """
+
+    __slots__ = ()
+
+    inherit_cache = True
+    """:meta private:"""
+
+    # hack to make __doc__ writeable on instances of
+    # InstrumentedAttribute, while still keeping classlevel
+    # __doc__ correct
+
+    @util.rw_hybridproperty
+    def __doc__(self) -> Optional[str]:
+        return self._doc
+
+    @__doc__.setter  # type: ignore
+    def __doc__(self, value: Optional[str]) -> None:
+        self._doc = value
+
+    @__doc__.classlevel  # type: ignore
+    def __doc__(cls) -> Optional[str]:
+        return super().__doc__
+
+    def __set__(self, instance: object, value: Any) -> None:
+        self.impl.set(
+            instance_state(instance), instance_dict(instance), value, None
+        )
+
+    def __delete__(self, instance: object) -> None:
+        self.impl.delete(instance_state(instance), instance_dict(instance))
+
+    @overload
+    def __get__(
+        self, instance: None, owner: Any
+    ) -> InstrumentedAttribute[_T_co]: ...
+
+    @overload
+    def __get__(self, instance: object, owner: Any) -> _T_co: ...
+
+    def __get__(
+        self, instance: Optional[object], owner: Any
+    ) -> Union[InstrumentedAttribute[_T_co], _T_co]:
+        if instance is None:
+            return self
+
+        dict_ = instance_dict(instance)
+        if self.impl.supports_population and self.key in dict_:
+            return dict_[self.key]  # type: ignore[no-any-return]
+        else:
+            try:
+                state = instance_state(instance)
+            except AttributeError as err:
+                raise orm_exc.UnmappedInstanceError(instance) from err
+            return self.impl.get(state, dict_)  # type: ignore[no-any-return]
+
+
+@dataclasses.dataclass(frozen=True)
+class AdHocHasEntityNamespace(HasCacheKey):
+    _traverse_internals: ClassVar[_TraverseInternalsType] = [
+        ("_entity_namespace", InternalTraversal.dp_has_cache_key),
+    ]
+
+    # py37 compat, no slots=True on dataclass
+    __slots__ = ("_entity_namespace",)
+    _entity_namespace: _InternalEntityType[Any]
+    is_mapper: ClassVar[bool] = False
+    is_aliased_class: ClassVar[bool] = False
+
+    @property
+    def entity_namespace(self):
+        return self._entity_namespace.entity_namespace
+
+
+def create_proxied_attribute(
+    descriptor: Any,
+) -> Callable[..., QueryableAttribute[Any]]:
+    """Create an QueryableAttribute / user descriptor hybrid.
+
+    Returns a new QueryableAttribute type that delegates descriptor
+    behavior and getattr() to the given descriptor.
+    """
+
+    # TODO: can move this to descriptor_props if the need for this
+    # function is removed from ext/hybrid.py
+
+    class Proxy(QueryableAttribute[Any]):
+        """Presents the :class:`.QueryableAttribute` interface as a
+        proxy on top of a Python descriptor / :class:`.PropComparator`
+        combination.
+
+        """
+
+        _extra_criteria = ()
+
+        # the attribute error catches inside of __getattr__ basically create a
+        # singularity if you try putting slots on this too
+        # __slots__ = ("descriptor", "original_property", "_comparator")
+
+        def __init__(
+            self,
+            class_,
+            key,
+            descriptor,
+            comparator,
+            adapt_to_entity=None,
+            doc=None,
+            original_property=None,
+        ):
+            self.class_ = class_
+            self.key = key
+            self.descriptor = descriptor
+            self.original_property = original_property
+            self._comparator = comparator
+            self._adapt_to_entity = adapt_to_entity
+            self._doc = self.__doc__ = doc
+
+        @property
+        def _parententity(self):
+            return inspection.inspect(self.class_, raiseerr=False)
+
+        @property
+        def parent(self):
+            return inspection.inspect(self.class_, raiseerr=False)
+
+        _is_internal_proxy = True
+
+        _cache_key_traversal = [
+            ("key", visitors.ExtendedInternalTraversal.dp_string),
+            ("_parententity", visitors.ExtendedInternalTraversal.dp_multi),
+        ]
+
+        @property
+        def _impl_uses_objects(self):
+            return (
+                self.original_property is not None
+                and getattr(self.class_, self.key).impl.uses_objects
+            )
+
+        @property
+        def _entity_namespace(self):
+            if hasattr(self._comparator, "_parententity"):
+                return self._comparator._parententity
+            else:
+                # used by hybrid attributes which try to remain
+                # agnostic of any ORM concepts like mappers
+                return AdHocHasEntityNamespace(self._parententity)
+
+        @property
+        def property(self):
+            return self.comparator.property
+
+        @util.memoized_property
+        def comparator(self):
+            if callable(self._comparator):
+                self._comparator = self._comparator()
+            if self._adapt_to_entity:
+                self._comparator = self._comparator.adapt_to_entity(
+                    self._adapt_to_entity
+                )
+            return self._comparator
+
+        def adapt_to_entity(self, adapt_to_entity):
+            return self.__class__(
+                adapt_to_entity.entity,
+                self.key,
+                self.descriptor,
+                self._comparator,
+                adapt_to_entity,
+            )
+
+        def _clone(self, **kw):
+            return self.__class__(
+                self.class_,
+                self.key,
+                self.descriptor,
+                self._comparator,
+                adapt_to_entity=self._adapt_to_entity,
+                original_property=self.original_property,
+            )
+
+        def __get__(self, instance, owner):
+            retval = self.descriptor.__get__(instance, owner)
+            # detect if this is a plain Python @property, which just returns
+            # itself for class level access.  If so, then return us.
+            # Otherwise, return the object returned by the descriptor.
+            if retval is self.descriptor and instance is None:
+                return self
+            else:
+                return retval
+
+        def __str__(self) -> str:
+            return f"{self.class_.__name__}.{self.key}"
+
+        def __getattr__(self, attribute):
+            """Delegate __getattr__ to the original descriptor and/or
+            comparator."""
+
+            # this is unfortunately very complicated, and is easily prone
+            # to recursion overflows when implementations of related
+            # __getattr__ schemes are changed
+
+            try:
+                return util.MemoizedSlots.__getattr__(self, attribute)
+            except AttributeError:
+                pass
+
+            try:
+                return getattr(descriptor, attribute)
+            except AttributeError as err:
+                if attribute == "comparator":
+                    raise AttributeError("comparator") from err
+                try:
+                    # comparator itself might be unreachable
+                    comparator = self.comparator
+                except AttributeError as err2:
+                    raise AttributeError(
+                        "Neither %r object nor unconfigured comparator "
+                        "object associated with %s has an attribute %r"
+                        % (type(descriptor).__name__, self, attribute)
+                    ) from err2
+                else:
+                    try:
+                        return getattr(comparator, attribute)
+                    except AttributeError as err3:
+                        raise AttributeError(
+                            "Neither %r object nor %r object "
+                            "associated with %s has an attribute %r"
+                            % (
+                                type(descriptor).__name__,
+                                type(comparator).__name__,
+                                self,
+                                attribute,
+                            )
+                        ) from err3
+
+    Proxy.__name__ = type(descriptor).__name__ + "Proxy"
+
+    util.monkeypatch_proxied_specials(
+        Proxy, type(descriptor), name="descriptor", from_instance=descriptor
+    )
+    return Proxy
+
+
+OP_REMOVE = util.symbol("REMOVE")
+OP_APPEND = util.symbol("APPEND")
+OP_REPLACE = util.symbol("REPLACE")
+OP_BULK_REPLACE = util.symbol("BULK_REPLACE")
+OP_MODIFIED = util.symbol("MODIFIED")
+
+
+class AttributeEventToken:
+    """A token propagated throughout the course of a chain of attribute
+    events.
+
+    Serves as an indicator of the source of the event and also provides
+    a means of controlling propagation across a chain of attribute
+    operations.
+
+    The :class:`.Event` object is sent as the ``initiator`` argument
+    when dealing with events such as :meth:`.AttributeEvents.append`,
+    :meth:`.AttributeEvents.set`,
+    and :meth:`.AttributeEvents.remove`.
+
+    The :class:`.Event` object is currently interpreted by the backref
+    event handlers, and is used to control the propagation of operations
+    across two mutually-dependent attributes.
+
+    .. versionchanged:: 2.0  Changed the name from ``AttributeEvent``
+       to ``AttributeEventToken``.
+
+    :attribute impl: The :class:`.AttributeImpl` which is the current event
+     initiator.
+
+    :attribute op: The symbol :attr:`.OP_APPEND`, :attr:`.OP_REMOVE`,
+     :attr:`.OP_REPLACE`, or :attr:`.OP_BULK_REPLACE`, indicating the
+     source operation.
+
+    """
+
+    __slots__ = "impl", "op", "parent_token"
+
+    def __init__(self, attribute_impl: AttributeImpl, op: util.symbol):
+        self.impl = attribute_impl
+        self.op = op
+        self.parent_token = self.impl.parent_token
+
+    def __eq__(self, other):
+        return (
+            isinstance(other, AttributeEventToken)
+            and other.impl is self.impl
+            and other.op == self.op
+        )
+
+    @property
+    def key(self):
+        return self.impl.key
+
+    def hasparent(self, state):
+        return self.impl.hasparent(state)
+
+
+AttributeEvent = AttributeEventToken  # legacy
+Event = AttributeEventToken  # legacy
+
+
+class AttributeImpl:
+    """internal implementation for instrumented attributes."""
+
+    collection: bool
+    default_accepts_scalar_loader: bool
+    uses_objects: bool
+    supports_population: bool
+    dynamic: bool
+
+    _is_has_collection_adapter = False
+
+    _replace_token: AttributeEventToken
+    _remove_token: AttributeEventToken
+    _append_token: AttributeEventToken
+
+    def __init__(
+        self,
+        class_: _ExternalEntityType[_O],
+        key: str,
+        callable_: Optional[_LoaderCallable],
+        dispatch: _Dispatch[QueryableAttribute[Any]],
+        trackparent: bool = False,
+        compare_function: Optional[Callable[..., bool]] = None,
+        active_history: bool = False,
+        parent_token: Optional[AttributeEventToken] = None,
+        load_on_unexpire: bool = True,
+        send_modified_events: bool = True,
+        accepts_scalar_loader: Optional[bool] = None,
+        **kwargs: Any,
+    ):
+        r"""Construct an AttributeImpl.
+
+        :param \class_: associated class
+
+        :param key: string name of the attribute
+
+        :param \callable_:
+          optional function which generates a callable based on a parent
+          instance, which produces the "default" values for a scalar or
+          collection attribute when it's first accessed, if not present
+          already.
+
+        :param trackparent:
+          if True, attempt to track if an instance has a parent attached
+          to it via this attribute.
+
+        :param compare_function:
+          a function that compares two values which are normally
+          assignable to this attribute.
+
+        :param active_history:
+          indicates that get_history() should always return the "old" value,
+          even if it means executing a lazy callable upon attribute change.
+
+        :param parent_token:
+          Usually references the MapperProperty, used as a key for
+          the hasparent() function to identify an "owning" attribute.
+          Allows multiple AttributeImpls to all match a single
+          owner attribute.
+
+        :param load_on_unexpire:
+          if False, don't include this attribute in a load-on-expired
+          operation, i.e. the "expired_attribute_loader" process.
+          The attribute can still be in the "expired" list and be
+          considered to be "expired".   Previously, this flag was called
+          "expire_missing" and is only used by a deferred column
+          attribute.
+
+        :param send_modified_events:
+          if False, the InstanceState._modified_event method will have no
+          effect; this means the attribute will never show up as changed in a
+          history entry.
+
+        """
+        self.class_ = class_
+        self.key = key
+        self.callable_ = callable_
+        self.dispatch = dispatch
+        self.trackparent = trackparent
+        self.parent_token = parent_token or self
+        self.send_modified_events = send_modified_events
+        if compare_function is None:
+            self.is_equal = operator.eq
+        else:
+            self.is_equal = compare_function
+
+        if accepts_scalar_loader is not None:
+            self.accepts_scalar_loader = accepts_scalar_loader
+        else:
+            self.accepts_scalar_loader = self.default_accepts_scalar_loader
+
+        _deferred_history = kwargs.pop("_deferred_history", False)
+        self._deferred_history = _deferred_history
+
+        if active_history:
+            self.dispatch._active_history = True
+
+        self.load_on_unexpire = load_on_unexpire
+        self._modified_token = AttributeEventToken(self, OP_MODIFIED)
+
+    __slots__ = (
+        "class_",
+        "key",
+        "callable_",
+        "dispatch",
+        "trackparent",
+        "parent_token",
+        "send_modified_events",
+        "is_equal",
+        "load_on_unexpire",
+        "_modified_token",
+        "accepts_scalar_loader",
+        "_deferred_history",
+    )
+
+    def __str__(self) -> str:
+        return f"{self.class_.__name__}.{self.key}"
+
+    def _get_active_history(self):
+        """Backwards compat for impl.active_history"""
+
+        return self.dispatch._active_history
+
+    def _set_active_history(self, value):
+        self.dispatch._active_history = value
+
+    active_history = property(_get_active_history, _set_active_history)
+
+    def hasparent(
+        self, state: InstanceState[Any], optimistic: bool = False
+    ) -> bool:
+        """Return the boolean value of a `hasparent` flag attached to
+        the given state.
+
+        The `optimistic` flag determines what the default return value
+        should be if no `hasparent` flag can be located.
+
+        As this function is used to determine if an instance is an
+        *orphan*, instances that were loaded from storage should be
+        assumed to not be orphans, until a True/False value for this
+        flag is set.
+
+        An instance attribute that is loaded by a callable function
+        will also not have a `hasparent` flag.
+
+        """
+        msg = "This AttributeImpl is not configured to track parents."
+        assert self.trackparent, msg
+
+        return (
+            state.parents.get(id(self.parent_token), optimistic) is not False
+        )
+
+    def sethasparent(
+        self,
+        state: InstanceState[Any],
+        parent_state: InstanceState[Any],
+        value: bool,
+    ) -> None:
+        """Set a boolean flag on the given item corresponding to
+        whether or not it is attached to a parent object via the
+        attribute represented by this ``InstrumentedAttribute``.
+
+        """
+        msg = "This AttributeImpl is not configured to track parents."
+        assert self.trackparent, msg
+
+        id_ = id(self.parent_token)
+        if value:
+            state.parents[id_] = parent_state
+        else:
+            if id_ in state.parents:
+                last_parent = state.parents[id_]
+
+                if (
+                    last_parent is not False
+                    and last_parent.key != parent_state.key
+                ):
+                    if last_parent.obj() is None:
+                        raise orm_exc.StaleDataError(
+                            "Removing state %s from parent "
+                            "state %s along attribute '%s', "
+                            "but the parent record "
+                            "has gone stale, can't be sure this "
+                            "is the most recent parent."
+                            % (
+                                state_str(state),
+                                state_str(parent_state),
+                                self.key,
+                            )
+                        )
+
+                    return
+
+            state.parents[id_] = False
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> History:
+        raise NotImplementedError()
+
+    def get_all_pending(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_NO_INITIALIZE,
+    ) -> _AllPendingType:
+        """Return a list of tuples of (state, obj)
+        for all objects in this attribute's current state
+        + history.
+
+        Only applies to object-based attributes.
+
+        This is an inlining of existing functionality
+        which roughly corresponds to:
+
+            get_state_history(
+                        state,
+                        key,
+                        passive=PASSIVE_NO_INITIALIZE).sum()
+
+        """
+        raise NotImplementedError()
+
+    def _default_value(
+        self, state: InstanceState[Any], dict_: _InstanceDict
+    ) -> Any:
+        """Produce an empty value for an uninitialized scalar attribute."""
+
+        assert self.key not in dict_, (
+            "_default_value should only be invoked for an "
+            "uninitialized or expired attribute"
+        )
+
+        value = None
+        for fn in self.dispatch.init_scalar:
+            ret = fn(state, value, dict_)
+            if ret is not ATTR_EMPTY:
+                value = ret
+
+        return value
+
+    def get(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> Any:
+        """Retrieve a value from the given object.
+        If a callable is assembled on this object's attribute, and
+        passive is False, the callable will be executed and the
+        resulting value will be set as the new value for this attribute.
+        """
+        if self.key in dict_:
+            return dict_[self.key]
+        else:
+            # if history present, don't load
+            key = self.key
+            if (
+                key not in state.committed_state
+                or state.committed_state[key] is NO_VALUE
+            ):
+                if not passive & CALLABLES_OK:
+                    return PASSIVE_NO_RESULT
+
+                value = self._fire_loader_callables(state, key, passive)
+
+                if value is PASSIVE_NO_RESULT or value is NO_VALUE:
+                    return value
+                elif value is ATTR_WAS_SET:
+                    try:
+                        return dict_[key]
+                    except KeyError as err:
+                        # TODO: no test coverage here.
+                        raise KeyError(
+                            "Deferred loader for attribute "
+                            "%r failed to populate "
+                            "correctly" % key
+                        ) from err
+                elif value is not ATTR_EMPTY:
+                    return self.set_committed_value(state, dict_, value)
+
+            if not passive & INIT_OK:
+                return NO_VALUE
+            else:
+                return self._default_value(state, dict_)
+
+    def _fire_loader_callables(
+        self, state: InstanceState[Any], key: str, passive: PassiveFlag
+    ) -> Any:
+        if (
+            self.accepts_scalar_loader
+            and self.load_on_unexpire
+            and key in state.expired_attributes
+        ):
+            return state._load_expired(state, passive)
+        elif key in state.callables:
+            callable_ = state.callables[key]
+            return callable_(state, passive)
+        elif self.callable_:
+            return self.callable_(state, passive)
+        else:
+            return ATTR_EMPTY
+
+    def append(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        self.set(state, dict_, value, initiator, passive=passive)
+
+    def remove(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        self.set(
+            state, dict_, None, initiator, passive=passive, check_old=value
+        )
+
+    def pop(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        self.set(
+            state,
+            dict_,
+            None,
+            initiator,
+            passive=passive,
+            check_old=value,
+            pop=True,
+        )
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PASSIVE_OFF,
+        check_old: Any = None,
+        pop: bool = False,
+    ) -> None:
+        raise NotImplementedError()
+
+    def delete(self, state: InstanceState[Any], dict_: _InstanceDict) -> None:
+        raise NotImplementedError()
+
+    def get_committed_value(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> Any:
+        """return the unchanged value of this attribute"""
+
+        if self.key in state.committed_state:
+            value = state.committed_state[self.key]
+            if value is NO_VALUE:
+                return None
+            else:
+                return value
+        else:
+            return self.get(state, dict_, passive=passive)
+
+    def set_committed_value(self, state, dict_, value):
+        """set an attribute value on the given instance and 'commit' it."""
+
+        dict_[self.key] = value
+        state._commit(dict_, [self.key])
+        return value
+
+
+class ScalarAttributeImpl(AttributeImpl):
+    """represents a scalar value-holding InstrumentedAttribute."""
+
+    default_accepts_scalar_loader = True
+    uses_objects = False
+    supports_population = True
+    collection = False
+    dynamic = False
+
+    __slots__ = "_replace_token", "_append_token", "_remove_token"
+
+    def __init__(self, *arg, **kw):
+        super().__init__(*arg, **kw)
+        self._replace_token = self._append_token = AttributeEventToken(
+            self, OP_REPLACE
+        )
+        self._remove_token = AttributeEventToken(self, OP_REMOVE)
+
+    def delete(self, state: InstanceState[Any], dict_: _InstanceDict) -> None:
+        if self.dispatch._active_history:
+            old = self.get(state, dict_, PASSIVE_RETURN_NO_VALUE)
+        else:
+            old = dict_.get(self.key, NO_VALUE)
+
+        if self.dispatch.remove:
+            self.fire_remove_event(state, dict_, old, self._remove_token)
+        state._modified_event(dict_, self, old)
+
+        existing = dict_.pop(self.key, NO_VALUE)
+        if (
+            existing is NO_VALUE
+            and old is NO_VALUE
+            and not state.expired
+            and self.key not in state.expired_attributes
+        ):
+            raise AttributeError("%s object does not have a value" % self)
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: Dict[str, Any],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> History:
+        if self.key in dict_:
+            return History.from_scalar_attribute(self, state, dict_[self.key])
+        elif self.key in state.committed_state:
+            return History.from_scalar_attribute(self, state, NO_VALUE)
+        else:
+            if passive & INIT_OK:
+                passive ^= INIT_OK
+            current = self.get(state, dict_, passive=passive)
+            if current is PASSIVE_NO_RESULT:
+                return HISTORY_BLANK
+            else:
+                return History.from_scalar_attribute(self, state, current)
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: Dict[str, Any],
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PASSIVE_OFF,
+        check_old: Optional[object] = None,
+        pop: bool = False,
+    ) -> None:
+        if self.dispatch._active_history:
+            old = self.get(state, dict_, PASSIVE_RETURN_NO_VALUE)
+        else:
+            old = dict_.get(self.key, NO_VALUE)
+
+        if self.dispatch.set:
+            value = self.fire_replace_event(
+                state, dict_, value, old, initiator
+            )
+        state._modified_event(dict_, self, old)
+        dict_[self.key] = value
+
+    def fire_replace_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: _T,
+        previous: Any,
+        initiator: Optional[AttributeEventToken],
+    ) -> _T:
+        for fn in self.dispatch.set:
+            value = fn(
+                state, value, previous, initiator or self._replace_token
+            )
+        return value
+
+    def fire_remove_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+    ) -> None:
+        for fn in self.dispatch.remove:
+            fn(state, value, initiator or self._remove_token)
+
+
+class ScalarObjectAttributeImpl(ScalarAttributeImpl):
+    """represents a scalar-holding InstrumentedAttribute,
+    where the target object is also instrumented.
+
+    Adds events to delete/set operations.
+
+    """
+
+    default_accepts_scalar_loader = False
+    uses_objects = True
+    supports_population = True
+    collection = False
+
+    __slots__ = ()
+
+    def delete(self, state: InstanceState[Any], dict_: _InstanceDict) -> None:
+        if self.dispatch._active_history:
+            old = self.get(
+                state,
+                dict_,
+                passive=PASSIVE_ONLY_PERSISTENT
+                | NO_AUTOFLUSH
+                | LOAD_AGAINST_COMMITTED,
+            )
+        else:
+            old = self.get(
+                state,
+                dict_,
+                passive=PASSIVE_NO_FETCH ^ INIT_OK
+                | LOAD_AGAINST_COMMITTED
+                | NO_RAISE,
+            )
+
+        self.fire_remove_event(state, dict_, old, self._remove_token)
+
+        existing = dict_.pop(self.key, NO_VALUE)
+
+        # if the attribute is expired, we currently have no way to tell
+        # that an object-attribute was expired vs. not loaded.   So
+        # for this test, we look to see if the object has a DB identity.
+        if (
+            existing is NO_VALUE
+            and old is not PASSIVE_NO_RESULT
+            and state.key is None
+        ):
+            raise AttributeError("%s object does not have a value" % self)
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> History:
+        if self.key in dict_:
+            current = dict_[self.key]
+        else:
+            if passive & INIT_OK:
+                passive ^= INIT_OK
+            current = self.get(state, dict_, passive=passive)
+            if current is PASSIVE_NO_RESULT:
+                return HISTORY_BLANK
+
+        if not self._deferred_history:
+            return History.from_object_attribute(self, state, current)
+        else:
+            original = state.committed_state.get(self.key, _NO_HISTORY)
+            if original is PASSIVE_NO_RESULT:
+                loader_passive = passive | (
+                    PASSIVE_ONLY_PERSISTENT
+                    | NO_AUTOFLUSH
+                    | LOAD_AGAINST_COMMITTED
+                    | NO_RAISE
+                    | DEFERRED_HISTORY_LOAD
+                )
+                original = self._fire_loader_callables(
+                    state, self.key, loader_passive
+                )
+            return History.from_object_attribute(
+                self, state, current, original=original
+            )
+
+    def get_all_pending(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_NO_INITIALIZE,
+    ) -> _AllPendingType:
+        if self.key in dict_:
+            current = dict_[self.key]
+        elif passive & CALLABLES_OK:
+            current = self.get(state, dict_, passive=passive)
+        else:
+            return []
+
+        ret: _AllPendingType
+
+        # can't use __hash__(), can't use __eq__() here
+        if (
+            current is not None
+            and current is not PASSIVE_NO_RESULT
+            and current is not NO_VALUE
+        ):
+            ret = [(instance_state(current), current)]
+        else:
+            ret = [(None, None)]
+
+        if self.key in state.committed_state:
+            original = state.committed_state[self.key]
+            if (
+                original is not None
+                and original is not PASSIVE_NO_RESULT
+                and original is not NO_VALUE
+                and original is not current
+            ):
+                ret.append((instance_state(original), original))
+        return ret
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PASSIVE_OFF,
+        check_old: Any = None,
+        pop: bool = False,
+    ) -> None:
+        """Set a value on the given InstanceState."""
+
+        if self.dispatch._active_history:
+            old = self.get(
+                state,
+                dict_,
+                passive=PASSIVE_ONLY_PERSISTENT
+                | NO_AUTOFLUSH
+                | LOAD_AGAINST_COMMITTED,
+            )
+        else:
+            old = self.get(
+                state,
+                dict_,
+                passive=PASSIVE_NO_FETCH ^ INIT_OK
+                | LOAD_AGAINST_COMMITTED
+                | NO_RAISE,
+            )
+
+        if (
+            check_old is not None
+            and old is not PASSIVE_NO_RESULT
+            and check_old is not old
+        ):
+            if pop:
+                return
+            else:
+                raise ValueError(
+                    "Object %s not associated with %s on attribute '%s'"
+                    % (instance_str(check_old), state_str(state), self.key)
+                )
+
+        value = self.fire_replace_event(state, dict_, value, old, initiator)
+        dict_[self.key] = value
+
+    def fire_remove_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+    ) -> None:
+        if self.trackparent and value not in (
+            None,
+            PASSIVE_NO_RESULT,
+            NO_VALUE,
+        ):
+            self.sethasparent(instance_state(value), state, False)
+
+        for fn in self.dispatch.remove:
+            fn(state, value, initiator or self._remove_token)
+
+        state._modified_event(dict_, self, value)
+
+    def fire_replace_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: _T,
+        previous: Any,
+        initiator: Optional[AttributeEventToken],
+    ) -> _T:
+        if self.trackparent:
+            if previous is not value and previous not in (
+                None,
+                PASSIVE_NO_RESULT,
+                NO_VALUE,
+            ):
+                self.sethasparent(instance_state(previous), state, False)
+
+        for fn in self.dispatch.set:
+            value = fn(
+                state, value, previous, initiator or self._replace_token
+            )
+
+        state._modified_event(dict_, self, previous)
+
+        if self.trackparent:
+            if value is not None:
+                self.sethasparent(instance_state(value), state, True)
+
+        return value
+
+
+class HasCollectionAdapter:
+    __slots__ = ()
+
+    collection: bool
+    _is_has_collection_adapter = True
+
+    def _dispose_previous_collection(
+        self,
+        state: InstanceState[Any],
+        collection: _AdaptedCollectionProtocol,
+        adapter: CollectionAdapter,
+        fire_event: bool,
+    ) -> None:
+        raise NotImplementedError()
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Literal[None] = ...,
+        passive: Literal[PassiveFlag.PASSIVE_OFF] = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: _AdaptedCollectionProtocol = ...,
+        passive: PassiveFlag = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = ...,
+        passive: PassiveFlag = ...,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]: ...
+
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]:
+        raise NotImplementedError()
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+        check_old: Any = None,
+        pop: bool = False,
+        _adapt: bool = True,
+    ) -> None:
+        raise NotImplementedError()
+
+
+if TYPE_CHECKING:
+
+    def _is_collection_attribute_impl(
+        impl: AttributeImpl,
+    ) -> TypeGuard[CollectionAttributeImpl]: ...
+
+else:
+    _is_collection_attribute_impl = operator.attrgetter("collection")
+
+
+class CollectionAttributeImpl(HasCollectionAdapter, AttributeImpl):
+    """A collection-holding attribute that instruments changes in membership.
+
+    Only handles collections of instrumented objects.
+
+    InstrumentedCollectionAttribute holds an arbitrary, user-specified
+    container object (defaulting to a list) and brokers access to the
+    CollectionAdapter, a "view" onto that object that presents consistent bag
+    semantics to the orm layer independent of the user data implementation.
+
+    """
+
+    uses_objects = True
+    collection = True
+    default_accepts_scalar_loader = False
+    supports_population = True
+    dynamic = False
+
+    _bulk_replace_token: AttributeEventToken
+
+    __slots__ = (
+        "copy",
+        "collection_factory",
+        "_append_token",
+        "_remove_token",
+        "_bulk_replace_token",
+        "_duck_typed_as",
+    )
+
+    def __init__(
+        self,
+        class_,
+        key,
+        callable_,
+        dispatch,
+        typecallable=None,
+        trackparent=False,
+        copy_function=None,
+        compare_function=None,
+        **kwargs,
+    ):
+        super().__init__(
+            class_,
+            key,
+            callable_,
+            dispatch,
+            trackparent=trackparent,
+            compare_function=compare_function,
+            **kwargs,
+        )
+
+        if copy_function is None:
+            copy_function = self.__copy
+        self.copy = copy_function
+        self.collection_factory = typecallable
+        self._append_token = AttributeEventToken(self, OP_APPEND)
+        self._remove_token = AttributeEventToken(self, OP_REMOVE)
+        self._bulk_replace_token = AttributeEventToken(self, OP_BULK_REPLACE)
+        self._duck_typed_as = util.duck_type_collection(
+            self.collection_factory()
+        )
+
+        if getattr(self.collection_factory, "_sa_linker", None):
+
+            @event.listens_for(self, "init_collection")
+            def link(target, collection, collection_adapter):
+                collection._sa_linker(collection_adapter)
+
+            @event.listens_for(self, "dispose_collection")
+            def unlink(target, collection, collection_adapter):
+                collection._sa_linker(None)
+
+    def __copy(self, item):
+        return [y for y in collections.collection_adapter(item)]
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> History:
+        current = self.get(state, dict_, passive=passive)
+
+        if current is PASSIVE_NO_RESULT:
+            if (
+                passive & PassiveFlag.INCLUDE_PENDING_MUTATIONS
+                and self.key in state._pending_mutations
+            ):
+                pending = state._pending_mutations[self.key]
+                return pending.merge_with_history(HISTORY_BLANK)
+            else:
+                return HISTORY_BLANK
+        else:
+            if passive & PassiveFlag.INCLUDE_PENDING_MUTATIONS:
+                # this collection is loaded / present.  should not be any
+                # pending mutations
+                assert self.key not in state._pending_mutations
+
+            return History.from_collection(self, state, current)
+
+    def get_all_pending(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PASSIVE_NO_INITIALIZE,
+    ) -> _AllPendingType:
+        # NOTE: passive is ignored here at the moment
+
+        if self.key not in dict_:
+            return []
+
+        current = dict_[self.key]
+        current = getattr(current, "_sa_adapter")
+
+        if self.key in state.committed_state:
+            original = state.committed_state[self.key]
+            if original is not NO_VALUE:
+                current_states = [
+                    ((c is not None) and instance_state(c) or None, c)
+                    for c in current
+                ]
+                original_states = [
+                    ((c is not None) and instance_state(c) or None, c)
+                    for c in original
+                ]
+
+                current_set = dict(current_states)
+                original_set = dict(original_states)
+
+                return (
+                    [
+                        (s, o)
+                        for s, o in current_states
+                        if s not in original_set
+                    ]
+                    + [(s, o) for s, o in current_states if s in original_set]
+                    + [
+                        (s, o)
+                        for s, o in original_states
+                        if s not in current_set
+                    ]
+                )
+
+        return [(instance_state(o), o) for o in current]
+
+    def fire_append_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: _T,
+        initiator: Optional[AttributeEventToken],
+        key: Optional[Any],
+    ) -> _T:
+        for fn in self.dispatch.append:
+            value = fn(state, value, initiator or self._append_token, key=key)
+
+        state._modified_event(dict_, self, NO_VALUE, True)
+
+        if self.trackparent and value is not None:
+            self.sethasparent(instance_state(value), state, True)
+
+        return value
+
+    def fire_append_wo_mutation_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: _T,
+        initiator: Optional[AttributeEventToken],
+        key: Optional[Any],
+    ) -> _T:
+        for fn in self.dispatch.append_wo_mutation:
+            value = fn(state, value, initiator or self._append_token, key=key)
+
+        return value
+
+    def fire_pre_remove_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        initiator: Optional[AttributeEventToken],
+        key: Optional[Any],
+    ) -> None:
+        """A special event used for pop() operations.
+
+        The "remove" event needs to have the item to be removed passed to
+        it, which in the case of pop from a set, we don't have a way to access
+        the item before the operation.   the event is used for all pop()
+        operations (even though set.pop is the one where it is really needed).
+
+        """
+        state._modified_event(dict_, self, NO_VALUE, True)
+
+    def fire_remove_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        key: Optional[Any],
+    ) -> None:
+        if self.trackparent and value is not None:
+            self.sethasparent(instance_state(value), state, False)
+
+        for fn in self.dispatch.remove:
+            fn(state, value, initiator or self._remove_token, key=key)
+
+        state._modified_event(dict_, self, NO_VALUE, True)
+
+    def delete(self, state: InstanceState[Any], dict_: _InstanceDict) -> None:
+        if self.key not in dict_:
+            return
+
+        state._modified_event(dict_, self, NO_VALUE, True)
+
+        collection = self.get_collection(state, state.dict)
+        collection.clear_with_event()
+
+        # key is always present because we checked above.  e.g.
+        # del is a no-op if collection not present.
+        del dict_[self.key]
+
+    def _default_value(
+        self, state: InstanceState[Any], dict_: _InstanceDict
+    ) -> _AdaptedCollectionProtocol:
+        """Produce an empty collection for an un-initialized attribute"""
+
+        assert self.key not in dict_, (
+            "_default_value should only be invoked for an "
+            "uninitialized or expired attribute"
+        )
+
+        if self.key in state._empty_collections:
+            return state._empty_collections[self.key]
+
+        adapter, user_data = self._initialize_collection(state)
+        adapter._set_empty(user_data)
+        return user_data
+
+    def _initialize_collection(
+        self, state: InstanceState[Any]
+    ) -> Tuple[CollectionAdapter, _AdaptedCollectionProtocol]:
+        adapter, collection = state.manager.initialize_collection(
+            self.key, state, self.collection_factory
+        )
+
+        self.dispatch.init_collection(state, collection, adapter)
+
+        return adapter, collection
+
+    def append(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        collection = self.get_collection(
+            state, dict_, user_data=None, passive=passive
+        )
+        if collection is PASSIVE_NO_RESULT:
+            value = self.fire_append_event(
+                state, dict_, value, initiator, key=NO_KEY
+            )
+            assert (
+                self.key not in dict_
+            ), "Collection was loaded during event handling."
+            state._get_pending_mutation(self.key).append(value)
+        else:
+            if TYPE_CHECKING:
+                assert isinstance(collection, CollectionAdapter)
+            collection.append_with_event(value, initiator)
+
+    def remove(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        collection = self.get_collection(
+            state, state.dict, user_data=None, passive=passive
+        )
+        if collection is PASSIVE_NO_RESULT:
+            self.fire_remove_event(state, dict_, value, initiator, key=NO_KEY)
+            assert (
+                self.key not in dict_
+            ), "Collection was loaded during event handling."
+            state._get_pending_mutation(self.key).remove(value)
+        else:
+            if TYPE_CHECKING:
+                assert isinstance(collection, CollectionAdapter)
+            collection.remove_with_event(value, initiator)
+
+    def pop(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> None:
+        try:
+            # TODO: better solution here would be to add
+            # a "popper" role to collections.py to complement
+            # "remover".
+            self.remove(state, dict_, value, initiator, passive=passive)
+        except (ValueError, KeyError, IndexError):
+            pass
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+        check_old: Any = None,
+        pop: bool = False,
+        _adapt: bool = True,
+    ) -> None:
+        iterable = orig_iterable = value
+        new_keys = None
+
+        # pulling a new collection first so that an adaptation exception does
+        # not trigger a lazy load of the old collection.
+        new_collection, user_data = self._initialize_collection(state)
+        if _adapt:
+            if new_collection._converter is not None:
+                iterable = new_collection._converter(iterable)
+            else:
+                setting_type = util.duck_type_collection(iterable)
+                receiving_type = self._duck_typed_as
+
+                if setting_type is not receiving_type:
+                    given = (
+                        iterable is None
+                        and "None"
+                        or iterable.__class__.__name__
+                    )
+                    wanted = self._duck_typed_as.__name__
+                    raise TypeError(
+                        "Incompatible collection type: %s is not %s-like"
+                        % (given, wanted)
+                    )
+
+                # If the object is an adapted collection, return the (iterable)
+                # adapter.
+                if hasattr(iterable, "_sa_iterator"):
+                    iterable = iterable._sa_iterator()
+                elif setting_type is dict:
+                    new_keys = list(iterable)
+                    iterable = iterable.values()
+                else:
+                    iterable = iter(iterable)
+        elif util.duck_type_collection(iterable) is dict:
+            new_keys = list(value)
+
+        new_values = list(iterable)
+
+        evt = self._bulk_replace_token
+
+        self.dispatch.bulk_replace(state, new_values, evt, keys=new_keys)
+
+        # propagate NO_RAISE in passive through to the get() for the
+        # existing object (ticket #8862)
+        old = self.get(
+            state,
+            dict_,
+            passive=PASSIVE_ONLY_PERSISTENT ^ (passive & PassiveFlag.NO_RAISE),
+        )
+        if old is PASSIVE_NO_RESULT:
+            old = self._default_value(state, dict_)
+        elif old is orig_iterable:
+            # ignore re-assignment of the current collection, as happens
+            # implicitly with in-place operators (foo.collection |= other)
+            return
+
+        # place a copy of "old" in state.committed_state
+        state._modified_event(dict_, self, old, True)
+
+        old_collection = old._sa_adapter
+
+        dict_[self.key] = user_data
+
+        collections.bulk_replace(
+            new_values, old_collection, new_collection, initiator=evt
+        )
+
+        self._dispose_previous_collection(state, old, old_collection, True)
+
+    def _dispose_previous_collection(
+        self,
+        state: InstanceState[Any],
+        collection: _AdaptedCollectionProtocol,
+        adapter: CollectionAdapter,
+        fire_event: bool,
+    ) -> None:
+        del collection._sa_adapter
+
+        # discarding old collection make sure it is not referenced in empty
+        # collections.
+        state._empty_collections.pop(self.key, None)
+        if fire_event:
+            self.dispatch.dispose_collection(state, collection, adapter)
+
+    def _invalidate_collection(
+        self, collection: _AdaptedCollectionProtocol
+    ) -> None:
+        adapter = getattr(collection, "_sa_adapter")
+        adapter.invalidated = True
+
+    def set_committed_value(
+        self, state: InstanceState[Any], dict_: _InstanceDict, value: Any
+    ) -> _AdaptedCollectionProtocol:
+        """Set an attribute value on the given instance and 'commit' it."""
+
+        collection, user_data = self._initialize_collection(state)
+
+        if value:
+            collection.append_multiple_without_event(value)
+
+        state.dict[self.key] = user_data
+
+        state._commit(dict_, [self.key])
+
+        if self.key in state._pending_mutations:
+            # pending items exist.  issue a modified event,
+            # add/remove new items.
+            state._modified_event(dict_, self, user_data, True)
+
+            pending = state._pending_mutations.pop(self.key)
+            added = pending.added_items
+            removed = pending.deleted_items
+            for item in added:
+                collection.append_without_event(item)
+            for item in removed:
+                collection.remove_without_event(item)
+
+        return user_data
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Literal[None] = ...,
+        passive: Literal[PassiveFlag.PASSIVE_OFF] = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: _AdaptedCollectionProtocol = ...,
+        passive: PassiveFlag = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = ...,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]: ...
+
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = None,
+        passive: PassiveFlag = PASSIVE_OFF,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]:
+        """Retrieve the CollectionAdapter associated with the given state.
+
+        if user_data is None, retrieves it from the state using normal
+        "get()" rules, which will fire lazy callables or return the "empty"
+        collection value.
+
+        """
+        if user_data is None:
+            fetch_user_data = self.get(state, dict_, passive=passive)
+            if fetch_user_data is LoaderCallableStatus.PASSIVE_NO_RESULT:
+                return fetch_user_data
+            else:
+                user_data = cast("_AdaptedCollectionProtocol", fetch_user_data)
+
+        return user_data._sa_adapter
+
+
+def backref_listeners(
+    attribute: QueryableAttribute[Any], key: str, uselist: bool
+) -> None:
+    """Apply listeners to synchronize a two-way relationship."""
+
+    # use easily recognizable names for stack traces.
+
+    # in the sections marked "tokens to test for a recursive loop",
+    # this is somewhat brittle and very performance-sensitive logic
+    # that is specific to how we might arrive at each event.  a marker
+    # that can target us directly to arguments being invoked against
+    # the impl might be simpler, but could interfere with other systems.
+
+    parent_token = attribute.impl.parent_token
+    parent_impl = attribute.impl
+
+    def _acceptable_key_err(child_state, initiator, child_impl):
+        raise ValueError(
+            "Bidirectional attribute conflict detected: "
+            'Passing object %s to attribute "%s" '
+            'triggers a modify event on attribute "%s" '
+            'via the backref "%s".'
+            % (
+                state_str(child_state),
+                initiator.parent_token,
+                child_impl.parent_token,
+                attribute.impl.parent_token,
+            )
+        )
+
+    def emit_backref_from_scalar_set_event(
+        state, child, oldchild, initiator, **kw
+    ):
+        if oldchild is child:
+            return child
+        if (
+            oldchild is not None
+            and oldchild is not PASSIVE_NO_RESULT
+            and oldchild is not NO_VALUE
+        ):
+            # With lazy=None, there's no guarantee that the full collection is
+            # present when updating via a backref.
+            old_state, old_dict = (
+                instance_state(oldchild),
+                instance_dict(oldchild),
+            )
+            impl = old_state.manager[key].impl
+
+            # tokens to test for a recursive loop.
+            if not impl.collection and not impl.dynamic:
+                check_recursive_token = impl._replace_token
+            else:
+                check_recursive_token = impl._remove_token
+
+            if initiator is not check_recursive_token:
+                impl.pop(
+                    old_state,
+                    old_dict,
+                    state.obj(),
+                    parent_impl._append_token,
+                    passive=PASSIVE_NO_FETCH,
+                )
+
+        if child is not None:
+            child_state, child_dict = (
+                instance_state(child),
+                instance_dict(child),
+            )
+            child_impl = child_state.manager[key].impl
+
+            if (
+                initiator.parent_token is not parent_token
+                and initiator.parent_token is not child_impl.parent_token
+            ):
+                _acceptable_key_err(state, initiator, child_impl)
+
+            # tokens to test for a recursive loop.
+            check_append_token = child_impl._append_token
+            check_bulk_replace_token = (
+                child_impl._bulk_replace_token
+                if _is_collection_attribute_impl(child_impl)
+                else None
+            )
+
+            if (
+                initiator is not check_append_token
+                and initiator is not check_bulk_replace_token
+            ):
+                child_impl.append(
+                    child_state,
+                    child_dict,
+                    state.obj(),
+                    initiator,
+                    passive=PASSIVE_NO_FETCH,
+                )
+        return child
+
+    def emit_backref_from_collection_append_event(
+        state, child, initiator, **kw
+    ):
+        if child is None:
+            return
+
+        child_state, child_dict = instance_state(child), instance_dict(child)
+        child_impl = child_state.manager[key].impl
+
+        if (
+            initiator.parent_token is not parent_token
+            and initiator.parent_token is not child_impl.parent_token
+        ):
+            _acceptable_key_err(state, initiator, child_impl)
+
+        # tokens to test for a recursive loop.
+        check_append_token = child_impl._append_token
+        check_bulk_replace_token = (
+            child_impl._bulk_replace_token
+            if _is_collection_attribute_impl(child_impl)
+            else None
+        )
+
+        if (
+            initiator is not check_append_token
+            and initiator is not check_bulk_replace_token
+        ):
+            child_impl.append(
+                child_state,
+                child_dict,
+                state.obj(),
+                initiator,
+                passive=PASSIVE_NO_FETCH,
+            )
+        return child
+
+    def emit_backref_from_collection_remove_event(
+        state, child, initiator, **kw
+    ):
+        if (
+            child is not None
+            and child is not PASSIVE_NO_RESULT
+            and child is not NO_VALUE
+        ):
+            child_state, child_dict = (
+                instance_state(child),
+                instance_dict(child),
+            )
+            child_impl = child_state.manager[key].impl
+
+            check_replace_token: Optional[AttributeEventToken]
+
+            # tokens to test for a recursive loop.
+            if not child_impl.collection and not child_impl.dynamic:
+                check_remove_token = child_impl._remove_token
+                check_replace_token = child_impl._replace_token
+                check_for_dupes_on_remove = uselist and not parent_impl.dynamic
+            else:
+                check_remove_token = child_impl._remove_token
+                check_replace_token = (
+                    child_impl._bulk_replace_token
+                    if _is_collection_attribute_impl(child_impl)
+                    else None
+                )
+                check_for_dupes_on_remove = False
+
+            if (
+                initiator is not check_remove_token
+                and initiator is not check_replace_token
+            ):
+                if not check_for_dupes_on_remove or not util.has_dupes(
+                    # when this event is called, the item is usually
+                    # present in the list, except for a pop() operation.
+                    state.dict[parent_impl.key],
+                    child,
+                ):
+                    child_impl.pop(
+                        child_state,
+                        child_dict,
+                        state.obj(),
+                        initiator,
+                        passive=PASSIVE_NO_FETCH,
+                    )
+
+    if uselist:
+        event.listen(
+            attribute,
+            "append",
+            emit_backref_from_collection_append_event,
+            retval=True,
+            raw=True,
+            include_key=True,
+        )
+    else:
+        event.listen(
+            attribute,
+            "set",
+            emit_backref_from_scalar_set_event,
+            retval=True,
+            raw=True,
+            include_key=True,
+        )
+    # TODO: need coverage in test/orm/ of remove event
+    event.listen(
+        attribute,
+        "remove",
+        emit_backref_from_collection_remove_event,
+        retval=True,
+        raw=True,
+        include_key=True,
+    )
+
+
+_NO_HISTORY = util.symbol("NO_HISTORY")
+_NO_STATE_SYMBOLS = frozenset([id(PASSIVE_NO_RESULT), id(NO_VALUE)])
+
+
+class History(NamedTuple):
+    """A 3-tuple of added, unchanged and deleted values,
+    representing the changes which have occurred on an instrumented
+    attribute.
+
+    The easiest way to get a :class:`.History` object for a particular
+    attribute on an object is to use the :func:`_sa.inspect` function::
+
+        from sqlalchemy import inspect
+
+        hist = inspect(myobject).attrs.myattribute.history
+
+    Each tuple member is an iterable sequence:
+
+    * ``added`` - the collection of items added to the attribute (the first
+      tuple element).
+
+    * ``unchanged`` - the collection of items that have not changed on the
+      attribute (the second tuple element).
+
+    * ``deleted`` - the collection of items that have been removed from the
+      attribute (the third tuple element).
+
+    """
+
+    added: Union[Tuple[()], List[Any]]
+    unchanged: Union[Tuple[()], List[Any]]
+    deleted: Union[Tuple[()], List[Any]]
+
+    def __bool__(self) -> bool:
+        return self != HISTORY_BLANK
+
+    def empty(self) -> bool:
+        """Return True if this :class:`.History` has no changes
+        and no existing, unchanged state.
+
+        """
+
+        return not bool((self.added or self.deleted) or self.unchanged)
+
+    def sum(self) -> Sequence[Any]:
+        """Return a collection of added + unchanged + deleted."""
+
+        return (
+            (self.added or []) + (self.unchanged or []) + (self.deleted or [])
+        )
+
+    def non_deleted(self) -> Sequence[Any]:
+        """Return a collection of added + unchanged."""
+
+        return (self.added or []) + (self.unchanged or [])
+
+    def non_added(self) -> Sequence[Any]:
+        """Return a collection of unchanged + deleted."""
+
+        return (self.unchanged or []) + (self.deleted or [])
+
+    def has_changes(self) -> bool:
+        """Return True if this :class:`.History` has changes."""
+
+        return bool(self.added or self.deleted)
+
+    def _merge(self, added: Iterable[Any], deleted: Iterable[Any]) -> History:
+        return History(
+            list(self.added) + list(added),
+            self.unchanged,
+            list(self.deleted) + list(deleted),
+        )
+
+    def as_state(self) -> History:
+        return History(
+            [
+                (c is not None) and instance_state(c) or None
+                for c in self.added
+            ],
+            [
+                (c is not None) and instance_state(c) or None
+                for c in self.unchanged
+            ],
+            [
+                (c is not None) and instance_state(c) or None
+                for c in self.deleted
+            ],
+        )
+
+    @classmethod
+    def from_scalar_attribute(
+        cls,
+        attribute: ScalarAttributeImpl,
+        state: InstanceState[Any],
+        current: Any,
+    ) -> History:
+        original = state.committed_state.get(attribute.key, _NO_HISTORY)
+
+        deleted: Union[Tuple[()], List[Any]]
+
+        if original is _NO_HISTORY:
+            if current is NO_VALUE:
+                return cls((), (), ())
+            else:
+                return cls((), [current], ())
+        # don't let ClauseElement expressions here trip things up
+        elif (
+            current is not NO_VALUE
+            and attribute.is_equal(current, original) is True
+        ):
+            return cls((), [current], ())
+        else:
+            # current convention on native scalars is to not
+            # include information
+            # about missing previous value in "deleted", but
+            # we do include None, which helps in some primary
+            # key situations
+            if id(original) in _NO_STATE_SYMBOLS:
+                deleted = ()
+                # indicate a "del" operation occurred when we don't have
+                # the previous value as: ([None], (), ())
+                if id(current) in _NO_STATE_SYMBOLS:
+                    current = None
+            else:
+                deleted = [original]
+            if current is NO_VALUE:
+                return cls((), (), deleted)
+            else:
+                return cls([current], (), deleted)
+
+    @classmethod
+    def from_object_attribute(
+        cls,
+        attribute: ScalarObjectAttributeImpl,
+        state: InstanceState[Any],
+        current: Any,
+        original: Any = _NO_HISTORY,
+    ) -> History:
+        deleted: Union[Tuple[()], List[Any]]
+
+        if original is _NO_HISTORY:
+            original = state.committed_state.get(attribute.key, _NO_HISTORY)
+
+        if original is _NO_HISTORY:
+            if current is NO_VALUE:
+                return cls((), (), ())
+            else:
+                return cls((), [current], ())
+        elif current is original and current is not NO_VALUE:
+            return cls((), [current], ())
+        else:
+            # current convention on related objects is to not
+            # include information
+            # about missing previous value in "deleted", and
+            # to also not include None - the dependency.py rules
+            # ignore the None in any case.
+            if id(original) in _NO_STATE_SYMBOLS or original is None:
+                deleted = ()
+                # indicate a "del" operation occurred when we don't have
+                # the previous value as: ([None], (), ())
+                if id(current) in _NO_STATE_SYMBOLS:
+                    current = None
+            else:
+                deleted = [original]
+            if current is NO_VALUE:
+                return cls((), (), deleted)
+            else:
+                return cls([current], (), deleted)
+
+    @classmethod
+    def from_collection(
+        cls,
+        attribute: CollectionAttributeImpl,
+        state: InstanceState[Any],
+        current: Any,
+    ) -> History:
+        original = state.committed_state.get(attribute.key, _NO_HISTORY)
+        if current is NO_VALUE:
+            return cls((), (), ())
+
+        current = getattr(current, "_sa_adapter")
+        if original is NO_VALUE:
+            return cls(list(current), (), ())
+        elif original is _NO_HISTORY:
+            return cls((), list(current), ())
+        else:
+            current_states = [
+                ((c is not None) and instance_state(c) or None, c)
+                for c in current
+            ]
+            original_states = [
+                ((c is not None) and instance_state(c) or None, c)
+                for c in original
+            ]
+
+            current_set = dict(current_states)
+            original_set = dict(original_states)
+
+            return cls(
+                [o for s, o in current_states if s not in original_set],
+                [o for s, o in current_states if s in original_set],
+                [o for s, o in original_states if s not in current_set],
+            )
+
+
+HISTORY_BLANK = History((), (), ())
+
+
+def get_history(
+    obj: object, key: str, passive: PassiveFlag = PASSIVE_OFF
+) -> History:
+    """Return a :class:`.History` record for the given object
+    and attribute key.
+
+    This is the **pre-flush** history for a given attribute, which is
+    reset each time the :class:`.Session` flushes changes to the
+    current database transaction.
+
+    .. note::
+
+        Prefer to use the :attr:`.AttributeState.history` and
+        :meth:`.AttributeState.load_history` accessors to retrieve the
+        :class:`.History` for instance attributes.
+
+
+    :param obj: an object whose class is instrumented by the
+      attributes package.
+
+    :param key: string attribute name.
+
+    :param passive: indicates loading behavior for the attribute
+       if the value is not already present.   This is a
+       bitflag attribute, which defaults to the symbol
+       :attr:`.PASSIVE_OFF` indicating all necessary SQL
+       should be emitted.
+
+    .. seealso::
+
+        :attr:`.AttributeState.history`
+
+        :meth:`.AttributeState.load_history` - retrieve history
+        using loader callables if the value is not locally present.
+
+    """
+
+    return get_state_history(instance_state(obj), key, passive)
+
+
+def get_state_history(
+    state: InstanceState[Any], key: str, passive: PassiveFlag = PASSIVE_OFF
+) -> History:
+    return state.get_history(key, passive)
+
+
+def has_parent(
+    cls: Type[_O], obj: _O, key: str, optimistic: bool = False
+) -> bool:
+    """TODO"""
+    manager = manager_of_class(cls)
+    state = instance_state(obj)
+    return manager.has_parent(state, key, optimistic)
+
+
+def register_attribute(
+    class_: Type[_O],
+    key: str,
+    *,
+    comparator: interfaces.PropComparator[_T],
+    parententity: _InternalEntityType[_O],
+    doc: Optional[str] = None,
+    **kw: Any,
+) -> InstrumentedAttribute[_T]:
+    desc = register_descriptor(
+        class_, key, comparator=comparator, parententity=parententity, doc=doc
+    )
+    register_attribute_impl(class_, key, **kw)
+    return desc
+
+
+def register_attribute_impl(
+    class_: Type[_O],
+    key: str,
+    uselist: bool = False,
+    callable_: Optional[_LoaderCallable] = None,
+    useobject: bool = False,
+    impl_class: Optional[Type[AttributeImpl]] = None,
+    backref: Optional[str] = None,
+    **kw: Any,
+) -> QueryableAttribute[Any]:
+    manager = manager_of_class(class_)
+    if uselist:
+        factory = kw.pop("typecallable", None)
+        typecallable = manager.instrument_collection_class(
+            key, factory or list
+        )
+    else:
+        typecallable = kw.pop("typecallable", None)
+
+    dispatch = cast(
+        "_Dispatch[QueryableAttribute[Any]]", manager[key].dispatch
+    )  # noqa: E501
+
+    impl: AttributeImpl
+
+    if impl_class:
+        # TODO: this appears to be the WriteOnlyAttributeImpl /
+        # DynamicAttributeImpl constructor which is hardcoded
+        impl = cast("Type[WriteOnlyAttributeImpl]", impl_class)(
+            class_, key, dispatch, **kw
+        )
+    elif uselist:
+        impl = CollectionAttributeImpl(
+            class_, key, callable_, dispatch, typecallable=typecallable, **kw
+        )
+    elif useobject:
+        impl = ScalarObjectAttributeImpl(
+            class_, key, callable_, dispatch, **kw
+        )
+    else:
+        impl = ScalarAttributeImpl(class_, key, callable_, dispatch, **kw)
+
+    manager[key].impl = impl
+
+    if backref:
+        backref_listeners(manager[key], backref, uselist)
+
+    manager.post_configure_attribute(key)
+    return manager[key]
+
+
+def register_descriptor(
+    class_: Type[Any],
+    key: str,
+    *,
+    comparator: interfaces.PropComparator[_T],
+    parententity: _InternalEntityType[Any],
+    doc: Optional[str] = None,
+) -> InstrumentedAttribute[_T]:
+    manager = manager_of_class(class_)
+
+    descriptor = InstrumentedAttribute(
+        class_, key, comparator=comparator, parententity=parententity
+    )
+
+    descriptor.__doc__ = doc  # type: ignore
+
+    manager.instrument_attribute(key, descriptor)
+    return descriptor
+
+
+def unregister_attribute(class_: Type[Any], key: str) -> None:
+    manager_of_class(class_).uninstrument_attribute(key)
+
+
+def init_collection(obj: object, key: str) -> CollectionAdapter:
+    """Initialize a collection attribute and return the collection adapter.
+
+    This function is used to provide direct access to collection internals
+    for a previously unloaded attribute.  e.g.::
+
+        collection_adapter = init_collection(someobject, "elements")
+        for elem in values:
+            collection_adapter.append_without_event(elem)
+
+    For an easier way to do the above, see
+    :func:`~sqlalchemy.orm.attributes.set_committed_value`.
+
+    :param obj: a mapped object
+
+    :param key: string attribute name where the collection is located.
+
+    """
+    state = instance_state(obj)
+    dict_ = state.dict
+    return init_state_collection(state, dict_, key)
+
+
+def init_state_collection(
+    state: InstanceState[Any], dict_: _InstanceDict, key: str
+) -> CollectionAdapter:
+    """Initialize a collection attribute and return the collection adapter.
+
+    Discards any existing collection which may be there.
+
+    """
+    attr = state.manager[key].impl
+
+    if TYPE_CHECKING:
+        assert isinstance(attr, HasCollectionAdapter)
+
+    old = dict_.pop(key, None)  # discard old collection
+    if old is not None:
+        old_collection = old._sa_adapter
+        attr._dispose_previous_collection(state, old, old_collection, False)
+
+    user_data = attr._default_value(state, dict_)
+    adapter: CollectionAdapter = attr.get_collection(
+        state, dict_, user_data, passive=PassiveFlag.PASSIVE_NO_FETCH
+    )
+    adapter._reset_empty()
+
+    return adapter
+
+
+def set_committed_value(instance, key, value):
+    """Set the value of an attribute with no history events.
+
+    Cancels any previous history present.  The value should be
+    a scalar value for scalar-holding attributes, or
+    an iterable for any collection-holding attribute.
+
+    This is the same underlying method used when a lazy loader
+    fires off and loads additional data from the database.
+    In particular, this method can be used by application code
+    which has loaded additional attributes or collections through
+    separate queries, which can then be attached to an instance
+    as though it were part of its original loaded state.
+
+    """
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    state.manager[key].impl.set_committed_value(state, dict_, value)
+
+
+def set_attribute(
+    instance: object,
+    key: str,
+    value: Any,
+    initiator: Optional[AttributeEventToken] = None,
+) -> None:
+    """Set the value of an attribute, firing history events.
+
+    This function may be used regardless of instrumentation
+    applied directly to the class, i.e. no descriptors are required.
+    Custom attribute management schemes will need to make usage
+    of this method to establish attribute state as understood
+    by SQLAlchemy.
+
+    :param instance: the object that will be modified
+
+    :param key: string name of the attribute
+
+    :param value: value to assign
+
+    :param initiator: an instance of :class:`.Event` that would have
+     been propagated from a previous event listener.  This argument
+     is used when the :func:`.set_attribute` function is being used within
+     an existing event listening function where an :class:`.Event` object
+     is being supplied; the object may be used to track the origin of the
+     chain of events.
+
+     .. versionadded:: 1.2.3
+
+    """
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    state.manager[key].impl.set(state, dict_, value, initiator)
+
+
+def get_attribute(instance: object, key: str) -> Any:
+    """Get the value of an attribute, firing any callables required.
+
+    This function may be used regardless of instrumentation
+    applied directly to the class, i.e. no descriptors are required.
+    Custom attribute management schemes will need to make usage
+    of this method to make usage of attribute state as understood
+    by SQLAlchemy.
+
+    """
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    return state.manager[key].impl.get(state, dict_)
+
+
+def del_attribute(instance: object, key: str) -> None:
+    """Delete the value of an attribute, firing history events.
+
+    This function may be used regardless of instrumentation
+    applied directly to the class, i.e. no descriptors are required.
+    Custom attribute management schemes will need to make usage
+    of this method to establish attribute state as understood
+    by SQLAlchemy.
+
+    """
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    state.manager[key].impl.delete(state, dict_)
+
+
+def flag_modified(instance: object, key: str) -> None:
+    """Mark an attribute on an instance as 'modified'.
+
+    This sets the 'modified' flag on the instance and
+    establishes an unconditional change event for the given attribute.
+    The attribute must have a value present, else an
+    :class:`.InvalidRequestError` is raised.
+
+    To mark an object "dirty" without referring to any specific attribute
+    so that it is considered within a flush, use the
+    :func:`.attributes.flag_dirty` call.
+
+    .. seealso::
+
+        :func:`.attributes.flag_dirty`
+
+    """
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    impl = state.manager[key].impl
+    impl.dispatch.modified(state, impl._modified_token)
+    state._modified_event(dict_, impl, NO_VALUE, is_userland=True)
+
+
+def flag_dirty(instance: object) -> None:
+    """Mark an instance as 'dirty' without any specific attribute mentioned.
+
+    This is a special operation that will allow the object to travel through
+    the flush process for interception by events such as
+    :meth:`.SessionEvents.before_flush`.   Note that no SQL will be emitted in
+    the flush process for an object that has no changes, even if marked dirty
+    via this method.  However, a :meth:`.SessionEvents.before_flush` handler
+    will be able to see the object in the :attr:`.Session.dirty` collection and
+    may establish changes on it, which will then be included in the SQL
+    emitted.
+
+    .. versionadded:: 1.2
+
+    .. seealso::
+
+        :func:`.attributes.flag_modified`
+
+    """
+
+    state, dict_ = instance_state(instance), instance_dict(instance)
+    state._modified_event(dict_, None, NO_VALUE, is_userland=True)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/base.py
new file mode 100644
index 00000000..ae0ba102
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/base.py
@@ -0,0 +1,973 @@
+# orm/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Constants and rudimental functions used throughout the ORM.
+
+"""
+
+from __future__ import annotations
+
+from enum import Enum
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import no_type_check
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import exc
+from ._typing import insp_is_mapper
+from .. import exc as sa_exc
+from .. import inspection
+from .. import util
+from ..sql import roles
+from ..sql.elements import SQLColumnExpression
+from ..sql.elements import SQLCoreOperations
+from ..util import FastIntFlag
+from ..util.langhelpers import TypingOnly
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _ExternalEntityType
+    from ._typing import _InternalEntityType
+    from .attributes import InstrumentedAttribute
+    from .dynamic import AppenderQuery
+    from .instrumentation import ClassManager
+    from .interfaces import PropComparator
+    from .mapper import Mapper
+    from .state import InstanceState
+    from .util import AliasedClass
+    from .writeonly import WriteOnlyCollection
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _InfoType
+    from ..sql.elements import ColumnElement
+    from ..sql.operators import OperatorType
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+_O = TypeVar("_O", bound=object)
+
+
+class LoaderCallableStatus(Enum):
+    PASSIVE_NO_RESULT = 0
+    """Symbol returned by a loader callable or other attribute/history
+    retrieval operation when a value could not be determined, based
+    on loader callable flags.
+    """
+
+    PASSIVE_CLASS_MISMATCH = 1
+    """Symbol indicating that an object is locally present for a given
+    primary key identity but it is not of the requested class.  The
+    return value is therefore None and no SQL should be emitted."""
+
+    ATTR_WAS_SET = 2
+    """Symbol returned by a loader callable to indicate the
+    retrieved value, or values, were assigned to their attributes
+    on the target object.
+    """
+
+    ATTR_EMPTY = 3
+    """Symbol used internally to indicate an attribute had no callable."""
+
+    NO_VALUE = 4
+    """Symbol which may be placed as the 'previous' value of an attribute,
+    indicating no value was loaded for an attribute when it was modified,
+    and flags indicated we were not to load it.
+    """
+
+    NEVER_SET = NO_VALUE
+    """
+    Synonymous with NO_VALUE
+
+    .. versionchanged:: 1.4   NEVER_SET was merged with NO_VALUE
+
+    """
+
+
+(
+    PASSIVE_NO_RESULT,
+    PASSIVE_CLASS_MISMATCH,
+    ATTR_WAS_SET,
+    ATTR_EMPTY,
+    NO_VALUE,
+) = tuple(LoaderCallableStatus)
+
+NEVER_SET = NO_VALUE
+
+
+class PassiveFlag(FastIntFlag):
+    """Bitflag interface that passes options onto loader callables"""
+
+    NO_CHANGE = 0
+    """No callables or SQL should be emitted on attribute access
+    and no state should change
+    """
+
+    CALLABLES_OK = 1
+    """Loader callables can be fired off if a value
+    is not present.
+    """
+
+    SQL_OK = 2
+    """Loader callables can emit SQL at least on scalar value attributes."""
+
+    RELATED_OBJECT_OK = 4
+    """Callables can use SQL to load related objects as well
+    as scalar value attributes.
+    """
+
+    INIT_OK = 8
+    """Attributes should be initialized with a blank
+    value (None or an empty collection) upon get, if no other
+    value can be obtained.
+    """
+
+    NON_PERSISTENT_OK = 16
+    """Callables can be emitted if the parent is not persistent."""
+
+    LOAD_AGAINST_COMMITTED = 32
+    """Callables should use committed values as primary/foreign keys during a
+    load.
+    """
+
+    NO_AUTOFLUSH = 64
+    """Loader callables should disable autoflush."""
+
+    NO_RAISE = 128
+    """Loader callables should not raise any assertions"""
+
+    DEFERRED_HISTORY_LOAD = 256
+    """indicates special load of the previous value of an attribute"""
+
+    INCLUDE_PENDING_MUTATIONS = 512
+
+    # pre-packaged sets of flags used as inputs
+    PASSIVE_OFF = (
+        RELATED_OBJECT_OK | NON_PERSISTENT_OK | INIT_OK | CALLABLES_OK | SQL_OK
+    )
+    "Callables can be emitted in all cases."
+
+    PASSIVE_RETURN_NO_VALUE = PASSIVE_OFF ^ INIT_OK
+    """PASSIVE_OFF ^ INIT_OK"""
+
+    PASSIVE_NO_INITIALIZE = PASSIVE_RETURN_NO_VALUE ^ CALLABLES_OK
+    "PASSIVE_RETURN_NO_VALUE ^ CALLABLES_OK"
+
+    PASSIVE_NO_FETCH = PASSIVE_OFF ^ SQL_OK
+    "PASSIVE_OFF ^ SQL_OK"
+
+    PASSIVE_NO_FETCH_RELATED = PASSIVE_OFF ^ RELATED_OBJECT_OK
+    "PASSIVE_OFF ^ RELATED_OBJECT_OK"
+
+    PASSIVE_ONLY_PERSISTENT = PASSIVE_OFF ^ NON_PERSISTENT_OK
+    "PASSIVE_OFF ^ NON_PERSISTENT_OK"
+
+    PASSIVE_MERGE = PASSIVE_OFF | NO_RAISE
+    """PASSIVE_OFF | NO_RAISE
+
+    Symbol used specifically for session.merge() and similar cases
+
+    """
+
+
+(
+    NO_CHANGE,
+    CALLABLES_OK,
+    SQL_OK,
+    RELATED_OBJECT_OK,
+    INIT_OK,
+    NON_PERSISTENT_OK,
+    LOAD_AGAINST_COMMITTED,
+    NO_AUTOFLUSH,
+    NO_RAISE,
+    DEFERRED_HISTORY_LOAD,
+    INCLUDE_PENDING_MUTATIONS,
+    PASSIVE_OFF,
+    PASSIVE_RETURN_NO_VALUE,
+    PASSIVE_NO_INITIALIZE,
+    PASSIVE_NO_FETCH,
+    PASSIVE_NO_FETCH_RELATED,
+    PASSIVE_ONLY_PERSISTENT,
+    PASSIVE_MERGE,
+) = PassiveFlag.__members__.values()
+
+DEFAULT_MANAGER_ATTR = "_sa_class_manager"
+DEFAULT_STATE_ATTR = "_sa_instance_state"
+
+
+class EventConstants(Enum):
+    EXT_CONTINUE = 1
+    EXT_STOP = 2
+    EXT_SKIP = 3
+    NO_KEY = 4
+    """indicates an :class:`.AttributeEvent` event that did not have any
+    key argument.
+
+    .. versionadded:: 2.0
+
+    """
+
+
+EXT_CONTINUE, EXT_STOP, EXT_SKIP, NO_KEY = tuple(EventConstants)
+
+
+class RelationshipDirection(Enum):
+    """enumeration which indicates the 'direction' of a
+    :class:`_orm.RelationshipProperty`.
+
+    :class:`.RelationshipDirection` is accessible from the
+    :attr:`_orm.Relationship.direction` attribute of
+    :class:`_orm.RelationshipProperty`.
+
+    """
+
+    ONETOMANY = 1
+    """Indicates the one-to-many direction for a :func:`_orm.relationship`.
+
+    This symbol is typically used by the internals but may be exposed within
+    certain API features.
+
+    """
+
+    MANYTOONE = 2
+    """Indicates the many-to-one direction for a :func:`_orm.relationship`.
+
+    This symbol is typically used by the internals but may be exposed within
+    certain API features.
+
+    """
+
+    MANYTOMANY = 3
+    """Indicates the many-to-many direction for a :func:`_orm.relationship`.
+
+    This symbol is typically used by the internals but may be exposed within
+    certain API features.
+
+    """
+
+
+ONETOMANY, MANYTOONE, MANYTOMANY = tuple(RelationshipDirection)
+
+
+class InspectionAttrExtensionType(Enum):
+    """Symbols indicating the type of extension that a
+    :class:`.InspectionAttr` is part of."""
+
+
+class NotExtension(InspectionAttrExtensionType):
+    NOT_EXTENSION = "not_extension"
+    """Symbol indicating an :class:`InspectionAttr` that's
+    not part of sqlalchemy.ext.
+
+    Is assigned to the :attr:`.InspectionAttr.extension_type`
+    attribute.
+
+    """
+
+
+_never_set = frozenset([NEVER_SET])
+
+_none_set = frozenset([None, NEVER_SET, PASSIVE_NO_RESULT])
+
+_none_only_set = frozenset([None])
+
+_SET_DEFERRED_EXPIRED = util.symbol("SET_DEFERRED_EXPIRED")
+
+_DEFER_FOR_STATE = util.symbol("DEFER_FOR_STATE")
+
+_RAISE_FOR_STATE = util.symbol("RAISE_FOR_STATE")
+
+
+_F = TypeVar("_F", bound=Callable[..., Any])
+_Self = TypeVar("_Self")
+
+
+def _assertions(
+    *assertions: Any,
+) -> Callable[[_F], _F]:
+    @util.decorator
+    def generate(fn: _F, self: _Self, *args: Any, **kw: Any) -> _Self:
+        for assertion in assertions:
+            assertion(self, fn.__name__)
+        fn(self, *args, **kw)
+        return self
+
+    return generate
+
+
+if TYPE_CHECKING:
+
+    def manager_of_class(cls: Type[_O]) -> ClassManager[_O]: ...
+
+    @overload
+    def opt_manager_of_class(cls: AliasedClass[Any]) -> None: ...
+
+    @overload
+    def opt_manager_of_class(
+        cls: _ExternalEntityType[_O],
+    ) -> Optional[ClassManager[_O]]: ...
+
+    def opt_manager_of_class(
+        cls: _ExternalEntityType[_O],
+    ) -> Optional[ClassManager[_O]]: ...
+
+    def instance_state(instance: _O) -> InstanceState[_O]: ...
+
+    def instance_dict(instance: object) -> Dict[str, Any]: ...
+
+else:
+    # these can be replaced by sqlalchemy.ext.instrumentation
+    # if augmented class instrumentation is enabled.
+
+    def manager_of_class(cls):
+        try:
+            return cls.__dict__[DEFAULT_MANAGER_ATTR]
+        except KeyError as ke:
+            raise exc.UnmappedClassError(
+                cls, f"Can't locate an instrumentation manager for class {cls}"
+            ) from ke
+
+    def opt_manager_of_class(cls):
+        return cls.__dict__.get(DEFAULT_MANAGER_ATTR)
+
+    instance_state = operator.attrgetter(DEFAULT_STATE_ATTR)
+
+    instance_dict = operator.attrgetter("__dict__")
+
+
+def instance_str(instance: object) -> str:
+    """Return a string describing an instance."""
+
+    return state_str(instance_state(instance))
+
+
+def state_str(state: InstanceState[Any]) -> str:
+    """Return a string describing an instance via its InstanceState."""
+
+    if state is None:
+        return "None"
+    else:
+        return "<%s at 0x%x>" % (state.class_.__name__, id(state.obj()))
+
+
+def state_class_str(state: InstanceState[Any]) -> str:
+    """Return a string describing an instance's class via its
+    InstanceState.
+    """
+
+    if state is None:
+        return "None"
+    else:
+        return "<%s>" % (state.class_.__name__,)
+
+
+def attribute_str(instance: object, attribute: str) -> str:
+    return instance_str(instance) + "." + attribute
+
+
+def state_attribute_str(state: InstanceState[Any], attribute: str) -> str:
+    return state_str(state) + "." + attribute
+
+
+def object_mapper(instance: _T) -> Mapper[_T]:
+    """Given an object, return the primary Mapper associated with the object
+    instance.
+
+    Raises :class:`sqlalchemy.orm.exc.UnmappedInstanceError`
+    if no mapping is configured.
+
+    This function is available via the inspection system as::
+
+        inspect(instance).mapper
+
+    Using the inspection system will raise
+    :class:`sqlalchemy.exc.NoInspectionAvailable` if the instance is
+    not part of a mapping.
+
+    """
+    return object_state(instance).mapper
+
+
+def object_state(instance: _T) -> InstanceState[_T]:
+    """Given an object, return the :class:`.InstanceState`
+    associated with the object.
+
+    Raises :class:`sqlalchemy.orm.exc.UnmappedInstanceError`
+    if no mapping is configured.
+
+    Equivalent functionality is available via the :func:`_sa.inspect`
+    function as::
+
+        inspect(instance)
+
+    Using the inspection system will raise
+    :class:`sqlalchemy.exc.NoInspectionAvailable` if the instance is
+    not part of a mapping.
+
+    """
+    state = _inspect_mapped_object(instance)
+    if state is None:
+        raise exc.UnmappedInstanceError(instance)
+    else:
+        return state
+
+
+@inspection._inspects(object)
+def _inspect_mapped_object(instance: _T) -> Optional[InstanceState[_T]]:
+    try:
+        return instance_state(instance)
+    except (exc.UnmappedClassError,) + exc.NO_STATE:
+        return None
+
+
+def _class_to_mapper(
+    class_or_mapper: Union[Mapper[_T], Type[_T]]
+) -> Mapper[_T]:
+    # can't get mypy to see an overload for this
+    insp = inspection.inspect(class_or_mapper, False)
+    if insp is not None:
+        return insp.mapper  # type: ignore
+    else:
+        assert isinstance(class_or_mapper, type)
+        raise exc.UnmappedClassError(class_or_mapper)
+
+
+def _mapper_or_none(
+    entity: Union[Type[_T], _InternalEntityType[_T]]
+) -> Optional[Mapper[_T]]:
+    """Return the :class:`_orm.Mapper` for the given class or None if the
+    class is not mapped.
+    """
+
+    # can't get mypy to see an overload for this
+    insp = inspection.inspect(entity, False)
+    if insp is not None:
+        return insp.mapper  # type: ignore
+    else:
+        return None
+
+
+def _is_mapped_class(entity: Any) -> bool:
+    """Return True if the given object is a mapped class,
+    :class:`_orm.Mapper`, or :class:`.AliasedClass`.
+    """
+
+    insp = inspection.inspect(entity, False)
+    return (
+        insp is not None
+        and not insp.is_clause_element
+        and (insp.is_mapper or insp.is_aliased_class)
+    )
+
+
+def _is_aliased_class(entity: Any) -> bool:
+    insp = inspection.inspect(entity, False)
+    return insp is not None and getattr(insp, "is_aliased_class", False)
+
+
+@no_type_check
+def _entity_descriptor(entity: _EntityType[Any], key: str) -> Any:
+    """Return a class attribute given an entity and string name.
+
+    May return :class:`.InstrumentedAttribute` or user-defined
+    attribute.
+
+    """
+    insp = inspection.inspect(entity)
+    if insp.is_selectable:
+        description = entity
+        entity = insp.c
+    elif insp.is_aliased_class:
+        entity = insp.entity
+        description = entity
+    elif hasattr(insp, "mapper"):
+        description = entity = insp.mapper.class_
+    else:
+        description = entity
+
+    try:
+        return getattr(entity, key)
+    except AttributeError as err:
+        raise sa_exc.InvalidRequestError(
+            "Entity '%s' has no property '%s'" % (description, key)
+        ) from err
+
+
+if TYPE_CHECKING:
+
+    def _state_mapper(state: InstanceState[_O]) -> Mapper[_O]: ...
+
+else:
+    _state_mapper = util.dottedgetter("manager.mapper")
+
+
+def _inspect_mapped_class(
+    class_: Type[_O], configure: bool = False
+) -> Optional[Mapper[_O]]:
+    try:
+        class_manager = opt_manager_of_class(class_)
+        if class_manager is None or not class_manager.is_mapped:
+            return None
+        mapper = class_manager.mapper
+    except exc.NO_STATE:
+        return None
+    else:
+        if configure:
+            mapper._check_configure()
+        return mapper
+
+
+def _parse_mapper_argument(arg: Union[Mapper[_O], Type[_O]]) -> Mapper[_O]:
+    insp = inspection.inspect(arg, raiseerr=False)
+    if insp_is_mapper(insp):
+        return insp
+
+    raise sa_exc.ArgumentError(f"Mapper or mapped class expected, got {arg!r}")
+
+
+def class_mapper(class_: Type[_O], configure: bool = True) -> Mapper[_O]:
+    """Given a class, return the primary :class:`_orm.Mapper` associated
+    with the key.
+
+    Raises :exc:`.UnmappedClassError` if no mapping is configured
+    on the given class, or :exc:`.ArgumentError` if a non-class
+    object is passed.
+
+    Equivalent functionality is available via the :func:`_sa.inspect`
+    function as::
+
+        inspect(some_mapped_class)
+
+    Using the inspection system will raise
+    :class:`sqlalchemy.exc.NoInspectionAvailable` if the class is not mapped.
+
+    """
+    mapper = _inspect_mapped_class(class_, configure=configure)
+    if mapper is None:
+        if not isinstance(class_, type):
+            raise sa_exc.ArgumentError(
+                "Class object expected, got '%r'." % (class_,)
+            )
+        raise exc.UnmappedClassError(class_)
+    else:
+        return mapper
+
+
+class InspectionAttr:
+    """A base class applied to all ORM objects and attributes that are
+    related to things that can be returned by the :func:`_sa.inspect` function.
+
+    The attributes defined here allow the usage of simple boolean
+    checks to test basic facts about the object returned.
+
+    While the boolean checks here are basically the same as using
+    the Python isinstance() function, the flags here can be used without
+    the need to import all of these classes, and also such that
+    the SQLAlchemy class system can change while leaving the flags
+    here intact for forwards-compatibility.
+
+    """
+
+    __slots__: Tuple[str, ...] = ()
+
+    is_selectable = False
+    """Return True if this object is an instance of
+    :class:`_expression.Selectable`."""
+
+    is_aliased_class = False
+    """True if this object is an instance of :class:`.AliasedClass`."""
+
+    is_instance = False
+    """True if this object is an instance of :class:`.InstanceState`."""
+
+    is_mapper = False
+    """True if this object is an instance of :class:`_orm.Mapper`."""
+
+    is_bundle = False
+    """True if this object is an instance of :class:`.Bundle`."""
+
+    is_property = False
+    """True if this object is an instance of :class:`.MapperProperty`."""
+
+    is_attribute = False
+    """True if this object is a Python :term:`descriptor`.
+
+    This can refer to one of many types.   Usually a
+    :class:`.QueryableAttribute` which handles attributes events on behalf
+    of a :class:`.MapperProperty`.   But can also be an extension type
+    such as :class:`.AssociationProxy` or :class:`.hybrid_property`.
+    The :attr:`.InspectionAttr.extension_type` will refer to a constant
+    identifying the specific subtype.
+
+    .. seealso::
+
+        :attr:`_orm.Mapper.all_orm_descriptors`
+
+    """
+
+    _is_internal_proxy = False
+    """True if this object is an internal proxy object.
+
+    .. versionadded:: 1.2.12
+
+    """
+
+    is_clause_element = False
+    """True if this object is an instance of
+    :class:`_expression.ClauseElement`."""
+
+    extension_type: InspectionAttrExtensionType = NotExtension.NOT_EXTENSION
+    """The extension type, if any.
+    Defaults to :attr:`.interfaces.NotExtension.NOT_EXTENSION`
+
+    .. seealso::
+
+        :class:`.HybridExtensionType`
+
+        :class:`.AssociationProxyExtensionType`
+
+    """
+
+
+class InspectionAttrInfo(InspectionAttr):
+    """Adds the ``.info`` attribute to :class:`.InspectionAttr`.
+
+    The rationale for :class:`.InspectionAttr` vs. :class:`.InspectionAttrInfo`
+    is that the former is compatible as a mixin for classes that specify
+    ``__slots__``; this is essentially an implementation artifact.
+
+    """
+
+    __slots__ = ()
+
+    @util.ro_memoized_property
+    def info(self) -> _InfoType:
+        """Info dictionary associated with the object, allowing user-defined
+        data to be associated with this :class:`.InspectionAttr`.
+
+        The dictionary is generated when first accessed.  Alternatively,
+        it can be specified as a constructor argument to the
+        :func:`.column_property`, :func:`_orm.relationship`, or
+        :func:`.composite`
+        functions.
+
+        .. seealso::
+
+            :attr:`.QueryableAttribute.info`
+
+            :attr:`.SchemaItem.info`
+
+        """
+        return {}
+
+
+class SQLORMOperations(SQLCoreOperations[_T_co], TypingOnly):
+    __slots__ = ()
+
+    if typing.TYPE_CHECKING:
+
+        def of_type(
+            self, class_: _EntityType[Any]
+        ) -> PropComparator[_T_co]: ...
+
+        def and_(
+            self, *criteria: _ColumnExpressionArgument[bool]
+        ) -> PropComparator[bool]: ...
+
+        def any(  # noqa: A001
+            self,
+            criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+            **kwargs: Any,
+        ) -> ColumnElement[bool]: ...
+
+        def has(
+            self,
+            criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+            **kwargs: Any,
+        ) -> ColumnElement[bool]: ...
+
+
+class ORMDescriptor(Generic[_T_co], TypingOnly):
+    """Represent any Python descriptor that provides a SQL expression
+    construct at the class level."""
+
+    __slots__ = ()
+
+    if typing.TYPE_CHECKING:
+
+        @overload
+        def __get__(
+            self, instance: Any, owner: Literal[None]
+        ) -> ORMDescriptor[_T_co]: ...
+
+        @overload
+        def __get__(
+            self, instance: Literal[None], owner: Any
+        ) -> SQLCoreOperations[_T_co]: ...
+
+        @overload
+        def __get__(self, instance: object, owner: Any) -> _T_co: ...
+
+        def __get__(
+            self, instance: object, owner: Any
+        ) -> Union[ORMDescriptor[_T_co], SQLCoreOperations[_T_co], _T_co]: ...
+
+
+class _MappedAnnotationBase(Generic[_T_co], TypingOnly):
+    """common class for Mapped and similar ORM container classes.
+
+    these are classes that can appear on the left side of an ORM declarative
+    mapping, containing a mapped class or in some cases a collection
+    surrounding a mapped class.
+
+    """
+
+    __slots__ = ()
+
+
+class SQLORMExpression(
+    SQLORMOperations[_T_co], SQLColumnExpression[_T_co], TypingOnly
+):
+    """A type that may be used to indicate any ORM-level attribute or
+    object that acts in place of one, in the context of SQL expression
+    construction.
+
+    :class:`.SQLORMExpression` extends from the Core
+    :class:`.SQLColumnExpression` to add additional SQL methods that are ORM
+    specific, such as :meth:`.PropComparator.of_type`, and is part of the bases
+    for :class:`.InstrumentedAttribute`. It may be used in :pep:`484` typing to
+    indicate arguments or return values that should behave as ORM-level
+    attribute expressions.
+
+    .. versionadded:: 2.0.0b4
+
+
+    """
+
+    __slots__ = ()
+
+
+class Mapped(
+    SQLORMExpression[_T_co],
+    ORMDescriptor[_T_co],
+    _MappedAnnotationBase[_T_co],
+    roles.DDLConstraintColumnRole,
+):
+    """Represent an ORM mapped attribute on a mapped class.
+
+    This class represents the complete descriptor interface for any class
+    attribute that will have been :term:`instrumented` by the ORM
+    :class:`_orm.Mapper` class.   Provides appropriate information to type
+    checkers such as pylance and mypy so that ORM-mapped attributes
+    are correctly typed.
+
+    The most prominent use of :class:`_orm.Mapped` is in
+    the :ref:`Declarative Mapping <orm_explicit_declarative_base>` form
+    of :class:`_orm.Mapper` configuration, where used explicitly it drives
+    the configuration of ORM attributes such as :func:`_orm.mapped_class`
+    and :func:`_orm.relationship`.
+
+    .. seealso::
+
+        :ref:`orm_explicit_declarative_base`
+
+        :ref:`orm_declarative_table`
+
+    .. tip::
+
+        The :class:`_orm.Mapped` class represents attributes that are handled
+        directly by the :class:`_orm.Mapper` class. It does not include other
+        Python descriptor classes that are provided as extensions, including
+        :ref:`hybrids_toplevel` and the :ref:`associationproxy_toplevel`.
+        While these systems still make use of ORM-specific superclasses
+        and structures, they are not :term:`instrumented` by the
+        :class:`_orm.Mapper` and instead provide their own functionality
+        when they are accessed on a class.
+
+    .. versionadded:: 1.4
+
+
+    """
+
+    __slots__ = ()
+
+    if typing.TYPE_CHECKING:
+
+        @overload
+        def __get__(
+            self, instance: None, owner: Any
+        ) -> InstrumentedAttribute[_T_co]: ...
+
+        @overload
+        def __get__(self, instance: object, owner: Any) -> _T_co: ...
+
+        def __get__(
+            self, instance: Optional[object], owner: Any
+        ) -> Union[InstrumentedAttribute[_T_co], _T_co]: ...
+
+        @classmethod
+        def _empty_constructor(cls, arg1: Any) -> Mapped[_T_co]: ...
+
+        def __set__(
+            self, instance: Any, value: Union[SQLCoreOperations[_T_co], _T_co]
+        ) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
+
+
+class _MappedAttribute(Generic[_T_co], TypingOnly):
+    """Mixin for attributes which should be replaced by mapper-assigned
+    attributes.
+
+    """
+
+    __slots__ = ()
+
+
+class _DeclarativeMapped(Mapped[_T_co], _MappedAttribute[_T_co]):
+    """Mixin for :class:`.MapperProperty` subclasses that allows them to
+    be compatible with ORM-annotated declarative mappings.
+
+    """
+
+    __slots__ = ()
+
+    # MappedSQLExpression, Relationship, Composite etc. dont actually do
+    # SQL expression behavior.  yet there is code that compares them with
+    # __eq__(), __ne__(), etc.   Since #8847 made Mapped even more full
+    # featured including ColumnOperators, we need to have those methods
+    # be no-ops for these objects, so return NotImplemented to fall back
+    # to normal comparison behavior.
+    def operate(self, op: OperatorType, *other: Any, **kwargs: Any) -> Any:
+        return NotImplemented
+
+    __sa_operate__ = operate
+
+    def reverse_operate(
+        self, op: OperatorType, other: Any, **kwargs: Any
+    ) -> Any:
+        return NotImplemented
+
+
+class DynamicMapped(_MappedAnnotationBase[_T_co]):
+    """Represent the ORM mapped attribute type for a "dynamic" relationship.
+
+    The :class:`_orm.DynamicMapped` type annotation may be used in an
+    :ref:`Annotated Declarative Table <orm_declarative_mapped_column>` mapping
+    to indicate that the ``lazy="dynamic"`` loader strategy should be used
+    for a particular :func:`_orm.relationship`.
+
+    .. legacy::  The "dynamic" lazy loader strategy is the legacy form of what
+       is now the "write_only" strategy described in the section
+       :ref:`write_only_relationship`.
+
+    E.g.::
+
+        class User(Base):
+            __tablename__ = "user"
+            id: Mapped[int] = mapped_column(primary_key=True)
+            addresses: DynamicMapped[Address] = relationship(
+                cascade="all,delete-orphan"
+            )
+
+    See the section :ref:`dynamic_relationship` for background.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`dynamic_relationship` - complete background
+
+        :class:`.WriteOnlyMapped` - fully 2.0 style version
+
+    """
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        @overload
+        def __get__(
+            self, instance: None, owner: Any
+        ) -> InstrumentedAttribute[_T_co]: ...
+
+        @overload
+        def __get__(
+            self, instance: object, owner: Any
+        ) -> AppenderQuery[_T_co]: ...
+
+        def __get__(
+            self, instance: Optional[object], owner: Any
+        ) -> Union[InstrumentedAttribute[_T_co], AppenderQuery[_T_co]]: ...
+
+        def __set__(
+            self, instance: Any, value: typing.Collection[_T_co]
+        ) -> None: ...
+
+
+class WriteOnlyMapped(_MappedAnnotationBase[_T_co]):
+    """Represent the ORM mapped attribute type for a "write only" relationship.
+
+    The :class:`_orm.WriteOnlyMapped` type annotation may be used in an
+    :ref:`Annotated Declarative Table <orm_declarative_mapped_column>` mapping
+    to indicate that the ``lazy="write_only"`` loader strategy should be used
+    for a particular :func:`_orm.relationship`.
+
+    E.g.::
+
+        class User(Base):
+            __tablename__ = "user"
+            id: Mapped[int] = mapped_column(primary_key=True)
+            addresses: WriteOnlyMapped[Address] = relationship(
+                cascade="all,delete-orphan"
+            )
+
+    See the section :ref:`write_only_relationship` for background.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`write_only_relationship` - complete background
+
+        :class:`.DynamicMapped` - includes legacy :class:`_orm.Query` support
+
+    """
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        @overload
+        def __get__(
+            self, instance: None, owner: Any
+        ) -> InstrumentedAttribute[_T_co]: ...
+
+        @overload
+        def __get__(
+            self, instance: object, owner: Any
+        ) -> WriteOnlyCollection[_T_co]: ...
+
+        def __get__(
+            self, instance: Optional[object], owner: Any
+        ) -> Union[
+            InstrumentedAttribute[_T_co], WriteOnlyCollection[_T_co]
+        ]: ...
+
+        def __set__(
+            self, instance: Any, value: typing.Collection[_T_co]
+        ) -> None: ...
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/bulk_persistence.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/bulk_persistence.py
new file mode 100644
index 00000000..402d7bed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/bulk_persistence.py
@@ -0,0 +1,2123 @@
+# orm/bulk_persistence.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""additional ORM persistence classes related to "bulk" operations,
+specifically outside of the flush() process.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import Optional
+from typing import overload
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import context
+from . import evaluator
+from . import exc as orm_exc
+from . import loading
+from . import persistence
+from .base import NO_VALUE
+from .context import AbstractORMCompileState
+from .context import FromStatement
+from .context import ORMFromStatementCompileState
+from .context import QueryContext
+from .. import exc as sa_exc
+from .. import util
+from ..engine import Dialect
+from ..engine import result as _result
+from ..sql import coercions
+from ..sql import dml
+from ..sql import expression
+from ..sql import roles
+from ..sql import select
+from ..sql import sqltypes
+from ..sql.base import _entity_namespace_key
+from ..sql.base import CompileState
+from ..sql.base import Options
+from ..sql.dml import DeleteDMLState
+from ..sql.dml import InsertDMLState
+from ..sql.dml import UpdateDMLState
+from ..util import EMPTY_DICT
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from ._typing import DMLStrategyArgument
+    from ._typing import OrmExecuteOptionsParameter
+    from ._typing import SynchronizeSessionArgument
+    from .mapper import Mapper
+    from .session import _BindArguments
+    from .session import ORMExecuteState
+    from .session import Session
+    from .session import SessionTransaction
+    from .state import InstanceState
+    from ..engine import Connection
+    from ..engine import cursor
+    from ..engine.interfaces import _CoreAnyExecuteParams
+
+_O = TypeVar("_O", bound=object)
+
+
+@overload
+def _bulk_insert(
+    mapper: Mapper[_O],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    return_defaults: bool,
+    render_nulls: bool,
+    use_orm_insert_stmt: Literal[None] = ...,
+    execution_options: Optional[OrmExecuteOptionsParameter] = ...,
+) -> None: ...
+
+
+@overload
+def _bulk_insert(
+    mapper: Mapper[_O],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    return_defaults: bool,
+    render_nulls: bool,
+    use_orm_insert_stmt: Optional[dml.Insert] = ...,
+    execution_options: Optional[OrmExecuteOptionsParameter] = ...,
+) -> cursor.CursorResult[Any]: ...
+
+
+def _bulk_insert(
+    mapper: Mapper[_O],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    return_defaults: bool,
+    render_nulls: bool,
+    use_orm_insert_stmt: Optional[dml.Insert] = None,
+    execution_options: Optional[OrmExecuteOptionsParameter] = None,
+) -> Optional[cursor.CursorResult[Any]]:
+    base_mapper = mapper.base_mapper
+
+    if session_transaction.session.connection_callable:
+        raise NotImplementedError(
+            "connection_callable / per-instance sharding "
+            "not supported in bulk_insert()"
+        )
+
+    if isstates:
+        if TYPE_CHECKING:
+            mappings = cast(Iterable[InstanceState[_O]], mappings)
+
+        if return_defaults:
+            # list of states allows us to attach .key for return_defaults case
+            states = [(state, state.dict) for state in mappings]
+            mappings = [dict_ for (state, dict_) in states]
+        else:
+            mappings = [state.dict for state in mappings]
+    else:
+        if TYPE_CHECKING:
+            mappings = cast(Iterable[Dict[str, Any]], mappings)
+
+        if return_defaults:
+            # use dictionaries given, so that newly populated defaults
+            # can be delivered back to the caller (see #11661). This is **not**
+            # compatible with other use cases such as a session-executed
+            # insert() construct, as this will confuse the case of
+            # insert-per-subclass for joined inheritance cases (see
+            # test_bulk_statements.py::BulkDMLReturningJoinedInhTest).
+            #
+            # So in this conditional, we have **only** called
+            # session.bulk_insert_mappings() which does not have this
+            # requirement
+            mappings = list(mappings)
+        else:
+            # for all other cases we need to establish a local dictionary
+            # so that the incoming dictionaries aren't mutated
+            mappings = [dict(m) for m in mappings]
+        _expand_composites(mapper, mappings)
+
+    connection = session_transaction.connection(base_mapper)
+
+    return_result: Optional[cursor.CursorResult[Any]] = None
+
+    mappers_to_run = [
+        (table, mp)
+        for table, mp in base_mapper._sorted_tables.items()
+        if table in mapper._pks_by_table
+    ]
+
+    if return_defaults:
+        # not used by new-style bulk inserts, only used for legacy
+        bookkeeping = True
+    elif len(mappers_to_run) > 1:
+        # if we have more than one table, mapper to run where we will be
+        # either horizontally splicing, or copying values between tables,
+        # we need the "bookkeeping" / deterministic returning order
+        bookkeeping = True
+    else:
+        bookkeeping = False
+
+    for table, super_mapper in mappers_to_run:
+        # find bindparams in the statement. For bulk, we don't really know if
+        # a key in the params applies to a different table since we are
+        # potentially inserting for multiple tables here; looking at the
+        # bindparam() is a lot more direct.   in most cases this will
+        # use _generate_cache_key() which is memoized, although in practice
+        # the ultimate statement that's executed is probably not the same
+        # object so that memoization might not matter much.
+        extra_bp_names = (
+            [
+                b.key
+                for b in use_orm_insert_stmt._get_embedded_bindparams()
+                if b.key in mappings[0]
+            ]
+            if use_orm_insert_stmt is not None
+            else ()
+        )
+
+        records = (
+            (
+                None,
+                state_dict,
+                params,
+                mapper,
+                connection,
+                value_params,
+                has_all_pks,
+                has_all_defaults,
+            )
+            for (
+                state,
+                state_dict,
+                params,
+                mp,
+                conn,
+                value_params,
+                has_all_pks,
+                has_all_defaults,
+            ) in persistence._collect_insert_commands(
+                table,
+                ((None, mapping, mapper, connection) for mapping in mappings),
+                bulk=True,
+                return_defaults=bookkeeping,
+                render_nulls=render_nulls,
+                include_bulk_keys=extra_bp_names,
+            )
+        )
+
+        result = persistence._emit_insert_statements(
+            base_mapper,
+            None,
+            super_mapper,
+            table,
+            records,
+            bookkeeping=bookkeeping,
+            use_orm_insert_stmt=use_orm_insert_stmt,
+            execution_options=execution_options,
+        )
+        if use_orm_insert_stmt is not None:
+            if not use_orm_insert_stmt._returning or return_result is None:
+                return_result = result
+            elif result.returns_rows:
+                assert bookkeeping
+                return_result = return_result.splice_horizontally(result)
+
+    if return_defaults and isstates:
+        identity_cls = mapper._identity_class
+        identity_props = [p.key for p in mapper._identity_key_props]
+        for state, dict_ in states:
+            state.key = (
+                identity_cls,
+                tuple([dict_[key] for key in identity_props]),
+                None,
+            )
+
+    if use_orm_insert_stmt is not None:
+        assert return_result is not None
+        return return_result
+
+
+@overload
+def _bulk_update(
+    mapper: Mapper[Any],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    update_changed_only: bool,
+    use_orm_update_stmt: Literal[None] = ...,
+    enable_check_rowcount: bool = True,
+) -> None: ...
+
+
+@overload
+def _bulk_update(
+    mapper: Mapper[Any],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    update_changed_only: bool,
+    use_orm_update_stmt: Optional[dml.Update] = ...,
+    enable_check_rowcount: bool = True,
+) -> _result.Result[Any]: ...
+
+
+def _bulk_update(
+    mapper: Mapper[Any],
+    mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+    session_transaction: SessionTransaction,
+    *,
+    isstates: bool,
+    update_changed_only: bool,
+    use_orm_update_stmt: Optional[dml.Update] = None,
+    enable_check_rowcount: bool = True,
+) -> Optional[_result.Result[Any]]:
+    base_mapper = mapper.base_mapper
+
+    search_keys = mapper._primary_key_propkeys
+    if mapper._version_id_prop:
+        search_keys = {mapper._version_id_prop.key}.union(search_keys)
+
+    def _changed_dict(mapper, state):
+        return {
+            k: v
+            for k, v in state.dict.items()
+            if k in state.committed_state or k in search_keys
+        }
+
+    if isstates:
+        if update_changed_only:
+            mappings = [_changed_dict(mapper, state) for state in mappings]
+        else:
+            mappings = [state.dict for state in mappings]
+    else:
+        mappings = [dict(m) for m in mappings]
+        _expand_composites(mapper, mappings)
+
+    if session_transaction.session.connection_callable:
+        raise NotImplementedError(
+            "connection_callable / per-instance sharding "
+            "not supported in bulk_update()"
+        )
+
+    connection = session_transaction.connection(base_mapper)
+
+    # find bindparams in the statement. see _bulk_insert for similar
+    # notes for the insert case
+    extra_bp_names = (
+        [
+            b.key
+            for b in use_orm_update_stmt._get_embedded_bindparams()
+            if b.key in mappings[0]
+        ]
+        if use_orm_update_stmt is not None
+        else ()
+    )
+
+    for table, super_mapper in base_mapper._sorted_tables.items():
+        if not mapper.isa(super_mapper) or table not in mapper._pks_by_table:
+            continue
+
+        records = persistence._collect_update_commands(
+            None,
+            table,
+            (
+                (
+                    None,
+                    mapping,
+                    mapper,
+                    connection,
+                    (
+                        mapping[mapper._version_id_prop.key]
+                        if mapper._version_id_prop
+                        else None
+                    ),
+                )
+                for mapping in mappings
+            ),
+            bulk=True,
+            use_orm_update_stmt=use_orm_update_stmt,
+            include_bulk_keys=extra_bp_names,
+        )
+        persistence._emit_update_statements(
+            base_mapper,
+            None,
+            super_mapper,
+            table,
+            records,
+            bookkeeping=False,
+            use_orm_update_stmt=use_orm_update_stmt,
+            enable_check_rowcount=enable_check_rowcount,
+        )
+
+    if use_orm_update_stmt is not None:
+        return _result.null_result()
+
+
+def _expand_composites(mapper, mappings):
+    composite_attrs = mapper.composites
+    if not composite_attrs:
+        return
+
+    composite_keys = set(composite_attrs.keys())
+    populators = {
+        key: composite_attrs[key]._populate_composite_bulk_save_mappings_fn()
+        for key in composite_keys
+    }
+    for mapping in mappings:
+        for key in composite_keys.intersection(mapping):
+            populators[key](mapping)
+
+
+class ORMDMLState(AbstractORMCompileState):
+    is_dml_returning = True
+    from_statement_ctx: Optional[ORMFromStatementCompileState] = None
+
+    @classmethod
+    def _get_orm_crud_kv_pairs(
+        cls, mapper, statement, kv_iterator, needs_to_be_cacheable
+    ):
+        core_get_crud_kv_pairs = UpdateDMLState._get_crud_kv_pairs
+
+        for k, v in kv_iterator:
+            k = coercions.expect(roles.DMLColumnRole, k)
+
+            if isinstance(k, str):
+                desc = _entity_namespace_key(mapper, k, default=NO_VALUE)
+                if desc is NO_VALUE:
+                    yield (
+                        coercions.expect(roles.DMLColumnRole, k),
+                        (
+                            coercions.expect(
+                                roles.ExpressionElementRole,
+                                v,
+                                type_=sqltypes.NullType(),
+                                is_crud=True,
+                            )
+                            if needs_to_be_cacheable
+                            else v
+                        ),
+                    )
+                else:
+                    yield from core_get_crud_kv_pairs(
+                        statement,
+                        desc._bulk_update_tuples(v),
+                        needs_to_be_cacheable,
+                    )
+            elif "entity_namespace" in k._annotations:
+                k_anno = k._annotations
+                attr = _entity_namespace_key(
+                    k_anno["entity_namespace"], k_anno["proxy_key"]
+                )
+                yield from core_get_crud_kv_pairs(
+                    statement,
+                    attr._bulk_update_tuples(v),
+                    needs_to_be_cacheable,
+                )
+            else:
+                yield (
+                    k,
+                    (
+                        v
+                        if not needs_to_be_cacheable
+                        else coercions.expect(
+                            roles.ExpressionElementRole,
+                            v,
+                            type_=sqltypes.NullType(),
+                            is_crud=True,
+                        )
+                    ),
+                )
+
+    @classmethod
+    def _get_multi_crud_kv_pairs(cls, statement, kv_iterator):
+        plugin_subject = statement._propagate_attrs["plugin_subject"]
+
+        if not plugin_subject or not plugin_subject.mapper:
+            return UpdateDMLState._get_multi_crud_kv_pairs(
+                statement, kv_iterator
+            )
+
+        return [
+            dict(
+                cls._get_orm_crud_kv_pairs(
+                    plugin_subject.mapper, statement, value_dict.items(), False
+                )
+            )
+            for value_dict in kv_iterator
+        ]
+
+    @classmethod
+    def _get_crud_kv_pairs(cls, statement, kv_iterator, needs_to_be_cacheable):
+        assert (
+            needs_to_be_cacheable
+        ), "no test coverage for needs_to_be_cacheable=False"
+
+        plugin_subject = statement._propagate_attrs["plugin_subject"]
+
+        if not plugin_subject or not plugin_subject.mapper:
+            return UpdateDMLState._get_crud_kv_pairs(
+                statement, kv_iterator, needs_to_be_cacheable
+            )
+
+        return list(
+            cls._get_orm_crud_kv_pairs(
+                plugin_subject.mapper,
+                statement,
+                kv_iterator,
+                needs_to_be_cacheable,
+            )
+        )
+
+    @classmethod
+    def get_entity_description(cls, statement):
+        ext_info = statement.table._annotations["parententity"]
+        mapper = ext_info.mapper
+        if ext_info.is_aliased_class:
+            _label_name = ext_info.name
+        else:
+            _label_name = mapper.class_.__name__
+
+        return {
+            "name": _label_name,
+            "type": mapper.class_,
+            "expr": ext_info.entity,
+            "entity": ext_info.entity,
+            "table": mapper.local_table,
+        }
+
+    @classmethod
+    def get_returning_column_descriptions(cls, statement):
+        def _ent_for_col(c):
+            return c._annotations.get("parententity", None)
+
+        def _attr_for_col(c, ent):
+            if ent is None:
+                return c
+            proxy_key = c._annotations.get("proxy_key", None)
+            if not proxy_key:
+                return c
+            else:
+                return getattr(ent.entity, proxy_key, c)
+
+        return [
+            {
+                "name": c.key,
+                "type": c.type,
+                "expr": _attr_for_col(c, ent),
+                "aliased": ent.is_aliased_class,
+                "entity": ent.entity,
+            }
+            for c, ent in [
+                (c, _ent_for_col(c)) for c in statement._all_selected_columns
+            ]
+        ]
+
+    def _setup_orm_returning(
+        self,
+        compiler,
+        orm_level_statement,
+        dml_level_statement,
+        dml_mapper,
+        *,
+        use_supplemental_cols=True,
+    ):
+        """establish ORM column handlers for an INSERT, UPDATE, or DELETE
+        which uses explicit returning().
+
+        called within compilation level create_for_statement.
+
+        The _return_orm_returning() method then receives the Result
+        after the statement was executed, and applies ORM loading to the
+        state that we first established here.
+
+        """
+
+        if orm_level_statement._returning:
+            fs = FromStatement(
+                orm_level_statement._returning,
+                dml_level_statement,
+                _adapt_on_names=False,
+            )
+            fs = fs.execution_options(**orm_level_statement._execution_options)
+            fs = fs.options(*orm_level_statement._with_options)
+            self.select_statement = fs
+            self.from_statement_ctx = fsc = (
+                ORMFromStatementCompileState.create_for_statement(fs, compiler)
+            )
+            fsc.setup_dml_returning_compile_state(dml_mapper)
+
+            dml_level_statement = dml_level_statement._generate()
+            dml_level_statement._returning = ()
+
+            cols_to_return = [c for c in fsc.primary_columns if c is not None]
+
+            # since we are splicing result sets together, make sure there
+            # are columns of some kind returned in each result set
+            if not cols_to_return:
+                cols_to_return.extend(dml_mapper.primary_key)
+
+            if use_supplemental_cols:
+                dml_level_statement = dml_level_statement.return_defaults(
+                    # this is a little weird looking, but by passing
+                    # primary key as the main list of cols, this tells
+                    # return_defaults to omit server-default cols (and
+                    # actually all cols, due to some weird thing we should
+                    # clean up in crud.py).
+                    # Since we have cols_to_return, just return what we asked
+                    # for (plus primary key, which ORM persistence needs since
+                    # we likely set bookkeeping=True here, which is another
+                    # whole thing...).   We dont want to clutter the
+                    # statement up with lots of other cols the user didn't
+                    # ask for.  see #9685
+                    *dml_mapper.primary_key,
+                    supplemental_cols=cols_to_return,
+                )
+            else:
+                dml_level_statement = dml_level_statement.returning(
+                    *cols_to_return
+                )
+
+        return dml_level_statement
+
+    @classmethod
+    def _return_orm_returning(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        result,
+    ):
+        execution_context = result.context
+        compile_state = execution_context.compiled.compile_state
+
+        if (
+            compile_state.from_statement_ctx
+            and not compile_state.from_statement_ctx.compile_options._is_star
+        ):
+            load_options = execution_options.get(
+                "_sa_orm_load_options", QueryContext.default_load_options
+            )
+
+            querycontext = QueryContext(
+                compile_state.from_statement_ctx,
+                compile_state.select_statement,
+                statement,
+                params,
+                session,
+                load_options,
+                execution_options,
+                bind_arguments,
+            )
+            return loading.instances(result, querycontext)
+        else:
+            return result
+
+
+class BulkUDCompileState(ORMDMLState):
+    class default_update_options(Options):
+        _dml_strategy: DMLStrategyArgument = "auto"
+        _synchronize_session: SynchronizeSessionArgument = "auto"
+        _can_use_returning: bool = False
+        _is_delete_using: bool = False
+        _is_update_from: bool = False
+        _autoflush: bool = True
+        _subject_mapper: Optional[Mapper[Any]] = None
+        _resolved_values = EMPTY_DICT
+        _eval_condition = None
+        _matched_rows = None
+        _identity_token = None
+        _populate_existing: bool = False
+
+    @classmethod
+    def can_use_returning(
+        cls,
+        dialect: Dialect,
+        mapper: Mapper[Any],
+        *,
+        is_multitable: bool = False,
+        is_update_from: bool = False,
+        is_delete_using: bool = False,
+        is_executemany: bool = False,
+    ) -> bool:
+        raise NotImplementedError()
+
+    @classmethod
+    def orm_pre_session_exec(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        is_pre_event,
+    ):
+        (
+            update_options,
+            execution_options,
+        ) = BulkUDCompileState.default_update_options.from_execution_options(
+            "_sa_orm_update_options",
+            {
+                "synchronize_session",
+                "autoflush",
+                "populate_existing",
+                "identity_token",
+                "is_delete_using",
+                "is_update_from",
+                "dml_strategy",
+            },
+            execution_options,
+            statement._execution_options,
+        )
+        bind_arguments["clause"] = statement
+        try:
+            plugin_subject = statement._propagate_attrs["plugin_subject"]
+        except KeyError:
+            assert False, "statement had 'orm' plugin but no plugin_subject"
+        else:
+            if plugin_subject:
+                bind_arguments["mapper"] = plugin_subject.mapper
+                update_options += {"_subject_mapper": plugin_subject.mapper}
+
+        if "parententity" not in statement.table._annotations:
+            update_options += {"_dml_strategy": "core_only"}
+        elif not isinstance(params, list):
+            if update_options._dml_strategy == "auto":
+                update_options += {"_dml_strategy": "orm"}
+            elif update_options._dml_strategy == "bulk":
+                raise sa_exc.InvalidRequestError(
+                    'Can\'t use "bulk" ORM insert strategy without '
+                    "passing separate parameters"
+                )
+        else:
+            if update_options._dml_strategy == "auto":
+                update_options += {"_dml_strategy": "bulk"}
+
+        sync = update_options._synchronize_session
+        if sync is not None:
+            if sync not in ("auto", "evaluate", "fetch", False):
+                raise sa_exc.ArgumentError(
+                    "Valid strategies for session synchronization "
+                    "are 'auto', 'evaluate', 'fetch', False"
+                )
+            if update_options._dml_strategy == "bulk" and sync == "fetch":
+                raise sa_exc.InvalidRequestError(
+                    "The 'fetch' synchronization strategy is not available "
+                    "for 'bulk' ORM updates (i.e. multiple parameter sets)"
+                )
+
+        if not is_pre_event:
+            if update_options._autoflush:
+                session._autoflush()
+
+            if update_options._dml_strategy == "orm":
+                if update_options._synchronize_session == "auto":
+                    update_options = cls._do_pre_synchronize_auto(
+                        session,
+                        statement,
+                        params,
+                        execution_options,
+                        bind_arguments,
+                        update_options,
+                    )
+                elif update_options._synchronize_session == "evaluate":
+                    update_options = cls._do_pre_synchronize_evaluate(
+                        session,
+                        statement,
+                        params,
+                        execution_options,
+                        bind_arguments,
+                        update_options,
+                    )
+                elif update_options._synchronize_session == "fetch":
+                    update_options = cls._do_pre_synchronize_fetch(
+                        session,
+                        statement,
+                        params,
+                        execution_options,
+                        bind_arguments,
+                        update_options,
+                    )
+            elif update_options._dml_strategy == "bulk":
+                if update_options._synchronize_session == "auto":
+                    update_options += {"_synchronize_session": "evaluate"}
+
+            # indicators from the "pre exec" step that are then
+            # added to the DML statement, which will also be part of the cache
+            # key.  The compile level create_for_statement() method will then
+            # consume these at compiler time.
+            statement = statement._annotate(
+                {
+                    "synchronize_session": update_options._synchronize_session,
+                    "is_delete_using": update_options._is_delete_using,
+                    "is_update_from": update_options._is_update_from,
+                    "dml_strategy": update_options._dml_strategy,
+                    "can_use_returning": update_options._can_use_returning,
+                }
+            )
+
+        return (
+            statement,
+            util.immutabledict(execution_options).union(
+                {"_sa_orm_update_options": update_options}
+            ),
+        )
+
+    @classmethod
+    def orm_setup_cursor_result(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        result,
+    ):
+        # this stage of the execution is called after the
+        # do_orm_execute event hook.  meaning for an extension like
+        # horizontal sharding, this step happens *within* the horizontal
+        # sharding event handler which calls session.execute() re-entrantly
+        # and will occur for each backend individually.
+        # the sharding extension then returns its own merged result from the
+        # individual ones we return here.
+
+        update_options = execution_options["_sa_orm_update_options"]
+        if update_options._dml_strategy == "orm":
+            if update_options._synchronize_session == "evaluate":
+                cls._do_post_synchronize_evaluate(
+                    session, statement, result, update_options
+                )
+            elif update_options._synchronize_session == "fetch":
+                cls._do_post_synchronize_fetch(
+                    session, statement, result, update_options
+                )
+        elif update_options._dml_strategy == "bulk":
+            if update_options._synchronize_session == "evaluate":
+                cls._do_post_synchronize_bulk_evaluate(
+                    session, params, result, update_options
+                )
+            return result
+
+        return cls._return_orm_returning(
+            session,
+            statement,
+            params,
+            execution_options,
+            bind_arguments,
+            result,
+        )
+
+    @classmethod
+    def _adjust_for_extra_criteria(cls, global_attributes, ext_info):
+        """Apply extra criteria filtering.
+
+        For all distinct single-table-inheritance mappers represented in the
+        table being updated or deleted, produce additional WHERE criteria such
+        that only the appropriate subtypes are selected from the total results.
+
+        Additionally, add WHERE criteria originating from LoaderCriteriaOptions
+        collected from the statement.
+
+        """
+
+        return_crit = ()
+
+        adapter = ext_info._adapter if ext_info.is_aliased_class else None
+
+        if (
+            "additional_entity_criteria",
+            ext_info.mapper,
+        ) in global_attributes:
+            return_crit += tuple(
+                ae._resolve_where_criteria(ext_info)
+                for ae in global_attributes[
+                    ("additional_entity_criteria", ext_info.mapper)
+                ]
+                if ae.include_aliases or ae.entity is ext_info
+            )
+
+        if ext_info.mapper._single_table_criterion is not None:
+            return_crit += (ext_info.mapper._single_table_criterion,)
+
+        if adapter:
+            return_crit = tuple(adapter.traverse(crit) for crit in return_crit)
+
+        return return_crit
+
+    @classmethod
+    def _interpret_returning_rows(cls, result, mapper, rows):
+        """return rows that indicate PK cols in mapper.primary_key position
+        for RETURNING rows.
+
+        Prior to 2.0.36, this method seemed to be written for some kind of
+        inheritance scenario but the scenario was unused for actual joined
+        inheritance, and the function instead seemed to perform some kind of
+        partial translation that would remove non-PK cols if the PK cols
+        happened to be first in the row, but not otherwise.  The joined
+        inheritance walk feature here seems to have never been used as it was
+        always skipped by the "local_table" check.
+
+        As of 2.0.36 the function strips away non-PK cols and provides the
+        PK cols for the table in mapper PK order.
+
+        """
+
+        try:
+            if mapper.local_table is not mapper.base_mapper.local_table:
+                # TODO: dive more into how a local table PK is used for fetch
+                # sync, not clear if this is correct as it depends on the
+                # downstream routine to fetch rows using
+                # local_table.primary_key order
+                pk_keys = result._tuple_getter(mapper.local_table.primary_key)
+            else:
+                pk_keys = result._tuple_getter(mapper.primary_key)
+        except KeyError:
+            # can't use these rows, they don't have PK cols in them
+            # this is an unusual case where the user would have used
+            # .return_defaults()
+            return []
+
+        return [pk_keys(row) for row in rows]
+
+    @classmethod
+    def _get_matched_objects_on_criteria(cls, update_options, states):
+        mapper = update_options._subject_mapper
+        eval_condition = update_options._eval_condition
+
+        raw_data = [
+            (state.obj(), state, state.dict)
+            for state in states
+            if state.mapper.isa(mapper) and not state.expired
+        ]
+
+        identity_token = update_options._identity_token
+        if identity_token is not None:
+            raw_data = [
+                (obj, state, dict_)
+                for obj, state, dict_ in raw_data
+                if state.identity_token == identity_token
+            ]
+
+        result = []
+        for obj, state, dict_ in raw_data:
+            evaled_condition = eval_condition(obj)
+
+            # caution: don't use "in ()" or == here, _EXPIRE_OBJECT
+            # evaluates as True for all comparisons
+            if (
+                evaled_condition is True
+                or evaled_condition is evaluator._EXPIRED_OBJECT
+            ):
+                result.append(
+                    (
+                        obj,
+                        state,
+                        dict_,
+                        evaled_condition is evaluator._EXPIRED_OBJECT,
+                    )
+                )
+        return result
+
+    @classmethod
+    def _eval_condition_from_statement(cls, update_options, statement):
+        mapper = update_options._subject_mapper
+        target_cls = mapper.class_
+
+        evaluator_compiler = evaluator._EvaluatorCompiler(target_cls)
+        crit = ()
+        if statement._where_criteria:
+            crit += statement._where_criteria
+
+        global_attributes = {}
+        for opt in statement._with_options:
+            if opt._is_criteria_option:
+                opt.get_global_criteria(global_attributes)
+
+        if global_attributes:
+            crit += cls._adjust_for_extra_criteria(global_attributes, mapper)
+
+        if crit:
+            eval_condition = evaluator_compiler.process(*crit)
+        else:
+            # workaround for mypy https://github.com/python/mypy/issues/14027
+            def _eval_condition(obj):
+                return True
+
+            eval_condition = _eval_condition
+
+        return eval_condition
+
+    @classmethod
+    def _do_pre_synchronize_auto(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        update_options,
+    ):
+        """setup auto sync strategy
+
+
+        "auto" checks if we can use "evaluate" first, then falls back
+        to "fetch"
+
+        evaluate is vastly more efficient for the common case
+        where session is empty, only has a few objects, and the UPDATE
+        statement can potentially match thousands/millions of rows.
+
+        OTOH more complex criteria that fails to work with "evaluate"
+        we would hope usually correlates with fewer net rows.
+
+        """
+
+        try:
+            eval_condition = cls._eval_condition_from_statement(
+                update_options, statement
+            )
+
+        except evaluator.UnevaluatableError:
+            pass
+        else:
+            return update_options + {
+                "_eval_condition": eval_condition,
+                "_synchronize_session": "evaluate",
+            }
+
+        update_options += {"_synchronize_session": "fetch"}
+        return cls._do_pre_synchronize_fetch(
+            session,
+            statement,
+            params,
+            execution_options,
+            bind_arguments,
+            update_options,
+        )
+
+    @classmethod
+    def _do_pre_synchronize_evaluate(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        update_options,
+    ):
+        try:
+            eval_condition = cls._eval_condition_from_statement(
+                update_options, statement
+            )
+
+        except evaluator.UnevaluatableError as err:
+            raise sa_exc.InvalidRequestError(
+                'Could not evaluate current criteria in Python: "%s". '
+                "Specify 'fetch' or False for the "
+                "synchronize_session execution option." % err
+            ) from err
+
+        return update_options + {
+            "_eval_condition": eval_condition,
+        }
+
+    @classmethod
+    def _get_resolved_values(cls, mapper, statement):
+        if statement._multi_values:
+            return []
+        elif statement._ordered_values:
+            return list(statement._ordered_values)
+        elif statement._values:
+            return list(statement._values.items())
+        else:
+            return []
+
+    @classmethod
+    def _resolved_keys_as_propnames(cls, mapper, resolved_values):
+        values = []
+        for k, v in resolved_values:
+            if mapper and isinstance(k, expression.ColumnElement):
+                try:
+                    attr = mapper._columntoproperty[k]
+                except orm_exc.UnmappedColumnError:
+                    pass
+                else:
+                    values.append((attr.key, v))
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "Attribute name not found, can't be "
+                    "synchronized back to objects: %r" % k
+                )
+        return values
+
+    @classmethod
+    def _do_pre_synchronize_fetch(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        update_options,
+    ):
+        mapper = update_options._subject_mapper
+
+        select_stmt = (
+            select(*(mapper.primary_key + (mapper.select_identity_token,)))
+            .select_from(mapper)
+            .options(*statement._with_options)
+        )
+        select_stmt._where_criteria = statement._where_criteria
+
+        # conditionally run the SELECT statement for pre-fetch, testing the
+        # "bind" for if we can use RETURNING or not using the do_orm_execute
+        # event.  If RETURNING is available, the do_orm_execute event
+        # will cancel the SELECT from being actually run.
+        #
+        # The way this is organized seems strange, why don't we just
+        # call can_use_returning() before invoking the statement and get
+        # answer?, why does this go through the whole execute phase using an
+        # event?  Answer: because we are integrating with extensions such
+        # as the horizontal sharding extention that "multiplexes" an individual
+        # statement run through multiple engines, and it uses
+        # do_orm_execute() to do that.
+
+        can_use_returning = None
+
+        def skip_for_returning(orm_context: ORMExecuteState) -> Any:
+            bind = orm_context.session.get_bind(**orm_context.bind_arguments)
+            nonlocal can_use_returning
+
+            per_bind_result = cls.can_use_returning(
+                bind.dialect,
+                mapper,
+                is_update_from=update_options._is_update_from,
+                is_delete_using=update_options._is_delete_using,
+                is_executemany=orm_context.is_executemany,
+            )
+
+            if can_use_returning is not None:
+                if can_use_returning != per_bind_result:
+                    raise sa_exc.InvalidRequestError(
+                        "For synchronize_session='fetch', can't mix multiple "
+                        "backends where some support RETURNING and others "
+                        "don't"
+                    )
+            elif orm_context.is_executemany and not per_bind_result:
+                raise sa_exc.InvalidRequestError(
+                    "For synchronize_session='fetch', can't use multiple "
+                    "parameter sets in ORM mode, which this backend does not "
+                    "support with RETURNING"
+                )
+            else:
+                can_use_returning = per_bind_result
+
+            if per_bind_result:
+                return _result.null_result()
+            else:
+                return None
+
+        result = session.execute(
+            select_stmt,
+            params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            _add_event=skip_for_returning,
+        )
+        matched_rows = result.fetchall()
+
+        return update_options + {
+            "_matched_rows": matched_rows,
+            "_can_use_returning": can_use_returning,
+        }
+
+
+@CompileState.plugin_for("orm", "insert")
+class BulkORMInsert(ORMDMLState, InsertDMLState):
+    class default_insert_options(Options):
+        _dml_strategy: DMLStrategyArgument = "auto"
+        _render_nulls: bool = False
+        _return_defaults: bool = False
+        _subject_mapper: Optional[Mapper[Any]] = None
+        _autoflush: bool = True
+        _populate_existing: bool = False
+
+    select_statement: Optional[FromStatement] = None
+
+    @classmethod
+    def orm_pre_session_exec(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        is_pre_event,
+    ):
+        (
+            insert_options,
+            execution_options,
+        ) = BulkORMInsert.default_insert_options.from_execution_options(
+            "_sa_orm_insert_options",
+            {"dml_strategy", "autoflush", "populate_existing", "render_nulls"},
+            execution_options,
+            statement._execution_options,
+        )
+        bind_arguments["clause"] = statement
+        try:
+            plugin_subject = statement._propagate_attrs["plugin_subject"]
+        except KeyError:
+            assert False, "statement had 'orm' plugin but no plugin_subject"
+        else:
+            if plugin_subject:
+                bind_arguments["mapper"] = plugin_subject.mapper
+                insert_options += {"_subject_mapper": plugin_subject.mapper}
+
+        if not params:
+            if insert_options._dml_strategy == "auto":
+                insert_options += {"_dml_strategy": "orm"}
+            elif insert_options._dml_strategy == "bulk":
+                raise sa_exc.InvalidRequestError(
+                    'Can\'t use "bulk" ORM insert strategy without '
+                    "passing separate parameters"
+                )
+        else:
+            if insert_options._dml_strategy == "auto":
+                insert_options += {"_dml_strategy": "bulk"}
+
+        if insert_options._dml_strategy != "raw":
+            # for ORM object loading, like ORMContext, we have to disable
+            # result set adapt_to_context, because we will be generating a
+            # new statement with specific columns that's cached inside of
+            # an ORMFromStatementCompileState, which we will re-use for
+            # each result.
+            if not execution_options:
+                execution_options = context._orm_load_exec_options
+            else:
+                execution_options = execution_options.union(
+                    context._orm_load_exec_options
+                )
+
+        if not is_pre_event and insert_options._autoflush:
+            session._autoflush()
+
+        statement = statement._annotate(
+            {"dml_strategy": insert_options._dml_strategy}
+        )
+
+        return (
+            statement,
+            util.immutabledict(execution_options).union(
+                {"_sa_orm_insert_options": insert_options}
+            ),
+        )
+
+    @classmethod
+    def orm_execute_statement(
+        cls,
+        session: Session,
+        statement: dml.Insert,
+        params: _CoreAnyExecuteParams,
+        execution_options: OrmExecuteOptionsParameter,
+        bind_arguments: _BindArguments,
+        conn: Connection,
+    ) -> _result.Result:
+        insert_options = execution_options.get(
+            "_sa_orm_insert_options", cls.default_insert_options
+        )
+
+        if insert_options._dml_strategy not in (
+            "raw",
+            "bulk",
+            "orm",
+            "auto",
+        ):
+            raise sa_exc.ArgumentError(
+                "Valid strategies for ORM insert strategy "
+                "are 'raw', 'orm', 'bulk', 'auto"
+            )
+
+        result: _result.Result[Any]
+
+        if insert_options._dml_strategy == "raw":
+            result = conn.execute(
+                statement, params or {}, execution_options=execution_options
+            )
+            return result
+
+        if insert_options._dml_strategy == "bulk":
+            mapper = insert_options._subject_mapper
+
+            if (
+                statement._post_values_clause is not None
+                and mapper._multiple_persistence_tables
+            ):
+                raise sa_exc.InvalidRequestError(
+                    "bulk INSERT with a 'post values' clause "
+                    "(typically upsert) not supported for multi-table "
+                    f"mapper {mapper}"
+                )
+
+            assert mapper is not None
+            assert session._transaction is not None
+            result = _bulk_insert(
+                mapper,
+                cast(
+                    "Iterable[Dict[str, Any]]",
+                    [params] if isinstance(params, dict) else params,
+                ),
+                session._transaction,
+                isstates=False,
+                return_defaults=insert_options._return_defaults,
+                render_nulls=insert_options._render_nulls,
+                use_orm_insert_stmt=statement,
+                execution_options=execution_options,
+            )
+        elif insert_options._dml_strategy == "orm":
+            result = conn.execute(
+                statement, params or {}, execution_options=execution_options
+            )
+        else:
+            raise AssertionError()
+
+        if not bool(statement._returning):
+            return result
+
+        if insert_options._populate_existing:
+            load_options = execution_options.get(
+                "_sa_orm_load_options", QueryContext.default_load_options
+            )
+            load_options += {"_populate_existing": True}
+            execution_options = execution_options.union(
+                {"_sa_orm_load_options": load_options}
+            )
+
+        return cls._return_orm_returning(
+            session,
+            statement,
+            params,
+            execution_options,
+            bind_arguments,
+            result,
+        )
+
+    @classmethod
+    def create_for_statement(cls, statement, compiler, **kw) -> BulkORMInsert:
+        self = cast(
+            BulkORMInsert,
+            super().create_for_statement(statement, compiler, **kw),
+        )
+
+        if compiler is not None:
+            toplevel = not compiler.stack
+        else:
+            toplevel = True
+        if not toplevel:
+            return self
+
+        mapper = statement._propagate_attrs["plugin_subject"]
+        dml_strategy = statement._annotations.get("dml_strategy", "raw")
+        if dml_strategy == "bulk":
+            self._setup_for_bulk_insert(compiler)
+        elif dml_strategy == "orm":
+            self._setup_for_orm_insert(compiler, mapper)
+
+        return self
+
+    @classmethod
+    def _resolved_keys_as_col_keys(cls, mapper, resolved_value_dict):
+        return {
+            col.key if col is not None else k: v
+            for col, k, v in (
+                (mapper.c.get(k), k, v) for k, v in resolved_value_dict.items()
+            )
+        }
+
+    def _setup_for_orm_insert(self, compiler, mapper):
+        statement = orm_level_statement = cast(dml.Insert, self.statement)
+
+        statement = self._setup_orm_returning(
+            compiler,
+            orm_level_statement,
+            statement,
+            dml_mapper=mapper,
+            use_supplemental_cols=False,
+        )
+        self.statement = statement
+
+    def _setup_for_bulk_insert(self, compiler):
+        """establish an INSERT statement within the context of
+        bulk insert.
+
+        This method will be within the "conn.execute()" call that is invoked
+        by persistence._emit_insert_statement().
+
+        """
+        statement = orm_level_statement = cast(dml.Insert, self.statement)
+        an = statement._annotations
+
+        emit_insert_table, emit_insert_mapper = (
+            an["_emit_insert_table"],
+            an["_emit_insert_mapper"],
+        )
+
+        statement = statement._clone()
+
+        statement.table = emit_insert_table
+        if self._dict_parameters:
+            self._dict_parameters = {
+                col: val
+                for col, val in self._dict_parameters.items()
+                if col.table is emit_insert_table
+            }
+
+        statement = self._setup_orm_returning(
+            compiler,
+            orm_level_statement,
+            statement,
+            dml_mapper=emit_insert_mapper,
+            use_supplemental_cols=True,
+        )
+
+        if (
+            self.from_statement_ctx is not None
+            and self.from_statement_ctx.compile_options._is_star
+        ):
+            raise sa_exc.CompileError(
+                "Can't use RETURNING * with bulk ORM INSERT.  "
+                "Please use a different INSERT form, such as INSERT..VALUES "
+                "or INSERT with a Core Connection"
+            )
+
+        self.statement = statement
+
+
+@CompileState.plugin_for("orm", "update")
+class BulkORMUpdate(BulkUDCompileState, UpdateDMLState):
+    @classmethod
+    def create_for_statement(cls, statement, compiler, **kw):
+        self = cls.__new__(cls)
+
+        dml_strategy = statement._annotations.get(
+            "dml_strategy", "unspecified"
+        )
+
+        toplevel = not compiler.stack
+
+        if toplevel and dml_strategy == "bulk":
+            self._setup_for_bulk_update(statement, compiler)
+        elif (
+            dml_strategy == "core_only"
+            or dml_strategy == "unspecified"
+            and "parententity" not in statement.table._annotations
+        ):
+            UpdateDMLState.__init__(self, statement, compiler, **kw)
+        elif not toplevel or dml_strategy in ("orm", "unspecified"):
+            self._setup_for_orm_update(statement, compiler)
+
+        return self
+
+    def _setup_for_orm_update(self, statement, compiler, **kw):
+        orm_level_statement = statement
+
+        toplevel = not compiler.stack
+
+        ext_info = statement.table._annotations["parententity"]
+
+        self.mapper = mapper = ext_info.mapper
+
+        self._resolved_values = self._get_resolved_values(mapper, statement)
+
+        self._init_global_attributes(
+            statement,
+            compiler,
+            toplevel=toplevel,
+            process_criteria_for_toplevel=toplevel,
+        )
+
+        if statement._values:
+            self._resolved_values = dict(self._resolved_values)
+
+        new_stmt = statement._clone()
+
+        if new_stmt.table._annotations["parententity"] is mapper:
+            new_stmt.table = mapper.local_table
+
+        # note if the statement has _multi_values, these
+        # are passed through to the new statement, which will then raise
+        # InvalidRequestError because UPDATE doesn't support multi_values
+        # right now.
+        if statement._ordered_values:
+            new_stmt._ordered_values = self._resolved_values
+        elif statement._values:
+            new_stmt._values = self._resolved_values
+
+        new_crit = self._adjust_for_extra_criteria(
+            self.global_attributes, mapper
+        )
+        if new_crit:
+            new_stmt = new_stmt.where(*new_crit)
+
+        # if we are against a lambda statement we might not be the
+        # topmost object that received per-execute annotations
+
+        # do this first as we need to determine if there is
+        # UPDATE..FROM
+
+        UpdateDMLState.__init__(self, new_stmt, compiler, **kw)
+
+        use_supplemental_cols = False
+
+        if not toplevel:
+            synchronize_session = None
+        else:
+            synchronize_session = compiler._annotations.get(
+                "synchronize_session", None
+            )
+        can_use_returning = compiler._annotations.get(
+            "can_use_returning", None
+        )
+        if can_use_returning is not False:
+            # even though pre_exec has determined basic
+            # can_use_returning for the dialect, if we are to use
+            # RETURNING we need to run can_use_returning() at this level
+            # unconditionally because is_delete_using was not known
+            # at the pre_exec level
+            can_use_returning = (
+                synchronize_session == "fetch"
+                and self.can_use_returning(
+                    compiler.dialect, mapper, is_multitable=self.is_multitable
+                )
+            )
+
+        if synchronize_session == "fetch" and can_use_returning:
+            use_supplemental_cols = True
+
+            # NOTE: we might want to RETURNING the actual columns to be
+            # synchronized also.  however this is complicated and difficult
+            # to align against the behavior of "evaluate".  Additionally,
+            # in a large number (if not the majority) of cases, we have the
+            # "evaluate" answer, usually a fixed value, in memory already and
+            # there's no need to re-fetch the same value
+            # over and over again.   so perhaps if it could be RETURNING just
+            # the elements that were based on a SQL expression and not
+            # a constant.   For now it doesn't quite seem worth it
+            new_stmt = new_stmt.return_defaults(*new_stmt.table.primary_key)
+
+        if toplevel:
+            new_stmt = self._setup_orm_returning(
+                compiler,
+                orm_level_statement,
+                new_stmt,
+                dml_mapper=mapper,
+                use_supplemental_cols=use_supplemental_cols,
+            )
+
+        self.statement = new_stmt
+
+    def _setup_for_bulk_update(self, statement, compiler, **kw):
+        """establish an UPDATE statement within the context of
+        bulk insert.
+
+        This method will be within the "conn.execute()" call that is invoked
+        by persistence._emit_update_statement().
+
+        """
+        statement = cast(dml.Update, statement)
+        an = statement._annotations
+
+        emit_update_table, _ = (
+            an["_emit_update_table"],
+            an["_emit_update_mapper"],
+        )
+
+        statement = statement._clone()
+        statement.table = emit_update_table
+
+        UpdateDMLState.__init__(self, statement, compiler, **kw)
+
+        if self._ordered_values:
+            raise sa_exc.InvalidRequestError(
+                "bulk ORM UPDATE does not support ordered_values() for "
+                "custom UPDATE statements with bulk parameter sets.  Use a "
+                "non-bulk UPDATE statement or use values()."
+            )
+
+        if self._dict_parameters:
+            self._dict_parameters = {
+                col: val
+                for col, val in self._dict_parameters.items()
+                if col.table is emit_update_table
+            }
+        self.statement = statement
+
+    @classmethod
+    def orm_execute_statement(
+        cls,
+        session: Session,
+        statement: dml.Update,
+        params: _CoreAnyExecuteParams,
+        execution_options: OrmExecuteOptionsParameter,
+        bind_arguments: _BindArguments,
+        conn: Connection,
+    ) -> _result.Result:
+
+        update_options = execution_options.get(
+            "_sa_orm_update_options", cls.default_update_options
+        )
+
+        if update_options._populate_existing:
+            load_options = execution_options.get(
+                "_sa_orm_load_options", QueryContext.default_load_options
+            )
+            load_options += {"_populate_existing": True}
+            execution_options = execution_options.union(
+                {"_sa_orm_load_options": load_options}
+            )
+
+        if update_options._dml_strategy not in (
+            "orm",
+            "auto",
+            "bulk",
+            "core_only",
+        ):
+            raise sa_exc.ArgumentError(
+                "Valid strategies for ORM UPDATE strategy "
+                "are 'orm', 'auto', 'bulk', 'core_only'"
+            )
+
+        result: _result.Result[Any]
+
+        if update_options._dml_strategy == "bulk":
+            enable_check_rowcount = not statement._where_criteria
+
+            assert update_options._synchronize_session != "fetch"
+
+            if (
+                statement._where_criteria
+                and update_options._synchronize_session == "evaluate"
+            ):
+                raise sa_exc.InvalidRequestError(
+                    "bulk synchronize of persistent objects not supported "
+                    "when using bulk update with additional WHERE "
+                    "criteria right now.  add synchronize_session=None "
+                    "execution option to bypass synchronize of persistent "
+                    "objects."
+                )
+            mapper = update_options._subject_mapper
+            assert mapper is not None
+            assert session._transaction is not None
+            result = _bulk_update(
+                mapper,
+                cast(
+                    "Iterable[Dict[str, Any]]",
+                    [params] if isinstance(params, dict) else params,
+                ),
+                session._transaction,
+                isstates=False,
+                update_changed_only=False,
+                use_orm_update_stmt=statement,
+                enable_check_rowcount=enable_check_rowcount,
+            )
+            return cls.orm_setup_cursor_result(
+                session,
+                statement,
+                params,
+                execution_options,
+                bind_arguments,
+                result,
+            )
+        else:
+            return super().orm_execute_statement(
+                session,
+                statement,
+                params,
+                execution_options,
+                bind_arguments,
+                conn,
+            )
+
+    @classmethod
+    def can_use_returning(
+        cls,
+        dialect: Dialect,
+        mapper: Mapper[Any],
+        *,
+        is_multitable: bool = False,
+        is_update_from: bool = False,
+        is_delete_using: bool = False,
+        is_executemany: bool = False,
+    ) -> bool:
+        # normal answer for "should we use RETURNING" at all.
+        normal_answer = (
+            dialect.update_returning and mapper.local_table.implicit_returning
+        )
+        if not normal_answer:
+            return False
+
+        if is_executemany:
+            return dialect.update_executemany_returning
+
+        # these workarounds are currently hypothetical for UPDATE,
+        # unlike DELETE where they impact MariaDB
+        if is_update_from:
+            return dialect.update_returning_multifrom
+
+        elif is_multitable and not dialect.update_returning_multifrom:
+            raise sa_exc.CompileError(
+                f'Dialect "{dialect.name}" does not support RETURNING '
+                "with UPDATE..FROM; for synchronize_session='fetch', "
+                "please add the additional execution option "
+                "'is_update_from=True' to the statement to indicate that "
+                "a separate SELECT should be used for this backend."
+            )
+
+        return True
+
+    @classmethod
+    def _do_post_synchronize_bulk_evaluate(
+        cls, session, params, result, update_options
+    ):
+        if not params:
+            return
+
+        mapper = update_options._subject_mapper
+        pk_keys = [prop.key for prop in mapper._identity_key_props]
+
+        identity_map = session.identity_map
+
+        for param in params:
+            identity_key = mapper.identity_key_from_primary_key(
+                (param[key] for key in pk_keys),
+                update_options._identity_token,
+            )
+            state = identity_map.fast_get_state(identity_key)
+            if not state:
+                continue
+
+            evaluated_keys = set(param).difference(pk_keys)
+
+            dict_ = state.dict
+            # only evaluate unmodified attributes
+            to_evaluate = state.unmodified.intersection(evaluated_keys)
+            for key in to_evaluate:
+                if key in dict_:
+                    dict_[key] = param[key]
+
+            state.manager.dispatch.refresh(state, None, to_evaluate)
+
+            state._commit(dict_, list(to_evaluate))
+
+            # attributes that were formerly modified instead get expired.
+            # this only gets hit if the session had pending changes
+            # and autoflush were set to False.
+            to_expire = evaluated_keys.intersection(dict_).difference(
+                to_evaluate
+            )
+            if to_expire:
+                state._expire_attributes(dict_, to_expire)
+
+    @classmethod
+    def _do_post_synchronize_evaluate(
+        cls, session, statement, result, update_options
+    ):
+        matched_objects = cls._get_matched_objects_on_criteria(
+            update_options,
+            session.identity_map.all_states(),
+        )
+
+        cls._apply_update_set_values_to_objects(
+            session,
+            update_options,
+            statement,
+            result.context.compiled_parameters[0],
+            [(obj, state, dict_) for obj, state, dict_, _ in matched_objects],
+            result.prefetch_cols(),
+            result.postfetch_cols(),
+        )
+
+    @classmethod
+    def _do_post_synchronize_fetch(
+        cls, session, statement, result, update_options
+    ):
+        target_mapper = update_options._subject_mapper
+
+        returned_defaults_rows = result.returned_defaults_rows
+        if returned_defaults_rows:
+            pk_rows = cls._interpret_returning_rows(
+                result, target_mapper, returned_defaults_rows
+            )
+            matched_rows = [
+                tuple(row) + (update_options._identity_token,)
+                for row in pk_rows
+            ]
+        else:
+            matched_rows = update_options._matched_rows
+
+        objs = [
+            session.identity_map[identity_key]
+            for identity_key in [
+                target_mapper.identity_key_from_primary_key(
+                    list(primary_key),
+                    identity_token=identity_token,
+                )
+                for primary_key, identity_token in [
+                    (row[0:-1], row[-1]) for row in matched_rows
+                ]
+                if update_options._identity_token is None
+                or identity_token == update_options._identity_token
+            ]
+            if identity_key in session.identity_map
+        ]
+
+        if not objs:
+            return
+
+        cls._apply_update_set_values_to_objects(
+            session,
+            update_options,
+            statement,
+            result.context.compiled_parameters[0],
+            [
+                (
+                    obj,
+                    attributes.instance_state(obj),
+                    attributes.instance_dict(obj),
+                )
+                for obj in objs
+            ],
+            result.prefetch_cols(),
+            result.postfetch_cols(),
+        )
+
+    @classmethod
+    def _apply_update_set_values_to_objects(
+        cls,
+        session,
+        update_options,
+        statement,
+        effective_params,
+        matched_objects,
+        prefetch_cols,
+        postfetch_cols,
+    ):
+        """apply values to objects derived from an update statement, e.g.
+        UPDATE..SET <values>
+
+        """
+
+        mapper = update_options._subject_mapper
+        target_cls = mapper.class_
+        evaluator_compiler = evaluator._EvaluatorCompiler(target_cls)
+        resolved_values = cls._get_resolved_values(mapper, statement)
+        resolved_keys_as_propnames = cls._resolved_keys_as_propnames(
+            mapper, resolved_values
+        )
+        value_evaluators = {}
+        for key, value in resolved_keys_as_propnames:
+            try:
+                _evaluator = evaluator_compiler.process(
+                    coercions.expect(roles.ExpressionElementRole, value)
+                )
+            except evaluator.UnevaluatableError:
+                pass
+            else:
+                value_evaluators[key] = _evaluator
+
+        evaluated_keys = list(value_evaluators.keys())
+        attrib = {k for k, v in resolved_keys_as_propnames}
+
+        states = set()
+
+        to_prefetch = {
+            c
+            for c in prefetch_cols
+            if c.key in effective_params
+            and c in mapper._columntoproperty
+            and c.key not in evaluated_keys
+        }
+        to_expire = {
+            mapper._columntoproperty[c].key
+            for c in postfetch_cols
+            if c in mapper._columntoproperty
+        }.difference(evaluated_keys)
+
+        prefetch_transfer = [
+            (mapper._columntoproperty[c].key, c.key) for c in to_prefetch
+        ]
+
+        for obj, state, dict_ in matched_objects:
+
+            dict_.update(
+                {
+                    col_to_prop: effective_params[c_key]
+                    for col_to_prop, c_key in prefetch_transfer
+                }
+            )
+
+            state._expire_attributes(state.dict, to_expire)
+
+            to_evaluate = state.unmodified.intersection(evaluated_keys)
+
+            for key in to_evaluate:
+                if key in dict_:
+                    # only run eval for attributes that are present.
+                    dict_[key] = value_evaluators[key](obj)
+
+            state.manager.dispatch.refresh(state, None, to_evaluate)
+
+            state._commit(dict_, list(to_evaluate))
+
+            # attributes that were formerly modified instead get expired.
+            # this only gets hit if the session had pending changes
+            # and autoflush were set to False.
+            to_expire = attrib.intersection(dict_).difference(to_evaluate)
+            if to_expire:
+                state._expire_attributes(dict_, to_expire)
+
+            states.add(state)
+        session._register_altered(states)
+
+
+@CompileState.plugin_for("orm", "delete")
+class BulkORMDelete(BulkUDCompileState, DeleteDMLState):
+    @classmethod
+    def create_for_statement(cls, statement, compiler, **kw):
+        self = cls.__new__(cls)
+
+        dml_strategy = statement._annotations.get(
+            "dml_strategy", "unspecified"
+        )
+
+        if (
+            dml_strategy == "core_only"
+            or dml_strategy == "unspecified"
+            and "parententity" not in statement.table._annotations
+        ):
+            DeleteDMLState.__init__(self, statement, compiler, **kw)
+            return self
+
+        toplevel = not compiler.stack
+
+        orm_level_statement = statement
+
+        ext_info = statement.table._annotations["parententity"]
+        self.mapper = mapper = ext_info.mapper
+
+        self._init_global_attributes(
+            statement,
+            compiler,
+            toplevel=toplevel,
+            process_criteria_for_toplevel=toplevel,
+        )
+
+        new_stmt = statement._clone()
+
+        if new_stmt.table._annotations["parententity"] is mapper:
+            new_stmt.table = mapper.local_table
+
+        new_crit = cls._adjust_for_extra_criteria(
+            self.global_attributes, mapper
+        )
+        if new_crit:
+            new_stmt = new_stmt.where(*new_crit)
+
+        # do this first as we need to determine if there is
+        # DELETE..FROM
+        DeleteDMLState.__init__(self, new_stmt, compiler, **kw)
+
+        use_supplemental_cols = False
+
+        if not toplevel:
+            synchronize_session = None
+        else:
+            synchronize_session = compiler._annotations.get(
+                "synchronize_session", None
+            )
+        can_use_returning = compiler._annotations.get(
+            "can_use_returning", None
+        )
+        if can_use_returning is not False:
+            # even though pre_exec has determined basic
+            # can_use_returning for the dialect, if we are to use
+            # RETURNING we need to run can_use_returning() at this level
+            # unconditionally because is_delete_using was not known
+            # at the pre_exec level
+            can_use_returning = (
+                synchronize_session == "fetch"
+                and self.can_use_returning(
+                    compiler.dialect,
+                    mapper,
+                    is_multitable=self.is_multitable,
+                    is_delete_using=compiler._annotations.get(
+                        "is_delete_using", False
+                    ),
+                )
+            )
+
+        if can_use_returning:
+            use_supplemental_cols = True
+
+            new_stmt = new_stmt.return_defaults(*new_stmt.table.primary_key)
+
+        if toplevel:
+            new_stmt = self._setup_orm_returning(
+                compiler,
+                orm_level_statement,
+                new_stmt,
+                dml_mapper=mapper,
+                use_supplemental_cols=use_supplemental_cols,
+            )
+
+        self.statement = new_stmt
+
+        return self
+
+    @classmethod
+    def orm_execute_statement(
+        cls,
+        session: Session,
+        statement: dml.Delete,
+        params: _CoreAnyExecuteParams,
+        execution_options: OrmExecuteOptionsParameter,
+        bind_arguments: _BindArguments,
+        conn: Connection,
+    ) -> _result.Result:
+        update_options = execution_options.get(
+            "_sa_orm_update_options", cls.default_update_options
+        )
+
+        if update_options._dml_strategy == "bulk":
+            raise sa_exc.InvalidRequestError(
+                "Bulk ORM DELETE not supported right now. "
+                "Statement may be invoked at the "
+                "Core level using "
+                "session.connection().execute(stmt, parameters)"
+            )
+
+        if update_options._dml_strategy not in ("orm", "auto", "core_only"):
+            raise sa_exc.ArgumentError(
+                "Valid strategies for ORM DELETE strategy are 'orm', 'auto', "
+                "'core_only'"
+            )
+
+        return super().orm_execute_statement(
+            session, statement, params, execution_options, bind_arguments, conn
+        )
+
+    @classmethod
+    def can_use_returning(
+        cls,
+        dialect: Dialect,
+        mapper: Mapper[Any],
+        *,
+        is_multitable: bool = False,
+        is_update_from: bool = False,
+        is_delete_using: bool = False,
+        is_executemany: bool = False,
+    ) -> bool:
+        # normal answer for "should we use RETURNING" at all.
+        normal_answer = (
+            dialect.delete_returning and mapper.local_table.implicit_returning
+        )
+        if not normal_answer:
+            return False
+
+        # now get into special workarounds because MariaDB supports
+        # DELETE...RETURNING but not DELETE...USING...RETURNING.
+        if is_delete_using:
+            # is_delete_using hint was passed.   use
+            # additional dialect feature (True for PG, False for MariaDB)
+            return dialect.delete_returning_multifrom
+
+        elif is_multitable and not dialect.delete_returning_multifrom:
+            # is_delete_using hint was not passed, but we determined
+            # at compile time that this is in fact a DELETE..USING.
+            # it's too late to continue since we did not pre-SELECT.
+            # raise that we need that hint up front.
+
+            raise sa_exc.CompileError(
+                f'Dialect "{dialect.name}" does not support RETURNING '
+                "with DELETE..USING; for synchronize_session='fetch', "
+                "please add the additional execution option "
+                "'is_delete_using=True' to the statement to indicate that "
+                "a separate SELECT should be used for this backend."
+            )
+
+        return True
+
+    @classmethod
+    def _do_post_synchronize_evaluate(
+        cls, session, statement, result, update_options
+    ):
+        matched_objects = cls._get_matched_objects_on_criteria(
+            update_options,
+            session.identity_map.all_states(),
+        )
+
+        to_delete = []
+
+        for _, state, dict_, is_partially_expired in matched_objects:
+            if is_partially_expired:
+                state._expire(dict_, session.identity_map._modified)
+            else:
+                to_delete.append(state)
+
+        if to_delete:
+            session._remove_newly_deleted(to_delete)
+
+    @classmethod
+    def _do_post_synchronize_fetch(
+        cls, session, statement, result, update_options
+    ):
+        target_mapper = update_options._subject_mapper
+
+        returned_defaults_rows = result.returned_defaults_rows
+
+        if returned_defaults_rows:
+            pk_rows = cls._interpret_returning_rows(
+                result, target_mapper, returned_defaults_rows
+            )
+
+            matched_rows = [
+                tuple(row) + (update_options._identity_token,)
+                for row in pk_rows
+            ]
+        else:
+            matched_rows = update_options._matched_rows
+
+        for row in matched_rows:
+            primary_key = row[0:-1]
+            identity_token = row[-1]
+
+            # TODO: inline this and call remove_newly_deleted
+            # once
+            identity_key = target_mapper.identity_key_from_primary_key(
+                list(primary_key),
+                identity_token=identity_token,
+            )
+            if identity_key in session.identity_map:
+                session._remove_newly_deleted(
+                    [
+                        attributes.instance_state(
+                            session.identity_map[identity_key]
+                        )
+                    ]
+                )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/clsregistry.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/clsregistry.py
new file mode 100644
index 00000000..70307ec7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/clsregistry.py
@@ -0,0 +1,571 @@
+# orm/clsregistry.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Routines to handle the string class registry used by declarative.
+
+This system allows specification of classes and expressions used in
+:func:`_orm.relationship` using strings.
+
+"""
+
+from __future__ import annotations
+
+import re
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generator
+from typing import Iterable
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import NoReturn
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import interfaces
+from .descriptor_props import SynonymProperty
+from .properties import ColumnProperty
+from .util import class_mapper
+from .. import exc
+from .. import inspection
+from .. import util
+from ..sql.schema import _get_table_key
+from ..util.typing import CallableReference
+
+if TYPE_CHECKING:
+    from .relationships import RelationshipProperty
+    from ..sql.schema import MetaData
+    from ..sql.schema import Table
+
+_T = TypeVar("_T", bound=Any)
+
+_ClsRegistryType = MutableMapping[str, Union[type, "ClsRegistryToken"]]
+
+# strong references to registries which we place in
+# the _decl_class_registry, which is usually weak referencing.
+# the internal registries here link to classes with weakrefs and remove
+# themselves when all references to contained classes are removed.
+_registries: Set[ClsRegistryToken] = set()
+
+
+def add_class(
+    classname: str, cls: Type[_T], decl_class_registry: _ClsRegistryType
+) -> None:
+    """Add a class to the _decl_class_registry associated with the
+    given declarative class.
+
+    """
+    if classname in decl_class_registry:
+        # class already exists.
+        existing = decl_class_registry[classname]
+        if not isinstance(existing, _MultipleClassMarker):
+            existing = decl_class_registry[classname] = _MultipleClassMarker(
+                [cls, cast("Type[Any]", existing)]
+            )
+    else:
+        decl_class_registry[classname] = cls
+
+    try:
+        root_module = cast(
+            _ModuleMarker, decl_class_registry["_sa_module_registry"]
+        )
+    except KeyError:
+        decl_class_registry["_sa_module_registry"] = root_module = (
+            _ModuleMarker("_sa_module_registry", None)
+        )
+
+    tokens = cls.__module__.split(".")
+
+    # build up a tree like this:
+    # modulename:  myapp.snacks.nuts
+    #
+    # myapp->snack->nuts->(classes)
+    # snack->nuts->(classes)
+    # nuts->(classes)
+    #
+    # this allows partial token paths to be used.
+    while tokens:
+        token = tokens.pop(0)
+        module = root_module.get_module(token)
+        for token in tokens:
+            module = module.get_module(token)
+
+        try:
+            module.add_class(classname, cls)
+        except AttributeError as ae:
+            if not isinstance(module, _ModuleMarker):
+                raise exc.InvalidRequestError(
+                    f'name "{classname}" matches both a '
+                    "class name and a module name"
+                ) from ae
+            else:
+                raise
+
+
+def remove_class(
+    classname: str, cls: Type[Any], decl_class_registry: _ClsRegistryType
+) -> None:
+    if classname in decl_class_registry:
+        existing = decl_class_registry[classname]
+        if isinstance(existing, _MultipleClassMarker):
+            existing.remove_item(cls)
+        else:
+            del decl_class_registry[classname]
+
+    try:
+        root_module = cast(
+            _ModuleMarker, decl_class_registry["_sa_module_registry"]
+        )
+    except KeyError:
+        return
+
+    tokens = cls.__module__.split(".")
+
+    while tokens:
+        token = tokens.pop(0)
+        module = root_module.get_module(token)
+        for token in tokens:
+            module = module.get_module(token)
+        try:
+            module.remove_class(classname, cls)
+        except AttributeError:
+            if not isinstance(module, _ModuleMarker):
+                pass
+            else:
+                raise
+
+
+def _key_is_empty(
+    key: str,
+    decl_class_registry: _ClsRegistryType,
+    test: Callable[[Any], bool],
+) -> bool:
+    """test if a key is empty of a certain object.
+
+    used for unit tests against the registry to see if garbage collection
+    is working.
+
+    "test" is a callable that will be passed an object should return True
+    if the given object is the one we were looking for.
+
+    We can't pass the actual object itself b.c. this is for testing garbage
+    collection; the caller will have to have removed references to the
+    object itself.
+
+    """
+    if key not in decl_class_registry:
+        return True
+
+    thing = decl_class_registry[key]
+    if isinstance(thing, _MultipleClassMarker):
+        for sub_thing in thing.contents:
+            if test(sub_thing):
+                return False
+        else:
+            raise NotImplementedError("unknown codepath")
+    else:
+        return not test(thing)
+
+
+class ClsRegistryToken:
+    """an object that can be in the registry._class_registry as a value."""
+
+    __slots__ = ()
+
+
+class _MultipleClassMarker(ClsRegistryToken):
+    """refers to multiple classes of the same name
+    within _decl_class_registry.
+
+    """
+
+    __slots__ = "on_remove", "contents", "__weakref__"
+
+    contents: Set[weakref.ref[Type[Any]]]
+    on_remove: CallableReference[Optional[Callable[[], None]]]
+
+    def __init__(
+        self,
+        classes: Iterable[Type[Any]],
+        on_remove: Optional[Callable[[], None]] = None,
+    ):
+        self.on_remove = on_remove
+        self.contents = {
+            weakref.ref(item, self._remove_item) for item in classes
+        }
+        _registries.add(self)
+
+    def remove_item(self, cls: Type[Any]) -> None:
+        self._remove_item(weakref.ref(cls))
+
+    def __iter__(self) -> Generator[Optional[Type[Any]], None, None]:
+        return (ref() for ref in self.contents)
+
+    def attempt_get(self, path: List[str], key: str) -> Type[Any]:
+        if len(self.contents) > 1:
+            raise exc.InvalidRequestError(
+                'Multiple classes found for path "%s" '
+                "in the registry of this declarative "
+                "base. Please use a fully module-qualified path."
+                % (".".join(path + [key]))
+            )
+        else:
+            ref = list(self.contents)[0]
+            cls = ref()
+            if cls is None:
+                raise NameError(key)
+            return cls
+
+    def _remove_item(self, ref: weakref.ref[Type[Any]]) -> None:
+        self.contents.discard(ref)
+        if not self.contents:
+            _registries.discard(self)
+            if self.on_remove:
+                self.on_remove()
+
+    def add_item(self, item: Type[Any]) -> None:
+        # protect against class registration race condition against
+        # asynchronous garbage collection calling _remove_item,
+        # [ticket:3208] and [ticket:10782]
+        modules = {
+            cls.__module__
+            for cls in [ref() for ref in list(self.contents)]
+            if cls is not None
+        }
+        if item.__module__ in modules:
+            util.warn(
+                "This declarative base already contains a class with the "
+                "same class name and module name as %s.%s, and will "
+                "be replaced in the string-lookup table."
+                % (item.__module__, item.__name__)
+            )
+        self.contents.add(weakref.ref(item, self._remove_item))
+
+
+class _ModuleMarker(ClsRegistryToken):
+    """Refers to a module name within
+    _decl_class_registry.
+
+    """
+
+    __slots__ = "parent", "name", "contents", "mod_ns", "path", "__weakref__"
+
+    parent: Optional[_ModuleMarker]
+    contents: Dict[str, Union[_ModuleMarker, _MultipleClassMarker]]
+    mod_ns: _ModNS
+    path: List[str]
+
+    def __init__(self, name: str, parent: Optional[_ModuleMarker]):
+        self.parent = parent
+        self.name = name
+        self.contents = {}
+        self.mod_ns = _ModNS(self)
+        if self.parent:
+            self.path = self.parent.path + [self.name]
+        else:
+            self.path = []
+        _registries.add(self)
+
+    def __contains__(self, name: str) -> bool:
+        return name in self.contents
+
+    def __getitem__(self, name: str) -> ClsRegistryToken:
+        return self.contents[name]
+
+    def _remove_item(self, name: str) -> None:
+        self.contents.pop(name, None)
+        if not self.contents:
+            if self.parent is not None:
+                self.parent._remove_item(self.name)
+            _registries.discard(self)
+
+    def resolve_attr(self, key: str) -> Union[_ModNS, Type[Any]]:
+        return self.mod_ns.__getattr__(key)
+
+    def get_module(self, name: str) -> _ModuleMarker:
+        if name not in self.contents:
+            marker = _ModuleMarker(name, self)
+            self.contents[name] = marker
+        else:
+            marker = cast(_ModuleMarker, self.contents[name])
+        return marker
+
+    def add_class(self, name: str, cls: Type[Any]) -> None:
+        if name in self.contents:
+            existing = cast(_MultipleClassMarker, self.contents[name])
+            try:
+                existing.add_item(cls)
+            except AttributeError as ae:
+                if not isinstance(existing, _MultipleClassMarker):
+                    raise exc.InvalidRequestError(
+                        f'name "{name}" matches both a '
+                        "class name and a module name"
+                    ) from ae
+                else:
+                    raise
+        else:
+            existing = self.contents[name] = _MultipleClassMarker(
+                [cls], on_remove=lambda: self._remove_item(name)
+            )
+
+    def remove_class(self, name: str, cls: Type[Any]) -> None:
+        if name in self.contents:
+            existing = cast(_MultipleClassMarker, self.contents[name])
+            existing.remove_item(cls)
+
+
+class _ModNS:
+    __slots__ = ("__parent",)
+
+    __parent: _ModuleMarker
+
+    def __init__(self, parent: _ModuleMarker):
+        self.__parent = parent
+
+    def __getattr__(self, key: str) -> Union[_ModNS, Type[Any]]:
+        try:
+            value = self.__parent.contents[key]
+        except KeyError:
+            pass
+        else:
+            if value is not None:
+                if isinstance(value, _ModuleMarker):
+                    return value.mod_ns
+                else:
+                    assert isinstance(value, _MultipleClassMarker)
+                    return value.attempt_get(self.__parent.path, key)
+        raise NameError(
+            "Module %r has no mapped classes "
+            "registered under the name %r" % (self.__parent.name, key)
+        )
+
+
+class _GetColumns:
+    __slots__ = ("cls",)
+
+    cls: Type[Any]
+
+    def __init__(self, cls: Type[Any]):
+        self.cls = cls
+
+    def __getattr__(self, key: str) -> Any:
+        mp = class_mapper(self.cls, configure=False)
+        if mp:
+            if key not in mp.all_orm_descriptors:
+                raise AttributeError(
+                    "Class %r does not have a mapped column named %r"
+                    % (self.cls, key)
+                )
+
+            desc = mp.all_orm_descriptors[key]
+            if desc.extension_type is interfaces.NotExtension.NOT_EXTENSION:
+                assert isinstance(desc, attributes.QueryableAttribute)
+                prop = desc.property
+                if isinstance(prop, SynonymProperty):
+                    key = prop.name
+                elif not isinstance(prop, ColumnProperty):
+                    raise exc.InvalidRequestError(
+                        "Property %r is not an instance of"
+                        " ColumnProperty (i.e. does not correspond"
+                        " directly to a Column)." % key
+                    )
+        return getattr(self.cls, key)
+
+
+inspection._inspects(_GetColumns)(
+    lambda target: inspection.inspect(target.cls)
+)
+
+
+class _GetTable:
+    __slots__ = "key", "metadata"
+
+    key: str
+    metadata: MetaData
+
+    def __init__(self, key: str, metadata: MetaData):
+        self.key = key
+        self.metadata = metadata
+
+    def __getattr__(self, key: str) -> Table:
+        return self.metadata.tables[_get_table_key(key, self.key)]
+
+
+def _determine_container(key: str, value: Any) -> _GetColumns:
+    if isinstance(value, _MultipleClassMarker):
+        value = value.attempt_get([], key)
+    return _GetColumns(value)
+
+
+class _class_resolver:
+    __slots__ = (
+        "cls",
+        "prop",
+        "arg",
+        "fallback",
+        "_dict",
+        "_resolvers",
+        "favor_tables",
+    )
+
+    cls: Type[Any]
+    prop: RelationshipProperty[Any]
+    fallback: Mapping[str, Any]
+    arg: str
+    favor_tables: bool
+    _resolvers: Tuple[Callable[[str], Any], ...]
+
+    def __init__(
+        self,
+        cls: Type[Any],
+        prop: RelationshipProperty[Any],
+        fallback: Mapping[str, Any],
+        arg: str,
+        favor_tables: bool = False,
+    ):
+        self.cls = cls
+        self.prop = prop
+        self.arg = arg
+        self.fallback = fallback
+        self._dict = util.PopulateDict(self._access_cls)
+        self._resolvers = ()
+        self.favor_tables = favor_tables
+
+    def _access_cls(self, key: str) -> Any:
+        cls = self.cls
+
+        manager = attributes.manager_of_class(cls)
+        decl_base = manager.registry
+        assert decl_base is not None
+        decl_class_registry = decl_base._class_registry
+        metadata = decl_base.metadata
+
+        if self.favor_tables:
+            if key in metadata.tables:
+                return metadata.tables[key]
+            elif key in metadata._schemas:
+                return _GetTable(key, getattr(cls, "metadata", metadata))
+
+        if key in decl_class_registry:
+            return _determine_container(key, decl_class_registry[key])
+
+        if not self.favor_tables:
+            if key in metadata.tables:
+                return metadata.tables[key]
+            elif key in metadata._schemas:
+                return _GetTable(key, getattr(cls, "metadata", metadata))
+
+        if "_sa_module_registry" in decl_class_registry and key in cast(
+            _ModuleMarker, decl_class_registry["_sa_module_registry"]
+        ):
+            registry = cast(
+                _ModuleMarker, decl_class_registry["_sa_module_registry"]
+            )
+            return registry.resolve_attr(key)
+        elif self._resolvers:
+            for resolv in self._resolvers:
+                value = resolv(key)
+                if value is not None:
+                    return value
+
+        return self.fallback[key]
+
+    def _raise_for_name(self, name: str, err: Exception) -> NoReturn:
+        generic_match = re.match(r"(.+)\[(.+)\]", name)
+
+        if generic_match:
+            clsarg = generic_match.group(2).strip("'")
+            raise exc.InvalidRequestError(
+                f"When initializing mapper {self.prop.parent}, "
+                f'expression "relationship({self.arg!r})" seems to be '
+                "using a generic class as the argument to relationship(); "
+                "please state the generic argument "
+                "using an annotation, e.g. "
+                f'"{self.prop.key}: Mapped[{generic_match.group(1)}'
+                f"['{clsarg}']] = relationship()\""
+            ) from err
+        else:
+            raise exc.InvalidRequestError(
+                "When initializing mapper %s, expression %r failed to "
+                "locate a name (%r). If this is a class name, consider "
+                "adding this relationship() to the %r class after "
+                "both dependent classes have been defined."
+                % (self.prop.parent, self.arg, name, self.cls)
+            ) from err
+
+    def _resolve_name(self) -> Union[Table, Type[Any], _ModNS]:
+        name = self.arg
+        d = self._dict
+        rval = None
+        try:
+            for token in name.split("."):
+                if rval is None:
+                    rval = d[token]
+                else:
+                    rval = getattr(rval, token)
+        except KeyError as err:
+            self._raise_for_name(name, err)
+        except NameError as n:
+            self._raise_for_name(n.args[0], n)
+        else:
+            if isinstance(rval, _GetColumns):
+                return rval.cls
+            else:
+                if TYPE_CHECKING:
+                    assert isinstance(rval, (type, Table, _ModNS))
+                return rval
+
+    def __call__(self) -> Any:
+        try:
+            x = eval(self.arg, globals(), self._dict)
+
+            if isinstance(x, _GetColumns):
+                return x.cls
+            else:
+                return x
+        except NameError as n:
+            self._raise_for_name(n.args[0], n)
+
+
+_fallback_dict: Mapping[str, Any] = None  # type: ignore
+
+
+def _resolver(cls: Type[Any], prop: RelationshipProperty[Any]) -> Tuple[
+    Callable[[str], Callable[[], Union[Type[Any], Table, _ModNS]]],
+    Callable[[str, bool], _class_resolver],
+]:
+    global _fallback_dict
+
+    if _fallback_dict is None:
+        import sqlalchemy
+        from . import foreign
+        from . import remote
+
+        _fallback_dict = util.immutabledict(sqlalchemy.__dict__).union(
+            {"foreign": foreign, "remote": remote}
+        )
+
+    def resolve_arg(arg: str, favor_tables: bool = False) -> _class_resolver:
+        return _class_resolver(
+            cls, prop, _fallback_dict, arg, favor_tables=favor_tables
+        )
+
+    def resolve_name(
+        arg: str,
+    ) -> Callable[[], Union[Type[Any], Table, _ModNS]]:
+        return _class_resolver(cls, prop, _fallback_dict, arg)._resolve_name
+
+    return resolve_name, resolve_arg
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/collections.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/collections.py
new file mode 100644
index 00000000..336b1133
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/collections.py
@@ -0,0 +1,1627 @@
+# orm/collections.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Support for collections of mapped entities.
+
+The collections package supplies the machinery used to inform the ORM of
+collection membership changes.  An instrumentation via decoration approach is
+used, allowing arbitrary types (including built-ins) to be used as entity
+collections without requiring inheritance from a base class.
+
+Instrumentation decoration relays membership change events to the
+:class:`.CollectionAttributeImpl` that is currently managing the collection.
+The decorators observe function call arguments and return values, tracking
+entities entering or leaving the collection.  Two decorator approaches are
+provided.  One is a bundle of generic decorators that map function arguments
+and return values to events::
+
+  from sqlalchemy.orm.collections import collection
+
+
+  class MyClass:
+      # ...
+
+      @collection.adds(1)
+      def store(self, item):
+          self.data.append(item)
+
+      @collection.removes_return()
+      def pop(self):
+          return self.data.pop()
+
+The second approach is a bundle of targeted decorators that wrap appropriate
+append and remove notifiers around the mutation methods present in the
+standard Python ``list``, ``set`` and ``dict`` interfaces.  These could be
+specified in terms of generic decorator recipes, but are instead hand-tooled
+for increased efficiency.  The targeted decorators occasionally implement
+adapter-like behavior, such as mapping bulk-set methods (``extend``,
+``update``, ``__setslice__``, etc.) into the series of atomic mutation events
+that the ORM requires.
+
+The targeted decorators are used internally for automatic instrumentation of
+entity collection classes.  Every collection class goes through a
+transformation process roughly like so:
+
+1. If the class is a built-in, substitute a trivial sub-class
+2. Is this class already instrumented?
+3. Add in generic decorators
+4. Sniff out the collection interface through duck-typing
+5. Add targeted decoration to any undecorated interface method
+
+This process modifies the class at runtime, decorating methods and adding some
+bookkeeping properties.  This isn't possible (or desirable) for built-in
+classes like ``list``, so trivial sub-classes are substituted to hold
+decoration::
+
+  class InstrumentedList(list):
+      pass
+
+Collection classes can be specified in ``relationship(collection_class=)`` as
+types or a function that returns an instance.  Collection classes are
+inspected and instrumented during the mapper compilation phase.  The
+collection_class callable will be executed once to produce a specimen
+instance, and the type of that specimen will be instrumented.  Functions that
+return built-in types like ``lists`` will be adapted to produce instrumented
+instances.
+
+When extending a known type like ``list``, additional decorations are not
+generally not needed.  Odds are, the extension method will delegate to a
+method that's already instrumented.  For example::
+
+  class QueueIsh(list):
+      def push(self, item):
+          self.append(item)
+
+      def shift(self):
+          return self.pop(0)
+
+There's no need to decorate these methods.  ``append`` and ``pop`` are already
+instrumented as part of the ``list`` interface.  Decorating them would fire
+duplicate events, which should be avoided.
+
+The targeted decoration tries not to rely on other methods in the underlying
+collection class, but some are unavoidable.  Many depend on 'read' methods
+being present to properly instrument a 'write', for example, ``__setitem__``
+needs ``__getitem__``.  "Bulk" methods like ``update`` and ``extend`` may also
+reimplemented in terms of atomic appends and removes, so the ``extend``
+decoration will actually perform many ``append`` operations and not call the
+underlying method at all.
+
+Tight control over bulk operation and the firing of events is also possible by
+implementing the instrumentation internally in your methods.  The basic
+instrumentation package works under the general assumption that collection
+mutation will not raise unusual exceptions.  If you want to closely
+orchestrate append and remove events with exception management, internal
+instrumentation may be the answer.  Within your method,
+``collection_adapter(self)`` will retrieve an object that you can use for
+explicit control over triggering append and remove events.
+
+The owning object and :class:`.CollectionAttributeImpl` are also reachable
+through the adapter, allowing for some very sophisticated behavior.
+
+"""
+from __future__ import annotations
+
+import operator
+import threading
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from .base import NO_KEY
+from .. import exc as sa_exc
+from .. import util
+from ..sql.base import NO_ARG
+from ..util.compat import inspect_getfullargspec
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from .attributes import AttributeEventToken
+    from .attributes import CollectionAttributeImpl
+    from .mapped_collection import attribute_keyed_dict
+    from .mapped_collection import column_keyed_dict
+    from .mapped_collection import keyfunc_mapping
+    from .mapped_collection import KeyFuncDict  # noqa: F401
+    from .state import InstanceState
+
+
+__all__ = [
+    "collection",
+    "collection_adapter",
+    "keyfunc_mapping",
+    "column_keyed_dict",
+    "attribute_keyed_dict",
+    "KeyFuncDict",
+    # old names in < 2.0
+    "mapped_collection",
+    "column_mapped_collection",
+    "attribute_mapped_collection",
+    "MappedCollection",
+]
+
+__instrumentation_mutex = threading.Lock()
+
+
+_CollectionFactoryType = Callable[[], "_AdaptedCollectionProtocol"]
+
+_T = TypeVar("_T", bound=Any)
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+_COL = TypeVar("_COL", bound="Collection[Any]")
+_FN = TypeVar("_FN", bound="Callable[..., Any]")
+
+
+class _CollectionConverterProtocol(Protocol):
+    def __call__(self, collection: _COL) -> _COL: ...
+
+
+class _AdaptedCollectionProtocol(Protocol):
+    _sa_adapter: CollectionAdapter
+    _sa_appender: Callable[..., Any]
+    _sa_remover: Callable[..., Any]
+    _sa_iterator: Callable[..., Iterable[Any]]
+    _sa_converter: _CollectionConverterProtocol
+
+
+class collection:
+    """Decorators for entity collection classes.
+
+    The decorators fall into two groups: annotations and interception recipes.
+
+    The annotating decorators (appender, remover, iterator, converter,
+    internally_instrumented) indicate the method's purpose and take no
+    arguments.  They are not written with parens::
+
+        @collection.appender
+        def append(self, append): ...
+
+    The recipe decorators all require parens, even those that take no
+    arguments::
+
+        @collection.adds("entity")
+        def insert(self, position, entity): ...
+
+
+        @collection.removes_return()
+        def popitem(self): ...
+
+    """
+
+    # Bundled as a class solely for ease of use: packaging, doc strings,
+    # importability.
+
+    @staticmethod
+    def appender(fn):
+        """Tag the method as the collection appender.
+
+        The appender method is called with one positional argument: the value
+        to append. The method will be automatically decorated with 'adds(1)'
+        if not already decorated::
+
+            @collection.appender
+            def add(self, append): ...
+
+
+            # or, equivalently
+            @collection.appender
+            @collection.adds(1)
+            def add(self, append): ...
+
+
+            # for mapping type, an 'append' may kick out a previous value
+            # that occupies that slot.  consider d['a'] = 'foo'- any previous
+            # value in d['a'] is discarded.
+            @collection.appender
+            @collection.replaces(1)
+            def add(self, entity):
+                key = some_key_func(entity)
+                previous = None
+                if key in self:
+                    previous = self[key]
+                self[key] = entity
+                return previous
+
+        If the value to append is not allowed in the collection, you may
+        raise an exception.  Something to remember is that the appender
+        will be called for each object mapped by a database query.  If the
+        database contains rows that violate your collection semantics, you
+        will need to get creative to fix the problem, as access via the
+        collection will not work.
+
+        If the appender method is internally instrumented, you must also
+        receive the keyword argument '_sa_initiator' and ensure its
+        promulgation to collection events.
+
+        """
+        fn._sa_instrument_role = "appender"
+        return fn
+
+    @staticmethod
+    def remover(fn):
+        """Tag the method as the collection remover.
+
+        The remover method is called with one positional argument: the value
+        to remove. The method will be automatically decorated with
+        :meth:`removes_return` if not already decorated::
+
+            @collection.remover
+            def zap(self, entity): ...
+
+
+            # or, equivalently
+            @collection.remover
+            @collection.removes_return()
+            def zap(self): ...
+
+        If the value to remove is not present in the collection, you may
+        raise an exception or return None to ignore the error.
+
+        If the remove method is internally instrumented, you must also
+        receive the keyword argument '_sa_initiator' and ensure its
+        promulgation to collection events.
+
+        """
+        fn._sa_instrument_role = "remover"
+        return fn
+
+    @staticmethod
+    def iterator(fn):
+        """Tag the method as the collection remover.
+
+        The iterator method is called with no arguments.  It is expected to
+        return an iterator over all collection members::
+
+            @collection.iterator
+            def __iter__(self): ...
+
+        """
+        fn._sa_instrument_role = "iterator"
+        return fn
+
+    @staticmethod
+    def internally_instrumented(fn):
+        """Tag the method as instrumented.
+
+        This tag will prevent any decoration from being applied to the
+        method. Use this if you are orchestrating your own calls to
+        :func:`.collection_adapter` in one of the basic SQLAlchemy
+        interface methods, or to prevent an automatic ABC method
+        decoration from wrapping your implementation::
+
+            # normally an 'extend' method on a list-like class would be
+            # automatically intercepted and re-implemented in terms of
+            # SQLAlchemy events and append().  your implementation will
+            # never be called, unless:
+            @collection.internally_instrumented
+            def extend(self, items): ...
+
+        """
+        fn._sa_instrumented = True
+        return fn
+
+    @staticmethod
+    @util.deprecated(
+        "1.3",
+        "The :meth:`.collection.converter` handler is deprecated and will "
+        "be removed in a future release.  Please refer to the "
+        ":class:`.AttributeEvents.bulk_replace` listener interface in "
+        "conjunction with the :func:`.event.listen` function.",
+    )
+    def converter(fn):
+        """Tag the method as the collection converter.
+
+        This optional method will be called when a collection is being
+        replaced entirely, as in::
+
+            myobj.acollection = [newvalue1, newvalue2]
+
+        The converter method will receive the object being assigned and should
+        return an iterable of values suitable for use by the ``appender``
+        method.  A converter must not assign values or mutate the collection,
+        its sole job is to adapt the value the user provides into an iterable
+        of values for the ORM's use.
+
+        The default converter implementation will use duck-typing to do the
+        conversion.  A dict-like collection will be convert into an iterable
+        of dictionary values, and other types will simply be iterated::
+
+            @collection.converter
+            def convert(self, other): ...
+
+        If the duck-typing of the object does not match the type of this
+        collection, a TypeError is raised.
+
+        Supply an implementation of this method if you want to expand the
+        range of possible types that can be assigned in bulk or perform
+        validation on the values about to be assigned.
+
+        """
+        fn._sa_instrument_role = "converter"
+        return fn
+
+    @staticmethod
+    def adds(arg):
+        """Mark the method as adding an entity to the collection.
+
+        Adds "add to collection" handling to the method.  The decorator
+        argument indicates which method argument holds the SQLAlchemy-relevant
+        value.  Arguments can be specified positionally (i.e. integer) or by
+        name::
+
+            @collection.adds(1)
+            def push(self, item): ...
+
+
+            @collection.adds("entity")
+            def do_stuff(self, thing, entity=None): ...
+
+        """
+
+        def decorator(fn):
+            fn._sa_instrument_before = ("fire_append_event", arg)
+            return fn
+
+        return decorator
+
+    @staticmethod
+    def replaces(arg):
+        """Mark the method as replacing an entity in the collection.
+
+        Adds "add to collection" and "remove from collection" handling to
+        the method.  The decorator argument indicates which method argument
+        holds the SQLAlchemy-relevant value to be added, and return value, if
+        any will be considered the value to remove.
+
+        Arguments can be specified positionally (i.e. integer) or by name::
+
+            @collection.replaces(2)
+            def __setitem__(self, index, item): ...
+
+        """
+
+        def decorator(fn):
+            fn._sa_instrument_before = ("fire_append_event", arg)
+            fn._sa_instrument_after = "fire_remove_event"
+            return fn
+
+        return decorator
+
+    @staticmethod
+    def removes(arg):
+        """Mark the method as removing an entity in the collection.
+
+        Adds "remove from collection" handling to the method.  The decorator
+        argument indicates which method argument holds the SQLAlchemy-relevant
+        value to be removed. Arguments can be specified positionally (i.e.
+        integer) or by name::
+
+            @collection.removes(1)
+            def zap(self, item): ...
+
+        For methods where the value to remove is not known at call-time, use
+        collection.removes_return.
+
+        """
+
+        def decorator(fn):
+            fn._sa_instrument_before = ("fire_remove_event", arg)
+            return fn
+
+        return decorator
+
+    @staticmethod
+    def removes_return():
+        """Mark the method as removing an entity in the collection.
+
+        Adds "remove from collection" handling to the method.  The return
+        value of the method, if any, is considered the value to remove.  The
+        method arguments are not inspected::
+
+            @collection.removes_return()
+            def pop(self): ...
+
+        For methods where the value to remove is known at call-time, use
+        collection.remove.
+
+        """
+
+        def decorator(fn):
+            fn._sa_instrument_after = "fire_remove_event"
+            return fn
+
+        return decorator
+
+
+if TYPE_CHECKING:
+
+    def collection_adapter(collection: Collection[Any]) -> CollectionAdapter:
+        """Fetch the :class:`.CollectionAdapter` for a collection."""
+
+else:
+    collection_adapter = operator.attrgetter("_sa_adapter")
+
+
+class CollectionAdapter:
+    """Bridges between the ORM and arbitrary Python collections.
+
+    Proxies base-level collection operations (append, remove, iterate)
+    to the underlying Python collection, and emits add/remove events for
+    entities entering or leaving the collection.
+
+    The ORM uses :class:`.CollectionAdapter` exclusively for interaction with
+    entity collections.
+
+
+    """
+
+    __slots__ = (
+        "attr",
+        "_key",
+        "_data",
+        "owner_state",
+        "_converter",
+        "invalidated",
+        "empty",
+    )
+
+    attr: CollectionAttributeImpl
+    _key: str
+
+    # this is actually a weakref; see note in constructor
+    _data: Callable[..., _AdaptedCollectionProtocol]
+
+    owner_state: InstanceState[Any]
+    _converter: _CollectionConverterProtocol
+    invalidated: bool
+    empty: bool
+
+    def __init__(
+        self,
+        attr: CollectionAttributeImpl,
+        owner_state: InstanceState[Any],
+        data: _AdaptedCollectionProtocol,
+    ):
+        self.attr = attr
+        self._key = attr.key
+
+        # this weakref stays referenced throughout the lifespan of
+        # CollectionAdapter.  so while the weakref can return None, this
+        # is realistically only during garbage collection of this object, so
+        # we type this as a callable that returns _AdaptedCollectionProtocol
+        # in all cases.
+        self._data = weakref.ref(data)  # type: ignore
+
+        self.owner_state = owner_state
+        data._sa_adapter = self
+        self._converter = data._sa_converter
+        self.invalidated = False
+        self.empty = False
+
+    def _warn_invalidated(self) -> None:
+        util.warn("This collection has been invalidated.")
+
+    @property
+    def data(self) -> _AdaptedCollectionProtocol:
+        "The entity collection being adapted."
+        return self._data()
+
+    @property
+    def _referenced_by_owner(self) -> bool:
+        """return True if the owner state still refers to this collection.
+
+        This will return False within a bulk replace operation,
+        where this collection is the one being replaced.
+
+        """
+        return self.owner_state.dict[self._key] is self._data()
+
+    def bulk_appender(self):
+        return self._data()._sa_appender
+
+    def append_with_event(
+        self, item: Any, initiator: Optional[AttributeEventToken] = None
+    ) -> None:
+        """Add an entity to the collection, firing mutation events."""
+
+        self._data()._sa_appender(item, _sa_initiator=initiator)
+
+    def _set_empty(self, user_data):
+        assert (
+            not self.empty
+        ), "This collection adapter is already in the 'empty' state"
+        self.empty = True
+        self.owner_state._empty_collections[self._key] = user_data
+
+    def _reset_empty(self) -> None:
+        assert (
+            self.empty
+        ), "This collection adapter is not in the 'empty' state"
+        self.empty = False
+        self.owner_state.dict[self._key] = (
+            self.owner_state._empty_collections.pop(self._key)
+        )
+
+    def _refuse_empty(self) -> NoReturn:
+        raise sa_exc.InvalidRequestError(
+            "This is a special 'empty' collection which cannot accommodate "
+            "internal mutation operations"
+        )
+
+    def append_without_event(self, item: Any) -> None:
+        """Add or restore an entity to the collection, firing no events."""
+
+        if self.empty:
+            self._refuse_empty()
+        self._data()._sa_appender(item, _sa_initiator=False)
+
+    def append_multiple_without_event(self, items: Iterable[Any]) -> None:
+        """Add or restore an entity to the collection, firing no events."""
+        if self.empty:
+            self._refuse_empty()
+        appender = self._data()._sa_appender
+        for item in items:
+            appender(item, _sa_initiator=False)
+
+    def bulk_remover(self):
+        return self._data()._sa_remover
+
+    def remove_with_event(
+        self, item: Any, initiator: Optional[AttributeEventToken] = None
+    ) -> None:
+        """Remove an entity from the collection, firing mutation events."""
+        self._data()._sa_remover(item, _sa_initiator=initiator)
+
+    def remove_without_event(self, item: Any) -> None:
+        """Remove an entity from the collection, firing no events."""
+        if self.empty:
+            self._refuse_empty()
+        self._data()._sa_remover(item, _sa_initiator=False)
+
+    def clear_with_event(
+        self, initiator: Optional[AttributeEventToken] = None
+    ) -> None:
+        """Empty the collection, firing a mutation event for each entity."""
+
+        if self.empty:
+            self._refuse_empty()
+        remover = self._data()._sa_remover
+        for item in list(self):
+            remover(item, _sa_initiator=initiator)
+
+    def clear_without_event(self) -> None:
+        """Empty the collection, firing no events."""
+
+        if self.empty:
+            self._refuse_empty()
+        remover = self._data()._sa_remover
+        for item in list(self):
+            remover(item, _sa_initiator=False)
+
+    def __iter__(self):
+        """Iterate over entities in the collection."""
+
+        return iter(self._data()._sa_iterator())
+
+    def __len__(self):
+        """Count entities in the collection."""
+        return len(list(self._data()._sa_iterator()))
+
+    def __bool__(self):
+        return True
+
+    def _fire_append_wo_mutation_event_bulk(
+        self, items, initiator=None, key=NO_KEY
+    ):
+        if not items:
+            return
+
+        if initiator is not False:
+            if self.invalidated:
+                self._warn_invalidated()
+
+            if self.empty:
+                self._reset_empty()
+
+            for item in items:
+                self.attr.fire_append_wo_mutation_event(
+                    self.owner_state,
+                    self.owner_state.dict,
+                    item,
+                    initiator,
+                    key,
+                )
+
+    def fire_append_wo_mutation_event(self, item, initiator=None, key=NO_KEY):
+        """Notify that a entity is entering the collection but is already
+        present.
+
+
+        Initiator is a token owned by the InstrumentedAttribute that
+        initiated the membership mutation, and should be left as None
+        unless you are passing along an initiator value from a chained
+        operation.
+
+        .. versionadded:: 1.4.15
+
+        """
+        if initiator is not False:
+            if self.invalidated:
+                self._warn_invalidated()
+
+            if self.empty:
+                self._reset_empty()
+
+            return self.attr.fire_append_wo_mutation_event(
+                self.owner_state, self.owner_state.dict, item, initiator, key
+            )
+        else:
+            return item
+
+    def fire_append_event(self, item, initiator=None, key=NO_KEY):
+        """Notify that a entity has entered the collection.
+
+        Initiator is a token owned by the InstrumentedAttribute that
+        initiated the membership mutation, and should be left as None
+        unless you are passing along an initiator value from a chained
+        operation.
+
+        """
+        if initiator is not False:
+            if self.invalidated:
+                self._warn_invalidated()
+
+            if self.empty:
+                self._reset_empty()
+
+            return self.attr.fire_append_event(
+                self.owner_state, self.owner_state.dict, item, initiator, key
+            )
+        else:
+            return item
+
+    def _fire_remove_event_bulk(self, items, initiator=None, key=NO_KEY):
+        if not items:
+            return
+
+        if initiator is not False:
+            if self.invalidated:
+                self._warn_invalidated()
+
+            if self.empty:
+                self._reset_empty()
+
+            for item in items:
+                self.attr.fire_remove_event(
+                    self.owner_state,
+                    self.owner_state.dict,
+                    item,
+                    initiator,
+                    key,
+                )
+
+    def fire_remove_event(self, item, initiator=None, key=NO_KEY):
+        """Notify that a entity has been removed from the collection.
+
+        Initiator is the InstrumentedAttribute that initiated the membership
+        mutation, and should be left as None unless you are passing along
+        an initiator value from a chained operation.
+
+        """
+        if initiator is not False:
+            if self.invalidated:
+                self._warn_invalidated()
+
+            if self.empty:
+                self._reset_empty()
+
+            self.attr.fire_remove_event(
+                self.owner_state, self.owner_state.dict, item, initiator, key
+            )
+
+    def fire_pre_remove_event(self, initiator=None, key=NO_KEY):
+        """Notify that an entity is about to be removed from the collection.
+
+        Only called if the entity cannot be removed after calling
+        fire_remove_event().
+
+        """
+        if self.invalidated:
+            self._warn_invalidated()
+        self.attr.fire_pre_remove_event(
+            self.owner_state,
+            self.owner_state.dict,
+            initiator=initiator,
+            key=key,
+        )
+
+    def __getstate__(self):
+        return {
+            "key": self._key,
+            "owner_state": self.owner_state,
+            "owner_cls": self.owner_state.class_,
+            "data": self.data,
+            "invalidated": self.invalidated,
+            "empty": self.empty,
+        }
+
+    def __setstate__(self, d):
+        self._key = d["key"]
+        self.owner_state = d["owner_state"]
+
+        # see note in constructor regarding this type: ignore
+        self._data = weakref.ref(d["data"])  # type: ignore
+
+        self._converter = d["data"]._sa_converter
+        d["data"]._sa_adapter = self
+        self.invalidated = d["invalidated"]
+        self.attr = getattr(d["owner_cls"], self._key).impl
+        self.empty = d.get("empty", False)
+
+
+def bulk_replace(values, existing_adapter, new_adapter, initiator=None):
+    """Load a new collection, firing events based on prior like membership.
+
+    Appends instances in ``values`` onto the ``new_adapter``. Events will be
+    fired for any instance not present in the ``existing_adapter``.  Any
+    instances in ``existing_adapter`` not present in ``values`` will have
+    remove events fired upon them.
+
+    :param values: An iterable of collection member instances
+
+    :param existing_adapter: A :class:`.CollectionAdapter` of
+     instances to be replaced
+
+    :param new_adapter: An empty :class:`.CollectionAdapter`
+     to load with ``values``
+
+
+    """
+
+    assert isinstance(values, list)
+
+    idset = util.IdentitySet
+    existing_idset = idset(existing_adapter or ())
+    constants = existing_idset.intersection(values or ())
+    additions = idset(values or ()).difference(constants)
+    removals = existing_idset.difference(constants)
+
+    appender = new_adapter.bulk_appender()
+
+    for member in values or ():
+        if member in additions:
+            appender(member, _sa_initiator=initiator)
+        elif member in constants:
+            appender(member, _sa_initiator=False)
+
+    if existing_adapter:
+        existing_adapter._fire_append_wo_mutation_event_bulk(
+            constants, initiator=initiator
+        )
+        existing_adapter._fire_remove_event_bulk(removals, initiator=initiator)
+
+
+def prepare_instrumentation(
+    factory: Union[Type[Collection[Any]], _CollectionFactoryType],
+) -> _CollectionFactoryType:
+    """Prepare a callable for future use as a collection class factory.
+
+    Given a collection class factory (either a type or no-arg callable),
+    return another factory that will produce compatible instances when
+    called.
+
+    This function is responsible for converting collection_class=list
+    into the run-time behavior of collection_class=InstrumentedList.
+
+    """
+
+    impl_factory: _CollectionFactoryType
+
+    # Convert a builtin to 'Instrumented*'
+    if factory in __canned_instrumentation:
+        impl_factory = __canned_instrumentation[factory]
+    else:
+        impl_factory = cast(_CollectionFactoryType, factory)
+
+    cls: Union[_CollectionFactoryType, Type[Collection[Any]]]
+
+    # Create a specimen
+    cls = type(impl_factory())
+
+    # Did factory callable return a builtin?
+    if cls in __canned_instrumentation:
+        # if so, just convert.
+        # in previous major releases, this codepath wasn't working and was
+        # not covered by tests.   prior to that it supplied a "wrapper"
+        # function that would return the class, though the rationale for this
+        # case is not known
+        impl_factory = __canned_instrumentation[cls]
+        cls = type(impl_factory())
+
+    # Instrument the class if needed.
+    if __instrumentation_mutex.acquire():
+        try:
+            if getattr(cls, "_sa_instrumented", None) != id(cls):
+                _instrument_class(cls)
+        finally:
+            __instrumentation_mutex.release()
+
+    return impl_factory
+
+
+def _instrument_class(cls):
+    """Modify methods in a class and install instrumentation."""
+
+    # In the normal call flow, a request for any of the 3 basic collection
+    # types is transformed into one of our trivial subclasses
+    # (e.g. InstrumentedList).  Catch anything else that sneaks in here...
+    if cls.__module__ == "__builtin__":
+        raise sa_exc.ArgumentError(
+            "Can not instrument a built-in type. Use a "
+            "subclass, even a trivial one."
+        )
+
+    roles, methods = _locate_roles_and_methods(cls)
+
+    _setup_canned_roles(cls, roles, methods)
+
+    _assert_required_roles(cls, roles, methods)
+
+    _set_collection_attributes(cls, roles, methods)
+
+
+def _locate_roles_and_methods(cls):
+    """search for _sa_instrument_role-decorated methods in
+    method resolution order, assign to roles.
+
+    """
+
+    roles: Dict[str, str] = {}
+    methods: Dict[str, Tuple[Optional[str], Optional[int], Optional[str]]] = {}
+
+    for supercls in cls.__mro__:
+        for name, method in vars(supercls).items():
+            if not callable(method):
+                continue
+
+            # note role declarations
+            if hasattr(method, "_sa_instrument_role"):
+                role = method._sa_instrument_role
+                assert role in (
+                    "appender",
+                    "remover",
+                    "iterator",
+                    "converter",
+                )
+                roles.setdefault(role, name)
+
+            # transfer instrumentation requests from decorated function
+            # to the combined queue
+            before: Optional[Tuple[str, int]] = None
+            after: Optional[str] = None
+
+            if hasattr(method, "_sa_instrument_before"):
+                op, argument = method._sa_instrument_before
+                assert op in ("fire_append_event", "fire_remove_event")
+                before = op, argument
+            if hasattr(method, "_sa_instrument_after"):
+                op = method._sa_instrument_after
+                assert op in ("fire_append_event", "fire_remove_event")
+                after = op
+            if before:
+                methods[name] = before + (after,)
+            elif after:
+                methods[name] = None, None, after
+    return roles, methods
+
+
+def _setup_canned_roles(cls, roles, methods):
+    """see if this class has "canned" roles based on a known
+    collection type (dict, set, list).  Apply those roles
+    as needed to the "roles" dictionary, and also
+    prepare "decorator" methods
+
+    """
+    collection_type = util.duck_type_collection(cls)
+    if collection_type in __interfaces:
+        assert collection_type is not None
+        canned_roles, decorators = __interfaces[collection_type]
+        for role, name in canned_roles.items():
+            roles.setdefault(role, name)
+
+        # apply ABC auto-decoration to methods that need it
+        for method, decorator in decorators.items():
+            fn = getattr(cls, method, None)
+            if (
+                fn
+                and method not in methods
+                and not hasattr(fn, "_sa_instrumented")
+            ):
+                setattr(cls, method, decorator(fn))
+
+
+def _assert_required_roles(cls, roles, methods):
+    """ensure all roles are present, and apply implicit instrumentation if
+    needed
+
+    """
+    if "appender" not in roles or not hasattr(cls, roles["appender"]):
+        raise sa_exc.ArgumentError(
+            "Type %s must elect an appender method to be "
+            "a collection class" % cls.__name__
+        )
+    elif roles["appender"] not in methods and not hasattr(
+        getattr(cls, roles["appender"]), "_sa_instrumented"
+    ):
+        methods[roles["appender"]] = ("fire_append_event", 1, None)
+
+    if "remover" not in roles or not hasattr(cls, roles["remover"]):
+        raise sa_exc.ArgumentError(
+            "Type %s must elect a remover method to be "
+            "a collection class" % cls.__name__
+        )
+    elif roles["remover"] not in methods and not hasattr(
+        getattr(cls, roles["remover"]), "_sa_instrumented"
+    ):
+        methods[roles["remover"]] = ("fire_remove_event", 1, None)
+
+    if "iterator" not in roles or not hasattr(cls, roles["iterator"]):
+        raise sa_exc.ArgumentError(
+            "Type %s must elect an iterator method to be "
+            "a collection class" % cls.__name__
+        )
+
+
+def _set_collection_attributes(cls, roles, methods):
+    """apply ad-hoc instrumentation from decorators, class-level defaults
+    and implicit role declarations
+
+    """
+    for method_name, (before, argument, after) in methods.items():
+        setattr(
+            cls,
+            method_name,
+            _instrument_membership_mutator(
+                getattr(cls, method_name), before, argument, after
+            ),
+        )
+    # intern the role map
+    for role, method_name in roles.items():
+        setattr(cls, "_sa_%s" % role, getattr(cls, method_name))
+
+    cls._sa_adapter = None
+
+    if not hasattr(cls, "_sa_converter"):
+        cls._sa_converter = None
+    cls._sa_instrumented = id(cls)
+
+
+def _instrument_membership_mutator(method, before, argument, after):
+    """Route method args and/or return value through the collection
+    adapter."""
+    # This isn't smart enough to handle @adds(1) for 'def fn(self, (a, b))'
+    if before:
+        fn_args = list(
+            util.flatten_iterator(inspect_getfullargspec(method)[0])
+        )
+        if isinstance(argument, int):
+            pos_arg = argument
+            named_arg = len(fn_args) > argument and fn_args[argument] or None
+        else:
+            if argument in fn_args:
+                pos_arg = fn_args.index(argument)
+            else:
+                pos_arg = None
+            named_arg = argument
+        del fn_args
+
+    def wrapper(*args, **kw):
+        if before:
+            if pos_arg is None:
+                if named_arg not in kw:
+                    raise sa_exc.ArgumentError(
+                        "Missing argument %s" % argument
+                    )
+                value = kw[named_arg]
+            else:
+                if len(args) > pos_arg:
+                    value = args[pos_arg]
+                elif named_arg in kw:
+                    value = kw[named_arg]
+                else:
+                    raise sa_exc.ArgumentError(
+                        "Missing argument %s" % argument
+                    )
+
+        initiator = kw.pop("_sa_initiator", None)
+        if initiator is False:
+            executor = None
+        else:
+            executor = args[0]._sa_adapter
+
+        if before and executor:
+            getattr(executor, before)(value, initiator)
+
+        if not after or not executor:
+            return method(*args, **kw)
+        else:
+            res = method(*args, **kw)
+            if res is not None:
+                getattr(executor, after)(res, initiator)
+            return res
+
+    wrapper._sa_instrumented = True  # type: ignore[attr-defined]
+    if hasattr(method, "_sa_instrument_role"):
+        wrapper._sa_instrument_role = method._sa_instrument_role  # type: ignore[attr-defined]  # noqa: E501
+    wrapper.__name__ = method.__name__
+    wrapper.__doc__ = method.__doc__
+    return wrapper
+
+
+def __set_wo_mutation(collection, item, _sa_initiator=None):
+    """Run set wo mutation events.
+
+    The collection is not mutated.
+
+    """
+    if _sa_initiator is not False:
+        executor = collection._sa_adapter
+        if executor:
+            executor.fire_append_wo_mutation_event(
+                item, _sa_initiator, key=None
+            )
+
+
+def __set(collection, item, _sa_initiator, key):
+    """Run set events.
+
+    This event always occurs before the collection is actually mutated.
+
+    """
+
+    if _sa_initiator is not False:
+        executor = collection._sa_adapter
+        if executor:
+            item = executor.fire_append_event(item, _sa_initiator, key=key)
+    return item
+
+
+def __del(collection, item, _sa_initiator, key):
+    """Run del events.
+
+    This event occurs before the collection is actually mutated, *except*
+    in the case of a pop operation, in which case it occurs afterwards.
+    For pop operations, the __before_pop hook is called before the
+    operation occurs.
+
+    """
+    if _sa_initiator is not False:
+        executor = collection._sa_adapter
+        if executor:
+            executor.fire_remove_event(item, _sa_initiator, key=key)
+
+
+def __before_pop(collection, _sa_initiator=None):
+    """An event which occurs on a before a pop() operation occurs."""
+    executor = collection._sa_adapter
+    if executor:
+        executor.fire_pre_remove_event(_sa_initiator)
+
+
+def _list_decorators() -> Dict[str, Callable[[_FN], _FN]]:
+    """Tailored instrumentation wrappers for any list-like class."""
+
+    def _tidy(fn):
+        fn._sa_instrumented = True
+        fn.__doc__ = getattr(list, fn.__name__).__doc__
+
+    def append(fn):
+        def append(self, item, _sa_initiator=None):
+            item = __set(self, item, _sa_initiator, NO_KEY)
+            fn(self, item)
+
+        _tidy(append)
+        return append
+
+    def remove(fn):
+        def remove(self, value, _sa_initiator=None):
+            __del(self, value, _sa_initiator, NO_KEY)
+            # testlib.pragma exempt:__eq__
+            fn(self, value)
+
+        _tidy(remove)
+        return remove
+
+    def insert(fn):
+        def insert(self, index, value):
+            value = __set(self, value, None, index)
+            fn(self, index, value)
+
+        _tidy(insert)
+        return insert
+
+    def __setitem__(fn):
+        def __setitem__(self, index, value):
+            if not isinstance(index, slice):
+                existing = self[index]
+                if existing is not None:
+                    __del(self, existing, None, index)
+                value = __set(self, value, None, index)
+                fn(self, index, value)
+            else:
+                # slice assignment requires __delitem__, insert, __len__
+                step = index.step or 1
+                start = index.start or 0
+                if start < 0:
+                    start += len(self)
+                if index.stop is not None:
+                    stop = index.stop
+                else:
+                    stop = len(self)
+                if stop < 0:
+                    stop += len(self)
+
+                if step == 1:
+                    if value is self:
+                        return
+                    for i in range(start, stop, step):
+                        if len(self) > start:
+                            del self[start]
+
+                    for i, item in enumerate(value):
+                        self.insert(i + start, item)
+                else:
+                    rng = list(range(start, stop, step))
+                    if len(value) != len(rng):
+                        raise ValueError(
+                            "attempt to assign sequence of size %s to "
+                            "extended slice of size %s"
+                            % (len(value), len(rng))
+                        )
+                    for i, item in zip(rng, value):
+                        self.__setitem__(i, item)
+
+        _tidy(__setitem__)
+        return __setitem__
+
+    def __delitem__(fn):
+        def __delitem__(self, index):
+            if not isinstance(index, slice):
+                item = self[index]
+                __del(self, item, None, index)
+                fn(self, index)
+            else:
+                # slice deletion requires __getslice__ and a slice-groking
+                # __getitem__ for stepped deletion
+                # note: not breaking this into atomic dels
+                for item in self[index]:
+                    __del(self, item, None, index)
+                fn(self, index)
+
+        _tidy(__delitem__)
+        return __delitem__
+
+    def extend(fn):
+        def extend(self, iterable):
+            for value in list(iterable):
+                self.append(value)
+
+        _tidy(extend)
+        return extend
+
+    def __iadd__(fn):
+        def __iadd__(self, iterable):
+            # list.__iadd__ takes any iterable and seems to let TypeError
+            # raise as-is instead of returning NotImplemented
+            for value in list(iterable):
+                self.append(value)
+            return self
+
+        _tidy(__iadd__)
+        return __iadd__
+
+    def pop(fn):
+        def pop(self, index=-1):
+            __before_pop(self)
+            item = fn(self, index)
+            __del(self, item, None, index)
+            return item
+
+        _tidy(pop)
+        return pop
+
+    def clear(fn):
+        def clear(self, index=-1):
+            for item in self:
+                __del(self, item, None, index)
+            fn(self)
+
+        _tidy(clear)
+        return clear
+
+    # __imul__ : not wrapping this.  all members of the collection are already
+    # present, so no need to fire appends... wrapping it with an explicit
+    # decorator is still possible, so events on *= can be had if they're
+    # desired.  hard to imagine a use case for __imul__, though.
+
+    l = locals().copy()
+    l.pop("_tidy")
+    return l
+
+
+def _dict_decorators() -> Dict[str, Callable[[_FN], _FN]]:
+    """Tailored instrumentation wrappers for any dict-like mapping class."""
+
+    def _tidy(fn):
+        fn._sa_instrumented = True
+        fn.__doc__ = getattr(dict, fn.__name__).__doc__
+
+    def __setitem__(fn):
+        def __setitem__(self, key, value, _sa_initiator=None):
+            if key in self:
+                __del(self, self[key], _sa_initiator, key)
+            value = __set(self, value, _sa_initiator, key)
+            fn(self, key, value)
+
+        _tidy(__setitem__)
+        return __setitem__
+
+    def __delitem__(fn):
+        def __delitem__(self, key, _sa_initiator=None):
+            if key in self:
+                __del(self, self[key], _sa_initiator, key)
+            fn(self, key)
+
+        _tidy(__delitem__)
+        return __delitem__
+
+    def clear(fn):
+        def clear(self):
+            for key in self:
+                __del(self, self[key], None, key)
+            fn(self)
+
+        _tidy(clear)
+        return clear
+
+    def pop(fn):
+        def pop(self, key, default=NO_ARG):
+            __before_pop(self)
+            _to_del = key in self
+            if default is NO_ARG:
+                item = fn(self, key)
+            else:
+                item = fn(self, key, default)
+            if _to_del:
+                __del(self, item, None, key)
+            return item
+
+        _tidy(pop)
+        return pop
+
+    def popitem(fn):
+        def popitem(self):
+            __before_pop(self)
+            item = fn(self)
+            __del(self, item[1], None, 1)
+            return item
+
+        _tidy(popitem)
+        return popitem
+
+    def setdefault(fn):
+        def setdefault(self, key, default=None):
+            if key not in self:
+                self.__setitem__(key, default)
+                return default
+            else:
+                value = self.__getitem__(key)
+                if value is default:
+                    __set_wo_mutation(self, value, None)
+
+                return value
+
+        _tidy(setdefault)
+        return setdefault
+
+    def update(fn):
+        def update(self, __other=NO_ARG, **kw):
+            if __other is not NO_ARG:
+                if hasattr(__other, "keys"):
+                    for key in list(__other):
+                        if key not in self or self[key] is not __other[key]:
+                            self[key] = __other[key]
+                        else:
+                            __set_wo_mutation(self, __other[key], None)
+                else:
+                    for key, value in __other:
+                        if key not in self or self[key] is not value:
+                            self[key] = value
+                        else:
+                            __set_wo_mutation(self, value, None)
+            for key in kw:
+                if key not in self or self[key] is not kw[key]:
+                    self[key] = kw[key]
+                else:
+                    __set_wo_mutation(self, kw[key], None)
+
+        _tidy(update)
+        return update
+
+    l = locals().copy()
+    l.pop("_tidy")
+    return l
+
+
+_set_binop_bases = (set, frozenset)
+
+
+def _set_binops_check_strict(self: Any, obj: Any) -> bool:
+    """Allow only set, frozenset and self.__class__-derived
+    objects in binops."""
+    return isinstance(obj, _set_binop_bases + (self.__class__,))
+
+
+def _set_binops_check_loose(self: Any, obj: Any) -> bool:
+    """Allow anything set-like to participate in set binops."""
+    return (
+        isinstance(obj, _set_binop_bases + (self.__class__,))
+        or util.duck_type_collection(obj) == set
+    )
+
+
+def _set_decorators() -> Dict[str, Callable[[_FN], _FN]]:
+    """Tailored instrumentation wrappers for any set-like class."""
+
+    def _tidy(fn):
+        fn._sa_instrumented = True
+        fn.__doc__ = getattr(set, fn.__name__).__doc__
+
+    def add(fn):
+        def add(self, value, _sa_initiator=None):
+            if value not in self:
+                value = __set(self, value, _sa_initiator, NO_KEY)
+            else:
+                __set_wo_mutation(self, value, _sa_initiator)
+            # testlib.pragma exempt:__hash__
+            fn(self, value)
+
+        _tidy(add)
+        return add
+
+    def discard(fn):
+        def discard(self, value, _sa_initiator=None):
+            # testlib.pragma exempt:__hash__
+            if value in self:
+                __del(self, value, _sa_initiator, NO_KEY)
+                # testlib.pragma exempt:__hash__
+            fn(self, value)
+
+        _tidy(discard)
+        return discard
+
+    def remove(fn):
+        def remove(self, value, _sa_initiator=None):
+            # testlib.pragma exempt:__hash__
+            if value in self:
+                __del(self, value, _sa_initiator, NO_KEY)
+            # testlib.pragma exempt:__hash__
+            fn(self, value)
+
+        _tidy(remove)
+        return remove
+
+    def pop(fn):
+        def pop(self):
+            __before_pop(self)
+            item = fn(self)
+            # for set in particular, we have no way to access the item
+            # that will be popped before pop is called.
+            __del(self, item, None, NO_KEY)
+            return item
+
+        _tidy(pop)
+        return pop
+
+    def clear(fn):
+        def clear(self):
+            for item in list(self):
+                self.remove(item)
+
+        _tidy(clear)
+        return clear
+
+    def update(fn):
+        def update(self, value):
+            for item in value:
+                self.add(item)
+
+        _tidy(update)
+        return update
+
+    def __ior__(fn):
+        def __ior__(self, value):
+            if not _set_binops_check_strict(self, value):
+                return NotImplemented
+            for item in value:
+                self.add(item)
+            return self
+
+        _tidy(__ior__)
+        return __ior__
+
+    def difference_update(fn):
+        def difference_update(self, value):
+            for item in value:
+                self.discard(item)
+
+        _tidy(difference_update)
+        return difference_update
+
+    def __isub__(fn):
+        def __isub__(self, value):
+            if not _set_binops_check_strict(self, value):
+                return NotImplemented
+            for item in value:
+                self.discard(item)
+            return self
+
+        _tidy(__isub__)
+        return __isub__
+
+    def intersection_update(fn):
+        def intersection_update(self, other):
+            want, have = self.intersection(other), set(self)
+            remove, add = have - want, want - have
+
+            for item in remove:
+                self.remove(item)
+            for item in add:
+                self.add(item)
+
+        _tidy(intersection_update)
+        return intersection_update
+
+    def __iand__(fn):
+        def __iand__(self, other):
+            if not _set_binops_check_strict(self, other):
+                return NotImplemented
+            want, have = self.intersection(other), set(self)
+            remove, add = have - want, want - have
+
+            for item in remove:
+                self.remove(item)
+            for item in add:
+                self.add(item)
+            return self
+
+        _tidy(__iand__)
+        return __iand__
+
+    def symmetric_difference_update(fn):
+        def symmetric_difference_update(self, other):
+            want, have = self.symmetric_difference(other), set(self)
+            remove, add = have - want, want - have
+
+            for item in remove:
+                self.remove(item)
+            for item in add:
+                self.add(item)
+
+        _tidy(symmetric_difference_update)
+        return symmetric_difference_update
+
+    def __ixor__(fn):
+        def __ixor__(self, other):
+            if not _set_binops_check_strict(self, other):
+                return NotImplemented
+            want, have = self.symmetric_difference(other), set(self)
+            remove, add = have - want, want - have
+
+            for item in remove:
+                self.remove(item)
+            for item in add:
+                self.add(item)
+            return self
+
+        _tidy(__ixor__)
+        return __ixor__
+
+    l = locals().copy()
+    l.pop("_tidy")
+    return l
+
+
+class InstrumentedList(List[_T]):
+    """An instrumented version of the built-in list."""
+
+
+class InstrumentedSet(Set[_T]):
+    """An instrumented version of the built-in set."""
+
+
+class InstrumentedDict(Dict[_KT, _VT]):
+    """An instrumented version of the built-in dict."""
+
+
+__canned_instrumentation: util.immutabledict[Any, _CollectionFactoryType] = (
+    util.immutabledict(
+        {
+            list: InstrumentedList,
+            set: InstrumentedSet,
+            dict: InstrumentedDict,
+        }
+    )
+)
+
+__interfaces: util.immutabledict[
+    Any,
+    Tuple[
+        Dict[str, str],
+        Dict[str, Callable[..., Any]],
+    ],
+] = util.immutabledict(
+    {
+        list: (
+            {
+                "appender": "append",
+                "remover": "remove",
+                "iterator": "__iter__",
+            },
+            _list_decorators(),
+        ),
+        set: (
+            {"appender": "add", "remover": "remove", "iterator": "__iter__"},
+            _set_decorators(),
+        ),
+        # decorators are required for dicts and object collections.
+        dict: ({"iterator": "values"}, _dict_decorators()),
+    }
+)
+
+
+def __go(lcls):
+    global keyfunc_mapping, mapped_collection
+    global column_keyed_dict, column_mapped_collection
+    global MappedCollection, KeyFuncDict
+    global attribute_keyed_dict, attribute_mapped_collection
+
+    from .mapped_collection import keyfunc_mapping
+    from .mapped_collection import column_keyed_dict
+    from .mapped_collection import attribute_keyed_dict
+    from .mapped_collection import KeyFuncDict
+
+    from .mapped_collection import mapped_collection
+    from .mapped_collection import column_mapped_collection
+    from .mapped_collection import attribute_mapped_collection
+    from .mapped_collection import MappedCollection
+
+    # ensure instrumentation is associated with
+    # these built-in classes; if a user-defined class
+    # subclasses these and uses @internally_instrumented,
+    # the superclass is otherwise not instrumented.
+    # see [ticket:2406].
+    _instrument_class(InstrumentedList)
+    _instrument_class(InstrumentedSet)
+    _instrument_class(KeyFuncDict)
+
+
+__go(locals())
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/context.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/context.py
new file mode 100644
index 00000000..c09c03b7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/context.py
@@ -0,0 +1,3268 @@
+# orm/context.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import itertools
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import interfaces
+from . import loading
+from .base import _is_aliased_class
+from .interfaces import ORMColumnDescription
+from .interfaces import ORMColumnsClauseRole
+from .path_registry import PathRegistry
+from .util import _entity_corresponds_to
+from .util import _ORMJoin
+from .util import _TraceAdaptRole
+from .util import AliasedClass
+from .util import Bundle
+from .util import ORMAdapter
+from .util import ORMStatementAdapter
+from .. import exc as sa_exc
+from .. import future
+from .. import inspect
+from .. import sql
+from .. import util
+from ..sql import coercions
+from ..sql import expression
+from ..sql import roles
+from ..sql import util as sql_util
+from ..sql import visitors
+from ..sql._typing import _TP
+from ..sql._typing import is_dml
+from ..sql._typing import is_insert_update
+from ..sql._typing import is_select_base
+from ..sql.base import _select_iterables
+from ..sql.base import CacheableOptions
+from ..sql.base import CompileState
+from ..sql.base import Executable
+from ..sql.base import Generative
+from ..sql.base import Options
+from ..sql.dml import UpdateBase
+from ..sql.elements import GroupedElement
+from ..sql.elements import TextClause
+from ..sql.selectable import CompoundSelectState
+from ..sql.selectable import LABEL_STYLE_DISAMBIGUATE_ONLY
+from ..sql.selectable import LABEL_STYLE_NONE
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..sql.selectable import Select
+from ..sql.selectable import SelectLabelStyle
+from ..sql.selectable import SelectState
+from ..sql.selectable import TypedReturnsRows
+from ..sql.visitors import InternalTraversal
+
+if TYPE_CHECKING:
+    from ._typing import _InternalEntityType
+    from ._typing import OrmExecuteOptionsParameter
+    from .loading import PostLoad
+    from .mapper import Mapper
+    from .query import Query
+    from .session import _BindArguments
+    from .session import Session
+    from ..engine import Result
+    from ..engine.interfaces import _CoreSingleExecuteParams
+    from ..sql._typing import _ColumnsClauseArgument
+    from ..sql.compiler import SQLCompiler
+    from ..sql.dml import _DMLTableElement
+    from ..sql.elements import ColumnElement
+    from ..sql.selectable import _JoinTargetElement
+    from ..sql.selectable import _LabelConventionCallable
+    from ..sql.selectable import _SetupJoinsElement
+    from ..sql.selectable import ExecutableReturnsRows
+    from ..sql.selectable import SelectBase
+    from ..sql.type_api import TypeEngine
+
+_T = TypeVar("_T", bound=Any)
+_path_registry = PathRegistry.root
+
+_EMPTY_DICT = util.immutabledict()
+
+
+LABEL_STYLE_LEGACY_ORM = SelectLabelStyle.LABEL_STYLE_LEGACY_ORM
+
+
+class QueryContext:
+    __slots__ = (
+        "top_level_context",
+        "compile_state",
+        "query",
+        "user_passed_query",
+        "params",
+        "load_options",
+        "bind_arguments",
+        "execution_options",
+        "session",
+        "autoflush",
+        "populate_existing",
+        "invoke_all_eagers",
+        "version_check",
+        "refresh_state",
+        "create_eager_joins",
+        "propagated_loader_options",
+        "attributes",
+        "runid",
+        "partials",
+        "post_load_paths",
+        "identity_token",
+        "yield_per",
+        "loaders_require_buffering",
+        "loaders_require_uniquing",
+    )
+
+    runid: int
+    post_load_paths: Dict[PathRegistry, PostLoad]
+    compile_state: ORMCompileState
+
+    class default_load_options(Options):
+        _only_return_tuples = False
+        _populate_existing = False
+        _version_check = False
+        _invoke_all_eagers = True
+        _autoflush = True
+        _identity_token = None
+        _yield_per = None
+        _refresh_state = None
+        _lazy_loaded_from = None
+        _legacy_uniquing = False
+        _sa_top_level_orm_context = None
+        _is_user_refresh = False
+
+    def __init__(
+        self,
+        compile_state: CompileState,
+        statement: Union[Select[Any], FromStatement[Any]],
+        user_passed_query: Union[
+            Select[Any],
+            FromStatement[Any],
+        ],
+        params: _CoreSingleExecuteParams,
+        session: Session,
+        load_options: Union[
+            Type[QueryContext.default_load_options],
+            QueryContext.default_load_options,
+        ],
+        execution_options: Optional[OrmExecuteOptionsParameter] = None,
+        bind_arguments: Optional[_BindArguments] = None,
+    ):
+        self.load_options = load_options
+        self.execution_options = execution_options or _EMPTY_DICT
+        self.bind_arguments = bind_arguments or _EMPTY_DICT
+        self.compile_state = compile_state
+        self.query = statement
+
+        # the query that the end user passed to Session.execute() or similar.
+        # this is usually the same as .query, except in the bulk_persistence
+        # routines where a separate FromStatement is manufactured in the
+        # compile stage; this allows differentiation in that case.
+        self.user_passed_query = user_passed_query
+
+        self.session = session
+        self.loaders_require_buffering = False
+        self.loaders_require_uniquing = False
+        self.params = params
+        self.top_level_context = load_options._sa_top_level_orm_context
+
+        cached_options = compile_state.select_statement._with_options
+        uncached_options = user_passed_query._with_options
+
+        # see issue #7447 , #8399 for some background
+        # propagated loader options will be present on loaded InstanceState
+        # objects under state.load_options and are typically used by
+        # LazyLoader to apply options to the SELECT statement it emits.
+        # For compile state options (i.e. loader strategy options), these
+        # need to line up with the ".load_path" attribute which in
+        # loader.py is pulled from context.compile_state.current_path.
+        # so, this means these options have to be the ones from the
+        # *cached* statement that's travelling with compile_state, not the
+        # *current* statement which won't match up for an ad-hoc
+        # AliasedClass
+        self.propagated_loader_options = tuple(
+            opt._adapt_cached_option_to_uncached_option(self, uncached_opt)
+            for opt, uncached_opt in zip(cached_options, uncached_options)
+            if opt.propagate_to_loaders
+        )
+
+        self.attributes = dict(compile_state.attributes)
+
+        self.autoflush = load_options._autoflush
+        self.populate_existing = load_options._populate_existing
+        self.invoke_all_eagers = load_options._invoke_all_eagers
+        self.version_check = load_options._version_check
+        self.refresh_state = load_options._refresh_state
+        self.yield_per = load_options._yield_per
+        self.identity_token = load_options._identity_token
+
+    def _get_top_level_context(self) -> QueryContext:
+        return self.top_level_context or self
+
+
+_orm_load_exec_options = util.immutabledict(
+    {"_result_disable_adapt_to_context": True}
+)
+
+
+class AbstractORMCompileState(CompileState):
+    is_dml_returning = False
+
+    def _init_global_attributes(
+        self, statement, compiler, *, toplevel, process_criteria_for_toplevel
+    ):
+        self.attributes = {}
+
+        if compiler is None:
+            # this is the legacy / testing only ORM _compile_state() use case.
+            # there is no need to apply criteria options for this.
+            self.global_attributes = ga = {}
+            assert toplevel
+            return
+        else:
+            self.global_attributes = ga = compiler._global_attributes
+
+        if toplevel:
+            ga["toplevel_orm"] = True
+
+            if process_criteria_for_toplevel:
+                for opt in statement._with_options:
+                    if opt._is_criteria_option:
+                        opt.process_compile_state(self)
+
+            return
+        elif ga.get("toplevel_orm", False):
+            return
+
+        stack_0 = compiler.stack[0]
+
+        try:
+            toplevel_stmt = stack_0["selectable"]
+        except KeyError:
+            pass
+        else:
+            for opt in toplevel_stmt._with_options:
+                if opt._is_compile_state and opt._is_criteria_option:
+                    opt.process_compile_state(self)
+
+        ga["toplevel_orm"] = True
+
+    @classmethod
+    def create_for_statement(
+        cls,
+        statement: Union[Select, FromStatement],
+        compiler: Optional[SQLCompiler],
+        **kw: Any,
+    ) -> AbstractORMCompileState:
+        """Create a context for a statement given a :class:`.Compiler`.
+
+        This method is always invoked in the context of SQLCompiler.process().
+
+        For a Select object, this would be invoked from
+        SQLCompiler.visit_select(). For the special FromStatement object used
+        by Query to indicate "Query.from_statement()", this is called by
+        FromStatement._compiler_dispatch() that would be called by
+        SQLCompiler.process().
+        """
+        return super().create_for_statement(statement, compiler, **kw)
+
+    @classmethod
+    def orm_pre_session_exec(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        is_pre_event,
+    ):
+        raise NotImplementedError()
+
+    @classmethod
+    def orm_execute_statement(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        conn,
+    ) -> Result:
+        result = conn.execute(
+            statement, params or {}, execution_options=execution_options
+        )
+        return cls.orm_setup_cursor_result(
+            session,
+            statement,
+            params,
+            execution_options,
+            bind_arguments,
+            result,
+        )
+
+    @classmethod
+    def orm_setup_cursor_result(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        result,
+    ):
+        raise NotImplementedError()
+
+
+class AutoflushOnlyORMCompileState(AbstractORMCompileState):
+    """ORM compile state that is a passthrough, except for autoflush."""
+
+    @classmethod
+    def orm_pre_session_exec(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        is_pre_event,
+    ):
+        # consume result-level load_options.  These may have been set up
+        # in an ORMExecuteState hook
+        (
+            load_options,
+            execution_options,
+        ) = QueryContext.default_load_options.from_execution_options(
+            "_sa_orm_load_options",
+            {
+                "autoflush",
+            },
+            execution_options,
+            statement._execution_options,
+        )
+
+        if not is_pre_event and load_options._autoflush:
+            session._autoflush()
+
+        return statement, execution_options
+
+    @classmethod
+    def orm_setup_cursor_result(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        result,
+    ):
+        return result
+
+
+class ORMCompileState(AbstractORMCompileState):
+    class default_compile_options(CacheableOptions):
+        _cache_key_traversal = [
+            ("_use_legacy_query_style", InternalTraversal.dp_boolean),
+            ("_for_statement", InternalTraversal.dp_boolean),
+            ("_bake_ok", InternalTraversal.dp_boolean),
+            ("_current_path", InternalTraversal.dp_has_cache_key),
+            ("_enable_single_crit", InternalTraversal.dp_boolean),
+            ("_enable_eagerloads", InternalTraversal.dp_boolean),
+            ("_only_load_props", InternalTraversal.dp_plain_obj),
+            ("_set_base_alias", InternalTraversal.dp_boolean),
+            ("_for_refresh_state", InternalTraversal.dp_boolean),
+            ("_render_for_subquery", InternalTraversal.dp_boolean),
+            ("_is_star", InternalTraversal.dp_boolean),
+        ]
+
+        # set to True by default from Query._statement_20(), to indicate
+        # the rendered query should look like a legacy ORM query.  right
+        # now this basically indicates we should use tablename_columnname
+        # style labels.    Generally indicates the statement originated
+        # from a Query object.
+        _use_legacy_query_style = False
+
+        # set *only* when we are coming from the Query.statement
+        # accessor, or a Query-level equivalent such as
+        # query.subquery().  this supersedes "toplevel".
+        _for_statement = False
+
+        _bake_ok = True
+        _current_path = _path_registry
+        _enable_single_crit = True
+        _enable_eagerloads = True
+        _only_load_props = None
+        _set_base_alias = False
+        _for_refresh_state = False
+        _render_for_subquery = False
+        _is_star = False
+
+    attributes: Dict[Any, Any]
+    global_attributes: Dict[Any, Any]
+
+    statement: Union[Select[Any], FromStatement[Any]]
+    select_statement: Union[Select[Any], FromStatement[Any]]
+    _entities: List[_QueryEntity]
+    _polymorphic_adapters: Dict[_InternalEntityType, ORMAdapter]
+    compile_options: Union[
+        Type[default_compile_options], default_compile_options
+    ]
+    _primary_entity: Optional[_QueryEntity]
+    use_legacy_query_style: bool
+    _label_convention: _LabelConventionCallable
+    primary_columns: List[ColumnElement[Any]]
+    secondary_columns: List[ColumnElement[Any]]
+    dedupe_columns: Set[ColumnElement[Any]]
+    create_eager_joins: List[
+        # TODO: this structure is set up by JoinedLoader
+        Tuple[Any, ...]
+    ]
+    current_path: PathRegistry = _path_registry
+    _has_mapper_entities = False
+
+    def __init__(self, *arg, **kw):
+        raise NotImplementedError()
+
+    if TYPE_CHECKING:
+
+        @classmethod
+        def create_for_statement(
+            cls,
+            statement: Union[Select, FromStatement],
+            compiler: Optional[SQLCompiler],
+            **kw: Any,
+        ) -> ORMCompileState: ...
+
+    def _append_dedupe_col_collection(self, obj, col_collection):
+        dedupe = self.dedupe_columns
+        if obj not in dedupe:
+            dedupe.add(obj)
+            col_collection.append(obj)
+
+    @classmethod
+    def _column_naming_convention(
+        cls, label_style: SelectLabelStyle, legacy: bool
+    ) -> _LabelConventionCallable:
+        if legacy:
+
+            def name(col, col_name=None):
+                if col_name:
+                    return col_name
+                else:
+                    return getattr(col, "key")
+
+            return name
+        else:
+            return SelectState._column_naming_convention(label_style)
+
+    @classmethod
+    def get_column_descriptions(cls, statement):
+        return _column_descriptions(statement)
+
+    @classmethod
+    def orm_pre_session_exec(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        is_pre_event,
+    ):
+        # consume result-level load_options.  These may have been set up
+        # in an ORMExecuteState hook
+        (
+            load_options,
+            execution_options,
+        ) = QueryContext.default_load_options.from_execution_options(
+            "_sa_orm_load_options",
+            {
+                "populate_existing",
+                "autoflush",
+                "yield_per",
+                "identity_token",
+                "sa_top_level_orm_context",
+            },
+            execution_options,
+            statement._execution_options,
+        )
+
+        # default execution options for ORM results:
+        # 1. _result_disable_adapt_to_context=True
+        #    this will disable the ResultSetMetadata._adapt_to_context()
+        #    step which we don't need, as we have result processors cached
+        #    against the original SELECT statement before caching.
+
+        if "sa_top_level_orm_context" in execution_options:
+            ctx = execution_options["sa_top_level_orm_context"]
+            execution_options = ctx.query._execution_options.merge_with(
+                ctx.execution_options, execution_options
+            )
+
+        if not execution_options:
+            execution_options = _orm_load_exec_options
+        else:
+            execution_options = execution_options.union(_orm_load_exec_options)
+
+        # would have been placed here by legacy Query only
+        if load_options._yield_per:
+            execution_options = execution_options.union(
+                {"yield_per": load_options._yield_per}
+            )
+
+        if (
+            getattr(statement._compile_options, "_current_path", None)
+            and len(statement._compile_options._current_path) > 10
+            and execution_options.get("compiled_cache", True) is not None
+        ):
+            execution_options: util.immutabledict[str, Any] = (
+                execution_options.union(
+                    {
+                        "compiled_cache": None,
+                        "_cache_disable_reason": "excess depth for "
+                        "ORM loader options",
+                    }
+                )
+            )
+
+        bind_arguments["clause"] = statement
+
+        # new in 1.4 - the coercions system is leveraged to allow the
+        # "subject" mapper of a statement be propagated to the top
+        # as the statement is built.   "subject" mapper is the generally
+        # standard object used as an identifier for multi-database schemes.
+
+        # we are here based on the fact that _propagate_attrs contains
+        # "compile_state_plugin": "orm".   The "plugin_subject"
+        # needs to be present as well.
+
+        try:
+            plugin_subject = statement._propagate_attrs["plugin_subject"]
+        except KeyError:
+            assert False, "statement had 'orm' plugin but no plugin_subject"
+        else:
+            if plugin_subject:
+                bind_arguments["mapper"] = plugin_subject.mapper
+
+        if not is_pre_event and load_options._autoflush:
+            session._autoflush()
+
+        return statement, execution_options
+
+    @classmethod
+    def orm_setup_cursor_result(
+        cls,
+        session,
+        statement,
+        params,
+        execution_options,
+        bind_arguments,
+        result,
+    ):
+        execution_context = result.context
+        compile_state = execution_context.compiled.compile_state
+
+        # cover edge case where ORM entities used in legacy select
+        # were passed to session.execute:
+        # session.execute(legacy_select([User.id, User.name]))
+        # see test_query->test_legacy_tuple_old_select
+
+        load_options = execution_options.get(
+            "_sa_orm_load_options", QueryContext.default_load_options
+        )
+
+        if compile_state.compile_options._is_star:
+            return result
+
+        querycontext = QueryContext(
+            compile_state,
+            statement,
+            statement,
+            params,
+            session,
+            load_options,
+            execution_options,
+            bind_arguments,
+        )
+        return loading.instances(result, querycontext)
+
+    @property
+    def _lead_mapper_entities(self):
+        """return all _MapperEntity objects in the lead entities collection.
+
+        Does **not** include entities that have been replaced by
+        with_entities(), with_only_columns()
+
+        """
+        return [
+            ent for ent in self._entities if isinstance(ent, _MapperEntity)
+        ]
+
+    def _create_with_polymorphic_adapter(self, ext_info, selectable):
+        """given MapperEntity or ORMColumnEntity, setup polymorphic loading
+        if called for by the Mapper.
+
+        As of #8168 in 2.0.0rc1, polymorphic adapters, which greatly increase
+        the complexity of the query creation process, are not used at all
+        except in the quasi-legacy cases of with_polymorphic referring to an
+        alias and/or subquery. This would apply to concrete polymorphic
+        loading, and joined inheritance where a subquery is
+        passed to with_polymorphic (which is completely unnecessary in modern
+        use).
+
+        """
+        if (
+            not ext_info.is_aliased_class
+            and ext_info.mapper.persist_selectable
+            not in self._polymorphic_adapters
+        ):
+            for mp in ext_info.mapper.iterate_to_root():
+                self._mapper_loads_polymorphically_with(
+                    mp,
+                    ORMAdapter(
+                        _TraceAdaptRole.WITH_POLYMORPHIC_ADAPTER,
+                        mp,
+                        equivalents=mp._equivalent_columns,
+                        selectable=selectable,
+                    ),
+                )
+
+    def _mapper_loads_polymorphically_with(self, mapper, adapter):
+        for m2 in mapper._with_polymorphic_mappers or [mapper]:
+            self._polymorphic_adapters[m2] = adapter
+
+            for m in m2.iterate_to_root():
+                self._polymorphic_adapters[m.local_table] = adapter
+
+    @classmethod
+    def _create_entities_collection(cls, query, legacy):
+        raise NotImplementedError(
+            "this method only works for ORMSelectCompileState"
+        )
+
+
+class DMLReturningColFilter:
+    """an adapter used for the DML RETURNING case.
+
+    Has a subset of the interface used by
+    :class:`.ORMAdapter` and is used for :class:`._QueryEntity`
+    instances to set up their columns as used in RETURNING for a
+    DML statement.
+
+    """
+
+    __slots__ = ("mapper", "columns", "__weakref__")
+
+    def __init__(self, target_mapper, immediate_dml_mapper):
+        if (
+            immediate_dml_mapper is not None
+            and target_mapper.local_table
+            is not immediate_dml_mapper.local_table
+        ):
+            # joined inh, or in theory other kinds of multi-table mappings
+            self.mapper = immediate_dml_mapper
+        else:
+            # single inh, normal mappings, etc.
+            self.mapper = target_mapper
+        self.columns = self.columns = util.WeakPopulateDict(
+            self.adapt_check_present  # type: ignore
+        )
+
+    def __call__(self, col, as_filter):
+        for cc in sql_util._find_columns(col):
+            c2 = self.adapt_check_present(cc)
+            if c2 is not None:
+                return col
+        else:
+            return None
+
+    def adapt_check_present(self, col):
+        mapper = self.mapper
+        prop = mapper._columntoproperty.get(col, None)
+        if prop is None:
+            return None
+        return mapper.local_table.c.corresponding_column(col)
+
+
+@sql.base.CompileState.plugin_for("orm", "orm_from_statement")
+class ORMFromStatementCompileState(ORMCompileState):
+    _from_obj_alias = None
+    _has_mapper_entities = False
+
+    statement_container: FromStatement
+    requested_statement: Union[SelectBase, TextClause, UpdateBase]
+    dml_table: Optional[_DMLTableElement] = None
+
+    _has_orm_entities = False
+    multi_row_eager_loaders = False
+    eager_adding_joins = False
+    compound_eager_adapter = None
+
+    extra_criteria_entities = _EMPTY_DICT
+    eager_joins = _EMPTY_DICT
+
+    @classmethod
+    def create_for_statement(
+        cls,
+        statement_container: Union[Select, FromStatement],
+        compiler: Optional[SQLCompiler],
+        **kw: Any,
+    ) -> ORMFromStatementCompileState:
+        assert isinstance(statement_container, FromStatement)
+
+        if compiler is not None and compiler.stack:
+            raise sa_exc.CompileError(
+                "The ORM FromStatement construct only supports being "
+                "invoked as the topmost statement, as it is only intended to "
+                "define how result rows should be returned."
+            )
+
+        self = cls.__new__(cls)
+        self._primary_entity = None
+
+        self.use_legacy_query_style = (
+            statement_container._compile_options._use_legacy_query_style
+        )
+        self.statement_container = self.select_statement = statement_container
+        self.requested_statement = statement = statement_container.element
+
+        if statement.is_dml:
+            self.dml_table = statement.table
+            self.is_dml_returning = True
+
+        self._entities = []
+        self._polymorphic_adapters = {}
+
+        self.compile_options = statement_container._compile_options
+
+        if (
+            self.use_legacy_query_style
+            and isinstance(statement, expression.SelectBase)
+            and not statement._is_textual
+            and not statement.is_dml
+            and statement._label_style is LABEL_STYLE_NONE
+        ):
+            self.statement = statement.set_label_style(
+                LABEL_STYLE_TABLENAME_PLUS_COL
+            )
+        else:
+            self.statement = statement
+
+        self._label_convention = self._column_naming_convention(
+            (
+                statement._label_style
+                if not statement._is_textual and not statement.is_dml
+                else LABEL_STYLE_NONE
+            ),
+            self.use_legacy_query_style,
+        )
+
+        _QueryEntity.to_compile_state(
+            self,
+            statement_container._raw_columns,
+            self._entities,
+            is_current_entities=True,
+        )
+
+        self.current_path = statement_container._compile_options._current_path
+
+        self._init_global_attributes(
+            statement_container,
+            compiler,
+            process_criteria_for_toplevel=False,
+            toplevel=True,
+        )
+
+        if statement_container._with_options:
+            for opt in statement_container._with_options:
+                if opt._is_compile_state:
+                    opt.process_compile_state(self)
+
+        if statement_container._with_context_options:
+            for fn, key in statement_container._with_context_options:
+                fn(self)
+
+        self.primary_columns = []
+        self.secondary_columns = []
+        self.dedupe_columns = set()
+        self.create_eager_joins = []
+        self._fallback_from_clauses = []
+
+        self.order_by = None
+
+        if isinstance(self.statement, expression.TextClause):
+            # TextClause has no "column" objects at all.  for this case,
+            # we generate columns from our _QueryEntity objects, then
+            # flip on all the "please match no matter what" parameters.
+            self.extra_criteria_entities = {}
+
+            for entity in self._entities:
+                entity.setup_compile_state(self)
+
+            compiler._ordered_columns = compiler._textual_ordered_columns = (
+                False
+            )
+
+            # enable looser result column matching.  this is shown to be
+            # needed by test_query.py::TextTest
+            compiler._loose_column_name_matching = True
+
+            for c in self.primary_columns:
+                compiler.process(
+                    c,
+                    within_columns_clause=True,
+                    add_to_result_map=compiler._add_to_result_map,
+                )
+        else:
+            # for everyone else, Select, Insert, Update, TextualSelect, they
+            # have column objects already.  After much
+            # experimentation here, the best approach seems to be, use
+            # those columns completely, don't interfere with the compiler
+            # at all; just in ORM land, use an adapter to convert from
+            # our ORM columns to whatever columns are in the statement,
+            # before we look in the result row. Adapt on names
+            # to accept cases such as issue #9217, however also allow
+            # this to be overridden for cases such as #9273.
+            self._from_obj_alias = ORMStatementAdapter(
+                _TraceAdaptRole.ADAPT_FROM_STATEMENT,
+                self.statement,
+                adapt_on_names=statement_container._adapt_on_names,
+            )
+
+        return self
+
+    def _adapt_col_list(self, cols, current_adapter):
+        return cols
+
+    def _get_current_adapter(self):
+        return None
+
+    def setup_dml_returning_compile_state(self, dml_mapper):
+        """used by BulkORMInsert (and Update / Delete?) to set up a handler
+        for RETURNING to return ORM objects and expressions
+
+        """
+        target_mapper = self.statement._propagate_attrs.get(
+            "plugin_subject", None
+        )
+        adapter = DMLReturningColFilter(target_mapper, dml_mapper)
+
+        if self.compile_options._is_star and (len(self._entities) != 1):
+            raise sa_exc.CompileError(
+                "Can't generate ORM query that includes multiple expressions "
+                "at the same time as '*'; query for '*' alone if present"
+            )
+
+        for entity in self._entities:
+            entity.setup_dml_returning_compile_state(self, adapter)
+
+
+class FromStatement(GroupedElement, Generative, TypedReturnsRows[_TP]):
+    """Core construct that represents a load of ORM objects from various
+    :class:`.ReturnsRows` and other classes including:
+
+    :class:`.Select`, :class:`.TextClause`, :class:`.TextualSelect`,
+    :class:`.CompoundSelect`, :class`.Insert`, :class:`.Update`,
+    and in theory, :class:`.Delete`.
+
+    """
+
+    __visit_name__ = "orm_from_statement"
+
+    _compile_options = ORMFromStatementCompileState.default_compile_options
+
+    _compile_state_factory = ORMFromStatementCompileState.create_for_statement
+
+    _for_update_arg = None
+
+    element: Union[ExecutableReturnsRows, TextClause]
+
+    _adapt_on_names: bool
+
+    _traverse_internals = [
+        ("_raw_columns", InternalTraversal.dp_clauseelement_list),
+        ("element", InternalTraversal.dp_clauseelement),
+    ] + Executable._executable_traverse_internals
+
+    _cache_key_traversal = _traverse_internals + [
+        ("_compile_options", InternalTraversal.dp_has_cache_key)
+    ]
+
+    is_from_statement = True
+
+    def __init__(
+        self,
+        entities: Iterable[_ColumnsClauseArgument[Any]],
+        element: Union[ExecutableReturnsRows, TextClause],
+        _adapt_on_names: bool = True,
+    ):
+        self._raw_columns = [
+            coercions.expect(
+                roles.ColumnsClauseRole,
+                ent,
+                apply_propagate_attrs=self,
+                post_inspect=True,
+            )
+            for ent in util.to_list(entities)
+        ]
+        self.element = element
+        self.is_dml = element.is_dml
+        self.is_select = element.is_select
+        self.is_delete = element.is_delete
+        self.is_insert = element.is_insert
+        self.is_update = element.is_update
+        self._label_style = (
+            element._label_style if is_select_base(element) else None
+        )
+        self._adapt_on_names = _adapt_on_names
+
+    def _compiler_dispatch(self, compiler, **kw):
+        """provide a fixed _compiler_dispatch method.
+
+        This is roughly similar to using the sqlalchemy.ext.compiler
+        ``@compiles`` extension.
+
+        """
+
+        compile_state = self._compile_state_factory(self, compiler, **kw)
+
+        toplevel = not compiler.stack
+
+        if toplevel:
+            compiler.compile_state = compile_state
+
+        return compiler.process(compile_state.statement, **kw)
+
+    @property
+    def column_descriptions(self):
+        """Return a :term:`plugin-enabled` 'column descriptions' structure
+        referring to the columns which are SELECTed by this statement.
+
+        See the section :ref:`queryguide_inspection` for an overview
+        of this feature.
+
+        .. seealso::
+
+            :ref:`queryguide_inspection` - ORM background
+
+        """
+        meth = cast(
+            ORMSelectCompileState, SelectState.get_plugin_class(self)
+        ).get_column_descriptions
+        return meth(self)
+
+    def _ensure_disambiguated_names(self):
+        return self
+
+    def get_children(self, **kw):
+        yield from itertools.chain.from_iterable(
+            element._from_objects for element in self._raw_columns
+        )
+        yield from super().get_children(**kw)
+
+    @property
+    def _all_selected_columns(self):
+        return self.element._all_selected_columns
+
+    @property
+    def _return_defaults(self):
+        return self.element._return_defaults if is_dml(self.element) else None
+
+    @property
+    def _returning(self):
+        return self.element._returning if is_dml(self.element) else None
+
+    @property
+    def _inline(self):
+        return self.element._inline if is_insert_update(self.element) else None
+
+
+@sql.base.CompileState.plugin_for("orm", "compound_select")
+class CompoundSelectCompileState(
+    AutoflushOnlyORMCompileState, CompoundSelectState
+):
+    pass
+
+
+@sql.base.CompileState.plugin_for("orm", "select")
+class ORMSelectCompileState(ORMCompileState, SelectState):
+    _already_joined_edges = ()
+
+    _memoized_entities = _EMPTY_DICT
+
+    _from_obj_alias = None
+    _has_mapper_entities = False
+
+    _has_orm_entities = False
+    multi_row_eager_loaders = False
+    eager_adding_joins = False
+    compound_eager_adapter = None
+
+    correlate = None
+    correlate_except = None
+    _where_criteria = ()
+    _having_criteria = ()
+
+    @classmethod
+    def create_for_statement(
+        cls,
+        statement: Union[Select, FromStatement],
+        compiler: Optional[SQLCompiler],
+        **kw: Any,
+    ) -> ORMSelectCompileState:
+        """compiler hook, we arrive here from compiler.visit_select() only."""
+
+        self = cls.__new__(cls)
+
+        if compiler is not None:
+            toplevel = not compiler.stack
+        else:
+            toplevel = True
+
+        select_statement = statement
+
+        # if we are a select() that was never a legacy Query, we won't
+        # have ORM level compile options.
+        statement._compile_options = cls.default_compile_options.safe_merge(
+            statement._compile_options
+        )
+
+        if select_statement._execution_options:
+            # execution options should not impact the compilation of a
+            # query, and at the moment subqueryloader is putting some things
+            # in here that we explicitly don't want stuck in a cache.
+            self.select_statement = select_statement._clone()
+            self.select_statement._execution_options = util.immutabledict()
+        else:
+            self.select_statement = select_statement
+
+        # indicates this select() came from Query.statement
+        self.for_statement = select_statement._compile_options._for_statement
+
+        # generally if we are from Query or directly from a select()
+        self.use_legacy_query_style = (
+            select_statement._compile_options._use_legacy_query_style
+        )
+
+        self._entities = []
+        self._primary_entity = None
+        self._polymorphic_adapters = {}
+
+        self.compile_options = select_statement._compile_options
+
+        if not toplevel:
+            # for subqueries, turn off eagerloads and set
+            # "render_for_subquery".
+            self.compile_options += {
+                "_enable_eagerloads": False,
+                "_render_for_subquery": True,
+            }
+
+        # determine label style.   we can make different decisions here.
+        # at the moment, trying to see if we can always use DISAMBIGUATE_ONLY
+        # rather than LABEL_STYLE_NONE, and if we can use disambiguate style
+        # for new style ORM selects too.
+        if (
+            self.use_legacy_query_style
+            and self.select_statement._label_style is LABEL_STYLE_LEGACY_ORM
+        ):
+            if not self.for_statement:
+                self.label_style = LABEL_STYLE_TABLENAME_PLUS_COL
+            else:
+                self.label_style = LABEL_STYLE_DISAMBIGUATE_ONLY
+        else:
+            self.label_style = self.select_statement._label_style
+
+        if select_statement._memoized_select_entities:
+            self._memoized_entities = {
+                memoized_entities: _QueryEntity.to_compile_state(
+                    self,
+                    memoized_entities._raw_columns,
+                    [],
+                    is_current_entities=False,
+                )
+                for memoized_entities in (
+                    select_statement._memoized_select_entities
+                )
+            }
+
+        # label_convention is stateful and will yield deduping keys if it
+        # sees the same key twice.  therefore it's important that it is not
+        # invoked for the above "memoized" entities that aren't actually
+        # in the columns clause
+        self._label_convention = self._column_naming_convention(
+            statement._label_style, self.use_legacy_query_style
+        )
+
+        _QueryEntity.to_compile_state(
+            self,
+            select_statement._raw_columns,
+            self._entities,
+            is_current_entities=True,
+        )
+
+        self.current_path = select_statement._compile_options._current_path
+
+        self.eager_order_by = ()
+
+        self._init_global_attributes(
+            select_statement,
+            compiler,
+            toplevel=toplevel,
+            process_criteria_for_toplevel=False,
+        )
+
+        if toplevel and (
+            select_statement._with_options
+            or select_statement._memoized_select_entities
+        ):
+            for (
+                memoized_entities
+            ) in select_statement._memoized_select_entities:
+                for opt in memoized_entities._with_options:
+                    if opt._is_compile_state:
+                        opt.process_compile_state_replaced_entities(
+                            self,
+                            [
+                                ent
+                                for ent in self._memoized_entities[
+                                    memoized_entities
+                                ]
+                                if isinstance(ent, _MapperEntity)
+                            ],
+                        )
+
+            for opt in self.select_statement._with_options:
+                if opt._is_compile_state:
+                    opt.process_compile_state(self)
+
+        # uncomment to print out the context.attributes structure
+        # after it's been set up above
+        # self._dump_option_struct()
+
+        if select_statement._with_context_options:
+            for fn, key in select_statement._with_context_options:
+                fn(self)
+
+        self.primary_columns = []
+        self.secondary_columns = []
+        self.dedupe_columns = set()
+        self.eager_joins = {}
+        self.extra_criteria_entities = {}
+        self.create_eager_joins = []
+        self._fallback_from_clauses = []
+
+        # normalize the FROM clauses early by themselves, as this makes
+        # it an easier job when we need to assemble a JOIN onto these,
+        # for select.join() as well as joinedload().   As of 1.4 there are now
+        # potentially more complex sets of FROM objects here as the use
+        # of lambda statements for lazyload, load_on_pk etc. uses more
+        # cloning of the select() construct.  See #6495
+        self.from_clauses = self._normalize_froms(
+            info.selectable for info in select_statement._from_obj
+        )
+
+        # this is a fairly arbitrary break into a second method,
+        # so it might be nicer to break up create_for_statement()
+        # and _setup_for_generate into three or four logical sections
+        self._setup_for_generate()
+
+        SelectState.__init__(self, self.statement, compiler, **kw)
+        return self
+
+    def _dump_option_struct(self):
+        print("\n---------------------------------------------------\n")
+        print(f"current path: {self.current_path}")
+        for key in self.attributes:
+            if isinstance(key, tuple) and key[0] == "loader":
+                print(f"\nLoader:           {PathRegistry.coerce(key[1])}")
+                print(f"    {self.attributes[key]}")
+                print(f"    {self.attributes[key].__dict__}")
+            elif isinstance(key, tuple) and key[0] == "path_with_polymorphic":
+                print(f"\nWith Polymorphic: {PathRegistry.coerce(key[1])}")
+                print(f"    {self.attributes[key]}")
+
+    def _setup_for_generate(self):
+        query = self.select_statement
+
+        self.statement = None
+        self._join_entities = ()
+
+        if self.compile_options._set_base_alias:
+            # legacy Query only
+            self._set_select_from_alias()
+
+        for memoized_entities in query._memoized_select_entities:
+            if memoized_entities._setup_joins:
+                self._join(
+                    memoized_entities._setup_joins,
+                    self._memoized_entities[memoized_entities],
+                )
+
+        if query._setup_joins:
+            self._join(query._setup_joins, self._entities)
+
+        current_adapter = self._get_current_adapter()
+
+        if query._where_criteria:
+            self._where_criteria = query._where_criteria
+
+            if current_adapter:
+                self._where_criteria = tuple(
+                    current_adapter(crit, True)
+                    for crit in self._where_criteria
+                )
+
+        # TODO: some complexity with order_by here was due to mapper.order_by.
+        # now that this is removed we can hopefully make order_by /
+        # group_by act identically to how they are in Core select.
+        self.order_by = (
+            self._adapt_col_list(query._order_by_clauses, current_adapter)
+            if current_adapter and query._order_by_clauses not in (None, False)
+            else query._order_by_clauses
+        )
+
+        if query._having_criteria:
+            self._having_criteria = tuple(
+                current_adapter(crit, True) if current_adapter else crit
+                for crit in query._having_criteria
+            )
+
+        self.group_by = (
+            self._adapt_col_list(
+                util.flatten_iterator(query._group_by_clauses), current_adapter
+            )
+            if current_adapter and query._group_by_clauses not in (None, False)
+            else query._group_by_clauses or None
+        )
+
+        if self.eager_order_by:
+            adapter = self.from_clauses[0]._target_adapter
+            self.eager_order_by = adapter.copy_and_process(self.eager_order_by)
+
+        if query._distinct_on:
+            self.distinct_on = self._adapt_col_list(
+                query._distinct_on, current_adapter
+            )
+        else:
+            self.distinct_on = ()
+
+        self.distinct = query._distinct
+
+        if query._correlate:
+            # ORM mapped entities that are mapped to joins can be passed
+            # to .correlate, so here they are broken into their component
+            # tables.
+            self.correlate = tuple(
+                util.flatten_iterator(
+                    sql_util.surface_selectables(s) if s is not None else None
+                    for s in query._correlate
+                )
+            )
+        elif query._correlate_except is not None:
+            self.correlate_except = tuple(
+                util.flatten_iterator(
+                    sql_util.surface_selectables(s) if s is not None else None
+                    for s in query._correlate_except
+                )
+            )
+        elif not query._auto_correlate:
+            self.correlate = (None,)
+
+        # PART II
+
+        self._for_update_arg = query._for_update_arg
+
+        if self.compile_options._is_star and (len(self._entities) != 1):
+            raise sa_exc.CompileError(
+                "Can't generate ORM query that includes multiple expressions "
+                "at the same time as '*'; query for '*' alone if present"
+            )
+        for entity in self._entities:
+            entity.setup_compile_state(self)
+
+        for rec in self.create_eager_joins:
+            strategy = rec[0]
+            strategy(self, *rec[1:])
+
+        # else "load from discrete FROMs" mode,
+        # i.e. when each _MappedEntity has its own FROM
+
+        if self.compile_options._enable_single_crit:
+            self._adjust_for_extra_criteria()
+
+        if not self.primary_columns:
+            if self.compile_options._only_load_props:
+                assert False, "no columns were included in _only_load_props"
+
+            raise sa_exc.InvalidRequestError(
+                "Query contains no columns with which to SELECT from."
+            )
+
+        if not self.from_clauses:
+            self.from_clauses = list(self._fallback_from_clauses)
+
+        if self.order_by is False:
+            self.order_by = None
+
+        if (
+            self.multi_row_eager_loaders
+            and self.eager_adding_joins
+            and self._should_nest_selectable
+        ):
+            self.statement = self._compound_eager_statement()
+        else:
+            self.statement = self._simple_statement()
+
+        if self.for_statement:
+            ezero = self._mapper_zero()
+            if ezero is not None:
+                # TODO: this goes away once we get rid of the deep entity
+                # thing
+                self.statement = self.statement._annotate(
+                    {"deepentity": ezero}
+                )
+
+    @classmethod
+    def _create_entities_collection(cls, query, legacy):
+        """Creates a partial ORMSelectCompileState that includes
+        the full collection of _MapperEntity and other _QueryEntity objects.
+
+        Supports a few remaining use cases that are pre-compilation
+        but still need to gather some of the column  / adaption information.
+
+        """
+        self = cls.__new__(cls)
+
+        self._entities = []
+        self._primary_entity = None
+        self._polymorphic_adapters = {}
+
+        self._label_convention = self._column_naming_convention(
+            query._label_style, legacy
+        )
+
+        # entities will also set up polymorphic adapters for mappers
+        # that have with_polymorphic configured
+        _QueryEntity.to_compile_state(
+            self, query._raw_columns, self._entities, is_current_entities=True
+        )
+        return self
+
+    @classmethod
+    def determine_last_joined_entity(cls, statement):
+        setup_joins = statement._setup_joins
+
+        return _determine_last_joined_entity(setup_joins, None)
+
+    @classmethod
+    def all_selected_columns(cls, statement):
+        for element in statement._raw_columns:
+            if (
+                element.is_selectable
+                and "entity_namespace" in element._annotations
+            ):
+                ens = element._annotations["entity_namespace"]
+                if not ens.is_mapper and not ens.is_aliased_class:
+                    yield from _select_iterables([element])
+                else:
+                    yield from _select_iterables(ens._all_column_expressions)
+            else:
+                yield from _select_iterables([element])
+
+    @classmethod
+    def get_columns_clause_froms(cls, statement):
+        return cls._normalize_froms(
+            itertools.chain.from_iterable(
+                (
+                    element._from_objects
+                    if "parententity" not in element._annotations
+                    else [
+                        element._annotations[
+                            "parententity"
+                        ].__clause_element__()
+                    ]
+                )
+                for element in statement._raw_columns
+            )
+        )
+
+    @classmethod
+    def from_statement(cls, statement, from_statement):
+        from_statement = coercions.expect(
+            roles.ReturnsRowsRole,
+            from_statement,
+            apply_propagate_attrs=statement,
+        )
+
+        stmt = FromStatement(statement._raw_columns, from_statement)
+
+        stmt.__dict__.update(
+            _with_options=statement._with_options,
+            _with_context_options=statement._with_context_options,
+            _execution_options=statement._execution_options,
+            _propagate_attrs=statement._propagate_attrs,
+        )
+        return stmt
+
+    def _set_select_from_alias(self):
+        """used only for legacy Query cases"""
+
+        query = self.select_statement  # query
+
+        assert self.compile_options._set_base_alias
+        assert len(query._from_obj) == 1
+
+        adapter = self._get_select_from_alias_from_obj(query._from_obj[0])
+        if adapter:
+            self.compile_options += {"_enable_single_crit": False}
+            self._from_obj_alias = adapter
+
+    def _get_select_from_alias_from_obj(self, from_obj):
+        """used only for legacy Query cases"""
+
+        info = from_obj
+
+        if "parententity" in info._annotations:
+            info = info._annotations["parententity"]
+
+        if hasattr(info, "mapper"):
+            if not info.is_aliased_class:
+                raise sa_exc.ArgumentError(
+                    "A selectable (FromClause) instance is "
+                    "expected when the base alias is being set."
+                )
+            else:
+                return info._adapter
+
+        elif isinstance(info.selectable, sql.selectable.AliasedReturnsRows):
+            equivs = self._all_equivs()
+            assert info is info.selectable
+            return ORMStatementAdapter(
+                _TraceAdaptRole.LEGACY_SELECT_FROM_ALIAS,
+                info.selectable,
+                equivalents=equivs,
+            )
+        else:
+            return None
+
+    def _mapper_zero(self):
+        """return the Mapper associated with the first QueryEntity."""
+        return self._entities[0].mapper
+
+    def _entity_zero(self):
+        """Return the 'entity' (mapper or AliasedClass) associated
+        with the first QueryEntity, or alternatively the 'select from'
+        entity if specified."""
+
+        for ent in self.from_clauses:
+            if "parententity" in ent._annotations:
+                return ent._annotations["parententity"]
+        for qent in self._entities:
+            if qent.entity_zero:
+                return qent.entity_zero
+
+        return None
+
+    def _only_full_mapper_zero(self, methname):
+        if self._entities != [self._primary_entity]:
+            raise sa_exc.InvalidRequestError(
+                "%s() can only be used against "
+                "a single mapped class." % methname
+            )
+        return self._primary_entity.entity_zero
+
+    def _only_entity_zero(self, rationale=None):
+        if len(self._entities) > 1:
+            raise sa_exc.InvalidRequestError(
+                rationale
+                or "This operation requires a Query "
+                "against a single mapper."
+            )
+        return self._entity_zero()
+
+    def _all_equivs(self):
+        equivs = {}
+
+        for memoized_entities in self._memoized_entities.values():
+            for ent in [
+                ent
+                for ent in memoized_entities
+                if isinstance(ent, _MapperEntity)
+            ]:
+                equivs.update(ent.mapper._equivalent_columns)
+
+        for ent in [
+            ent for ent in self._entities if isinstance(ent, _MapperEntity)
+        ]:
+            equivs.update(ent.mapper._equivalent_columns)
+        return equivs
+
+    def _compound_eager_statement(self):
+        # for eager joins present and LIMIT/OFFSET/DISTINCT,
+        # wrap the query inside a select,
+        # then append eager joins onto that
+
+        if self.order_by:
+            # the default coercion for ORDER BY is now the OrderByRole,
+            # which adds an additional post coercion to ByOfRole in that
+            # elements are converted into label references.  For the
+            # eager load / subquery wrapping case, we need to un-coerce
+            # the original expressions outside of the label references
+            # in order to have them render.
+            unwrapped_order_by = [
+                (
+                    elem.element
+                    if isinstance(elem, sql.elements._label_reference)
+                    else elem
+                )
+                for elem in self.order_by
+            ]
+
+            order_by_col_expr = sql_util.expand_column_list_from_order_by(
+                self.primary_columns, unwrapped_order_by
+            )
+        else:
+            order_by_col_expr = []
+            unwrapped_order_by = None
+
+        # put FOR UPDATE on the inner query, where MySQL will honor it,
+        # as well as if it has an OF so PostgreSQL can use it.
+        inner = self._select_statement(
+            self.primary_columns
+            + [c for c in order_by_col_expr if c not in self.dedupe_columns],
+            self.from_clauses,
+            self._where_criteria,
+            self._having_criteria,
+            self.label_style,
+            self.order_by,
+            for_update=self._for_update_arg,
+            hints=self.select_statement._hints,
+            statement_hints=self.select_statement._statement_hints,
+            correlate=self.correlate,
+            correlate_except=self.correlate_except,
+            **self._select_args,
+        )
+
+        inner = inner.alias()
+
+        equivs = self._all_equivs()
+
+        self.compound_eager_adapter = ORMStatementAdapter(
+            _TraceAdaptRole.COMPOUND_EAGER_STATEMENT, inner, equivalents=equivs
+        )
+
+        statement = future.select(
+            *([inner] + self.secondary_columns)  # use_labels=self.labels
+        )
+        statement._label_style = self.label_style
+
+        # Oracle Database however does not allow FOR UPDATE on the subquery,
+        # and the Oracle Database dialects ignore it, plus for PostgreSQL,
+        # MySQL we expect that all elements of the row are locked, so also put
+        # it on the outside (except in the case of PG when OF is used)
+        if (
+            self._for_update_arg is not None
+            and self._for_update_arg.of is None
+        ):
+            statement._for_update_arg = self._for_update_arg
+
+        from_clause = inner
+        for eager_join in self.eager_joins.values():
+            # EagerLoader places a 'stop_on' attribute on the join,
+            # giving us a marker as to where the "splice point" of
+            # the join should be
+            from_clause = sql_util.splice_joins(
+                from_clause, eager_join, eager_join.stop_on
+            )
+
+        statement.select_from.non_generative(statement, from_clause)
+
+        if unwrapped_order_by:
+            statement.order_by.non_generative(
+                statement,
+                *self.compound_eager_adapter.copy_and_process(
+                    unwrapped_order_by
+                ),
+            )
+
+        statement.order_by.non_generative(statement, *self.eager_order_by)
+        return statement
+
+    def _simple_statement(self):
+        statement = self._select_statement(
+            self.primary_columns + self.secondary_columns,
+            tuple(self.from_clauses) + tuple(self.eager_joins.values()),
+            self._where_criteria,
+            self._having_criteria,
+            self.label_style,
+            self.order_by,
+            for_update=self._for_update_arg,
+            hints=self.select_statement._hints,
+            statement_hints=self.select_statement._statement_hints,
+            correlate=self.correlate,
+            correlate_except=self.correlate_except,
+            **self._select_args,
+        )
+
+        if self.eager_order_by:
+            statement.order_by.non_generative(statement, *self.eager_order_by)
+        return statement
+
+    def _select_statement(
+        self,
+        raw_columns,
+        from_obj,
+        where_criteria,
+        having_criteria,
+        label_style,
+        order_by,
+        for_update,
+        hints,
+        statement_hints,
+        correlate,
+        correlate_except,
+        limit_clause,
+        offset_clause,
+        fetch_clause,
+        fetch_clause_options,
+        distinct,
+        distinct_on,
+        prefixes,
+        suffixes,
+        group_by,
+        independent_ctes,
+        independent_ctes_opts,
+    ):
+        statement = Select._create_raw_select(
+            _raw_columns=raw_columns,
+            _from_obj=from_obj,
+            _label_style=label_style,
+        )
+
+        if where_criteria:
+            statement._where_criteria = where_criteria
+        if having_criteria:
+            statement._having_criteria = having_criteria
+
+        if order_by:
+            statement._order_by_clauses += tuple(order_by)
+
+        if distinct_on:
+            statement.distinct.non_generative(statement, *distinct_on)
+        elif distinct:
+            statement.distinct.non_generative(statement)
+
+        if group_by:
+            statement._group_by_clauses += tuple(group_by)
+
+        statement._limit_clause = limit_clause
+        statement._offset_clause = offset_clause
+        statement._fetch_clause = fetch_clause
+        statement._fetch_clause_options = fetch_clause_options
+        statement._independent_ctes = independent_ctes
+        statement._independent_ctes_opts = independent_ctes_opts
+
+        if prefixes:
+            statement._prefixes = prefixes
+
+        if suffixes:
+            statement._suffixes = suffixes
+
+        statement._for_update_arg = for_update
+
+        if hints:
+            statement._hints = hints
+        if statement_hints:
+            statement._statement_hints = statement_hints
+
+        if correlate:
+            statement.correlate.non_generative(statement, *correlate)
+
+        if correlate_except is not None:
+            statement.correlate_except.non_generative(
+                statement, *correlate_except
+            )
+
+        return statement
+
+    def _adapt_polymorphic_element(self, element):
+        if "parententity" in element._annotations:
+            search = element._annotations["parententity"]
+            alias = self._polymorphic_adapters.get(search, None)
+            if alias:
+                return alias.adapt_clause(element)
+
+        if isinstance(element, expression.FromClause):
+            search = element
+        elif hasattr(element, "table"):
+            search = element.table
+        else:
+            return None
+
+        alias = self._polymorphic_adapters.get(search, None)
+        if alias:
+            return alias.adapt_clause(element)
+
+    def _adapt_col_list(self, cols, current_adapter):
+        if current_adapter:
+            return [current_adapter(o, True) for o in cols]
+        else:
+            return cols
+
+    def _get_current_adapter(self):
+        adapters = []
+
+        if self._from_obj_alias:
+            # used for legacy going forward for query set_ops, e.g.
+            # union(), union_all(), etc.
+            # 1.4 and previously, also used for from_self(),
+            # select_entity_from()
+            #
+            # for the "from obj" alias, apply extra rule to the
+            # 'ORM only' check, if this query were generated from a
+            # subquery of itself, i.e. _from_selectable(), apply adaption
+            # to all SQL constructs.
+            adapters.append(
+                (
+                    True,
+                    self._from_obj_alias.replace,
+                )
+            )
+
+        # this was *hopefully* the only adapter we were going to need
+        # going forward...however, we unfortunately need _from_obj_alias
+        # for query.union(), which we can't drop
+        if self._polymorphic_adapters:
+            adapters.append((False, self._adapt_polymorphic_element))
+
+        if not adapters:
+            return None
+
+        def _adapt_clause(clause, as_filter):
+            # do we adapt all expression elements or only those
+            # tagged as 'ORM' constructs ?
+
+            def replace(elem):
+                is_orm_adapt = (
+                    "_orm_adapt" in elem._annotations
+                    or "parententity" in elem._annotations
+                )
+                for always_adapt, adapter in adapters:
+                    if is_orm_adapt or always_adapt:
+                        e = adapter(elem)
+                        if e is not None:
+                            return e
+
+            return visitors.replacement_traverse(clause, {}, replace)
+
+        return _adapt_clause
+
+    def _join(self, args, entities_collection):
+        for right, onclause, from_, flags in args:
+            isouter = flags["isouter"]
+            full = flags["full"]
+
+            right = inspect(right)
+            if onclause is not None:
+                onclause = inspect(onclause)
+
+            if isinstance(right, interfaces.PropComparator):
+                if onclause is not None:
+                    raise sa_exc.InvalidRequestError(
+                        "No 'on clause' argument may be passed when joining "
+                        "to a relationship path as a target"
+                    )
+
+                onclause = right
+                right = None
+            elif "parententity" in right._annotations:
+                right = right._annotations["parententity"]
+
+            if onclause is None:
+                if not right.is_selectable and not hasattr(right, "mapper"):
+                    raise sa_exc.ArgumentError(
+                        "Expected mapped entity or "
+                        "selectable/table as join target"
+                    )
+
+            of_type = None
+
+            if isinstance(onclause, interfaces.PropComparator):
+                # descriptor/property given (or determined); this tells us
+                # explicitly what the expected "left" side of the join is.
+
+                of_type = getattr(onclause, "_of_type", None)
+
+                if right is None:
+                    if of_type:
+                        right = of_type
+                    else:
+                        right = onclause.property
+
+                        try:
+                            right = right.entity
+                        except AttributeError as err:
+                            raise sa_exc.ArgumentError(
+                                "Join target %s does not refer to a "
+                                "mapped entity" % right
+                            ) from err
+
+                left = onclause._parententity
+
+                prop = onclause.property
+                if not isinstance(onclause, attributes.QueryableAttribute):
+                    onclause = prop
+
+                # check for this path already present.  don't render in that
+                # case.
+                if (left, right, prop.key) in self._already_joined_edges:
+                    continue
+
+                if from_ is not None:
+                    if (
+                        from_ is not left
+                        and from_._annotations.get("parententity", None)
+                        is not left
+                    ):
+                        raise sa_exc.InvalidRequestError(
+                            "explicit from clause %s does not match left side "
+                            "of relationship attribute %s"
+                            % (
+                                from_._annotations.get("parententity", from_),
+                                onclause,
+                            )
+                        )
+            elif from_ is not None:
+                prop = None
+                left = from_
+            else:
+                # no descriptor/property given; we will need to figure out
+                # what the effective "left" side is
+                prop = left = None
+
+            # figure out the final "left" and "right" sides and create an
+            # ORMJoin to add to our _from_obj tuple
+            self._join_left_to_right(
+                entities_collection,
+                left,
+                right,
+                onclause,
+                prop,
+                isouter,
+                full,
+            )
+
+    def _join_left_to_right(
+        self,
+        entities_collection,
+        left,
+        right,
+        onclause,
+        prop,
+        outerjoin,
+        full,
+    ):
+        """given raw "left", "right", "onclause" parameters consumed from
+        a particular key within _join(), add a real ORMJoin object to
+        our _from_obj list (or augment an existing one)
+
+        """
+
+        if left is None:
+            # left not given (e.g. no relationship object/name specified)
+            # figure out the best "left" side based on our existing froms /
+            # entities
+            assert prop is None
+            (
+                left,
+                replace_from_obj_index,
+                use_entity_index,
+            ) = self._join_determine_implicit_left_side(
+                entities_collection, left, right, onclause
+            )
+        else:
+            # left is given via a relationship/name, or as explicit left side.
+            # Determine where in our
+            # "froms" list it should be spliced/appended as well as what
+            # existing entity it corresponds to.
+            (
+                replace_from_obj_index,
+                use_entity_index,
+            ) = self._join_place_explicit_left_side(entities_collection, left)
+
+        if left is right:
+            raise sa_exc.InvalidRequestError(
+                "Can't construct a join from %s to %s, they "
+                "are the same entity" % (left, right)
+            )
+
+        # the right side as given often needs to be adapted.  additionally
+        # a lot of things can be wrong with it.  handle all that and
+        # get back the new effective "right" side
+        r_info, right, onclause = self._join_check_and_adapt_right_side(
+            left, right, onclause, prop
+        )
+
+        if not r_info.is_selectable:
+            extra_criteria = self._get_extra_criteria(r_info)
+        else:
+            extra_criteria = ()
+
+        if replace_from_obj_index is not None:
+            # splice into an existing element in the
+            # self._from_obj list
+            left_clause = self.from_clauses[replace_from_obj_index]
+
+            self.from_clauses = (
+                self.from_clauses[:replace_from_obj_index]
+                + [
+                    _ORMJoin(
+                        left_clause,
+                        right,
+                        onclause,
+                        isouter=outerjoin,
+                        full=full,
+                        _extra_criteria=extra_criteria,
+                    )
+                ]
+                + self.from_clauses[replace_from_obj_index + 1 :]
+            )
+        else:
+            # add a new element to the self._from_obj list
+            if use_entity_index is not None:
+                # make use of _MapperEntity selectable, which is usually
+                # entity_zero.selectable, but if with_polymorphic() were used
+                # might be distinct
+                assert isinstance(
+                    entities_collection[use_entity_index], _MapperEntity
+                )
+                left_clause = entities_collection[use_entity_index].selectable
+            else:
+                left_clause = left
+
+            self.from_clauses = self.from_clauses + [
+                _ORMJoin(
+                    left_clause,
+                    r_info,
+                    onclause,
+                    isouter=outerjoin,
+                    full=full,
+                    _extra_criteria=extra_criteria,
+                )
+            ]
+
+    def _join_determine_implicit_left_side(
+        self, entities_collection, left, right, onclause
+    ):
+        """When join conditions don't express the left side explicitly,
+        determine if an existing FROM or entity in this query
+        can serve as the left hand side.
+
+        """
+
+        # when we are here, it means join() was called without an ORM-
+        # specific way of telling us what the "left" side is, e.g.:
+        #
+        # join(RightEntity)
+        #
+        # or
+        #
+        # join(RightEntity, RightEntity.foo == LeftEntity.bar)
+        #
+
+        r_info = inspect(right)
+
+        replace_from_obj_index = use_entity_index = None
+
+        if self.from_clauses:
+            # we have a list of FROMs already.  So by definition this
+            # join has to connect to one of those FROMs.
+
+            indexes = sql_util.find_left_clause_to_join_from(
+                self.from_clauses, r_info.selectable, onclause
+            )
+
+            if len(indexes) == 1:
+                replace_from_obj_index = indexes[0]
+                left = self.from_clauses[replace_from_obj_index]
+            elif len(indexes) > 1:
+                raise sa_exc.InvalidRequestError(
+                    "Can't determine which FROM clause to join "
+                    "from, there are multiple FROMS which can "
+                    "join to this entity. Please use the .select_from() "
+                    "method to establish an explicit left side, as well as "
+                    "providing an explicit ON clause if not present already "
+                    "to help resolve the ambiguity."
+                )
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "Don't know how to join to %r. "
+                    "Please use the .select_from() "
+                    "method to establish an explicit left side, as well as "
+                    "providing an explicit ON clause if not present already "
+                    "to help resolve the ambiguity." % (right,)
+                )
+
+        elif entities_collection:
+            # we have no explicit FROMs, so the implicit left has to
+            # come from our list of entities.
+
+            potential = {}
+            for entity_index, ent in enumerate(entities_collection):
+                entity = ent.entity_zero_or_selectable
+                if entity is None:
+                    continue
+                ent_info = inspect(entity)
+                if ent_info is r_info:  # left and right are the same, skip
+                    continue
+
+                # by using a dictionary with the selectables as keys this
+                # de-duplicates those selectables as occurs when the query is
+                # against a series of columns from the same selectable
+                if isinstance(ent, _MapperEntity):
+                    potential[ent.selectable] = (entity_index, entity)
+                else:
+                    potential[ent_info.selectable] = (None, entity)
+
+            all_clauses = list(potential.keys())
+            indexes = sql_util.find_left_clause_to_join_from(
+                all_clauses, r_info.selectable, onclause
+            )
+
+            if len(indexes) == 1:
+                use_entity_index, left = potential[all_clauses[indexes[0]]]
+            elif len(indexes) > 1:
+                raise sa_exc.InvalidRequestError(
+                    "Can't determine which FROM clause to join "
+                    "from, there are multiple FROMS which can "
+                    "join to this entity. Please use the .select_from() "
+                    "method to establish an explicit left side, as well as "
+                    "providing an explicit ON clause if not present already "
+                    "to help resolve the ambiguity."
+                )
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "Don't know how to join to %r. "
+                    "Please use the .select_from() "
+                    "method to establish an explicit left side, as well as "
+                    "providing an explicit ON clause if not present already "
+                    "to help resolve the ambiguity." % (right,)
+                )
+        else:
+            raise sa_exc.InvalidRequestError(
+                "No entities to join from; please use "
+                "select_from() to establish the left "
+                "entity/selectable of this join"
+            )
+
+        return left, replace_from_obj_index, use_entity_index
+
+    def _join_place_explicit_left_side(self, entities_collection, left):
+        """When join conditions express a left side explicitly, determine
+        where in our existing list of FROM clauses we should join towards,
+        or if we need to make a new join, and if so is it from one of our
+        existing entities.
+
+        """
+
+        # when we are here, it means join() was called with an indicator
+        # as to an exact left side, which means a path to a
+        # Relationship was given, e.g.:
+        #
+        # join(RightEntity, LeftEntity.right)
+        #
+        # or
+        #
+        # join(LeftEntity.right)
+        #
+        # as well as string forms:
+        #
+        # join(RightEntity, "right")
+        #
+        # etc.
+        #
+
+        replace_from_obj_index = use_entity_index = None
+
+        l_info = inspect(left)
+        if self.from_clauses:
+            indexes = sql_util.find_left_clause_that_matches_given(
+                self.from_clauses, l_info.selectable
+            )
+
+            if len(indexes) > 1:
+                raise sa_exc.InvalidRequestError(
+                    "Can't identify which entity in which to assign the "
+                    "left side of this join.   Please use a more specific "
+                    "ON clause."
+                )
+
+            # have an index, means the left side is already present in
+            # an existing FROM in the self._from_obj tuple
+            if indexes:
+                replace_from_obj_index = indexes[0]
+
+            # no index, means we need to add a new element to the
+            # self._from_obj tuple
+
+        # no from element present, so we will have to add to the
+        # self._from_obj tuple.  Determine if this left side matches up
+        # with existing mapper entities, in which case we want to apply the
+        # aliasing / adaptation rules present on that entity if any
+        if (
+            replace_from_obj_index is None
+            and entities_collection
+            and hasattr(l_info, "mapper")
+        ):
+            for idx, ent in enumerate(entities_collection):
+                # TODO: should we be checking for multiple mapper entities
+                # matching?
+                if isinstance(ent, _MapperEntity) and ent.corresponds_to(left):
+                    use_entity_index = idx
+                    break
+
+        return replace_from_obj_index, use_entity_index
+
+    def _join_check_and_adapt_right_side(self, left, right, onclause, prop):
+        """transform the "right" side of the join as well as the onclause
+        according to polymorphic mapping translations, aliasing on the query
+        or on the join, special cases where the right and left side have
+        overlapping tables.
+
+        """
+
+        l_info = inspect(left)
+        r_info = inspect(right)
+
+        overlap = False
+
+        right_mapper = getattr(r_info, "mapper", None)
+        # if the target is a joined inheritance mapping,
+        # be more liberal about auto-aliasing.
+        if right_mapper and (
+            right_mapper.with_polymorphic
+            or isinstance(right_mapper.persist_selectable, expression.Join)
+        ):
+            for from_obj in self.from_clauses or [l_info.selectable]:
+                if sql_util.selectables_overlap(
+                    l_info.selectable, from_obj
+                ) and sql_util.selectables_overlap(
+                    from_obj, r_info.selectable
+                ):
+                    overlap = True
+                    break
+
+        if overlap and l_info.selectable is r_info.selectable:
+            raise sa_exc.InvalidRequestError(
+                "Can't join table/selectable '%s' to itself"
+                % l_info.selectable
+            )
+
+        right_mapper, right_selectable, right_is_aliased = (
+            getattr(r_info, "mapper", None),
+            r_info.selectable,
+            getattr(r_info, "is_aliased_class", False),
+        )
+
+        if (
+            right_mapper
+            and prop
+            and not right_mapper.common_parent(prop.mapper)
+        ):
+            raise sa_exc.InvalidRequestError(
+                "Join target %s does not correspond to "
+                "the right side of join condition %s" % (right, onclause)
+            )
+
+        # _join_entities is used as a hint for single-table inheritance
+        # purposes at the moment
+        if hasattr(r_info, "mapper"):
+            self._join_entities += (r_info,)
+
+        need_adapter = False
+
+        # test for joining to an unmapped selectable as the target
+        if r_info.is_clause_element:
+            if prop:
+                right_mapper = prop.mapper
+
+            if right_selectable._is_lateral:
+                # orm_only is disabled to suit the case where we have to
+                # adapt an explicit correlate(Entity) - the select() loses
+                # the ORM-ness in this case right now, ideally it would not
+                current_adapter = self._get_current_adapter()
+                if current_adapter is not None:
+                    # TODO: we had orm_only=False here before, removing
+                    # it didn't break things.   if we identify the rationale,
+                    # may need to apply "_orm_only" annotation here.
+                    right = current_adapter(right, True)
+
+            elif prop:
+                # joining to selectable with a mapper property given
+                # as the ON clause
+
+                if not right_selectable.is_derived_from(
+                    right_mapper.persist_selectable
+                ):
+                    raise sa_exc.InvalidRequestError(
+                        "Selectable '%s' is not derived from '%s'"
+                        % (
+                            right_selectable.description,
+                            right_mapper.persist_selectable.description,
+                        )
+                    )
+
+                # if the destination selectable is a plain select(),
+                # turn it into an alias().
+                if isinstance(right_selectable, expression.SelectBase):
+                    right_selectable = coercions.expect(
+                        roles.FromClauseRole, right_selectable
+                    )
+                    need_adapter = True
+
+                # make the right hand side target into an ORM entity
+                right = AliasedClass(right_mapper, right_selectable)
+
+                util.warn_deprecated(
+                    "An alias is being generated automatically against "
+                    "joined entity %s for raw clauseelement, which is "
+                    "deprecated and will be removed in a later release. "
+                    "Use the aliased() "
+                    "construct explicitly, see the linked example."
+                    % right_mapper,
+                    "1.4",
+                    code="xaj1",
+                )
+
+        # test for overlap:
+        # orm/inheritance/relationships.py
+        # SelfReferentialM2MTest
+        aliased_entity = right_mapper and not right_is_aliased and overlap
+
+        if not need_adapter and aliased_entity:
+            # there are a few places in the ORM that automatic aliasing
+            # is still desirable, and can't be automatic with a Core
+            # only approach.  For illustrations of "overlaps" see
+            # test/orm/inheritance/test_relationships.py.  There are also
+            # general overlap cases with many-to-many tables where automatic
+            # aliasing is desirable.
+            right = AliasedClass(right, flat=True)
+            need_adapter = True
+
+            util.warn(
+                "An alias is being generated automatically against "
+                "joined entity %s due to overlapping tables.  This is a "
+                "legacy pattern which may be "
+                "deprecated in a later release.  Use the "
+                "aliased(<entity>, flat=True) "
+                "construct explicitly, see the linked example." % right_mapper,
+                code="xaj2",
+            )
+
+        if need_adapter:
+            # if need_adapter is True, we are in a deprecated case and
+            # a warning has been emitted.
+            assert right_mapper
+
+            adapter = ORMAdapter(
+                _TraceAdaptRole.DEPRECATED_JOIN_ADAPT_RIGHT_SIDE,
+                inspect(right),
+                equivalents=right_mapper._equivalent_columns,
+            )
+
+            # if an alias() on the right side was generated,
+            # which is intended to wrap a the right side in a subquery,
+            # ensure that columns retrieved from this target in the result
+            # set are also adapted.
+            self._mapper_loads_polymorphically_with(right_mapper, adapter)
+        elif (
+            not r_info.is_clause_element
+            and not right_is_aliased
+            and right_mapper._has_aliased_polymorphic_fromclause
+        ):
+            # for the case where the target mapper has a with_polymorphic
+            # set up, ensure an adapter is set up for criteria that works
+            # against this mapper.  Previously, this logic used to
+            # use the "create_aliases or aliased_entity" case to generate
+            # an aliased() object, but this creates an alias that isn't
+            # strictly necessary.
+            # see test/orm/test_core_compilation.py
+            # ::RelNaturalAliasedJoinsTest::test_straight
+            # and similar
+            self._mapper_loads_polymorphically_with(
+                right_mapper,
+                ORMAdapter(
+                    _TraceAdaptRole.WITH_POLYMORPHIC_ADAPTER_RIGHT_JOIN,
+                    right_mapper,
+                    selectable=right_mapper.selectable,
+                    equivalents=right_mapper._equivalent_columns,
+                ),
+            )
+        # if the onclause is a ClauseElement, adapt it with any
+        # adapters that are in place right now
+        if isinstance(onclause, expression.ClauseElement):
+            current_adapter = self._get_current_adapter()
+            if current_adapter:
+                onclause = current_adapter(onclause, True)
+
+        # if joining on a MapperProperty path,
+        # track the path to prevent redundant joins
+        if prop:
+            self._already_joined_edges += ((left, right, prop.key),)
+
+        return inspect(right), right, onclause
+
+    @property
+    def _select_args(self):
+        return {
+            "limit_clause": self.select_statement._limit_clause,
+            "offset_clause": self.select_statement._offset_clause,
+            "distinct": self.distinct,
+            "distinct_on": self.distinct_on,
+            "prefixes": self.select_statement._prefixes,
+            "suffixes": self.select_statement._suffixes,
+            "group_by": self.group_by or None,
+            "fetch_clause": self.select_statement._fetch_clause,
+            "fetch_clause_options": (
+                self.select_statement._fetch_clause_options
+            ),
+            "independent_ctes": self.select_statement._independent_ctes,
+            "independent_ctes_opts": (
+                self.select_statement._independent_ctes_opts
+            ),
+        }
+
+    @property
+    def _should_nest_selectable(self):
+        kwargs = self._select_args
+        return (
+            kwargs.get("limit_clause") is not None
+            or kwargs.get("offset_clause") is not None
+            or kwargs.get("distinct", False)
+            or kwargs.get("distinct_on", ())
+            or kwargs.get("group_by", False)
+        )
+
+    def _get_extra_criteria(self, ext_info):
+        if (
+            "additional_entity_criteria",
+            ext_info.mapper,
+        ) in self.global_attributes:
+            return tuple(
+                ae._resolve_where_criteria(ext_info)
+                for ae in self.global_attributes[
+                    ("additional_entity_criteria", ext_info.mapper)
+                ]
+                if (ae.include_aliases or ae.entity is ext_info)
+                and ae._should_include(self)
+            )
+        else:
+            return ()
+
+    def _adjust_for_extra_criteria(self):
+        """Apply extra criteria filtering.
+
+        For all distinct single-table-inheritance mappers represented in
+        the columns clause of this query, as well as the "select from entity",
+        add criterion to the WHERE
+        clause of the given QueryContext such that only the appropriate
+        subtypes are selected from the total results.
+
+        Additionally, add WHERE criteria originating from LoaderCriteriaOptions
+        associated with the global context.
+
+        """
+
+        for fromclause in self.from_clauses:
+            ext_info = fromclause._annotations.get("parententity", None)
+
+            if (
+                ext_info
+                and (
+                    ext_info.mapper._single_table_criterion is not None
+                    or ("additional_entity_criteria", ext_info.mapper)
+                    in self.global_attributes
+                )
+                and ext_info not in self.extra_criteria_entities
+            ):
+                self.extra_criteria_entities[ext_info] = (
+                    ext_info,
+                    ext_info._adapter if ext_info.is_aliased_class else None,
+                )
+
+        search = set(self.extra_criteria_entities.values())
+
+        for ext_info, adapter in search:
+            if ext_info in self._join_entities:
+                continue
+
+            single_crit = ext_info.mapper._single_table_criterion
+
+            if self.compile_options._for_refresh_state:
+                additional_entity_criteria = []
+            else:
+                additional_entity_criteria = self._get_extra_criteria(ext_info)
+
+            if single_crit is not None:
+                additional_entity_criteria += (single_crit,)
+
+            current_adapter = self._get_current_adapter()
+            for crit in additional_entity_criteria:
+                if adapter:
+                    crit = adapter.traverse(crit)
+
+                if current_adapter:
+                    crit = sql_util._deep_annotate(crit, {"_orm_adapt": True})
+                    crit = current_adapter(crit, False)
+                self._where_criteria += (crit,)
+
+
+def _column_descriptions(
+    query_or_select_stmt: Union[Query, Select, FromStatement],
+    compile_state: Optional[ORMSelectCompileState] = None,
+    legacy: bool = False,
+) -> List[ORMColumnDescription]:
+    if compile_state is None:
+        compile_state = ORMSelectCompileState._create_entities_collection(
+            query_or_select_stmt, legacy=legacy
+        )
+    ctx = compile_state
+    d = [
+        {
+            "name": ent._label_name,
+            "type": ent.type,
+            "aliased": getattr(insp_ent, "is_aliased_class", False),
+            "expr": ent.expr,
+            "entity": (
+                getattr(insp_ent, "entity", None)
+                if ent.entity_zero is not None
+                and not insp_ent.is_clause_element
+                else None
+            ),
+        }
+        for ent, insp_ent in [
+            (_ent, _ent.entity_zero) for _ent in ctx._entities
+        ]
+    ]
+    return d
+
+
+def _legacy_filter_by_entity_zero(
+    query_or_augmented_select: Union[Query[Any], Select[Any]]
+) -> Optional[_InternalEntityType[Any]]:
+    self = query_or_augmented_select
+    if self._setup_joins:
+        _last_joined_entity = self._last_joined_entity
+        if _last_joined_entity is not None:
+            return _last_joined_entity
+
+    if self._from_obj and "parententity" in self._from_obj[0]._annotations:
+        return self._from_obj[0]._annotations["parententity"]
+
+    return _entity_from_pre_ent_zero(self)
+
+
+def _entity_from_pre_ent_zero(
+    query_or_augmented_select: Union[Query[Any], Select[Any]]
+) -> Optional[_InternalEntityType[Any]]:
+    self = query_or_augmented_select
+    if not self._raw_columns:
+        return None
+
+    ent = self._raw_columns[0]
+
+    if "parententity" in ent._annotations:
+        return ent._annotations["parententity"]
+    elif isinstance(ent, ORMColumnsClauseRole):
+        return ent.entity
+    elif "bundle" in ent._annotations:
+        return ent._annotations["bundle"]
+    else:
+        return ent
+
+
+def _determine_last_joined_entity(
+    setup_joins: Tuple[_SetupJoinsElement, ...],
+    entity_zero: Optional[_InternalEntityType[Any]] = None,
+) -> Optional[Union[_InternalEntityType[Any], _JoinTargetElement]]:
+    if not setup_joins:
+        return None
+
+    (target, onclause, from_, flags) = setup_joins[-1]
+
+    if isinstance(
+        target,
+        attributes.QueryableAttribute,
+    ):
+        return target.entity
+    else:
+        return target
+
+
+class _QueryEntity:
+    """represent an entity column returned within a Query result."""
+
+    __slots__ = ()
+
+    supports_single_entity: bool
+
+    _non_hashable_value = False
+    _null_column_type = False
+    use_id_for_hash = False
+
+    _label_name: Optional[str]
+    type: Union[Type[Any], TypeEngine[Any]]
+    expr: Union[_InternalEntityType, ColumnElement[Any]]
+    entity_zero: Optional[_InternalEntityType]
+
+    def setup_compile_state(self, compile_state: ORMCompileState) -> None:
+        raise NotImplementedError()
+
+    def setup_dml_returning_compile_state(
+        self,
+        compile_state: ORMCompileState,
+        adapter: DMLReturningColFilter,
+    ) -> None:
+        raise NotImplementedError()
+
+    def row_processor(self, context, result):
+        raise NotImplementedError()
+
+    @classmethod
+    def to_compile_state(
+        cls, compile_state, entities, entities_collection, is_current_entities
+    ):
+        for idx, entity in enumerate(entities):
+            if entity._is_lambda_element:
+                if entity._is_sequence:
+                    cls.to_compile_state(
+                        compile_state,
+                        entity._resolved,
+                        entities_collection,
+                        is_current_entities,
+                    )
+                    continue
+                else:
+                    entity = entity._resolved
+
+            if entity.is_clause_element:
+                if entity.is_selectable:
+                    if "parententity" in entity._annotations:
+                        _MapperEntity(
+                            compile_state,
+                            entity,
+                            entities_collection,
+                            is_current_entities,
+                        )
+                    else:
+                        _ColumnEntity._for_columns(
+                            compile_state,
+                            entity._select_iterable,
+                            entities_collection,
+                            idx,
+                            is_current_entities,
+                        )
+                else:
+                    if entity._annotations.get("bundle", False):
+                        _BundleEntity(
+                            compile_state,
+                            entity,
+                            entities_collection,
+                            is_current_entities,
+                        )
+                    elif entity._is_clause_list:
+                        # this is legacy only - test_composites.py
+                        # test_query_cols_legacy
+                        _ColumnEntity._for_columns(
+                            compile_state,
+                            entity._select_iterable,
+                            entities_collection,
+                            idx,
+                            is_current_entities,
+                        )
+                    else:
+                        _ColumnEntity._for_columns(
+                            compile_state,
+                            [entity],
+                            entities_collection,
+                            idx,
+                            is_current_entities,
+                        )
+            elif entity.is_bundle:
+                _BundleEntity(compile_state, entity, entities_collection)
+
+        return entities_collection
+
+
+class _MapperEntity(_QueryEntity):
+    """mapper/class/AliasedClass entity"""
+
+    __slots__ = (
+        "expr",
+        "mapper",
+        "entity_zero",
+        "is_aliased_class",
+        "path",
+        "_extra_entities",
+        "_label_name",
+        "_with_polymorphic_mappers",
+        "selectable",
+        "_polymorphic_discriminator",
+    )
+
+    expr: _InternalEntityType
+    mapper: Mapper[Any]
+    entity_zero: _InternalEntityType
+    is_aliased_class: bool
+    path: PathRegistry
+    _label_name: str
+
+    def __init__(
+        self, compile_state, entity, entities_collection, is_current_entities
+    ):
+        entities_collection.append(self)
+        if is_current_entities:
+            if compile_state._primary_entity is None:
+                compile_state._primary_entity = self
+            compile_state._has_mapper_entities = True
+            compile_state._has_orm_entities = True
+
+        entity = entity._annotations["parententity"]
+        entity._post_inspect
+        ext_info = self.entity_zero = entity
+        entity = ext_info.entity
+
+        self.expr = entity
+        self.mapper = mapper = ext_info.mapper
+
+        self._extra_entities = (self.expr,)
+
+        if ext_info.is_aliased_class:
+            self._label_name = ext_info.name
+        else:
+            self._label_name = mapper.class_.__name__
+
+        self.is_aliased_class = ext_info.is_aliased_class
+        self.path = ext_info._path_registry
+
+        self.selectable = ext_info.selectable
+        self._with_polymorphic_mappers = ext_info.with_polymorphic_mappers
+        self._polymorphic_discriminator = ext_info.polymorphic_on
+
+        if mapper._should_select_with_poly_adapter:
+            compile_state._create_with_polymorphic_adapter(
+                ext_info, self.selectable
+            )
+
+    supports_single_entity = True
+
+    _non_hashable_value = True
+    use_id_for_hash = True
+
+    @property
+    def type(self):
+        return self.mapper.class_
+
+    @property
+    def entity_zero_or_selectable(self):
+        return self.entity_zero
+
+    def corresponds_to(self, entity):
+        return _entity_corresponds_to(self.entity_zero, entity)
+
+    def _get_entity_clauses(self, compile_state):
+        adapter = None
+
+        if not self.is_aliased_class:
+            if compile_state._polymorphic_adapters:
+                adapter = compile_state._polymorphic_adapters.get(
+                    self.mapper, None
+                )
+        else:
+            adapter = self.entity_zero._adapter
+
+        if adapter:
+            if compile_state._from_obj_alias:
+                ret = adapter.wrap(compile_state._from_obj_alias)
+            else:
+                ret = adapter
+        else:
+            ret = compile_state._from_obj_alias
+
+        return ret
+
+    def row_processor(self, context, result):
+        compile_state = context.compile_state
+        adapter = self._get_entity_clauses(compile_state)
+
+        if compile_state.compound_eager_adapter and adapter:
+            adapter = adapter.wrap(compile_state.compound_eager_adapter)
+        elif not adapter:
+            adapter = compile_state.compound_eager_adapter
+
+        if compile_state._primary_entity is self:
+            only_load_props = compile_state.compile_options._only_load_props
+            refresh_state = context.refresh_state
+        else:
+            only_load_props = refresh_state = None
+
+        _instance = loading._instance_processor(
+            self,
+            self.mapper,
+            context,
+            result,
+            self.path,
+            adapter,
+            only_load_props=only_load_props,
+            refresh_state=refresh_state,
+            polymorphic_discriminator=self._polymorphic_discriminator,
+        )
+
+        return _instance, self._label_name, self._extra_entities
+
+    def setup_dml_returning_compile_state(
+        self,
+        compile_state: ORMCompileState,
+        adapter: DMLReturningColFilter,
+    ) -> None:
+        loading._setup_entity_query(
+            compile_state,
+            self.mapper,
+            self,
+            self.path,
+            adapter,
+            compile_state.primary_columns,
+            with_polymorphic=self._with_polymorphic_mappers,
+            only_load_props=compile_state.compile_options._only_load_props,
+            polymorphic_discriminator=self._polymorphic_discriminator,
+        )
+
+    def setup_compile_state(self, compile_state):
+        adapter = self._get_entity_clauses(compile_state)
+
+        single_table_crit = self.mapper._single_table_criterion
+        if (
+            single_table_crit is not None
+            or ("additional_entity_criteria", self.mapper)
+            in compile_state.global_attributes
+        ):
+            ext_info = self.entity_zero
+            compile_state.extra_criteria_entities[ext_info] = (
+                ext_info,
+                ext_info._adapter if ext_info.is_aliased_class else None,
+            )
+
+        loading._setup_entity_query(
+            compile_state,
+            self.mapper,
+            self,
+            self.path,
+            adapter,
+            compile_state.primary_columns,
+            with_polymorphic=self._with_polymorphic_mappers,
+            only_load_props=compile_state.compile_options._only_load_props,
+            polymorphic_discriminator=self._polymorphic_discriminator,
+        )
+        compile_state._fallback_from_clauses.append(self.selectable)
+
+
+class _BundleEntity(_QueryEntity):
+    _extra_entities = ()
+
+    __slots__ = (
+        "bundle",
+        "expr",
+        "type",
+        "_label_name",
+        "_entities",
+        "supports_single_entity",
+    )
+
+    _entities: List[_QueryEntity]
+    bundle: Bundle
+    type: Type[Any]
+    _label_name: str
+    supports_single_entity: bool
+    expr: Bundle
+
+    def __init__(
+        self,
+        compile_state,
+        expr,
+        entities_collection,
+        is_current_entities,
+        setup_entities=True,
+        parent_bundle=None,
+    ):
+        compile_state._has_orm_entities = True
+
+        expr = expr._annotations["bundle"]
+        if parent_bundle:
+            parent_bundle._entities.append(self)
+        else:
+            entities_collection.append(self)
+
+        if isinstance(
+            expr, (attributes.QueryableAttribute, interfaces.PropComparator)
+        ):
+            bundle = expr.__clause_element__()
+        else:
+            bundle = expr
+
+        self.bundle = self.expr = bundle
+        self.type = type(bundle)
+        self._label_name = bundle.name
+        self._entities = []
+
+        if setup_entities:
+            for expr in bundle.exprs:
+                if "bundle" in expr._annotations:
+                    _BundleEntity(
+                        compile_state,
+                        expr,
+                        entities_collection,
+                        is_current_entities,
+                        parent_bundle=self,
+                    )
+                elif isinstance(expr, Bundle):
+                    _BundleEntity(
+                        compile_state,
+                        expr,
+                        entities_collection,
+                        is_current_entities,
+                        parent_bundle=self,
+                    )
+                else:
+                    _ORMColumnEntity._for_columns(
+                        compile_state,
+                        [expr],
+                        entities_collection,
+                        None,
+                        is_current_entities,
+                        parent_bundle=self,
+                    )
+
+        self.supports_single_entity = self.bundle.single_entity
+
+    @property
+    def mapper(self):
+        ezero = self.entity_zero
+        if ezero is not None:
+            return ezero.mapper
+        else:
+            return None
+
+    @property
+    def entity_zero(self):
+        for ent in self._entities:
+            ezero = ent.entity_zero
+            if ezero is not None:
+                return ezero
+        else:
+            return None
+
+    def corresponds_to(self, entity):
+        # TODO: we might be able to implement this but for now
+        # we are working around it
+        return False
+
+    @property
+    def entity_zero_or_selectable(self):
+        for ent in self._entities:
+            ezero = ent.entity_zero_or_selectable
+            if ezero is not None:
+                return ezero
+        else:
+            return None
+
+    def setup_compile_state(self, compile_state):
+        for ent in self._entities:
+            ent.setup_compile_state(compile_state)
+
+    def setup_dml_returning_compile_state(
+        self,
+        compile_state: ORMCompileState,
+        adapter: DMLReturningColFilter,
+    ) -> None:
+        return self.setup_compile_state(compile_state)
+
+    def row_processor(self, context, result):
+        procs, labels, extra = zip(
+            *[ent.row_processor(context, result) for ent in self._entities]
+        )
+
+        proc = self.bundle.create_row_processor(context.query, procs, labels)
+
+        return proc, self._label_name, self._extra_entities
+
+
+class _ColumnEntity(_QueryEntity):
+    __slots__ = (
+        "_fetch_column",
+        "_row_processor",
+        "raw_column_index",
+        "translate_raw_column",
+    )
+
+    @classmethod
+    def _for_columns(
+        cls,
+        compile_state,
+        columns,
+        entities_collection,
+        raw_column_index,
+        is_current_entities,
+        parent_bundle=None,
+    ):
+        for column in columns:
+            annotations = column._annotations
+            if "parententity" in annotations:
+                _entity = annotations["parententity"]
+            else:
+                _entity = sql_util.extract_first_column_annotation(
+                    column, "parententity"
+                )
+
+            if _entity:
+                if "identity_token" in column._annotations:
+                    _IdentityTokenEntity(
+                        compile_state,
+                        column,
+                        entities_collection,
+                        _entity,
+                        raw_column_index,
+                        is_current_entities,
+                        parent_bundle=parent_bundle,
+                    )
+                else:
+                    _ORMColumnEntity(
+                        compile_state,
+                        column,
+                        entities_collection,
+                        _entity,
+                        raw_column_index,
+                        is_current_entities,
+                        parent_bundle=parent_bundle,
+                    )
+            else:
+                _RawColumnEntity(
+                    compile_state,
+                    column,
+                    entities_collection,
+                    raw_column_index,
+                    is_current_entities,
+                    parent_bundle=parent_bundle,
+                )
+
+    @property
+    def type(self):
+        return self.column.type
+
+    @property
+    def _non_hashable_value(self):
+        return not self.column.type.hashable
+
+    @property
+    def _null_column_type(self):
+        return self.column.type._isnull
+
+    def row_processor(self, context, result):
+        compile_state = context.compile_state
+
+        # the resulting callable is entirely cacheable so just return
+        # it if we already made one
+        if self._row_processor is not None:
+            getter, label_name, extra_entities = self._row_processor
+            if self.translate_raw_column:
+                extra_entities += (
+                    context.query._raw_columns[self.raw_column_index],
+                )
+
+            return getter, label_name, extra_entities
+
+        # retrieve the column that would have been set up in
+        # setup_compile_state, to avoid doing redundant work
+        if self._fetch_column is not None:
+            column = self._fetch_column
+        else:
+            # fetch_column will be None when we are doing a from_statement
+            # and setup_compile_state may not have been called.
+            column = self.column
+
+            # previously, the RawColumnEntity didn't look for from_obj_alias
+            # however I can't think of a case where we would be here and
+            # we'd want to ignore it if this is the from_statement use case.
+            # it's not really a use case to have raw columns + from_statement
+            if compile_state._from_obj_alias:
+                column = compile_state._from_obj_alias.columns[column]
+
+            if column._annotations:
+                # annotated columns perform more slowly in compiler and
+                # result due to the __eq__() method, so use deannotated
+                column = column._deannotate()
+
+        if compile_state.compound_eager_adapter:
+            column = compile_state.compound_eager_adapter.columns[column]
+
+        getter = result._getter(column)
+        ret = getter, self._label_name, self._extra_entities
+        self._row_processor = ret
+
+        if self.translate_raw_column:
+            extra_entities = self._extra_entities + (
+                context.query._raw_columns[self.raw_column_index],
+            )
+            return getter, self._label_name, extra_entities
+        else:
+            return ret
+
+
+class _RawColumnEntity(_ColumnEntity):
+    entity_zero = None
+    mapper = None
+    supports_single_entity = False
+
+    __slots__ = (
+        "expr",
+        "column",
+        "_label_name",
+        "entity_zero_or_selectable",
+        "_extra_entities",
+    )
+
+    def __init__(
+        self,
+        compile_state,
+        column,
+        entities_collection,
+        raw_column_index,
+        is_current_entities,
+        parent_bundle=None,
+    ):
+        self.expr = column
+        self.raw_column_index = raw_column_index
+        self.translate_raw_column = raw_column_index is not None
+
+        if column._is_star:
+            compile_state.compile_options += {"_is_star": True}
+
+        if not is_current_entities or column._is_text_clause:
+            self._label_name = None
+        else:
+            if parent_bundle:
+                self._label_name = column._proxy_key
+            else:
+                self._label_name = compile_state._label_convention(column)
+
+        if parent_bundle:
+            parent_bundle._entities.append(self)
+        else:
+            entities_collection.append(self)
+
+        self.column = column
+        self.entity_zero_or_selectable = (
+            self.column._from_objects[0] if self.column._from_objects else None
+        )
+        self._extra_entities = (self.expr, self.column)
+        self._fetch_column = self._row_processor = None
+
+    def corresponds_to(self, entity):
+        return False
+
+    def setup_dml_returning_compile_state(
+        self,
+        compile_state: ORMCompileState,
+        adapter: DMLReturningColFilter,
+    ) -> None:
+        return self.setup_compile_state(compile_state)
+
+    def setup_compile_state(self, compile_state):
+        current_adapter = compile_state._get_current_adapter()
+        if current_adapter:
+            column = current_adapter(self.column, False)
+            if column is None:
+                return
+        else:
+            column = self.column
+
+        if column._annotations:
+            # annotated columns perform more slowly in compiler and
+            # result due to the __eq__() method, so use deannotated
+            column = column._deannotate()
+
+        compile_state.dedupe_columns.add(column)
+        compile_state.primary_columns.append(column)
+        self._fetch_column = column
+
+
+class _ORMColumnEntity(_ColumnEntity):
+    """Column/expression based entity."""
+
+    supports_single_entity = False
+
+    __slots__ = (
+        "expr",
+        "mapper",
+        "column",
+        "_label_name",
+        "entity_zero_or_selectable",
+        "entity_zero",
+        "_extra_entities",
+    )
+
+    def __init__(
+        self,
+        compile_state,
+        column,
+        entities_collection,
+        parententity,
+        raw_column_index,
+        is_current_entities,
+        parent_bundle=None,
+    ):
+        annotations = column._annotations
+
+        _entity = parententity
+
+        # an AliasedClass won't have proxy_key in the annotations for
+        # a column if it was acquired using the class' adapter directly,
+        # such as using AliasedInsp._adapt_element().  this occurs
+        # within internal loaders.
+
+        orm_key = annotations.get("proxy_key", None)
+        proxy_owner = annotations.get("proxy_owner", _entity)
+        if orm_key:
+            self.expr = getattr(proxy_owner.entity, orm_key)
+            self.translate_raw_column = False
+        else:
+            # if orm_key is not present, that means this is an ad-hoc
+            # SQL ColumnElement, like a CASE() or other expression.
+            # include this column position from the invoked statement
+            # in the ORM-level ResultSetMetaData on each execute, so that
+            # it can be targeted by identity after caching
+            self.expr = column
+            self.translate_raw_column = raw_column_index is not None
+
+        self.raw_column_index = raw_column_index
+
+        if is_current_entities:
+            if parent_bundle:
+                self._label_name = orm_key if orm_key else column._proxy_key
+            else:
+                self._label_name = compile_state._label_convention(
+                    column, col_name=orm_key
+                )
+        else:
+            self._label_name = None
+
+        _entity._post_inspect
+        self.entity_zero = self.entity_zero_or_selectable = ezero = _entity
+        self.mapper = mapper = _entity.mapper
+
+        if parent_bundle:
+            parent_bundle._entities.append(self)
+        else:
+            entities_collection.append(self)
+
+        compile_state._has_orm_entities = True
+
+        self.column = column
+
+        self._fetch_column = self._row_processor = None
+
+        self._extra_entities = (self.expr, self.column)
+
+        if mapper._should_select_with_poly_adapter:
+            compile_state._create_with_polymorphic_adapter(
+                ezero, ezero.selectable
+            )
+
+    def corresponds_to(self, entity):
+        if _is_aliased_class(entity):
+            # TODO: polymorphic subclasses ?
+            return entity is self.entity_zero
+        else:
+            return not _is_aliased_class(
+                self.entity_zero
+            ) and entity.common_parent(self.entity_zero)
+
+    def setup_dml_returning_compile_state(
+        self,
+        compile_state: ORMCompileState,
+        adapter: DMLReturningColFilter,
+    ) -> None:
+        self._fetch_column = self.column
+        column = adapter(self.column, False)
+        if column is not None:
+            compile_state.dedupe_columns.add(column)
+            compile_state.primary_columns.append(column)
+
+    def setup_compile_state(self, compile_state):
+        current_adapter = compile_state._get_current_adapter()
+        if current_adapter:
+            column = current_adapter(self.column, False)
+            if column is None:
+                assert compile_state.is_dml_returning
+                self._fetch_column = self.column
+                return
+        else:
+            column = self.column
+
+        ezero = self.entity_zero
+
+        single_table_crit = self.mapper._single_table_criterion
+        if (
+            single_table_crit is not None
+            or ("additional_entity_criteria", self.mapper)
+            in compile_state.global_attributes
+        ):
+            compile_state.extra_criteria_entities[ezero] = (
+                ezero,
+                ezero._adapter if ezero.is_aliased_class else None,
+            )
+
+        if column._annotations and not column._expression_label:
+            # annotated columns perform more slowly in compiler and
+            # result due to the __eq__() method, so use deannotated
+            column = column._deannotate()
+
+        # use entity_zero as the from if we have it. this is necessary
+        # for polymorphic scenarios where our FROM is based on ORM entity,
+        # not the FROM of the column.  but also, don't use it if our column
+        # doesn't actually have any FROMs that line up, such as when its
+        # a scalar subquery.
+        if set(self.column._from_objects).intersection(
+            ezero.selectable._from_objects
+        ):
+            compile_state._fallback_from_clauses.append(ezero.selectable)
+
+        compile_state.dedupe_columns.add(column)
+        compile_state.primary_columns.append(column)
+        self._fetch_column = column
+
+
+class _IdentityTokenEntity(_ORMColumnEntity):
+    translate_raw_column = False
+
+    def setup_compile_state(self, compile_state):
+        pass
+
+    def row_processor(self, context, result):
+        def getter(row):
+            return context.load_options._identity_token
+
+        return getter, self._label_name, self._extra_entities
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_api.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_api.py
new file mode 100644
index 00000000..c32851de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_api.py
@@ -0,0 +1,1917 @@
+# orm/decl_api.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Public API functions and helpers for declarative."""
+
+from __future__ import annotations
+
+import itertools
+import re
+import typing
+from typing import Any
+from typing import Callable
+from typing import ClassVar
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import clsregistry
+from . import instrumentation
+from . import interfaces
+from . import mapperlib
+from ._orm_constructors import composite
+from ._orm_constructors import deferred
+from ._orm_constructors import mapped_column
+from ._orm_constructors import relationship
+from ._orm_constructors import synonym
+from .attributes import InstrumentedAttribute
+from .base import _inspect_mapped_class
+from .base import _is_mapped_class
+from .base import Mapped
+from .base import ORMDescriptor
+from .decl_base import _add_attribute
+from .decl_base import _as_declarative
+from .decl_base import _ClassScanMapperConfig
+from .decl_base import _declarative_constructor
+from .decl_base import _DeferredMapperConfig
+from .decl_base import _del_attribute
+from .decl_base import _mapper
+from .descriptor_props import Composite
+from .descriptor_props import Synonym
+from .descriptor_props import Synonym as _orm_synonym
+from .mapper import Mapper
+from .properties import MappedColumn
+from .relationships import RelationshipProperty
+from .state import InstanceState
+from .. import exc
+from .. import inspection
+from .. import util
+from ..sql import sqltypes
+from ..sql.base import _NoArg
+from ..sql.elements import SQLCoreOperations
+from ..sql.schema import MetaData
+from ..sql.selectable import FromClause
+from ..util import hybridmethod
+from ..util import hybridproperty
+from ..util import typing as compat_typing
+from ..util import warn_deprecated
+from ..util.typing import CallableReference
+from ..util.typing import de_optionalize_union_types
+from ..util.typing import flatten_newtype
+from ..util.typing import is_generic
+from ..util.typing import is_literal
+from ..util.typing import is_newtype
+from ..util.typing import is_pep695
+from ..util.typing import Literal
+from ..util.typing import LITERAL_TYPES
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from ._typing import _O
+    from ._typing import _RegistryType
+    from .decl_base import _DataclassArguments
+    from .instrumentation import ClassManager
+    from .interfaces import MapperProperty
+    from .state import InstanceState  # noqa
+    from ..sql._typing import _TypeEngineArgument
+    from ..sql.type_api import _MatchedOnType
+
+_T = TypeVar("_T", bound=Any)
+
+_TT = TypeVar("_TT", bound=Any)
+
+# it's not clear how to have Annotated, Union objects etc. as keys here
+# from a typing perspective so just leave it open ended for now
+_TypeAnnotationMapType = Mapping[Any, "_TypeEngineArgument[Any]"]
+_MutableTypeAnnotationMapType = Dict[Any, "_TypeEngineArgument[Any]"]
+
+_DeclaredAttrDecorated = Callable[
+    ..., Union[Mapped[_T], ORMDescriptor[_T], SQLCoreOperations[_T]]
+]
+
+
+def has_inherited_table(cls: Type[_O]) -> bool:
+    """Given a class, return True if any of the classes it inherits from has a
+    mapped table, otherwise return False.
+
+    This is used in declarative mixins to build attributes that behave
+    differently for the base class vs. a subclass in an inheritance
+    hierarchy.
+
+    .. seealso::
+
+        :ref:`decl_mixin_inheritance`
+
+    """
+    for class_ in cls.__mro__[1:]:
+        if getattr(class_, "__table__", None) is not None:
+            return True
+    return False
+
+
+class _DynamicAttributesType(type):
+    def __setattr__(cls, key: str, value: Any) -> None:
+        if "__mapper__" in cls.__dict__:
+            _add_attribute(cls, key, value)
+        else:
+            type.__setattr__(cls, key, value)
+
+    def __delattr__(cls, key: str) -> None:
+        if "__mapper__" in cls.__dict__:
+            _del_attribute(cls, key)
+        else:
+            type.__delattr__(cls, key)
+
+
+class DeclarativeAttributeIntercept(
+    _DynamicAttributesType,
+    # Inspectable is used only by the mypy plugin
+    inspection.Inspectable[Mapper[Any]],
+):
+    """Metaclass that may be used in conjunction with the
+    :class:`_orm.DeclarativeBase` class to support addition of class
+    attributes dynamically.
+
+    """
+
+
+@compat_typing.dataclass_transform(
+    field_specifiers=(
+        MappedColumn,
+        RelationshipProperty,
+        Composite,
+        Synonym,
+        mapped_column,
+        relationship,
+        composite,
+        synonym,
+        deferred,
+    ),
+)
+class DCTransformDeclarative(DeclarativeAttributeIntercept):
+    """metaclass that includes @dataclass_transforms"""
+
+
+class DeclarativeMeta(DeclarativeAttributeIntercept):
+    metadata: MetaData
+    registry: RegistryType
+
+    def __init__(
+        cls, classname: Any, bases: Any, dict_: Any, **kw: Any
+    ) -> None:
+        # use cls.__dict__, which can be modified by an
+        # __init_subclass__() method (#7900)
+        dict_ = cls.__dict__
+
+        # early-consume registry from the initial declarative base,
+        # assign privately to not conflict with subclass attributes named
+        # "registry"
+        reg = getattr(cls, "_sa_registry", None)
+        if reg is None:
+            reg = dict_.get("registry", None)
+            if not isinstance(reg, registry):
+                raise exc.InvalidRequestError(
+                    "Declarative base class has no 'registry' attribute, "
+                    "or registry is not a sqlalchemy.orm.registry() object"
+                )
+            else:
+                cls._sa_registry = reg
+
+        if not cls.__dict__.get("__abstract__", False):
+            _as_declarative(reg, cls, dict_)
+        type.__init__(cls, classname, bases, dict_)
+
+
+def synonym_for(
+    name: str, map_column: bool = False
+) -> Callable[[Callable[..., Any]], Synonym[Any]]:
+    """Decorator that produces an :func:`_orm.synonym`
+    attribute in conjunction with a Python descriptor.
+
+    The function being decorated is passed to :func:`_orm.synonym` as the
+    :paramref:`.orm.synonym.descriptor` parameter::
+
+        class MyClass(Base):
+            __tablename__ = "my_table"
+
+            id = Column(Integer, primary_key=True)
+            _job_status = Column("job_status", String(50))
+
+            @synonym_for("job_status")
+            @property
+            def job_status(self):
+                return "Status: %s" % self._job_status
+
+    The :ref:`hybrid properties <mapper_hybrids>` feature of SQLAlchemy
+    is typically preferred instead of synonyms, which is a more legacy
+    feature.
+
+    .. seealso::
+
+        :ref:`synonyms` - Overview of synonyms
+
+        :func:`_orm.synonym` - the mapper-level function
+
+        :ref:`mapper_hybrids` - The Hybrid Attribute extension provides an
+        updated approach to augmenting attribute behavior more flexibly than
+        can be achieved with synonyms.
+
+    """
+
+    def decorate(fn: Callable[..., Any]) -> Synonym[Any]:
+        return _orm_synonym(name, map_column=map_column, descriptor=fn)
+
+    return decorate
+
+
+class _declared_attr_common:
+    def __init__(
+        self,
+        fn: Callable[..., Any],
+        cascading: bool = False,
+        quiet: bool = False,
+    ):
+        # suppport
+        # @declared_attr
+        # @classmethod
+        # def foo(cls) -> Mapped[thing]:
+        #    ...
+        # which seems to help typing tools interpret the fn as a classmethod
+        # for situations where needed
+        if isinstance(fn, classmethod):
+            fn = fn.__func__
+
+        self.fget = fn
+        self._cascading = cascading
+        self._quiet = quiet
+        self.__doc__ = fn.__doc__
+
+    def _collect_return_annotation(self) -> Optional[Type[Any]]:
+        return util.get_annotations(self.fget).get("return")
+
+    def __get__(self, instance: Optional[object], owner: Any) -> Any:
+        # the declared_attr needs to make use of a cache that exists
+        # for the span of the declarative scan_attributes() phase.
+        # to achieve this we look at the class manager that's configured.
+
+        # note this method should not be called outside of the declarative
+        # setup phase
+
+        cls = owner
+        manager = attributes.opt_manager_of_class(cls)
+        if manager is None:
+            if not re.match(r"^__.+__$", self.fget.__name__):
+                # if there is no manager at all, then this class hasn't been
+                # run through declarative or mapper() at all, emit a warning.
+                util.warn(
+                    "Unmanaged access of declarative attribute %s from "
+                    "non-mapped class %s" % (self.fget.__name__, cls.__name__)
+                )
+            return self.fget(cls)
+        elif manager.is_mapped:
+            # the class is mapped, which means we're outside of the declarative
+            # scan setup, just run the function.
+            return self.fget(cls)
+
+        # here, we are inside of the declarative scan.  use the registry
+        # that is tracking the values of these attributes.
+        declarative_scan = manager.declarative_scan()
+
+        # assert that we are in fact in the declarative scan
+        assert declarative_scan is not None
+
+        reg = declarative_scan.declared_attr_reg
+
+        if self in reg:
+            return reg[self]
+        else:
+            reg[self] = obj = self.fget(cls)
+            return obj
+
+
+class _declared_directive(_declared_attr_common, Generic[_T]):
+    # see mapping_api.rst for docstring
+
+    if typing.TYPE_CHECKING:
+
+        def __init__(
+            self,
+            fn: Callable[..., _T],
+            cascading: bool = False,
+        ): ...
+
+        def __get__(self, instance: Optional[object], owner: Any) -> _T: ...
+
+        def __set__(self, instance: Any, value: Any) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
+
+        def __call__(self, fn: Callable[..., _TT]) -> _declared_directive[_TT]:
+            # extensive fooling of mypy underway...
+            ...
+
+
+class declared_attr(interfaces._MappedAttribute[_T], _declared_attr_common):
+    """Mark a class-level method as representing the definition of
+    a mapped property or Declarative directive.
+
+    :class:`_orm.declared_attr` is typically applied as a decorator to a class
+    level method, turning the attribute into a scalar-like property that can be
+    invoked from the uninstantiated class. The Declarative mapping process
+    looks for these :class:`_orm.declared_attr` callables as it scans classes,
+    and assumes any attribute marked with :class:`_orm.declared_attr` will be a
+    callable that will produce an object specific to the Declarative mapping or
+    table configuration.
+
+    :class:`_orm.declared_attr` is usually applicable to
+    :ref:`mixins <orm_mixins_toplevel>`, to define relationships that are to be
+    applied to different implementors of the class. It may also be used to
+    define dynamically generated column expressions and other Declarative
+    attributes.
+
+    Example::
+
+        class ProvidesUserMixin:
+            "A mixin that adds a 'user' relationship to classes."
+
+            user_id: Mapped[int] = mapped_column(ForeignKey("user_table.id"))
+
+            @declared_attr
+            def user(cls) -> Mapped["User"]:
+                return relationship("User")
+
+    When used with Declarative directives such as ``__tablename__``, the
+    :meth:`_orm.declared_attr.directive` modifier may be used which indicates
+    to :pep:`484` typing tools that the given method is not dealing with
+    :class:`_orm.Mapped` attributes::
+
+        class CreateTableName:
+            @declared_attr.directive
+            def __tablename__(cls) -> str:
+                return cls.__name__.lower()
+
+    :class:`_orm.declared_attr` can also be applied directly to mapped
+    classes, to allow for attributes that dynamically configure themselves
+    on subclasses when using mapped inheritance schemes.   Below
+    illustrates :class:`_orm.declared_attr` to create a dynamic scheme
+    for generating the :paramref:`_orm.Mapper.polymorphic_identity` parameter
+    for subclasses::
+
+        class Employee(Base):
+            __tablename__ = "employee"
+
+            id: Mapped[int] = mapped_column(primary_key=True)
+            type: Mapped[str] = mapped_column(String(50))
+
+            @declared_attr.directive
+            def __mapper_args__(cls) -> Dict[str, Any]:
+                if cls.__name__ == "Employee":
+                    return {
+                        "polymorphic_on": cls.type,
+                        "polymorphic_identity": "Employee",
+                    }
+                else:
+                    return {"polymorphic_identity": cls.__name__}
+
+
+        class Engineer(Employee):
+            pass
+
+    :class:`_orm.declared_attr` supports decorating functions that are
+    explicitly decorated with ``@classmethod``. This is never necessary from a
+    runtime perspective, however may be needed in order to support :pep:`484`
+    typing tools that don't otherwise recognize the decorated function as
+    having class-level behaviors for the ``cls`` parameter::
+
+        class SomethingMixin:
+            x: Mapped[int]
+            y: Mapped[int]
+
+            @declared_attr
+            @classmethod
+            def x_plus_y(cls) -> Mapped[int]:
+                return column_property(cls.x + cls.y)
+
+    .. versionadded:: 2.0 - :class:`_orm.declared_attr` can accommodate a
+       function decorated with ``@classmethod`` to help with :pep:`484`
+       integration where needed.
+
+
+    .. seealso::
+
+        :ref:`orm_mixins_toplevel` - Declarative Mixin documentation with
+        background on use patterns for :class:`_orm.declared_attr`.
+
+    """  # noqa: E501
+
+    if typing.TYPE_CHECKING:
+
+        def __init__(
+            self,
+            fn: _DeclaredAttrDecorated[_T],
+            cascading: bool = False,
+        ): ...
+
+        def __set__(self, instance: Any, value: Any) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
+
+        # this is the Mapped[] API where at class descriptor get time we want
+        # the type checker to see InstrumentedAttribute[_T].   However the
+        # callable function prior to mapping in fact calls the given
+        # declarative function that does not return InstrumentedAttribute
+        @overload
+        def __get__(
+            self, instance: None, owner: Any
+        ) -> InstrumentedAttribute[_T]: ...
+
+        @overload
+        def __get__(self, instance: object, owner: Any) -> _T: ...
+
+        def __get__(
+            self, instance: Optional[object], owner: Any
+        ) -> Union[InstrumentedAttribute[_T], _T]: ...
+
+    @hybridmethod
+    def _stateful(cls, **kw: Any) -> _stateful_declared_attr[_T]:
+        return _stateful_declared_attr(**kw)
+
+    @hybridproperty
+    def directive(cls) -> _declared_directive[Any]:
+        # see mapping_api.rst for docstring
+        return _declared_directive  # type: ignore
+
+    @hybridproperty
+    def cascading(cls) -> _stateful_declared_attr[_T]:
+        # see mapping_api.rst for docstring
+        return cls._stateful(cascading=True)
+
+
+class _stateful_declared_attr(declared_attr[_T]):
+    kw: Dict[str, Any]
+
+    def __init__(self, **kw: Any):
+        self.kw = kw
+
+    @hybridmethod
+    def _stateful(self, **kw: Any) -> _stateful_declared_attr[_T]:
+        new_kw = self.kw.copy()
+        new_kw.update(kw)
+        return _stateful_declared_attr(**new_kw)
+
+    def __call__(self, fn: _DeclaredAttrDecorated[_T]) -> declared_attr[_T]:
+        return declared_attr(fn, **self.kw)
+
+
+def declarative_mixin(cls: Type[_T]) -> Type[_T]:
+    """Mark a class as providing the feature of "declarative mixin".
+
+    E.g.::
+
+        from sqlalchemy.orm import declared_attr
+        from sqlalchemy.orm import declarative_mixin
+
+
+        @declarative_mixin
+        class MyMixin:
+
+            @declared_attr
+            def __tablename__(cls):
+                return cls.__name__.lower()
+
+            __table_args__ = {"mysql_engine": "InnoDB"}
+            __mapper_args__ = {"always_refresh": True}
+
+            id = Column(Integer, primary_key=True)
+
+
+        class MyModel(MyMixin, Base):
+            name = Column(String(1000))
+
+    The :func:`_orm.declarative_mixin` decorator currently does not modify
+    the given class in any way; it's current purpose is strictly to assist
+    the :ref:`Mypy plugin <mypy_toplevel>` in being able to identify
+    SQLAlchemy declarative mixin classes when no other context is present.
+
+    .. versionadded:: 1.4.6
+
+    .. seealso::
+
+        :ref:`orm_mixins_toplevel`
+
+        :ref:`mypy_declarative_mixins` - in the
+        :ref:`Mypy plugin documentation <mypy_toplevel>`
+
+    """  # noqa: E501
+
+    return cls
+
+
+def _setup_declarative_base(cls: Type[Any]) -> None:
+    if "metadata" in cls.__dict__:
+        metadata = cls.__dict__["metadata"]
+    else:
+        metadata = None
+
+    if "type_annotation_map" in cls.__dict__:
+        type_annotation_map = cls.__dict__["type_annotation_map"]
+    else:
+        type_annotation_map = None
+
+    reg = cls.__dict__.get("registry", None)
+    if reg is not None:
+        if not isinstance(reg, registry):
+            raise exc.InvalidRequestError(
+                "Declarative base class has a 'registry' attribute that is "
+                "not an instance of sqlalchemy.orm.registry()"
+            )
+        elif type_annotation_map is not None:
+            raise exc.InvalidRequestError(
+                "Declarative base class has both a 'registry' attribute and a "
+                "type_annotation_map entry.  Per-base type_annotation_maps "
+                "are not supported.  Please apply the type_annotation_map "
+                "to this registry directly."
+            )
+
+    else:
+        reg = registry(
+            metadata=metadata, type_annotation_map=type_annotation_map
+        )
+        cls.registry = reg
+
+    cls._sa_registry = reg
+
+    if "metadata" not in cls.__dict__:
+        cls.metadata = cls.registry.metadata
+
+    if getattr(cls, "__init__", object.__init__) is object.__init__:
+        cls.__init__ = cls.registry.constructor
+
+
+class MappedAsDataclass(metaclass=DCTransformDeclarative):
+    """Mixin class to indicate when mapping this class, also convert it to be
+    a dataclass.
+
+    .. seealso::
+
+        :ref:`orm_declarative_native_dataclasses` - complete background
+        on SQLAlchemy native dataclass mapping
+
+    .. versionadded:: 2.0
+
+    """
+
+    def __init_subclass__(
+        cls,
+        init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+        eq: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        order: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        unsafe_hash: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        match_args: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        dataclass_callable: Union[
+            _NoArg, Callable[..., Type[Any]]
+        ] = _NoArg.NO_ARG,
+        **kw: Any,
+    ) -> None:
+        apply_dc_transforms: _DataclassArguments = {
+            "init": init,
+            "repr": repr,
+            "eq": eq,
+            "order": order,
+            "unsafe_hash": unsafe_hash,
+            "match_args": match_args,
+            "kw_only": kw_only,
+            "dataclass_callable": dataclass_callable,
+        }
+
+        current_transforms: _DataclassArguments
+
+        if hasattr(cls, "_sa_apply_dc_transforms"):
+            current = cls._sa_apply_dc_transforms
+
+            _ClassScanMapperConfig._assert_dc_arguments(current)
+
+            cls._sa_apply_dc_transforms = current_transforms = {  # type: ignore  # noqa: E501
+                k: current.get(k, _NoArg.NO_ARG) if v is _NoArg.NO_ARG else v
+                for k, v in apply_dc_transforms.items()
+            }
+        else:
+            cls._sa_apply_dc_transforms = current_transforms = (
+                apply_dc_transforms
+            )
+
+        super().__init_subclass__(**kw)
+
+        if not _is_mapped_class(cls):
+            new_anno = (
+                _ClassScanMapperConfig._update_annotations_for_non_mapped_class
+            )(cls)
+            _ClassScanMapperConfig._apply_dataclasses_to_any_class(
+                current_transforms, cls, new_anno
+            )
+
+
+class DeclarativeBase(
+    # Inspectable is used only by the mypy plugin
+    inspection.Inspectable[InstanceState[Any]],
+    metaclass=DeclarativeAttributeIntercept,
+):
+    """Base class used for declarative class definitions.
+
+    The :class:`_orm.DeclarativeBase` allows for the creation of new
+    declarative bases in such a way that is compatible with type checkers::
+
+
+        from sqlalchemy.orm import DeclarativeBase
+
+
+        class Base(DeclarativeBase):
+            pass
+
+    The above ``Base`` class is now usable as the base for new declarative
+    mappings.  The superclass makes use of the ``__init_subclass__()``
+    method to set up new classes and metaclasses aren't used.
+
+    When first used, the :class:`_orm.DeclarativeBase` class instantiates a new
+    :class:`_orm.registry` to be used with the base, assuming one was not
+    provided explicitly. The :class:`_orm.DeclarativeBase` class supports
+    class-level attributes which act as parameters for the construction of this
+    registry; such as to indicate a specific :class:`_schema.MetaData`
+    collection as well as a specific value for
+    :paramref:`_orm.registry.type_annotation_map`::
+
+        from typing_extensions import Annotated
+
+        from sqlalchemy import BigInteger
+        from sqlalchemy import MetaData
+        from sqlalchemy import String
+        from sqlalchemy.orm import DeclarativeBase
+
+        bigint = Annotated[int, "bigint"]
+        my_metadata = MetaData()
+
+
+        class Base(DeclarativeBase):
+            metadata = my_metadata
+            type_annotation_map = {
+                str: String().with_variant(String(255), "mysql", "mariadb"),
+                bigint: BigInteger(),
+            }
+
+    Class-level attributes which may be specified include:
+
+    :param metadata: optional :class:`_schema.MetaData` collection.
+     If a :class:`_orm.registry` is constructed automatically, this
+     :class:`_schema.MetaData` collection will be used to construct it.
+     Otherwise, the local :class:`_schema.MetaData` collection will supercede
+     that used by an existing :class:`_orm.registry` passed using the
+     :paramref:`_orm.DeclarativeBase.registry` parameter.
+    :param type_annotation_map: optional type annotation map that will be
+     passed to the :class:`_orm.registry` as
+     :paramref:`_orm.registry.type_annotation_map`.
+    :param registry: supply a pre-existing :class:`_orm.registry` directly.
+
+    .. versionadded:: 2.0  Added :class:`.DeclarativeBase`, so that declarative
+       base classes may be constructed in such a way that is also recognized
+       by :pep:`484` type checkers.   As a result, :class:`.DeclarativeBase`
+       and other subclassing-oriented APIs should be seen as
+       superseding previous "class returned by a function" APIs, namely
+       :func:`_orm.declarative_base` and :meth:`_orm.registry.generate_base`,
+       where the base class returned cannot be recognized by type checkers
+       without using plugins.
+
+    **__init__ behavior**
+
+    In a plain Python class, the base-most ``__init__()`` method in the class
+    hierarchy is ``object.__init__()``, which accepts no arguments. However,
+    when the :class:`_orm.DeclarativeBase` subclass is first declared, the
+    class is given an ``__init__()`` method that links to the
+    :paramref:`_orm.registry.constructor` constructor function, if no
+    ``__init__()`` method is already present; this is the usual declarative
+    constructor that will assign keyword arguments as attributes on the
+    instance, assuming those attributes are established at the class level
+    (i.e. are mapped, or are linked to a descriptor). This constructor is
+    **never accessed by a mapped class without being called explicitly via
+    super()**, as mapped classes are themselves given an ``__init__()`` method
+    directly which calls :paramref:`_orm.registry.constructor`, so in the
+    default case works independently of what the base-most ``__init__()``
+    method does.
+
+    .. versionchanged:: 2.0.1  :class:`_orm.DeclarativeBase` has a default
+       constructor that links to :paramref:`_orm.registry.constructor` by
+       default, so that calls to ``super().__init__()`` can access this
+       constructor. Previously, due to an implementation mistake, this default
+       constructor was missing, and calling ``super().__init__()`` would invoke
+       ``object.__init__()``.
+
+    The :class:`_orm.DeclarativeBase` subclass may also declare an explicit
+    ``__init__()`` method which will replace the use of the
+    :paramref:`_orm.registry.constructor` function at this level::
+
+        class Base(DeclarativeBase):
+            def __init__(self, id=None):
+                self.id = id
+
+    Mapped classes still will not invoke this constructor implicitly; it
+    remains only accessible by calling ``super().__init__()``::
+
+        class MyClass(Base):
+            def __init__(self, id=None, name=None):
+                self.name = name
+                super().__init__(id=id)
+
+    Note that this is a different behavior from what functions like the legacy
+    :func:`_orm.declarative_base` would do; the base created by those functions
+    would always install :paramref:`_orm.registry.constructor` for
+    ``__init__()``.
+
+
+    """
+
+    if typing.TYPE_CHECKING:
+
+        def _sa_inspect_type(self) -> Mapper[Self]: ...
+
+        def _sa_inspect_instance(self) -> InstanceState[Self]: ...
+
+        _sa_registry: ClassVar[_RegistryType]
+
+        registry: ClassVar[_RegistryType]
+        """Refers to the :class:`_orm.registry` in use where new
+        :class:`_orm.Mapper` objects will be associated."""
+
+        metadata: ClassVar[MetaData]
+        """Refers to the :class:`_schema.MetaData` collection that will be used
+        for new :class:`_schema.Table` objects.
+
+        .. seealso::
+
+            :ref:`orm_declarative_metadata`
+
+        """
+
+        __name__: ClassVar[str]
+
+        # this ideally should be Mapper[Self], but mypy as of 1.4.1 does not
+        # like it, and breaks the declared_attr_one test. Pyright/pylance is
+        # ok with it.
+        __mapper__: ClassVar[Mapper[Any]]
+        """The :class:`_orm.Mapper` object to which a particular class is
+        mapped.
+
+        May also be acquired using :func:`_sa.inspect`, e.g.
+        ``inspect(klass)``.
+
+        """
+
+        __table__: ClassVar[FromClause]
+        """The :class:`_sql.FromClause` to which a particular subclass is
+        mapped.
+
+        This is usually an instance of :class:`_schema.Table` but may also
+        refer to other kinds of :class:`_sql.FromClause` such as
+        :class:`_sql.Subquery`, depending on how the class is mapped.
+
+        .. seealso::
+
+            :ref:`orm_declarative_metadata`
+
+        """
+
+        # pyright/pylance do not consider a classmethod a ClassVar so use Any
+        # https://github.com/microsoft/pylance-release/issues/3484
+        __tablename__: Any
+        """String name to assign to the generated
+        :class:`_schema.Table` object, if not specified directly via
+        :attr:`_orm.DeclarativeBase.__table__`.
+
+        .. seealso::
+
+            :ref:`orm_declarative_table`
+
+        """
+
+        __mapper_args__: Any
+        """Dictionary of arguments which will be passed to the
+        :class:`_orm.Mapper` constructor.
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapper_options`
+
+        """
+
+        __table_args__: Any
+        """A dictionary or tuple of arguments that will be passed to the
+        :class:`_schema.Table` constructor.  See
+        :ref:`orm_declarative_table_configuration`
+        for background on the specific structure of this collection.
+
+        .. seealso::
+
+            :ref:`orm_declarative_table_configuration`
+
+        """
+
+        def __init__(self, **kw: Any): ...
+
+    def __init_subclass__(cls, **kw: Any) -> None:
+        if DeclarativeBase in cls.__bases__:
+            _check_not_declarative(cls, DeclarativeBase)
+            _setup_declarative_base(cls)
+        else:
+            _as_declarative(cls._sa_registry, cls, cls.__dict__)
+        super().__init_subclass__(**kw)
+
+
+def _check_not_declarative(cls: Type[Any], base: Type[Any]) -> None:
+    cls_dict = cls.__dict__
+    if (
+        "__table__" in cls_dict
+        and not (
+            callable(cls_dict["__table__"])
+            or hasattr(cls_dict["__table__"], "__get__")
+        )
+    ) or isinstance(cls_dict.get("__tablename__", None), str):
+        raise exc.InvalidRequestError(
+            f"Cannot use {base.__name__!r} directly as a declarative base "
+            "class. Create a Base by creating a subclass of it."
+        )
+
+
+class DeclarativeBaseNoMeta(
+    # Inspectable is used only by the mypy plugin
+    inspection.Inspectable[InstanceState[Any]]
+):
+    """Same as :class:`_orm.DeclarativeBase`, but does not use a metaclass
+    to intercept new attributes.
+
+    The :class:`_orm.DeclarativeBaseNoMeta` base may be used when use of
+    custom metaclasses is desirable.
+
+    .. versionadded:: 2.0
+
+
+    """
+
+    _sa_registry: ClassVar[_RegistryType]
+
+    registry: ClassVar[_RegistryType]
+    """Refers to the :class:`_orm.registry` in use where new
+    :class:`_orm.Mapper` objects will be associated."""
+
+    metadata: ClassVar[MetaData]
+    """Refers to the :class:`_schema.MetaData` collection that will be used
+    for new :class:`_schema.Table` objects.
+
+    .. seealso::
+
+        :ref:`orm_declarative_metadata`
+
+    """
+
+    # this ideally should be Mapper[Self], but mypy as of 1.4.1 does not
+    # like it, and breaks the declared_attr_one test. Pyright/pylance is
+    # ok with it.
+    __mapper__: ClassVar[Mapper[Any]]
+    """The :class:`_orm.Mapper` object to which a particular class is
+    mapped.
+
+    May also be acquired using :func:`_sa.inspect`, e.g.
+    ``inspect(klass)``.
+
+    """
+
+    __table__: Optional[FromClause]
+    """The :class:`_sql.FromClause` to which a particular subclass is
+    mapped.
+
+    This is usually an instance of :class:`_schema.Table` but may also
+    refer to other kinds of :class:`_sql.FromClause` such as
+    :class:`_sql.Subquery`, depending on how the class is mapped.
+
+    .. seealso::
+
+        :ref:`orm_declarative_metadata`
+
+    """
+
+    if typing.TYPE_CHECKING:
+
+        def _sa_inspect_type(self) -> Mapper[Self]: ...
+
+        def _sa_inspect_instance(self) -> InstanceState[Self]: ...
+
+        __tablename__: Any
+        """String name to assign to the generated
+        :class:`_schema.Table` object, if not specified directly via
+        :attr:`_orm.DeclarativeBase.__table__`.
+
+        .. seealso::
+
+            :ref:`orm_declarative_table`
+
+        """
+
+        __mapper_args__: Any
+        """Dictionary of arguments which will be passed to the
+        :class:`_orm.Mapper` constructor.
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapper_options`
+
+        """
+
+        __table_args__: Any
+        """A dictionary or tuple of arguments that will be passed to the
+        :class:`_schema.Table` constructor.  See
+        :ref:`orm_declarative_table_configuration`
+        for background on the specific structure of this collection.
+
+        .. seealso::
+
+            :ref:`orm_declarative_table_configuration`
+
+        """
+
+        def __init__(self, **kw: Any): ...
+
+    def __init_subclass__(cls, **kw: Any) -> None:
+        if DeclarativeBaseNoMeta in cls.__bases__:
+            _check_not_declarative(cls, DeclarativeBaseNoMeta)
+            _setup_declarative_base(cls)
+        else:
+            _as_declarative(cls._sa_registry, cls, cls.__dict__)
+        super().__init_subclass__(**kw)
+
+
+def add_mapped_attribute(
+    target: Type[_O], key: str, attr: MapperProperty[Any]
+) -> None:
+    """Add a new mapped attribute to an ORM mapped class.
+
+    E.g.::
+
+        add_mapped_attribute(User, "addresses", relationship(Address))
+
+    This may be used for ORM mappings that aren't using a declarative
+    metaclass that intercepts attribute set operations.
+
+    .. versionadded:: 2.0
+
+
+    """
+    _add_attribute(target, key, attr)
+
+
+def declarative_base(
+    *,
+    metadata: Optional[MetaData] = None,
+    mapper: Optional[Callable[..., Mapper[Any]]] = None,
+    cls: Type[Any] = object,
+    name: str = "Base",
+    class_registry: Optional[clsregistry._ClsRegistryType] = None,
+    type_annotation_map: Optional[_TypeAnnotationMapType] = None,
+    constructor: Callable[..., None] = _declarative_constructor,
+    metaclass: Type[Any] = DeclarativeMeta,
+) -> Any:
+    r"""Construct a base class for declarative class definitions.
+
+    The new base class will be given a metaclass that produces
+    appropriate :class:`~sqlalchemy.schema.Table` objects and makes
+    the appropriate :class:`_orm.Mapper` calls based on the
+    information provided declaratively in the class and any subclasses
+    of the class.
+
+    .. versionchanged:: 2.0 Note that the :func:`_orm.declarative_base`
+       function is superseded by the new :class:`_orm.DeclarativeBase` class,
+       which generates a new "base" class using subclassing, rather than
+       return value of a function.  This allows an approach that is compatible
+       with :pep:`484` typing tools.
+
+    The :func:`_orm.declarative_base` function is a shorthand version
+    of using the :meth:`_orm.registry.generate_base`
+    method.  That is, the following::
+
+        from sqlalchemy.orm import declarative_base
+
+        Base = declarative_base()
+
+    Is equivalent to::
+
+        from sqlalchemy.orm import registry
+
+        mapper_registry = registry()
+        Base = mapper_registry.generate_base()
+
+    See the docstring for :class:`_orm.registry`
+    and :meth:`_orm.registry.generate_base`
+    for more details.
+
+    .. versionchanged:: 1.4  The :func:`_orm.declarative_base`
+       function is now a specialization of the more generic
+       :class:`_orm.registry` class.  The function also moves to the
+       ``sqlalchemy.orm`` package from the ``declarative.ext`` package.
+
+
+    :param metadata:
+      An optional :class:`~sqlalchemy.schema.MetaData` instance.  All
+      :class:`~sqlalchemy.schema.Table` objects implicitly declared by
+      subclasses of the base will share this MetaData.  A MetaData instance
+      will be created if none is provided.  The
+      :class:`~sqlalchemy.schema.MetaData` instance will be available via the
+      ``metadata`` attribute of the generated declarative base class.
+
+    :param mapper:
+      An optional callable, defaults to :class:`_orm.Mapper`. Will
+      be used to map subclasses to their Tables.
+
+    :param cls:
+      Defaults to :class:`object`. A type to use as the base for the generated
+      declarative base class. May be a class or tuple of classes.
+
+    :param name:
+      Defaults to ``Base``.  The display name for the generated
+      class.  Customizing this is not required, but can improve clarity in
+      tracebacks and debugging.
+
+    :param constructor:
+      Specify the implementation for the ``__init__`` function on a mapped
+      class that has no ``__init__`` of its own.  Defaults to an
+      implementation that assigns \**kwargs for declared
+      fields and relationships to an instance.  If ``None`` is supplied,
+      no __init__ will be provided and construction will fall back to
+      cls.__init__ by way of the normal Python semantics.
+
+    :param class_registry: optional dictionary that will serve as the
+      registry of class names-> mapped classes when string names
+      are used to identify classes inside of :func:`_orm.relationship`
+      and others.  Allows two or more declarative base classes
+      to share the same registry of class names for simplified
+      inter-base relationships.
+
+    :param type_annotation_map: optional dictionary of Python types to
+        SQLAlchemy :class:`_types.TypeEngine` classes or instances.  This
+        is used exclusively by the :class:`_orm.MappedColumn` construct
+        to produce column types based on annotations within the
+        :class:`_orm.Mapped` type.
+
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapped_column_type_map`
+
+    :param metaclass:
+      Defaults to :class:`.DeclarativeMeta`.  A metaclass or __metaclass__
+      compatible callable to use as the meta type of the generated
+      declarative base class.
+
+    .. seealso::
+
+        :class:`_orm.registry`
+
+    """
+
+    return registry(
+        metadata=metadata,
+        class_registry=class_registry,
+        constructor=constructor,
+        type_annotation_map=type_annotation_map,
+    ).generate_base(
+        mapper=mapper,
+        cls=cls,
+        name=name,
+        metaclass=metaclass,
+    )
+
+
+class registry:
+    """Generalized registry for mapping classes.
+
+    The :class:`_orm.registry` serves as the basis for maintaining a collection
+    of mappings, and provides configurational hooks used to map classes.
+
+    The three general kinds of mappings supported are Declarative Base,
+    Declarative Decorator, and Imperative Mapping.   All of these mapping
+    styles may be used interchangeably:
+
+    * :meth:`_orm.registry.generate_base` returns a new declarative base
+      class, and is the underlying implementation of the
+      :func:`_orm.declarative_base` function.
+
+    * :meth:`_orm.registry.mapped` provides a class decorator that will
+      apply declarative mapping to a class without the use of a declarative
+      base class.
+
+    * :meth:`_orm.registry.map_imperatively` will produce a
+      :class:`_orm.Mapper` for a class without scanning the class for
+      declarative class attributes. This method suits the use case historically
+      provided by the ``sqlalchemy.orm.mapper()`` classical mapping function,
+      which is removed as of SQLAlchemy 2.0.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`orm_mapping_classes_toplevel` - overview of class mapping
+        styles.
+
+    """
+
+    _class_registry: clsregistry._ClsRegistryType
+    _managers: weakref.WeakKeyDictionary[ClassManager[Any], Literal[True]]
+    _non_primary_mappers: weakref.WeakKeyDictionary[Mapper[Any], Literal[True]]
+    metadata: MetaData
+    constructor: CallableReference[Callable[..., None]]
+    type_annotation_map: _MutableTypeAnnotationMapType
+    _dependents: Set[_RegistryType]
+    _dependencies: Set[_RegistryType]
+    _new_mappers: bool
+
+    def __init__(
+        self,
+        *,
+        metadata: Optional[MetaData] = None,
+        class_registry: Optional[clsregistry._ClsRegistryType] = None,
+        type_annotation_map: Optional[_TypeAnnotationMapType] = None,
+        constructor: Callable[..., None] = _declarative_constructor,
+    ):
+        r"""Construct a new :class:`_orm.registry`
+
+        :param metadata:
+          An optional :class:`_schema.MetaData` instance.  All
+          :class:`_schema.Table` objects generated using declarative
+          table mapping will make use of this :class:`_schema.MetaData`
+          collection.  If this argument is left at its default of ``None``,
+          a blank :class:`_schema.MetaData` collection is created.
+
+        :param constructor:
+          Specify the implementation for the ``__init__`` function on a mapped
+          class that has no ``__init__`` of its own.  Defaults to an
+          implementation that assigns \**kwargs for declared
+          fields and relationships to an instance.  If ``None`` is supplied,
+          no __init__ will be provided and construction will fall back to
+          cls.__init__ by way of the normal Python semantics.
+
+        :param class_registry: optional dictionary that will serve as the
+          registry of class names-> mapped classes when string names
+          are used to identify classes inside of :func:`_orm.relationship`
+          and others.  Allows two or more declarative base classes
+          to share the same registry of class names for simplified
+          inter-base relationships.
+
+        :param type_annotation_map: optional dictionary of Python types to
+          SQLAlchemy :class:`_types.TypeEngine` classes or instances.
+          The provided dict will update the default type mapping.  This
+          is used exclusively by the :class:`_orm.MappedColumn` construct
+          to produce column types based on annotations within the
+          :class:`_orm.Mapped` type.
+
+          .. versionadded:: 2.0
+
+          .. seealso::
+
+              :ref:`orm_declarative_mapped_column_type_map`
+
+
+        """
+        lcl_metadata = metadata or MetaData()
+
+        if class_registry is None:
+            class_registry = weakref.WeakValueDictionary()
+
+        self._class_registry = class_registry
+        self._managers = weakref.WeakKeyDictionary()
+        self._non_primary_mappers = weakref.WeakKeyDictionary()
+        self.metadata = lcl_metadata
+        self.constructor = constructor
+        self.type_annotation_map = {}
+        if type_annotation_map is not None:
+            self.update_type_annotation_map(type_annotation_map)
+        self._dependents = set()
+        self._dependencies = set()
+
+        self._new_mappers = False
+
+        with mapperlib._CONFIGURE_MUTEX:
+            mapperlib._mapper_registries[self] = True
+
+    def update_type_annotation_map(
+        self,
+        type_annotation_map: _TypeAnnotationMapType,
+    ) -> None:
+        """update the :paramref:`_orm.registry.type_annotation_map` with new
+        values."""
+
+        self.type_annotation_map.update(
+            {
+                de_optionalize_union_types(typ): sqltype
+                for typ, sqltype in type_annotation_map.items()
+            }
+        )
+
+    def _resolve_type(
+        self, python_type: _MatchedOnType, _do_fallbacks: bool = True
+    ) -> Optional[sqltypes.TypeEngine[Any]]:
+        python_type_type: Type[Any]
+        search: Iterable[Tuple[_MatchedOnType, Type[Any]]]
+
+        if is_generic(python_type):
+            if is_literal(python_type):
+                python_type_type = python_type  # type: ignore[assignment]
+
+                search = (
+                    (python_type, python_type_type),
+                    *((lt, python_type_type) for lt in LITERAL_TYPES),  # type: ignore[arg-type] # noqa: E501
+                )
+            else:
+                python_type_type = python_type.__origin__
+                search = ((python_type, python_type_type),)
+        elif isinstance(python_type, type):
+            python_type_type = python_type
+            search = ((pt, pt) for pt in python_type_type.__mro__)
+        else:
+            python_type_type = python_type  # type: ignore[assignment]
+            search = ((python_type, python_type_type),)
+
+        for pt, flattened in search:
+            # we search through full __mro__ for types.  however...
+            sql_type = self.type_annotation_map.get(pt)
+            if sql_type is None:
+                sql_type = sqltypes._type_map_get(pt)  # type: ignore  # noqa: E501
+
+            if sql_type is not None:
+                sql_type_inst = sqltypes.to_instance(sql_type)
+
+                # ... this additional step will reject most
+                # type -> supertype matches, such as if we had
+                # a MyInt(int) subclass.  note also we pass NewType()
+                # here directly; these always have to be in the
+                # type_annotation_map to be useful
+                resolved_sql_type = sql_type_inst._resolve_for_python_type(
+                    python_type_type,
+                    pt,
+                    flattened,
+                )
+                if resolved_sql_type is not None:
+                    return resolved_sql_type
+
+        # 2.0 fallbacks
+        if _do_fallbacks:
+            python_type_to_check: Any = None
+            kind = None
+            if is_pep695(python_type):
+                # NOTE: assume there aren't type alias types of new types.
+                python_type_to_check = python_type
+                while is_pep695(python_type_to_check):
+                    python_type_to_check = python_type_to_check.__value__
+                python_type_to_check = de_optionalize_union_types(
+                    python_type_to_check
+                )
+                kind = "TypeAliasType"
+            if is_newtype(python_type):
+                python_type_to_check = flatten_newtype(python_type)
+                kind = "NewType"
+
+            if python_type_to_check is not None:
+                res_after_fallback = self._resolve_type(
+                    python_type_to_check, False
+                )
+                if res_after_fallback is not None:
+                    assert kind is not None
+                    warn_deprecated(
+                        f"Matching the provided {kind} '{python_type}' on "
+                        "its resolved value without matching it in the "
+                        "type_annotation_map is deprecated; add this type to "
+                        "the type_annotation_map to allow it to match "
+                        "explicitly.",
+                        "2.0",
+                    )
+                    return res_after_fallback
+
+        return None
+
+    @property
+    def mappers(self) -> FrozenSet[Mapper[Any]]:
+        """read only collection of all :class:`_orm.Mapper` objects."""
+
+        return frozenset(manager.mapper for manager in self._managers).union(
+            self._non_primary_mappers
+        )
+
+    def _set_depends_on(self, registry: RegistryType) -> None:
+        if registry is self:
+            return
+        registry._dependents.add(self)
+        self._dependencies.add(registry)
+
+    def _flag_new_mapper(self, mapper: Mapper[Any]) -> None:
+        mapper._ready_for_configure = True
+        if self._new_mappers:
+            return
+
+        for reg in self._recurse_with_dependents({self}):
+            reg._new_mappers = True
+
+    @classmethod
+    def _recurse_with_dependents(
+        cls, registries: Set[RegistryType]
+    ) -> Iterator[RegistryType]:
+        todo = registries
+        done = set()
+        while todo:
+            reg = todo.pop()
+            done.add(reg)
+
+            # if yielding would remove dependents, make sure we have
+            # them before
+            todo.update(reg._dependents.difference(done))
+            yield reg
+
+            # if yielding would add dependents, make sure we have them
+            # after
+            todo.update(reg._dependents.difference(done))
+
+    @classmethod
+    def _recurse_with_dependencies(
+        cls, registries: Set[RegistryType]
+    ) -> Iterator[RegistryType]:
+        todo = registries
+        done = set()
+        while todo:
+            reg = todo.pop()
+            done.add(reg)
+
+            # if yielding would remove dependencies, make sure we have
+            # them before
+            todo.update(reg._dependencies.difference(done))
+
+            yield reg
+
+            # if yielding would remove dependencies, make sure we have
+            # them before
+            todo.update(reg._dependencies.difference(done))
+
+    def _mappers_to_configure(self) -> Iterator[Mapper[Any]]:
+        return itertools.chain(
+            (
+                manager.mapper
+                for manager in list(self._managers)
+                if manager.is_mapped
+                and not manager.mapper.configured
+                and manager.mapper._ready_for_configure
+            ),
+            (
+                npm
+                for npm in list(self._non_primary_mappers)
+                if not npm.configured and npm._ready_for_configure
+            ),
+        )
+
+    def _add_non_primary_mapper(self, np_mapper: Mapper[Any]) -> None:
+        self._non_primary_mappers[np_mapper] = True
+
+    def _dispose_cls(self, cls: Type[_O]) -> None:
+        clsregistry.remove_class(cls.__name__, cls, self._class_registry)
+
+    def _add_manager(self, manager: ClassManager[Any]) -> None:
+        self._managers[manager] = True
+        if manager.is_mapped:
+            raise exc.ArgumentError(
+                "Class '%s' already has a primary mapper defined. "
+                % manager.class_
+            )
+        assert manager.registry is None
+        manager.registry = self
+
+    def configure(self, cascade: bool = False) -> None:
+        """Configure all as-yet unconfigured mappers in this
+        :class:`_orm.registry`.
+
+        The configure step is used to reconcile and initialize the
+        :func:`_orm.relationship` linkages between mapped classes, as well as
+        to invoke configuration events such as the
+        :meth:`_orm.MapperEvents.before_configured` and
+        :meth:`_orm.MapperEvents.after_configured`, which may be used by ORM
+        extensions or user-defined extension hooks.
+
+        If one or more mappers in this registry contain
+        :func:`_orm.relationship` constructs that refer to mapped classes in
+        other registries, this registry is said to be *dependent* on those
+        registries. In order to configure those dependent registries
+        automatically, the :paramref:`_orm.registry.configure.cascade` flag
+        should be set to ``True``. Otherwise, if they are not configured, an
+        exception will be raised.  The rationale behind this behavior is to
+        allow an application to programmatically invoke configuration of
+        registries while controlling whether or not the process implicitly
+        reaches other registries.
+
+        As an alternative to invoking :meth:`_orm.registry.configure`, the ORM
+        function :func:`_orm.configure_mappers` function may be used to ensure
+        configuration is complete for all :class:`_orm.registry` objects in
+        memory. This is generally simpler to use and also predates the usage of
+        :class:`_orm.registry` objects overall. However, this function will
+        impact all mappings throughout the running Python process and may be
+        more memory/time consuming for an application that has many registries
+        in use for different purposes that may not be needed immediately.
+
+        .. seealso::
+
+            :func:`_orm.configure_mappers`
+
+
+        .. versionadded:: 1.4.0b2
+
+        """
+        mapperlib._configure_registries({self}, cascade=cascade)
+
+    def dispose(self, cascade: bool = False) -> None:
+        """Dispose of all mappers in this :class:`_orm.registry`.
+
+        After invocation, all the classes that were mapped within this registry
+        will no longer have class instrumentation associated with them. This
+        method is the per-:class:`_orm.registry` analogue to the
+        application-wide :func:`_orm.clear_mappers` function.
+
+        If this registry contains mappers that are dependencies of other
+        registries, typically via :func:`_orm.relationship` links, then those
+        registries must be disposed as well. When such registries exist in
+        relation to this one, their :meth:`_orm.registry.dispose` method will
+        also be called, if the :paramref:`_orm.registry.dispose.cascade` flag
+        is set to ``True``; otherwise, an error is raised if those registries
+        were not already disposed.
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :func:`_orm.clear_mappers`
+
+        """
+
+        mapperlib._dispose_registries({self}, cascade=cascade)
+
+    def _dispose_manager_and_mapper(self, manager: ClassManager[Any]) -> None:
+        if "mapper" in manager.__dict__:
+            mapper = manager.mapper
+
+            mapper._set_dispose_flags()
+
+        class_ = manager.class_
+        self._dispose_cls(class_)
+        instrumentation._instrumentation_factory.unregister(class_)
+
+    def generate_base(
+        self,
+        mapper: Optional[Callable[..., Mapper[Any]]] = None,
+        cls: Type[Any] = object,
+        name: str = "Base",
+        metaclass: Type[Any] = DeclarativeMeta,
+    ) -> Any:
+        """Generate a declarative base class.
+
+        Classes that inherit from the returned class object will be
+        automatically mapped using declarative mapping.
+
+        E.g.::
+
+            from sqlalchemy.orm import registry
+
+            mapper_registry = registry()
+
+            Base = mapper_registry.generate_base()
+
+
+            class MyClass(Base):
+                __tablename__ = "my_table"
+                id = Column(Integer, primary_key=True)
+
+        The above dynamically generated class is equivalent to the
+        non-dynamic example below::
+
+            from sqlalchemy.orm import registry
+            from sqlalchemy.orm.decl_api import DeclarativeMeta
+
+            mapper_registry = registry()
+
+
+            class Base(metaclass=DeclarativeMeta):
+                __abstract__ = True
+                registry = mapper_registry
+                metadata = mapper_registry.metadata
+
+                __init__ = mapper_registry.constructor
+
+        .. versionchanged:: 2.0 Note that the
+           :meth:`_orm.registry.generate_base` method is superseded by the new
+           :class:`_orm.DeclarativeBase` class, which generates a new "base"
+           class using subclassing, rather than return value of a function.
+           This allows an approach that is compatible with :pep:`484` typing
+           tools.
+
+        The :meth:`_orm.registry.generate_base` method provides the
+        implementation for the :func:`_orm.declarative_base` function, which
+        creates the :class:`_orm.registry` and base class all at once.
+
+        See the section :ref:`orm_declarative_mapping` for background and
+        examples.
+
+        :param mapper:
+          An optional callable, defaults to :class:`_orm.Mapper`.
+          This function is used to generate new :class:`_orm.Mapper` objects.
+
+        :param cls:
+          Defaults to :class:`object`. A type to use as the base for the
+          generated declarative base class. May be a class or tuple of classes.
+
+        :param name:
+          Defaults to ``Base``.  The display name for the generated
+          class.  Customizing this is not required, but can improve clarity in
+          tracebacks and debugging.
+
+        :param metaclass:
+          Defaults to :class:`.DeclarativeMeta`.  A metaclass or __metaclass__
+          compatible callable to use as the meta type of the generated
+          declarative base class.
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapping`
+
+            :func:`_orm.declarative_base`
+
+        """
+        metadata = self.metadata
+
+        bases = not isinstance(cls, tuple) and (cls,) or cls
+
+        class_dict: Dict[str, Any] = dict(registry=self, metadata=metadata)
+        if isinstance(cls, type):
+            class_dict["__doc__"] = cls.__doc__
+
+        if self.constructor is not None:
+            class_dict["__init__"] = self.constructor
+
+        class_dict["__abstract__"] = True
+        if mapper:
+            class_dict["__mapper_cls__"] = mapper
+
+        if hasattr(cls, "__class_getitem__"):
+
+            def __class_getitem__(cls: Type[_T], key: Any) -> Type[_T]:
+                # allow generic classes in py3.9+
+                return cls
+
+            class_dict["__class_getitem__"] = __class_getitem__
+
+        return metaclass(name, bases, class_dict)
+
+    @compat_typing.dataclass_transform(
+        field_specifiers=(
+            MappedColumn,
+            RelationshipProperty,
+            Composite,
+            Synonym,
+            mapped_column,
+            relationship,
+            composite,
+            synonym,
+            deferred,
+        ),
+    )
+    @overload
+    def mapped_as_dataclass(self, __cls: Type[_O]) -> Type[_O]: ...
+
+    @overload
+    def mapped_as_dataclass(
+        self,
+        __cls: Literal[None] = ...,
+        *,
+        init: Union[_NoArg, bool] = ...,
+        repr: Union[_NoArg, bool] = ...,  # noqa: A002
+        eq: Union[_NoArg, bool] = ...,
+        order: Union[_NoArg, bool] = ...,
+        unsafe_hash: Union[_NoArg, bool] = ...,
+        match_args: Union[_NoArg, bool] = ...,
+        kw_only: Union[_NoArg, bool] = ...,
+        dataclass_callable: Union[_NoArg, Callable[..., Type[Any]]] = ...,
+    ) -> Callable[[Type[_O]], Type[_O]]: ...
+
+    def mapped_as_dataclass(
+        self,
+        __cls: Optional[Type[_O]] = None,
+        *,
+        init: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        repr: Union[_NoArg, bool] = _NoArg.NO_ARG,  # noqa: A002
+        eq: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        order: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        unsafe_hash: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        match_args: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        kw_only: Union[_NoArg, bool] = _NoArg.NO_ARG,
+        dataclass_callable: Union[
+            _NoArg, Callable[..., Type[Any]]
+        ] = _NoArg.NO_ARG,
+    ) -> Union[Type[_O], Callable[[Type[_O]], Type[_O]]]:
+        """Class decorator that will apply the Declarative mapping process
+        to a given class, and additionally convert the class to be a
+        Python dataclass.
+
+        .. seealso::
+
+            :ref:`orm_declarative_native_dataclasses` - complete background
+            on SQLAlchemy native dataclass mapping
+
+
+        .. versionadded:: 2.0
+
+
+        """
+
+        def decorate(cls: Type[_O]) -> Type[_O]:
+            setattr(
+                cls,
+                "_sa_apply_dc_transforms",
+                {
+                    "init": init,
+                    "repr": repr,
+                    "eq": eq,
+                    "order": order,
+                    "unsafe_hash": unsafe_hash,
+                    "match_args": match_args,
+                    "kw_only": kw_only,
+                    "dataclass_callable": dataclass_callable,
+                },
+            )
+            _as_declarative(self, cls, cls.__dict__)
+            return cls
+
+        if __cls:
+            return decorate(__cls)
+        else:
+            return decorate
+
+    def mapped(self, cls: Type[_O]) -> Type[_O]:
+        """Class decorator that will apply the Declarative mapping process
+        to a given class.
+
+        E.g.::
+
+            from sqlalchemy.orm import registry
+
+            mapper_registry = registry()
+
+
+            @mapper_registry.mapped
+            class Foo:
+                __tablename__ = "some_table"
+
+                id = Column(Integer, primary_key=True)
+                name = Column(String)
+
+        See the section :ref:`orm_declarative_mapping` for complete
+        details and examples.
+
+        :param cls: class to be mapped.
+
+        :return: the class that was passed.
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapping`
+
+            :meth:`_orm.registry.generate_base` - generates a base class
+            that will apply Declarative mapping to subclasses automatically
+            using a Python metaclass.
+
+        .. seealso::
+
+            :meth:`_orm.registry.mapped_as_dataclass`
+
+        """
+        _as_declarative(self, cls, cls.__dict__)
+        return cls
+
+    def as_declarative_base(self, **kw: Any) -> Callable[[Type[_T]], Type[_T]]:
+        """
+        Class decorator which will invoke
+        :meth:`_orm.registry.generate_base`
+        for a given base class.
+
+        E.g.::
+
+            from sqlalchemy.orm import registry
+
+            mapper_registry = registry()
+
+
+            @mapper_registry.as_declarative_base()
+            class Base:
+                @declared_attr
+                def __tablename__(cls):
+                    return cls.__name__.lower()
+
+                id = Column(Integer, primary_key=True)
+
+
+            class MyMappedClass(Base): ...
+
+        All keyword arguments passed to
+        :meth:`_orm.registry.as_declarative_base` are passed
+        along to :meth:`_orm.registry.generate_base`.
+
+        """
+
+        def decorate(cls: Type[_T]) -> Type[_T]:
+            kw["cls"] = cls
+            kw["name"] = cls.__name__
+            return self.generate_base(**kw)  # type: ignore
+
+        return decorate
+
+    def map_declaratively(self, cls: Type[_O]) -> Mapper[_O]:
+        """Map a class declaratively.
+
+        In this form of mapping, the class is scanned for mapping information,
+        including for columns to be associated with a table, and/or an
+        actual table object.
+
+        Returns the :class:`_orm.Mapper` object.
+
+        E.g.::
+
+            from sqlalchemy.orm import registry
+
+            mapper_registry = registry()
+
+
+            class Foo:
+                __tablename__ = "some_table"
+
+                id = Column(Integer, primary_key=True)
+                name = Column(String)
+
+
+            mapper = mapper_registry.map_declaratively(Foo)
+
+        This function is more conveniently invoked indirectly via either the
+        :meth:`_orm.registry.mapped` class decorator or by subclassing a
+        declarative metaclass generated from
+        :meth:`_orm.registry.generate_base`.
+
+        See the section :ref:`orm_declarative_mapping` for complete
+        details and examples.
+
+        :param cls: class to be mapped.
+
+        :return: a :class:`_orm.Mapper` object.
+
+        .. seealso::
+
+            :ref:`orm_declarative_mapping`
+
+            :meth:`_orm.registry.mapped` - more common decorator interface
+            to this function.
+
+            :meth:`_orm.registry.map_imperatively`
+
+        """
+        _as_declarative(self, cls, cls.__dict__)
+        return cls.__mapper__  # type: ignore
+
+    def map_imperatively(
+        self,
+        class_: Type[_O],
+        local_table: Optional[FromClause] = None,
+        **kw: Any,
+    ) -> Mapper[_O]:
+        r"""Map a class imperatively.
+
+        In this form of mapping, the class is not scanned for any mapping
+        information.  Instead, all mapping constructs are passed as
+        arguments.
+
+        This method is intended to be fully equivalent to the now-removed
+        SQLAlchemy ``mapper()`` function, except that it's in terms of
+        a particular registry.
+
+        E.g.::
+
+            from sqlalchemy.orm import registry
+
+            mapper_registry = registry()
+
+            my_table = Table(
+                "my_table",
+                mapper_registry.metadata,
+                Column("id", Integer, primary_key=True),
+            )
+
+
+            class MyClass:
+                pass
+
+
+            mapper_registry.map_imperatively(MyClass, my_table)
+
+        See the section :ref:`orm_imperative_mapping` for complete background
+        and usage examples.
+
+        :param class\_: The class to be mapped.  Corresponds to the
+         :paramref:`_orm.Mapper.class_` parameter.
+
+        :param local_table: the :class:`_schema.Table` or other
+         :class:`_sql.FromClause` object that is the subject of the mapping.
+         Corresponds to the
+         :paramref:`_orm.Mapper.local_table` parameter.
+
+        :param \**kw: all other keyword arguments are passed to the
+         :class:`_orm.Mapper` constructor directly.
+
+        .. seealso::
+
+            :ref:`orm_imperative_mapping`
+
+            :ref:`orm_declarative_mapping`
+
+        """
+        return _mapper(self, class_, local_table, kw)
+
+
+RegistryType = registry
+
+if not TYPE_CHECKING:
+    # allow for runtime type resolution of ``ClassVar[_RegistryType]``
+    _RegistryType = registry  # noqa
+
+
+def as_declarative(**kw: Any) -> Callable[[Type[_T]], Type[_T]]:
+    """
+    Class decorator which will adapt a given class into a
+    :func:`_orm.declarative_base`.
+
+    This function makes use of the :meth:`_orm.registry.as_declarative_base`
+    method, by first creating a :class:`_orm.registry` automatically
+    and then invoking the decorator.
+
+    E.g.::
+
+        from sqlalchemy.orm import as_declarative
+
+
+        @as_declarative()
+        class Base:
+            @declared_attr
+            def __tablename__(cls):
+                return cls.__name__.lower()
+
+            id = Column(Integer, primary_key=True)
+
+
+        class MyMappedClass(Base): ...
+
+    .. seealso::
+
+        :meth:`_orm.registry.as_declarative_base`
+
+    """
+    metadata, class_registry = (
+        kw.pop("metadata", None),
+        kw.pop("class_registry", None),
+    )
+
+    return registry(
+        metadata=metadata, class_registry=class_registry
+    ).as_declarative_base(**kw)
+
+
+@inspection._inspects(
+    DeclarativeMeta, DeclarativeBase, DeclarativeAttributeIntercept
+)
+def _inspect_decl_meta(cls: Type[Any]) -> Optional[Mapper[Any]]:
+    mp: Optional[Mapper[Any]] = _inspect_mapped_class(cls)
+    if mp is None:
+        if _DeferredMapperConfig.has_cls(cls):
+            _DeferredMapperConfig.raise_unmapped_for_cls(cls)
+    return mp
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_base.py
new file mode 100644
index 00000000..c480994d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/decl_base.py
@@ -0,0 +1,2188 @@
+# orm/decl_base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Internal implementation for declarative."""
+
+from __future__ import annotations
+
+import collections
+import dataclasses
+import re
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Mapping
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import clsregistry
+from . import exc as orm_exc
+from . import instrumentation
+from . import mapperlib
+from ._typing import _O
+from ._typing import attr_is_internal_proxy
+from .attributes import InstrumentedAttribute
+from .attributes import QueryableAttribute
+from .base import _is_mapped_class
+from .base import InspectionAttr
+from .descriptor_props import CompositeProperty
+from .descriptor_props import SynonymProperty
+from .interfaces import _AttributeOptions
+from .interfaces import _DCAttributeOptions
+from .interfaces import _IntrospectsAnnotations
+from .interfaces import _MappedAttribute
+from .interfaces import _MapsColumns
+from .interfaces import MapperProperty
+from .mapper import Mapper
+from .properties import ColumnProperty
+from .properties import MappedColumn
+from .util import _extract_mapped_subtype
+from .util import _is_mapped_annotation
+from .util import class_mapper
+from .util import de_stringify_annotation
+from .. import event
+from .. import exc
+from .. import util
+from ..sql import expression
+from ..sql.base import _NoArg
+from ..sql.schema import Column
+from ..sql.schema import Table
+from ..util import topological
+from ..util.typing import _AnnotationScanType
+from ..util.typing import get_args
+from ..util.typing import is_fwd_ref
+from ..util.typing import is_literal
+from ..util.typing import Protocol
+from ..util.typing import TypedDict
+
+if TYPE_CHECKING:
+    from ._typing import _ClassDict
+    from ._typing import _RegistryType
+    from .base import Mapped
+    from .decl_api import declared_attr
+    from .instrumentation import ClassManager
+    from ..sql.elements import NamedColumn
+    from ..sql.schema import MetaData
+    from ..sql.selectable import FromClause
+
+_T = TypeVar("_T", bound=Any)
+
+_MapperKwArgs = Mapping[str, Any]
+_TableArgsType = Union[Tuple[Any, ...], Dict[str, Any]]
+
+
+class MappedClassProtocol(Protocol[_O]):
+    """A protocol representing a SQLAlchemy mapped class.
+
+    The protocol is generic on the type of class, use
+    ``MappedClassProtocol[Any]`` to allow any mapped class.
+    """
+
+    __name__: str
+    __mapper__: Mapper[_O]
+    __table__: FromClause
+
+    def __call__(self, **kw: Any) -> _O: ...
+
+
+class _DeclMappedClassProtocol(MappedClassProtocol[_O], Protocol):
+    "Internal more detailed version of ``MappedClassProtocol``."
+    metadata: MetaData
+    __tablename__: str
+    __mapper_args__: _MapperKwArgs
+    __table_args__: Optional[_TableArgsType]
+
+    _sa_apply_dc_transforms: Optional[_DataclassArguments]
+
+    def __declare_first__(self) -> None: ...
+
+    def __declare_last__(self) -> None: ...
+
+
+class _DataclassArguments(TypedDict):
+    init: Union[_NoArg, bool]
+    repr: Union[_NoArg, bool]
+    eq: Union[_NoArg, bool]
+    order: Union[_NoArg, bool]
+    unsafe_hash: Union[_NoArg, bool]
+    match_args: Union[_NoArg, bool]
+    kw_only: Union[_NoArg, bool]
+    dataclass_callable: Union[_NoArg, Callable[..., Type[Any]]]
+
+
+def _declared_mapping_info(
+    cls: Type[Any],
+) -> Optional[Union[_DeferredMapperConfig, Mapper[Any]]]:
+    # deferred mapping
+    if _DeferredMapperConfig.has_cls(cls):
+        return _DeferredMapperConfig.config_for_cls(cls)
+    # regular mapping
+    elif _is_mapped_class(cls):
+        return class_mapper(cls, configure=False)
+    else:
+        return None
+
+
+def _is_supercls_for_inherits(cls: Type[Any]) -> bool:
+    """return True if this class will be used as a superclass to set in
+    'inherits'.
+
+    This includes deferred mapper configs that aren't mapped yet, however does
+    not include classes with _sa_decl_prepare_nocascade (e.g.
+    ``AbstractConcreteBase``); these concrete-only classes are not set up as
+    "inherits" until after mappers are configured using
+    mapper._set_concrete_base()
+
+    """
+    if _DeferredMapperConfig.has_cls(cls):
+        return not _get_immediate_cls_attr(
+            cls, "_sa_decl_prepare_nocascade", strict=True
+        )
+    # regular mapping
+    elif _is_mapped_class(cls):
+        return True
+    else:
+        return False
+
+
+def _resolve_for_abstract_or_classical(cls: Type[Any]) -> Optional[Type[Any]]:
+    if cls is object:
+        return None
+
+    sup: Optional[Type[Any]]
+
+    if cls.__dict__.get("__abstract__", False):
+        for base_ in cls.__bases__:
+            sup = _resolve_for_abstract_or_classical(base_)
+            if sup is not None:
+                return sup
+        else:
+            return None
+    else:
+        clsmanager = _dive_for_cls_manager(cls)
+
+        if clsmanager:
+            return clsmanager.class_
+        else:
+            return cls
+
+
+def _get_immediate_cls_attr(
+    cls: Type[Any], attrname: str, strict: bool = False
+) -> Optional[Any]:
+    """return an attribute of the class that is either present directly
+    on the class, e.g. not on a superclass, or is from a superclass but
+    this superclass is a non-mapped mixin, that is, not a descendant of
+    the declarative base and is also not classically mapped.
+
+    This is used to detect attributes that indicate something about
+    a mapped class independently from any mapped classes that it may
+    inherit from.
+
+    """
+
+    # the rules are different for this name than others,
+    # make sure we've moved it out.  transitional
+    assert attrname != "__abstract__"
+
+    if not issubclass(cls, object):
+        return None
+
+    if attrname in cls.__dict__:
+        return getattr(cls, attrname)
+
+    for base in cls.__mro__[1:]:
+        _is_classical_inherits = _dive_for_cls_manager(base) is not None
+
+        if attrname in base.__dict__ and (
+            base is cls
+            or (
+                (base in cls.__bases__ if strict else True)
+                and not _is_classical_inherits
+            )
+        ):
+            return getattr(base, attrname)
+    else:
+        return None
+
+
+def _dive_for_cls_manager(cls: Type[_O]) -> Optional[ClassManager[_O]]:
+    # because the class manager registration is pluggable,
+    # we need to do the search for every class in the hierarchy,
+    # rather than just a simple "cls._sa_class_manager"
+
+    for base in cls.__mro__:
+        manager: Optional[ClassManager[_O]] = attributes.opt_manager_of_class(
+            base
+        )
+        if manager:
+            return manager
+    return None
+
+
+def _as_declarative(
+    registry: _RegistryType, cls: Type[Any], dict_: _ClassDict
+) -> Optional[_MapperConfig]:
+    # declarative scans the class for attributes.  no table or mapper
+    # args passed separately.
+    return _MapperConfig.setup_mapping(registry, cls, dict_, None, {})
+
+
+def _mapper(
+    registry: _RegistryType,
+    cls: Type[_O],
+    table: Optional[FromClause],
+    mapper_kw: _MapperKwArgs,
+) -> Mapper[_O]:
+    _ImperativeMapperConfig(registry, cls, table, mapper_kw)
+    return cast("MappedClassProtocol[_O]", cls).__mapper__
+
+
+@util.preload_module("sqlalchemy.orm.decl_api")
+def _is_declarative_props(obj: Any) -> bool:
+    _declared_attr_common = util.preloaded.orm_decl_api._declared_attr_common
+
+    return isinstance(obj, (_declared_attr_common, util.classproperty))
+
+
+def _check_declared_props_nocascade(
+    obj: Any, name: str, cls: Type[_O]
+) -> bool:
+    if _is_declarative_props(obj):
+        if getattr(obj, "_cascading", False):
+            util.warn(
+                "@declared_attr.cascading is not supported on the %s "
+                "attribute on class %s.  This attribute invokes for "
+                "subclasses in any case." % (name, cls)
+            )
+        return True
+    else:
+        return False
+
+
+class _MapperConfig:
+    __slots__ = (
+        "cls",
+        "classname",
+        "properties",
+        "declared_attr_reg",
+        "__weakref__",
+    )
+
+    cls: Type[Any]
+    classname: str
+    properties: util.OrderedDict[
+        str,
+        Union[
+            Sequence[NamedColumn[Any]], NamedColumn[Any], MapperProperty[Any]
+        ],
+    ]
+    declared_attr_reg: Dict[declared_attr[Any], Any]
+
+    @classmethod
+    def setup_mapping(
+        cls,
+        registry: _RegistryType,
+        cls_: Type[_O],
+        dict_: _ClassDict,
+        table: Optional[FromClause],
+        mapper_kw: _MapperKwArgs,
+    ) -> Optional[_MapperConfig]:
+        manager = attributes.opt_manager_of_class(cls)
+        if manager and manager.class_ is cls_:
+            raise exc.InvalidRequestError(
+                f"Class {cls!r} already has been instrumented declaratively"
+            )
+
+        if cls_.__dict__.get("__abstract__", False):
+            return None
+
+        defer_map = _get_immediate_cls_attr(
+            cls_, "_sa_decl_prepare_nocascade", strict=True
+        ) or hasattr(cls_, "_sa_decl_prepare")
+
+        if defer_map:
+            return _DeferredMapperConfig(
+                registry, cls_, dict_, table, mapper_kw
+            )
+        else:
+            return _ClassScanMapperConfig(
+                registry, cls_, dict_, table, mapper_kw
+            )
+
+    def __init__(
+        self,
+        registry: _RegistryType,
+        cls_: Type[Any],
+        mapper_kw: _MapperKwArgs,
+    ):
+        self.cls = util.assert_arg_type(cls_, type, "cls_")
+        self.classname = cls_.__name__
+        self.properties = util.OrderedDict()
+        self.declared_attr_reg = {}
+
+        if not mapper_kw.get("non_primary", False):
+            instrumentation.register_class(
+                self.cls,
+                finalize=False,
+                registry=registry,
+                declarative_scan=self,
+                init_method=registry.constructor,
+            )
+        else:
+            manager = attributes.opt_manager_of_class(self.cls)
+            if not manager or not manager.is_mapped:
+                raise exc.InvalidRequestError(
+                    "Class %s has no primary mapper configured.  Configure "
+                    "a primary mapper first before setting up a non primary "
+                    "Mapper." % self.cls
+                )
+
+    def set_cls_attribute(self, attrname: str, value: _T) -> _T:
+        manager = instrumentation.manager_of_class(self.cls)
+        manager.install_member(attrname, value)
+        return value
+
+    def map(self, mapper_kw: _MapperKwArgs = ...) -> Mapper[Any]:
+        raise NotImplementedError()
+
+    def _early_mapping(self, mapper_kw: _MapperKwArgs) -> None:
+        self.map(mapper_kw)
+
+
+class _ImperativeMapperConfig(_MapperConfig):
+    __slots__ = ("local_table", "inherits")
+
+    def __init__(
+        self,
+        registry: _RegistryType,
+        cls_: Type[_O],
+        table: Optional[FromClause],
+        mapper_kw: _MapperKwArgs,
+    ):
+        super().__init__(registry, cls_, mapper_kw)
+
+        self.local_table = self.set_cls_attribute("__table__", table)
+
+        with mapperlib._CONFIGURE_MUTEX:
+            if not mapper_kw.get("non_primary", False):
+                clsregistry.add_class(
+                    self.classname, self.cls, registry._class_registry
+                )
+
+            self._setup_inheritance(mapper_kw)
+
+            self._early_mapping(mapper_kw)
+
+    def map(self, mapper_kw: _MapperKwArgs = util.EMPTY_DICT) -> Mapper[Any]:
+        mapper_cls = Mapper
+
+        return self.set_cls_attribute(
+            "__mapper__",
+            mapper_cls(self.cls, self.local_table, **mapper_kw),
+        )
+
+    def _setup_inheritance(self, mapper_kw: _MapperKwArgs) -> None:
+        cls = self.cls
+
+        inherits = mapper_kw.get("inherits", None)
+
+        if inherits is None:
+            # since we search for classical mappings now, search for
+            # multiple mapped bases as well and raise an error.
+            inherits_search = []
+            for base_ in cls.__bases__:
+                c = _resolve_for_abstract_or_classical(base_)
+                if c is None:
+                    continue
+
+                if _is_supercls_for_inherits(c) and c not in inherits_search:
+                    inherits_search.append(c)
+
+            if inherits_search:
+                if len(inherits_search) > 1:
+                    raise exc.InvalidRequestError(
+                        "Class %s has multiple mapped bases: %r"
+                        % (cls, inherits_search)
+                    )
+                inherits = inherits_search[0]
+        elif isinstance(inherits, Mapper):
+            inherits = inherits.class_
+
+        self.inherits = inherits
+
+
+class _CollectedAnnotation(NamedTuple):
+    raw_annotation: _AnnotationScanType
+    mapped_container: Optional[Type[Mapped[Any]]]
+    extracted_mapped_annotation: Union[_AnnotationScanType, str]
+    is_dataclass: bool
+    attr_value: Any
+    originating_module: str
+    originating_class: Type[Any]
+
+
+class _ClassScanMapperConfig(_MapperConfig):
+    __slots__ = (
+        "registry",
+        "clsdict_view",
+        "collected_attributes",
+        "collected_annotations",
+        "local_table",
+        "persist_selectable",
+        "declared_columns",
+        "column_ordering",
+        "column_copies",
+        "table_args",
+        "tablename",
+        "mapper_args",
+        "mapper_args_fn",
+        "table_fn",
+        "inherits",
+        "single",
+        "allow_dataclass_fields",
+        "dataclass_setup_arguments",
+        "is_dataclass_prior_to_mapping",
+        "allow_unmapped_annotations",
+    )
+
+    is_deferred = False
+    registry: _RegistryType
+    clsdict_view: _ClassDict
+    collected_annotations: Dict[str, _CollectedAnnotation]
+    collected_attributes: Dict[str, Any]
+    local_table: Optional[FromClause]
+    persist_selectable: Optional[FromClause]
+    declared_columns: util.OrderedSet[Column[Any]]
+    column_ordering: Dict[Column[Any], int]
+    column_copies: Dict[
+        Union[MappedColumn[Any], Column[Any]],
+        Union[MappedColumn[Any], Column[Any]],
+    ]
+    tablename: Optional[str]
+    mapper_args: Mapping[str, Any]
+    table_args: Optional[_TableArgsType]
+    mapper_args_fn: Optional[Callable[[], Dict[str, Any]]]
+    inherits: Optional[Type[Any]]
+    single: bool
+
+    is_dataclass_prior_to_mapping: bool
+    allow_unmapped_annotations: bool
+
+    dataclass_setup_arguments: Optional[_DataclassArguments]
+    """if the class has SQLAlchemy native dataclass parameters, where
+    we will turn the class into a dataclass within the declarative mapping
+    process.
+
+    """
+
+    allow_dataclass_fields: bool
+    """if true, look for dataclass-processed Field objects on the target
+    class as well as superclasses and extract ORM mapping directives from
+    the "metadata" attribute of each Field.
+
+    if False, dataclass fields can still be used, however they won't be
+    mapped.
+
+    """
+
+    def __init__(
+        self,
+        registry: _RegistryType,
+        cls_: Type[_O],
+        dict_: _ClassDict,
+        table: Optional[FromClause],
+        mapper_kw: _MapperKwArgs,
+    ):
+        # grab class dict before the instrumentation manager has been added.
+        # reduces cycles
+        self.clsdict_view = (
+            util.immutabledict(dict_) if dict_ else util.EMPTY_DICT
+        )
+        super().__init__(registry, cls_, mapper_kw)
+        self.registry = registry
+        self.persist_selectable = None
+
+        self.collected_attributes = {}
+        self.collected_annotations = {}
+        self.declared_columns = util.OrderedSet()
+        self.column_ordering = {}
+        self.column_copies = {}
+        self.single = False
+        self.dataclass_setup_arguments = dca = getattr(
+            self.cls, "_sa_apply_dc_transforms", None
+        )
+
+        self.allow_unmapped_annotations = getattr(
+            self.cls, "__allow_unmapped__", False
+        ) or bool(self.dataclass_setup_arguments)
+
+        self.is_dataclass_prior_to_mapping = cld = dataclasses.is_dataclass(
+            cls_
+        )
+
+        sdk = _get_immediate_cls_attr(cls_, "__sa_dataclass_metadata_key__")
+
+        # we don't want to consume Field objects from a not-already-dataclass.
+        # the Field objects won't have their "name" or "type" populated,
+        # and while it seems like we could just set these on Field as we
+        # read them, Field is documented as "user read only" and we need to
+        # stay far away from any off-label use of dataclasses APIs.
+        if (not cld or dca) and sdk:
+            raise exc.InvalidRequestError(
+                "SQLAlchemy mapped dataclasses can't consume mapping "
+                "information from dataclass.Field() objects if the immediate "
+                "class is not already a dataclass."
+            )
+
+        # if already a dataclass, and __sa_dataclass_metadata_key__ present,
+        # then also look inside of dataclass.Field() objects yielded by
+        # dataclasses.get_fields(cls) when scanning for attributes
+        self.allow_dataclass_fields = bool(sdk and cld)
+
+        self._setup_declared_events()
+
+        self._scan_attributes()
+
+        self._setup_dataclasses_transforms()
+
+        with mapperlib._CONFIGURE_MUTEX:
+            clsregistry.add_class(
+                self.classname, self.cls, registry._class_registry
+            )
+
+            self._setup_inheriting_mapper(mapper_kw)
+
+            self._extract_mappable_attributes()
+
+            self._extract_declared_columns()
+
+            self._setup_table(table)
+
+            self._setup_inheriting_columns(mapper_kw)
+
+            self._early_mapping(mapper_kw)
+
+    def _setup_declared_events(self) -> None:
+        if _get_immediate_cls_attr(self.cls, "__declare_last__"):
+
+            @event.listens_for(Mapper, "after_configured")
+            def after_configured() -> None:
+                cast(
+                    "_DeclMappedClassProtocol[Any]", self.cls
+                ).__declare_last__()
+
+        if _get_immediate_cls_attr(self.cls, "__declare_first__"):
+
+            @event.listens_for(Mapper, "before_configured")
+            def before_configured() -> None:
+                cast(
+                    "_DeclMappedClassProtocol[Any]", self.cls
+                ).__declare_first__()
+
+    def _cls_attr_override_checker(
+        self, cls: Type[_O]
+    ) -> Callable[[str, Any], bool]:
+        """Produce a function that checks if a class has overridden an
+        attribute, taking SQLAlchemy-enabled dataclass fields into account.
+
+        """
+
+        if self.allow_dataclass_fields:
+            sa_dataclass_metadata_key = _get_immediate_cls_attr(
+                cls, "__sa_dataclass_metadata_key__"
+            )
+        else:
+            sa_dataclass_metadata_key = None
+
+        if not sa_dataclass_metadata_key:
+
+            def attribute_is_overridden(key: str, obj: Any) -> bool:
+                return getattr(cls, key, obj) is not obj
+
+        else:
+            all_datacls_fields = {
+                f.name: f.metadata[sa_dataclass_metadata_key]
+                for f in util.dataclass_fields(cls)
+                if sa_dataclass_metadata_key in f.metadata
+            }
+            local_datacls_fields = {
+                f.name: f.metadata[sa_dataclass_metadata_key]
+                for f in util.local_dataclass_fields(cls)
+                if sa_dataclass_metadata_key in f.metadata
+            }
+
+            absent = object()
+
+            def attribute_is_overridden(key: str, obj: Any) -> bool:
+                if _is_declarative_props(obj):
+                    obj = obj.fget
+
+                # this function likely has some failure modes still if
+                # someone is doing a deep mixing of the same attribute
+                # name as plain Python attribute vs. dataclass field.
+
+                ret = local_datacls_fields.get(key, absent)
+                if _is_declarative_props(ret):
+                    ret = ret.fget
+
+                if ret is obj:
+                    return False
+                elif ret is not absent:
+                    return True
+
+                all_field = all_datacls_fields.get(key, absent)
+
+                ret = getattr(cls, key, obj)
+
+                if ret is obj:
+                    return False
+
+                # for dataclasses, this could be the
+                # 'default' of the field.  so filter more specifically
+                # for an already-mapped InstrumentedAttribute
+                if ret is not absent and isinstance(
+                    ret, InstrumentedAttribute
+                ):
+                    return True
+
+                if all_field is obj:
+                    return False
+                elif all_field is not absent:
+                    return True
+
+                # can't find another attribute
+                return False
+
+        return attribute_is_overridden
+
+    _include_dunders = {
+        "__table__",
+        "__mapper_args__",
+        "__tablename__",
+        "__table_args__",
+    }
+
+    _match_exclude_dunders = re.compile(r"^(?:_sa_|__)")
+
+    def _cls_attr_resolver(
+        self, cls: Type[Any]
+    ) -> Callable[[], Iterable[Tuple[str, Any, Any, bool]]]:
+        """produce a function to iterate the "attributes" of a class
+        which we want to consider for mapping, adjusting for SQLAlchemy fields
+        embedded in dataclass fields.
+
+        """
+        cls_annotations = util.get_annotations(cls)
+
+        cls_vars = vars(cls)
+
+        _include_dunders = self._include_dunders
+        _match_exclude_dunders = self._match_exclude_dunders
+
+        names = [
+            n
+            for n in util.merge_lists_w_ordering(
+                list(cls_vars), list(cls_annotations)
+            )
+            if not _match_exclude_dunders.match(n) or n in _include_dunders
+        ]
+
+        if self.allow_dataclass_fields:
+            sa_dataclass_metadata_key: Optional[str] = _get_immediate_cls_attr(
+                cls, "__sa_dataclass_metadata_key__"
+            )
+        else:
+            sa_dataclass_metadata_key = None
+
+        if not sa_dataclass_metadata_key:
+
+            def local_attributes_for_class() -> (
+                Iterable[Tuple[str, Any, Any, bool]]
+            ):
+                return (
+                    (
+                        name,
+                        cls_vars.get(name),
+                        cls_annotations.get(name),
+                        False,
+                    )
+                    for name in names
+                )
+
+        else:
+            dataclass_fields = {
+                field.name: field for field in util.local_dataclass_fields(cls)
+            }
+
+            fixed_sa_dataclass_metadata_key = sa_dataclass_metadata_key
+
+            def local_attributes_for_class() -> (
+                Iterable[Tuple[str, Any, Any, bool]]
+            ):
+                for name in names:
+                    field = dataclass_fields.get(name, None)
+                    if field and sa_dataclass_metadata_key in field.metadata:
+                        yield field.name, _as_dc_declaredattr(
+                            field.metadata, fixed_sa_dataclass_metadata_key
+                        ), cls_annotations.get(field.name), True
+                    else:
+                        yield name, cls_vars.get(name), cls_annotations.get(
+                            name
+                        ), False
+
+        return local_attributes_for_class
+
+    def _scan_attributes(self) -> None:
+        cls = self.cls
+
+        cls_as_Decl = cast("_DeclMappedClassProtocol[Any]", cls)
+
+        clsdict_view = self.clsdict_view
+        collected_attributes = self.collected_attributes
+        column_copies = self.column_copies
+        _include_dunders = self._include_dunders
+        mapper_args_fn = None
+        table_args = inherited_table_args = None
+        table_fn = None
+        tablename = None
+        fixed_table = "__table__" in clsdict_view
+
+        attribute_is_overridden = self._cls_attr_override_checker(self.cls)
+
+        bases = []
+
+        for base in cls.__mro__:
+            # collect bases and make sure standalone columns are copied
+            # to be the column they will ultimately be on the class,
+            # so that declared_attr functions use the right columns.
+            # need to do this all the way up the hierarchy first
+            # (see #8190)
+
+            class_mapped = base is not cls and _is_supercls_for_inherits(base)
+
+            local_attributes_for_class = self._cls_attr_resolver(base)
+
+            if not class_mapped and base is not cls:
+                locally_collected_columns = self._produce_column_copies(
+                    local_attributes_for_class,
+                    attribute_is_overridden,
+                    fixed_table,
+                    base,
+                )
+            else:
+                locally_collected_columns = {}
+
+            bases.append(
+                (
+                    base,
+                    class_mapped,
+                    local_attributes_for_class,
+                    locally_collected_columns,
+                )
+            )
+
+        for (
+            base,
+            class_mapped,
+            local_attributes_for_class,
+            locally_collected_columns,
+        ) in bases:
+            # this transfer can also take place as we scan each name
+            # for finer-grained control of how collected_attributes is
+            # populated, as this is what impacts column ordering.
+            # however it's simpler to get it out of the way here.
+            collected_attributes.update(locally_collected_columns)
+
+            for (
+                name,
+                obj,
+                annotation,
+                is_dataclass_field,
+            ) in local_attributes_for_class():
+                if name in _include_dunders:
+                    if name == "__mapper_args__":
+                        check_decl = _check_declared_props_nocascade(
+                            obj, name, cls
+                        )
+                        if not mapper_args_fn and (
+                            not class_mapped or check_decl
+                        ):
+                            # don't even invoke __mapper_args__ until
+                            # after we've determined everything about the
+                            # mapped table.
+                            # make a copy of it so a class-level dictionary
+                            # is not overwritten when we update column-based
+                            # arguments.
+                            def _mapper_args_fn() -> Dict[str, Any]:
+                                return dict(cls_as_Decl.__mapper_args__)
+
+                            mapper_args_fn = _mapper_args_fn
+
+                    elif name == "__tablename__":
+                        check_decl = _check_declared_props_nocascade(
+                            obj, name, cls
+                        )
+                        if not tablename and (not class_mapped or check_decl):
+                            tablename = cls_as_Decl.__tablename__
+                    elif name == "__table__":
+                        check_decl = _check_declared_props_nocascade(
+                            obj, name, cls
+                        )
+                        # if a @declared_attr using "__table__" is detected,
+                        # wrap up a callable to look for "__table__" from
+                        # the final concrete class when we set up a table.
+                        # this was fixed by
+                        # #11509, regression in 2.0 from version 1.4.
+                        if check_decl and not table_fn:
+                            # don't even invoke __table__ until we're ready
+                            def _table_fn() -> FromClause:
+                                return cls_as_Decl.__table__
+
+                            table_fn = _table_fn
+
+                    elif name == "__table_args__":
+                        check_decl = _check_declared_props_nocascade(
+                            obj, name, cls
+                        )
+                        if not table_args and (not class_mapped or check_decl):
+                            table_args = cls_as_Decl.__table_args__
+                            if not isinstance(
+                                table_args, (tuple, dict, type(None))
+                            ):
+                                raise exc.ArgumentError(
+                                    "__table_args__ value must be a tuple, "
+                                    "dict, or None"
+                                )
+                            if base is not cls:
+                                inherited_table_args = True
+                    else:
+                        # any other dunder names; should not be here
+                        # as we have tested for all four names in
+                        # _include_dunders
+                        assert False
+                elif class_mapped:
+                    if _is_declarative_props(obj) and not obj._quiet:
+                        util.warn(
+                            "Regular (i.e. not __special__) "
+                            "attribute '%s.%s' uses @declared_attr, "
+                            "but owning class %s is mapped - "
+                            "not applying to subclass %s."
+                            % (base.__name__, name, base, cls)
+                        )
+
+                    continue
+                elif base is not cls:
+                    # we're a mixin, abstract base, or something that is
+                    # acting like that for now.
+
+                    if isinstance(obj, (Column, MappedColumn)):
+                        # already copied columns to the mapped class.
+                        continue
+                    elif isinstance(obj, MapperProperty):
+                        raise exc.InvalidRequestError(
+                            "Mapper properties (i.e. deferred,"
+                            "column_property(), relationship(), etc.) must "
+                            "be declared as @declared_attr callables "
+                            "on declarative mixin classes.  For dataclass "
+                            "field() objects, use a lambda:"
+                        )
+                    elif _is_declarative_props(obj):
+                        # tried to get overloads to tell this to
+                        # pylance, no luck
+                        assert obj is not None
+
+                        if obj._cascading:
+                            if name in clsdict_view:
+                                # unfortunately, while we can use the user-
+                                # defined attribute here to allow a clean
+                                # override, if there's another
+                                # subclass below then it still tries to use
+                                # this.  not sure if there is enough
+                                # information here to add this as a feature
+                                # later on.
+                                util.warn(
+                                    "Attribute '%s' on class %s cannot be "
+                                    "processed due to "
+                                    "@declared_attr.cascading; "
+                                    "skipping" % (name, cls)
+                                )
+                            collected_attributes[name] = column_copies[obj] = (
+                                ret
+                            ) = obj.__get__(obj, cls)
+                            setattr(cls, name, ret)
+                        else:
+                            if is_dataclass_field:
+                                # access attribute using normal class access
+                                # first, to see if it's been mapped on a
+                                # superclass.   note if the dataclasses.field()
+                                # has "default", this value can be anything.
+                                ret = getattr(cls, name, None)
+
+                                # so, if it's anything that's not ORM
+                                # mapped, assume we should invoke the
+                                # declared_attr
+                                if not isinstance(ret, InspectionAttr):
+                                    ret = obj.fget()
+                            else:
+                                # access attribute using normal class access.
+                                # if the declared attr already took place
+                                # on a superclass that is mapped, then
+                                # this is no longer a declared_attr, it will
+                                # be the InstrumentedAttribute
+                                ret = getattr(cls, name)
+
+                            # correct for proxies created from hybrid_property
+                            # or similar.  note there is no known case that
+                            # produces nested proxies, so we are only
+                            # looking one level deep right now.
+
+                            if (
+                                isinstance(ret, InspectionAttr)
+                                and attr_is_internal_proxy(ret)
+                                and not isinstance(
+                                    ret.original_property, MapperProperty
+                                )
+                            ):
+                                ret = ret.descriptor
+
+                            collected_attributes[name] = column_copies[obj] = (
+                                ret
+                            )
+
+                        if (
+                            isinstance(ret, (Column, MapperProperty))
+                            and ret.doc is None
+                        ):
+                            ret.doc = obj.__doc__
+
+                        self._collect_annotation(
+                            name,
+                            obj._collect_return_annotation(),
+                            base,
+                            True,
+                            obj,
+                        )
+                    elif _is_mapped_annotation(annotation, cls, base):
+                        # Mapped annotation without any object.
+                        # product_column_copies should have handled this.
+                        # if future support for other MapperProperty,
+                        # then test if this name is already handled and
+                        # otherwise proceed to generate.
+                        if not fixed_table:
+                            assert (
+                                name in collected_attributes
+                                or attribute_is_overridden(name, None)
+                            )
+                        continue
+                    else:
+                        # here, the attribute is some other kind of
+                        # property that we assume is not part of the
+                        # declarative mapping.  however, check for some
+                        # more common mistakes
+                        self._warn_for_decl_attributes(base, name, obj)
+                elif is_dataclass_field and (
+                    name not in clsdict_view or clsdict_view[name] is not obj
+                ):
+                    # here, we are definitely looking at the target class
+                    # and not a superclass.   this is currently a
+                    # dataclass-only path.  if the name is only
+                    # a dataclass field and isn't in local cls.__dict__,
+                    # put the object there.
+                    # assert that the dataclass-enabled resolver agrees
+                    # with what we are seeing
+
+                    assert not attribute_is_overridden(name, obj)
+
+                    if _is_declarative_props(obj):
+                        obj = obj.fget()
+
+                    collected_attributes[name] = obj
+                    self._collect_annotation(
+                        name, annotation, base, False, obj
+                    )
+                else:
+                    collected_annotation = self._collect_annotation(
+                        name, annotation, base, None, obj
+                    )
+                    is_mapped = (
+                        collected_annotation is not None
+                        and collected_annotation.mapped_container is not None
+                    )
+                    generated_obj = (
+                        collected_annotation.attr_value
+                        if collected_annotation is not None
+                        else obj
+                    )
+                    if obj is None and not fixed_table and is_mapped:
+                        collected_attributes[name] = (
+                            generated_obj
+                            if generated_obj is not None
+                            else MappedColumn()
+                        )
+                    elif name in clsdict_view:
+                        collected_attributes[name] = obj
+                    # else if the name is not in the cls.__dict__,
+                    # don't collect it as an attribute.
+                    # we will see the annotation only, which is meaningful
+                    # both for mapping and dataclasses setup
+
+        if inherited_table_args and not tablename:
+            table_args = None
+
+        self.table_args = table_args
+        self.tablename = tablename
+        self.mapper_args_fn = mapper_args_fn
+        self.table_fn = table_fn
+
+    def _setup_dataclasses_transforms(self) -> None:
+        dataclass_setup_arguments = self.dataclass_setup_arguments
+        if not dataclass_setup_arguments:
+            return
+
+        # can't use is_dataclass since it uses hasattr
+        if "__dataclass_fields__" in self.cls.__dict__:
+            raise exc.InvalidRequestError(
+                f"Class {self.cls} is already a dataclass; ensure that "
+                "base classes / decorator styles of establishing dataclasses "
+                "are not being mixed. "
+                "This can happen if a class that inherits from "
+                "'MappedAsDataclass', even indirectly, is been mapped with "
+                "'@registry.mapped_as_dataclass'"
+            )
+
+        # can't create a dataclass if __table__ is already there. This would
+        # fail an assertion when calling _get_arguments_for_make_dataclass:
+        # assert False, "Mapped[] received without a mapping declaration"
+        if "__table__" in self.cls.__dict__:
+            raise exc.InvalidRequestError(
+                f"Class {self.cls} already defines a '__table__'. "
+                "ORM Annotated Dataclasses do not support a pre-existing "
+                "'__table__' element"
+            )
+
+        warn_for_non_dc_attrs = collections.defaultdict(list)
+
+        def _allow_dataclass_field(
+            key: str, originating_class: Type[Any]
+        ) -> bool:
+            if (
+                originating_class is not self.cls
+                and "__dataclass_fields__" not in originating_class.__dict__
+            ):
+                warn_for_non_dc_attrs[originating_class].append(key)
+
+            return True
+
+        manager = instrumentation.manager_of_class(self.cls)
+        assert manager is not None
+
+        field_list = [
+            _AttributeOptions._get_arguments_for_make_dataclass(
+                key,
+                anno,
+                mapped_container,
+                self.collected_attributes.get(key, _NoArg.NO_ARG),
+            )
+            for key, anno, mapped_container in (
+                (
+                    key,
+                    mapped_anno if mapped_anno else raw_anno,
+                    mapped_container,
+                )
+                for key, (
+                    raw_anno,
+                    mapped_container,
+                    mapped_anno,
+                    is_dc,
+                    attr_value,
+                    originating_module,
+                    originating_class,
+                ) in self.collected_annotations.items()
+                if _allow_dataclass_field(key, originating_class)
+                and (
+                    key not in self.collected_attributes
+                    # issue #9226; check for attributes that we've collected
+                    # which are already instrumented, which we would assume
+                    # mean we are in an ORM inheritance mapping and this
+                    # attribute is already mapped on the superclass.   Under
+                    # no circumstance should any QueryableAttribute be sent to
+                    # the dataclass() function; anything that's mapped should
+                    # be Field and that's it
+                    or not isinstance(
+                        self.collected_attributes[key], QueryableAttribute
+                    )
+                )
+            )
+        ]
+
+        if warn_for_non_dc_attrs:
+            for (
+                originating_class,
+                non_dc_attrs,
+            ) in warn_for_non_dc_attrs.items():
+                util.warn_deprecated(
+                    f"When transforming {self.cls} to a dataclass, "
+                    f"attribute(s) "
+                    f"{', '.join(repr(key) for key in non_dc_attrs)} "
+                    f"originates from superclass "
+                    f"{originating_class}, which is not a dataclass.  This "
+                    f"usage is deprecated and will raise an error in "
+                    f"SQLAlchemy 2.1.  When declaring SQLAlchemy Declarative "
+                    f"Dataclasses, ensure that all mixin classes and other "
+                    f"superclasses which include attributes are also a "
+                    f"subclass of MappedAsDataclass.",
+                    "2.0",
+                    code="dcmx",
+                )
+
+        annotations = {}
+        defaults = {}
+        for item in field_list:
+            if len(item) == 2:
+                name, tp = item
+            elif len(item) == 3:
+                name, tp, spec = item
+                defaults[name] = spec
+            else:
+                assert False
+            annotations[name] = tp
+
+        for k, v in defaults.items():
+            setattr(self.cls, k, v)
+
+        self._apply_dataclasses_to_any_class(
+            dataclass_setup_arguments, self.cls, annotations
+        )
+
+    @classmethod
+    def _update_annotations_for_non_mapped_class(
+        cls, klass: Type[_O]
+    ) -> Mapping[str, _AnnotationScanType]:
+        cls_annotations = util.get_annotations(klass)
+
+        new_anno = {}
+        for name, annotation in cls_annotations.items():
+            if _is_mapped_annotation(annotation, klass, klass):
+                extracted = _extract_mapped_subtype(
+                    annotation,
+                    klass,
+                    klass.__module__,
+                    name,
+                    type(None),
+                    required=False,
+                    is_dataclass_field=False,
+                    expect_mapped=False,
+                )
+                if extracted:
+                    inner, _ = extracted
+                    new_anno[name] = inner
+            else:
+                new_anno[name] = annotation
+        return new_anno
+
+    @classmethod
+    def _apply_dataclasses_to_any_class(
+        cls,
+        dataclass_setup_arguments: _DataclassArguments,
+        klass: Type[_O],
+        use_annotations: Mapping[str, _AnnotationScanType],
+    ) -> None:
+        cls._assert_dc_arguments(dataclass_setup_arguments)
+
+        dataclass_callable = dataclass_setup_arguments["dataclass_callable"]
+        if dataclass_callable is _NoArg.NO_ARG:
+            dataclass_callable = dataclasses.dataclass
+
+        restored: Optional[Any]
+
+        if use_annotations:
+            # apply constructed annotations that should look "normal" to a
+            # dataclasses callable, based on the fields present.  This
+            # means remove the Mapped[] container and ensure all Field
+            # entries have an annotation
+            restored = getattr(klass, "__annotations__", None)
+            klass.__annotations__ = cast("Dict[str, Any]", use_annotations)
+        else:
+            restored = None
+
+        try:
+            dataclass_callable(
+                klass,
+                **{
+                    k: v
+                    for k, v in dataclass_setup_arguments.items()
+                    if v is not _NoArg.NO_ARG and k != "dataclass_callable"
+                },
+            )
+        except (TypeError, ValueError) as ex:
+            raise exc.InvalidRequestError(
+                f"Python dataclasses error encountered when creating "
+                f"dataclass for {klass.__name__!r}: "
+                f"{ex!r}. Please refer to Python dataclasses "
+                "documentation for additional information.",
+                code="dcte",
+            ) from ex
+        finally:
+            # restore original annotations outside of the dataclasses
+            # process; for mixins and __abstract__ superclasses, SQLAlchemy
+            # Declarative will need to see the Mapped[] container inside the
+            # annotations in order to map subclasses
+            if use_annotations:
+                if restored is None:
+                    del klass.__annotations__
+                else:
+                    klass.__annotations__ = restored
+
+    @classmethod
+    def _assert_dc_arguments(cls, arguments: _DataclassArguments) -> None:
+        allowed = {
+            "init",
+            "repr",
+            "order",
+            "eq",
+            "unsafe_hash",
+            "kw_only",
+            "match_args",
+            "dataclass_callable",
+        }
+        disallowed_args = set(arguments).difference(allowed)
+        if disallowed_args:
+            msg = ", ".join(f"{arg!r}" for arg in sorted(disallowed_args))
+            raise exc.ArgumentError(
+                f"Dataclass argument(s) {msg} are not accepted"
+            )
+
+    def _collect_annotation(
+        self,
+        name: str,
+        raw_annotation: _AnnotationScanType,
+        originating_class: Type[Any],
+        expect_mapped: Optional[bool],
+        attr_value: Any,
+    ) -> Optional[_CollectedAnnotation]:
+        if name in self.collected_annotations:
+            return self.collected_annotations[name]
+
+        if raw_annotation is None:
+            return None
+
+        is_dataclass = self.is_dataclass_prior_to_mapping
+        allow_unmapped = self.allow_unmapped_annotations
+
+        if expect_mapped is None:
+            is_dataclass_field = isinstance(attr_value, dataclasses.Field)
+            expect_mapped = (
+                not is_dataclass_field
+                and not allow_unmapped
+                and (
+                    attr_value is None
+                    or isinstance(attr_value, _MappedAttribute)
+                )
+            )
+        else:
+            is_dataclass_field = False
+
+        is_dataclass_field = False
+        extracted = _extract_mapped_subtype(
+            raw_annotation,
+            self.cls,
+            originating_class.__module__,
+            name,
+            type(attr_value),
+            required=False,
+            is_dataclass_field=is_dataclass_field,
+            expect_mapped=expect_mapped and not is_dataclass,
+        )
+        if extracted is None:
+            # ClassVar can come out here
+            return None
+
+        extracted_mapped_annotation, mapped_container = extracted
+
+        if attr_value is None and not is_literal(extracted_mapped_annotation):
+            for elem in get_args(extracted_mapped_annotation):
+                if is_fwd_ref(
+                    elem, check_generic=True, check_for_plain_string=True
+                ):
+                    elem = de_stringify_annotation(
+                        self.cls,
+                        elem,
+                        originating_class.__module__,
+                        include_generic=True,
+                    )
+                # look in Annotated[...] for an ORM construct,
+                # such as Annotated[int, mapped_column(primary_key=True)]
+                if isinstance(elem, _IntrospectsAnnotations):
+                    attr_value = elem.found_in_pep593_annotated()
+
+        self.collected_annotations[name] = ca = _CollectedAnnotation(
+            raw_annotation,
+            mapped_container,
+            extracted_mapped_annotation,
+            is_dataclass,
+            attr_value,
+            originating_class.__module__,
+            originating_class,
+        )
+        return ca
+
+    def _warn_for_decl_attributes(
+        self, cls: Type[Any], key: str, c: Any
+    ) -> None:
+        if isinstance(c, expression.ColumnElement):
+            util.warn(
+                f"Attribute '{key}' on class {cls} appears to "
+                "be a non-schema SQLAlchemy expression "
+                "object; this won't be part of the declarative mapping. "
+                "To map arbitrary expressions, use ``column_property()`` "
+                "or a similar function such as ``deferred()``, "
+                "``query_expression()`` etc. "
+            )
+
+    def _produce_column_copies(
+        self,
+        attributes_for_class: Callable[
+            [], Iterable[Tuple[str, Any, Any, bool]]
+        ],
+        attribute_is_overridden: Callable[[str, Any], bool],
+        fixed_table: bool,
+        originating_class: Type[Any],
+    ) -> Dict[str, Union[Column[Any], MappedColumn[Any]]]:
+        cls = self.cls
+        dict_ = self.clsdict_view
+        locally_collected_attributes = {}
+        column_copies = self.column_copies
+        # copy mixin columns to the mapped class
+
+        for name, obj, annotation, is_dataclass in attributes_for_class():
+            if (
+                not fixed_table
+                and obj is None
+                and _is_mapped_annotation(annotation, cls, originating_class)
+            ):
+                # obj is None means this is the annotation only path
+
+                if attribute_is_overridden(name, obj):
+                    # perform same "overridden" check as we do for
+                    # Column/MappedColumn, this is how a mixin col is not
+                    # applied to an inherited subclass that does not have
+                    # the mixin.   the anno-only path added here for
+                    # #9564
+                    continue
+
+                collected_annotation = self._collect_annotation(
+                    name, annotation, originating_class, True, obj
+                )
+                obj = (
+                    collected_annotation.attr_value
+                    if collected_annotation is not None
+                    else obj
+                )
+                if obj is None:
+                    obj = MappedColumn()
+
+                locally_collected_attributes[name] = obj
+                setattr(cls, name, obj)
+
+            elif isinstance(obj, (Column, MappedColumn)):
+                if attribute_is_overridden(name, obj):
+                    # if column has been overridden
+                    # (like by the InstrumentedAttribute of the
+                    # superclass), skip.  don't collect the annotation
+                    # either (issue #8718)
+                    continue
+
+                collected_annotation = self._collect_annotation(
+                    name, annotation, originating_class, True, obj
+                )
+                obj = (
+                    collected_annotation.attr_value
+                    if collected_annotation is not None
+                    else obj
+                )
+
+                if name not in dict_ and not (
+                    "__table__" in dict_
+                    and (getattr(obj, "name", None) or name)
+                    in dict_["__table__"].c
+                ):
+                    if obj.foreign_keys:
+                        for fk in obj.foreign_keys:
+                            if (
+                                fk._table_column is not None
+                                and fk._table_column.table is None
+                            ):
+                                raise exc.InvalidRequestError(
+                                    "Columns with foreign keys to "
+                                    "non-table-bound "
+                                    "columns must be declared as "
+                                    "@declared_attr callables "
+                                    "on declarative mixin classes.  "
+                                    "For dataclass "
+                                    "field() objects, use a lambda:."
+                                )
+
+                    column_copies[obj] = copy_ = obj._copy()
+
+                    locally_collected_attributes[name] = copy_
+                    setattr(cls, name, copy_)
+
+        return locally_collected_attributes
+
+    def _extract_mappable_attributes(self) -> None:
+        cls = self.cls
+        collected_attributes = self.collected_attributes
+
+        our_stuff = self.properties
+
+        _include_dunders = self._include_dunders
+
+        late_mapped = _get_immediate_cls_attr(
+            cls, "_sa_decl_prepare_nocascade", strict=True
+        )
+
+        allow_unmapped_annotations = self.allow_unmapped_annotations
+        expect_annotations_wo_mapped = (
+            allow_unmapped_annotations or self.is_dataclass_prior_to_mapping
+        )
+
+        look_for_dataclass_things = bool(self.dataclass_setup_arguments)
+
+        for k in list(collected_attributes):
+            if k in _include_dunders:
+                continue
+
+            value = collected_attributes[k]
+
+            if _is_declarative_props(value):
+                # @declared_attr in collected_attributes only occurs here for a
+                # @declared_attr that's directly on the mapped class;
+                # for a mixin, these have already been evaluated
+                if value._cascading:
+                    util.warn(
+                        "Use of @declared_attr.cascading only applies to "
+                        "Declarative 'mixin' and 'abstract' classes.  "
+                        "Currently, this flag is ignored on mapped class "
+                        "%s" % self.cls
+                    )
+
+                value = getattr(cls, k)
+
+            elif (
+                isinstance(value, QueryableAttribute)
+                and value.class_ is not cls
+                and value.key != k
+            ):
+                # detect a QueryableAttribute that's already mapped being
+                # assigned elsewhere in userland, turn into a synonym()
+                value = SynonymProperty(value.key)
+                setattr(cls, k, value)
+
+            if (
+                isinstance(value, tuple)
+                and len(value) == 1
+                and isinstance(value[0], (Column, _MappedAttribute))
+            ):
+                util.warn(
+                    "Ignoring declarative-like tuple value of attribute "
+                    "'%s': possibly a copy-and-paste error with a comma "
+                    "accidentally placed at the end of the line?" % k
+                )
+                continue
+            elif look_for_dataclass_things and isinstance(
+                value, dataclasses.Field
+            ):
+                # we collected a dataclass Field; dataclasses would have
+                # set up the correct state on the class
+                continue
+            elif not isinstance(value, (Column, _DCAttributeOptions)):
+                # using @declared_attr for some object that
+                # isn't Column/MapperProperty/_DCAttributeOptions; remove
+                # from the clsdict_view
+                # and place the evaluated value onto the class.
+                collected_attributes.pop(k)
+                self._warn_for_decl_attributes(cls, k, value)
+                if not late_mapped:
+                    setattr(cls, k, value)
+                continue
+            # we expect to see the name 'metadata' in some valid cases;
+            # however at this point we see it's assigned to something trying
+            # to be mapped, so raise for that.
+            # TODO: should "registry" here be also?   might be too late
+            # to change that now (2.0 betas)
+            elif k in ("metadata",):
+                raise exc.InvalidRequestError(
+                    f"Attribute name '{k}' is reserved when using the "
+                    "Declarative API."
+                )
+            elif isinstance(value, Column):
+                _undefer_column_name(
+                    k, self.column_copies.get(value, value)  # type: ignore
+                )
+            else:
+                if isinstance(value, _IntrospectsAnnotations):
+                    (
+                        annotation,
+                        mapped_container,
+                        extracted_mapped_annotation,
+                        is_dataclass,
+                        attr_value,
+                        originating_module,
+                        originating_class,
+                    ) = self.collected_annotations.get(
+                        k, (None, None, None, False, None, None, None)
+                    )
+
+                    # issue #8692 - don't do any annotation interpretation if
+                    # an annotation were present and a container such as
+                    # Mapped[] etc. were not used.  If annotation is None,
+                    # do declarative_scan so that the property can raise
+                    # for required
+                    if (
+                        mapped_container is not None
+                        or annotation is None
+                        # issue #10516: need to do declarative_scan even with
+                        # a non-Mapped annotation if we are doing
+                        # __allow_unmapped__, for things like col.name
+                        # assignment
+                        or allow_unmapped_annotations
+                    ):
+                        try:
+                            value.declarative_scan(
+                                self,
+                                self.registry,
+                                cls,
+                                originating_module,
+                                k,
+                                mapped_container,
+                                annotation,
+                                extracted_mapped_annotation,
+                                is_dataclass,
+                            )
+                        except NameError as ne:
+                            raise exc.ArgumentError(
+                                f"Could not resolve all types within mapped "
+                                f'annotation: "{annotation}".  Ensure all '
+                                f"types are written correctly and are "
+                                f"imported within the module in use."
+                            ) from ne
+                    else:
+                        # assert that we were expecting annotations
+                        # without Mapped[] were going to be passed.
+                        # otherwise an error should have been raised
+                        # by util._extract_mapped_subtype before we got here.
+                        assert expect_annotations_wo_mapped
+
+                if isinstance(value, _DCAttributeOptions):
+                    if (
+                        value._has_dataclass_arguments
+                        and not look_for_dataclass_things
+                    ):
+                        if isinstance(value, MapperProperty):
+                            argnames = [
+                                "init",
+                                "default_factory",
+                                "repr",
+                                "default",
+                            ]
+                        else:
+                            argnames = ["init", "default_factory", "repr"]
+
+                        args = {
+                            a
+                            for a in argnames
+                            if getattr(
+                                value._attribute_options, f"dataclasses_{a}"
+                            )
+                            is not _NoArg.NO_ARG
+                        }
+
+                        raise exc.ArgumentError(
+                            f"Attribute '{k}' on class {cls} includes "
+                            f"dataclasses argument(s): "
+                            f"{', '.join(sorted(repr(a) for a in args))} but "
+                            f"class does not specify "
+                            "SQLAlchemy native dataclass configuration."
+                        )
+
+                    if not isinstance(value, (MapperProperty, _MapsColumns)):
+                        # filter for _DCAttributeOptions objects that aren't
+                        # MapperProperty / mapped_column().  Currently this
+                        # includes AssociationProxy.   pop it from the things
+                        # we're going to map and set it up as a descriptor
+                        # on the class.
+                        collected_attributes.pop(k)
+
+                        # Assoc Prox (or other descriptor object that may
+                        # use _DCAttributeOptions) is usually here, except if
+                        # 1. we're a
+                        # dataclass, dataclasses would have removed the
+                        # attr here or 2. assoc proxy is coming from a
+                        # superclass, we want it to be direct here so it
+                        # tracks state or 3. assoc prox comes from
+                        # declared_attr, uncommon case
+                        setattr(cls, k, value)
+                        continue
+
+            our_stuff[k] = value
+
+    def _extract_declared_columns(self) -> None:
+        our_stuff = self.properties
+
+        # extract columns from the class dict
+        declared_columns = self.declared_columns
+        column_ordering = self.column_ordering
+        name_to_prop_key = collections.defaultdict(set)
+
+        for key, c in list(our_stuff.items()):
+            if isinstance(c, _MapsColumns):
+                mp_to_assign = c.mapper_property_to_assign
+                if mp_to_assign:
+                    our_stuff[key] = mp_to_assign
+                else:
+                    # if no mapper property to assign, this currently means
+                    # this is a MappedColumn that will produce a Column for us
+                    del our_stuff[key]
+
+                for col, sort_order in c.columns_to_assign:
+                    if not isinstance(c, CompositeProperty):
+                        name_to_prop_key[col.name].add(key)
+                    declared_columns.add(col)
+
+                    # we would assert this, however we want the below
+                    # warning to take effect instead.  See #9630
+                    # assert col not in column_ordering
+
+                    column_ordering[col] = sort_order
+
+                    # if this is a MappedColumn and the attribute key we
+                    # have is not what the column has for its key, map the
+                    # Column explicitly under the attribute key name.
+                    # otherwise, Mapper will map it under the column key.
+                    if mp_to_assign is None and key != col.key:
+                        our_stuff[key] = col
+            elif isinstance(c, Column):
+                # undefer previously occurred here, and now occurs earlier.
+                # ensure every column we get here has been named
+                assert c.name is not None
+                name_to_prop_key[c.name].add(key)
+                declared_columns.add(c)
+                # if the column is the same name as the key,
+                # remove it from the explicit properties dict.
+                # the normal rules for assigning column-based properties
+                # will take over, including precedence of columns
+                # in multi-column ColumnProperties.
+                if key == c.key:
+                    del our_stuff[key]
+
+        for name, keys in name_to_prop_key.items():
+            if len(keys) > 1:
+                util.warn(
+                    "On class %r, Column object %r named "
+                    "directly multiple times, "
+                    "only one will be used: %s. "
+                    "Consider using orm.synonym instead"
+                    % (self.classname, name, (", ".join(sorted(keys))))
+                )
+
+    def _setup_table(self, table: Optional[FromClause] = None) -> None:
+        cls = self.cls
+        cls_as_Decl = cast("MappedClassProtocol[Any]", cls)
+
+        tablename = self.tablename
+        table_args = self.table_args
+        clsdict_view = self.clsdict_view
+        declared_columns = self.declared_columns
+        column_ordering = self.column_ordering
+
+        manager = attributes.manager_of_class(cls)
+
+        if (
+            self.table_fn is None
+            and "__table__" not in clsdict_view
+            and table is None
+        ):
+            if hasattr(cls, "__table_cls__"):
+                table_cls = cast(
+                    Type[Table],
+                    util.unbound_method_to_callable(cls.__table_cls__),  # type: ignore  # noqa: E501
+                )
+            else:
+                table_cls = Table
+
+            if tablename is not None:
+                args: Tuple[Any, ...] = ()
+                table_kw: Dict[str, Any] = {}
+
+                if table_args:
+                    if isinstance(table_args, dict):
+                        table_kw = table_args
+                    elif isinstance(table_args, tuple):
+                        if isinstance(table_args[-1], dict):
+                            args, table_kw = table_args[0:-1], table_args[-1]
+                        else:
+                            args = table_args
+
+                autoload_with = clsdict_view.get("__autoload_with__")
+                if autoload_with:
+                    table_kw["autoload_with"] = autoload_with
+
+                autoload = clsdict_view.get("__autoload__")
+                if autoload:
+                    table_kw["autoload"] = True
+
+                sorted_columns = sorted(
+                    declared_columns,
+                    key=lambda c: column_ordering.get(c, 0),
+                )
+                table = self.set_cls_attribute(
+                    "__table__",
+                    table_cls(
+                        tablename,
+                        self._metadata_for_cls(manager),
+                        *sorted_columns,
+                        *args,
+                        **table_kw,
+                    ),
+                )
+        else:
+            if table is None:
+                if self.table_fn:
+                    table = self.set_cls_attribute(
+                        "__table__", self.table_fn()
+                    )
+                else:
+                    table = cls_as_Decl.__table__
+            if declared_columns:
+                for c in declared_columns:
+                    if not table.c.contains_column(c):
+                        raise exc.ArgumentError(
+                            "Can't add additional column %r when "
+                            "specifying __table__" % c.key
+                        )
+
+        self.local_table = table
+
+    def _metadata_for_cls(self, manager: ClassManager[Any]) -> MetaData:
+        meta: Optional[MetaData] = getattr(self.cls, "metadata", None)
+        if meta is not None:
+            return meta
+        else:
+            return manager.registry.metadata
+
+    def _setup_inheriting_mapper(self, mapper_kw: _MapperKwArgs) -> None:
+        cls = self.cls
+
+        inherits = mapper_kw.get("inherits", None)
+
+        if inherits is None:
+            # since we search for classical mappings now, search for
+            # multiple mapped bases as well and raise an error.
+            inherits_search = []
+            for base_ in cls.__bases__:
+                c = _resolve_for_abstract_or_classical(base_)
+                if c is None:
+                    continue
+
+                if _is_supercls_for_inherits(c) and c not in inherits_search:
+                    inherits_search.append(c)
+
+            if inherits_search:
+                if len(inherits_search) > 1:
+                    raise exc.InvalidRequestError(
+                        "Class %s has multiple mapped bases: %r"
+                        % (cls, inherits_search)
+                    )
+                inherits = inherits_search[0]
+        elif isinstance(inherits, Mapper):
+            inherits = inherits.class_
+
+        self.inherits = inherits
+
+        clsdict_view = self.clsdict_view
+        if "__table__" not in clsdict_view and self.tablename is None:
+            self.single = True
+
+    def _setup_inheriting_columns(self, mapper_kw: _MapperKwArgs) -> None:
+        table = self.local_table
+        cls = self.cls
+        table_args = self.table_args
+        declared_columns = self.declared_columns
+
+        if (
+            table is None
+            and self.inherits is None
+            and not _get_immediate_cls_attr(cls, "__no_table__")
+        ):
+            raise exc.InvalidRequestError(
+                "Class %r does not have a __table__ or __tablename__ "
+                "specified and does not inherit from an existing "
+                "table-mapped class." % cls
+            )
+        elif self.inherits:
+            inherited_mapper_or_config = _declared_mapping_info(self.inherits)
+            assert inherited_mapper_or_config is not None
+            inherited_table = inherited_mapper_or_config.local_table
+            inherited_persist_selectable = (
+                inherited_mapper_or_config.persist_selectable
+            )
+
+            if table is None:
+                # single table inheritance.
+                # ensure no table args
+                if table_args:
+                    raise exc.ArgumentError(
+                        "Can't place __table_args__ on an inherited class "
+                        "with no table."
+                    )
+
+                # add any columns declared here to the inherited table.
+                if declared_columns and not isinstance(inherited_table, Table):
+                    raise exc.ArgumentError(
+                        f"Can't declare columns on single-table-inherited "
+                        f"subclass {self.cls}; superclass {self.inherits} "
+                        "is not mapped to a Table"
+                    )
+
+                for col in declared_columns:
+                    assert inherited_table is not None
+                    if col.name in inherited_table.c:
+                        if inherited_table.c[col.name] is col:
+                            continue
+                        raise exc.ArgumentError(
+                            f"Column '{col}' on class {cls.__name__} "
+                            f"conflicts with existing column "
+                            f"'{inherited_table.c[col.name]}'.  If using "
+                            f"Declarative, consider using the "
+                            "use_existing_column parameter of mapped_column() "
+                            "to resolve conflicts."
+                        )
+                    if col.primary_key:
+                        raise exc.ArgumentError(
+                            "Can't place primary key columns on an inherited "
+                            "class with no table."
+                        )
+
+                    if TYPE_CHECKING:
+                        assert isinstance(inherited_table, Table)
+
+                    inherited_table.append_column(col)
+                    if (
+                        inherited_persist_selectable is not None
+                        and inherited_persist_selectable is not inherited_table
+                    ):
+                        inherited_persist_selectable._refresh_for_new_column(
+                            col
+                        )
+
+    def _prepare_mapper_arguments(self, mapper_kw: _MapperKwArgs) -> None:
+        properties = self.properties
+
+        if self.mapper_args_fn:
+            mapper_args = self.mapper_args_fn()
+        else:
+            mapper_args = {}
+
+        if mapper_kw:
+            mapper_args.update(mapper_kw)
+
+        if "properties" in mapper_args:
+            properties = dict(properties)
+            properties.update(mapper_args["properties"])
+
+        # make sure that column copies are used rather
+        # than the original columns from any mixins
+        for k in ("version_id_col", "polymorphic_on"):
+            if k in mapper_args:
+                v = mapper_args[k]
+                mapper_args[k] = self.column_copies.get(v, v)
+
+        if "primary_key" in mapper_args:
+            mapper_args["primary_key"] = [
+                self.column_copies.get(v, v)
+                for v in util.to_list(mapper_args["primary_key"])
+            ]
+
+        if "inherits" in mapper_args:
+            inherits_arg = mapper_args["inherits"]
+            if isinstance(inherits_arg, Mapper):
+                inherits_arg = inherits_arg.class_
+
+            if inherits_arg is not self.inherits:
+                raise exc.InvalidRequestError(
+                    "mapper inherits argument given for non-inheriting "
+                    "class %s" % (mapper_args["inherits"])
+                )
+
+        if self.inherits:
+            mapper_args["inherits"] = self.inherits
+
+        if self.inherits and not mapper_args.get("concrete", False):
+            # note the superclass is expected to have a Mapper assigned and
+            # not be a deferred config, as this is called within map()
+            inherited_mapper = class_mapper(self.inherits, False)
+            inherited_table = inherited_mapper.local_table
+
+            # single or joined inheritance
+            # exclude any cols on the inherited table which are
+            # not mapped on the parent class, to avoid
+            # mapping columns specific to sibling/nephew classes
+            if "exclude_properties" not in mapper_args:
+                mapper_args["exclude_properties"] = exclude_properties = {
+                    c.key
+                    for c in inherited_table.c
+                    if c not in inherited_mapper._columntoproperty
+                }.union(inherited_mapper.exclude_properties or ())
+                exclude_properties.difference_update(
+                    [c.key for c in self.declared_columns]
+                )
+
+            # look through columns in the current mapper that
+            # are keyed to a propname different than the colname
+            # (if names were the same, we'd have popped it out above,
+            # in which case the mapper makes this combination).
+            # See if the superclass has a similar column property.
+            # If so, join them together.
+            for k, col in list(properties.items()):
+                if not isinstance(col, expression.ColumnElement):
+                    continue
+                if k in inherited_mapper._props:
+                    p = inherited_mapper._props[k]
+                    if isinstance(p, ColumnProperty):
+                        # note here we place the subclass column
+                        # first.  See [ticket:1892] for background.
+                        properties[k] = [col] + p.columns
+        result_mapper_args = mapper_args.copy()
+        result_mapper_args["properties"] = properties
+        self.mapper_args = result_mapper_args
+
+    def map(self, mapper_kw: _MapperKwArgs = util.EMPTY_DICT) -> Mapper[Any]:
+        self._prepare_mapper_arguments(mapper_kw)
+        if hasattr(self.cls, "__mapper_cls__"):
+            mapper_cls = cast(
+                "Type[Mapper[Any]]",
+                util.unbound_method_to_callable(
+                    self.cls.__mapper_cls__  # type: ignore
+                ),
+            )
+        else:
+            mapper_cls = Mapper
+
+        return self.set_cls_attribute(
+            "__mapper__",
+            mapper_cls(self.cls, self.local_table, **self.mapper_args),
+        )
+
+
+@util.preload_module("sqlalchemy.orm.decl_api")
+def _as_dc_declaredattr(
+    field_metadata: Mapping[str, Any], sa_dataclass_metadata_key: str
+) -> Any:
+    # wrap lambdas inside dataclass fields inside an ad-hoc declared_attr.
+    # we can't write it because field.metadata is immutable :( so we have
+    # to go through extra trouble to compare these
+    decl_api = util.preloaded.orm_decl_api
+    obj = field_metadata[sa_dataclass_metadata_key]
+    if callable(obj) and not isinstance(obj, decl_api.declared_attr):
+        return decl_api.declared_attr(obj)
+    else:
+        return obj
+
+
+class _DeferredMapperConfig(_ClassScanMapperConfig):
+    _cls: weakref.ref[Type[Any]]
+
+    is_deferred = True
+
+    _configs: util.OrderedDict[
+        weakref.ref[Type[Any]], _DeferredMapperConfig
+    ] = util.OrderedDict()
+
+    def _early_mapping(self, mapper_kw: _MapperKwArgs) -> None:
+        pass
+
+    # mypy disallows plain property override of variable
+    @property  # type: ignore
+    def cls(self) -> Type[Any]:
+        return self._cls()  # type: ignore
+
+    @cls.setter
+    def cls(self, class_: Type[Any]) -> None:
+        self._cls = weakref.ref(class_, self._remove_config_cls)
+        self._configs[self._cls] = self
+
+    @classmethod
+    def _remove_config_cls(cls, ref: weakref.ref[Type[Any]]) -> None:
+        cls._configs.pop(ref, None)
+
+    @classmethod
+    def has_cls(cls, class_: Type[Any]) -> bool:
+        # 2.6 fails on weakref if class_ is an old style class
+        return isinstance(class_, type) and weakref.ref(class_) in cls._configs
+
+    @classmethod
+    def raise_unmapped_for_cls(cls, class_: Type[Any]) -> NoReturn:
+        if hasattr(class_, "_sa_raise_deferred_config"):
+            class_._sa_raise_deferred_config()
+
+        raise orm_exc.UnmappedClassError(
+            class_,
+            msg=(
+                f"Class {orm_exc._safe_cls_name(class_)} has a deferred "
+                "mapping on it.  It is not yet usable as a mapped class."
+            ),
+        )
+
+    @classmethod
+    def config_for_cls(cls, class_: Type[Any]) -> _DeferredMapperConfig:
+        return cls._configs[weakref.ref(class_)]
+
+    @classmethod
+    def classes_for_base(
+        cls, base_cls: Type[Any], sort: bool = True
+    ) -> List[_DeferredMapperConfig]:
+        classes_for_base = [
+            m
+            for m, cls_ in [(m, m.cls) for m in cls._configs.values()]
+            if cls_ is not None and issubclass(cls_, base_cls)
+        ]
+
+        if not sort:
+            return classes_for_base
+
+        all_m_by_cls = {m.cls: m for m in classes_for_base}
+
+        tuples: List[Tuple[_DeferredMapperConfig, _DeferredMapperConfig]] = []
+        for m_cls in all_m_by_cls:
+            tuples.extend(
+                (all_m_by_cls[base_cls], all_m_by_cls[m_cls])
+                for base_cls in m_cls.__bases__
+                if base_cls in all_m_by_cls
+            )
+        return list(topological.sort(tuples, classes_for_base))
+
+    def map(self, mapper_kw: _MapperKwArgs = util.EMPTY_DICT) -> Mapper[Any]:
+        self._configs.pop(self._cls, None)
+        return super().map(mapper_kw)
+
+
+def _add_attribute(
+    cls: Type[Any], key: str, value: MapperProperty[Any]
+) -> None:
+    """add an attribute to an existing declarative class.
+
+    This runs through the logic to determine MapperProperty,
+    adds it to the Mapper, adds a column to the mapped Table, etc.
+
+    """
+
+    if "__mapper__" in cls.__dict__:
+        mapped_cls = cast("MappedClassProtocol[Any]", cls)
+
+        def _table_or_raise(mc: MappedClassProtocol[Any]) -> Table:
+            if isinstance(mc.__table__, Table):
+                return mc.__table__
+            raise exc.InvalidRequestError(
+                f"Cannot add a new attribute to mapped class {mc.__name__!r} "
+                "because it's not mapped against a table."
+            )
+
+        if isinstance(value, Column):
+            _undefer_column_name(key, value)
+            _table_or_raise(mapped_cls).append_column(
+                value, replace_existing=True
+            )
+            mapped_cls.__mapper__.add_property(key, value)
+        elif isinstance(value, _MapsColumns):
+            mp = value.mapper_property_to_assign
+            for col, _ in value.columns_to_assign:
+                _undefer_column_name(key, col)
+                _table_or_raise(mapped_cls).append_column(
+                    col, replace_existing=True
+                )
+                if not mp:
+                    mapped_cls.__mapper__.add_property(key, col)
+            if mp:
+                mapped_cls.__mapper__.add_property(key, mp)
+        elif isinstance(value, MapperProperty):
+            mapped_cls.__mapper__.add_property(key, value)
+        elif isinstance(value, QueryableAttribute) and value.key != key:
+            # detect a QueryableAttribute that's already mapped being
+            # assigned elsewhere in userland, turn into a synonym()
+            value = SynonymProperty(value.key)
+            mapped_cls.__mapper__.add_property(key, value)
+        else:
+            type.__setattr__(cls, key, value)
+            mapped_cls.__mapper__._expire_memoizations()
+    else:
+        type.__setattr__(cls, key, value)
+
+
+def _del_attribute(cls: Type[Any], key: str) -> None:
+    if (
+        "__mapper__" in cls.__dict__
+        and key in cls.__dict__
+        and not cast(
+            "MappedClassProtocol[Any]", cls
+        ).__mapper__._dispose_called
+    ):
+        value = cls.__dict__[key]
+        if isinstance(
+            value, (Column, _MapsColumns, MapperProperty, QueryableAttribute)
+        ):
+            raise NotImplementedError(
+                "Can't un-map individual mapped attributes on a mapped class."
+            )
+        else:
+            type.__delattr__(cls, key)
+            cast(
+                "MappedClassProtocol[Any]", cls
+            ).__mapper__._expire_memoizations()
+    else:
+        type.__delattr__(cls, key)
+
+
+def _declarative_constructor(self: Any, **kwargs: Any) -> None:
+    """A simple constructor that allows initialization from kwargs.
+
+    Sets attributes on the constructed instance using the names and
+    values in ``kwargs``.
+
+    Only keys that are present as
+    attributes of the instance's class are allowed. These could be,
+    for example, any mapped columns or relationships.
+    """
+    cls_ = type(self)
+    for k in kwargs:
+        if not hasattr(cls_, k):
+            raise TypeError(
+                "%r is an invalid keyword argument for %s" % (k, cls_.__name__)
+            )
+        setattr(self, k, kwargs[k])
+
+
+_declarative_constructor.__name__ = "__init__"
+
+
+def _undefer_column_name(key: str, column: Column[Any]) -> None:
+    if column.key is None:
+        column.key = key
+    if column.name is None:
+        column.name = key
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dependency.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dependency.py
new file mode 100644
index 00000000..59530624
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dependency.py
@@ -0,0 +1,1304 @@
+# orm/dependency.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Relationship dependencies.
+
+"""
+
+from __future__ import annotations
+
+from . import attributes
+from . import exc
+from . import sync
+from . import unitofwork
+from . import util as mapperutil
+from .interfaces import MANYTOMANY
+from .interfaces import MANYTOONE
+from .interfaces import ONETOMANY
+from .. import exc as sa_exc
+from .. import sql
+from .. import util
+
+
+class DependencyProcessor:
+    def __init__(self, prop):
+        self.prop = prop
+        self.cascade = prop.cascade
+        self.mapper = prop.mapper
+        self.parent = prop.parent
+        self.secondary = prop.secondary
+        self.direction = prop.direction
+        self.post_update = prop.post_update
+        self.passive_deletes = prop.passive_deletes
+        self.passive_updates = prop.passive_updates
+        self.enable_typechecks = prop.enable_typechecks
+        if self.passive_deletes:
+            self._passive_delete_flag = attributes.PASSIVE_NO_INITIALIZE
+        else:
+            self._passive_delete_flag = attributes.PASSIVE_OFF
+        if self.passive_updates:
+            self._passive_update_flag = attributes.PASSIVE_NO_INITIALIZE
+        else:
+            self._passive_update_flag = attributes.PASSIVE_OFF
+
+        self.sort_key = "%s_%s" % (self.parent._sort_key, prop.key)
+        self.key = prop.key
+        if not self.prop.synchronize_pairs:
+            raise sa_exc.ArgumentError(
+                "Can't build a DependencyProcessor for relationship %s. "
+                "No target attributes to populate between parent and "
+                "child are present" % self.prop
+            )
+
+    @classmethod
+    def from_relationship(cls, prop):
+        return _direction_to_processor[prop.direction](prop)
+
+    def hasparent(self, state):
+        """return True if the given object instance has a parent,
+        according to the ``InstrumentedAttribute`` handled by this
+        ``DependencyProcessor``.
+
+        """
+        return self.parent.class_manager.get_impl(self.key).hasparent(state)
+
+    def per_property_preprocessors(self, uow):
+        """establish actions and dependencies related to a flush.
+
+        These actions will operate on all relevant states in
+        the aggregate.
+
+        """
+        uow.register_preprocessor(self, True)
+
+    def per_property_flush_actions(self, uow):
+        after_save = unitofwork.ProcessAll(uow, self, False, True)
+        before_delete = unitofwork.ProcessAll(uow, self, True, True)
+
+        parent_saves = unitofwork.SaveUpdateAll(
+            uow, self.parent.primary_base_mapper
+        )
+        child_saves = unitofwork.SaveUpdateAll(
+            uow, self.mapper.primary_base_mapper
+        )
+
+        parent_deletes = unitofwork.DeleteAll(
+            uow, self.parent.primary_base_mapper
+        )
+        child_deletes = unitofwork.DeleteAll(
+            uow, self.mapper.primary_base_mapper
+        )
+
+        self.per_property_dependencies(
+            uow,
+            parent_saves,
+            child_saves,
+            parent_deletes,
+            child_deletes,
+            after_save,
+            before_delete,
+        )
+
+    def per_state_flush_actions(self, uow, states, isdelete):
+        """establish actions and dependencies related to a flush.
+
+        These actions will operate on all relevant states
+        individually.    This occurs only if there are cycles
+        in the 'aggregated' version of events.
+
+        """
+
+        child_base_mapper = self.mapper.primary_base_mapper
+        child_saves = unitofwork.SaveUpdateAll(uow, child_base_mapper)
+        child_deletes = unitofwork.DeleteAll(uow, child_base_mapper)
+
+        # locate and disable the aggregate processors
+        # for this dependency
+
+        if isdelete:
+            before_delete = unitofwork.ProcessAll(uow, self, True, True)
+            before_delete.disabled = True
+        else:
+            after_save = unitofwork.ProcessAll(uow, self, False, True)
+            after_save.disabled = True
+
+        # check if the "child" side is part of the cycle
+
+        if child_saves not in uow.cycles:
+            # based on the current dependencies we use, the saves/
+            # deletes should always be in the 'cycles' collection
+            # together.   if this changes, we will have to break up
+            # this method a bit more.
+            assert child_deletes not in uow.cycles
+
+            # child side is not part of the cycle, so we will link per-state
+            # actions to the aggregate "saves", "deletes" actions
+            child_actions = [(child_saves, False), (child_deletes, True)]
+            child_in_cycles = False
+        else:
+            child_in_cycles = True
+
+        # check if the "parent" side is part of the cycle
+        if not isdelete:
+            parent_saves = unitofwork.SaveUpdateAll(
+                uow, self.parent.base_mapper
+            )
+            parent_deletes = before_delete = None
+            if parent_saves in uow.cycles:
+                parent_in_cycles = True
+        else:
+            parent_deletes = unitofwork.DeleteAll(uow, self.parent.base_mapper)
+            parent_saves = after_save = None
+            if parent_deletes in uow.cycles:
+                parent_in_cycles = True
+
+        # now create actions /dependencies for each state.
+
+        for state in states:
+            # detect if there's anything changed or loaded
+            # by a preprocessor on this state/attribute.   In the
+            # case of deletes we may try to load missing items here as well.
+            sum_ = state.manager[self.key].impl.get_all_pending(
+                state,
+                state.dict,
+                (
+                    self._passive_delete_flag
+                    if isdelete
+                    else attributes.PASSIVE_NO_INITIALIZE
+                ),
+            )
+
+            if not sum_:
+                continue
+
+            if isdelete:
+                before_delete = unitofwork.ProcessState(uow, self, True, state)
+                if parent_in_cycles:
+                    parent_deletes = unitofwork.DeleteState(uow, state)
+            else:
+                after_save = unitofwork.ProcessState(uow, self, False, state)
+                if parent_in_cycles:
+                    parent_saves = unitofwork.SaveUpdateState(uow, state)
+
+            if child_in_cycles:
+                child_actions = []
+                for child_state, child in sum_:
+                    if child_state not in uow.states:
+                        child_action = (None, None)
+                    else:
+                        (deleted, listonly) = uow.states[child_state]
+                        if deleted:
+                            child_action = (
+                                unitofwork.DeleteState(uow, child_state),
+                                True,
+                            )
+                        else:
+                            child_action = (
+                                unitofwork.SaveUpdateState(uow, child_state),
+                                False,
+                            )
+                    child_actions.append(child_action)
+
+            # establish dependencies between our possibly per-state
+            # parent action and our possibly per-state child action.
+            for child_action, childisdelete in child_actions:
+                self.per_state_dependencies(
+                    uow,
+                    parent_saves,
+                    parent_deletes,
+                    child_action,
+                    after_save,
+                    before_delete,
+                    isdelete,
+                    childisdelete,
+                )
+
+    def presort_deletes(self, uowcommit, states):
+        return False
+
+    def presort_saves(self, uowcommit, states):
+        return False
+
+    def process_deletes(self, uowcommit, states):
+        pass
+
+    def process_saves(self, uowcommit, states):
+        pass
+
+    def prop_has_changes(self, uowcommit, states, isdelete):
+        if not isdelete or self.passive_deletes:
+            passive = (
+                attributes.PASSIVE_NO_INITIALIZE
+                | attributes.INCLUDE_PENDING_MUTATIONS
+            )
+        elif self.direction is MANYTOONE:
+            # here, we were hoping to optimize having to fetch many-to-one
+            # for history and ignore it, if there's no further cascades
+            # to take place.  however there are too many less common conditions
+            # that still take place and tests in test_relationships /
+            # test_cascade etc. will still fail.
+            passive = attributes.PASSIVE_NO_FETCH_RELATED
+        else:
+            passive = (
+                attributes.PASSIVE_OFF | attributes.INCLUDE_PENDING_MUTATIONS
+            )
+
+        for s in states:
+            # TODO: add a high speed method
+            # to InstanceState which returns:  attribute
+            # has a non-None value, or had one
+            history = uowcommit.get_attribute_history(s, self.key, passive)
+            if history and not history.empty():
+                return True
+        else:
+            return (
+                states
+                and not self.prop._is_self_referential
+                and self.mapper in uowcommit.mappers
+            )
+
+    def _verify_canload(self, state):
+        if self.prop.uselist and state is None:
+            raise exc.FlushError(
+                "Can't flush None value found in "
+                "collection %s" % (self.prop,)
+            )
+        elif state is not None and not self.mapper._canload(
+            state, allow_subtypes=not self.enable_typechecks
+        ):
+            if self.mapper._canload(state, allow_subtypes=True):
+                raise exc.FlushError(
+                    "Attempting to flush an item of type "
+                    "%(x)s as a member of collection "
+                    '"%(y)s". Expected an object of type '
+                    "%(z)s or a polymorphic subclass of "
+                    "this type. If %(x)s is a subclass of "
+                    '%(z)s, configure mapper "%(zm)s" to '
+                    "load this subtype polymorphically, or "
+                    "set enable_typechecks=False to allow "
+                    "any subtype to be accepted for flush. "
+                    % {
+                        "x": state.class_,
+                        "y": self.prop,
+                        "z": self.mapper.class_,
+                        "zm": self.mapper,
+                    }
+                )
+            else:
+                raise exc.FlushError(
+                    "Attempting to flush an item of type "
+                    "%(x)s as a member of collection "
+                    '"%(y)s". Expected an object of type '
+                    "%(z)s or a polymorphic subclass of "
+                    "this type."
+                    % {
+                        "x": state.class_,
+                        "y": self.prop,
+                        "z": self.mapper.class_,
+                    }
+                )
+
+    def _synchronize(self, state, child, associationrow, clearkeys, uowcommit):
+        raise NotImplementedError()
+
+    def _get_reversed_processed_set(self, uow):
+        if not self.prop._reverse_property:
+            return None
+
+        process_key = tuple(
+            sorted([self.key] + [p.key for p in self.prop._reverse_property])
+        )
+        return uow.memo(("reverse_key", process_key), set)
+
+    def _post_update(self, state, uowcommit, related, is_m2o_delete=False):
+        for x in related:
+            if not is_m2o_delete or x is not None:
+                uowcommit.register_post_update(
+                    state, [r for l, r in self.prop.synchronize_pairs]
+                )
+                break
+
+    def _pks_changed(self, uowcommit, state):
+        raise NotImplementedError()
+
+    def __repr__(self):
+        return "%s(%s)" % (self.__class__.__name__, self.prop)
+
+
+class OneToManyDP(DependencyProcessor):
+    def per_property_dependencies(
+        self,
+        uow,
+        parent_saves,
+        child_saves,
+        parent_deletes,
+        child_deletes,
+        after_save,
+        before_delete,
+    ):
+        if self.post_update:
+            child_post_updates = unitofwork.PostUpdateAll(
+                uow, self.mapper.primary_base_mapper, False
+            )
+            child_pre_updates = unitofwork.PostUpdateAll(
+                uow, self.mapper.primary_base_mapper, True
+            )
+
+            uow.dependencies.update(
+                [
+                    (child_saves, after_save),
+                    (parent_saves, after_save),
+                    (after_save, child_post_updates),
+                    (before_delete, child_pre_updates),
+                    (child_pre_updates, parent_deletes),
+                    (child_pre_updates, child_deletes),
+                ]
+            )
+        else:
+            uow.dependencies.update(
+                [
+                    (parent_saves, after_save),
+                    (after_save, child_saves),
+                    (after_save, child_deletes),
+                    (child_saves, parent_deletes),
+                    (child_deletes, parent_deletes),
+                    (before_delete, child_saves),
+                    (before_delete, child_deletes),
+                ]
+            )
+
+    def per_state_dependencies(
+        self,
+        uow,
+        save_parent,
+        delete_parent,
+        child_action,
+        after_save,
+        before_delete,
+        isdelete,
+        childisdelete,
+    ):
+        if self.post_update:
+            child_post_updates = unitofwork.PostUpdateAll(
+                uow, self.mapper.primary_base_mapper, False
+            )
+            child_pre_updates = unitofwork.PostUpdateAll(
+                uow, self.mapper.primary_base_mapper, True
+            )
+
+            # TODO: this whole block is not covered
+            # by any tests
+            if not isdelete:
+                if childisdelete:
+                    uow.dependencies.update(
+                        [
+                            (child_action, after_save),
+                            (after_save, child_post_updates),
+                        ]
+                    )
+                else:
+                    uow.dependencies.update(
+                        [
+                            (save_parent, after_save),
+                            (child_action, after_save),
+                            (after_save, child_post_updates),
+                        ]
+                    )
+            else:
+                if childisdelete:
+                    uow.dependencies.update(
+                        [
+                            (before_delete, child_pre_updates),
+                            (child_pre_updates, delete_parent),
+                        ]
+                    )
+                else:
+                    uow.dependencies.update(
+                        [
+                            (before_delete, child_pre_updates),
+                            (child_pre_updates, delete_parent),
+                        ]
+                    )
+        elif not isdelete:
+            uow.dependencies.update(
+                [
+                    (save_parent, after_save),
+                    (after_save, child_action),
+                    (save_parent, child_action),
+                ]
+            )
+        else:
+            uow.dependencies.update(
+                [(before_delete, child_action), (child_action, delete_parent)]
+            )
+
+    def presort_deletes(self, uowcommit, states):
+        # head object is being deleted, and we manage its list of
+        # child objects the child objects have to have their
+        # foreign key to the parent set to NULL
+        should_null_fks = (
+            not self.cascade.delete and not self.passive_deletes == "all"
+        )
+
+        for state in states:
+            history = uowcommit.get_attribute_history(
+                state, self.key, self._passive_delete_flag
+            )
+            if history:
+                for child in history.deleted:
+                    if child is not None and self.hasparent(child) is False:
+                        if self.cascade.delete_orphan:
+                            uowcommit.register_object(child, isdelete=True)
+                        else:
+                            uowcommit.register_object(child)
+
+                if should_null_fks:
+                    for child in history.unchanged:
+                        if child is not None:
+                            uowcommit.register_object(
+                                child, operation="delete", prop=self.prop
+                            )
+
+    def presort_saves(self, uowcommit, states):
+        children_added = uowcommit.memo(("children_added", self), set)
+
+        should_null_fks = (
+            not self.cascade.delete_orphan
+            and not self.passive_deletes == "all"
+        )
+
+        for state in states:
+            pks_changed = self._pks_changed(uowcommit, state)
+
+            if not pks_changed or self.passive_updates:
+                passive = (
+                    attributes.PASSIVE_NO_INITIALIZE
+                    | attributes.INCLUDE_PENDING_MUTATIONS
+                )
+            else:
+                passive = (
+                    attributes.PASSIVE_OFF
+                    | attributes.INCLUDE_PENDING_MUTATIONS
+                )
+
+            history = uowcommit.get_attribute_history(state, self.key, passive)
+            if history:
+                for child in history.added:
+                    if child is not None:
+                        uowcommit.register_object(
+                            child,
+                            cancel_delete=True,
+                            operation="add",
+                            prop=self.prop,
+                        )
+
+                children_added.update(history.added)
+
+                for child in history.deleted:
+                    if not self.cascade.delete_orphan:
+                        if should_null_fks:
+                            uowcommit.register_object(
+                                child,
+                                isdelete=False,
+                                operation="delete",
+                                prop=self.prop,
+                            )
+                    elif self.hasparent(child) is False:
+                        uowcommit.register_object(
+                            child,
+                            isdelete=True,
+                            operation="delete",
+                            prop=self.prop,
+                        )
+                        for c, m, st_, dct_ in self.mapper.cascade_iterator(
+                            "delete", child
+                        ):
+                            uowcommit.register_object(st_, isdelete=True)
+
+            if pks_changed:
+                if history:
+                    for child in history.unchanged:
+                        if child is not None:
+                            uowcommit.register_object(
+                                child,
+                                False,
+                                self.passive_updates,
+                                operation="pk change",
+                                prop=self.prop,
+                            )
+
+    def process_deletes(self, uowcommit, states):
+        # head object is being deleted, and we manage its list of
+        # child objects the child objects have to have their foreign
+        # key to the parent set to NULL this phase can be called
+        # safely for any cascade but is unnecessary if delete cascade
+        # is on.
+
+        if self.post_update or not self.passive_deletes == "all":
+            children_added = uowcommit.memo(("children_added", self), set)
+
+            for state in states:
+                history = uowcommit.get_attribute_history(
+                    state, self.key, self._passive_delete_flag
+                )
+                if history:
+                    for child in history.deleted:
+                        if (
+                            child is not None
+                            and self.hasparent(child) is False
+                        ):
+                            self._synchronize(
+                                state, child, None, True, uowcommit, False
+                            )
+                            if self.post_update and child:
+                                self._post_update(child, uowcommit, [state])
+
+                    if self.post_update or not self.cascade.delete:
+                        for child in set(history.unchanged).difference(
+                            children_added
+                        ):
+                            if child is not None:
+                                self._synchronize(
+                                    state, child, None, True, uowcommit, False
+                                )
+                                if self.post_update and child:
+                                    self._post_update(
+                                        child, uowcommit, [state]
+                                    )
+
+                    # technically, we can even remove each child from the
+                    # collection here too.  but this would be a somewhat
+                    # inconsistent behavior since it wouldn't happen
+                    # if the old parent wasn't deleted but child was moved.
+
+    def process_saves(self, uowcommit, states):
+        should_null_fks = (
+            not self.cascade.delete_orphan
+            and not self.passive_deletes == "all"
+        )
+
+        for state in states:
+            history = uowcommit.get_attribute_history(
+                state, self.key, attributes.PASSIVE_NO_INITIALIZE
+            )
+            if history:
+                for child in history.added:
+                    self._synchronize(
+                        state, child, None, False, uowcommit, False
+                    )
+                    if child is not None and self.post_update:
+                        self._post_update(child, uowcommit, [state])
+
+                for child in history.deleted:
+                    if (
+                        should_null_fks
+                        and not self.cascade.delete_orphan
+                        and not self.hasparent(child)
+                    ):
+                        self._synchronize(
+                            state, child, None, True, uowcommit, False
+                        )
+
+                if self._pks_changed(uowcommit, state):
+                    for child in history.unchanged:
+                        self._synchronize(
+                            state, child, None, False, uowcommit, True
+                        )
+
+    def _synchronize(
+        self, state, child, associationrow, clearkeys, uowcommit, pks_changed
+    ):
+        source = state
+        dest = child
+        self._verify_canload(child)
+        if dest is None or (
+            not self.post_update and uowcommit.is_deleted(dest)
+        ):
+            return
+        if clearkeys:
+            sync.clear(dest, self.mapper, self.prop.synchronize_pairs)
+        else:
+            sync.populate(
+                source,
+                self.parent,
+                dest,
+                self.mapper,
+                self.prop.synchronize_pairs,
+                uowcommit,
+                self.passive_updates and pks_changed,
+            )
+
+    def _pks_changed(self, uowcommit, state):
+        return sync.source_modified(
+            uowcommit, state, self.parent, self.prop.synchronize_pairs
+        )
+
+
+class ManyToOneDP(DependencyProcessor):
+    def __init__(self, prop):
+        DependencyProcessor.__init__(self, prop)
+        for mapper in self.mapper.self_and_descendants:
+            mapper._dependency_processors.append(DetectKeySwitch(prop))
+
+    def per_property_dependencies(
+        self,
+        uow,
+        parent_saves,
+        child_saves,
+        parent_deletes,
+        child_deletes,
+        after_save,
+        before_delete,
+    ):
+        if self.post_update:
+            parent_post_updates = unitofwork.PostUpdateAll(
+                uow, self.parent.primary_base_mapper, False
+            )
+            parent_pre_updates = unitofwork.PostUpdateAll(
+                uow, self.parent.primary_base_mapper, True
+            )
+
+            uow.dependencies.update(
+                [
+                    (child_saves, after_save),
+                    (parent_saves, after_save),
+                    (after_save, parent_post_updates),
+                    (after_save, parent_pre_updates),
+                    (before_delete, parent_pre_updates),
+                    (parent_pre_updates, child_deletes),
+                    (parent_pre_updates, parent_deletes),
+                ]
+            )
+        else:
+            uow.dependencies.update(
+                [
+                    (child_saves, after_save),
+                    (after_save, parent_saves),
+                    (parent_saves, child_deletes),
+                    (parent_deletes, child_deletes),
+                ]
+            )
+
+    def per_state_dependencies(
+        self,
+        uow,
+        save_parent,
+        delete_parent,
+        child_action,
+        after_save,
+        before_delete,
+        isdelete,
+        childisdelete,
+    ):
+        if self.post_update:
+            if not isdelete:
+                parent_post_updates = unitofwork.PostUpdateAll(
+                    uow, self.parent.primary_base_mapper, False
+                )
+                if childisdelete:
+                    uow.dependencies.update(
+                        [
+                            (after_save, parent_post_updates),
+                            (parent_post_updates, child_action),
+                        ]
+                    )
+                else:
+                    uow.dependencies.update(
+                        [
+                            (save_parent, after_save),
+                            (child_action, after_save),
+                            (after_save, parent_post_updates),
+                        ]
+                    )
+            else:
+                parent_pre_updates = unitofwork.PostUpdateAll(
+                    uow, self.parent.primary_base_mapper, True
+                )
+
+                uow.dependencies.update(
+                    [
+                        (before_delete, parent_pre_updates),
+                        (parent_pre_updates, delete_parent),
+                        (parent_pre_updates, child_action),
+                    ]
+                )
+
+        elif not isdelete:
+            if not childisdelete:
+                uow.dependencies.update(
+                    [(child_action, after_save), (after_save, save_parent)]
+                )
+            else:
+                uow.dependencies.update([(after_save, save_parent)])
+
+        else:
+            if childisdelete:
+                uow.dependencies.update([(delete_parent, child_action)])
+
+    def presort_deletes(self, uowcommit, states):
+        if self.cascade.delete or self.cascade.delete_orphan:
+            for state in states:
+                history = uowcommit.get_attribute_history(
+                    state, self.key, self._passive_delete_flag
+                )
+                if history:
+                    if self.cascade.delete_orphan:
+                        todelete = history.sum()
+                    else:
+                        todelete = history.non_deleted()
+                    for child in todelete:
+                        if child is None:
+                            continue
+                        uowcommit.register_object(
+                            child,
+                            isdelete=True,
+                            operation="delete",
+                            prop=self.prop,
+                        )
+                        t = self.mapper.cascade_iterator("delete", child)
+                        for c, m, st_, dct_ in t:
+                            uowcommit.register_object(st_, isdelete=True)
+
+    def presort_saves(self, uowcommit, states):
+        for state in states:
+            uowcommit.register_object(state, operation="add", prop=self.prop)
+            if self.cascade.delete_orphan:
+                history = uowcommit.get_attribute_history(
+                    state, self.key, self._passive_delete_flag
+                )
+                if history:
+                    for child in history.deleted:
+                        if self.hasparent(child) is False:
+                            uowcommit.register_object(
+                                child,
+                                isdelete=True,
+                                operation="delete",
+                                prop=self.prop,
+                            )
+
+                            t = self.mapper.cascade_iterator("delete", child)
+                            for c, m, st_, dct_ in t:
+                                uowcommit.register_object(st_, isdelete=True)
+
+    def process_deletes(self, uowcommit, states):
+        if (
+            self.post_update
+            and not self.cascade.delete_orphan
+            and not self.passive_deletes == "all"
+        ):
+            # post_update means we have to update our
+            # row to not reference the child object
+            # before we can DELETE the row
+            for state in states:
+                self._synchronize(state, None, None, True, uowcommit)
+                if state and self.post_update:
+                    history = uowcommit.get_attribute_history(
+                        state, self.key, self._passive_delete_flag
+                    )
+                    if history:
+                        self._post_update(
+                            state, uowcommit, history.sum(), is_m2o_delete=True
+                        )
+
+    def process_saves(self, uowcommit, states):
+        for state in states:
+            history = uowcommit.get_attribute_history(
+                state, self.key, attributes.PASSIVE_NO_INITIALIZE
+            )
+            if history:
+                if history.added:
+                    for child in history.added:
+                        self._synchronize(
+                            state, child, None, False, uowcommit, "add"
+                        )
+                elif history.deleted:
+                    self._synchronize(
+                        state, None, None, True, uowcommit, "delete"
+                    )
+                if self.post_update:
+                    self._post_update(state, uowcommit, history.sum())
+
+    def _synchronize(
+        self,
+        state,
+        child,
+        associationrow,
+        clearkeys,
+        uowcommit,
+        operation=None,
+    ):
+        if state is None or (
+            not self.post_update and uowcommit.is_deleted(state)
+        ):
+            return
+
+        if (
+            operation is not None
+            and child is not None
+            and not uowcommit.session._contains_state(child)
+        ):
+            util.warn(
+                "Object of type %s not in session, %s "
+                "operation along '%s' won't proceed"
+                % (mapperutil.state_class_str(child), operation, self.prop)
+            )
+            return
+
+        if clearkeys or child is None:
+            sync.clear(state, self.parent, self.prop.synchronize_pairs)
+        else:
+            self._verify_canload(child)
+            sync.populate(
+                child,
+                self.mapper,
+                state,
+                self.parent,
+                self.prop.synchronize_pairs,
+                uowcommit,
+                False,
+            )
+
+
+class DetectKeySwitch(DependencyProcessor):
+    """For many-to-one relationships with no one-to-many backref,
+    searches for parents through the unit of work when a primary
+    key has changed and updates them.
+
+    Theoretically, this approach could be expanded to support transparent
+    deletion of objects referenced via many-to-one as well, although
+    the current attribute system doesn't do enough bookkeeping for this
+    to be efficient.
+
+    """
+
+    def per_property_preprocessors(self, uow):
+        if self.prop._reverse_property:
+            if self.passive_updates:
+                return
+            else:
+                if False in (
+                    prop.passive_updates
+                    for prop in self.prop._reverse_property
+                ):
+                    return
+
+        uow.register_preprocessor(self, False)
+
+    def per_property_flush_actions(self, uow):
+        parent_saves = unitofwork.SaveUpdateAll(uow, self.parent.base_mapper)
+        after_save = unitofwork.ProcessAll(uow, self, False, False)
+        uow.dependencies.update([(parent_saves, after_save)])
+
+    def per_state_flush_actions(self, uow, states, isdelete):
+        pass
+
+    def presort_deletes(self, uowcommit, states):
+        pass
+
+    def presort_saves(self, uow, states):
+        if not self.passive_updates:
+            # for non-passive updates, register in the preprocess stage
+            # so that mapper save_obj() gets a hold of changes
+            self._process_key_switches(states, uow)
+
+    def prop_has_changes(self, uow, states, isdelete):
+        if not isdelete and self.passive_updates:
+            d = self._key_switchers(uow, states)
+            return bool(d)
+
+        return False
+
+    def process_deletes(self, uowcommit, states):
+        assert False
+
+    def process_saves(self, uowcommit, states):
+        # for passive updates, register objects in the process stage
+        # so that we avoid ManyToOneDP's registering the object without
+        # the listonly flag in its own preprocess stage (results in UPDATE)
+        # statements being emitted
+        assert self.passive_updates
+        self._process_key_switches(states, uowcommit)
+
+    def _key_switchers(self, uow, states):
+        switched, notswitched = uow.memo(
+            ("pk_switchers", self), lambda: (set(), set())
+        )
+
+        allstates = switched.union(notswitched)
+        for s in states:
+            if s not in allstates:
+                if self._pks_changed(uow, s):
+                    switched.add(s)
+                else:
+                    notswitched.add(s)
+        return switched
+
+    def _process_key_switches(self, deplist, uowcommit):
+        switchers = self._key_switchers(uowcommit, deplist)
+        if switchers:
+            # if primary key values have actually changed somewhere, perform
+            # a linear search through the UOW in search of a parent.
+            for state in uowcommit.session.identity_map.all_states():
+                if not issubclass(state.class_, self.parent.class_):
+                    continue
+                dict_ = state.dict
+                related = state.get_impl(self.key).get(
+                    state, dict_, passive=self._passive_update_flag
+                )
+                if (
+                    related is not attributes.PASSIVE_NO_RESULT
+                    and related is not None
+                ):
+                    if self.prop.uselist:
+                        if not related:
+                            continue
+                        related_obj = related[0]
+                    else:
+                        related_obj = related
+                    related_state = attributes.instance_state(related_obj)
+                    if related_state in switchers:
+                        uowcommit.register_object(
+                            state, False, self.passive_updates
+                        )
+                        sync.populate(
+                            related_state,
+                            self.mapper,
+                            state,
+                            self.parent,
+                            self.prop.synchronize_pairs,
+                            uowcommit,
+                            self.passive_updates,
+                        )
+
+    def _pks_changed(self, uowcommit, state):
+        return bool(state.key) and sync.source_modified(
+            uowcommit, state, self.mapper, self.prop.synchronize_pairs
+        )
+
+
+class ManyToManyDP(DependencyProcessor):
+    def per_property_dependencies(
+        self,
+        uow,
+        parent_saves,
+        child_saves,
+        parent_deletes,
+        child_deletes,
+        after_save,
+        before_delete,
+    ):
+        uow.dependencies.update(
+            [
+                (parent_saves, after_save),
+                (child_saves, after_save),
+                (after_save, child_deletes),
+                # a rowswitch on the parent from  deleted to saved
+                # can make this one occur, as the "save" may remove
+                # an element from the
+                # "deleted" list before we have a chance to
+                # process its child rows
+                (before_delete, parent_saves),
+                (before_delete, parent_deletes),
+                (before_delete, child_deletes),
+                (before_delete, child_saves),
+            ]
+        )
+
+    def per_state_dependencies(
+        self,
+        uow,
+        save_parent,
+        delete_parent,
+        child_action,
+        after_save,
+        before_delete,
+        isdelete,
+        childisdelete,
+    ):
+        if not isdelete:
+            if childisdelete:
+                uow.dependencies.update(
+                    [(save_parent, after_save), (after_save, child_action)]
+                )
+            else:
+                uow.dependencies.update(
+                    [(save_parent, after_save), (child_action, after_save)]
+                )
+        else:
+            uow.dependencies.update(
+                [(before_delete, child_action), (before_delete, delete_parent)]
+            )
+
+    def presort_deletes(self, uowcommit, states):
+        # TODO: no tests fail if this whole
+        # thing is removed !!!!
+        if not self.passive_deletes:
+            # if no passive deletes, load history on
+            # the collection, so that prop_has_changes()
+            # returns True
+            for state in states:
+                uowcommit.get_attribute_history(
+                    state, self.key, self._passive_delete_flag
+                )
+
+    def presort_saves(self, uowcommit, states):
+        if not self.passive_updates:
+            # if no passive updates, load history on
+            # each collection where parent has changed PK,
+            # so that prop_has_changes() returns True
+            for state in states:
+                if self._pks_changed(uowcommit, state):
+                    history = uowcommit.get_attribute_history(
+                        state, self.key, attributes.PASSIVE_OFF
+                    )
+
+        if not self.cascade.delete_orphan:
+            return
+
+        # check for child items removed from the collection
+        # if delete_orphan check is turned on.
+        for state in states:
+            history = uowcommit.get_attribute_history(
+                state, self.key, attributes.PASSIVE_NO_INITIALIZE
+            )
+            if history:
+                for child in history.deleted:
+                    if self.hasparent(child) is False:
+                        uowcommit.register_object(
+                            child,
+                            isdelete=True,
+                            operation="delete",
+                            prop=self.prop,
+                        )
+                        for c, m, st_, dct_ in self.mapper.cascade_iterator(
+                            "delete", child
+                        ):
+                            uowcommit.register_object(st_, isdelete=True)
+
+    def process_deletes(self, uowcommit, states):
+        secondary_delete = []
+        secondary_insert = []
+        secondary_update = []
+
+        processed = self._get_reversed_processed_set(uowcommit)
+        tmp = set()
+        for state in states:
+            # this history should be cached already, as
+            # we loaded it in preprocess_deletes
+            history = uowcommit.get_attribute_history(
+                state, self.key, self._passive_delete_flag
+            )
+            if history:
+                for child in history.non_added():
+                    if child is None or (
+                        processed is not None and (state, child) in processed
+                    ):
+                        continue
+                    associationrow = {}
+                    if not self._synchronize(
+                        state,
+                        child,
+                        associationrow,
+                        False,
+                        uowcommit,
+                        "delete",
+                    ):
+                        continue
+                    secondary_delete.append(associationrow)
+
+                tmp.update((c, state) for c in history.non_added())
+
+        if processed is not None:
+            processed.update(tmp)
+
+        self._run_crud(
+            uowcommit, secondary_insert, secondary_update, secondary_delete
+        )
+
+    def process_saves(self, uowcommit, states):
+        secondary_delete = []
+        secondary_insert = []
+        secondary_update = []
+
+        processed = self._get_reversed_processed_set(uowcommit)
+        tmp = set()
+
+        for state in states:
+            need_cascade_pks = not self.passive_updates and self._pks_changed(
+                uowcommit, state
+            )
+            if need_cascade_pks:
+                passive = (
+                    attributes.PASSIVE_OFF
+                    | attributes.INCLUDE_PENDING_MUTATIONS
+                )
+            else:
+                passive = (
+                    attributes.PASSIVE_NO_INITIALIZE
+                    | attributes.INCLUDE_PENDING_MUTATIONS
+                )
+            history = uowcommit.get_attribute_history(state, self.key, passive)
+            if history:
+                for child in history.added:
+                    if processed is not None and (state, child) in processed:
+                        continue
+                    associationrow = {}
+                    if not self._synchronize(
+                        state, child, associationrow, False, uowcommit, "add"
+                    ):
+                        continue
+                    secondary_insert.append(associationrow)
+                for child in history.deleted:
+                    if processed is not None and (state, child) in processed:
+                        continue
+                    associationrow = {}
+                    if not self._synchronize(
+                        state,
+                        child,
+                        associationrow,
+                        False,
+                        uowcommit,
+                        "delete",
+                    ):
+                        continue
+                    secondary_delete.append(associationrow)
+
+                tmp.update((c, state) for c in history.added + history.deleted)
+
+                if need_cascade_pks:
+                    for child in history.unchanged:
+                        associationrow = {}
+                        sync.update(
+                            state,
+                            self.parent,
+                            associationrow,
+                            "old_",
+                            self.prop.synchronize_pairs,
+                        )
+                        sync.update(
+                            child,
+                            self.mapper,
+                            associationrow,
+                            "old_",
+                            self.prop.secondary_synchronize_pairs,
+                        )
+
+                        secondary_update.append(associationrow)
+
+        if processed is not None:
+            processed.update(tmp)
+
+        self._run_crud(
+            uowcommit, secondary_insert, secondary_update, secondary_delete
+        )
+
+    def _run_crud(
+        self, uowcommit, secondary_insert, secondary_update, secondary_delete
+    ):
+        connection = uowcommit.transaction.connection(self.mapper)
+
+        if secondary_delete:
+            associationrow = secondary_delete[0]
+            statement = self.secondary.delete().where(
+                sql.and_(
+                    *[
+                        c == sql.bindparam(c.key, type_=c.type)
+                        for c in self.secondary.c
+                        if c.key in associationrow
+                    ]
+                )
+            )
+            result = connection.execute(statement, secondary_delete)
+
+            if (
+                result.supports_sane_multi_rowcount()
+            ) and result.rowcount != len(secondary_delete):
+                raise exc.StaleDataError(
+                    "DELETE statement on table '%s' expected to delete "
+                    "%d row(s); Only %d were matched."
+                    % (
+                        self.secondary.description,
+                        len(secondary_delete),
+                        result.rowcount,
+                    )
+                )
+
+        if secondary_update:
+            associationrow = secondary_update[0]
+            statement = self.secondary.update().where(
+                sql.and_(
+                    *[
+                        c == sql.bindparam("old_" + c.key, type_=c.type)
+                        for c in self.secondary.c
+                        if c.key in associationrow
+                    ]
+                )
+            )
+            result = connection.execute(statement, secondary_update)
+
+            if (
+                result.supports_sane_multi_rowcount()
+            ) and result.rowcount != len(secondary_update):
+                raise exc.StaleDataError(
+                    "UPDATE statement on table '%s' expected to update "
+                    "%d row(s); Only %d were matched."
+                    % (
+                        self.secondary.description,
+                        len(secondary_update),
+                        result.rowcount,
+                    )
+                )
+
+        if secondary_insert:
+            statement = self.secondary.insert()
+            connection.execute(statement, secondary_insert)
+
+    def _synchronize(
+        self, state, child, associationrow, clearkeys, uowcommit, operation
+    ):
+        # this checks for None if uselist=True
+        self._verify_canload(child)
+
+        # but if uselist=False we get here.   If child is None,
+        # no association row can be generated, so return.
+        if child is None:
+            return False
+
+        if child is not None and not uowcommit.session._contains_state(child):
+            if not child.deleted:
+                util.warn(
+                    "Object of type %s not in session, %s "
+                    "operation along '%s' won't proceed"
+                    % (mapperutil.state_class_str(child), operation, self.prop)
+                )
+            return False
+
+        sync.populate_dict(
+            state, self.parent, associationrow, self.prop.synchronize_pairs
+        )
+        sync.populate_dict(
+            child,
+            self.mapper,
+            associationrow,
+            self.prop.secondary_synchronize_pairs,
+        )
+
+        return True
+
+    def _pks_changed(self, uowcommit, state):
+        return sync.source_modified(
+            uowcommit, state, self.parent, self.prop.synchronize_pairs
+        )
+
+
+_direction_to_processor = {
+    ONETOMANY: OneToManyDP,
+    MANYTOONE: ManyToOneDP,
+    MANYTOMANY: ManyToManyDP,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/descriptor_props.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/descriptor_props.py
new file mode 100644
index 00000000..f01cc178
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/descriptor_props.py
@@ -0,0 +1,1077 @@
+# orm/descriptor_props.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Descriptor properties are more "auxiliary" properties
+that exist as configurational elements, but don't participate
+as actively in the load/persist ORM loop.
+
+"""
+from __future__ import annotations
+
+from dataclasses import is_dataclass
+import inspect
+import itertools
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import util as orm_util
+from .base import _DeclarativeMapped
+from .base import LoaderCallableStatus
+from .base import Mapped
+from .base import PassiveFlag
+from .base import SQLORMOperations
+from .interfaces import _AttributeOptions
+from .interfaces import _IntrospectsAnnotations
+from .interfaces import _MapsColumns
+from .interfaces import MapperProperty
+from .interfaces import PropComparator
+from .util import _none_set
+from .util import de_stringify_annotation
+from .. import event
+from .. import exc as sa_exc
+from .. import schema
+from .. import sql
+from .. import util
+from ..sql import expression
+from ..sql import operators
+from ..sql.elements import BindParameter
+from ..util.typing import get_args
+from ..util.typing import is_fwd_ref
+from ..util.typing import is_pep593
+
+
+if typing.TYPE_CHECKING:
+    from ._typing import _InstanceDict
+    from ._typing import _RegistryType
+    from .attributes import History
+    from .attributes import InstrumentedAttribute
+    from .attributes import QueryableAttribute
+    from .context import ORMCompileState
+    from .decl_base import _ClassScanMapperConfig
+    from .mapper import Mapper
+    from .properties import ColumnProperty
+    from .properties import MappedColumn
+    from .state import InstanceState
+    from ..engine.base import Connection
+    from ..engine.row import Row
+    from ..sql._typing import _DMLColumnArgument
+    from ..sql._typing import _InfoType
+    from ..sql.elements import ClauseList
+    from ..sql.elements import ColumnElement
+    from ..sql.operators import OperatorType
+    from ..sql.schema import Column
+    from ..sql.selectable import Select
+    from ..util.typing import _AnnotationScanType
+    from ..util.typing import CallableReference
+    from ..util.typing import DescriptorReference
+    from ..util.typing import RODescriptorReference
+
+_T = TypeVar("_T", bound=Any)
+_PT = TypeVar("_PT", bound=Any)
+
+
+class DescriptorProperty(MapperProperty[_T]):
+    """:class:`.MapperProperty` which proxies access to a
+    user-defined descriptor."""
+
+    doc: Optional[str] = None
+
+    uses_objects = False
+    _links_to_entity = False
+
+    descriptor: DescriptorReference[Any]
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> History:
+        raise NotImplementedError()
+
+    def instrument_class(self, mapper: Mapper[Any]) -> None:
+        prop = self
+
+        class _ProxyImpl(attributes.AttributeImpl):
+            accepts_scalar_loader = False
+            load_on_unexpire = True
+            collection = False
+
+            @property
+            def uses_objects(self) -> bool:  # type: ignore
+                return prop.uses_objects
+
+            def __init__(self, key: str):
+                self.key = key
+
+            def get_history(
+                self,
+                state: InstanceState[Any],
+                dict_: _InstanceDict,
+                passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+            ) -> History:
+                return prop.get_history(state, dict_, passive)
+
+        if self.descriptor is None:
+            desc = getattr(mapper.class_, self.key, None)
+            if mapper._is_userland_descriptor(self.key, desc):
+                self.descriptor = desc
+
+        if self.descriptor is None:
+
+            def fset(obj: Any, value: Any) -> None:
+                setattr(obj, self.name, value)
+
+            def fdel(obj: Any) -> None:
+                delattr(obj, self.name)
+
+            def fget(obj: Any) -> Any:
+                return getattr(obj, self.name)
+
+            self.descriptor = property(fget=fget, fset=fset, fdel=fdel)
+
+        proxy_attr = attributes.create_proxied_attribute(self.descriptor)(
+            self.parent.class_,
+            self.key,
+            self.descriptor,
+            lambda: self._comparator_factory(mapper),
+            doc=self.doc,
+            original_property=self,
+        )
+        proxy_attr.impl = _ProxyImpl(self.key)
+        mapper.class_manager.instrument_attribute(self.key, proxy_attr)
+
+
+_CompositeAttrType = Union[
+    str,
+    "Column[_T]",
+    "MappedColumn[_T]",
+    "InstrumentedAttribute[_T]",
+    "Mapped[_T]",
+]
+
+
+_CC = TypeVar("_CC", bound=Any)
+
+
+_composite_getters: weakref.WeakKeyDictionary[
+    Type[Any], Callable[[Any], Tuple[Any, ...]]
+] = weakref.WeakKeyDictionary()
+
+
+class CompositeProperty(
+    _MapsColumns[_CC], _IntrospectsAnnotations, DescriptorProperty[_CC]
+):
+    """Defines a "composite" mapped attribute, representing a collection
+    of columns as one attribute.
+
+    :class:`.CompositeProperty` is constructed using the :func:`.composite`
+    function.
+
+    .. seealso::
+
+        :ref:`mapper_composite`
+
+    """
+
+    composite_class: Union[Type[_CC], Callable[..., _CC]]
+    attrs: Tuple[_CompositeAttrType[Any], ...]
+
+    _generated_composite_accessor: CallableReference[
+        Optional[Callable[[_CC], Tuple[Any, ...]]]
+    ]
+
+    comparator_factory: Type[Comparator[_CC]]
+
+    def __init__(
+        self,
+        _class_or_attr: Union[
+            None, Type[_CC], Callable[..., _CC], _CompositeAttrType[Any]
+        ] = None,
+        *attrs: _CompositeAttrType[Any],
+        attribute_options: Optional[_AttributeOptions] = None,
+        active_history: bool = False,
+        deferred: bool = False,
+        group: Optional[str] = None,
+        comparator_factory: Optional[Type[Comparator[_CC]]] = None,
+        info: Optional[_InfoType] = None,
+        **kwargs: Any,
+    ):
+        super().__init__(attribute_options=attribute_options)
+
+        if isinstance(_class_or_attr, (Mapped, str, sql.ColumnElement)):
+            self.attrs = (_class_or_attr,) + attrs
+            # will initialize within declarative_scan
+            self.composite_class = None  # type: ignore
+        else:
+            self.composite_class = _class_or_attr  # type: ignore
+            self.attrs = attrs
+
+        self.active_history = active_history
+        self.deferred = deferred
+        self.group = group
+        self.comparator_factory = (
+            comparator_factory
+            if comparator_factory is not None
+            else self.__class__.Comparator
+        )
+        self._generated_composite_accessor = None
+        if info is not None:
+            self.info.update(info)
+
+        util.set_creation_order(self)
+        self._create_descriptor()
+        self._init_accessor()
+
+    def instrument_class(self, mapper: Mapper[Any]) -> None:
+        super().instrument_class(mapper)
+        self._setup_event_handlers()
+
+    def _composite_values_from_instance(self, value: _CC) -> Tuple[Any, ...]:
+        if self._generated_composite_accessor:
+            return self._generated_composite_accessor(value)
+        else:
+            try:
+                accessor = value.__composite_values__
+            except AttributeError as ae:
+                raise sa_exc.InvalidRequestError(
+                    f"Composite class {self.composite_class.__name__} is not "
+                    f"a dataclass and does not define a __composite_values__()"
+                    " method; can't get state"
+                ) from ae
+            else:
+                return accessor()  # type: ignore
+
+    def do_init(self) -> None:
+        """Initialization which occurs after the :class:`.Composite`
+        has been associated with its parent mapper.
+
+        """
+        self._setup_arguments_on_columns()
+
+    _COMPOSITE_FGET = object()
+
+    def _create_descriptor(self) -> None:
+        """Create the Python descriptor that will serve as
+        the access point on instances of the mapped class.
+
+        """
+
+        def fget(instance: Any) -> Any:
+            dict_ = attributes.instance_dict(instance)
+            state = attributes.instance_state(instance)
+
+            if self.key not in dict_:
+                # key not present.  Iterate through related
+                # attributes, retrieve their values.  This
+                # ensures they all load.
+                values = [
+                    getattr(instance, key) for key in self._attribute_keys
+                ]
+
+                # current expected behavior here is that the composite is
+                # created on access if the object is persistent or if
+                # col attributes have non-None.  This would be better
+                # if the composite were created unconditionally,
+                # but that would be a behavioral change.
+                if self.key not in dict_ and (
+                    state.key is not None or not _none_set.issuperset(values)
+                ):
+                    dict_[self.key] = self.composite_class(*values)
+                    state.manager.dispatch.refresh(
+                        state, self._COMPOSITE_FGET, [self.key]
+                    )
+
+            return dict_.get(self.key, None)
+
+        def fset(instance: Any, value: Any) -> None:
+            dict_ = attributes.instance_dict(instance)
+            state = attributes.instance_state(instance)
+            attr = state.manager[self.key]
+
+            if attr.dispatch._active_history:
+                previous = fget(instance)
+            else:
+                previous = dict_.get(self.key, LoaderCallableStatus.NO_VALUE)
+
+            for fn in attr.dispatch.set:
+                value = fn(state, value, previous, attr.impl)
+            dict_[self.key] = value
+            if value is None:
+                for key in self._attribute_keys:
+                    setattr(instance, key, None)
+            else:
+                for key, value in zip(
+                    self._attribute_keys,
+                    self._composite_values_from_instance(value),
+                ):
+                    setattr(instance, key, value)
+
+        def fdel(instance: Any) -> None:
+            state = attributes.instance_state(instance)
+            dict_ = attributes.instance_dict(instance)
+            attr = state.manager[self.key]
+
+            if attr.dispatch._active_history:
+                previous = fget(instance)
+                dict_.pop(self.key, None)
+            else:
+                previous = dict_.pop(self.key, LoaderCallableStatus.NO_VALUE)
+
+            attr = state.manager[self.key]
+            attr.dispatch.remove(state, previous, attr.impl)
+            for key in self._attribute_keys:
+                setattr(instance, key, None)
+
+        self.descriptor = property(fget, fset, fdel)
+
+    @util.preload_module("sqlalchemy.orm.properties")
+    def declarative_scan(
+        self,
+        decl_scan: _ClassScanMapperConfig,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        mapped_container: Optional[Type[Mapped[Any]]],
+        annotation: Optional[_AnnotationScanType],
+        extracted_mapped_annotation: Optional[_AnnotationScanType],
+        is_dataclass_field: bool,
+    ) -> None:
+        MappedColumn = util.preloaded.orm_properties.MappedColumn
+        if (
+            self.composite_class is None
+            and extracted_mapped_annotation is None
+        ):
+            self._raise_for_required(key, cls)
+        argument = extracted_mapped_annotation
+
+        if is_pep593(argument):
+            argument = get_args(argument)[0]
+
+        if argument and self.composite_class is None:
+            if isinstance(argument, str) or is_fwd_ref(
+                argument, check_generic=True
+            ):
+                if originating_module is None:
+                    str_arg = (
+                        argument.__forward_arg__
+                        if hasattr(argument, "__forward_arg__")
+                        else str(argument)
+                    )
+                    raise sa_exc.ArgumentError(
+                        f"Can't use forward ref {argument} for composite "
+                        f"class argument; set up the type as Mapped[{str_arg}]"
+                    )
+                argument = de_stringify_annotation(
+                    cls, argument, originating_module, include_generic=True
+                )
+
+            self.composite_class = argument
+
+        if is_dataclass(self.composite_class):
+            self._setup_for_dataclass(registry, cls, originating_module, key)
+        else:
+            for attr in self.attrs:
+                if (
+                    isinstance(attr, (MappedColumn, schema.Column))
+                    and attr.name is None
+                ):
+                    raise sa_exc.ArgumentError(
+                        "Composite class column arguments must be named "
+                        "unless a dataclass is used"
+                    )
+        self._init_accessor()
+
+    def _init_accessor(self) -> None:
+        if is_dataclass(self.composite_class) and not hasattr(
+            self.composite_class, "__composite_values__"
+        ):
+            insp = inspect.signature(self.composite_class)
+            getter = operator.attrgetter(
+                *[p.name for p in insp.parameters.values()]
+            )
+            if len(insp.parameters) == 1:
+                self._generated_composite_accessor = lambda obj: (getter(obj),)
+            else:
+                self._generated_composite_accessor = getter
+
+        if (
+            self.composite_class is not None
+            and isinstance(self.composite_class, type)
+            and self.composite_class not in _composite_getters
+        ):
+            if self._generated_composite_accessor is not None:
+                _composite_getters[self.composite_class] = (
+                    self._generated_composite_accessor
+                )
+            elif hasattr(self.composite_class, "__composite_values__"):
+                _composite_getters[self.composite_class] = (
+                    lambda obj: obj.__composite_values__()
+                )
+
+    @util.preload_module("sqlalchemy.orm.properties")
+    @util.preload_module("sqlalchemy.orm.decl_base")
+    def _setup_for_dataclass(
+        self,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+    ) -> None:
+        MappedColumn = util.preloaded.orm_properties.MappedColumn
+
+        decl_base = util.preloaded.orm_decl_base
+
+        insp = inspect.signature(self.composite_class)
+        for param, attr in itertools.zip_longest(
+            insp.parameters.values(), self.attrs
+        ):
+            if param is None:
+                raise sa_exc.ArgumentError(
+                    f"number of composite attributes "
+                    f"{len(self.attrs)} exceeds "
+                    f"that of the number of attributes in class "
+                    f"{self.composite_class.__name__} {len(insp.parameters)}"
+                )
+            if attr is None:
+                # fill in missing attr spots with empty MappedColumn
+                attr = MappedColumn()
+                self.attrs += (attr,)
+
+            if isinstance(attr, MappedColumn):
+                attr.declarative_scan_for_composite(
+                    registry,
+                    cls,
+                    originating_module,
+                    key,
+                    param.name,
+                    param.annotation,
+                )
+            elif isinstance(attr, schema.Column):
+                decl_base._undefer_column_name(param.name, attr)
+
+    @util.memoized_property
+    def _comparable_elements(self) -> Sequence[QueryableAttribute[Any]]:
+        return [getattr(self.parent.class_, prop.key) for prop in self.props]
+
+    @util.memoized_property
+    @util.preload_module("orm.properties")
+    def props(self) -> Sequence[MapperProperty[Any]]:
+        props = []
+        MappedColumn = util.preloaded.orm_properties.MappedColumn
+
+        for attr in self.attrs:
+            if isinstance(attr, str):
+                prop = self.parent.get_property(attr, _configure_mappers=False)
+            elif isinstance(attr, schema.Column):
+                prop = self.parent._columntoproperty[attr]
+            elif isinstance(attr, MappedColumn):
+                prop = self.parent._columntoproperty[attr.column]
+            elif isinstance(attr, attributes.InstrumentedAttribute):
+                prop = attr.property
+            else:
+                prop = None
+
+            if not isinstance(prop, MapperProperty):
+                raise sa_exc.ArgumentError(
+                    "Composite expects Column objects or mapped "
+                    f"attributes/attribute names as arguments, got: {attr!r}"
+                )
+
+            props.append(prop)
+        return props
+
+    @util.non_memoized_property
+    @util.preload_module("orm.properties")
+    def columns(self) -> Sequence[Column[Any]]:
+        MappedColumn = util.preloaded.orm_properties.MappedColumn
+        return [
+            a.column if isinstance(a, MappedColumn) else a
+            for a in self.attrs
+            if isinstance(a, (schema.Column, MappedColumn))
+        ]
+
+    @property
+    def mapper_property_to_assign(self) -> Optional[MapperProperty[_CC]]:
+        return self
+
+    @property
+    def columns_to_assign(self) -> List[Tuple[schema.Column[Any], int]]:
+        return [(c, 0) for c in self.columns if c.table is None]
+
+    @util.preload_module("orm.properties")
+    def _setup_arguments_on_columns(self) -> None:
+        """Propagate configuration arguments made on this composite
+        to the target columns, for those that apply.
+
+        """
+        ColumnProperty = util.preloaded.orm_properties.ColumnProperty
+
+        for prop in self.props:
+            if not isinstance(prop, ColumnProperty):
+                continue
+            else:
+                cprop = prop
+
+            cprop.active_history = self.active_history
+            if self.deferred:
+                cprop.deferred = self.deferred
+                cprop.strategy_key = (("deferred", True), ("instrument", True))
+            cprop.group = self.group
+
+    def _setup_event_handlers(self) -> None:
+        """Establish events that populate/expire the composite attribute."""
+
+        def load_handler(
+            state: InstanceState[Any], context: ORMCompileState
+        ) -> None:
+            _load_refresh_handler(state, context, None, is_refresh=False)
+
+        def refresh_handler(
+            state: InstanceState[Any],
+            context: ORMCompileState,
+            to_load: Optional[Sequence[str]],
+        ) -> None:
+            # note this corresponds to sqlalchemy.ext.mutable load_attrs()
+
+            if not to_load or (
+                {self.key}.union(self._attribute_keys)
+            ).intersection(to_load):
+                _load_refresh_handler(state, context, to_load, is_refresh=True)
+
+        def _load_refresh_handler(
+            state: InstanceState[Any],
+            context: ORMCompileState,
+            to_load: Optional[Sequence[str]],
+            is_refresh: bool,
+        ) -> None:
+            dict_ = state.dict
+
+            # if context indicates we are coming from the
+            # fget() handler, this already set the value; skip the
+            # handler here. (other handlers like mutablecomposite will still
+            # want to catch it)
+            # there's an insufficiency here in that the fget() handler
+            # really should not be using the refresh event and there should
+            # be some other event that mutablecomposite can subscribe
+            # towards for this.
+
+            if (
+                not is_refresh or context is self._COMPOSITE_FGET
+            ) and self.key in dict_:
+                return
+
+            # if column elements aren't loaded, skip.
+            # __get__() will initiate a load for those
+            # columns
+            for k in self._attribute_keys:
+                if k not in dict_:
+                    return
+
+            dict_[self.key] = self.composite_class(
+                *[state.dict[key] for key in self._attribute_keys]
+            )
+
+        def expire_handler(
+            state: InstanceState[Any], keys: Optional[Sequence[str]]
+        ) -> None:
+            if keys is None or set(self._attribute_keys).intersection(keys):
+                state.dict.pop(self.key, None)
+
+        def insert_update_handler(
+            mapper: Mapper[Any],
+            connection: Connection,
+            state: InstanceState[Any],
+        ) -> None:
+            """After an insert or update, some columns may be expired due
+            to server side defaults, or re-populated due to client side
+            defaults.  Pop out the composite value here so that it
+            recreates.
+
+            """
+
+            state.dict.pop(self.key, None)
+
+        event.listen(
+            self.parent, "after_insert", insert_update_handler, raw=True
+        )
+        event.listen(
+            self.parent, "after_update", insert_update_handler, raw=True
+        )
+        event.listen(
+            self.parent, "load", load_handler, raw=True, propagate=True
+        )
+        event.listen(
+            self.parent, "refresh", refresh_handler, raw=True, propagate=True
+        )
+        event.listen(
+            self.parent, "expire", expire_handler, raw=True, propagate=True
+        )
+
+        proxy_attr = self.parent.class_manager[self.key]
+        proxy_attr.impl.dispatch = proxy_attr.dispatch  # type: ignore
+        proxy_attr.impl.dispatch._active_history = self.active_history
+
+        # TODO: need a deserialize hook here
+
+    @util.memoized_property
+    def _attribute_keys(self) -> Sequence[str]:
+        return [prop.key for prop in self.props]
+
+    def _populate_composite_bulk_save_mappings_fn(
+        self,
+    ) -> Callable[[Dict[str, Any]], None]:
+        if self._generated_composite_accessor:
+            get_values = self._generated_composite_accessor
+        else:
+
+            def get_values(val: Any) -> Tuple[Any]:
+                return val.__composite_values__()  # type: ignore
+
+        attrs = [prop.key for prop in self.props]
+
+        def populate(dest_dict: Dict[str, Any]) -> None:
+            dest_dict.update(
+                {
+                    key: val
+                    for key, val in zip(
+                        attrs, get_values(dest_dict.pop(self.key))
+                    )
+                }
+            )
+
+        return populate
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> History:
+        """Provided for userland code that uses attributes.get_history()."""
+
+        added: List[Any] = []
+        deleted: List[Any] = []
+
+        has_history = False
+        for prop in self.props:
+            key = prop.key
+            hist = state.manager[key].impl.get_history(state, dict_)
+            if hist.has_changes():
+                has_history = True
+
+            non_deleted = hist.non_deleted()
+            if non_deleted:
+                added.extend(non_deleted)
+            else:
+                added.append(None)
+            if hist.deleted:
+                deleted.extend(hist.deleted)
+            else:
+                deleted.append(None)
+
+        if has_history:
+            return attributes.History(
+                [self.composite_class(*added)],
+                (),
+                [self.composite_class(*deleted)],
+            )
+        else:
+            return attributes.History((), [self.composite_class(*added)], ())
+
+    def _comparator_factory(
+        self, mapper: Mapper[Any]
+    ) -> Composite.Comparator[_CC]:
+        return self.comparator_factory(self, mapper)
+
+    class CompositeBundle(orm_util.Bundle[_T]):
+        def __init__(
+            self,
+            property_: Composite[_T],
+            expr: ClauseList,
+        ):
+            self.property = property_
+            super().__init__(property_.key, *expr)
+
+        def create_row_processor(
+            self,
+            query: Select[Any],
+            procs: Sequence[Callable[[Row[Any]], Any]],
+            labels: Sequence[str],
+        ) -> Callable[[Row[Any]], Any]:
+            def proc(row: Row[Any]) -> Any:
+                return self.property.composite_class(
+                    *[proc(row) for proc in procs]
+                )
+
+            return proc
+
+    class Comparator(PropComparator[_PT]):
+        """Produce boolean, comparison, and other operators for
+        :class:`.Composite` attributes.
+
+        See the example in :ref:`composite_operations` for an overview
+        of usage , as well as the documentation for :class:`.PropComparator`.
+
+        .. seealso::
+
+            :class:`.PropComparator`
+
+            :class:`.ColumnOperators`
+
+            :ref:`types_operators`
+
+            :attr:`.TypeEngine.comparator_factory`
+
+        """
+
+        # https://github.com/python/mypy/issues/4266
+        __hash__ = None  # type: ignore
+
+        prop: RODescriptorReference[Composite[_PT]]
+
+        @util.memoized_property
+        def clauses(self) -> ClauseList:
+            return expression.ClauseList(
+                group=False, *self._comparable_elements
+            )
+
+        def __clause_element__(self) -> CompositeProperty.CompositeBundle[_PT]:
+            return self.expression
+
+        @util.memoized_property
+        def expression(self) -> CompositeProperty.CompositeBundle[_PT]:
+            clauses = self.clauses._annotate(
+                {
+                    "parententity": self._parententity,
+                    "parentmapper": self._parententity,
+                    "proxy_key": self.prop.key,
+                }
+            )
+            return CompositeProperty.CompositeBundle(self.prop, clauses)
+
+        def _bulk_update_tuples(
+            self, value: Any
+        ) -> Sequence[Tuple[_DMLColumnArgument, Any]]:
+            if isinstance(value, BindParameter):
+                value = value.value
+
+            values: Sequence[Any]
+
+            if value is None:
+                values = [None for key in self.prop._attribute_keys]
+            elif isinstance(self.prop.composite_class, type) and isinstance(
+                value, self.prop.composite_class
+            ):
+                values = self.prop._composite_values_from_instance(
+                    value  # type: ignore[arg-type]
+                )
+            else:
+                raise sa_exc.ArgumentError(
+                    "Can't UPDATE composite attribute %s to %r"
+                    % (self.prop, value)
+                )
+
+            return list(zip(self._comparable_elements, values))
+
+        @util.memoized_property
+        def _comparable_elements(self) -> Sequence[QueryableAttribute[Any]]:
+            if self._adapt_to_entity:
+                return [
+                    getattr(self._adapt_to_entity.entity, prop.key)
+                    for prop in self.prop._comparable_elements
+                ]
+            else:
+                return self.prop._comparable_elements
+
+        def __eq__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            return self._compare(operators.eq, other)
+
+        def __ne__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            return self._compare(operators.ne, other)
+
+        def __lt__(self, other: Any) -> ColumnElement[bool]:
+            return self._compare(operators.lt, other)
+
+        def __gt__(self, other: Any) -> ColumnElement[bool]:
+            return self._compare(operators.gt, other)
+
+        def __le__(self, other: Any) -> ColumnElement[bool]:
+            return self._compare(operators.le, other)
+
+        def __ge__(self, other: Any) -> ColumnElement[bool]:
+            return self._compare(operators.ge, other)
+
+        # what might be interesting would be if we create
+        # an instance of the composite class itself with
+        # the columns as data members, then use "hybrid style" comparison
+        # to create these comparisons.  then your Point.__eq__() method could
+        # be where comparison behavior is defined for SQL also.   Likely
+        # not a good choice for default behavior though, not clear how it would
+        # work w/ dataclasses, etc.  also no demand for any of this anyway.
+        def _compare(
+            self, operator: OperatorType, other: Any
+        ) -> ColumnElement[bool]:
+            values: Sequence[Any]
+            if other is None:
+                values = [None] * len(self.prop._comparable_elements)
+            else:
+                values = self.prop._composite_values_from_instance(other)
+            comparisons = [
+                operator(a, b)
+                for a, b in zip(self.prop._comparable_elements, values)
+            ]
+            if self._adapt_to_entity:
+                assert self.adapter is not None
+                comparisons = [self.adapter(x) for x in comparisons]
+            return sql.and_(*comparisons)
+
+    def __str__(self) -> str:
+        return str(self.parent.class_.__name__) + "." + self.key
+
+
+class Composite(CompositeProperty[_T], _DeclarativeMapped[_T]):
+    """Declarative-compatible front-end for the :class:`.CompositeProperty`
+    class.
+
+    Public constructor is the :func:`_orm.composite` function.
+
+    .. versionchanged:: 2.0 Added :class:`_orm.Composite` as a Declarative
+       compatible subclass of :class:`_orm.CompositeProperty`.
+
+    .. seealso::
+
+        :ref:`mapper_composite`
+
+    """
+
+    inherit_cache = True
+    """:meta private:"""
+
+
+class ConcreteInheritedProperty(DescriptorProperty[_T]):
+    """A 'do nothing' :class:`.MapperProperty` that disables
+    an attribute on a concrete subclass that is only present
+    on the inherited mapper, not the concrete classes' mapper.
+
+    Cases where this occurs include:
+
+    * When the superclass mapper is mapped against a
+      "polymorphic union", which includes all attributes from
+      all subclasses.
+    * When a relationship() is configured on an inherited mapper,
+      but not on the subclass mapper.  Concrete mappers require
+      that relationship() is configured explicitly on each
+      subclass.
+
+    """
+
+    def _comparator_factory(
+        self, mapper: Mapper[Any]
+    ) -> Type[PropComparator[_T]]:
+        comparator_callable = None
+
+        for m in self.parent.iterate_to_root():
+            p = m._props[self.key]
+            if getattr(p, "comparator_factory", None) is not None:
+                comparator_callable = p.comparator_factory
+                break
+        assert comparator_callable is not None
+        return comparator_callable(p, mapper)  # type: ignore
+
+    def __init__(self) -> None:
+        super().__init__()
+
+        def warn() -> NoReturn:
+            raise AttributeError(
+                "Concrete %s does not implement "
+                "attribute %r at the instance level.  Add "
+                "this property explicitly to %s."
+                % (self.parent, self.key, self.parent)
+            )
+
+        class NoninheritedConcreteProp:
+            def __set__(s: Any, obj: Any, value: Any) -> NoReturn:
+                warn()
+
+            def __delete__(s: Any, obj: Any) -> NoReturn:
+                warn()
+
+            def __get__(s: Any, obj: Any, owner: Any) -> Any:
+                if obj is None:
+                    return self.descriptor
+                warn()
+
+        self.descriptor = NoninheritedConcreteProp()
+
+
+class SynonymProperty(DescriptorProperty[_T]):
+    """Denote an attribute name as a synonym to a mapped property,
+    in that the attribute will mirror the value and expression behavior
+    of another attribute.
+
+    :class:`.Synonym` is constructed using the :func:`_orm.synonym`
+    function.
+
+    .. seealso::
+
+        :ref:`synonyms` - Overview of synonyms
+
+    """
+
+    comparator_factory: Optional[Type[PropComparator[_T]]]
+
+    def __init__(
+        self,
+        name: str,
+        map_column: Optional[bool] = None,
+        descriptor: Optional[Any] = None,
+        comparator_factory: Optional[Type[PropComparator[_T]]] = None,
+        attribute_options: Optional[_AttributeOptions] = None,
+        info: Optional[_InfoType] = None,
+        doc: Optional[str] = None,
+    ):
+        super().__init__(attribute_options=attribute_options)
+
+        self.name = name
+        self.map_column = map_column
+        self.descriptor = descriptor
+        self.comparator_factory = comparator_factory
+        if doc:
+            self.doc = doc
+        elif descriptor and descriptor.__doc__:
+            self.doc = descriptor.__doc__
+        else:
+            self.doc = None
+        if info:
+            self.info.update(info)
+
+        util.set_creation_order(self)
+
+    if not TYPE_CHECKING:
+
+        @property
+        def uses_objects(self) -> bool:
+            return getattr(self.parent.class_, self.name).impl.uses_objects
+
+    # TODO: when initialized, check _proxied_object,
+    # emit a warning if its not a column-based property
+
+    @util.memoized_property
+    def _proxied_object(
+        self,
+    ) -> Union[MapperProperty[_T], SQLORMOperations[_T]]:
+        attr = getattr(self.parent.class_, self.name)
+        if not hasattr(attr, "property") or not isinstance(
+            attr.property, MapperProperty
+        ):
+            # attribute is a non-MapperProprerty proxy such as
+            # hybrid or association proxy
+            if isinstance(attr, attributes.QueryableAttribute):
+                return attr.comparator
+            elif isinstance(attr, SQLORMOperations):
+                # assocaition proxy comes here
+                return attr
+
+            raise sa_exc.InvalidRequestError(
+                """synonym() attribute "%s.%s" only supports """
+                """ORM mapped attributes, got %r"""
+                % (self.parent.class_.__name__, self.name, attr)
+            )
+        return attr.property
+
+    def _comparator_factory(self, mapper: Mapper[Any]) -> SQLORMOperations[_T]:
+        prop = self._proxied_object
+
+        if isinstance(prop, MapperProperty):
+            if self.comparator_factory:
+                comp = self.comparator_factory(prop, mapper)
+            else:
+                comp = prop.comparator_factory(prop, mapper)
+            return comp
+        else:
+            return prop
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> History:
+        attr: QueryableAttribute[Any] = getattr(self.parent.class_, self.name)
+        return attr.impl.get_history(state, dict_, passive=passive)
+
+    @util.preload_module("sqlalchemy.orm.properties")
+    def set_parent(self, parent: Mapper[Any], init: bool) -> None:
+        properties = util.preloaded.orm_properties
+
+        if self.map_column:
+            # implement the 'map_column' option.
+            if self.key not in parent.persist_selectable.c:
+                raise sa_exc.ArgumentError(
+                    "Can't compile synonym '%s': no column on table "
+                    "'%s' named '%s'"
+                    % (
+                        self.name,
+                        parent.persist_selectable.description,
+                        self.key,
+                    )
+                )
+            elif (
+                parent.persist_selectable.c[self.key]
+                in parent._columntoproperty
+                and parent._columntoproperty[
+                    parent.persist_selectable.c[self.key]
+                ].key
+                == self.name
+            ):
+                raise sa_exc.ArgumentError(
+                    "Can't call map_column=True for synonym %r=%r, "
+                    "a ColumnProperty already exists keyed to the name "
+                    "%r for column %r"
+                    % (self.key, self.name, self.name, self.key)
+                )
+            p: ColumnProperty[Any] = properties.ColumnProperty(
+                parent.persist_selectable.c[self.key]
+            )
+            parent._configure_property(self.name, p, init=init, setparent=True)
+            p._mapped_by_synonym = self.key
+
+        self.parent = parent
+
+
+class Synonym(SynonymProperty[_T], _DeclarativeMapped[_T]):
+    """Declarative front-end for the :class:`.SynonymProperty` class.
+
+    Public constructor is the :func:`_orm.synonym` function.
+
+    .. versionchanged:: 2.0 Added :class:`_orm.Synonym` as a Declarative
+       compatible subclass for :class:`_orm.SynonymProperty`
+
+    .. seealso::
+
+        :ref:`synonyms` - Overview of synonyms
+
+    """
+
+    inherit_cache = True
+    """:meta private:"""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dynamic.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dynamic.py
new file mode 100644
index 00000000..3c81c396
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/dynamic.py
@@ -0,0 +1,300 @@
+# orm/dynamic.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+"""Dynamic collection API.
+
+Dynamic collections act like Query() objects for read operations and support
+basic add/delete mutation.
+
+.. legacy:: the "dynamic" loader is a legacy feature, superseded by the
+ "write_only" loader.
+
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import exc as orm_exc
+from . import relationships
+from . import util as orm_util
+from .base import PassiveFlag
+from .query import Query
+from .session import object_session
+from .writeonly import AbstractCollectionWriter
+from .writeonly import WriteOnlyAttributeImpl
+from .writeonly import WriteOnlyHistory
+from .writeonly import WriteOnlyLoader
+from .. import util
+from ..engine import result
+
+
+if TYPE_CHECKING:
+    from . import QueryableAttribute
+    from .mapper import Mapper
+    from .relationships import _RelationshipOrderByArg
+    from .session import Session
+    from .state import InstanceState
+    from .util import AliasedClass
+    from ..event import _Dispatch
+    from ..sql.elements import ColumnElement
+
+_T = TypeVar("_T", bound=Any)
+
+
+class DynamicCollectionHistory(WriteOnlyHistory[_T]):
+    def __init__(
+        self,
+        attr: DynamicAttributeImpl,
+        state: InstanceState[_T],
+        passive: PassiveFlag,
+        apply_to: Optional[DynamicCollectionHistory[_T]] = None,
+    ) -> None:
+        if apply_to:
+            coll = AppenderQuery(attr, state).autoflush(False)
+            self.unchanged_items = util.OrderedIdentitySet(coll)
+            self.added_items = apply_to.added_items
+            self.deleted_items = apply_to.deleted_items
+            self._reconcile_collection = True
+        else:
+            self.deleted_items = util.OrderedIdentitySet()
+            self.added_items = util.OrderedIdentitySet()
+            self.unchanged_items = util.OrderedIdentitySet()
+            self._reconcile_collection = False
+
+
+class DynamicAttributeImpl(WriteOnlyAttributeImpl):
+    _supports_dynamic_iteration = True
+    collection_history_cls = DynamicCollectionHistory[Any]
+    query_class: Type[AppenderMixin[Any]]  # type: ignore[assignment]
+
+    def __init__(
+        self,
+        class_: Union[Type[Any], AliasedClass[Any]],
+        key: str,
+        dispatch: _Dispatch[QueryableAttribute[Any]],
+        target_mapper: Mapper[_T],
+        order_by: _RelationshipOrderByArg,
+        query_class: Optional[Type[AppenderMixin[_T]]] = None,
+        **kw: Any,
+    ) -> None:
+        attributes.AttributeImpl.__init__(
+            self, class_, key, None, dispatch, **kw
+        )
+        self.target_mapper = target_mapper
+        if order_by:
+            self.order_by = tuple(order_by)
+        if not query_class:
+            self.query_class = AppenderQuery
+        elif AppenderMixin in query_class.mro():
+            self.query_class = query_class
+        else:
+            self.query_class = mixin_user_query(query_class)
+
+
+@relationships.RelationshipProperty.strategy_for(lazy="dynamic")
+class DynaLoader(WriteOnlyLoader):
+    impl_class = DynamicAttributeImpl
+
+
+class AppenderMixin(AbstractCollectionWriter[_T]):
+    """A mixin that expects to be mixing in a Query class with
+    AbstractAppender.
+
+
+    """
+
+    query_class: Optional[Type[Query[_T]]] = None
+    _order_by_clauses: Tuple[ColumnElement[Any], ...]
+
+    def __init__(
+        self, attr: DynamicAttributeImpl, state: InstanceState[_T]
+    ) -> None:
+        Query.__init__(
+            self,  # type: ignore[arg-type]
+            attr.target_mapper,
+            None,
+        )
+        super().__init__(attr, state)
+
+    @property
+    def session(self) -> Optional[Session]:
+        sess = object_session(self.instance)
+        if sess is not None and sess.autoflush and self.instance in sess:
+            sess.flush()
+        if not orm_util.has_identity(self.instance):
+            return None
+        else:
+            return sess
+
+    @session.setter
+    def session(self, session: Session) -> None:
+        self.sess = session
+
+    def _iter(self) -> Union[result.ScalarResult[_T], result.Result[_T]]:
+        sess = self.session
+        if sess is None:
+            state = attributes.instance_state(self.instance)
+            if state.detached:
+                util.warn(
+                    "Instance %s is detached, dynamic relationship cannot "
+                    "return a correct result.   This warning will become "
+                    "a DetachedInstanceError in a future release."
+                    % (orm_util.state_str(state))
+                )
+
+            return result.IteratorResult(
+                result.SimpleResultMetaData([self.attr.class_.__name__]),
+                iter(
+                    self.attr._get_collection_history(
+                        attributes.instance_state(self.instance),
+                        PassiveFlag.PASSIVE_NO_INITIALIZE,
+                    ).added_items
+                ),
+                _source_supports_scalars=True,
+            ).scalars()
+        else:
+            return self._generate(sess)._iter()
+
+    if TYPE_CHECKING:
+
+        def __iter__(self) -> Iterator[_T]: ...
+
+    def __getitem__(self, index: Any) -> Union[_T, List[_T]]:
+        sess = self.session
+        if sess is None:
+            return self.attr._get_collection_history(
+                attributes.instance_state(self.instance),
+                PassiveFlag.PASSIVE_NO_INITIALIZE,
+            ).indexed(index)
+        else:
+            return self._generate(sess).__getitem__(index)  # type: ignore[no-any-return] # noqa: E501
+
+    def count(self) -> int:
+        sess = self.session
+        if sess is None:
+            return len(
+                self.attr._get_collection_history(
+                    attributes.instance_state(self.instance),
+                    PassiveFlag.PASSIVE_NO_INITIALIZE,
+                ).added_items
+            )
+        else:
+            return self._generate(sess).count()
+
+    def _generate(
+        self,
+        sess: Optional[Session] = None,
+    ) -> Query[_T]:
+        # note we're returning an entirely new Query class instance
+        # here without any assignment capabilities; the class of this
+        # query is determined by the session.
+        instance = self.instance
+        if sess is None:
+            sess = object_session(instance)
+            if sess is None:
+                raise orm_exc.DetachedInstanceError(
+                    "Parent instance %s is not bound to a Session, and no "
+                    "contextual session is established; lazy load operation "
+                    "of attribute '%s' cannot proceed"
+                    % (orm_util.instance_str(instance), self.attr.key)
+                )
+
+        if self.query_class:
+            query = self.query_class(self.attr.target_mapper, session=sess)
+        else:
+            query = sess.query(self.attr.target_mapper)
+
+        query._where_criteria = self._where_criteria
+        query._from_obj = self._from_obj
+        query._order_by_clauses = self._order_by_clauses
+
+        return query
+
+    def add_all(self, iterator: Iterable[_T]) -> None:
+        """Add an iterable of items to this :class:`_orm.AppenderQuery`.
+
+        The given items will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        This method is provided to assist in delivering forwards-compatibility
+        with the :class:`_orm.WriteOnlyCollection` collection class.
+
+        .. versionadded:: 2.0
+
+        """
+        self._add_all_impl(iterator)
+
+    def add(self, item: _T) -> None:
+        """Add an item to this :class:`_orm.AppenderQuery`.
+
+        The given item will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        This method is provided to assist in delivering forwards-compatibility
+        with the :class:`_orm.WriteOnlyCollection` collection class.
+
+        .. versionadded:: 2.0
+
+        """
+        self._add_all_impl([item])
+
+    def extend(self, iterator: Iterable[_T]) -> None:
+        """Add an iterable of items to this :class:`_orm.AppenderQuery`.
+
+        The given items will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        """
+        self._add_all_impl(iterator)
+
+    def append(self, item: _T) -> None:
+        """Append an item to this :class:`_orm.AppenderQuery`.
+
+        The given item will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        """
+        self._add_all_impl([item])
+
+    def remove(self, item: _T) -> None:
+        """Remove an item from this :class:`_orm.AppenderQuery`.
+
+        The given item will be removed from the parent instance's collection on
+        the next flush.
+
+        """
+        self._remove_impl(item)
+
+
+class AppenderQuery(AppenderMixin[_T], Query[_T]):  # type: ignore[misc]
+    """A dynamic query that supports basic collection storage operations.
+
+    Methods on :class:`.AppenderQuery` include all methods of
+    :class:`_orm.Query`, plus additional methods used for collection
+    persistence.
+
+
+    """
+
+
+def mixin_user_query(cls: Any) -> type[AppenderMixin[Any]]:
+    """Return a new class with AppenderQuery functionality layered over."""
+    name = "Appender" + cls.__name__
+    return type(name, (AppenderMixin, cls), {"query_class": cls})
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/evaluator.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/evaluator.py
new file mode 100644
index 00000000..57aae5a3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/evaluator.py
@@ -0,0 +1,379 @@
+# orm/evaluator.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""Evaluation functions used **INTERNALLY** by ORM DML use cases.
+
+
+This module is **private, for internal use by SQLAlchemy**.
+
+.. versionchanged:: 2.0.4 renamed ``EvaluatorCompiler`` to
+   ``_EvaluatorCompiler``.
+
+"""
+
+
+from __future__ import annotations
+
+from typing import Type
+
+from . import exc as orm_exc
+from .base import LoaderCallableStatus
+from .base import PassiveFlag
+from .. import exc
+from .. import inspect
+from ..sql import and_
+from ..sql import operators
+from ..sql.sqltypes import Concatenable
+from ..sql.sqltypes import Integer
+from ..sql.sqltypes import Numeric
+from ..util import warn_deprecated
+
+
+class UnevaluatableError(exc.InvalidRequestError):
+    pass
+
+
+class _NoObject(operators.ColumnOperators):
+    def operate(self, *arg, **kw):
+        return None
+
+    def reverse_operate(self, *arg, **kw):
+        return None
+
+
+class _ExpiredObject(operators.ColumnOperators):
+    def operate(self, *arg, **kw):
+        return self
+
+    def reverse_operate(self, *arg, **kw):
+        return self
+
+
+_NO_OBJECT = _NoObject()
+_EXPIRED_OBJECT = _ExpiredObject()
+
+
+class _EvaluatorCompiler:
+    def __init__(self, target_cls=None):
+        self.target_cls = target_cls
+
+    def process(self, clause, *clauses):
+        if clauses:
+            clause = and_(clause, *clauses)
+
+        meth = getattr(self, f"visit_{clause.__visit_name__}", None)
+        if not meth:
+            raise UnevaluatableError(
+                f"Cannot evaluate {type(clause).__name__}"
+            )
+        return meth(clause)
+
+    def visit_grouping(self, clause):
+        return self.process(clause.element)
+
+    def visit_null(self, clause):
+        return lambda obj: None
+
+    def visit_false(self, clause):
+        return lambda obj: False
+
+    def visit_true(self, clause):
+        return lambda obj: True
+
+    def visit_column(self, clause):
+        try:
+            parentmapper = clause._annotations["parentmapper"]
+        except KeyError as ke:
+            raise UnevaluatableError(
+                f"Cannot evaluate column: {clause}"
+            ) from ke
+
+        if self.target_cls and not issubclass(
+            self.target_cls, parentmapper.class_
+        ):
+            raise UnevaluatableError(
+                "Can't evaluate criteria against "
+                f"alternate class {parentmapper.class_}"
+            )
+
+        parentmapper._check_configure()
+
+        # we'd like to use "proxy_key" annotation to get the "key", however
+        # in relationship primaryjoin cases proxy_key is sometimes deannotated
+        # and sometimes apparently not present in the first place (?).
+        # While I can stop it from being deannotated (though need to see if
+        # this breaks other things), not sure right now  about cases where it's
+        # not there in the first place.  can fix at some later point.
+        # key = clause._annotations["proxy_key"]
+
+        # for now, use the old way
+        try:
+            key = parentmapper._columntoproperty[clause].key
+        except orm_exc.UnmappedColumnError as err:
+            raise UnevaluatableError(
+                f"Cannot evaluate expression: {err}"
+            ) from err
+
+        # note this used to fall back to a simple `getattr(obj, key)` evaluator
+        # if impl was None; as of #8656, we ensure mappers are configured
+        # so that impl is available
+        impl = parentmapper.class_manager[key].impl
+
+        def get_corresponding_attr(obj):
+            if obj is None:
+                return _NO_OBJECT
+            state = inspect(obj)
+            dict_ = state.dict
+
+            value = impl.get(
+                state, dict_, passive=PassiveFlag.PASSIVE_NO_FETCH
+            )
+            if value is LoaderCallableStatus.PASSIVE_NO_RESULT:
+                return _EXPIRED_OBJECT
+            return value
+
+        return get_corresponding_attr
+
+    def visit_tuple(self, clause):
+        return self.visit_clauselist(clause)
+
+    def visit_expression_clauselist(self, clause):
+        return self.visit_clauselist(clause)
+
+    def visit_clauselist(self, clause):
+        evaluators = [self.process(clause) for clause in clause.clauses]
+
+        dispatch = (
+            f"visit_{clause.operator.__name__.rstrip('_')}_clauselist_op"
+        )
+        meth = getattr(self, dispatch, None)
+        if meth:
+            return meth(clause.operator, evaluators, clause)
+        else:
+            raise UnevaluatableError(
+                f"Cannot evaluate clauselist with operator {clause.operator}"
+            )
+
+    def visit_binary(self, clause):
+        eval_left = self.process(clause.left)
+        eval_right = self.process(clause.right)
+
+        dispatch = f"visit_{clause.operator.__name__.rstrip('_')}_binary_op"
+        meth = getattr(self, dispatch, None)
+        if meth:
+            return meth(clause.operator, eval_left, eval_right, clause)
+        else:
+            raise UnevaluatableError(
+                f"Cannot evaluate {type(clause).__name__} with "
+                f"operator {clause.operator}"
+            )
+
+    def visit_or_clauselist_op(self, operator, evaluators, clause):
+        def evaluate(obj):
+            has_null = False
+            for sub_evaluate in evaluators:
+                value = sub_evaluate(obj)
+                if value is _EXPIRED_OBJECT:
+                    return _EXPIRED_OBJECT
+                elif value:
+                    return True
+                has_null = has_null or value is None
+            if has_null:
+                return None
+            return False
+
+        return evaluate
+
+    def visit_and_clauselist_op(self, operator, evaluators, clause):
+        def evaluate(obj):
+            for sub_evaluate in evaluators:
+                value = sub_evaluate(obj)
+                if value is _EXPIRED_OBJECT:
+                    return _EXPIRED_OBJECT
+
+                if not value:
+                    if value is None or value is _NO_OBJECT:
+                        return None
+                    return False
+            return True
+
+        return evaluate
+
+    def visit_comma_op_clauselist_op(self, operator, evaluators, clause):
+        def evaluate(obj):
+            values = []
+            for sub_evaluate in evaluators:
+                value = sub_evaluate(obj)
+                if value is _EXPIRED_OBJECT:
+                    return _EXPIRED_OBJECT
+                elif value is None or value is _NO_OBJECT:
+                    return None
+                values.append(value)
+            return tuple(values)
+
+        return evaluate
+
+    def visit_custom_op_binary_op(
+        self, operator, eval_left, eval_right, clause
+    ):
+        if operator.python_impl:
+            return self._straight_evaluate(
+                operator, eval_left, eval_right, clause
+            )
+        else:
+            raise UnevaluatableError(
+                f"Custom operator {operator.opstring!r} can't be evaluated "
+                "in Python unless it specifies a callable using "
+                "`.python_impl`."
+            )
+
+    def visit_is_binary_op(self, operator, eval_left, eval_right, clause):
+        def evaluate(obj):
+            left_val = eval_left(obj)
+            right_val = eval_right(obj)
+            if left_val is _EXPIRED_OBJECT or right_val is _EXPIRED_OBJECT:
+                return _EXPIRED_OBJECT
+            return left_val == right_val
+
+        return evaluate
+
+    def visit_is_not_binary_op(self, operator, eval_left, eval_right, clause):
+        def evaluate(obj):
+            left_val = eval_left(obj)
+            right_val = eval_right(obj)
+            if left_val is _EXPIRED_OBJECT or right_val is _EXPIRED_OBJECT:
+                return _EXPIRED_OBJECT
+            return left_val != right_val
+
+        return evaluate
+
+    def _straight_evaluate(self, operator, eval_left, eval_right, clause):
+        def evaluate(obj):
+            left_val = eval_left(obj)
+            right_val = eval_right(obj)
+            if left_val is _EXPIRED_OBJECT or right_val is _EXPIRED_OBJECT:
+                return _EXPIRED_OBJECT
+            elif left_val is None or right_val is None:
+                return None
+
+            return operator(eval_left(obj), eval_right(obj))
+
+        return evaluate
+
+    def _straight_evaluate_numeric_only(
+        self, operator, eval_left, eval_right, clause
+    ):
+        if clause.left.type._type_affinity not in (
+            Numeric,
+            Integer,
+        ) or clause.right.type._type_affinity not in (Numeric, Integer):
+            raise UnevaluatableError(
+                f'Cannot evaluate math operator "{operator.__name__}" for '
+                f"datatypes {clause.left.type}, {clause.right.type}"
+            )
+
+        return self._straight_evaluate(operator, eval_left, eval_right, clause)
+
+    visit_add_binary_op = _straight_evaluate_numeric_only
+    visit_mul_binary_op = _straight_evaluate_numeric_only
+    visit_sub_binary_op = _straight_evaluate_numeric_only
+    visit_mod_binary_op = _straight_evaluate_numeric_only
+    visit_truediv_binary_op = _straight_evaluate_numeric_only
+    visit_lt_binary_op = _straight_evaluate
+    visit_le_binary_op = _straight_evaluate
+    visit_ne_binary_op = _straight_evaluate
+    visit_gt_binary_op = _straight_evaluate
+    visit_ge_binary_op = _straight_evaluate
+    visit_eq_binary_op = _straight_evaluate
+
+    def visit_in_op_binary_op(self, operator, eval_left, eval_right, clause):
+        return self._straight_evaluate(
+            lambda a, b: a in b if a is not _NO_OBJECT else None,
+            eval_left,
+            eval_right,
+            clause,
+        )
+
+    def visit_not_in_op_binary_op(
+        self, operator, eval_left, eval_right, clause
+    ):
+        return self._straight_evaluate(
+            lambda a, b: a not in b if a is not _NO_OBJECT else None,
+            eval_left,
+            eval_right,
+            clause,
+        )
+
+    def visit_concat_op_binary_op(
+        self, operator, eval_left, eval_right, clause
+    ):
+
+        if not issubclass(
+            clause.left.type._type_affinity, Concatenable
+        ) or not issubclass(clause.right.type._type_affinity, Concatenable):
+            raise UnevaluatableError(
+                f"Cannot evaluate concatenate operator "
+                f'"{operator.__name__}" for '
+                f"datatypes {clause.left.type}, {clause.right.type}"
+            )
+
+        return self._straight_evaluate(
+            lambda a, b: a + b, eval_left, eval_right, clause
+        )
+
+    def visit_startswith_op_binary_op(
+        self, operator, eval_left, eval_right, clause
+    ):
+        return self._straight_evaluate(
+            lambda a, b: a.startswith(b), eval_left, eval_right, clause
+        )
+
+    def visit_endswith_op_binary_op(
+        self, operator, eval_left, eval_right, clause
+    ):
+        return self._straight_evaluate(
+            lambda a, b: a.endswith(b), eval_left, eval_right, clause
+        )
+
+    def visit_unary(self, clause):
+        eval_inner = self.process(clause.element)
+        if clause.operator is operators.inv:
+
+            def evaluate(obj):
+                value = eval_inner(obj)
+                if value is _EXPIRED_OBJECT:
+                    return _EXPIRED_OBJECT
+                elif value is None:
+                    return None
+                return not value
+
+            return evaluate
+        raise UnevaluatableError(
+            f"Cannot evaluate {type(clause).__name__} "
+            f"with operator {clause.operator}"
+        )
+
+    def visit_bindparam(self, clause):
+        if clause.callable:
+            val = clause.callable()
+        else:
+            val = clause.value
+        return lambda obj: val
+
+
+def __getattr__(name: str) -> Type[_EvaluatorCompiler]:
+    if name == "EvaluatorCompiler":
+        warn_deprecated(
+            "Direct use of 'EvaluatorCompiler' is not supported, and this "
+            "name will be removed in a future release.  "
+            "'_EvaluatorCompiler' is for internal use only",
+            "2.0",
+        )
+        return _EvaluatorCompiler
+    else:
+        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/events.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/events.py
new file mode 100644
index 00000000..f161760e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/events.py
@@ -0,0 +1,3271 @@
+# orm/events.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""ORM event interfaces.
+
+"""
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Collection
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import instrumentation
+from . import interfaces
+from . import mapperlib
+from .attributes import QueryableAttribute
+from .base import _mapper_or_none
+from .base import NO_KEY
+from .instrumentation import ClassManager
+from .instrumentation import InstrumentationFactory
+from .query import BulkDelete
+from .query import BulkUpdate
+from .query import Query
+from .scoping import scoped_session
+from .session import Session
+from .session import sessionmaker
+from .. import event
+from .. import exc
+from .. import util
+from ..event import EventTarget
+from ..event.registry import _ET
+from ..util.compat import inspect_getfullargspec
+
+if TYPE_CHECKING:
+    from weakref import ReferenceType
+
+    from ._typing import _InstanceDict
+    from ._typing import _InternalEntityType
+    from ._typing import _O
+    from ._typing import _T
+    from .attributes import Event
+    from .base import EventConstants
+    from .session import ORMExecuteState
+    from .session import SessionTransaction
+    from .unitofwork import UOWTransaction
+    from ..engine import Connection
+    from ..event.base import _Dispatch
+    from ..event.base import _HasEventsDispatch
+    from ..event.registry import _EventKey
+    from ..orm.collections import CollectionAdapter
+    from ..orm.context import QueryContext
+    from ..orm.decl_api import DeclarativeAttributeIntercept
+    from ..orm.decl_api import DeclarativeMeta
+    from ..orm.mapper import Mapper
+    from ..orm.state import InstanceState
+
+_KT = TypeVar("_KT", bound=Any)
+_ET2 = TypeVar("_ET2", bound=EventTarget)
+
+
+class InstrumentationEvents(event.Events[InstrumentationFactory]):
+    """Events related to class instrumentation events.
+
+    The listeners here support being established against
+    any new style class, that is any object that is a subclass
+    of 'type'.  Events will then be fired off for events
+    against that class.  If the "propagate=True" flag is passed
+    to event.listen(), the event will fire off for subclasses
+    of that class as well.
+
+    The Python ``type`` builtin is also accepted as a target,
+    which when used has the effect of events being emitted
+    for all classes.
+
+    Note the "propagate" flag here is defaulted to ``True``,
+    unlike the other class level events where it defaults
+    to ``False``.  This means that new subclasses will also
+    be the subject of these events, when a listener
+    is established on a superclass.
+
+    """
+
+    _target_class_doc = "SomeBaseClass"
+    _dispatch_target = InstrumentationFactory
+
+    @classmethod
+    def _accept_with(
+        cls,
+        target: Union[
+            InstrumentationFactory,
+            Type[InstrumentationFactory],
+        ],
+        identifier: str,
+    ) -> Optional[
+        Union[
+            InstrumentationFactory,
+            Type[InstrumentationFactory],
+        ]
+    ]:
+        if isinstance(target, type):
+            return _InstrumentationEventsHold(target)  # type: ignore [return-value] # noqa: E501
+        else:
+            return None
+
+    @classmethod
+    def _listen(
+        cls, event_key: _EventKey[_T], propagate: bool = True, **kw: Any
+    ) -> None:
+        target, identifier, fn = (
+            event_key.dispatch_target,
+            event_key.identifier,
+            event_key._listen_fn,
+        )
+
+        def listen(target_cls: type, *arg: Any) -> Optional[Any]:
+            listen_cls = target()
+
+            # if weakref were collected, however this is not something
+            # that normally happens.   it was occurring during test teardown
+            # between mapper/registry/instrumentation_manager, however this
+            # interaction was changed to not rely upon the event system.
+            if listen_cls is None:
+                return None
+
+            if propagate and issubclass(target_cls, listen_cls):
+                return fn(target_cls, *arg)
+            elif not propagate and target_cls is listen_cls:
+                return fn(target_cls, *arg)
+            else:
+                return None
+
+        def remove(ref: ReferenceType[_T]) -> None:
+            key = event.registry._EventKey(  # type: ignore [type-var]
+                None,
+                identifier,
+                listen,
+                instrumentation._instrumentation_factory,
+            )
+            getattr(
+                instrumentation._instrumentation_factory.dispatch, identifier
+            ).remove(key)
+
+        target = weakref.ref(target.class_, remove)
+
+        event_key.with_dispatch_target(
+            instrumentation._instrumentation_factory
+        ).with_wrapper(listen).base_listen(**kw)
+
+    @classmethod
+    def _clear(cls) -> None:
+        super()._clear()
+        instrumentation._instrumentation_factory.dispatch._clear()
+
+    def class_instrument(self, cls: ClassManager[_O]) -> None:
+        """Called after the given class is instrumented.
+
+        To get at the :class:`.ClassManager`, use
+        :func:`.manager_of_class`.
+
+        """
+
+    def class_uninstrument(self, cls: ClassManager[_O]) -> None:
+        """Called before the given class is uninstrumented.
+
+        To get at the :class:`.ClassManager`, use
+        :func:`.manager_of_class`.
+
+        """
+
+    def attribute_instrument(
+        self, cls: ClassManager[_O], key: _KT, inst: _O
+    ) -> None:
+        """Called when an attribute is instrumented."""
+
+
+class _InstrumentationEventsHold:
+    """temporary marker object used to transfer from _accept_with() to
+    _listen() on the InstrumentationEvents class.
+
+    """
+
+    def __init__(self, class_: type) -> None:
+        self.class_ = class_
+
+    dispatch = event.dispatcher(InstrumentationEvents)
+
+
+class InstanceEvents(event.Events[ClassManager[Any]]):
+    """Define events specific to object lifecycle.
+
+    e.g.::
+
+        from sqlalchemy import event
+
+
+        def my_load_listener(target, context):
+            print("on load!")
+
+
+        event.listen(SomeClass, "load", my_load_listener)
+
+    Available targets include:
+
+    * mapped classes
+    * unmapped superclasses of mapped or to-be-mapped classes
+      (using the ``propagate=True`` flag)
+    * :class:`_orm.Mapper` objects
+    * the :class:`_orm.Mapper` class itself indicates listening for all
+      mappers.
+
+    Instance events are closely related to mapper events, but
+    are more specific to the instance and its instrumentation,
+    rather than its system of persistence.
+
+    When using :class:`.InstanceEvents`, several modifiers are
+    available to the :func:`.event.listen` function.
+
+    :param propagate=False: When True, the event listener should
+       be applied to all inheriting classes as well as the
+       class which is the target of this listener.
+    :param raw=False: When True, the "target" argument passed
+       to applicable event listener functions will be the
+       instance's :class:`.InstanceState` management
+       object, rather than the mapped instance itself.
+    :param restore_load_context=False: Applies to the
+       :meth:`.InstanceEvents.load` and :meth:`.InstanceEvents.refresh`
+       events.  Restores the loader context of the object when the event
+       hook is complete, so that ongoing eager load operations continue
+       to target the object appropriately.  A warning is emitted if the
+       object is moved to a new loader context from within one of these
+       events if this flag is not set.
+
+       .. versionadded:: 1.3.14
+
+
+    """
+
+    _target_class_doc = "SomeClass"
+
+    _dispatch_target = ClassManager
+
+    @classmethod
+    def _new_classmanager_instance(
+        cls,
+        class_: Union[DeclarativeAttributeIntercept, DeclarativeMeta, type],
+        classmanager: ClassManager[_O],
+    ) -> None:
+        _InstanceEventsHold.populate(class_, classmanager)
+
+    @classmethod
+    @util.preload_module("sqlalchemy.orm")
+    def _accept_with(
+        cls,
+        target: Union[
+            ClassManager[Any],
+            Type[ClassManager[Any]],
+        ],
+        identifier: str,
+    ) -> Optional[Union[ClassManager[Any], Type[ClassManager[Any]]]]:
+        orm = util.preloaded.orm
+
+        if isinstance(target, ClassManager):
+            return target
+        elif isinstance(target, mapperlib.Mapper):
+            return target.class_manager
+        elif target is orm.mapper:  # type: ignore [attr-defined]
+            util.warn_deprecated(
+                "The `sqlalchemy.orm.mapper()` symbol is deprecated and "
+                "will be removed in a future release. For the mapper-wide "
+                "event target, use the 'sqlalchemy.orm.Mapper' class.",
+                "2.0",
+            )
+            return ClassManager
+        elif isinstance(target, type):
+            if issubclass(target, mapperlib.Mapper):
+                return ClassManager
+            else:
+                manager = instrumentation.opt_manager_of_class(target)
+                if manager:
+                    return manager
+                else:
+                    return _InstanceEventsHold(target)  # type: ignore [return-value] # noqa: E501
+        return None
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: _EventKey[ClassManager[Any]],
+        raw: bool = False,
+        propagate: bool = False,
+        restore_load_context: bool = False,
+        **kw: Any,
+    ) -> None:
+        target, fn = (event_key.dispatch_target, event_key._listen_fn)
+
+        if not raw or restore_load_context:
+
+            def wrap(
+                state: InstanceState[_O], *arg: Any, **kw: Any
+            ) -> Optional[Any]:
+                if not raw:
+                    target: Any = state.obj()
+                else:
+                    target = state
+                if restore_load_context:
+                    runid = state.runid
+                try:
+                    return fn(target, *arg, **kw)
+                finally:
+                    if restore_load_context:
+                        state.runid = runid
+
+            event_key = event_key.with_wrapper(wrap)
+
+        event_key.base_listen(propagate=propagate, **kw)
+
+        if propagate:
+            for mgr in target.subclass_managers(True):
+                event_key.with_dispatch_target(mgr).base_listen(propagate=True)
+
+    @classmethod
+    def _clear(cls) -> None:
+        super()._clear()
+        _InstanceEventsHold._clear()
+
+    def first_init(self, manager: ClassManager[_O], cls: Type[_O]) -> None:
+        """Called when the first instance of a particular mapping is called.
+
+        This event is called when the ``__init__`` method of a class
+        is called the first time for that particular class.    The event
+        invokes before ``__init__`` actually proceeds as well as before
+        the :meth:`.InstanceEvents.init` event is invoked.
+
+        """
+
+    def init(self, target: _O, args: Any, kwargs: Any) -> None:
+        """Receive an instance when its constructor is called.
+
+        This method is only called during a userland construction of
+        an object, in conjunction with the object's constructor, e.g.
+        its ``__init__`` method.  It is not called when an object is
+        loaded from the database; see the :meth:`.InstanceEvents.load`
+        event in order to intercept a database load.
+
+        The event is called before the actual ``__init__`` constructor
+        of the object is called.  The ``kwargs`` dictionary may be
+        modified in-place in order to affect what is passed to
+        ``__init__``.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param args: positional arguments passed to the ``__init__`` method.
+         This is passed as a tuple and is currently immutable.
+        :param kwargs: keyword arguments passed to the ``__init__`` method.
+         This structure *can* be altered in place.
+
+        .. seealso::
+
+            :meth:`.InstanceEvents.init_failure`
+
+            :meth:`.InstanceEvents.load`
+
+        """
+
+    def init_failure(self, target: _O, args: Any, kwargs: Any) -> None:
+        """Receive an instance when its constructor has been called,
+        and raised an exception.
+
+        This method is only called during a userland construction of
+        an object, in conjunction with the object's constructor, e.g.
+        its ``__init__`` method. It is not called when an object is loaded
+        from the database.
+
+        The event is invoked after an exception raised by the ``__init__``
+        method is caught.  After the event
+        is invoked, the original exception is re-raised outwards, so that
+        the construction of the object still raises an exception.   The
+        actual exception and stack trace raised should be present in
+        ``sys.exc_info()``.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param args: positional arguments that were passed to the ``__init__``
+         method.
+        :param kwargs: keyword arguments that were passed to the ``__init__``
+         method.
+
+        .. seealso::
+
+            :meth:`.InstanceEvents.init`
+
+            :meth:`.InstanceEvents.load`
+
+        """
+
+    def _sa_event_merge_wo_load(
+        self, target: _O, context: QueryContext
+    ) -> None:
+        """receive an object instance after it was the subject of a merge()
+        call, when load=False was passed.
+
+        The target would be the already-loaded object in the Session which
+        would have had its attributes overwritten by the incoming object. This
+        overwrite operation does not use attribute events, instead just
+        populating dict directly. Therefore the purpose of this event is so
+        that extensions like sqlalchemy.ext.mutable know that object state has
+        changed and incoming state needs to be set up for "parents" etc.
+
+        This functionality is acceptable to be made public in a later release.
+
+        .. versionadded:: 1.4.41
+
+        """
+
+    def load(self, target: _O, context: QueryContext) -> None:
+        """Receive an object instance after it has been created via
+        ``__new__``, and after initial attribute population has
+        occurred.
+
+        This typically occurs when the instance is created based on
+        incoming result rows, and is only called once for that
+        instance's lifetime.
+
+        .. warning::
+
+            During a result-row load, this event is invoked when the
+            first row received for this instance is processed.  When using
+            eager loading with collection-oriented attributes, the additional
+            rows that are to be loaded / processed in order to load subsequent
+            collection items have not occurred yet.   This has the effect
+            both that collections will not be fully loaded, as well as that
+            if an operation occurs within this event handler that emits
+            another database load operation for the object, the "loading
+            context" for the object can change and interfere with the
+            existing eager loaders still in progress.
+
+            Examples of what can cause the "loading context" to change within
+            the event handler include, but are not necessarily limited to:
+
+            * accessing deferred attributes that weren't part of the row,
+              will trigger an "undefer" operation and refresh the object
+
+            * accessing attributes on a joined-inheritance subclass that
+              weren't part of the row, will trigger a refresh operation.
+
+            As of SQLAlchemy 1.3.14, a warning is emitted when this occurs. The
+            :paramref:`.InstanceEvents.restore_load_context` option may  be
+            used on the event to prevent this warning; this will ensure that
+            the existing loading context is maintained for the object after the
+            event is called::
+
+                @event.listens_for(SomeClass, "load", restore_load_context=True)
+                def on_load(instance, context):
+                    instance.some_unloaded_attribute
+
+            .. versionchanged:: 1.3.14 Added
+               :paramref:`.InstanceEvents.restore_load_context`
+               and :paramref:`.SessionEvents.restore_load_context` flags which
+               apply to "on load" events, which will ensure that the loading
+               context for an object is restored when the event hook is
+               complete; a warning is emitted if the load context of the object
+               changes without this flag being set.
+
+
+        The :meth:`.InstanceEvents.load` event is also available in a
+        class-method decorator format called :func:`_orm.reconstructor`.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param context: the :class:`.QueryContext` corresponding to the
+         current :class:`_query.Query` in progress.  This argument may be
+         ``None`` if the load does not correspond to a :class:`_query.Query`,
+         such as during :meth:`.Session.merge`.
+
+        .. seealso::
+
+            :ref:`mapped_class_load_events`
+
+            :meth:`.InstanceEvents.init`
+
+            :meth:`.InstanceEvents.refresh`
+
+            :meth:`.SessionEvents.loaded_as_persistent`
+
+        """  # noqa: E501
+
+    def refresh(
+        self, target: _O, context: QueryContext, attrs: Optional[Iterable[str]]
+    ) -> None:
+        """Receive an object instance after one or more attributes have
+        been refreshed from a query.
+
+        Contrast this to the :meth:`.InstanceEvents.load` method, which
+        is invoked when the object is first loaded from a query.
+
+        .. note:: This event is invoked within the loader process before
+           eager loaders may have been completed, and the object's state may
+           not be complete.  Additionally, invoking row-level refresh
+           operations on the object will place the object into a new loader
+           context, interfering with the existing load context.   See the note
+           on :meth:`.InstanceEvents.load` for background on making use of the
+           :paramref:`.InstanceEvents.restore_load_context` parameter, in
+           order to resolve this scenario.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param context: the :class:`.QueryContext` corresponding to the
+         current :class:`_query.Query` in progress.
+        :param attrs: sequence of attribute names which
+         were populated, or None if all column-mapped, non-deferred
+         attributes were populated.
+
+        .. seealso::
+
+            :ref:`mapped_class_load_events`
+
+            :meth:`.InstanceEvents.load`
+
+        """
+
+    def refresh_flush(
+        self,
+        target: _O,
+        flush_context: UOWTransaction,
+        attrs: Optional[Iterable[str]],
+    ) -> None:
+        """Receive an object instance after one or more attributes that
+        contain a column-level default or onupdate handler have been refreshed
+        during persistence of the object's state.
+
+        This event is the same as :meth:`.InstanceEvents.refresh` except
+        it is invoked within the unit of work flush process, and includes
+        only non-primary-key columns that have column level default or
+        onupdate handlers, including Python callables as well as server side
+        defaults and triggers which may be fetched via the RETURNING clause.
+
+        .. note::
+
+            While the :meth:`.InstanceEvents.refresh_flush` event is triggered
+            for an object that was INSERTed as well as for an object that was
+            UPDATEd, the event is geared primarily  towards the UPDATE process;
+            it is mostly an internal artifact that INSERT actions can also
+            trigger this event, and note that **primary key columns for an
+            INSERTed row are explicitly omitted** from this event.  In order to
+            intercept the newly INSERTed state of an object, the
+            :meth:`.SessionEvents.pending_to_persistent` and
+            :meth:`.MapperEvents.after_insert` are better choices.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param flush_context: Internal :class:`.UOWTransaction` object
+         which handles the details of the flush.
+        :param attrs: sequence of attribute names which
+         were populated.
+
+        .. seealso::
+
+            :ref:`mapped_class_load_events`
+
+            :ref:`orm_server_defaults`
+
+            :ref:`metadata_defaults_toplevel`
+
+        """
+
+    def expire(self, target: _O, attrs: Optional[Iterable[str]]) -> None:
+        """Receive an object instance after its attributes or some subset
+        have been expired.
+
+        'keys' is a list of attribute names.  If None, the entire
+        state was expired.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param attrs: sequence of attribute
+         names which were expired, or None if all attributes were
+         expired.
+
+        """
+
+    def pickle(self, target: _O, state_dict: _InstanceDict) -> None:
+        """Receive an object instance when its associated state is
+        being pickled.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param state_dict: the dictionary returned by
+         :class:`.InstanceState.__getstate__`, containing the state
+         to be pickled.
+
+        """
+
+    def unpickle(self, target: _O, state_dict: _InstanceDict) -> None:
+        """Receive an object instance after its associated state has
+        been unpickled.
+
+        :param target: the mapped instance.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :param state_dict: the dictionary sent to
+         :class:`.InstanceState.__setstate__`, containing the state
+         dictionary which was pickled.
+
+        """
+
+
+class _EventsHold(event.RefCollection[_ET]):
+    """Hold onto listeners against unmapped, uninstrumented classes.
+
+    Establish _listen() for that class' mapper/instrumentation when
+    those objects are created for that class.
+
+    """
+
+    all_holds: weakref.WeakKeyDictionary[Any, Any]
+
+    def __init__(
+        self,
+        class_: Union[DeclarativeAttributeIntercept, DeclarativeMeta, type],
+    ) -> None:
+        self.class_ = class_
+
+    @classmethod
+    def _clear(cls) -> None:
+        cls.all_holds.clear()
+
+    class HoldEvents(Generic[_ET2]):
+        _dispatch_target: Optional[Type[_ET2]] = None
+
+        @classmethod
+        def _listen(
+            cls,
+            event_key: _EventKey[_ET2],
+            raw: bool = False,
+            propagate: bool = False,
+            retval: bool = False,
+            **kw: Any,
+        ) -> None:
+            target = event_key.dispatch_target
+
+            if target.class_ in target.all_holds:
+                collection = target.all_holds[target.class_]
+            else:
+                collection = target.all_holds[target.class_] = {}
+
+            event.registry._stored_in_collection(event_key, target)
+            collection[event_key._key] = (
+                event_key,
+                raw,
+                propagate,
+                retval,
+                kw,
+            )
+
+            if propagate:
+                stack = list(target.class_.__subclasses__())
+                while stack:
+                    subclass = stack.pop(0)
+                    stack.extend(subclass.__subclasses__())
+                    subject = target.resolve(subclass)
+                    if subject is not None:
+                        # we are already going through __subclasses__()
+                        # so leave generic propagate flag False
+                        event_key.with_dispatch_target(subject).listen(
+                            raw=raw, propagate=False, retval=retval, **kw
+                        )
+
+    def remove(self, event_key: _EventKey[_ET]) -> None:
+        target = event_key.dispatch_target
+
+        if isinstance(target, _EventsHold):
+            collection = target.all_holds[target.class_]
+            del collection[event_key._key]
+
+    @classmethod
+    def populate(
+        cls,
+        class_: Union[DeclarativeAttributeIntercept, DeclarativeMeta, type],
+        subject: Union[ClassManager[_O], Mapper[_O]],
+    ) -> None:
+        for subclass in class_.__mro__:
+            if subclass in cls.all_holds:
+                collection = cls.all_holds[subclass]
+                for (
+                    event_key,
+                    raw,
+                    propagate,
+                    retval,
+                    kw,
+                ) in collection.values():
+                    if propagate or subclass is class_:
+                        # since we can't be sure in what order different
+                        # classes in a hierarchy are triggered with
+                        # populate(), we rely upon _EventsHold for all event
+                        # assignment, instead of using the generic propagate
+                        # flag.
+                        event_key.with_dispatch_target(subject).listen(
+                            raw=raw, propagate=False, retval=retval, **kw
+                        )
+
+
+class _InstanceEventsHold(_EventsHold[_ET]):
+    all_holds: weakref.WeakKeyDictionary[Any, Any] = (
+        weakref.WeakKeyDictionary()
+    )
+
+    def resolve(self, class_: Type[_O]) -> Optional[ClassManager[_O]]:
+        return instrumentation.opt_manager_of_class(class_)
+
+    class HoldInstanceEvents(_EventsHold.HoldEvents[_ET], InstanceEvents):  # type: ignore [misc] # noqa: E501
+        pass
+
+    dispatch = event.dispatcher(HoldInstanceEvents)
+
+
+class MapperEvents(event.Events[mapperlib.Mapper[Any]]):
+    """Define events specific to mappings.
+
+    e.g.::
+
+        from sqlalchemy import event
+
+
+        def my_before_insert_listener(mapper, connection, target):
+            # execute a stored procedure upon INSERT,
+            # apply the value to the row to be inserted
+            target.calculated_value = connection.execute(
+                text("select my_special_function(%d)" % target.special_number)
+            ).scalar()
+
+
+        # associate the listener function with SomeClass,
+        # to execute during the "before_insert" hook
+        event.listen(SomeClass, "before_insert", my_before_insert_listener)
+
+    Available targets include:
+
+    * mapped classes
+    * unmapped superclasses of mapped or to-be-mapped classes
+      (using the ``propagate=True`` flag)
+    * :class:`_orm.Mapper` objects
+    * the :class:`_orm.Mapper` class itself indicates listening for all
+      mappers.
+
+    Mapper events provide hooks into critical sections of the
+    mapper, including those related to object instrumentation,
+    object loading, and object persistence. In particular, the
+    persistence methods :meth:`~.MapperEvents.before_insert`,
+    and :meth:`~.MapperEvents.before_update` are popular
+    places to augment the state being persisted - however, these
+    methods operate with several significant restrictions. The
+    user is encouraged to evaluate the
+    :meth:`.SessionEvents.before_flush` and
+    :meth:`.SessionEvents.after_flush` methods as more
+    flexible and user-friendly hooks in which to apply
+    additional database state during a flush.
+
+    When using :class:`.MapperEvents`, several modifiers are
+    available to the :func:`.event.listen` function.
+
+    :param propagate=False: When True, the event listener should
+       be applied to all inheriting mappers and/or the mappers of
+       inheriting classes, as well as any
+       mapper which is the target of this listener.
+    :param raw=False: When True, the "target" argument passed
+       to applicable event listener functions will be the
+       instance's :class:`.InstanceState` management
+       object, rather than the mapped instance itself.
+    :param retval=False: when True, the user-defined event function
+       must have a return value, the purpose of which is either to
+       control subsequent event propagation, or to otherwise alter
+       the operation in progress by the mapper.   Possible return
+       values are:
+
+       * ``sqlalchemy.orm.interfaces.EXT_CONTINUE`` - continue event
+         processing normally.
+       * ``sqlalchemy.orm.interfaces.EXT_STOP`` - cancel all subsequent
+         event handlers in the chain.
+       * other values - the return value specified by specific listeners.
+
+    """
+
+    _target_class_doc = "SomeClass"
+    _dispatch_target = mapperlib.Mapper
+
+    @classmethod
+    def _new_mapper_instance(
+        cls,
+        class_: Union[DeclarativeAttributeIntercept, DeclarativeMeta, type],
+        mapper: Mapper[_O],
+    ) -> None:
+        _MapperEventsHold.populate(class_, mapper)
+
+    @classmethod
+    @util.preload_module("sqlalchemy.orm")
+    def _accept_with(
+        cls,
+        target: Union[mapperlib.Mapper[Any], Type[mapperlib.Mapper[Any]]],
+        identifier: str,
+    ) -> Optional[Union[mapperlib.Mapper[Any], Type[mapperlib.Mapper[Any]]]]:
+        orm = util.preloaded.orm
+
+        if target is orm.mapper:  # type: ignore [attr-defined]
+            util.warn_deprecated(
+                "The `sqlalchemy.orm.mapper()` symbol is deprecated and "
+                "will be removed in a future release. For the mapper-wide "
+                "event target, use the 'sqlalchemy.orm.Mapper' class.",
+                "2.0",
+            )
+            return mapperlib.Mapper
+        elif isinstance(target, type):
+            if issubclass(target, mapperlib.Mapper):
+                return target
+            else:
+                mapper = _mapper_or_none(target)
+                if mapper is not None:
+                    return mapper
+                else:
+                    return _MapperEventsHold(target)
+        else:
+            return target
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: _EventKey[_ET],
+        raw: bool = False,
+        retval: bool = False,
+        propagate: bool = False,
+        **kw: Any,
+    ) -> None:
+        target, identifier, fn = (
+            event_key.dispatch_target,
+            event_key.identifier,
+            event_key._listen_fn,
+        )
+
+        if (
+            identifier in ("before_configured", "after_configured")
+            and target is not mapperlib.Mapper
+        ):
+            util.warn(
+                "'before_configured' and 'after_configured' ORM events "
+                "only invoke with the Mapper class "
+                "as the target."
+            )
+
+        if not raw or not retval:
+            if not raw:
+                meth = getattr(cls, identifier)
+                try:
+                    target_index = (
+                        inspect_getfullargspec(meth)[0].index("target") - 1
+                    )
+                except ValueError:
+                    target_index = None
+
+            def wrap(*arg: Any, **kw: Any) -> Any:
+                if not raw and target_index is not None:
+                    arg = list(arg)  # type: ignore [assignment]
+                    arg[target_index] = arg[target_index].obj()  # type: ignore [index] # noqa: E501
+                if not retval:
+                    fn(*arg, **kw)
+                    return interfaces.EXT_CONTINUE
+                else:
+                    return fn(*arg, **kw)
+
+            event_key = event_key.with_wrapper(wrap)
+
+        if propagate:
+            for mapper in target.self_and_descendants:
+                event_key.with_dispatch_target(mapper).base_listen(
+                    propagate=True, **kw
+                )
+        else:
+            event_key.base_listen(**kw)
+
+    @classmethod
+    def _clear(cls) -> None:
+        super()._clear()
+        _MapperEventsHold._clear()
+
+    def instrument_class(self, mapper: Mapper[_O], class_: Type[_O]) -> None:
+        r"""Receive a class when the mapper is first constructed,
+        before instrumentation is applied to the mapped class.
+
+        This event is the earliest phase of mapper construction.
+        Most attributes of the mapper are not yet initialized.   To
+        receive an event within initial mapper construction where basic
+        state is available such as the :attr:`_orm.Mapper.attrs` collection,
+        the :meth:`_orm.MapperEvents.after_mapper_constructed` event may
+        be a better choice.
+
+        This listener can either be applied to the :class:`_orm.Mapper`
+        class overall, or to any un-mapped class which serves as a base
+        for classes that will be mapped (using the ``propagate=True`` flag)::
+
+            Base = declarative_base()
+
+
+            @event.listens_for(Base, "instrument_class", propagate=True)
+            def on_new_class(mapper, cls_):
+                "..."
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param class\_: the mapped class.
+
+        .. seealso::
+
+            :meth:`_orm.MapperEvents.after_mapper_constructed`
+
+        """
+
+    def after_mapper_constructed(
+        self, mapper: Mapper[_O], class_: Type[_O]
+    ) -> None:
+        """Receive a class and mapper when the :class:`_orm.Mapper` has been
+        fully constructed.
+
+        This event is called after the initial constructor for
+        :class:`_orm.Mapper` completes.  This occurs after the
+        :meth:`_orm.MapperEvents.instrument_class` event and after the
+        :class:`_orm.Mapper` has done an initial pass of its arguments
+        to generate its collection of :class:`_orm.MapperProperty` objects,
+        which are accessible via the :meth:`_orm.Mapper.get_property`
+        method and the :attr:`_orm.Mapper.iterate_properties` attribute.
+
+        This event differs from the
+        :meth:`_orm.MapperEvents.before_mapper_configured` event in that it
+        is invoked within the constructor for :class:`_orm.Mapper`, rather
+        than within the :meth:`_orm.registry.configure` process.   Currently,
+        this event is the only one which is appropriate for handlers that
+        wish to create additional mapped classes in response to the
+        construction of this :class:`_orm.Mapper`, which will be part of the
+        same configure step when :meth:`_orm.registry.configure` next runs.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`examples_versioning` - an example which illustrates the use
+            of the :meth:`_orm.MapperEvents.before_mapper_configured`
+            event to create new mappers to record change-audit histories on
+            objects.
+
+        """
+
+    def before_mapper_configured(
+        self, mapper: Mapper[_O], class_: Type[_O]
+    ) -> None:
+        """Called right before a specific mapper is to be configured.
+
+        This event is intended to allow a specific mapper to be skipped during
+        the configure step, by returning the :attr:`.orm.interfaces.EXT_SKIP`
+        symbol which indicates to the :func:`.configure_mappers` call that this
+        particular mapper (or hierarchy of mappers, if ``propagate=True`` is
+        used) should be skipped in the current configuration run. When one or
+        more mappers are skipped, the "new mappers" flag will remain set,
+        meaning the :func:`.configure_mappers` function will continue to be
+        called when mappers are used, to continue to try to configure all
+        available mappers.
+
+        In comparison to the other configure-level events,
+        :meth:`.MapperEvents.before_configured`,
+        :meth:`.MapperEvents.after_configured`, and
+        :meth:`.MapperEvents.mapper_configured`, the
+        :meth:`.MapperEvents.before_mapper_configured` event provides for a
+        meaningful return value when it is registered with the ``retval=True``
+        parameter.
+
+        .. versionadded:: 1.3
+
+        e.g.::
+
+            from sqlalchemy.orm import EXT_SKIP
+
+            Base = declarative_base()
+
+            DontConfigureBase = declarative_base()
+
+
+            @event.listens_for(
+                DontConfigureBase,
+                "before_mapper_configured",
+                retval=True,
+                propagate=True,
+            )
+            def dont_configure(mapper, cls):
+                return EXT_SKIP
+
+        .. seealso::
+
+            :meth:`.MapperEvents.before_configured`
+
+            :meth:`.MapperEvents.after_configured`
+
+            :meth:`.MapperEvents.mapper_configured`
+
+        """
+
+    def mapper_configured(self, mapper: Mapper[_O], class_: Type[_O]) -> None:
+        r"""Called when a specific mapper has completed its own configuration
+        within the scope of the :func:`.configure_mappers` call.
+
+        The :meth:`.MapperEvents.mapper_configured` event is invoked
+        for each mapper that is encountered when the
+        :func:`_orm.configure_mappers` function proceeds through the current
+        list of not-yet-configured mappers.
+        :func:`_orm.configure_mappers` is typically invoked
+        automatically as mappings are first used, as well as each time
+        new mappers have been made available and new mapper use is
+        detected.
+
+        When the event is called, the mapper should be in its final
+        state, but **not including backrefs** that may be invoked from
+        other mappers; they might still be pending within the
+        configuration operation.    Bidirectional relationships that
+        are instead configured via the
+        :paramref:`.orm.relationship.back_populates` argument
+        *will* be fully available, since this style of relationship does not
+        rely upon other possibly-not-configured mappers to know that they
+        exist.
+
+        For an event that is guaranteed to have **all** mappers ready
+        to go including backrefs that are defined only on other
+        mappings, use the :meth:`.MapperEvents.after_configured`
+        event; this event invokes only after all known mappings have been
+        fully configured.
+
+        The :meth:`.MapperEvents.mapper_configured` event, unlike
+        :meth:`.MapperEvents.before_configured` or
+        :meth:`.MapperEvents.after_configured`,
+        is called for each mapper/class individually, and the mapper is
+        passed to the event itself.  It also is called exactly once for
+        a particular mapper.  The event is therefore useful for
+        configurational steps that benefit from being invoked just once
+        on a specific mapper basis, which don't require that "backref"
+        configurations are necessarily ready yet.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param class\_: the mapped class.
+
+        .. seealso::
+
+            :meth:`.MapperEvents.before_configured`
+
+            :meth:`.MapperEvents.after_configured`
+
+            :meth:`.MapperEvents.before_mapper_configured`
+
+        """
+        # TODO: need coverage for this event
+
+    def before_configured(self) -> None:
+        """Called before a series of mappers have been configured.
+
+        The :meth:`.MapperEvents.before_configured` event is invoked
+        each time the :func:`_orm.configure_mappers` function is
+        invoked, before the function has done any of its work.
+        :func:`_orm.configure_mappers` is typically invoked
+        automatically as mappings are first used, as well as each time
+        new mappers have been made available and new mapper use is
+        detected.
+
+        This event can **only** be applied to the :class:`_orm.Mapper` class,
+        and not to individual mappings or mapped classes. It is only invoked
+        for all mappings as a whole::
+
+            from sqlalchemy.orm import Mapper
+
+
+            @event.listens_for(Mapper, "before_configured")
+            def go(): ...
+
+        Contrast this event to :meth:`.MapperEvents.after_configured`,
+        which is invoked after the series of mappers has been configured,
+        as well as :meth:`.MapperEvents.before_mapper_configured`
+        and :meth:`.MapperEvents.mapper_configured`, which are both invoked
+        on a per-mapper basis.
+
+        Theoretically this event is called once per
+        application, but is actually called any time new mappers
+        are to be affected by a :func:`_orm.configure_mappers`
+        call.   If new mappings are constructed after existing ones have
+        already been used, this event will likely be called again.  To ensure
+        that a particular event is only called once and no further, the
+        ``once=True`` argument (new in 0.9.4) can be applied::
+
+            from sqlalchemy.orm import mapper
+
+
+            @event.listens_for(mapper, "before_configured", once=True)
+            def go(): ...
+
+        .. seealso::
+
+            :meth:`.MapperEvents.before_mapper_configured`
+
+            :meth:`.MapperEvents.mapper_configured`
+
+            :meth:`.MapperEvents.after_configured`
+
+        """
+
+    def after_configured(self) -> None:
+        """Called after a series of mappers have been configured.
+
+        The :meth:`.MapperEvents.after_configured` event is invoked
+        each time the :func:`_orm.configure_mappers` function is
+        invoked, after the function has completed its work.
+        :func:`_orm.configure_mappers` is typically invoked
+        automatically as mappings are first used, as well as each time
+        new mappers have been made available and new mapper use is
+        detected.
+
+        Contrast this event to the :meth:`.MapperEvents.mapper_configured`
+        event, which is called on a per-mapper basis while the configuration
+        operation proceeds; unlike that event, when this event is invoked,
+        all cross-configurations (e.g. backrefs) will also have been made
+        available for any mappers that were pending.
+        Also contrast to :meth:`.MapperEvents.before_configured`,
+        which is invoked before the series of mappers has been configured.
+
+        This event can **only** be applied to the :class:`_orm.Mapper` class,
+        and not to individual mappings or
+        mapped classes.  It is only invoked for all mappings as a whole::
+
+            from sqlalchemy.orm import Mapper
+
+
+            @event.listens_for(Mapper, "after_configured")
+            def go(): ...
+
+        Theoretically this event is called once per
+        application, but is actually called any time new mappers
+        have been affected by a :func:`_orm.configure_mappers`
+        call.   If new mappings are constructed after existing ones have
+        already been used, this event will likely be called again.  To ensure
+        that a particular event is only called once and no further, the
+        ``once=True`` argument (new in 0.9.4) can be applied::
+
+            from sqlalchemy.orm import mapper
+
+
+            @event.listens_for(mapper, "after_configured", once=True)
+            def go(): ...
+
+        .. seealso::
+
+            :meth:`.MapperEvents.before_mapper_configured`
+
+            :meth:`.MapperEvents.mapper_configured`
+
+            :meth:`.MapperEvents.before_configured`
+
+        """
+
+    def before_insert(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance before an INSERT statement
+        is emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to modify local, non-object related
+        attributes on the instance before an INSERT occurs, as well
+        as to emit additional SQL statements on the given
+        connection.
+
+        The event is often called for a batch of objects of the
+        same class before their INSERT statements are emitted at
+        once in a later step. In the extremely rare case that
+        this is not desirable, the :class:`_orm.Mapper` object can be
+        configured with ``batch=False``, which will cause
+        batches of instances to be broken up into individual
+        (and more poorly performing) event->persist->event
+        steps.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit INSERT statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being persisted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_insert(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance after an INSERT statement
+        is emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to modify in-Python-only
+        state on the instance after an INSERT occurs, as well
+        as to emit additional SQL statements on the given
+        connection.
+
+        The event is often called for a batch of objects of the
+        same class after their INSERT statements have been
+        emitted at once in a previous step. In the extremely
+        rare case that this is not desirable, the
+        :class:`_orm.Mapper` object can be configured with ``batch=False``,
+        which will cause batches of instances to be broken up
+        into individual (and more poorly performing)
+        event->persist->event steps.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit INSERT statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being persisted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def before_update(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance before an UPDATE statement
+        is emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to modify local, non-object related
+        attributes on the instance before an UPDATE occurs, as well
+        as to emit additional SQL statements on the given
+        connection.
+
+        This method is called for all instances that are
+        marked as "dirty", *even those which have no net changes
+        to their column-based attributes*. An object is marked
+        as dirty when any of its column-based attributes have a
+        "set attribute" operation called or when any of its
+        collections are modified. If, at update time, no
+        column-based attributes have any net changes, no UPDATE
+        statement will be issued. This means that an instance
+        being sent to :meth:`~.MapperEvents.before_update` is
+        *not* a guarantee that an UPDATE statement will be
+        issued, although you can affect the outcome here by
+        modifying attributes so that a net change in value does
+        exist.
+
+        To detect if the column-based attributes on the object have net
+        changes, and will therefore generate an UPDATE statement, use
+        ``object_session(instance).is_modified(instance,
+        include_collections=False)``.
+
+        The event is often called for a batch of objects of the
+        same class before their UPDATE statements are emitted at
+        once in a later step. In the extremely rare case that
+        this is not desirable, the :class:`_orm.Mapper` can be
+        configured with ``batch=False``, which will cause
+        batches of instances to be broken up into individual
+        (and more poorly performing) event->persist->event
+        steps.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit UPDATE statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being persisted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_update(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance after an UPDATE statement
+        is emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to modify in-Python-only
+        state on the instance after an UPDATE occurs, as well
+        as to emit additional SQL statements on the given
+        connection.
+
+        This method is called for all instances that are
+        marked as "dirty", *even those which have no net changes
+        to their column-based attributes*, and for which
+        no UPDATE statement has proceeded. An object is marked
+        as dirty when any of its column-based attributes have a
+        "set attribute" operation called or when any of its
+        collections are modified. If, at update time, no
+        column-based attributes have any net changes, no UPDATE
+        statement will be issued. This means that an instance
+        being sent to :meth:`~.MapperEvents.after_update` is
+        *not* a guarantee that an UPDATE statement has been
+        issued.
+
+        To detect if the column-based attributes on the object have net
+        changes, and therefore resulted in an UPDATE statement, use
+        ``object_session(instance).is_modified(instance,
+        include_collections=False)``.
+
+        The event is often called for a batch of objects of the
+        same class after their UPDATE statements have been emitted at
+        once in a previous step. In the extremely rare case that
+        this is not desirable, the :class:`_orm.Mapper` can be
+        configured with ``batch=False``, which will cause
+        batches of instances to be broken up into individual
+        (and more poorly performing) event->persist->event
+        steps.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit UPDATE statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being persisted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def before_delete(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance before a DELETE statement
+        is emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to emit additional SQL statements on
+        the given connection as well as to perform application
+        specific bookkeeping related to a deletion event.
+
+        The event is often called for a batch of objects of the
+        same class before their DELETE statements are emitted at
+        once in a later step.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit DELETE statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being deleted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_delete(
+        self, mapper: Mapper[_O], connection: Connection, target: _O
+    ) -> None:
+        """Receive an object instance after a DELETE statement
+        has been emitted corresponding to that instance.
+
+        .. note:: this event **only** applies to the
+           :ref:`session flush operation <session_flushing>`
+           and does **not** apply to the ORM DML operations described at
+           :ref:`orm_expression_update_delete`.  To intercept ORM
+           DML events, use :meth:`_orm.SessionEvents.do_orm_execute`.
+
+        This event is used to emit additional SQL statements on
+        the given connection as well as to perform application
+        specific bookkeeping related to a deletion event.
+
+        The event is often called for a batch of objects of the
+        same class after their DELETE statements have been emitted at
+        once in a previous step.
+
+        .. warning::
+
+            Mapper-level flush events only allow **very limited operations**,
+            on attributes local to the row being operated upon only,
+            as well as allowing any SQL to be emitted on the given
+            :class:`_engine.Connection`.  **Please read fully** the notes
+            at :ref:`session_persistence_mapper` for guidelines on using
+            these methods; generally, the :meth:`.SessionEvents.before_flush`
+            method should be preferred for general on-flush changes.
+
+        :param mapper: the :class:`_orm.Mapper` which is the target
+         of this event.
+        :param connection: the :class:`_engine.Connection` being used to
+         emit DELETE statements for this instance.  This
+         provides a handle into the current transaction on the
+         target database specific to this instance.
+        :param target: the mapped instance being deleted.  If
+         the event is configured with ``raw=True``, this will
+         instead be the :class:`.InstanceState` state-management
+         object associated with the instance.
+        :return: No return value is supported by this event.
+
+        .. seealso::
+
+            :ref:`session_persistence_events`
+
+        """
+
+
+class _MapperEventsHold(_EventsHold[_ET]):
+    all_holds = weakref.WeakKeyDictionary()
+
+    def resolve(
+        self, class_: Union[Type[_T], _InternalEntityType[_T]]
+    ) -> Optional[Mapper[_T]]:
+        return _mapper_or_none(class_)
+
+    class HoldMapperEvents(_EventsHold.HoldEvents[_ET], MapperEvents):  # type: ignore [misc] # noqa: E501
+        pass
+
+    dispatch = event.dispatcher(HoldMapperEvents)
+
+
+_sessionevents_lifecycle_event_names: Set[str] = set()
+
+
+class SessionEvents(event.Events[Session]):
+    """Define events specific to :class:`.Session` lifecycle.
+
+    e.g.::
+
+        from sqlalchemy import event
+        from sqlalchemy.orm import sessionmaker
+
+
+        def my_before_commit(session):
+            print("before commit!")
+
+
+        Session = sessionmaker()
+
+        event.listen(Session, "before_commit", my_before_commit)
+
+    The :func:`~.event.listen` function will accept
+    :class:`.Session` objects as well as the return result
+    of :class:`~.sessionmaker()` and :class:`~.scoped_session()`.
+
+    Additionally, it accepts the :class:`.Session` class which
+    will apply listeners to all :class:`.Session` instances
+    globally.
+
+    :param raw=False: When True, the "target" argument passed
+       to applicable event listener functions that work on individual
+       objects will be the instance's :class:`.InstanceState` management
+       object, rather than the mapped instance itself.
+
+       .. versionadded:: 1.3.14
+
+    :param restore_load_context=False: Applies to the
+       :meth:`.SessionEvents.loaded_as_persistent` event.  Restores the loader
+       context of the object when the event hook is complete, so that ongoing
+       eager load operations continue to target the object appropriately.  A
+       warning is emitted if the object is moved to a new loader context from
+       within this event if this flag is not set.
+
+       .. versionadded:: 1.3.14
+
+    """
+
+    _target_class_doc = "SomeSessionClassOrObject"
+
+    _dispatch_target = Session
+
+    def _lifecycle_event(  # type: ignore [misc]
+        fn: Callable[[SessionEvents, Session, Any], None]
+    ) -> Callable[[SessionEvents, Session, Any], None]:
+        _sessionevents_lifecycle_event_names.add(fn.__name__)
+        return fn
+
+    @classmethod
+    def _accept_with(  # type: ignore [return]
+        cls, target: Any, identifier: str
+    ) -> Union[Session, type]:
+        if isinstance(target, scoped_session):
+            target = target.session_factory
+            if not isinstance(target, sessionmaker) and (
+                not isinstance(target, type) or not issubclass(target, Session)
+            ):
+                raise exc.ArgumentError(
+                    "Session event listen on a scoped_session "
+                    "requires that its creation callable "
+                    "is associated with the Session class."
+                )
+
+        if isinstance(target, sessionmaker):
+            return target.class_
+        elif isinstance(target, type):
+            if issubclass(target, scoped_session):
+                return Session
+            elif issubclass(target, Session):
+                return target
+        elif isinstance(target, Session):
+            return target
+        elif hasattr(target, "_no_async_engine_events"):
+            target._no_async_engine_events()
+        else:
+            # allows alternate SessionEvents-like-classes to be consulted
+            return event.Events._accept_with(target, identifier)  # type: ignore [return-value] # noqa: E501
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: Any,
+        *,
+        raw: bool = False,
+        restore_load_context: bool = False,
+        **kw: Any,
+    ) -> None:
+        is_instance_event = (
+            event_key.identifier in _sessionevents_lifecycle_event_names
+        )
+
+        if is_instance_event:
+            if not raw or restore_load_context:
+                fn = event_key._listen_fn
+
+                def wrap(
+                    session: Session,
+                    state: InstanceState[_O],
+                    *arg: Any,
+                    **kw: Any,
+                ) -> Optional[Any]:
+                    if not raw:
+                        target = state.obj()
+                        if target is None:
+                            # existing behavior is that if the object is
+                            # garbage collected, no event is emitted
+                            return None
+                    else:
+                        target = state  # type: ignore [assignment]
+                    if restore_load_context:
+                        runid = state.runid
+                    try:
+                        return fn(session, target, *arg, **kw)
+                    finally:
+                        if restore_load_context:
+                            state.runid = runid
+
+                event_key = event_key.with_wrapper(wrap)
+
+        event_key.base_listen(**kw)
+
+    def do_orm_execute(self, orm_execute_state: ORMExecuteState) -> None:
+        """Intercept statement executions that occur on behalf of an
+        ORM :class:`.Session` object.
+
+        This event is invoked for all top-level SQL statements invoked from the
+        :meth:`_orm.Session.execute` method, as well as related methods such as
+        :meth:`_orm.Session.scalars` and :meth:`_orm.Session.scalar`. As of
+        SQLAlchemy 1.4, all ORM queries that run through the
+        :meth:`_orm.Session.execute` method as well as related methods
+        :meth:`_orm.Session.scalars`, :meth:`_orm.Session.scalar` etc.
+        will participate in this event.
+        This event hook does **not** apply to the queries that are
+        emitted internally within the ORM flush process, i.e. the
+        process described at :ref:`session_flushing`.
+
+        .. note::  The :meth:`_orm.SessionEvents.do_orm_execute` event hook
+           is triggered **for ORM statement executions only**, meaning those
+           invoked via the :meth:`_orm.Session.execute` and similar methods on
+           the :class:`_orm.Session` object. It does **not** trigger for
+           statements that are invoked by SQLAlchemy Core only, i.e. statements
+           invoked directly using :meth:`_engine.Connection.execute` or
+           otherwise originating from an :class:`_engine.Engine` object without
+           any :class:`_orm.Session` involved. To intercept **all** SQL
+           executions regardless of whether the Core or ORM APIs are in use,
+           see the event hooks at :class:`.ConnectionEvents`, such as
+           :meth:`.ConnectionEvents.before_execute` and
+           :meth:`.ConnectionEvents.before_cursor_execute`.
+
+           Also, this event hook does **not** apply to queries that are
+           emitted internally within the ORM flush process,
+           i.e. the process described at :ref:`session_flushing`; to
+           intercept steps within the flush process, see the event
+           hooks described at :ref:`session_persistence_events` as
+           well as :ref:`session_persistence_mapper`.
+
+        This event is a ``do_`` event, meaning it has the capability to replace
+        the operation that the :meth:`_orm.Session.execute` method normally
+        performs.  The intended use for this includes sharding and
+        result-caching schemes which may seek to invoke the same statement
+        across  multiple database connections, returning a result that is
+        merged from each of them, or which don't invoke the statement at all,
+        instead returning data from a cache.
+
+        The hook intends to replace the use of the
+        ``Query._execute_and_instances`` method that could be subclassed prior
+        to SQLAlchemy 1.4.
+
+        :param orm_execute_state: an instance of :class:`.ORMExecuteState`
+         which contains all information about the current execution, as well
+         as helper functions used to derive other commonly required
+         information.   See that object for details.
+
+        .. seealso::
+
+            :ref:`session_execute_events` - top level documentation on how
+            to use :meth:`_orm.SessionEvents.do_orm_execute`
+
+            :class:`.ORMExecuteState` - the object passed to the
+            :meth:`_orm.SessionEvents.do_orm_execute` event which contains
+            all information about the statement to be invoked.  It also
+            provides an interface to extend the current statement, options,
+            and parameters as well as an option that allows programmatic
+            invocation of the statement at any point.
+
+            :ref:`examples_session_orm_events` - includes examples of using
+            :meth:`_orm.SessionEvents.do_orm_execute`
+
+            :ref:`examples_caching` - an example of how to integrate
+            Dogpile caching with the ORM :class:`_orm.Session` making use
+            of the :meth:`_orm.SessionEvents.do_orm_execute` event hook.
+
+            :ref:`examples_sharding` - the Horizontal Sharding example /
+            extension relies upon the
+            :meth:`_orm.SessionEvents.do_orm_execute` event hook to invoke a
+            SQL statement on multiple backends and return a merged result.
+
+
+        .. versionadded:: 1.4
+
+        """
+
+    def after_transaction_create(
+        self, session: Session, transaction: SessionTransaction
+    ) -> None:
+        """Execute when a new :class:`.SessionTransaction` is created.
+
+        This event differs from :meth:`~.SessionEvents.after_begin`
+        in that it occurs for each :class:`.SessionTransaction`
+        overall, as opposed to when transactions are begun
+        on individual database connections.  It is also invoked
+        for nested transactions and subtransactions, and is always
+        matched by a corresponding
+        :meth:`~.SessionEvents.after_transaction_end` event
+        (assuming normal operation of the :class:`.Session`).
+
+        :param session: the target :class:`.Session`.
+        :param transaction: the target :class:`.SessionTransaction`.
+
+         To detect if this is the outermost
+         :class:`.SessionTransaction`, as opposed to a "subtransaction" or a
+         SAVEPOINT, test that the :attr:`.SessionTransaction.parent` attribute
+         is ``None``::
+
+                @event.listens_for(session, "after_transaction_create")
+                def after_transaction_create(session, transaction):
+                    if transaction.parent is None:
+                        ...  # work with top-level transaction
+
+         To detect if the :class:`.SessionTransaction` is a SAVEPOINT, use the
+         :attr:`.SessionTransaction.nested` attribute::
+
+                @event.listens_for(session, "after_transaction_create")
+                def after_transaction_create(session, transaction):
+                    if transaction.nested:
+                        ...  # work with SAVEPOINT transaction
+
+        .. seealso::
+
+            :class:`.SessionTransaction`
+
+            :meth:`~.SessionEvents.after_transaction_end`
+
+        """
+
+    def after_transaction_end(
+        self, session: Session, transaction: SessionTransaction
+    ) -> None:
+        """Execute when the span of a :class:`.SessionTransaction` ends.
+
+        This event differs from :meth:`~.SessionEvents.after_commit`
+        in that it corresponds to all :class:`.SessionTransaction`
+        objects in use, including those for nested transactions
+        and subtransactions, and is always matched by a corresponding
+        :meth:`~.SessionEvents.after_transaction_create` event.
+
+        :param session: the target :class:`.Session`.
+        :param transaction: the target :class:`.SessionTransaction`.
+
+         To detect if this is the outermost
+         :class:`.SessionTransaction`, as opposed to a "subtransaction" or a
+         SAVEPOINT, test that the :attr:`.SessionTransaction.parent` attribute
+         is ``None``::
+
+                @event.listens_for(session, "after_transaction_create")
+                def after_transaction_end(session, transaction):
+                    if transaction.parent is None:
+                        ...  # work with top-level transaction
+
+         To detect if the :class:`.SessionTransaction` is a SAVEPOINT, use the
+         :attr:`.SessionTransaction.nested` attribute::
+
+                @event.listens_for(session, "after_transaction_create")
+                def after_transaction_end(session, transaction):
+                    if transaction.nested:
+                        ...  # work with SAVEPOINT transaction
+
+        .. seealso::
+
+            :class:`.SessionTransaction`
+
+            :meth:`~.SessionEvents.after_transaction_create`
+
+        """
+
+    def before_commit(self, session: Session) -> None:
+        """Execute before commit is called.
+
+        .. note::
+
+            The :meth:`~.SessionEvents.before_commit` hook is *not* per-flush,
+            that is, the :class:`.Session` can emit SQL to the database
+            many times within the scope of a transaction.
+            For interception of these events, use the
+            :meth:`~.SessionEvents.before_flush`,
+            :meth:`~.SessionEvents.after_flush`, or
+            :meth:`~.SessionEvents.after_flush_postexec`
+            events.
+
+        :param session: The target :class:`.Session`.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.after_commit`
+
+            :meth:`~.SessionEvents.after_begin`
+
+            :meth:`~.SessionEvents.after_transaction_create`
+
+            :meth:`~.SessionEvents.after_transaction_end`
+
+        """
+
+    def after_commit(self, session: Session) -> None:
+        """Execute after a commit has occurred.
+
+        .. note::
+
+            The :meth:`~.SessionEvents.after_commit` hook is *not* per-flush,
+            that is, the :class:`.Session` can emit SQL to the database
+            many times within the scope of a transaction.
+            For interception of these events, use the
+            :meth:`~.SessionEvents.before_flush`,
+            :meth:`~.SessionEvents.after_flush`, or
+            :meth:`~.SessionEvents.after_flush_postexec`
+            events.
+
+        .. note::
+
+            The :class:`.Session` is not in an active transaction
+            when the :meth:`~.SessionEvents.after_commit` event is invoked,
+            and therefore can not emit SQL.  To emit SQL corresponding to
+            every transaction, use the :meth:`~.SessionEvents.before_commit`
+            event.
+
+        :param session: The target :class:`.Session`.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.before_commit`
+
+            :meth:`~.SessionEvents.after_begin`
+
+            :meth:`~.SessionEvents.after_transaction_create`
+
+            :meth:`~.SessionEvents.after_transaction_end`
+
+        """
+
+    def after_rollback(self, session: Session) -> None:
+        """Execute after a real DBAPI rollback has occurred.
+
+        Note that this event only fires when the *actual* rollback against
+        the database occurs - it does *not* fire each time the
+        :meth:`.Session.rollback` method is called, if the underlying
+        DBAPI transaction has already been rolled back.  In many
+        cases, the :class:`.Session` will not be in
+        an "active" state during this event, as the current
+        transaction is not valid.   To acquire a :class:`.Session`
+        which is active after the outermost rollback has proceeded,
+        use the :meth:`.SessionEvents.after_soft_rollback` event, checking the
+        :attr:`.Session.is_active` flag.
+
+        :param session: The target :class:`.Session`.
+
+        """
+
+    def after_soft_rollback(
+        self, session: Session, previous_transaction: SessionTransaction
+    ) -> None:
+        """Execute after any rollback has occurred, including "soft"
+        rollbacks that don't actually emit at the DBAPI level.
+
+        This corresponds to both nested and outer rollbacks, i.e.
+        the innermost rollback that calls the DBAPI's
+        rollback() method, as well as the enclosing rollback
+        calls that only pop themselves from the transaction stack.
+
+        The given :class:`.Session` can be used to invoke SQL and
+        :meth:`.Session.query` operations after an outermost rollback
+        by first checking the :attr:`.Session.is_active` flag::
+
+            @event.listens_for(Session, "after_soft_rollback")
+            def do_something(session, previous_transaction):
+                if session.is_active:
+                    session.execute(text("select * from some_table"))
+
+        :param session: The target :class:`.Session`.
+        :param previous_transaction: The :class:`.SessionTransaction`
+         transactional marker object which was just closed.   The current
+         :class:`.SessionTransaction` for the given :class:`.Session` is
+         available via the :attr:`.Session.transaction` attribute.
+
+        """
+
+    def before_flush(
+        self,
+        session: Session,
+        flush_context: UOWTransaction,
+        instances: Optional[Sequence[_O]],
+    ) -> None:
+        """Execute before flush process has started.
+
+        :param session: The target :class:`.Session`.
+        :param flush_context: Internal :class:`.UOWTransaction` object
+         which handles the details of the flush.
+        :param instances: Usually ``None``, this is the collection of
+         objects which can be passed to the :meth:`.Session.flush` method
+         (note this usage is deprecated).
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.after_flush`
+
+            :meth:`~.SessionEvents.after_flush_postexec`
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_flush(
+        self, session: Session, flush_context: UOWTransaction
+    ) -> None:
+        """Execute after flush has completed, but before commit has been
+        called.
+
+        Note that the session's state is still in pre-flush, i.e. 'new',
+        'dirty', and 'deleted' lists still show pre-flush state as well
+        as the history settings on instance attributes.
+
+        .. warning:: This event runs after the :class:`.Session` has emitted
+           SQL to modify the database, but **before** it has altered its
+           internal state to reflect those changes, including that newly
+           inserted objects are placed into the identity map.  ORM operations
+           emitted within this event such as loads of related items
+           may produce new identity map entries that will immediately
+           be replaced, sometimes causing confusing results.  SQLAlchemy will
+           emit a warning for this condition as of version 1.3.9.
+
+        :param session: The target :class:`.Session`.
+        :param flush_context: Internal :class:`.UOWTransaction` object
+         which handles the details of the flush.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.before_flush`
+
+            :meth:`~.SessionEvents.after_flush_postexec`
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_flush_postexec(
+        self, session: Session, flush_context: UOWTransaction
+    ) -> None:
+        """Execute after flush has completed, and after the post-exec
+        state occurs.
+
+        This will be when the 'new', 'dirty', and 'deleted' lists are in
+        their final state.  An actual commit() may or may not have
+        occurred, depending on whether or not the flush started its own
+        transaction or participated in a larger transaction.
+
+        :param session: The target :class:`.Session`.
+        :param flush_context: Internal :class:`.UOWTransaction` object
+         which handles the details of the flush.
+
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.before_flush`
+
+            :meth:`~.SessionEvents.after_flush`
+
+            :ref:`session_persistence_events`
+
+        """
+
+    def after_begin(
+        self,
+        session: Session,
+        transaction: SessionTransaction,
+        connection: Connection,
+    ) -> None:
+        """Execute after a transaction is begun on a connection.
+
+        .. note:: This event is called within the process of the
+          :class:`_orm.Session` modifying its own internal state.
+          To invoke SQL operations within this hook, use the
+          :class:`_engine.Connection` provided to the event;
+          do not run SQL operations using the :class:`_orm.Session`
+          directly.
+
+        :param session: The target :class:`.Session`.
+        :param transaction: The :class:`.SessionTransaction`.
+        :param connection: The :class:`_engine.Connection` object
+         which will be used for SQL statements.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.before_commit`
+
+            :meth:`~.SessionEvents.after_commit`
+
+            :meth:`~.SessionEvents.after_transaction_create`
+
+            :meth:`~.SessionEvents.after_transaction_end`
+
+        """
+
+    @_lifecycle_event
+    def before_attach(self, session: Session, instance: _O) -> None:
+        """Execute before an instance is attached to a session.
+
+        This is called before an add, delete or merge causes
+        the object to be part of the session.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.after_attach`
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def after_attach(self, session: Session, instance: _O) -> None:
+        """Execute after an instance is attached to a session.
+
+        This is called after an add, delete or merge.
+
+        .. note::
+
+           As of 0.8, this event fires off *after* the item
+           has been fully associated with the session, which is
+           different than previous releases.  For event
+           handlers that require the object not yet
+           be part of session state (such as handlers which
+           may autoflush while the target object is not
+           yet complete) consider the
+           new :meth:`.before_attach` event.
+
+        .. seealso::
+
+            :meth:`~.SessionEvents.before_attach`
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @event._legacy_signature(
+        "0.9",
+        ["session", "query", "query_context", "result"],
+        lambda update_context: (
+            update_context.session,
+            update_context.query,
+            None,
+            update_context.result,
+        ),
+    )
+    def after_bulk_update(self, update_context: _O) -> None:
+        """Event for after the legacy :meth:`_orm.Query.update` method
+        has been called.
+
+        .. legacy:: The :meth:`_orm.SessionEvents.after_bulk_update` method
+           is a legacy event hook as of SQLAlchemy 2.0.   The event
+           **does not participate** in :term:`2.0 style` invocations
+           using :func:`_dml.update` documented at
+           :ref:`orm_queryguide_update_delete_where`.  For 2.0 style use,
+           the :meth:`_orm.SessionEvents.do_orm_execute` hook will intercept
+           these calls.
+
+        :param update_context: an "update context" object which contains
+         details about the update, including these attributes:
+
+            * ``session`` - the :class:`.Session` involved
+            * ``query`` -the :class:`_query.Query`
+              object that this update operation
+              was called upon.
+            * ``values`` The "values" dictionary that was passed to
+              :meth:`_query.Query.update`.
+            * ``result`` the :class:`_engine.CursorResult`
+              returned as a result of the
+              bulk UPDATE operation.
+
+        .. versionchanged:: 1.4 the update_context no longer has a
+           ``QueryContext`` object associated with it.
+
+        .. seealso::
+
+            :meth:`.QueryEvents.before_compile_update`
+
+            :meth:`.SessionEvents.after_bulk_delete`
+
+        """
+
+    @event._legacy_signature(
+        "0.9",
+        ["session", "query", "query_context", "result"],
+        lambda delete_context: (
+            delete_context.session,
+            delete_context.query,
+            None,
+            delete_context.result,
+        ),
+    )
+    def after_bulk_delete(self, delete_context: _O) -> None:
+        """Event for after the legacy :meth:`_orm.Query.delete` method
+        has been called.
+
+        .. legacy:: The :meth:`_orm.SessionEvents.after_bulk_delete` method
+           is a legacy event hook as of SQLAlchemy 2.0.   The event
+           **does not participate** in :term:`2.0 style` invocations
+           using :func:`_dml.delete` documented at
+           :ref:`orm_queryguide_update_delete_where`.  For 2.0 style use,
+           the :meth:`_orm.SessionEvents.do_orm_execute` hook will intercept
+           these calls.
+
+        :param delete_context: a "delete context" object which contains
+         details about the update, including these attributes:
+
+            * ``session`` - the :class:`.Session` involved
+            * ``query`` -the :class:`_query.Query`
+              object that this update operation
+              was called upon.
+            * ``result`` the :class:`_engine.CursorResult`
+              returned as a result of the
+              bulk DELETE operation.
+
+        .. versionchanged:: 1.4 the update_context no longer has a
+           ``QueryContext`` object associated with it.
+
+        .. seealso::
+
+            :meth:`.QueryEvents.before_compile_delete`
+
+            :meth:`.SessionEvents.after_bulk_update`
+
+        """
+
+    @_lifecycle_event
+    def transient_to_pending(self, session: Session, instance: _O) -> None:
+        """Intercept the "transient to pending" transition for a specific
+        object.
+
+        This event is a specialization of the
+        :meth:`.SessionEvents.after_attach` event which is only invoked
+        for this specific transition.  It is invoked typically during the
+        :meth:`.Session.add` call.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def pending_to_transient(self, session: Session, instance: _O) -> None:
+        """Intercept the "pending to transient" transition for a specific
+        object.
+
+        This less common transition occurs when an pending object that has
+        not been flushed is evicted from the session; this can occur
+        when the :meth:`.Session.rollback` method rolls back the transaction,
+        or when the :meth:`.Session.expunge` method is used.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def persistent_to_transient(self, session: Session, instance: _O) -> None:
+        """Intercept the "persistent to transient" transition for a specific
+        object.
+
+        This less common transition occurs when an pending object that has
+        has been flushed is evicted from the session; this can occur
+        when the :meth:`.Session.rollback` method rolls back the transaction.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def pending_to_persistent(self, session: Session, instance: _O) -> None:
+        """Intercept the "pending to persistent"" transition for a specific
+        object.
+
+        This event is invoked within the flush process, and is
+        similar to scanning the :attr:`.Session.new` collection within
+        the :meth:`.SessionEvents.after_flush` event.  However, in this
+        case the object has already been moved to the persistent state
+        when the event is called.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def detached_to_persistent(self, session: Session, instance: _O) -> None:
+        """Intercept the "detached to persistent" transition for a specific
+        object.
+
+        This event is a specialization of the
+        :meth:`.SessionEvents.after_attach` event which is only invoked
+        for this specific transition.  It is invoked typically during the
+        :meth:`.Session.add` call, as well as during the
+        :meth:`.Session.delete` call if the object was not previously
+        associated with the
+        :class:`.Session` (note that an object marked as "deleted" remains
+        in the "persistent" state until the flush proceeds).
+
+        .. note::
+
+            If the object becomes persistent as part of a call to
+            :meth:`.Session.delete`, the object is **not** yet marked as
+            deleted when this event is called.  To detect deleted objects,
+            check the ``deleted`` flag sent to the
+            :meth:`.SessionEvents.persistent_to_detached` to event after the
+            flush proceeds, or check the :attr:`.Session.deleted` collection
+            within the :meth:`.SessionEvents.before_flush` event if deleted
+            objects need to be intercepted before the flush.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def loaded_as_persistent(self, session: Session, instance: _O) -> None:
+        """Intercept the "loaded as persistent" transition for a specific
+        object.
+
+        This event is invoked within the ORM loading process, and is invoked
+        very similarly to the :meth:`.InstanceEvents.load` event.  However,
+        the event here is linkable to a :class:`.Session` class or instance,
+        rather than to a mapper or class hierarchy, and integrates
+        with the other session lifecycle events smoothly.  The object
+        is guaranteed to be present in the session's identity map when
+        this event is called.
+
+        .. note:: This event is invoked within the loader process before
+           eager loaders may have been completed, and the object's state may
+           not be complete.  Additionally, invoking row-level refresh
+           operations on the object will place the object into a new loader
+           context, interfering with the existing load context.   See the note
+           on :meth:`.InstanceEvents.load` for background on making use of the
+           :paramref:`.SessionEvents.restore_load_context` parameter, which
+           works in the same manner as that of
+           :paramref:`.InstanceEvents.restore_load_context`, in  order to
+           resolve this scenario.
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def persistent_to_deleted(self, session: Session, instance: _O) -> None:
+        """Intercept the "persistent to deleted" transition for a specific
+        object.
+
+        This event is invoked when a persistent object's identity
+        is deleted from the database within a flush, however the object
+        still remains associated with the :class:`.Session` until the
+        transaction completes.
+
+        If the transaction is rolled back, the object moves again
+        to the persistent state, and the
+        :meth:`.SessionEvents.deleted_to_persistent` event is called.
+        If the transaction is committed, the object becomes detached,
+        which will emit the :meth:`.SessionEvents.deleted_to_detached`
+        event.
+
+        Note that while the :meth:`.Session.delete` method is the primary
+        public interface to mark an object as deleted, many objects
+        get deleted due to cascade rules, which are not always determined
+        until flush time.  Therefore, there's no way to catch
+        every object that will be deleted until the flush has proceeded.
+        the :meth:`.SessionEvents.persistent_to_deleted` event is therefore
+        invoked at the end of a flush.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def deleted_to_persistent(self, session: Session, instance: _O) -> None:
+        """Intercept the "deleted to persistent" transition for a specific
+        object.
+
+        This transition occurs only when an object that's been deleted
+        successfully in a flush is restored due to a call to
+        :meth:`.Session.rollback`.   The event is not called under
+        any other circumstances.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def deleted_to_detached(self, session: Session, instance: _O) -> None:
+        """Intercept the "deleted to detached" transition for a specific
+        object.
+
+        This event is invoked when a deleted object is evicted
+        from the session.   The typical case when this occurs is when
+        the transaction for a :class:`.Session` in which the object
+        was deleted is committed; the object moves from the deleted
+        state to the detached state.
+
+        It is also invoked for objects that were deleted in a flush
+        when the :meth:`.Session.expunge_all` or :meth:`.Session.close`
+        events are called, as well as if the object is individually
+        expunged from its deleted state via :meth:`.Session.expunge`.
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+    @_lifecycle_event
+    def persistent_to_detached(self, session: Session, instance: _O) -> None:
+        """Intercept the "persistent to detached" transition for a specific
+        object.
+
+        This event is invoked when a persistent object is evicted
+        from the session.  There are many conditions that cause this
+        to happen, including:
+
+        * using a method such as :meth:`.Session.expunge`
+          or :meth:`.Session.close`
+
+        * Calling the :meth:`.Session.rollback` method, when the object
+          was part of an INSERT statement for that session's transaction
+
+
+        :param session: target :class:`.Session`
+
+        :param instance: the ORM-mapped instance being operated upon.
+
+        :param deleted: boolean.  If True, indicates this object moved
+         to the detached state because it was marked as deleted and flushed.
+
+
+        .. seealso::
+
+            :ref:`session_lifecycle_events`
+
+        """
+
+
+class AttributeEvents(event.Events[QueryableAttribute[Any]]):
+    r"""Define events for object attributes.
+
+    These are typically defined on the class-bound descriptor for the
+    target class.
+
+    For example, to register a listener that will receive the
+    :meth:`_orm.AttributeEvents.append` event::
+
+        from sqlalchemy import event
+
+
+        @event.listens_for(MyClass.collection, "append", propagate=True)
+        def my_append_listener(target, value, initiator):
+            print("received append event for target: %s" % target)
+
+    Listeners have the option to return a possibly modified version of the
+    value, when the :paramref:`.AttributeEvents.retval` flag is passed to
+    :func:`.event.listen` or :func:`.event.listens_for`, such as below,
+    illustrated using the :meth:`_orm.AttributeEvents.set` event::
+
+        def validate_phone(target, value, oldvalue, initiator):
+            "Strip non-numeric characters from a phone number"
+
+            return re.sub(r"\D", "", value)
+
+
+        # setup listener on UserContact.phone attribute, instructing
+        # it to use the return value
+        listen(UserContact.phone, "set", validate_phone, retval=True)
+
+    A validation function like the above can also raise an exception
+    such as :exc:`ValueError` to halt the operation.
+
+    The :paramref:`.AttributeEvents.propagate` flag is also important when
+    applying listeners to mapped classes that also have mapped subclasses,
+    as when using mapper inheritance patterns::
+
+
+        @event.listens_for(MySuperClass.attr, "set", propagate=True)
+        def receive_set(target, value, initiator):
+            print("value set: %s" % target)
+
+    The full list of modifiers available to the :func:`.event.listen`
+    and :func:`.event.listens_for` functions are below.
+
+    :param active_history=False: When True, indicates that the
+      "set" event would like to receive the "old" value being
+      replaced unconditionally, even if this requires firing off
+      database loads. Note that ``active_history`` can also be
+      set directly via :func:`.column_property` and
+      :func:`_orm.relationship`.
+
+    :param propagate=False: When True, the listener function will
+      be established not just for the class attribute given, but
+      for attributes of the same name on all current subclasses
+      of that class, as well as all future subclasses of that
+      class, using an additional listener that listens for
+      instrumentation events.
+    :param raw=False: When True, the "target" argument to the
+      event will be the :class:`.InstanceState` management
+      object, rather than the mapped instance itself.
+    :param retval=False: when True, the user-defined event
+      listening must return the "value" argument from the
+      function.  This gives the listening function the opportunity
+      to change the value that is ultimately used for a "set"
+      or "append" event.
+
+    """
+
+    _target_class_doc = "SomeClass.some_attribute"
+    _dispatch_target = QueryableAttribute
+
+    @staticmethod
+    def _set_dispatch(
+        cls: Type[_HasEventsDispatch[Any]], dispatch_cls: Type[_Dispatch[Any]]
+    ) -> _Dispatch[Any]:
+        dispatch = event.Events._set_dispatch(cls, dispatch_cls)
+        dispatch_cls._active_history = False
+        return dispatch
+
+    @classmethod
+    def _accept_with(
+        cls,
+        target: Union[QueryableAttribute[Any], Type[QueryableAttribute[Any]]],
+        identifier: str,
+    ) -> Union[QueryableAttribute[Any], Type[QueryableAttribute[Any]]]:
+        # TODO: coverage
+        if isinstance(target, interfaces.MapperProperty):
+            return getattr(target.parent.class_, target.key)
+        else:
+            return target
+
+    @classmethod
+    def _listen(  # type: ignore [override]
+        cls,
+        event_key: _EventKey[QueryableAttribute[Any]],
+        active_history: bool = False,
+        raw: bool = False,
+        retval: bool = False,
+        propagate: bool = False,
+        include_key: bool = False,
+    ) -> None:
+        target, fn = event_key.dispatch_target, event_key._listen_fn
+
+        if active_history:
+            target.dispatch._active_history = True
+
+        if not raw or not retval or not include_key:
+
+            def wrap(target: InstanceState[_O], *arg: Any, **kw: Any) -> Any:
+                if not raw:
+                    target = target.obj()  # type: ignore [assignment]
+                if not retval:
+                    if arg:
+                        value = arg[0]
+                    else:
+                        value = None
+                    if include_key:
+                        fn(target, *arg, **kw)
+                    else:
+                        fn(target, *arg)
+                    return value
+                else:
+                    if include_key:
+                        return fn(target, *arg, **kw)
+                    else:
+                        return fn(target, *arg)
+
+            event_key = event_key.with_wrapper(wrap)
+
+        event_key.base_listen(propagate=propagate)
+
+        if propagate:
+            manager = instrumentation.manager_of_class(target.class_)
+
+            for mgr in manager.subclass_managers(True):  # type: ignore [no-untyped-call] # noqa: E501
+                event_key.with_dispatch_target(mgr[target.key]).base_listen(
+                    propagate=True
+                )
+                if active_history:
+                    mgr[target.key].dispatch._active_history = True
+
+    def append(
+        self,
+        target: _O,
+        value: _T,
+        initiator: Event,
+        *,
+        key: EventConstants = NO_KEY,
+    ) -> Optional[_T]:
+        """Receive a collection append event.
+
+        The append event is invoked for each element as it is appended
+        to the collection.  This occurs for single-item appends as well
+        as for a "bulk replace" operation.
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+        :param value: the value being appended.  If this listener
+          is registered with ``retval=True``, the listener
+          function must return this value, or a new value which
+          replaces it.
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.  May be modified
+          from its original value by backref handlers in order to control
+          chained event propagation, as well as be inspected for information
+          about the source of the event.
+        :param key: When the event is established using the
+         :paramref:`.AttributeEvents.include_key` parameter set to
+         True, this will be the key used in the operation, such as
+         ``collection[some_key_or_index] = value``.
+         The parameter is not passed
+         to the event at all if the the
+         :paramref:`.AttributeEvents.include_key`
+         was not used to set up the event; this is to allow backwards
+         compatibility with existing event handlers that don't include the
+         ``key`` parameter.
+
+         .. versionadded:: 2.0
+
+        :return: if the event was registered with ``retval=True``,
+         the given value, or a new effective value, should be returned.
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+            :meth:`.AttributeEvents.bulk_replace`
+
+        """
+
+    def append_wo_mutation(
+        self,
+        target: _O,
+        value: _T,
+        initiator: Event,
+        *,
+        key: EventConstants = NO_KEY,
+    ) -> None:
+        """Receive a collection append event where the collection was not
+        actually mutated.
+
+        This event differs from :meth:`_orm.AttributeEvents.append` in that
+        it is fired off for de-duplicating collections such as sets and
+        dictionaries, when the object already exists in the target collection.
+        The event does not have a return value and the identity of the
+        given object cannot be changed.
+
+        The event is used for cascading objects into a :class:`_orm.Session`
+        when the collection has already been mutated via a backref event.
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+        :param value: the value that would be appended if the object did not
+          already exist in the collection.
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.  May be modified
+          from its original value by backref handlers in order to control
+          chained event propagation, as well as be inspected for information
+          about the source of the event.
+        :param key: When the event is established using the
+         :paramref:`.AttributeEvents.include_key` parameter set to
+         True, this will be the key used in the operation, such as
+         ``collection[some_key_or_index] = value``.
+         The parameter is not passed
+         to the event at all if the the
+         :paramref:`.AttributeEvents.include_key`
+         was not used to set up the event; this is to allow backwards
+         compatibility with existing event handlers that don't include the
+         ``key`` parameter.
+
+         .. versionadded:: 2.0
+
+        :return: No return value is defined for this event.
+
+        .. versionadded:: 1.4.15
+
+        """
+
+    def bulk_replace(
+        self,
+        target: _O,
+        values: Iterable[_T],
+        initiator: Event,
+        *,
+        keys: Optional[Iterable[EventConstants]] = None,
+    ) -> None:
+        """Receive a collection 'bulk replace' event.
+
+        This event is invoked for a sequence of values as they are incoming
+        to a bulk collection set operation, which can be
+        modified in place before the values are treated as ORM objects.
+        This is an "early hook" that runs before the bulk replace routine
+        attempts to reconcile which objects are already present in the
+        collection and which are being removed by the net replace operation.
+
+        It is typical that this method be combined with use of the
+        :meth:`.AttributeEvents.append` event.    When using both of these
+        events, note that a bulk replace operation will invoke
+        the :meth:`.AttributeEvents.append` event for all new items,
+        even after :meth:`.AttributeEvents.bulk_replace` has been invoked
+        for the collection as a whole.  In order to determine if an
+        :meth:`.AttributeEvents.append` event is part of a bulk replace,
+        use the symbol :attr:`~.attributes.OP_BULK_REPLACE` to test the
+        incoming initiator::
+
+            from sqlalchemy.orm.attributes import OP_BULK_REPLACE
+
+
+            @event.listens_for(SomeObject.collection, "bulk_replace")
+            def process_collection(target, values, initiator):
+                values[:] = [_make_value(value) for value in values]
+
+
+            @event.listens_for(SomeObject.collection, "append", retval=True)
+            def process_collection(target, value, initiator):
+                # make sure bulk_replace didn't already do it
+                if initiator is None or initiator.op is not OP_BULK_REPLACE:
+                    return _make_value(value)
+                else:
+                    return value
+
+        .. versionadded:: 1.2
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+        :param value: a sequence (e.g. a list) of the values being set.  The
+          handler can modify this list in place.
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.
+        :param keys: When the event is established using the
+         :paramref:`.AttributeEvents.include_key` parameter set to
+         True, this will be the sequence of keys used in the operation,
+         typically only for a dictionary update.  The parameter is not passed
+         to the event at all if the the
+         :paramref:`.AttributeEvents.include_key`
+         was not used to set up the event; this is to allow backwards
+         compatibility with existing event handlers that don't include the
+         ``key`` parameter.
+
+         .. versionadded:: 2.0
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+
+        """
+
+    def remove(
+        self,
+        target: _O,
+        value: _T,
+        initiator: Event,
+        *,
+        key: EventConstants = NO_KEY,
+    ) -> None:
+        """Receive a collection remove event.
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+        :param value: the value being removed.
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.  May be modified
+          from its original value by backref handlers in order to control
+          chained event propagation.
+
+        :param key: When the event is established using the
+         :paramref:`.AttributeEvents.include_key` parameter set to
+         True, this will be the key used in the operation, such as
+         ``del collection[some_key_or_index]``.  The parameter is not passed
+         to the event at all if the the
+         :paramref:`.AttributeEvents.include_key`
+         was not used to set up the event; this is to allow backwards
+         compatibility with existing event handlers that don't include the
+         ``key`` parameter.
+
+         .. versionadded:: 2.0
+
+        :return: No return value is defined for this event.
+
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+        """
+
+    def set(
+        self, target: _O, value: _T, oldvalue: _T, initiator: Event
+    ) -> None:
+        """Receive a scalar set event.
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+        :param value: the value being set.  If this listener
+          is registered with ``retval=True``, the listener
+          function must return this value, or a new value which
+          replaces it.
+        :param oldvalue: the previous value being replaced.  This
+          may also be the symbol ``NEVER_SET`` or ``NO_VALUE``.
+          If the listener is registered with ``active_history=True``,
+          the previous value of the attribute will be loaded from
+          the database if the existing value is currently unloaded
+          or expired.
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.  May be modified
+          from its original value by backref handlers in order to control
+          chained event propagation.
+
+        :return: if the event was registered with ``retval=True``,
+         the given value, or a new effective value, should be returned.
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+        """
+
+    def init_scalar(
+        self, target: _O, value: _T, dict_: Dict[Any, Any]
+    ) -> None:
+        r"""Receive a scalar "init" event.
+
+        This event is invoked when an uninitialized, unpersisted scalar
+        attribute is accessed, e.g. read::
+
+
+            x = my_object.some_attribute
+
+        The ORM's default behavior when this occurs for an un-initialized
+        attribute is to return the value ``None``; note this differs from
+        Python's usual behavior of raising ``AttributeError``.    The
+        event here can be used to customize what value is actually returned,
+        with the assumption that the event listener would be mirroring
+        a default generator that is configured on the Core
+        :class:`_schema.Column`
+        object as well.
+
+        Since a default generator on a :class:`_schema.Column`
+        might also produce
+        a changing value such as a timestamp, the
+        :meth:`.AttributeEvents.init_scalar`
+        event handler can also be used to **set** the newly returned value, so
+        that a Core-level default generation function effectively fires off
+        only once, but at the moment the attribute is accessed on the
+        non-persisted object.   Normally, no change to the object's state
+        is made when an uninitialized attribute is accessed (much older
+        SQLAlchemy versions did in fact change the object's state).
+
+        If a default generator on a column returned a particular constant,
+        a handler might be used as follows::
+
+            SOME_CONSTANT = 3.1415926
+
+
+            class MyClass(Base):
+                # ...
+
+                some_attribute = Column(Numeric, default=SOME_CONSTANT)
+
+
+            @event.listens_for(
+                MyClass.some_attribute, "init_scalar", retval=True, propagate=True
+            )
+            def _init_some_attribute(target, dict_, value):
+                dict_["some_attribute"] = SOME_CONSTANT
+                return SOME_CONSTANT
+
+        Above, we initialize the attribute ``MyClass.some_attribute`` to the
+        value of ``SOME_CONSTANT``.   The above code includes the following
+        features:
+
+        * By setting the value ``SOME_CONSTANT`` in the given ``dict_``,
+          we indicate that this value is to be persisted to the database.
+          This supersedes the use of ``SOME_CONSTANT`` in the default generator
+          for the :class:`_schema.Column`.  The ``active_column_defaults.py``
+          example given at :ref:`examples_instrumentation` illustrates using
+          the same approach for a changing default, e.g. a timestamp
+          generator.    In this particular example, it is not strictly
+          necessary to do this since ``SOME_CONSTANT`` would be part of the
+          INSERT statement in either case.
+
+        * By establishing the ``retval=True`` flag, the value we return
+          from the function will be returned by the attribute getter.
+          Without this flag, the event is assumed to be a passive observer
+          and the return value of our function is ignored.
+
+        * The ``propagate=True`` flag is significant if the mapped class
+          includes inheriting subclasses, which would also make use of this
+          event listener.  Without this flag, an inheriting subclass will
+          not use our event handler.
+
+        In the above example, the attribute set event
+        :meth:`.AttributeEvents.set` as well as the related validation feature
+        provided by :obj:`_orm.validates` is **not** invoked when we apply our
+        value to the given ``dict_``.  To have these events to invoke in
+        response to our newly generated value, apply the value to the given
+        object as a normal attribute set operation::
+
+            SOME_CONSTANT = 3.1415926
+
+
+            @event.listens_for(
+                MyClass.some_attribute, "init_scalar", retval=True, propagate=True
+            )
+            def _init_some_attribute(target, dict_, value):
+                # will also fire off attribute set events
+                target.some_attribute = SOME_CONSTANT
+                return SOME_CONSTANT
+
+        When multiple listeners are set up, the generation of the value
+        is "chained" from one listener to the next by passing the value
+        returned by the previous listener that specifies ``retval=True``
+        as the ``value`` argument of the next listener.
+
+        :param target: the object instance receiving the event.
+         If the listener is registered with ``raw=True``, this will
+         be the :class:`.InstanceState` object.
+        :param value: the value that is to be returned before this event
+         listener were invoked.  This value begins as the value ``None``,
+         however will be the return value of the previous event handler
+         function if multiple listeners are present.
+        :param dict\_: the attribute dictionary of this mapped object.
+         This is normally the ``__dict__`` of the object, but in all cases
+         represents the destination that the attribute system uses to get
+         at the actual value of this attribute.  Placing the value in this
+         dictionary has the effect that the value will be used in the
+         INSERT statement generated by the unit of work.
+
+
+        .. seealso::
+
+            :meth:`.AttributeEvents.init_collection` - collection version
+            of this event
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+            :ref:`examples_instrumentation` - see the
+            ``active_column_defaults.py`` example.
+
+        """  # noqa: E501
+
+    def init_collection(
+        self,
+        target: _O,
+        collection: Type[Collection[Any]],
+        collection_adapter: CollectionAdapter,
+    ) -> None:
+        """Receive a 'collection init' event.
+
+        This event is triggered for a collection-based attribute, when
+        the initial "empty collection" is first generated for a blank
+        attribute, as well as for when the collection is replaced with
+        a new one, such as via a set event.
+
+        E.g., given that ``User.addresses`` is a relationship-based
+        collection, the event is triggered here::
+
+            u1 = User()
+            u1.addresses.append(a1)  #  <- new collection
+
+        and also during replace operations::
+
+            u1.addresses = [a2, a3]  #  <- new collection
+
+        :param target: the object instance receiving the event.
+         If the listener is registered with ``raw=True``, this will
+         be the :class:`.InstanceState` object.
+        :param collection: the new collection.  This will always be generated
+         from what was specified as
+         :paramref:`_orm.relationship.collection_class`, and will always
+         be empty.
+        :param collection_adapter: the :class:`.CollectionAdapter` that will
+         mediate internal access to the collection.
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+            :meth:`.AttributeEvents.init_scalar` - "scalar" version of this
+            event.
+
+        """
+
+    def dispose_collection(
+        self,
+        target: _O,
+        collection: Collection[Any],
+        collection_adapter: CollectionAdapter,
+    ) -> None:
+        """Receive a 'collection dispose' event.
+
+        This event is triggered for a collection-based attribute when
+        a collection is replaced, that is::
+
+            u1.addresses.append(a1)
+
+            u1.addresses = [a2, a3]  # <- old collection is disposed
+
+        The old collection received will contain its previous contents.
+
+        .. versionchanged:: 1.2 The collection passed to
+           :meth:`.AttributeEvents.dispose_collection` will now have its
+           contents before the dispose intact; previously, the collection
+           would be empty.
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+        """
+
+    def modified(self, target: _O, initiator: Event) -> None:
+        """Receive a 'modified' event.
+
+        This event is triggered when the :func:`.attributes.flag_modified`
+        function is used to trigger a modify event on an attribute without
+        any specific value being set.
+
+        .. versionadded:: 1.2
+
+        :param target: the object instance receiving the event.
+          If the listener is registered with ``raw=True``, this will
+          be the :class:`.InstanceState` object.
+
+        :param initiator: An instance of :class:`.attributes.Event`
+          representing the initiation of the event.
+
+        .. seealso::
+
+            :class:`.AttributeEvents` - background on listener options such
+            as propagation to subclasses.
+
+        """
+
+
+class QueryEvents(event.Events[Query[Any]]):
+    """Represent events within the construction of a :class:`_query.Query`
+    object.
+
+    .. legacy:: The :class:`_orm.QueryEvents` event methods are legacy
+        as of SQLAlchemy 2.0, and only apply to direct use of the
+        :class:`_orm.Query` object. They are not used for :term:`2.0 style`
+        statements. For events to intercept and modify 2.0 style ORM use,
+        use the :meth:`_orm.SessionEvents.do_orm_execute` hook.
+
+
+    The :class:`_orm.QueryEvents` hooks are now superseded by the
+    :meth:`_orm.SessionEvents.do_orm_execute` event hook.
+
+    """
+
+    _target_class_doc = "SomeQuery"
+    _dispatch_target = Query
+
+    def before_compile(self, query: Query[Any]) -> None:
+        """Receive the :class:`_query.Query`
+        object before it is composed into a
+        core :class:`_expression.Select` object.
+
+        .. deprecated:: 1.4  The :meth:`_orm.QueryEvents.before_compile` event
+           is superseded by the much more capable
+           :meth:`_orm.SessionEvents.do_orm_execute` hook.   In version 1.4,
+           the :meth:`_orm.QueryEvents.before_compile` event is **no longer
+           used** for ORM-level attribute loads, such as loads of deferred
+           or expired attributes as well as relationship loaders.   See the
+           new examples in :ref:`examples_session_orm_events` which
+           illustrate new ways of intercepting and modifying ORM queries
+           for the most common purpose of adding arbitrary filter criteria.
+
+
+        This event is intended to allow changes to the query given::
+
+            @event.listens_for(Query, "before_compile", retval=True)
+            def no_deleted(query):
+                for desc in query.column_descriptions:
+                    if desc["type"] is User:
+                        entity = desc["entity"]
+                        query = query.filter(entity.deleted == False)
+                return query
+
+        The event should normally be listened with the ``retval=True``
+        parameter set, so that the modified query may be returned.
+
+        The :meth:`.QueryEvents.before_compile` event by default
+        will disallow "baked" queries from caching a query, if the event
+        hook returns a new :class:`_query.Query` object.
+        This affects both direct
+        use of the baked query extension as well as its operation within
+        lazy loaders and eager loaders for relationships.  In order to
+        re-establish the query being cached, apply the event adding the
+        ``bake_ok`` flag::
+
+            @event.listens_for(Query, "before_compile", retval=True, bake_ok=True)
+            def my_event(query):
+                for desc in query.column_descriptions:
+                    if desc["type"] is User:
+                        entity = desc["entity"]
+                        query = query.filter(entity.deleted == False)
+                return query
+
+        When ``bake_ok`` is set to True, the event hook will only be invoked
+        once, and not called for subsequent invocations of a particular query
+        that is being cached.
+
+        .. versionadded:: 1.3.11  - added the "bake_ok" flag to the
+           :meth:`.QueryEvents.before_compile` event and disallowed caching via
+           the "baked" extension from occurring for event handlers that
+           return  a new :class:`_query.Query` object if this flag is not set.
+
+        .. seealso::
+
+            :meth:`.QueryEvents.before_compile_update`
+
+            :meth:`.QueryEvents.before_compile_delete`
+
+            :ref:`baked_with_before_compile`
+
+        """  # noqa: E501
+
+    def before_compile_update(
+        self, query: Query[Any], update_context: BulkUpdate
+    ) -> None:
+        """Allow modifications to the :class:`_query.Query` object within
+        :meth:`_query.Query.update`.
+
+        .. deprecated:: 1.4  The :meth:`_orm.QueryEvents.before_compile_update`
+           event is superseded by the much more capable
+           :meth:`_orm.SessionEvents.do_orm_execute` hook.
+
+        Like the :meth:`.QueryEvents.before_compile` event, if the event
+        is to be used to alter the :class:`_query.Query` object, it should
+        be configured with ``retval=True``, and the modified
+        :class:`_query.Query` object returned, as in ::
+
+            @event.listens_for(Query, "before_compile_update", retval=True)
+            def no_deleted(query, update_context):
+                for desc in query.column_descriptions:
+                    if desc["type"] is User:
+                        entity = desc["entity"]
+                        query = query.filter(entity.deleted == False)
+
+                        update_context.values["timestamp"] = datetime.datetime.now(
+                            datetime.UTC
+                        )
+                return query
+
+        The ``.values`` dictionary of the "update context" object can also
+        be modified in place as illustrated above.
+
+        :param query: a :class:`_query.Query` instance; this is also
+         the ``.query`` attribute of the given "update context"
+         object.
+
+        :param update_context: an "update context" object which is
+         the same kind of object as described in
+         :paramref:`.QueryEvents.after_bulk_update.update_context`.
+         The object has a ``.values`` attribute in an UPDATE context which is
+         the dictionary of parameters passed to :meth:`_query.Query.update`.
+         This
+         dictionary can be modified to alter the VALUES clause of the
+         resulting UPDATE statement.
+
+        .. versionadded:: 1.2.17
+
+        .. seealso::
+
+            :meth:`.QueryEvents.before_compile`
+
+            :meth:`.QueryEvents.before_compile_delete`
+
+
+        """  # noqa: E501
+
+    def before_compile_delete(
+        self, query: Query[Any], delete_context: BulkDelete
+    ) -> None:
+        """Allow modifications to the :class:`_query.Query` object within
+        :meth:`_query.Query.delete`.
+
+        .. deprecated:: 1.4  The :meth:`_orm.QueryEvents.before_compile_delete`
+           event is superseded by the much more capable
+           :meth:`_orm.SessionEvents.do_orm_execute` hook.
+
+        Like the :meth:`.QueryEvents.before_compile` event, this event
+        should be configured with ``retval=True``, and the modified
+        :class:`_query.Query` object returned, as in ::
+
+            @event.listens_for(Query, "before_compile_delete", retval=True)
+            def no_deleted(query, delete_context):
+                for desc in query.column_descriptions:
+                    if desc["type"] is User:
+                        entity = desc["entity"]
+                        query = query.filter(entity.deleted == False)
+                return query
+
+        :param query: a :class:`_query.Query` instance; this is also
+         the ``.query`` attribute of the given "delete context"
+         object.
+
+        :param delete_context: a "delete context" object which is
+         the same kind of object as described in
+         :paramref:`.QueryEvents.after_bulk_delete.delete_context`.
+
+        .. versionadded:: 1.2.17
+
+        .. seealso::
+
+            :meth:`.QueryEvents.before_compile`
+
+            :meth:`.QueryEvents.before_compile_update`
+
+
+        """
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: _EventKey[_ET],
+        retval: bool = False,
+        bake_ok: bool = False,
+        **kw: Any,
+    ) -> None:
+        fn = event_key._listen_fn
+
+        if not retval:
+
+            def wrap(*arg: Any, **kw: Any) -> Any:
+                if not retval:
+                    query = arg[0]
+                    fn(*arg, **kw)
+                    return query
+                else:
+                    return fn(*arg, **kw)
+
+            event_key = event_key.with_wrapper(wrap)
+        else:
+            # don't assume we can apply an attribute to the callable
+            def wrap(*arg: Any, **kw: Any) -> Any:
+                return fn(*arg, **kw)
+
+            event_key = event_key.with_wrapper(wrap)
+
+        wrap._bake_ok = bake_ok  # type: ignore [attr-defined]
+
+        event_key.base_listen(**kw)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/exc.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/exc.py
new file mode 100644
index 00000000..0494edf9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/exc.py
@@ -0,0 +1,228 @@
+# orm/exc.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""SQLAlchemy ORM exceptions."""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from .util import _mapper_property_as_plain_name
+from .. import exc as sa_exc
+from .. import util
+from ..exc import MultipleResultsFound  # noqa
+from ..exc import NoResultFound  # noqa
+
+if TYPE_CHECKING:
+    from .interfaces import LoaderStrategy
+    from .interfaces import MapperProperty
+    from .state import InstanceState
+
+_T = TypeVar("_T", bound=Any)
+
+NO_STATE = (AttributeError, KeyError)
+"""Exception types that may be raised by instrumentation implementations."""
+
+
+class StaleDataError(sa_exc.SQLAlchemyError):
+    """An operation encountered database state that is unaccounted for.
+
+    Conditions which cause this to happen include:
+
+    * A flush may have attempted to update or delete rows
+      and an unexpected number of rows were matched during
+      the UPDATE or DELETE statement.   Note that when
+      version_id_col is used, rows in UPDATE or DELETE statements
+      are also matched against the current known version
+      identifier.
+
+    * A mapped object with version_id_col was refreshed,
+      and the version number coming back from the database does
+      not match that of the object itself.
+
+    * A object is detached from its parent object, however
+      the object was previously attached to a different parent
+      identity which was garbage collected, and a decision
+      cannot be made if the new parent was really the most
+      recent "parent".
+
+    """
+
+
+ConcurrentModificationError = StaleDataError
+
+
+class FlushError(sa_exc.SQLAlchemyError):
+    """A invalid condition was detected during flush()."""
+
+
+class UnmappedError(sa_exc.InvalidRequestError):
+    """Base for exceptions that involve expected mappings not present."""
+
+
+class ObjectDereferencedError(sa_exc.SQLAlchemyError):
+    """An operation cannot complete due to an object being garbage
+    collected.
+
+    """
+
+
+class DetachedInstanceError(sa_exc.SQLAlchemyError):
+    """An attempt to access unloaded attributes on a
+    mapped instance that is detached."""
+
+    code = "bhk3"
+
+
+class UnmappedInstanceError(UnmappedError):
+    """An mapping operation was requested for an unknown instance."""
+
+    @util.preload_module("sqlalchemy.orm.base")
+    def __init__(self, obj: object, msg: Optional[str] = None):
+        base = util.preloaded.orm_base
+
+        if not msg:
+            try:
+                base.class_mapper(type(obj))
+                name = _safe_cls_name(type(obj))
+                msg = (
+                    "Class %r is mapped, but this instance lacks "
+                    "instrumentation.  This occurs when the instance "
+                    "is created before sqlalchemy.orm.mapper(%s) "
+                    "was called." % (name, name)
+                )
+            except UnmappedClassError:
+                msg = f"Class '{_safe_cls_name(type(obj))}' is not mapped"
+                if isinstance(obj, type):
+                    msg += (
+                        "; was a class (%s) supplied where an instance was "
+                        "required?" % _safe_cls_name(obj)
+                    )
+        UnmappedError.__init__(self, msg)
+
+    def __reduce__(self) -> Any:
+        return self.__class__, (None, self.args[0])
+
+
+class UnmappedClassError(UnmappedError):
+    """An mapping operation was requested for an unknown class."""
+
+    def __init__(self, cls: Type[_T], msg: Optional[str] = None):
+        if not msg:
+            msg = _default_unmapped(cls)
+        UnmappedError.__init__(self, msg)
+
+    def __reduce__(self) -> Any:
+        return self.__class__, (None, self.args[0])
+
+
+class ObjectDeletedError(sa_exc.InvalidRequestError):
+    """A refresh operation failed to retrieve the database
+    row corresponding to an object's known primary key identity.
+
+    A refresh operation proceeds when an expired attribute is
+    accessed on an object, or when :meth:`_query.Query.get` is
+    used to retrieve an object which is, upon retrieval, detected
+    as expired.   A SELECT is emitted for the target row
+    based on primary key; if no row is returned, this
+    exception is raised.
+
+    The true meaning of this exception is simply that
+    no row exists for the primary key identifier associated
+    with a persistent object.   The row may have been
+    deleted, or in some cases the primary key updated
+    to a new value, outside of the ORM's management of the target
+    object.
+
+    """
+
+    @util.preload_module("sqlalchemy.orm.base")
+    def __init__(self, state: InstanceState[Any], msg: Optional[str] = None):
+        base = util.preloaded.orm_base
+
+        if not msg:
+            msg = (
+                "Instance '%s' has been deleted, or its "
+                "row is otherwise not present." % base.state_str(state)
+            )
+
+        sa_exc.InvalidRequestError.__init__(self, msg)
+
+    def __reduce__(self) -> Any:
+        return self.__class__, (None, self.args[0])
+
+
+class UnmappedColumnError(sa_exc.InvalidRequestError):
+    """Mapping operation was requested on an unknown column."""
+
+
+class LoaderStrategyException(sa_exc.InvalidRequestError):
+    """A loader strategy for an attribute does not exist."""
+
+    def __init__(
+        self,
+        applied_to_property_type: Type[Any],
+        requesting_property: MapperProperty[Any],
+        applies_to: Optional[Type[MapperProperty[Any]]],
+        actual_strategy_type: Optional[Type[LoaderStrategy]],
+        strategy_key: Tuple[Any, ...],
+    ):
+        if actual_strategy_type is None:
+            sa_exc.InvalidRequestError.__init__(
+                self,
+                "Can't find strategy %s for %s"
+                % (strategy_key, requesting_property),
+            )
+        else:
+            assert applies_to is not None
+            sa_exc.InvalidRequestError.__init__(
+                self,
+                'Can\'t apply "%s" strategy to property "%s", '
+                'which is a "%s"; this loader strategy is intended '
+                'to be used with a "%s".'
+                % (
+                    util.clsname_as_plain_name(actual_strategy_type),
+                    requesting_property,
+                    _mapper_property_as_plain_name(applied_to_property_type),
+                    _mapper_property_as_plain_name(applies_to),
+                ),
+            )
+
+
+def _safe_cls_name(cls: Type[Any]) -> str:
+    cls_name: Optional[str]
+    try:
+        cls_name = ".".join((cls.__module__, cls.__name__))
+    except AttributeError:
+        cls_name = getattr(cls, "__name__", None)
+        if cls_name is None:
+            cls_name = repr(cls)
+    return cls_name
+
+
+@util.preload_module("sqlalchemy.orm.base")
+def _default_unmapped(cls: Type[Any]) -> Optional[str]:
+    base = util.preloaded.orm_base
+
+    try:
+        mappers = base.manager_of_class(cls).mappers  # type: ignore
+    except (
+        UnmappedClassError,
+        TypeError,
+    ) + NO_STATE:
+        mappers = {}
+    name = _safe_cls_name(cls)
+
+    if not mappers:
+        return f"Class '{name}' is not mapped"
+    else:
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/identity.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/identity.py
new file mode 100644
index 00000000..1808b2d5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/identity.py
@@ -0,0 +1,302 @@
+# orm/identity.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+import weakref
+
+from . import util as orm_util
+from .. import exc as sa_exc
+
+if TYPE_CHECKING:
+    from ._typing import _IdentityKeyType
+    from .state import InstanceState
+
+
+_T = TypeVar("_T", bound=Any)
+
+_O = TypeVar("_O", bound=object)
+
+
+class IdentityMap:
+    _wr: weakref.ref[IdentityMap]
+
+    _dict: Dict[_IdentityKeyType[Any], Any]
+    _modified: Set[InstanceState[Any]]
+
+    def __init__(self) -> None:
+        self._dict = {}
+        self._modified = set()
+        self._wr = weakref.ref(self)
+
+    def _kill(self) -> None:
+        self._add_unpresent = _killed  # type: ignore
+
+    def all_states(self) -> List[InstanceState[Any]]:
+        raise NotImplementedError()
+
+    def contains_state(self, state: InstanceState[Any]) -> bool:
+        raise NotImplementedError()
+
+    def __contains__(self, key: _IdentityKeyType[Any]) -> bool:
+        raise NotImplementedError()
+
+    def safe_discard(self, state: InstanceState[Any]) -> None:
+        raise NotImplementedError()
+
+    def __getitem__(self, key: _IdentityKeyType[_O]) -> _O:
+        raise NotImplementedError()
+
+    def get(
+        self, key: _IdentityKeyType[_O], default: Optional[_O] = None
+    ) -> Optional[_O]:
+        raise NotImplementedError()
+
+    def fast_get_state(
+        self, key: _IdentityKeyType[_O]
+    ) -> Optional[InstanceState[_O]]:
+        raise NotImplementedError()
+
+    def keys(self) -> Iterable[_IdentityKeyType[Any]]:
+        return self._dict.keys()
+
+    def values(self) -> Iterable[object]:
+        raise NotImplementedError()
+
+    def replace(self, state: InstanceState[_O]) -> Optional[InstanceState[_O]]:
+        raise NotImplementedError()
+
+    def add(self, state: InstanceState[Any]) -> bool:
+        raise NotImplementedError()
+
+    def _fast_discard(self, state: InstanceState[Any]) -> None:
+        raise NotImplementedError()
+
+    def _add_unpresent(
+        self, state: InstanceState[Any], key: _IdentityKeyType[Any]
+    ) -> None:
+        """optional inlined form of add() which can assume item isn't present
+        in the map"""
+        self.add(state)
+
+    def _manage_incoming_state(self, state: InstanceState[Any]) -> None:
+        state._instance_dict = self._wr
+
+        if state.modified:
+            self._modified.add(state)
+
+    def _manage_removed_state(self, state: InstanceState[Any]) -> None:
+        del state._instance_dict
+        if state.modified:
+            self._modified.discard(state)
+
+    def _dirty_states(self) -> Set[InstanceState[Any]]:
+        return self._modified
+
+    def check_modified(self) -> bool:
+        """return True if any InstanceStates present have been marked
+        as 'modified'.
+
+        """
+        return bool(self._modified)
+
+    def has_key(self, key: _IdentityKeyType[Any]) -> bool:
+        return key in self
+
+    def __len__(self) -> int:
+        return len(self._dict)
+
+
+class WeakInstanceDict(IdentityMap):
+    _dict: Dict[_IdentityKeyType[Any], InstanceState[Any]]
+
+    def __getitem__(self, key: _IdentityKeyType[_O]) -> _O:
+        state = cast("InstanceState[_O]", self._dict[key])
+        o = state.obj()
+        if o is None:
+            raise KeyError(key)
+        return o
+
+    def __contains__(self, key: _IdentityKeyType[Any]) -> bool:
+        try:
+            if key in self._dict:
+                state = self._dict[key]
+                o = state.obj()
+            else:
+                return False
+        except KeyError:
+            return False
+        else:
+            return o is not None
+
+    def contains_state(self, state: InstanceState[Any]) -> bool:
+        if state.key in self._dict:
+            if TYPE_CHECKING:
+                assert state.key is not None
+            try:
+                return self._dict[state.key] is state
+            except KeyError:
+                return False
+        else:
+            return False
+
+    def replace(
+        self, state: InstanceState[Any]
+    ) -> Optional[InstanceState[Any]]:
+        assert state.key is not None
+        if state.key in self._dict:
+            try:
+                existing = existing_non_none = self._dict[state.key]
+            except KeyError:
+                # catch gc removed the key after we just checked for it
+                existing = None
+            else:
+                if existing_non_none is not state:
+                    self._manage_removed_state(existing_non_none)
+                else:
+                    return None
+        else:
+            existing = None
+
+        self._dict[state.key] = state
+        self._manage_incoming_state(state)
+        return existing
+
+    def add(self, state: InstanceState[Any]) -> bool:
+        key = state.key
+        assert key is not None
+        # inline of self.__contains__
+        if key in self._dict:
+            try:
+                existing_state = self._dict[key]
+            except KeyError:
+                # catch gc removed the key after we just checked for it
+                pass
+            else:
+                if existing_state is not state:
+                    o = existing_state.obj()
+                    if o is not None:
+                        raise sa_exc.InvalidRequestError(
+                            "Can't attach instance "
+                            "%s; another instance with key %s is already "
+                            "present in this session."
+                            % (orm_util.state_str(state), state.key)
+                        )
+                else:
+                    return False
+        self._dict[key] = state
+        self._manage_incoming_state(state)
+        return True
+
+    def _add_unpresent(
+        self, state: InstanceState[Any], key: _IdentityKeyType[Any]
+    ) -> None:
+        # inlined form of add() called by loading.py
+        self._dict[key] = state
+        state._instance_dict = self._wr
+
+    def fast_get_state(
+        self, key: _IdentityKeyType[_O]
+    ) -> Optional[InstanceState[_O]]:
+        return self._dict.get(key)
+
+    def get(
+        self, key: _IdentityKeyType[_O], default: Optional[_O] = None
+    ) -> Optional[_O]:
+        if key not in self._dict:
+            return default
+        try:
+            state = cast("InstanceState[_O]", self._dict[key])
+        except KeyError:
+            # catch gc removed the key after we just checked for it
+            return default
+        else:
+            o = state.obj()
+            if o is None:
+                return default
+            return o
+
+    def items(self) -> List[Tuple[_IdentityKeyType[Any], InstanceState[Any]]]:
+        values = self.all_states()
+        result = []
+        for state in values:
+            value = state.obj()
+            key = state.key
+            assert key is not None
+            if value is not None:
+                result.append((key, value))
+        return result
+
+    def values(self) -> List[object]:
+        values = self.all_states()
+        result = []
+        for state in values:
+            value = state.obj()
+            if value is not None:
+                result.append(value)
+
+        return result
+
+    def __iter__(self) -> Iterator[_IdentityKeyType[Any]]:
+        return iter(self.keys())
+
+    def all_states(self) -> List[InstanceState[Any]]:
+        return list(self._dict.values())
+
+    def _fast_discard(self, state: InstanceState[Any]) -> None:
+        # used by InstanceState for state being
+        # GC'ed, inlines _managed_removed_state
+        key = state.key
+        assert key is not None
+        try:
+            st = self._dict[key]
+        except KeyError:
+            # catch gc removed the key after we just checked for it
+            pass
+        else:
+            if st is state:
+                self._dict.pop(key, None)
+
+    def discard(self, state: InstanceState[Any]) -> None:
+        self.safe_discard(state)
+
+    def safe_discard(self, state: InstanceState[Any]) -> None:
+        key = state.key
+        if key in self._dict:
+            assert key is not None
+            try:
+                st = self._dict[key]
+            except KeyError:
+                # catch gc removed the key after we just checked for it
+                pass
+            else:
+                if st is state:
+                    self._dict.pop(key, None)
+                    self._manage_removed_state(state)
+
+
+def _killed(state: InstanceState[Any], key: _IdentityKeyType[Any]) -> NoReturn:
+    # external function to avoid creating cycles when assigned to
+    # the IdentityMap
+    raise sa_exc.InvalidRequestError(
+        "Object %s cannot be converted to 'persistent' state, as this "
+        "identity map is no longer valid.  Has the owning Session "
+        "been closed?" % orm_util.state_str(state),
+        code="lkrp",
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/instrumentation.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/instrumentation.py
new file mode 100644
index 00000000..f87023f1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/instrumentation.py
@@ -0,0 +1,754 @@
+# orm/instrumentation.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Defines SQLAlchemy's system of class instrumentation.
+
+This module is usually not directly visible to user applications, but
+defines a large part of the ORM's interactivity.
+
+instrumentation.py deals with registration of end-user classes
+for state tracking.   It interacts closely with state.py
+and attributes.py which establish per-instance and per-class-attribute
+instrumentation, respectively.
+
+The class instrumentation system can be customized on a per-class
+or global basis using the :mod:`sqlalchemy.ext.instrumentation`
+module, which provides the means to build and specify
+alternate instrumentation forms.
+
+.. versionchanged: 0.8
+   The instrumentation extension system was moved out of the
+   ORM and into the external :mod:`sqlalchemy.ext.instrumentation`
+   package.  When that package is imported, it installs
+   itself within sqlalchemy.orm so that its more comprehensive
+   resolution mechanics take effect.
+
+"""
+
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import base
+from . import collections
+from . import exc
+from . import interfaces
+from . import state
+from ._typing import _O
+from .attributes import _is_collection_attribute_impl
+from .. import util
+from ..event import EventTarget
+from ..util import HasMemoized
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if TYPE_CHECKING:
+    from ._typing import _RegistryType
+    from .attributes import AttributeImpl
+    from .attributes import QueryableAttribute
+    from .collections import _AdaptedCollectionProtocol
+    from .collections import _CollectionFactoryType
+    from .decl_base import _MapperConfig
+    from .events import InstanceEvents
+    from .mapper import Mapper
+    from .state import InstanceState
+    from ..event import dispatcher
+
+_T = TypeVar("_T", bound=Any)
+DEL_ATTR = util.symbol("DEL_ATTR")
+
+
+class _ExpiredAttributeLoaderProto(Protocol):
+    def __call__(
+        self,
+        state: state.InstanceState[Any],
+        toload: Set[str],
+        passive: base.PassiveFlag,
+    ) -> None: ...
+
+
+class _ManagerFactory(Protocol):
+    def __call__(self, class_: Type[_O]) -> ClassManager[_O]: ...
+
+
+class ClassManager(
+    HasMemoized,
+    Dict[str, "QueryableAttribute[Any]"],
+    Generic[_O],
+    EventTarget,
+):
+    """Tracks state information at the class level."""
+
+    dispatch: dispatcher[ClassManager[_O]]
+
+    MANAGER_ATTR = base.DEFAULT_MANAGER_ATTR
+    STATE_ATTR = base.DEFAULT_STATE_ATTR
+
+    _state_setter = staticmethod(util.attrsetter(STATE_ATTR))
+
+    expired_attribute_loader: _ExpiredAttributeLoaderProto
+    "previously known as deferred_scalar_loader"
+
+    init_method: Optional[Callable[..., None]]
+    original_init: Optional[Callable[..., None]] = None
+
+    factory: Optional[_ManagerFactory]
+
+    declarative_scan: Optional[weakref.ref[_MapperConfig]] = None
+
+    registry: _RegistryType
+
+    if not TYPE_CHECKING:
+        # starts as None during setup
+        registry = None
+
+    class_: Type[_O]
+
+    _bases: List[ClassManager[Any]]
+
+    @property
+    @util.deprecated(
+        "1.4",
+        message="The ClassManager.deferred_scalar_loader attribute is now "
+        "named expired_attribute_loader",
+    )
+    def deferred_scalar_loader(self):
+        return self.expired_attribute_loader
+
+    @deferred_scalar_loader.setter
+    @util.deprecated(
+        "1.4",
+        message="The ClassManager.deferred_scalar_loader attribute is now "
+        "named expired_attribute_loader",
+    )
+    def deferred_scalar_loader(self, obj):
+        self.expired_attribute_loader = obj
+
+    def __init__(self, class_):
+        self.class_ = class_
+        self.info = {}
+        self.new_init = None
+        self.local_attrs = {}
+        self.originals = {}
+        self._finalized = False
+        self.factory = None
+        self.init_method = None
+
+        self._bases = [
+            mgr
+            for mgr in cast(
+                "List[Optional[ClassManager[Any]]]",
+                [
+                    opt_manager_of_class(base)
+                    for base in self.class_.__bases__
+                    if isinstance(base, type)
+                ],
+            )
+            if mgr is not None
+        ]
+
+        for base_ in self._bases:
+            self.update(base_)
+
+        cast(
+            "InstanceEvents", self.dispatch._events
+        )._new_classmanager_instance(class_, self)
+
+        for basecls in class_.__mro__:
+            mgr = opt_manager_of_class(basecls)
+            if mgr is not None:
+                self.dispatch._update(mgr.dispatch)
+
+        self.manage()
+
+        if "__del__" in class_.__dict__:
+            util.warn(
+                "__del__() method on class %s will "
+                "cause unreachable cycles and memory leaks, "
+                "as SQLAlchemy instrumentation often creates "
+                "reference cycles.  Please remove this method." % class_
+            )
+
+    def _update_state(
+        self,
+        finalize: bool = False,
+        mapper: Optional[Mapper[_O]] = None,
+        registry: Optional[_RegistryType] = None,
+        declarative_scan: Optional[_MapperConfig] = None,
+        expired_attribute_loader: Optional[
+            _ExpiredAttributeLoaderProto
+        ] = None,
+        init_method: Optional[Callable[..., None]] = None,
+    ) -> None:
+        if mapper:
+            self.mapper = mapper  #
+        if registry:
+            registry._add_manager(self)
+        if declarative_scan:
+            self.declarative_scan = weakref.ref(declarative_scan)
+        if expired_attribute_loader:
+            self.expired_attribute_loader = expired_attribute_loader
+
+        if init_method:
+            assert not self._finalized, (
+                "class is already instrumented, "
+                "init_method %s can't be applied" % init_method
+            )
+            self.init_method = init_method
+
+        if not self._finalized:
+            self.original_init = (
+                self.init_method
+                if self.init_method is not None
+                and self.class_.__init__ is object.__init__
+                else self.class_.__init__
+            )
+
+        if finalize and not self._finalized:
+            self._finalize()
+
+    def _finalize(self) -> None:
+        if self._finalized:
+            return
+        self._finalized = True
+
+        self._instrument_init()
+
+        _instrumentation_factory.dispatch.class_instrument(self.class_)
+
+    def __hash__(self) -> int:  # type: ignore[override]
+        return id(self)
+
+    def __eq__(self, other: Any) -> bool:
+        return other is self
+
+    @property
+    def is_mapped(self) -> bool:
+        return "mapper" in self.__dict__
+
+    @HasMemoized.memoized_attribute
+    def _all_key_set(self):
+        return frozenset(self)
+
+    @HasMemoized.memoized_attribute
+    def _collection_impl_keys(self):
+        return frozenset(
+            [attr.key for attr in self.values() if attr.impl.collection]
+        )
+
+    @HasMemoized.memoized_attribute
+    def _scalar_loader_impls(self):
+        return frozenset(
+            [
+                attr.impl
+                for attr in self.values()
+                if attr.impl.accepts_scalar_loader
+            ]
+        )
+
+    @HasMemoized.memoized_attribute
+    def _loader_impls(self):
+        return frozenset([attr.impl for attr in self.values()])
+
+    @util.memoized_property
+    def mapper(self) -> Mapper[_O]:
+        # raises unless self.mapper has been assigned
+        raise exc.UnmappedClassError(self.class_)
+
+    def _all_sqla_attributes(self, exclude=None):
+        """return an iterator of all classbound attributes that are
+        implement :class:`.InspectionAttr`.
+
+        This includes :class:`.QueryableAttribute` as well as extension
+        types such as :class:`.hybrid_property` and
+        :class:`.AssociationProxy`.
+
+        """
+
+        found: Dict[str, Any] = {}
+
+        # constraints:
+        # 1. yield keys in cls.__dict__ order
+        # 2. if a subclass has the same key as a superclass, include that
+        #    key as part of the ordering of the superclass, because an
+        #    overridden key is usually installed by the mapper which is going
+        #    on a different ordering
+        # 3. don't use getattr() as this fires off descriptors
+
+        for supercls in self.class_.__mro__[0:-1]:
+            inherits = supercls.__mro__[1]
+            for key in supercls.__dict__:
+                found.setdefault(key, supercls)
+                if key in inherits.__dict__:
+                    continue
+                val = found[key].__dict__[key]
+                if (
+                    isinstance(val, interfaces.InspectionAttr)
+                    and val.is_attribute
+                ):
+                    yield key, val
+
+    def _get_class_attr_mro(self, key, default=None):
+        """return an attribute on the class without tripping it."""
+
+        for supercls in self.class_.__mro__:
+            if key in supercls.__dict__:
+                return supercls.__dict__[key]
+        else:
+            return default
+
+    def _attr_has_impl(self, key: str) -> bool:
+        """Return True if the given attribute is fully initialized.
+
+        i.e. has an impl.
+        """
+
+        return key in self and self[key].impl is not None
+
+    def _subclass_manager(self, cls: Type[_T]) -> ClassManager[_T]:
+        """Create a new ClassManager for a subclass of this ClassManager's
+        class.
+
+        This is called automatically when attributes are instrumented so that
+        the attributes can be propagated to subclasses against their own
+        class-local manager, without the need for mappers etc. to have already
+        pre-configured managers for the full class hierarchy.   Mappers
+        can post-configure the auto-generated ClassManager when needed.
+
+        """
+        return register_class(cls, finalize=False)
+
+    def _instrument_init(self):
+        self.new_init = _generate_init(self.class_, self, self.original_init)
+        self.install_member("__init__", self.new_init)
+
+    @util.memoized_property
+    def _state_constructor(self) -> Type[state.InstanceState[_O]]:
+        self.dispatch.first_init(self, self.class_)
+        return state.InstanceState
+
+    def manage(self):
+        """Mark this instance as the manager for its class."""
+
+        setattr(self.class_, self.MANAGER_ATTR, self)
+
+    @util.hybridmethod
+    def manager_getter(self):
+        return _default_manager_getter
+
+    @util.hybridmethod
+    def state_getter(self):
+        """Return a (instance) -> InstanceState callable.
+
+        "state getter" callables should raise either KeyError or
+        AttributeError if no InstanceState could be found for the
+        instance.
+        """
+
+        return _default_state_getter
+
+    @util.hybridmethod
+    def dict_getter(self):
+        return _default_dict_getter
+
+    def instrument_attribute(
+        self,
+        key: str,
+        inst: QueryableAttribute[Any],
+        propagated: bool = False,
+    ) -> None:
+        if propagated:
+            if key in self.local_attrs:
+                return  # don't override local attr with inherited attr
+        else:
+            self.local_attrs[key] = inst
+            self.install_descriptor(key, inst)
+        self._reset_memoizations()
+        self[key] = inst
+
+        for cls in self.class_.__subclasses__():
+            manager = self._subclass_manager(cls)
+            manager.instrument_attribute(key, inst, True)
+
+    def subclass_managers(self, recursive):
+        for cls in self.class_.__subclasses__():
+            mgr = opt_manager_of_class(cls)
+            if mgr is not None and mgr is not self:
+                yield mgr
+                if recursive:
+                    yield from mgr.subclass_managers(True)
+
+    def post_configure_attribute(self, key):
+        _instrumentation_factory.dispatch.attribute_instrument(
+            self.class_, key, self[key]
+        )
+
+    def uninstrument_attribute(self, key, propagated=False):
+        if key not in self:
+            return
+        if propagated:
+            if key in self.local_attrs:
+                return  # don't get rid of local attr
+        else:
+            del self.local_attrs[key]
+            self.uninstall_descriptor(key)
+        self._reset_memoizations()
+        del self[key]
+        for cls in self.class_.__subclasses__():
+            manager = opt_manager_of_class(cls)
+            if manager:
+                manager.uninstrument_attribute(key, True)
+
+    def unregister(self) -> None:
+        """remove all instrumentation established by this ClassManager."""
+
+        for key in list(self.originals):
+            self.uninstall_member(key)
+
+        self.mapper = None
+        self.dispatch = None  # type: ignore
+        self.new_init = None
+        self.info.clear()
+
+        for key in list(self):
+            if key in self.local_attrs:
+                self.uninstrument_attribute(key)
+
+        if self.MANAGER_ATTR in self.class_.__dict__:
+            delattr(self.class_, self.MANAGER_ATTR)
+
+    def install_descriptor(
+        self, key: str, inst: QueryableAttribute[Any]
+    ) -> None:
+        if key in (self.STATE_ATTR, self.MANAGER_ATTR):
+            raise KeyError(
+                "%r: requested attribute name conflicts with "
+                "instrumentation attribute of the same name." % key
+            )
+        setattr(self.class_, key, inst)
+
+    def uninstall_descriptor(self, key: str) -> None:
+        delattr(self.class_, key)
+
+    def install_member(self, key: str, implementation: Any) -> None:
+        if key in (self.STATE_ATTR, self.MANAGER_ATTR):
+            raise KeyError(
+                "%r: requested attribute name conflicts with "
+                "instrumentation attribute of the same name." % key
+            )
+        self.originals.setdefault(key, self.class_.__dict__.get(key, DEL_ATTR))
+        setattr(self.class_, key, implementation)
+
+    def uninstall_member(self, key: str) -> None:
+        original = self.originals.pop(key, None)
+        if original is not DEL_ATTR:
+            setattr(self.class_, key, original)
+        else:
+            delattr(self.class_, key)
+
+    def instrument_collection_class(
+        self, key: str, collection_class: Type[Collection[Any]]
+    ) -> _CollectionFactoryType:
+        return collections.prepare_instrumentation(collection_class)
+
+    def initialize_collection(
+        self,
+        key: str,
+        state: InstanceState[_O],
+        factory: _CollectionFactoryType,
+    ) -> Tuple[collections.CollectionAdapter, _AdaptedCollectionProtocol]:
+        user_data = factory()
+        impl = self.get_impl(key)
+        assert _is_collection_attribute_impl(impl)
+        adapter = collections.CollectionAdapter(impl, state, user_data)
+        return adapter, user_data
+
+    def is_instrumented(self, key: str, search: bool = False) -> bool:
+        if search:
+            return key in self
+        else:
+            return key in self.local_attrs
+
+    def get_impl(self, key: str) -> AttributeImpl:
+        return self[key].impl
+
+    @property
+    def attributes(self) -> Iterable[Any]:
+        return iter(self.values())
+
+    # InstanceState management
+
+    def new_instance(self, state: Optional[InstanceState[_O]] = None) -> _O:
+        # here, we would prefer _O to be bound to "object"
+        # so that mypy sees that __new__ is present.   currently
+        # it's bound to Any as there were other problems not having
+        # it that way but these can be revisited
+        instance = self.class_.__new__(self.class_)
+        if state is None:
+            state = self._state_constructor(instance, self)
+        self._state_setter(instance, state)
+        return instance
+
+    def setup_instance(
+        self, instance: _O, state: Optional[InstanceState[_O]] = None
+    ) -> None:
+        if state is None:
+            state = self._state_constructor(instance, self)
+        self._state_setter(instance, state)
+
+    def teardown_instance(self, instance: _O) -> None:
+        delattr(instance, self.STATE_ATTR)
+
+    def _serialize(
+        self, state: InstanceState[_O], state_dict: Dict[str, Any]
+    ) -> _SerializeManager:
+        return _SerializeManager(state, state_dict)
+
+    def _new_state_if_none(
+        self, instance: _O
+    ) -> Union[Literal[False], InstanceState[_O]]:
+        """Install a default InstanceState if none is present.
+
+        A private convenience method used by the __init__ decorator.
+
+        """
+        if hasattr(instance, self.STATE_ATTR):
+            return False
+        elif self.class_ is not instance.__class__ and self.is_mapped:
+            # this will create a new ClassManager for the
+            # subclass, without a mapper.  This is likely a
+            # user error situation but allow the object
+            # to be constructed, so that it is usable
+            # in a non-ORM context at least.
+            return self._subclass_manager(
+                instance.__class__
+            )._new_state_if_none(instance)
+        else:
+            state = self._state_constructor(instance, self)
+            self._state_setter(instance, state)
+            return state
+
+    def has_state(self, instance: _O) -> bool:
+        return hasattr(instance, self.STATE_ATTR)
+
+    def has_parent(
+        self, state: InstanceState[_O], key: str, optimistic: bool = False
+    ) -> bool:
+        """TODO"""
+        return self.get_impl(key).hasparent(state, optimistic=optimistic)
+
+    def __bool__(self) -> bool:
+        """All ClassManagers are non-zero regardless of attribute state."""
+        return True
+
+    def __repr__(self) -> str:
+        return "<%s of %r at %x>" % (
+            self.__class__.__name__,
+            self.class_,
+            id(self),
+        )
+
+
+class _SerializeManager:
+    """Provide serialization of a :class:`.ClassManager`.
+
+    The :class:`.InstanceState` uses ``__init__()`` on serialize
+    and ``__call__()`` on deserialize.
+
+    """
+
+    def __init__(self, state: state.InstanceState[Any], d: Dict[str, Any]):
+        self.class_ = state.class_
+        manager = state.manager
+        manager.dispatch.pickle(state, d)
+
+    def __call__(self, state, inst, state_dict):
+        state.manager = manager = opt_manager_of_class(self.class_)
+        if manager is None:
+            raise exc.UnmappedInstanceError(
+                inst,
+                "Cannot deserialize object of type %r - "
+                "no mapper() has "
+                "been configured for this class within the current "
+                "Python process!" % self.class_,
+            )
+        elif manager.is_mapped and not manager.mapper.configured:
+            manager.mapper._check_configure()
+
+        # setup _sa_instance_state ahead of time so that
+        # unpickle events can access the object normally.
+        # see [ticket:2362]
+        if inst is not None:
+            manager.setup_instance(inst, state)
+        manager.dispatch.unpickle(state, state_dict)
+
+
+class InstrumentationFactory(EventTarget):
+    """Factory for new ClassManager instances."""
+
+    dispatch: dispatcher[InstrumentationFactory]
+
+    def create_manager_for_cls(self, class_: Type[_O]) -> ClassManager[_O]:
+        assert class_ is not None
+        assert opt_manager_of_class(class_) is None
+
+        # give a more complicated subclass
+        # a chance to do what it wants here
+        manager, factory = self._locate_extended_factory(class_)
+
+        if factory is None:
+            factory = ClassManager
+            manager = ClassManager(class_)
+        else:
+            assert manager is not None
+
+        self._check_conflicts(class_, factory)
+
+        manager.factory = factory
+
+        return manager
+
+    def _locate_extended_factory(
+        self, class_: Type[_O]
+    ) -> Tuple[Optional[ClassManager[_O]], Optional[_ManagerFactory]]:
+        """Overridden by a subclass to do an extended lookup."""
+        return None, None
+
+    def _check_conflicts(
+        self, class_: Type[_O], factory: Callable[[Type[_O]], ClassManager[_O]]
+    ) -> None:
+        """Overridden by a subclass to test for conflicting factories."""
+
+    def unregister(self, class_: Type[_O]) -> None:
+        manager = manager_of_class(class_)
+        manager.unregister()
+        self.dispatch.class_uninstrument(class_)
+
+
+# this attribute is replaced by sqlalchemy.ext.instrumentation
+# when imported.
+_instrumentation_factory = InstrumentationFactory()
+
+# these attributes are replaced by sqlalchemy.ext.instrumentation
+# when a non-standard InstrumentationManager class is first
+# used to instrument a class.
+instance_state = _default_state_getter = base.instance_state
+
+instance_dict = _default_dict_getter = base.instance_dict
+
+manager_of_class = _default_manager_getter = base.manager_of_class
+opt_manager_of_class = _default_opt_manager_getter = base.opt_manager_of_class
+
+
+def register_class(
+    class_: Type[_O],
+    finalize: bool = True,
+    mapper: Optional[Mapper[_O]] = None,
+    registry: Optional[_RegistryType] = None,
+    declarative_scan: Optional[_MapperConfig] = None,
+    expired_attribute_loader: Optional[_ExpiredAttributeLoaderProto] = None,
+    init_method: Optional[Callable[..., None]] = None,
+) -> ClassManager[_O]:
+    """Register class instrumentation.
+
+    Returns the existing or newly created class manager.
+
+    """
+
+    manager = opt_manager_of_class(class_)
+    if manager is None:
+        manager = _instrumentation_factory.create_manager_for_cls(class_)
+    manager._update_state(
+        mapper=mapper,
+        registry=registry,
+        declarative_scan=declarative_scan,
+        expired_attribute_loader=expired_attribute_loader,
+        init_method=init_method,
+        finalize=finalize,
+    )
+
+    return manager
+
+
+def unregister_class(class_):
+    """Unregister class instrumentation."""
+
+    _instrumentation_factory.unregister(class_)
+
+
+def is_instrumented(instance, key):
+    """Return True if the given attribute on the given instance is
+    instrumented by the attributes package.
+
+    This function may be used regardless of instrumentation
+    applied directly to the class, i.e. no descriptors are required.
+
+    """
+    return manager_of_class(instance.__class__).is_instrumented(
+        key, search=True
+    )
+
+
+def _generate_init(class_, class_manager, original_init):
+    """Build an __init__ decorator that triggers ClassManager events."""
+
+    # TODO: we should use the ClassManager's notion of the
+    # original '__init__' method, once ClassManager is fixed
+    # to always reference that.
+
+    if original_init is None:
+        original_init = class_.__init__
+
+    # Go through some effort here and don't change the user's __init__
+    # calling signature, including the unlikely case that it has
+    # a return value.
+    # FIXME: need to juggle local names to avoid constructor argument
+    # clashes.
+    func_body = """\
+def __init__(%(apply_pos)s):
+    new_state = class_manager._new_state_if_none(%(self_arg)s)
+    if new_state:
+        return new_state._initialize_instance(%(apply_kw)s)
+    else:
+        return original_init(%(apply_kw)s)
+"""
+    func_vars = util.format_argspec_init(original_init, grouped=False)
+    func_text = func_body % func_vars
+
+    func_defaults = getattr(original_init, "__defaults__", None)
+    func_kw_defaults = getattr(original_init, "__kwdefaults__", None)
+
+    env = locals().copy()
+    env["__name__"] = __name__
+    exec(func_text, env)
+    __init__ = env["__init__"]
+    __init__.__doc__ = original_init.__doc__
+    __init__._sa_original_init = original_init
+
+    if func_defaults:
+        __init__.__defaults__ = func_defaults
+    if func_kw_defaults:
+        __init__.__kwdefaults__ = func_kw_defaults
+
+    return __init__
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/interfaces.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/interfaces.py
new file mode 100644
index 00000000..b4462e54
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/interfaces.py
@@ -0,0 +1,1490 @@
+# orm/interfaces.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""
+
+Contains various base classes used throughout the ORM.
+
+Defines some key base classes prominent within the internals.
+
+This module and the classes within are mostly private, though some attributes
+are exposed when inspecting mappings.
+
+"""
+
+from __future__ import annotations
+
+import collections
+import dataclasses
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import Generic
+from typing import Iterator
+from typing import List
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import exc as orm_exc
+from . import path_registry
+from .base import _MappedAttribute as _MappedAttribute
+from .base import EXT_CONTINUE as EXT_CONTINUE  # noqa: F401
+from .base import EXT_SKIP as EXT_SKIP  # noqa: F401
+from .base import EXT_STOP as EXT_STOP  # noqa: F401
+from .base import InspectionAttr as InspectionAttr  # noqa: F401
+from .base import InspectionAttrInfo as InspectionAttrInfo
+from .base import MANYTOMANY as MANYTOMANY  # noqa: F401
+from .base import MANYTOONE as MANYTOONE  # noqa: F401
+from .base import NO_KEY as NO_KEY  # noqa: F401
+from .base import NO_VALUE as NO_VALUE  # noqa: F401
+from .base import NotExtension as NotExtension  # noqa: F401
+from .base import ONETOMANY as ONETOMANY  # noqa: F401
+from .base import RelationshipDirection as RelationshipDirection  # noqa: F401
+from .base import SQLORMOperations
+from .. import ColumnElement
+from .. import exc as sa_exc
+from .. import inspection
+from .. import util
+from ..sql import operators
+from ..sql import roles
+from ..sql import visitors
+from ..sql.base import _NoArg
+from ..sql.base import ExecutableOption
+from ..sql.cache_key import HasCacheKey
+from ..sql.operators import ColumnOperators
+from ..sql.schema import Column
+from ..sql.type_api import TypeEngine
+from ..util import warn_deprecated
+from ..util.typing import RODescriptorReference
+from ..util.typing import TypedDict
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import _InternalEntityType
+    from ._typing import _ORMAdapterProto
+    from .attributes import InstrumentedAttribute
+    from .base import Mapped
+    from .context import _MapperEntity
+    from .context import ORMCompileState
+    from .context import QueryContext
+    from .decl_api import RegistryType
+    from .decl_base import _ClassScanMapperConfig
+    from .loading import _PopulatorDict
+    from .mapper import Mapper
+    from .path_registry import AbstractEntityRegistry
+    from .query import Query
+    from .session import Session
+    from .state import InstanceState
+    from .strategy_options import _LoadElement
+    from .util import AliasedInsp
+    from .util import ORMAdapter
+    from ..engine.result import Result
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _ColumnsClauseArgument
+    from ..sql._typing import _DMLColumnArgument
+    from ..sql._typing import _InfoType
+    from ..sql.operators import OperatorType
+    from ..sql.visitors import _TraverseInternalsType
+    from ..util.typing import _AnnotationScanType
+
+_StrategyKey = Tuple[Any, ...]
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+_TLS = TypeVar("_TLS", bound="Type[LoaderStrategy]")
+
+
+class ORMStatementRole(roles.StatementRole):
+    __slots__ = ()
+    _role_name = (
+        "Executable SQL or text() construct, including ORM aware objects"
+    )
+
+
+class ORMColumnsClauseRole(
+    roles.ColumnsClauseRole, roles.TypedColumnsClauseRole[_T]
+):
+    __slots__ = ()
+    _role_name = "ORM mapped entity, aliased entity, or Column expression"
+
+
+class ORMEntityColumnsClauseRole(ORMColumnsClauseRole[_T]):
+    __slots__ = ()
+    _role_name = "ORM mapped or aliased entity"
+
+
+class ORMFromClauseRole(roles.StrictFromClauseRole):
+    __slots__ = ()
+    _role_name = "ORM mapped entity, aliased entity, or FROM expression"
+
+
+class ORMColumnDescription(TypedDict):
+    name: str
+    # TODO: add python_type and sql_type here; combining them
+    # into "type" is a bad idea
+    type: Union[Type[Any], TypeEngine[Any]]
+    aliased: bool
+    expr: _ColumnsClauseArgument[Any]
+    entity: Optional[_ColumnsClauseArgument[Any]]
+
+
+class _IntrospectsAnnotations:
+    __slots__ = ()
+
+    @classmethod
+    def _mapper_property_name(cls) -> str:
+        return cls.__name__
+
+    def found_in_pep593_annotated(self) -> Any:
+        """return a copy of this object to use in declarative when the
+        object is found inside of an Annotated object."""
+
+        raise NotImplementedError(
+            f"Use of the {self._mapper_property_name()!r} "
+            "construct inside of an Annotated object is not yet supported."
+        )
+
+    def declarative_scan(
+        self,
+        decl_scan: _ClassScanMapperConfig,
+        registry: RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        mapped_container: Optional[Type[Mapped[Any]]],
+        annotation: Optional[_AnnotationScanType],
+        extracted_mapped_annotation: Optional[_AnnotationScanType],
+        is_dataclass_field: bool,
+    ) -> None:
+        """Perform class-specific initializaton at early declarative scanning
+        time.
+
+        .. versionadded:: 2.0
+
+        """
+
+    def _raise_for_required(self, key: str, cls: Type[Any]) -> NoReturn:
+        raise sa_exc.ArgumentError(
+            f"Python typing annotation is required for attribute "
+            f'"{cls.__name__}.{key}" when primary argument(s) for '
+            f'"{self._mapper_property_name()}" '
+            "construct are None or not present"
+        )
+
+
+class _AttributeOptions(NamedTuple):
+    """define Python-local attribute behavior options common to all
+    :class:`.MapperProperty` objects.
+
+    Currently this includes dataclass-generation arguments.
+
+    .. versionadded:: 2.0
+
+    """
+
+    dataclasses_init: Union[_NoArg, bool]
+    dataclasses_repr: Union[_NoArg, bool]
+    dataclasses_default: Union[_NoArg, Any]
+    dataclasses_default_factory: Union[_NoArg, Callable[[], Any]]
+    dataclasses_compare: Union[_NoArg, bool]
+    dataclasses_kw_only: Union[_NoArg, bool]
+    dataclasses_hash: Union[_NoArg, bool, None]
+
+    def _as_dataclass_field(self, key: str) -> Any:
+        """Return a ``dataclasses.Field`` object given these arguments."""
+
+        kw: Dict[str, Any] = {}
+        if self.dataclasses_default_factory is not _NoArg.NO_ARG:
+            kw["default_factory"] = self.dataclasses_default_factory
+        if self.dataclasses_default is not _NoArg.NO_ARG:
+            kw["default"] = self.dataclasses_default
+        if self.dataclasses_init is not _NoArg.NO_ARG:
+            kw["init"] = self.dataclasses_init
+        if self.dataclasses_repr is not _NoArg.NO_ARG:
+            kw["repr"] = self.dataclasses_repr
+        if self.dataclasses_compare is not _NoArg.NO_ARG:
+            kw["compare"] = self.dataclasses_compare
+        if self.dataclasses_kw_only is not _NoArg.NO_ARG:
+            kw["kw_only"] = self.dataclasses_kw_only
+        if self.dataclasses_hash is not _NoArg.NO_ARG:
+            kw["hash"] = self.dataclasses_hash
+
+        if "default" in kw and callable(kw["default"]):
+            # callable defaults are ambiguous. deprecate them in favour of
+            # insert_default or default_factory. #9936
+            warn_deprecated(
+                f"Callable object passed to the ``default`` parameter for "
+                f"attribute {key!r} in a ORM-mapped Dataclasses context is "
+                "ambiguous, "
+                "and this use will raise an error in a future release.  "
+                "If this callable is intended to produce Core level INSERT "
+                "default values for an underlying ``Column``, use "
+                "the ``mapped_column.insert_default`` parameter instead.  "
+                "To establish this callable as providing a default value "
+                "for instances of the dataclass itself, use the "
+                "``default_factory`` dataclasses parameter.",
+                "2.0",
+            )
+
+        if (
+            "init" in kw
+            and not kw["init"]
+            and "default" in kw
+            and not callable(kw["default"])  # ignore callable defaults. #9936
+            and "default_factory" not in kw  # illegal but let dc.field raise
+        ):
+            # fix for #9879
+            default = kw.pop("default")
+            kw["default_factory"] = lambda: default
+
+        return dataclasses.field(**kw)
+
+    @classmethod
+    def _get_arguments_for_make_dataclass(
+        cls,
+        key: str,
+        annotation: _AnnotationScanType,
+        mapped_container: Optional[Any],
+        elem: _T,
+    ) -> Union[
+        Tuple[str, _AnnotationScanType],
+        Tuple[str, _AnnotationScanType, dataclasses.Field[Any]],
+    ]:
+        """given attribute key, annotation, and value from a class, return
+        the argument tuple we would pass to dataclasses.make_dataclass()
+        for this attribute.
+
+        """
+        if isinstance(elem, _DCAttributeOptions):
+            dc_field = elem._attribute_options._as_dataclass_field(key)
+
+            return (key, annotation, dc_field)
+        elif elem is not _NoArg.NO_ARG:
+            # why is typing not erroring on this?
+            return (key, annotation, elem)
+        elif mapped_container is not None:
+            # it's Mapped[], but there's no "element", which means declarative
+            # did not actually do anything for this field.  this shouldn't
+            # happen.
+            # previously, this would occur because _scan_attributes would
+            # skip a field that's on an already mapped superclass, but it
+            # would still include it in the annotations, leading
+            # to issue #8718
+
+            assert False, "Mapped[] received without a mapping declaration"
+
+        else:
+            # plain dataclass field, not mapped.  Is only possible
+            # if __allow_unmapped__ is set up.  I can see this mode causing
+            # problems...
+            return (key, annotation)
+
+
+_DEFAULT_ATTRIBUTE_OPTIONS = _AttributeOptions(
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+)
+
+_DEFAULT_READONLY_ATTRIBUTE_OPTIONS = _AttributeOptions(
+    False,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+    _NoArg.NO_ARG,
+)
+
+
+class _DCAttributeOptions:
+    """mixin for descriptors or configurational objects that include dataclass
+    field options.
+
+    This includes :class:`.MapperProperty`, :class:`._MapsColumn` within
+    the ORM, but also includes :class:`.AssociationProxy` within ext.
+    Can in theory be used for other descriptors that serve a similar role
+    as association proxy.   (*maybe* hybrids, not sure yet.)
+
+    """
+
+    __slots__ = ()
+
+    _attribute_options: _AttributeOptions
+    """behavioral options for ORM-enabled Python attributes
+
+    .. versionadded:: 2.0
+
+    """
+
+    _has_dataclass_arguments: bool
+
+
+class _MapsColumns(_DCAttributeOptions, _MappedAttribute[_T]):
+    """interface for declarative-capable construct that delivers one or more
+    Column objects to the declarative process to be part of a Table.
+    """
+
+    __slots__ = ()
+
+    @property
+    def mapper_property_to_assign(self) -> Optional[MapperProperty[_T]]:
+        """return a MapperProperty to be assigned to the declarative mapping"""
+        raise NotImplementedError()
+
+    @property
+    def columns_to_assign(self) -> List[Tuple[Column[_T], int]]:
+        """A list of Column objects that should be declaratively added to the
+        new Table object.
+
+        """
+        raise NotImplementedError()
+
+
+# NOTE: MapperProperty needs to extend _MappedAttribute so that declarative
+# typing works, i.e. "Mapped[A] = relationship()".   This introduces an
+# inconvenience which is that all the MapperProperty objects are treated
+# as descriptors by typing tools, which are misled by this as assignment /
+# access to a descriptor attribute wants to move through __get__.
+# Therefore, references to MapperProperty as an instance variable, such
+# as in PropComparator, may have some special typing workarounds such as the
+# use of sqlalchemy.util.typing.DescriptorReference to avoid mis-interpretation
+# by typing tools
+@inspection._self_inspects
+class MapperProperty(
+    HasCacheKey,
+    _DCAttributeOptions,
+    _MappedAttribute[_T],
+    InspectionAttrInfo,
+    util.MemoizedSlots,
+):
+    """Represent a particular class attribute mapped by :class:`_orm.Mapper`.
+
+    The most common occurrences of :class:`.MapperProperty` are the
+    mapped :class:`_schema.Column`, which is represented in a mapping as
+    an instance of :class:`.ColumnProperty`,
+    and a reference to another class produced by :func:`_orm.relationship`,
+    represented in the mapping as an instance of
+    :class:`.Relationship`.
+
+    """
+
+    __slots__ = (
+        "_configure_started",
+        "_configure_finished",
+        "_attribute_options",
+        "_has_dataclass_arguments",
+        "parent",
+        "key",
+        "info",
+        "doc",
+    )
+
+    _cache_key_traversal: _TraverseInternalsType = [
+        ("parent", visitors.ExtendedInternalTraversal.dp_has_cache_key),
+        ("key", visitors.ExtendedInternalTraversal.dp_string),
+    ]
+
+    if not TYPE_CHECKING:
+        cascade = None
+
+    is_property = True
+    """Part of the InspectionAttr interface; states this object is a
+    mapper property.
+
+    """
+
+    comparator: PropComparator[_T]
+    """The :class:`_orm.PropComparator` instance that implements SQL
+    expression construction on behalf of this mapped attribute."""
+
+    key: str
+    """name of class attribute"""
+
+    parent: Mapper[Any]
+    """the :class:`.Mapper` managing this property."""
+
+    _is_relationship = False
+
+    _links_to_entity: bool
+    """True if this MapperProperty refers to a mapped entity.
+
+    Should only be True for Relationship, False for all others.
+
+    """
+
+    doc: Optional[str]
+    """optional documentation string"""
+
+    info: _InfoType
+    """Info dictionary associated with the object, allowing user-defined
+    data to be associated with this :class:`.InspectionAttr`.
+
+    The dictionary is generated when first accessed.  Alternatively,
+    it can be specified as a constructor argument to the
+    :func:`.column_property`, :func:`_orm.relationship`, or :func:`.composite`
+    functions.
+
+    .. seealso::
+
+        :attr:`.QueryableAttribute.info`
+
+        :attr:`.SchemaItem.info`
+
+    """
+
+    def _memoized_attr_info(self) -> _InfoType:
+        """Info dictionary associated with the object, allowing user-defined
+        data to be associated with this :class:`.InspectionAttr`.
+
+        The dictionary is generated when first accessed.  Alternatively,
+        it can be specified as a constructor argument to the
+        :func:`.column_property`, :func:`_orm.relationship`, or
+        :func:`.composite`
+        functions.
+
+        .. seealso::
+
+            :attr:`.QueryableAttribute.info`
+
+            :attr:`.SchemaItem.info`
+
+        """
+        return {}
+
+    def setup(
+        self,
+        context: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        adapter: Optional[ORMAdapter],
+        **kwargs: Any,
+    ) -> None:
+        """Called by Query for the purposes of constructing a SQL statement.
+
+        Each MapperProperty associated with the target mapper processes the
+        statement referenced by the query context, adding columns and/or
+        criterion as appropriate.
+
+        """
+
+    def create_row_processor(
+        self,
+        context: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        mapper: Mapper[Any],
+        result: Result[Any],
+        adapter: Optional[ORMAdapter],
+        populators: _PopulatorDict,
+    ) -> None:
+        """Produce row processing functions and append to the given
+        set of populators lists.
+
+        """
+
+    def cascade_iterator(
+        self,
+        type_: str,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        visited_states: Set[InstanceState[Any]],
+        halt_on: Optional[Callable[[InstanceState[Any]], bool]] = None,
+    ) -> Iterator[
+        Tuple[object, Mapper[Any], InstanceState[Any], _InstanceDict]
+    ]:
+        """Iterate through instances related to the given instance for
+        a particular 'cascade', starting with this MapperProperty.
+
+        Return an iterator3-tuples (instance, mapper, state).
+
+        Note that the 'cascade' collection on this MapperProperty is
+        checked first for the given type before cascade_iterator is called.
+
+        This method typically only applies to Relationship.
+
+        """
+
+        return iter(())
+
+    def set_parent(self, parent: Mapper[Any], init: bool) -> None:
+        """Set the parent mapper that references this MapperProperty.
+
+        This method is overridden by some subclasses to perform extra
+        setup when the mapper is first known.
+
+        """
+        self.parent = parent
+
+    def instrument_class(self, mapper: Mapper[Any]) -> None:
+        """Hook called by the Mapper to the property to initiate
+        instrumentation of the class attribute managed by this
+        MapperProperty.
+
+        The MapperProperty here will typically call out to the
+        attributes module to set up an InstrumentedAttribute.
+
+        This step is the first of two steps to set up an InstrumentedAttribute,
+        and is called early in the mapper setup process.
+
+        The second step is typically the init_class_attribute step,
+        called from StrategizedProperty via the post_instrument_class()
+        hook.  This step assigns additional state to the InstrumentedAttribute
+        (specifically the "impl") which has been determined after the
+        MapperProperty has determined what kind of persistence
+        management it needs to do (e.g. scalar, object, collection, etc).
+
+        """
+
+    def __init__(
+        self,
+        attribute_options: Optional[_AttributeOptions] = None,
+        _assume_readonly_dc_attributes: bool = False,
+    ) -> None:
+        self._configure_started = False
+        self._configure_finished = False
+
+        if _assume_readonly_dc_attributes:
+            default_attrs = _DEFAULT_READONLY_ATTRIBUTE_OPTIONS
+        else:
+            default_attrs = _DEFAULT_ATTRIBUTE_OPTIONS
+
+        if attribute_options and attribute_options != default_attrs:
+            self._has_dataclass_arguments = True
+            self._attribute_options = attribute_options
+        else:
+            self._has_dataclass_arguments = False
+            self._attribute_options = default_attrs
+
+    def init(self) -> None:
+        """Called after all mappers are created to assemble
+        relationships between mappers and perform other post-mapper-creation
+        initialization steps.
+
+
+        """
+        self._configure_started = True
+        self.do_init()
+        self._configure_finished = True
+
+    @property
+    def class_attribute(self) -> InstrumentedAttribute[_T]:
+        """Return the class-bound descriptor corresponding to this
+        :class:`.MapperProperty`.
+
+        This is basically a ``getattr()`` call::
+
+            return getattr(self.parent.class_, self.key)
+
+        I.e. if this :class:`.MapperProperty` were named ``addresses``,
+        and the class to which it is mapped is ``User``, this sequence
+        is possible::
+
+            >>> from sqlalchemy import inspect
+            >>> mapper = inspect(User)
+            >>> addresses_property = mapper.attrs.addresses
+            >>> addresses_property.class_attribute is User.addresses
+            True
+            >>> User.addresses.property is addresses_property
+            True
+
+
+        """
+
+        return getattr(self.parent.class_, self.key)  # type: ignore
+
+    def do_init(self) -> None:
+        """Perform subclass-specific initialization post-mapper-creation
+        steps.
+
+        This is a template method called by the ``MapperProperty``
+        object's init() method.
+
+        """
+
+    def post_instrument_class(self, mapper: Mapper[Any]) -> None:
+        """Perform instrumentation adjustments that need to occur
+        after init() has completed.
+
+        The given Mapper is the Mapper invoking the operation, which
+        may not be the same Mapper as self.parent in an inheritance
+        scenario; however, Mapper will always at least be a sub-mapper of
+        self.parent.
+
+        This method is typically used by StrategizedProperty, which delegates
+        it to LoaderStrategy.init_class_attribute() to perform final setup
+        on the class-bound InstrumentedAttribute.
+
+        """
+
+    def merge(
+        self,
+        session: Session,
+        source_state: InstanceState[Any],
+        source_dict: _InstanceDict,
+        dest_state: InstanceState[Any],
+        dest_dict: _InstanceDict,
+        load: bool,
+        _recursive: Dict[Any, object],
+        _resolve_conflict_map: Dict[_IdentityKeyType[Any], object],
+    ) -> None:
+        """Merge the attribute represented by this ``MapperProperty``
+        from source to destination object.
+
+        """
+
+    def __repr__(self) -> str:
+        return "<%s at 0x%x; %s>" % (
+            self.__class__.__name__,
+            id(self),
+            getattr(self, "key", "no key"),
+        )
+
+
+@inspection._self_inspects
+class PropComparator(SQLORMOperations[_T_co], Generic[_T_co], ColumnOperators):
+    r"""Defines SQL operations for ORM mapped attributes.
+
+    SQLAlchemy allows for operators to
+    be redefined at both the Core and ORM level.  :class:`.PropComparator`
+    is the base class of operator redefinition for ORM-level operations,
+    including those of :class:`.ColumnProperty`,
+    :class:`.Relationship`, and :class:`.Composite`.
+
+    User-defined subclasses of :class:`.PropComparator` may be created. The
+    built-in Python comparison and math operator methods, such as
+    :meth:`.operators.ColumnOperators.__eq__`,
+    :meth:`.operators.ColumnOperators.__lt__`, and
+    :meth:`.operators.ColumnOperators.__add__`, can be overridden to provide
+    new operator behavior. The custom :class:`.PropComparator` is passed to
+    the :class:`.MapperProperty` instance via the ``comparator_factory``
+    argument. In each case,
+    the appropriate subclass of :class:`.PropComparator` should be used::
+
+        # definition of custom PropComparator subclasses
+
+        from sqlalchemy.orm.properties import (
+            ColumnProperty,
+            Composite,
+            Relationship,
+        )
+
+
+        class MyColumnComparator(ColumnProperty.Comparator):
+            def __eq__(self, other):
+                return self.__clause_element__() == other
+
+
+        class MyRelationshipComparator(Relationship.Comparator):
+            def any(self, expression):
+                "define the 'any' operation"
+                # ...
+
+
+        class MyCompositeComparator(Composite.Comparator):
+            def __gt__(self, other):
+                "redefine the 'greater than' operation"
+
+                return sql.and_(
+                    *[
+                        a > b
+                        for a, b in zip(
+                            self.__clause_element__().clauses,
+                            other.__composite_values__(),
+                        )
+                    ]
+                )
+
+
+        # application of custom PropComparator subclasses
+
+        from sqlalchemy.orm import column_property, relationship, composite
+        from sqlalchemy import Column, String
+
+
+        class SomeMappedClass(Base):
+            some_column = column_property(
+                Column("some_column", String),
+                comparator_factory=MyColumnComparator,
+            )
+
+            some_relationship = relationship(
+                SomeOtherClass, comparator_factory=MyRelationshipComparator
+            )
+
+            some_composite = composite(
+                Column("a", String),
+                Column("b", String),
+                comparator_factory=MyCompositeComparator,
+            )
+
+    Note that for column-level operator redefinition, it's usually
+    simpler to define the operators at the Core level, using the
+    :attr:`.TypeEngine.comparator_factory` attribute.  See
+    :ref:`types_operators` for more detail.
+
+    .. seealso::
+
+        :class:`.ColumnProperty.Comparator`
+
+        :class:`.Relationship.Comparator`
+
+        :class:`.Composite.Comparator`
+
+        :class:`.ColumnOperators`
+
+        :ref:`types_operators`
+
+        :attr:`.TypeEngine.comparator_factory`
+
+    """
+
+    __slots__ = "prop", "_parententity", "_adapt_to_entity"
+
+    __visit_name__ = "orm_prop_comparator"
+
+    _parententity: _InternalEntityType[Any]
+    _adapt_to_entity: Optional[AliasedInsp[Any]]
+    prop: RODescriptorReference[MapperProperty[_T_co]]
+
+    def __init__(
+        self,
+        prop: MapperProperty[_T],
+        parentmapper: _InternalEntityType[Any],
+        adapt_to_entity: Optional[AliasedInsp[Any]] = None,
+    ):
+        self.prop = prop
+        self._parententity = adapt_to_entity or parentmapper
+        self._adapt_to_entity = adapt_to_entity
+
+    @util.non_memoized_property
+    def property(self) -> MapperProperty[_T_co]:
+        """Return the :class:`.MapperProperty` associated with this
+        :class:`.PropComparator`.
+
+
+        Return values here will commonly be instances of
+        :class:`.ColumnProperty` or :class:`.Relationship`.
+
+
+        """
+        return self.prop
+
+    def __clause_element__(self) -> roles.ColumnsClauseRole:
+        raise NotImplementedError("%r" % self)
+
+    def _bulk_update_tuples(
+        self, value: Any
+    ) -> Sequence[Tuple[_DMLColumnArgument, Any]]:
+        """Receive a SQL expression that represents a value in the SET
+        clause of an UPDATE statement.
+
+        Return a tuple that can be passed to a :class:`_expression.Update`
+        construct.
+
+        """
+
+        return [(cast("_DMLColumnArgument", self.__clause_element__()), value)]
+
+    def adapt_to_entity(
+        self, adapt_to_entity: AliasedInsp[Any]
+    ) -> PropComparator[_T_co]:
+        """Return a copy of this PropComparator which will use the given
+        :class:`.AliasedInsp` to produce corresponding expressions.
+        """
+        return self.__class__(self.prop, self._parententity, adapt_to_entity)
+
+    @util.ro_non_memoized_property
+    def _parentmapper(self) -> Mapper[Any]:
+        """legacy; this is renamed to _parententity to be
+        compatible with QueryableAttribute."""
+        return self._parententity.mapper
+
+    def _criterion_exists(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[Any]:
+        return self.prop.comparator._criterion_exists(criterion, **kwargs)
+
+    @util.ro_non_memoized_property
+    def adapter(self) -> Optional[_ORMAdapterProto]:
+        """Produce a callable that adapts column expressions
+        to suit an aliased version of this comparator.
+
+        """
+        if self._adapt_to_entity is None:
+            return None
+        else:
+            return self._adapt_to_entity._orm_adapt_element
+
+    @util.ro_non_memoized_property
+    def info(self) -> _InfoType:
+        return self.prop.info
+
+    @staticmethod
+    def _any_op(a: Any, b: Any, **kwargs: Any) -> Any:
+        return a.any(b, **kwargs)
+
+    @staticmethod
+    def _has_op(left: Any, other: Any, **kwargs: Any) -> Any:
+        return left.has(other, **kwargs)
+
+    @staticmethod
+    def _of_type_op(a: Any, class_: Any) -> Any:
+        return a.of_type(class_)
+
+    any_op = cast(operators.OperatorType, _any_op)
+    has_op = cast(operators.OperatorType, _has_op)
+    of_type_op = cast(operators.OperatorType, _of_type_op)
+
+    if typing.TYPE_CHECKING:
+
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]: ...
+
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]: ...
+
+    def of_type(self, class_: _EntityType[Any]) -> PropComparator[_T_co]:
+        r"""Redefine this object in terms of a polymorphic subclass,
+        :func:`_orm.with_polymorphic` construct, or :func:`_orm.aliased`
+        construct.
+
+        Returns a new PropComparator from which further criterion can be
+        evaluated.
+
+        e.g.::
+
+            query.join(Company.employees.of_type(Engineer)).filter(
+                Engineer.name == "foo"
+            )
+
+        :param \class_: a class or mapper indicating that criterion will be
+            against this specific subclass.
+
+        .. seealso::
+
+            :ref:`orm_queryguide_joining_relationships_aliased` - in the
+            :ref:`queryguide_toplevel`
+
+            :ref:`inheritance_of_type`
+
+        """
+
+        return self.operate(PropComparator.of_type_op, class_)  # type: ignore
+
+    def and_(
+        self, *criteria: _ColumnExpressionArgument[bool]
+    ) -> PropComparator[bool]:
+        """Add additional criteria to the ON clause that's represented by this
+        relationship attribute.
+
+        E.g.::
+
+
+            stmt = select(User).join(
+                User.addresses.and_(Address.email_address != "foo")
+            )
+
+            stmt = select(User).options(
+                joinedload(User.addresses.and_(Address.email_address != "foo"))
+            )
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :ref:`orm_queryguide_join_on_augmented`
+
+            :ref:`loader_option_criteria`
+
+            :func:`.with_loader_criteria`
+
+        """
+        return self.operate(operators.and_, *criteria)  # type: ignore
+
+    def any(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[bool]:
+        r"""Return a SQL expression representing true if this element
+        references a member which meets the given criterion.
+
+        The usual implementation of ``any()`` is
+        :meth:`.Relationship.Comparator.any`.
+
+        :param criterion: an optional ClauseElement formulated against the
+          member class' table or attributes.
+
+        :param \**kwargs: key/value pairs corresponding to member class
+          attribute names which will be compared via equality to the
+          corresponding values.
+
+        """
+
+        return self.operate(PropComparator.any_op, criterion, **kwargs)
+
+    def has(
+        self,
+        criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+        **kwargs: Any,
+    ) -> ColumnElement[bool]:
+        r"""Return a SQL expression representing true if this element
+        references a member which meets the given criterion.
+
+        The usual implementation of ``has()`` is
+        :meth:`.Relationship.Comparator.has`.
+
+        :param criterion: an optional ClauseElement formulated against the
+          member class' table or attributes.
+
+        :param \**kwargs: key/value pairs corresponding to member class
+          attribute names which will be compared via equality to the
+          corresponding values.
+
+        """
+
+        return self.operate(PropComparator.has_op, criterion, **kwargs)
+
+
+class StrategizedProperty(MapperProperty[_T]):
+    """A MapperProperty which uses selectable strategies to affect
+    loading behavior.
+
+    There is a single strategy selected by default.  Alternate
+    strategies can be selected at Query time through the usage of
+    ``StrategizedOption`` objects via the Query.options() method.
+
+    The mechanics of StrategizedProperty are used for every Query
+    invocation for every mapped attribute participating in that Query,
+    to determine first how the attribute will be rendered in SQL
+    and secondly how the attribute will retrieve a value from a result
+    row and apply it to a mapped object.  The routines here are very
+    performance-critical.
+
+    """
+
+    __slots__ = (
+        "_strategies",
+        "strategy",
+        "_wildcard_token",
+        "_default_path_loader_key",
+        "strategy_key",
+    )
+    inherit_cache = True
+    strategy_wildcard_key: ClassVar[str]
+
+    strategy_key: _StrategyKey
+
+    _strategies: Dict[_StrategyKey, LoaderStrategy]
+
+    def _memoized_attr__wildcard_token(self) -> Tuple[str]:
+        return (
+            f"{self.strategy_wildcard_key}:{path_registry._WILDCARD_TOKEN}",
+        )
+
+    def _memoized_attr__default_path_loader_key(
+        self,
+    ) -> Tuple[str, Tuple[str]]:
+        return (
+            "loader",
+            (f"{self.strategy_wildcard_key}:{path_registry._DEFAULT_TOKEN}",),
+        )
+
+    def _get_context_loader(
+        self, context: ORMCompileState, path: AbstractEntityRegistry
+    ) -> Optional[_LoadElement]:
+        load: Optional[_LoadElement] = None
+
+        search_path = path[self]
+
+        # search among: exact match, "attr.*", "default" strategy
+        # if any.
+        for path_key in (
+            search_path._loader_key,
+            search_path._wildcard_path_loader_key,
+            search_path._default_path_loader_key,
+        ):
+            if path_key in context.attributes:
+                load = context.attributes[path_key]
+                break
+
+                # note that if strategy_options.Load is placing non-actionable
+                # objects in the context like defaultload(), we would
+                # need to continue the loop here if we got such an
+                # option as below.
+                # if load.strategy or load.local_opts:
+                #    break
+
+        return load
+
+    def _get_strategy(self, key: _StrategyKey) -> LoaderStrategy:
+        try:
+            return self._strategies[key]
+        except KeyError:
+            pass
+
+        # run outside to prevent transfer of exception context
+        cls = self._strategy_lookup(self, *key)
+        # this previously was setting self._strategies[cls], that's
+        # a bad idea; should use strategy key at all times because every
+        # strategy has multiple keys at this point
+        self._strategies[key] = strategy = cls(self, key)
+        return strategy
+
+    def setup(
+        self,
+        context: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        adapter: Optional[ORMAdapter],
+        **kwargs: Any,
+    ) -> None:
+        loader = self._get_context_loader(context, path)
+        if loader and loader.strategy:
+            strat = self._get_strategy(loader.strategy)
+        else:
+            strat = self.strategy
+        strat.setup_query(
+            context, query_entity, path, loader, adapter, **kwargs
+        )
+
+    def create_row_processor(
+        self,
+        context: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        mapper: Mapper[Any],
+        result: Result[Any],
+        adapter: Optional[ORMAdapter],
+        populators: _PopulatorDict,
+    ) -> None:
+        loader = self._get_context_loader(context, path)
+        if loader and loader.strategy:
+            strat = self._get_strategy(loader.strategy)
+        else:
+            strat = self.strategy
+        strat.create_row_processor(
+            context,
+            query_entity,
+            path,
+            loader,
+            mapper,
+            result,
+            adapter,
+            populators,
+        )
+
+    def do_init(self) -> None:
+        self._strategies = {}
+        self.strategy = self._get_strategy(self.strategy_key)
+
+    def post_instrument_class(self, mapper: Mapper[Any]) -> None:
+        if (
+            not self.parent.non_primary
+            and not mapper.class_manager._attr_has_impl(self.key)
+        ):
+            self.strategy.init_class_attribute(mapper)
+
+    _all_strategies: collections.defaultdict[
+        Type[MapperProperty[Any]], Dict[_StrategyKey, Type[LoaderStrategy]]
+    ] = collections.defaultdict(dict)
+
+    @classmethod
+    def strategy_for(cls, **kw: Any) -> Callable[[_TLS], _TLS]:
+        def decorate(dec_cls: _TLS) -> _TLS:
+            # ensure each subclass of the strategy has its
+            # own _strategy_keys collection
+            if "_strategy_keys" not in dec_cls.__dict__:
+                dec_cls._strategy_keys = []
+            key = tuple(sorted(kw.items()))
+            cls._all_strategies[cls][key] = dec_cls
+            dec_cls._strategy_keys.append(key)
+            return dec_cls
+
+        return decorate
+
+    @classmethod
+    def _strategy_lookup(
+        cls, requesting_property: MapperProperty[Any], *key: Any
+    ) -> Type[LoaderStrategy]:
+        requesting_property.parent._with_polymorphic_mappers
+
+        for prop_cls in cls.__mro__:
+            if prop_cls in cls._all_strategies:
+                if TYPE_CHECKING:
+                    assert issubclass(prop_cls, MapperProperty)
+                strategies = cls._all_strategies[prop_cls]
+                try:
+                    return strategies[key]
+                except KeyError:
+                    pass
+
+        for property_type, strats in cls._all_strategies.items():
+            if key in strats:
+                intended_property_type = property_type
+                actual_strategy = strats[key]
+                break
+        else:
+            intended_property_type = None
+            actual_strategy = None
+
+        raise orm_exc.LoaderStrategyException(
+            cls,
+            requesting_property,
+            intended_property_type,
+            actual_strategy,
+            key,
+        )
+
+
+class ORMOption(ExecutableOption):
+    """Base class for option objects that are passed to ORM queries.
+
+    These options may be consumed by :meth:`.Query.options`,
+    :meth:`.Select.options`, or in a more general sense by any
+    :meth:`.Executable.options` method.   They are interpreted at
+    statement compile time or execution time in modern use.  The
+    deprecated :class:`.MapperOption` is consumed at ORM query construction
+    time.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    _is_legacy_option = False
+
+    propagate_to_loaders = False
+    """if True, indicate this option should be carried along
+    to "secondary" SELECT statements that occur for relationship
+    lazy loaders as well as attribute load / refresh operations.
+
+    """
+
+    _is_core = False
+
+    _is_user_defined = False
+
+    _is_compile_state = False
+
+    _is_criteria_option = False
+
+    _is_strategy_option = False
+
+    def _adapt_cached_option_to_uncached_option(
+        self, context: QueryContext, uncached_opt: ORMOption
+    ) -> ORMOption:
+        """adapt this option to the "uncached" version of itself in a
+        loader strategy context.
+
+        given "self" which is an option from a cached query, as well as the
+        corresponding option from the uncached version of the same query,
+        return the option we should use in a new query, in the context of a
+        loader strategy being asked to load related rows on behalf of that
+        cached query, which is assumed to be building a new query based on
+        entities passed to us from the cached query.
+
+        Currently this routine chooses between "self" and "uncached" without
+        manufacturing anything new. If the option is itself a loader strategy
+        option which has a path, that path needs to match to the entities being
+        passed to us by the cached query, so the :class:`_orm.Load` subclass
+        overrides this to return "self". For all other options, we return the
+        uncached form which may have changing state, such as a
+        with_loader_criteria() option which will very often have new state.
+
+        This routine could in the future involve
+        generating a new option based on both inputs if use cases arise,
+        such as if with_loader_criteria() needed to match up to
+        ``AliasedClass`` instances given in the parent query.
+
+        However, longer term it might be better to restructure things such that
+        ``AliasedClass`` entities are always matched up on their cache key,
+        instead of identity, in things like paths and such, so that this whole
+        issue of "the uncached option does not match the entities" goes away.
+        However this would make ``PathRegistry`` more complicated and difficult
+        to debug as well as potentially less performant in that it would be
+        hashing enormous cache keys rather than a simple AliasedInsp. UNLESS,
+        we could get cache keys overall to be reliably hashed into something
+        like an md5 key.
+
+        .. versionadded:: 1.4.41
+
+        """
+        if uncached_opt is not None:
+            return uncached_opt
+        else:
+            return self
+
+
+class CompileStateOption(HasCacheKey, ORMOption):
+    """base for :class:`.ORMOption` classes that affect the compilation of
+    a SQL query and therefore need to be part of the cache key.
+
+    .. note::  :class:`.CompileStateOption` is generally non-public and
+       should not be used as a base class for user-defined options; instead,
+       use :class:`.UserDefinedOption`, which is easier to use as it does not
+       interact with ORM compilation internals or caching.
+
+    :class:`.CompileStateOption` defines an internal attribute
+    ``_is_compile_state=True`` which has the effect of the ORM compilation
+    routines for SELECT and other statements will call upon these options when
+    a SQL string is being compiled. As such, these classes implement
+    :class:`.HasCacheKey` and need to provide robust ``_cache_key_traversal``
+    structures.
+
+    The :class:`.CompileStateOption` class is used to implement the ORM
+    :class:`.LoaderOption` and :class:`.CriteriaOption` classes.
+
+    .. versionadded:: 1.4.28
+
+
+    """
+
+    __slots__ = ()
+
+    _is_compile_state = True
+
+    def process_compile_state(self, compile_state: ORMCompileState) -> None:
+        """Apply a modification to a given :class:`.ORMCompileState`.
+
+        This method is part of the implementation of a particular
+        :class:`.CompileStateOption` and is only invoked internally
+        when an ORM query is compiled.
+
+        """
+
+    def process_compile_state_replaced_entities(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Sequence[_MapperEntity],
+    ) -> None:
+        """Apply a modification to a given :class:`.ORMCompileState`,
+        given entities that were replaced by with_only_columns() or
+        with_entities().
+
+        This method is part of the implementation of a particular
+        :class:`.CompileStateOption` and is only invoked internally
+        when an ORM query is compiled.
+
+        .. versionadded:: 1.4.19
+
+        """
+
+
+class LoaderOption(CompileStateOption):
+    """Describe a loader modification to an ORM statement at compilation time.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    def process_compile_state_replaced_entities(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Sequence[_MapperEntity],
+    ) -> None:
+        self.process_compile_state(compile_state)
+
+
+class CriteriaOption(CompileStateOption):
+    """Describe a WHERE criteria modification to an ORM statement at
+    compilation time.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+
+    _is_criteria_option = True
+
+    def get_global_criteria(self, attributes: Dict[str, Any]) -> None:
+        """update additional entity criteria options in the given
+        attributes dictionary.
+
+        """
+
+
+class UserDefinedOption(ORMOption):
+    """Base class for a user-defined option that can be consumed from the
+    :meth:`.SessionEvents.do_orm_execute` event hook.
+
+    """
+
+    __slots__ = ("payload",)
+
+    _is_legacy_option = False
+
+    _is_user_defined = True
+
+    propagate_to_loaders = False
+    """if True, indicate this option should be carried along
+    to "secondary" Query objects produced during lazy loads
+    or refresh operations.
+
+    """
+
+    def __init__(self, payload: Optional[Any] = None):
+        self.payload = payload
+
+
+@util.deprecated_cls(
+    "1.4",
+    "The :class:`.MapperOption class is deprecated and will be removed "
+    "in a future release.   For "
+    "modifications to queries on a per-execution basis, use the "
+    ":class:`.UserDefinedOption` class to establish state within a "
+    ":class:`.Query` or other Core statement, then use the "
+    ":meth:`.SessionEvents.before_orm_execute` hook to consume them.",
+    constructor=None,
+)
+class MapperOption(ORMOption):
+    """Describe a modification to a Query"""
+
+    __slots__ = ()
+
+    _is_legacy_option = True
+
+    propagate_to_loaders = False
+    """if True, indicate this option should be carried along
+    to "secondary" Query objects produced during lazy loads
+    or refresh operations.
+
+    """
+
+    def process_query(self, query: Query[Any]) -> None:
+        """Apply a modification to the given :class:`_query.Query`."""
+
+    def process_query_conditionally(self, query: Query[Any]) -> None:
+        """same as process_query(), except that this option may not
+        apply to the given query.
+
+        This is typically applied during a lazy load or scalar refresh
+        operation to propagate options stated in the original Query to the
+        new Query being used for the load.  It occurs for those options that
+        specify propagate_to_loaders=True.
+
+        """
+
+        self.process_query(query)
+
+
+class LoaderStrategy:
+    """Describe the loading behavior of a StrategizedProperty object.
+
+    The ``LoaderStrategy`` interacts with the querying process in three
+    ways:
+
+    * it controls the configuration of the ``InstrumentedAttribute``
+      placed on a class to handle the behavior of the attribute.  this
+      may involve setting up class-level callable functions to fire
+      off a select operation when the attribute is first accessed
+      (i.e. a lazy load)
+
+    * it processes the ``QueryContext`` at statement construction time,
+      where it can modify the SQL statement that is being produced.
+      For example, simple column attributes will add their represented
+      column to the list of selected columns, a joined eager loader
+      may establish join clauses to add to the statement.
+
+    * It produces "row processor" functions at result fetching time.
+      These "row processor" functions populate a particular attribute
+      on a particular mapped instance.
+
+    """
+
+    __slots__ = (
+        "parent_property",
+        "is_class_level",
+        "parent",
+        "key",
+        "strategy_key",
+        "strategy_opts",
+    )
+
+    _strategy_keys: ClassVar[List[_StrategyKey]]
+
+    def __init__(
+        self, parent: MapperProperty[Any], strategy_key: _StrategyKey
+    ):
+        self.parent_property = parent
+        self.is_class_level = False
+        self.parent = self.parent_property.parent
+        self.key = self.parent_property.key
+        self.strategy_key = strategy_key
+        self.strategy_opts = dict(strategy_key)
+
+    def init_class_attribute(self, mapper: Mapper[Any]) -> None:
+        pass
+
+    def setup_query(
+        self,
+        compile_state: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        loadopt: Optional[_LoadElement],
+        adapter: Optional[ORMAdapter],
+        **kwargs: Any,
+    ) -> None:
+        """Establish column and other state for a given QueryContext.
+
+        This method fulfills the contract specified by MapperProperty.setup().
+
+        StrategizedProperty delegates its setup() method
+        directly to this method.
+
+        """
+
+    def create_row_processor(
+        self,
+        context: ORMCompileState,
+        query_entity: _MapperEntity,
+        path: AbstractEntityRegistry,
+        loadopt: Optional[_LoadElement],
+        mapper: Mapper[Any],
+        result: Result[Any],
+        adapter: Optional[ORMAdapter],
+        populators: _PopulatorDict,
+    ) -> None:
+        """Establish row processing functions for a given QueryContext.
+
+        This method fulfills the contract specified by
+        MapperProperty.create_row_processor().
+
+        StrategizedProperty delegates its create_row_processor() method
+        directly to this method.
+
+        """
+
+    def __str__(self) -> str:
+        return str(self.parent_property)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/loading.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/loading.py
new file mode 100644
index 00000000..679286f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/loading.py
@@ -0,0 +1,1682 @@
+# orm/loading.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""private module containing functions used to convert database
+rows into object instances and associated state.
+
+the functions here are called primarily by Query, Mapper,
+as well as some of the attribute loading strategies.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import exc as orm_exc
+from . import path_registry
+from .base import _DEFER_FOR_STATE
+from .base import _RAISE_FOR_STATE
+from .base import _SET_DEFERRED_EXPIRED
+from .base import PassiveFlag
+from .context import FromStatement
+from .context import ORMCompileState
+from .context import QueryContext
+from .util import _none_set
+from .util import state_str
+from .. import exc as sa_exc
+from .. import util
+from ..engine import result_tuple
+from ..engine.result import ChunkedIteratorResult
+from ..engine.result import FrozenResult
+from ..engine.result import SimpleResultMetaData
+from ..sql import select
+from ..sql import util as sql_util
+from ..sql.selectable import ForUpdateArg
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..sql.selectable import SelectState
+from ..util import EMPTY_DICT
+
+if TYPE_CHECKING:
+    from ._typing import _IdentityKeyType
+    from .base import LoaderCallableStatus
+    from .interfaces import ORMOption
+    from .mapper import Mapper
+    from .query import Query
+    from .session import Session
+    from .state import InstanceState
+    from ..engine.cursor import CursorResult
+    from ..engine.interfaces import _ExecuteOptions
+    from ..engine.result import Result
+    from ..sql import Select
+
+_T = TypeVar("_T", bound=Any)
+_O = TypeVar("_O", bound=object)
+_new_runid = util.counter()
+
+
+_PopulatorDict = Dict[str, List[Tuple[str, Any]]]
+
+
+def instances(cursor: CursorResult[Any], context: QueryContext) -> Result[Any]:
+    """Return a :class:`.Result` given an ORM query context.
+
+    :param cursor: a :class:`.CursorResult`, generated by a statement
+     which came from :class:`.ORMCompileState`
+
+    :param context: a :class:`.QueryContext` object
+
+    :return: a :class:`.Result` object representing ORM results
+
+    .. versionchanged:: 1.4 The instances() function now uses
+       :class:`.Result` objects and has an all new interface.
+
+    """
+
+    context.runid = _new_runid()
+
+    if context.top_level_context:
+        is_top_level = False
+        context.post_load_paths = context.top_level_context.post_load_paths
+    else:
+        is_top_level = True
+        context.post_load_paths = {}
+
+    compile_state = context.compile_state
+    filtered = compile_state._has_mapper_entities
+    single_entity = (
+        not context.load_options._only_return_tuples
+        and len(compile_state._entities) == 1
+        and compile_state._entities[0].supports_single_entity
+    )
+
+    try:
+        (process, labels, extra) = list(
+            zip(
+                *[
+                    query_entity.row_processor(context, cursor)
+                    for query_entity in context.compile_state._entities
+                ]
+            )
+        )
+
+        if context.yield_per and (
+            context.loaders_require_buffering
+            or context.loaders_require_uniquing
+        ):
+            raise sa_exc.InvalidRequestError(
+                "Can't use yield_per with eager loaders that require uniquing "
+                "or row buffering, e.g. joinedload() against collections "
+                "or subqueryload().  Consider the selectinload() strategy "
+                "for better flexibility in loading objects."
+            )
+
+    except Exception:
+        with util.safe_reraise():
+            cursor.close()
+
+    def _no_unique(entry):
+        raise sa_exc.InvalidRequestError(
+            "Can't use the ORM yield_per feature in conjunction with unique()"
+        )
+
+    def _not_hashable(datatype, *, legacy=False, uncertain=False):
+        if not legacy:
+
+            def go(obj):
+                if uncertain:
+                    try:
+                        return hash(obj)
+                    except:
+                        pass
+
+                raise sa_exc.InvalidRequestError(
+                    "Can't apply uniqueness to row tuple containing value of "
+                    f"""type {datatype!r}; {
+                        'the values returned appear to be'
+                        if uncertain
+                        else 'this datatype produces'
+                    } non-hashable values"""
+                )
+
+            return go
+        elif not uncertain:
+            return id
+        else:
+            _use_id = False
+
+            def go(obj):
+                nonlocal _use_id
+
+                if not _use_id:
+                    try:
+                        return hash(obj)
+                    except:
+                        pass
+
+                    # in #10459, we considered using a warning here, however
+                    # as legacy query uses result.unique() in all cases, this
+                    # would lead to too many warning cases.
+                    _use_id = True
+
+                return id(obj)
+
+            return go
+
+    unique_filters = [
+        (
+            _no_unique
+            if context.yield_per
+            else (
+                _not_hashable(
+                    ent.column.type,  # type: ignore
+                    legacy=context.load_options._legacy_uniquing,
+                    uncertain=ent._null_column_type,
+                )
+                if (
+                    not ent.use_id_for_hash
+                    and (ent._non_hashable_value or ent._null_column_type)
+                )
+                else id if ent.use_id_for_hash else None
+            )
+        )
+        for ent in context.compile_state._entities
+    ]
+
+    row_metadata = SimpleResultMetaData(
+        labels, extra, _unique_filters=unique_filters
+    )
+
+    def chunks(size):  # type: ignore
+        while True:
+            yield_per = size
+
+            context.partials = {}
+
+            if yield_per:
+                fetch = cursor.fetchmany(yield_per)
+
+                if not fetch:
+                    break
+            else:
+                fetch = cursor._raw_all_rows()
+
+            if single_entity:
+                proc = process[0]
+                rows = [proc(row) for row in fetch]
+            else:
+                rows = [
+                    tuple([proc(row) for proc in process]) for row in fetch
+                ]
+
+            # if we are the originating load from a query, meaning we
+            # aren't being called as a result of a nested "post load",
+            # iterate through all the collected post loaders and fire them
+            # off.  Previously this used to work recursively, however that
+            # prevented deeply nested structures from being loadable
+            if is_top_level:
+                if yield_per:
+                    # if using yield per, memoize the state of the
+                    # collection so that it can be restored
+                    top_level_post_loads = list(
+                        context.post_load_paths.items()
+                    )
+
+                while context.post_load_paths:
+                    post_loads = list(context.post_load_paths.items())
+                    context.post_load_paths.clear()
+                    for path, post_load in post_loads:
+                        post_load.invoke(context, path)
+
+                if yield_per:
+                    context.post_load_paths.clear()
+                    context.post_load_paths.update(top_level_post_loads)
+
+            yield rows
+
+            if not yield_per:
+                break
+
+    if context.execution_options.get("prebuffer_rows", False):
+        # this is a bit of a hack at the moment.
+        # I would rather have some option in the result to pre-buffer
+        # internally.
+        _prebuffered = list(chunks(None))
+
+        def chunks(size):
+            return iter(_prebuffered)
+
+    result = ChunkedIteratorResult(
+        row_metadata,
+        chunks,
+        source_supports_scalars=single_entity,
+        raw=cursor,
+        dynamic_yield_per=cursor.context._is_server_side,
+    )
+
+    # filtered and single_entity are used to indicate to legacy Query that the
+    # query has ORM entities, so legacy deduping and scalars should be called
+    # on the result.
+    result._attributes = result._attributes.union(
+        dict(filtered=filtered, is_single_entity=single_entity)
+    )
+
+    # multi_row_eager_loaders OTOH is specific to joinedload.
+    if context.compile_state.multi_row_eager_loaders:
+
+        def require_unique(obj):
+            raise sa_exc.InvalidRequestError(
+                "The unique() method must be invoked on this Result, "
+                "as it contains results that include joined eager loads "
+                "against collections"
+            )
+
+        result._unique_filter_state = (None, require_unique)
+
+    if context.yield_per:
+        result.yield_per(context.yield_per)
+
+    return result
+
+
+@util.preload_module("sqlalchemy.orm.context")
+def merge_frozen_result(session, statement, frozen_result, load=True):
+    """Merge a :class:`_engine.FrozenResult` back into a :class:`_orm.Session`,
+    returning a new :class:`_engine.Result` object with :term:`persistent`
+    objects.
+
+    See the section :ref:`do_orm_execute_re_executing` for an example.
+
+    .. seealso::
+
+        :ref:`do_orm_execute_re_executing`
+
+        :meth:`_engine.Result.freeze`
+
+        :class:`_engine.FrozenResult`
+
+    """
+    querycontext = util.preloaded.orm_context
+
+    if load:
+        # flush current contents if we expect to load data
+        session._autoflush()
+
+    ctx = querycontext.ORMSelectCompileState._create_entities_collection(
+        statement, legacy=False
+    )
+
+    autoflush = session.autoflush
+    try:
+        session.autoflush = False
+        mapped_entities = [
+            i
+            for i, e in enumerate(ctx._entities)
+            if isinstance(e, querycontext._MapperEntity)
+        ]
+        keys = [ent._label_name for ent in ctx._entities]
+
+        keyed_tuple = result_tuple(
+            keys, [ent._extra_entities for ent in ctx._entities]
+        )
+
+        result = []
+        for newrow in frozen_result.rewrite_rows():
+            for i in mapped_entities:
+                if newrow[i] is not None:
+                    newrow[i] = session._merge(
+                        attributes.instance_state(newrow[i]),
+                        attributes.instance_dict(newrow[i]),
+                        load=load,
+                        _recursive={},
+                        _resolve_conflict_map={},
+                    )
+
+            result.append(keyed_tuple(newrow))
+
+        return frozen_result.with_new_rows(result)
+    finally:
+        session.autoflush = autoflush
+
+
+@util.became_legacy_20(
+    ":func:`_orm.merge_result`",
+    alternative="The function as well as the method on :class:`_orm.Query` "
+    "is superseded by the :func:`_orm.merge_frozen_result` function.",
+)
+@util.preload_module("sqlalchemy.orm.context")
+def merge_result(
+    query: Query[Any],
+    iterator: Union[FrozenResult, Iterable[Sequence[Any]], Iterable[object]],
+    load: bool = True,
+) -> Union[FrozenResult, Iterable[Any]]:
+    """Merge a result into the given :class:`.Query` object's Session.
+
+    See :meth:`_orm.Query.merge_result` for top-level documentation on this
+    function.
+
+    """
+
+    querycontext = util.preloaded.orm_context
+
+    session = query.session
+    if load:
+        # flush current contents if we expect to load data
+        session._autoflush()
+
+    # TODO: need test coverage and documentation for the FrozenResult
+    # use case.
+    if isinstance(iterator, FrozenResult):
+        frozen_result = iterator
+        iterator = iter(frozen_result.data)
+    else:
+        frozen_result = None
+
+    ctx = querycontext.ORMSelectCompileState._create_entities_collection(
+        query, legacy=True
+    )
+
+    autoflush = session.autoflush
+    try:
+        session.autoflush = False
+        single_entity = not frozen_result and len(ctx._entities) == 1
+
+        if single_entity:
+            if isinstance(ctx._entities[0], querycontext._MapperEntity):
+                result = [
+                    session._merge(
+                        attributes.instance_state(instance),
+                        attributes.instance_dict(instance),
+                        load=load,
+                        _recursive={},
+                        _resolve_conflict_map={},
+                    )
+                    for instance in iterator
+                ]
+            else:
+                result = list(iterator)
+        else:
+            mapped_entities = [
+                i
+                for i, e in enumerate(ctx._entities)
+                if isinstance(e, querycontext._MapperEntity)
+            ]
+            result = []
+            keys = [ent._label_name for ent in ctx._entities]
+
+            keyed_tuple = result_tuple(
+                keys, [ent._extra_entities for ent in ctx._entities]
+            )
+
+            for row in iterator:
+                newrow = list(row)
+                for i in mapped_entities:
+                    if newrow[i] is not None:
+                        newrow[i] = session._merge(
+                            attributes.instance_state(newrow[i]),
+                            attributes.instance_dict(newrow[i]),
+                            load=load,
+                            _recursive={},
+                            _resolve_conflict_map={},
+                        )
+                result.append(keyed_tuple(newrow))
+
+        if frozen_result:
+            return frozen_result.with_new_rows(result)
+        else:
+            return iter(result)
+    finally:
+        session.autoflush = autoflush
+
+
+def get_from_identity(
+    session: Session,
+    mapper: Mapper[_O],
+    key: _IdentityKeyType[_O],
+    passive: PassiveFlag,
+) -> Union[LoaderCallableStatus, Optional[_O]]:
+    """Look up the given key in the given session's identity map,
+    check the object for expired state if found.
+
+    """
+    instance = session.identity_map.get(key)
+    if instance is not None:
+        state = attributes.instance_state(instance)
+
+        if mapper.inherits and not state.mapper.isa(mapper):
+            return attributes.PASSIVE_CLASS_MISMATCH
+
+        # expired - ensure it still exists
+        if state.expired:
+            if not passive & attributes.SQL_OK:
+                # TODO: no coverage here
+                return attributes.PASSIVE_NO_RESULT
+            elif not passive & attributes.RELATED_OBJECT_OK:
+                # this mode is used within a flush and the instance's
+                # expired state will be checked soon enough, if necessary.
+                # also used by immediateloader for a mutually-dependent
+                # o2m->m2m load, :ticket:`6301`
+                return instance
+            try:
+                state._load_expired(state, passive)
+            except orm_exc.ObjectDeletedError:
+                session._remove_newly_deleted([state])
+                return None
+        return instance
+    else:
+        return None
+
+
+def load_on_ident(
+    session: Session,
+    statement: Union[Select, FromStatement],
+    key: Optional[_IdentityKeyType],
+    *,
+    load_options: Optional[Sequence[ORMOption]] = None,
+    refresh_state: Optional[InstanceState[Any]] = None,
+    with_for_update: Optional[ForUpdateArg] = None,
+    only_load_props: Optional[Iterable[str]] = None,
+    no_autoflush: bool = False,
+    bind_arguments: Mapping[str, Any] = util.EMPTY_DICT,
+    execution_options: _ExecuteOptions = util.EMPTY_DICT,
+    require_pk_cols: bool = False,
+    is_user_refresh: bool = False,
+):
+    """Load the given identity key from the database."""
+    if key is not None:
+        ident = key[1]
+        identity_token = key[2]
+    else:
+        ident = identity_token = None
+
+    return load_on_pk_identity(
+        session,
+        statement,
+        ident,
+        load_options=load_options,
+        refresh_state=refresh_state,
+        with_for_update=with_for_update,
+        only_load_props=only_load_props,
+        identity_token=identity_token,
+        no_autoflush=no_autoflush,
+        bind_arguments=bind_arguments,
+        execution_options=execution_options,
+        require_pk_cols=require_pk_cols,
+        is_user_refresh=is_user_refresh,
+    )
+
+
+def load_on_pk_identity(
+    session: Session,
+    statement: Union[Select, FromStatement],
+    primary_key_identity: Optional[Tuple[Any, ...]],
+    *,
+    load_options: Optional[Sequence[ORMOption]] = None,
+    refresh_state: Optional[InstanceState[Any]] = None,
+    with_for_update: Optional[ForUpdateArg] = None,
+    only_load_props: Optional[Iterable[str]] = None,
+    identity_token: Optional[Any] = None,
+    no_autoflush: bool = False,
+    bind_arguments: Mapping[str, Any] = util.EMPTY_DICT,
+    execution_options: _ExecuteOptions = util.EMPTY_DICT,
+    require_pk_cols: bool = False,
+    is_user_refresh: bool = False,
+):
+    """Load the given primary key identity from the database."""
+
+    query = statement
+    q = query._clone()
+
+    assert not q._is_lambda_element
+
+    if load_options is None:
+        load_options = QueryContext.default_load_options
+
+    if (
+        statement._compile_options
+        is SelectState.default_select_compile_options
+    ):
+        compile_options = ORMCompileState.default_compile_options
+    else:
+        compile_options = statement._compile_options
+
+    if primary_key_identity is not None:
+        mapper = query._propagate_attrs["plugin_subject"]
+
+        (_get_clause, _get_params) = mapper._get_clause
+
+        # None present in ident - turn those comparisons
+        # into "IS NULL"
+        if None in primary_key_identity:
+            nones = {
+                _get_params[col].key
+                for col, value in zip(mapper.primary_key, primary_key_identity)
+                if value is None
+            }
+
+            _get_clause = sql_util.adapt_criterion_to_null(_get_clause, nones)
+
+            if len(nones) == len(primary_key_identity):
+                util.warn(
+                    "fully NULL primary key identity cannot load any "
+                    "object.  This condition may raise an error in a future "
+                    "release."
+                )
+
+        q._where_criteria = (
+            sql_util._deep_annotate(_get_clause, {"_orm_adapt": True}),
+        )
+
+        params = {
+            _get_params[primary_key].key: id_val
+            for id_val, primary_key in zip(
+                primary_key_identity, mapper.primary_key
+            )
+        }
+    else:
+        params = None
+
+    if with_for_update is not None:
+        version_check = True
+        q._for_update_arg = with_for_update
+    elif query._for_update_arg is not None:
+        version_check = True
+        q._for_update_arg = query._for_update_arg
+    else:
+        version_check = False
+
+    if require_pk_cols and only_load_props:
+        if not refresh_state:
+            raise sa_exc.ArgumentError(
+                "refresh_state is required when require_pk_cols is present"
+            )
+
+        refresh_state_prokeys = refresh_state.mapper._primary_key_propkeys
+        has_changes = {
+            key
+            for key in refresh_state_prokeys.difference(only_load_props)
+            if refresh_state.attrs[key].history.has_changes()
+        }
+        if has_changes:
+            # raise if pending pk changes are present.
+            # technically, this could be limited to the case where we have
+            # relationships in the only_load_props collection to be refreshed
+            # also (and only ones that have a secondary eager loader, at that).
+            # however, the error is in place across the board so that behavior
+            # here is easier to predict.   The use case it prevents is one
+            # of mutating PK attrs, leaving them unflushed,
+            # calling session.refresh(), and expecting those attrs to remain
+            # still unflushed.   It seems likely someone doing all those
+            # things would be better off having the PK attributes flushed
+            # to the database before tinkering like that (session.refresh() is
+            # tinkering).
+            raise sa_exc.InvalidRequestError(
+                f"Please flush pending primary key changes on "
+                "attributes "
+                f"{has_changes} for mapper {refresh_state.mapper} before "
+                "proceeding with a refresh"
+            )
+
+        # overall, the ORM has no internal flow right now for "dont load the
+        # primary row of an object at all, but fire off
+        # selectinload/subqueryload/immediateload for some relationships".
+        # It would probably be a pretty big effort to add such a flow.  So
+        # here, the case for #8703 is introduced; user asks to refresh some
+        # relationship attributes only which are
+        # selectinload/subqueryload/immediateload/ etc. (not joinedload).
+        # ORM complains there's no columns in the primary row to load.
+        # So here, we just add the PK cols if that
+        # case is detected, so that there is a SELECT emitted for the primary
+        # row.
+        #
+        # Let's just state right up front, for this one little case,
+        # the ORM here is adding a whole extra SELECT just to satisfy
+        # limitations in the internal flow.  This is really not a thing
+        # SQLAlchemy finds itself doing like, ever, obviously, we are
+        # constantly working to *remove* SELECTs we don't need.   We
+        # rationalize this for now based on 1. session.refresh() is not
+        # commonly used 2. session.refresh() with only relationship attrs is
+        # even less commonly used 3. the SELECT in question is very low
+        # latency.
+        #
+        # to add the flow to not include the SELECT, the quickest way
+        # might be to just manufacture a single-row result set to send off to
+        # instances(), but we'd have to weave that into context.py and all
+        # that.  For 2.0.0, we have enough big changes to navigate for now.
+        #
+        mp = refresh_state.mapper._props
+        for p in only_load_props:
+            if mp[p]._is_relationship:
+                only_load_props = refresh_state_prokeys.union(only_load_props)
+                break
+
+    if refresh_state and refresh_state.load_options:
+        compile_options += {"_current_path": refresh_state.load_path.parent}
+        q = q.options(*refresh_state.load_options)
+
+    new_compile_options, load_options = _set_get_options(
+        compile_options,
+        load_options,
+        version_check=version_check,
+        only_load_props=only_load_props,
+        refresh_state=refresh_state,
+        identity_token=identity_token,
+        is_user_refresh=is_user_refresh,
+    )
+
+    q._compile_options = new_compile_options
+    q._order_by = None
+
+    if no_autoflush:
+        load_options += {"_autoflush": False}
+
+    execution_options = util.EMPTY_DICT.merge_with(
+        execution_options, {"_sa_orm_load_options": load_options}
+    )
+    result = (
+        session.execute(
+            q,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+        .unique()
+        .scalars()
+    )
+
+    try:
+        return result.one()
+    except orm_exc.NoResultFound:
+        return None
+
+
+def _set_get_options(
+    compile_opt,
+    load_opt,
+    populate_existing=None,
+    version_check=None,
+    only_load_props=None,
+    refresh_state=None,
+    identity_token=None,
+    is_user_refresh=None,
+):
+    compile_options = {}
+    load_options = {}
+    if version_check:
+        load_options["_version_check"] = version_check
+    if populate_existing:
+        load_options["_populate_existing"] = populate_existing
+    if refresh_state:
+        load_options["_refresh_state"] = refresh_state
+        compile_options["_for_refresh_state"] = True
+    if only_load_props:
+        compile_options["_only_load_props"] = frozenset(only_load_props)
+    if identity_token:
+        load_options["_identity_token"] = identity_token
+
+    if is_user_refresh:
+        load_options["_is_user_refresh"] = is_user_refresh
+    if load_options:
+        load_opt += load_options
+    if compile_options:
+        compile_opt += compile_options
+
+    return compile_opt, load_opt
+
+
+def _setup_entity_query(
+    compile_state,
+    mapper,
+    query_entity,
+    path,
+    adapter,
+    column_collection,
+    with_polymorphic=None,
+    only_load_props=None,
+    polymorphic_discriminator=None,
+    **kw,
+):
+    if with_polymorphic:
+        poly_properties = mapper._iterate_polymorphic_properties(
+            with_polymorphic
+        )
+    else:
+        poly_properties = mapper._polymorphic_properties
+
+    quick_populators = {}
+
+    path.set(compile_state.attributes, "memoized_setups", quick_populators)
+
+    # for the lead entities in the path, e.g. not eager loads, and
+    # assuming a user-passed aliased class, e.g. not a from_self() or any
+    # implicit aliasing, don't add columns to the SELECT that aren't
+    # in the thing that's aliased.
+    check_for_adapt = adapter and len(path) == 1 and path[-1].is_aliased_class
+
+    for value in poly_properties:
+        if only_load_props and value.key not in only_load_props:
+            continue
+        value.setup(
+            compile_state,
+            query_entity,
+            path,
+            adapter,
+            only_load_props=only_load_props,
+            column_collection=column_collection,
+            memoized_populators=quick_populators,
+            check_for_adapt=check_for_adapt,
+            **kw,
+        )
+
+    if (
+        polymorphic_discriminator is not None
+        and polymorphic_discriminator is not mapper.polymorphic_on
+    ):
+        if adapter:
+            pd = adapter.columns[polymorphic_discriminator]
+        else:
+            pd = polymorphic_discriminator
+        column_collection.append(pd)
+
+
+def _warn_for_runid_changed(state):
+    util.warn(
+        "Loading context for %s has changed within a load/refresh "
+        "handler, suggesting a row refresh operation took place. If this "
+        "event handler is expected to be "
+        "emitting row refresh operations within an existing load or refresh "
+        "operation, set restore_load_context=True when establishing the "
+        "listener to ensure the context remains unchanged when the event "
+        "handler completes." % (state_str(state),)
+    )
+
+
+def _instance_processor(
+    query_entity,
+    mapper,
+    context,
+    result,
+    path,
+    adapter,
+    only_load_props=None,
+    refresh_state=None,
+    polymorphic_discriminator=None,
+    _polymorphic_from=None,
+):
+    """Produce a mapper level row processor callable
+    which processes rows into mapped instances."""
+
+    # note that this method, most of which exists in a closure
+    # called _instance(), resists being broken out, as
+    # attempts to do so tend to add significant function
+    # call overhead.  _instance() is the most
+    # performance-critical section in the whole ORM.
+
+    identity_class = mapper._identity_class
+    compile_state = context.compile_state
+
+    # look for "row getter" functions that have been assigned along
+    # with the compile state that were cached from a previous load.
+    # these are operator.itemgetter() objects that each will extract a
+    # particular column from each row.
+
+    getter_key = ("getters", mapper)
+    getters = path.get(compile_state.attributes, getter_key, None)
+
+    if getters is None:
+        # no getters, so go through a list of attributes we are loading for,
+        # and the ones that are column based will have already put information
+        # for us in another collection "memoized_setups", which represents the
+        # output of the LoaderStrategy.setup_query() method.  We can just as
+        # easily call LoaderStrategy.create_row_processor for each, but by
+        # getting it all at once from setup_query we save another method call
+        # per attribute.
+        props = mapper._prop_set
+        if only_load_props is not None:
+            props = props.intersection(
+                mapper._props[k] for k in only_load_props
+            )
+
+        quick_populators = path.get(
+            context.attributes, "memoized_setups", EMPTY_DICT
+        )
+
+        todo = []
+        cached_populators = {
+            "new": [],
+            "quick": [],
+            "deferred": [],
+            "expire": [],
+            "existing": [],
+            "eager": [],
+        }
+
+        if refresh_state is None:
+            # we can also get the "primary key" tuple getter function
+            pk_cols = mapper.primary_key
+
+            if adapter:
+                pk_cols = [adapter.columns[c] for c in pk_cols]
+            primary_key_getter = result._tuple_getter(pk_cols)
+        else:
+            primary_key_getter = None
+
+        getters = {
+            "cached_populators": cached_populators,
+            "todo": todo,
+            "primary_key_getter": primary_key_getter,
+        }
+        for prop in props:
+            if prop in quick_populators:
+                # this is an inlined path just for column-based attributes.
+                col = quick_populators[prop]
+                if col is _DEFER_FOR_STATE:
+                    cached_populators["new"].append(
+                        (prop.key, prop._deferred_column_loader)
+                    )
+                elif col is _SET_DEFERRED_EXPIRED:
+                    # note that in this path, we are no longer
+                    # searching in the result to see if the column might
+                    # be present in some unexpected way.
+                    cached_populators["expire"].append((prop.key, False))
+                elif col is _RAISE_FOR_STATE:
+                    cached_populators["new"].append(
+                        (prop.key, prop._raise_column_loader)
+                    )
+                else:
+                    getter = None
+                    if adapter:
+                        # this logic had been removed for all 1.4 releases
+                        # up until 1.4.18; the adapter here is particularly
+                        # the compound eager adapter which isn't accommodated
+                        # in the quick_populators right now.  The "fallback"
+                        # logic below instead took over in many more cases
+                        # until issue #6596 was identified.
+
+                        # note there is still an issue where this codepath
+                        # produces no "getter" for cases where a joined-inh
+                        # mapping includes a labeled column property, meaning
+                        # KeyError is caught internally and we fall back to
+                        # _getter(col), which works anyway.   The adapter
+                        # here for joined inh without any aliasing might not
+                        # be useful.  Tests which see this include
+                        # test.orm.inheritance.test_basic ->
+                        # EagerTargetingTest.test_adapt_stringency
+                        # OptimizedLoadTest.test_column_expression_joined
+                        # PolymorphicOnNotLocalTest.test_polymorphic_on_column_prop  # noqa: E501
+                        #
+
+                        adapted_col = adapter.columns[col]
+                        if adapted_col is not None:
+                            getter = result._getter(adapted_col, False)
+                    if not getter:
+                        getter = result._getter(col, False)
+                    if getter:
+                        cached_populators["quick"].append((prop.key, getter))
+                    else:
+                        # fall back to the ColumnProperty itself, which
+                        # will iterate through all of its columns
+                        # to see if one fits
+                        prop.create_row_processor(
+                            context,
+                            query_entity,
+                            path,
+                            mapper,
+                            result,
+                            adapter,
+                            cached_populators,
+                        )
+            else:
+                # loader strategies like subqueryload, selectinload,
+                # joinedload, basically relationships, these need to interact
+                # with the context each time to work correctly.
+                todo.append(prop)
+
+        path.set(compile_state.attributes, getter_key, getters)
+
+    cached_populators = getters["cached_populators"]
+
+    populators = {key: list(value) for key, value in cached_populators.items()}
+    for prop in getters["todo"]:
+        prop.create_row_processor(
+            context, query_entity, path, mapper, result, adapter, populators
+        )
+
+    propagated_loader_options = context.propagated_loader_options
+    load_path = (
+        context.compile_state.current_path + path
+        if context.compile_state.current_path.path
+        else path
+    )
+
+    session_identity_map = context.session.identity_map
+
+    populate_existing = context.populate_existing or mapper.always_refresh
+    load_evt = bool(mapper.class_manager.dispatch.load)
+    refresh_evt = bool(mapper.class_manager.dispatch.refresh)
+    persistent_evt = bool(context.session.dispatch.loaded_as_persistent)
+    if persistent_evt:
+        loaded_as_persistent = context.session.dispatch.loaded_as_persistent
+    instance_state = attributes.instance_state
+    instance_dict = attributes.instance_dict
+    session_id = context.session.hash_key
+    runid = context.runid
+    identity_token = context.identity_token
+
+    version_check = context.version_check
+    if version_check:
+        version_id_col = mapper.version_id_col
+        if version_id_col is not None:
+            if adapter:
+                version_id_col = adapter.columns[version_id_col]
+            version_id_getter = result._getter(version_id_col)
+        else:
+            version_id_getter = None
+
+    if not refresh_state and _polymorphic_from is not None:
+        key = ("loader", path.path)
+
+        if key in context.attributes and context.attributes[key].strategy == (
+            ("selectinload_polymorphic", True),
+        ):
+            option_entities = context.attributes[key].local_opts["entities"]
+        else:
+            option_entities = None
+        selectin_load_via = mapper._should_selectin_load(
+            option_entities,
+            _polymorphic_from,
+        )
+
+        if selectin_load_via and selectin_load_via is not _polymorphic_from:
+            # only_load_props goes w/ refresh_state only, and in a refresh
+            # we are a single row query for the exact entity; polymorphic
+            # loading does not apply
+            assert only_load_props is None
+
+            if selectin_load_via.is_mapper:
+                _load_supers = []
+                _endmost_mapper = selectin_load_via
+                while (
+                    _endmost_mapper
+                    and _endmost_mapper is not _polymorphic_from
+                ):
+                    _load_supers.append(_endmost_mapper)
+                    _endmost_mapper = _endmost_mapper.inherits
+            else:
+                _load_supers = [selectin_load_via]
+
+            for _selectinload_entity in _load_supers:
+                if PostLoad.path_exists(
+                    context, load_path, _selectinload_entity
+                ):
+                    continue
+                callable_ = _load_subclass_via_in(
+                    context,
+                    path,
+                    _selectinload_entity,
+                    _polymorphic_from,
+                    option_entities,
+                )
+                PostLoad.callable_for_path(
+                    context,
+                    load_path,
+                    _selectinload_entity.mapper,
+                    _selectinload_entity,
+                    callable_,
+                    _selectinload_entity,
+                )
+
+    post_load = PostLoad.for_context(context, load_path, only_load_props)
+
+    if refresh_state:
+        refresh_identity_key = refresh_state.key
+        if refresh_identity_key is None:
+            # super-rare condition; a refresh is being called
+            # on a non-instance-key instance; this is meant to only
+            # occur within a flush()
+            refresh_identity_key = mapper._identity_key_from_state(
+                refresh_state
+            )
+    else:
+        refresh_identity_key = None
+
+        primary_key_getter = getters["primary_key_getter"]
+
+    if mapper.allow_partial_pks:
+        is_not_primary_key = _none_set.issuperset
+    else:
+        is_not_primary_key = _none_set.intersection
+
+    def _instance(row):
+        # determine the state that we'll be populating
+        if refresh_identity_key:
+            # fixed state that we're refreshing
+            state = refresh_state
+            instance = state.obj()
+            dict_ = instance_dict(instance)
+            isnew = state.runid != runid
+            currentload = True
+            loaded_instance = False
+        else:
+            # look at the row, see if that identity is in the
+            # session, or we have to create a new one
+            identitykey = (
+                identity_class,
+                primary_key_getter(row),
+                identity_token,
+            )
+
+            instance = session_identity_map.get(identitykey)
+
+            if instance is not None:
+                # existing instance
+                state = instance_state(instance)
+                dict_ = instance_dict(instance)
+
+                isnew = state.runid != runid
+                currentload = not isnew
+                loaded_instance = False
+
+                if version_check and version_id_getter and not currentload:
+                    _validate_version_id(
+                        mapper, state, dict_, row, version_id_getter
+                    )
+
+            else:
+                # create a new instance
+
+                # check for non-NULL values in the primary key columns,
+                # else no entity is returned for the row
+                if is_not_primary_key(identitykey[1]):
+                    return None
+
+                isnew = True
+                currentload = True
+                loaded_instance = True
+
+                instance = mapper.class_manager.new_instance()
+
+                dict_ = instance_dict(instance)
+                state = instance_state(instance)
+                state.key = identitykey
+                state.identity_token = identity_token
+
+                # attach instance to session.
+                state.session_id = session_id
+                session_identity_map._add_unpresent(state, identitykey)
+
+        effective_populate_existing = populate_existing
+        if refresh_state is state:
+            effective_populate_existing = True
+
+        # populate.  this looks at whether this state is new
+        # for this load or was existing, and whether or not this
+        # row is the first row with this identity.
+        if currentload or effective_populate_existing:
+            # full population routines.  Objects here are either
+            # just created, or we are doing a populate_existing
+
+            # be conservative about setting load_path when populate_existing
+            # is in effect; want to maintain options from the original
+            # load.  see test_expire->test_refresh_maintains_deferred_options
+            if isnew and (
+                propagated_loader_options or not effective_populate_existing
+            ):
+                state.load_options = propagated_loader_options
+                state.load_path = load_path
+
+            _populate_full(
+                context,
+                row,
+                state,
+                dict_,
+                isnew,
+                load_path,
+                loaded_instance,
+                effective_populate_existing,
+                populators,
+            )
+
+            if isnew:
+                # state.runid should be equal to context.runid / runid
+                # here, however for event checks we are being more conservative
+                # and checking against existing run id
+                # assert state.runid == runid
+
+                existing_runid = state.runid
+
+                if loaded_instance:
+                    if load_evt:
+                        state.manager.dispatch.load(state, context)
+                        if state.runid != existing_runid:
+                            _warn_for_runid_changed(state)
+                    if persistent_evt:
+                        loaded_as_persistent(context.session, state)
+                        if state.runid != existing_runid:
+                            _warn_for_runid_changed(state)
+                elif refresh_evt:
+                    state.manager.dispatch.refresh(
+                        state, context, only_load_props
+                    )
+                    if state.runid != runid:
+                        _warn_for_runid_changed(state)
+
+                if effective_populate_existing or state.modified:
+                    if refresh_state and only_load_props:
+                        state._commit(dict_, only_load_props)
+                    else:
+                        state._commit_all(dict_, session_identity_map)
+
+            if post_load:
+                post_load.add_state(state, True)
+
+        else:
+            # partial population routines, for objects that were already
+            # in the Session, but a row matches them; apply eager loaders
+            # on existing objects, etc.
+            unloaded = state.unloaded
+            isnew = state not in context.partials
+
+            if not isnew or unloaded or populators["eager"]:
+                # state is having a partial set of its attributes
+                # refreshed.  Populate those attributes,
+                # and add to the "context.partials" collection.
+
+                to_load = _populate_partial(
+                    context,
+                    row,
+                    state,
+                    dict_,
+                    isnew,
+                    load_path,
+                    unloaded,
+                    populators,
+                )
+
+                if isnew:
+                    if refresh_evt:
+                        existing_runid = state.runid
+                        state.manager.dispatch.refresh(state, context, to_load)
+                        if state.runid != existing_runid:
+                            _warn_for_runid_changed(state)
+
+                    state._commit(dict_, to_load)
+
+            if post_load and context.invoke_all_eagers:
+                post_load.add_state(state, False)
+
+        return instance
+
+    if mapper.polymorphic_map and not _polymorphic_from and not refresh_state:
+        # if we are doing polymorphic, dispatch to a different _instance()
+        # method specific to the subclass mapper
+        def ensure_no_pk(row):
+            identitykey = (
+                identity_class,
+                primary_key_getter(row),
+                identity_token,
+            )
+            if not is_not_primary_key(identitykey[1]):
+                return identitykey
+            else:
+                return None
+
+        _instance = _decorate_polymorphic_switch(
+            _instance,
+            context,
+            query_entity,
+            mapper,
+            result,
+            path,
+            polymorphic_discriminator,
+            adapter,
+            ensure_no_pk,
+        )
+
+    return _instance
+
+
+def _load_subclass_via_in(
+    context, path, entity, polymorphic_from, option_entities
+):
+    mapper = entity.mapper
+
+    # TODO: polymorphic_from seems to be a Mapper in all cases.
+    # this is likely not needed, but as we dont have typing in loading.py
+    # yet, err on the safe side
+    polymorphic_from_mapper = polymorphic_from.mapper
+    not_against_basemost = polymorphic_from_mapper.inherits is not None
+
+    zero_idx = len(mapper.base_mapper.primary_key) == 1
+
+    if entity.is_aliased_class or not_against_basemost:
+        q, enable_opt, disable_opt = mapper._subclass_load_via_in(
+            entity, polymorphic_from
+        )
+    else:
+        q, enable_opt, disable_opt = mapper._subclass_load_via_in_mapper
+
+    def do_load(context, path, states, load_only, effective_entity):
+        if not option_entities:
+            # filter out states for those that would have selectinloaded
+            # from another loader
+            # TODO: we are currently ignoring the case where the
+            # "selectin_polymorphic" option is used, as this is much more
+            # complex / specific / very uncommon API use
+            states = [
+                (s, v)
+                for s, v in states
+                if s.mapper._would_selectin_load_only_from_given_mapper(mapper)
+            ]
+
+            if not states:
+                return
+
+        orig_query = context.query
+
+        if path.parent:
+            enable_opt_lcl = enable_opt._prepend_path(path)
+            disable_opt_lcl = disable_opt._prepend_path(path)
+        else:
+            enable_opt_lcl = enable_opt
+            disable_opt_lcl = disable_opt
+        options = (
+            (enable_opt_lcl,) + orig_query._with_options + (disable_opt_lcl,)
+        )
+
+        q2 = q.options(*options)
+
+        q2._compile_options = context.compile_state.default_compile_options
+        q2._compile_options += {"_current_path": path.parent}
+
+        if context.populate_existing:
+            q2 = q2.execution_options(populate_existing=True)
+
+        context.session.execute(
+            q2,
+            dict(
+                primary_keys=[
+                    state.key[1][0] if zero_idx else state.key[1]
+                    for state, load_attrs in states
+                ]
+            ),
+        ).unique().scalars().all()
+
+    return do_load
+
+
+def _populate_full(
+    context,
+    row,
+    state,
+    dict_,
+    isnew,
+    load_path,
+    loaded_instance,
+    populate_existing,
+    populators,
+):
+    if isnew:
+        # first time we are seeing a row with this identity.
+        state.runid = context.runid
+
+        for key, getter in populators["quick"]:
+            dict_[key] = getter(row)
+        if populate_existing:
+            for key, set_callable in populators["expire"]:
+                dict_.pop(key, None)
+                if set_callable:
+                    state.expired_attributes.add(key)
+        else:
+            for key, set_callable in populators["expire"]:
+                if set_callable:
+                    state.expired_attributes.add(key)
+
+        for key, populator in populators["new"]:
+            populator(state, dict_, row)
+
+    elif load_path != state.load_path:
+        # new load path, e.g. object is present in more than one
+        # column position in a series of rows
+        state.load_path = load_path
+
+        # if we have data, and the data isn't in the dict, OK, let's put
+        # it in.
+        for key, getter in populators["quick"]:
+            if key not in dict_:
+                dict_[key] = getter(row)
+
+        # otherwise treat like an "already seen" row
+        for key, populator in populators["existing"]:
+            populator(state, dict_, row)
+            # TODO:  allow "existing" populator to know this is
+            # a new path for the state:
+            # populator(state, dict_, row, new_path=True)
+
+    else:
+        # have already seen rows with this identity in this same path.
+        for key, populator in populators["existing"]:
+            populator(state, dict_, row)
+
+            # TODO: same path
+            # populator(state, dict_, row, new_path=False)
+
+
+def _populate_partial(
+    context, row, state, dict_, isnew, load_path, unloaded, populators
+):
+    if not isnew:
+        if unloaded:
+            # extra pass, see #8166
+            for key, getter in populators["quick"]:
+                if key in unloaded:
+                    dict_[key] = getter(row)
+
+        to_load = context.partials[state]
+        for key, populator in populators["existing"]:
+            if key in to_load:
+                populator(state, dict_, row)
+    else:
+        to_load = unloaded
+        context.partials[state] = to_load
+
+        for key, getter in populators["quick"]:
+            if key in to_load:
+                dict_[key] = getter(row)
+        for key, set_callable in populators["expire"]:
+            if key in to_load:
+                dict_.pop(key, None)
+                if set_callable:
+                    state.expired_attributes.add(key)
+        for key, populator in populators["new"]:
+            if key in to_load:
+                populator(state, dict_, row)
+
+    for key, populator in populators["eager"]:
+        if key not in unloaded:
+            populator(state, dict_, row)
+
+    return to_load
+
+
+def _validate_version_id(mapper, state, dict_, row, getter):
+    if mapper._get_state_attr_by_column(
+        state, dict_, mapper.version_id_col
+    ) != getter(row):
+        raise orm_exc.StaleDataError(
+            "Instance '%s' has version id '%s' which "
+            "does not match database-loaded version id '%s'."
+            % (
+                state_str(state),
+                mapper._get_state_attr_by_column(
+                    state, dict_, mapper.version_id_col
+                ),
+                getter(row),
+            )
+        )
+
+
+def _decorate_polymorphic_switch(
+    instance_fn,
+    context,
+    query_entity,
+    mapper,
+    result,
+    path,
+    polymorphic_discriminator,
+    adapter,
+    ensure_no_pk,
+):
+    if polymorphic_discriminator is not None:
+        polymorphic_on = polymorphic_discriminator
+    else:
+        polymorphic_on = mapper.polymorphic_on
+    if polymorphic_on is None:
+        return instance_fn
+
+    if adapter:
+        polymorphic_on = adapter.columns[polymorphic_on]
+
+    def configure_subclass_mapper(discriminator):
+        try:
+            sub_mapper = mapper.polymorphic_map[discriminator]
+        except KeyError:
+            raise AssertionError(
+                "No such polymorphic_identity %r is defined" % discriminator
+            )
+        else:
+            if sub_mapper is mapper:
+                return None
+            elif not sub_mapper.isa(mapper):
+                return False
+
+            return _instance_processor(
+                query_entity,
+                sub_mapper,
+                context,
+                result,
+                path,
+                adapter,
+                _polymorphic_from=mapper,
+            )
+
+    polymorphic_instances = util.PopulateDict(configure_subclass_mapper)
+
+    getter = result._getter(polymorphic_on)
+
+    def polymorphic_instance(row):
+        discriminator = getter(row)
+        if discriminator is not None:
+            _instance = polymorphic_instances[discriminator]
+            if _instance:
+                return _instance(row)
+            elif _instance is False:
+                identitykey = ensure_no_pk(row)
+
+                if identitykey:
+                    raise sa_exc.InvalidRequestError(
+                        "Row with identity key %s can't be loaded into an "
+                        "object; the polymorphic discriminator column '%s' "
+                        "refers to %s, which is not a sub-mapper of "
+                        "the requested %s"
+                        % (
+                            identitykey,
+                            polymorphic_on,
+                            mapper.polymorphic_map[discriminator],
+                            mapper,
+                        )
+                    )
+                else:
+                    return None
+            else:
+                return instance_fn(row)
+        else:
+            identitykey = ensure_no_pk(row)
+
+            if identitykey:
+                raise sa_exc.InvalidRequestError(
+                    "Row with identity key %s can't be loaded into an "
+                    "object; the polymorphic discriminator column '%s' is "
+                    "NULL" % (identitykey, polymorphic_on)
+                )
+            else:
+                return None
+
+    return polymorphic_instance
+
+
+class PostLoad:
+    """Track loaders and states for "post load" operations."""
+
+    __slots__ = "loaders", "states", "load_keys"
+
+    def __init__(self):
+        self.loaders = {}
+        self.states = util.OrderedDict()
+        self.load_keys = None
+
+    def add_state(self, state, overwrite):
+        # the states for a polymorphic load here are all shared
+        # within a single PostLoad object among multiple subtypes.
+        # Filtering of callables on a per-subclass basis needs to be done at
+        # the invocation level
+        self.states[state] = overwrite
+
+    def invoke(self, context, path):
+        if not self.states:
+            return
+        path = path_registry.PathRegistry.coerce(path)
+        for (
+            effective_context,
+            token,
+            limit_to_mapper,
+            loader,
+            arg,
+            kw,
+        ) in self.loaders.values():
+            states = [
+                (state, overwrite)
+                for state, overwrite in self.states.items()
+                if state.manager.mapper.isa(limit_to_mapper)
+            ]
+            if states:
+                loader(
+                    effective_context, path, states, self.load_keys, *arg, **kw
+                )
+        self.states.clear()
+
+    @classmethod
+    def for_context(cls, context, path, only_load_props):
+        pl = context.post_load_paths.get(path.path)
+        if pl is not None and only_load_props:
+            pl.load_keys = only_load_props
+        return pl
+
+    @classmethod
+    def path_exists(self, context, path, key):
+        return (
+            path.path in context.post_load_paths
+            and key in context.post_load_paths[path.path].loaders
+        )
+
+    @classmethod
+    def callable_for_path(
+        cls, context, path, limit_to_mapper, token, loader_callable, *arg, **kw
+    ):
+        if path.path in context.post_load_paths:
+            pl = context.post_load_paths[path.path]
+        else:
+            pl = context.post_load_paths[path.path] = PostLoad()
+        pl.loaders[token] = (
+            context,
+            token,
+            limit_to_mapper,
+            loader_callable,
+            arg,
+            kw,
+        )
+
+
+def load_scalar_attributes(mapper, state, attribute_names, passive):
+    """initiate a column-based attribute refresh operation."""
+
+    # assert mapper is _state_mapper(state)
+    session = state.session
+    if not session:
+        raise orm_exc.DetachedInstanceError(
+            "Instance %s is not bound to a Session; "
+            "attribute refresh operation cannot proceed" % (state_str(state))
+        )
+
+    no_autoflush = bool(passive & attributes.NO_AUTOFLUSH)
+
+    # in the case of inheritance, particularly concrete and abstract
+    # concrete inheritance, the class manager might have some keys
+    # of attributes on the superclass that we didn't actually map.
+    # These could be mapped as "concrete, don't load" or could be completely
+    # excluded from the mapping and we know nothing about them.  Filter them
+    # here to prevent them from coming through.
+    if attribute_names:
+        attribute_names = attribute_names.intersection(mapper.attrs.keys())
+
+    if mapper.inherits and not mapper.concrete:
+        # load based on committed attributes in the object, formed into
+        # a truncated SELECT that only includes relevant tables.  does not
+        # currently use state.key
+        statement = mapper._optimized_get_statement(state, attribute_names)
+        if statement is not None:
+            # undefer() isn't needed here because statement has the
+            # columns needed already, this implicitly undefers that column
+            stmt = FromStatement(mapper, statement)
+
+            return load_on_ident(
+                session,
+                stmt,
+                None,
+                only_load_props=attribute_names,
+                refresh_state=state,
+                no_autoflush=no_autoflush,
+            )
+
+    # normal load, use state.key as the identity to SELECT
+    has_key = bool(state.key)
+
+    if has_key:
+        identity_key = state.key
+    else:
+        # this codepath is rare - only valid when inside a flush, and the
+        # object is becoming persistent but hasn't yet been assigned
+        # an identity_key.
+        # check here to ensure we have the attrs we need.
+        pk_attrs = [
+            mapper._columntoproperty[col].key for col in mapper.primary_key
+        ]
+        if state.expired_attributes.intersection(pk_attrs):
+            raise sa_exc.InvalidRequestError(
+                "Instance %s cannot be refreshed - it's not "
+                " persistent and does not "
+                "contain a full primary key." % state_str(state)
+            )
+        identity_key = mapper._identity_key_from_state(state)
+
+    if (
+        _none_set.issubset(identity_key) and not mapper.allow_partial_pks
+    ) or _none_set.issuperset(identity_key):
+        util.warn_limited(
+            "Instance %s to be refreshed doesn't "
+            "contain a full primary key - can't be refreshed "
+            "(and shouldn't be expired, either).",
+            state_str(state),
+        )
+        return
+
+    result = load_on_ident(
+        session,
+        select(mapper).set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL),
+        identity_key,
+        refresh_state=state,
+        only_load_props=attribute_names,
+        no_autoflush=no_autoflush,
+    )
+
+    # if instance is pending, a refresh operation
+    # may not complete (even if PK attributes are assigned)
+    if has_key and result is None:
+        raise orm_exc.ObjectDeletedError(state)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapped_collection.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapped_collection.py
new file mode 100644
index 00000000..ca085c40
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapped_collection.py
@@ -0,0 +1,557 @@
+# orm/mapped_collection.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import operator
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import base
+from .collections import collection
+from .collections import collection_adapter
+from .. import exc as sa_exc
+from .. import util
+from ..sql import coercions
+from ..sql import expression
+from ..sql import roles
+from ..util.langhelpers import Missing
+from ..util.langhelpers import MissingOr
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from . import AttributeEventToken
+    from . import Mapper
+    from .collections import CollectionAdapter
+    from ..sql.elements import ColumnElement
+
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+
+
+class _PlainColumnGetter(Generic[_KT]):
+    """Plain column getter, stores collection of Column objects
+    directly.
+
+    Serializes to a :class:`._SerializableColumnGetterV2`
+    which has more expensive __call__() performance
+    and some rare caveats.
+
+    """
+
+    __slots__ = ("cols", "composite")
+
+    def __init__(self, cols: Sequence[ColumnElement[_KT]]) -> None:
+        self.cols = cols
+        self.composite = len(cols) > 1
+
+    def __reduce__(
+        self,
+    ) -> Tuple[
+        Type[_SerializableColumnGetterV2[_KT]],
+        Tuple[Sequence[Tuple[Optional[str], Optional[str]]]],
+    ]:
+        return _SerializableColumnGetterV2._reduce_from_cols(self.cols)
+
+    def _cols(self, mapper: Mapper[_KT]) -> Sequence[ColumnElement[_KT]]:
+        return self.cols
+
+    def __call__(self, value: _KT) -> MissingOr[Union[_KT, Tuple[_KT, ...]]]:
+        state = base.instance_state(value)
+        m = base._state_mapper(state)
+
+        key: List[_KT] = [
+            m._get_state_attr_by_column(state, state.dict, col)
+            for col in self._cols(m)
+        ]
+        if self.composite:
+            return tuple(key)
+        else:
+            obj = key[0]
+            if obj is None:
+                return Missing
+            else:
+                return obj
+
+
+class _SerializableColumnGetterV2(_PlainColumnGetter[_KT]):
+    """Updated serializable getter which deals with
+    multi-table mapped classes.
+
+    Two extremely unusual cases are not supported.
+    Mappings which have tables across multiple metadata
+    objects, or which are mapped to non-Table selectables
+    linked across inheriting mappers may fail to function
+    here.
+
+    """
+
+    __slots__ = ("colkeys",)
+
+    def __init__(
+        self, colkeys: Sequence[Tuple[Optional[str], Optional[str]]]
+    ) -> None:
+        self.colkeys = colkeys
+        self.composite = len(colkeys) > 1
+
+    def __reduce__(
+        self,
+    ) -> Tuple[
+        Type[_SerializableColumnGetterV2[_KT]],
+        Tuple[Sequence[Tuple[Optional[str], Optional[str]]]],
+    ]:
+        return self.__class__, (self.colkeys,)
+
+    @classmethod
+    def _reduce_from_cols(cls, cols: Sequence[ColumnElement[_KT]]) -> Tuple[
+        Type[_SerializableColumnGetterV2[_KT]],
+        Tuple[Sequence[Tuple[Optional[str], Optional[str]]]],
+    ]:
+        def _table_key(c: ColumnElement[_KT]) -> Optional[str]:
+            if not isinstance(c.table, expression.TableClause):
+                return None
+            else:
+                return c.table.key  # type: ignore
+
+        colkeys = [(c.key, _table_key(c)) for c in cols]
+        return _SerializableColumnGetterV2, (colkeys,)
+
+    def _cols(self, mapper: Mapper[_KT]) -> Sequence[ColumnElement[_KT]]:
+        cols: List[ColumnElement[_KT]] = []
+        metadata = getattr(mapper.local_table, "metadata", None)
+        for ckey, tkey in self.colkeys:
+            if tkey is None or metadata is None or tkey not in metadata:
+                cols.append(mapper.local_table.c[ckey])  # type: ignore
+            else:
+                cols.append(metadata.tables[tkey].c[ckey])
+        return cols
+
+
+def column_keyed_dict(
+    mapping_spec: Union[Type[_KT], Callable[[_KT], _VT]],
+    *,
+    ignore_unpopulated_attribute: bool = False,
+) -> Type[KeyFuncDict[_KT, _KT]]:
+    """A dictionary-based collection type with column-based keying.
+
+    .. versionchanged:: 2.0 Renamed :data:`.column_mapped_collection` to
+       :class:`.column_keyed_dict`.
+
+    Returns a :class:`.KeyFuncDict` factory which will produce new
+    dictionary keys based on the value of a particular :class:`.Column`-mapped
+    attribute on ORM mapped instances to be added to the dictionary.
+
+    .. note:: the value of the target attribute must be assigned with its
+       value at the time that the object is being added to the
+       dictionary collection.   Additionally, changes to the key attribute
+       are **not tracked**, which means the key in the dictionary is not
+       automatically synchronized with the key value on the target object
+       itself.  See :ref:`key_collections_mutations` for further details.
+
+    .. seealso::
+
+        :ref:`orm_dictionary_collection` - background on use
+
+    :param mapping_spec: a :class:`_schema.Column` object that is expected
+     to be mapped by the target mapper to a particular attribute on the
+     mapped class, the value of which on a particular instance is to be used
+     as the key for a new dictionary entry for that instance.
+    :param ignore_unpopulated_attribute:  if True, and the mapped attribute
+     indicated by the given :class:`_schema.Column` target attribute
+     on an object is not populated at all, the operation will be silently
+     skipped.  By default, an error is raised.
+
+     .. versionadded:: 2.0 an error is raised by default if the attribute
+        being used for the dictionary key is determined that it was never
+        populated with any value.  The
+        :paramref:`_orm.column_keyed_dict.ignore_unpopulated_attribute`
+        parameter may be set which will instead indicate that this condition
+        should be ignored, and the append operation silently skipped.
+        This is in contrast to the behavior of the 1.x series which would
+        erroneously populate the value in the dictionary with an arbitrary key
+        value of ``None``.
+
+
+    """
+    cols = [
+        coercions.expect(roles.ColumnArgumentRole, q, argname="mapping_spec")
+        for q in util.to_list(mapping_spec)
+    ]
+    keyfunc = _PlainColumnGetter(cols)
+    return _mapped_collection_cls(
+        keyfunc,
+        ignore_unpopulated_attribute=ignore_unpopulated_attribute,
+    )
+
+
+class _AttrGetter:
+    __slots__ = ("attr_name", "getter")
+
+    def __init__(self, attr_name: str):
+        self.attr_name = attr_name
+        self.getter = operator.attrgetter(attr_name)
+
+    def __call__(self, mapped_object: Any) -> Any:
+        obj = self.getter(mapped_object)
+        if obj is None:
+            state = base.instance_state(mapped_object)
+            mp = state.mapper
+            if self.attr_name in mp.attrs:
+                dict_ = state.dict
+                obj = dict_.get(self.attr_name, base.NO_VALUE)
+                if obj is None:
+                    return Missing
+            else:
+                return Missing
+
+        return obj
+
+    def __reduce__(self) -> Tuple[Type[_AttrGetter], Tuple[str]]:
+        return _AttrGetter, (self.attr_name,)
+
+
+def attribute_keyed_dict(
+    attr_name: str, *, ignore_unpopulated_attribute: bool = False
+) -> Type[KeyFuncDict[Any, Any]]:
+    """A dictionary-based collection type with attribute-based keying.
+
+    .. versionchanged:: 2.0 Renamed :data:`.attribute_mapped_collection` to
+       :func:`.attribute_keyed_dict`.
+
+    Returns a :class:`.KeyFuncDict` factory which will produce new
+    dictionary keys based on the value of a particular named attribute on
+    ORM mapped instances to be added to the dictionary.
+
+    .. note:: the value of the target attribute must be assigned with its
+       value at the time that the object is being added to the
+       dictionary collection.   Additionally, changes to the key attribute
+       are **not tracked**, which means the key in the dictionary is not
+       automatically synchronized with the key value on the target object
+       itself.  See :ref:`key_collections_mutations` for further details.
+
+    .. seealso::
+
+        :ref:`orm_dictionary_collection` - background on use
+
+    :param attr_name: string name of an ORM-mapped attribute
+     on the mapped class, the value of which on a particular instance
+     is to be used as the key for a new dictionary entry for that instance.
+    :param ignore_unpopulated_attribute:  if True, and the target attribute
+     on an object is not populated at all, the operation will be silently
+     skipped.  By default, an error is raised.
+
+     .. versionadded:: 2.0 an error is raised by default if the attribute
+        being used for the dictionary key is determined that it was never
+        populated with any value.  The
+        :paramref:`_orm.attribute_keyed_dict.ignore_unpopulated_attribute`
+        parameter may be set which will instead indicate that this condition
+        should be ignored, and the append operation silently skipped.
+        This is in contrast to the behavior of the 1.x series which would
+        erroneously populate the value in the dictionary with an arbitrary key
+        value of ``None``.
+
+
+    """
+
+    return _mapped_collection_cls(
+        _AttrGetter(attr_name),
+        ignore_unpopulated_attribute=ignore_unpopulated_attribute,
+    )
+
+
+def keyfunc_mapping(
+    keyfunc: Callable[[Any], Any],
+    *,
+    ignore_unpopulated_attribute: bool = False,
+) -> Type[KeyFuncDict[_KT, Any]]:
+    """A dictionary-based collection type with arbitrary keying.
+
+    .. versionchanged:: 2.0 Renamed :data:`.mapped_collection` to
+       :func:`.keyfunc_mapping`.
+
+    Returns a :class:`.KeyFuncDict` factory with a keying function
+    generated from keyfunc, a callable that takes an entity and returns a
+    key value.
+
+    .. note:: the given keyfunc is called only once at the time that the
+       target object is being added to the collection.   Changes to the
+       effective value returned by the function are not tracked.
+
+
+    .. seealso::
+
+        :ref:`orm_dictionary_collection` - background on use
+
+    :param keyfunc: a callable that will be passed the ORM-mapped instance
+     which should then generate a new key to use in the dictionary.
+     If the value returned is :attr:`.LoaderCallableStatus.NO_VALUE`, an error
+     is raised.
+    :param ignore_unpopulated_attribute:  if True, and the callable returns
+     :attr:`.LoaderCallableStatus.NO_VALUE` for a particular instance, the
+     operation will be silently skipped.  By default, an error is raised.
+
+     .. versionadded:: 2.0 an error is raised by default if the callable
+        being used for the dictionary key returns
+        :attr:`.LoaderCallableStatus.NO_VALUE`, which in an ORM attribute
+        context indicates an attribute that was never populated with any value.
+        The :paramref:`_orm.mapped_collection.ignore_unpopulated_attribute`
+        parameter may be set which will instead indicate that this condition
+        should be ignored, and the append operation silently skipped. This is
+        in contrast to the behavior of the 1.x series which would erroneously
+        populate the value in the dictionary with an arbitrary key value of
+        ``None``.
+
+
+    """
+    return _mapped_collection_cls(
+        keyfunc, ignore_unpopulated_attribute=ignore_unpopulated_attribute
+    )
+
+
+class KeyFuncDict(Dict[_KT, _VT]):
+    """Base for ORM mapped dictionary classes.
+
+    Extends the ``dict`` type with additional methods needed by SQLAlchemy ORM
+    collection classes. Use of :class:`_orm.KeyFuncDict` is most directly
+    by using the :func:`.attribute_keyed_dict` or
+    :func:`.column_keyed_dict` class factories.
+    :class:`_orm.KeyFuncDict` may also serve as the base for user-defined
+    custom dictionary classes.
+
+    .. versionchanged:: 2.0 Renamed :class:`.MappedCollection` to
+       :class:`.KeyFuncDict`.
+
+    .. seealso::
+
+        :func:`_orm.attribute_keyed_dict`
+
+        :func:`_orm.column_keyed_dict`
+
+        :ref:`orm_dictionary_collection`
+
+        :ref:`orm_custom_collection`
+
+
+    """
+
+    def __init__(
+        self,
+        keyfunc: Callable[[Any], Any],
+        *dict_args: Any,
+        ignore_unpopulated_attribute: bool = False,
+    ) -> None:
+        """Create a new collection with keying provided by keyfunc.
+
+        keyfunc may be any callable that takes an object and returns an object
+        for use as a dictionary key.
+
+        The keyfunc will be called every time the ORM needs to add a member by
+        value-only (such as when loading instances from the database) or
+        remove a member.  The usual cautions about dictionary keying apply-
+        ``keyfunc(object)`` should return the same output for the life of the
+        collection.  Keying based on mutable properties can result in
+        unreachable instances "lost" in the collection.
+
+        """
+        self.keyfunc = keyfunc
+        self.ignore_unpopulated_attribute = ignore_unpopulated_attribute
+        super().__init__(*dict_args)
+
+    @classmethod
+    def _unreduce(
+        cls,
+        keyfunc: Callable[[Any], Any],
+        values: Dict[_KT, _KT],
+        adapter: Optional[CollectionAdapter] = None,
+    ) -> "KeyFuncDict[_KT, _KT]":
+        mp: KeyFuncDict[_KT, _KT] = KeyFuncDict(keyfunc)
+        mp.update(values)
+        # note that the adapter sets itself up onto this collection
+        # when its `__setstate__` method is called
+        return mp
+
+    def __reduce__(
+        self,
+    ) -> Tuple[
+        Callable[[_KT, _KT], KeyFuncDict[_KT, _KT]],
+        Tuple[Any, Union[Dict[_KT, _KT], Dict[_KT, _KT]], CollectionAdapter],
+    ]:
+        return (
+            KeyFuncDict._unreduce,
+            (
+                self.keyfunc,
+                dict(self),
+                collection_adapter(self),
+            ),
+        )
+
+    @util.preload_module("sqlalchemy.orm.attributes")
+    def _raise_for_unpopulated(
+        self,
+        value: _KT,
+        initiator: Union[AttributeEventToken, Literal[None, False]] = None,
+        *,
+        warn_only: bool,
+    ) -> None:
+        mapper = base.instance_state(value).mapper
+
+        attributes = util.preloaded.orm_attributes
+
+        if not isinstance(initiator, attributes.AttributeEventToken):
+            relationship = "unknown relationship"
+        elif initiator.key in mapper.attrs:
+            relationship = f"{mapper.attrs[initiator.key]}"
+        else:
+            relationship = initiator.key
+
+        if warn_only:
+            util.warn(
+                f"Attribute keyed dictionary value for "
+                f"attribute '{relationship}' was None; this will raise "
+                "in a future release. "
+                f"To skip this assignment entirely, "
+                f'Set the "ignore_unpopulated_attribute=True" '
+                f"parameter on the mapped collection factory."
+            )
+        else:
+            raise sa_exc.InvalidRequestError(
+                "In event triggered from population of "
+                f"attribute '{relationship}' "
+                "(potentially from a backref), "
+                f"can't populate value in KeyFuncDict; "
+                "dictionary key "
+                f"derived from {base.instance_str(value)} is not "
+                f"populated. Ensure appropriate state is set up on "
+                f"the {base.instance_str(value)} object "
+                f"before assigning to the {relationship} attribute. "
+                f"To skip this assignment entirely, "
+                f'Set the "ignore_unpopulated_attribute=True" '
+                f"parameter on the mapped collection factory."
+            )
+
+    @collection.appender  # type: ignore[misc]
+    @collection.internally_instrumented  # type: ignore[misc]
+    def set(
+        self,
+        value: _KT,
+        _sa_initiator: Union[AttributeEventToken, Literal[None, False]] = None,
+    ) -> None:
+        """Add an item by value, consulting the keyfunc for the key."""
+
+        key = self.keyfunc(value)
+
+        if key is base.NO_VALUE:
+            if not self.ignore_unpopulated_attribute:
+                self._raise_for_unpopulated(
+                    value, _sa_initiator, warn_only=False
+                )
+            else:
+                return
+        elif key is Missing:
+            if not self.ignore_unpopulated_attribute:
+                self._raise_for_unpopulated(
+                    value, _sa_initiator, warn_only=True
+                )
+                key = None
+            else:
+                return
+
+        self.__setitem__(key, value, _sa_initiator)  # type: ignore[call-arg]
+
+    @collection.remover  # type: ignore[misc]
+    @collection.internally_instrumented  # type: ignore[misc]
+    def remove(
+        self,
+        value: _KT,
+        _sa_initiator: Union[AttributeEventToken, Literal[None, False]] = None,
+    ) -> None:
+        """Remove an item by value, consulting the keyfunc for the key."""
+
+        key = self.keyfunc(value)
+
+        if key is base.NO_VALUE:
+            if not self.ignore_unpopulated_attribute:
+                self._raise_for_unpopulated(
+                    value, _sa_initiator, warn_only=False
+                )
+            return
+        elif key is Missing:
+            if not self.ignore_unpopulated_attribute:
+                self._raise_for_unpopulated(
+                    value, _sa_initiator, warn_only=True
+                )
+                key = None
+            else:
+                return
+
+        # Let self[key] raise if key is not in this collection
+        # testlib.pragma exempt:__ne__
+        if self[key] != value:
+            raise sa_exc.InvalidRequestError(
+                "Can not remove '%s': collection holds '%s' for key '%s'. "
+                "Possible cause: is the KeyFuncDict key function "
+                "based on mutable properties or properties that only obtain "
+                "values after flush?" % (value, self[key], key)
+            )
+        self.__delitem__(key, _sa_initiator)  # type: ignore[call-arg]
+
+
+def _mapped_collection_cls(
+    keyfunc: Callable[[Any], Any], ignore_unpopulated_attribute: bool
+) -> Type[KeyFuncDict[_KT, _KT]]:
+    class _MKeyfuncMapped(KeyFuncDict[_KT, _KT]):
+        def __init__(self, *dict_args: Any) -> None:
+            super().__init__(
+                keyfunc,
+                *dict_args,
+                ignore_unpopulated_attribute=ignore_unpopulated_attribute,
+            )
+
+    return _MKeyfuncMapped
+
+
+MappedCollection = KeyFuncDict
+"""A synonym for :class:`.KeyFuncDict`.
+
+.. versionchanged:: 2.0 Renamed :class:`.MappedCollection` to
+   :class:`.KeyFuncDict`.
+
+"""
+
+mapped_collection = keyfunc_mapping
+"""A synonym for :func:`_orm.keyfunc_mapping`.
+
+.. versionchanged:: 2.0 Renamed :data:`.mapped_collection` to
+   :func:`_orm.keyfunc_mapping`
+
+"""
+
+attribute_mapped_collection = attribute_keyed_dict
+"""A synonym for :func:`_orm.attribute_keyed_dict`.
+
+.. versionchanged:: 2.0 Renamed :data:`.attribute_mapped_collection` to
+   :func:`_orm.attribute_keyed_dict`
+
+"""
+
+column_mapped_collection = column_keyed_dict
+"""A synonym for :func:`_orm.column_keyed_dict.
+
+.. versionchanged:: 2.0 Renamed :func:`.column_mapped_collection` to
+   :func:`_orm.column_keyed_dict`
+
+"""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapper.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapper.py
new file mode 100644
index 00000000..11010efb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/mapper.py
@@ -0,0 +1,4430 @@
+# orm/mapper.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Logic to map Python classes to and from selectables.
+
+Defines the :class:`~sqlalchemy.orm.mapper.Mapper` class, the central
+configurational unit which associates a class with a database table.
+
+This is a semi-private module; the main configurational API of the ORM is
+available in :class:`~sqlalchemy.orm.`.
+
+"""
+from __future__ import annotations
+
+from collections import deque
+from functools import reduce
+from itertools import chain
+import sys
+import threading
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Deque
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import exc as orm_exc
+from . import instrumentation
+from . import loading
+from . import properties
+from . import util as orm_util
+from ._typing import _O
+from .base import _class_to_mapper
+from .base import _parse_mapper_argument
+from .base import _state_mapper
+from .base import PassiveFlag
+from .base import state_str
+from .interfaces import _MappedAttribute
+from .interfaces import EXT_SKIP
+from .interfaces import InspectionAttr
+from .interfaces import MapperProperty
+from .interfaces import ORMEntityColumnsClauseRole
+from .interfaces import ORMFromClauseRole
+from .interfaces import StrategizedProperty
+from .path_registry import PathRegistry
+from .. import event
+from .. import exc as sa_exc
+from .. import inspection
+from .. import log
+from .. import schema
+from .. import sql
+from .. import util
+from ..event import dispatcher
+from ..event import EventTarget
+from ..sql import base as sql_base
+from ..sql import coercions
+from ..sql import expression
+from ..sql import operators
+from ..sql import roles
+from ..sql import TableClause
+from ..sql import util as sql_util
+from ..sql import visitors
+from ..sql.cache_key import MemoizedHasCacheKey
+from ..sql.elements import KeyedColumnElement
+from ..sql.schema import Column
+from ..sql.schema import Table
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..util import HasMemoized
+from ..util import HasMemoized_ro_memoized_attribute
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import _ORMColumnExprArgument
+    from ._typing import _RegistryType
+    from .decl_api import registry
+    from .dependency import DependencyProcessor
+    from .descriptor_props import CompositeProperty
+    from .descriptor_props import SynonymProperty
+    from .events import MapperEvents
+    from .instrumentation import ClassManager
+    from .path_registry import CachingEntityRegistry
+    from .properties import ColumnProperty
+    from .relationships import RelationshipProperty
+    from .state import InstanceState
+    from .util import ORMAdapter
+    from ..engine import Row
+    from ..engine import RowMapping
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _EquivalentColumnMap
+    from ..sql.base import ReadOnlyColumnCollection
+    from ..sql.elements import ColumnClause
+    from ..sql.elements import ColumnElement
+    from ..sql.selectable import FromClause
+    from ..util import OrderedSet
+
+
+_T = TypeVar("_T", bound=Any)
+_MP = TypeVar("_MP", bound="MapperProperty[Any]")
+_Fn = TypeVar("_Fn", bound="Callable[..., Any]")
+
+
+_WithPolymorphicArg = Union[
+    Literal["*"],
+    Tuple[
+        Union[Literal["*"], Sequence[Union["Mapper[Any]", Type[Any]]]],
+        Optional["FromClause"],
+    ],
+    Sequence[Union["Mapper[Any]", Type[Any]]],
+]
+
+
+_mapper_registries: weakref.WeakKeyDictionary[_RegistryType, bool] = (
+    weakref.WeakKeyDictionary()
+)
+
+
+def _all_registries() -> Set[registry]:
+    with _CONFIGURE_MUTEX:
+        return set(_mapper_registries)
+
+
+def _unconfigured_mappers() -> Iterator[Mapper[Any]]:
+    for reg in _all_registries():
+        yield from reg._mappers_to_configure()
+
+
+_already_compiling = False
+
+
+# a constant returned by _get_attr_by_column to indicate
+# this mapper is not handling an attribute for a particular
+# column
+NO_ATTRIBUTE = util.symbol("NO_ATTRIBUTE")
+
+# lock used to synchronize the "mapper configure" step
+_CONFIGURE_MUTEX = threading.RLock()
+
+
+@inspection._self_inspects
+@log.class_logger
+class Mapper(
+    ORMFromClauseRole,
+    ORMEntityColumnsClauseRole[_O],
+    MemoizedHasCacheKey,
+    InspectionAttr,
+    log.Identified,
+    inspection.Inspectable["Mapper[_O]"],
+    EventTarget,
+    Generic[_O],
+):
+    """Defines an association between a Python class and a database table or
+    other relational structure, so that ORM operations against the class may
+    proceed.
+
+    The :class:`_orm.Mapper` object is instantiated using mapping methods
+    present on the :class:`_orm.registry` object.  For information
+    about instantiating new :class:`_orm.Mapper` objects, see
+    :ref:`orm_mapping_classes_toplevel`.
+
+    """
+
+    dispatch: dispatcher[Mapper[_O]]
+
+    _dispose_called = False
+    _configure_failed: Any = False
+    _ready_for_configure = False
+
+    @util.deprecated_params(
+        non_primary=(
+            "1.3",
+            "The :paramref:`.mapper.non_primary` parameter is deprecated, "
+            "and will be removed in a future release.  The functionality "
+            "of non primary mappers is now better suited using the "
+            ":class:`.AliasedClass` construct, which can also be used "
+            "as the target of a :func:`_orm.relationship` in 1.3.",
+        ),
+    )
+    def __init__(
+        self,
+        class_: Type[_O],
+        local_table: Optional[FromClause] = None,
+        properties: Optional[Mapping[str, MapperProperty[Any]]] = None,
+        primary_key: Optional[Iterable[_ORMColumnExprArgument[Any]]] = None,
+        non_primary: bool = False,
+        inherits: Optional[Union[Mapper[Any], Type[Any]]] = None,
+        inherit_condition: Optional[_ColumnExpressionArgument[bool]] = None,
+        inherit_foreign_keys: Optional[
+            Sequence[_ORMColumnExprArgument[Any]]
+        ] = None,
+        always_refresh: bool = False,
+        version_id_col: Optional[_ORMColumnExprArgument[Any]] = None,
+        version_id_generator: Optional[
+            Union[Literal[False], Callable[[Any], Any]]
+        ] = None,
+        polymorphic_on: Optional[
+            Union[_ORMColumnExprArgument[Any], str, MapperProperty[Any]]
+        ] = None,
+        _polymorphic_map: Optional[Dict[Any, Mapper[Any]]] = None,
+        polymorphic_identity: Optional[Any] = None,
+        concrete: bool = False,
+        with_polymorphic: Optional[_WithPolymorphicArg] = None,
+        polymorphic_abstract: bool = False,
+        polymorphic_load: Optional[Literal["selectin", "inline"]] = None,
+        allow_partial_pks: bool = True,
+        batch: bool = True,
+        column_prefix: Optional[str] = None,
+        include_properties: Optional[Sequence[str]] = None,
+        exclude_properties: Optional[Sequence[str]] = None,
+        passive_updates: bool = True,
+        passive_deletes: bool = False,
+        confirm_deleted_rows: bool = True,
+        eager_defaults: Literal[True, False, "auto"] = "auto",
+        legacy_is_orphan: bool = False,
+        _compiled_cache_size: int = 100,
+    ):
+        r"""Direct constructor for a new :class:`_orm.Mapper` object.
+
+        The :class:`_orm.Mapper` constructor is not called directly, and
+        is normally invoked through the
+        use of the :class:`_orm.registry` object through either the
+        :ref:`Declarative <orm_declarative_mapping>` or
+        :ref:`Imperative <orm_imperative_mapping>` mapping styles.
+
+        .. versionchanged:: 2.0 The public facing ``mapper()`` function is
+           removed; for a classical mapping configuration, use the
+           :meth:`_orm.registry.map_imperatively` method.
+
+        Parameters documented below may be passed to either the
+        :meth:`_orm.registry.map_imperatively` method, or may be passed in the
+        ``__mapper_args__`` declarative class attribute described at
+        :ref:`orm_declarative_mapper_options`.
+
+        :param class\_: The class to be mapped.  When using Declarative,
+          this argument is automatically passed as the declared class
+          itself.
+
+        :param local_table: The :class:`_schema.Table` or other
+           :class:`_sql.FromClause` (i.e. selectable) to which the class is
+           mapped. May be ``None`` if this mapper inherits from another mapper
+           using single-table inheritance. When using Declarative, this
+           argument is automatically passed by the extension, based on what is
+           configured via the :attr:`_orm.DeclarativeBase.__table__` attribute
+           or via the :class:`_schema.Table` produced as a result of
+           the :attr:`_orm.DeclarativeBase.__tablename__` attribute being
+           present.
+
+        :param polymorphic_abstract: Indicates this class will be mapped in a
+            polymorphic hierarchy, but not directly instantiated. The class is
+            mapped normally, except that it has no requirement for a
+            :paramref:`_orm.Mapper.polymorphic_identity` within an inheritance
+            hierarchy. The class however must be part of a polymorphic
+            inheritance scheme which uses
+            :paramref:`_orm.Mapper.polymorphic_on` at the base.
+
+            .. versionadded:: 2.0
+
+            .. seealso::
+
+                :ref:`orm_inheritance_abstract_poly`
+
+        :param always_refresh: If True, all query operations for this mapped
+           class will overwrite all data within object instances that already
+           exist within the session, erasing any in-memory changes with
+           whatever information was loaded from the database. Usage of this
+           flag is highly discouraged; as an alternative, see the method
+           :meth:`_query.Query.populate_existing`.
+
+        :param allow_partial_pks: Defaults to True.  Indicates that a
+           composite primary key with some NULL values should be considered as
+           possibly existing within the database. This affects whether a
+           mapper will assign an incoming row to an existing identity, as well
+           as if :meth:`.Session.merge` will check the database first for a
+           particular primary key value. A "partial primary key" can occur if
+           one has mapped to an OUTER JOIN, for example.
+
+           The :paramref:`.orm.Mapper.allow_partial_pks` parameter also
+           indicates to the ORM relationship lazy loader, when loading a
+           many-to-one related object, if a composite primary key that has
+           partial NULL values should result in an attempt to load from the
+           database, or if a load attempt is not necessary.
+
+           .. versionadded:: 2.0.36 :paramref:`.orm.Mapper.allow_partial_pks`
+              is consulted by the relationship lazy loader strategy, such that
+              when set to False, a SELECT for a composite primary key that
+              has partial NULL values will not be emitted.
+
+        :param batch: Defaults to ``True``, indicating that save operations
+           of multiple entities can be batched together for efficiency.
+           Setting to False indicates
+           that an instance will be fully saved before saving the next
+           instance.  This is used in the extremely rare case that a
+           :class:`.MapperEvents` listener requires being called
+           in between individual row persistence operations.
+
+        :param column_prefix: A string which will be prepended
+           to the mapped attribute name when :class:`_schema.Column`
+           objects are automatically assigned as attributes to the
+           mapped class.  Does not affect :class:`.Column` objects that
+           are mapped explicitly in the :paramref:`.Mapper.properties`
+           dictionary.
+
+           This parameter is typically useful with imperative mappings
+           that keep the :class:`.Table` object separate.  Below, assuming
+           the ``user_table`` :class:`.Table` object has columns named
+           ``user_id``, ``user_name``, and ``password``::
+
+                class User(Base):
+                    __table__ = user_table
+                    __mapper_args__ = {"column_prefix": "_"}
+
+           The above mapping will assign the ``user_id``, ``user_name``, and
+           ``password`` columns to attributes named ``_user_id``,
+           ``_user_name``, and ``_password`` on the mapped ``User`` class.
+
+           The :paramref:`.Mapper.column_prefix` parameter is uncommon in
+           modern use. For dealing with reflected tables, a more flexible
+           approach to automating a naming scheme is to intercept the
+           :class:`.Column` objects as they are reflected; see the section
+           :ref:`mapper_automated_reflection_schemes` for notes on this usage
+           pattern.
+
+        :param concrete: If True, indicates this mapper should use concrete
+           table inheritance with its parent mapper.
+
+           See the section :ref:`concrete_inheritance` for an example.
+
+        :param confirm_deleted_rows: defaults to True; when a DELETE occurs
+          of one more rows based on specific primary keys, a warning is
+          emitted when the number of rows matched does not equal the number
+          of rows expected.  This parameter may be set to False to handle the
+          case where database ON DELETE CASCADE rules may be deleting some of
+          those rows automatically.  The warning may be changed to an
+          exception in a future release.
+
+        :param eager_defaults: if True, the ORM will immediately fetch the
+          value of server-generated default values after an INSERT or UPDATE,
+          rather than leaving them as expired to be fetched on next access.
+          This can be used for event schemes where the server-generated values
+          are needed immediately before the flush completes.
+
+          The fetch of values occurs either by using ``RETURNING`` inline
+          with the ``INSERT`` or ``UPDATE`` statement, or by adding an
+          additional ``SELECT`` statement subsequent to the ``INSERT`` or
+          ``UPDATE``, if the backend does not support ``RETURNING``.
+
+          The use of ``RETURNING`` is extremely performant in particular for
+          ``INSERT`` statements where SQLAlchemy can take advantage of
+          :ref:`insertmanyvalues <engine_insertmanyvalues>`, whereas the use of
+          an additional ``SELECT`` is relatively poor performing, adding
+          additional SQL round trips which would be unnecessary if these new
+          attributes are not to be accessed in any case.
+
+          For this reason, :paramref:`.Mapper.eager_defaults` defaults to the
+          string value ``"auto"``, which indicates that server defaults for
+          INSERT should be fetched using ``RETURNING`` if the backing database
+          supports it and if the dialect in use supports "insertmanyreturning"
+          for an INSERT statement. If the backing database does not support
+          ``RETURNING`` or "insertmanyreturning" is not available, server
+          defaults will not be fetched.
+
+          .. versionchanged:: 2.0.0rc1 added the "auto" option for
+             :paramref:`.Mapper.eager_defaults`
+
+          .. seealso::
+
+                :ref:`orm_server_defaults`
+
+          .. versionchanged:: 2.0.0  RETURNING now works with multiple rows
+             INSERTed at once using the
+             :ref:`insertmanyvalues <engine_insertmanyvalues>` feature, which
+             among other things allows the :paramref:`.Mapper.eager_defaults`
+             feature to be very performant on supporting backends.
+
+        :param exclude_properties: A list or set of string column names to
+          be excluded from mapping.
+
+          .. seealso::
+
+            :ref:`include_exclude_cols`
+
+        :param include_properties: An inclusive list or set of string column
+          names to map.
+
+          .. seealso::
+
+            :ref:`include_exclude_cols`
+
+        :param inherits: A mapped class or the corresponding
+          :class:`_orm.Mapper`
+          of one indicating a superclass to which this :class:`_orm.Mapper`
+          should *inherit* from.   The mapped class here must be a subclass
+          of the other mapper's class.   When using Declarative, this argument
+          is passed automatically as a result of the natural class
+          hierarchy of the declared classes.
+
+          .. seealso::
+
+            :ref:`inheritance_toplevel`
+
+        :param inherit_condition: For joined table inheritance, a SQL
+           expression which will
+           define how the two tables are joined; defaults to a natural join
+           between the two tables.
+
+        :param inherit_foreign_keys: When ``inherit_condition`` is used and
+           the columns present are missing a :class:`_schema.ForeignKey`
+           configuration, this parameter can be used to specify which columns
+           are "foreign".  In most cases can be left as ``None``.
+
+        :param legacy_is_orphan: Boolean, defaults to ``False``.
+          When ``True``, specifies that "legacy" orphan consideration
+          is to be applied to objects mapped by this mapper, which means
+          that a pending (that is, not persistent) object is auto-expunged
+          from an owning :class:`.Session` only when it is de-associated
+          from *all* parents that specify a ``delete-orphan`` cascade towards
+          this mapper.  The new default behavior is that the object is
+          auto-expunged when it is de-associated with *any* of its parents
+          that specify ``delete-orphan`` cascade.  This behavior is more
+          consistent with that of a persistent object, and allows behavior to
+          be consistent in more scenarios independently of whether or not an
+          orphan object has been flushed yet or not.
+
+          See the change note and example at :ref:`legacy_is_orphan_addition`
+          for more detail on this change.
+
+        :param non_primary: Specify that this :class:`_orm.Mapper`
+          is in addition
+          to the "primary" mapper, that is, the one used for persistence.
+          The :class:`_orm.Mapper` created here may be used for ad-hoc
+          mapping of the class to an alternate selectable, for loading
+          only.
+
+          .. seealso::
+
+            :ref:`relationship_aliased_class` - the new pattern that removes
+            the need for the :paramref:`_orm.Mapper.non_primary` flag.
+
+        :param passive_deletes: Indicates DELETE behavior of foreign key
+           columns when a joined-table inheritance entity is being deleted.
+           Defaults to ``False`` for a base mapper; for an inheriting mapper,
+           defaults to ``False`` unless the value is set to ``True``
+           on the superclass mapper.
+
+           When ``True``, it is assumed that ON DELETE CASCADE is configured
+           on the foreign key relationships that link this mapper's table
+           to its superclass table, so that when the unit of work attempts
+           to delete the entity, it need only emit a DELETE statement for the
+           superclass table, and not this table.
+
+           When ``False``, a DELETE statement is emitted for this mapper's
+           table individually.  If the primary key attributes local to this
+           table are unloaded, then a SELECT must be emitted in order to
+           validate these attributes; note that the primary key columns
+           of a joined-table subclass are not part of the "primary key" of
+           the object as a whole.
+
+           Note that a value of ``True`` is **always** forced onto the
+           subclass mappers; that is, it's not possible for a superclass
+           to specify passive_deletes without this taking effect for
+           all subclass mappers.
+
+           .. seealso::
+
+               :ref:`passive_deletes` - description of similar feature as
+               used with :func:`_orm.relationship`
+
+               :paramref:`.mapper.passive_updates` - supporting ON UPDATE
+               CASCADE for joined-table inheritance mappers
+
+        :param passive_updates: Indicates UPDATE behavior of foreign key
+           columns when a primary key column changes on a joined-table
+           inheritance mapping.   Defaults to ``True``.
+
+           When True, it is assumed that ON UPDATE CASCADE is configured on
+           the foreign key in the database, and that the database will handle
+           propagation of an UPDATE from a source column to dependent columns
+           on joined-table rows.
+
+           When False, it is assumed that the database does not enforce
+           referential integrity and will not be issuing its own CASCADE
+           operation for an update.  The unit of work process will
+           emit an UPDATE statement for the dependent columns during a
+           primary key change.
+
+           .. seealso::
+
+               :ref:`passive_updates` - description of a similar feature as
+               used with :func:`_orm.relationship`
+
+               :paramref:`.mapper.passive_deletes` - supporting ON DELETE
+               CASCADE for joined-table inheritance mappers
+
+        :param polymorphic_load: Specifies "polymorphic loading" behavior
+         for a subclass in an inheritance hierarchy (joined and single
+         table inheritance only).   Valid values are:
+
+          * "'inline'" - specifies this class should be part of
+            the "with_polymorphic" mappers, e.g. its columns will be included
+            in a SELECT query against the base.
+
+          * "'selectin'" - specifies that when instances of this class
+            are loaded, an additional SELECT will be emitted to retrieve
+            the columns specific to this subclass.  The SELECT uses
+            IN to fetch multiple subclasses at once.
+
+         .. versionadded:: 1.2
+
+         .. seealso::
+
+            :ref:`with_polymorphic_mapper_config`
+
+            :ref:`polymorphic_selectin`
+
+        :param polymorphic_on: Specifies the column, attribute, or
+          SQL expression used to determine the target class for an
+          incoming row, when inheriting classes are present.
+
+          May be specified as a string attribute name, or as a SQL
+          expression such as a :class:`_schema.Column` or in a Declarative
+          mapping a :func:`_orm.mapped_column` object.  It is typically
+          expected that the SQL expression corresponds to a column in the
+          base-most mapped :class:`.Table`::
+
+            class Employee(Base):
+                __tablename__ = "employee"
+
+                id: Mapped[int] = mapped_column(primary_key=True)
+                discriminator: Mapped[str] = mapped_column(String(50))
+
+                __mapper_args__ = {
+                    "polymorphic_on": discriminator,
+                    "polymorphic_identity": "employee",
+                }
+
+          It may also be specified
+          as a SQL expression, as in this example where we
+          use the :func:`.case` construct to provide a conditional
+          approach::
+
+            class Employee(Base):
+                __tablename__ = "employee"
+
+                id: Mapped[int] = mapped_column(primary_key=True)
+                discriminator: Mapped[str] = mapped_column(String(50))
+
+                __mapper_args__ = {
+                    "polymorphic_on": case(
+                        (discriminator == "EN", "engineer"),
+                        (discriminator == "MA", "manager"),
+                        else_="employee",
+                    ),
+                    "polymorphic_identity": "employee",
+                }
+
+          It may also refer to any attribute using its string name,
+          which is of particular use when using annotated column
+          configurations::
+
+                class Employee(Base):
+                    __tablename__ = "employee"
+
+                    id: Mapped[int] = mapped_column(primary_key=True)
+                    discriminator: Mapped[str]
+
+                    __mapper_args__ = {
+                        "polymorphic_on": "discriminator",
+                        "polymorphic_identity": "employee",
+                    }
+
+          When setting ``polymorphic_on`` to reference an
+          attribute or expression that's not present in the
+          locally mapped :class:`_schema.Table`, yet the value
+          of the discriminator should be persisted to the database,
+          the value of the
+          discriminator is not automatically set on new
+          instances; this must be handled by the user,
+          either through manual means or via event listeners.
+          A typical approach to establishing such a listener
+          looks like::
+
+                from sqlalchemy import event
+                from sqlalchemy.orm import object_mapper
+
+
+                @event.listens_for(Employee, "init", propagate=True)
+                def set_identity(instance, *arg, **kw):
+                    mapper = object_mapper(instance)
+                    instance.discriminator = mapper.polymorphic_identity
+
+          Where above, we assign the value of ``polymorphic_identity``
+          for the mapped class to the ``discriminator`` attribute,
+          thus persisting the value to the ``discriminator`` column
+          in the database.
+
+          .. warning::
+
+             Currently, **only one discriminator column may be set**, typically
+             on the base-most class in the hierarchy. "Cascading" polymorphic
+             columns are not yet supported.
+
+          .. seealso::
+
+            :ref:`inheritance_toplevel`
+
+        :param polymorphic_identity: Specifies the value which
+          identifies this particular class as returned by the column expression
+          referred to by the :paramref:`_orm.Mapper.polymorphic_on` setting. As
+          rows are received, the value corresponding to the
+          :paramref:`_orm.Mapper.polymorphic_on` column expression is compared
+          to this value, indicating which subclass should be used for the newly
+          reconstructed object.
+
+          .. seealso::
+
+            :ref:`inheritance_toplevel`
+
+        :param properties: A dictionary mapping the string names of object
+           attributes to :class:`.MapperProperty` instances, which define the
+           persistence behavior of that attribute.  Note that
+           :class:`_schema.Column`
+           objects present in
+           the mapped :class:`_schema.Table` are automatically placed into
+           ``ColumnProperty`` instances upon mapping, unless overridden.
+           When using Declarative, this argument is passed automatically,
+           based on all those :class:`.MapperProperty` instances declared
+           in the declared class body.
+
+           .. seealso::
+
+               :ref:`orm_mapping_properties` - in the
+               :ref:`orm_mapping_classes_toplevel`
+
+        :param primary_key: A list of :class:`_schema.Column`
+           objects, or alternatively string names of attribute names which
+           refer to :class:`_schema.Column`, which define
+           the primary key to be used against this mapper's selectable unit.
+           This is normally simply the primary key of the ``local_table``, but
+           can be overridden here.
+
+           .. versionchanged:: 2.0.2 :paramref:`_orm.Mapper.primary_key`
+              arguments may be indicated as string attribute names as well.
+
+           .. seealso::
+
+                :ref:`mapper_primary_key` - background and example use
+
+        :param version_id_col: A :class:`_schema.Column`
+           that will be used to keep a running version id of rows
+           in the table.  This is used to detect concurrent updates or
+           the presence of stale data in a flush.  The methodology is to
+           detect if an UPDATE statement does not match the last known
+           version id, a
+           :class:`~sqlalchemy.orm.exc.StaleDataError` exception is
+           thrown.
+           By default, the column must be of :class:`.Integer` type,
+           unless ``version_id_generator`` specifies an alternative version
+           generator.
+
+           .. seealso::
+
+              :ref:`mapper_version_counter` - discussion of version counting
+              and rationale.
+
+        :param version_id_generator: Define how new version ids should
+          be generated.  Defaults to ``None``, which indicates that
+          a simple integer counting scheme be employed.  To provide a custom
+          versioning scheme, provide a callable function of the form::
+
+              def generate_version(version):
+                  return next_version
+
+          Alternatively, server-side versioning functions such as triggers,
+          or programmatic versioning schemes outside of the version id
+          generator may be used, by specifying the value ``False``.
+          Please see :ref:`server_side_version_counter` for a discussion
+          of important points when using this option.
+
+          .. seealso::
+
+             :ref:`custom_version_counter`
+
+             :ref:`server_side_version_counter`
+
+
+        :param with_polymorphic: A tuple in the form ``(<classes>,
+            <selectable>)`` indicating the default style of "polymorphic"
+            loading, that is, which tables are queried at once. <classes> is
+            any single or list of mappers and/or classes indicating the
+            inherited classes that should be loaded at once. The special value
+            ``'*'`` may be used to indicate all descending classes should be
+            loaded immediately. The second tuple argument <selectable>
+            indicates a selectable that will be used to query for multiple
+            classes.
+
+            The :paramref:`_orm.Mapper.polymorphic_load` parameter may be
+            preferable over the use of :paramref:`_orm.Mapper.with_polymorphic`
+            in modern mappings to indicate a per-subclass technique of
+            indicating polymorphic loading styles.
+
+            .. seealso::
+
+                :ref:`with_polymorphic_mapper_config`
+
+        """
+        self.class_ = util.assert_arg_type(class_, type, "class_")
+        self._sort_key = "%s.%s" % (
+            self.class_.__module__,
+            self.class_.__name__,
+        )
+
+        self._primary_key_argument = util.to_list(primary_key)
+        self.non_primary = non_primary
+
+        self.always_refresh = always_refresh
+
+        if isinstance(version_id_col, MapperProperty):
+            self.version_id_prop = version_id_col
+            self.version_id_col = None
+        else:
+            self.version_id_col = (
+                coercions.expect(
+                    roles.ColumnArgumentOrKeyRole,
+                    version_id_col,
+                    argname="version_id_col",
+                )
+                if version_id_col is not None
+                else None
+            )
+
+        if version_id_generator is False:
+            self.version_id_generator = False
+        elif version_id_generator is None:
+            self.version_id_generator = lambda x: (x or 0) + 1
+        else:
+            self.version_id_generator = version_id_generator
+
+        self.concrete = concrete
+        self.single = False
+
+        if inherits is not None:
+            self.inherits = _parse_mapper_argument(inherits)
+        else:
+            self.inherits = None
+
+        if local_table is not None:
+            self.local_table = coercions.expect(
+                roles.StrictFromClauseRole,
+                local_table,
+                disable_inspection=True,
+                argname="local_table",
+            )
+        elif self.inherits:
+            # note this is a new flow as of 2.0 so that
+            # .local_table need not be Optional
+            self.local_table = self.inherits.local_table
+            self.single = True
+        else:
+            raise sa_exc.ArgumentError(
+                f"Mapper[{self.class_.__name__}(None)] has None for a "
+                "primary table argument and does not specify 'inherits'"
+            )
+
+        if inherit_condition is not None:
+            self.inherit_condition = coercions.expect(
+                roles.OnClauseRole, inherit_condition
+            )
+        else:
+            self.inherit_condition = None
+
+        self.inherit_foreign_keys = inherit_foreign_keys
+        self._init_properties = dict(properties) if properties else {}
+        self._delete_orphans = []
+        self.batch = batch
+        self.eager_defaults = eager_defaults
+        self.column_prefix = column_prefix
+
+        # interim - polymorphic_on is further refined in
+        # _configure_polymorphic_setter
+        self.polymorphic_on = (
+            coercions.expect(  # type: ignore
+                roles.ColumnArgumentOrKeyRole,
+                polymorphic_on,
+                argname="polymorphic_on",
+            )
+            if polymorphic_on is not None
+            else None
+        )
+        self.polymorphic_abstract = polymorphic_abstract
+        self._dependency_processors = []
+        self.validators = util.EMPTY_DICT
+        self.passive_updates = passive_updates
+        self.passive_deletes = passive_deletes
+        self.legacy_is_orphan = legacy_is_orphan
+        self._clause_adapter = None
+        self._requires_row_aliasing = False
+        self._inherits_equated_pairs = None
+        self._memoized_values = {}
+        self._compiled_cache_size = _compiled_cache_size
+        self._reconstructor = None
+        self.allow_partial_pks = allow_partial_pks
+
+        if self.inherits and not self.concrete:
+            self.confirm_deleted_rows = False
+        else:
+            self.confirm_deleted_rows = confirm_deleted_rows
+
+        self._set_with_polymorphic(with_polymorphic)
+        self.polymorphic_load = polymorphic_load
+
+        # our 'polymorphic identity', a string name that when located in a
+        #  result set row indicates this Mapper should be used to construct
+        # the object instance for that row.
+        self.polymorphic_identity = polymorphic_identity
+
+        # a dictionary of 'polymorphic identity' names, associating those
+        # names with Mappers that will be used to construct object instances
+        # upon a select operation.
+        if _polymorphic_map is None:
+            self.polymorphic_map = {}
+        else:
+            self.polymorphic_map = _polymorphic_map
+
+        if include_properties is not None:
+            self.include_properties = util.to_set(include_properties)
+        else:
+            self.include_properties = None
+        if exclude_properties:
+            self.exclude_properties = util.to_set(exclude_properties)
+        else:
+            self.exclude_properties = None
+
+        # prevent this mapper from being constructed
+        # while a configure_mappers() is occurring (and defer a
+        # configure_mappers() until construction succeeds)
+        with _CONFIGURE_MUTEX:
+            cast("MapperEvents", self.dispatch._events)._new_mapper_instance(
+                class_, self
+            )
+            self._configure_inheritance()
+            self._configure_class_instrumentation()
+            self._configure_properties()
+            self._configure_polymorphic_setter()
+            self._configure_pks()
+            self.registry._flag_new_mapper(self)
+            self._log("constructed")
+            self._expire_memoizations()
+
+        self.dispatch.after_mapper_constructed(self, self.class_)
+
+    def _prefer_eager_defaults(self, dialect, table):
+        if self.eager_defaults == "auto":
+            if not table.implicit_returning:
+                return False
+
+            return (
+                table in self._server_default_col_keys
+                and dialect.insert_executemany_returning
+            )
+        else:
+            return self.eager_defaults
+
+    def _gen_cache_key(self, anon_map, bindparams):
+        return (self,)
+
+    # ### BEGIN
+    # ATTRIBUTE DECLARATIONS START HERE
+
+    is_mapper = True
+    """Part of the inspection API."""
+
+    represents_outer_join = False
+
+    registry: _RegistryType
+
+    @property
+    def mapper(self) -> Mapper[_O]:
+        """Part of the inspection API.
+
+        Returns self.
+
+        """
+        return self
+
+    @property
+    def entity(self):
+        r"""Part of the inspection API.
+
+        Returns self.class\_.
+
+        """
+        return self.class_
+
+    class_: Type[_O]
+    """The class to which this :class:`_orm.Mapper` is mapped."""
+
+    _identity_class: Type[_O]
+
+    _delete_orphans: List[Tuple[str, Type[Any]]]
+    _dependency_processors: List[DependencyProcessor]
+    _memoized_values: Dict[Any, Callable[[], Any]]
+    _inheriting_mappers: util.WeakSequence[Mapper[Any]]
+    _all_tables: Set[TableClause]
+    _polymorphic_attr_key: Optional[str]
+
+    _pks_by_table: Dict[FromClause, OrderedSet[ColumnClause[Any]]]
+    _cols_by_table: Dict[FromClause, OrderedSet[ColumnElement[Any]]]
+
+    _props: util.OrderedDict[str, MapperProperty[Any]]
+    _init_properties: Dict[str, MapperProperty[Any]]
+
+    _columntoproperty: _ColumnMapping
+
+    _set_polymorphic_identity: Optional[Callable[[InstanceState[_O]], None]]
+    _validate_polymorphic_identity: Optional[
+        Callable[[Mapper[_O], InstanceState[_O], _InstanceDict], None]
+    ]
+
+    tables: Sequence[TableClause]
+    """A sequence containing the collection of :class:`_schema.Table`
+    or :class:`_schema.TableClause` objects which this :class:`_orm.Mapper`
+    is aware of.
+
+    If the mapper is mapped to a :class:`_expression.Join`, or an
+    :class:`_expression.Alias`
+    representing a :class:`_expression.Select`, the individual
+    :class:`_schema.Table`
+    objects that comprise the full construct will be represented here.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    validators: util.immutabledict[str, Tuple[str, Dict[str, Any]]]
+    """An immutable dictionary of attributes which have been decorated
+    using the :func:`_orm.validates` decorator.
+
+    The dictionary contains string attribute names as keys
+    mapped to the actual validation method.
+
+    """
+
+    always_refresh: bool
+    allow_partial_pks: bool
+    version_id_col: Optional[ColumnElement[Any]]
+
+    with_polymorphic: Optional[
+        Tuple[
+            Union[Literal["*"], Sequence[Union[Mapper[Any], Type[Any]]]],
+            Optional[FromClause],
+        ]
+    ]
+
+    version_id_generator: Optional[Union[Literal[False], Callable[[Any], Any]]]
+
+    local_table: FromClause
+    """The immediate :class:`_expression.FromClause` to which this
+    :class:`_orm.Mapper` refers.
+
+    Typically is an instance of :class:`_schema.Table`, may be any
+    :class:`.FromClause`.
+
+    The "local" table is the
+    selectable that the :class:`_orm.Mapper` is directly responsible for
+    managing from an attribute access and flush perspective.   For
+    non-inheriting mappers, :attr:`.Mapper.local_table` will be the same
+    as :attr:`.Mapper.persist_selectable`.  For inheriting mappers,
+    :attr:`.Mapper.local_table` refers to the specific portion of
+    :attr:`.Mapper.persist_selectable` that includes the columns to which
+    this :class:`.Mapper` is loading/persisting, such as a particular
+    :class:`.Table` within a join.
+
+    .. seealso::
+
+        :attr:`_orm.Mapper.persist_selectable`.
+
+        :attr:`_orm.Mapper.selectable`.
+
+    """
+
+    persist_selectable: FromClause
+    """The :class:`_expression.FromClause` to which this :class:`_orm.Mapper`
+    is mapped.
+
+    Typically is an instance of :class:`_schema.Table`, may be any
+    :class:`.FromClause`.
+
+    The :attr:`_orm.Mapper.persist_selectable` is similar to
+    :attr:`.Mapper.local_table`, but represents the :class:`.FromClause` that
+    represents the inheriting class hierarchy overall in an inheritance
+    scenario.
+
+    :attr.`.Mapper.persist_selectable` is also separate from the
+    :attr:`.Mapper.selectable` attribute, the latter of which may be an
+    alternate subquery used for selecting columns.
+    :attr.`.Mapper.persist_selectable` is oriented towards columns that
+    will be written on a persist operation.
+
+    .. seealso::
+
+        :attr:`_orm.Mapper.selectable`.
+
+        :attr:`_orm.Mapper.local_table`.
+
+    """
+
+    inherits: Optional[Mapper[Any]]
+    """References the :class:`_orm.Mapper` which this :class:`_orm.Mapper`
+    inherits from, if any.
+
+    """
+
+    inherit_condition: Optional[ColumnElement[bool]]
+
+    configured: bool = False
+    """Represent ``True`` if this :class:`_orm.Mapper` has been configured.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    .. seealso::
+
+        :func:`.configure_mappers`.
+
+    """
+
+    concrete: bool
+    """Represent ``True`` if this :class:`_orm.Mapper` is a concrete
+    inheritance mapper.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    primary_key: Tuple[Column[Any], ...]
+    """An iterable containing the collection of :class:`_schema.Column`
+    objects
+    which comprise the 'primary key' of the mapped table, from the
+    perspective of this :class:`_orm.Mapper`.
+
+    This list is against the selectable in
+    :attr:`_orm.Mapper.persist_selectable`.
+    In the case of inheriting mappers, some columns may be managed by a
+    superclass mapper.  For example, in the case of a
+    :class:`_expression.Join`, the
+    primary key is determined by all of the primary key columns across all
+    tables referenced by the :class:`_expression.Join`.
+
+    The list is also not necessarily the same as the primary key column
+    collection associated with the underlying tables; the :class:`_orm.Mapper`
+    features a ``primary_key`` argument that can override what the
+    :class:`_orm.Mapper` considers as primary key columns.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    class_manager: ClassManager[_O]
+    """The :class:`.ClassManager` which maintains event listeners
+    and class-bound descriptors for this :class:`_orm.Mapper`.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    single: bool
+    """Represent ``True`` if this :class:`_orm.Mapper` is a single table
+    inheritance mapper.
+
+    :attr:`_orm.Mapper.local_table` will be ``None`` if this flag is set.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    non_primary: bool
+    """Represent ``True`` if this :class:`_orm.Mapper` is a "non-primary"
+    mapper, e.g. a mapper that is used only to select rows but not for
+    persistence management.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    polymorphic_on: Optional[KeyedColumnElement[Any]]
+    """The :class:`_schema.Column` or SQL expression specified as the
+    ``polymorphic_on`` argument
+    for this :class:`_orm.Mapper`, within an inheritance scenario.
+
+    This attribute is normally a :class:`_schema.Column` instance but
+    may also be an expression, such as one derived from
+    :func:`.cast`.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    polymorphic_map: Dict[Any, Mapper[Any]]
+    """A mapping of "polymorphic identity" identifiers mapped to
+    :class:`_orm.Mapper` instances, within an inheritance scenario.
+
+    The identifiers can be of any type which is comparable to the
+    type of column represented by :attr:`_orm.Mapper.polymorphic_on`.
+
+    An inheritance chain of mappers will all reference the same
+    polymorphic map object.  The object is used to correlate incoming
+    result rows to target mappers.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    polymorphic_identity: Optional[Any]
+    """Represent an identifier which is matched against the
+    :attr:`_orm.Mapper.polymorphic_on` column during result row loading.
+
+    Used only with inheritance, this object can be of any type which is
+    comparable to the type of column represented by
+    :attr:`_orm.Mapper.polymorphic_on`.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    base_mapper: Mapper[Any]
+    """The base-most :class:`_orm.Mapper` in an inheritance chain.
+
+    In a non-inheriting scenario, this attribute will always be this
+    :class:`_orm.Mapper`.   In an inheritance scenario, it references
+    the :class:`_orm.Mapper` which is parent to all other :class:`_orm.Mapper`
+    objects in the inheritance chain.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    columns: ReadOnlyColumnCollection[str, Column[Any]]
+    """A collection of :class:`_schema.Column` or other scalar expression
+    objects maintained by this :class:`_orm.Mapper`.
+
+    The collection behaves the same as that of the ``c`` attribute on
+    any :class:`_schema.Table` object,
+    except that only those columns included in
+    this mapping are present, and are keyed based on the attribute name
+    defined in the mapping, not necessarily the ``key`` attribute of the
+    :class:`_schema.Column` itself.   Additionally, scalar expressions mapped
+    by :func:`.column_property` are also present here.
+
+    This is a *read only* attribute determined during mapper construction.
+    Behavior is undefined if directly modified.
+
+    """
+
+    c: ReadOnlyColumnCollection[str, Column[Any]]
+    """A synonym for :attr:`_orm.Mapper.columns`."""
+
+    @util.non_memoized_property
+    @util.deprecated("1.3", "Use .persist_selectable")
+    def mapped_table(self):
+        return self.persist_selectable
+
+    @util.memoized_property
+    def _path_registry(self) -> CachingEntityRegistry:
+        return PathRegistry.per_mapper(self)
+
+    def _configure_inheritance(self):
+        """Configure settings related to inheriting and/or inherited mappers
+        being present."""
+
+        # a set of all mappers which inherit from this one.
+        self._inheriting_mappers = util.WeakSequence()
+
+        if self.inherits:
+            if not issubclass(self.class_, self.inherits.class_):
+                raise sa_exc.ArgumentError(
+                    "Class '%s' does not inherit from '%s'"
+                    % (self.class_.__name__, self.inherits.class_.__name__)
+                )
+
+            self.dispatch._update(self.inherits.dispatch)
+
+            if self.non_primary != self.inherits.non_primary:
+                np = not self.non_primary and "primary" or "non-primary"
+                raise sa_exc.ArgumentError(
+                    "Inheritance of %s mapper for class '%s' is "
+                    "only allowed from a %s mapper"
+                    % (np, self.class_.__name__, np)
+                )
+
+            if self.single:
+                self.persist_selectable = self.inherits.persist_selectable
+            elif self.local_table is not self.inherits.local_table:
+                if self.concrete:
+                    self.persist_selectable = self.local_table
+                    for mapper in self.iterate_to_root():
+                        if mapper.polymorphic_on is not None:
+                            mapper._requires_row_aliasing = True
+                else:
+                    if self.inherit_condition is None:
+                        # figure out inherit condition from our table to the
+                        # immediate table of the inherited mapper, not its
+                        # full table which could pull in other stuff we don't
+                        # want (allows test/inheritance.InheritTest4 to pass)
+                        try:
+                            self.inherit_condition = sql_util.join_condition(
+                                self.inherits.local_table, self.local_table
+                            )
+                        except sa_exc.NoForeignKeysError as nfe:
+                            assert self.inherits.local_table is not None
+                            assert self.local_table is not None
+                            raise sa_exc.NoForeignKeysError(
+                                "Can't determine the inherit condition "
+                                "between inherited table '%s' and "
+                                "inheriting "
+                                "table '%s'; tables have no "
+                                "foreign key relationships established.  "
+                                "Please ensure the inheriting table has "
+                                "a foreign key relationship to the "
+                                "inherited "
+                                "table, or provide an "
+                                "'on clause' using "
+                                "the 'inherit_condition' mapper argument."
+                                % (
+                                    self.inherits.local_table.description,
+                                    self.local_table.description,
+                                )
+                            ) from nfe
+                        except sa_exc.AmbiguousForeignKeysError as afe:
+                            assert self.inherits.local_table is not None
+                            assert self.local_table is not None
+                            raise sa_exc.AmbiguousForeignKeysError(
+                                "Can't determine the inherit condition "
+                                "between inherited table '%s' and "
+                                "inheriting "
+                                "table '%s'; tables have more than one "
+                                "foreign key relationship established.  "
+                                "Please specify the 'on clause' using "
+                                "the 'inherit_condition' mapper argument."
+                                % (
+                                    self.inherits.local_table.description,
+                                    self.local_table.description,
+                                )
+                            ) from afe
+                    assert self.inherits.persist_selectable is not None
+                    self.persist_selectable = sql.join(
+                        self.inherits.persist_selectable,
+                        self.local_table,
+                        self.inherit_condition,
+                    )
+
+                    fks = util.to_set(self.inherit_foreign_keys)
+                    self._inherits_equated_pairs = sql_util.criterion_as_pairs(
+                        self.persist_selectable.onclause,
+                        consider_as_foreign_keys=fks,
+                    )
+            else:
+                self.persist_selectable = self.local_table
+
+            if self.polymorphic_identity is None:
+                self._identity_class = self.class_
+
+                if (
+                    not self.polymorphic_abstract
+                    and self.inherits.base_mapper.polymorphic_on is not None
+                ):
+                    util.warn(
+                        f"{self} does not indicate a 'polymorphic_identity', "
+                        "yet is part of an inheritance hierarchy that has a "
+                        f"'polymorphic_on' column of "
+                        f"'{self.inherits.base_mapper.polymorphic_on}'. "
+                        "If this is an intermediary class that should not be "
+                        "instantiated, the class may either be left unmapped, "
+                        "or may include the 'polymorphic_abstract=True' "
+                        "parameter in its Mapper arguments. To leave the "
+                        "class unmapped when using Declarative, set the "
+                        "'__abstract__ = True' attribute on the class."
+                    )
+            elif self.concrete:
+                self._identity_class = self.class_
+            else:
+                self._identity_class = self.inherits._identity_class
+
+            if self.version_id_col is None:
+                self.version_id_col = self.inherits.version_id_col
+                self.version_id_generator = self.inherits.version_id_generator
+            elif (
+                self.inherits.version_id_col is not None
+                and self.version_id_col is not self.inherits.version_id_col
+            ):
+                util.warn(
+                    "Inheriting version_id_col '%s' does not match inherited "
+                    "version_id_col '%s' and will not automatically populate "
+                    "the inherited versioning column. "
+                    "version_id_col should only be specified on "
+                    "the base-most mapper that includes versioning."
+                    % (
+                        self.version_id_col.description,
+                        self.inherits.version_id_col.description,
+                    )
+                )
+
+            self.polymorphic_map = self.inherits.polymorphic_map
+            self.batch = self.inherits.batch
+            self.inherits._inheriting_mappers.append(self)
+            self.base_mapper = self.inherits.base_mapper
+            self.passive_updates = self.inherits.passive_updates
+            self.passive_deletes = (
+                self.inherits.passive_deletes or self.passive_deletes
+            )
+            self._all_tables = self.inherits._all_tables
+
+            if self.polymorphic_identity is not None:
+                if self.polymorphic_identity in self.polymorphic_map:
+                    util.warn(
+                        "Reassigning polymorphic association for identity %r "
+                        "from %r to %r: Check for duplicate use of %r as "
+                        "value for polymorphic_identity."
+                        % (
+                            self.polymorphic_identity,
+                            self.polymorphic_map[self.polymorphic_identity],
+                            self,
+                            self.polymorphic_identity,
+                        )
+                    )
+                self.polymorphic_map[self.polymorphic_identity] = self
+
+            if self.polymorphic_load and self.concrete:
+                raise sa_exc.ArgumentError(
+                    "polymorphic_load is not currently supported "
+                    "with concrete table inheritance"
+                )
+            if self.polymorphic_load == "inline":
+                self.inherits._add_with_polymorphic_subclass(self)
+            elif self.polymorphic_load == "selectin":
+                pass
+            elif self.polymorphic_load is not None:
+                raise sa_exc.ArgumentError(
+                    "unknown argument for polymorphic_load: %r"
+                    % self.polymorphic_load
+                )
+
+        else:
+            self._all_tables = set()
+            self.base_mapper = self
+            assert self.local_table is not None
+            self.persist_selectable = self.local_table
+            if self.polymorphic_identity is not None:
+                self.polymorphic_map[self.polymorphic_identity] = self
+            self._identity_class = self.class_
+
+        if self.persist_selectable is None:
+            raise sa_exc.ArgumentError(
+                "Mapper '%s' does not have a persist_selectable specified."
+                % self
+            )
+
+    def _set_with_polymorphic(
+        self, with_polymorphic: Optional[_WithPolymorphicArg]
+    ) -> None:
+        if with_polymorphic == "*":
+            self.with_polymorphic = ("*", None)
+        elif isinstance(with_polymorphic, (tuple, list)):
+            if isinstance(with_polymorphic[0], (str, tuple, list)):
+                self.with_polymorphic = cast(
+                    """Tuple[
+                        Union[
+                            Literal["*"],
+                            Sequence[Union["Mapper[Any]", Type[Any]]],
+                        ],
+                        Optional["FromClause"],
+                    ]""",
+                    with_polymorphic,
+                )
+            else:
+                self.with_polymorphic = (with_polymorphic, None)
+        elif with_polymorphic is not None:
+            raise sa_exc.ArgumentError(
+                f"Invalid setting for with_polymorphic: {with_polymorphic!r}"
+            )
+        else:
+            self.with_polymorphic = None
+
+        if self.with_polymorphic and self.with_polymorphic[1] is not None:
+            self.with_polymorphic = (
+                self.with_polymorphic[0],
+                coercions.expect(
+                    roles.StrictFromClauseRole,
+                    self.with_polymorphic[1],
+                    allow_select=True,
+                ),
+            )
+
+        if self.configured:
+            self._expire_memoizations()
+
+    def _add_with_polymorphic_subclass(self, mapper):
+        subcl = mapper.class_
+        if self.with_polymorphic is None:
+            self._set_with_polymorphic((subcl,))
+        elif self.with_polymorphic[0] != "*":
+            assert isinstance(self.with_polymorphic[0], tuple)
+            self._set_with_polymorphic(
+                (self.with_polymorphic[0] + (subcl,), self.with_polymorphic[1])
+            )
+
+    def _set_concrete_base(self, mapper):
+        """Set the given :class:`_orm.Mapper` as the 'inherits' for this
+        :class:`_orm.Mapper`, assuming this :class:`_orm.Mapper` is concrete
+        and does not already have an inherits."""
+
+        assert self.concrete
+        assert not self.inherits
+        assert isinstance(mapper, Mapper)
+        self.inherits = mapper
+        self.inherits.polymorphic_map.update(self.polymorphic_map)
+        self.polymorphic_map = self.inherits.polymorphic_map
+        for mapper in self.iterate_to_root():
+            if mapper.polymorphic_on is not None:
+                mapper._requires_row_aliasing = True
+        self.batch = self.inherits.batch
+        for mp in self.self_and_descendants:
+            mp.base_mapper = self.inherits.base_mapper
+        self.inherits._inheriting_mappers.append(self)
+        self.passive_updates = self.inherits.passive_updates
+        self._all_tables = self.inherits._all_tables
+
+        for key, prop in mapper._props.items():
+            if key not in self._props and not self._should_exclude(
+                key, key, local=False, column=None
+            ):
+                self._adapt_inherited_property(key, prop, False)
+
+    def _set_polymorphic_on(self, polymorphic_on):
+        self.polymorphic_on = polymorphic_on
+        self._configure_polymorphic_setter(True)
+
+    def _configure_class_instrumentation(self):
+        """If this mapper is to be a primary mapper (i.e. the
+        non_primary flag is not set), associate this Mapper with the
+        given class and entity name.
+
+        Subsequent calls to ``class_mapper()`` for the ``class_`` / ``entity``
+        name combination will return this mapper.  Also decorate the
+        `__init__` method on the mapped class to include optional
+        auto-session attachment logic.
+
+        """
+
+        # we expect that declarative has applied the class manager
+        # already and set up a registry.  if this is None,
+        # this raises as of 2.0.
+        manager = attributes.opt_manager_of_class(self.class_)
+
+        if self.non_primary:
+            if not manager or not manager.is_mapped:
+                raise sa_exc.InvalidRequestError(
+                    "Class %s has no primary mapper configured.  Configure "
+                    "a primary mapper first before setting up a non primary "
+                    "Mapper." % self.class_
+                )
+            self.class_manager = manager
+
+            assert manager.registry is not None
+            self.registry = manager.registry
+            self._identity_class = manager.mapper._identity_class
+            manager.registry._add_non_primary_mapper(self)
+            return
+
+        if manager is None or not manager.registry:
+            raise sa_exc.InvalidRequestError(
+                "The _mapper() function and Mapper() constructor may not be "
+                "invoked directly outside of a declarative registry."
+                " Please use the sqlalchemy.orm.registry.map_imperatively() "
+                "function for a classical mapping."
+            )
+
+        self.dispatch.instrument_class(self, self.class_)
+
+        # this invokes the class_instrument event and sets up
+        # the __init__ method.  documented behavior is that this must
+        # occur after the instrument_class event above.
+        # yes two events with the same two words reversed and different APIs.
+        # :(
+
+        manager = instrumentation.register_class(
+            self.class_,
+            mapper=self,
+            expired_attribute_loader=util.partial(
+                loading.load_scalar_attributes, self
+            ),
+            # finalize flag means instrument the __init__ method
+            # and call the class_instrument event
+            finalize=True,
+        )
+
+        self.class_manager = manager
+
+        assert manager.registry is not None
+        self.registry = manager.registry
+
+        # The remaining members can be added by any mapper,
+        # e_name None or not.
+        if manager.mapper is None:
+            return
+
+        event.listen(manager, "init", _event_on_init, raw=True)
+
+        for key, method in util.iterate_attributes(self.class_):
+            if key == "__init__" and hasattr(method, "_sa_original_init"):
+                method = method._sa_original_init
+                if hasattr(method, "__func__"):
+                    method = method.__func__
+            if callable(method):
+                if hasattr(method, "__sa_reconstructor__"):
+                    self._reconstructor = method
+                    event.listen(manager, "load", _event_on_load, raw=True)
+                elif hasattr(method, "__sa_validators__"):
+                    validation_opts = method.__sa_validation_opts__
+                    for name in method.__sa_validators__:
+                        if name in self.validators:
+                            raise sa_exc.InvalidRequestError(
+                                "A validation function for mapped "
+                                "attribute %r on mapper %s already exists."
+                                % (name, self)
+                            )
+                        self.validators = self.validators.union(
+                            {name: (method, validation_opts)}
+                        )
+
+    def _set_dispose_flags(self) -> None:
+        self.configured = True
+        self._ready_for_configure = True
+        self._dispose_called = True
+
+        self.__dict__.pop("_configure_failed", None)
+
+    def _str_arg_to_mapped_col(self, argname: str, key: str) -> Column[Any]:
+        try:
+            prop = self._props[key]
+        except KeyError as err:
+            raise sa_exc.ArgumentError(
+                f"Can't determine {argname} column '{key}' - "
+                "no attribute is mapped to this name."
+            ) from err
+        try:
+            expr = prop.expression
+        except AttributeError as ae:
+            raise sa_exc.ArgumentError(
+                f"Can't determine {argname} column '{key}'; "
+                "property does not refer to a single mapped Column"
+            ) from ae
+        if not isinstance(expr, Column):
+            raise sa_exc.ArgumentError(
+                f"Can't determine {argname} column '{key}'; "
+                "property does not refer to a single "
+                "mapped Column"
+            )
+        return expr
+
+    def _configure_pks(self) -> None:
+        self.tables = sql_util.find_tables(self.persist_selectable)
+
+        self._all_tables.update(t for t in self.tables)
+
+        self._pks_by_table = {}
+        self._cols_by_table = {}
+
+        all_cols = util.column_set(
+            chain(*[col.proxy_set for col in self._columntoproperty])
+        )
+
+        pk_cols = util.column_set(c for c in all_cols if c.primary_key)
+
+        # identify primary key columns which are also mapped by this mapper.
+        for fc in set(self.tables).union([self.persist_selectable]):
+            if fc.primary_key and pk_cols.issuperset(fc.primary_key):
+                # ordering is important since it determines the ordering of
+                # mapper.primary_key (and therefore query.get())
+                self._pks_by_table[fc] = util.ordered_column_set(  # type: ignore  # noqa: E501
+                    fc.primary_key
+                ).intersection(
+                    pk_cols
+                )
+            self._cols_by_table[fc] = util.ordered_column_set(fc.c).intersection(  # type: ignore  # noqa: E501
+                all_cols
+            )
+
+        if self._primary_key_argument:
+            coerced_pk_arg = [
+                (
+                    self._str_arg_to_mapped_col("primary_key", c)
+                    if isinstance(c, str)
+                    else c
+                )
+                for c in (
+                    coercions.expect(
+                        roles.DDLConstraintColumnRole,
+                        coerce_pk,
+                        argname="primary_key",
+                    )
+                    for coerce_pk in self._primary_key_argument
+                )
+            ]
+        else:
+            coerced_pk_arg = None
+
+        # if explicit PK argument sent, add those columns to the
+        # primary key mappings
+        if coerced_pk_arg:
+            for k in coerced_pk_arg:
+                if k.table not in self._pks_by_table:
+                    self._pks_by_table[k.table] = util.OrderedSet()
+                self._pks_by_table[k.table].add(k)
+
+        # otherwise, see that we got a full PK for the mapped table
+        elif (
+            self.persist_selectable not in self._pks_by_table
+            or len(self._pks_by_table[self.persist_selectable]) == 0
+        ):
+            raise sa_exc.ArgumentError(
+                "Mapper %s could not assemble any primary "
+                "key columns for mapped table '%s'"
+                % (self, self.persist_selectable.description)
+            )
+        elif self.local_table not in self._pks_by_table and isinstance(
+            self.local_table, schema.Table
+        ):
+            util.warn(
+                "Could not assemble any primary "
+                "keys for locally mapped table '%s' - "
+                "no rows will be persisted in this Table."
+                % self.local_table.description
+            )
+
+        if (
+            self.inherits
+            and not self.concrete
+            and not self._primary_key_argument
+        ):
+            # if inheriting, the "primary key" for this mapper is
+            # that of the inheriting (unless concrete or explicit)
+            self.primary_key = self.inherits.primary_key
+        else:
+            # determine primary key from argument or persist_selectable pks
+            primary_key: Collection[ColumnElement[Any]]
+
+            if coerced_pk_arg:
+                primary_key = [
+                    cc if cc is not None else c
+                    for cc, c in (
+                        (self.persist_selectable.corresponding_column(c), c)
+                        for c in coerced_pk_arg
+                    )
+                ]
+            else:
+                # if heuristically determined PKs, reduce to the minimal set
+                # of columns by eliminating FK->PK pairs for a multi-table
+                # expression.   May over-reduce for some kinds of UNIONs
+                # / CTEs; use explicit PK argument for these special cases
+                primary_key = sql_util.reduce_columns(
+                    self._pks_by_table[self.persist_selectable],
+                    ignore_nonexistent_tables=True,
+                )
+
+            if len(primary_key) == 0:
+                raise sa_exc.ArgumentError(
+                    "Mapper %s could not assemble any primary "
+                    "key columns for mapped table '%s'"
+                    % (self, self.persist_selectable.description)
+                )
+
+            self.primary_key = tuple(primary_key)
+            self._log("Identified primary key columns: %s", primary_key)
+
+        # determine cols that aren't expressed within our tables; mark these
+        # as "read only" properties which are refreshed upon INSERT/UPDATE
+        self._readonly_props = {
+            self._columntoproperty[col]
+            for col in self._columntoproperty
+            if self._columntoproperty[col] not in self._identity_key_props
+            and (
+                not hasattr(col, "table")
+                or col.table not in self._cols_by_table
+            )
+        }
+
+    def _configure_properties(self) -> None:
+        self.columns = self.c = sql_base.ColumnCollection()  # type: ignore
+
+        # object attribute names mapped to MapperProperty objects
+        self._props = util.OrderedDict()
+
+        # table columns mapped to MapperProperty
+        self._columntoproperty = _ColumnMapping(self)
+
+        explicit_col_props_by_column: Dict[
+            KeyedColumnElement[Any], Tuple[str, ColumnProperty[Any]]
+        ] = {}
+        explicit_col_props_by_key: Dict[str, ColumnProperty[Any]] = {}
+
+        # step 1: go through properties that were explicitly passed
+        # in the properties dictionary.  For Columns that are local, put them
+        # aside in a separate collection we will reconcile with the Table
+        # that's given.  For other properties, set them up in _props now.
+        if self._init_properties:
+            for key, prop_arg in self._init_properties.items():
+                if not isinstance(prop_arg, MapperProperty):
+                    possible_col_prop = self._make_prop_from_column(
+                        key, prop_arg
+                    )
+                else:
+                    possible_col_prop = prop_arg
+
+                # issue #8705.  if the explicit property is actually a
+                # Column that is local to the local Table, don't set it up
+                # in ._props yet, integrate it into the order given within
+                # the Table.
+
+                _map_as_property_now = True
+                if isinstance(possible_col_prop, properties.ColumnProperty):
+                    for given_col in possible_col_prop.columns:
+                        if self.local_table.c.contains_column(given_col):
+                            _map_as_property_now = False
+                            explicit_col_props_by_key[key] = possible_col_prop
+                            explicit_col_props_by_column[given_col] = (
+                                key,
+                                possible_col_prop,
+                            )
+
+                if _map_as_property_now:
+                    self._configure_property(
+                        key,
+                        possible_col_prop,
+                        init=False,
+                    )
+
+        # step 2: pull properties from the inherited mapper.  reconcile
+        # columns with those which are explicit above.  for properties that
+        # are only in the inheriting mapper, set them up as local props
+        if self.inherits:
+            for key, inherited_prop in self.inherits._props.items():
+                if self._should_exclude(key, key, local=False, column=None):
+                    continue
+
+                incoming_prop = explicit_col_props_by_key.get(key)
+                if incoming_prop:
+                    new_prop = self._reconcile_prop_with_incoming_columns(
+                        key,
+                        inherited_prop,
+                        warn_only=False,
+                        incoming_prop=incoming_prop,
+                    )
+                    explicit_col_props_by_key[key] = new_prop
+
+                    for inc_col in incoming_prop.columns:
+                        explicit_col_props_by_column[inc_col] = (
+                            key,
+                            new_prop,
+                        )
+                elif key not in self._props:
+                    self._adapt_inherited_property(key, inherited_prop, False)
+
+        # step 3.  Iterate through all columns in the persist selectable.
+        # this includes not only columns in the local table / fromclause,
+        # but also those columns in the superclass table if we are joined
+        # inh or single inh mapper.  map these columns as well. additional
+        # reconciliation against inherited columns occurs here also.
+
+        for column in self.persist_selectable.columns:
+            if column in explicit_col_props_by_column:
+                # column was explicitly passed to properties; configure
+                # it now in the order in which it corresponds to the
+                # Table / selectable
+                key, prop = explicit_col_props_by_column[column]
+                self._configure_property(key, prop, init=False)
+                continue
+
+            elif column in self._columntoproperty:
+                continue
+
+            column_key = (self.column_prefix or "") + column.key
+            if self._should_exclude(
+                column.key,
+                column_key,
+                local=self.local_table.c.contains_column(column),
+                column=column,
+            ):
+                continue
+
+            # adjust the "key" used for this column to that
+            # of the inheriting mapper
+            for mapper in self.iterate_to_root():
+                if column in mapper._columntoproperty:
+                    column_key = mapper._columntoproperty[column].key
+
+            self._configure_property(
+                column_key,
+                column,
+                init=False,
+                setparent=True,
+            )
+
+    def _configure_polymorphic_setter(self, init=False):
+        """Configure an attribute on the mapper representing the
+        'polymorphic_on' column, if applicable, and not
+        already generated by _configure_properties (which is typical).
+
+        Also create a setter function which will assign this
+        attribute to the value of the 'polymorphic_identity'
+        upon instance construction, also if applicable.  This
+        routine will run when an instance is created.
+
+        """
+        setter = False
+        polymorphic_key: Optional[str] = None
+
+        if self.polymorphic_on is not None:
+            setter = True
+
+            if isinstance(self.polymorphic_on, str):
+                # polymorphic_on specified as a string - link
+                # it to mapped ColumnProperty
+                try:
+                    self.polymorphic_on = self._props[self.polymorphic_on]
+                except KeyError as err:
+                    raise sa_exc.ArgumentError(
+                        "Can't determine polymorphic_on "
+                        "value '%s' - no attribute is "
+                        "mapped to this name." % self.polymorphic_on
+                    ) from err
+
+            if self.polymorphic_on in self._columntoproperty:
+                # polymorphic_on is a column that is already mapped
+                # to a ColumnProperty
+                prop = self._columntoproperty[self.polymorphic_on]
+            elif isinstance(self.polymorphic_on, MapperProperty):
+                # polymorphic_on is directly a MapperProperty,
+                # ensure it's a ColumnProperty
+                if not isinstance(
+                    self.polymorphic_on, properties.ColumnProperty
+                ):
+                    raise sa_exc.ArgumentError(
+                        "Only direct column-mapped "
+                        "property or SQL expression "
+                        "can be passed for polymorphic_on"
+                    )
+                prop = self.polymorphic_on
+            else:
+                # polymorphic_on is a Column or SQL expression and
+                # doesn't appear to be mapped. this means it can be 1.
+                # only present in the with_polymorphic selectable or
+                # 2. a totally standalone SQL expression which we'd
+                # hope is compatible with this mapper's persist_selectable
+                col = self.persist_selectable.corresponding_column(
+                    self.polymorphic_on
+                )
+                if col is None:
+                    # polymorphic_on doesn't derive from any
+                    # column/expression isn't present in the mapped
+                    # table. we will make a "hidden" ColumnProperty
+                    # for it. Just check that if it's directly a
+                    # schema.Column and we have with_polymorphic, it's
+                    # likely a user error if the schema.Column isn't
+                    # represented somehow in either persist_selectable or
+                    # with_polymorphic.   Otherwise as of 0.7.4 we
+                    # just go with it and assume the user wants it
+                    # that way (i.e. a CASE statement)
+                    setter = False
+                    instrument = False
+                    col = self.polymorphic_on
+                    if isinstance(col, schema.Column) and (
+                        self.with_polymorphic is None
+                        or self.with_polymorphic[1] is None
+                        or self.with_polymorphic[1].corresponding_column(col)
+                        is None
+                    ):
+                        raise sa_exc.InvalidRequestError(
+                            "Could not map polymorphic_on column "
+                            "'%s' to the mapped table - polymorphic "
+                            "loads will not function properly"
+                            % col.description
+                        )
+                else:
+                    # column/expression that polymorphic_on derives from
+                    # is present in our mapped table
+                    # and is probably mapped, but polymorphic_on itself
+                    # is not.  This happens when
+                    # the polymorphic_on is only directly present in the
+                    # with_polymorphic selectable, as when use
+                    # polymorphic_union.
+                    # we'll make a separate ColumnProperty for it.
+                    instrument = True
+                key = getattr(col, "key", None)
+                if key:
+                    if self._should_exclude(key, key, False, col):
+                        raise sa_exc.InvalidRequestError(
+                            "Cannot exclude or override the "
+                            "discriminator column %r" % key
+                        )
+                else:
+                    self.polymorphic_on = col = col.label("_sa_polymorphic_on")
+                    key = col.key
+
+                prop = properties.ColumnProperty(col, _instrument=instrument)
+                self._configure_property(key, prop, init=init, setparent=True)
+
+            # the actual polymorphic_on should be the first public-facing
+            # column in the property
+            self.polymorphic_on = prop.columns[0]
+            polymorphic_key = prop.key
+        else:
+            # no polymorphic_on was set.
+            # check inheriting mappers for one.
+            for mapper in self.iterate_to_root():
+                # determine if polymorphic_on of the parent
+                # should be propagated here.   If the col
+                # is present in our mapped table, or if our mapped
+                # table is the same as the parent (i.e. single table
+                # inheritance), we can use it
+                if mapper.polymorphic_on is not None:
+                    if self.persist_selectable is mapper.persist_selectable:
+                        self.polymorphic_on = mapper.polymorphic_on
+                    else:
+                        self.polymorphic_on = (
+                            self.persist_selectable
+                        ).corresponding_column(mapper.polymorphic_on)
+                    # we can use the parent mapper's _set_polymorphic_identity
+                    # directly; it ensures the polymorphic_identity of the
+                    # instance's mapper is used so is portable to subclasses.
+                    if self.polymorphic_on is not None:
+                        self._set_polymorphic_identity = (
+                            mapper._set_polymorphic_identity
+                        )
+                        self._polymorphic_attr_key = (
+                            mapper._polymorphic_attr_key
+                        )
+                        self._validate_polymorphic_identity = (
+                            mapper._validate_polymorphic_identity
+                        )
+                    else:
+                        self._set_polymorphic_identity = None
+                        self._polymorphic_attr_key = None
+                    return
+
+        if self.polymorphic_abstract and self.polymorphic_on is None:
+            raise sa_exc.InvalidRequestError(
+                "The Mapper.polymorphic_abstract parameter may only be used "
+                "on a mapper hierarchy which includes the "
+                "Mapper.polymorphic_on parameter at the base of the hierarchy."
+            )
+
+        if setter:
+
+            def _set_polymorphic_identity(state):
+                dict_ = state.dict
+                # TODO: what happens if polymorphic_on column attribute name
+                # does not match .key?
+
+                polymorphic_identity = (
+                    state.manager.mapper.polymorphic_identity
+                )
+                if (
+                    polymorphic_identity is None
+                    and state.manager.mapper.polymorphic_abstract
+                ):
+                    raise sa_exc.InvalidRequestError(
+                        f"Can't instantiate class for {state.manager.mapper}; "
+                        "mapper is marked polymorphic_abstract=True"
+                    )
+
+                state.get_impl(polymorphic_key).set(
+                    state,
+                    dict_,
+                    polymorphic_identity,
+                    None,
+                )
+
+            self._polymorphic_attr_key = polymorphic_key
+
+            def _validate_polymorphic_identity(mapper, state, dict_):
+                if (
+                    polymorphic_key in dict_
+                    and dict_[polymorphic_key]
+                    not in mapper._acceptable_polymorphic_identities
+                ):
+                    util.warn_limited(
+                        "Flushing object %s with "
+                        "incompatible polymorphic identity %r; the "
+                        "object may not refresh and/or load correctly",
+                        (state_str(state), dict_[polymorphic_key]),
+                    )
+
+            self._set_polymorphic_identity = _set_polymorphic_identity
+            self._validate_polymorphic_identity = (
+                _validate_polymorphic_identity
+            )
+        else:
+            self._polymorphic_attr_key = None
+            self._set_polymorphic_identity = None
+
+    _validate_polymorphic_identity = None
+
+    @HasMemoized.memoized_attribute
+    def _version_id_prop(self):
+        if self.version_id_col is not None:
+            return self._columntoproperty[self.version_id_col]
+        else:
+            return None
+
+    @HasMemoized.memoized_attribute
+    def _acceptable_polymorphic_identities(self):
+        identities = set()
+
+        stack = deque([self])
+        while stack:
+            item = stack.popleft()
+            if item.persist_selectable is self.persist_selectable:
+                identities.add(item.polymorphic_identity)
+                stack.extend(item._inheriting_mappers)
+
+        return identities
+
+    @HasMemoized.memoized_attribute
+    def _prop_set(self):
+        return frozenset(self._props.values())
+
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def _adapt_inherited_property(self, key, prop, init):
+        descriptor_props = util.preloaded.orm_descriptor_props
+
+        if not self.concrete:
+            self._configure_property(key, prop, init=False, setparent=False)
+        elif key not in self._props:
+            # determine if the class implements this attribute; if not,
+            # or if it is implemented by the attribute that is handling the
+            # given superclass-mapped property, then we need to report that we
+            # can't use this at the instance level since we are a concrete
+            # mapper and we don't map this.  don't trip user-defined
+            # descriptors that might have side effects when invoked.
+            implementing_attribute = self.class_manager._get_class_attr_mro(
+                key, prop
+            )
+            if implementing_attribute is prop or (
+                isinstance(
+                    implementing_attribute, attributes.InstrumentedAttribute
+                )
+                and implementing_attribute._parententity is prop.parent
+            ):
+                self._configure_property(
+                    key,
+                    descriptor_props.ConcreteInheritedProperty(),
+                    init=init,
+                    setparent=True,
+                )
+
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def _configure_property(
+        self,
+        key: str,
+        prop_arg: Union[KeyedColumnElement[Any], MapperProperty[Any]],
+        *,
+        init: bool = True,
+        setparent: bool = True,
+        warn_for_existing: bool = False,
+    ) -> MapperProperty[Any]:
+        descriptor_props = util.preloaded.orm_descriptor_props
+        self._log(
+            "_configure_property(%s, %s)", key, prop_arg.__class__.__name__
+        )
+
+        if not isinstance(prop_arg, MapperProperty):
+            prop: MapperProperty[Any] = self._property_from_column(
+                key, prop_arg
+            )
+        else:
+            prop = prop_arg
+
+        if isinstance(prop, properties.ColumnProperty):
+            col = self.persist_selectable.corresponding_column(prop.columns[0])
+
+            # if the column is not present in the mapped table,
+            # test if a column has been added after the fact to the
+            # parent table (or their parent, etc.) [ticket:1570]
+            if col is None and self.inherits:
+                path = [self]
+                for m in self.inherits.iterate_to_root():
+                    col = m.local_table.corresponding_column(prop.columns[0])
+                    if col is not None:
+                        for m2 in path:
+                            m2.persist_selectable._refresh_for_new_column(col)
+                        col = self.persist_selectable.corresponding_column(
+                            prop.columns[0]
+                        )
+                        break
+                    path.append(m)
+
+            # subquery expression, column not present in the mapped
+            # selectable.
+            if col is None:
+                col = prop.columns[0]
+
+                # column is coming in after _readonly_props was
+                # initialized; check for 'readonly'
+                if hasattr(self, "_readonly_props") and (
+                    not hasattr(col, "table")
+                    or col.table not in self._cols_by_table
+                ):
+                    self._readonly_props.add(prop)
+
+            else:
+                # if column is coming in after _cols_by_table was
+                # initialized, ensure the col is in the right set
+                if (
+                    hasattr(self, "_cols_by_table")
+                    and col.table in self._cols_by_table
+                    and col not in self._cols_by_table[col.table]
+                ):
+                    self._cols_by_table[col.table].add(col)
+
+            # if this properties.ColumnProperty represents the "polymorphic
+            # discriminator" column, mark it.  We'll need this when rendering
+            # columns in SELECT statements.
+            if not hasattr(prop, "_is_polymorphic_discriminator"):
+                prop._is_polymorphic_discriminator = (
+                    col is self.polymorphic_on
+                    or prop.columns[0] is self.polymorphic_on
+                )
+
+            if isinstance(col, expression.Label):
+                # new in 1.4, get column property against expressions
+                # to be addressable in subqueries
+                col.key = col._tq_key_label = key
+
+            self.columns.add(col, key)
+
+            for col in prop.columns:
+                for proxy_col in col.proxy_set:
+                    self._columntoproperty[proxy_col] = prop
+
+        if getattr(prop, "key", key) != key:
+            util.warn(
+                f"ORM mapped property {self.class_.__name__}.{prop.key} being "
+                "assigned to attribute "
+                f"{key!r} is already associated with "
+                f"attribute {prop.key!r}. The attribute will be de-associated "
+                f"from {prop.key!r}."
+            )
+
+        prop.key = key
+
+        if setparent:
+            prop.set_parent(self, init)
+
+        if key in self._props and getattr(
+            self._props[key], "_mapped_by_synonym", False
+        ):
+            syn = self._props[key]._mapped_by_synonym
+            raise sa_exc.ArgumentError(
+                "Can't call map_column=True for synonym %r=%r, "
+                "a ColumnProperty already exists keyed to the name "
+                "%r for column %r" % (syn, key, key, syn)
+            )
+
+        # replacement cases
+
+        # case one: prop is replacing a prop that we have mapped.  this is
+        # independent of whatever might be in the actual class dictionary
+        if (
+            key in self._props
+            and not isinstance(
+                self._props[key], descriptor_props.ConcreteInheritedProperty
+            )
+            and not isinstance(prop, descriptor_props.SynonymProperty)
+        ):
+            if warn_for_existing:
+                util.warn_deprecated(
+                    f"User-placed attribute {self.class_.__name__}.{key} on "
+                    f"{self} is replacing an existing ORM-mapped attribute.  "
+                    "Behavior is not fully defined in this case.  This "
+                    "use is deprecated and will raise an error in a future "
+                    "release",
+                    "2.0",
+                )
+            oldprop = self._props[key]
+            self._path_registry.pop(oldprop, None)
+
+        # case two: prop is replacing an attribute on the class of some kind.
+        # we have to be more careful here since it's normal when using
+        # Declarative that all the "declared attributes" on the class
+        # get replaced.
+        elif (
+            warn_for_existing
+            and self.class_.__dict__.get(key, None) is not None
+            and not isinstance(prop, descriptor_props.SynonymProperty)
+            and not isinstance(
+                self._props.get(key, None),
+                descriptor_props.ConcreteInheritedProperty,
+            )
+        ):
+            util.warn_deprecated(
+                f"User-placed attribute {self.class_.__name__}.{key} on "
+                f"{self} is replacing an existing class-bound "
+                "attribute of the same name.  "
+                "Behavior is not fully defined in this case.  This "
+                "use is deprecated and will raise an error in a future "
+                "release",
+                "2.0",
+            )
+
+        self._props[key] = prop
+
+        if not self.non_primary:
+            prop.instrument_class(self)
+
+        for mapper in self._inheriting_mappers:
+            mapper._adapt_inherited_property(key, prop, init)
+
+        if init:
+            prop.init()
+            prop.post_instrument_class(self)
+
+        if self.configured:
+            self._expire_memoizations()
+
+        return prop
+
+    def _make_prop_from_column(
+        self,
+        key: str,
+        column: Union[
+            Sequence[KeyedColumnElement[Any]], KeyedColumnElement[Any]
+        ],
+    ) -> ColumnProperty[Any]:
+        columns = util.to_list(column)
+        mapped_column = []
+        for c in columns:
+            mc = self.persist_selectable.corresponding_column(c)
+            if mc is None:
+                mc = self.local_table.corresponding_column(c)
+                if mc is not None:
+                    # if the column is in the local table but not the
+                    # mapped table, this corresponds to adding a
+                    # column after the fact to the local table.
+                    # [ticket:1523]
+                    self.persist_selectable._refresh_for_new_column(mc)
+                mc = self.persist_selectable.corresponding_column(c)
+                if mc is None:
+                    raise sa_exc.ArgumentError(
+                        "When configuring property '%s' on %s, "
+                        "column '%s' is not represented in the mapper's "
+                        "table. Use the `column_property()` function to "
+                        "force this column to be mapped as a read-only "
+                        "attribute." % (key, self, c)
+                    )
+            mapped_column.append(mc)
+        return properties.ColumnProperty(*mapped_column)
+
+    def _reconcile_prop_with_incoming_columns(
+        self,
+        key: str,
+        existing_prop: MapperProperty[Any],
+        warn_only: bool,
+        incoming_prop: Optional[ColumnProperty[Any]] = None,
+        single_column: Optional[KeyedColumnElement[Any]] = None,
+    ) -> ColumnProperty[Any]:
+        if incoming_prop and (
+            self.concrete
+            or not isinstance(existing_prop, properties.ColumnProperty)
+        ):
+            return incoming_prop
+
+        existing_column = existing_prop.columns[0]
+
+        if incoming_prop and existing_column in incoming_prop.columns:
+            return incoming_prop
+
+        if incoming_prop is None:
+            assert single_column is not None
+            incoming_column = single_column
+            equated_pair_key = (existing_prop.columns[0], incoming_column)
+        else:
+            assert single_column is None
+            incoming_column = incoming_prop.columns[0]
+            equated_pair_key = (incoming_column, existing_prop.columns[0])
+
+        if (
+            (
+                not self._inherits_equated_pairs
+                or (equated_pair_key not in self._inherits_equated_pairs)
+            )
+            and not existing_column.shares_lineage(incoming_column)
+            and existing_column is not self.version_id_col
+            and incoming_column is not self.version_id_col
+        ):
+            msg = (
+                "Implicitly combining column %s with column "
+                "%s under attribute '%s'.  Please configure one "
+                "or more attributes for these same-named columns "
+                "explicitly."
+                % (
+                    existing_prop.columns[-1],
+                    incoming_column,
+                    key,
+                )
+            )
+            if warn_only:
+                util.warn(msg)
+            else:
+                raise sa_exc.InvalidRequestError(msg)
+
+        # existing properties.ColumnProperty from an inheriting
+        # mapper. make a copy and append our column to it
+        # breakpoint()
+        new_prop = existing_prop.copy()
+
+        new_prop.columns.insert(0, incoming_column)
+        self._log(
+            "inserting column to existing list "
+            "in properties.ColumnProperty %s",
+            key,
+        )
+        return new_prop  # type: ignore
+
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def _property_from_column(
+        self,
+        key: str,
+        column: KeyedColumnElement[Any],
+    ) -> ColumnProperty[Any]:
+        """generate/update a :class:`.ColumnProperty` given a
+        :class:`_schema.Column` or other SQL expression object."""
+
+        descriptor_props = util.preloaded.orm_descriptor_props
+
+        prop = self._props.get(key)
+
+        if isinstance(prop, properties.ColumnProperty):
+            return self._reconcile_prop_with_incoming_columns(
+                key,
+                prop,
+                single_column=column,
+                warn_only=prop.parent is not self,
+            )
+        elif prop is None or isinstance(
+            prop, descriptor_props.ConcreteInheritedProperty
+        ):
+            return self._make_prop_from_column(key, column)
+        else:
+            raise sa_exc.ArgumentError(
+                "WARNING: when configuring property '%s' on %s, "
+                "column '%s' conflicts with property '%r'. "
+                "To resolve this, map the column to the class under a "
+                "different name in the 'properties' dictionary.  Or, "
+                "to remove all awareness of the column entirely "
+                "(including its availability as a foreign key), "
+                "use the 'include_properties' or 'exclude_properties' "
+                "mapper arguments to control specifically which table "
+                "columns get mapped." % (key, self, column.key, prop)
+            )
+
+    @util.langhelpers.tag_method_for_warnings(
+        "This warning originated from the `configure_mappers()` process, "
+        "which was invoked automatically in response to a user-initiated "
+        "operation.",
+        sa_exc.SAWarning,
+    )
+    def _check_configure(self) -> None:
+        if self.registry._new_mappers:
+            _configure_registries({self.registry}, cascade=True)
+
+    def _post_configure_properties(self) -> None:
+        """Call the ``init()`` method on all ``MapperProperties``
+        attached to this mapper.
+
+        This is a deferred configuration step which is intended
+        to execute once all mappers have been constructed.
+
+        """
+
+        self._log("_post_configure_properties() started")
+        l = [(key, prop) for key, prop in self._props.items()]
+        for key, prop in l:
+            self._log("initialize prop %s", key)
+
+            if prop.parent is self and not prop._configure_started:
+                prop.init()
+
+            if prop._configure_finished:
+                prop.post_instrument_class(self)
+
+        self._log("_post_configure_properties() complete")
+        self.configured = True
+
+    def add_properties(self, dict_of_properties):
+        """Add the given dictionary of properties to this mapper,
+        using `add_property`.
+
+        """
+        for key, value in dict_of_properties.items():
+            self.add_property(key, value)
+
+    def add_property(
+        self, key: str, prop: Union[Column[Any], MapperProperty[Any]]
+    ) -> None:
+        """Add an individual MapperProperty to this mapper.
+
+        If the mapper has not been configured yet, just adds the
+        property to the initial properties dictionary sent to the
+        constructor.  If this Mapper has already been configured, then
+        the given MapperProperty is configured immediately.
+
+        """
+        prop = self._configure_property(
+            key, prop, init=self.configured, warn_for_existing=True
+        )
+        assert isinstance(prop, MapperProperty)
+        self._init_properties[key] = prop
+
+    def _expire_memoizations(self) -> None:
+        for mapper in self.iterate_to_root():
+            mapper._reset_memoizations()
+
+    @property
+    def _log_desc(self) -> str:
+        return (
+            "("
+            + self.class_.__name__
+            + "|"
+            + (
+                self.local_table is not None
+                and self.local_table.description
+                or str(self.local_table)
+            )
+            + (self.non_primary and "|non-primary" or "")
+            + ")"
+        )
+
+    def _log(self, msg: str, *args: Any) -> None:
+        self.logger.info("%s " + msg, *((self._log_desc,) + args))
+
+    def _log_debug(self, msg: str, *args: Any) -> None:
+        self.logger.debug("%s " + msg, *((self._log_desc,) + args))
+
+    def __repr__(self) -> str:
+        return "<Mapper at 0x%x; %s>" % (id(self), self.class_.__name__)
+
+    def __str__(self) -> str:
+        return "Mapper[%s%s(%s)]" % (
+            self.class_.__name__,
+            self.non_primary and " (non-primary)" or "",
+            (
+                self.local_table.description
+                if self.local_table is not None
+                else self.persist_selectable.description
+            ),
+        )
+
+    def _is_orphan(self, state: InstanceState[_O]) -> bool:
+        orphan_possible = False
+        for mapper in self.iterate_to_root():
+            for key, cls in mapper._delete_orphans:
+                orphan_possible = True
+
+                has_parent = attributes.manager_of_class(cls).has_parent(
+                    state, key, optimistic=state.has_identity
+                )
+
+                if self.legacy_is_orphan and has_parent:
+                    return False
+                elif not self.legacy_is_orphan and not has_parent:
+                    return True
+
+        if self.legacy_is_orphan:
+            return orphan_possible
+        else:
+            return False
+
+    def has_property(self, key: str) -> bool:
+        return key in self._props
+
+    def get_property(
+        self, key: str, _configure_mappers: bool = False
+    ) -> MapperProperty[Any]:
+        """return a MapperProperty associated with the given key."""
+
+        if _configure_mappers:
+            self._check_configure()
+
+        try:
+            return self._props[key]
+        except KeyError as err:
+            raise sa_exc.InvalidRequestError(
+                f"Mapper '{self}' has no property '{key}'.  If this property "
+                "was indicated from other mappers or configure events, ensure "
+                "registry.configure() has been called."
+            ) from err
+
+    def get_property_by_column(
+        self, column: ColumnElement[_T]
+    ) -> MapperProperty[_T]:
+        """Given a :class:`_schema.Column` object, return the
+        :class:`.MapperProperty` which maps this column."""
+
+        return self._columntoproperty[column]
+
+    @property
+    def iterate_properties(self):
+        """return an iterator of all MapperProperty objects."""
+
+        return iter(self._props.values())
+
+    def _mappers_from_spec(
+        self, spec: Any, selectable: Optional[FromClause]
+    ) -> Sequence[Mapper[Any]]:
+        """given a with_polymorphic() argument, return the set of mappers it
+        represents.
+
+        Trims the list of mappers to just those represented within the given
+        selectable, if present. This helps some more legacy-ish mappings.
+
+        """
+        if spec == "*":
+            mappers = list(self.self_and_descendants)
+        elif spec:
+            mapper_set = set()
+            for m in util.to_list(spec):
+                m = _class_to_mapper(m)
+                if not m.isa(self):
+                    raise sa_exc.InvalidRequestError(
+                        "%r does not inherit from %r" % (m, self)
+                    )
+
+                if selectable is None:
+                    mapper_set.update(m.iterate_to_root())
+                else:
+                    mapper_set.add(m)
+            mappers = [m for m in self.self_and_descendants if m in mapper_set]
+        else:
+            mappers = []
+
+        if selectable is not None:
+            tables = set(
+                sql_util.find_tables(selectable, include_aliases=True)
+            )
+            mappers = [m for m in mappers if m.local_table in tables]
+        return mappers
+
+    def _selectable_from_mappers(
+        self, mappers: Iterable[Mapper[Any]], innerjoin: bool
+    ) -> FromClause:
+        """given a list of mappers (assumed to be within this mapper's
+        inheritance hierarchy), construct an outerjoin amongst those mapper's
+        mapped tables.
+
+        """
+        from_obj = self.persist_selectable
+        for m in mappers:
+            if m is self:
+                continue
+            if m.concrete:
+                raise sa_exc.InvalidRequestError(
+                    "'with_polymorphic()' requires 'selectable' argument "
+                    "when concrete-inheriting mappers are used."
+                )
+            elif not m.single:
+                if innerjoin:
+                    from_obj = from_obj.join(
+                        m.local_table, m.inherit_condition
+                    )
+                else:
+                    from_obj = from_obj.outerjoin(
+                        m.local_table, m.inherit_condition
+                    )
+
+        return from_obj
+
+    @HasMemoized.memoized_attribute
+    def _version_id_has_server_side_value(self) -> bool:
+        vid_col = self.version_id_col
+
+        if vid_col is None:
+            return False
+
+        elif not isinstance(vid_col, Column):
+            return True
+        else:
+            return vid_col.server_default is not None or (
+                vid_col.default is not None
+                and (
+                    not vid_col.default.is_scalar
+                    and not vid_col.default.is_callable
+                )
+            )
+
+    @HasMemoized.memoized_attribute
+    def _single_table_criterion(self):
+        if self.single and self.inherits and self.polymorphic_on is not None:
+            return self.polymorphic_on._annotate(
+                {"parententity": self, "parentmapper": self}
+            ).in_(
+                [
+                    m.polymorphic_identity
+                    for m in self.self_and_descendants
+                    if not m.polymorphic_abstract
+                ]
+            )
+        else:
+            return None
+
+    @HasMemoized.memoized_attribute
+    def _has_aliased_polymorphic_fromclause(self):
+        """return True if with_polymorphic[1] is an aliased fromclause,
+        like a subquery.
+
+        As of #8168, polymorphic adaption with ORMAdapter is used only
+        if this is present.
+
+        """
+        return self.with_polymorphic and isinstance(
+            self.with_polymorphic[1],
+            expression.AliasedReturnsRows,
+        )
+
+    @HasMemoized.memoized_attribute
+    def _should_select_with_poly_adapter(self):
+        """determine if _MapperEntity or _ORMColumnEntity will need to use
+        polymorphic adaption when setting up a SELECT as well as fetching
+        rows for mapped classes and subclasses against this Mapper.
+
+        moved here from context.py for #8456 to generalize the ruleset
+        for this condition.
+
+        """
+
+        # this has been simplified as of #8456.
+        # rule is: if we have a with_polymorphic or a concrete-style
+        # polymorphic selectable, *or* if the base mapper has either of those,
+        # we turn on the adaption thing.  if not, we do *no* adaption.
+        #
+        # (UPDATE for #8168: the above comment was not accurate, as we were
+        # still saying "do polymorphic" if we were using an auto-generated
+        # flattened JOIN for with_polymorphic.)
+        #
+        # this splits the behavior among the "regular" joined inheritance
+        # and single inheritance mappers, vs. the "weird / difficult"
+        # concrete and joined inh mappings that use a with_polymorphic of
+        # some kind or polymorphic_union.
+        #
+        # note we have some tests in test_polymorphic_rel that query against
+        # a subclass, then refer to the superclass that has a with_polymorphic
+        # on it (such as test_join_from_polymorphic_explicit_aliased_three).
+        # these tests actually adapt the polymorphic selectable (like, the
+        # UNION or the SELECT subquery with JOIN in it) to be just the simple
+        # subclass table.   Hence even if we are a "plain" inheriting mapper
+        # but our base has a wpoly on it, we turn on adaption.  This is a
+        # legacy case we should probably disable.
+        #
+        #
+        # UPDATE: simplified way more as of #8168.   polymorphic adaption
+        # is turned off even if with_polymorphic is set, as long as there
+        # is no user-defined aliased selectable / subquery configured.
+        # this scales back the use of polymorphic adaption in practice
+        # to basically no cases except for concrete inheritance with a
+        # polymorphic base class.
+        #
+        return (
+            self._has_aliased_polymorphic_fromclause
+            or self._requires_row_aliasing
+            or (self.base_mapper._has_aliased_polymorphic_fromclause)
+            or self.base_mapper._requires_row_aliasing
+        )
+
+    @HasMemoized.memoized_attribute
+    def _with_polymorphic_mappers(self) -> Sequence[Mapper[Any]]:
+        self._check_configure()
+
+        if not self.with_polymorphic:
+            return []
+        return self._mappers_from_spec(*self.with_polymorphic)
+
+    @HasMemoized.memoized_attribute
+    def _post_inspect(self):
+        """This hook is invoked by attribute inspection.
+
+        E.g. when Query calls:
+
+            coercions.expect(roles.ColumnsClauseRole, ent, keep_inspect=True)
+
+        This allows the inspection process run a configure mappers hook.
+
+        """
+        self._check_configure()
+
+    @HasMemoized_ro_memoized_attribute
+    def _with_polymorphic_selectable(self) -> FromClause:
+        if not self.with_polymorphic:
+            return self.persist_selectable
+
+        spec, selectable = self.with_polymorphic
+        if selectable is not None:
+            return selectable
+        else:
+            return self._selectable_from_mappers(
+                self._mappers_from_spec(spec, selectable), False
+            )
+
+    with_polymorphic_mappers = _with_polymorphic_mappers
+    """The list of :class:`_orm.Mapper` objects included in the
+    default "polymorphic" query.
+
+    """
+
+    @HasMemoized_ro_memoized_attribute
+    def _insert_cols_evaluating_none(self):
+        return {
+            table: frozenset(
+                col for col in columns if col.type.should_evaluate_none
+            )
+            for table, columns in self._cols_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _insert_cols_as_none(self):
+        return {
+            table: frozenset(
+                col.key
+                for col in columns
+                if not col.primary_key
+                and not col.server_default
+                and not col.default
+                and not col.type.should_evaluate_none
+            )
+            for table, columns in self._cols_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _propkey_to_col(self):
+        return {
+            table: {self._columntoproperty[col].key: col for col in columns}
+            for table, columns in self._cols_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _pk_keys_by_table(self):
+        return {
+            table: frozenset([col.key for col in pks])
+            for table, pks in self._pks_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _pk_attr_keys_by_table(self):
+        return {
+            table: frozenset([self._columntoproperty[col].key for col in pks])
+            for table, pks in self._pks_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _server_default_cols(
+        self,
+    ) -> Mapping[FromClause, FrozenSet[Column[Any]]]:
+        return {
+            table: frozenset(
+                [
+                    col
+                    for col in cast("Iterable[Column[Any]]", columns)
+                    if col.server_default is not None
+                    or (
+                        col.default is not None
+                        and col.default.is_clause_element
+                    )
+                ]
+            )
+            for table, columns in self._cols_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _server_onupdate_default_cols(
+        self,
+    ) -> Mapping[FromClause, FrozenSet[Column[Any]]]:
+        return {
+            table: frozenset(
+                [
+                    col
+                    for col in cast("Iterable[Column[Any]]", columns)
+                    if col.server_onupdate is not None
+                    or (
+                        col.onupdate is not None
+                        and col.onupdate.is_clause_element
+                    )
+                ]
+            )
+            for table, columns in self._cols_by_table.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _server_default_col_keys(self) -> Mapping[FromClause, FrozenSet[str]]:
+        return {
+            table: frozenset(col.key for col in cols if col.key is not None)
+            for table, cols in self._server_default_cols.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _server_onupdate_default_col_keys(
+        self,
+    ) -> Mapping[FromClause, FrozenSet[str]]:
+        return {
+            table: frozenset(col.key for col in cols if col.key is not None)
+            for table, cols in self._server_onupdate_default_cols.items()
+        }
+
+    @HasMemoized.memoized_attribute
+    def _server_default_plus_onupdate_propkeys(self) -> Set[str]:
+        result: Set[str] = set()
+
+        col_to_property = self._columntoproperty
+        for table, columns in self._server_default_cols.items():
+            result.update(
+                col_to_property[col].key
+                for col in columns.intersection(col_to_property)
+            )
+        for table, columns in self._server_onupdate_default_cols.items():
+            result.update(
+                col_to_property[col].key
+                for col in columns.intersection(col_to_property)
+            )
+        return result
+
+    @HasMemoized.memoized_instancemethod
+    def __clause_element__(self):
+        annotations: Dict[str, Any] = {
+            "entity_namespace": self,
+            "parententity": self,
+            "parentmapper": self,
+        }
+        if self.persist_selectable is not self.local_table:
+            # joined table inheritance, with polymorphic selectable,
+            # etc.
+            annotations["dml_table"] = self.local_table._annotate(
+                {
+                    "entity_namespace": self,
+                    "parententity": self,
+                    "parentmapper": self,
+                }
+            )._set_propagate_attrs(
+                {"compile_state_plugin": "orm", "plugin_subject": self}
+            )
+
+        return self.selectable._annotate(annotations)._set_propagate_attrs(
+            {"compile_state_plugin": "orm", "plugin_subject": self}
+        )
+
+    @util.memoized_property
+    def select_identity_token(self):
+        return (
+            expression.null()
+            ._annotate(
+                {
+                    "entity_namespace": self,
+                    "parententity": self,
+                    "parentmapper": self,
+                    "identity_token": True,
+                }
+            )
+            ._set_propagate_attrs(
+                {"compile_state_plugin": "orm", "plugin_subject": self}
+            )
+        )
+
+    @property
+    def selectable(self) -> FromClause:
+        """The :class:`_schema.FromClause` construct this
+        :class:`_orm.Mapper` selects from by default.
+
+        Normally, this is equivalent to :attr:`.persist_selectable`, unless
+        the ``with_polymorphic`` feature is in use, in which case the
+        full "polymorphic" selectable is returned.
+
+        """
+        return self._with_polymorphic_selectable
+
+    def _with_polymorphic_args(
+        self,
+        spec: Any = None,
+        selectable: Union[Literal[False, None], FromClause] = False,
+        innerjoin: bool = False,
+    ) -> Tuple[Sequence[Mapper[Any]], FromClause]:
+        if selectable not in (None, False):
+            selectable = coercions.expect(
+                roles.StrictFromClauseRole, selectable, allow_select=True
+            )
+
+        if self.with_polymorphic:
+            if not spec:
+                spec = self.with_polymorphic[0]
+            if selectable is False:
+                selectable = self.with_polymorphic[1]
+        elif selectable is False:
+            selectable = None
+        mappers = self._mappers_from_spec(spec, selectable)
+        if selectable is not None:
+            return mappers, selectable
+        else:
+            return mappers, self._selectable_from_mappers(mappers, innerjoin)
+
+    @HasMemoized.memoized_attribute
+    def _polymorphic_properties(self):
+        return list(
+            self._iterate_polymorphic_properties(
+                self._with_polymorphic_mappers
+            )
+        )
+
+    @property
+    def _all_column_expressions(self):
+        poly_properties = self._polymorphic_properties
+        adapter = self._polymorphic_adapter
+
+        return [
+            adapter.columns[c] if adapter else c
+            for prop in poly_properties
+            if isinstance(prop, properties.ColumnProperty)
+            and prop._renders_in_subqueries
+            for c in prop.columns
+        ]
+
+    def _columns_plus_keys(self, polymorphic_mappers=()):
+        if polymorphic_mappers:
+            poly_properties = self._iterate_polymorphic_properties(
+                polymorphic_mappers
+            )
+        else:
+            poly_properties = self._polymorphic_properties
+
+        return [
+            (prop.key, prop.columns[0])
+            for prop in poly_properties
+            if isinstance(prop, properties.ColumnProperty)
+        ]
+
+    @HasMemoized.memoized_attribute
+    def _polymorphic_adapter(self) -> Optional[orm_util.ORMAdapter]:
+        if self._has_aliased_polymorphic_fromclause:
+            return orm_util.ORMAdapter(
+                orm_util._TraceAdaptRole.MAPPER_POLYMORPHIC_ADAPTER,
+                self,
+                selectable=self.selectable,
+                equivalents=self._equivalent_columns,
+                limit_on_entity=False,
+            )
+        else:
+            return None
+
+    def _iterate_polymorphic_properties(self, mappers=None):
+        """Return an iterator of MapperProperty objects which will render into
+        a SELECT."""
+        if mappers is None:
+            mappers = self._with_polymorphic_mappers
+
+        if not mappers:
+            for c in self.iterate_properties:
+                yield c
+        else:
+            # in the polymorphic case, filter out discriminator columns
+            # from other mappers, as these are sometimes dependent on that
+            # mapper's polymorphic selectable (which we don't want rendered)
+            for c in util.unique_list(
+                chain(
+                    *[
+                        list(mapper.iterate_properties)
+                        for mapper in [self] + mappers
+                    ]
+                )
+            ):
+                if getattr(c, "_is_polymorphic_discriminator", False) and (
+                    self.polymorphic_on is None
+                    or c.columns[0] is not self.polymorphic_on
+                ):
+                    continue
+                yield c
+
+    @HasMemoized.memoized_attribute
+    def attrs(self) -> util.ReadOnlyProperties[MapperProperty[Any]]:
+        """A namespace of all :class:`.MapperProperty` objects
+        associated this mapper.
+
+        This is an object that provides each property based on
+        its key name.  For instance, the mapper for a
+        ``User`` class which has ``User.name`` attribute would
+        provide ``mapper.attrs.name``, which would be the
+        :class:`.ColumnProperty` representing the ``name``
+        column.   The namespace object can also be iterated,
+        which would yield each :class:`.MapperProperty`.
+
+        :class:`_orm.Mapper` has several pre-filtered views
+        of this attribute which limit the types of properties
+        returned, including :attr:`.synonyms`, :attr:`.column_attrs`,
+        :attr:`.relationships`, and :attr:`.composites`.
+
+        .. warning::
+
+            The :attr:`_orm.Mapper.attrs` accessor namespace is an
+            instance of :class:`.OrderedProperties`.  This is
+            a dictionary-like object which includes a small number of
+            named methods such as :meth:`.OrderedProperties.items`
+            and :meth:`.OrderedProperties.values`.  When
+            accessing attributes dynamically, favor using the dict-access
+            scheme, e.g. ``mapper.attrs[somename]`` over
+            ``getattr(mapper.attrs, somename)`` to avoid name collisions.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.all_orm_descriptors`
+
+        """
+
+        self._check_configure()
+        return util.ReadOnlyProperties(self._props)
+
+    @HasMemoized.memoized_attribute
+    def all_orm_descriptors(self) -> util.ReadOnlyProperties[InspectionAttr]:
+        """A namespace of all :class:`.InspectionAttr` attributes associated
+        with the mapped class.
+
+        These attributes are in all cases Python :term:`descriptors`
+        associated with the mapped class or its superclasses.
+
+        This namespace includes attributes that are mapped to the class
+        as well as attributes declared by extension modules.
+        It includes any Python descriptor type that inherits from
+        :class:`.InspectionAttr`.  This includes
+        :class:`.QueryableAttribute`, as well as extension types such as
+        :class:`.hybrid_property`, :class:`.hybrid_method` and
+        :class:`.AssociationProxy`.
+
+        To distinguish between mapped attributes and extension attributes,
+        the attribute :attr:`.InspectionAttr.extension_type` will refer
+        to a constant that distinguishes between different extension types.
+
+        The sorting of the attributes is based on the following rules:
+
+        1. Iterate through the class and its superclasses in order from
+           subclass to superclass (i.e. iterate through ``cls.__mro__``)
+
+        2. For each class, yield the attributes in the order in which they
+           appear in ``__dict__``, with the exception of those in step
+           3 below.  In Python 3.6 and above this ordering will be the
+           same as that of the class' construction, with the exception
+           of attributes that were added after the fact by the application
+           or the mapper.
+
+        3. If a certain attribute key is also in the superclass ``__dict__``,
+           then it's included in the iteration for that class, and not the
+           class in which it first appeared.
+
+        The above process produces an ordering that is deterministic in terms
+        of the order in which attributes were assigned to the class.
+
+        .. versionchanged:: 1.3.19 ensured deterministic ordering for
+           :meth:`_orm.Mapper.all_orm_descriptors`.
+
+        When dealing with a :class:`.QueryableAttribute`, the
+        :attr:`.QueryableAttribute.property` attribute refers to the
+        :class:`.MapperProperty` property, which is what you get when
+        referring to the collection of mapped properties via
+        :attr:`_orm.Mapper.attrs`.
+
+        .. warning::
+
+            The :attr:`_orm.Mapper.all_orm_descriptors`
+            accessor namespace is an
+            instance of :class:`.OrderedProperties`.  This is
+            a dictionary-like object which includes a small number of
+            named methods such as :meth:`.OrderedProperties.items`
+            and :meth:`.OrderedProperties.values`.  When
+            accessing attributes dynamically, favor using the dict-access
+            scheme, e.g. ``mapper.all_orm_descriptors[somename]`` over
+            ``getattr(mapper.all_orm_descriptors, somename)`` to avoid name
+            collisions.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.attrs`
+
+        """
+        return util.ReadOnlyProperties(
+            dict(self.class_manager._all_sqla_attributes())
+        )
+
+    @HasMemoized.memoized_attribute
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def _pk_synonyms(self) -> Dict[str, str]:
+        """return a dictionary of {syn_attribute_name: pk_attr_name} for
+        all synonyms that refer to primary key columns
+
+        """
+        descriptor_props = util.preloaded.orm_descriptor_props
+
+        pk_keys = {prop.key for prop in self._identity_key_props}
+
+        return {
+            syn.key: syn.name
+            for k, syn in self._props.items()
+            if isinstance(syn, descriptor_props.SynonymProperty)
+            and syn.name in pk_keys
+        }
+
+    @HasMemoized.memoized_attribute
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def synonyms(self) -> util.ReadOnlyProperties[SynonymProperty[Any]]:
+        """Return a namespace of all :class:`.Synonym`
+        properties maintained by this :class:`_orm.Mapper`.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.attrs` - namespace of all
+            :class:`.MapperProperty`
+            objects.
+
+        """
+        descriptor_props = util.preloaded.orm_descriptor_props
+
+        return self._filter_properties(descriptor_props.SynonymProperty)
+
+    @property
+    def entity_namespace(self):
+        return self.class_
+
+    @HasMemoized.memoized_attribute
+    def column_attrs(self) -> util.ReadOnlyProperties[ColumnProperty[Any]]:
+        """Return a namespace of all :class:`.ColumnProperty`
+        properties maintained by this :class:`_orm.Mapper`.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.attrs` - namespace of all
+            :class:`.MapperProperty`
+            objects.
+
+        """
+        return self._filter_properties(properties.ColumnProperty)
+
+    @HasMemoized.memoized_attribute
+    @util.preload_module("sqlalchemy.orm.relationships")
+    def relationships(
+        self,
+    ) -> util.ReadOnlyProperties[RelationshipProperty[Any]]:
+        """A namespace of all :class:`.Relationship` properties
+        maintained by this :class:`_orm.Mapper`.
+
+        .. warning::
+
+            the :attr:`_orm.Mapper.relationships` accessor namespace is an
+            instance of :class:`.OrderedProperties`.  This is
+            a dictionary-like object which includes a small number of
+            named methods such as :meth:`.OrderedProperties.items`
+            and :meth:`.OrderedProperties.values`.  When
+            accessing attributes dynamically, favor using the dict-access
+            scheme, e.g. ``mapper.relationships[somename]`` over
+            ``getattr(mapper.relationships, somename)`` to avoid name
+            collisions.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.attrs` - namespace of all
+            :class:`.MapperProperty`
+            objects.
+
+        """
+        return self._filter_properties(
+            util.preloaded.orm_relationships.RelationshipProperty
+        )
+
+    @HasMemoized.memoized_attribute
+    @util.preload_module("sqlalchemy.orm.descriptor_props")
+    def composites(self) -> util.ReadOnlyProperties[CompositeProperty[Any]]:
+        """Return a namespace of all :class:`.Composite`
+        properties maintained by this :class:`_orm.Mapper`.
+
+        .. seealso::
+
+            :attr:`_orm.Mapper.attrs` - namespace of all
+            :class:`.MapperProperty`
+            objects.
+
+        """
+        return self._filter_properties(
+            util.preloaded.orm_descriptor_props.CompositeProperty
+        )
+
+    def _filter_properties(
+        self, type_: Type[_MP]
+    ) -> util.ReadOnlyProperties[_MP]:
+        self._check_configure()
+        return util.ReadOnlyProperties(
+            util.OrderedDict(
+                (k, v) for k, v in self._props.items() if isinstance(v, type_)
+            )
+        )
+
+    @HasMemoized.memoized_attribute
+    def _get_clause(self):
+        """create a "get clause" based on the primary key.  this is used
+        by query.get() and many-to-one lazyloads to load this item
+        by primary key.
+
+        """
+        params = [
+            (
+                primary_key,
+                sql.bindparam("pk_%d" % idx, type_=primary_key.type),
+            )
+            for idx, primary_key in enumerate(self.primary_key, 1)
+        ]
+        return (
+            sql.and_(*[k == v for (k, v) in params]),
+            util.column_dict(params),
+        )
+
+    @HasMemoized.memoized_attribute
+    def _equivalent_columns(self) -> _EquivalentColumnMap:
+        """Create a map of all equivalent columns, based on
+        the determination of column pairs that are equated to
+        one another based on inherit condition.  This is designed
+        to work with the queries that util.polymorphic_union
+        comes up with, which often don't include the columns from
+        the base table directly (including the subclass table columns
+        only).
+
+        The resulting structure is a dictionary of columns mapped
+        to lists of equivalent columns, e.g.::
+
+            {tablea.col1: {tableb.col1, tablec.col1}, tablea.col2: {tabled.col2}}
+
+        """  # noqa: E501
+        result: _EquivalentColumnMap = {}
+
+        def visit_binary(binary):
+            if binary.operator == operators.eq:
+                if binary.left in result:
+                    result[binary.left].add(binary.right)
+                else:
+                    result[binary.left] = {binary.right}
+                if binary.right in result:
+                    result[binary.right].add(binary.left)
+                else:
+                    result[binary.right] = {binary.left}
+
+        for mapper in self.base_mapper.self_and_descendants:
+            if mapper.inherit_condition is not None:
+                visitors.traverse(
+                    mapper.inherit_condition, {}, {"binary": visit_binary}
+                )
+
+        return result
+
+    def _is_userland_descriptor(self, assigned_name: str, obj: Any) -> bool:
+        if isinstance(
+            obj,
+            (
+                _MappedAttribute,
+                instrumentation.ClassManager,
+                expression.ColumnElement,
+            ),
+        ):
+            return False
+        else:
+            return assigned_name not in self._dataclass_fields
+
+    @HasMemoized.memoized_attribute
+    def _dataclass_fields(self):
+        return [f.name for f in util.dataclass_fields(self.class_)]
+
+    def _should_exclude(self, name, assigned_name, local, column):
+        """determine whether a particular property should be implicitly
+        present on the class.
+
+        This occurs when properties are propagated from an inherited class, or
+        are applied from the columns present in the mapped table.
+
+        """
+
+        if column is not None and sql_base._never_select_column(column):
+            return True
+
+        # check for class-bound attributes and/or descriptors,
+        # either local or from an inherited class
+        # ignore dataclass field default values
+        if local:
+            if self.class_.__dict__.get(
+                assigned_name, None
+            ) is not None and self._is_userland_descriptor(
+                assigned_name, self.class_.__dict__[assigned_name]
+            ):
+                return True
+        else:
+            attr = self.class_manager._get_class_attr_mro(assigned_name, None)
+            if attr is not None and self._is_userland_descriptor(
+                assigned_name, attr
+            ):
+                return True
+
+        if (
+            self.include_properties is not None
+            and name not in self.include_properties
+            and (column is None or column not in self.include_properties)
+        ):
+            self._log("not including property %s" % (name))
+            return True
+
+        if self.exclude_properties is not None and (
+            name in self.exclude_properties
+            or (column is not None and column in self.exclude_properties)
+        ):
+            self._log("excluding property %s" % (name))
+            return True
+
+        return False
+
+    def common_parent(self, other: Mapper[Any]) -> bool:
+        """Return true if the given mapper shares a
+        common inherited parent as this mapper."""
+
+        return self.base_mapper is other.base_mapper
+
+    def is_sibling(self, other: Mapper[Any]) -> bool:
+        """return true if the other mapper is an inheriting sibling to this
+        one.  common parent but different branch
+
+        """
+        return (
+            self.base_mapper is other.base_mapper
+            and not self.isa(other)
+            and not other.isa(self)
+        )
+
+    def _canload(
+        self, state: InstanceState[Any], allow_subtypes: bool
+    ) -> bool:
+        s = self.primary_mapper()
+        if self.polymorphic_on is not None or allow_subtypes:
+            return _state_mapper(state).isa(s)
+        else:
+            return _state_mapper(state) is s
+
+    def isa(self, other: Mapper[Any]) -> bool:
+        """Return True if the this mapper inherits from the given mapper."""
+
+        m: Optional[Mapper[Any]] = self
+        while m and m is not other:
+            m = m.inherits
+        return bool(m)
+
+    def iterate_to_root(self) -> Iterator[Mapper[Any]]:
+        m: Optional[Mapper[Any]] = self
+        while m:
+            yield m
+            m = m.inherits
+
+    @HasMemoized.memoized_attribute
+    def self_and_descendants(self) -> Sequence[Mapper[Any]]:
+        """The collection including this mapper and all descendant mappers.
+
+        This includes not just the immediately inheriting mappers but
+        all their inheriting mappers as well.
+
+        """
+        descendants = []
+        stack = deque([self])
+        while stack:
+            item = stack.popleft()
+            descendants.append(item)
+            stack.extend(item._inheriting_mappers)
+        return util.WeakSequence(descendants)
+
+    def polymorphic_iterator(self) -> Iterator[Mapper[Any]]:
+        """Iterate through the collection including this mapper and
+        all descendant mappers.
+
+        This includes not just the immediately inheriting mappers but
+        all their inheriting mappers as well.
+
+        To iterate through an entire hierarchy, use
+        ``mapper.base_mapper.polymorphic_iterator()``.
+
+        """
+        return iter(self.self_and_descendants)
+
+    def primary_mapper(self) -> Mapper[Any]:
+        """Return the primary mapper corresponding to this mapper's class key
+        (class)."""
+
+        return self.class_manager.mapper
+
+    @property
+    def primary_base_mapper(self) -> Mapper[Any]:
+        return self.class_manager.mapper.base_mapper
+
+    def _result_has_identity_key(self, result, adapter=None):
+        pk_cols: Sequence[ColumnClause[Any]] = self.primary_key
+        if adapter:
+            pk_cols = [adapter.columns[c] for c in pk_cols]
+        rk = result.keys()
+        for col in pk_cols:
+            if col not in rk:
+                return False
+        else:
+            return True
+
+    def identity_key_from_row(
+        self,
+        row: Optional[Union[Row[Any], RowMapping]],
+        identity_token: Optional[Any] = None,
+        adapter: Optional[ORMAdapter] = None,
+    ) -> _IdentityKeyType[_O]:
+        """Return an identity-map key for use in storing/retrieving an
+        item from the identity map.
+
+        :param row: A :class:`.Row` or :class:`.RowMapping` produced from a
+         result set that selected from the ORM mapped primary key columns.
+
+         .. versionchanged:: 2.0
+            :class:`.Row` or :class:`.RowMapping` are accepted
+            for the "row" argument
+
+        """
+        pk_cols: Sequence[ColumnClause[Any]] = self.primary_key
+        if adapter:
+            pk_cols = [adapter.columns[c] for c in pk_cols]
+
+        if hasattr(row, "_mapping"):
+            mapping = row._mapping  # type: ignore
+        else:
+            mapping = cast("Mapping[Any, Any]", row)
+
+        return (
+            self._identity_class,
+            tuple(mapping[column] for column in pk_cols),  # type: ignore
+            identity_token,
+        )
+
+    def identity_key_from_primary_key(
+        self,
+        primary_key: Tuple[Any, ...],
+        identity_token: Optional[Any] = None,
+    ) -> _IdentityKeyType[_O]:
+        """Return an identity-map key for use in storing/retrieving an
+        item from an identity map.
+
+        :param primary_key: A list of values indicating the identifier.
+
+        """
+        return (
+            self._identity_class,
+            tuple(primary_key),
+            identity_token,
+        )
+
+    def identity_key_from_instance(self, instance: _O) -> _IdentityKeyType[_O]:
+        """Return the identity key for the given instance, based on
+        its primary key attributes.
+
+        If the instance's state is expired, calling this method
+        will result in a database check to see if the object has been deleted.
+        If the row no longer exists,
+        :class:`~sqlalchemy.orm.exc.ObjectDeletedError` is raised.
+
+        This value is typically also found on the instance state under the
+        attribute name `key`.
+
+        """
+        state = attributes.instance_state(instance)
+        return self._identity_key_from_state(state, PassiveFlag.PASSIVE_OFF)
+
+    def _identity_key_from_state(
+        self,
+        state: InstanceState[_O],
+        passive: PassiveFlag = PassiveFlag.PASSIVE_RETURN_NO_VALUE,
+    ) -> _IdentityKeyType[_O]:
+        dict_ = state.dict
+        manager = state.manager
+        return (
+            self._identity_class,
+            tuple(
+                [
+                    manager[prop.key].impl.get(state, dict_, passive)
+                    for prop in self._identity_key_props
+                ]
+            ),
+            state.identity_token,
+        )
+
+    def primary_key_from_instance(self, instance: _O) -> Tuple[Any, ...]:
+        """Return the list of primary key values for the given
+        instance.
+
+        If the instance's state is expired, calling this method
+        will result in a database check to see if the object has been deleted.
+        If the row no longer exists,
+        :class:`~sqlalchemy.orm.exc.ObjectDeletedError` is raised.
+
+        """
+        state = attributes.instance_state(instance)
+        identity_key = self._identity_key_from_state(
+            state, PassiveFlag.PASSIVE_OFF
+        )
+        return identity_key[1]
+
+    @HasMemoized.memoized_attribute
+    def _persistent_sortkey_fn(self):
+        key_fns = [col.type.sort_key_function for col in self.primary_key]
+
+        if set(key_fns).difference([None]):
+
+            def key(state):
+                return tuple(
+                    key_fn(val) if key_fn is not None else val
+                    for key_fn, val in zip(key_fns, state.key[1])
+                )
+
+        else:
+
+            def key(state):
+                return state.key[1]
+
+        return key
+
+    @HasMemoized.memoized_attribute
+    def _identity_key_props(self):
+        return [self._columntoproperty[col] for col in self.primary_key]
+
+    @HasMemoized.memoized_attribute
+    def _all_pk_cols(self):
+        collection: Set[ColumnClause[Any]] = set()
+        for table in self.tables:
+            collection.update(self._pks_by_table[table])
+        return collection
+
+    @HasMemoized.memoized_attribute
+    def _should_undefer_in_wildcard(self):
+        cols: Set[ColumnElement[Any]] = set(self.primary_key)
+        if self.polymorphic_on is not None:
+            cols.add(self.polymorphic_on)
+        return cols
+
+    @HasMemoized.memoized_attribute
+    def _primary_key_propkeys(self):
+        return {self._columntoproperty[col].key for col in self._all_pk_cols}
+
+    def _get_state_attr_by_column(
+        self,
+        state: InstanceState[_O],
+        dict_: _InstanceDict,
+        column: ColumnElement[Any],
+        passive: PassiveFlag = PassiveFlag.PASSIVE_RETURN_NO_VALUE,
+    ) -> Any:
+        prop = self._columntoproperty[column]
+        return state.manager[prop.key].impl.get(state, dict_, passive=passive)
+
+    def _set_committed_state_attr_by_column(self, state, dict_, column, value):
+        prop = self._columntoproperty[column]
+        state.manager[prop.key].impl.set_committed_value(state, dict_, value)
+
+    def _set_state_attr_by_column(self, state, dict_, column, value):
+        prop = self._columntoproperty[column]
+        state.manager[prop.key].impl.set(state, dict_, value, None)
+
+    def _get_committed_attr_by_column(self, obj, column):
+        state = attributes.instance_state(obj)
+        dict_ = attributes.instance_dict(obj)
+        return self._get_committed_state_attr_by_column(
+            state, dict_, column, passive=PassiveFlag.PASSIVE_OFF
+        )
+
+    def _get_committed_state_attr_by_column(
+        self, state, dict_, column, passive=PassiveFlag.PASSIVE_RETURN_NO_VALUE
+    ):
+        prop = self._columntoproperty[column]
+        return state.manager[prop.key].impl.get_committed_value(
+            state, dict_, passive=passive
+        )
+
+    def _optimized_get_statement(self, state, attribute_names):
+        """assemble a WHERE clause which retrieves a given state by primary
+        key, using a minimized set of tables.
+
+        Applies to a joined-table inheritance mapper where the
+        requested attribute names are only present on joined tables,
+        not the base table.  The WHERE clause attempts to include
+        only those tables to minimize joins.
+
+        """
+        props = self._props
+
+        col_attribute_names = set(attribute_names).intersection(
+            state.mapper.column_attrs.keys()
+        )
+        tables: Set[FromClause] = set(
+            chain(
+                *[
+                    sql_util.find_tables(c, check_columns=True)
+                    for key in col_attribute_names
+                    for c in props[key].columns
+                ]
+            )
+        )
+
+        if self.base_mapper.local_table in tables:
+            return None
+
+        def visit_binary(binary):
+            leftcol = binary.left
+            rightcol = binary.right
+            if leftcol is None or rightcol is None:
+                return
+
+            if leftcol.table not in tables:
+                leftval = self._get_committed_state_attr_by_column(
+                    state,
+                    state.dict,
+                    leftcol,
+                    passive=PassiveFlag.PASSIVE_NO_INITIALIZE,
+                )
+                if leftval in orm_util._none_set:
+                    raise _OptGetColumnsNotAvailable()
+                binary.left = sql.bindparam(
+                    None, leftval, type_=binary.right.type
+                )
+            elif rightcol.table not in tables:
+                rightval = self._get_committed_state_attr_by_column(
+                    state,
+                    state.dict,
+                    rightcol,
+                    passive=PassiveFlag.PASSIVE_NO_INITIALIZE,
+                )
+                if rightval in orm_util._none_set:
+                    raise _OptGetColumnsNotAvailable()
+                binary.right = sql.bindparam(
+                    None, rightval, type_=binary.right.type
+                )
+
+        allconds: List[ColumnElement[bool]] = []
+
+        start = False
+
+        # as of #7507, from the lowest base table on upwards,
+        # we include all intermediary tables.
+
+        for mapper in reversed(list(self.iterate_to_root())):
+            if mapper.local_table in tables:
+                start = True
+            elif not isinstance(mapper.local_table, expression.TableClause):
+                return None
+            if start and not mapper.single:
+                assert mapper.inherits
+                assert not mapper.concrete
+                assert mapper.inherit_condition is not None
+                allconds.append(mapper.inherit_condition)
+                tables.add(mapper.local_table)
+
+        # only the bottom table needs its criteria to be altered to fit
+        # the primary key ident - the rest of the tables upwards to the
+        # descendant-most class should all be present and joined to each
+        # other.
+        try:
+            _traversed = visitors.cloned_traverse(
+                allconds[0], {}, {"binary": visit_binary}
+            )
+        except _OptGetColumnsNotAvailable:
+            return None
+        else:
+            allconds[0] = _traversed
+
+        cond = sql.and_(*allconds)
+
+        cols = []
+        for key in col_attribute_names:
+            cols.extend(props[key].columns)
+        return (
+            sql.select(*cols)
+            .where(cond)
+            .set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+        )
+
+    def _iterate_to_target_viawpoly(self, mapper):
+        if self.isa(mapper):
+            prev = self
+            for m in self.iterate_to_root():
+                yield m
+
+                if m is not prev and prev not in m._with_polymorphic_mappers:
+                    break
+
+                prev = m
+                if m is mapper:
+                    break
+
+    @HasMemoized.memoized_attribute
+    def _would_selectinload_combinations_cache(self):
+        return {}
+
+    def _would_selectin_load_only_from_given_mapper(self, super_mapper):
+        """return True if this mapper would "selectin" polymorphic load based
+        on the given super mapper, and not from a setting from a subclass.
+
+        given::
+
+            class A: ...
+
+
+            class B(A):
+                __mapper_args__ = {"polymorphic_load": "selectin"}
+
+
+            class C(B): ...
+
+
+            class D(B):
+                __mapper_args__ = {"polymorphic_load": "selectin"}
+
+        ``inspect(C)._would_selectin_load_only_from_given_mapper(inspect(B))``
+        returns True, because C does selectin loading because of B's setting.
+
+        OTOH, ``inspect(D)
+        ._would_selectin_load_only_from_given_mapper(inspect(B))``
+        returns False, because D does selectin loading because of its own
+        setting; when we are doing a selectin poly load from B, we want to
+        filter out D because it would already have its own selectin poly load
+        set up separately.
+
+        Added as part of #9373.
+
+        """
+        cache = self._would_selectinload_combinations_cache
+
+        try:
+            return cache[super_mapper]
+        except KeyError:
+            pass
+
+        # assert that given object is a supermapper, meaning we already
+        # strong reference it directly or indirectly.  this allows us
+        # to not worry that we are creating new strongrefs to unrelated
+        # mappers or other objects.
+        assert self.isa(super_mapper)
+
+        mapper = super_mapper
+        for m in self._iterate_to_target_viawpoly(mapper):
+            if m.polymorphic_load == "selectin":
+                retval = m is super_mapper
+                break
+        else:
+            retval = False
+
+        cache[super_mapper] = retval
+        return retval
+
+    def _should_selectin_load(self, enabled_via_opt, polymorphic_from):
+        if not enabled_via_opt:
+            # common case, takes place for all polymorphic loads
+            mapper = polymorphic_from
+            for m in self._iterate_to_target_viawpoly(mapper):
+                if m.polymorphic_load == "selectin":
+                    return m
+        else:
+            # uncommon case, selectin load options were used
+            enabled_via_opt = set(enabled_via_opt)
+            enabled_via_opt_mappers = {e.mapper: e for e in enabled_via_opt}
+            for entity in enabled_via_opt.union([polymorphic_from]):
+                mapper = entity.mapper
+                for m in self._iterate_to_target_viawpoly(mapper):
+                    if (
+                        m.polymorphic_load == "selectin"
+                        or m in enabled_via_opt_mappers
+                    ):
+                        return enabled_via_opt_mappers.get(m, m)
+
+        return None
+
+    @util.preload_module("sqlalchemy.orm.strategy_options")
+    def _subclass_load_via_in(self, entity, polymorphic_from):
+        """Assemble a that can load the columns local to
+        this subclass as a SELECT with IN.
+
+        """
+
+        strategy_options = util.preloaded.orm_strategy_options
+
+        assert self.inherits
+
+        if self.polymorphic_on is not None:
+            polymorphic_prop = self._columntoproperty[self.polymorphic_on]
+            keep_props = set([polymorphic_prop] + self._identity_key_props)
+        else:
+            keep_props = set(self._identity_key_props)
+
+        disable_opt = strategy_options.Load(entity)
+        enable_opt = strategy_options.Load(entity)
+
+        classes_to_include = {self}
+        m: Optional[Mapper[Any]] = self.inherits
+        while (
+            m is not None
+            and m is not polymorphic_from
+            and m.polymorphic_load == "selectin"
+        ):
+            classes_to_include.add(m)
+            m = m.inherits
+
+        for prop in self.column_attrs + self.relationships:
+            # skip prop keys that are not instrumented on the mapped class.
+            # this is primarily the "_sa_polymorphic_on" property that gets
+            # created for an ad-hoc polymorphic_on SQL expression, issue #8704
+            if prop.key not in self.class_manager:
+                continue
+
+            if prop.parent in classes_to_include or prop in keep_props:
+                # "enable" options, to turn on the properties that we want to
+                # load by default (subject to options from the query)
+                if not isinstance(prop, StrategizedProperty):
+                    continue
+
+                enable_opt = enable_opt._set_generic_strategy(
+                    # convert string name to an attribute before passing
+                    # to loader strategy.   note this must be in terms
+                    # of given entity, such as AliasedClass, etc.
+                    (getattr(entity.entity_namespace, prop.key),),
+                    dict(prop.strategy_key),
+                    _reconcile_to_other=True,
+                )
+            else:
+                # "disable" options, to turn off the properties from the
+                # superclass that we *don't* want to load, applied after
+                # the options from the query to override them
+                disable_opt = disable_opt._set_generic_strategy(
+                    # convert string name to an attribute before passing
+                    # to loader strategy.   note this must be in terms
+                    # of given entity, such as AliasedClass, etc.
+                    (getattr(entity.entity_namespace, prop.key),),
+                    {"do_nothing": True},
+                    _reconcile_to_other=False,
+                )
+
+        primary_key = [
+            sql_util._deep_annotate(pk, {"_orm_adapt": True})
+            for pk in self.primary_key
+        ]
+
+        in_expr: ColumnElement[Any]
+
+        if len(primary_key) > 1:
+            in_expr = sql.tuple_(*primary_key)
+        else:
+            in_expr = primary_key[0]
+
+        if entity.is_aliased_class:
+            assert entity.mapper is self
+
+            q = sql.select(entity).set_label_style(
+                LABEL_STYLE_TABLENAME_PLUS_COL
+            )
+
+            in_expr = entity._adapter.traverse(in_expr)
+            primary_key = [entity._adapter.traverse(k) for k in primary_key]
+            q = q.where(
+                in_expr.in_(sql.bindparam("primary_keys", expanding=True))
+            ).order_by(*primary_key)
+        else:
+            q = sql.select(self).set_label_style(
+                LABEL_STYLE_TABLENAME_PLUS_COL
+            )
+            q = q.where(
+                in_expr.in_(sql.bindparam("primary_keys", expanding=True))
+            ).order_by(*primary_key)
+
+        return q, enable_opt, disable_opt
+
+    @HasMemoized.memoized_attribute
+    def _subclass_load_via_in_mapper(self):
+        # the default is loading this mapper against the basemost mapper
+        return self._subclass_load_via_in(self, self.base_mapper)
+
+    def cascade_iterator(
+        self,
+        type_: str,
+        state: InstanceState[_O],
+        halt_on: Optional[Callable[[InstanceState[Any]], bool]] = None,
+    ) -> Iterator[
+        Tuple[object, Mapper[Any], InstanceState[Any], _InstanceDict]
+    ]:
+        r"""Iterate each element and its mapper in an object graph,
+        for all relationships that meet the given cascade rule.
+
+        :param type\_:
+          The name of the cascade rule (i.e. ``"save-update"``, ``"delete"``,
+          etc.).
+
+          .. note::  the ``"all"`` cascade is not accepted here.  For a generic
+             object traversal function, see :ref:`faq_walk_objects`.
+
+        :param state:
+          The lead InstanceState.  child items will be processed per
+          the relationships defined for this object's mapper.
+
+        :return: the method yields individual object instances.
+
+        .. seealso::
+
+            :ref:`unitofwork_cascades`
+
+            :ref:`faq_walk_objects` - illustrates a generic function to
+            traverse all objects without relying on cascades.
+
+        """
+        visited_states: Set[InstanceState[Any]] = set()
+        prp, mpp = object(), object()
+
+        assert state.mapper.isa(self)
+
+        # this is actually a recursive structure, fully typing it seems
+        # a little too difficult for what it's worth here
+        visitables: Deque[
+            Tuple[
+                Deque[Any],
+                object,
+                Optional[InstanceState[Any]],
+                Optional[_InstanceDict],
+            ]
+        ]
+
+        visitables = deque(
+            [(deque(state.mapper._props.values()), prp, state, state.dict)]
+        )
+
+        while visitables:
+            iterator, item_type, parent_state, parent_dict = visitables[-1]
+            if not iterator:
+                visitables.pop()
+                continue
+
+            if item_type is prp:
+                prop = iterator.popleft()
+                if not prop.cascade or type_ not in prop.cascade:
+                    continue
+                assert parent_state is not None
+                assert parent_dict is not None
+                queue = deque(
+                    prop.cascade_iterator(
+                        type_,
+                        parent_state,
+                        parent_dict,
+                        visited_states,
+                        halt_on,
+                    )
+                )
+                if queue:
+                    visitables.append((queue, mpp, None, None))
+            elif item_type is mpp:
+                (
+                    instance,
+                    instance_mapper,
+                    corresponding_state,
+                    corresponding_dict,
+                ) = iterator.popleft()
+                yield (
+                    instance,
+                    instance_mapper,
+                    corresponding_state,
+                    corresponding_dict,
+                )
+                visitables.append(
+                    (
+                        deque(instance_mapper._props.values()),
+                        prp,
+                        corresponding_state,
+                        corresponding_dict,
+                    )
+                )
+
+    @HasMemoized.memoized_attribute
+    def _compiled_cache(self):
+        return util.LRUCache(self._compiled_cache_size)
+
+    @HasMemoized.memoized_attribute
+    def _multiple_persistence_tables(self):
+        return len(self.tables) > 1
+
+    @HasMemoized.memoized_attribute
+    def _sorted_tables(self):
+        table_to_mapper: Dict[TableClause, Mapper[Any]] = {}
+
+        for mapper in self.base_mapper.self_and_descendants:
+            for t in mapper.tables:
+                table_to_mapper.setdefault(t, mapper)
+
+        extra_dependencies = []
+        for table, mapper in table_to_mapper.items():
+            super_ = mapper.inherits
+            if super_:
+                extra_dependencies.extend(
+                    [(super_table, table) for super_table in super_.tables]
+                )
+
+        def skip(fk):
+            # attempt to skip dependencies that are not
+            # significant to the inheritance chain
+            # for two tables that are related by inheritance.
+            # while that dependency may be important, it's technically
+            # not what we mean to sort on here.
+            parent = table_to_mapper.get(fk.parent.table)
+            dep = table_to_mapper.get(fk.column.table)
+            if (
+                parent is not None
+                and dep is not None
+                and dep is not parent
+                and dep.inherit_condition is not None
+            ):
+                cols = set(sql_util._find_columns(dep.inherit_condition))
+                if parent.inherit_condition is not None:
+                    cols = cols.union(
+                        sql_util._find_columns(parent.inherit_condition)
+                    )
+                    return fk.parent not in cols and fk.column not in cols
+                else:
+                    return fk.parent not in cols
+            return False
+
+        sorted_ = sql_util.sort_tables(
+            table_to_mapper,
+            skip_fn=skip,
+            extra_dependencies=extra_dependencies,
+        )
+
+        ret = util.OrderedDict()
+        for t in sorted_:
+            ret[t] = table_to_mapper[t]
+        return ret
+
+    def _memo(self, key: Any, callable_: Callable[[], _T]) -> _T:
+        if key in self._memoized_values:
+            return cast(_T, self._memoized_values[key])
+        else:
+            self._memoized_values[key] = value = callable_()
+            return value
+
+    @util.memoized_property
+    def _table_to_equated(self):
+        """memoized map of tables to collections of columns to be
+        synchronized upwards to the base mapper."""
+
+        result: util.defaultdict[
+            Table,
+            List[
+                Tuple[
+                    Mapper[Any],
+                    List[Tuple[ColumnElement[Any], ColumnElement[Any]]],
+                ]
+            ],
+        ] = util.defaultdict(list)
+
+        def set_union(x, y):
+            return x.union(y)
+
+        for table in self._sorted_tables:
+            cols = set(table.c)
+
+            for m in self.iterate_to_root():
+                if m._inherits_equated_pairs and cols.intersection(
+                    reduce(
+                        set_union,
+                        [l.proxy_set for l, r in m._inherits_equated_pairs],
+                    )
+                ):
+                    result[table].append((m, m._inherits_equated_pairs))
+
+        return result
+
+
+class _OptGetColumnsNotAvailable(Exception):
+    pass
+
+
+def configure_mappers() -> None:
+    """Initialize the inter-mapper relationships of all mappers that
+    have been constructed thus far across all :class:`_orm.registry`
+    collections.
+
+    The configure step is used to reconcile and initialize the
+    :func:`_orm.relationship` linkages between mapped classes, as well as to
+    invoke configuration events such as the
+    :meth:`_orm.MapperEvents.before_configured` and
+    :meth:`_orm.MapperEvents.after_configured`, which may be used by ORM
+    extensions or user-defined extension hooks.
+
+    Mapper configuration is normally invoked automatically, the first time
+    mappings from a particular :class:`_orm.registry` are used, as well as
+    whenever mappings are used and additional not-yet-configured mappers have
+    been constructed. The automatic configuration process however is local only
+    to the :class:`_orm.registry` involving the target mapper and any related
+    :class:`_orm.registry` objects which it may depend on; this is
+    equivalent to invoking the :meth:`_orm.registry.configure` method
+    on a particular :class:`_orm.registry`.
+
+    By contrast, the :func:`_orm.configure_mappers` function will invoke the
+    configuration process on all :class:`_orm.registry` objects that
+    exist in memory, and may be useful for scenarios where many individual
+    :class:`_orm.registry` objects that are nonetheless interrelated are
+    in use.
+
+    .. versionchanged:: 1.4
+
+        As of SQLAlchemy 1.4.0b2, this function works on a
+        per-:class:`_orm.registry` basis, locating all :class:`_orm.registry`
+        objects present and invoking the :meth:`_orm.registry.configure` method
+        on each. The :meth:`_orm.registry.configure` method may be preferred to
+        limit the configuration of mappers to those local to a particular
+        :class:`_orm.registry` and/or declarative base class.
+
+    Points at which automatic configuration is invoked include when a mapped
+    class is instantiated into an instance, as well as when ORM queries
+    are emitted using :meth:`.Session.query` or :meth:`_orm.Session.execute`
+    with an ORM-enabled statement.
+
+    The mapper configure process, whether invoked by
+    :func:`_orm.configure_mappers` or from :meth:`_orm.registry.configure`,
+    provides several event hooks that can be used to augment the mapper
+    configuration step. These hooks include:
+
+    * :meth:`.MapperEvents.before_configured` - called once before
+      :func:`.configure_mappers` or :meth:`_orm.registry.configure` does any
+      work; this can be used to establish additional options, properties, or
+      related mappings before the operation proceeds.
+
+    * :meth:`.MapperEvents.mapper_configured` - called as each individual
+      :class:`_orm.Mapper` is configured within the process; will include all
+      mapper state except for backrefs set up by other mappers that are still
+      to be configured.
+
+    * :meth:`.MapperEvents.after_configured` - called once after
+      :func:`.configure_mappers` or :meth:`_orm.registry.configure` is
+      complete; at this stage, all :class:`_orm.Mapper` objects that fall
+      within the scope of the configuration operation will be fully configured.
+      Note that the calling application may still have other mappings that
+      haven't been produced yet, such as if they are in modules as yet
+      unimported, and may also have mappings that are still to be configured,
+      if they are in other :class:`_orm.registry` collections not part of the
+      current scope of configuration.
+
+    """
+
+    _configure_registries(_all_registries(), cascade=True)
+
+
+def _configure_registries(
+    registries: Set[_RegistryType], cascade: bool
+) -> None:
+    for reg in registries:
+        if reg._new_mappers:
+            break
+    else:
+        return
+
+    with _CONFIGURE_MUTEX:
+        global _already_compiling
+        if _already_compiling:
+            return
+        _already_compiling = True
+        try:
+            # double-check inside mutex
+            for reg in registries:
+                if reg._new_mappers:
+                    break
+            else:
+                return
+
+            Mapper.dispatch._for_class(Mapper).before_configured()  # type: ignore # noqa: E501
+            # initialize properties on all mappers
+            # note that _mapper_registry is unordered, which
+            # may randomly conceal/reveal issues related to
+            # the order of mapper compilation
+
+            _do_configure_registries(registries, cascade)
+        finally:
+            _already_compiling = False
+    Mapper.dispatch._for_class(Mapper).after_configured()  # type: ignore
+
+
+@util.preload_module("sqlalchemy.orm.decl_api")
+def _do_configure_registries(
+    registries: Set[_RegistryType], cascade: bool
+) -> None:
+    registry = util.preloaded.orm_decl_api.registry
+
+    orig = set(registries)
+
+    for reg in registry._recurse_with_dependencies(registries):
+        has_skip = False
+
+        for mapper in reg._mappers_to_configure():
+            run_configure = None
+
+            for fn in mapper.dispatch.before_mapper_configured:
+                run_configure = fn(mapper, mapper.class_)
+                if run_configure is EXT_SKIP:
+                    has_skip = True
+                    break
+            if run_configure is EXT_SKIP:
+                continue
+
+            if getattr(mapper, "_configure_failed", False):
+                e = sa_exc.InvalidRequestError(
+                    "One or more mappers failed to initialize - "
+                    "can't proceed with initialization of other "
+                    "mappers. Triggering mapper: '%s'. "
+                    "Original exception was: %s"
+                    % (mapper, mapper._configure_failed)
+                )
+                e._configure_failed = mapper._configure_failed  # type: ignore
+                raise e
+
+            if not mapper.configured:
+                try:
+                    mapper._post_configure_properties()
+                    mapper._expire_memoizations()
+                    mapper.dispatch.mapper_configured(mapper, mapper.class_)
+                except Exception:
+                    exc = sys.exc_info()[1]
+                    if not hasattr(exc, "_configure_failed"):
+                        mapper._configure_failed = exc
+                    raise
+        if not has_skip:
+            reg._new_mappers = False
+
+        if not cascade and reg._dependencies.difference(orig):
+            raise sa_exc.InvalidRequestError(
+                "configure was called with cascade=False but "
+                "additional registries remain"
+            )
+
+
+@util.preload_module("sqlalchemy.orm.decl_api")
+def _dispose_registries(registries: Set[_RegistryType], cascade: bool) -> None:
+    registry = util.preloaded.orm_decl_api.registry
+
+    orig = set(registries)
+
+    for reg in registry._recurse_with_dependents(registries):
+        if not cascade and reg._dependents.difference(orig):
+            raise sa_exc.InvalidRequestError(
+                "Registry has dependent registries that are not disposed; "
+                "pass cascade=True to clear these also"
+            )
+
+        while reg._managers:
+            try:
+                manager, _ = reg._managers.popitem()
+            except KeyError:
+                # guard against race between while and popitem
+                pass
+            else:
+                reg._dispose_manager_and_mapper(manager)
+
+        reg._non_primary_mappers.clear()
+        reg._dependents.clear()
+        for dep in reg._dependencies:
+            dep._dependents.discard(reg)
+        reg._dependencies.clear()
+        # this wasn't done in the 1.3 clear_mappers() and in fact it
+        # was a bug, as it could cause configure_mappers() to invoke
+        # the "before_configured" event even though mappers had all been
+        # disposed.
+        reg._new_mappers = False
+
+
+def reconstructor(fn):
+    """Decorate a method as the 'reconstructor' hook.
+
+    Designates a single method as the "reconstructor", an ``__init__``-like
+    method that will be called by the ORM after the instance has been
+    loaded from the database or otherwise reconstituted.
+
+    .. tip::
+
+        The :func:`_orm.reconstructor` decorator makes use of the
+        :meth:`_orm.InstanceEvents.load` event hook, which can be
+        used directly.
+
+    The reconstructor will be invoked with no arguments.  Scalar
+    (non-collection) database-mapped attributes of the instance will
+    be available for use within the function.  Eagerly-loaded
+    collections are generally not yet available and will usually only
+    contain the first element.  ORM state changes made to objects at
+    this stage will not be recorded for the next flush() operation, so
+    the activity within a reconstructor should be conservative.
+
+    .. seealso::
+
+        :meth:`.InstanceEvents.load`
+
+    """
+    fn.__sa_reconstructor__ = True
+    return fn
+
+
+def validates(
+    *names: str, include_removes: bool = False, include_backrefs: bool = True
+) -> Callable[[_Fn], _Fn]:
+    r"""Decorate a method as a 'validator' for one or more named properties.
+
+    Designates a method as a validator, a method which receives the
+    name of the attribute as well as a value to be assigned, or in the
+    case of a collection, the value to be added to the collection.
+    The function can then raise validation exceptions to halt the
+    process from continuing (where Python's built-in ``ValueError``
+    and ``AssertionError`` exceptions are reasonable choices), or can
+    modify or replace the value before proceeding. The function should
+    otherwise return the given value.
+
+    Note that a validator for a collection **cannot** issue a load of that
+    collection within the validation routine - this usage raises
+    an assertion to avoid recursion overflows.  This is a reentrant
+    condition which is not supported.
+
+    :param \*names: list of attribute names to be validated.
+    :param include_removes: if True, "remove" events will be
+     sent as well - the validation function must accept an additional
+     argument "is_remove" which will be a boolean.
+
+    :param include_backrefs: defaults to ``True``; if ``False``, the
+     validation function will not emit if the originator is an attribute
+     event related via a backref.  This can be used for bi-directional
+     :func:`.validates` usage where only one validator should emit per
+     attribute operation.
+
+     .. versionchanged:: 2.0.16 This paramter inadvertently defaulted to
+        ``False`` for releases 2.0.0 through 2.0.15.  Its correct default
+        of ``True`` is restored in 2.0.16.
+
+    .. seealso::
+
+      :ref:`simple_validators` - usage examples for :func:`.validates`
+
+    """
+
+    def wrap(fn: _Fn) -> _Fn:
+        fn.__sa_validators__ = names  # type: ignore[attr-defined]
+        fn.__sa_validation_opts__ = {  # type: ignore[attr-defined]
+            "include_removes": include_removes,
+            "include_backrefs": include_backrefs,
+        }
+        return fn
+
+    return wrap
+
+
+def _event_on_load(state, ctx):
+    instrumenting_mapper = state.manager.mapper
+
+    if instrumenting_mapper._reconstructor:
+        instrumenting_mapper._reconstructor(state.obj())
+
+
+def _event_on_init(state, args, kwargs):
+    """Run init_instance hooks.
+
+    This also includes mapper compilation, normally not needed
+    here but helps with some piecemeal configuration
+    scenarios (such as in the ORM tutorial).
+
+    """
+
+    instrumenting_mapper = state.manager.mapper
+    if instrumenting_mapper:
+        instrumenting_mapper._check_configure()
+        if instrumenting_mapper._set_polymorphic_identity:
+            instrumenting_mapper._set_polymorphic_identity(state)
+
+
+class _ColumnMapping(Dict["ColumnElement[Any]", "MapperProperty[Any]"]):
+    """Error reporting helper for mapper._columntoproperty."""
+
+    __slots__ = ("mapper",)
+
+    def __init__(self, mapper):
+        # TODO: weakref would be a good idea here
+        self.mapper = mapper
+
+    def __missing__(self, column):
+        prop = self.mapper._props.get(column)
+        if prop:
+            raise orm_exc.UnmappedColumnError(
+                "Column '%s.%s' is not available, due to "
+                "conflicting property '%s':%r"
+                % (column.table.name, column.name, column.key, prop)
+            )
+        raise orm_exc.UnmappedColumnError(
+            "No column %s is configured on mapper %s..."
+            % (column, self.mapper)
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/path_registry.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/path_registry.py
new file mode 100644
index 00000000..388e4609
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/path_registry.py
@@ -0,0 +1,811 @@
+# orm/path_registry.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+"""Path tracking utilities, representing mapper graph traversals.
+
+"""
+
+from __future__ import annotations
+
+from functools import reduce
+from itertools import chain
+import logging
+import operator
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterator
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import base as orm_base
+from ._typing import insp_is_mapper_property
+from .. import exc
+from .. import util
+from ..sql import visitors
+from ..sql.cache_key import HasCacheKey
+
+if TYPE_CHECKING:
+    from ._typing import _InternalEntityType
+    from .interfaces import StrategizedProperty
+    from .mapper import Mapper
+    from .relationships import RelationshipProperty
+    from .util import AliasedInsp
+    from ..sql.cache_key import _CacheKeyTraversalType
+    from ..sql.elements import BindParameter
+    from ..sql.visitors import anon_map
+    from ..util.typing import _LiteralStar
+    from ..util.typing import TypeGuard
+
+    def is_root(path: PathRegistry) -> TypeGuard[RootRegistry]: ...
+
+    def is_entity(path: PathRegistry) -> TypeGuard[AbstractEntityRegistry]: ...
+
+else:
+    is_root = operator.attrgetter("is_root")
+    is_entity = operator.attrgetter("is_entity")
+
+
+_SerializedPath = List[Any]
+_StrPathToken = str
+_PathElementType = Union[
+    _StrPathToken, "_InternalEntityType[Any]", "StrategizedProperty[Any]"
+]
+
+# the representation is in fact
+# a tuple with alternating:
+# [_InternalEntityType[Any], Union[str, StrategizedProperty[Any]],
+# _InternalEntityType[Any], Union[str, StrategizedProperty[Any]], ...]
+# this might someday be a tuple of 2-tuples instead, but paths can be
+# chopped at odd intervals as well so this is less flexible
+_PathRepresentation = Tuple[_PathElementType, ...]
+
+# NOTE: these names are weird since the array is 0-indexed,
+# the "_Odd" entries are at 0, 2, 4, etc
+_OddPathRepresentation = Sequence["_InternalEntityType[Any]"]
+_EvenPathRepresentation = Sequence[Union["StrategizedProperty[Any]", str]]
+
+
+log = logging.getLogger(__name__)
+
+
+def _unreduce_path(path: _SerializedPath) -> PathRegistry:
+    return PathRegistry.deserialize(path)
+
+
+_WILDCARD_TOKEN: _LiteralStar = "*"
+_DEFAULT_TOKEN = "_sa_default"
+
+
+class PathRegistry(HasCacheKey):
+    """Represent query load paths and registry functions.
+
+    Basically represents structures like:
+
+    (<User mapper>, "orders", <Order mapper>, "items", <Item mapper>)
+
+    These structures are generated by things like
+    query options (joinedload(), subqueryload(), etc.) and are
+    used to compose keys stored in the query._attributes dictionary
+    for various options.
+
+    They are then re-composed at query compile/result row time as
+    the query is formed and as rows are fetched, where they again
+    serve to compose keys to look up options in the context.attributes
+    dictionary, which is copied from query._attributes.
+
+    The path structure has a limited amount of caching, where each
+    "root" ultimately pulls from a fixed registry associated with
+    the first mapper, that also contains elements for each of its
+    property keys.  However paths longer than two elements, which
+    are the exception rather than the rule, are generated on an
+    as-needed basis.
+
+    """
+
+    __slots__ = ()
+
+    is_token = False
+    is_root = False
+    has_entity = False
+    is_property = False
+    is_entity = False
+
+    is_unnatural: bool
+
+    path: _PathRepresentation
+    natural_path: _PathRepresentation
+    parent: Optional[PathRegistry]
+    root: RootRegistry
+
+    _cache_key_traversal: _CacheKeyTraversalType = [
+        ("path", visitors.ExtendedInternalTraversal.dp_has_cache_key_list)
+    ]
+
+    def __eq__(self, other: Any) -> bool:
+        try:
+            return other is not None and self.path == other._path_for_compare
+        except AttributeError:
+            util.warn(
+                "Comparison of PathRegistry to %r is not supported"
+                % (type(other))
+            )
+            return False
+
+    def __ne__(self, other: Any) -> bool:
+        try:
+            return other is None or self.path != other._path_for_compare
+        except AttributeError:
+            util.warn(
+                "Comparison of PathRegistry to %r is not supported"
+                % (type(other))
+            )
+            return True
+
+    @property
+    def _path_for_compare(self) -> Optional[_PathRepresentation]:
+        return self.path
+
+    def odd_element(self, index: int) -> _InternalEntityType[Any]:
+        return self.path[index]  # type: ignore
+
+    def set(self, attributes: Dict[Any, Any], key: Any, value: Any) -> None:
+        log.debug("set '%s' on path '%s' to '%s'", key, self, value)
+        attributes[(key, self.natural_path)] = value
+
+    def setdefault(
+        self, attributes: Dict[Any, Any], key: Any, value: Any
+    ) -> None:
+        log.debug("setdefault '%s' on path '%s' to '%s'", key, self, value)
+        attributes.setdefault((key, self.natural_path), value)
+
+    def get(
+        self, attributes: Dict[Any, Any], key: Any, value: Optional[Any] = None
+    ) -> Any:
+        key = (key, self.natural_path)
+        if key in attributes:
+            return attributes[key]
+        else:
+            return value
+
+    def __len__(self) -> int:
+        return len(self.path)
+
+    def __hash__(self) -> int:
+        return id(self)
+
+    @overload
+    def __getitem__(self, entity: _StrPathToken) -> TokenRegistry: ...
+
+    @overload
+    def __getitem__(self, entity: int) -> _PathElementType: ...
+
+    @overload
+    def __getitem__(self, entity: slice) -> _PathRepresentation: ...
+
+    @overload
+    def __getitem__(
+        self, entity: _InternalEntityType[Any]
+    ) -> AbstractEntityRegistry: ...
+
+    @overload
+    def __getitem__(
+        self, entity: StrategizedProperty[Any]
+    ) -> PropRegistry: ...
+
+    def __getitem__(
+        self,
+        entity: Union[
+            _StrPathToken,
+            int,
+            slice,
+            _InternalEntityType[Any],
+            StrategizedProperty[Any],
+        ],
+    ) -> Union[
+        TokenRegistry,
+        _PathElementType,
+        _PathRepresentation,
+        PropRegistry,
+        AbstractEntityRegistry,
+    ]:
+        raise NotImplementedError()
+
+    # TODO: what are we using this for?
+    @property
+    def length(self) -> int:
+        return len(self.path)
+
+    def pairs(
+        self,
+    ) -> Iterator[
+        Tuple[_InternalEntityType[Any], Union[str, StrategizedProperty[Any]]]
+    ]:
+        odd_path = cast(_OddPathRepresentation, self.path)
+        even_path = cast(_EvenPathRepresentation, odd_path)
+        for i in range(0, len(odd_path), 2):
+            yield odd_path[i], even_path[i + 1]
+
+    def contains_mapper(self, mapper: Mapper[Any]) -> bool:
+        _m_path = cast(_OddPathRepresentation, self.path)
+        for path_mapper in [_m_path[i] for i in range(0, len(_m_path), 2)]:
+            if path_mapper.mapper.isa(mapper):
+                return True
+        else:
+            return False
+
+    def contains(self, attributes: Dict[Any, Any], key: Any) -> bool:
+        return (key, self.path) in attributes
+
+    def __reduce__(self) -> Any:
+        return _unreduce_path, (self.serialize(),)
+
+    @classmethod
+    def _serialize_path(cls, path: _PathRepresentation) -> _SerializedPath:
+        _m_path = cast(_OddPathRepresentation, path)
+        _p_path = cast(_EvenPathRepresentation, path)
+
+        return list(
+            zip(
+                tuple(
+                    m.class_ if (m.is_mapper or m.is_aliased_class) else str(m)
+                    for m in [_m_path[i] for i in range(0, len(_m_path), 2)]
+                ),
+                tuple(
+                    p.key if insp_is_mapper_property(p) else str(p)
+                    for p in [_p_path[i] for i in range(1, len(_p_path), 2)]
+                )
+                + (None,),
+            )
+        )
+
+    @classmethod
+    def _deserialize_path(cls, path: _SerializedPath) -> _PathRepresentation:
+        def _deserialize_mapper_token(mcls: Any) -> Any:
+            return (
+                # note: we likely dont want configure=True here however
+                # this is maintained at the moment for backwards compatibility
+                orm_base._inspect_mapped_class(mcls, configure=True)
+                if mcls not in PathToken._intern
+                else PathToken._intern[mcls]
+            )
+
+        def _deserialize_key_token(mcls: Any, key: Any) -> Any:
+            if key is None:
+                return None
+            elif key in PathToken._intern:
+                return PathToken._intern[key]
+            else:
+                mp = orm_base._inspect_mapped_class(mcls, configure=True)
+                assert mp is not None
+                return mp.attrs[key]
+
+        p = tuple(
+            chain(
+                *[
+                    (
+                        _deserialize_mapper_token(mcls),
+                        _deserialize_key_token(mcls, key),
+                    )
+                    for mcls, key in path
+                ]
+            )
+        )
+        if p and p[-1] is None:
+            p = p[0:-1]
+        return p
+
+    def serialize(self) -> _SerializedPath:
+        path = self.path
+        return self._serialize_path(path)
+
+    @classmethod
+    def deserialize(cls, path: _SerializedPath) -> PathRegistry:
+        assert path is not None
+        p = cls._deserialize_path(path)
+        return cls.coerce(p)
+
+    @overload
+    @classmethod
+    def per_mapper(cls, mapper: Mapper[Any]) -> CachingEntityRegistry: ...
+
+    @overload
+    @classmethod
+    def per_mapper(cls, mapper: AliasedInsp[Any]) -> SlotsEntityRegistry: ...
+
+    @classmethod
+    def per_mapper(
+        cls, mapper: _InternalEntityType[Any]
+    ) -> AbstractEntityRegistry:
+        if mapper.is_mapper:
+            return CachingEntityRegistry(cls.root, mapper)
+        else:
+            return SlotsEntityRegistry(cls.root, mapper)
+
+    @classmethod
+    def coerce(cls, raw: _PathRepresentation) -> PathRegistry:
+        def _red(prev: PathRegistry, next_: _PathElementType) -> PathRegistry:
+            return prev[next_]
+
+        # can't quite get mypy to appreciate this one :)
+        return reduce(_red, raw, cls.root)  # type: ignore
+
+    def __add__(self, other: PathRegistry) -> PathRegistry:
+        def _red(prev: PathRegistry, next_: _PathElementType) -> PathRegistry:
+            return prev[next_]
+
+        return reduce(_red, other.path, self)
+
+    def __str__(self) -> str:
+        return f"ORM Path[{' -> '.join(str(elem) for elem in self.path)}]"
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.path!r})"
+
+
+class CreatesToken(PathRegistry):
+    __slots__ = ()
+
+    is_aliased_class: bool
+    is_root: bool
+
+    def token(self, token: _StrPathToken) -> TokenRegistry:
+        if token.endswith(f":{_WILDCARD_TOKEN}"):
+            return TokenRegistry(self, token)
+        elif token.endswith(f":{_DEFAULT_TOKEN}"):
+            return TokenRegistry(self.root, token)
+        else:
+            raise exc.ArgumentError(f"invalid token: {token}")
+
+
+class RootRegistry(CreatesToken):
+    """Root registry, defers to mappers so that
+    paths are maintained per-root-mapper.
+
+    """
+
+    __slots__ = ()
+
+    inherit_cache = True
+
+    path = natural_path = ()
+    has_entity = False
+    is_aliased_class = False
+    is_root = True
+    is_unnatural = False
+
+    def _getitem(
+        self, entity: Any
+    ) -> Union[TokenRegistry, AbstractEntityRegistry]:
+        if entity in PathToken._intern:
+            if TYPE_CHECKING:
+                assert isinstance(entity, _StrPathToken)
+            return TokenRegistry(self, PathToken._intern[entity])
+        else:
+            try:
+                return entity._path_registry  # type: ignore
+            except AttributeError:
+                raise IndexError(
+                    f"invalid argument for RootRegistry.__getitem__: {entity}"
+                )
+
+    def _truncate_recursive(self) -> RootRegistry:
+        return self
+
+    if not TYPE_CHECKING:
+        __getitem__ = _getitem
+
+
+PathRegistry.root = RootRegistry()
+
+
+class PathToken(orm_base.InspectionAttr, HasCacheKey, str):
+    """cacheable string token"""
+
+    _intern: Dict[str, PathToken] = {}
+
+    def _gen_cache_key(
+        self, anon_map: anon_map, bindparams: List[BindParameter[Any]]
+    ) -> Tuple[Any, ...]:
+        return (str(self),)
+
+    @property
+    def _path_for_compare(self) -> Optional[_PathRepresentation]:
+        return None
+
+    @classmethod
+    def intern(cls, strvalue: str) -> PathToken:
+        if strvalue in cls._intern:
+            return cls._intern[strvalue]
+        else:
+            cls._intern[strvalue] = result = PathToken(strvalue)
+            return result
+
+
+class TokenRegistry(PathRegistry):
+    __slots__ = ("token", "parent", "path", "natural_path")
+
+    inherit_cache = True
+
+    token: _StrPathToken
+    parent: CreatesToken
+
+    def __init__(self, parent: CreatesToken, token: _StrPathToken):
+        token = PathToken.intern(token)
+
+        self.token = token
+        self.parent = parent
+        self.path = parent.path + (token,)
+        self.natural_path = parent.natural_path + (token,)
+
+    has_entity = False
+
+    is_token = True
+
+    def generate_for_superclasses(self) -> Iterator[PathRegistry]:
+        # NOTE: this method is no longer used.  consider removal
+        parent = self.parent
+        if is_root(parent):
+            yield self
+            return
+
+        if TYPE_CHECKING:
+            assert isinstance(parent, AbstractEntityRegistry)
+        if not parent.is_aliased_class:
+            for mp_ent in parent.mapper.iterate_to_root():
+                yield TokenRegistry(parent.parent[mp_ent], self.token)
+        elif (
+            parent.is_aliased_class
+            and cast(
+                "AliasedInsp[Any]",
+                parent.entity,
+            )._is_with_polymorphic
+        ):
+            yield self
+            for ent in cast(
+                "AliasedInsp[Any]", parent.entity
+            )._with_polymorphic_entities:
+                yield TokenRegistry(parent.parent[ent], self.token)
+        else:
+            yield self
+
+    def _generate_natural_for_superclasses(
+        self,
+    ) -> Iterator[_PathRepresentation]:
+        parent = self.parent
+        if is_root(parent):
+            yield self.natural_path
+            return
+
+        if TYPE_CHECKING:
+            assert isinstance(parent, AbstractEntityRegistry)
+        for mp_ent in parent.mapper.iterate_to_root():
+            yield TokenRegistry(parent.parent[mp_ent], self.token).natural_path
+        if (
+            parent.is_aliased_class
+            and cast(
+                "AliasedInsp[Any]",
+                parent.entity,
+            )._is_with_polymorphic
+        ):
+            yield self.natural_path
+            for ent in cast(
+                "AliasedInsp[Any]", parent.entity
+            )._with_polymorphic_entities:
+                yield (
+                    TokenRegistry(parent.parent[ent], self.token).natural_path
+                )
+        else:
+            yield self.natural_path
+
+    def _getitem(self, entity: Any) -> Any:
+        try:
+            return self.path[entity]
+        except TypeError as err:
+            raise IndexError(f"{entity}") from err
+
+    if not TYPE_CHECKING:
+        __getitem__ = _getitem
+
+
+class PropRegistry(PathRegistry):
+    __slots__ = (
+        "prop",
+        "parent",
+        "path",
+        "natural_path",
+        "has_entity",
+        "entity",
+        "mapper",
+        "_wildcard_path_loader_key",
+        "_default_path_loader_key",
+        "_loader_key",
+        "is_unnatural",
+    )
+    inherit_cache = True
+    is_property = True
+
+    prop: StrategizedProperty[Any]
+    mapper: Optional[Mapper[Any]]
+    entity: Optional[_InternalEntityType[Any]]
+
+    def __init__(
+        self, parent: AbstractEntityRegistry, prop: StrategizedProperty[Any]
+    ):
+
+        # restate this path in terms of the
+        # given StrategizedProperty's parent.
+        insp = cast("_InternalEntityType[Any]", parent[-1])
+        natural_parent: AbstractEntityRegistry = parent
+
+        # inherit "is_unnatural" from the parent
+        self.is_unnatural = parent.parent.is_unnatural or bool(
+            parent.mapper.inherits
+        )
+
+        if not insp.is_aliased_class or insp._use_mapper_path:  # type: ignore
+            parent = natural_parent = parent.parent[prop.parent]
+        elif (
+            insp.is_aliased_class
+            and insp.with_polymorphic_mappers
+            and prop.parent in insp.with_polymorphic_mappers
+        ):
+            subclass_entity: _InternalEntityType[Any] = parent[-1]._entity_for_mapper(prop.parent)  # type: ignore  # noqa: E501
+            parent = parent.parent[subclass_entity]
+
+            # when building a path where with_polymorphic() is in use,
+            # special logic to determine the "natural path" when subclass
+            # entities are used.
+            #
+            # here we are trying to distinguish between a path that starts
+            # on a with_polymorphic entity vs. one that starts on a
+            # normal entity that introduces a with_polymorphic() in the
+            # middle using of_type():
+            #
+            #  # as in test_polymorphic_rel->
+            #  #    test_subqueryload_on_subclass_uses_path_correctly
+            #  wp = with_polymorphic(RegularEntity, "*")
+            #  sess.query(wp).options(someload(wp.SomeSubEntity.foos))
+            #
+            # vs
+            #
+            #  # as in test_relationship->JoinedloadWPolyOfTypeContinued
+            #  wp = with_polymorphic(SomeFoo, "*")
+            #  sess.query(RegularEntity).options(
+            #       someload(RegularEntity.foos.of_type(wp))
+            #       .someload(wp.SubFoo.bar)
+            #   )
+            #
+            # in the former case, the Query as it generates a path that we
+            # want to match will be in terms of the with_polymorphic at the
+            # beginning.  in the latter case, Query will generate simple
+            # paths that don't know about this with_polymorphic, so we must
+            # use a separate natural path.
+            #
+            #
+            if parent.parent:
+                natural_parent = parent.parent[subclass_entity.mapper]
+                self.is_unnatural = True
+            else:
+                natural_parent = parent
+        elif (
+            natural_parent.parent
+            and insp.is_aliased_class
+            and prop.parent  # this should always be the case here
+            is not insp.mapper
+            and insp.mapper.isa(prop.parent)
+        ):
+            natural_parent = parent.parent[prop.parent]
+
+        self.prop = prop
+        self.parent = parent
+        self.path = parent.path + (prop,)
+        self.natural_path = natural_parent.natural_path + (prop,)
+
+        self.has_entity = prop._links_to_entity
+        if prop._is_relationship:
+            if TYPE_CHECKING:
+                assert isinstance(prop, RelationshipProperty)
+            self.entity = prop.entity
+            self.mapper = prop.mapper
+        else:
+            self.entity = None
+            self.mapper = None
+
+        self._wildcard_path_loader_key = (
+            "loader",
+            parent.natural_path + self.prop._wildcard_token,
+        )
+        self._default_path_loader_key = self.prop._default_path_loader_key
+        self._loader_key = ("loader", self.natural_path)
+
+    def _truncate_recursive(self) -> PropRegistry:
+        earliest = None
+        for i, token in enumerate(reversed(self.path[:-1])):
+            if token is self.prop:
+                earliest = i
+
+        if earliest is None:
+            return self
+        else:
+            return self.coerce(self.path[0 : -(earliest + 1)])  # type: ignore
+
+    @property
+    def entity_path(self) -> AbstractEntityRegistry:
+        assert self.entity is not None
+        return self[self.entity]
+
+    def _getitem(
+        self, entity: Union[int, slice, _InternalEntityType[Any]]
+    ) -> Union[AbstractEntityRegistry, _PathElementType, _PathRepresentation]:
+        if isinstance(entity, (int, slice)):
+            return self.path[entity]
+        else:
+            return SlotsEntityRegistry(self, entity)
+
+    if not TYPE_CHECKING:
+        __getitem__ = _getitem
+
+
+class AbstractEntityRegistry(CreatesToken):
+    __slots__ = (
+        "key",
+        "parent",
+        "is_aliased_class",
+        "path",
+        "entity",
+        "natural_path",
+    )
+
+    has_entity = True
+    is_entity = True
+
+    parent: Union[RootRegistry, PropRegistry]
+    key: _InternalEntityType[Any]
+    entity: _InternalEntityType[Any]
+    is_aliased_class: bool
+
+    def __init__(
+        self,
+        parent: Union[RootRegistry, PropRegistry],
+        entity: _InternalEntityType[Any],
+    ):
+        self.key = entity
+        self.parent = parent
+        self.is_aliased_class = entity.is_aliased_class
+        self.entity = entity
+        self.path = parent.path + (entity,)
+
+        # the "natural path" is the path that we get when Query is traversing
+        # from the lead entities into the various relationships; it corresponds
+        # to the structure of mappers and relationships. when we are given a
+        # path that comes from loader options, as of 1.3 it can have ac-hoc
+        # with_polymorphic() and other AliasedInsp objects inside of it, which
+        # are usually not present in mappings.  So here we track both the
+        # "enhanced" path in self.path and the "natural" path that doesn't
+        # include those objects so these two traversals can be matched up.
+
+        # the test here for "(self.is_aliased_class or parent.is_unnatural)"
+        # are to avoid the more expensive conditional logic that follows if we
+        # know we don't have to do it.   This conditional can just as well be
+        # "if parent.path:", it just is more function calls.
+        #
+        # This is basically the only place that the "is_unnatural" flag
+        # actually changes behavior.
+        if parent.path and (self.is_aliased_class or parent.is_unnatural):
+            # this is an infrequent code path used only for loader strategies
+            # that also make use of of_type().
+            if entity.mapper.isa(parent.natural_path[-1].mapper):  # type: ignore # noqa: E501
+                self.natural_path = parent.natural_path + (entity.mapper,)
+            else:
+                self.natural_path = parent.natural_path + (
+                    parent.natural_path[-1].entity,  # type: ignore
+                )
+        # it seems to make sense that since these paths get mixed up
+        # with statements that are cached or not, we should make
+        # sure the natural path is cacheable across different occurrences
+        # of equivalent AliasedClass objects.  however, so far this
+        # does not seem to be needed for whatever reason.
+        # elif not parent.path and self.is_aliased_class:
+        #     self.natural_path = (self.entity._generate_cache_key()[0], )
+        else:
+            self.natural_path = self.path
+
+    def _truncate_recursive(self) -> AbstractEntityRegistry:
+        return self.parent._truncate_recursive()[self.entity]
+
+    @property
+    def root_entity(self) -> _InternalEntityType[Any]:
+        return self.odd_element(0)
+
+    @property
+    def entity_path(self) -> PathRegistry:
+        return self
+
+    @property
+    def mapper(self) -> Mapper[Any]:
+        return self.entity.mapper
+
+    def __bool__(self) -> bool:
+        return True
+
+    def _getitem(
+        self, entity: Any
+    ) -> Union[_PathElementType, _PathRepresentation, PathRegistry]:
+        if isinstance(entity, (int, slice)):
+            return self.path[entity]
+        elif entity in PathToken._intern:
+            return TokenRegistry(self, PathToken._intern[entity])
+        else:
+            return PropRegistry(self, entity)
+
+    if not TYPE_CHECKING:
+        __getitem__ = _getitem
+
+
+class SlotsEntityRegistry(AbstractEntityRegistry):
+    # for aliased class, return lightweight, no-cycles created
+    # version
+    inherit_cache = True
+
+
+class _ERDict(Dict[Any, Any]):
+    def __init__(self, registry: CachingEntityRegistry):
+        self.registry = registry
+
+    def __missing__(self, key: Any) -> PropRegistry:
+        self[key] = item = PropRegistry(self.registry, key)
+
+        return item
+
+
+class CachingEntityRegistry(AbstractEntityRegistry):
+    # for long lived mapper, return dict based caching
+    # version that creates reference cycles
+
+    __slots__ = ("_cache",)
+
+    inherit_cache = True
+
+    def __init__(
+        self,
+        parent: Union[RootRegistry, PropRegistry],
+        entity: _InternalEntityType[Any],
+    ):
+        super().__init__(parent, entity)
+        self._cache = _ERDict(self)
+
+    def pop(self, key: Any, default: Any) -> Any:
+        return self._cache.pop(key, default)
+
+    def _getitem(self, entity: Any) -> Any:
+        if isinstance(entity, (int, slice)):
+            return self.path[entity]
+        elif isinstance(entity, PathToken):
+            return TokenRegistry(self, entity)
+        else:
+            return self._cache[entity]
+
+    if not TYPE_CHECKING:
+        __getitem__ = _getitem
+
+
+if TYPE_CHECKING:
+
+    def path_is_entity(
+        path: PathRegistry,
+    ) -> TypeGuard[AbstractEntityRegistry]: ...
+
+    def path_is_property(path: PathRegistry) -> TypeGuard[PropRegistry]: ...
+
+else:
+    path_is_entity = operator.attrgetter("is_entity")
+    path_is_property = operator.attrgetter("is_property")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/persistence.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/persistence.py
new file mode 100644
index 00000000..cbe8557a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/persistence.py
@@ -0,0 +1,1782 @@
+# orm/persistence.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""private module containing functions used to emit INSERT, UPDATE
+and DELETE statements on behalf of a :class:`_orm.Mapper` and its descending
+mappers.
+
+The functions here are called only by the unit of work functions
+in unitofwork.py.
+
+"""
+from __future__ import annotations
+
+from itertools import chain
+from itertools import groupby
+from itertools import zip_longest
+import operator
+
+from . import attributes
+from . import exc as orm_exc
+from . import loading
+from . import sync
+from .base import state_str
+from .. import exc as sa_exc
+from .. import future
+from .. import sql
+from .. import util
+from ..engine import cursor as _cursor
+from ..sql import operators
+from ..sql.elements import BooleanClauseList
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+
+
+def save_obj(base_mapper, states, uowtransaction, single=False):
+    """Issue ``INSERT`` and/or ``UPDATE`` statements for a list
+    of objects.
+
+    This is called within the context of a UOWTransaction during a
+    flush operation, given a list of states to be flushed.  The
+    base mapper in an inheritance hierarchy handles the inserts/
+    updates for all descendant mappers.
+
+    """
+
+    # if batch=false, call _save_obj separately for each object
+    if not single and not base_mapper.batch:
+        for state in _sort_states(base_mapper, states):
+            save_obj(base_mapper, [state], uowtransaction, single=True)
+        return
+
+    states_to_update = []
+    states_to_insert = []
+
+    for (
+        state,
+        dict_,
+        mapper,
+        connection,
+        has_identity,
+        row_switch,
+        update_version_id,
+    ) in _organize_states_for_save(base_mapper, states, uowtransaction):
+        if has_identity or row_switch:
+            states_to_update.append(
+                (state, dict_, mapper, connection, update_version_id)
+            )
+        else:
+            states_to_insert.append((state, dict_, mapper, connection))
+
+    for table, mapper in base_mapper._sorted_tables.items():
+        if table not in mapper._pks_by_table:
+            continue
+        insert = _collect_insert_commands(table, states_to_insert)
+
+        update = _collect_update_commands(
+            uowtransaction, table, states_to_update
+        )
+
+        _emit_update_statements(
+            base_mapper,
+            uowtransaction,
+            mapper,
+            table,
+            update,
+        )
+
+        _emit_insert_statements(
+            base_mapper,
+            uowtransaction,
+            mapper,
+            table,
+            insert,
+        )
+
+    _finalize_insert_update_commands(
+        base_mapper,
+        uowtransaction,
+        chain(
+            (
+                (state, state_dict, mapper, connection, False)
+                for (state, state_dict, mapper, connection) in states_to_insert
+            ),
+            (
+                (state, state_dict, mapper, connection, True)
+                for (
+                    state,
+                    state_dict,
+                    mapper,
+                    connection,
+                    update_version_id,
+                ) in states_to_update
+            ),
+        ),
+    )
+
+
+def post_update(base_mapper, states, uowtransaction, post_update_cols):
+    """Issue UPDATE statements on behalf of a relationship() which
+    specifies post_update.
+
+    """
+
+    states_to_update = list(
+        _organize_states_for_post_update(base_mapper, states, uowtransaction)
+    )
+
+    for table, mapper in base_mapper._sorted_tables.items():
+        if table not in mapper._pks_by_table:
+            continue
+
+        update = (
+            (
+                state,
+                state_dict,
+                sub_mapper,
+                connection,
+                (
+                    mapper._get_committed_state_attr_by_column(
+                        state, state_dict, mapper.version_id_col
+                    )
+                    if mapper.version_id_col is not None
+                    else None
+                ),
+            )
+            for state, state_dict, sub_mapper, connection in states_to_update
+            if table in sub_mapper._pks_by_table
+        )
+
+        update = _collect_post_update_commands(
+            base_mapper, uowtransaction, table, update, post_update_cols
+        )
+
+        _emit_post_update_statements(
+            base_mapper,
+            uowtransaction,
+            mapper,
+            table,
+            update,
+        )
+
+
+def delete_obj(base_mapper, states, uowtransaction):
+    """Issue ``DELETE`` statements for a list of objects.
+
+    This is called within the context of a UOWTransaction during a
+    flush operation.
+
+    """
+
+    states_to_delete = list(
+        _organize_states_for_delete(base_mapper, states, uowtransaction)
+    )
+
+    table_to_mapper = base_mapper._sorted_tables
+
+    for table in reversed(list(table_to_mapper.keys())):
+        mapper = table_to_mapper[table]
+        if table not in mapper._pks_by_table:
+            continue
+        elif mapper.inherits and mapper.passive_deletes:
+            continue
+
+        delete = _collect_delete_commands(
+            base_mapper, uowtransaction, table, states_to_delete
+        )
+
+        _emit_delete_statements(
+            base_mapper,
+            uowtransaction,
+            mapper,
+            table,
+            delete,
+        )
+
+    for (
+        state,
+        state_dict,
+        mapper,
+        connection,
+        update_version_id,
+    ) in states_to_delete:
+        mapper.dispatch.after_delete(mapper, connection, state)
+
+
+def _organize_states_for_save(base_mapper, states, uowtransaction):
+    """Make an initial pass across a set of states for INSERT or
+    UPDATE.
+
+    This includes splitting out into distinct lists for
+    each, calling before_insert/before_update, obtaining
+    key information for each state including its dictionary,
+    mapper, the connection to use for the execution per state,
+    and the identity flag.
+
+    """
+
+    for state, dict_, mapper, connection in _connections_for_states(
+        base_mapper, uowtransaction, states
+    ):
+        has_identity = bool(state.key)
+
+        instance_key = state.key or mapper._identity_key_from_state(state)
+
+        row_switch = update_version_id = None
+
+        # call before_XXX extensions
+        if not has_identity:
+            mapper.dispatch.before_insert(mapper, connection, state)
+        else:
+            mapper.dispatch.before_update(mapper, connection, state)
+
+        if mapper._validate_polymorphic_identity:
+            mapper._validate_polymorphic_identity(mapper, state, dict_)
+
+        # detect if we have a "pending" instance (i.e. has
+        # no instance_key attached to it), and another instance
+        # with the same identity key already exists as persistent.
+        # convert to an UPDATE if so.
+        if (
+            not has_identity
+            and instance_key in uowtransaction.session.identity_map
+        ):
+            instance = uowtransaction.session.identity_map[instance_key]
+            existing = attributes.instance_state(instance)
+
+            if not uowtransaction.was_already_deleted(existing):
+                if not uowtransaction.is_deleted(existing):
+                    util.warn(
+                        "New instance %s with identity key %s conflicts "
+                        "with persistent instance %s"
+                        % (state_str(state), instance_key, state_str(existing))
+                    )
+                else:
+                    base_mapper._log_debug(
+                        "detected row switch for identity %s.  "
+                        "will update %s, remove %s from "
+                        "transaction",
+                        instance_key,
+                        state_str(state),
+                        state_str(existing),
+                    )
+
+                    # remove the "delete" flag from the existing element
+                    uowtransaction.remove_state_actions(existing)
+                    row_switch = existing
+
+        if (has_identity or row_switch) and mapper.version_id_col is not None:
+            update_version_id = mapper._get_committed_state_attr_by_column(
+                row_switch if row_switch else state,
+                row_switch.dict if row_switch else dict_,
+                mapper.version_id_col,
+            )
+
+        yield (
+            state,
+            dict_,
+            mapper,
+            connection,
+            has_identity,
+            row_switch,
+            update_version_id,
+        )
+
+
+def _organize_states_for_post_update(base_mapper, states, uowtransaction):
+    """Make an initial pass across a set of states for UPDATE
+    corresponding to post_update.
+
+    This includes obtaining key information for each state
+    including its dictionary, mapper, the connection to use for
+    the execution per state.
+
+    """
+    return _connections_for_states(base_mapper, uowtransaction, states)
+
+
+def _organize_states_for_delete(base_mapper, states, uowtransaction):
+    """Make an initial pass across a set of states for DELETE.
+
+    This includes calling out before_delete and obtaining
+    key information for each state including its dictionary,
+    mapper, the connection to use for the execution per state.
+
+    """
+    for state, dict_, mapper, connection in _connections_for_states(
+        base_mapper, uowtransaction, states
+    ):
+        mapper.dispatch.before_delete(mapper, connection, state)
+
+        if mapper.version_id_col is not None:
+            update_version_id = mapper._get_committed_state_attr_by_column(
+                state, dict_, mapper.version_id_col
+            )
+        else:
+            update_version_id = None
+
+        yield (state, dict_, mapper, connection, update_version_id)
+
+
+def _collect_insert_commands(
+    table,
+    states_to_insert,
+    *,
+    bulk=False,
+    return_defaults=False,
+    render_nulls=False,
+    include_bulk_keys=(),
+):
+    """Identify sets of values to use in INSERT statements for a
+    list of states.
+
+    """
+    for state, state_dict, mapper, connection in states_to_insert:
+        if table not in mapper._pks_by_table:
+            continue
+
+        params = {}
+        value_params = {}
+
+        propkey_to_col = mapper._propkey_to_col[table]
+
+        eval_none = mapper._insert_cols_evaluating_none[table]
+
+        for propkey in set(propkey_to_col).intersection(state_dict):
+            value = state_dict[propkey]
+            col = propkey_to_col[propkey]
+            if value is None and col not in eval_none and not render_nulls:
+                continue
+            elif not bulk and (
+                hasattr(value, "__clause_element__")
+                or isinstance(value, sql.ClauseElement)
+            ):
+                value_params[col] = (
+                    value.__clause_element__()
+                    if hasattr(value, "__clause_element__")
+                    else value
+                )
+            else:
+                params[col.key] = value
+
+        if not bulk:
+            # for all the columns that have no default and we don't have
+            # a value and where "None" is not a special value, add
+            # explicit None to the INSERT.   This is a legacy behavior
+            # which might be worth removing, as it should not be necessary
+            # and also produces confusion, given that "missing" and None
+            # now have distinct meanings
+            for colkey in (
+                mapper._insert_cols_as_none[table]
+                .difference(params)
+                .difference([c.key for c in value_params])
+            ):
+                params[colkey] = None
+
+        if not bulk or return_defaults:
+            # params are in terms of Column key objects, so
+            # compare to pk_keys_by_table
+            has_all_pks = mapper._pk_keys_by_table[table].issubset(params)
+
+            if mapper.base_mapper._prefer_eager_defaults(
+                connection.dialect, table
+            ):
+                has_all_defaults = mapper._server_default_col_keys[
+                    table
+                ].issubset(params)
+            else:
+                has_all_defaults = True
+        else:
+            has_all_defaults = has_all_pks = True
+
+        if (
+            mapper.version_id_generator is not False
+            and mapper.version_id_col is not None
+            and mapper.version_id_col in mapper._cols_by_table[table]
+        ):
+            params[mapper.version_id_col.key] = mapper.version_id_generator(
+                None
+            )
+
+        if bulk:
+            if mapper._set_polymorphic_identity:
+                params.setdefault(
+                    mapper._polymorphic_attr_key, mapper.polymorphic_identity
+                )
+
+            if include_bulk_keys:
+                params.update((k, state_dict[k]) for k in include_bulk_keys)
+
+        yield (
+            state,
+            state_dict,
+            params,
+            mapper,
+            connection,
+            value_params,
+            has_all_pks,
+            has_all_defaults,
+        )
+
+
+def _collect_update_commands(
+    uowtransaction,
+    table,
+    states_to_update,
+    *,
+    bulk=False,
+    use_orm_update_stmt=None,
+    include_bulk_keys=(),
+):
+    """Identify sets of values to use in UPDATE statements for a
+    list of states.
+
+    This function works intricately with the history system
+    to determine exactly what values should be updated
+    as well as how the row should be matched within an UPDATE
+    statement.  Includes some tricky scenarios where the primary
+    key of an object might have been changed.
+
+    """
+
+    for (
+        state,
+        state_dict,
+        mapper,
+        connection,
+        update_version_id,
+    ) in states_to_update:
+        if table not in mapper._pks_by_table:
+            continue
+
+        pks = mapper._pks_by_table[table]
+
+        if use_orm_update_stmt is not None:
+            # TODO: ordered values, etc
+            value_params = use_orm_update_stmt._values
+        else:
+            value_params = {}
+
+        propkey_to_col = mapper._propkey_to_col[table]
+
+        if bulk:
+            # keys here are mapped attribute keys, so
+            # look at mapper attribute keys for pk
+            params = {
+                propkey_to_col[propkey].key: state_dict[propkey]
+                for propkey in set(propkey_to_col)
+                .intersection(state_dict)
+                .difference(mapper._pk_attr_keys_by_table[table])
+            }
+            has_all_defaults = True
+        else:
+            params = {}
+            for propkey in set(propkey_to_col).intersection(
+                state.committed_state
+            ):
+                value = state_dict[propkey]
+                col = propkey_to_col[propkey]
+
+                if hasattr(value, "__clause_element__") or isinstance(
+                    value, sql.ClauseElement
+                ):
+                    value_params[col] = (
+                        value.__clause_element__()
+                        if hasattr(value, "__clause_element__")
+                        else value
+                    )
+                # guard against values that generate non-__nonzero__
+                # objects for __eq__()
+                elif (
+                    state.manager[propkey].impl.is_equal(
+                        value, state.committed_state[propkey]
+                    )
+                    is not True
+                ):
+                    params[col.key] = value
+
+            if mapper.base_mapper.eager_defaults is True:
+                has_all_defaults = (
+                    mapper._server_onupdate_default_col_keys[table]
+                ).issubset(params)
+            else:
+                has_all_defaults = True
+
+        if (
+            update_version_id is not None
+            and mapper.version_id_col in mapper._cols_by_table[table]
+        ):
+            if not bulk and not (params or value_params):
+                # HACK: check for history in other tables, in case the
+                # history is only in a different table than the one
+                # where the version_id_col is.  This logic was lost
+                # from 0.9 -> 1.0.0 and restored in 1.0.6.
+                for prop in mapper._columntoproperty.values():
+                    history = state.manager[prop.key].impl.get_history(
+                        state, state_dict, attributes.PASSIVE_NO_INITIALIZE
+                    )
+                    if history.added:
+                        break
+                else:
+                    # no net change, break
+                    continue
+
+            col = mapper.version_id_col
+            no_params = not params and not value_params
+            params[col._label] = update_version_id
+
+            if (
+                bulk or col.key not in params
+            ) and mapper.version_id_generator is not False:
+                val = mapper.version_id_generator(update_version_id)
+                params[col.key] = val
+            elif mapper.version_id_generator is False and no_params:
+                # no version id generator, no values set on the table,
+                # and version id wasn't manually incremented.
+                # set version id to itself so we get an UPDATE
+                # statement
+                params[col.key] = update_version_id
+
+        elif not (params or value_params):
+            continue
+
+        has_all_pks = True
+        expect_pk_cascaded = False
+        if bulk:
+            # keys here are mapped attribute keys, so
+            # look at mapper attribute keys for pk
+            pk_params = {
+                propkey_to_col[propkey]._label: state_dict.get(propkey)
+                for propkey in set(propkey_to_col).intersection(
+                    mapper._pk_attr_keys_by_table[table]
+                )
+            }
+            if util.NONE_SET.intersection(pk_params.values()):
+                raise sa_exc.InvalidRequestError(
+                    f"No primary key value supplied for column(s) "
+                    f"""{
+                        ', '.join(
+                            str(c) for c in pks if pk_params[c._label] is None
+                        )
+                    }; """
+                    "per-row ORM Bulk UPDATE by Primary Key requires that "
+                    "records contain primary key values",
+                    code="bupq",
+                )
+
+        else:
+            pk_params = {}
+            for col in pks:
+                propkey = mapper._columntoproperty[col].key
+
+                history = state.manager[propkey].impl.get_history(
+                    state, state_dict, attributes.PASSIVE_OFF
+                )
+
+                if history.added:
+                    if (
+                        not history.deleted
+                        or ("pk_cascaded", state, col)
+                        in uowtransaction.attributes
+                    ):
+                        expect_pk_cascaded = True
+                        pk_params[col._label] = history.added[0]
+                        params.pop(col.key, None)
+                    else:
+                        # else, use the old value to locate the row
+                        pk_params[col._label] = history.deleted[0]
+                        if col in value_params:
+                            has_all_pks = False
+                else:
+                    pk_params[col._label] = history.unchanged[0]
+                if pk_params[col._label] is None:
+                    raise orm_exc.FlushError(
+                        "Can't update table %s using NULL for primary "
+                        "key value on column %s" % (table, col)
+                    )
+
+        if include_bulk_keys:
+            params.update((k, state_dict[k]) for k in include_bulk_keys)
+
+        if params or value_params:
+            params.update(pk_params)
+            yield (
+                state,
+                state_dict,
+                params,
+                mapper,
+                connection,
+                value_params,
+                has_all_defaults,
+                has_all_pks,
+            )
+        elif expect_pk_cascaded:
+            # no UPDATE occurs on this table, but we expect that CASCADE rules
+            # have changed the primary key of the row; propagate this event to
+            # other columns that expect to have been modified. this normally
+            # occurs after the UPDATE is emitted however we invoke it here
+            # explicitly in the absence of our invoking an UPDATE
+            for m, equated_pairs in mapper._table_to_equated[table]:
+                sync.populate(
+                    state,
+                    m,
+                    state,
+                    m,
+                    equated_pairs,
+                    uowtransaction,
+                    mapper.passive_updates,
+                )
+
+
+def _collect_post_update_commands(
+    base_mapper, uowtransaction, table, states_to_update, post_update_cols
+):
+    """Identify sets of values to use in UPDATE statements for a
+    list of states within a post_update operation.
+
+    """
+
+    for (
+        state,
+        state_dict,
+        mapper,
+        connection,
+        update_version_id,
+    ) in states_to_update:
+        # assert table in mapper._pks_by_table
+
+        pks = mapper._pks_by_table[table]
+        params = {}
+        hasdata = False
+
+        for col in mapper._cols_by_table[table]:
+            if col in pks:
+                params[col._label] = mapper._get_state_attr_by_column(
+                    state, state_dict, col, passive=attributes.PASSIVE_OFF
+                )
+
+            elif col in post_update_cols or col.onupdate is not None:
+                prop = mapper._columntoproperty[col]
+                history = state.manager[prop.key].impl.get_history(
+                    state, state_dict, attributes.PASSIVE_NO_INITIALIZE
+                )
+                if history.added:
+                    value = history.added[0]
+                    params[col.key] = value
+                    hasdata = True
+        if hasdata:
+            if (
+                update_version_id is not None
+                and mapper.version_id_col in mapper._cols_by_table[table]
+            ):
+                col = mapper.version_id_col
+                params[col._label] = update_version_id
+
+                if (
+                    bool(state.key)
+                    and col.key not in params
+                    and mapper.version_id_generator is not False
+                ):
+                    val = mapper.version_id_generator(update_version_id)
+                    params[col.key] = val
+            yield state, state_dict, mapper, connection, params
+
+
+def _collect_delete_commands(
+    base_mapper, uowtransaction, table, states_to_delete
+):
+    """Identify values to use in DELETE statements for a list of
+    states to be deleted."""
+
+    for (
+        state,
+        state_dict,
+        mapper,
+        connection,
+        update_version_id,
+    ) in states_to_delete:
+        if table not in mapper._pks_by_table:
+            continue
+
+        params = {}
+        for col in mapper._pks_by_table[table]:
+            params[col.key] = value = (
+                mapper._get_committed_state_attr_by_column(
+                    state, state_dict, col
+                )
+            )
+            if value is None:
+                raise orm_exc.FlushError(
+                    "Can't delete from table %s "
+                    "using NULL for primary "
+                    "key value on column %s" % (table, col)
+                )
+
+        if (
+            update_version_id is not None
+            and mapper.version_id_col in mapper._cols_by_table[table]
+        ):
+            params[mapper.version_id_col.key] = update_version_id
+        yield params, connection
+
+
+def _emit_update_statements(
+    base_mapper,
+    uowtransaction,
+    mapper,
+    table,
+    update,
+    *,
+    bookkeeping=True,
+    use_orm_update_stmt=None,
+    enable_check_rowcount=True,
+):
+    """Emit UPDATE statements corresponding to value lists collected
+    by _collect_update_commands()."""
+
+    needs_version_id = (
+        mapper.version_id_col is not None
+        and mapper.version_id_col in mapper._cols_by_table[table]
+    )
+
+    execution_options = {"compiled_cache": base_mapper._compiled_cache}
+
+    def update_stmt(existing_stmt=None):
+        clauses = BooleanClauseList._construct_raw(operators.and_)
+
+        for col in mapper._pks_by_table[table]:
+            clauses._append_inplace(
+                col == sql.bindparam(col._label, type_=col.type)
+            )
+
+        if needs_version_id:
+            clauses._append_inplace(
+                mapper.version_id_col
+                == sql.bindparam(
+                    mapper.version_id_col._label,
+                    type_=mapper.version_id_col.type,
+                )
+            )
+
+        if existing_stmt is not None:
+            stmt = existing_stmt.where(clauses)
+        else:
+            stmt = table.update().where(clauses)
+        return stmt
+
+    if use_orm_update_stmt is not None:
+        cached_stmt = update_stmt(use_orm_update_stmt)
+
+    else:
+        cached_stmt = base_mapper._memo(("update", table), update_stmt)
+
+    for (
+        (connection, paramkeys, hasvalue, has_all_defaults, has_all_pks),
+        records,
+    ) in groupby(
+        update,
+        lambda rec: (
+            rec[4],  # connection
+            set(rec[2]),  # set of parameter keys
+            bool(rec[5]),  # whether or not we have "value" parameters
+            rec[6],  # has_all_defaults
+            rec[7],  # has all pks
+        ),
+    ):
+        rows = 0
+        records = list(records)
+
+        statement = cached_stmt
+
+        if use_orm_update_stmt is not None:
+            statement = statement._annotate(
+                {
+                    "_emit_update_table": table,
+                    "_emit_update_mapper": mapper,
+                }
+            )
+
+        return_defaults = False
+
+        if not has_all_pks:
+            statement = statement.return_defaults(*mapper._pks_by_table[table])
+            return_defaults = True
+
+        if (
+            bookkeeping
+            and not has_all_defaults
+            and mapper.base_mapper.eager_defaults is True
+            # change as of #8889 - if RETURNING is not going to be used anyway,
+            # (applies to MySQL, MariaDB which lack UPDATE RETURNING) ensure
+            # we can do an executemany UPDATE which is more efficient
+            and table.implicit_returning
+            and connection.dialect.update_returning
+        ):
+            statement = statement.return_defaults(
+                *mapper._server_onupdate_default_cols[table]
+            )
+            return_defaults = True
+
+        if mapper._version_id_has_server_side_value:
+            statement = statement.return_defaults(mapper.version_id_col)
+            return_defaults = True
+
+        assert_singlerow = connection.dialect.supports_sane_rowcount
+
+        assert_multirow = (
+            assert_singlerow
+            and connection.dialect.supports_sane_multi_rowcount
+        )
+
+        # change as of #8889 - if RETURNING is not going to be used anyway,
+        # (applies to MySQL, MariaDB which lack UPDATE RETURNING) ensure
+        # we can do an executemany UPDATE which is more efficient
+        allow_executemany = not return_defaults and not needs_version_id
+
+        if hasvalue:
+            for (
+                state,
+                state_dict,
+                params,
+                mapper,
+                connection,
+                value_params,
+                has_all_defaults,
+                has_all_pks,
+            ) in records:
+                c = connection.execute(
+                    statement.values(value_params),
+                    params,
+                    execution_options=execution_options,
+                )
+                if bookkeeping:
+                    _postfetch(
+                        mapper,
+                        uowtransaction,
+                        table,
+                        state,
+                        state_dict,
+                        c,
+                        c.context.compiled_parameters[0],
+                        value_params,
+                        True,
+                        c.returned_defaults,
+                    )
+                rows += c.rowcount
+                check_rowcount = enable_check_rowcount and assert_singlerow
+        else:
+            if not allow_executemany:
+                check_rowcount = enable_check_rowcount and assert_singlerow
+                for (
+                    state,
+                    state_dict,
+                    params,
+                    mapper,
+                    connection,
+                    value_params,
+                    has_all_defaults,
+                    has_all_pks,
+                ) in records:
+                    c = connection.execute(
+                        statement, params, execution_options=execution_options
+                    )
+
+                    # TODO: why with bookkeeping=False?
+                    if bookkeeping:
+                        _postfetch(
+                            mapper,
+                            uowtransaction,
+                            table,
+                            state,
+                            state_dict,
+                            c,
+                            c.context.compiled_parameters[0],
+                            value_params,
+                            True,
+                            c.returned_defaults,
+                        )
+                    rows += c.rowcount
+            else:
+                multiparams = [rec[2] for rec in records]
+
+                check_rowcount = enable_check_rowcount and (
+                    assert_multirow
+                    or (assert_singlerow and len(multiparams) == 1)
+                )
+
+                c = connection.execute(
+                    statement, multiparams, execution_options=execution_options
+                )
+
+                rows += c.rowcount
+
+                for (
+                    state,
+                    state_dict,
+                    params,
+                    mapper,
+                    connection,
+                    value_params,
+                    has_all_defaults,
+                    has_all_pks,
+                ) in records:
+                    if bookkeeping:
+                        _postfetch(
+                            mapper,
+                            uowtransaction,
+                            table,
+                            state,
+                            state_dict,
+                            c,
+                            c.context.compiled_parameters[0],
+                            value_params,
+                            True,
+                            (
+                                c.returned_defaults
+                                if not c.context.executemany
+                                else None
+                            ),
+                        )
+
+        if check_rowcount:
+            if rows != len(records):
+                raise orm_exc.StaleDataError(
+                    "UPDATE statement on table '%s' expected to "
+                    "update %d row(s); %d were matched."
+                    % (table.description, len(records), rows)
+                )
+
+        elif needs_version_id:
+            util.warn(
+                "Dialect %s does not support updated rowcount "
+                "- versioning cannot be verified."
+                % c.dialect.dialect_description
+            )
+
+
+def _emit_insert_statements(
+    base_mapper,
+    uowtransaction,
+    mapper,
+    table,
+    insert,
+    *,
+    bookkeeping=True,
+    use_orm_insert_stmt=None,
+    execution_options=None,
+):
+    """Emit INSERT statements corresponding to value lists collected
+    by _collect_insert_commands()."""
+
+    if use_orm_insert_stmt is not None:
+        cached_stmt = use_orm_insert_stmt
+        exec_opt = util.EMPTY_DICT
+
+        # if a user query with RETURNING was passed, we definitely need
+        # to use RETURNING.
+        returning_is_required_anyway = bool(use_orm_insert_stmt._returning)
+        deterministic_results_reqd = (
+            returning_is_required_anyway
+            and use_orm_insert_stmt._sort_by_parameter_order
+        ) or bookkeeping
+    else:
+        returning_is_required_anyway = False
+        deterministic_results_reqd = bookkeeping
+        cached_stmt = base_mapper._memo(("insert", table), table.insert)
+        exec_opt = {"compiled_cache": base_mapper._compiled_cache}
+
+    if execution_options:
+        execution_options = util.EMPTY_DICT.merge_with(
+            exec_opt, execution_options
+        )
+    else:
+        execution_options = exec_opt
+
+    return_result = None
+
+    for (
+        (connection, _, hasvalue, has_all_pks, has_all_defaults),
+        records,
+    ) in groupby(
+        insert,
+        lambda rec: (
+            rec[4],  # connection
+            set(rec[2]),  # parameter keys
+            bool(rec[5]),  # whether we have "value" parameters
+            rec[6],
+            rec[7],
+        ),
+    ):
+        statement = cached_stmt
+
+        if use_orm_insert_stmt is not None:
+            statement = statement._annotate(
+                {
+                    "_emit_insert_table": table,
+                    "_emit_insert_mapper": mapper,
+                }
+            )
+
+        if (
+            (
+                not bookkeeping
+                or (
+                    has_all_defaults
+                    or not base_mapper._prefer_eager_defaults(
+                        connection.dialect, table
+                    )
+                    or not table.implicit_returning
+                    or not connection.dialect.insert_returning
+                )
+            )
+            and not returning_is_required_anyway
+            and has_all_pks
+            and not hasvalue
+        ):
+            # the "we don't need newly generated values back" section.
+            # here we have all the PKs, all the defaults or we don't want
+            # to fetch them, or the dialect doesn't support RETURNING at all
+            # so we have to post-fetch / use lastrowid anyway.
+            records = list(records)
+            multiparams = [rec[2] for rec in records]
+
+            result = connection.execute(
+                statement, multiparams, execution_options=execution_options
+            )
+            if bookkeeping:
+                for (
+                    (
+                        state,
+                        state_dict,
+                        params,
+                        mapper_rec,
+                        conn,
+                        value_params,
+                        has_all_pks,
+                        has_all_defaults,
+                    ),
+                    last_inserted_params,
+                ) in zip(records, result.context.compiled_parameters):
+                    if state:
+                        _postfetch(
+                            mapper_rec,
+                            uowtransaction,
+                            table,
+                            state,
+                            state_dict,
+                            result,
+                            last_inserted_params,
+                            value_params,
+                            False,
+                            (
+                                result.returned_defaults
+                                if not result.context.executemany
+                                else None
+                            ),
+                        )
+                    else:
+                        _postfetch_bulk_save(mapper_rec, state_dict, table)
+
+        else:
+            # here, we need defaults and/or pk values back or we otherwise
+            # know that we are using RETURNING in any case
+
+            records = list(records)
+
+            if returning_is_required_anyway or (
+                table.implicit_returning and not hasvalue and len(records) > 1
+            ):
+                if (
+                    deterministic_results_reqd
+                    and connection.dialect.insert_executemany_returning_sort_by_parameter_order  # noqa: E501
+                ) or (
+                    not deterministic_results_reqd
+                    and connection.dialect.insert_executemany_returning
+                ):
+                    do_executemany = True
+                elif returning_is_required_anyway:
+                    if deterministic_results_reqd:
+                        dt = " with RETURNING and sort by parameter order"
+                    else:
+                        dt = " with RETURNING"
+                    raise sa_exc.InvalidRequestError(
+                        f"Can't use explicit RETURNING for bulk INSERT "
+                        f"operation with "
+                        f"{connection.dialect.dialect_description} backend; "
+                        f"executemany{dt} is not enabled for this dialect."
+                    )
+                else:
+                    do_executemany = False
+            else:
+                do_executemany = False
+
+            if use_orm_insert_stmt is None:
+                if (
+                    not has_all_defaults
+                    and base_mapper._prefer_eager_defaults(
+                        connection.dialect, table
+                    )
+                ):
+                    statement = statement.return_defaults(
+                        *mapper._server_default_cols[table],
+                        sort_by_parameter_order=bookkeeping,
+                    )
+
+            if mapper.version_id_col is not None:
+                statement = statement.return_defaults(
+                    mapper.version_id_col,
+                    sort_by_parameter_order=bookkeeping,
+                )
+            elif do_executemany:
+                statement = statement.return_defaults(
+                    *table.primary_key, sort_by_parameter_order=bookkeeping
+                )
+
+            if do_executemany:
+                multiparams = [rec[2] for rec in records]
+
+                result = connection.execute(
+                    statement, multiparams, execution_options=execution_options
+                )
+
+                if use_orm_insert_stmt is not None:
+                    if return_result is None:
+                        return_result = result
+                    else:
+                        return_result = return_result.splice_vertically(result)
+
+                if bookkeeping:
+                    for (
+                        (
+                            state,
+                            state_dict,
+                            params,
+                            mapper_rec,
+                            conn,
+                            value_params,
+                            has_all_pks,
+                            has_all_defaults,
+                        ),
+                        last_inserted_params,
+                        inserted_primary_key,
+                        returned_defaults,
+                    ) in zip_longest(
+                        records,
+                        result.context.compiled_parameters,
+                        result.inserted_primary_key_rows,
+                        result.returned_defaults_rows or (),
+                    ):
+                        if inserted_primary_key is None:
+                            # this is a real problem and means that we didn't
+                            # get back as many PK rows.  we can't continue
+                            # since this indicates PK rows were missing, which
+                            # means we likely mis-populated records starting
+                            # at that point with incorrectly matched PK
+                            # values.
+                            raise orm_exc.FlushError(
+                                "Multi-row INSERT statement for %s did not "
+                                "produce "
+                                "the correct number of INSERTed rows for "
+                                "RETURNING.  Ensure there are no triggers or "
+                                "special driver issues preventing INSERT from "
+                                "functioning properly." % mapper_rec
+                            )
+
+                        for pk, col in zip(
+                            inserted_primary_key,
+                            mapper._pks_by_table[table],
+                        ):
+                            prop = mapper_rec._columntoproperty[col]
+                            if state_dict.get(prop.key) is None:
+                                state_dict[prop.key] = pk
+
+                        if state:
+                            _postfetch(
+                                mapper_rec,
+                                uowtransaction,
+                                table,
+                                state,
+                                state_dict,
+                                result,
+                                last_inserted_params,
+                                value_params,
+                                False,
+                                returned_defaults,
+                            )
+                        else:
+                            _postfetch_bulk_save(mapper_rec, state_dict, table)
+            else:
+                assert not returning_is_required_anyway
+
+                for (
+                    state,
+                    state_dict,
+                    params,
+                    mapper_rec,
+                    connection,
+                    value_params,
+                    has_all_pks,
+                    has_all_defaults,
+                ) in records:
+                    if value_params:
+                        result = connection.execute(
+                            statement.values(value_params),
+                            params,
+                            execution_options=execution_options,
+                        )
+                    else:
+                        result = connection.execute(
+                            statement,
+                            params,
+                            execution_options=execution_options,
+                        )
+
+                    primary_key = result.inserted_primary_key
+                    if primary_key is None:
+                        raise orm_exc.FlushError(
+                            "Single-row INSERT statement for %s "
+                            "did not produce a "
+                            "new primary key result "
+                            "being invoked.  Ensure there are no triggers or "
+                            "special driver issues preventing INSERT from "
+                            "functioning properly." % (mapper_rec,)
+                        )
+                    for pk, col in zip(
+                        primary_key, mapper._pks_by_table[table]
+                    ):
+                        prop = mapper_rec._columntoproperty[col]
+                        if (
+                            col in value_params
+                            or state_dict.get(prop.key) is None
+                        ):
+                            state_dict[prop.key] = pk
+                    if bookkeeping:
+                        if state:
+                            _postfetch(
+                                mapper_rec,
+                                uowtransaction,
+                                table,
+                                state,
+                                state_dict,
+                                result,
+                                result.context.compiled_parameters[0],
+                                value_params,
+                                False,
+                                (
+                                    result.returned_defaults
+                                    if not result.context.executemany
+                                    else None
+                                ),
+                            )
+                        else:
+                            _postfetch_bulk_save(mapper_rec, state_dict, table)
+
+    if use_orm_insert_stmt is not None:
+        if return_result is None:
+            return _cursor.null_dml_result()
+        else:
+            return return_result
+
+
+def _emit_post_update_statements(
+    base_mapper, uowtransaction, mapper, table, update
+):
+    """Emit UPDATE statements corresponding to value lists collected
+    by _collect_post_update_commands()."""
+
+    execution_options = {"compiled_cache": base_mapper._compiled_cache}
+
+    needs_version_id = (
+        mapper.version_id_col is not None
+        and mapper.version_id_col in mapper._cols_by_table[table]
+    )
+
+    def update_stmt():
+        clauses = BooleanClauseList._construct_raw(operators.and_)
+
+        for col in mapper._pks_by_table[table]:
+            clauses._append_inplace(
+                col == sql.bindparam(col._label, type_=col.type)
+            )
+
+        if needs_version_id:
+            clauses._append_inplace(
+                mapper.version_id_col
+                == sql.bindparam(
+                    mapper.version_id_col._label,
+                    type_=mapper.version_id_col.type,
+                )
+            )
+
+        stmt = table.update().where(clauses)
+
+        return stmt
+
+    statement = base_mapper._memo(("post_update", table), update_stmt)
+
+    if mapper._version_id_has_server_side_value:
+        statement = statement.return_defaults(mapper.version_id_col)
+
+    # execute each UPDATE in the order according to the original
+    # list of states to guarantee row access order, but
+    # also group them into common (connection, cols) sets
+    # to support executemany().
+    for key, records in groupby(
+        update,
+        lambda rec: (rec[3], set(rec[4])),  # connection  # parameter keys
+    ):
+        rows = 0
+
+        records = list(records)
+        connection = key[0]
+
+        assert_singlerow = connection.dialect.supports_sane_rowcount
+        assert_multirow = (
+            assert_singlerow
+            and connection.dialect.supports_sane_multi_rowcount
+        )
+        allow_executemany = not needs_version_id or assert_multirow
+
+        if not allow_executemany:
+            check_rowcount = assert_singlerow
+            for state, state_dict, mapper_rec, connection, params in records:
+                c = connection.execute(
+                    statement, params, execution_options=execution_options
+                )
+
+                _postfetch_post_update(
+                    mapper_rec,
+                    uowtransaction,
+                    table,
+                    state,
+                    state_dict,
+                    c,
+                    c.context.compiled_parameters[0],
+                )
+                rows += c.rowcount
+        else:
+            multiparams = [
+                params
+                for state, state_dict, mapper_rec, conn, params in records
+            ]
+
+            check_rowcount = assert_multirow or (
+                assert_singlerow and len(multiparams) == 1
+            )
+
+            c = connection.execute(
+                statement, multiparams, execution_options=execution_options
+            )
+
+            rows += c.rowcount
+            for state, state_dict, mapper_rec, connection, params in records:
+                _postfetch_post_update(
+                    mapper_rec,
+                    uowtransaction,
+                    table,
+                    state,
+                    state_dict,
+                    c,
+                    c.context.compiled_parameters[0],
+                )
+
+        if check_rowcount:
+            if rows != len(records):
+                raise orm_exc.StaleDataError(
+                    "UPDATE statement on table '%s' expected to "
+                    "update %d row(s); %d were matched."
+                    % (table.description, len(records), rows)
+                )
+
+        elif needs_version_id:
+            util.warn(
+                "Dialect %s does not support updated rowcount "
+                "- versioning cannot be verified."
+                % c.dialect.dialect_description
+            )
+
+
+def _emit_delete_statements(
+    base_mapper, uowtransaction, mapper, table, delete
+):
+    """Emit DELETE statements corresponding to value lists collected
+    by _collect_delete_commands()."""
+
+    need_version_id = (
+        mapper.version_id_col is not None
+        and mapper.version_id_col in mapper._cols_by_table[table]
+    )
+
+    def delete_stmt():
+        clauses = BooleanClauseList._construct_raw(operators.and_)
+
+        for col in mapper._pks_by_table[table]:
+            clauses._append_inplace(
+                col == sql.bindparam(col.key, type_=col.type)
+            )
+
+        if need_version_id:
+            clauses._append_inplace(
+                mapper.version_id_col
+                == sql.bindparam(
+                    mapper.version_id_col.key, type_=mapper.version_id_col.type
+                )
+            )
+
+        return table.delete().where(clauses)
+
+    statement = base_mapper._memo(("delete", table), delete_stmt)
+    for connection, recs in groupby(delete, lambda rec: rec[1]):  # connection
+        del_objects = [params for params, connection in recs]
+
+        execution_options = {"compiled_cache": base_mapper._compiled_cache}
+        expected = len(del_objects)
+        rows_matched = -1
+        only_warn = False
+
+        if (
+            need_version_id
+            and not connection.dialect.supports_sane_multi_rowcount
+        ):
+            if connection.dialect.supports_sane_rowcount:
+                rows_matched = 0
+                # execute deletes individually so that versioned
+                # rows can be verified
+                for params in del_objects:
+                    c = connection.execute(
+                        statement, params, execution_options=execution_options
+                    )
+                    rows_matched += c.rowcount
+            else:
+                util.warn(
+                    "Dialect %s does not support deleted rowcount "
+                    "- versioning cannot be verified."
+                    % connection.dialect.dialect_description
+                )
+                connection.execute(
+                    statement, del_objects, execution_options=execution_options
+                )
+        else:
+            c = connection.execute(
+                statement, del_objects, execution_options=execution_options
+            )
+
+            if not need_version_id:
+                only_warn = True
+
+            rows_matched = c.rowcount
+
+        if (
+            base_mapper.confirm_deleted_rows
+            and rows_matched > -1
+            and expected != rows_matched
+            and (
+                connection.dialect.supports_sane_multi_rowcount
+                or len(del_objects) == 1
+            )
+        ):
+            # TODO: why does this "only warn" if versioning is turned off,
+            # whereas the UPDATE raises?
+            if only_warn:
+                util.warn(
+                    "DELETE statement on table '%s' expected to "
+                    "delete %d row(s); %d were matched.  Please set "
+                    "confirm_deleted_rows=False within the mapper "
+                    "configuration to prevent this warning."
+                    % (table.description, expected, rows_matched)
+                )
+            else:
+                raise orm_exc.StaleDataError(
+                    "DELETE statement on table '%s' expected to "
+                    "delete %d row(s); %d were matched.  Please set "
+                    "confirm_deleted_rows=False within the mapper "
+                    "configuration to prevent this warning."
+                    % (table.description, expected, rows_matched)
+                )
+
+
+def _finalize_insert_update_commands(base_mapper, uowtransaction, states):
+    """finalize state on states that have been inserted or updated,
+    including calling after_insert/after_update events.
+
+    """
+    for state, state_dict, mapper, connection, has_identity in states:
+        if mapper._readonly_props:
+            readonly = state.unmodified_intersection(
+                [
+                    p.key
+                    for p in mapper._readonly_props
+                    if (
+                        p.expire_on_flush
+                        and (not p.deferred or p.key in state.dict)
+                    )
+                    or (
+                        not p.expire_on_flush
+                        and not p.deferred
+                        and p.key not in state.dict
+                    )
+                ]
+            )
+            if readonly:
+                state._expire_attributes(state.dict, readonly)
+
+        # if eager_defaults option is enabled, load
+        # all expired cols.  Else if we have a version_id_col, make sure
+        # it isn't expired.
+        toload_now = []
+
+        # this is specifically to emit a second SELECT for eager_defaults,
+        # so only if it's set to True, not "auto"
+        if base_mapper.eager_defaults is True:
+            toload_now.extend(
+                state._unloaded_non_object.intersection(
+                    mapper._server_default_plus_onupdate_propkeys
+                )
+            )
+
+        if (
+            mapper.version_id_col is not None
+            and mapper.version_id_generator is False
+        ):
+            if mapper._version_id_prop.key in state.unloaded:
+                toload_now.extend([mapper._version_id_prop.key])
+
+        if toload_now:
+            state.key = base_mapper._identity_key_from_state(state)
+            stmt = future.select(mapper).set_label_style(
+                LABEL_STYLE_TABLENAME_PLUS_COL
+            )
+            loading.load_on_ident(
+                uowtransaction.session,
+                stmt,
+                state.key,
+                refresh_state=state,
+                only_load_props=toload_now,
+            )
+
+        # call after_XXX extensions
+        if not has_identity:
+            mapper.dispatch.after_insert(mapper, connection, state)
+        else:
+            mapper.dispatch.after_update(mapper, connection, state)
+
+        if (
+            mapper.version_id_generator is False
+            and mapper.version_id_col is not None
+        ):
+            if state_dict[mapper._version_id_prop.key] is None:
+                raise orm_exc.FlushError(
+                    "Instance does not contain a non-NULL version value"
+                )
+
+
+def _postfetch_post_update(
+    mapper, uowtransaction, table, state, dict_, result, params
+):
+    needs_version_id = (
+        mapper.version_id_col is not None
+        and mapper.version_id_col in mapper._cols_by_table[table]
+    )
+
+    if not uowtransaction.is_deleted(state):
+        # post updating after a regular INSERT or UPDATE, do a full postfetch
+        prefetch_cols = result.context.compiled.prefetch
+        postfetch_cols = result.context.compiled.postfetch
+    elif needs_version_id:
+        # post updating before a DELETE with a version_id_col, need to
+        # postfetch just version_id_col
+        prefetch_cols = postfetch_cols = ()
+    else:
+        # post updating before a DELETE without a version_id_col,
+        # don't need to postfetch
+        return
+
+    if needs_version_id:
+        prefetch_cols = list(prefetch_cols) + [mapper.version_id_col]
+
+    refresh_flush = bool(mapper.class_manager.dispatch.refresh_flush)
+    if refresh_flush:
+        load_evt_attrs = []
+
+    for c in prefetch_cols:
+        if c.key in params and c in mapper._columntoproperty:
+            dict_[mapper._columntoproperty[c].key] = params[c.key]
+            if refresh_flush:
+                load_evt_attrs.append(mapper._columntoproperty[c].key)
+
+    if refresh_flush and load_evt_attrs:
+        mapper.class_manager.dispatch.refresh_flush(
+            state, uowtransaction, load_evt_attrs
+        )
+
+    if postfetch_cols:
+        state._expire_attributes(
+            state.dict,
+            [
+                mapper._columntoproperty[c].key
+                for c in postfetch_cols
+                if c in mapper._columntoproperty
+            ],
+        )
+
+
+def _postfetch(
+    mapper,
+    uowtransaction,
+    table,
+    state,
+    dict_,
+    result,
+    params,
+    value_params,
+    isupdate,
+    returned_defaults,
+):
+    """Expire attributes in need of newly persisted database state,
+    after an INSERT or UPDATE statement has proceeded for that
+    state."""
+
+    prefetch_cols = result.context.compiled.prefetch
+    postfetch_cols = result.context.compiled.postfetch
+    returning_cols = result.context.compiled.effective_returning
+
+    if (
+        mapper.version_id_col is not None
+        and mapper.version_id_col in mapper._cols_by_table[table]
+    ):
+        prefetch_cols = list(prefetch_cols) + [mapper.version_id_col]
+
+    refresh_flush = bool(mapper.class_manager.dispatch.refresh_flush)
+    if refresh_flush:
+        load_evt_attrs = []
+
+    if returning_cols:
+        row = returned_defaults
+        if row is not None:
+            for row_value, col in zip(row, returning_cols):
+                # pk cols returned from insert are handled
+                # distinctly, don't step on the values here
+                if col.primary_key and result.context.isinsert:
+                    continue
+
+                # note that columns can be in the "return defaults" that are
+                # not mapped to this mapper, typically because they are
+                # "excluded", which can be specified directly or also occurs
+                # when using declarative w/ single table inheritance
+                prop = mapper._columntoproperty.get(col)
+                if prop:
+                    dict_[prop.key] = row_value
+                    if refresh_flush:
+                        load_evt_attrs.append(prop.key)
+
+    for c in prefetch_cols:
+        if c.key in params and c in mapper._columntoproperty:
+            pkey = mapper._columntoproperty[c].key
+
+            # set prefetched value in dict and also pop from committed_state,
+            # since this is new database state that replaces whatever might
+            # have previously been fetched (see #10800).  this is essentially a
+            # shorthand version of set_committed_value(), which could also be
+            # used here directly (with more overhead)
+            dict_[pkey] = params[c.key]
+            state.committed_state.pop(pkey, None)
+
+            if refresh_flush:
+                load_evt_attrs.append(pkey)
+
+    if refresh_flush and load_evt_attrs:
+        mapper.class_manager.dispatch.refresh_flush(
+            state, uowtransaction, load_evt_attrs
+        )
+
+    if isupdate and value_params:
+        # explicitly suit the use case specified by
+        # [ticket:3801], PK SQL expressions for UPDATE on non-RETURNING
+        # database which are set to themselves in order to do a version bump.
+        postfetch_cols.extend(
+            [
+                col
+                for col in value_params
+                if col.primary_key and col not in returning_cols
+            ]
+        )
+
+    if postfetch_cols:
+        state._expire_attributes(
+            state.dict,
+            [
+                mapper._columntoproperty[c].key
+                for c in postfetch_cols
+                if c in mapper._columntoproperty
+            ],
+        )
+
+    # synchronize newly inserted ids from one table to the next
+    # TODO: this still goes a little too often.  would be nice to
+    # have definitive list of "columns that changed" here
+    for m, equated_pairs in mapper._table_to_equated[table]:
+        sync.populate(
+            state,
+            m,
+            state,
+            m,
+            equated_pairs,
+            uowtransaction,
+            mapper.passive_updates,
+        )
+
+
+def _postfetch_bulk_save(mapper, dict_, table):
+    for m, equated_pairs in mapper._table_to_equated[table]:
+        sync.bulk_populate_inherit_keys(dict_, m, equated_pairs)
+
+
+def _connections_for_states(base_mapper, uowtransaction, states):
+    """Return an iterator of (state, state.dict, mapper, connection).
+
+    The states are sorted according to _sort_states, then paired
+    with the connection they should be using for the given
+    unit of work transaction.
+
+    """
+    # if session has a connection callable,
+    # organize individual states with the connection
+    # to use for update
+    if uowtransaction.session.connection_callable:
+        connection_callable = uowtransaction.session.connection_callable
+    else:
+        connection = uowtransaction.transaction.connection(base_mapper)
+        connection_callable = None
+
+    for state in _sort_states(base_mapper, states):
+        if connection_callable:
+            connection = connection_callable(base_mapper, state.obj())
+
+        mapper = state.manager.mapper
+
+        yield state, state.dict, mapper, connection
+
+
+def _sort_states(mapper, states):
+    pending = set(states)
+    persistent = {s for s in pending if s.key is not None}
+    pending.difference_update(persistent)
+
+    try:
+        persistent_sorted = sorted(
+            persistent, key=mapper._persistent_sortkey_fn
+        )
+    except TypeError as err:
+        raise sa_exc.InvalidRequestError(
+            "Could not sort objects by primary key; primary key "
+            "values must be sortable in Python (was: %s)" % err
+        ) from err
+    return (
+        sorted(pending, key=operator.attrgetter("insert_order"))
+        + persistent_sorted
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/properties.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/properties.py
new file mode 100644
index 00000000..a41c520c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/properties.py
@@ -0,0 +1,877 @@
+# orm/properties.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""MapperProperty implementations.
+
+This is a private module which defines the behavior of individual ORM-
+mapped attributes.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import strategy_options
+from .base import _DeclarativeMapped
+from .base import class_mapper
+from .descriptor_props import CompositeProperty
+from .descriptor_props import ConcreteInheritedProperty
+from .descriptor_props import SynonymProperty
+from .interfaces import _AttributeOptions
+from .interfaces import _DEFAULT_ATTRIBUTE_OPTIONS
+from .interfaces import _IntrospectsAnnotations
+from .interfaces import _MapsColumns
+from .interfaces import MapperProperty
+from .interfaces import PropComparator
+from .interfaces import StrategizedProperty
+from .relationships import RelationshipProperty
+from .util import de_stringify_annotation
+from .. import exc as sa_exc
+from .. import ForeignKey
+from .. import log
+from .. import util
+from ..sql import coercions
+from ..sql import roles
+from ..sql.base import _NoArg
+from ..sql.schema import Column
+from ..sql.schema import SchemaConst
+from ..sql.type_api import TypeEngine
+from ..util.typing import de_optionalize_union_types
+from ..util.typing import get_args
+from ..util.typing import includes_none
+from ..util.typing import is_fwd_ref
+from ..util.typing import is_pep593
+from ..util.typing import is_pep695
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import _ORMColumnExprArgument
+    from ._typing import _RegistryType
+    from .base import Mapped
+    from .decl_base import _ClassScanMapperConfig
+    from .mapper import Mapper
+    from .session import Session
+    from .state import _InstallLoaderCallableProto
+    from .state import InstanceState
+    from ..sql._typing import _InfoType
+    from ..sql.elements import ColumnElement
+    from ..sql.elements import NamedColumn
+    from ..sql.operators import OperatorType
+    from ..util.typing import _AnnotationScanType
+    from ..util.typing import RODescriptorReference
+
+_T = TypeVar("_T", bound=Any)
+_PT = TypeVar("_PT", bound=Any)
+_NC = TypeVar("_NC", bound="NamedColumn[Any]")
+
+__all__ = [
+    "ColumnProperty",
+    "CompositeProperty",
+    "ConcreteInheritedProperty",
+    "RelationshipProperty",
+    "SynonymProperty",
+]
+
+
+@log.class_logger
+class ColumnProperty(
+    _MapsColumns[_T],
+    StrategizedProperty[_T],
+    _IntrospectsAnnotations,
+    log.Identified,
+):
+    """Describes an object attribute that corresponds to a table column
+    or other column expression.
+
+    Public constructor is the :func:`_orm.column_property` function.
+
+    """
+
+    strategy_wildcard_key = strategy_options._COLUMN_TOKEN
+    inherit_cache = True
+    """:meta private:"""
+
+    _links_to_entity = False
+
+    columns: List[NamedColumn[Any]]
+
+    _is_polymorphic_discriminator: bool
+
+    _mapped_by_synonym: Optional[str]
+
+    comparator_factory: Type[PropComparator[_T]]
+
+    __slots__ = (
+        "columns",
+        "group",
+        "deferred",
+        "instrument",
+        "comparator_factory",
+        "active_history",
+        "expire_on_flush",
+        "_creation_order",
+        "_is_polymorphic_discriminator",
+        "_mapped_by_synonym",
+        "_deferred_column_loader",
+        "_raise_column_loader",
+        "_renders_in_subqueries",
+        "raiseload",
+    )
+
+    def __init__(
+        self,
+        column: _ORMColumnExprArgument[_T],
+        *additional_columns: _ORMColumnExprArgument[Any],
+        attribute_options: Optional[_AttributeOptions] = None,
+        group: Optional[str] = None,
+        deferred: bool = False,
+        raiseload: bool = False,
+        comparator_factory: Optional[Type[PropComparator[_T]]] = None,
+        active_history: bool = False,
+        expire_on_flush: bool = True,
+        info: Optional[_InfoType] = None,
+        doc: Optional[str] = None,
+        _instrument: bool = True,
+        _assume_readonly_dc_attributes: bool = False,
+    ):
+        super().__init__(
+            attribute_options=attribute_options,
+            _assume_readonly_dc_attributes=_assume_readonly_dc_attributes,
+        )
+        columns = (column,) + additional_columns
+        self.columns = [
+            coercions.expect(roles.LabeledColumnExprRole, c) for c in columns
+        ]
+        self.group = group
+        self.deferred = deferred
+        self.raiseload = raiseload
+        self.instrument = _instrument
+        self.comparator_factory = (
+            comparator_factory
+            if comparator_factory is not None
+            else self.__class__.Comparator
+        )
+        self.active_history = active_history
+        self.expire_on_flush = expire_on_flush
+
+        if info is not None:
+            self.info.update(info)
+
+        if doc is not None:
+            self.doc = doc
+        else:
+            for col in reversed(self.columns):
+                doc = getattr(col, "doc", None)
+                if doc is not None:
+                    self.doc = doc
+                    break
+            else:
+                self.doc = None
+
+        util.set_creation_order(self)
+
+        self.strategy_key = (
+            ("deferred", self.deferred),
+            ("instrument", self.instrument),
+        )
+        if self.raiseload:
+            self.strategy_key += (("raiseload", True),)
+
+    def declarative_scan(
+        self,
+        decl_scan: _ClassScanMapperConfig,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        mapped_container: Optional[Type[Mapped[Any]]],
+        annotation: Optional[_AnnotationScanType],
+        extracted_mapped_annotation: Optional[_AnnotationScanType],
+        is_dataclass_field: bool,
+    ) -> None:
+        column = self.columns[0]
+        if column.key is None:
+            column.key = key
+        if column.name is None:
+            column.name = key
+
+    @property
+    def mapper_property_to_assign(self) -> Optional[MapperProperty[_T]]:
+        return self
+
+    @property
+    def columns_to_assign(self) -> List[Tuple[Column[Any], int]]:
+        # mypy doesn't care about the isinstance here
+        return [
+            (c, 0)  # type: ignore
+            for c in self.columns
+            if isinstance(c, Column) and c.table is None
+        ]
+
+    def _memoized_attr__renders_in_subqueries(self) -> bool:
+        if ("query_expression", True) in self.strategy_key:
+            return self.strategy._have_default_expression  # type: ignore
+
+        return ("deferred", True) not in self.strategy_key or (
+            self not in self.parent._readonly_props  # type: ignore
+        )
+
+    @util.preload_module("sqlalchemy.orm.state", "sqlalchemy.orm.strategies")
+    def _memoized_attr__deferred_column_loader(
+        self,
+    ) -> _InstallLoaderCallableProto[Any]:
+        state = util.preloaded.orm_state
+        strategies = util.preloaded.orm_strategies
+        return state.InstanceState._instance_level_callable_processor(
+            self.parent.class_manager,
+            strategies.LoadDeferredColumns(self.key),
+            self.key,
+        )
+
+    @util.preload_module("sqlalchemy.orm.state", "sqlalchemy.orm.strategies")
+    def _memoized_attr__raise_column_loader(
+        self,
+    ) -> _InstallLoaderCallableProto[Any]:
+        state = util.preloaded.orm_state
+        strategies = util.preloaded.orm_strategies
+        return state.InstanceState._instance_level_callable_processor(
+            self.parent.class_manager,
+            strategies.LoadDeferredColumns(self.key, True),
+            self.key,
+        )
+
+    def __clause_element__(self) -> roles.ColumnsClauseRole:
+        """Allow the ColumnProperty to work in expression before it is turned
+        into an instrumented attribute.
+        """
+
+        return self.expression
+
+    @property
+    def expression(self) -> roles.ColumnsClauseRole:
+        """Return the primary column or expression for this ColumnProperty.
+
+        E.g.::
+
+
+            class File(Base):
+                # ...
+
+                name = Column(String(64))
+                extension = Column(String(8))
+                filename = column_property(name + "." + extension)
+                path = column_property("C:/" + filename.expression)
+
+        .. seealso::
+
+            :ref:`mapper_column_property_sql_expressions_composed`
+
+        """
+        return self.columns[0]
+
+    def instrument_class(self, mapper: Mapper[Any]) -> None:
+        if not self.instrument:
+            return
+
+        attributes.register_descriptor(
+            mapper.class_,
+            self.key,
+            comparator=self.comparator_factory(self, mapper),
+            parententity=mapper,
+            doc=self.doc,
+        )
+
+    def do_init(self) -> None:
+        super().do_init()
+
+        if len(self.columns) > 1 and set(self.parent.primary_key).issuperset(
+            self.columns
+        ):
+            util.warn(
+                (
+                    "On mapper %s, primary key column '%s' is being combined "
+                    "with distinct primary key column '%s' in attribute '%s'. "
+                    "Use explicit properties to give each column its own "
+                    "mapped attribute name."
+                )
+                % (self.parent, self.columns[1], self.columns[0], self.key)
+            )
+
+    def copy(self) -> ColumnProperty[_T]:
+        return ColumnProperty(
+            *self.columns,
+            deferred=self.deferred,
+            group=self.group,
+            active_history=self.active_history,
+        )
+
+    def merge(
+        self,
+        session: Session,
+        source_state: InstanceState[Any],
+        source_dict: _InstanceDict,
+        dest_state: InstanceState[Any],
+        dest_dict: _InstanceDict,
+        load: bool,
+        _recursive: Dict[Any, object],
+        _resolve_conflict_map: Dict[_IdentityKeyType[Any], object],
+    ) -> None:
+        if not self.instrument:
+            return
+        elif self.key in source_dict:
+            value = source_dict[self.key]
+
+            if not load:
+                dest_dict[self.key] = value
+            else:
+                impl = dest_state.get_impl(self.key)
+                impl.set(dest_state, dest_dict, value, None)
+        elif dest_state.has_identity and self.key not in dest_dict:
+            dest_state._expire_attributes(
+                dest_dict, [self.key], no_loader=True
+            )
+
+    class Comparator(util.MemoizedSlots, PropComparator[_PT]):
+        """Produce boolean, comparison, and other operators for
+        :class:`.ColumnProperty` attributes.
+
+        See the documentation for :class:`.PropComparator` for a brief
+        overview.
+
+        .. seealso::
+
+            :class:`.PropComparator`
+
+            :class:`.ColumnOperators`
+
+            :ref:`types_operators`
+
+            :attr:`.TypeEngine.comparator_factory`
+
+        """
+
+        if not TYPE_CHECKING:
+            # prevent pylance from being clever about slots
+            __slots__ = "__clause_element__", "info", "expressions"
+
+        prop: RODescriptorReference[ColumnProperty[_PT]]
+
+        expressions: Sequence[NamedColumn[Any]]
+        """The full sequence of columns referenced by this
+         attribute, adjusted for any aliasing in progress.
+
+        .. versionadded:: 1.3.17
+
+        .. seealso::
+
+           :ref:`maptojoin` - usage example
+        """
+
+        def _orm_annotate_column(self, column: _NC) -> _NC:
+            """annotate and possibly adapt a column to be returned
+            as the mapped-attribute exposed version of the column.
+
+            The column in this context needs to act as much like the
+            column in an ORM mapped context as possible, so includes
+            annotations to give hints to various ORM functions as to
+            the source entity of this column.   It also adapts it
+            to the mapper's with_polymorphic selectable if one is
+            present.
+
+            """
+
+            pe = self._parententity
+            annotations: Dict[str, Any] = {
+                "entity_namespace": pe,
+                "parententity": pe,
+                "parentmapper": pe,
+                "proxy_key": self.prop.key,
+            }
+
+            col = column
+
+            # for a mapper with polymorphic_on and an adapter, return
+            # the column against the polymorphic selectable.
+            # see also orm.util._orm_downgrade_polymorphic_columns
+            # for the reverse operation.
+            if self._parentmapper._polymorphic_adapter:
+                mapper_local_col = col
+                col = self._parentmapper._polymorphic_adapter.traverse(col)
+
+                # this is a clue to the ORM Query etc. that this column
+                # was adapted to the mapper's polymorphic_adapter.  the
+                # ORM uses this hint to know which column its adapting.
+                annotations["adapt_column"] = mapper_local_col
+
+            return col._annotate(annotations)._set_propagate_attrs(
+                {"compile_state_plugin": "orm", "plugin_subject": pe}
+            )
+
+        if TYPE_CHECKING:
+
+            def __clause_element__(self) -> NamedColumn[_PT]: ...
+
+        def _memoized_method___clause_element__(
+            self,
+        ) -> NamedColumn[_PT]:
+            if self.adapter:
+                return self.adapter(self.prop.columns[0], self.prop.key)
+            else:
+                return self._orm_annotate_column(self.prop.columns[0])
+
+        def _memoized_attr_info(self) -> _InfoType:
+            """The .info dictionary for this attribute."""
+
+            ce = self.__clause_element__()
+            try:
+                return ce.info  # type: ignore
+            except AttributeError:
+                return self.prop.info
+
+        def _memoized_attr_expressions(self) -> Sequence[NamedColumn[Any]]:
+            """The full sequence of columns referenced by this
+            attribute, adjusted for any aliasing in progress.
+
+            .. versionadded:: 1.3.17
+
+            """
+            if self.adapter:
+                return [
+                    self.adapter(col, self.prop.key)
+                    for col in self.prop.columns
+                ]
+            else:
+                return [
+                    self._orm_annotate_column(col) for col in self.prop.columns
+                ]
+
+        def _fallback_getattr(self, key: str) -> Any:
+            """proxy attribute access down to the mapped column.
+
+            this allows user-defined comparison methods to be accessed.
+            """
+            return getattr(self.__clause_element__(), key)
+
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]:
+            return op(self.__clause_element__(), *other, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]:
+            col = self.__clause_element__()
+            return op(col._bind_param(op, other), col, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def __str__(self) -> str:
+        if not self.parent or not self.key:
+            return object.__repr__(self)
+        return str(self.parent.class_.__name__) + "." + self.key
+
+
+class MappedSQLExpression(ColumnProperty[_T], _DeclarativeMapped[_T]):
+    """Declarative front-end for the :class:`.ColumnProperty` class.
+
+    Public constructor is the :func:`_orm.column_property` function.
+
+    .. versionchanged:: 2.0 Added :class:`_orm.MappedSQLExpression` as
+       a Declarative compatible subclass for :class:`_orm.ColumnProperty`.
+
+    .. seealso::
+
+        :class:`.MappedColumn`
+
+    """
+
+    inherit_cache = True
+    """:meta private:"""
+
+
+class MappedColumn(
+    _IntrospectsAnnotations,
+    _MapsColumns[_T],
+    _DeclarativeMapped[_T],
+):
+    """Maps a single :class:`_schema.Column` on a class.
+
+    :class:`_orm.MappedColumn` is a specialization of the
+    :class:`_orm.ColumnProperty` class and is oriented towards declarative
+    configuration.
+
+    To construct :class:`_orm.MappedColumn` objects, use the
+    :func:`_orm.mapped_column` constructor function.
+
+    .. versionadded:: 2.0
+
+
+    """
+
+    __slots__ = (
+        "column",
+        "_creation_order",
+        "_sort_order",
+        "foreign_keys",
+        "_has_nullable",
+        "_has_insert_default",
+        "deferred",
+        "deferred_group",
+        "deferred_raiseload",
+        "active_history",
+        "_attribute_options",
+        "_has_dataclass_arguments",
+        "_use_existing_column",
+    )
+
+    deferred: Union[_NoArg, bool]
+    deferred_raiseload: bool
+    deferred_group: Optional[str]
+
+    column: Column[_T]
+    foreign_keys: Optional[Set[ForeignKey]]
+    _attribute_options: _AttributeOptions
+
+    def __init__(self, *arg: Any, **kw: Any):
+        self._attribute_options = attr_opts = kw.pop(
+            "attribute_options", _DEFAULT_ATTRIBUTE_OPTIONS
+        )
+
+        self._use_existing_column = kw.pop("use_existing_column", False)
+
+        self._has_dataclass_arguments = (
+            attr_opts is not None
+            and attr_opts != _DEFAULT_ATTRIBUTE_OPTIONS
+            and any(
+                attr_opts[i] is not _NoArg.NO_ARG
+                for i, attr in enumerate(attr_opts._fields)
+                if attr != "dataclasses_default"
+            )
+        )
+
+        insert_default = kw.pop("insert_default", _NoArg.NO_ARG)
+        self._has_insert_default = insert_default is not _NoArg.NO_ARG
+
+        if self._has_insert_default:
+            kw["default"] = insert_default
+        elif attr_opts.dataclasses_default is not _NoArg.NO_ARG:
+            kw["default"] = attr_opts.dataclasses_default
+
+        self.deferred_group = kw.pop("deferred_group", None)
+        self.deferred_raiseload = kw.pop("deferred_raiseload", None)
+        self.deferred = kw.pop("deferred", _NoArg.NO_ARG)
+        self.active_history = kw.pop("active_history", False)
+
+        self._sort_order = kw.pop("sort_order", _NoArg.NO_ARG)
+        self.column = cast("Column[_T]", Column(*arg, **kw))
+        self.foreign_keys = self.column.foreign_keys
+        self._has_nullable = "nullable" in kw and kw.get("nullable") not in (
+            None,
+            SchemaConst.NULL_UNSPECIFIED,
+        )
+        util.set_creation_order(self)
+
+    def _copy(self, **kw: Any) -> Self:
+        new = self.__class__.__new__(self.__class__)
+        new.column = self.column._copy(**kw)
+        new.deferred = self.deferred
+        new.deferred_group = self.deferred_group
+        new.deferred_raiseload = self.deferred_raiseload
+        new.foreign_keys = new.column.foreign_keys
+        new.active_history = self.active_history
+        new._has_nullable = self._has_nullable
+        new._attribute_options = self._attribute_options
+        new._has_insert_default = self._has_insert_default
+        new._has_dataclass_arguments = self._has_dataclass_arguments
+        new._use_existing_column = self._use_existing_column
+        new._sort_order = self._sort_order
+        util.set_creation_order(new)
+        return new
+
+    @property
+    def name(self) -> str:
+        return self.column.name
+
+    @property
+    def mapper_property_to_assign(self) -> Optional[MapperProperty[_T]]:
+        effective_deferred = self.deferred
+        if effective_deferred is _NoArg.NO_ARG:
+            effective_deferred = bool(
+                self.deferred_group or self.deferred_raiseload
+            )
+
+        if effective_deferred or self.active_history:
+            return ColumnProperty(
+                self.column,
+                deferred=effective_deferred,
+                group=self.deferred_group,
+                raiseload=self.deferred_raiseload,
+                attribute_options=self._attribute_options,
+                active_history=self.active_history,
+            )
+        else:
+            return None
+
+    @property
+    def columns_to_assign(self) -> List[Tuple[Column[Any], int]]:
+        return [
+            (
+                self.column,
+                (
+                    self._sort_order
+                    if self._sort_order is not _NoArg.NO_ARG
+                    else 0
+                ),
+            )
+        ]
+
+    def __clause_element__(self) -> Column[_T]:
+        return self.column
+
+    def operate(
+        self, op: OperatorType, *other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(self.__clause_element__(), *other, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def reverse_operate(
+        self, op: OperatorType, other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        col = self.__clause_element__()
+        return op(col._bind_param(op, other), col, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def found_in_pep593_annotated(self) -> Any:
+        # return a blank mapped_column().  This mapped_column()'s
+        # Column will be merged into it in _init_column_for_annotation().
+        return MappedColumn()
+
+    def declarative_scan(
+        self,
+        decl_scan: _ClassScanMapperConfig,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        mapped_container: Optional[Type[Mapped[Any]]],
+        annotation: Optional[_AnnotationScanType],
+        extracted_mapped_annotation: Optional[_AnnotationScanType],
+        is_dataclass_field: bool,
+    ) -> None:
+        column = self.column
+
+        if (
+            self._use_existing_column
+            and decl_scan.inherits
+            and decl_scan.single
+        ):
+            if decl_scan.is_deferred:
+                raise sa_exc.ArgumentError(
+                    "Can't use use_existing_column with deferred mappers"
+                )
+            supercls_mapper = class_mapper(decl_scan.inherits, False)
+
+            colname = column.name if column.name is not None else key
+            column = self.column = supercls_mapper.local_table.c.get(  # type: ignore[assignment] # noqa: E501
+                colname, column
+            )
+
+        if column.key is None:
+            column.key = key
+        if column.name is None:
+            column.name = key
+
+        sqltype = column.type
+
+        if extracted_mapped_annotation is None:
+            if sqltype._isnull and not self.column.foreign_keys:
+                self._raise_for_required(key, cls)
+            else:
+                return
+
+        self._init_column_for_annotation(
+            cls,
+            registry,
+            extracted_mapped_annotation,
+            originating_module,
+        )
+
+    @util.preload_module("sqlalchemy.orm.decl_base")
+    def declarative_scan_for_composite(
+        self,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        param_name: str,
+        param_annotation: _AnnotationScanType,
+    ) -> None:
+        decl_base = util.preloaded.orm_decl_base
+        decl_base._undefer_column_name(param_name, self.column)
+        self._init_column_for_annotation(
+            cls, registry, param_annotation, originating_module
+        )
+
+    def _init_column_for_annotation(
+        self,
+        cls: Type[Any],
+        registry: _RegistryType,
+        argument: _AnnotationScanType,
+        originating_module: Optional[str],
+    ) -> None:
+        sqltype = self.column.type
+
+        if is_fwd_ref(
+            argument, check_generic=True, check_for_plain_string=True
+        ):
+            assert originating_module is not None
+            argument = de_stringify_annotation(
+                cls, argument, originating_module, include_generic=True
+            )
+
+        nullable = includes_none(argument)
+
+        if not self._has_nullable:
+            self.column.nullable = nullable
+
+        our_type = de_optionalize_union_types(argument)
+
+        find_mapped_in: Tuple[Any, ...] = ()
+        our_type_is_pep593 = False
+        raw_pep_593_type = None
+
+        if is_pep593(our_type):
+            our_type_is_pep593 = True
+
+            pep_593_components = get_args(our_type)
+            raw_pep_593_type = pep_593_components[0]
+            if nullable:
+                raw_pep_593_type = de_optionalize_union_types(raw_pep_593_type)
+            find_mapped_in = pep_593_components[1:]
+        elif is_pep695(argument) and is_pep593(argument.__value__):
+            # do not support nested annotation inside unions ets
+            find_mapped_in = get_args(argument.__value__)[1:]
+
+        use_args_from: Optional[MappedColumn[Any]]
+        for elem in find_mapped_in:
+            if isinstance(elem, MappedColumn):
+                use_args_from = elem
+                break
+        else:
+            use_args_from = None
+
+        if use_args_from is not None:
+            if (
+                not self._has_insert_default
+                and use_args_from.column.default is not None
+            ):
+                self.column.default = None
+
+            use_args_from.column._merge(self.column)
+            sqltype = self.column.type
+
+            if (
+                use_args_from.deferred is not _NoArg.NO_ARG
+                and self.deferred is _NoArg.NO_ARG
+            ):
+                self.deferred = use_args_from.deferred
+
+            if (
+                use_args_from.deferred_group is not None
+                and self.deferred_group is None
+            ):
+                self.deferred_group = use_args_from.deferred_group
+
+            if (
+                use_args_from.deferred_raiseload is not None
+                and self.deferred_raiseload is None
+            ):
+                self.deferred_raiseload = use_args_from.deferred_raiseload
+
+            if (
+                use_args_from._use_existing_column
+                and not self._use_existing_column
+            ):
+                self._use_existing_column = True
+
+            if use_args_from.active_history:
+                self.active_history = use_args_from.active_history
+
+            if (
+                use_args_from._sort_order is not None
+                and self._sort_order is _NoArg.NO_ARG
+            ):
+                self._sort_order = use_args_from._sort_order
+
+            if (
+                use_args_from.column.key is not None
+                or use_args_from.column.name is not None
+            ):
+                util.warn_deprecated(
+                    "Can't use the 'key' or 'name' arguments in "
+                    "Annotated with mapped_column(); this will be ignored",
+                    "2.0.22",
+                )
+
+            if use_args_from._has_dataclass_arguments:
+                for idx, arg in enumerate(
+                    use_args_from._attribute_options._fields
+                ):
+                    if (
+                        use_args_from._attribute_options[idx]
+                        is not _NoArg.NO_ARG
+                    ):
+                        arg = arg.replace("dataclasses_", "")
+                        util.warn_deprecated(
+                            f"Argument '{arg}' is a dataclass argument and "
+                            "cannot be specified within a mapped_column() "
+                            "bundled inside of an Annotated object",
+                            "2.0.22",
+                        )
+
+        if sqltype._isnull and not self.column.foreign_keys:
+            new_sqltype = None
+
+            checks: List[Any]
+            if our_type_is_pep593:
+                checks = [our_type, raw_pep_593_type]
+            else:
+                checks = [our_type]
+
+            for check_type in checks:
+                new_sqltype = registry._resolve_type(check_type)
+                if new_sqltype is not None:
+                    break
+            else:
+                if isinstance(our_type, TypeEngine) or (
+                    isinstance(our_type, type)
+                    and issubclass(our_type, TypeEngine)
+                ):
+                    raise sa_exc.ArgumentError(
+                        f"The type provided inside the {self.column.key!r} "
+                        "attribute Mapped annotation is the SQLAlchemy type "
+                        f"{our_type}. Expected a Python type instead"
+                    )
+                else:
+                    raise sa_exc.ArgumentError(
+                        "Could not locate SQLAlchemy Core type for Python "
+                        f"type {our_type} inside the {self.column.key!r} "
+                        "attribute Mapped annotation"
+                    )
+
+            self.column._set_type(new_sqltype)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/query.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/query.py
new file mode 100644
index 00000000..c7e1ca8a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/query.py
@@ -0,0 +1,3447 @@
+# orm/query.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""The Query class and support.
+
+Defines the :class:`_query.Query` class, the central
+construct used by the ORM to construct database queries.
+
+The :class:`_query.Query` class should not be confused with the
+:class:`_expression.Select` class, which defines database
+SELECT operations at the SQL (non-ORM) level.  ``Query`` differs from
+``Select`` in that it returns ORM-mapped objects and interacts with an
+ORM session, whereas the ``Select`` construct interacts directly with the
+database to return iterable result sets.
+
+"""
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import operator
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import attributes
+from . import interfaces
+from . import loading
+from . import util as orm_util
+from ._typing import _O
+from .base import _assertions
+from .context import _column_descriptions
+from .context import _determine_last_joined_entity
+from .context import _legacy_filter_by_entity_zero
+from .context import FromStatement
+from .context import ORMCompileState
+from .context import QueryContext
+from .interfaces import ORMColumnDescription
+from .interfaces import ORMColumnsClauseRole
+from .util import AliasedClass
+from .util import object_mapper
+from .util import with_parent
+from .. import exc as sa_exc
+from .. import inspect
+from .. import inspection
+from .. import log
+from .. import sql
+from .. import util
+from ..engine import Result
+from ..engine import Row
+from ..event import dispatcher
+from ..event import EventTarget
+from ..sql import coercions
+from ..sql import expression
+from ..sql import roles
+from ..sql import Select
+from ..sql import util as sql_util
+from ..sql import visitors
+from ..sql._typing import _FromClauseArgument
+from ..sql._typing import _TP
+from ..sql.annotation import SupportsCloneAnnotations
+from ..sql.base import _entity_namespace_key
+from ..sql.base import _generative
+from ..sql.base import _NoArg
+from ..sql.base import Executable
+from ..sql.base import Generative
+from ..sql.elements import BooleanClauseList
+from ..sql.expression import Exists
+from ..sql.selectable import _MemoizedSelectEntities
+from ..sql.selectable import _SelectFromElements
+from ..sql.selectable import ForUpdateArg
+from ..sql.selectable import HasHints
+from ..sql.selectable import HasPrefixes
+from ..sql.selectable import HasSuffixes
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..sql.selectable import SelectLabelStyle
+from ..util.typing import Literal
+from ..util.typing import Self
+
+
+if TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _ExternalEntityType
+    from ._typing import _InternalEntityType
+    from ._typing import SynchronizeSessionArgument
+    from .mapper import Mapper
+    from .path_registry import PathRegistry
+    from .session import _PKIdentityArgument
+    from .session import Session
+    from .state import InstanceState
+    from ..engine.cursor import CursorResult
+    from ..engine.interfaces import _ImmutableExecuteOptions
+    from ..engine.interfaces import CompiledCacheType
+    from ..engine.interfaces import IsolationLevel
+    from ..engine.interfaces import SchemaTranslateMapType
+    from ..engine.result import FrozenResult
+    from ..engine.result import ScalarResult
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _ColumnExpressionOrStrLabelArgument
+    from ..sql._typing import _ColumnsClauseArgument
+    from ..sql._typing import _DMLColumnArgument
+    from ..sql._typing import _JoinTargetArgument
+    from ..sql._typing import _LimitOffsetType
+    from ..sql._typing import _MAYBE_ENTITY
+    from ..sql._typing import _no_kw
+    from ..sql._typing import _NOT_ENTITY
+    from ..sql._typing import _OnClauseArgument
+    from ..sql._typing import _PropagateAttrsType
+    from ..sql._typing import _T0
+    from ..sql._typing import _T1
+    from ..sql._typing import _T2
+    from ..sql._typing import _T3
+    from ..sql._typing import _T4
+    from ..sql._typing import _T5
+    from ..sql._typing import _T6
+    from ..sql._typing import _T7
+    from ..sql._typing import _TypedColumnClauseArgument as _TCCA
+    from ..sql.base import CacheableOptions
+    from ..sql.base import ExecutableOption
+    from ..sql.elements import ColumnElement
+    from ..sql.elements import Label
+    from ..sql.selectable import _ForUpdateOfArgument
+    from ..sql.selectable import _JoinTargetElement
+    from ..sql.selectable import _SetupJoinsElement
+    from ..sql.selectable import Alias
+    from ..sql.selectable import CTE
+    from ..sql.selectable import ExecutableReturnsRows
+    from ..sql.selectable import FromClause
+    from ..sql.selectable import ScalarSelect
+    from ..sql.selectable import Subquery
+
+
+__all__ = ["Query", "QueryContext"]
+
+_T = TypeVar("_T", bound=Any)
+
+
+@inspection._self_inspects
+@log.class_logger
+class Query(
+    _SelectFromElements,
+    SupportsCloneAnnotations,
+    HasPrefixes,
+    HasSuffixes,
+    HasHints,
+    EventTarget,
+    log.Identified,
+    Generative,
+    Executable,
+    Generic[_T],
+):
+    """ORM-level SQL construction object.
+
+    .. legacy:: The ORM :class:`.Query` object is a legacy construct
+       as of SQLAlchemy 2.0.   See the notes at the top of
+       :ref:`query_api_toplevel` for an overview, including links to migration
+       documentation.
+
+    :class:`_query.Query` objects are normally initially generated using the
+    :meth:`~.Session.query` method of :class:`.Session`, and in
+    less common cases by instantiating the :class:`_query.Query` directly and
+    associating with a :class:`.Session` using the
+    :meth:`_query.Query.with_session`
+    method.
+
+    """
+
+    # elements that are in Core and can be cached in the same way
+    _where_criteria: Tuple[ColumnElement[Any], ...] = ()
+    _having_criteria: Tuple[ColumnElement[Any], ...] = ()
+
+    _order_by_clauses: Tuple[ColumnElement[Any], ...] = ()
+    _group_by_clauses: Tuple[ColumnElement[Any], ...] = ()
+    _limit_clause: Optional[ColumnElement[Any]] = None
+    _offset_clause: Optional[ColumnElement[Any]] = None
+
+    _distinct: bool = False
+    _distinct_on: Tuple[ColumnElement[Any], ...] = ()
+
+    _for_update_arg: Optional[ForUpdateArg] = None
+    _correlate: Tuple[FromClause, ...] = ()
+    _auto_correlate: bool = True
+    _from_obj: Tuple[FromClause, ...] = ()
+    _setup_joins: Tuple[_SetupJoinsElement, ...] = ()
+
+    _label_style: SelectLabelStyle = SelectLabelStyle.LABEL_STYLE_LEGACY_ORM
+
+    _memoized_select_entities = ()
+
+    _compile_options: Union[Type[CacheableOptions], CacheableOptions] = (
+        ORMCompileState.default_compile_options
+    )
+
+    _with_options: Tuple[ExecutableOption, ...]
+    load_options = QueryContext.default_load_options + {
+        "_legacy_uniquing": True
+    }
+
+    _params: util.immutabledict[str, Any] = util.EMPTY_DICT
+
+    # local Query builder state, not needed for
+    # compilation or execution
+    _enable_assertions = True
+
+    _statement: Optional[ExecutableReturnsRows] = None
+
+    session: Session
+
+    dispatch: dispatcher[Query[_T]]
+
+    # mirrors that of ClauseElement, used to propagate the "orm"
+    # plugin as well as the "subject" of the plugin, e.g. the mapper
+    # we are querying against.
+    @util.memoized_property
+    def _propagate_attrs(self) -> _PropagateAttrsType:
+        return util.EMPTY_DICT
+
+    def __init__(
+        self,
+        entities: Union[
+            _ColumnsClauseArgument[Any], Sequence[_ColumnsClauseArgument[Any]]
+        ],
+        session: Optional[Session] = None,
+    ):
+        """Construct a :class:`_query.Query` directly.
+
+        E.g.::
+
+            q = Query([User, Address], session=some_session)
+
+        The above is equivalent to::
+
+            q = some_session.query(User, Address)
+
+        :param entities: a sequence of entities and/or SQL expressions.
+
+        :param session: a :class:`.Session` with which the
+         :class:`_query.Query`
+         will be associated.   Optional; a :class:`_query.Query`
+         can be associated
+         with a :class:`.Session` generatively via the
+         :meth:`_query.Query.with_session` method as well.
+
+        .. seealso::
+
+            :meth:`.Session.query`
+
+            :meth:`_query.Query.with_session`
+
+        """
+
+        # session is usually present.  There's one case in subqueryloader
+        # where it stores a Query without a Session and also there are tests
+        # for the query(Entity).with_session(session) API which is likely in
+        # some old recipes, however these are legacy as select() can now be
+        # used.
+        self.session = session  # type: ignore
+        self._set_entities(entities)
+
+    def _set_propagate_attrs(self, values: Mapping[str, Any]) -> Self:
+        self._propagate_attrs = util.immutabledict(values)
+        return self
+
+    def _set_entities(
+        self,
+        entities: Union[
+            _ColumnsClauseArgument[Any], Iterable[_ColumnsClauseArgument[Any]]
+        ],
+    ) -> None:
+        self._raw_columns = [
+            coercions.expect(
+                roles.ColumnsClauseRole,
+                ent,
+                apply_propagate_attrs=self,
+                post_inspect=True,
+            )
+            for ent in util.to_list(entities)
+        ]
+
+    def tuples(self: Query[_O]) -> Query[Tuple[_O]]:
+        """return a tuple-typed form of this :class:`.Query`.
+
+        This method invokes the :meth:`.Query.only_return_tuples`
+        method with a value of ``True``, which by itself ensures that this
+        :class:`.Query` will always return :class:`.Row` objects, even
+        if the query is made against a single entity.  It then also
+        at the typing level will return a "typed" query, if possible,
+        that will type result rows as ``Tuple`` objects with typed
+        elements.
+
+        This method can be compared to the :meth:`.Result.tuples` method,
+        which returns "self", but from a typing perspective returns an object
+        that will yield typed ``Tuple`` objects for results.   Typing
+        takes effect only if this :class:`.Query` object is a typed
+        query object already.
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :meth:`.Result.tuples` - v2 equivalent method.
+
+        """
+        return self.only_return_tuples(True)  # type: ignore
+
+    def _entity_from_pre_ent_zero(self) -> Optional[_InternalEntityType[Any]]:
+        if not self._raw_columns:
+            return None
+
+        ent = self._raw_columns[0]
+
+        if "parententity" in ent._annotations:
+            return ent._annotations["parententity"]  # type: ignore
+        elif "bundle" in ent._annotations:
+            return ent._annotations["bundle"]  # type: ignore
+        else:
+            # label, other SQL expression
+            for element in visitors.iterate(ent):
+                if "parententity" in element._annotations:
+                    return element._annotations["parententity"]  # type: ignore  # noqa: E501
+            else:
+                return None
+
+    def _only_full_mapper_zero(self, methname: str) -> Mapper[Any]:
+        if (
+            len(self._raw_columns) != 1
+            or "parententity" not in self._raw_columns[0]._annotations
+            or not self._raw_columns[0].is_selectable
+        ):
+            raise sa_exc.InvalidRequestError(
+                "%s() can only be used against "
+                "a single mapped class." % methname
+            )
+
+        return self._raw_columns[0]._annotations["parententity"]  # type: ignore  # noqa: E501
+
+    def _set_select_from(
+        self, obj: Iterable[_FromClauseArgument], set_base_alias: bool
+    ) -> None:
+        fa = [
+            coercions.expect(
+                roles.StrictFromClauseRole,
+                elem,
+                allow_select=True,
+                apply_propagate_attrs=self,
+            )
+            for elem in obj
+        ]
+
+        self._compile_options += {"_set_base_alias": set_base_alias}
+        self._from_obj = tuple(fa)
+
+    @_generative
+    def _set_lazyload_from(self, state: InstanceState[Any]) -> Self:
+        self.load_options += {"_lazy_loaded_from": state}
+        return self
+
+    def _get_condition(self) -> None:
+        """used by legacy BakedQuery"""
+        self._no_criterion_condition("get", order_by=False, distinct=False)
+
+    def _get_existing_condition(self) -> None:
+        self._no_criterion_assertion("get", order_by=False, distinct=False)
+
+    def _no_criterion_assertion(
+        self, meth: str, order_by: bool = True, distinct: bool = True
+    ) -> None:
+        if not self._enable_assertions:
+            return
+        if (
+            self._where_criteria
+            or self._statement is not None
+            or self._from_obj
+            or self._setup_joins
+            or self._limit_clause is not None
+            or self._offset_clause is not None
+            or self._group_by_clauses
+            or (order_by and self._order_by_clauses)
+            or (distinct and self._distinct)
+        ):
+            raise sa_exc.InvalidRequestError(
+                "Query.%s() being called on a "
+                "Query with existing criterion. " % meth
+            )
+
+    def _no_criterion_condition(
+        self, meth: str, order_by: bool = True, distinct: bool = True
+    ) -> None:
+        self._no_criterion_assertion(meth, order_by, distinct)
+
+        self._from_obj = self._setup_joins = ()
+        if self._statement is not None:
+            self._compile_options += {"_statement": None}
+        self._where_criteria = ()
+        self._distinct = False
+
+        self._order_by_clauses = self._group_by_clauses = ()
+
+    def _no_clauseelement_condition(self, meth: str) -> None:
+        if not self._enable_assertions:
+            return
+        if self._order_by_clauses:
+            raise sa_exc.InvalidRequestError(
+                "Query.%s() being called on a "
+                "Query with existing criterion. " % meth
+            )
+        self._no_criterion_condition(meth)
+
+    def _no_statement_condition(self, meth: str) -> None:
+        if not self._enable_assertions:
+            return
+        if self._statement is not None:
+            raise sa_exc.InvalidRequestError(
+                (
+                    "Query.%s() being called on a Query with an existing full "
+                    "statement - can't apply criterion."
+                )
+                % meth
+            )
+
+    def _no_limit_offset(self, meth: str) -> None:
+        if not self._enable_assertions:
+            return
+        if self._limit_clause is not None or self._offset_clause is not None:
+            raise sa_exc.InvalidRequestError(
+                "Query.%s() being called on a Query which already has LIMIT "
+                "or OFFSET applied.  Call %s() before limit() or offset() "
+                "are applied." % (meth, meth)
+            )
+
+    @property
+    def _has_row_limiting_clause(self) -> bool:
+        return (
+            self._limit_clause is not None or self._offset_clause is not None
+        )
+
+    def _get_options(
+        self,
+        populate_existing: Optional[bool] = None,
+        version_check: Optional[bool] = None,
+        only_load_props: Optional[Sequence[str]] = None,
+        refresh_state: Optional[InstanceState[Any]] = None,
+        identity_token: Optional[Any] = None,
+    ) -> Self:
+        load_options: Dict[str, Any] = {}
+        compile_options: Dict[str, Any] = {}
+
+        if version_check:
+            load_options["_version_check"] = version_check
+        if populate_existing:
+            load_options["_populate_existing"] = populate_existing
+        if refresh_state:
+            load_options["_refresh_state"] = refresh_state
+            compile_options["_for_refresh_state"] = True
+        if only_load_props:
+            compile_options["_only_load_props"] = frozenset(only_load_props)
+        if identity_token:
+            load_options["_identity_token"] = identity_token
+
+        if load_options:
+            self.load_options += load_options
+        if compile_options:
+            self._compile_options += compile_options
+
+        return self
+
+    def _clone(self, **kw: Any) -> Self:
+        return self._generate()
+
+    def _get_select_statement_only(self) -> Select[_T]:
+        if self._statement is not None:
+            raise sa_exc.InvalidRequestError(
+                "Can't call this method on a Query that uses from_statement()"
+            )
+        return cast("Select[_T]", self.statement)
+
+    @property
+    def statement(self) -> Union[Select[_T], FromStatement[_T]]:
+        """The full SELECT statement represented by this Query.
+
+        The statement by default will not have disambiguating labels
+        applied to the construct unless with_labels(True) is called
+        first.
+
+        """
+
+        # .statement can return the direct future.Select() construct here, as
+        # long as we are not using subsequent adaption features that
+        # are made against raw entities, e.g. from_self(), with_polymorphic(),
+        # select_entity_from().  If these features are being used, then
+        # the Select() we return will not have the correct .selected_columns
+        # collection and will not embed in subsequent queries correctly.
+        # We could find a way to make this collection "correct", however
+        # this would not be too different from doing the full compile as
+        # we are doing in any case, the Select() would still not have the
+        # proper state for other attributes like whereclause, order_by,
+        # and these features are all deprecated in any case.
+        #
+        # for these reasons, Query is not a Select, it remains an ORM
+        # object for which __clause_element__() must be called in order for
+        # it to provide a real expression object.
+        #
+        # from there, it starts to look much like Query itself won't be
+        # passed into the execute process and won't generate its own cache
+        # key; this will all occur in terms of the ORM-enabled Select.
+        if not self._compile_options._set_base_alias:
+            # if we don't have legacy top level aliasing features in use
+            # then convert to a future select() directly
+            stmt = self._statement_20(for_statement=True)
+        else:
+            stmt = self._compile_state(for_statement=True).statement
+
+        if self._params:
+            stmt = stmt.params(self._params)
+
+        return stmt
+
+    def _final_statement(self, legacy_query_style: bool = True) -> Select[Any]:
+        """Return the 'final' SELECT statement for this :class:`.Query`.
+
+        This is used by the testing suite only and is fairly inefficient.
+
+        This is the Core-only select() that will be rendered by a complete
+        compilation of this query, and is what .statement used to return
+        in 1.3.
+
+
+        """
+
+        q = self._clone()
+
+        return q._compile_state(
+            use_legacy_query_style=legacy_query_style
+        ).statement  # type: ignore
+
+    def _statement_20(
+        self, for_statement: bool = False, use_legacy_query_style: bool = True
+    ) -> Union[Select[_T], FromStatement[_T]]:
+        # TODO: this event needs to be deprecated, as it currently applies
+        # only to ORM query and occurs at this spot that is now more
+        # or less an artificial spot
+        if self.dispatch.before_compile:
+            for fn in self.dispatch.before_compile:
+                new_query = fn(self)
+                if new_query is not None and new_query is not self:
+                    self = new_query
+                    if not fn._bake_ok:  # type: ignore
+                        self._compile_options += {"_bake_ok": False}
+
+        compile_options = self._compile_options
+        compile_options += {
+            "_for_statement": for_statement,
+            "_use_legacy_query_style": use_legacy_query_style,
+        }
+
+        stmt: Union[Select[_T], FromStatement[_T]]
+
+        if self._statement is not None:
+            stmt = FromStatement(self._raw_columns, self._statement)
+            stmt.__dict__.update(
+                _with_options=self._with_options,
+                _with_context_options=self._with_context_options,
+                _compile_options=compile_options,
+                _execution_options=self._execution_options,
+                _propagate_attrs=self._propagate_attrs,
+            )
+        else:
+            # Query / select() internal attributes are 99% cross-compatible
+            stmt = Select._create_raw_select(**self.__dict__)
+            stmt.__dict__.update(
+                _label_style=self._label_style,
+                _compile_options=compile_options,
+                _propagate_attrs=self._propagate_attrs,
+            )
+            stmt.__dict__.pop("session", None)
+
+        # ensure the ORM context is used to compile the statement, even
+        # if it has no ORM entities.  This is so ORM-only things like
+        # _legacy_joins are picked up that wouldn't be picked up by the
+        # Core statement context
+        if "compile_state_plugin" not in stmt._propagate_attrs:
+            stmt._propagate_attrs = stmt._propagate_attrs.union(
+                {"compile_state_plugin": "orm", "plugin_subject": None}
+            )
+
+        return stmt
+
+    def subquery(
+        self,
+        name: Optional[str] = None,
+        with_labels: bool = False,
+        reduce_columns: bool = False,
+    ) -> Subquery:
+        """Return the full SELECT statement represented by
+        this :class:`_query.Query`, embedded within an
+        :class:`_expression.Alias`.
+
+        Eager JOIN generation within the query is disabled.
+
+        .. seealso::
+
+            :meth:`_sql.Select.subquery` - v2 comparable method.
+
+        :param name: string name to be assigned as the alias;
+            this is passed through to :meth:`_expression.FromClause.alias`.
+            If ``None``, a name will be deterministically generated
+            at compile time.
+
+        :param with_labels: if True, :meth:`.with_labels` will be called
+         on the :class:`_query.Query` first to apply table-qualified labels
+         to all columns.
+
+        :param reduce_columns: if True,
+         :meth:`_expression.Select.reduce_columns` will
+         be called on the resulting :func:`_expression.select` construct,
+         to remove same-named columns where one also refers to the other
+         via foreign key or WHERE clause equivalence.
+
+        """
+        q = self.enable_eagerloads(False)
+        if with_labels:
+            q = q.set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+
+        stmt = q._get_select_statement_only()
+
+        if TYPE_CHECKING:
+            assert isinstance(stmt, Select)
+
+        if reduce_columns:
+            stmt = stmt.reduce_columns()
+        return stmt.subquery(name=name)
+
+    def cte(
+        self,
+        name: Optional[str] = None,
+        recursive: bool = False,
+        nesting: bool = False,
+    ) -> CTE:
+        r"""Return the full SELECT statement represented by this
+        :class:`_query.Query` represented as a common table expression (CTE).
+
+        Parameters and usage are the same as those of the
+        :meth:`_expression.SelectBase.cte` method; see that method for
+        further details.
+
+        Here is the `PostgreSQL WITH
+        RECURSIVE example
+        <https://www.postgresql.org/docs/current/static/queries-with.html>`_.
+        Note that, in this example, the ``included_parts`` cte and the
+        ``incl_alias`` alias of it are Core selectables, which
+        means the columns are accessed via the ``.c.`` attribute.  The
+        ``parts_alias`` object is an :func:`_orm.aliased` instance of the
+        ``Part`` entity, so column-mapped attributes are available
+        directly::
+
+            from sqlalchemy.orm import aliased
+
+
+            class Part(Base):
+                __tablename__ = "part"
+                part = Column(String, primary_key=True)
+                sub_part = Column(String, primary_key=True)
+                quantity = Column(Integer)
+
+
+            included_parts = (
+                session.query(Part.sub_part, Part.part, Part.quantity)
+                .filter(Part.part == "our part")
+                .cte(name="included_parts", recursive=True)
+            )
+
+            incl_alias = aliased(included_parts, name="pr")
+            parts_alias = aliased(Part, name="p")
+            included_parts = included_parts.union_all(
+                session.query(
+                    parts_alias.sub_part, parts_alias.part, parts_alias.quantity
+                ).filter(parts_alias.part == incl_alias.c.sub_part)
+            )
+
+            q = session.query(
+                included_parts.c.sub_part,
+                func.sum(included_parts.c.quantity).label("total_quantity"),
+            ).group_by(included_parts.c.sub_part)
+
+        .. seealso::
+
+            :meth:`_sql.Select.cte` - v2 equivalent method.
+
+        """  # noqa: E501
+        return (
+            self.enable_eagerloads(False)
+            ._get_select_statement_only()
+            .cte(name=name, recursive=recursive, nesting=nesting)
+        )
+
+    def label(self, name: Optional[str]) -> Label[Any]:
+        """Return the full SELECT statement represented by this
+        :class:`_query.Query`, converted
+        to a scalar subquery with a label of the given name.
+
+        .. seealso::
+
+            :meth:`_sql.Select.label` - v2 comparable method.
+
+        """
+
+        return (
+            self.enable_eagerloads(False)
+            ._get_select_statement_only()
+            .label(name)
+        )
+
+    @overload
+    def as_scalar(  # type: ignore[overload-overlap]
+        self: Query[Tuple[_MAYBE_ENTITY]],
+    ) -> ScalarSelect[_MAYBE_ENTITY]: ...
+
+    @overload
+    def as_scalar(
+        self: Query[Tuple[_NOT_ENTITY]],
+    ) -> ScalarSelect[_NOT_ENTITY]: ...
+
+    @overload
+    def as_scalar(self) -> ScalarSelect[Any]: ...
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_query.Query.as_scalar` method is deprecated and will be "
+        "removed in a future release.  Please refer to "
+        ":meth:`_query.Query.scalar_subquery`.",
+    )
+    def as_scalar(self) -> ScalarSelect[Any]:
+        """Return the full SELECT statement represented by this
+        :class:`_query.Query`, converted to a scalar subquery.
+
+        """
+        return self.scalar_subquery()
+
+    @overload
+    def scalar_subquery(
+        self: Query[Tuple[_MAYBE_ENTITY]],
+    ) -> ScalarSelect[Any]: ...
+
+    @overload
+    def scalar_subquery(
+        self: Query[Tuple[_NOT_ENTITY]],
+    ) -> ScalarSelect[_NOT_ENTITY]: ...
+
+    @overload
+    def scalar_subquery(self) -> ScalarSelect[Any]: ...
+
+    def scalar_subquery(self) -> ScalarSelect[Any]:
+        """Return the full SELECT statement represented by this
+        :class:`_query.Query`, converted to a scalar subquery.
+
+        Analogous to
+        :meth:`sqlalchemy.sql.expression.SelectBase.scalar_subquery`.
+
+        .. versionchanged:: 1.4 The :meth:`_query.Query.scalar_subquery`
+           method replaces the :meth:`_query.Query.as_scalar` method.
+
+        .. seealso::
+
+            :meth:`_sql.Select.scalar_subquery` - v2 comparable method.
+
+        """
+
+        return (
+            self.enable_eagerloads(False)
+            ._get_select_statement_only()
+            .scalar_subquery()
+        )
+
+    @property
+    def selectable(self) -> Union[Select[_T], FromStatement[_T]]:
+        """Return the :class:`_expression.Select` object emitted by this
+        :class:`_query.Query`.
+
+        Used for :func:`_sa.inspect` compatibility, this is equivalent to::
+
+            query.enable_eagerloads(False).with_labels().statement
+
+        """
+        return self.__clause_element__()
+
+    def __clause_element__(self) -> Union[Select[_T], FromStatement[_T]]:
+        return (
+            self._with_compile_options(
+                _enable_eagerloads=False, _render_for_subquery=True
+            )
+            .set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+            .statement
+        )
+
+    @overload
+    def only_return_tuples(
+        self: Query[_O], value: Literal[True]
+    ) -> RowReturningQuery[Tuple[_O]]: ...
+
+    @overload
+    def only_return_tuples(
+        self: Query[_O], value: Literal[False]
+    ) -> Query[_O]: ...
+
+    @_generative
+    def only_return_tuples(self, value: bool) -> Query[Any]:
+        """When set to True, the query results will always be a
+        :class:`.Row` object.
+
+        This can change a query that normally returns a single entity
+        as a scalar to return a :class:`.Row` result in all cases.
+
+        .. seealso::
+
+            :meth:`.Query.tuples` - returns tuples, but also at the typing
+            level will type results as ``Tuple``.
+
+            :meth:`_query.Query.is_single_entity`
+
+            :meth:`_engine.Result.tuples` - v2 comparable method.
+
+        """
+        self.load_options += dict(_only_return_tuples=value)
+        return self
+
+    @property
+    def is_single_entity(self) -> bool:
+        """Indicates if this :class:`_query.Query`
+        returns tuples or single entities.
+
+        Returns True if this query returns a single entity for each instance
+        in its result list, and False if this query returns a tuple of entities
+        for each result.
+
+        .. versionadded:: 1.3.11
+
+        .. seealso::
+
+            :meth:`_query.Query.only_return_tuples`
+
+        """
+        return (
+            not self.load_options._only_return_tuples
+            and len(self._raw_columns) == 1
+            and "parententity" in self._raw_columns[0]._annotations
+            and isinstance(
+                self._raw_columns[0]._annotations["parententity"],
+                ORMColumnsClauseRole,
+            )
+        )
+
+    @_generative
+    def enable_eagerloads(self, value: bool) -> Self:
+        """Control whether or not eager joins and subqueries are
+        rendered.
+
+        When set to False, the returned Query will not render
+        eager joins regardless of :func:`~sqlalchemy.orm.joinedload`,
+        :func:`~sqlalchemy.orm.subqueryload` options
+        or mapper-level ``lazy='joined'``/``lazy='subquery'``
+        configurations.
+
+        This is used primarily when nesting the Query's
+        statement into a subquery or other
+        selectable, or when using :meth:`_query.Query.yield_per`.
+
+        """
+        self._compile_options += {"_enable_eagerloads": value}
+        return self
+
+    @_generative
+    def _with_compile_options(self, **opt: Any) -> Self:
+        self._compile_options += opt
+        return self
+
+    @util.became_legacy_20(
+        ":meth:`_orm.Query.with_labels` and :meth:`_orm.Query.apply_labels`",
+        alternative="Use set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL) "
+        "instead.",
+    )
+    def with_labels(self) -> Self:
+        return self.set_label_style(
+            SelectLabelStyle.LABEL_STYLE_TABLENAME_PLUS_COL
+        )
+
+    apply_labels = with_labels
+
+    @property
+    def get_label_style(self) -> SelectLabelStyle:
+        """
+        Retrieve the current label style.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :meth:`_sql.Select.get_label_style` - v2 equivalent method.
+
+        """
+        return self._label_style
+
+    def set_label_style(self, style: SelectLabelStyle) -> Self:
+        """Apply column labels to the return value of Query.statement.
+
+        Indicates that this Query's `statement` accessor should return
+        a SELECT statement that applies labels to all columns in the
+        form <tablename>_<columnname>; this is commonly used to
+        disambiguate columns from multiple tables which have the same
+        name.
+
+        When the `Query` actually issues SQL to load rows, it always
+        uses column labeling.
+
+        .. note:: The :meth:`_query.Query.set_label_style` method *only* applies
+           the output of :attr:`_query.Query.statement`, and *not* to any of
+           the result-row invoking systems of :class:`_query.Query` itself,
+           e.g.
+           :meth:`_query.Query.first`, :meth:`_query.Query.all`, etc.
+           To execute
+           a query using :meth:`_query.Query.set_label_style`, invoke the
+           :attr:`_query.Query.statement` using :meth:`.Session.execute`::
+
+                result = session.execute(
+                    query.set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL).statement
+                )
+
+        .. versionadded:: 1.4
+
+
+        .. seealso::
+
+            :meth:`_sql.Select.set_label_style` - v2 equivalent method.
+
+        """  # noqa
+        if self._label_style is not style:
+            self = self._generate()
+            self._label_style = style
+        return self
+
+    @_generative
+    def enable_assertions(self, value: bool) -> Self:
+        """Control whether assertions are generated.
+
+        When set to False, the returned Query will
+        not assert its state before certain operations,
+        including that LIMIT/OFFSET has not been applied
+        when filter() is called, no criterion exists
+        when get() is called, and no "from_statement()"
+        exists when filter()/order_by()/group_by() etc.
+        is called.  This more permissive mode is used by
+        custom Query subclasses to specify criterion or
+        other modifiers outside of the usual usage patterns.
+
+        Care should be taken to ensure that the usage
+        pattern is even possible.  A statement applied
+        by from_statement() will override any criterion
+        set by filter() or order_by(), for example.
+
+        """
+        self._enable_assertions = value
+        return self
+
+    @property
+    def whereclause(self) -> Optional[ColumnElement[bool]]:
+        """A readonly attribute which returns the current WHERE criterion for
+        this Query.
+
+        This returned value is a SQL expression construct, or ``None`` if no
+        criterion has been established.
+
+        .. seealso::
+
+            :attr:`_sql.Select.whereclause` - v2 equivalent property.
+
+        """
+        return BooleanClauseList._construct_for_whereclause(
+            self._where_criteria
+        )
+
+    @_generative
+    def _with_current_path(self, path: PathRegistry) -> Self:
+        """indicate that this query applies to objects loaded
+        within a certain path.
+
+        Used by deferred loaders (see strategies.py) which transfer
+        query options from an originating query to a newly generated
+        query intended for the deferred load.
+
+        """
+        self._compile_options += {"_current_path": path}
+        return self
+
+    @_generative
+    def yield_per(self, count: int) -> Self:
+        r"""Yield only ``count`` rows at a time.
+
+        The purpose of this method is when fetching very large result sets
+        (> 10K rows), to batch results in sub-collections and yield them
+        out partially, so that the Python interpreter doesn't need to declare
+        very large areas of memory which is both time consuming and leads
+        to excessive memory use.   The performance from fetching hundreds of
+        thousands of rows can often double when a suitable yield-per setting
+        (e.g. approximately 1000) is used, even with DBAPIs that buffer
+        rows (which are most).
+
+        As of SQLAlchemy 1.4, the :meth:`_orm.Query.yield_per` method is
+        equivalent to using the ``yield_per`` execution option at the ORM
+        level. See the section :ref:`orm_queryguide_yield_per` for further
+        background on this option.
+
+        .. seealso::
+
+            :ref:`orm_queryguide_yield_per`
+
+        """
+        self.load_options += {"_yield_per": count}
+        return self
+
+    @util.became_legacy_20(
+        ":meth:`_orm.Query.get`",
+        alternative="The method is now available as :meth:`_orm.Session.get`",
+    )
+    def get(self, ident: _PKIdentityArgument) -> Optional[Any]:
+        """Return an instance based on the given primary key identifier,
+        or ``None`` if not found.
+
+        E.g.::
+
+            my_user = session.query(User).get(5)
+
+            some_object = session.query(VersionedFoo).get((5, 10))
+
+            some_object = session.query(VersionedFoo).get({"id": 5, "version_id": 10})
+
+        :meth:`_query.Query.get` is special in that it provides direct
+        access to the identity map of the owning :class:`.Session`.
+        If the given primary key identifier is present
+        in the local identity map, the object is returned
+        directly from this collection and no SQL is emitted,
+        unless the object has been marked fully expired.
+        If not present,
+        a SELECT is performed in order to locate the object.
+
+        :meth:`_query.Query.get` also will perform a check if
+        the object is present in the identity map and
+        marked as expired - a SELECT
+        is emitted to refresh the object as well as to
+        ensure that the row is still present.
+        If not, :class:`~sqlalchemy.orm.exc.ObjectDeletedError` is raised.
+
+        :meth:`_query.Query.get` is only used to return a single
+        mapped instance, not multiple instances or
+        individual column constructs, and strictly
+        on a single primary key value.  The originating
+        :class:`_query.Query` must be constructed in this way,
+        i.e. against a single mapped entity,
+        with no additional filtering criterion.  Loading
+        options via :meth:`_query.Query.options` may be applied
+        however, and will be used if the object is not
+        yet locally present.
+
+        :param ident: A scalar, tuple, or dictionary representing the
+         primary key.  For a composite (e.g. multiple column) primary key,
+         a tuple or dictionary should be passed.
+
+         For a single-column primary key, the scalar calling form is typically
+         the most expedient.  If the primary key of a row is the value "5",
+         the call looks like::
+
+            my_object = query.get(5)
+
+         The tuple form contains primary key values typically in
+         the order in which they correspond to the mapped
+         :class:`_schema.Table`
+         object's primary key columns, or if the
+         :paramref:`_orm.Mapper.primary_key` configuration parameter were
+         used, in
+         the order used for that parameter. For example, if the primary key
+         of a row is represented by the integer
+         digits "5, 10" the call would look like::
+
+             my_object = query.get((5, 10))
+
+         The dictionary form should include as keys the mapped attribute names
+         corresponding to each element of the primary key.  If the mapped class
+         has the attributes ``id``, ``version_id`` as the attributes which
+         store the object's primary key value, the call would look like::
+
+            my_object = query.get({"id": 5, "version_id": 10})
+
+         .. versionadded:: 1.3 the :meth:`_query.Query.get`
+            method now optionally
+            accepts a dictionary of attribute names to values in order to
+            indicate a primary key identifier.
+
+
+        :return: The object instance, or ``None``.
+
+        """  # noqa: E501
+        self._no_criterion_assertion("get", order_by=False, distinct=False)
+
+        # we still implement _get_impl() so that baked query can override
+        # it
+        return self._get_impl(ident, loading.load_on_pk_identity)
+
+    def _get_impl(
+        self,
+        primary_key_identity: _PKIdentityArgument,
+        db_load_fn: Callable[..., Any],
+        identity_token: Optional[Any] = None,
+    ) -> Optional[Any]:
+        mapper = self._only_full_mapper_zero("get")
+        return self.session._get_impl(
+            mapper,
+            primary_key_identity,
+            db_load_fn,
+            populate_existing=self.load_options._populate_existing,
+            with_for_update=self._for_update_arg,
+            options=self._with_options,
+            identity_token=identity_token,
+            execution_options=self._execution_options,
+        )
+
+    @property
+    def lazy_loaded_from(self) -> Optional[InstanceState[Any]]:
+        """An :class:`.InstanceState` that is using this :class:`_query.Query`
+        for a lazy load operation.
+
+        .. deprecated:: 1.4  This attribute should be viewed via the
+           :attr:`.ORMExecuteState.lazy_loaded_from` attribute, within
+           the context of the :meth:`.SessionEvents.do_orm_execute`
+           event.
+
+        .. seealso::
+
+            :attr:`.ORMExecuteState.lazy_loaded_from`
+
+        """
+        return self.load_options._lazy_loaded_from  # type: ignore
+
+    @property
+    def _current_path(self) -> PathRegistry:
+        return self._compile_options._current_path  # type: ignore
+
+    @_generative
+    def correlate(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        """Return a :class:`.Query` construct which will correlate the given
+        FROM clauses to that of an enclosing :class:`.Query` or
+        :func:`~.expression.select`.
+
+        The method here accepts mapped classes, :func:`.aliased` constructs,
+        and :class:`_orm.Mapper` constructs as arguments, which are resolved
+        into expression constructs, in addition to appropriate expression
+        constructs.
+
+        The correlation arguments are ultimately passed to
+        :meth:`_expression.Select.correlate`
+        after coercion to expression constructs.
+
+        The correlation arguments take effect in such cases
+        as when :meth:`_query.Query.from_self` is used, or when
+        a subquery as returned by :meth:`_query.Query.subquery` is
+        embedded in another :func:`_expression.select` construct.
+
+        .. seealso::
+
+            :meth:`_sql.Select.correlate` - v2 equivalent method.
+
+        """
+
+        self._auto_correlate = False
+        if fromclauses and fromclauses[0] in {None, False}:
+            self._correlate = ()
+        else:
+            self._correlate = self._correlate + tuple(
+                coercions.expect(roles.FromClauseRole, f) for f in fromclauses
+            )
+        return self
+
+    @_generative
+    def autoflush(self, setting: bool) -> Self:
+        """Return a Query with a specific 'autoflush' setting.
+
+        As of SQLAlchemy 1.4, the :meth:`_orm.Query.autoflush` method
+        is equivalent to using the ``autoflush`` execution option at the
+        ORM level. See the section :ref:`orm_queryguide_autoflush` for
+        further background on this option.
+
+        """
+        self.load_options += {"_autoflush": setting}
+        return self
+
+    @_generative
+    def populate_existing(self) -> Self:
+        """Return a :class:`_query.Query`
+        that will expire and refresh all instances
+        as they are loaded, or reused from the current :class:`.Session`.
+
+        As of SQLAlchemy 1.4, the :meth:`_orm.Query.populate_existing` method
+        is equivalent to using the ``populate_existing`` execution option at
+        the ORM level. See the section :ref:`orm_queryguide_populate_existing`
+        for further background on this option.
+
+        """
+        self.load_options += {"_populate_existing": True}
+        return self
+
+    @_generative
+    def _with_invoke_all_eagers(self, value: bool) -> Self:
+        """Set the 'invoke all eagers' flag which causes joined- and
+        subquery loaders to traverse into already-loaded related objects
+        and collections.
+
+        Default is that of :attr:`_query.Query._invoke_all_eagers`.
+
+        """
+        self.load_options += {"_invoke_all_eagers": value}
+        return self
+
+    @util.became_legacy_20(
+        ":meth:`_orm.Query.with_parent`",
+        alternative="Use the :func:`_orm.with_parent` standalone construct.",
+    )
+    @util.preload_module("sqlalchemy.orm.relationships")
+    def with_parent(
+        self,
+        instance: object,
+        property: Optional[  # noqa: A002
+            attributes.QueryableAttribute[Any]
+        ] = None,
+        from_entity: Optional[_ExternalEntityType[Any]] = None,
+    ) -> Self:
+        """Add filtering criterion that relates the given instance
+        to a child object or collection, using its attribute state
+        as well as an established :func:`_orm.relationship()`
+        configuration.
+
+        The method uses the :func:`.with_parent` function to generate
+        the clause, the result of which is passed to
+        :meth:`_query.Query.filter`.
+
+        Parameters are the same as :func:`.with_parent`, with the exception
+        that the given property can be None, in which case a search is
+        performed against this :class:`_query.Query` object's target mapper.
+
+        :param instance:
+          An instance which has some :func:`_orm.relationship`.
+
+        :param property:
+          Class bound attribute which indicates
+          what relationship from the instance should be used to reconcile the
+          parent/child relationship.
+
+        :param from_entity:
+          Entity in which to consider as the left side.  This defaults to the
+          "zero" entity of the :class:`_query.Query` itself.
+
+        """
+        relationships = util.preloaded.orm_relationships
+
+        if from_entity:
+            entity_zero = inspect(from_entity)
+        else:
+            entity_zero = _legacy_filter_by_entity_zero(self)
+        if property is None:
+            # TODO: deprecate, property has to be supplied
+            mapper = object_mapper(instance)
+
+            for prop in mapper.iterate_properties:
+                if (
+                    isinstance(prop, relationships.RelationshipProperty)
+                    and prop.mapper is entity_zero.mapper  # type: ignore
+                ):
+                    property = prop  # type: ignore  # noqa: A001
+                    break
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "Could not locate a property which relates instances "
+                    "of class '%s' to instances of class '%s'"
+                    % (
+                        entity_zero.mapper.class_.__name__,  # type: ignore
+                        instance.__class__.__name__,
+                    )
+                )
+
+        return self.filter(
+            with_parent(
+                instance,
+                property,  # type: ignore
+                entity_zero.entity,  # type: ignore
+            )
+        )
+
+    @_generative
+    def add_entity(
+        self,
+        entity: _EntityType[Any],
+        alias: Optional[Union[Alias, Subquery]] = None,
+    ) -> Query[Any]:
+        """add a mapped entity to the list of result columns
+        to be returned.
+
+        .. seealso::
+
+            :meth:`_sql.Select.add_columns` - v2 comparable method.
+        """
+
+        if alias is not None:
+            # TODO: deprecate
+            entity = AliasedClass(entity, alias)
+
+        self._raw_columns = list(self._raw_columns)
+
+        self._raw_columns.append(
+            coercions.expect(
+                roles.ColumnsClauseRole, entity, apply_propagate_attrs=self
+            )
+        )
+        return self
+
+    @_generative
+    def with_session(self, session: Session) -> Self:
+        """Return a :class:`_query.Query` that will use the given
+        :class:`.Session`.
+
+        While the :class:`_query.Query`
+        object is normally instantiated using the
+        :meth:`.Session.query` method, it is legal to build the
+        :class:`_query.Query`
+        directly without necessarily using a :class:`.Session`.  Such a
+        :class:`_query.Query` object, or any :class:`_query.Query`
+        already associated
+        with a different :class:`.Session`, can produce a new
+        :class:`_query.Query`
+        object associated with a target session using this method::
+
+            from sqlalchemy.orm import Query
+
+            query = Query([MyClass]).filter(MyClass.id == 5)
+
+            result = query.with_session(my_session).one()
+
+        """
+
+        self.session = session
+        return self
+
+    def _legacy_from_self(
+        self, *entities: _ColumnsClauseArgument[Any]
+    ) -> Self:
+        # used for query.count() as well as for the same
+        # function in BakedQuery, as well as some old tests in test_baked.py.
+
+        fromclause = (
+            self.set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+            .correlate(None)
+            .subquery()
+            ._anonymous_fromclause()
+        )
+
+        q = self._from_selectable(fromclause)
+
+        if entities:
+            q._set_entities(entities)
+        return q
+
+    @_generative
+    def _set_enable_single_crit(self, val: bool) -> Self:
+        self._compile_options += {"_enable_single_crit": val}
+        return self
+
+    @_generative
+    def _from_selectable(
+        self, fromclause: FromClause, set_entity_from: bool = True
+    ) -> Self:
+        for attr in (
+            "_where_criteria",
+            "_order_by_clauses",
+            "_group_by_clauses",
+            "_limit_clause",
+            "_offset_clause",
+            "_last_joined_entity",
+            "_setup_joins",
+            "_memoized_select_entities",
+            "_distinct",
+            "_distinct_on",
+            "_having_criteria",
+            "_prefixes",
+            "_suffixes",
+        ):
+            self.__dict__.pop(attr, None)
+        self._set_select_from([fromclause], set_entity_from)
+        self._compile_options += {
+            "_enable_single_crit": False,
+        }
+
+        return self
+
+    @util.deprecated(
+        "1.4",
+        ":meth:`_query.Query.values` "
+        "is deprecated and will be removed in a "
+        "future release.  Please use :meth:`_query.Query.with_entities`",
+    )
+    def values(self, *columns: _ColumnsClauseArgument[Any]) -> Iterable[Any]:
+        """Return an iterator yielding result tuples corresponding
+        to the given list of columns
+
+        """
+        return self._values_no_warn(*columns)
+
+    _values = values
+
+    def _values_no_warn(
+        self, *columns: _ColumnsClauseArgument[Any]
+    ) -> Iterable[Any]:
+        if not columns:
+            return iter(())
+        q = self._clone().enable_eagerloads(False)
+        q._set_entities(columns)
+        if not q.load_options._yield_per:
+            q.load_options += {"_yield_per": 10}
+        return iter(q)
+
+    @util.deprecated(
+        "1.4",
+        ":meth:`_query.Query.value` "
+        "is deprecated and will be removed in a "
+        "future release.  Please use :meth:`_query.Query.with_entities` "
+        "in combination with :meth:`_query.Query.scalar`",
+    )
+    def value(self, column: _ColumnExpressionArgument[Any]) -> Any:
+        """Return a scalar result corresponding to the given
+        column expression.
+
+        """
+        try:
+            return next(self._values_no_warn(column))[0]  # type: ignore
+        except StopIteration:
+            return None
+
+    @overload
+    def with_entities(self, _entity: _EntityType[_O]) -> Query[_O]: ...
+
+    @overload
+    def with_entities(
+        self,
+        _colexpr: roles.TypedColumnsClauseRole[_T],
+    ) -> RowReturningQuery[Tuple[_T]]: ...
+
+    # START OVERLOADED FUNCTIONS self.with_entities RowReturningQuery 2-8
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_tuple_map_overloads.py
+
+    @overload
+    def with_entities(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+    ) -> RowReturningQuery[Tuple[_T0, _T1]]: ...
+
+    @overload
+    def with_entities(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2]]: ...
+
+    @overload
+    def with_entities(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+    @overload
+    def with_entities(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+    @overload
+    def with_entities(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+    @overload
+    def with_entities(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+    @overload
+    def with_entities(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+        __ent7: _TCCA[_T7],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]]: ...
+
+    # END OVERLOADED FUNCTIONS self.with_entities
+
+    @overload
+    def with_entities(
+        self, *entities: _ColumnsClauseArgument[Any]
+    ) -> Query[Any]: ...
+
+    @_generative
+    def with_entities(
+        self, *entities: _ColumnsClauseArgument[Any], **__kw: Any
+    ) -> Query[Any]:
+        r"""Return a new :class:`_query.Query`
+        replacing the SELECT list with the
+        given entities.
+
+        e.g.::
+
+            # Users, filtered on some arbitrary criterion
+            # and then ordered by related email address
+            q = (
+                session.query(User)
+                .join(User.address)
+                .filter(User.name.like("%ed%"))
+                .order_by(Address.email)
+            )
+
+            # given *only* User.id==5, Address.email, and 'q', what
+            # would the *next* User in the result be ?
+            subq = (
+                q.with_entities(Address.email)
+                .order_by(None)
+                .filter(User.id == 5)
+                .subquery()
+            )
+            q = q.join((subq, subq.c.email < Address.email)).limit(1)
+
+        .. seealso::
+
+            :meth:`_sql.Select.with_only_columns` - v2 comparable method.
+        """
+        if __kw:
+            raise _no_kw()
+
+        # Query has all the same fields as Select for this operation
+        # this could in theory be based on a protocol but not sure if it's
+        # worth it
+        _MemoizedSelectEntities._generate_for_statement(self)  # type: ignore
+        self._set_entities(entities)
+        return self
+
+    @_generative
+    def add_columns(
+        self, *column: _ColumnExpressionArgument[Any]
+    ) -> Query[Any]:
+        """Add one or more column expressions to the list
+        of result columns to be returned.
+
+        .. seealso::
+
+            :meth:`_sql.Select.add_columns` - v2 comparable method.
+        """
+
+        self._raw_columns = list(self._raw_columns)
+
+        self._raw_columns.extend(
+            coercions.expect(
+                roles.ColumnsClauseRole,
+                c,
+                apply_propagate_attrs=self,
+                post_inspect=True,
+            )
+            for c in column
+        )
+        return self
+
+    @util.deprecated(
+        "1.4",
+        ":meth:`_query.Query.add_column` "
+        "is deprecated and will be removed in a "
+        "future release.  Please use :meth:`_query.Query.add_columns`",
+    )
+    def add_column(self, column: _ColumnExpressionArgument[Any]) -> Query[Any]:
+        """Add a column expression to the list of result columns to be
+        returned.
+
+        """
+        return self.add_columns(column)
+
+    @_generative
+    def options(self, *args: ExecutableOption) -> Self:
+        """Return a new :class:`_query.Query` object,
+        applying the given list of
+        mapper options.
+
+        Most supplied options regard changing how column- and
+        relationship-mapped attributes are loaded.
+
+        .. seealso::
+
+            :ref:`loading_columns`
+
+            :ref:`relationship_loader_options`
+
+        """
+
+        opts = tuple(util.flatten_iterator(args))
+        if self._compile_options._current_path:
+            # opting for lower method overhead for the checks
+            for opt in opts:
+                if not opt._is_core and opt._is_legacy_option:  # type: ignore
+                    opt.process_query_conditionally(self)  # type: ignore
+        else:
+            for opt in opts:
+                if not opt._is_core and opt._is_legacy_option:  # type: ignore
+                    opt.process_query(self)  # type: ignore
+
+        self._with_options += opts
+        return self
+
+    def with_transformation(
+        self, fn: Callable[[Query[Any]], Query[Any]]
+    ) -> Query[Any]:
+        """Return a new :class:`_query.Query` object transformed by
+        the given function.
+
+        E.g.::
+
+            def filter_something(criterion):
+                def transform(q):
+                    return q.filter(criterion)
+
+                return transform
+
+
+            q = q.with_transformation(filter_something(x == 5))
+
+        This allows ad-hoc recipes to be created for :class:`_query.Query`
+        objects.
+
+        """
+        return fn(self)
+
+    def get_execution_options(self) -> _ImmutableExecuteOptions:
+        """Get the non-SQL options which will take effect during execution.
+
+        .. versionadded:: 1.3
+
+        .. seealso::
+
+            :meth:`_query.Query.execution_options`
+
+            :meth:`_sql.Select.get_execution_options` - v2 comparable method.
+
+        """
+        return self._execution_options
+
+    @overload
+    def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        no_parameters: bool = False,
+        stream_results: bool = False,
+        max_row_buffer: int = ...,
+        yield_per: int = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        populate_existing: bool = False,
+        autoflush: bool = False,
+        preserve_rowcount: bool = False,
+        **opt: Any,
+    ) -> Self: ...
+
+    @overload
+    def execution_options(self, **opt: Any) -> Self: ...
+
+    @_generative
+    def execution_options(self, **kwargs: Any) -> Self:
+        """Set non-SQL options which take effect during execution.
+
+        Options allowed here include all of those accepted by
+        :meth:`_engine.Connection.execution_options`, as well as a series
+        of ORM specific options:
+
+        ``populate_existing=True`` - equivalent to using
+        :meth:`_orm.Query.populate_existing`
+
+        ``autoflush=True|False`` - equivalent to using
+        :meth:`_orm.Query.autoflush`
+
+        ``yield_per=<value>`` - equivalent to using
+        :meth:`_orm.Query.yield_per`
+
+        Note that the ``stream_results`` execution option is enabled
+        automatically if the :meth:`~sqlalchemy.orm.query.Query.yield_per()`
+        method or execution option is used.
+
+        .. versionadded:: 1.4 - added ORM options to
+           :meth:`_orm.Query.execution_options`
+
+        The execution options may also be specified on a per execution basis
+        when using :term:`2.0 style` queries via the
+        :paramref:`_orm.Session.execution_options` parameter.
+
+        .. warning:: The
+           :paramref:`_engine.Connection.execution_options.stream_results`
+           parameter should not be used at the level of individual ORM
+           statement executions, as the :class:`_orm.Session` will not track
+           objects from different schema translate maps within a single
+           session.  For multiple schema translate maps within the scope of a
+           single :class:`_orm.Session`, see :ref:`examples_sharding`.
+
+
+        .. seealso::
+
+            :ref:`engine_stream_results`
+
+            :meth:`_query.Query.get_execution_options`
+
+            :meth:`_sql.Select.execution_options` - v2 equivalent method.
+
+        """
+        self._execution_options = self._execution_options.union(kwargs)
+        return self
+
+    @_generative
+    def with_for_update(
+        self,
+        *,
+        nowait: bool = False,
+        read: bool = False,
+        of: Optional[_ForUpdateOfArgument] = None,
+        skip_locked: bool = False,
+        key_share: bool = False,
+    ) -> Self:
+        """return a new :class:`_query.Query`
+        with the specified options for the
+        ``FOR UPDATE`` clause.
+
+        The behavior of this method is identical to that of
+        :meth:`_expression.GenerativeSelect.with_for_update`.
+        When called with no arguments,
+        the resulting ``SELECT`` statement will have a ``FOR UPDATE`` clause
+        appended.  When additional arguments are specified, backend-specific
+        options such as ``FOR UPDATE NOWAIT`` or ``LOCK IN SHARE MODE``
+        can take effect.
+
+        E.g.::
+
+            q = (
+                sess.query(User)
+                .populate_existing()
+                .with_for_update(nowait=True, of=User)
+            )
+
+        The above query on a PostgreSQL backend will render like:
+
+        .. sourcecode:: sql
+
+            SELECT users.id AS users_id FROM users FOR UPDATE OF users NOWAIT
+
+        .. warning::
+
+            Using ``with_for_update`` in the context of eager loading
+            relationships is not officially supported or recommended by
+            SQLAlchemy and may not work with certain queries on various
+            database backends.  When ``with_for_update`` is successfully used
+            with a query that involves :func:`_orm.joinedload`, SQLAlchemy will
+            attempt to emit SQL that locks all involved tables.
+
+        .. note::  It is generally a good idea to combine the use of the
+           :meth:`_orm.Query.populate_existing` method when using the
+           :meth:`_orm.Query.with_for_update` method.   The purpose of
+           :meth:`_orm.Query.populate_existing` is to force all the data read
+           from the SELECT to be populated into the ORM objects returned,
+           even if these objects are already in the :term:`identity map`.
+
+        .. seealso::
+
+            :meth:`_expression.GenerativeSelect.with_for_update`
+            - Core level method with
+            full argument and behavioral description.
+
+            :meth:`_orm.Query.populate_existing` - overwrites attributes of
+            objects already loaded in the identity map.
+
+        """  # noqa: E501
+
+        self._for_update_arg = ForUpdateArg(
+            read=read,
+            nowait=nowait,
+            of=of,
+            skip_locked=skip_locked,
+            key_share=key_share,
+        )
+        return self
+
+    @_generative
+    def params(
+        self, __params: Optional[Dict[str, Any]] = None, **kw: Any
+    ) -> Self:
+        r"""Add values for bind parameters which may have been
+        specified in filter().
+
+        Parameters may be specified using \**kwargs, or optionally a single
+        dictionary as the first positional argument. The reason for both is
+        that \**kwargs is convenient, however some parameter dictionaries
+        contain unicode keys in which case \**kwargs cannot be used.
+
+        """
+        if __params:
+            kw.update(__params)
+        self._params = self._params.union(kw)
+        return self
+
+    def where(self, *criterion: _ColumnExpressionArgument[bool]) -> Self:
+        """A synonym for :meth:`.Query.filter`.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :meth:`_sql.Select.where` - v2 equivalent method.
+
+        """
+        return self.filter(*criterion)
+
+    @_generative
+    @_assertions(_no_statement_condition, _no_limit_offset)
+    def filter(self, *criterion: _ColumnExpressionArgument[bool]) -> Self:
+        r"""Apply the given filtering criterion to a copy
+        of this :class:`_query.Query`, using SQL expressions.
+
+        e.g.::
+
+            session.query(MyClass).filter(MyClass.name == "some name")
+
+        Multiple criteria may be specified as comma separated; the effect
+        is that they will be joined together using the :func:`.and_`
+        function::
+
+            session.query(MyClass).filter(MyClass.name == "some name", MyClass.id > 5)
+
+        The criterion is any SQL expression object applicable to the
+        WHERE clause of a select.   String expressions are coerced
+        into SQL expression constructs via the :func:`_expression.text`
+        construct.
+
+        .. seealso::
+
+            :meth:`_query.Query.filter_by` - filter on keyword expressions.
+
+            :meth:`_sql.Select.where` - v2 equivalent method.
+
+        """  # noqa: E501
+        for crit in list(criterion):
+            crit = coercions.expect(
+                roles.WhereHavingRole, crit, apply_propagate_attrs=self
+            )
+
+            self._where_criteria += (crit,)
+        return self
+
+    @util.memoized_property
+    def _last_joined_entity(
+        self,
+    ) -> Optional[Union[_InternalEntityType[Any], _JoinTargetElement]]:
+        if self._setup_joins:
+            return _determine_last_joined_entity(
+                self._setup_joins,
+            )
+        else:
+            return None
+
+    def _filter_by_zero(self) -> Any:
+        """for the filter_by() method, return the target entity for which
+        we will attempt to derive an expression from based on string name.
+
+        """
+
+        if self._setup_joins:
+            _last_joined_entity = self._last_joined_entity
+            if _last_joined_entity is not None:
+                return _last_joined_entity
+
+        # discussion related to #7239
+        # special check determines if we should try to derive attributes
+        # for filter_by() from the "from object", i.e., if the user
+        # called query.select_from(some selectable).filter_by(some_attr=value).
+        # We don't want to do that in the case that methods like
+        # from_self(), select_entity_from(), or a set op like union() were
+        # called; while these methods also place a
+        # selectable in the _from_obj collection, they also set up
+        # the _set_base_alias boolean which turns on the whole "adapt the
+        # entity to this selectable" thing, meaning the query still continues
+        # to construct itself in terms of the lead entity that was passed
+        # to query(), e.g. query(User).from_self() is still in terms of User,
+        # and not the subquery that from_self() created.   This feature of
+        # "implicitly adapt all occurrences of entity X to some arbitrary
+        # subquery" is the main thing I am trying to do away with in 2.0 as
+        # users should now used aliased() for that, but I can't entirely get
+        # rid of it due to query.union() and other set ops relying upon it.
+        #
+        # compare this to the base Select()._filter_by_zero() which can
+        # just return self._from_obj[0] if present, because there is no
+        # "_set_base_alias" feature.
+        #
+        # IOW, this conditional essentially detects if
+        # "select_from(some_selectable)" has been called, as opposed to
+        # "select_entity_from()", "from_self()"
+        # or "union() / some_set_op()".
+        if self._from_obj and not self._compile_options._set_base_alias:
+            return self._from_obj[0]
+
+        return self._raw_columns[0]
+
+    def filter_by(self, **kwargs: Any) -> Self:
+        r"""Apply the given filtering criterion to a copy
+        of this :class:`_query.Query`, using keyword expressions.
+
+        e.g.::
+
+            session.query(MyClass).filter_by(name="some name")
+
+        Multiple criteria may be specified as comma separated; the effect
+        is that they will be joined together using the :func:`.and_`
+        function::
+
+            session.query(MyClass).filter_by(name="some name", id=5)
+
+        The keyword expressions are extracted from the primary
+        entity of the query, or the last entity that was the
+        target of a call to :meth:`_query.Query.join`.
+
+        .. seealso::
+
+            :meth:`_query.Query.filter` - filter on SQL expressions.
+
+            :meth:`_sql.Select.filter_by` - v2 comparable method.
+
+        """
+        from_entity = self._filter_by_zero()
+
+        clauses = [
+            _entity_namespace_key(from_entity, key) == value
+            for key, value in kwargs.items()
+        ]
+        return self.filter(*clauses)
+
+    @_generative
+    def order_by(
+        self,
+        __first: Union[
+            Literal[None, False, _NoArg.NO_ARG],
+            _ColumnExpressionOrStrLabelArgument[Any],
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionOrStrLabelArgument[Any],
+    ) -> Self:
+        """Apply one or more ORDER BY criteria to the query and return
+        the newly resulting :class:`_query.Query`.
+
+        e.g.::
+
+            q = session.query(Entity).order_by(Entity.id, Entity.name)
+
+        Calling this method multiple times is equivalent to calling it once
+        with all the clauses concatenated. All existing ORDER BY criteria may
+        be cancelled by passing ``None`` by itself.  New ORDER BY criteria may
+        then be added by invoking :meth:`_orm.Query.order_by` again, e.g.::
+
+            # will erase all ORDER BY and ORDER BY new_col alone
+            q = q.order_by(None).order_by(new_col)
+
+        .. seealso::
+
+            These sections describe ORDER BY in terms of :term:`2.0 style`
+            invocation but apply to :class:`_orm.Query` as well:
+
+            :ref:`tutorial_order_by` - in the :ref:`unified_tutorial`
+
+            :ref:`tutorial_order_by_label` - in the :ref:`unified_tutorial`
+
+            :meth:`_sql.Select.order_by` - v2 equivalent method.
+
+        """
+
+        for assertion in (self._no_statement_condition, self._no_limit_offset):
+            assertion("order_by")
+
+        if not clauses and (__first is None or __first is False):
+            self._order_by_clauses = ()
+        elif __first is not _NoArg.NO_ARG:
+            criterion = tuple(
+                coercions.expect(roles.OrderByRole, clause)
+                for clause in (__first,) + clauses
+            )
+            self._order_by_clauses += criterion
+
+        return self
+
+    @_generative
+    def group_by(
+        self,
+        __first: Union[
+            Literal[None, False, _NoArg.NO_ARG],
+            _ColumnExpressionOrStrLabelArgument[Any],
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionOrStrLabelArgument[Any],
+    ) -> Self:
+        """Apply one or more GROUP BY criterion to the query and return
+        the newly resulting :class:`_query.Query`.
+
+        All existing GROUP BY settings can be suppressed by
+        passing ``None`` - this will suppress any GROUP BY configured
+        on mappers as well.
+
+        .. seealso::
+
+            These sections describe GROUP BY in terms of :term:`2.0 style`
+            invocation but apply to :class:`_orm.Query` as well:
+
+            :ref:`tutorial_group_by_w_aggregates` - in the
+            :ref:`unified_tutorial`
+
+            :ref:`tutorial_order_by_label` - in the :ref:`unified_tutorial`
+
+            :meth:`_sql.Select.group_by` - v2 equivalent method.
+
+        """
+
+        for assertion in (self._no_statement_condition, self._no_limit_offset):
+            assertion("group_by")
+
+        if not clauses and (__first is None or __first is False):
+            self._group_by_clauses = ()
+        elif __first is not _NoArg.NO_ARG:
+            criterion = tuple(
+                coercions.expect(roles.GroupByRole, clause)
+                for clause in (__first,) + clauses
+            )
+            self._group_by_clauses += criterion
+        return self
+
+    @_generative
+    @_assertions(_no_statement_condition, _no_limit_offset)
+    def having(self, *having: _ColumnExpressionArgument[bool]) -> Self:
+        r"""Apply a HAVING criterion to the query and return the
+        newly resulting :class:`_query.Query`.
+
+        :meth:`_query.Query.having` is used in conjunction with
+        :meth:`_query.Query.group_by`.
+
+        HAVING criterion makes it possible to use filters on aggregate
+        functions like COUNT, SUM, AVG, MAX, and MIN, eg.::
+
+            q = (
+                session.query(User.id)
+                .join(User.addresses)
+                .group_by(User.id)
+                .having(func.count(Address.id) > 2)
+            )
+
+        .. seealso::
+
+            :meth:`_sql.Select.having` - v2 equivalent method.
+
+        """
+
+        for criterion in having:
+            having_criteria = coercions.expect(
+                roles.WhereHavingRole, criterion
+            )
+            self._having_criteria += (having_criteria,)
+        return self
+
+    def _set_op(self, expr_fn: Any, *q: Query[Any]) -> Self:
+        list_of_queries = (self,) + q
+        return self._from_selectable(expr_fn(*(list_of_queries)).subquery())
+
+    def union(self, *q: Query[Any]) -> Self:
+        """Produce a UNION of this Query against one or more queries.
+
+        e.g.::
+
+            q1 = sess.query(SomeClass).filter(SomeClass.foo == "bar")
+            q2 = sess.query(SomeClass).filter(SomeClass.bar == "foo")
+
+            q3 = q1.union(q2)
+
+        The method accepts multiple Query objects so as to control
+        the level of nesting.  A series of ``union()`` calls such as::
+
+            x.union(y).union(z).all()
+
+        will nest on each ``union()``, and produces:
+
+        .. sourcecode:: sql
+
+            SELECT * FROM (SELECT * FROM (SELECT * FROM X UNION
+                            SELECT * FROM y) UNION SELECT * FROM Z)
+
+        Whereas::
+
+            x.union(y, z).all()
+
+        produces:
+
+        .. sourcecode:: sql
+
+            SELECT * FROM (SELECT * FROM X UNION SELECT * FROM y UNION
+                            SELECT * FROM Z)
+
+        Note that many database backends do not allow ORDER BY to
+        be rendered on a query called within UNION, EXCEPT, etc.
+        To disable all ORDER BY clauses including those configured
+        on mappers, issue ``query.order_by(None)`` - the resulting
+        :class:`_query.Query` object will not render ORDER BY within
+        its SELECT statement.
+
+        .. seealso::
+
+            :meth:`_sql.Select.union` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.union, *q)
+
+    def union_all(self, *q: Query[Any]) -> Self:
+        """Produce a UNION ALL of this Query against one or more queries.
+
+        Works the same way as :meth:`~sqlalchemy.orm.query.Query.union`. See
+        that method for usage examples.
+
+        .. seealso::
+
+            :meth:`_sql.Select.union_all` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.union_all, *q)
+
+    def intersect(self, *q: Query[Any]) -> Self:
+        """Produce an INTERSECT of this Query against one or more queries.
+
+        Works the same way as :meth:`~sqlalchemy.orm.query.Query.union`. See
+        that method for usage examples.
+
+        .. seealso::
+
+            :meth:`_sql.Select.intersect` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.intersect, *q)
+
+    def intersect_all(self, *q: Query[Any]) -> Self:
+        """Produce an INTERSECT ALL of this Query against one or more queries.
+
+        Works the same way as :meth:`~sqlalchemy.orm.query.Query.union`. See
+        that method for usage examples.
+
+        .. seealso::
+
+            :meth:`_sql.Select.intersect_all` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.intersect_all, *q)
+
+    def except_(self, *q: Query[Any]) -> Self:
+        """Produce an EXCEPT of this Query against one or more queries.
+
+        Works the same way as :meth:`~sqlalchemy.orm.query.Query.union`. See
+        that method for usage examples.
+
+        .. seealso::
+
+            :meth:`_sql.Select.except_` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.except_, *q)
+
+    def except_all(self, *q: Query[Any]) -> Self:
+        """Produce an EXCEPT ALL of this Query against one or more queries.
+
+        Works the same way as :meth:`~sqlalchemy.orm.query.Query.union`. See
+        that method for usage examples.
+
+        .. seealso::
+
+            :meth:`_sql.Select.except_all` - v2 equivalent method.
+
+        """
+        return self._set_op(expression.except_all, *q)
+
+    @_generative
+    @_assertions(_no_statement_condition, _no_limit_offset)
+    def join(
+        self,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        isouter: bool = False,
+        full: bool = False,
+    ) -> Self:
+        r"""Create a SQL JOIN against this :class:`_query.Query`
+        object's criterion
+        and apply generatively, returning the newly resulting
+        :class:`_query.Query`.
+
+        **Simple Relationship Joins**
+
+        Consider a mapping between two classes ``User`` and ``Address``,
+        with a relationship ``User.addresses`` representing a collection
+        of ``Address`` objects associated with each ``User``.   The most
+        common usage of :meth:`_query.Query.join`
+        is to create a JOIN along this
+        relationship, using the ``User.addresses`` attribute as an indicator
+        for how this should occur::
+
+            q = session.query(User).join(User.addresses)
+
+        Where above, the call to :meth:`_query.Query.join` along
+        ``User.addresses`` will result in SQL approximately equivalent to:
+
+        .. sourcecode:: sql
+
+            SELECT user.id, user.name
+            FROM user JOIN address ON user.id = address.user_id
+
+        In the above example we refer to ``User.addresses`` as passed to
+        :meth:`_query.Query.join` as the "on clause", that is, it indicates
+        how the "ON" portion of the JOIN should be constructed.
+
+        To construct a chain of joins, multiple :meth:`_query.Query.join`
+        calls may be used.  The relationship-bound attribute implies both
+        the left and right side of the join at once::
+
+            q = (
+                session.query(User)
+                .join(User.orders)
+                .join(Order.items)
+                .join(Item.keywords)
+            )
+
+        .. note:: as seen in the above example, **the order in which each
+           call to the join() method occurs is important**.    Query would not,
+           for example, know how to join correctly if we were to specify
+           ``User``, then ``Item``, then ``Order``, in our chain of joins; in
+           such a case, depending on the arguments passed, it may raise an
+           error that it doesn't know how to join, or it may produce invalid
+           SQL in which case the database will raise an error. In correct
+           practice, the
+           :meth:`_query.Query.join` method is invoked in such a way that lines
+           up with how we would want the JOIN clauses in SQL to be
+           rendered, and each call should represent a clear link from what
+           precedes it.
+
+        **Joins to a Target Entity or Selectable**
+
+        A second form of :meth:`_query.Query.join` allows any mapped entity or
+        core selectable construct as a target.   In this usage,
+        :meth:`_query.Query.join` will attempt to create a JOIN along the
+        natural foreign key relationship between two entities::
+
+            q = session.query(User).join(Address)
+
+        In the above calling form, :meth:`_query.Query.join` is called upon to
+        create the "on clause" automatically for us.  This calling form will
+        ultimately raise an error if either there are no foreign keys between
+        the two entities, or if there are multiple foreign key linkages between
+        the target entity and the entity or entities already present on the
+        left side such that creating a join requires more information.  Note
+        that when indicating a join to a target without any ON clause, ORM
+        configured relationships are not taken into account.
+
+        **Joins to a Target with an ON Clause**
+
+        The third calling form allows both the target entity as well
+        as the ON clause to be passed explicitly.    A example that includes
+        a SQL expression as the ON clause is as follows::
+
+            q = session.query(User).join(Address, User.id == Address.user_id)
+
+        The above form may also use a relationship-bound attribute as the
+        ON clause as well::
+
+            q = session.query(User).join(Address, User.addresses)
+
+        The above syntax can be useful for the case where we wish
+        to join to an alias of a particular target entity.  If we wanted
+        to join to ``Address`` twice, it could be achieved using two
+        aliases set up using the :func:`~sqlalchemy.orm.aliased` function::
+
+            a1 = aliased(Address)
+            a2 = aliased(Address)
+
+            q = (
+                session.query(User)
+                .join(a1, User.addresses)
+                .join(a2, User.addresses)
+                .filter(a1.email_address == "ed@foo.com")
+                .filter(a2.email_address == "ed@bar.com")
+            )
+
+        The relationship-bound calling form can also specify a target entity
+        using the :meth:`_orm.PropComparator.of_type` method; a query
+        equivalent to the one above would be::
+
+            a1 = aliased(Address)
+            a2 = aliased(Address)
+
+            q = (
+                session.query(User)
+                .join(User.addresses.of_type(a1))
+                .join(User.addresses.of_type(a2))
+                .filter(a1.email_address == "ed@foo.com")
+                .filter(a2.email_address == "ed@bar.com")
+            )
+
+        **Augmenting Built-in ON Clauses**
+
+        As a substitute for providing a full custom ON condition for an
+        existing relationship, the :meth:`_orm.PropComparator.and_` function
+        may be applied to a relationship attribute to augment additional
+        criteria into the ON clause; the additional criteria will be combined
+        with the default criteria using AND::
+
+            q = session.query(User).join(
+                User.addresses.and_(Address.email_address != "foo@bar.com")
+            )
+
+        .. versionadded:: 1.4
+
+        **Joining to Tables and Subqueries**
+
+
+        The target of a join may also be any table or SELECT statement,
+        which may be related to a target entity or not.   Use the
+        appropriate ``.subquery()`` method in order to make a subquery
+        out of a query::
+
+            subq = (
+                session.query(Address)
+                .filter(Address.email_address == "ed@foo.com")
+                .subquery()
+            )
+
+
+            q = session.query(User).join(subq, User.id == subq.c.user_id)
+
+        Joining to a subquery in terms of a specific relationship and/or
+        target entity may be achieved by linking the subquery to the
+        entity using :func:`_orm.aliased`::
+
+            subq = (
+                session.query(Address)
+                .filter(Address.email_address == "ed@foo.com")
+                .subquery()
+            )
+
+            address_subq = aliased(Address, subq)
+
+            q = session.query(User).join(User.addresses.of_type(address_subq))
+
+        **Controlling what to Join From**
+
+        In cases where the left side of the current state of
+        :class:`_query.Query` is not in line with what we want to join from,
+        the :meth:`_query.Query.select_from` method may be used::
+
+            q = (
+                session.query(Address)
+                .select_from(User)
+                .join(User.addresses)
+                .filter(User.name == "ed")
+            )
+
+        Which will produce SQL similar to:
+
+        .. sourcecode:: sql
+
+            SELECT address.* FROM user
+                JOIN address ON user.id=address.user_id
+                WHERE user.name = :name_1
+
+        .. seealso::
+
+            :meth:`_sql.Select.join` - v2 equivalent method.
+
+        :param \*props: Incoming arguments for :meth:`_query.Query.join`,
+         the props collection in modern use should be considered to be a  one
+         or two argument form, either as a single "target" entity or ORM
+         attribute-bound relationship, or as a target entity plus an "on
+         clause" which  may be a SQL expression or ORM attribute-bound
+         relationship.
+
+        :param isouter=False: If True, the join used will be a left outer join,
+         just as if the :meth:`_query.Query.outerjoin` method were called.
+
+        :param full=False: render FULL OUTER JOIN; implies ``isouter``.
+
+        """
+
+        join_target = coercions.expect(
+            roles.JoinTargetRole,
+            target,
+            apply_propagate_attrs=self,
+            legacy=True,
+        )
+        if onclause is not None:
+            onclause_element = coercions.expect(
+                roles.OnClauseRole, onclause, legacy=True
+            )
+        else:
+            onclause_element = None
+
+        self._setup_joins += (
+            (
+                join_target,
+                onclause_element,
+                None,
+                {
+                    "isouter": isouter,
+                    "full": full,
+                },
+            ),
+        )
+
+        self.__dict__.pop("_last_joined_entity", None)
+        return self
+
+    def outerjoin(
+        self,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        full: bool = False,
+    ) -> Self:
+        """Create a left outer join against this ``Query`` object's criterion
+        and apply generatively, returning the newly resulting ``Query``.
+
+        Usage is the same as the ``join()`` method.
+
+        .. seealso::
+
+            :meth:`_sql.Select.outerjoin` - v2 equivalent method.
+
+        """
+        return self.join(target, onclause=onclause, isouter=True, full=full)
+
+    @_generative
+    @_assertions(_no_statement_condition)
+    def reset_joinpoint(self) -> Self:
+        """Return a new :class:`.Query`, where the "join point" has
+        been reset back to the base FROM entities of the query.
+
+        This method is usually used in conjunction with the
+        ``aliased=True`` feature of the :meth:`~.Query.join`
+        method.  See the example in :meth:`~.Query.join` for how
+        this is used.
+
+        """
+        self._last_joined_entity = None
+
+        return self
+
+    @_generative
+    @_assertions(_no_clauseelement_condition)
+    def select_from(self, *from_obj: _FromClauseArgument) -> Self:
+        r"""Set the FROM clause of this :class:`.Query` explicitly.
+
+        :meth:`.Query.select_from` is often used in conjunction with
+        :meth:`.Query.join` in order to control which entity is selected
+        from on the "left" side of the join.
+
+        The entity or selectable object here effectively replaces the
+        "left edge" of any calls to :meth:`~.Query.join`, when no
+        joinpoint is otherwise established - usually, the default "join
+        point" is the leftmost entity in the :class:`~.Query` object's
+        list of entities to be selected.
+
+        A typical example::
+
+            q = (
+                session.query(Address)
+                .select_from(User)
+                .join(User.addresses)
+                .filter(User.name == "ed")
+            )
+
+        Which produces SQL equivalent to:
+
+        .. sourcecode:: sql
+
+            SELECT address.* FROM user
+            JOIN address ON user.id=address.user_id
+            WHERE user.name = :name_1
+
+        :param \*from_obj: collection of one or more entities to apply
+         to the FROM clause.  Entities can be mapped classes,
+         :class:`.AliasedClass` objects, :class:`.Mapper` objects
+         as well as core :class:`.FromClause` elements like subqueries.
+
+        .. seealso::
+
+            :meth:`~.Query.join`
+
+            :meth:`.Query.select_entity_from`
+
+            :meth:`_sql.Select.select_from` - v2 equivalent method.
+
+        """
+
+        self._set_select_from(from_obj, False)
+        return self
+
+    def __getitem__(self, item: Any) -> Any:
+        return orm_util._getitem(
+            self,
+            item,
+        )
+
+    @_generative
+    @_assertions(_no_statement_condition)
+    def slice(
+        self,
+        start: int,
+        stop: int,
+    ) -> Self:
+        """Computes the "slice" of the :class:`_query.Query` represented by
+        the given indices and returns the resulting :class:`_query.Query`.
+
+        The start and stop indices behave like the argument to Python's
+        built-in :func:`range` function. This method provides an
+        alternative to using ``LIMIT``/``OFFSET`` to get a slice of the
+        query.
+
+        For example, ::
+
+            session.query(User).order_by(User.id).slice(1, 3)
+
+        renders as
+
+        .. sourcecode:: sql
+
+           SELECT users.id AS users_id,
+                  users.name AS users_name
+           FROM users ORDER BY users.id
+           LIMIT ? OFFSET ?
+           (2, 1)
+
+        .. seealso::
+
+           :meth:`_query.Query.limit`
+
+           :meth:`_query.Query.offset`
+
+           :meth:`_sql.Select.slice` - v2 equivalent method.
+
+        """
+
+        self._limit_clause, self._offset_clause = sql_util._make_slice(
+            self._limit_clause, self._offset_clause, start, stop
+        )
+        return self
+
+    @_generative
+    @_assertions(_no_statement_condition)
+    def limit(self, limit: _LimitOffsetType) -> Self:
+        """Apply a ``LIMIT`` to the query and return the newly resulting
+        ``Query``.
+
+        .. seealso::
+
+            :meth:`_sql.Select.limit` - v2 equivalent method.
+
+        """
+        self._limit_clause = sql_util._offset_or_limit_clause(limit)
+        return self
+
+    @_generative
+    @_assertions(_no_statement_condition)
+    def offset(self, offset: _LimitOffsetType) -> Self:
+        """Apply an ``OFFSET`` to the query and return the newly resulting
+        ``Query``.
+
+        .. seealso::
+
+            :meth:`_sql.Select.offset` - v2 equivalent method.
+        """
+        self._offset_clause = sql_util._offset_or_limit_clause(offset)
+        return self
+
+    @_generative
+    @_assertions(_no_statement_condition)
+    def distinct(self, *expr: _ColumnExpressionArgument[Any]) -> Self:
+        r"""Apply a ``DISTINCT`` to the query and return the newly resulting
+        ``Query``.
+
+
+        .. note::
+
+            The ORM-level :meth:`.distinct` call includes logic that will
+            automatically add columns from the ORDER BY of the query to the
+            columns clause of the SELECT statement, to satisfy the common need
+            of the database backend that ORDER BY columns be part of the SELECT
+            list when DISTINCT is used.   These columns *are not* added to the
+            list of columns actually fetched by the :class:`_query.Query`,
+            however,
+            so would not affect results. The columns are passed through when
+            using the :attr:`_query.Query.statement` accessor, however.
+
+            .. deprecated:: 2.0  This logic is deprecated and will be removed
+               in SQLAlchemy 2.0.     See :ref:`migration_20_query_distinct`
+               for a description of this use case in 2.0.
+
+        .. seealso::
+
+            :meth:`_sql.Select.distinct` - v2 equivalent method.
+
+        :param \*expr: optional column expressions.  When present,
+         the PostgreSQL dialect will render a ``DISTINCT ON (<expressions>)``
+         construct.
+
+         .. deprecated:: 1.4 Using \*expr in other dialects is deprecated
+            and will raise :class:`_exc.CompileError` in a future version.
+
+        """
+        if expr:
+            self._distinct = True
+            self._distinct_on = self._distinct_on + tuple(
+                coercions.expect(roles.ByOfRole, e) for e in expr
+            )
+        else:
+            self._distinct = True
+        return self
+
+    def all(self) -> List[_T]:
+        """Return the results represented by this :class:`_query.Query`
+        as a list.
+
+        This results in an execution of the underlying SQL statement.
+
+        .. warning::  The :class:`_query.Query` object,
+           when asked to return either
+           a sequence or iterator that consists of full ORM-mapped entities,
+           will **deduplicate entries based on primary key**.  See the FAQ for
+           more details.
+
+            .. seealso::
+
+                :ref:`faq_query_deduplicating`
+
+        .. seealso::
+
+            :meth:`_engine.Result.all` - v2 comparable method.
+
+            :meth:`_engine.Result.scalars` - v2 comparable method.
+        """
+        return self._iter().all()  # type: ignore
+
+    @_generative
+    @_assertions(_no_clauseelement_condition)
+    def from_statement(self, statement: ExecutableReturnsRows) -> Self:
+        """Execute the given SELECT statement and return results.
+
+        This method bypasses all internal statement compilation, and the
+        statement is executed without modification.
+
+        The statement is typically either a :func:`_expression.text`
+        or :func:`_expression.select` construct, and should return the set
+        of columns
+        appropriate to the entity class represented by this
+        :class:`_query.Query`.
+
+        .. seealso::
+
+            :meth:`_sql.Select.from_statement` - v2 comparable method.
+
+        """
+        statement = coercions.expect(
+            roles.SelectStatementRole, statement, apply_propagate_attrs=self
+        )
+        self._statement = statement
+        return self
+
+    def first(self) -> Optional[_T]:
+        """Return the first result of this ``Query`` or
+        None if the result doesn't contain any row.
+
+        first() applies a limit of one within the generated SQL, so that
+        only one primary entity row is generated on the server side
+        (note this may consist of multiple result rows if join-loaded
+        collections are present).
+
+        Calling :meth:`_query.Query.first`
+        results in an execution of the underlying
+        query.
+
+        .. seealso::
+
+            :meth:`_query.Query.one`
+
+            :meth:`_query.Query.one_or_none`
+
+            :meth:`_engine.Result.first` - v2 comparable method.
+
+            :meth:`_engine.Result.scalars` - v2 comparable method.
+
+        """
+        # replicates limit(1) behavior
+        if self._statement is not None:
+            return self._iter().first()  # type: ignore
+        else:
+            return self.limit(1)._iter().first()  # type: ignore
+
+    def one_or_none(self) -> Optional[_T]:
+        """Return at most one result or raise an exception.
+
+        Returns ``None`` if the query selects
+        no rows.  Raises ``sqlalchemy.orm.exc.MultipleResultsFound``
+        if multiple object identities are returned, or if multiple
+        rows are returned for a query that returns only scalar values
+        as opposed to full identity-mapped entities.
+
+        Calling :meth:`_query.Query.one_or_none`
+        results in an execution of the
+        underlying query.
+
+        .. seealso::
+
+            :meth:`_query.Query.first`
+
+            :meth:`_query.Query.one`
+
+            :meth:`_engine.Result.one_or_none` - v2 comparable method.
+
+            :meth:`_engine.Result.scalar_one_or_none` - v2 comparable method.
+
+        """
+        return self._iter().one_or_none()  # type: ignore
+
+    def one(self) -> _T:
+        """Return exactly one result or raise an exception.
+
+        Raises ``sqlalchemy.orm.exc.NoResultFound`` if the query selects
+        no rows.  Raises ``sqlalchemy.orm.exc.MultipleResultsFound``
+        if multiple object identities are returned, or if multiple
+        rows are returned for a query that returns only scalar values
+        as opposed to full identity-mapped entities.
+
+        Calling :meth:`.one` results in an execution of the underlying query.
+
+        .. seealso::
+
+            :meth:`_query.Query.first`
+
+            :meth:`_query.Query.one_or_none`
+
+            :meth:`_engine.Result.one` - v2 comparable method.
+
+            :meth:`_engine.Result.scalar_one` - v2 comparable method.
+
+        """
+        return self._iter().one()  # type: ignore
+
+    def scalar(self) -> Any:
+        """Return the first element of the first result or None
+        if no rows present.  If multiple rows are returned,
+        raises MultipleResultsFound.
+
+          >>> session.query(Item).scalar()
+          <Item>
+          >>> session.query(Item.id).scalar()
+          1
+          >>> session.query(Item.id).filter(Item.id < 0).scalar()
+          None
+          >>> session.query(Item.id, Item.name).scalar()
+          1
+          >>> session.query(func.count(Parent.id)).scalar()
+          20
+
+        This results in an execution of the underlying query.
+
+        .. seealso::
+
+            :meth:`_engine.Result.scalar` - v2 comparable method.
+
+        """
+        # TODO: not sure why we can't use result.scalar() here
+        try:
+            ret = self.one()
+            if not isinstance(ret, collections_abc.Sequence):
+                return ret
+            return ret[0]
+        except sa_exc.NoResultFound:
+            return None
+
+    def __iter__(self) -> Iterator[_T]:
+        result = self._iter()
+        try:
+            yield from result  # type: ignore
+        except GeneratorExit:
+            # issue #8710 - direct iteration is not re-usable after
+            # an iterable block is broken, so close the result
+            result._soft_close()
+            raise
+
+    def _iter(self) -> Union[ScalarResult[_T], Result[_T]]:
+        # new style execution.
+        params = self._params
+
+        statement = self._statement_20()
+        result: Union[ScalarResult[_T], Result[_T]] = self.session.execute(
+            statement,
+            params,
+            execution_options={"_sa_orm_load_options": self.load_options},
+        )
+
+        # legacy: automatically set scalars, unique
+        if result._attributes.get("is_single_entity", False):
+            result = cast("Result[_T]", result).scalars()
+
+        if (
+            result._attributes.get("filtered", False)
+            and not self.load_options._yield_per
+        ):
+            result = result.unique()
+
+        return result
+
+    def __str__(self) -> str:
+        statement = self._statement_20()
+
+        try:
+            bind = (
+                self._get_bind_args(statement, self.session.get_bind)
+                if self.session
+                else None
+            )
+        except sa_exc.UnboundExecutionError:
+            bind = None
+
+        return str(statement.compile(bind))
+
+    def _get_bind_args(self, statement: Any, fn: Any, **kw: Any) -> Any:
+        return fn(clause=statement, **kw)
+
+    @property
+    def column_descriptions(self) -> List[ORMColumnDescription]:
+        """Return metadata about the columns which would be
+        returned by this :class:`_query.Query`.
+
+        Format is a list of dictionaries::
+
+            user_alias = aliased(User, name="user2")
+            q = sess.query(User, User.id, user_alias)
+
+            # this expression:
+            q.column_descriptions
+
+            # would return:
+            [
+                {
+                    "name": "User",
+                    "type": User,
+                    "aliased": False,
+                    "expr": User,
+                    "entity": User,
+                },
+                {
+                    "name": "id",
+                    "type": Integer(),
+                    "aliased": False,
+                    "expr": User.id,
+                    "entity": User,
+                },
+                {
+                    "name": "user2",
+                    "type": User,
+                    "aliased": True,
+                    "expr": user_alias,
+                    "entity": user_alias,
+                },
+            ]
+
+        .. seealso::
+
+            This API is available using :term:`2.0 style` queries as well,
+            documented at:
+
+            * :ref:`queryguide_inspection`
+
+            * :attr:`.Select.column_descriptions`
+
+        """
+
+        return _column_descriptions(self, legacy=True)
+
+    @util.deprecated(
+        "2.0",
+        "The :meth:`_orm.Query.instances` method is deprecated and will "
+        "be removed in a future release. "
+        "Use the Select.from_statement() method or aliased() construct in "
+        "conjunction with Session.execute() instead.",
+    )
+    def instances(
+        self,
+        result_proxy: CursorResult[Any],
+        context: Optional[QueryContext] = None,
+    ) -> Any:
+        """Return an ORM result given a :class:`_engine.CursorResult` and
+        :class:`.QueryContext`.
+
+        """
+        if context is None:
+            util.warn_deprecated(
+                "Using the Query.instances() method without a context "
+                "is deprecated and will be disallowed in a future release.  "
+                "Please make use of :meth:`_query.Query.from_statement` "
+                "for linking ORM results to arbitrary select constructs.",
+                version="1.4",
+            )
+            compile_state = self._compile_state(for_statement=False)
+
+            context = QueryContext(
+                compile_state,
+                compile_state.statement,
+                compile_state.statement,
+                self._params,
+                self.session,
+                self.load_options,
+            )
+
+        result = loading.instances(result_proxy, context)
+
+        # legacy: automatically set scalars, unique
+        if result._attributes.get("is_single_entity", False):
+            result = result.scalars()  # type: ignore
+
+        if result._attributes.get("filtered", False):
+            result = result.unique()
+
+        # TODO: isn't this supposed to be a list?
+        return result
+
+    @util.became_legacy_20(
+        ":meth:`_orm.Query.merge_result`",
+        alternative="The method is superseded by the "
+        ":func:`_orm.merge_frozen_result` function.",
+        enable_warnings=False,  # warnings occur via loading.merge_result
+    )
+    def merge_result(
+        self,
+        iterator: Union[
+            FrozenResult[Any], Iterable[Sequence[Any]], Iterable[object]
+        ],
+        load: bool = True,
+    ) -> Union[FrozenResult[Any], Iterable[Any]]:
+        """Merge a result into this :class:`_query.Query` object's Session.
+
+        Given an iterator returned by a :class:`_query.Query`
+        of the same structure
+        as this one, return an identical iterator of results, with all mapped
+        instances merged into the session using :meth:`.Session.merge`. This
+        is an optimized method which will merge all mapped instances,
+        preserving the structure of the result rows and unmapped columns with
+        less method overhead than that of calling :meth:`.Session.merge`
+        explicitly for each value.
+
+        The structure of the results is determined based on the column list of
+        this :class:`_query.Query` - if these do not correspond,
+        unchecked errors
+        will occur.
+
+        The 'load' argument is the same as that of :meth:`.Session.merge`.
+
+        For an example of how :meth:`_query.Query.merge_result` is used, see
+        the source code for the example :ref:`examples_caching`, where
+        :meth:`_query.Query.merge_result` is used to efficiently restore state
+        from a cache back into a target :class:`.Session`.
+
+        """
+
+        return loading.merge_result(self, iterator, load)
+
+    def exists(self) -> Exists:
+        """A convenience method that turns a query into an EXISTS subquery
+        of the form EXISTS (SELECT 1 FROM ... WHERE ...).
+
+        e.g.::
+
+            q = session.query(User).filter(User.name == "fred")
+            session.query(q.exists())
+
+        Producing SQL similar to:
+
+        .. sourcecode:: sql
+
+            SELECT EXISTS (
+                SELECT 1 FROM users WHERE users.name = :name_1
+            ) AS anon_1
+
+        The EXISTS construct is usually used in the WHERE clause::
+
+            session.query(User.id).filter(q.exists()).scalar()
+
+        Note that some databases such as SQL Server don't allow an
+        EXISTS expression to be present in the columns clause of a
+        SELECT.    To select a simple boolean value based on the exists
+        as a WHERE, use :func:`.literal`::
+
+            from sqlalchemy import literal
+
+            session.query(literal(True)).filter(q.exists()).scalar()
+
+        .. seealso::
+
+            :meth:`_sql.Select.exists` - v2 comparable method.
+
+        """
+
+        # .add_columns() for the case that we are a query().select_from(X),
+        # so that ".statement" can be produced (#2995) but also without
+        # omitting the FROM clause from a query(X) (#2818);
+        # .with_only_columns() after we have a core select() so that
+        # we get just "SELECT 1" without any entities.
+
+        inner = (
+            self.enable_eagerloads(False)
+            .add_columns(sql.literal_column("1"))
+            .set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+            ._get_select_statement_only()
+            .with_only_columns(1)
+        )
+
+        ezero = self._entity_from_pre_ent_zero()
+        if ezero is not None:
+            inner = inner.select_from(ezero)
+
+        return sql.exists(inner)
+
+    def count(self) -> int:
+        r"""Return a count of rows this the SQL formed by this :class:`Query`
+        would return.
+
+        This generates the SQL for this Query as follows:
+
+        .. sourcecode:: sql
+
+            SELECT count(1) AS count_1 FROM (
+                SELECT <rest of query follows...>
+            ) AS anon_1
+
+        The above SQL returns a single row, which is the aggregate value
+        of the count function; the :meth:`_query.Query.count`
+        method then returns
+        that single integer value.
+
+        .. warning::
+
+            It is important to note that the value returned by
+            count() is **not the same as the number of ORM objects that this
+            Query would return from a method such as the .all() method**.
+            The :class:`_query.Query` object,
+            when asked to return full entities,
+            will **deduplicate entries based on primary key**, meaning if the
+            same primary key value would appear in the results more than once,
+            only one object of that primary key would be present.  This does
+            not apply to a query that is against individual columns.
+
+            .. seealso::
+
+                :ref:`faq_query_deduplicating`
+
+        For fine grained control over specific columns to count, to skip the
+        usage of a subquery or otherwise control of the FROM clause, or to use
+        other aggregate functions, use :attr:`~sqlalchemy.sql.expression.func`
+        expressions in conjunction with :meth:`~.Session.query`, i.e.::
+
+            from sqlalchemy import func
+
+            # count User records, without
+            # using a subquery.
+            session.query(func.count(User.id))
+
+            # return count of user "id" grouped
+            # by "name"
+            session.query(func.count(User.id)).group_by(User.name)
+
+            from sqlalchemy import distinct
+
+            # count distinct "name" values
+            session.query(func.count(distinct(User.name)))
+
+        .. seealso::
+
+            :ref:`migration_20_query_usage`
+
+        """
+        col = sql.func.count(sql.literal_column("*"))
+        return (  # type: ignore
+            self._legacy_from_self(col).enable_eagerloads(False).scalar()
+        )
+
+    def delete(
+        self,
+        synchronize_session: SynchronizeSessionArgument = "auto",
+        delete_args: Optional[Dict[Any, Any]] = None,
+    ) -> int:
+        r"""Perform a DELETE with an arbitrary WHERE clause.
+
+        Deletes rows matched by this query from the database.
+
+        E.g.::
+
+            sess.query(User).filter(User.age == 25).delete(synchronize_session=False)
+
+            sess.query(User).filter(User.age == 25).delete(
+                synchronize_session="evaluate"
+            )
+
+        .. warning::
+
+            See the section :ref:`orm_expression_update_delete` for important
+            caveats and warnings, including limitations when using bulk UPDATE
+            and DELETE with mapper inheritance configurations.
+
+        :param synchronize_session: chooses the strategy to update the
+         attributes on objects in the session.   See the section
+         :ref:`orm_expression_update_delete` for a discussion of these
+         strategies.
+
+        :param delete_args: Optional dictionary, if present will be passed
+         to the underlying :func:`_expression.delete` construct as the ``**kw``
+         for the object.  May be used to pass dialect-specific arguments such
+         as ``mysql_limit``.
+
+         .. versionadded:: 2.0.37
+
+        :return: the count of rows matched as returned by the database's
+          "row count" feature.
+
+        .. seealso::
+
+            :ref:`orm_expression_update_delete`
+
+        """  # noqa: E501
+
+        bulk_del = BulkDelete(self, delete_args)
+        if self.dispatch.before_compile_delete:
+            for fn in self.dispatch.before_compile_delete:
+                new_query = fn(bulk_del.query, bulk_del)
+                if new_query is not None:
+                    bulk_del.query = new_query
+
+                self = bulk_del.query
+
+        delete_ = sql.delete(*self._raw_columns)  # type: ignore
+
+        if delete_args:
+            delete_ = delete_.with_dialect_options(**delete_args)
+
+        delete_._where_criteria = self._where_criteria
+        result: CursorResult[Any] = self.session.execute(
+            delete_,
+            self._params,
+            execution_options=self._execution_options.union(
+                {"synchronize_session": synchronize_session}
+            ),
+        )
+        bulk_del.result = result  # type: ignore
+        self.session.dispatch.after_bulk_delete(bulk_del)
+        result.close()
+
+        return result.rowcount
+
+    def update(
+        self,
+        values: Dict[_DMLColumnArgument, Any],
+        synchronize_session: SynchronizeSessionArgument = "auto",
+        update_args: Optional[Dict[Any, Any]] = None,
+    ) -> int:
+        r"""Perform an UPDATE with an arbitrary WHERE clause.
+
+        Updates rows matched by this query in the database.
+
+        E.g.::
+
+            sess.query(User).filter(User.age == 25).update(
+                {User.age: User.age - 10}, synchronize_session=False
+            )
+
+            sess.query(User).filter(User.age == 25).update(
+                {"age": User.age - 10}, synchronize_session="evaluate"
+            )
+
+        .. warning::
+
+            See the section :ref:`orm_expression_update_delete` for important
+            caveats and warnings, including limitations when using arbitrary
+            UPDATE and DELETE with mapper inheritance configurations.
+
+        :param values: a dictionary with attributes names, or alternatively
+         mapped attributes or SQL expressions, as keys, and literal
+         values or sql expressions as values.   If :ref:`parameter-ordered
+         mode <tutorial_parameter_ordered_updates>` is desired, the values can
+         be passed as a list of 2-tuples; this requires that the
+         :paramref:`~sqlalchemy.sql.expression.update.preserve_parameter_order`
+         flag is passed to the :paramref:`.Query.update.update_args` dictionary
+         as well.
+
+        :param synchronize_session: chooses the strategy to update the
+         attributes on objects in the session.   See the section
+         :ref:`orm_expression_update_delete` for a discussion of these
+         strategies.
+
+        :param update_args: Optional dictionary, if present will be passed
+         to the underlying :func:`_expression.update` construct as the ``**kw``
+         for the object.  May be used to pass dialect-specific arguments such
+         as ``mysql_limit``, as well as other special arguments such as
+         :paramref:`~sqlalchemy.sql.expression.update.preserve_parameter_order`.
+
+        :return: the count of rows matched as returned by the database's
+         "row count" feature.
+
+
+        .. seealso::
+
+            :ref:`orm_expression_update_delete`
+
+        """
+
+        update_args = update_args or {}
+
+        bulk_ud = BulkUpdate(self, values, update_args)
+
+        if self.dispatch.before_compile_update:
+            for fn in self.dispatch.before_compile_update:
+                new_query = fn(bulk_ud.query, bulk_ud)
+                if new_query is not None:
+                    bulk_ud.query = new_query
+            self = bulk_ud.query
+
+        upd = sql.update(*self._raw_columns)  # type: ignore
+
+        ppo = update_args.pop("preserve_parameter_order", False)
+        if ppo:
+            upd = upd.ordered_values(*values)  # type: ignore
+        else:
+            upd = upd.values(values)
+        if update_args:
+            upd = upd.with_dialect_options(**update_args)
+
+        upd._where_criteria = self._where_criteria
+        result: CursorResult[Any] = self.session.execute(
+            upd,
+            self._params,
+            execution_options=self._execution_options.union(
+                {"synchronize_session": synchronize_session}
+            ),
+        )
+        bulk_ud.result = result  # type: ignore
+        self.session.dispatch.after_bulk_update(bulk_ud)
+        result.close()
+        return result.rowcount
+
+    def _compile_state(
+        self, for_statement: bool = False, **kw: Any
+    ) -> ORMCompileState:
+        """Create an out-of-compiler ORMCompileState object.
+
+        The ORMCompileState object is normally created directly as a result
+        of the SQLCompiler.process() method being handed a Select()
+        or FromStatement() object that uses the "orm" plugin.   This method
+        provides a means of creating this ORMCompileState object directly
+        without using the compiler.
+
+        This method is used only for deprecated cases, which include
+        the .from_self() method for a Query that has multiple levels
+        of .from_self() in use, as well as the instances() method.  It is
+        also used within the test suite to generate ORMCompileState objects
+        for test purposes.
+
+        """
+
+        stmt = self._statement_20(for_statement=for_statement, **kw)
+        assert for_statement == stmt._compile_options._for_statement
+
+        # this chooses between ORMFromStatementCompileState and
+        # ORMSelectCompileState.  We could also base this on
+        # query._statement is not None as we have the ORM Query here
+        # however this is the more general path.
+        compile_state_cls = cast(
+            ORMCompileState,
+            ORMCompileState._get_plugin_class_for_plugin(stmt, "orm"),
+        )
+
+        return compile_state_cls.create_for_statement(stmt, None)
+
+    def _compile_context(self, for_statement: bool = False) -> QueryContext:
+        compile_state = self._compile_state(for_statement=for_statement)
+        context = QueryContext(
+            compile_state,
+            compile_state.statement,
+            compile_state.statement,
+            self._params,
+            self.session,
+            self.load_options,
+        )
+
+        return context
+
+
+class AliasOption(interfaces.LoaderOption):
+    inherit_cache = False
+
+    @util.deprecated(
+        "1.4",
+        "The :class:`.AliasOption` object is not necessary "
+        "for entities to be matched up to a query that is established "
+        "via :meth:`.Query.from_statement` and now does nothing.",
+    )
+    def __init__(self, alias: Union[Alias, Subquery]):
+        r"""Return a :class:`.MapperOption` that will indicate to the
+        :class:`_query.Query`
+        that the main table has been aliased.
+
+        """
+
+    def process_compile_state(self, compile_state: ORMCompileState) -> None:
+        pass
+
+
+class BulkUD:
+    """State used for the orm.Query version of update() / delete().
+
+    This object is now specific to Query only.
+
+    """
+
+    def __init__(self, query: Query[Any]):
+        self.query = query.enable_eagerloads(False)
+        self._validate_query_state()
+        self.mapper = self.query._entity_from_pre_ent_zero()
+
+    def _validate_query_state(self) -> None:
+        for attr, methname, notset, op in (
+            ("_limit_clause", "limit()", None, operator.is_),
+            ("_offset_clause", "offset()", None, operator.is_),
+            ("_order_by_clauses", "order_by()", (), operator.eq),
+            ("_group_by_clauses", "group_by()", (), operator.eq),
+            ("_distinct", "distinct()", False, operator.is_),
+            (
+                "_from_obj",
+                "join(), outerjoin(), select_from(), or from_self()",
+                (),
+                operator.eq,
+            ),
+            (
+                "_setup_joins",
+                "join(), outerjoin(), select_from(), or from_self()",
+                (),
+                operator.eq,
+            ),
+        ):
+            if not op(getattr(self.query, attr), notset):
+                raise sa_exc.InvalidRequestError(
+                    "Can't call Query.update() or Query.delete() "
+                    "when %s has been called" % (methname,)
+                )
+
+    @property
+    def session(self) -> Session:
+        return self.query.session
+
+
+class BulkUpdate(BulkUD):
+    """BulkUD which handles UPDATEs."""
+
+    def __init__(
+        self,
+        query: Query[Any],
+        values: Dict[_DMLColumnArgument, Any],
+        update_kwargs: Optional[Dict[Any, Any]],
+    ):
+        super().__init__(query)
+        self.values = values
+        self.update_kwargs = update_kwargs
+
+
+class BulkDelete(BulkUD):
+    """BulkUD which handles DELETEs."""
+
+    def __init__(
+        self,
+        query: Query[Any],
+        delete_kwargs: Optional[Dict[Any, Any]],
+    ):
+        super().__init__(query)
+        self.delete_kwargs = delete_kwargs
+
+
+class RowReturningQuery(Query[Row[_TP]]):
+    if TYPE_CHECKING:
+
+        def tuples(self) -> Query[_TP]:  # type: ignore
+            ...
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/relationships.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/relationships.py
new file mode 100644
index 00000000..0d0bc708
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/relationships.py
@@ -0,0 +1,3514 @@
+# orm/relationships.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Heuristics related to join conditions as used in
+:func:`_orm.relationship`.
+
+Provides the :class:`.JoinCondition` object, which encapsulates
+SQL annotation and aliasing behavior focused on the `primaryjoin`
+and `secondaryjoin` aspects of :func:`_orm.relationship`.
+
+"""
+from __future__ import annotations
+
+import collections
+from collections import abc
+import dataclasses
+import inspect as _py_inspect
+import itertools
+import re
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import strategy_options
+from ._typing import insp_is_aliased_class
+from ._typing import is_has_collection_adapter
+from .base import _DeclarativeMapped
+from .base import _is_mapped_class
+from .base import class_mapper
+from .base import DynamicMapped
+from .base import LoaderCallableStatus
+from .base import PassiveFlag
+from .base import state_str
+from .base import WriteOnlyMapped
+from .interfaces import _AttributeOptions
+from .interfaces import _IntrospectsAnnotations
+from .interfaces import MANYTOMANY
+from .interfaces import MANYTOONE
+from .interfaces import ONETOMANY
+from .interfaces import PropComparator
+from .interfaces import RelationshipDirection
+from .interfaces import StrategizedProperty
+from .util import _orm_annotate
+from .util import _orm_deannotate
+from .util import CascadeOptions
+from .. import exc as sa_exc
+from .. import Exists
+from .. import log
+from .. import schema
+from .. import sql
+from .. import util
+from ..inspection import inspect
+from ..sql import coercions
+from ..sql import expression
+from ..sql import operators
+from ..sql import roles
+from ..sql import visitors
+from ..sql._typing import _ColumnExpressionArgument
+from ..sql._typing import _HasClauseElement
+from ..sql.annotation import _safe_annotate
+from ..sql.elements import ColumnClause
+from ..sql.elements import ColumnElement
+from ..sql.util import _deep_annotate
+from ..sql.util import _deep_deannotate
+from ..sql.util import _shallow_annotate
+from ..sql.util import adapt_criterion_to_null
+from ..sql.util import ClauseAdapter
+from ..sql.util import join_condition
+from ..sql.util import selectables_overlap
+from ..sql.util import visit_binary_product
+from ..util.typing import de_optionalize_union_types
+from ..util.typing import Literal
+from ..util.typing import resolve_name_to_real_class_name
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _ExternalEntityType
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import _InternalEntityType
+    from ._typing import _O
+    from ._typing import _RegistryType
+    from .base import Mapped
+    from .clsregistry import _class_resolver
+    from .clsregistry import _ModNS
+    from .decl_base import _ClassScanMapperConfig
+    from .dependency import DependencyProcessor
+    from .mapper import Mapper
+    from .query import Query
+    from .session import Session
+    from .state import InstanceState
+    from .strategies import LazyLoader
+    from .util import AliasedClass
+    from .util import AliasedInsp
+    from ..sql._typing import _CoreAdapterProto
+    from ..sql._typing import _EquivalentColumnMap
+    from ..sql._typing import _InfoType
+    from ..sql.annotation import _AnnotationDict
+    from ..sql.annotation import SupportsAnnotations
+    from ..sql.elements import BinaryExpression
+    from ..sql.elements import BindParameter
+    from ..sql.elements import ClauseElement
+    from ..sql.schema import Table
+    from ..sql.selectable import FromClause
+    from ..util.typing import _AnnotationScanType
+    from ..util.typing import RODescriptorReference
+
+_T = TypeVar("_T", bound=Any)
+_T1 = TypeVar("_T1", bound=Any)
+_T2 = TypeVar("_T2", bound=Any)
+
+_PT = TypeVar("_PT", bound=Any)
+
+_PT2 = TypeVar("_PT2", bound=Any)
+
+
+_RelationshipArgumentType = Union[
+    str,
+    Type[_T],
+    Callable[[], Type[_T]],
+    "Mapper[_T]",
+    "AliasedClass[_T]",
+    Callable[[], "Mapper[_T]"],
+    Callable[[], "AliasedClass[_T]"],
+]
+
+_LazyLoadArgumentType = Literal[
+    "select",
+    "joined",
+    "selectin",
+    "subquery",
+    "raise",
+    "raise_on_sql",
+    "noload",
+    "immediate",
+    "write_only",
+    "dynamic",
+    True,
+    False,
+    None,
+]
+
+
+_RelationshipJoinConditionArgument = Union[
+    str, _ColumnExpressionArgument[bool]
+]
+_RelationshipSecondaryArgument = Union[
+    "FromClause", str, Callable[[], "FromClause"]
+]
+_ORMOrderByArgument = Union[
+    Literal[False],
+    str,
+    _ColumnExpressionArgument[Any],
+    Callable[[], _ColumnExpressionArgument[Any]],
+    Callable[[], Iterable[_ColumnExpressionArgument[Any]]],
+    Iterable[Union[str, _ColumnExpressionArgument[Any]]],
+]
+ORMBackrefArgument = Union[str, Tuple[str, Dict[str, Any]]]
+
+_ORMColCollectionElement = Union[
+    ColumnClause[Any],
+    _HasClauseElement[Any],
+    roles.DMLColumnRole,
+    "Mapped[Any]",
+]
+_ORMColCollectionArgument = Union[
+    str,
+    Sequence[_ORMColCollectionElement],
+    Callable[[], Sequence[_ORMColCollectionElement]],
+    Callable[[], _ORMColCollectionElement],
+    _ORMColCollectionElement,
+]
+
+
+_CEA = TypeVar("_CEA", bound=_ColumnExpressionArgument[Any])
+
+_CE = TypeVar("_CE", bound="ColumnElement[Any]")
+
+
+_ColumnPairIterable = Iterable[Tuple[ColumnElement[Any], ColumnElement[Any]]]
+
+_ColumnPairs = Sequence[Tuple[ColumnElement[Any], ColumnElement[Any]]]
+
+_MutableColumnPairs = List[Tuple[ColumnElement[Any], ColumnElement[Any]]]
+
+
+def remote(expr: _CEA) -> _CEA:
+    """Annotate a portion of a primaryjoin expression
+    with a 'remote' annotation.
+
+    See the section :ref:`relationship_custom_foreign` for a
+    description of use.
+
+    .. seealso::
+
+        :ref:`relationship_custom_foreign`
+
+        :func:`.foreign`
+
+    """
+    return _annotate_columns(  # type: ignore
+        coercions.expect(roles.ColumnArgumentRole, expr), {"remote": True}
+    )
+
+
+def foreign(expr: _CEA) -> _CEA:
+    """Annotate a portion of a primaryjoin expression
+    with a 'foreign' annotation.
+
+    See the section :ref:`relationship_custom_foreign` for a
+    description of use.
+
+    .. seealso::
+
+        :ref:`relationship_custom_foreign`
+
+        :func:`.remote`
+
+    """
+
+    return _annotate_columns(  # type: ignore
+        coercions.expect(roles.ColumnArgumentRole, expr), {"foreign": True}
+    )
+
+
+@dataclasses.dataclass
+class _RelationshipArg(Generic[_T1, _T2]):
+    """stores a user-defined parameter value that must be resolved and
+    parsed later at mapper configuration time.
+
+    """
+
+    __slots__ = "name", "argument", "resolved"
+    name: str
+    argument: _T1
+    resolved: Optional[_T2]
+
+    def _is_populated(self) -> bool:
+        return self.argument is not None
+
+    def _resolve_against_registry(
+        self, clsregistry_resolver: Callable[[str, bool], _class_resolver]
+    ) -> None:
+        attr_value = self.argument
+
+        if isinstance(attr_value, str):
+            self.resolved = clsregistry_resolver(
+                attr_value, self.name == "secondary"
+            )()
+        elif callable(attr_value) and not _is_mapped_class(attr_value):
+            self.resolved = attr_value()
+        else:
+            self.resolved = attr_value
+
+
+_RelationshipOrderByArg = Union[Literal[False], Tuple[ColumnElement[Any], ...]]
+
+
+class _RelationshipArgs(NamedTuple):
+    """stores user-passed parameters that are resolved at mapper configuration
+    time.
+
+    """
+
+    secondary: _RelationshipArg[
+        Optional[_RelationshipSecondaryArgument],
+        Optional[FromClause],
+    ]
+    primaryjoin: _RelationshipArg[
+        Optional[_RelationshipJoinConditionArgument],
+        Optional[ColumnElement[Any]],
+    ]
+    secondaryjoin: _RelationshipArg[
+        Optional[_RelationshipJoinConditionArgument],
+        Optional[ColumnElement[Any]],
+    ]
+    order_by: _RelationshipArg[_ORMOrderByArgument, _RelationshipOrderByArg]
+    foreign_keys: _RelationshipArg[
+        Optional[_ORMColCollectionArgument], Set[ColumnElement[Any]]
+    ]
+    remote_side: _RelationshipArg[
+        Optional[_ORMColCollectionArgument], Set[ColumnElement[Any]]
+    ]
+
+
+@log.class_logger
+class RelationshipProperty(
+    _IntrospectsAnnotations, StrategizedProperty[_T], log.Identified
+):
+    """Describes an object property that holds a single item or list
+    of items that correspond to a related database table.
+
+    Public constructor is the :func:`_orm.relationship` function.
+
+    .. seealso::
+
+        :ref:`relationship_config_toplevel`
+
+    """
+
+    strategy_wildcard_key = strategy_options._RELATIONSHIP_TOKEN
+    inherit_cache = True
+    """:meta private:"""
+
+    _links_to_entity = True
+    _is_relationship = True
+
+    _overlaps: Sequence[str]
+
+    _lazy_strategy: LazyLoader
+
+    _persistence_only = dict(
+        passive_deletes=False,
+        passive_updates=True,
+        enable_typechecks=True,
+        active_history=False,
+        cascade_backrefs=False,
+    )
+
+    _dependency_processor: Optional[DependencyProcessor] = None
+
+    primaryjoin: ColumnElement[bool]
+    secondaryjoin: Optional[ColumnElement[bool]]
+    secondary: Optional[FromClause]
+    _join_condition: JoinCondition
+    order_by: _RelationshipOrderByArg
+
+    _user_defined_foreign_keys: Set[ColumnElement[Any]]
+    _calculated_foreign_keys: Set[ColumnElement[Any]]
+
+    remote_side: Set[ColumnElement[Any]]
+    local_columns: Set[ColumnElement[Any]]
+
+    synchronize_pairs: _ColumnPairs
+    secondary_synchronize_pairs: Optional[_ColumnPairs]
+
+    local_remote_pairs: Optional[_ColumnPairs]
+
+    direction: RelationshipDirection
+
+    _init_args: _RelationshipArgs
+
+    def __init__(
+        self,
+        argument: Optional[_RelationshipArgumentType[_T]] = None,
+        secondary: Optional[_RelationshipSecondaryArgument] = None,
+        *,
+        uselist: Optional[bool] = None,
+        collection_class: Optional[
+            Union[Type[Collection[Any]], Callable[[], Collection[Any]]]
+        ] = None,
+        primaryjoin: Optional[_RelationshipJoinConditionArgument] = None,
+        secondaryjoin: Optional[_RelationshipJoinConditionArgument] = None,
+        back_populates: Optional[str] = None,
+        order_by: _ORMOrderByArgument = False,
+        backref: Optional[ORMBackrefArgument] = None,
+        overlaps: Optional[str] = None,
+        post_update: bool = False,
+        cascade: str = "save-update, merge",
+        viewonly: bool = False,
+        attribute_options: Optional[_AttributeOptions] = None,
+        lazy: _LazyLoadArgumentType = "select",
+        passive_deletes: Union[Literal["all"], bool] = False,
+        passive_updates: bool = True,
+        active_history: bool = False,
+        enable_typechecks: bool = True,
+        foreign_keys: Optional[_ORMColCollectionArgument] = None,
+        remote_side: Optional[_ORMColCollectionArgument] = None,
+        join_depth: Optional[int] = None,
+        comparator_factory: Optional[
+            Type[RelationshipProperty.Comparator[Any]]
+        ] = None,
+        single_parent: bool = False,
+        innerjoin: bool = False,
+        distinct_target_key: Optional[bool] = None,
+        load_on_pending: bool = False,
+        query_class: Optional[Type[Query[Any]]] = None,
+        info: Optional[_InfoType] = None,
+        omit_join: Literal[None, False] = None,
+        sync_backref: Optional[bool] = None,
+        doc: Optional[str] = None,
+        bake_queries: Literal[True] = True,
+        cascade_backrefs: Literal[False] = False,
+        _local_remote_pairs: Optional[_ColumnPairs] = None,
+        _legacy_inactive_history_style: bool = False,
+    ):
+        super().__init__(attribute_options=attribute_options)
+
+        self.uselist = uselist
+        self.argument = argument
+
+        self._init_args = _RelationshipArgs(
+            _RelationshipArg("secondary", secondary, None),
+            _RelationshipArg("primaryjoin", primaryjoin, None),
+            _RelationshipArg("secondaryjoin", secondaryjoin, None),
+            _RelationshipArg("order_by", order_by, None),
+            _RelationshipArg("foreign_keys", foreign_keys, None),
+            _RelationshipArg("remote_side", remote_side, None),
+        )
+
+        self.post_update = post_update
+        self.viewonly = viewonly
+        if viewonly:
+            self._warn_for_persistence_only_flags(
+                passive_deletes=passive_deletes,
+                passive_updates=passive_updates,
+                enable_typechecks=enable_typechecks,
+                active_history=active_history,
+                cascade_backrefs=cascade_backrefs,
+            )
+        if viewonly and sync_backref:
+            raise sa_exc.ArgumentError(
+                "sync_backref and viewonly cannot both be True"
+            )
+        self.sync_backref = sync_backref
+        self.lazy = lazy
+        self.single_parent = single_parent
+        self.collection_class = collection_class
+        self.passive_deletes = passive_deletes
+
+        if cascade_backrefs:
+            raise sa_exc.ArgumentError(
+                "The 'cascade_backrefs' parameter passed to "
+                "relationship() may only be set to False."
+            )
+
+        self.passive_updates = passive_updates
+        self.enable_typechecks = enable_typechecks
+        self.query_class = query_class
+        self.innerjoin = innerjoin
+        self.distinct_target_key = distinct_target_key
+        self.doc = doc
+        self.active_history = active_history
+        self._legacy_inactive_history_style = _legacy_inactive_history_style
+
+        self.join_depth = join_depth
+        if omit_join:
+            util.warn(
+                "setting omit_join to True is not supported; selectin "
+                "loading of this relationship may not work correctly if this "
+                "flag is set explicitly.  omit_join optimization is "
+                "automatically detected for conditions under which it is "
+                "supported."
+            )
+
+        self.omit_join = omit_join
+        self.local_remote_pairs = _local_remote_pairs
+        self.load_on_pending = load_on_pending
+        self.comparator_factory = (
+            comparator_factory or RelationshipProperty.Comparator
+        )
+        util.set_creation_order(self)
+
+        if info is not None:
+            self.info.update(info)
+
+        self.strategy_key = (("lazy", self.lazy),)
+
+        self._reverse_property: Set[RelationshipProperty[Any]] = set()
+
+        if overlaps:
+            self._overlaps = set(re.split(r"\s*,\s*", overlaps))  # type: ignore  # noqa: E501
+        else:
+            self._overlaps = ()
+
+        # mypy ignoring the @property setter
+        self.cascade = cascade  # type: ignore
+
+        self.back_populates = back_populates
+
+        if self.back_populates:
+            if backref:
+                raise sa_exc.ArgumentError(
+                    "backref and back_populates keyword arguments "
+                    "are mutually exclusive"
+                )
+            self.backref = None
+        else:
+            self.backref = backref
+
+    def _warn_for_persistence_only_flags(self, **kw: Any) -> None:
+        for k, v in kw.items():
+            if v != self._persistence_only[k]:
+                # we are warning here rather than warn deprecated as this is a
+                # configuration mistake, and Python shows regular warnings more
+                # aggressively than deprecation warnings by default. Unlike the
+                # case of setting viewonly with cascade, the settings being
+                # warned about here are not actively doing the wrong thing
+                # against viewonly=True, so it is not as urgent to have these
+                # raise an error.
+                util.warn(
+                    "Setting %s on relationship() while also "
+                    "setting viewonly=True does not make sense, as a "
+                    "viewonly=True relationship does not perform persistence "
+                    "operations. This configuration may raise an error "
+                    "in a future release." % (k,)
+                )
+
+    def instrument_class(self, mapper: Mapper[Any]) -> None:
+        attributes.register_descriptor(
+            mapper.class_,
+            self.key,
+            comparator=self.comparator_factory(self, mapper),
+            parententity=mapper,
+            doc=self.doc,
+        )
+
+    class Comparator(util.MemoizedSlots, PropComparator[_PT]):
+        """Produce boolean, comparison, and other operators for
+        :class:`.RelationshipProperty` attributes.
+
+        See the documentation for :class:`.PropComparator` for a brief
+        overview of ORM level operator definition.
+
+        .. seealso::
+
+            :class:`.PropComparator`
+
+            :class:`.ColumnProperty.Comparator`
+
+            :class:`.ColumnOperators`
+
+            :ref:`types_operators`
+
+            :attr:`.TypeEngine.comparator_factory`
+
+        """
+
+        __slots__ = (
+            "entity",
+            "mapper",
+            "property",
+            "_of_type",
+            "_extra_criteria",
+        )
+
+        prop: RODescriptorReference[RelationshipProperty[_PT]]
+        _of_type: Optional[_EntityType[_PT]]
+
+        def __init__(
+            self,
+            prop: RelationshipProperty[_PT],
+            parentmapper: _InternalEntityType[Any],
+            adapt_to_entity: Optional[AliasedInsp[Any]] = None,
+            of_type: Optional[_EntityType[_PT]] = None,
+            extra_criteria: Tuple[ColumnElement[bool], ...] = (),
+        ):
+            """Construction of :class:`.RelationshipProperty.Comparator`
+            is internal to the ORM's attribute mechanics.
+
+            """
+            self.prop = prop
+            self._parententity = parentmapper
+            self._adapt_to_entity = adapt_to_entity
+            if of_type:
+                self._of_type = of_type
+            else:
+                self._of_type = None
+            self._extra_criteria = extra_criteria
+
+        def adapt_to_entity(
+            self, adapt_to_entity: AliasedInsp[Any]
+        ) -> RelationshipProperty.Comparator[Any]:
+            return self.__class__(
+                self.prop,
+                self._parententity,
+                adapt_to_entity=adapt_to_entity,
+                of_type=self._of_type,
+            )
+
+        entity: _InternalEntityType[_PT]
+        """The target entity referred to by this
+        :class:`.RelationshipProperty.Comparator`.
+
+        This is either a :class:`_orm.Mapper` or :class:`.AliasedInsp`
+        object.
+
+        This is the "target" or "remote" side of the
+        :func:`_orm.relationship`.
+
+        """
+
+        mapper: Mapper[_PT]
+        """The target :class:`_orm.Mapper` referred to by this
+        :class:`.RelationshipProperty.Comparator`.
+
+        This is the "target" or "remote" side of the
+        :func:`_orm.relationship`.
+
+        """
+
+        def _memoized_attr_entity(self) -> _InternalEntityType[_PT]:
+            if self._of_type:
+                return inspect(self._of_type)  # type: ignore
+            else:
+                return self.prop.entity
+
+        def _memoized_attr_mapper(self) -> Mapper[_PT]:
+            return self.entity.mapper
+
+        def _source_selectable(self) -> FromClause:
+            if self._adapt_to_entity:
+                return self._adapt_to_entity.selectable
+            else:
+                return self.property.parent._with_polymorphic_selectable
+
+        def __clause_element__(self) -> ColumnElement[bool]:
+            adapt_from = self._source_selectable()
+            if self._of_type:
+                of_type_entity = inspect(self._of_type)
+            else:
+                of_type_entity = None
+
+            (
+                pj,
+                sj,
+                source,
+                dest,
+                secondary,
+                target_adapter,
+            ) = self.prop._create_joins(
+                source_selectable=adapt_from,
+                source_polymorphic=True,
+                of_type_entity=of_type_entity,
+                alias_secondary=True,
+                extra_criteria=self._extra_criteria,
+            )
+            if sj is not None:
+                return pj & sj
+            else:
+                return pj
+
+        def of_type(self, class_: _EntityType[Any]) -> PropComparator[_PT]:
+            r"""Redefine this object in terms of a polymorphic subclass.
+
+            See :meth:`.PropComparator.of_type` for an example.
+
+
+            """
+            return RelationshipProperty.Comparator(
+                self.prop,
+                self._parententity,
+                adapt_to_entity=self._adapt_to_entity,
+                of_type=class_,
+                extra_criteria=self._extra_criteria,
+            )
+
+        def and_(
+            self, *criteria: _ColumnExpressionArgument[bool]
+        ) -> PropComparator[Any]:
+            """Add AND criteria.
+
+            See :meth:`.PropComparator.and_` for an example.
+
+            .. versionadded:: 1.4
+
+            """
+            exprs = tuple(
+                coercions.expect(roles.WhereHavingRole, clause)
+                for clause in util.coerce_generator_arg(criteria)
+            )
+
+            return RelationshipProperty.Comparator(
+                self.prop,
+                self._parententity,
+                adapt_to_entity=self._adapt_to_entity,
+                of_type=self._of_type,
+                extra_criteria=self._extra_criteria + exprs,
+            )
+
+        def in_(self, other: Any) -> NoReturn:
+            """Produce an IN clause - this is not implemented
+            for :func:`_orm.relationship`-based attributes at this time.
+
+            """
+            raise NotImplementedError(
+                "in_() not yet supported for "
+                "relationships.  For a simple "
+                "many-to-one, use in_() against "
+                "the set of foreign key values."
+            )
+
+        # https://github.com/python/mypy/issues/4266
+        __hash__ = None  # type: ignore
+
+        def __eq__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            """Implement the ``==`` operator.
+
+            In a many-to-one context, such as:
+
+            .. sourcecode:: text
+
+              MyClass.some_prop == <some object>
+
+            this will typically produce a
+            clause such as:
+
+            .. sourcecode:: text
+
+              mytable.related_id == <some id>
+
+            Where ``<some id>`` is the primary key of the given
+            object.
+
+            The ``==`` operator provides partial functionality for non-
+            many-to-one comparisons:
+
+            * Comparisons against collections are not supported.
+              Use :meth:`~.Relationship.Comparator.contains`.
+            * Compared to a scalar one-to-many, will produce a
+              clause that compares the target columns in the parent to
+              the given target.
+            * Compared to a scalar many-to-many, an alias
+              of the association table will be rendered as
+              well, forming a natural join that is part of the
+              main body of the query. This will not work for
+              queries that go beyond simple AND conjunctions of
+              comparisons, such as those which use OR. Use
+              explicit joins, outerjoins, or
+              :meth:`~.Relationship.Comparator.has` for
+              more comprehensive non-many-to-one scalar
+              membership tests.
+            * Comparisons against ``None`` given in a one-to-many
+              or many-to-many context produce a NOT EXISTS clause.
+
+            """
+            if other is None or isinstance(other, expression.Null):
+                if self.property.direction in [ONETOMANY, MANYTOMANY]:
+                    return ~self._criterion_exists()
+                else:
+                    return _orm_annotate(
+                        self.property._optimized_compare(
+                            None, adapt_source=self.adapter
+                        )
+                    )
+            elif self.property.uselist:
+                raise sa_exc.InvalidRequestError(
+                    "Can't compare a collection to an object or collection; "
+                    "use contains() to test for membership."
+                )
+            else:
+                return _orm_annotate(
+                    self.property._optimized_compare(
+                        other, adapt_source=self.adapter
+                    )
+                )
+
+        def _criterion_exists(
+            self,
+            criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+            **kwargs: Any,
+        ) -> Exists:
+            where_criteria = (
+                coercions.expect(roles.WhereHavingRole, criterion)
+                if criterion is not None
+                else None
+            )
+
+            if getattr(self, "_of_type", None):
+                info: Optional[_InternalEntityType[Any]] = inspect(
+                    self._of_type
+                )
+                assert info is not None
+                target_mapper, to_selectable, is_aliased_class = (
+                    info.mapper,
+                    info.selectable,
+                    info.is_aliased_class,
+                )
+                if self.property._is_self_referential and not is_aliased_class:
+                    to_selectable = to_selectable._anonymous_fromclause()
+
+                single_crit = target_mapper._single_table_criterion
+                if single_crit is not None:
+                    if where_criteria is not None:
+                        where_criteria = single_crit & where_criteria
+                    else:
+                        where_criteria = single_crit
+            else:
+                is_aliased_class = False
+                to_selectable = None
+
+            if self.adapter:
+                source_selectable = self._source_selectable()
+            else:
+                source_selectable = None
+
+            (
+                pj,
+                sj,
+                source,
+                dest,
+                secondary,
+                target_adapter,
+            ) = self.property._create_joins(
+                dest_selectable=to_selectable,
+                source_selectable=source_selectable,
+            )
+
+            for k in kwargs:
+                crit = getattr(self.property.mapper.class_, k) == kwargs[k]
+                if where_criteria is None:
+                    where_criteria = crit
+                else:
+                    where_criteria = where_criteria & crit
+
+            # annotate the *local* side of the join condition, in the case
+            # of pj + sj this is the full primaryjoin, in the case of just
+            # pj its the local side of the primaryjoin.
+            if sj is not None:
+                j = _orm_annotate(pj) & sj
+            else:
+                j = _orm_annotate(pj, exclude=self.property.remote_side)
+
+            if (
+                where_criteria is not None
+                and target_adapter
+                and not is_aliased_class
+            ):
+                # limit this adapter to annotated only?
+                where_criteria = target_adapter.traverse(where_criteria)
+
+            # only have the "joined left side" of what we
+            # return be subject to Query adaption.  The right
+            # side of it is used for an exists() subquery and
+            # should not correlate or otherwise reach out
+            # to anything in the enclosing query.
+            if where_criteria is not None:
+                where_criteria = where_criteria._annotate(
+                    {"no_replacement_traverse": True}
+                )
+
+            crit = j & sql.True_._ifnone(where_criteria)
+
+            if secondary is not None:
+                ex = (
+                    sql.exists(1)
+                    .where(crit)
+                    .select_from(dest, secondary)
+                    .correlate_except(dest, secondary)
+                )
+            else:
+                ex = (
+                    sql.exists(1)
+                    .where(crit)
+                    .select_from(dest)
+                    .correlate_except(dest)
+                )
+            return ex
+
+        def any(
+            self,
+            criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+            **kwargs: Any,
+        ) -> ColumnElement[bool]:
+            """Produce an expression that tests a collection against
+            particular criterion, using EXISTS.
+
+            An expression like::
+
+                session.query(MyClass).filter(
+                    MyClass.somereference.any(SomeRelated.x == 2)
+                )
+
+            Will produce a query like:
+
+            .. sourcecode:: sql
+
+                SELECT * FROM my_table WHERE
+                EXISTS (SELECT 1 FROM related WHERE related.my_id=my_table.id
+                AND related.x=2)
+
+            Because :meth:`~.Relationship.Comparator.any` uses
+            a correlated subquery, its performance is not nearly as
+            good when compared against large target tables as that of
+            using a join.
+
+            :meth:`~.Relationship.Comparator.any` is particularly
+            useful for testing for empty collections::
+
+                session.query(MyClass).filter(~MyClass.somereference.any())
+
+            will produce:
+
+            .. sourcecode:: sql
+
+                SELECT * FROM my_table WHERE
+                NOT (EXISTS (SELECT 1 FROM related WHERE
+                related.my_id=my_table.id))
+
+            :meth:`~.Relationship.Comparator.any` is only
+            valid for collections, i.e. a :func:`_orm.relationship`
+            that has ``uselist=True``.  For scalar references,
+            use :meth:`~.Relationship.Comparator.has`.
+
+            """
+            if not self.property.uselist:
+                raise sa_exc.InvalidRequestError(
+                    "'any()' not implemented for scalar "
+                    "attributes. Use has()."
+                )
+
+            return self._criterion_exists(criterion, **kwargs)
+
+        def has(
+            self,
+            criterion: Optional[_ColumnExpressionArgument[bool]] = None,
+            **kwargs: Any,
+        ) -> ColumnElement[bool]:
+            """Produce an expression that tests a scalar reference against
+            particular criterion, using EXISTS.
+
+            An expression like::
+
+                session.query(MyClass).filter(
+                    MyClass.somereference.has(SomeRelated.x == 2)
+                )
+
+            Will produce a query like:
+
+            .. sourcecode:: sql
+
+                SELECT * FROM my_table WHERE
+                EXISTS (SELECT 1 FROM related WHERE
+                related.id==my_table.related_id AND related.x=2)
+
+            Because :meth:`~.Relationship.Comparator.has` uses
+            a correlated subquery, its performance is not nearly as
+            good when compared against large target tables as that of
+            using a join.
+
+            :meth:`~.Relationship.Comparator.has` is only
+            valid for scalar references, i.e. a :func:`_orm.relationship`
+            that has ``uselist=False``.  For collection references,
+            use :meth:`~.Relationship.Comparator.any`.
+
+            """
+            if self.property.uselist:
+                raise sa_exc.InvalidRequestError(
+                    "'has()' not implemented for collections. Use any()."
+                )
+            return self._criterion_exists(criterion, **kwargs)
+
+        def contains(
+            self, other: _ColumnExpressionArgument[Any], **kwargs: Any
+        ) -> ColumnElement[bool]:
+            """Return a simple expression that tests a collection for
+            containment of a particular item.
+
+            :meth:`~.Relationship.Comparator.contains` is
+            only valid for a collection, i.e. a
+            :func:`_orm.relationship` that implements
+            one-to-many or many-to-many with ``uselist=True``.
+
+            When used in a simple one-to-many context, an
+            expression like::
+
+                MyClass.contains(other)
+
+            Produces a clause like:
+
+            .. sourcecode:: sql
+
+                mytable.id == <some id>
+
+            Where ``<some id>`` is the value of the foreign key
+            attribute on ``other`` which refers to the primary
+            key of its parent object. From this it follows that
+            :meth:`~.Relationship.Comparator.contains` is
+            very useful when used with simple one-to-many
+            operations.
+
+            For many-to-many operations, the behavior of
+            :meth:`~.Relationship.Comparator.contains`
+            has more caveats. The association table will be
+            rendered in the statement, producing an "implicit"
+            join, that is, includes multiple tables in the FROM
+            clause which are equated in the WHERE clause::
+
+                query(MyClass).filter(MyClass.contains(other))
+
+            Produces a query like:
+
+            .. sourcecode:: sql
+
+                SELECT * FROM my_table, my_association_table AS
+                my_association_table_1 WHERE
+                my_table.id = my_association_table_1.parent_id
+                AND my_association_table_1.child_id = <some id>
+
+            Where ``<some id>`` would be the primary key of
+            ``other``. From the above, it is clear that
+            :meth:`~.Relationship.Comparator.contains`
+            will **not** work with many-to-many collections when
+            used in queries that move beyond simple AND
+            conjunctions, such as multiple
+            :meth:`~.Relationship.Comparator.contains`
+            expressions joined by OR. In such cases subqueries or
+            explicit "outer joins" will need to be used instead.
+            See :meth:`~.Relationship.Comparator.any` for
+            a less-performant alternative using EXISTS, or refer
+            to :meth:`_query.Query.outerjoin`
+            as well as :ref:`orm_queryguide_joins`
+            for more details on constructing outer joins.
+
+            kwargs may be ignored by this operator but are required for API
+            conformance.
+            """
+            if not self.prop.uselist:
+                raise sa_exc.InvalidRequestError(
+                    "'contains' not implemented for scalar "
+                    "attributes.  Use =="
+                )
+
+            clause = self.prop._optimized_compare(
+                other, adapt_source=self.adapter
+            )
+
+            if self.prop.secondaryjoin is not None:
+                clause.negation_clause = self.__negated_contains_or_equals(
+                    other
+                )
+
+            return clause
+
+        def __negated_contains_or_equals(
+            self, other: Any
+        ) -> ColumnElement[bool]:
+            if self.prop.direction == MANYTOONE:
+                state = attributes.instance_state(other)
+
+                def state_bindparam(
+                    local_col: ColumnElement[Any],
+                    state: InstanceState[Any],
+                    remote_col: ColumnElement[Any],
+                ) -> BindParameter[Any]:
+                    dict_ = state.dict
+                    return sql.bindparam(
+                        local_col.key,
+                        type_=local_col.type,
+                        unique=True,
+                        callable_=self.prop._get_attr_w_warn_on_none(
+                            self.prop.mapper, state, dict_, remote_col
+                        ),
+                    )
+
+                def adapt(col: _CE) -> _CE:
+                    if self.adapter:
+                        return self.adapter(col)
+                    else:
+                        return col
+
+                if self.property._use_get:
+                    return sql.and_(
+                        *[
+                            sql.or_(
+                                adapt(x)
+                                != state_bindparam(adapt(x), state, y),
+                                adapt(x) == None,
+                            )
+                            for (x, y) in self.property.local_remote_pairs
+                        ]
+                    )
+
+            criterion = sql.and_(
+                *[
+                    x == y
+                    for (x, y) in zip(
+                        self.property.mapper.primary_key,
+                        self.property.mapper.primary_key_from_instance(other),
+                    )
+                ]
+            )
+
+            return ~self._criterion_exists(criterion)
+
+        def __ne__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            """Implement the ``!=`` operator.
+
+            In a many-to-one context, such as:
+
+            .. sourcecode:: text
+
+              MyClass.some_prop != <some object>
+
+            This will typically produce a clause such as:
+
+            .. sourcecode:: sql
+
+              mytable.related_id != <some id>
+
+            Where ``<some id>`` is the primary key of the
+            given object.
+
+            The ``!=`` operator provides partial functionality for non-
+            many-to-one comparisons:
+
+            * Comparisons against collections are not supported.
+              Use
+              :meth:`~.Relationship.Comparator.contains`
+              in conjunction with :func:`_expression.not_`.
+            * Compared to a scalar one-to-many, will produce a
+              clause that compares the target columns in the parent to
+              the given target.
+            * Compared to a scalar many-to-many, an alias
+              of the association table will be rendered as
+              well, forming a natural join that is part of the
+              main body of the query. This will not work for
+              queries that go beyond simple AND conjunctions of
+              comparisons, such as those which use OR. Use
+              explicit joins, outerjoins, or
+              :meth:`~.Relationship.Comparator.has` in
+              conjunction with :func:`_expression.not_` for
+              more comprehensive non-many-to-one scalar
+              membership tests.
+            * Comparisons against ``None`` given in a one-to-many
+              or many-to-many context produce an EXISTS clause.
+
+            """
+            if other is None or isinstance(other, expression.Null):
+                if self.property.direction == MANYTOONE:
+                    return _orm_annotate(
+                        ~self.property._optimized_compare(
+                            None, adapt_source=self.adapter
+                        )
+                    )
+
+                else:
+                    return self._criterion_exists()
+            elif self.property.uselist:
+                raise sa_exc.InvalidRequestError(
+                    "Can't compare a collection"
+                    " to an object or collection; use "
+                    "contains() to test for membership."
+                )
+            else:
+                return _orm_annotate(self.__negated_contains_or_equals(other))
+
+        def _memoized_attr_property(self) -> RelationshipProperty[_PT]:
+            self.prop.parent._check_configure()
+            return self.prop
+
+    def _with_parent(
+        self,
+        instance: object,
+        alias_secondary: bool = True,
+        from_entity: Optional[_EntityType[Any]] = None,
+    ) -> ColumnElement[bool]:
+        assert instance is not None
+        adapt_source: Optional[_CoreAdapterProto] = None
+        if from_entity is not None:
+            insp: Optional[_InternalEntityType[Any]] = inspect(from_entity)
+            assert insp is not None
+            if insp_is_aliased_class(insp):
+                adapt_source = insp._adapter.adapt_clause
+        return self._optimized_compare(
+            instance,
+            value_is_parent=True,
+            adapt_source=adapt_source,
+            alias_secondary=alias_secondary,
+        )
+
+    def _optimized_compare(
+        self,
+        state: Any,
+        value_is_parent: bool = False,
+        adapt_source: Optional[_CoreAdapterProto] = None,
+        alias_secondary: bool = True,
+    ) -> ColumnElement[bool]:
+        if state is not None:
+            try:
+                state = inspect(state)
+            except sa_exc.NoInspectionAvailable:
+                state = None
+
+            if state is None or not getattr(state, "is_instance", False):
+                raise sa_exc.ArgumentError(
+                    "Mapped instance expected for relationship "
+                    "comparison to object.   Classes, queries and other "
+                    "SQL elements are not accepted in this context; for "
+                    "comparison with a subquery, "
+                    "use %s.has(**criteria)." % self
+                )
+        reverse_direction = not value_is_parent
+
+        if state is None:
+            return self._lazy_none_clause(
+                reverse_direction, adapt_source=adapt_source
+            )
+
+        if not reverse_direction:
+            criterion, bind_to_col = (
+                self._lazy_strategy._lazywhere,
+                self._lazy_strategy._bind_to_col,
+            )
+        else:
+            criterion, bind_to_col = (
+                self._lazy_strategy._rev_lazywhere,
+                self._lazy_strategy._rev_bind_to_col,
+            )
+
+        if reverse_direction:
+            mapper = self.mapper
+        else:
+            mapper = self.parent
+
+        dict_ = attributes.instance_dict(state.obj())
+
+        def visit_bindparam(bindparam: BindParameter[Any]) -> None:
+            if bindparam._identifying_key in bind_to_col:
+                bindparam.callable = self._get_attr_w_warn_on_none(
+                    mapper,
+                    state,
+                    dict_,
+                    bind_to_col[bindparam._identifying_key],
+                )
+
+        if self.secondary is not None and alias_secondary:
+            criterion = ClauseAdapter(
+                self.secondary._anonymous_fromclause()
+            ).traverse(criterion)
+
+        criterion = visitors.cloned_traverse(
+            criterion, {}, {"bindparam": visit_bindparam}
+        )
+
+        if adapt_source:
+            criterion = adapt_source(criterion)
+        return criterion
+
+    def _get_attr_w_warn_on_none(
+        self,
+        mapper: Mapper[Any],
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        column: ColumnElement[Any],
+    ) -> Callable[[], Any]:
+        """Create the callable that is used in a many-to-one expression.
+
+        E.g.::
+
+            u1 = s.query(User).get(5)
+
+            expr = Address.user == u1
+
+        Above, the SQL should be "address.user_id = 5". The callable
+        returned by this method produces the value "5" based on the identity
+        of ``u1``.
+
+        """
+
+        # in this callable, we're trying to thread the needle through
+        # a wide variety of scenarios, including:
+        #
+        # * the object hasn't been flushed yet and there's no value for
+        #   the attribute as of yet
+        #
+        # * the object hasn't been flushed yet but it has a user-defined
+        #   value
+        #
+        # * the object has a value but it's expired and not locally present
+        #
+        # * the object has a value but it's expired and not locally present,
+        #   and the object is also detached
+        #
+        # * The object hadn't been flushed yet, there was no value, but
+        #   later, the object has been expired and detached, and *now*
+        #   they're trying to evaluate it
+        #
+        # * the object had a value, but it was changed to a new value, and
+        #   then expired
+        #
+        # * the object had a value, but it was changed to a new value, and
+        #   then expired, then the object was detached
+        #
+        # * the object has a user-set value, but it's None and we don't do
+        #   the comparison correctly for that so warn
+        #
+
+        prop = mapper.get_property_by_column(column)
+
+        # by invoking this method, InstanceState will track the last known
+        # value for this key each time the attribute is to be expired.
+        # this feature was added explicitly for use in this method.
+        state._track_last_known_value(prop.key)
+
+        lkv_fixed = state._last_known_values
+
+        def _go() -> Any:
+            assert lkv_fixed is not None
+            last_known = to_return = lkv_fixed[prop.key]
+            existing_is_available = (
+                last_known is not LoaderCallableStatus.NO_VALUE
+            )
+
+            # we support that the value may have changed.  so here we
+            # try to get the most recent value including re-fetching.
+            # only if we can't get a value now due to detachment do we return
+            # the last known value
+            current_value = mapper._get_state_attr_by_column(
+                state,
+                dict_,
+                column,
+                passive=(
+                    PassiveFlag.PASSIVE_OFF
+                    if state.persistent
+                    else PassiveFlag.PASSIVE_NO_FETCH ^ PassiveFlag.INIT_OK
+                ),
+            )
+
+            if current_value is LoaderCallableStatus.NEVER_SET:
+                if not existing_is_available:
+                    raise sa_exc.InvalidRequestError(
+                        "Can't resolve value for column %s on object "
+                        "%s; no value has been set for this column"
+                        % (column, state_str(state))
+                    )
+            elif current_value is LoaderCallableStatus.PASSIVE_NO_RESULT:
+                if not existing_is_available:
+                    raise sa_exc.InvalidRequestError(
+                        "Can't resolve value for column %s on object "
+                        "%s; the object is detached and the value was "
+                        "expired" % (column, state_str(state))
+                    )
+            else:
+                to_return = current_value
+            if to_return is None:
+                util.warn(
+                    "Got None for value of column %s; this is unsupported "
+                    "for a relationship comparison and will not "
+                    "currently produce an IS comparison "
+                    "(but may in a future release)" % column
+                )
+            return to_return
+
+        return _go
+
+    def _lazy_none_clause(
+        self,
+        reverse_direction: bool = False,
+        adapt_source: Optional[_CoreAdapterProto] = None,
+    ) -> ColumnElement[bool]:
+        if not reverse_direction:
+            criterion, bind_to_col = (
+                self._lazy_strategy._lazywhere,
+                self._lazy_strategy._bind_to_col,
+            )
+        else:
+            criterion, bind_to_col = (
+                self._lazy_strategy._rev_lazywhere,
+                self._lazy_strategy._rev_bind_to_col,
+            )
+
+        criterion = adapt_criterion_to_null(criterion, bind_to_col)
+
+        if adapt_source:
+            criterion = adapt_source(criterion)
+        return criterion
+
+    def __str__(self) -> str:
+        return str(self.parent.class_.__name__) + "." + self.key
+
+    def merge(
+        self,
+        session: Session,
+        source_state: InstanceState[Any],
+        source_dict: _InstanceDict,
+        dest_state: InstanceState[Any],
+        dest_dict: _InstanceDict,
+        load: bool,
+        _recursive: Dict[Any, object],
+        _resolve_conflict_map: Dict[_IdentityKeyType[Any], object],
+    ) -> None:
+        if load:
+            for r in self._reverse_property:
+                if (source_state, r) in _recursive:
+                    return
+
+        if "merge" not in self._cascade:
+            return
+
+        if self.key not in source_dict:
+            return
+
+        if self.uselist:
+            impl = source_state.get_impl(self.key)
+
+            assert is_has_collection_adapter(impl)
+            instances_iterable = impl.get_collection(source_state, source_dict)
+
+            # if this is a CollectionAttributeImpl, then empty should
+            # be False, otherwise "self.key in source_dict" should not be
+            # True
+            assert not instances_iterable.empty if impl.collection else True
+
+            if load:
+                # for a full merge, pre-load the destination collection,
+                # so that individual _merge of each item pulls from identity
+                # map for those already present.
+                # also assumes CollectionAttributeImpl behavior of loading
+                # "old" list in any case
+                dest_state.get_impl(self.key).get(
+                    dest_state, dest_dict, passive=PassiveFlag.PASSIVE_MERGE
+                )
+
+            dest_list = []
+            for current in instances_iterable:
+                current_state = attributes.instance_state(current)
+                current_dict = attributes.instance_dict(current)
+                _recursive[(current_state, self)] = True
+                obj = session._merge(
+                    current_state,
+                    current_dict,
+                    load=load,
+                    _recursive=_recursive,
+                    _resolve_conflict_map=_resolve_conflict_map,
+                )
+                if obj is not None:
+                    dest_list.append(obj)
+
+            if not load:
+                coll = attributes.init_state_collection(
+                    dest_state, dest_dict, self.key
+                )
+                for c in dest_list:
+                    coll.append_without_event(c)
+            else:
+                dest_impl = dest_state.get_impl(self.key)
+                assert is_has_collection_adapter(dest_impl)
+                dest_impl.set(
+                    dest_state,
+                    dest_dict,
+                    dest_list,
+                    _adapt=False,
+                    passive=PassiveFlag.PASSIVE_MERGE,
+                )
+        else:
+            current = source_dict[self.key]
+            if current is not None:
+                current_state = attributes.instance_state(current)
+                current_dict = attributes.instance_dict(current)
+                _recursive[(current_state, self)] = True
+                obj = session._merge(
+                    current_state,
+                    current_dict,
+                    load=load,
+                    _recursive=_recursive,
+                    _resolve_conflict_map=_resolve_conflict_map,
+                )
+            else:
+                obj = None
+
+            if not load:
+                dest_dict[self.key] = obj
+            else:
+                dest_state.get_impl(self.key).set(
+                    dest_state, dest_dict, obj, None
+                )
+
+    def _value_as_iterable(
+        self,
+        state: InstanceState[_O],
+        dict_: _InstanceDict,
+        key: str,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> Sequence[Tuple[InstanceState[_O], _O]]:
+        """Return a list of tuples (state, obj) for the given
+        key.
+
+        returns an empty list if the value is None/empty/PASSIVE_NO_RESULT
+        """
+
+        impl = state.manager[key].impl
+        x = impl.get(state, dict_, passive=passive)
+        if x is LoaderCallableStatus.PASSIVE_NO_RESULT or x is None:
+            return []
+        elif is_has_collection_adapter(impl):
+            return [
+                (attributes.instance_state(o), o)
+                for o in impl.get_collection(state, dict_, x, passive=passive)
+            ]
+        else:
+            return [(attributes.instance_state(x), x)]
+
+    def cascade_iterator(
+        self,
+        type_: str,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        visited_states: Set[InstanceState[Any]],
+        halt_on: Optional[Callable[[InstanceState[Any]], bool]] = None,
+    ) -> Iterator[Tuple[Any, Mapper[Any], InstanceState[Any], _InstanceDict]]:
+        # assert type_ in self._cascade
+
+        # only actively lazy load on the 'delete' cascade
+        if type_ != "delete" or self.passive_deletes:
+            passive = PassiveFlag.PASSIVE_NO_INITIALIZE
+        else:
+            passive = PassiveFlag.PASSIVE_OFF | PassiveFlag.NO_RAISE
+
+        if type_ == "save-update":
+            tuples = state.manager[self.key].impl.get_all_pending(state, dict_)
+        else:
+            tuples = self._value_as_iterable(
+                state, dict_, self.key, passive=passive
+            )
+
+        skip_pending = (
+            type_ == "refresh-expire" and "delete-orphan" not in self._cascade
+        )
+
+        for instance_state, c in tuples:
+            if instance_state in visited_states:
+                continue
+
+            if c is None:
+                # would like to emit a warning here, but
+                # would not be consistent with collection.append(None)
+                # current behavior of silently skipping.
+                # see [ticket:2229]
+                continue
+
+            assert instance_state is not None
+            instance_dict = attributes.instance_dict(c)
+
+            if halt_on and halt_on(instance_state):
+                continue
+
+            if skip_pending and not instance_state.key:
+                continue
+
+            instance_mapper = instance_state.manager.mapper
+
+            if not instance_mapper.isa(self.mapper.class_manager.mapper):
+                raise AssertionError(
+                    "Attribute '%s' on class '%s' "
+                    "doesn't handle objects "
+                    "of type '%s'"
+                    % (self.key, self.parent.class_, c.__class__)
+                )
+
+            visited_states.add(instance_state)
+
+            yield c, instance_mapper, instance_state, instance_dict
+
+    @property
+    def _effective_sync_backref(self) -> bool:
+        if self.viewonly:
+            return False
+        else:
+            return self.sync_backref is not False
+
+    @staticmethod
+    def _check_sync_backref(
+        rel_a: RelationshipProperty[Any], rel_b: RelationshipProperty[Any]
+    ) -> None:
+        if rel_a.viewonly and rel_b.sync_backref:
+            raise sa_exc.InvalidRequestError(
+                "Relationship %s cannot specify sync_backref=True since %s "
+                "includes viewonly=True." % (rel_b, rel_a)
+            )
+        if (
+            rel_a.viewonly
+            and not rel_b.viewonly
+            and rel_b.sync_backref is not False
+        ):
+            rel_b.sync_backref = False
+
+    def _add_reverse_property(self, key: str) -> None:
+        other = self.mapper.get_property(key, _configure_mappers=False)
+        if not isinstance(other, RelationshipProperty):
+            raise sa_exc.InvalidRequestError(
+                "back_populates on relationship '%s' refers to attribute '%s' "
+                "that is not a relationship.  The back_populates parameter "
+                "should refer to the name of a relationship on the target "
+                "class." % (self, other)
+            )
+        # viewonly and sync_backref cases
+        # 1. self.viewonly==True and other.sync_backref==True -> error
+        # 2. self.viewonly==True and other.viewonly==False and
+        #    other.sync_backref==None -> warn sync_backref=False, set to False
+        self._check_sync_backref(self, other)
+        # 3. other.viewonly==True and self.sync_backref==True -> error
+        # 4. other.viewonly==True and self.viewonly==False and
+        #    self.sync_backref==None -> warn sync_backref=False, set to False
+        self._check_sync_backref(other, self)
+
+        self._reverse_property.add(other)
+        other._reverse_property.add(self)
+
+        other._setup_entity()
+
+        if not other.mapper.common_parent(self.parent):
+            raise sa_exc.ArgumentError(
+                "reverse_property %r on "
+                "relationship %s references relationship %s, which "
+                "does not reference mapper %s"
+                % (key, self, other, self.parent)
+            )
+
+        if (
+            other._configure_started
+            and self.direction in (ONETOMANY, MANYTOONE)
+            and self.direction == other.direction
+        ):
+            raise sa_exc.ArgumentError(
+                "%s and back-reference %s are "
+                "both of the same direction %r.  Did you mean to "
+                "set remote_side on the many-to-one side ?"
+                % (other, self, self.direction)
+            )
+
+    @util.memoized_property
+    def entity(self) -> _InternalEntityType[_T]:
+        """Return the target mapped entity, which is an inspect() of the
+        class or aliased class that is referenced by this
+        :class:`.RelationshipProperty`.
+
+        """
+        self.parent._check_configure()
+        return self.entity
+
+    @util.memoized_property
+    def mapper(self) -> Mapper[_T]:
+        """Return the targeted :class:`_orm.Mapper` for this
+        :class:`.RelationshipProperty`.
+
+        """
+        return self.entity.mapper
+
+    def do_init(self) -> None:
+        self._check_conflicts()
+        self._process_dependent_arguments()
+        self._setup_entity()
+        self._setup_registry_dependencies()
+        self._setup_join_conditions()
+        self._check_cascade_settings(self._cascade)
+        self._post_init()
+        self._generate_backref()
+        self._join_condition._warn_for_conflicting_sync_targets()
+        super().do_init()
+        self._lazy_strategy = cast(
+            "LazyLoader", self._get_strategy((("lazy", "select"),))
+        )
+
+    def _setup_registry_dependencies(self) -> None:
+        self.parent.mapper.registry._set_depends_on(
+            self.entity.mapper.registry
+        )
+
+    def _process_dependent_arguments(self) -> None:
+        """Convert incoming configuration arguments to their
+        proper form.
+
+        Callables are resolved, ORM annotations removed.
+
+        """
+
+        # accept callables for other attributes which may require
+        # deferred initialization.  This technique is used
+        # by declarative "string configs" and some recipes.
+        init_args = self._init_args
+
+        for attr in (
+            "order_by",
+            "primaryjoin",
+            "secondaryjoin",
+            "secondary",
+            "foreign_keys",
+            "remote_side",
+        ):
+            rel_arg = getattr(init_args, attr)
+
+            rel_arg._resolve_against_registry(self._clsregistry_resolvers[1])
+
+        # remove "annotations" which are present if mapped class
+        # descriptors are used to create the join expression.
+        for attr in "primaryjoin", "secondaryjoin":
+            rel_arg = getattr(init_args, attr)
+            val = rel_arg.resolved
+            if val is not None:
+                rel_arg.resolved = _orm_deannotate(
+                    coercions.expect(
+                        roles.ColumnArgumentRole, val, argname=attr
+                    )
+                )
+
+        secondary = init_args.secondary.resolved
+        if secondary is not None and _is_mapped_class(secondary):
+            raise sa_exc.ArgumentError(
+                "secondary argument %s passed to to relationship() %s must "
+                "be a Table object or other FROM clause; can't send a mapped "
+                "class directly as rows in 'secondary' are persisted "
+                "independently of a class that is mapped "
+                "to that same table." % (secondary, self)
+            )
+
+        # ensure expressions in self.order_by, foreign_keys,
+        # remote_side are all columns, not strings.
+        if (
+            init_args.order_by.resolved is not False
+            and init_args.order_by.resolved is not None
+        ):
+            self.order_by = tuple(
+                coercions.expect(
+                    roles.ColumnArgumentRole, x, argname="order_by"
+                )
+                for x in util.to_list(init_args.order_by.resolved)
+            )
+        else:
+            self.order_by = False
+
+        self._user_defined_foreign_keys = util.column_set(
+            coercions.expect(
+                roles.ColumnArgumentRole, x, argname="foreign_keys"
+            )
+            for x in util.to_column_set(init_args.foreign_keys.resolved)
+        )
+
+        self.remote_side = util.column_set(
+            coercions.expect(
+                roles.ColumnArgumentRole, x, argname="remote_side"
+            )
+            for x in util.to_column_set(init_args.remote_side.resolved)
+        )
+
+    def declarative_scan(
+        self,
+        decl_scan: _ClassScanMapperConfig,
+        registry: _RegistryType,
+        cls: Type[Any],
+        originating_module: Optional[str],
+        key: str,
+        mapped_container: Optional[Type[Mapped[Any]]],
+        annotation: Optional[_AnnotationScanType],
+        extracted_mapped_annotation: Optional[_AnnotationScanType],
+        is_dataclass_field: bool,
+    ) -> None:
+        argument = extracted_mapped_annotation
+
+        if extracted_mapped_annotation is None:
+            if self.argument is None:
+                self._raise_for_required(key, cls)
+            else:
+                return
+
+        argument = extracted_mapped_annotation
+        assert originating_module is not None
+
+        if mapped_container is not None:
+            is_write_only = issubclass(mapped_container, WriteOnlyMapped)
+            is_dynamic = issubclass(mapped_container, DynamicMapped)
+            if is_write_only:
+                self.lazy = "write_only"
+                self.strategy_key = (("lazy", self.lazy),)
+            elif is_dynamic:
+                self.lazy = "dynamic"
+                self.strategy_key = (("lazy", self.lazy),)
+        else:
+            is_write_only = is_dynamic = False
+
+        argument = de_optionalize_union_types(argument)
+
+        if hasattr(argument, "__origin__"):
+            arg_origin = argument.__origin__
+            if isinstance(arg_origin, type) and issubclass(
+                arg_origin, abc.Collection
+            ):
+                if self.collection_class is None:
+                    if _py_inspect.isabstract(arg_origin):
+                        raise sa_exc.ArgumentError(
+                            f"Collection annotation type {arg_origin} cannot "
+                            "be instantiated; please provide an explicit "
+                            "'collection_class' parameter "
+                            "(e.g. list, set, etc.) to the "
+                            "relationship() function to accompany this "
+                            "annotation"
+                        )
+
+                    self.collection_class = arg_origin
+
+            elif not is_write_only and not is_dynamic:
+                self.uselist = False
+
+            if argument.__args__:  # type: ignore
+                if isinstance(arg_origin, type) and issubclass(
+                    arg_origin, typing.Mapping
+                ):
+                    type_arg = argument.__args__[-1]  # type: ignore
+                else:
+                    type_arg = argument.__args__[0]  # type: ignore
+                if hasattr(type_arg, "__forward_arg__"):
+                    str_argument = type_arg.__forward_arg__
+
+                    argument = resolve_name_to_real_class_name(
+                        str_argument, originating_module
+                    )
+                else:
+                    argument = type_arg
+            else:
+                raise sa_exc.ArgumentError(
+                    f"Generic alias {argument} requires an argument"
+                )
+        elif hasattr(argument, "__forward_arg__"):
+            argument = argument.__forward_arg__
+
+            argument = resolve_name_to_real_class_name(
+                argument, originating_module
+            )
+
+        if (
+            self.collection_class is None
+            and not is_write_only
+            and not is_dynamic
+        ):
+            self.uselist = False
+
+        # ticket #8759
+        # if a lead argument was given to relationship(), like
+        # `relationship("B")`, use that, don't replace it with class we
+        # found in the annotation.  The declarative_scan() method call here is
+        # still useful, as we continue to derive collection type and do
+        # checking of the annotation in any case.
+        if self.argument is None:
+            self.argument = cast("_RelationshipArgumentType[_T]", argument)
+
+    @util.preload_module("sqlalchemy.orm.mapper")
+    def _setup_entity(self, __argument: Any = None) -> None:
+        if "entity" in self.__dict__:
+            return
+
+        mapperlib = util.preloaded.orm_mapper
+
+        if __argument:
+            argument = __argument
+        else:
+            argument = self.argument
+
+        resolved_argument: _ExternalEntityType[Any]
+
+        if isinstance(argument, str):
+            # we might want to cleanup clsregistry API to make this
+            # more straightforward
+            resolved_argument = cast(
+                "_ExternalEntityType[Any]",
+                self._clsregistry_resolve_name(argument)(),
+            )
+        elif callable(argument) and not isinstance(
+            argument, (type, mapperlib.Mapper)
+        ):
+            resolved_argument = argument()
+        else:
+            resolved_argument = argument
+
+        entity: _InternalEntityType[Any]
+
+        if isinstance(resolved_argument, type):
+            entity = class_mapper(resolved_argument, configure=False)
+        else:
+            try:
+                entity = inspect(resolved_argument)
+            except sa_exc.NoInspectionAvailable:
+                entity = None  # type: ignore
+
+            if not hasattr(entity, "mapper"):
+                raise sa_exc.ArgumentError(
+                    "relationship '%s' expects "
+                    "a class or a mapper argument (received: %s)"
+                    % (self.key, type(resolved_argument))
+                )
+
+        self.entity = entity
+        self.target = self.entity.persist_selectable
+
+    def _setup_join_conditions(self) -> None:
+        self._join_condition = jc = JoinCondition(
+            parent_persist_selectable=self.parent.persist_selectable,
+            child_persist_selectable=self.entity.persist_selectable,
+            parent_local_selectable=self.parent.local_table,
+            child_local_selectable=self.entity.local_table,
+            primaryjoin=self._init_args.primaryjoin.resolved,
+            secondary=self._init_args.secondary.resolved,
+            secondaryjoin=self._init_args.secondaryjoin.resolved,
+            parent_equivalents=self.parent._equivalent_columns,
+            child_equivalents=self.mapper._equivalent_columns,
+            consider_as_foreign_keys=self._user_defined_foreign_keys,
+            local_remote_pairs=self.local_remote_pairs,
+            remote_side=self.remote_side,
+            self_referential=self._is_self_referential,
+            prop=self,
+            support_sync=not self.viewonly,
+            can_be_synced_fn=self._columns_are_mapped,
+        )
+        self.primaryjoin = jc.primaryjoin
+        self.secondaryjoin = jc.secondaryjoin
+        self.secondary = jc.secondary
+        self.direction = jc.direction
+        self.local_remote_pairs = jc.local_remote_pairs
+        self.remote_side = jc.remote_columns
+        self.local_columns = jc.local_columns
+        self.synchronize_pairs = jc.synchronize_pairs
+        self._calculated_foreign_keys = jc.foreign_key_columns
+        self.secondary_synchronize_pairs = jc.secondary_synchronize_pairs
+
+    @property
+    def _clsregistry_resolve_arg(
+        self,
+    ) -> Callable[[str, bool], _class_resolver]:
+        return self._clsregistry_resolvers[1]
+
+    @property
+    def _clsregistry_resolve_name(
+        self,
+    ) -> Callable[[str], Callable[[], Union[Type[Any], Table, _ModNS]]]:
+        return self._clsregistry_resolvers[0]
+
+    @util.memoized_property
+    @util.preload_module("sqlalchemy.orm.clsregistry")
+    def _clsregistry_resolvers(
+        self,
+    ) -> Tuple[
+        Callable[[str], Callable[[], Union[Type[Any], Table, _ModNS]]],
+        Callable[[str, bool], _class_resolver],
+    ]:
+        _resolver = util.preloaded.orm_clsregistry._resolver
+
+        return _resolver(self.parent.class_, self)
+
+    def _check_conflicts(self) -> None:
+        """Test that this relationship is legal, warn about
+        inheritance conflicts."""
+        if self.parent.non_primary and not class_mapper(
+            self.parent.class_, configure=False
+        ).has_property(self.key):
+            raise sa_exc.ArgumentError(
+                "Attempting to assign a new "
+                "relationship '%s' to a non-primary mapper on "
+                "class '%s'.  New relationships can only be added "
+                "to the primary mapper, i.e. the very first mapper "
+                "created for class '%s' "
+                % (
+                    self.key,
+                    self.parent.class_.__name__,
+                    self.parent.class_.__name__,
+                )
+            )
+
+    @property
+    def cascade(self) -> CascadeOptions:
+        """Return the current cascade setting for this
+        :class:`.RelationshipProperty`.
+        """
+        return self._cascade
+
+    @cascade.setter
+    def cascade(self, cascade: Union[str, CascadeOptions]) -> None:
+        self._set_cascade(cascade)
+
+    def _set_cascade(self, cascade_arg: Union[str, CascadeOptions]) -> None:
+        cascade = CascadeOptions(cascade_arg)
+
+        if self.viewonly:
+            cascade = CascadeOptions(
+                cascade.intersection(CascadeOptions._viewonly_cascades)
+            )
+
+        if "mapper" in self.__dict__:
+            self._check_cascade_settings(cascade)
+        self._cascade = cascade
+
+        if self._dependency_processor:
+            self._dependency_processor.cascade = cascade
+
+    def _check_cascade_settings(self, cascade: CascadeOptions) -> None:
+        if (
+            cascade.delete_orphan
+            and not self.single_parent
+            and (self.direction is MANYTOMANY or self.direction is MANYTOONE)
+        ):
+            raise sa_exc.ArgumentError(
+                "For %(direction)s relationship %(rel)s, delete-orphan "
+                "cascade is normally "
+                'configured only on the "one" side of a one-to-many '
+                "relationship, "
+                'and not on the "many" side of a many-to-one or many-to-many '
+                "relationship.  "
+                "To force this relationship to allow a particular "
+                '"%(relatedcls)s" object to be referenced by only '
+                'a single "%(clsname)s" object at a time via the '
+                "%(rel)s relationship, which "
+                "would allow "
+                "delete-orphan cascade to take place in this direction, set "
+                "the single_parent=True flag."
+                % {
+                    "rel": self,
+                    "direction": (
+                        "many-to-one"
+                        if self.direction is MANYTOONE
+                        else "many-to-many"
+                    ),
+                    "clsname": self.parent.class_.__name__,
+                    "relatedcls": self.mapper.class_.__name__,
+                },
+                code="bbf0",
+            )
+
+        if self.passive_deletes == "all" and (
+            "delete" in cascade or "delete-orphan" in cascade
+        ):
+            raise sa_exc.ArgumentError(
+                "On %s, can't set passive_deletes='all' in conjunction "
+                "with 'delete' or 'delete-orphan' cascade" % self
+            )
+
+        if cascade.delete_orphan:
+            self.mapper.primary_mapper()._delete_orphans.append(
+                (self.key, self.parent.class_)
+            )
+
+    def _persists_for(self, mapper: Mapper[Any]) -> bool:
+        """Return True if this property will persist values on behalf
+        of the given mapper.
+
+        """
+
+        return (
+            self.key in mapper.relationships
+            and mapper.relationships[self.key] is self
+        )
+
+    def _columns_are_mapped(self, *cols: ColumnElement[Any]) -> bool:
+        """Return True if all columns in the given collection are
+        mapped by the tables referenced by this :class:`.RelationshipProperty`.
+
+        """
+
+        secondary = self._init_args.secondary.resolved
+        for c in cols:
+            if secondary is not None and secondary.c.contains_column(c):
+                continue
+            if not self.parent.persist_selectable.c.contains_column(
+                c
+            ) and not self.target.c.contains_column(c):
+                return False
+        return True
+
+    def _generate_backref(self) -> None:
+        """Interpret the 'backref' instruction to create a
+        :func:`_orm.relationship` complementary to this one."""
+
+        if self.parent.non_primary:
+            return
+        if self.backref is not None and not self.back_populates:
+            kwargs: Dict[str, Any]
+            if isinstance(self.backref, str):
+                backref_key, kwargs = self.backref, {}
+            else:
+                backref_key, kwargs = self.backref
+            mapper = self.mapper.primary_mapper()
+
+            if not mapper.concrete:
+                check = set(mapper.iterate_to_root()).union(
+                    mapper.self_and_descendants
+                )
+                for m in check:
+                    if m.has_property(backref_key) and not m.concrete:
+                        raise sa_exc.ArgumentError(
+                            "Error creating backref "
+                            "'%s' on relationship '%s': property of that "
+                            "name exists on mapper '%s'"
+                            % (backref_key, self, m)
+                        )
+
+            # determine primaryjoin/secondaryjoin for the
+            # backref.  Use the one we had, so that
+            # a custom join doesn't have to be specified in
+            # both directions.
+            if self.secondary is not None:
+                # for many to many, just switch primaryjoin/
+                # secondaryjoin.   use the annotated
+                # pj/sj on the _join_condition.
+                pj = kwargs.pop(
+                    "primaryjoin",
+                    self._join_condition.secondaryjoin_minus_local,
+                )
+                sj = kwargs.pop(
+                    "secondaryjoin",
+                    self._join_condition.primaryjoin_minus_local,
+                )
+            else:
+                pj = kwargs.pop(
+                    "primaryjoin",
+                    self._join_condition.primaryjoin_reverse_remote,
+                )
+                sj = kwargs.pop("secondaryjoin", None)
+                if sj:
+                    raise sa_exc.InvalidRequestError(
+                        "Can't assign 'secondaryjoin' on a backref "
+                        "against a non-secondary relationship."
+                    )
+
+            foreign_keys = kwargs.pop(
+                "foreign_keys", self._user_defined_foreign_keys
+            )
+            parent = self.parent.primary_mapper()
+            kwargs.setdefault("viewonly", self.viewonly)
+            kwargs.setdefault("post_update", self.post_update)
+            kwargs.setdefault("passive_updates", self.passive_updates)
+            kwargs.setdefault("sync_backref", self.sync_backref)
+            self.back_populates = backref_key
+            relationship = RelationshipProperty(
+                parent,
+                self.secondary,
+                primaryjoin=pj,
+                secondaryjoin=sj,
+                foreign_keys=foreign_keys,
+                back_populates=self.key,
+                **kwargs,
+            )
+            mapper._configure_property(
+                backref_key, relationship, warn_for_existing=True
+            )
+
+        if self.back_populates:
+            self._add_reverse_property(self.back_populates)
+
+    @util.preload_module("sqlalchemy.orm.dependency")
+    def _post_init(self) -> None:
+        dependency = util.preloaded.orm_dependency
+
+        if self.uselist is None:
+            self.uselist = self.direction is not MANYTOONE
+        if not self.viewonly:
+            self._dependency_processor = (  # type: ignore
+                dependency.DependencyProcessor.from_relationship
+            )(self)
+
+    @util.memoized_property
+    def _use_get(self) -> bool:
+        """memoize the 'use_get' attribute of this RelationshipLoader's
+        lazyloader."""
+
+        strategy = self._lazy_strategy
+        return strategy.use_get
+
+    @util.memoized_property
+    def _is_self_referential(self) -> bool:
+        return self.mapper.common_parent(self.parent)
+
+    def _create_joins(
+        self,
+        source_polymorphic: bool = False,
+        source_selectable: Optional[FromClause] = None,
+        dest_selectable: Optional[FromClause] = None,
+        of_type_entity: Optional[_InternalEntityType[Any]] = None,
+        alias_secondary: bool = False,
+        extra_criteria: Tuple[ColumnElement[bool], ...] = (),
+    ) -> Tuple[
+        ColumnElement[bool],
+        Optional[ColumnElement[bool]],
+        FromClause,
+        FromClause,
+        Optional[FromClause],
+        Optional[ClauseAdapter],
+    ]:
+        aliased = False
+
+        if alias_secondary and self.secondary is not None:
+            aliased = True
+
+        if source_selectable is None:
+            if source_polymorphic and self.parent.with_polymorphic:
+                source_selectable = self.parent._with_polymorphic_selectable
+
+        if of_type_entity:
+            dest_mapper = of_type_entity.mapper
+            if dest_selectable is None:
+                dest_selectable = of_type_entity.selectable
+                aliased = True
+        else:
+            dest_mapper = self.mapper
+
+        if dest_selectable is None:
+            dest_selectable = self.entity.selectable
+            if self.mapper.with_polymorphic:
+                aliased = True
+
+            if self._is_self_referential and source_selectable is None:
+                dest_selectable = dest_selectable._anonymous_fromclause()
+                aliased = True
+        elif (
+            dest_selectable is not self.mapper._with_polymorphic_selectable
+            or self.mapper.with_polymorphic
+        ):
+            aliased = True
+
+        single_crit = dest_mapper._single_table_criterion
+        aliased = aliased or (
+            source_selectable is not None
+            and (
+                source_selectable
+                is not self.parent._with_polymorphic_selectable
+                or source_selectable._is_subquery
+            )
+        )
+
+        (
+            primaryjoin,
+            secondaryjoin,
+            secondary,
+            target_adapter,
+            dest_selectable,
+        ) = self._join_condition.join_targets(
+            source_selectable,
+            dest_selectable,
+            aliased,
+            single_crit,
+            extra_criteria,
+        )
+        if source_selectable is None:
+            source_selectable = self.parent.local_table
+        if dest_selectable is None:
+            dest_selectable = self.entity.local_table
+        return (
+            primaryjoin,
+            secondaryjoin,
+            source_selectable,
+            dest_selectable,
+            secondary,
+            target_adapter,
+        )
+
+
+def _annotate_columns(element: _CE, annotations: _AnnotationDict) -> _CE:
+    def clone(elem: _CE) -> _CE:
+        if isinstance(elem, expression.ColumnClause):
+            elem = elem._annotate(annotations.copy())  # type: ignore
+        elem._copy_internals(clone=clone)
+        return elem
+
+    if element is not None:
+        element = clone(element)
+    clone = None  # type: ignore # remove gc cycles
+    return element
+
+
+class JoinCondition:
+    primaryjoin_initial: Optional[ColumnElement[bool]]
+    primaryjoin: ColumnElement[bool]
+    secondaryjoin: Optional[ColumnElement[bool]]
+    secondary: Optional[FromClause]
+    prop: RelationshipProperty[Any]
+
+    synchronize_pairs: _ColumnPairs
+    secondary_synchronize_pairs: _ColumnPairs
+    direction: RelationshipDirection
+
+    parent_persist_selectable: FromClause
+    child_persist_selectable: FromClause
+    parent_local_selectable: FromClause
+    child_local_selectable: FromClause
+
+    _local_remote_pairs: Optional[_ColumnPairs]
+
+    def __init__(
+        self,
+        parent_persist_selectable: FromClause,
+        child_persist_selectable: FromClause,
+        parent_local_selectable: FromClause,
+        child_local_selectable: FromClause,
+        *,
+        primaryjoin: Optional[ColumnElement[bool]] = None,
+        secondary: Optional[FromClause] = None,
+        secondaryjoin: Optional[ColumnElement[bool]] = None,
+        parent_equivalents: Optional[_EquivalentColumnMap] = None,
+        child_equivalents: Optional[_EquivalentColumnMap] = None,
+        consider_as_foreign_keys: Any = None,
+        local_remote_pairs: Optional[_ColumnPairs] = None,
+        remote_side: Any = None,
+        self_referential: Any = False,
+        prop: RelationshipProperty[Any],
+        support_sync: bool = True,
+        can_be_synced_fn: Callable[..., bool] = lambda *c: True,
+    ):
+        self.parent_persist_selectable = parent_persist_selectable
+        self.parent_local_selectable = parent_local_selectable
+        self.child_persist_selectable = child_persist_selectable
+        self.child_local_selectable = child_local_selectable
+        self.parent_equivalents = parent_equivalents
+        self.child_equivalents = child_equivalents
+        self.primaryjoin_initial = primaryjoin
+        self.secondaryjoin = secondaryjoin
+        self.secondary = secondary
+        self.consider_as_foreign_keys = consider_as_foreign_keys
+        self._local_remote_pairs = local_remote_pairs
+        self._remote_side = remote_side
+        self.prop = prop
+        self.self_referential = self_referential
+        self.support_sync = support_sync
+        self.can_be_synced_fn = can_be_synced_fn
+
+        self._determine_joins()
+        assert self.primaryjoin is not None
+
+        self._sanitize_joins()
+        self._annotate_fks()
+        self._annotate_remote()
+        self._annotate_local()
+        self._annotate_parentmapper()
+        self._setup_pairs()
+        self._check_foreign_cols(self.primaryjoin, True)
+        if self.secondaryjoin is not None:
+            self._check_foreign_cols(self.secondaryjoin, False)
+        self._determine_direction()
+        self._check_remote_side()
+        self._log_joins()
+
+    def _log_joins(self) -> None:
+        log = self.prop.logger
+        log.info("%s setup primary join %s", self.prop, self.primaryjoin)
+        log.info("%s setup secondary join %s", self.prop, self.secondaryjoin)
+        log.info(
+            "%s synchronize pairs [%s]",
+            self.prop,
+            ",".join(
+                "(%s => %s)" % (l, r) for (l, r) in self.synchronize_pairs
+            ),
+        )
+        log.info(
+            "%s secondary synchronize pairs [%s]",
+            self.prop,
+            ",".join(
+                "(%s => %s)" % (l, r)
+                for (l, r) in self.secondary_synchronize_pairs or []
+            ),
+        )
+        log.info(
+            "%s local/remote pairs [%s]",
+            self.prop,
+            ",".join(
+                "(%s / %s)" % (l, r) for (l, r) in self.local_remote_pairs
+            ),
+        )
+        log.info(
+            "%s remote columns [%s]",
+            self.prop,
+            ",".join("%s" % col for col in self.remote_columns),
+        )
+        log.info(
+            "%s local columns [%s]",
+            self.prop,
+            ",".join("%s" % col for col in self.local_columns),
+        )
+        log.info("%s relationship direction %s", self.prop, self.direction)
+
+    def _sanitize_joins(self) -> None:
+        """remove the parententity annotation from our join conditions which
+        can leak in here based on some declarative patterns and maybe others.
+
+        "parentmapper" is relied upon both by the ORM evaluator as well as
+        the use case in _join_fixture_inh_selfref_w_entity
+        that relies upon it being present, see :ticket:`3364`.
+
+        """
+
+        self.primaryjoin = _deep_deannotate(
+            self.primaryjoin, values=("parententity", "proxy_key")
+        )
+        if self.secondaryjoin is not None:
+            self.secondaryjoin = _deep_deannotate(
+                self.secondaryjoin, values=("parententity", "proxy_key")
+            )
+
+    def _determine_joins(self) -> None:
+        """Determine the 'primaryjoin' and 'secondaryjoin' attributes,
+        if not passed to the constructor already.
+
+        This is based on analysis of the foreign key relationships
+        between the parent and target mapped selectables.
+
+        """
+        if self.secondaryjoin is not None and self.secondary is None:
+            raise sa_exc.ArgumentError(
+                "Property %s specified with secondary "
+                "join condition but "
+                "no secondary argument" % self.prop
+            )
+
+        # find a join between the given mapper's mapped table and
+        # the given table. will try the mapper's local table first
+        # for more specificity, then if not found will try the more
+        # general mapped table, which in the case of inheritance is
+        # a join.
+        try:
+            consider_as_foreign_keys = self.consider_as_foreign_keys or None
+            if self.secondary is not None:
+                if self.secondaryjoin is None:
+                    self.secondaryjoin = join_condition(
+                        self.child_persist_selectable,
+                        self.secondary,
+                        a_subset=self.child_local_selectable,
+                        consider_as_foreign_keys=consider_as_foreign_keys,
+                    )
+                if self.primaryjoin_initial is None:
+                    self.primaryjoin = join_condition(
+                        self.parent_persist_selectable,
+                        self.secondary,
+                        a_subset=self.parent_local_selectable,
+                        consider_as_foreign_keys=consider_as_foreign_keys,
+                    )
+                else:
+                    self.primaryjoin = self.primaryjoin_initial
+            else:
+                if self.primaryjoin_initial is None:
+                    self.primaryjoin = join_condition(
+                        self.parent_persist_selectable,
+                        self.child_persist_selectable,
+                        a_subset=self.parent_local_selectable,
+                        consider_as_foreign_keys=consider_as_foreign_keys,
+                    )
+                else:
+                    self.primaryjoin = self.primaryjoin_initial
+        except sa_exc.NoForeignKeysError as nfe:
+            if self.secondary is not None:
+                raise sa_exc.NoForeignKeysError(
+                    "Could not determine join "
+                    "condition between parent/child tables on "
+                    "relationship %s - there are no foreign keys "
+                    "linking these tables via secondary table '%s'.  "
+                    "Ensure that referencing columns are associated "
+                    "with a ForeignKey or ForeignKeyConstraint, or "
+                    "specify 'primaryjoin' and 'secondaryjoin' "
+                    "expressions." % (self.prop, self.secondary)
+                ) from nfe
+            else:
+                raise sa_exc.NoForeignKeysError(
+                    "Could not determine join "
+                    "condition between parent/child tables on "
+                    "relationship %s - there are no foreign keys "
+                    "linking these tables.  "
+                    "Ensure that referencing columns are associated "
+                    "with a ForeignKey or ForeignKeyConstraint, or "
+                    "specify a 'primaryjoin' expression." % self.prop
+                ) from nfe
+        except sa_exc.AmbiguousForeignKeysError as afe:
+            if self.secondary is not None:
+                raise sa_exc.AmbiguousForeignKeysError(
+                    "Could not determine join "
+                    "condition between parent/child tables on "
+                    "relationship %s - there are multiple foreign key "
+                    "paths linking the tables via secondary table '%s'.  "
+                    "Specify the 'foreign_keys' "
+                    "argument, providing a list of those columns which "
+                    "should be counted as containing a foreign key "
+                    "reference from the secondary table to each of the "
+                    "parent and child tables." % (self.prop, self.secondary)
+                ) from afe
+            else:
+                raise sa_exc.AmbiguousForeignKeysError(
+                    "Could not determine join "
+                    "condition between parent/child tables on "
+                    "relationship %s - there are multiple foreign key "
+                    "paths linking the tables.  Specify the "
+                    "'foreign_keys' argument, providing a list of those "
+                    "columns which should be counted as containing a "
+                    "foreign key reference to the parent table." % self.prop
+                ) from afe
+
+    @property
+    def primaryjoin_minus_local(self) -> ColumnElement[bool]:
+        return _deep_deannotate(self.primaryjoin, values=("local", "remote"))
+
+    @property
+    def secondaryjoin_minus_local(self) -> ColumnElement[bool]:
+        assert self.secondaryjoin is not None
+        return _deep_deannotate(self.secondaryjoin, values=("local", "remote"))
+
+    @util.memoized_property
+    def primaryjoin_reverse_remote(self) -> ColumnElement[bool]:
+        """Return the primaryjoin condition suitable for the
+        "reverse" direction.
+
+        If the primaryjoin was delivered here with pre-existing
+        "remote" annotations, the local/remote annotations
+        are reversed.  Otherwise, the local/remote annotations
+        are removed.
+
+        """
+        if self._has_remote_annotations:
+
+            def replace(element: _CE, **kw: Any) -> Optional[_CE]:
+                if "remote" in element._annotations:
+                    v = dict(element._annotations)
+                    del v["remote"]
+                    v["local"] = True
+                    return element._with_annotations(v)
+                elif "local" in element._annotations:
+                    v = dict(element._annotations)
+                    del v["local"]
+                    v["remote"] = True
+                    return element._with_annotations(v)
+
+                return None
+
+            return visitors.replacement_traverse(self.primaryjoin, {}, replace)
+        else:
+            if self._has_foreign_annotations:
+                # TODO: coverage
+                return _deep_deannotate(
+                    self.primaryjoin, values=("local", "remote")
+                )
+            else:
+                return _deep_deannotate(self.primaryjoin)
+
+    def _has_annotation(self, clause: ClauseElement, annotation: str) -> bool:
+        for col in visitors.iterate(clause, {}):
+            if annotation in col._annotations:
+                return True
+        else:
+            return False
+
+    @util.memoized_property
+    def _has_foreign_annotations(self) -> bool:
+        return self._has_annotation(self.primaryjoin, "foreign")
+
+    @util.memoized_property
+    def _has_remote_annotations(self) -> bool:
+        return self._has_annotation(self.primaryjoin, "remote")
+
+    def _annotate_fks(self) -> None:
+        """Annotate the primaryjoin and secondaryjoin
+        structures with 'foreign' annotations marking columns
+        considered as foreign.
+
+        """
+        if self._has_foreign_annotations:
+            return
+
+        if self.consider_as_foreign_keys:
+            self._annotate_from_fk_list()
+        else:
+            self._annotate_present_fks()
+
+    def _annotate_from_fk_list(self) -> None:
+        def check_fk(element: _CE, **kw: Any) -> Optional[_CE]:
+            if element in self.consider_as_foreign_keys:
+                return element._annotate({"foreign": True})
+            return None
+
+        self.primaryjoin = visitors.replacement_traverse(
+            self.primaryjoin, {}, check_fk
+        )
+        if self.secondaryjoin is not None:
+            self.secondaryjoin = visitors.replacement_traverse(
+                self.secondaryjoin, {}, check_fk
+            )
+
+    def _annotate_present_fks(self) -> None:
+        if self.secondary is not None:
+            secondarycols = util.column_set(self.secondary.c)
+        else:
+            secondarycols = set()
+
+        def is_foreign(
+            a: ColumnElement[Any], b: ColumnElement[Any]
+        ) -> Optional[ColumnElement[Any]]:
+            if isinstance(a, schema.Column) and isinstance(b, schema.Column):
+                if a.references(b):
+                    return a
+                elif b.references(a):
+                    return b
+
+            if secondarycols:
+                if a in secondarycols and b not in secondarycols:
+                    return a
+                elif b in secondarycols and a not in secondarycols:
+                    return b
+
+            return None
+
+        def visit_binary(binary: BinaryExpression[Any]) -> None:
+            if not isinstance(
+                binary.left, sql.ColumnElement
+            ) or not isinstance(binary.right, sql.ColumnElement):
+                return
+
+            if (
+                "foreign" not in binary.left._annotations
+                and "foreign" not in binary.right._annotations
+            ):
+                col = is_foreign(binary.left, binary.right)
+                if col is not None:
+                    if col.compare(binary.left):
+                        binary.left = binary.left._annotate({"foreign": True})
+                    elif col.compare(binary.right):
+                        binary.right = binary.right._annotate(
+                            {"foreign": True}
+                        )
+
+        self.primaryjoin = visitors.cloned_traverse(
+            self.primaryjoin, {}, {"binary": visit_binary}
+        )
+        if self.secondaryjoin is not None:
+            self.secondaryjoin = visitors.cloned_traverse(
+                self.secondaryjoin, {}, {"binary": visit_binary}
+            )
+
+    def _refers_to_parent_table(self) -> bool:
+        """Return True if the join condition contains column
+        comparisons where both columns are in both tables.
+
+        """
+        pt = self.parent_persist_selectable
+        mt = self.child_persist_selectable
+        result = False
+
+        def visit_binary(binary: BinaryExpression[Any]) -> None:
+            nonlocal result
+            c, f = binary.left, binary.right
+            if (
+                isinstance(c, expression.ColumnClause)
+                and isinstance(f, expression.ColumnClause)
+                and pt.is_derived_from(c.table)
+                and pt.is_derived_from(f.table)
+                and mt.is_derived_from(c.table)
+                and mt.is_derived_from(f.table)
+            ):
+                result = True
+
+        visitors.traverse(self.primaryjoin, {}, {"binary": visit_binary})
+        return result
+
+    def _tables_overlap(self) -> bool:
+        """Return True if parent/child tables have some overlap."""
+
+        return selectables_overlap(
+            self.parent_persist_selectable, self.child_persist_selectable
+        )
+
+    def _annotate_remote(self) -> None:
+        """Annotate the primaryjoin and secondaryjoin
+        structures with 'remote' annotations marking columns
+        considered as part of the 'remote' side.
+
+        """
+        if self._has_remote_annotations:
+            return
+
+        if self.secondary is not None:
+            self._annotate_remote_secondary()
+        elif self._local_remote_pairs or self._remote_side:
+            self._annotate_remote_from_args()
+        elif self._refers_to_parent_table():
+            self._annotate_selfref(
+                lambda col: "foreign" in col._annotations, False
+            )
+        elif self._tables_overlap():
+            self._annotate_remote_with_overlap()
+        else:
+            self._annotate_remote_distinct_selectables()
+
+    def _annotate_remote_secondary(self) -> None:
+        """annotate 'remote' in primaryjoin, secondaryjoin
+        when 'secondary' is present.
+
+        """
+
+        assert self.secondary is not None
+        fixed_secondary = self.secondary
+
+        def repl(element: _CE, **kw: Any) -> Optional[_CE]:
+            if fixed_secondary.c.contains_column(element):
+                return element._annotate({"remote": True})
+            return None
+
+        self.primaryjoin = visitors.replacement_traverse(
+            self.primaryjoin, {}, repl
+        )
+
+        assert self.secondaryjoin is not None
+        self.secondaryjoin = visitors.replacement_traverse(
+            self.secondaryjoin, {}, repl
+        )
+
+    def _annotate_selfref(
+        self, fn: Callable[[ColumnElement[Any]], bool], remote_side_given: bool
+    ) -> None:
+        """annotate 'remote' in primaryjoin, secondaryjoin
+        when the relationship is detected as self-referential.
+
+        """
+
+        def visit_binary(binary: BinaryExpression[Any]) -> None:
+            equated = binary.left.compare(binary.right)
+            if isinstance(binary.left, expression.ColumnClause) and isinstance(
+                binary.right, expression.ColumnClause
+            ):
+                # assume one to many - FKs are "remote"
+                if fn(binary.left):
+                    binary.left = binary.left._annotate({"remote": True})
+                if fn(binary.right) and not equated:
+                    binary.right = binary.right._annotate({"remote": True})
+            elif not remote_side_given:
+                self._warn_non_column_elements()
+
+        self.primaryjoin = visitors.cloned_traverse(
+            self.primaryjoin, {}, {"binary": visit_binary}
+        )
+
+    def _annotate_remote_from_args(self) -> None:
+        """annotate 'remote' in primaryjoin, secondaryjoin
+        when the 'remote_side' or '_local_remote_pairs'
+        arguments are used.
+
+        """
+        if self._local_remote_pairs:
+            if self._remote_side:
+                raise sa_exc.ArgumentError(
+                    "remote_side argument is redundant "
+                    "against more detailed _local_remote_side "
+                    "argument."
+                )
+
+            remote_side = [r for (l, r) in self._local_remote_pairs]
+        else:
+            remote_side = self._remote_side
+
+        if self._refers_to_parent_table():
+            self._annotate_selfref(lambda col: col in remote_side, True)
+        else:
+
+            def repl(element: _CE, **kw: Any) -> Optional[_CE]:
+                # use set() to avoid generating ``__eq__()`` expressions
+                # against each element
+                if element in set(remote_side):
+                    return element._annotate({"remote": True})
+                return None
+
+            self.primaryjoin = visitors.replacement_traverse(
+                self.primaryjoin, {}, repl
+            )
+
+    def _annotate_remote_with_overlap(self) -> None:
+        """annotate 'remote' in primaryjoin, secondaryjoin
+        when the parent/child tables have some set of
+        tables in common, though is not a fully self-referential
+        relationship.
+
+        """
+
+        def visit_binary(binary: BinaryExpression[Any]) -> None:
+            binary.left, binary.right = proc_left_right(
+                binary.left, binary.right
+            )
+            binary.right, binary.left = proc_left_right(
+                binary.right, binary.left
+            )
+
+        check_entities = (
+            self.prop is not None and self.prop.mapper is not self.prop.parent
+        )
+
+        def proc_left_right(
+            left: ColumnElement[Any], right: ColumnElement[Any]
+        ) -> Tuple[ColumnElement[Any], ColumnElement[Any]]:
+            if isinstance(left, expression.ColumnClause) and isinstance(
+                right, expression.ColumnClause
+            ):
+                if self.child_persist_selectable.c.contains_column(
+                    right
+                ) and self.parent_persist_selectable.c.contains_column(left):
+                    right = right._annotate({"remote": True})
+            elif (
+                check_entities
+                and right._annotations.get("parentmapper") is self.prop.mapper
+            ):
+                right = right._annotate({"remote": True})
+            elif (
+                check_entities
+                and left._annotations.get("parentmapper") is self.prop.mapper
+            ):
+                left = left._annotate({"remote": True})
+            else:
+                self._warn_non_column_elements()
+
+            return left, right
+
+        self.primaryjoin = visitors.cloned_traverse(
+            self.primaryjoin, {}, {"binary": visit_binary}
+        )
+
+    def _annotate_remote_distinct_selectables(self) -> None:
+        """annotate 'remote' in primaryjoin, secondaryjoin
+        when the parent/child tables are entirely
+        separate.
+
+        """
+
+        def repl(element: _CE, **kw: Any) -> Optional[_CE]:
+            if self.child_persist_selectable.c.contains_column(element) and (
+                not self.parent_local_selectable.c.contains_column(element)
+                or self.child_local_selectable.c.contains_column(element)
+            ):
+                return element._annotate({"remote": True})
+            return None
+
+        self.primaryjoin = visitors.replacement_traverse(
+            self.primaryjoin, {}, repl
+        )
+
+    def _warn_non_column_elements(self) -> None:
+        util.warn(
+            "Non-simple column elements in primary "
+            "join condition for property %s - consider using "
+            "remote() annotations to mark the remote side." % self.prop
+        )
+
+    def _annotate_local(self) -> None:
+        """Annotate the primaryjoin and secondaryjoin
+        structures with 'local' annotations.
+
+        This annotates all column elements found
+        simultaneously in the parent table
+        and the join condition that don't have a
+        'remote' annotation set up from
+        _annotate_remote() or user-defined.
+
+        """
+        if self._has_annotation(self.primaryjoin, "local"):
+            return
+
+        if self._local_remote_pairs:
+            local_side = util.column_set(
+                [l for (l, r) in self._local_remote_pairs]
+            )
+        else:
+            local_side = util.column_set(self.parent_persist_selectable.c)
+
+        def locals_(element: _CE, **kw: Any) -> Optional[_CE]:
+            if "remote" not in element._annotations and element in local_side:
+                return element._annotate({"local": True})
+            return None
+
+        self.primaryjoin = visitors.replacement_traverse(
+            self.primaryjoin, {}, locals_
+        )
+
+    def _annotate_parentmapper(self) -> None:
+        def parentmappers_(element: _CE, **kw: Any) -> Optional[_CE]:
+            if "remote" in element._annotations:
+                return element._annotate({"parentmapper": self.prop.mapper})
+            elif "local" in element._annotations:
+                return element._annotate({"parentmapper": self.prop.parent})
+            return None
+
+        self.primaryjoin = visitors.replacement_traverse(
+            self.primaryjoin, {}, parentmappers_
+        )
+
+    def _check_remote_side(self) -> None:
+        if not self.local_remote_pairs:
+            raise sa_exc.ArgumentError(
+                "Relationship %s could "
+                "not determine any unambiguous local/remote column "
+                "pairs based on join condition and remote_side "
+                "arguments.  "
+                "Consider using the remote() annotation to "
+                "accurately mark those elements of the join "
+                "condition that are on the remote side of "
+                "the relationship." % (self.prop,)
+            )
+        else:
+            not_target = util.column_set(
+                self.parent_persist_selectable.c
+            ).difference(self.child_persist_selectable.c)
+
+            for _, rmt in self.local_remote_pairs:
+                if rmt in not_target:
+                    util.warn(
+                        "Expression %s is marked as 'remote', but these "
+                        "column(s) are local to the local side.  The "
+                        "remote() annotation is needed only for a "
+                        "self-referential relationship where both sides "
+                        "of the relationship refer to the same tables."
+                        % (rmt,)
+                    )
+
+    def _check_foreign_cols(
+        self, join_condition: ColumnElement[bool], primary: bool
+    ) -> None:
+        """Check the foreign key columns collected and emit error
+        messages."""
+
+        can_sync = False
+
+        foreign_cols = self._gather_columns_with_annotation(
+            join_condition, "foreign"
+        )
+
+        has_foreign = bool(foreign_cols)
+
+        if primary:
+            can_sync = bool(self.synchronize_pairs)
+        else:
+            can_sync = bool(self.secondary_synchronize_pairs)
+
+        if (
+            self.support_sync
+            and can_sync
+            or (not self.support_sync and has_foreign)
+        ):
+            return
+
+        # from here below is just determining the best error message
+        # to report.  Check for a join condition using any operator
+        # (not just ==), perhaps they need to turn on "viewonly=True".
+        if self.support_sync and has_foreign and not can_sync:
+            err = (
+                "Could not locate any simple equality expressions "
+                "involving locally mapped foreign key columns for "
+                "%s join condition "
+                "'%s' on relationship %s."
+                % (
+                    primary and "primary" or "secondary",
+                    join_condition,
+                    self.prop,
+                )
+            )
+            err += (
+                "  Ensure that referencing columns are associated "
+                "with a ForeignKey or ForeignKeyConstraint, or are "
+                "annotated in the join condition with the foreign() "
+                "annotation. To allow comparison operators other than "
+                "'==', the relationship can be marked as viewonly=True."
+            )
+
+            raise sa_exc.ArgumentError(err)
+        else:
+            err = (
+                "Could not locate any relevant foreign key columns "
+                "for %s join condition '%s' on relationship %s."
+                % (
+                    primary and "primary" or "secondary",
+                    join_condition,
+                    self.prop,
+                )
+            )
+            err += (
+                "  Ensure that referencing columns are associated "
+                "with a ForeignKey or ForeignKeyConstraint, or are "
+                "annotated in the join condition with the foreign() "
+                "annotation."
+            )
+            raise sa_exc.ArgumentError(err)
+
+    def _determine_direction(self) -> None:
+        """Determine if this relationship is one to many, many to one,
+        many to many.
+
+        """
+        if self.secondaryjoin is not None:
+            self.direction = MANYTOMANY
+        else:
+            parentcols = util.column_set(self.parent_persist_selectable.c)
+            targetcols = util.column_set(self.child_persist_selectable.c)
+
+            # fk collection which suggests ONETOMANY.
+            onetomany_fk = targetcols.intersection(self.foreign_key_columns)
+
+            # fk collection which suggests MANYTOONE.
+
+            manytoone_fk = parentcols.intersection(self.foreign_key_columns)
+
+            if onetomany_fk and manytoone_fk:
+                # fks on both sides.  test for overlap of local/remote
+                # with foreign key.
+                # we will gather columns directly from their annotations
+                # without deannotating, so that we can distinguish on a column
+                # that refers to itself.
+
+                # 1. columns that are both remote and FK suggest
+                # onetomany.
+                onetomany_local = self._gather_columns_with_annotation(
+                    self.primaryjoin, "remote", "foreign"
+                )
+
+                # 2. columns that are FK but are not remote (e.g. local)
+                # suggest manytoone.
+                manytoone_local = {
+                    c
+                    for c in self._gather_columns_with_annotation(
+                        self.primaryjoin, "foreign"
+                    )
+                    if "remote" not in c._annotations
+                }
+
+                # 3. if both collections are present, remove columns that
+                # refer to themselves.  This is for the case of
+                # and_(Me.id == Me.remote_id, Me.version == Me.version)
+                if onetomany_local and manytoone_local:
+                    self_equated = self.remote_columns.intersection(
+                        self.local_columns
+                    )
+                    onetomany_local = onetomany_local.difference(self_equated)
+                    manytoone_local = manytoone_local.difference(self_equated)
+
+                # at this point, if only one or the other collection is
+                # present, we know the direction, otherwise it's still
+                # ambiguous.
+
+                if onetomany_local and not manytoone_local:
+                    self.direction = ONETOMANY
+                elif manytoone_local and not onetomany_local:
+                    self.direction = MANYTOONE
+                else:
+                    raise sa_exc.ArgumentError(
+                        "Can't determine relationship"
+                        " direction for relationship '%s' - foreign "
+                        "key columns within the join condition are present "
+                        "in both the parent and the child's mapped tables.  "
+                        "Ensure that only those columns referring "
+                        "to a parent column are marked as foreign, "
+                        "either via the foreign() annotation or "
+                        "via the foreign_keys argument." % self.prop
+                    )
+            elif onetomany_fk:
+                self.direction = ONETOMANY
+            elif manytoone_fk:
+                self.direction = MANYTOONE
+            else:
+                raise sa_exc.ArgumentError(
+                    "Can't determine relationship "
+                    "direction for relationship '%s' - foreign "
+                    "key columns are present in neither the parent "
+                    "nor the child's mapped tables" % self.prop
+                )
+
+    def _deannotate_pairs(
+        self, collection: _ColumnPairIterable
+    ) -> _MutableColumnPairs:
+        """provide deannotation for the various lists of
+        pairs, so that using them in hashes doesn't incur
+        high-overhead __eq__() comparisons against
+        original columns mapped.
+
+        """
+        return [(x._deannotate(), y._deannotate()) for x, y in collection]
+
+    def _setup_pairs(self) -> None:
+        sync_pairs: _MutableColumnPairs = []
+        lrp: util.OrderedSet[Tuple[ColumnElement[Any], ColumnElement[Any]]] = (
+            util.OrderedSet([])
+        )
+        secondary_sync_pairs: _MutableColumnPairs = []
+
+        def go(
+            joincond: ColumnElement[bool],
+            collection: _MutableColumnPairs,
+        ) -> None:
+            def visit_binary(
+                binary: BinaryExpression[Any],
+                left: ColumnElement[Any],
+                right: ColumnElement[Any],
+            ) -> None:
+                if (
+                    "remote" in right._annotations
+                    and "remote" not in left._annotations
+                    and self.can_be_synced_fn(left)
+                ):
+                    lrp.add((left, right))
+                elif (
+                    "remote" in left._annotations
+                    and "remote" not in right._annotations
+                    and self.can_be_synced_fn(right)
+                ):
+                    lrp.add((right, left))
+                if binary.operator is operators.eq and self.can_be_synced_fn(
+                    left, right
+                ):
+                    if "foreign" in right._annotations:
+                        collection.append((left, right))
+                    elif "foreign" in left._annotations:
+                        collection.append((right, left))
+
+            visit_binary_product(visit_binary, joincond)
+
+        for joincond, collection in [
+            (self.primaryjoin, sync_pairs),
+            (self.secondaryjoin, secondary_sync_pairs),
+        ]:
+            if joincond is None:
+                continue
+            go(joincond, collection)
+
+        self.local_remote_pairs = self._deannotate_pairs(lrp)
+        self.synchronize_pairs = self._deannotate_pairs(sync_pairs)
+        self.secondary_synchronize_pairs = self._deannotate_pairs(
+            secondary_sync_pairs
+        )
+
+    _track_overlapping_sync_targets: weakref.WeakKeyDictionary[
+        ColumnElement[Any],
+        weakref.WeakKeyDictionary[
+            RelationshipProperty[Any], ColumnElement[Any]
+        ],
+    ] = weakref.WeakKeyDictionary()
+
+    def _warn_for_conflicting_sync_targets(self) -> None:
+        if not self.support_sync:
+            return
+
+        # we would like to detect if we are synchronizing any column
+        # pairs in conflict with another relationship that wishes to sync
+        # an entirely different column to the same target.   This is a
+        # very rare edge case so we will try to minimize the memory/overhead
+        # impact of this check
+        for from_, to_ in [
+            (from_, to_) for (from_, to_) in self.synchronize_pairs
+        ] + [
+            (from_, to_) for (from_, to_) in self.secondary_synchronize_pairs
+        ]:
+            # save ourselves a ton of memory and overhead by only
+            # considering columns that are subject to a overlapping
+            # FK constraints at the core level.   This condition can arise
+            # if multiple relationships overlap foreign() directly, but
+            # we're going to assume it's typically a ForeignKeyConstraint-
+            # level configuration that benefits from this warning.
+
+            if to_ not in self._track_overlapping_sync_targets:
+                self._track_overlapping_sync_targets[to_] = (
+                    weakref.WeakKeyDictionary({self.prop: from_})
+                )
+            else:
+                other_props = []
+                prop_to_from = self._track_overlapping_sync_targets[to_]
+
+                for pr, fr_ in prop_to_from.items():
+                    if (
+                        not pr.mapper._dispose_called
+                        and pr not in self.prop._reverse_property
+                        and pr.key not in self.prop._overlaps
+                        and self.prop.key not in pr._overlaps
+                        # note: the "__*" symbol is used internally by
+                        # SQLAlchemy as a general means of suppressing the
+                        # overlaps warning for some extension cases, however
+                        # this is not currently
+                        # a publicly supported symbol and may change at
+                        # any time.
+                        and "__*" not in self.prop._overlaps
+                        and "__*" not in pr._overlaps
+                        and not self.prop.parent.is_sibling(pr.parent)
+                        and not self.prop.mapper.is_sibling(pr.mapper)
+                        and not self.prop.parent.is_sibling(pr.mapper)
+                        and not self.prop.mapper.is_sibling(pr.parent)
+                        and (
+                            self.prop.key != pr.key
+                            or not self.prop.parent.common_parent(pr.parent)
+                        )
+                    ):
+                        other_props.append((pr, fr_))
+
+                if other_props:
+                    util.warn(
+                        "relationship '%s' will copy column %s to column %s, "
+                        "which conflicts with relationship(s): %s. "
+                        "If this is not the intention, consider if these "
+                        "relationships should be linked with "
+                        "back_populates, or if viewonly=True should be "
+                        "applied to one or more if they are read-only. "
+                        "For the less common case that foreign key "
+                        "constraints are partially overlapping, the "
+                        "orm.foreign() "
+                        "annotation can be used to isolate the columns that "
+                        "should be written towards.   To silence this "
+                        "warning, add the parameter 'overlaps=\"%s\"' to the "
+                        "'%s' relationship."
+                        % (
+                            self.prop,
+                            from_,
+                            to_,
+                            ", ".join(
+                                sorted(
+                                    "'%s' (copies %s to %s)" % (pr, fr_, to_)
+                                    for (pr, fr_) in other_props
+                                )
+                            ),
+                            ",".join(sorted(pr.key for pr, fr in other_props)),
+                            self.prop,
+                        ),
+                        code="qzyx",
+                    )
+                self._track_overlapping_sync_targets[to_][self.prop] = from_
+
+    @util.memoized_property
+    def remote_columns(self) -> Set[ColumnElement[Any]]:
+        return self._gather_join_annotations("remote")
+
+    @util.memoized_property
+    def local_columns(self) -> Set[ColumnElement[Any]]:
+        return self._gather_join_annotations("local")
+
+    @util.memoized_property
+    def foreign_key_columns(self) -> Set[ColumnElement[Any]]:
+        return self._gather_join_annotations("foreign")
+
+    def _gather_join_annotations(
+        self, annotation: str
+    ) -> Set[ColumnElement[Any]]:
+        s = set(
+            self._gather_columns_with_annotation(self.primaryjoin, annotation)
+        )
+        if self.secondaryjoin is not None:
+            s.update(
+                self._gather_columns_with_annotation(
+                    self.secondaryjoin, annotation
+                )
+            )
+        return {x._deannotate() for x in s}
+
+    def _gather_columns_with_annotation(
+        self, clause: ColumnElement[Any], *annotation: Iterable[str]
+    ) -> Set[ColumnElement[Any]]:
+        annotation_set = set(annotation)
+        return {
+            cast(ColumnElement[Any], col)
+            for col in visitors.iterate(clause, {})
+            if annotation_set.issubset(col._annotations)
+        }
+
+    @util.memoized_property
+    def _secondary_lineage_set(self) -> FrozenSet[ColumnElement[Any]]:
+        if self.secondary is not None:
+            return frozenset(
+                itertools.chain(*[c.proxy_set for c in self.secondary.c])
+            )
+        else:
+            return util.EMPTY_SET
+
+    def join_targets(
+        self,
+        source_selectable: Optional[FromClause],
+        dest_selectable: FromClause,
+        aliased: bool,
+        single_crit: Optional[ColumnElement[bool]] = None,
+        extra_criteria: Tuple[ColumnElement[bool], ...] = (),
+    ) -> Tuple[
+        ColumnElement[bool],
+        Optional[ColumnElement[bool]],
+        Optional[FromClause],
+        Optional[ClauseAdapter],
+        FromClause,
+    ]:
+        """Given a source and destination selectable, create a
+        join between them.
+
+        This takes into account aliasing the join clause
+        to reference the appropriate corresponding columns
+        in the target objects, as well as the extra child
+        criterion, equivalent column sets, etc.
+
+        """
+        # place a barrier on the destination such that
+        # replacement traversals won't ever dig into it.
+        # its internal structure remains fixed
+        # regardless of context.
+        dest_selectable = _shallow_annotate(
+            dest_selectable, {"no_replacement_traverse": True}
+        )
+
+        primaryjoin, secondaryjoin, secondary = (
+            self.primaryjoin,
+            self.secondaryjoin,
+            self.secondary,
+        )
+
+        # adjust the join condition for single table inheritance,
+        # in the case that the join is to a subclass
+        # this is analogous to the
+        # "_adjust_for_single_table_inheritance()" method in Query.
+
+        if single_crit is not None:
+            if secondaryjoin is not None:
+                secondaryjoin = secondaryjoin & single_crit
+            else:
+                primaryjoin = primaryjoin & single_crit
+
+        if extra_criteria:
+
+            def mark_exclude_cols(
+                elem: SupportsAnnotations, annotations: _AnnotationDict
+            ) -> SupportsAnnotations:
+                """note unrelated columns in the "extra criteria" as either
+                should be adapted or not adapted, even though they are not
+                part of our "local" or "remote" side.
+
+                see #9779 for this case, as well as #11010 for a follow up
+
+                """
+
+                parentmapper_for_element = elem._annotations.get(
+                    "parentmapper", None
+                )
+
+                if (
+                    parentmapper_for_element is not self.prop.parent
+                    and parentmapper_for_element is not self.prop.mapper
+                    and elem not in self._secondary_lineage_set
+                ):
+                    return _safe_annotate(elem, annotations)
+                else:
+                    return elem
+
+            extra_criteria = tuple(
+                _deep_annotate(
+                    elem,
+                    {"should_not_adapt": True},
+                    annotate_callable=mark_exclude_cols,
+                )
+                for elem in extra_criteria
+            )
+
+            if secondaryjoin is not None:
+                secondaryjoin = secondaryjoin & sql.and_(*extra_criteria)
+            else:
+                primaryjoin = primaryjoin & sql.and_(*extra_criteria)
+
+        if aliased:
+            if secondary is not None:
+                secondary = secondary._anonymous_fromclause(flat=True)
+                primary_aliasizer = ClauseAdapter(
+                    secondary,
+                    exclude_fn=_local_col_exclude,
+                )
+                secondary_aliasizer = ClauseAdapter(
+                    dest_selectable, equivalents=self.child_equivalents
+                ).chain(primary_aliasizer)
+                if source_selectable is not None:
+                    primary_aliasizer = ClauseAdapter(
+                        secondary,
+                        exclude_fn=_local_col_exclude,
+                    ).chain(
+                        ClauseAdapter(
+                            source_selectable,
+                            equivalents=self.parent_equivalents,
+                        )
+                    )
+
+                secondaryjoin = secondary_aliasizer.traverse(secondaryjoin)
+            else:
+                primary_aliasizer = ClauseAdapter(
+                    dest_selectable,
+                    exclude_fn=_local_col_exclude,
+                    equivalents=self.child_equivalents,
+                )
+                if source_selectable is not None:
+                    primary_aliasizer.chain(
+                        ClauseAdapter(
+                            source_selectable,
+                            exclude_fn=_remote_col_exclude,
+                            equivalents=self.parent_equivalents,
+                        )
+                    )
+                secondary_aliasizer = None
+
+            primaryjoin = primary_aliasizer.traverse(primaryjoin)
+            target_adapter = secondary_aliasizer or primary_aliasizer
+            target_adapter.exclude_fn = None
+        else:
+            target_adapter = None
+        return (
+            primaryjoin,
+            secondaryjoin,
+            secondary,
+            target_adapter,
+            dest_selectable,
+        )
+
+    def create_lazy_clause(self, reverse_direction: bool = False) -> Tuple[
+        ColumnElement[bool],
+        Dict[str, ColumnElement[Any]],
+        Dict[ColumnElement[Any], ColumnElement[Any]],
+    ]:
+        binds: Dict[ColumnElement[Any], BindParameter[Any]] = {}
+        equated_columns: Dict[ColumnElement[Any], ColumnElement[Any]] = {}
+
+        has_secondary = self.secondaryjoin is not None
+
+        if has_secondary:
+            lookup = collections.defaultdict(list)
+            for l, r in self.local_remote_pairs:
+                lookup[l].append((l, r))
+                equated_columns[r] = l
+        elif not reverse_direction:
+            for l, r in self.local_remote_pairs:
+                equated_columns[r] = l
+        else:
+            for l, r in self.local_remote_pairs:
+                equated_columns[l] = r
+
+        def col_to_bind(
+            element: ColumnElement[Any], **kw: Any
+        ) -> Optional[BindParameter[Any]]:
+            if (
+                (not reverse_direction and "local" in element._annotations)
+                or reverse_direction
+                and (
+                    (has_secondary and element in lookup)
+                    or (not has_secondary and "remote" in element._annotations)
+                )
+            ):
+                if element not in binds:
+                    binds[element] = sql.bindparam(
+                        None, None, type_=element.type, unique=True
+                    )
+                return binds[element]
+            return None
+
+        lazywhere = self.primaryjoin
+        if self.secondaryjoin is None or not reverse_direction:
+            lazywhere = visitors.replacement_traverse(
+                lazywhere, {}, col_to_bind
+            )
+
+        if self.secondaryjoin is not None:
+            secondaryjoin = self.secondaryjoin
+            if reverse_direction:
+                secondaryjoin = visitors.replacement_traverse(
+                    secondaryjoin, {}, col_to_bind
+                )
+            lazywhere = sql.and_(lazywhere, secondaryjoin)
+
+        bind_to_col = {binds[col].key: col for col in binds}
+
+        return lazywhere, bind_to_col, equated_columns
+
+
+class _ColInAnnotations:
+    """Serializable object that tests for names in c._annotations.
+
+    TODO: does this need to be serializable anymore?  can we find what the
+    use case was for that?
+
+    """
+
+    __slots__ = ("names",)
+
+    def __init__(self, *names: str):
+        self.names = frozenset(names)
+
+    def __call__(self, c: ClauseElement) -> bool:
+        return bool(self.names.intersection(c._annotations))
+
+
+_local_col_exclude = _ColInAnnotations("local", "should_not_adapt")
+_remote_col_exclude = _ColInAnnotations("remote", "should_not_adapt")
+
+
+class Relationship(
+    RelationshipProperty[_T],
+    _DeclarativeMapped[_T],
+):
+    """Describes an object property that holds a single item or list
+    of items that correspond to a related database table.
+
+    Public constructor is the :func:`_orm.relationship` function.
+
+    .. seealso::
+
+        :ref:`relationship_config_toplevel`
+
+    .. versionchanged:: 2.0 Added :class:`_orm.Relationship` as a Declarative
+       compatible subclass for :class:`_orm.RelationshipProperty`.
+
+    """
+
+    inherit_cache = True
+    """:meta private:"""
+
+
+class _RelationshipDeclared(  # type: ignore[misc]
+    Relationship[_T],
+    WriteOnlyMapped[_T],  # not compatible with Mapped[_T]
+    DynamicMapped[_T],  # not compatible with Mapped[_T]
+):
+    """Relationship subclass used implicitly for declarative mapping."""
+
+    inherit_cache = True
+    """:meta private:"""
+
+    @classmethod
+    def _mapper_property_name(cls) -> str:
+        return "Relationship"
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/scoping.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/scoping.py
new file mode 100644
index 00000000..a0e9f17e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/scoping.py
@@ -0,0 +1,2163 @@
+# orm/scoping.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .session import _S
+from .session import Session
+from .. import exc as sa_exc
+from .. import util
+from ..util import create_proxy_methods
+from ..util import ScopedRegistry
+from ..util import ThreadLocalRegistry
+from ..util import warn
+from ..util import warn_deprecated
+from ..util.typing import Protocol
+
+if TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _IdentityKeyType
+    from ._typing import OrmExecuteOptionsParameter
+    from .identity import IdentityMap
+    from .interfaces import ORMOption
+    from .mapper import Mapper
+    from .query import Query
+    from .query import RowReturningQuery
+    from .session import _BindArguments
+    from .session import _EntityBindKey
+    from .session import _PKIdentityArgument
+    from .session import _SessionBind
+    from .session import sessionmaker
+    from .session import SessionTransaction
+    from ..engine import Connection
+    from ..engine import CursorResult
+    from ..engine import Engine
+    from ..engine import Result
+    from ..engine import Row
+    from ..engine import RowMapping
+    from ..engine.interfaces import _CoreAnyExecuteParams
+    from ..engine.interfaces import _CoreSingleExecuteParams
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..engine.result import ScalarResult
+    from ..sql._typing import _ColumnsClauseArgument
+    from ..sql._typing import _T0
+    from ..sql._typing import _T1
+    from ..sql._typing import _T2
+    from ..sql._typing import _T3
+    from ..sql._typing import _T4
+    from ..sql._typing import _T5
+    from ..sql._typing import _T6
+    from ..sql._typing import _T7
+    from ..sql._typing import _TypedColumnClauseArgument as _TCCA
+    from ..sql.base import Executable
+    from ..sql.dml import UpdateBase
+    from ..sql.elements import ClauseElement
+    from ..sql.roles import TypedColumnsClauseRole
+    from ..sql.selectable import ForUpdateParameter
+    from ..sql.selectable import TypedReturnsRows
+
+_T = TypeVar("_T", bound=Any)
+
+
+class QueryPropertyDescriptor(Protocol):
+    """Describes the type applied to a class-level
+    :meth:`_orm.scoped_session.query_property` attribute.
+
+    .. versionadded:: 2.0.5
+
+    """
+
+    def __get__(self, instance: Any, owner: Type[_T]) -> Query[_T]: ...
+
+
+_O = TypeVar("_O", bound=object)
+
+__all__ = ["scoped_session"]
+
+
+@create_proxy_methods(
+    Session,
+    ":class:`_orm.Session`",
+    ":class:`_orm.scoping.scoped_session`",
+    classmethods=["close_all", "object_session", "identity_key"],
+    methods=[
+        "__contains__",
+        "__iter__",
+        "add",
+        "add_all",
+        "begin",
+        "begin_nested",
+        "close",
+        "reset",
+        "commit",
+        "connection",
+        "delete",
+        "execute",
+        "expire",
+        "expire_all",
+        "expunge",
+        "expunge_all",
+        "flush",
+        "get",
+        "get_one",
+        "get_bind",
+        "is_modified",
+        "bulk_save_objects",
+        "bulk_insert_mappings",
+        "bulk_update_mappings",
+        "merge",
+        "query",
+        "refresh",
+        "rollback",
+        "scalar",
+        "scalars",
+    ],
+    attributes=[
+        "bind",
+        "dirty",
+        "deleted",
+        "new",
+        "identity_map",
+        "is_active",
+        "autoflush",
+        "no_autoflush",
+        "info",
+    ],
+)
+class scoped_session(Generic[_S]):
+    """Provides scoped management of :class:`.Session` objects.
+
+    See :ref:`unitofwork_contextual` for a tutorial.
+
+    .. note::
+
+       When using :ref:`asyncio_toplevel`, the async-compatible
+       :class:`_asyncio.async_scoped_session` class should be
+       used in place of :class:`.scoped_session`.
+
+    """
+
+    _support_async: bool = False
+
+    session_factory: sessionmaker[_S]
+    """The `session_factory` provided to `__init__` is stored in this
+    attribute and may be accessed at a later time.  This can be useful when
+    a new non-scoped :class:`.Session` is needed."""
+
+    registry: ScopedRegistry[_S]
+
+    def __init__(
+        self,
+        session_factory: sessionmaker[_S],
+        scopefunc: Optional[Callable[[], Any]] = None,
+    ):
+        """Construct a new :class:`.scoped_session`.
+
+        :param session_factory: a factory to create new :class:`.Session`
+         instances. This is usually, but not necessarily, an instance
+         of :class:`.sessionmaker`.
+        :param scopefunc: optional function which defines
+         the current scope.   If not passed, the :class:`.scoped_session`
+         object assumes "thread-local" scope, and will use
+         a Python ``threading.local()`` in order to maintain the current
+         :class:`.Session`.  If passed, the function should return
+         a hashable token; this token will be used as the key in a
+         dictionary in order to store and retrieve the current
+         :class:`.Session`.
+
+        """
+        self.session_factory = session_factory
+
+        if scopefunc:
+            self.registry = ScopedRegistry(session_factory, scopefunc)
+        else:
+            self.registry = ThreadLocalRegistry(session_factory)
+
+    @property
+    def _proxied(self) -> _S:
+        return self.registry()
+
+    def __call__(self, **kw: Any) -> _S:
+        r"""Return the current :class:`.Session`, creating it
+        using the :attr:`.scoped_session.session_factory` if not present.
+
+        :param \**kw: Keyword arguments will be passed to the
+         :attr:`.scoped_session.session_factory` callable, if an existing
+         :class:`.Session` is not present.  If the :class:`.Session` is present
+         and keyword arguments have been passed,
+         :exc:`~sqlalchemy.exc.InvalidRequestError` is raised.
+
+        """
+        if kw:
+            if self.registry.has():
+                raise sa_exc.InvalidRequestError(
+                    "Scoped session is already present; "
+                    "no new arguments may be specified."
+                )
+            else:
+                sess = self.session_factory(**kw)
+                self.registry.set(sess)
+        else:
+            sess = self.registry()
+        if not self._support_async and sess._is_asyncio:
+            warn_deprecated(
+                "Using `scoped_session` with asyncio is deprecated and "
+                "will raise an error in a future version. "
+                "Please use `async_scoped_session` instead.",
+                "1.4.23",
+            )
+        return sess
+
+    def configure(self, **kwargs: Any) -> None:
+        """reconfigure the :class:`.sessionmaker` used by this
+        :class:`.scoped_session`.
+
+        See :meth:`.sessionmaker.configure`.
+
+        """
+
+        if self.registry.has():
+            warn(
+                "At least one scoped session is already present. "
+                " configure() can not affect sessions that have "
+                "already been created."
+            )
+
+        self.session_factory.configure(**kwargs)
+
+    def remove(self) -> None:
+        """Dispose of the current :class:`.Session`, if present.
+
+        This will first call :meth:`.Session.close` method
+        on the current :class:`.Session`, which releases any existing
+        transactional/connection resources still being held; transactions
+        specifically are rolled back.  The :class:`.Session` is then
+        discarded.   Upon next usage within the same scope,
+        the :class:`.scoped_session` will produce a new
+        :class:`.Session` object.
+
+        """
+
+        if self.registry.has():
+            self.registry().close()
+        self.registry.clear()
+
+    def query_property(
+        self, query_cls: Optional[Type[Query[_T]]] = None
+    ) -> QueryPropertyDescriptor:
+        """return a class property which produces a legacy
+        :class:`_query.Query` object against the class and the current
+        :class:`.Session` when called.
+
+        .. legacy:: The :meth:`_orm.scoped_session.query_property` accessor
+           is specific to the legacy :class:`.Query` object and is not
+           considered to be part of :term:`2.0-style` ORM use.
+
+        e.g.::
+
+            from sqlalchemy.orm import QueryPropertyDescriptor
+            from sqlalchemy.orm import scoped_session
+            from sqlalchemy.orm import sessionmaker
+
+            Session = scoped_session(sessionmaker())
+
+
+            class MyClass:
+                query: QueryPropertyDescriptor = Session.query_property()
+
+
+            # after mappers are defined
+            result = MyClass.query.filter(MyClass.name == "foo").all()
+
+        Produces instances of the session's configured query class by
+        default.  To override and use a custom implementation, provide
+        a ``query_cls`` callable.  The callable will be invoked with
+        the class's mapper as a positional argument and a session
+        keyword argument.
+
+        There is no limit to the number of query properties placed on
+        a class.
+
+        """
+
+        class query:
+            def __get__(s, instance: Any, owner: Type[_O]) -> Query[_O]:
+                if query_cls:
+                    # custom query class
+                    return query_cls(owner, session=self.registry())  # type: ignore  # noqa: E501
+                else:
+                    # session's configured query class
+                    return self.registry().query(owner)
+
+        return query()
+
+    # START PROXY METHODS scoped_session
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_proxy_methods.py
+
+    def __contains__(self, instance: object) -> bool:
+        r"""Return True if the instance is associated with this session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The instance may be pending or persistent within the Session for a
+        result of True.
+
+
+        """  # noqa: E501
+
+        return self._proxied.__contains__(instance)
+
+    def __iter__(self) -> Iterator[object]:
+        r"""Iterate over all pending or persistent instances within this
+        Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+
+        """  # noqa: E501
+
+        return self._proxied.__iter__()
+
+    def add(self, instance: object, _warn: bool = True) -> None:
+        r"""Place an object into this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Objects that are in the :term:`transient` state when passed to the
+        :meth:`_orm.Session.add` method will move to the
+        :term:`pending` state, until the next flush, at which point they
+        will move to the :term:`persistent` state.
+
+        Objects that are in the :term:`detached` state when passed to the
+        :meth:`_orm.Session.add` method will move to the :term:`persistent`
+        state directly.
+
+        If the transaction used by the :class:`_orm.Session` is rolled back,
+        objects which were transient when they were passed to
+        :meth:`_orm.Session.add` will be moved back to the
+        :term:`transient` state, and will no longer be present within this
+        :class:`_orm.Session`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add_all`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+        """  # noqa: E501
+
+        return self._proxied.add(instance, _warn=_warn)
+
+    def add_all(self, instances: Iterable[object]) -> None:
+        r"""Add the given collection of instances to this :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        See the documentation for :meth:`_orm.Session.add` for a general
+        behavioral description.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+
+        """  # noqa: E501
+
+        return self._proxied.add_all(instances)
+
+    def begin(self, nested: bool = False) -> SessionTransaction:
+        r"""Begin a transaction, or nested transaction,
+        on this :class:`.Session`, if one is not already begun.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The :class:`_orm.Session` object features **autobegin** behavior,
+        so that normally it is not necessary to call the
+        :meth:`_orm.Session.begin`
+        method explicitly. However, it may be used in order to control
+        the scope of when the transactional state is begun.
+
+        When used to begin the outermost transaction, an error is raised
+        if this :class:`.Session` is already inside of a transaction.
+
+        :param nested: if True, begins a SAVEPOINT transaction and is
+         equivalent to calling :meth:`~.Session.begin_nested`. For
+         documentation on SAVEPOINT transactions, please see
+         :ref:`session_begin_nested`.
+
+        :return: the :class:`.SessionTransaction` object.  Note that
+         :class:`.SessionTransaction`
+         acts as a Python context manager, allowing :meth:`.Session.begin`
+         to be used in a "with" block.  See :ref:`session_explicit_begin` for
+         an example.
+
+        .. seealso::
+
+            :ref:`session_autobegin`
+
+            :ref:`unitofwork_transaction`
+
+            :meth:`.Session.begin_nested`
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.begin(nested=nested)
+
+    def begin_nested(self) -> SessionTransaction:
+        r"""Begin a "nested" transaction on this Session, e.g. SAVEPOINT.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The target database(s) and associated drivers must support SQL
+        SAVEPOINT for this method to function correctly.
+
+        For documentation on SAVEPOINT
+        transactions, please see :ref:`session_begin_nested`.
+
+        :return: the :class:`.SessionTransaction` object.  Note that
+         :class:`.SessionTransaction` acts as a context manager, allowing
+         :meth:`.Session.begin_nested` to be used in a "with" block.
+         See :ref:`session_begin_nested` for a usage example.
+
+        .. seealso::
+
+            :ref:`session_begin_nested`
+
+            :ref:`pysqlite_serializable` - special workarounds required
+            with the SQLite driver in order for SAVEPOINT to work
+            correctly. For asyncio use cases, see the section
+            :ref:`aiosqlite_serializable`.
+
+
+        """  # noqa: E501
+
+        return self._proxied.begin_nested()
+
+    def close(self) -> None:
+        r"""Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This expunges all ORM objects associated with this
+        :class:`_orm.Session`, ends any transaction in progress and
+        :term:`releases` any :class:`_engine.Connection` objects which this
+        :class:`_orm.Session` itself has checked out from associated
+        :class:`_engine.Engine` objects. The operation then leaves the
+        :class:`_orm.Session` in a state which it may be used again.
+
+        .. tip::
+
+            In the default running mode the :meth:`_orm.Session.close`
+            method **does not prevent the Session from being used again**.
+            The :class:`_orm.Session` itself does not actually have a
+            distinct "closed" state; it merely means
+            the :class:`_orm.Session` will release all database connections
+            and ORM objects.
+
+            Setting the parameter :paramref:`_orm.Session.close_resets_only`
+            to ``False`` will instead make the ``close`` final, meaning that
+            any further action on the session will be forbidden.
+
+        .. versionchanged:: 1.4  The :meth:`.Session.close` method does not
+           immediately create a new :class:`.SessionTransaction` object;
+           instead, the new :class:`.SessionTransaction` is created only if
+           the :class:`.Session` is used again for a database operation.
+
+        .. seealso::
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_orm.Session.close` and :meth:`_orm.Session.reset`.
+
+            :meth:`_orm.Session.reset` - a similar method that behaves like
+            ``close()`` with  the parameter
+            :paramref:`_orm.Session.close_resets_only` set to ``True``.
+
+
+        """  # noqa: E501
+
+        return self._proxied.close()
+
+    def reset(self) -> None:
+        r"""Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`, resetting the session to its initial state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This method provides for same "reset-only" behavior that the
+        :meth:`_orm.Session.close` method has provided historically, where the
+        state of the :class:`_orm.Session` is reset as though the object were
+        brand new, and ready to be used again.
+        This method may then be useful for :class:`_orm.Session` objects
+        which set :paramref:`_orm.Session.close_resets_only` to ``False``,
+        so that "reset only" behavior is still available.
+
+        .. versionadded:: 2.0.22
+
+        .. seealso::
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_orm.Session.close` and :meth:`_orm.Session.reset`.
+
+            :meth:`_orm.Session.close` - a similar method will additionally
+            prevent re-use of the Session when the parameter
+            :paramref:`_orm.Session.close_resets_only` is set to ``False``.
+
+        """  # noqa: E501
+
+        return self._proxied.reset()
+
+    def commit(self) -> None:
+        r"""Flush pending changes and commit the current transaction.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        When the COMMIT operation is complete, all objects are fully
+        :term:`expired`, erasing their internal contents, which will be
+        automatically re-loaded when the objects are next accessed. In the
+        interim, these objects are in an expired state and will not function if
+        they are :term:`detached` from the :class:`.Session`. Additionally,
+        this re-load operation is not supported when using asyncio-oriented
+        APIs. The :paramref:`.Session.expire_on_commit` parameter may be used
+        to disable this behavior.
+
+        When there is no transaction in place for the :class:`.Session`,
+        indicating that no operations were invoked on this :class:`.Session`
+        since the previous call to :meth:`.Session.commit`, the method will
+        begin and commit an internal-only "logical" transaction, that does not
+        normally affect the database unless pending flush changes were
+        detected, but will still invoke event handlers and object expiration
+        rules.
+
+        The outermost database transaction is committed unconditionally,
+        automatically releasing any SAVEPOINTs in effect.
+
+        .. seealso::
+
+            :ref:`session_committing`
+
+            :ref:`unitofwork_transaction`
+
+            :ref:`asyncio_orm_avoid_lazyloads`
+
+
+        """  # noqa: E501
+
+        return self._proxied.commit()
+
+    def connection(
+        self,
+        bind_arguments: Optional[_BindArguments] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Connection:
+        r"""Return a :class:`_engine.Connection` object corresponding to this
+        :class:`.Session` object's transactional state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Either the :class:`_engine.Connection` corresponding to the current
+        transaction is returned, or if no transaction is in progress, a new
+        one is begun and the :class:`_engine.Connection`
+        returned (note that no
+        transactional state is established with the DBAPI until the first
+        SQL statement is emitted).
+
+        Ambiguity in multi-bind or unbound :class:`.Session` objects can be
+        resolved through any of the optional keyword arguments.   This
+        ultimately makes usage of the :meth:`.get_bind` method for resolution.
+
+        :param bind_arguments: dictionary of bind arguments.  May include
+         "mapper", "bind", "clause", other custom arguments that are passed
+         to :meth:`.Session.get_bind`.
+
+        :param execution_options: a dictionary of execution options that will
+         be passed to :meth:`_engine.Connection.execution_options`, **when the
+         connection is first procured only**.   If the connection is already
+         present within the :class:`.Session`, a warning is emitted and
+         the arguments are ignored.
+
+         .. seealso::
+
+            :ref:`session_transaction_isolation`
+
+
+        """  # noqa: E501
+
+        return self._proxied.connection(
+            bind_arguments=bind_arguments, execution_options=execution_options
+        )
+
+    def delete(self, instance: object) -> None:
+        r"""Mark an instance as deleted.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The object is assumed to be either :term:`persistent` or
+        :term:`detached` when passed; after the method is called, the
+        object will remain in the :term:`persistent` state until the next
+        flush proceeds.  During this time, the object will also be a member
+        of the :attr:`_orm.Session.deleted` collection.
+
+        When the next flush proceeds, the object will move to the
+        :term:`deleted` state, indicating a ``DELETE`` statement was emitted
+        for its row within the current transaction.   When the transaction
+        is successfully committed,
+        the deleted object is moved to the :term:`detached` state and is
+        no longer present within this :class:`_orm.Session`.
+
+        .. seealso::
+
+            :ref:`session_deleting` - at :ref:`session_basics`
+
+
+        """  # noqa: E501
+
+        return self._proxied.delete(instance)
+
+    @overload
+    def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[_T]: ...
+
+    @overload
+    def execute(
+        self,
+        statement: UpdateBase,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> CursorResult[Any]: ...
+
+    @overload
+    def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]: ...
+
+    def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]:
+        r"""Execute a SQL expression construct.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Returns a :class:`_engine.Result` object representing
+        results of the statement execution.
+
+        E.g.::
+
+            from sqlalchemy import select
+
+            result = session.execute(select(User).where(User.id == 5))
+
+        The API contract of :meth:`_orm.Session.execute` is similar to that
+        of :meth:`_engine.Connection.execute`, the :term:`2.0 style` version
+        of :class:`_engine.Connection`.
+
+        .. versionchanged:: 1.4 the :meth:`_orm.Session.execute` method is
+           now the primary point of ORM statement execution when using
+           :term:`2.0 style` ORM usage.
+
+        :param statement:
+            An executable statement (i.e. an :class:`.Executable` expression
+            such as :func:`_expression.select`).
+
+        :param params:
+            Optional dictionary, or list of dictionaries, containing
+            bound parameter values.   If a single dictionary, single-row
+            execution occurs; if a list of dictionaries, an
+            "executemany" will be invoked.  The keys in each dictionary
+            must correspond to parameter names present in the statement.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the statement execution.  This
+         dictionary can provide a subset of the options that are accepted
+         by :meth:`_engine.Connection.execution_options`, and may also
+         provide additional options understood only in an ORM context.
+
+         .. seealso::
+
+            :ref:`orm_queryguide_execution_options` - ORM-specific execution
+            options
+
+        :param bind_arguments: dictionary of additional arguments to determine
+         the bind.  May include "mapper", "bind", or other custom arguments.
+         Contents of this dictionary are passed to the
+         :meth:`.Session.get_bind` method.
+
+        :return: a :class:`_engine.Result` object.
+
+
+
+        """  # noqa: E501
+
+        return self._proxied.execute(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            _parent_execute_state=_parent_execute_state,
+            _add_event=_add_event,
+        )
+
+    def expire(
+        self, instance: object, attribute_names: Optional[Iterable[str]] = None
+    ) -> None:
+        r"""Expire the attributes on an instance.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Marks the attributes of an instance as out of date. When an expired
+        attribute is next accessed, a query will be issued to the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire all objects in the :class:`.Session` simultaneously,
+        use :meth:`Session.expire_all`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire` only makes sense for the specific
+        case that a non-ORM SQL statement was emitted in the current
+        transaction.
+
+        :param instance: The instance to be refreshed.
+        :param attribute_names: optional list of string attribute names
+          indicating a subset of attributes to be expired.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire(instance, attribute_names=attribute_names)
+
+    def expire_all(self) -> None:
+        r"""Expires all persistent instances within this Session.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        When any attributes on a persistent instance is next accessed,
+        a query will be issued using the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire individual objects and individual attributes
+        on those objects, use :meth:`Session.expire`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire_all` is not usually needed,
+        assuming the transaction is isolated.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+
+        """  # noqa: E501
+
+        return self._proxied.expire_all()
+
+    def expunge(self, instance: object) -> None:
+        r"""Remove the `instance` from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This will free all internal references to the instance.  Cascading
+        will be applied according to the *expunge* cascade rule.
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge(instance)
+
+    def expunge_all(self) -> None:
+        r"""Remove all object instances from this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This is equivalent to calling ``expunge(obj)`` on all objects in this
+        ``Session``.
+
+
+        """  # noqa: E501
+
+        return self._proxied.expunge_all()
+
+    def flush(self, objects: Optional[Sequence[Any]] = None) -> None:
+        r"""Flush all the object changes to the database.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Writes out all pending object creations, deletions and modifications
+        to the database as INSERTs, DELETEs, UPDATEs, etc.  Operations are
+        automatically ordered by the Session's unit of work dependency
+        solver.
+
+        Database operations will be issued in the current transactional
+        context and do not affect the state of the transaction, unless an
+        error occurs, in which case the entire transaction is rolled back.
+        You may flush() as often as you like within a transaction to move
+        changes from Python to the database's transaction buffer.
+
+        :param objects: Optional; restricts the flush operation to operate
+          only on elements that are in the given collection.
+
+          This feature is for an extremely narrow set of use cases where
+          particular objects may need to be operated upon before the
+          full flush() occurs.  It is not intended for general use.
+
+
+        """  # noqa: E501
+
+        return self._proxied.flush(objects=objects)
+
+    def get(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> Optional[_O]:
+        r"""Return an instance based on the given primary key identifier,
+        or ``None`` if not found.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        E.g.::
+
+            my_user = session.get(User, 5)
+
+            some_object = session.get(VersionedFoo, (5, 10))
+
+            some_object = session.get(VersionedFoo, {"id": 5, "version_id": 10})
+
+        .. versionadded:: 1.4 Added :meth:`_orm.Session.get`, which is moved
+           from the now legacy :meth:`_orm.Query.get` method.
+
+        :meth:`_orm.Session.get` is special in that it provides direct
+        access to the identity map of the :class:`.Session`.
+        If the given primary key identifier is present
+        in the local identity map, the object is returned
+        directly from this collection and no SQL is emitted,
+        unless the object has been marked fully expired.
+        If not present,
+        a SELECT is performed in order to locate the object.
+
+        :meth:`_orm.Session.get` also will perform a check if
+        the object is present in the identity map and
+        marked as expired - a SELECT
+        is emitted to refresh the object as well as to
+        ensure that the row is still present.
+        If not, :class:`~sqlalchemy.orm.exc.ObjectDeletedError` is raised.
+
+        :param entity: a mapped class or :class:`.Mapper` indicating the
+         type of entity to be loaded.
+
+        :param ident: A scalar, tuple, or dictionary representing the
+         primary key.  For a composite (e.g. multiple column) primary key,
+         a tuple or dictionary should be passed.
+
+         For a single-column primary key, the scalar calling form is typically
+         the most expedient.  If the primary key of a row is the value "5",
+         the call looks like::
+
+            my_object = session.get(SomeClass, 5)
+
+         The tuple form contains primary key values typically in
+         the order in which they correspond to the mapped
+         :class:`_schema.Table`
+         object's primary key columns, or if the
+         :paramref:`_orm.Mapper.primary_key` configuration parameter were
+         used, in
+         the order used for that parameter. For example, if the primary key
+         of a row is represented by the integer
+         digits "5, 10" the call would look like::
+
+             my_object = session.get(SomeClass, (5, 10))
+
+         The dictionary form should include as keys the mapped attribute names
+         corresponding to each element of the primary key.  If the mapped class
+         has the attributes ``id``, ``version_id`` as the attributes which
+         store the object's primary key value, the call would look like::
+
+            my_object = session.get(SomeClass, {"id": 5, "version_id": 10})
+
+        :param options: optional sequence of loader options which will be
+         applied to the query, if one is emitted.
+
+        :param populate_existing: causes the method to unconditionally emit
+         a SQL query and refresh the object with the newly loaded data,
+         regardless of whether or not the object is already present.
+
+        :param with_for_update: optional boolean ``True`` indicating FOR UPDATE
+          should be used, or may be a dictionary containing flags to
+          indicate a more specific set of FOR UPDATE flags for the SELECT;
+          flags should match the parameters of
+          :meth:`_query.Query.with_for_update`.
+          Supersedes the :paramref:`.Session.refresh.lockmode` parameter.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the query execution if one is emitted.
+         This dictionary can provide a subset of the options that are
+         accepted by :meth:`_engine.Connection.execution_options`, and may
+         also provide additional options understood only in an ORM context.
+
+         .. versionadded:: 1.4.29
+
+         .. seealso::
+
+            :ref:`orm_queryguide_execution_options` - ORM-specific execution
+            options
+
+        :param bind_arguments: dictionary of additional arguments to determine
+         the bind.  May include "mapper", "bind", or other custom arguments.
+         Contents of this dictionary are passed to the
+         :meth:`.Session.get_bind` method.
+
+         .. versionadded: 2.0.0rc1
+
+        :return: The object instance, or ``None``.
+
+
+        """  # noqa: E501
+
+        return self._proxied.get(
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+
+    def get_one(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> _O:
+        r"""Return exactly one instance based on the given primary key
+        identifier, or raise an exception if not found.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Raises ``sqlalchemy.orm.exc.NoResultFound`` if the query
+        selects no rows.
+
+        For a detailed documentation of the arguments see the
+        method :meth:`.Session.get`.
+
+        .. versionadded:: 2.0.22
+
+        :return: The object instance.
+
+        .. seealso::
+
+            :meth:`.Session.get` - equivalent method that instead
+              returns ``None`` if no row was found with the provided primary
+              key
+
+
+        """  # noqa: E501
+
+        return self._proxied.get_one(
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+
+    def get_bind(
+        self,
+        mapper: Optional[_EntityBindKey[_O]] = None,
+        *,
+        clause: Optional[ClauseElement] = None,
+        bind: Optional[_SessionBind] = None,
+        _sa_skip_events: Optional[bool] = None,
+        _sa_skip_for_implicit_returning: bool = False,
+        **kw: Any,
+    ) -> Union[Engine, Connection]:
+        r"""Return a "bind" to which this :class:`.Session` is bound.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The "bind" is usually an instance of :class:`_engine.Engine`,
+        except in the case where the :class:`.Session` has been
+        explicitly bound directly to a :class:`_engine.Connection`.
+
+        For a multiply-bound or unbound :class:`.Session`, the
+        ``mapper`` or ``clause`` arguments are used to determine the
+        appropriate bind to return.
+
+        Note that the "mapper" argument is usually present
+        when :meth:`.Session.get_bind` is called via an ORM
+        operation such as a :meth:`.Session.query`, each
+        individual INSERT/UPDATE/DELETE operation within a
+        :meth:`.Session.flush`, call, etc.
+
+        The order of resolution is:
+
+        1. if mapper given and :paramref:`.Session.binds` is present,
+           locate a bind based first on the mapper in use, then
+           on the mapped class in use, then on any base classes that are
+           present in the ``__mro__`` of the mapped class, from more specific
+           superclasses to more general.
+        2. if clause given and ``Session.binds`` is present,
+           locate a bind based on :class:`_schema.Table` objects
+           found in the given clause present in ``Session.binds``.
+        3. if ``Session.binds`` is present, return that.
+        4. if clause given, attempt to return a bind
+           linked to the :class:`_schema.MetaData` ultimately
+           associated with the clause.
+        5. if mapper given, attempt to return a bind
+           linked to the :class:`_schema.MetaData` ultimately
+           associated with the :class:`_schema.Table` or other
+           selectable to which the mapper is mapped.
+        6. No bind can be found, :exc:`~sqlalchemy.exc.UnboundExecutionError`
+           is raised.
+
+        Note that the :meth:`.Session.get_bind` method can be overridden on
+        a user-defined subclass of :class:`.Session` to provide any kind
+        of bind resolution scheme.  See the example at
+        :ref:`session_custom_partitioning`.
+
+        :param mapper:
+          Optional mapped class or corresponding :class:`_orm.Mapper` instance.
+          The bind can be derived from a :class:`_orm.Mapper` first by
+          consulting the "binds" map associated with this :class:`.Session`,
+          and secondly by consulting the :class:`_schema.MetaData` associated
+          with the :class:`_schema.Table` to which the :class:`_orm.Mapper` is
+          mapped for a bind.
+
+        :param clause:
+            A :class:`_expression.ClauseElement` (i.e.
+            :func:`_expression.select`,
+            :func:`_expression.text`,
+            etc.).  If the ``mapper`` argument is not present or could not
+            produce a bind, the given expression construct will be searched
+            for a bound element, typically a :class:`_schema.Table`
+            associated with
+            bound :class:`_schema.MetaData`.
+
+        .. seealso::
+
+             :ref:`session_partitioning`
+
+             :paramref:`.Session.binds`
+
+             :meth:`.Session.bind_mapper`
+
+             :meth:`.Session.bind_table`
+
+
+        """  # noqa: E501
+
+        return self._proxied.get_bind(
+            mapper=mapper,
+            clause=clause,
+            bind=bind,
+            _sa_skip_events=_sa_skip_events,
+            _sa_skip_for_implicit_returning=_sa_skip_for_implicit_returning,
+            **kw,
+        )
+
+    def is_modified(
+        self, instance: object, include_collections: bool = True
+    ) -> bool:
+        r"""Return ``True`` if the given instance has locally
+        modified attributes.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This method retrieves the history for each instrumented
+        attribute on the instance and performs a comparison of the current
+        value to its previously flushed or committed value, if any.
+
+        It is in effect a more expensive and accurate
+        version of checking for the given instance in the
+        :attr:`.Session.dirty` collection; a full test for
+        each attribute's net "dirty" status is performed.
+
+        E.g.::
+
+            return session.is_modified(someobject)
+
+        A few caveats to this method apply:
+
+        * Instances present in the :attr:`.Session.dirty` collection may
+          report ``False`` when tested with this method.  This is because
+          the object may have received change events via attribute mutation,
+          thus placing it in :attr:`.Session.dirty`, but ultimately the state
+          is the same as that loaded from the database, resulting in no net
+          change here.
+        * Scalar attributes may not have recorded the previously set
+          value when a new value was applied, if the attribute was not loaded,
+          or was expired, at the time the new value was received - in these
+          cases, the attribute is assumed to have a change, even if there is
+          ultimately no net change against its database value. SQLAlchemy in
+          most cases does not need the "old" value when a set event occurs, so
+          it skips the expense of a SQL call if the old value isn't present,
+          based on the assumption that an UPDATE of the scalar value is
+          usually needed, and in those few cases where it isn't, is less
+          expensive on average than issuing a defensive SELECT.
+
+          The "old" value is fetched unconditionally upon set only if the
+          attribute container has the ``active_history`` flag set to ``True``.
+          This flag is set typically for primary key attributes and scalar
+          object references that are not a simple many-to-one.  To set this
+          flag for any arbitrary mapped column, use the ``active_history``
+          argument with :func:`.column_property`.
+
+        :param instance: mapped instance to be tested for pending changes.
+        :param include_collections: Indicates if multivalued collections
+         should be included in the operation.  Setting this to ``False`` is a
+         way to detect only local-column based properties (i.e. scalar columns
+         or many-to-one foreign keys) that would result in an UPDATE for this
+         instance upon flush.
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_modified(
+            instance, include_collections=include_collections
+        )
+
+    def bulk_save_objects(
+        self,
+        objects: Iterable[object],
+        return_defaults: bool = False,
+        update_changed_only: bool = True,
+        preserve_order: bool = True,
+    ) -> None:
+        r"""Perform a bulk save of the given list of objects.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.
+
+            For general INSERT and UPDATE of existing ORM mapped objects,
+            prefer standard :term:`unit of work` data management patterns,
+            introduced in the :ref:`unified_tutorial` at
+            :ref:`tutorial_orm_data_manipulation`.  SQLAlchemy 2.0
+            now uses :ref:`engine_insertmanyvalues` with modern dialects
+            which solves previous issues of bulk INSERT slowness.
+
+        :param objects: a sequence of mapped object instances.  The mapped
+         objects are persisted as is, and are **not** associated with the
+         :class:`.Session` afterwards.
+
+         For each object, whether the object is sent as an INSERT or an
+         UPDATE is dependent on the same rules used by the :class:`.Session`
+         in traditional operation; if the object has the
+         :attr:`.InstanceState.key`
+         attribute set, then the object is assumed to be "detached" and
+         will result in an UPDATE.  Otherwise, an INSERT is used.
+
+         In the case of an UPDATE, statements are grouped based on which
+         attributes have changed, and are thus to be the subject of each
+         SET clause.  If ``update_changed_only`` is False, then all
+         attributes present within each object are applied to the UPDATE
+         statement, which may help in allowing the statements to be grouped
+         together into a larger executemany(), and will also reduce the
+         overhead of checking history on attributes.
+
+        :param return_defaults: when True, rows that are missing values which
+         generate defaults, namely integer primary key defaults and sequences,
+         will be inserted **one at a time**, so that the primary key value
+         is available.  In particular this will allow joined-inheritance
+         and other multi-table mappings to insert correctly without the need
+         to provide primary key values ahead of time; however,
+         :paramref:`.Session.bulk_save_objects.return_defaults` **greatly
+         reduces the performance gains** of the method overall.  It is strongly
+         advised to please use the standard :meth:`_orm.Session.add_all`
+         approach.
+
+        :param update_changed_only: when True, UPDATE statements are rendered
+         based on those attributes in each state that have logged changes.
+         When False, all attributes present are rendered into the SET clause
+         with the exception of primary key attributes.
+
+        :param preserve_order: when True, the order of inserts and updates
+         matches exactly the order in which the objects are given.   When
+         False, common types of objects are grouped into inserts
+         and updates, to allow for more batching opportunities.
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_insert_mappings`
+
+            :meth:`.Session.bulk_update_mappings`
+
+
+        """  # noqa: E501
+
+        return self._proxied.bulk_save_objects(
+            objects,
+            return_defaults=return_defaults,
+            update_changed_only=update_changed_only,
+            preserve_order=preserve_order,
+        )
+
+    def bulk_insert_mappings(
+        self,
+        mapper: Mapper[Any],
+        mappings: Iterable[Dict[str, Any]],
+        return_defaults: bool = False,
+        render_nulls: bool = False,
+    ) -> None:
+        r"""Perform a bulk insert of the given list of mapping dictionaries.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.  The 2.0 API shares
+            implementation details with this method and adds new features
+            as well.
+
+        :param mapper: a mapped class, or the actual :class:`_orm.Mapper`
+         object,
+         representing the single kind of object represented within the mapping
+         list.
+
+        :param mappings: a sequence of dictionaries, each one containing the
+         state of the mapped row to be inserted, in terms of the attribute
+         names on the mapped class.   If the mapping refers to multiple tables,
+         such as a joined-inheritance mapping, each dictionary must contain all
+         keys to be populated into all tables.
+
+        :param return_defaults: when True, the INSERT process will be altered
+         to ensure that newly generated primary key values will be fetched.
+         The rationale for this parameter is typically to enable
+         :ref:`Joined Table Inheritance <joined_inheritance>` mappings to
+         be bulk inserted.
+
+         .. note:: for backends that don't support RETURNING, the
+            :paramref:`_orm.Session.bulk_insert_mappings.return_defaults`
+            parameter can significantly decrease performance as INSERT
+            statements can no longer be batched.   See
+            :ref:`engine_insertmanyvalues`
+            for background on which backends are affected.
+
+        :param render_nulls: When True, a value of ``None`` will result
+         in a NULL value being included in the INSERT statement, rather
+         than the column being omitted from the INSERT.   This allows all
+         the rows being INSERTed to have the identical set of columns which
+         allows the full set of rows to be batched to the DBAPI.  Normally,
+         each column-set that contains a different combination of NULL values
+         than the previous row must omit a different series of columns from
+         the rendered INSERT statement, which means it must be emitted as a
+         separate statement.   By passing this flag, the full set of rows
+         are guaranteed to be batchable into one batch; the cost however is
+         that server-side defaults which are invoked by an omitted column will
+         be skipped, so care must be taken to ensure that these are not
+         necessary.
+
+         .. warning::
+
+            When this flag is set, **server side default SQL values will
+            not be invoked** for those columns that are inserted as NULL;
+            the NULL value will be sent explicitly.   Care must be taken
+            to ensure that no server-side default functions need to be
+            invoked for the operation as a whole.
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_save_objects`
+
+            :meth:`.Session.bulk_update_mappings`
+
+
+        """  # noqa: E501
+
+        return self._proxied.bulk_insert_mappings(
+            mapper,
+            mappings,
+            return_defaults=return_defaults,
+            render_nulls=render_nulls,
+        )
+
+    def bulk_update_mappings(
+        self, mapper: Mapper[Any], mappings: Iterable[Dict[str, Any]]
+    ) -> None:
+        r"""Perform a bulk update of the given list of mapping dictionaries.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.  The 2.0 API shares
+            implementation details with this method and adds new features
+            as well.
+
+        :param mapper: a mapped class, or the actual :class:`_orm.Mapper`
+         object,
+         representing the single kind of object represented within the mapping
+         list.
+
+        :param mappings: a sequence of dictionaries, each one containing the
+         state of the mapped row to be updated, in terms of the attribute names
+         on the mapped class.   If the mapping refers to multiple tables, such
+         as a joined-inheritance mapping, each dictionary may contain keys
+         corresponding to all tables.   All those keys which are present and
+         are not part of the primary key are applied to the SET clause of the
+         UPDATE statement; the primary key values, which are required, are
+         applied to the WHERE clause.
+
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_insert_mappings`
+
+            :meth:`.Session.bulk_save_objects`
+
+
+        """  # noqa: E501
+
+        return self._proxied.bulk_update_mappings(mapper, mappings)
+
+    def merge(
+        self,
+        instance: _O,
+        *,
+        load: bool = True,
+        options: Optional[Sequence[ORMOption]] = None,
+    ) -> _O:
+        r"""Copy the state of a given instance into a corresponding instance
+        within this :class:`.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        :meth:`.Session.merge` examines the primary key attributes of the
+        source instance, and attempts to reconcile it with an instance of the
+        same primary key in the session.   If not found locally, it attempts
+        to load the object from the database based on primary key, and if
+        none can be located, creates a new instance.  The state of each
+        attribute on the source instance is then copied to the target
+        instance.  The resulting target instance is then returned by the
+        method; the original source instance is left unmodified, and
+        un-associated with the :class:`.Session` if not already.
+
+        This operation cascades to associated instances if the association is
+        mapped with ``cascade="merge"``.
+
+        See :ref:`unitofwork_merging` for a detailed discussion of merging.
+
+        :param instance: Instance to be merged.
+        :param load: Boolean, when False, :meth:`.merge` switches into
+         a "high performance" mode which causes it to forego emitting history
+         events as well as all database access.  This flag is used for
+         cases such as transferring graphs of objects into a :class:`.Session`
+         from a second level cache, or to transfer just-loaded objects
+         into the :class:`.Session` owned by a worker thread or process
+         without re-querying the database.
+
+         The ``load=False`` use case adds the caveat that the given
+         object has to be in a "clean" state, that is, has no pending changes
+         to be flushed - even if the incoming object is detached from any
+         :class:`.Session`.   This is so that when
+         the merge operation populates local attributes and
+         cascades to related objects and
+         collections, the values can be "stamped" onto the
+         target object as is, without generating any history or attribute
+         events, and without the need to reconcile the incoming data with
+         any existing related objects or collections that might not
+         be loaded.  The resulting objects from ``load=False`` are always
+         produced as "clean", so it is only appropriate that the given objects
+         should be "clean" as well, else this suggests a mis-use of the
+         method.
+        :param options: optional sequence of loader options which will be
+         applied to the :meth:`_orm.Session.get` method when the merge
+         operation loads the existing version of the object from the database.
+
+         .. versionadded:: 1.4.24
+
+
+        .. seealso::
+
+            :func:`.make_transient_to_detached` - provides for an alternative
+            means of "merging" a single object into the :class:`.Session`
+
+
+        """  # noqa: E501
+
+        return self._proxied.merge(instance, load=load, options=options)
+
+    @overload
+    def query(self, _entity: _EntityType[_O]) -> Query[_O]: ...
+
+    @overload
+    def query(
+        self, _colexpr: TypedColumnsClauseRole[_T]
+    ) -> RowReturningQuery[Tuple[_T]]: ...
+
+    # START OVERLOADED FUNCTIONS self.query RowReturningQuery 2-8
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_tuple_map_overloads.py
+
+    @overload
+    def query(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+    ) -> RowReturningQuery[Tuple[_T0, _T1]]: ...
+
+    @overload
+    def query(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+        __ent7: _TCCA[_T7],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]]: ...
+
+    # END OVERLOADED FUNCTIONS self.query
+
+    @overload
+    def query(
+        self, *entities: _ColumnsClauseArgument[Any], **kwargs: Any
+    ) -> Query[Any]: ...
+
+    def query(
+        self, *entities: _ColumnsClauseArgument[Any], **kwargs: Any
+    ) -> Query[Any]:
+        r"""Return a new :class:`_query.Query` object corresponding to this
+        :class:`_orm.Session`.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Note that the :class:`_query.Query` object is legacy as of
+        SQLAlchemy 2.0; the :func:`_sql.select` construct is now used
+        to construct ORM queries.
+
+        .. seealso::
+
+            :ref:`unified_tutorial`
+
+            :ref:`queryguide_toplevel`
+
+            :ref:`query_api_toplevel` - legacy API doc
+
+
+        """  # noqa: E501
+
+        return self._proxied.query(*entities, **kwargs)
+
+    def refresh(
+        self,
+        instance: object,
+        attribute_names: Optional[Iterable[str]] = None,
+        with_for_update: ForUpdateParameter = None,
+    ) -> None:
+        r"""Expire and refresh attributes on the given instance.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The selected attributes will first be expired as they would when using
+        :meth:`_orm.Session.expire`; then a SELECT statement will be issued to
+        the database to refresh column-oriented attributes with the current
+        value available in the current transaction.
+
+        :func:`_orm.relationship` oriented attributes will also be immediately
+        loaded if they were already eagerly loaded on the object, using the
+        same eager loading strategy that they were loaded with originally.
+
+        .. versionadded:: 1.4 - the :meth:`_orm.Session.refresh` method
+           can also refresh eagerly loaded attributes.
+
+        :func:`_orm.relationship` oriented attributes that would normally
+        load using the ``select`` (or "lazy") loader strategy will also
+        load **if they are named explicitly in the attribute_names
+        collection**, emitting a SELECT statement for the attribute using the
+        ``immediate`` loader strategy.  If lazy-loaded relationships are not
+        named in :paramref:`_orm.Session.refresh.attribute_names`, then
+        they remain as "lazy loaded" attributes and are not implicitly
+        refreshed.
+
+        .. versionchanged:: 2.0.4  The :meth:`_orm.Session.refresh` method
+           will now refresh lazy-loaded :func:`_orm.relationship` oriented
+           attributes for those which are named explicitly in the
+           :paramref:`_orm.Session.refresh.attribute_names` collection.
+
+        .. tip::
+
+            While the :meth:`_orm.Session.refresh` method is capable of
+            refreshing both column and relationship oriented attributes, its
+            primary focus is on refreshing of local column-oriented attributes
+            on a single instance. For more open ended "refresh" functionality,
+            including the ability to refresh the attributes on many objects at
+            once while having explicit control over relationship loader
+            strategies, use the
+            :ref:`populate existing <orm_queryguide_populate_existing>` feature
+            instead.
+
+        Note that a highly isolated transaction will return the same values as
+        were previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.   Refreshing
+        attributes usually only makes sense at the start of a transaction
+        where database rows have not yet been accessed.
+
+        :param attribute_names: optional.  An iterable collection of
+          string attribute names indicating a subset of attributes to
+          be refreshed.
+
+        :param with_for_update: optional boolean ``True`` indicating FOR UPDATE
+          should be used, or may be a dictionary containing flags to
+          indicate a more specific set of FOR UPDATE flags for the SELECT;
+          flags should match the parameters of
+          :meth:`_query.Query.with_for_update`.
+          Supersedes the :paramref:`.Session.refresh.lockmode` parameter.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.expire_all`
+
+            :ref:`orm_queryguide_populate_existing` - allows any ORM query
+            to refresh objects as they would be loaded normally.
+
+
+        """  # noqa: E501
+
+        return self._proxied.refresh(
+            instance,
+            attribute_names=attribute_names,
+            with_for_update=with_for_update,
+        )
+
+    def rollback(self) -> None:
+        r"""Rollback the current transaction in progress.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        If no transaction is in progress, this method is a pass-through.
+
+        The method always rolls back
+        the topmost database transaction, discarding any nested
+        transactions that may be in progress.
+
+        .. seealso::
+
+            :ref:`session_rollback`
+
+            :ref:`unitofwork_transaction`
+
+
+        """  # noqa: E501
+
+        return self._proxied.rollback()
+
+    @overload
+    def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Optional[_T]: ...
+
+    @overload
+    def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any: ...
+
+    def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any:
+        r"""Execute a statement and return a scalar result.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Usage and parameters are the same as that of
+        :meth:`_orm.Session.execute`; the return result is a scalar Python
+        value.
+
+
+        """  # noqa: E501
+
+        return self._proxied.scalar(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @overload
+    def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]: ...
+
+    def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]:
+        r"""Execute a statement and return the results as scalars.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        Usage and parameters are the same as that of
+        :meth:`_orm.Session.execute`; the return result is a
+        :class:`_result.ScalarResult` filtering object which
+        will return single elements rather than :class:`_row.Row` objects.
+
+        :return:  a :class:`_result.ScalarResult` object
+
+        .. versionadded:: 1.4.24 Added :meth:`_orm.Session.scalars`
+
+        .. versionadded:: 1.4.26 Added :meth:`_orm.scoped_session.scalars`
+
+        .. seealso::
+
+            :ref:`orm_queryguide_select_orm_entities` - contrasts the behavior
+            of :meth:`_orm.Session.execute` to :meth:`_orm.Session.scalars`
+
+
+        """  # noqa: E501
+
+        return self._proxied.scalars(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            **kw,
+        )
+
+    @property
+    def bind(self) -> Optional[Union[Engine, Connection]]:
+        r"""Proxy for the :attr:`_orm.Session.bind` attribute
+        on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.bind
+
+    @bind.setter
+    def bind(self, attr: Optional[Union[Engine, Connection]]) -> None:
+        self._proxied.bind = attr
+
+    @property
+    def dirty(self) -> Any:
+        r"""The set of all persistent instances considered dirty.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        E.g.::
+
+            some_mapped_object in session.dirty
+
+        Instances are considered dirty when they were modified but not
+        deleted.
+
+        Note that this 'dirty' calculation is 'optimistic'; most
+        attribute-setting or collection modification operations will
+        mark an instance as 'dirty' and place it in this set, even if
+        there is no net change to the attribute's value.  At flush
+        time, the value of each attribute is compared to its
+        previously saved value, and if there's no net change, no SQL
+        operation will occur (this is a more expensive operation so
+        it's only done at flush time).
+
+        To check if an instance has actionable net changes to its
+        attributes, use the :meth:`.Session.is_modified` method.
+
+
+        """  # noqa: E501
+
+        return self._proxied.dirty
+
+    @property
+    def deleted(self) -> Any:
+        r"""The set of all instances marked as 'deleted' within this ``Session``
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.deleted
+
+    @property
+    def new(self) -> Any:
+        r"""The set of all instances marked as 'new' within this ``Session``.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.new
+
+    @property
+    def identity_map(self) -> IdentityMap:
+        r"""Proxy for the :attr:`_orm.Session.identity_map` attribute
+        on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.identity_map
+
+    @identity_map.setter
+    def identity_map(self, attr: IdentityMap) -> None:
+        self._proxied.identity_map = attr
+
+    @property
+    def is_active(self) -> Any:
+        r"""True if this :class:`.Session` not in "partial rollback" state.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        .. versionchanged:: 1.4 The :class:`_orm.Session` no longer begins
+           a new transaction immediately, so this attribute will be False
+           when the :class:`_orm.Session` is first instantiated.
+
+        "partial rollback" state typically indicates that the flush process
+        of the :class:`_orm.Session` has failed, and that the
+        :meth:`_orm.Session.rollback` method must be emitted in order to
+        fully roll back the transaction.
+
+        If this :class:`_orm.Session` is not in a transaction at all, the
+        :class:`_orm.Session` will autobegin when it is first used, so in this
+        case :attr:`_orm.Session.is_active` will return True.
+
+        Otherwise, if this :class:`_orm.Session` is within a transaction,
+        and that transaction has not been rolled back internally, the
+        :attr:`_orm.Session.is_active` will also return True.
+
+        .. seealso::
+
+            :ref:`faq_session_rollback`
+
+            :meth:`_orm.Session.in_transaction`
+
+
+        """  # noqa: E501
+
+        return self._proxied.is_active
+
+    @property
+    def autoflush(self) -> bool:
+        r"""Proxy for the :attr:`_orm.Session.autoflush` attribute
+        on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        """  # noqa: E501
+
+        return self._proxied.autoflush
+
+    @autoflush.setter
+    def autoflush(self, attr: bool) -> None:
+        self._proxied.autoflush = attr
+
+    @property
+    def no_autoflush(self) -> Any:
+        r"""Return a context manager that disables autoflush.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        e.g.::
+
+            with session.no_autoflush:
+
+                some_object = SomeClass()
+                session.add(some_object)
+                # won't autoflush
+                some_object.related_thing = session.query(SomeRelated).first()
+
+        Operations that proceed within the ``with:`` block
+        will not be subject to flushes occurring upon query
+        access.  This is useful when initializing a series
+        of objects which involve existing database queries,
+        where the uncompleted object should not yet be flushed.
+
+
+        """  # noqa: E501
+
+        return self._proxied.no_autoflush
+
+    @property
+    def info(self) -> Any:
+        r"""A user-modifiable dictionary.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class
+            on behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        The initial value of this dictionary can be populated using the
+        ``info`` argument to the :class:`.Session` constructor or
+        :class:`.sessionmaker` constructor or factory methods.  The dictionary
+        here is always local to this :class:`.Session` and can be modified
+        independently of all other :class:`.Session` objects.
+
+
+        """  # noqa: E501
+
+        return self._proxied.info
+
+    @classmethod
+    def close_all(cls) -> None:
+        r"""Close *all* sessions in memory.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        .. deprecated:: 1.3 The :meth:`.Session.close_all` method is deprecated and will be removed in a future release.  Please refer to :func:`.session.close_all_sessions`.
+
+        """  # noqa: E501
+
+        return Session.close_all()
+
+    @classmethod
+    def object_session(cls, instance: object) -> Optional[Session]:
+        r"""Return the :class:`.Session` to which an object belongs.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This is an alias of :func:`.object_session`.
+
+
+        """  # noqa: E501
+
+        return Session.object_session(instance)
+
+    @classmethod
+    def identity_key(
+        cls,
+        class_: Optional[Type[Any]] = None,
+        ident: Union[Any, Tuple[Any, ...]] = None,
+        *,
+        instance: Optional[Any] = None,
+        row: Optional[Union[Row[Any], RowMapping]] = None,
+        identity_token: Optional[Any] = None,
+    ) -> _IdentityKeyType[Any]:
+        r"""Return an identity key.
+
+        .. container:: class_bases
+
+            Proxied for the :class:`_orm.Session` class on
+            behalf of the :class:`_orm.scoping.scoped_session` class.
+
+        This is an alias of :func:`.util.identity_key`.
+
+
+        """  # noqa: E501
+
+        return Session.identity_key(
+            class_=class_,
+            ident=ident,
+            instance=instance,
+            row=row,
+            identity_token=identity_token,
+        )
+
+    # END PROXY METHODS scoped_session
+
+
+ScopedSession = scoped_session
+"""Old name for backwards compatibility."""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/session.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/session.py
new file mode 100644
index 00000000..6cd7cd63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/session.py
@@ -0,0 +1,5302 @@
+# orm/session.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Provides the Session class and related utilities."""
+
+from __future__ import annotations
+
+import contextlib
+from enum import Enum
+import itertools
+import sys
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes
+from . import bulk_persistence
+from . import context
+from . import descriptor_props
+from . import exc
+from . import identity
+from . import loading
+from . import query
+from . import state as statelib
+from ._typing import _O
+from ._typing import insp_is_mapper
+from ._typing import is_composite_class
+from ._typing import is_orm_option
+from ._typing import is_user_defined_option
+from .base import _class_to_mapper
+from .base import _none_set
+from .base import _state_mapper
+from .base import instance_str
+from .base import LoaderCallableStatus
+from .base import object_mapper
+from .base import object_state
+from .base import PassiveFlag
+from .base import state_str
+from .context import FromStatement
+from .context import ORMCompileState
+from .identity import IdentityMap
+from .query import Query
+from .state import InstanceState
+from .state_changes import _StateChange
+from .state_changes import _StateChangeState
+from .state_changes import _StateChangeStates
+from .unitofwork import UOWTransaction
+from .. import engine
+from .. import exc as sa_exc
+from .. import sql
+from .. import util
+from ..engine import Connection
+from ..engine import Engine
+from ..engine.util import TransactionalContext
+from ..event import dispatcher
+from ..event import EventTarget
+from ..inspection import inspect
+from ..inspection import Inspectable
+from ..sql import coercions
+from ..sql import dml
+from ..sql import roles
+from ..sql import Select
+from ..sql import TableClause
+from ..sql import visitors
+from ..sql.base import _NoArg
+from ..sql.base import CompileState
+from ..sql.schema import Table
+from ..sql.selectable import ForUpdateArg
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..util import IdentitySet
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import OrmExecuteOptionsParameter
+    from .interfaces import ORMOption
+    from .interfaces import UserDefinedOption
+    from .mapper import Mapper
+    from .path_registry import PathRegistry
+    from .query import RowReturningQuery
+    from ..engine import CursorResult
+    from ..engine import Result
+    from ..engine import Row
+    from ..engine import RowMapping
+    from ..engine.base import Transaction
+    from ..engine.base import TwoPhaseTransaction
+    from ..engine.interfaces import _CoreAnyExecuteParams
+    from ..engine.interfaces import _CoreSingleExecuteParams
+    from ..engine.interfaces import _ExecuteOptions
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..engine.result import ScalarResult
+    from ..event import _InstanceLevelDispatch
+    from ..sql._typing import _ColumnsClauseArgument
+    from ..sql._typing import _InfoType
+    from ..sql._typing import _T0
+    from ..sql._typing import _T1
+    from ..sql._typing import _T2
+    from ..sql._typing import _T3
+    from ..sql._typing import _T4
+    from ..sql._typing import _T5
+    from ..sql._typing import _T6
+    from ..sql._typing import _T7
+    from ..sql._typing import _TypedColumnClauseArgument as _TCCA
+    from ..sql.base import Executable
+    from ..sql.base import ExecutableOption
+    from ..sql.dml import UpdateBase
+    from ..sql.elements import ClauseElement
+    from ..sql.roles import TypedColumnsClauseRole
+    from ..sql.selectable import ForUpdateParameter
+    from ..sql.selectable import TypedReturnsRows
+
+_T = TypeVar("_T", bound=Any)
+
+__all__ = [
+    "Session",
+    "SessionTransaction",
+    "sessionmaker",
+    "ORMExecuteState",
+    "close_all_sessions",
+    "make_transient",
+    "make_transient_to_detached",
+    "object_session",
+]
+
+_sessions: weakref.WeakValueDictionary[int, Session] = (
+    weakref.WeakValueDictionary()
+)
+"""Weak-referencing dictionary of :class:`.Session` objects.
+"""
+
+statelib._sessions = _sessions
+
+_PKIdentityArgument = Union[Any, Tuple[Any, ...]]
+
+_BindArguments = Dict[str, Any]
+
+_EntityBindKey = Union[Type[_O], "Mapper[_O]"]
+_SessionBindKey = Union[Type[Any], "Mapper[Any]", "TableClause", str]
+_SessionBind = Union["Engine", "Connection"]
+
+JoinTransactionMode = Literal[
+    "conditional_savepoint",
+    "rollback_only",
+    "control_fully",
+    "create_savepoint",
+]
+
+
+class _ConnectionCallableProto(Protocol):
+    """a callable that returns a :class:`.Connection` given an instance.
+
+    This callable, when present on a :class:`.Session`, is called only from the
+    ORM's persistence mechanism (i.e. the unit of work flush process) to allow
+    for connection-per-instance schemes (i.e. horizontal sharding) to be used
+    as persistence time.
+
+    This callable is not present on a plain :class:`.Session`, however
+    is established when using the horizontal sharding extension.
+
+    """
+
+    def __call__(
+        self,
+        mapper: Optional[Mapper[Any]] = None,
+        instance: Optional[object] = None,
+        **kw: Any,
+    ) -> Connection: ...
+
+
+def _state_session(state: InstanceState[Any]) -> Optional[Session]:
+    """Given an :class:`.InstanceState`, return the :class:`.Session`
+    associated, if any.
+    """
+    return state.session
+
+
+class _SessionClassMethods:
+    """Class-level methods for :class:`.Session`, :class:`.sessionmaker`."""
+
+    @classmethod
+    @util.deprecated(
+        "1.3",
+        "The :meth:`.Session.close_all` method is deprecated and will be "
+        "removed in a future release.  Please refer to "
+        ":func:`.session.close_all_sessions`.",
+    )
+    def close_all(cls) -> None:
+        """Close *all* sessions in memory."""
+
+        close_all_sessions()
+
+    @classmethod
+    @util.preload_module("sqlalchemy.orm.util")
+    def identity_key(
+        cls,
+        class_: Optional[Type[Any]] = None,
+        ident: Union[Any, Tuple[Any, ...]] = None,
+        *,
+        instance: Optional[Any] = None,
+        row: Optional[Union[Row[Any], RowMapping]] = None,
+        identity_token: Optional[Any] = None,
+    ) -> _IdentityKeyType[Any]:
+        """Return an identity key.
+
+        This is an alias of :func:`.util.identity_key`.
+
+        """
+        return util.preloaded.orm_util.identity_key(
+            class_,
+            ident,
+            instance=instance,
+            row=row,
+            identity_token=identity_token,
+        )
+
+    @classmethod
+    def object_session(cls, instance: object) -> Optional[Session]:
+        """Return the :class:`.Session` to which an object belongs.
+
+        This is an alias of :func:`.object_session`.
+
+        """
+
+        return object_session(instance)
+
+
+class SessionTransactionState(_StateChangeState):
+    ACTIVE = 1
+    PREPARED = 2
+    COMMITTED = 3
+    DEACTIVE = 4
+    CLOSED = 5
+    PROVISIONING_CONNECTION = 6
+
+
+# backwards compatibility
+ACTIVE, PREPARED, COMMITTED, DEACTIVE, CLOSED, PROVISIONING_CONNECTION = tuple(
+    SessionTransactionState
+)
+
+
+class ORMExecuteState(util.MemoizedSlots):
+    """Represents a call to the :meth:`_orm.Session.execute` method, as passed
+    to the :meth:`.SessionEvents.do_orm_execute` event hook.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`session_execute_events` - top level documentation on how
+        to use :meth:`_orm.SessionEvents.do_orm_execute`
+
+    """
+
+    __slots__ = (
+        "session",
+        "statement",
+        "parameters",
+        "execution_options",
+        "local_execution_options",
+        "bind_arguments",
+        "identity_token",
+        "_compile_state_cls",
+        "_starting_event_idx",
+        "_events_todo",
+        "_update_execution_options",
+    )
+
+    session: Session
+    """The :class:`_orm.Session` in use."""
+
+    statement: Executable
+    """The SQL statement being invoked.
+
+    For an ORM selection as would
+    be retrieved from :class:`_orm.Query`, this is an instance of
+    :class:`_sql.select` that was generated from the ORM query.
+    """
+
+    parameters: Optional[_CoreAnyExecuteParams]
+    """Dictionary of parameters that was passed to
+    :meth:`_orm.Session.execute`."""
+
+    execution_options: _ExecuteOptions
+    """The complete dictionary of current execution options.
+
+    This is a merge of the statement level options with the
+    locally passed execution options.
+
+    .. seealso::
+
+        :attr:`_orm.ORMExecuteState.local_execution_options`
+
+        :meth:`_sql.Executable.execution_options`
+
+        :ref:`orm_queryguide_execution_options`
+
+    """
+
+    local_execution_options: _ExecuteOptions
+    """Dictionary view of the execution options passed to the
+    :meth:`.Session.execute` method.
+
+    This does not include options that may be associated with the statement
+    being invoked.
+
+    .. seealso::
+
+        :attr:`_orm.ORMExecuteState.execution_options`
+
+    """
+
+    bind_arguments: _BindArguments
+    """The dictionary passed as the
+    :paramref:`_orm.Session.execute.bind_arguments` dictionary.
+
+    This dictionary may be used by extensions to :class:`_orm.Session` to pass
+    arguments that will assist in determining amongst a set of database
+    connections which one should be used to invoke this statement.
+
+    """
+
+    _compile_state_cls: Optional[Type[ORMCompileState]]
+    _starting_event_idx: int
+    _events_todo: List[Any]
+    _update_execution_options: Optional[_ExecuteOptions]
+
+    def __init__(
+        self,
+        session: Session,
+        statement: Executable,
+        parameters: Optional[_CoreAnyExecuteParams],
+        execution_options: _ExecuteOptions,
+        bind_arguments: _BindArguments,
+        compile_state_cls: Optional[Type[ORMCompileState]],
+        events_todo: List[_InstanceLevelDispatch[Session]],
+    ):
+        """Construct a new :class:`_orm.ORMExecuteState`.
+
+        this object is constructed internally.
+
+        """
+        self.session = session
+        self.statement = statement
+        self.parameters = parameters
+        self.local_execution_options = execution_options
+        self.execution_options = statement._execution_options.union(
+            execution_options
+        )
+        self.bind_arguments = bind_arguments
+        self._compile_state_cls = compile_state_cls
+        self._events_todo = list(events_todo)
+
+    def _remaining_events(self) -> List[_InstanceLevelDispatch[Session]]:
+        return self._events_todo[self._starting_event_idx + 1 :]
+
+    def invoke_statement(
+        self,
+        statement: Optional[Executable] = None,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        execution_options: Optional[OrmExecuteOptionsParameter] = None,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> Result[Any]:
+        """Execute the statement represented by this
+        :class:`.ORMExecuteState`, without re-invoking events that have
+        already proceeded.
+
+        This method essentially performs a re-entrant execution of the current
+        statement for which the :meth:`.SessionEvents.do_orm_execute` event is
+        being currently invoked.    The use case for this is for event handlers
+        that want to override how the ultimate
+        :class:`_engine.Result` object is returned, such as for schemes that
+        retrieve results from an offline cache or which concatenate results
+        from multiple executions.
+
+        When the :class:`_engine.Result` object is returned by the actual
+        handler function within :meth:`_orm.SessionEvents.do_orm_execute` and
+        is propagated to the calling
+        :meth:`_orm.Session.execute` method, the remainder of the
+        :meth:`_orm.Session.execute` method is preempted and the
+        :class:`_engine.Result` object is returned to the caller of
+        :meth:`_orm.Session.execute` immediately.
+
+        :param statement: optional statement to be invoked, in place of the
+         statement currently represented by :attr:`.ORMExecuteState.statement`.
+
+        :param params: optional dictionary of parameters or list of parameters
+         which will be merged into the existing
+         :attr:`.ORMExecuteState.parameters` of this :class:`.ORMExecuteState`.
+
+         .. versionchanged:: 2.0 a list of parameter dictionaries is accepted
+            for executemany executions.
+
+        :param execution_options: optional dictionary of execution options
+         will be merged into the existing
+         :attr:`.ORMExecuteState.execution_options` of this
+         :class:`.ORMExecuteState`.
+
+        :param bind_arguments: optional dictionary of bind_arguments
+         which will be merged amongst the current
+         :attr:`.ORMExecuteState.bind_arguments`
+         of this :class:`.ORMExecuteState`.
+
+        :return: a :class:`_engine.Result` object with ORM-level results.
+
+        .. seealso::
+
+            :ref:`do_orm_execute_re_executing` - background and examples on the
+            appropriate usage of :meth:`_orm.ORMExecuteState.invoke_statement`.
+
+
+        """
+
+        if statement is None:
+            statement = self.statement
+
+        _bind_arguments = dict(self.bind_arguments)
+        if bind_arguments:
+            _bind_arguments.update(bind_arguments)
+        _bind_arguments["_sa_skip_events"] = True
+
+        _params: Optional[_CoreAnyExecuteParams]
+        if params:
+            if self.is_executemany:
+                _params = []
+                exec_many_parameters = cast(
+                    "List[Dict[str, Any]]", self.parameters
+                )
+                for _existing_params, _new_params in itertools.zip_longest(
+                    exec_many_parameters,
+                    cast("List[Dict[str, Any]]", params),
+                ):
+                    if _existing_params is None or _new_params is None:
+                        raise sa_exc.InvalidRequestError(
+                            f"Can't apply executemany parameters to "
+                            f"statement; number of parameter sets passed to "
+                            f"Session.execute() ({len(exec_many_parameters)}) "
+                            f"does not match number of parameter sets given "
+                            f"to ORMExecuteState.invoke_statement() "
+                            f"({len(params)})"
+                        )
+                    _existing_params = dict(_existing_params)
+                    _existing_params.update(_new_params)
+                    _params.append(_existing_params)
+            else:
+                _params = dict(cast("Dict[str, Any]", self.parameters))
+                _params.update(cast("Dict[str, Any]", params))
+        else:
+            _params = self.parameters
+
+        _execution_options = self.local_execution_options
+        if execution_options:
+            _execution_options = _execution_options.union(execution_options)
+
+        return self.session._execute_internal(
+            statement,
+            _params,
+            execution_options=_execution_options,
+            bind_arguments=_bind_arguments,
+            _parent_execute_state=self,
+        )
+
+    @property
+    def bind_mapper(self) -> Optional[Mapper[Any]]:
+        """Return the :class:`_orm.Mapper` that is the primary "bind" mapper.
+
+        For an :class:`_orm.ORMExecuteState` object invoking an ORM
+        statement, that is, the :attr:`_orm.ORMExecuteState.is_orm_statement`
+        attribute is ``True``, this attribute will return the
+        :class:`_orm.Mapper` that is considered to be the "primary" mapper
+        of the statement.   The term "bind mapper" refers to the fact that
+        a :class:`_orm.Session` object may be "bound" to multiple
+        :class:`_engine.Engine` objects keyed to mapped classes, and the
+        "bind mapper" determines which of those :class:`_engine.Engine` objects
+        would be selected.
+
+        For a statement that is invoked against a single mapped class,
+        :attr:`_orm.ORMExecuteState.bind_mapper` is intended to be a reliable
+        way of getting this mapper.
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :attr:`_orm.ORMExecuteState.all_mappers`
+
+
+        """
+        mp: Optional[Mapper[Any]] = self.bind_arguments.get("mapper", None)
+        return mp
+
+    @property
+    def all_mappers(self) -> Sequence[Mapper[Any]]:
+        """Return a sequence of all :class:`_orm.Mapper` objects that are
+        involved at the top level of this statement.
+
+        By "top level" we mean those :class:`_orm.Mapper` objects that would
+        be represented in the result set rows for a :func:`_sql.select`
+        query, or for a :func:`_dml.update` or :func:`_dml.delete` query,
+        the mapper that is the main subject of the UPDATE or DELETE.
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :attr:`_orm.ORMExecuteState.bind_mapper`
+
+
+
+        """
+        if not self.is_orm_statement:
+            return []
+        elif isinstance(self.statement, (Select, FromStatement)):
+            result = []
+            seen = set()
+            for d in self.statement.column_descriptions:
+                ent = d["entity"]
+                if ent:
+                    insp = inspect(ent, raiseerr=False)
+                    if insp and insp.mapper and insp.mapper not in seen:
+                        seen.add(insp.mapper)
+                        result.append(insp.mapper)
+            return result
+        elif self.statement.is_dml and self.bind_mapper:
+            return [self.bind_mapper]
+        else:
+            return []
+
+    @property
+    def is_orm_statement(self) -> bool:
+        """return True if the operation is an ORM statement.
+
+        This indicates that the select(), insert(), update(), or delete()
+        being invoked contains ORM entities as subjects.   For a statement
+        that does not have ORM entities and instead refers only to
+        :class:`.Table` metadata, it is invoked as a Core SQL statement
+        and no ORM-level automation takes place.
+
+        """
+        return self._compile_state_cls is not None
+
+    @property
+    def is_executemany(self) -> bool:
+        """return True if the parameters are a multi-element list of
+        dictionaries with more than one dictionary.
+
+        .. versionadded:: 2.0
+
+        """
+        return isinstance(self.parameters, list)
+
+    @property
+    def is_select(self) -> bool:
+        """return True if this is a SELECT operation.
+
+        .. versionchanged:: 2.0.30 - the attribute is also True for a
+           :meth:`_sql.Select.from_statement` construct that is itself against
+           a :class:`_sql.Select` construct, such as
+           ``select(Entity).from_statement(select(..))``
+
+        """
+        return self.statement.is_select
+
+    @property
+    def is_from_statement(self) -> bool:
+        """return True if this operation is a
+        :meth:`_sql.Select.from_statement` operation.
+
+        This is independent from :attr:`_orm.ORMExecuteState.is_select`, as a
+        ``select().from_statement()`` construct can be used with
+        INSERT/UPDATE/DELETE RETURNING types of statements as well.
+        :attr:`_orm.ORMExecuteState.is_select` will only be set if the
+        :meth:`_sql.Select.from_statement` is itself against a
+        :class:`_sql.Select` construct.
+
+        .. versionadded:: 2.0.30
+
+        """
+        return self.statement.is_from_statement
+
+    @property
+    def is_insert(self) -> bool:
+        """return True if this is an INSERT operation.
+
+        .. versionchanged:: 2.0.30 - the attribute is also True for a
+           :meth:`_sql.Select.from_statement` construct that is itself against
+           a :class:`_sql.Insert` construct, such as
+           ``select(Entity).from_statement(insert(..))``
+
+        """
+        return self.statement.is_dml and self.statement.is_insert
+
+    @property
+    def is_update(self) -> bool:
+        """return True if this is an UPDATE operation.
+
+        .. versionchanged:: 2.0.30 - the attribute is also True for a
+           :meth:`_sql.Select.from_statement` construct that is itself against
+           a :class:`_sql.Update` construct, such as
+           ``select(Entity).from_statement(update(..))``
+
+        """
+        return self.statement.is_dml and self.statement.is_update
+
+    @property
+    def is_delete(self) -> bool:
+        """return True if this is a DELETE operation.
+
+        .. versionchanged:: 2.0.30 - the attribute is also True for a
+           :meth:`_sql.Select.from_statement` construct that is itself against
+           a :class:`_sql.Delete` construct, such as
+           ``select(Entity).from_statement(delete(..))``
+
+        """
+        return self.statement.is_dml and self.statement.is_delete
+
+    @property
+    def _is_crud(self) -> bool:
+        return isinstance(self.statement, (dml.Update, dml.Delete))
+
+    def update_execution_options(self, **opts: Any) -> None:
+        """Update the local execution options with new values."""
+        self.local_execution_options = self.local_execution_options.union(opts)
+
+    def _orm_compile_options(
+        self,
+    ) -> Optional[
+        Union[
+            context.ORMCompileState.default_compile_options,
+            Type[context.ORMCompileState.default_compile_options],
+        ]
+    ]:
+        if not self.is_select:
+            return None
+        try:
+            opts = self.statement._compile_options
+        except AttributeError:
+            return None
+
+        if opts is not None and opts.isinstance(
+            context.ORMCompileState.default_compile_options
+        ):
+            return opts  # type: ignore
+        else:
+            return None
+
+    @property
+    def lazy_loaded_from(self) -> Optional[InstanceState[Any]]:
+        """An :class:`.InstanceState` that is using this statement execution
+        for a lazy load operation.
+
+        The primary rationale for this attribute is to support the horizontal
+        sharding extension, where it is available within specific query
+        execution time hooks created by this extension.   To that end, the
+        attribute is only intended to be meaningful at **query execution
+        time**, and importantly not any time prior to that, including query
+        compilation time.
+
+        """
+        return self.load_options._lazy_loaded_from
+
+    @property
+    def loader_strategy_path(self) -> Optional[PathRegistry]:
+        """Return the :class:`.PathRegistry` for the current load path.
+
+        This object represents the "path" in a query along relationships
+        when a particular object or collection is being loaded.
+
+        """
+        opts = self._orm_compile_options()
+        if opts is not None:
+            return opts._current_path
+        else:
+            return None
+
+    @property
+    def is_column_load(self) -> bool:
+        """Return True if the operation is refreshing column-oriented
+        attributes on an existing ORM object.
+
+        This occurs during operations such as :meth:`_orm.Session.refresh`,
+        as well as when an attribute deferred by :func:`_orm.defer` is
+        being loaded, or an attribute that was expired either directly
+        by :meth:`_orm.Session.expire` or via a commit operation is being
+        loaded.
+
+        Handlers will very likely not want to add any options to queries
+        when such an operation is occurring as the query should be a straight
+        primary key fetch which should not have any additional WHERE criteria,
+        and loader options travelling with the instance
+        will have already been added to the query.
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :attr:`_orm.ORMExecuteState.is_relationship_load`
+
+        """
+        opts = self._orm_compile_options()
+        return opts is not None and opts._for_refresh_state
+
+    @property
+    def is_relationship_load(self) -> bool:
+        """Return True if this load is loading objects on behalf of a
+        relationship.
+
+        This means, the loader in effect is either a LazyLoader,
+        SelectInLoader, SubqueryLoader, or similar, and the entire
+        SELECT statement being emitted is on behalf of a relationship
+        load.
+
+        Handlers will very likely not want to add any options to queries
+        when such an operation is occurring, as loader options are already
+        capable of being propagated to relationship loaders and should
+        be already present.
+
+        .. seealso::
+
+            :attr:`_orm.ORMExecuteState.is_column_load`
+
+        """
+        opts = self._orm_compile_options()
+        if opts is None:
+            return False
+        path = self.loader_strategy_path
+        return path is not None and not path.is_root
+
+    @property
+    def load_options(
+        self,
+    ) -> Union[
+        context.QueryContext.default_load_options,
+        Type[context.QueryContext.default_load_options],
+    ]:
+        """Return the load_options that will be used for this execution."""
+
+        if not self.is_select:
+            raise sa_exc.InvalidRequestError(
+                "This ORM execution is not against a SELECT statement "
+                "so there are no load options."
+            )
+
+        lo: Union[
+            context.QueryContext.default_load_options,
+            Type[context.QueryContext.default_load_options],
+        ] = self.execution_options.get(
+            "_sa_orm_load_options", context.QueryContext.default_load_options
+        )
+        return lo
+
+    @property
+    def update_delete_options(
+        self,
+    ) -> Union[
+        bulk_persistence.BulkUDCompileState.default_update_options,
+        Type[bulk_persistence.BulkUDCompileState.default_update_options],
+    ]:
+        """Return the update_delete_options that will be used for this
+        execution."""
+
+        if not self._is_crud:
+            raise sa_exc.InvalidRequestError(
+                "This ORM execution is not against an UPDATE or DELETE "
+                "statement so there are no update options."
+            )
+        uo: Union[
+            bulk_persistence.BulkUDCompileState.default_update_options,
+            Type[bulk_persistence.BulkUDCompileState.default_update_options],
+        ] = self.execution_options.get(
+            "_sa_orm_update_options",
+            bulk_persistence.BulkUDCompileState.default_update_options,
+        )
+        return uo
+
+    @property
+    def _non_compile_orm_options(self) -> Sequence[ORMOption]:
+        return [
+            opt
+            for opt in self.statement._with_options
+            if is_orm_option(opt) and not opt._is_compile_state
+        ]
+
+    @property
+    def user_defined_options(self) -> Sequence[UserDefinedOption]:
+        """The sequence of :class:`.UserDefinedOptions` that have been
+        associated with the statement being invoked.
+
+        """
+        return [
+            opt
+            for opt in self.statement._with_options
+            if is_user_defined_option(opt)
+        ]
+
+
+class SessionTransactionOrigin(Enum):
+    """indicates the origin of a :class:`.SessionTransaction`.
+
+    This enumeration is present on the
+    :attr:`.SessionTransaction.origin` attribute of any
+    :class:`.SessionTransaction` object.
+
+    .. versionadded:: 2.0
+
+    """
+
+    AUTOBEGIN = 0
+    """transaction were started by autobegin"""
+
+    BEGIN = 1
+    """transaction were started by calling :meth:`_orm.Session.begin`"""
+
+    BEGIN_NESTED = 2
+    """tranaction were started by :meth:`_orm.Session.begin_nested`"""
+
+    SUBTRANSACTION = 3
+    """transaction is an internal "subtransaction" """
+
+
+class SessionTransaction(_StateChange, TransactionalContext):
+    """A :class:`.Session`-level transaction.
+
+    :class:`.SessionTransaction` is produced from the
+    :meth:`_orm.Session.begin`
+    and :meth:`_orm.Session.begin_nested` methods.   It's largely an internal
+    object that in modern use provides a context manager for session
+    transactions.
+
+    Documentation on interacting with :class:`_orm.SessionTransaction` is
+    at: :ref:`unitofwork_transaction`.
+
+
+    .. versionchanged:: 1.4  The scoping and API methods to work with the
+       :class:`_orm.SessionTransaction` object directly have been simplified.
+
+    .. seealso::
+
+        :ref:`unitofwork_transaction`
+
+        :meth:`.Session.begin`
+
+        :meth:`.Session.begin_nested`
+
+        :meth:`.Session.rollback`
+
+        :meth:`.Session.commit`
+
+        :meth:`.Session.in_transaction`
+
+        :meth:`.Session.in_nested_transaction`
+
+        :meth:`.Session.get_transaction`
+
+        :meth:`.Session.get_nested_transaction`
+
+
+    """
+
+    _rollback_exception: Optional[BaseException] = None
+
+    _connections: Dict[
+        Union[Engine, Connection], Tuple[Connection, Transaction, bool, bool]
+    ]
+    session: Session
+    _parent: Optional[SessionTransaction]
+
+    _state: SessionTransactionState
+
+    _new: weakref.WeakKeyDictionary[InstanceState[Any], object]
+    _deleted: weakref.WeakKeyDictionary[InstanceState[Any], object]
+    _dirty: weakref.WeakKeyDictionary[InstanceState[Any], object]
+    _key_switches: weakref.WeakKeyDictionary[
+        InstanceState[Any], Tuple[Any, Any]
+    ]
+
+    origin: SessionTransactionOrigin
+    """Origin of this :class:`_orm.SessionTransaction`.
+
+    Refers to a :class:`.SessionTransactionOrigin` instance which is an
+    enumeration indicating the source event that led to constructing
+    this :class:`_orm.SessionTransaction`.
+
+    .. versionadded:: 2.0
+
+    """
+
+    nested: bool = False
+    """Indicates if this is a nested, or SAVEPOINT, transaction.
+
+    When :attr:`.SessionTransaction.nested` is True, it is expected
+    that :attr:`.SessionTransaction.parent` will be present as well,
+    linking to the enclosing :class:`.SessionTransaction`.
+
+    .. seealso::
+
+        :attr:`.SessionTransaction.origin`
+
+    """
+
+    def __init__(
+        self,
+        session: Session,
+        origin: SessionTransactionOrigin,
+        parent: Optional[SessionTransaction] = None,
+    ):
+        TransactionalContext._trans_ctx_check(session)
+
+        self.session = session
+        self._connections = {}
+        self._parent = parent
+        self.nested = nested = origin is SessionTransactionOrigin.BEGIN_NESTED
+        self.origin = origin
+
+        if session._close_state is _SessionCloseState.CLOSED:
+            raise sa_exc.InvalidRequestError(
+                "This Session has been permanently closed and is unable "
+                "to handle any more transaction requests."
+            )
+
+        if nested:
+            if not parent:
+                raise sa_exc.InvalidRequestError(
+                    "Can't start a SAVEPOINT transaction when no existing "
+                    "transaction is in progress"
+                )
+
+            self._previous_nested_transaction = session._nested_transaction
+        elif origin is SessionTransactionOrigin.SUBTRANSACTION:
+            assert parent is not None
+        else:
+            assert parent is None
+
+        self._state = SessionTransactionState.ACTIVE
+
+        self._take_snapshot()
+
+        # make sure transaction is assigned before we call the
+        # dispatch
+        self.session._transaction = self
+
+        self.session.dispatch.after_transaction_create(self.session, self)
+
+    def _raise_for_prerequisite_state(
+        self, operation_name: str, state: _StateChangeState
+    ) -> NoReturn:
+        if state is SessionTransactionState.DEACTIVE:
+            if self._rollback_exception:
+                raise sa_exc.PendingRollbackError(
+                    "This Session's transaction has been rolled back "
+                    "due to a previous exception during flush."
+                    " To begin a new transaction with this Session, "
+                    "first issue Session.rollback()."
+                    f" Original exception was: {self._rollback_exception}",
+                    code="7s2a",
+                )
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "This session is in 'inactive' state, due to the "
+                    "SQL transaction being rolled back; no further SQL "
+                    "can be emitted within this transaction."
+                )
+        elif state is SessionTransactionState.CLOSED:
+            raise sa_exc.ResourceClosedError("This transaction is closed")
+        elif state is SessionTransactionState.PROVISIONING_CONNECTION:
+            raise sa_exc.InvalidRequestError(
+                "This session is provisioning a new connection; concurrent "
+                "operations are not permitted",
+                code="isce",
+            )
+        else:
+            raise sa_exc.InvalidRequestError(
+                f"This session is in '{state.name.lower()}' state; no "
+                "further SQL can be emitted within this transaction."
+            )
+
+    @property
+    def parent(self) -> Optional[SessionTransaction]:
+        """The parent :class:`.SessionTransaction` of this
+        :class:`.SessionTransaction`.
+
+        If this attribute is ``None``, indicates this
+        :class:`.SessionTransaction` is at the top of the stack, and
+        corresponds to a real "COMMIT"/"ROLLBACK"
+        block.  If non-``None``, then this is either a "subtransaction"
+        (an internal marker object used by the flush process) or a
+        "nested" / SAVEPOINT transaction.  If the
+        :attr:`.SessionTransaction.nested` attribute is ``True``, then
+        this is a SAVEPOINT, and if ``False``, indicates this a subtransaction.
+
+        """
+        return self._parent
+
+    @property
+    def is_active(self) -> bool:
+        return (
+            self.session is not None
+            and self._state is SessionTransactionState.ACTIVE
+        )
+
+    @property
+    def _is_transaction_boundary(self) -> bool:
+        return self.nested or not self._parent
+
+    @_StateChange.declare_states(
+        (SessionTransactionState.ACTIVE,), _StateChangeStates.NO_CHANGE
+    )
+    def connection(
+        self,
+        bindkey: Optional[Mapper[Any]],
+        execution_options: Optional[_ExecuteOptions] = None,
+        **kwargs: Any,
+    ) -> Connection:
+        bind = self.session.get_bind(bindkey, **kwargs)
+        return self._connection_for_bind(bind, execution_options)
+
+    @_StateChange.declare_states(
+        (SessionTransactionState.ACTIVE,), _StateChangeStates.NO_CHANGE
+    )
+    def _begin(self, nested: bool = False) -> SessionTransaction:
+        return SessionTransaction(
+            self.session,
+            (
+                SessionTransactionOrigin.BEGIN_NESTED
+                if nested
+                else SessionTransactionOrigin.SUBTRANSACTION
+            ),
+            self,
+        )
+
+    def _iterate_self_and_parents(
+        self, upto: Optional[SessionTransaction] = None
+    ) -> Iterable[SessionTransaction]:
+        current = self
+        result: Tuple[SessionTransaction, ...] = ()
+        while current:
+            result += (current,)
+            if current._parent is upto:
+                break
+            elif current._parent is None:
+                raise sa_exc.InvalidRequestError(
+                    "Transaction %s is not on the active transaction list"
+                    % (upto)
+                )
+            else:
+                current = current._parent
+
+        return result
+
+    def _take_snapshot(self) -> None:
+        if not self._is_transaction_boundary:
+            parent = self._parent
+            assert parent is not None
+            self._new = parent._new
+            self._deleted = parent._deleted
+            self._dirty = parent._dirty
+            self._key_switches = parent._key_switches
+            return
+
+        is_begin = self.origin in (
+            SessionTransactionOrigin.BEGIN,
+            SessionTransactionOrigin.AUTOBEGIN,
+        )
+        if not is_begin and not self.session._flushing:
+            self.session.flush()
+
+        self._new = weakref.WeakKeyDictionary()
+        self._deleted = weakref.WeakKeyDictionary()
+        self._dirty = weakref.WeakKeyDictionary()
+        self._key_switches = weakref.WeakKeyDictionary()
+
+    def _restore_snapshot(self, dirty_only: bool = False) -> None:
+        """Restore the restoration state taken before a transaction began.
+
+        Corresponds to a rollback.
+
+        """
+        assert self._is_transaction_boundary
+
+        to_expunge = set(self._new).union(self.session._new)
+        self.session._expunge_states(to_expunge, to_transient=True)
+
+        for s, (oldkey, newkey) in self._key_switches.items():
+            # we probably can do this conditionally based on
+            # if we expunged or not, but safe_discard does that anyway
+            self.session.identity_map.safe_discard(s)
+
+            # restore the old key
+            s.key = oldkey
+
+            # now restore the object, but only if we didn't expunge
+            if s not in to_expunge:
+                self.session.identity_map.replace(s)
+
+        for s in set(self._deleted).union(self.session._deleted):
+            self.session._update_impl(s, revert_deletion=True)
+
+        assert not self.session._deleted
+
+        for s in self.session.identity_map.all_states():
+            if not dirty_only or s.modified or s in self._dirty:
+                s._expire(s.dict, self.session.identity_map._modified)
+
+    def _remove_snapshot(self) -> None:
+        """Remove the restoration state taken before a transaction began.
+
+        Corresponds to a commit.
+
+        """
+        assert self._is_transaction_boundary
+
+        if not self.nested and self.session.expire_on_commit:
+            for s in self.session.identity_map.all_states():
+                s._expire(s.dict, self.session.identity_map._modified)
+
+            statelib.InstanceState._detach_states(
+                list(self._deleted), self.session
+            )
+            self._deleted.clear()
+        elif self.nested:
+            parent = self._parent
+            assert parent is not None
+            parent._new.update(self._new)
+            parent._dirty.update(self._dirty)
+            parent._deleted.update(self._deleted)
+            parent._key_switches.update(self._key_switches)
+
+    @_StateChange.declare_states(
+        (SessionTransactionState.ACTIVE,), _StateChangeStates.NO_CHANGE
+    )
+    def _connection_for_bind(
+        self,
+        bind: _SessionBind,
+        execution_options: Optional[CoreExecuteOptionsParameter],
+    ) -> Connection:
+        if bind in self._connections:
+            if execution_options:
+                util.warn(
+                    "Connection is already established for the "
+                    "given bind; execution_options ignored"
+                )
+            return self._connections[bind][0]
+
+        self._state = SessionTransactionState.PROVISIONING_CONNECTION
+
+        local_connect = False
+        should_commit = True
+
+        try:
+            if self._parent:
+                conn = self._parent._connection_for_bind(
+                    bind, execution_options
+                )
+                if not self.nested:
+                    return conn
+            else:
+                if isinstance(bind, engine.Connection):
+                    conn = bind
+                    if conn.engine in self._connections:
+                        raise sa_exc.InvalidRequestError(
+                            "Session already has a Connection associated "
+                            "for the given Connection's Engine"
+                        )
+                else:
+                    conn = bind.connect()
+                    local_connect = True
+
+            try:
+                if execution_options:
+                    conn = conn.execution_options(**execution_options)
+
+                transaction: Transaction
+                if self.session.twophase and self._parent is None:
+                    # TODO: shouldn't we only be here if not
+                    # conn.in_transaction() ?
+                    # if twophase is set and conn.in_transaction(), validate
+                    # that it is in fact twophase.
+                    transaction = conn.begin_twophase()
+                elif self.nested:
+                    transaction = conn.begin_nested()
+                elif conn.in_transaction():
+                    join_transaction_mode = self.session.join_transaction_mode
+
+                    if join_transaction_mode == "conditional_savepoint":
+                        if conn.in_nested_transaction():
+                            join_transaction_mode = "create_savepoint"
+                        else:
+                            join_transaction_mode = "rollback_only"
+
+                        if local_connect:
+                            util.warn(
+                                "The engine provided as bind produced a "
+                                "connection that is already in a transaction. "
+                                "This is usually caused by a core event, "
+                                "such as 'engine_connect', that has left a "
+                                "transaction open. The effective join "
+                                "transaction mode used by this session is "
+                                f"{join_transaction_mode!r}. To silence this "
+                                "warning, do not leave transactions open"
+                            )
+                    if join_transaction_mode in (
+                        "control_fully",
+                        "rollback_only",
+                    ):
+                        if conn.in_nested_transaction():
+                            transaction = (
+                                conn._get_required_nested_transaction()
+                            )
+                        else:
+                            transaction = conn._get_required_transaction()
+                        if join_transaction_mode == "rollback_only":
+                            should_commit = False
+                    elif join_transaction_mode == "create_savepoint":
+                        transaction = conn.begin_nested()
+                    else:
+                        assert False, join_transaction_mode
+                else:
+                    transaction = conn.begin()
+            except:
+                # connection will not not be associated with this Session;
+                # close it immediately so that it isn't closed under GC
+                if local_connect:
+                    conn.close()
+                raise
+            else:
+                bind_is_connection = isinstance(bind, engine.Connection)
+
+                self._connections[conn] = self._connections[conn.engine] = (
+                    conn,
+                    transaction,
+                    should_commit,
+                    not bind_is_connection,
+                )
+                self.session.dispatch.after_begin(self.session, self, conn)
+                return conn
+        finally:
+            self._state = SessionTransactionState.ACTIVE
+
+    def prepare(self) -> None:
+        if self._parent is not None or not self.session.twophase:
+            raise sa_exc.InvalidRequestError(
+                "'twophase' mode not enabled, or not root transaction; "
+                "can't prepare."
+            )
+        self._prepare_impl()
+
+    @_StateChange.declare_states(
+        (SessionTransactionState.ACTIVE,), SessionTransactionState.PREPARED
+    )
+    def _prepare_impl(self) -> None:
+        if self._parent is None or self.nested:
+            self.session.dispatch.before_commit(self.session)
+
+        stx = self.session._transaction
+        assert stx is not None
+        if stx is not self:
+            for subtransaction in stx._iterate_self_and_parents(upto=self):
+                subtransaction.commit()
+
+        if not self.session._flushing:
+            for _flush_guard in range(100):
+                if self.session._is_clean():
+                    break
+                self.session.flush()
+            else:
+                raise exc.FlushError(
+                    "Over 100 subsequent flushes have occurred within "
+                    "session.commit() - is an after_flush() hook "
+                    "creating new objects?"
+                )
+
+        if self._parent is None and self.session.twophase:
+            try:
+                for t in set(self._connections.values()):
+                    cast("TwoPhaseTransaction", t[1]).prepare()
+            except:
+                with util.safe_reraise():
+                    self.rollback()
+
+        self._state = SessionTransactionState.PREPARED
+
+    @_StateChange.declare_states(
+        (SessionTransactionState.ACTIVE, SessionTransactionState.PREPARED),
+        SessionTransactionState.CLOSED,
+    )
+    def commit(self, _to_root: bool = False) -> None:
+        if self._state is not SessionTransactionState.PREPARED:
+            with self._expect_state(SessionTransactionState.PREPARED):
+                self._prepare_impl()
+
+        if self._parent is None or self.nested:
+            for conn, trans, should_commit, autoclose in set(
+                self._connections.values()
+            ):
+                if should_commit:
+                    trans.commit()
+
+            self._state = SessionTransactionState.COMMITTED
+            self.session.dispatch.after_commit(self.session)
+
+            self._remove_snapshot()
+
+        with self._expect_state(SessionTransactionState.CLOSED):
+            self.close()
+
+        if _to_root and self._parent:
+            self._parent.commit(_to_root=True)
+
+    @_StateChange.declare_states(
+        (
+            SessionTransactionState.ACTIVE,
+            SessionTransactionState.DEACTIVE,
+            SessionTransactionState.PREPARED,
+        ),
+        SessionTransactionState.CLOSED,
+    )
+    def rollback(
+        self, _capture_exception: bool = False, _to_root: bool = False
+    ) -> None:
+        stx = self.session._transaction
+        assert stx is not None
+        if stx is not self:
+            for subtransaction in stx._iterate_self_and_parents(upto=self):
+                subtransaction.close()
+
+        boundary = self
+        rollback_err = None
+        if self._state in (
+            SessionTransactionState.ACTIVE,
+            SessionTransactionState.PREPARED,
+        ):
+            for transaction in self._iterate_self_and_parents():
+                if transaction._parent is None or transaction.nested:
+                    try:
+                        for t in set(transaction._connections.values()):
+                            t[1].rollback()
+
+                        transaction._state = SessionTransactionState.DEACTIVE
+                        self.session.dispatch.after_rollback(self.session)
+                    except:
+                        rollback_err = sys.exc_info()
+                    finally:
+                        transaction._state = SessionTransactionState.DEACTIVE
+                        transaction._restore_snapshot(
+                            dirty_only=transaction.nested
+                        )
+                    boundary = transaction
+                    break
+                else:
+                    transaction._state = SessionTransactionState.DEACTIVE
+
+        sess = self.session
+
+        if not rollback_err and not sess._is_clean():
+            # if items were added, deleted, or mutated
+            # here, we need to re-restore the snapshot
+            util.warn(
+                "Session's state has been changed on "
+                "a non-active transaction - this state "
+                "will be discarded."
+            )
+            boundary._restore_snapshot(dirty_only=boundary.nested)
+
+        with self._expect_state(SessionTransactionState.CLOSED):
+            self.close()
+
+        if self._parent and _capture_exception:
+            self._parent._rollback_exception = sys.exc_info()[1]
+
+        if rollback_err and rollback_err[1]:
+            raise rollback_err[1].with_traceback(rollback_err[2])
+
+        sess.dispatch.after_soft_rollback(sess, self)
+
+        if _to_root and self._parent:
+            self._parent.rollback(_to_root=True)
+
+    @_StateChange.declare_states(
+        _StateChangeStates.ANY, SessionTransactionState.CLOSED
+    )
+    def close(self, invalidate: bool = False) -> None:
+        if self.nested:
+            self.session._nested_transaction = (
+                self._previous_nested_transaction
+            )
+
+        self.session._transaction = self._parent
+
+        for connection, transaction, should_commit, autoclose in set(
+            self._connections.values()
+        ):
+            if invalidate and self._parent is None:
+                connection.invalidate()
+            if should_commit and transaction.is_active:
+                transaction.close()
+            if autoclose and self._parent is None:
+                connection.close()
+
+        self._state = SessionTransactionState.CLOSED
+        sess = self.session
+
+        # TODO: these two None sets were historically after the
+        # event hook below, and in 2.0 I changed it this way for some reason,
+        # and I remember there being a reason, but not what it was.
+        # Why do we need to get rid of them at all?  test_memusage::CycleTest
+        # passes with these commented out.
+        # self.session = None  # type: ignore
+        # self._connections = None  # type: ignore
+
+        sess.dispatch.after_transaction_end(sess, self)
+
+    def _get_subject(self) -> Session:
+        return self.session
+
+    def _transaction_is_active(self) -> bool:
+        return self._state is SessionTransactionState.ACTIVE
+
+    def _transaction_is_closed(self) -> bool:
+        return self._state is SessionTransactionState.CLOSED
+
+    def _rollback_can_be_called(self) -> bool:
+        return self._state not in (COMMITTED, CLOSED)
+
+
+class _SessionCloseState(Enum):
+    ACTIVE = 1
+    CLOSED = 2
+    CLOSE_IS_RESET = 3
+
+
+class Session(_SessionClassMethods, EventTarget):
+    """Manages persistence operations for ORM-mapped objects.
+
+    The :class:`_orm.Session` is **not safe for use in concurrent threads.**.
+    See :ref:`session_faq_threadsafe` for background.
+
+    The Session's usage paradigm is described at :doc:`/orm/session`.
+
+
+    """
+
+    _is_asyncio = False
+
+    dispatch: dispatcher[Session]
+
+    identity_map: IdentityMap
+    """A mapping of object identities to objects themselves.
+
+    Iterating through ``Session.identity_map.values()`` provides
+    access to the full set of persistent objects (i.e., those
+    that have row identity) currently in the session.
+
+    .. seealso::
+
+        :func:`.identity_key` - helper function to produce the keys used
+        in this dictionary.
+
+    """
+
+    _new: Dict[InstanceState[Any], Any]
+    _deleted: Dict[InstanceState[Any], Any]
+    bind: Optional[Union[Engine, Connection]]
+    __binds: Dict[_SessionBindKey, _SessionBind]
+    _flushing: bool
+    _warn_on_events: bool
+    _transaction: Optional[SessionTransaction]
+    _nested_transaction: Optional[SessionTransaction]
+    hash_key: int
+    autoflush: bool
+    expire_on_commit: bool
+    enable_baked_queries: bool
+    twophase: bool
+    join_transaction_mode: JoinTransactionMode
+    _query_cls: Type[Query[Any]]
+    _close_state: _SessionCloseState
+
+    def __init__(
+        self,
+        bind: Optional[_SessionBind] = None,
+        *,
+        autoflush: bool = True,
+        future: Literal[True] = True,
+        expire_on_commit: bool = True,
+        autobegin: bool = True,
+        twophase: bool = False,
+        binds: Optional[Dict[_SessionBindKey, _SessionBind]] = None,
+        enable_baked_queries: bool = True,
+        info: Optional[_InfoType] = None,
+        query_cls: Optional[Type[Query[Any]]] = None,
+        autocommit: Literal[False] = False,
+        join_transaction_mode: JoinTransactionMode = "conditional_savepoint",
+        close_resets_only: Union[bool, _NoArg] = _NoArg.NO_ARG,
+    ):
+        r"""Construct a new :class:`_orm.Session`.
+
+        See also the :class:`.sessionmaker` function which is used to
+        generate a :class:`.Session`-producing callable with a given
+        set of arguments.
+
+        :param autoflush: When ``True``, all query operations will issue a
+           :meth:`~.Session.flush` call to this ``Session`` before proceeding.
+           This is a convenience feature so that :meth:`~.Session.flush` need
+           not be called repeatedly in order for database queries to retrieve
+           results.
+
+           .. seealso::
+
+               :ref:`session_flushing` - additional background on autoflush
+
+        :param autobegin: Automatically start transactions (i.e. equivalent to
+           invoking :meth:`_orm.Session.begin`) when database access is
+           requested by an operation.   Defaults to ``True``.    Set to
+           ``False`` to prevent a :class:`_orm.Session` from implicitly
+           beginning transactions after construction, as well as after any of
+           the :meth:`_orm.Session.rollback`, :meth:`_orm.Session.commit`,
+           or :meth:`_orm.Session.close` methods are called.
+
+           .. versionadded:: 2.0
+
+           .. seealso::
+
+                :ref:`session_autobegin_disable`
+
+        :param bind: An optional :class:`_engine.Engine` or
+           :class:`_engine.Connection` to
+           which this ``Session`` should be bound. When specified, all SQL
+           operations performed by this session will execute via this
+           connectable.
+
+        :param binds: A dictionary which may specify any number of
+           :class:`_engine.Engine` or :class:`_engine.Connection`
+           objects as the source of
+           connectivity for SQL operations on a per-entity basis.   The keys
+           of the dictionary consist of any series of mapped classes,
+           arbitrary Python classes that are bases for mapped classes,
+           :class:`_schema.Table` objects and :class:`_orm.Mapper` objects.
+           The
+           values of the dictionary are then instances of
+           :class:`_engine.Engine`
+           or less commonly :class:`_engine.Connection` objects.
+           Operations which
+           proceed relative to a particular mapped class will consult this
+           dictionary for the closest matching entity in order to determine
+           which :class:`_engine.Engine` should be used for a particular SQL
+           operation.    The complete heuristics for resolution are
+           described at :meth:`.Session.get_bind`.  Usage looks like::
+
+            Session = sessionmaker(
+                binds={
+                    SomeMappedClass: create_engine("postgresql+psycopg2://engine1"),
+                    SomeDeclarativeBase: create_engine(
+                        "postgresql+psycopg2://engine2"
+                    ),
+                    some_mapper: create_engine("postgresql+psycopg2://engine3"),
+                    some_table: create_engine("postgresql+psycopg2://engine4"),
+                }
+            )
+
+           .. seealso::
+
+                :ref:`session_partitioning`
+
+                :meth:`.Session.bind_mapper`
+
+                :meth:`.Session.bind_table`
+
+                :meth:`.Session.get_bind`
+
+
+        :param \class_: Specify an alternate class other than
+           ``sqlalchemy.orm.session.Session`` which should be used by the
+           returned class. This is the only argument that is local to the
+           :class:`.sessionmaker` function, and is not sent directly to the
+           constructor for ``Session``.
+
+        :param enable_baked_queries: legacy; defaults to ``True``.
+           A parameter consumed
+           by the :mod:`sqlalchemy.ext.baked` extension to determine if
+           "baked queries" should be cached, as is the normal operation
+           of this extension.  When set to ``False``, caching as used by
+           this particular extension is disabled.
+
+           .. versionchanged:: 1.4 The ``sqlalchemy.ext.baked`` extension is
+              legacy and is not used by any of SQLAlchemy's internals. This
+              flag therefore only affects applications that are making explicit
+              use of this extension within their own code.
+
+        :param expire_on_commit:  Defaults to ``True``. When ``True``, all
+           instances will be fully expired after each :meth:`~.commit`,
+           so that all attribute/object access subsequent to a completed
+           transaction will load from the most recent database state.
+
+            .. seealso::
+
+                :ref:`session_committing`
+
+        :param future: Deprecated; this flag is always True.
+
+          .. seealso::
+
+            :ref:`migration_20_toplevel`
+
+        :param info: optional dictionary of arbitrary data to be associated
+           with this :class:`.Session`.  Is available via the
+           :attr:`.Session.info` attribute.  Note the dictionary is copied at
+           construction time so that modifications to the per-
+           :class:`.Session` dictionary will be local to that
+           :class:`.Session`.
+
+        :param query_cls:  Class which should be used to create new Query
+          objects, as returned by the :meth:`~.Session.query` method.
+          Defaults to :class:`_query.Query`.
+
+        :param twophase:  When ``True``, all transactions will be started as
+            a "two phase" transaction, i.e. using the "two phase" semantics
+            of the database in use along with an XID.  During a
+            :meth:`~.commit`, after :meth:`~.flush` has been issued for all
+            attached databases, the :meth:`~.TwoPhaseTransaction.prepare`
+            method on each database's :class:`.TwoPhaseTransaction` will be
+            called. This allows each database to roll back the entire
+            transaction, before each transaction is committed.
+
+        :param autocommit: the "autocommit" keyword is present for backwards
+            compatibility but must remain at its default value of ``False``.
+
+        :param join_transaction_mode: Describes the transactional behavior to
+          take when a given bind is a :class:`_engine.Connection` that
+          has already begun a transaction outside the scope of this
+          :class:`_orm.Session`; in other words the
+          :meth:`_engine.Connection.in_transaction()` method returns True.
+
+          The following behaviors only take effect when the :class:`_orm.Session`
+          **actually makes use of the connection given**; that is, a method
+          such as :meth:`_orm.Session.execute`, :meth:`_orm.Session.connection`,
+          etc. are actually invoked:
+
+          * ``"conditional_savepoint"`` - this is the default.  if the given
+            :class:`_engine.Connection` is begun within a transaction but
+            does not have a SAVEPOINT, then ``"rollback_only"`` is used.
+            If the :class:`_engine.Connection` is additionally within
+            a SAVEPOINT, in other words
+            :meth:`_engine.Connection.in_nested_transaction()` method returns
+            True, then ``"create_savepoint"`` is used.
+
+            ``"conditional_savepoint"`` behavior attempts to make use of
+            savepoints in order to keep the state of the existing transaction
+            unchanged, but only if there is already a savepoint in progress;
+            otherwise, it is not assumed that the backend in use has adequate
+            support for SAVEPOINT, as availability of this feature varies.
+            ``"conditional_savepoint"`` also seeks to establish approximate
+            backwards compatibility with previous :class:`_orm.Session`
+            behavior, for applications that are not setting a specific mode. It
+            is recommended that one of the explicit settings be used.
+
+          * ``"create_savepoint"`` - the :class:`_orm.Session` will use
+            :meth:`_engine.Connection.begin_nested()` in all cases to create
+            its own transaction.  This transaction by its nature rides
+            "on top" of any existing transaction that's opened on the given
+            :class:`_engine.Connection`; if the underlying database and
+            the driver in use has full, non-broken support for SAVEPOINT, the
+            external transaction will remain unaffected throughout the
+            lifespan of the :class:`_orm.Session`.
+
+            The ``"create_savepoint"`` mode is the most useful for integrating
+            a :class:`_orm.Session` into a test suite where an externally
+            initiated transaction should remain unaffected; however, it relies
+            on proper SAVEPOINT support from the underlying driver and
+            database.
+
+            .. tip:: When using SQLite, the SQLite driver included through
+               Python 3.11 does not handle SAVEPOINTs correctly in all cases
+               without workarounds. See the sections
+               :ref:`pysqlite_serializable` and :ref:`aiosqlite_serializable`
+               for details on current workarounds.
+
+          * ``"control_fully"`` - the :class:`_orm.Session` will take
+            control of the given transaction as its own;
+            :meth:`_orm.Session.commit` will call ``.commit()`` on the
+            transaction, :meth:`_orm.Session.rollback` will call
+            ``.rollback()`` on the transaction, :meth:`_orm.Session.close` will
+            call ``.rollback`` on the transaction.
+
+            .. tip:: This mode of use is equivalent to how SQLAlchemy 1.4 would
+               handle a :class:`_engine.Connection` given with an existing
+               SAVEPOINT (i.e. :meth:`_engine.Connection.begin_nested`); the
+               :class:`_orm.Session` would take full control of the existing
+               SAVEPOINT.
+
+          * ``"rollback_only"`` - the :class:`_orm.Session` will take control
+            of the given transaction for ``.rollback()`` calls only;
+            ``.commit()`` calls will not be propagated to the given
+            transaction.  ``.close()`` calls will have no effect on the
+            given transaction.
+
+            .. tip:: This mode of use is equivalent to how SQLAlchemy 1.4 would
+               handle a :class:`_engine.Connection` given with an existing
+               regular database transaction (i.e.
+               :meth:`_engine.Connection.begin`); the :class:`_orm.Session`
+               would propagate :meth:`_orm.Session.rollback` calls to the
+               underlying transaction, but not :meth:`_orm.Session.commit` or
+               :meth:`_orm.Session.close` calls.
+
+          .. versionadded:: 2.0.0rc1
+
+        :param close_resets_only: Defaults to ``True``. Determines if
+          the session should reset itself after calling ``.close()``
+          or should pass in a no longer usable state, disabling re-use.
+
+          .. versionadded:: 2.0.22 added flag ``close_resets_only``.
+            A future SQLAlchemy version may change the default value of
+            this flag to ``False``.
+
+          .. seealso::
+
+            :ref:`session_closing` - Detail on the semantics of
+            :meth:`_orm.Session.close` and :meth:`_orm.Session.reset`.
+
+        """  # noqa
+
+        # considering allowing the "autocommit" keyword to still be accepted
+        # as long as it's False, so that external test suites, oslo.db etc
+        # continue to function as the argument appears to be passed in lots
+        # of cases including in our own test suite
+        if autocommit:
+            raise sa_exc.ArgumentError(
+                "autocommit=True is no longer supported"
+            )
+        self.identity_map = identity.WeakInstanceDict()
+
+        if not future:
+            raise sa_exc.ArgumentError(
+                "The 'future' parameter passed to "
+                "Session() may only be set to True."
+            )
+
+        self._new = {}  # InstanceState->object, strong refs object
+        self._deleted = {}  # same
+        self.bind = bind
+        self.__binds = {}
+        self._flushing = False
+        self._warn_on_events = False
+        self._transaction = None
+        self._nested_transaction = None
+        self.hash_key = _new_sessionid()
+        self.autobegin = autobegin
+        self.autoflush = autoflush
+        self.expire_on_commit = expire_on_commit
+        self.enable_baked_queries = enable_baked_queries
+
+        # the idea is that at some point NO_ARG will warn that in the future
+        # the default will switch to close_resets_only=False.
+        if close_resets_only in (True, _NoArg.NO_ARG):
+            self._close_state = _SessionCloseState.CLOSE_IS_RESET
+        else:
+            self._close_state = _SessionCloseState.ACTIVE
+        if (
+            join_transaction_mode
+            and join_transaction_mode
+            not in JoinTransactionMode.__args__  # type: ignore
+        ):
+            raise sa_exc.ArgumentError(
+                f"invalid selection for join_transaction_mode: "
+                f'"{join_transaction_mode}"'
+            )
+        self.join_transaction_mode = join_transaction_mode
+
+        self.twophase = twophase
+        self._query_cls = query_cls if query_cls else query.Query
+        if info:
+            self.info.update(info)
+
+        if binds is not None:
+            for key, bind in binds.items():
+                self._add_bind(key, bind)
+
+        _sessions[self.hash_key] = self
+
+    # used by sqlalchemy.engine.util.TransactionalContext
+    _trans_context_manager: Optional[TransactionalContext] = None
+
+    connection_callable: Optional[_ConnectionCallableProto] = None
+
+    def __enter__(self: _S) -> _S:
+        return self
+
+    def __exit__(self, type_: Any, value: Any, traceback: Any) -> None:
+        self.close()
+
+    @contextlib.contextmanager
+    def _maker_context_manager(self: _S) -> Iterator[_S]:
+        with self:
+            with self.begin():
+                yield self
+
+    def in_transaction(self) -> bool:
+        """Return True if this :class:`_orm.Session` has begun a transaction.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_orm.Session.is_active`
+
+
+        """
+        return self._transaction is not None
+
+    def in_nested_transaction(self) -> bool:
+        """Return True if this :class:`_orm.Session` has begun a nested
+        transaction, e.g. SAVEPOINT.
+
+        .. versionadded:: 1.4
+
+        """
+        return self._nested_transaction is not None
+
+    def get_transaction(self) -> Optional[SessionTransaction]:
+        """Return the current root transaction in progress, if any.
+
+        .. versionadded:: 1.4
+
+        """
+        trans = self._transaction
+        while trans is not None and trans._parent is not None:
+            trans = trans._parent
+        return trans
+
+    def get_nested_transaction(self) -> Optional[SessionTransaction]:
+        """Return the current nested transaction in progress, if any.
+
+        .. versionadded:: 1.4
+
+        """
+
+        return self._nested_transaction
+
+    @util.memoized_property
+    def info(self) -> _InfoType:
+        """A user-modifiable dictionary.
+
+        The initial value of this dictionary can be populated using the
+        ``info`` argument to the :class:`.Session` constructor or
+        :class:`.sessionmaker` constructor or factory methods.  The dictionary
+        here is always local to this :class:`.Session` and can be modified
+        independently of all other :class:`.Session` objects.
+
+        """
+        return {}
+
+    def _autobegin_t(self, begin: bool = False) -> SessionTransaction:
+        if self._transaction is None:
+            if not begin and not self.autobegin:
+                raise sa_exc.InvalidRequestError(
+                    "Autobegin is disabled on this Session; please call "
+                    "session.begin() to start a new transaction"
+                )
+            trans = SessionTransaction(
+                self,
+                (
+                    SessionTransactionOrigin.BEGIN
+                    if begin
+                    else SessionTransactionOrigin.AUTOBEGIN
+                ),
+            )
+            assert self._transaction is trans
+            return trans
+
+        return self._transaction
+
+    def begin(self, nested: bool = False) -> SessionTransaction:
+        """Begin a transaction, or nested transaction,
+        on this :class:`.Session`, if one is not already begun.
+
+        The :class:`_orm.Session` object features **autobegin** behavior,
+        so that normally it is not necessary to call the
+        :meth:`_orm.Session.begin`
+        method explicitly. However, it may be used in order to control
+        the scope of when the transactional state is begun.
+
+        When used to begin the outermost transaction, an error is raised
+        if this :class:`.Session` is already inside of a transaction.
+
+        :param nested: if True, begins a SAVEPOINT transaction and is
+         equivalent to calling :meth:`~.Session.begin_nested`. For
+         documentation on SAVEPOINT transactions, please see
+         :ref:`session_begin_nested`.
+
+        :return: the :class:`.SessionTransaction` object.  Note that
+         :class:`.SessionTransaction`
+         acts as a Python context manager, allowing :meth:`.Session.begin`
+         to be used in a "with" block.  See :ref:`session_explicit_begin` for
+         an example.
+
+        .. seealso::
+
+            :ref:`session_autobegin`
+
+            :ref:`unitofwork_transaction`
+
+            :meth:`.Session.begin_nested`
+
+
+        """
+
+        trans = self._transaction
+        if trans is None:
+            trans = self._autobegin_t(begin=True)
+
+            if not nested:
+                return trans
+
+        assert trans is not None
+
+        if nested:
+            trans = trans._begin(nested=nested)
+            assert self._transaction is trans
+            self._nested_transaction = trans
+        else:
+            raise sa_exc.InvalidRequestError(
+                "A transaction is already begun on this Session."
+            )
+
+        return trans  # needed for __enter__/__exit__ hook
+
+    def begin_nested(self) -> SessionTransaction:
+        """Begin a "nested" transaction on this Session, e.g. SAVEPOINT.
+
+        The target database(s) and associated drivers must support SQL
+        SAVEPOINT for this method to function correctly.
+
+        For documentation on SAVEPOINT
+        transactions, please see :ref:`session_begin_nested`.
+
+        :return: the :class:`.SessionTransaction` object.  Note that
+         :class:`.SessionTransaction` acts as a context manager, allowing
+         :meth:`.Session.begin_nested` to be used in a "with" block.
+         See :ref:`session_begin_nested` for a usage example.
+
+        .. seealso::
+
+            :ref:`session_begin_nested`
+
+            :ref:`pysqlite_serializable` - special workarounds required
+            with the SQLite driver in order for SAVEPOINT to work
+            correctly. For asyncio use cases, see the section
+            :ref:`aiosqlite_serializable`.
+
+        """
+        return self.begin(nested=True)
+
+    def rollback(self) -> None:
+        """Rollback the current transaction in progress.
+
+        If no transaction is in progress, this method is a pass-through.
+
+        The method always rolls back
+        the topmost database transaction, discarding any nested
+        transactions that may be in progress.
+
+        .. seealso::
+
+            :ref:`session_rollback`
+
+            :ref:`unitofwork_transaction`
+
+        """
+        if self._transaction is None:
+            pass
+        else:
+            self._transaction.rollback(_to_root=True)
+
+    def commit(self) -> None:
+        """Flush pending changes and commit the current transaction.
+
+        When the COMMIT operation is complete, all objects are fully
+        :term:`expired`, erasing their internal contents, which will be
+        automatically re-loaded when the objects are next accessed. In the
+        interim, these objects are in an expired state and will not function if
+        they are :term:`detached` from the :class:`.Session`. Additionally,
+        this re-load operation is not supported when using asyncio-oriented
+        APIs. The :paramref:`.Session.expire_on_commit` parameter may be used
+        to disable this behavior.
+
+        When there is no transaction in place for the :class:`.Session`,
+        indicating that no operations were invoked on this :class:`.Session`
+        since the previous call to :meth:`.Session.commit`, the method will
+        begin and commit an internal-only "logical" transaction, that does not
+        normally affect the database unless pending flush changes were
+        detected, but will still invoke event handlers and object expiration
+        rules.
+
+        The outermost database transaction is committed unconditionally,
+        automatically releasing any SAVEPOINTs in effect.
+
+        .. seealso::
+
+            :ref:`session_committing`
+
+            :ref:`unitofwork_transaction`
+
+            :ref:`asyncio_orm_avoid_lazyloads`
+
+        """
+        trans = self._transaction
+        if trans is None:
+            trans = self._autobegin_t()
+
+        trans.commit(_to_root=True)
+
+    def prepare(self) -> None:
+        """Prepare the current transaction in progress for two phase commit.
+
+        If no transaction is in progress, this method raises an
+        :exc:`~sqlalchemy.exc.InvalidRequestError`.
+
+        Only root transactions of two phase sessions can be prepared. If the
+        current transaction is not such, an
+        :exc:`~sqlalchemy.exc.InvalidRequestError` is raised.
+
+        """
+        trans = self._transaction
+        if trans is None:
+            trans = self._autobegin_t()
+
+        trans.prepare()
+
+    def connection(
+        self,
+        bind_arguments: Optional[_BindArguments] = None,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+    ) -> Connection:
+        r"""Return a :class:`_engine.Connection` object corresponding to this
+        :class:`.Session` object's transactional state.
+
+        Either the :class:`_engine.Connection` corresponding to the current
+        transaction is returned, or if no transaction is in progress, a new
+        one is begun and the :class:`_engine.Connection`
+        returned (note that no
+        transactional state is established with the DBAPI until the first
+        SQL statement is emitted).
+
+        Ambiguity in multi-bind or unbound :class:`.Session` objects can be
+        resolved through any of the optional keyword arguments.   This
+        ultimately makes usage of the :meth:`.get_bind` method for resolution.
+
+        :param bind_arguments: dictionary of bind arguments.  May include
+         "mapper", "bind", "clause", other custom arguments that are passed
+         to :meth:`.Session.get_bind`.
+
+        :param execution_options: a dictionary of execution options that will
+         be passed to :meth:`_engine.Connection.execution_options`, **when the
+         connection is first procured only**.   If the connection is already
+         present within the :class:`.Session`, a warning is emitted and
+         the arguments are ignored.
+
+         .. seealso::
+
+            :ref:`session_transaction_isolation`
+
+        """
+
+        if bind_arguments:
+            bind = bind_arguments.pop("bind", None)
+
+            if bind is None:
+                bind = self.get_bind(**bind_arguments)
+        else:
+            bind = self.get_bind()
+
+        return self._connection_for_bind(
+            bind,
+            execution_options=execution_options,
+        )
+
+    def _connection_for_bind(
+        self,
+        engine: _SessionBind,
+        execution_options: Optional[CoreExecuteOptionsParameter] = None,
+        **kw: Any,
+    ) -> Connection:
+        TransactionalContext._trans_ctx_check(self)
+
+        trans = self._transaction
+        if trans is None:
+            trans = self._autobegin_t()
+        return trans._connection_for_bind(engine, execution_options)
+
+    @overload
+    def _execute_internal(
+        self,
+        statement: Executable,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+        _scalar_result: Literal[True] = ...,
+    ) -> Any: ...
+
+    @overload
+    def _execute_internal(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+        _scalar_result: bool = ...,
+    ) -> Result[Any]: ...
+
+    def _execute_internal(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+        _scalar_result: bool = False,
+    ) -> Any:
+        statement = coercions.expect(roles.StatementRole, statement)
+
+        if not bind_arguments:
+            bind_arguments = {}
+        else:
+            bind_arguments = dict(bind_arguments)
+
+        if (
+            statement._propagate_attrs.get("compile_state_plugin", None)
+            == "orm"
+        ):
+            compile_state_cls = CompileState._get_plugin_class_for_plugin(
+                statement, "orm"
+            )
+            if TYPE_CHECKING:
+                assert isinstance(
+                    compile_state_cls, context.AbstractORMCompileState
+                )
+        else:
+            compile_state_cls = None
+            bind_arguments.setdefault("clause", statement)
+
+        execution_options = util.coerce_to_immutabledict(execution_options)
+
+        if _parent_execute_state:
+            events_todo = _parent_execute_state._remaining_events()
+        else:
+            events_todo = self.dispatch.do_orm_execute
+            if _add_event:
+                events_todo = list(events_todo) + [_add_event]
+
+        if events_todo:
+            if compile_state_cls is not None:
+                # for event handlers, do the orm_pre_session_exec
+                # pass ahead of the event handlers, so that things like
+                # .load_options, .update_delete_options etc. are populated.
+                # is_pre_event=True allows the hook to hold off on things
+                # it doesn't want to do twice, including autoflush as well
+                # as "pre fetch" for DML, etc.
+                (
+                    statement,
+                    execution_options,
+                ) = compile_state_cls.orm_pre_session_exec(
+                    self,
+                    statement,
+                    params,
+                    execution_options,
+                    bind_arguments,
+                    True,
+                )
+
+            orm_exec_state = ORMExecuteState(
+                self,
+                statement,
+                params,
+                execution_options,
+                bind_arguments,
+                compile_state_cls,
+                events_todo,
+            )
+            for idx, fn in enumerate(events_todo):
+                orm_exec_state._starting_event_idx = idx
+                fn_result: Optional[Result[Any]] = fn(orm_exec_state)
+                if fn_result:
+                    if _scalar_result:
+                        return fn_result.scalar()
+                    else:
+                        return fn_result
+
+            statement = orm_exec_state.statement
+            execution_options = orm_exec_state.local_execution_options
+
+        if compile_state_cls is not None:
+            # now run orm_pre_session_exec() "for real".   if there were
+            # event hooks, this will re-run the steps that interpret
+            # new execution_options into load_options / update_delete_options,
+            # which we assume the event hook might have updated.
+            # autoflush will also be invoked in this step if enabled.
+            (
+                statement,
+                execution_options,
+            ) = compile_state_cls.orm_pre_session_exec(
+                self,
+                statement,
+                params,
+                execution_options,
+                bind_arguments,
+                False,
+            )
+
+        bind = self.get_bind(**bind_arguments)
+
+        conn = self._connection_for_bind(bind)
+
+        if _scalar_result and not compile_state_cls:
+            if TYPE_CHECKING:
+                params = cast(_CoreSingleExecuteParams, params)
+            return conn.scalar(
+                statement, params or {}, execution_options=execution_options
+            )
+
+        if compile_state_cls:
+            result: Result[Any] = compile_state_cls.orm_execute_statement(
+                self,
+                statement,
+                params or {},
+                execution_options,
+                bind_arguments,
+                conn,
+            )
+        else:
+            result = conn.execute(
+                statement, params or {}, execution_options=execution_options
+            )
+
+        if _scalar_result:
+            return result.scalar()
+        else:
+            return result
+
+    @overload
+    def execute(
+        self,
+        statement: TypedReturnsRows[_T],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[_T]: ...
+
+    @overload
+    def execute(
+        self,
+        statement: UpdateBase,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> CursorResult[Any]: ...
+
+    @overload
+    def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]: ...
+
+    def execute(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        _parent_execute_state: Optional[Any] = None,
+        _add_event: Optional[Any] = None,
+    ) -> Result[Any]:
+        r"""Execute a SQL expression construct.
+
+        Returns a :class:`_engine.Result` object representing
+        results of the statement execution.
+
+        E.g.::
+
+            from sqlalchemy import select
+
+            result = session.execute(select(User).where(User.id == 5))
+
+        The API contract of :meth:`_orm.Session.execute` is similar to that
+        of :meth:`_engine.Connection.execute`, the :term:`2.0 style` version
+        of :class:`_engine.Connection`.
+
+        .. versionchanged:: 1.4 the :meth:`_orm.Session.execute` method is
+           now the primary point of ORM statement execution when using
+           :term:`2.0 style` ORM usage.
+
+        :param statement:
+            An executable statement (i.e. an :class:`.Executable` expression
+            such as :func:`_expression.select`).
+
+        :param params:
+            Optional dictionary, or list of dictionaries, containing
+            bound parameter values.   If a single dictionary, single-row
+            execution occurs; if a list of dictionaries, an
+            "executemany" will be invoked.  The keys in each dictionary
+            must correspond to parameter names present in the statement.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the statement execution.  This
+         dictionary can provide a subset of the options that are accepted
+         by :meth:`_engine.Connection.execution_options`, and may also
+         provide additional options understood only in an ORM context.
+
+         .. seealso::
+
+            :ref:`orm_queryguide_execution_options` - ORM-specific execution
+            options
+
+        :param bind_arguments: dictionary of additional arguments to determine
+         the bind.  May include "mapper", "bind", or other custom arguments.
+         Contents of this dictionary are passed to the
+         :meth:`.Session.get_bind` method.
+
+        :return: a :class:`_engine.Result` object.
+
+
+        """
+        return self._execute_internal(
+            statement,
+            params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            _parent_execute_state=_parent_execute_state,
+            _add_event=_add_event,
+        )
+
+    @overload
+    def scalar(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Optional[_T]: ...
+
+    @overload
+    def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any: ...
+
+    def scalar(
+        self,
+        statement: Executable,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> Any:
+        """Execute a statement and return a scalar result.
+
+        Usage and parameters are the same as that of
+        :meth:`_orm.Session.execute`; the return result is a scalar Python
+        value.
+
+        """
+
+        return self._execute_internal(
+            statement,
+            params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            _scalar_result=True,
+            **kw,
+        )
+
+    @overload
+    def scalars(
+        self,
+        statement: TypedReturnsRows[Tuple[_T]],
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[_T]: ...
+
+    @overload
+    def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]: ...
+
+    def scalars(
+        self,
+        statement: Executable,
+        params: Optional[_CoreAnyExecuteParams] = None,
+        *,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+        **kw: Any,
+    ) -> ScalarResult[Any]:
+        """Execute a statement and return the results as scalars.
+
+        Usage and parameters are the same as that of
+        :meth:`_orm.Session.execute`; the return result is a
+        :class:`_result.ScalarResult` filtering object which
+        will return single elements rather than :class:`_row.Row` objects.
+
+        :return:  a :class:`_result.ScalarResult` object
+
+        .. versionadded:: 1.4.24 Added :meth:`_orm.Session.scalars`
+
+        .. versionadded:: 1.4.26 Added :meth:`_orm.scoped_session.scalars`
+
+        .. seealso::
+
+            :ref:`orm_queryguide_select_orm_entities` - contrasts the behavior
+            of :meth:`_orm.Session.execute` to :meth:`_orm.Session.scalars`
+
+        """
+
+        return self._execute_internal(
+            statement,
+            params=params,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+            _scalar_result=False,  # mypy appreciates this
+            **kw,
+        ).scalars()
+
+    def close(self) -> None:
+        """Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`.
+
+        This expunges all ORM objects associated with this
+        :class:`_orm.Session`, ends any transaction in progress and
+        :term:`releases` any :class:`_engine.Connection` objects which this
+        :class:`_orm.Session` itself has checked out from associated
+        :class:`_engine.Engine` objects. The operation then leaves the
+        :class:`_orm.Session` in a state which it may be used again.
+
+        .. tip::
+
+            In the default running mode the :meth:`_orm.Session.close`
+            method **does not prevent the Session from being used again**.
+            The :class:`_orm.Session` itself does not actually have a
+            distinct "closed" state; it merely means
+            the :class:`_orm.Session` will release all database connections
+            and ORM objects.
+
+            Setting the parameter :paramref:`_orm.Session.close_resets_only`
+            to ``False`` will instead make the ``close`` final, meaning that
+            any further action on the session will be forbidden.
+
+        .. versionchanged:: 1.4  The :meth:`.Session.close` method does not
+           immediately create a new :class:`.SessionTransaction` object;
+           instead, the new :class:`.SessionTransaction` is created only if
+           the :class:`.Session` is used again for a database operation.
+
+        .. seealso::
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_orm.Session.close` and :meth:`_orm.Session.reset`.
+
+            :meth:`_orm.Session.reset` - a similar method that behaves like
+            ``close()`` with  the parameter
+            :paramref:`_orm.Session.close_resets_only` set to ``True``.
+
+        """
+        self._close_impl(invalidate=False)
+
+    def reset(self) -> None:
+        """Close out the transactional resources and ORM objects used by this
+        :class:`_orm.Session`, resetting the session to its initial state.
+
+        This method provides for same "reset-only" behavior that the
+        :meth:`_orm.Session.close` method has provided historically, where the
+        state of the :class:`_orm.Session` is reset as though the object were
+        brand new, and ready to be used again.
+        This method may then be useful for :class:`_orm.Session` objects
+        which set :paramref:`_orm.Session.close_resets_only` to ``False``,
+        so that "reset only" behavior is still available.
+
+        .. versionadded:: 2.0.22
+
+        .. seealso::
+
+            :ref:`session_closing` - detail on the semantics of
+            :meth:`_orm.Session.close` and :meth:`_orm.Session.reset`.
+
+            :meth:`_orm.Session.close` - a similar method will additionally
+            prevent re-use of the Session when the parameter
+            :paramref:`_orm.Session.close_resets_only` is set to ``False``.
+        """
+        self._close_impl(invalidate=False, is_reset=True)
+
+    def invalidate(self) -> None:
+        """Close this Session, using connection invalidation.
+
+        This is a variant of :meth:`.Session.close` that will additionally
+        ensure that the :meth:`_engine.Connection.invalidate`
+        method will be called on each :class:`_engine.Connection` object
+        that is currently in use for a transaction (typically there is only
+        one connection unless the :class:`_orm.Session` is used with
+        multiple engines).
+
+        This can be called when the database is known to be in a state where
+        the connections are no longer safe to be used.
+
+        Below illustrates a scenario when using `gevent
+        <https://www.gevent.org/>`_, which can produce ``Timeout`` exceptions
+        that may mean the underlying connection should be discarded::
+
+            import gevent
+
+            try:
+                sess = Session()
+                sess.add(User())
+                sess.commit()
+            except gevent.Timeout:
+                sess.invalidate()
+                raise
+            except:
+                sess.rollback()
+                raise
+
+        The method additionally does everything that :meth:`_orm.Session.close`
+        does, including that all ORM objects are expunged.
+
+        """
+        self._close_impl(invalidate=True)
+
+    def _close_impl(self, invalidate: bool, is_reset: bool = False) -> None:
+        if not is_reset and self._close_state is _SessionCloseState.ACTIVE:
+            self._close_state = _SessionCloseState.CLOSED
+        self.expunge_all()
+        if self._transaction is not None:
+            for transaction in self._transaction._iterate_self_and_parents():
+                transaction.close(invalidate)
+
+    def expunge_all(self) -> None:
+        """Remove all object instances from this ``Session``.
+
+        This is equivalent to calling ``expunge(obj)`` on all objects in this
+        ``Session``.
+
+        """
+
+        all_states = self.identity_map.all_states() + list(self._new)
+        self.identity_map._kill()
+        self.identity_map = identity.WeakInstanceDict()
+        self._new = {}
+        self._deleted = {}
+
+        statelib.InstanceState._detach_states(all_states, self)
+
+    def _add_bind(self, key: _SessionBindKey, bind: _SessionBind) -> None:
+        try:
+            insp = inspect(key)
+        except sa_exc.NoInspectionAvailable as err:
+            if not isinstance(key, type):
+                raise sa_exc.ArgumentError(
+                    "Not an acceptable bind target: %s" % key
+                ) from err
+            else:
+                self.__binds[key] = bind
+        else:
+            if TYPE_CHECKING:
+                assert isinstance(insp, Inspectable)
+
+            if isinstance(insp, TableClause):
+                self.__binds[insp] = bind
+            elif insp_is_mapper(insp):
+                self.__binds[insp.class_] = bind
+                for _selectable in insp._all_tables:
+                    self.__binds[_selectable] = bind
+            else:
+                raise sa_exc.ArgumentError(
+                    "Not an acceptable bind target: %s" % key
+                )
+
+    def bind_mapper(
+        self, mapper: _EntityBindKey[_O], bind: _SessionBind
+    ) -> None:
+        """Associate a :class:`_orm.Mapper` or arbitrary Python class with a
+        "bind", e.g. an :class:`_engine.Engine` or
+        :class:`_engine.Connection`.
+
+        The given entity is added to a lookup used by the
+        :meth:`.Session.get_bind` method.
+
+        :param mapper: a :class:`_orm.Mapper` object,
+         or an instance of a mapped
+         class, or any Python class that is the base of a set of mapped
+         classes.
+
+        :param bind: an :class:`_engine.Engine` or :class:`_engine.Connection`
+                    object.
+
+        .. seealso::
+
+            :ref:`session_partitioning`
+
+            :paramref:`.Session.binds`
+
+            :meth:`.Session.bind_table`
+
+
+        """
+        self._add_bind(mapper, bind)
+
+    def bind_table(self, table: TableClause, bind: _SessionBind) -> None:
+        """Associate a :class:`_schema.Table` with a "bind", e.g. an
+        :class:`_engine.Engine`
+        or :class:`_engine.Connection`.
+
+        The given :class:`_schema.Table` is added to a lookup used by the
+        :meth:`.Session.get_bind` method.
+
+        :param table: a :class:`_schema.Table` object,
+         which is typically the target
+         of an ORM mapping, or is present within a selectable that is
+         mapped.
+
+        :param bind: an :class:`_engine.Engine` or :class:`_engine.Connection`
+         object.
+
+        .. seealso::
+
+            :ref:`session_partitioning`
+
+            :paramref:`.Session.binds`
+
+            :meth:`.Session.bind_mapper`
+
+
+        """
+        self._add_bind(table, bind)
+
+    def get_bind(
+        self,
+        mapper: Optional[_EntityBindKey[_O]] = None,
+        *,
+        clause: Optional[ClauseElement] = None,
+        bind: Optional[_SessionBind] = None,
+        _sa_skip_events: Optional[bool] = None,
+        _sa_skip_for_implicit_returning: bool = False,
+        **kw: Any,
+    ) -> Union[Engine, Connection]:
+        """Return a "bind" to which this :class:`.Session` is bound.
+
+        The "bind" is usually an instance of :class:`_engine.Engine`,
+        except in the case where the :class:`.Session` has been
+        explicitly bound directly to a :class:`_engine.Connection`.
+
+        For a multiply-bound or unbound :class:`.Session`, the
+        ``mapper`` or ``clause`` arguments are used to determine the
+        appropriate bind to return.
+
+        Note that the "mapper" argument is usually present
+        when :meth:`.Session.get_bind` is called via an ORM
+        operation such as a :meth:`.Session.query`, each
+        individual INSERT/UPDATE/DELETE operation within a
+        :meth:`.Session.flush`, call, etc.
+
+        The order of resolution is:
+
+        1. if mapper given and :paramref:`.Session.binds` is present,
+           locate a bind based first on the mapper in use, then
+           on the mapped class in use, then on any base classes that are
+           present in the ``__mro__`` of the mapped class, from more specific
+           superclasses to more general.
+        2. if clause given and ``Session.binds`` is present,
+           locate a bind based on :class:`_schema.Table` objects
+           found in the given clause present in ``Session.binds``.
+        3. if ``Session.binds`` is present, return that.
+        4. if clause given, attempt to return a bind
+           linked to the :class:`_schema.MetaData` ultimately
+           associated with the clause.
+        5. if mapper given, attempt to return a bind
+           linked to the :class:`_schema.MetaData` ultimately
+           associated with the :class:`_schema.Table` or other
+           selectable to which the mapper is mapped.
+        6. No bind can be found, :exc:`~sqlalchemy.exc.UnboundExecutionError`
+           is raised.
+
+        Note that the :meth:`.Session.get_bind` method can be overridden on
+        a user-defined subclass of :class:`.Session` to provide any kind
+        of bind resolution scheme.  See the example at
+        :ref:`session_custom_partitioning`.
+
+        :param mapper:
+          Optional mapped class or corresponding :class:`_orm.Mapper` instance.
+          The bind can be derived from a :class:`_orm.Mapper` first by
+          consulting the "binds" map associated with this :class:`.Session`,
+          and secondly by consulting the :class:`_schema.MetaData` associated
+          with the :class:`_schema.Table` to which the :class:`_orm.Mapper` is
+          mapped for a bind.
+
+        :param clause:
+            A :class:`_expression.ClauseElement` (i.e.
+            :func:`_expression.select`,
+            :func:`_expression.text`,
+            etc.).  If the ``mapper`` argument is not present or could not
+            produce a bind, the given expression construct will be searched
+            for a bound element, typically a :class:`_schema.Table`
+            associated with
+            bound :class:`_schema.MetaData`.
+
+        .. seealso::
+
+             :ref:`session_partitioning`
+
+             :paramref:`.Session.binds`
+
+             :meth:`.Session.bind_mapper`
+
+             :meth:`.Session.bind_table`
+
+        """
+
+        # this function is documented as a subclassing hook, so we have
+        # to call this method even if the return is simple
+        if bind:
+            return bind
+        elif not self.__binds and self.bind:
+            # simplest and most common case, we have a bind and no
+            # per-mapper/table binds, we're done
+            return self.bind
+
+        # we don't have self.bind and either have self.__binds
+        # or we don't have self.__binds (which is legacy).  Look at the
+        # mapper and the clause
+        if mapper is None and clause is None:
+            if self.bind:
+                return self.bind
+            else:
+                raise sa_exc.UnboundExecutionError(
+                    "This session is not bound to a single Engine or "
+                    "Connection, and no context was provided to locate "
+                    "a binding."
+                )
+
+        # look more closely at the mapper.
+        if mapper is not None:
+            try:
+                inspected_mapper = inspect(mapper)
+            except sa_exc.NoInspectionAvailable as err:
+                if isinstance(mapper, type):
+                    raise exc.UnmappedClassError(mapper) from err
+                else:
+                    raise
+        else:
+            inspected_mapper = None
+
+        # match up the mapper or clause in the __binds
+        if self.__binds:
+            # matching mappers and selectables to entries in the
+            # binds dictionary; supported use case.
+            if inspected_mapper:
+                for cls in inspected_mapper.class_.__mro__:
+                    if cls in self.__binds:
+                        return self.__binds[cls]
+                if clause is None:
+                    clause = inspected_mapper.persist_selectable
+
+            if clause is not None:
+                plugin_subject = clause._propagate_attrs.get(
+                    "plugin_subject", None
+                )
+
+                if plugin_subject is not None:
+                    for cls in plugin_subject.mapper.class_.__mro__:
+                        if cls in self.__binds:
+                            return self.__binds[cls]
+
+                for obj in visitors.iterate(clause):
+                    if obj in self.__binds:
+                        if TYPE_CHECKING:
+                            assert isinstance(obj, Table)
+                        return self.__binds[obj]
+
+        # none of the __binds matched, but we have a fallback bind.
+        # return that
+        if self.bind:
+            return self.bind
+
+        context = []
+        if inspected_mapper is not None:
+            context.append(f"mapper {inspected_mapper}")
+        if clause is not None:
+            context.append("SQL expression")
+
+        raise sa_exc.UnboundExecutionError(
+            f"Could not locate a bind configured on "
+            f'{", ".join(context)} or this Session.'
+        )
+
+    @overload
+    def query(self, _entity: _EntityType[_O]) -> Query[_O]: ...
+
+    @overload
+    def query(
+        self, _colexpr: TypedColumnsClauseRole[_T]
+    ) -> RowReturningQuery[Tuple[_T]]: ...
+
+    # START OVERLOADED FUNCTIONS self.query RowReturningQuery 2-8
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_tuple_map_overloads.py
+
+    @overload
+    def query(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+    ) -> RowReturningQuery[Tuple[_T0, _T1]]: ...
+
+    @overload
+    def query(
+        self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+    @overload
+    def query(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+        __ent7: _TCCA[_T7],
+    ) -> RowReturningQuery[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]]: ...
+
+    # END OVERLOADED FUNCTIONS self.query
+
+    @overload
+    def query(
+        self, *entities: _ColumnsClauseArgument[Any], **kwargs: Any
+    ) -> Query[Any]: ...
+
+    def query(
+        self, *entities: _ColumnsClauseArgument[Any], **kwargs: Any
+    ) -> Query[Any]:
+        """Return a new :class:`_query.Query` object corresponding to this
+        :class:`_orm.Session`.
+
+        Note that the :class:`_query.Query` object is legacy as of
+        SQLAlchemy 2.0; the :func:`_sql.select` construct is now used
+        to construct ORM queries.
+
+        .. seealso::
+
+            :ref:`unified_tutorial`
+
+            :ref:`queryguide_toplevel`
+
+            :ref:`query_api_toplevel` - legacy API doc
+
+        """
+
+        return self._query_cls(entities, self, **kwargs)
+
+    def _identity_lookup(
+        self,
+        mapper: Mapper[_O],
+        primary_key_identity: Union[Any, Tuple[Any, ...]],
+        identity_token: Any = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+        lazy_loaded_from: Optional[InstanceState[Any]] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> Union[Optional[_O], LoaderCallableStatus]:
+        """Locate an object in the identity map.
+
+        Given a primary key identity, constructs an identity key and then
+        looks in the session's identity map.  If present, the object may
+        be run through unexpiration rules (e.g. load unloaded attributes,
+        check if was deleted).
+
+        e.g.::
+
+            obj = session._identity_lookup(inspect(SomeClass), (1,))
+
+        :param mapper: mapper in use
+        :param primary_key_identity: the primary key we are searching for, as
+         a tuple.
+        :param identity_token: identity token that should be used to create
+         the identity key.  Used as is, however overriding subclasses can
+         repurpose this in order to interpret the value in a special way,
+         such as if None then look among multiple target tokens.
+        :param passive: passive load flag passed to
+         :func:`.loading.get_from_identity`, which impacts the behavior if
+         the object is found; the object may be validated and/or unexpired
+         if the flag allows for SQL to be emitted.
+        :param lazy_loaded_from: an :class:`.InstanceState` that is
+         specifically asking for this identity as a related identity.  Used
+         for sharding schemes where there is a correspondence between an object
+         and a related object being lazy-loaded (or otherwise
+         relationship-loaded).
+
+        :return: None if the object is not found in the identity map, *or*
+         if the object was unexpired and found to have been deleted.
+         if passive flags disallow SQL and the object is expired, returns
+         PASSIVE_NO_RESULT.   In all other cases the instance is returned.
+
+        .. versionchanged:: 1.4.0 - the :meth:`.Session._identity_lookup`
+           method was moved from :class:`_query.Query` to
+           :class:`.Session`, to avoid having to instantiate the
+           :class:`_query.Query` object.
+
+
+        """
+
+        key = mapper.identity_key_from_primary_key(
+            primary_key_identity, identity_token=identity_token
+        )
+
+        # work around: https://github.com/python/typing/discussions/1143
+        return_value = loading.get_from_identity(self, mapper, key, passive)
+        return return_value
+
+    @util.non_memoized_property
+    @contextlib.contextmanager
+    def no_autoflush(self) -> Iterator[Session]:
+        """Return a context manager that disables autoflush.
+
+        e.g.::
+
+            with session.no_autoflush:
+
+                some_object = SomeClass()
+                session.add(some_object)
+                # won't autoflush
+                some_object.related_thing = session.query(SomeRelated).first()
+
+        Operations that proceed within the ``with:`` block
+        will not be subject to flushes occurring upon query
+        access.  This is useful when initializing a series
+        of objects which involve existing database queries,
+        where the uncompleted object should not yet be flushed.
+
+        """
+        autoflush = self.autoflush
+        self.autoflush = False
+        try:
+            yield self
+        finally:
+            self.autoflush = autoflush
+
+    @util.langhelpers.tag_method_for_warnings(
+        "This warning originated from the Session 'autoflush' process, "
+        "which was invoked automatically in response to a user-initiated "
+        "operation. Consider using ``no_autoflush`` context manager if this "
+        "warning happended while initializing objects.",
+        sa_exc.SAWarning,
+    )
+    def _autoflush(self) -> None:
+        if self.autoflush and not self._flushing:
+            try:
+                self.flush()
+            except sa_exc.StatementError as e:
+                # note we are reraising StatementError as opposed to
+                # raising FlushError with "chaining" to remain compatible
+                # with code that catches StatementError, IntegrityError,
+                # etc.
+                e.add_detail(
+                    "raised as a result of Query-invoked autoflush; "
+                    "consider using a session.no_autoflush block if this "
+                    "flush is occurring prematurely"
+                )
+                raise e.with_traceback(sys.exc_info()[2])
+
+    def refresh(
+        self,
+        instance: object,
+        attribute_names: Optional[Iterable[str]] = None,
+        with_for_update: ForUpdateParameter = None,
+    ) -> None:
+        """Expire and refresh attributes on the given instance.
+
+        The selected attributes will first be expired as they would when using
+        :meth:`_orm.Session.expire`; then a SELECT statement will be issued to
+        the database to refresh column-oriented attributes with the current
+        value available in the current transaction.
+
+        :func:`_orm.relationship` oriented attributes will also be immediately
+        loaded if they were already eagerly loaded on the object, using the
+        same eager loading strategy that they were loaded with originally.
+
+        .. versionadded:: 1.4 - the :meth:`_orm.Session.refresh` method
+           can also refresh eagerly loaded attributes.
+
+        :func:`_orm.relationship` oriented attributes that would normally
+        load using the ``select`` (or "lazy") loader strategy will also
+        load **if they are named explicitly in the attribute_names
+        collection**, emitting a SELECT statement for the attribute using the
+        ``immediate`` loader strategy.  If lazy-loaded relationships are not
+        named in :paramref:`_orm.Session.refresh.attribute_names`, then
+        they remain as "lazy loaded" attributes and are not implicitly
+        refreshed.
+
+        .. versionchanged:: 2.0.4  The :meth:`_orm.Session.refresh` method
+           will now refresh lazy-loaded :func:`_orm.relationship` oriented
+           attributes for those which are named explicitly in the
+           :paramref:`_orm.Session.refresh.attribute_names` collection.
+
+        .. tip::
+
+            While the :meth:`_orm.Session.refresh` method is capable of
+            refreshing both column and relationship oriented attributes, its
+            primary focus is on refreshing of local column-oriented attributes
+            on a single instance. For more open ended "refresh" functionality,
+            including the ability to refresh the attributes on many objects at
+            once while having explicit control over relationship loader
+            strategies, use the
+            :ref:`populate existing <orm_queryguide_populate_existing>` feature
+            instead.
+
+        Note that a highly isolated transaction will return the same values as
+        were previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.   Refreshing
+        attributes usually only makes sense at the start of a transaction
+        where database rows have not yet been accessed.
+
+        :param attribute_names: optional.  An iterable collection of
+          string attribute names indicating a subset of attributes to
+          be refreshed.
+
+        :param with_for_update: optional boolean ``True`` indicating FOR UPDATE
+          should be used, or may be a dictionary containing flags to
+          indicate a more specific set of FOR UPDATE flags for the SELECT;
+          flags should match the parameters of
+          :meth:`_query.Query.with_for_update`.
+          Supersedes the :paramref:`.Session.refresh.lockmode` parameter.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.expire_all`
+
+            :ref:`orm_queryguide_populate_existing` - allows any ORM query
+            to refresh objects as they would be loaded normally.
+
+        """
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+
+        self._expire_state(state, attribute_names)
+
+        # this autoflush previously used to occur as a secondary effect
+        # of the load_on_ident below.   Meaning we'd organize the SELECT
+        # based on current DB pks, then flush, then if pks changed in that
+        # flush, crash.  this was unticketed but discovered as part of
+        # #8703.  So here, autoflush up front, dont autoflush inside
+        # load_on_ident.
+        self._autoflush()
+
+        if with_for_update == {}:
+            raise sa_exc.ArgumentError(
+                "with_for_update should be the boolean value "
+                "True, or a dictionary with options.  "
+                "A blank dictionary is ambiguous."
+            )
+
+        with_for_update = ForUpdateArg._from_argument(with_for_update)
+
+        stmt: Select[Any] = sql.select(object_mapper(instance))
+        if (
+            loading.load_on_ident(
+                self,
+                stmt,
+                state.key,
+                refresh_state=state,
+                with_for_update=with_for_update,
+                only_load_props=attribute_names,
+                require_pk_cols=True,
+                # technically unnecessary as we just did autoflush
+                # above, however removes the additional unnecessary
+                # call to _autoflush()
+                no_autoflush=True,
+                is_user_refresh=True,
+            )
+            is None
+        ):
+            raise sa_exc.InvalidRequestError(
+                "Could not refresh instance '%s'" % instance_str(instance)
+            )
+
+    def expire_all(self) -> None:
+        """Expires all persistent instances within this Session.
+
+        When any attributes on a persistent instance is next accessed,
+        a query will be issued using the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire individual objects and individual attributes
+        on those objects, use :meth:`Session.expire`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire_all` is not usually needed,
+        assuming the transaction is isolated.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+        """
+        for state in self.identity_map.all_states():
+            state._expire(state.dict, self.identity_map._modified)
+
+    def expire(
+        self, instance: object, attribute_names: Optional[Iterable[str]] = None
+    ) -> None:
+        """Expire the attributes on an instance.
+
+        Marks the attributes of an instance as out of date. When an expired
+        attribute is next accessed, a query will be issued to the
+        :class:`.Session` object's current transactional context in order to
+        load all expired attributes for the given instance.   Note that
+        a highly isolated transaction will return the same values as were
+        previously read in that same transaction, regardless of changes
+        in database state outside of that transaction.
+
+        To expire all objects in the :class:`.Session` simultaneously,
+        use :meth:`Session.expire_all`.
+
+        The :class:`.Session` object's default behavior is to
+        expire all state whenever the :meth:`Session.rollback`
+        or :meth:`Session.commit` methods are called, so that new
+        state can be loaded for the new transaction.   For this reason,
+        calling :meth:`Session.expire` only makes sense for the specific
+        case that a non-ORM SQL statement was emitted in the current
+        transaction.
+
+        :param instance: The instance to be refreshed.
+        :param attribute_names: optional list of string attribute names
+          indicating a subset of attributes to be expired.
+
+        .. seealso::
+
+            :ref:`session_expire` - introductory material
+
+            :meth:`.Session.expire`
+
+            :meth:`.Session.refresh`
+
+            :meth:`_orm.Query.populate_existing`
+
+        """
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+        self._expire_state(state, attribute_names)
+
+    def _expire_state(
+        self,
+        state: InstanceState[Any],
+        attribute_names: Optional[Iterable[str]],
+    ) -> None:
+        self._validate_persistent(state)
+        if attribute_names:
+            state._expire_attributes(state.dict, attribute_names)
+        else:
+            # pre-fetch the full cascade since the expire is going to
+            # remove associations
+            cascaded = list(
+                state.manager.mapper.cascade_iterator("refresh-expire", state)
+            )
+            self._conditional_expire(state)
+            for o, m, st_, dct_ in cascaded:
+                self._conditional_expire(st_)
+
+    def _conditional_expire(
+        self, state: InstanceState[Any], autoflush: Optional[bool] = None
+    ) -> None:
+        """Expire a state if persistent, else expunge if pending"""
+
+        if state.key:
+            state._expire(state.dict, self.identity_map._modified)
+        elif state in self._new:
+            self._new.pop(state)
+            state._detach(self)
+
+    def expunge(self, instance: object) -> None:
+        """Remove the `instance` from this ``Session``.
+
+        This will free all internal references to the instance.  Cascading
+        will be applied according to the *expunge* cascade rule.
+
+        """
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+        if state.session_id is not self.hash_key:
+            raise sa_exc.InvalidRequestError(
+                "Instance %s is not present in this Session" % state_str(state)
+            )
+
+        cascaded = list(
+            state.manager.mapper.cascade_iterator("expunge", state)
+        )
+        self._expunge_states([state] + [st_ for o, m, st_, dct_ in cascaded])
+
+    def _expunge_states(
+        self, states: Iterable[InstanceState[Any]], to_transient: bool = False
+    ) -> None:
+        for state in states:
+            if state in self._new:
+                self._new.pop(state)
+            elif self.identity_map.contains_state(state):
+                self.identity_map.safe_discard(state)
+                self._deleted.pop(state, None)
+            elif self._transaction:
+                # state is "detached" from being deleted, but still present
+                # in the transaction snapshot
+                self._transaction._deleted.pop(state, None)
+        statelib.InstanceState._detach_states(
+            states, self, to_transient=to_transient
+        )
+
+    def _register_persistent(self, states: Set[InstanceState[Any]]) -> None:
+        """Register all persistent objects from a flush.
+
+        This is used both for pending objects moving to the persistent
+        state as well as already persistent objects.
+
+        """
+
+        pending_to_persistent = self.dispatch.pending_to_persistent or None
+        for state in states:
+            mapper = _state_mapper(state)
+
+            # prevent against last minute dereferences of the object
+            obj = state.obj()
+            if obj is not None:
+                instance_key = mapper._identity_key_from_state(state)
+
+                if (
+                    _none_set.intersection(instance_key[1])
+                    and not mapper.allow_partial_pks
+                    or _none_set.issuperset(instance_key[1])
+                ):
+                    raise exc.FlushError(
+                        "Instance %s has a NULL identity key.  If this is an "
+                        "auto-generated value, check that the database table "
+                        "allows generation of new primary key values, and "
+                        "that the mapped Column object is configured to "
+                        "expect these generated values.  Ensure also that "
+                        "this flush() is not occurring at an inappropriate "
+                        "time, such as within a load() event."
+                        % state_str(state)
+                    )
+
+                if state.key is None:
+                    state.key = instance_key
+                elif state.key != instance_key:
+                    # primary key switch. use safe_discard() in case another
+                    # state has already replaced this one in the identity
+                    # map (see test/orm/test_naturalpks.py ReversePKsTest)
+                    self.identity_map.safe_discard(state)
+                    trans = self._transaction
+                    assert trans is not None
+                    if state in trans._key_switches:
+                        orig_key = trans._key_switches[state][0]
+                    else:
+                        orig_key = state.key
+                    trans._key_switches[state] = (
+                        orig_key,
+                        instance_key,
+                    )
+                    state.key = instance_key
+
+                # there can be an existing state in the identity map
+                # that is replaced when the primary keys of two instances
+                # are swapped; see test/orm/test_naturalpks.py -> test_reverse
+                old = self.identity_map.replace(state)
+                if (
+                    old is not None
+                    and mapper._identity_key_from_state(old) == instance_key
+                    and old.obj() is not None
+                ):
+                    util.warn(
+                        "Identity map already had an identity for %s, "
+                        "replacing it with newly flushed object.   Are there "
+                        "load operations occurring inside of an event handler "
+                        "within the flush?" % (instance_key,)
+                    )
+                state._orphaned_outside_of_session = False
+
+        statelib.InstanceState._commit_all_states(
+            ((state, state.dict) for state in states), self.identity_map
+        )
+
+        self._register_altered(states)
+
+        if pending_to_persistent is not None:
+            for state in states.intersection(self._new):
+                pending_to_persistent(self, state)
+
+        # remove from new last, might be the last strong ref
+        for state in set(states).intersection(self._new):
+            self._new.pop(state)
+
+    def _register_altered(self, states: Iterable[InstanceState[Any]]) -> None:
+        if self._transaction:
+            for state in states:
+                if state in self._new:
+                    self._transaction._new[state] = True
+                else:
+                    self._transaction._dirty[state] = True
+
+    def _remove_newly_deleted(
+        self, states: Iterable[InstanceState[Any]]
+    ) -> None:
+        persistent_to_deleted = self.dispatch.persistent_to_deleted or None
+        for state in states:
+            if self._transaction:
+                self._transaction._deleted[state] = True
+
+            if persistent_to_deleted is not None:
+                # get a strong reference before we pop out of
+                # self._deleted
+                obj = state.obj()  # noqa
+
+            self.identity_map.safe_discard(state)
+            self._deleted.pop(state, None)
+            state._deleted = True
+            # can't call state._detach() here, because this state
+            # is still in the transaction snapshot and needs to be
+            # tracked as part of that
+            if persistent_to_deleted is not None:
+                persistent_to_deleted(self, state)
+
+    def add(self, instance: object, _warn: bool = True) -> None:
+        """Place an object into this :class:`_orm.Session`.
+
+        Objects that are in the :term:`transient` state when passed to the
+        :meth:`_orm.Session.add` method will move to the
+        :term:`pending` state, until the next flush, at which point they
+        will move to the :term:`persistent` state.
+
+        Objects that are in the :term:`detached` state when passed to the
+        :meth:`_orm.Session.add` method will move to the :term:`persistent`
+        state directly.
+
+        If the transaction used by the :class:`_orm.Session` is rolled back,
+        objects which were transient when they were passed to
+        :meth:`_orm.Session.add` will be moved back to the
+        :term:`transient` state, and will no longer be present within this
+        :class:`_orm.Session`.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add_all`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+        """
+        if _warn and self._warn_on_events:
+            self._flush_warning("Session.add()")
+
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+
+        self._save_or_update_state(state)
+
+    def add_all(self, instances: Iterable[object]) -> None:
+        """Add the given collection of instances to this :class:`_orm.Session`.
+
+        See the documentation for :meth:`_orm.Session.add` for a general
+        behavioral description.
+
+        .. seealso::
+
+            :meth:`_orm.Session.add`
+
+            :ref:`session_adding` - at :ref:`session_basics`
+
+        """
+
+        if self._warn_on_events:
+            self._flush_warning("Session.add_all()")
+
+        for instance in instances:
+            self.add(instance, _warn=False)
+
+    def _save_or_update_state(self, state: InstanceState[Any]) -> None:
+        state._orphaned_outside_of_session = False
+        self._save_or_update_impl(state)
+
+        mapper = _state_mapper(state)
+        for o, m, st_, dct_ in mapper.cascade_iterator(
+            "save-update", state, halt_on=self._contains_state
+        ):
+            self._save_or_update_impl(st_)
+
+    def delete(self, instance: object) -> None:
+        """Mark an instance as deleted.
+
+        The object is assumed to be either :term:`persistent` or
+        :term:`detached` when passed; after the method is called, the
+        object will remain in the :term:`persistent` state until the next
+        flush proceeds.  During this time, the object will also be a member
+        of the :attr:`_orm.Session.deleted` collection.
+
+        When the next flush proceeds, the object will move to the
+        :term:`deleted` state, indicating a ``DELETE`` statement was emitted
+        for its row within the current transaction.   When the transaction
+        is successfully committed,
+        the deleted object is moved to the :term:`detached` state and is
+        no longer present within this :class:`_orm.Session`.
+
+        .. seealso::
+
+            :ref:`session_deleting` - at :ref:`session_basics`
+
+        """
+        if self._warn_on_events:
+            self._flush_warning("Session.delete()")
+
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+
+        self._delete_impl(state, instance, head=True)
+
+    def _delete_impl(
+        self, state: InstanceState[Any], obj: object, head: bool
+    ) -> None:
+        if state.key is None:
+            if head:
+                raise sa_exc.InvalidRequestError(
+                    "Instance '%s' is not persisted" % state_str(state)
+                )
+            else:
+                return
+
+        to_attach = self._before_attach(state, obj)
+
+        if state in self._deleted:
+            return
+
+        self.identity_map.add(state)
+
+        if to_attach:
+            self._after_attach(state, obj)
+
+        if head:
+            # grab the cascades before adding the item to the deleted list
+            # so that autoflush does not delete the item
+            # the strong reference to the instance itself is significant here
+            cascade_states = list(
+                state.manager.mapper.cascade_iterator("delete", state)
+            )
+        else:
+            cascade_states = None
+
+        self._deleted[state] = obj
+
+        if head:
+            if TYPE_CHECKING:
+                assert cascade_states is not None
+            for o, m, st_, dct_ in cascade_states:
+                self._delete_impl(st_, o, False)
+
+    def get(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> Optional[_O]:
+        """Return an instance based on the given primary key identifier,
+        or ``None`` if not found.
+
+        E.g.::
+
+            my_user = session.get(User, 5)
+
+            some_object = session.get(VersionedFoo, (5, 10))
+
+            some_object = session.get(VersionedFoo, {"id": 5, "version_id": 10})
+
+        .. versionadded:: 1.4 Added :meth:`_orm.Session.get`, which is moved
+           from the now legacy :meth:`_orm.Query.get` method.
+
+        :meth:`_orm.Session.get` is special in that it provides direct
+        access to the identity map of the :class:`.Session`.
+        If the given primary key identifier is present
+        in the local identity map, the object is returned
+        directly from this collection and no SQL is emitted,
+        unless the object has been marked fully expired.
+        If not present,
+        a SELECT is performed in order to locate the object.
+
+        :meth:`_orm.Session.get` also will perform a check if
+        the object is present in the identity map and
+        marked as expired - a SELECT
+        is emitted to refresh the object as well as to
+        ensure that the row is still present.
+        If not, :class:`~sqlalchemy.orm.exc.ObjectDeletedError` is raised.
+
+        :param entity: a mapped class or :class:`.Mapper` indicating the
+         type of entity to be loaded.
+
+        :param ident: A scalar, tuple, or dictionary representing the
+         primary key.  For a composite (e.g. multiple column) primary key,
+         a tuple or dictionary should be passed.
+
+         For a single-column primary key, the scalar calling form is typically
+         the most expedient.  If the primary key of a row is the value "5",
+         the call looks like::
+
+            my_object = session.get(SomeClass, 5)
+
+         The tuple form contains primary key values typically in
+         the order in which they correspond to the mapped
+         :class:`_schema.Table`
+         object's primary key columns, or if the
+         :paramref:`_orm.Mapper.primary_key` configuration parameter were
+         used, in
+         the order used for that parameter. For example, if the primary key
+         of a row is represented by the integer
+         digits "5, 10" the call would look like::
+
+             my_object = session.get(SomeClass, (5, 10))
+
+         The dictionary form should include as keys the mapped attribute names
+         corresponding to each element of the primary key.  If the mapped class
+         has the attributes ``id``, ``version_id`` as the attributes which
+         store the object's primary key value, the call would look like::
+
+            my_object = session.get(SomeClass, {"id": 5, "version_id": 10})
+
+        :param options: optional sequence of loader options which will be
+         applied to the query, if one is emitted.
+
+        :param populate_existing: causes the method to unconditionally emit
+         a SQL query and refresh the object with the newly loaded data,
+         regardless of whether or not the object is already present.
+
+        :param with_for_update: optional boolean ``True`` indicating FOR UPDATE
+          should be used, or may be a dictionary containing flags to
+          indicate a more specific set of FOR UPDATE flags for the SELECT;
+          flags should match the parameters of
+          :meth:`_query.Query.with_for_update`.
+          Supersedes the :paramref:`.Session.refresh.lockmode` parameter.
+
+        :param execution_options: optional dictionary of execution options,
+         which will be associated with the query execution if one is emitted.
+         This dictionary can provide a subset of the options that are
+         accepted by :meth:`_engine.Connection.execution_options`, and may
+         also provide additional options understood only in an ORM context.
+
+         .. versionadded:: 1.4.29
+
+         .. seealso::
+
+            :ref:`orm_queryguide_execution_options` - ORM-specific execution
+            options
+
+        :param bind_arguments: dictionary of additional arguments to determine
+         the bind.  May include "mapper", "bind", or other custom arguments.
+         Contents of this dictionary are passed to the
+         :meth:`.Session.get_bind` method.
+
+         .. versionadded: 2.0.0rc1
+
+        :return: The object instance, or ``None``.
+
+        """  # noqa: E501
+        return self._get_impl(
+            entity,
+            ident,
+            loading.load_on_pk_identity,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+
+    def get_one(
+        self,
+        entity: _EntityBindKey[_O],
+        ident: _PKIdentityArgument,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> _O:
+        """Return exactly one instance based on the given primary key
+        identifier, or raise an exception if not found.
+
+        Raises ``sqlalchemy.orm.exc.NoResultFound`` if the query
+        selects no rows.
+
+        For a detailed documentation of the arguments see the
+        method :meth:`.Session.get`.
+
+        .. versionadded:: 2.0.22
+
+        :return: The object instance.
+
+        .. seealso::
+
+            :meth:`.Session.get` - equivalent method that instead
+              returns ``None`` if no row was found with the provided primary
+              key
+
+        """
+
+        instance = self.get(
+            entity,
+            ident,
+            options=options,
+            populate_existing=populate_existing,
+            with_for_update=with_for_update,
+            identity_token=identity_token,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+
+        if instance is None:
+            raise sa_exc.NoResultFound(
+                "No row was found when one was required"
+            )
+
+        return instance
+
+    def _get_impl(
+        self,
+        entity: _EntityBindKey[_O],
+        primary_key_identity: _PKIdentityArgument,
+        db_load_fn: Callable[..., _O],
+        *,
+        options: Optional[Sequence[ExecutableOption]] = None,
+        populate_existing: bool = False,
+        with_for_update: ForUpdateParameter = None,
+        identity_token: Optional[Any] = None,
+        execution_options: OrmExecuteOptionsParameter = util.EMPTY_DICT,
+        bind_arguments: Optional[_BindArguments] = None,
+    ) -> Optional[_O]:
+        # convert composite types to individual args
+        if (
+            is_composite_class(primary_key_identity)
+            and type(primary_key_identity)
+            in descriptor_props._composite_getters
+        ):
+            getter = descriptor_props._composite_getters[
+                type(primary_key_identity)
+            ]
+            primary_key_identity = getter(primary_key_identity)
+
+        mapper: Optional[Mapper[_O]] = inspect(entity)
+
+        if mapper is None or not mapper.is_mapper:
+            raise sa_exc.ArgumentError(
+                "Expected mapped class or mapper, got: %r" % entity
+            )
+
+        is_dict = isinstance(primary_key_identity, dict)
+        if not is_dict:
+            primary_key_identity = util.to_list(
+                primary_key_identity, default=[None]
+            )
+
+        if len(primary_key_identity) != len(mapper.primary_key):
+            raise sa_exc.InvalidRequestError(
+                "Incorrect number of values in identifier to formulate "
+                "primary key for session.get(); primary key columns "
+                "are %s" % ",".join("'%s'" % c for c in mapper.primary_key)
+            )
+
+        if is_dict:
+            pk_synonyms = mapper._pk_synonyms
+
+            if pk_synonyms:
+                correct_keys = set(pk_synonyms).intersection(
+                    primary_key_identity
+                )
+
+                if correct_keys:
+                    primary_key_identity = dict(primary_key_identity)
+                    for k in correct_keys:
+                        primary_key_identity[pk_synonyms[k]] = (
+                            primary_key_identity[k]
+                        )
+
+            try:
+                primary_key_identity = list(
+                    primary_key_identity[prop.key]
+                    for prop in mapper._identity_key_props
+                )
+
+            except KeyError as err:
+                raise sa_exc.InvalidRequestError(
+                    "Incorrect names of values in identifier to formulate "
+                    "primary key for session.get(); primary key attribute "
+                    "names are %s (synonym names are also accepted)"
+                    % ",".join(
+                        "'%s'" % prop.key
+                        for prop in mapper._identity_key_props
+                    )
+                ) from err
+
+        if (
+            not populate_existing
+            and not mapper.always_refresh
+            and with_for_update is None
+        ):
+            instance = self._identity_lookup(
+                mapper,
+                primary_key_identity,
+                identity_token=identity_token,
+                execution_options=execution_options,
+                bind_arguments=bind_arguments,
+            )
+
+            if instance is not None:
+                # reject calls for id in identity map but class
+                # mismatch.
+                if not isinstance(instance, mapper.class_):
+                    return None
+                return instance
+
+            # TODO: this was being tested before, but this is not possible
+            assert instance is not LoaderCallableStatus.PASSIVE_CLASS_MISMATCH
+
+        # set_label_style() not strictly necessary, however this will ensure
+        # that tablename_colname style is used which at the moment is
+        # asserted in a lot of unit tests :)
+
+        load_options = context.QueryContext.default_load_options
+
+        if populate_existing:
+            load_options += {"_populate_existing": populate_existing}
+        statement = sql.select(mapper).set_label_style(
+            LABEL_STYLE_TABLENAME_PLUS_COL
+        )
+        if with_for_update is not None:
+            statement._for_update_arg = ForUpdateArg._from_argument(
+                with_for_update
+            )
+
+        if options:
+            statement = statement.options(*options)
+        return db_load_fn(
+            self,
+            statement,
+            primary_key_identity,
+            load_options=load_options,
+            identity_token=identity_token,
+            execution_options=execution_options,
+            bind_arguments=bind_arguments,
+        )
+
+    def merge(
+        self,
+        instance: _O,
+        *,
+        load: bool = True,
+        options: Optional[Sequence[ORMOption]] = None,
+    ) -> _O:
+        """Copy the state of a given instance into a corresponding instance
+        within this :class:`.Session`.
+
+        :meth:`.Session.merge` examines the primary key attributes of the
+        source instance, and attempts to reconcile it with an instance of the
+        same primary key in the session.   If not found locally, it attempts
+        to load the object from the database based on primary key, and if
+        none can be located, creates a new instance.  The state of each
+        attribute on the source instance is then copied to the target
+        instance.  The resulting target instance is then returned by the
+        method; the original source instance is left unmodified, and
+        un-associated with the :class:`.Session` if not already.
+
+        This operation cascades to associated instances if the association is
+        mapped with ``cascade="merge"``.
+
+        See :ref:`unitofwork_merging` for a detailed discussion of merging.
+
+        :param instance: Instance to be merged.
+        :param load: Boolean, when False, :meth:`.merge` switches into
+         a "high performance" mode which causes it to forego emitting history
+         events as well as all database access.  This flag is used for
+         cases such as transferring graphs of objects into a :class:`.Session`
+         from a second level cache, or to transfer just-loaded objects
+         into the :class:`.Session` owned by a worker thread or process
+         without re-querying the database.
+
+         The ``load=False`` use case adds the caveat that the given
+         object has to be in a "clean" state, that is, has no pending changes
+         to be flushed - even if the incoming object is detached from any
+         :class:`.Session`.   This is so that when
+         the merge operation populates local attributes and
+         cascades to related objects and
+         collections, the values can be "stamped" onto the
+         target object as is, without generating any history or attribute
+         events, and without the need to reconcile the incoming data with
+         any existing related objects or collections that might not
+         be loaded.  The resulting objects from ``load=False`` are always
+         produced as "clean", so it is only appropriate that the given objects
+         should be "clean" as well, else this suggests a mis-use of the
+         method.
+        :param options: optional sequence of loader options which will be
+         applied to the :meth:`_orm.Session.get` method when the merge
+         operation loads the existing version of the object from the database.
+
+         .. versionadded:: 1.4.24
+
+
+        .. seealso::
+
+            :func:`.make_transient_to_detached` - provides for an alternative
+            means of "merging" a single object into the :class:`.Session`
+
+        """
+
+        if self._warn_on_events:
+            self._flush_warning("Session.merge()")
+
+        _recursive: Dict[InstanceState[Any], object] = {}
+        _resolve_conflict_map: Dict[_IdentityKeyType[Any], object] = {}
+
+        if load:
+            # flush current contents if we expect to load data
+            self._autoflush()
+
+        object_mapper(instance)  # verify mapped
+        autoflush = self.autoflush
+        try:
+            self.autoflush = False
+            return self._merge(
+                attributes.instance_state(instance),
+                attributes.instance_dict(instance),
+                load=load,
+                options=options,
+                _recursive=_recursive,
+                _resolve_conflict_map=_resolve_conflict_map,
+            )
+        finally:
+            self.autoflush = autoflush
+
+    def _merge(
+        self,
+        state: InstanceState[_O],
+        state_dict: _InstanceDict,
+        *,
+        options: Optional[Sequence[ORMOption]] = None,
+        load: bool,
+        _recursive: Dict[Any, object],
+        _resolve_conflict_map: Dict[_IdentityKeyType[Any], object],
+    ) -> _O:
+        mapper: Mapper[_O] = _state_mapper(state)
+        if state in _recursive:
+            return cast(_O, _recursive[state])
+
+        new_instance = False
+        key = state.key
+
+        merged: Optional[_O]
+
+        if key is None:
+            if state in self._new:
+                util.warn(
+                    "Instance %s is already pending in this Session yet is "
+                    "being merged again; this is probably not what you want "
+                    "to do" % state_str(state)
+                )
+
+            if not load:
+                raise sa_exc.InvalidRequestError(
+                    "merge() with load=False option does not support "
+                    "objects transient (i.e. unpersisted) objects.  flush() "
+                    "all changes on mapped instances before merging with "
+                    "load=False."
+                )
+            key = mapper._identity_key_from_state(state)
+            key_is_persistent = LoaderCallableStatus.NEVER_SET not in key[
+                1
+            ] and (
+                not _none_set.intersection(key[1])
+                or (
+                    mapper.allow_partial_pks
+                    and not _none_set.issuperset(key[1])
+                )
+            )
+        else:
+            key_is_persistent = True
+
+        if key in self.identity_map:
+            try:
+                merged = self.identity_map[key]
+            except KeyError:
+                # object was GC'ed right as we checked for it
+                merged = None
+        else:
+            merged = None
+
+        if merged is None:
+            if key_is_persistent and key in _resolve_conflict_map:
+                merged = cast(_O, _resolve_conflict_map[key])
+
+            elif not load:
+                if state.modified:
+                    raise sa_exc.InvalidRequestError(
+                        "merge() with load=False option does not support "
+                        "objects marked as 'dirty'.  flush() all changes on "
+                        "mapped instances before merging with load=False."
+                    )
+                merged = mapper.class_manager.new_instance()
+                merged_state = attributes.instance_state(merged)
+                merged_state.key = key
+                self._update_impl(merged_state)
+                new_instance = True
+
+            elif key_is_persistent:
+                merged = self.get(
+                    mapper.class_,
+                    key[1],
+                    identity_token=key[2],
+                    options=options,
+                )
+
+        if merged is None:
+            merged = mapper.class_manager.new_instance()
+            merged_state = attributes.instance_state(merged)
+            merged_dict = attributes.instance_dict(merged)
+            new_instance = True
+            self._save_or_update_state(merged_state)
+        else:
+            merged_state = attributes.instance_state(merged)
+            merged_dict = attributes.instance_dict(merged)
+
+        _recursive[state] = merged
+        _resolve_conflict_map[key] = merged
+
+        # check that we didn't just pull the exact same
+        # state out.
+        if state is not merged_state:
+            # version check if applicable
+            if mapper.version_id_col is not None:
+                existing_version = mapper._get_state_attr_by_column(
+                    state,
+                    state_dict,
+                    mapper.version_id_col,
+                    passive=PassiveFlag.PASSIVE_NO_INITIALIZE,
+                )
+
+                merged_version = mapper._get_state_attr_by_column(
+                    merged_state,
+                    merged_dict,
+                    mapper.version_id_col,
+                    passive=PassiveFlag.PASSIVE_NO_INITIALIZE,
+                )
+
+                if (
+                    existing_version
+                    is not LoaderCallableStatus.PASSIVE_NO_RESULT
+                    and merged_version
+                    is not LoaderCallableStatus.PASSIVE_NO_RESULT
+                    and existing_version != merged_version
+                ):
+                    raise exc.StaleDataError(
+                        "Version id '%s' on merged state %s "
+                        "does not match existing version '%s'. "
+                        "Leave the version attribute unset when "
+                        "merging to update the most recent version."
+                        % (
+                            existing_version,
+                            state_str(merged_state),
+                            merged_version,
+                        )
+                    )
+
+            merged_state.load_path = state.load_path
+            merged_state.load_options = state.load_options
+
+            # since we are copying load_options, we need to copy
+            # the callables_ that would have been generated by those
+            # load_options.
+            # assumes that the callables we put in state.callables_
+            # are not instance-specific (which they should not be)
+            merged_state._copy_callables(state)
+
+            for prop in mapper.iterate_properties:
+                prop.merge(
+                    self,
+                    state,
+                    state_dict,
+                    merged_state,
+                    merged_dict,
+                    load,
+                    _recursive,
+                    _resolve_conflict_map,
+                )
+
+        if not load:
+            # remove any history
+            merged_state._commit_all(merged_dict, self.identity_map)
+            merged_state.manager.dispatch._sa_event_merge_wo_load(
+                merged_state, None
+            )
+
+        if new_instance:
+            merged_state.manager.dispatch.load(merged_state, None)
+
+        return merged
+
+    def _validate_persistent(self, state: InstanceState[Any]) -> None:
+        if not self.identity_map.contains_state(state):
+            raise sa_exc.InvalidRequestError(
+                "Instance '%s' is not persistent within this Session"
+                % state_str(state)
+            )
+
+    def _save_impl(self, state: InstanceState[Any]) -> None:
+        if state.key is not None:
+            raise sa_exc.InvalidRequestError(
+                "Object '%s' already has an identity - "
+                "it can't be registered as pending" % state_str(state)
+            )
+
+        obj = state.obj()
+        to_attach = self._before_attach(state, obj)
+        if state not in self._new:
+            self._new[state] = obj
+            state.insert_order = len(self._new)
+        if to_attach:
+            self._after_attach(state, obj)
+
+    def _update_impl(
+        self, state: InstanceState[Any], revert_deletion: bool = False
+    ) -> None:
+        if state.key is None:
+            raise sa_exc.InvalidRequestError(
+                "Instance '%s' is not persisted" % state_str(state)
+            )
+
+        if state._deleted:
+            if revert_deletion:
+                if not state._attached:
+                    return
+                del state._deleted
+            else:
+                raise sa_exc.InvalidRequestError(
+                    "Instance '%s' has been deleted.  "
+                    "Use the make_transient() "
+                    "function to send this object back "
+                    "to the transient state." % state_str(state)
+                )
+
+        obj = state.obj()
+
+        # check for late gc
+        if obj is None:
+            return
+
+        to_attach = self._before_attach(state, obj)
+
+        self._deleted.pop(state, None)
+        if revert_deletion:
+            self.identity_map.replace(state)
+        else:
+            self.identity_map.add(state)
+
+        if to_attach:
+            self._after_attach(state, obj)
+        elif revert_deletion:
+            self.dispatch.deleted_to_persistent(self, state)
+
+    def _save_or_update_impl(self, state: InstanceState[Any]) -> None:
+        if state.key is None:
+            self._save_impl(state)
+        else:
+            self._update_impl(state)
+
+    def enable_relationship_loading(self, obj: object) -> None:
+        """Associate an object with this :class:`.Session` for related
+        object loading.
+
+        .. warning::
+
+            :meth:`.enable_relationship_loading` exists to serve special
+            use cases and is not recommended for general use.
+
+        Accesses of attributes mapped with :func:`_orm.relationship`
+        will attempt to load a value from the database using this
+        :class:`.Session` as the source of connectivity.  The values
+        will be loaded based on foreign key and primary key values
+        present on this object - if not present, then those relationships
+        will be unavailable.
+
+        The object will be attached to this session, but will
+        **not** participate in any persistence operations; its state
+        for almost all purposes will remain either "transient" or
+        "detached", except for the case of relationship loading.
+
+        Also note that backrefs will often not work as expected.
+        Altering a relationship-bound attribute on the target object
+        may not fire off a backref event, if the effective value
+        is what was already loaded from a foreign-key-holding value.
+
+        The :meth:`.Session.enable_relationship_loading` method is
+        similar to the ``load_on_pending`` flag on :func:`_orm.relationship`.
+        Unlike that flag, :meth:`.Session.enable_relationship_loading` allows
+        an object to remain transient while still being able to load
+        related items.
+
+        To make a transient object associated with a :class:`.Session`
+        via :meth:`.Session.enable_relationship_loading` pending, add
+        it to the :class:`.Session` using :meth:`.Session.add` normally.
+        If the object instead represents an existing identity in the database,
+        it should be merged using :meth:`.Session.merge`.
+
+        :meth:`.Session.enable_relationship_loading` does not improve
+        behavior when the ORM is used normally - object references should be
+        constructed at the object level, not at the foreign key level, so
+        that they are present in an ordinary way before flush()
+        proceeds.  This method is not intended for general use.
+
+        .. seealso::
+
+            :paramref:`_orm.relationship.load_on_pending` - this flag
+            allows per-relationship loading of many-to-ones on items that
+            are pending.
+
+            :func:`.make_transient_to_detached` - allows for an object to
+            be added to a :class:`.Session` without SQL emitted, which then
+            will unexpire attributes on access.
+
+        """
+        try:
+            state = attributes.instance_state(obj)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(obj) from err
+
+        to_attach = self._before_attach(state, obj)
+        state._load_pending = True
+        if to_attach:
+            self._after_attach(state, obj)
+
+    def _before_attach(self, state: InstanceState[Any], obj: object) -> bool:
+        self._autobegin_t()
+
+        if state.session_id == self.hash_key:
+            return False
+
+        if state.session_id and state.session_id in _sessions:
+            raise sa_exc.InvalidRequestError(
+                "Object '%s' is already attached to session '%s' "
+                "(this is '%s')"
+                % (state_str(state), state.session_id, self.hash_key)
+            )
+
+        self.dispatch.before_attach(self, state)
+
+        return True
+
+    def _after_attach(self, state: InstanceState[Any], obj: object) -> None:
+        state.session_id = self.hash_key
+        if state.modified and state._strong_obj is None:
+            state._strong_obj = obj
+        self.dispatch.after_attach(self, state)
+
+        if state.key:
+            self.dispatch.detached_to_persistent(self, state)
+        else:
+            self.dispatch.transient_to_pending(self, state)
+
+    def __contains__(self, instance: object) -> bool:
+        """Return True if the instance is associated with this session.
+
+        The instance may be pending or persistent within the Session for a
+        result of True.
+
+        """
+        try:
+            state = attributes.instance_state(instance)
+        except exc.NO_STATE as err:
+            raise exc.UnmappedInstanceError(instance) from err
+        return self._contains_state(state)
+
+    def __iter__(self) -> Iterator[object]:
+        """Iterate over all pending or persistent instances within this
+        Session.
+
+        """
+        return iter(
+            list(self._new.values()) + list(self.identity_map.values())
+        )
+
+    def _contains_state(self, state: InstanceState[Any]) -> bool:
+        return state in self._new or self.identity_map.contains_state(state)
+
+    def flush(self, objects: Optional[Sequence[Any]] = None) -> None:
+        """Flush all the object changes to the database.
+
+        Writes out all pending object creations, deletions and modifications
+        to the database as INSERTs, DELETEs, UPDATEs, etc.  Operations are
+        automatically ordered by the Session's unit of work dependency
+        solver.
+
+        Database operations will be issued in the current transactional
+        context and do not affect the state of the transaction, unless an
+        error occurs, in which case the entire transaction is rolled back.
+        You may flush() as often as you like within a transaction to move
+        changes from Python to the database's transaction buffer.
+
+        :param objects: Optional; restricts the flush operation to operate
+          only on elements that are in the given collection.
+
+          This feature is for an extremely narrow set of use cases where
+          particular objects may need to be operated upon before the
+          full flush() occurs.  It is not intended for general use.
+
+        """
+
+        if self._flushing:
+            raise sa_exc.InvalidRequestError("Session is already flushing")
+
+        if self._is_clean():
+            return
+        try:
+            self._flushing = True
+            self._flush(objects)
+        finally:
+            self._flushing = False
+
+    def _flush_warning(self, method: Any) -> None:
+        util.warn(
+            "Usage of the '%s' operation is not currently supported "
+            "within the execution stage of the flush process. "
+            "Results may not be consistent.  Consider using alternative "
+            "event listeners or connection-level operations instead." % method
+        )
+
+    def _is_clean(self) -> bool:
+        return (
+            not self.identity_map.check_modified()
+            and not self._deleted
+            and not self._new
+        )
+
+    def _flush(self, objects: Optional[Sequence[object]] = None) -> None:
+        dirty = self._dirty_states
+        if not dirty and not self._deleted and not self._new:
+            self.identity_map._modified.clear()
+            return
+
+        flush_context = UOWTransaction(self)
+
+        if self.dispatch.before_flush:
+            self.dispatch.before_flush(self, flush_context, objects)
+            # re-establish "dirty states" in case the listeners
+            # added
+            dirty = self._dirty_states
+
+        deleted = set(self._deleted)
+        new = set(self._new)
+
+        dirty = set(dirty).difference(deleted)
+
+        # create the set of all objects we want to operate upon
+        if objects:
+            # specific list passed in
+            objset = set()
+            for o in objects:
+                try:
+                    state = attributes.instance_state(o)
+
+                except exc.NO_STATE as err:
+                    raise exc.UnmappedInstanceError(o) from err
+                objset.add(state)
+        else:
+            objset = None
+
+        # store objects whose fate has been decided
+        processed = set()
+
+        # put all saves/updates into the flush context.  detect top-level
+        # orphans and throw them into deleted.
+        if objset:
+            proc = new.union(dirty).intersection(objset).difference(deleted)
+        else:
+            proc = new.union(dirty).difference(deleted)
+
+        for state in proc:
+            is_orphan = _state_mapper(state)._is_orphan(state)
+
+            is_persistent_orphan = is_orphan and state.has_identity
+
+            if (
+                is_orphan
+                and not is_persistent_orphan
+                and state._orphaned_outside_of_session
+            ):
+                self._expunge_states([state])
+            else:
+                _reg = flush_context.register_object(
+                    state, isdelete=is_persistent_orphan
+                )
+                assert _reg, "Failed to add object to the flush context!"
+                processed.add(state)
+
+        # put all remaining deletes into the flush context.
+        if objset:
+            proc = deleted.intersection(objset).difference(processed)
+        else:
+            proc = deleted.difference(processed)
+        for state in proc:
+            _reg = flush_context.register_object(state, isdelete=True)
+            assert _reg, "Failed to add object to the flush context!"
+
+        if not flush_context.has_work:
+            return
+
+        flush_context.transaction = transaction = self._autobegin_t()._begin()
+        try:
+            self._warn_on_events = True
+            try:
+                flush_context.execute()
+            finally:
+                self._warn_on_events = False
+
+            self.dispatch.after_flush(self, flush_context)
+
+            flush_context.finalize_flush_changes()
+
+            if not objects and self.identity_map._modified:
+                len_ = len(self.identity_map._modified)
+
+                statelib.InstanceState._commit_all_states(
+                    [
+                        (state, state.dict)
+                        for state in self.identity_map._modified
+                    ],
+                    instance_dict=self.identity_map,
+                )
+                util.warn(
+                    "Attribute history events accumulated on %d "
+                    "previously clean instances "
+                    "within inner-flush event handlers have been "
+                    "reset, and will not result in database updates. "
+                    "Consider using set_committed_value() within "
+                    "inner-flush event handlers to avoid this warning." % len_
+                )
+
+            # useful assertions:
+            # if not objects:
+            #    assert not self.identity_map._modified
+            # else:
+            #    assert self.identity_map._modified == \
+            #            self.identity_map._modified.difference(objects)
+
+            self.dispatch.after_flush_postexec(self, flush_context)
+
+            transaction.commit()
+
+        except:
+            with util.safe_reraise():
+                transaction.rollback(_capture_exception=True)
+
+    def bulk_save_objects(
+        self,
+        objects: Iterable[object],
+        return_defaults: bool = False,
+        update_changed_only: bool = True,
+        preserve_order: bool = True,
+    ) -> None:
+        """Perform a bulk save of the given list of objects.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.
+
+            For general INSERT and UPDATE of existing ORM mapped objects,
+            prefer standard :term:`unit of work` data management patterns,
+            introduced in the :ref:`unified_tutorial` at
+            :ref:`tutorial_orm_data_manipulation`.  SQLAlchemy 2.0
+            now uses :ref:`engine_insertmanyvalues` with modern dialects
+            which solves previous issues of bulk INSERT slowness.
+
+        :param objects: a sequence of mapped object instances.  The mapped
+         objects are persisted as is, and are **not** associated with the
+         :class:`.Session` afterwards.
+
+         For each object, whether the object is sent as an INSERT or an
+         UPDATE is dependent on the same rules used by the :class:`.Session`
+         in traditional operation; if the object has the
+         :attr:`.InstanceState.key`
+         attribute set, then the object is assumed to be "detached" and
+         will result in an UPDATE.  Otherwise, an INSERT is used.
+
+         In the case of an UPDATE, statements are grouped based on which
+         attributes have changed, and are thus to be the subject of each
+         SET clause.  If ``update_changed_only`` is False, then all
+         attributes present within each object are applied to the UPDATE
+         statement, which may help in allowing the statements to be grouped
+         together into a larger executemany(), and will also reduce the
+         overhead of checking history on attributes.
+
+        :param return_defaults: when True, rows that are missing values which
+         generate defaults, namely integer primary key defaults and sequences,
+         will be inserted **one at a time**, so that the primary key value
+         is available.  In particular this will allow joined-inheritance
+         and other multi-table mappings to insert correctly without the need
+         to provide primary key values ahead of time; however,
+         :paramref:`.Session.bulk_save_objects.return_defaults` **greatly
+         reduces the performance gains** of the method overall.  It is strongly
+         advised to please use the standard :meth:`_orm.Session.add_all`
+         approach.
+
+        :param update_changed_only: when True, UPDATE statements are rendered
+         based on those attributes in each state that have logged changes.
+         When False, all attributes present are rendered into the SET clause
+         with the exception of primary key attributes.
+
+        :param preserve_order: when True, the order of inserts and updates
+         matches exactly the order in which the objects are given.   When
+         False, common types of objects are grouped into inserts
+         and updates, to allow for more batching opportunities.
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_insert_mappings`
+
+            :meth:`.Session.bulk_update_mappings`
+
+        """
+
+        obj_states: Iterable[InstanceState[Any]]
+
+        obj_states = (attributes.instance_state(obj) for obj in objects)
+
+        if not preserve_order:
+            # the purpose of this sort is just so that common mappers
+            # and persistence states are grouped together, so that groupby
+            # will return a single group for a particular type of mapper.
+            # it's not trying to be deterministic beyond that.
+            obj_states = sorted(
+                obj_states,
+                key=lambda state: (id(state.mapper), state.key is not None),
+            )
+
+        def grouping_key(
+            state: InstanceState[_O],
+        ) -> Tuple[Mapper[_O], bool]:
+            return (state.mapper, state.key is not None)
+
+        for (mapper, isupdate), states in itertools.groupby(
+            obj_states, grouping_key
+        ):
+            self._bulk_save_mappings(
+                mapper,
+                states,
+                isupdate=isupdate,
+                isstates=True,
+                return_defaults=return_defaults,
+                update_changed_only=update_changed_only,
+                render_nulls=False,
+            )
+
+    def bulk_insert_mappings(
+        self,
+        mapper: Mapper[Any],
+        mappings: Iterable[Dict[str, Any]],
+        return_defaults: bool = False,
+        render_nulls: bool = False,
+    ) -> None:
+        """Perform a bulk insert of the given list of mapping dictionaries.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.  The 2.0 API shares
+            implementation details with this method and adds new features
+            as well.
+
+        :param mapper: a mapped class, or the actual :class:`_orm.Mapper`
+         object,
+         representing the single kind of object represented within the mapping
+         list.
+
+        :param mappings: a sequence of dictionaries, each one containing the
+         state of the mapped row to be inserted, in terms of the attribute
+         names on the mapped class.   If the mapping refers to multiple tables,
+         such as a joined-inheritance mapping, each dictionary must contain all
+         keys to be populated into all tables.
+
+        :param return_defaults: when True, the INSERT process will be altered
+         to ensure that newly generated primary key values will be fetched.
+         The rationale for this parameter is typically to enable
+         :ref:`Joined Table Inheritance <joined_inheritance>` mappings to
+         be bulk inserted.
+
+         .. note:: for backends that don't support RETURNING, the
+            :paramref:`_orm.Session.bulk_insert_mappings.return_defaults`
+            parameter can significantly decrease performance as INSERT
+            statements can no longer be batched.   See
+            :ref:`engine_insertmanyvalues`
+            for background on which backends are affected.
+
+        :param render_nulls: When True, a value of ``None`` will result
+         in a NULL value being included in the INSERT statement, rather
+         than the column being omitted from the INSERT.   This allows all
+         the rows being INSERTed to have the identical set of columns which
+         allows the full set of rows to be batched to the DBAPI.  Normally,
+         each column-set that contains a different combination of NULL values
+         than the previous row must omit a different series of columns from
+         the rendered INSERT statement, which means it must be emitted as a
+         separate statement.   By passing this flag, the full set of rows
+         are guaranteed to be batchable into one batch; the cost however is
+         that server-side defaults which are invoked by an omitted column will
+         be skipped, so care must be taken to ensure that these are not
+         necessary.
+
+         .. warning::
+
+            When this flag is set, **server side default SQL values will
+            not be invoked** for those columns that are inserted as NULL;
+            the NULL value will be sent explicitly.   Care must be taken
+            to ensure that no server-side default functions need to be
+            invoked for the operation as a whole.
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_save_objects`
+
+            :meth:`.Session.bulk_update_mappings`
+
+        """
+        self._bulk_save_mappings(
+            mapper,
+            mappings,
+            isupdate=False,
+            isstates=False,
+            return_defaults=return_defaults,
+            update_changed_only=False,
+            render_nulls=render_nulls,
+        )
+
+    def bulk_update_mappings(
+        self, mapper: Mapper[Any], mappings: Iterable[Dict[str, Any]]
+    ) -> None:
+        """Perform a bulk update of the given list of mapping dictionaries.
+
+        .. legacy::
+
+            This method is a legacy feature as of the 2.0 series of
+            SQLAlchemy.   For modern bulk INSERT and UPDATE, see
+            the sections :ref:`orm_queryguide_bulk_insert` and
+            :ref:`orm_queryguide_bulk_update`.  The 2.0 API shares
+            implementation details with this method and adds new features
+            as well.
+
+        :param mapper: a mapped class, or the actual :class:`_orm.Mapper`
+         object,
+         representing the single kind of object represented within the mapping
+         list.
+
+        :param mappings: a sequence of dictionaries, each one containing the
+         state of the mapped row to be updated, in terms of the attribute names
+         on the mapped class.   If the mapping refers to multiple tables, such
+         as a joined-inheritance mapping, each dictionary may contain keys
+         corresponding to all tables.   All those keys which are present and
+         are not part of the primary key are applied to the SET clause of the
+         UPDATE statement; the primary key values, which are required, are
+         applied to the WHERE clause.
+
+
+        .. seealso::
+
+            :doc:`queryguide/dml`
+
+            :meth:`.Session.bulk_insert_mappings`
+
+            :meth:`.Session.bulk_save_objects`
+
+        """
+        self._bulk_save_mappings(
+            mapper,
+            mappings,
+            isupdate=True,
+            isstates=False,
+            return_defaults=False,
+            update_changed_only=False,
+            render_nulls=False,
+        )
+
+    def _bulk_save_mappings(
+        self,
+        mapper: Mapper[_O],
+        mappings: Union[Iterable[InstanceState[_O]], Iterable[Dict[str, Any]]],
+        *,
+        isupdate: bool,
+        isstates: bool,
+        return_defaults: bool,
+        update_changed_only: bool,
+        render_nulls: bool,
+    ) -> None:
+        mapper = _class_to_mapper(mapper)
+        self._flushing = True
+
+        transaction = self._autobegin_t()._begin()
+        try:
+            if isupdate:
+                bulk_persistence._bulk_update(
+                    mapper,
+                    mappings,
+                    transaction,
+                    isstates=isstates,
+                    update_changed_only=update_changed_only,
+                )
+            else:
+                bulk_persistence._bulk_insert(
+                    mapper,
+                    mappings,
+                    transaction,
+                    isstates=isstates,
+                    return_defaults=return_defaults,
+                    render_nulls=render_nulls,
+                )
+            transaction.commit()
+
+        except:
+            with util.safe_reraise():
+                transaction.rollback(_capture_exception=True)
+        finally:
+            self._flushing = False
+
+    def is_modified(
+        self, instance: object, include_collections: bool = True
+    ) -> bool:
+        r"""Return ``True`` if the given instance has locally
+        modified attributes.
+
+        This method retrieves the history for each instrumented
+        attribute on the instance and performs a comparison of the current
+        value to its previously flushed or committed value, if any.
+
+        It is in effect a more expensive and accurate
+        version of checking for the given instance in the
+        :attr:`.Session.dirty` collection; a full test for
+        each attribute's net "dirty" status is performed.
+
+        E.g.::
+
+            return session.is_modified(someobject)
+
+        A few caveats to this method apply:
+
+        * Instances present in the :attr:`.Session.dirty` collection may
+          report ``False`` when tested with this method.  This is because
+          the object may have received change events via attribute mutation,
+          thus placing it in :attr:`.Session.dirty`, but ultimately the state
+          is the same as that loaded from the database, resulting in no net
+          change here.
+        * Scalar attributes may not have recorded the previously set
+          value when a new value was applied, if the attribute was not loaded,
+          or was expired, at the time the new value was received - in these
+          cases, the attribute is assumed to have a change, even if there is
+          ultimately no net change against its database value. SQLAlchemy in
+          most cases does not need the "old" value when a set event occurs, so
+          it skips the expense of a SQL call if the old value isn't present,
+          based on the assumption that an UPDATE of the scalar value is
+          usually needed, and in those few cases where it isn't, is less
+          expensive on average than issuing a defensive SELECT.
+
+          The "old" value is fetched unconditionally upon set only if the
+          attribute container has the ``active_history`` flag set to ``True``.
+          This flag is set typically for primary key attributes and scalar
+          object references that are not a simple many-to-one.  To set this
+          flag for any arbitrary mapped column, use the ``active_history``
+          argument with :func:`.column_property`.
+
+        :param instance: mapped instance to be tested for pending changes.
+        :param include_collections: Indicates if multivalued collections
+         should be included in the operation.  Setting this to ``False`` is a
+         way to detect only local-column based properties (i.e. scalar columns
+         or many-to-one foreign keys) that would result in an UPDATE for this
+         instance upon flush.
+
+        """
+        state = object_state(instance)
+
+        if not state.modified:
+            return False
+
+        dict_ = state.dict
+
+        for attr in state.manager.attributes:
+            if (
+                not include_collections
+                and hasattr(attr.impl, "get_collection")
+            ) or not hasattr(attr.impl, "get_history"):
+                continue
+
+            (added, unchanged, deleted) = attr.impl.get_history(
+                state, dict_, passive=PassiveFlag.NO_CHANGE
+            )
+
+            if added or deleted:
+                return True
+        else:
+            return False
+
+    @property
+    def is_active(self) -> bool:
+        """True if this :class:`.Session` not in "partial rollback" state.
+
+        .. versionchanged:: 1.4 The :class:`_orm.Session` no longer begins
+           a new transaction immediately, so this attribute will be False
+           when the :class:`_orm.Session` is first instantiated.
+
+        "partial rollback" state typically indicates that the flush process
+        of the :class:`_orm.Session` has failed, and that the
+        :meth:`_orm.Session.rollback` method must be emitted in order to
+        fully roll back the transaction.
+
+        If this :class:`_orm.Session` is not in a transaction at all, the
+        :class:`_orm.Session` will autobegin when it is first used, so in this
+        case :attr:`_orm.Session.is_active` will return True.
+
+        Otherwise, if this :class:`_orm.Session` is within a transaction,
+        and that transaction has not been rolled back internally, the
+        :attr:`_orm.Session.is_active` will also return True.
+
+        .. seealso::
+
+            :ref:`faq_session_rollback`
+
+            :meth:`_orm.Session.in_transaction`
+
+        """
+        return self._transaction is None or self._transaction.is_active
+
+    @property
+    def _dirty_states(self) -> Iterable[InstanceState[Any]]:
+        """The set of all persistent states considered dirty.
+
+        This method returns all states that were modified including
+        those that were possibly deleted.
+
+        """
+        return self.identity_map._dirty_states()
+
+    @property
+    def dirty(self) -> IdentitySet:
+        """The set of all persistent instances considered dirty.
+
+        E.g.::
+
+            some_mapped_object in session.dirty
+
+        Instances are considered dirty when they were modified but not
+        deleted.
+
+        Note that this 'dirty' calculation is 'optimistic'; most
+        attribute-setting or collection modification operations will
+        mark an instance as 'dirty' and place it in this set, even if
+        there is no net change to the attribute's value.  At flush
+        time, the value of each attribute is compared to its
+        previously saved value, and if there's no net change, no SQL
+        operation will occur (this is a more expensive operation so
+        it's only done at flush time).
+
+        To check if an instance has actionable net changes to its
+        attributes, use the :meth:`.Session.is_modified` method.
+
+        """
+        return IdentitySet(
+            [
+                state.obj()
+                for state in self._dirty_states
+                if state not in self._deleted
+            ]
+        )
+
+    @property
+    def deleted(self) -> IdentitySet:
+        "The set of all instances marked as 'deleted' within this ``Session``"
+
+        return util.IdentitySet(list(self._deleted.values()))
+
+    @property
+    def new(self) -> IdentitySet:
+        "The set of all instances marked as 'new' within this ``Session``."
+
+        return util.IdentitySet(list(self._new.values()))
+
+
+_S = TypeVar("_S", bound="Session")
+
+
+class sessionmaker(_SessionClassMethods, Generic[_S]):
+    """A configurable :class:`.Session` factory.
+
+    The :class:`.sessionmaker` factory generates new
+    :class:`.Session` objects when called, creating them given
+    the configurational arguments established here.
+
+    e.g.::
+
+        from sqlalchemy import create_engine
+        from sqlalchemy.orm import sessionmaker
+
+        # an Engine, which the Session will use for connection
+        # resources
+        engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/")
+
+        Session = sessionmaker(engine)
+
+        with Session() as session:
+            session.add(some_object)
+            session.add(some_other_object)
+            session.commit()
+
+    Context manager use is optional; otherwise, the returned
+    :class:`_orm.Session` object may be closed explicitly via the
+    :meth:`_orm.Session.close` method.   Using a
+    ``try:/finally:`` block is optional, however will ensure that the close
+    takes place even if there are database errors::
+
+        session = Session()
+        try:
+            session.add(some_object)
+            session.add(some_other_object)
+            session.commit()
+        finally:
+            session.close()
+
+    :class:`.sessionmaker` acts as a factory for :class:`_orm.Session`
+    objects in the same way as an :class:`_engine.Engine` acts as a factory
+    for :class:`_engine.Connection` objects.  In this way it also includes
+    a :meth:`_orm.sessionmaker.begin` method, that provides a context
+    manager which both begins and commits a transaction, as well as closes
+    out the :class:`_orm.Session` when complete, rolling back the transaction
+    if any errors occur::
+
+        Session = sessionmaker(engine)
+
+        with Session.begin() as session:
+            session.add(some_object)
+            session.add(some_other_object)
+        # commits transaction, closes session
+
+    .. versionadded:: 1.4
+
+    When calling upon :class:`_orm.sessionmaker` to construct a
+    :class:`_orm.Session`, keyword arguments may also be passed to the
+    method; these arguments will override that of the globally configured
+    parameters.  Below we use a :class:`_orm.sessionmaker` bound to a certain
+    :class:`_engine.Engine` to produce a :class:`_orm.Session` that is instead
+    bound to a specific :class:`_engine.Connection` procured from that engine::
+
+        Session = sessionmaker(engine)
+
+        # bind an individual session to a connection
+
+        with engine.connect() as connection:
+            with Session(bind=connection) as session:
+                ...  # work with session
+
+    The class also includes a method :meth:`_orm.sessionmaker.configure`, which
+    can be used to specify additional keyword arguments to the factory, which
+    will take effect for subsequent :class:`.Session` objects generated. This
+    is usually used to associate one or more :class:`_engine.Engine` objects
+    with an existing
+    :class:`.sessionmaker` factory before it is first used::
+
+        # application starts, sessionmaker does not have
+        # an engine bound yet
+        Session = sessionmaker()
+
+        # ... later, when an engine URL is read from a configuration
+        # file or other events allow the engine to be created
+        engine = create_engine("sqlite:///foo.db")
+        Session.configure(bind=engine)
+
+        sess = Session()
+        # work with session
+
+    .. seealso::
+
+        :ref:`session_getting` - introductory text on creating
+        sessions using :class:`.sessionmaker`.
+
+    """
+
+    class_: Type[_S]
+
+    @overload
+    def __init__(
+        self,
+        bind: Optional[_SessionBind] = ...,
+        *,
+        class_: Type[_S],
+        autoflush: bool = ...,
+        expire_on_commit: bool = ...,
+        info: Optional[_InfoType] = ...,
+        **kw: Any,
+    ): ...
+
+    @overload
+    def __init__(
+        self: "sessionmaker[Session]",
+        bind: Optional[_SessionBind] = ...,
+        *,
+        autoflush: bool = ...,
+        expire_on_commit: bool = ...,
+        info: Optional[_InfoType] = ...,
+        **kw: Any,
+    ): ...
+
+    def __init__(
+        self,
+        bind: Optional[_SessionBind] = None,
+        *,
+        class_: Type[_S] = Session,  # type: ignore
+        autoflush: bool = True,
+        expire_on_commit: bool = True,
+        info: Optional[_InfoType] = None,
+        **kw: Any,
+    ):
+        r"""Construct a new :class:`.sessionmaker`.
+
+        All arguments here except for ``class_`` correspond to arguments
+        accepted by :class:`.Session` directly.  See the
+        :meth:`.Session.__init__` docstring for more details on parameters.
+
+        :param bind: a :class:`_engine.Engine` or other :class:`.Connectable`
+         with
+         which newly created :class:`.Session` objects will be associated.
+        :param class\_: class to use in order to create new :class:`.Session`
+         objects.  Defaults to :class:`.Session`.
+        :param autoflush: The autoflush setting to use with newly created
+         :class:`.Session` objects.
+
+         .. seealso::
+
+            :ref:`session_flushing` - additional background on autoflush
+
+        :param expire_on_commit=True: the
+         :paramref:`_orm.Session.expire_on_commit` setting to use
+         with newly created :class:`.Session` objects.
+
+        :param info: optional dictionary of information that will be available
+         via :attr:`.Session.info`.  Note this dictionary is *updated*, not
+         replaced, when the ``info`` parameter is specified to the specific
+         :class:`.Session` construction operation.
+
+        :param \**kw: all other keyword arguments are passed to the
+         constructor of newly created :class:`.Session` objects.
+
+        """
+        kw["bind"] = bind
+        kw["autoflush"] = autoflush
+        kw["expire_on_commit"] = expire_on_commit
+        if info is not None:
+            kw["info"] = info
+        self.kw = kw
+        # make our own subclass of the given class, so that
+        # events can be associated with it specifically.
+        self.class_ = type(class_.__name__, (class_,), {})
+
+    def begin(self) -> contextlib.AbstractContextManager[_S]:
+        """Produce a context manager that both provides a new
+        :class:`_orm.Session` as well as a transaction that commits.
+
+
+        e.g.::
+
+            Session = sessionmaker(some_engine)
+
+            with Session.begin() as session:
+                session.add(some_object)
+
+            # commits transaction, closes session
+
+        .. versionadded:: 1.4
+
+
+        """
+
+        session = self()
+        return session._maker_context_manager()
+
+    def __call__(self, **local_kw: Any) -> _S:
+        """Produce a new :class:`.Session` object using the configuration
+        established in this :class:`.sessionmaker`.
+
+        In Python, the ``__call__`` method is invoked on an object when
+        it is "called" in the same way as a function::
+
+            Session = sessionmaker(some_engine)
+            session = Session()  # invokes sessionmaker.__call__()
+
+        """
+        for k, v in self.kw.items():
+            if k == "info" and "info" in local_kw:
+                d = v.copy()
+                d.update(local_kw["info"])
+                local_kw["info"] = d
+            else:
+                local_kw.setdefault(k, v)
+        return self.class_(**local_kw)
+
+    def configure(self, **new_kw: Any) -> None:
+        """(Re)configure the arguments for this sessionmaker.
+
+        e.g.::
+
+            Session = sessionmaker()
+
+            Session.configure(bind=create_engine("sqlite://"))
+        """
+        self.kw.update(new_kw)
+
+    def __repr__(self) -> str:
+        return "%s(class_=%r, %s)" % (
+            self.__class__.__name__,
+            self.class_.__name__,
+            ", ".join("%s=%r" % (k, v) for k, v in self.kw.items()),
+        )
+
+
+def close_all_sessions() -> None:
+    """Close all sessions in memory.
+
+    This function consults a global registry of all :class:`.Session` objects
+    and calls :meth:`.Session.close` on them, which resets them to a clean
+    state.
+
+    This function is not for general use but may be useful for test suites
+    within the teardown scheme.
+
+    .. versionadded:: 1.3
+
+    """
+
+    for sess in _sessions.values():
+        sess.close()
+
+
+def make_transient(instance: object) -> None:
+    """Alter the state of the given instance so that it is :term:`transient`.
+
+    .. note::
+
+        :func:`.make_transient` is a special-case function for
+        advanced use cases only.
+
+    The given mapped instance is assumed to be in the :term:`persistent` or
+    :term:`detached` state.   The function will remove its association with any
+    :class:`.Session` as well as its :attr:`.InstanceState.identity`. The
+    effect is that the object will behave as though it were newly constructed,
+    except retaining any attribute / collection values that were loaded at the
+    time of the call.   The :attr:`.InstanceState.deleted` flag is also reset
+    if this object had been deleted as a result of using
+    :meth:`.Session.delete`.
+
+    .. warning::
+
+        :func:`.make_transient` does **not** "unexpire" or otherwise eagerly
+        load ORM-mapped attributes that are not currently loaded at the time
+        the function is called.   This includes attributes which:
+
+        * were expired via :meth:`.Session.expire`
+
+        * were expired as the natural effect of committing a session
+          transaction, e.g. :meth:`.Session.commit`
+
+        * are normally :term:`lazy loaded` but are not currently loaded
+
+        * are "deferred" (see :ref:`orm_queryguide_column_deferral`) and are
+          not yet loaded
+
+        * were not present in the query which loaded this object, such as that
+          which is common in joined table inheritance and other scenarios.
+
+        After :func:`.make_transient` is called, unloaded attributes such
+        as those above will normally resolve to the value ``None`` when
+        accessed, or an empty collection for a collection-oriented attribute.
+        As the object is transient and un-associated with any database
+        identity, it will no longer retrieve these values.
+
+    .. seealso::
+
+        :func:`.make_transient_to_detached`
+
+    """
+    state = attributes.instance_state(instance)
+    s = _state_session(state)
+    if s:
+        s._expunge_states([state])
+
+    # remove expired state
+    state.expired_attributes.clear()
+
+    # remove deferred callables
+    if state.callables:
+        del state.callables
+
+    if state.key:
+        del state.key
+    if state._deleted:
+        del state._deleted
+
+
+def make_transient_to_detached(instance: object) -> None:
+    """Make the given transient instance :term:`detached`.
+
+    .. note::
+
+        :func:`.make_transient_to_detached` is a special-case function for
+        advanced use cases only.
+
+    All attribute history on the given instance
+    will be reset as though the instance were freshly loaded
+    from a query.  Missing attributes will be marked as expired.
+    The primary key attributes of the object, which are required, will be made
+    into the "key" of the instance.
+
+    The object can then be added to a session, or merged
+    possibly with the load=False flag, at which point it will look
+    as if it were loaded that way, without emitting SQL.
+
+    This is a special use case function that differs from a normal
+    call to :meth:`.Session.merge` in that a given persistent state
+    can be manufactured without any SQL calls.
+
+    .. seealso::
+
+        :func:`.make_transient`
+
+        :meth:`.Session.enable_relationship_loading`
+
+    """
+    state = attributes.instance_state(instance)
+    if state.session_id or state.key:
+        raise sa_exc.InvalidRequestError("Given object must be transient")
+    state.key = state.mapper._identity_key_from_state(state)
+    if state._deleted:
+        del state._deleted
+    state._commit_all(state.dict)
+    state._expire_attributes(state.dict, state.unloaded)
+
+
+def object_session(instance: object) -> Optional[Session]:
+    """Return the :class:`.Session` to which the given instance belongs.
+
+    This is essentially the same as the :attr:`.InstanceState.session`
+    accessor.  See that attribute for details.
+
+    """
+
+    try:
+        state = attributes.instance_state(instance)
+    except exc.NO_STATE as err:
+        raise exc.UnmappedInstanceError(instance) from err
+    else:
+        return _state_session(state)
+
+
+_new_sessionid = util.counter()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state.py
new file mode 100644
index 00000000..d4bbf920
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state.py
@@ -0,0 +1,1143 @@
+# orm/state.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Defines instrumentation of instances.
+
+This module is usually not directly visible to user applications, but
+defines a large part of the ORM's interactivity.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+import weakref
+
+from . import base
+from . import exc as orm_exc
+from . import interfaces
+from ._typing import _O
+from ._typing import is_collection_impl
+from .base import ATTR_WAS_SET
+from .base import INIT_OK
+from .base import LoaderCallableStatus
+from .base import NEVER_SET
+from .base import NO_VALUE
+from .base import PASSIVE_NO_INITIALIZE
+from .base import PASSIVE_NO_RESULT
+from .base import PASSIVE_OFF
+from .base import SQL_OK
+from .path_registry import PathRegistry
+from .. import exc as sa_exc
+from .. import inspection
+from .. import util
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if TYPE_CHECKING:
+    from ._typing import _IdentityKeyType
+    from ._typing import _InstanceDict
+    from ._typing import _LoaderCallable
+    from .attributes import AttributeImpl
+    from .attributes import History
+    from .base import PassiveFlag
+    from .collections import _AdaptedCollectionProtocol
+    from .identity import IdentityMap
+    from .instrumentation import ClassManager
+    from .interfaces import ORMOption
+    from .mapper import Mapper
+    from .session import Session
+    from ..engine import Row
+    from ..ext.asyncio.session import async_session as _async_provider
+    from ..ext.asyncio.session import AsyncSession
+
+if TYPE_CHECKING:
+    _sessions: weakref.WeakValueDictionary[int, Session]
+else:
+    # late-populated by session.py
+    _sessions = None
+
+
+if not TYPE_CHECKING:
+    # optionally late-provided by sqlalchemy.ext.asyncio.session
+
+    _async_provider = None  # noqa
+
+
+class _InstanceDictProto(Protocol):
+    def __call__(self) -> Optional[IdentityMap]: ...
+
+
+class _InstallLoaderCallableProto(Protocol[_O]):
+    """used at result loading time to install a _LoaderCallable callable
+    upon a specific InstanceState, which will be used to populate an
+    attribute when that attribute is accessed.
+
+    Concrete examples are per-instance deferred column loaders and
+    relationship lazy loaders.
+
+    """
+
+    def __call__(
+        self, state: InstanceState[_O], dict_: _InstanceDict, row: Row[Any]
+    ) -> None: ...
+
+
+@inspection._self_inspects
+class InstanceState(interfaces.InspectionAttrInfo, Generic[_O]):
+    """Tracks state information at the instance level.
+
+    The :class:`.InstanceState` is a key object used by the
+    SQLAlchemy ORM in order to track the state of an object;
+    it is created the moment an object is instantiated, typically
+    as a result of :term:`instrumentation` which SQLAlchemy applies
+    to the ``__init__()`` method of the class.
+
+    :class:`.InstanceState` is also a semi-public object,
+    available for runtime inspection as to the state of a
+    mapped instance, including information such as its current
+    status within a particular :class:`.Session` and details
+    about data on individual attributes.  The public API
+    in order to acquire a :class:`.InstanceState` object
+    is to use the :func:`_sa.inspect` system::
+
+        >>> from sqlalchemy import inspect
+        >>> insp = inspect(some_mapped_object)
+        >>> insp.attrs.nickname.history
+        History(added=['new nickname'], unchanged=(), deleted=['nickname'])
+
+    .. seealso::
+
+        :ref:`orm_mapper_inspection_instancestate`
+
+    """
+
+    __slots__ = (
+        "__dict__",
+        "__weakref__",
+        "class_",
+        "manager",
+        "obj",
+        "committed_state",
+        "expired_attributes",
+    )
+
+    manager: ClassManager[_O]
+    session_id: Optional[int] = None
+    key: Optional[_IdentityKeyType[_O]] = None
+    runid: Optional[int] = None
+    load_options: Tuple[ORMOption, ...] = ()
+    load_path: PathRegistry = PathRegistry.root
+    insert_order: Optional[int] = None
+    _strong_obj: Optional[object] = None
+    obj: weakref.ref[_O]
+
+    committed_state: Dict[str, Any]
+
+    modified: bool = False
+    """When ``True`` the object was modified."""
+    expired: bool = False
+    """When ``True`` the object is :term:`expired`.
+
+    .. seealso::
+
+        :ref:`session_expire`
+    """
+    _deleted: bool = False
+    _load_pending: bool = False
+    _orphaned_outside_of_session: bool = False
+    is_instance: bool = True
+    identity_token: object = None
+    _last_known_values: Optional[Dict[str, Any]] = None
+
+    _instance_dict: _InstanceDictProto
+    """A weak reference, or in the default case a plain callable, that
+    returns a reference to the current :class:`.IdentityMap`, if any.
+
+    """
+    if not TYPE_CHECKING:
+
+        def _instance_dict(self):
+            """default 'weak reference' for _instance_dict"""
+            return None
+
+    expired_attributes: Set[str]
+    """The set of keys which are 'expired' to be loaded by
+    the manager's deferred scalar loader, assuming no pending
+    changes.
+
+    See also the ``unmodified`` collection which is intersected
+    against this set when a refresh operation occurs.
+    """
+
+    callables: Dict[str, Callable[[InstanceState[_O], PassiveFlag], Any]]
+    """A namespace where a per-state loader callable can be associated.
+
+    In SQLAlchemy 1.0, this is only used for lazy loaders / deferred
+    loaders that were set up via query option.
+
+    Previously, callables was used also to indicate expired attributes
+    by storing a link to the InstanceState itself in this dictionary.
+    This role is now handled by the expired_attributes set.
+
+    """
+
+    if not TYPE_CHECKING:
+        callables = util.EMPTY_DICT
+
+    def __init__(self, obj: _O, manager: ClassManager[_O]):
+        self.class_ = obj.__class__
+        self.manager = manager
+        self.obj = weakref.ref(obj, self._cleanup)
+        self.committed_state = {}
+        self.expired_attributes = set()
+
+    @util.memoized_property
+    def attrs(self) -> util.ReadOnlyProperties[AttributeState]:
+        """Return a namespace representing each attribute on
+        the mapped object, including its current value
+        and history.
+
+        The returned object is an instance of :class:`.AttributeState`.
+        This object allows inspection of the current data
+        within an attribute as well as attribute history
+        since the last flush.
+
+        """
+        return util.ReadOnlyProperties(
+            {key: AttributeState(self, key) for key in self.manager}
+        )
+
+    @property
+    def transient(self) -> bool:
+        """Return ``True`` if the object is :term:`transient`.
+
+        .. seealso::
+
+            :ref:`session_object_states`
+
+        """
+        return self.key is None and not self._attached
+
+    @property
+    def pending(self) -> bool:
+        """Return ``True`` if the object is :term:`pending`.
+
+        .. seealso::
+
+            :ref:`session_object_states`
+
+        """
+        return self.key is None and self._attached
+
+    @property
+    def deleted(self) -> bool:
+        """Return ``True`` if the object is :term:`deleted`.
+
+        An object that is in the deleted state is guaranteed to
+        not be within the :attr:`.Session.identity_map` of its parent
+        :class:`.Session`; however if the session's transaction is rolled
+        back, the object will be restored to the persistent state and
+        the identity map.
+
+        .. note::
+
+            The :attr:`.InstanceState.deleted` attribute refers to a specific
+            state of the object that occurs between the "persistent" and
+            "detached" states; once the object is :term:`detached`, the
+            :attr:`.InstanceState.deleted` attribute **no longer returns
+            True**; in order to detect that a state was deleted, regardless
+            of whether or not the object is associated with a
+            :class:`.Session`, use the :attr:`.InstanceState.was_deleted`
+            accessor.
+
+        .. versionadded: 1.1
+
+        .. seealso::
+
+            :ref:`session_object_states`
+
+        """
+        return self.key is not None and self._attached and self._deleted
+
+    @property
+    def was_deleted(self) -> bool:
+        """Return True if this object is or was previously in the
+        "deleted" state and has not been reverted to persistent.
+
+        This flag returns True once the object was deleted in flush.
+        When the object is expunged from the session either explicitly
+        or via transaction commit and enters the "detached" state,
+        this flag will continue to report True.
+
+        .. seealso::
+
+            :attr:`.InstanceState.deleted` - refers to the "deleted" state
+
+            :func:`.orm.util.was_deleted` - standalone function
+
+            :ref:`session_object_states`
+
+        """
+        return self._deleted
+
+    @property
+    def persistent(self) -> bool:
+        """Return ``True`` if the object is :term:`persistent`.
+
+        An object that is in the persistent state is guaranteed to
+        be within the :attr:`.Session.identity_map` of its parent
+        :class:`.Session`.
+
+        .. seealso::
+
+            :ref:`session_object_states`
+
+        """
+        return self.key is not None and self._attached and not self._deleted
+
+    @property
+    def detached(self) -> bool:
+        """Return ``True`` if the object is :term:`detached`.
+
+        .. seealso::
+
+            :ref:`session_object_states`
+
+        """
+        return self.key is not None and not self._attached
+
+    @util.non_memoized_property
+    @util.preload_module("sqlalchemy.orm.session")
+    def _attached(self) -> bool:
+        return (
+            self.session_id is not None
+            and self.session_id in util.preloaded.orm_session._sessions
+        )
+
+    def _track_last_known_value(self, key: str) -> None:
+        """Track the last known value of a particular key after expiration
+        operations.
+
+        .. versionadded:: 1.3
+
+        """
+
+        lkv = self._last_known_values
+        if lkv is None:
+            self._last_known_values = lkv = {}
+        if key not in lkv:
+            lkv[key] = NO_VALUE
+
+    @property
+    def session(self) -> Optional[Session]:
+        """Return the owning :class:`.Session` for this instance,
+        or ``None`` if none available.
+
+        Note that the result here can in some cases be *different*
+        from that of ``obj in session``; an object that's been deleted
+        will report as not ``in session``, however if the transaction is
+        still in progress, this attribute will still refer to that session.
+        Only when the transaction is completed does the object become
+        fully detached under normal circumstances.
+
+        .. seealso::
+
+            :attr:`_orm.InstanceState.async_session`
+
+        """
+        if self.session_id:
+            try:
+                return _sessions[self.session_id]
+            except KeyError:
+                pass
+        return None
+
+    @property
+    def async_session(self) -> Optional[AsyncSession]:
+        """Return the owning :class:`_asyncio.AsyncSession` for this instance,
+        or ``None`` if none available.
+
+        This attribute is only non-None when the :mod:`sqlalchemy.ext.asyncio`
+        API is in use for this ORM object. The returned
+        :class:`_asyncio.AsyncSession` object will be a proxy for the
+        :class:`_orm.Session` object that would be returned from the
+        :attr:`_orm.InstanceState.session` attribute for this
+        :class:`_orm.InstanceState`.
+
+        .. versionadded:: 1.4.18
+
+        .. seealso::
+
+            :ref:`asyncio_toplevel`
+
+        """
+        if _async_provider is None:
+            return None
+
+        sess = self.session
+        if sess is not None:
+            return _async_provider(sess)
+        else:
+            return None
+
+    @property
+    def object(self) -> Optional[_O]:
+        """Return the mapped object represented by this
+        :class:`.InstanceState`.
+
+        Returns None if the object has been garbage collected
+
+        """
+        return self.obj()
+
+    @property
+    def identity(self) -> Optional[Tuple[Any, ...]]:
+        """Return the mapped identity of the mapped object.
+        This is the primary key identity as persisted by the ORM
+        which can always be passed directly to
+        :meth:`_query.Query.get`.
+
+        Returns ``None`` if the object has no primary key identity.
+
+        .. note::
+            An object which is :term:`transient` or :term:`pending`
+            does **not** have a mapped identity until it is flushed,
+            even if its attributes include primary key values.
+
+        """
+        if self.key is None:
+            return None
+        else:
+            return self.key[1]
+
+    @property
+    def identity_key(self) -> Optional[_IdentityKeyType[_O]]:
+        """Return the identity key for the mapped object.
+
+        This is the key used to locate the object within
+        the :attr:`.Session.identity_map` mapping.   It contains
+        the identity as returned by :attr:`.identity` within it.
+
+
+        """
+        return self.key
+
+    @util.memoized_property
+    def parents(self) -> Dict[int, Union[Literal[False], InstanceState[Any]]]:
+        return {}
+
+    @util.memoized_property
+    def _pending_mutations(self) -> Dict[str, PendingCollection]:
+        return {}
+
+    @util.memoized_property
+    def _empty_collections(self) -> Dict[str, _AdaptedCollectionProtocol]:
+        return {}
+
+    @util.memoized_property
+    def mapper(self) -> Mapper[_O]:
+        """Return the :class:`_orm.Mapper` used for this mapped object."""
+        return self.manager.mapper
+
+    @property
+    def has_identity(self) -> bool:
+        """Return ``True`` if this object has an identity key.
+
+        This should always have the same value as the
+        expression ``state.persistent`` or ``state.detached``.
+
+        """
+        return bool(self.key)
+
+    @classmethod
+    def _detach_states(
+        self,
+        states: Iterable[InstanceState[_O]],
+        session: Session,
+        to_transient: bool = False,
+    ) -> None:
+        persistent_to_detached = (
+            session.dispatch.persistent_to_detached or None
+        )
+        deleted_to_detached = session.dispatch.deleted_to_detached or None
+        pending_to_transient = session.dispatch.pending_to_transient or None
+        persistent_to_transient = (
+            session.dispatch.persistent_to_transient or None
+        )
+
+        for state in states:
+            deleted = state._deleted
+            pending = state.key is None
+            persistent = not pending and not deleted
+
+            state.session_id = None
+
+            if to_transient and state.key:
+                del state.key
+            if persistent:
+                if to_transient:
+                    if persistent_to_transient is not None:
+                        persistent_to_transient(session, state)
+                elif persistent_to_detached is not None:
+                    persistent_to_detached(session, state)
+            elif deleted and deleted_to_detached is not None:
+                deleted_to_detached(session, state)
+            elif pending and pending_to_transient is not None:
+                pending_to_transient(session, state)
+
+            state._strong_obj = None
+
+    def _detach(self, session: Optional[Session] = None) -> None:
+        if session:
+            InstanceState._detach_states([self], session)
+        else:
+            self.session_id = self._strong_obj = None
+
+    def _dispose(self) -> None:
+        # used by the test suite, apparently
+        self._detach()
+
+    def _cleanup(self, ref: weakref.ref[_O]) -> None:
+        """Weakref callback cleanup.
+
+        This callable cleans out the state when it is being garbage
+        collected.
+
+        this _cleanup **assumes** that there are no strong refs to us!
+        Will not work otherwise!
+
+        """
+
+        # Python builtins become undefined during interpreter shutdown.
+        # Guard against exceptions during this phase, as the method cannot
+        # proceed in any case if builtins have been undefined.
+        if dict is None:
+            return
+
+        instance_dict = self._instance_dict()
+        if instance_dict is not None:
+            instance_dict._fast_discard(self)
+            del self._instance_dict
+
+            # we can't possibly be in instance_dict._modified
+            # b.c. this is weakref cleanup only, that set
+            # is strong referencing!
+            # assert self not in instance_dict._modified
+
+        self.session_id = self._strong_obj = None
+
+    @property
+    def dict(self) -> _InstanceDict:
+        """Return the instance dict used by the object.
+
+        Under normal circumstances, this is always synonymous
+        with the ``__dict__`` attribute of the mapped object,
+        unless an alternative instrumentation system has been
+        configured.
+
+        In the case that the actual object has been garbage
+        collected, this accessor returns a blank dictionary.
+
+        """
+        o = self.obj()
+        if o is not None:
+            return base.instance_dict(o)
+        else:
+            return {}
+
+    def _initialize_instance(*mixed: Any, **kwargs: Any) -> None:
+        self, instance, args = mixed[0], mixed[1], mixed[2:]  # noqa
+        manager = self.manager
+
+        manager.dispatch.init(self, args, kwargs)
+
+        try:
+            manager.original_init(*mixed[1:], **kwargs)
+        except:
+            with util.safe_reraise():
+                manager.dispatch.init_failure(self, args, kwargs)
+
+    def get_history(self, key: str, passive: PassiveFlag) -> History:
+        return self.manager[key].impl.get_history(self, self.dict, passive)
+
+    def get_impl(self, key: str) -> AttributeImpl:
+        return self.manager[key].impl
+
+    def _get_pending_mutation(self, key: str) -> PendingCollection:
+        if key not in self._pending_mutations:
+            self._pending_mutations[key] = PendingCollection()
+        return self._pending_mutations[key]
+
+    def __getstate__(self) -> Dict[str, Any]:
+        state_dict: Dict[str, Any] = {
+            "instance": self.obj(),
+            "class_": self.class_,
+            "committed_state": self.committed_state,
+            "expired_attributes": self.expired_attributes,
+        }
+        state_dict.update(
+            (k, self.__dict__[k])
+            for k in (
+                "_pending_mutations",
+                "modified",
+                "expired",
+                "callables",
+                "key",
+                "parents",
+                "load_options",
+                "class_",
+                "expired_attributes",
+                "info",
+            )
+            if k in self.__dict__
+        )
+        if self.load_path:
+            state_dict["load_path"] = self.load_path.serialize()
+
+        state_dict["manager"] = self.manager._serialize(self, state_dict)
+
+        return state_dict
+
+    def __setstate__(self, state_dict: Dict[str, Any]) -> None:
+        inst = state_dict["instance"]
+        if inst is not None:
+            self.obj = weakref.ref(inst, self._cleanup)
+            self.class_ = inst.__class__
+        else:
+            self.obj = lambda: None  # type: ignore
+            self.class_ = state_dict["class_"]
+
+        self.committed_state = state_dict.get("committed_state", {})
+        self._pending_mutations = state_dict.get("_pending_mutations", {})
+        self.parents = state_dict.get("parents", {})
+        self.modified = state_dict.get("modified", False)
+        self.expired = state_dict.get("expired", False)
+        if "info" in state_dict:
+            self.info.update(state_dict["info"])
+        if "callables" in state_dict:
+            self.callables = state_dict["callables"]
+
+            self.expired_attributes = state_dict["expired_attributes"]
+        else:
+            if "expired_attributes" in state_dict:
+                self.expired_attributes = state_dict["expired_attributes"]
+            else:
+                self.expired_attributes = set()
+
+        self.__dict__.update(
+            [
+                (k, state_dict[k])
+                for k in ("key", "load_options")
+                if k in state_dict
+            ]
+        )
+        if self.key:
+            self.identity_token = self.key[2]
+
+        if "load_path" in state_dict:
+            self.load_path = PathRegistry.deserialize(state_dict["load_path"])
+
+        state_dict["manager"](self, inst, state_dict)
+
+    def _reset(self, dict_: _InstanceDict, key: str) -> None:
+        """Remove the given attribute and any
+        callables associated with it."""
+
+        old = dict_.pop(key, None)
+        manager_impl = self.manager[key].impl
+        if old is not None and is_collection_impl(manager_impl):
+            manager_impl._invalidate_collection(old)
+        self.expired_attributes.discard(key)
+        if self.callables:
+            self.callables.pop(key, None)
+
+    def _copy_callables(self, from_: InstanceState[Any]) -> None:
+        if "callables" in from_.__dict__:
+            self.callables = dict(from_.callables)
+
+    @classmethod
+    def _instance_level_callable_processor(
+        cls, manager: ClassManager[_O], fn: _LoaderCallable, key: Any
+    ) -> _InstallLoaderCallableProto[_O]:
+        impl = manager[key].impl
+        if is_collection_impl(impl):
+            fixed_impl = impl
+
+            def _set_callable(
+                state: InstanceState[_O], dict_: _InstanceDict, row: Row[Any]
+            ) -> None:
+                if "callables" not in state.__dict__:
+                    state.callables = {}
+                old = dict_.pop(key, None)
+                if old is not None:
+                    fixed_impl._invalidate_collection(old)
+                state.callables[key] = fn
+
+        else:
+
+            def _set_callable(
+                state: InstanceState[_O], dict_: _InstanceDict, row: Row[Any]
+            ) -> None:
+                if "callables" not in state.__dict__:
+                    state.callables = {}
+                state.callables[key] = fn
+
+        return _set_callable
+
+    def _expire(
+        self, dict_: _InstanceDict, modified_set: Set[InstanceState[Any]]
+    ) -> None:
+        self.expired = True
+        if self.modified:
+            modified_set.discard(self)
+            self.committed_state.clear()
+            self.modified = False
+
+        self._strong_obj = None
+
+        if "_pending_mutations" in self.__dict__:
+            del self.__dict__["_pending_mutations"]
+
+        if "parents" in self.__dict__:
+            del self.__dict__["parents"]
+
+        self.expired_attributes.update(
+            [impl.key for impl in self.manager._loader_impls]
+        )
+
+        if self.callables:
+            # the per state loader callables we can remove here are
+            # LoadDeferredColumns, which undefers a column at the instance
+            # level that is mapped with deferred, and LoadLazyAttribute,
+            # which lazy loads a relationship at the instance level that
+            # is mapped with "noload" or perhaps "immediateload".
+            # Before 1.4, only column-based
+            # attributes could be considered to be "expired", so here they
+            # were the only ones "unexpired", which means to make them deferred
+            # again.   For the moment, as of 1.4 we also apply the same
+            # treatment relationships now, that is, an instance level lazy
+            # loader is reset in the same way as a column loader.
+            for k in self.expired_attributes.intersection(self.callables):
+                del self.callables[k]
+
+        for k in self.manager._collection_impl_keys.intersection(dict_):
+            collection = dict_.pop(k)
+            collection._sa_adapter.invalidated = True
+
+        if self._last_known_values:
+            self._last_known_values.update(
+                {k: dict_[k] for k in self._last_known_values if k in dict_}
+            )
+
+        for key in self.manager._all_key_set.intersection(dict_):
+            del dict_[key]
+
+        self.manager.dispatch.expire(self, None)
+
+    def _expire_attributes(
+        self,
+        dict_: _InstanceDict,
+        attribute_names: Iterable[str],
+        no_loader: bool = False,
+    ) -> None:
+        pending = self.__dict__.get("_pending_mutations", None)
+
+        callables = self.callables
+
+        for key in attribute_names:
+            impl = self.manager[key].impl
+            if impl.accepts_scalar_loader:
+                if no_loader and (impl.callable_ or key in callables):
+                    continue
+
+                self.expired_attributes.add(key)
+                if callables and key in callables:
+                    del callables[key]
+            old = dict_.pop(key, NO_VALUE)
+            if is_collection_impl(impl) and old is not NO_VALUE:
+                impl._invalidate_collection(old)
+
+            lkv = self._last_known_values
+            if lkv is not None and key in lkv and old is not NO_VALUE:
+                lkv[key] = old
+
+            self.committed_state.pop(key, None)
+            if pending:
+                pending.pop(key, None)
+
+        self.manager.dispatch.expire(self, attribute_names)
+
+    def _load_expired(
+        self, state: InstanceState[_O], passive: PassiveFlag
+    ) -> LoaderCallableStatus:
+        """__call__ allows the InstanceState to act as a deferred
+        callable for loading expired attributes, which is also
+        serializable (picklable).
+
+        """
+
+        if not passive & SQL_OK:
+            return PASSIVE_NO_RESULT
+
+        toload = self.expired_attributes.intersection(self.unmodified)
+        toload = toload.difference(
+            attr
+            for attr in toload
+            if not self.manager[attr].impl.load_on_unexpire
+        )
+
+        self.manager.expired_attribute_loader(self, toload, passive)
+
+        # if the loader failed, or this
+        # instance state didn't have an identity,
+        # the attributes still might be in the callables
+        # dict.  ensure they are removed.
+        self.expired_attributes.clear()
+
+        return ATTR_WAS_SET
+
+    @property
+    def unmodified(self) -> Set[str]:
+        """Return the set of keys which have no uncommitted changes"""
+
+        return set(self.manager).difference(self.committed_state)
+
+    def unmodified_intersection(self, keys: Iterable[str]) -> Set[str]:
+        """Return self.unmodified.intersection(keys)."""
+
+        return (
+            set(keys)
+            .intersection(self.manager)
+            .difference(self.committed_state)
+        )
+
+    @property
+    def unloaded(self) -> Set[str]:
+        """Return the set of keys which do not have a loaded value.
+
+        This includes expired attributes and any other attribute that was never
+        populated or modified.
+
+        """
+        return (
+            set(self.manager)
+            .difference(self.committed_state)
+            .difference(self.dict)
+        )
+
+    @property
+    @util.deprecated(
+        "2.0",
+        "The :attr:`.InstanceState.unloaded_expirable` attribute is "
+        "deprecated.  Please use :attr:`.InstanceState.unloaded`.",
+    )
+    def unloaded_expirable(self) -> Set[str]:
+        """Synonymous with :attr:`.InstanceState.unloaded`.
+
+        This attribute was added as an implementation-specific detail at some
+        point and should be considered to be private.
+
+        """
+        return self.unloaded
+
+    @property
+    def _unloaded_non_object(self) -> Set[str]:
+        return self.unloaded.intersection(
+            attr
+            for attr in self.manager
+            if self.manager[attr].impl.accepts_scalar_loader
+        )
+
+    def _modified_event(
+        self,
+        dict_: _InstanceDict,
+        attr: Optional[AttributeImpl],
+        previous: Any,
+        collection: bool = False,
+        is_userland: bool = False,
+    ) -> None:
+        if attr:
+            if not attr.send_modified_events:
+                return
+            if is_userland and attr.key not in dict_:
+                raise sa_exc.InvalidRequestError(
+                    "Can't flag attribute '%s' modified; it's not present in "
+                    "the object state" % attr.key
+                )
+            if attr.key not in self.committed_state or is_userland:
+                if collection:
+                    if TYPE_CHECKING:
+                        assert is_collection_impl(attr)
+                    if previous is NEVER_SET:
+                        if attr.key in dict_:
+                            previous = dict_[attr.key]
+
+                    if previous not in (None, NO_VALUE, NEVER_SET):
+                        previous = attr.copy(previous)
+                self.committed_state[attr.key] = previous
+
+            lkv = self._last_known_values
+            if lkv is not None and attr.key in lkv:
+                lkv[attr.key] = NO_VALUE
+
+        # assert self._strong_obj is None or self.modified
+
+        if (self.session_id and self._strong_obj is None) or not self.modified:
+            self.modified = True
+            instance_dict = self._instance_dict()
+            if instance_dict:
+                has_modified = bool(instance_dict._modified)
+                instance_dict._modified.add(self)
+            else:
+                has_modified = False
+
+            # only create _strong_obj link if attached
+            # to a session
+
+            inst = self.obj()
+            if self.session_id:
+                self._strong_obj = inst
+
+                # if identity map already had modified objects,
+                # assume autobegin already occurred, else check
+                # for autobegin
+                if not has_modified:
+                    # inline of autobegin, to ensure session transaction
+                    # snapshot is established
+                    try:
+                        session = _sessions[self.session_id]
+                    except KeyError:
+                        pass
+                    else:
+                        if session._transaction is None:
+                            session._autobegin_t()
+
+            if inst is None and attr:
+                raise orm_exc.ObjectDereferencedError(
+                    "Can't emit change event for attribute '%s' - "
+                    "parent object of type %s has been garbage "
+                    "collected."
+                    % (self.manager[attr.key], base.state_class_str(self))
+                )
+
+    def _commit(self, dict_: _InstanceDict, keys: Iterable[str]) -> None:
+        """Commit attributes.
+
+        This is used by a partial-attribute load operation to mark committed
+        those attributes which were refreshed from the database.
+
+        Attributes marked as "expired" can potentially remain "expired" after
+        this step if a value was not populated in state.dict.
+
+        """
+        for key in keys:
+            self.committed_state.pop(key, None)
+
+        self.expired = False
+
+        self.expired_attributes.difference_update(
+            set(keys).intersection(dict_)
+        )
+
+        # the per-keys commit removes object-level callables,
+        # while that of commit_all does not.  it's not clear
+        # if this behavior has a clear rationale, however tests do
+        # ensure this is what it does.
+        if self.callables:
+            for key in (
+                set(self.callables).intersection(keys).intersection(dict_)
+            ):
+                del self.callables[key]
+
+    def _commit_all(
+        self, dict_: _InstanceDict, instance_dict: Optional[IdentityMap] = None
+    ) -> None:
+        """commit all attributes unconditionally.
+
+        This is used after a flush() or a full load/refresh
+        to remove all pending state from the instance.
+
+         - all attributes are marked as "committed"
+         - the "strong dirty reference" is removed
+         - the "modified" flag is set to False
+         - any "expired" markers for scalar attributes loaded are removed.
+         - lazy load callables for objects / collections *stay*
+
+        Attributes marked as "expired" can potentially remain
+        "expired" after this step if a value was not populated in state.dict.
+
+        """
+        self._commit_all_states([(self, dict_)], instance_dict)
+
+    @classmethod
+    def _commit_all_states(
+        self,
+        iter_: Iterable[Tuple[InstanceState[Any], _InstanceDict]],
+        instance_dict: Optional[IdentityMap] = None,
+    ) -> None:
+        """Mass / highly inlined version of commit_all()."""
+
+        for state, dict_ in iter_:
+            state_dict = state.__dict__
+
+            state.committed_state.clear()
+
+            if "_pending_mutations" in state_dict:
+                del state_dict["_pending_mutations"]
+
+            state.expired_attributes.difference_update(dict_)
+
+            if instance_dict and state.modified:
+                instance_dict._modified.discard(state)
+
+            state.modified = state.expired = False
+            state._strong_obj = None
+
+
+class AttributeState:
+    """Provide an inspection interface corresponding
+    to a particular attribute on a particular mapped object.
+
+    The :class:`.AttributeState` object is accessed
+    via the :attr:`.InstanceState.attrs` collection
+    of a particular :class:`.InstanceState`::
+
+        from sqlalchemy import inspect
+
+        insp = inspect(some_mapped_object)
+        attr_state = insp.attrs.some_attribute
+
+    """
+
+    __slots__ = ("state", "key")
+
+    state: InstanceState[Any]
+    key: str
+
+    def __init__(self, state: InstanceState[Any], key: str):
+        self.state = state
+        self.key = key
+
+    @property
+    def loaded_value(self) -> Any:
+        """The current value of this attribute as loaded from the database.
+
+        If the value has not been loaded, or is otherwise not present
+        in the object's dictionary, returns NO_VALUE.
+
+        """
+        return self.state.dict.get(self.key, NO_VALUE)
+
+    @property
+    def value(self) -> Any:
+        """Return the value of this attribute.
+
+        This operation is equivalent to accessing the object's
+        attribute directly or via ``getattr()``, and will fire
+        off any pending loader callables if needed.
+
+        """
+        return self.state.manager[self.key].__get__(
+            self.state.obj(), self.state.class_
+        )
+
+    @property
+    def history(self) -> History:
+        """Return the current **pre-flush** change history for
+        this attribute, via the :class:`.History` interface.
+
+        This method will **not** emit loader callables if the value of the
+        attribute is unloaded.
+
+        .. note::
+
+            The attribute history system tracks changes on a **per flush
+            basis**. Each time the :class:`.Session` is flushed, the history
+            of each attribute is reset to empty.   The :class:`.Session` by
+            default autoflushes each time a :class:`_query.Query` is invoked.
+            For
+            options on how to control this, see :ref:`session_flushing`.
+
+
+        .. seealso::
+
+            :meth:`.AttributeState.load_history` - retrieve history
+            using loader callables if the value is not locally present.
+
+            :func:`.attributes.get_history` - underlying function
+
+        """
+        return self.state.get_history(self.key, PASSIVE_NO_INITIALIZE)
+
+    def load_history(self) -> History:
+        """Return the current **pre-flush** change history for
+        this attribute, via the :class:`.History` interface.
+
+        This method **will** emit loader callables if the value of the
+        attribute is unloaded.
+
+        .. note::
+
+            The attribute history system tracks changes on a **per flush
+            basis**. Each time the :class:`.Session` is flushed, the history
+            of each attribute is reset to empty.   The :class:`.Session` by
+            default autoflushes each time a :class:`_query.Query` is invoked.
+            For
+            options on how to control this, see :ref:`session_flushing`.
+
+        .. seealso::
+
+            :attr:`.AttributeState.history`
+
+            :func:`.attributes.get_history` - underlying function
+
+        """
+        return self.state.get_history(self.key, PASSIVE_OFF ^ INIT_OK)
+
+
+class PendingCollection:
+    """A writable placeholder for an unloaded collection.
+
+    Stores items appended to and removed from a collection that has not yet
+    been loaded. When the collection is loaded, the changes stored in
+    PendingCollection are applied to it to produce the final result.
+
+    """
+
+    __slots__ = ("deleted_items", "added_items")
+
+    deleted_items: util.IdentitySet
+    added_items: util.OrderedIdentitySet
+
+    def __init__(self) -> None:
+        self.deleted_items = util.IdentitySet()
+        self.added_items = util.OrderedIdentitySet()
+
+    def merge_with_history(self, history: History) -> History:
+        return history._merge(self.added_items, self.deleted_items)
+
+    def append(self, value: Any) -> None:
+        if value in self.deleted_items:
+            self.deleted_items.remove(value)
+        else:
+            self.added_items.add(value)
+
+    def remove(self, value: Any) -> None:
+        if value in self.added_items:
+            self.added_items.remove(value)
+        else:
+            self.deleted_items.add(value)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state_changes.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state_changes.py
new file mode 100644
index 00000000..10e417e8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/state_changes.py
@@ -0,0 +1,198 @@
+# orm/state_changes.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""State tracking utilities used by :class:`_orm.Session`.
+
+"""
+
+from __future__ import annotations
+
+import contextlib
+from enum import Enum
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Iterator
+from typing import NoReturn
+from typing import Optional
+from typing import Tuple
+from typing import TypeVar
+from typing import Union
+
+from .. import exc as sa_exc
+from .. import util
+from ..util.typing import Literal
+
+_F = TypeVar("_F", bound=Callable[..., Any])
+
+
+class _StateChangeState(Enum):
+    pass
+
+
+class _StateChangeStates(_StateChangeState):
+    ANY = 1
+    NO_CHANGE = 2
+    CHANGE_IN_PROGRESS = 3
+
+
+class _StateChange:
+    """Supplies state assertion decorators.
+
+    The current use case is for the :class:`_orm.SessionTransaction` class. The
+    :class:`_StateChange` class itself is agnostic of the
+    :class:`_orm.SessionTransaction` class so could in theory be generalized
+    for other systems as well.
+
+    """
+
+    _next_state: _StateChangeState = _StateChangeStates.ANY
+    _state: _StateChangeState = _StateChangeStates.NO_CHANGE
+    _current_fn: Optional[Callable[..., Any]] = None
+
+    def _raise_for_prerequisite_state(
+        self, operation_name: str, state: _StateChangeState
+    ) -> NoReturn:
+        raise sa_exc.IllegalStateChangeError(
+            f"Can't run operation '{operation_name}()' when Session "
+            f"is in state {state!r}",
+            code="isce",
+        )
+
+    @classmethod
+    def declare_states(
+        cls,
+        prerequisite_states: Union[
+            Literal[_StateChangeStates.ANY], Tuple[_StateChangeState, ...]
+        ],
+        moves_to: _StateChangeState,
+    ) -> Callable[[_F], _F]:
+        """Method decorator declaring valid states.
+
+        :param prerequisite_states: sequence of acceptable prerequisite
+         states.   Can be the single constant _State.ANY to indicate no
+         prerequisite state
+
+        :param moves_to: the expected state at the end of the method, assuming
+         no exceptions raised.   Can be the constant _State.NO_CHANGE to
+         indicate state should not change at the end of the method.
+
+        """
+        assert prerequisite_states, "no prequisite states sent"
+        has_prerequisite_states = (
+            prerequisite_states is not _StateChangeStates.ANY
+        )
+
+        prerequisite_state_collection = cast(
+            "Tuple[_StateChangeState, ...]", prerequisite_states
+        )
+        expect_state_change = moves_to is not _StateChangeStates.NO_CHANGE
+
+        @util.decorator
+        def _go(fn: _F, self: Any, *arg: Any, **kw: Any) -> Any:
+            current_state = self._state
+
+            if (
+                has_prerequisite_states
+                and current_state not in prerequisite_state_collection
+            ):
+                self._raise_for_prerequisite_state(fn.__name__, current_state)
+
+            next_state = self._next_state
+            existing_fn = self._current_fn
+            expect_state = moves_to if expect_state_change else current_state
+
+            if (
+                # destination states are restricted
+                next_state is not _StateChangeStates.ANY
+                # method seeks to change state
+                and expect_state_change
+                # destination state incorrect
+                and next_state is not expect_state
+            ):
+                if existing_fn and next_state in (
+                    _StateChangeStates.NO_CHANGE,
+                    _StateChangeStates.CHANGE_IN_PROGRESS,
+                ):
+                    raise sa_exc.IllegalStateChangeError(
+                        f"Method '{fn.__name__}()' can't be called here; "
+                        f"method '{existing_fn.__name__}()' is already "
+                        f"in progress and this would cause an unexpected "
+                        f"state change to {moves_to!r}",
+                        code="isce",
+                    )
+                else:
+                    raise sa_exc.IllegalStateChangeError(
+                        f"Cant run operation '{fn.__name__}()' here; "
+                        f"will move to state {moves_to!r} where we are "
+                        f"expecting {next_state!r}",
+                        code="isce",
+                    )
+
+            self._current_fn = fn
+            self._next_state = _StateChangeStates.CHANGE_IN_PROGRESS
+            try:
+                ret_value = fn(self, *arg, **kw)
+            except:
+                raise
+            else:
+                if self._state is expect_state:
+                    return ret_value
+
+                if self._state is current_state:
+                    raise sa_exc.IllegalStateChangeError(
+                        f"Method '{fn.__name__}()' failed to "
+                        "change state "
+                        f"to {moves_to!r} as expected",
+                        code="isce",
+                    )
+                elif existing_fn:
+                    raise sa_exc.IllegalStateChangeError(
+                        f"While method '{existing_fn.__name__}()' was "
+                        "running, "
+                        f"method '{fn.__name__}()' caused an "
+                        "unexpected "
+                        f"state change to {self._state!r}",
+                        code="isce",
+                    )
+                else:
+                    raise sa_exc.IllegalStateChangeError(
+                        f"Method '{fn.__name__}()' caused an unexpected "
+                        f"state change to {self._state!r}",
+                        code="isce",
+                    )
+
+            finally:
+                self._next_state = next_state
+                self._current_fn = existing_fn
+
+        return _go
+
+    @contextlib.contextmanager
+    def _expect_state(self, expected: _StateChangeState) -> Iterator[Any]:
+        """called within a method that changes states.
+
+        method must also use the ``@declare_states()`` decorator.
+
+        """
+        assert self._next_state is _StateChangeStates.CHANGE_IN_PROGRESS, (
+            "Unexpected call to _expect_state outside of "
+            "state-changing method"
+        )
+
+        self._next_state = expected
+        try:
+            yield
+        except:
+            raise
+        else:
+            if self._state is not expected:
+                raise sa_exc.IllegalStateChangeError(
+                    f"Unexpected state change to {self._state!r}", code="isce"
+                )
+        finally:
+            self._next_state = _StateChangeStates.CHANGE_IN_PROGRESS
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategies.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategies.py
new file mode 100644
index 00000000..f2d16514
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategies.py
@@ -0,0 +1,3473 @@
+# orm/strategies.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""sqlalchemy.orm.interfaces.LoaderStrategy
+   implementations, and related MapperOptions."""
+
+from __future__ import annotations
+
+import collections
+import itertools
+from typing import Any
+from typing import Dict
+from typing import Optional
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import attributes
+from . import exc as orm_exc
+from . import interfaces
+from . import loading
+from . import path_registry
+from . import properties
+from . import query
+from . import relationships
+from . import unitofwork
+from . import util as orm_util
+from .base import _DEFER_FOR_STATE
+from .base import _RAISE_FOR_STATE
+from .base import _SET_DEFERRED_EXPIRED
+from .base import ATTR_WAS_SET
+from .base import LoaderCallableStatus
+from .base import PASSIVE_OFF
+from .base import PassiveFlag
+from .context import _column_descriptions
+from .context import ORMCompileState
+from .context import ORMSelectCompileState
+from .context import QueryContext
+from .interfaces import LoaderStrategy
+from .interfaces import StrategizedProperty
+from .session import _state_session
+from .state import InstanceState
+from .strategy_options import Load
+from .util import _none_only_set
+from .util import AliasedClass
+from .. import event
+from .. import exc as sa_exc
+from .. import inspect
+from .. import log
+from .. import sql
+from .. import util
+from ..sql import util as sql_util
+from ..sql import visitors
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..sql.selectable import Select
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from .mapper import Mapper
+    from .relationships import RelationshipProperty
+    from ..sql.elements import ColumnElement
+
+
+def _register_attribute(
+    prop,
+    mapper,
+    useobject,
+    compare_function=None,
+    typecallable=None,
+    callable_=None,
+    proxy_property=None,
+    active_history=False,
+    impl_class=None,
+    **kw,
+):
+    listen_hooks = []
+
+    uselist = useobject and prop.uselist
+
+    if useobject and prop.single_parent:
+        listen_hooks.append(single_parent_validator)
+
+    if prop.key in prop.parent.validators:
+        fn, opts = prop.parent.validators[prop.key]
+        listen_hooks.append(
+            lambda desc, prop: orm_util._validator_events(
+                desc, prop.key, fn, **opts
+            )
+        )
+
+    if useobject:
+        listen_hooks.append(unitofwork.track_cascade_events)
+
+    # need to assemble backref listeners
+    # after the singleparentvalidator, mapper validator
+    if useobject:
+        backref = prop.back_populates
+        if backref and prop._effective_sync_backref:
+            listen_hooks.append(
+                lambda desc, prop: attributes.backref_listeners(
+                    desc, backref, uselist
+                )
+            )
+
+    # a single MapperProperty is shared down a class inheritance
+    # hierarchy, so we set up attribute instrumentation and backref event
+    # for each mapper down the hierarchy.
+
+    # typically, "mapper" is the same as prop.parent, due to the way
+    # the configure_mappers() process runs, however this is not strongly
+    # enforced, and in the case of a second configure_mappers() run the
+    # mapper here might not be prop.parent; also, a subclass mapper may
+    # be called here before a superclass mapper.  That is, can't depend
+    # on mappers not already being set up so we have to check each one.
+
+    for m in mapper.self_and_descendants:
+        if prop is m._props.get(
+            prop.key
+        ) and not m.class_manager._attr_has_impl(prop.key):
+            desc = attributes.register_attribute_impl(
+                m.class_,
+                prop.key,
+                parent_token=prop,
+                uselist=uselist,
+                compare_function=compare_function,
+                useobject=useobject,
+                trackparent=useobject
+                and (
+                    prop.single_parent
+                    or prop.direction is interfaces.ONETOMANY
+                ),
+                typecallable=typecallable,
+                callable_=callable_,
+                active_history=active_history,
+                impl_class=impl_class,
+                send_modified_events=not useobject or not prop.viewonly,
+                doc=prop.doc,
+                **kw,
+            )
+
+            for hook in listen_hooks:
+                hook(desc, prop)
+
+
+@properties.ColumnProperty.strategy_for(instrument=False, deferred=False)
+class UninstrumentedColumnLoader(LoaderStrategy):
+    """Represent a non-instrumented MapperProperty.
+
+    The polymorphic_on argument of mapper() often results in this,
+    if the argument is against the with_polymorphic selectable.
+
+    """
+
+    __slots__ = ("columns",)
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.columns = self.parent_property.columns
+
+    def setup_query(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection=None,
+        **kwargs,
+    ):
+        for c in self.columns:
+            if adapter:
+                c = adapter.columns[c]
+            compile_state._append_dedupe_col_collection(c, column_collection)
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        pass
+
+
+@log.class_logger
+@properties.ColumnProperty.strategy_for(instrument=True, deferred=False)
+class ColumnLoader(LoaderStrategy):
+    """Provide loading behavior for a :class:`.ColumnProperty`."""
+
+    __slots__ = "columns", "is_composite"
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.columns = self.parent_property.columns
+        self.is_composite = hasattr(self.parent_property, "composite_class")
+
+    def setup_query(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection,
+        memoized_populators,
+        check_for_adapt=False,
+        **kwargs,
+    ):
+        for c in self.columns:
+            if adapter:
+                if check_for_adapt:
+                    c = adapter.adapt_check_present(c)
+                    if c is None:
+                        return
+                else:
+                    c = adapter.columns[c]
+
+            compile_state._append_dedupe_col_collection(c, column_collection)
+
+        fetch = self.columns[0]
+        if adapter:
+            fetch = adapter.columns[fetch]
+            if fetch is None:
+                # None happens here only for dml bulk_persistence cases
+                # when context.DMLReturningColFilter is used
+                return
+
+        memoized_populators[self.parent_property] = fetch
+
+    def init_class_attribute(self, mapper):
+        self.is_class_level = True
+        coltype = self.columns[0].type
+        # TODO: check all columns ?  check for foreign key as well?
+        active_history = (
+            self.parent_property.active_history
+            or self.columns[0].primary_key
+            or (
+                mapper.version_id_col is not None
+                and mapper._columntoproperty.get(mapper.version_id_col, None)
+                is self.parent_property
+            )
+        )
+
+        _register_attribute(
+            self.parent_property,
+            mapper,
+            useobject=False,
+            compare_function=coltype.compare_values,
+            active_history=active_history,
+        )
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        # look through list of columns represented here
+        # to see which, if any, is present in the row.
+
+        for col in self.columns:
+            if adapter:
+                col = adapter.columns[col]
+            getter = result._getter(col, False)
+            if getter:
+                populators["quick"].append((self.key, getter))
+                break
+        else:
+            populators["expire"].append((self.key, True))
+
+
+@log.class_logger
+@properties.ColumnProperty.strategy_for(query_expression=True)
+class ExpressionColumnLoader(ColumnLoader):
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+
+        # compare to the "default" expression that is mapped in
+        # the column.   If it's sql.null, we don't need to render
+        # unless an expr is passed in the options.
+        null = sql.null().label(None)
+        self._have_default_expression = any(
+            not c.compare(null) for c in self.parent_property.columns
+        )
+
+    def setup_query(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection,
+        memoized_populators,
+        **kwargs,
+    ):
+        columns = None
+        if loadopt and loadopt._extra_criteria:
+            columns = loadopt._extra_criteria
+
+        elif self._have_default_expression:
+            columns = self.parent_property.columns
+
+        if columns is None:
+            return
+
+        for c in columns:
+            if adapter:
+                c = adapter.columns[c]
+            compile_state._append_dedupe_col_collection(c, column_collection)
+
+        fetch = columns[0]
+        if adapter:
+            fetch = adapter.columns[fetch]
+            if fetch is None:
+                # None is not expected to be the result of any
+                # adapter implementation here, however there may be theoretical
+                # usages of returning() with context.DMLReturningColFilter
+                return
+
+        memoized_populators[self.parent_property] = fetch
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        # look through list of columns represented here
+        # to see which, if any, is present in the row.
+        if loadopt and loadopt._extra_criteria:
+            columns = loadopt._extra_criteria
+
+            for col in columns:
+                if adapter:
+                    col = adapter.columns[col]
+                getter = result._getter(col, False)
+                if getter:
+                    populators["quick"].append((self.key, getter))
+                    break
+            else:
+                populators["expire"].append((self.key, True))
+
+    def init_class_attribute(self, mapper):
+        self.is_class_level = True
+
+        _register_attribute(
+            self.parent_property,
+            mapper,
+            useobject=False,
+            compare_function=self.columns[0].type.compare_values,
+            accepts_scalar_loader=False,
+        )
+
+
+@log.class_logger
+@properties.ColumnProperty.strategy_for(deferred=True, instrument=True)
+@properties.ColumnProperty.strategy_for(
+    deferred=True, instrument=True, raiseload=True
+)
+@properties.ColumnProperty.strategy_for(do_nothing=True)
+class DeferredColumnLoader(LoaderStrategy):
+    """Provide loading behavior for a deferred :class:`.ColumnProperty`."""
+
+    __slots__ = "columns", "group", "raiseload"
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        if hasattr(self.parent_property, "composite_class"):
+            raise NotImplementedError(
+                "Deferred loading for composite types not implemented yet"
+            )
+        self.raiseload = self.strategy_opts.get("raiseload", False)
+        self.columns = self.parent_property.columns
+        self.group = self.parent_property.group
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        # for a DeferredColumnLoader, this method is only used during a
+        # "row processor only" query; see test_deferred.py ->
+        # tests with "rowproc_only" in their name.  As of the 1.0 series,
+        # loading._instance_processor doesn't use a "row processing" function
+        # to populate columns, instead it uses data in the "populators"
+        # dictionary.  Normally, the DeferredColumnLoader.setup_query()
+        # sets up that data in the "memoized_populators" dictionary
+        # and "create_row_processor()" here is never invoked.
+
+        if (
+            context.refresh_state
+            and context.query._compile_options._only_load_props
+            and self.key in context.query._compile_options._only_load_props
+        ):
+            self.parent_property._get_strategy(
+                (("deferred", False), ("instrument", True))
+            ).create_row_processor(
+                context,
+                query_entity,
+                path,
+                loadopt,
+                mapper,
+                result,
+                adapter,
+                populators,
+            )
+
+        elif not self.is_class_level:
+            if self.raiseload:
+                set_deferred_for_local_state = (
+                    self.parent_property._raise_column_loader
+                )
+            else:
+                set_deferred_for_local_state = (
+                    self.parent_property._deferred_column_loader
+                )
+            populators["new"].append((self.key, set_deferred_for_local_state))
+        else:
+            populators["expire"].append((self.key, False))
+
+    def init_class_attribute(self, mapper):
+        self.is_class_level = True
+
+        _register_attribute(
+            self.parent_property,
+            mapper,
+            useobject=False,
+            compare_function=self.columns[0].type.compare_values,
+            callable_=self._load_for_state,
+            load_on_unexpire=False,
+        )
+
+    def setup_query(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection,
+        memoized_populators,
+        only_load_props=None,
+        **kw,
+    ):
+        if (
+            (
+                compile_state.compile_options._render_for_subquery
+                and self.parent_property._renders_in_subqueries
+            )
+            or (
+                loadopt
+                and set(self.columns).intersection(
+                    self.parent._should_undefer_in_wildcard
+                )
+            )
+            or (
+                loadopt
+                and self.group
+                and loadopt.local_opts.get(
+                    "undefer_group_%s" % self.group, False
+                )
+            )
+            or (only_load_props and self.key in only_load_props)
+        ):
+            self.parent_property._get_strategy(
+                (("deferred", False), ("instrument", True))
+            ).setup_query(
+                compile_state,
+                query_entity,
+                path,
+                loadopt,
+                adapter,
+                column_collection,
+                memoized_populators,
+                **kw,
+            )
+        elif self.is_class_level:
+            memoized_populators[self.parent_property] = _SET_DEFERRED_EXPIRED
+        elif not self.raiseload:
+            memoized_populators[self.parent_property] = _DEFER_FOR_STATE
+        else:
+            memoized_populators[self.parent_property] = _RAISE_FOR_STATE
+
+    def _load_for_state(self, state, passive):
+        if not state.key:
+            return LoaderCallableStatus.ATTR_EMPTY
+
+        if not passive & PassiveFlag.SQL_OK:
+            return LoaderCallableStatus.PASSIVE_NO_RESULT
+
+        localparent = state.manager.mapper
+
+        if self.group:
+            toload = [
+                p.key
+                for p in localparent.iterate_properties
+                if isinstance(p, StrategizedProperty)
+                and isinstance(p.strategy, DeferredColumnLoader)
+                and p.group == self.group
+            ]
+        else:
+            toload = [self.key]
+
+        # narrow the keys down to just those which have no history
+        group = [k for k in toload if k in state.unmodified]
+
+        session = _state_session(state)
+        if session is None:
+            raise orm_exc.DetachedInstanceError(
+                "Parent instance %s is not bound to a Session; "
+                "deferred load operation of attribute '%s' cannot proceed"
+                % (orm_util.state_str(state), self.key)
+            )
+
+        if self.raiseload:
+            self._invoke_raise_load(state, passive, "raise")
+
+        loading.load_scalar_attributes(
+            state.mapper, state, set(group), PASSIVE_OFF
+        )
+
+        return LoaderCallableStatus.ATTR_WAS_SET
+
+    def _invoke_raise_load(self, state, passive, lazy):
+        raise sa_exc.InvalidRequestError(
+            "'%s' is not available due to raiseload=True" % (self,)
+        )
+
+
+class LoadDeferredColumns:
+    """serializable loader object used by DeferredColumnLoader"""
+
+    def __init__(self, key: str, raiseload: bool = False):
+        self.key = key
+        self.raiseload = raiseload
+
+    def __call__(self, state, passive=attributes.PASSIVE_OFF):
+        key = self.key
+
+        localparent = state.manager.mapper
+        prop = localparent._props[key]
+        if self.raiseload:
+            strategy_key = (
+                ("deferred", True),
+                ("instrument", True),
+                ("raiseload", True),
+            )
+        else:
+            strategy_key = (("deferred", True), ("instrument", True))
+        strategy = prop._get_strategy(strategy_key)
+        return strategy._load_for_state(state, passive)
+
+
+class AbstractRelationshipLoader(LoaderStrategy):
+    """LoaderStratgies which deal with related objects."""
+
+    __slots__ = "mapper", "target", "uselist", "entity"
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.mapper = self.parent_property.mapper
+        self.entity = self.parent_property.entity
+        self.target = self.parent_property.target
+        self.uselist = self.parent_property.uselist
+
+    def _immediateload_create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        return self.parent_property._get_strategy(
+            (("lazy", "immediate"),)
+        ).create_row_processor(
+            context,
+            query_entity,
+            path,
+            loadopt,
+            mapper,
+            result,
+            adapter,
+            populators,
+        )
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(do_nothing=True)
+class DoNothingLoader(LoaderStrategy):
+    """Relationship loader that makes no change to the object's state.
+
+    Compared to NoLoader, this loader does not initialize the
+    collection/attribute to empty/none; the usual default LazyLoader will
+    take effect.
+
+    """
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy="noload")
+@relationships.RelationshipProperty.strategy_for(lazy=None)
+class NoLoader(AbstractRelationshipLoader):
+    """Provide loading behavior for a :class:`.Relationship`
+    with "lazy=None".
+
+    """
+
+    __slots__ = ()
+
+    def init_class_attribute(self, mapper):
+        self.is_class_level = True
+
+        _register_attribute(
+            self.parent_property,
+            mapper,
+            useobject=True,
+            typecallable=self.parent_property.collection_class,
+        )
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        def invoke_no_load(state, dict_, row):
+            if self.uselist:
+                attributes.init_state_collection(state, dict_, self.key)
+            else:
+                dict_[self.key] = None
+
+        populators["new"].append((self.key, invoke_no_load))
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy=True)
+@relationships.RelationshipProperty.strategy_for(lazy="select")
+@relationships.RelationshipProperty.strategy_for(lazy="raise")
+@relationships.RelationshipProperty.strategy_for(lazy="raise_on_sql")
+@relationships.RelationshipProperty.strategy_for(lazy="baked_select")
+class LazyLoader(
+    AbstractRelationshipLoader, util.MemoizedSlots, log.Identified
+):
+    """Provide loading behavior for a :class:`.Relationship`
+    with "lazy=True", that is loads when first accessed.
+
+    """
+
+    __slots__ = (
+        "_lazywhere",
+        "_rev_lazywhere",
+        "_lazyload_reverse_option",
+        "_order_by",
+        "use_get",
+        "is_aliased_class",
+        "_bind_to_col",
+        "_equated_columns",
+        "_rev_bind_to_col",
+        "_rev_equated_columns",
+        "_simple_lazy_clause",
+        "_raise_always",
+        "_raise_on_sql",
+    )
+
+    _lazywhere: ColumnElement[bool]
+    _bind_to_col: Dict[str, ColumnElement[Any]]
+    _rev_lazywhere: ColumnElement[bool]
+    _rev_bind_to_col: Dict[str, ColumnElement[Any]]
+
+    parent_property: RelationshipProperty[Any]
+
+    def __init__(
+        self, parent: RelationshipProperty[Any], strategy_key: Tuple[Any, ...]
+    ):
+        super().__init__(parent, strategy_key)
+        self._raise_always = self.strategy_opts["lazy"] == "raise"
+        self._raise_on_sql = self.strategy_opts["lazy"] == "raise_on_sql"
+
+        self.is_aliased_class = inspect(self.entity).is_aliased_class
+
+        join_condition = self.parent_property._join_condition
+        (
+            self._lazywhere,
+            self._bind_to_col,
+            self._equated_columns,
+        ) = join_condition.create_lazy_clause()
+
+        (
+            self._rev_lazywhere,
+            self._rev_bind_to_col,
+            self._rev_equated_columns,
+        ) = join_condition.create_lazy_clause(reverse_direction=True)
+
+        if self.parent_property.order_by:
+            self._order_by = [
+                sql_util._deep_annotate(elem, {"_orm_adapt": True})
+                for elem in util.to_list(self.parent_property.order_by)
+            ]
+        else:
+            self._order_by = None
+
+        self.logger.info("%s lazy loading clause %s", self, self._lazywhere)
+
+        # determine if our "lazywhere" clause is the same as the mapper's
+        # get() clause.  then we can just use mapper.get()
+        #
+        # TODO: the "not self.uselist" can be taken out entirely; a m2o
+        # load that populates for a list (very unusual, but is possible with
+        # the API) can still set for "None" and the attribute system will
+        # populate as an empty list.
+        self.use_get = (
+            not self.is_aliased_class
+            and not self.uselist
+            and self.entity._get_clause[0].compare(
+                self._lazywhere,
+                use_proxies=True,
+                compare_keys=False,
+                equivalents=self.mapper._equivalent_columns,
+            )
+        )
+
+        if self.use_get:
+            for col in list(self._equated_columns):
+                if col in self.mapper._equivalent_columns:
+                    for c in self.mapper._equivalent_columns[col]:
+                        self._equated_columns[c] = self._equated_columns[col]
+
+            self.logger.info(
+                "%s will use Session.get() to optimize instance loads", self
+            )
+
+    def init_class_attribute(self, mapper):
+        self.is_class_level = True
+
+        _legacy_inactive_history_style = (
+            self.parent_property._legacy_inactive_history_style
+        )
+
+        if self.parent_property.active_history:
+            active_history = True
+            _deferred_history = False
+
+        elif (
+            self.parent_property.direction is not interfaces.MANYTOONE
+            or not self.use_get
+        ):
+            if _legacy_inactive_history_style:
+                active_history = True
+                _deferred_history = False
+            else:
+                active_history = False
+                _deferred_history = True
+        else:
+            active_history = _deferred_history = False
+
+        _register_attribute(
+            self.parent_property,
+            mapper,
+            useobject=True,
+            callable_=self._load_for_state,
+            typecallable=self.parent_property.collection_class,
+            active_history=active_history,
+            _deferred_history=_deferred_history,
+        )
+
+    def _memoized_attr__simple_lazy_clause(self):
+        lazywhere = sql_util._deep_annotate(
+            self._lazywhere, {"_orm_adapt": True}
+        )
+
+        criterion, bind_to_col = (lazywhere, self._bind_to_col)
+
+        params = []
+
+        def visit_bindparam(bindparam):
+            bindparam.unique = False
+
+        visitors.traverse(criterion, {}, {"bindparam": visit_bindparam})
+
+        def visit_bindparam(bindparam):
+            if bindparam._identifying_key in bind_to_col:
+                params.append(
+                    (
+                        bindparam.key,
+                        bind_to_col[bindparam._identifying_key],
+                        None,
+                    )
+                )
+            elif bindparam.callable is None:
+                params.append((bindparam.key, None, bindparam.value))
+
+        criterion = visitors.cloned_traverse(
+            criterion, {}, {"bindparam": visit_bindparam}
+        )
+
+        return criterion, params
+
+    def _generate_lazy_clause(self, state, passive):
+        criterion, param_keys = self._simple_lazy_clause
+
+        if state is None:
+            return sql_util.adapt_criterion_to_null(
+                criterion, [key for key, ident, value in param_keys]
+            )
+
+        mapper = self.parent_property.parent
+
+        o = state.obj()  # strong ref
+        dict_ = attributes.instance_dict(o)
+
+        if passive & PassiveFlag.INIT_OK:
+            passive ^= PassiveFlag.INIT_OK
+
+        params = {}
+        for key, ident, value in param_keys:
+            if ident is not None:
+                if passive and passive & PassiveFlag.LOAD_AGAINST_COMMITTED:
+                    value = mapper._get_committed_state_attr_by_column(
+                        state, dict_, ident, passive
+                    )
+                else:
+                    value = mapper._get_state_attr_by_column(
+                        state, dict_, ident, passive
+                    )
+
+            params[key] = value
+
+        return criterion, params
+
+    def _invoke_raise_load(self, state, passive, lazy):
+        raise sa_exc.InvalidRequestError(
+            "'%s' is not available due to lazy='%s'" % (self, lazy)
+        )
+
+    def _load_for_state(
+        self,
+        state,
+        passive,
+        loadopt=None,
+        extra_criteria=(),
+        extra_options=(),
+        alternate_effective_path=None,
+        execution_options=util.EMPTY_DICT,
+    ):
+        if not state.key and (
+            (
+                not self.parent_property.load_on_pending
+                and not state._load_pending
+            )
+            or not state.session_id
+        ):
+            return LoaderCallableStatus.ATTR_EMPTY
+
+        pending = not state.key
+        primary_key_identity = None
+
+        use_get = self.use_get and (not loadopt or not loadopt._extra_criteria)
+
+        if (not passive & PassiveFlag.SQL_OK and not use_get) or (
+            not passive & attributes.NON_PERSISTENT_OK and pending
+        ):
+            return LoaderCallableStatus.PASSIVE_NO_RESULT
+
+        if (
+            # we were given lazy="raise"
+            self._raise_always
+            # the no_raise history-related flag was not passed
+            and not passive & PassiveFlag.NO_RAISE
+            and (
+                # if we are use_get and related_object_ok is disabled,
+                # which means we are at most looking in the identity map
+                # for history purposes or otherwise returning
+                # PASSIVE_NO_RESULT, don't raise.  This is also a
+                # history-related flag
+                not use_get
+                or passive & PassiveFlag.RELATED_OBJECT_OK
+            )
+        ):
+            self._invoke_raise_load(state, passive, "raise")
+
+        session = _state_session(state)
+        if not session:
+            if passive & PassiveFlag.NO_RAISE:
+                return LoaderCallableStatus.PASSIVE_NO_RESULT
+
+            raise orm_exc.DetachedInstanceError(
+                "Parent instance %s is not bound to a Session; "
+                "lazy load operation of attribute '%s' cannot proceed"
+                % (orm_util.state_str(state), self.key)
+            )
+
+        # if we have a simple primary key load, check the
+        # identity map without generating a Query at all
+        if use_get:
+            primary_key_identity = self._get_ident_for_use_get(
+                session, state, passive
+            )
+            if LoaderCallableStatus.PASSIVE_NO_RESULT in primary_key_identity:
+                return LoaderCallableStatus.PASSIVE_NO_RESULT
+            elif LoaderCallableStatus.NEVER_SET in primary_key_identity:
+                return LoaderCallableStatus.NEVER_SET
+
+            # test for None alone in primary_key_identity based on
+            # allow_partial_pks preference.   PASSIVE_NO_RESULT and NEVER_SET
+            # have already been tested above
+            if not self.mapper.allow_partial_pks:
+                if _none_only_set.intersection(primary_key_identity):
+                    return None
+            else:
+                if _none_only_set.issuperset(primary_key_identity):
+                    return None
+
+            if (
+                self.key in state.dict
+                and not passive & PassiveFlag.DEFERRED_HISTORY_LOAD
+            ):
+                return LoaderCallableStatus.ATTR_WAS_SET
+
+            # look for this identity in the identity map.  Delegate to the
+            # Query class in use, as it may have special rules for how it
+            # does this, including how it decides what the correct
+            # identity_token would be for this identity.
+
+            instance = session._identity_lookup(
+                self.entity,
+                primary_key_identity,
+                passive=passive,
+                lazy_loaded_from=state,
+            )
+
+            if instance is not None:
+                if instance is LoaderCallableStatus.PASSIVE_CLASS_MISMATCH:
+                    return None
+                else:
+                    return instance
+            elif (
+                not passive & PassiveFlag.SQL_OK
+                or not passive & PassiveFlag.RELATED_OBJECT_OK
+            ):
+                return LoaderCallableStatus.PASSIVE_NO_RESULT
+
+        return self._emit_lazyload(
+            session,
+            state,
+            primary_key_identity,
+            passive,
+            loadopt,
+            extra_criteria,
+            extra_options,
+            alternate_effective_path,
+            execution_options,
+        )
+
+    def _get_ident_for_use_get(self, session, state, passive):
+        instance_mapper = state.manager.mapper
+
+        if passive & PassiveFlag.LOAD_AGAINST_COMMITTED:
+            get_attr = instance_mapper._get_committed_state_attr_by_column
+        else:
+            get_attr = instance_mapper._get_state_attr_by_column
+
+        dict_ = state.dict
+
+        return [
+            get_attr(state, dict_, self._equated_columns[pk], passive=passive)
+            for pk in self.mapper.primary_key
+        ]
+
+    @util.preload_module("sqlalchemy.orm.strategy_options")
+    def _emit_lazyload(
+        self,
+        session,
+        state,
+        primary_key_identity,
+        passive,
+        loadopt,
+        extra_criteria,
+        extra_options,
+        alternate_effective_path,
+        execution_options,
+    ):
+        strategy_options = util.preloaded.orm_strategy_options
+
+        clauseelement = self.entity.__clause_element__()
+        stmt = Select._create_raw_select(
+            _raw_columns=[clauseelement],
+            _propagate_attrs=clauseelement._propagate_attrs,
+            _label_style=LABEL_STYLE_TABLENAME_PLUS_COL,
+            _compile_options=ORMCompileState.default_compile_options,
+        )
+        load_options = QueryContext.default_load_options
+
+        load_options += {
+            "_invoke_all_eagers": False,
+            "_lazy_loaded_from": state,
+        }
+
+        if self.parent_property.secondary is not None:
+            stmt = stmt.select_from(
+                self.mapper, self.parent_property.secondary
+            )
+
+        pending = not state.key
+
+        # don't autoflush on pending
+        if pending or passive & attributes.NO_AUTOFLUSH:
+            stmt._execution_options = util.immutabledict({"autoflush": False})
+
+        use_get = self.use_get
+
+        if state.load_options or (loadopt and loadopt._extra_criteria):
+            if alternate_effective_path is None:
+                effective_path = state.load_path[self.parent_property]
+            else:
+                effective_path = alternate_effective_path[self.parent_property]
+
+            opts = state.load_options
+
+            if loadopt and loadopt._extra_criteria:
+                use_get = False
+                opts += (
+                    orm_util.LoaderCriteriaOption(self.entity, extra_criteria),
+                )
+
+            stmt._with_options = opts
+        elif alternate_effective_path is None:
+            # this path is used if there are not already any options
+            # in the query, but an event may want to add them
+            effective_path = state.mapper._path_registry[self.parent_property]
+        else:
+            # added by immediateloader
+            effective_path = alternate_effective_path[self.parent_property]
+
+        if extra_options:
+            stmt._with_options += extra_options
+
+        stmt._compile_options += {"_current_path": effective_path}
+
+        if use_get:
+            if self._raise_on_sql and not passive & PassiveFlag.NO_RAISE:
+                self._invoke_raise_load(state, passive, "raise_on_sql")
+
+            return loading.load_on_pk_identity(
+                session,
+                stmt,
+                primary_key_identity,
+                load_options=load_options,
+                execution_options=execution_options,
+            )
+
+        if self._order_by:
+            stmt._order_by_clauses = self._order_by
+
+        def _lazyload_reverse(compile_context):
+            for rev in self.parent_property._reverse_property:
+                # reverse props that are MANYTOONE are loading *this*
+                # object from get(), so don't need to eager out to those.
+                if (
+                    rev.direction is interfaces.MANYTOONE
+                    and rev._use_get
+                    and not isinstance(rev.strategy, LazyLoader)
+                ):
+                    strategy_options.Load._construct_for_existing_path(
+                        compile_context.compile_options._current_path[
+                            rev.parent
+                        ]
+                    ).lazyload(rev).process_compile_state(compile_context)
+
+        stmt._with_context_options += (
+            (_lazyload_reverse, self.parent_property),
+        )
+
+        lazy_clause, params = self._generate_lazy_clause(state, passive)
+
+        if execution_options:
+            execution_options = util.EMPTY_DICT.merge_with(
+                execution_options,
+                {
+                    "_sa_orm_load_options": load_options,
+                },
+            )
+        else:
+            execution_options = {
+                "_sa_orm_load_options": load_options,
+            }
+
+        if (
+            self.key in state.dict
+            and not passive & PassiveFlag.DEFERRED_HISTORY_LOAD
+        ):
+            return LoaderCallableStatus.ATTR_WAS_SET
+
+        if pending:
+            if util.has_intersection(orm_util._none_set, params.values()):
+                return None
+
+        elif util.has_intersection(orm_util._never_set, params.values()):
+            return None
+
+        if self._raise_on_sql and not passive & PassiveFlag.NO_RAISE:
+            self._invoke_raise_load(state, passive, "raise_on_sql")
+
+        stmt._where_criteria = (lazy_clause,)
+
+        result = session.execute(
+            stmt, params, execution_options=execution_options
+        )
+
+        result = result.unique().scalars().all()
+
+        if self.uselist:
+            return result
+        else:
+            l = len(result)
+            if l:
+                if l > 1:
+                    util.warn(
+                        "Multiple rows returned with "
+                        "uselist=False for lazily-loaded attribute '%s' "
+                        % self.parent_property
+                    )
+
+                return result[0]
+            else:
+                return None
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        key = self.key
+
+        if (
+            context.load_options._is_user_refresh
+            and context.query._compile_options._only_load_props
+            and self.key in context.query._compile_options._only_load_props
+        ):
+            return self._immediateload_create_row_processor(
+                context,
+                query_entity,
+                path,
+                loadopt,
+                mapper,
+                result,
+                adapter,
+                populators,
+            )
+
+        if not self.is_class_level or (loadopt and loadopt._extra_criteria):
+            # we are not the primary manager for this attribute
+            # on this class - set up a
+            # per-instance lazyloader, which will override the
+            # class-level behavior.
+            # this currently only happens when using a
+            # "lazyload" option on a "no load"
+            # attribute - "eager" attributes always have a
+            # class-level lazyloader installed.
+            set_lazy_callable = (
+                InstanceState._instance_level_callable_processor
+            )(
+                mapper.class_manager,
+                LoadLazyAttribute(
+                    key,
+                    self,
+                    loadopt,
+                    (
+                        loadopt._generate_extra_criteria(context)
+                        if loadopt._extra_criteria
+                        else None
+                    ),
+                ),
+                key,
+            )
+
+            populators["new"].append((self.key, set_lazy_callable))
+        elif context.populate_existing or mapper.always_refresh:
+
+            def reset_for_lazy_callable(state, dict_, row):
+                # we are the primary manager for this attribute on
+                # this class - reset its
+                # per-instance attribute state, so that the class-level
+                # lazy loader is
+                # executed when next referenced on this instance.
+                # this is needed in
+                # populate_existing() types of scenarios to reset
+                # any existing state.
+                state._reset(dict_, key)
+
+            populators["new"].append((self.key, reset_for_lazy_callable))
+
+
+class LoadLazyAttribute:
+    """semi-serializable loader object used by LazyLoader
+
+    Historically, this object would be carried along with instances that
+    needed to run lazyloaders, so it had to be serializable to support
+    cached instances.
+
+    this is no longer a general requirement, and the case where this object
+    is used is exactly the case where we can't really serialize easily,
+    which is when extra criteria in the loader option is present.
+
+    We can't reliably serialize that as it refers to mapped entities and
+    AliasedClass objects that are local to the current process, which would
+    need to be matched up on deserialize e.g. the sqlalchemy.ext.serializer
+    approach.
+
+    """
+
+    def __init__(self, key, initiating_strategy, loadopt, extra_criteria):
+        self.key = key
+        self.strategy_key = initiating_strategy.strategy_key
+        self.loadopt = loadopt
+        self.extra_criteria = extra_criteria
+
+    def __getstate__(self):
+        if self.extra_criteria is not None:
+            util.warn(
+                "Can't reliably serialize a lazyload() option that "
+                "contains additional criteria; please use eager loading "
+                "for this case"
+            )
+        return {
+            "key": self.key,
+            "strategy_key": self.strategy_key,
+            "loadopt": self.loadopt,
+            "extra_criteria": (),
+        }
+
+    def __call__(self, state, passive=attributes.PASSIVE_OFF):
+        key = self.key
+        instance_mapper = state.manager.mapper
+        prop = instance_mapper._props[key]
+        strategy = prop._strategies[self.strategy_key]
+
+        return strategy._load_for_state(
+            state,
+            passive,
+            loadopt=self.loadopt,
+            extra_criteria=self.extra_criteria,
+        )
+
+
+class PostLoader(AbstractRelationshipLoader):
+    """A relationship loader that emits a second SELECT statement."""
+
+    __slots__ = ()
+
+    def _setup_for_recursion(self, context, path, loadopt, join_depth=None):
+        effective_path = (
+            context.compile_state.current_path or orm_util.PathRegistry.root
+        ) + path
+
+        top_level_context = context._get_top_level_context()
+        execution_options = util.immutabledict(
+            {"sa_top_level_orm_context": top_level_context}
+        )
+
+        if loadopt:
+            recursion_depth = loadopt.local_opts.get("recursion_depth", None)
+            unlimited_recursion = recursion_depth == -1
+        else:
+            recursion_depth = None
+            unlimited_recursion = False
+
+        if recursion_depth is not None:
+            if not self.parent_property._is_self_referential:
+                raise sa_exc.InvalidRequestError(
+                    f"recursion_depth option on relationship "
+                    f"{self.parent_property} not valid for "
+                    "non-self-referential relationship"
+                )
+            recursion_depth = context.execution_options.get(
+                f"_recursion_depth_{id(self)}", recursion_depth
+            )
+
+            if not unlimited_recursion and recursion_depth < 0:
+                return (
+                    effective_path,
+                    False,
+                    execution_options,
+                    recursion_depth,
+                )
+
+            if not unlimited_recursion:
+                execution_options = execution_options.union(
+                    {
+                        f"_recursion_depth_{id(self)}": recursion_depth - 1,
+                    }
+                )
+
+        if loading.PostLoad.path_exists(
+            context, effective_path, self.parent_property
+        ):
+            return effective_path, False, execution_options, recursion_depth
+
+        path_w_prop = path[self.parent_property]
+        effective_path_w_prop = effective_path[self.parent_property]
+
+        if not path_w_prop.contains(context.attributes, "loader"):
+            if join_depth:
+                if effective_path_w_prop.length / 2 > join_depth:
+                    return (
+                        effective_path,
+                        False,
+                        execution_options,
+                        recursion_depth,
+                    )
+            elif effective_path_w_prop.contains_mapper(self.mapper):
+                return (
+                    effective_path,
+                    False,
+                    execution_options,
+                    recursion_depth,
+                )
+
+        return effective_path, True, execution_options, recursion_depth
+
+
+@relationships.RelationshipProperty.strategy_for(lazy="immediate")
+class ImmediateLoader(PostLoader):
+    __slots__ = ("join_depth",)
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.join_depth = self.parent_property.join_depth
+
+    def init_class_attribute(self, mapper):
+        self.parent_property._get_strategy(
+            (("lazy", "select"),)
+        ).init_class_attribute(mapper)
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        if not context.compile_state.compile_options._enable_eagerloads:
+            return
+
+        (
+            effective_path,
+            run_loader,
+            execution_options,
+            recursion_depth,
+        ) = self._setup_for_recursion(context, path, loadopt, self.join_depth)
+
+        if not run_loader:
+            # this will not emit SQL and will only emit for a many-to-one
+            # "use get" load.   the "_RELATED" part means it may return
+            # instance even if its expired, since this is a mutually-recursive
+            # load operation.
+            flags = attributes.PASSIVE_NO_FETCH_RELATED | PassiveFlag.NO_RAISE
+        else:
+            flags = attributes.PASSIVE_OFF | PassiveFlag.NO_RAISE
+
+        loading.PostLoad.callable_for_path(
+            context,
+            effective_path,
+            self.parent,
+            self.parent_property,
+            self._load_for_path,
+            loadopt,
+            flags,
+            recursion_depth,
+            execution_options,
+        )
+
+    def _load_for_path(
+        self,
+        context,
+        path,
+        states,
+        load_only,
+        loadopt,
+        flags,
+        recursion_depth,
+        execution_options,
+    ):
+        if recursion_depth:
+            new_opt = Load(loadopt.path.entity)
+            new_opt.context = (
+                loadopt,
+                loadopt._recurse(),
+            )
+            alternate_effective_path = path._truncate_recursive()
+            extra_options = (new_opt,)
+        else:
+            new_opt = None
+            alternate_effective_path = path
+            extra_options = ()
+
+        key = self.key
+        lazyloader = self.parent_property._get_strategy((("lazy", "select"),))
+        for state, overwrite in states:
+            dict_ = state.dict
+
+            if overwrite or key not in dict_:
+                value = lazyloader._load_for_state(
+                    state,
+                    flags,
+                    extra_options=extra_options,
+                    alternate_effective_path=alternate_effective_path,
+                    execution_options=execution_options,
+                )
+                if value not in (
+                    ATTR_WAS_SET,
+                    LoaderCallableStatus.PASSIVE_NO_RESULT,
+                ):
+                    state.get_impl(key).set_committed_value(
+                        state, dict_, value
+                    )
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy="subquery")
+class SubqueryLoader(PostLoader):
+    __slots__ = ("join_depth",)
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.join_depth = self.parent_property.join_depth
+
+    def init_class_attribute(self, mapper):
+        self.parent_property._get_strategy(
+            (("lazy", "select"),)
+        ).init_class_attribute(mapper)
+
+    def _get_leftmost(
+        self,
+        orig_query_entity_index,
+        subq_path,
+        current_compile_state,
+        is_root,
+    ):
+        given_subq_path = subq_path
+        subq_path = subq_path.path
+        subq_mapper = orm_util._class_to_mapper(subq_path[0])
+
+        # determine attributes of the leftmost mapper
+        if (
+            self.parent.isa(subq_mapper)
+            and self.parent_property is subq_path[1]
+        ):
+            leftmost_mapper, leftmost_prop = self.parent, self.parent_property
+        else:
+            leftmost_mapper, leftmost_prop = subq_mapper, subq_path[1]
+
+        if is_root:
+            # the subq_path is also coming from cached state, so when we start
+            # building up this path, it has to also be converted to be in terms
+            # of the current state. this is for the specific case of the entity
+            # is an AliasedClass against a subquery that's not otherwise going
+            # to adapt
+            new_subq_path = current_compile_state._entities[
+                orig_query_entity_index
+            ].entity_zero._path_registry[leftmost_prop]
+            additional = len(subq_path) - len(new_subq_path)
+            if additional:
+                new_subq_path += path_registry.PathRegistry.coerce(
+                    subq_path[-additional:]
+                )
+        else:
+            new_subq_path = given_subq_path
+
+        leftmost_cols = leftmost_prop.local_columns
+
+        leftmost_attr = [
+            getattr(
+                new_subq_path.path[0].entity,
+                leftmost_mapper._columntoproperty[c].key,
+            )
+            for c in leftmost_cols
+        ]
+
+        return leftmost_mapper, leftmost_attr, leftmost_prop, new_subq_path
+
+    def _generate_from_original_query(
+        self,
+        orig_compile_state,
+        orig_query,
+        leftmost_mapper,
+        leftmost_attr,
+        leftmost_relationship,
+        orig_entity,
+    ):
+        # reformat the original query
+        # to look only for significant columns
+        q = orig_query._clone().correlate(None)
+
+        # LEGACY: make a Query back from the select() !!
+        # This suits at least two legacy cases:
+        # 1. applications which expect before_compile() to be called
+        #    below when we run .subquery() on this query (Keystone)
+        # 2. applications which are doing subqueryload with complex
+        #    from_self() queries, as query.subquery() / .statement
+        #    has to do the full compile context for multiply-nested
+        #    from_self() (Neutron) - see test_subqload_from_self
+        #    for demo.
+        q2 = query.Query.__new__(query.Query)
+        q2.__dict__.update(q.__dict__)
+        q = q2
+
+        # set the query's "FROM" list explicitly to what the
+        # FROM list would be in any case, as we will be limiting
+        # the columns in the SELECT list which may no longer include
+        # all entities mentioned in things like WHERE, JOIN, etc.
+        if not q._from_obj:
+            q._enable_assertions = False
+            q.select_from.non_generative(
+                q,
+                *{
+                    ent["entity"]
+                    for ent in _column_descriptions(
+                        orig_query, compile_state=orig_compile_state
+                    )
+                    if ent["entity"] is not None
+                },
+            )
+
+        # select from the identity columns of the outer (specifically, these
+        # are the 'local_cols' of the property).  This will remove other
+        # columns from the query that might suggest the right entity which is
+        # why we do set select_from above.   The attributes we have are
+        # coerced and adapted using the original query's adapter, which is
+        # needed only for the case of adapting a subclass column to
+        # that of a polymorphic selectable, e.g. we have
+        # Engineer.primary_language and the entity is Person.  All other
+        # adaptations, e.g. from_self, select_entity_from(), will occur
+        # within the new query when it compiles, as the compile_state we are
+        # using here is only a partial one.  If the subqueryload is from a
+        # with_polymorphic() or other aliased() object, left_attr will already
+        # be the correct attributes so no adaptation is needed.
+        target_cols = orig_compile_state._adapt_col_list(
+            [
+                sql.coercions.expect(sql.roles.ColumnsClauseRole, o)
+                for o in leftmost_attr
+            ],
+            orig_compile_state._get_current_adapter(),
+        )
+        q._raw_columns = target_cols
+
+        distinct_target_key = leftmost_relationship.distinct_target_key
+
+        if distinct_target_key is True:
+            q._distinct = True
+        elif distinct_target_key is None:
+            # if target_cols refer to a non-primary key or only
+            # part of a composite primary key, set the q as distinct
+            for t in {c.table for c in target_cols}:
+                if not set(target_cols).issuperset(t.primary_key):
+                    q._distinct = True
+                    break
+
+        # don't need ORDER BY if no limit/offset
+        if not q._has_row_limiting_clause:
+            q._order_by_clauses = ()
+
+        if q._distinct is True and q._order_by_clauses:
+            # the logic to automatically add the order by columns to the query
+            # when distinct is True is deprecated in the query
+            to_add = sql_util.expand_column_list_from_order_by(
+                target_cols, q._order_by_clauses
+            )
+            if to_add:
+                q._set_entities(target_cols + to_add)
+
+        # the original query now becomes a subquery
+        # which we'll join onto.
+        # LEGACY: as "q" is a Query, the before_compile() event is invoked
+        # here.
+        embed_q = q.set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL).subquery()
+        left_alias = orm_util.AliasedClass(
+            leftmost_mapper, embed_q, use_mapper_path=True
+        )
+        return left_alias
+
+    def _prep_for_joins(self, left_alias, subq_path):
+        # figure out what's being joined.  a.k.a. the fun part
+        to_join = []
+        pairs = list(subq_path.pairs())
+
+        for i, (mapper, prop) in enumerate(pairs):
+            if i > 0:
+                # look at the previous mapper in the chain -
+                # if it is as or more specific than this prop's
+                # mapper, use that instead.
+                # note we have an assumption here that
+                # the non-first element is always going to be a mapper,
+                # not an AliasedClass
+
+                prev_mapper = pairs[i - 1][1].mapper
+                to_append = prev_mapper if prev_mapper.isa(mapper) else mapper
+            else:
+                to_append = mapper
+
+            to_join.append((to_append, prop.key))
+
+        # determine the immediate parent class we are joining from,
+        # which needs to be aliased.
+
+        if len(to_join) < 2:
+            # in the case of a one level eager load, this is the
+            # leftmost "left_alias".
+            parent_alias = left_alias
+        else:
+            info = inspect(to_join[-1][0])
+            if info.is_aliased_class:
+                parent_alias = info.entity
+            else:
+                # alias a plain mapper as we may be
+                # joining multiple times
+                parent_alias = orm_util.AliasedClass(
+                    info.entity, use_mapper_path=True
+                )
+
+        local_cols = self.parent_property.local_columns
+
+        local_attr = [
+            getattr(parent_alias, self.parent._columntoproperty[c].key)
+            for c in local_cols
+        ]
+        return to_join, local_attr, parent_alias
+
+    def _apply_joins(
+        self, q, to_join, left_alias, parent_alias, effective_entity
+    ):
+        ltj = len(to_join)
+        if ltj == 1:
+            to_join = [
+                getattr(left_alias, to_join[0][1]).of_type(effective_entity)
+            ]
+        elif ltj == 2:
+            to_join = [
+                getattr(left_alias, to_join[0][1]).of_type(parent_alias),
+                getattr(parent_alias, to_join[-1][1]).of_type(
+                    effective_entity
+                ),
+            ]
+        elif ltj > 2:
+            middle = [
+                (
+                    (
+                        orm_util.AliasedClass(item[0])
+                        if not inspect(item[0]).is_aliased_class
+                        else item[0].entity
+                    ),
+                    item[1],
+                )
+                for item in to_join[1:-1]
+            ]
+            inner = []
+
+            while middle:
+                item = middle.pop(0)
+                attr = getattr(item[0], item[1])
+                if middle:
+                    attr = attr.of_type(middle[0][0])
+                else:
+                    attr = attr.of_type(parent_alias)
+
+                inner.append(attr)
+
+            to_join = (
+                [getattr(left_alias, to_join[0][1]).of_type(inner[0].parent)]
+                + inner
+                + [
+                    getattr(parent_alias, to_join[-1][1]).of_type(
+                        effective_entity
+                    )
+                ]
+            )
+
+        for attr in to_join:
+            q = q.join(attr)
+
+        return q
+
+    def _setup_options(
+        self,
+        context,
+        q,
+        subq_path,
+        rewritten_path,
+        orig_query,
+        effective_entity,
+        loadopt,
+    ):
+        # note that because the subqueryload object
+        # does not re-use the cached query, instead always making
+        # use of the current invoked query, while we have two queries
+        # here (orig and context.query), they are both non-cached
+        # queries and we can transfer the options as is without
+        # adjusting for new criteria.   Some work on #6881 / #6889
+        # brought this into question.
+        new_options = orig_query._with_options
+
+        if loadopt and loadopt._extra_criteria:
+            new_options += (
+                orm_util.LoaderCriteriaOption(
+                    self.entity,
+                    loadopt._generate_extra_criteria(context),
+                ),
+            )
+
+        # propagate loader options etc. to the new query.
+        # these will fire relative to subq_path.
+        q = q._with_current_path(rewritten_path)
+        q = q.options(*new_options)
+
+        return q
+
+    def _setup_outermost_orderby(self, q):
+        if self.parent_property.order_by:
+
+            def _setup_outermost_orderby(compile_context):
+                compile_context.eager_order_by += tuple(
+                    util.to_list(self.parent_property.order_by)
+                )
+
+            q = q._add_context_option(
+                _setup_outermost_orderby, self.parent_property
+            )
+
+        return q
+
+    class _SubqCollections:
+        """Given a :class:`_query.Query` used to emit the "subquery load",
+        provide a load interface that executes the query at the
+        first moment a value is needed.
+
+        """
+
+        __slots__ = (
+            "session",
+            "execution_options",
+            "load_options",
+            "params",
+            "subq",
+            "_data",
+        )
+
+        def __init__(self, context, subq):
+            # avoid creating a cycle by storing context
+            # even though that's preferable
+            self.session = context.session
+            self.execution_options = context.execution_options
+            self.load_options = context.load_options
+            self.params = context.params or {}
+            self.subq = subq
+            self._data = None
+
+        def get(self, key, default):
+            if self._data is None:
+                self._load()
+            return self._data.get(key, default)
+
+        def _load(self):
+            self._data = collections.defaultdict(list)
+
+            q = self.subq
+            assert q.session is None
+
+            q = q.with_session(self.session)
+
+            if self.load_options._populate_existing:
+                q = q.populate_existing()
+            # to work with baked query, the parameters may have been
+            # updated since this query was created, so take these into account
+
+            rows = list(q.params(self.params))
+            for k, v in itertools.groupby(rows, lambda x: x[1:]):
+                self._data[k].extend(vv[0] for vv in v)
+
+        def loader(self, state, dict_, row):
+            if self._data is None:
+                self._load()
+
+    def _setup_query_from_rowproc(
+        self,
+        context,
+        query_entity,
+        path,
+        entity,
+        loadopt,
+        adapter,
+    ):
+        compile_state = context.compile_state
+        if (
+            not compile_state.compile_options._enable_eagerloads
+            or compile_state.compile_options._for_refresh_state
+        ):
+            return
+
+        orig_query_entity_index = compile_state._entities.index(query_entity)
+        context.loaders_require_buffering = True
+
+        path = path[self.parent_property]
+
+        # build up a path indicating the path from the leftmost
+        # entity to the thing we're subquery loading.
+        with_poly_entity = path.get(
+            compile_state.attributes, "path_with_polymorphic", None
+        )
+        if with_poly_entity is not None:
+            effective_entity = with_poly_entity
+        else:
+            effective_entity = self.entity
+
+        subq_path, rewritten_path = context.query._execution_options.get(
+            ("subquery_paths", None),
+            (orm_util.PathRegistry.root, orm_util.PathRegistry.root),
+        )
+        is_root = subq_path is orm_util.PathRegistry.root
+        subq_path = subq_path + path
+        rewritten_path = rewritten_path + path
+
+        # use the current query being invoked, not the compile state
+        # one.  this is so that we get the current parameters.  however,
+        # it means we can't use the existing compile state, we have to make
+        # a new one.    other approaches include possibly using the
+        # compiled query but swapping the params, seems only marginally
+        # less time spent but more complicated
+        orig_query = context.query._execution_options.get(
+            ("orig_query", SubqueryLoader), context.query
+        )
+
+        # make a new compile_state for the query that's probably cached, but
+        # we're sort of undoing a bit of that caching :(
+        compile_state_cls = ORMCompileState._get_plugin_class_for_plugin(
+            orig_query, "orm"
+        )
+
+        if orig_query._is_lambda_element:
+            if context.load_options._lazy_loaded_from is None:
+                util.warn(
+                    'subqueryloader for "%s" must invoke lambda callable '
+                    "at %r in "
+                    "order to produce a new query, decreasing the efficiency "
+                    "of caching for this statement.  Consider using "
+                    "selectinload() for more effective full-lambda caching"
+                    % (self, orig_query)
+                )
+            orig_query = orig_query._resolved
+
+        # this is the more "quick" version, however it's not clear how
+        # much of this we need.    in particular I can't get a test to
+        # fail if the "set_base_alias" is missing and not sure why that is.
+        orig_compile_state = compile_state_cls._create_entities_collection(
+            orig_query, legacy=False
+        )
+
+        (
+            leftmost_mapper,
+            leftmost_attr,
+            leftmost_relationship,
+            rewritten_path,
+        ) = self._get_leftmost(
+            orig_query_entity_index,
+            rewritten_path,
+            orig_compile_state,
+            is_root,
+        )
+
+        # generate a new Query from the original, then
+        # produce a subquery from it.
+        left_alias = self._generate_from_original_query(
+            orig_compile_state,
+            orig_query,
+            leftmost_mapper,
+            leftmost_attr,
+            leftmost_relationship,
+            entity,
+        )
+
+        # generate another Query that will join the
+        # left alias to the target relationships.
+        # basically doing a longhand
+        # "from_self()".  (from_self() itself not quite industrial
+        # strength enough for all contingencies...but very close)
+
+        q = query.Query(effective_entity)
+
+        q._execution_options = context.query._execution_options.merge_with(
+            context.execution_options,
+            {
+                ("orig_query", SubqueryLoader): orig_query,
+                ("subquery_paths", None): (subq_path, rewritten_path),
+            },
+        )
+
+        q = q._set_enable_single_crit(False)
+        to_join, local_attr, parent_alias = self._prep_for_joins(
+            left_alias, subq_path
+        )
+
+        q = q.add_columns(*local_attr)
+        q = self._apply_joins(
+            q, to_join, left_alias, parent_alias, effective_entity
+        )
+
+        q = self._setup_options(
+            context,
+            q,
+            subq_path,
+            rewritten_path,
+            orig_query,
+            effective_entity,
+            loadopt,
+        )
+        q = self._setup_outermost_orderby(q)
+
+        return q
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        if (
+            loadopt
+            and context.compile_state.statement is not None
+            and context.compile_state.statement.is_dml
+        ):
+            util.warn_deprecated(
+                "The subqueryload loader option is not compatible with DML "
+                "statements such as INSERT, UPDATE.  Only SELECT may be used."
+                "This warning will become an exception in a future release.",
+                "2.0",
+            )
+
+        if context.refresh_state:
+            return self._immediateload_create_row_processor(
+                context,
+                query_entity,
+                path,
+                loadopt,
+                mapper,
+                result,
+                adapter,
+                populators,
+            )
+
+        _, run_loader, _, _ = self._setup_for_recursion(
+            context, path, loadopt, self.join_depth
+        )
+        if not run_loader:
+            return
+
+        if not isinstance(context.compile_state, ORMSelectCompileState):
+            # issue 7505 - subqueryload() in 1.3 and previous would silently
+            # degrade for from_statement() without warning. this behavior
+            # is restored here
+            return
+
+        if not self.parent.class_manager[self.key].impl.supports_population:
+            raise sa_exc.InvalidRequestError(
+                "'%s' does not support object "
+                "population - eager loading cannot be applied." % self
+            )
+
+        # a little dance here as the "path" is still something that only
+        # semi-tracks the exact series of things we are loading, still not
+        # telling us about with_polymorphic() and stuff like that when it's at
+        # the root..  the initial MapperEntity is more accurate for this case.
+        if len(path) == 1:
+            if not orm_util._entity_isa(query_entity.entity_zero, self.parent):
+                return
+        elif not orm_util._entity_isa(path[-1], self.parent):
+            return
+
+        subq = self._setup_query_from_rowproc(
+            context,
+            query_entity,
+            path,
+            path[-1],
+            loadopt,
+            adapter,
+        )
+
+        if subq is None:
+            return
+
+        assert subq.session is None
+
+        path = path[self.parent_property]
+
+        local_cols = self.parent_property.local_columns
+
+        # cache the loaded collections in the context
+        # so that inheriting mappers don't re-load when they
+        # call upon create_row_processor again
+        collections = path.get(context.attributes, "collections")
+        if collections is None:
+            collections = self._SubqCollections(context, subq)
+            path.set(context.attributes, "collections", collections)
+
+        if adapter:
+            local_cols = [adapter.columns[c] for c in local_cols]
+
+        if self.uselist:
+            self._create_collection_loader(
+                context, result, collections, local_cols, populators
+            )
+        else:
+            self._create_scalar_loader(
+                context, result, collections, local_cols, populators
+            )
+
+    def _create_collection_loader(
+        self, context, result, collections, local_cols, populators
+    ):
+        tuple_getter = result._tuple_getter(local_cols)
+
+        def load_collection_from_subq(state, dict_, row):
+            collection = collections.get(tuple_getter(row), ())
+            state.get_impl(self.key).set_committed_value(
+                state, dict_, collection
+            )
+
+        def load_collection_from_subq_existing_row(state, dict_, row):
+            if self.key not in dict_:
+                load_collection_from_subq(state, dict_, row)
+
+        populators["new"].append((self.key, load_collection_from_subq))
+        populators["existing"].append(
+            (self.key, load_collection_from_subq_existing_row)
+        )
+
+        if context.invoke_all_eagers:
+            populators["eager"].append((self.key, collections.loader))
+
+    def _create_scalar_loader(
+        self, context, result, collections, local_cols, populators
+    ):
+        tuple_getter = result._tuple_getter(local_cols)
+
+        def load_scalar_from_subq(state, dict_, row):
+            collection = collections.get(tuple_getter(row), (None,))
+            if len(collection) > 1:
+                util.warn(
+                    "Multiple rows returned with "
+                    "uselist=False for eagerly-loaded attribute '%s' " % self
+                )
+
+            scalar = collection[0]
+            state.get_impl(self.key).set_committed_value(state, dict_, scalar)
+
+        def load_scalar_from_subq_existing_row(state, dict_, row):
+            if self.key not in dict_:
+                load_scalar_from_subq(state, dict_, row)
+
+        populators["new"].append((self.key, load_scalar_from_subq))
+        populators["existing"].append(
+            (self.key, load_scalar_from_subq_existing_row)
+        )
+        if context.invoke_all_eagers:
+            populators["eager"].append((self.key, collections.loader))
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy="joined")
+@relationships.RelationshipProperty.strategy_for(lazy=False)
+class JoinedLoader(AbstractRelationshipLoader):
+    """Provide loading behavior for a :class:`.Relationship`
+    using joined eager loading.
+
+    """
+
+    __slots__ = "join_depth"
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.join_depth = self.parent_property.join_depth
+
+    def init_class_attribute(self, mapper):
+        self.parent_property._get_strategy(
+            (("lazy", "select"),)
+        ).init_class_attribute(mapper)
+
+    def setup_query(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection=None,
+        parentmapper=None,
+        chained_from_outerjoin=False,
+        **kwargs,
+    ):
+        """Add a left outer join to the statement that's being constructed."""
+
+        if not compile_state.compile_options._enable_eagerloads:
+            return
+        elif (
+            loadopt
+            and compile_state.statement is not None
+            and compile_state.statement.is_dml
+        ):
+            util.warn_deprecated(
+                "The joinedload loader option is not compatible with DML "
+                "statements such as INSERT, UPDATE.  Only SELECT may be used."
+                "This warning will become an exception in a future release.",
+                "2.0",
+            )
+        elif self.uselist:
+            compile_state.multi_row_eager_loaders = True
+
+        path = path[self.parent_property]
+
+        with_polymorphic = None
+
+        user_defined_adapter = (
+            self._init_user_defined_eager_proc(
+                loadopt, compile_state, compile_state.attributes
+            )
+            if loadopt
+            else False
+        )
+
+        if user_defined_adapter is not False:
+            # setup an adapter but dont create any JOIN, assume it's already
+            # in the query
+            (
+                clauses,
+                adapter,
+                add_to_collection,
+            ) = self._setup_query_on_user_defined_adapter(
+                compile_state,
+                query_entity,
+                path,
+                adapter,
+                user_defined_adapter,
+            )
+
+            # don't do "wrap" for multi-row, we want to wrap
+            # limited/distinct SELECT,
+            # because we want to put the JOIN on the outside.
+
+        else:
+            # if not via query option, check for
+            # a cycle
+            if not path.contains(compile_state.attributes, "loader"):
+                if self.join_depth:
+                    if path.length / 2 > self.join_depth:
+                        return
+                elif path.contains_mapper(self.mapper):
+                    return
+
+            # add the JOIN and create an adapter
+            (
+                clauses,
+                adapter,
+                add_to_collection,
+                chained_from_outerjoin,
+            ) = self._generate_row_adapter(
+                compile_state,
+                query_entity,
+                path,
+                loadopt,
+                adapter,
+                column_collection,
+                parentmapper,
+                chained_from_outerjoin,
+            )
+
+            # for multi-row, we want to wrap limited/distinct SELECT,
+            # because we want to put the JOIN on the outside.
+            compile_state.eager_adding_joins = True
+
+        with_poly_entity = path.get(
+            compile_state.attributes, "path_with_polymorphic", None
+        )
+        if with_poly_entity is not None:
+            with_polymorphic = inspect(
+                with_poly_entity
+            ).with_polymorphic_mappers
+        else:
+            with_polymorphic = None
+
+        path = path[self.entity]
+
+        loading._setup_entity_query(
+            compile_state,
+            self.mapper,
+            query_entity,
+            path,
+            clauses,
+            add_to_collection,
+            with_polymorphic=with_polymorphic,
+            parentmapper=self.mapper,
+            chained_from_outerjoin=chained_from_outerjoin,
+        )
+
+        has_nones = util.NONE_SET.intersection(compile_state.secondary_columns)
+
+        if has_nones:
+            if with_poly_entity is not None:
+                raise sa_exc.InvalidRequestError(
+                    "Detected unaliased columns when generating joined "
+                    "load.  Make sure to use aliased=True or flat=True "
+                    "when using joined loading with with_polymorphic()."
+                )
+            else:
+                compile_state.secondary_columns = [
+                    c for c in compile_state.secondary_columns if c is not None
+                ]
+
+    def _init_user_defined_eager_proc(
+        self, loadopt, compile_state, target_attributes
+    ):
+        # check if the opt applies at all
+        if "eager_from_alias" not in loadopt.local_opts:
+            # nope
+            return False
+
+        path = loadopt.path.parent
+
+        # the option applies.  check if the "user_defined_eager_row_processor"
+        # has been built up.
+        adapter = path.get(
+            compile_state.attributes, "user_defined_eager_row_processor", False
+        )
+        if adapter is not False:
+            # just return it
+            return adapter
+
+        # otherwise figure it out.
+        alias = loadopt.local_opts["eager_from_alias"]
+        root_mapper, prop = path[-2:]
+
+        if alias is not None:
+            if isinstance(alias, str):
+                alias = prop.target.alias(alias)
+            adapter = orm_util.ORMAdapter(
+                orm_util._TraceAdaptRole.JOINEDLOAD_USER_DEFINED_ALIAS,
+                prop.mapper,
+                selectable=alias,
+                equivalents=prop.mapper._equivalent_columns,
+                limit_on_entity=False,
+            )
+        else:
+            if path.contains(
+                compile_state.attributes, "path_with_polymorphic"
+            ):
+                with_poly_entity = path.get(
+                    compile_state.attributes, "path_with_polymorphic"
+                )
+                adapter = orm_util.ORMAdapter(
+                    orm_util._TraceAdaptRole.JOINEDLOAD_PATH_WITH_POLYMORPHIC,
+                    with_poly_entity,
+                    equivalents=prop.mapper._equivalent_columns,
+                )
+            else:
+                adapter = compile_state._polymorphic_adapters.get(
+                    prop.mapper, None
+                )
+        path.set(
+            target_attributes,
+            "user_defined_eager_row_processor",
+            adapter,
+        )
+
+        return adapter
+
+    def _setup_query_on_user_defined_adapter(
+        self, context, entity, path, adapter, user_defined_adapter
+    ):
+        # apply some more wrapping to the "user defined adapter"
+        # if we are setting up the query for SQL render.
+        adapter = entity._get_entity_clauses(context)
+
+        if adapter and user_defined_adapter:
+            user_defined_adapter = user_defined_adapter.wrap(adapter)
+            path.set(
+                context.attributes,
+                "user_defined_eager_row_processor",
+                user_defined_adapter,
+            )
+        elif adapter:
+            user_defined_adapter = adapter
+            path.set(
+                context.attributes,
+                "user_defined_eager_row_processor",
+                user_defined_adapter,
+            )
+
+        add_to_collection = context.primary_columns
+        return user_defined_adapter, adapter, add_to_collection
+
+    def _generate_row_adapter(
+        self,
+        compile_state,
+        entity,
+        path,
+        loadopt,
+        adapter,
+        column_collection,
+        parentmapper,
+        chained_from_outerjoin,
+    ):
+        with_poly_entity = path.get(
+            compile_state.attributes, "path_with_polymorphic", None
+        )
+        if with_poly_entity:
+            to_adapt = with_poly_entity
+        else:
+            insp = inspect(self.entity)
+            if insp.is_aliased_class:
+                alt_selectable = insp.selectable
+            else:
+                alt_selectable = None
+
+            to_adapt = orm_util.AliasedClass(
+                self.mapper,
+                alias=(
+                    alt_selectable._anonymous_fromclause(flat=True)
+                    if alt_selectable is not None
+                    else None
+                ),
+                flat=True,
+                use_mapper_path=True,
+            )
+
+        to_adapt_insp = inspect(to_adapt)
+
+        clauses = to_adapt_insp._memo(
+            ("joinedloader_ormadapter", self),
+            orm_util.ORMAdapter,
+            orm_util._TraceAdaptRole.JOINEDLOAD_MEMOIZED_ADAPTER,
+            to_adapt_insp,
+            equivalents=self.mapper._equivalent_columns,
+            adapt_required=True,
+            allow_label_resolve=False,
+            anonymize_labels=True,
+        )
+
+        assert clauses.is_aliased_class
+
+        innerjoin = (
+            loadopt.local_opts.get("innerjoin", self.parent_property.innerjoin)
+            if loadopt is not None
+            else self.parent_property.innerjoin
+        )
+
+        if not innerjoin:
+            # if this is an outer join, all non-nested eager joins from
+            # this path must also be outer joins
+            chained_from_outerjoin = True
+
+        compile_state.create_eager_joins.append(
+            (
+                self._create_eager_join,
+                entity,
+                path,
+                adapter,
+                parentmapper,
+                clauses,
+                innerjoin,
+                chained_from_outerjoin,
+                loadopt._extra_criteria if loadopt else (),
+            )
+        )
+
+        add_to_collection = compile_state.secondary_columns
+        path.set(compile_state.attributes, "eager_row_processor", clauses)
+
+        return clauses, adapter, add_to_collection, chained_from_outerjoin
+
+    def _create_eager_join(
+        self,
+        compile_state,
+        query_entity,
+        path,
+        adapter,
+        parentmapper,
+        clauses,
+        innerjoin,
+        chained_from_outerjoin,
+        extra_criteria,
+    ):
+        if parentmapper is None:
+            localparent = query_entity.mapper
+        else:
+            localparent = parentmapper
+
+        # whether or not the Query will wrap the selectable in a subquery,
+        # and then attach eager load joins to that (i.e., in the case of
+        # LIMIT/OFFSET etc.)
+        should_nest_selectable = (
+            compile_state.multi_row_eager_loaders
+            and compile_state._should_nest_selectable
+        )
+
+        query_entity_key = None
+
+        if (
+            query_entity not in compile_state.eager_joins
+            and not should_nest_selectable
+            and compile_state.from_clauses
+        ):
+            indexes = sql_util.find_left_clause_that_matches_given(
+                compile_state.from_clauses, query_entity.selectable
+            )
+
+            if len(indexes) > 1:
+                # for the eager load case, I can't reproduce this right
+                # now.   For query.join() I can.
+                raise sa_exc.InvalidRequestError(
+                    "Can't identify which query entity in which to joined "
+                    "eager load from.   Please use an exact match when "
+                    "specifying the join path."
+                )
+
+            if indexes:
+                clause = compile_state.from_clauses[indexes[0]]
+                # join to an existing FROM clause on the query.
+                # key it to its list index in the eager_joins dict.
+                # Query._compile_context will adapt as needed and
+                # append to the FROM clause of the select().
+                query_entity_key, default_towrap = indexes[0], clause
+
+        if query_entity_key is None:
+            query_entity_key, default_towrap = (
+                query_entity,
+                query_entity.selectable,
+            )
+
+        towrap = compile_state.eager_joins.setdefault(
+            query_entity_key, default_towrap
+        )
+
+        if adapter:
+            if getattr(adapter, "is_aliased_class", False):
+                # joining from an adapted entity.  The adapted entity
+                # might be a "with_polymorphic", so resolve that to our
+                # specific mapper's entity before looking for our attribute
+                # name on it.
+                efm = adapter.aliased_insp._entity_for_mapper(
+                    localparent
+                    if localparent.isa(self.parent)
+                    else self.parent
+                )
+
+                # look for our attribute on the adapted entity, else fall back
+                # to our straight property
+                onclause = getattr(efm.entity, self.key, self.parent_property)
+            else:
+                onclause = getattr(
+                    orm_util.AliasedClass(
+                        self.parent, adapter.selectable, use_mapper_path=True
+                    ),
+                    self.key,
+                    self.parent_property,
+                )
+
+        else:
+            onclause = self.parent_property
+
+        assert clauses.is_aliased_class
+
+        attach_on_outside = (
+            not chained_from_outerjoin
+            or not innerjoin
+            or innerjoin == "unnested"
+            or query_entity.entity_zero.represents_outer_join
+        )
+
+        extra_join_criteria = extra_criteria
+        additional_entity_criteria = compile_state.global_attributes.get(
+            ("additional_entity_criteria", self.mapper), ()
+        )
+        if additional_entity_criteria:
+            extra_join_criteria += tuple(
+                ae._resolve_where_criteria(self.mapper)
+                for ae in additional_entity_criteria
+                if ae.propagate_to_loaders
+            )
+
+        if attach_on_outside:
+            # this is the "classic" eager join case.
+            eagerjoin = orm_util._ORMJoin(
+                towrap,
+                clauses.aliased_insp,
+                onclause,
+                isouter=not innerjoin
+                or query_entity.entity_zero.represents_outer_join
+                or (chained_from_outerjoin and isinstance(towrap, sql.Join)),
+                _left_memo=self.parent,
+                _right_memo=path[self.mapper],
+                _extra_criteria=extra_join_criteria,
+            )
+        else:
+            # all other cases are innerjoin=='nested' approach
+            eagerjoin = self._splice_nested_inner_join(
+                path, path[-2], towrap, clauses, onclause, extra_join_criteria
+            )
+
+        compile_state.eager_joins[query_entity_key] = eagerjoin
+
+        # send a hint to the Query as to where it may "splice" this join
+        eagerjoin.stop_on = query_entity.selectable
+
+        if not parentmapper:
+            # for parentclause that is the non-eager end of the join,
+            # ensure all the parent cols in the primaryjoin are actually
+            # in the
+            # columns clause (i.e. are not deferred), so that aliasing applied
+            # by the Query propagates those columns outward.
+            # This has the effect
+            # of "undefering" those columns.
+            for col in sql_util._find_columns(
+                self.parent_property.primaryjoin
+            ):
+                if localparent.persist_selectable.c.contains_column(col):
+                    if adapter:
+                        col = adapter.columns[col]
+                    compile_state._append_dedupe_col_collection(
+                        col, compile_state.primary_columns
+                    )
+
+        if self.parent_property.order_by:
+            compile_state.eager_order_by += tuple(
+                (eagerjoin._target_adapter.copy_and_process)(
+                    util.to_list(self.parent_property.order_by)
+                )
+            )
+
+    def _splice_nested_inner_join(
+        self,
+        path,
+        entity_we_want_to_splice_onto,
+        join_obj,
+        clauses,
+        onclause,
+        extra_criteria,
+        entity_inside_join_structure: Union[
+            Mapper, None, Literal[False]
+        ] = False,
+        detected_existing_path: Optional[path_registry.PathRegistry] = None,
+    ):
+        # recursive fn to splice a nested join into an existing one.
+        # entity_inside_join_structure=False means this is the outermost call,
+        # and it should return a value.  entity_inside_join_structure=<mapper>
+        # indicates we've descended into a join and are looking at a FROM
+        # clause representing this mapper; if this is not
+        # entity_we_want_to_splice_onto then return None to end the recursive
+        # branch
+
+        assert entity_we_want_to_splice_onto is path[-2]
+
+        if entity_inside_join_structure is False:
+            assert isinstance(join_obj, orm_util._ORMJoin)
+
+        if isinstance(join_obj, sql.selectable.FromGrouping):
+            # FromGrouping - continue descending into the structure
+            return self._splice_nested_inner_join(
+                path,
+                entity_we_want_to_splice_onto,
+                join_obj.element,
+                clauses,
+                onclause,
+                extra_criteria,
+                entity_inside_join_structure,
+            )
+        elif isinstance(join_obj, orm_util._ORMJoin):
+            # _ORMJoin - continue descending into the structure
+
+            join_right_path = join_obj._right_memo
+
+            # see if right side of join is viable
+            target_join = self._splice_nested_inner_join(
+                path,
+                entity_we_want_to_splice_onto,
+                join_obj.right,
+                clauses,
+                onclause,
+                extra_criteria,
+                entity_inside_join_structure=(
+                    join_right_path[-1].mapper
+                    if join_right_path is not None
+                    else None
+                ),
+            )
+
+            if target_join is not None:
+                # for a right splice, attempt to flatten out
+                # a JOIN b JOIN c JOIN .. to avoid needless
+                # parenthesis nesting
+                if not join_obj.isouter and not target_join.isouter:
+                    eagerjoin = join_obj._splice_into_center(target_join)
+                else:
+                    eagerjoin = orm_util._ORMJoin(
+                        join_obj.left,
+                        target_join,
+                        join_obj.onclause,
+                        isouter=join_obj.isouter,
+                        _left_memo=join_obj._left_memo,
+                    )
+
+                eagerjoin._target_adapter = target_join._target_adapter
+                return eagerjoin
+
+            else:
+                # see if left side of join is viable
+                target_join = self._splice_nested_inner_join(
+                    path,
+                    entity_we_want_to_splice_onto,
+                    join_obj.left,
+                    clauses,
+                    onclause,
+                    extra_criteria,
+                    entity_inside_join_structure=join_obj._left_memo,
+                    detected_existing_path=join_right_path,
+                )
+
+                if target_join is not None:
+                    eagerjoin = orm_util._ORMJoin(
+                        target_join,
+                        join_obj.right,
+                        join_obj.onclause,
+                        isouter=join_obj.isouter,
+                        _right_memo=join_obj._right_memo,
+                    )
+                    eagerjoin._target_adapter = target_join._target_adapter
+                    return eagerjoin
+
+            # neither side viable, return None, or fail if this was the top
+            # most call
+            if entity_inside_join_structure is False:
+                assert (
+                    False
+                ), "assertion failed attempting to produce joined eager loads"
+            return None
+
+        # reached an endpoint (e.g. a table that's mapped, or an alias of that
+        # table).  determine if we can use this endpoint to splice onto
+
+        # is this the entity we want to splice onto in the first place?
+        if not entity_we_want_to_splice_onto.isa(entity_inside_join_structure):
+            return None
+
+        # path check.  if we know the path how this join endpoint got here,
+        # lets look at our path we are satisfying and see if we're in the
+        # wrong place.  This is specifically for when our entity may
+        # appear more than once in the path, issue #11449
+        # updated in issue #11965.
+        if detected_existing_path and len(detected_existing_path) > 2:
+            # this assertion is currently based on how this call is made,
+            # where given a join_obj, the call will have these parameters as
+            # entity_inside_join_structure=join_obj._left_memo
+            # and entity_inside_join_structure=join_obj._right_memo.mapper
+            assert detected_existing_path[-3] is entity_inside_join_structure
+
+            # from that, see if the path we are targeting matches the
+            # "existing" path of this join all the way up to the midpoint
+            # of this join object (e.g. the relationship).
+            # if not, then this is not our target
+            #
+            # a test condition where this test is false looks like:
+            #
+            # desired splice:         Node->kind->Kind
+            # path of desired splice: NodeGroup->nodes->Node->kind
+            # path we've located:     NodeGroup->nodes->Node->common_node->Node
+            #
+            # above, because we want to splice kind->Kind onto
+            # NodeGroup->nodes->Node, this is not our path because it actually
+            # goes more steps than we want into self-referential
+            # ->common_node->Node
+            #
+            # a test condition where this test is true looks like:
+            #
+            # desired splice:         B->c2s->C2
+            # path of desired splice: A->bs->B->c2s
+            # path we've located:     A->bs->B->c1s->C1
+            #
+            # above, we want to splice c2s->C2 onto B, and the located path
+            # shows that the join ends with B->c1s->C1.  so we will
+            # add another join onto that, which would create a "branch" that
+            # we might represent in a pseudopath as:
+            #
+            # B->c1s->C1
+            #  ->c2s->C2
+            #
+            # i.e. A JOIN B ON <bs> JOIN C1 ON <c1s>
+            #                       JOIN C2 ON <c2s>
+            #
+
+            if detected_existing_path[0:-2] != path.path[0:-1]:
+                return None
+
+        return orm_util._ORMJoin(
+            join_obj,
+            clauses.aliased_insp,
+            onclause,
+            isouter=False,
+            _left_memo=entity_inside_join_structure,
+            _right_memo=path[path[-1].mapper],
+            _extra_criteria=extra_criteria,
+        )
+
+    def _create_eager_adapter(self, context, result, adapter, path, loadopt):
+        compile_state = context.compile_state
+
+        user_defined_adapter = (
+            self._init_user_defined_eager_proc(
+                loadopt, compile_state, context.attributes
+            )
+            if loadopt
+            else False
+        )
+
+        if user_defined_adapter is not False:
+            decorator = user_defined_adapter
+            # user defined eagerloads are part of the "primary"
+            # portion of the load.
+            # the adapters applied to the Query should be honored.
+            if compile_state.compound_eager_adapter and decorator:
+                decorator = decorator.wrap(
+                    compile_state.compound_eager_adapter
+                )
+            elif compile_state.compound_eager_adapter:
+                decorator = compile_state.compound_eager_adapter
+        else:
+            decorator = path.get(
+                compile_state.attributes, "eager_row_processor"
+            )
+            if decorator is None:
+                return False
+
+        if self.mapper._result_has_identity_key(result, decorator):
+            return decorator
+        else:
+            # no identity key - don't return a row
+            # processor, will cause a degrade to lazy
+            return False
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+
+        if not context.compile_state.compile_options._enable_eagerloads:
+            return
+
+        if not self.parent.class_manager[self.key].impl.supports_population:
+            raise sa_exc.InvalidRequestError(
+                "'%s' does not support object "
+                "population - eager loading cannot be applied." % self
+            )
+
+        if self.uselist:
+            context.loaders_require_uniquing = True
+
+        our_path = path[self.parent_property]
+
+        eager_adapter = self._create_eager_adapter(
+            context, result, adapter, our_path, loadopt
+        )
+
+        if eager_adapter is not False:
+            key = self.key
+
+            _instance = loading._instance_processor(
+                query_entity,
+                self.mapper,
+                context,
+                result,
+                our_path[self.entity],
+                eager_adapter,
+            )
+
+            if not self.uselist:
+                self._create_scalar_loader(context, key, _instance, populators)
+            else:
+                self._create_collection_loader(
+                    context, key, _instance, populators
+                )
+        else:
+            self.parent_property._get_strategy(
+                (("lazy", "select"),)
+            ).create_row_processor(
+                context,
+                query_entity,
+                path,
+                loadopt,
+                mapper,
+                result,
+                adapter,
+                populators,
+            )
+
+    def _create_collection_loader(self, context, key, _instance, populators):
+        def load_collection_from_joined_new_row(state, dict_, row):
+            # note this must unconditionally clear out any existing collection.
+            # an existing collection would be present only in the case of
+            # populate_existing().
+            collection = attributes.init_state_collection(state, dict_, key)
+            result_list = util.UniqueAppender(
+                collection, "append_without_event"
+            )
+            context.attributes[(state, key)] = result_list
+            inst = _instance(row)
+            if inst is not None:
+                result_list.append(inst)
+
+        def load_collection_from_joined_existing_row(state, dict_, row):
+            if (state, key) in context.attributes:
+                result_list = context.attributes[(state, key)]
+            else:
+                # appender_key can be absent from context.attributes
+                # with isnew=False when self-referential eager loading
+                # is used; the same instance may be present in two
+                # distinct sets of result columns
+                collection = attributes.init_state_collection(
+                    state, dict_, key
+                )
+                result_list = util.UniqueAppender(
+                    collection, "append_without_event"
+                )
+                context.attributes[(state, key)] = result_list
+            inst = _instance(row)
+            if inst is not None:
+                result_list.append(inst)
+
+        def load_collection_from_joined_exec(state, dict_, row):
+            _instance(row)
+
+        populators["new"].append(
+            (self.key, load_collection_from_joined_new_row)
+        )
+        populators["existing"].append(
+            (self.key, load_collection_from_joined_existing_row)
+        )
+        if context.invoke_all_eagers:
+            populators["eager"].append(
+                (self.key, load_collection_from_joined_exec)
+            )
+
+    def _create_scalar_loader(self, context, key, _instance, populators):
+        def load_scalar_from_joined_new_row(state, dict_, row):
+            # set a scalar object instance directly on the parent
+            # object, bypassing InstrumentedAttribute event handlers.
+            dict_[key] = _instance(row)
+
+        def load_scalar_from_joined_existing_row(state, dict_, row):
+            # call _instance on the row, even though the object has
+            # been created, so that we further descend into properties
+            existing = _instance(row)
+
+            # conflicting value already loaded, this shouldn't happen
+            if key in dict_:
+                if existing is not dict_[key]:
+                    util.warn(
+                        "Multiple rows returned with "
+                        "uselist=False for eagerly-loaded attribute '%s' "
+                        % self
+                    )
+            else:
+                # this case is when one row has multiple loads of the
+                # same entity (e.g. via aliasing), one has an attribute
+                # that the other doesn't.
+                dict_[key] = existing
+
+        def load_scalar_from_joined_exec(state, dict_, row):
+            _instance(row)
+
+        populators["new"].append((self.key, load_scalar_from_joined_new_row))
+        populators["existing"].append(
+            (self.key, load_scalar_from_joined_existing_row)
+        )
+        if context.invoke_all_eagers:
+            populators["eager"].append(
+                (self.key, load_scalar_from_joined_exec)
+            )
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy="selectin")
+class SelectInLoader(PostLoader, util.MemoizedSlots):
+    __slots__ = (
+        "join_depth",
+        "omit_join",
+        "_parent_alias",
+        "_query_info",
+        "_fallback_query_info",
+    )
+
+    query_info = collections.namedtuple(
+        "queryinfo",
+        [
+            "load_only_child",
+            "load_with_join",
+            "in_expr",
+            "pk_cols",
+            "zero_idx",
+            "child_lookup_cols",
+        ],
+    )
+
+    _chunksize = 500
+
+    def __init__(self, parent, strategy_key):
+        super().__init__(parent, strategy_key)
+        self.join_depth = self.parent_property.join_depth
+        is_m2o = self.parent_property.direction is interfaces.MANYTOONE
+
+        if self.parent_property.omit_join is not None:
+            self.omit_join = self.parent_property.omit_join
+        else:
+            lazyloader = self.parent_property._get_strategy(
+                (("lazy", "select"),)
+            )
+            if is_m2o:
+                self.omit_join = lazyloader.use_get
+            else:
+                self.omit_join = self.parent._get_clause[0].compare(
+                    lazyloader._rev_lazywhere,
+                    use_proxies=True,
+                    compare_keys=False,
+                    equivalents=self.parent._equivalent_columns,
+                )
+
+        if self.omit_join:
+            if is_m2o:
+                self._query_info = self._init_for_omit_join_m2o()
+                self._fallback_query_info = self._init_for_join()
+            else:
+                self._query_info = self._init_for_omit_join()
+        else:
+            self._query_info = self._init_for_join()
+
+    def _init_for_omit_join(self):
+        pk_to_fk = dict(
+            self.parent_property._join_condition.local_remote_pairs
+        )
+        pk_to_fk.update(
+            (equiv, pk_to_fk[k])
+            for k in list(pk_to_fk)
+            for equiv in self.parent._equivalent_columns.get(k, ())
+        )
+
+        pk_cols = fk_cols = [
+            pk_to_fk[col] for col in self.parent.primary_key if col in pk_to_fk
+        ]
+        if len(fk_cols) > 1:
+            in_expr = sql.tuple_(*fk_cols)
+            zero_idx = False
+        else:
+            in_expr = fk_cols[0]
+            zero_idx = True
+
+        return self.query_info(False, False, in_expr, pk_cols, zero_idx, None)
+
+    def _init_for_omit_join_m2o(self):
+        pk_cols = self.mapper.primary_key
+        if len(pk_cols) > 1:
+            in_expr = sql.tuple_(*pk_cols)
+            zero_idx = False
+        else:
+            in_expr = pk_cols[0]
+            zero_idx = True
+
+        lazyloader = self.parent_property._get_strategy((("lazy", "select"),))
+        lookup_cols = [lazyloader._equated_columns[pk] for pk in pk_cols]
+
+        return self.query_info(
+            True, False, in_expr, pk_cols, zero_idx, lookup_cols
+        )
+
+    def _init_for_join(self):
+        self._parent_alias = AliasedClass(self.parent.class_)
+        pa_insp = inspect(self._parent_alias)
+        pk_cols = [
+            pa_insp._adapt_element(col) for col in self.parent.primary_key
+        ]
+        if len(pk_cols) > 1:
+            in_expr = sql.tuple_(*pk_cols)
+            zero_idx = False
+        else:
+            in_expr = pk_cols[0]
+            zero_idx = True
+        return self.query_info(False, True, in_expr, pk_cols, zero_idx, None)
+
+    def init_class_attribute(self, mapper):
+        self.parent_property._get_strategy(
+            (("lazy", "select"),)
+        ).init_class_attribute(mapper)
+
+    def create_row_processor(
+        self,
+        context,
+        query_entity,
+        path,
+        loadopt,
+        mapper,
+        result,
+        adapter,
+        populators,
+    ):
+        if context.refresh_state:
+            return self._immediateload_create_row_processor(
+                context,
+                query_entity,
+                path,
+                loadopt,
+                mapper,
+                result,
+                adapter,
+                populators,
+            )
+
+        (
+            effective_path,
+            run_loader,
+            execution_options,
+            recursion_depth,
+        ) = self._setup_for_recursion(
+            context, path, loadopt, join_depth=self.join_depth
+        )
+
+        if not run_loader:
+            return
+
+        if not context.compile_state.compile_options._enable_eagerloads:
+            return
+
+        if not self.parent.class_manager[self.key].impl.supports_population:
+            raise sa_exc.InvalidRequestError(
+                "'%s' does not support object "
+                "population - eager loading cannot be applied." % self
+            )
+
+        # a little dance here as the "path" is still something that only
+        # semi-tracks the exact series of things we are loading, still not
+        # telling us about with_polymorphic() and stuff like that when it's at
+        # the root..  the initial MapperEntity is more accurate for this case.
+        if len(path) == 1:
+            if not orm_util._entity_isa(query_entity.entity_zero, self.parent):
+                return
+        elif not orm_util._entity_isa(path[-1], self.parent):
+            return
+
+        selectin_path = effective_path
+
+        path_w_prop = path[self.parent_property]
+
+        # build up a path indicating the path from the leftmost
+        # entity to the thing we're subquery loading.
+        with_poly_entity = path_w_prop.get(
+            context.attributes, "path_with_polymorphic", None
+        )
+        if with_poly_entity is not None:
+            effective_entity = inspect(with_poly_entity)
+        else:
+            effective_entity = self.entity
+
+        loading.PostLoad.callable_for_path(
+            context,
+            selectin_path,
+            self.parent,
+            self.parent_property,
+            self._load_for_path,
+            effective_entity,
+            loadopt,
+            recursion_depth,
+            execution_options,
+        )
+
+    def _load_for_path(
+        self,
+        context,
+        path,
+        states,
+        load_only,
+        effective_entity,
+        loadopt,
+        recursion_depth,
+        execution_options,
+    ):
+        if load_only and self.key not in load_only:
+            return
+
+        query_info = self._query_info
+
+        if query_info.load_only_child:
+            our_states = collections.defaultdict(list)
+            none_states = []
+
+            mapper = self.parent
+
+            for state, overwrite in states:
+                state_dict = state.dict
+                related_ident = tuple(
+                    mapper._get_state_attr_by_column(
+                        state,
+                        state_dict,
+                        lk,
+                        passive=attributes.PASSIVE_NO_FETCH,
+                    )
+                    for lk in query_info.child_lookup_cols
+                )
+                # if the loaded parent objects do not have the foreign key
+                # to the related item loaded, then degrade into the joined
+                # version of selectinload
+                if LoaderCallableStatus.PASSIVE_NO_RESULT in related_ident:
+                    query_info = self._fallback_query_info
+                    break
+
+                # organize states into lists keyed to particular foreign
+                # key values.
+                if None not in related_ident:
+                    our_states[related_ident].append(
+                        (state, state_dict, overwrite)
+                    )
+                else:
+                    # For FK values that have None, add them to a
+                    # separate collection that will be populated separately
+                    none_states.append((state, state_dict, overwrite))
+
+        # note the above conditional may have changed query_info
+        if not query_info.load_only_child:
+            our_states = [
+                (state.key[1], state, state.dict, overwrite)
+                for state, overwrite in states
+            ]
+
+        pk_cols = query_info.pk_cols
+        in_expr = query_info.in_expr
+
+        if not query_info.load_with_join:
+            # in "omit join" mode, the primary key column and the
+            # "in" expression are in terms of the related entity.  So
+            # if the related entity is polymorphic or otherwise aliased,
+            # we need to adapt our "pk_cols" and "in_expr" to that
+            # entity.   in non-"omit join" mode, these are against the
+            # parent entity and do not need adaption.
+            if effective_entity.is_aliased_class:
+                pk_cols = [
+                    effective_entity._adapt_element(col) for col in pk_cols
+                ]
+                in_expr = effective_entity._adapt_element(in_expr)
+
+        bundle_ent = orm_util.Bundle("pk", *pk_cols)
+        bundle_sql = bundle_ent.__clause_element__()
+
+        entity_sql = effective_entity.__clause_element__()
+        q = Select._create_raw_select(
+            _raw_columns=[bundle_sql, entity_sql],
+            _label_style=LABEL_STYLE_TABLENAME_PLUS_COL,
+            _compile_options=ORMCompileState.default_compile_options,
+            _propagate_attrs={
+                "compile_state_plugin": "orm",
+                "plugin_subject": effective_entity,
+            },
+        )
+
+        if not query_info.load_with_join:
+            # the Bundle we have in the "omit_join" case is against raw, non
+            # annotated columns, so to ensure the Query knows its primary
+            # entity, we add it explicitly.  If we made the Bundle against
+            # annotated columns, we hit a performance issue in this specific
+            # case, which is detailed in issue #4347.
+            q = q.select_from(effective_entity)
+        else:
+            # in the non-omit_join case, the Bundle is against the annotated/
+            # mapped column of the parent entity, but the #4347 issue does not
+            # occur in this case.
+            q = q.select_from(self._parent_alias).join(
+                getattr(self._parent_alias, self.parent_property.key).of_type(
+                    effective_entity
+                )
+            )
+
+        q = q.filter(in_expr.in_(sql.bindparam("primary_keys")))
+
+        # a test which exercises what these comments talk about is
+        # test_selectin_relations.py -> test_twolevel_selectin_w_polymorphic
+        #
+        # effective_entity above is given to us in terms of the cached
+        # statement, namely this one:
+        orig_query = context.compile_state.select_statement
+
+        # the actual statement that was requested is this one:
+        #  context_query = context.user_passed_query
+        #
+        # that's not the cached one, however.  So while it is of the identical
+        # structure, if it has entities like AliasedInsp, which we get from
+        # aliased() or with_polymorphic(), the AliasedInsp will likely be a
+        # different object identity each time, and will not match up
+        # hashing-wise to the corresponding AliasedInsp that's in the
+        # cached query, meaning it won't match on paths and loader lookups
+        # and loaders like this one will be skipped if it is used in options.
+        #
+        # as it turns out, standard loader options like selectinload(),
+        # lazyload() that have a path need
+        # to come from the cached query so that the AliasedInsp etc. objects
+        # that are in the query line up with the object that's in the path
+        # of the strategy object. however other options like
+        # with_loader_criteria() that doesn't have a path (has a fixed entity)
+        # and needs to have access to the latest closure state in order to
+        # be correct, we need to use the uncached one.
+        #
+        # as of #8399 we let the loader option itself figure out what it
+        # wants to do given cached and uncached version of itself.
+
+        effective_path = path[self.parent_property]
+
+        if orig_query is context.user_passed_query:
+            new_options = orig_query._with_options
+        else:
+            cached_options = orig_query._with_options
+            uncached_options = context.user_passed_query._with_options
+
+            # propagate compile state options from the original query,
+            # updating their "extra_criteria" as necessary.
+            # note this will create a different cache key than
+            # "orig" options if extra_criteria is present, because the copy
+            # of extra_criteria will have different boundparam than that of
+            # the QueryableAttribute in the path
+            new_options = [
+                orig_opt._adapt_cached_option_to_uncached_option(
+                    context, uncached_opt
+                )
+                for orig_opt, uncached_opt in zip(
+                    cached_options, uncached_options
+                )
+            ]
+
+        if loadopt and loadopt._extra_criteria:
+            new_options += (
+                orm_util.LoaderCriteriaOption(
+                    effective_entity,
+                    loadopt._generate_extra_criteria(context),
+                ),
+            )
+
+        if recursion_depth is not None:
+            effective_path = effective_path._truncate_recursive()
+
+        q = q.options(*new_options)
+
+        q = q._update_compile_options({"_current_path": effective_path})
+        if context.populate_existing:
+            q = q.execution_options(populate_existing=True)
+
+        if self.parent_property.order_by:
+            if not query_info.load_with_join:
+                eager_order_by = self.parent_property.order_by
+                if effective_entity.is_aliased_class:
+                    eager_order_by = [
+                        effective_entity._adapt_element(elem)
+                        for elem in eager_order_by
+                    ]
+                q = q.order_by(*eager_order_by)
+            else:
+
+                def _setup_outermost_orderby(compile_context):
+                    compile_context.eager_order_by += tuple(
+                        util.to_list(self.parent_property.order_by)
+                    )
+
+                q = q._add_context_option(
+                    _setup_outermost_orderby, self.parent_property
+                )
+
+        if query_info.load_only_child:
+            self._load_via_child(
+                our_states,
+                none_states,
+                query_info,
+                q,
+                context,
+                execution_options,
+            )
+        else:
+            self._load_via_parent(
+                our_states, query_info, q, context, execution_options
+            )
+
+    def _load_via_child(
+        self,
+        our_states,
+        none_states,
+        query_info,
+        q,
+        context,
+        execution_options,
+    ):
+        uselist = self.uselist
+
+        # this sort is really for the benefit of the unit tests
+        our_keys = sorted(our_states)
+        while our_keys:
+            chunk = our_keys[0 : self._chunksize]
+            our_keys = our_keys[self._chunksize :]
+            data = {
+                k: v
+                for k, v in context.session.execute(
+                    q,
+                    params={
+                        "primary_keys": [
+                            key[0] if query_info.zero_idx else key
+                            for key in chunk
+                        ]
+                    },
+                    execution_options=execution_options,
+                ).unique()
+            }
+
+            for key in chunk:
+                # for a real foreign key and no concurrent changes to the
+                # DB while running this method, "key" is always present in
+                # data.  However, for primaryjoins without real foreign keys
+                # a non-None primaryjoin condition may still refer to no
+                # related object.
+                related_obj = data.get(key, None)
+                for state, dict_, overwrite in our_states[key]:
+                    if not overwrite and self.key in dict_:
+                        continue
+
+                    state.get_impl(self.key).set_committed_value(
+                        state,
+                        dict_,
+                        related_obj if not uselist else [related_obj],
+                    )
+        # populate none states with empty value / collection
+        for state, dict_, overwrite in none_states:
+            if not overwrite and self.key in dict_:
+                continue
+
+            # note it's OK if this is a uselist=True attribute, the empty
+            # collection will be populated
+            state.get_impl(self.key).set_committed_value(state, dict_, None)
+
+    def _load_via_parent(
+        self, our_states, query_info, q, context, execution_options
+    ):
+        uselist = self.uselist
+        _empty_result = () if uselist else None
+
+        while our_states:
+            chunk = our_states[0 : self._chunksize]
+            our_states = our_states[self._chunksize :]
+
+            primary_keys = [
+                key[0] if query_info.zero_idx else key
+                for key, state, state_dict, overwrite in chunk
+            ]
+
+            data = collections.defaultdict(list)
+            for k, v in itertools.groupby(
+                context.session.execute(
+                    q,
+                    params={"primary_keys": primary_keys},
+                    execution_options=execution_options,
+                ).unique(),
+                lambda x: x[0],
+            ):
+                data[k].extend(vv[1] for vv in v)
+
+            for key, state, state_dict, overwrite in chunk:
+                if not overwrite and self.key in state_dict:
+                    continue
+
+                collection = data.get(key, _empty_result)
+
+                if not uselist and collection:
+                    if len(collection) > 1:
+                        util.warn(
+                            "Multiple rows returned with "
+                            "uselist=False for eagerly-loaded "
+                            "attribute '%s' " % self
+                        )
+                    state.get_impl(self.key).set_committed_value(
+                        state, state_dict, collection[0]
+                    )
+                else:
+                    # note that empty tuple set on uselist=False sets the
+                    # value to None
+                    state.get_impl(self.key).set_committed_value(
+                        state, state_dict, collection
+                    )
+
+
+def single_parent_validator(desc, prop):
+    def _do_check(state, value, oldvalue, initiator):
+        if value is not None and initiator.key == prop.key:
+            hasparent = initiator.hasparent(attributes.instance_state(value))
+            if hasparent and oldvalue is not value:
+                raise sa_exc.InvalidRequestError(
+                    "Instance %s is already associated with an instance "
+                    "of %s via its %s attribute, and is only allowed a "
+                    "single parent."
+                    % (orm_util.instance_str(value), state.class_, prop),
+                    code="bbf1",
+                )
+        return value
+
+    def append(state, value, initiator):
+        return _do_check(state, value, None, initiator)
+
+    def set_(state, value, oldvalue, initiator):
+        return _do_check(state, value, oldvalue, initiator)
+
+    event.listen(
+        desc, "append", append, raw=True, retval=True, active_history=True
+    )
+    event.listen(desc, "set", set_, raw=True, retval=True, active_history=True)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategy_options.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategy_options.py
new file mode 100644
index 00000000..f4f292ee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/strategy_options.py
@@ -0,0 +1,2550 @@
+# orm/strategy_options.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""
+
+"""
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+from . import util as orm_util
+from ._typing import insp_is_aliased_class
+from ._typing import insp_is_attribute
+from ._typing import insp_is_mapper
+from ._typing import insp_is_mapper_property
+from .attributes import QueryableAttribute
+from .base import InspectionAttr
+from .interfaces import LoaderOption
+from .path_registry import _DEFAULT_TOKEN
+from .path_registry import _StrPathToken
+from .path_registry import _WILDCARD_TOKEN
+from .path_registry import AbstractEntityRegistry
+from .path_registry import path_is_property
+from .path_registry import PathRegistry
+from .path_registry import TokenRegistry
+from .util import _orm_full_deannotate
+from .util import AliasedInsp
+from .. import exc as sa_exc
+from .. import inspect
+from .. import util
+from ..sql import and_
+from ..sql import cache_key
+from ..sql import coercions
+from ..sql import roles
+from ..sql import traversals
+from ..sql import visitors
+from ..sql.base import _generative
+from ..util.typing import Final
+from ..util.typing import Literal
+from ..util.typing import Self
+
+_RELATIONSHIP_TOKEN: Final[Literal["relationship"]] = "relationship"
+_COLUMN_TOKEN: Final[Literal["column"]] = "column"
+
+_FN = TypeVar("_FN", bound="Callable[..., Any]")
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _InternalEntityType
+    from .context import _MapperEntity
+    from .context import ORMCompileState
+    from .context import QueryContext
+    from .interfaces import _StrategyKey
+    from .interfaces import MapperProperty
+    from .interfaces import ORMOption
+    from .mapper import Mapper
+    from .path_registry import _PathRepresentation
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _FromClauseArgument
+    from ..sql.cache_key import _CacheKeyTraversalType
+    from ..sql.cache_key import CacheKey
+
+
+_AttrType = Union[Literal["*"], "QueryableAttribute[Any]"]
+
+_WildcardKeyType = Literal["relationship", "column"]
+_StrategySpec = Dict[str, Any]
+_OptsType = Dict[str, Any]
+_AttrGroupType = Tuple[_AttrType, ...]
+
+
+class _AbstractLoad(traversals.GenerativeOnTraversal, LoaderOption):
+    __slots__ = ("propagate_to_loaders",)
+
+    _is_strategy_option = True
+    propagate_to_loaders: bool
+
+    def contains_eager(
+        self,
+        attr: _AttrType,
+        alias: Optional[_FromClauseArgument] = None,
+        _is_chain: bool = False,
+        _propagate_to_loaders: bool = False,
+    ) -> Self:
+        r"""Indicate that the given attribute should be eagerly loaded from
+        columns stated manually in the query.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        The option is used in conjunction with an explicit join that loads
+        the desired rows, i.e.::
+
+            sess.query(Order).join(Order.user).options(contains_eager(Order.user))
+
+        The above query would join from the ``Order`` entity to its related
+        ``User`` entity, and the returned ``Order`` objects would have the
+        ``Order.user`` attribute pre-populated.
+
+        It may also be used for customizing the entries in an eagerly loaded
+        collection; queries will normally want to use the
+        :ref:`orm_queryguide_populate_existing` execution option assuming the
+        primary collection of parent objects may already have been loaded::
+
+            sess.query(User).join(User.addresses).filter(
+                Address.email_address.like("%@aol.com")
+            ).options(contains_eager(User.addresses)).populate_existing()
+
+        See the section :ref:`contains_eager` for complete usage details.
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`contains_eager`
+
+        """
+        if alias is not None:
+            if not isinstance(alias, str):
+                coerced_alias = coercions.expect(roles.FromClauseRole, alias)
+            else:
+                util.warn_deprecated(
+                    "Passing a string name for the 'alias' argument to "
+                    "'contains_eager()` is deprecated, and will not work in a "
+                    "future release.  Please use a sqlalchemy.alias() or "
+                    "sqlalchemy.orm.aliased() construct.",
+                    version="1.4",
+                )
+                coerced_alias = alias
+
+        elif getattr(attr, "_of_type", None):
+            assert isinstance(attr, QueryableAttribute)
+            ot: Optional[_InternalEntityType[Any]] = inspect(attr._of_type)
+            assert ot is not None
+            coerced_alias = ot.selectable
+        else:
+            coerced_alias = None
+
+        cloned = self._set_relationship_strategy(
+            attr,
+            {"lazy": "joined"},
+            propagate_to_loaders=_propagate_to_loaders,
+            opts={"eager_from_alias": coerced_alias},
+            _reconcile_to_other=True if _is_chain else None,
+        )
+        return cloned
+
+    def load_only(self, *attrs: _AttrType, raiseload: bool = False) -> Self:
+        r"""Indicate that for a particular entity, only the given list
+        of column-based attribute names should be loaded; all others will be
+        deferred.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        Example - given a class ``User``, load only the ``name`` and
+        ``fullname`` attributes::
+
+            session.query(User).options(load_only(User.name, User.fullname))
+
+        Example - given a relationship ``User.addresses -> Address``, specify
+        subquery loading for the ``User.addresses`` collection, but on each
+        ``Address`` object load only the ``email_address`` attribute::
+
+            session.query(User).options(
+                subqueryload(User.addresses).load_only(Address.email_address)
+            )
+
+        For a statement that has multiple entities,
+        the lead entity can be
+        specifically referred to using the :class:`_orm.Load` constructor::
+
+            stmt = (
+                select(User, Address)
+                .join(User.addresses)
+                .options(
+                    Load(User).load_only(User.name, User.fullname),
+                    Load(Address).load_only(Address.email_address),
+                )
+            )
+
+        When used together with the
+        :ref:`populate_existing <orm_queryguide_populate_existing>`
+        execution option only the attributes listed will be refreshed.
+
+        :param \*attrs: Attributes to be loaded, all others will be deferred.
+
+        :param raiseload: raise :class:`.InvalidRequestError` rather than
+         lazy loading a value when a deferred attribute is accessed. Used
+         to prevent unwanted SQL from being emitted.
+
+         .. versionadded:: 2.0
+
+        .. seealso::
+
+            :ref:`orm_queryguide_column_deferral` - in the
+            :ref:`queryguide_toplevel`
+
+        :param \*attrs: Attributes to be loaded, all others will be deferred.
+
+        :param raiseload: raise :class:`.InvalidRequestError` rather than
+         lazy loading a value when a deferred attribute is accessed. Used
+         to prevent unwanted SQL from being emitted.
+
+         .. versionadded:: 2.0
+
+        """
+        cloned = self._set_column_strategy(
+            attrs,
+            {"deferred": False, "instrument": True},
+        )
+
+        wildcard_strategy = {"deferred": True, "instrument": True}
+        if raiseload:
+            wildcard_strategy["raiseload"] = True
+
+        cloned = cloned._set_column_strategy(
+            ("*",),
+            wildcard_strategy,
+        )
+        return cloned
+
+    def joinedload(
+        self,
+        attr: _AttrType,
+        innerjoin: Optional[bool] = None,
+    ) -> Self:
+        """Indicate that the given attribute should be loaded using joined
+        eager loading.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        examples::
+
+            # joined-load the "orders" collection on "User"
+            select(User).options(joinedload(User.orders))
+
+            # joined-load Order.items and then Item.keywords
+            select(Order).options(joinedload(Order.items).joinedload(Item.keywords))
+
+            # lazily load Order.items, but when Items are loaded,
+            # joined-load the keywords collection
+            select(Order).options(lazyload(Order.items).joinedload(Item.keywords))
+
+        :param innerjoin: if ``True``, indicates that the joined eager load
+         should use an inner join instead of the default of left outer join::
+
+            select(Order).options(joinedload(Order.user, innerjoin=True))
+
+        In order to chain multiple eager joins together where some may be
+        OUTER and others INNER, right-nested joins are used to link them::
+
+            select(A).options(
+                joinedload(A.bs, innerjoin=False).joinedload(B.cs, innerjoin=True)
+            )
+
+        The above query, linking A.bs via "outer" join and B.cs via "inner"
+        join would render the joins as "a LEFT OUTER JOIN (b JOIN c)". When
+        using older versions of SQLite (< 3.7.16), this form of JOIN is
+        translated to use full subqueries as this syntax is otherwise not
+        directly supported.
+
+        The ``innerjoin`` flag can also be stated with the term ``"unnested"``.
+        This indicates that an INNER JOIN should be used, *unless* the join
+        is linked to a LEFT OUTER JOIN to the left, in which case it
+        will render as LEFT OUTER JOIN.  For example, supposing ``A.bs``
+        is an outerjoin::
+
+            select(A).options(joinedload(A.bs).joinedload(B.cs, innerjoin="unnested"))
+
+        The above join will render as "a LEFT OUTER JOIN b LEFT OUTER JOIN c",
+        rather than as "a LEFT OUTER JOIN (b JOIN c)".
+
+        .. note:: The "unnested" flag does **not** affect the JOIN rendered
+            from a many-to-many association table, e.g. a table configured as
+            :paramref:`_orm.relationship.secondary`, to the target table; for
+            correctness of results, these joins are always INNER and are
+            therefore right-nested if linked to an OUTER join.
+
+        .. note::
+
+            The joins produced by :func:`_orm.joinedload` are **anonymously
+            aliased**. The criteria by which the join proceeds cannot be
+            modified, nor can the ORM-enabled :class:`_sql.Select` or legacy
+            :class:`_query.Query` refer to these joins in any way, including
+            ordering. See :ref:`zen_of_eager_loading` for further detail.
+
+            To produce a specific SQL JOIN which is explicitly available, use
+            :meth:`_sql.Select.join` and :meth:`_query.Query.join`. To combine
+            explicit JOINs with eager loading of collections, use
+            :func:`_orm.contains_eager`; see :ref:`contains_eager`.
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`joined_eager_loading`
+
+        """  # noqa: E501
+        loader = self._set_relationship_strategy(
+            attr,
+            {"lazy": "joined"},
+            opts=(
+                {"innerjoin": innerjoin}
+                if innerjoin is not None
+                else util.EMPTY_DICT
+            ),
+        )
+        return loader
+
+    def subqueryload(self, attr: _AttrType) -> Self:
+        """Indicate that the given attribute should be loaded using
+        subquery eager loading.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        examples::
+
+            # subquery-load the "orders" collection on "User"
+            select(User).options(subqueryload(User.orders))
+
+            # subquery-load Order.items and then Item.keywords
+            select(Order).options(
+                subqueryload(Order.items).subqueryload(Item.keywords)
+            )
+
+            # lazily load Order.items, but when Items are loaded,
+            # subquery-load the keywords collection
+            select(Order).options(lazyload(Order.items).subqueryload(Item.keywords))
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`subquery_eager_loading`
+
+        """
+        return self._set_relationship_strategy(attr, {"lazy": "subquery"})
+
+    def selectinload(
+        self,
+        attr: _AttrType,
+        recursion_depth: Optional[int] = None,
+    ) -> Self:
+        """Indicate that the given attribute should be loaded using
+        SELECT IN eager loading.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        examples::
+
+            # selectin-load the "orders" collection on "User"
+            select(User).options(selectinload(User.orders))
+
+            # selectin-load Order.items and then Item.keywords
+            select(Order).options(
+                selectinload(Order.items).selectinload(Item.keywords)
+            )
+
+            # lazily load Order.items, but when Items are loaded,
+            # selectin-load the keywords collection
+            select(Order).options(lazyload(Order.items).selectinload(Item.keywords))
+
+        :param recursion_depth: optional int; when set to a positive integer
+         in conjunction with a self-referential relationship,
+         indicates "selectin" loading will continue that many levels deep
+         automatically until no items are found.
+
+         .. note:: The :paramref:`_orm.selectinload.recursion_depth` option
+            currently supports only self-referential relationships.  There
+            is not yet an option to automatically traverse recursive structures
+            with more than one relationship involved.
+
+            Additionally, the :paramref:`_orm.selectinload.recursion_depth`
+            parameter is new and experimental and should be treated as "alpha"
+            status for the 2.0 series.
+
+         .. versionadded:: 2.0 added
+            :paramref:`_orm.selectinload.recursion_depth`
+
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`selectin_eager_loading`
+
+        """
+        return self._set_relationship_strategy(
+            attr,
+            {"lazy": "selectin"},
+            opts={"recursion_depth": recursion_depth},
+        )
+
+    def lazyload(self, attr: _AttrType) -> Self:
+        """Indicate that the given attribute should be loaded using "lazy"
+        loading.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`lazy_loading`
+
+        """
+        return self._set_relationship_strategy(attr, {"lazy": "select"})
+
+    def immediateload(
+        self,
+        attr: _AttrType,
+        recursion_depth: Optional[int] = None,
+    ) -> Self:
+        """Indicate that the given attribute should be loaded using
+        an immediate load with a per-attribute SELECT statement.
+
+        The load is achieved using the "lazyloader" strategy and does not
+        fire off any additional eager loaders.
+
+        The :func:`.immediateload` option is superseded in general
+        by the :func:`.selectinload` option, which performs the same task
+        more efficiently by emitting a SELECT for all loaded objects.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        :param recursion_depth: optional int; when set to a positive integer
+         in conjunction with a self-referential relationship,
+         indicates "selectin" loading will continue that many levels deep
+         automatically until no items are found.
+
+         .. note:: The :paramref:`_orm.immediateload.recursion_depth` option
+            currently supports only self-referential relationships.  There
+            is not yet an option to automatically traverse recursive structures
+            with more than one relationship involved.
+
+         .. warning:: This parameter is new and experimental and should be
+            treated as "alpha" status
+
+         .. versionadded:: 2.0 added
+            :paramref:`_orm.immediateload.recursion_depth`
+
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`selectin_eager_loading`
+
+        """
+        loader = self._set_relationship_strategy(
+            attr,
+            {"lazy": "immediate"},
+            opts={"recursion_depth": recursion_depth},
+        )
+        return loader
+
+    def noload(self, attr: _AttrType) -> Self:
+        """Indicate that the given relationship attribute should remain
+        unloaded.
+
+        The relationship attribute will return ``None`` when accessed without
+        producing any loading effect.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        :func:`_orm.noload` applies to :func:`_orm.relationship` attributes
+        only.
+
+        .. legacy:: The :func:`_orm.noload` option is **legacy**.  As it
+           forces collections to be empty, which invariably leads to
+           non-intuitive and difficult to predict results.  There are no
+           legitimate uses for this option in modern SQLAlchemy.
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+        """
+
+        return self._set_relationship_strategy(attr, {"lazy": "noload"})
+
+    def raiseload(self, attr: _AttrType, sql_only: bool = False) -> Self:
+        """Indicate that the given attribute should raise an error if accessed.
+
+        A relationship attribute configured with :func:`_orm.raiseload` will
+        raise an :exc:`~sqlalchemy.exc.InvalidRequestError` upon access. The
+        typical way this is useful is when an application is attempting to
+        ensure that all relationship attributes that are accessed in a
+        particular context would have been already loaded via eager loading.
+        Instead of having to read through SQL logs to ensure lazy loads aren't
+        occurring, this strategy will cause them to raise immediately.
+
+        :func:`_orm.raiseload` applies to :func:`_orm.relationship` attributes
+        only. In order to apply raise-on-SQL behavior to a column-based
+        attribute, use the :paramref:`.orm.defer.raiseload` parameter on the
+        :func:`.defer` loader option.
+
+        :param sql_only: if True, raise only if the lazy load would emit SQL,
+         but not if it is only checking the identity map, or determining that
+         the related value should just be None due to missing keys. When False,
+         the strategy will raise for all varieties of relationship loading.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        .. seealso::
+
+            :ref:`loading_toplevel`
+
+            :ref:`prevent_lazy_with_raiseload`
+
+            :ref:`orm_queryguide_deferred_raiseload`
+
+        """
+
+        return self._set_relationship_strategy(
+            attr, {"lazy": "raise_on_sql" if sql_only else "raise"}
+        )
+
+    def defaultload(self, attr: _AttrType) -> Self:
+        """Indicate an attribute should load using its predefined loader style.
+
+        The behavior of this loading option is to not change the current
+        loading style of the attribute, meaning that the previously configured
+        one is used or, if no previous style was selected, the default
+        loading will be used.
+
+        This method is used to link to other loader options further into
+        a chain of attributes without altering the loader style of the links
+        along the chain.  For example, to set joined eager loading for an
+        element of an element::
+
+            session.query(MyClass).options(
+                defaultload(MyClass.someattribute).joinedload(
+                    MyOtherClass.someotherattribute
+                )
+            )
+
+        :func:`.defaultload` is also useful for setting column-level options on
+        a related class, namely that of :func:`.defer` and :func:`.undefer`::
+
+            session.scalars(
+                select(MyClass).options(
+                    defaultload(MyClass.someattribute)
+                    .defer("some_column")
+                    .undefer("some_other_column")
+                )
+            )
+
+        .. seealso::
+
+            :ref:`orm_queryguide_relationship_sub_options`
+
+            :meth:`_orm.Load.options`
+
+        """
+        return self._set_relationship_strategy(attr, None)
+
+    def defer(self, key: _AttrType, raiseload: bool = False) -> Self:
+        r"""Indicate that the given column-oriented attribute should be
+        deferred, e.g. not loaded until accessed.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        e.g.::
+
+            from sqlalchemy.orm import defer
+
+            session.query(MyClass).options(
+                defer(MyClass.attribute_one), defer(MyClass.attribute_two)
+            )
+
+        To specify a deferred load of an attribute on a related class,
+        the path can be specified one token at a time, specifying the loading
+        style for each link along the chain.  To leave the loading style
+        for a link unchanged, use :func:`_orm.defaultload`::
+
+            session.query(MyClass).options(
+                defaultload(MyClass.someattr).defer(RelatedClass.some_column)
+            )
+
+        Multiple deferral options related to a relationship can be bundled
+        at once using :meth:`_orm.Load.options`::
+
+
+            select(MyClass).options(
+                defaultload(MyClass.someattr).options(
+                    defer(RelatedClass.some_column),
+                    defer(RelatedClass.some_other_column),
+                    defer(RelatedClass.another_column),
+                )
+            )
+
+        :param key: Attribute to be deferred.
+
+        :param raiseload: raise :class:`.InvalidRequestError` rather than
+         lazy loading a value when the deferred attribute is accessed. Used
+         to prevent unwanted SQL from being emitted.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :ref:`orm_queryguide_column_deferral` - in the
+            :ref:`queryguide_toplevel`
+
+            :func:`_orm.load_only`
+
+            :func:`_orm.undefer`
+
+        """
+        strategy = {"deferred": True, "instrument": True}
+        if raiseload:
+            strategy["raiseload"] = True
+        return self._set_column_strategy((key,), strategy)
+
+    def undefer(self, key: _AttrType) -> Self:
+        r"""Indicate that the given column-oriented attribute should be
+        undeferred, e.g. specified within the SELECT statement of the entity
+        as a whole.
+
+        The column being undeferred is typically set up on the mapping as a
+        :func:`.deferred` attribute.
+
+        This function is part of the :class:`_orm.Load` interface and supports
+        both method-chained and standalone operation.
+
+        Examples::
+
+            # undefer two columns
+            session.query(MyClass).options(
+                undefer(MyClass.col1), undefer(MyClass.col2)
+            )
+
+            # undefer all columns specific to a single class using Load + *
+            session.query(MyClass, MyOtherClass).options(Load(MyClass).undefer("*"))
+
+            # undefer a column on a related object
+            select(MyClass).options(defaultload(MyClass.items).undefer(MyClass.text))
+
+        :param key: Attribute to be undeferred.
+
+        .. seealso::
+
+            :ref:`orm_queryguide_column_deferral` - in the
+            :ref:`queryguide_toplevel`
+
+            :func:`_orm.defer`
+
+            :func:`_orm.undefer_group`
+
+        """  # noqa: E501
+        return self._set_column_strategy(
+            (key,), {"deferred": False, "instrument": True}
+        )
+
+    def undefer_group(self, name: str) -> Self:
+        """Indicate that columns within the given deferred group name should be
+        undeferred.
+
+        The columns being undeferred are set up on the mapping as
+        :func:`.deferred` attributes and include a "group" name.
+
+        E.g::
+
+            session.query(MyClass).options(undefer_group("large_attrs"))
+
+        To undefer a group of attributes on a related entity, the path can be
+        spelled out using relationship loader options, such as
+        :func:`_orm.defaultload`::
+
+            select(MyClass).options(
+                defaultload("someattr").undefer_group("large_attrs")
+            )
+
+        .. seealso::
+
+            :ref:`orm_queryguide_column_deferral` - in the
+            :ref:`queryguide_toplevel`
+
+            :func:`_orm.defer`
+
+            :func:`_orm.undefer`
+
+        """
+        return self._set_column_strategy(
+            (_WILDCARD_TOKEN,), None, {f"undefer_group_{name}": True}
+        )
+
+    def with_expression(
+        self,
+        key: _AttrType,
+        expression: _ColumnExpressionArgument[Any],
+    ) -> Self:
+        r"""Apply an ad-hoc SQL expression to a "deferred expression"
+        attribute.
+
+        This option is used in conjunction with the
+        :func:`_orm.query_expression` mapper-level construct that indicates an
+        attribute which should be the target of an ad-hoc SQL expression.
+
+        E.g.::
+
+            stmt = select(SomeClass).options(
+                with_expression(SomeClass.x_y_expr, SomeClass.x + SomeClass.y)
+            )
+
+        .. versionadded:: 1.2
+
+        :param key: Attribute to be populated
+
+        :param expr: SQL expression to be applied to the attribute.
+
+        .. seealso::
+
+            :ref:`orm_queryguide_with_expression` - background and usage
+            examples
+
+        """
+
+        expression = _orm_full_deannotate(
+            coercions.expect(roles.LabeledColumnExprRole, expression)
+        )
+
+        return self._set_column_strategy(
+            (key,), {"query_expression": True}, extra_criteria=(expression,)
+        )
+
+    def selectin_polymorphic(self, classes: Iterable[Type[Any]]) -> Self:
+        """Indicate an eager load should take place for all attributes
+        specific to a subclass.
+
+        This uses an additional SELECT with IN against all matched primary
+        key values, and is the per-query analogue to the ``"selectin"``
+        setting on the :paramref:`.mapper.polymorphic_load` parameter.
+
+        .. versionadded:: 1.2
+
+        .. seealso::
+
+            :ref:`polymorphic_selectin`
+
+        """
+        self = self._set_class_strategy(
+            {"selectinload_polymorphic": True},
+            opts={
+                "entities": tuple(
+                    sorted((inspect(cls) for cls in classes), key=id)
+                )
+            },
+        )
+        return self
+
+    @overload
+    def _coerce_strat(self, strategy: _StrategySpec) -> _StrategyKey: ...
+
+    @overload
+    def _coerce_strat(self, strategy: Literal[None]) -> None: ...
+
+    def _coerce_strat(
+        self, strategy: Optional[_StrategySpec]
+    ) -> Optional[_StrategyKey]:
+        if strategy is not None:
+            strategy_key = tuple(sorted(strategy.items()))
+        else:
+            strategy_key = None
+        return strategy_key
+
+    @_generative
+    def _set_relationship_strategy(
+        self,
+        attr: _AttrType,
+        strategy: Optional[_StrategySpec],
+        propagate_to_loaders: bool = True,
+        opts: Optional[_OptsType] = None,
+        _reconcile_to_other: Optional[bool] = None,
+    ) -> Self:
+        strategy_key = self._coerce_strat(strategy)
+
+        self._clone_for_bind_strategy(
+            (attr,),
+            strategy_key,
+            _RELATIONSHIP_TOKEN,
+            opts=opts,
+            propagate_to_loaders=propagate_to_loaders,
+            reconcile_to_other=_reconcile_to_other,
+        )
+        return self
+
+    @_generative
+    def _set_column_strategy(
+        self,
+        attrs: Tuple[_AttrType, ...],
+        strategy: Optional[_StrategySpec],
+        opts: Optional[_OptsType] = None,
+        extra_criteria: Optional[Tuple[Any, ...]] = None,
+    ) -> Self:
+        strategy_key = self._coerce_strat(strategy)
+
+        self._clone_for_bind_strategy(
+            attrs,
+            strategy_key,
+            _COLUMN_TOKEN,
+            opts=opts,
+            attr_group=attrs,
+            extra_criteria=extra_criteria,
+        )
+        return self
+
+    @_generative
+    def _set_generic_strategy(
+        self,
+        attrs: Tuple[_AttrType, ...],
+        strategy: _StrategySpec,
+        _reconcile_to_other: Optional[bool] = None,
+    ) -> Self:
+        strategy_key = self._coerce_strat(strategy)
+        self._clone_for_bind_strategy(
+            attrs,
+            strategy_key,
+            None,
+            propagate_to_loaders=True,
+            reconcile_to_other=_reconcile_to_other,
+        )
+        return self
+
+    @_generative
+    def _set_class_strategy(
+        self, strategy: _StrategySpec, opts: _OptsType
+    ) -> Self:
+        strategy_key = self._coerce_strat(strategy)
+
+        self._clone_for_bind_strategy(None, strategy_key, None, opts=opts)
+        return self
+
+    def _apply_to_parent(self, parent: Load) -> None:
+        """apply this :class:`_orm._AbstractLoad` object as a sub-option o
+        a :class:`_orm.Load` object.
+
+        Implementation is provided by subclasses.
+
+        """
+        raise NotImplementedError()
+
+    def options(self, *opts: _AbstractLoad) -> Self:
+        r"""Apply a series of options as sub-options to this
+        :class:`_orm._AbstractLoad` object.
+
+        Implementation is provided by subclasses.
+
+        """
+        raise NotImplementedError()
+
+    def _clone_for_bind_strategy(
+        self,
+        attrs: Optional[Tuple[_AttrType, ...]],
+        strategy: Optional[_StrategyKey],
+        wildcard_key: Optional[_WildcardKeyType],
+        opts: Optional[_OptsType] = None,
+        attr_group: Optional[_AttrGroupType] = None,
+        propagate_to_loaders: bool = True,
+        reconcile_to_other: Optional[bool] = None,
+        extra_criteria: Optional[Tuple[Any, ...]] = None,
+    ) -> Self:
+        raise NotImplementedError()
+
+    def process_compile_state_replaced_entities(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Sequence[_MapperEntity],
+    ) -> None:
+        if not compile_state.compile_options._enable_eagerloads:
+            return
+
+        # process is being run here so that the options given are validated
+        # against what the lead entities were, as well as to accommodate
+        # for the entities having been replaced with equivalents
+        self._process(
+            compile_state,
+            mapper_entities,
+            not bool(compile_state.current_path),
+        )
+
+    def process_compile_state(self, compile_state: ORMCompileState) -> None:
+        if not compile_state.compile_options._enable_eagerloads:
+            return
+
+        self._process(
+            compile_state,
+            compile_state._lead_mapper_entities,
+            not bool(compile_state.current_path)
+            and not compile_state.compile_options._for_refresh_state,
+        )
+
+    def _process(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Sequence[_MapperEntity],
+        raiseerr: bool,
+    ) -> None:
+        """implemented by subclasses"""
+        raise NotImplementedError()
+
+    @classmethod
+    def _chop_path(
+        cls,
+        to_chop: _PathRepresentation,
+        path: PathRegistry,
+        debug: bool = False,
+    ) -> Optional[_PathRepresentation]:
+        i = -1
+
+        for i, (c_token, p_token) in enumerate(
+            zip(to_chop, path.natural_path)
+        ):
+            if isinstance(c_token, str):
+                if i == 0 and (
+                    c_token.endswith(f":{_DEFAULT_TOKEN}")
+                    or c_token.endswith(f":{_WILDCARD_TOKEN}")
+                ):
+                    return to_chop
+                elif (
+                    c_token != f"{_RELATIONSHIP_TOKEN}:{_WILDCARD_TOKEN}"
+                    and c_token != p_token.key  # type: ignore
+                ):
+                    return None
+
+            if c_token is p_token:
+                continue
+            elif (
+                isinstance(c_token, InspectionAttr)
+                and insp_is_mapper(c_token)
+                and insp_is_mapper(p_token)
+                and c_token.isa(p_token)
+            ):
+                continue
+
+            else:
+                return None
+        return to_chop[i + 1 :]
+
+
+class Load(_AbstractLoad):
+    """Represents loader options which modify the state of a
+    ORM-enabled :class:`_sql.Select` or a legacy :class:`_query.Query` in
+    order to affect how various mapped attributes are loaded.
+
+    The :class:`_orm.Load` object is in most cases used implicitly behind the
+    scenes when one makes use of a query option like :func:`_orm.joinedload`,
+    :func:`_orm.defer`, or similar.   It typically is not instantiated directly
+    except for in some very specific cases.
+
+    .. seealso::
+
+        :ref:`orm_queryguide_relationship_per_entity_wildcard` - illustrates an
+        example where direct use of :class:`_orm.Load` may be useful
+
+    """
+
+    __slots__ = (
+        "path",
+        "context",
+        "additional_source_entities",
+    )
+
+    _traverse_internals = [
+        ("path", visitors.ExtendedInternalTraversal.dp_has_cache_key),
+        (
+            "context",
+            visitors.InternalTraversal.dp_has_cache_key_list,
+        ),
+        ("propagate_to_loaders", visitors.InternalTraversal.dp_boolean),
+        (
+            "additional_source_entities",
+            visitors.InternalTraversal.dp_has_cache_key_list,
+        ),
+    ]
+    _cache_key_traversal = None
+
+    path: PathRegistry
+    context: Tuple[_LoadElement, ...]
+    additional_source_entities: Tuple[_InternalEntityType[Any], ...]
+
+    def __init__(self, entity: _EntityType[Any]):
+        insp = cast("Union[Mapper[Any], AliasedInsp[Any]]", inspect(entity))
+        insp._post_inspect
+
+        self.path = insp._path_registry
+        self.context = ()
+        self.propagate_to_loaders = False
+        self.additional_source_entities = ()
+
+    def __str__(self) -> str:
+        return f"Load({self.path[0]})"
+
+    @classmethod
+    def _construct_for_existing_path(
+        cls, path: AbstractEntityRegistry
+    ) -> Load:
+        load = cls.__new__(cls)
+        load.path = path
+        load.context = ()
+        load.propagate_to_loaders = False
+        load.additional_source_entities = ()
+        return load
+
+    def _adapt_cached_option_to_uncached_option(
+        self, context: QueryContext, uncached_opt: ORMOption
+    ) -> ORMOption:
+        if uncached_opt is self:
+            return self
+        return self._adjust_for_extra_criteria(context)
+
+    def _prepend_path(self, path: PathRegistry) -> Load:
+        cloned = self._clone()
+        cloned.context = tuple(
+            element._prepend_path(path) for element in self.context
+        )
+        return cloned
+
+    def _adjust_for_extra_criteria(self, context: QueryContext) -> Load:
+        """Apply the current bound parameters in a QueryContext to all
+        occurrences "extra_criteria" stored within this ``Load`` object,
+        returning a new instance of this ``Load`` object.
+
+        """
+
+        # avoid generating cache keys for the queries if we don't
+        # actually have any extra_criteria options, which is the
+        # common case
+        for value in self.context:
+            if value._extra_criteria:
+                break
+        else:
+            return self
+
+        replacement_cache_key = context.user_passed_query._generate_cache_key()
+
+        if replacement_cache_key is None:
+            return self
+
+        orig_query = context.compile_state.select_statement
+        orig_cache_key = orig_query._generate_cache_key()
+        assert orig_cache_key is not None
+
+        def process(
+            opt: _LoadElement,
+            replacement_cache_key: CacheKey,
+            orig_cache_key: CacheKey,
+        ) -> _LoadElement:
+            cloned_opt = opt._clone()
+
+            cloned_opt._extra_criteria = tuple(
+                replacement_cache_key._apply_params_to_element(
+                    orig_cache_key, crit
+                )
+                for crit in cloned_opt._extra_criteria
+            )
+
+            return cloned_opt
+
+        cloned = self._clone()
+        cloned.context = tuple(
+            (
+                process(value, replacement_cache_key, orig_cache_key)
+                if value._extra_criteria
+                else value
+            )
+            for value in self.context
+        )
+        return cloned
+
+    def _reconcile_query_entities_with_us(self, mapper_entities, raiseerr):
+        """called at process time to allow adjustment of the root
+        entity inside of _LoadElement objects.
+
+        """
+        path = self.path
+
+        ezero = None
+        for ent in mapper_entities:
+            ezero = ent.entity_zero
+            if ezero and orm_util._entity_corresponds_to(
+                # technically this can be a token also, but this is
+                # safe to pass to _entity_corresponds_to()
+                ezero,
+                cast("_InternalEntityType[Any]", path[0]),
+            ):
+                return ezero
+
+        return None
+
+    def _process(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Sequence[_MapperEntity],
+        raiseerr: bool,
+    ) -> None:
+        reconciled_lead_entity = self._reconcile_query_entities_with_us(
+            mapper_entities, raiseerr
+        )
+
+        # if the context has a current path, this is a lazy load
+        has_current_path = bool(compile_state.compile_options._current_path)
+
+        for loader in self.context:
+            # issue #11292
+            # historically, propagate_to_loaders was only considered at
+            # object loading time, whether or not to carry along options
+            # onto an object's loaded state where it would be used by lazyload.
+            # however, the defaultload() option needs to propagate in case
+            # its sub-options propagate_to_loaders, but its sub-options
+            # that dont propagate should not be applied for lazy loaders.
+            # so we check again
+            if has_current_path and not loader.propagate_to_loaders:
+                continue
+            loader.process_compile_state(
+                self,
+                compile_state,
+                mapper_entities,
+                reconciled_lead_entity,
+                raiseerr,
+            )
+
+    def _apply_to_parent(self, parent: Load) -> None:
+        """apply this :class:`_orm.Load` object as a sub-option of another
+        :class:`_orm.Load` object.
+
+        This method is used by the :meth:`_orm.Load.options` method.
+
+        """
+        cloned = self._generate()
+
+        assert cloned.propagate_to_loaders == self.propagate_to_loaders
+
+        if not any(
+            orm_util._entity_corresponds_to_use_path_impl(
+                elem, cloned.path.odd_element(0)
+            )
+            for elem in (parent.path.odd_element(-1),)
+            + parent.additional_source_entities
+        ):
+            if len(cloned.path) > 1:
+                attrname = cloned.path[1]
+                parent_entity = cloned.path[0]
+            else:
+                attrname = cloned.path[0]
+                parent_entity = cloned.path[0]
+            _raise_for_does_not_link(parent.path, attrname, parent_entity)
+
+        cloned.path = PathRegistry.coerce(parent.path[0:-1] + cloned.path[:])
+
+        if self.context:
+            cloned.context = tuple(
+                value._prepend_path_from(parent) for value in self.context
+            )
+
+        if cloned.context:
+            parent.context += cloned.context
+            parent.additional_source_entities += (
+                cloned.additional_source_entities
+            )
+
+    @_generative
+    def options(self, *opts: _AbstractLoad) -> Self:
+        r"""Apply a series of options as sub-options to this
+        :class:`_orm.Load`
+        object.
+
+        E.g.::
+
+            query = session.query(Author)
+            query = query.options(
+                joinedload(Author.book).options(
+                    load_only(Book.summary, Book.excerpt),
+                    joinedload(Book.citations).options(joinedload(Citation.author)),
+                )
+            )
+
+        :param \*opts: A series of loader option objects (ultimately
+         :class:`_orm.Load` objects) which should be applied to the path
+         specified by this :class:`_orm.Load` object.
+
+        .. versionadded:: 1.3.6
+
+        .. seealso::
+
+            :func:`.defaultload`
+
+            :ref:`orm_queryguide_relationship_sub_options`
+
+        """
+        for opt in opts:
+            try:
+                opt._apply_to_parent(self)
+            except AttributeError as ae:
+                if not isinstance(opt, _AbstractLoad):
+                    raise sa_exc.ArgumentError(
+                        f"Loader option {opt} is not compatible with the "
+                        "Load.options() method."
+                    ) from ae
+                else:
+                    raise
+        return self
+
+    def _clone_for_bind_strategy(
+        self,
+        attrs: Optional[Tuple[_AttrType, ...]],
+        strategy: Optional[_StrategyKey],
+        wildcard_key: Optional[_WildcardKeyType],
+        opts: Optional[_OptsType] = None,
+        attr_group: Optional[_AttrGroupType] = None,
+        propagate_to_loaders: bool = True,
+        reconcile_to_other: Optional[bool] = None,
+        extra_criteria: Optional[Tuple[Any, ...]] = None,
+    ) -> Self:
+        # for individual strategy that needs to propagate, set the whole
+        # Load container to also propagate, so that it shows up in
+        # InstanceState.load_options
+        if propagate_to_loaders:
+            self.propagate_to_loaders = True
+
+        if self.path.is_token:
+            raise sa_exc.ArgumentError(
+                "Wildcard token cannot be followed by another entity"
+            )
+
+        elif path_is_property(self.path):
+            # re-use the lookup which will raise a nicely formatted
+            # LoaderStrategyException
+            if strategy:
+                self.path.prop._strategy_lookup(self.path.prop, strategy[0])
+            else:
+                raise sa_exc.ArgumentError(
+                    f"Mapped attribute '{self.path.prop}' does not "
+                    "refer to a mapped entity"
+                )
+
+        if attrs is None:
+            load_element = _ClassStrategyLoad.create(
+                self.path,
+                None,
+                strategy,
+                wildcard_key,
+                opts,
+                propagate_to_loaders,
+                attr_group=attr_group,
+                reconcile_to_other=reconcile_to_other,
+                extra_criteria=extra_criteria,
+            )
+            if load_element:
+                self.context += (load_element,)
+                assert opts is not None
+                self.additional_source_entities += cast(
+                    "Tuple[_InternalEntityType[Any]]", opts["entities"]
+                )
+
+        else:
+            for attr in attrs:
+                if isinstance(attr, str):
+                    load_element = _TokenStrategyLoad.create(
+                        self.path,
+                        attr,
+                        strategy,
+                        wildcard_key,
+                        opts,
+                        propagate_to_loaders,
+                        attr_group=attr_group,
+                        reconcile_to_other=reconcile_to_other,
+                        extra_criteria=extra_criteria,
+                    )
+                else:
+                    load_element = _AttributeStrategyLoad.create(
+                        self.path,
+                        attr,
+                        strategy,
+                        wildcard_key,
+                        opts,
+                        propagate_to_loaders,
+                        attr_group=attr_group,
+                        reconcile_to_other=reconcile_to_other,
+                        extra_criteria=extra_criteria,
+                    )
+
+                if load_element:
+                    # for relationship options, update self.path on this Load
+                    # object with the latest path.
+                    if wildcard_key is _RELATIONSHIP_TOKEN:
+                        self.path = load_element.path
+                    self.context += (load_element,)
+
+                    # this seems to be effective for selectinloader,
+                    # giving the extra match to one more level deep.
+                    # but does not work for immediateloader, which still
+                    # must add additional options at load time
+                    if load_element.local_opts.get("recursion_depth", False):
+                        r1 = load_element._recurse()
+                        self.context += (r1,)
+
+        return self
+
+    def __getstate__(self):
+        d = self._shallow_to_dict()
+        d["path"] = self.path.serialize()
+        return d
+
+    def __setstate__(self, state):
+        state["path"] = PathRegistry.deserialize(state["path"])
+        self._shallow_from_dict(state)
+
+
+class _WildcardLoad(_AbstractLoad):
+    """represent a standalone '*' load operation"""
+
+    __slots__ = ("strategy", "path", "local_opts")
+
+    _traverse_internals = [
+        ("strategy", visitors.ExtendedInternalTraversal.dp_plain_obj),
+        ("path", visitors.ExtendedInternalTraversal.dp_plain_obj),
+        (
+            "local_opts",
+            visitors.ExtendedInternalTraversal.dp_string_multi_dict,
+        ),
+    ]
+    cache_key_traversal: _CacheKeyTraversalType = None
+
+    strategy: Optional[Tuple[Any, ...]]
+    local_opts: _OptsType
+    path: Union[Tuple[()], Tuple[str]]
+    propagate_to_loaders = False
+
+    def __init__(self) -> None:
+        self.path = ()
+        self.strategy = None
+        self.local_opts = util.EMPTY_DICT
+
+    def _clone_for_bind_strategy(
+        self,
+        attrs,
+        strategy,
+        wildcard_key,
+        opts=None,
+        attr_group=None,
+        propagate_to_loaders=True,
+        reconcile_to_other=None,
+        extra_criteria=None,
+    ):
+        assert attrs is not None
+        attr = attrs[0]
+        assert (
+            wildcard_key
+            and isinstance(attr, str)
+            and attr in (_WILDCARD_TOKEN, _DEFAULT_TOKEN)
+        )
+
+        attr = f"{wildcard_key}:{attr}"
+
+        self.strategy = strategy
+        self.path = (attr,)
+        if opts:
+            self.local_opts = util.immutabledict(opts)
+
+        assert extra_criteria is None
+
+    def options(self, *opts: _AbstractLoad) -> Self:
+        raise NotImplementedError("Star option does not support sub-options")
+
+    def _apply_to_parent(self, parent: Load) -> None:
+        """apply this :class:`_orm._WildcardLoad` object as a sub-option of
+        a :class:`_orm.Load` object.
+
+        This method is used by the :meth:`_orm.Load.options` method.   Note
+        that :class:`_orm.WildcardLoad` itself can't have sub-options, but
+        it may be used as the sub-option of a :class:`_orm.Load` object.
+
+        """
+        assert self.path
+        attr = self.path[0]
+        if attr.endswith(_DEFAULT_TOKEN):
+            attr = f"{attr.split(':')[0]}:{_WILDCARD_TOKEN}"
+
+        effective_path = cast(AbstractEntityRegistry, parent.path).token(attr)
+
+        assert effective_path.is_token
+
+        loader = _TokenStrategyLoad.create(
+            effective_path,
+            None,
+            self.strategy,
+            None,
+            self.local_opts,
+            self.propagate_to_loaders,
+        )
+
+        parent.context += (loader,)
+
+    def _process(self, compile_state, mapper_entities, raiseerr):
+        is_refresh = compile_state.compile_options._for_refresh_state
+
+        if is_refresh and not self.propagate_to_loaders:
+            return
+
+        entities = [ent.entity_zero for ent in mapper_entities]
+        current_path = compile_state.current_path
+
+        start_path: _PathRepresentation = self.path
+
+        if current_path:
+            # TODO: no cases in test suite where we actually get
+            # None back here
+            new_path = self._chop_path(start_path, current_path)
+            if new_path is None:
+                return
+
+            # chop_path does not actually "chop" a wildcard token path,
+            # just returns it
+            assert new_path == start_path
+
+        # start_path is a single-token tuple
+        assert start_path and len(start_path) == 1
+
+        token = start_path[0]
+        assert isinstance(token, str)
+        entity = self._find_entity_basestring(entities, token, raiseerr)
+
+        if not entity:
+            return
+
+        path_element = entity
+
+        # transfer our entity-less state into a Load() object
+        # with a real entity path.  Start with the lead entity
+        # we just located, then go through the rest of our path
+        # tokens and populate into the Load().
+
+        assert isinstance(token, str)
+        loader = _TokenStrategyLoad.create(
+            path_element._path_registry,
+            token,
+            self.strategy,
+            None,
+            self.local_opts,
+            self.propagate_to_loaders,
+            raiseerr=raiseerr,
+        )
+        if not loader:
+            return
+
+        assert loader.path.is_token
+
+        # don't pass a reconciled lead entity here
+        loader.process_compile_state(
+            self, compile_state, mapper_entities, None, raiseerr
+        )
+
+        return loader
+
+    def _find_entity_basestring(
+        self,
+        entities: Iterable[_InternalEntityType[Any]],
+        token: str,
+        raiseerr: bool,
+    ) -> Optional[_InternalEntityType[Any]]:
+        if token.endswith(f":{_WILDCARD_TOKEN}"):
+            if len(list(entities)) != 1:
+                if raiseerr:
+                    raise sa_exc.ArgumentError(
+                        "Can't apply wildcard ('*') or load_only() "
+                        f"loader option to multiple entities "
+                        f"{', '.join(str(ent) for ent in entities)}. Specify "
+                        "loader options for each entity individually, such as "
+                        f"""{
+                            ", ".join(
+                                f"Load({ent}).some_option('*')"
+                                for ent in entities
+                            )
+                        }."""
+                    )
+        elif token.endswith(_DEFAULT_TOKEN):
+            raiseerr = False
+
+        for ent in entities:
+            # return only the first _MapperEntity when searching
+            # based on string prop name.   Ideally object
+            # attributes are used to specify more exactly.
+            return ent
+        else:
+            if raiseerr:
+                raise sa_exc.ArgumentError(
+                    "Query has only expression-based entities - "
+                    f'can\'t find property named "{token}".'
+                )
+            else:
+                return None
+
+    def __getstate__(self) -> Dict[str, Any]:
+        d = self._shallow_to_dict()
+        return d
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        self._shallow_from_dict(state)
+
+
+class _LoadElement(
+    cache_key.HasCacheKey, traversals.HasShallowCopy, visitors.Traversible
+):
+    """represents strategy information to select for a LoaderStrategy
+    and pass options to it.
+
+    :class:`._LoadElement` objects provide the inner datastructure
+    stored by a :class:`_orm.Load` object and are also the object passed
+    to methods like :meth:`.LoaderStrategy.setup_query`.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __slots__ = (
+        "path",
+        "strategy",
+        "propagate_to_loaders",
+        "local_opts",
+        "_extra_criteria",
+        "_reconcile_to_other",
+    )
+    __visit_name__ = "load_element"
+
+    _traverse_internals = [
+        ("path", visitors.ExtendedInternalTraversal.dp_has_cache_key),
+        ("strategy", visitors.ExtendedInternalTraversal.dp_plain_obj),
+        (
+            "local_opts",
+            visitors.ExtendedInternalTraversal.dp_string_multi_dict,
+        ),
+        ("_extra_criteria", visitors.InternalTraversal.dp_clauseelement_list),
+        ("propagate_to_loaders", visitors.InternalTraversal.dp_plain_obj),
+        ("_reconcile_to_other", visitors.InternalTraversal.dp_plain_obj),
+    ]
+    _cache_key_traversal = None
+
+    _extra_criteria: Tuple[Any, ...]
+
+    _reconcile_to_other: Optional[bool]
+    strategy: Optional[_StrategyKey]
+    path: PathRegistry
+    propagate_to_loaders: bool
+
+    local_opts: util.immutabledict[str, Any]
+
+    is_token_strategy: bool
+    is_class_strategy: bool
+
+    def __hash__(self) -> int:
+        return id(self)
+
+    def __eq__(self, other):
+        return traversals.compare(self, other)
+
+    @property
+    def is_opts_only(self) -> bool:
+        return bool(self.local_opts and self.strategy is None)
+
+    def _clone(self, **kw: Any) -> _LoadElement:
+        cls = self.__class__
+        s = cls.__new__(cls)
+
+        self._shallow_copy_to(s)
+        return s
+
+    def _update_opts(self, **kw: Any) -> _LoadElement:
+        new = self._clone()
+        new.local_opts = new.local_opts.union(kw)
+        return new
+
+    def __getstate__(self) -> Dict[str, Any]:
+        d = self._shallow_to_dict()
+        d["path"] = self.path.serialize()
+        return d
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        state["path"] = PathRegistry.deserialize(state["path"])
+        self._shallow_from_dict(state)
+
+    def _raise_for_no_match(self, parent_loader, mapper_entities):
+        path = parent_loader.path
+
+        found_entities = False
+        for ent in mapper_entities:
+            ezero = ent.entity_zero
+            if ezero:
+                found_entities = True
+                break
+
+        if not found_entities:
+            raise sa_exc.ArgumentError(
+                "Query has only expression-based entities; "
+                f"attribute loader options for {path[0]} can't "
+                "be applied here."
+            )
+        else:
+            raise sa_exc.ArgumentError(
+                f"Mapped class {path[0]} does not apply to any of the "
+                f"root entities in this query, e.g. "
+                f"""{
+                    ", ".join(
+                        str(x.entity_zero)
+                        for x in mapper_entities if x.entity_zero
+                    )}. Please """
+                "specify the full path "
+                "from one of the root entities to the target "
+                "attribute. "
+            )
+
+    def _adjust_effective_path_for_current_path(
+        self, effective_path: PathRegistry, current_path: PathRegistry
+    ) -> Optional[PathRegistry]:
+        """receives the 'current_path' entry from an :class:`.ORMCompileState`
+        instance, which is set during lazy loads and secondary loader strategy
+        loads, and adjusts the given path to be relative to the
+        current_path.
+
+        E.g. given a loader path and current path:
+
+        .. sourcecode:: text
+
+            lp: User -> orders -> Order -> items -> Item -> keywords -> Keyword
+
+            cp: User -> orders -> Order -> items
+
+        The adjusted path would be:
+
+        .. sourcecode:: text
+
+            Item -> keywords -> Keyword
+
+
+        """
+        chopped_start_path = Load._chop_path(
+            effective_path.natural_path, current_path
+        )
+        if not chopped_start_path:
+            return None
+
+        tokens_removed_from_start_path = len(effective_path) - len(
+            chopped_start_path
+        )
+
+        loader_lead_path_element = self.path[tokens_removed_from_start_path]
+
+        effective_path = PathRegistry.coerce(
+            (loader_lead_path_element,) + chopped_start_path[1:]
+        )
+
+        return effective_path
+
+    def _init_path(
+        self, path, attr, wildcard_key, attr_group, raiseerr, extra_criteria
+    ):
+        """Apply ORM attributes and/or wildcard to an existing path, producing
+        a new path.
+
+        This method is used within the :meth:`.create` method to initialize
+        a :class:`._LoadElement` object.
+
+        """
+        raise NotImplementedError()
+
+    def _prepare_for_compile_state(
+        self,
+        parent_loader,
+        compile_state,
+        mapper_entities,
+        reconciled_lead_entity,
+        raiseerr,
+    ):
+        """implemented by subclasses."""
+        raise NotImplementedError()
+
+    def process_compile_state(
+        self,
+        parent_loader,
+        compile_state,
+        mapper_entities,
+        reconciled_lead_entity,
+        raiseerr,
+    ):
+        """populate ORMCompileState.attributes with loader state for this
+        _LoadElement.
+
+        """
+        keys = self._prepare_for_compile_state(
+            parent_loader,
+            compile_state,
+            mapper_entities,
+            reconciled_lead_entity,
+            raiseerr,
+        )
+        for key in keys:
+            if key in compile_state.attributes:
+                compile_state.attributes[key] = _LoadElement._reconcile(
+                    self, compile_state.attributes[key]
+                )
+            else:
+                compile_state.attributes[key] = self
+
+    @classmethod
+    def create(
+        cls,
+        path: PathRegistry,
+        attr: Union[_AttrType, _StrPathToken, None],
+        strategy: Optional[_StrategyKey],
+        wildcard_key: Optional[_WildcardKeyType],
+        local_opts: Optional[_OptsType],
+        propagate_to_loaders: bool,
+        raiseerr: bool = True,
+        attr_group: Optional[_AttrGroupType] = None,
+        reconcile_to_other: Optional[bool] = None,
+        extra_criteria: Optional[Tuple[Any, ...]] = None,
+    ) -> _LoadElement:
+        """Create a new :class:`._LoadElement` object."""
+
+        opt = cls.__new__(cls)
+        opt.path = path
+        opt.strategy = strategy
+        opt.propagate_to_loaders = propagate_to_loaders
+        opt.local_opts = (
+            util.immutabledict(local_opts) if local_opts else util.EMPTY_DICT
+        )
+        opt._extra_criteria = ()
+
+        if reconcile_to_other is not None:
+            opt._reconcile_to_other = reconcile_to_other
+        elif strategy is None and not local_opts:
+            opt._reconcile_to_other = True
+        else:
+            opt._reconcile_to_other = None
+
+        path = opt._init_path(
+            path, attr, wildcard_key, attr_group, raiseerr, extra_criteria
+        )
+
+        if not path:
+            return None  # type: ignore
+
+        assert opt.is_token_strategy == path.is_token
+
+        opt.path = path
+        return opt
+
+    def __init__(self) -> None:
+        raise NotImplementedError()
+
+    def _recurse(self) -> _LoadElement:
+        cloned = self._clone()
+        cloned.path = PathRegistry.coerce(self.path[:] + self.path[-2:])
+
+        return cloned
+
+    def _prepend_path_from(self, parent: Load) -> _LoadElement:
+        """adjust the path of this :class:`._LoadElement` to be
+        a subpath of that of the given parent :class:`_orm.Load` object's
+        path.
+
+        This is used by the :meth:`_orm.Load._apply_to_parent` method,
+        which is in turn part of the :meth:`_orm.Load.options` method.
+
+        """
+
+        if not any(
+            orm_util._entity_corresponds_to_use_path_impl(
+                elem,
+                self.path.odd_element(0),
+            )
+            for elem in (parent.path.odd_element(-1),)
+            + parent.additional_source_entities
+        ):
+            raise sa_exc.ArgumentError(
+                f'Attribute "{self.path[1]}" does not link '
+                f'from element "{parent.path[-1]}".'
+            )
+
+        return self._prepend_path(parent.path)
+
+    def _prepend_path(self, path: PathRegistry) -> _LoadElement:
+        cloned = self._clone()
+
+        assert cloned.strategy == self.strategy
+        assert cloned.local_opts == self.local_opts
+        assert cloned.is_class_strategy == self.is_class_strategy
+
+        cloned.path = PathRegistry.coerce(path[0:-1] + cloned.path[:])
+
+        return cloned
+
+    @staticmethod
+    def _reconcile(
+        replacement: _LoadElement, existing: _LoadElement
+    ) -> _LoadElement:
+        """define behavior for when two Load objects are to be put into
+        the context.attributes under the same key.
+
+        :param replacement: ``_LoadElement`` that seeks to replace the
+         existing one
+
+        :param existing: ``_LoadElement`` that is already present.
+
+        """
+        # mapper inheritance loading requires fine-grained "block other
+        # options" / "allow these options to be overridden" behaviors
+        # see test_poly_loading.py
+
+        if replacement._reconcile_to_other:
+            return existing
+        elif replacement._reconcile_to_other is False:
+            return replacement
+        elif existing._reconcile_to_other:
+            return replacement
+        elif existing._reconcile_to_other is False:
+            return existing
+
+        if existing is replacement:
+            return replacement
+        elif (
+            existing.strategy == replacement.strategy
+            and existing.local_opts == replacement.local_opts
+        ):
+            return replacement
+        elif replacement.is_opts_only:
+            existing = existing._clone()
+            existing.local_opts = existing.local_opts.union(
+                replacement.local_opts
+            )
+            existing._extra_criteria += replacement._extra_criteria
+            return existing
+        elif existing.is_opts_only:
+            replacement = replacement._clone()
+            replacement.local_opts = replacement.local_opts.union(
+                existing.local_opts
+            )
+            replacement._extra_criteria += existing._extra_criteria
+            return replacement
+        elif replacement.path.is_token:
+            # use 'last one wins' logic for wildcard options.  this is also
+            # kind of inconsistent vs. options that are specific paths which
+            # will raise as below
+            return replacement
+
+        raise sa_exc.InvalidRequestError(
+            f"Loader strategies for {replacement.path} conflict"
+        )
+
+
+class _AttributeStrategyLoad(_LoadElement):
+    """Loader strategies against specific relationship or column paths.
+
+    e.g.::
+
+        joinedload(User.addresses)
+        defer(Order.name)
+        selectinload(User.orders).lazyload(Order.items)
+
+    """
+
+    __slots__ = ("_of_type", "_path_with_polymorphic_path")
+
+    __visit_name__ = "attribute_strategy_load_element"
+
+    _traverse_internals = _LoadElement._traverse_internals + [
+        ("_of_type", visitors.ExtendedInternalTraversal.dp_multi),
+        (
+            "_path_with_polymorphic_path",
+            visitors.ExtendedInternalTraversal.dp_has_cache_key,
+        ),
+    ]
+
+    _of_type: Union[Mapper[Any], AliasedInsp[Any], None]
+    _path_with_polymorphic_path: Optional[PathRegistry]
+
+    is_class_strategy = False
+    is_token_strategy = False
+
+    def _init_path(
+        self, path, attr, wildcard_key, attr_group, raiseerr, extra_criteria
+    ):
+        assert attr is not None
+        self._of_type = None
+        self._path_with_polymorphic_path = None
+        insp, _, prop = _parse_attr_argument(attr)
+
+        if insp.is_property:
+            # direct property can be sent from internal strategy logic
+            # that sets up specific loaders, such as
+            # emit_lazyload->_lazyload_reverse
+            # prop = found_property = attr
+            prop = attr
+            path = path[prop]
+
+            if path.has_entity:
+                path = path.entity_path
+            return path
+
+        elif not insp.is_attribute:
+            # should not reach here;
+            assert False
+
+        # here we assume we have user-passed InstrumentedAttribute
+        if not orm_util._entity_corresponds_to_use_path_impl(
+            path[-1], attr.parent
+        ):
+            if raiseerr:
+                if attr_group and attr is not attr_group[0]:
+                    raise sa_exc.ArgumentError(
+                        "Can't apply wildcard ('*') or load_only() "
+                        "loader option to multiple entities in the "
+                        "same option. Use separate options per entity."
+                    )
+                else:
+                    _raise_for_does_not_link(path, str(attr), attr.parent)
+            else:
+                return None
+
+        # note the essential logic of this attribute was very different in
+        # 1.4, where there were caching failures in e.g.
+        # test_relationship_criteria.py::RelationshipCriteriaTest::
+        # test_selectinload_nested_criteria[True] if an existing
+        # "_extra_criteria" on a Load object were replaced with that coming
+        # from an attribute.   This appears to have been an artifact of how
+        # _UnboundLoad / Load interacted together, which was opaque and
+        # poorly defined.
+        if extra_criteria:
+            assert not attr._extra_criteria
+            self._extra_criteria = extra_criteria
+        else:
+            self._extra_criteria = attr._extra_criteria
+
+        if getattr(attr, "_of_type", None):
+            ac = attr._of_type
+            ext_info = inspect(ac)
+            self._of_type = ext_info
+
+            self._path_with_polymorphic_path = path.entity_path[prop]
+
+            path = path[prop][ext_info]
+
+        else:
+            path = path[prop]
+
+        if path.has_entity:
+            path = path.entity_path
+
+        return path
+
+    def _generate_extra_criteria(self, context):
+        """Apply the current bound parameters in a QueryContext to the
+        immediate "extra_criteria" stored with this Load object.
+
+        Load objects are typically pulled from the cached version of
+        the statement from a QueryContext.  The statement currently being
+        executed will have new values (and keys) for bound parameters in the
+        extra criteria which need to be applied by loader strategies when
+        they handle this criteria for a result set.
+
+        """
+
+        assert (
+            self._extra_criteria
+        ), "this should only be called if _extra_criteria is present"
+
+        orig_query = context.compile_state.select_statement
+        current_query = context.query
+
+        # NOTE: while it seems like we should not do the "apply" operation
+        # here if orig_query is current_query, skipping it in the "optimized"
+        # case causes the query to be different from a cache key perspective,
+        # because we are creating a copy of the criteria which is no longer
+        # the same identity of the _extra_criteria in the loader option
+        # itself.  cache key logic produces a different key for
+        # (A, copy_of_A) vs. (A, A), because in the latter case it shortens
+        # the second part of the key to just indicate on identity.
+
+        # if orig_query is current_query:
+        # not cached yet.   just do the and_()
+        #    return and_(*self._extra_criteria)
+
+        k1 = orig_query._generate_cache_key()
+        k2 = current_query._generate_cache_key()
+
+        return k2._apply_params_to_element(k1, and_(*self._extra_criteria))
+
+    def _set_of_type_info(self, context, current_path):
+        assert self._path_with_polymorphic_path
+
+        pwpi = self._of_type
+        assert pwpi
+        if not pwpi.is_aliased_class:
+            pwpi = inspect(
+                orm_util.AliasedInsp._with_polymorphic_factory(
+                    pwpi.mapper.base_mapper,
+                    (pwpi.mapper,),
+                    aliased=True,
+                    _use_mapper_path=True,
+                )
+            )
+        start_path = self._path_with_polymorphic_path
+        if current_path:
+            new_path = self._adjust_effective_path_for_current_path(
+                start_path, current_path
+            )
+            if new_path is None:
+                return
+            start_path = new_path
+
+        key = ("path_with_polymorphic", start_path.natural_path)
+        if key in context:
+            existing_aliased_insp = context[key]
+            this_aliased_insp = pwpi
+            new_aliased_insp = existing_aliased_insp._merge_with(
+                this_aliased_insp
+            )
+            context[key] = new_aliased_insp
+        else:
+            context[key] = pwpi
+
+    def _prepare_for_compile_state(
+        self,
+        parent_loader,
+        compile_state,
+        mapper_entities,
+        reconciled_lead_entity,
+        raiseerr,
+    ):
+        # _AttributeStrategyLoad
+
+        current_path = compile_state.current_path
+        is_refresh = compile_state.compile_options._for_refresh_state
+        assert not self.path.is_token
+
+        if is_refresh and not self.propagate_to_loaders:
+            return []
+
+        if self._of_type:
+            # apply additional with_polymorphic alias that may have been
+            # generated.  this has to happen even if this is a defaultload
+            self._set_of_type_info(compile_state.attributes, current_path)
+
+        # omit setting loader attributes for a "defaultload" type of option
+        if not self.strategy and not self.local_opts:
+            return []
+
+        if raiseerr and not reconciled_lead_entity:
+            self._raise_for_no_match(parent_loader, mapper_entities)
+
+        if self.path.has_entity:
+            effective_path = self.path.parent
+        else:
+            effective_path = self.path
+
+        if current_path:
+            assert effective_path is not None
+            effective_path = self._adjust_effective_path_for_current_path(
+                effective_path, current_path
+            )
+            if effective_path is None:
+                return []
+
+        return [("loader", cast(PathRegistry, effective_path).natural_path)]
+
+    def __getstate__(self):
+        d = super().__getstate__()
+
+        # can't pickle this.  See
+        # test_pickled.py -> test_lazyload_extra_criteria_not_supported
+        # where we should be emitting a warning for the usual case where this
+        # would be non-None
+        d["_extra_criteria"] = ()
+
+        if self._path_with_polymorphic_path:
+            d["_path_with_polymorphic_path"] = (
+                self._path_with_polymorphic_path.serialize()
+            )
+
+        if self._of_type:
+            if self._of_type.is_aliased_class:
+                d["_of_type"] = None
+            elif self._of_type.is_mapper:
+                d["_of_type"] = self._of_type.class_
+            else:
+                assert False, "unexpected object for _of_type"
+
+        return d
+
+    def __setstate__(self, state):
+        super().__setstate__(state)
+
+        if state.get("_path_with_polymorphic_path", None):
+            self._path_with_polymorphic_path = PathRegistry.deserialize(
+                state["_path_with_polymorphic_path"]
+            )
+        else:
+            self._path_with_polymorphic_path = None
+
+        if state.get("_of_type", None):
+            self._of_type = inspect(state["_of_type"])
+        else:
+            self._of_type = None
+
+
+class _TokenStrategyLoad(_LoadElement):
+    """Loader strategies against wildcard attributes
+
+    e.g.::
+
+        raiseload("*")
+        Load(User).lazyload("*")
+        defer("*")
+        load_only(User.name, User.email)  # will create a defer('*')
+        joinedload(User.addresses).raiseload("*")
+
+    """
+
+    __visit_name__ = "token_strategy_load_element"
+
+    inherit_cache = True
+    is_class_strategy = False
+    is_token_strategy = True
+
+    def _init_path(
+        self, path, attr, wildcard_key, attr_group, raiseerr, extra_criteria
+    ):
+        # assert isinstance(attr, str) or attr is None
+        if attr is not None:
+            default_token = attr.endswith(_DEFAULT_TOKEN)
+            if attr.endswith(_WILDCARD_TOKEN) or default_token:
+                if wildcard_key:
+                    attr = f"{wildcard_key}:{attr}"
+
+                path = path.token(attr)
+                return path
+            else:
+                raise sa_exc.ArgumentError(
+                    "Strings are not accepted for attribute names in loader "
+                    "options; please use class-bound attributes directly."
+                )
+        return path
+
+    def _prepare_for_compile_state(
+        self,
+        parent_loader,
+        compile_state,
+        mapper_entities,
+        reconciled_lead_entity,
+        raiseerr,
+    ):
+        # _TokenStrategyLoad
+
+        current_path = compile_state.current_path
+        is_refresh = compile_state.compile_options._for_refresh_state
+
+        assert self.path.is_token
+
+        if is_refresh and not self.propagate_to_loaders:
+            return []
+
+        # omit setting attributes for a "defaultload" type of option
+        if not self.strategy and not self.local_opts:
+            return []
+
+        effective_path = self.path
+        if reconciled_lead_entity:
+            effective_path = PathRegistry.coerce(
+                (reconciled_lead_entity,) + effective_path.path[1:]
+            )
+
+        if current_path:
+            new_effective_path = self._adjust_effective_path_for_current_path(
+                effective_path, current_path
+            )
+            if new_effective_path is None:
+                return []
+            effective_path = new_effective_path
+
+        # for a wildcard token, expand out the path we set
+        # to encompass everything from the query entity on
+        # forward.  not clear if this is necessary when current_path
+        # is set.
+
+        return [
+            ("loader", natural_path)
+            for natural_path in (
+                cast(
+                    TokenRegistry, effective_path
+                )._generate_natural_for_superclasses()
+            )
+        ]
+
+
+class _ClassStrategyLoad(_LoadElement):
+    """Loader strategies that deals with a class as a target, not
+    an attribute path
+
+    e.g.::
+
+        q = s.query(Person).options(
+            selectin_polymorphic(Person, [Engineer, Manager])
+        )
+
+    """
+
+    inherit_cache = True
+    is_class_strategy = True
+    is_token_strategy = False
+
+    __visit_name__ = "class_strategy_load_element"
+
+    def _init_path(
+        self, path, attr, wildcard_key, attr_group, raiseerr, extra_criteria
+    ):
+        return path
+
+    def _prepare_for_compile_state(
+        self,
+        parent_loader,
+        compile_state,
+        mapper_entities,
+        reconciled_lead_entity,
+        raiseerr,
+    ):
+        # _ClassStrategyLoad
+
+        current_path = compile_state.current_path
+        is_refresh = compile_state.compile_options._for_refresh_state
+
+        if is_refresh and not self.propagate_to_loaders:
+            return []
+
+        # omit setting attributes for a "defaultload" type of option
+        if not self.strategy and not self.local_opts:
+            return []
+
+        effective_path = self.path
+
+        if current_path:
+            new_effective_path = self._adjust_effective_path_for_current_path(
+                effective_path, current_path
+            )
+            if new_effective_path is None:
+                return []
+            effective_path = new_effective_path
+
+        return [("loader", effective_path.natural_path)]
+
+
+def _generate_from_keys(
+    meth: Callable[..., _AbstractLoad],
+    keys: Tuple[_AttrType, ...],
+    chained: bool,
+    kw: Any,
+) -> _AbstractLoad:
+    lead_element: Optional[_AbstractLoad] = None
+
+    attr: Any
+    for is_default, _keys in (True, keys[0:-1]), (False, keys[-1:]):
+        for attr in _keys:
+            if isinstance(attr, str):
+                if attr.startswith("." + _WILDCARD_TOKEN):
+                    util.warn_deprecated(
+                        "The undocumented `.{WILDCARD}` format is "
+                        "deprecated "
+                        "and will be removed in a future version as "
+                        "it is "
+                        "believed to be unused. "
+                        "If you have been using this functionality, "
+                        "please "
+                        "comment on Issue #4390 on the SQLAlchemy project "
+                        "tracker.",
+                        version="1.4",
+                    )
+                    attr = attr[1:]
+
+                if attr == _WILDCARD_TOKEN:
+                    if is_default:
+                        raise sa_exc.ArgumentError(
+                            "Wildcard token cannot be followed by "
+                            "another entity",
+                        )
+
+                    if lead_element is None:
+                        lead_element = _WildcardLoad()
+
+                    lead_element = meth(lead_element, _DEFAULT_TOKEN, **kw)
+
+                else:
+                    raise sa_exc.ArgumentError(
+                        "Strings are not accepted for attribute names in "
+                        "loader options; please use class-bound "
+                        "attributes directly.",
+                    )
+            else:
+                if lead_element is None:
+                    _, lead_entity, _ = _parse_attr_argument(attr)
+                    lead_element = Load(lead_entity)
+
+                if is_default:
+                    if not chained:
+                        lead_element = lead_element.defaultload(attr)
+                    else:
+                        lead_element = meth(
+                            lead_element, attr, _is_chain=True, **kw
+                        )
+                else:
+                    lead_element = meth(lead_element, attr, **kw)
+
+    assert lead_element
+    return lead_element
+
+
+def _parse_attr_argument(
+    attr: _AttrType,
+) -> Tuple[InspectionAttr, _InternalEntityType[Any], MapperProperty[Any]]:
+    """parse an attribute or wildcard argument to produce an
+    :class:`._AbstractLoad` instance.
+
+    This is used by the standalone loader strategy functions like
+    ``joinedload()``, ``defer()``, etc. to produce :class:`_orm.Load` or
+    :class:`._WildcardLoad` objects.
+
+    """
+    try:
+        # TODO: need to figure out this None thing being returned by
+        # inspect(), it should not have None as an option in most cases
+        # if at all
+        insp: InspectionAttr = inspect(attr)  # type: ignore
+    except sa_exc.NoInspectionAvailable as err:
+        raise sa_exc.ArgumentError(
+            "expected ORM mapped attribute for loader strategy argument"
+        ) from err
+
+    lead_entity: _InternalEntityType[Any]
+
+    if insp_is_mapper_property(insp):
+        lead_entity = insp.parent
+        prop = insp
+    elif insp_is_attribute(insp):
+        lead_entity = insp.parent
+        prop = insp.prop
+    else:
+        raise sa_exc.ArgumentError(
+            "expected ORM mapped attribute for loader strategy argument"
+        )
+
+    return insp, lead_entity, prop
+
+
+def loader_unbound_fn(fn: _FN) -> _FN:
+    """decorator that applies docstrings between standalone loader functions
+    and the loader methods on :class:`._AbstractLoad`.
+
+    """
+    bound_fn = getattr(_AbstractLoad, fn.__name__)
+    fn_doc = bound_fn.__doc__
+    bound_fn.__doc__ = f"""Produce a new :class:`_orm.Load` object with the
+:func:`_orm.{fn.__name__}` option applied.
+
+See :func:`_orm.{fn.__name__}` for usage examples.
+
+"""
+
+    fn.__doc__ = fn_doc
+    return fn
+
+
+# standalone functions follow.  docstrings are filled in
+# by the ``@loader_unbound_fn`` decorator.
+
+
+@loader_unbound_fn
+def contains_eager(*keys: _AttrType, **kw: Any) -> _AbstractLoad:
+    return _generate_from_keys(Load.contains_eager, keys, True, kw)
+
+
+@loader_unbound_fn
+def load_only(*attrs: _AttrType, raiseload: bool = False) -> _AbstractLoad:
+    # TODO: attrs against different classes.  we likely have to
+    # add some extra state to Load of some kind
+    _, lead_element, _ = _parse_attr_argument(attrs[0])
+    return Load(lead_element).load_only(*attrs, raiseload=raiseload)
+
+
+@loader_unbound_fn
+def joinedload(*keys: _AttrType, **kw: Any) -> _AbstractLoad:
+    return _generate_from_keys(Load.joinedload, keys, False, kw)
+
+
+@loader_unbound_fn
+def subqueryload(*keys: _AttrType) -> _AbstractLoad:
+    return _generate_from_keys(Load.subqueryload, keys, False, {})
+
+
+@loader_unbound_fn
+def selectinload(
+    *keys: _AttrType, recursion_depth: Optional[int] = None
+) -> _AbstractLoad:
+    return _generate_from_keys(
+        Load.selectinload, keys, False, {"recursion_depth": recursion_depth}
+    )
+
+
+@loader_unbound_fn
+def lazyload(*keys: _AttrType) -> _AbstractLoad:
+    return _generate_from_keys(Load.lazyload, keys, False, {})
+
+
+@loader_unbound_fn
+def immediateload(
+    *keys: _AttrType, recursion_depth: Optional[int] = None
+) -> _AbstractLoad:
+    return _generate_from_keys(
+        Load.immediateload, keys, False, {"recursion_depth": recursion_depth}
+    )
+
+
+@loader_unbound_fn
+def noload(*keys: _AttrType) -> _AbstractLoad:
+    return _generate_from_keys(Load.noload, keys, False, {})
+
+
+@loader_unbound_fn
+def raiseload(*keys: _AttrType, **kw: Any) -> _AbstractLoad:
+    return _generate_from_keys(Load.raiseload, keys, False, kw)
+
+
+@loader_unbound_fn
+def defaultload(*keys: _AttrType) -> _AbstractLoad:
+    return _generate_from_keys(Load.defaultload, keys, False, {})
+
+
+@loader_unbound_fn
+def defer(
+    key: _AttrType, *addl_attrs: _AttrType, raiseload: bool = False
+) -> _AbstractLoad:
+    if addl_attrs:
+        util.warn_deprecated(
+            "The *addl_attrs on orm.defer is deprecated.  Please use "
+            "method chaining in conjunction with defaultload() to "
+            "indicate a path.",
+            version="1.3",
+        )
+
+    if raiseload:
+        kw = {"raiseload": raiseload}
+    else:
+        kw = {}
+
+    return _generate_from_keys(Load.defer, (key,) + addl_attrs, False, kw)
+
+
+@loader_unbound_fn
+def undefer(key: _AttrType, *addl_attrs: _AttrType) -> _AbstractLoad:
+    if addl_attrs:
+        util.warn_deprecated(
+            "The *addl_attrs on orm.undefer is deprecated.  Please use "
+            "method chaining in conjunction with defaultload() to "
+            "indicate a path.",
+            version="1.3",
+        )
+    return _generate_from_keys(Load.undefer, (key,) + addl_attrs, False, {})
+
+
+@loader_unbound_fn
+def undefer_group(name: str) -> _AbstractLoad:
+    element = _WildcardLoad()
+    return element.undefer_group(name)
+
+
+@loader_unbound_fn
+def with_expression(
+    key: _AttrType, expression: _ColumnExpressionArgument[Any]
+) -> _AbstractLoad:
+    return _generate_from_keys(
+        Load.with_expression, (key,), False, {"expression": expression}
+    )
+
+
+@loader_unbound_fn
+def selectin_polymorphic(
+    base_cls: _EntityType[Any], classes: Iterable[Type[Any]]
+) -> _AbstractLoad:
+    ul = Load(base_cls)
+    return ul.selectin_polymorphic(classes)
+
+
+def _raise_for_does_not_link(path, attrname, parent_entity):
+    if len(path) > 1:
+        path_is_of_type = path[-1].entity is not path[-2].mapper.class_
+        if insp_is_aliased_class(parent_entity):
+            parent_entity_str = str(parent_entity)
+        else:
+            parent_entity_str = parent_entity.class_.__name__
+
+        raise sa_exc.ArgumentError(
+            f'ORM mapped entity or attribute "{attrname}" does not '
+            f'link from relationship "{path[-2]}%s".%s'
+            % (
+                f".of_type({path[-1]})" if path_is_of_type else "",
+                (
+                    "  Did you mean to use "
+                    f'"{path[-2]}'
+                    f'.of_type({parent_entity_str})" or "loadopt.options('
+                    f"selectin_polymorphic({path[-2].mapper.class_.__name__}, "
+                    f'[{parent_entity_str}]), ...)" ?'
+                    if not path_is_of_type
+                    and not path[-1].is_aliased_class
+                    and orm_util._entity_corresponds_to(
+                        path.entity, inspect(parent_entity).mapper
+                    )
+                    else ""
+                ),
+            )
+        )
+    else:
+        raise sa_exc.ArgumentError(
+            f'ORM mapped attribute "{attrname}" does not '
+            f'link mapped class "{path[-1]}"'
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/sync.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/sync.py
new file mode 100644
index 00000000..8f85a41a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/sync.py
@@ -0,0 +1,164 @@
+# orm/sync.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+
+"""private module containing functions used for copying data
+between instances based on join conditions.
+
+"""
+
+from __future__ import annotations
+
+from . import exc
+from . import util as orm_util
+from .base import PassiveFlag
+
+
+def populate(
+    source,
+    source_mapper,
+    dest,
+    dest_mapper,
+    synchronize_pairs,
+    uowcommit,
+    flag_cascaded_pks,
+):
+    source_dict = source.dict
+    dest_dict = dest.dict
+
+    for l, r in synchronize_pairs:
+        try:
+            # inline of source_mapper._get_state_attr_by_column
+            prop = source_mapper._columntoproperty[l]
+            value = source.manager[prop.key].impl.get(
+                source, source_dict, PassiveFlag.PASSIVE_OFF
+            )
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(False, source_mapper, l, dest_mapper, r, err)
+
+        try:
+            # inline of dest_mapper._set_state_attr_by_column
+            prop = dest_mapper._columntoproperty[r]
+            dest.manager[prop.key].impl.set(dest, dest_dict, value, None)
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(True, source_mapper, l, dest_mapper, r, err)
+
+        # technically the "r.primary_key" check isn't
+        # needed here, but we check for this condition to limit
+        # how often this logic is invoked for memory/performance
+        # reasons, since we only need this info for a primary key
+        # destination.
+        if (
+            flag_cascaded_pks
+            and l.primary_key
+            and r.primary_key
+            and r.references(l)
+        ):
+            uowcommit.attributes[("pk_cascaded", dest, r)] = True
+
+
+def bulk_populate_inherit_keys(source_dict, source_mapper, synchronize_pairs):
+    # a simplified version of populate() used by bulk insert mode
+    for l, r in synchronize_pairs:
+        try:
+            prop = source_mapper._columntoproperty[l]
+            value = source_dict[prop.key]
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(False, source_mapper, l, source_mapper, r, err)
+
+        try:
+            prop = source_mapper._columntoproperty[r]
+            source_dict[prop.key] = value
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(True, source_mapper, l, source_mapper, r, err)
+
+
+def clear(dest, dest_mapper, synchronize_pairs):
+    for l, r in synchronize_pairs:
+        if (
+            r.primary_key
+            and dest_mapper._get_state_attr_by_column(dest, dest.dict, r)
+            not in orm_util._none_set
+        ):
+            raise AssertionError(
+                f"Dependency rule on column '{l}' "
+                "tried to blank-out primary key "
+                f"column '{r}' on instance '{orm_util.state_str(dest)}'"
+            )
+        try:
+            dest_mapper._set_state_attr_by_column(dest, dest.dict, r, None)
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(True, None, l, dest_mapper, r, err)
+
+
+def update(source, source_mapper, dest, old_prefix, synchronize_pairs):
+    for l, r in synchronize_pairs:
+        try:
+            oldvalue = source_mapper._get_committed_attr_by_column(
+                source.obj(), l
+            )
+            value = source_mapper._get_state_attr_by_column(
+                source, source.dict, l, passive=PassiveFlag.PASSIVE_OFF
+            )
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(False, source_mapper, l, None, r, err)
+        dest[r.key] = value
+        dest[old_prefix + r.key] = oldvalue
+
+
+def populate_dict(source, source_mapper, dict_, synchronize_pairs):
+    for l, r in synchronize_pairs:
+        try:
+            value = source_mapper._get_state_attr_by_column(
+                source, source.dict, l, passive=PassiveFlag.PASSIVE_OFF
+            )
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(False, source_mapper, l, None, r, err)
+
+        dict_[r.key] = value
+
+
+def source_modified(uowcommit, source, source_mapper, synchronize_pairs):
+    """return true if the source object has changes from an old to a
+    new value on the given synchronize pairs
+
+    """
+    for l, r in synchronize_pairs:
+        try:
+            prop = source_mapper._columntoproperty[l]
+        except exc.UnmappedColumnError as err:
+            _raise_col_to_prop(False, source_mapper, l, None, r, err)
+        history = uowcommit.get_attribute_history(
+            source, prop.key, PassiveFlag.PASSIVE_NO_INITIALIZE
+        )
+        if bool(history.deleted):
+            return True
+    else:
+        return False
+
+
+def _raise_col_to_prop(
+    isdest, source_mapper, source_column, dest_mapper, dest_column, err
+):
+    if isdest:
+        raise exc.UnmappedColumnError(
+            "Can't execute sync rule for "
+            "destination column '%s'; mapper '%s' does not map "
+            "this column.  Try using an explicit `foreign_keys` "
+            "collection which does not include this column (or use "
+            "a viewonly=True relation)." % (dest_column, dest_mapper)
+        ) from err
+    else:
+        raise exc.UnmappedColumnError(
+            "Can't execute sync rule for "
+            "source column '%s'; mapper '%s' does not map this "
+            "column.  Try using an explicit `foreign_keys` "
+            "collection which does not include destination column "
+            "'%s' (or use a viewonly=True relation)."
+            % (source_column, source_mapper, dest_column)
+        ) from err
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/unitofwork.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/unitofwork.py
new file mode 100644
index 00000000..80897f29
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/unitofwork.py
@@ -0,0 +1,796 @@
+# orm/unitofwork.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""The internals for the unit of work system.
+
+The session's flush() process passes objects to a contextual object
+here, which assembles flush tasks based on mappers and their properties,
+organizes them in order of dependency, and executes.
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Dict
+from typing import Optional
+from typing import Set
+from typing import TYPE_CHECKING
+
+from . import attributes
+from . import exc as orm_exc
+from . import util as orm_util
+from .. import event
+from .. import util
+from ..util import topological
+
+
+if TYPE_CHECKING:
+    from .dependency import DependencyProcessor
+    from .interfaces import MapperProperty
+    from .mapper import Mapper
+    from .session import Session
+    from .session import SessionTransaction
+    from .state import InstanceState
+
+
+def track_cascade_events(descriptor, prop):
+    """Establish event listeners on object attributes which handle
+    cascade-on-set/append.
+
+    """
+    key = prop.key
+
+    def append(state, item, initiator, **kw):
+        # process "save_update" cascade rules for when
+        # an instance is appended to the list of another instance
+
+        if item is None:
+            return
+
+        sess = state.session
+        if sess:
+            if sess._warn_on_events:
+                sess._flush_warning("collection append")
+
+            prop = state.manager.mapper._props[key]
+            item_state = attributes.instance_state(item)
+
+            if (
+                prop._cascade.save_update
+                and (key == initiator.key)
+                and not sess._contains_state(item_state)
+            ):
+                sess._save_or_update_state(item_state)
+        return item
+
+    def remove(state, item, initiator, **kw):
+        if item is None:
+            return
+
+        sess = state.session
+
+        prop = state.manager.mapper._props[key]
+
+        if sess and sess._warn_on_events:
+            sess._flush_warning(
+                "collection remove"
+                if prop.uselist
+                else "related attribute delete"
+            )
+
+        if (
+            item is not None
+            and item is not attributes.NEVER_SET
+            and item is not attributes.PASSIVE_NO_RESULT
+            and prop._cascade.delete_orphan
+        ):
+            # expunge pending orphans
+            item_state = attributes.instance_state(item)
+
+            if prop.mapper._is_orphan(item_state):
+                if sess and item_state in sess._new:
+                    sess.expunge(item)
+                else:
+                    # the related item may or may not itself be in a
+                    # Session, however the parent for which we are catching
+                    # the event is not in a session, so memoize this on the
+                    # item
+                    item_state._orphaned_outside_of_session = True
+
+    def set_(state, newvalue, oldvalue, initiator, **kw):
+        # process "save_update" cascade rules for when an instance
+        # is attached to another instance
+        if oldvalue is newvalue:
+            return newvalue
+
+        sess = state.session
+        if sess:
+            if sess._warn_on_events:
+                sess._flush_warning("related attribute set")
+
+            prop = state.manager.mapper._props[key]
+            if newvalue is not None:
+                newvalue_state = attributes.instance_state(newvalue)
+                if (
+                    prop._cascade.save_update
+                    and (key == initiator.key)
+                    and not sess._contains_state(newvalue_state)
+                ):
+                    sess._save_or_update_state(newvalue_state)
+
+            if (
+                oldvalue is not None
+                and oldvalue is not attributes.NEVER_SET
+                and oldvalue is not attributes.PASSIVE_NO_RESULT
+                and prop._cascade.delete_orphan
+            ):
+                # possible to reach here with attributes.NEVER_SET ?
+                oldvalue_state = attributes.instance_state(oldvalue)
+
+                if oldvalue_state in sess._new and prop.mapper._is_orphan(
+                    oldvalue_state
+                ):
+                    sess.expunge(oldvalue)
+        return newvalue
+
+    event.listen(
+        descriptor, "append_wo_mutation", append, raw=True, include_key=True
+    )
+    event.listen(
+        descriptor, "append", append, raw=True, retval=True, include_key=True
+    )
+    event.listen(
+        descriptor, "remove", remove, raw=True, retval=True, include_key=True
+    )
+    event.listen(
+        descriptor, "set", set_, raw=True, retval=True, include_key=True
+    )
+
+
+class UOWTransaction:
+    session: Session
+    transaction: SessionTransaction
+    attributes: Dict[str, Any]
+    deps: util.defaultdict[Mapper[Any], Set[DependencyProcessor]]
+    mappers: util.defaultdict[Mapper[Any], Set[InstanceState[Any]]]
+
+    def __init__(self, session: Session):
+        self.session = session
+
+        # dictionary used by external actors to
+        # store arbitrary state information.
+        self.attributes = {}
+
+        # dictionary of mappers to sets of
+        # DependencyProcessors, which are also
+        # set to be part of the sorted flush actions,
+        # which have that mapper as a parent.
+        self.deps = util.defaultdict(set)
+
+        # dictionary of mappers to sets of InstanceState
+        # items pending for flush which have that mapper
+        # as a parent.
+        self.mappers = util.defaultdict(set)
+
+        # a dictionary of Preprocess objects, which gather
+        # additional states impacted by the flush
+        # and determine if a flush action is needed
+        self.presort_actions = {}
+
+        # dictionary of PostSortRec objects, each
+        # one issues work during the flush within
+        # a certain ordering.
+        self.postsort_actions = {}
+
+        # a set of 2-tuples, each containing two
+        # PostSortRec objects where the second
+        # is dependent on the first being executed
+        # first
+        self.dependencies = set()
+
+        # dictionary of InstanceState-> (isdelete, listonly)
+        # tuples, indicating if this state is to be deleted
+        # or insert/updated, or just refreshed
+        self.states = {}
+
+        # tracks InstanceStates which will be receiving
+        # a "post update" call.  Keys are mappers,
+        # values are a set of states and a set of the
+        # columns which should be included in the update.
+        self.post_update_states = util.defaultdict(lambda: (set(), set()))
+
+    @property
+    def has_work(self):
+        return bool(self.states)
+
+    def was_already_deleted(self, state):
+        """Return ``True`` if the given state is expired and was deleted
+        previously.
+        """
+        if state.expired:
+            try:
+                state._load_expired(state, attributes.PASSIVE_OFF)
+            except orm_exc.ObjectDeletedError:
+                self.session._remove_newly_deleted([state])
+                return True
+        return False
+
+    def is_deleted(self, state):
+        """Return ``True`` if the given state is marked as deleted
+        within this uowtransaction."""
+
+        return state in self.states and self.states[state][0]
+
+    def memo(self, key, callable_):
+        if key in self.attributes:
+            return self.attributes[key]
+        else:
+            self.attributes[key] = ret = callable_()
+            return ret
+
+    def remove_state_actions(self, state):
+        """Remove pending actions for a state from the uowtransaction."""
+
+        isdelete = self.states[state][0]
+
+        self.states[state] = (isdelete, True)
+
+    def get_attribute_history(
+        self, state, key, passive=attributes.PASSIVE_NO_INITIALIZE
+    ):
+        """Facade to attributes.get_state_history(), including
+        caching of results."""
+
+        hashkey = ("history", state, key)
+
+        # cache the objects, not the states; the strong reference here
+        # prevents newly loaded objects from being dereferenced during the
+        # flush process
+
+        if hashkey in self.attributes:
+            history, state_history, cached_passive = self.attributes[hashkey]
+            # if the cached lookup was "passive" and now
+            # we want non-passive, do a non-passive lookup and re-cache
+
+            if (
+                not cached_passive & attributes.SQL_OK
+                and passive & attributes.SQL_OK
+            ):
+                impl = state.manager[key].impl
+                history = impl.get_history(
+                    state,
+                    state.dict,
+                    attributes.PASSIVE_OFF
+                    | attributes.LOAD_AGAINST_COMMITTED
+                    | attributes.NO_RAISE,
+                )
+                if history and impl.uses_objects:
+                    state_history = history.as_state()
+                else:
+                    state_history = history
+                self.attributes[hashkey] = (history, state_history, passive)
+        else:
+            impl = state.manager[key].impl
+            # TODO: store the history as (state, object) tuples
+            # so we don't have to keep converting here
+            history = impl.get_history(
+                state,
+                state.dict,
+                passive
+                | attributes.LOAD_AGAINST_COMMITTED
+                | attributes.NO_RAISE,
+            )
+            if history and impl.uses_objects:
+                state_history = history.as_state()
+            else:
+                state_history = history
+            self.attributes[hashkey] = (history, state_history, passive)
+
+        return state_history
+
+    def has_dep(self, processor):
+        return (processor, True) in self.presort_actions
+
+    def register_preprocessor(self, processor, fromparent):
+        key = (processor, fromparent)
+        if key not in self.presort_actions:
+            self.presort_actions[key] = Preprocess(processor, fromparent)
+
+    def register_object(
+        self,
+        state: InstanceState[Any],
+        isdelete: bool = False,
+        listonly: bool = False,
+        cancel_delete: bool = False,
+        operation: Optional[str] = None,
+        prop: Optional[MapperProperty] = None,
+    ) -> bool:
+        if not self.session._contains_state(state):
+            # this condition is normal when objects are registered
+            # as part of a relationship cascade operation.  it should
+            # not occur for the top-level register from Session.flush().
+            if not state.deleted and operation is not None:
+                util.warn(
+                    "Object of type %s not in session, %s operation "
+                    "along '%s' will not proceed"
+                    % (orm_util.state_class_str(state), operation, prop)
+                )
+            return False
+
+        if state not in self.states:
+            mapper = state.manager.mapper
+
+            if mapper not in self.mappers:
+                self._per_mapper_flush_actions(mapper)
+
+            self.mappers[mapper].add(state)
+            self.states[state] = (isdelete, listonly)
+        else:
+            if not listonly and (isdelete or cancel_delete):
+                self.states[state] = (isdelete, False)
+        return True
+
+    def register_post_update(self, state, post_update_cols):
+        mapper = state.manager.mapper.base_mapper
+        states, cols = self.post_update_states[mapper]
+        states.add(state)
+        cols.update(post_update_cols)
+
+    def _per_mapper_flush_actions(self, mapper):
+        saves = SaveUpdateAll(self, mapper.base_mapper)
+        deletes = DeleteAll(self, mapper.base_mapper)
+        self.dependencies.add((saves, deletes))
+
+        for dep in mapper._dependency_processors:
+            dep.per_property_preprocessors(self)
+
+        for prop in mapper.relationships:
+            if prop.viewonly:
+                continue
+            dep = prop._dependency_processor
+            dep.per_property_preprocessors(self)
+
+    @util.memoized_property
+    def _mapper_for_dep(self):
+        """return a dynamic mapping of (Mapper, DependencyProcessor) to
+        True or False, indicating if the DependencyProcessor operates
+        on objects of that Mapper.
+
+        The result is stored in the dictionary persistently once
+        calculated.
+
+        """
+        return util.PopulateDict(
+            lambda tup: tup[0]._props.get(tup[1].key) is tup[1].prop
+        )
+
+    def filter_states_for_dep(self, dep, states):
+        """Filter the given list of InstanceStates to those relevant to the
+        given DependencyProcessor.
+
+        """
+        mapper_for_dep = self._mapper_for_dep
+        return [s for s in states if mapper_for_dep[(s.manager.mapper, dep)]]
+
+    def states_for_mapper_hierarchy(self, mapper, isdelete, listonly):
+        checktup = (isdelete, listonly)
+        for mapper in mapper.base_mapper.self_and_descendants:
+            for state in self.mappers[mapper]:
+                if self.states[state] == checktup:
+                    yield state
+
+    def _generate_actions(self):
+        """Generate the full, unsorted collection of PostSortRecs as
+        well as dependency pairs for this UOWTransaction.
+
+        """
+        # execute presort_actions, until all states
+        # have been processed.   a presort_action might
+        # add new states to the uow.
+        while True:
+            ret = False
+            for action in list(self.presort_actions.values()):
+                if action.execute(self):
+                    ret = True
+            if not ret:
+                break
+
+        # see if the graph of mapper dependencies has cycles.
+        self.cycles = cycles = topological.find_cycles(
+            self.dependencies, list(self.postsort_actions.values())
+        )
+
+        if cycles:
+            # if yes, break the per-mapper actions into
+            # per-state actions
+            convert = {
+                rec: set(rec.per_state_flush_actions(self)) for rec in cycles
+            }
+
+            # rewrite the existing dependencies to point to
+            # the per-state actions for those per-mapper actions
+            # that were broken up.
+            for edge in list(self.dependencies):
+                if (
+                    None in edge
+                    or edge[0].disabled
+                    or edge[1].disabled
+                    or cycles.issuperset(edge)
+                ):
+                    self.dependencies.remove(edge)
+                elif edge[0] in cycles:
+                    self.dependencies.remove(edge)
+                    for dep in convert[edge[0]]:
+                        self.dependencies.add((dep, edge[1]))
+                elif edge[1] in cycles:
+                    self.dependencies.remove(edge)
+                    for dep in convert[edge[1]]:
+                        self.dependencies.add((edge[0], dep))
+
+        return {
+            a for a in self.postsort_actions.values() if not a.disabled
+        }.difference(cycles)
+
+    def execute(self) -> None:
+        postsort_actions = self._generate_actions()
+
+        postsort_actions = sorted(
+            postsort_actions,
+            key=lambda item: item.sort_key,
+        )
+        # sort = topological.sort(self.dependencies, postsort_actions)
+        # print "--------------"
+        # print "\ndependencies:", self.dependencies
+        # print "\ncycles:", self.cycles
+        # print "\nsort:", list(sort)
+        # print "\nCOUNT OF POSTSORT ACTIONS", len(postsort_actions)
+
+        # execute
+        if self.cycles:
+            for subset in topological.sort_as_subsets(
+                self.dependencies, postsort_actions
+            ):
+                set_ = set(subset)
+                while set_:
+                    n = set_.pop()
+                    n.execute_aggregate(self, set_)
+        else:
+            for rec in topological.sort(self.dependencies, postsort_actions):
+                rec.execute(self)
+
+    def finalize_flush_changes(self) -> None:
+        """Mark processed objects as clean / deleted after a successful
+        flush().
+
+        This method is called within the flush() method after the
+        execute() method has succeeded and the transaction has been committed.
+
+        """
+        if not self.states:
+            return
+
+        states = set(self.states)
+        isdel = {
+            s for (s, (isdelete, listonly)) in self.states.items() if isdelete
+        }
+        other = states.difference(isdel)
+        if isdel:
+            self.session._remove_newly_deleted(isdel)
+        if other:
+            self.session._register_persistent(other)
+
+
+class IterateMappersMixin:
+    __slots__ = ()
+
+    def _mappers(self, uow):
+        if self.fromparent:
+            return iter(
+                m
+                for m in self.dependency_processor.parent.self_and_descendants
+                if uow._mapper_for_dep[(m, self.dependency_processor)]
+            )
+        else:
+            return self.dependency_processor.mapper.self_and_descendants
+
+
+class Preprocess(IterateMappersMixin):
+    __slots__ = (
+        "dependency_processor",
+        "fromparent",
+        "processed",
+        "setup_flush_actions",
+    )
+
+    def __init__(self, dependency_processor, fromparent):
+        self.dependency_processor = dependency_processor
+        self.fromparent = fromparent
+        self.processed = set()
+        self.setup_flush_actions = False
+
+    def execute(self, uow):
+        delete_states = set()
+        save_states = set()
+
+        for mapper in self._mappers(uow):
+            for state in uow.mappers[mapper].difference(self.processed):
+                (isdelete, listonly) = uow.states[state]
+                if not listonly:
+                    if isdelete:
+                        delete_states.add(state)
+                    else:
+                        save_states.add(state)
+
+        if delete_states:
+            self.dependency_processor.presort_deletes(uow, delete_states)
+            self.processed.update(delete_states)
+        if save_states:
+            self.dependency_processor.presort_saves(uow, save_states)
+            self.processed.update(save_states)
+
+        if delete_states or save_states:
+            if not self.setup_flush_actions and (
+                self.dependency_processor.prop_has_changes(
+                    uow, delete_states, True
+                )
+                or self.dependency_processor.prop_has_changes(
+                    uow, save_states, False
+                )
+            ):
+                self.dependency_processor.per_property_flush_actions(uow)
+                self.setup_flush_actions = True
+            return True
+        else:
+            return False
+
+
+class PostSortRec:
+    __slots__ = ("disabled",)
+
+    def __new__(cls, uow, *args):
+        key = (cls,) + args
+        if key in uow.postsort_actions:
+            return uow.postsort_actions[key]
+        else:
+            uow.postsort_actions[key] = ret = object.__new__(cls)
+            ret.disabled = False
+            return ret
+
+    def execute_aggregate(self, uow, recs):
+        self.execute(uow)
+
+
+class ProcessAll(IterateMappersMixin, PostSortRec):
+    __slots__ = "dependency_processor", "isdelete", "fromparent", "sort_key"
+
+    def __init__(self, uow, dependency_processor, isdelete, fromparent):
+        self.dependency_processor = dependency_processor
+        self.sort_key = (
+            "ProcessAll",
+            self.dependency_processor.sort_key,
+            isdelete,
+        )
+        self.isdelete = isdelete
+        self.fromparent = fromparent
+        uow.deps[dependency_processor.parent.base_mapper].add(
+            dependency_processor
+        )
+
+    def execute(self, uow):
+        states = self._elements(uow)
+        if self.isdelete:
+            self.dependency_processor.process_deletes(uow, states)
+        else:
+            self.dependency_processor.process_saves(uow, states)
+
+    def per_state_flush_actions(self, uow):
+        # this is handled by SaveUpdateAll and DeleteAll,
+        # since a ProcessAll should unconditionally be pulled
+        # into per-state if either the parent/child mappers
+        # are part of a cycle
+        return iter([])
+
+    def __repr__(self):
+        return "%s(%s, isdelete=%s)" % (
+            self.__class__.__name__,
+            self.dependency_processor,
+            self.isdelete,
+        )
+
+    def _elements(self, uow):
+        for mapper in self._mappers(uow):
+            for state in uow.mappers[mapper]:
+                (isdelete, listonly) = uow.states[state]
+                if isdelete == self.isdelete and not listonly:
+                    yield state
+
+
+class PostUpdateAll(PostSortRec):
+    __slots__ = "mapper", "isdelete", "sort_key"
+
+    def __init__(self, uow, mapper, isdelete):
+        self.mapper = mapper
+        self.isdelete = isdelete
+        self.sort_key = ("PostUpdateAll", mapper._sort_key, isdelete)
+
+    @util.preload_module("sqlalchemy.orm.persistence")
+    def execute(self, uow):
+        persistence = util.preloaded.orm_persistence
+        states, cols = uow.post_update_states[self.mapper]
+        states = [s for s in states if uow.states[s][0] == self.isdelete]
+
+        persistence.post_update(self.mapper, states, uow, cols)
+
+
+class SaveUpdateAll(PostSortRec):
+    __slots__ = ("mapper", "sort_key")
+
+    def __init__(self, uow, mapper):
+        self.mapper = mapper
+        self.sort_key = ("SaveUpdateAll", mapper._sort_key)
+        assert mapper is mapper.base_mapper
+
+    @util.preload_module("sqlalchemy.orm.persistence")
+    def execute(self, uow):
+        util.preloaded.orm_persistence.save_obj(
+            self.mapper,
+            uow.states_for_mapper_hierarchy(self.mapper, False, False),
+            uow,
+        )
+
+    def per_state_flush_actions(self, uow):
+        states = list(
+            uow.states_for_mapper_hierarchy(self.mapper, False, False)
+        )
+        base_mapper = self.mapper.base_mapper
+        delete_all = DeleteAll(uow, base_mapper)
+        for state in states:
+            # keep saves before deletes -
+            # this ensures 'row switch' operations work
+            action = SaveUpdateState(uow, state)
+            uow.dependencies.add((action, delete_all))
+            yield action
+
+        for dep in uow.deps[self.mapper]:
+            states_for_prop = uow.filter_states_for_dep(dep, states)
+            dep.per_state_flush_actions(uow, states_for_prop, False)
+
+    def __repr__(self):
+        return "%s(%s)" % (self.__class__.__name__, self.mapper)
+
+
+class DeleteAll(PostSortRec):
+    __slots__ = ("mapper", "sort_key")
+
+    def __init__(self, uow, mapper):
+        self.mapper = mapper
+        self.sort_key = ("DeleteAll", mapper._sort_key)
+        assert mapper is mapper.base_mapper
+
+    @util.preload_module("sqlalchemy.orm.persistence")
+    def execute(self, uow):
+        util.preloaded.orm_persistence.delete_obj(
+            self.mapper,
+            uow.states_for_mapper_hierarchy(self.mapper, True, False),
+            uow,
+        )
+
+    def per_state_flush_actions(self, uow):
+        states = list(
+            uow.states_for_mapper_hierarchy(self.mapper, True, False)
+        )
+        base_mapper = self.mapper.base_mapper
+        save_all = SaveUpdateAll(uow, base_mapper)
+        for state in states:
+            # keep saves before deletes -
+            # this ensures 'row switch' operations work
+            action = DeleteState(uow, state)
+            uow.dependencies.add((save_all, action))
+            yield action
+
+        for dep in uow.deps[self.mapper]:
+            states_for_prop = uow.filter_states_for_dep(dep, states)
+            dep.per_state_flush_actions(uow, states_for_prop, True)
+
+    def __repr__(self):
+        return "%s(%s)" % (self.__class__.__name__, self.mapper)
+
+
+class ProcessState(PostSortRec):
+    __slots__ = "dependency_processor", "isdelete", "state", "sort_key"
+
+    def __init__(self, uow, dependency_processor, isdelete, state):
+        self.dependency_processor = dependency_processor
+        self.sort_key = ("ProcessState", dependency_processor.sort_key)
+        self.isdelete = isdelete
+        self.state = state
+
+    def execute_aggregate(self, uow, recs):
+        cls_ = self.__class__
+        dependency_processor = self.dependency_processor
+        isdelete = self.isdelete
+        our_recs = [
+            r
+            for r in recs
+            if r.__class__ is cls_
+            and r.dependency_processor is dependency_processor
+            and r.isdelete is isdelete
+        ]
+        recs.difference_update(our_recs)
+        states = [self.state] + [r.state for r in our_recs]
+        if isdelete:
+            dependency_processor.process_deletes(uow, states)
+        else:
+            dependency_processor.process_saves(uow, states)
+
+    def __repr__(self):
+        return "%s(%s, %s, delete=%s)" % (
+            self.__class__.__name__,
+            self.dependency_processor,
+            orm_util.state_str(self.state),
+            self.isdelete,
+        )
+
+
+class SaveUpdateState(PostSortRec):
+    __slots__ = "state", "mapper", "sort_key"
+
+    def __init__(self, uow, state):
+        self.state = state
+        self.mapper = state.mapper.base_mapper
+        self.sort_key = ("ProcessState", self.mapper._sort_key)
+
+    @util.preload_module("sqlalchemy.orm.persistence")
+    def execute_aggregate(self, uow, recs):
+        persistence = util.preloaded.orm_persistence
+        cls_ = self.__class__
+        mapper = self.mapper
+        our_recs = [
+            r for r in recs if r.__class__ is cls_ and r.mapper is mapper
+        ]
+        recs.difference_update(our_recs)
+        persistence.save_obj(
+            mapper, [self.state] + [r.state for r in our_recs], uow
+        )
+
+    def __repr__(self):
+        return "%s(%s)" % (
+            self.__class__.__name__,
+            orm_util.state_str(self.state),
+        )
+
+
+class DeleteState(PostSortRec):
+    __slots__ = "state", "mapper", "sort_key"
+
+    def __init__(self, uow, state):
+        self.state = state
+        self.mapper = state.mapper.base_mapper
+        self.sort_key = ("DeleteState", self.mapper._sort_key)
+
+    @util.preload_module("sqlalchemy.orm.persistence")
+    def execute_aggregate(self, uow, recs):
+        persistence = util.preloaded.orm_persistence
+        cls_ = self.__class__
+        mapper = self.mapper
+        our_recs = [
+            r for r in recs if r.__class__ is cls_ and r.mapper is mapper
+        ]
+        recs.difference_update(our_recs)
+        states = [self.state] + [r.state for r in our_recs]
+        persistence.delete_obj(
+            mapper, [s for s in states if uow.states[s][0]], uow
+        )
+
+    def __repr__(self):
+        return "%s(%s)" % (
+            self.__class__.__name__,
+            orm_util.state_str(self.state),
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/util.py
new file mode 100644
index 00000000..48282b2d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/util.py
@@ -0,0 +1,2402 @@
+# orm/util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import enum
+import functools
+import re
+import types
+import typing
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Match
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import attributes  # noqa
+from . import exc
+from ._typing import _O
+from ._typing import insp_is_aliased_class
+from ._typing import insp_is_mapper
+from ._typing import prop_is_relationship
+from .base import _class_to_mapper as _class_to_mapper
+from .base import _MappedAnnotationBase
+from .base import _never_set as _never_set  # noqa: F401
+from .base import _none_only_set as _none_only_set  # noqa: F401
+from .base import _none_set as _none_set  # noqa: F401
+from .base import attribute_str as attribute_str  # noqa: F401
+from .base import class_mapper as class_mapper
+from .base import DynamicMapped
+from .base import InspectionAttr as InspectionAttr
+from .base import instance_str as instance_str  # noqa: F401
+from .base import Mapped
+from .base import object_mapper as object_mapper
+from .base import object_state as object_state  # noqa: F401
+from .base import opt_manager_of_class
+from .base import ORMDescriptor
+from .base import state_attribute_str as state_attribute_str  # noqa: F401
+from .base import state_class_str as state_class_str  # noqa: F401
+from .base import state_str as state_str  # noqa: F401
+from .base import WriteOnlyMapped
+from .interfaces import CriteriaOption
+from .interfaces import MapperProperty as MapperProperty
+from .interfaces import ORMColumnsClauseRole
+from .interfaces import ORMEntityColumnsClauseRole
+from .interfaces import ORMFromClauseRole
+from .path_registry import PathRegistry as PathRegistry
+from .. import event
+from .. import exc as sa_exc
+from .. import inspection
+from .. import sql
+from .. import util
+from ..engine.result import result_tuple
+from ..sql import coercions
+from ..sql import expression
+from ..sql import lambdas
+from ..sql import roles
+from ..sql import util as sql_util
+from ..sql import visitors
+from ..sql._typing import is_selectable
+from ..sql.annotation import SupportsCloneAnnotations
+from ..sql.base import ColumnCollection
+from ..sql.cache_key import HasCacheKey
+from ..sql.cache_key import MemoizedHasCacheKey
+from ..sql.elements import ColumnElement
+from ..sql.elements import KeyedColumnElement
+from ..sql.selectable import FromClause
+from ..util.langhelpers import MemoizedSlots
+from ..util.typing import de_stringify_annotation as _de_stringify_annotation
+from ..util.typing import eval_name_only as _eval_name_only
+from ..util.typing import fixup_container_fwd_refs
+from ..util.typing import get_origin
+from ..util.typing import is_origin_of_cls
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EntityType
+    from ._typing import _IdentityKeyType
+    from ._typing import _InternalEntityType
+    from ._typing import _ORMCOLEXPR
+    from .context import _MapperEntity
+    from .context import ORMCompileState
+    from .mapper import Mapper
+    from .path_registry import AbstractEntityRegistry
+    from .query import Query
+    from .relationships import RelationshipProperty
+    from ..engine import Row
+    from ..engine import RowMapping
+    from ..sql._typing import _CE
+    from ..sql._typing import _ColumnExpressionArgument
+    from ..sql._typing import _EquivalentColumnMap
+    from ..sql._typing import _FromClauseArgument
+    from ..sql._typing import _OnClauseArgument
+    from ..sql._typing import _PropagateAttrsType
+    from ..sql.annotation import _SA
+    from ..sql.base import ReadOnlyColumnCollection
+    from ..sql.elements import BindParameter
+    from ..sql.selectable import _ColumnsClauseElement
+    from ..sql.selectable import Select
+    from ..sql.selectable import Selectable
+    from ..sql.visitors import anon_map
+    from ..util.typing import _AnnotationScanType
+
+_T = TypeVar("_T", bound=Any)
+
+all_cascades = frozenset(
+    (
+        "delete",
+        "delete-orphan",
+        "all",
+        "merge",
+        "expunge",
+        "save-update",
+        "refresh-expire",
+        "none",
+    )
+)
+
+_de_stringify_partial = functools.partial(
+    functools.partial,
+    locals_=util.immutabledict(
+        {
+            "Mapped": Mapped,
+            "WriteOnlyMapped": WriteOnlyMapped,
+            "DynamicMapped": DynamicMapped,
+        }
+    ),
+)
+
+# partial is practically useless as we have to write out the whole
+# function and maintain the signature anyway
+
+
+class _DeStringifyAnnotation(Protocol):
+    def __call__(
+        self,
+        cls: Type[Any],
+        annotation: _AnnotationScanType,
+        originating_module: str,
+        *,
+        str_cleanup_fn: Optional[Callable[[str, str], str]] = None,
+        include_generic: bool = False,
+    ) -> Type[Any]: ...
+
+
+de_stringify_annotation = cast(
+    _DeStringifyAnnotation, _de_stringify_partial(_de_stringify_annotation)
+)
+
+
+class _EvalNameOnly(Protocol):
+    def __call__(self, name: str, module_name: str) -> Any: ...
+
+
+eval_name_only = cast(_EvalNameOnly, _de_stringify_partial(_eval_name_only))
+
+
+class CascadeOptions(FrozenSet[str]):
+    """Keeps track of the options sent to
+    :paramref:`.relationship.cascade`"""
+
+    _add_w_all_cascades = all_cascades.difference(
+        ["all", "none", "delete-orphan"]
+    )
+    _allowed_cascades = all_cascades
+
+    _viewonly_cascades = ["expunge", "all", "none", "refresh-expire", "merge"]
+
+    __slots__ = (
+        "save_update",
+        "delete",
+        "refresh_expire",
+        "merge",
+        "expunge",
+        "delete_orphan",
+    )
+
+    save_update: bool
+    delete: bool
+    refresh_expire: bool
+    merge: bool
+    expunge: bool
+    delete_orphan: bool
+
+    def __new__(
+        cls, value_list: Optional[Union[Iterable[str], str]]
+    ) -> CascadeOptions:
+        if isinstance(value_list, str) or value_list is None:
+            return cls.from_string(value_list)  # type: ignore
+        values = set(value_list)
+        if values.difference(cls._allowed_cascades):
+            raise sa_exc.ArgumentError(
+                "Invalid cascade option(s): %s"
+                % ", ".join(
+                    [
+                        repr(x)
+                        for x in sorted(
+                            values.difference(cls._allowed_cascades)
+                        )
+                    ]
+                )
+            )
+
+        if "all" in values:
+            values.update(cls._add_w_all_cascades)
+        if "none" in values:
+            values.clear()
+        values.discard("all")
+
+        self = super().__new__(cls, values)
+        self.save_update = "save-update" in values
+        self.delete = "delete" in values
+        self.refresh_expire = "refresh-expire" in values
+        self.merge = "merge" in values
+        self.expunge = "expunge" in values
+        self.delete_orphan = "delete-orphan" in values
+
+        if self.delete_orphan and not self.delete:
+            util.warn("The 'delete-orphan' cascade option requires 'delete'.")
+        return self
+
+    def __repr__(self):
+        return "CascadeOptions(%r)" % (",".join([x for x in sorted(self)]))
+
+    @classmethod
+    def from_string(cls, arg):
+        values = [c for c in re.split(r"\s*,\s*", arg or "") if c]
+        return cls(values)
+
+
+def _validator_events(desc, key, validator, include_removes, include_backrefs):
+    """Runs a validation method on an attribute value to be set or
+    appended.
+    """
+
+    if not include_backrefs:
+
+        def detect_is_backref(state, initiator):
+            impl = state.manager[key].impl
+            return initiator.impl is not impl
+
+    if include_removes:
+
+        def append(state, value, initiator):
+            if initiator.op is not attributes.OP_BULK_REPLACE and (
+                include_backrefs or not detect_is_backref(state, initiator)
+            ):
+                return validator(state.obj(), key, value, False)
+            else:
+                return value
+
+        def bulk_set(state, values, initiator):
+            if include_backrefs or not detect_is_backref(state, initiator):
+                obj = state.obj()
+                values[:] = [
+                    validator(obj, key, value, False) for value in values
+                ]
+
+        def set_(state, value, oldvalue, initiator):
+            if include_backrefs or not detect_is_backref(state, initiator):
+                return validator(state.obj(), key, value, False)
+            else:
+                return value
+
+        def remove(state, value, initiator):
+            if include_backrefs or not detect_is_backref(state, initiator):
+                validator(state.obj(), key, value, True)
+
+    else:
+
+        def append(state, value, initiator):
+            if initiator.op is not attributes.OP_BULK_REPLACE and (
+                include_backrefs or not detect_is_backref(state, initiator)
+            ):
+                return validator(state.obj(), key, value)
+            else:
+                return value
+
+        def bulk_set(state, values, initiator):
+            if include_backrefs or not detect_is_backref(state, initiator):
+                obj = state.obj()
+                values[:] = [validator(obj, key, value) for value in values]
+
+        def set_(state, value, oldvalue, initiator):
+            if include_backrefs or not detect_is_backref(state, initiator):
+                return validator(state.obj(), key, value)
+            else:
+                return value
+
+    event.listen(desc, "append", append, raw=True, retval=True)
+    event.listen(desc, "bulk_replace", bulk_set, raw=True)
+    event.listen(desc, "set", set_, raw=True, retval=True)
+    if include_removes:
+        event.listen(desc, "remove", remove, raw=True, retval=True)
+
+
+def polymorphic_union(
+    table_map, typecolname, aliasname="p_union", cast_nulls=True
+):
+    """Create a ``UNION`` statement used by a polymorphic mapper.
+
+    See  :ref:`concrete_inheritance` for an example of how
+    this is used.
+
+    :param table_map: mapping of polymorphic identities to
+     :class:`_schema.Table` objects.
+    :param typecolname: string name of a "discriminator" column, which will be
+     derived from the query, producing the polymorphic identity for
+     each row.  If ``None``, no polymorphic discriminator is generated.
+    :param aliasname: name of the :func:`~sqlalchemy.sql.expression.alias()`
+     construct generated.
+    :param cast_nulls: if True, non-existent columns, which are represented
+     as labeled NULLs, will be passed into CAST.   This is a legacy behavior
+     that is problematic on some backends such as Oracle - in which case it
+     can be set to False.
+
+    """
+
+    colnames: util.OrderedSet[str] = util.OrderedSet()
+    colnamemaps = {}
+    types = {}
+    for key in table_map:
+        table = table_map[key]
+
+        table = coercions.expect(
+            roles.StrictFromClauseRole, table, allow_select=True
+        )
+        table_map[key] = table
+
+        m = {}
+        for c in table.c:
+            if c.key == typecolname:
+                raise sa_exc.InvalidRequestError(
+                    "Polymorphic union can't use '%s' as the discriminator "
+                    "column due to mapped column %r; please apply the "
+                    "'typecolname' "
+                    "argument; this is available on "
+                    "ConcreteBase as '_concrete_discriminator_name'"
+                    % (typecolname, c)
+                )
+            colnames.add(c.key)
+            m[c.key] = c
+            types[c.key] = c.type
+        colnamemaps[table] = m
+
+    def col(name, table):
+        try:
+            return colnamemaps[table][name]
+        except KeyError:
+            if cast_nulls:
+                return sql.cast(sql.null(), types[name]).label(name)
+            else:
+                return sql.type_coerce(sql.null(), types[name]).label(name)
+
+    result = []
+    for type_, table in table_map.items():
+        if typecolname is not None:
+            result.append(
+                sql.select(
+                    *(
+                        [col(name, table) for name in colnames]
+                        + [
+                            sql.literal_column(
+                                sql_util._quote_ddl_expr(type_)
+                            ).label(typecolname)
+                        ]
+                    )
+                ).select_from(table)
+            )
+        else:
+            result.append(
+                sql.select(
+                    *[col(name, table) for name in colnames]
+                ).select_from(table)
+            )
+    return sql.union_all(*result).alias(aliasname)
+
+
+def identity_key(
+    class_: Optional[Type[_T]] = None,
+    ident: Union[Any, Tuple[Any, ...]] = None,
+    *,
+    instance: Optional[_T] = None,
+    row: Optional[Union[Row[Any], RowMapping]] = None,
+    identity_token: Optional[Any] = None,
+) -> _IdentityKeyType[_T]:
+    r"""Generate "identity key" tuples, as are used as keys in the
+    :attr:`.Session.identity_map` dictionary.
+
+    This function has several call styles:
+
+    * ``identity_key(class, ident, identity_token=token)``
+
+      This form receives a mapped class and a primary key scalar or
+      tuple as an argument.
+
+      E.g.::
+
+        >>> identity_key(MyClass, (1, 2))
+        (<class '__main__.MyClass'>, (1, 2), None)
+
+      :param class: mapped class (must be a positional argument)
+      :param ident: primary key, may be a scalar or tuple argument.
+      :param identity_token: optional identity token
+
+        .. versionadded:: 1.2 added identity_token
+
+
+    * ``identity_key(instance=instance)``
+
+      This form will produce the identity key for a given instance.  The
+      instance need not be persistent, only that its primary key attributes
+      are populated (else the key will contain ``None`` for those missing
+      values).
+
+      E.g.::
+
+        >>> instance = MyClass(1, 2)
+        >>> identity_key(instance=instance)
+        (<class '__main__.MyClass'>, (1, 2), None)
+
+      In this form, the given instance is ultimately run though
+      :meth:`_orm.Mapper.identity_key_from_instance`, which will have the
+      effect of performing a database check for the corresponding row
+      if the object is expired.
+
+      :param instance: object instance (must be given as a keyword arg)
+
+    * ``identity_key(class, row=row, identity_token=token)``
+
+      This form is similar to the class/tuple form, except is passed a
+      database result row as a :class:`.Row` or :class:`.RowMapping` object.
+
+      E.g.::
+
+        >>> row = engine.execute(text("select * from table where a=1 and b=2")).first()
+        >>> identity_key(MyClass, row=row)
+        (<class '__main__.MyClass'>, (1, 2), None)
+
+      :param class: mapped class (must be a positional argument)
+      :param row: :class:`.Row` row returned by a :class:`_engine.CursorResult`
+       (must be given as a keyword arg)
+      :param identity_token: optional identity token
+
+        .. versionadded:: 1.2 added identity_token
+
+    """  # noqa: E501
+    if class_ is not None:
+        mapper = class_mapper(class_)
+        if row is None:
+            if ident is None:
+                raise sa_exc.ArgumentError("ident or row is required")
+            return mapper.identity_key_from_primary_key(
+                tuple(util.to_list(ident)), identity_token=identity_token
+            )
+        else:
+            return mapper.identity_key_from_row(
+                row, identity_token=identity_token
+            )
+    elif instance is not None:
+        mapper = object_mapper(instance)
+        return mapper.identity_key_from_instance(instance)
+    else:
+        raise sa_exc.ArgumentError("class or instance is required")
+
+
+class _TraceAdaptRole(enum.Enum):
+    """Enumeration of all the use cases for ORMAdapter.
+
+    ORMAdapter remains one of the most complicated aspects of the ORM, as it is
+    used for in-place adaption of column expressions to be applied to a SELECT,
+    replacing :class:`.Table` and other objects that are mapped to classes with
+    aliases of those tables in the case of joined eager loading, or in the case
+    of polymorphic loading as used with concrete mappings or other custom "with
+    polymorphic" parameters, with whole user-defined subqueries. The
+    enumerations provide an overview of all the use cases used by ORMAdapter, a
+    layer of formality as to the introduction of new ORMAdapter use cases (of
+    which none are anticipated), as well as a means to trace the origins of a
+    particular ORMAdapter within runtime debugging.
+
+    SQLAlchemy 2.0 has greatly scaled back ORM features which relied heavily on
+    open-ended statement adaption, including the ``Query.with_polymorphic()``
+    method and the ``Query.select_from_entity()`` methods, favoring
+    user-explicit aliasing schemes using the ``aliased()`` and
+    ``with_polymorphic()`` standalone constructs; these still use adaption,
+    however the adaption is applied in a narrower scope.
+
+    """
+
+    # aliased() use that is used to adapt individual attributes at query
+    # construction time
+    ALIASED_INSP = enum.auto()
+
+    # joinedload cases; typically adapt an ON clause of a relationship
+    # join
+    JOINEDLOAD_USER_DEFINED_ALIAS = enum.auto()
+    JOINEDLOAD_PATH_WITH_POLYMORPHIC = enum.auto()
+    JOINEDLOAD_MEMOIZED_ADAPTER = enum.auto()
+
+    # polymorphic cases - these are complex ones that replace FROM
+    # clauses, replacing tables with subqueries
+    MAPPER_POLYMORPHIC_ADAPTER = enum.auto()
+    WITH_POLYMORPHIC_ADAPTER = enum.auto()
+    WITH_POLYMORPHIC_ADAPTER_RIGHT_JOIN = enum.auto()
+    DEPRECATED_JOIN_ADAPT_RIGHT_SIDE = enum.auto()
+
+    # the from_statement() case, used only to adapt individual attributes
+    # from a given statement to local ORM attributes at result fetching
+    # time.  assigned to ORMCompileState._from_obj_alias
+    ADAPT_FROM_STATEMENT = enum.auto()
+
+    # the joinedload for queries that have LIMIT/OFFSET/DISTINCT case;
+    # the query is placed inside of a subquery with the LIMIT/OFFSET/etc.,
+    # joinedloads are then placed on the outside.
+    # assigned to ORMCompileState.compound_eager_adapter
+    COMPOUND_EAGER_STATEMENT = enum.auto()
+
+    # the legacy Query._set_select_from() case.
+    # this is needed for Query's set operations (i.e. UNION, etc. )
+    # as well as "legacy from_self()", which while removed from 2.0 as
+    # public API, is used for the Query.count() method.  this one
+    # still does full statement traversal
+    # assigned to ORMCompileState._from_obj_alias
+    LEGACY_SELECT_FROM_ALIAS = enum.auto()
+
+
+class ORMStatementAdapter(sql_util.ColumnAdapter):
+    """ColumnAdapter which includes a role attribute."""
+
+    __slots__ = ("role",)
+
+    def __init__(
+        self,
+        role: _TraceAdaptRole,
+        selectable: Selectable,
+        *,
+        equivalents: Optional[_EquivalentColumnMap] = None,
+        adapt_required: bool = False,
+        allow_label_resolve: bool = True,
+        anonymize_labels: bool = False,
+        adapt_on_names: bool = False,
+        adapt_from_selectables: Optional[AbstractSet[FromClause]] = None,
+    ):
+        self.role = role
+        super().__init__(
+            selectable,
+            equivalents=equivalents,
+            adapt_required=adapt_required,
+            allow_label_resolve=allow_label_resolve,
+            anonymize_labels=anonymize_labels,
+            adapt_on_names=adapt_on_names,
+            adapt_from_selectables=adapt_from_selectables,
+        )
+
+
+class ORMAdapter(sql_util.ColumnAdapter):
+    """ColumnAdapter subclass which excludes adaptation of entities from
+    non-matching mappers.
+
+    """
+
+    __slots__ = ("role", "mapper", "is_aliased_class", "aliased_insp")
+
+    is_aliased_class: bool
+    aliased_insp: Optional[AliasedInsp[Any]]
+
+    def __init__(
+        self,
+        role: _TraceAdaptRole,
+        entity: _InternalEntityType[Any],
+        *,
+        equivalents: Optional[_EquivalentColumnMap] = None,
+        adapt_required: bool = False,
+        allow_label_resolve: bool = True,
+        anonymize_labels: bool = False,
+        selectable: Optional[Selectable] = None,
+        limit_on_entity: bool = True,
+        adapt_on_names: bool = False,
+        adapt_from_selectables: Optional[AbstractSet[FromClause]] = None,
+    ):
+        self.role = role
+        self.mapper = entity.mapper
+        if selectable is None:
+            selectable = entity.selectable
+        if insp_is_aliased_class(entity):
+            self.is_aliased_class = True
+            self.aliased_insp = entity
+        else:
+            self.is_aliased_class = False
+            self.aliased_insp = None
+
+        super().__init__(
+            selectable,
+            equivalents,
+            adapt_required=adapt_required,
+            allow_label_resolve=allow_label_resolve,
+            anonymize_labels=anonymize_labels,
+            include_fn=self._include_fn if limit_on_entity else None,
+            adapt_on_names=adapt_on_names,
+            adapt_from_selectables=adapt_from_selectables,
+        )
+
+    def _include_fn(self, elem):
+        entity = elem._annotations.get("parentmapper", None)
+
+        return not entity or entity.isa(self.mapper) or self.mapper.isa(entity)
+
+
+class AliasedClass(
+    inspection.Inspectable["AliasedInsp[_O]"], ORMColumnsClauseRole[_O]
+):
+    r"""Represents an "aliased" form of a mapped class for usage with Query.
+
+    The ORM equivalent of a :func:`~sqlalchemy.sql.expression.alias`
+    construct, this object mimics the mapped class using a
+    ``__getattr__`` scheme and maintains a reference to a
+    real :class:`~sqlalchemy.sql.expression.Alias` object.
+
+    A primary purpose of :class:`.AliasedClass` is to serve as an alternate
+    within a SQL statement generated by the ORM, such that an existing
+    mapped entity can be used in multiple contexts.   A simple example::
+
+        # find all pairs of users with the same name
+        user_alias = aliased(User)
+        session.query(User, user_alias).join(
+            (user_alias, User.id > user_alias.id)
+        ).filter(User.name == user_alias.name)
+
+    :class:`.AliasedClass` is also capable of mapping an existing mapped
+    class to an entirely new selectable, provided this selectable is column-
+    compatible with the existing mapped selectable, and it can also be
+    configured in a mapping as the target of a :func:`_orm.relationship`.
+    See the links below for examples.
+
+    The :class:`.AliasedClass` object is constructed typically using the
+    :func:`_orm.aliased` function.   It also is produced with additional
+    configuration when using the :func:`_orm.with_polymorphic` function.
+
+    The resulting object is an instance of :class:`.AliasedClass`.
+    This object implements an attribute scheme which produces the
+    same attribute and method interface as the original mapped
+    class, allowing :class:`.AliasedClass` to be compatible
+    with any attribute technique which works on the original class,
+    including hybrid attributes (see :ref:`hybrids_toplevel`).
+
+    The :class:`.AliasedClass` can be inspected for its underlying
+    :class:`_orm.Mapper`, aliased selectable, and other information
+    using :func:`_sa.inspect`::
+
+        from sqlalchemy import inspect
+
+        my_alias = aliased(MyClass)
+        insp = inspect(my_alias)
+
+    The resulting inspection object is an instance of :class:`.AliasedInsp`.
+
+
+    .. seealso::
+
+        :func:`.aliased`
+
+        :func:`.with_polymorphic`
+
+        :ref:`relationship_aliased_class`
+
+        :ref:`relationship_to_window_function`
+
+
+    """
+
+    __name__: str
+
+    def __init__(
+        self,
+        mapped_class_or_ac: _EntityType[_O],
+        alias: Optional[FromClause] = None,
+        name: Optional[str] = None,
+        flat: bool = False,
+        adapt_on_names: bool = False,
+        with_polymorphic_mappers: Optional[Sequence[Mapper[Any]]] = None,
+        with_polymorphic_discriminator: Optional[ColumnElement[Any]] = None,
+        base_alias: Optional[AliasedInsp[Any]] = None,
+        use_mapper_path: bool = False,
+        represents_outer_join: bool = False,
+    ):
+        insp = cast(
+            "_InternalEntityType[_O]", inspection.inspect(mapped_class_or_ac)
+        )
+        mapper = insp.mapper
+
+        nest_adapters = False
+
+        if alias is None:
+            if insp.is_aliased_class and insp.selectable._is_subquery:
+                alias = insp.selectable.alias()
+            else:
+                alias = (
+                    mapper._with_polymorphic_selectable._anonymous_fromclause(
+                        name=name,
+                        flat=flat,
+                    )
+                )
+        elif insp.is_aliased_class:
+            nest_adapters = True
+
+        assert alias is not None
+        self._aliased_insp = AliasedInsp(
+            self,
+            insp,
+            alias,
+            name,
+            (
+                with_polymorphic_mappers
+                if with_polymorphic_mappers
+                else mapper.with_polymorphic_mappers
+            ),
+            (
+                with_polymorphic_discriminator
+                if with_polymorphic_discriminator is not None
+                else mapper.polymorphic_on
+            ),
+            base_alias,
+            use_mapper_path,
+            adapt_on_names,
+            represents_outer_join,
+            nest_adapters,
+        )
+
+        self.__name__ = f"aliased({mapper.class_.__name__})"
+
+    @classmethod
+    def _reconstitute_from_aliased_insp(
+        cls, aliased_insp: AliasedInsp[_O]
+    ) -> AliasedClass[_O]:
+        obj = cls.__new__(cls)
+        obj.__name__ = f"aliased({aliased_insp.mapper.class_.__name__})"
+        obj._aliased_insp = aliased_insp
+
+        if aliased_insp._is_with_polymorphic:
+            for sub_aliased_insp in aliased_insp._with_polymorphic_entities:
+                if sub_aliased_insp is not aliased_insp:
+                    ent = AliasedClass._reconstitute_from_aliased_insp(
+                        sub_aliased_insp
+                    )
+                    setattr(obj, sub_aliased_insp.class_.__name__, ent)
+
+        return obj
+
+    def __getattr__(self, key: str) -> Any:
+        try:
+            _aliased_insp = self.__dict__["_aliased_insp"]
+        except KeyError:
+            raise AttributeError()
+        else:
+            target = _aliased_insp._target
+            # maintain all getattr mechanics
+            attr = getattr(target, key)
+
+        # attribute is a method, that will be invoked against a
+        # "self"; so just return a new method with the same function and
+        # new self
+        if hasattr(attr, "__call__") and hasattr(attr, "__self__"):
+            return types.MethodType(attr.__func__, self)
+
+        # attribute is a descriptor, that will be invoked against a
+        # "self"; so invoke the descriptor against this self
+        if hasattr(attr, "__get__"):
+            attr = attr.__get__(None, self)
+
+        # attributes within the QueryableAttribute system will want this
+        # to be invoked so the object can be adapted
+        if hasattr(attr, "adapt_to_entity"):
+            attr = attr.adapt_to_entity(_aliased_insp)
+            setattr(self, key, attr)
+
+        return attr
+
+    def _get_from_serialized(
+        self, key: str, mapped_class: _O, aliased_insp: AliasedInsp[_O]
+    ) -> Any:
+        # this method is only used in terms of the
+        # sqlalchemy.ext.serializer extension
+        attr = getattr(mapped_class, key)
+        if hasattr(attr, "__call__") and hasattr(attr, "__self__"):
+            return types.MethodType(attr.__func__, self)
+
+        # attribute is a descriptor, that will be invoked against a
+        # "self"; so invoke the descriptor against this self
+        if hasattr(attr, "__get__"):
+            attr = attr.__get__(None, self)
+
+        # attributes within the QueryableAttribute system will want this
+        # to be invoked so the object can be adapted
+        if hasattr(attr, "adapt_to_entity"):
+            aliased_insp._weak_entity = weakref.ref(self)
+            attr = attr.adapt_to_entity(aliased_insp)
+            setattr(self, key, attr)
+
+        return attr
+
+    def __repr__(self) -> str:
+        return "<AliasedClass at 0x%x; %s>" % (
+            id(self),
+            self._aliased_insp._target.__name__,
+        )
+
+    def __str__(self) -> str:
+        return str(self._aliased_insp)
+
+
+@inspection._self_inspects
+class AliasedInsp(
+    ORMEntityColumnsClauseRole[_O],
+    ORMFromClauseRole,
+    HasCacheKey,
+    InspectionAttr,
+    MemoizedSlots,
+    inspection.Inspectable["AliasedInsp[_O]"],
+    Generic[_O],
+):
+    """Provide an inspection interface for an
+    :class:`.AliasedClass` object.
+
+    The :class:`.AliasedInsp` object is returned
+    given an :class:`.AliasedClass` using the
+    :func:`_sa.inspect` function::
+
+        from sqlalchemy import inspect
+        from sqlalchemy.orm import aliased
+
+        my_alias = aliased(MyMappedClass)
+        insp = inspect(my_alias)
+
+    Attributes on :class:`.AliasedInsp`
+    include:
+
+    * ``entity`` - the :class:`.AliasedClass` represented.
+    * ``mapper`` - the :class:`_orm.Mapper` mapping the underlying class.
+    * ``selectable`` - the :class:`_expression.Alias`
+      construct which ultimately
+      represents an aliased :class:`_schema.Table` or
+      :class:`_expression.Select`
+      construct.
+    * ``name`` - the name of the alias.  Also is used as the attribute
+      name when returned in a result tuple from :class:`_query.Query`.
+    * ``with_polymorphic_mappers`` - collection of :class:`_orm.Mapper`
+      objects
+      indicating all those mappers expressed in the select construct
+      for the :class:`.AliasedClass`.
+    * ``polymorphic_on`` - an alternate column or SQL expression which
+      will be used as the "discriminator" for a polymorphic load.
+
+    .. seealso::
+
+        :ref:`inspection_toplevel`
+
+    """
+
+    __slots__ = (
+        "__weakref__",
+        "_weak_entity",
+        "mapper",
+        "selectable",
+        "name",
+        "_adapt_on_names",
+        "with_polymorphic_mappers",
+        "polymorphic_on",
+        "_use_mapper_path",
+        "_base_alias",
+        "represents_outer_join",
+        "persist_selectable",
+        "local_table",
+        "_is_with_polymorphic",
+        "_with_polymorphic_entities",
+        "_adapter",
+        "_target",
+        "__clause_element__",
+        "_memoized_values",
+        "_all_column_expressions",
+        "_nest_adapters",
+    )
+
+    _cache_key_traversal = [
+        ("name", visitors.ExtendedInternalTraversal.dp_string),
+        ("_adapt_on_names", visitors.ExtendedInternalTraversal.dp_boolean),
+        ("_use_mapper_path", visitors.ExtendedInternalTraversal.dp_boolean),
+        ("_target", visitors.ExtendedInternalTraversal.dp_inspectable),
+        ("selectable", visitors.ExtendedInternalTraversal.dp_clauseelement),
+        (
+            "with_polymorphic_mappers",
+            visitors.InternalTraversal.dp_has_cache_key_list,
+        ),
+        ("polymorphic_on", visitors.InternalTraversal.dp_clauseelement),
+    ]
+
+    mapper: Mapper[_O]
+    selectable: FromClause
+    _adapter: ORMAdapter
+    with_polymorphic_mappers: Sequence[Mapper[Any]]
+    _with_polymorphic_entities: Sequence[AliasedInsp[Any]]
+
+    _weak_entity: weakref.ref[AliasedClass[_O]]
+    """the AliasedClass that refers to this AliasedInsp"""
+
+    _target: Union[Type[_O], AliasedClass[_O]]
+    """the thing referenced by the AliasedClass/AliasedInsp.
+
+    In the vast majority of cases, this is the mapped class.  However
+    it may also be another AliasedClass (alias of alias).
+
+    """
+
+    def __init__(
+        self,
+        entity: AliasedClass[_O],
+        inspected: _InternalEntityType[_O],
+        selectable: FromClause,
+        name: Optional[str],
+        with_polymorphic_mappers: Optional[Sequence[Mapper[Any]]],
+        polymorphic_on: Optional[ColumnElement[Any]],
+        _base_alias: Optional[AliasedInsp[Any]],
+        _use_mapper_path: bool,
+        adapt_on_names: bool,
+        represents_outer_join: bool,
+        nest_adapters: bool,
+    ):
+        mapped_class_or_ac = inspected.entity
+        mapper = inspected.mapper
+
+        self._weak_entity = weakref.ref(entity)
+        self.mapper = mapper
+        self.selectable = self.persist_selectable = self.local_table = (
+            selectable
+        )
+        self.name = name
+        self.polymorphic_on = polymorphic_on
+        self._base_alias = weakref.ref(_base_alias or self)
+        self._use_mapper_path = _use_mapper_path
+        self.represents_outer_join = represents_outer_join
+        self._nest_adapters = nest_adapters
+
+        if with_polymorphic_mappers:
+            self._is_with_polymorphic = True
+            self.with_polymorphic_mappers = with_polymorphic_mappers
+            self._with_polymorphic_entities = []
+            for poly in self.with_polymorphic_mappers:
+                if poly is not mapper:
+                    ent = AliasedClass(
+                        poly.class_,
+                        selectable,
+                        base_alias=self,
+                        adapt_on_names=adapt_on_names,
+                        use_mapper_path=_use_mapper_path,
+                    )
+
+                    setattr(self.entity, poly.class_.__name__, ent)
+                    self._with_polymorphic_entities.append(ent._aliased_insp)
+
+        else:
+            self._is_with_polymorphic = False
+            self.with_polymorphic_mappers = [mapper]
+
+        self._adapter = ORMAdapter(
+            _TraceAdaptRole.ALIASED_INSP,
+            mapper,
+            selectable=selectable,
+            equivalents=mapper._equivalent_columns,
+            adapt_on_names=adapt_on_names,
+            anonymize_labels=True,
+            # make sure the adapter doesn't try to grab other tables that
+            # are not even the thing we are mapping, such as embedded
+            # selectables in subqueries or CTEs.  See issue #6060
+            adapt_from_selectables={
+                m.selectable
+                for m in self.with_polymorphic_mappers
+                if not adapt_on_names
+            },
+            limit_on_entity=False,
+        )
+
+        if nest_adapters:
+            # supports "aliased class of aliased class" use case
+            assert isinstance(inspected, AliasedInsp)
+            self._adapter = inspected._adapter.wrap(self._adapter)
+
+        self._adapt_on_names = adapt_on_names
+        self._target = mapped_class_or_ac
+
+    @classmethod
+    def _alias_factory(
+        cls,
+        element: Union[_EntityType[_O], FromClause],
+        alias: Optional[FromClause] = None,
+        name: Optional[str] = None,
+        flat: bool = False,
+        adapt_on_names: bool = False,
+    ) -> Union[AliasedClass[_O], FromClause]:
+        if isinstance(element, FromClause):
+            if adapt_on_names:
+                raise sa_exc.ArgumentError(
+                    "adapt_on_names only applies to ORM elements"
+                )
+            if name:
+                return element.alias(name=name, flat=flat)
+            else:
+                return coercions.expect(
+                    roles.AnonymizedFromClauseRole, element, flat=flat
+                )
+        else:
+            return AliasedClass(
+                element,
+                alias=alias,
+                flat=flat,
+                name=name,
+                adapt_on_names=adapt_on_names,
+            )
+
+    @classmethod
+    def _with_polymorphic_factory(
+        cls,
+        base: Union[Type[_O], Mapper[_O]],
+        classes: Union[Literal["*"], Iterable[_EntityType[Any]]],
+        selectable: Union[Literal[False, None], FromClause] = False,
+        flat: bool = False,
+        polymorphic_on: Optional[ColumnElement[Any]] = None,
+        aliased: bool = False,
+        innerjoin: bool = False,
+        adapt_on_names: bool = False,
+        name: Optional[str] = None,
+        _use_mapper_path: bool = False,
+    ) -> AliasedClass[_O]:
+        primary_mapper = _class_to_mapper(base)
+
+        if selectable not in (None, False) and flat:
+            raise sa_exc.ArgumentError(
+                "the 'flat' and 'selectable' arguments cannot be passed "
+                "simultaneously to with_polymorphic()"
+            )
+
+        mappers, selectable = primary_mapper._with_polymorphic_args(
+            classes, selectable, innerjoin=innerjoin
+        )
+        if aliased or flat:
+            assert selectable is not None
+            selectable = selectable._anonymous_fromclause(flat=flat)
+
+        return AliasedClass(
+            base,
+            selectable,
+            name=name,
+            with_polymorphic_mappers=mappers,
+            adapt_on_names=adapt_on_names,
+            with_polymorphic_discriminator=polymorphic_on,
+            use_mapper_path=_use_mapper_path,
+            represents_outer_join=not innerjoin,
+        )
+
+    @property
+    def entity(self) -> AliasedClass[_O]:
+        # to eliminate reference cycles, the AliasedClass is held weakly.
+        # this produces some situations where the AliasedClass gets lost,
+        # particularly when one is created internally and only the AliasedInsp
+        # is passed around.
+        # to work around this case, we just generate a new one when we need
+        # it, as it is a simple class with very little initial state on it.
+        ent = self._weak_entity()
+        if ent is None:
+            ent = AliasedClass._reconstitute_from_aliased_insp(self)
+            self._weak_entity = weakref.ref(ent)
+        return ent
+
+    is_aliased_class = True
+    "always returns True"
+
+    def _memoized_method___clause_element__(self) -> FromClause:
+        return self.selectable._annotate(
+            {
+                "parentmapper": self.mapper,
+                "parententity": self,
+                "entity_namespace": self,
+            }
+        )._set_propagate_attrs(
+            {"compile_state_plugin": "orm", "plugin_subject": self}
+        )
+
+    @property
+    def entity_namespace(self) -> AliasedClass[_O]:
+        return self.entity
+
+    @property
+    def class_(self) -> Type[_O]:
+        """Return the mapped class ultimately represented by this
+        :class:`.AliasedInsp`."""
+        return self.mapper.class_
+
+    @property
+    def _path_registry(self) -> AbstractEntityRegistry:
+        if self._use_mapper_path:
+            return self.mapper._path_registry
+        else:
+            return PathRegistry.per_mapper(self)
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {
+            "entity": self.entity,
+            "mapper": self.mapper,
+            "alias": self.selectable,
+            "name": self.name,
+            "adapt_on_names": self._adapt_on_names,
+            "with_polymorphic_mappers": self.with_polymorphic_mappers,
+            "with_polymorphic_discriminator": self.polymorphic_on,
+            "base_alias": self._base_alias(),
+            "use_mapper_path": self._use_mapper_path,
+            "represents_outer_join": self.represents_outer_join,
+            "nest_adapters": self._nest_adapters,
+        }
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        self.__init__(  # type: ignore
+            state["entity"],
+            state["mapper"],
+            state["alias"],
+            state["name"],
+            state["with_polymorphic_mappers"],
+            state["with_polymorphic_discriminator"],
+            state["base_alias"],
+            state["use_mapper_path"],
+            state["adapt_on_names"],
+            state["represents_outer_join"],
+            state["nest_adapters"],
+        )
+
+    def _merge_with(self, other: AliasedInsp[_O]) -> AliasedInsp[_O]:
+        # assert self._is_with_polymorphic
+        # assert other._is_with_polymorphic
+
+        primary_mapper = other.mapper
+
+        assert self.mapper is primary_mapper
+
+        our_classes = util.to_set(
+            mp.class_ for mp in self.with_polymorphic_mappers
+        )
+        new_classes = {mp.class_ for mp in other.with_polymorphic_mappers}
+        if our_classes == new_classes:
+            return other
+        else:
+            classes = our_classes.union(new_classes)
+
+        mappers, selectable = primary_mapper._with_polymorphic_args(
+            classes, None, innerjoin=not other.represents_outer_join
+        )
+        selectable = selectable._anonymous_fromclause(flat=True)
+        return AliasedClass(
+            primary_mapper,
+            selectable,
+            with_polymorphic_mappers=mappers,
+            with_polymorphic_discriminator=other.polymorphic_on,
+            use_mapper_path=other._use_mapper_path,
+            represents_outer_join=other.represents_outer_join,
+        )._aliased_insp
+
+    def _adapt_element(
+        self, expr: _ORMCOLEXPR, key: Optional[str] = None
+    ) -> _ORMCOLEXPR:
+        assert isinstance(expr, ColumnElement)
+        d: Dict[str, Any] = {
+            "parententity": self,
+            "parentmapper": self.mapper,
+        }
+        if key:
+            d["proxy_key"] = key
+
+        # IMO mypy should see this one also as returning the same type
+        # we put into it, but it's not
+        return (
+            self._adapter.traverse(expr)
+            ._annotate(d)
+            ._set_propagate_attrs(
+                {"compile_state_plugin": "orm", "plugin_subject": self}
+            )
+        )
+
+    if TYPE_CHECKING:
+        # establish compatibility with the _ORMAdapterProto protocol,
+        # which in turn is compatible with _CoreAdapterProto.
+
+        def _orm_adapt_element(
+            self,
+            obj: _CE,
+            key: Optional[str] = None,
+        ) -> _CE: ...
+
+    else:
+        _orm_adapt_element = _adapt_element
+
+    def _entity_for_mapper(self, mapper):
+        self_poly = self.with_polymorphic_mappers
+        if mapper in self_poly:
+            if mapper is self.mapper:
+                return self
+            else:
+                return getattr(
+                    self.entity, mapper.class_.__name__
+                )._aliased_insp
+        elif mapper.isa(self.mapper):
+            return self
+        else:
+            assert False, "mapper %s doesn't correspond to %s" % (mapper, self)
+
+    def _memoized_attr__get_clause(self):
+        onclause, replacemap = self.mapper._get_clause
+        return (
+            self._adapter.traverse(onclause),
+            {
+                self._adapter.traverse(col): param
+                for col, param in replacemap.items()
+            },
+        )
+
+    def _memoized_attr__memoized_values(self):
+        return {}
+
+    def _memoized_attr__all_column_expressions(self):
+        if self._is_with_polymorphic:
+            cols_plus_keys = self.mapper._columns_plus_keys(
+                [ent.mapper for ent in self._with_polymorphic_entities]
+            )
+        else:
+            cols_plus_keys = self.mapper._columns_plus_keys()
+
+        cols_plus_keys = [
+            (key, self._adapt_element(col)) for key, col in cols_plus_keys
+        ]
+
+        return ColumnCollection(cols_plus_keys)
+
+    def _memo(self, key, callable_, *args, **kw):
+        if key in self._memoized_values:
+            return self._memoized_values[key]
+        else:
+            self._memoized_values[key] = value = callable_(*args, **kw)
+            return value
+
+    def __repr__(self):
+        if self.with_polymorphic_mappers:
+            with_poly = "(%s)" % ", ".join(
+                mp.class_.__name__ for mp in self.with_polymorphic_mappers
+            )
+        else:
+            with_poly = ""
+        return "<AliasedInsp at 0x%x; %s%s>" % (
+            id(self),
+            self.class_.__name__,
+            with_poly,
+        )
+
+    def __str__(self):
+        if self._is_with_polymorphic:
+            return "with_polymorphic(%s, [%s])" % (
+                self._target.__name__,
+                ", ".join(
+                    mp.class_.__name__
+                    for mp in self.with_polymorphic_mappers
+                    if mp is not self.mapper
+                ),
+            )
+        else:
+            return "aliased(%s)" % (self._target.__name__,)
+
+
+class _WrapUserEntity:
+    """A wrapper used within the loader_criteria lambda caller so that
+    we can bypass declared_attr descriptors on unmapped mixins, which
+    normally emit a warning for such use.
+
+    might also be useful for other per-lambda instrumentations should
+    the need arise.
+
+    """
+
+    __slots__ = ("subject",)
+
+    def __init__(self, subject):
+        self.subject = subject
+
+    @util.preload_module("sqlalchemy.orm.decl_api")
+    def __getattribute__(self, name):
+        decl_api = util.preloaded.orm.decl_api
+
+        subject = object.__getattribute__(self, "subject")
+        if name in subject.__dict__ and isinstance(
+            subject.__dict__[name], decl_api.declared_attr
+        ):
+            return subject.__dict__[name].fget(subject)
+        else:
+            return getattr(subject, name)
+
+
+class LoaderCriteriaOption(CriteriaOption):
+    """Add additional WHERE criteria to the load for all occurrences of
+    a particular entity.
+
+    :class:`_orm.LoaderCriteriaOption` is invoked using the
+    :func:`_orm.with_loader_criteria` function; see that function for
+    details.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = (
+        "root_entity",
+        "entity",
+        "deferred_where_criteria",
+        "where_criteria",
+        "_where_crit_orig",
+        "include_aliases",
+        "propagate_to_loaders",
+    )
+
+    _traverse_internals = [
+        ("root_entity", visitors.ExtendedInternalTraversal.dp_plain_obj),
+        ("entity", visitors.ExtendedInternalTraversal.dp_has_cache_key),
+        ("where_criteria", visitors.InternalTraversal.dp_clauseelement),
+        ("include_aliases", visitors.InternalTraversal.dp_boolean),
+        ("propagate_to_loaders", visitors.InternalTraversal.dp_boolean),
+    ]
+
+    root_entity: Optional[Type[Any]]
+    entity: Optional[_InternalEntityType[Any]]
+    where_criteria: Union[ColumnElement[bool], lambdas.DeferredLambdaElement]
+    deferred_where_criteria: bool
+    include_aliases: bool
+    propagate_to_loaders: bool
+
+    _where_crit_orig: Any
+
+    def __init__(
+        self,
+        entity_or_base: _EntityType[Any],
+        where_criteria: Union[
+            _ColumnExpressionArgument[bool],
+            Callable[[Any], _ColumnExpressionArgument[bool]],
+        ],
+        loader_only: bool = False,
+        include_aliases: bool = False,
+        propagate_to_loaders: bool = True,
+        track_closure_variables: bool = True,
+    ):
+        entity = cast(
+            "_InternalEntityType[Any]",
+            inspection.inspect(entity_or_base, False),
+        )
+        if entity is None:
+            self.root_entity = cast("Type[Any]", entity_or_base)
+            self.entity = None
+        else:
+            self.root_entity = None
+            self.entity = entity
+
+        self._where_crit_orig = where_criteria
+        if callable(where_criteria):
+            if self.root_entity is not None:
+                wrap_entity = self.root_entity
+            else:
+                assert entity is not None
+                wrap_entity = entity.entity
+
+            self.deferred_where_criteria = True
+            self.where_criteria = lambdas.DeferredLambdaElement(
+                where_criteria,
+                roles.WhereHavingRole,
+                lambda_args=(_WrapUserEntity(wrap_entity),),
+                opts=lambdas.LambdaOptions(
+                    track_closure_variables=track_closure_variables
+                ),
+            )
+        else:
+            self.deferred_where_criteria = False
+            self.where_criteria = coercions.expect(
+                roles.WhereHavingRole, where_criteria
+            )
+
+        self.include_aliases = include_aliases
+        self.propagate_to_loaders = propagate_to_loaders
+
+    @classmethod
+    def _unreduce(
+        cls, entity, where_criteria, include_aliases, propagate_to_loaders
+    ):
+        return LoaderCriteriaOption(
+            entity,
+            where_criteria,
+            include_aliases=include_aliases,
+            propagate_to_loaders=propagate_to_loaders,
+        )
+
+    def __reduce__(self):
+        return (
+            LoaderCriteriaOption._unreduce,
+            (
+                self.entity.class_ if self.entity else self.root_entity,
+                self._where_crit_orig,
+                self.include_aliases,
+                self.propagate_to_loaders,
+            ),
+        )
+
+    def _all_mappers(self) -> Iterator[Mapper[Any]]:
+        if self.entity:
+            yield from self.entity.mapper.self_and_descendants
+        else:
+            assert self.root_entity
+            stack = list(self.root_entity.__subclasses__())
+            while stack:
+                subclass = stack.pop(0)
+                ent = cast(
+                    "_InternalEntityType[Any]",
+                    inspection.inspect(subclass, raiseerr=False),
+                )
+                if ent:
+                    yield from ent.mapper.self_and_descendants
+                else:
+                    stack.extend(subclass.__subclasses__())
+
+    def _should_include(self, compile_state: ORMCompileState) -> bool:
+        if (
+            compile_state.select_statement._annotations.get(
+                "for_loader_criteria", None
+            )
+            is self
+        ):
+            return False
+        return True
+
+    def _resolve_where_criteria(
+        self, ext_info: _InternalEntityType[Any]
+    ) -> ColumnElement[bool]:
+        if self.deferred_where_criteria:
+            crit = cast(
+                "ColumnElement[bool]",
+                self.where_criteria._resolve_with_args(ext_info.entity),
+            )
+        else:
+            crit = self.where_criteria  # type: ignore
+        assert isinstance(crit, ColumnElement)
+        return sql_util._deep_annotate(
+            crit,
+            {"for_loader_criteria": self},
+            detect_subquery_cols=True,
+            ind_cols_on_fromclause=True,
+        )
+
+    def process_compile_state_replaced_entities(
+        self,
+        compile_state: ORMCompileState,
+        mapper_entities: Iterable[_MapperEntity],
+    ) -> None:
+        self.process_compile_state(compile_state)
+
+    def process_compile_state(self, compile_state: ORMCompileState) -> None:
+        """Apply a modification to a given :class:`.CompileState`."""
+
+        # if options to limit the criteria to immediate query only,
+        # use compile_state.attributes instead
+
+        self.get_global_criteria(compile_state.global_attributes)
+
+    def get_global_criteria(self, attributes: Dict[Any, Any]) -> None:
+        for mp in self._all_mappers():
+            load_criteria = attributes.setdefault(
+                ("additional_entity_criteria", mp), []
+            )
+
+            load_criteria.append(self)
+
+
+inspection._inspects(AliasedClass)(lambda target: target._aliased_insp)
+
+
+@inspection._inspects(type)
+def _inspect_mc(
+    class_: Type[_O],
+) -> Optional[Mapper[_O]]:
+    try:
+        class_manager = opt_manager_of_class(class_)
+        if class_manager is None or not class_manager.is_mapped:
+            return None
+        mapper = class_manager.mapper
+    except exc.NO_STATE:
+        return None
+    else:
+        return mapper
+
+
+GenericAlias = type(List[Any])
+
+
+@inspection._inspects(GenericAlias)
+def _inspect_generic_alias(
+    class_: Type[_O],
+) -> Optional[Mapper[_O]]:
+    origin = cast("Type[_O]", get_origin(class_))
+    return _inspect_mc(origin)
+
+
+@inspection._self_inspects
+class Bundle(
+    ORMColumnsClauseRole[_T],
+    SupportsCloneAnnotations,
+    MemoizedHasCacheKey,
+    inspection.Inspectable["Bundle[_T]"],
+    InspectionAttr,
+):
+    """A grouping of SQL expressions that are returned by a :class:`.Query`
+    under one namespace.
+
+    The :class:`.Bundle` essentially allows nesting of the tuple-based
+    results returned by a column-oriented :class:`_query.Query` object.
+    It also
+    is extensible via simple subclassing, where the primary capability
+    to override is that of how the set of expressions should be returned,
+    allowing post-processing as well as custom return types, without
+    involving ORM identity-mapped classes.
+
+    .. seealso::
+
+        :ref:`bundles`
+
+
+    """
+
+    single_entity = False
+    """If True, queries for a single Bundle will be returned as a single
+    entity, rather than an element within a keyed tuple."""
+
+    is_clause_element = False
+
+    is_mapper = False
+
+    is_aliased_class = False
+
+    is_bundle = True
+
+    _propagate_attrs: _PropagateAttrsType = util.immutabledict()
+
+    proxy_set = util.EMPTY_SET  # type: ignore
+
+    exprs: List[_ColumnsClauseElement]
+
+    def __init__(
+        self, name: str, *exprs: _ColumnExpressionArgument[Any], **kw: Any
+    ):
+        r"""Construct a new :class:`.Bundle`.
+
+        e.g.::
+
+            bn = Bundle("mybundle", MyClass.x, MyClass.y)
+
+            for row in session.query(bn).filter(bn.c.x == 5).filter(bn.c.y == 4):
+                print(row.mybundle.x, row.mybundle.y)
+
+        :param name: name of the bundle.
+        :param \*exprs: columns or SQL expressions comprising the bundle.
+        :param single_entity=False: if True, rows for this :class:`.Bundle`
+         can be returned as a "single entity" outside of any enclosing tuple
+         in the same manner as a mapped entity.
+
+        """  # noqa: E501
+        self.name = self._label = name
+        coerced_exprs = [
+            coercions.expect(
+                roles.ColumnsClauseRole, expr, apply_propagate_attrs=self
+            )
+            for expr in exprs
+        ]
+        self.exprs = coerced_exprs
+
+        self.c = self.columns = ColumnCollection(
+            (getattr(col, "key", col._label), col)
+            for col in [e._annotations.get("bundle", e) for e in coerced_exprs]
+        ).as_readonly()
+        self.single_entity = kw.pop("single_entity", self.single_entity)
+
+    def _gen_cache_key(
+        self, anon_map: anon_map, bindparams: List[BindParameter[Any]]
+    ) -> Tuple[Any, ...]:
+        return (self.__class__, self.name, self.single_entity) + tuple(
+            [expr._gen_cache_key(anon_map, bindparams) for expr in self.exprs]
+        )
+
+    @property
+    def mapper(self) -> Optional[Mapper[Any]]:
+        mp: Optional[Mapper[Any]] = self.exprs[0]._annotations.get(
+            "parentmapper", None
+        )
+        return mp
+
+    @property
+    def entity(self) -> Optional[_InternalEntityType[Any]]:
+        ie: Optional[_InternalEntityType[Any]] = self.exprs[
+            0
+        ]._annotations.get("parententity", None)
+        return ie
+
+    @property
+    def entity_namespace(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        return self.c
+
+    columns: ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]
+
+    """A namespace of SQL expressions referred to by this :class:`.Bundle`.
+
+        e.g.::
+
+            bn = Bundle("mybundle", MyClass.x, MyClass.y)
+
+            q = sess.query(bn).filter(bn.c.x == 5)
+
+        Nesting of bundles is also supported::
+
+            b1 = Bundle(
+                "b1",
+                Bundle("b2", MyClass.a, MyClass.b),
+                Bundle("b3", MyClass.x, MyClass.y),
+            )
+
+            q = sess.query(b1).filter(b1.c.b2.c.a == 5).filter(b1.c.b3.c.y == 9)
+
+    .. seealso::
+
+        :attr:`.Bundle.c`
+
+    """  # noqa: E501
+
+    c: ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]
+    """An alias for :attr:`.Bundle.columns`."""
+
+    def _clone(self, **kw):
+        cloned = self.__class__.__new__(self.__class__)
+        cloned.__dict__.update(self.__dict__)
+        return cloned
+
+    def __clause_element__(self):
+        # ensure existing entity_namespace remains
+        annotations = {"bundle": self, "entity_namespace": self}
+        annotations.update(self._annotations)
+
+        plugin_subject = self.exprs[0]._propagate_attrs.get(
+            "plugin_subject", self.entity
+        )
+        return (
+            expression.ClauseList(
+                _literal_as_text_role=roles.ColumnsClauseRole,
+                group=False,
+                *[e._annotations.get("bundle", e) for e in self.exprs],
+            )
+            ._annotate(annotations)
+            ._set_propagate_attrs(
+                # the Bundle *must* use the orm plugin no matter what.  the
+                # subject can be None but it's much better if it's not.
+                {
+                    "compile_state_plugin": "orm",
+                    "plugin_subject": plugin_subject,
+                }
+            )
+        )
+
+    @property
+    def clauses(self):
+        return self.__clause_element__().clauses
+
+    def label(self, name):
+        """Provide a copy of this :class:`.Bundle` passing a new label."""
+
+        cloned = self._clone()
+        cloned.name = name
+        return cloned
+
+    def create_row_processor(
+        self,
+        query: Select[Any],
+        procs: Sequence[Callable[[Row[Any]], Any]],
+        labels: Sequence[str],
+    ) -> Callable[[Row[Any]], Any]:
+        """Produce the "row processing" function for this :class:`.Bundle`.
+
+        May be overridden by subclasses to provide custom behaviors when
+        results are fetched. The method is passed the statement object and a
+        set of "row processor" functions at query execution time; these
+        processor functions when given a result row will return the individual
+        attribute value, which can then be adapted into any kind of return data
+        structure.
+
+        The example below illustrates replacing the usual :class:`.Row`
+        return structure with a straight Python dictionary::
+
+            from sqlalchemy.orm import Bundle
+
+
+            class DictBundle(Bundle):
+                def create_row_processor(self, query, procs, labels):
+                    "Override create_row_processor to return values as dictionaries"
+
+                    def proc(row):
+                        return dict(zip(labels, (proc(row) for proc in procs)))
+
+                    return proc
+
+        A result from the above :class:`_orm.Bundle` will return dictionary
+        values::
+
+            bn = DictBundle("mybundle", MyClass.data1, MyClass.data2)
+            for row in session.execute(select(bn)).where(bn.c.data1 == "d1"):
+                print(row.mybundle["data1"], row.mybundle["data2"])
+
+        """  # noqa: E501
+        keyed_tuple = result_tuple(labels, [() for l in labels])
+
+        def proc(row: Row[Any]) -> Any:
+            return keyed_tuple([proc(row) for proc in procs])
+
+        return proc
+
+
+def _orm_annotate(element: _SA, exclude: Optional[Any] = None) -> _SA:
+    """Deep copy the given ClauseElement, annotating each element with the
+    "_orm_adapt" flag.
+
+    Elements within the exclude collection will be cloned but not annotated.
+
+    """
+    return sql_util._deep_annotate(element, {"_orm_adapt": True}, exclude)
+
+
+def _orm_deannotate(element: _SA) -> _SA:
+    """Remove annotations that link a column to a particular mapping.
+
+    Note this doesn't affect "remote" and "foreign" annotations
+    passed by the :func:`_orm.foreign` and :func:`_orm.remote`
+    annotators.
+
+    """
+
+    return sql_util._deep_deannotate(
+        element, values=("_orm_adapt", "parententity")
+    )
+
+
+def _orm_full_deannotate(element: _SA) -> _SA:
+    return sql_util._deep_deannotate(element)
+
+
+class _ORMJoin(expression.Join):
+    """Extend Join to support ORM constructs as input."""
+
+    __visit_name__ = expression.Join.__visit_name__
+
+    inherit_cache = True
+
+    def __init__(
+        self,
+        left: _FromClauseArgument,
+        right: _FromClauseArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        isouter: bool = False,
+        full: bool = False,
+        _left_memo: Optional[Any] = None,
+        _right_memo: Optional[Any] = None,
+        _extra_criteria: Tuple[ColumnElement[bool], ...] = (),
+    ):
+        left_info = cast(
+            "Union[FromClause, _InternalEntityType[Any]]",
+            inspection.inspect(left),
+        )
+
+        right_info = cast(
+            "Union[FromClause, _InternalEntityType[Any]]",
+            inspection.inspect(right),
+        )
+        adapt_to = right_info.selectable
+
+        # used by joined eager loader
+        self._left_memo = _left_memo
+        self._right_memo = _right_memo
+
+        if isinstance(onclause, attributes.QueryableAttribute):
+            if TYPE_CHECKING:
+                assert isinstance(
+                    onclause.comparator, RelationshipProperty.Comparator
+                )
+            on_selectable = onclause.comparator._source_selectable()
+            prop = onclause.property
+            _extra_criteria += onclause._extra_criteria
+        elif isinstance(onclause, MapperProperty):
+            # used internally by joined eager loader...possibly not ideal
+            prop = onclause
+            on_selectable = prop.parent.selectable
+        else:
+            prop = None
+            on_selectable = None
+
+        left_selectable = left_info.selectable
+        if prop:
+            adapt_from: Optional[FromClause]
+            if sql_util.clause_is_present(on_selectable, left_selectable):
+                adapt_from = on_selectable
+            else:
+                assert isinstance(left_selectable, FromClause)
+                adapt_from = left_selectable
+
+            (
+                pj,
+                sj,
+                source,
+                dest,
+                secondary,
+                target_adapter,
+            ) = prop._create_joins(
+                source_selectable=adapt_from,
+                dest_selectable=adapt_to,
+                source_polymorphic=True,
+                of_type_entity=right_info,
+                alias_secondary=True,
+                extra_criteria=_extra_criteria,
+            )
+
+            if sj is not None:
+                if isouter:
+                    # note this is an inner join from secondary->right
+                    right = sql.join(secondary, right, sj)
+                    onclause = pj
+                else:
+                    left = sql.join(left, secondary, pj, isouter)
+                    onclause = sj
+            else:
+                onclause = pj
+
+            self._target_adapter = target_adapter
+
+        # we don't use the normal coercions logic for _ORMJoin
+        # (probably should), so do some gymnastics to get the entity.
+        # logic here is for #8721, which was a major bug in 1.4
+        # for almost two years, not reported/fixed until 1.4.43 (!)
+        if is_selectable(left_info):
+            parententity = left_selectable._annotations.get(
+                "parententity", None
+            )
+        elif insp_is_mapper(left_info) or insp_is_aliased_class(left_info):
+            parententity = left_info
+        else:
+            parententity = None
+
+        if parententity is not None:
+            self._annotations = self._annotations.union(
+                {"parententity": parententity}
+            )
+
+        augment_onclause = bool(_extra_criteria) and not prop
+        expression.Join.__init__(self, left, right, onclause, isouter, full)
+
+        assert self.onclause is not None
+
+        if augment_onclause:
+            self.onclause &= sql.and_(*_extra_criteria)
+
+        if (
+            not prop
+            and getattr(right_info, "mapper", None)
+            and right_info.mapper.single  # type: ignore
+        ):
+            right_info = cast("_InternalEntityType[Any]", right_info)
+            # if single inheritance target and we are using a manual
+            # or implicit ON clause, augment it the same way we'd augment the
+            # WHERE.
+            single_crit = right_info.mapper._single_table_criterion
+            if single_crit is not None:
+                if insp_is_aliased_class(right_info):
+                    single_crit = right_info._adapter.traverse(single_crit)
+                self.onclause = self.onclause & single_crit
+
+    def _splice_into_center(self, other):
+        """Splice a join into the center.
+
+        Given join(a, b) and join(b, c), return join(a, b).join(c)
+
+        """
+        leftmost = other
+        while isinstance(leftmost, sql.Join):
+            leftmost = leftmost.left
+
+        assert self.right is leftmost
+
+        left = _ORMJoin(
+            self.left,
+            other.left,
+            self.onclause,
+            isouter=self.isouter,
+            _left_memo=self._left_memo,
+            _right_memo=other._left_memo._path_registry,
+        )
+
+        return _ORMJoin(
+            left,
+            other.right,
+            other.onclause,
+            isouter=other.isouter,
+            _right_memo=other._right_memo,
+        )
+
+    def join(
+        self,
+        right: _FromClauseArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        isouter: bool = False,
+        full: bool = False,
+    ) -> _ORMJoin:
+        return _ORMJoin(self, right, onclause, full=full, isouter=isouter)
+
+    def outerjoin(
+        self,
+        right: _FromClauseArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        full: bool = False,
+    ) -> _ORMJoin:
+        return _ORMJoin(self, right, onclause, isouter=True, full=full)
+
+
+def with_parent(
+    instance: object,
+    prop: attributes.QueryableAttribute[Any],
+    from_entity: Optional[_EntityType[Any]] = None,
+) -> ColumnElement[bool]:
+    """Create filtering criterion that relates this query's primary entity
+    to the given related instance, using established
+    :func:`_orm.relationship()`
+    configuration.
+
+    E.g.::
+
+        stmt = select(Address).where(with_parent(some_user, User.addresses))
+
+    The SQL rendered is the same as that rendered when a lazy loader
+    would fire off from the given parent on that attribute, meaning
+    that the appropriate state is taken from the parent object in
+    Python without the need to render joins to the parent table
+    in the rendered statement.
+
+    The given property may also make use of :meth:`_orm.PropComparator.of_type`
+    to indicate the left side of the criteria::
+
+
+        a1 = aliased(Address)
+        a2 = aliased(Address)
+        stmt = select(a1, a2).where(with_parent(u1, User.addresses.of_type(a2)))
+
+    The above use is equivalent to using the
+    :func:`_orm.with_parent.from_entity` argument::
+
+        a1 = aliased(Address)
+        a2 = aliased(Address)
+        stmt = select(a1, a2).where(
+            with_parent(u1, User.addresses, from_entity=a2)
+        )
+
+    :param instance:
+      An instance which has some :func:`_orm.relationship`.
+
+    :param property:
+      Class-bound attribute, which indicates
+      what relationship from the instance should be used to reconcile the
+      parent/child relationship.
+
+    :param from_entity:
+      Entity in which to consider as the left side.  This defaults to the
+      "zero" entity of the :class:`_query.Query` itself.
+
+      .. versionadded:: 1.2
+
+    """  # noqa: E501
+    prop_t: RelationshipProperty[Any]
+
+    if isinstance(prop, str):
+        raise sa_exc.ArgumentError(
+            "with_parent() accepts class-bound mapped attributes, not strings"
+        )
+    elif isinstance(prop, attributes.QueryableAttribute):
+        if prop._of_type:
+            from_entity = prop._of_type
+        mapper_property = prop.property
+        if mapper_property is None or not prop_is_relationship(
+            mapper_property
+        ):
+            raise sa_exc.ArgumentError(
+                f"Expected relationship property for with_parent(), "
+                f"got {mapper_property}"
+            )
+        prop_t = mapper_property
+    else:
+        prop_t = prop
+
+    return prop_t._with_parent(instance, from_entity=from_entity)
+
+
+def has_identity(object_: object) -> bool:
+    """Return True if the given object has a database
+    identity.
+
+    This typically corresponds to the object being
+    in either the persistent or detached state.
+
+    .. seealso::
+
+        :func:`.was_deleted`
+
+    """
+    state = attributes.instance_state(object_)
+    return state.has_identity
+
+
+def was_deleted(object_: object) -> bool:
+    """Return True if the given object was deleted
+    within a session flush.
+
+    This is regardless of whether or not the object is
+    persistent or detached.
+
+    .. seealso::
+
+        :attr:`.InstanceState.was_deleted`
+
+    """
+
+    state = attributes.instance_state(object_)
+    return state.was_deleted
+
+
+def _entity_corresponds_to(
+    given: _InternalEntityType[Any], entity: _InternalEntityType[Any]
+) -> bool:
+    """determine if 'given' corresponds to 'entity', in terms
+    of an entity passed to Query that would match the same entity
+    being referred to elsewhere in the query.
+
+    """
+    if insp_is_aliased_class(entity):
+        if insp_is_aliased_class(given):
+            if entity._base_alias() is given._base_alias():
+                return True
+        return False
+    elif insp_is_aliased_class(given):
+        if given._use_mapper_path:
+            return entity in given.with_polymorphic_mappers
+        else:
+            return entity is given
+
+    assert insp_is_mapper(given)
+    return entity.common_parent(given)
+
+
+def _entity_corresponds_to_use_path_impl(
+    given: _InternalEntityType[Any], entity: _InternalEntityType[Any]
+) -> bool:
+    """determine if 'given' corresponds to 'entity', in terms
+    of a path of loader options where a mapped attribute is taken to
+    be a member of a parent entity.
+
+    e.g.::
+
+        someoption(A).someoption(A.b)  # -> fn(A, A) -> True
+        someoption(A).someoption(C.d)  # -> fn(A, C) -> False
+
+        a1 = aliased(A)
+        someoption(a1).someoption(A.b)  # -> fn(a1, A) -> False
+        someoption(a1).someoption(a1.b)  # -> fn(a1, a1) -> True
+
+        wp = with_polymorphic(A, [A1, A2])
+        someoption(wp).someoption(A1.foo)  # -> fn(wp, A1) -> False
+        someoption(wp).someoption(wp.A1.foo)  # -> fn(wp, wp.A1) -> True
+
+    """
+    if insp_is_aliased_class(given):
+        return (
+            insp_is_aliased_class(entity)
+            and not entity._use_mapper_path
+            and (given is entity or entity in given._with_polymorphic_entities)
+        )
+    elif not insp_is_aliased_class(entity):
+        return given.isa(entity.mapper)
+    else:
+        return (
+            entity._use_mapper_path
+            and given in entity.with_polymorphic_mappers
+        )
+
+
+def _entity_isa(given: _InternalEntityType[Any], mapper: Mapper[Any]) -> bool:
+    """determine if 'given' "is a" mapper, in terms of the given
+    would load rows of type 'mapper'.
+
+    """
+    if given.is_aliased_class:
+        return mapper in given.with_polymorphic_mappers or given.mapper.isa(
+            mapper
+        )
+    elif given.with_polymorphic_mappers:
+        return mapper in given.with_polymorphic_mappers or given.isa(mapper)
+    else:
+        return given.isa(mapper)
+
+
+def _getitem(iterable_query: Query[Any], item: Any) -> Any:
+    """calculate __getitem__ in terms of an iterable query object
+    that also has a slice() method.
+
+    """
+
+    def _no_negative_indexes():
+        raise IndexError(
+            "negative indexes are not accepted by SQL "
+            "index / slice operators"
+        )
+
+    if isinstance(item, slice):
+        start, stop, step = util.decode_slice(item)
+
+        if (
+            isinstance(stop, int)
+            and isinstance(start, int)
+            and stop - start <= 0
+        ):
+            return []
+
+        elif (isinstance(start, int) and start < 0) or (
+            isinstance(stop, int) and stop < 0
+        ):
+            _no_negative_indexes()
+
+        res = iterable_query.slice(start, stop)
+        if step is not None:
+            return list(res)[None : None : item.step]
+        else:
+            return list(res)
+    else:
+        if item == -1:
+            _no_negative_indexes()
+        else:
+            return list(iterable_query[item : item + 1])[0]
+
+
+def _is_mapped_annotation(
+    raw_annotation: _AnnotationScanType,
+    cls: Type[Any],
+    originating_cls: Type[Any],
+) -> bool:
+    try:
+        annotated = de_stringify_annotation(
+            cls, raw_annotation, originating_cls.__module__
+        )
+    except NameError:
+        # in most cases, at least within our own tests, we can raise
+        # here, which is more accurate as it prevents us from returning
+        # false negatives.  However, in the real world, try to avoid getting
+        # involved with end-user annotations that have nothing to do with us.
+        # see issue #8888 where we bypass using this function in the case
+        # that we want to detect an unresolvable Mapped[] type.
+        return False
+    else:
+        return is_origin_of_cls(annotated, _MappedAnnotationBase)
+
+
+class _CleanupError(Exception):
+    pass
+
+
+def _cleanup_mapped_str_annotation(
+    annotation: str, originating_module: str
+) -> str:
+    # fix up an annotation that comes in as the form:
+    # 'Mapped[List[Address]]'  so that it instead looks like:
+    # 'Mapped[List["Address"]]' , which will allow us to get
+    # "Address" as a string
+
+    # additionally, resolve symbols for these names since this is where
+    # we'd have to do it
+
+    inner: Optional[Match[str]]
+
+    mm = re.match(r"^([^ \|]+?)\[(.+)\]$", annotation)
+
+    if not mm:
+        return annotation
+
+    # ticket #8759.  Resolve the Mapped name to a real symbol.
+    # originally this just checked the name.
+    try:
+        obj = eval_name_only(mm.group(1), originating_module)
+    except NameError as ne:
+        raise _CleanupError(
+            f'For annotation "{annotation}", could not resolve '
+            f'container type "{mm.group(1)}".  '
+            "Please ensure this type is imported at the module level "
+            "outside of TYPE_CHECKING blocks"
+        ) from ne
+
+    if obj is typing.ClassVar:
+        real_symbol = "ClassVar"
+    else:
+        try:
+            if issubclass(obj, _MappedAnnotationBase):
+                real_symbol = obj.__name__
+            else:
+                return annotation
+        except TypeError:
+            # avoid isinstance(obj, type) check, just catch TypeError
+            return annotation
+
+    # note: if one of the codepaths above didn't define real_symbol and
+    # then didn't return, real_symbol raises UnboundLocalError
+    # which is actually a NameError, and the calling routines don't
+    # notice this since they are catching NameError anyway.   Just in case
+    # this is being modified in the future, something to be aware of.
+
+    stack = []
+    inner = mm
+    while True:
+        stack.append(real_symbol if mm is inner else inner.group(1))
+        g2 = inner.group(2)
+        inner = re.match(r"^([^ \|]+?)\[(.+)\]$", g2)
+        if inner is None:
+            stack.append(g2)
+            break
+
+    # stacks we want to rewrite, that is, quote the last entry which
+    # we think is a relationship class name:
+    #
+    #   ['Mapped', 'List', 'Address']
+    #   ['Mapped', 'A']
+    #
+    # stacks we dont want to rewrite, which are generally MappedColumn
+    # use cases:
+    #
+    # ['Mapped', "'Optional[Dict[str, str]]'"]
+    # ['Mapped', 'dict[str, str] | None']
+
+    if (
+        # avoid already quoted symbols such as
+        # ['Mapped', "'Optional[Dict[str, str]]'"]
+        not re.match(r"""^["'].*["']$""", stack[-1])
+        # avoid further generics like Dict[] such as
+        # ['Mapped', 'dict[str, str] | None'],
+        # ['Mapped', 'list[int] | list[str]'],
+        # ['Mapped', 'Union[list[int], list[str]]'],
+        and not re.search(r"[\[\]]", stack[-1])
+    ):
+        stripchars = "\"' "
+        stack[-1] = ", ".join(
+            f'"{elem.strip(stripchars)}"' for elem in stack[-1].split(",")
+        )
+
+        annotation = "[".join(stack) + ("]" * (len(stack) - 1))
+
+    return annotation
+
+
+def _extract_mapped_subtype(
+    raw_annotation: Optional[_AnnotationScanType],
+    cls: type,
+    originating_module: str,
+    key: str,
+    attr_cls: Type[Any],
+    required: bool,
+    is_dataclass_field: bool,
+    expect_mapped: bool = True,
+    raiseerr: bool = True,
+) -> Optional[Tuple[Union[_AnnotationScanType, str], Optional[type]]]:
+    """given an annotation, figure out if it's ``Mapped[something]`` and if
+    so, return the ``something`` part.
+
+    Includes error raise scenarios and other options.
+
+    """
+
+    if raw_annotation is None:
+        if required:
+            raise sa_exc.ArgumentError(
+                f"Python typing annotation is required for attribute "
+                f'"{cls.__name__}.{key}" when primary argument(s) for '
+                f'"{attr_cls.__name__}" construct are None or not present'
+            )
+        return None
+
+    try:
+        # destringify the "outside" of the annotation.  note we are not
+        # adding include_generic so it will *not* dig into generic contents,
+        # which will remain as ForwardRef or plain str under future annotations
+        # mode.  The full destringify happens later when mapped_column goes
+        # to do a full lookup in the registry type_annotations_map.
+        annotated = de_stringify_annotation(
+            cls,
+            raw_annotation,
+            originating_module,
+            str_cleanup_fn=_cleanup_mapped_str_annotation,
+        )
+    except _CleanupError as ce:
+        raise sa_exc.ArgumentError(
+            f"Could not interpret annotation {raw_annotation}.  "
+            "Check that it uses names that are correctly imported at the "
+            "module level. See chained stack trace for more hints."
+        ) from ce
+    except NameError as ne:
+        if raiseerr and "Mapped[" in raw_annotation:  # type: ignore
+            raise sa_exc.ArgumentError(
+                f"Could not interpret annotation {raw_annotation}.  "
+                "Check that it uses names that are correctly imported at the "
+                "module level. See chained stack trace for more hints."
+            ) from ne
+
+        annotated = raw_annotation  # type: ignore
+
+    if is_dataclass_field:
+        return annotated, None
+    else:
+        if not hasattr(annotated, "__origin__") or not is_origin_of_cls(
+            annotated, _MappedAnnotationBase
+        ):
+            if expect_mapped:
+                if not raiseerr:
+                    return None
+
+                origin = getattr(annotated, "__origin__", None)
+                if origin is typing.ClassVar:
+                    return None
+
+                # check for other kind of ORM descriptor like AssociationProxy,
+                # don't raise for that (issue #9957)
+                elif isinstance(origin, type) and issubclass(
+                    origin, ORMDescriptor
+                ):
+                    return None
+
+                raise sa_exc.ArgumentError(
+                    f'Type annotation for "{cls.__name__}.{key}" '
+                    "can't be correctly interpreted for "
+                    "Annotated Declarative Table form.  ORM annotations "
+                    "should normally make use of the ``Mapped[]`` generic "
+                    "type, or other ORM-compatible generic type, as a "
+                    "container for the actual type, which indicates the "
+                    "intent that the attribute is mapped. "
+                    "Class variables that are not intended to be mapped "
+                    "by the ORM should use ClassVar[].  "
+                    "To allow Annotated Declarative to disregard legacy "
+                    "annotations which don't use Mapped[] to pass, set "
+                    '"__allow_unmapped__ = True" on the class or a '
+                    "superclass this class.",
+                    code="zlpr",
+                )
+
+            else:
+                return annotated, None
+
+        if len(annotated.__args__) != 1:
+            raise sa_exc.ArgumentError(
+                "Expected sub-type for Mapped[] annotation"
+            )
+
+        return (
+            # fix dict/list/set args to be ForwardRef, see #11814
+            fixup_container_fwd_refs(annotated.__args__[0]),
+            annotated.__origin__,
+        )
+
+
+def _mapper_property_as_plain_name(prop: Type[Any]) -> str:
+    if hasattr(prop, "_mapper_property_name"):
+        name = prop._mapper_property_name()
+    else:
+        name = None
+    return util.clsname_as_plain_name(prop, name)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/writeonly.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/writeonly.py
new file mode 100644
index 00000000..ac034a09
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/orm/writeonly.py
@@ -0,0 +1,678 @@
+# orm/writeonly.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Write-only collection API.
+
+This is an alternate mapped attribute style that only supports single-item
+collection mutation operations.   To read the collection, a select()
+object must be executed each time.
+
+.. versionadded:: 2.0
+
+
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Collection
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from sqlalchemy.sql import bindparam
+from . import attributes
+from . import interfaces
+from . import relationships
+from . import strategies
+from .base import NEVER_SET
+from .base import object_mapper
+from .base import PassiveFlag
+from .base import RelationshipDirection
+from .. import exc
+from .. import inspect
+from .. import log
+from .. import util
+from ..sql import delete
+from ..sql import insert
+from ..sql import select
+from ..sql import update
+from ..sql.dml import Delete
+from ..sql.dml import Insert
+from ..sql.dml import Update
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from . import QueryableAttribute
+    from ._typing import _InstanceDict
+    from .attributes import AttributeEventToken
+    from .base import LoaderCallableStatus
+    from .collections import _AdaptedCollectionProtocol
+    from .collections import CollectionAdapter
+    from .mapper import Mapper
+    from .relationships import _RelationshipOrderByArg
+    from .state import InstanceState
+    from .util import AliasedClass
+    from ..event import _Dispatch
+    from ..sql.selectable import FromClause
+    from ..sql.selectable import Select
+
+_T = TypeVar("_T", bound=Any)
+
+
+class WriteOnlyHistory(Generic[_T]):
+    """Overrides AttributeHistory to receive append/remove events directly."""
+
+    unchanged_items: util.OrderedIdentitySet
+    added_items: util.OrderedIdentitySet
+    deleted_items: util.OrderedIdentitySet
+    _reconcile_collection: bool
+
+    def __init__(
+        self,
+        attr: WriteOnlyAttributeImpl,
+        state: InstanceState[_T],
+        passive: PassiveFlag,
+        apply_to: Optional[WriteOnlyHistory[_T]] = None,
+    ) -> None:
+        if apply_to:
+            if passive & PassiveFlag.SQL_OK:
+                raise exc.InvalidRequestError(
+                    f"Attribute {attr} can't load the existing state from the "
+                    "database for this operation; full iteration is not "
+                    "permitted.  If this is a delete operation, configure "
+                    f"passive_deletes=True on the {attr} relationship in "
+                    "order to resolve this error."
+                )
+
+            self.unchanged_items = apply_to.unchanged_items
+            self.added_items = apply_to.added_items
+            self.deleted_items = apply_to.deleted_items
+            self._reconcile_collection = apply_to._reconcile_collection
+        else:
+            self.deleted_items = util.OrderedIdentitySet()
+            self.added_items = util.OrderedIdentitySet()
+            self.unchanged_items = util.OrderedIdentitySet()
+            self._reconcile_collection = False
+
+    @property
+    def added_plus_unchanged(self) -> List[_T]:
+        return list(self.added_items.union(self.unchanged_items))
+
+    @property
+    def all_items(self) -> List[_T]:
+        return list(
+            self.added_items.union(self.unchanged_items).union(
+                self.deleted_items
+            )
+        )
+
+    def as_history(self) -> attributes.History:
+        if self._reconcile_collection:
+            added = self.added_items.difference(self.unchanged_items)
+            deleted = self.deleted_items.intersection(self.unchanged_items)
+            unchanged = self.unchanged_items.difference(deleted)
+        else:
+            added, unchanged, deleted = (
+                self.added_items,
+                self.unchanged_items,
+                self.deleted_items,
+            )
+        return attributes.History(list(added), list(unchanged), list(deleted))
+
+    def indexed(self, index: Union[int, slice]) -> Union[List[_T], _T]:
+        return list(self.added_items)[index]
+
+    def add_added(self, value: _T) -> None:
+        self.added_items.add(value)
+
+    def add_removed(self, value: _T) -> None:
+        if value in self.added_items:
+            self.added_items.remove(value)
+        else:
+            self.deleted_items.add(value)
+
+
+class WriteOnlyAttributeImpl(
+    attributes.HasCollectionAdapter, attributes.AttributeImpl
+):
+    uses_objects: bool = True
+    default_accepts_scalar_loader: bool = False
+    supports_population: bool = False
+    _supports_dynamic_iteration: bool = False
+    collection: bool = False
+    dynamic: bool = True
+    order_by: _RelationshipOrderByArg = ()
+    collection_history_cls: Type[WriteOnlyHistory[Any]] = WriteOnlyHistory
+
+    query_class: Type[WriteOnlyCollection[Any]]
+
+    def __init__(
+        self,
+        class_: Union[Type[Any], AliasedClass[Any]],
+        key: str,
+        dispatch: _Dispatch[QueryableAttribute[Any]],
+        target_mapper: Mapper[_T],
+        order_by: _RelationshipOrderByArg,
+        **kw: Any,
+    ):
+        super().__init__(class_, key, None, dispatch, **kw)
+        self.target_mapper = target_mapper
+        self.query_class = WriteOnlyCollection
+        if order_by:
+            self.order_by = tuple(order_by)
+
+    def get(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> Union[util.OrderedIdentitySet, WriteOnlyCollection[Any]]:
+        if not passive & PassiveFlag.SQL_OK:
+            return self._get_collection_history(
+                state, PassiveFlag.PASSIVE_NO_INITIALIZE
+            ).added_items
+        else:
+            return self.query_class(self, state)
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Literal[None] = ...,
+        passive: Literal[PassiveFlag.PASSIVE_OFF] = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: _AdaptedCollectionProtocol = ...,
+        passive: PassiveFlag = ...,
+    ) -> CollectionAdapter: ...
+
+    @overload
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = ...,
+        passive: PassiveFlag = ...,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]: ...
+
+    def get_collection(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        user_data: Optional[_AdaptedCollectionProtocol] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+    ) -> Union[
+        Literal[LoaderCallableStatus.PASSIVE_NO_RESULT], CollectionAdapter
+    ]:
+        data: Collection[Any]
+        if not passive & PassiveFlag.SQL_OK:
+            data = self._get_collection_history(state, passive).added_items
+        else:
+            history = self._get_collection_history(state, passive)
+            data = history.added_plus_unchanged
+        return DynamicCollectionAdapter(data)  # type: ignore[return-value]
+
+    @util.memoized_property
+    def _append_token(  # type:ignore[override]
+        self,
+    ) -> attributes.AttributeEventToken:
+        return attributes.AttributeEventToken(self, attributes.OP_APPEND)
+
+    @util.memoized_property
+    def _remove_token(  # type:ignore[override]
+        self,
+    ) -> attributes.AttributeEventToken:
+        return attributes.AttributeEventToken(self, attributes.OP_REMOVE)
+
+    def fire_append_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        collection_history: Optional[WriteOnlyHistory[Any]] = None,
+    ) -> None:
+        if collection_history is None:
+            collection_history = self._modified_event(state, dict_)
+
+        collection_history.add_added(value)
+
+        for fn in self.dispatch.append:
+            value = fn(state, value, initiator or self._append_token)
+
+        if self.trackparent and value is not None:
+            self.sethasparent(attributes.instance_state(value), state, True)
+
+    def fire_remove_event(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        collection_history: Optional[WriteOnlyHistory[Any]] = None,
+    ) -> None:
+        if collection_history is None:
+            collection_history = self._modified_event(state, dict_)
+
+        collection_history.add_removed(value)
+
+        if self.trackparent and value is not None:
+            self.sethasparent(attributes.instance_state(value), state, False)
+
+        for fn in self.dispatch.remove:
+            fn(state, value, initiator or self._remove_token)
+
+    def _modified_event(
+        self, state: InstanceState[Any], dict_: _InstanceDict
+    ) -> WriteOnlyHistory[Any]:
+        if self.key not in state.committed_state:
+            state.committed_state[self.key] = self.collection_history_cls(
+                self, state, PassiveFlag.PASSIVE_NO_FETCH
+            )
+
+        state._modified_event(dict_, self, NEVER_SET)
+
+        # this is a hack to allow the entities.ComparableEntity fixture
+        # to work
+        dict_[self.key] = True
+        return state.committed_state[self.key]  # type: ignore[no-any-return]
+
+    def set(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken] = None,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_OFF,
+        check_old: Any = None,
+        pop: bool = False,
+        _adapt: bool = True,
+    ) -> None:
+        if initiator and initiator.parent_token is self.parent_token:
+            return
+
+        if pop and value is None:
+            return
+
+        iterable = value
+        new_values = list(iterable)
+        if state.has_identity:
+            if not self._supports_dynamic_iteration:
+                raise exc.InvalidRequestError(
+                    f'Collection "{self}" does not support implicit '
+                    "iteration; collection replacement operations "
+                    "can't be used"
+                )
+            old_collection = util.IdentitySet(
+                self.get(state, dict_, passive=passive)
+            )
+
+        collection_history = self._modified_event(state, dict_)
+        if not state.has_identity:
+            old_collection = collection_history.added_items
+        else:
+            old_collection = old_collection.union(
+                collection_history.added_items
+            )
+
+        constants = old_collection.intersection(new_values)
+        additions = util.IdentitySet(new_values).difference(constants)
+        removals = old_collection.difference(constants)
+
+        for member in new_values:
+            if member in additions:
+                self.fire_append_event(
+                    state,
+                    dict_,
+                    member,
+                    None,
+                    collection_history=collection_history,
+                )
+
+        for member in removals:
+            self.fire_remove_event(
+                state,
+                dict_,
+                member,
+                None,
+                collection_history=collection_history,
+            )
+
+    def delete(self, *args: Any, **kwargs: Any) -> NoReturn:
+        raise NotImplementedError()
+
+    def set_committed_value(
+        self, state: InstanceState[Any], dict_: _InstanceDict, value: Any
+    ) -> NoReturn:
+        raise NotImplementedError(
+            "Dynamic attributes don't support collection population."
+        )
+
+    def get_history(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_NO_FETCH,
+    ) -> attributes.History:
+        c = self._get_collection_history(state, passive)
+        return c.as_history()
+
+    def get_all_pending(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        passive: PassiveFlag = PassiveFlag.PASSIVE_NO_INITIALIZE,
+    ) -> List[Tuple[InstanceState[Any], Any]]:
+        c = self._get_collection_history(state, passive)
+        return [(attributes.instance_state(x), x) for x in c.all_items]
+
+    def _get_collection_history(
+        self, state: InstanceState[Any], passive: PassiveFlag
+    ) -> WriteOnlyHistory[Any]:
+        c: WriteOnlyHistory[Any]
+        if self.key in state.committed_state:
+            c = state.committed_state[self.key]
+        else:
+            c = self.collection_history_cls(
+                self, state, PassiveFlag.PASSIVE_NO_FETCH
+            )
+
+        if state.has_identity and (passive & PassiveFlag.INIT_OK):
+            return self.collection_history_cls(
+                self, state, passive, apply_to=c
+            )
+        else:
+            return c
+
+    def append(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PassiveFlag.PASSIVE_NO_FETCH,
+    ) -> None:
+        if initiator is not self:
+            self.fire_append_event(state, dict_, value, initiator)
+
+    def remove(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PassiveFlag.PASSIVE_NO_FETCH,
+    ) -> None:
+        if initiator is not self:
+            self.fire_remove_event(state, dict_, value, initiator)
+
+    def pop(
+        self,
+        state: InstanceState[Any],
+        dict_: _InstanceDict,
+        value: Any,
+        initiator: Optional[AttributeEventToken],
+        passive: PassiveFlag = PassiveFlag.PASSIVE_NO_FETCH,
+    ) -> None:
+        self.remove(state, dict_, value, initiator, passive=passive)
+
+
+@log.class_logger
+@relationships.RelationshipProperty.strategy_for(lazy="write_only")
+class WriteOnlyLoader(strategies.AbstractRelationshipLoader, log.Identified):
+    impl_class = WriteOnlyAttributeImpl
+
+    def init_class_attribute(self, mapper: Mapper[Any]) -> None:
+        self.is_class_level = True
+        if not self.uselist or self.parent_property.direction not in (
+            interfaces.ONETOMANY,
+            interfaces.MANYTOMANY,
+        ):
+            raise exc.InvalidRequestError(
+                "On relationship %s, 'dynamic' loaders cannot be used with "
+                "many-to-one/one-to-one relationships and/or "
+                "uselist=False." % self.parent_property
+            )
+
+        strategies._register_attribute(  # type: ignore[no-untyped-call]
+            self.parent_property,
+            mapper,
+            useobject=True,
+            impl_class=self.impl_class,
+            target_mapper=self.parent_property.mapper,
+            order_by=self.parent_property.order_by,
+            query_class=self.parent_property.query_class,
+        )
+
+
+class DynamicCollectionAdapter:
+    """simplified CollectionAdapter for internal API consistency"""
+
+    data: Collection[Any]
+
+    def __init__(self, data: Collection[Any]):
+        self.data = data
+
+    def __iter__(self) -> Iterator[Any]:
+        return iter(self.data)
+
+    def _reset_empty(self) -> None:
+        pass
+
+    def __len__(self) -> int:
+        return len(self.data)
+
+    def __bool__(self) -> bool:
+        return True
+
+
+class AbstractCollectionWriter(Generic[_T]):
+    """Virtual collection which includes append/remove methods that synchronize
+    into the attribute event system.
+
+    """
+
+    if not TYPE_CHECKING:
+        __slots__ = ()
+
+    instance: _T
+    _from_obj: Tuple[FromClause, ...]
+
+    def __init__(self, attr: WriteOnlyAttributeImpl, state: InstanceState[_T]):
+        instance = state.obj()
+        if TYPE_CHECKING:
+            assert instance
+        self.instance = instance
+        self.attr = attr
+
+        mapper = object_mapper(instance)
+        prop = mapper._props[self.attr.key]
+
+        if prop.secondary is not None:
+            # this is a hack right now.  The Query only knows how to
+            # make subsequent joins() without a given left-hand side
+            # from self._from_obj[0].  We need to ensure prop.secondary
+            # is in the FROM.  So we purposely put the mapper selectable
+            # in _from_obj[0] to ensure a user-defined join() later on
+            # doesn't fail, and secondary is then in _from_obj[1].
+
+            # note also, we are using the official ORM-annotated selectable
+            # from __clause_element__(), see #7868
+            self._from_obj = (prop.mapper.__clause_element__(), prop.secondary)
+        else:
+            self._from_obj = ()
+
+        self._where_criteria = (
+            prop._with_parent(instance, alias_secondary=False),
+        )
+
+        if self.attr.order_by:
+            self._order_by_clauses = self.attr.order_by
+        else:
+            self._order_by_clauses = ()
+
+    def _add_all_impl(self, iterator: Iterable[_T]) -> None:
+        for item in iterator:
+            self.attr.append(
+                attributes.instance_state(self.instance),
+                attributes.instance_dict(self.instance),
+                item,
+                None,
+            )
+
+    def _remove_impl(self, item: _T) -> None:
+        self.attr.remove(
+            attributes.instance_state(self.instance),
+            attributes.instance_dict(self.instance),
+            item,
+            None,
+        )
+
+
+class WriteOnlyCollection(AbstractCollectionWriter[_T]):
+    """Write-only collection which can synchronize changes into the
+    attribute event system.
+
+    The :class:`.WriteOnlyCollection` is used in a mapping by
+    using the ``"write_only"`` lazy loading strategy with
+    :func:`_orm.relationship`.     For background on this configuration,
+    see :ref:`write_only_relationship`.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :ref:`write_only_relationship`
+
+    """
+
+    __slots__ = (
+        "instance",
+        "attr",
+        "_where_criteria",
+        "_from_obj",
+        "_order_by_clauses",
+    )
+
+    def __iter__(self) -> NoReturn:
+        raise TypeError(
+            "WriteOnly collections don't support iteration in-place; "
+            "to query for collection items, use the select() method to "
+            "produce a SQL statement and execute it with session.scalars()."
+        )
+
+    def select(self) -> Select[Tuple[_T]]:
+        """Produce a :class:`_sql.Select` construct that represents the
+        rows within this instance-local :class:`_orm.WriteOnlyCollection`.
+
+        """
+        stmt = select(self.attr.target_mapper).where(*self._where_criteria)
+        if self._from_obj:
+            stmt = stmt.select_from(*self._from_obj)
+        if self._order_by_clauses:
+            stmt = stmt.order_by(*self._order_by_clauses)
+        return stmt
+
+    def insert(self) -> Insert:
+        """For one-to-many collections, produce a :class:`_dml.Insert` which
+        will insert new rows in terms of this this instance-local
+        :class:`_orm.WriteOnlyCollection`.
+
+        This construct is only supported for a :class:`_orm.Relationship`
+        that does **not** include the :paramref:`_orm.relationship.secondary`
+        parameter.  For relationships that refer to a many-to-many table,
+        use ordinary bulk insert techniques to produce new objects, then
+        use :meth:`_orm.AbstractCollectionWriter.add_all` to associate them
+        with the collection.
+
+
+        """
+
+        state = inspect(self.instance)
+        mapper = state.mapper
+        prop = mapper._props[self.attr.key]
+
+        if prop.direction is not RelationshipDirection.ONETOMANY:
+            raise exc.InvalidRequestError(
+                "Write only bulk INSERT only supported for one-to-many "
+                "collections; for many-to-many, use a separate bulk "
+                "INSERT along with add_all()."
+            )
+
+        dict_: Dict[str, Any] = {}
+
+        for l, r in prop.synchronize_pairs:
+            fn = prop._get_attr_w_warn_on_none(
+                mapper,
+                state,
+                state.dict,
+                l,
+            )
+
+            dict_[r.key] = bindparam(None, callable_=fn)
+
+        return insert(self.attr.target_mapper).values(**dict_)
+
+    def update(self) -> Update:
+        """Produce a :class:`_dml.Update` which will refer to rows in terms
+        of this instance-local :class:`_orm.WriteOnlyCollection`.
+
+        """
+        return update(self.attr.target_mapper).where(*self._where_criteria)
+
+    def delete(self) -> Delete:
+        """Produce a :class:`_dml.Delete` which will refer to rows in terms
+        of this instance-local :class:`_orm.WriteOnlyCollection`.
+
+        """
+        return delete(self.attr.target_mapper).where(*self._where_criteria)
+
+    def add_all(self, iterator: Iterable[_T]) -> None:
+        """Add an iterable of items to this :class:`_orm.WriteOnlyCollection`.
+
+        The given items will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        """
+        self._add_all_impl(iterator)
+
+    def add(self, item: _T) -> None:
+        """Add an item to this :class:`_orm.WriteOnlyCollection`.
+
+        The given item will be persisted to the database in terms of
+        the parent instance's collection on the next flush.
+
+        """
+        self._add_all_impl([item])
+
+    def remove(self, item: _T) -> None:
+        """Remove an item from this :class:`_orm.WriteOnlyCollection`.
+
+        The given item will be removed from the parent instance's collection on
+        the next flush.
+
+        """
+        self._remove_impl(item)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__init__.py
new file mode 100644
index 00000000..51bf0ec7
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__init__.py
@@ -0,0 +1,44 @@
+# pool/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+"""Connection pooling for DB-API connections.
+
+Provides a number of connection pool implementations for a variety of
+usage scenarios and thread behavior requirements imposed by the
+application, DB-API or database itself.
+
+Also provides a DB-API 2.0 connection proxying mechanism allowing
+regular DB-API connect() methods to be transparently managed by a
+SQLAlchemy connection pool.
+"""
+
+from . import events
+from .base import _AdhocProxiedConnection as _AdhocProxiedConnection
+from .base import _ConnectionFairy as _ConnectionFairy
+from .base import _ConnectionRecord
+from .base import _CreatorFnType as _CreatorFnType
+from .base import _CreatorWRecFnType as _CreatorWRecFnType
+from .base import _finalize_fairy
+from .base import _ResetStyleArgType as _ResetStyleArgType
+from .base import ConnectionPoolEntry as ConnectionPoolEntry
+from .base import ManagesConnection as ManagesConnection
+from .base import Pool as Pool
+from .base import PoolProxiedConnection as PoolProxiedConnection
+from .base import PoolResetState as PoolResetState
+from .base import reset_commit as reset_commit
+from .base import reset_none as reset_none
+from .base import reset_rollback as reset_rollback
+from .impl import AssertionPool as AssertionPool
+from .impl import AsyncAdaptedQueuePool as AsyncAdaptedQueuePool
+from .impl import (
+    FallbackAsyncAdaptedQueuePool as FallbackAsyncAdaptedQueuePool,
+)
+from .impl import NullPool as NullPool
+from .impl import QueuePool as QueuePool
+from .impl import SingletonThreadPool as SingletonThreadPool
+from .impl import StaticPool as StaticPool
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d1921c9e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..b4d51e8c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/events.cpython-312.pyc
new file mode 100644
index 00000000..1eac2763
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/impl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/impl.cpython-312.pyc
new file mode 100644
index 00000000..29551cd2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/__pycache__/impl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/base.py
new file mode 100644
index 00000000..34d02254
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/base.py
@@ -0,0 +1,1515 @@
+# pool/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+"""Base constructs for connection pools.
+
+"""
+
+from __future__ import annotations
+
+from collections import deque
+import dataclasses
+from enum import Enum
+import threading
+import time
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Deque
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+import weakref
+
+from .. import event
+from .. import exc
+from .. import log
+from .. import util
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if TYPE_CHECKING:
+    from ..engine.interfaces import DBAPIConnection
+    from ..engine.interfaces import DBAPICursor
+    from ..engine.interfaces import Dialect
+    from ..event import _DispatchCommon
+    from ..event import _ListenerFnType
+    from ..event import dispatcher
+    from ..sql._typing import _InfoType
+
+
+@dataclasses.dataclass(frozen=True)
+class PoolResetState:
+    """describes the state of a DBAPI connection as it is being passed to
+    the :meth:`.PoolEvents.reset` connection pool event.
+
+    .. versionadded:: 2.0.0b3
+
+    """
+
+    __slots__ = ("transaction_was_reset", "terminate_only", "asyncio_safe")
+
+    transaction_was_reset: bool
+    """Indicates if the transaction on the DBAPI connection was already
+    essentially "reset" back by the :class:`.Connection` object.
+
+    This boolean is True if the :class:`.Connection` had transactional
+    state present upon it, which was then not closed using the
+    :meth:`.Connection.rollback` or :meth:`.Connection.commit` method;
+    instead, the transaction was closed inline within the
+    :meth:`.Connection.close` method so is guaranteed to remain non-present
+    when this event is reached.
+
+    """
+
+    terminate_only: bool
+    """indicates if the connection is to be immediately terminated and
+    not checked in to the pool.
+
+    This occurs for connections that were invalidated, as well as asyncio
+    connections that were not cleanly handled by the calling code that
+    are instead being garbage collected.   In the latter case,
+    operations can't be safely run on asyncio connections within garbage
+    collection as there is not necessarily an event loop present.
+
+    """
+
+    asyncio_safe: bool
+    """Indicates if the reset operation is occurring within a scope where
+    an enclosing event loop is expected to be present for asyncio applications.
+
+    Will be False in the case that the connection is being garbage collected.
+
+    """
+
+
+class ResetStyle(Enum):
+    """Describe options for "reset on return" behaviors."""
+
+    reset_rollback = 0
+    reset_commit = 1
+    reset_none = 2
+
+
+_ResetStyleArgType = Union[
+    ResetStyle,
+    Literal[True, None, False, "commit", "rollback"],
+]
+reset_rollback, reset_commit, reset_none = list(ResetStyle)
+
+
+class _ConnDialect:
+    """partial implementation of :class:`.Dialect`
+    which provides DBAPI connection methods.
+
+    When a :class:`_pool.Pool` is combined with an :class:`_engine.Engine`,
+    the :class:`_engine.Engine` replaces this with its own
+    :class:`.Dialect`.
+
+    """
+
+    is_async = False
+    has_terminate = False
+
+    def do_rollback(self, dbapi_connection: PoolProxiedConnection) -> None:
+        dbapi_connection.rollback()
+
+    def do_commit(self, dbapi_connection: PoolProxiedConnection) -> None:
+        dbapi_connection.commit()
+
+    def do_terminate(self, dbapi_connection: DBAPIConnection) -> None:
+        dbapi_connection.close()
+
+    def do_close(self, dbapi_connection: DBAPIConnection) -> None:
+        dbapi_connection.close()
+
+    def _do_ping_w_event(self, dbapi_connection: DBAPIConnection) -> bool:
+        raise NotImplementedError(
+            "The ping feature requires that a dialect is "
+            "passed to the connection pool."
+        )
+
+    def get_driver_connection(self, connection: DBAPIConnection) -> Any:
+        return connection
+
+
+class _AsyncConnDialect(_ConnDialect):
+    is_async = True
+
+
+class _CreatorFnType(Protocol):
+    def __call__(self) -> DBAPIConnection: ...
+
+
+class _CreatorWRecFnType(Protocol):
+    def __call__(self, rec: ConnectionPoolEntry) -> DBAPIConnection: ...
+
+
+class Pool(log.Identified, event.EventTarget):
+    """Abstract base class for connection pools."""
+
+    dispatch: dispatcher[Pool]
+    echo: log._EchoFlagType
+
+    _orig_logging_name: Optional[str]
+    _dialect: Union[_ConnDialect, Dialect] = _ConnDialect()
+    _creator_arg: Union[_CreatorFnType, _CreatorWRecFnType]
+    _invoke_creator: _CreatorWRecFnType
+    _invalidate_time: float
+
+    def __init__(
+        self,
+        creator: Union[_CreatorFnType, _CreatorWRecFnType],
+        recycle: int = -1,
+        echo: log._EchoFlagType = None,
+        logging_name: Optional[str] = None,
+        reset_on_return: _ResetStyleArgType = True,
+        events: Optional[List[Tuple[_ListenerFnType, str]]] = None,
+        dialect: Optional[Union[_ConnDialect, Dialect]] = None,
+        pre_ping: bool = False,
+        _dispatch: Optional[_DispatchCommon[Pool]] = None,
+    ):
+        """
+        Construct a Pool.
+
+        :param creator: a callable function that returns a DB-API
+          connection object.  The function will be called with
+          parameters.
+
+        :param recycle: If set to a value other than -1, number of
+          seconds between connection recycling, which means upon
+          checkout, if this timeout is surpassed the connection will be
+          closed and replaced with a newly opened connection. Defaults to -1.
+
+        :param logging_name:  String identifier which will be used within
+          the "name" field of logging records generated within the
+          "sqlalchemy.pool" logger. Defaults to a hexstring of the object's
+          id.
+
+        :param echo: if True, the connection pool will log
+         informational output such as when connections are invalidated
+         as well as when connections are recycled to the default log handler,
+         which defaults to ``sys.stdout`` for output..   If set to the string
+         ``"debug"``, the logging will include pool checkouts and checkins.
+
+         The :paramref:`_pool.Pool.echo` parameter can also be set from the
+         :func:`_sa.create_engine` call by using the
+         :paramref:`_sa.create_engine.echo_pool` parameter.
+
+         .. seealso::
+
+             :ref:`dbengine_logging` - further detail on how to configure
+             logging.
+
+        :param reset_on_return: Determine steps to take on
+         connections as they are returned to the pool, which were
+         not otherwise handled by a :class:`_engine.Connection`.
+         Available from :func:`_sa.create_engine` via the
+         :paramref:`_sa.create_engine.pool_reset_on_return` parameter.
+
+         :paramref:`_pool.Pool.reset_on_return` can have any of these values:
+
+         * ``"rollback"`` - call rollback() on the connection,
+           to release locks and transaction resources.
+           This is the default value.  The vast majority
+           of use cases should leave this value set.
+         * ``"commit"`` - call commit() on the connection,
+           to release locks and transaction resources.
+           A commit here may be desirable for databases that
+           cache query plans if a commit is emitted,
+           such as Microsoft SQL Server.  However, this
+           value is more dangerous than 'rollback' because
+           any data changes present on the transaction
+           are committed unconditionally.
+         * ``None`` - don't do anything on the connection.
+           This setting may be appropriate if the database / DBAPI
+           works in pure "autocommit" mode at all times, or if
+           a custom reset handler is established using the
+           :meth:`.PoolEvents.reset` event handler.
+
+         * ``True`` - same as 'rollback', this is here for
+           backwards compatibility.
+         * ``False`` - same as None, this is here for
+           backwards compatibility.
+
+         For further customization of reset on return, the
+         :meth:`.PoolEvents.reset` event hook may be used which can perform
+         any connection activity desired on reset.
+
+         .. seealso::
+
+            :ref:`pool_reset_on_return`
+
+            :meth:`.PoolEvents.reset`
+
+        :param events: a list of 2-tuples, each of the form
+         ``(callable, target)`` which will be passed to :func:`.event.listen`
+         upon construction.   Provided here so that event listeners
+         can be assigned via :func:`_sa.create_engine` before dialect-level
+         listeners are applied.
+
+        :param dialect: a :class:`.Dialect` that will handle the job
+         of calling rollback(), close(), or commit() on DBAPI connections.
+         If omitted, a built-in "stub" dialect is used.   Applications that
+         make use of :func:`_sa.create_engine` should not use this parameter
+         as it is handled by the engine creation strategy.
+
+        :param pre_ping: if True, the pool will emit a "ping" (typically
+         "SELECT 1", but is dialect-specific) on the connection
+         upon checkout, to test if the connection is alive or not.   If not,
+         the connection is transparently re-connected and upon success, all
+         other pooled connections established prior to that timestamp are
+         invalidated.     Requires that a dialect is passed as well to
+         interpret the disconnection error.
+
+         .. versionadded:: 1.2
+
+        """
+        if logging_name:
+            self.logging_name = self._orig_logging_name = logging_name
+        else:
+            self._orig_logging_name = None
+
+        log.instance_logger(self, echoflag=echo)
+        self._creator = creator
+        self._recycle = recycle
+        self._invalidate_time = 0
+        self._pre_ping = pre_ping
+        self._reset_on_return = util.parse_user_argument_for_enum(
+            reset_on_return,
+            {
+                ResetStyle.reset_rollback: ["rollback", True],
+                ResetStyle.reset_none: ["none", None, False],
+                ResetStyle.reset_commit: ["commit"],
+            },
+            "reset_on_return",
+        )
+
+        self.echo = echo
+
+        if _dispatch:
+            self.dispatch._update(_dispatch, only_propagate=False)
+        if dialect:
+            self._dialect = dialect
+        if events:
+            for fn, target in events:
+                event.listen(self, target, fn)
+
+    @util.hybridproperty
+    def _is_asyncio(self) -> bool:
+        return self._dialect.is_async
+
+    @property
+    def _creator(self) -> Union[_CreatorFnType, _CreatorWRecFnType]:
+        return self._creator_arg
+
+    @_creator.setter
+    def _creator(
+        self, creator: Union[_CreatorFnType, _CreatorWRecFnType]
+    ) -> None:
+        self._creator_arg = creator
+
+        # mypy seems to get super confused assigning functions to
+        # attributes
+        self._invoke_creator = self._should_wrap_creator(creator)
+
+    @_creator.deleter
+    def _creator(self) -> None:
+        # needed for mock testing
+        del self._creator_arg
+        del self._invoke_creator
+
+    def _should_wrap_creator(
+        self, creator: Union[_CreatorFnType, _CreatorWRecFnType]
+    ) -> _CreatorWRecFnType:
+        """Detect if creator accepts a single argument, or is sent
+        as a legacy style no-arg function.
+
+        """
+
+        try:
+            argspec = util.get_callable_argspec(self._creator, no_self=True)
+        except TypeError:
+            creator_fn = cast(_CreatorFnType, creator)
+            return lambda rec: creator_fn()
+
+        if argspec.defaults is not None:
+            defaulted = len(argspec.defaults)
+        else:
+            defaulted = 0
+        positionals = len(argspec[0]) - defaulted
+
+        # look for the exact arg signature that DefaultStrategy
+        # sends us
+        if (argspec[0], argspec[3]) == (["connection_record"], (None,)):
+            return cast(_CreatorWRecFnType, creator)
+        # or just a single positional
+        elif positionals == 1:
+            return cast(_CreatorWRecFnType, creator)
+        # all other cases, just wrap and assume legacy "creator" callable
+        # thing
+        else:
+            creator_fn = cast(_CreatorFnType, creator)
+            return lambda rec: creator_fn()
+
+    def _close_connection(
+        self, connection: DBAPIConnection, *, terminate: bool = False
+    ) -> None:
+        self.logger.debug(
+            "%s connection %r",
+            "Hard-closing" if terminate else "Closing",
+            connection,
+        )
+        try:
+            if terminate:
+                self._dialect.do_terminate(connection)
+            else:
+                self._dialect.do_close(connection)
+        except BaseException as e:
+            self.logger.error(
+                f"Exception {'terminating' if terminate else 'closing'} "
+                f"connection %r",
+                connection,
+                exc_info=True,
+            )
+            if not isinstance(e, Exception):
+                raise
+
+    def _create_connection(self) -> ConnectionPoolEntry:
+        """Called by subclasses to create a new ConnectionRecord."""
+
+        return _ConnectionRecord(self)
+
+    def _invalidate(
+        self,
+        connection: PoolProxiedConnection,
+        exception: Optional[BaseException] = None,
+        _checkin: bool = True,
+    ) -> None:
+        """Mark all connections established within the generation
+        of the given connection as invalidated.
+
+        If this pool's last invalidate time is before when the given
+        connection was created, update the timestamp til now.  Otherwise,
+        no action is performed.
+
+        Connections with a start time prior to this pool's invalidation
+        time will be recycled upon next checkout.
+        """
+        rec = getattr(connection, "_connection_record", None)
+        if not rec or self._invalidate_time < rec.starttime:
+            self._invalidate_time = time.time()
+        if _checkin and getattr(connection, "is_valid", False):
+            connection.invalidate(exception)
+
+    def recreate(self) -> Pool:
+        """Return a new :class:`_pool.Pool`, of the same class as this one
+        and configured with identical creation arguments.
+
+        This method is used in conjunction with :meth:`dispose`
+        to close out an entire :class:`_pool.Pool` and create a new one in
+        its place.
+
+        """
+
+        raise NotImplementedError()
+
+    def dispose(self) -> None:
+        """Dispose of this pool.
+
+        This method leaves the possibility of checked-out connections
+        remaining open, as it only affects connections that are
+        idle in the pool.
+
+        .. seealso::
+
+            :meth:`Pool.recreate`
+
+        """
+
+        raise NotImplementedError()
+
+    def connect(self) -> PoolProxiedConnection:
+        """Return a DBAPI connection from the pool.
+
+        The connection is instrumented such that when its
+        ``close()`` method is called, the connection will be returned to
+        the pool.
+
+        """
+        return _ConnectionFairy._checkout(self)
+
+    def _return_conn(self, record: ConnectionPoolEntry) -> None:
+        """Given a _ConnectionRecord, return it to the :class:`_pool.Pool`.
+
+        This method is called when an instrumented DBAPI connection
+        has its ``close()`` method called.
+
+        """
+        self._do_return_conn(record)
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        """Implementation for :meth:`get`, supplied by subclasses."""
+
+        raise NotImplementedError()
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        """Implementation for :meth:`return_conn`, supplied by subclasses."""
+
+        raise NotImplementedError()
+
+    def status(self) -> str:
+        raise NotImplementedError()
+
+
+class ManagesConnection:
+    """Common base for the two connection-management interfaces
+    :class:`.PoolProxiedConnection` and :class:`.ConnectionPoolEntry`.
+
+    These two objects are typically exposed in the public facing API
+    via the connection pool event hooks, documented at :class:`.PoolEvents`.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __slots__ = ()
+
+    dbapi_connection: Optional[DBAPIConnection]
+    """A reference to the actual DBAPI connection being tracked.
+
+    This is a :pep:`249`-compliant object that for traditional sync-style
+    dialects is provided by the third-party
+    DBAPI implementation in use.  For asyncio dialects, the implementation
+    is typically an adapter object provided by the SQLAlchemy dialect
+    itself; the underlying asyncio object is available via the
+    :attr:`.ManagesConnection.driver_connection` attribute.
+
+    SQLAlchemy's interface for the DBAPI connection is based on the
+    :class:`.DBAPIConnection` protocol object
+
+    .. seealso::
+
+        :attr:`.ManagesConnection.driver_connection`
+
+        :ref:`faq_dbapi_connection`
+
+    """
+
+    driver_connection: Optional[Any]
+    """The "driver level" connection object as used by the Python
+    DBAPI or database driver.
+
+    For traditional :pep:`249` DBAPI implementations, this object will
+    be the same object as that of
+    :attr:`.ManagesConnection.dbapi_connection`.   For an asyncio database
+    driver, this will be the ultimate "connection" object used by that
+    driver, such as the ``asyncpg.Connection`` object which will not have
+    standard pep-249 methods.
+
+    .. versionadded:: 1.4.24
+
+    .. seealso::
+
+        :attr:`.ManagesConnection.dbapi_connection`
+
+        :ref:`faq_dbapi_connection`
+
+    """
+
+    @util.ro_memoized_property
+    def info(self) -> _InfoType:
+        """Info dictionary associated with the underlying DBAPI connection
+        referred to by this :class:`.ManagesConnection` instance, allowing
+        user-defined data to be associated with the connection.
+
+        The data in this dictionary is persistent for the lifespan
+        of the DBAPI connection itself, including across pool checkins
+        and checkouts.  When the connection is invalidated
+        and replaced with a new one, this dictionary is cleared.
+
+        For a :class:`.PoolProxiedConnection` instance that's not associated
+        with a :class:`.ConnectionPoolEntry`, such as if it were detached, the
+        attribute returns a dictionary that is local to that
+        :class:`.ConnectionPoolEntry`. Therefore the
+        :attr:`.ManagesConnection.info` attribute will always provide a Python
+        dictionary.
+
+        .. seealso::
+
+            :attr:`.ManagesConnection.record_info`
+
+
+        """
+        raise NotImplementedError()
+
+    @util.ro_memoized_property
+    def record_info(self) -> Optional[_InfoType]:
+        """Persistent info dictionary associated with this
+        :class:`.ManagesConnection`.
+
+        Unlike the :attr:`.ManagesConnection.info` dictionary, the lifespan
+        of this dictionary is that of the :class:`.ConnectionPoolEntry`
+        which owns it; therefore this dictionary will persist across
+        reconnects and connection invalidation for a particular entry
+        in the connection pool.
+
+        For a :class:`.PoolProxiedConnection` instance that's not associated
+        with a :class:`.ConnectionPoolEntry`, such as if it were detached, the
+        attribute returns None. Contrast to the :attr:`.ManagesConnection.info`
+        dictionary which is never None.
+
+
+        .. seealso::
+
+            :attr:`.ManagesConnection.info`
+
+        """
+        raise NotImplementedError()
+
+    def invalidate(
+        self, e: Optional[BaseException] = None, soft: bool = False
+    ) -> None:
+        """Mark the managed connection as invalidated.
+
+        :param e: an exception object indicating a reason for the invalidation.
+
+        :param soft: if True, the connection isn't closed; instead, this
+         connection will be recycled on next checkout.
+
+        .. seealso::
+
+            :ref:`pool_connection_invalidation`
+
+
+        """
+        raise NotImplementedError()
+
+
+class ConnectionPoolEntry(ManagesConnection):
+    """Interface for the object that maintains an individual database
+    connection on behalf of a :class:`_pool.Pool` instance.
+
+    The :class:`.ConnectionPoolEntry` object represents the long term
+    maintainance of a particular connection for a pool, including expiring or
+    invalidating that connection to have it replaced with a new one, which will
+    continue to be maintained by that same :class:`.ConnectionPoolEntry`
+    instance. Compared to :class:`.PoolProxiedConnection`, which is the
+    short-term, per-checkout connection manager, this object lasts for the
+    lifespan of a particular "slot" within a connection pool.
+
+    The :class:`.ConnectionPoolEntry` object is mostly visible to public-facing
+    API code when it is delivered to connection pool event hooks, such as
+    :meth:`_events.PoolEvents.connect` and :meth:`_events.PoolEvents.checkout`.
+
+    .. versionadded:: 2.0  :class:`.ConnectionPoolEntry` provides the public
+       facing interface for the :class:`._ConnectionRecord` internal class.
+
+    """
+
+    __slots__ = ()
+
+    @property
+    def in_use(self) -> bool:
+        """Return True the connection is currently checked out"""
+
+        raise NotImplementedError()
+
+    def close(self) -> None:
+        """Close the DBAPI connection managed by this connection pool entry."""
+        raise NotImplementedError()
+
+
+class _ConnectionRecord(ConnectionPoolEntry):
+    """Maintains a position in a connection pool which references a pooled
+    connection.
+
+    This is an internal object used by the :class:`_pool.Pool` implementation
+    to provide context management to a DBAPI connection maintained by
+    that :class:`_pool.Pool`.   The public facing interface for this class
+    is described by the :class:`.ConnectionPoolEntry` class.  See that
+    class for public API details.
+
+    .. seealso::
+
+        :class:`.ConnectionPoolEntry`
+
+        :class:`.PoolProxiedConnection`
+
+    """
+
+    __slots__ = (
+        "__pool",
+        "fairy_ref",
+        "finalize_callback",
+        "fresh",
+        "starttime",
+        "dbapi_connection",
+        "__weakref__",
+        "__dict__",
+    )
+
+    finalize_callback: Deque[Callable[[DBAPIConnection], None]]
+    fresh: bool
+    fairy_ref: Optional[weakref.ref[_ConnectionFairy]]
+    starttime: float
+
+    def __init__(self, pool: Pool, connect: bool = True):
+        self.fresh = False
+        self.fairy_ref = None
+        self.starttime = 0
+        self.dbapi_connection = None
+
+        self.__pool = pool
+        if connect:
+            self.__connect()
+        self.finalize_callback = deque()
+
+    dbapi_connection: Optional[DBAPIConnection]
+
+    @property
+    def driver_connection(self) -> Optional[Any]:  # type: ignore[override]  # mypy#4125  # noqa: E501
+        if self.dbapi_connection is None:
+            return None
+        else:
+            return self.__pool._dialect.get_driver_connection(
+                self.dbapi_connection
+            )
+
+    @property
+    @util.deprecated(
+        "2.0",
+        "The _ConnectionRecord.connection attribute is deprecated; "
+        "please use 'driver_connection'",
+    )
+    def connection(self) -> Optional[DBAPIConnection]:
+        return self.dbapi_connection
+
+    _soft_invalidate_time: float = 0
+
+    @util.ro_memoized_property
+    def info(self) -> _InfoType:
+        return {}
+
+    @util.ro_memoized_property
+    def record_info(self) -> Optional[_InfoType]:
+        return {}
+
+    @classmethod
+    def checkout(cls, pool: Pool) -> _ConnectionFairy:
+        if TYPE_CHECKING:
+            rec = cast(_ConnectionRecord, pool._do_get())
+        else:
+            rec = pool._do_get()
+
+        try:
+            dbapi_connection = rec.get_connection()
+        except BaseException as err:
+            with util.safe_reraise():
+                rec._checkin_failed(err, _fairy_was_created=False)
+
+            # not reached, for code linters only
+            raise
+
+        echo = pool._should_log_debug()
+        fairy = _ConnectionFairy(pool, dbapi_connection, rec, echo)
+
+        rec.fairy_ref = ref = weakref.ref(
+            fairy,
+            lambda ref: (
+                _finalize_fairy(
+                    None, rec, pool, ref, echo, transaction_was_reset=False
+                )
+                if _finalize_fairy is not None
+                else None
+            ),
+        )
+        _strong_ref_connection_records[ref] = rec
+        if echo:
+            pool.logger.debug(
+                "Connection %r checked out from pool", dbapi_connection
+            )
+        return fairy
+
+    def _checkin_failed(
+        self, err: BaseException, _fairy_was_created: bool = True
+    ) -> None:
+        self.invalidate(e=err)
+        self.checkin(
+            _fairy_was_created=_fairy_was_created,
+        )
+
+    def checkin(self, _fairy_was_created: bool = True) -> None:
+        if self.fairy_ref is None and _fairy_was_created:
+            # _fairy_was_created is False for the initial get connection phase;
+            # meaning there was no _ConnectionFairy and we must unconditionally
+            # do a checkin.
+            #
+            # otherwise, if fairy_was_created==True, if fairy_ref is None here
+            # that means we were checked in already, so this looks like
+            # a double checkin.
+            util.warn("Double checkin attempted on %s" % self)
+            return
+        self.fairy_ref = None
+        connection = self.dbapi_connection
+        pool = self.__pool
+        while self.finalize_callback:
+            finalizer = self.finalize_callback.pop()
+            if connection is not None:
+                finalizer(connection)
+        if pool.dispatch.checkin:
+            pool.dispatch.checkin(connection, self)
+
+        pool._return_conn(self)
+
+    @property
+    def in_use(self) -> bool:
+        return self.fairy_ref is not None
+
+    @property
+    def last_connect_time(self) -> float:
+        return self.starttime
+
+    def close(self) -> None:
+        if self.dbapi_connection is not None:
+            self.__close()
+
+    def invalidate(
+        self, e: Optional[BaseException] = None, soft: bool = False
+    ) -> None:
+        # already invalidated
+        if self.dbapi_connection is None:
+            return
+        if soft:
+            self.__pool.dispatch.soft_invalidate(
+                self.dbapi_connection, self, e
+            )
+        else:
+            self.__pool.dispatch.invalidate(self.dbapi_connection, self, e)
+        if e is not None:
+            self.__pool.logger.info(
+                "%sInvalidate connection %r (reason: %s:%s)",
+                "Soft " if soft else "",
+                self.dbapi_connection,
+                e.__class__.__name__,
+                e,
+            )
+        else:
+            self.__pool.logger.info(
+                "%sInvalidate connection %r",
+                "Soft " if soft else "",
+                self.dbapi_connection,
+            )
+
+        if soft:
+            self._soft_invalidate_time = time.time()
+        else:
+            self.__close(terminate=True)
+            self.dbapi_connection = None
+
+    def get_connection(self) -> DBAPIConnection:
+        recycle = False
+
+        # NOTE: the various comparisons here are assuming that measurable time
+        # passes between these state changes.  however, time.time() is not
+        # guaranteed to have sub-second precision.  comparisons of
+        # "invalidation time" to "starttime" should perhaps use >= so that the
+        # state change can take place assuming no measurable  time has passed,
+        # however this does not guarantee correct behavior here as if time
+        # continues to not pass, it will try to reconnect repeatedly until
+        # these timestamps diverge, so in that sense using > is safer.  Per
+        # https://stackoverflow.com/a/1938096/34549, Windows time.time() may be
+        # within 16 milliseconds accuracy, so unit tests for connection
+        # invalidation need a sleep of at least this long between initial start
+        # time and invalidation for the logic below to work reliably.
+
+        if self.dbapi_connection is None:
+            self.info.clear()
+            self.__connect()
+        elif (
+            self.__pool._recycle > -1
+            and time.time() - self.starttime > self.__pool._recycle
+        ):
+            self.__pool.logger.info(
+                "Connection %r exceeded timeout; recycling",
+                self.dbapi_connection,
+            )
+            recycle = True
+        elif self.__pool._invalidate_time > self.starttime:
+            self.__pool.logger.info(
+                "Connection %r invalidated due to pool invalidation; "
+                + "recycling",
+                self.dbapi_connection,
+            )
+            recycle = True
+        elif self._soft_invalidate_time > self.starttime:
+            self.__pool.logger.info(
+                "Connection %r invalidated due to local soft invalidation; "
+                + "recycling",
+                self.dbapi_connection,
+            )
+            recycle = True
+
+        if recycle:
+            self.__close(terminate=True)
+            self.info.clear()
+
+            self.__connect()
+
+        assert self.dbapi_connection is not None
+        return self.dbapi_connection
+
+    def _is_hard_or_soft_invalidated(self) -> bool:
+        return (
+            self.dbapi_connection is None
+            or self.__pool._invalidate_time > self.starttime
+            or (self._soft_invalidate_time > self.starttime)
+        )
+
+    def __close(self, *, terminate: bool = False) -> None:
+        self.finalize_callback.clear()
+        if self.__pool.dispatch.close:
+            self.__pool.dispatch.close(self.dbapi_connection, self)
+        assert self.dbapi_connection is not None
+        self.__pool._close_connection(
+            self.dbapi_connection, terminate=terminate
+        )
+        self.dbapi_connection = None
+
+    def __connect(self) -> None:
+        pool = self.__pool
+
+        # ensure any existing connection is removed, so that if
+        # creator fails, this attribute stays None
+        self.dbapi_connection = None
+        try:
+            self.starttime = time.time()
+            self.dbapi_connection = connection = pool._invoke_creator(self)
+            pool.logger.debug("Created new connection %r", connection)
+            self.fresh = True
+        except BaseException as e:
+            with util.safe_reraise():
+                pool.logger.debug("Error on connect(): %s", e)
+        else:
+            # in SQLAlchemy 1.4 the first_connect event is not used by
+            # the engine, so this will usually not be set
+            if pool.dispatch.first_connect:
+                pool.dispatch.first_connect.for_modify(
+                    pool.dispatch
+                ).exec_once_unless_exception(self.dbapi_connection, self)
+
+            # init of the dialect now takes place within the connect
+            # event, so ensure a mutex is used on the first run
+            pool.dispatch.connect.for_modify(
+                pool.dispatch
+            )._exec_w_sync_on_first_run(self.dbapi_connection, self)
+
+
+def _finalize_fairy(
+    dbapi_connection: Optional[DBAPIConnection],
+    connection_record: Optional[_ConnectionRecord],
+    pool: Pool,
+    ref: Optional[
+        weakref.ref[_ConnectionFairy]
+    ],  # this is None when called directly, not by the gc
+    echo: Optional[log._EchoFlagType],
+    transaction_was_reset: bool = False,
+    fairy: Optional[_ConnectionFairy] = None,
+) -> None:
+    """Cleanup for a :class:`._ConnectionFairy` whether or not it's already
+    been garbage collected.
+
+    When using an async dialect no IO can happen here (without using
+    a dedicated thread), since this is called outside the greenlet
+    context and with an already running loop. In this case function
+    will only log a message and raise a warning.
+    """
+
+    is_gc_cleanup = ref is not None
+
+    if is_gc_cleanup:
+        assert ref is not None
+        _strong_ref_connection_records.pop(ref, None)
+        assert connection_record is not None
+        if connection_record.fairy_ref is not ref:
+            return
+        assert dbapi_connection is None
+        dbapi_connection = connection_record.dbapi_connection
+
+    elif fairy:
+        _strong_ref_connection_records.pop(weakref.ref(fairy), None)
+
+    # null pool is not _is_asyncio but can be used also with async dialects
+    dont_restore_gced = pool._dialect.is_async
+
+    if dont_restore_gced:
+        detach = connection_record is None or is_gc_cleanup
+        can_manipulate_connection = not is_gc_cleanup
+        can_close_or_terminate_connection = (
+            not pool._dialect.is_async or pool._dialect.has_terminate
+        )
+        requires_terminate_for_close = (
+            pool._dialect.is_async and pool._dialect.has_terminate
+        )
+
+    else:
+        detach = connection_record is None
+        can_manipulate_connection = can_close_or_terminate_connection = True
+        requires_terminate_for_close = False
+
+    if dbapi_connection is not None:
+        if connection_record and echo:
+            pool.logger.debug(
+                "Connection %r being returned to pool", dbapi_connection
+            )
+
+        try:
+            if not fairy:
+                assert connection_record is not None
+                fairy = _ConnectionFairy(
+                    pool,
+                    dbapi_connection,
+                    connection_record,
+                    echo,
+                )
+            assert fairy.dbapi_connection is dbapi_connection
+
+            fairy._reset(
+                pool,
+                transaction_was_reset=transaction_was_reset,
+                terminate_only=detach,
+                asyncio_safe=can_manipulate_connection,
+            )
+
+            if detach:
+                if connection_record:
+                    fairy._pool = pool
+                    fairy.detach()
+
+                if can_close_or_terminate_connection:
+                    if pool.dispatch.close_detached:
+                        pool.dispatch.close_detached(dbapi_connection)
+
+                    pool._close_connection(
+                        dbapi_connection,
+                        terminate=requires_terminate_for_close,
+                    )
+
+        except BaseException as e:
+            pool.logger.error(
+                "Exception during reset or similar", exc_info=True
+            )
+            if connection_record:
+                connection_record.invalidate(e=e)
+            if not isinstance(e, Exception):
+                raise
+        finally:
+            if detach and is_gc_cleanup and dont_restore_gced:
+                message = (
+                    "The garbage collector is trying to clean up "
+                    f"non-checked-in connection {dbapi_connection!r}, "
+                    f"""which will be {
+                        'dropped, as it cannot be safely terminated'
+                        if not can_close_or_terminate_connection
+                        else 'terminated'
+                    }.  """
+                    "Please ensure that SQLAlchemy pooled connections are "
+                    "returned to "
+                    "the pool explicitly, either by calling ``close()`` "
+                    "or by using appropriate context managers to manage "
+                    "their lifecycle."
+                )
+                pool.logger.error(message)
+                util.warn(message)
+
+    if connection_record and connection_record.fairy_ref is not None:
+        connection_record.checkin()
+
+    # give gc some help.  See
+    # test/engine/test_pool.py::PoolEventsTest::test_checkin_event_gc[True]
+    # which actually started failing when pytest warnings plugin was
+    # turned on, due to util.warn() above
+    if fairy is not None:
+        fairy.dbapi_connection = None  # type: ignore
+        fairy._connection_record = None
+    del dbapi_connection
+    del connection_record
+    del fairy
+
+
+# a dictionary of the _ConnectionFairy weakrefs to _ConnectionRecord, so that
+# GC under pypy will call ConnectionFairy finalizers.  linked directly to the
+# weakref that will empty itself when collected so that it should not create
+# any unmanaged memory references.
+_strong_ref_connection_records: Dict[
+    weakref.ref[_ConnectionFairy], _ConnectionRecord
+] = {}
+
+
+class PoolProxiedConnection(ManagesConnection):
+    """A connection-like adapter for a :pep:`249` DBAPI connection, which
+    includes additional methods specific to the :class:`.Pool` implementation.
+
+    :class:`.PoolProxiedConnection` is the public-facing interface for the
+    internal :class:`._ConnectionFairy` implementation object; users familiar
+    with :class:`._ConnectionFairy` can consider this object to be equivalent.
+
+    .. versionadded:: 2.0  :class:`.PoolProxiedConnection` provides the public-
+       facing interface for the :class:`._ConnectionFairy` internal class.
+
+    """
+
+    __slots__ = ()
+
+    if typing.TYPE_CHECKING:
+
+        def commit(self) -> None: ...
+
+        def cursor(self) -> DBAPICursor: ...
+
+        def rollback(self) -> None: ...
+
+    @property
+    def is_valid(self) -> bool:
+        """Return True if this :class:`.PoolProxiedConnection` still refers
+        to an active DBAPI connection."""
+
+        raise NotImplementedError()
+
+    @property
+    def is_detached(self) -> bool:
+        """Return True if this :class:`.PoolProxiedConnection` is detached
+        from its pool."""
+
+        raise NotImplementedError()
+
+    def detach(self) -> None:
+        """Separate this connection from its Pool.
+
+        This means that the connection will no longer be returned to the
+        pool when closed, and will instead be literally closed.  The
+        associated :class:`.ConnectionPoolEntry` is de-associated from this
+        DBAPI connection.
+
+        Note that any overall connection limiting constraints imposed by a
+        Pool implementation may be violated after a detach, as the detached
+        connection is removed from the pool's knowledge and control.
+
+        """
+
+        raise NotImplementedError()
+
+    def close(self) -> None:
+        """Release this connection back to the pool.
+
+        The :meth:`.PoolProxiedConnection.close` method shadows the
+        :pep:`249` ``.close()`` method, altering its behavior to instead
+        :term:`release` the proxied connection back to the connection pool.
+
+        Upon release to the pool, whether the connection stays "opened" and
+        pooled in the Python process, versus actually closed out and removed
+        from the Python process, is based on the pool implementation in use and
+        its configuration and current state.
+
+        """
+        raise NotImplementedError()
+
+
+class _AdhocProxiedConnection(PoolProxiedConnection):
+    """provides the :class:`.PoolProxiedConnection` interface for cases where
+    the DBAPI connection is not actually proxied.
+
+    This is used by the engine internals to pass a consistent
+    :class:`.PoolProxiedConnection` object to consuming dialects in response to
+    pool events that may not always have the :class:`._ConnectionFairy`
+    available.
+
+    """
+
+    __slots__ = ("dbapi_connection", "_connection_record", "_is_valid")
+
+    dbapi_connection: DBAPIConnection
+    _connection_record: ConnectionPoolEntry
+
+    def __init__(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+    ):
+        self.dbapi_connection = dbapi_connection
+        self._connection_record = connection_record
+        self._is_valid = True
+
+    @property
+    def driver_connection(self) -> Any:  # type: ignore[override]  # mypy#4125
+        return self._connection_record.driver_connection
+
+    @property
+    def connection(self) -> DBAPIConnection:
+        return self.dbapi_connection
+
+    @property
+    def is_valid(self) -> bool:
+        """Implement is_valid state attribute.
+
+        for the adhoc proxied connection it's assumed the connection is valid
+        as there is no "invalidate" routine.
+
+        """
+        return self._is_valid
+
+    def invalidate(
+        self, e: Optional[BaseException] = None, soft: bool = False
+    ) -> None:
+        self._is_valid = False
+
+    @util.ro_non_memoized_property
+    def record_info(self) -> Optional[_InfoType]:
+        return self._connection_record.record_info
+
+    def cursor(self, *args: Any, **kwargs: Any) -> DBAPICursor:
+        return self.dbapi_connection.cursor(*args, **kwargs)
+
+    def __getattr__(self, key: Any) -> Any:
+        return getattr(self.dbapi_connection, key)
+
+
+class _ConnectionFairy(PoolProxiedConnection):
+    """Proxies a DBAPI connection and provides return-on-dereference
+    support.
+
+    This is an internal object used by the :class:`_pool.Pool` implementation
+    to provide context management to a DBAPI connection delivered by
+    that :class:`_pool.Pool`.   The public facing interface for this class
+    is described by the :class:`.PoolProxiedConnection` class.  See that
+    class for public API details.
+
+    The name "fairy" is inspired by the fact that the
+    :class:`._ConnectionFairy` object's lifespan is transitory, as it lasts
+    only for the length of a specific DBAPI connection being checked out from
+    the pool, and additionally that as a transparent proxy, it is mostly
+    invisible.
+
+    .. seealso::
+
+        :class:`.PoolProxiedConnection`
+
+        :class:`.ConnectionPoolEntry`
+
+
+    """
+
+    __slots__ = (
+        "dbapi_connection",
+        "_connection_record",
+        "_echo",
+        "_pool",
+        "_counter",
+        "__weakref__",
+        "__dict__",
+    )
+
+    pool: Pool
+    dbapi_connection: DBAPIConnection
+    _echo: log._EchoFlagType
+
+    def __init__(
+        self,
+        pool: Pool,
+        dbapi_connection: DBAPIConnection,
+        connection_record: _ConnectionRecord,
+        echo: log._EchoFlagType,
+    ):
+        self._pool = pool
+        self._counter = 0
+        self.dbapi_connection = dbapi_connection
+        self._connection_record = connection_record
+        self._echo = echo
+
+    _connection_record: Optional[_ConnectionRecord]
+
+    @property
+    def driver_connection(self) -> Optional[Any]:  # type: ignore[override]  # mypy#4125  # noqa: E501
+        if self._connection_record is None:
+            return None
+        return self._connection_record.driver_connection
+
+    @property
+    @util.deprecated(
+        "2.0",
+        "The _ConnectionFairy.connection attribute is deprecated; "
+        "please use 'driver_connection'",
+    )
+    def connection(self) -> DBAPIConnection:
+        return self.dbapi_connection
+
+    @classmethod
+    def _checkout(
+        cls,
+        pool: Pool,
+        threadconns: Optional[threading.local] = None,
+        fairy: Optional[_ConnectionFairy] = None,
+    ) -> _ConnectionFairy:
+        if not fairy:
+            fairy = _ConnectionRecord.checkout(pool)
+
+            if threadconns is not None:
+                threadconns.current = weakref.ref(fairy)
+
+        assert (
+            fairy._connection_record is not None
+        ), "can't 'checkout' a detached connection fairy"
+        assert (
+            fairy.dbapi_connection is not None
+        ), "can't 'checkout' an invalidated connection fairy"
+
+        fairy._counter += 1
+        if (
+            not pool.dispatch.checkout and not pool._pre_ping
+        ) or fairy._counter != 1:
+            return fairy
+
+        # Pool listeners can trigger a reconnection on checkout, as well
+        # as the pre-pinger.
+        # there are three attempts made here, but note that if the database
+        # is not accessible from a connection standpoint, those won't proceed
+        # here.
+
+        attempts = 2
+
+        while attempts > 0:
+            connection_is_fresh = fairy._connection_record.fresh
+            fairy._connection_record.fresh = False
+            try:
+                if pool._pre_ping:
+                    if not connection_is_fresh:
+                        if fairy._echo:
+                            pool.logger.debug(
+                                "Pool pre-ping on connection %s",
+                                fairy.dbapi_connection,
+                            )
+                        result = pool._dialect._do_ping_w_event(
+                            fairy.dbapi_connection
+                        )
+                        if not result:
+                            if fairy._echo:
+                                pool.logger.debug(
+                                    "Pool pre-ping on connection %s failed, "
+                                    "will invalidate pool",
+                                    fairy.dbapi_connection,
+                                )
+                            raise exc.InvalidatePoolError()
+                    elif fairy._echo:
+                        pool.logger.debug(
+                            "Connection %s is fresh, skipping pre-ping",
+                            fairy.dbapi_connection,
+                        )
+
+                pool.dispatch.checkout(
+                    fairy.dbapi_connection, fairy._connection_record, fairy
+                )
+                return fairy
+            except exc.DisconnectionError as e:
+                if e.invalidate_pool:
+                    pool.logger.info(
+                        "Disconnection detected on checkout, "
+                        "invalidating all pooled connections prior to "
+                        "current timestamp (reason: %r)",
+                        e,
+                    )
+                    fairy._connection_record.invalidate(e)
+                    pool._invalidate(fairy, e, _checkin=False)
+                else:
+                    pool.logger.info(
+                        "Disconnection detected on checkout, "
+                        "invalidating individual connection %s (reason: %r)",
+                        fairy.dbapi_connection,
+                        e,
+                    )
+                    fairy._connection_record.invalidate(e)
+                try:
+                    fairy.dbapi_connection = (
+                        fairy._connection_record.get_connection()
+                    )
+                except BaseException as err:
+                    with util.safe_reraise():
+                        fairy._connection_record._checkin_failed(
+                            err,
+                            _fairy_was_created=True,
+                        )
+
+                        # prevent _ConnectionFairy from being carried
+                        # in the stack trace.  Do this after the
+                        # connection record has been checked in, so that
+                        # if the del triggers a finalize fairy, it won't
+                        # try to checkin a second time.
+                        del fairy
+
+                    # never called, this is for code linters
+                    raise
+
+                attempts -= 1
+            except BaseException as be_outer:
+                with util.safe_reraise():
+                    rec = fairy._connection_record
+                    if rec is not None:
+                        rec._checkin_failed(
+                            be_outer,
+                            _fairy_was_created=True,
+                        )
+
+                    # prevent _ConnectionFairy from being carried
+                    # in the stack trace, see above
+                    del fairy
+
+                # never called, this is for code linters
+                raise
+
+        pool.logger.info("Reconnection attempts exhausted on checkout")
+        fairy.invalidate()
+        raise exc.InvalidRequestError("This connection is closed")
+
+    def _checkout_existing(self) -> _ConnectionFairy:
+        return _ConnectionFairy._checkout(self._pool, fairy=self)
+
+    def _checkin(self, transaction_was_reset: bool = False) -> None:
+        _finalize_fairy(
+            self.dbapi_connection,
+            self._connection_record,
+            self._pool,
+            None,
+            self._echo,
+            transaction_was_reset=transaction_was_reset,
+            fairy=self,
+        )
+
+    def _close(self) -> None:
+        self._checkin()
+
+    def _reset(
+        self,
+        pool: Pool,
+        transaction_was_reset: bool,
+        terminate_only: bool,
+        asyncio_safe: bool,
+    ) -> None:
+        if pool.dispatch.reset:
+            pool.dispatch.reset(
+                self.dbapi_connection,
+                self._connection_record,
+                PoolResetState(
+                    transaction_was_reset=transaction_was_reset,
+                    terminate_only=terminate_only,
+                    asyncio_safe=asyncio_safe,
+                ),
+            )
+
+        if not asyncio_safe:
+            return
+
+        if pool._reset_on_return is reset_rollback:
+            if transaction_was_reset:
+                if self._echo:
+                    pool.logger.debug(
+                        "Connection %s reset, transaction already reset",
+                        self.dbapi_connection,
+                    )
+            else:
+                if self._echo:
+                    pool.logger.debug(
+                        "Connection %s rollback-on-return",
+                        self.dbapi_connection,
+                    )
+                pool._dialect.do_rollback(self)
+        elif pool._reset_on_return is reset_commit:
+            if self._echo:
+                pool.logger.debug(
+                    "Connection %s commit-on-return",
+                    self.dbapi_connection,
+                )
+            pool._dialect.do_commit(self)
+
+    @property
+    def _logger(self) -> log._IdentifiedLoggerType:
+        return self._pool.logger
+
+    @property
+    def is_valid(self) -> bool:
+        return self.dbapi_connection is not None
+
+    @property
+    def is_detached(self) -> bool:
+        return self._connection_record is None
+
+    @util.ro_memoized_property
+    def info(self) -> _InfoType:
+        if self._connection_record is None:
+            return {}
+        else:
+            return self._connection_record.info
+
+    @util.ro_non_memoized_property
+    def record_info(self) -> Optional[_InfoType]:
+        if self._connection_record is None:
+            return None
+        else:
+            return self._connection_record.record_info
+
+    def invalidate(
+        self, e: Optional[BaseException] = None, soft: bool = False
+    ) -> None:
+        if self.dbapi_connection is None:
+            util.warn("Can't invalidate an already-closed connection.")
+            return
+        if self._connection_record:
+            self._connection_record.invalidate(e=e, soft=soft)
+        if not soft:
+            # prevent any rollback / reset actions etc. on
+            # the connection
+            self.dbapi_connection = None  # type: ignore
+
+            # finalize
+            self._checkin()
+
+    def cursor(self, *args: Any, **kwargs: Any) -> DBAPICursor:
+        assert self.dbapi_connection is not None
+        return self.dbapi_connection.cursor(*args, **kwargs)
+
+    def __getattr__(self, key: str) -> Any:
+        return getattr(self.dbapi_connection, key)
+
+    def detach(self) -> None:
+        if self._connection_record is not None:
+            rec = self._connection_record
+            rec.fairy_ref = None
+            rec.dbapi_connection = None
+            # TODO: should this be _return_conn?
+            self._pool._do_return_conn(self._connection_record)
+
+            # can't get the descriptor assignment to work here
+            # in pylance.  mypy is OK w/ it
+            self.info = self.info.copy()  # type: ignore
+
+            self._connection_record = None
+
+            if self._pool.dispatch.detach:
+                self._pool.dispatch.detach(self.dbapi_connection, rec)
+
+    def close(self) -> None:
+        self._counter -= 1
+        if self._counter == 0:
+            self._checkin()
+
+    def _close_special(self, transaction_reset: bool = False) -> None:
+        self._counter -= 1
+        if self._counter == 0:
+            self._checkin(transaction_was_reset=transaction_reset)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/events.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/events.py
new file mode 100644
index 00000000..4ceb260f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/events.py
@@ -0,0 +1,372 @@
+# pool/events.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Optional
+from typing import Type
+from typing import Union
+
+from .base import ConnectionPoolEntry
+from .base import Pool
+from .base import PoolProxiedConnection
+from .base import PoolResetState
+from .. import event
+from .. import util
+
+if typing.TYPE_CHECKING:
+    from ..engine import Engine
+    from ..engine.interfaces import DBAPIConnection
+
+
+class PoolEvents(event.Events[Pool]):
+    """Available events for :class:`_pool.Pool`.
+
+    The methods here define the name of an event as well
+    as the names of members that are passed to listener
+    functions.
+
+    e.g.::
+
+        from sqlalchemy import event
+
+
+        def my_on_checkout(dbapi_conn, connection_rec, connection_proxy):
+            "handle an on checkout event"
+
+
+        event.listen(Pool, "checkout", my_on_checkout)
+
+    In addition to accepting the :class:`_pool.Pool` class and
+    :class:`_pool.Pool` instances, :class:`_events.PoolEvents` also accepts
+    :class:`_engine.Engine` objects and the :class:`_engine.Engine` class as
+    targets, which will be resolved to the ``.pool`` attribute of the
+    given engine or the :class:`_pool.Pool` class::
+
+        engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test")
+
+        # will associate with engine.pool
+        event.listen(engine, "checkout", my_on_checkout)
+
+    """  # noqa: E501
+
+    _target_class_doc = "SomeEngineOrPool"
+    _dispatch_target = Pool
+
+    @util.preload_module("sqlalchemy.engine")
+    @classmethod
+    def _accept_with(
+        cls,
+        target: Union[Pool, Type[Pool], Engine, Type[Engine]],
+        identifier: str,
+    ) -> Optional[Union[Pool, Type[Pool]]]:
+        if not typing.TYPE_CHECKING:
+            Engine = util.preloaded.engine.Engine
+
+        if isinstance(target, type):
+            if issubclass(target, Engine):
+                return Pool
+            else:
+                assert issubclass(target, Pool)
+                return target
+        elif isinstance(target, Engine):
+            return target.pool
+        elif isinstance(target, Pool):
+            return target
+        elif hasattr(target, "_no_async_engine_events"):
+            target._no_async_engine_events()
+        else:
+            return None
+
+    @classmethod
+    def _listen(
+        cls,
+        event_key: event._EventKey[Pool],
+        **kw: Any,
+    ) -> None:
+        target = event_key.dispatch_target
+
+        kw.setdefault("asyncio", target._is_asyncio)
+
+        event_key.base_listen(**kw)
+
+    def connect(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+    ) -> None:
+        """Called at the moment a particular DBAPI connection is first
+        created for a given :class:`_pool.Pool`.
+
+        This event allows one to capture the point directly after which
+        the DBAPI module-level ``.connect()`` method has been used in order
+        to produce a new DBAPI connection.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        """
+
+    def first_connect(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+    ) -> None:
+        """Called exactly once for the first time a DBAPI connection is
+        checked out from a particular :class:`_pool.Pool`.
+
+        The rationale for :meth:`_events.PoolEvents.first_connect`
+        is to determine
+        information about a particular series of database connections based
+        on the settings used for all connections.  Since a particular
+        :class:`_pool.Pool`
+        refers to a single "creator" function (which in terms
+        of a :class:`_engine.Engine`
+        refers to the URL and connection options used),
+        it is typically valid to make observations about a single connection
+        that can be safely assumed to be valid about all subsequent
+        connections, such as the database version, the server and client
+        encoding settings, collation settings, and many others.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        """
+
+    def checkout(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+        connection_proxy: PoolProxiedConnection,
+    ) -> None:
+        """Called when a connection is retrieved from the Pool.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        :param connection_proxy: the :class:`.PoolProxiedConnection` object
+          which will proxy the public interface of the DBAPI connection for the
+          lifespan of the checkout.
+
+        If you raise a :class:`~sqlalchemy.exc.DisconnectionError`, the current
+        connection will be disposed and a fresh connection retrieved.
+        Processing of all checkout listeners will abort and restart
+        using the new connection.
+
+        .. seealso:: :meth:`_events.ConnectionEvents.engine_connect`
+           - a similar event
+           which occurs upon creation of a new :class:`_engine.Connection`.
+
+        """
+
+    def checkin(
+        self,
+        dbapi_connection: Optional[DBAPIConnection],
+        connection_record: ConnectionPoolEntry,
+    ) -> None:
+        """Called when a connection returns to the pool.
+
+        Note that the connection may be closed, and may be None if the
+        connection has been invalidated.  ``checkin`` will not be called
+        for detached connections.  (They do not return to the pool.)
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        """
+
+    @event._legacy_signature(
+        "2.0",
+        ["dbapi_connection", "connection_record"],
+        lambda dbapi_connection, connection_record, reset_state: (
+            dbapi_connection,
+            connection_record,
+        ),
+    )
+    def reset(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+        reset_state: PoolResetState,
+    ) -> None:
+        """Called before the "reset" action occurs for a pooled connection.
+
+        This event represents
+        when the ``rollback()`` method is called on the DBAPI connection
+        before it is returned to the pool or discarded.
+        A custom "reset" strategy may be implemented using this event hook,
+        which may also be combined with disabling the default "reset"
+        behavior using the :paramref:`_pool.Pool.reset_on_return` parameter.
+
+        The primary difference between the :meth:`_events.PoolEvents.reset` and
+        :meth:`_events.PoolEvents.checkin` events are that
+        :meth:`_events.PoolEvents.reset` is called not just for pooled
+        connections that are being returned to the pool, but also for
+        connections that were detached using the
+        :meth:`_engine.Connection.detach` method as well as asyncio connections
+        that are being discarded due to garbage collection taking place on
+        connections before the connection was checked in.
+
+        Note that the event **is not** invoked for connections that were
+        invalidated using :meth:`_engine.Connection.invalidate`.    These
+        events may be intercepted using the :meth:`.PoolEvents.soft_invalidate`
+        and :meth:`.PoolEvents.invalidate` event hooks, and all "connection
+        close" events may be intercepted using :meth:`.PoolEvents.close`.
+
+        The :meth:`_events.PoolEvents.reset` event is usually followed by the
+        :meth:`_events.PoolEvents.checkin` event, except in those
+        cases where the connection is discarded immediately after reset.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        :param reset_state: :class:`.PoolResetState` instance which provides
+         information about the circumstances under which the connection
+         is being reset.
+
+         .. versionadded:: 2.0
+
+        .. seealso::
+
+            :ref:`pool_reset_on_return`
+
+            :meth:`_events.ConnectionEvents.rollback`
+
+            :meth:`_events.ConnectionEvents.commit`
+
+        """
+
+    def invalidate(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+        exception: Optional[BaseException],
+    ) -> None:
+        """Called when a DBAPI connection is to be "invalidated".
+
+        This event is called any time the
+        :meth:`.ConnectionPoolEntry.invalidate` method is invoked, either from
+        API usage or via "auto-invalidation", without the ``soft`` flag.
+
+        The event occurs before a final attempt to call ``.close()`` on the
+        connection occurs.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        :param exception: the exception object corresponding to the reason
+         for this invalidation, if any.  May be ``None``.
+
+        .. seealso::
+
+            :ref:`pool_connection_invalidation`
+
+        """
+
+    def soft_invalidate(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+        exception: Optional[BaseException],
+    ) -> None:
+        """Called when a DBAPI connection is to be "soft invalidated".
+
+        This event is called any time the
+        :meth:`.ConnectionPoolEntry.invalidate`
+        method is invoked with the ``soft`` flag.
+
+        Soft invalidation refers to when the connection record that tracks
+        this connection will force a reconnect after the current connection
+        is checked in.   It does not actively close the dbapi_connection
+        at the point at which it is called.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        :param exception: the exception object corresponding to the reason
+         for this invalidation, if any.  May be ``None``.
+
+        """
+
+    def close(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+    ) -> None:
+        """Called when a DBAPI connection is closed.
+
+        The event is emitted before the close occurs.
+
+        The close of a connection can fail; typically this is because
+        the connection is already closed.  If the close operation fails,
+        the connection is discarded.
+
+        The :meth:`.close` event corresponds to a connection that's still
+        associated with the pool. To intercept close events for detached
+        connections use :meth:`.close_detached`.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        """
+
+    def detach(
+        self,
+        dbapi_connection: DBAPIConnection,
+        connection_record: ConnectionPoolEntry,
+    ) -> None:
+        """Called when a DBAPI connection is "detached" from a pool.
+
+        This event is emitted after the detach occurs.  The connection
+        is no longer associated with the given connection record.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        :param connection_record: the :class:`.ConnectionPoolEntry` managing
+         the DBAPI connection.
+
+        """
+
+    def close_detached(self, dbapi_connection: DBAPIConnection) -> None:
+        """Called when a detached DBAPI connection is closed.
+
+        The event is emitted before the close occurs.
+
+        The close of a connection can fail; typically this is because
+        the connection is already closed.  If the close operation fails,
+        the connection is discarded.
+
+        :param dbapi_connection: a DBAPI connection.
+         The :attr:`.ConnectionPoolEntry.dbapi_connection` attribute.
+
+        """
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/impl.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/impl.py
new file mode 100644
index 00000000..f2b951d8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/pool/impl.py
@@ -0,0 +1,581 @@
+# pool/impl.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+"""Pool implementation classes.
+
+"""
+from __future__ import annotations
+
+import threading
+import traceback
+import typing
+from typing import Any
+from typing import cast
+from typing import List
+from typing import Optional
+from typing import Set
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+import weakref
+
+from .base import _AsyncConnDialect
+from .base import _ConnectionFairy
+from .base import _ConnectionRecord
+from .base import _CreatorFnType
+from .base import _CreatorWRecFnType
+from .base import ConnectionPoolEntry
+from .base import Pool
+from .base import PoolProxiedConnection
+from .. import exc
+from .. import util
+from ..util import chop_traceback
+from ..util import queue as sqla_queue
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from ..engine.interfaces import DBAPIConnection
+
+
+class QueuePool(Pool):
+    """A :class:`_pool.Pool`
+    that imposes a limit on the number of open connections.
+
+    :class:`.QueuePool` is the default pooling implementation used for
+    all :class:`_engine.Engine` objects other than SQLite with a ``:memory:``
+    database.
+
+    The :class:`.QueuePool` class **is not compatible** with asyncio and
+    :func:`_asyncio.create_async_engine`.  The
+    :class:`.AsyncAdaptedQueuePool` class is used automatically when
+    using :func:`_asyncio.create_async_engine`, if no other kind of pool
+    is specified.
+
+    .. seealso::
+
+        :class:`.AsyncAdaptedQueuePool`
+
+    """
+
+    _is_asyncio = False  # type: ignore[assignment]
+
+    _queue_class: Type[sqla_queue.QueueCommon[ConnectionPoolEntry]] = (
+        sqla_queue.Queue
+    )
+
+    _pool: sqla_queue.QueueCommon[ConnectionPoolEntry]
+
+    def __init__(
+        self,
+        creator: Union[_CreatorFnType, _CreatorWRecFnType],
+        pool_size: int = 5,
+        max_overflow: int = 10,
+        timeout: float = 30.0,
+        use_lifo: bool = False,
+        **kw: Any,
+    ):
+        r"""
+        Construct a QueuePool.
+
+        :param creator: a callable function that returns a DB-API
+          connection object, same as that of :paramref:`_pool.Pool.creator`.
+
+        :param pool_size: The size of the pool to be maintained,
+          defaults to 5. This is the largest number of connections that
+          will be kept persistently in the pool. Note that the pool
+          begins with no connections; once this number of connections
+          is requested, that number of connections will remain.
+          ``pool_size`` can be set to 0 to indicate no size limit; to
+          disable pooling, use a :class:`~sqlalchemy.pool.NullPool`
+          instead.
+
+        :param max_overflow: The maximum overflow size of the
+          pool. When the number of checked-out connections reaches the
+          size set in pool_size, additional connections will be
+          returned up to this limit. When those additional connections
+          are returned to the pool, they are disconnected and
+          discarded. It follows then that the total number of
+          simultaneous connections the pool will allow is pool_size +
+          `max_overflow`, and the total number of "sleeping"
+          connections the pool will allow is pool_size. `max_overflow`
+          can be set to -1 to indicate no overflow limit; no limit
+          will be placed on the total number of concurrent
+          connections. Defaults to 10.
+
+        :param timeout: The number of seconds to wait before giving up
+          on returning a connection. Defaults to 30.0. This can be a float
+          but is subject to the limitations of Python time functions which
+          may not be reliable in the tens of milliseconds.
+
+        :param use_lifo: use LIFO (last-in-first-out) when retrieving
+          connections instead of FIFO (first-in-first-out). Using LIFO, a
+          server-side timeout scheme can reduce the number of connections used
+          during non-peak periods of use.   When planning for server-side
+          timeouts, ensure that a recycle or pre-ping strategy is in use to
+          gracefully handle stale connections.
+
+          .. versionadded:: 1.3
+
+          .. seealso::
+
+            :ref:`pool_use_lifo`
+
+            :ref:`pool_disconnects`
+
+        :param \**kw: Other keyword arguments including
+          :paramref:`_pool.Pool.recycle`, :paramref:`_pool.Pool.echo`,
+          :paramref:`_pool.Pool.reset_on_return` and others are passed to the
+          :class:`_pool.Pool` constructor.
+
+        """
+
+        Pool.__init__(self, creator, **kw)
+        self._pool = self._queue_class(pool_size, use_lifo=use_lifo)
+        self._overflow = 0 - pool_size
+        self._max_overflow = -1 if pool_size == 0 else max_overflow
+        self._timeout = timeout
+        self._overflow_lock = threading.Lock()
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        try:
+            self._pool.put(record, False)
+        except sqla_queue.Full:
+            try:
+                record.close()
+            finally:
+                self._dec_overflow()
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        use_overflow = self._max_overflow > -1
+
+        wait = use_overflow and self._overflow >= self._max_overflow
+        try:
+            return self._pool.get(wait, self._timeout)
+        except sqla_queue.Empty:
+            # don't do things inside of "except Empty", because when we say
+            # we timed out or can't connect and raise, Python 3 tells
+            # people the real error is queue.Empty which it isn't.
+            pass
+        if use_overflow and self._overflow >= self._max_overflow:
+            if not wait:
+                return self._do_get()
+            else:
+                raise exc.TimeoutError(
+                    "QueuePool limit of size %d overflow %d reached, "
+                    "connection timed out, timeout %0.2f"
+                    % (self.size(), self.overflow(), self._timeout),
+                    code="3o7r",
+                )
+
+        if self._inc_overflow():
+            try:
+                return self._create_connection()
+            except:
+                with util.safe_reraise():
+                    self._dec_overflow()
+                raise
+        else:
+            return self._do_get()
+
+    def _inc_overflow(self) -> bool:
+        if self._max_overflow == -1:
+            self._overflow += 1
+            return True
+        with self._overflow_lock:
+            if self._overflow < self._max_overflow:
+                self._overflow += 1
+                return True
+            else:
+                return False
+
+    def _dec_overflow(self) -> Literal[True]:
+        if self._max_overflow == -1:
+            self._overflow -= 1
+            return True
+        with self._overflow_lock:
+            self._overflow -= 1
+            return True
+
+    def recreate(self) -> QueuePool:
+        self.logger.info("Pool recreating")
+        return self.__class__(
+            self._creator,
+            pool_size=self._pool.maxsize,
+            max_overflow=self._max_overflow,
+            pre_ping=self._pre_ping,
+            use_lifo=self._pool.use_lifo,
+            timeout=self._timeout,
+            recycle=self._recycle,
+            echo=self.echo,
+            logging_name=self._orig_logging_name,
+            reset_on_return=self._reset_on_return,
+            _dispatch=self.dispatch,
+            dialect=self._dialect,
+        )
+
+    def dispose(self) -> None:
+        while True:
+            try:
+                conn = self._pool.get(False)
+                conn.close()
+            except sqla_queue.Empty:
+                break
+
+        self._overflow = 0 - self.size()
+        self.logger.info("Pool disposed. %s", self.status())
+
+    def status(self) -> str:
+        return (
+            "Pool size: %d  Connections in pool: %d "
+            "Current Overflow: %d Current Checked out "
+            "connections: %d"
+            % (
+                self.size(),
+                self.checkedin(),
+                self.overflow(),
+                self.checkedout(),
+            )
+        )
+
+    def size(self) -> int:
+        return self._pool.maxsize
+
+    def timeout(self) -> float:
+        return self._timeout
+
+    def checkedin(self) -> int:
+        return self._pool.qsize()
+
+    def overflow(self) -> int:
+        return self._overflow if self._pool.maxsize else 0
+
+    def checkedout(self) -> int:
+        return self._pool.maxsize - self._pool.qsize() + self._overflow
+
+
+class AsyncAdaptedQueuePool(QueuePool):
+    """An asyncio-compatible version of :class:`.QueuePool`.
+
+    This pool is used by default when using :class:`.AsyncEngine` engines that
+    were generated from :func:`_asyncio.create_async_engine`.   It uses an
+    asyncio-compatible queue implementation that does not use
+    ``threading.Lock``.
+
+    The arguments and operation of :class:`.AsyncAdaptedQueuePool` are
+    otherwise identical to that of :class:`.QueuePool`.
+
+    """
+
+    _is_asyncio = True  # type: ignore[assignment]
+    _queue_class: Type[sqla_queue.QueueCommon[ConnectionPoolEntry]] = (
+        sqla_queue.AsyncAdaptedQueue
+    )
+
+    _dialect = _AsyncConnDialect()
+
+
+class FallbackAsyncAdaptedQueuePool(AsyncAdaptedQueuePool):
+    _queue_class = sqla_queue.FallbackAsyncAdaptedQueue
+
+
+class NullPool(Pool):
+    """A Pool which does not pool connections.
+
+    Instead it literally opens and closes the underlying DB-API connection
+    per each connection open/close.
+
+    Reconnect-related functions such as ``recycle`` and connection
+    invalidation are not supported by this Pool implementation, since
+    no connections are held persistently.
+
+    The :class:`.NullPool` class **is compatible** with asyncio and
+    :func:`_asyncio.create_async_engine`.
+
+    """
+
+    def status(self) -> str:
+        return "NullPool"
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        record.close()
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        return self._create_connection()
+
+    def recreate(self) -> NullPool:
+        self.logger.info("Pool recreating")
+
+        return self.__class__(
+            self._creator,
+            recycle=self._recycle,
+            echo=self.echo,
+            logging_name=self._orig_logging_name,
+            reset_on_return=self._reset_on_return,
+            pre_ping=self._pre_ping,
+            _dispatch=self.dispatch,
+            dialect=self._dialect,
+        )
+
+    def dispose(self) -> None:
+        pass
+
+
+class SingletonThreadPool(Pool):
+    """A Pool that maintains one connection per thread.
+
+    Maintains one connection per each thread, never moving a connection to a
+    thread other than the one which it was created in.
+
+    .. warning::  the :class:`.SingletonThreadPool` will call ``.close()``
+       on arbitrary connections that exist beyond the size setting of
+       ``pool_size``, e.g. if more unique **thread identities**
+       than what ``pool_size`` states are used.   This cleanup is
+       non-deterministic and not sensitive to whether or not the connections
+       linked to those thread identities are currently in use.
+
+       :class:`.SingletonThreadPool` may be improved in a future release,
+       however in its current status it is generally used only for test
+       scenarios using a SQLite ``:memory:`` database and is not recommended
+       for production use.
+
+    The :class:`.SingletonThreadPool` class **is not compatible** with asyncio
+    and :func:`_asyncio.create_async_engine`.
+
+
+    Options are the same as those of :class:`_pool.Pool`, as well as:
+
+    :param pool_size: The number of threads in which to maintain connections
+        at once.  Defaults to five.
+
+    :class:`.SingletonThreadPool` is used by the SQLite dialect
+    automatically when a memory-based database is used.
+    See :ref:`sqlite_toplevel`.
+
+    """
+
+    _is_asyncio = False  # type: ignore[assignment]
+
+    def __init__(
+        self,
+        creator: Union[_CreatorFnType, _CreatorWRecFnType],
+        pool_size: int = 5,
+        **kw: Any,
+    ):
+        Pool.__init__(self, creator, **kw)
+        self._conn = threading.local()
+        self._fairy = threading.local()
+        self._all_conns: Set[ConnectionPoolEntry] = set()
+        self.size = pool_size
+
+    def recreate(self) -> SingletonThreadPool:
+        self.logger.info("Pool recreating")
+        return self.__class__(
+            self._creator,
+            pool_size=self.size,
+            recycle=self._recycle,
+            echo=self.echo,
+            pre_ping=self._pre_ping,
+            logging_name=self._orig_logging_name,
+            reset_on_return=self._reset_on_return,
+            _dispatch=self.dispatch,
+            dialect=self._dialect,
+        )
+
+    def dispose(self) -> None:
+        """Dispose of this pool."""
+
+        for conn in self._all_conns:
+            try:
+                conn.close()
+            except Exception:
+                # pysqlite won't even let you close a conn from a thread
+                # that didn't create it
+                pass
+
+        self._all_conns.clear()
+
+    def _cleanup(self) -> None:
+        while len(self._all_conns) >= self.size:
+            c = self._all_conns.pop()
+            c.close()
+
+    def status(self) -> str:
+        return "SingletonThreadPool id:%d size: %d" % (
+            id(self),
+            len(self._all_conns),
+        )
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        try:
+            del self._fairy.current
+        except AttributeError:
+            pass
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        try:
+            if TYPE_CHECKING:
+                c = cast(ConnectionPoolEntry, self._conn.current())
+            else:
+                c = self._conn.current()
+            if c:
+                return c
+        except AttributeError:
+            pass
+        c = self._create_connection()
+        self._conn.current = weakref.ref(c)
+        if len(self._all_conns) >= self.size:
+            self._cleanup()
+        self._all_conns.add(c)
+        return c
+
+    def connect(self) -> PoolProxiedConnection:
+        # vendored from Pool to include the now removed use_threadlocal
+        # behavior
+        try:
+            rec = cast(_ConnectionFairy, self._fairy.current())
+        except AttributeError:
+            pass
+        else:
+            if rec is not None:
+                return rec._checkout_existing()
+
+        return _ConnectionFairy._checkout(self, self._fairy)
+
+
+class StaticPool(Pool):
+    """A Pool of exactly one connection, used for all requests.
+
+    Reconnect-related functions such as ``recycle`` and connection
+    invalidation (which is also used to support auto-reconnect) are only
+    partially supported right now and may not yield good results.
+
+    The :class:`.StaticPool` class **is compatible** with asyncio and
+    :func:`_asyncio.create_async_engine`.
+
+    """
+
+    @util.memoized_property
+    def connection(self) -> _ConnectionRecord:
+        return _ConnectionRecord(self)
+
+    def status(self) -> str:
+        return "StaticPool"
+
+    def dispose(self) -> None:
+        if (
+            "connection" in self.__dict__
+            and self.connection.dbapi_connection is not None
+        ):
+            self.connection.close()
+            del self.__dict__["connection"]
+
+    def recreate(self) -> StaticPool:
+        self.logger.info("Pool recreating")
+        return self.__class__(
+            creator=self._creator,
+            recycle=self._recycle,
+            reset_on_return=self._reset_on_return,
+            pre_ping=self._pre_ping,
+            echo=self.echo,
+            logging_name=self._orig_logging_name,
+            _dispatch=self.dispatch,
+            dialect=self._dialect,
+        )
+
+    def _transfer_from(self, other_static_pool: StaticPool) -> None:
+        # used by the test suite to make a new engine / pool without
+        # losing the state of an existing SQLite :memory: connection
+        def creator(rec: ConnectionPoolEntry) -> DBAPIConnection:
+            conn = other_static_pool.connection.dbapi_connection
+            assert conn is not None
+            return conn
+
+        self._invoke_creator = creator
+
+    def _create_connection(self) -> ConnectionPoolEntry:
+        raise NotImplementedError()
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        pass
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        rec = self.connection
+        if rec._is_hard_or_soft_invalidated():
+            del self.__dict__["connection"]
+            rec = self.connection
+
+        return rec
+
+
+class AssertionPool(Pool):
+    """A :class:`_pool.Pool` that allows at most one checked out connection at
+    any given time.
+
+    This will raise an exception if more than one connection is checked out
+    at a time.  Useful for debugging code that is using more connections
+    than desired.
+
+    The :class:`.AssertionPool` class **is compatible** with asyncio and
+    :func:`_asyncio.create_async_engine`.
+
+    """
+
+    _conn: Optional[ConnectionPoolEntry]
+    _checkout_traceback: Optional[List[str]]
+
+    def __init__(self, *args: Any, **kw: Any):
+        self._conn = None
+        self._checked_out = False
+        self._store_traceback = kw.pop("store_traceback", True)
+        self._checkout_traceback = None
+        Pool.__init__(self, *args, **kw)
+
+    def status(self) -> str:
+        return "AssertionPool"
+
+    def _do_return_conn(self, record: ConnectionPoolEntry) -> None:
+        if not self._checked_out:
+            raise AssertionError("connection is not checked out")
+        self._checked_out = False
+        assert record is self._conn
+
+    def dispose(self) -> None:
+        self._checked_out = False
+        if self._conn:
+            self._conn.close()
+
+    def recreate(self) -> AssertionPool:
+        self.logger.info("Pool recreating")
+        return self.__class__(
+            self._creator,
+            echo=self.echo,
+            pre_ping=self._pre_ping,
+            recycle=self._recycle,
+            reset_on_return=self._reset_on_return,
+            logging_name=self._orig_logging_name,
+            _dispatch=self.dispatch,
+            dialect=self._dialect,
+        )
+
+    def _do_get(self) -> ConnectionPoolEntry:
+        if self._checked_out:
+            if self._checkout_traceback:
+                suffix = " at:\n%s" % "".join(
+                    chop_traceback(self._checkout_traceback)
+                )
+            else:
+                suffix = ""
+            raise AssertionError("connection is already checked out" + suffix)
+
+        if not self._conn:
+            self._conn = self._create_connection()
+
+        self._checked_out = True
+        if self._store_traceback:
+            self._checkout_traceback = traceback.format_stack()
+        return self._conn
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/py.typed b/backend/venv/lib/python3.12/site-packages/sqlalchemy/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/schema.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/schema.py
new file mode 100644
index 00000000..32adc9bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/schema.py
@@ -0,0 +1,70 @@
+# schema.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Compatibility namespace for sqlalchemy.sql.schema and related.
+
+"""
+
+from __future__ import annotations
+
+from .sql.base import SchemaVisitor as SchemaVisitor
+from .sql.ddl import _CreateDropBase as _CreateDropBase
+from .sql.ddl import _DropView as _DropView
+from .sql.ddl import AddConstraint as AddConstraint
+from .sql.ddl import BaseDDLElement as BaseDDLElement
+from .sql.ddl import CreateColumn as CreateColumn
+from .sql.ddl import CreateIndex as CreateIndex
+from .sql.ddl import CreateSchema as CreateSchema
+from .sql.ddl import CreateSequence as CreateSequence
+from .sql.ddl import CreateTable as CreateTable
+from .sql.ddl import DDL as DDL
+from .sql.ddl import DDLElement as DDLElement
+from .sql.ddl import DropColumnComment as DropColumnComment
+from .sql.ddl import DropConstraint as DropConstraint
+from .sql.ddl import DropConstraintComment as DropConstraintComment
+from .sql.ddl import DropIndex as DropIndex
+from .sql.ddl import DropSchema as DropSchema
+from .sql.ddl import DropSequence as DropSequence
+from .sql.ddl import DropTable as DropTable
+from .sql.ddl import DropTableComment as DropTableComment
+from .sql.ddl import ExecutableDDLElement as ExecutableDDLElement
+from .sql.ddl import InvokeDDLBase as InvokeDDLBase
+from .sql.ddl import SetColumnComment as SetColumnComment
+from .sql.ddl import SetConstraintComment as SetConstraintComment
+from .sql.ddl import SetTableComment as SetTableComment
+from .sql.ddl import sort_tables as sort_tables
+from .sql.ddl import (
+    sort_tables_and_constraints as sort_tables_and_constraints,
+)
+from .sql.naming import conv as conv
+from .sql.schema import _get_table_key as _get_table_key
+from .sql.schema import BLANK_SCHEMA as BLANK_SCHEMA
+from .sql.schema import CheckConstraint as CheckConstraint
+from .sql.schema import Column as Column
+from .sql.schema import (
+    ColumnCollectionConstraint as ColumnCollectionConstraint,
+)
+from .sql.schema import ColumnCollectionMixin as ColumnCollectionMixin
+from .sql.schema import ColumnDefault as ColumnDefault
+from .sql.schema import Computed as Computed
+from .sql.schema import Constraint as Constraint
+from .sql.schema import DefaultClause as DefaultClause
+from .sql.schema import DefaultGenerator as DefaultGenerator
+from .sql.schema import FetchedValue as FetchedValue
+from .sql.schema import ForeignKey as ForeignKey
+from .sql.schema import ForeignKeyConstraint as ForeignKeyConstraint
+from .sql.schema import HasConditionalDDL as HasConditionalDDL
+from .sql.schema import Identity as Identity
+from .sql.schema import Index as Index
+from .sql.schema import insert_sentinel as insert_sentinel
+from .sql.schema import MetaData as MetaData
+from .sql.schema import PrimaryKeyConstraint as PrimaryKeyConstraint
+from .sql.schema import SchemaConst as SchemaConst
+from .sql.schema import SchemaItem as SchemaItem
+from .sql.schema import Sequence as Sequence
+from .sql.schema import Table as Table
+from .sql.schema import UniqueConstraint as UniqueConstraint
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__init__.py
new file mode 100644
index 00000000..188f709d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__init__.py
@@ -0,0 +1,145 @@
+# sql/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from typing import Any
+from typing import TYPE_CHECKING
+
+from ._typing import ColumnExpressionArgument as ColumnExpressionArgument
+from ._typing import NotNullable as NotNullable
+from ._typing import Nullable as Nullable
+from .base import Executable as Executable
+from .compiler import COLLECT_CARTESIAN_PRODUCTS as COLLECT_CARTESIAN_PRODUCTS
+from .compiler import FROM_LINTING as FROM_LINTING
+from .compiler import NO_LINTING as NO_LINTING
+from .compiler import WARN_LINTING as WARN_LINTING
+from .ddl import BaseDDLElement as BaseDDLElement
+from .ddl import DDL as DDL
+from .ddl import DDLElement as DDLElement
+from .ddl import ExecutableDDLElement as ExecutableDDLElement
+from .expression import Alias as Alias
+from .expression import alias as alias
+from .expression import all_ as all_
+from .expression import and_ as and_
+from .expression import any_ as any_
+from .expression import asc as asc
+from .expression import between as between
+from .expression import bindparam as bindparam
+from .expression import case as case
+from .expression import cast as cast
+from .expression import ClauseElement as ClauseElement
+from .expression import collate as collate
+from .expression import column as column
+from .expression import ColumnCollection as ColumnCollection
+from .expression import ColumnElement as ColumnElement
+from .expression import CompoundSelect as CompoundSelect
+from .expression import cte as cte
+from .expression import Delete as Delete
+from .expression import delete as delete
+from .expression import desc as desc
+from .expression import distinct as distinct
+from .expression import except_ as except_
+from .expression import except_all as except_all
+from .expression import exists as exists
+from .expression import extract as extract
+from .expression import false as false
+from .expression import False_ as False_
+from .expression import FromClause as FromClause
+from .expression import func as func
+from .expression import funcfilter as funcfilter
+from .expression import Insert as Insert
+from .expression import insert as insert
+from .expression import intersect as intersect
+from .expression import intersect_all as intersect_all
+from .expression import Join as Join
+from .expression import join as join
+from .expression import label as label
+from .expression import LABEL_STYLE_DEFAULT as LABEL_STYLE_DEFAULT
+from .expression import (
+    LABEL_STYLE_DISAMBIGUATE_ONLY as LABEL_STYLE_DISAMBIGUATE_ONLY,
+)
+from .expression import LABEL_STYLE_NONE as LABEL_STYLE_NONE
+from .expression import (
+    LABEL_STYLE_TABLENAME_PLUS_COL as LABEL_STYLE_TABLENAME_PLUS_COL,
+)
+from .expression import lambda_stmt as lambda_stmt
+from .expression import LambdaElement as LambdaElement
+from .expression import lateral as lateral
+from .expression import literal as literal
+from .expression import literal_column as literal_column
+from .expression import modifier as modifier
+from .expression import not_ as not_
+from .expression import null as null
+from .expression import nulls_first as nulls_first
+from .expression import nulls_last as nulls_last
+from .expression import nullsfirst as nullsfirst
+from .expression import nullslast as nullslast
+from .expression import or_ as or_
+from .expression import outerjoin as outerjoin
+from .expression import outparam as outparam
+from .expression import over as over
+from .expression import quoted_name as quoted_name
+from .expression import Select as Select
+from .expression import select as select
+from .expression import Selectable as Selectable
+from .expression import SelectLabelStyle as SelectLabelStyle
+from .expression import SQLColumnExpression as SQLColumnExpression
+from .expression import StatementLambdaElement as StatementLambdaElement
+from .expression import Subquery as Subquery
+from .expression import table as table
+from .expression import TableClause as TableClause
+from .expression import TableSample as TableSample
+from .expression import tablesample as tablesample
+from .expression import text as text
+from .expression import true as true
+from .expression import True_ as True_
+from .expression import try_cast as try_cast
+from .expression import tuple_ as tuple_
+from .expression import type_coerce as type_coerce
+from .expression import union as union
+from .expression import union_all as union_all
+from .expression import Update as Update
+from .expression import update as update
+from .expression import Values as Values
+from .expression import values as values
+from .expression import within_group as within_group
+from .visitors import ClauseVisitor as ClauseVisitor
+
+
+def __go(lcls: Any) -> None:
+    from .. import util as _sa_util
+
+    from . import base
+    from . import coercions
+    from . import elements
+    from . import lambdas
+    from . import selectable
+    from . import schema
+    from . import traversals
+    from . import type_api
+
+    if not TYPE_CHECKING:
+        base.coercions = elements.coercions = coercions
+        base.elements = elements
+        base.type_api = type_api
+        coercions.elements = elements
+        coercions.lambdas = lambdas
+        coercions.schema = schema
+        coercions.selectable = selectable
+
+    from .annotation import _prepare_annotations
+    from .annotation import Annotated
+    from .elements import AnnotatedColumnElement
+    from .elements import ClauseList
+    from .selectable import AnnotatedFromClause
+
+    _prepare_annotations(ColumnElement, AnnotatedColumnElement)
+    _prepare_annotations(FromClause, AnnotatedFromClause)
+    _prepare_annotations(ClauseList, Annotated)
+
+    _sa_util.preloaded.import_prefix("sqlalchemy.sql")
+
+
+__go(locals())
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7690f1a1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_dml_constructors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_dml_constructors.cpython-312.pyc
new file mode 100644
index 00000000..efc6b3c1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_dml_constructors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_elements_constructors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_elements_constructors.cpython-312.pyc
new file mode 100644
index 00000000..e8058339
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_elements_constructors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_orm_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_orm_types.cpython-312.pyc
new file mode 100644
index 00000000..1a2f93d9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_orm_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_py_util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_py_util.cpython-312.pyc
new file mode 100644
index 00000000..9a785be7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_py_util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_selectable_constructors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_selectable_constructors.cpython-312.pyc
new file mode 100644
index 00000000..28a55995
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_selectable_constructors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_typing.cpython-312.pyc
new file mode 100644
index 00000000..651f0e11
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/_typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/annotation.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/annotation.cpython-312.pyc
new file mode 100644
index 00000000..d7a291e0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/annotation.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..e097120c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/cache_key.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/cache_key.cpython-312.pyc
new file mode 100644
index 00000000..15392153
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/cache_key.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/coercions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/coercions.cpython-312.pyc
new file mode 100644
index 00000000..ddbad680
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/coercions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/compiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/compiler.cpython-312.pyc
new file mode 100644
index 00000000..3c5cc959
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/compiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/crud.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/crud.cpython-312.pyc
new file mode 100644
index 00000000..7ccd9263
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/crud.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/ddl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/ddl.cpython-312.pyc
new file mode 100644
index 00000000..4a247e84
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/ddl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/default_comparator.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/default_comparator.cpython-312.pyc
new file mode 100644
index 00000000..9a96ac5a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/default_comparator.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/dml.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/dml.cpython-312.pyc
new file mode 100644
index 00000000..7c8e5639
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/dml.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/elements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/elements.cpython-312.pyc
new file mode 100644
index 00000000..bfb5e511
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/elements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/events.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/events.cpython-312.pyc
new file mode 100644
index 00000000..d85f3212
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/events.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/expression.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/expression.cpython-312.pyc
new file mode 100644
index 00000000..b6688d98
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/expression.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/functions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/functions.cpython-312.pyc
new file mode 100644
index 00000000..788667a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/functions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/lambdas.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/lambdas.cpython-312.pyc
new file mode 100644
index 00000000..1d1dff4a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/lambdas.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/naming.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/naming.cpython-312.pyc
new file mode 100644
index 00000000..5cf299ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/naming.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/operators.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/operators.cpython-312.pyc
new file mode 100644
index 00000000..fb6da730
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/operators.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/roles.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/roles.cpython-312.pyc
new file mode 100644
index 00000000..82a8bd74
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/roles.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/schema.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/schema.cpython-312.pyc
new file mode 100644
index 00000000..86bb71b4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/schema.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/selectable.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/selectable.cpython-312.pyc
new file mode 100644
index 00000000..c255d720
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/selectable.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/sqltypes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/sqltypes.cpython-312.pyc
new file mode 100644
index 00000000..656442d2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/sqltypes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/traversals.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/traversals.cpython-312.pyc
new file mode 100644
index 00000000..fe36c02e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/traversals.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/type_api.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/type_api.cpython-312.pyc
new file mode 100644
index 00000000..831444cd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/type_api.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..2e52851e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/visitors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/visitors.cpython-312.pyc
new file mode 100644
index 00000000..24230aa7
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/__pycache__/visitors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_dml_constructors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_dml_constructors.py
new file mode 100644
index 00000000..0a6f6011
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_dml_constructors.py
@@ -0,0 +1,132 @@
+# sql/_dml_constructors.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from .dml import Delete
+from .dml import Insert
+from .dml import Update
+
+if TYPE_CHECKING:
+    from ._typing import _DMLTableArgument
+
+
+def insert(table: _DMLTableArgument) -> Insert:
+    """Construct an :class:`_expression.Insert` object.
+
+    E.g.::
+
+        from sqlalchemy import insert
+
+        stmt = insert(user_table).values(name="username", fullname="Full Username")
+
+    Similar functionality is available via the
+    :meth:`_expression.TableClause.insert` method on
+    :class:`_schema.Table`.
+
+    .. seealso::
+
+        :ref:`tutorial_core_insert` - in the :ref:`unified_tutorial`
+
+
+    :param table: :class:`_expression.TableClause`
+     which is the subject of the
+     insert.
+
+    :param values: collection of values to be inserted; see
+     :meth:`_expression.Insert.values`
+     for a description of allowed formats here.
+     Can be omitted entirely; a :class:`_expression.Insert` construct
+     will also dynamically render the VALUES clause at execution time
+     based on the parameters passed to :meth:`_engine.Connection.execute`.
+
+    :param inline: if True, no attempt will be made to retrieve the
+     SQL-generated default values to be provided within the statement;
+     in particular,
+     this allows SQL expressions to be rendered 'inline' within the
+     statement without the need to pre-execute them beforehand; for
+     backends that support "returning", this turns off the "implicit
+     returning" feature for the statement.
+
+    If both :paramref:`_expression.insert.values` and compile-time bind
+    parameters are present, the compile-time bind parameters override the
+    information specified within :paramref:`_expression.insert.values` on a
+    per-key basis.
+
+    The keys within :paramref:`_expression.Insert.values` can be either
+    :class:`~sqlalchemy.schema.Column` objects or their string
+    identifiers. Each key may reference one of:
+
+    * a literal data value (i.e. string, number, etc.);
+    * a Column object;
+    * a SELECT statement.
+
+    If a ``SELECT`` statement is specified which references this
+    ``INSERT`` statement's table, the statement will be correlated
+    against the ``INSERT`` statement.
+
+    .. seealso::
+
+        :ref:`tutorial_core_insert` - in the :ref:`unified_tutorial`
+
+    """  # noqa: E501
+    return Insert(table)
+
+
+def update(table: _DMLTableArgument) -> Update:
+    r"""Construct an :class:`_expression.Update` object.
+
+    E.g.::
+
+        from sqlalchemy import update
+
+        stmt = (
+            update(user_table).where(user_table.c.id == 5).values(name="user #5")
+        )
+
+    Similar functionality is available via the
+    :meth:`_expression.TableClause.update` method on
+    :class:`_schema.Table`.
+
+    :param table: A :class:`_schema.Table`
+     object representing the database
+     table to be updated.
+
+
+    .. seealso::
+
+        :ref:`tutorial_core_update_delete` - in the :ref:`unified_tutorial`
+
+
+    """  # noqa: E501
+    return Update(table)
+
+
+def delete(table: _DMLTableArgument) -> Delete:
+    r"""Construct :class:`_expression.Delete` object.
+
+    E.g.::
+
+        from sqlalchemy import delete
+
+        stmt = delete(user_table).where(user_table.c.id == 5)
+
+    Similar functionality is available via the
+    :meth:`_expression.TableClause.delete` method on
+    :class:`_schema.Table`.
+
+    :param table: The table to delete rows from.
+
+    .. seealso::
+
+        :ref:`tutorial_core_update_delete` - in the :ref:`unified_tutorial`
+
+
+    """
+    return Delete(table)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_elements_constructors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_elements_constructors.py
new file mode 100644
index 00000000..b628fcc9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_elements_constructors.py
@@ -0,0 +1,1862 @@
+# sql/_elements_constructors.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple as typing_Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import roles
+from .base import _NoArg
+from .coercions import _document_text_coercion
+from .elements import BindParameter
+from .elements import BooleanClauseList
+from .elements import Case
+from .elements import Cast
+from .elements import CollationClause
+from .elements import CollectionAggregate
+from .elements import ColumnClause
+from .elements import ColumnElement
+from .elements import Extract
+from .elements import False_
+from .elements import FunctionFilter
+from .elements import Label
+from .elements import Null
+from .elements import Over
+from .elements import TextClause
+from .elements import True_
+from .elements import TryCast
+from .elements import Tuple
+from .elements import TypeCoerce
+from .elements import UnaryExpression
+from .elements import WithinGroup
+from .functions import FunctionElement
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from ._typing import _ByArgument
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnExpressionOrLiteralArgument
+    from ._typing import _ColumnExpressionOrStrLabelArgument
+    from ._typing import _TypeEngineArgument
+    from .elements import BinaryExpression
+    from .selectable import FromClause
+    from .type_api import TypeEngine
+
+_T = TypeVar("_T")
+
+
+def all_(expr: _ColumnExpressionArgument[_T]) -> CollectionAggregate[bool]:
+    """Produce an ALL expression.
+
+    For dialects such as that of PostgreSQL, this operator applies
+    to usage of the :class:`_types.ARRAY` datatype, for that of
+    MySQL, it may apply to a subquery.  e.g.::
+
+        # renders on PostgreSQL:
+        # '5 = ALL (somearray)'
+        expr = 5 == all_(mytable.c.somearray)
+
+        # renders on MySQL:
+        # '5 = ALL (SELECT value FROM table)'
+        expr = 5 == all_(select(table.c.value))
+
+    Comparison to NULL may work using ``None``::
+
+        None == all_(mytable.c.somearray)
+
+    The any_() / all_() operators also feature a special "operand flipping"
+    behavior such that if any_() / all_() are used on the left side of a
+    comparison using a standalone operator such as ``==``, ``!=``, etc.
+    (not including operator methods such as
+    :meth:`_sql.ColumnOperators.is_`) the rendered expression is flipped::
+
+        # would render '5 = ALL (column)`
+        all_(mytable.c.column) == 5
+
+    Or with ``None``, which note will not perform
+    the usual step of rendering "IS" as is normally the case for NULL::
+
+        # would render 'NULL = ALL(somearray)'
+        all_(mytable.c.somearray) == None
+
+    .. versionchanged:: 1.4.26  repaired the use of any_() / all_()
+       comparing to NULL on the right side to be flipped to the left.
+
+    The column-level :meth:`_sql.ColumnElement.all_` method (not to be
+    confused with :class:`_types.ARRAY` level
+    :meth:`_types.ARRAY.Comparator.all`) is shorthand for
+    ``all_(col)``::
+
+        5 == mytable.c.somearray.all_()
+
+    .. seealso::
+
+        :meth:`_sql.ColumnOperators.all_`
+
+        :func:`_expression.any_`
+
+    """
+    return CollectionAggregate._create_all(expr)
+
+
+def and_(  # type: ignore[empty-body]
+    initial_clause: Union[Literal[True], _ColumnExpressionArgument[bool]],
+    *clauses: _ColumnExpressionArgument[bool],
+) -> ColumnElement[bool]:
+    r"""Produce a conjunction of expressions joined by ``AND``.
+
+    E.g.::
+
+        from sqlalchemy import and_
+
+        stmt = select(users_table).where(
+            and_(users_table.c.name == "wendy", users_table.c.enrolled == True)
+        )
+
+    The :func:`.and_` conjunction is also available using the
+    Python ``&`` operator (though note that compound expressions
+    need to be parenthesized in order to function with Python
+    operator precedence behavior)::
+
+        stmt = select(users_table).where(
+            (users_table.c.name == "wendy") & (users_table.c.enrolled == True)
+        )
+
+    The :func:`.and_` operation is also implicit in some cases;
+    the :meth:`_expression.Select.where`
+    method for example can be invoked multiple
+    times against a statement, which will have the effect of each
+    clause being combined using :func:`.and_`::
+
+        stmt = (
+            select(users_table)
+            .where(users_table.c.name == "wendy")
+            .where(users_table.c.enrolled == True)
+        )
+
+    The :func:`.and_` construct must be given at least one positional
+    argument in order to be valid; a :func:`.and_` construct with no
+    arguments is ambiguous.   To produce an "empty" or dynamically
+    generated :func:`.and_`  expression, from a given list of expressions,
+    a "default" element of :func:`_sql.true` (or just ``True``) should be
+    specified::
+
+        from sqlalchemy import true
+
+        criteria = and_(true(), *expressions)
+
+    The above expression will compile to SQL as the expression ``true``
+    or ``1 = 1``, depending on backend, if no other expressions are
+    present.  If expressions are present, then the :func:`_sql.true` value is
+    ignored as it does not affect the outcome of an AND expression that
+    has other elements.
+
+    .. deprecated:: 1.4  The :func:`.and_` element now requires that at
+       least one argument is passed; creating the :func:`.and_` construct
+       with no arguments is deprecated, and will emit a deprecation warning
+       while continuing to produce a blank SQL string.
+
+    .. seealso::
+
+        :func:`.or_`
+
+    """
+    ...
+
+
+if not TYPE_CHECKING:
+    # handle deprecated case which allows zero-arguments
+    def and_(*clauses):  # noqa: F811
+        r"""Produce a conjunction of expressions joined by ``AND``.
+
+        E.g.::
+
+            from sqlalchemy import and_
+
+            stmt = select(users_table).where(
+                and_(users_table.c.name == "wendy", users_table.c.enrolled == True)
+            )
+
+        The :func:`.and_` conjunction is also available using the
+        Python ``&`` operator (though note that compound expressions
+        need to be parenthesized in order to function with Python
+        operator precedence behavior)::
+
+            stmt = select(users_table).where(
+                (users_table.c.name == "wendy") & (users_table.c.enrolled == True)
+            )
+
+        The :func:`.and_` operation is also implicit in some cases;
+        the :meth:`_expression.Select.where`
+        method for example can be invoked multiple
+        times against a statement, which will have the effect of each
+        clause being combined using :func:`.and_`::
+
+            stmt = (
+                select(users_table)
+                .where(users_table.c.name == "wendy")
+                .where(users_table.c.enrolled == True)
+            )
+
+        The :func:`.and_` construct must be given at least one positional
+        argument in order to be valid; a :func:`.and_` construct with no
+        arguments is ambiguous.   To produce an "empty" or dynamically
+        generated :func:`.and_`  expression, from a given list of expressions,
+        a "default" element of :func:`_sql.true` (or just ``True``) should be
+        specified::
+
+            from sqlalchemy import true
+
+            criteria = and_(true(), *expressions)
+
+        The above expression will compile to SQL as the expression ``true``
+        or ``1 = 1``, depending on backend, if no other expressions are
+        present.  If expressions are present, then the :func:`_sql.true` value
+        is ignored as it does not affect the outcome of an AND expression that
+        has other elements.
+
+        .. deprecated:: 1.4  The :func:`.and_` element now requires that at
+          least one argument is passed; creating the :func:`.and_` construct
+          with no arguments is deprecated, and will emit a deprecation warning
+          while continuing to produce a blank SQL string.
+
+        .. seealso::
+
+            :func:`.or_`
+
+        """  # noqa: E501
+        return BooleanClauseList.and_(*clauses)
+
+
+def any_(expr: _ColumnExpressionArgument[_T]) -> CollectionAggregate[bool]:
+    """Produce an ANY expression.
+
+    For dialects such as that of PostgreSQL, this operator applies
+    to usage of the :class:`_types.ARRAY` datatype, for that of
+    MySQL, it may apply to a subquery.  e.g.::
+
+        # renders on PostgreSQL:
+        # '5 = ANY (somearray)'
+        expr = 5 == any_(mytable.c.somearray)
+
+        # renders on MySQL:
+        # '5 = ANY (SELECT value FROM table)'
+        expr = 5 == any_(select(table.c.value))
+
+    Comparison to NULL may work using ``None`` or :func:`_sql.null`::
+
+        None == any_(mytable.c.somearray)
+
+    The any_() / all_() operators also feature a special "operand flipping"
+    behavior such that if any_() / all_() are used on the left side of a
+    comparison using a standalone operator such as ``==``, ``!=``, etc.
+    (not including operator methods such as
+    :meth:`_sql.ColumnOperators.is_`) the rendered expression is flipped::
+
+        # would render '5 = ANY (column)`
+        any_(mytable.c.column) == 5
+
+    Or with ``None``, which note will not perform
+    the usual step of rendering "IS" as is normally the case for NULL::
+
+        # would render 'NULL = ANY(somearray)'
+        any_(mytable.c.somearray) == None
+
+    .. versionchanged:: 1.4.26  repaired the use of any_() / all_()
+       comparing to NULL on the right side to be flipped to the left.
+
+    The column-level :meth:`_sql.ColumnElement.any_` method (not to be
+    confused with :class:`_types.ARRAY` level
+    :meth:`_types.ARRAY.Comparator.any`) is shorthand for
+    ``any_(col)``::
+
+        5 = mytable.c.somearray.any_()
+
+    .. seealso::
+
+        :meth:`_sql.ColumnOperators.any_`
+
+        :func:`_expression.all_`
+
+    """
+    return CollectionAggregate._create_any(expr)
+
+
+def asc(
+    column: _ColumnExpressionOrStrLabelArgument[_T],
+) -> UnaryExpression[_T]:
+    """Produce an ascending ``ORDER BY`` clause element.
+
+    e.g.::
+
+        from sqlalchemy import asc
+
+        stmt = select(users_table).order_by(asc(users_table.c.name))
+
+    will produce SQL as:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user ORDER BY name ASC
+
+    The :func:`.asc` function is a standalone version of the
+    :meth:`_expression.ColumnElement.asc`
+    method available on all SQL expressions,
+    e.g.::
+
+
+        stmt = select(users_table).order_by(users_table.c.name.asc())
+
+    :param column: A :class:`_expression.ColumnElement` (e.g.
+     scalar SQL expression)
+     with which to apply the :func:`.asc` operation.
+
+    .. seealso::
+
+        :func:`.desc`
+
+        :func:`.nulls_first`
+
+        :func:`.nulls_last`
+
+        :meth:`_expression.Select.order_by`
+
+    """
+    return UnaryExpression._create_asc(column)
+
+
+def collate(
+    expression: _ColumnExpressionArgument[str], collation: str
+) -> BinaryExpression[str]:
+    """Return the clause ``expression COLLATE collation``.
+
+    e.g.::
+
+        collate(mycolumn, "utf8_bin")
+
+    produces:
+
+    .. sourcecode:: sql
+
+        mycolumn COLLATE utf8_bin
+
+    The collation expression is also quoted if it is a case sensitive
+    identifier, e.g. contains uppercase characters.
+
+    .. versionchanged:: 1.2 quoting is automatically applied to COLLATE
+       expressions if they are case sensitive.
+
+    """
+    return CollationClause._create_collation_expression(expression, collation)
+
+
+def between(
+    expr: _ColumnExpressionOrLiteralArgument[_T],
+    lower_bound: Any,
+    upper_bound: Any,
+    symmetric: bool = False,
+) -> BinaryExpression[bool]:
+    """Produce a ``BETWEEN`` predicate clause.
+
+    E.g.::
+
+        from sqlalchemy import between
+
+        stmt = select(users_table).where(between(users_table.c.id, 5, 7))
+
+    Would produce SQL resembling:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user WHERE id BETWEEN :id_1 AND :id_2
+
+    The :func:`.between` function is a standalone version of the
+    :meth:`_expression.ColumnElement.between` method available on all
+    SQL expressions, as in::
+
+        stmt = select(users_table).where(users_table.c.id.between(5, 7))
+
+    All arguments passed to :func:`.between`, including the left side
+    column expression, are coerced from Python scalar values if a
+    the value is not a :class:`_expression.ColumnElement` subclass.
+    For example,
+    three fixed values can be compared as in::
+
+        print(between(5, 3, 7))
+
+    Which would produce::
+
+        :param_1 BETWEEN :param_2 AND :param_3
+
+    :param expr: a column expression, typically a
+     :class:`_expression.ColumnElement`
+     instance or alternatively a Python scalar expression to be coerced
+     into a column expression, serving as the left side of the ``BETWEEN``
+     expression.
+
+    :param lower_bound: a column or Python scalar expression serving as the
+     lower bound of the right side of the ``BETWEEN`` expression.
+
+    :param upper_bound: a column or Python scalar expression serving as the
+     upper bound of the right side of the ``BETWEEN`` expression.
+
+    :param symmetric: if True, will render " BETWEEN SYMMETRIC ". Note
+     that not all databases support this syntax.
+
+    .. seealso::
+
+        :meth:`_expression.ColumnElement.between`
+
+    """
+    col_expr = coercions.expect(roles.ExpressionElementRole, expr)
+    return col_expr.between(lower_bound, upper_bound, symmetric=symmetric)
+
+
+def outparam(
+    key: str, type_: Optional[TypeEngine[_T]] = None
+) -> BindParameter[_T]:
+    """Create an 'OUT' parameter for usage in functions (stored procedures),
+    for databases which support them.
+
+    The ``outparam`` can be used like a regular function parameter.
+    The "output" value will be available from the
+    :class:`~sqlalchemy.engine.CursorResult` object via its ``out_parameters``
+    attribute, which returns a dictionary containing the values.
+
+    """
+    return BindParameter(key, None, type_=type_, unique=False, isoutparam=True)
+
+
+@overload
+def not_(clause: BinaryExpression[_T]) -> BinaryExpression[_T]: ...
+
+
+@overload
+def not_(clause: _ColumnExpressionArgument[_T]) -> ColumnElement[_T]: ...
+
+
+def not_(clause: _ColumnExpressionArgument[_T]) -> ColumnElement[_T]:
+    """Return a negation of the given clause, i.e. ``NOT(clause)``.
+
+    The ``~`` operator is also overloaded on all
+    :class:`_expression.ColumnElement` subclasses to produce the
+    same result.
+
+    """
+
+    return coercions.expect(roles.ExpressionElementRole, clause).__invert__()
+
+
+def bindparam(
+    key: Optional[str],
+    value: Any = _NoArg.NO_ARG,
+    type_: Optional[_TypeEngineArgument[_T]] = None,
+    unique: bool = False,
+    required: Union[bool, Literal[_NoArg.NO_ARG]] = _NoArg.NO_ARG,
+    quote: Optional[bool] = None,
+    callable_: Optional[Callable[[], Any]] = None,
+    expanding: bool = False,
+    isoutparam: bool = False,
+    literal_execute: bool = False,
+) -> BindParameter[_T]:
+    r"""Produce a "bound expression".
+
+    The return value is an instance of :class:`.BindParameter`; this
+    is a :class:`_expression.ColumnElement`
+    subclass which represents a so-called
+    "placeholder" value in a SQL expression, the value of which is
+    supplied at the point at which the statement in executed against a
+    database connection.
+
+    In SQLAlchemy, the :func:`.bindparam` construct has
+    the ability to carry along the actual value that will be ultimately
+    used at expression time.  In this way, it serves not just as
+    a "placeholder" for eventual population, but also as a means of
+    representing so-called "unsafe" values which should not be rendered
+    directly in a SQL statement, but rather should be passed along
+    to the :term:`DBAPI` as values which need to be correctly escaped
+    and potentially handled for type-safety.
+
+    When using :func:`.bindparam` explicitly, the use case is typically
+    one of traditional deferment of parameters; the :func:`.bindparam`
+    construct accepts a name which can then be referred to at execution
+    time::
+
+        from sqlalchemy import bindparam
+
+        stmt = select(users_table).where(
+            users_table.c.name == bindparam("username")
+        )
+
+    The above statement, when rendered, will produce SQL similar to:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user WHERE name = :username
+
+    In order to populate the value of ``:username`` above, the value
+    would typically be applied at execution time to a method
+    like :meth:`_engine.Connection.execute`::
+
+        result = connection.execute(stmt, {"username": "wendy"})
+
+    Explicit use of :func:`.bindparam` is also common when producing
+    UPDATE or DELETE statements that are to be invoked multiple times,
+    where the WHERE criterion of the statement is to change on each
+    invocation, such as::
+
+        stmt = (
+            users_table.update()
+            .where(user_table.c.name == bindparam("username"))
+            .values(fullname=bindparam("fullname"))
+        )
+
+        connection.execute(
+            stmt,
+            [
+                {"username": "wendy", "fullname": "Wendy Smith"},
+                {"username": "jack", "fullname": "Jack Jones"},
+            ],
+        )
+
+    SQLAlchemy's Core expression system makes wide use of
+    :func:`.bindparam` in an implicit sense.   It is typical that Python
+    literal values passed to virtually all SQL expression functions are
+    coerced into fixed :func:`.bindparam` constructs.  For example, given
+    a comparison operation such as::
+
+        expr = users_table.c.name == "Wendy"
+
+    The above expression will produce a :class:`.BinaryExpression`
+    construct, where the left side is the :class:`_schema.Column` object
+    representing the ``name`` column, and the right side is a
+    :class:`.BindParameter` representing the literal value::
+
+        print(repr(expr.right))
+        BindParameter("%(4327771088 name)s", "Wendy", type_=String())
+
+    The expression above will render SQL such as:
+
+    .. sourcecode:: sql
+
+        user.name = :name_1
+
+    Where the ``:name_1`` parameter name is an anonymous name.  The
+    actual string ``Wendy`` is not in the rendered string, but is carried
+    along where it is later used within statement execution.  If we
+    invoke a statement like the following::
+
+        stmt = select(users_table).where(users_table.c.name == "Wendy")
+        result = connection.execute(stmt)
+
+    We would see SQL logging output as:
+
+    .. sourcecode:: sql
+
+        SELECT "user".id, "user".name
+        FROM "user"
+        WHERE "user".name = %(name_1)s
+        {'name_1': 'Wendy'}
+
+    Above, we see that ``Wendy`` is passed as a parameter to the database,
+    while the placeholder ``:name_1`` is rendered in the appropriate form
+    for the target database, in this case the PostgreSQL database.
+
+    Similarly, :func:`.bindparam` is invoked automatically when working
+    with :term:`CRUD` statements as far as the "VALUES" portion is
+    concerned.   The :func:`_expression.insert` construct produces an
+    ``INSERT`` expression which will, at statement execution time, generate
+    bound placeholders based on the arguments passed, as in::
+
+        stmt = users_table.insert()
+        result = connection.execute(stmt, {"name": "Wendy"})
+
+    The above will produce SQL output as:
+
+    .. sourcecode:: sql
+
+        INSERT INTO "user" (name) VALUES (%(name)s)
+        {'name': 'Wendy'}
+
+    The :class:`_expression.Insert` construct, at
+    compilation/execution time, rendered a single :func:`.bindparam`
+    mirroring the column name ``name`` as a result of the single ``name``
+    parameter we passed to the :meth:`_engine.Connection.execute` method.
+
+    :param key:
+      the key (e.g. the name) for this bind param.
+      Will be used in the generated
+      SQL statement for dialects that use named parameters.  This
+      value may be modified when part of a compilation operation,
+      if other :class:`BindParameter` objects exist with the same
+      key, or if its length is too long and truncation is
+      required.
+
+      If omitted, an "anonymous" name is generated for the bound parameter;
+      when given a value to bind, the end result is equivalent to calling upon
+      the :func:`.literal` function with a value to bind, particularly
+      if the :paramref:`.bindparam.unique` parameter is also provided.
+
+    :param value:
+      Initial value for this bind param.  Will be used at statement
+      execution time as the value for this parameter passed to the
+      DBAPI, if no other value is indicated to the statement execution
+      method for this particular parameter name.  Defaults to ``None``.
+
+    :param callable\_:
+      A callable function that takes the place of "value".  The function
+      will be called at statement execution time to determine the
+      ultimate value.   Used for scenarios where the actual bind
+      value cannot be determined at the point at which the clause
+      construct is created, but embedded bind values are still desirable.
+
+    :param type\_:
+      A :class:`.TypeEngine` class or instance representing an optional
+      datatype for this :func:`.bindparam`.  If not passed, a type
+      may be determined automatically for the bind, based on the given
+      value; for example, trivial Python types such as ``str``,
+      ``int``, ``bool``
+      may result in the :class:`.String`, :class:`.Integer` or
+      :class:`.Boolean` types being automatically selected.
+
+      The type of a :func:`.bindparam` is significant especially in that
+      the type will apply pre-processing to the value before it is
+      passed to the database.  For example, a :func:`.bindparam` which
+      refers to a datetime value, and is specified as holding the
+      :class:`.DateTime` type, may apply conversion needed to the
+      value (such as stringification on SQLite) before passing the value
+      to the database.
+
+    :param unique:
+      if True, the key name of this :class:`.BindParameter` will be
+      modified if another :class:`.BindParameter` of the same name
+      already has been located within the containing
+      expression.  This flag is used generally by the internals
+      when producing so-called "anonymous" bound expressions, it
+      isn't generally applicable to explicitly-named :func:`.bindparam`
+      constructs.
+
+    :param required:
+      If ``True``, a value is required at execution time.  If not passed,
+      it defaults to ``True`` if neither :paramref:`.bindparam.value`
+      or :paramref:`.bindparam.callable` were passed.  If either of these
+      parameters are present, then :paramref:`.bindparam.required`
+      defaults to ``False``.
+
+    :param quote:
+      True if this parameter name requires quoting and is not
+      currently known as a SQLAlchemy reserved word; this currently
+      only applies to the Oracle Database backends, where bound names must
+      sometimes be quoted.
+
+    :param isoutparam:
+      if True, the parameter should be treated like a stored procedure
+      "OUT" parameter.  This applies to backends such as Oracle Database which
+      support OUT parameters.
+
+    :param expanding:
+      if True, this parameter will be treated as an "expanding" parameter
+      at execution time; the parameter value is expected to be a sequence,
+      rather than a scalar value, and the string SQL statement will
+      be transformed on a per-execution basis to accommodate the sequence
+      with a variable number of parameter slots passed to the DBAPI.
+      This is to allow statement caching to be used in conjunction with
+      an IN clause.
+
+      .. seealso::
+
+        :meth:`.ColumnOperators.in_`
+
+        :ref:`baked_in` - with baked queries
+
+      .. note:: The "expanding" feature does not support "executemany"-
+         style parameter sets.
+
+      .. versionadded:: 1.2
+
+      .. versionchanged:: 1.3 the "expanding" bound parameter feature now
+         supports empty lists.
+
+    :param literal_execute:
+      if True, the bound parameter will be rendered in the compile phase
+      with a special "POSTCOMPILE" token, and the SQLAlchemy compiler will
+      render the final value of the parameter into the SQL statement at
+      statement execution time, omitting the value from the parameter
+      dictionary / list passed to DBAPI ``cursor.execute()``.  This
+      produces a similar effect as that of using the ``literal_binds``,
+      compilation flag,  however takes place as the statement is sent to
+      the DBAPI ``cursor.execute()`` method, rather than when the statement
+      is compiled.   The primary use of this
+      capability is for rendering LIMIT / OFFSET clauses for database
+      drivers that can't accommodate for bound parameters in these
+      contexts, while allowing SQL constructs to be cacheable at the
+      compilation level.
+
+      .. versionadded:: 1.4 Added "post compile" bound parameters
+
+        .. seealso::
+
+            :ref:`change_4808`.
+
+    .. seealso::
+
+        :ref:`tutorial_sending_parameters` - in the
+        :ref:`unified_tutorial`
+
+
+    """
+    return BindParameter(
+        key,
+        value,
+        type_,
+        unique,
+        required,
+        quote,
+        callable_,
+        expanding,
+        isoutparam,
+        literal_execute,
+    )
+
+
+def case(
+    *whens: Union[
+        typing_Tuple[_ColumnExpressionArgument[bool], Any], Mapping[Any, Any]
+    ],
+    value: Optional[Any] = None,
+    else_: Optional[Any] = None,
+) -> Case[Any]:
+    r"""Produce a ``CASE`` expression.
+
+    The ``CASE`` construct in SQL is a conditional object that
+    acts somewhat analogously to an "if/then" construct in other
+    languages.  It returns an instance of :class:`.Case`.
+
+    :func:`.case` in its usual form is passed a series of "when"
+    constructs, that is, a list of conditions and results as tuples::
+
+        from sqlalchemy import case
+
+        stmt = select(users_table).where(
+            case(
+                (users_table.c.name == "wendy", "W"),
+                (users_table.c.name == "jack", "J"),
+                else_="E",
+            )
+        )
+
+    The above statement will produce SQL resembling:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user
+        WHERE CASE
+            WHEN (name = :name_1) THEN :param_1
+            WHEN (name = :name_2) THEN :param_2
+            ELSE :param_3
+        END
+
+    When simple equality expressions of several values against a single
+    parent column are needed, :func:`.case` also has a "shorthand" format
+    used via the
+    :paramref:`.case.value` parameter, which is passed a column
+    expression to be compared.  In this form, the :paramref:`.case.whens`
+    parameter is passed as a dictionary containing expressions to be
+    compared against keyed to result expressions.  The statement below is
+    equivalent to the preceding statement::
+
+        stmt = select(users_table).where(
+            case({"wendy": "W", "jack": "J"}, value=users_table.c.name, else_="E")
+        )
+
+    The values which are accepted as result values in
+    :paramref:`.case.whens` as well as with :paramref:`.case.else_` are
+    coerced from Python literals into :func:`.bindparam` constructs.
+    SQL expressions, e.g. :class:`_expression.ColumnElement` constructs,
+    are accepted
+    as well.  To coerce a literal string expression into a constant
+    expression rendered inline, use the :func:`_expression.literal_column`
+    construct,
+    as in::
+
+        from sqlalchemy import case, literal_column
+
+        case(
+            (orderline.c.qty > 100, literal_column("'greaterthan100'")),
+            (orderline.c.qty > 10, literal_column("'greaterthan10'")),
+            else_=literal_column("'lessthan10'"),
+        )
+
+    The above will render the given constants without using bound
+    parameters for the result values (but still for the comparison
+    values), as in:
+
+    .. sourcecode:: sql
+
+        CASE
+            WHEN (orderline.qty > :qty_1) THEN 'greaterthan100'
+            WHEN (orderline.qty > :qty_2) THEN 'greaterthan10'
+            ELSE 'lessthan10'
+        END
+
+    :param \*whens: The criteria to be compared against,
+     :paramref:`.case.whens` accepts two different forms, based on
+     whether or not :paramref:`.case.value` is used.
+
+     .. versionchanged:: 1.4 the :func:`_sql.case`
+        function now accepts the series of WHEN conditions positionally
+
+     In the first form, it accepts multiple 2-tuples passed as positional
+     arguments; each 2-tuple consists of ``(<sql expression>, <value>)``,
+     where the SQL expression is a boolean expression and "value" is a
+     resulting value, e.g.::
+
+        case(
+            (users_table.c.name == "wendy", "W"),
+            (users_table.c.name == "jack", "J"),
+        )
+
+     In the second form, it accepts a Python dictionary of comparison
+     values mapped to a resulting value; this form requires
+     :paramref:`.case.value` to be present, and values will be compared
+     using the ``==`` operator, e.g.::
+
+        case({"wendy": "W", "jack": "J"}, value=users_table.c.name)
+
+    :param value: An optional SQL expression which will be used as a
+      fixed "comparison point" for candidate values within a dictionary
+      passed to :paramref:`.case.whens`.
+
+    :param else\_: An optional SQL expression which will be the evaluated
+      result of the ``CASE`` construct if all expressions within
+      :paramref:`.case.whens` evaluate to false.  When omitted, most
+      databases will produce a result of NULL if none of the "when"
+      expressions evaluate to true.
+
+
+    """  # noqa: E501
+    return Case(*whens, value=value, else_=else_)
+
+
+def cast(
+    expression: _ColumnExpressionOrLiteralArgument[Any],
+    type_: _TypeEngineArgument[_T],
+) -> Cast[_T]:
+    r"""Produce a ``CAST`` expression.
+
+    :func:`.cast` returns an instance of :class:`.Cast`.
+
+    E.g.::
+
+        from sqlalchemy import cast, Numeric
+
+        stmt = select(cast(product_table.c.unit_price, Numeric(10, 4)))
+
+    The above statement will produce SQL resembling:
+
+    .. sourcecode:: sql
+
+        SELECT CAST(unit_price AS NUMERIC(10, 4)) FROM product
+
+    The :func:`.cast` function performs two distinct functions when
+    used.  The first is that it renders the ``CAST`` expression within
+    the resulting SQL string.  The second is that it associates the given
+    type (e.g. :class:`.TypeEngine` class or instance) with the column
+    expression on the Python side, which means the expression will take
+    on the expression operator behavior associated with that type,
+    as well as the bound-value handling and result-row-handling behavior
+    of the type.
+
+    An alternative to :func:`.cast` is the :func:`.type_coerce` function.
+    This function performs the second task of associating an expression
+    with a specific type, but does not render the ``CAST`` expression
+    in SQL.
+
+    :param expression: A SQL expression, such as a
+     :class:`_expression.ColumnElement`
+     expression or a Python string which will be coerced into a bound
+     literal value.
+
+    :param type\_: A :class:`.TypeEngine` class or instance indicating
+     the type to which the ``CAST`` should apply.
+
+    .. seealso::
+
+        :ref:`tutorial_casts`
+
+        :func:`.try_cast` - an alternative to CAST that results in
+        NULLs when the cast fails, instead of raising an error.
+        Only supported by some dialects.
+
+        :func:`.type_coerce` - an alternative to CAST that coerces the type
+        on the Python side only, which is often sufficient to generate the
+        correct SQL and data coercion.
+
+
+    """
+    return Cast(expression, type_)
+
+
+def try_cast(
+    expression: _ColumnExpressionOrLiteralArgument[Any],
+    type_: _TypeEngineArgument[_T],
+) -> TryCast[_T]:
+    """Produce a ``TRY_CAST`` expression for backends which support it;
+    this is a ``CAST`` which returns NULL for un-castable conversions.
+
+    In SQLAlchemy, this construct is supported **only** by the SQL Server
+    dialect, and will raise a :class:`.CompileError` if used on other
+    included backends.  However, third party backends may also support
+    this construct.
+
+    .. tip:: As :func:`_sql.try_cast` originates from the SQL Server dialect,
+       it's importable both from ``sqlalchemy.`` as well as from
+       ``sqlalchemy.dialects.mssql``.
+
+    :func:`_sql.try_cast` returns an instance of :class:`.TryCast` and
+    generally behaves similarly to the :class:`.Cast` construct;
+    at the SQL level, the difference between ``CAST`` and ``TRY_CAST``
+    is that ``TRY_CAST`` returns NULL for an un-castable expression,
+    such as attempting to cast a string ``"hi"`` to an integer value.
+
+    E.g.::
+
+        from sqlalchemy import select, try_cast, Numeric
+
+        stmt = select(try_cast(product_table.c.unit_price, Numeric(10, 4)))
+
+    The above would render on Microsoft SQL Server as:
+
+    .. sourcecode:: sql
+
+        SELECT TRY_CAST (product_table.unit_price AS NUMERIC(10, 4))
+        FROM product_table
+
+    .. versionadded:: 2.0.14  :func:`.try_cast` has been
+       generalized from the SQL Server dialect into a general use
+       construct that may be supported by additional dialects.
+
+    """
+    return TryCast(expression, type_)
+
+
+def column(
+    text: str,
+    type_: Optional[_TypeEngineArgument[_T]] = None,
+    is_literal: bool = False,
+    _selectable: Optional[FromClause] = None,
+) -> ColumnClause[_T]:
+    """Produce a :class:`.ColumnClause` object.
+
+    The :class:`.ColumnClause` is a lightweight analogue to the
+    :class:`_schema.Column` class.  The :func:`_expression.column`
+    function can
+    be invoked with just a name alone, as in::
+
+        from sqlalchemy import column
+
+        id, name = column("id"), column("name")
+        stmt = select(id, name).select_from("user")
+
+    The above statement would produce SQL like:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user
+
+    Once constructed, :func:`_expression.column`
+    may be used like any other SQL
+    expression element such as within :func:`_expression.select`
+    constructs::
+
+        from sqlalchemy.sql import column
+
+        id, name = column("id"), column("name")
+        stmt = select(id, name).select_from("user")
+
+    The text handled by :func:`_expression.column`
+    is assumed to be handled
+    like the name of a database column; if the string contains mixed case,
+    special characters, or matches a known reserved word on the target
+    backend, the column expression will render using the quoting
+    behavior determined by the backend.  To produce a textual SQL
+    expression that is rendered exactly without any quoting,
+    use :func:`_expression.literal_column` instead,
+    or pass ``True`` as the
+    value of :paramref:`_expression.column.is_literal`.   Additionally,
+    full SQL
+    statements are best handled using the :func:`_expression.text`
+    construct.
+
+    :func:`_expression.column` can be used in a table-like
+    fashion by combining it with the :func:`.table` function
+    (which is the lightweight analogue to :class:`_schema.Table`
+    ) to produce
+    a working table construct with minimal boilerplate::
+
+        from sqlalchemy import table, column, select
+
+        user = table(
+            "user",
+            column("id"),
+            column("name"),
+            column("description"),
+        )
+
+        stmt = select(user.c.description).where(user.c.name == "wendy")
+
+    A :func:`_expression.column` / :func:`.table`
+    construct like that illustrated
+    above can be created in an
+    ad-hoc fashion and is not associated with any
+    :class:`_schema.MetaData`, DDL, or events, unlike its
+    :class:`_schema.Table` counterpart.
+
+    :param text: the text of the element.
+
+    :param type: :class:`_types.TypeEngine` object which can associate
+      this :class:`.ColumnClause` with a type.
+
+    :param is_literal: if True, the :class:`.ColumnClause` is assumed to
+      be an exact expression that will be delivered to the output with no
+      quoting rules applied regardless of case sensitive settings. the
+      :func:`_expression.literal_column()` function essentially invokes
+      :func:`_expression.column` while passing ``is_literal=True``.
+
+    .. seealso::
+
+        :class:`_schema.Column`
+
+        :func:`_expression.literal_column`
+
+        :func:`.table`
+
+        :func:`_expression.text`
+
+        :ref:`tutorial_select_arbitrary_text`
+
+    """
+    return ColumnClause(text, type_, is_literal, _selectable)
+
+
+def desc(
+    column: _ColumnExpressionOrStrLabelArgument[_T],
+) -> UnaryExpression[_T]:
+    """Produce a descending ``ORDER BY`` clause element.
+
+    e.g.::
+
+        from sqlalchemy import desc
+
+        stmt = select(users_table).order_by(desc(users_table.c.name))
+
+    will produce SQL as:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user ORDER BY name DESC
+
+    The :func:`.desc` function is a standalone version of the
+    :meth:`_expression.ColumnElement.desc`
+    method available on all SQL expressions,
+    e.g.::
+
+
+        stmt = select(users_table).order_by(users_table.c.name.desc())
+
+    :param column: A :class:`_expression.ColumnElement` (e.g.
+     scalar SQL expression)
+     with which to apply the :func:`.desc` operation.
+
+    .. seealso::
+
+        :func:`.asc`
+
+        :func:`.nulls_first`
+
+        :func:`.nulls_last`
+
+        :meth:`_expression.Select.order_by`
+
+    """
+    return UnaryExpression._create_desc(column)
+
+
+def distinct(expr: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]:
+    """Produce an column-expression-level unary ``DISTINCT`` clause.
+
+    This applies the ``DISTINCT`` keyword to an **individual column
+    expression** (e.g. not the whole statement), and renders **specifically
+    in that column position**; this is used for containment within
+    an aggregate function, as in::
+
+        from sqlalchemy import distinct, func
+
+        stmt = select(users_table.c.id, func.count(distinct(users_table.c.name)))
+
+    The above would produce an statement resembling:
+
+    .. sourcecode:: sql
+
+        SELECT user.id, count(DISTINCT user.name) FROM user
+
+    .. tip:: The :func:`_sql.distinct` function does **not** apply DISTINCT
+       to the full SELECT statement, instead applying a DISTINCT modifier
+       to **individual column expressions**.  For general ``SELECT DISTINCT``
+       support, use the
+       :meth:`_sql.Select.distinct` method on :class:`_sql.Select`.
+
+    The :func:`.distinct` function is also available as a column-level
+    method, e.g. :meth:`_expression.ColumnElement.distinct`, as in::
+
+        stmt = select(func.count(users_table.c.name.distinct()))
+
+    The :func:`.distinct` operator is different from the
+    :meth:`_expression.Select.distinct` method of
+    :class:`_expression.Select`,
+    which produces a ``SELECT`` statement
+    with ``DISTINCT`` applied to the result set as a whole,
+    e.g. a ``SELECT DISTINCT`` expression.  See that method for further
+    information.
+
+    .. seealso::
+
+        :meth:`_expression.ColumnElement.distinct`
+
+        :meth:`_expression.Select.distinct`
+
+        :data:`.func`
+
+    """  # noqa: E501
+    return UnaryExpression._create_distinct(expr)
+
+
+def bitwise_not(expr: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]:
+    """Produce a unary bitwise NOT clause, typically via the ``~`` operator.
+
+    Not to be confused with boolean negation :func:`_sql.not_`.
+
+    .. versionadded:: 2.0.2
+
+    .. seealso::
+
+        :ref:`operators_bitwise`
+
+
+    """
+
+    return UnaryExpression._create_bitwise_not(expr)
+
+
+def extract(field: str, expr: _ColumnExpressionArgument[Any]) -> Extract:
+    """Return a :class:`.Extract` construct.
+
+    This is typically available as :func:`.extract`
+    as well as ``func.extract`` from the
+    :data:`.func` namespace.
+
+    :param field: The field to extract.
+
+     .. warning:: This field is used as a literal SQL string.
+         **DO NOT PASS UNTRUSTED INPUT TO THIS STRING**.
+
+    :param expr: A column or Python scalar expression serving as the
+      right side of the ``EXTRACT`` expression.
+
+    E.g.::
+
+        from sqlalchemy import extract
+        from sqlalchemy import table, column
+
+        logged_table = table(
+            "user",
+            column("id"),
+            column("date_created"),
+        )
+
+        stmt = select(logged_table.c.id).where(
+            extract("YEAR", logged_table.c.date_created) == 2021
+        )
+
+    In the above example, the statement is used to select ids from the
+    database where the ``YEAR`` component matches a specific value.
+
+    Similarly, one can also select an extracted component::
+
+        stmt = select(extract("YEAR", logged_table.c.date_created)).where(
+            logged_table.c.id == 1
+        )
+
+    The implementation of ``EXTRACT`` may vary across database backends.
+    Users are reminded to consult their database documentation.
+    """
+    return Extract(field, expr)
+
+
+def false() -> False_:
+    """Return a :class:`.False_` construct.
+
+    E.g.:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import false
+        >>> print(select(t.c.x).where(false()))
+        {printsql}SELECT x FROM t WHERE false
+
+    A backend which does not support true/false constants will render as
+    an expression against 1 or 0:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(t.c.x).where(false()))
+        {printsql}SELECT x FROM t WHERE 0 = 1
+
+    The :func:`.true` and :func:`.false` constants also feature
+    "short circuit" operation within an :func:`.and_` or :func:`.or_`
+    conjunction:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(t.c.x).where(or_(t.c.x > 5, true())))
+        {printsql}SELECT x FROM t WHERE true{stop}
+
+        >>> print(select(t.c.x).where(and_(t.c.x > 5, false())))
+        {printsql}SELECT x FROM t WHERE false{stop}
+
+    .. seealso::
+
+        :func:`.true`
+
+    """
+
+    return False_._instance()
+
+
+def funcfilter(
+    func: FunctionElement[_T], *criterion: _ColumnExpressionArgument[bool]
+) -> FunctionFilter[_T]:
+    """Produce a :class:`.FunctionFilter` object against a function.
+
+    Used against aggregate and window functions,
+    for database backends that support the "FILTER" clause.
+
+    E.g.::
+
+        from sqlalchemy import funcfilter
+
+        funcfilter(func.count(1), MyClass.name == "some name")
+
+    Would produce "COUNT(1) FILTER (WHERE myclass.name = 'some name')".
+
+    This function is also available from the :data:`~.expression.func`
+    construct itself via the :meth:`.FunctionElement.filter` method.
+
+    .. seealso::
+
+        :ref:`tutorial_functions_within_group` - in the
+        :ref:`unified_tutorial`
+
+        :meth:`.FunctionElement.filter`
+
+    """
+    return FunctionFilter(func, *criterion)
+
+
+def label(
+    name: str,
+    element: _ColumnExpressionArgument[_T],
+    type_: Optional[_TypeEngineArgument[_T]] = None,
+) -> Label[_T]:
+    """Return a :class:`Label` object for the
+    given :class:`_expression.ColumnElement`.
+
+    A label changes the name of an element in the columns clause of a
+    ``SELECT`` statement, typically via the ``AS`` SQL keyword.
+
+    This functionality is more conveniently available via the
+    :meth:`_expression.ColumnElement.label` method on
+    :class:`_expression.ColumnElement`.
+
+    :param name: label name
+
+    :param obj: a :class:`_expression.ColumnElement`.
+
+    """
+    return Label(name, element, type_)
+
+
+def null() -> Null:
+    """Return a constant :class:`.Null` construct."""
+
+    return Null._instance()
+
+
+def nulls_first(column: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]:
+    """Produce the ``NULLS FIRST`` modifier for an ``ORDER BY`` expression.
+
+    :func:`.nulls_first` is intended to modify the expression produced
+    by :func:`.asc` or :func:`.desc`, and indicates how NULL values
+    should be handled when they are encountered during ordering::
+
+
+        from sqlalchemy import desc, nulls_first
+
+        stmt = select(users_table).order_by(nulls_first(desc(users_table.c.name)))
+
+    The SQL expression from the above would resemble:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user ORDER BY name DESC NULLS FIRST
+
+    Like :func:`.asc` and :func:`.desc`, :func:`.nulls_first` is typically
+    invoked from the column expression itself using
+    :meth:`_expression.ColumnElement.nulls_first`,
+    rather than as its standalone
+    function version, as in::
+
+        stmt = select(users_table).order_by(
+            users_table.c.name.desc().nulls_first()
+        )
+
+    .. versionchanged:: 1.4 :func:`.nulls_first` is renamed from
+        :func:`.nullsfirst` in previous releases.
+        The previous name remains available for backwards compatibility.
+
+    .. seealso::
+
+        :func:`.asc`
+
+        :func:`.desc`
+
+        :func:`.nulls_last`
+
+        :meth:`_expression.Select.order_by`
+
+    """  # noqa: E501
+    return UnaryExpression._create_nulls_first(column)
+
+
+def nulls_last(column: _ColumnExpressionArgument[_T]) -> UnaryExpression[_T]:
+    """Produce the ``NULLS LAST`` modifier for an ``ORDER BY`` expression.
+
+    :func:`.nulls_last` is intended to modify the expression produced
+    by :func:`.asc` or :func:`.desc`, and indicates how NULL values
+    should be handled when they are encountered during ordering::
+
+
+        from sqlalchemy import desc, nulls_last
+
+        stmt = select(users_table).order_by(nulls_last(desc(users_table.c.name)))
+
+    The SQL expression from the above would resemble:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user ORDER BY name DESC NULLS LAST
+
+    Like :func:`.asc` and :func:`.desc`, :func:`.nulls_last` is typically
+    invoked from the column expression itself using
+    :meth:`_expression.ColumnElement.nulls_last`,
+    rather than as its standalone
+    function version, as in::
+
+        stmt = select(users_table).order_by(users_table.c.name.desc().nulls_last())
+
+    .. versionchanged:: 1.4 :func:`.nulls_last` is renamed from
+        :func:`.nullslast` in previous releases.
+        The previous name remains available for backwards compatibility.
+
+    .. seealso::
+
+        :func:`.asc`
+
+        :func:`.desc`
+
+        :func:`.nulls_first`
+
+        :meth:`_expression.Select.order_by`
+
+    """  # noqa: E501
+    return UnaryExpression._create_nulls_last(column)
+
+
+def or_(  # type: ignore[empty-body]
+    initial_clause: Union[Literal[False], _ColumnExpressionArgument[bool]],
+    *clauses: _ColumnExpressionArgument[bool],
+) -> ColumnElement[bool]:
+    """Produce a conjunction of expressions joined by ``OR``.
+
+    E.g.::
+
+        from sqlalchemy import or_
+
+        stmt = select(users_table).where(
+            or_(users_table.c.name == "wendy", users_table.c.name == "jack")
+        )
+
+    The :func:`.or_` conjunction is also available using the
+    Python ``|`` operator (though note that compound expressions
+    need to be parenthesized in order to function with Python
+    operator precedence behavior)::
+
+        stmt = select(users_table).where(
+            (users_table.c.name == "wendy") | (users_table.c.name == "jack")
+        )
+
+    The :func:`.or_` construct must be given at least one positional
+    argument in order to be valid; a :func:`.or_` construct with no
+    arguments is ambiguous.   To produce an "empty" or dynamically
+    generated :func:`.or_`  expression, from a given list of expressions,
+    a "default" element of :func:`_sql.false` (or just ``False``) should be
+    specified::
+
+        from sqlalchemy import false
+
+        or_criteria = or_(false(), *expressions)
+
+    The above expression will compile to SQL as the expression ``false``
+    or ``0 = 1``, depending on backend, if no other expressions are
+    present.  If expressions are present, then the :func:`_sql.false` value is
+    ignored as it does not affect the outcome of an OR expression which
+    has other elements.
+
+    .. deprecated:: 1.4  The :func:`.or_` element now requires that at
+       least one argument is passed; creating the :func:`.or_` construct
+       with no arguments is deprecated, and will emit a deprecation warning
+       while continuing to produce a blank SQL string.
+
+    .. seealso::
+
+        :func:`.and_`
+
+    """
+    ...
+
+
+if not TYPE_CHECKING:
+    # handle deprecated case which allows zero-arguments
+    def or_(*clauses):  # noqa: F811
+        """Produce a conjunction of expressions joined by ``OR``.
+
+        E.g.::
+
+            from sqlalchemy import or_
+
+            stmt = select(users_table).where(
+                or_(users_table.c.name == "wendy", users_table.c.name == "jack")
+            )
+
+        The :func:`.or_` conjunction is also available using the
+        Python ``|`` operator (though note that compound expressions
+        need to be parenthesized in order to function with Python
+        operator precedence behavior)::
+
+            stmt = select(users_table).where(
+                (users_table.c.name == "wendy") | (users_table.c.name == "jack")
+            )
+
+        The :func:`.or_` construct must be given at least one positional
+        argument in order to be valid; a :func:`.or_` construct with no
+        arguments is ambiguous.   To produce an "empty" or dynamically
+        generated :func:`.or_`  expression, from a given list of expressions,
+        a "default" element of :func:`_sql.false` (or just ``False``) should be
+        specified::
+
+            from sqlalchemy import false
+
+            or_criteria = or_(false(), *expressions)
+
+        The above expression will compile to SQL as the expression ``false``
+        or ``0 = 1``, depending on backend, if no other expressions are
+        present.  If expressions are present, then the :func:`_sql.false` value
+        is ignored as it does not affect the outcome of an OR expression which
+        has other elements.
+
+        .. deprecated:: 1.4  The :func:`.or_` element now requires that at
+           least one argument is passed; creating the :func:`.or_` construct
+           with no arguments is deprecated, and will emit a deprecation warning
+           while continuing to produce a blank SQL string.
+
+        .. seealso::
+
+            :func:`.and_`
+
+        """  # noqa: E501
+        return BooleanClauseList.or_(*clauses)
+
+
+def over(
+    element: FunctionElement[_T],
+    partition_by: Optional[_ByArgument] = None,
+    order_by: Optional[_ByArgument] = None,
+    range_: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+    rows: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+) -> Over[_T]:
+    r"""Produce an :class:`.Over` object against a function.
+
+    Used against aggregate or so-called "window" functions,
+    for database backends that support window functions.
+
+    :func:`_expression.over` is usually called using
+    the :meth:`.FunctionElement.over` method, e.g.::
+
+        func.row_number().over(order_by=mytable.c.some_column)
+
+    Would produce:
+
+    .. sourcecode:: sql
+
+        ROW_NUMBER() OVER(ORDER BY some_column)
+
+    Ranges are also possible using the :paramref:`.expression.over.range_`
+    and :paramref:`.expression.over.rows` parameters.  These
+    mutually-exclusive parameters each accept a 2-tuple, which contains
+    a combination of integers and None::
+
+        func.row_number().over(order_by=my_table.c.some_column, range_=(None, 0))
+
+    The above would produce:
+
+    .. sourcecode:: sql
+
+        ROW_NUMBER() OVER(ORDER BY some_column
+        RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
+
+    A value of ``None`` indicates "unbounded", a
+    value of zero indicates "current row", and negative / positive
+    integers indicate "preceding" and "following":
+
+    * RANGE BETWEEN 5 PRECEDING AND 10 FOLLOWING::
+
+        func.row_number().over(order_by="x", range_=(-5, 10))
+
+    * ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW::
+
+        func.row_number().over(order_by="x", rows=(None, 0))
+
+    * RANGE BETWEEN 2 PRECEDING AND UNBOUNDED FOLLOWING::
+
+        func.row_number().over(order_by="x", range_=(-2, None))
+
+    * RANGE BETWEEN 1 FOLLOWING AND 3 FOLLOWING::
+
+        func.row_number().over(order_by="x", range_=(1, 3))
+
+    :param element: a :class:`.FunctionElement`, :class:`.WithinGroup`,
+     or other compatible construct.
+    :param partition_by: a column element or string, or a list
+     of such, that will be used as the PARTITION BY clause
+     of the OVER construct.
+    :param order_by: a column element or string, or a list
+     of such, that will be used as the ORDER BY clause
+     of the OVER construct.
+    :param range\_: optional range clause for the window.  This is a
+     tuple value which can contain integer values or ``None``,
+     and will render a RANGE BETWEEN PRECEDING / FOLLOWING clause.
+
+    :param rows: optional rows clause for the window.  This is a tuple
+     value which can contain integer values or None, and will render
+     a ROWS BETWEEN PRECEDING / FOLLOWING clause.
+
+    This function is also available from the :data:`~.expression.func`
+    construct itself via the :meth:`.FunctionElement.over` method.
+
+    .. seealso::
+
+        :ref:`tutorial_window_functions` - in the :ref:`unified_tutorial`
+
+        :data:`.expression.func`
+
+        :func:`_expression.within_group`
+
+    """  # noqa: E501
+    return Over(element, partition_by, order_by, range_, rows)
+
+
+@_document_text_coercion("text", ":func:`.text`", ":paramref:`.text.text`")
+def text(text: str) -> TextClause:
+    r"""Construct a new :class:`_expression.TextClause` clause,
+    representing
+    a textual SQL string directly.
+
+    E.g.::
+
+        from sqlalchemy import text
+
+        t = text("SELECT * FROM users")
+        result = connection.execute(t)
+
+    The advantages :func:`_expression.text`
+    provides over a plain string are
+    backend-neutral support for bind parameters, per-statement
+    execution options, as well as
+    bind parameter and result-column typing behavior, allowing
+    SQLAlchemy type constructs to play a role when executing
+    a statement that is specified literally.  The construct can also
+    be provided with a ``.c`` collection of column elements, allowing
+    it to be embedded in other SQL expression constructs as a subquery.
+
+    Bind parameters are specified by name, using the format ``:name``.
+    E.g.::
+
+        t = text("SELECT * FROM users WHERE id=:user_id")
+        result = connection.execute(t, {"user_id": 12})
+
+    For SQL statements where a colon is required verbatim, as within
+    an inline string, use a backslash to escape::
+
+        t = text(r"SELECT * FROM users WHERE name='\:username'")
+
+    The :class:`_expression.TextClause`
+    construct includes methods which can
+    provide information about the bound parameters as well as the column
+    values which would be returned from the textual statement, assuming
+    it's an executable SELECT type of statement.  The
+    :meth:`_expression.TextClause.bindparams`
+    method is used to provide bound
+    parameter detail, and :meth:`_expression.TextClause.columns`
+    method allows
+    specification of return columns including names and types::
+
+        t = (
+            text("SELECT * FROM users WHERE id=:user_id")
+            .bindparams(user_id=7)
+            .columns(id=Integer, name=String)
+        )
+
+        for id, name in connection.execute(t):
+            print(id, name)
+
+    The :func:`_expression.text` construct is used in cases when
+    a literal string SQL fragment is specified as part of a larger query,
+    such as for the WHERE clause of a SELECT statement::
+
+        s = select(users.c.id, users.c.name).where(text("id=:user_id"))
+        result = connection.execute(s, {"user_id": 12})
+
+    :func:`_expression.text` is also used for the construction
+    of a full, standalone statement using plain text.
+    As such, SQLAlchemy refers
+    to it as an :class:`.Executable` object and may be used
+    like any other statement passed to an ``.execute()`` method.
+
+    :param text:
+      the text of the SQL statement to be created.  Use ``:<param>``
+      to specify bind parameters; they will be compiled to their
+      engine-specific format.
+
+    .. seealso::
+
+        :ref:`tutorial_select_arbitrary_text`
+
+    """
+    return TextClause(text)
+
+
+def true() -> True_:
+    """Return a constant :class:`.True_` construct.
+
+    E.g.:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import true
+        >>> print(select(t.c.x).where(true()))
+        {printsql}SELECT x FROM t WHERE true
+
+    A backend which does not support true/false constants will render as
+    an expression against 1 or 0:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(t.c.x).where(true()))
+        {printsql}SELECT x FROM t WHERE 1 = 1
+
+    The :func:`.true` and :func:`.false` constants also feature
+    "short circuit" operation within an :func:`.and_` or :func:`.or_`
+    conjunction:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(t.c.x).where(or_(t.c.x > 5, true())))
+        {printsql}SELECT x FROM t WHERE true{stop}
+
+        >>> print(select(t.c.x).where(and_(t.c.x > 5, false())))
+        {printsql}SELECT x FROM t WHERE false{stop}
+
+    .. seealso::
+
+        :func:`.false`
+
+    """
+
+    return True_._instance()
+
+
+def tuple_(
+    *clauses: _ColumnExpressionArgument[Any],
+    types: Optional[Sequence[_TypeEngineArgument[Any]]] = None,
+) -> Tuple:
+    """Return a :class:`.Tuple`.
+
+    Main usage is to produce a composite IN construct using
+    :meth:`.ColumnOperators.in_` ::
+
+        from sqlalchemy import tuple_
+
+        tuple_(table.c.col1, table.c.col2).in_([(1, 2), (5, 12), (10, 19)])
+
+    .. versionchanged:: 1.3.6 Added support for SQLite IN tuples.
+
+    .. warning::
+
+        The composite IN construct is not supported by all backends, and is
+        currently known to work on PostgreSQL, MySQL, and SQLite.
+        Unsupported backends will raise a subclass of
+        :class:`~sqlalchemy.exc.DBAPIError` when such an expression is
+        invoked.
+
+    """
+    return Tuple(*clauses, types=types)
+
+
+def type_coerce(
+    expression: _ColumnExpressionOrLiteralArgument[Any],
+    type_: _TypeEngineArgument[_T],
+) -> TypeCoerce[_T]:
+    r"""Associate a SQL expression with a particular type, without rendering
+    ``CAST``.
+
+    E.g.::
+
+        from sqlalchemy import type_coerce
+
+        stmt = select(type_coerce(log_table.date_string, StringDateTime()))
+
+    The above construct will produce a :class:`.TypeCoerce` object, which
+    does not modify the rendering in any way on the SQL side, with the
+    possible exception of a generated label if used in a columns clause
+    context:
+
+    .. sourcecode:: sql
+
+        SELECT date_string AS date_string FROM log
+
+    When result rows are fetched, the ``StringDateTime`` type processor
+    will be applied to result rows on behalf of the ``date_string`` column.
+
+    .. note:: the :func:`.type_coerce` construct does not render any
+       SQL syntax of its own, including that it does not imply
+       parenthesization.   Please use :meth:`.TypeCoerce.self_group`
+       if explicit parenthesization is required.
+
+    In order to provide a named label for the expression, use
+    :meth:`_expression.ColumnElement.label`::
+
+        stmt = select(
+            type_coerce(log_table.date_string, StringDateTime()).label("date")
+        )
+
+    A type that features bound-value handling will also have that behavior
+    take effect when literal values or :func:`.bindparam` constructs are
+    passed to :func:`.type_coerce` as targets.
+    For example, if a type implements the
+    :meth:`.TypeEngine.bind_expression`
+    method or :meth:`.TypeEngine.bind_processor` method or equivalent,
+    these functions will take effect at statement compilation/execution
+    time when a literal value is passed, as in::
+
+        # bound-value handling of MyStringType will be applied to the
+        # literal value "some string"
+        stmt = select(type_coerce("some string", MyStringType))
+
+    When using :func:`.type_coerce` with composed expressions, note that
+    **parenthesis are not applied**.   If :func:`.type_coerce` is being
+    used in an operator context where the parenthesis normally present from
+    CAST are necessary, use the :meth:`.TypeCoerce.self_group` method:
+
+    .. sourcecode:: pycon+sql
+
+        >>> some_integer = column("someint", Integer)
+        >>> some_string = column("somestr", String)
+        >>> expr = type_coerce(some_integer + 5, String) + some_string
+        >>> print(expr)
+        {printsql}someint + :someint_1 || somestr{stop}
+        >>> expr = type_coerce(some_integer + 5, String).self_group() + some_string
+        >>> print(expr)
+        {printsql}(someint + :someint_1) || somestr{stop}
+
+    :param expression: A SQL expression, such as a
+     :class:`_expression.ColumnElement`
+     expression or a Python string which will be coerced into a bound
+     literal value.
+
+    :param type\_: A :class:`.TypeEngine` class or instance indicating
+     the type to which the expression is coerced.
+
+    .. seealso::
+
+        :ref:`tutorial_casts`
+
+        :func:`.cast`
+
+    """  # noqa
+    return TypeCoerce(expression, type_)
+
+
+def within_group(
+    element: FunctionElement[_T], *order_by: _ColumnExpressionArgument[Any]
+) -> WithinGroup[_T]:
+    r"""Produce a :class:`.WithinGroup` object against a function.
+
+    Used against so-called "ordered set aggregate" and "hypothetical
+    set aggregate" functions, including :class:`.percentile_cont`,
+    :class:`.rank`, :class:`.dense_rank`, etc.
+
+    :func:`_expression.within_group` is usually called using
+    the :meth:`.FunctionElement.within_group` method, e.g.::
+
+        from sqlalchemy import within_group
+
+        stmt = select(
+            department.c.id,
+            func.percentile_cont(0.5).within_group(department.c.salary.desc()),
+        )
+
+    The above statement would produce SQL similar to
+    ``SELECT department.id, percentile_cont(0.5)
+    WITHIN GROUP (ORDER BY department.salary DESC)``.
+
+    :param element: a :class:`.FunctionElement` construct, typically
+     generated by :data:`~.expression.func`.
+    :param \*order_by: one or more column elements that will be used
+     as the ORDER BY clause of the WITHIN GROUP construct.
+
+    .. seealso::
+
+        :ref:`tutorial_functions_within_group` - in the
+        :ref:`unified_tutorial`
+
+        :data:`.expression.func`
+
+        :func:`_expression.over`
+
+    """
+    return WithinGroup(element, *order_by)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_orm_types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_orm_types.py
new file mode 100644
index 00000000..c37d805e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_orm_types.py
@@ -0,0 +1,20 @@
+# sql/_orm_types.py
+# Copyright (C) 2022-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""ORM types that need to present specifically for **documentation only** of
+the Executable.execution_options() method, which includes options that
+are meaningful to the ORM.
+
+"""
+
+
+from __future__ import annotations
+
+from ..util.typing import Literal
+
+SynchronizeSessionArgument = Literal[False, "auto", "evaluate", "fetch"]
+DMLStrategyArgument = Literal["bulk", "raw", "orm", "auto"]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_py_util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_py_util.py
new file mode 100644
index 00000000..9e1a084a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_py_util.py
@@ -0,0 +1,75 @@
+# sql/_py_util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Dict
+from typing import Tuple
+from typing import Union
+
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    from .cache_key import CacheConst
+
+
+class prefix_anon_map(Dict[str, str]):
+    """A map that creates new keys for missing key access.
+
+    Considers keys of the form "<ident> <name>" to produce
+    new symbols "<name>_<index>", where "index" is an incrementing integer
+    corresponding to <name>.
+
+    Inlines the approach taken by :class:`sqlalchemy.util.PopulateDict` which
+    is otherwise usually used for this type of operation.
+
+    """
+
+    def __missing__(self, key: str) -> str:
+        (ident, derived) = key.split(" ", 1)
+        anonymous_counter = self.get(derived, 1)
+        self[derived] = anonymous_counter + 1  # type: ignore
+        value = f"{derived}_{anonymous_counter}"
+        self[key] = value
+        return value
+
+
+class cache_anon_map(
+    Dict[Union[int, "Literal[CacheConst.NO_CACHE]"], Union[Literal[True], str]]
+):
+    """A map that creates new keys for missing key access.
+
+    Produces an incrementing sequence given a series of unique keys.
+
+    This is similar to the compiler prefix_anon_map class although simpler.
+
+    Inlines the approach taken by :class:`sqlalchemy.util.PopulateDict` which
+    is otherwise usually used for this type of operation.
+
+    """
+
+    _index = 0
+
+    def get_anon(self, object_: Any) -> Tuple[str, bool]:
+        idself = id(object_)
+        if idself in self:
+            s_val = self[idself]
+            assert s_val is not True
+            return s_val, True
+        else:
+            # inline of __missing__
+            self[idself] = id_ = str(self._index)
+            self._index += 1
+
+            return id_, False
+
+    def __missing__(self, key: int) -> str:
+        self[key] = val = str(self._index)
+        self._index += 1
+        return val
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_selectable_constructors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_selectable_constructors.py
new file mode 100644
index 00000000..1660778c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_selectable_constructors.py
@@ -0,0 +1,637 @@
+# sql/_selectable_constructors.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import roles
+from ._typing import _ColumnsClauseArgument
+from ._typing import _no_kw
+from .elements import ColumnClause
+from .selectable import Alias
+from .selectable import CompoundSelect
+from .selectable import Exists
+from .selectable import FromClause
+from .selectable import Join
+from .selectable import Lateral
+from .selectable import LateralFromClause
+from .selectable import NamedFromClause
+from .selectable import Select
+from .selectable import TableClause
+from .selectable import TableSample
+from .selectable import Values
+
+if TYPE_CHECKING:
+    from ._typing import _FromClauseArgument
+    from ._typing import _OnClauseArgument
+    from ._typing import _SelectStatementForCompoundArgument
+    from ._typing import _T0
+    from ._typing import _T1
+    from ._typing import _T2
+    from ._typing import _T3
+    from ._typing import _T4
+    from ._typing import _T5
+    from ._typing import _T6
+    from ._typing import _T7
+    from ._typing import _T8
+    from ._typing import _T9
+    from ._typing import _TypedColumnClauseArgument as _TCCA
+    from .functions import Function
+    from .selectable import CTE
+    from .selectable import HasCTE
+    from .selectable import ScalarSelect
+    from .selectable import SelectBase
+
+
+_T = TypeVar("_T", bound=Any)
+
+
+def alias(
+    selectable: FromClause, name: Optional[str] = None, flat: bool = False
+) -> NamedFromClause:
+    """Return a named alias of the given :class:`.FromClause`.
+
+    For :class:`.Table` and :class:`.Join` objects, the return type is the
+    :class:`_expression.Alias` object. Other kinds of :class:`.NamedFromClause`
+    objects may be returned for other kinds of :class:`.FromClause` objects.
+
+    The named alias represents any :class:`_expression.FromClause` with an
+    alternate name assigned within SQL, typically using the ``AS`` clause when
+    generated, e.g. ``SELECT * FROM table AS aliasname``.
+
+    Equivalent functionality is available via the
+    :meth:`_expression.FromClause.alias`
+    method available on all :class:`_expression.FromClause` objects.
+
+    :param selectable: any :class:`_expression.FromClause` subclass,
+        such as a table, select statement, etc.
+
+    :param name: string name to be assigned as the alias.
+        If ``None``, a name will be deterministically generated at compile
+        time. Deterministic means the name is guaranteed to be unique against
+        other constructs used in the same statement, and will also be the same
+        name for each successive compilation of the same statement object.
+
+    :param flat: Will be passed through to if the given selectable
+     is an instance of :class:`_expression.Join` - see
+     :meth:`_expression.Join.alias` for details.
+
+    """
+    return Alias._factory(selectable, name=name, flat=flat)
+
+
+def cte(
+    selectable: HasCTE, name: Optional[str] = None, recursive: bool = False
+) -> CTE:
+    r"""Return a new :class:`_expression.CTE`,
+    or Common Table Expression instance.
+
+    Please see :meth:`_expression.HasCTE.cte` for detail on CTE usage.
+
+    """
+    return coercions.expect(roles.HasCTERole, selectable).cte(
+        name=name, recursive=recursive
+    )
+
+
+def except_(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return an ``EXCEPT`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+    """
+    return CompoundSelect._create_except(*selects)
+
+
+def except_all(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return an ``EXCEPT ALL`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+    """
+    return CompoundSelect._create_except_all(*selects)
+
+
+def exists(
+    __argument: Optional[
+        Union[_ColumnsClauseArgument[Any], SelectBase, ScalarSelect[Any]]
+    ] = None,
+) -> Exists:
+    """Construct a new :class:`_expression.Exists` construct.
+
+    The :func:`_sql.exists` can be invoked by itself to produce an
+    :class:`_sql.Exists` construct, which will accept simple WHERE
+    criteria::
+
+        exists_criteria = exists().where(table1.c.col1 == table2.c.col2)
+
+    However, for greater flexibility in constructing the SELECT, an
+    existing :class:`_sql.Select` construct may be converted to an
+    :class:`_sql.Exists`, most conveniently by making use of the
+    :meth:`_sql.SelectBase.exists` method::
+
+        exists_criteria = (
+            select(table2.c.col2).where(table1.c.col1 == table2.c.col2).exists()
+        )
+
+    The EXISTS criteria is then used inside of an enclosing SELECT::
+
+        stmt = select(table1.c.col1).where(exists_criteria)
+
+    The above statement will then be of the form:
+
+    .. sourcecode:: sql
+
+        SELECT col1 FROM table1 WHERE EXISTS
+        (SELECT table2.col2 FROM table2 WHERE table2.col2 = table1.col1)
+
+    .. seealso::
+
+        :ref:`tutorial_exists` - in the :term:`2.0 style` tutorial.
+
+        :meth:`_sql.SelectBase.exists` - method to transform a ``SELECT`` to an
+        ``EXISTS`` clause.
+
+    """  # noqa: E501
+
+    return Exists(__argument)
+
+
+def intersect(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return an ``INTERSECT`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+    """
+    return CompoundSelect._create_intersect(*selects)
+
+
+def intersect_all(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return an ``INTERSECT ALL`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+
+    """
+    return CompoundSelect._create_intersect_all(*selects)
+
+
+def join(
+    left: _FromClauseArgument,
+    right: _FromClauseArgument,
+    onclause: Optional[_OnClauseArgument] = None,
+    isouter: bool = False,
+    full: bool = False,
+) -> Join:
+    """Produce a :class:`_expression.Join` object, given two
+    :class:`_expression.FromClause`
+    expressions.
+
+    E.g.::
+
+        j = join(
+            user_table, address_table, user_table.c.id == address_table.c.user_id
+        )
+        stmt = select(user_table).select_from(j)
+
+    would emit SQL along the lines of:
+
+    .. sourcecode:: sql
+
+        SELECT user.id, user.name FROM user
+        JOIN address ON user.id = address.user_id
+
+    Similar functionality is available given any
+    :class:`_expression.FromClause` object (e.g. such as a
+    :class:`_schema.Table`) using
+    the :meth:`_expression.FromClause.join` method.
+
+    :param left: The left side of the join.
+
+    :param right: the right side of the join; this is any
+     :class:`_expression.FromClause` object such as a
+     :class:`_schema.Table` object, and
+     may also be a selectable-compatible object such as an ORM-mapped
+     class.
+
+    :param onclause: a SQL expression representing the ON clause of the
+     join.  If left at ``None``, :meth:`_expression.FromClause.join`
+     will attempt to
+     join the two tables based on a foreign key relationship.
+
+    :param isouter: if True, render a LEFT OUTER JOIN, instead of JOIN.
+
+    :param full: if True, render a FULL OUTER JOIN, instead of JOIN.
+
+    .. seealso::
+
+        :meth:`_expression.FromClause.join` - method form,
+        based on a given left side.
+
+        :class:`_expression.Join` - the type of object produced.
+
+    """  # noqa: E501
+
+    return Join(left, right, onclause, isouter, full)
+
+
+def lateral(
+    selectable: Union[SelectBase, _FromClauseArgument],
+    name: Optional[str] = None,
+) -> LateralFromClause:
+    """Return a :class:`_expression.Lateral` object.
+
+    :class:`_expression.Lateral` is an :class:`_expression.Alias`
+    subclass that represents
+    a subquery with the LATERAL keyword applied to it.
+
+    The special behavior of a LATERAL subquery is that it appears in the
+    FROM clause of an enclosing SELECT, but may correlate to other
+    FROM clauses of that SELECT.   It is a special case of subquery
+    only supported by a small number of backends, currently more recent
+    PostgreSQL versions.
+
+    .. seealso::
+
+        :ref:`tutorial_lateral_correlation` -  overview of usage.
+
+    """
+    return Lateral._factory(selectable, name=name)
+
+
+def outerjoin(
+    left: _FromClauseArgument,
+    right: _FromClauseArgument,
+    onclause: Optional[_OnClauseArgument] = None,
+    full: bool = False,
+) -> Join:
+    """Return an ``OUTER JOIN`` clause element.
+
+    The returned object is an instance of :class:`_expression.Join`.
+
+    Similar functionality is also available via the
+    :meth:`_expression.FromClause.outerjoin` method on any
+    :class:`_expression.FromClause`.
+
+    :param left: The left side of the join.
+
+    :param right: The right side of the join.
+
+    :param onclause:  Optional criterion for the ``ON`` clause, is
+      derived from foreign key relationships established between
+      left and right otherwise.
+
+    To chain joins together, use the :meth:`_expression.FromClause.join`
+    or
+    :meth:`_expression.FromClause.outerjoin` methods on the resulting
+    :class:`_expression.Join` object.
+
+    """
+    return Join(left, right, onclause, isouter=True, full=full)
+
+
+# START OVERLOADED FUNCTIONS select Select 1-10
+
+# code within this block is **programmatically,
+# statically generated** by tools/generate_tuple_map_overloads.py
+
+
+@overload
+def select(__ent0: _TCCA[_T0]) -> Select[Tuple[_T0]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+) -> Select[Tuple[_T0, _T1]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+) -> Select[Tuple[_T0, _T1, _T2]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+) -> Select[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+    __ent5: _TCCA[_T5],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+    __ent5: _TCCA[_T5],
+    __ent6: _TCCA[_T6],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+    __ent5: _TCCA[_T5],
+    __ent6: _TCCA[_T6],
+    __ent7: _TCCA[_T7],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+    __ent5: _TCCA[_T5],
+    __ent6: _TCCA[_T6],
+    __ent7: _TCCA[_T7],
+    __ent8: _TCCA[_T8],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7, _T8]]: ...
+
+
+@overload
+def select(
+    __ent0: _TCCA[_T0],
+    __ent1: _TCCA[_T1],
+    __ent2: _TCCA[_T2],
+    __ent3: _TCCA[_T3],
+    __ent4: _TCCA[_T4],
+    __ent5: _TCCA[_T5],
+    __ent6: _TCCA[_T6],
+    __ent7: _TCCA[_T7],
+    __ent8: _TCCA[_T8],
+    __ent9: _TCCA[_T9],
+) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7, _T8, _T9]]: ...
+
+
+# END OVERLOADED FUNCTIONS select
+
+
+@overload
+def select(
+    *entities: _ColumnsClauseArgument[Any], **__kw: Any
+) -> Select[Any]: ...
+
+
+def select(*entities: _ColumnsClauseArgument[Any], **__kw: Any) -> Select[Any]:
+    r"""Construct a new :class:`_expression.Select`.
+
+
+    .. versionadded:: 1.4 - The :func:`_sql.select` function now accepts
+       column arguments positionally.   The top-level :func:`_sql.select`
+       function will automatically use the 1.x or 2.x style API based on
+       the incoming arguments; using :func:`_sql.select` from the
+       ``sqlalchemy.future`` module will enforce that only the 2.x style
+       constructor is used.
+
+    Similar functionality is also available via the
+    :meth:`_expression.FromClause.select` method on any
+    :class:`_expression.FromClause`.
+
+    .. seealso::
+
+        :ref:`tutorial_selecting_data` - in the :ref:`unified_tutorial`
+
+    :param \*entities:
+      Entities to SELECT from.  For Core usage, this is typically a series
+      of :class:`_expression.ColumnElement` and / or
+      :class:`_expression.FromClause`
+      objects which will form the columns clause of the resulting
+      statement.   For those objects that are instances of
+      :class:`_expression.FromClause` (typically :class:`_schema.Table`
+      or :class:`_expression.Alias`
+      objects), the :attr:`_expression.FromClause.c`
+      collection is extracted
+      to form a collection of :class:`_expression.ColumnElement` objects.
+
+      This parameter will also accept :class:`_expression.TextClause`
+      constructs as
+      given, as well as ORM-mapped classes.
+
+    """
+    # the keyword args are a necessary element in order for the typing
+    # to work out w/ the varargs vs. having named "keyword" arguments that
+    # aren't always present.
+    if __kw:
+        raise _no_kw()
+    return Select(*entities)
+
+
+def table(name: str, *columns: ColumnClause[Any], **kw: Any) -> TableClause:
+    """Produce a new :class:`_expression.TableClause`.
+
+    The object returned is an instance of
+    :class:`_expression.TableClause`, which
+    represents the "syntactical" portion of the schema-level
+    :class:`_schema.Table` object.
+    It may be used to construct lightweight table constructs.
+
+    :param name: Name of the table.
+
+    :param columns: A collection of :func:`_expression.column` constructs.
+
+    :param schema: The schema name for this table.
+
+        .. versionadded:: 1.3.18 :func:`_expression.table` can now
+           accept a ``schema`` argument.
+    """
+
+    return TableClause(name, *columns, **kw)
+
+
+def tablesample(
+    selectable: _FromClauseArgument,
+    sampling: Union[float, Function[Any]],
+    name: Optional[str] = None,
+    seed: Optional[roles.ExpressionElementRole[Any]] = None,
+) -> TableSample:
+    """Return a :class:`_expression.TableSample` object.
+
+    :class:`_expression.TableSample` is an :class:`_expression.Alias`
+    subclass that represents
+    a table with the TABLESAMPLE clause applied to it.
+    :func:`_expression.tablesample`
+    is also available from the :class:`_expression.FromClause`
+    class via the
+    :meth:`_expression.FromClause.tablesample` method.
+
+    The TABLESAMPLE clause allows selecting a randomly selected approximate
+    percentage of rows from a table. It supports multiple sampling methods,
+    most commonly BERNOULLI and SYSTEM.
+
+    e.g.::
+
+        from sqlalchemy import func
+
+        selectable = people.tablesample(
+            func.bernoulli(1), name="alias", seed=func.random()
+        )
+        stmt = select(selectable.c.people_id)
+
+    Assuming ``people`` with a column ``people_id``, the above
+    statement would render as:
+
+    .. sourcecode:: sql
+
+        SELECT alias.people_id FROM
+        people AS alias TABLESAMPLE bernoulli(:bernoulli_1)
+        REPEATABLE (random())
+
+    :param sampling: a ``float`` percentage between 0 and 100 or
+        :class:`_functions.Function`.
+
+    :param name: optional alias name
+
+    :param seed: any real-valued SQL expression.  When specified, the
+     REPEATABLE sub-clause is also rendered.
+
+    """
+    return TableSample._factory(selectable, sampling, name=name, seed=seed)
+
+
+def union(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return a ``UNION`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    A similar :func:`union()` method is available on all
+    :class:`_expression.FromClause` subclasses.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+    :param \**kwargs:
+      available keyword arguments are the same as those of
+      :func:`select`.
+
+    """
+    return CompoundSelect._create_union(*selects)
+
+
+def union_all(
+    *selects: _SelectStatementForCompoundArgument,
+) -> CompoundSelect:
+    r"""Return a ``UNION ALL`` of multiple selectables.
+
+    The returned object is an instance of
+    :class:`_expression.CompoundSelect`.
+
+    A similar :func:`union_all()` method is available on all
+    :class:`_expression.FromClause` subclasses.
+
+    :param \*selects:
+      a list of :class:`_expression.Select` instances.
+
+    """
+    return CompoundSelect._create_union_all(*selects)
+
+
+def values(
+    *columns: ColumnClause[Any],
+    name: Optional[str] = None,
+    literal_binds: bool = False,
+) -> Values:
+    r"""Construct a :class:`_expression.Values` construct.
+
+    The column expressions and the actual data for
+    :class:`_expression.Values` are given in two separate steps.  The
+    constructor receives the column expressions typically as
+    :func:`_expression.column` constructs,
+    and the data is then passed via the
+    :meth:`_expression.Values.data` method as a list,
+    which can be called multiple
+    times to add more data, e.g.::
+
+        from sqlalchemy import column
+        from sqlalchemy import values
+
+        value_expr = values(
+            column("id", Integer),
+            column("name", String),
+            name="my_values",
+        ).data([(1, "name1"), (2, "name2"), (3, "name3")])
+
+    :param \*columns: column expressions, typically composed using
+     :func:`_expression.column` objects.
+
+    :param name: the name for this VALUES construct.  If omitted, the
+     VALUES construct will be unnamed in a SQL expression.   Different
+     backends may have different requirements here.
+
+    :param literal_binds: Defaults to False.  Whether or not to render
+     the data values inline in the SQL output, rather than using bound
+     parameters.
+
+    """
+    return Values(*columns, literal_binds=literal_binds, name=name)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_typing.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_typing.py
new file mode 100644
index 00000000..cf9129b4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/_typing.py
@@ -0,0 +1,460 @@
+# sql/_typing.py
+# Copyright (C) 2022-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import operator
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import roles
+from .. import exc
+from .. import util
+from ..inspection import Inspectable
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import TypeAlias
+
+if TYPE_CHECKING:
+    from datetime import date
+    from datetime import datetime
+    from datetime import time
+    from datetime import timedelta
+    from decimal import Decimal
+    from uuid import UUID
+
+    from .base import Executable
+    from .compiler import Compiled
+    from .compiler import DDLCompiler
+    from .compiler import SQLCompiler
+    from .dml import UpdateBase
+    from .dml import ValuesBase
+    from .elements import ClauseElement
+    from .elements import ColumnElement
+    from .elements import KeyedColumnElement
+    from .elements import quoted_name
+    from .elements import SQLCoreOperations
+    from .elements import TextClause
+    from .lambdas import LambdaElement
+    from .roles import FromClauseRole
+    from .schema import Column
+    from .selectable import Alias
+    from .selectable import CTE
+    from .selectable import FromClause
+    from .selectable import Join
+    from .selectable import NamedFromClause
+    from .selectable import ReturnsRows
+    from .selectable import Select
+    from .selectable import Selectable
+    from .selectable import SelectBase
+    from .selectable import Subquery
+    from .selectable import TableClause
+    from .sqltypes import TableValueType
+    from .sqltypes import TupleType
+    from .type_api import TypeEngine
+    from ..engine import Dialect
+    from ..util.typing import TypeGuard
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+
+_CE = TypeVar("_CE", bound="ColumnElement[Any]")
+
+_CLE = TypeVar("_CLE", bound="ClauseElement")
+
+
+class _HasClauseElement(Protocol, Generic[_T_co]):
+    """indicates a class that has a __clause_element__() method"""
+
+    def __clause_element__(self) -> roles.ExpressionElementRole[_T_co]: ...
+
+
+class _CoreAdapterProto(Protocol):
+    """protocol for the ClauseAdapter/ColumnAdapter.traverse() method."""
+
+    def __call__(self, obj: _CE) -> _CE: ...
+
+
+class _HasDialect(Protocol):
+    """protocol for Engine/Connection-like objects that have dialect
+    attribute.
+    """
+
+    @property
+    def dialect(self) -> Dialect: ...
+
+
+# match column types that are not ORM entities
+_NOT_ENTITY = TypeVar(
+    "_NOT_ENTITY",
+    int,
+    str,
+    bool,
+    "datetime",
+    "date",
+    "time",
+    "timedelta",
+    "UUID",
+    float,
+    "Decimal",
+)
+
+_StarOrOne = Literal["*", 1]
+
+_MAYBE_ENTITY = TypeVar(
+    "_MAYBE_ENTITY",
+    roles.ColumnsClauseRole,
+    _StarOrOne,
+    Type[Any],
+    Inspectable[_HasClauseElement[Any]],
+    _HasClauseElement[Any],
+)
+
+
+# convention:
+# XYZArgument - something that the end user is passing to a public API method
+# XYZElement - the internal representation that we use for the thing.
+# the coercions system is responsible for converting from XYZArgument to
+# XYZElement.
+
+_TextCoercedExpressionArgument = Union[
+    str,
+    "TextClause",
+    "ColumnElement[_T]",
+    _HasClauseElement[_T],
+    roles.ExpressionElementRole[_T],
+]
+
+_ColumnsClauseArgument = Union[
+    roles.TypedColumnsClauseRole[_T],
+    roles.ColumnsClauseRole,
+    "SQLCoreOperations[_T]",
+    _StarOrOne,
+    Type[_T],
+    Inspectable[_HasClauseElement[_T]],
+    _HasClauseElement[_T],
+]
+"""open-ended SELECT columns clause argument.
+
+Includes column expressions, tables, ORM mapped entities, a few literal values.
+
+This type is used for lists of columns  / entities to be returned in result
+sets; select(...), insert().returning(...), etc.
+
+
+"""
+
+_TypedColumnClauseArgument = Union[
+    roles.TypedColumnsClauseRole[_T],
+    "SQLCoreOperations[_T]",
+    Type[_T],
+]
+
+_TP = TypeVar("_TP", bound=Tuple[Any, ...])
+
+_T0 = TypeVar("_T0", bound=Any)
+_T1 = TypeVar("_T1", bound=Any)
+_T2 = TypeVar("_T2", bound=Any)
+_T3 = TypeVar("_T3", bound=Any)
+_T4 = TypeVar("_T4", bound=Any)
+_T5 = TypeVar("_T5", bound=Any)
+_T6 = TypeVar("_T6", bound=Any)
+_T7 = TypeVar("_T7", bound=Any)
+_T8 = TypeVar("_T8", bound=Any)
+_T9 = TypeVar("_T9", bound=Any)
+
+
+_ColumnExpressionArgument = Union[
+    "ColumnElement[_T]",
+    _HasClauseElement[_T],
+    "SQLCoreOperations[_T]",
+    roles.ExpressionElementRole[_T],
+    roles.TypedColumnsClauseRole[_T],
+    Callable[[], "ColumnElement[_T]"],
+    "LambdaElement",
+]
+"See docs in public alias ColumnExpressionArgument."
+
+ColumnExpressionArgument: TypeAlias = _ColumnExpressionArgument[_T]
+"""Narrower "column expression" argument.
+
+This type is used for all the other "column" kinds of expressions that
+typically represent a single SQL column expression, not a set of columns the
+way a table or ORM entity does.
+
+This includes ColumnElement, or ORM-mapped attributes that will have a
+``__clause_element__()`` method, it also has the ExpressionElementRole
+overall which brings in the TextClause object also.
+
+.. versionadded:: 2.0.13
+
+"""
+
+_ColumnExpressionOrLiteralArgument = Union[Any, _ColumnExpressionArgument[_T]]
+
+_ColumnExpressionOrStrLabelArgument = Union[str, _ColumnExpressionArgument[_T]]
+
+_ByArgument = Union[
+    Iterable[_ColumnExpressionOrStrLabelArgument[Any]],
+    _ColumnExpressionOrStrLabelArgument[Any],
+]
+"""Used for keyword-based ``order_by`` and ``partition_by`` parameters."""
+
+
+_InfoType = Dict[Any, Any]
+"""the .info dictionary accepted and used throughout Core /ORM"""
+
+_FromClauseArgument = Union[
+    roles.FromClauseRole,
+    Type[Any],
+    Inspectable[_HasClauseElement[Any]],
+    _HasClauseElement[Any],
+]
+"""A FROM clause, like we would send to select().select_from().
+
+Also accommodates ORM entities and related constructs.
+
+"""
+
+_JoinTargetArgument = Union[_FromClauseArgument, roles.JoinTargetRole]
+"""target for join() builds on _FromClauseArgument to include additional
+join target roles such as those which come from the ORM.
+
+"""
+
+_OnClauseArgument = Union[_ColumnExpressionArgument[Any], roles.OnClauseRole]
+"""target for an ON clause, includes additional roles such as those which
+come from the ORM.
+
+"""
+
+_SelectStatementForCompoundArgument = Union[
+    "SelectBase", roles.CompoundElementRole
+]
+"""SELECT statement acceptable by ``union()`` and other SQL set operations"""
+
+_DMLColumnArgument = Union[
+    str,
+    _HasClauseElement[Any],
+    roles.DMLColumnRole,
+    "SQLCoreOperations[Any]",
+]
+"""A DML column expression.  This is a "key" inside of insert().values(),
+update().values(), and related.
+
+These are usually strings or SQL table columns.
+
+There's also edge cases like JSON expression assignment, which we would want
+the DMLColumnRole to be able to accommodate.
+
+"""
+
+_DMLKey = TypeVar("_DMLKey", bound=_DMLColumnArgument)
+_DMLColumnKeyMapping = Mapping[_DMLKey, Any]
+
+
+_DDLColumnArgument = Union[str, "Column[Any]", roles.DDLConstraintColumnRole]
+"""DDL column.
+
+used for :class:`.PrimaryKeyConstraint`, :class:`.UniqueConstraint`, etc.
+
+"""
+
+_DMLTableArgument = Union[
+    "TableClause",
+    "Join",
+    "Alias",
+    "CTE",
+    Type[Any],
+    Inspectable[_HasClauseElement[Any]],
+    _HasClauseElement[Any],
+]
+
+_PropagateAttrsType = util.immutabledict[str, Any]
+
+_TypeEngineArgument = Union[Type["TypeEngine[_T]"], "TypeEngine[_T]"]
+
+_EquivalentColumnMap = Dict["ColumnElement[Any]", Set["ColumnElement[Any]"]]
+
+_LimitOffsetType = Union[int, _ColumnExpressionArgument[int], None]
+
+_AutoIncrementType = Union[bool, Literal["auto", "ignore_fk"]]
+
+if TYPE_CHECKING:
+
+    def is_sql_compiler(c: Compiled) -> TypeGuard[SQLCompiler]: ...
+
+    def is_ddl_compiler(c: Compiled) -> TypeGuard[DDLCompiler]: ...
+
+    def is_named_from_clause(
+        t: FromClauseRole,
+    ) -> TypeGuard[NamedFromClause]: ...
+
+    def is_column_element(
+        c: ClauseElement,
+    ) -> TypeGuard[ColumnElement[Any]]: ...
+
+    def is_keyed_column_element(
+        c: ClauseElement,
+    ) -> TypeGuard[KeyedColumnElement[Any]]: ...
+
+    def is_text_clause(c: ClauseElement) -> TypeGuard[TextClause]: ...
+
+    def is_from_clause(c: ClauseElement) -> TypeGuard[FromClause]: ...
+
+    def is_tuple_type(t: TypeEngine[Any]) -> TypeGuard[TupleType]: ...
+
+    def is_table_value_type(
+        t: TypeEngine[Any],
+    ) -> TypeGuard[TableValueType]: ...
+
+    def is_selectable(t: Any) -> TypeGuard[Selectable]: ...
+
+    def is_select_base(
+        t: Union[Executable, ReturnsRows]
+    ) -> TypeGuard[SelectBase]: ...
+
+    def is_select_statement(
+        t: Union[Executable, ReturnsRows]
+    ) -> TypeGuard[Select[Any]]: ...
+
+    def is_table(t: FromClause) -> TypeGuard[TableClause]: ...
+
+    def is_subquery(t: FromClause) -> TypeGuard[Subquery]: ...
+
+    def is_dml(c: ClauseElement) -> TypeGuard[UpdateBase]: ...
+
+else:
+    is_sql_compiler = operator.attrgetter("is_sql")
+    is_ddl_compiler = operator.attrgetter("is_ddl")
+    is_named_from_clause = operator.attrgetter("named_with_column")
+    is_column_element = operator.attrgetter("_is_column_element")
+    is_keyed_column_element = operator.attrgetter("_is_keyed_column_element")
+    is_text_clause = operator.attrgetter("_is_text_clause")
+    is_from_clause = operator.attrgetter("_is_from_clause")
+    is_tuple_type = operator.attrgetter("_is_tuple_type")
+    is_table_value_type = operator.attrgetter("_is_table_value")
+    is_selectable = operator.attrgetter("is_selectable")
+    is_select_base = operator.attrgetter("_is_select_base")
+    is_select_statement = operator.attrgetter("_is_select_statement")
+    is_table = operator.attrgetter("_is_table")
+    is_subquery = operator.attrgetter("_is_subquery")
+    is_dml = operator.attrgetter("is_dml")
+
+
+def has_schema_attr(t: FromClauseRole) -> TypeGuard[TableClause]:
+    return hasattr(t, "schema")
+
+
+def is_quoted_name(s: str) -> TypeGuard[quoted_name]:
+    return hasattr(s, "quote")
+
+
+def is_has_clause_element(s: object) -> TypeGuard[_HasClauseElement[Any]]:
+    return hasattr(s, "__clause_element__")
+
+
+def is_insert_update(c: ClauseElement) -> TypeGuard[ValuesBase]:
+    return c.is_dml and (c.is_insert or c.is_update)  # type: ignore
+
+
+def _no_kw() -> exc.ArgumentError:
+    return exc.ArgumentError(
+        "Additional keyword arguments are not accepted by this "
+        "function/method.  The presence of **kw is for pep-484 typing purposes"
+    )
+
+
+def _unexpected_kw(methname: str, kw: Dict[str, Any]) -> NoReturn:
+    k = list(kw)[0]
+    raise TypeError(f"{methname} got an unexpected keyword argument '{k}'")
+
+
+@overload
+def Nullable(
+    val: "SQLCoreOperations[_T]",
+) -> "SQLCoreOperations[Optional[_T]]": ...
+
+
+@overload
+def Nullable(
+    val: roles.ExpressionElementRole[_T],
+) -> roles.ExpressionElementRole[Optional[_T]]: ...
+
+
+@overload
+def Nullable(val: Type[_T]) -> Type[Optional[_T]]: ...
+
+
+def Nullable(
+    val: _TypedColumnClauseArgument[_T],
+) -> _TypedColumnClauseArgument[Optional[_T]]:
+    """Types a column or ORM class as nullable.
+
+    This can be used in select and other contexts to express that the value of
+    a column can be null, for example due to an outer join::
+
+        stmt1 = select(A, Nullable(B)).outerjoin(A.bs)
+        stmt2 = select(A.data, Nullable(B.data)).outerjoin(A.bs)
+
+    At runtime this method returns the input unchanged.
+
+    .. versionadded:: 2.0.20
+    """
+    return val
+
+
+@overload
+def NotNullable(
+    val: "SQLCoreOperations[Optional[_T]]",
+) -> "SQLCoreOperations[_T]": ...
+
+
+@overload
+def NotNullable(
+    val: roles.ExpressionElementRole[Optional[_T]],
+) -> roles.ExpressionElementRole[_T]: ...
+
+
+@overload
+def NotNullable(val: Type[Optional[_T]]) -> Type[_T]: ...
+
+
+@overload
+def NotNullable(val: Optional[Type[_T]]) -> Type[_T]: ...
+
+
+def NotNullable(
+    val: Union[_TypedColumnClauseArgument[Optional[_T]], Optional[Type[_T]]],
+) -> _TypedColumnClauseArgument[_T]:
+    """Types a column or ORM class as not nullable.
+
+    This can be used in select and other contexts to express that the value of
+    a column cannot be null, for example due to a where condition on a
+    nullable column::
+
+        stmt = select(NotNullable(A.value)).where(A.value.is_not(None))
+
+    At runtime this method returns the input unchanged.
+
+    .. versionadded:: 2.0.20
+    """
+    return val  # type: ignore
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/annotation.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/annotation.py
new file mode 100644
index 00000000..bf445ff3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/annotation.py
@@ -0,0 +1,585 @@
+# sql/annotation.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""The :class:`.Annotated` class and related routines; creates hash-equivalent
+copies of SQL constructs which contain context-specific markers and
+associations.
+
+Note that the :class:`.Annotated` concept as implemented in this module is not
+related in any way to the pep-593 concept of "Annotated".
+
+
+"""
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import FrozenSet
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from . import operators
+from .cache_key import HasCacheKey
+from .visitors import anon_map
+from .visitors import ExternallyTraversible
+from .visitors import InternalTraversal
+from .. import util
+from ..util.typing import Literal
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from .base import _EntityNamespace
+    from .visitors import _TraverseInternalsType
+
+_AnnotationDict = Mapping[str, Any]
+
+EMPTY_ANNOTATIONS: util.immutabledict[str, Any] = util.EMPTY_DICT
+
+
+class SupportsAnnotations(ExternallyTraversible):
+    __slots__ = ()
+
+    _annotations: util.immutabledict[str, Any] = EMPTY_ANNOTATIONS
+
+    proxy_set: util.generic_fn_descriptor[FrozenSet[Any]]
+
+    _is_immutable: bool
+
+    def _annotate(self, values: _AnnotationDict) -> Self:
+        raise NotImplementedError()
+
+    @overload
+    def _deannotate(
+        self,
+        values: Literal[None] = ...,
+        clone: bool = ...,
+    ) -> Self: ...
+
+    @overload
+    def _deannotate(
+        self,
+        values: Sequence[str] = ...,
+        clone: bool = ...,
+    ) -> SupportsAnnotations: ...
+
+    def _deannotate(
+        self,
+        values: Optional[Sequence[str]] = None,
+        clone: bool = False,
+    ) -> SupportsAnnotations:
+        raise NotImplementedError()
+
+    @util.memoized_property
+    def _annotations_cache_key(self) -> Tuple[Any, ...]:
+        anon_map_ = anon_map()
+
+        return self._gen_annotations_cache_key(anon_map_)
+
+    def _gen_annotations_cache_key(
+        self, anon_map: anon_map
+    ) -> Tuple[Any, ...]:
+        return (
+            "_annotations",
+            tuple(
+                (
+                    key,
+                    (
+                        value._gen_cache_key(anon_map, [])
+                        if isinstance(value, HasCacheKey)
+                        else value
+                    ),
+                )
+                for key, value in [
+                    (key, self._annotations[key])
+                    for key in sorted(self._annotations)
+                ]
+            ),
+        )
+
+
+class SupportsWrappingAnnotations(SupportsAnnotations):
+    __slots__ = ()
+
+    _constructor: Callable[..., SupportsWrappingAnnotations]
+
+    if TYPE_CHECKING:
+
+        @util.ro_non_memoized_property
+        def entity_namespace(self) -> _EntityNamespace: ...
+
+    def _annotate(self, values: _AnnotationDict) -> Self:
+        """return a copy of this ClauseElement with annotations
+        updated by the given dictionary.
+
+        """
+        return Annotated._as_annotated_instance(self, values)  # type: ignore
+
+    def _with_annotations(self, values: _AnnotationDict) -> Self:
+        """return a copy of this ClauseElement with annotations
+        replaced by the given dictionary.
+
+        """
+        return Annotated._as_annotated_instance(self, values)  # type: ignore
+
+    @overload
+    def _deannotate(
+        self,
+        values: Literal[None] = ...,
+        clone: bool = ...,
+    ) -> Self: ...
+
+    @overload
+    def _deannotate(
+        self,
+        values: Sequence[str] = ...,
+        clone: bool = ...,
+    ) -> SupportsAnnotations: ...
+
+    def _deannotate(
+        self,
+        values: Optional[Sequence[str]] = None,
+        clone: bool = False,
+    ) -> SupportsAnnotations:
+        """return a copy of this :class:`_expression.ClauseElement`
+        with annotations
+        removed.
+
+        :param values: optional tuple of individual values
+         to remove.
+
+        """
+        if clone:
+            s = self._clone()
+            return s
+        else:
+            return self
+
+
+class SupportsCloneAnnotations(SupportsWrappingAnnotations):
+    # SupportsCloneAnnotations extends from SupportsWrappingAnnotations
+    # to support the structure of having the base ClauseElement
+    # be a subclass of SupportsWrappingAnnotations.  Any ClauseElement
+    # subclass that wants to extend from SupportsCloneAnnotations
+    # will inherently also be subclassing SupportsWrappingAnnotations, so
+    # make that specific here.
+
+    if not typing.TYPE_CHECKING:
+        __slots__ = ()
+
+    _clone_annotations_traverse_internals: _TraverseInternalsType = [
+        ("_annotations", InternalTraversal.dp_annotations_key)
+    ]
+
+    def _annotate(self, values: _AnnotationDict) -> Self:
+        """return a copy of this ClauseElement with annotations
+        updated by the given dictionary.
+
+        """
+        new = self._clone()
+        new._annotations = new._annotations.union(values)
+        new.__dict__.pop("_annotations_cache_key", None)
+        new.__dict__.pop("_generate_cache_key", None)
+        return new
+
+    def _with_annotations(self, values: _AnnotationDict) -> Self:
+        """return a copy of this ClauseElement with annotations
+        replaced by the given dictionary.
+
+        """
+        new = self._clone()
+        new._annotations = util.immutabledict(values)
+        new.__dict__.pop("_annotations_cache_key", None)
+        new.__dict__.pop("_generate_cache_key", None)
+        return new
+
+    @overload
+    def _deannotate(
+        self,
+        values: Literal[None] = ...,
+        clone: bool = ...,
+    ) -> Self: ...
+
+    @overload
+    def _deannotate(
+        self,
+        values: Sequence[str] = ...,
+        clone: bool = ...,
+    ) -> SupportsAnnotations: ...
+
+    def _deannotate(
+        self,
+        values: Optional[Sequence[str]] = None,
+        clone: bool = False,
+    ) -> SupportsAnnotations:
+        """return a copy of this :class:`_expression.ClauseElement`
+        with annotations
+        removed.
+
+        :param values: optional tuple of individual values
+         to remove.
+
+        """
+        if clone or self._annotations:
+            # clone is used when we are also copying
+            # the expression for a deep deannotation
+            new = self._clone()
+            new._annotations = util.immutabledict()
+            new.__dict__.pop("_annotations_cache_key", None)
+            return new
+        else:
+            return self
+
+
+class Annotated(SupportsAnnotations):
+    """clones a SupportsAnnotations and applies an 'annotations' dictionary.
+
+    Unlike regular clones, this clone also mimics __hash__() and
+    __eq__() of the original element so that it takes its place
+    in hashed collections.
+
+    A reference to the original element is maintained, for the important
+    reason of keeping its hash value current.  When GC'ed, the
+    hash value may be reused, causing conflicts.
+
+    .. note::  The rationale for Annotated producing a brand new class,
+       rather than placing the functionality directly within ClauseElement,
+       is **performance**.  The __hash__() method is absent on plain
+       ClauseElement which leads to significantly reduced function call
+       overhead, as the use of sets and dictionaries against ClauseElement
+       objects is prevalent, but most are not "annotated".
+
+    """
+
+    _is_column_operators = False
+
+    @classmethod
+    def _as_annotated_instance(
+        cls, element: SupportsWrappingAnnotations, values: _AnnotationDict
+    ) -> Annotated:
+        try:
+            cls = annotated_classes[element.__class__]
+        except KeyError:
+            cls = _new_annotation_type(element.__class__, cls)
+        return cls(element, values)
+
+    _annotations: util.immutabledict[str, Any]
+    __element: SupportsWrappingAnnotations
+    _hash: int
+
+    def __new__(cls: Type[Self], *args: Any) -> Self:
+        return object.__new__(cls)
+
+    def __init__(
+        self, element: SupportsWrappingAnnotations, values: _AnnotationDict
+    ):
+        self.__dict__ = element.__dict__.copy()
+        self.__dict__.pop("_annotations_cache_key", None)
+        self.__dict__.pop("_generate_cache_key", None)
+        self.__element = element
+        self._annotations = util.immutabledict(values)
+        self._hash = hash(element)
+
+    def _annotate(self, values: _AnnotationDict) -> Self:
+        _values = self._annotations.union(values)
+        new = self._with_annotations(_values)
+        return new
+
+    def _with_annotations(self, values: _AnnotationDict) -> Self:
+        clone = self.__class__.__new__(self.__class__)
+        clone.__dict__ = self.__dict__.copy()
+        clone.__dict__.pop("_annotations_cache_key", None)
+        clone.__dict__.pop("_generate_cache_key", None)
+        clone._annotations = util.immutabledict(values)
+        return clone
+
+    @overload
+    def _deannotate(
+        self,
+        values: Literal[None] = ...,
+        clone: bool = ...,
+    ) -> Self: ...
+
+    @overload
+    def _deannotate(
+        self,
+        values: Sequence[str] = ...,
+        clone: bool = ...,
+    ) -> Annotated: ...
+
+    def _deannotate(
+        self,
+        values: Optional[Sequence[str]] = None,
+        clone: bool = True,
+    ) -> SupportsAnnotations:
+        if values is None:
+            return self.__element
+        else:
+            return self._with_annotations(
+                util.immutabledict(
+                    {
+                        key: value
+                        for key, value in self._annotations.items()
+                        if key not in values
+                    }
+                )
+            )
+
+    if not typing.TYPE_CHECKING:
+        # manually proxy some methods that need extra attention
+        def _compiler_dispatch(self, visitor: Any, **kw: Any) -> Any:
+            return self.__element.__class__._compiler_dispatch(
+                self, visitor, **kw
+            )
+
+        @property
+        def _constructor(self):
+            return self.__element._constructor
+
+    def _clone(self, **kw: Any) -> Self:
+        clone = self.__element._clone(**kw)
+        if clone is self.__element:
+            # detect immutable, don't change anything
+            return self
+        else:
+            # update the clone with any changes that have occurred
+            # to this object's __dict__.
+            clone.__dict__.update(self.__dict__)
+            return self.__class__(clone, self._annotations)
+
+    def __reduce__(self) -> Tuple[Type[Annotated], Tuple[Any, ...]]:
+        return self.__class__, (self.__element, self._annotations)
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __eq__(self, other: Any) -> bool:
+        if self._is_column_operators:
+            return self.__element.__class__.__eq__(self, other)
+        else:
+            return hash(other) == hash(self)
+
+    @util.ro_non_memoized_property
+    def entity_namespace(self) -> _EntityNamespace:
+        if "entity_namespace" in self._annotations:
+            return cast(
+                SupportsWrappingAnnotations,
+                self._annotations["entity_namespace"],
+            ).entity_namespace
+        else:
+            return self.__element.entity_namespace
+
+
+# hard-generate Annotated subclasses.  this technique
+# is used instead of on-the-fly types (i.e. type.__new__())
+# so that the resulting objects are pickleable; additionally, other
+# decisions can be made up front about the type of object being annotated
+# just once per class rather than per-instance.
+annotated_classes: Dict[Type[SupportsWrappingAnnotations], Type[Annotated]] = (
+    {}
+)
+
+_SA = TypeVar("_SA", bound="SupportsAnnotations")
+
+
+def _safe_annotate(to_annotate: _SA, annotations: _AnnotationDict) -> _SA:
+    try:
+        _annotate = to_annotate._annotate
+    except AttributeError:
+        # skip objects that don't actually have an `_annotate`
+        # attribute, namely QueryableAttribute inside of a join
+        # condition
+        return to_annotate
+    else:
+        return _annotate(annotations)
+
+
+def _deep_annotate(
+    element: _SA,
+    annotations: _AnnotationDict,
+    exclude: Optional[Sequence[SupportsAnnotations]] = None,
+    *,
+    detect_subquery_cols: bool = False,
+    ind_cols_on_fromclause: bool = False,
+    annotate_callable: Optional[
+        Callable[[SupportsAnnotations, _AnnotationDict], SupportsAnnotations]
+    ] = None,
+) -> _SA:
+    """Deep copy the given ClauseElement, annotating each element
+    with the given annotations dictionary.
+
+    Elements within the exclude collection will be cloned but not annotated.
+
+    """
+
+    # annotated objects hack the __hash__() method so if we want to
+    # uniquely process them we have to use id()
+
+    cloned_ids: Dict[int, SupportsAnnotations] = {}
+
+    def clone(elem: SupportsAnnotations, **kw: Any) -> SupportsAnnotations:
+        # ind_cols_on_fromclause means make sure an AnnotatedFromClause
+        # has its own .c collection independent of that which its proxying.
+        # this is used specifically by orm.LoaderCriteriaOption to break
+        # a reference cycle that it's otherwise prone to building,
+        # see test_relationship_criteria->
+        # test_loader_criteria_subquery_w_same_entity.  logic here was
+        # changed for #8796 and made explicit; previously it occurred
+        # by accident
+
+        kw["detect_subquery_cols"] = detect_subquery_cols
+        id_ = id(elem)
+
+        if id_ in cloned_ids:
+            return cloned_ids[id_]
+
+        if (
+            exclude
+            and hasattr(elem, "proxy_set")
+            and elem.proxy_set.intersection(exclude)
+        ):
+            newelem = elem._clone(clone=clone, **kw)
+        elif annotations != elem._annotations:
+            if detect_subquery_cols and elem._is_immutable:
+                to_annotate = elem._clone(clone=clone, **kw)
+            else:
+                to_annotate = elem
+            if annotate_callable:
+                newelem = annotate_callable(to_annotate, annotations)
+            else:
+                newelem = _safe_annotate(to_annotate, annotations)
+        else:
+            newelem = elem
+
+        newelem._copy_internals(
+            clone=clone, ind_cols_on_fromclause=ind_cols_on_fromclause
+        )
+
+        cloned_ids[id_] = newelem
+        return newelem
+
+    if element is not None:
+        element = cast(_SA, clone(element))
+    clone = None  # type: ignore  # remove gc cycles
+    return element
+
+
+@overload
+def _deep_deannotate(
+    element: Literal[None], values: Optional[Sequence[str]] = None
+) -> Literal[None]: ...
+
+
+@overload
+def _deep_deannotate(
+    element: _SA, values: Optional[Sequence[str]] = None
+) -> _SA: ...
+
+
+def _deep_deannotate(
+    element: Optional[_SA], values: Optional[Sequence[str]] = None
+) -> Optional[_SA]:
+    """Deep copy the given element, removing annotations."""
+
+    cloned: Dict[Any, SupportsAnnotations] = {}
+
+    def clone(elem: SupportsAnnotations, **kw: Any) -> SupportsAnnotations:
+        key: Any
+        if values:
+            key = id(elem)
+        else:
+            key = elem
+
+        if key not in cloned:
+            newelem = elem._deannotate(values=values, clone=True)
+            newelem._copy_internals(clone=clone)
+            cloned[key] = newelem
+            return newelem
+        else:
+            return cloned[key]
+
+    if element is not None:
+        element = cast(_SA, clone(element))
+    clone = None  # type: ignore  # remove gc cycles
+    return element
+
+
+def _shallow_annotate(element: _SA, annotations: _AnnotationDict) -> _SA:
+    """Annotate the given ClauseElement and copy its internals so that
+    internal objects refer to the new annotated object.
+
+    Basically used to apply a "don't traverse" annotation to a
+    selectable, without digging throughout the whole
+    structure wasting time.
+    """
+    element = element._annotate(annotations)
+    element._copy_internals()
+    return element
+
+
+def _new_annotation_type(
+    cls: Type[SupportsWrappingAnnotations], base_cls: Type[Annotated]
+) -> Type[Annotated]:
+    """Generates a new class that subclasses Annotated and proxies a given
+    element type.
+
+    """
+    if issubclass(cls, Annotated):
+        return cls
+    elif cls in annotated_classes:
+        return annotated_classes[cls]
+
+    for super_ in cls.__mro__:
+        # check if an Annotated subclass more specific than
+        # the given base_cls is already registered, such
+        # as AnnotatedColumnElement.
+        if super_ in annotated_classes:
+            base_cls = annotated_classes[super_]
+            break
+
+    annotated_classes[cls] = anno_cls = cast(
+        Type[Annotated],
+        type("Annotated%s" % cls.__name__, (base_cls, cls), {}),
+    )
+    globals()["Annotated%s" % cls.__name__] = anno_cls
+
+    if "_traverse_internals" in cls.__dict__:
+        anno_cls._traverse_internals = list(cls._traverse_internals) + [
+            ("_annotations", InternalTraversal.dp_annotations_key)
+        ]
+    elif cls.__dict__.get("inherit_cache", False):
+        anno_cls._traverse_internals = list(cls._traverse_internals) + [
+            ("_annotations", InternalTraversal.dp_annotations_key)
+        ]
+
+    # some classes include this even if they have traverse_internals
+    # e.g. BindParameter, add it if present.
+    if cls.__dict__.get("inherit_cache", False):
+        anno_cls.inherit_cache = True  # type: ignore
+    elif "inherit_cache" in cls.__dict__:
+        anno_cls.inherit_cache = cls.__dict__["inherit_cache"]  # type: ignore
+
+    anno_cls._is_column_operators = issubclass(cls, operators.ColumnOperators)
+
+    return anno_cls
+
+
+def _prepare_annotations(
+    target_hierarchy: Type[SupportsWrappingAnnotations],
+    base_cls: Type[Annotated],
+) -> None:
+    for cls in util.walk_subclasses(target_hierarchy):
+        _new_annotation_type(cls, base_cls)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/base.py
new file mode 100644
index 00000000..6d409a9f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/base.py
@@ -0,0 +1,2182 @@
+# sql/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Foundational utilities common to many sql modules.
+
+"""
+
+
+from __future__ import annotations
+
+import collections
+from enum import Enum
+import itertools
+from itertools import zip_longest
+import operator
+import re
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import roles
+from . import visitors
+from .cache_key import HasCacheKey  # noqa
+from .cache_key import MemoizedHasCacheKey  # noqa
+from .traversals import HasCopyInternals  # noqa
+from .visitors import ClauseVisitor
+from .visitors import ExtendedInternalTraversal
+from .visitors import ExternallyTraversible
+from .visitors import InternalTraversal
+from .. import event
+from .. import exc
+from .. import util
+from ..util import HasMemoized as HasMemoized
+from ..util import hybridmethod
+from ..util import typing as compat_typing
+from ..util.typing import Protocol
+from ..util.typing import Self
+from ..util.typing import TypeGuard
+
+if TYPE_CHECKING:
+    from . import coercions
+    from . import elements
+    from . import type_api
+    from ._orm_types import DMLStrategyArgument
+    from ._orm_types import SynchronizeSessionArgument
+    from ._typing import _CLE
+    from .elements import BindParameter
+    from .elements import ClauseList
+    from .elements import ColumnClause  # noqa
+    from .elements import ColumnElement
+    from .elements import NamedColumn
+    from .elements import SQLCoreOperations
+    from .elements import TextClause
+    from .schema import Column
+    from .schema import DefaultGenerator
+    from .selectable import _JoinTargetElement
+    from .selectable import _SelectIterable
+    from .selectable import FromClause
+    from ..engine import Connection
+    from ..engine import CursorResult
+    from ..engine.interfaces import _CoreMultiExecuteParams
+    from ..engine.interfaces import _ExecuteOptions
+    from ..engine.interfaces import _ImmutableExecuteOptions
+    from ..engine.interfaces import CacheStats
+    from ..engine.interfaces import Compiled
+    from ..engine.interfaces import CompiledCacheType
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..engine.interfaces import Dialect
+    from ..engine.interfaces import IsolationLevel
+    from ..engine.interfaces import SchemaTranslateMapType
+    from ..event import dispatcher
+
+if not TYPE_CHECKING:
+    coercions = None  # noqa
+    elements = None  # noqa
+    type_api = None  # noqa
+
+
+class _NoArg(Enum):
+    NO_ARG = 0
+
+    def __repr__(self):
+        return f"_NoArg.{self.name}"
+
+
+NO_ARG = _NoArg.NO_ARG
+
+
+class _NoneName(Enum):
+    NONE_NAME = 0
+    """indicate a 'deferred' name that was ultimately the value None."""
+
+
+_NONE_NAME = _NoneName.NONE_NAME
+
+_T = TypeVar("_T", bound=Any)
+
+_Fn = TypeVar("_Fn", bound=Callable[..., Any])
+
+_AmbiguousTableNameMap = MutableMapping[str, str]
+
+
+class _DefaultDescriptionTuple(NamedTuple):
+    arg: Any
+    is_scalar: Optional[bool]
+    is_callable: Optional[bool]
+    is_sentinel: Optional[bool]
+
+    @classmethod
+    def _from_column_default(
+        cls, default: Optional[DefaultGenerator]
+    ) -> _DefaultDescriptionTuple:
+        return (
+            _DefaultDescriptionTuple(
+                default.arg,  # type: ignore
+                default.is_scalar,
+                default.is_callable,
+                default.is_sentinel,
+            )
+            if default
+            and (
+                default.has_arg
+                or (not default.for_update and default.is_sentinel)
+            )
+            else _DefaultDescriptionTuple(None, None, None, None)
+        )
+
+
+_never_select_column = operator.attrgetter("_omit_from_statements")
+
+
+class _EntityNamespace(Protocol):
+    def __getattr__(self, key: str) -> SQLCoreOperations[Any]: ...
+
+
+class _HasEntityNamespace(Protocol):
+    @util.ro_non_memoized_property
+    def entity_namespace(self) -> _EntityNamespace: ...
+
+
+def _is_has_entity_namespace(element: Any) -> TypeGuard[_HasEntityNamespace]:
+    return hasattr(element, "entity_namespace")
+
+
+# Remove when https://github.com/python/mypy/issues/14640 will be fixed
+_Self = TypeVar("_Self", bound=Any)
+
+
+class Immutable:
+    """mark a ClauseElement as 'immutable' when expressions are cloned.
+
+    "immutable" objects refers to the "mutability" of an object in the
+    context of SQL DQL and DML generation.   Such as, in DQL, one can
+    compose a SELECT or subquery of varied forms, but one cannot modify
+    the structure of a specific table or column within DQL.
+    :class:`.Immutable` is mostly intended to follow this concept, and as
+    such the primary "immutable" objects are :class:`.ColumnClause`,
+    :class:`.Column`, :class:`.TableClause`, :class:`.Table`.
+
+    """
+
+    __slots__ = ()
+
+    _is_immutable = True
+
+    def unique_params(self, *optionaldict, **kwargs):
+        raise NotImplementedError("Immutable objects do not support copying")
+
+    def params(self, *optionaldict, **kwargs):
+        raise NotImplementedError("Immutable objects do not support copying")
+
+    def _clone(self: _Self, **kw: Any) -> _Self:
+        return self
+
+    def _copy_internals(
+        self, *, omit_attrs: Iterable[str] = (), **kw: Any
+    ) -> None:
+        pass
+
+
+class SingletonConstant(Immutable):
+    """Represent SQL constants like NULL, TRUE, FALSE"""
+
+    _is_singleton_constant = True
+
+    _singleton: SingletonConstant
+
+    def __new__(cls: _T, *arg: Any, **kw: Any) -> _T:
+        return cast(_T, cls._singleton)
+
+    @util.non_memoized_property
+    def proxy_set(self) -> FrozenSet[ColumnElement[Any]]:
+        raise NotImplementedError()
+
+    @classmethod
+    def _create_singleton(cls):
+        obj = object.__new__(cls)
+        obj.__init__()  # type: ignore
+
+        # for a long time this was an empty frozenset, meaning
+        # a SingletonConstant would never be a "corresponding column" in
+        # a statement.  This referred to #6259.  However, in #7154 we see
+        # that we do in fact need "correspondence" to work when matching cols
+        # in result sets, so the non-correspondence was moved to a more
+        # specific level when we are actually adapting expressions for SQL
+        # render only.
+        obj.proxy_set = frozenset([obj])
+        cls._singleton = obj
+
+
+def _from_objects(
+    *elements: Union[
+        ColumnElement[Any], FromClause, TextClause, _JoinTargetElement
+    ]
+) -> Iterator[FromClause]:
+    return itertools.chain.from_iterable(
+        [element._from_objects for element in elements]
+    )
+
+
+def _select_iterables(
+    elements: Iterable[roles.ColumnsClauseRole],
+) -> _SelectIterable:
+    """expand tables into individual columns in the
+    given list of column expressions.
+
+    """
+    return itertools.chain.from_iterable(
+        [c._select_iterable for c in elements]
+    )
+
+
+_SelfGenerativeType = TypeVar("_SelfGenerativeType", bound="_GenerativeType")
+
+
+class _GenerativeType(compat_typing.Protocol):
+    def _generate(self) -> Self: ...
+
+
+def _generative(fn: _Fn) -> _Fn:
+    """non-caching _generative() decorator.
+
+    This is basically the legacy decorator that copies the object and
+    runs a method on the new copy.
+
+    """
+
+    @util.decorator
+    def _generative(
+        fn: _Fn, self: _SelfGenerativeType, *args: Any, **kw: Any
+    ) -> _SelfGenerativeType:
+        """Mark a method as generative."""
+
+        self = self._generate()
+        x = fn(self, *args, **kw)
+        assert x is self, "generative methods must return self"
+        return self
+
+    decorated = _generative(fn)
+    decorated.non_generative = fn  # type: ignore
+    return decorated
+
+
+def _exclusive_against(*names: str, **kw: Any) -> Callable[[_Fn], _Fn]:
+    msgs = kw.pop("msgs", {})
+
+    defaults = kw.pop("defaults", {})
+
+    getters = [
+        (name, operator.attrgetter(name), defaults.get(name, None))
+        for name in names
+    ]
+
+    @util.decorator
+    def check(fn, *args, **kw):
+        # make pylance happy by not including "self" in the argument
+        # list
+        self = args[0]
+        args = args[1:]
+        for name, getter, default_ in getters:
+            if getter(self) is not default_:
+                msg = msgs.get(
+                    name,
+                    "Method %s() has already been invoked on this %s construct"
+                    % (fn.__name__, self.__class__),
+                )
+                raise exc.InvalidRequestError(msg)
+        return fn(self, *args, **kw)
+
+    return check
+
+
+def _clone(element, **kw):
+    return element._clone(**kw)
+
+
+def _expand_cloned(
+    elements: Iterable[_CLE],
+) -> Iterable[_CLE]:
+    """expand the given set of ClauseElements to be the set of all 'cloned'
+    predecessors.
+
+    """
+    # TODO: cython candidate
+    return itertools.chain(*[x._cloned_set for x in elements])
+
+
+def _de_clone(
+    elements: Iterable[_CLE],
+) -> Iterable[_CLE]:
+    for x in elements:
+        while x._is_clone_of is not None:
+            x = x._is_clone_of
+        yield x
+
+
+def _cloned_intersection(a: Iterable[_CLE], b: Iterable[_CLE]) -> Set[_CLE]:
+    """return the intersection of sets a and b, counting
+    any overlap between 'cloned' predecessors.
+
+    The returned set is in terms of the entities present within 'a'.
+
+    """
+    all_overlap = set(_expand_cloned(a)).intersection(_expand_cloned(b))
+    return {elem for elem in a if all_overlap.intersection(elem._cloned_set)}
+
+
+def _cloned_difference(a: Iterable[_CLE], b: Iterable[_CLE]) -> Set[_CLE]:
+    all_overlap = set(_expand_cloned(a)).intersection(_expand_cloned(b))
+    return {
+        elem for elem in a if not all_overlap.intersection(elem._cloned_set)
+    }
+
+
+class _DialectArgView(MutableMapping[str, Any]):
+    """A dictionary view of dialect-level arguments in the form
+    <dialectname>_<argument_name>.
+
+    """
+
+    def __init__(self, obj):
+        self.obj = obj
+
+    def _key(self, key):
+        try:
+            dialect, value_key = key.split("_", 1)
+        except ValueError as err:
+            raise KeyError(key) from err
+        else:
+            return dialect, value_key
+
+    def __getitem__(self, key):
+        dialect, value_key = self._key(key)
+
+        try:
+            opt = self.obj.dialect_options[dialect]
+        except exc.NoSuchModuleError as err:
+            raise KeyError(key) from err
+        else:
+            return opt[value_key]
+
+    def __setitem__(self, key, value):
+        try:
+            dialect, value_key = self._key(key)
+        except KeyError as err:
+            raise exc.ArgumentError(
+                "Keys must be of the form <dialectname>_<argname>"
+            ) from err
+        else:
+            self.obj.dialect_options[dialect][value_key] = value
+
+    def __delitem__(self, key):
+        dialect, value_key = self._key(key)
+        del self.obj.dialect_options[dialect][value_key]
+
+    def __len__(self):
+        return sum(
+            len(args._non_defaults)
+            for args in self.obj.dialect_options.values()
+        )
+
+    def __iter__(self):
+        return (
+            "%s_%s" % (dialect_name, value_name)
+            for dialect_name in self.obj.dialect_options
+            for value_name in self.obj.dialect_options[
+                dialect_name
+            ]._non_defaults
+        )
+
+
+class _DialectArgDict(MutableMapping[str, Any]):
+    """A dictionary view of dialect-level arguments for a specific
+    dialect.
+
+    Maintains a separate collection of user-specified arguments
+    and dialect-specified default arguments.
+
+    """
+
+    def __init__(self):
+        self._non_defaults = {}
+        self._defaults = {}
+
+    def __len__(self):
+        return len(set(self._non_defaults).union(self._defaults))
+
+    def __iter__(self):
+        return iter(set(self._non_defaults).union(self._defaults))
+
+    def __getitem__(self, key):
+        if key in self._non_defaults:
+            return self._non_defaults[key]
+        else:
+            return self._defaults[key]
+
+    def __setitem__(self, key, value):
+        self._non_defaults[key] = value
+
+    def __delitem__(self, key):
+        del self._non_defaults[key]
+
+
+@util.preload_module("sqlalchemy.dialects")
+def _kw_reg_for_dialect(dialect_name):
+    dialect_cls = util.preloaded.dialects.registry.load(dialect_name)
+    if dialect_cls.construct_arguments is None:
+        return None
+    return dict(dialect_cls.construct_arguments)
+
+
+class DialectKWArgs:
+    """Establish the ability for a class to have dialect-specific arguments
+    with defaults and constructor validation.
+
+    The :class:`.DialectKWArgs` interacts with the
+    :attr:`.DefaultDialect.construct_arguments` present on a dialect.
+
+    .. seealso::
+
+        :attr:`.DefaultDialect.construct_arguments`
+
+    """
+
+    __slots__ = ()
+
+    _dialect_kwargs_traverse_internals = [
+        ("dialect_options", InternalTraversal.dp_dialect_options)
+    ]
+
+    @classmethod
+    def argument_for(cls, dialect_name, argument_name, default):
+        """Add a new kind of dialect-specific keyword argument for this class.
+
+        E.g.::
+
+            Index.argument_for("mydialect", "length", None)
+
+            some_index = Index("a", "b", mydialect_length=5)
+
+        The :meth:`.DialectKWArgs.argument_for` method is a per-argument
+        way adding extra arguments to the
+        :attr:`.DefaultDialect.construct_arguments` dictionary. This
+        dictionary provides a list of argument names accepted by various
+        schema-level constructs on behalf of a dialect.
+
+        New dialects should typically specify this dictionary all at once as a
+        data member of the dialect class.  The use case for ad-hoc addition of
+        argument names is typically for end-user code that is also using
+        a custom compilation scheme which consumes the additional arguments.
+
+        :param dialect_name: name of a dialect.  The dialect must be
+         locatable, else a :class:`.NoSuchModuleError` is raised.   The
+         dialect must also include an existing
+         :attr:`.DefaultDialect.construct_arguments` collection, indicating
+         that it participates in the keyword-argument validation and default
+         system, else :class:`.ArgumentError` is raised.  If the dialect does
+         not include this collection, then any keyword argument can be
+         specified on behalf of this dialect already.  All dialects packaged
+         within SQLAlchemy include this collection, however for third party
+         dialects, support may vary.
+
+        :param argument_name: name of the parameter.
+
+        :param default: default value of the parameter.
+
+        """
+
+        construct_arg_dictionary = DialectKWArgs._kw_registry[dialect_name]
+        if construct_arg_dictionary is None:
+            raise exc.ArgumentError(
+                "Dialect '%s' does have keyword-argument "
+                "validation and defaults enabled configured" % dialect_name
+            )
+        if cls not in construct_arg_dictionary:
+            construct_arg_dictionary[cls] = {}
+        construct_arg_dictionary[cls][argument_name] = default
+
+    @util.memoized_property
+    def dialect_kwargs(self):
+        """A collection of keyword arguments specified as dialect-specific
+        options to this construct.
+
+        The arguments are present here in their original ``<dialect>_<kwarg>``
+        format.  Only arguments that were actually passed are included;
+        unlike the :attr:`.DialectKWArgs.dialect_options` collection, which
+        contains all options known by this dialect including defaults.
+
+        The collection is also writable; keys are accepted of the
+        form ``<dialect>_<kwarg>`` where the value will be assembled
+        into the list of options.
+
+        .. seealso::
+
+            :attr:`.DialectKWArgs.dialect_options` - nested dictionary form
+
+        """
+        return _DialectArgView(self)
+
+    @property
+    def kwargs(self):
+        """A synonym for :attr:`.DialectKWArgs.dialect_kwargs`."""
+        return self.dialect_kwargs
+
+    _kw_registry = util.PopulateDict(_kw_reg_for_dialect)
+
+    def _kw_reg_for_dialect_cls(self, dialect_name):
+        construct_arg_dictionary = DialectKWArgs._kw_registry[dialect_name]
+        d = _DialectArgDict()
+
+        if construct_arg_dictionary is None:
+            d._defaults.update({"*": None})
+        else:
+            for cls in reversed(self.__class__.__mro__):
+                if cls in construct_arg_dictionary:
+                    d._defaults.update(construct_arg_dictionary[cls])
+        return d
+
+    @util.memoized_property
+    def dialect_options(self):
+        """A collection of keyword arguments specified as dialect-specific
+        options to this construct.
+
+        This is a two-level nested registry, keyed to ``<dialect_name>``
+        and ``<argument_name>``.  For example, the ``postgresql_where``
+        argument would be locatable as::
+
+            arg = my_object.dialect_options["postgresql"]["where"]
+
+        .. versionadded:: 0.9.2
+
+        .. seealso::
+
+            :attr:`.DialectKWArgs.dialect_kwargs` - flat dictionary form
+
+        """
+
+        return util.PopulateDict(
+            util.portable_instancemethod(self._kw_reg_for_dialect_cls)
+        )
+
+    def _validate_dialect_kwargs(self, kwargs: Dict[str, Any]) -> None:
+        # validate remaining kwargs that they all specify DB prefixes
+
+        if not kwargs:
+            return
+
+        for k in kwargs:
+            m = re.match("^(.+?)_(.+)$", k)
+            if not m:
+                raise TypeError(
+                    "Additional arguments should be "
+                    "named <dialectname>_<argument>, got '%s'" % k
+                )
+            dialect_name, arg_name = m.group(1, 2)
+
+            try:
+                construct_arg_dictionary = self.dialect_options[dialect_name]
+            except exc.NoSuchModuleError:
+                util.warn(
+                    "Can't validate argument %r; can't "
+                    "locate any SQLAlchemy dialect named %r"
+                    % (k, dialect_name)
+                )
+                self.dialect_options[dialect_name] = d = _DialectArgDict()
+                d._defaults.update({"*": None})
+                d._non_defaults[arg_name] = kwargs[k]
+            else:
+                if (
+                    "*" not in construct_arg_dictionary
+                    and arg_name not in construct_arg_dictionary
+                ):
+                    raise exc.ArgumentError(
+                        "Argument %r is not accepted by "
+                        "dialect %r on behalf of %r"
+                        % (k, dialect_name, self.__class__)
+                    )
+                else:
+                    construct_arg_dictionary[arg_name] = kwargs[k]
+
+
+class CompileState:
+    """Produces additional object state necessary for a statement to be
+    compiled.
+
+    the :class:`.CompileState` class is at the base of classes that assemble
+    state for a particular statement object that is then used by the
+    compiler.   This process is essentially an extension of the process that
+    the SQLCompiler.visit_XYZ() method takes, however there is an emphasis
+    on converting raw user intent into more organized structures rather than
+    producing string output.   The top-level :class:`.CompileState` for the
+    statement being executed is also accessible when the execution context
+    works with invoking the statement and collecting results.
+
+    The production of :class:`.CompileState` is specific to the compiler,  such
+    as within the :meth:`.SQLCompiler.visit_insert`,
+    :meth:`.SQLCompiler.visit_select` etc. methods.  These methods are also
+    responsible for associating the :class:`.CompileState` with the
+    :class:`.SQLCompiler` itself, if the statement is the "toplevel" statement,
+    i.e. the outermost SQL statement that's actually being executed.
+    There can be other :class:`.CompileState` objects that are not the
+    toplevel, such as when a SELECT subquery or CTE-nested
+    INSERT/UPDATE/DELETE is generated.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ("statement", "_ambiguous_table_name_map")
+
+    plugins: Dict[Tuple[str, str], Type[CompileState]] = {}
+
+    _ambiguous_table_name_map: Optional[_AmbiguousTableNameMap]
+
+    @classmethod
+    def create_for_statement(cls, statement, compiler, **kw):
+        # factory construction.
+
+        if statement._propagate_attrs:
+            plugin_name = statement._propagate_attrs.get(
+                "compile_state_plugin", "default"
+            )
+            klass = cls.plugins.get(
+                (plugin_name, statement._effective_plugin_target), None
+            )
+            if klass is None:
+                klass = cls.plugins[
+                    ("default", statement._effective_plugin_target)
+                ]
+
+        else:
+            klass = cls.plugins[
+                ("default", statement._effective_plugin_target)
+            ]
+
+        if klass is cls:
+            return cls(statement, compiler, **kw)
+        else:
+            return klass.create_for_statement(statement, compiler, **kw)
+
+    def __init__(self, statement, compiler, **kw):
+        self.statement = statement
+
+    @classmethod
+    def get_plugin_class(
+        cls, statement: Executable
+    ) -> Optional[Type[CompileState]]:
+        plugin_name = statement._propagate_attrs.get(
+            "compile_state_plugin", None
+        )
+
+        if plugin_name:
+            key = (plugin_name, statement._effective_plugin_target)
+            if key in cls.plugins:
+                return cls.plugins[key]
+
+        # there's no case where we call upon get_plugin_class() and want
+        # to get None back, there should always be a default.  return that
+        # if there was no plugin-specific class  (e.g. Insert with "orm"
+        # plugin)
+        try:
+            return cls.plugins[("default", statement._effective_plugin_target)]
+        except KeyError:
+            return None
+
+    @classmethod
+    def _get_plugin_class_for_plugin(
+        cls, statement: Executable, plugin_name: str
+    ) -> Optional[Type[CompileState]]:
+        try:
+            return cls.plugins[
+                (plugin_name, statement._effective_plugin_target)
+            ]
+        except KeyError:
+            return None
+
+    @classmethod
+    def plugin_for(
+        cls, plugin_name: str, visit_name: str
+    ) -> Callable[[_Fn], _Fn]:
+        def decorate(cls_to_decorate):
+            cls.plugins[(plugin_name, visit_name)] = cls_to_decorate
+            return cls_to_decorate
+
+        return decorate
+
+
+class Generative(HasMemoized):
+    """Provide a method-chaining pattern in conjunction with the
+    @_generative decorator."""
+
+    def _generate(self) -> Self:
+        skip = self._memoized_keys
+        cls = self.__class__
+        s = cls.__new__(cls)
+        if skip:
+            # ensure this iteration remains atomic
+            s.__dict__ = {
+                k: v for k, v in self.__dict__.copy().items() if k not in skip
+            }
+        else:
+            s.__dict__ = self.__dict__.copy()
+        return s
+
+
+class InPlaceGenerative(HasMemoized):
+    """Provide a method-chaining pattern in conjunction with the
+    @_generative decorator that mutates in place."""
+
+    __slots__ = ()
+
+    def _generate(self):
+        skip = self._memoized_keys
+        # note __dict__ needs to be in __slots__ if this is used
+        for k in skip:
+            self.__dict__.pop(k, None)
+        return self
+
+
+class HasCompileState(Generative):
+    """A class that has a :class:`.CompileState` associated with it."""
+
+    _compile_state_plugin: Optional[Type[CompileState]] = None
+
+    _attributes: util.immutabledict[str, Any] = util.EMPTY_DICT
+
+    _compile_state_factory = CompileState.create_for_statement
+
+
+class _MetaOptions(type):
+    """metaclass for the Options class.
+
+    This metaclass is actually necessary despite the availability of the
+    ``__init_subclass__()`` hook as this type also provides custom class-level
+    behavior for the ``__add__()`` method.
+
+    """
+
+    _cache_attrs: Tuple[str, ...]
+
+    def __add__(self, other):
+        o1 = self()
+
+        if set(other).difference(self._cache_attrs):
+            raise TypeError(
+                "dictionary contains attributes not covered by "
+                "Options class %s: %r"
+                % (self, set(other).difference(self._cache_attrs))
+            )
+
+        o1.__dict__.update(other)
+        return o1
+
+    if TYPE_CHECKING:
+
+        def __getattr__(self, key: str) -> Any: ...
+
+        def __setattr__(self, key: str, value: Any) -> None: ...
+
+        def __delattr__(self, key: str) -> None: ...
+
+
+class Options(metaclass=_MetaOptions):
+    """A cacheable option dictionary with defaults."""
+
+    __slots__ = ()
+
+    _cache_attrs: Tuple[str, ...]
+
+    def __init_subclass__(cls) -> None:
+        dict_ = cls.__dict__
+        cls._cache_attrs = tuple(
+            sorted(
+                d
+                for d in dict_
+                if not d.startswith("__")
+                and d not in ("_cache_key_traversal",)
+            )
+        )
+        super().__init_subclass__()
+
+    def __init__(self, **kw):
+        self.__dict__.update(kw)
+
+    def __add__(self, other):
+        o1 = self.__class__.__new__(self.__class__)
+        o1.__dict__.update(self.__dict__)
+
+        if set(other).difference(self._cache_attrs):
+            raise TypeError(
+                "dictionary contains attributes not covered by "
+                "Options class %s: %r"
+                % (self, set(other).difference(self._cache_attrs))
+            )
+
+        o1.__dict__.update(other)
+        return o1
+
+    def __eq__(self, other):
+        # TODO: very inefficient.  This is used only in test suites
+        # right now.
+        for a, b in zip_longest(self._cache_attrs, other._cache_attrs):
+            if getattr(self, a) != getattr(other, b):
+                return False
+        return True
+
+    def __repr__(self):
+        # TODO: fairly inefficient, used only in debugging right now.
+
+        return "%s(%s)" % (
+            self.__class__.__name__,
+            ", ".join(
+                "%s=%r" % (k, self.__dict__[k])
+                for k in self._cache_attrs
+                if k in self.__dict__
+            ),
+        )
+
+    @classmethod
+    def isinstance(cls, klass: Type[Any]) -> bool:
+        return issubclass(cls, klass)
+
+    @hybridmethod
+    def add_to_element(self, name, value):
+        return self + {name: getattr(self, name) + value}
+
+    @hybridmethod
+    def _state_dict_inst(self) -> Mapping[str, Any]:
+        return self.__dict__
+
+    _state_dict_const: util.immutabledict[str, Any] = util.EMPTY_DICT
+
+    @_state_dict_inst.classlevel
+    def _state_dict(cls) -> Mapping[str, Any]:
+        return cls._state_dict_const
+
+    @classmethod
+    def safe_merge(cls, other):
+        d = other._state_dict()
+
+        # only support a merge with another object of our class
+        # and which does not have attrs that we don't.   otherwise
+        # we risk having state that might not be part of our cache
+        # key strategy
+
+        if (
+            cls is not other.__class__
+            and other._cache_attrs
+            and set(other._cache_attrs).difference(cls._cache_attrs)
+        ):
+            raise TypeError(
+                "other element %r is not empty, is not of type %s, "
+                "and contains attributes not covered here %r"
+                % (
+                    other,
+                    cls,
+                    set(other._cache_attrs).difference(cls._cache_attrs),
+                )
+            )
+        return cls + d
+
+    @classmethod
+    def from_execution_options(
+        cls, key, attrs, exec_options, statement_exec_options
+    ):
+        """process Options argument in terms of execution options.
+
+
+        e.g.::
+
+            (
+                load_options,
+                execution_options,
+            ) = QueryContext.default_load_options.from_execution_options(
+                "_sa_orm_load_options",
+                {"populate_existing", "autoflush", "yield_per"},
+                execution_options,
+                statement._execution_options,
+            )
+
+        get back the Options and refresh "_sa_orm_load_options" in the
+        exec options dict w/ the Options as well
+
+        """
+
+        # common case is that no options we are looking for are
+        # in either dictionary, so cancel for that first
+        check_argnames = attrs.intersection(
+            set(exec_options).union(statement_exec_options)
+        )
+
+        existing_options = exec_options.get(key, cls)
+
+        if check_argnames:
+            result = {}
+            for argname in check_argnames:
+                local = "_" + argname
+                if argname in exec_options:
+                    result[local] = exec_options[argname]
+                elif argname in statement_exec_options:
+                    result[local] = statement_exec_options[argname]
+
+            new_options = existing_options + result
+            exec_options = util.immutabledict().merge_with(
+                exec_options, {key: new_options}
+            )
+            return new_options, exec_options
+
+        else:
+            return existing_options, exec_options
+
+    if TYPE_CHECKING:
+
+        def __getattr__(self, key: str) -> Any: ...
+
+        def __setattr__(self, key: str, value: Any) -> None: ...
+
+        def __delattr__(self, key: str) -> None: ...
+
+
+class CacheableOptions(Options, HasCacheKey):
+    __slots__ = ()
+
+    @hybridmethod
+    def _gen_cache_key_inst(self, anon_map, bindparams):
+        return HasCacheKey._gen_cache_key(self, anon_map, bindparams)
+
+    @_gen_cache_key_inst.classlevel
+    def _gen_cache_key(cls, anon_map, bindparams):
+        return (cls, ())
+
+    @hybridmethod
+    def _generate_cache_key(self):
+        return HasCacheKey._generate_cache_key_for_object(self)
+
+
+class ExecutableOption(HasCopyInternals):
+    __slots__ = ()
+
+    _annotations = util.EMPTY_DICT
+
+    __visit_name__ = "executable_option"
+
+    _is_has_cache_key = False
+
+    _is_core = True
+
+    def _clone(self, **kw):
+        """Create a shallow copy of this ExecutableOption."""
+        c = self.__class__.__new__(self.__class__)
+        c.__dict__ = dict(self.__dict__)  # type: ignore
+        return c
+
+
+class Executable(roles.StatementRole):
+    """Mark a :class:`_expression.ClauseElement` as supporting execution.
+
+    :class:`.Executable` is a superclass for all "statement" types
+    of objects, including :func:`select`, :func:`delete`, :func:`update`,
+    :func:`insert`, :func:`text`.
+
+    """
+
+    supports_execution: bool = True
+    _execution_options: _ImmutableExecuteOptions = util.EMPTY_DICT
+    _is_default_generator = False
+    _with_options: Tuple[ExecutableOption, ...] = ()
+    _with_context_options: Tuple[
+        Tuple[Callable[[CompileState], None], Any], ...
+    ] = ()
+    _compile_options: Optional[Union[Type[CacheableOptions], CacheableOptions]]
+
+    _executable_traverse_internals = [
+        ("_with_options", InternalTraversal.dp_executable_options),
+        (
+            "_with_context_options",
+            ExtendedInternalTraversal.dp_with_context_options,
+        ),
+        ("_propagate_attrs", ExtendedInternalTraversal.dp_propagate_attrs),
+    ]
+
+    is_select = False
+    is_from_statement = False
+    is_update = False
+    is_insert = False
+    is_text = False
+    is_delete = False
+    is_dml = False
+
+    if TYPE_CHECKING:
+        __visit_name__: str
+
+        def _compile_w_cache(
+            self,
+            dialect: Dialect,
+            *,
+            compiled_cache: Optional[CompiledCacheType],
+            column_keys: List[str],
+            for_executemany: bool = False,
+            schema_translate_map: Optional[SchemaTranslateMapType] = None,
+            **kw: Any,
+        ) -> Tuple[
+            Compiled, Optional[Sequence[BindParameter[Any]]], CacheStats
+        ]: ...
+
+        def _execute_on_connection(
+            self,
+            connection: Connection,
+            distilled_params: _CoreMultiExecuteParams,
+            execution_options: CoreExecuteOptionsParameter,
+        ) -> CursorResult[Any]: ...
+
+        def _execute_on_scalar(
+            self,
+            connection: Connection,
+            distilled_params: _CoreMultiExecuteParams,
+            execution_options: CoreExecuteOptionsParameter,
+        ) -> Any: ...
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self):
+        raise NotImplementedError()
+
+    @property
+    def _effective_plugin_target(self) -> str:
+        return self.__visit_name__
+
+    @_generative
+    def options(self, *options: ExecutableOption) -> Self:
+        """Apply options to this statement.
+
+        In the general sense, options are any kind of Python object
+        that can be interpreted by the SQL compiler for the statement.
+        These options can be consumed by specific dialects or specific kinds
+        of compilers.
+
+        The most commonly known kind of option are the ORM level options
+        that apply "eager load" and other loading behaviors to an ORM
+        query.   However, options can theoretically be used for many other
+        purposes.
+
+        For background on specific kinds of options for specific kinds of
+        statements, refer to the documentation for those option objects.
+
+        .. versionchanged:: 1.4 - added :meth:`.Executable.options` to
+           Core statement objects towards the goal of allowing unified
+           Core / ORM querying capabilities.
+
+        .. seealso::
+
+            :ref:`loading_columns` - refers to options specific to the usage
+            of ORM queries
+
+            :ref:`relationship_loader_options` - refers to options specific
+            to the usage of ORM queries
+
+        """
+        self._with_options += tuple(
+            coercions.expect(roles.ExecutableOptionRole, opt)
+            for opt in options
+        )
+        return self
+
+    @_generative
+    def _set_compile_options(self, compile_options: CacheableOptions) -> Self:
+        """Assign the compile options to a new value.
+
+        :param compile_options: appropriate CacheableOptions structure
+
+        """
+
+        self._compile_options = compile_options
+        return self
+
+    @_generative
+    def _update_compile_options(self, options: CacheableOptions) -> Self:
+        """update the _compile_options with new keys."""
+
+        assert self._compile_options is not None
+        self._compile_options += options
+        return self
+
+    @_generative
+    def _add_context_option(
+        self,
+        callable_: Callable[[CompileState], None],
+        cache_args: Any,
+    ) -> Self:
+        """Add a context option to this statement.
+
+        These are callable functions that will
+        be given the CompileState object upon compilation.
+
+        A second argument cache_args is required, which will be combined with
+        the ``__code__`` identity of the function itself in order to produce a
+        cache key.
+
+        """
+        self._with_context_options += ((callable_, cache_args),)
+        return self
+
+    @overload
+    def execution_options(
+        self,
+        *,
+        compiled_cache: Optional[CompiledCacheType] = ...,
+        logging_token: str = ...,
+        isolation_level: IsolationLevel = ...,
+        no_parameters: bool = False,
+        stream_results: bool = False,
+        max_row_buffer: int = ...,
+        yield_per: int = ...,
+        insertmanyvalues_page_size: int = ...,
+        schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+        populate_existing: bool = False,
+        autoflush: bool = False,
+        synchronize_session: SynchronizeSessionArgument = ...,
+        dml_strategy: DMLStrategyArgument = ...,
+        render_nulls: bool = ...,
+        is_delete_using: bool = ...,
+        is_update_from: bool = ...,
+        preserve_rowcount: bool = False,
+        **opt: Any,
+    ) -> Self: ...
+
+    @overload
+    def execution_options(self, **opt: Any) -> Self: ...
+
+    @_generative
+    def execution_options(self, **kw: Any) -> Self:
+        """Set non-SQL options for the statement which take effect during
+        execution.
+
+        Execution options can be set at many scopes, including per-statement,
+        per-connection, or per execution, using methods such as
+        :meth:`_engine.Connection.execution_options` and parameters which
+        accept a dictionary of options such as
+        :paramref:`_engine.Connection.execute.execution_options` and
+        :paramref:`_orm.Session.execute.execution_options`.
+
+        The primary characteristic of an execution option, as opposed to
+        other kinds of options such as ORM loader options, is that
+        **execution options never affect the compiled SQL of a query, only
+        things that affect how the SQL statement itself is invoked or how
+        results are fetched**.  That is, execution options are not part of
+        what's accommodated by SQL compilation nor are they considered part of
+        the cached state of a statement.
+
+        The :meth:`_sql.Executable.execution_options` method is
+        :term:`generative`, as
+        is the case for the method as applied to the :class:`_engine.Engine`
+        and :class:`_orm.Query` objects, which means when the method is called,
+        a copy of the object is returned, which applies the given parameters to
+        that new copy, but leaves the original unchanged::
+
+            statement = select(table.c.x, table.c.y)
+            new_statement = statement.execution_options(my_option=True)
+
+        An exception to this behavior is the :class:`_engine.Connection`
+        object, where the :meth:`_engine.Connection.execution_options` method
+        is explicitly **not** generative.
+
+        The kinds of options that may be passed to
+        :meth:`_sql.Executable.execution_options` and other related methods and
+        parameter dictionaries include parameters that are explicitly consumed
+        by SQLAlchemy Core or ORM, as well as arbitrary keyword arguments not
+        defined by SQLAlchemy, which means the methods and/or parameter
+        dictionaries may be used for user-defined parameters that interact with
+        custom code, which may access the parameters using methods such as
+        :meth:`_sql.Executable.get_execution_options` and
+        :meth:`_engine.Connection.get_execution_options`, or within selected
+        event hooks using a dedicated ``execution_options`` event parameter
+        such as
+        :paramref:`_events.ConnectionEvents.before_execute.execution_options`
+        or :attr:`_orm.ORMExecuteState.execution_options`, e.g.::
+
+             from sqlalchemy import event
+
+
+             @event.listens_for(some_engine, "before_execute")
+             def _process_opt(conn, statement, multiparams, params, execution_options):
+                 "run a SQL function before invoking a statement"
+
+                 if execution_options.get("do_special_thing", False):
+                     conn.exec_driver_sql("run_special_function()")
+
+        Within the scope of options that are explicitly recognized by
+        SQLAlchemy, most apply to specific classes of objects and not others.
+        The most common execution options include:
+
+        * :paramref:`_engine.Connection.execution_options.isolation_level` -
+          sets the isolation level for a connection or a class of connections
+          via an :class:`_engine.Engine`.  This option is accepted only
+          by :class:`_engine.Connection` or :class:`_engine.Engine`.
+
+        * :paramref:`_engine.Connection.execution_options.stream_results` -
+          indicates results should be fetched using a server side cursor;
+          this option is accepted by :class:`_engine.Connection`, by the
+          :paramref:`_engine.Connection.execute.execution_options` parameter
+          on :meth:`_engine.Connection.execute`, and additionally by
+          :meth:`_sql.Executable.execution_options` on a SQL statement object,
+          as well as by ORM constructs like :meth:`_orm.Session.execute`.
+
+        * :paramref:`_engine.Connection.execution_options.compiled_cache` -
+          indicates a dictionary that will serve as the
+          :ref:`SQL compilation cache <sql_caching>`
+          for a :class:`_engine.Connection` or :class:`_engine.Engine`, as
+          well as for ORM methods like :meth:`_orm.Session.execute`.
+          Can be passed as ``None`` to disable caching for statements.
+          This option is not accepted by
+          :meth:`_sql.Executable.execution_options` as it is inadvisable to
+          carry along a compilation cache within a statement object.
+
+        * :paramref:`_engine.Connection.execution_options.schema_translate_map`
+          - a mapping of schema names used by the
+          :ref:`Schema Translate Map <schema_translating>` feature, accepted
+          by :class:`_engine.Connection`, :class:`_engine.Engine`,
+          :class:`_sql.Executable`, as well as by ORM constructs
+          like :meth:`_orm.Session.execute`.
+
+        .. seealso::
+
+            :meth:`_engine.Connection.execution_options`
+
+            :paramref:`_engine.Connection.execute.execution_options`
+
+            :paramref:`_orm.Session.execute.execution_options`
+
+            :ref:`orm_queryguide_execution_options` - documentation on all
+            ORM-specific execution options
+
+        """  # noqa: E501
+        if "isolation_level" in kw:
+            raise exc.ArgumentError(
+                "'isolation_level' execution option may only be specified "
+                "on Connection.execution_options(), or "
+                "per-engine using the isolation_level "
+                "argument to create_engine()."
+            )
+        if "compiled_cache" in kw:
+            raise exc.ArgumentError(
+                "'compiled_cache' execution option may only be specified "
+                "on Connection.execution_options(), not per statement."
+            )
+        self._execution_options = self._execution_options.union(kw)
+        return self
+
+    def get_execution_options(self) -> _ExecuteOptions:
+        """Get the non-SQL options which will take effect during execution.
+
+        .. versionadded:: 1.3
+
+        .. seealso::
+
+            :meth:`.Executable.execution_options`
+        """
+        return self._execution_options
+
+
+class SchemaEventTarget(event.EventTarget):
+    """Base class for elements that are the targets of :class:`.DDLEvents`
+    events.
+
+    This includes :class:`.SchemaItem` as well as :class:`.SchemaType`.
+
+    """
+
+    dispatch: dispatcher[SchemaEventTarget]
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        """Associate with this SchemaEvent's parent object."""
+
+    def _set_parent_with_dispatch(
+        self, parent: SchemaEventTarget, **kw: Any
+    ) -> None:
+        self.dispatch.before_parent_attach(self, parent)
+        self._set_parent(parent, **kw)
+        self.dispatch.after_parent_attach(self, parent)
+
+
+class SchemaVisitor(ClauseVisitor):
+    """Define the visiting for ``SchemaItem`` objects."""
+
+    __traverse_options__ = {"schema_visitor": True}
+
+
+class _SentinelDefaultCharacterization(Enum):
+    NONE = "none"
+    UNKNOWN = "unknown"
+    CLIENTSIDE = "clientside"
+    SENTINEL_DEFAULT = "sentinel_default"
+    SERVERSIDE = "serverside"
+    IDENTITY = "identity"
+    SEQUENCE = "sequence"
+
+
+class _SentinelColumnCharacterization(NamedTuple):
+    columns: Optional[Sequence[Column[Any]]] = None
+    is_explicit: bool = False
+    is_autoinc: bool = False
+    default_characterization: _SentinelDefaultCharacterization = (
+        _SentinelDefaultCharacterization.NONE
+    )
+
+
+_COLKEY = TypeVar("_COLKEY", Union[None, str], str)
+
+_COL_co = TypeVar("_COL_co", bound="ColumnElement[Any]", covariant=True)
+_COL = TypeVar("_COL", bound="ColumnElement[Any]")
+
+
+class _ColumnMetrics(Generic[_COL_co]):
+    __slots__ = ("column",)
+
+    column: _COL_co
+
+    def __init__(
+        self, collection: ColumnCollection[Any, _COL_co], col: _COL_co
+    ):
+        self.column = col
+
+        # proxy_index being non-empty means it was initialized.
+        # so we need to update it
+        pi = collection._proxy_index
+        if pi:
+            for eps_col in col._expanded_proxy_set:
+                pi[eps_col].add(self)
+
+    def get_expanded_proxy_set(self):
+        return self.column._expanded_proxy_set
+
+    def dispose(self, collection):
+        pi = collection._proxy_index
+        if not pi:
+            return
+        for col in self.column._expanded_proxy_set:
+            colset = pi.get(col, None)
+            if colset:
+                colset.discard(self)
+            if colset is not None and not colset:
+                del pi[col]
+
+    def embedded(
+        self,
+        target_set: Union[
+            Set[ColumnElement[Any]], FrozenSet[ColumnElement[Any]]
+        ],
+    ) -> bool:
+        expanded_proxy_set = self.column._expanded_proxy_set
+        for t in target_set.difference(expanded_proxy_set):
+            if not expanded_proxy_set.intersection(_expand_cloned([t])):
+                return False
+        return True
+
+
+class ColumnCollection(Generic[_COLKEY, _COL_co]):
+    """Collection of :class:`_expression.ColumnElement` instances,
+    typically for
+    :class:`_sql.FromClause` objects.
+
+    The :class:`_sql.ColumnCollection` object is most commonly available
+    as the :attr:`_schema.Table.c` or :attr:`_schema.Table.columns` collection
+    on the :class:`_schema.Table` object, introduced at
+    :ref:`metadata_tables_and_columns`.
+
+    The :class:`_expression.ColumnCollection` has both mapping- and sequence-
+    like behaviors. A :class:`_expression.ColumnCollection` usually stores
+    :class:`_schema.Column` objects, which are then accessible both via mapping
+    style access as well as attribute access style.
+
+    To access :class:`_schema.Column` objects using ordinary attribute-style
+    access, specify the name like any other object attribute, such as below
+    a column named ``employee_name`` is accessed::
+
+        >>> employee_table.c.employee_name
+
+    To access columns that have names with special characters or spaces,
+    index-style access is used, such as below which illustrates a column named
+    ``employee ' payment`` is accessed::
+
+        >>> employee_table.c["employee ' payment"]
+
+    As the :class:`_sql.ColumnCollection` object provides a Python dictionary
+    interface, common dictionary method names like
+    :meth:`_sql.ColumnCollection.keys`, :meth:`_sql.ColumnCollection.values`,
+    and :meth:`_sql.ColumnCollection.items` are available, which means that
+    database columns that are keyed under these names also need to use indexed
+    access::
+
+        >>> employee_table.c["values"]
+
+
+    The name for which a :class:`_schema.Column` would be present is normally
+    that of the :paramref:`_schema.Column.key` parameter.  In some contexts,
+    such as a :class:`_sql.Select` object that uses a label style set
+    using the :meth:`_sql.Select.set_label_style` method, a column of a certain
+    key may instead be represented under a particular label name such
+    as ``tablename_columnname``::
+
+        >>> from sqlalchemy import select, column, table
+        >>> from sqlalchemy import LABEL_STYLE_TABLENAME_PLUS_COL
+        >>> t = table("t", column("c"))
+        >>> stmt = select(t).set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+        >>> subq = stmt.subquery()
+        >>> subq.c.t_c
+        <sqlalchemy.sql.elements.ColumnClause at 0x7f59dcf04fa0; t_c>
+
+    :class:`.ColumnCollection` also indexes the columns in order and allows
+    them to be accessible by their integer position::
+
+        >>> cc[0]
+        Column('x', Integer(), table=None)
+        >>> cc[1]
+        Column('y', Integer(), table=None)
+
+    .. versionadded:: 1.4 :class:`_expression.ColumnCollection`
+       allows integer-based
+       index access to the collection.
+
+    Iterating the collection yields the column expressions in order::
+
+        >>> list(cc)
+        [Column('x', Integer(), table=None),
+         Column('y', Integer(), table=None)]
+
+    The base :class:`_expression.ColumnCollection` object can store
+    duplicates, which can
+    mean either two columns with the same key, in which case the column
+    returned by key  access is **arbitrary**::
+
+        >>> x1, x2 = Column("x", Integer), Column("x", Integer)
+        >>> cc = ColumnCollection(columns=[(x1.name, x1), (x2.name, x2)])
+        >>> list(cc)
+        [Column('x', Integer(), table=None),
+         Column('x', Integer(), table=None)]
+        >>> cc["x"] is x1
+        False
+        >>> cc["x"] is x2
+        True
+
+    Or it can also mean the same column multiple times.   These cases are
+    supported as :class:`_expression.ColumnCollection`
+    is used to represent the columns in
+    a SELECT statement which may include duplicates.
+
+    A special subclass :class:`.DedupeColumnCollection` exists which instead
+    maintains SQLAlchemy's older behavior of not allowing duplicates; this
+    collection is used for schema level objects like :class:`_schema.Table`
+    and
+    :class:`.PrimaryKeyConstraint` where this deduping is helpful.  The
+    :class:`.DedupeColumnCollection` class also has additional mutation methods
+    as the schema constructs have more use cases that require removal and
+    replacement of columns.
+
+    .. versionchanged:: 1.4 :class:`_expression.ColumnCollection`
+       now stores duplicate
+       column keys as well as the same column in multiple positions.  The
+       :class:`.DedupeColumnCollection` class is added to maintain the
+       former behavior in those cases where deduplication as well as
+       additional replace/remove operations are needed.
+
+
+    """
+
+    __slots__ = "_collection", "_index", "_colset", "_proxy_index"
+
+    _collection: List[Tuple[_COLKEY, _COL_co, _ColumnMetrics[_COL_co]]]
+    _index: Dict[Union[None, str, int], Tuple[_COLKEY, _COL_co]]
+    _proxy_index: Dict[ColumnElement[Any], Set[_ColumnMetrics[_COL_co]]]
+    _colset: Set[_COL_co]
+
+    def __init__(
+        self, columns: Optional[Iterable[Tuple[_COLKEY, _COL_co]]] = None
+    ):
+        object.__setattr__(self, "_colset", set())
+        object.__setattr__(self, "_index", {})
+        object.__setattr__(
+            self, "_proxy_index", collections.defaultdict(util.OrderedSet)
+        )
+        object.__setattr__(self, "_collection", [])
+        if columns:
+            self._initial_populate(columns)
+
+    @util.preload_module("sqlalchemy.sql.elements")
+    def __clause_element__(self) -> ClauseList:
+        elements = util.preloaded.sql_elements
+
+        return elements.ClauseList(
+            _literal_as_text_role=roles.ColumnsClauseRole,
+            group=False,
+            *self._all_columns,
+        )
+
+    def _initial_populate(
+        self, iter_: Iterable[Tuple[_COLKEY, _COL_co]]
+    ) -> None:
+        self._populate_separate_keys(iter_)
+
+    @property
+    def _all_columns(self) -> List[_COL_co]:
+        return [col for (_, col, _) in self._collection]
+
+    def keys(self) -> List[_COLKEY]:
+        """Return a sequence of string key names for all columns in this
+        collection."""
+        return [k for (k, _, _) in self._collection]
+
+    def values(self) -> List[_COL_co]:
+        """Return a sequence of :class:`_sql.ColumnClause` or
+        :class:`_schema.Column` objects for all columns in this
+        collection."""
+        return [col for (_, col, _) in self._collection]
+
+    def items(self) -> List[Tuple[_COLKEY, _COL_co]]:
+        """Return a sequence of (key, column) tuples for all columns in this
+        collection each consisting of a string key name and a
+        :class:`_sql.ColumnClause` or
+        :class:`_schema.Column` object.
+        """
+
+        return [(k, col) for (k, col, _) in self._collection]
+
+    def __bool__(self) -> bool:
+        return bool(self._collection)
+
+    def __len__(self) -> int:
+        return len(self._collection)
+
+    def __iter__(self) -> Iterator[_COL_co]:
+        # turn to a list first to maintain over a course of changes
+        return iter([col for _, col, _ in self._collection])
+
+    @overload
+    def __getitem__(self, key: Union[str, int]) -> _COL_co: ...
+
+    @overload
+    def __getitem__(
+        self, key: Tuple[Union[str, int], ...]
+    ) -> ReadOnlyColumnCollection[_COLKEY, _COL_co]: ...
+
+    @overload
+    def __getitem__(
+        self, key: slice
+    ) -> ReadOnlyColumnCollection[_COLKEY, _COL_co]: ...
+
+    def __getitem__(
+        self, key: Union[str, int, slice, Tuple[Union[str, int], ...]]
+    ) -> Union[ReadOnlyColumnCollection[_COLKEY, _COL_co], _COL_co]:
+        try:
+            if isinstance(key, (tuple, slice)):
+                if isinstance(key, slice):
+                    cols = (
+                        (sub_key, col)
+                        for (sub_key, col, _) in self._collection[key]
+                    )
+                else:
+                    cols = (self._index[sub_key] for sub_key in key)
+
+                return ColumnCollection(cols).as_readonly()
+            else:
+                return self._index[key][1]
+        except KeyError as err:
+            if isinstance(err.args[0], int):
+                raise IndexError(err.args[0]) from err
+            else:
+                raise
+
+    def __getattr__(self, key: str) -> _COL_co:
+        try:
+            return self._index[key][1]
+        except KeyError as err:
+            raise AttributeError(key) from err
+
+    def __contains__(self, key: str) -> bool:
+        if key not in self._index:
+            if not isinstance(key, str):
+                raise exc.ArgumentError(
+                    "__contains__ requires a string argument"
+                )
+            return False
+        else:
+            return True
+
+    def compare(self, other: ColumnCollection[Any, Any]) -> bool:
+        """Compare this :class:`_expression.ColumnCollection` to another
+        based on the names of the keys"""
+
+        for l, r in zip_longest(self, other):
+            if l is not r:
+                return False
+        else:
+            return True
+
+    def __eq__(self, other: Any) -> bool:
+        return self.compare(other)
+
+    @overload
+    def get(self, key: str, default: None = None) -> Optional[_COL_co]: ...
+
+    @overload
+    def get(self, key: str, default: _COL) -> Union[_COL_co, _COL]: ...
+
+    def get(
+        self, key: str, default: Optional[_COL] = None
+    ) -> Optional[Union[_COL_co, _COL]]:
+        """Get a :class:`_sql.ColumnClause` or :class:`_schema.Column` object
+        based on a string key name from this
+        :class:`_expression.ColumnCollection`."""
+
+        if key in self._index:
+            return self._index[key][1]
+        else:
+            return default
+
+    def __str__(self) -> str:
+        return "%s(%s)" % (
+            self.__class__.__name__,
+            ", ".join(str(c) for c in self),
+        )
+
+    def __setitem__(self, key: str, value: Any) -> NoReturn:
+        raise NotImplementedError()
+
+    def __delitem__(self, key: str) -> NoReturn:
+        raise NotImplementedError()
+
+    def __setattr__(self, key: str, obj: Any) -> NoReturn:
+        raise NotImplementedError()
+
+    def clear(self) -> NoReturn:
+        """Dictionary clear() is not implemented for
+        :class:`_sql.ColumnCollection`."""
+        raise NotImplementedError()
+
+    def remove(self, column: Any) -> None:
+        raise NotImplementedError()
+
+    def update(self, iter_: Any) -> NoReturn:
+        """Dictionary update() is not implemented for
+        :class:`_sql.ColumnCollection`."""
+        raise NotImplementedError()
+
+    # https://github.com/python/mypy/issues/4266
+    __hash__ = None  # type: ignore
+
+    def _populate_separate_keys(
+        self, iter_: Iterable[Tuple[_COLKEY, _COL_co]]
+    ) -> None:
+        """populate from an iterator of (key, column)"""
+
+        self._collection[:] = collection = [
+            (k, c, _ColumnMetrics(self, c)) for k, c in iter_
+        ]
+        self._colset.update(c._deannotate() for _, c, _ in collection)
+        self._index.update(
+            {idx: (k, c) for idx, (k, c, _) in enumerate(collection)}
+        )
+        self._index.update({k: (k, col) for k, col, _ in reversed(collection)})
+
+    def add(
+        self, column: ColumnElement[Any], key: Optional[_COLKEY] = None
+    ) -> None:
+        """Add a column to this :class:`_sql.ColumnCollection`.
+
+        .. note::
+
+            This method is **not normally used by user-facing code**, as the
+            :class:`_sql.ColumnCollection` is usually part of an existing
+            object such as a :class:`_schema.Table`. To add a
+            :class:`_schema.Column` to an existing :class:`_schema.Table`
+            object, use the :meth:`_schema.Table.append_column` method.
+
+        """
+        colkey: _COLKEY
+
+        if key is None:
+            colkey = column.key  # type: ignore
+        else:
+            colkey = key
+
+        l = len(self._collection)
+
+        # don't really know how this part is supposed to work w/ the
+        # covariant thing
+
+        _column = cast(_COL_co, column)
+
+        self._collection.append(
+            (colkey, _column, _ColumnMetrics(self, _column))
+        )
+        self._colset.add(_column._deannotate())
+        self._index[l] = (colkey, _column)
+        if colkey not in self._index:
+            self._index[colkey] = (colkey, _column)
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {
+            "_collection": [(k, c) for k, c, _ in self._collection],
+            "_index": self._index,
+        }
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        object.__setattr__(self, "_index", state["_index"])
+        object.__setattr__(
+            self, "_proxy_index", collections.defaultdict(util.OrderedSet)
+        )
+        object.__setattr__(
+            self,
+            "_collection",
+            [
+                (k, c, _ColumnMetrics(self, c))
+                for (k, c) in state["_collection"]
+            ],
+        )
+        object.__setattr__(
+            self, "_colset", {col for k, col, _ in self._collection}
+        )
+
+    def contains_column(self, col: ColumnElement[Any]) -> bool:
+        """Checks if a column object exists in this collection"""
+        if col not in self._colset:
+            if isinstance(col, str):
+                raise exc.ArgumentError(
+                    "contains_column cannot be used with string arguments. "
+                    "Use ``col_name in table.c`` instead."
+                )
+            return False
+        else:
+            return True
+
+    def as_readonly(self) -> ReadOnlyColumnCollection[_COLKEY, _COL_co]:
+        """Return a "read only" form of this
+        :class:`_sql.ColumnCollection`."""
+
+        return ReadOnlyColumnCollection(self)
+
+    def _init_proxy_index(self):
+        """populate the "proxy index", if empty.
+
+        proxy index is added in 2.0 to provide more efficient operation
+        for the corresponding_column() method.
+
+        For reasons of both time to construct new .c collections as well as
+        memory conservation for large numbers of large .c collections, the
+        proxy_index is only filled if corresponding_column() is called. once
+        filled it stays that way, and new _ColumnMetrics objects created after
+        that point will populate it with new data. Note this case would be
+        unusual, if not nonexistent, as it means a .c collection is being
+        mutated after corresponding_column() were used, however it is tested in
+        test/base/test_utils.py.
+
+        """
+        pi = self._proxy_index
+        if pi:
+            return
+
+        for _, _, metrics in self._collection:
+            eps = metrics.column._expanded_proxy_set
+
+            for eps_col in eps:
+                pi[eps_col].add(metrics)
+
+    def corresponding_column(
+        self, column: _COL, require_embedded: bool = False
+    ) -> Optional[Union[_COL, _COL_co]]:
+        """Given a :class:`_expression.ColumnElement`, return the exported
+        :class:`_expression.ColumnElement` object from this
+        :class:`_expression.ColumnCollection`
+        which corresponds to that original :class:`_expression.ColumnElement`
+        via a common
+        ancestor column.
+
+        :param column: the target :class:`_expression.ColumnElement`
+                      to be matched.
+
+        :param require_embedded: only return corresponding columns for
+         the given :class:`_expression.ColumnElement`, if the given
+         :class:`_expression.ColumnElement`
+         is actually present within a sub-element
+         of this :class:`_expression.Selectable`.
+         Normally the column will match if
+         it merely shares a common ancestor with one of the exported
+         columns of this :class:`_expression.Selectable`.
+
+        .. seealso::
+
+            :meth:`_expression.Selectable.corresponding_column`
+            - invokes this method
+            against the collection returned by
+            :attr:`_expression.Selectable.exported_columns`.
+
+        .. versionchanged:: 1.4 the implementation for ``corresponding_column``
+           was moved onto the :class:`_expression.ColumnCollection` itself.
+
+        """
+        # TODO: cython candidate
+
+        # don't dig around if the column is locally present
+        if column in self._colset:
+            return column
+
+        selected_intersection, selected_metrics = None, None
+        target_set = column.proxy_set
+
+        pi = self._proxy_index
+        if not pi:
+            self._init_proxy_index()
+
+        for current_metrics in (
+            mm for ts in target_set if ts in pi for mm in pi[ts]
+        ):
+            if not require_embedded or current_metrics.embedded(target_set):
+                if selected_metrics is None:
+                    # no corresponding column yet, pick this one.
+                    selected_metrics = current_metrics
+                    continue
+
+                current_intersection = target_set.intersection(
+                    current_metrics.column._expanded_proxy_set
+                )
+                if selected_intersection is None:
+                    selected_intersection = target_set.intersection(
+                        selected_metrics.column._expanded_proxy_set
+                    )
+
+                if len(current_intersection) > len(selected_intersection):
+                    # 'current' has a larger field of correspondence than
+                    # 'selected'. i.e. selectable.c.a1_x->a1.c.x->table.c.x
+                    # matches a1.c.x->table.c.x better than
+                    # selectable.c.x->table.c.x does.
+
+                    selected_metrics = current_metrics
+                    selected_intersection = current_intersection
+                elif current_intersection == selected_intersection:
+                    # they have the same field of correspondence. see
+                    # which proxy_set has fewer columns in it, which
+                    # indicates a closer relationship with the root
+                    # column. Also take into account the "weight"
+                    # attribute which CompoundSelect() uses to give
+                    # higher precedence to columns based on vertical
+                    # position in the compound statement, and discard
+                    # columns that have no reference to the target
+                    # column (also occurs with CompoundSelect)
+
+                    selected_col_distance = sum(
+                        [
+                            sc._annotations.get("weight", 1)
+                            for sc in (
+                                selected_metrics.column._uncached_proxy_list()
+                            )
+                            if sc.shares_lineage(column)
+                        ],
+                    )
+                    current_col_distance = sum(
+                        [
+                            sc._annotations.get("weight", 1)
+                            for sc in (
+                                current_metrics.column._uncached_proxy_list()
+                            )
+                            if sc.shares_lineage(column)
+                        ],
+                    )
+                    if current_col_distance < selected_col_distance:
+                        selected_metrics = current_metrics
+                        selected_intersection = current_intersection
+
+        return selected_metrics.column if selected_metrics else None
+
+
+_NAMEDCOL = TypeVar("_NAMEDCOL", bound="NamedColumn[Any]")
+
+
+class DedupeColumnCollection(ColumnCollection[str, _NAMEDCOL]):
+    """A :class:`_expression.ColumnCollection`
+    that maintains deduplicating behavior.
+
+    This is useful by schema level objects such as :class:`_schema.Table` and
+    :class:`.PrimaryKeyConstraint`.    The collection includes more
+    sophisticated mutator methods as well to suit schema objects which
+    require mutable column collections.
+
+    .. versionadded:: 1.4
+
+    """
+
+    def add(  # type: ignore[override]
+        self, column: _NAMEDCOL, key: Optional[str] = None
+    ) -> None:
+        if key is not None and column.key != key:
+            raise exc.ArgumentError(
+                "DedupeColumnCollection requires columns be under "
+                "the same key as their .key"
+            )
+        key = column.key
+
+        if key is None:
+            raise exc.ArgumentError(
+                "Can't add unnamed column to column collection"
+            )
+
+        if key in self._index:
+            existing = self._index[key][1]
+
+            if existing is column:
+                return
+
+            self.replace(column)
+
+            # pop out memoized proxy_set as this
+            # operation may very well be occurring
+            # in a _make_proxy operation
+            util.memoized_property.reset(column, "proxy_set")
+        else:
+            self._append_new_column(key, column)
+
+    def _append_new_column(self, key: str, named_column: _NAMEDCOL) -> None:
+        l = len(self._collection)
+        self._collection.append(
+            (key, named_column, _ColumnMetrics(self, named_column))
+        )
+        self._colset.add(named_column._deannotate())
+        self._index[l] = (key, named_column)
+        self._index[key] = (key, named_column)
+
+    def _populate_separate_keys(
+        self, iter_: Iterable[Tuple[str, _NAMEDCOL]]
+    ) -> None:
+        """populate from an iterator of (key, column)"""
+        cols = list(iter_)
+
+        replace_col = []
+        for k, col in cols:
+            if col.key != k:
+                raise exc.ArgumentError(
+                    "DedupeColumnCollection requires columns be under "
+                    "the same key as their .key"
+                )
+            if col.name in self._index and col.key != col.name:
+                replace_col.append(col)
+            elif col.key in self._index:
+                replace_col.append(col)
+            else:
+                self._index[k] = (k, col)
+                self._collection.append((k, col, _ColumnMetrics(self, col)))
+        self._colset.update(c._deannotate() for (k, c, _) in self._collection)
+
+        self._index.update(
+            (idx, (k, c)) for idx, (k, c, _) in enumerate(self._collection)
+        )
+        for col in replace_col:
+            self.replace(col)
+
+    def extend(self, iter_: Iterable[_NAMEDCOL]) -> None:
+        self._populate_separate_keys((col.key, col) for col in iter_)
+
+    def remove(self, column: _NAMEDCOL) -> None:
+        if column not in self._colset:
+            raise ValueError(
+                "Can't remove column %r; column is not in this collection"
+                % column
+            )
+        del self._index[column.key]
+        self._colset.remove(column)
+        self._collection[:] = [
+            (k, c, metrics)
+            for (k, c, metrics) in self._collection
+            if c is not column
+        ]
+        for metrics in self._proxy_index.get(column, ()):
+            metrics.dispose(self)
+
+        self._index.update(
+            {idx: (k, col) for idx, (k, col, _) in enumerate(self._collection)}
+        )
+        # delete higher index
+        del self._index[len(self._collection)]
+
+    def replace(
+        self,
+        column: _NAMEDCOL,
+        extra_remove: Optional[Iterable[_NAMEDCOL]] = None,
+    ) -> None:
+        """add the given column to this collection, removing unaliased
+        versions of this column  as well as existing columns with the
+        same key.
+
+        e.g.::
+
+            t = Table("sometable", metadata, Column("col1", Integer))
+            t.columns.replace(Column("col1", Integer, key="columnone"))
+
+        will remove the original 'col1' from the collection, and add
+        the new column under the name 'columnname'.
+
+        Used by schema.Column to override columns during table reflection.
+
+        """
+
+        if extra_remove:
+            remove_col = set(extra_remove)
+        else:
+            remove_col = set()
+        # remove up to two columns based on matches of name as well as key
+        if column.name in self._index and column.key != column.name:
+            other = self._index[column.name][1]
+            if other.name == other.key:
+                remove_col.add(other)
+
+        if column.key in self._index:
+            remove_col.add(self._index[column.key][1])
+
+        if not remove_col:
+            self._append_new_column(column.key, column)
+            return
+        new_cols: List[Tuple[str, _NAMEDCOL, _ColumnMetrics[_NAMEDCOL]]] = []
+        replaced = False
+        for k, col, metrics in self._collection:
+            if col in remove_col:
+                if not replaced:
+                    replaced = True
+                    new_cols.append(
+                        (column.key, column, _ColumnMetrics(self, column))
+                    )
+            else:
+                new_cols.append((k, col, metrics))
+
+        if remove_col:
+            self._colset.difference_update(remove_col)
+
+            for rc in remove_col:
+                for metrics in self._proxy_index.get(rc, ()):
+                    metrics.dispose(self)
+
+        if not replaced:
+            new_cols.append((column.key, column, _ColumnMetrics(self, column)))
+
+        self._colset.add(column._deannotate())
+        self._collection[:] = new_cols
+
+        self._index.clear()
+
+        self._index.update(
+            {idx: (k, col) for idx, (k, col, _) in enumerate(self._collection)}
+        )
+        self._index.update({k: (k, col) for (k, col, _) in self._collection})
+
+
+class ReadOnlyColumnCollection(
+    util.ReadOnlyContainer, ColumnCollection[_COLKEY, _COL_co]
+):
+    __slots__ = ("_parent",)
+
+    def __init__(self, collection):
+        object.__setattr__(self, "_parent", collection)
+        object.__setattr__(self, "_colset", collection._colset)
+        object.__setattr__(self, "_index", collection._index)
+        object.__setattr__(self, "_collection", collection._collection)
+        object.__setattr__(self, "_proxy_index", collection._proxy_index)
+
+    def __getstate__(self):
+        return {"_parent": self._parent}
+
+    def __setstate__(self, state):
+        parent = state["_parent"]
+        self.__init__(parent)  # type: ignore
+
+    def add(self, column: Any, key: Any = ...) -> Any:
+        self._readonly()
+
+    def extend(self, elements: Any) -> NoReturn:
+        self._readonly()
+
+    def remove(self, item: Any) -> NoReturn:
+        self._readonly()
+
+
+class ColumnSet(util.OrderedSet["ColumnClause[Any]"]):
+    def contains_column(self, col):
+        return col in self
+
+    def extend(self, cols):
+        for col in cols:
+            self.add(col)
+
+    def __eq__(self, other):
+        l = []
+        for c in other:
+            for local in self:
+                if c.shares_lineage(local):
+                    l.append(c == local)
+        return elements.and_(*l)
+
+    def __hash__(self):  # type: ignore[override]
+        return hash(tuple(x for x in self))
+
+
+def _entity_namespace(
+    entity: Union[_HasEntityNamespace, ExternallyTraversible]
+) -> _EntityNamespace:
+    """Return the nearest .entity_namespace for the given entity.
+
+    If not immediately available, does an iterate to find a sub-element
+    that has one, if any.
+
+    """
+    try:
+        return cast(_HasEntityNamespace, entity).entity_namespace
+    except AttributeError:
+        for elem in visitors.iterate(cast(ExternallyTraversible, entity)):
+            if _is_has_entity_namespace(elem):
+                return elem.entity_namespace
+        else:
+            raise
+
+
+def _entity_namespace_key(
+    entity: Union[_HasEntityNamespace, ExternallyTraversible],
+    key: str,
+    default: Union[SQLCoreOperations[Any], _NoArg] = NO_ARG,
+) -> SQLCoreOperations[Any]:
+    """Return an entry from an entity_namespace.
+
+
+    Raises :class:`_exc.InvalidRequestError` rather than attribute error
+    on not found.
+
+    """
+
+    try:
+        ns = _entity_namespace(entity)
+        if default is not NO_ARG:
+            return getattr(ns, key, default)
+        else:
+            return getattr(ns, key)  # type: ignore
+    except AttributeError as err:
+        raise exc.InvalidRequestError(
+            'Entity namespace for "%s" has no property "%s"' % (entity, key)
+        ) from err
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/cache_key.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/cache_key.py
new file mode 100644
index 00000000..1f562f2e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/cache_key.py
@@ -0,0 +1,1057 @@
+# sql/cache_key.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+import enum
+from itertools import zip_longest
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import MutableMapping
+from typing import NamedTuple
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+from typing import Union
+
+from .visitors import anon_map
+from .visitors import HasTraversalDispatch
+from .visitors import HasTraverseInternals
+from .visitors import InternalTraversal
+from .visitors import prefix_anon_map
+from .. import util
+from ..inspection import inspect
+from ..util import HasMemoized
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from .elements import BindParameter
+    from .elements import ClauseElement
+    from .elements import ColumnElement
+    from .visitors import _TraverseInternalsType
+    from ..engine.interfaces import _CoreSingleExecuteParams
+
+
+class _CacheKeyTraversalDispatchType(Protocol):
+    def __call__(
+        s, self: HasCacheKey, visitor: _CacheKeyTraversal
+    ) -> _CacheKeyTraversalDispatchTypeReturn: ...
+
+
+class CacheConst(enum.Enum):
+    NO_CACHE = 0
+
+
+NO_CACHE = CacheConst.NO_CACHE
+
+
+_CacheKeyTraversalType = Union[
+    "_TraverseInternalsType", Literal[CacheConst.NO_CACHE], Literal[None]
+]
+
+
+class CacheTraverseTarget(enum.Enum):
+    CACHE_IN_PLACE = 0
+    CALL_GEN_CACHE_KEY = 1
+    STATIC_CACHE_KEY = 2
+    PROPAGATE_ATTRS = 3
+    ANON_NAME = 4
+
+
+(
+    CACHE_IN_PLACE,
+    CALL_GEN_CACHE_KEY,
+    STATIC_CACHE_KEY,
+    PROPAGATE_ATTRS,
+    ANON_NAME,
+) = tuple(CacheTraverseTarget)
+
+_CacheKeyTraversalDispatchTypeReturn = Sequence[
+    Tuple[
+        str,
+        Any,
+        Union[
+            Callable[..., Tuple[Any, ...]],
+            CacheTraverseTarget,
+            InternalTraversal,
+        ],
+    ]
+]
+
+
+class HasCacheKey:
+    """Mixin for objects which can produce a cache key.
+
+    This class is usually in a hierarchy that starts with the
+    :class:`.HasTraverseInternals` base, but this is optional.  Currently,
+    the class should be able to work on its own without including
+    :class:`.HasTraverseInternals`.
+
+    .. seealso::
+
+        :class:`.CacheKey`
+
+        :ref:`sql_caching`
+
+    """
+
+    __slots__ = ()
+
+    _cache_key_traversal: _CacheKeyTraversalType = NO_CACHE
+
+    _is_has_cache_key = True
+
+    _hierarchy_supports_caching = True
+    """private attribute which may be set to False to prevent the
+    inherit_cache warning from being emitted for a hierarchy of subclasses.
+
+    Currently applies to the :class:`.ExecutableDDLElement` hierarchy which
+    does not implement caching.
+
+    """
+
+    inherit_cache: Optional[bool] = None
+    """Indicate if this :class:`.HasCacheKey` instance should make use of the
+    cache key generation scheme used by its immediate superclass.
+
+    The attribute defaults to ``None``, which indicates that a construct has
+    not yet taken into account whether or not its appropriate for it to
+    participate in caching; this is functionally equivalent to setting the
+    value to ``False``, except that a warning is also emitted.
+
+    This flag can be set to ``True`` on a particular class, if the SQL that
+    corresponds to the object does not change based on attributes which
+    are local to this class, and not its superclass.
+
+    .. seealso::
+
+        :ref:`compilerext_caching` - General guideslines for setting the
+        :attr:`.HasCacheKey.inherit_cache` attribute for third-party or user
+        defined SQL constructs.
+
+    """
+
+    __slots__ = ()
+
+    _generated_cache_key_traversal: Any
+
+    @classmethod
+    def _generate_cache_attrs(
+        cls,
+    ) -> Union[_CacheKeyTraversalDispatchType, Literal[CacheConst.NO_CACHE]]:
+        """generate cache key dispatcher for a new class.
+
+        This sets the _generated_cache_key_traversal attribute once called
+        so should only be called once per class.
+
+        """
+        inherit_cache = cls.__dict__.get("inherit_cache", None)
+        inherit = bool(inherit_cache)
+
+        if inherit:
+            _cache_key_traversal = getattr(cls, "_cache_key_traversal", None)
+            if _cache_key_traversal is None:
+                try:
+                    assert issubclass(cls, HasTraverseInternals)
+                    _cache_key_traversal = cls._traverse_internals
+                except AttributeError:
+                    cls._generated_cache_key_traversal = NO_CACHE
+                    return NO_CACHE
+
+            assert _cache_key_traversal is not NO_CACHE, (
+                f"class {cls} has _cache_key_traversal=NO_CACHE, "
+                "which conflicts with inherit_cache=True"
+            )
+
+            # TODO: wouldn't we instead get this from our superclass?
+            # also, our superclass may not have this yet, but in any case,
+            # we'd generate for the superclass that has it.   this is a little
+            # more complicated, so for the moment this is a little less
+            # efficient on startup but simpler.
+            return _cache_key_traversal_visitor.generate_dispatch(
+                cls,
+                _cache_key_traversal,
+                "_generated_cache_key_traversal",
+            )
+        else:
+            _cache_key_traversal = cls.__dict__.get(
+                "_cache_key_traversal", None
+            )
+            if _cache_key_traversal is None:
+                _cache_key_traversal = cls.__dict__.get(
+                    "_traverse_internals", None
+                )
+                if _cache_key_traversal is None:
+                    cls._generated_cache_key_traversal = NO_CACHE
+                    if (
+                        inherit_cache is None
+                        and cls._hierarchy_supports_caching
+                    ):
+                        util.warn(
+                            "Class %s will not make use of SQL compilation "
+                            "caching as it does not set the 'inherit_cache' "
+                            "attribute to ``True``.  This can have "
+                            "significant performance implications including "
+                            "some performance degradations in comparison to "
+                            "prior SQLAlchemy versions.  Set this attribute "
+                            "to True if this object can make use of the cache "
+                            "key generated by the superclass.  Alternatively, "
+                            "this attribute may be set to False which will "
+                            "disable this warning." % (cls.__name__),
+                            code="cprf",
+                        )
+                    return NO_CACHE
+
+            return _cache_key_traversal_visitor.generate_dispatch(
+                cls,
+                _cache_key_traversal,
+                "_generated_cache_key_traversal",
+            )
+
+    @util.preload_module("sqlalchemy.sql.elements")
+    def _gen_cache_key(
+        self, anon_map: anon_map, bindparams: List[BindParameter[Any]]
+    ) -> Optional[Tuple[Any, ...]]:
+        """return an optional cache key.
+
+        The cache key is a tuple which can contain any series of
+        objects that are hashable and also identifies
+        this object uniquely within the presence of a larger SQL expression
+        or statement, for the purposes of caching the resulting query.
+
+        The cache key should be based on the SQL compiled structure that would
+        ultimately be produced.   That is, two structures that are composed in
+        exactly the same way should produce the same cache key; any difference
+        in the structures that would affect the SQL string or the type handlers
+        should result in a different cache key.
+
+        If a structure cannot produce a useful cache key, the NO_CACHE
+        symbol should be added to the anon_map and the method should
+        return None.
+
+        """
+
+        cls = self.__class__
+
+        id_, found = anon_map.get_anon(self)
+        if found:
+            return (id_, cls)
+
+        dispatcher: Union[
+            Literal[CacheConst.NO_CACHE],
+            _CacheKeyTraversalDispatchType,
+        ]
+
+        try:
+            dispatcher = cls.__dict__["_generated_cache_key_traversal"]
+        except KeyError:
+            # traversals.py -> _preconfigure_traversals()
+            # may be used to run these ahead of time, but
+            # is not enabled right now.
+            # this block will generate any remaining dispatchers.
+            dispatcher = cls._generate_cache_attrs()
+
+        if dispatcher is NO_CACHE:
+            anon_map[NO_CACHE] = True
+            return None
+
+        result: Tuple[Any, ...] = (id_, cls)
+
+        # inline of _cache_key_traversal_visitor.run_generated_dispatch()
+
+        for attrname, obj, meth in dispatcher(
+            self, _cache_key_traversal_visitor
+        ):
+            if obj is not None:
+                # TODO: see if C code can help here as Python lacks an
+                # efficient switch construct
+
+                if meth is STATIC_CACHE_KEY:
+                    sck = obj._static_cache_key
+                    if sck is NO_CACHE:
+                        anon_map[NO_CACHE] = True
+                        return None
+                    result += (attrname, sck)
+                elif meth is ANON_NAME:
+                    elements = util.preloaded.sql_elements
+                    if isinstance(obj, elements._anonymous_label):
+                        obj = obj.apply_map(anon_map)  # type: ignore
+                    result += (attrname, obj)
+                elif meth is CALL_GEN_CACHE_KEY:
+                    result += (
+                        attrname,
+                        obj._gen_cache_key(anon_map, bindparams),
+                    )
+
+                # remaining cache functions are against
+                # Python tuples, dicts, lists, etc. so we can skip
+                # if they are empty
+                elif obj:
+                    if meth is CACHE_IN_PLACE:
+                        result += (attrname, obj)
+                    elif meth is PROPAGATE_ATTRS:
+                        result += (
+                            attrname,
+                            obj["compile_state_plugin"],
+                            (
+                                obj["plugin_subject"]._gen_cache_key(
+                                    anon_map, bindparams
+                                )
+                                if obj["plugin_subject"]
+                                else None
+                            ),
+                        )
+                    elif meth is InternalTraversal.dp_annotations_key:
+                        # obj is here is the _annotations dict.  Table uses
+                        # a memoized version of it.  however in other cases,
+                        # we generate it given anon_map as we may be from a
+                        # Join, Aliased, etc.
+                        # see #8790
+
+                        if self._gen_static_annotations_cache_key:  # type: ignore  # noqa: E501
+                            result += self._annotations_cache_key  # type: ignore  # noqa: E501
+                        else:
+                            result += self._gen_annotations_cache_key(anon_map)  # type: ignore  # noqa: E501
+
+                    elif (
+                        meth is InternalTraversal.dp_clauseelement_list
+                        or meth is InternalTraversal.dp_clauseelement_tuple
+                        or meth
+                        is InternalTraversal.dp_memoized_select_entities
+                    ):
+                        result += (
+                            attrname,
+                            tuple(
+                                [
+                                    elem._gen_cache_key(anon_map, bindparams)
+                                    for elem in obj
+                                ]
+                            ),
+                        )
+                    else:
+                        result += meth(  # type: ignore
+                            attrname, obj, self, anon_map, bindparams
+                        )
+        return result
+
+    def _generate_cache_key(self) -> Optional[CacheKey]:
+        """return a cache key.
+
+        The cache key is a tuple which can contain any series of
+        objects that are hashable and also identifies
+        this object uniquely within the presence of a larger SQL expression
+        or statement, for the purposes of caching the resulting query.
+
+        The cache key should be based on the SQL compiled structure that would
+        ultimately be produced.   That is, two structures that are composed in
+        exactly the same way should produce the same cache key; any difference
+        in the structures that would affect the SQL string or the type handlers
+        should result in a different cache key.
+
+        The cache key returned by this method is an instance of
+        :class:`.CacheKey`, which consists of a tuple representing the
+        cache key, as well as a list of :class:`.BindParameter` objects
+        which are extracted from the expression.   While two expressions
+        that produce identical cache key tuples will themselves generate
+        identical SQL strings, the list of :class:`.BindParameter` objects
+        indicates the bound values which may have different values in
+        each one; these bound parameters must be consulted in order to
+        execute the statement with the correct parameters.
+
+        a :class:`_expression.ClauseElement` structure that does not implement
+        a :meth:`._gen_cache_key` method and does not implement a
+        :attr:`.traverse_internals` attribute will not be cacheable; when
+        such an element is embedded into a larger structure, this method
+        will return None, indicating no cache key is available.
+
+        """
+
+        bindparams: List[BindParameter[Any]] = []
+
+        _anon_map = anon_map()
+        key = self._gen_cache_key(_anon_map, bindparams)
+        if NO_CACHE in _anon_map:
+            return None
+        else:
+            assert key is not None
+            return CacheKey(key, bindparams)
+
+    @classmethod
+    def _generate_cache_key_for_object(
+        cls, obj: HasCacheKey
+    ) -> Optional[CacheKey]:
+        bindparams: List[BindParameter[Any]] = []
+
+        _anon_map = anon_map()
+        key = obj._gen_cache_key(_anon_map, bindparams)
+        if NO_CACHE in _anon_map:
+            return None
+        else:
+            assert key is not None
+            return CacheKey(key, bindparams)
+
+
+class HasCacheKeyTraverse(HasTraverseInternals, HasCacheKey):
+    pass
+
+
+class MemoizedHasCacheKey(HasCacheKey, HasMemoized):
+    __slots__ = ()
+
+    @HasMemoized.memoized_instancemethod
+    def _generate_cache_key(self) -> Optional[CacheKey]:
+        return HasCacheKey._generate_cache_key(self)
+
+
+class SlotsMemoizedHasCacheKey(HasCacheKey, util.MemoizedSlots):
+    __slots__ = ()
+
+    def _memoized_method__generate_cache_key(self) -> Optional[CacheKey]:
+        return HasCacheKey._generate_cache_key(self)
+
+
+class CacheKey(NamedTuple):
+    """The key used to identify a SQL statement construct in the
+    SQL compilation cache.
+
+    .. seealso::
+
+        :ref:`sql_caching`
+
+    """
+
+    key: Tuple[Any, ...]
+    bindparams: Sequence[BindParameter[Any]]
+
+    # can't set __hash__ attribute because it interferes
+    # with namedtuple
+    # can't use "if not TYPE_CHECKING" because mypy rejects it
+    # inside of a NamedTuple
+    def __hash__(self) -> Optional[int]:  # type: ignore
+        """CacheKey itself is not hashable - hash the .key portion"""
+        return None
+
+    def to_offline_string(
+        self,
+        statement_cache: MutableMapping[Any, str],
+        statement: ClauseElement,
+        parameters: _CoreSingleExecuteParams,
+    ) -> str:
+        """Generate an "offline string" form of this :class:`.CacheKey`
+
+        The "offline string" is basically the string SQL for the
+        statement plus a repr of the bound parameter values in series.
+        Whereas the :class:`.CacheKey` object is dependent on in-memory
+        identities in order to work as a cache key, the "offline" version
+        is suitable for a cache that will work for other processes as well.
+
+        The given ``statement_cache`` is a dictionary-like object where the
+        string form of the statement itself will be cached.  This dictionary
+        should be in a longer lived scope in order to reduce the time spent
+        stringifying statements.
+
+
+        """
+        if self.key not in statement_cache:
+            statement_cache[self.key] = sql_str = str(statement)
+        else:
+            sql_str = statement_cache[self.key]
+
+        if not self.bindparams:
+            param_tuple = tuple(parameters[key] for key in sorted(parameters))
+        else:
+            param_tuple = tuple(
+                parameters.get(bindparam.key, bindparam.value)
+                for bindparam in self.bindparams
+            )
+
+        return repr((sql_str, param_tuple))
+
+    def __eq__(self, other: Any) -> bool:
+        return bool(self.key == other.key)
+
+    def __ne__(self, other: Any) -> bool:
+        return not (self.key == other.key)
+
+    @classmethod
+    def _diff_tuples(cls, left: CacheKey, right: CacheKey) -> str:
+        ck1 = CacheKey(left, [])
+        ck2 = CacheKey(right, [])
+        return ck1._diff(ck2)
+
+    def _whats_different(self, other: CacheKey) -> Iterator[str]:
+        k1 = self.key
+        k2 = other.key
+
+        stack: List[int] = []
+        pickup_index = 0
+        while True:
+            s1, s2 = k1, k2
+            for idx in stack:
+                s1 = s1[idx]
+                s2 = s2[idx]
+
+            for idx, (e1, e2) in enumerate(zip_longest(s1, s2)):
+                if idx < pickup_index:
+                    continue
+                if e1 != e2:
+                    if isinstance(e1, tuple) and isinstance(e2, tuple):
+                        stack.append(idx)
+                        break
+                    else:
+                        yield "key%s[%d]:  %s != %s" % (
+                            "".join("[%d]" % id_ for id_ in stack),
+                            idx,
+                            e1,
+                            e2,
+                        )
+            else:
+                pickup_index = stack.pop(-1)
+                break
+
+    def _diff(self, other: CacheKey) -> str:
+        return ", ".join(self._whats_different(other))
+
+    def __str__(self) -> str:
+        stack: List[Union[Tuple[Any, ...], HasCacheKey]] = [self.key]
+
+        output = []
+        sentinel = object()
+        indent = -1
+        while stack:
+            elem = stack.pop(0)
+            if elem is sentinel:
+                output.append((" " * (indent * 2)) + "),")
+                indent -= 1
+            elif isinstance(elem, tuple):
+                if not elem:
+                    output.append((" " * ((indent + 1) * 2)) + "()")
+                else:
+                    indent += 1
+                    stack = list(elem) + [sentinel] + stack
+                    output.append((" " * (indent * 2)) + "(")
+            else:
+                if isinstance(elem, HasCacheKey):
+                    repr_ = "<%s object at %s>" % (
+                        type(elem).__name__,
+                        hex(id(elem)),
+                    )
+                else:
+                    repr_ = repr(elem)
+                output.append((" " * (indent * 2)) + "  " + repr_ + ", ")
+
+        return "CacheKey(key=%s)" % ("\n".join(output),)
+
+    def _generate_param_dict(self) -> Dict[str, Any]:
+        """used for testing"""
+
+        _anon_map = prefix_anon_map()
+        return {b.key % _anon_map: b.effective_value for b in self.bindparams}
+
+    @util.preload_module("sqlalchemy.sql.elements")
+    def _apply_params_to_element(
+        self, original_cache_key: CacheKey, target_element: ColumnElement[Any]
+    ) -> ColumnElement[Any]:
+        if target_element._is_immutable or original_cache_key is self:
+            return target_element
+
+        elements = util.preloaded.sql_elements
+        return elements._OverrideBinds(
+            target_element, self.bindparams, original_cache_key.bindparams
+        )
+
+
+def _ad_hoc_cache_key_from_args(
+    tokens: Tuple[Any, ...],
+    traverse_args: Iterable[Tuple[str, InternalTraversal]],
+    args: Iterable[Any],
+) -> Tuple[Any, ...]:
+    """a quick cache key generator used by reflection.flexi_cache."""
+    bindparams: List[BindParameter[Any]] = []
+
+    _anon_map = anon_map()
+
+    tup = tokens
+
+    for (attrname, sym), arg in zip(traverse_args, args):
+        key = sym.name
+        visit_key = key.replace("dp_", "visit_")
+
+        if arg is None:
+            tup += (attrname, None)
+            continue
+
+        meth = getattr(_cache_key_traversal_visitor, visit_key)
+        if meth is CACHE_IN_PLACE:
+            tup += (attrname, arg)
+        elif meth in (
+            CALL_GEN_CACHE_KEY,
+            STATIC_CACHE_KEY,
+            ANON_NAME,
+            PROPAGATE_ATTRS,
+        ):
+            raise NotImplementedError(
+                f"Haven't implemented symbol {meth} for ad-hoc key from args"
+            )
+        else:
+            tup += meth(attrname, arg, None, _anon_map, bindparams)
+    return tup
+
+
+class _CacheKeyTraversal(HasTraversalDispatch):
+    # very common elements are inlined into the main _get_cache_key() method
+    # to produce a dramatic savings in Python function call overhead
+
+    visit_has_cache_key = visit_clauseelement = CALL_GEN_CACHE_KEY
+    visit_clauseelement_list = InternalTraversal.dp_clauseelement_list
+    visit_annotations_key = InternalTraversal.dp_annotations_key
+    visit_clauseelement_tuple = InternalTraversal.dp_clauseelement_tuple
+    visit_memoized_select_entities = (
+        InternalTraversal.dp_memoized_select_entities
+    )
+
+    visit_string = visit_boolean = visit_operator = visit_plain_obj = (
+        CACHE_IN_PLACE
+    )
+    visit_statement_hint_list = CACHE_IN_PLACE
+    visit_type = STATIC_CACHE_KEY
+    visit_anon_name = ANON_NAME
+
+    visit_propagate_attrs = PROPAGATE_ATTRS
+
+    def visit_with_context_options(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return tuple((fn.__code__, c_key) for fn, c_key in obj)
+
+    def visit_inspectable(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (attrname, inspect(obj)._gen_cache_key(anon_map, bindparams))
+
+    def visit_string_list(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return tuple(obj)
+
+    def visit_multi(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            (
+                obj._gen_cache_key(anon_map, bindparams)
+                if isinstance(obj, HasCacheKey)
+                else obj
+            ),
+        )
+
+    def visit_multi_list(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            tuple(
+                (
+                    elem._gen_cache_key(anon_map, bindparams)
+                    if isinstance(elem, HasCacheKey)
+                    else elem
+                )
+                for elem in obj
+            ),
+        )
+
+    def visit_has_cache_key_tuples(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+        return (
+            attrname,
+            tuple(
+                tuple(
+                    elem._gen_cache_key(anon_map, bindparams)
+                    for elem in tup_elem
+                )
+                for tup_elem in obj
+            ),
+        )
+
+    def visit_has_cache_key_list(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+        return (
+            attrname,
+            tuple(elem._gen_cache_key(anon_map, bindparams) for elem in obj),
+        )
+
+    def visit_executable_options(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+        return (
+            attrname,
+            tuple(
+                elem._gen_cache_key(anon_map, bindparams)
+                for elem in obj
+                if elem._is_has_cache_key
+            ),
+        )
+
+    def visit_inspectable_list(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return self.visit_has_cache_key_list(
+            attrname, [inspect(o) for o in obj], parent, anon_map, bindparams
+        )
+
+    def visit_clauseelement_tuples(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return self.visit_has_cache_key_tuples(
+            attrname, obj, parent, anon_map, bindparams
+        )
+
+    def visit_fromclause_ordered_set(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+        return (
+            attrname,
+            tuple([elem._gen_cache_key(anon_map, bindparams) for elem in obj]),
+        )
+
+    def visit_clauseelement_unordered_set(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+        cache_keys = [
+            elem._gen_cache_key(anon_map, bindparams) for elem in obj
+        ]
+        return (
+            attrname,
+            tuple(
+                sorted(cache_keys)
+            ),  # cache keys all start with (id_, class)
+        )
+
+    def visit_named_ddl_element(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (attrname, obj.name)
+
+    def visit_prefix_sequence(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+
+        return (
+            attrname,
+            tuple(
+                [
+                    (clause._gen_cache_key(anon_map, bindparams), strval)
+                    for clause, strval in obj
+                ]
+            ),
+        )
+
+    def visit_setup_join_tuple(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return tuple(
+            (
+                target._gen_cache_key(anon_map, bindparams),
+                (
+                    onclause._gen_cache_key(anon_map, bindparams)
+                    if onclause is not None
+                    else None
+                ),
+                (
+                    from_._gen_cache_key(anon_map, bindparams)
+                    if from_ is not None
+                    else None
+                ),
+                tuple([(key, flags[key]) for key in sorted(flags)]),
+            )
+            for (target, onclause, from_, flags) in obj
+        )
+
+    def visit_table_hint_list(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        if not obj:
+            return ()
+
+        return (
+            attrname,
+            tuple(
+                [
+                    (
+                        clause._gen_cache_key(anon_map, bindparams),
+                        dialect_name,
+                        text,
+                    )
+                    for (clause, dialect_name), text in obj.items()
+                ]
+            ),
+        )
+
+    def visit_plain_dict(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (attrname, tuple([(key, obj[key]) for key in sorted(obj)]))
+
+    def visit_dialect_options(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            tuple(
+                (
+                    dialect_name,
+                    tuple(
+                        [
+                            (key, obj[dialect_name][key])
+                            for key in sorted(obj[dialect_name])
+                        ]
+                    ),
+                )
+                for dialect_name in sorted(obj)
+            ),
+        )
+
+    def visit_string_clauseelement_dict(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            tuple(
+                (key, obj[key]._gen_cache_key(anon_map, bindparams))
+                for key in sorted(obj)
+            ),
+        )
+
+    def visit_string_multi_dict(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            tuple(
+                (
+                    key,
+                    (
+                        value._gen_cache_key(anon_map, bindparams)
+                        if isinstance(value, HasCacheKey)
+                        else value
+                    ),
+                )
+                for key, value in [(key, obj[key]) for key in sorted(obj)]
+            ),
+        )
+
+    def visit_fromclause_canonical_column_collection(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        # inlining into the internals of ColumnCollection
+        return (
+            attrname,
+            tuple(
+                col._gen_cache_key(anon_map, bindparams)
+                for k, col, _ in obj._collection
+            ),
+        )
+
+    def visit_unknown_structure(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        anon_map[NO_CACHE] = True
+        return ()
+
+    def visit_dml_ordered_values(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        return (
+            attrname,
+            tuple(
+                (
+                    (
+                        key._gen_cache_key(anon_map, bindparams)
+                        if hasattr(key, "__clause_element__")
+                        else key
+                    ),
+                    value._gen_cache_key(anon_map, bindparams),
+                )
+                for key, value in obj
+            ),
+        )
+
+    def visit_dml_values(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        # in py37 we can assume two dictionaries created in the same
+        # insert ordering will retain that sorting
+        return (
+            attrname,
+            tuple(
+                (
+                    (
+                        k._gen_cache_key(anon_map, bindparams)
+                        if hasattr(k, "__clause_element__")
+                        else k
+                    ),
+                    obj[k]._gen_cache_key(anon_map, bindparams),
+                )
+                for k in obj
+            ),
+        )
+
+    def visit_dml_multi_values(
+        self,
+        attrname: str,
+        obj: Any,
+        parent: Any,
+        anon_map: anon_map,
+        bindparams: List[BindParameter[Any]],
+    ) -> Tuple[Any, ...]:
+        # multivalues are simply not cacheable right now
+        anon_map[NO_CACHE] = True
+        return ()
+
+
+_cache_key_traversal_visitor = _CacheKeyTraversal()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/coercions.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/coercions.py
new file mode 100644
index 00000000..802ce757
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/coercions.py
@@ -0,0 +1,1403 @@
+# sql/coercions.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import numbers
+import re
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import roles
+from . import visitors
+from ._typing import is_from_clause
+from .base import ExecutableOption
+from .base import Options
+from .cache_key import HasCacheKey
+from .visitors import Visitable
+from .. import exc
+from .. import inspection
+from .. import util
+from ..util.typing import Literal
+
+if typing.TYPE_CHECKING:
+    # elements lambdas schema selectable are set by __init__
+    from . import elements
+    from . import lambdas
+    from . import schema
+    from . import selectable
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnsClauseArgument
+    from ._typing import _DDLColumnArgument
+    from ._typing import _DMLTableArgument
+    from ._typing import _FromClauseArgument
+    from .dml import _DMLTableElement
+    from .elements import BindParameter
+    from .elements import ClauseElement
+    from .elements import ColumnClause
+    from .elements import ColumnElement
+    from .elements import NamedColumn
+    from .elements import SQLCoreOperations
+    from .elements import TextClause
+    from .schema import Column
+    from .selectable import _ColumnsClauseElement
+    from .selectable import _JoinTargetProtocol
+    from .selectable import FromClause
+    from .selectable import HasCTE
+    from .selectable import SelectBase
+    from .selectable import Subquery
+    from .visitors import _TraverseCallableType
+
+_SR = TypeVar("_SR", bound=roles.SQLRole)
+_F = TypeVar("_F", bound=Callable[..., Any])
+_StringOnlyR = TypeVar("_StringOnlyR", bound=roles.StringRole)
+_T = TypeVar("_T", bound=Any)
+
+
+def _is_literal(element):
+    """Return whether or not the element is a "literal" in the context
+    of a SQL expression construct.
+
+    """
+
+    return not isinstance(
+        element,
+        (Visitable, schema.SchemaEventTarget),
+    ) and not hasattr(element, "__clause_element__")
+
+
+def _deep_is_literal(element):
+    """Return whether or not the element is a "literal" in the context
+    of a SQL expression construct.
+
+    does a deeper more esoteric check than _is_literal.   is used
+    for lambda elements that have to distinguish values that would
+    be bound vs. not without any context.
+
+    """
+
+    if isinstance(element, collections_abc.Sequence) and not isinstance(
+        element, str
+    ):
+        for elem in element:
+            if not _deep_is_literal(elem):
+                return False
+        else:
+            return True
+
+    return (
+        not isinstance(
+            element,
+            (
+                Visitable,
+                schema.SchemaEventTarget,
+                HasCacheKey,
+                Options,
+                util.langhelpers.symbol,
+            ),
+        )
+        and not hasattr(element, "__clause_element__")
+        and (
+            not isinstance(element, type)
+            or not issubclass(element, HasCacheKey)
+        )
+    )
+
+
+def _document_text_coercion(
+    paramname: str, meth_rst: str, param_rst: str
+) -> Callable[[_F], _F]:
+    return util.add_parameter_text(
+        paramname,
+        (
+            ".. warning:: "
+            "The %s argument to %s can be passed as a Python string argument, "
+            "which will be treated "
+            "as **trusted SQL text** and rendered as given.  **DO NOT PASS "
+            "UNTRUSTED INPUT TO THIS PARAMETER**."
+        )
+        % (param_rst, meth_rst),
+    )
+
+
+def _expression_collection_was_a_list(
+    attrname: str,
+    fnname: str,
+    args: Union[Sequence[_T], Sequence[Sequence[_T]]],
+) -> Sequence[_T]:
+    if args and isinstance(args[0], (list, set, dict)) and len(args) == 1:
+        if isinstance(args[0], list):
+            raise exc.ArgumentError(
+                f'The "{attrname}" argument to {fnname}(), when '
+                "referring to a sequence "
+                "of items, is now passed as a series of positional "
+                "elements, rather than as a list. "
+            )
+        return cast("Sequence[_T]", args[0])
+
+    return cast("Sequence[_T]", args)
+
+
+@overload
+def expect(
+    role: Type[roles.TruncatedLabelRole],
+    element: Any,
+    **kw: Any,
+) -> str: ...
+
+
+@overload
+def expect(
+    role: Type[roles.DMLColumnRole],
+    element: Any,
+    *,
+    as_key: Literal[True] = ...,
+    **kw: Any,
+) -> str: ...
+
+
+@overload
+def expect(
+    role: Type[roles.LiteralValueRole],
+    element: Any,
+    **kw: Any,
+) -> BindParameter[Any]: ...
+
+
+@overload
+def expect(
+    role: Type[roles.DDLReferredColumnRole],
+    element: Any,
+    **kw: Any,
+) -> Union[Column[Any], str]: ...
+
+
+@overload
+def expect(
+    role: Type[roles.DDLConstraintColumnRole],
+    element: Any,
+    **kw: Any,
+) -> Union[Column[Any], str]: ...
+
+
+@overload
+def expect(
+    role: Type[roles.StatementOptionRole],
+    element: Any,
+    **kw: Any,
+) -> Union[ColumnElement[Any], TextClause]: ...
+
+
+@overload
+def expect(
+    role: Type[roles.LabeledColumnExprRole[Any]],
+    element: _ColumnExpressionArgument[_T],
+    **kw: Any,
+) -> NamedColumn[_T]: ...
+
+
+@overload
+def expect(
+    role: Union[
+        Type[roles.ExpressionElementRole[Any]],
+        Type[roles.LimitOffsetRole],
+        Type[roles.WhereHavingRole],
+    ],
+    element: _ColumnExpressionArgument[_T],
+    **kw: Any,
+) -> ColumnElement[_T]: ...
+
+
+@overload
+def expect(
+    role: Union[
+        Type[roles.ExpressionElementRole[Any]],
+        Type[roles.LimitOffsetRole],
+        Type[roles.WhereHavingRole],
+        Type[roles.OnClauseRole],
+        Type[roles.ColumnArgumentRole],
+    ],
+    element: Any,
+    **kw: Any,
+) -> ColumnElement[Any]: ...
+
+
+@overload
+def expect(
+    role: Type[roles.DMLTableRole],
+    element: _DMLTableArgument,
+    **kw: Any,
+) -> _DMLTableElement: ...
+
+
+@overload
+def expect(
+    role: Type[roles.HasCTERole],
+    element: HasCTE,
+    **kw: Any,
+) -> HasCTE: ...
+
+
+@overload
+def expect(
+    role: Type[roles.SelectStatementRole],
+    element: SelectBase,
+    **kw: Any,
+) -> SelectBase: ...
+
+
+@overload
+def expect(
+    role: Type[roles.FromClauseRole],
+    element: _FromClauseArgument,
+    **kw: Any,
+) -> FromClause: ...
+
+
+@overload
+def expect(
+    role: Type[roles.FromClauseRole],
+    element: SelectBase,
+    *,
+    explicit_subquery: Literal[True] = ...,
+    **kw: Any,
+) -> Subquery: ...
+
+
+@overload
+def expect(
+    role: Type[roles.ColumnsClauseRole],
+    element: _ColumnsClauseArgument[Any],
+    **kw: Any,
+) -> _ColumnsClauseElement: ...
+
+
+@overload
+def expect(
+    role: Type[roles.JoinTargetRole],
+    element: _JoinTargetProtocol,
+    **kw: Any,
+) -> _JoinTargetProtocol: ...
+
+
+# catchall for not-yet-implemented overloads
+@overload
+def expect(
+    role: Type[_SR],
+    element: Any,
+    **kw: Any,
+) -> Any: ...
+
+
+def expect(
+    role: Type[_SR],
+    element: Any,
+    *,
+    apply_propagate_attrs: Optional[ClauseElement] = None,
+    argname: Optional[str] = None,
+    post_inspect: bool = False,
+    disable_inspection: bool = False,
+    **kw: Any,
+) -> Any:
+    if (
+        role.allows_lambda
+        # note callable() will not invoke a __getattr__() method, whereas
+        # hasattr(obj, "__call__") will. by keeping the callable() check here
+        # we prevent most needless calls to hasattr()  and therefore
+        # __getattr__(), which is present on ColumnElement.
+        and callable(element)
+        and hasattr(element, "__code__")
+    ):
+        return lambdas.LambdaElement(
+            element,
+            role,
+            lambdas.LambdaOptions(**kw),
+            apply_propagate_attrs=apply_propagate_attrs,
+        )
+
+    # major case is that we are given a ClauseElement already, skip more
+    # elaborate logic up front if possible
+    impl = _impl_lookup[role]
+
+    original_element = element
+
+    if not isinstance(
+        element,
+        (
+            elements.CompilerElement,
+            schema.SchemaItem,
+            schema.FetchedValue,
+            lambdas.PyWrapper,
+        ),
+    ):
+        resolved = None
+
+        if impl._resolve_literal_only:
+            resolved = impl._literal_coercion(element, **kw)
+        else:
+            original_element = element
+
+            is_clause_element = False
+
+            # this is a special performance optimization for ORM
+            # joins used by JoinTargetImpl that we don't go through the
+            # work of creating __clause_element__() when we only need the
+            # original QueryableAttribute, as the former will do clause
+            # adaption and all that which is just thrown away here.
+            if (
+                impl._skip_clauseelement_for_target_match
+                and isinstance(element, role)
+                and hasattr(element, "__clause_element__")
+            ):
+                is_clause_element = True
+            else:
+                while hasattr(element, "__clause_element__"):
+                    is_clause_element = True
+
+                    if not getattr(element, "is_clause_element", False):
+                        element = element.__clause_element__()
+                    else:
+                        break
+
+            if not is_clause_element:
+                if impl._use_inspection and not disable_inspection:
+                    insp = inspection.inspect(element, raiseerr=False)
+                    if insp is not None:
+                        if post_inspect:
+                            insp._post_inspect
+                        try:
+                            resolved = insp.__clause_element__()
+                        except AttributeError:
+                            impl._raise_for_expected(original_element, argname)
+
+                if resolved is None:
+                    resolved = impl._literal_coercion(
+                        element, argname=argname, **kw
+                    )
+            else:
+                resolved = element
+    elif isinstance(element, lambdas.PyWrapper):
+        resolved = element._sa__py_wrapper_literal(**kw)
+    else:
+        resolved = element
+
+    if apply_propagate_attrs is not None:
+        if typing.TYPE_CHECKING:
+            assert isinstance(resolved, (SQLCoreOperations, ClauseElement))
+
+        if not apply_propagate_attrs._propagate_attrs and getattr(
+            resolved, "_propagate_attrs", None
+        ):
+            apply_propagate_attrs._propagate_attrs = resolved._propagate_attrs
+
+    if impl._role_class in resolved.__class__.__mro__:
+        if impl._post_coercion:
+            resolved = impl._post_coercion(
+                resolved,
+                argname=argname,
+                original_element=original_element,
+                **kw,
+            )
+        return resolved
+    else:
+        return impl._implicit_coercions(
+            original_element, resolved, argname=argname, **kw
+        )
+
+
+def expect_as_key(
+    role: Type[roles.DMLColumnRole], element: Any, **kw: Any
+) -> str:
+    kw.pop("as_key", None)
+    return expect(role, element, as_key=True, **kw)
+
+
+def expect_col_expression_collection(
+    role: Type[roles.DDLConstraintColumnRole],
+    expressions: Iterable[_DDLColumnArgument],
+) -> Iterator[
+    Tuple[
+        Union[str, Column[Any]],
+        Optional[ColumnClause[Any]],
+        Optional[str],
+        Optional[Union[Column[Any], str]],
+    ]
+]:
+    for expr in expressions:
+        strname = None
+        column = None
+
+        resolved: Union[Column[Any], str] = expect(role, expr)
+        if isinstance(resolved, str):
+            assert isinstance(expr, str)
+            strname = resolved = expr
+        else:
+            cols: List[Column[Any]] = []
+            col_append: _TraverseCallableType[Column[Any]] = cols.append
+            visitors.traverse(resolved, {}, {"column": col_append})
+            if cols:
+                column = cols[0]
+        add_element = column if column is not None else strname
+
+        yield resolved, column, strname, add_element
+
+
+class RoleImpl:
+    __slots__ = ("_role_class", "name", "_use_inspection")
+
+    def _literal_coercion(self, element, **kw):
+        raise NotImplementedError()
+
+    _post_coercion: Any = None
+    _resolve_literal_only = False
+    _skip_clauseelement_for_target_match = False
+
+    def __init__(self, role_class):
+        self._role_class = role_class
+        self.name = role_class._role_name
+        self._use_inspection = issubclass(role_class, roles.UsesInspection)
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        self._raise_for_expected(element, argname, resolved)
+
+    def _raise_for_expected(
+        self,
+        element: Any,
+        argname: Optional[str] = None,
+        resolved: Optional[Any] = None,
+        *,
+        advice: Optional[str] = None,
+        code: Optional[str] = None,
+        err: Optional[Exception] = None,
+        **kw: Any,
+    ) -> NoReturn:
+        if resolved is not None and resolved is not element:
+            got = "%r object resolved from %r object" % (resolved, element)
+        else:
+            got = repr(element)
+
+        if argname:
+            msg = "%s expected for argument %r; got %s." % (
+                self.name,
+                argname,
+                got,
+            )
+        else:
+            msg = "%s expected, got %s." % (self.name, got)
+
+        if advice:
+            msg += " " + advice
+
+        raise exc.ArgumentError(msg, code=code) from err
+
+
+class _Deannotate:
+    __slots__ = ()
+
+    def _post_coercion(self, resolved, **kw):
+        from .util import _deep_deannotate
+
+        return _deep_deannotate(resolved)
+
+
+class _StringOnly:
+    __slots__ = ()
+
+    _resolve_literal_only = True
+
+
+class _ReturnsStringKey(RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(self, element, resolved, argname=None, **kw):
+        if isinstance(element, str):
+            return element
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _literal_coercion(self, element, **kw):
+        return element
+
+
+class _ColumnCoercions(RoleImpl):
+    __slots__ = ()
+
+    def _warn_for_scalar_subquery_coercion(self):
+        util.warn(
+            "implicitly coercing SELECT object to scalar subquery; "
+            "please use the .scalar_subquery() method to produce a scalar "
+            "subquery.",
+        )
+
+    def _implicit_coercions(self, element, resolved, argname=None, **kw):
+        original_element = element
+        if not getattr(resolved, "is_clause_element", False):
+            self._raise_for_expected(original_element, argname, resolved)
+        elif resolved._is_select_base:
+            self._warn_for_scalar_subquery_coercion()
+            return resolved.scalar_subquery()
+        elif resolved._is_from_clause and isinstance(
+            resolved, selectable.Subquery
+        ):
+            self._warn_for_scalar_subquery_coercion()
+            return resolved.element.scalar_subquery()
+        elif self._role_class.allows_lambda and resolved._is_lambda_element:
+            return resolved
+        else:
+            self._raise_for_expected(original_element, argname, resolved)
+
+
+def _no_text_coercion(
+    element: Any,
+    argname: Optional[str] = None,
+    exc_cls: Type[exc.SQLAlchemyError] = exc.ArgumentError,
+    extra: Optional[str] = None,
+    err: Optional[Exception] = None,
+) -> NoReturn:
+    raise exc_cls(
+        "%(extra)sTextual SQL expression %(expr)r %(argname)sshould be "
+        "explicitly declared as text(%(expr)r)"
+        % {
+            "expr": util.ellipses_string(element),
+            "argname": "for argument %s" % (argname,) if argname else "",
+            "extra": "%s " % extra if extra else "",
+        }
+    ) from err
+
+
+class _NoTextCoercion(RoleImpl):
+    __slots__ = ()
+
+    def _literal_coercion(self, element, *, argname=None, **kw):
+        if isinstance(element, str) and issubclass(
+            elements.TextClause, self._role_class
+        ):
+            _no_text_coercion(element, argname)
+        else:
+            self._raise_for_expected(element, argname)
+
+
+class _CoerceLiterals(RoleImpl):
+    __slots__ = ()
+    _coerce_consts = False
+    _coerce_star = False
+    _coerce_numerics = False
+
+    def _text_coercion(self, element, argname=None):
+        return _no_text_coercion(element, argname)
+
+    def _literal_coercion(self, element, *, argname=None, **kw):
+        if isinstance(element, str):
+            if self._coerce_star and element == "*":
+                return elements.ColumnClause("*", is_literal=True)
+            else:
+                return self._text_coercion(element, argname, **kw)
+
+        if self._coerce_consts:
+            if element is None:
+                return elements.Null()
+            elif element is False:
+                return elements.False_()
+            elif element is True:
+                return elements.True_()
+
+        if self._coerce_numerics and isinstance(element, (numbers.Number)):
+            return elements.ColumnClause(str(element), is_literal=True)
+
+        self._raise_for_expected(element, argname)
+
+
+class LiteralValueImpl(RoleImpl):
+    _resolve_literal_only = True
+
+    def _implicit_coercions(
+        self,
+        element,
+        resolved,
+        argname=None,
+        *,
+        type_=None,
+        literal_execute=False,
+        **kw,
+    ):
+        if not _is_literal(resolved):
+            self._raise_for_expected(
+                element, resolved=resolved, argname=argname, **kw
+            )
+
+        return elements.BindParameter(
+            None,
+            element,
+            type_=type_,
+            unique=True,
+            literal_execute=literal_execute,
+        )
+
+    def _literal_coercion(self, element, **kw):
+        return element
+
+
+class _SelectIsNotFrom(RoleImpl):
+    __slots__ = ()
+
+    def _raise_for_expected(
+        self,
+        element: Any,
+        argname: Optional[str] = None,
+        resolved: Optional[Any] = None,
+        *,
+        advice: Optional[str] = None,
+        code: Optional[str] = None,
+        err: Optional[Exception] = None,
+        **kw: Any,
+    ) -> NoReturn:
+        if (
+            not advice
+            and isinstance(element, roles.SelectStatementRole)
+            or isinstance(resolved, roles.SelectStatementRole)
+        ):
+            advice = (
+                "To create a "
+                "FROM clause from a %s object, use the .subquery() method."
+                % (resolved.__class__ if resolved is not None else element,)
+            )
+            code = "89ve"
+        else:
+            code = None
+
+        super()._raise_for_expected(
+            element,
+            argname=argname,
+            resolved=resolved,
+            advice=advice,
+            code=code,
+            err=err,
+            **kw,
+        )
+        # never reached
+        assert False
+
+
+class HasCacheKeyImpl(RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if isinstance(element, HasCacheKey):
+            return element
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _literal_coercion(self, element, **kw):
+        return element
+
+
+class ExecutableOptionImpl(RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if isinstance(element, ExecutableOption):
+            return element
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _literal_coercion(self, element, **kw):
+        return element
+
+
+class ExpressionElementImpl(_ColumnCoercions, RoleImpl):
+    __slots__ = ()
+
+    def _literal_coercion(
+        self, element, *, name=None, type_=None, is_crud=False, **kw
+    ):
+        if (
+            element is None
+            and not is_crud
+            and (type_ is None or not type_.should_evaluate_none)
+        ):
+            # TODO: there's no test coverage now for the
+            # "should_evaluate_none" part of this, as outside of "crud" this
+            # codepath is not normally used except in some special cases
+            return elements.Null()
+        else:
+            try:
+                return elements.BindParameter(
+                    name, element, type_, unique=True, _is_crud=is_crud
+                )
+            except exc.ArgumentError as err:
+                self._raise_for_expected(element, err=err)
+
+    def _raise_for_expected(self, element, argname=None, resolved=None, **kw):
+        # select uses implicit coercion with warning instead of raising
+        if isinstance(element, selectable.Values):
+            advice = (
+                "To create a column expression from a VALUES clause, "
+                "use the .scalar_values() method."
+            )
+        elif isinstance(element, roles.AnonymizedFromClauseRole):
+            advice = (
+                "To create a column expression from a FROM clause row "
+                "as a whole, use the .table_valued() method."
+            )
+        else:
+            advice = None
+
+        return super()._raise_for_expected(
+            element, argname=argname, resolved=resolved, advice=advice, **kw
+        )
+
+
+class BinaryElementImpl(ExpressionElementImpl, RoleImpl):
+    __slots__ = ()
+
+    def _literal_coercion(  # type: ignore[override]
+        self,
+        element,
+        *,
+        expr,
+        operator,
+        bindparam_type=None,
+        argname=None,
+        **kw,
+    ):
+        try:
+            return expr._bind_param(operator, element, type_=bindparam_type)
+        except exc.ArgumentError as err:
+            self._raise_for_expected(element, err=err)
+
+    def _post_coercion(self, resolved, *, expr, bindparam_type=None, **kw):
+        if resolved.type._isnull and not expr.type._isnull:
+            resolved = resolved._with_binary_element_type(
+                bindparam_type if bindparam_type is not None else expr.type
+            )
+        return resolved
+
+
+class InElementImpl(RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_from_clause:
+            if (
+                isinstance(resolved, selectable.Alias)
+                and resolved.element._is_select_base
+            ):
+                self._warn_for_implicit_coercion(resolved)
+                return self._post_coercion(resolved.element, **kw)
+            else:
+                self._warn_for_implicit_coercion(resolved)
+                return self._post_coercion(resolved.select(), **kw)
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _warn_for_implicit_coercion(self, elem):
+        util.warn(
+            "Coercing %s object into a select() for use in IN(); "
+            "please pass a select() construct explicitly"
+            % (elem.__class__.__name__)
+        )
+
+    @util.preload_module("sqlalchemy.sql.elements")
+    def _literal_coercion(self, element, *, expr, operator, **kw):
+        if util.is_non_string_iterable(element):
+            non_literal_expressions: Dict[
+                Optional[_ColumnExpressionArgument[Any]],
+                _ColumnExpressionArgument[Any],
+            ] = {}
+            element = list(element)
+            for o in element:
+                if not _is_literal(o):
+                    if not isinstance(
+                        o, util.preloaded.sql_elements.ColumnElement
+                    ) and not hasattr(o, "__clause_element__"):
+                        self._raise_for_expected(element, **kw)
+
+                    else:
+                        non_literal_expressions[o] = o
+
+            if non_literal_expressions:
+                return elements.ClauseList(
+                    *[
+                        (
+                            non_literal_expressions[o]
+                            if o in non_literal_expressions
+                            else expr._bind_param(operator, o)
+                        )
+                        for o in element
+                    ]
+                )
+            else:
+                return expr._bind_param(operator, element, expanding=True)
+
+        else:
+            self._raise_for_expected(element, **kw)
+
+    def _post_coercion(self, element, *, expr, operator, **kw):
+        if element._is_select_base:
+            # for IN, we are doing scalar_subquery() coercion without
+            # a warning
+            return element.scalar_subquery()
+        elif isinstance(element, elements.ClauseList):
+            assert not len(element.clauses) == 0
+            return element.self_group(against=operator)
+
+        elif isinstance(element, elements.BindParameter):
+            element = element._clone(maintain_key=True)
+            element.expanding = True
+            element.expand_op = operator
+
+            return element
+        elif isinstance(element, selectable.Values):
+            return element.scalar_values()
+        else:
+            return element
+
+
+class OnClauseImpl(_ColumnCoercions, RoleImpl):
+    __slots__ = ()
+
+    _coerce_consts = True
+
+    def _literal_coercion(self, element, **kw):
+        self._raise_for_expected(element)
+
+    def _post_coercion(self, resolved, *, original_element=None, **kw):
+        # this is a hack right now as we want to use coercion on an
+        # ORM InstrumentedAttribute, but we want to return the object
+        # itself if it is one, not its clause element.
+        # ORM context _join and _legacy_join() would need to be improved
+        # to look for annotations in a clause element form.
+        if isinstance(original_element, roles.JoinTargetRole):
+            return original_element
+        return resolved
+
+
+class WhereHavingImpl(_CoerceLiterals, _ColumnCoercions, RoleImpl):
+    __slots__ = ()
+
+    _coerce_consts = True
+
+    def _text_coercion(self, element, argname=None):
+        return _no_text_coercion(element, argname)
+
+
+class StatementOptionImpl(_CoerceLiterals, RoleImpl):
+    __slots__ = ()
+
+    _coerce_consts = True
+
+    def _text_coercion(self, element, argname=None):
+        return elements.TextClause(element)
+
+
+class ColumnArgumentImpl(_NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+
+class ColumnArgumentOrKeyImpl(_ReturnsStringKey, RoleImpl):
+    __slots__ = ()
+
+
+class StrAsPlainColumnImpl(_CoerceLiterals, RoleImpl):
+    __slots__ = ()
+
+    def _text_coercion(self, element, argname=None):
+        return elements.ColumnClause(element)
+
+
+class ByOfImpl(_CoerceLiterals, _ColumnCoercions, RoleImpl, roles.ByOfRole):
+    __slots__ = ()
+
+    _coerce_consts = True
+
+    def _text_coercion(self, element, argname=None):
+        return elements._textual_label_reference(element)
+
+
+class OrderByImpl(ByOfImpl, RoleImpl):
+    __slots__ = ()
+
+    def _post_coercion(self, resolved, **kw):
+        if (
+            isinstance(resolved, self._role_class)
+            and resolved._order_by_label_element is not None
+        ):
+            return elements._label_reference(resolved)
+        else:
+            return resolved
+
+
+class GroupByImpl(ByOfImpl, RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if is_from_clause(resolved):
+            return elements.ClauseList(*resolved.c)
+        else:
+            return resolved
+
+
+class DMLColumnImpl(_ReturnsStringKey, RoleImpl):
+    __slots__ = ()
+
+    def _post_coercion(self, element, *, as_key=False, **kw):
+        if as_key:
+            return element.key
+        else:
+            return element
+
+
+class ConstExprImpl(RoleImpl):
+    __slots__ = ()
+
+    def _literal_coercion(self, element, *, argname=None, **kw):
+        if element is None:
+            return elements.Null()
+        elif element is False:
+            return elements.False_()
+        elif element is True:
+            return elements.True_()
+        else:
+            self._raise_for_expected(element, argname)
+
+
+class TruncatedLabelImpl(_StringOnly, RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if isinstance(element, str):
+            return resolved
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _literal_coercion(self, element, **kw):
+        """coerce the given value to :class:`._truncated_label`.
+
+        Existing :class:`._truncated_label` and
+        :class:`._anonymous_label` objects are passed
+        unchanged.
+        """
+
+        if isinstance(element, elements._truncated_label):
+            return element
+        else:
+            return elements._truncated_label(element)
+
+
+class DDLExpressionImpl(_Deannotate, _CoerceLiterals, RoleImpl):
+    __slots__ = ()
+
+    _coerce_consts = True
+
+    def _text_coercion(self, element, argname=None):
+        # see #5754 for why we can't easily deprecate this coercion.
+        # essentially expressions like postgresql_where would have to be
+        # text() as they come back from reflection and we don't want to
+        # have text() elements wired into the inspection dictionaries.
+        return elements.TextClause(element)
+
+
+class DDLConstraintColumnImpl(_Deannotate, _ReturnsStringKey, RoleImpl):
+    __slots__ = ()
+
+
+class DDLReferredColumnImpl(DDLConstraintColumnImpl):
+    __slots__ = ()
+
+
+class LimitOffsetImpl(RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if resolved is None:
+            return None
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _literal_coercion(  # type: ignore[override]
+        self, element, *, name, type_, **kw
+    ):
+        if element is None:
+            return None
+        else:
+            value = util.asint(element)
+            return selectable._OffsetLimitParam(
+                name, value, type_=type_, unique=True
+            )
+
+
+class LabeledColumnExprImpl(ExpressionElementImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if isinstance(resolved, roles.ExpressionElementRole):
+            return resolved.label(None)
+        else:
+            new = super()._implicit_coercions(
+                element, resolved, argname=argname, **kw
+            )
+            if isinstance(new, roles.ExpressionElementRole):
+                return new.label(None)
+            else:
+                self._raise_for_expected(element, argname, resolved)
+
+
+class ColumnsClauseImpl(_SelectIsNotFrom, _CoerceLiterals, RoleImpl):
+    __slots__ = ()
+
+    _coerce_consts = True
+    _coerce_numerics = True
+    _coerce_star = True
+
+    _guess_straight_column = re.compile(r"^\w\S*$", re.I)
+
+    def _raise_for_expected(
+        self, element, argname=None, resolved=None, *, advice=None, **kw
+    ):
+        if not advice and isinstance(element, list):
+            advice = (
+                f"Did you mean to say select("
+                f"{', '.join(repr(e) for e in element)})?"
+            )
+
+        return super()._raise_for_expected(
+            element, argname=argname, resolved=resolved, advice=advice, **kw
+        )
+
+    def _text_coercion(self, element, argname=None):
+        element = str(element)
+
+        guess_is_literal = not self._guess_straight_column.match(element)
+        raise exc.ArgumentError(
+            "Textual column expression %(column)r %(argname)sshould be "
+            "explicitly declared with text(%(column)r), "
+            "or use %(literal_column)s(%(column)r) "
+            "for more specificity"
+            % {
+                "column": util.ellipses_string(element),
+                "argname": "for argument %s" % (argname,) if argname else "",
+                "literal_column": (
+                    "literal_column" if guess_is_literal else "column"
+                ),
+            }
+        )
+
+
+class ReturnsRowsImpl(RoleImpl):
+    __slots__ = ()
+
+
+class StatementImpl(_CoerceLiterals, RoleImpl):
+    __slots__ = ()
+
+    def _post_coercion(
+        self, resolved, *, original_element, argname=None, **kw
+    ):
+        if resolved is not original_element and not isinstance(
+            original_element, str
+        ):
+            # use same method as Connection uses; this will later raise
+            # ObjectNotExecutableError
+            try:
+                original_element._execute_on_connection
+            except AttributeError:
+                util.warn_deprecated(
+                    "Object %r should not be used directly in a SQL statement "
+                    "context, such as passing to methods such as "
+                    "session.execute().  This usage will be disallowed in a "
+                    "future release.  "
+                    "Please use Core select() / update() / delete() etc. "
+                    "with Session.execute() and other statement execution "
+                    "methods." % original_element,
+                    "1.4",
+                )
+
+        return resolved
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_lambda_element:
+            return resolved
+        else:
+            return super()._implicit_coercions(
+                element, resolved, argname=argname, **kw
+            )
+
+
+class SelectStatementImpl(_NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_text_clause:
+            return resolved.columns()
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+
+class HasCTEImpl(ReturnsRowsImpl):
+    __slots__ = ()
+
+
+class IsCTEImpl(RoleImpl):
+    __slots__ = ()
+
+
+class JoinTargetImpl(RoleImpl):
+    __slots__ = ()
+
+    _skip_clauseelement_for_target_match = True
+
+    def _literal_coercion(self, element, *, argname=None, **kw):
+        self._raise_for_expected(element, argname)
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        *,
+        legacy: bool = False,
+        **kw: Any,
+    ) -> Any:
+        if isinstance(element, roles.JoinTargetRole):
+            # note that this codepath no longer occurs as of
+            # #6550, unless JoinTargetImpl._skip_clauseelement_for_target_match
+            # were set to False.
+            return element
+        elif legacy and resolved._is_select_base:
+            util.warn_deprecated(
+                "Implicit coercion of SELECT and textual SELECT "
+                "constructs into FROM clauses is deprecated; please call "
+                ".subquery() on any Core select or ORM Query object in "
+                "order to produce a subquery object.",
+                version="1.4",
+            )
+            # TODO: doing _implicit_subquery here causes tests to fail,
+            # how was this working before?  probably that ORM
+            # join logic treated it as a select and subquery would happen
+            # in _ORMJoin->Join
+            return resolved
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+
+class FromClauseImpl(_SelectIsNotFrom, _NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        *,
+        explicit_subquery: bool = False,
+        allow_select: bool = True,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_select_base:
+            if explicit_subquery:
+                return resolved.subquery()
+            elif allow_select:
+                util.warn_deprecated(
+                    "Implicit coercion of SELECT and textual SELECT "
+                    "constructs into FROM clauses is deprecated; please call "
+                    ".subquery() on any Core select or ORM Query object in "
+                    "order to produce a subquery object.",
+                    version="1.4",
+                )
+                return resolved._implicit_subquery
+        elif resolved._is_text_clause:
+            return resolved
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+    def _post_coercion(self, element, *, deannotate=False, **kw):
+        if deannotate:
+            return element._deannotate()
+        else:
+            return element
+
+
+class StrictFromClauseImpl(FromClauseImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        *,
+        allow_select: bool = False,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_select_base and allow_select:
+            util.warn_deprecated(
+                "Implicit coercion of SELECT and textual SELECT constructs "
+                "into FROM clauses is deprecated; please call .subquery() "
+                "on any Core select or ORM Query object in order to produce a "
+                "subquery object.",
+                version="1.4",
+            )
+            return resolved._implicit_subquery
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+
+class AnonymizedFromClauseImpl(StrictFromClauseImpl):
+    __slots__ = ()
+
+    def _post_coercion(self, element, *, flat=False, name=None, **kw):
+        assert name is None
+
+        return element._anonymous_fromclause(flat=flat)
+
+
+class DMLTableImpl(_SelectIsNotFrom, _NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+    def _post_coercion(self, element, **kw):
+        if "dml_table" in element._annotations:
+            return element._annotations["dml_table"]
+        else:
+            return element
+
+
+class DMLSelectImpl(_NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+    def _implicit_coercions(
+        self,
+        element: Any,
+        resolved: Any,
+        argname: Optional[str] = None,
+        **kw: Any,
+    ) -> Any:
+        if resolved._is_from_clause:
+            if (
+                isinstance(resolved, selectable.Alias)
+                and resolved.element._is_select_base
+            ):
+                return resolved.element
+            else:
+                return resolved.select()
+        else:
+            self._raise_for_expected(element, argname, resolved)
+
+
+class CompoundElementImpl(_NoTextCoercion, RoleImpl):
+    __slots__ = ()
+
+    def _raise_for_expected(self, element, argname=None, resolved=None, **kw):
+        if isinstance(element, roles.FromClauseRole):
+            if element._is_subquery:
+                advice = (
+                    "Use the plain select() object without "
+                    "calling .subquery() or .alias()."
+                )
+            else:
+                advice = (
+                    "To SELECT from any FROM clause, use the .select() method."
+                )
+        else:
+            advice = None
+        return super()._raise_for_expected(
+            element, argname=argname, resolved=resolved, advice=advice, **kw
+        )
+
+
+_impl_lookup = {}
+
+
+for name in dir(roles):
+    cls = getattr(roles, name)
+    if name.endswith("Role"):
+        name = name.replace("Role", "Impl")
+        if name in globals():
+            impl = globals()[name](cls)
+            _impl_lookup[cls] = impl
+
+if not TYPE_CHECKING:
+    ee_impl = _impl_lookup[roles.ExpressionElementRole]
+
+    for py_type in (int, bool, str, float):
+        _impl_lookup[roles.ExpressionElementRole[py_type]] = ee_impl
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/compiler.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/compiler.py
new file mode 100644
index 00000000..49e8ce50
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/compiler.py
@@ -0,0 +1,7827 @@
+# sql/compiler.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Base SQL and DDL compiler implementations.
+
+Classes provided include:
+
+:class:`.compiler.SQLCompiler` - renders SQL
+strings
+
+:class:`.compiler.DDLCompiler` - renders DDL
+(data definition language) strings
+
+:class:`.compiler.GenericTypeCompiler` - renders
+type specification strings.
+
+To generate user-defined SQL strings, see
+:doc:`/ext/compiler`.
+
+"""
+from __future__ import annotations
+
+import collections
+import collections.abc as collections_abc
+import contextlib
+from enum import IntEnum
+import functools
+import itertools
+import operator
+import re
+from time import perf_counter
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import FrozenSet
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import MutableMapping
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import Pattern
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import base
+from . import coercions
+from . import crud
+from . import elements
+from . import functions
+from . import operators
+from . import roles
+from . import schema
+from . import selectable
+from . import sqltypes
+from . import util as sql_util
+from ._typing import is_column_element
+from ._typing import is_dml
+from .base import _de_clone
+from .base import _from_objects
+from .base import _NONE_NAME
+from .base import _SentinelDefaultCharacterization
+from .base import Executable
+from .base import NO_ARG
+from .elements import ClauseElement
+from .elements import quoted_name
+from .schema import Column
+from .sqltypes import TupleType
+from .type_api import TypeEngine
+from .visitors import prefix_anon_map
+from .visitors import Visitable
+from .. import exc
+from .. import util
+from ..util import FastIntFlag
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import TypedDict
+
+if typing.TYPE_CHECKING:
+    from .annotation import _AnnotationDict
+    from .base import _AmbiguousTableNameMap
+    from .base import CompileState
+    from .cache_key import CacheKey
+    from .ddl import ExecutableDDLElement
+    from .dml import Insert
+    from .dml import UpdateBase
+    from .dml import ValuesBase
+    from .elements import _truncated_label
+    from .elements import BindParameter
+    from .elements import ColumnClause
+    from .elements import ColumnElement
+    from .elements import Label
+    from .functions import Function
+    from .schema import Table
+    from .selectable import AliasedReturnsRows
+    from .selectable import CompoundSelectState
+    from .selectable import CTE
+    from .selectable import FromClause
+    from .selectable import NamedFromClause
+    from .selectable import ReturnsRows
+    from .selectable import Select
+    from .selectable import SelectState
+    from .type_api import _BindProcessorType
+    from ..engine.cursor import CursorResultMetaData
+    from ..engine.interfaces import _CoreSingleExecuteParams
+    from ..engine.interfaces import _DBAPIAnyExecuteParams
+    from ..engine.interfaces import _DBAPIMultiExecuteParams
+    from ..engine.interfaces import _DBAPISingleExecuteParams
+    from ..engine.interfaces import _ExecuteOptions
+    from ..engine.interfaces import _GenericSetInputSizesType
+    from ..engine.interfaces import _MutableCoreSingleExecuteParams
+    from ..engine.interfaces import Dialect
+    from ..engine.interfaces import SchemaTranslateMapType
+
+_FromHintsType = Dict["FromClause", str]
+
+RESERVED_WORDS = {
+    "all",
+    "analyse",
+    "analyze",
+    "and",
+    "any",
+    "array",
+    "as",
+    "asc",
+    "asymmetric",
+    "authorization",
+    "between",
+    "binary",
+    "both",
+    "case",
+    "cast",
+    "check",
+    "collate",
+    "column",
+    "constraint",
+    "create",
+    "cross",
+    "current_date",
+    "current_role",
+    "current_time",
+    "current_timestamp",
+    "current_user",
+    "default",
+    "deferrable",
+    "desc",
+    "distinct",
+    "do",
+    "else",
+    "end",
+    "except",
+    "false",
+    "for",
+    "foreign",
+    "freeze",
+    "from",
+    "full",
+    "grant",
+    "group",
+    "having",
+    "ilike",
+    "in",
+    "initially",
+    "inner",
+    "intersect",
+    "into",
+    "is",
+    "isnull",
+    "join",
+    "leading",
+    "left",
+    "like",
+    "limit",
+    "localtime",
+    "localtimestamp",
+    "natural",
+    "new",
+    "not",
+    "notnull",
+    "null",
+    "off",
+    "offset",
+    "old",
+    "on",
+    "only",
+    "or",
+    "order",
+    "outer",
+    "overlaps",
+    "placing",
+    "primary",
+    "references",
+    "right",
+    "select",
+    "session_user",
+    "set",
+    "similar",
+    "some",
+    "symmetric",
+    "table",
+    "then",
+    "to",
+    "trailing",
+    "true",
+    "union",
+    "unique",
+    "user",
+    "using",
+    "verbose",
+    "when",
+    "where",
+}
+
+LEGAL_CHARACTERS = re.compile(r"^[A-Z0-9_$]+$", re.I)
+LEGAL_CHARACTERS_PLUS_SPACE = re.compile(r"^[A-Z0-9_ $]+$", re.I)
+ILLEGAL_INITIAL_CHARACTERS = {str(x) for x in range(0, 10)}.union(["$"])
+
+FK_ON_DELETE = re.compile(
+    r"^(?:RESTRICT|CASCADE|SET NULL|NO ACTION|SET DEFAULT)$", re.I
+)
+FK_ON_UPDATE = re.compile(
+    r"^(?:RESTRICT|CASCADE|SET NULL|NO ACTION|SET DEFAULT)$", re.I
+)
+FK_INITIALLY = re.compile(r"^(?:DEFERRED|IMMEDIATE)$", re.I)
+BIND_PARAMS = re.compile(r"(?<![:\w\$\x5c]):([\w\$]+)(?![:\w\$])", re.UNICODE)
+BIND_PARAMS_ESC = re.compile(r"\x5c(:[\w\$]*)(?![:\w\$])", re.UNICODE)
+
+_pyformat_template = "%%(%(name)s)s"
+BIND_TEMPLATES = {
+    "pyformat": _pyformat_template,
+    "qmark": "?",
+    "format": "%%s",
+    "numeric": ":[_POSITION]",
+    "numeric_dollar": "$[_POSITION]",
+    "named": ":%(name)s",
+}
+
+
+OPERATORS = {
+    # binary
+    operators.and_: " AND ",
+    operators.or_: " OR ",
+    operators.add: " + ",
+    operators.mul: " * ",
+    operators.sub: " - ",
+    operators.mod: " % ",
+    operators.neg: "-",
+    operators.lt: " < ",
+    operators.le: " <= ",
+    operators.ne: " != ",
+    operators.gt: " > ",
+    operators.ge: " >= ",
+    operators.eq: " = ",
+    operators.is_distinct_from: " IS DISTINCT FROM ",
+    operators.is_not_distinct_from: " IS NOT DISTINCT FROM ",
+    operators.concat_op: " || ",
+    operators.match_op: " MATCH ",
+    operators.not_match_op: " NOT MATCH ",
+    operators.in_op: " IN ",
+    operators.not_in_op: " NOT IN ",
+    operators.comma_op: ", ",
+    operators.from_: " FROM ",
+    operators.as_: " AS ",
+    operators.is_: " IS ",
+    operators.is_not: " IS NOT ",
+    operators.collate: " COLLATE ",
+    # unary
+    operators.exists: "EXISTS ",
+    operators.distinct_op: "DISTINCT ",
+    operators.inv: "NOT ",
+    operators.any_op: "ANY ",
+    operators.all_op: "ALL ",
+    # modifiers
+    operators.desc_op: " DESC",
+    operators.asc_op: " ASC",
+    operators.nulls_first_op: " NULLS FIRST",
+    operators.nulls_last_op: " NULLS LAST",
+    # bitwise
+    operators.bitwise_xor_op: " ^ ",
+    operators.bitwise_or_op: " | ",
+    operators.bitwise_and_op: " & ",
+    operators.bitwise_not_op: "~",
+    operators.bitwise_lshift_op: " << ",
+    operators.bitwise_rshift_op: " >> ",
+}
+
+FUNCTIONS: Dict[Type[Function[Any]], str] = {
+    functions.coalesce: "coalesce",
+    functions.current_date: "CURRENT_DATE",
+    functions.current_time: "CURRENT_TIME",
+    functions.current_timestamp: "CURRENT_TIMESTAMP",
+    functions.current_user: "CURRENT_USER",
+    functions.localtime: "LOCALTIME",
+    functions.localtimestamp: "LOCALTIMESTAMP",
+    functions.random: "random",
+    functions.sysdate: "sysdate",
+    functions.session_user: "SESSION_USER",
+    functions.user: "USER",
+    functions.cube: "CUBE",
+    functions.rollup: "ROLLUP",
+    functions.grouping_sets: "GROUPING SETS",
+}
+
+
+EXTRACT_MAP = {
+    "month": "month",
+    "day": "day",
+    "year": "year",
+    "second": "second",
+    "hour": "hour",
+    "doy": "doy",
+    "minute": "minute",
+    "quarter": "quarter",
+    "dow": "dow",
+    "week": "week",
+    "epoch": "epoch",
+    "milliseconds": "milliseconds",
+    "microseconds": "microseconds",
+    "timezone_hour": "timezone_hour",
+    "timezone_minute": "timezone_minute",
+}
+
+COMPOUND_KEYWORDS = {
+    selectable._CompoundSelectKeyword.UNION: "UNION",
+    selectable._CompoundSelectKeyword.UNION_ALL: "UNION ALL",
+    selectable._CompoundSelectKeyword.EXCEPT: "EXCEPT",
+    selectable._CompoundSelectKeyword.EXCEPT_ALL: "EXCEPT ALL",
+    selectable._CompoundSelectKeyword.INTERSECT: "INTERSECT",
+    selectable._CompoundSelectKeyword.INTERSECT_ALL: "INTERSECT ALL",
+}
+
+
+class ResultColumnsEntry(NamedTuple):
+    """Tracks a column expression that is expected to be represented
+    in the result rows for this statement.
+
+    This normally refers to the columns clause of a SELECT statement
+    but may also refer to a RETURNING clause, as well as for dialect-specific
+    emulations.
+
+    """
+
+    keyname: str
+    """string name that's expected in cursor.description"""
+
+    name: str
+    """column name, may be labeled"""
+
+    objects: Tuple[Any, ...]
+    """sequence of objects that should be able to locate this column
+    in a RowMapping.  This is typically string names and aliases
+    as well as Column objects.
+
+    """
+
+    type: TypeEngine[Any]
+    """Datatype to be associated with this column.   This is where
+    the "result processing" logic directly links the compiled statement
+    to the rows that come back from the cursor.
+
+    """
+
+
+class _ResultMapAppender(Protocol):
+    def __call__(
+        self,
+        keyname: str,
+        name: str,
+        objects: Sequence[Any],
+        type_: TypeEngine[Any],
+    ) -> None: ...
+
+
+# integer indexes into ResultColumnsEntry used by cursor.py.
+# some profiling showed integer access faster than named tuple
+RM_RENDERED_NAME: Literal[0] = 0
+RM_NAME: Literal[1] = 1
+RM_OBJECTS: Literal[2] = 2
+RM_TYPE: Literal[3] = 3
+
+
+class _BaseCompilerStackEntry(TypedDict):
+    asfrom_froms: Set[FromClause]
+    correlate_froms: Set[FromClause]
+    selectable: ReturnsRows
+
+
+class _CompilerStackEntry(_BaseCompilerStackEntry, total=False):
+    compile_state: CompileState
+    need_result_map_for_nested: bool
+    need_result_map_for_compound: bool
+    select_0: ReturnsRows
+    insert_from_select: Select[Any]
+
+
+class ExpandedState(NamedTuple):
+    """represents state to use when producing "expanded" and
+    "post compile" bound parameters for a statement.
+
+    "expanded" parameters are parameters that are generated at
+    statement execution time to suit a number of parameters passed, the most
+    prominent example being the individual elements inside of an IN expression.
+
+    "post compile" parameters are parameters where the SQL literal value
+    will be rendered into the SQL statement at execution time, rather than
+    being passed as separate parameters to the driver.
+
+    To create an :class:`.ExpandedState` instance, use the
+    :meth:`.SQLCompiler.construct_expanded_state` method on any
+    :class:`.SQLCompiler` instance.
+
+    """
+
+    statement: str
+    """String SQL statement with parameters fully expanded"""
+
+    parameters: _CoreSingleExecuteParams
+    """Parameter dictionary with parameters fully expanded.
+
+    For a statement that uses named parameters, this dictionary will map
+    exactly to the names in the statement.  For a statement that uses
+    positional parameters, the :attr:`.ExpandedState.positional_parameters`
+    will yield a tuple with the positional parameter set.
+
+    """
+
+    processors: Mapping[str, _BindProcessorType[Any]]
+    """mapping of bound value processors"""
+
+    positiontup: Optional[Sequence[str]]
+    """Sequence of string names indicating the order of positional
+    parameters"""
+
+    parameter_expansion: Mapping[str, List[str]]
+    """Mapping representing the intermediary link from original parameter
+    name to list of "expanded" parameter names, for those parameters that
+    were expanded."""
+
+    @property
+    def positional_parameters(self) -> Tuple[Any, ...]:
+        """Tuple of positional parameters, for statements that were compiled
+        using a positional paramstyle.
+
+        """
+        if self.positiontup is None:
+            raise exc.InvalidRequestError(
+                "statement does not use a positional paramstyle"
+            )
+        return tuple(self.parameters[key] for key in self.positiontup)
+
+    @property
+    def additional_parameters(self) -> _CoreSingleExecuteParams:
+        """synonym for :attr:`.ExpandedState.parameters`."""
+        return self.parameters
+
+
+class _InsertManyValues(NamedTuple):
+    """represents state to use for executing an "insertmanyvalues" statement.
+
+    The primary consumers of this object are the
+    :meth:`.SQLCompiler._deliver_insertmanyvalues_batches` and
+    :meth:`.DefaultDialect._deliver_insertmanyvalues_batches` methods.
+
+    .. versionadded:: 2.0
+
+    """
+
+    is_default_expr: bool
+    """if True, the statement is of the form
+    ``INSERT INTO TABLE DEFAULT VALUES``, and can't be rewritten as a "batch"
+
+    """
+
+    single_values_expr: str
+    """The rendered "values" clause of the INSERT statement.
+
+    This is typically the parenthesized section e.g. "(?, ?, ?)" or similar.
+    The insertmanyvalues logic uses this string as a search and replace
+    target.
+
+    """
+
+    insert_crud_params: List[crud._CrudParamElementStr]
+    """List of Column / bind names etc. used while rewriting the statement"""
+
+    num_positional_params_counted: int
+    """the number of bound parameters in a single-row statement.
+
+    This count may be larger or smaller than the actual number of columns
+    targeted in the INSERT, as it accommodates for SQL expressions
+    in the values list that may have zero or more parameters embedded
+    within them.
+
+    This count is part of what's used to organize rewritten parameter lists
+    when batching.
+
+    """
+
+    sort_by_parameter_order: bool = False
+    """if the deterministic_returnined_order parameter were used on the
+    insert.
+
+    All of the attributes following this will only be used if this is True.
+
+    """
+
+    includes_upsert_behaviors: bool = False
+    """if True, we have to accommodate for upsert behaviors.
+
+    This will in some cases downgrade "insertmanyvalues" that requests
+    deterministic ordering.
+
+    """
+
+    sentinel_columns: Optional[Sequence[Column[Any]]] = None
+    """List of sentinel columns that were located.
+
+    This list is only here if the INSERT asked for
+    sort_by_parameter_order=True,
+    and dialect-appropriate sentinel columns were located.
+
+    .. versionadded:: 2.0.10
+
+    """
+
+    num_sentinel_columns: int = 0
+    """how many sentinel columns are in the above list, if any.
+
+    This is the same as
+    ``len(sentinel_columns) if sentinel_columns is not None else 0``
+
+    """
+
+    sentinel_param_keys: Optional[Sequence[str]] = None
+    """parameter str keys in each param dictionary / tuple
+    that would link to the client side "sentinel" values for that row, which
+    we can use to match up parameter sets to result rows.
+
+    This is only present if sentinel_columns is present and the INSERT
+    statement actually refers to client side values for these sentinel
+    columns.
+
+    .. versionadded:: 2.0.10
+
+    .. versionchanged:: 2.0.29 - the sequence is now string dictionary keys
+       only, used against the "compiled parameteters" collection before
+       the parameters were converted by bound parameter processors
+
+    """
+
+    implicit_sentinel: bool = False
+    """if True, we have exactly one sentinel column and it uses a server side
+    value, currently has to generate an incrementing integer value.
+
+    The dialect in question would have asserted that it supports receiving
+    these values back and sorting on that value as a means of guaranteeing
+    correlation with the incoming parameter list.
+
+    .. versionadded:: 2.0.10
+
+    """
+
+    embed_values_counter: bool = False
+    """Whether to embed an incrementing integer counter in each parameter
+    set within the VALUES clause as parameters are batched over.
+
+    This is only used for a specific INSERT..SELECT..VALUES..RETURNING syntax
+    where a subquery is used to produce value tuples.  Current support
+    includes PostgreSQL, Microsoft SQL Server.
+
+    .. versionadded:: 2.0.10
+
+    """
+
+
+class _InsertManyValuesBatch(NamedTuple):
+    """represents an individual batch SQL statement for insertmanyvalues.
+
+    This is passed through the
+    :meth:`.SQLCompiler._deliver_insertmanyvalues_batches` and
+    :meth:`.DefaultDialect._deliver_insertmanyvalues_batches` methods out
+    to the :class:`.Connection` within the
+    :meth:`.Connection._exec_insertmany_context` method.
+
+    .. versionadded:: 2.0.10
+
+    """
+
+    replaced_statement: str
+    replaced_parameters: _DBAPIAnyExecuteParams
+    processed_setinputsizes: Optional[_GenericSetInputSizesType]
+    batch: Sequence[_DBAPISingleExecuteParams]
+    sentinel_values: Sequence[Tuple[Any, ...]]
+    current_batch_size: int
+    batchnum: int
+    total_batches: int
+    rows_sorted: bool
+    is_downgraded: bool
+
+
+class InsertmanyvaluesSentinelOpts(FastIntFlag):
+    """bitflag enum indicating styles of PK defaults
+    which can work as implicit sentinel columns
+
+    """
+
+    NOT_SUPPORTED = 1
+    AUTOINCREMENT = 2
+    IDENTITY = 4
+    SEQUENCE = 8
+
+    ANY_AUTOINCREMENT = AUTOINCREMENT | IDENTITY | SEQUENCE
+    _SUPPORTED_OR_NOT = NOT_SUPPORTED | ANY_AUTOINCREMENT
+
+    USE_INSERT_FROM_SELECT = 16
+    RENDER_SELECT_COL_CASTS = 64
+
+
+class CompilerState(IntEnum):
+    COMPILING = 0
+    """statement is present, compilation phase in progress"""
+
+    STRING_APPLIED = 1
+    """statement is present, string form of the statement has been applied.
+
+    Additional processors by subclasses may still be pending.
+
+    """
+
+    NO_STATEMENT = 2
+    """compiler does not have a statement to compile, is used
+    for method access"""
+
+
+class Linting(IntEnum):
+    """represent preferences for the 'SQL linting' feature.
+
+    this feature currently includes support for flagging cartesian products
+    in SQL statements.
+
+    """
+
+    NO_LINTING = 0
+    "Disable all linting."
+
+    COLLECT_CARTESIAN_PRODUCTS = 1
+    """Collect data on FROMs and cartesian products and gather into
+    'self.from_linter'"""
+
+    WARN_LINTING = 2
+    "Emit warnings for linters that find problems"
+
+    FROM_LINTING = COLLECT_CARTESIAN_PRODUCTS | WARN_LINTING
+    """Warn for cartesian products; combines COLLECT_CARTESIAN_PRODUCTS
+    and WARN_LINTING"""
+
+
+NO_LINTING, COLLECT_CARTESIAN_PRODUCTS, WARN_LINTING, FROM_LINTING = tuple(
+    Linting
+)
+
+
+class FromLinter(collections.namedtuple("FromLinter", ["froms", "edges"])):
+    """represents current state for the "cartesian product" detection
+    feature."""
+
+    def lint(self, start=None):
+        froms = self.froms
+        if not froms:
+            return None, None
+
+        edges = set(self.edges)
+        the_rest = set(froms)
+
+        if start is not None:
+            start_with = start
+            the_rest.remove(start_with)
+        else:
+            start_with = the_rest.pop()
+
+        stack = collections.deque([start_with])
+
+        while stack and the_rest:
+            node = stack.popleft()
+            the_rest.discard(node)
+
+            # comparison of nodes in edges here is based on hash equality, as
+            # there are "annotated" elements that match the non-annotated ones.
+            #   to remove the need for in-python hash() calls, use native
+            # containment routines (e.g. "node in edge", "edge.index(node)")
+            to_remove = {edge for edge in edges if node in edge}
+
+            # appendleft the node in each edge that is not
+            # the one that matched.
+            stack.extendleft(edge[not edge.index(node)] for edge in to_remove)
+            edges.difference_update(to_remove)
+
+        # FROMS left over?  boom
+        if the_rest:
+            return the_rest, start_with
+        else:
+            return None, None
+
+    def warn(self, stmt_type="SELECT"):
+        the_rest, start_with = self.lint()
+
+        # FROMS left over?  boom
+        if the_rest:
+            froms = the_rest
+            if froms:
+                template = (
+                    "{stmt_type} statement has a cartesian product between "
+                    "FROM element(s) {froms} and "
+                    'FROM element "{start}".  Apply join condition(s) '
+                    "between each element to resolve."
+                )
+                froms_str = ", ".join(
+                    f'"{self.froms[from_]}"' for from_ in froms
+                )
+                message = template.format(
+                    stmt_type=stmt_type,
+                    froms=froms_str,
+                    start=self.froms[start_with],
+                )
+
+                util.warn(message)
+
+
+class Compiled:
+    """Represent a compiled SQL or DDL expression.
+
+    The ``__str__`` method of the ``Compiled`` object should produce
+    the actual text of the statement.  ``Compiled`` objects are
+    specific to their underlying database dialect, and also may
+    or may not be specific to the columns referenced within a
+    particular set of bind parameters.  In no case should the
+    ``Compiled`` object be dependent on the actual values of those
+    bind parameters, even though it may reference those values as
+    defaults.
+    """
+
+    statement: Optional[ClauseElement] = None
+    "The statement to compile."
+    string: str = ""
+    "The string representation of the ``statement``"
+
+    state: CompilerState
+    """description of the compiler's state"""
+
+    is_sql = False
+    is_ddl = False
+
+    _cached_metadata: Optional[CursorResultMetaData] = None
+
+    _result_columns: Optional[List[ResultColumnsEntry]] = None
+
+    schema_translate_map: Optional[SchemaTranslateMapType] = None
+
+    execution_options: _ExecuteOptions = util.EMPTY_DICT
+    """
+    Execution options propagated from the statement.   In some cases,
+    sub-elements of the statement can modify these.
+    """
+
+    preparer: IdentifierPreparer
+
+    _annotations: _AnnotationDict = util.EMPTY_DICT
+
+    compile_state: Optional[CompileState] = None
+    """Optional :class:`.CompileState` object that maintains additional
+    state used by the compiler.
+
+    Major executable objects such as :class:`_expression.Insert`,
+    :class:`_expression.Update`, :class:`_expression.Delete`,
+    :class:`_expression.Select` will generate this
+    state when compiled in order to calculate additional information about the
+    object.   For the top level object that is to be executed, the state can be
+    stored here where it can also have applicability towards result set
+    processing.
+
+    .. versionadded:: 1.4
+
+    """
+
+    dml_compile_state: Optional[CompileState] = None
+    """Optional :class:`.CompileState` assigned at the same point that
+    .isinsert, .isupdate, or .isdelete is assigned.
+
+    This will normally be the same object as .compile_state, with the
+    exception of cases like the :class:`.ORMFromStatementCompileState`
+    object.
+
+    .. versionadded:: 1.4.40
+
+    """
+
+    cache_key: Optional[CacheKey] = None
+    """The :class:`.CacheKey` that was generated ahead of creating this
+    :class:`.Compiled` object.
+
+    This is used for routines that need access to the original
+    :class:`.CacheKey` instance generated when the :class:`.Compiled`
+    instance was first cached, typically in order to reconcile
+    the original list of :class:`.BindParameter` objects with a
+    per-statement list that's generated on each call.
+
+    """
+
+    _gen_time: float
+    """Generation time of this :class:`.Compiled`, used for reporting
+    cache stats."""
+
+    def __init__(
+        self,
+        dialect: Dialect,
+        statement: Optional[ClauseElement],
+        schema_translate_map: Optional[SchemaTranslateMapType] = None,
+        render_schema_translate: bool = False,
+        compile_kwargs: Mapping[str, Any] = util.immutabledict(),
+    ):
+        """Construct a new :class:`.Compiled` object.
+
+        :param dialect: :class:`.Dialect` to compile against.
+
+        :param statement: :class:`_expression.ClauseElement` to be compiled.
+
+        :param schema_translate_map: dictionary of schema names to be
+         translated when forming the resultant SQL
+
+         .. seealso::
+
+            :ref:`schema_translating`
+
+        :param compile_kwargs: additional kwargs that will be
+         passed to the initial call to :meth:`.Compiled.process`.
+
+
+        """
+        self.dialect = dialect
+        self.preparer = self.dialect.identifier_preparer
+        if schema_translate_map:
+            self.schema_translate_map = schema_translate_map
+            self.preparer = self.preparer._with_schema_translate(
+                schema_translate_map
+            )
+
+        if statement is not None:
+            self.state = CompilerState.COMPILING
+            self.statement = statement
+            self.can_execute = statement.supports_execution
+            self._annotations = statement._annotations
+            if self.can_execute:
+                if TYPE_CHECKING:
+                    assert isinstance(statement, Executable)
+                self.execution_options = statement._execution_options
+            self.string = self.process(self.statement, **compile_kwargs)
+
+            if render_schema_translate:
+                self.string = self.preparer._render_schema_translates(
+                    self.string, schema_translate_map
+                )
+
+            self.state = CompilerState.STRING_APPLIED
+        else:
+            self.state = CompilerState.NO_STATEMENT
+
+        self._gen_time = perf_counter()
+
+    def __init_subclass__(cls) -> None:
+        cls._init_compiler_cls()
+        return super().__init_subclass__()
+
+    @classmethod
+    def _init_compiler_cls(cls):
+        pass
+
+    def _execute_on_connection(
+        self, connection, distilled_params, execution_options
+    ):
+        if self.can_execute:
+            return connection._execute_compiled(
+                self, distilled_params, execution_options
+            )
+        else:
+            raise exc.ObjectNotExecutableError(self.statement)
+
+    def visit_unsupported_compilation(self, element, err, **kw):
+        raise exc.UnsupportedCompilationError(self, type(element)) from err
+
+    @property
+    def sql_compiler(self):
+        """Return a Compiled that is capable of processing SQL expressions.
+
+        If this compiler is one, it would likely just return 'self'.
+
+        """
+
+        raise NotImplementedError()
+
+    def process(self, obj: Visitable, **kwargs: Any) -> str:
+        return obj._compiler_dispatch(self, **kwargs)
+
+    def __str__(self) -> str:
+        """Return the string text of the generated SQL or DDL."""
+
+        if self.state is CompilerState.STRING_APPLIED:
+            return self.string
+        else:
+            return ""
+
+    def construct_params(
+        self,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        extracted_parameters: Optional[Sequence[BindParameter[Any]]] = None,
+        escape_names: bool = True,
+    ) -> Optional[_MutableCoreSingleExecuteParams]:
+        """Return the bind params for this compiled object.
+
+        :param params: a dict of string/object pairs whose values will
+                       override bind values compiled in to the
+                       statement.
+        """
+
+        raise NotImplementedError()
+
+    @property
+    def params(self):
+        """Return the bind params for this compiled object."""
+        return self.construct_params()
+
+
+class TypeCompiler(util.EnsureKWArg):
+    """Produces DDL specification for TypeEngine objects."""
+
+    ensure_kwarg = r"visit_\w+"
+
+    def __init__(self, dialect: Dialect):
+        self.dialect = dialect
+
+    def process(self, type_: TypeEngine[Any], **kw: Any) -> str:
+        if (
+            type_._variant_mapping
+            and self.dialect.name in type_._variant_mapping
+        ):
+            type_ = type_._variant_mapping[self.dialect.name]
+        return type_._compiler_dispatch(self, **kw)
+
+    def visit_unsupported_compilation(
+        self, element: Any, err: Exception, **kw: Any
+    ) -> NoReturn:
+        raise exc.UnsupportedCompilationError(self, element) from err
+
+
+# this was a Visitable, but to allow accurate detection of
+# column elements this is actually a column element
+class _CompileLabel(
+    roles.BinaryElementRole[Any], elements.CompilerColumnElement
+):
+    """lightweight label object which acts as an expression.Label."""
+
+    __visit_name__ = "label"
+    __slots__ = "element", "name", "_alt_names"
+
+    def __init__(self, col, name, alt_names=()):
+        self.element = col
+        self.name = name
+        self._alt_names = (col,) + alt_names
+
+    @property
+    def proxy_set(self):
+        return self.element.proxy_set
+
+    @property
+    def type(self):
+        return self.element.type
+
+    def self_group(self, **kw):
+        return self
+
+
+class ilike_case_insensitive(
+    roles.BinaryElementRole[Any], elements.CompilerColumnElement
+):
+    """produce a wrapping element for a case-insensitive portion of
+    an ILIKE construct.
+
+    The construct usually renders the ``lower()`` function, but on
+    PostgreSQL will pass silently with the assumption that "ILIKE"
+    is being used.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __visit_name__ = "ilike_case_insensitive_operand"
+    __slots__ = "element", "comparator"
+
+    def __init__(self, element):
+        self.element = element
+        self.comparator = element.comparator
+
+    @property
+    def proxy_set(self):
+        return self.element.proxy_set
+
+    @property
+    def type(self):
+        return self.element.type
+
+    def self_group(self, **kw):
+        return self
+
+    def _with_binary_element_type(self, type_):
+        return ilike_case_insensitive(
+            self.element._with_binary_element_type(type_)
+        )
+
+
+class SQLCompiler(Compiled):
+    """Default implementation of :class:`.Compiled`.
+
+    Compiles :class:`_expression.ClauseElement` objects into SQL strings.
+
+    """
+
+    extract_map = EXTRACT_MAP
+
+    bindname_escape_characters: ClassVar[Mapping[str, str]] = (
+        util.immutabledict(
+            {
+                "%": "P",
+                "(": "A",
+                ")": "Z",
+                ":": "C",
+                ".": "_",
+                "[": "_",
+                "]": "_",
+                " ": "_",
+            }
+        )
+    )
+    """A mapping (e.g. dict or similar) containing a lookup of
+    characters keyed to replacement characters which will be applied to all
+    'bind names' used in SQL statements as a form of 'escaping'; the given
+    characters are replaced entirely with the 'replacement' character when
+    rendered in the SQL statement, and a similar translation is performed
+    on the incoming names used in parameter dictionaries passed to methods
+    like :meth:`_engine.Connection.execute`.
+
+    This allows bound parameter names used in :func:`_sql.bindparam` and
+    other constructs to have any arbitrary characters present without any
+    concern for characters that aren't allowed at all on the target database.
+
+    Third party dialects can establish their own dictionary here to replace the
+    default mapping, which will ensure that the particular characters in the
+    mapping will never appear in a bound parameter name.
+
+    The dictionary is evaluated at **class creation time**, so cannot be
+    modified at runtime; it must be present on the class when the class
+    is first declared.
+
+    Note that for dialects that have additional bound parameter rules such
+    as additional restrictions on leading characters, the
+    :meth:`_sql.SQLCompiler.bindparam_string` method may need to be augmented.
+    See the cx_Oracle compiler for an example of this.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    _bind_translate_re: ClassVar[Pattern[str]]
+    _bind_translate_chars: ClassVar[Mapping[str, str]]
+
+    is_sql = True
+
+    compound_keywords = COMPOUND_KEYWORDS
+
+    isdelete: bool = False
+    isinsert: bool = False
+    isupdate: bool = False
+    """class-level defaults which can be set at the instance
+    level to define if this Compiled instance represents
+    INSERT/UPDATE/DELETE
+    """
+
+    postfetch: Optional[List[Column[Any]]]
+    """list of columns that can be post-fetched after INSERT or UPDATE to
+    receive server-updated values"""
+
+    insert_prefetch: Sequence[Column[Any]] = ()
+    """list of columns for which default values should be evaluated before
+    an INSERT takes place"""
+
+    update_prefetch: Sequence[Column[Any]] = ()
+    """list of columns for which onupdate default values should be evaluated
+    before an UPDATE takes place"""
+
+    implicit_returning: Optional[Sequence[ColumnElement[Any]]] = None
+    """list of "implicit" returning columns for a toplevel INSERT or UPDATE
+    statement, used to receive newly generated values of columns.
+
+    .. versionadded:: 2.0  ``implicit_returning`` replaces the previous
+       ``returning`` collection, which was not a generalized RETURNING
+       collection and instead was in fact specific to the "implicit returning"
+       feature.
+
+    """
+
+    isplaintext: bool = False
+
+    binds: Dict[str, BindParameter[Any]]
+    """a dictionary of bind parameter keys to BindParameter instances."""
+
+    bind_names: Dict[BindParameter[Any], str]
+    """a dictionary of BindParameter instances to "compiled" names
+    that are actually present in the generated SQL"""
+
+    stack: List[_CompilerStackEntry]
+    """major statements such as SELECT, INSERT, UPDATE, DELETE are
+    tracked in this stack using an entry format."""
+
+    returning_precedes_values: bool = False
+    """set to True classwide to generate RETURNING
+    clauses before the VALUES or WHERE clause (i.e. MSSQL)
+    """
+
+    render_table_with_column_in_update_from: bool = False
+    """set to True classwide to indicate the SET clause
+    in a multi-table UPDATE statement should qualify
+    columns with the table name (i.e. MySQL only)
+    """
+
+    ansi_bind_rules: bool = False
+    """SQL 92 doesn't allow bind parameters to be used
+    in the columns clause of a SELECT, nor does it allow
+    ambiguous expressions like "? = ?".  A compiler
+    subclass can set this flag to False if the target
+    driver/DB enforces this
+    """
+
+    bindtemplate: str
+    """template to render bound parameters based on paramstyle."""
+
+    compilation_bindtemplate: str
+    """template used by compiler to render parameters before positional
+    paramstyle application"""
+
+    _numeric_binds_identifier_char: str
+    """Character that's used to as the identifier of a numerical bind param.
+    For example if this char is set to ``$``, numerical binds will be rendered
+    in the form ``$1, $2, $3``.
+    """
+
+    _result_columns: List[ResultColumnsEntry]
+    """relates label names in the final SQL to a tuple of local
+    column/label name, ColumnElement object (if any) and
+    TypeEngine. CursorResult uses this for type processing and
+    column targeting"""
+
+    _textual_ordered_columns: bool = False
+    """tell the result object that the column names as rendered are important,
+    but they are also "ordered" vs. what is in the compiled object here.
+
+    As of 1.4.42 this condition is only present when the statement is a
+    TextualSelect, e.g. text("....").columns(...), where it is required
+    that the columns are considered positionally and not by name.
+
+    """
+
+    _ad_hoc_textual: bool = False
+    """tell the result that we encountered text() or '*' constructs in the
+    middle of the result columns, but we also have compiled columns, so
+    if the number of columns in cursor.description does not match how many
+    expressions we have, that means we can't rely on positional at all and
+    should match on name.
+
+    """
+
+    _ordered_columns: bool = True
+    """
+    if False, means we can't be sure the list of entries
+    in _result_columns is actually the rendered order.  Usually
+    True unless using an unordered TextualSelect.
+    """
+
+    _loose_column_name_matching: bool = False
+    """tell the result object that the SQL statement is textual, wants to match
+    up to Column objects, and may be using the ._tq_label in the SELECT rather
+    than the base name.
+
+    """
+
+    _numeric_binds: bool = False
+    """
+    True if paramstyle is "numeric".  This paramstyle is trickier than
+    all the others.
+
+    """
+
+    _render_postcompile: bool = False
+    """
+    whether to render out POSTCOMPILE params during the compile phase.
+
+    This attribute is used only for end-user invocation of stmt.compile();
+    it's never used for actual statement execution, where instead the
+    dialect internals access and render the internal postcompile structure
+    directly.
+
+    """
+
+    _post_compile_expanded_state: Optional[ExpandedState] = None
+    """When render_postcompile is used, the ``ExpandedState`` used to create
+    the "expanded" SQL is assigned here, and then used by the ``.params``
+    accessor and ``.construct_params()`` methods for their return values.
+
+    .. versionadded:: 2.0.0rc1
+
+    """
+
+    _pre_expanded_string: Optional[str] = None
+    """Stores the original string SQL before 'post_compile' is applied,
+    for cases where 'post_compile' were used.
+
+    """
+
+    _pre_expanded_positiontup: Optional[List[str]] = None
+
+    _insertmanyvalues: Optional[_InsertManyValues] = None
+
+    _insert_crud_params: Optional[crud._CrudParamSequence] = None
+
+    literal_execute_params: FrozenSet[BindParameter[Any]] = frozenset()
+    """bindparameter objects that are rendered as literal values at statement
+    execution time.
+
+    """
+
+    post_compile_params: FrozenSet[BindParameter[Any]] = frozenset()
+    """bindparameter objects that are rendered as bound parameter placeholders
+    at statement execution time.
+
+    """
+
+    escaped_bind_names: util.immutabledict[str, str] = util.EMPTY_DICT
+    """Late escaping of bound parameter names that has to be converted
+    to the original name when looking in the parameter dictionary.
+
+    """
+
+    has_out_parameters = False
+    """if True, there are bindparam() objects that have the isoutparam
+    flag set."""
+
+    postfetch_lastrowid = False
+    """if True, and this in insert, use cursor.lastrowid to populate
+    result.inserted_primary_key. """
+
+    _cache_key_bind_match: Optional[
+        Tuple[
+            Dict[
+                BindParameter[Any],
+                List[BindParameter[Any]],
+            ],
+            Dict[
+                str,
+                BindParameter[Any],
+            ],
+        ]
+    ] = None
+    """a mapping that will relate the BindParameter object we compile
+    to those that are part of the extracted collection of parameters
+    in the cache key, if we were given a cache key.
+
+    """
+
+    positiontup: Optional[List[str]] = None
+    """for a compiled construct that uses a positional paramstyle, will be
+    a sequence of strings, indicating the names of bound parameters in order.
+
+    This is used in order to render bound parameters in their correct order,
+    and is combined with the :attr:`_sql.Compiled.params` dictionary to
+    render parameters.
+
+    This sequence always contains the unescaped name of the parameters.
+
+    .. seealso::
+
+        :ref:`faq_sql_expression_string` - includes a usage example for
+        debugging use cases.
+
+    """
+    _values_bindparam: Optional[List[str]] = None
+
+    _visited_bindparam: Optional[List[str]] = None
+
+    inline: bool = False
+
+    ctes: Optional[MutableMapping[CTE, str]]
+
+    # Detect same CTE references - Dict[(level, name), cte]
+    # Level is required for supporting nesting
+    ctes_by_level_name: Dict[Tuple[int, str], CTE]
+
+    # To retrieve key/level in ctes_by_level_name -
+    # Dict[cte_reference, (level, cte_name, cte_opts)]
+    level_name_by_cte: Dict[CTE, Tuple[int, str, selectable._CTEOpts]]
+
+    ctes_recursive: bool
+
+    _post_compile_pattern = re.compile(r"__\[POSTCOMPILE_(\S+?)(~~.+?~~)?\]")
+    _pyformat_pattern = re.compile(r"%\(([^)]+?)\)s")
+    _positional_pattern = re.compile(
+        f"{_pyformat_pattern.pattern}|{_post_compile_pattern.pattern}"
+    )
+
+    @classmethod
+    def _init_compiler_cls(cls):
+        cls._init_bind_translate()
+
+    @classmethod
+    def _init_bind_translate(cls):
+        reg = re.escape("".join(cls.bindname_escape_characters))
+        cls._bind_translate_re = re.compile(f"[{reg}]")
+        cls._bind_translate_chars = cls.bindname_escape_characters
+
+    def __init__(
+        self,
+        dialect: Dialect,
+        statement: Optional[ClauseElement],
+        cache_key: Optional[CacheKey] = None,
+        column_keys: Optional[Sequence[str]] = None,
+        for_executemany: bool = False,
+        linting: Linting = NO_LINTING,
+        _supporting_against: Optional[SQLCompiler] = None,
+        **kwargs: Any,
+    ):
+        """Construct a new :class:`.SQLCompiler` object.
+
+        :param dialect: :class:`.Dialect` to be used
+
+        :param statement: :class:`_expression.ClauseElement` to be compiled
+
+        :param column_keys:  a list of column names to be compiled into an
+         INSERT or UPDATE statement.
+
+        :param for_executemany: whether INSERT / UPDATE statements should
+         expect that they are to be invoked in an "executemany" style,
+         which may impact how the statement will be expected to return the
+         values of defaults and autoincrement / sequences and similar.
+         Depending on the backend and driver in use, support for retrieving
+         these values may be disabled which means SQL expressions may
+         be rendered inline, RETURNING may not be rendered, etc.
+
+        :param kwargs: additional keyword arguments to be consumed by the
+         superclass.
+
+        """
+        self.column_keys = column_keys
+
+        self.cache_key = cache_key
+
+        if cache_key:
+            cksm = {b.key: b for b in cache_key[1]}
+            ckbm = {b: [b] for b in cache_key[1]}
+            self._cache_key_bind_match = (ckbm, cksm)
+
+        # compile INSERT/UPDATE defaults/sequences to expect executemany
+        # style execution, which may mean no pre-execute of defaults,
+        # or no RETURNING
+        self.for_executemany = for_executemany
+
+        self.linting = linting
+
+        # a dictionary of bind parameter keys to BindParameter
+        # instances.
+        self.binds = {}
+
+        # a dictionary of BindParameter instances to "compiled" names
+        # that are actually present in the generated SQL
+        self.bind_names = util.column_dict()
+
+        # stack which keeps track of nested SELECT statements
+        self.stack = []
+
+        self._result_columns = []
+
+        # true if the paramstyle is positional
+        self.positional = dialect.positional
+        if self.positional:
+            self._numeric_binds = nb = dialect.paramstyle.startswith("numeric")
+            if nb:
+                self._numeric_binds_identifier_char = (
+                    "$" if dialect.paramstyle == "numeric_dollar" else ":"
+                )
+
+            self.compilation_bindtemplate = _pyformat_template
+        else:
+            self.compilation_bindtemplate = BIND_TEMPLATES[dialect.paramstyle]
+
+        self.ctes = None
+
+        self.label_length = (
+            dialect.label_length or dialect.max_identifier_length
+        )
+
+        # a map which tracks "anonymous" identifiers that are created on
+        # the fly here
+        self.anon_map = prefix_anon_map()
+
+        # a map which tracks "truncated" names based on
+        # dialect.label_length or dialect.max_identifier_length
+        self.truncated_names: Dict[Tuple[str, str], str] = {}
+        self._truncated_counters: Dict[str, int] = {}
+
+        Compiled.__init__(self, dialect, statement, **kwargs)
+
+        if self.isinsert or self.isupdate or self.isdelete:
+            if TYPE_CHECKING:
+                assert isinstance(statement, UpdateBase)
+
+            if self.isinsert or self.isupdate:
+                if TYPE_CHECKING:
+                    assert isinstance(statement, ValuesBase)
+                if statement._inline:
+                    self.inline = True
+                elif self.for_executemany and (
+                    not self.isinsert
+                    or (
+                        self.dialect.insert_executemany_returning
+                        and statement._return_defaults
+                    )
+                ):
+                    self.inline = True
+
+        self.bindtemplate = BIND_TEMPLATES[dialect.paramstyle]
+
+        if _supporting_against:
+            self.__dict__.update(
+                {
+                    k: v
+                    for k, v in _supporting_against.__dict__.items()
+                    if k
+                    not in {
+                        "state",
+                        "dialect",
+                        "preparer",
+                        "positional",
+                        "_numeric_binds",
+                        "compilation_bindtemplate",
+                        "bindtemplate",
+                    }
+                }
+            )
+
+        if self.state is CompilerState.STRING_APPLIED:
+            if self.positional:
+                if self._numeric_binds:
+                    self._process_numeric()
+                else:
+                    self._process_positional()
+
+            if self._render_postcompile:
+                parameters = self.construct_params(
+                    escape_names=False,
+                    _no_postcompile=True,
+                )
+
+                self._process_parameters_for_postcompile(
+                    parameters, _populate_self=True
+                )
+
+    @property
+    def insert_single_values_expr(self) -> Optional[str]:
+        """When an INSERT is compiled with a single set of parameters inside
+        a VALUES expression, the string is assigned here, where it can be
+        used for insert batching schemes to rewrite the VALUES expression.
+
+        .. versionadded:: 1.3.8
+
+        .. versionchanged:: 2.0 This collection is no longer used by
+           SQLAlchemy's built-in dialects, in favor of the currently
+           internal ``_insertmanyvalues`` collection that is used only by
+           :class:`.SQLCompiler`.
+
+        """
+        if self._insertmanyvalues is None:
+            return None
+        else:
+            return self._insertmanyvalues.single_values_expr
+
+    @util.ro_memoized_property
+    def effective_returning(self) -> Optional[Sequence[ColumnElement[Any]]]:
+        """The effective "returning" columns for INSERT, UPDATE or DELETE.
+
+        This is either the so-called "implicit returning" columns which are
+        calculated by the compiler on the fly, or those present based on what's
+        present in ``self.statement._returning`` (expanded into individual
+        columns using the ``._all_selected_columns`` attribute) i.e. those set
+        explicitly using the :meth:`.UpdateBase.returning` method.
+
+        .. versionadded:: 2.0
+
+        """
+        if self.implicit_returning:
+            return self.implicit_returning
+        elif self.statement is not None and is_dml(self.statement):
+            return [
+                c
+                for c in self.statement._all_selected_columns
+                if is_column_element(c)
+            ]
+
+        else:
+            return None
+
+    @property
+    def returning(self):
+        """backwards compatibility; returns the
+        effective_returning collection.
+
+        """
+        return self.effective_returning
+
+    @property
+    def current_executable(self):
+        """Return the current 'executable' that is being compiled.
+
+        This is currently the :class:`_sql.Select`, :class:`_sql.Insert`,
+        :class:`_sql.Update`, :class:`_sql.Delete`,
+        :class:`_sql.CompoundSelect` object that is being compiled.
+        Specifically it's assigned to the ``self.stack`` list of elements.
+
+        When a statement like the above is being compiled, it normally
+        is also assigned to the ``.statement`` attribute of the
+        :class:`_sql.Compiler` object.   However, all SQL constructs are
+        ultimately nestable, and this attribute should never be consulted
+        by a ``visit_`` method, as it is not guaranteed to be assigned
+        nor guaranteed to correspond to the current statement being compiled.
+
+        .. versionadded:: 1.3.21
+
+            For compatibility with previous versions, use the following
+            recipe::
+
+                statement = getattr(self, "current_executable", False)
+                if statement is False:
+                    statement = self.stack[-1]["selectable"]
+
+            For versions 1.4 and above, ensure only .current_executable
+            is used; the format of "self.stack" may change.
+
+
+        """
+        try:
+            return self.stack[-1]["selectable"]
+        except IndexError as ie:
+            raise IndexError("Compiler does not have a stack entry") from ie
+
+    @property
+    def prefetch(self):
+        return list(self.insert_prefetch) + list(self.update_prefetch)
+
+    @util.memoized_property
+    def _global_attributes(self) -> Dict[Any, Any]:
+        return {}
+
+    @util.memoized_instancemethod
+    def _init_cte_state(self) -> MutableMapping[CTE, str]:
+        """Initialize collections related to CTEs only if
+        a CTE is located, to save on the overhead of
+        these collections otherwise.
+
+        """
+        # collect CTEs to tack on top of a SELECT
+        # To store the query to print - Dict[cte, text_query]
+        ctes: MutableMapping[CTE, str] = util.OrderedDict()
+        self.ctes = ctes
+
+        # Detect same CTE references - Dict[(level, name), cte]
+        # Level is required for supporting nesting
+        self.ctes_by_level_name = {}
+
+        # To retrieve key/level in ctes_by_level_name -
+        # Dict[cte_reference, (level, cte_name, cte_opts)]
+        self.level_name_by_cte = {}
+
+        self.ctes_recursive = False
+
+        return ctes
+
+    @contextlib.contextmanager
+    def _nested_result(self):
+        """special API to support the use case of 'nested result sets'"""
+        result_columns, ordered_columns = (
+            self._result_columns,
+            self._ordered_columns,
+        )
+        self._result_columns, self._ordered_columns = [], False
+
+        try:
+            if self.stack:
+                entry = self.stack[-1]
+                entry["need_result_map_for_nested"] = True
+            else:
+                entry = None
+            yield self._result_columns, self._ordered_columns
+        finally:
+            if entry:
+                entry.pop("need_result_map_for_nested")
+            self._result_columns, self._ordered_columns = (
+                result_columns,
+                ordered_columns,
+            )
+
+    def _process_positional(self):
+        assert not self.positiontup
+        assert self.state is CompilerState.STRING_APPLIED
+        assert not self._numeric_binds
+
+        if self.dialect.paramstyle == "format":
+            placeholder = "%s"
+        else:
+            assert self.dialect.paramstyle == "qmark"
+            placeholder = "?"
+
+        positions = []
+
+        def find_position(m: re.Match[str]) -> str:
+            normal_bind = m.group(1)
+            if normal_bind:
+                positions.append(normal_bind)
+                return placeholder
+            else:
+                # this a post-compile bind
+                positions.append(m.group(2))
+                return m.group(0)
+
+        self.string = re.sub(
+            self._positional_pattern, find_position, self.string
+        )
+
+        if self.escaped_bind_names:
+            reverse_escape = {v: k for k, v in self.escaped_bind_names.items()}
+            assert len(self.escaped_bind_names) == len(reverse_escape)
+            self.positiontup = [
+                reverse_escape.get(name, name) for name in positions
+            ]
+        else:
+            self.positiontup = positions
+
+        if self._insertmanyvalues:
+            positions = []
+
+            single_values_expr = re.sub(
+                self._positional_pattern,
+                find_position,
+                self._insertmanyvalues.single_values_expr,
+            )
+            insert_crud_params = [
+                (
+                    v[0],
+                    v[1],
+                    re.sub(self._positional_pattern, find_position, v[2]),
+                    v[3],
+                )
+                for v in self._insertmanyvalues.insert_crud_params
+            ]
+
+            self._insertmanyvalues = self._insertmanyvalues._replace(
+                single_values_expr=single_values_expr,
+                insert_crud_params=insert_crud_params,
+            )
+
+    def _process_numeric(self):
+        assert self._numeric_binds
+        assert self.state is CompilerState.STRING_APPLIED
+
+        num = 1
+        param_pos: Dict[str, str] = {}
+        order: Iterable[str]
+        if self._insertmanyvalues and self._values_bindparam is not None:
+            # bindparams that are not in values are always placed first.
+            # this avoids the need of changing them when using executemany
+            # values () ()
+            order = itertools.chain(
+                (
+                    name
+                    for name in self.bind_names.values()
+                    if name not in self._values_bindparam
+                ),
+                self.bind_names.values(),
+            )
+        else:
+            order = self.bind_names.values()
+
+        for bind_name in order:
+            if bind_name in param_pos:
+                continue
+            bind = self.binds[bind_name]
+            if (
+                bind in self.post_compile_params
+                or bind in self.literal_execute_params
+            ):
+                # set to None to just mark the in positiontup, it will not
+                # be replaced below.
+                param_pos[bind_name] = None  # type: ignore
+            else:
+                ph = f"{self._numeric_binds_identifier_char}{num}"
+                num += 1
+                param_pos[bind_name] = ph
+
+        self.next_numeric_pos = num
+
+        self.positiontup = list(param_pos)
+        if self.escaped_bind_names:
+            len_before = len(param_pos)
+            param_pos = {
+                self.escaped_bind_names.get(name, name): pos
+                for name, pos in param_pos.items()
+            }
+            assert len(param_pos) == len_before
+
+        # Can't use format here since % chars are not escaped.
+        self.string = self._pyformat_pattern.sub(
+            lambda m: param_pos[m.group(1)], self.string
+        )
+
+        if self._insertmanyvalues:
+            single_values_expr = (
+                # format is ok here since single_values_expr includes only
+                # place-holders
+                self._insertmanyvalues.single_values_expr
+                % param_pos
+            )
+            insert_crud_params = [
+                (v[0], v[1], "%s", v[3])
+                for v in self._insertmanyvalues.insert_crud_params
+            ]
+
+            self._insertmanyvalues = self._insertmanyvalues._replace(
+                # This has the numbers (:1, :2)
+                single_values_expr=single_values_expr,
+                # The single binds are instead %s so they can be formatted
+                insert_crud_params=insert_crud_params,
+            )
+
+    @util.memoized_property
+    def _bind_processors(
+        self,
+    ) -> MutableMapping[
+        str, Union[_BindProcessorType[Any], Sequence[_BindProcessorType[Any]]]
+    ]:
+        # mypy is not able to see the two value types as the above Union,
+        # it just sees "object".  don't know how to resolve
+        return {
+            key: value  # type: ignore
+            for key, value in (
+                (
+                    self.bind_names[bindparam],
+                    (
+                        bindparam.type._cached_bind_processor(self.dialect)
+                        if not bindparam.type._is_tuple_type
+                        else tuple(
+                            elem_type._cached_bind_processor(self.dialect)
+                            for elem_type in cast(
+                                TupleType, bindparam.type
+                            ).types
+                        )
+                    ),
+                )
+                for bindparam in self.bind_names
+            )
+            if value is not None
+        }
+
+    def is_subquery(self):
+        return len(self.stack) > 1
+
+    @property
+    def sql_compiler(self):
+        return self
+
+    def construct_expanded_state(
+        self,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        escape_names: bool = True,
+    ) -> ExpandedState:
+        """Return a new :class:`.ExpandedState` for a given parameter set.
+
+        For queries that use "expanding" or other late-rendered parameters,
+        this method will provide for both the finalized SQL string as well
+        as the parameters that would be used for a particular parameter set.
+
+        .. versionadded:: 2.0.0rc1
+
+        """
+        parameters = self.construct_params(
+            params,
+            escape_names=escape_names,
+            _no_postcompile=True,
+        )
+        return self._process_parameters_for_postcompile(
+            parameters,
+        )
+
+    def construct_params(
+        self,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        extracted_parameters: Optional[Sequence[BindParameter[Any]]] = None,
+        escape_names: bool = True,
+        _group_number: Optional[int] = None,
+        _check: bool = True,
+        _no_postcompile: bool = False,
+    ) -> _MutableCoreSingleExecuteParams:
+        """return a dictionary of bind parameter keys and values"""
+
+        if self._render_postcompile and not _no_postcompile:
+            assert self._post_compile_expanded_state is not None
+            if not params:
+                return dict(self._post_compile_expanded_state.parameters)
+            else:
+                raise exc.InvalidRequestError(
+                    "can't construct new parameters when render_postcompile "
+                    "is used; the statement is hard-linked to the original "
+                    "parameters.  Use construct_expanded_state to generate a "
+                    "new statement and parameters."
+                )
+
+        has_escaped_names = escape_names and bool(self.escaped_bind_names)
+
+        if extracted_parameters:
+            # related the bound parameters collected in the original cache key
+            # to those collected in the incoming cache key.  They will not have
+            # matching names but they will line up positionally in the same
+            # way.   The parameters present in self.bind_names may be clones of
+            # these original cache key params in the case of DML but the .key
+            # will be guaranteed to match.
+            if self.cache_key is None:
+                raise exc.CompileError(
+                    "This compiled object has no original cache key; "
+                    "can't pass extracted_parameters to construct_params"
+                )
+            else:
+                orig_extracted = self.cache_key[1]
+
+            ckbm_tuple = self._cache_key_bind_match
+            assert ckbm_tuple is not None
+            ckbm, _ = ckbm_tuple
+            resolved_extracted = {
+                bind: extracted
+                for b, extracted in zip(orig_extracted, extracted_parameters)
+                for bind in ckbm[b]
+            }
+        else:
+            resolved_extracted = None
+
+        if params:
+            pd = {}
+            for bindparam, name in self.bind_names.items():
+                escaped_name = (
+                    self.escaped_bind_names.get(name, name)
+                    if has_escaped_names
+                    else name
+                )
+
+                if bindparam.key in params:
+                    pd[escaped_name] = params[bindparam.key]
+                elif name in params:
+                    pd[escaped_name] = params[name]
+
+                elif _check and bindparam.required:
+                    if _group_number:
+                        raise exc.InvalidRequestError(
+                            "A value is required for bind parameter %r, "
+                            "in parameter group %d"
+                            % (bindparam.key, _group_number),
+                            code="cd3x",
+                        )
+                    else:
+                        raise exc.InvalidRequestError(
+                            "A value is required for bind parameter %r"
+                            % bindparam.key,
+                            code="cd3x",
+                        )
+                else:
+                    if resolved_extracted:
+                        value_param = resolved_extracted.get(
+                            bindparam, bindparam
+                        )
+                    else:
+                        value_param = bindparam
+
+                    if bindparam.callable:
+                        pd[escaped_name] = value_param.effective_value
+                    else:
+                        pd[escaped_name] = value_param.value
+            return pd
+        else:
+            pd = {}
+            for bindparam, name in self.bind_names.items():
+                escaped_name = (
+                    self.escaped_bind_names.get(name, name)
+                    if has_escaped_names
+                    else name
+                )
+
+                if _check and bindparam.required:
+                    if _group_number:
+                        raise exc.InvalidRequestError(
+                            "A value is required for bind parameter %r, "
+                            "in parameter group %d"
+                            % (bindparam.key, _group_number),
+                            code="cd3x",
+                        )
+                    else:
+                        raise exc.InvalidRequestError(
+                            "A value is required for bind parameter %r"
+                            % bindparam.key,
+                            code="cd3x",
+                        )
+
+                if resolved_extracted:
+                    value_param = resolved_extracted.get(bindparam, bindparam)
+                else:
+                    value_param = bindparam
+
+                if bindparam.callable:
+                    pd[escaped_name] = value_param.effective_value
+                else:
+                    pd[escaped_name] = value_param.value
+
+            return pd
+
+    @util.memoized_instancemethod
+    def _get_set_input_sizes_lookup(self):
+        dialect = self.dialect
+
+        include_types = dialect.include_set_input_sizes
+        exclude_types = dialect.exclude_set_input_sizes
+
+        dbapi = dialect.dbapi
+
+        def lookup_type(typ):
+            dbtype = typ._unwrapped_dialect_impl(dialect).get_dbapi_type(dbapi)
+
+            if (
+                dbtype is not None
+                and (exclude_types is None or dbtype not in exclude_types)
+                and (include_types is None or dbtype in include_types)
+            ):
+                return dbtype
+            else:
+                return None
+
+        inputsizes = {}
+
+        literal_execute_params = self.literal_execute_params
+
+        for bindparam in self.bind_names:
+            if bindparam in literal_execute_params:
+                continue
+
+            if bindparam.type._is_tuple_type:
+                inputsizes[bindparam] = [
+                    lookup_type(typ)
+                    for typ in cast(TupleType, bindparam.type).types
+                ]
+            else:
+                inputsizes[bindparam] = lookup_type(bindparam.type)
+
+        return inputsizes
+
+    @property
+    def params(self):
+        """Return the bind param dictionary embedded into this
+        compiled object, for those values that are present.
+
+        .. seealso::
+
+            :ref:`faq_sql_expression_string` - includes a usage example for
+            debugging use cases.
+
+        """
+        return self.construct_params(_check=False)
+
+    def _process_parameters_for_postcompile(
+        self,
+        parameters: _MutableCoreSingleExecuteParams,
+        _populate_self: bool = False,
+    ) -> ExpandedState:
+        """handle special post compile parameters.
+
+        These include:
+
+        * "expanding" parameters -typically IN tuples that are rendered
+          on a per-parameter basis for an otherwise fixed SQL statement string.
+
+        * literal_binds compiled with the literal_execute flag.  Used for
+          things like SQL Server "TOP N" where the driver does not accommodate
+          N as a bound parameter.
+
+        """
+
+        expanded_parameters = {}
+        new_positiontup: Optional[List[str]]
+
+        pre_expanded_string = self._pre_expanded_string
+        if pre_expanded_string is None:
+            pre_expanded_string = self.string
+
+        if self.positional:
+            new_positiontup = []
+
+            pre_expanded_positiontup = self._pre_expanded_positiontup
+            if pre_expanded_positiontup is None:
+                pre_expanded_positiontup = self.positiontup
+
+        else:
+            new_positiontup = pre_expanded_positiontup = None
+
+        processors = self._bind_processors
+        single_processors = cast(
+            "Mapping[str, _BindProcessorType[Any]]", processors
+        )
+        tuple_processors = cast(
+            "Mapping[str, Sequence[_BindProcessorType[Any]]]", processors
+        )
+
+        new_processors: Dict[str, _BindProcessorType[Any]] = {}
+
+        replacement_expressions: Dict[str, Any] = {}
+        to_update_sets: Dict[str, Any] = {}
+
+        # notes:
+        # *unescaped* parameter names in:
+        # self.bind_names, self.binds, self._bind_processors, self.positiontup
+        #
+        # *escaped* parameter names in:
+        # construct_params(), replacement_expressions
+
+        numeric_positiontup: Optional[List[str]] = None
+
+        if self.positional and pre_expanded_positiontup is not None:
+            names: Iterable[str] = pre_expanded_positiontup
+            if self._numeric_binds:
+                numeric_positiontup = []
+        else:
+            names = self.bind_names.values()
+
+        ebn = self.escaped_bind_names
+        for name in names:
+            escaped_name = ebn.get(name, name) if ebn else name
+            parameter = self.binds[name]
+
+            if parameter in self.literal_execute_params:
+                if escaped_name not in replacement_expressions:
+                    replacement_expressions[escaped_name] = (
+                        self.render_literal_bindparam(
+                            parameter,
+                            render_literal_value=parameters.pop(escaped_name),
+                        )
+                    )
+                continue
+
+            if parameter in self.post_compile_params:
+                if escaped_name in replacement_expressions:
+                    to_update = to_update_sets[escaped_name]
+                    values = None
+                else:
+                    # we are removing the parameter from parameters
+                    # because it is a list value, which is not expected by
+                    # TypeEngine objects that would otherwise be asked to
+                    # process it. the single name is being replaced with
+                    # individual numbered parameters for each value in the
+                    # param.
+                    #
+                    # note we are also inserting *escaped* parameter names
+                    # into the given dictionary.   default dialect will
+                    # use these param names directly as they will not be
+                    # in the escaped_bind_names dictionary.
+                    values = parameters.pop(name)
+
+                    leep_res = self._literal_execute_expanding_parameter(
+                        escaped_name, parameter, values
+                    )
+                    (to_update, replacement_expr) = leep_res
+
+                    to_update_sets[escaped_name] = to_update
+                    replacement_expressions[escaped_name] = replacement_expr
+
+                if not parameter.literal_execute:
+                    parameters.update(to_update)
+                    if parameter.type._is_tuple_type:
+                        assert values is not None
+                        new_processors.update(
+                            (
+                                "%s_%s_%s" % (name, i, j),
+                                tuple_processors[name][j - 1],
+                            )
+                            for i, tuple_element in enumerate(values, 1)
+                            for j, _ in enumerate(tuple_element, 1)
+                            if name in tuple_processors
+                            and tuple_processors[name][j - 1] is not None
+                        )
+                    else:
+                        new_processors.update(
+                            (key, single_processors[name])
+                            for key, _ in to_update
+                            if name in single_processors
+                        )
+                    if numeric_positiontup is not None:
+                        numeric_positiontup.extend(
+                            name for name, _ in to_update
+                        )
+                    elif new_positiontup is not None:
+                        # to_update has escaped names, but that's ok since
+                        # these are new names, that aren't in the
+                        # escaped_bind_names dict.
+                        new_positiontup.extend(name for name, _ in to_update)
+                    expanded_parameters[name] = [
+                        expand_key for expand_key, _ in to_update
+                    ]
+            elif new_positiontup is not None:
+                new_positiontup.append(name)
+
+        def process_expanding(m):
+            key = m.group(1)
+            expr = replacement_expressions[key]
+
+            # if POSTCOMPILE included a bind_expression, render that
+            # around each element
+            if m.group(2):
+                tok = m.group(2).split("~~")
+                be_left, be_right = tok[1], tok[3]
+                expr = ", ".join(
+                    "%s%s%s" % (be_left, exp, be_right)
+                    for exp in expr.split(", ")
+                )
+            return expr
+
+        statement = re.sub(
+            self._post_compile_pattern, process_expanding, pre_expanded_string
+        )
+
+        if numeric_positiontup is not None:
+            assert new_positiontup is not None
+            param_pos = {
+                key: f"{self._numeric_binds_identifier_char}{num}"
+                for num, key in enumerate(
+                    numeric_positiontup, self.next_numeric_pos
+                )
+            }
+            # Can't use format here since % chars are not escaped.
+            statement = self._pyformat_pattern.sub(
+                lambda m: param_pos[m.group(1)], statement
+            )
+            new_positiontup.extend(numeric_positiontup)
+
+        expanded_state = ExpandedState(
+            statement,
+            parameters,
+            new_processors,
+            new_positiontup,
+            expanded_parameters,
+        )
+
+        if _populate_self:
+            # this is for the "render_postcompile" flag, which is not
+            # otherwise used internally and is for end-user debugging and
+            # special use cases.
+            self._pre_expanded_string = pre_expanded_string
+            self._pre_expanded_positiontup = pre_expanded_positiontup
+            self.string = expanded_state.statement
+            self.positiontup = (
+                list(expanded_state.positiontup or ())
+                if self.positional
+                else None
+            )
+            self._post_compile_expanded_state = expanded_state
+
+        return expanded_state
+
+    @util.preload_module("sqlalchemy.engine.cursor")
+    def _create_result_map(self):
+        """utility method used for unit tests only."""
+        cursor = util.preloaded.engine_cursor
+        return cursor.CursorResultMetaData._create_description_match_map(
+            self._result_columns
+        )
+
+    # assigned by crud.py for insert/update statements
+    _get_bind_name_for_col: _BindNameForColProtocol
+
+    @util.memoized_property
+    def _within_exec_param_key_getter(self) -> Callable[[Any], str]:
+        getter = self._get_bind_name_for_col
+        return getter
+
+    @util.memoized_property
+    @util.preload_module("sqlalchemy.engine.result")
+    def _inserted_primary_key_from_lastrowid_getter(self):
+        result = util.preloaded.engine_result
+
+        param_key_getter = self._within_exec_param_key_getter
+
+        assert self.compile_state is not None
+        statement = self.compile_state.statement
+
+        if TYPE_CHECKING:
+            assert isinstance(statement, Insert)
+
+        table = statement.table
+
+        getters = [
+            (operator.methodcaller("get", param_key_getter(col), None), col)
+            for col in table.primary_key
+        ]
+
+        autoinc_getter = None
+        autoinc_col = table._autoincrement_column
+        if autoinc_col is not None:
+            # apply type post processors to the lastrowid
+            lastrowid_processor = autoinc_col.type._cached_result_processor(
+                self.dialect, None
+            )
+            autoinc_key = param_key_getter(autoinc_col)
+
+            # if a bind value is present for the autoincrement column
+            # in the parameters, we need to do the logic dictated by
+            # #7998; honor a non-None user-passed parameter over lastrowid.
+            # previously in the 1.4 series we weren't fetching lastrowid
+            # at all if the key were present in the parameters
+            if autoinc_key in self.binds:
+
+                def _autoinc_getter(lastrowid, parameters):
+                    param_value = parameters.get(autoinc_key, lastrowid)
+                    if param_value is not None:
+                        # they supplied non-None parameter, use that.
+                        # SQLite at least is observed to return the wrong
+                        # cursor.lastrowid for INSERT..ON CONFLICT so it
+                        # can't be used in all cases
+                        return param_value
+                    else:
+                        # use lastrowid
+                        return lastrowid
+
+                # work around mypy https://github.com/python/mypy/issues/14027
+                autoinc_getter = _autoinc_getter
+
+        else:
+            lastrowid_processor = None
+
+        row_fn = result.result_tuple([col.key for col in table.primary_key])
+
+        def get(lastrowid, parameters):
+            """given cursor.lastrowid value and the parameters used for INSERT,
+            return a "row" that represents the primary key, either by
+            using the "lastrowid" or by extracting values from the parameters
+            that were sent along with the INSERT.
+
+            """
+            if lastrowid_processor is not None:
+                lastrowid = lastrowid_processor(lastrowid)
+
+            if lastrowid is None:
+                return row_fn(getter(parameters) for getter, col in getters)
+            else:
+                return row_fn(
+                    (
+                        (
+                            autoinc_getter(lastrowid, parameters)
+                            if autoinc_getter is not None
+                            else lastrowid
+                        )
+                        if col is autoinc_col
+                        else getter(parameters)
+                    )
+                    for getter, col in getters
+                )
+
+        return get
+
+    @util.memoized_property
+    @util.preload_module("sqlalchemy.engine.result")
+    def _inserted_primary_key_from_returning_getter(self):
+        if typing.TYPE_CHECKING:
+            from ..engine import result
+        else:
+            result = util.preloaded.engine_result
+
+        assert self.compile_state is not None
+        statement = self.compile_state.statement
+
+        if TYPE_CHECKING:
+            assert isinstance(statement, Insert)
+
+        param_key_getter = self._within_exec_param_key_getter
+        table = statement.table
+
+        returning = self.implicit_returning
+        assert returning is not None
+        ret = {col: idx for idx, col in enumerate(returning)}
+
+        getters = cast(
+            "List[Tuple[Callable[[Any], Any], bool]]",
+            [
+                (
+                    (operator.itemgetter(ret[col]), True)
+                    if col in ret
+                    else (
+                        operator.methodcaller(
+                            "get", param_key_getter(col), None
+                        ),
+                        False,
+                    )
+                )
+                for col in table.primary_key
+            ],
+        )
+
+        row_fn = result.result_tuple([col.key for col in table.primary_key])
+
+        def get(row, parameters):
+            return row_fn(
+                getter(row) if use_row else getter(parameters)
+                for getter, use_row in getters
+            )
+
+        return get
+
+    def default_from(self):
+        """Called when a SELECT statement has no froms, and no FROM clause is
+        to be appended.
+
+        Gives Oracle Database a chance to tack on a ``FROM DUAL`` to the string
+        output.
+
+        """
+        return ""
+
+    def visit_override_binds(self, override_binds, **kw):
+        """SQL compile the nested element of an _OverrideBinds with
+        bindparams swapped out.
+
+        The _OverrideBinds is not normally expected to be compiled; it
+        is meant to be used when an already cached statement is to be used,
+        the compilation was already performed, and only the bound params should
+        be swapped in at execution time.
+
+        However, there are test cases that exericise this object, and
+        additionally the ORM subquery loader is known to feed in expressions
+        which include this construct into new queries (discovered in #11173),
+        so it has to do the right thing at compile time as well.
+
+        """
+
+        # get SQL text first
+        sqltext = override_binds.element._compiler_dispatch(self, **kw)
+
+        # for a test compile that is not for caching, change binds after the
+        # fact.  note that we don't try to
+        # swap the bindparam as we compile, because our element may be
+        # elsewhere in the statement already (e.g. a subquery or perhaps a
+        # CTE) and was already visited / compiled. See
+        # test_relationship_criteria.py ->
+        #    test_selectinload_local_criteria_subquery
+        for k in override_binds.translate:
+            if k not in self.binds:
+                continue
+            bp = self.binds[k]
+
+            # so this would work, just change the value of bp in place.
+            # but we dont want to mutate things outside.
+            # bp.value = override_binds.translate[bp.key]
+            # continue
+
+            # instead, need to replace bp with new_bp or otherwise accommodate
+            # in all internal collections
+            new_bp = bp._with_value(
+                override_binds.translate[bp.key],
+                maintain_key=True,
+                required=False,
+            )
+
+            name = self.bind_names[bp]
+            self.binds[k] = self.binds[name] = new_bp
+            self.bind_names[new_bp] = name
+            self.bind_names.pop(bp, None)
+
+            if bp in self.post_compile_params:
+                self.post_compile_params |= {new_bp}
+            if bp in self.literal_execute_params:
+                self.literal_execute_params |= {new_bp}
+
+            ckbm_tuple = self._cache_key_bind_match
+            if ckbm_tuple:
+                ckbm, cksm = ckbm_tuple
+                for bp in bp._cloned_set:
+                    if bp.key in cksm:
+                        cb = cksm[bp.key]
+                        ckbm[cb].append(new_bp)
+
+        return sqltext
+
+    def visit_grouping(self, grouping, asfrom=False, **kwargs):
+        return "(" + grouping.element._compiler_dispatch(self, **kwargs) + ")"
+
+    def visit_select_statement_grouping(self, grouping, **kwargs):
+        return "(" + grouping.element._compiler_dispatch(self, **kwargs) + ")"
+
+    def visit_label_reference(
+        self, element, within_columns_clause=False, **kwargs
+    ):
+        if self.stack and self.dialect.supports_simple_order_by_label:
+            try:
+                compile_state = cast(
+                    "Union[SelectState, CompoundSelectState]",
+                    self.stack[-1]["compile_state"],
+                )
+            except KeyError as ke:
+                raise exc.CompileError(
+                    "Can't resolve label reference for ORDER BY / "
+                    "GROUP BY / DISTINCT etc."
+                ) from ke
+
+            (
+                with_cols,
+                only_froms,
+                only_cols,
+            ) = compile_state._label_resolve_dict
+            if within_columns_clause:
+                resolve_dict = only_froms
+            else:
+                resolve_dict = only_cols
+
+            # this can be None in the case that a _label_reference()
+            # were subject to a replacement operation, in which case
+            # the replacement of the Label element may have changed
+            # to something else like a ColumnClause expression.
+            order_by_elem = element.element._order_by_label_element
+
+            if (
+                order_by_elem is not None
+                and order_by_elem.name in resolve_dict
+                and order_by_elem.shares_lineage(
+                    resolve_dict[order_by_elem.name]
+                )
+            ):
+                kwargs["render_label_as_label"] = (
+                    element.element._order_by_label_element
+                )
+        return self.process(
+            element.element,
+            within_columns_clause=within_columns_clause,
+            **kwargs,
+        )
+
+    def visit_textual_label_reference(
+        self, element, within_columns_clause=False, **kwargs
+    ):
+        if not self.stack:
+            # compiling the element outside of the context of a SELECT
+            return self.process(element._text_clause)
+
+        try:
+            compile_state = cast(
+                "Union[SelectState, CompoundSelectState]",
+                self.stack[-1]["compile_state"],
+            )
+        except KeyError as ke:
+            coercions._no_text_coercion(
+                element.element,
+                extra=(
+                    "Can't resolve label reference for ORDER BY / "
+                    "GROUP BY / DISTINCT etc."
+                ),
+                exc_cls=exc.CompileError,
+                err=ke,
+            )
+
+        with_cols, only_froms, only_cols = compile_state._label_resolve_dict
+        try:
+            if within_columns_clause:
+                col = only_froms[element.element]
+            else:
+                col = with_cols[element.element]
+        except KeyError as err:
+            coercions._no_text_coercion(
+                element.element,
+                extra=(
+                    "Can't resolve label reference for ORDER BY / "
+                    "GROUP BY / DISTINCT etc."
+                ),
+                exc_cls=exc.CompileError,
+                err=err,
+            )
+        else:
+            kwargs["render_label_as_label"] = col
+            return self.process(
+                col, within_columns_clause=within_columns_clause, **kwargs
+            )
+
+    def visit_label(
+        self,
+        label,
+        add_to_result_map=None,
+        within_label_clause=False,
+        within_columns_clause=False,
+        render_label_as_label=None,
+        result_map_targets=(),
+        **kw,
+    ):
+        # only render labels within the columns clause
+        # or ORDER BY clause of a select.  dialect-specific compilers
+        # can modify this behavior.
+        render_label_with_as = (
+            within_columns_clause and not within_label_clause
+        )
+        render_label_only = render_label_as_label is label
+
+        if render_label_only or render_label_with_as:
+            if isinstance(label.name, elements._truncated_label):
+                labelname = self._truncated_identifier("colident", label.name)
+            else:
+                labelname = label.name
+
+        if render_label_with_as:
+            if add_to_result_map is not None:
+                add_to_result_map(
+                    labelname,
+                    label.name,
+                    (label, labelname) + label._alt_names + result_map_targets,
+                    label.type,
+                )
+            return (
+                label.element._compiler_dispatch(
+                    self,
+                    within_columns_clause=True,
+                    within_label_clause=True,
+                    **kw,
+                )
+                + OPERATORS[operators.as_]
+                + self.preparer.format_label(label, labelname)
+            )
+        elif render_label_only:
+            return self.preparer.format_label(label, labelname)
+        else:
+            return label.element._compiler_dispatch(
+                self, within_columns_clause=False, **kw
+            )
+
+    def _fallback_column_name(self, column):
+        raise exc.CompileError(
+            "Cannot compile Column object until its 'name' is assigned."
+        )
+
+    def visit_lambda_element(self, element, **kw):
+        sql_element = element._resolved
+        return self.process(sql_element, **kw)
+
+    def visit_column(
+        self,
+        column: ColumnClause[Any],
+        add_to_result_map: Optional[_ResultMapAppender] = None,
+        include_table: bool = True,
+        result_map_targets: Tuple[Any, ...] = (),
+        ambiguous_table_name_map: Optional[_AmbiguousTableNameMap] = None,
+        **kwargs: Any,
+    ) -> str:
+        name = orig_name = column.name
+        if name is None:
+            name = self._fallback_column_name(column)
+
+        is_literal = column.is_literal
+        if not is_literal and isinstance(name, elements._truncated_label):
+            name = self._truncated_identifier("colident", name)
+
+        if add_to_result_map is not None:
+            targets = (column, name, column.key) + result_map_targets
+            if column._tq_label:
+                targets += (column._tq_label,)
+
+            add_to_result_map(name, orig_name, targets, column.type)
+
+        if is_literal:
+            # note we are not currently accommodating for
+            # literal_column(quoted_name('ident', True)) here
+            name = self.escape_literal_column(name)
+        else:
+            name = self.preparer.quote(name)
+        table = column.table
+        if table is None or not include_table or not table.named_with_column:
+            return name
+        else:
+            effective_schema = self.preparer.schema_for_object(table)
+
+            if effective_schema:
+                schema_prefix = (
+                    self.preparer.quote_schema(effective_schema) + "."
+                )
+            else:
+                schema_prefix = ""
+
+            if TYPE_CHECKING:
+                assert isinstance(table, NamedFromClause)
+            tablename = table.name
+
+            if (
+                not effective_schema
+                and ambiguous_table_name_map
+                and tablename in ambiguous_table_name_map
+            ):
+                tablename = ambiguous_table_name_map[tablename]
+
+            if isinstance(tablename, elements._truncated_label):
+                tablename = self._truncated_identifier("alias", tablename)
+
+            return schema_prefix + self.preparer.quote(tablename) + "." + name
+
+    def visit_collation(self, element, **kw):
+        return self.preparer.format_collation(element.collation)
+
+    def visit_fromclause(self, fromclause, **kwargs):
+        return fromclause.name
+
+    def visit_index(self, index, **kwargs):
+        return index.name
+
+    def visit_typeclause(self, typeclause, **kw):
+        kw["type_expression"] = typeclause
+        kw["identifier_preparer"] = self.preparer
+        return self.dialect.type_compiler_instance.process(
+            typeclause.type, **kw
+        )
+
+    def post_process_text(self, text):
+        if self.preparer._double_percents:
+            text = text.replace("%", "%%")
+        return text
+
+    def escape_literal_column(self, text):
+        if self.preparer._double_percents:
+            text = text.replace("%", "%%")
+        return text
+
+    def visit_textclause(self, textclause, add_to_result_map=None, **kw):
+        def do_bindparam(m):
+            name = m.group(1)
+            if name in textclause._bindparams:
+                return self.process(textclause._bindparams[name], **kw)
+            else:
+                return self.bindparam_string(name, **kw)
+
+        if not self.stack:
+            self.isplaintext = True
+
+        if add_to_result_map:
+            # text() object is present in the columns clause of a
+            # select().   Add a no-name entry to the result map so that
+            # row[text()] produces a result
+            add_to_result_map(None, None, (textclause,), sqltypes.NULLTYPE)
+
+        # un-escape any \:params
+        return BIND_PARAMS_ESC.sub(
+            lambda m: m.group(1),
+            BIND_PARAMS.sub(
+                do_bindparam, self.post_process_text(textclause.text)
+            ),
+        )
+
+    def visit_textual_select(
+        self, taf, compound_index=None, asfrom=False, **kw
+    ):
+        toplevel = not self.stack
+        entry = self._default_stack_entry if toplevel else self.stack[-1]
+
+        new_entry: _CompilerStackEntry = {
+            "correlate_froms": set(),
+            "asfrom_froms": set(),
+            "selectable": taf,
+        }
+        self.stack.append(new_entry)
+
+        if taf._independent_ctes:
+            self._dispatch_independent_ctes(taf, kw)
+
+        populate_result_map = (
+            toplevel
+            or (
+                compound_index == 0
+                and entry.get("need_result_map_for_compound", False)
+            )
+            or entry.get("need_result_map_for_nested", False)
+        )
+
+        if populate_result_map:
+            self._ordered_columns = self._textual_ordered_columns = (
+                taf.positional
+            )
+
+            # enable looser result column matching when the SQL text links to
+            # Column objects by name only
+            self._loose_column_name_matching = not taf.positional and bool(
+                taf.column_args
+            )
+
+            for c in taf.column_args:
+                self.process(
+                    c,
+                    within_columns_clause=True,
+                    add_to_result_map=self._add_to_result_map,
+                )
+
+        text = self.process(taf.element, **kw)
+        if self.ctes:
+            nesting_level = len(self.stack) if not toplevel else None
+            text = self._render_cte_clause(nesting_level=nesting_level) + text
+
+        self.stack.pop(-1)
+
+        return text
+
+    def visit_null(self, expr, **kw):
+        return "NULL"
+
+    def visit_true(self, expr, **kw):
+        if self.dialect.supports_native_boolean:
+            return "true"
+        else:
+            return "1"
+
+    def visit_false(self, expr, **kw):
+        if self.dialect.supports_native_boolean:
+            return "false"
+        else:
+            return "0"
+
+    def _generate_delimited_list(self, elements, separator, **kw):
+        return separator.join(
+            s
+            for s in (c._compiler_dispatch(self, **kw) for c in elements)
+            if s
+        )
+
+    def _generate_delimited_and_list(self, clauses, **kw):
+        lcc, clauses = elements.BooleanClauseList._process_clauses_for_boolean(
+            operators.and_,
+            elements.True_._singleton,
+            elements.False_._singleton,
+            clauses,
+        )
+        if lcc == 1:
+            return clauses[0]._compiler_dispatch(self, **kw)
+        else:
+            separator = OPERATORS[operators.and_]
+            return separator.join(
+                s
+                for s in (c._compiler_dispatch(self, **kw) for c in clauses)
+                if s
+            )
+
+    def visit_tuple(self, clauselist, **kw):
+        return "(%s)" % self.visit_clauselist(clauselist, **kw)
+
+    def visit_clauselist(self, clauselist, **kw):
+        sep = clauselist.operator
+        if sep is None:
+            sep = " "
+        else:
+            sep = OPERATORS[clauselist.operator]
+
+        return self._generate_delimited_list(clauselist.clauses, sep, **kw)
+
+    def visit_expression_clauselist(self, clauselist, **kw):
+        operator_ = clauselist.operator
+
+        disp = self._get_operator_dispatch(
+            operator_, "expression_clauselist", None
+        )
+        if disp:
+            return disp(clauselist, operator_, **kw)
+
+        try:
+            opstring = OPERATORS[operator_]
+        except KeyError as err:
+            raise exc.UnsupportedCompilationError(self, operator_) from err
+        else:
+            kw["_in_operator_expression"] = True
+            return self._generate_delimited_list(
+                clauselist.clauses, opstring, **kw
+            )
+
+    def visit_case(self, clause, **kwargs):
+        x = "CASE "
+        if clause.value is not None:
+            x += clause.value._compiler_dispatch(self, **kwargs) + " "
+        for cond, result in clause.whens:
+            x += (
+                "WHEN "
+                + cond._compiler_dispatch(self, **kwargs)
+                + " THEN "
+                + result._compiler_dispatch(self, **kwargs)
+                + " "
+            )
+        if clause.else_ is not None:
+            x += (
+                "ELSE " + clause.else_._compiler_dispatch(self, **kwargs) + " "
+            )
+        x += "END"
+        return x
+
+    def visit_type_coerce(self, type_coerce, **kw):
+        return type_coerce.typed_expression._compiler_dispatch(self, **kw)
+
+    def visit_cast(self, cast, **kwargs):
+        type_clause = cast.typeclause._compiler_dispatch(self, **kwargs)
+        match = re.match("(.*)( COLLATE .*)", type_clause)
+        return "CAST(%s AS %s)%s" % (
+            cast.clause._compiler_dispatch(self, **kwargs),
+            match.group(1) if match else type_clause,
+            match.group(2) if match else "",
+        )
+
+    def _format_frame_clause(self, range_, **kw):
+        return "%s AND %s" % (
+            (
+                "UNBOUNDED PRECEDING"
+                if range_[0] is elements.RANGE_UNBOUNDED
+                else (
+                    "CURRENT ROW"
+                    if range_[0] is elements.RANGE_CURRENT
+                    else (
+                        "%s PRECEDING"
+                        % (
+                            self.process(
+                                elements.literal(abs(range_[0])), **kw
+                            ),
+                        )
+                        if range_[0] < 0
+                        else "%s FOLLOWING"
+                        % (self.process(elements.literal(range_[0]), **kw),)
+                    )
+                )
+            ),
+            (
+                "UNBOUNDED FOLLOWING"
+                if range_[1] is elements.RANGE_UNBOUNDED
+                else (
+                    "CURRENT ROW"
+                    if range_[1] is elements.RANGE_CURRENT
+                    else (
+                        "%s PRECEDING"
+                        % (
+                            self.process(
+                                elements.literal(abs(range_[1])), **kw
+                            ),
+                        )
+                        if range_[1] < 0
+                        else "%s FOLLOWING"
+                        % (self.process(elements.literal(range_[1]), **kw),)
+                    )
+                )
+            ),
+        )
+
+    def visit_over(self, over, **kwargs):
+        text = over.element._compiler_dispatch(self, **kwargs)
+        if over.range_:
+            range_ = "RANGE BETWEEN %s" % self._format_frame_clause(
+                over.range_, **kwargs
+            )
+        elif over.rows:
+            range_ = "ROWS BETWEEN %s" % self._format_frame_clause(
+                over.rows, **kwargs
+            )
+        else:
+            range_ = None
+
+        return "%s OVER (%s)" % (
+            text,
+            " ".join(
+                [
+                    "%s BY %s"
+                    % (word, clause._compiler_dispatch(self, **kwargs))
+                    for word, clause in (
+                        ("PARTITION", over.partition_by),
+                        ("ORDER", over.order_by),
+                    )
+                    if clause is not None and len(clause)
+                ]
+                + ([range_] if range_ else [])
+            ),
+        )
+
+    def visit_withingroup(self, withingroup, **kwargs):
+        return "%s WITHIN GROUP (ORDER BY %s)" % (
+            withingroup.element._compiler_dispatch(self, **kwargs),
+            withingroup.order_by._compiler_dispatch(self, **kwargs),
+        )
+
+    def visit_funcfilter(self, funcfilter, **kwargs):
+        return "%s FILTER (WHERE %s)" % (
+            funcfilter.func._compiler_dispatch(self, **kwargs),
+            funcfilter.criterion._compiler_dispatch(self, **kwargs),
+        )
+
+    def visit_extract(self, extract, **kwargs):
+        field = self.extract_map.get(extract.field, extract.field)
+        return "EXTRACT(%s FROM %s)" % (
+            field,
+            extract.expr._compiler_dispatch(self, **kwargs),
+        )
+
+    def visit_scalar_function_column(self, element, **kw):
+        compiled_fn = self.visit_function(element.fn, **kw)
+        compiled_col = self.visit_column(element, **kw)
+        return "(%s).%s" % (compiled_fn, compiled_col)
+
+    def visit_function(
+        self,
+        func: Function[Any],
+        add_to_result_map: Optional[_ResultMapAppender] = None,
+        **kwargs: Any,
+    ) -> str:
+        if add_to_result_map is not None:
+            add_to_result_map(func.name, func.name, (func.name,), func.type)
+
+        disp = getattr(self, "visit_%s_func" % func.name.lower(), None)
+
+        text: str
+
+        if disp:
+            text = disp(func, **kwargs)
+        else:
+            name = FUNCTIONS.get(func._deannotate().__class__, None)
+            if name:
+                if func._has_args:
+                    name += "%(expr)s"
+            else:
+                name = func.name
+                name = (
+                    self.preparer.quote(name)
+                    if self.preparer._requires_quotes_illegal_chars(name)
+                    or isinstance(name, elements.quoted_name)
+                    else name
+                )
+                name = name + "%(expr)s"
+            text = ".".join(
+                [
+                    (
+                        self.preparer.quote(tok)
+                        if self.preparer._requires_quotes_illegal_chars(tok)
+                        or isinstance(name, elements.quoted_name)
+                        else tok
+                    )
+                    for tok in func.packagenames
+                ]
+                + [name]
+            ) % {"expr": self.function_argspec(func, **kwargs)}
+
+        if func._with_ordinality:
+            text += " WITH ORDINALITY"
+        return text
+
+    def visit_next_value_func(self, next_value, **kw):
+        return self.visit_sequence(next_value.sequence)
+
+    def visit_sequence(self, sequence, **kw):
+        raise NotImplementedError(
+            "Dialect '%s' does not support sequence increments."
+            % self.dialect.name
+        )
+
+    def function_argspec(self, func, **kwargs):
+        return func.clause_expr._compiler_dispatch(self, **kwargs)
+
+    def visit_compound_select(
+        self, cs, asfrom=False, compound_index=None, **kwargs
+    ):
+        toplevel = not self.stack
+
+        compile_state = cs._compile_state_factory(cs, self, **kwargs)
+
+        if toplevel and not self.compile_state:
+            self.compile_state = compile_state
+
+        compound_stmt = compile_state.statement
+
+        entry = self._default_stack_entry if toplevel else self.stack[-1]
+        need_result_map = toplevel or (
+            not compound_index
+            and entry.get("need_result_map_for_compound", False)
+        )
+
+        # indicates there is already a CompoundSelect in play
+        if compound_index == 0:
+            entry["select_0"] = cs
+
+        self.stack.append(
+            {
+                "correlate_froms": entry["correlate_froms"],
+                "asfrom_froms": entry["asfrom_froms"],
+                "selectable": cs,
+                "compile_state": compile_state,
+                "need_result_map_for_compound": need_result_map,
+            }
+        )
+
+        if compound_stmt._independent_ctes:
+            self._dispatch_independent_ctes(compound_stmt, kwargs)
+
+        keyword = self.compound_keywords[cs.keyword]
+
+        text = (" " + keyword + " ").join(
+            (
+                c._compiler_dispatch(
+                    self, asfrom=asfrom, compound_index=i, **kwargs
+                )
+                for i, c in enumerate(cs.selects)
+            )
+        )
+
+        kwargs["include_table"] = False
+        text += self.group_by_clause(cs, **dict(asfrom=asfrom, **kwargs))
+        text += self.order_by_clause(cs, **kwargs)
+        if cs._has_row_limiting_clause:
+            text += self._row_limit_clause(cs, **kwargs)
+
+        if self.ctes:
+            nesting_level = len(self.stack) if not toplevel else None
+            text = (
+                self._render_cte_clause(
+                    nesting_level=nesting_level,
+                    include_following_stack=True,
+                )
+                + text
+            )
+
+        self.stack.pop(-1)
+        return text
+
+    def _row_limit_clause(self, cs, **kwargs):
+        if cs._fetch_clause is not None:
+            return self.fetch_clause(cs, **kwargs)
+        else:
+            return self.limit_clause(cs, **kwargs)
+
+    def _get_operator_dispatch(self, operator_, qualifier1, qualifier2):
+        attrname = "visit_%s_%s%s" % (
+            operator_.__name__,
+            qualifier1,
+            "_" + qualifier2 if qualifier2 else "",
+        )
+        return getattr(self, attrname, None)
+
+    def visit_unary(
+        self, unary, add_to_result_map=None, result_map_targets=(), **kw
+    ):
+        if add_to_result_map is not None:
+            result_map_targets += (unary,)
+            kw["add_to_result_map"] = add_to_result_map
+            kw["result_map_targets"] = result_map_targets
+
+        if unary.operator:
+            if unary.modifier:
+                raise exc.CompileError(
+                    "Unary expression does not support operator "
+                    "and modifier simultaneously"
+                )
+            disp = self._get_operator_dispatch(
+                unary.operator, "unary", "operator"
+            )
+            if disp:
+                return disp(unary, unary.operator, **kw)
+            else:
+                return self._generate_generic_unary_operator(
+                    unary, OPERATORS[unary.operator], **kw
+                )
+        elif unary.modifier:
+            disp = self._get_operator_dispatch(
+                unary.modifier, "unary", "modifier"
+            )
+            if disp:
+                return disp(unary, unary.modifier, **kw)
+            else:
+                return self._generate_generic_unary_modifier(
+                    unary, OPERATORS[unary.modifier], **kw
+                )
+        else:
+            raise exc.CompileError(
+                "Unary expression has no operator or modifier"
+            )
+
+    def visit_truediv_binary(self, binary, operator, **kw):
+        if self.dialect.div_is_floordiv:
+            return (
+                self.process(binary.left, **kw)
+                + " / "
+                # TODO: would need a fast cast again here,
+                # unless we want to use an implicit cast like "+ 0.0"
+                + self.process(
+                    elements.Cast(
+                        binary.right,
+                        (
+                            binary.right.type
+                            if binary.right.type._type_affinity
+                            is sqltypes.Numeric
+                            else sqltypes.Numeric()
+                        ),
+                    ),
+                    **kw,
+                )
+            )
+        else:
+            return (
+                self.process(binary.left, **kw)
+                + " / "
+                + self.process(binary.right, **kw)
+            )
+
+    def visit_floordiv_binary(self, binary, operator, **kw):
+        if (
+            self.dialect.div_is_floordiv
+            and binary.right.type._type_affinity is sqltypes.Integer
+        ):
+            return (
+                self.process(binary.left, **kw)
+                + " / "
+                + self.process(binary.right, **kw)
+            )
+        else:
+            return "FLOOR(%s)" % (
+                self.process(binary.left, **kw)
+                + " / "
+                + self.process(binary.right, **kw)
+            )
+
+    def visit_is_true_unary_operator(self, element, operator, **kw):
+        if (
+            element._is_implicitly_boolean
+            or self.dialect.supports_native_boolean
+        ):
+            return self.process(element.element, **kw)
+        else:
+            return "%s = 1" % self.process(element.element, **kw)
+
+    def visit_is_false_unary_operator(self, element, operator, **kw):
+        if (
+            element._is_implicitly_boolean
+            or self.dialect.supports_native_boolean
+        ):
+            return "NOT %s" % self.process(element.element, **kw)
+        else:
+            return "%s = 0" % self.process(element.element, **kw)
+
+    def visit_not_match_op_binary(self, binary, operator, **kw):
+        return "NOT %s" % self.visit_binary(
+            binary, override_operator=operators.match_op
+        )
+
+    def visit_not_in_op_binary(self, binary, operator, **kw):
+        # The brackets are required in the NOT IN operation because the empty
+        # case is handled using the form "(col NOT IN (null) OR 1 = 1)".
+        # The presence of the OR makes the brackets required.
+        return "(%s)" % self._generate_generic_binary(
+            binary, OPERATORS[operator], **kw
+        )
+
+    def visit_empty_set_op_expr(self, type_, expand_op, **kw):
+        if expand_op is operators.not_in_op:
+            if len(type_) > 1:
+                return "(%s)) OR (1 = 1" % (
+                    ", ".join("NULL" for element in type_)
+                )
+            else:
+                return "NULL) OR (1 = 1"
+        elif expand_op is operators.in_op:
+            if len(type_) > 1:
+                return "(%s)) AND (1 != 1" % (
+                    ", ".join("NULL" for element in type_)
+                )
+            else:
+                return "NULL) AND (1 != 1"
+        else:
+            return self.visit_empty_set_expr(type_)
+
+    def visit_empty_set_expr(self, element_types, **kw):
+        raise NotImplementedError(
+            "Dialect '%s' does not support empty set expression."
+            % self.dialect.name
+        )
+
+    def _literal_execute_expanding_parameter_literal_binds(
+        self, parameter, values, bind_expression_template=None
+    ):
+        typ_dialect_impl = parameter.type._unwrapped_dialect_impl(self.dialect)
+
+        if not values:
+            # empty IN expression.  note we don't need to use
+            # bind_expression_template here because there are no
+            # expressions to render.
+
+            if typ_dialect_impl._is_tuple_type:
+                replacement_expression = (
+                    "VALUES " if self.dialect.tuple_in_values else ""
+                ) + self.visit_empty_set_op_expr(
+                    parameter.type.types, parameter.expand_op
+                )
+
+            else:
+                replacement_expression = self.visit_empty_set_op_expr(
+                    [parameter.type], parameter.expand_op
+                )
+
+        elif typ_dialect_impl._is_tuple_type or (
+            typ_dialect_impl._isnull
+            and isinstance(values[0], collections_abc.Sequence)
+            and not isinstance(values[0], (str, bytes))
+        ):
+            if typ_dialect_impl._has_bind_expression:
+                raise NotImplementedError(
+                    "bind_expression() on TupleType not supported with "
+                    "literal_binds"
+                )
+
+            replacement_expression = (
+                "VALUES " if self.dialect.tuple_in_values else ""
+            ) + ", ".join(
+                "(%s)"
+                % (
+                    ", ".join(
+                        self.render_literal_value(value, param_type)
+                        for value, param_type in zip(
+                            tuple_element, parameter.type.types
+                        )
+                    )
+                )
+                for i, tuple_element in enumerate(values)
+            )
+        else:
+            if bind_expression_template:
+                post_compile_pattern = self._post_compile_pattern
+                m = post_compile_pattern.search(bind_expression_template)
+                assert m and m.group(
+                    2
+                ), "unexpected format for expanding parameter"
+
+                tok = m.group(2).split("~~")
+                be_left, be_right = tok[1], tok[3]
+                replacement_expression = ", ".join(
+                    "%s%s%s"
+                    % (
+                        be_left,
+                        self.render_literal_value(value, parameter.type),
+                        be_right,
+                    )
+                    for value in values
+                )
+            else:
+                replacement_expression = ", ".join(
+                    self.render_literal_value(value, parameter.type)
+                    for value in values
+                )
+
+        return (), replacement_expression
+
+    def _literal_execute_expanding_parameter(self, name, parameter, values):
+        if parameter.literal_execute:
+            return self._literal_execute_expanding_parameter_literal_binds(
+                parameter, values
+            )
+
+        dialect = self.dialect
+        typ_dialect_impl = parameter.type._unwrapped_dialect_impl(dialect)
+
+        if self._numeric_binds:
+            bind_template = self.compilation_bindtemplate
+        else:
+            bind_template = self.bindtemplate
+
+        if (
+            self.dialect._bind_typing_render_casts
+            and typ_dialect_impl.render_bind_cast
+        ):
+
+            def _render_bindtemplate(name):
+                return self.render_bind_cast(
+                    parameter.type,
+                    typ_dialect_impl,
+                    bind_template % {"name": name},
+                )
+
+        else:
+
+            def _render_bindtemplate(name):
+                return bind_template % {"name": name}
+
+        if not values:
+            to_update = []
+            if typ_dialect_impl._is_tuple_type:
+                replacement_expression = self.visit_empty_set_op_expr(
+                    parameter.type.types, parameter.expand_op
+                )
+            else:
+                replacement_expression = self.visit_empty_set_op_expr(
+                    [parameter.type], parameter.expand_op
+                )
+
+        elif typ_dialect_impl._is_tuple_type or (
+            typ_dialect_impl._isnull
+            and isinstance(values[0], collections_abc.Sequence)
+            and not isinstance(values[0], (str, bytes))
+        ):
+            assert not typ_dialect_impl._is_array
+            to_update = [
+                ("%s_%s_%s" % (name, i, j), value)
+                for i, tuple_element in enumerate(values, 1)
+                for j, value in enumerate(tuple_element, 1)
+            ]
+
+            replacement_expression = (
+                "VALUES " if dialect.tuple_in_values else ""
+            ) + ", ".join(
+                "(%s)"
+                % (
+                    ", ".join(
+                        _render_bindtemplate(
+                            to_update[i * len(tuple_element) + j][0]
+                        )
+                        for j, value in enumerate(tuple_element)
+                    )
+                )
+                for i, tuple_element in enumerate(values)
+            )
+        else:
+            to_update = [
+                ("%s_%s" % (name, i), value)
+                for i, value in enumerate(values, 1)
+            ]
+            replacement_expression = ", ".join(
+                _render_bindtemplate(key) for key, value in to_update
+            )
+
+        return to_update, replacement_expression
+
+    def visit_binary(
+        self,
+        binary,
+        override_operator=None,
+        eager_grouping=False,
+        from_linter=None,
+        lateral_from_linter=None,
+        **kw,
+    ):
+        if from_linter and operators.is_comparison(binary.operator):
+            if lateral_from_linter is not None:
+                enclosing_lateral = kw["enclosing_lateral"]
+                lateral_from_linter.edges.update(
+                    itertools.product(
+                        _de_clone(
+                            binary.left._from_objects + [enclosing_lateral]
+                        ),
+                        _de_clone(
+                            binary.right._from_objects + [enclosing_lateral]
+                        ),
+                    )
+                )
+            else:
+                from_linter.edges.update(
+                    itertools.product(
+                        _de_clone(binary.left._from_objects),
+                        _de_clone(binary.right._from_objects),
+                    )
+                )
+
+        # don't allow "? = ?" to render
+        if (
+            self.ansi_bind_rules
+            and isinstance(binary.left, elements.BindParameter)
+            and isinstance(binary.right, elements.BindParameter)
+        ):
+            kw["literal_execute"] = True
+
+        operator_ = override_operator or binary.operator
+        disp = self._get_operator_dispatch(operator_, "binary", None)
+        if disp:
+            return disp(binary, operator_, **kw)
+        else:
+            try:
+                opstring = OPERATORS[operator_]
+            except KeyError as err:
+                raise exc.UnsupportedCompilationError(self, operator_) from err
+            else:
+                return self._generate_generic_binary(
+                    binary,
+                    opstring,
+                    from_linter=from_linter,
+                    lateral_from_linter=lateral_from_linter,
+                    **kw,
+                )
+
+    def visit_function_as_comparison_op_binary(self, element, operator, **kw):
+        return self.process(element.sql_function, **kw)
+
+    def visit_mod_binary(self, binary, operator, **kw):
+        if self.preparer._double_percents:
+            return (
+                self.process(binary.left, **kw)
+                + " %% "
+                + self.process(binary.right, **kw)
+            )
+        else:
+            return (
+                self.process(binary.left, **kw)
+                + " % "
+                + self.process(binary.right, **kw)
+            )
+
+    def visit_custom_op_binary(self, element, operator, **kw):
+        kw["eager_grouping"] = operator.eager_grouping
+        return self._generate_generic_binary(
+            element,
+            " " + self.escape_literal_column(operator.opstring) + " ",
+            **kw,
+        )
+
+    def visit_custom_op_unary_operator(self, element, operator, **kw):
+        return self._generate_generic_unary_operator(
+            element, self.escape_literal_column(operator.opstring) + " ", **kw
+        )
+
+    def visit_custom_op_unary_modifier(self, element, operator, **kw):
+        return self._generate_generic_unary_modifier(
+            element, " " + self.escape_literal_column(operator.opstring), **kw
+        )
+
+    def _generate_generic_binary(
+        self, binary, opstring, eager_grouping=False, **kw
+    ):
+        _in_operator_expression = kw.get("_in_operator_expression", False)
+
+        kw["_in_operator_expression"] = True
+        kw["_binary_op"] = binary.operator
+        text = (
+            binary.left._compiler_dispatch(
+                self, eager_grouping=eager_grouping, **kw
+            )
+            + opstring
+            + binary.right._compiler_dispatch(
+                self, eager_grouping=eager_grouping, **kw
+            )
+        )
+
+        if _in_operator_expression and eager_grouping:
+            text = "(%s)" % text
+        return text
+
+    def _generate_generic_unary_operator(self, unary, opstring, **kw):
+        return opstring + unary.element._compiler_dispatch(self, **kw)
+
+    def _generate_generic_unary_modifier(self, unary, opstring, **kw):
+        return unary.element._compiler_dispatch(self, **kw) + opstring
+
+    @util.memoized_property
+    def _like_percent_literal(self):
+        return elements.literal_column("'%'", type_=sqltypes.STRINGTYPE)
+
+    def visit_ilike_case_insensitive_operand(self, element, **kw):
+        return f"lower({element.element._compiler_dispatch(self, **kw)})"
+
+    def visit_contains_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent.concat(binary.right).concat(percent)
+        return self.visit_like_op_binary(binary, operator, **kw)
+
+    def visit_not_contains_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent.concat(binary.right).concat(percent)
+        return self.visit_not_like_op_binary(binary, operator, **kw)
+
+    def visit_icontains_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent.concat(
+            ilike_case_insensitive(binary.right)
+        ).concat(percent)
+        return self.visit_ilike_op_binary(binary, operator, **kw)
+
+    def visit_not_icontains_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent.concat(
+            ilike_case_insensitive(binary.right)
+        ).concat(percent)
+        return self.visit_not_ilike_op_binary(binary, operator, **kw)
+
+    def visit_startswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent._rconcat(binary.right)
+        return self.visit_like_op_binary(binary, operator, **kw)
+
+    def visit_not_startswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent._rconcat(binary.right)
+        return self.visit_not_like_op_binary(binary, operator, **kw)
+
+    def visit_istartswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent._rconcat(ilike_case_insensitive(binary.right))
+        return self.visit_ilike_op_binary(binary, operator, **kw)
+
+    def visit_not_istartswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent._rconcat(ilike_case_insensitive(binary.right))
+        return self.visit_not_ilike_op_binary(binary, operator, **kw)
+
+    def visit_endswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent.concat(binary.right)
+        return self.visit_like_op_binary(binary, operator, **kw)
+
+    def visit_not_endswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.right = percent.concat(binary.right)
+        return self.visit_not_like_op_binary(binary, operator, **kw)
+
+    def visit_iendswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent.concat(ilike_case_insensitive(binary.right))
+        return self.visit_ilike_op_binary(binary, operator, **kw)
+
+    def visit_not_iendswith_op_binary(self, binary, operator, **kw):
+        binary = binary._clone()
+        percent = self._like_percent_literal
+        binary.left = ilike_case_insensitive(binary.left)
+        binary.right = percent.concat(ilike_case_insensitive(binary.right))
+        return self.visit_not_ilike_op_binary(binary, operator, **kw)
+
+    def visit_like_op_binary(self, binary, operator, **kw):
+        escape = binary.modifiers.get("escape", None)
+
+        return "%s LIKE %s" % (
+            binary.left._compiler_dispatch(self, **kw),
+            binary.right._compiler_dispatch(self, **kw),
+        ) + (
+            " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE)
+            if escape is not None
+            else ""
+        )
+
+    def visit_not_like_op_binary(self, binary, operator, **kw):
+        escape = binary.modifiers.get("escape", None)
+        return "%s NOT LIKE %s" % (
+            binary.left._compiler_dispatch(self, **kw),
+            binary.right._compiler_dispatch(self, **kw),
+        ) + (
+            " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE)
+            if escape is not None
+            else ""
+        )
+
+    def visit_ilike_op_binary(self, binary, operator, **kw):
+        if operator is operators.ilike_op:
+            binary = binary._clone()
+            binary.left = ilike_case_insensitive(binary.left)
+            binary.right = ilike_case_insensitive(binary.right)
+        # else we assume ilower() has been applied
+
+        return self.visit_like_op_binary(binary, operator, **kw)
+
+    def visit_not_ilike_op_binary(self, binary, operator, **kw):
+        if operator is operators.not_ilike_op:
+            binary = binary._clone()
+            binary.left = ilike_case_insensitive(binary.left)
+            binary.right = ilike_case_insensitive(binary.right)
+        # else we assume ilower() has been applied
+
+        return self.visit_not_like_op_binary(binary, operator, **kw)
+
+    def visit_between_op_binary(self, binary, operator, **kw):
+        symmetric = binary.modifiers.get("symmetric", False)
+        return self._generate_generic_binary(
+            binary, " BETWEEN SYMMETRIC " if symmetric else " BETWEEN ", **kw
+        )
+
+    def visit_not_between_op_binary(self, binary, operator, **kw):
+        symmetric = binary.modifiers.get("symmetric", False)
+        return self._generate_generic_binary(
+            binary,
+            " NOT BETWEEN SYMMETRIC " if symmetric else " NOT BETWEEN ",
+            **kw,
+        )
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        raise exc.CompileError(
+            "%s dialect does not support regular expressions"
+            % self.dialect.name
+        )
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        raise exc.CompileError(
+            "%s dialect does not support regular expressions"
+            % self.dialect.name
+        )
+
+    def visit_regexp_replace_op_binary(self, binary, operator, **kw):
+        raise exc.CompileError(
+            "%s dialect does not support regular expression replacements"
+            % self.dialect.name
+        )
+
+    def visit_bindparam(
+        self,
+        bindparam,
+        within_columns_clause=False,
+        literal_binds=False,
+        skip_bind_expression=False,
+        literal_execute=False,
+        render_postcompile=False,
+        **kwargs,
+    ):
+
+        if not skip_bind_expression:
+            impl = bindparam.type.dialect_impl(self.dialect)
+            if impl._has_bind_expression:
+                bind_expression = impl.bind_expression(bindparam)
+                wrapped = self.process(
+                    bind_expression,
+                    skip_bind_expression=True,
+                    within_columns_clause=within_columns_clause,
+                    literal_binds=literal_binds and not bindparam.expanding,
+                    literal_execute=literal_execute,
+                    render_postcompile=render_postcompile,
+                    **kwargs,
+                )
+                if bindparam.expanding:
+                    # for postcompile w/ expanding, move the "wrapped" part
+                    # of this into the inside
+
+                    m = re.match(
+                        r"^(.*)\(__\[POSTCOMPILE_(\S+?)\]\)(.*)$", wrapped
+                    )
+                    assert m, "unexpected format for expanding parameter"
+                    wrapped = "(__[POSTCOMPILE_%s~~%s~~REPL~~%s~~])" % (
+                        m.group(2),
+                        m.group(1),
+                        m.group(3),
+                    )
+
+                    if literal_binds:
+                        ret = self.render_literal_bindparam(
+                            bindparam,
+                            within_columns_clause=True,
+                            bind_expression_template=wrapped,
+                            **kwargs,
+                        )
+                        return "(%s)" % ret
+
+                return wrapped
+
+        if not literal_binds:
+            literal_execute = (
+                literal_execute
+                or bindparam.literal_execute
+                or (within_columns_clause and self.ansi_bind_rules)
+            )
+            post_compile = literal_execute or bindparam.expanding
+        else:
+            post_compile = False
+
+        if literal_binds:
+            ret = self.render_literal_bindparam(
+                bindparam, within_columns_clause=True, **kwargs
+            )
+            if bindparam.expanding:
+                ret = "(%s)" % ret
+            return ret
+
+        name = self._truncate_bindparam(bindparam)
+
+        if name in self.binds:
+            existing = self.binds[name]
+            if existing is not bindparam:
+                if (
+                    (existing.unique or bindparam.unique)
+                    and not existing.proxy_set.intersection(
+                        bindparam.proxy_set
+                    )
+                    and not existing._cloned_set.intersection(
+                        bindparam._cloned_set
+                    )
+                ):
+                    raise exc.CompileError(
+                        "Bind parameter '%s' conflicts with "
+                        "unique bind parameter of the same name" % name
+                    )
+                elif existing.expanding != bindparam.expanding:
+                    raise exc.CompileError(
+                        "Can't reuse bound parameter name '%s' in both "
+                        "'expanding' (e.g. within an IN expression) and "
+                        "non-expanding contexts.  If this parameter is to "
+                        "receive a list/array value, set 'expanding=True' on "
+                        "it for expressions that aren't IN, otherwise use "
+                        "a different parameter name." % (name,)
+                    )
+                elif existing._is_crud or bindparam._is_crud:
+                    if existing._is_crud and bindparam._is_crud:
+                        # TODO: this condition is not well understood.
+                        # see tests in test/sql/test_update.py
+                        raise exc.CompileError(
+                            "Encountered unsupported case when compiling an "
+                            "INSERT or UPDATE statement.  If this is a "
+                            "multi-table "
+                            "UPDATE statement, please provide string-named "
+                            "arguments to the "
+                            "values() method with distinct names; support for "
+                            "multi-table UPDATE statements that "
+                            "target multiple tables for UPDATE is very "
+                            "limited",
+                        )
+                    else:
+                        raise exc.CompileError(
+                            f"bindparam() name '{bindparam.key}' is reserved "
+                            "for automatic usage in the VALUES or SET "
+                            "clause of this "
+                            "insert/update statement.   Please use a "
+                            "name other than column name when using "
+                            "bindparam() "
+                            "with insert() or update() (for example, "
+                            f"'b_{bindparam.key}')."
+                        )
+
+        self.binds[bindparam.key] = self.binds[name] = bindparam
+
+        # if we are given a cache key that we're going to match against,
+        # relate the bindparam here to one that is most likely present
+        # in the "extracted params" portion of the cache key.  this is used
+        # to set up a positional mapping that is used to determine the
+        # correct parameters for a subsequent use of this compiled with
+        # a different set of parameter values.   here, we accommodate for
+        # parameters that may have been cloned both before and after the cache
+        # key was been generated.
+        ckbm_tuple = self._cache_key_bind_match
+
+        if ckbm_tuple:
+            ckbm, cksm = ckbm_tuple
+            for bp in bindparam._cloned_set:
+                if bp.key in cksm:
+                    cb = cksm[bp.key]
+                    ckbm[cb].append(bindparam)
+
+        if bindparam.isoutparam:
+            self.has_out_parameters = True
+
+        if post_compile:
+            if render_postcompile:
+                self._render_postcompile = True
+
+            if literal_execute:
+                self.literal_execute_params |= {bindparam}
+            else:
+                self.post_compile_params |= {bindparam}
+
+        ret = self.bindparam_string(
+            name,
+            post_compile=post_compile,
+            expanding=bindparam.expanding,
+            bindparam_type=bindparam.type,
+            **kwargs,
+        )
+
+        if bindparam.expanding:
+            ret = "(%s)" % ret
+
+        return ret
+
+    def render_bind_cast(self, type_, dbapi_type, sqltext):
+        raise NotImplementedError()
+
+    def render_literal_bindparam(
+        self,
+        bindparam,
+        render_literal_value=NO_ARG,
+        bind_expression_template=None,
+        **kw,
+    ):
+        if render_literal_value is not NO_ARG:
+            value = render_literal_value
+        else:
+            if bindparam.value is None and bindparam.callable is None:
+                op = kw.get("_binary_op", None)
+                if op and op not in (operators.is_, operators.is_not):
+                    util.warn_limited(
+                        "Bound parameter '%s' rendering literal NULL in a SQL "
+                        "expression; comparisons to NULL should not use "
+                        "operators outside of 'is' or 'is not'",
+                        (bindparam.key,),
+                    )
+                return self.process(sqltypes.NULLTYPE, **kw)
+            value = bindparam.effective_value
+
+        if bindparam.expanding:
+            leep = self._literal_execute_expanding_parameter_literal_binds
+            to_update, replacement_expr = leep(
+                bindparam,
+                value,
+                bind_expression_template=bind_expression_template,
+            )
+            return replacement_expr
+        else:
+            return self.render_literal_value(value, bindparam.type)
+
+    def render_literal_value(self, value, type_):
+        """Render the value of a bind parameter as a quoted literal.
+
+        This is used for statement sections that do not accept bind parameters
+        on the target driver/database.
+
+        This should be implemented by subclasses using the quoting services
+        of the DBAPI.
+
+        """
+
+        if value is None and not type_.should_evaluate_none:
+            # issue #10535 - handle NULL in the compiler without placing
+            # this onto each type, except for "evaluate None" types
+            # (e.g. JSON)
+            return self.process(elements.Null._instance())
+
+        processor = type_._cached_literal_processor(self.dialect)
+        if processor:
+            try:
+                return processor(value)
+            except Exception as e:
+                raise exc.CompileError(
+                    f"Could not render literal value "
+                    f'"{sql_util._repr_single_value(value)}" '
+                    f"with datatype "
+                    f"{type_}; see parent stack trace for "
+                    "more detail."
+                ) from e
+
+        else:
+            raise exc.CompileError(
+                f"No literal value renderer is available for literal value "
+                f'"{sql_util._repr_single_value(value)}" '
+                f"with datatype {type_}"
+            )
+
+    def _truncate_bindparam(self, bindparam):
+        if bindparam in self.bind_names:
+            return self.bind_names[bindparam]
+
+        bind_name = bindparam.key
+        if isinstance(bind_name, elements._truncated_label):
+            bind_name = self._truncated_identifier("bindparam", bind_name)
+
+        # add to bind_names for translation
+        self.bind_names[bindparam] = bind_name
+
+        return bind_name
+
+    def _truncated_identifier(
+        self, ident_class: str, name: _truncated_label
+    ) -> str:
+        if (ident_class, name) in self.truncated_names:
+            return self.truncated_names[(ident_class, name)]
+
+        anonname = name.apply_map(self.anon_map)
+
+        if len(anonname) > self.label_length - 6:
+            counter = self._truncated_counters.get(ident_class, 1)
+            truncname = (
+                anonname[0 : max(self.label_length - 6, 0)]
+                + "_"
+                + hex(counter)[2:]
+            )
+            self._truncated_counters[ident_class] = counter + 1
+        else:
+            truncname = anonname
+        self.truncated_names[(ident_class, name)] = truncname
+        return truncname
+
+    def _anonymize(self, name: str) -> str:
+        return name % self.anon_map
+
+    def bindparam_string(
+        self,
+        name: str,
+        post_compile: bool = False,
+        expanding: bool = False,
+        escaped_from: Optional[str] = None,
+        bindparam_type: Optional[TypeEngine[Any]] = None,
+        accumulate_bind_names: Optional[Set[str]] = None,
+        visited_bindparam: Optional[List[str]] = None,
+        **kw: Any,
+    ) -> str:
+        # TODO: accumulate_bind_names is passed by crud.py to gather
+        # names on a per-value basis, visited_bindparam is passed by
+        # visit_insert() to collect all parameters in the statement.
+        # see if this gathering can be simplified somehow
+        if accumulate_bind_names is not None:
+            accumulate_bind_names.add(name)
+        if visited_bindparam is not None:
+            visited_bindparam.append(name)
+
+        if not escaped_from:
+            if self._bind_translate_re.search(name):
+                # not quite the translate use case as we want to
+                # also get a quick boolean if we even found
+                # unusual characters in the name
+                new_name = self._bind_translate_re.sub(
+                    lambda m: self._bind_translate_chars[m.group(0)],
+                    name,
+                )
+                escaped_from = name
+                name = new_name
+
+        if escaped_from:
+            self.escaped_bind_names = self.escaped_bind_names.union(
+                {escaped_from: name}
+            )
+        if post_compile:
+            ret = "__[POSTCOMPILE_%s]" % name
+            if expanding:
+                # for expanding, bound parameters or literal values will be
+                # rendered per item
+                return ret
+
+            # otherwise, for non-expanding "literal execute", apply
+            # bind casts as determined by the datatype
+            if bindparam_type is not None:
+                type_impl = bindparam_type._unwrapped_dialect_impl(
+                    self.dialect
+                )
+                if type_impl.render_literal_cast:
+                    ret = self.render_bind_cast(bindparam_type, type_impl, ret)
+            return ret
+        elif self.state is CompilerState.COMPILING:
+            ret = self.compilation_bindtemplate % {"name": name}
+        else:
+            ret = self.bindtemplate % {"name": name}
+
+        if (
+            bindparam_type is not None
+            and self.dialect._bind_typing_render_casts
+        ):
+            type_impl = bindparam_type._unwrapped_dialect_impl(self.dialect)
+            if type_impl.render_bind_cast:
+                ret = self.render_bind_cast(bindparam_type, type_impl, ret)
+
+        return ret
+
+    def _dispatch_independent_ctes(self, stmt, kw):
+        local_kw = kw.copy()
+        local_kw.pop("cte_opts", None)
+        for cte, opt in zip(
+            stmt._independent_ctes, stmt._independent_ctes_opts
+        ):
+            cte._compiler_dispatch(self, cte_opts=opt, **local_kw)
+
+    def visit_cte(
+        self,
+        cte: CTE,
+        asfrom: bool = False,
+        ashint: bool = False,
+        fromhints: Optional[_FromHintsType] = None,
+        visiting_cte: Optional[CTE] = None,
+        from_linter: Optional[FromLinter] = None,
+        cte_opts: selectable._CTEOpts = selectable._CTEOpts(False),
+        **kwargs: Any,
+    ) -> Optional[str]:
+        self_ctes = self._init_cte_state()
+        assert self_ctes is self.ctes
+
+        kwargs["visiting_cte"] = cte
+
+        cte_name = cte.name
+
+        if isinstance(cte_name, elements._truncated_label):
+            cte_name = self._truncated_identifier("alias", cte_name)
+
+        is_new_cte = True
+        embedded_in_current_named_cte = False
+
+        _reference_cte = cte._get_reference_cte()
+
+        nesting = cte.nesting or cte_opts.nesting
+
+        # check for CTE already encountered
+        if _reference_cte in self.level_name_by_cte:
+            cte_level, _, existing_cte_opts = self.level_name_by_cte[
+                _reference_cte
+            ]
+            assert _ == cte_name
+
+            cte_level_name = (cte_level, cte_name)
+            existing_cte = self.ctes_by_level_name[cte_level_name]
+
+            # check if we are receiving it here with a specific
+            # "nest_here" location; if so, move it to this location
+
+            if cte_opts.nesting:
+                if existing_cte_opts.nesting:
+                    raise exc.CompileError(
+                        "CTE is stated as 'nest_here' in "
+                        "more than one location"
+                    )
+
+                old_level_name = (cte_level, cte_name)
+                cte_level = len(self.stack) if nesting else 1
+                cte_level_name = new_level_name = (cte_level, cte_name)
+
+                del self.ctes_by_level_name[old_level_name]
+                self.ctes_by_level_name[new_level_name] = existing_cte
+                self.level_name_by_cte[_reference_cte] = new_level_name + (
+                    cte_opts,
+                )
+
+        else:
+            cte_level = len(self.stack) if nesting else 1
+            cte_level_name = (cte_level, cte_name)
+
+            if cte_level_name in self.ctes_by_level_name:
+                existing_cte = self.ctes_by_level_name[cte_level_name]
+            else:
+                existing_cte = None
+
+        if existing_cte is not None:
+            embedded_in_current_named_cte = visiting_cte is existing_cte
+
+            # we've generated a same-named CTE that we are enclosed in,
+            # or this is the same CTE.  just return the name.
+            if cte is existing_cte._restates or cte is existing_cte:
+                is_new_cte = False
+            elif existing_cte is cte._restates:
+                # we've generated a same-named CTE that is
+                # enclosed in us - we take precedence, so
+                # discard the text for the "inner".
+                del self_ctes[existing_cte]
+
+                existing_cte_reference_cte = existing_cte._get_reference_cte()
+
+                assert existing_cte_reference_cte is _reference_cte
+                assert existing_cte_reference_cte is existing_cte
+
+                del self.level_name_by_cte[existing_cte_reference_cte]
+            else:
+                # if the two CTEs are deep-copy identical, consider them
+                # the same, **if** they are clones, that is, they came from
+                # the ORM or other visit method
+                if (
+                    cte._is_clone_of is not None
+                    or existing_cte._is_clone_of is not None
+                ) and cte.compare(existing_cte):
+                    is_new_cte = False
+                else:
+                    raise exc.CompileError(
+                        "Multiple, unrelated CTEs found with "
+                        "the same name: %r" % cte_name
+                    )
+
+        if not asfrom and not is_new_cte:
+            return None
+
+        if cte._cte_alias is not None:
+            pre_alias_cte = cte._cte_alias
+            cte_pre_alias_name = cte._cte_alias.name
+            if isinstance(cte_pre_alias_name, elements._truncated_label):
+                cte_pre_alias_name = self._truncated_identifier(
+                    "alias", cte_pre_alias_name
+                )
+        else:
+            pre_alias_cte = cte
+            cte_pre_alias_name = None
+
+        if is_new_cte:
+            self.ctes_by_level_name[cte_level_name] = cte
+            self.level_name_by_cte[_reference_cte] = cte_level_name + (
+                cte_opts,
+            )
+
+            if pre_alias_cte not in self.ctes:
+                self.visit_cte(pre_alias_cte, **kwargs)
+
+            if not cte_pre_alias_name and cte not in self_ctes:
+                if cte.recursive:
+                    self.ctes_recursive = True
+                text = self.preparer.format_alias(cte, cte_name)
+                if cte.recursive:
+                    col_source = cte.element
+
+                    # TODO: can we get at the .columns_plus_names collection
+                    # that is already (or will be?) generated for the SELECT
+                    # rather than calling twice?
+                    recur_cols = [
+                        # TODO: proxy_name is not technically safe,
+                        # see test_cte->
+                        # test_with_recursive_no_name_currently_buggy.  not
+                        # clear what should be done with such a case
+                        fallback_label_name or proxy_name
+                        for (
+                            _,
+                            proxy_name,
+                            fallback_label_name,
+                            c,
+                            repeated,
+                        ) in (col_source._generate_columns_plus_names(True))
+                        if not repeated
+                    ]
+
+                    text += "(%s)" % (
+                        ", ".join(
+                            self.preparer.format_label_name(
+                                ident, anon_map=self.anon_map
+                            )
+                            for ident in recur_cols
+                        )
+                    )
+
+                assert kwargs.get("subquery", False) is False
+
+                if not self.stack:
+                    # toplevel, this is a stringify of the
+                    # cte directly.  just compile the inner
+                    # the way alias() does.
+                    return cte.element._compiler_dispatch(
+                        self, asfrom=asfrom, **kwargs
+                    )
+                else:
+                    prefixes = self._generate_prefixes(
+                        cte, cte._prefixes, **kwargs
+                    )
+                    inner = cte.element._compiler_dispatch(
+                        self, asfrom=True, **kwargs
+                    )
+
+                    text += " AS %s\n(%s)" % (prefixes, inner)
+
+                if cte._suffixes:
+                    text += " " + self._generate_prefixes(
+                        cte, cte._suffixes, **kwargs
+                    )
+
+                self_ctes[cte] = text
+
+        if asfrom:
+            if from_linter:
+                from_linter.froms[cte._de_clone()] = cte_name
+
+            if not is_new_cte and embedded_in_current_named_cte:
+                return self.preparer.format_alias(cte, cte_name)
+
+            if cte_pre_alias_name:
+                text = self.preparer.format_alias(cte, cte_pre_alias_name)
+                if self.preparer._requires_quotes(cte_name):
+                    cte_name = self.preparer.quote(cte_name)
+                text += self.get_render_as_alias_suffix(cte_name)
+                return text
+            else:
+                return self.preparer.format_alias(cte, cte_name)
+
+        return None
+
+    def visit_table_valued_alias(self, element, **kw):
+        if element.joins_implicitly:
+            kw["from_linter"] = None
+        if element._is_lateral:
+            return self.visit_lateral(element, **kw)
+        else:
+            return self.visit_alias(element, **kw)
+
+    def visit_table_valued_column(self, element, **kw):
+        return self.visit_column(element, **kw)
+
+    def visit_alias(
+        self,
+        alias,
+        asfrom=False,
+        ashint=False,
+        iscrud=False,
+        fromhints=None,
+        subquery=False,
+        lateral=False,
+        enclosing_alias=None,
+        from_linter=None,
+        **kwargs,
+    ):
+        if lateral:
+            if "enclosing_lateral" not in kwargs:
+                # if lateral is set and enclosing_lateral is not
+                # present, we assume we are being called directly
+                # from visit_lateral() and we need to set enclosing_lateral.
+                assert alias._is_lateral
+                kwargs["enclosing_lateral"] = alias
+
+            # for lateral objects, we track a second from_linter that is...
+            # lateral!  to the level above us.
+            if (
+                from_linter
+                and "lateral_from_linter" not in kwargs
+                and "enclosing_lateral" in kwargs
+            ):
+                kwargs["lateral_from_linter"] = from_linter
+
+        if enclosing_alias is not None and enclosing_alias.element is alias:
+            inner = alias.element._compiler_dispatch(
+                self,
+                asfrom=asfrom,
+                ashint=ashint,
+                iscrud=iscrud,
+                fromhints=fromhints,
+                lateral=lateral,
+                enclosing_alias=alias,
+                **kwargs,
+            )
+            if subquery and (asfrom or lateral):
+                inner = "(%s)" % (inner,)
+            return inner
+        else:
+            enclosing_alias = kwargs["enclosing_alias"] = alias
+
+        if asfrom or ashint:
+            if isinstance(alias.name, elements._truncated_label):
+                alias_name = self._truncated_identifier("alias", alias.name)
+            else:
+                alias_name = alias.name
+
+        if ashint:
+            return self.preparer.format_alias(alias, alias_name)
+        elif asfrom:
+            if from_linter:
+                from_linter.froms[alias._de_clone()] = alias_name
+
+            inner = alias.element._compiler_dispatch(
+                self, asfrom=True, lateral=lateral, **kwargs
+            )
+            if subquery:
+                inner = "(%s)" % (inner,)
+
+            ret = inner + self.get_render_as_alias_suffix(
+                self.preparer.format_alias(alias, alias_name)
+            )
+
+            if alias._supports_derived_columns and alias._render_derived:
+                ret += "(%s)" % (
+                    ", ".join(
+                        "%s%s"
+                        % (
+                            self.preparer.quote(col.name),
+                            (
+                                " %s"
+                                % self.dialect.type_compiler_instance.process(
+                                    col.type, **kwargs
+                                )
+                                if alias._render_derived_w_types
+                                else ""
+                            ),
+                        )
+                        for col in alias.c
+                    )
+                )
+
+            if fromhints and alias in fromhints:
+                ret = self.format_from_hint_text(
+                    ret, alias, fromhints[alias], iscrud
+                )
+
+            return ret
+        else:
+            # note we cancel the "subquery" flag here as well
+            return alias.element._compiler_dispatch(
+                self, lateral=lateral, **kwargs
+            )
+
+    def visit_subquery(self, subquery, **kw):
+        kw["subquery"] = True
+        return self.visit_alias(subquery, **kw)
+
+    def visit_lateral(self, lateral_, **kw):
+        kw["lateral"] = True
+        return "LATERAL %s" % self.visit_alias(lateral_, **kw)
+
+    def visit_tablesample(self, tablesample, asfrom=False, **kw):
+        text = "%s TABLESAMPLE %s" % (
+            self.visit_alias(tablesample, asfrom=True, **kw),
+            tablesample._get_method()._compiler_dispatch(self, **kw),
+        )
+
+        if tablesample.seed is not None:
+            text += " REPEATABLE (%s)" % (
+                tablesample.seed._compiler_dispatch(self, **kw)
+            )
+
+        return text
+
+    def _render_values(self, element, **kw):
+        kw.setdefault("literal_binds", element.literal_binds)
+        tuples = ", ".join(
+            self.process(
+                elements.Tuple(
+                    types=element._column_types, *elem
+                ).self_group(),
+                **kw,
+            )
+            for chunk in element._data
+            for elem in chunk
+        )
+        return f"VALUES {tuples}"
+
+    def visit_values(self, element, asfrom=False, from_linter=None, **kw):
+        v = self._render_values(element, **kw)
+
+        if element._unnamed:
+            name = None
+        elif isinstance(element.name, elements._truncated_label):
+            name = self._truncated_identifier("values", element.name)
+        else:
+            name = element.name
+
+        if element._is_lateral:
+            lateral = "LATERAL "
+        else:
+            lateral = ""
+
+        if asfrom:
+            if from_linter:
+                from_linter.froms[element._de_clone()] = (
+                    name if name is not None else "(unnamed VALUES element)"
+                )
+
+            if name:
+                kw["include_table"] = False
+                v = "%s(%s)%s (%s)" % (
+                    lateral,
+                    v,
+                    self.get_render_as_alias_suffix(self.preparer.quote(name)),
+                    (
+                        ", ".join(
+                            c._compiler_dispatch(self, **kw)
+                            for c in element.columns
+                        )
+                    ),
+                )
+            else:
+                v = "%s(%s)" % (lateral, v)
+        return v
+
+    def visit_scalar_values(self, element, **kw):
+        return f"({self._render_values(element, **kw)})"
+
+    def get_render_as_alias_suffix(self, alias_name_text):
+        return " AS " + alias_name_text
+
+    def _add_to_result_map(
+        self,
+        keyname: str,
+        name: str,
+        objects: Tuple[Any, ...],
+        type_: TypeEngine[Any],
+    ) -> None:
+
+        # note objects must be non-empty for cursor.py to handle the
+        # collection properly
+        assert objects
+
+        if keyname is None or keyname == "*":
+            self._ordered_columns = False
+            self._ad_hoc_textual = True
+        if type_._is_tuple_type:
+            raise exc.CompileError(
+                "Most backends don't support SELECTing "
+                "from a tuple() object.  If this is an ORM query, "
+                "consider using the Bundle object."
+            )
+        self._result_columns.append(
+            ResultColumnsEntry(keyname, name, objects, type_)
+        )
+
+    def _label_returning_column(
+        self, stmt, column, populate_result_map, column_clause_args=None, **kw
+    ):
+        """Render a column with necessary labels inside of a RETURNING clause.
+
+        This method is provided for individual dialects in place of calling
+        the _label_select_column method directly, so that the two use cases
+        of RETURNING vs. SELECT can be disambiguated going forward.
+
+        .. versionadded:: 1.4.21
+
+        """
+        return self._label_select_column(
+            None,
+            column,
+            populate_result_map,
+            False,
+            {} if column_clause_args is None else column_clause_args,
+            **kw,
+        )
+
+    def _label_select_column(
+        self,
+        select,
+        column,
+        populate_result_map,
+        asfrom,
+        column_clause_args,
+        name=None,
+        proxy_name=None,
+        fallback_label_name=None,
+        within_columns_clause=True,
+        column_is_repeated=False,
+        need_column_expressions=False,
+        include_table=True,
+    ):
+        """produce labeled columns present in a select()."""
+        impl = column.type.dialect_impl(self.dialect)
+
+        if impl._has_column_expression and (
+            need_column_expressions or populate_result_map
+        ):
+            col_expr = impl.column_expression(column)
+        else:
+            col_expr = column
+
+        if populate_result_map:
+            # pass an "add_to_result_map" callable into the compilation
+            # of embedded columns.  this collects information about the
+            # column as it will be fetched in the result and is coordinated
+            # with cursor.description when the query is executed.
+            add_to_result_map = self._add_to_result_map
+
+            # if the SELECT statement told us this column is a repeat,
+            # wrap the callable with one that prevents the addition of the
+            # targets
+            if column_is_repeated:
+                _add_to_result_map = add_to_result_map
+
+                def add_to_result_map(keyname, name, objects, type_):
+                    _add_to_result_map(keyname, name, (keyname,), type_)
+
+            # if we redefined col_expr for type expressions, wrap the
+            # callable with one that adds the original column to the targets
+            elif col_expr is not column:
+                _add_to_result_map = add_to_result_map
+
+                def add_to_result_map(keyname, name, objects, type_):
+                    _add_to_result_map(
+                        keyname, name, (column,) + objects, type_
+                    )
+
+        else:
+            add_to_result_map = None
+
+        # this method is used by some of the dialects for RETURNING,
+        # which has different inputs.  _label_returning_column was added
+        # as the better target for this now however for 1.4 we will keep
+        # _label_select_column directly compatible with this use case.
+        # these assertions right now set up the current expected inputs
+        assert within_columns_clause, (
+            "_label_select_column is only relevant within "
+            "the columns clause of a SELECT or RETURNING"
+        )
+        if isinstance(column, elements.Label):
+            if col_expr is not column:
+                result_expr = _CompileLabel(
+                    col_expr, column.name, alt_names=(column.element,)
+                )
+            else:
+                result_expr = col_expr
+
+        elif name:
+            # here, _columns_plus_names has determined there's an explicit
+            # label name we need to use.  this is the default for
+            # tablenames_plus_columnnames as well as when columns are being
+            # deduplicated on name
+
+            assert (
+                proxy_name is not None
+            ), "proxy_name is required if 'name' is passed"
+
+            result_expr = _CompileLabel(
+                col_expr,
+                name,
+                alt_names=(
+                    proxy_name,
+                    # this is a hack to allow legacy result column lookups
+                    # to work as they did before; this goes away in 2.0.
+                    # TODO: this only seems to be tested indirectly
+                    # via test/orm/test_deprecations.py.   should be a
+                    # resultset test for this
+                    column._tq_label,
+                ),
+            )
+        else:
+            # determine here whether this column should be rendered in
+            # a labelled context or not, as we were given no required label
+            # name from the caller. Here we apply heuristics based on the kind
+            # of SQL expression involved.
+
+            if col_expr is not column:
+                # type-specific expression wrapping the given column,
+                # so we render a label
+                render_with_label = True
+            elif isinstance(column, elements.ColumnClause):
+                # table-bound column, we render its name as a label if we are
+                # inside of a subquery only
+                render_with_label = (
+                    asfrom
+                    and not column.is_literal
+                    and column.table is not None
+                )
+            elif isinstance(column, elements.TextClause):
+                render_with_label = False
+            elif isinstance(column, elements.UnaryExpression):
+                render_with_label = column.wraps_column_expression or asfrom
+            elif (
+                # general class of expressions that don't have a SQL-column
+                # addressible name.  includes scalar selects, bind parameters,
+                # SQL functions, others
+                not isinstance(column, elements.NamedColumn)
+                # deeper check that indicates there's no natural "name" to
+                # this element, which accommodates for custom SQL constructs
+                # that might have a ".name" attribute (but aren't SQL
+                # functions) but are not implementing this more recently added
+                # base class.  in theory the "NamedColumn" check should be
+                # enough, however here we seek to maintain legacy behaviors
+                # as well.
+                and column._non_anon_label is None
+            ):
+                render_with_label = True
+            else:
+                render_with_label = False
+
+            if render_with_label:
+                if not fallback_label_name:
+                    # used by the RETURNING case right now.  we generate it
+                    # here as 3rd party dialects may be referring to
+                    # _label_select_column method directly instead of the
+                    # just-added _label_returning_column method
+                    assert not column_is_repeated
+                    fallback_label_name = column._anon_name_label
+
+                fallback_label_name = (
+                    elements._truncated_label(fallback_label_name)
+                    if not isinstance(
+                        fallback_label_name, elements._truncated_label
+                    )
+                    else fallback_label_name
+                )
+
+                result_expr = _CompileLabel(
+                    col_expr, fallback_label_name, alt_names=(proxy_name,)
+                )
+            else:
+                result_expr = col_expr
+
+        column_clause_args.update(
+            within_columns_clause=within_columns_clause,
+            add_to_result_map=add_to_result_map,
+            include_table=include_table,
+        )
+        return result_expr._compiler_dispatch(self, **column_clause_args)
+
+    def format_from_hint_text(self, sqltext, table, hint, iscrud):
+        hinttext = self.get_from_hint_text(table, hint)
+        if hinttext:
+            sqltext += " " + hinttext
+        return sqltext
+
+    def get_select_hint_text(self, byfroms):
+        return None
+
+    def get_from_hint_text(self, table, text):
+        return None
+
+    def get_crud_hint_text(self, table, text):
+        return None
+
+    def get_statement_hint_text(self, hint_texts):
+        return " ".join(hint_texts)
+
+    _default_stack_entry: _CompilerStackEntry
+
+    if not typing.TYPE_CHECKING:
+        _default_stack_entry = util.immutabledict(
+            [("correlate_froms", frozenset()), ("asfrom_froms", frozenset())]
+        )
+
+    def _display_froms_for_select(
+        self, select_stmt, asfrom, lateral=False, **kw
+    ):
+        # utility method to help external dialects
+        # get the correct from list for a select.
+        # specifically the oracle dialect needs this feature
+        # right now.
+        toplevel = not self.stack
+        entry = self._default_stack_entry if toplevel else self.stack[-1]
+
+        compile_state = select_stmt._compile_state_factory(select_stmt, self)
+
+        correlate_froms = entry["correlate_froms"]
+        asfrom_froms = entry["asfrom_froms"]
+
+        if asfrom and not lateral:
+            froms = compile_state._get_display_froms(
+                explicit_correlate_froms=correlate_froms.difference(
+                    asfrom_froms
+                ),
+                implicit_correlate_froms=(),
+            )
+        else:
+            froms = compile_state._get_display_froms(
+                explicit_correlate_froms=correlate_froms,
+                implicit_correlate_froms=asfrom_froms,
+            )
+        return froms
+
+    translate_select_structure: Any = None
+    """if not ``None``, should be a callable which accepts ``(select_stmt,
+    **kw)`` and returns a select object.   this is used for structural changes
+    mostly to accommodate for LIMIT/OFFSET schemes
+
+    """
+
+    def visit_select(
+        self,
+        select_stmt,
+        asfrom=False,
+        insert_into=False,
+        fromhints=None,
+        compound_index=None,
+        select_wraps_for=None,
+        lateral=False,
+        from_linter=None,
+        **kwargs,
+    ):
+        assert select_wraps_for is None, (
+            "SQLAlchemy 1.4 requires use of "
+            "the translate_select_structure hook for structural "
+            "translations of SELECT objects"
+        )
+
+        # initial setup of SELECT.  the compile_state_factory may now
+        # be creating a totally different SELECT from the one that was
+        # passed in.  for ORM use this will convert from an ORM-state
+        # SELECT to a regular "Core" SELECT.  other composed operations
+        # such as computation of joins will be performed.
+
+        kwargs["within_columns_clause"] = False
+
+        compile_state = select_stmt._compile_state_factory(
+            select_stmt, self, **kwargs
+        )
+        kwargs["ambiguous_table_name_map"] = (
+            compile_state._ambiguous_table_name_map
+        )
+
+        select_stmt = compile_state.statement
+
+        toplevel = not self.stack
+
+        if toplevel and not self.compile_state:
+            self.compile_state = compile_state
+
+        is_embedded_select = compound_index is not None or insert_into
+
+        # translate step for Oracle, SQL Server which often need to
+        # restructure the SELECT to allow for LIMIT/OFFSET and possibly
+        # other conditions
+        if self.translate_select_structure:
+            new_select_stmt = self.translate_select_structure(
+                select_stmt, asfrom=asfrom, **kwargs
+            )
+
+            # if SELECT was restructured, maintain a link to the originals
+            # and assemble a new compile state
+            if new_select_stmt is not select_stmt:
+                compile_state_wraps_for = compile_state
+                select_wraps_for = select_stmt
+                select_stmt = new_select_stmt
+
+                compile_state = select_stmt._compile_state_factory(
+                    select_stmt, self, **kwargs
+                )
+                select_stmt = compile_state.statement
+
+        entry = self._default_stack_entry if toplevel else self.stack[-1]
+
+        populate_result_map = need_column_expressions = (
+            toplevel
+            or entry.get("need_result_map_for_compound", False)
+            or entry.get("need_result_map_for_nested", False)
+        )
+
+        # indicates there is a CompoundSelect in play and we are not the
+        # first select
+        if compound_index:
+            populate_result_map = False
+
+        # this was first proposed as part of #3372; however, it is not
+        # reached in current tests and could possibly be an assertion
+        # instead.
+        if not populate_result_map and "add_to_result_map" in kwargs:
+            del kwargs["add_to_result_map"]
+
+        froms = self._setup_select_stack(
+            select_stmt, compile_state, entry, asfrom, lateral, compound_index
+        )
+
+        column_clause_args = kwargs.copy()
+        column_clause_args.update(
+            {"within_label_clause": False, "within_columns_clause": False}
+        )
+
+        text = "SELECT "  # we're off to a good start !
+
+        if select_stmt._hints:
+            hint_text, byfrom = self._setup_select_hints(select_stmt)
+            if hint_text:
+                text += hint_text + " "
+        else:
+            byfrom = None
+
+        if select_stmt._independent_ctes:
+            self._dispatch_independent_ctes(select_stmt, kwargs)
+
+        if select_stmt._prefixes:
+            text += self._generate_prefixes(
+                select_stmt, select_stmt._prefixes, **kwargs
+            )
+
+        text += self.get_select_precolumns(select_stmt, **kwargs)
+        # the actual list of columns to print in the SELECT column list.
+        inner_columns = [
+            c
+            for c in [
+                self._label_select_column(
+                    select_stmt,
+                    column,
+                    populate_result_map,
+                    asfrom,
+                    column_clause_args,
+                    name=name,
+                    proxy_name=proxy_name,
+                    fallback_label_name=fallback_label_name,
+                    column_is_repeated=repeated,
+                    need_column_expressions=need_column_expressions,
+                )
+                for (
+                    name,
+                    proxy_name,
+                    fallback_label_name,
+                    column,
+                    repeated,
+                ) in compile_state.columns_plus_names
+            ]
+            if c is not None
+        ]
+
+        if populate_result_map and select_wraps_for is not None:
+            # if this select was generated from translate_select,
+            # rewrite the targeted columns in the result map
+
+            translate = dict(
+                zip(
+                    [
+                        name
+                        for (
+                            key,
+                            proxy_name,
+                            fallback_label_name,
+                            name,
+                            repeated,
+                        ) in compile_state.columns_plus_names
+                    ],
+                    [
+                        name
+                        for (
+                            key,
+                            proxy_name,
+                            fallback_label_name,
+                            name,
+                            repeated,
+                        ) in compile_state_wraps_for.columns_plus_names
+                    ],
+                )
+            )
+
+            self._result_columns = [
+                ResultColumnsEntry(
+                    key, name, tuple(translate.get(o, o) for o in obj), type_
+                )
+                for key, name, obj, type_ in self._result_columns
+            ]
+
+        text = self._compose_select_body(
+            text,
+            select_stmt,
+            compile_state,
+            inner_columns,
+            froms,
+            byfrom,
+            toplevel,
+            kwargs,
+        )
+
+        if select_stmt._statement_hints:
+            per_dialect = [
+                ht
+                for (dialect_name, ht) in select_stmt._statement_hints
+                if dialect_name in ("*", self.dialect.name)
+            ]
+            if per_dialect:
+                text += " " + self.get_statement_hint_text(per_dialect)
+
+        # In compound query, CTEs are shared at the compound level
+        if self.ctes and (not is_embedded_select or toplevel):
+            nesting_level = len(self.stack) if not toplevel else None
+            text = self._render_cte_clause(nesting_level=nesting_level) + text
+
+        if select_stmt._suffixes:
+            text += " " + self._generate_prefixes(
+                select_stmt, select_stmt._suffixes, **kwargs
+            )
+
+        self.stack.pop(-1)
+
+        return text
+
+    def _setup_select_hints(
+        self, select: Select[Any]
+    ) -> Tuple[str, _FromHintsType]:
+        byfrom = {
+            from_: hinttext
+            % {"name": from_._compiler_dispatch(self, ashint=True)}
+            for (from_, dialect), hinttext in select._hints.items()
+            if dialect in ("*", self.dialect.name)
+        }
+        hint_text = self.get_select_hint_text(byfrom)
+        return hint_text, byfrom
+
+    def _setup_select_stack(
+        self, select, compile_state, entry, asfrom, lateral, compound_index
+    ):
+        correlate_froms = entry["correlate_froms"]
+        asfrom_froms = entry["asfrom_froms"]
+
+        if compound_index == 0:
+            entry["select_0"] = select
+        elif compound_index:
+            select_0 = entry["select_0"]
+            numcols = len(select_0._all_selected_columns)
+
+            if len(compile_state.columns_plus_names) != numcols:
+                raise exc.CompileError(
+                    "All selectables passed to "
+                    "CompoundSelect must have identical numbers of "
+                    "columns; select #%d has %d columns, select "
+                    "#%d has %d"
+                    % (
+                        1,
+                        numcols,
+                        compound_index + 1,
+                        len(select._all_selected_columns),
+                    )
+                )
+
+        if asfrom and not lateral:
+            froms = compile_state._get_display_froms(
+                explicit_correlate_froms=correlate_froms.difference(
+                    asfrom_froms
+                ),
+                implicit_correlate_froms=(),
+            )
+        else:
+            froms = compile_state._get_display_froms(
+                explicit_correlate_froms=correlate_froms,
+                implicit_correlate_froms=asfrom_froms,
+            )
+
+        new_correlate_froms = set(_from_objects(*froms))
+        all_correlate_froms = new_correlate_froms.union(correlate_froms)
+
+        new_entry: _CompilerStackEntry = {
+            "asfrom_froms": new_correlate_froms,
+            "correlate_froms": all_correlate_froms,
+            "selectable": select,
+            "compile_state": compile_state,
+        }
+        self.stack.append(new_entry)
+
+        return froms
+
+    def _compose_select_body(
+        self,
+        text,
+        select,
+        compile_state,
+        inner_columns,
+        froms,
+        byfrom,
+        toplevel,
+        kwargs,
+    ):
+        text += ", ".join(inner_columns)
+
+        if self.linting & COLLECT_CARTESIAN_PRODUCTS:
+            from_linter = FromLinter({}, set())
+            warn_linting = self.linting & WARN_LINTING
+            if toplevel:
+                self.from_linter = from_linter
+        else:
+            from_linter = None
+            warn_linting = False
+
+        # adjust the whitespace for no inner columns, part of #9440,
+        # so that a no-col SELECT comes out as "SELECT WHERE..." or
+        # "SELECT FROM ...".
+        # while it would be better to have built the SELECT starting string
+        # without trailing whitespace first, then add whitespace only if inner
+        # cols were present, this breaks compatibility with various custom
+        # compilation schemes that are currently being tested.
+        if not inner_columns:
+            text = text.rstrip()
+
+        if froms:
+            text += " \nFROM "
+
+            if select._hints:
+                text += ", ".join(
+                    [
+                        f._compiler_dispatch(
+                            self,
+                            asfrom=True,
+                            fromhints=byfrom,
+                            from_linter=from_linter,
+                            **kwargs,
+                        )
+                        for f in froms
+                    ]
+                )
+            else:
+                text += ", ".join(
+                    [
+                        f._compiler_dispatch(
+                            self,
+                            asfrom=True,
+                            from_linter=from_linter,
+                            **kwargs,
+                        )
+                        for f in froms
+                    ]
+                )
+        else:
+            text += self.default_from()
+
+        if select._where_criteria:
+            t = self._generate_delimited_and_list(
+                select._where_criteria, from_linter=from_linter, **kwargs
+            )
+            if t:
+                text += " \nWHERE " + t
+
+        if warn_linting:
+            assert from_linter is not None
+            from_linter.warn()
+
+        if select._group_by_clauses:
+            text += self.group_by_clause(select, **kwargs)
+
+        if select._having_criteria:
+            t = self._generate_delimited_and_list(
+                select._having_criteria, **kwargs
+            )
+            if t:
+                text += " \nHAVING " + t
+
+        if select._order_by_clauses:
+            text += self.order_by_clause(select, **kwargs)
+
+        if select._has_row_limiting_clause:
+            text += self._row_limit_clause(select, **kwargs)
+
+        if select._for_update_arg is not None:
+            text += self.for_update_clause(select, **kwargs)
+
+        return text
+
+    def _generate_prefixes(self, stmt, prefixes, **kw):
+        clause = " ".join(
+            prefix._compiler_dispatch(self, **kw)
+            for prefix, dialect_name in prefixes
+            if dialect_name in (None, "*") or dialect_name == self.dialect.name
+        )
+        if clause:
+            clause += " "
+        return clause
+
+    def _render_cte_clause(
+        self,
+        nesting_level=None,
+        include_following_stack=False,
+    ):
+        """
+        include_following_stack
+            Also render the nesting CTEs on the next stack. Useful for
+            SQL structures like UNION or INSERT that can wrap SELECT
+            statements containing nesting CTEs.
+        """
+        if not self.ctes:
+            return ""
+
+        ctes: MutableMapping[CTE, str]
+
+        if nesting_level and nesting_level > 1:
+            ctes = util.OrderedDict()
+            for cte in list(self.ctes.keys()):
+                cte_level, cte_name, cte_opts = self.level_name_by_cte[
+                    cte._get_reference_cte()
+                ]
+                nesting = cte.nesting or cte_opts.nesting
+                is_rendered_level = cte_level == nesting_level or (
+                    include_following_stack and cte_level == nesting_level + 1
+                )
+                if not (nesting and is_rendered_level):
+                    continue
+
+                ctes[cte] = self.ctes[cte]
+
+        else:
+            ctes = self.ctes
+
+        if not ctes:
+            return ""
+        ctes_recursive = any([cte.recursive for cte in ctes])
+
+        cte_text = self.get_cte_preamble(ctes_recursive) + " "
+        cte_text += ", \n".join([txt for txt in ctes.values()])
+        cte_text += "\n "
+
+        if nesting_level and nesting_level > 1:
+            for cte in list(ctes.keys()):
+                cte_level, cte_name, cte_opts = self.level_name_by_cte[
+                    cte._get_reference_cte()
+                ]
+                del self.ctes[cte]
+                del self.ctes_by_level_name[(cte_level, cte_name)]
+                del self.level_name_by_cte[cte._get_reference_cte()]
+
+        return cte_text
+
+    def get_cte_preamble(self, recursive):
+        if recursive:
+            return "WITH RECURSIVE"
+        else:
+            return "WITH"
+
+    def get_select_precolumns(self, select, **kw):
+        """Called when building a ``SELECT`` statement, position is just
+        before column list.
+
+        """
+        if select._distinct_on:
+            util.warn_deprecated(
+                "DISTINCT ON is currently supported only by the PostgreSQL "
+                "dialect.  Use of DISTINCT ON for other backends is currently "
+                "silently ignored, however this usage is deprecated, and will "
+                "raise CompileError in a future release for all backends "
+                "that do not support this syntax.",
+                version="1.4",
+            )
+        return "DISTINCT " if select._distinct else ""
+
+    def group_by_clause(self, select, **kw):
+        """allow dialects to customize how GROUP BY is rendered."""
+
+        group_by = self._generate_delimited_list(
+            select._group_by_clauses, OPERATORS[operators.comma_op], **kw
+        )
+        if group_by:
+            return " GROUP BY " + group_by
+        else:
+            return ""
+
+    def order_by_clause(self, select, **kw):
+        """allow dialects to customize how ORDER BY is rendered."""
+
+        order_by = self._generate_delimited_list(
+            select._order_by_clauses, OPERATORS[operators.comma_op], **kw
+        )
+
+        if order_by:
+            return " ORDER BY " + order_by
+        else:
+            return ""
+
+    def for_update_clause(self, select, **kw):
+        return " FOR UPDATE"
+
+    def returning_clause(
+        self,
+        stmt: UpdateBase,
+        returning_cols: Sequence[ColumnElement[Any]],
+        *,
+        populate_result_map: bool,
+        **kw: Any,
+    ) -> str:
+        columns = [
+            self._label_returning_column(
+                stmt,
+                column,
+                populate_result_map,
+                fallback_label_name=fallback_label_name,
+                column_is_repeated=repeated,
+                name=name,
+                proxy_name=proxy_name,
+                **kw,
+            )
+            for (
+                name,
+                proxy_name,
+                fallback_label_name,
+                column,
+                repeated,
+            ) in stmt._generate_columns_plus_names(
+                True, cols=base._select_iterables(returning_cols)
+            )
+        ]
+
+        return "RETURNING " + ", ".join(columns)
+
+    def limit_clause(self, select, **kw):
+        text = ""
+        if select._limit_clause is not None:
+            text += "\n LIMIT " + self.process(select._limit_clause, **kw)
+        if select._offset_clause is not None:
+            if select._limit_clause is None:
+                text += "\n LIMIT -1"
+            text += " OFFSET " + self.process(select._offset_clause, **kw)
+        return text
+
+    def fetch_clause(
+        self,
+        select,
+        fetch_clause=None,
+        require_offset=False,
+        use_literal_execute_for_simple_int=False,
+        **kw,
+    ):
+        if fetch_clause is None:
+            fetch_clause = select._fetch_clause
+            fetch_clause_options = select._fetch_clause_options
+        else:
+            fetch_clause_options = {"percent": False, "with_ties": False}
+
+        text = ""
+
+        if select._offset_clause is not None:
+            offset_clause = select._offset_clause
+            if (
+                use_literal_execute_for_simple_int
+                and select._simple_int_clause(offset_clause)
+            ):
+                offset_clause = offset_clause.render_literal_execute()
+            offset_str = self.process(offset_clause, **kw)
+            text += "\n OFFSET %s ROWS" % offset_str
+        elif require_offset:
+            text += "\n OFFSET 0 ROWS"
+
+        if fetch_clause is not None:
+            if (
+                use_literal_execute_for_simple_int
+                and select._simple_int_clause(fetch_clause)
+            ):
+                fetch_clause = fetch_clause.render_literal_execute()
+            text += "\n FETCH FIRST %s%s ROWS %s" % (
+                self.process(fetch_clause, **kw),
+                " PERCENT" if fetch_clause_options["percent"] else "",
+                "WITH TIES" if fetch_clause_options["with_ties"] else "ONLY",
+            )
+        return text
+
+    def visit_table(
+        self,
+        table,
+        asfrom=False,
+        iscrud=False,
+        ashint=False,
+        fromhints=None,
+        use_schema=True,
+        from_linter=None,
+        ambiguous_table_name_map=None,
+        **kwargs,
+    ):
+        if from_linter:
+            from_linter.froms[table] = table.fullname
+
+        if asfrom or ashint:
+            effective_schema = self.preparer.schema_for_object(table)
+
+            if use_schema and effective_schema:
+                ret = (
+                    self.preparer.quote_schema(effective_schema)
+                    + "."
+                    + self.preparer.quote(table.name)
+                )
+            else:
+                ret = self.preparer.quote(table.name)
+
+                if (
+                    not effective_schema
+                    and ambiguous_table_name_map
+                    and table.name in ambiguous_table_name_map
+                ):
+                    anon_name = self._truncated_identifier(
+                        "alias", ambiguous_table_name_map[table.name]
+                    )
+
+                    ret = ret + self.get_render_as_alias_suffix(
+                        self.preparer.format_alias(None, anon_name)
+                    )
+
+            if fromhints and table in fromhints:
+                ret = self.format_from_hint_text(
+                    ret, table, fromhints[table], iscrud
+                )
+            return ret
+        else:
+            return ""
+
+    def visit_join(self, join, asfrom=False, from_linter=None, **kwargs):
+        if from_linter:
+            from_linter.edges.update(
+                itertools.product(
+                    _de_clone(join.left._from_objects),
+                    _de_clone(join.right._from_objects),
+                )
+            )
+
+        if join.full:
+            join_type = " FULL OUTER JOIN "
+        elif join.isouter:
+            join_type = " LEFT OUTER JOIN "
+        else:
+            join_type = " JOIN "
+        return (
+            join.left._compiler_dispatch(
+                self, asfrom=True, from_linter=from_linter, **kwargs
+            )
+            + join_type
+            + join.right._compiler_dispatch(
+                self, asfrom=True, from_linter=from_linter, **kwargs
+            )
+            + " ON "
+            # TODO: likely need asfrom=True here?
+            + join.onclause._compiler_dispatch(
+                self, from_linter=from_linter, **kwargs
+            )
+        )
+
+    def _setup_crud_hints(self, stmt, table_text):
+        dialect_hints = {
+            table: hint_text
+            for (table, dialect), hint_text in stmt._hints.items()
+            if dialect in ("*", self.dialect.name)
+        }
+        if stmt.table in dialect_hints:
+            table_text = self.format_from_hint_text(
+                table_text, stmt.table, dialect_hints[stmt.table], True
+            )
+        return dialect_hints, table_text
+
+    # within the realm of "insertmanyvalues sentinel columns",
+    # these lookups match different kinds of Column() configurations
+    # to specific backend capabilities.  they are broken into two
+    # lookups, one for autoincrement columns and the other for non
+    # autoincrement columns
+    _sentinel_col_non_autoinc_lookup = util.immutabledict(
+        {
+            _SentinelDefaultCharacterization.CLIENTSIDE: (
+                InsertmanyvaluesSentinelOpts._SUPPORTED_OR_NOT
+            ),
+            _SentinelDefaultCharacterization.SENTINEL_DEFAULT: (
+                InsertmanyvaluesSentinelOpts._SUPPORTED_OR_NOT
+            ),
+            _SentinelDefaultCharacterization.NONE: (
+                InsertmanyvaluesSentinelOpts._SUPPORTED_OR_NOT
+            ),
+            _SentinelDefaultCharacterization.IDENTITY: (
+                InsertmanyvaluesSentinelOpts.IDENTITY
+            ),
+            _SentinelDefaultCharacterization.SEQUENCE: (
+                InsertmanyvaluesSentinelOpts.SEQUENCE
+            ),
+        }
+    )
+    _sentinel_col_autoinc_lookup = _sentinel_col_non_autoinc_lookup.union(
+        {
+            _SentinelDefaultCharacterization.NONE: (
+                InsertmanyvaluesSentinelOpts.AUTOINCREMENT
+            ),
+        }
+    )
+
+    def _get_sentinel_column_for_table(
+        self, table: Table
+    ) -> Optional[Sequence[Column[Any]]]:
+        """given a :class:`.Table`, return a usable sentinel column or
+        columns for this dialect if any.
+
+        Return None if no sentinel columns could be identified, or raise an
+        error if a column was marked as a sentinel explicitly but isn't
+        compatible with this dialect.
+
+        """
+
+        sentinel_opts = self.dialect.insertmanyvalues_implicit_sentinel
+        sentinel_characteristics = table._sentinel_column_characteristics
+
+        sent_cols = sentinel_characteristics.columns
+
+        if sent_cols is None:
+            return None
+
+        if sentinel_characteristics.is_autoinc:
+            bitmask = self._sentinel_col_autoinc_lookup.get(
+                sentinel_characteristics.default_characterization, 0
+            )
+        else:
+            bitmask = self._sentinel_col_non_autoinc_lookup.get(
+                sentinel_characteristics.default_characterization, 0
+            )
+
+        if sentinel_opts & bitmask:
+            return sent_cols
+
+        if sentinel_characteristics.is_explicit:
+            # a column was explicitly marked as insert_sentinel=True,
+            # however it is not compatible with this dialect.   they should
+            # not indicate this column as a sentinel if they need to include
+            # this dialect.
+
+            # TODO: do we want non-primary key explicit sentinel cols
+            # that can gracefully degrade for some backends?
+            # insert_sentinel="degrade" perhaps.  not for the initial release.
+            # I am hoping people are generally not dealing with this sentinel
+            # business at all.
+
+            # if is_explicit is True, there will be only one sentinel column.
+
+            raise exc.InvalidRequestError(
+                f"Column {sent_cols[0]} can't be explicitly "
+                "marked as a sentinel column when using the "
+                f"{self.dialect.name} dialect, as the "
+                "particular type of default generation on this column is "
+                "not currently compatible with this dialect's specific "
+                f"INSERT..RETURNING syntax which can receive the "
+                "server-generated value in "
+                "a deterministic way.  To remove this error, remove "
+                "insert_sentinel=True from primary key autoincrement "
+                "columns; these columns are automatically used as "
+                "sentinels for supported dialects in any case."
+            )
+
+        return None
+
+    def _deliver_insertmanyvalues_batches(
+        self,
+        statement: str,
+        parameters: _DBAPIMultiExecuteParams,
+        compiled_parameters: List[_MutableCoreSingleExecuteParams],
+        generic_setinputsizes: Optional[_GenericSetInputSizesType],
+        batch_size: int,
+        sort_by_parameter_order: bool,
+        schema_translate_map: Optional[SchemaTranslateMapType],
+    ) -> Iterator[_InsertManyValuesBatch]:
+        imv = self._insertmanyvalues
+        assert imv is not None
+
+        if not imv.sentinel_param_keys:
+            _sentinel_from_params = None
+        else:
+            _sentinel_from_params = operator.itemgetter(
+                *imv.sentinel_param_keys
+            )
+
+        lenparams = len(parameters)
+        if imv.is_default_expr and not self.dialect.supports_default_metavalue:
+            # backend doesn't support
+            # INSERT INTO table (pk_col) VALUES (DEFAULT), (DEFAULT), ...
+            # at the moment this is basically SQL Server due to
+            # not being able to use DEFAULT for identity column
+            # just yield out that many single statements!  still
+            # faster than a whole connection.execute() call ;)
+            #
+            # note we still are taking advantage of the fact that we know
+            # we are using RETURNING.   The generalized approach of fetching
+            # cursor.lastrowid etc. still goes through the more heavyweight
+            # "ExecutionContext per statement" system as it isn't usable
+            # as a generic "RETURNING" approach
+            use_row_at_a_time = True
+            downgraded = False
+        elif not self.dialect.supports_multivalues_insert or (
+            sort_by_parameter_order
+            and self._result_columns
+            and (imv.sentinel_columns is None or imv.includes_upsert_behaviors)
+        ):
+            # deterministic order was requested and the compiler could
+            # not organize sentinel columns for this dialect/statement.
+            # use row at a time
+            use_row_at_a_time = True
+            downgraded = True
+        else:
+            use_row_at_a_time = False
+            downgraded = False
+
+        if use_row_at_a_time:
+            for batchnum, (param, compiled_param) in enumerate(
+                cast(
+                    "Sequence[Tuple[_DBAPISingleExecuteParams, _MutableCoreSingleExecuteParams]]",  # noqa: E501
+                    zip(parameters, compiled_parameters),
+                ),
+                1,
+            ):
+                yield _InsertManyValuesBatch(
+                    statement,
+                    param,
+                    generic_setinputsizes,
+                    [param],
+                    (
+                        [_sentinel_from_params(compiled_param)]
+                        if _sentinel_from_params
+                        else []
+                    ),
+                    1,
+                    batchnum,
+                    lenparams,
+                    sort_by_parameter_order,
+                    downgraded,
+                )
+            return
+
+        if schema_translate_map:
+            rst = functools.partial(
+                self.preparer._render_schema_translates,
+                schema_translate_map=schema_translate_map,
+            )
+        else:
+            rst = None
+
+        imv_single_values_expr = imv.single_values_expr
+        if rst:
+            imv_single_values_expr = rst(imv_single_values_expr)
+
+        executemany_values = f"({imv_single_values_expr})"
+        statement = statement.replace(executemany_values, "__EXECMANY_TOKEN__")
+
+        # Use optional insertmanyvalues_max_parameters
+        # to further shrink the batch size so that there are no more than
+        # insertmanyvalues_max_parameters params.
+        # Currently used by SQL Server, which limits statements to 2100 bound
+        # parameters (actually 2099).
+        max_params = self.dialect.insertmanyvalues_max_parameters
+        if max_params:
+            total_num_of_params = len(self.bind_names)
+            num_params_per_batch = len(imv.insert_crud_params)
+            num_params_outside_of_batch = (
+                total_num_of_params - num_params_per_batch
+            )
+            batch_size = min(
+                batch_size,
+                (
+                    (max_params - num_params_outside_of_batch)
+                    // num_params_per_batch
+                ),
+            )
+
+        batches = cast("List[Sequence[Any]]", list(parameters))
+        compiled_batches = cast(
+            "List[Sequence[Any]]", list(compiled_parameters)
+        )
+
+        processed_setinputsizes: Optional[_GenericSetInputSizesType] = None
+        batchnum = 1
+        total_batches = lenparams // batch_size + (
+            1 if lenparams % batch_size else 0
+        )
+
+        insert_crud_params = imv.insert_crud_params
+        assert insert_crud_params is not None
+
+        if rst:
+            insert_crud_params = [
+                (col, key, rst(expr), st)
+                for col, key, expr, st in insert_crud_params
+            ]
+
+        escaped_bind_names: Mapping[str, str]
+        expand_pos_lower_index = expand_pos_upper_index = 0
+
+        if not self.positional:
+            if self.escaped_bind_names:
+                escaped_bind_names = self.escaped_bind_names
+            else:
+                escaped_bind_names = {}
+
+            all_keys = set(parameters[0])
+
+            def apply_placeholders(keys, formatted):
+                for key in keys:
+                    key = escaped_bind_names.get(key, key)
+                    formatted = formatted.replace(
+                        self.bindtemplate % {"name": key},
+                        self.bindtemplate
+                        % {"name": f"{key}__EXECMANY_INDEX__"},
+                    )
+                return formatted
+
+            if imv.embed_values_counter:
+                imv_values_counter = ", _IMV_VALUES_COUNTER"
+            else:
+                imv_values_counter = ""
+            formatted_values_clause = f"""({', '.join(
+                apply_placeholders(bind_keys, formatted)
+                for _, _, formatted, bind_keys in insert_crud_params
+            )}{imv_values_counter})"""
+
+            keys_to_replace = all_keys.intersection(
+                escaped_bind_names.get(key, key)
+                for _, _, _, bind_keys in insert_crud_params
+                for key in bind_keys
+            )
+            base_parameters = {
+                key: parameters[0][key]
+                for key in all_keys.difference(keys_to_replace)
+            }
+            executemany_values_w_comma = ""
+        else:
+            formatted_values_clause = ""
+            keys_to_replace = set()
+            base_parameters = {}
+
+            if imv.embed_values_counter:
+                executemany_values_w_comma = (
+                    f"({imv_single_values_expr}, _IMV_VALUES_COUNTER), "
+                )
+            else:
+                executemany_values_w_comma = f"({imv_single_values_expr}), "
+
+            all_names_we_will_expand: Set[str] = set()
+            for elem in imv.insert_crud_params:
+                all_names_we_will_expand.update(elem[3])
+
+            # get the start and end position in a particular list
+            # of parameters where we will be doing the "expanding".
+            # statements can have params on either side or both sides,
+            # given RETURNING and CTEs
+            if all_names_we_will_expand:
+                positiontup = self.positiontup
+                assert positiontup is not None
+
+                all_expand_positions = {
+                    idx
+                    for idx, name in enumerate(positiontup)
+                    if name in all_names_we_will_expand
+                }
+                expand_pos_lower_index = min(all_expand_positions)
+                expand_pos_upper_index = max(all_expand_positions) + 1
+                assert (
+                    len(all_expand_positions)
+                    == expand_pos_upper_index - expand_pos_lower_index
+                )
+
+            if self._numeric_binds:
+                escaped = re.escape(self._numeric_binds_identifier_char)
+                executemany_values_w_comma = re.sub(
+                    rf"{escaped}\d+", "%s", executemany_values_w_comma
+                )
+
+        while batches:
+            batch = batches[0:batch_size]
+            compiled_batch = compiled_batches[0:batch_size]
+
+            batches[0:batch_size] = []
+            compiled_batches[0:batch_size] = []
+
+            if batches:
+                current_batch_size = batch_size
+            else:
+                current_batch_size = len(batch)
+
+            if generic_setinputsizes:
+                # if setinputsizes is present, expand this collection to
+                # suit the batch length as well
+                # currently this will be mssql+pyodbc for internal dialects
+                processed_setinputsizes = [
+                    (new_key, len_, typ)
+                    for new_key, len_, typ in (
+                        (f"{key}_{index}", len_, typ)
+                        for index in range(current_batch_size)
+                        for key, len_, typ in generic_setinputsizes
+                    )
+                ]
+
+            replaced_parameters: Any
+            if self.positional:
+                num_ins_params = imv.num_positional_params_counted
+
+                batch_iterator: Iterable[Sequence[Any]]
+                extra_params_left: Sequence[Any]
+                extra_params_right: Sequence[Any]
+
+                if num_ins_params == len(batch[0]):
+                    extra_params_left = extra_params_right = ()
+                    batch_iterator = batch
+                else:
+                    extra_params_left = batch[0][:expand_pos_lower_index]
+                    extra_params_right = batch[0][expand_pos_upper_index:]
+                    batch_iterator = (
+                        b[expand_pos_lower_index:expand_pos_upper_index]
+                        for b in batch
+                    )
+
+                if imv.embed_values_counter:
+                    expanded_values_string = (
+                        "".join(
+                            executemany_values_w_comma.replace(
+                                "_IMV_VALUES_COUNTER", str(i)
+                            )
+                            for i, _ in enumerate(batch)
+                        )
+                    )[:-2]
+                else:
+                    expanded_values_string = (
+                        (executemany_values_w_comma * current_batch_size)
+                    )[:-2]
+
+                if self._numeric_binds and num_ins_params > 0:
+                    # numeric will always number the parameters inside of
+                    # VALUES (and thus order self.positiontup) to be higher
+                    # than non-VALUES parameters, no matter where in the
+                    # statement those non-VALUES parameters appear (this is
+                    # ensured in _process_numeric by numbering first all
+                    # params that are not in _values_bindparam)
+                    # therefore all extra params are always
+                    # on the left side and numbered lower than the VALUES
+                    # parameters
+                    assert not extra_params_right
+
+                    start = expand_pos_lower_index + 1
+                    end = num_ins_params * (current_batch_size) + start
+
+                    # need to format here, since statement may contain
+                    # unescaped %, while values_string contains just (%s, %s)
+                    positions = tuple(
+                        f"{self._numeric_binds_identifier_char}{i}"
+                        for i in range(start, end)
+                    )
+                    expanded_values_string = expanded_values_string % positions
+
+                replaced_statement = statement.replace(
+                    "__EXECMANY_TOKEN__", expanded_values_string
+                )
+
+                replaced_parameters = tuple(
+                    itertools.chain.from_iterable(batch_iterator)
+                )
+
+                replaced_parameters = (
+                    extra_params_left
+                    + replaced_parameters
+                    + extra_params_right
+                )
+
+            else:
+                replaced_values_clauses = []
+                replaced_parameters = base_parameters.copy()
+
+                for i, param in enumerate(batch):
+                    fmv = formatted_values_clause.replace(
+                        "EXECMANY_INDEX__", str(i)
+                    )
+                    if imv.embed_values_counter:
+                        fmv = fmv.replace("_IMV_VALUES_COUNTER", str(i))
+
+                    replaced_values_clauses.append(fmv)
+                    replaced_parameters.update(
+                        {f"{key}__{i}": param[key] for key in keys_to_replace}
+                    )
+
+                replaced_statement = statement.replace(
+                    "__EXECMANY_TOKEN__",
+                    ", ".join(replaced_values_clauses),
+                )
+
+            yield _InsertManyValuesBatch(
+                replaced_statement,
+                replaced_parameters,
+                processed_setinputsizes,
+                batch,
+                (
+                    [_sentinel_from_params(cb) for cb in compiled_batch]
+                    if _sentinel_from_params
+                    else []
+                ),
+                current_batch_size,
+                batchnum,
+                total_batches,
+                sort_by_parameter_order,
+                False,
+            )
+            batchnum += 1
+
+    def visit_insert(
+        self, insert_stmt, visited_bindparam=None, visiting_cte=None, **kw
+    ):
+        compile_state = insert_stmt._compile_state_factory(
+            insert_stmt, self, **kw
+        )
+        insert_stmt = compile_state.statement
+
+        if visiting_cte is not None:
+            kw["visiting_cte"] = visiting_cte
+            toplevel = False
+        else:
+            toplevel = not self.stack
+
+        if toplevel:
+            self.isinsert = True
+            if not self.dml_compile_state:
+                self.dml_compile_state = compile_state
+            if not self.compile_state:
+                self.compile_state = compile_state
+
+        self.stack.append(
+            {
+                "correlate_froms": set(),
+                "asfrom_froms": set(),
+                "selectable": insert_stmt,
+            }
+        )
+
+        counted_bindparam = 0
+
+        # reset any incoming "visited_bindparam" collection
+        visited_bindparam = None
+
+        # for positional, insertmanyvalues needs to know how many
+        # bound parameters are in the VALUES sequence; there's no simple
+        # rule because default expressions etc. can have zero or more
+        # params inside them.   After multiple attempts to figure this out,
+        # this very simplistic "count after" works and is
+        # likely the least amount of callcounts, though looks clumsy
+        if self.positional and visiting_cte is None:
+            # if we are inside a CTE, don't count parameters
+            # here since they wont be for insertmanyvalues. keep
+            # visited_bindparam at None so no counting happens.
+            # see #9173
+            visited_bindparam = []
+
+        crud_params_struct = crud._get_crud_params(
+            self,
+            insert_stmt,
+            compile_state,
+            toplevel,
+            visited_bindparam=visited_bindparam,
+            **kw,
+        )
+
+        if self.positional and visited_bindparam is not None:
+            counted_bindparam = len(visited_bindparam)
+            if self._numeric_binds:
+                if self._values_bindparam is not None:
+                    self._values_bindparam += visited_bindparam
+                else:
+                    self._values_bindparam = visited_bindparam
+
+        crud_params_single = crud_params_struct.single_params
+
+        if (
+            not crud_params_single
+            and not self.dialect.supports_default_values
+            and not self.dialect.supports_default_metavalue
+            and not self.dialect.supports_empty_insert
+        ):
+            raise exc.CompileError(
+                "The '%s' dialect with current database "
+                "version settings does not support empty "
+                "inserts." % self.dialect.name
+            )
+
+        if compile_state._has_multi_parameters:
+            if not self.dialect.supports_multivalues_insert:
+                raise exc.CompileError(
+                    "The '%s' dialect with current database "
+                    "version settings does not support "
+                    "in-place multirow inserts." % self.dialect.name
+                )
+            elif (
+                self.implicit_returning or insert_stmt._returning
+            ) and insert_stmt._sort_by_parameter_order:
+                raise exc.CompileError(
+                    "RETURNING cannot be determinstically sorted when "
+                    "using an INSERT which includes multi-row values()."
+                )
+            crud_params_single = crud_params_struct.single_params
+        else:
+            crud_params_single = crud_params_struct.single_params
+
+        preparer = self.preparer
+        supports_default_values = self.dialect.supports_default_values
+
+        text = "INSERT "
+
+        if insert_stmt._prefixes:
+            text += self._generate_prefixes(
+                insert_stmt, insert_stmt._prefixes, **kw
+            )
+
+        text += "INTO "
+        table_text = preparer.format_table(insert_stmt.table)
+
+        if insert_stmt._hints:
+            _, table_text = self._setup_crud_hints(insert_stmt, table_text)
+
+        if insert_stmt._independent_ctes:
+            self._dispatch_independent_ctes(insert_stmt, kw)
+
+        text += table_text
+
+        if crud_params_single or not supports_default_values:
+            text += " (%s)" % ", ".join(
+                [expr for _, expr, _, _ in crud_params_single]
+            )
+
+        # look for insertmanyvalues attributes that would have been configured
+        # by crud.py as it scanned through the columns to be part of the
+        # INSERT
+        use_insertmanyvalues = crud_params_struct.use_insertmanyvalues
+        named_sentinel_params: Optional[Sequence[str]] = None
+        add_sentinel_cols = None
+        implicit_sentinel = False
+
+        returning_cols = self.implicit_returning or insert_stmt._returning
+        if returning_cols:
+            add_sentinel_cols = crud_params_struct.use_sentinel_columns
+            if add_sentinel_cols is not None:
+                assert use_insertmanyvalues
+
+                # search for the sentinel column explicitly present
+                # in the INSERT columns list, and additionally check that
+                # this column has a bound parameter name set up that's in the
+                # parameter list.  If both of these cases are present, it means
+                # we will have a client side value for the sentinel in each
+                # parameter set.
+
+                _params_by_col = {
+                    col: param_names
+                    for col, _, _, param_names in crud_params_single
+                }
+                named_sentinel_params = []
+                for _add_sentinel_col in add_sentinel_cols:
+                    if _add_sentinel_col not in _params_by_col:
+                        named_sentinel_params = None
+                        break
+                    param_name = self._within_exec_param_key_getter(
+                        _add_sentinel_col
+                    )
+                    if param_name not in _params_by_col[_add_sentinel_col]:
+                        named_sentinel_params = None
+                        break
+                    named_sentinel_params.append(param_name)
+
+                if named_sentinel_params is None:
+                    # if we are not going to have a client side value for
+                    # the sentinel in the parameter set, that means it's
+                    # an autoincrement, an IDENTITY, or a server-side SQL
+                    # expression like nextval('seqname').  So this is
+                    # an "implicit" sentinel; we will look for it in
+                    # RETURNING
+                    # only, and then sort on it.  For this case on PG,
+                    # SQL Server we have to use a special INSERT form
+                    # that guarantees the server side function lines up with
+                    # the entries in the VALUES.
+                    if (
+                        self.dialect.insertmanyvalues_implicit_sentinel
+                        & InsertmanyvaluesSentinelOpts.ANY_AUTOINCREMENT
+                    ):
+                        implicit_sentinel = True
+                    else:
+                        # here, we are not using a sentinel at all
+                        # and we are likely the SQLite dialect.
+                        # The first add_sentinel_col that we have should not
+                        # be marked as "insert_sentinel=True".  if it was,
+                        # an error should have been raised in
+                        # _get_sentinel_column_for_table.
+                        assert not add_sentinel_cols[0]._insert_sentinel, (
+                            "sentinel selection rules should have prevented "
+                            "us from getting here for this dialect"
+                        )
+
+                # always put the sentinel columns last.  even if they are
+                # in the returning list already, they will be there twice
+                # then.
+                returning_cols = list(returning_cols) + list(add_sentinel_cols)
+
+            returning_clause = self.returning_clause(
+                insert_stmt,
+                returning_cols,
+                populate_result_map=toplevel,
+            )
+
+            if self.returning_precedes_values:
+                text += " " + returning_clause
+
+        else:
+            returning_clause = None
+
+        if insert_stmt.select is not None:
+            # placed here by crud.py
+            select_text = self.process(
+                self.stack[-1]["insert_from_select"], insert_into=True, **kw
+            )
+
+            if self.ctes and self.dialect.cte_follows_insert:
+                nesting_level = len(self.stack) if not toplevel else None
+                text += " %s%s" % (
+                    self._render_cte_clause(
+                        nesting_level=nesting_level,
+                        include_following_stack=True,
+                    ),
+                    select_text,
+                )
+            else:
+                text += " %s" % select_text
+        elif not crud_params_single and supports_default_values:
+            text += " DEFAULT VALUES"
+            if use_insertmanyvalues:
+                self._insertmanyvalues = _InsertManyValues(
+                    True,
+                    self.dialect.default_metavalue_token,
+                    cast(
+                        "List[crud._CrudParamElementStr]", crud_params_single
+                    ),
+                    counted_bindparam,
+                    sort_by_parameter_order=(
+                        insert_stmt._sort_by_parameter_order
+                    ),
+                    includes_upsert_behaviors=(
+                        insert_stmt._post_values_clause is not None
+                    ),
+                    sentinel_columns=add_sentinel_cols,
+                    num_sentinel_columns=(
+                        len(add_sentinel_cols) if add_sentinel_cols else 0
+                    ),
+                    implicit_sentinel=implicit_sentinel,
+                )
+        elif compile_state._has_multi_parameters:
+            text += " VALUES %s" % (
+                ", ".join(
+                    "(%s)"
+                    % (", ".join(value for _, _, value, _ in crud_param_set))
+                    for crud_param_set in crud_params_struct.all_multi_params
+                ),
+            )
+        else:
+            insert_single_values_expr = ", ".join(
+                [
+                    value
+                    for _, _, value, _ in cast(
+                        "List[crud._CrudParamElementStr]",
+                        crud_params_single,
+                    )
+                ]
+            )
+
+            if use_insertmanyvalues:
+                if (
+                    implicit_sentinel
+                    and (
+                        self.dialect.insertmanyvalues_implicit_sentinel
+                        & InsertmanyvaluesSentinelOpts.USE_INSERT_FROM_SELECT
+                    )
+                    # this is checking if we have
+                    # INSERT INTO table (id) VALUES (DEFAULT).
+                    and not (crud_params_struct.is_default_metavalue_only)
+                ):
+                    # if we have a sentinel column that is server generated,
+                    # then for selected backends render the VALUES list as a
+                    # subquery.  This is the orderable form supported by
+                    # PostgreSQL and SQL Server.
+                    embed_sentinel_value = True
+
+                    render_bind_casts = (
+                        self.dialect.insertmanyvalues_implicit_sentinel
+                        & InsertmanyvaluesSentinelOpts.RENDER_SELECT_COL_CASTS
+                    )
+
+                    colnames = ", ".join(
+                        f"p{i}" for i, _ in enumerate(crud_params_single)
+                    )
+
+                    if render_bind_casts:
+                        # render casts for the SELECT list.  For PG, we are
+                        # already rendering bind casts in the parameter list,
+                        # selectively for the more "tricky" types like ARRAY.
+                        # however, even for the "easy" types, if the parameter
+                        # is NULL for every entry, PG gives up and says
+                        # "it must be TEXT", which fails for other easy types
+                        # like ints.  So we cast on this side too.
+                        colnames_w_cast = ", ".join(
+                            self.render_bind_cast(
+                                col.type,
+                                col.type._unwrapped_dialect_impl(self.dialect),
+                                f"p{i}",
+                            )
+                            for i, (col, *_) in enumerate(crud_params_single)
+                        )
+                    else:
+                        colnames_w_cast = colnames
+
+                    text += (
+                        f" SELECT {colnames_w_cast} FROM "
+                        f"(VALUES ({insert_single_values_expr})) "
+                        f"AS imp_sen({colnames}, sen_counter) "
+                        "ORDER BY sen_counter"
+                    )
+                else:
+                    # otherwise, if no sentinel or backend doesn't support
+                    # orderable subquery form, use a plain VALUES list
+                    embed_sentinel_value = False
+                    text += f" VALUES ({insert_single_values_expr})"
+
+                self._insertmanyvalues = _InsertManyValues(
+                    is_default_expr=False,
+                    single_values_expr=insert_single_values_expr,
+                    insert_crud_params=cast(
+                        "List[crud._CrudParamElementStr]",
+                        crud_params_single,
+                    ),
+                    num_positional_params_counted=counted_bindparam,
+                    sort_by_parameter_order=(
+                        insert_stmt._sort_by_parameter_order
+                    ),
+                    includes_upsert_behaviors=(
+                        insert_stmt._post_values_clause is not None
+                    ),
+                    sentinel_columns=add_sentinel_cols,
+                    num_sentinel_columns=(
+                        len(add_sentinel_cols) if add_sentinel_cols else 0
+                    ),
+                    sentinel_param_keys=named_sentinel_params,
+                    implicit_sentinel=implicit_sentinel,
+                    embed_values_counter=embed_sentinel_value,
+                )
+
+            else:
+                text += f" VALUES ({insert_single_values_expr})"
+
+        if insert_stmt._post_values_clause is not None:
+            post_values_clause = self.process(
+                insert_stmt._post_values_clause, **kw
+            )
+            if post_values_clause:
+                text += " " + post_values_clause
+
+        if returning_clause and not self.returning_precedes_values:
+            text += " " + returning_clause
+
+        if self.ctes and not self.dialect.cte_follows_insert:
+            nesting_level = len(self.stack) if not toplevel else None
+            text = (
+                self._render_cte_clause(
+                    nesting_level=nesting_level,
+                    include_following_stack=True,
+                )
+                + text
+            )
+
+        self.stack.pop(-1)
+
+        return text
+
+    def update_limit_clause(self, update_stmt):
+        """Provide a hook for MySQL to add LIMIT to the UPDATE"""
+        return None
+
+    def delete_limit_clause(self, delete_stmt):
+        """Provide a hook for MySQL to add LIMIT to the DELETE"""
+        return None
+
+    def update_tables_clause(self, update_stmt, from_table, extra_froms, **kw):
+        """Provide a hook to override the initial table clause
+        in an UPDATE statement.
+
+        MySQL overrides this.
+
+        """
+        kw["asfrom"] = True
+        return from_table._compiler_dispatch(self, iscrud=True, **kw)
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Provide a hook to override the generation of an
+        UPDATE..FROM clause.
+
+        MySQL and MSSQL override this.
+
+        """
+        raise NotImplementedError(
+            "This backend does not support multiple-table "
+            "criteria within UPDATE"
+        )
+
+    def visit_update(self, update_stmt, visiting_cte=None, **kw):
+        compile_state = update_stmt._compile_state_factory(
+            update_stmt, self, **kw
+        )
+        update_stmt = compile_state.statement
+
+        if visiting_cte is not None:
+            kw["visiting_cte"] = visiting_cte
+            toplevel = False
+        else:
+            toplevel = not self.stack
+
+        if toplevel:
+            self.isupdate = True
+            if not self.dml_compile_state:
+                self.dml_compile_state = compile_state
+            if not self.compile_state:
+                self.compile_state = compile_state
+
+        if self.linting & COLLECT_CARTESIAN_PRODUCTS:
+            from_linter = FromLinter({}, set())
+            warn_linting = self.linting & WARN_LINTING
+            if toplevel:
+                self.from_linter = from_linter
+        else:
+            from_linter = None
+            warn_linting = False
+
+        extra_froms = compile_state._extra_froms
+        is_multitable = bool(extra_froms)
+
+        if is_multitable:
+            # main table might be a JOIN
+            main_froms = set(_from_objects(update_stmt.table))
+            render_extra_froms = [
+                f for f in extra_froms if f not in main_froms
+            ]
+            correlate_froms = main_froms.union(extra_froms)
+        else:
+            render_extra_froms = []
+            correlate_froms = {update_stmt.table}
+
+        self.stack.append(
+            {
+                "correlate_froms": correlate_froms,
+                "asfrom_froms": correlate_froms,
+                "selectable": update_stmt,
+            }
+        )
+
+        text = "UPDATE "
+
+        if update_stmt._prefixes:
+            text += self._generate_prefixes(
+                update_stmt, update_stmt._prefixes, **kw
+            )
+
+        table_text = self.update_tables_clause(
+            update_stmt,
+            update_stmt.table,
+            render_extra_froms,
+            from_linter=from_linter,
+            **kw,
+        )
+        crud_params_struct = crud._get_crud_params(
+            self, update_stmt, compile_state, toplevel, **kw
+        )
+        crud_params = crud_params_struct.single_params
+
+        if update_stmt._hints:
+            dialect_hints, table_text = self._setup_crud_hints(
+                update_stmt, table_text
+            )
+        else:
+            dialect_hints = None
+
+        if update_stmt._independent_ctes:
+            self._dispatch_independent_ctes(update_stmt, kw)
+
+        text += table_text
+
+        text += " SET "
+        text += ", ".join(
+            expr + "=" + value
+            for _, expr, value, _ in cast(
+                "List[Tuple[Any, str, str, Any]]", crud_params
+            )
+        )
+
+        if self.implicit_returning or update_stmt._returning:
+            if self.returning_precedes_values:
+                text += " " + self.returning_clause(
+                    update_stmt,
+                    self.implicit_returning or update_stmt._returning,
+                    populate_result_map=toplevel,
+                )
+
+        if extra_froms:
+            extra_from_text = self.update_from_clause(
+                update_stmt,
+                update_stmt.table,
+                render_extra_froms,
+                dialect_hints,
+                from_linter=from_linter,
+                **kw,
+            )
+            if extra_from_text:
+                text += " " + extra_from_text
+
+        if update_stmt._where_criteria:
+            t = self._generate_delimited_and_list(
+                update_stmt._where_criteria, from_linter=from_linter, **kw
+            )
+            if t:
+                text += " WHERE " + t
+
+        limit_clause = self.update_limit_clause(update_stmt)
+        if limit_clause:
+            text += " " + limit_clause
+
+        if (
+            self.implicit_returning or update_stmt._returning
+        ) and not self.returning_precedes_values:
+            text += " " + self.returning_clause(
+                update_stmt,
+                self.implicit_returning or update_stmt._returning,
+                populate_result_map=toplevel,
+            )
+
+        if self.ctes:
+            nesting_level = len(self.stack) if not toplevel else None
+            text = self._render_cte_clause(nesting_level=nesting_level) + text
+
+        if warn_linting:
+            assert from_linter is not None
+            from_linter.warn(stmt_type="UPDATE")
+
+        self.stack.pop(-1)
+
+        return text
+
+    def delete_extra_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        """Provide a hook to override the generation of an
+        DELETE..FROM clause.
+
+        This can be used to implement DELETE..USING for example.
+
+        MySQL and MSSQL override this.
+
+        """
+        raise NotImplementedError(
+            "This backend does not support multiple-table "
+            "criteria within DELETE"
+        )
+
+    def delete_table_clause(self, delete_stmt, from_table, extra_froms, **kw):
+        return from_table._compiler_dispatch(
+            self, asfrom=True, iscrud=True, **kw
+        )
+
+    def visit_delete(self, delete_stmt, visiting_cte=None, **kw):
+        compile_state = delete_stmt._compile_state_factory(
+            delete_stmt, self, **kw
+        )
+        delete_stmt = compile_state.statement
+
+        if visiting_cte is not None:
+            kw["visiting_cte"] = visiting_cte
+            toplevel = False
+        else:
+            toplevel = not self.stack
+
+        if toplevel:
+            self.isdelete = True
+            if not self.dml_compile_state:
+                self.dml_compile_state = compile_state
+            if not self.compile_state:
+                self.compile_state = compile_state
+
+        if self.linting & COLLECT_CARTESIAN_PRODUCTS:
+            from_linter = FromLinter({}, set())
+            warn_linting = self.linting & WARN_LINTING
+            if toplevel:
+                self.from_linter = from_linter
+        else:
+            from_linter = None
+            warn_linting = False
+
+        extra_froms = compile_state._extra_froms
+
+        correlate_froms = {delete_stmt.table}.union(extra_froms)
+        self.stack.append(
+            {
+                "correlate_froms": correlate_froms,
+                "asfrom_froms": correlate_froms,
+                "selectable": delete_stmt,
+            }
+        )
+
+        text = "DELETE "
+
+        if delete_stmt._prefixes:
+            text += self._generate_prefixes(
+                delete_stmt, delete_stmt._prefixes, **kw
+            )
+
+        text += "FROM "
+
+        try:
+            table_text = self.delete_table_clause(
+                delete_stmt,
+                delete_stmt.table,
+                extra_froms,
+                from_linter=from_linter,
+            )
+        except TypeError:
+            # anticipate 3rd party dialects that don't include **kw
+            # TODO: remove in 2.1
+            table_text = self.delete_table_clause(
+                delete_stmt, delete_stmt.table, extra_froms
+            )
+            if from_linter:
+                _ = self.process(delete_stmt.table, from_linter=from_linter)
+
+        crud._get_crud_params(self, delete_stmt, compile_state, toplevel, **kw)
+
+        if delete_stmt._hints:
+            dialect_hints, table_text = self._setup_crud_hints(
+                delete_stmt, table_text
+            )
+        else:
+            dialect_hints = None
+
+        if delete_stmt._independent_ctes:
+            self._dispatch_independent_ctes(delete_stmt, kw)
+
+        text += table_text
+
+        if (
+            self.implicit_returning or delete_stmt._returning
+        ) and self.returning_precedes_values:
+            text += " " + self.returning_clause(
+                delete_stmt,
+                self.implicit_returning or delete_stmt._returning,
+                populate_result_map=toplevel,
+            )
+
+        if extra_froms:
+            extra_from_text = self.delete_extra_from_clause(
+                delete_stmt,
+                delete_stmt.table,
+                extra_froms,
+                dialect_hints,
+                from_linter=from_linter,
+                **kw,
+            )
+            if extra_from_text:
+                text += " " + extra_from_text
+
+        if delete_stmt._where_criteria:
+            t = self._generate_delimited_and_list(
+                delete_stmt._where_criteria, from_linter=from_linter, **kw
+            )
+            if t:
+                text += " WHERE " + t
+
+        limit_clause = self.delete_limit_clause(delete_stmt)
+        if limit_clause:
+            text += " " + limit_clause
+
+        if (
+            self.implicit_returning or delete_stmt._returning
+        ) and not self.returning_precedes_values:
+            text += " " + self.returning_clause(
+                delete_stmt,
+                self.implicit_returning or delete_stmt._returning,
+                populate_result_map=toplevel,
+            )
+
+        if self.ctes:
+            nesting_level = len(self.stack) if not toplevel else None
+            text = self._render_cte_clause(nesting_level=nesting_level) + text
+
+        if warn_linting:
+            assert from_linter is not None
+            from_linter.warn(stmt_type="DELETE")
+
+        self.stack.pop(-1)
+
+        return text
+
+    def visit_savepoint(self, savepoint_stmt, **kw):
+        return "SAVEPOINT %s" % self.preparer.format_savepoint(savepoint_stmt)
+
+    def visit_rollback_to_savepoint(self, savepoint_stmt, **kw):
+        return "ROLLBACK TO SAVEPOINT %s" % self.preparer.format_savepoint(
+            savepoint_stmt
+        )
+
+    def visit_release_savepoint(self, savepoint_stmt, **kw):
+        return "RELEASE SAVEPOINT %s" % self.preparer.format_savepoint(
+            savepoint_stmt
+        )
+
+
+class StrSQLCompiler(SQLCompiler):
+    """A :class:`.SQLCompiler` subclass which allows a small selection
+    of non-standard SQL features to render into a string value.
+
+    The :class:`.StrSQLCompiler` is invoked whenever a Core expression
+    element is directly stringified without calling upon the
+    :meth:`_expression.ClauseElement.compile` method.
+    It can render a limited set
+    of non-standard SQL constructs to assist in basic stringification,
+    however for more substantial custom or dialect-specific SQL constructs,
+    it will be necessary to make use of
+    :meth:`_expression.ClauseElement.compile`
+    directly.
+
+    .. seealso::
+
+        :ref:`faq_sql_expression_string`
+
+    """
+
+    def _fallback_column_name(self, column):
+        return "<name unknown>"
+
+    @util.preload_module("sqlalchemy.engine.url")
+    def visit_unsupported_compilation(self, element, err, **kw):
+        if element.stringify_dialect != "default":
+            url = util.preloaded.engine_url
+            dialect = url.URL.create(element.stringify_dialect).get_dialect()()
+
+            compiler = dialect.statement_compiler(
+                dialect, None, _supporting_against=self
+            )
+            if not isinstance(compiler, StrSQLCompiler):
+                return compiler.process(element, **kw)
+
+        return super().visit_unsupported_compilation(element, err)
+
+    def visit_getitem_binary(self, binary, operator, **kw):
+        return "%s[%s]" % (
+            self.process(binary.left, **kw),
+            self.process(binary.right, **kw),
+        )
+
+    def visit_json_getitem_op_binary(self, binary, operator, **kw):
+        return self.visit_getitem_binary(binary, operator, **kw)
+
+    def visit_json_path_getitem_op_binary(self, binary, operator, **kw):
+        return self.visit_getitem_binary(binary, operator, **kw)
+
+    def visit_sequence(self, sequence, **kw):
+        return (
+            f"<next sequence value: {self.preparer.format_sequence(sequence)}>"
+        )
+
+    def returning_clause(
+        self,
+        stmt: UpdateBase,
+        returning_cols: Sequence[ColumnElement[Any]],
+        *,
+        populate_result_map: bool,
+        **kw: Any,
+    ) -> str:
+        columns = [
+            self._label_select_column(None, c, True, False, {})
+            for c in base._select_iterables(returning_cols)
+        ]
+        return "RETURNING " + ", ".join(columns)
+
+    def update_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        kw["asfrom"] = True
+        return "FROM " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in extra_froms
+        )
+
+    def delete_extra_from_clause(
+        self, update_stmt, from_table, extra_froms, from_hints, **kw
+    ):
+        kw["asfrom"] = True
+        return ", " + ", ".join(
+            t._compiler_dispatch(self, fromhints=from_hints, **kw)
+            for t in extra_froms
+        )
+
+    def visit_empty_set_expr(self, element_types, **kw):
+        return "SELECT 1 WHERE 1!=1"
+
+    def get_from_hint_text(self, table, text):
+        return "[%s]" % text
+
+    def visit_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._generate_generic_binary(binary, " <regexp> ", **kw)
+
+    def visit_not_regexp_match_op_binary(self, binary, operator, **kw):
+        return self._generate_generic_binary(binary, " <not regexp> ", **kw)
+
+    def visit_regexp_replace_op_binary(self, binary, operator, **kw):
+        return "<regexp replace>(%s, %s)" % (
+            binary.left._compiler_dispatch(self, **kw),
+            binary.right._compiler_dispatch(self, **kw),
+        )
+
+    def visit_try_cast(self, cast, **kwargs):
+        return "TRY_CAST(%s AS %s)" % (
+            cast.clause._compiler_dispatch(self, **kwargs),
+            cast.typeclause._compiler_dispatch(self, **kwargs),
+        )
+
+
+class DDLCompiler(Compiled):
+    is_ddl = True
+
+    if TYPE_CHECKING:
+
+        def __init__(
+            self,
+            dialect: Dialect,
+            statement: ExecutableDDLElement,
+            schema_translate_map: Optional[SchemaTranslateMapType] = ...,
+            render_schema_translate: bool = ...,
+            compile_kwargs: Mapping[str, Any] = ...,
+        ): ...
+
+    @util.memoized_property
+    def sql_compiler(self):
+        return self.dialect.statement_compiler(
+            self.dialect, None, schema_translate_map=self.schema_translate_map
+        )
+
+    @util.memoized_property
+    def type_compiler(self):
+        return self.dialect.type_compiler_instance
+
+    def construct_params(
+        self,
+        params: Optional[_CoreSingleExecuteParams] = None,
+        extracted_parameters: Optional[Sequence[BindParameter[Any]]] = None,
+        escape_names: bool = True,
+    ) -> Optional[_MutableCoreSingleExecuteParams]:
+        return None
+
+    def visit_ddl(self, ddl, **kwargs):
+        # table events can substitute table and schema name
+        context = ddl.context
+        if isinstance(ddl.target, schema.Table):
+            context = context.copy()
+
+            preparer = self.preparer
+            path = preparer.format_table_seq(ddl.target)
+            if len(path) == 1:
+                table, sch = path[0], ""
+            else:
+                table, sch = path[-1], path[0]
+
+            context.setdefault("table", table)
+            context.setdefault("schema", sch)
+            context.setdefault("fullname", preparer.format_table(ddl.target))
+
+        return self.sql_compiler.post_process_text(ddl.statement % context)
+
+    def visit_create_schema(self, create, **kw):
+        text = "CREATE SCHEMA "
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+        return text + self.preparer.format_schema(create.element)
+
+    def visit_drop_schema(self, drop, **kw):
+        text = "DROP SCHEMA "
+        if drop.if_exists:
+            text += "IF EXISTS "
+        text += self.preparer.format_schema(drop.element)
+        if drop.cascade:
+            text += " CASCADE"
+        return text
+
+    def visit_create_table(self, create, **kw):
+        table = create.element
+        preparer = self.preparer
+
+        text = "\nCREATE "
+        if table._prefixes:
+            text += " ".join(table._prefixes) + " "
+
+        text += "TABLE "
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+
+        text += preparer.format_table(table) + " "
+
+        create_table_suffix = self.create_table_suffix(table)
+        if create_table_suffix:
+            text += create_table_suffix + " "
+
+        text += "("
+
+        separator = "\n"
+
+        # if only one primary key, specify it along with the column
+        first_pk = False
+        for create_column in create.columns:
+            column = create_column.element
+            try:
+                processed = self.process(
+                    create_column, first_pk=column.primary_key and not first_pk
+                )
+                if processed is not None:
+                    text += separator
+                    separator = ", \n"
+                    text += "\t" + processed
+                if column.primary_key:
+                    first_pk = True
+            except exc.CompileError as ce:
+                raise exc.CompileError(
+                    "(in table '%s', column '%s'): %s"
+                    % (table.description, column.name, ce.args[0])
+                ) from ce
+
+        const = self.create_table_constraints(
+            table,
+            _include_foreign_key_constraints=create.include_foreign_key_constraints,  # noqa
+        )
+        if const:
+            text += separator + "\t" + const
+
+        text += "\n)%s\n\n" % self.post_create_table(table)
+        return text
+
+    def visit_create_column(self, create, first_pk=False, **kw):
+        column = create.element
+
+        if column.system:
+            return None
+
+        text = self.get_column_specification(column, first_pk=first_pk)
+        const = " ".join(
+            self.process(constraint) for constraint in column.constraints
+        )
+        if const:
+            text += " " + const
+
+        return text
+
+    def create_table_constraints(
+        self, table, _include_foreign_key_constraints=None, **kw
+    ):
+        # On some DB order is significant: visit PK first, then the
+        # other constraints (engine.ReflectionTest.testbasic failed on FB2)
+        constraints = []
+        if table.primary_key:
+            constraints.append(table.primary_key)
+
+        all_fkcs = table.foreign_key_constraints
+        if _include_foreign_key_constraints is not None:
+            omit_fkcs = all_fkcs.difference(_include_foreign_key_constraints)
+        else:
+            omit_fkcs = set()
+
+        constraints.extend(
+            [
+                c
+                for c in table._sorted_constraints
+                if c is not table.primary_key and c not in omit_fkcs
+            ]
+        )
+
+        return ", \n\t".join(
+            p
+            for p in (
+                self.process(constraint)
+                for constraint in constraints
+                if (constraint._should_create_for_compiler(self))
+                and (
+                    not self.dialect.supports_alter
+                    or not getattr(constraint, "use_alter", False)
+                )
+            )
+            if p is not None
+        )
+
+    def visit_drop_table(self, drop, **kw):
+        text = "\nDROP TABLE "
+        if drop.if_exists:
+            text += "IF EXISTS "
+        return text + self.preparer.format_table(drop.element)
+
+    def visit_drop_view(self, drop, **kw):
+        return "\nDROP VIEW " + self.preparer.format_table(drop.element)
+
+    def _verify_index_table(self, index):
+        if index.table is None:
+            raise exc.CompileError(
+                "Index '%s' is not associated with any table." % index.name
+            )
+
+    def visit_create_index(
+        self, create, include_schema=False, include_table_schema=True, **kw
+    ):
+        index = create.element
+        self._verify_index_table(index)
+        preparer = self.preparer
+        text = "CREATE "
+        if index.unique:
+            text += "UNIQUE "
+        if index.name is None:
+            raise exc.CompileError(
+                "CREATE INDEX requires that the index have a name"
+            )
+
+        text += "INDEX "
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+
+        text += "%s ON %s (%s)" % (
+            self._prepared_index_name(index, include_schema=include_schema),
+            preparer.format_table(
+                index.table, use_schema=include_table_schema
+            ),
+            ", ".join(
+                self.sql_compiler.process(
+                    expr, include_table=False, literal_binds=True
+                )
+                for expr in index.expressions
+            ),
+        )
+        return text
+
+    def visit_drop_index(self, drop, **kw):
+        index = drop.element
+
+        if index.name is None:
+            raise exc.CompileError(
+                "DROP INDEX requires that the index have a name"
+            )
+        text = "\nDROP INDEX "
+        if drop.if_exists:
+            text += "IF EXISTS "
+
+        return text + self._prepared_index_name(index, include_schema=True)
+
+    def _prepared_index_name(self, index, include_schema=False):
+        if index.table is not None:
+            effective_schema = self.preparer.schema_for_object(index.table)
+        else:
+            effective_schema = None
+        if include_schema and effective_schema:
+            schema_name = self.preparer.quote_schema(effective_schema)
+        else:
+            schema_name = None
+
+        index_name = self.preparer.format_index(index)
+
+        if schema_name:
+            index_name = schema_name + "." + index_name
+        return index_name
+
+    def visit_add_constraint(self, create, **kw):
+        return "ALTER TABLE %s ADD %s" % (
+            self.preparer.format_table(create.element.table),
+            self.process(create.element),
+        )
+
+    def visit_set_table_comment(self, create, **kw):
+        return "COMMENT ON TABLE %s IS %s" % (
+            self.preparer.format_table(create.element),
+            self.sql_compiler.render_literal_value(
+                create.element.comment, sqltypes.String()
+            ),
+        )
+
+    def visit_drop_table_comment(self, drop, **kw):
+        return "COMMENT ON TABLE %s IS NULL" % self.preparer.format_table(
+            drop.element
+        )
+
+    def visit_set_column_comment(self, create, **kw):
+        return "COMMENT ON COLUMN %s IS %s" % (
+            self.preparer.format_column(
+                create.element, use_table=True, use_schema=True
+            ),
+            self.sql_compiler.render_literal_value(
+                create.element.comment, sqltypes.String()
+            ),
+        )
+
+    def visit_drop_column_comment(self, drop, **kw):
+        return "COMMENT ON COLUMN %s IS NULL" % self.preparer.format_column(
+            drop.element, use_table=True
+        )
+
+    def visit_set_constraint_comment(self, create, **kw):
+        raise exc.UnsupportedCompilationError(self, type(create))
+
+    def visit_drop_constraint_comment(self, drop, **kw):
+        raise exc.UnsupportedCompilationError(self, type(drop))
+
+    def get_identity_options(self, identity_options):
+        text = []
+        if identity_options.increment is not None:
+            text.append("INCREMENT BY %d" % identity_options.increment)
+        if identity_options.start is not None:
+            text.append("START WITH %d" % identity_options.start)
+        if identity_options.minvalue is not None:
+            text.append("MINVALUE %d" % identity_options.minvalue)
+        if identity_options.maxvalue is not None:
+            text.append("MAXVALUE %d" % identity_options.maxvalue)
+        if identity_options.nominvalue is not None:
+            text.append("NO MINVALUE")
+        if identity_options.nomaxvalue is not None:
+            text.append("NO MAXVALUE")
+        if identity_options.cache is not None:
+            text.append("CACHE %d" % identity_options.cache)
+        if identity_options.cycle is not None:
+            text.append("CYCLE" if identity_options.cycle else "NO CYCLE")
+        return " ".join(text)
+
+    def visit_create_sequence(self, create, prefix=None, **kw):
+        text = "CREATE SEQUENCE "
+        if create.if_not_exists:
+            text += "IF NOT EXISTS "
+        text += self.preparer.format_sequence(create.element)
+
+        if prefix:
+            text += prefix
+        options = self.get_identity_options(create.element)
+        if options:
+            text += " " + options
+        return text
+
+    def visit_drop_sequence(self, drop, **kw):
+        text = "DROP SEQUENCE "
+        if drop.if_exists:
+            text += "IF EXISTS "
+        return text + self.preparer.format_sequence(drop.element)
+
+    def visit_drop_constraint(self, drop, **kw):
+        constraint = drop.element
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+        else:
+            formatted_name = None
+
+        if formatted_name is None:
+            raise exc.CompileError(
+                "Can't emit DROP CONSTRAINT for constraint %r; "
+                "it has no name" % drop.element
+            )
+        return "ALTER TABLE %s DROP CONSTRAINT %s%s%s" % (
+            self.preparer.format_table(drop.element.table),
+            "IF EXISTS " if drop.if_exists else "",
+            formatted_name,
+            " CASCADE" if drop.cascade else "",
+        )
+
+    def get_column_specification(self, column, **kwargs):
+        colspec = (
+            self.preparer.format_column(column)
+            + " "
+            + self.dialect.type_compiler_instance.process(
+                column.type, type_expression=column
+            )
+        )
+        default = self.get_column_default_string(column)
+        if default is not None:
+            colspec += " DEFAULT " + default
+
+        if column.computed is not None:
+            colspec += " " + self.process(column.computed)
+
+        if (
+            column.identity is not None
+            and self.dialect.supports_identity_columns
+        ):
+            colspec += " " + self.process(column.identity)
+
+        if not column.nullable and (
+            not column.identity or not self.dialect.supports_identity_columns
+        ):
+            colspec += " NOT NULL"
+        return colspec
+
+    def create_table_suffix(self, table):
+        return ""
+
+    def post_create_table(self, table):
+        return ""
+
+    def get_column_default_string(self, column):
+        if isinstance(column.server_default, schema.DefaultClause):
+            return self.render_default_string(column.server_default.arg)
+        else:
+            return None
+
+    def render_default_string(self, default):
+        if isinstance(default, str):
+            return self.sql_compiler.render_literal_value(
+                default, sqltypes.STRINGTYPE
+            )
+        else:
+            return self.sql_compiler.process(default, literal_binds=True)
+
+    def visit_table_or_column_check_constraint(self, constraint, **kw):
+        if constraint.is_column_level:
+            return self.visit_column_check_constraint(constraint)
+        else:
+            return self.visit_check_constraint(constraint)
+
+    def visit_check_constraint(self, constraint, **kw):
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        text += "CHECK (%s)" % self.sql_compiler.process(
+            constraint.sqltext, include_table=False, literal_binds=True
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def visit_column_check_constraint(self, constraint, **kw):
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        text += "CHECK (%s)" % self.sql_compiler.process(
+            constraint.sqltext, include_table=False, literal_binds=True
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def visit_primary_key_constraint(self, constraint, **kw):
+        if len(constraint) == 0:
+            return ""
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        text += "PRIMARY KEY "
+        text += "(%s)" % ", ".join(
+            self.preparer.quote(c.name)
+            for c in (
+                constraint.columns_autoinc_first
+                if constraint._implicit_generated
+                else constraint.columns
+            )
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def visit_foreign_key_constraint(self, constraint, **kw):
+        preparer = self.preparer
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        remote_table = list(constraint.elements)[0].column.table
+        text += "FOREIGN KEY(%s) REFERENCES %s (%s)" % (
+            ", ".join(
+                preparer.quote(f.parent.name) for f in constraint.elements
+            ),
+            self.define_constraint_remote_table(
+                constraint, remote_table, preparer
+            ),
+            ", ".join(
+                preparer.quote(f.column.name) for f in constraint.elements
+            ),
+        )
+        text += self.define_constraint_match(constraint)
+        text += self.define_constraint_cascades(constraint)
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def define_constraint_remote_table(self, constraint, table, preparer):
+        """Format the remote table clause of a CREATE CONSTRAINT clause."""
+
+        return preparer.format_table(table)
+
+    def visit_unique_constraint(self, constraint, **kw):
+        if len(constraint) == 0:
+            return ""
+        text = ""
+        if constraint.name is not None:
+            formatted_name = self.preparer.format_constraint(constraint)
+            if formatted_name is not None:
+                text += "CONSTRAINT %s " % formatted_name
+        text += "UNIQUE %s(%s)" % (
+            self.define_unique_constraint_distinct(constraint, **kw),
+            ", ".join(self.preparer.quote(c.name) for c in constraint),
+        )
+        text += self.define_constraint_deferrability(constraint)
+        return text
+
+    def define_unique_constraint_distinct(self, constraint, **kw):
+        return ""
+
+    def define_constraint_cascades(self, constraint):
+        text = ""
+        if constraint.ondelete is not None:
+            text += " ON DELETE %s" % self.preparer.validate_sql_phrase(
+                constraint.ondelete, FK_ON_DELETE
+            )
+        if constraint.onupdate is not None:
+            text += " ON UPDATE %s" % self.preparer.validate_sql_phrase(
+                constraint.onupdate, FK_ON_UPDATE
+            )
+        return text
+
+    def define_constraint_deferrability(self, constraint):
+        text = ""
+        if constraint.deferrable is not None:
+            if constraint.deferrable:
+                text += " DEFERRABLE"
+            else:
+                text += " NOT DEFERRABLE"
+        if constraint.initially is not None:
+            text += " INITIALLY %s" % self.preparer.validate_sql_phrase(
+                constraint.initially, FK_INITIALLY
+            )
+        return text
+
+    def define_constraint_match(self, constraint):
+        text = ""
+        if constraint.match is not None:
+            text += " MATCH %s" % constraint.match
+        return text
+
+    def visit_computed_column(self, generated, **kw):
+        text = "GENERATED ALWAYS AS (%s)" % self.sql_compiler.process(
+            generated.sqltext, include_table=False, literal_binds=True
+        )
+        if generated.persisted is True:
+            text += " STORED"
+        elif generated.persisted is False:
+            text += " VIRTUAL"
+        return text
+
+    def visit_identity_column(self, identity, **kw):
+        text = "GENERATED %s AS IDENTITY" % (
+            "ALWAYS" if identity.always else "BY DEFAULT",
+        )
+        options = self.get_identity_options(identity)
+        if options:
+            text += " (%s)" % options
+        return text
+
+
+class GenericTypeCompiler(TypeCompiler):
+    def visit_FLOAT(self, type_, **kw):
+        return "FLOAT"
+
+    def visit_DOUBLE(self, type_, **kw):
+        return "DOUBLE"
+
+    def visit_DOUBLE_PRECISION(self, type_, **kw):
+        return "DOUBLE PRECISION"
+
+    def visit_REAL(self, type_, **kw):
+        return "REAL"
+
+    def visit_NUMERIC(self, type_, **kw):
+        if type_.precision is None:
+            return "NUMERIC"
+        elif type_.scale is None:
+            return "NUMERIC(%(precision)s)" % {"precision": type_.precision}
+        else:
+            return "NUMERIC(%(precision)s, %(scale)s)" % {
+                "precision": type_.precision,
+                "scale": type_.scale,
+            }
+
+    def visit_DECIMAL(self, type_, **kw):
+        if type_.precision is None:
+            return "DECIMAL"
+        elif type_.scale is None:
+            return "DECIMAL(%(precision)s)" % {"precision": type_.precision}
+        else:
+            return "DECIMAL(%(precision)s, %(scale)s)" % {
+                "precision": type_.precision,
+                "scale": type_.scale,
+            }
+
+    def visit_INTEGER(self, type_, **kw):
+        return "INTEGER"
+
+    def visit_SMALLINT(self, type_, **kw):
+        return "SMALLINT"
+
+    def visit_BIGINT(self, type_, **kw):
+        return "BIGINT"
+
+    def visit_TIMESTAMP(self, type_, **kw):
+        return "TIMESTAMP"
+
+    def visit_DATETIME(self, type_, **kw):
+        return "DATETIME"
+
+    def visit_DATE(self, type_, **kw):
+        return "DATE"
+
+    def visit_TIME(self, type_, **kw):
+        return "TIME"
+
+    def visit_CLOB(self, type_, **kw):
+        return "CLOB"
+
+    def visit_NCLOB(self, type_, **kw):
+        return "NCLOB"
+
+    def _render_string_type(self, type_, name, length_override=None):
+        text = name
+        if length_override:
+            text += "(%d)" % length_override
+        elif type_.length:
+            text += "(%d)" % type_.length
+        if type_.collation:
+            text += ' COLLATE "%s"' % type_.collation
+        return text
+
+    def visit_CHAR(self, type_, **kw):
+        return self._render_string_type(type_, "CHAR")
+
+    def visit_NCHAR(self, type_, **kw):
+        return self._render_string_type(type_, "NCHAR")
+
+    def visit_VARCHAR(self, type_, **kw):
+        return self._render_string_type(type_, "VARCHAR")
+
+    def visit_NVARCHAR(self, type_, **kw):
+        return self._render_string_type(type_, "NVARCHAR")
+
+    def visit_TEXT(self, type_, **kw):
+        return self._render_string_type(type_, "TEXT")
+
+    def visit_UUID(self, type_, **kw):
+        return "UUID"
+
+    def visit_BLOB(self, type_, **kw):
+        return "BLOB"
+
+    def visit_BINARY(self, type_, **kw):
+        return "BINARY" + (type_.length and "(%d)" % type_.length or "")
+
+    def visit_VARBINARY(self, type_, **kw):
+        return "VARBINARY" + (type_.length and "(%d)" % type_.length or "")
+
+    def visit_BOOLEAN(self, type_, **kw):
+        return "BOOLEAN"
+
+    def visit_uuid(self, type_, **kw):
+        if not type_.native_uuid or not self.dialect.supports_native_uuid:
+            return self._render_string_type(type_, "CHAR", length_override=32)
+        else:
+            return self.visit_UUID(type_, **kw)
+
+    def visit_large_binary(self, type_, **kw):
+        return self.visit_BLOB(type_, **kw)
+
+    def visit_boolean(self, type_, **kw):
+        return self.visit_BOOLEAN(type_, **kw)
+
+    def visit_time(self, type_, **kw):
+        return self.visit_TIME(type_, **kw)
+
+    def visit_datetime(self, type_, **kw):
+        return self.visit_DATETIME(type_, **kw)
+
+    def visit_date(self, type_, **kw):
+        return self.visit_DATE(type_, **kw)
+
+    def visit_big_integer(self, type_, **kw):
+        return self.visit_BIGINT(type_, **kw)
+
+    def visit_small_integer(self, type_, **kw):
+        return self.visit_SMALLINT(type_, **kw)
+
+    def visit_integer(self, type_, **kw):
+        return self.visit_INTEGER(type_, **kw)
+
+    def visit_real(self, type_, **kw):
+        return self.visit_REAL(type_, **kw)
+
+    def visit_float(self, type_, **kw):
+        return self.visit_FLOAT(type_, **kw)
+
+    def visit_double(self, type_, **kw):
+        return self.visit_DOUBLE(type_, **kw)
+
+    def visit_numeric(self, type_, **kw):
+        return self.visit_NUMERIC(type_, **kw)
+
+    def visit_string(self, type_, **kw):
+        return self.visit_VARCHAR(type_, **kw)
+
+    def visit_unicode(self, type_, **kw):
+        return self.visit_VARCHAR(type_, **kw)
+
+    def visit_text(self, type_, **kw):
+        return self.visit_TEXT(type_, **kw)
+
+    def visit_unicode_text(self, type_, **kw):
+        return self.visit_TEXT(type_, **kw)
+
+    def visit_enum(self, type_, **kw):
+        return self.visit_VARCHAR(type_, **kw)
+
+    def visit_null(self, type_, **kw):
+        raise exc.CompileError(
+            "Can't generate DDL for %r; "
+            "did you forget to specify a "
+            "type on this Column?" % type_
+        )
+
+    def visit_type_decorator(self, type_, **kw):
+        return self.process(type_.type_engine(self.dialect), **kw)
+
+    def visit_user_defined(self, type_, **kw):
+        return type_.get_col_spec(**kw)
+
+
+class StrSQLTypeCompiler(GenericTypeCompiler):
+    def process(self, type_, **kw):
+        try:
+            _compiler_dispatch = type_._compiler_dispatch
+        except AttributeError:
+            return self._visit_unknown(type_, **kw)
+        else:
+            return _compiler_dispatch(self, **kw)
+
+    def __getattr__(self, key):
+        if key.startswith("visit_"):
+            return self._visit_unknown
+        else:
+            raise AttributeError(key)
+
+    def _visit_unknown(self, type_, **kw):
+        if type_.__class__.__name__ == type_.__class__.__name__.upper():
+            return type_.__class__.__name__
+        else:
+            return repr(type_)
+
+    def visit_null(self, type_, **kw):
+        return "NULL"
+
+    def visit_user_defined(self, type_, **kw):
+        try:
+            get_col_spec = type_.get_col_spec
+        except AttributeError:
+            return repr(type_)
+        else:
+            return get_col_spec(**kw)
+
+
+class _SchemaForObjectCallable(Protocol):
+    def __call__(self, __obj: Any) -> str: ...
+
+
+class _BindNameForColProtocol(Protocol):
+    def __call__(self, col: ColumnClause[Any]) -> str: ...
+
+
+class IdentifierPreparer:
+    """Handle quoting and case-folding of identifiers based on options."""
+
+    reserved_words = RESERVED_WORDS
+
+    legal_characters = LEGAL_CHARACTERS
+
+    illegal_initial_characters = ILLEGAL_INITIAL_CHARACTERS
+
+    initial_quote: str
+
+    final_quote: str
+
+    _strings: MutableMapping[str, str]
+
+    schema_for_object: _SchemaForObjectCallable = operator.attrgetter("schema")
+    """Return the .schema attribute for an object.
+
+    For the default IdentifierPreparer, the schema for an object is always
+    the value of the ".schema" attribute.   if the preparer is replaced
+    with one that has a non-empty schema_translate_map, the value of the
+    ".schema" attribute is rendered a symbol that will be converted to a
+    real schema name from the mapping post-compile.
+
+    """
+
+    _includes_none_schema_translate: bool = False
+
+    def __init__(
+        self,
+        dialect,
+        initial_quote='"',
+        final_quote=None,
+        escape_quote='"',
+        quote_case_sensitive_collations=True,
+        omit_schema=False,
+    ):
+        """Construct a new ``IdentifierPreparer`` object.
+
+        initial_quote
+          Character that begins a delimited identifier.
+
+        final_quote
+          Character that ends a delimited identifier. Defaults to
+          `initial_quote`.
+
+        omit_schema
+          Prevent prepending schema name. Useful for databases that do
+          not support schemae.
+        """
+
+        self.dialect = dialect
+        self.initial_quote = initial_quote
+        self.final_quote = final_quote or self.initial_quote
+        self.escape_quote = escape_quote
+        self.escape_to_quote = self.escape_quote * 2
+        self.omit_schema = omit_schema
+        self.quote_case_sensitive_collations = quote_case_sensitive_collations
+        self._strings = {}
+        self._double_percents = self.dialect.paramstyle in (
+            "format",
+            "pyformat",
+        )
+
+    def _with_schema_translate(self, schema_translate_map):
+        prep = self.__class__.__new__(self.__class__)
+        prep.__dict__.update(self.__dict__)
+
+        includes_none = None in schema_translate_map
+
+        def symbol_getter(obj):
+            name = obj.schema
+            if obj._use_schema_map and (name is not None or includes_none):
+                if name is not None and ("[" in name or "]" in name):
+                    raise exc.CompileError(
+                        "Square bracket characters ([]) not supported "
+                        "in schema translate name '%s'" % name
+                    )
+                return quoted_name(
+                    "__[SCHEMA_%s]" % (name or "_none"), quote=False
+                )
+            else:
+                return obj.schema
+
+        prep.schema_for_object = symbol_getter
+        prep._includes_none_schema_translate = includes_none
+        return prep
+
+    def _render_schema_translates(self, statement, schema_translate_map):
+        d = schema_translate_map
+        if None in d:
+            if not self._includes_none_schema_translate:
+                raise exc.InvalidRequestError(
+                    "schema translate map which previously did not have "
+                    "`None` present as a key now has `None` present; compiled "
+                    "statement may lack adequate placeholders.  Please use "
+                    "consistent keys in successive "
+                    "schema_translate_map dictionaries."
+                )
+
+            d["_none"] = d[None]
+
+        def replace(m):
+            name = m.group(2)
+            if name in d:
+                effective_schema = d[name]
+            else:
+                if name in (None, "_none"):
+                    raise exc.InvalidRequestError(
+                        "schema translate map which previously had `None` "
+                        "present as a key now no longer has it present; don't "
+                        "know how to apply schema for compiled statement. "
+                        "Please use consistent keys in successive "
+                        "schema_translate_map dictionaries."
+                    )
+                effective_schema = name
+
+            if not effective_schema:
+                effective_schema = self.dialect.default_schema_name
+                if not effective_schema:
+                    # TODO: no coverage here
+                    raise exc.CompileError(
+                        "Dialect has no default schema name; can't "
+                        "use None as dynamic schema target."
+                    )
+            return self.quote_schema(effective_schema)
+
+        return re.sub(r"(__\[SCHEMA_([^\]]+)\])", replace, statement)
+
+    def _escape_identifier(self, value: str) -> str:
+        """Escape an identifier.
+
+        Subclasses should override this to provide database-dependent
+        escaping behavior.
+        """
+
+        value = value.replace(self.escape_quote, self.escape_to_quote)
+        if self._double_percents:
+            value = value.replace("%", "%%")
+        return value
+
+    def _unescape_identifier(self, value: str) -> str:
+        """Canonicalize an escaped identifier.
+
+        Subclasses should override this to provide database-dependent
+        unescaping behavior that reverses _escape_identifier.
+        """
+
+        return value.replace(self.escape_to_quote, self.escape_quote)
+
+    def validate_sql_phrase(self, element, reg):
+        """keyword sequence filter.
+
+        a filter for elements that are intended to represent keyword sequences,
+        such as "INITIALLY", "INITIALLY DEFERRED", etc.   no special characters
+        should be present.
+
+        .. versionadded:: 1.3
+
+        """
+
+        if element is not None and not reg.match(element):
+            raise exc.CompileError(
+                "Unexpected SQL phrase: %r (matching against %r)"
+                % (element, reg.pattern)
+            )
+        return element
+
+    def quote_identifier(self, value: str) -> str:
+        """Quote an identifier.
+
+        Subclasses should override this to provide database-dependent
+        quoting behavior.
+        """
+
+        return (
+            self.initial_quote
+            + self._escape_identifier(value)
+            + self.final_quote
+        )
+
+    def _requires_quotes(self, value: str) -> bool:
+        """Return True if the given identifier requires quoting."""
+        lc_value = value.lower()
+        return (
+            lc_value in self.reserved_words
+            or value[0] in self.illegal_initial_characters
+            or not self.legal_characters.match(str(value))
+            or (lc_value != value)
+        )
+
+    def _requires_quotes_illegal_chars(self, value):
+        """Return True if the given identifier requires quoting, but
+        not taking case convention into account."""
+        return not self.legal_characters.match(str(value))
+
+    def quote_schema(self, schema: str, force: Any = None) -> str:
+        """Conditionally quote a schema name.
+
+
+        The name is quoted if it is a reserved word, contains quote-necessary
+        characters, or is an instance of :class:`.quoted_name` which includes
+        ``quote`` set to ``True``.
+
+        Subclasses can override this to provide database-dependent
+        quoting behavior for schema names.
+
+        :param schema: string schema name
+        :param force: unused
+
+            .. deprecated:: 0.9
+
+                The :paramref:`.IdentifierPreparer.quote_schema.force`
+                parameter is deprecated and will be removed in a future
+                release.  This flag has no effect on the behavior of the
+                :meth:`.IdentifierPreparer.quote` method; please refer to
+                :class:`.quoted_name`.
+
+        """
+        if force is not None:
+            # not using the util.deprecated_params() decorator in this
+            # case because of the additional function call overhead on this
+            # very performance-critical spot.
+            util.warn_deprecated(
+                "The IdentifierPreparer.quote_schema.force parameter is "
+                "deprecated and will be removed in a future release.  This "
+                "flag has no effect on the behavior of the "
+                "IdentifierPreparer.quote method; please refer to "
+                "quoted_name().",
+                # deprecated 0.9. warning from 1.3
+                version="0.9",
+            )
+
+        return self.quote(schema)
+
+    def quote(self, ident: str, force: Any = None) -> str:
+        """Conditionally quote an identifier.
+
+        The identifier is quoted if it is a reserved word, contains
+        quote-necessary characters, or is an instance of
+        :class:`.quoted_name` which includes ``quote`` set to ``True``.
+
+        Subclasses can override this to provide database-dependent
+        quoting behavior for identifier names.
+
+        :param ident: string identifier
+        :param force: unused
+
+            .. deprecated:: 0.9
+
+                The :paramref:`.IdentifierPreparer.quote.force`
+                parameter is deprecated and will be removed in a future
+                release.  This flag has no effect on the behavior of the
+                :meth:`.IdentifierPreparer.quote` method; please refer to
+                :class:`.quoted_name`.
+
+        """
+        if force is not None:
+            # not using the util.deprecated_params() decorator in this
+            # case because of the additional function call overhead on this
+            # very performance-critical spot.
+            util.warn_deprecated(
+                "The IdentifierPreparer.quote.force parameter is "
+                "deprecated and will be removed in a future release.  This "
+                "flag has no effect on the behavior of the "
+                "IdentifierPreparer.quote method; please refer to "
+                "quoted_name().",
+                # deprecated 0.9. warning from 1.3
+                version="0.9",
+            )
+
+        force = getattr(ident, "quote", None)
+
+        if force is None:
+            if ident in self._strings:
+                return self._strings[ident]
+            else:
+                if self._requires_quotes(ident):
+                    self._strings[ident] = self.quote_identifier(ident)
+                else:
+                    self._strings[ident] = ident
+                return self._strings[ident]
+        elif force:
+            return self.quote_identifier(ident)
+        else:
+            return ident
+
+    def format_collation(self, collation_name):
+        if self.quote_case_sensitive_collations:
+            return self.quote(collation_name)
+        else:
+            return collation_name
+
+    def format_sequence(self, sequence, use_schema=True):
+        name = self.quote(sequence.name)
+
+        effective_schema = self.schema_for_object(sequence)
+
+        if (
+            not self.omit_schema
+            and use_schema
+            and effective_schema is not None
+        ):
+            name = self.quote_schema(effective_schema) + "." + name
+        return name
+
+    def format_label(
+        self, label: Label[Any], name: Optional[str] = None
+    ) -> str:
+        return self.quote(name or label.name)
+
+    def format_alias(
+        self, alias: Optional[AliasedReturnsRows], name: Optional[str] = None
+    ) -> str:
+        if name is None:
+            assert alias is not None
+            return self.quote(alias.name)
+        else:
+            return self.quote(name)
+
+    def format_savepoint(self, savepoint, name=None):
+        # Running the savepoint name through quoting is unnecessary
+        # for all known dialects.  This is here to support potential
+        # third party use cases
+        ident = name or savepoint.ident
+        if self._requires_quotes(ident):
+            ident = self.quote_identifier(ident)
+        return ident
+
+    @util.preload_module("sqlalchemy.sql.naming")
+    def format_constraint(self, constraint, _alembic_quote=True):
+        naming = util.preloaded.sql_naming
+
+        if constraint.name is _NONE_NAME:
+            name = naming._constraint_name_for_table(
+                constraint, constraint.table
+            )
+
+            if name is None:
+                return None
+        else:
+            name = constraint.name
+
+        if constraint.__visit_name__ == "index":
+            return self.truncate_and_render_index_name(
+                name, _alembic_quote=_alembic_quote
+            )
+        else:
+            return self.truncate_and_render_constraint_name(
+                name, _alembic_quote=_alembic_quote
+            )
+
+    def truncate_and_render_index_name(self, name, _alembic_quote=True):
+        # calculate these at format time so that ad-hoc changes
+        # to dialect.max_identifier_length etc. can be reflected
+        # as IdentifierPreparer is long lived
+        max_ = (
+            self.dialect.max_index_name_length
+            or self.dialect.max_identifier_length
+        )
+        return self._truncate_and_render_maxlen_name(
+            name, max_, _alembic_quote
+        )
+
+    def truncate_and_render_constraint_name(self, name, _alembic_quote=True):
+        # calculate these at format time so that ad-hoc changes
+        # to dialect.max_identifier_length etc. can be reflected
+        # as IdentifierPreparer is long lived
+        max_ = (
+            self.dialect.max_constraint_name_length
+            or self.dialect.max_identifier_length
+        )
+        return self._truncate_and_render_maxlen_name(
+            name, max_, _alembic_quote
+        )
+
+    def _truncate_and_render_maxlen_name(self, name, max_, _alembic_quote):
+        if isinstance(name, elements._truncated_label):
+            if len(name) > max_:
+                name = name[0 : max_ - 8] + "_" + util.md5_hex(name)[-4:]
+        else:
+            self.dialect.validate_identifier(name)
+
+        if not _alembic_quote:
+            return name
+        else:
+            return self.quote(name)
+
+    def format_index(self, index):
+        return self.format_constraint(index)
+
+    def format_table(self, table, use_schema=True, name=None):
+        """Prepare a quoted table and schema name."""
+
+        if name is None:
+            name = table.name
+
+        result = self.quote(name)
+
+        effective_schema = self.schema_for_object(table)
+
+        if not self.omit_schema and use_schema and effective_schema:
+            result = self.quote_schema(effective_schema) + "." + result
+        return result
+
+    def format_schema(self, name):
+        """Prepare a quoted schema name."""
+
+        return self.quote(name)
+
+    def format_label_name(
+        self,
+        name,
+        anon_map=None,
+    ):
+        """Prepare a quoted column name."""
+
+        if anon_map is not None and isinstance(
+            name, elements._truncated_label
+        ):
+            name = name.apply_map(anon_map)
+
+        return self.quote(name)
+
+    def format_column(
+        self,
+        column,
+        use_table=False,
+        name=None,
+        table_name=None,
+        use_schema=False,
+        anon_map=None,
+    ):
+        """Prepare a quoted column name."""
+
+        if name is None:
+            name = column.name
+
+        if anon_map is not None and isinstance(
+            name, elements._truncated_label
+        ):
+            name = name.apply_map(anon_map)
+
+        if not getattr(column, "is_literal", False):
+            if use_table:
+                return (
+                    self.format_table(
+                        column.table, use_schema=use_schema, name=table_name
+                    )
+                    + "."
+                    + self.quote(name)
+                )
+            else:
+                return self.quote(name)
+        else:
+            # literal textual elements get stuck into ColumnClause a lot,
+            # which shouldn't get quoted
+
+            if use_table:
+                return (
+                    self.format_table(
+                        column.table, use_schema=use_schema, name=table_name
+                    )
+                    + "."
+                    + name
+                )
+            else:
+                return name
+
+    def format_table_seq(self, table, use_schema=True):
+        """Format table name and schema as a tuple."""
+
+        # Dialects with more levels in their fully qualified references
+        # ('database', 'owner', etc.) could override this and return
+        # a longer sequence.
+
+        effective_schema = self.schema_for_object(table)
+
+        if not self.omit_schema and use_schema and effective_schema:
+            return (
+                self.quote_schema(effective_schema),
+                self.format_table(table, use_schema=False),
+            )
+        else:
+            return (self.format_table(table, use_schema=False),)
+
+    @util.memoized_property
+    def _r_identifiers(self):
+        initial, final, escaped_final = (
+            re.escape(s)
+            for s in (
+                self.initial_quote,
+                self.final_quote,
+                self._escape_identifier(self.final_quote),
+            )
+        )
+        r = re.compile(
+            r"(?:"
+            r"(?:%(initial)s((?:%(escaped)s|[^%(final)s])+)%(final)s"
+            r"|([^\.]+))(?=\.|$))+"
+            % {"initial": initial, "final": final, "escaped": escaped_final}
+        )
+        return r
+
+    def unformat_identifiers(self, identifiers):
+        """Unpack 'schema.table.column'-like strings into components."""
+
+        r = self._r_identifiers
+        return [
+            self._unescape_identifier(i)
+            for i in [a or b for a, b in r.findall(identifiers)]
+        ]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/crud.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/crud.py
new file mode 100644
index 00000000..19af40ff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/crud.py
@@ -0,0 +1,1669 @@
+# sql/crud.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Functions used by compiler.py to determine the parameters rendered
+within INSERT and UPDATE statements.
+
+"""
+from __future__ import annotations
+
+import functools
+import operator
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import MutableMapping
+from typing import NamedTuple
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import Union
+
+from . import coercions
+from . import dml
+from . import elements
+from . import roles
+from .base import _DefaultDescriptionTuple
+from .dml import isinsert as _compile_state_isinsert
+from .elements import ColumnClause
+from .schema import default_is_clause_element
+from .schema import default_is_sequence
+from .selectable import Select
+from .selectable import TableClause
+from .. import exc
+from .. import util
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from .compiler import _BindNameForColProtocol
+    from .compiler import SQLCompiler
+    from .dml import _DMLColumnElement
+    from .dml import DMLState
+    from .dml import ValuesBase
+    from .elements import ColumnElement
+    from .elements import KeyedColumnElement
+    from .schema import _SQLExprDefault
+    from .schema import Column
+
+REQUIRED = util.symbol(
+    "REQUIRED",
+    """
+Placeholder for the value within a :class:`.BindParameter`
+which is required to be present when the statement is passed
+to :meth:`_engine.Connection.execute`.
+
+This symbol is typically used when a :func:`_expression.insert`
+or :func:`_expression.update` statement is compiled without parameter
+values present.
+
+""",
+)
+
+
+def _as_dml_column(c: ColumnElement[Any]) -> ColumnClause[Any]:
+    if not isinstance(c, ColumnClause):
+        raise exc.CompileError(
+            f"Can't create DML statement against column expression {c!r}"
+        )
+    return c
+
+
+_CrudParamElement = Tuple[
+    "ColumnElement[Any]",
+    str,  # column name
+    Optional[
+        Union[str, "_SQLExprDefault"]
+    ],  # bound parameter string or SQL expression to apply
+    Iterable[str],
+]
+_CrudParamElementStr = Tuple[
+    "KeyedColumnElement[Any]",
+    str,  # column name
+    str,  # bound parameter string
+    Iterable[str],
+]
+_CrudParamElementSQLExpr = Tuple[
+    "ColumnClause[Any]",
+    str,
+    "_SQLExprDefault",  # SQL expression to apply
+    Iterable[str],
+]
+
+_CrudParamSequence = List[_CrudParamElement]
+
+
+class _CrudParams(NamedTuple):
+    single_params: _CrudParamSequence
+    all_multi_params: List[Sequence[_CrudParamElementStr]]
+    is_default_metavalue_only: bool = False
+    use_insertmanyvalues: bool = False
+    use_sentinel_columns: Optional[Sequence[Column[Any]]] = None
+
+
+def _get_crud_params(
+    compiler: SQLCompiler,
+    stmt: ValuesBase,
+    compile_state: DMLState,
+    toplevel: bool,
+    **kw: Any,
+) -> _CrudParams:
+    """create a set of tuples representing column/string pairs for use
+    in an INSERT or UPDATE statement.
+
+    Also generates the Compiled object's postfetch, prefetch, and
+    returning column collections, used for default handling and ultimately
+    populating the CursorResult's prefetch_cols() and postfetch_cols()
+    collections.
+
+    """
+
+    # note: the _get_crud_params() system was written with the notion in mind
+    # that INSERT, UPDATE, DELETE are always the top level statement and
+    # that there is only one of them.  With the addition of CTEs that can
+    # make use of DML, this assumption is no longer accurate; the DML
+    # statement is not necessarily the top-level "row returning" thing
+    # and it is also theoretically possible (fortunately nobody has asked yet)
+    # to have a single statement with multiple DMLs inside of it via CTEs.
+
+    # the current _get_crud_params() design doesn't accommodate these cases
+    # right now.  It "just works" for a CTE that has a single DML inside of
+    # it, and for a CTE with multiple DML, it's not clear what would happen.
+
+    # overall, the "compiler.XYZ" collections here would need to be in a
+    # per-DML structure of some kind, and DefaultDialect would need to
+    # navigate these collections on a per-statement basis, with additional
+    # emphasis on the "toplevel returning data" statement.  However we
+    # still need to run through _get_crud_params() for all DML as we have
+    # Python / SQL generated column defaults that need to be rendered.
+
+    # if there is user need for this kind of thing, it's likely a post 2.0
+    # kind of change as it would require deep changes to DefaultDialect
+    # as well as here.
+
+    compiler.postfetch = []
+    compiler.insert_prefetch = []
+    compiler.update_prefetch = []
+    compiler.implicit_returning = []
+
+    visiting_cte = kw.get("visiting_cte", None)
+    if visiting_cte is not None:
+        # for insert -> CTE -> insert, don't populate an incoming
+        # _crud_accumulate_bind_names collection; the INSERT we process here
+        # will not be inline within the VALUES of the enclosing INSERT as the
+        # CTE is placed on the outside.  See issue #9173
+        kw.pop("accumulate_bind_names", None)
+    assert (
+        "accumulate_bind_names" not in kw
+    ), "Don't know how to handle insert within insert without a CTE"
+
+    # getters - these are normally just column.key,
+    # but in the case of mysql multi-table update, the rules for
+    # .key must conditionally take tablename into account
+    (
+        _column_as_key,
+        _getattr_col_key,
+        _col_bind_name,
+    ) = _key_getters_for_crud_column(compiler, stmt, compile_state)
+
+    compiler._get_bind_name_for_col = _col_bind_name
+
+    if stmt._returning and stmt._return_defaults:
+        raise exc.CompileError(
+            "Can't compile statement that includes returning() and "
+            "return_defaults() simultaneously"
+        )
+
+    if compile_state.isdelete:
+        _setup_delete_return_defaults(
+            compiler,
+            stmt,
+            compile_state,
+            (),
+            _getattr_col_key,
+            _column_as_key,
+            _col_bind_name,
+            (),
+            (),
+            toplevel,
+            kw,
+        )
+        return _CrudParams([], [])
+
+    # no parameters in the statement, no parameters in the
+    # compiled params - return binds for all columns
+    if compiler.column_keys is None and compile_state._no_parameters:
+        return _CrudParams(
+            [
+                (
+                    c,
+                    compiler.preparer.format_column(c),
+                    _create_bind_param(compiler, c, None, required=True),
+                    (c.key,),
+                )
+                for c in stmt.table.columns
+                if not c._omit_from_statements
+            ],
+            [],
+        )
+
+    stmt_parameter_tuples: Optional[
+        List[Tuple[Union[str, ColumnClause[Any]], Any]]
+    ]
+    spd: Optional[MutableMapping[_DMLColumnElement, Any]]
+
+    if (
+        _compile_state_isinsert(compile_state)
+        and compile_state._has_multi_parameters
+    ):
+        mp = compile_state._multi_parameters
+        assert mp is not None
+        spd = mp[0]
+        stmt_parameter_tuples = list(spd.items())
+        spd_str_key = {_column_as_key(key) for key in spd}
+    elif compile_state._ordered_values:
+        spd = compile_state._dict_parameters
+        stmt_parameter_tuples = compile_state._ordered_values
+        assert spd is not None
+        spd_str_key = {_column_as_key(key) for key in spd}
+    elif compile_state._dict_parameters:
+        spd = compile_state._dict_parameters
+        stmt_parameter_tuples = list(spd.items())
+        spd_str_key = {_column_as_key(key) for key in spd}
+    else:
+        stmt_parameter_tuples = spd = spd_str_key = None
+
+    # if we have statement parameters - set defaults in the
+    # compiled params
+    if compiler.column_keys is None:
+        parameters = {}
+    elif stmt_parameter_tuples:
+        assert spd_str_key is not None
+        parameters = {
+            _column_as_key(key): REQUIRED
+            for key in compiler.column_keys
+            if key not in spd_str_key
+        }
+    else:
+        parameters = {
+            _column_as_key(key): REQUIRED for key in compiler.column_keys
+        }
+
+    # create a list of column assignment clauses as tuples
+    values: List[_CrudParamElement] = []
+
+    if stmt_parameter_tuples is not None:
+        _get_stmt_parameter_tuples_params(
+            compiler,
+            compile_state,
+            parameters,
+            stmt_parameter_tuples,
+            _column_as_key,
+            values,
+            kw,
+        )
+
+    check_columns: Dict[str, ColumnClause[Any]] = {}
+
+    # special logic that only occurs for multi-table UPDATE
+    # statements
+    if dml.isupdate(compile_state) and compile_state.is_multitable:
+        _get_update_multitable_params(
+            compiler,
+            stmt,
+            compile_state,
+            stmt_parameter_tuples,
+            check_columns,
+            _col_bind_name,
+            _getattr_col_key,
+            values,
+            kw,
+        )
+
+    if _compile_state_isinsert(compile_state) and stmt._select_names:
+        # is an insert from select, is not a multiparams
+
+        assert not compile_state._has_multi_parameters
+
+        _scan_insert_from_select_cols(
+            compiler,
+            stmt,
+            compile_state,
+            parameters,
+            _getattr_col_key,
+            _column_as_key,
+            _col_bind_name,
+            check_columns,
+            values,
+            toplevel,
+            kw,
+        )
+        use_insertmanyvalues = False
+        use_sentinel_columns = None
+    else:
+        use_insertmanyvalues, use_sentinel_columns = _scan_cols(
+            compiler,
+            stmt,
+            compile_state,
+            parameters,
+            _getattr_col_key,
+            _column_as_key,
+            _col_bind_name,
+            check_columns,
+            values,
+            toplevel,
+            kw,
+        )
+
+    if parameters and stmt_parameter_tuples:
+        check = (
+            set(parameters)
+            .intersection(_column_as_key(k) for k, v in stmt_parameter_tuples)
+            .difference(check_columns)
+        )
+        if check:
+            raise exc.CompileError(
+                "Unconsumed column names: %s"
+                % (", ".join("%s" % (c,) for c in check))
+            )
+
+    is_default_metavalue_only = False
+
+    if (
+        _compile_state_isinsert(compile_state)
+        and compile_state._has_multi_parameters
+    ):
+        # is a multiparams, is not an insert from a select
+        assert not stmt._select_names
+        multi_extended_values = _extend_values_for_multiparams(
+            compiler,
+            stmt,
+            compile_state,
+            cast(
+                "Sequence[_CrudParamElementStr]",
+                values,
+            ),
+            cast("Callable[..., str]", _column_as_key),
+            kw,
+        )
+        return _CrudParams(values, multi_extended_values)
+    elif (
+        not values
+        and compiler.for_executemany
+        and compiler.dialect.supports_default_metavalue
+    ):
+        # convert an "INSERT DEFAULT VALUES"
+        # into INSERT (firstcol) VALUES (DEFAULT) which can be turned
+        # into an in-place multi values.  This supports
+        # insert_executemany_returning mode :)
+        values = [
+            (
+                _as_dml_column(stmt.table.columns[0]),
+                compiler.preparer.format_column(stmt.table.columns[0]),
+                compiler.dialect.default_metavalue_token,
+                (),
+            )
+        ]
+        is_default_metavalue_only = True
+
+    return _CrudParams(
+        values,
+        [],
+        is_default_metavalue_only=is_default_metavalue_only,
+        use_insertmanyvalues=use_insertmanyvalues,
+        use_sentinel_columns=use_sentinel_columns,
+    )
+
+
+@overload
+def _create_bind_param(
+    compiler: SQLCompiler,
+    col: ColumnElement[Any],
+    value: Any,
+    process: Literal[True] = ...,
+    required: bool = False,
+    name: Optional[str] = None,
+    **kw: Any,
+) -> str: ...
+
+
+@overload
+def _create_bind_param(
+    compiler: SQLCompiler,
+    col: ColumnElement[Any],
+    value: Any,
+    **kw: Any,
+) -> str: ...
+
+
+def _create_bind_param(
+    compiler: SQLCompiler,
+    col: ColumnElement[Any],
+    value: Any,
+    process: bool = True,
+    required: bool = False,
+    name: Optional[str] = None,
+    **kw: Any,
+) -> Union[str, elements.BindParameter[Any]]:
+    if name is None:
+        name = col.key
+    bindparam = elements.BindParameter(
+        name, value, type_=col.type, required=required
+    )
+    bindparam._is_crud = True
+    if process:
+        return bindparam._compiler_dispatch(compiler, **kw)
+    else:
+        return bindparam
+
+
+def _handle_values_anonymous_param(compiler, col, value, name, **kw):
+    # the insert() and update() constructs as of 1.4 will now produce anonymous
+    # bindparam() objects in the values() collections up front when given plain
+    # literal values.  This is so that cache key behaviors, which need to
+    # produce bound parameters in deterministic order without invoking any
+    # compilation here, can be applied to these constructs when they include
+    # values() (but not yet multi-values, which are not included in caching
+    # right now).
+    #
+    # in order to produce the desired "crud" style name for these parameters,
+    # which will also be targetable in engine/default.py through the usual
+    # conventions, apply our desired name to these unique parameters by
+    # populating the compiler truncated names cache with the desired name,
+    # rather than having
+    # compiler.visit_bindparam()->compiler._truncated_identifier make up a
+    # name.  Saves on call counts also.
+
+    # for INSERT/UPDATE that's a CTE, we don't need names to match to
+    # external parameters and these would also conflict in the case where
+    # multiple insert/update are combined together using CTEs
+    is_cte = "visiting_cte" in kw
+
+    if (
+        not is_cte
+        and value.unique
+        and isinstance(value.key, elements._truncated_label)
+    ):
+        compiler.truncated_names[("bindparam", value.key)] = name
+
+    if value.type._isnull:
+        # either unique parameter, or other bound parameters that were
+        # passed in directly
+        # set type to that of the column unconditionally
+        value = value._with_binary_element_type(col.type)
+
+    return value._compiler_dispatch(compiler, **kw)
+
+
+def _key_getters_for_crud_column(
+    compiler: SQLCompiler, stmt: ValuesBase, compile_state: DMLState
+) -> Tuple[
+    Callable[[Union[str, ColumnClause[Any]]], Union[str, Tuple[str, str]]],
+    Callable[[ColumnClause[Any]], Union[str, Tuple[str, str]]],
+    _BindNameForColProtocol,
+]:
+    if dml.isupdate(compile_state) and compile_state._extra_froms:
+        # when extra tables are present, refer to the columns
+        # in those extra tables as table-qualified, including in
+        # dictionaries and when rendering bind param names.
+        # the "main" table of the statement remains unqualified,
+        # allowing the most compatibility with a non-multi-table
+        # statement.
+        _et = set(compile_state._extra_froms)
+
+        c_key_role = functools.partial(
+            coercions.expect_as_key, roles.DMLColumnRole
+        )
+
+        def _column_as_key(
+            key: Union[ColumnClause[Any], str]
+        ) -> Union[str, Tuple[str, str]]:
+            str_key = c_key_role(key)
+            if hasattr(key, "table") and key.table in _et:
+                return (key.table.name, str_key)  # type: ignore
+            else:
+                return str_key
+
+        def _getattr_col_key(
+            col: ColumnClause[Any],
+        ) -> Union[str, Tuple[str, str]]:
+            if col.table in _et:
+                return (col.table.name, col.key)  # type: ignore
+            else:
+                return col.key
+
+        def _col_bind_name(col: ColumnClause[Any]) -> str:
+            if col.table in _et:
+                if TYPE_CHECKING:
+                    assert isinstance(col.table, TableClause)
+                return "%s_%s" % (col.table.name, col.key)
+            else:
+                return col.key
+
+    else:
+        _column_as_key = functools.partial(
+            coercions.expect_as_key, roles.DMLColumnRole
+        )
+        _getattr_col_key = _col_bind_name = operator.attrgetter("key")  # type: ignore  # noqa: E501
+
+    return _column_as_key, _getattr_col_key, _col_bind_name
+
+
+def _scan_insert_from_select_cols(
+    compiler,
+    stmt,
+    compile_state,
+    parameters,
+    _getattr_col_key,
+    _column_as_key,
+    _col_bind_name,
+    check_columns,
+    values,
+    toplevel,
+    kw,
+):
+    cols = [stmt.table.c[_column_as_key(name)] for name in stmt._select_names]
+
+    assert compiler.stack[-1]["selectable"] is stmt
+
+    compiler.stack[-1]["insert_from_select"] = stmt.select
+
+    add_select_cols: List[_CrudParamElementSQLExpr] = []
+    if stmt.include_insert_from_select_defaults:
+        col_set = set(cols)
+        for col in stmt.table.columns:
+            # omit columns that were not in the SELECT statement.
+            # this will omit columns marked as omit_from_statements naturally,
+            # as long as that col was not explicit in the SELECT.
+            # if an omit_from_statements col has a "default" on it, then
+            # we need to include it, as these defaults should still fire off.
+            # but, if it has that default and it's the "sentinel" default,
+            # we don't do sentinel default operations for insert_from_select
+            # here so we again omit it.
+            if (
+                col not in col_set
+                and col.default
+                and not col.default.is_sentinel
+            ):
+                cols.append(col)
+
+    for c in cols:
+        col_key = _getattr_col_key(c)
+        if col_key in parameters and col_key not in check_columns:
+            parameters.pop(col_key)
+            values.append((c, compiler.preparer.format_column(c), None, ()))
+        else:
+            _append_param_insert_select_hasdefault(
+                compiler, stmt, c, add_select_cols, kw
+            )
+
+    if add_select_cols:
+        values.extend(add_select_cols)
+        ins_from_select = compiler.stack[-1]["insert_from_select"]
+        if not isinstance(ins_from_select, Select):
+            raise exc.CompileError(
+                f"Can't extend statement for INSERT..FROM SELECT to include "
+                f"additional default-holding column(s) "
+                f"""{
+                    ', '.join(repr(key) for _, key, _, _ in add_select_cols)
+                }.  Convert the selectable to a subquery() first, or pass """
+                "include_defaults=False to Insert.from_select() to skip these "
+                "columns."
+            )
+        ins_from_select = ins_from_select._generate()
+        # copy raw_columns
+        ins_from_select._raw_columns = list(ins_from_select._raw_columns) + [
+            expr for _, _, expr, _ in add_select_cols
+        ]
+        compiler.stack[-1]["insert_from_select"] = ins_from_select
+
+
+def _scan_cols(
+    compiler,
+    stmt,
+    compile_state,
+    parameters,
+    _getattr_col_key,
+    _column_as_key,
+    _col_bind_name,
+    check_columns,
+    values,
+    toplevel,
+    kw,
+):
+    (
+        need_pks,
+        implicit_returning,
+        implicit_return_defaults,
+        postfetch_lastrowid,
+        use_insertmanyvalues,
+        use_sentinel_columns,
+    ) = _get_returning_modifiers(compiler, stmt, compile_state, toplevel)
+
+    assert compile_state.isupdate or compile_state.isinsert
+
+    if compile_state._parameter_ordering:
+        parameter_ordering = [
+            _column_as_key(key) for key in compile_state._parameter_ordering
+        ]
+        ordered_keys = set(parameter_ordering)
+        cols = [
+            stmt.table.c[key]
+            for key in parameter_ordering
+            if isinstance(key, str) and key in stmt.table.c
+        ] + [c for c in stmt.table.c if c.key not in ordered_keys]
+
+    else:
+        cols = stmt.table.columns
+
+    isinsert = _compile_state_isinsert(compile_state)
+    if isinsert and not compile_state._has_multi_parameters:
+        # new rules for #7998.  fetch lastrowid or implicit returning
+        # for autoincrement column even if parameter is NULL, for DBs that
+        # override NULL param for primary key (sqlite, mysql/mariadb)
+        autoincrement_col = stmt.table._autoincrement_column
+        insert_null_pk_still_autoincrements = (
+            compiler.dialect.insert_null_pk_still_autoincrements
+        )
+    else:
+        autoincrement_col = insert_null_pk_still_autoincrements = None
+
+    if stmt._supplemental_returning:
+        supplemental_returning = set(stmt._supplemental_returning)
+    else:
+        supplemental_returning = set()
+
+    compiler_implicit_returning = compiler.implicit_returning
+
+    # TODO - see TODO(return_defaults_columns) below
+    # cols_in_params = set()
+
+    for c in cols:
+        # scan through every column in the target table
+
+        col_key = _getattr_col_key(c)
+
+        if col_key in parameters and col_key not in check_columns:
+            # parameter is present for the column.  use that.
+
+            _append_param_parameter(
+                compiler,
+                stmt,
+                compile_state,
+                c,
+                col_key,
+                parameters,
+                _col_bind_name,
+                implicit_returning,
+                implicit_return_defaults,
+                postfetch_lastrowid,
+                values,
+                autoincrement_col,
+                insert_null_pk_still_autoincrements,
+                kw,
+            )
+
+            # TODO - see TODO(return_defaults_columns) below
+            # cols_in_params.add(c)
+
+        elif isinsert:
+            # no parameter is present and it's an insert.
+
+            if c.primary_key and need_pks:
+                # it's a primary key column, it will need to be generated by a
+                # default generator of some kind, and the statement expects
+                # inserted_primary_key to be available.
+
+                if implicit_returning:
+                    # we can use RETURNING, find out how to invoke this
+                    # column and get the value where RETURNING is an option.
+                    # we can inline server-side functions in this case.
+
+                    _append_param_insert_pk_returning(
+                        compiler, stmt, c, values, kw
+                    )
+                else:
+                    # otherwise, find out how to invoke this column
+                    # and get its value where RETURNING is not an option.
+                    # if we have to invoke a server-side function, we need
+                    # to pre-execute it.   or if this is a straight
+                    # autoincrement column and the dialect supports it
+                    # we can use cursor.lastrowid.
+
+                    _append_param_insert_pk_no_returning(
+                        compiler, stmt, c, values, kw
+                    )
+
+            elif c.default is not None:
+                # column has a default, but it's not a pk column, or it is but
+                # we don't need to get the pk back.
+                if not c.default.is_sentinel or (
+                    use_sentinel_columns is not None
+                ):
+                    _append_param_insert_hasdefault(
+                        compiler, stmt, c, implicit_return_defaults, values, kw
+                    )
+
+            elif c.server_default is not None:
+                # column has a DDL-level default, and is either not a pk
+                # column or we don't need the pk.
+                if implicit_return_defaults and c in implicit_return_defaults:
+                    compiler_implicit_returning.append(c)
+                elif not c.primary_key:
+                    compiler.postfetch.append(c)
+
+            elif implicit_return_defaults and c in implicit_return_defaults:
+                compiler_implicit_returning.append(c)
+
+            elif (
+                c.primary_key
+                and c is not stmt.table._autoincrement_column
+                and not c.nullable
+            ):
+                _warn_pk_with_no_anticipated_value(c)
+
+        elif compile_state.isupdate:
+            # no parameter is present and it's an insert.
+
+            _append_param_update(
+                compiler,
+                compile_state,
+                stmt,
+                c,
+                implicit_return_defaults,
+                values,
+                kw,
+            )
+
+        # adding supplemental cols to implicit_returning in table
+        # order so that order is maintained between multiple INSERT
+        # statements which may have different parameters included, but all
+        # have the same RETURNING clause
+        if (
+            c in supplemental_returning
+            and c not in compiler_implicit_returning
+        ):
+            compiler_implicit_returning.append(c)
+
+    if supplemental_returning:
+        # we should have gotten every col into implicit_returning,
+        # however supplemental returning can also have SQL functions etc.
+        # in it
+        remaining_supplemental = supplemental_returning.difference(
+            compiler_implicit_returning
+        )
+        compiler_implicit_returning.extend(
+            c
+            for c in stmt._supplemental_returning
+            if c in remaining_supplemental
+        )
+
+    # TODO(return_defaults_columns): there can still be more columns in
+    # _return_defaults_columns in the case that they are from something like an
+    # aliased of the table. we can add them here, however this breaks other ORM
+    # things. so this is for another day. see
+    # test/orm/dml/test_update_delete_where.py -> test_update_from_alias
+
+    # if stmt._return_defaults_columns:
+    #     compiler_implicit_returning.extend(
+    #         set(stmt._return_defaults_columns)
+    #         .difference(compiler_implicit_returning)
+    #         .difference(cols_in_params)
+    #     )
+
+    return (use_insertmanyvalues, use_sentinel_columns)
+
+
+def _setup_delete_return_defaults(
+    compiler,
+    stmt,
+    compile_state,
+    parameters,
+    _getattr_col_key,
+    _column_as_key,
+    _col_bind_name,
+    check_columns,
+    values,
+    toplevel,
+    kw,
+):
+    (_, _, implicit_return_defaults, *_) = _get_returning_modifiers(
+        compiler, stmt, compile_state, toplevel
+    )
+
+    if not implicit_return_defaults:
+        return
+
+    if stmt._return_defaults_columns:
+        compiler.implicit_returning.extend(implicit_return_defaults)
+
+    if stmt._supplemental_returning:
+        ir_set = set(compiler.implicit_returning)
+        compiler.implicit_returning.extend(
+            c for c in stmt._supplemental_returning if c not in ir_set
+        )
+
+
+def _append_param_parameter(
+    compiler,
+    stmt,
+    compile_state,
+    c,
+    col_key,
+    parameters,
+    _col_bind_name,
+    implicit_returning,
+    implicit_return_defaults,
+    postfetch_lastrowid,
+    values,
+    autoincrement_col,
+    insert_null_pk_still_autoincrements,
+    kw,
+):
+    value = parameters.pop(col_key)
+
+    col_value = compiler.preparer.format_column(
+        c, use_table=compile_state.include_table_with_column_exprs
+    )
+
+    accumulated_bind_names: Set[str] = set()
+
+    if coercions._is_literal(value):
+        if (
+            insert_null_pk_still_autoincrements
+            and c.primary_key
+            and c is autoincrement_col
+        ):
+            # support use case for #7998, fetch autoincrement cols
+            # even if value was given.
+
+            if postfetch_lastrowid:
+                compiler.postfetch_lastrowid = True
+            elif implicit_returning:
+                compiler.implicit_returning.append(c)
+
+        value = _create_bind_param(
+            compiler,
+            c,
+            value,
+            required=value is REQUIRED,
+            name=(
+                _col_bind_name(c)
+                if not _compile_state_isinsert(compile_state)
+                or not compile_state._has_multi_parameters
+                else "%s_m0" % _col_bind_name(c)
+            ),
+            accumulate_bind_names=accumulated_bind_names,
+            **kw,
+        )
+    elif value._is_bind_parameter:
+        if (
+            insert_null_pk_still_autoincrements
+            and value.value is None
+            and c.primary_key
+            and c is autoincrement_col
+        ):
+            # support use case for #7998, fetch autoincrement cols
+            # even if value was given
+            if implicit_returning:
+                compiler.implicit_returning.append(c)
+            elif compiler.dialect.postfetch_lastrowid:
+                compiler.postfetch_lastrowid = True
+
+        value = _handle_values_anonymous_param(
+            compiler,
+            c,
+            value,
+            name=(
+                _col_bind_name(c)
+                if not _compile_state_isinsert(compile_state)
+                or not compile_state._has_multi_parameters
+                else "%s_m0" % _col_bind_name(c)
+            ),
+            accumulate_bind_names=accumulated_bind_names,
+            **kw,
+        )
+    else:
+        # value is a SQL expression
+        value = compiler.process(
+            value.self_group(),
+            accumulate_bind_names=accumulated_bind_names,
+            **kw,
+        )
+
+        if compile_state.isupdate:
+            if implicit_return_defaults and c in implicit_return_defaults:
+                compiler.implicit_returning.append(c)
+
+            else:
+                compiler.postfetch.append(c)
+        else:
+            if c.primary_key:
+                if implicit_returning:
+                    compiler.implicit_returning.append(c)
+                elif compiler.dialect.postfetch_lastrowid:
+                    compiler.postfetch_lastrowid = True
+
+            elif implicit_return_defaults and (c in implicit_return_defaults):
+                compiler.implicit_returning.append(c)
+
+            else:
+                # postfetch specifically means, "we can SELECT the row we just
+                # inserted by primary key to get back the server generated
+                # defaults". so by definition this can't be used to get the
+                # primary key value back, because we need to have it ahead of
+                # time.
+
+                compiler.postfetch.append(c)
+
+    values.append((c, col_value, value, accumulated_bind_names))
+
+
+def _append_param_insert_pk_returning(compiler, stmt, c, values, kw):
+    """Create a primary key expression in the INSERT statement where
+    we want to populate result.inserted_primary_key and RETURNING
+    is available.
+
+    """
+    if c.default is not None:
+        if c.default.is_sequence:
+            if compiler.dialect.supports_sequences and (
+                not c.default.optional
+                or not compiler.dialect.sequences_optional
+            ):
+                accumulated_bind_names: Set[str] = set()
+                values.append(
+                    (
+                        c,
+                        compiler.preparer.format_column(c),
+                        compiler.process(
+                            c.default,
+                            accumulate_bind_names=accumulated_bind_names,
+                            **kw,
+                        ),
+                        accumulated_bind_names,
+                    )
+                )
+            compiler.implicit_returning.append(c)
+        elif c.default.is_clause_element:
+            accumulated_bind_names = set()
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(c),
+                    compiler.process(
+                        c.default.arg.self_group(),
+                        accumulate_bind_names=accumulated_bind_names,
+                        **kw,
+                    ),
+                    accumulated_bind_names,
+                )
+            )
+            compiler.implicit_returning.append(c)
+        else:
+            # client side default.  OK we can't use RETURNING, need to
+            # do a "prefetch", which in fact fetches the default value
+            # on the Python side
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(c),
+                    _create_insert_prefetch_bind_param(compiler, c, **kw),
+                    (c.key,),
+                )
+            )
+    elif c is stmt.table._autoincrement_column or c.server_default is not None:
+        compiler.implicit_returning.append(c)
+    elif not c.nullable:
+        # no .default, no .server_default, not autoincrement, we have
+        # no indication this primary key column will have any value
+        _warn_pk_with_no_anticipated_value(c)
+
+
+def _append_param_insert_pk_no_returning(compiler, stmt, c, values, kw):
+    """Create a primary key expression in the INSERT statement where
+    we want to populate result.inserted_primary_key and we cannot use
+    RETURNING.
+
+    Depending on the kind of default here we may create a bound parameter
+    in the INSERT statement and pre-execute a default generation function,
+    or we may use cursor.lastrowid if supported by the dialect.
+
+
+    """
+
+    if (
+        # column has a Python-side default
+        c.default is not None
+        and (
+            # and it either is not a sequence, or it is and we support
+            # sequences and want to invoke it
+            not c.default.is_sequence
+            or (
+                compiler.dialect.supports_sequences
+                and (
+                    not c.default.optional
+                    or not compiler.dialect.sequences_optional
+                )
+            )
+        )
+    ) or (
+        # column is the "autoincrement column"
+        c is stmt.table._autoincrement_column
+        and (
+            # dialect can't use cursor.lastrowid
+            not compiler.dialect.postfetch_lastrowid
+            and (
+                # column has a Sequence and we support those
+                (
+                    c.default is not None
+                    and c.default.is_sequence
+                    and compiler.dialect.supports_sequences
+                )
+                or
+                # column has no default on it, but dialect can run the
+                # "autoincrement" mechanism explicitly, e.g. PostgreSQL
+                # SERIAL we know the sequence name
+                (
+                    c.default is None
+                    and compiler.dialect.preexecute_autoincrement_sequences
+                )
+            )
+        )
+    ):
+        # do a pre-execute of the default
+        values.append(
+            (
+                c,
+                compiler.preparer.format_column(c),
+                _create_insert_prefetch_bind_param(compiler, c, **kw),
+                (c.key,),
+            )
+        )
+    elif (
+        c.default is None
+        and c.server_default is None
+        and not c.nullable
+        and c is not stmt.table._autoincrement_column
+    ):
+        # no .default, no .server_default, not autoincrement, we have
+        # no indication this primary key column will have any value
+        _warn_pk_with_no_anticipated_value(c)
+    elif compiler.dialect.postfetch_lastrowid:
+        # finally, where it seems like there will be a generated primary key
+        # value and we haven't set up any other way to fetch it, and the
+        # dialect supports cursor.lastrowid, switch on the lastrowid flag so
+        # that the DefaultExecutionContext calls upon cursor.lastrowid
+        compiler.postfetch_lastrowid = True
+
+
+def _append_param_insert_hasdefault(
+    compiler, stmt, c, implicit_return_defaults, values, kw
+):
+    if c.default.is_sequence:
+        if compiler.dialect.supports_sequences and (
+            not c.default.optional or not compiler.dialect.sequences_optional
+        ):
+            accumulated_bind_names: Set[str] = set()
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(c),
+                    compiler.process(
+                        c.default,
+                        accumulate_bind_names=accumulated_bind_names,
+                        **kw,
+                    ),
+                    accumulated_bind_names,
+                )
+            )
+            if implicit_return_defaults and c in implicit_return_defaults:
+                compiler.implicit_returning.append(c)
+            elif not c.primary_key:
+                compiler.postfetch.append(c)
+    elif c.default.is_clause_element:
+        accumulated_bind_names = set()
+        values.append(
+            (
+                c,
+                compiler.preparer.format_column(c),
+                compiler.process(
+                    c.default.arg.self_group(),
+                    accumulate_bind_names=accumulated_bind_names,
+                    **kw,
+                ),
+                accumulated_bind_names,
+            )
+        )
+
+        if implicit_return_defaults and c in implicit_return_defaults:
+            compiler.implicit_returning.append(c)
+        elif not c.primary_key:
+            # don't add primary key column to postfetch
+            compiler.postfetch.append(c)
+    else:
+        values.append(
+            (
+                c,
+                compiler.preparer.format_column(c),
+                _create_insert_prefetch_bind_param(compiler, c, **kw),
+                (c.key,),
+            )
+        )
+
+
+def _append_param_insert_select_hasdefault(
+    compiler: SQLCompiler,
+    stmt: ValuesBase,
+    c: ColumnClause[Any],
+    values: List[_CrudParamElementSQLExpr],
+    kw: Dict[str, Any],
+) -> None:
+    if default_is_sequence(c.default):
+        if compiler.dialect.supports_sequences and (
+            not c.default.optional or not compiler.dialect.sequences_optional
+        ):
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(c),
+                    c.default.next_value(),
+                    (),
+                )
+            )
+    elif default_is_clause_element(c.default):
+        values.append(
+            (
+                c,
+                compiler.preparer.format_column(c),
+                c.default.arg.self_group(),
+                (),
+            )
+        )
+    else:
+        values.append(
+            (
+                c,
+                compiler.preparer.format_column(c),
+                _create_insert_prefetch_bind_param(
+                    compiler, c, process=False, **kw
+                ),
+                (c.key,),
+            )
+        )
+
+
+def _append_param_update(
+    compiler, compile_state, stmt, c, implicit_return_defaults, values, kw
+):
+    include_table = compile_state.include_table_with_column_exprs
+    if c.onupdate is not None and not c.onupdate.is_sequence:
+        if c.onupdate.is_clause_element:
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(
+                        c,
+                        use_table=include_table,
+                    ),
+                    compiler.process(c.onupdate.arg.self_group(), **kw),
+                    (),
+                )
+            )
+            if implicit_return_defaults and c in implicit_return_defaults:
+                compiler.implicit_returning.append(c)
+            else:
+                compiler.postfetch.append(c)
+        else:
+            values.append(
+                (
+                    c,
+                    compiler.preparer.format_column(
+                        c,
+                        use_table=include_table,
+                    ),
+                    _create_update_prefetch_bind_param(compiler, c, **kw),
+                    (c.key,),
+                )
+            )
+    elif c.server_onupdate is not None:
+        if implicit_return_defaults and c in implicit_return_defaults:
+            compiler.implicit_returning.append(c)
+        else:
+            compiler.postfetch.append(c)
+    elif (
+        implicit_return_defaults
+        and (stmt._return_defaults_columns or not stmt._return_defaults)
+        and c in implicit_return_defaults
+    ):
+        compiler.implicit_returning.append(c)
+
+
+@overload
+def _create_insert_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: Literal[True] = ...,
+    **kw: Any,
+) -> str: ...
+
+
+@overload
+def _create_insert_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: Literal[False],
+    **kw: Any,
+) -> elements.BindParameter[Any]: ...
+
+
+def _create_insert_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: bool = True,
+    name: Optional[str] = None,
+    **kw: Any,
+) -> Union[elements.BindParameter[Any], str]:
+    param = _create_bind_param(
+        compiler, c, None, process=process, name=name, **kw
+    )
+    compiler.insert_prefetch.append(c)  # type: ignore
+    return param
+
+
+@overload
+def _create_update_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: Literal[True] = ...,
+    **kw: Any,
+) -> str: ...
+
+
+@overload
+def _create_update_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: Literal[False],
+    **kw: Any,
+) -> elements.BindParameter[Any]: ...
+
+
+def _create_update_prefetch_bind_param(
+    compiler: SQLCompiler,
+    c: ColumnElement[Any],
+    process: bool = True,
+    name: Optional[str] = None,
+    **kw: Any,
+) -> Union[elements.BindParameter[Any], str]:
+    param = _create_bind_param(
+        compiler, c, None, process=process, name=name, **kw
+    )
+    compiler.update_prefetch.append(c)  # type: ignore
+    return param
+
+
+class _multiparam_column(elements.ColumnElement[Any]):
+    _is_multiparam_column = True
+
+    def __init__(self, original, index):
+        self.index = index
+        self.key = "%s_m%d" % (original.key, index + 1)
+        self.original = original
+        self.default = original.default
+        self.type = original.type
+
+    def compare(self, other, **kw):
+        raise NotImplementedError()
+
+    def _copy_internals(self, **kw):
+        raise NotImplementedError()
+
+    def __eq__(self, other):
+        return (
+            isinstance(other, _multiparam_column)
+            and other.key == self.key
+            and other.original == self.original
+        )
+
+    @util.memoized_property
+    def _default_description_tuple(self) -> _DefaultDescriptionTuple:
+        """used by default.py -> _process_execute_defaults()"""
+
+        return _DefaultDescriptionTuple._from_column_default(self.default)
+
+    @util.memoized_property
+    def _onupdate_description_tuple(self) -> _DefaultDescriptionTuple:
+        """used by default.py -> _process_execute_defaults()"""
+
+        return _DefaultDescriptionTuple._from_column_default(self.onupdate)
+
+
+def _process_multiparam_default_bind(
+    compiler: SQLCompiler,
+    stmt: ValuesBase,
+    c: KeyedColumnElement[Any],
+    index: int,
+    kw: Dict[str, Any],
+) -> str:
+    if not c.default:
+        raise exc.CompileError(
+            "INSERT value for column %s is explicitly rendered as a bound"
+            "parameter in the VALUES clause; "
+            "a Python-side value or SQL expression is required" % c
+        )
+    elif default_is_clause_element(c.default):
+        return compiler.process(c.default.arg.self_group(), **kw)
+    elif c.default.is_sequence:
+        # these conditions would have been established
+        # by append_param_insert_(?:hasdefault|pk_returning|pk_no_returning)
+        # in order for us to be here, so these don't need to be
+        # checked
+        # assert compiler.dialect.supports_sequences and (
+        #    not c.default.optional
+        #    or not compiler.dialect.sequences_optional
+        # )
+        return compiler.process(c.default, **kw)
+    else:
+        col = _multiparam_column(c, index)
+        assert isinstance(stmt, dml.Insert)
+        return _create_insert_prefetch_bind_param(
+            compiler, col, process=True, **kw
+        )
+
+
+def _get_update_multitable_params(
+    compiler,
+    stmt,
+    compile_state,
+    stmt_parameter_tuples,
+    check_columns,
+    _col_bind_name,
+    _getattr_col_key,
+    values,
+    kw,
+):
+    normalized_params = {
+        coercions.expect(roles.DMLColumnRole, c): param
+        for c, param in stmt_parameter_tuples or ()
+    }
+
+    include_table = compile_state.include_table_with_column_exprs
+
+    affected_tables = set()
+    for t in compile_state._extra_froms:
+        for c in t.c:
+            if c in normalized_params:
+                affected_tables.add(t)
+                check_columns[_getattr_col_key(c)] = c
+                value = normalized_params[c]
+
+                col_value = compiler.process(c, include_table=include_table)
+                if coercions._is_literal(value):
+                    value = _create_bind_param(
+                        compiler,
+                        c,
+                        value,
+                        required=value is REQUIRED,
+                        name=_col_bind_name(c),
+                        **kw,  # TODO: no test coverage for literal binds here
+                    )
+                    accumulated_bind_names: Iterable[str] = (c.key,)
+                elif value._is_bind_parameter:
+                    cbn = _col_bind_name(c)
+                    value = _handle_values_anonymous_param(
+                        compiler, c, value, name=cbn, **kw
+                    )
+                    accumulated_bind_names = (cbn,)
+                else:
+                    compiler.postfetch.append(c)
+                    value = compiler.process(value.self_group(), **kw)
+                    accumulated_bind_names = ()
+                values.append((c, col_value, value, accumulated_bind_names))
+    # determine tables which are actually to be updated - process onupdate
+    # and server_onupdate for these
+    for t in affected_tables:
+        for c in t.c:
+            if c in normalized_params:
+                continue
+            elif c.onupdate is not None and not c.onupdate.is_sequence:
+                if c.onupdate.is_clause_element:
+                    values.append(
+                        (
+                            c,
+                            compiler.process(c, include_table=include_table),
+                            compiler.process(
+                                c.onupdate.arg.self_group(), **kw
+                            ),
+                            (),
+                        )
+                    )
+                    compiler.postfetch.append(c)
+                else:
+                    values.append(
+                        (
+                            c,
+                            compiler.process(c, include_table=include_table),
+                            _create_update_prefetch_bind_param(
+                                compiler, c, name=_col_bind_name(c), **kw
+                            ),
+                            (c.key,),
+                        )
+                    )
+            elif c.server_onupdate is not None:
+                compiler.postfetch.append(c)
+
+
+def _extend_values_for_multiparams(
+    compiler: SQLCompiler,
+    stmt: ValuesBase,
+    compile_state: DMLState,
+    initial_values: Sequence[_CrudParamElementStr],
+    _column_as_key: Callable[..., str],
+    kw: Dict[str, Any],
+) -> List[Sequence[_CrudParamElementStr]]:
+    values_0 = initial_values
+    values = [initial_values]
+
+    mp = compile_state._multi_parameters
+    assert mp is not None
+    for i, row in enumerate(mp[1:]):
+        extension: List[_CrudParamElementStr] = []
+
+        row = {_column_as_key(key): v for key, v in row.items()}
+
+        for col, col_expr, param, accumulated_names in values_0:
+            if col.key in row:
+                key = col.key
+
+                if coercions._is_literal(row[key]):
+                    new_param = _create_bind_param(
+                        compiler,
+                        col,
+                        row[key],
+                        name="%s_m%d" % (col.key, i + 1),
+                        **kw,
+                    )
+                else:
+                    new_param = compiler.process(row[key].self_group(), **kw)
+            else:
+                new_param = _process_multiparam_default_bind(
+                    compiler, stmt, col, i, kw
+                )
+
+            extension.append((col, col_expr, new_param, accumulated_names))
+
+        values.append(extension)
+
+    return values
+
+
+def _get_stmt_parameter_tuples_params(
+    compiler,
+    compile_state,
+    parameters,
+    stmt_parameter_tuples,
+    _column_as_key,
+    values,
+    kw,
+):
+    for k, v in stmt_parameter_tuples:
+        colkey = _column_as_key(k)
+        if colkey is not None:
+            parameters.setdefault(colkey, v)
+        else:
+            # a non-Column expression on the left side;
+            # add it to values() in an "as-is" state,
+            # coercing right side to bound param
+
+            # note one of the main use cases for this is array slice
+            # updates on PostgreSQL, as the left side is also an expression.
+
+            col_expr = compiler.process(
+                k, include_table=compile_state.include_table_with_column_exprs
+            )
+
+            if coercions._is_literal(v):
+                v = compiler.process(
+                    elements.BindParameter(None, v, type_=k.type), **kw
+                )
+            else:
+                if v._is_bind_parameter and v.type._isnull:
+                    # either unique parameter, or other bound parameters that
+                    # were passed in directly
+                    # set type to that of the column unconditionally
+                    v = v._with_binary_element_type(k.type)
+
+                v = compiler.process(v.self_group(), **kw)
+
+            # TODO: not sure if accumulated_bind_names applies here
+            values.append((k, col_expr, v, ()))
+
+
+def _get_returning_modifiers(compiler, stmt, compile_state, toplevel):
+    """determines RETURNING strategy, if any, for the statement.
+
+    This is where it's determined what we need to fetch from the
+    INSERT or UPDATE statement after it's invoked.
+
+    """
+
+    dialect = compiler.dialect
+
+    need_pks = (
+        toplevel
+        and _compile_state_isinsert(compile_state)
+        and not stmt._inline
+        and (
+            not compiler.for_executemany
+            or (dialect.insert_executemany_returning and stmt._return_defaults)
+        )
+        and not stmt._returning
+        # and (not stmt._returning or stmt._return_defaults)
+        and not compile_state._has_multi_parameters
+    )
+
+    # check if we have access to simple cursor.lastrowid.  we can use that
+    # after the INSERT if that's all we need.
+    postfetch_lastrowid = (
+        need_pks
+        and dialect.postfetch_lastrowid
+        and stmt.table._autoincrement_column is not None
+    )
+
+    # see if we want to add RETURNING to an INSERT in order to get
+    # primary key columns back.  This would be instead of postfetch_lastrowid
+    # if that's set.
+    implicit_returning = (
+        # statement itself can veto it
+        need_pks
+        # the dialect can veto it if it just doesnt support RETURNING
+        # with INSERT
+        and dialect.insert_returning
+        # user-defined implicit_returning on Table can veto it
+        and compile_state._primary_table.implicit_returning
+        # the compile_state can veto it (SQlite uses this to disable
+        # RETURNING for an ON CONFLICT insert, as SQLite does not return
+        # for rows that were updated, which is wrong)
+        and compile_state._supports_implicit_returning
+        and (
+            # since we support MariaDB and SQLite which also support lastrowid,
+            # decide if we should use lastrowid or RETURNING.  for insert
+            # that didnt call return_defaults() and has just one set of
+            # parameters, we can use lastrowid.   this is more "traditional"
+            # and a lot of weird use cases are supported by it.
+            # SQLite lastrowid times 3x faster than returning,
+            # Mariadb lastrowid 2x faster than returning
+            (not postfetch_lastrowid or dialect.favor_returning_over_lastrowid)
+            or compile_state._has_multi_parameters
+            or stmt._return_defaults
+        )
+    )
+    if implicit_returning:
+        postfetch_lastrowid = False
+
+    if _compile_state_isinsert(compile_state):
+        should_implicit_return_defaults = (
+            implicit_returning and stmt._return_defaults
+        )
+        explicit_returning = (
+            should_implicit_return_defaults
+            or stmt._returning
+            or stmt._supplemental_returning
+        )
+        use_insertmanyvalues = (
+            toplevel
+            and compiler.for_executemany
+            and dialect.use_insertmanyvalues
+            and (
+                explicit_returning or dialect.use_insertmanyvalues_wo_returning
+            )
+        )
+
+        use_sentinel_columns = None
+        if (
+            use_insertmanyvalues
+            and explicit_returning
+            and stmt._sort_by_parameter_order
+        ):
+            use_sentinel_columns = compiler._get_sentinel_column_for_table(
+                stmt.table
+            )
+
+    elif compile_state.isupdate:
+        should_implicit_return_defaults = (
+            stmt._return_defaults
+            and compile_state._primary_table.implicit_returning
+            and compile_state._supports_implicit_returning
+            and dialect.update_returning
+        )
+        use_insertmanyvalues = False
+        use_sentinel_columns = None
+    elif compile_state.isdelete:
+        should_implicit_return_defaults = (
+            stmt._return_defaults
+            and compile_state._primary_table.implicit_returning
+            and compile_state._supports_implicit_returning
+            and dialect.delete_returning
+        )
+        use_insertmanyvalues = False
+        use_sentinel_columns = None
+    else:
+        should_implicit_return_defaults = False  # pragma: no cover
+        use_insertmanyvalues = False
+        use_sentinel_columns = None
+
+    if should_implicit_return_defaults:
+        if not stmt._return_defaults_columns:
+            # TODO: this is weird.  See #9685 where we have to
+            # take an extra step to prevent this from happening.  why
+            # would this ever be *all* columns?  but if we set to blank, then
+            # that seems to break things also in the ORM.  So we should
+            # try to clean this up and figure out what return_defaults
+            # needs to do w/ the ORM etc. here
+            implicit_return_defaults = set(stmt.table.c)
+        else:
+            implicit_return_defaults = set(stmt._return_defaults_columns)
+    else:
+        implicit_return_defaults = None
+
+    return (
+        need_pks,
+        implicit_returning or should_implicit_return_defaults,
+        implicit_return_defaults,
+        postfetch_lastrowid,
+        use_insertmanyvalues,
+        use_sentinel_columns,
+    )
+
+
+def _warn_pk_with_no_anticipated_value(c):
+    msg = (
+        "Column '%s.%s' is marked as a member of the "
+        "primary key for table '%s', "
+        "but has no Python-side or server-side default generator indicated, "
+        "nor does it indicate 'autoincrement=True' or 'nullable=True', "
+        "and no explicit value is passed.  "
+        "Primary key columns typically may not store NULL."
+        % (c.table.fullname, c.name, c.table.fullname)
+    )
+    if len(c.table.primary_key) > 1:
+        msg += (
+            " Note that as of SQLAlchemy 1.1, 'autoincrement=True' must be "
+            "indicated explicitly for composite (e.g. multicolumn) primary "
+            "keys if AUTO_INCREMENT/SERIAL/IDENTITY "
+            "behavior is expected for one of the columns in the primary key. "
+            "CREATE TABLE statements are impacted by this change as well on "
+            "most backends."
+        )
+    util.warn(msg)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/ddl.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/ddl.py
new file mode 100644
index 00000000..0950043b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/ddl.py
@@ -0,0 +1,1388 @@
+# sql/ddl.py
+# Copyright (C) 2009-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""
+Provides the hierarchy of DDL-defining schema items as well as routines
+to invoke them for a create/drop call.
+
+"""
+from __future__ import annotations
+
+import contextlib
+import typing
+from typing import Any
+from typing import Callable
+from typing import Iterable
+from typing import List
+from typing import Optional
+from typing import Sequence as typing_Sequence
+from typing import Tuple
+
+from . import roles
+from .base import _generative
+from .base import Executable
+from .base import SchemaVisitor
+from .elements import ClauseElement
+from .. import exc
+from .. import util
+from ..util import topological
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+if typing.TYPE_CHECKING:
+    from .compiler import Compiled
+    from .compiler import DDLCompiler
+    from .elements import BindParameter
+    from .schema import Constraint
+    from .schema import ForeignKeyConstraint
+    from .schema import SchemaItem
+    from .schema import Sequence
+    from .schema import Table
+    from .selectable import TableClause
+    from ..engine.base import Connection
+    from ..engine.interfaces import CacheStats
+    from ..engine.interfaces import CompiledCacheType
+    from ..engine.interfaces import Dialect
+    from ..engine.interfaces import SchemaTranslateMapType
+
+
+class BaseDDLElement(ClauseElement):
+    """The root of DDL constructs, including those that are sub-elements
+    within the "create table" and other processes.
+
+    .. versionadded:: 2.0
+
+    """
+
+    _hierarchy_supports_caching = False
+    """disable cache warnings for all _DDLCompiles subclasses. """
+
+    def _compiler(self, dialect, **kw):
+        """Return a compiler appropriate for this ClauseElement, given a
+        Dialect."""
+
+        return dialect.ddl_compiler(dialect, self, **kw)
+
+    def _compile_w_cache(
+        self,
+        dialect: Dialect,
+        *,
+        compiled_cache: Optional[CompiledCacheType],
+        column_keys: List[str],
+        for_executemany: bool = False,
+        schema_translate_map: Optional[SchemaTranslateMapType] = None,
+        **kw: Any,
+    ) -> Tuple[
+        Compiled, Optional[typing_Sequence[BindParameter[Any]]], CacheStats
+    ]:
+        raise NotImplementedError()
+
+
+class DDLIfCallable(Protocol):
+    def __call__(
+        self,
+        ddl: BaseDDLElement,
+        target: SchemaItem,
+        bind: Optional[Connection],
+        tables: Optional[List[Table]] = None,
+        state: Optional[Any] = None,
+        *,
+        dialect: Dialect,
+        compiler: Optional[DDLCompiler] = ...,
+        checkfirst: bool,
+    ) -> bool: ...
+
+
+class DDLIf(typing.NamedTuple):
+    dialect: Optional[str]
+    callable_: Optional[DDLIfCallable]
+    state: Optional[Any]
+
+    def _should_execute(
+        self,
+        ddl: BaseDDLElement,
+        target: SchemaItem,
+        bind: Optional[Connection],
+        compiler: Optional[DDLCompiler] = None,
+        **kw: Any,
+    ) -> bool:
+        if bind is not None:
+            dialect = bind.dialect
+        elif compiler is not None:
+            dialect = compiler.dialect
+        else:
+            assert False, "compiler or dialect is required"
+
+        if isinstance(self.dialect, str):
+            if self.dialect != dialect.name:
+                return False
+        elif isinstance(self.dialect, (tuple, list, set)):
+            if dialect.name not in self.dialect:
+                return False
+        if self.callable_ is not None and not self.callable_(
+            ddl,
+            target,
+            bind,
+            state=self.state,
+            dialect=dialect,
+            compiler=compiler,
+            **kw,
+        ):
+            return False
+
+        return True
+
+
+class ExecutableDDLElement(roles.DDLRole, Executable, BaseDDLElement):
+    """Base class for standalone executable DDL expression constructs.
+
+    This class is the base for the general purpose :class:`.DDL` class,
+    as well as the various create/drop clause constructs such as
+    :class:`.CreateTable`, :class:`.DropTable`, :class:`.AddConstraint`,
+    etc.
+
+    .. versionchanged:: 2.0  :class:`.ExecutableDDLElement` is renamed from
+       :class:`.DDLElement`, which still exists for backwards compatibility.
+
+    :class:`.ExecutableDDLElement` integrates closely with SQLAlchemy events,
+    introduced in :ref:`event_toplevel`.  An instance of one is
+    itself an event receiving callable::
+
+        event.listen(
+            users,
+            "after_create",
+            AddConstraint(constraint).execute_if(dialect="postgresql"),
+        )
+
+    .. seealso::
+
+        :class:`.DDL`
+
+        :class:`.DDLEvents`
+
+        :ref:`event_toplevel`
+
+        :ref:`schema_ddl_sequences`
+
+    """
+
+    _ddl_if: Optional[DDLIf] = None
+    target: Optional[SchemaItem] = None
+
+    def _execute_on_connection(
+        self, connection, distilled_params, execution_options
+    ):
+        return connection._execute_ddl(
+            self, distilled_params, execution_options
+        )
+
+    @_generative
+    def against(self, target: SchemaItem) -> Self:
+        """Return a copy of this :class:`_schema.ExecutableDDLElement` which
+        will include the given target.
+
+        This essentially applies the given item to the ``.target`` attribute of
+        the returned :class:`_schema.ExecutableDDLElement` object. This target
+        is then usable by event handlers and compilation routines in order to
+        provide services such as tokenization of a DDL string in terms of a
+        particular :class:`_schema.Table`.
+
+        When a :class:`_schema.ExecutableDDLElement` object is established as
+        an event handler for the :meth:`_events.DDLEvents.before_create` or
+        :meth:`_events.DDLEvents.after_create` events, and the event then
+        occurs for a given target such as a :class:`_schema.Constraint` or
+        :class:`_schema.Table`, that target is established with a copy of the
+        :class:`_schema.ExecutableDDLElement` object using this method, which
+        then proceeds to the :meth:`_schema.ExecutableDDLElement.execute`
+        method in order to invoke the actual DDL instruction.
+
+        :param target: a :class:`_schema.SchemaItem` that will be the subject
+         of a DDL operation.
+
+        :return: a copy of this :class:`_schema.ExecutableDDLElement` with the
+         ``.target`` attribute assigned to the given
+         :class:`_schema.SchemaItem`.
+
+        .. seealso::
+
+            :class:`_schema.DDL` - uses tokenization against the "target" when
+            processing the DDL string.
+
+        """
+        self.target = target
+        return self
+
+    @_generative
+    def execute_if(
+        self,
+        dialect: Optional[str] = None,
+        callable_: Optional[DDLIfCallable] = None,
+        state: Optional[Any] = None,
+    ) -> Self:
+        r"""Return a callable that will execute this
+        :class:`_ddl.ExecutableDDLElement` conditionally within an event
+        handler.
+
+        Used to provide a wrapper for event listening::
+
+            event.listen(
+                metadata,
+                "before_create",
+                DDL("my_ddl").execute_if(dialect="postgresql"),
+            )
+
+        :param dialect: May be a string or tuple of strings.
+          If a string, it will be compared to the name of the
+          executing database dialect::
+
+            DDL("something").execute_if(dialect="postgresql")
+
+          If a tuple, specifies multiple dialect names::
+
+            DDL("something").execute_if(dialect=("postgresql", "mysql"))
+
+        :param callable\_: A callable, which will be invoked with
+          three positional arguments as well as optional keyword
+          arguments:
+
+            :ddl:
+              This DDL element.
+
+            :target:
+              The :class:`_schema.Table` or :class:`_schema.MetaData`
+              object which is the
+              target of this event. May be None if the DDL is executed
+              explicitly.
+
+            :bind:
+              The :class:`_engine.Connection` being used for DDL execution.
+              May be None if this construct is being created inline within
+              a table, in which case ``compiler`` will be present.
+
+            :tables:
+              Optional keyword argument - a list of Table objects which are to
+              be created/ dropped within a MetaData.create_all() or drop_all()
+              method call.
+
+            :dialect: keyword argument, but always present - the
+              :class:`.Dialect` involved in the operation.
+
+            :compiler: keyword argument.  Will be ``None`` for an engine
+              level DDL invocation, but will refer to a :class:`.DDLCompiler`
+              if this DDL element is being created inline within a table.
+
+            :state:
+              Optional keyword argument - will be the ``state`` argument
+              passed to this function.
+
+            :checkfirst:
+             Keyword argument, will be True if the 'checkfirst' flag was
+             set during the call to ``create()``, ``create_all()``,
+             ``drop()``, ``drop_all()``.
+
+          If the callable returns a True value, the DDL statement will be
+          executed.
+
+        :param state: any value which will be passed to the callable\_
+          as the ``state`` keyword argument.
+
+        .. seealso::
+
+            :meth:`.SchemaItem.ddl_if`
+
+            :class:`.DDLEvents`
+
+            :ref:`event_toplevel`
+
+        """
+        self._ddl_if = DDLIf(dialect, callable_, state)
+        return self
+
+    def _should_execute(self, target, bind, **kw):
+        if self._ddl_if is None:
+            return True
+        else:
+            return self._ddl_if._should_execute(self, target, bind, **kw)
+
+    def _invoke_with(self, bind):
+        if self._should_execute(self.target, bind):
+            return bind.execute(self)
+
+    def __call__(self, target, bind, **kw):
+        """Execute the DDL as a ddl_listener."""
+
+        self.against(target)._invoke_with(bind)
+
+    def _generate(self):
+        s = self.__class__.__new__(self.__class__)
+        s.__dict__ = self.__dict__.copy()
+        return s
+
+
+DDLElement = ExecutableDDLElement
+""":class:`.DDLElement` is renamed to :class:`.ExecutableDDLElement`."""
+
+
+class DDL(ExecutableDDLElement):
+    """A literal DDL statement.
+
+    Specifies literal SQL DDL to be executed by the database.  DDL objects
+    function as DDL event listeners, and can be subscribed to those events
+    listed in :class:`.DDLEvents`, using either :class:`_schema.Table` or
+    :class:`_schema.MetaData` objects as targets.
+    Basic templating support allows
+    a single DDL instance to handle repetitive tasks for multiple tables.
+
+    Examples::
+
+      from sqlalchemy import event, DDL
+
+      tbl = Table("users", metadata, Column("uid", Integer))
+      event.listen(tbl, "before_create", DDL("DROP TRIGGER users_trigger"))
+
+      spow = DDL("ALTER TABLE %(table)s SET secretpowers TRUE")
+      event.listen(tbl, "after_create", spow.execute_if(dialect="somedb"))
+
+      drop_spow = DDL("ALTER TABLE users SET secretpowers FALSE")
+      connection.execute(drop_spow)
+
+    When operating on Table events, the following ``statement``
+    string substitutions are available:
+
+    .. sourcecode:: text
+
+      %(table)s  - the Table name, with any required quoting applied
+      %(schema)s - the schema name, with any required quoting applied
+      %(fullname)s - the Table name including schema, quoted if needed
+
+    The DDL's "context", if any, will be combined with the standard
+    substitutions noted above.  Keys present in the context will override
+    the standard substitutions.
+
+    """
+
+    __visit_name__ = "ddl"
+
+    def __init__(self, statement, context=None):
+        """Create a DDL statement.
+
+        :param statement:
+          A string or unicode string to be executed.  Statements will be
+          processed with Python's string formatting operator using
+          a fixed set of string substitutions, as well as additional
+          substitutions provided by the optional :paramref:`.DDL.context`
+          parameter.
+
+          A literal '%' in a statement must be escaped as '%%'.
+
+          SQL bind parameters are not available in DDL statements.
+
+        :param context:
+          Optional dictionary, defaults to None.  These values will be
+          available for use in string substitutions on the DDL statement.
+
+        .. seealso::
+
+            :class:`.DDLEvents`
+
+            :ref:`event_toplevel`
+
+        """
+
+        if not isinstance(statement, str):
+            raise exc.ArgumentError(
+                "Expected a string or unicode SQL statement, got '%r'"
+                % statement
+            )
+
+        self.statement = statement
+        self.context = context or {}
+
+    def __repr__(self):
+        parts = [repr(self.statement)]
+        if self.context:
+            parts.append(f"context={self.context}")
+
+        return "<%s@%s; %s>" % (
+            type(self).__name__,
+            id(self),
+            ", ".join(parts),
+        )
+
+
+class _CreateDropBase(ExecutableDDLElement):
+    """Base class for DDL constructs that represent CREATE and DROP or
+    equivalents.
+
+    The common theme of _CreateDropBase is a single
+    ``element`` attribute which refers to the element
+    to be created or dropped.
+
+    """
+
+    def __init__(
+        self,
+        element,
+    ):
+        self.element = self.target = element
+        self._ddl_if = getattr(element, "_ddl_if", None)
+
+    @property
+    def stringify_dialect(self):
+        return self.element.create_drop_stringify_dialect
+
+    def _create_rule_disable(self, compiler):
+        """Allow disable of _create_rule using a callable.
+
+        Pass to _create_rule using
+        util.portable_instancemethod(self._create_rule_disable)
+        to retain serializability.
+
+        """
+        return False
+
+
+class _CreateBase(_CreateDropBase):
+    def __init__(self, element, if_not_exists=False):
+        super().__init__(element)
+        self.if_not_exists = if_not_exists
+
+
+class _DropBase(_CreateDropBase):
+    def __init__(self, element, if_exists=False):
+        super().__init__(element)
+        self.if_exists = if_exists
+
+
+class CreateSchema(_CreateBase):
+    """Represent a CREATE SCHEMA statement.
+
+    The argument here is the string name of the schema.
+
+    """
+
+    __visit_name__ = "create_schema"
+
+    stringify_dialect = "default"
+
+    def __init__(
+        self,
+        name: str,
+        if_not_exists: bool = False,
+    ):
+        """Create a new :class:`.CreateSchema` construct."""
+
+        super().__init__(element=name, if_not_exists=if_not_exists)
+
+
+class DropSchema(_DropBase):
+    """Represent a DROP SCHEMA statement.
+
+    The argument here is the string name of the schema.
+
+    """
+
+    __visit_name__ = "drop_schema"
+
+    stringify_dialect = "default"
+
+    def __init__(
+        self,
+        name: str,
+        cascade: bool = False,
+        if_exists: bool = False,
+    ):
+        """Create a new :class:`.DropSchema` construct."""
+
+        super().__init__(element=name, if_exists=if_exists)
+        self.cascade = cascade
+
+
+class CreateTable(_CreateBase):
+    """Represent a CREATE TABLE statement."""
+
+    __visit_name__ = "create_table"
+
+    def __init__(
+        self,
+        element: Table,
+        include_foreign_key_constraints: Optional[
+            typing_Sequence[ForeignKeyConstraint]
+        ] = None,
+        if_not_exists: bool = False,
+    ):
+        """Create a :class:`.CreateTable` construct.
+
+        :param element: a :class:`_schema.Table` that's the subject
+         of the CREATE
+        :param on: See the description for 'on' in :class:`.DDL`.
+        :param include_foreign_key_constraints: optional sequence of
+         :class:`_schema.ForeignKeyConstraint` objects that will be included
+         inline within the CREATE construct; if omitted, all foreign key
+         constraints that do not specify use_alter=True are included.
+
+        :param if_not_exists: if True, an IF NOT EXISTS operator will be
+         applied to the construct.
+
+         .. versionadded:: 1.4.0b2
+
+        """
+        super().__init__(element, if_not_exists=if_not_exists)
+        self.columns = [CreateColumn(column) for column in element.columns]
+        self.include_foreign_key_constraints = include_foreign_key_constraints
+
+
+class _DropView(_DropBase):
+    """Semi-public 'DROP VIEW' construct.
+
+    Used by the test suite for dialect-agnostic drops of views.
+    This object will eventually be part of a public "view" API.
+
+    """
+
+    __visit_name__ = "drop_view"
+
+
+class CreateConstraint(BaseDDLElement):
+    def __init__(self, element: Constraint):
+        self.element = element
+
+
+class CreateColumn(BaseDDLElement):
+    """Represent a :class:`_schema.Column`
+    as rendered in a CREATE TABLE statement,
+    via the :class:`.CreateTable` construct.
+
+    This is provided to support custom column DDL within the generation
+    of CREATE TABLE statements, by using the
+    compiler extension documented in :ref:`sqlalchemy.ext.compiler_toplevel`
+    to extend :class:`.CreateColumn`.
+
+    Typical integration is to examine the incoming :class:`_schema.Column`
+    object, and to redirect compilation if a particular flag or condition
+    is found::
+
+        from sqlalchemy import schema
+        from sqlalchemy.ext.compiler import compiles
+
+
+        @compiles(schema.CreateColumn)
+        def compile(element, compiler, **kw):
+            column = element.element
+
+            if "special" not in column.info:
+                return compiler.visit_create_column(element, **kw)
+
+            text = "%s SPECIAL DIRECTIVE %s" % (
+                column.name,
+                compiler.type_compiler.process(column.type),
+            )
+            default = compiler.get_column_default_string(column)
+            if default is not None:
+                text += " DEFAULT " + default
+
+            if not column.nullable:
+                text += " NOT NULL"
+
+            if column.constraints:
+                text += " ".join(
+                    compiler.process(const) for const in column.constraints
+                )
+            return text
+
+    The above construct can be applied to a :class:`_schema.Table`
+    as follows::
+
+        from sqlalchemy import Table, Metadata, Column, Integer, String
+        from sqlalchemy import schema
+
+        metadata = MetaData()
+
+        table = Table(
+            "mytable",
+            MetaData(),
+            Column("x", Integer, info={"special": True}, primary_key=True),
+            Column("y", String(50)),
+            Column("z", String(20), info={"special": True}),
+        )
+
+        metadata.create_all(conn)
+
+    Above, the directives we've added to the :attr:`_schema.Column.info`
+    collection
+    will be detected by our custom compilation scheme:
+
+    .. sourcecode:: sql
+
+        CREATE TABLE mytable (
+                x SPECIAL DIRECTIVE INTEGER NOT NULL,
+                y VARCHAR(50),
+                z SPECIAL DIRECTIVE VARCHAR(20),
+            PRIMARY KEY (x)
+        )
+
+    The :class:`.CreateColumn` construct can also be used to skip certain
+    columns when producing a ``CREATE TABLE``.  This is accomplished by
+    creating a compilation rule that conditionally returns ``None``.
+    This is essentially how to produce the same effect as using the
+    ``system=True`` argument on :class:`_schema.Column`, which marks a column
+    as an implicitly-present "system" column.
+
+    For example, suppose we wish to produce a :class:`_schema.Table`
+    which skips
+    rendering of the PostgreSQL ``xmin`` column against the PostgreSQL
+    backend, but on other backends does render it, in anticipation of a
+    triggered rule.  A conditional compilation rule could skip this name only
+    on PostgreSQL::
+
+        from sqlalchemy.schema import CreateColumn
+
+
+        @compiles(CreateColumn, "postgresql")
+        def skip_xmin(element, compiler, **kw):
+            if element.element.name == "xmin":
+                return None
+            else:
+                return compiler.visit_create_column(element, **kw)
+
+
+        my_table = Table(
+            "mytable",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("xmin", Integer),
+        )
+
+    Above, a :class:`.CreateTable` construct will generate a ``CREATE TABLE``
+    which only includes the ``id`` column in the string; the ``xmin`` column
+    will be omitted, but only against the PostgreSQL backend.
+
+    """
+
+    __visit_name__ = "create_column"
+
+    def __init__(self, element):
+        self.element = element
+
+
+class DropTable(_DropBase):
+    """Represent a DROP TABLE statement."""
+
+    __visit_name__ = "drop_table"
+
+    def __init__(self, element: Table, if_exists: bool = False):
+        """Create a :class:`.DropTable` construct.
+
+        :param element: a :class:`_schema.Table` that's the subject
+         of the DROP.
+        :param on: See the description for 'on' in :class:`.DDL`.
+        :param if_exists: if True, an IF EXISTS operator will be applied to the
+         construct.
+
+         .. versionadded:: 1.4.0b2
+
+        """
+        super().__init__(element, if_exists=if_exists)
+
+
+class CreateSequence(_CreateBase):
+    """Represent a CREATE SEQUENCE statement."""
+
+    __visit_name__ = "create_sequence"
+
+    def __init__(self, element: Sequence, if_not_exists: bool = False):
+        super().__init__(element, if_not_exists=if_not_exists)
+
+
+class DropSequence(_DropBase):
+    """Represent a DROP SEQUENCE statement."""
+
+    __visit_name__ = "drop_sequence"
+
+    def __init__(self, element: Sequence, if_exists: bool = False):
+        super().__init__(element, if_exists=if_exists)
+
+
+class CreateIndex(_CreateBase):
+    """Represent a CREATE INDEX statement."""
+
+    __visit_name__ = "create_index"
+
+    def __init__(self, element, if_not_exists=False):
+        """Create a :class:`.Createindex` construct.
+
+        :param element: a :class:`_schema.Index` that's the subject
+         of the CREATE.
+        :param if_not_exists: if True, an IF NOT EXISTS operator will be
+         applied to the construct.
+
+         .. versionadded:: 1.4.0b2
+
+        """
+        super().__init__(element, if_not_exists=if_not_exists)
+
+
+class DropIndex(_DropBase):
+    """Represent a DROP INDEX statement."""
+
+    __visit_name__ = "drop_index"
+
+    def __init__(self, element, if_exists=False):
+        """Create a :class:`.DropIndex` construct.
+
+        :param element: a :class:`_schema.Index` that's the subject
+         of the DROP.
+        :param if_exists: if True, an IF EXISTS operator will be applied to the
+         construct.
+
+         .. versionadded:: 1.4.0b2
+
+        """
+        super().__init__(element, if_exists=if_exists)
+
+
+class AddConstraint(_CreateBase):
+    """Represent an ALTER TABLE ADD CONSTRAINT statement."""
+
+    __visit_name__ = "add_constraint"
+
+    def __init__(self, element):
+        super().__init__(element)
+        element._create_rule = util.portable_instancemethod(
+            self._create_rule_disable
+        )
+
+
+class DropConstraint(_DropBase):
+    """Represent an ALTER TABLE DROP CONSTRAINT statement."""
+
+    __visit_name__ = "drop_constraint"
+
+    def __init__(self, element, cascade=False, if_exists=False, **kw):
+        self.cascade = cascade
+        super().__init__(element, if_exists=if_exists, **kw)
+        element._create_rule = util.portable_instancemethod(
+            self._create_rule_disable
+        )
+
+
+class SetTableComment(_CreateDropBase):
+    """Represent a COMMENT ON TABLE IS statement."""
+
+    __visit_name__ = "set_table_comment"
+
+
+class DropTableComment(_CreateDropBase):
+    """Represent a COMMENT ON TABLE '' statement.
+
+    Note this varies a lot across database backends.
+
+    """
+
+    __visit_name__ = "drop_table_comment"
+
+
+class SetColumnComment(_CreateDropBase):
+    """Represent a COMMENT ON COLUMN IS statement."""
+
+    __visit_name__ = "set_column_comment"
+
+
+class DropColumnComment(_CreateDropBase):
+    """Represent a COMMENT ON COLUMN IS NULL statement."""
+
+    __visit_name__ = "drop_column_comment"
+
+
+class SetConstraintComment(_CreateDropBase):
+    """Represent a COMMENT ON CONSTRAINT IS statement."""
+
+    __visit_name__ = "set_constraint_comment"
+
+
+class DropConstraintComment(_CreateDropBase):
+    """Represent a COMMENT ON CONSTRAINT IS NULL statement."""
+
+    __visit_name__ = "drop_constraint_comment"
+
+
+class InvokeDDLBase(SchemaVisitor):
+    def __init__(self, connection):
+        self.connection = connection
+
+    @contextlib.contextmanager
+    def with_ddl_events(self, target, **kw):
+        """helper context manager that will apply appropriate DDL events
+        to a CREATE or DROP operation."""
+
+        raise NotImplementedError()
+
+
+class InvokeCreateDDLBase(InvokeDDLBase):
+    @contextlib.contextmanager
+    def with_ddl_events(self, target, **kw):
+        """helper context manager that will apply appropriate DDL events
+        to a CREATE or DROP operation."""
+
+        target.dispatch.before_create(
+            target, self.connection, _ddl_runner=self, **kw
+        )
+        yield
+        target.dispatch.after_create(
+            target, self.connection, _ddl_runner=self, **kw
+        )
+
+
+class InvokeDropDDLBase(InvokeDDLBase):
+    @contextlib.contextmanager
+    def with_ddl_events(self, target, **kw):
+        """helper context manager that will apply appropriate DDL events
+        to a CREATE or DROP operation."""
+
+        target.dispatch.before_drop(
+            target, self.connection, _ddl_runner=self, **kw
+        )
+        yield
+        target.dispatch.after_drop(
+            target, self.connection, _ddl_runner=self, **kw
+        )
+
+
+class SchemaGenerator(InvokeCreateDDLBase):
+    def __init__(
+        self, dialect, connection, checkfirst=False, tables=None, **kwargs
+    ):
+        super().__init__(connection, **kwargs)
+        self.checkfirst = checkfirst
+        self.tables = tables
+        self.preparer = dialect.identifier_preparer
+        self.dialect = dialect
+        self.memo = {}
+
+    def _can_create_table(self, table):
+        self.dialect.validate_identifier(table.name)
+        effective_schema = self.connection.schema_for_object(table)
+        if effective_schema:
+            self.dialect.validate_identifier(effective_schema)
+        return not self.checkfirst or not self.dialect.has_table(
+            self.connection, table.name, schema=effective_schema
+        )
+
+    def _can_create_index(self, index):
+        effective_schema = self.connection.schema_for_object(index.table)
+        if effective_schema:
+            self.dialect.validate_identifier(effective_schema)
+        return not self.checkfirst or not self.dialect.has_index(
+            self.connection,
+            index.table.name,
+            index.name,
+            schema=effective_schema,
+        )
+
+    def _can_create_sequence(self, sequence):
+        effective_schema = self.connection.schema_for_object(sequence)
+
+        return self.dialect.supports_sequences and (
+            (not self.dialect.sequences_optional or not sequence.optional)
+            and (
+                not self.checkfirst
+                or not self.dialect.has_sequence(
+                    self.connection, sequence.name, schema=effective_schema
+                )
+            )
+        )
+
+    def visit_metadata(self, metadata):
+        if self.tables is not None:
+            tables = self.tables
+        else:
+            tables = list(metadata.tables.values())
+
+        collection = sort_tables_and_constraints(
+            [t for t in tables if self._can_create_table(t)]
+        )
+
+        seq_coll = [
+            s
+            for s in metadata._sequences.values()
+            if s.column is None and self._can_create_sequence(s)
+        ]
+
+        event_collection = [t for (t, fks) in collection if t is not None]
+
+        with self.with_ddl_events(
+            metadata,
+            tables=event_collection,
+            checkfirst=self.checkfirst,
+        ):
+            for seq in seq_coll:
+                self.traverse_single(seq, create_ok=True)
+
+            for table, fkcs in collection:
+                if table is not None:
+                    self.traverse_single(
+                        table,
+                        create_ok=True,
+                        include_foreign_key_constraints=fkcs,
+                        _is_metadata_operation=True,
+                    )
+                else:
+                    for fkc in fkcs:
+                        self.traverse_single(fkc)
+
+    def visit_table(
+        self,
+        table,
+        create_ok=False,
+        include_foreign_key_constraints=None,
+        _is_metadata_operation=False,
+    ):
+        if not create_ok and not self._can_create_table(table):
+            return
+
+        with self.with_ddl_events(
+            table,
+            checkfirst=self.checkfirst,
+            _is_metadata_operation=_is_metadata_operation,
+        ):
+            for column in table.columns:
+                if column.default is not None:
+                    self.traverse_single(column.default)
+
+            if not self.dialect.supports_alter:
+                # e.g., don't omit any foreign key constraints
+                include_foreign_key_constraints = None
+
+            CreateTable(
+                table,
+                include_foreign_key_constraints=(
+                    include_foreign_key_constraints
+                ),
+            )._invoke_with(self.connection)
+
+            if hasattr(table, "indexes"):
+                for index in table.indexes:
+                    self.traverse_single(index, create_ok=True)
+
+            if (
+                self.dialect.supports_comments
+                and not self.dialect.inline_comments
+            ):
+                if table.comment is not None:
+                    SetTableComment(table)._invoke_with(self.connection)
+
+                for column in table.columns:
+                    if column.comment is not None:
+                        SetColumnComment(column)._invoke_with(self.connection)
+
+                if self.dialect.supports_constraint_comments:
+                    for constraint in table.constraints:
+                        if constraint.comment is not None:
+                            self.connection.execute(
+                                SetConstraintComment(constraint)
+                            )
+
+    def visit_foreign_key_constraint(self, constraint):
+        if not self.dialect.supports_alter:
+            return
+
+        with self.with_ddl_events(constraint):
+            AddConstraint(constraint)._invoke_with(self.connection)
+
+    def visit_sequence(self, sequence, create_ok=False):
+        if not create_ok and not self._can_create_sequence(sequence):
+            return
+        with self.with_ddl_events(sequence):
+            CreateSequence(sequence)._invoke_with(self.connection)
+
+    def visit_index(self, index, create_ok=False):
+        if not create_ok and not self._can_create_index(index):
+            return
+        with self.with_ddl_events(index):
+            CreateIndex(index)._invoke_with(self.connection)
+
+
+class SchemaDropper(InvokeDropDDLBase):
+    def __init__(
+        self, dialect, connection, checkfirst=False, tables=None, **kwargs
+    ):
+        super().__init__(connection, **kwargs)
+        self.checkfirst = checkfirst
+        self.tables = tables
+        self.preparer = dialect.identifier_preparer
+        self.dialect = dialect
+        self.memo = {}
+
+    def visit_metadata(self, metadata):
+        if self.tables is not None:
+            tables = self.tables
+        else:
+            tables = list(metadata.tables.values())
+
+        try:
+            unsorted_tables = [t for t in tables if self._can_drop_table(t)]
+            collection = list(
+                reversed(
+                    sort_tables_and_constraints(
+                        unsorted_tables,
+                        filter_fn=lambda constraint: (
+                            False
+                            if not self.dialect.supports_alter
+                            or constraint.name is None
+                            else None
+                        ),
+                    )
+                )
+            )
+        except exc.CircularDependencyError as err2:
+            if not self.dialect.supports_alter:
+                util.warn(
+                    "Can't sort tables for DROP; an "
+                    "unresolvable foreign key "
+                    "dependency exists between tables: %s; and backend does "
+                    "not support ALTER.  To restore at least a partial sort, "
+                    "apply use_alter=True to ForeignKey and "
+                    "ForeignKeyConstraint "
+                    "objects involved in the cycle to mark these as known "
+                    "cycles that will be ignored."
+                    % (", ".join(sorted([t.fullname for t in err2.cycles])))
+                )
+                collection = [(t, ()) for t in unsorted_tables]
+            else:
+                raise exc.CircularDependencyError(
+                    err2.args[0],
+                    err2.cycles,
+                    err2.edges,
+                    msg="Can't sort tables for DROP; an "
+                    "unresolvable foreign key "
+                    "dependency exists between tables: %s.  Please ensure "
+                    "that the ForeignKey and ForeignKeyConstraint objects "
+                    "involved in the cycle have "
+                    "names so that they can be dropped using "
+                    "DROP CONSTRAINT."
+                    % (", ".join(sorted([t.fullname for t in err2.cycles]))),
+                ) from err2
+
+        seq_coll = [
+            s
+            for s in metadata._sequences.values()
+            if self._can_drop_sequence(s)
+        ]
+
+        event_collection = [t for (t, fks) in collection if t is not None]
+
+        with self.with_ddl_events(
+            metadata,
+            tables=event_collection,
+            checkfirst=self.checkfirst,
+        ):
+            for table, fkcs in collection:
+                if table is not None:
+                    self.traverse_single(
+                        table,
+                        drop_ok=True,
+                        _is_metadata_operation=True,
+                        _ignore_sequences=seq_coll,
+                    )
+                else:
+                    for fkc in fkcs:
+                        self.traverse_single(fkc)
+
+            for seq in seq_coll:
+                self.traverse_single(seq, drop_ok=seq.column is None)
+
+    def _can_drop_table(self, table):
+        self.dialect.validate_identifier(table.name)
+        effective_schema = self.connection.schema_for_object(table)
+        if effective_schema:
+            self.dialect.validate_identifier(effective_schema)
+        return not self.checkfirst or self.dialect.has_table(
+            self.connection, table.name, schema=effective_schema
+        )
+
+    def _can_drop_index(self, index):
+        effective_schema = self.connection.schema_for_object(index.table)
+        if effective_schema:
+            self.dialect.validate_identifier(effective_schema)
+        return not self.checkfirst or self.dialect.has_index(
+            self.connection,
+            index.table.name,
+            index.name,
+            schema=effective_schema,
+        )
+
+    def _can_drop_sequence(self, sequence):
+        effective_schema = self.connection.schema_for_object(sequence)
+        return self.dialect.supports_sequences and (
+            (not self.dialect.sequences_optional or not sequence.optional)
+            and (
+                not self.checkfirst
+                or self.dialect.has_sequence(
+                    self.connection, sequence.name, schema=effective_schema
+                )
+            )
+        )
+
+    def visit_index(self, index, drop_ok=False):
+        if not drop_ok and not self._can_drop_index(index):
+            return
+
+        with self.with_ddl_events(index):
+            DropIndex(index)(index, self.connection)
+
+    def visit_table(
+        self,
+        table,
+        drop_ok=False,
+        _is_metadata_operation=False,
+        _ignore_sequences=(),
+    ):
+        if not drop_ok and not self._can_drop_table(table):
+            return
+
+        with self.with_ddl_events(
+            table,
+            checkfirst=self.checkfirst,
+            _is_metadata_operation=_is_metadata_operation,
+        ):
+            DropTable(table)._invoke_with(self.connection)
+
+            # traverse client side defaults which may refer to server-side
+            # sequences. noting that some of these client side defaults may
+            # also be set up as server side defaults
+            # (see https://docs.sqlalchemy.org/en/
+            # latest/core/defaults.html
+            # #associating-a-sequence-as-the-server-side-
+            # default), so have to be dropped after the table is dropped.
+            for column in table.columns:
+                if (
+                    column.default is not None
+                    and column.default not in _ignore_sequences
+                ):
+                    self.traverse_single(column.default)
+
+    def visit_foreign_key_constraint(self, constraint):
+        if not self.dialect.supports_alter:
+            return
+        with self.with_ddl_events(constraint):
+            DropConstraint(constraint)._invoke_with(self.connection)
+
+    def visit_sequence(self, sequence, drop_ok=False):
+        if not drop_ok and not self._can_drop_sequence(sequence):
+            return
+        with self.with_ddl_events(sequence):
+            DropSequence(sequence)._invoke_with(self.connection)
+
+
+def sort_tables(
+    tables: Iterable[TableClause],
+    skip_fn: Optional[Callable[[ForeignKeyConstraint], bool]] = None,
+    extra_dependencies: Optional[
+        typing_Sequence[Tuple[TableClause, TableClause]]
+    ] = None,
+) -> List[Table]:
+    """Sort a collection of :class:`_schema.Table` objects based on
+    dependency.
+
+    This is a dependency-ordered sort which will emit :class:`_schema.Table`
+    objects such that they will follow their dependent :class:`_schema.Table`
+    objects.
+    Tables are dependent on another based on the presence of
+    :class:`_schema.ForeignKeyConstraint`
+    objects as well as explicit dependencies
+    added by :meth:`_schema.Table.add_is_dependent_on`.
+
+    .. warning::
+
+        The :func:`._schema.sort_tables` function cannot by itself
+        accommodate automatic resolution of dependency cycles between
+        tables, which are usually caused by mutually dependent foreign key
+        constraints. When these cycles are detected, the foreign keys
+        of these tables are omitted from consideration in the sort.
+        A warning is emitted when this condition occurs, which will be an
+        exception raise in a future release.   Tables which are not part
+        of the cycle will still be returned in dependency order.
+
+        To resolve these cycles, the
+        :paramref:`_schema.ForeignKeyConstraint.use_alter` parameter may be
+        applied to those constraints which create a cycle.  Alternatively,
+        the :func:`_schema.sort_tables_and_constraints` function will
+        automatically return foreign key constraints in a separate
+        collection when cycles are detected so that they may be applied
+        to a schema separately.
+
+        .. versionchanged:: 1.3.17 - a warning is emitted when
+           :func:`_schema.sort_tables` cannot perform a proper sort due to
+           cyclical dependencies.  This will be an exception in a future
+           release.  Additionally, the sort will continue to return
+           other tables not involved in the cycle in dependency order
+           which was not the case previously.
+
+    :param tables: a sequence of :class:`_schema.Table` objects.
+
+    :param skip_fn: optional callable which will be passed a
+     :class:`_schema.ForeignKeyConstraint` object; if it returns True, this
+     constraint will not be considered as a dependency.  Note this is
+     **different** from the same parameter in
+     :func:`.sort_tables_and_constraints`, which is
+     instead passed the owning :class:`_schema.ForeignKeyConstraint` object.
+
+    :param extra_dependencies: a sequence of 2-tuples of tables which will
+     also be considered as dependent on each other.
+
+    .. seealso::
+
+        :func:`.sort_tables_and_constraints`
+
+        :attr:`_schema.MetaData.sorted_tables` - uses this function to sort
+
+
+    """
+
+    if skip_fn is not None:
+        fixed_skip_fn = skip_fn
+
+        def _skip_fn(fkc):
+            for fk in fkc.elements:
+                if fixed_skip_fn(fk):
+                    return True
+            else:
+                return None
+
+    else:
+        _skip_fn = None  # type: ignore
+
+    return [
+        t
+        for (t, fkcs) in sort_tables_and_constraints(
+            tables,
+            filter_fn=_skip_fn,
+            extra_dependencies=extra_dependencies,
+            _warn_for_cycles=True,
+        )
+        if t is not None
+    ]
+
+
+def sort_tables_and_constraints(
+    tables, filter_fn=None, extra_dependencies=None, _warn_for_cycles=False
+):
+    """Sort a collection of :class:`_schema.Table`  /
+    :class:`_schema.ForeignKeyConstraint`
+    objects.
+
+    This is a dependency-ordered sort which will emit tuples of
+    ``(Table, [ForeignKeyConstraint, ...])`` such that each
+    :class:`_schema.Table` follows its dependent :class:`_schema.Table`
+    objects.
+    Remaining :class:`_schema.ForeignKeyConstraint`
+    objects that are separate due to
+    dependency rules not satisfied by the sort are emitted afterwards
+    as ``(None, [ForeignKeyConstraint ...])``.
+
+    Tables are dependent on another based on the presence of
+    :class:`_schema.ForeignKeyConstraint` objects, explicit dependencies
+    added by :meth:`_schema.Table.add_is_dependent_on`,
+    as well as dependencies
+    stated here using the :paramref:`~.sort_tables_and_constraints.skip_fn`
+    and/or :paramref:`~.sort_tables_and_constraints.extra_dependencies`
+    parameters.
+
+    :param tables: a sequence of :class:`_schema.Table` objects.
+
+    :param filter_fn: optional callable which will be passed a
+     :class:`_schema.ForeignKeyConstraint` object,
+     and returns a value based on
+     whether this constraint should definitely be included or excluded as
+     an inline constraint, or neither.   If it returns False, the constraint
+     will definitely be included as a dependency that cannot be subject
+     to ALTER; if True, it will **only** be included as an ALTER result at
+     the end.   Returning None means the constraint is included in the
+     table-based result unless it is detected as part of a dependency cycle.
+
+    :param extra_dependencies: a sequence of 2-tuples of tables which will
+     also be considered as dependent on each other.
+
+    .. seealso::
+
+        :func:`.sort_tables`
+
+
+    """
+
+    fixed_dependencies = set()
+    mutable_dependencies = set()
+
+    if extra_dependencies is not None:
+        fixed_dependencies.update(extra_dependencies)
+
+    remaining_fkcs = set()
+    for table in tables:
+        for fkc in table.foreign_key_constraints:
+            if fkc.use_alter is True:
+                remaining_fkcs.add(fkc)
+                continue
+
+            if filter_fn:
+                filtered = filter_fn(fkc)
+
+                if filtered is True:
+                    remaining_fkcs.add(fkc)
+                    continue
+
+            dependent_on = fkc.referred_table
+            if dependent_on is not table:
+                mutable_dependencies.add((dependent_on, table))
+
+        fixed_dependencies.update(
+            (parent, table) for parent in table._extra_dependencies
+        )
+
+    try:
+        candidate_sort = list(
+            topological.sort(
+                fixed_dependencies.union(mutable_dependencies),
+                tables,
+            )
+        )
+    except exc.CircularDependencyError as err:
+        if _warn_for_cycles:
+            util.warn(
+                "Cannot correctly sort tables; there are unresolvable cycles "
+                'between tables "%s", which is usually caused by mutually '
+                "dependent foreign key constraints.  Foreign key constraints "
+                "involving these tables will not be considered; this warning "
+                "may raise an error in a future release."
+                % (", ".join(sorted(t.fullname for t in err.cycles)),)
+            )
+        for edge in err.edges:
+            if edge in mutable_dependencies:
+                table = edge[1]
+                if table not in err.cycles:
+                    continue
+                can_remove = [
+                    fkc
+                    for fkc in table.foreign_key_constraints
+                    if filter_fn is None or filter_fn(fkc) is not False
+                ]
+                remaining_fkcs.update(can_remove)
+                for fkc in can_remove:
+                    dependent_on = fkc.referred_table
+                    if dependent_on is not table:
+                        mutable_dependencies.discard((dependent_on, table))
+        candidate_sort = list(
+            topological.sort(
+                fixed_dependencies.union(mutable_dependencies),
+                tables,
+            )
+        )
+
+    return [
+        (table, table.foreign_key_constraints.difference(remaining_fkcs))
+        for table in candidate_sort
+    ] + [(None, list(remaining_fkcs))]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/default_comparator.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/default_comparator.py
new file mode 100644
index 00000000..7fa5dafe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/default_comparator.py
@@ -0,0 +1,552 @@
+# sql/default_comparator.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Default implementation of SQL comparison operations.
+"""
+
+from __future__ import annotations
+
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import NoReturn
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import Union
+
+from . import coercions
+from . import operators
+from . import roles
+from . import type_api
+from .elements import and_
+from .elements import BinaryExpression
+from .elements import ClauseElement
+from .elements import CollationClause
+from .elements import CollectionAggregate
+from .elements import ExpressionClauseList
+from .elements import False_
+from .elements import Null
+from .elements import OperatorExpression
+from .elements import or_
+from .elements import True_
+from .elements import UnaryExpression
+from .operators import OperatorType
+from .. import exc
+from .. import util
+
+_T = typing.TypeVar("_T", bound=Any)
+
+if typing.TYPE_CHECKING:
+    from .elements import ColumnElement
+    from .operators import custom_op
+    from .type_api import TypeEngine
+
+
+def _boolean_compare(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    obj: Any,
+    *,
+    negate_op: Optional[OperatorType] = None,
+    reverse: bool = False,
+    _python_is_types: Tuple[Type[Any], ...] = (type(None), bool),
+    result_type: Optional[TypeEngine[bool]] = None,
+    **kwargs: Any,
+) -> OperatorExpression[bool]:
+    if result_type is None:
+        result_type = type_api.BOOLEANTYPE
+
+    if isinstance(obj, _python_is_types + (Null, True_, False_)):
+        # allow x ==/!= True/False to be treated as a literal.
+        # this comes out to "== / != true/false" or "1/0" if those
+        # constants aren't supported and works on all platforms
+        if op in (operators.eq, operators.ne) and isinstance(
+            obj, (bool, True_, False_)
+        ):
+            return OperatorExpression._construct_for_op(
+                expr,
+                coercions.expect(roles.ConstExprRole, obj),
+                op,
+                type_=result_type,
+                negate=negate_op,
+                modifiers=kwargs,
+            )
+        elif op in (
+            operators.is_distinct_from,
+            operators.is_not_distinct_from,
+        ):
+            return OperatorExpression._construct_for_op(
+                expr,
+                coercions.expect(roles.ConstExprRole, obj),
+                op,
+                type_=result_type,
+                negate=negate_op,
+                modifiers=kwargs,
+            )
+        elif expr._is_collection_aggregate:
+            obj = coercions.expect(
+                roles.ConstExprRole, element=obj, operator=op, expr=expr
+            )
+        else:
+            # all other None uses IS, IS NOT
+            if op in (operators.eq, operators.is_):
+                return OperatorExpression._construct_for_op(
+                    expr,
+                    coercions.expect(roles.ConstExprRole, obj),
+                    operators.is_,
+                    negate=operators.is_not,
+                    type_=result_type,
+                )
+            elif op in (operators.ne, operators.is_not):
+                return OperatorExpression._construct_for_op(
+                    expr,
+                    coercions.expect(roles.ConstExprRole, obj),
+                    operators.is_not,
+                    negate=operators.is_,
+                    type_=result_type,
+                )
+            else:
+                raise exc.ArgumentError(
+                    "Only '=', '!=', 'is_()', 'is_not()', "
+                    "'is_distinct_from()', 'is_not_distinct_from()' "
+                    "operators can be used with None/True/False"
+                )
+    else:
+        obj = coercions.expect(
+            roles.BinaryElementRole, element=obj, operator=op, expr=expr
+        )
+
+    if reverse:
+        return OperatorExpression._construct_for_op(
+            obj,
+            expr,
+            op,
+            type_=result_type,
+            negate=negate_op,
+            modifiers=kwargs,
+        )
+    else:
+        return OperatorExpression._construct_for_op(
+            expr,
+            obj,
+            op,
+            type_=result_type,
+            negate=negate_op,
+            modifiers=kwargs,
+        )
+
+
+def _custom_op_operate(
+    expr: ColumnElement[Any],
+    op: custom_op[Any],
+    obj: Any,
+    reverse: bool = False,
+    result_type: Optional[TypeEngine[Any]] = None,
+    **kw: Any,
+) -> ColumnElement[Any]:
+    if result_type is None:
+        if op.return_type:
+            result_type = op.return_type
+        elif op.is_comparison:
+            result_type = type_api.BOOLEANTYPE
+
+    return _binary_operate(
+        expr, op, obj, reverse=reverse, result_type=result_type, **kw
+    )
+
+
+def _binary_operate(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    obj: roles.BinaryElementRole[Any],
+    *,
+    reverse: bool = False,
+    result_type: Optional[TypeEngine[_T]] = None,
+    **kw: Any,
+) -> OperatorExpression[_T]:
+    coerced_obj = coercions.expect(
+        roles.BinaryElementRole, obj, expr=expr, operator=op
+    )
+
+    if reverse:
+        left, right = coerced_obj, expr
+    else:
+        left, right = expr, coerced_obj
+
+    if result_type is None:
+        op, result_type = left.comparator._adapt_expression(
+            op, right.comparator
+        )
+
+    return OperatorExpression._construct_for_op(
+        left, right, op, type_=result_type, modifiers=kw
+    )
+
+
+def _conjunction_operate(
+    expr: ColumnElement[Any], op: OperatorType, other: Any, **kw: Any
+) -> ColumnElement[Any]:
+    if op is operators.and_:
+        return and_(expr, other)
+    elif op is operators.or_:
+        return or_(expr, other)
+    else:
+        raise NotImplementedError()
+
+
+def _scalar(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    fn: Callable[[ColumnElement[Any]], ColumnElement[Any]],
+    **kw: Any,
+) -> ColumnElement[Any]:
+    return fn(expr)
+
+
+def _in_impl(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    seq_or_selectable: ClauseElement,
+    negate_op: OperatorType,
+    **kw: Any,
+) -> ColumnElement[Any]:
+    seq_or_selectable = coercions.expect(
+        roles.InElementRole, seq_or_selectable, expr=expr, operator=op
+    )
+    if "in_ops" in seq_or_selectable._annotations:
+        op, negate_op = seq_or_selectable._annotations["in_ops"]
+
+    return _boolean_compare(
+        expr, op, seq_or_selectable, negate_op=negate_op, **kw
+    )
+
+
+def _getitem_impl(
+    expr: ColumnElement[Any], op: OperatorType, other: Any, **kw: Any
+) -> ColumnElement[Any]:
+    if (
+        isinstance(expr.type, type_api.INDEXABLE)
+        or isinstance(expr.type, type_api.TypeDecorator)
+        and isinstance(expr.type.impl_instance, type_api.INDEXABLE)
+    ):
+        other = coercions.expect(
+            roles.BinaryElementRole, other, expr=expr, operator=op
+        )
+        return _binary_operate(expr, op, other, **kw)
+    else:
+        _unsupported_impl(expr, op, other, **kw)
+
+
+def _unsupported_impl(
+    expr: ColumnElement[Any], op: OperatorType, *arg: Any, **kw: Any
+) -> NoReturn:
+    raise NotImplementedError(
+        "Operator '%s' is not supported on this expression" % op.__name__
+    )
+
+
+def _inv_impl(
+    expr: ColumnElement[Any], op: OperatorType, **kw: Any
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.__inv__`."""
+
+    # undocumented element currently used by the ORM for
+    # relationship.contains()
+    if hasattr(expr, "negation_clause"):
+        return expr.negation_clause
+    else:
+        return expr._negate()
+
+
+def _neg_impl(
+    expr: ColumnElement[Any], op: OperatorType, **kw: Any
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.__neg__`."""
+    return UnaryExpression(expr, operator=operators.neg, type_=expr.type)
+
+
+def _bitwise_not_impl(
+    expr: ColumnElement[Any], op: OperatorType, **kw: Any
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.bitwise_not`."""
+
+    return UnaryExpression(
+        expr, operator=operators.bitwise_not_op, type_=expr.type
+    )
+
+
+def _match_impl(
+    expr: ColumnElement[Any], op: OperatorType, other: Any, **kw: Any
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.match`."""
+
+    return _boolean_compare(
+        expr,
+        operators.match_op,
+        coercions.expect(
+            roles.BinaryElementRole,
+            other,
+            expr=expr,
+            operator=operators.match_op,
+        ),
+        result_type=type_api.MATCHTYPE,
+        negate_op=(
+            operators.not_match_op
+            if op is operators.match_op
+            else operators.match_op
+        ),
+        **kw,
+    )
+
+
+def _distinct_impl(
+    expr: ColumnElement[Any], op: OperatorType, **kw: Any
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.distinct`."""
+    return UnaryExpression(
+        expr, operator=operators.distinct_op, type_=expr.type
+    )
+
+
+def _between_impl(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    cleft: Any,
+    cright: Any,
+    **kw: Any,
+) -> ColumnElement[Any]:
+    """See :meth:`.ColumnOperators.between`."""
+    return BinaryExpression(
+        expr,
+        ExpressionClauseList._construct_for_list(
+            operators.and_,
+            type_api.NULLTYPE,
+            coercions.expect(
+                roles.BinaryElementRole,
+                cleft,
+                expr=expr,
+                operator=operators.and_,
+            ),
+            coercions.expect(
+                roles.BinaryElementRole,
+                cright,
+                expr=expr,
+                operator=operators.and_,
+            ),
+            group=False,
+        ),
+        op,
+        negate=(
+            operators.not_between_op
+            if op is operators.between_op
+            else operators.between_op
+        ),
+        modifiers=kw,
+    )
+
+
+def _collate_impl(
+    expr: ColumnElement[str], op: OperatorType, collation: str, **kw: Any
+) -> ColumnElement[str]:
+    return CollationClause._create_collation_expression(expr, collation)
+
+
+def _regexp_match_impl(
+    expr: ColumnElement[str],
+    op: OperatorType,
+    pattern: Any,
+    flags: Optional[str],
+    **kw: Any,
+) -> ColumnElement[Any]:
+    return BinaryExpression(
+        expr,
+        coercions.expect(
+            roles.BinaryElementRole,
+            pattern,
+            expr=expr,
+            operator=operators.comma_op,
+        ),
+        op,
+        negate=operators.not_regexp_match_op,
+        modifiers={"flags": flags},
+    )
+
+
+def _regexp_replace_impl(
+    expr: ColumnElement[Any],
+    op: OperatorType,
+    pattern: Any,
+    replacement: Any,
+    flags: Optional[str],
+    **kw: Any,
+) -> ColumnElement[Any]:
+    return BinaryExpression(
+        expr,
+        ExpressionClauseList._construct_for_list(
+            operators.comma_op,
+            type_api.NULLTYPE,
+            coercions.expect(
+                roles.BinaryElementRole,
+                pattern,
+                expr=expr,
+                operator=operators.comma_op,
+            ),
+            coercions.expect(
+                roles.BinaryElementRole,
+                replacement,
+                expr=expr,
+                operator=operators.comma_op,
+            ),
+            group=False,
+        ),
+        op,
+        modifiers={"flags": flags},
+    )
+
+
+# a mapping of operators with the method they use, along with
+# additional keyword arguments to be passed
+operator_lookup: Dict[
+    str,
+    Tuple[
+        Callable[..., ColumnElement[Any]],
+        util.immutabledict[
+            str, Union[OperatorType, Callable[..., ColumnElement[Any]]]
+        ],
+    ],
+] = {
+    "and_": (_conjunction_operate, util.EMPTY_DICT),
+    "or_": (_conjunction_operate, util.EMPTY_DICT),
+    "inv": (_inv_impl, util.EMPTY_DICT),
+    "add": (_binary_operate, util.EMPTY_DICT),
+    "mul": (_binary_operate, util.EMPTY_DICT),
+    "sub": (_binary_operate, util.EMPTY_DICT),
+    "div": (_binary_operate, util.EMPTY_DICT),
+    "mod": (_binary_operate, util.EMPTY_DICT),
+    "bitwise_xor_op": (_binary_operate, util.EMPTY_DICT),
+    "bitwise_or_op": (_binary_operate, util.EMPTY_DICT),
+    "bitwise_and_op": (_binary_operate, util.EMPTY_DICT),
+    "bitwise_not_op": (_bitwise_not_impl, util.EMPTY_DICT),
+    "bitwise_lshift_op": (_binary_operate, util.EMPTY_DICT),
+    "bitwise_rshift_op": (_binary_operate, util.EMPTY_DICT),
+    "truediv": (_binary_operate, util.EMPTY_DICT),
+    "floordiv": (_binary_operate, util.EMPTY_DICT),
+    "custom_op": (_custom_op_operate, util.EMPTY_DICT),
+    "json_path_getitem_op": (_binary_operate, util.EMPTY_DICT),
+    "json_getitem_op": (_binary_operate, util.EMPTY_DICT),
+    "concat_op": (_binary_operate, util.EMPTY_DICT),
+    "any_op": (
+        _scalar,
+        util.immutabledict({"fn": CollectionAggregate._create_any}),
+    ),
+    "all_op": (
+        _scalar,
+        util.immutabledict({"fn": CollectionAggregate._create_all}),
+    ),
+    "lt": (_boolean_compare, util.immutabledict({"negate_op": operators.ge})),
+    "le": (_boolean_compare, util.immutabledict({"negate_op": operators.gt})),
+    "ne": (_boolean_compare, util.immutabledict({"negate_op": operators.eq})),
+    "gt": (_boolean_compare, util.immutabledict({"negate_op": operators.le})),
+    "ge": (_boolean_compare, util.immutabledict({"negate_op": operators.lt})),
+    "eq": (_boolean_compare, util.immutabledict({"negate_op": operators.ne})),
+    "is_distinct_from": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.is_not_distinct_from}),
+    ),
+    "is_not_distinct_from": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.is_distinct_from}),
+    ),
+    "like_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_like_op}),
+    ),
+    "ilike_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_ilike_op}),
+    ),
+    "not_like_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.like_op}),
+    ),
+    "not_ilike_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.ilike_op}),
+    ),
+    "contains_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_contains_op}),
+    ),
+    "icontains_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_icontains_op}),
+    ),
+    "startswith_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_startswith_op}),
+    ),
+    "istartswith_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_istartswith_op}),
+    ),
+    "endswith_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_endswith_op}),
+    ),
+    "iendswith_op": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.not_iendswith_op}),
+    ),
+    "desc_op": (
+        _scalar,
+        util.immutabledict({"fn": UnaryExpression._create_desc}),
+    ),
+    "asc_op": (
+        _scalar,
+        util.immutabledict({"fn": UnaryExpression._create_asc}),
+    ),
+    "nulls_first_op": (
+        _scalar,
+        util.immutabledict({"fn": UnaryExpression._create_nulls_first}),
+    ),
+    "nulls_last_op": (
+        _scalar,
+        util.immutabledict({"fn": UnaryExpression._create_nulls_last}),
+    ),
+    "in_op": (
+        _in_impl,
+        util.immutabledict({"negate_op": operators.not_in_op}),
+    ),
+    "not_in_op": (
+        _in_impl,
+        util.immutabledict({"negate_op": operators.in_op}),
+    ),
+    "is_": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.is_}),
+    ),
+    "is_not": (
+        _boolean_compare,
+        util.immutabledict({"negate_op": operators.is_not}),
+    ),
+    "collate": (_collate_impl, util.EMPTY_DICT),
+    "match_op": (_match_impl, util.EMPTY_DICT),
+    "not_match_op": (_match_impl, util.EMPTY_DICT),
+    "distinct_op": (_distinct_impl, util.EMPTY_DICT),
+    "between_op": (_between_impl, util.EMPTY_DICT),
+    "not_between_op": (_between_impl, util.EMPTY_DICT),
+    "neg": (_neg_impl, util.EMPTY_DICT),
+    "getitem": (_getitem_impl, util.EMPTY_DICT),
+    "lshift": (_unsupported_impl, util.EMPTY_DICT),
+    "rshift": (_unsupported_impl, util.EMPTY_DICT),
+    "contains": (_unsupported_impl, util.EMPTY_DICT),
+    "regexp_match_op": (_regexp_match_impl, util.EMPTY_DICT),
+    "not_regexp_match_op": (_regexp_match_impl, util.EMPTY_DICT),
+    "regexp_replace_op": (_regexp_replace_impl, util.EMPTY_DICT),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/dml.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/dml.py
new file mode 100644
index 00000000..f0e6edbb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/dml.py
@@ -0,0 +1,1827 @@
+# sql/dml.py
+# Copyright (C) 2009-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+"""
+Provide :class:`_expression.Insert`, :class:`_expression.Update` and
+:class:`_expression.Delete`.
+
+"""
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import operator
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import MutableMapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import roles
+from . import util as sql_util
+from ._typing import _TP
+from ._typing import _unexpected_kw
+from ._typing import is_column_element
+from ._typing import is_named_from_clause
+from .base import _entity_namespace_key
+from .base import _exclusive_against
+from .base import _from_objects
+from .base import _generative
+from .base import _select_iterables
+from .base import ColumnCollection
+from .base import ColumnSet
+from .base import CompileState
+from .base import DialectKWArgs
+from .base import Executable
+from .base import Generative
+from .base import HasCompileState
+from .elements import BooleanClauseList
+from .elements import ClauseElement
+from .elements import ColumnClause
+from .elements import ColumnElement
+from .elements import Null
+from .selectable import Alias
+from .selectable import ExecutableReturnsRows
+from .selectable import FromClause
+from .selectable import HasCTE
+from .selectable import HasPrefixes
+from .selectable import Join
+from .selectable import SelectLabelStyle
+from .selectable import TableClause
+from .selectable import TypedReturnsRows
+from .sqltypes import NullType
+from .visitors import InternalTraversal
+from .. import exc
+from .. import util
+from ..util.typing import Self
+from ..util.typing import TypeGuard
+
+if TYPE_CHECKING:
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnsClauseArgument
+    from ._typing import _DMLColumnArgument
+    from ._typing import _DMLColumnKeyMapping
+    from ._typing import _DMLTableArgument
+    from ._typing import _T0  # noqa
+    from ._typing import _T1  # noqa
+    from ._typing import _T2  # noqa
+    from ._typing import _T3  # noqa
+    from ._typing import _T4  # noqa
+    from ._typing import _T5  # noqa
+    from ._typing import _T6  # noqa
+    from ._typing import _T7  # noqa
+    from ._typing import _TypedColumnClauseArgument as _TCCA  # noqa
+    from .base import ReadOnlyColumnCollection
+    from .compiler import SQLCompiler
+    from .elements import KeyedColumnElement
+    from .selectable import _ColumnsClauseElement
+    from .selectable import _SelectIterable
+    from .selectable import Select
+    from .selectable import Selectable
+
+    def isupdate(dml: DMLState) -> TypeGuard[UpdateDMLState]: ...
+
+    def isdelete(dml: DMLState) -> TypeGuard[DeleteDMLState]: ...
+
+    def isinsert(dml: DMLState) -> TypeGuard[InsertDMLState]: ...
+
+else:
+    isupdate = operator.attrgetter("isupdate")
+    isdelete = operator.attrgetter("isdelete")
+    isinsert = operator.attrgetter("isinsert")
+
+
+_T = TypeVar("_T", bound=Any)
+
+_DMLColumnElement = Union[str, ColumnClause[Any]]
+_DMLTableElement = Union[TableClause, Alias, Join]
+
+
+class DMLState(CompileState):
+    _no_parameters = True
+    _dict_parameters: Optional[MutableMapping[_DMLColumnElement, Any]] = None
+    _multi_parameters: Optional[
+        List[MutableMapping[_DMLColumnElement, Any]]
+    ] = None
+    _ordered_values: Optional[List[Tuple[_DMLColumnElement, Any]]] = None
+    _parameter_ordering: Optional[List[_DMLColumnElement]] = None
+    _primary_table: FromClause
+    _supports_implicit_returning = True
+
+    isupdate = False
+    isdelete = False
+    isinsert = False
+
+    statement: UpdateBase
+
+    def __init__(
+        self, statement: UpdateBase, compiler: SQLCompiler, **kw: Any
+    ):
+        raise NotImplementedError()
+
+    @classmethod
+    def get_entity_description(cls, statement: UpdateBase) -> Dict[str, Any]:
+        return {
+            "name": (
+                statement.table.name
+                if is_named_from_clause(statement.table)
+                else None
+            ),
+            "table": statement.table,
+        }
+
+    @classmethod
+    def get_returning_column_descriptions(
+        cls, statement: UpdateBase
+    ) -> List[Dict[str, Any]]:
+        return [
+            {
+                "name": c.key,
+                "type": c.type,
+                "expr": c,
+            }
+            for c in statement._all_selected_columns
+        ]
+
+    @property
+    def dml_table(self) -> _DMLTableElement:
+        return self.statement.table
+
+    if TYPE_CHECKING:
+
+        @classmethod
+        def get_plugin_class(cls, statement: Executable) -> Type[DMLState]: ...
+
+    @classmethod
+    def _get_multi_crud_kv_pairs(
+        cls,
+        statement: UpdateBase,
+        multi_kv_iterator: Iterable[Dict[_DMLColumnArgument, Any]],
+    ) -> List[Dict[_DMLColumnElement, Any]]:
+        return [
+            {
+                coercions.expect(roles.DMLColumnRole, k): v
+                for k, v in mapping.items()
+            }
+            for mapping in multi_kv_iterator
+        ]
+
+    @classmethod
+    def _get_crud_kv_pairs(
+        cls,
+        statement: UpdateBase,
+        kv_iterator: Iterable[Tuple[_DMLColumnArgument, Any]],
+        needs_to_be_cacheable: bool,
+    ) -> List[Tuple[_DMLColumnElement, Any]]:
+        return [
+            (
+                coercions.expect(roles.DMLColumnRole, k),
+                (
+                    v
+                    if not needs_to_be_cacheable
+                    else coercions.expect(
+                        roles.ExpressionElementRole,
+                        v,
+                        type_=NullType(),
+                        is_crud=True,
+                    )
+                ),
+            )
+            for k, v in kv_iterator
+        ]
+
+    def _make_extra_froms(
+        self, statement: DMLWhereBase
+    ) -> Tuple[FromClause, List[FromClause]]:
+        froms: List[FromClause] = []
+
+        all_tables = list(sql_util.tables_from_leftmost(statement.table))
+        primary_table = all_tables[0]
+        seen = {primary_table}
+
+        consider = statement._where_criteria
+        if self._dict_parameters:
+            consider += tuple(self._dict_parameters.values())
+
+        for crit in consider:
+            for item in _from_objects(crit):
+                if not seen.intersection(item._cloned_set):
+                    froms.append(item)
+                seen.update(item._cloned_set)
+
+        froms.extend(all_tables[1:])
+        return primary_table, froms
+
+    def _process_values(self, statement: ValuesBase) -> None:
+        if self._no_parameters:
+            self._dict_parameters = statement._values
+            self._no_parameters = False
+
+    def _process_select_values(self, statement: ValuesBase) -> None:
+        assert statement._select_names is not None
+        parameters: MutableMapping[_DMLColumnElement, Any] = {
+            name: Null() for name in statement._select_names
+        }
+
+        if self._no_parameters:
+            self._no_parameters = False
+            self._dict_parameters = parameters
+        else:
+            # this condition normally not reachable as the Insert
+            # does not allow this construction to occur
+            assert False, "This statement already has parameters"
+
+    def _no_multi_values_supported(self, statement: ValuesBase) -> NoReturn:
+        raise exc.InvalidRequestError(
+            "%s construct does not support "
+            "multiple parameter sets." % statement.__visit_name__.upper()
+        )
+
+    def _cant_mix_formats_error(self) -> NoReturn:
+        raise exc.InvalidRequestError(
+            "Can't mix single and multiple VALUES "
+            "formats in one INSERT statement; one style appends to a "
+            "list while the other replaces values, so the intent is "
+            "ambiguous."
+        )
+
+
+@CompileState.plugin_for("default", "insert")
+class InsertDMLState(DMLState):
+    isinsert = True
+
+    include_table_with_column_exprs = False
+
+    _has_multi_parameters = False
+
+    def __init__(
+        self,
+        statement: Insert,
+        compiler: SQLCompiler,
+        disable_implicit_returning: bool = False,
+        **kw: Any,
+    ):
+        self.statement = statement
+        self._primary_table = statement.table
+
+        if disable_implicit_returning:
+            self._supports_implicit_returning = False
+
+        self.isinsert = True
+        if statement._select_names:
+            self._process_select_values(statement)
+        if statement._values is not None:
+            self._process_values(statement)
+        if statement._multi_values:
+            self._process_multi_values(statement)
+
+    @util.memoized_property
+    def _insert_col_keys(self) -> List[str]:
+        # this is also done in crud.py -> _key_getters_for_crud_column
+        return [
+            coercions.expect(roles.DMLColumnRole, col, as_key=True)
+            for col in self._dict_parameters or ()
+        ]
+
+    def _process_values(self, statement: ValuesBase) -> None:
+        if self._no_parameters:
+            self._has_multi_parameters = False
+            self._dict_parameters = statement._values
+            self._no_parameters = False
+        elif self._has_multi_parameters:
+            self._cant_mix_formats_error()
+
+    def _process_multi_values(self, statement: ValuesBase) -> None:
+        for parameters in statement._multi_values:
+            multi_parameters: List[MutableMapping[_DMLColumnElement, Any]] = [
+                (
+                    {
+                        c.key: value
+                        for c, value in zip(statement.table.c, parameter_set)
+                    }
+                    if isinstance(parameter_set, collections_abc.Sequence)
+                    else parameter_set
+                )
+                for parameter_set in parameters
+            ]
+
+            if self._no_parameters:
+                self._no_parameters = False
+                self._has_multi_parameters = True
+                self._multi_parameters = multi_parameters
+                self._dict_parameters = self._multi_parameters[0]
+            elif not self._has_multi_parameters:
+                self._cant_mix_formats_error()
+            else:
+                assert self._multi_parameters
+                self._multi_parameters.extend(multi_parameters)
+
+
+@CompileState.plugin_for("default", "update")
+class UpdateDMLState(DMLState):
+    isupdate = True
+
+    include_table_with_column_exprs = False
+
+    def __init__(self, statement: Update, compiler: SQLCompiler, **kw: Any):
+        self.statement = statement
+
+        self.isupdate = True
+        if statement._ordered_values is not None:
+            self._process_ordered_values(statement)
+        elif statement._values is not None:
+            self._process_values(statement)
+        elif statement._multi_values:
+            self._no_multi_values_supported(statement)
+        t, ef = self._make_extra_froms(statement)
+        self._primary_table = t
+        self._extra_froms = ef
+
+        self.is_multitable = mt = ef
+        self.include_table_with_column_exprs = bool(
+            mt and compiler.render_table_with_column_in_update_from
+        )
+
+    def _process_ordered_values(self, statement: ValuesBase) -> None:
+        parameters = statement._ordered_values
+
+        if self._no_parameters:
+            self._no_parameters = False
+            assert parameters is not None
+            self._dict_parameters = dict(parameters)
+            self._ordered_values = parameters
+            self._parameter_ordering = [key for key, value in parameters]
+        else:
+            raise exc.InvalidRequestError(
+                "Can only invoke ordered_values() once, and not mixed "
+                "with any other values() call"
+            )
+
+
+@CompileState.plugin_for("default", "delete")
+class DeleteDMLState(DMLState):
+    isdelete = True
+
+    def __init__(self, statement: Delete, compiler: SQLCompiler, **kw: Any):
+        self.statement = statement
+
+        self.isdelete = True
+        t, ef = self._make_extra_froms(statement)
+        self._primary_table = t
+        self._extra_froms = ef
+        self.is_multitable = ef
+
+
+class UpdateBase(
+    roles.DMLRole,
+    HasCTE,
+    HasCompileState,
+    DialectKWArgs,
+    HasPrefixes,
+    Generative,
+    ExecutableReturnsRows,
+    ClauseElement,
+):
+    """Form the base for ``INSERT``, ``UPDATE``, and ``DELETE`` statements."""
+
+    __visit_name__ = "update_base"
+
+    _hints: util.immutabledict[Tuple[_DMLTableElement, str], str] = (
+        util.EMPTY_DICT
+    )
+    named_with_column = False
+
+    _label_style: SelectLabelStyle = (
+        SelectLabelStyle.LABEL_STYLE_DISAMBIGUATE_ONLY
+    )
+    table: _DMLTableElement
+
+    _return_defaults = False
+    _return_defaults_columns: Optional[Tuple[_ColumnsClauseElement, ...]] = (
+        None
+    )
+    _supplemental_returning: Optional[Tuple[_ColumnsClauseElement, ...]] = None
+    _returning: Tuple[_ColumnsClauseElement, ...] = ()
+
+    is_dml = True
+
+    def _generate_fromclause_column_proxies(
+        self,
+        fromclause: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        columns._populate_separate_keys(
+            col._make_proxy(
+                fromclause, primary_key=primary_key, foreign_keys=foreign_keys
+            )
+            for col in self._all_selected_columns
+            if is_column_element(col)
+        )
+
+    def params(self, *arg: Any, **kw: Any) -> NoReturn:
+        """Set the parameters for the statement.
+
+        This method raises ``NotImplementedError`` on the base class,
+        and is overridden by :class:`.ValuesBase` to provide the
+        SET/VALUES clause of UPDATE and INSERT.
+
+        """
+        raise NotImplementedError(
+            "params() is not supported for INSERT/UPDATE/DELETE statements."
+            " To set the values for an INSERT or UPDATE statement, use"
+            " stmt.values(**parameters)."
+        )
+
+    @_generative
+    def with_dialect_options(self, **opt: Any) -> Self:
+        """Add dialect options to this INSERT/UPDATE/DELETE object.
+
+        e.g.::
+
+            upd = table.update().dialect_options(mysql_limit=10)
+
+        .. versionadded: 1.4 - this method supersedes the dialect options
+           associated with the constructor.
+
+
+        """
+        self._validate_dialect_kwargs(opt)
+        return self
+
+    @_generative
+    def return_defaults(
+        self,
+        *cols: _DMLColumnArgument,
+        supplemental_cols: Optional[Iterable[_DMLColumnArgument]] = None,
+        sort_by_parameter_order: bool = False,
+    ) -> Self:
+        """Make use of a :term:`RETURNING` clause for the purpose
+        of fetching server-side expressions and defaults, for supporting
+        backends only.
+
+        .. deepalchemy::
+
+            The :meth:`.UpdateBase.return_defaults` method is used by the ORM
+            for its internal work in fetching newly generated primary key
+            and server default values, in particular to provide the underyling
+            implementation of the :paramref:`_orm.Mapper.eager_defaults`
+            ORM feature as well as to allow RETURNING support with bulk
+            ORM inserts.  Its behavior is fairly idiosyncratic
+            and is not really intended for general use.  End users should
+            stick with using :meth:`.UpdateBase.returning` in order to
+            add RETURNING clauses to their INSERT, UPDATE and DELETE
+            statements.
+
+        Normally, a single row INSERT statement will automatically populate the
+        :attr:`.CursorResult.inserted_primary_key` attribute when executed,
+        which stores the primary key of the row that was just inserted in the
+        form of a :class:`.Row` object with column names as named tuple keys
+        (and the :attr:`.Row._mapping` view fully populated as well). The
+        dialect in use chooses the strategy to use in order to populate this
+        data; if it was generated using server-side defaults and / or SQL
+        expressions, dialect-specific approaches such as ``cursor.lastrowid``
+        or ``RETURNING`` are typically used to acquire the new primary key
+        value.
+
+        However, when the statement is modified by calling
+        :meth:`.UpdateBase.return_defaults` before executing the statement,
+        additional behaviors take place **only** for backends that support
+        RETURNING and for :class:`.Table` objects that maintain the
+        :paramref:`.Table.implicit_returning` parameter at its default value of
+        ``True``. In these cases, when the :class:`.CursorResult` is returned
+        from the statement's execution, not only will
+        :attr:`.CursorResult.inserted_primary_key` be populated as always, the
+        :attr:`.CursorResult.returned_defaults` attribute will also be
+        populated with a :class:`.Row` named-tuple representing the full range
+        of server generated
+        values from that single row, including values for any columns that
+        specify :paramref:`_schema.Column.server_default` or which make use of
+        :paramref:`_schema.Column.default` using a SQL expression.
+
+        When invoking INSERT statements with multiple rows using
+        :ref:`insertmanyvalues <engine_insertmanyvalues>`, the
+        :meth:`.UpdateBase.return_defaults` modifier will have the effect of
+        the :attr:`_engine.CursorResult.inserted_primary_key_rows` and
+        :attr:`_engine.CursorResult.returned_defaults_rows` attributes being
+        fully populated with lists of :class:`.Row` objects representing newly
+        inserted primary key values as well as newly inserted server generated
+        values for each row inserted. The
+        :attr:`.CursorResult.inserted_primary_key` and
+        :attr:`.CursorResult.returned_defaults` attributes will also continue
+        to be populated with the first row of these two collections.
+
+        If the backend does not support RETURNING or the :class:`.Table` in use
+        has disabled :paramref:`.Table.implicit_returning`, then no RETURNING
+        clause is added and no additional data is fetched, however the
+        INSERT, UPDATE or DELETE statement proceeds normally.
+
+        E.g.::
+
+            stmt = table.insert().values(data="newdata").return_defaults()
+
+            result = connection.execute(stmt)
+
+            server_created_at = result.returned_defaults["created_at"]
+
+        When used against an UPDATE statement
+        :meth:`.UpdateBase.return_defaults` instead looks for columns that
+        include :paramref:`_schema.Column.onupdate` or
+        :paramref:`_schema.Column.server_onupdate` parameters assigned, when
+        constructing the columns that will be included in the RETURNING clause
+        by default if explicit columns were not specified. When used against a
+        DELETE statement, no columns are included in RETURNING by default, they
+        instead must be specified explicitly as there are no columns that
+        normally change values when a DELETE statement proceeds.
+
+        .. versionadded:: 2.0  :meth:`.UpdateBase.return_defaults` is supported
+           for DELETE statements also and has been moved from
+           :class:`.ValuesBase` to :class:`.UpdateBase`.
+
+        The :meth:`.UpdateBase.return_defaults` method is mutually exclusive
+        against the :meth:`.UpdateBase.returning` method and errors will be
+        raised during the SQL compilation process if both are used at the same
+        time on one statement. The RETURNING clause of the INSERT, UPDATE or
+        DELETE statement is therefore controlled by only one of these methods
+        at a time.
+
+        The :meth:`.UpdateBase.return_defaults` method differs from
+        :meth:`.UpdateBase.returning` in these ways:
+
+        1. :meth:`.UpdateBase.return_defaults` method causes the
+           :attr:`.CursorResult.returned_defaults` collection to be populated
+           with the first row from the RETURNING result. This attribute is not
+           populated when using :meth:`.UpdateBase.returning`.
+
+        2. :meth:`.UpdateBase.return_defaults` is compatible with existing
+           logic used to fetch auto-generated primary key values that are then
+           populated into the :attr:`.CursorResult.inserted_primary_key`
+           attribute. By contrast, using :meth:`.UpdateBase.returning` will
+           have the effect of the :attr:`.CursorResult.inserted_primary_key`
+           attribute being left unpopulated.
+
+        3. :meth:`.UpdateBase.return_defaults` can be called against any
+           backend. Backends that don't support RETURNING will skip the usage
+           of the feature, rather than raising an exception, *unless*
+           ``supplemental_cols`` is passed. The return value
+           of :attr:`_engine.CursorResult.returned_defaults` will be ``None``
+           for backends that don't support RETURNING or for which the target
+           :class:`.Table` sets :paramref:`.Table.implicit_returning` to
+           ``False``.
+
+        4. An INSERT statement invoked with executemany() is supported if the
+           backend database driver supports the
+           :ref:`insertmanyvalues <engine_insertmanyvalues>`
+           feature which is now supported by most SQLAlchemy-included backends.
+           When executemany is used, the
+           :attr:`_engine.CursorResult.returned_defaults_rows` and
+           :attr:`_engine.CursorResult.inserted_primary_key_rows` accessors
+           will return the inserted defaults and primary keys.
+
+           .. versionadded:: 1.4 Added
+              :attr:`_engine.CursorResult.returned_defaults_rows` and
+              :attr:`_engine.CursorResult.inserted_primary_key_rows` accessors.
+              In version 2.0, the underlying implementation which fetches and
+              populates the data for these attributes was generalized to be
+              supported by most backends, whereas in 1.4 they were only
+              supported by the ``psycopg2`` driver.
+
+
+        :param cols: optional list of column key names or
+         :class:`_schema.Column` that acts as a filter for those columns that
+         will be fetched.
+        :param supplemental_cols: optional list of RETURNING expressions,
+          in the same form as one would pass to the
+          :meth:`.UpdateBase.returning` method. When present, the additional
+          columns will be included in the RETURNING clause, and the
+          :class:`.CursorResult` object will be "rewound" when returned, so
+          that methods like :meth:`.CursorResult.all` will return new rows
+          mostly as though the statement used :meth:`.UpdateBase.returning`
+          directly. However, unlike when using :meth:`.UpdateBase.returning`
+          directly, the **order of the columns is undefined**, so can only be
+          targeted using names or :attr:`.Row._mapping` keys; they cannot
+          reliably be targeted positionally.
+
+          .. versionadded:: 2.0
+
+        :param sort_by_parameter_order: for a batch INSERT that is being
+         executed against multiple parameter sets, organize the results of
+         RETURNING so that the returned rows correspond to the order of
+         parameter sets passed in.  This applies only to an :term:`executemany`
+         execution for supporting dialects and typically makes use of the
+         :term:`insertmanyvalues` feature.
+
+         .. versionadded:: 2.0.10
+
+         .. seealso::
+
+            :ref:`engine_insertmanyvalues_returning_order` - background on
+            sorting of RETURNING rows for bulk INSERT
+
+        .. seealso::
+
+            :meth:`.UpdateBase.returning`
+
+            :attr:`_engine.CursorResult.returned_defaults`
+
+            :attr:`_engine.CursorResult.returned_defaults_rows`
+
+            :attr:`_engine.CursorResult.inserted_primary_key`
+
+            :attr:`_engine.CursorResult.inserted_primary_key_rows`
+
+        """
+
+        if self._return_defaults:
+            # note _return_defaults_columns = () means return all columns,
+            # so if we have been here before, only update collection if there
+            # are columns in the collection
+            if self._return_defaults_columns and cols:
+                self._return_defaults_columns = tuple(
+                    util.OrderedSet(self._return_defaults_columns).union(
+                        coercions.expect(roles.ColumnsClauseRole, c)
+                        for c in cols
+                    )
+                )
+            else:
+                # set for all columns
+                self._return_defaults_columns = ()
+        else:
+            self._return_defaults_columns = tuple(
+                coercions.expect(roles.ColumnsClauseRole, c) for c in cols
+            )
+        self._return_defaults = True
+        if sort_by_parameter_order:
+            if not self.is_insert:
+                raise exc.ArgumentError(
+                    "The 'sort_by_parameter_order' argument to "
+                    "return_defaults() only applies to INSERT statements"
+                )
+            self._sort_by_parameter_order = True
+        if supplemental_cols:
+            # uniquifying while also maintaining order (the maintain of order
+            # is for test suites but also for vertical splicing
+            supplemental_col_tup = (
+                coercions.expect(roles.ColumnsClauseRole, c)
+                for c in supplemental_cols
+            )
+
+            if self._supplemental_returning is None:
+                self._supplemental_returning = tuple(
+                    util.unique_list(supplemental_col_tup)
+                )
+            else:
+                self._supplemental_returning = tuple(
+                    util.unique_list(
+                        self._supplemental_returning
+                        + tuple(supplemental_col_tup)
+                    )
+                )
+
+        return self
+
+    @_generative
+    def returning(
+        self,
+        *cols: _ColumnsClauseArgument[Any],
+        sort_by_parameter_order: bool = False,
+        **__kw: Any,
+    ) -> UpdateBase:
+        r"""Add a :term:`RETURNING` or equivalent clause to this statement.
+
+        e.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> stmt = (
+            ...     table.update()
+            ...     .where(table.c.data == "value")
+            ...     .values(status="X")
+            ...     .returning(table.c.server_flag, table.c.updated_timestamp)
+            ... )
+            >>> print(stmt)
+            {printsql}UPDATE some_table SET status=:status
+            WHERE some_table.data = :data_1
+            RETURNING some_table.server_flag, some_table.updated_timestamp
+
+        The method may be invoked multiple times to add new entries to the
+        list of expressions to be returned.
+
+        .. versionadded:: 1.4.0b2 The method may be invoked multiple times to
+         add new entries to the list of expressions to be returned.
+
+        The given collection of column expressions should be derived from the
+        table that is the target of the INSERT, UPDATE, or DELETE.  While
+        :class:`_schema.Column` objects are typical, the elements can also be
+        expressions:
+
+        .. sourcecode:: pycon+sql
+
+            >>> stmt = table.insert().returning(
+            ...     (table.c.first_name + " " + table.c.last_name).label("fullname")
+            ... )
+            >>> print(stmt)
+            {printsql}INSERT INTO some_table (first_name, last_name)
+            VALUES (:first_name, :last_name)
+            RETURNING some_table.first_name || :first_name_1 || some_table.last_name AS fullname
+
+        Upon compilation, a RETURNING clause, or database equivalent,
+        will be rendered within the statement.   For INSERT and UPDATE,
+        the values are the newly inserted/updated values.  For DELETE,
+        the values are those of the rows which were deleted.
+
+        Upon execution, the values of the columns to be returned are made
+        available via the result set and can be iterated using
+        :meth:`_engine.CursorResult.fetchone` and similar.
+        For DBAPIs which do not
+        natively support returning values (i.e. cx_oracle), SQLAlchemy will
+        approximate this behavior at the result level so that a reasonable
+        amount of behavioral neutrality is provided.
+
+        Note that not all databases/DBAPIs
+        support RETURNING.   For those backends with no support,
+        an exception is raised upon compilation and/or execution.
+        For those who do support it, the functionality across backends
+        varies greatly, including restrictions on executemany()
+        and other statements which return multiple rows. Please
+        read the documentation notes for the database in use in
+        order to determine the availability of RETURNING.
+
+        :param \*cols: series of columns, SQL expressions, or whole tables
+         entities to be returned.
+        :param sort_by_parameter_order: for a batch INSERT that is being
+         executed against multiple parameter sets, organize the results of
+         RETURNING so that the returned rows correspond to the order of
+         parameter sets passed in.  This applies only to an :term:`executemany`
+         execution for supporting dialects and typically makes use of the
+         :term:`insertmanyvalues` feature.
+
+         .. versionadded:: 2.0.10
+
+         .. seealso::
+
+            :ref:`engine_insertmanyvalues_returning_order` - background on
+            sorting of RETURNING rows for bulk INSERT (Core level discussion)
+
+            :ref:`orm_queryguide_bulk_insert_returning_ordered` - example of
+            use with :ref:`orm_queryguide_bulk_insert` (ORM level discussion)
+
+        .. seealso::
+
+          :meth:`.UpdateBase.return_defaults` - an alternative method tailored
+          towards efficient fetching of server-side defaults and triggers
+          for single-row INSERTs or UPDATEs.
+
+          :ref:`tutorial_insert_returning` - in the :ref:`unified_tutorial`
+
+        """  # noqa: E501
+        if __kw:
+            raise _unexpected_kw("UpdateBase.returning()", __kw)
+        if self._return_defaults:
+            raise exc.InvalidRequestError(
+                "return_defaults() is already configured on this statement"
+            )
+        self._returning += tuple(
+            coercions.expect(roles.ColumnsClauseRole, c) for c in cols
+        )
+        if sort_by_parameter_order:
+            if not self.is_insert:
+                raise exc.ArgumentError(
+                    "The 'sort_by_parameter_order' argument to returning() "
+                    "only applies to INSERT statements"
+                )
+            self._sort_by_parameter_order = True
+        return self
+
+    def corresponding_column(
+        self, column: KeyedColumnElement[Any], require_embedded: bool = False
+    ) -> Optional[ColumnElement[Any]]:
+        return self.exported_columns.corresponding_column(
+            column, require_embedded=require_embedded
+        )
+
+    @util.ro_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        return [c for c in _select_iterables(self._returning)]
+
+    @util.ro_memoized_property
+    def exported_columns(
+        self,
+    ) -> ReadOnlyColumnCollection[Optional[str], ColumnElement[Any]]:
+        """Return the RETURNING columns as a column collection for this
+        statement.
+
+        .. versionadded:: 1.4
+
+        """
+        return ColumnCollection(
+            (c.key, c)
+            for c in self._all_selected_columns
+            if is_column_element(c)
+        ).as_readonly()
+
+    @_generative
+    def with_hint(
+        self,
+        text: str,
+        selectable: Optional[_DMLTableArgument] = None,
+        dialect_name: str = "*",
+    ) -> Self:
+        """Add a table hint for a single table to this
+        INSERT/UPDATE/DELETE statement.
+
+        .. note::
+
+         :meth:`.UpdateBase.with_hint` currently applies only to
+         Microsoft SQL Server.  For MySQL INSERT/UPDATE/DELETE hints, use
+         :meth:`.UpdateBase.prefix_with`.
+
+        The text of the hint is rendered in the appropriate
+        location for the database backend in use, relative
+        to the :class:`_schema.Table` that is the subject of this
+        statement, or optionally to that of the given
+        :class:`_schema.Table` passed as the ``selectable`` argument.
+
+        The ``dialect_name`` option will limit the rendering of a particular
+        hint to a particular backend. Such as, to add a hint
+        that only takes effect for SQL Server::
+
+            mytable.insert().with_hint("WITH (PAGLOCK)", dialect_name="mssql")
+
+        :param text: Text of the hint.
+        :param selectable: optional :class:`_schema.Table` that specifies
+         an element of the FROM clause within an UPDATE or DELETE
+         to be the subject of the hint - applies only to certain backends.
+        :param dialect_name: defaults to ``*``, if specified as the name
+         of a particular dialect, will apply these hints only when
+         that dialect is in use.
+        """
+        if selectable is None:
+            selectable = self.table
+        else:
+            selectable = coercions.expect(roles.DMLTableRole, selectable)
+        self._hints = self._hints.union({(selectable, dialect_name): text})
+        return self
+
+    @property
+    def entity_description(self) -> Dict[str, Any]:
+        """Return a :term:`plugin-enabled` description of the table and/or
+        entity which this DML construct is operating against.
+
+        This attribute is generally useful when using the ORM, as an
+        extended structure which includes information about mapped
+        entities is returned.  The section :ref:`queryguide_inspection`
+        contains more background.
+
+        For a Core statement, the structure returned by this accessor
+        is derived from the :attr:`.UpdateBase.table` attribute, and
+        refers to the :class:`.Table` being inserted, updated, or deleted::
+
+            >>> stmt = insert(user_table)
+            >>> stmt.entity_description
+            {
+                "name": "user_table",
+                "table": Table("user_table", ...)
+            }
+
+        .. versionadded:: 1.4.33
+
+        .. seealso::
+
+            :attr:`.UpdateBase.returning_column_descriptions`
+
+            :attr:`.Select.column_descriptions` - entity information for
+            a :func:`.select` construct
+
+            :ref:`queryguide_inspection` - ORM background
+
+        """
+        meth = DMLState.get_plugin_class(self).get_entity_description
+        return meth(self)
+
+    @property
+    def returning_column_descriptions(self) -> List[Dict[str, Any]]:
+        """Return a :term:`plugin-enabled` description of the columns
+        which this DML construct is RETURNING against, in other words
+        the expressions established as part of :meth:`.UpdateBase.returning`.
+
+        This attribute is generally useful when using the ORM, as an
+        extended structure which includes information about mapped
+        entities is returned.  The section :ref:`queryguide_inspection`
+        contains more background.
+
+        For a Core statement, the structure returned by this accessor is
+        derived from the same objects that are returned by the
+        :attr:`.UpdateBase.exported_columns` accessor::
+
+            >>> stmt = insert(user_table).returning(user_table.c.id, user_table.c.name)
+            >>> stmt.entity_description
+            [
+                {
+                    "name": "id",
+                    "type": Integer,
+                    "expr": Column("id", Integer(), table=<user>, ...)
+                },
+                {
+                    "name": "name",
+                    "type": String(),
+                    "expr": Column("name", String(), table=<user>, ...)
+                },
+            ]
+
+        .. versionadded:: 1.4.33
+
+        .. seealso::
+
+            :attr:`.UpdateBase.entity_description`
+
+            :attr:`.Select.column_descriptions` - entity information for
+            a :func:`.select` construct
+
+            :ref:`queryguide_inspection` - ORM background
+
+        """  # noqa: E501
+        meth = DMLState.get_plugin_class(
+            self
+        ).get_returning_column_descriptions
+        return meth(self)
+
+
+class ValuesBase(UpdateBase):
+    """Supplies support for :meth:`.ValuesBase.values` to
+    INSERT and UPDATE constructs."""
+
+    __visit_name__ = "values_base"
+
+    _supports_multi_parameters = False
+
+    select: Optional[Select[Any]] = None
+    """SELECT statement for INSERT .. FROM SELECT"""
+
+    _post_values_clause: Optional[ClauseElement] = None
+    """used by extensions to Insert etc. to add additional syntacitcal
+    constructs, e.g. ON CONFLICT etc."""
+
+    _values: Optional[util.immutabledict[_DMLColumnElement, Any]] = None
+    _multi_values: Tuple[
+        Union[
+            Sequence[Dict[_DMLColumnElement, Any]],
+            Sequence[Sequence[Any]],
+        ],
+        ...,
+    ] = ()
+
+    _ordered_values: Optional[List[Tuple[_DMLColumnElement, Any]]] = None
+
+    _select_names: Optional[List[str]] = None
+    _inline: bool = False
+
+    def __init__(self, table: _DMLTableArgument):
+        self.table = coercions.expect(
+            roles.DMLTableRole, table, apply_propagate_attrs=self
+        )
+
+    @_generative
+    @_exclusive_against(
+        "_select_names",
+        "_ordered_values",
+        msgs={
+            "_select_names": "This construct already inserts from a SELECT",
+            "_ordered_values": "This statement already has ordered "
+            "values present",
+        },
+    )
+    def values(
+        self,
+        *args: Union[
+            _DMLColumnKeyMapping[Any],
+            Sequence[Any],
+        ],
+        **kwargs: Any,
+    ) -> Self:
+        r"""Specify a fixed VALUES clause for an INSERT statement, or the SET
+        clause for an UPDATE.
+
+        Note that the :class:`_expression.Insert` and
+        :class:`_expression.Update`
+        constructs support
+        per-execution time formatting of the VALUES and/or SET clauses,
+        based on the arguments passed to :meth:`_engine.Connection.execute`.
+        However, the :meth:`.ValuesBase.values` method can be used to "fix" a
+        particular set of parameters into the statement.
+
+        Multiple calls to :meth:`.ValuesBase.values` will produce a new
+        construct, each one with the parameter list modified to include
+        the new parameters sent.  In the typical case of a single
+        dictionary of parameters, the newly passed keys will replace
+        the same keys in the previous construct.  In the case of a list-based
+        "multiple values" construct, each new list of values is extended
+        onto the existing list of values.
+
+        :param \**kwargs: key value pairs representing the string key
+          of a :class:`_schema.Column`
+          mapped to the value to be rendered into the
+          VALUES or SET clause::
+
+                users.insert().values(name="some name")
+
+                users.update().where(users.c.id == 5).values(name="some name")
+
+        :param \*args: As an alternative to passing key/value parameters,
+         a dictionary, tuple, or list of dictionaries or tuples can be passed
+         as a single positional argument in order to form the VALUES or
+         SET clause of the statement.  The forms that are accepted vary
+         based on whether this is an :class:`_expression.Insert` or an
+         :class:`_expression.Update` construct.
+
+         For either an :class:`_expression.Insert` or
+         :class:`_expression.Update`
+         construct, a single dictionary can be passed, which works the same as
+         that of the kwargs form::
+
+            users.insert().values({"name": "some name"})
+
+            users.update().values({"name": "some new name"})
+
+         Also for either form but more typically for the
+         :class:`_expression.Insert` construct, a tuple that contains an
+         entry for every column in the table is also accepted::
+
+            users.insert().values((5, "some name"))
+
+         The :class:`_expression.Insert` construct also supports being
+         passed a list of dictionaries or full-table-tuples, which on the
+         server will render the less common SQL syntax of "multiple values" -
+         this syntax is supported on backends such as SQLite, PostgreSQL,
+         MySQL, but not necessarily others::
+
+            users.insert().values(
+                [
+                    {"name": "some name"},
+                    {"name": "some other name"},
+                    {"name": "yet another name"},
+                ]
+            )
+
+         The above form would render a multiple VALUES statement similar to:
+
+         .. sourcecode:: sql
+
+                INSERT INTO users (name) VALUES
+                                (:name_1),
+                                (:name_2),
+                                (:name_3)
+
+         It is essential to note that **passing multiple values is
+         NOT the same as using traditional executemany() form**.  The above
+         syntax is a **special** syntax not typically used.  To emit an
+         INSERT statement against multiple rows, the normal method is
+         to pass a multiple values list to the
+         :meth:`_engine.Connection.execute`
+         method, which is supported by all database backends and is generally
+         more efficient for a very large number of parameters.
+
+           .. seealso::
+
+               :ref:`tutorial_multiple_parameters` - an introduction to
+               the traditional Core method of multiple parameter set
+               invocation for INSERTs and other statements.
+
+          The UPDATE construct also supports rendering the SET parameters
+          in a specific order.  For this feature refer to the
+          :meth:`_expression.Update.ordered_values` method.
+
+           .. seealso::
+
+              :meth:`_expression.Update.ordered_values`
+
+
+        """
+        if args:
+            # positional case.  this is currently expensive.   we don't
+            # yet have positional-only args so we have to check the length.
+            # then we need to check multiparams vs. single dictionary.
+            # since the parameter format is needed in order to determine
+            # a cache key, we need to determine this up front.
+            arg = args[0]
+
+            if kwargs:
+                raise exc.ArgumentError(
+                    "Can't pass positional and kwargs to values() "
+                    "simultaneously"
+                )
+            elif len(args) > 1:
+                raise exc.ArgumentError(
+                    "Only a single dictionary/tuple or list of "
+                    "dictionaries/tuples is accepted positionally."
+                )
+
+            elif isinstance(arg, collections_abc.Sequence):
+                if arg and isinstance(arg[0], dict):
+                    multi_kv_generator = DMLState.get_plugin_class(
+                        self
+                    )._get_multi_crud_kv_pairs
+                    self._multi_values += (multi_kv_generator(self, arg),)
+                    return self
+
+                if arg and isinstance(arg[0], (list, tuple)):
+                    self._multi_values += (arg,)
+                    return self
+
+                if TYPE_CHECKING:
+                    # crud.py raises during compilation if this is not the
+                    # case
+                    assert isinstance(self, Insert)
+
+                # tuple values
+                arg = {c.key: value for c, value in zip(self.table.c, arg)}
+
+        else:
+            # kwarg path.  this is the most common path for non-multi-params
+            # so this is fairly quick.
+            arg = cast("Dict[_DMLColumnArgument, Any]", kwargs)
+            if args:
+                raise exc.ArgumentError(
+                    "Only a single dictionary/tuple or list of "
+                    "dictionaries/tuples is accepted positionally."
+                )
+
+        # for top level values(), convert literals to anonymous bound
+        # parameters at statement construction time, so that these values can
+        # participate in the cache key process like any other ClauseElement.
+        # crud.py now intercepts bound parameters with unique=True from here
+        # and ensures they get the "crud"-style name when rendered.
+
+        kv_generator = DMLState.get_plugin_class(self)._get_crud_kv_pairs
+        coerced_arg = dict(kv_generator(self, arg.items(), True))
+        if self._values:
+            self._values = self._values.union(coerced_arg)
+        else:
+            self._values = util.immutabledict(coerced_arg)
+        return self
+
+
+class Insert(ValuesBase):
+    """Represent an INSERT construct.
+
+    The :class:`_expression.Insert` object is created using the
+    :func:`_expression.insert()` function.
+
+    """
+
+    __visit_name__ = "insert"
+
+    _supports_multi_parameters = True
+
+    select = None
+    include_insert_from_select_defaults = False
+
+    _sort_by_parameter_order: bool = False
+
+    is_insert = True
+
+    table: TableClause
+
+    _traverse_internals = (
+        [
+            ("table", InternalTraversal.dp_clauseelement),
+            ("_inline", InternalTraversal.dp_boolean),
+            ("_select_names", InternalTraversal.dp_string_list),
+            ("_values", InternalTraversal.dp_dml_values),
+            ("_multi_values", InternalTraversal.dp_dml_multi_values),
+            ("select", InternalTraversal.dp_clauseelement),
+            ("_post_values_clause", InternalTraversal.dp_clauseelement),
+            ("_returning", InternalTraversal.dp_clauseelement_tuple),
+            ("_hints", InternalTraversal.dp_table_hint_list),
+            ("_return_defaults", InternalTraversal.dp_boolean),
+            (
+                "_return_defaults_columns",
+                InternalTraversal.dp_clauseelement_tuple,
+            ),
+            ("_sort_by_parameter_order", InternalTraversal.dp_boolean),
+        ]
+        + HasPrefixes._has_prefixes_traverse_internals
+        + DialectKWArgs._dialect_kwargs_traverse_internals
+        + Executable._executable_traverse_internals
+        + HasCTE._has_ctes_traverse_internals
+    )
+
+    def __init__(self, table: _DMLTableArgument):
+        super().__init__(table)
+
+    @_generative
+    def inline(self) -> Self:
+        """Make this :class:`_expression.Insert` construct "inline" .
+
+        When set, no attempt will be made to retrieve the
+        SQL-generated default values to be provided within the statement;
+        in particular,
+        this allows SQL expressions to be rendered 'inline' within the
+        statement without the need to pre-execute them beforehand; for
+        backends that support "returning", this turns off the "implicit
+        returning" feature for the statement.
+
+
+        .. versionchanged:: 1.4 the :paramref:`_expression.Insert.inline`
+           parameter
+           is now superseded by the :meth:`_expression.Insert.inline` method.
+
+        """
+        self._inline = True
+        return self
+
+    @_generative
+    def from_select(
+        self,
+        names: Sequence[_DMLColumnArgument],
+        select: Selectable,
+        include_defaults: bool = True,
+    ) -> Self:
+        """Return a new :class:`_expression.Insert` construct which represents
+        an ``INSERT...FROM SELECT`` statement.
+
+        e.g.::
+
+            sel = select(table1.c.a, table1.c.b).where(table1.c.c > 5)
+            ins = table2.insert().from_select(["a", "b"], sel)
+
+        :param names: a sequence of string column names or
+         :class:`_schema.Column`
+         objects representing the target columns.
+        :param select: a :func:`_expression.select` construct,
+         :class:`_expression.FromClause`
+         or other construct which resolves into a
+         :class:`_expression.FromClause`,
+         such as an ORM :class:`_query.Query` object, etc.  The order of
+         columns returned from this FROM clause should correspond to the
+         order of columns sent as the ``names`` parameter;  while this
+         is not checked before passing along to the database, the database
+         would normally raise an exception if these column lists don't
+         correspond.
+        :param include_defaults: if True, non-server default values and
+         SQL expressions as specified on :class:`_schema.Column` objects
+         (as documented in :ref:`metadata_defaults_toplevel`) not
+         otherwise specified in the list of names will be rendered
+         into the INSERT and SELECT statements, so that these values are also
+         included in the data to be inserted.
+
+         .. note:: A Python-side default that uses a Python callable function
+            will only be invoked **once** for the whole statement, and **not
+            per row**.
+
+        """
+
+        if self._values:
+            raise exc.InvalidRequestError(
+                "This construct already inserts value expressions"
+            )
+
+        self._select_names = [
+            coercions.expect(roles.DMLColumnRole, name, as_key=True)
+            for name in names
+        ]
+        self._inline = True
+        self.include_insert_from_select_defaults = include_defaults
+        self.select = coercions.expect(roles.DMLSelectRole, select)
+        return self
+
+    if TYPE_CHECKING:
+        # START OVERLOADED FUNCTIONS self.returning ReturningInsert 1-8 ", *, sort_by_parameter_order: bool = False"  # noqa: E501
+
+        # code within this block is **programmatically,
+        # statically generated** by tools/generate_tuple_map_overloads.py
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0], *, sort_by_parameter_order: bool = False
+        ) -> ReturningInsert[Tuple[_T0]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1, _T2]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+            __ent7: _TCCA[_T7],
+            *,
+            sort_by_parameter_order: bool = False,
+        ) -> ReturningInsert[
+            Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]
+        ]: ...
+
+        # END OVERLOADED FUNCTIONS self.returning
+
+        @overload
+        def returning(
+            self,
+            *cols: _ColumnsClauseArgument[Any],
+            sort_by_parameter_order: bool = False,
+            **__kw: Any,
+        ) -> ReturningInsert[Any]: ...
+
+        def returning(
+            self,
+            *cols: _ColumnsClauseArgument[Any],
+            sort_by_parameter_order: bool = False,
+            **__kw: Any,
+        ) -> ReturningInsert[Any]: ...
+
+
+class ReturningInsert(Insert, TypedReturnsRows[_TP]):
+    """Typing-only class that establishes a generic type form of
+    :class:`.Insert` which tracks returned column types.
+
+    This datatype is delivered when calling the
+    :meth:`.Insert.returning` method.
+
+    .. versionadded:: 2.0
+
+    """
+
+
+class DMLWhereBase:
+    table: _DMLTableElement
+    _where_criteria: Tuple[ColumnElement[Any], ...] = ()
+
+    @_generative
+    def where(self, *whereclause: _ColumnExpressionArgument[bool]) -> Self:
+        """Return a new construct with the given expression(s) added to
+        its WHERE clause, joined to the existing clause via AND, if any.
+
+        Both :meth:`_dml.Update.where` and :meth:`_dml.Delete.where`
+        support multiple-table forms, including database-specific
+        ``UPDATE...FROM`` as well as ``DELETE..USING``.  For backends that
+        don't have multiple-table support, a backend agnostic approach
+        to using multiple tables is to make use of correlated subqueries.
+        See the linked tutorial sections below for examples.
+
+        .. seealso::
+
+            :ref:`tutorial_correlated_updates`
+
+            :ref:`tutorial_update_from`
+
+            :ref:`tutorial_multi_table_deletes`
+
+        """
+
+        for criterion in whereclause:
+            where_criteria: ColumnElement[Any] = coercions.expect(
+                roles.WhereHavingRole, criterion, apply_propagate_attrs=self
+            )
+            self._where_criteria += (where_criteria,)
+        return self
+
+    def filter(self, *criteria: roles.ExpressionElementRole[Any]) -> Self:
+        """A synonym for the :meth:`_dml.DMLWhereBase.where` method.
+
+        .. versionadded:: 1.4
+
+        """
+
+        return self.where(*criteria)
+
+    def _filter_by_zero(self) -> _DMLTableElement:
+        return self.table
+
+    def filter_by(self, **kwargs: Any) -> Self:
+        r"""apply the given filtering criterion as a WHERE clause
+        to this select.
+
+        """
+        from_entity = self._filter_by_zero()
+
+        clauses = [
+            _entity_namespace_key(from_entity, key) == value
+            for key, value in kwargs.items()
+        ]
+        return self.filter(*clauses)
+
+    @property
+    def whereclause(self) -> Optional[ColumnElement[Any]]:
+        """Return the completed WHERE clause for this :class:`.DMLWhereBase`
+        statement.
+
+        This assembles the current collection of WHERE criteria
+        into a single :class:`_expression.BooleanClauseList` construct.
+
+
+        .. versionadded:: 1.4
+
+        """
+
+        return BooleanClauseList._construct_for_whereclause(
+            self._where_criteria
+        )
+
+
+class Update(DMLWhereBase, ValuesBase):
+    """Represent an Update construct.
+
+    The :class:`_expression.Update` object is created using the
+    :func:`_expression.update()` function.
+
+    """
+
+    __visit_name__ = "update"
+
+    is_update = True
+
+    _traverse_internals = (
+        [
+            ("table", InternalTraversal.dp_clauseelement),
+            ("_where_criteria", InternalTraversal.dp_clauseelement_tuple),
+            ("_inline", InternalTraversal.dp_boolean),
+            ("_ordered_values", InternalTraversal.dp_dml_ordered_values),
+            ("_values", InternalTraversal.dp_dml_values),
+            ("_returning", InternalTraversal.dp_clauseelement_tuple),
+            ("_hints", InternalTraversal.dp_table_hint_list),
+            ("_return_defaults", InternalTraversal.dp_boolean),
+            (
+                "_return_defaults_columns",
+                InternalTraversal.dp_clauseelement_tuple,
+            ),
+        ]
+        + HasPrefixes._has_prefixes_traverse_internals
+        + DialectKWArgs._dialect_kwargs_traverse_internals
+        + Executable._executable_traverse_internals
+        + HasCTE._has_ctes_traverse_internals
+    )
+
+    def __init__(self, table: _DMLTableArgument):
+        super().__init__(table)
+
+    @_generative
+    def ordered_values(self, *args: Tuple[_DMLColumnArgument, Any]) -> Self:
+        """Specify the VALUES clause of this UPDATE statement with an explicit
+        parameter ordering that will be maintained in the SET clause of the
+        resulting UPDATE statement.
+
+        E.g.::
+
+            stmt = table.update().ordered_values(("name", "ed"), ("ident", "foo"))
+
+        .. seealso::
+
+           :ref:`tutorial_parameter_ordered_updates` - full example of the
+           :meth:`_expression.Update.ordered_values` method.
+
+        .. versionchanged:: 1.4 The :meth:`_expression.Update.ordered_values`
+           method
+           supersedes the
+           :paramref:`_expression.update.preserve_parameter_order`
+           parameter, which will be removed in SQLAlchemy 2.0.
+
+        """  # noqa: E501
+        if self._values:
+            raise exc.ArgumentError(
+                "This statement already has values present"
+            )
+        elif self._ordered_values:
+            raise exc.ArgumentError(
+                "This statement already has ordered values present"
+            )
+
+        kv_generator = DMLState.get_plugin_class(self)._get_crud_kv_pairs
+        self._ordered_values = kv_generator(self, args, True)
+        return self
+
+    @_generative
+    def inline(self) -> Self:
+        """Make this :class:`_expression.Update` construct "inline" .
+
+        When set, SQL defaults present on :class:`_schema.Column`
+        objects via the
+        ``default`` keyword will be compiled 'inline' into the statement and
+        not pre-executed.  This means that their values will not be available
+        in the dictionary returned from
+        :meth:`_engine.CursorResult.last_updated_params`.
+
+        .. versionchanged:: 1.4 the :paramref:`_expression.update.inline`
+           parameter
+           is now superseded by the :meth:`_expression.Update.inline` method.
+
+        """
+        self._inline = True
+        return self
+
+    if TYPE_CHECKING:
+        # START OVERLOADED FUNCTIONS self.returning ReturningUpdate 1-8
+
+        # code within this block is **programmatically,
+        # statically generated** by tools/generate_tuple_map_overloads.py
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0]
+        ) -> ReturningUpdate[Tuple[_T0]]: ...
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+        ) -> ReturningUpdate[Tuple[_T0, _T1]]: ...
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+        ) -> ReturningUpdate[Tuple[_T0, _T1, _T2]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+        ) -> ReturningUpdate[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+        ) -> ReturningUpdate[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+        ) -> ReturningUpdate[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+        ) -> ReturningUpdate[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+            __ent7: _TCCA[_T7],
+        ) -> ReturningUpdate[
+            Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]
+        ]: ...
+
+        # END OVERLOADED FUNCTIONS self.returning
+
+        @overload
+        def returning(
+            self, *cols: _ColumnsClauseArgument[Any], **__kw: Any
+        ) -> ReturningUpdate[Any]: ...
+
+        def returning(
+            self, *cols: _ColumnsClauseArgument[Any], **__kw: Any
+        ) -> ReturningUpdate[Any]: ...
+
+
+class ReturningUpdate(Update, TypedReturnsRows[_TP]):
+    """Typing-only class that establishes a generic type form of
+    :class:`.Update` which tracks returned column types.
+
+    This datatype is delivered when calling the
+    :meth:`.Update.returning` method.
+
+    .. versionadded:: 2.0
+
+    """
+
+
+class Delete(DMLWhereBase, UpdateBase):
+    """Represent a DELETE construct.
+
+    The :class:`_expression.Delete` object is created using the
+    :func:`_expression.delete()` function.
+
+    """
+
+    __visit_name__ = "delete"
+
+    is_delete = True
+
+    _traverse_internals = (
+        [
+            ("table", InternalTraversal.dp_clauseelement),
+            ("_where_criteria", InternalTraversal.dp_clauseelement_tuple),
+            ("_returning", InternalTraversal.dp_clauseelement_tuple),
+            ("_hints", InternalTraversal.dp_table_hint_list),
+        ]
+        + HasPrefixes._has_prefixes_traverse_internals
+        + DialectKWArgs._dialect_kwargs_traverse_internals
+        + Executable._executable_traverse_internals
+        + HasCTE._has_ctes_traverse_internals
+    )
+
+    def __init__(self, table: _DMLTableArgument):
+        self.table = coercions.expect(
+            roles.DMLTableRole, table, apply_propagate_attrs=self
+        )
+
+    if TYPE_CHECKING:
+        # START OVERLOADED FUNCTIONS self.returning ReturningDelete 1-8
+
+        # code within this block is **programmatically,
+        # statically generated** by tools/generate_tuple_map_overloads.py
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0]
+        ) -> ReturningDelete[Tuple[_T0]]: ...
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1]
+        ) -> ReturningDelete[Tuple[_T0, _T1]]: ...
+
+        @overload
+        def returning(
+            self, __ent0: _TCCA[_T0], __ent1: _TCCA[_T1], __ent2: _TCCA[_T2]
+        ) -> ReturningDelete[Tuple[_T0, _T1, _T2]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+        ) -> ReturningDelete[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+        ) -> ReturningDelete[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+        ) -> ReturningDelete[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+        ) -> ReturningDelete[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+        @overload
+        def returning(
+            self,
+            __ent0: _TCCA[_T0],
+            __ent1: _TCCA[_T1],
+            __ent2: _TCCA[_T2],
+            __ent3: _TCCA[_T3],
+            __ent4: _TCCA[_T4],
+            __ent5: _TCCA[_T5],
+            __ent6: _TCCA[_T6],
+            __ent7: _TCCA[_T7],
+        ) -> ReturningDelete[
+            Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]
+        ]: ...
+
+        # END OVERLOADED FUNCTIONS self.returning
+
+        @overload
+        def returning(
+            self, *cols: _ColumnsClauseArgument[Any], **__kw: Any
+        ) -> ReturningDelete[Any]: ...
+
+        def returning(
+            self, *cols: _ColumnsClauseArgument[Any], **__kw: Any
+        ) -> ReturningDelete[Any]: ...
+
+
+class ReturningDelete(Update, TypedReturnsRows[_TP]):
+    """Typing-only class that establishes a generic type form of
+    :class:`.Delete` which tracks returned column types.
+
+    This datatype is delivered when calling the
+    :meth:`.Delete.returning` method.
+
+    .. versionadded:: 2.0
+
+    """
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/elements.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/elements.py
new file mode 100644
index 00000000..6f20d7ef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/elements.py
@@ -0,0 +1,5526 @@
+# sql/elements.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Core SQL expression elements, including :class:`_expression.ClauseElement`,
+:class:`_expression.ColumnElement`, and derived classes.
+
+"""
+
+from __future__ import annotations
+
+from decimal import Decimal
+from enum import Enum
+import itertools
+import operator
+import re
+import typing
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple as typing_Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import operators
+from . import roles
+from . import traversals
+from . import type_api
+from ._typing import has_schema_attr
+from ._typing import is_named_from_clause
+from ._typing import is_quoted_name
+from ._typing import is_tuple_type
+from .annotation import Annotated
+from .annotation import SupportsWrappingAnnotations
+from .base import _clone
+from .base import _expand_cloned
+from .base import _generative
+from .base import _NoArg
+from .base import Executable
+from .base import Generative
+from .base import HasMemoized
+from .base import Immutable
+from .base import NO_ARG
+from .base import SingletonConstant
+from .cache_key import MemoizedHasCacheKey
+from .cache_key import NO_CACHE
+from .coercions import _document_text_coercion  # noqa
+from .operators import ColumnOperators
+from .traversals import HasCopyInternals
+from .visitors import cloned_traverse
+from .visitors import ExternallyTraversible
+from .visitors import InternalTraversal
+from .visitors import traverse
+from .visitors import Visitable
+from .. import exc
+from .. import inspection
+from .. import util
+from ..util import HasMemoized_ro_memoized_attribute
+from ..util import TypingOnly
+from ..util.typing import Literal
+from ..util.typing import ParamSpec
+from ..util.typing import Self
+
+if typing.TYPE_CHECKING:
+    from ._typing import _ByArgument
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnExpressionOrStrLabelArgument
+    from ._typing import _HasDialect
+    from ._typing import _InfoType
+    from ._typing import _PropagateAttrsType
+    from ._typing import _TypeEngineArgument
+    from .base import ColumnSet
+    from .cache_key import _CacheKeyTraversalType
+    from .cache_key import CacheKey
+    from .compiler import Compiled
+    from .compiler import SQLCompiler
+    from .functions import FunctionElement
+    from .operators import OperatorType
+    from .schema import Column
+    from .schema import DefaultGenerator
+    from .schema import FetchedValue
+    from .schema import ForeignKey
+    from .selectable import _SelectIterable
+    from .selectable import FromClause
+    from .selectable import NamedFromClause
+    from .selectable import TextualSelect
+    from .sqltypes import TupleType
+    from .type_api import TypeEngine
+    from .visitors import _CloneCallableType
+    from .visitors import _TraverseInternalsType
+    from .visitors import anon_map
+    from ..engine import Connection
+    from ..engine import Dialect
+    from ..engine.interfaces import _CoreMultiExecuteParams
+    from ..engine.interfaces import CacheStats
+    from ..engine.interfaces import CompiledCacheType
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..engine.interfaces import SchemaTranslateMapType
+    from ..engine.result import Result
+
+_NUMERIC = Union[float, Decimal]
+_NUMBER = Union[float, int, Decimal]
+
+_T = TypeVar("_T", bound="Any")
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+_OPT = TypeVar("_OPT", bound="Any")
+_NT = TypeVar("_NT", bound="_NUMERIC")
+
+_NMT = TypeVar("_NMT", bound="_NUMBER")
+
+
+@overload
+def literal(
+    value: Any,
+    type_: _TypeEngineArgument[_T],
+    literal_execute: bool = False,
+) -> BindParameter[_T]: ...
+
+
+@overload
+def literal(
+    value: _T,
+    type_: None = None,
+    literal_execute: bool = False,
+) -> BindParameter[_T]: ...
+
+
+@overload
+def literal(
+    value: Any,
+    type_: Optional[_TypeEngineArgument[Any]] = None,
+    literal_execute: bool = False,
+) -> BindParameter[Any]: ...
+
+
+def literal(
+    value: Any,
+    type_: Optional[_TypeEngineArgument[Any]] = None,
+    literal_execute: bool = False,
+) -> BindParameter[Any]:
+    r"""Return a literal clause, bound to a bind parameter.
+
+    Literal clauses are created automatically when non-
+    :class:`_expression.ClauseElement` objects (such as strings, ints, dates,
+    etc.) are
+    used in a comparison operation with a :class:`_expression.ColumnElement`
+    subclass,
+    such as a :class:`~sqlalchemy.schema.Column` object.  Use this function
+    to force the generation of a literal clause, which will be created as a
+    :class:`BindParameter` with a bound value.
+
+    :param value: the value to be bound. Can be any Python object supported by
+     the underlying DB-API, or is translatable via the given type argument.
+
+    :param type\_: an optional :class:`~sqlalchemy.types.TypeEngine` which will
+     provide bind-parameter translation for this literal.
+
+    :param literal_execute: optional bool, when True, the SQL engine will
+     attempt to render the bound value directly in the SQL statement at
+     execution time rather than providing as a parameter value.
+
+     .. versionadded:: 2.0
+
+    """
+    return coercions.expect(
+        roles.LiteralValueRole,
+        value,
+        type_=type_,
+        literal_execute=literal_execute,
+    )
+
+
+def literal_column(
+    text: str, type_: Optional[_TypeEngineArgument[_T]] = None
+) -> ColumnClause[_T]:
+    r"""Produce a :class:`.ColumnClause` object that has the
+    :paramref:`_expression.column.is_literal` flag set to True.
+
+    :func:`_expression.literal_column` is similar to
+    :func:`_expression.column`, except that
+    it is more often used as a "standalone" column expression that renders
+    exactly as stated; while :func:`_expression.column`
+    stores a string name that
+    will be assumed to be part of a table and may be quoted as such,
+    :func:`_expression.literal_column` can be that,
+    or any other arbitrary column-oriented
+    expression.
+
+    :param text: the text of the expression; can be any SQL expression.
+      Quoting rules will not be applied. To specify a column-name expression
+      which should be subject to quoting rules, use the :func:`column`
+      function.
+
+    :param type\_: an optional :class:`~sqlalchemy.types.TypeEngine`
+      object which will
+      provide result-set translation and additional expression semantics for
+      this column. If left as ``None`` the type will be :class:`.NullType`.
+
+    .. seealso::
+
+        :func:`_expression.column`
+
+        :func:`_expression.text`
+
+        :ref:`tutorial_select_arbitrary_text`
+
+    """
+    return ColumnClause(text, type_=type_, is_literal=True)
+
+
+class CompilerElement(Visitable):
+    """base class for SQL elements that can be compiled to produce a
+    SQL string.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __slots__ = ()
+    __visit_name__ = "compiler_element"
+
+    supports_execution = False
+
+    stringify_dialect = "default"
+
+    @util.preload_module("sqlalchemy.engine.default")
+    @util.preload_module("sqlalchemy.engine.url")
+    def compile(
+        self,
+        bind: Optional[_HasDialect] = None,
+        dialect: Optional[Dialect] = None,
+        **kw: Any,
+    ) -> Compiled:
+        """Compile this SQL expression.
+
+        The return value is a :class:`~.Compiled` object.
+        Calling ``str()`` or ``unicode()`` on the returned value will yield a
+        string representation of the result. The
+        :class:`~.Compiled` object also can return a
+        dictionary of bind parameter names and values
+        using the ``params`` accessor.
+
+        :param bind: An :class:`.Connection` or :class:`.Engine` which
+           can provide a :class:`.Dialect` in order to generate a
+           :class:`.Compiled` object.  If the ``bind`` and
+           ``dialect`` parameters are both omitted, a default SQL compiler
+           is used.
+
+        :param column_keys: Used for INSERT and UPDATE statements, a list of
+            column names which should be present in the VALUES clause of the
+            compiled statement. If ``None``, all columns from the target table
+            object are rendered.
+
+        :param dialect: A :class:`.Dialect` instance which can generate
+            a :class:`.Compiled` object.  This argument takes precedence over
+            the ``bind`` argument.
+
+        :param compile_kwargs: optional dictionary of additional parameters
+            that will be passed through to the compiler within all "visit"
+            methods.  This allows any custom flag to be passed through to
+            a custom compilation construct, for example.  It is also used
+            for the case of passing the ``literal_binds`` flag through::
+
+                from sqlalchemy.sql import table, column, select
+
+                t = table("t", column("x"))
+
+                s = select(t).where(t.c.x == 5)
+
+                print(s.compile(compile_kwargs={"literal_binds": True}))
+
+        .. seealso::
+
+            :ref:`faq_sql_expression_string`
+
+        """
+
+        if dialect is None:
+            if bind:
+                dialect = bind.dialect
+            elif self.stringify_dialect == "default":
+                default = util.preloaded.engine_default
+                dialect = default.StrCompileDialect()
+            else:
+                url = util.preloaded.engine_url
+                dialect = url.URL.create(
+                    self.stringify_dialect
+                ).get_dialect()()
+
+        return self._compiler(dialect, **kw)
+
+    def _compiler(self, dialect: Dialect, **kw: Any) -> Compiled:
+        """Return a compiler appropriate for this ClauseElement, given a
+        Dialect."""
+
+        if TYPE_CHECKING:
+            assert isinstance(self, ClauseElement)
+        return dialect.statement_compiler(dialect, self, **kw)
+
+    def __str__(self) -> str:
+        return str(self.compile())
+
+
+@inspection._self_inspects
+class ClauseElement(
+    SupportsWrappingAnnotations,
+    MemoizedHasCacheKey,
+    HasCopyInternals,
+    ExternallyTraversible,
+    CompilerElement,
+):
+    """Base class for elements of a programmatically constructed SQL
+    expression.
+
+    """
+
+    __visit_name__ = "clause"
+
+    if TYPE_CHECKING:
+
+        @util.memoized_property
+        def _propagate_attrs(self) -> _PropagateAttrsType:
+            """like annotations, however these propagate outwards liberally
+            as SQL constructs are built, and are set up at construction time.
+
+            """
+            ...
+
+    else:
+        _propagate_attrs = util.EMPTY_DICT
+
+    @util.ro_memoized_property
+    def description(self) -> Optional[str]:
+        return None
+
+    _is_clone_of: Optional[Self] = None
+
+    is_clause_element = True
+    is_selectable = False
+    is_dml = False
+    _is_column_element = False
+    _is_keyed_column_element = False
+    _is_table = False
+    _gen_static_annotations_cache_key = False
+    _is_textual = False
+    _is_from_clause = False
+    _is_returns_rows = False
+    _is_text_clause = False
+    _is_from_container = False
+    _is_select_container = False
+    _is_select_base = False
+    _is_select_statement = False
+    _is_bind_parameter = False
+    _is_clause_list = False
+    _is_lambda_element = False
+    _is_singleton_constant = False
+    _is_immutable = False
+    _is_star = False
+
+    @property
+    def _order_by_label_element(self) -> Optional[Label[Any]]:
+        return None
+
+    _cache_key_traversal: _CacheKeyTraversalType = None
+
+    negation_clause: ColumnElement[bool]
+
+    if typing.TYPE_CHECKING:
+
+        def get_children(
+            self, *, omit_attrs: typing_Tuple[str, ...] = ..., **kw: Any
+        ) -> Iterable[ClauseElement]: ...
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return []
+
+    def _set_propagate_attrs(self, values: Mapping[str, Any]) -> Self:
+        # usually, self._propagate_attrs is empty here.  one case where it's
+        # not is a subquery against ORM select, that is then pulled as a
+        # property of an aliased class.   should all be good
+
+        # assert not self._propagate_attrs
+
+        self._propagate_attrs = util.immutabledict(values)
+        return self
+
+    def _clone(self, **kw: Any) -> Self:
+        """Create a shallow copy of this ClauseElement.
+
+        This method may be used by a generative API.  Its also used as
+        part of the "deep" copy afforded by a traversal that combines
+        the _copy_internals() method.
+
+        """
+
+        skip = self._memoized_keys
+        c = self.__class__.__new__(self.__class__)
+
+        if skip:
+            # ensure this iteration remains atomic
+            c.__dict__ = {
+                k: v for k, v in self.__dict__.copy().items() if k not in skip
+            }
+        else:
+            c.__dict__ = self.__dict__.copy()
+
+        # this is a marker that helps to "equate" clauses to each other
+        # when a Select returns its list of FROM clauses.  the cloning
+        # process leaves around a lot of remnants of the previous clause
+        # typically in the form of column expressions still attached to the
+        # old table.
+        cc = self._is_clone_of
+        c._is_clone_of = cc if cc is not None else self
+        return c
+
+    def _negate_in_binary(self, negated_op, original_op):
+        """a hook to allow the right side of a binary expression to respond
+        to a negation of the binary expression.
+
+        Used for the special case of expanding bind parameter with IN.
+
+        """
+        return self
+
+    def _with_binary_element_type(self, type_):
+        """in the context of binary expression, convert the type of this
+        object to the one given.
+
+        applies only to :class:`_expression.ColumnElement` classes.
+
+        """
+        return self
+
+    @property
+    def _constructor(self):
+        """return the 'constructor' for this ClauseElement.
+
+        This is for the purposes for creating a new object of
+        this type.   Usually, its just the element's __class__.
+        However, the "Annotated" version of the object overrides
+        to return the class of its proxied element.
+
+        """
+        return self.__class__
+
+    @HasMemoized.memoized_attribute
+    def _cloned_set(self):
+        """Return the set consisting all cloned ancestors of this
+        ClauseElement.
+
+        Includes this ClauseElement.  This accessor tends to be used for
+        FromClause objects to identify 'equivalent' FROM clauses, regardless
+        of transformative operations.
+
+        """
+        s = util.column_set()
+        f: Optional[ClauseElement] = self
+
+        # note this creates a cycle, asserted in test_memusage. however,
+        # turning this into a plain @property adds tends of thousands of method
+        # calls to Core / ORM performance tests, so the small overhead
+        # introduced by the relatively small amount of short term cycles
+        # produced here is preferable
+        while f is not None:
+            s.add(f)
+            f = f._is_clone_of
+        return s
+
+    def _de_clone(self):
+        while self._is_clone_of is not None:
+            self = self._is_clone_of
+        return self
+
+    @property
+    def entity_namespace(self):
+        raise AttributeError(
+            "This SQL expression has no entity namespace "
+            "with which to filter from."
+        )
+
+    def __getstate__(self):
+        d = self.__dict__.copy()
+        d.pop("_is_clone_of", None)
+        d.pop("_generate_cache_key", None)
+        return d
+
+    def _execute_on_connection(
+        self,
+        connection: Connection,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> Result[Any]:
+        if self.supports_execution:
+            if TYPE_CHECKING:
+                assert isinstance(self, Executable)
+            return connection._execute_clauseelement(
+                self, distilled_params, execution_options
+            )
+        else:
+            raise exc.ObjectNotExecutableError(self)
+
+    def _execute_on_scalar(
+        self,
+        connection: Connection,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> Any:
+        """an additional hook for subclasses to provide a different
+        implementation for connection.scalar() vs. connection.execute().
+
+        .. versionadded:: 2.0
+
+        """
+        return self._execute_on_connection(
+            connection, distilled_params, execution_options
+        ).scalar()
+
+    def _get_embedded_bindparams(self) -> Sequence[BindParameter[Any]]:
+        """Return the list of :class:`.BindParameter` objects embedded in the
+        object.
+
+        This accomplishes the same purpose as ``visitors.traverse()`` or
+        similar would provide, however by making use of the cache key
+        it takes advantage of memoization of the key to result in fewer
+        net method calls, assuming the statement is also going to be
+        executed.
+
+        """
+
+        key = self._generate_cache_key()
+        if key is None:
+            bindparams: List[BindParameter[Any]] = []
+
+            traverse(self, {}, {"bindparam": bindparams.append})
+            return bindparams
+
+        else:
+            return key.bindparams
+
+    def unique_params(
+        self,
+        __optionaldict: Optional[Dict[str, Any]] = None,
+        **kwargs: Any,
+    ) -> Self:
+        """Return a copy with :func:`_expression.bindparam` elements
+        replaced.
+
+        Same functionality as :meth:`_expression.ClauseElement.params`,
+        except adds `unique=True`
+        to affected bind parameters so that multiple statements can be
+        used.
+
+        """
+        return self._replace_params(True, __optionaldict, kwargs)
+
+    def params(
+        self,
+        __optionaldict: Optional[Mapping[str, Any]] = None,
+        **kwargs: Any,
+    ) -> Self:
+        """Return a copy with :func:`_expression.bindparam` elements
+        replaced.
+
+        Returns a copy of this ClauseElement with
+        :func:`_expression.bindparam`
+        elements replaced with values taken from the given dictionary::
+
+          >>> clause = column("x") + bindparam("foo")
+          >>> print(clause.compile().params)
+          {'foo':None}
+          >>> print(clause.params({"foo": 7}).compile().params)
+          {'foo':7}
+
+        """
+        return self._replace_params(False, __optionaldict, kwargs)
+
+    def _replace_params(
+        self,
+        unique: bool,
+        optionaldict: Optional[Mapping[str, Any]],
+        kwargs: Dict[str, Any],
+    ) -> Self:
+        if optionaldict:
+            kwargs.update(optionaldict)
+
+        def visit_bindparam(bind: BindParameter[Any]) -> None:
+            if bind.key in kwargs:
+                bind.value = kwargs[bind.key]
+                bind.required = False
+            if unique:
+                bind._convert_to_unique()
+
+        return cloned_traverse(
+            self,
+            {"maintain_key": True, "detect_subquery_cols": True},
+            {"bindparam": visit_bindparam},
+        )
+
+    def compare(self, other: ClauseElement, **kw: Any) -> bool:
+        r"""Compare this :class:`_expression.ClauseElement` to
+        the given :class:`_expression.ClauseElement`.
+
+        Subclasses should override the default behavior, which is a
+        straight identity comparison.
+
+        \**kw are arguments consumed by subclass ``compare()`` methods and
+        may be used to modify the criteria for comparison
+        (see :class:`_expression.ColumnElement`).
+
+        """
+        return traversals.compare(self, other, **kw)
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> ClauseElement:
+        """Apply a 'grouping' to this :class:`_expression.ClauseElement`.
+
+        This method is overridden by subclasses to return a "grouping"
+        construct, i.e. parenthesis.   In particular it's used by "binary"
+        expressions to provide a grouping around themselves when placed into a
+        larger expression, as well as by :func:`_expression.select`
+        constructs when placed into the FROM clause of another
+        :func:`_expression.select`.  (Note that subqueries should be
+        normally created using the :meth:`_expression.Select.alias` method,
+        as many
+        platforms require nested SELECT statements to be named).
+
+        As expressions are composed together, the application of
+        :meth:`self_group` is automatic - end-user code should never
+        need to use this method directly.  Note that SQLAlchemy's
+        clause constructs take operator precedence into account -
+        so parenthesis might not be needed, for example, in
+        an expression like ``x OR (y AND z)`` - AND takes precedence
+        over OR.
+
+        The base :meth:`self_group` method of
+        :class:`_expression.ClauseElement`
+        just returns self.
+        """
+        return self
+
+    def _ungroup(self) -> ClauseElement:
+        """Return this :class:`_expression.ClauseElement`
+        without any groupings.
+        """
+
+        return self
+
+    def _compile_w_cache(
+        self,
+        dialect: Dialect,
+        *,
+        compiled_cache: Optional[CompiledCacheType],
+        column_keys: List[str],
+        for_executemany: bool = False,
+        schema_translate_map: Optional[SchemaTranslateMapType] = None,
+        **kw: Any,
+    ) -> typing_Tuple[
+        Compiled, Optional[Sequence[BindParameter[Any]]], CacheStats
+    ]:
+        elem_cache_key: Optional[CacheKey]
+
+        if compiled_cache is not None and dialect._supports_statement_cache:
+            elem_cache_key = self._generate_cache_key()
+        else:
+            elem_cache_key = None
+
+        if elem_cache_key is not None:
+            if TYPE_CHECKING:
+                assert compiled_cache is not None
+
+            cache_key, extracted_params = elem_cache_key
+            key = (
+                dialect,
+                cache_key,
+                tuple(column_keys),
+                bool(schema_translate_map),
+                for_executemany,
+            )
+            compiled_sql = compiled_cache.get(key)
+
+            if compiled_sql is None:
+                cache_hit = dialect.CACHE_MISS
+                compiled_sql = self._compiler(
+                    dialect,
+                    cache_key=elem_cache_key,
+                    column_keys=column_keys,
+                    for_executemany=for_executemany,
+                    schema_translate_map=schema_translate_map,
+                    **kw,
+                )
+                compiled_cache[key] = compiled_sql
+            else:
+                cache_hit = dialect.CACHE_HIT
+        else:
+            extracted_params = None
+            compiled_sql = self._compiler(
+                dialect,
+                cache_key=elem_cache_key,
+                column_keys=column_keys,
+                for_executemany=for_executemany,
+                schema_translate_map=schema_translate_map,
+                **kw,
+            )
+
+            if not dialect._supports_statement_cache:
+                cache_hit = dialect.NO_DIALECT_SUPPORT
+            elif compiled_cache is None:
+                cache_hit = dialect.CACHING_DISABLED
+            else:
+                cache_hit = dialect.NO_CACHE_KEY
+
+        return compiled_sql, extracted_params, cache_hit
+
+    def __invert__(self):
+        # undocumented element currently used by the ORM for
+        # relationship.contains()
+        if hasattr(self, "negation_clause"):
+            return self.negation_clause
+        else:
+            return self._negate()
+
+    def _negate(self) -> ClauseElement:
+        grouped = self.self_group(against=operators.inv)
+        assert isinstance(grouped, ColumnElement)
+        return UnaryExpression(grouped, operator=operators.inv)
+
+    def __bool__(self):
+        raise TypeError("Boolean value of this clause is not defined")
+
+    def __repr__(self):
+        friendly = self.description
+        if friendly is None:
+            return object.__repr__(self)
+        else:
+            return "<%s.%s at 0x%x; %s>" % (
+                self.__module__,
+                self.__class__.__name__,
+                id(self),
+                friendly,
+            )
+
+
+class DQLDMLClauseElement(ClauseElement):
+    """represents a :class:`.ClauseElement` that compiles to a DQL or DML
+    expression, not DDL.
+
+    .. versionadded:: 2.0
+
+    """
+
+    if typing.TYPE_CHECKING:
+
+        def _compiler(self, dialect: Dialect, **kw: Any) -> SQLCompiler:
+            """Return a compiler appropriate for this ClauseElement, given a
+            Dialect."""
+            ...
+
+        def compile(  # noqa: A001
+            self,
+            bind: Optional[_HasDialect] = None,
+            dialect: Optional[Dialect] = None,
+            **kw: Any,
+        ) -> SQLCompiler: ...
+
+
+class CompilerColumnElement(
+    roles.DMLColumnRole,
+    roles.DDLConstraintColumnRole,
+    roles.ColumnsClauseRole,
+    CompilerElement,
+):
+    """A compiler-only column element used for ad-hoc string compilations.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __slots__ = ()
+
+    _propagate_attrs = util.EMPTY_DICT
+    _is_collection_aggregate = False
+
+
+# SQLCoreOperations should be suiting the ExpressionElementRole
+# and ColumnsClauseRole.   however the MRO issues become too elaborate
+# at the moment.
+class SQLCoreOperations(Generic[_T_co], ColumnOperators, TypingOnly):
+    __slots__ = ()
+
+    # annotations for comparison methods
+    # these are from operators->Operators / ColumnOperators,
+    # redefined with the specific types returned by ColumnElement hierarchies
+    if typing.TYPE_CHECKING:
+
+        @util.non_memoized_property
+        def _propagate_attrs(self) -> _PropagateAttrsType: ...
+
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]: ...
+
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnElement[Any]: ...
+
+        @overload
+        def op(
+            self,
+            opstring: str,
+            precedence: int = ...,
+            is_comparison: bool = ...,
+            *,
+            return_type: _TypeEngineArgument[_OPT],
+            python_impl: Optional[Callable[..., Any]] = None,
+        ) -> Callable[[Any], BinaryExpression[_OPT]]: ...
+
+        @overload
+        def op(
+            self,
+            opstring: str,
+            precedence: int = ...,
+            is_comparison: bool = ...,
+            return_type: Optional[_TypeEngineArgument[Any]] = ...,
+            python_impl: Optional[Callable[..., Any]] = ...,
+        ) -> Callable[[Any], BinaryExpression[Any]]: ...
+
+        def op(
+            self,
+            opstring: str,
+            precedence: int = 0,
+            is_comparison: bool = False,
+            return_type: Optional[_TypeEngineArgument[Any]] = None,
+            python_impl: Optional[Callable[..., Any]] = None,
+        ) -> Callable[[Any], BinaryExpression[Any]]: ...
+
+        def bool_op(
+            self,
+            opstring: str,
+            precedence: int = 0,
+            python_impl: Optional[Callable[..., Any]] = None,
+        ) -> Callable[[Any], BinaryExpression[bool]]: ...
+
+        def __and__(self, other: Any) -> BooleanClauseList: ...
+
+        def __or__(self, other: Any) -> BooleanClauseList: ...
+
+        def __invert__(self) -> ColumnElement[_T_co]: ...
+
+        def __lt__(self, other: Any) -> ColumnElement[bool]: ...
+
+        def __le__(self, other: Any) -> ColumnElement[bool]: ...
+
+        # declare also that this class has an hash method otherwise
+        # it may be assumed to be None by type checkers since the
+        # object defines __eq__ and python sets it to None in that case:
+        # https://docs.python.org/3/reference/datamodel.html#object.__hash__
+        def __hash__(self) -> int: ...
+
+        def __eq__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            ...
+
+        def __ne__(self, other: Any) -> ColumnElement[bool]:  # type: ignore[override]  # noqa: E501
+            ...
+
+        def is_distinct_from(self, other: Any) -> ColumnElement[bool]: ...
+
+        def is_not_distinct_from(self, other: Any) -> ColumnElement[bool]: ...
+
+        def __gt__(self, other: Any) -> ColumnElement[bool]: ...
+
+        def __ge__(self, other: Any) -> ColumnElement[bool]: ...
+
+        def __neg__(self) -> UnaryExpression[_T_co]: ...
+
+        def __contains__(self, other: Any) -> ColumnElement[bool]: ...
+
+        def __getitem__(self, index: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __lshift__(self: _SQO[int], other: Any) -> ColumnElement[int]: ...
+
+        @overload
+        def __lshift__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __lshift__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rshift__(self: _SQO[int], other: Any) -> ColumnElement[int]: ...
+
+        @overload
+        def __rshift__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rshift__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def concat(self: _SQO[str], other: Any) -> ColumnElement[str]: ...
+
+        @overload
+        def concat(self, other: Any) -> ColumnElement[Any]: ...
+
+        def concat(self, other: Any) -> ColumnElement[Any]: ...
+
+        def like(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def ilike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def bitwise_xor(self, other: Any) -> BinaryExpression[Any]: ...
+
+        def bitwise_or(self, other: Any) -> BinaryExpression[Any]: ...
+
+        def bitwise_and(self, other: Any) -> BinaryExpression[Any]: ...
+
+        def bitwise_not(self) -> UnaryExpression[_T_co]: ...
+
+        def bitwise_lshift(self, other: Any) -> BinaryExpression[Any]: ...
+
+        def bitwise_rshift(self, other: Any) -> BinaryExpression[Any]: ...
+
+        def in_(
+            self,
+            other: Union[
+                Iterable[Any], BindParameter[Any], roles.InElementRole
+            ],
+        ) -> BinaryExpression[bool]: ...
+
+        def not_in(
+            self,
+            other: Union[
+                Iterable[Any], BindParameter[Any], roles.InElementRole
+            ],
+        ) -> BinaryExpression[bool]: ...
+
+        def notin_(
+            self,
+            other: Union[
+                Iterable[Any], BindParameter[Any], roles.InElementRole
+            ],
+        ) -> BinaryExpression[bool]: ...
+
+        def not_like(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def notlike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def not_ilike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def notilike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> BinaryExpression[bool]: ...
+
+        def is_(self, other: Any) -> BinaryExpression[bool]: ...
+
+        def is_not(self, other: Any) -> BinaryExpression[bool]: ...
+
+        def isnot(self, other: Any) -> BinaryExpression[bool]: ...
+
+        def startswith(
+            self,
+            other: Any,
+            escape: Optional[str] = None,
+            autoescape: bool = False,
+        ) -> ColumnElement[bool]: ...
+
+        def istartswith(
+            self,
+            other: Any,
+            escape: Optional[str] = None,
+            autoescape: bool = False,
+        ) -> ColumnElement[bool]: ...
+
+        def endswith(
+            self,
+            other: Any,
+            escape: Optional[str] = None,
+            autoescape: bool = False,
+        ) -> ColumnElement[bool]: ...
+
+        def iendswith(
+            self,
+            other: Any,
+            escape: Optional[str] = None,
+            autoescape: bool = False,
+        ) -> ColumnElement[bool]: ...
+
+        def contains(self, other: Any, **kw: Any) -> ColumnElement[bool]: ...
+
+        def icontains(self, other: Any, **kw: Any) -> ColumnElement[bool]: ...
+
+        def match(self, other: Any, **kwargs: Any) -> ColumnElement[bool]: ...
+
+        def regexp_match(
+            self, pattern: Any, flags: Optional[str] = None
+        ) -> ColumnElement[bool]: ...
+
+        def regexp_replace(
+            self, pattern: Any, replacement: Any, flags: Optional[str] = None
+        ) -> ColumnElement[str]: ...
+
+        def desc(self) -> UnaryExpression[_T_co]: ...
+
+        def asc(self) -> UnaryExpression[_T_co]: ...
+
+        def nulls_first(self) -> UnaryExpression[_T_co]: ...
+
+        def nullsfirst(self) -> UnaryExpression[_T_co]: ...
+
+        def nulls_last(self) -> UnaryExpression[_T_co]: ...
+
+        def nullslast(self) -> UnaryExpression[_T_co]: ...
+
+        def collate(self, collation: str) -> CollationClause: ...
+
+        def between(
+            self, cleft: Any, cright: Any, symmetric: bool = False
+        ) -> BinaryExpression[bool]: ...
+
+        def distinct(self: _SQO[_T_co]) -> UnaryExpression[_T_co]: ...
+
+        def any_(self) -> CollectionAggregate[Any]: ...
+
+        def all_(self) -> CollectionAggregate[Any]: ...
+
+        # numeric overloads.  These need more tweaking
+        # in particular they all need to have a variant for Optiona[_T]
+        # because Optional only applies to the data side, not the expression
+        # side
+
+        @overload
+        def __add__(
+            self: _SQO[_NMT],
+            other: Any,
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __add__(
+            self: _SQO[str],
+            other: Any,
+        ) -> ColumnElement[str]: ...
+
+        @overload
+        def __add__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __add__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __radd__(self: _SQO[_NMT], other: Any) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __radd__(self: _SQO[str], other: Any) -> ColumnElement[str]: ...
+
+        def __radd__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __sub__(
+            self: _SQO[_NMT],
+            other: Any,
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __sub__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __sub__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rsub__(
+            self: _SQO[_NMT],
+            other: Any,
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __rsub__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rsub__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __mul__(
+            self: _SQO[_NMT],
+            other: Any,
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __mul__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __mul__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rmul__(
+            self: _SQO[_NMT],
+            other: Any,
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __rmul__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rmul__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __mod__(self: _SQO[_NMT], other: Any) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __mod__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __mod__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rmod__(self: _SQO[_NMT], other: Any) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __rmod__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rmod__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __truediv__(
+            self: _SQO[int], other: Any
+        ) -> ColumnElement[_NUMERIC]: ...
+
+        @overload
+        def __truediv__(self: _SQO[_NT], other: Any) -> ColumnElement[_NT]: ...
+
+        @overload
+        def __truediv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __truediv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rtruediv__(
+            self: _SQO[_NMT], other: Any
+        ) -> ColumnElement[_NUMERIC]: ...
+
+        @overload
+        def __rtruediv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rtruediv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __floordiv__(
+            self: _SQO[_NMT], other: Any
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __floordiv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __floordiv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        @overload
+        def __rfloordiv__(
+            self: _SQO[_NMT], other: Any
+        ) -> ColumnElement[_NMT]: ...
+
+        @overload
+        def __rfloordiv__(self, other: Any) -> ColumnElement[Any]: ...
+
+        def __rfloordiv__(self, other: Any) -> ColumnElement[Any]: ...
+
+
+class SQLColumnExpression(
+    SQLCoreOperations[_T_co], roles.ExpressionElementRole[_T_co], TypingOnly
+):
+    """A type that may be used to indicate any SQL column element or object
+    that acts in place of one.
+
+    :class:`.SQLColumnExpression` is a base of
+    :class:`.ColumnElement`, as well as within the bases of ORM elements
+    such as :class:`.InstrumentedAttribute`, and may be used in :pep:`484`
+    typing to indicate arguments or return values that should behave
+    as column expressions.
+
+    .. versionadded:: 2.0.0b4
+
+
+    """
+
+    __slots__ = ()
+
+
+_SQO = SQLCoreOperations
+
+
+class ColumnElement(
+    roles.ColumnArgumentOrKeyRole,
+    roles.StatementOptionRole,
+    roles.WhereHavingRole,
+    roles.BinaryElementRole[_T],
+    roles.OrderByRole,
+    roles.ColumnsClauseRole,
+    roles.LimitOffsetRole,
+    roles.DMLColumnRole,
+    roles.DDLConstraintColumnRole,
+    roles.DDLExpressionRole,
+    SQLColumnExpression[_T],
+    DQLDMLClauseElement,
+):
+    """Represent a column-oriented SQL expression suitable for usage in the
+    "columns" clause, WHERE clause etc. of a statement.
+
+    While the most familiar kind of :class:`_expression.ColumnElement` is the
+    :class:`_schema.Column` object, :class:`_expression.ColumnElement`
+    serves as the basis
+    for any unit that may be present in a SQL expression, including
+    the expressions themselves, SQL functions, bound parameters,
+    literal expressions, keywords such as ``NULL``, etc.
+    :class:`_expression.ColumnElement`
+    is the ultimate base class for all such elements.
+
+    A wide variety of SQLAlchemy Core functions work at the SQL expression
+    level, and are intended to accept instances of
+    :class:`_expression.ColumnElement` as
+    arguments.  These functions will typically document that they accept a
+    "SQL expression" as an argument.  What this means in terms of SQLAlchemy
+    usually refers to an input which is either already in the form of a
+    :class:`_expression.ColumnElement` object,
+    or a value which can be **coerced** into
+    one.  The coercion rules followed by most, but not all, SQLAlchemy Core
+    functions with regards to SQL expressions are as follows:
+
+        * a literal Python value, such as a string, integer or floating
+          point value, boolean, datetime, ``Decimal`` object, or virtually
+          any other Python object, will be coerced into a "literal bound
+          value".  This generally means that a :func:`.bindparam` will be
+          produced featuring the given value embedded into the construct; the
+          resulting :class:`.BindParameter` object is an instance of
+          :class:`_expression.ColumnElement`.
+          The Python value will ultimately be sent
+          to the DBAPI at execution time as a parameterized argument to the
+          ``execute()`` or ``executemany()`` methods, after SQLAlchemy
+          type-specific converters (e.g. those provided by any associated
+          :class:`.TypeEngine` objects) are applied to the value.
+
+        * any special object value, typically ORM-level constructs, which
+          feature an accessor called ``__clause_element__()``.  The Core
+          expression system looks for this method when an object of otherwise
+          unknown type is passed to a function that is looking to coerce the
+          argument into a :class:`_expression.ColumnElement` and sometimes a
+          :class:`_expression.SelectBase` expression.
+          It is used within the ORM to
+          convert from ORM-specific objects like mapped classes and
+          mapped attributes into Core expression objects.
+
+        * The Python ``None`` value is typically interpreted as ``NULL``,
+          which in SQLAlchemy Core produces an instance of :func:`.null`.
+
+    A :class:`_expression.ColumnElement` provides the ability to generate new
+    :class:`_expression.ColumnElement`
+    objects using Python expressions.  This means that Python operators
+    such as ``==``, ``!=`` and ``<`` are overloaded to mimic SQL operations,
+    and allow the instantiation of further :class:`_expression.ColumnElement`
+    instances
+    which are composed from other, more fundamental
+    :class:`_expression.ColumnElement`
+    objects.  For example, two :class:`.ColumnClause` objects can be added
+    together with the addition operator ``+`` to produce
+    a :class:`.BinaryExpression`.
+    Both :class:`.ColumnClause` and :class:`.BinaryExpression` are subclasses
+    of :class:`_expression.ColumnElement`:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy.sql import column
+        >>> column("a") + column("b")
+        <sqlalchemy.sql.expression.BinaryExpression object at 0x101029dd0>
+        >>> print(column("a") + column("b"))
+        {printsql}a + b
+
+    .. seealso::
+
+        :class:`_schema.Column`
+
+        :func:`_expression.column`
+
+    """
+
+    __visit_name__ = "column_element"
+
+    primary_key: bool = False
+    _is_clone_of: Optional[ColumnElement[_T]]
+    _is_column_element = True
+    _insert_sentinel: bool = False
+    _omit_from_statements = False
+    _is_collection_aggregate = False
+
+    foreign_keys: AbstractSet[ForeignKey] = frozenset()
+
+    @util.memoized_property
+    def _proxies(self) -> List[ColumnElement[Any]]:
+        return []
+
+    @util.non_memoized_property
+    def _tq_label(self) -> Optional[str]:
+        """The named label that can be used to target
+        this column in a result set in a "table qualified" context.
+
+        This label is almost always the label used when
+        rendering <expr> AS <label> in a SELECT statement when using
+        the LABEL_STYLE_TABLENAME_PLUS_COL label style, which is what the
+        legacy ORM ``Query`` object uses as well.
+
+        For a regular Column bound to a Table, this is typically the label
+        <tablename>_<columnname>.  For other constructs, different rules
+        may apply, such as anonymized labels and others.
+
+        .. versionchanged:: 1.4.21 renamed from ``._label``
+
+        """
+        return None
+
+    key: Optional[str] = None
+    """The 'key' that in some circumstances refers to this object in a
+    Python namespace.
+
+    This typically refers to the "key" of the column as present in the
+    ``.c`` collection of a selectable, e.g. ``sometable.c["somekey"]`` would
+    return a :class:`_schema.Column` with a ``.key`` of "somekey".
+
+    """
+
+    @HasMemoized.memoized_attribute
+    def _tq_key_label(self) -> Optional[str]:
+        """A label-based version of 'key' that in some circumstances refers
+        to this object in a Python namespace.
+
+
+        _tq_key_label comes into play when a select() statement is constructed
+        with apply_labels(); in this case, all Column objects in the ``.c``
+        collection are rendered as <tablename>_<columnname> in SQL; this is
+        essentially the value of ._label. But to locate those columns in the
+        ``.c`` collection, the name is along the lines of <tablename>_<key>;
+        that's the typical value of .key_label.
+
+        .. versionchanged:: 1.4.21 renamed from ``._key_label``
+
+        """
+        return self._proxy_key
+
+    @property
+    def _key_label(self) -> Optional[str]:
+        """legacy; renamed to _tq_key_label"""
+        return self._tq_key_label
+
+    @property
+    def _label(self) -> Optional[str]:
+        """legacy; renamed to _tq_label"""
+        return self._tq_label
+
+    @property
+    def _non_anon_label(self) -> Optional[str]:
+        """the 'name' that naturally applies this element when rendered in
+        SQL.
+
+        Concretely, this is the "name" of a column or a label in a
+        SELECT statement; ``<columnname>`` and ``<labelname>`` below:
+
+        .. sourcecode:: sql
+
+            SELECT <columnmame> FROM table
+
+            SELECT column AS <labelname> FROM table
+
+        Above, the two names noted will be what's present in the DBAPI
+        ``cursor.description`` as the names.
+
+        If this attribute returns ``None``, it means that the SQL element as
+        written does not have a 100% fully predictable "name" that would appear
+        in the ``cursor.description``. Examples include SQL functions, CAST
+        functions, etc. While such things do return names in
+        ``cursor.description``, they are only predictable on a
+        database-specific basis; e.g. an expression like ``MAX(table.col)`` may
+        appear as the string ``max`` on one database (like PostgreSQL) or may
+        appear as the whole expression ``max(table.col)`` on SQLite.
+
+        The default implementation looks for a ``.name`` attribute on the
+        object, as has been the precedent established in SQLAlchemy for many
+        years.  An exception is made on the ``FunctionElement`` subclass
+        so that the return value is always ``None``.
+
+        .. versionadded:: 1.4.21
+
+
+
+        """
+        return getattr(self, "name", None)
+
+    _render_label_in_columns_clause = True
+    """A flag used by select._columns_plus_names that helps to determine
+    we are actually going to render in terms of "SELECT <col> AS <label>".
+    This flag can be returned as False for some Column objects that want
+    to be rendered as simple "SELECT <col>"; typically columns that don't have
+    any parent table and are named the same as what the label would be
+    in any case.
+
+    """
+
+    _allow_label_resolve = True
+    """A flag that can be flipped to prevent a column from being resolvable
+    by string label name.
+
+    The joined eager loader strategy in the ORM uses this, for example.
+
+    """
+
+    _is_implicitly_boolean = False
+
+    _alt_names: Sequence[str] = ()
+
+    @overload
+    def self_group(self, against: None = None) -> ColumnElement[_T]: ...
+
+    @overload
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> ColumnElement[Any]: ...
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> ColumnElement[Any]:
+        if (
+            against in (operators.and_, operators.or_, operators._asbool)
+            and self.type._type_affinity is type_api.BOOLEANTYPE._type_affinity
+        ):
+            return AsBoolean(self, operators.is_true, operators.is_false)
+        elif against in (operators.any_op, operators.all_op):
+            return Grouping(self)
+        else:
+            return self
+
+    @overload
+    def _negate(self: ColumnElement[bool]) -> ColumnElement[bool]: ...
+
+    @overload
+    def _negate(self: ColumnElement[_T]) -> ColumnElement[_T]: ...
+
+    def _negate(self) -> ColumnElement[Any]:
+        if self.type._type_affinity is type_api.BOOLEANTYPE._type_affinity:
+            return AsBoolean(self, operators.is_false, operators.is_true)
+        else:
+            grouped = self.self_group(against=operators.inv)
+            assert isinstance(grouped, ColumnElement)
+            return UnaryExpression(
+                grouped, operator=operators.inv, wraps_column_expression=True
+            )
+
+    type: TypeEngine[_T]
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            # used for delayed setup of
+            # type_api
+            return type_api.NULLTYPE
+
+    @HasMemoized.memoized_attribute
+    def comparator(self) -> TypeEngine.Comparator[_T]:
+        try:
+            comparator_factory = self.type.comparator_factory
+        except AttributeError as err:
+            raise TypeError(
+                "Object %r associated with '.type' attribute "
+                "is not a TypeEngine class or object" % self.type
+            ) from err
+        else:
+            return comparator_factory(self)
+
+    def __setstate__(self, state):
+        self.__dict__.update(state)
+
+    def __getattr__(self, key: str) -> Any:
+        try:
+            return getattr(self.comparator, key)
+        except AttributeError as err:
+            raise AttributeError(
+                "Neither %r object nor %r object has an attribute %r"
+                % (
+                    type(self).__name__,
+                    type(self.comparator).__name__,
+                    key,
+                )
+            ) from err
+
+    def operate(
+        self,
+        op: operators.OperatorType,
+        *other: Any,
+        **kwargs: Any,
+    ) -> ColumnElement[Any]:
+        return op(self.comparator, *other, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def reverse_operate(
+        self, op: operators.OperatorType, other: Any, **kwargs: Any
+    ) -> ColumnElement[Any]:
+        return op(other, self.comparator, **kwargs)  # type: ignore[no-any-return]  # noqa: E501
+
+    def _bind_param(
+        self,
+        operator: operators.OperatorType,
+        obj: Any,
+        type_: Optional[TypeEngine[_T]] = None,
+        expanding: bool = False,
+    ) -> BindParameter[_T]:
+        return BindParameter(
+            None,
+            obj,
+            _compared_to_operator=operator,
+            type_=type_,
+            _compared_to_type=self.type,
+            unique=True,
+            expanding=expanding,
+        )
+
+    @property
+    def expression(self) -> ColumnElement[Any]:
+        """Return a column expression.
+
+        Part of the inspection interface; returns self.
+
+        """
+        return self
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return (self,)
+
+    @util.memoized_property
+    def base_columns(self) -> FrozenSet[ColumnElement[Any]]:
+        return frozenset(c for c in self.proxy_set if not c._proxies)
+
+    @util.memoized_property
+    def proxy_set(self) -> FrozenSet[ColumnElement[Any]]:
+        """set of all columns we are proxying
+
+        as of 2.0 this is explicitly deannotated columns.  previously it was
+        effectively deannotated columns but wasn't enforced.  annotated
+        columns should basically not go into sets if at all possible because
+        their hashing behavior is very non-performant.
+
+        """
+        return frozenset([self._deannotate()]).union(
+            itertools.chain(*[c.proxy_set for c in self._proxies])
+        )
+
+    @util.memoized_property
+    def _expanded_proxy_set(self) -> FrozenSet[ColumnElement[Any]]:
+        return frozenset(_expand_cloned(self.proxy_set))
+
+    def _uncached_proxy_list(self) -> List[ColumnElement[Any]]:
+        """An 'uncached' version of proxy set.
+
+        This list includes annotated columns which perform very poorly in
+        set operations.
+
+        """
+
+        return [self] + list(
+            itertools.chain(*[c._uncached_proxy_list() for c in self._proxies])
+        )
+
+    def shares_lineage(self, othercolumn: ColumnElement[Any]) -> bool:
+        """Return True if the given :class:`_expression.ColumnElement`
+        has a common ancestor to this :class:`_expression.ColumnElement`."""
+
+        return bool(self.proxy_set.intersection(othercolumn.proxy_set))
+
+    def _compare_name_for_result(self, other: ColumnElement[Any]) -> bool:
+        """Return True if the given column element compares to this one
+        when targeting within a result row."""
+
+        return (
+            hasattr(other, "name")
+            and hasattr(self, "name")
+            and other.name == self.name
+        )
+
+    @HasMemoized.memoized_attribute
+    def _proxy_key(self) -> Optional[str]:
+        if self._annotations and "proxy_key" in self._annotations:
+            return cast(str, self._annotations["proxy_key"])
+
+        name = self.key
+        if not name:
+            # there's a bit of a seeming contradiction which is that the
+            # "_non_anon_label" of a column can in fact be an
+            # "_anonymous_label"; this is when it's on a column that is
+            # proxying for an anonymous expression in a subquery.
+            name = self._non_anon_label
+
+        if isinstance(name, _anonymous_label):
+            return None
+        else:
+            return name
+
+    @HasMemoized.memoized_attribute
+    def _expression_label(self) -> Optional[str]:
+        """a suggested label to use in the case that the column has no name,
+        which should be used if possible as the explicit 'AS <label>'
+        where this expression would normally have an anon label.
+
+        this is essentially mostly what _proxy_key does except it returns
+        None if the column has a normal name that can be used.
+
+        """
+
+        if getattr(self, "name", None) is not None:
+            return None
+        elif self._annotations and "proxy_key" in self._annotations:
+            return cast(str, self._annotations["proxy_key"])
+        else:
+            return None
+
+    def _make_proxy(
+        self,
+        selectable: FromClause,
+        *,
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        name: Optional[str] = None,
+        key: Optional[str] = None,
+        name_is_truncatable: bool = False,
+        compound_select_cols: Optional[Sequence[ColumnElement[Any]]] = None,
+        **kw: Any,
+    ) -> typing_Tuple[str, ColumnClause[_T]]:
+        """Create a new :class:`_expression.ColumnElement` representing this
+        :class:`_expression.ColumnElement` as it appears in the select list of
+        a descending selectable.
+
+        """
+        if name is None:
+            name = self._anon_name_label
+            if key is None:
+                key = self._proxy_key
+        else:
+            key = name
+
+        assert key is not None
+
+        co: ColumnClause[_T] = ColumnClause(
+            (
+                coercions.expect(roles.TruncatedLabelRole, name)
+                if name_is_truncatable
+                else name
+            ),
+            type_=getattr(self, "type", None),
+            _selectable=selectable,
+        )
+
+        co._propagate_attrs = selectable._propagate_attrs
+        if compound_select_cols:
+            co._proxies = list(compound_select_cols)
+        else:
+            co._proxies = [self]
+        if selectable._is_clone_of is not None:
+            co._is_clone_of = selectable._is_clone_of.columns.get(key)
+        return key, co
+
+    def cast(self, type_: _TypeEngineArgument[_OPT]) -> Cast[_OPT]:
+        """Produce a type cast, i.e. ``CAST(<expression> AS <type>)``.
+
+        This is a shortcut to the :func:`_expression.cast` function.
+
+        .. seealso::
+
+            :ref:`tutorial_casts`
+
+            :func:`_expression.cast`
+
+            :func:`_expression.type_coerce`
+
+        """
+        return Cast(self, type_)
+
+    def label(self, name: Optional[str]) -> Label[_T]:
+        """Produce a column label, i.e. ``<columnname> AS <name>``.
+
+        This is a shortcut to the :func:`_expression.label` function.
+
+        If 'name' is ``None``, an anonymous label name will be generated.
+
+        """
+        return Label(name, self, self.type)
+
+    def _anon_label(
+        self, seed: Optional[str], add_hash: Optional[int] = None
+    ) -> _anonymous_label:
+        while self._is_clone_of is not None:
+            self = self._is_clone_of
+
+        # as of 1.4 anonymous label for ColumnElement uses hash(), not id(),
+        # as the identifier, because a column and its annotated version are
+        # the same thing in a SQL statement
+        hash_value = hash(self)
+
+        if add_hash:
+            # this path is used for disambiguating anon labels that would
+            # otherwise be the same name for the same element repeated.
+            # an additional numeric value is factored in for each label.
+
+            # shift hash(self) (which is id(self), typically 8 byte integer)
+            # 16 bits leftward.  fill extra add_hash on right
+            assert add_hash < (2 << 15)
+            assert seed
+            hash_value = (hash_value << 16) | add_hash
+
+            # extra underscore is added for labels with extra hash
+            # values, to isolate the "deduped anon" namespace from the
+            # regular namespace.  eliminates chance of these
+            # manufactured hash values overlapping with regular ones for some
+            # undefined python interpreter
+            seed = seed + "_"
+
+        if isinstance(seed, _anonymous_label):
+            return _anonymous_label.safe_construct(
+                hash_value, "", enclosing_label=seed
+            )
+
+        return _anonymous_label.safe_construct(hash_value, seed or "anon")
+
+    @util.memoized_property
+    def _anon_name_label(self) -> str:
+        """Provides a constant 'anonymous label' for this ColumnElement.
+
+        This is a label() expression which will be named at compile time.
+        The same label() is returned each time ``anon_label`` is called so
+        that expressions can reference ``anon_label`` multiple times,
+        producing the same label name at compile time.
+
+        The compiler uses this function automatically at compile time
+        for expressions that are known to be 'unnamed' like binary
+        expressions and function calls.
+
+        .. versionchanged:: 1.4.9 - this attribute was not intended to be
+           public and is renamed to _anon_name_label.  anon_name exists
+           for backwards compat
+
+        """
+        name = getattr(self, "name", None)
+        return self._anon_label(name)
+
+    @util.memoized_property
+    def _anon_key_label(self) -> _anonymous_label:
+        """Provides a constant 'anonymous key label' for this ColumnElement.
+
+        Compare to ``anon_label``, except that the "key" of the column,
+        if available, is used to generate the label.
+
+        This is used when a deduplicating key is placed into the columns
+        collection of a selectable.
+
+        .. versionchanged:: 1.4.9 - this attribute was not intended to be
+           public and is renamed to _anon_key_label.  anon_key_label exists
+           for backwards compat
+
+        """
+        return self._anon_label(self._proxy_key)
+
+    @property
+    @util.deprecated(
+        "1.4",
+        "The :attr:`_expression.ColumnElement.anon_label` attribute is now "
+        "private, and the public accessor is deprecated.",
+    )
+    def anon_label(self) -> str:
+        return self._anon_name_label
+
+    @property
+    @util.deprecated(
+        "1.4",
+        "The :attr:`_expression.ColumnElement.anon_key_label` attribute is "
+        "now private, and the public accessor is deprecated.",
+    )
+    def anon_key_label(self) -> str:
+        return self._anon_key_label
+
+    def _dedupe_anon_label_idx(self, idx: int) -> str:
+        """label to apply to a column that is anon labeled, but repeated
+        in the SELECT, so that we have to make an "extra anon" label that
+        disambiguates it from the previous appearance.
+
+        these labels come out like "foo_bar_id__1" and have double underscores
+        in them.
+
+        """
+        label = getattr(self, "name", None)
+
+        # current convention is that if the element doesn't have a
+        # ".name" (usually because it is not NamedColumn), we try to
+        # use a "table qualified" form for the "dedupe anon" label,
+        # based on the notion that a label like
+        # "CAST(casttest.v1 AS DECIMAL) AS casttest_v1__1" looks better than
+        # "CAST(casttest.v1 AS DECIMAL) AS anon__1"
+
+        if label is None:
+            return self._dedupe_anon_tq_label_idx(idx)
+        else:
+            return self._anon_label(label, add_hash=idx)
+
+    @util.memoized_property
+    def _anon_tq_label(self) -> _anonymous_label:
+        return self._anon_label(getattr(self, "_tq_label", None))
+
+    @util.memoized_property
+    def _anon_tq_key_label(self) -> _anonymous_label:
+        return self._anon_label(getattr(self, "_tq_key_label", None))
+
+    def _dedupe_anon_tq_label_idx(self, idx: int) -> _anonymous_label:
+        label = getattr(self, "_tq_label", None) or "anon"
+
+        return self._anon_label(label, add_hash=idx)
+
+
+class KeyedColumnElement(ColumnElement[_T]):
+    """ColumnElement where ``.key`` is non-None."""
+
+    _is_keyed_column_element = True
+
+    key: str
+
+
+class WrapsColumnExpression(ColumnElement[_T]):
+    """Mixin that defines a :class:`_expression.ColumnElement`
+    as a wrapper with special
+    labeling behavior for an expression that already has a name.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`change_4449`
+
+
+    """
+
+    @property
+    def wrapped_column_expression(self) -> ColumnElement[_T]:
+        raise NotImplementedError()
+
+    @util.non_memoized_property
+    def _tq_label(self) -> Optional[str]:
+        wce = self.wrapped_column_expression
+        if hasattr(wce, "_tq_label"):
+            return wce._tq_label
+        else:
+            return None
+
+    @property
+    def _label(self) -> Optional[str]:
+        return self._tq_label
+
+    @property
+    def _non_anon_label(self) -> Optional[str]:
+        return None
+
+    @util.non_memoized_property
+    def _anon_name_label(self) -> str:
+        wce = self.wrapped_column_expression
+
+        # this logic tries to get the WrappedColumnExpression to render
+        # with "<expr> AS <name>", where "<name>" is the natural name
+        # within the expression itself.   e.g. "CAST(table.foo) AS foo".
+        if not wce._is_text_clause:
+            nal = wce._non_anon_label
+            if nal:
+                return nal
+            elif hasattr(wce, "_anon_name_label"):
+                return wce._anon_name_label
+        return super()._anon_name_label
+
+    def _dedupe_anon_label_idx(self, idx: int) -> str:
+        wce = self.wrapped_column_expression
+        nal = wce._non_anon_label
+        if nal:
+            return self._anon_label(nal + "_")
+        else:
+            return self._dedupe_anon_tq_label_idx(idx)
+
+    @property
+    def _proxy_key(self):
+        wce = self.wrapped_column_expression
+
+        if not wce._is_text_clause:
+            return wce._proxy_key
+        return super()._proxy_key
+
+
+class BindParameter(roles.InElementRole, KeyedColumnElement[_T]):
+    r"""Represent a "bound expression".
+
+    :class:`.BindParameter` is invoked explicitly using the
+    :func:`.bindparam` function, as in::
+
+        from sqlalchemy import bindparam
+
+        stmt = select(users_table).where(
+            users_table.c.name == bindparam("username")
+        )
+
+    Detailed discussion of how :class:`.BindParameter` is used is
+    at :func:`.bindparam`.
+
+    .. seealso::
+
+        :func:`.bindparam`
+
+    """
+
+    __visit_name__ = "bindparam"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("key", InternalTraversal.dp_anon_name),
+        ("type", InternalTraversal.dp_type),
+        ("callable", InternalTraversal.dp_plain_dict),
+        ("value", InternalTraversal.dp_plain_obj),
+        ("literal_execute", InternalTraversal.dp_boolean),
+    ]
+
+    key: str
+    type: TypeEngine[_T]
+    value: Optional[_T]
+
+    _is_crud = False
+    _is_bind_parameter = True
+    _key_is_anon = False
+
+    # bindparam implements its own _gen_cache_key() method however
+    # we check subclasses for this flag, else no cache key is generated
+    inherit_cache = True
+
+    def __init__(
+        self,
+        key: Optional[str],
+        value: Any = _NoArg.NO_ARG,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        unique: bool = False,
+        required: Union[bool, Literal[_NoArg.NO_ARG]] = _NoArg.NO_ARG,
+        quote: Optional[bool] = None,
+        callable_: Optional[Callable[[], Any]] = None,
+        expanding: bool = False,
+        isoutparam: bool = False,
+        literal_execute: bool = False,
+        _compared_to_operator: Optional[OperatorType] = None,
+        _compared_to_type: Optional[TypeEngine[Any]] = None,
+        _is_crud: bool = False,
+    ):
+        if required is _NoArg.NO_ARG:
+            required = value is _NoArg.NO_ARG and callable_ is None
+        if value is _NoArg.NO_ARG:
+            value = None
+
+        if quote is not None:
+            key = quoted_name.construct(key, quote)
+
+        if unique:
+            self.key = _anonymous_label.safe_construct(
+                id(self),
+                (
+                    key
+                    if key is not None
+                    and not isinstance(key, _anonymous_label)
+                    else "param"
+                ),
+                sanitize_key=True,
+            )
+            self._key_is_anon = True
+        elif key:
+            self.key = key
+        else:
+            self.key = _anonymous_label.safe_construct(id(self), "param")
+            self._key_is_anon = True
+
+        # identifying key that won't change across
+        # clones, used to identify the bind's logical
+        # identity
+        self._identifying_key = self.key
+
+        # key that was passed in the first place, used to
+        # generate new keys
+        self._orig_key = key or "param"
+
+        self.unique = unique
+        self.value = value
+        self.callable = callable_
+        self.isoutparam = isoutparam
+        self.required = required
+
+        # indicate an "expanding" parameter; the compiler sets this
+        # automatically in the compiler _render_in_expr_w_bindparam method
+        # for an IN expression
+        self.expanding = expanding
+
+        # this is another hint to help w/ expanding and is typically
+        # set in the compiler _render_in_expr_w_bindparam method for an
+        # IN expression
+        self.expand_op = None
+
+        self.literal_execute = literal_execute
+        if _is_crud:
+            self._is_crud = True
+
+        if type_ is None:
+            if expanding:
+                if value:
+                    check_value = value[0]
+                else:
+                    check_value = type_api._NO_VALUE_IN_LIST
+            else:
+                check_value = value
+            if _compared_to_type is not None:
+                self.type = _compared_to_type.coerce_compared_value(
+                    _compared_to_operator, check_value
+                )
+            else:
+                self.type = type_api._resolve_value_to_type(check_value)
+        elif isinstance(type_, type):
+            self.type = type_()
+        elif is_tuple_type(type_):
+            if value:
+                if expanding:
+                    check_value = value[0]
+                else:
+                    check_value = value
+                cast("BindParameter[typing_Tuple[Any, ...]]", self).type = (
+                    type_._resolve_values_to_types(check_value)
+                )
+            else:
+                cast("BindParameter[typing_Tuple[Any, ...]]", self).type = (
+                    type_
+                )
+        else:
+            self.type = type_
+
+    def _with_value(self, value, maintain_key=False, required=NO_ARG):
+        """Return a copy of this :class:`.BindParameter` with the given value
+        set.
+        """
+        cloned = self._clone(maintain_key=maintain_key)
+        cloned.value = value
+        cloned.callable = None
+        cloned.required = required if required is not NO_ARG else self.required
+        if cloned.type is type_api.NULLTYPE:
+            cloned.type = type_api._resolve_value_to_type(value)
+        return cloned
+
+    @property
+    def effective_value(self) -> Optional[_T]:
+        """Return the value of this bound parameter,
+        taking into account if the ``callable`` parameter
+        was set.
+
+        The ``callable`` value will be evaluated
+        and returned if present, else ``value``.
+
+        """
+        if self.callable:
+            # TODO: set up protocol for bind parameter callable
+            return self.callable()  # type: ignore
+        else:
+            return self.value
+
+    def render_literal_execute(self) -> BindParameter[_T]:
+        """Produce a copy of this bound parameter that will enable the
+        :paramref:`_sql.BindParameter.literal_execute` flag.
+
+        The :paramref:`_sql.BindParameter.literal_execute` flag will
+        have the effect of the parameter rendered in the compiled SQL
+        string using ``[POSTCOMPILE]`` form, which is a special form that
+        is converted to be a rendering of the literal value of the parameter
+        at SQL execution time.    The rationale is to support caching
+        of SQL statement strings that can embed per-statement literal values,
+        such as LIMIT and OFFSET parameters, in the final SQL string that
+        is passed to the DBAPI.   Dialects in particular may want to use
+        this method within custom compilation schemes.
+
+        .. versionadded:: 1.4.5
+
+        .. seealso::
+
+            :ref:`engine_thirdparty_caching`
+
+        """
+        c = ClauseElement._clone(self)
+        c.literal_execute = True
+        return c
+
+    def _negate_in_binary(self, negated_op, original_op):
+        if self.expand_op is original_op:
+            bind = self._clone()
+            bind.expand_op = negated_op
+            return bind
+        else:
+            return self
+
+    def _with_binary_element_type(self, type_):
+        c = ClauseElement._clone(self)
+        c.type = type_
+        return c
+
+    def _clone(self, maintain_key: bool = False, **kw: Any) -> Self:
+        c = ClauseElement._clone(self, **kw)
+        # ensure all the BindParameter objects stay in cloned set.
+        # in #7823, we changed "clone" so that a clone only keeps a reference
+        # to the "original" element, since for column correspondence, that's
+        # all we need.   However, for BindParam, _cloned_set is used by
+        # the "cache key bind match" lookup, which means if any of those
+        # interim BindParameter objects became part of a cache key in the
+        # cache, we need it.  So here, make sure all clones keep carrying
+        # forward.
+        c._cloned_set.update(self._cloned_set)
+        if not maintain_key and self.unique:
+            c.key = _anonymous_label.safe_construct(
+                id(c), c._orig_key or "param", sanitize_key=True
+            )
+        return c
+
+    def _gen_cache_key(self, anon_map, bindparams):
+        _gen_cache_ok = self.__class__.__dict__.get("inherit_cache", False)
+
+        if not _gen_cache_ok:
+            if anon_map is not None:
+                anon_map[NO_CACHE] = True
+            return None
+
+        id_, found = anon_map.get_anon(self)
+        if found:
+            return (id_, self.__class__)
+
+        if bindparams is not None:
+            bindparams.append(self)
+
+        return (
+            id_,
+            self.__class__,
+            self.type._static_cache_key,
+            self.key % anon_map if self._key_is_anon else self.key,
+            self.literal_execute,
+        )
+
+    def _convert_to_unique(self):
+        if not self.unique:
+            self.unique = True
+            self.key = _anonymous_label.safe_construct(
+                id(self), self._orig_key or "param", sanitize_key=True
+            )
+
+    def __getstate__(self):
+        """execute a deferred value for serialization purposes."""
+
+        d = self.__dict__.copy()
+        v = self.value
+        if self.callable:
+            v = self.callable()
+            d["callable"] = None
+        d["value"] = v
+        return d
+
+    def __setstate__(self, state):
+        if state.get("unique", False):
+            state["key"] = _anonymous_label.safe_construct(
+                id(self), state.get("_orig_key", "param"), sanitize_key=True
+            )
+        self.__dict__.update(state)
+
+    def __repr__(self):
+        return "%s(%r, %r, type_=%r)" % (
+            self.__class__.__name__,
+            self.key,
+            self.value,
+            self.type,
+        )
+
+
+class TypeClause(DQLDMLClauseElement):
+    """Handle a type keyword in a SQL statement.
+
+    Used by the ``Case`` statement.
+
+    """
+
+    __visit_name__ = "typeclause"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("type", InternalTraversal.dp_type)
+    ]
+
+    def __init__(self, type_):
+        self.type = type_
+
+
+class TextClause(
+    roles.DDLConstraintColumnRole,
+    roles.DDLExpressionRole,
+    roles.StatementOptionRole,
+    roles.WhereHavingRole,
+    roles.OrderByRole,
+    roles.FromClauseRole,
+    roles.SelectStatementRole,
+    roles.InElementRole,
+    Generative,
+    Executable,
+    DQLDMLClauseElement,
+    roles.BinaryElementRole[Any],
+    inspection.Inspectable["TextClause"],
+):
+    """Represent a literal SQL text fragment.
+
+    E.g.::
+
+        from sqlalchemy import text
+
+        t = text("SELECT * FROM users")
+        result = connection.execute(t)
+
+    The :class:`_expression.TextClause` construct is produced using the
+    :func:`_expression.text`
+    function; see that function for full documentation.
+
+    .. seealso::
+
+        :func:`_expression.text`
+
+    """
+
+    __visit_name__ = "textclause"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("_bindparams", InternalTraversal.dp_string_clauseelement_dict),
+        ("text", InternalTraversal.dp_string),
+    ]
+
+    _is_text_clause = True
+
+    _is_textual = True
+
+    _bind_params_regex = re.compile(r"(?<![:\w\x5c]):(\w+)(?!:)", re.UNICODE)
+    _is_implicitly_boolean = False
+
+    _render_label_in_columns_clause = False
+
+    _omit_from_statements = False
+
+    _is_collection_aggregate = False
+
+    @property
+    def _hide_froms(self) -> Iterable[FromClause]:
+        return ()
+
+    def __and__(self, other):
+        # support use in select.where(), query.filter()
+        return and_(self, other)
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return (self,)
+
+    # help in those cases where text() is
+    # interpreted in a column expression situation
+    key: Optional[str] = None
+    _label: Optional[str] = None
+
+    _allow_label_resolve = False
+
+    @property
+    def _is_star(self):
+        return self.text == "*"
+
+    def __init__(self, text: str):
+        self._bindparams: Dict[str, BindParameter[Any]] = {}
+
+        def repl(m):
+            self._bindparams[m.group(1)] = BindParameter(m.group(1))
+            return ":%s" % m.group(1)
+
+        # scan the string and search for bind parameter names, add them
+        # to the list of bindparams
+        self.text = self._bind_params_regex.sub(repl, text)
+
+    @_generative
+    def bindparams(
+        self,
+        *binds: BindParameter[Any],
+        **names_to_values: Any,
+    ) -> Self:
+        """Establish the values and/or types of bound parameters within
+        this :class:`_expression.TextClause` construct.
+
+        Given a text construct such as::
+
+            from sqlalchemy import text
+
+            stmt = text(
+                "SELECT id, name FROM user WHERE name=:name AND timestamp=:timestamp"
+            )
+
+        the :meth:`_expression.TextClause.bindparams`
+        method can be used to establish
+        the initial value of ``:name`` and ``:timestamp``,
+        using simple keyword arguments::
+
+            stmt = stmt.bindparams(
+                name="jack", timestamp=datetime.datetime(2012, 10, 8, 15, 12, 5)
+            )
+
+        Where above, new :class:`.BindParameter` objects
+        will be generated with the names ``name`` and ``timestamp``, and
+        values of ``jack`` and ``datetime.datetime(2012, 10, 8, 15, 12, 5)``,
+        respectively.  The types will be
+        inferred from the values given, in this case :class:`.String` and
+        :class:`.DateTime`.
+
+        When specific typing behavior is needed, the positional ``*binds``
+        argument can be used in which to specify :func:`.bindparam` constructs
+        directly.  These constructs must include at least the ``key``
+        argument, then an optional value and type::
+
+            from sqlalchemy import bindparam
+
+            stmt = stmt.bindparams(
+                bindparam("name", value="jack", type_=String),
+                bindparam("timestamp", type_=DateTime),
+            )
+
+        Above, we specified the type of :class:`.DateTime` for the
+        ``timestamp`` bind, and the type of :class:`.String` for the ``name``
+        bind.  In the case of ``name`` we also set the default value of
+        ``"jack"``.
+
+        Additional bound parameters can be supplied at statement execution
+        time, e.g.::
+
+            result = connection.execute(
+                stmt, timestamp=datetime.datetime(2012, 10, 8, 15, 12, 5)
+            )
+
+        The :meth:`_expression.TextClause.bindparams`
+        method can be called repeatedly,
+        where it will re-use existing :class:`.BindParameter` objects to add
+        new information.  For example, we can call
+        :meth:`_expression.TextClause.bindparams`
+        first with typing information, and a
+        second time with value information, and it will be combined::
+
+            stmt = text(
+                "SELECT id, name FROM user WHERE name=:name "
+                "AND timestamp=:timestamp"
+            )
+            stmt = stmt.bindparams(
+                bindparam("name", type_=String), bindparam("timestamp", type_=DateTime)
+            )
+            stmt = stmt.bindparams(
+                name="jack", timestamp=datetime.datetime(2012, 10, 8, 15, 12, 5)
+            )
+
+        The :meth:`_expression.TextClause.bindparams`
+        method also supports the concept of
+        **unique** bound parameters.  These are parameters that are
+        "uniquified" on name at statement compilation time, so that  multiple
+        :func:`_expression.text`
+        constructs may be combined together without the names
+        conflicting.  To use this feature, specify the
+        :paramref:`.BindParameter.unique` flag on each :func:`.bindparam`
+        object::
+
+            stmt1 = text("select id from table where name=:name").bindparams(
+                bindparam("name", value="name1", unique=True)
+            )
+            stmt2 = text("select id from table where name=:name").bindparams(
+                bindparam("name", value="name2", unique=True)
+            )
+
+            union = union_all(stmt1.columns(column("id")), stmt2.columns(column("id")))
+
+        The above statement will render as:
+
+        .. sourcecode:: sql
+
+            select id from table where name=:name_1
+            UNION ALL select id from table where name=:name_2
+
+        .. versionadded:: 1.3.11  Added support for the
+           :paramref:`.BindParameter.unique` flag to work with
+           :func:`_expression.text`
+           constructs.
+
+        """  # noqa: E501
+        self._bindparams = new_params = self._bindparams.copy()
+
+        for bind in binds:
+            try:
+                # the regex used for text() currently will not match
+                # a unique/anonymous key in any case, so use the _orig_key
+                # so that a text() construct can support unique parameters
+                existing = new_params[bind._orig_key]
+            except KeyError as err:
+                raise exc.ArgumentError(
+                    "This text() construct doesn't define a "
+                    "bound parameter named %r" % bind._orig_key
+                ) from err
+            else:
+                new_params[existing._orig_key] = bind
+
+        for key, value in names_to_values.items():
+            try:
+                existing = new_params[key]
+            except KeyError as err:
+                raise exc.ArgumentError(
+                    "This text() construct doesn't define a "
+                    "bound parameter named %r" % key
+                ) from err
+            else:
+                new_params[key] = existing._with_value(value, required=False)
+        return self
+
+    @util.preload_module("sqlalchemy.sql.selectable")
+    def columns(
+        self,
+        *cols: _ColumnExpressionArgument[Any],
+        **types: _TypeEngineArgument[Any],
+    ) -> TextualSelect:
+        r"""Turn this :class:`_expression.TextClause` object into a
+        :class:`_expression.TextualSelect`
+        object that serves the same role as a SELECT
+        statement.
+
+        The :class:`_expression.TextualSelect` is part of the
+        :class:`_expression.SelectBase`
+        hierarchy and can be embedded into another statement by using the
+        :meth:`_expression.TextualSelect.subquery` method to produce a
+        :class:`.Subquery`
+        object, which can then be SELECTed from.
+
+        This function essentially bridges the gap between an entirely
+        textual SELECT statement and the SQL expression language concept
+        of a "selectable"::
+
+            from sqlalchemy.sql import column, text
+
+            stmt = text("SELECT id, name FROM some_table")
+            stmt = stmt.columns(column("id"), column("name")).subquery("st")
+
+            stmt = (
+                select(mytable)
+                .select_from(mytable.join(stmt, mytable.c.name == stmt.c.name))
+                .where(stmt.c.id > 5)
+            )
+
+        Above, we pass a series of :func:`_expression.column` elements to the
+        :meth:`_expression.TextClause.columns` method positionally.  These
+        :func:`_expression.column`
+        elements now become first class elements upon the
+        :attr:`_expression.TextualSelect.selected_columns` column collection,
+        which then
+        become part of the :attr:`.Subquery.c` collection after
+        :meth:`_expression.TextualSelect.subquery` is invoked.
+
+        The column expressions we pass to
+        :meth:`_expression.TextClause.columns` may
+        also be typed; when we do so, these :class:`.TypeEngine` objects become
+        the effective return type of the column, so that SQLAlchemy's
+        result-set-processing systems may be used on the return values.
+        This is often needed for types such as date or boolean types, as well
+        as for unicode processing on some dialect configurations::
+
+            stmt = text("SELECT id, name, timestamp FROM some_table")
+            stmt = stmt.columns(
+                column("id", Integer),
+                column("name", Unicode),
+                column("timestamp", DateTime),
+            )
+
+            for id, name, timestamp in connection.execute(stmt):
+                print(id, name, timestamp)
+
+        As a shortcut to the above syntax, keyword arguments referring to
+        types alone may be used, if only type conversion is needed::
+
+            stmt = text("SELECT id, name, timestamp FROM some_table")
+            stmt = stmt.columns(id=Integer, name=Unicode, timestamp=DateTime)
+
+            for id, name, timestamp in connection.execute(stmt):
+                print(id, name, timestamp)
+
+        The positional form of :meth:`_expression.TextClause.columns`
+        also provides the
+        unique feature of **positional column targeting**, which is
+        particularly useful when using the ORM with complex textual queries. If
+        we specify the columns from our model to
+        :meth:`_expression.TextClause.columns`,
+        the result set will match to those columns positionally, meaning the
+        name or origin of the column in the textual SQL doesn't matter::
+
+            stmt = text(
+                "SELECT users.id, addresses.id, users.id, "
+                "users.name, addresses.email_address AS email "
+                "FROM users JOIN addresses ON users.id=addresses.user_id "
+                "WHERE users.id = 1"
+            ).columns(
+                User.id,
+                Address.id,
+                Address.user_id,
+                User.name,
+                Address.email_address,
+            )
+
+            query = (
+                session.query(User)
+                .from_statement(stmt)
+                .options(contains_eager(User.addresses))
+            )
+
+        The :meth:`_expression.TextClause.columns` method provides a direct
+        route to calling :meth:`_expression.FromClause.subquery` as well as
+        :meth:`_expression.SelectBase.cte`
+        against a textual SELECT statement::
+
+            stmt = stmt.columns(id=Integer, name=String).cte("st")
+
+            stmt = select(sometable).where(sometable.c.id == stmt.c.id)
+
+        :param \*cols: A series of :class:`_expression.ColumnElement` objects,
+         typically
+         :class:`_schema.Column` objects from a :class:`_schema.Table`
+         or ORM level
+         column-mapped attributes, representing a set of columns that this
+         textual string will SELECT from.
+
+        :param \**types: A mapping of string names to :class:`.TypeEngine`
+         type objects indicating the datatypes to use for names that are
+         SELECTed from the textual string.  Prefer to use the ``*cols``
+         argument as it also indicates positional ordering.
+
+        """
+        selectable = util.preloaded.sql_selectable
+
+        input_cols: List[NamedColumn[Any]] = [
+            coercions.expect(roles.LabeledColumnExprRole, col) for col in cols
+        ]
+
+        positional_input_cols = [
+            (
+                ColumnClause(col.key, types.pop(col.key))
+                if col.key in types
+                else col
+            )
+            for col in input_cols
+        ]
+        keyed_input_cols: List[NamedColumn[Any]] = [
+            ColumnClause(key, type_) for key, type_ in types.items()
+        ]
+
+        elem = selectable.TextualSelect.__new__(selectable.TextualSelect)
+        elem._init(
+            self,
+            positional_input_cols + keyed_input_cols,
+            positional=bool(positional_input_cols) and not keyed_input_cols,
+        )
+        return elem
+
+    @property
+    def type(self) -> TypeEngine[Any]:
+        return type_api.NULLTYPE
+
+    @property
+    def comparator(self):
+        # TODO: this seems wrong, it seems like we might not
+        # be using this method.
+        return self.type.comparator_factory(self)  # type: ignore
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[Any]]:
+        if against is operators.in_op:
+            return Grouping(self)
+        else:
+            return self
+
+
+class Null(SingletonConstant, roles.ConstExprRole[None], ColumnElement[None]):
+    """Represent the NULL keyword in a SQL statement.
+
+    :class:`.Null` is accessed as a constant via the
+    :func:`.null` function.
+
+    """
+
+    __visit_name__ = "null"
+
+    _traverse_internals: _TraverseInternalsType = []
+    _singleton: Null
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            return type_api.NULLTYPE
+
+    @classmethod
+    def _instance(cls) -> Null:
+        """Return a constant :class:`.Null` construct."""
+
+        return Null._singleton
+
+
+Null._create_singleton()
+
+
+class False_(
+    SingletonConstant, roles.ConstExprRole[bool], ColumnElement[bool]
+):
+    """Represent the ``false`` keyword, or equivalent, in a SQL statement.
+
+    :class:`.False_` is accessed as a constant via the
+    :func:`.false` function.
+
+    """
+
+    __visit_name__ = "false"
+    _traverse_internals: _TraverseInternalsType = []
+    _singleton: False_
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            return type_api.BOOLEANTYPE
+
+    def _negate(self) -> True_:
+        return True_._singleton
+
+    @classmethod
+    def _instance(cls) -> False_:
+        return False_._singleton
+
+
+False_._create_singleton()
+
+
+class True_(SingletonConstant, roles.ConstExprRole[bool], ColumnElement[bool]):
+    """Represent the ``true`` keyword, or equivalent, in a SQL statement.
+
+    :class:`.True_` is accessed as a constant via the
+    :func:`.true` function.
+
+    """
+
+    __visit_name__ = "true"
+
+    _traverse_internals: _TraverseInternalsType = []
+    _singleton: True_
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            return type_api.BOOLEANTYPE
+
+    def _negate(self) -> False_:
+        return False_._singleton
+
+    @classmethod
+    def _ifnone(
+        cls, other: Optional[ColumnElement[Any]]
+    ) -> ColumnElement[Any]:
+        if other is None:
+            return cls._instance()
+        else:
+            return other
+
+    @classmethod
+    def _instance(cls) -> True_:
+        return True_._singleton
+
+
+True_._create_singleton()
+
+
+class ClauseList(
+    roles.InElementRole,
+    roles.OrderByRole,
+    roles.ColumnsClauseRole,
+    roles.DMLColumnRole,
+    DQLDMLClauseElement,
+):
+    """Describe a list of clauses, separated by an operator.
+
+    By default, is comma-separated, such as a column listing.
+
+    """
+
+    __visit_name__ = "clauselist"
+
+    # this is used only by the ORM in a legacy use case for
+    # composite attributes
+    _is_clause_list = True
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("clauses", InternalTraversal.dp_clauseelement_list),
+        ("operator", InternalTraversal.dp_operator),
+    ]
+
+    clauses: List[ColumnElement[Any]]
+
+    def __init__(
+        self,
+        *clauses: _ColumnExpressionArgument[Any],
+        operator: OperatorType = operators.comma_op,
+        group: bool = True,
+        group_contents: bool = True,
+        _literal_as_text_role: Type[roles.SQLRole] = roles.WhereHavingRole,
+    ):
+        self.operator = operator
+        self.group = group
+        self.group_contents = group_contents
+        clauses_iterator: Iterable[_ColumnExpressionArgument[Any]] = clauses
+        text_converter_role: Type[roles.SQLRole] = _literal_as_text_role
+        self._text_converter_role = text_converter_role
+
+        if self.group_contents:
+            self.clauses = [
+                coercions.expect(
+                    text_converter_role, clause, apply_propagate_attrs=self
+                ).self_group(against=self.operator)
+                for clause in clauses_iterator
+            ]
+        else:
+            self.clauses = [
+                coercions.expect(
+                    text_converter_role, clause, apply_propagate_attrs=self
+                )
+                for clause in clauses_iterator
+            ]
+        self._is_implicitly_boolean = operators.is_boolean(self.operator)
+
+    @classmethod
+    def _construct_raw(
+        cls,
+        operator: OperatorType,
+        clauses: Optional[Sequence[ColumnElement[Any]]] = None,
+    ) -> ClauseList:
+        self = cls.__new__(cls)
+        self.clauses = list(clauses) if clauses else []
+        self.group = True
+        self.operator = operator
+        self.group_contents = True
+        self._is_implicitly_boolean = False
+        return self
+
+    def __iter__(self) -> Iterator[ColumnElement[Any]]:
+        return iter(self.clauses)
+
+    def __len__(self) -> int:
+        return len(self.clauses)
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return itertools.chain.from_iterable(
+            [elem._select_iterable for elem in self.clauses]
+        )
+
+    def append(self, clause):
+        if self.group_contents:
+            self.clauses.append(
+                coercions.expect(self._text_converter_role, clause).self_group(
+                    against=self.operator
+                )
+            )
+        else:
+            self.clauses.append(
+                coercions.expect(self._text_converter_role, clause)
+            )
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(itertools.chain(*[c._from_objects for c in self.clauses]))
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[Any]]:
+        if self.group and operators.is_precedent(self.operator, against):
+            return Grouping(self)
+        else:
+            return self
+
+
+class OperatorExpression(ColumnElement[_T]):
+    """base for expressions that contain an operator and operands
+
+    .. versionadded:: 2.0
+
+    """
+
+    operator: OperatorType
+    type: TypeEngine[_T]
+
+    group: bool = True
+
+    @property
+    def is_comparison(self):
+        return operators.is_comparison(self.operator)
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[_T]]:
+        if (
+            self.group
+            and operators.is_precedent(self.operator, against)
+            or (
+                # a negate against a non-boolean operator
+                # doesn't make too much sense but we should
+                # group for that
+                against is operators.inv
+                and not operators.is_boolean(self.operator)
+            )
+        ):
+            return Grouping(self)
+        else:
+            return self
+
+    @property
+    def _flattened_operator_clauses(
+        self,
+    ) -> typing_Tuple[ColumnElement[Any], ...]:
+        raise NotImplementedError()
+
+    @classmethod
+    def _construct_for_op(
+        cls,
+        left: ColumnElement[Any],
+        right: ColumnElement[Any],
+        op: OperatorType,
+        *,
+        type_: TypeEngine[_T],
+        negate: Optional[OperatorType] = None,
+        modifiers: Optional[Mapping[str, Any]] = None,
+    ) -> OperatorExpression[_T]:
+        if operators.is_associative(op):
+            assert (
+                negate is None
+            ), f"negate not supported for associative operator {op}"
+
+            multi = False
+            if getattr(
+                left, "operator", None
+            ) is op and type_._compare_type_affinity(left.type):
+                multi = True
+                left_flattened = left._flattened_operator_clauses
+            else:
+                left_flattened = (left,)
+
+            if getattr(
+                right, "operator", None
+            ) is op and type_._compare_type_affinity(right.type):
+                multi = True
+                right_flattened = right._flattened_operator_clauses
+            else:
+                right_flattened = (right,)
+
+            if multi:
+                return ExpressionClauseList._construct_for_list(
+                    op,
+                    type_,
+                    *(left_flattened + right_flattened),
+                )
+
+        if right._is_collection_aggregate:
+            negate = None
+
+        return BinaryExpression(
+            left, right, op, type_=type_, negate=negate, modifiers=modifiers
+        )
+
+
+class ExpressionClauseList(OperatorExpression[_T]):
+    """Describe a list of clauses, separated by an operator,
+    in a column expression context.
+
+    :class:`.ExpressionClauseList` differs from :class:`.ClauseList` in that
+    it represents a column-oriented DQL expression only, not an open ended
+    list of anything comma separated.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __visit_name__ = "expression_clauselist"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("clauses", InternalTraversal.dp_clauseelement_tuple),
+        ("operator", InternalTraversal.dp_operator),
+    ]
+
+    clauses: typing_Tuple[ColumnElement[Any], ...]
+
+    group: bool
+
+    def __init__(
+        self,
+        operator: OperatorType,
+        *clauses: _ColumnExpressionArgument[Any],
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+    ):
+        self.operator = operator
+
+        self.clauses = tuple(
+            coercions.expect(
+                roles.ExpressionElementRole, clause, apply_propagate_attrs=self
+            )
+            for clause in clauses
+        )
+        self._is_implicitly_boolean = operators.is_boolean(self.operator)
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+    @property
+    def _flattened_operator_clauses(
+        self,
+    ) -> typing_Tuple[ColumnElement[Any], ...]:
+        return self.clauses
+
+    def __iter__(self) -> Iterator[ColumnElement[Any]]:
+        return iter(self.clauses)
+
+    def __len__(self) -> int:
+        return len(self.clauses)
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return (self,)
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(itertools.chain(*[c._from_objects for c in self.clauses]))
+
+    def _append_inplace(self, clause: ColumnElement[Any]) -> None:
+        self.clauses += (clause,)
+
+    @classmethod
+    def _construct_for_list(
+        cls,
+        operator: OperatorType,
+        type_: TypeEngine[_T],
+        *clauses: ColumnElement[Any],
+        group: bool = True,
+    ) -> ExpressionClauseList[_T]:
+        self = cls.__new__(cls)
+        self.group = group
+        if group:
+            self.clauses = tuple(
+                c.self_group(against=operator) for c in clauses
+            )
+        else:
+            self.clauses = clauses
+        self.operator = operator
+        self.type = type_
+        return self
+
+    def _negate(self) -> Any:
+        grouped = self.self_group(against=operators.inv)
+        assert isinstance(grouped, ColumnElement)
+        return UnaryExpression(
+            grouped, operator=operators.inv, wraps_column_expression=True
+        )
+
+
+class BooleanClauseList(ExpressionClauseList[bool]):
+    __visit_name__ = "expression_clauselist"
+    inherit_cache = True
+
+    def __init__(self, *arg, **kw):
+        raise NotImplementedError(
+            "BooleanClauseList has a private constructor"
+        )
+
+    @classmethod
+    def _process_clauses_for_boolean(
+        cls,
+        operator: OperatorType,
+        continue_on: Any,
+        skip_on: Any,
+        clauses: Iterable[ColumnElement[Any]],
+    ) -> typing_Tuple[int, List[ColumnElement[Any]]]:
+        has_continue_on = None
+
+        convert_clauses = []
+
+        against = operators._asbool
+        lcc = 0
+
+        for clause in clauses:
+            if clause is continue_on:
+                # instance of continue_on, like and_(x, y, True, z), store it
+                # if we didn't find one already, we will use it if there
+                # are no other expressions here.
+                has_continue_on = clause
+            elif clause is skip_on:
+                # instance of skip_on, e.g. and_(x, y, False, z), cancels
+                # the rest out
+                convert_clauses = [clause]
+                lcc = 1
+                break
+            else:
+                if not lcc:
+                    lcc = 1
+                else:
+                    against = operator
+                    # technically this would be len(convert_clauses) + 1
+                    # however this only needs to indicate "greater than one"
+                    lcc = 2
+                convert_clauses.append(clause)
+
+        if not convert_clauses and has_continue_on is not None:
+            convert_clauses = [has_continue_on]
+            lcc = 1
+
+        return lcc, [c.self_group(against=against) for c in convert_clauses]
+
+    @classmethod
+    def _construct(
+        cls,
+        operator: OperatorType,
+        continue_on: Any,
+        skip_on: Any,
+        initial_clause: Any = _NoArg.NO_ARG,
+        *clauses: Any,
+        **kw: Any,
+    ) -> ColumnElement[Any]:
+        if initial_clause is _NoArg.NO_ARG:
+            # no elements period.  deprecated use case.  return an empty
+            # ClauseList construct that generates nothing unless it has
+            # elements added to it.
+            name = operator.__name__
+
+            util.warn_deprecated(
+                f"Invoking {name}() without arguments is deprecated, and "
+                f"will be disallowed in a future release.   For an empty "
+                f"""{name}() construct, use '{name}({
+                    'true()' if continue_on is True_._singleton else 'false()'
+                }, *args)' """
+                f"""or '{name}({
+                    'True' if continue_on is True_._singleton else 'False'
+                }, *args)'.""",
+                version="1.4",
+            )
+            return cls._construct_raw(operator)
+
+        lcc, convert_clauses = cls._process_clauses_for_boolean(
+            operator,
+            continue_on,
+            skip_on,
+            [
+                coercions.expect(roles.WhereHavingRole, clause)
+                for clause in util.coerce_generator_arg(
+                    (initial_clause,) + clauses
+                )
+            ],
+        )
+
+        if lcc > 1:
+            # multiple elements.  Return regular BooleanClauseList
+            # which will link elements against the operator.
+
+            flattened_clauses = itertools.chain.from_iterable(
+                (
+                    (c for c in to_flat._flattened_operator_clauses)
+                    if getattr(to_flat, "operator", None) is operator
+                    else (to_flat,)
+                )
+                for to_flat in convert_clauses
+            )
+
+            return cls._construct_raw(operator, flattened_clauses)  # type: ignore # noqa: E501
+        else:
+            assert lcc
+            # just one element.  return it as a single boolean element,
+            # not a list and discard the operator.
+            return convert_clauses[0]
+
+    @classmethod
+    def _construct_for_whereclause(
+        cls, clauses: Iterable[ColumnElement[Any]]
+    ) -> Optional[ColumnElement[bool]]:
+        operator, continue_on, skip_on = (
+            operators.and_,
+            True_._singleton,
+            False_._singleton,
+        )
+
+        lcc, convert_clauses = cls._process_clauses_for_boolean(
+            operator,
+            continue_on,
+            skip_on,
+            clauses,  # these are assumed to be coerced already
+        )
+
+        if lcc > 1:
+            # multiple elements.  Return regular BooleanClauseList
+            # which will link elements against the operator.
+            return cls._construct_raw(operator, convert_clauses)
+        elif lcc == 1:
+            # just one element.  return it as a single boolean element,
+            # not a list and discard the operator.
+            return convert_clauses[0]
+        else:
+            return None
+
+    @classmethod
+    def _construct_raw(
+        cls,
+        operator: OperatorType,
+        clauses: Optional[Sequence[ColumnElement[Any]]] = None,
+    ) -> BooleanClauseList:
+        self = cls.__new__(cls)
+        self.clauses = tuple(clauses) if clauses else ()
+        self.group = True
+        self.operator = operator
+        self.type = type_api.BOOLEANTYPE
+        self._is_implicitly_boolean = True
+        return self
+
+    @classmethod
+    def and_(
+        cls,
+        initial_clause: Union[
+            Literal[True], _ColumnExpressionArgument[bool], _NoArg
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionArgument[bool],
+    ) -> ColumnElement[bool]:
+        r"""Produce a conjunction of expressions joined by ``AND``.
+
+        See :func:`_sql.and_` for full documentation.
+        """
+        return cls._construct(
+            operators.and_,
+            True_._singleton,
+            False_._singleton,
+            initial_clause,
+            *clauses,
+        )
+
+    @classmethod
+    def or_(
+        cls,
+        initial_clause: Union[
+            Literal[False], _ColumnExpressionArgument[bool], _NoArg
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionArgument[bool],
+    ) -> ColumnElement[bool]:
+        """Produce a conjunction of expressions joined by ``OR``.
+
+        See :func:`_sql.or_` for full documentation.
+        """
+        return cls._construct(
+            operators.or_,
+            False_._singleton,
+            True_._singleton,
+            initial_clause,
+            *clauses,
+        )
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return (self,)
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[bool]]:
+        if not self.clauses:
+            return self
+        else:
+            return super().self_group(against=against)
+
+
+and_ = BooleanClauseList.and_
+or_ = BooleanClauseList.or_
+
+
+class Tuple(ClauseList, ColumnElement[typing_Tuple[Any, ...]]):
+    """Represent a SQL tuple."""
+
+    __visit_name__ = "tuple"
+
+    _traverse_internals: _TraverseInternalsType = (
+        ClauseList._traverse_internals + []
+    )
+
+    type: TupleType
+
+    @util.preload_module("sqlalchemy.sql.sqltypes")
+    def __init__(
+        self,
+        *clauses: _ColumnExpressionArgument[Any],
+        types: Optional[Sequence[_TypeEngineArgument[Any]]] = None,
+    ):
+        sqltypes = util.preloaded.sql_sqltypes
+
+        if types is None:
+            init_clauses: List[ColumnElement[Any]] = [
+                coercions.expect(roles.ExpressionElementRole, c)
+                for c in clauses
+            ]
+        else:
+            if len(types) != len(clauses):
+                raise exc.ArgumentError(
+                    "Wrong number of elements for %d-tuple: %r "
+                    % (len(types), clauses)
+                )
+            init_clauses = [
+                coercions.expect(
+                    roles.ExpressionElementRole,
+                    c,
+                    type_=typ if not typ._isnull else None,
+                )
+                for typ, c in zip(types, clauses)
+            ]
+
+        self.type = sqltypes.TupleType(*[arg.type for arg in init_clauses])
+        super().__init__(*init_clauses)
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        return (self,)
+
+    def _bind_param(self, operator, obj, type_=None, expanding=False):
+        if expanding:
+            return BindParameter(
+                None,
+                value=obj,
+                _compared_to_operator=operator,
+                unique=True,
+                expanding=True,
+                type_=type_,
+                _compared_to_type=self.type,
+            )
+        else:
+            return Tuple(
+                *[
+                    BindParameter(
+                        None,
+                        o,
+                        _compared_to_operator=operator,
+                        _compared_to_type=compared_to_type,
+                        unique=True,
+                        type_=type_,
+                    )
+                    for o, compared_to_type in zip(obj, self.type.types)
+                ]
+            )
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        # Tuple is parenthesized by definition.
+        return self
+
+
+class Case(ColumnElement[_T]):
+    """Represent a ``CASE`` expression.
+
+    :class:`.Case` is produced using the :func:`.case` factory function,
+    as in::
+
+        from sqlalchemy import case
+
+        stmt = select(users_table).where(
+            case(
+                (users_table.c.name == "wendy", "W"),
+                (users_table.c.name == "jack", "J"),
+                else_="E",
+            )
+        )
+
+    Details on :class:`.Case` usage is at :func:`.case`.
+
+    .. seealso::
+
+        :func:`.case`
+
+    """
+
+    __visit_name__ = "case"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("value", InternalTraversal.dp_clauseelement),
+        ("whens", InternalTraversal.dp_clauseelement_tuples),
+        ("else_", InternalTraversal.dp_clauseelement),
+    ]
+
+    # for case(), the type is derived from the whens.  so for the moment
+    # users would have to cast() the case to get a specific type
+
+    whens: List[typing_Tuple[ColumnElement[bool], ColumnElement[_T]]]
+    else_: Optional[ColumnElement[_T]]
+    value: Optional[ColumnElement[Any]]
+
+    def __init__(
+        self,
+        *whens: Union[
+            typing_Tuple[_ColumnExpressionArgument[bool], Any],
+            Mapping[Any, Any],
+        ],
+        value: Optional[Any] = None,
+        else_: Optional[Any] = None,
+    ):
+        new_whens: Iterable[Any] = coercions._expression_collection_was_a_list(
+            "whens", "case", whens
+        )
+        try:
+            new_whens = util.dictlike_iteritems(new_whens)
+        except TypeError:
+            pass
+
+        self.whens = [
+            (
+                coercions.expect(
+                    roles.ExpressionElementRole,
+                    c,
+                    apply_propagate_attrs=self,
+                ).self_group(),
+                coercions.expect(roles.ExpressionElementRole, r),
+            )
+            for (c, r) in new_whens
+        ]
+
+        if value is None:
+            self.value = None
+        else:
+            self.value = coercions.expect(roles.ExpressionElementRole, value)
+
+        if else_ is not None:
+            self.else_ = coercions.expect(roles.ExpressionElementRole, else_)
+        else:
+            self.else_ = None
+
+        type_ = next(
+            (
+                then.type
+                # Iterate `whens` in reverse to match previous behaviour
+                # where type of final element took priority
+                for *_, then in reversed(self.whens)
+                if not then.type._isnull
+            ),
+            self.else_.type if self.else_ is not None else type_api.NULLTYPE,
+        )
+        self.type = cast(_T, type_)
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(
+            itertools.chain(*[x._from_objects for x in self.get_children()])
+        )
+
+
+class Cast(WrapsColumnExpression[_T]):
+    """Represent a ``CAST`` expression.
+
+    :class:`.Cast` is produced using the :func:`.cast` factory function,
+    as in::
+
+        from sqlalchemy import cast, Numeric
+
+        stmt = select(cast(product_table.c.unit_price, Numeric(10, 4)))
+
+    Details on :class:`.Cast` usage is at :func:`.cast`.
+
+    .. seealso::
+
+        :ref:`tutorial_casts`
+
+        :func:`.cast`
+
+        :func:`.try_cast`
+
+        :func:`.type_coerce` - an alternative to CAST that coerces the type
+        on the Python side only, which is often sufficient to generate the
+        correct SQL and data coercion.
+
+    """
+
+    __visit_name__ = "cast"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("clause", InternalTraversal.dp_clauseelement),
+        ("type", InternalTraversal.dp_type),
+    ]
+
+    clause: ColumnElement[Any]
+    type: TypeEngine[_T]
+    typeclause: TypeClause
+
+    def __init__(
+        self,
+        expression: _ColumnExpressionArgument[Any],
+        type_: _TypeEngineArgument[_T],
+    ):
+        self.type = type_api.to_instance(type_)
+        self.clause = coercions.expect(
+            roles.ExpressionElementRole,
+            expression,
+            type_=self.type,
+            apply_propagate_attrs=self,
+        )
+        self.typeclause = TypeClause(self.type)
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.clause._from_objects
+
+    @property
+    def wrapped_column_expression(self):
+        return self.clause
+
+
+class TryCast(Cast[_T]):
+    """Represent a TRY_CAST expression.
+
+    Details on :class:`.TryCast` usage is at :func:`.try_cast`.
+
+    .. seealso::
+
+        :func:`.try_cast`
+
+        :ref:`tutorial_casts`
+    """
+
+    __visit_name__ = "try_cast"
+    inherit_cache = True
+
+
+class TypeCoerce(WrapsColumnExpression[_T]):
+    """Represent a Python-side type-coercion wrapper.
+
+    :class:`.TypeCoerce` supplies the :func:`_expression.type_coerce`
+    function; see that function for usage details.
+
+    .. seealso::
+
+        :func:`_expression.type_coerce`
+
+        :func:`.cast`
+
+    """
+
+    __visit_name__ = "type_coerce"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("clause", InternalTraversal.dp_clauseelement),
+        ("type", InternalTraversal.dp_type),
+    ]
+
+    clause: ColumnElement[Any]
+    type: TypeEngine[_T]
+
+    def __init__(
+        self,
+        expression: _ColumnExpressionArgument[Any],
+        type_: _TypeEngineArgument[_T],
+    ):
+        self.type = type_api.to_instance(type_)
+        self.clause = coercions.expect(
+            roles.ExpressionElementRole,
+            expression,
+            type_=self.type,
+            apply_propagate_attrs=self,
+        )
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.clause._from_objects
+
+    @HasMemoized.memoized_attribute
+    def typed_expression(self):
+        if isinstance(self.clause, BindParameter):
+            bp = self.clause._clone()
+            bp.type = self.type
+            return bp
+        else:
+            return self.clause
+
+    @property
+    def wrapped_column_expression(self):
+        return self.clause
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> TypeCoerce[_T]:
+        grouped = self.clause.self_group(against=against)
+        if grouped is not self.clause:
+            return TypeCoerce(grouped, self.type)
+        else:
+            return self
+
+
+class Extract(ColumnElement[int]):
+    """Represent a SQL EXTRACT clause, ``extract(field FROM expr)``."""
+
+    __visit_name__ = "extract"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("expr", InternalTraversal.dp_clauseelement),
+        ("field", InternalTraversal.dp_string),
+    ]
+
+    expr: ColumnElement[Any]
+    field: str
+
+    def __init__(self, field: str, expr: _ColumnExpressionArgument[Any]):
+        self.type = type_api.INTEGERTYPE
+        self.field = field
+        self.expr = coercions.expect(roles.ExpressionElementRole, expr)
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.expr._from_objects
+
+
+class _label_reference(ColumnElement[_T]):
+    """Wrap a column expression as it appears in a 'reference' context.
+
+    This expression is any that includes an _order_by_label_element,
+    which is a Label, or a DESC / ASC construct wrapping a Label.
+
+    The production of _label_reference() should occur when an expression
+    is added to this context; this includes the ORDER BY or GROUP BY of a
+    SELECT statement, as well as a few other places, such as the ORDER BY
+    within an OVER clause.
+
+    """
+
+    __visit_name__ = "label_reference"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement)
+    ]
+
+    element: ColumnElement[_T]
+
+    def __init__(self, element: ColumnElement[_T]):
+        self.element = element
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return []
+
+
+class _textual_label_reference(ColumnElement[Any]):
+    __visit_name__ = "textual_label_reference"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_string)
+    ]
+
+    def __init__(self, element: str):
+        self.element = element
+
+    @util.memoized_property
+    def _text_clause(self) -> TextClause:
+        return TextClause(self.element)
+
+
+class UnaryExpression(ColumnElement[_T]):
+    """Define a 'unary' expression.
+
+    A unary expression has a single column expression
+    and an operator.  The operator can be placed on the left
+    (where it is called the 'operator') or right (where it is called the
+    'modifier') of the column expression.
+
+    :class:`.UnaryExpression` is the basis for several unary operators
+    including those used by :func:`.desc`, :func:`.asc`, :func:`.distinct`,
+    :func:`.nulls_first` and :func:`.nulls_last`.
+
+    """
+
+    __visit_name__ = "unary"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("operator", InternalTraversal.dp_operator),
+        ("modifier", InternalTraversal.dp_operator),
+    ]
+
+    element: ClauseElement
+
+    def __init__(
+        self,
+        element: ColumnElement[Any],
+        operator: Optional[OperatorType] = None,
+        modifier: Optional[OperatorType] = None,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        wraps_column_expression: bool = False,
+    ):
+        self.operator = operator
+        self.modifier = modifier
+        self._propagate_attrs = element._propagate_attrs
+        self.element = element.self_group(
+            against=self.operator or self.modifier
+        )
+
+        # if type is None, we get NULLTYPE, which is our _T.  But I don't
+        # know how to get the overloads to express that correctly
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+        self.wraps_column_expression = wraps_column_expression
+
+    @classmethod
+    def _create_nulls_first(
+        cls,
+        column: _ColumnExpressionArgument[_T],
+    ) -> UnaryExpression[_T]:
+        return UnaryExpression(
+            coercions.expect(roles.ByOfRole, column),
+            modifier=operators.nulls_first_op,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_nulls_last(
+        cls,
+        column: _ColumnExpressionArgument[_T],
+    ) -> UnaryExpression[_T]:
+        return UnaryExpression(
+            coercions.expect(roles.ByOfRole, column),
+            modifier=operators.nulls_last_op,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_desc(
+        cls, column: _ColumnExpressionOrStrLabelArgument[_T]
+    ) -> UnaryExpression[_T]:
+        return UnaryExpression(
+            coercions.expect(roles.ByOfRole, column),
+            modifier=operators.desc_op,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_asc(
+        cls,
+        column: _ColumnExpressionOrStrLabelArgument[_T],
+    ) -> UnaryExpression[_T]:
+        return UnaryExpression(
+            coercions.expect(roles.ByOfRole, column),
+            modifier=operators.asc_op,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_distinct(
+        cls,
+        expr: _ColumnExpressionArgument[_T],
+    ) -> UnaryExpression[_T]:
+        col_expr: ColumnElement[_T] = coercions.expect(
+            roles.ExpressionElementRole, expr
+        )
+        return UnaryExpression(
+            col_expr,
+            operator=operators.distinct_op,
+            type_=col_expr.type,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_bitwise_not(
+        cls,
+        expr: _ColumnExpressionArgument[_T],
+    ) -> UnaryExpression[_T]:
+        col_expr: ColumnElement[_T] = coercions.expect(
+            roles.ExpressionElementRole, expr
+        )
+        return UnaryExpression(
+            col_expr,
+            operator=operators.bitwise_not_op,
+            type_=col_expr.type,
+            wraps_column_expression=False,
+        )
+
+    @property
+    def _order_by_label_element(self) -> Optional[Label[Any]]:
+        if operators.is_order_by_modifier(self.modifier):
+            return self.element._order_by_label_element
+        else:
+            return None
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.element._from_objects
+
+    def _negate(self):
+        if self.type._type_affinity is type_api.BOOLEANTYPE._type_affinity:
+            return UnaryExpression(
+                self.self_group(against=operators.inv),
+                operator=operators.inv,
+                type_=type_api.BOOLEANTYPE,
+                wraps_column_expression=self.wraps_column_expression,
+            )
+        else:
+            return ClauseElement._negate(self)
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[_T]]:
+        if self.operator and operators.is_precedent(self.operator, against):
+            return Grouping(self)
+        else:
+            return self
+
+
+class CollectionAggregate(UnaryExpression[_T]):
+    """Forms the basis for right-hand collection operator modifiers
+    ANY and ALL.
+
+    The ANY and ALL keywords are available in different ways on different
+    backends.  On PostgreSQL, they only work for an ARRAY type.  On
+    MySQL, they only work for subqueries.
+
+    """
+
+    inherit_cache = True
+    _is_collection_aggregate = True
+
+    @classmethod
+    def _create_any(
+        cls, expr: _ColumnExpressionArgument[_T]
+    ) -> CollectionAggregate[bool]:
+        col_expr: ColumnElement[_T] = coercions.expect(
+            roles.ExpressionElementRole,
+            expr,
+        )
+        col_expr = col_expr.self_group()
+        return CollectionAggregate(
+            col_expr,
+            operator=operators.any_op,
+            type_=type_api.BOOLEANTYPE,
+            wraps_column_expression=False,
+        )
+
+    @classmethod
+    def _create_all(
+        cls, expr: _ColumnExpressionArgument[_T]
+    ) -> CollectionAggregate[bool]:
+        col_expr: ColumnElement[_T] = coercions.expect(
+            roles.ExpressionElementRole,
+            expr,
+        )
+        col_expr = col_expr.self_group()
+        return CollectionAggregate(
+            col_expr,
+            operator=operators.all_op,
+            type_=type_api.BOOLEANTYPE,
+            wraps_column_expression=False,
+        )
+
+    # operate and reverse_operate are hardwired to
+    # dispatch onto the type comparator directly, so that we can
+    # ensure "reversed" behavior.
+    def operate(self, op, *other, **kwargs):
+        if not operators.is_comparison(op):
+            raise exc.ArgumentError(
+                "Only comparison operators may be used with ANY/ALL"
+            )
+        kwargs["reverse"] = True
+        return self.comparator.operate(operators.mirror(op), *other, **kwargs)
+
+    def reverse_operate(self, op, other, **kwargs):
+        # comparison operators should never call reverse_operate
+        assert not operators.is_comparison(op)
+        raise exc.ArgumentError(
+            "Only comparison operators may be used with ANY/ALL"
+        )
+
+
+class AsBoolean(WrapsColumnExpression[bool], UnaryExpression[bool]):
+    inherit_cache = True
+
+    def __init__(self, element, operator, negate):
+        self.element = element
+        self.type = type_api.BOOLEANTYPE
+        self.operator = operator
+        self.negate = negate
+        self.modifier = None
+        self.wraps_column_expression = True
+        self._is_implicitly_boolean = element._is_implicitly_boolean
+
+    @property
+    def wrapped_column_expression(self):
+        return self.element
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        return self
+
+    def _negate(self):
+        if isinstance(self.element, (True_, False_)):
+            return self.element._negate()
+        else:
+            return AsBoolean(self.element, self.negate, self.operator)
+
+
+class BinaryExpression(OperatorExpression[_T]):
+    """Represent an expression that is ``LEFT <operator> RIGHT``.
+
+    A :class:`.BinaryExpression` is generated automatically
+    whenever two column expressions are used in a Python binary expression:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy.sql import column
+        >>> column("a") + column("b")
+        <sqlalchemy.sql.expression.BinaryExpression object at 0x101029dd0>
+        >>> print(column("a") + column("b"))
+        {printsql}a + b
+
+    """
+
+    __visit_name__ = "binary"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("left", InternalTraversal.dp_clauseelement),
+        ("right", InternalTraversal.dp_clauseelement),
+        ("operator", InternalTraversal.dp_operator),
+        ("negate", InternalTraversal.dp_operator),
+        ("modifiers", InternalTraversal.dp_plain_dict),
+        (
+            "type",
+            InternalTraversal.dp_type,
+        ),
+    ]
+
+    _cache_key_traversal = [
+        ("left", InternalTraversal.dp_clauseelement),
+        ("right", InternalTraversal.dp_clauseelement),
+        ("operator", InternalTraversal.dp_operator),
+        ("modifiers", InternalTraversal.dp_plain_dict),
+        # "type" affects JSON CAST operators, so while redundant in most cases,
+        # is needed for that one
+        (
+            "type",
+            InternalTraversal.dp_type,
+        ),
+    ]
+
+    _is_implicitly_boolean = True
+    """Indicates that any database will know this is a boolean expression
+    even if the database does not have an explicit boolean datatype.
+
+    """
+
+    modifiers: Optional[Mapping[str, Any]]
+
+    left: ColumnElement[Any]
+    right: ColumnElement[Any]
+
+    def __init__(
+        self,
+        left: ColumnElement[Any],
+        right: ColumnElement[Any],
+        operator: OperatorType,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        negate: Optional[OperatorType] = None,
+        modifiers: Optional[Mapping[str, Any]] = None,
+    ):
+        # allow compatibility with libraries that
+        # refer to BinaryExpression directly and pass strings
+        if isinstance(operator, str):
+            operator = operators.custom_op(operator)
+        self._orig = (left.__hash__(), right.__hash__())
+        self._propagate_attrs = left._propagate_attrs or right._propagate_attrs
+        self.left = left.self_group(against=operator)
+        self.right = right.self_group(against=operator)
+        self.operator = operator
+
+        # if type is None, we get NULLTYPE, which is our _T.  But I don't
+        # know how to get the overloads to express that correctly
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+        self.negate = negate
+        self._is_implicitly_boolean = operators.is_boolean(operator)
+
+        if modifiers is None:
+            self.modifiers = {}
+        else:
+            self.modifiers = modifiers
+
+    @property
+    def _flattened_operator_clauses(
+        self,
+    ) -> typing_Tuple[ColumnElement[Any], ...]:
+        return (self.left, self.right)
+
+    def __bool__(self):
+        """Implement Python-side "bool" for BinaryExpression as a
+        simple "identity" check for the left and right attributes,
+        if the operator is "eq" or "ne".  Otherwise the expression
+        continues to not support "bool" like all other column expressions.
+
+        The rationale here is so that ColumnElement objects can be hashable.
+        What?  Well, suppose you do this::
+
+            c1, c2 = column("x"), column("y")
+            s1 = set([c1, c2])
+
+        We do that **a lot**, columns inside of sets is an extremely basic
+        thing all over the ORM for example.
+
+        So what happens if we do this? ::
+
+            c1 in s1
+
+        Hashing means it will normally use ``__hash__()`` of the object,
+        but in case of hash collision, it's going to also do ``c1 == c1``
+        and/or ``c1 == c2`` inside.  Those operations need to return a
+        True/False value.   But because we override ``==`` and ``!=``, they're
+        going to get a BinaryExpression.  Hence we implement ``__bool__`` here
+        so that these comparisons behave in this particular context mostly
+        like regular object comparisons.  Thankfully Python is OK with
+        that!  Otherwise we'd have to use special set classes for columns
+        (which we used to do, decades ago).
+
+        """
+        if self.operator in (operators.eq, operators.ne):
+            # this is using the eq/ne operator given int hash values,
+            # rather than Operator, so that "bool" can be based on
+            # identity
+            return self.operator(*self._orig)  # type: ignore
+        else:
+            raise TypeError("Boolean value of this clause is not defined")
+
+    if typing.TYPE_CHECKING:
+
+        def __invert__(
+            self: BinaryExpression[_T],
+        ) -> BinaryExpression[_T]: ...
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.left._from_objects + self.right._from_objects
+
+    def _negate(self):
+        if self.negate is not None:
+            return BinaryExpression(
+                self.left,
+                self.right._negate_in_binary(self.negate, self.operator),
+                self.negate,
+                negate=self.operator,
+                type_=self.type,
+                modifiers=self.modifiers,
+            )
+        else:
+            return self.self_group()._negate()
+
+
+class Slice(ColumnElement[Any]):
+    """Represent SQL for a Python array-slice object.
+
+    This is not a specific SQL construct at this level, but
+    may be interpreted by specific dialects, e.g. PostgreSQL.
+
+    """
+
+    __visit_name__ = "slice"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("start", InternalTraversal.dp_clauseelement),
+        ("stop", InternalTraversal.dp_clauseelement),
+        ("step", InternalTraversal.dp_clauseelement),
+    ]
+
+    def __init__(self, start, stop, step, _name=None):
+        self.start = coercions.expect(
+            roles.ExpressionElementRole,
+            start,
+            name=_name,
+            type_=type_api.INTEGERTYPE,
+        )
+        self.stop = coercions.expect(
+            roles.ExpressionElementRole,
+            stop,
+            name=_name,
+            type_=type_api.INTEGERTYPE,
+        )
+        self.step = coercions.expect(
+            roles.ExpressionElementRole,
+            step,
+            name=_name,
+            type_=type_api.INTEGERTYPE,
+        )
+        self.type = type_api.NULLTYPE
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        assert against is operator.getitem
+        return self
+
+
+class IndexExpression(BinaryExpression[Any]):
+    """Represent the class of expressions that are like an "index"
+    operation."""
+
+    inherit_cache = True
+
+
+class GroupedElement(DQLDMLClauseElement):
+    """Represent any parenthesized expression"""
+
+    __visit_name__ = "grouping"
+
+    element: ClauseElement
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        return self
+
+    def _ungroup(self):
+        return self.element._ungroup()
+
+
+class Grouping(GroupedElement, ColumnElement[_T]):
+    """Represent a grouping within a column expression"""
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("type", InternalTraversal.dp_type),
+    ]
+
+    _cache_key_traversal = [
+        ("element", InternalTraversal.dp_clauseelement),
+    ]
+
+    element: Union[TextClause, ClauseList, ColumnElement[_T]]
+
+    def __init__(
+        self, element: Union[TextClause, ClauseList, ColumnElement[_T]]
+    ):
+        self.element = element
+
+        # nulltype assignment issue
+        self.type = getattr(element, "type", type_api.NULLTYPE)  # type: ignore
+        self._propagate_attrs = element._propagate_attrs
+
+    def _with_binary_element_type(self, type_):
+        return self.__class__(self.element._with_binary_element_type(type_))
+
+    @util.memoized_property
+    def _is_implicitly_boolean(self):
+        return self.element._is_implicitly_boolean
+
+    @util.non_memoized_property
+    def _tq_label(self) -> Optional[str]:
+        return (
+            getattr(self.element, "_tq_label", None) or self._anon_name_label
+        )
+
+    @util.non_memoized_property
+    def _proxies(self) -> List[ColumnElement[Any]]:
+        if isinstance(self.element, ColumnElement):
+            return [self.element]
+        else:
+            return []
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.element._from_objects
+
+    def __getattr__(self, attr):
+        return getattr(self.element, attr)
+
+    def __getstate__(self):
+        return {"element": self.element, "type": self.type}
+
+    def __setstate__(self, state):
+        self.element = state["element"]
+        self.type = state["type"]
+
+    if TYPE_CHECKING:
+
+        def self_group(
+            self, against: Optional[OperatorType] = None
+        ) -> Self: ...
+
+
+class _OverrideBinds(Grouping[_T]):
+    """used by cache_key->_apply_params_to_element to allow compilation /
+    execution of a SQL element that's been cached, using an alternate set of
+    bound parameter values.
+
+    This is used by the ORM to swap new parameter values into expressions
+    that are embedded into loader options like with_expression(),
+    selectinload().  Previously, this task was accomplished using the
+    .params() method which would perform a deep-copy instead.  This deep
+    copy proved to be too expensive for more complex expressions.
+
+    See #11085
+
+    """
+
+    __visit_name__ = "override_binds"
+
+    def __init__(
+        self,
+        element: ColumnElement[_T],
+        bindparams: Sequence[BindParameter[Any]],
+        replaces_params: Sequence[BindParameter[Any]],
+    ):
+        self.element = element
+        self.translate = {
+            k.key: v.value for k, v in zip(replaces_params, bindparams)
+        }
+
+    def _gen_cache_key(
+        self, anon_map: anon_map, bindparams: List[BindParameter[Any]]
+    ) -> Optional[typing_Tuple[Any, ...]]:
+        """generate a cache key for the given element, substituting its bind
+        values for the translation values present."""
+
+        existing_bps: List[BindParameter[Any]] = []
+        ck = self.element._gen_cache_key(anon_map, existing_bps)
+
+        bindparams.extend(
+            (
+                bp._with_value(
+                    self.translate[bp.key], maintain_key=True, required=False
+                )
+                if bp.key in self.translate
+                else bp
+            )
+            for bp in existing_bps
+        )
+
+        return ck
+
+
+class _OverRange(Enum):
+    RANGE_UNBOUNDED = 0
+    RANGE_CURRENT = 1
+
+
+RANGE_UNBOUNDED = _OverRange.RANGE_UNBOUNDED
+RANGE_CURRENT = _OverRange.RANGE_CURRENT
+
+_IntOrRange = Union[int, _OverRange]
+
+
+class Over(ColumnElement[_T]):
+    """Represent an OVER clause.
+
+    This is a special operator against a so-called
+    "window" function, as well as any aggregate function,
+    which produces results relative to the result set
+    itself.  Most modern SQL backends now support window functions.
+
+    """
+
+    __visit_name__ = "over"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("order_by", InternalTraversal.dp_clauseelement),
+        ("partition_by", InternalTraversal.dp_clauseelement),
+        ("range_", InternalTraversal.dp_plain_obj),
+        ("rows", InternalTraversal.dp_plain_obj),
+    ]
+
+    order_by: Optional[ClauseList] = None
+    partition_by: Optional[ClauseList] = None
+
+    element: ColumnElement[_T]
+    """The underlying expression object to which this :class:`.Over`
+    object refers."""
+
+    range_: Optional[typing_Tuple[_IntOrRange, _IntOrRange]]
+    rows: Optional[typing_Tuple[_IntOrRange, _IntOrRange]]
+
+    def __init__(
+        self,
+        element: ColumnElement[_T],
+        partition_by: Optional[_ByArgument] = None,
+        order_by: Optional[_ByArgument] = None,
+        range_: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+        rows: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+    ):
+        self.element = element
+        if order_by is not None:
+            self.order_by = ClauseList(
+                *util.to_list(order_by), _literal_as_text_role=roles.ByOfRole
+            )
+        if partition_by is not None:
+            self.partition_by = ClauseList(
+                *util.to_list(partition_by),
+                _literal_as_text_role=roles.ByOfRole,
+            )
+
+        if range_:
+            self.range_ = self._interpret_range(range_)
+            if rows:
+                raise exc.ArgumentError(
+                    "'range_' and 'rows' are mutually exclusive"
+                )
+            else:
+                self.rows = None
+        elif rows:
+            self.rows = self._interpret_range(rows)
+            self.range_ = None
+        else:
+            self.rows = self.range_ = None
+
+    def __reduce__(self):
+        return self.__class__, (
+            self.element,
+            self.partition_by,
+            self.order_by,
+            self.range_,
+            self.rows,
+        )
+
+    def _interpret_range(
+        self,
+        range_: typing_Tuple[Optional[_IntOrRange], Optional[_IntOrRange]],
+    ) -> typing_Tuple[_IntOrRange, _IntOrRange]:
+        if not isinstance(range_, tuple) or len(range_) != 2:
+            raise exc.ArgumentError("2-tuple expected for range/rows")
+
+        r0, r1 = range_
+
+        lower: _IntOrRange
+        upper: _IntOrRange
+
+        if r0 is None:
+            lower = RANGE_UNBOUNDED
+        elif isinstance(r0, _OverRange):
+            lower = r0
+        else:
+            try:
+                lower = int(r0)
+            except ValueError as err:
+                raise exc.ArgumentError(
+                    "Integer or None expected for range value"
+                ) from err
+            else:
+                if lower == 0:
+                    lower = RANGE_CURRENT
+
+        if r1 is None:
+            upper = RANGE_UNBOUNDED
+        elif isinstance(r1, _OverRange):
+            upper = r1
+        else:
+            try:
+                upper = int(r1)
+            except ValueError as err:
+                raise exc.ArgumentError(
+                    "Integer or None expected for range value"
+                ) from err
+            else:
+                if upper == 0:
+                    upper = RANGE_CURRENT
+
+        return lower, upper
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            return self.element.type
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(
+            itertools.chain(
+                *[
+                    c._from_objects
+                    for c in (self.element, self.partition_by, self.order_by)
+                    if c is not None
+                ]
+            )
+        )
+
+
+class WithinGroup(ColumnElement[_T]):
+    """Represent a WITHIN GROUP (ORDER BY) clause.
+
+    This is a special operator against so-called
+    "ordered set aggregate" and "hypothetical
+    set aggregate" functions, including ``percentile_cont()``,
+    ``rank()``, ``dense_rank()``, etc.
+
+    It's supported only by certain database backends, such as PostgreSQL,
+    Oracle Database and MS SQL Server.
+
+    The :class:`.WithinGroup` construct extracts its type from the
+    method :meth:`.FunctionElement.within_group_type`.  If this returns
+    ``None``, the function's ``.type`` is used.
+
+    """
+
+    __visit_name__ = "withingroup"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("order_by", InternalTraversal.dp_clauseelement),
+    ]
+
+    order_by: Optional[ClauseList] = None
+
+    def __init__(
+        self,
+        element: Union[FunctionElement[_T], FunctionFilter[_T]],
+        *order_by: _ColumnExpressionArgument[Any],
+    ):
+        self.element = element
+        if order_by is not None:
+            self.order_by = ClauseList(
+                *util.to_list(order_by), _literal_as_text_role=roles.ByOfRole
+            )
+
+    def __reduce__(self):
+        return self.__class__, (self.element,) + (
+            tuple(self.order_by) if self.order_by is not None else ()
+        )
+
+    def over(
+        self,
+        *,
+        partition_by: Optional[_ByArgument] = None,
+        order_by: Optional[_ByArgument] = None,
+        rows: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+        range_: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+    ) -> Over[_T]:
+        """Produce an OVER clause against this :class:`.WithinGroup`
+        construct.
+
+        This function has the same signature as that of
+        :meth:`.FunctionElement.over`.
+
+        """
+        return Over(
+            self,
+            partition_by=partition_by,
+            order_by=order_by,
+            range_=range_,
+            rows=rows,
+        )
+
+    @overload
+    def filter(self) -> Self: ...
+
+    @overload
+    def filter(
+        self,
+        __criterion0: _ColumnExpressionArgument[bool],
+        *criterion: _ColumnExpressionArgument[bool],
+    ) -> FunctionFilter[_T]: ...
+
+    def filter(
+        self, *criterion: _ColumnExpressionArgument[bool]
+    ) -> Union[Self, FunctionFilter[_T]]:
+        """Produce a FILTER clause against this function."""
+        if not criterion:
+            return self
+        return FunctionFilter(self, *criterion)
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            wgt = self.element.within_group_type(self)
+            if wgt is not None:
+                return wgt
+            else:
+                return self.element.type
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(
+            itertools.chain(
+                *[
+                    c._from_objects
+                    for c in (self.element, self.order_by)
+                    if c is not None
+                ]
+            )
+        )
+
+
+class FunctionFilter(Generative, ColumnElement[_T]):
+    """Represent a function FILTER clause.
+
+    This is a special operator against aggregate and window functions,
+    which controls which rows are passed to it.
+    It's supported only by certain database backends.
+
+    Invocation of :class:`.FunctionFilter` is via
+    :meth:`.FunctionElement.filter`::
+
+        func.count(1).filter(True)
+
+    .. seealso::
+
+        :meth:`.FunctionElement.filter`
+
+    """
+
+    __visit_name__ = "funcfilter"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("func", InternalTraversal.dp_clauseelement),
+        ("criterion", InternalTraversal.dp_clauseelement),
+    ]
+
+    criterion: Optional[ColumnElement[bool]] = None
+
+    def __init__(
+        self,
+        func: Union[FunctionElement[_T], WithinGroup[_T]],
+        *criterion: _ColumnExpressionArgument[bool],
+    ):
+        self.func = func
+        self.filter.non_generative(self, *criterion)  # type: ignore
+
+    @_generative
+    def filter(self, *criterion: _ColumnExpressionArgument[bool]) -> Self:
+        """Produce an additional FILTER against the function.
+
+        This method adds additional criteria to the initial criteria
+        set up by :meth:`.FunctionElement.filter`.
+
+        Multiple criteria are joined together at SQL render time
+        via ``AND``.
+
+
+        """
+
+        for crit in list(criterion):
+            crit = coercions.expect(roles.WhereHavingRole, crit)
+
+            if self.criterion is not None:
+                self.criterion = self.criterion & crit
+            else:
+                self.criterion = crit
+
+        return self
+
+    def over(
+        self,
+        partition_by: Optional[
+            Union[
+                Iterable[_ColumnExpressionArgument[Any]],
+                _ColumnExpressionArgument[Any],
+            ]
+        ] = None,
+        order_by: Optional[
+            Union[
+                Iterable[_ColumnExpressionArgument[Any]],
+                _ColumnExpressionArgument[Any],
+            ]
+        ] = None,
+        range_: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+        rows: Optional[typing_Tuple[Optional[int], Optional[int]]] = None,
+    ) -> Over[_T]:
+        """Produce an OVER clause against this filtered function.
+
+        Used against aggregate or so-called "window" functions,
+        for database backends that support window functions.
+
+        The expression::
+
+            func.rank().filter(MyClass.y > 5).over(order_by="x")
+
+        is shorthand for::
+
+            from sqlalchemy import over, funcfilter
+
+            over(funcfilter(func.rank(), MyClass.y > 5), order_by="x")
+
+        See :func:`_expression.over` for a full description.
+
+        """
+        return Over(
+            self,
+            partition_by=partition_by,
+            order_by=order_by,
+            range_=range_,
+            rows=rows,
+        )
+
+    def within_group(
+        self, *order_by: _ColumnExpressionArgument[Any]
+    ) -> WithinGroup[_T]:
+        """Produce a WITHIN GROUP (ORDER BY expr) clause against
+        this function.
+        """
+        return WithinGroup(self, *order_by)
+
+    def within_group_type(
+        self, within_group: WithinGroup[_T]
+    ) -> Optional[TypeEngine[_T]]:
+        return None
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[Self, Grouping[_T]]:
+        if operators.is_precedent(operators.filter_op, against):
+            return Grouping(self)
+        else:
+            return self
+
+    if not TYPE_CHECKING:
+
+        @util.memoized_property
+        def type(self) -> TypeEngine[_T]:  # noqa: A001
+            return self.func.type
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return list(
+            itertools.chain(
+                *[
+                    c._from_objects
+                    for c in (self.func, self.criterion)
+                    if c is not None
+                ]
+            )
+        )
+
+
+class NamedColumn(KeyedColumnElement[_T]):
+    is_literal = False
+    table: Optional[FromClause] = None
+    name: str
+    key: str
+
+    def _compare_name_for_result(self, other):
+        return (hasattr(other, "name") and self.name == other.name) or (
+            hasattr(other, "_label") and self._label == other._label
+        )
+
+    @util.ro_memoized_property
+    def description(self) -> str:
+        return self.name
+
+    @HasMemoized.memoized_attribute
+    def _tq_key_label(self) -> Optional[str]:
+        """table qualified label based on column key.
+
+        for table-bound columns this is <tablename>_<column key/proxy key>;
+
+        all other expressions it resolves to key/proxy key.
+
+        """
+        proxy_key = self._proxy_key
+        if proxy_key and proxy_key != self.name:
+            return self._gen_tq_label(proxy_key)
+        else:
+            return self._tq_label
+
+    @HasMemoized.memoized_attribute
+    def _tq_label(self) -> Optional[str]:
+        """table qualified label based on column name.
+
+        for table-bound columns this is <tablename>_<columnname>; all other
+        expressions it resolves to .name.
+
+        """
+        return self._gen_tq_label(self.name)
+
+    @HasMemoized.memoized_attribute
+    def _render_label_in_columns_clause(self):
+        return True
+
+    @HasMemoized.memoized_attribute
+    def _non_anon_label(self):
+        return self.name
+
+    def _gen_tq_label(
+        self, name: str, dedupe_on_key: bool = True
+    ) -> Optional[str]:
+        return name
+
+    def _bind_param(
+        self,
+        operator: OperatorType,
+        obj: Any,
+        type_: Optional[TypeEngine[_T]] = None,
+        expanding: bool = False,
+    ) -> BindParameter[_T]:
+        return BindParameter(
+            self.key,
+            obj,
+            _compared_to_operator=operator,
+            _compared_to_type=self.type,
+            type_=type_,
+            unique=True,
+            expanding=expanding,
+        )
+
+    def _make_proxy(
+        self,
+        selectable: FromClause,
+        *,
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        name: Optional[str] = None,
+        key: Optional[str] = None,
+        name_is_truncatable: bool = False,
+        compound_select_cols: Optional[Sequence[ColumnElement[Any]]] = None,
+        disallow_is_literal: bool = False,
+        **kw: Any,
+    ) -> typing_Tuple[str, ColumnClause[_T]]:
+        c = ColumnClause(
+            (
+                coercions.expect(roles.TruncatedLabelRole, name or self.name)
+                if name_is_truncatable
+                else (name or self.name)
+            ),
+            type_=self.type,
+            _selectable=selectable,
+            is_literal=False,
+        )
+
+        c._propagate_attrs = selectable._propagate_attrs
+        if name is None:
+            c.key = self.key
+        if compound_select_cols:
+            c._proxies = list(compound_select_cols)
+        else:
+            c._proxies = [self]
+
+        if selectable._is_clone_of is not None:
+            c._is_clone_of = selectable._is_clone_of.columns.get(c.key)
+        return c.key, c
+
+
+_PS = ParamSpec("_PS")
+
+
+class Label(roles.LabeledColumnExprRole[_T], NamedColumn[_T]):
+    """Represents a column label (AS).
+
+    Represent a label, as typically applied to any column-level
+    element using the ``AS`` sql keyword.
+
+    """
+
+    __visit_name__ = "label"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("name", InternalTraversal.dp_anon_name),
+        ("type", InternalTraversal.dp_type),
+        ("_element", InternalTraversal.dp_clauseelement),
+    ]
+
+    _cache_key_traversal = [
+        ("name", InternalTraversal.dp_anon_name),
+        ("_element", InternalTraversal.dp_clauseelement),
+    ]
+
+    _element: ColumnElement[_T]
+    name: str
+
+    def __init__(
+        self,
+        name: Optional[str],
+        element: _ColumnExpressionArgument[_T],
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+    ):
+        orig_element = element
+        element = coercions.expect(
+            roles.ExpressionElementRole,
+            element,
+            apply_propagate_attrs=self,
+        )
+        while isinstance(element, Label):
+            # TODO: this is only covered in test_text.py, but nothing
+            # fails if it's removed.  determine rationale
+            element = element.element
+
+        if name:
+            self.name = name
+        else:
+            self.name = _anonymous_label.safe_construct(
+                id(self), getattr(element, "name", "anon")
+            )
+            if isinstance(orig_element, Label):
+                # TODO: no coverage for this block, again would be in
+                # test_text.py where the resolve_label concept is important
+                self._resolve_label = orig_element._label
+
+        self.key = self._tq_label = self._tq_key_label = self.name
+        self._element = element
+
+        self.type = (
+            type_api.to_instance(type_)
+            if type_ is not None
+            else self._element.type
+        )
+
+        self._proxies = [element]
+
+    def __reduce__(self):
+        return self.__class__, (self.name, self._element, self.type)
+
+    @HasMemoized.memoized_attribute
+    def _render_label_in_columns_clause(self):
+        return True
+
+    def _bind_param(self, operator, obj, type_=None, expanding=False):
+        return BindParameter(
+            None,
+            obj,
+            _compared_to_operator=operator,
+            type_=type_,
+            _compared_to_type=self.type,
+            unique=True,
+            expanding=expanding,
+        )
+
+    @util.memoized_property
+    def _is_implicitly_boolean(self):
+        return self.element._is_implicitly_boolean
+
+    @HasMemoized.memoized_attribute
+    def _allow_label_resolve(self):
+        return self.element._allow_label_resolve
+
+    @property
+    def _order_by_label_element(self):
+        return self
+
+    @HasMemoized.memoized_attribute
+    def element(self) -> ColumnElement[_T]:
+        return self._element.self_group(against=operators.as_)
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Label[_T]:
+        return self._apply_to_inner(self._element.self_group, against=against)
+
+    def _negate(self):
+        return self._apply_to_inner(self._element._negate)
+
+    def _apply_to_inner(
+        self,
+        fn: Callable[_PS, ColumnElement[_T]],
+        *arg: _PS.args,
+        **kw: _PS.kwargs,
+    ) -> Label[_T]:
+        sub_element = fn(*arg, **kw)
+        if sub_element is not self._element:
+            return Label(self.name, sub_element, type_=self.type)
+        else:
+            return self
+
+    @property
+    def primary_key(self):
+        return self.element.primary_key
+
+    @property
+    def foreign_keys(self):
+        return self.element.foreign_keys
+
+    def _copy_internals(
+        self,
+        *,
+        clone: _CloneCallableType = _clone,
+        anonymize_labels: bool = False,
+        **kw: Any,
+    ) -> None:
+        self._reset_memoizations()
+        self._element = clone(self._element, **kw)
+        if anonymize_labels:
+            self.name = _anonymous_label.safe_construct(
+                id(self), getattr(self.element, "name", "anon")
+            )
+            self.key = self._tq_label = self._tq_key_label = self.name
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.element._from_objects
+
+    def _make_proxy(
+        self,
+        selectable: FromClause,
+        *,
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        name: Optional[str] = None,
+        compound_select_cols: Optional[Sequence[ColumnElement[Any]]] = None,
+        **kw: Any,
+    ) -> typing_Tuple[str, ColumnClause[_T]]:
+        name = self.name if not name else name
+
+        key, e = self.element._make_proxy(
+            selectable,
+            name=name,
+            disallow_is_literal=True,
+            name_is_truncatable=isinstance(name, _truncated_label),
+            compound_select_cols=compound_select_cols,
+            primary_key=primary_key,
+            foreign_keys=foreign_keys,
+        )
+
+        # there was a note here to remove this assertion, which was here
+        # to determine if we later could support a use case where
+        # the key and name of a label are separate.  But I don't know what
+        # that case was.  For now, this is an unexpected case that occurs
+        # when a label name conflicts with other columns and select()
+        # is attempting to disambiguate an explicit label, which is not what
+        # the user would want.   See issue #6090.
+        if key != self.name and not isinstance(self.name, _anonymous_label):
+            raise exc.InvalidRequestError(
+                "Label name %s is being renamed to an anonymous label due "
+                "to disambiguation "
+                "which is not supported right now.  Please use unique names "
+                "for explicit labels." % (self.name)
+            )
+
+        e._propagate_attrs = selectable._propagate_attrs
+        e._proxies.append(self)
+        if self.type is not None:
+            e.type = self.type
+
+        return self.key, e
+
+
+class ColumnClause(
+    roles.DDLReferredColumnRole,
+    roles.LabeledColumnExprRole[_T],
+    roles.StrAsPlainColumnRole,
+    Immutable,
+    NamedColumn[_T],
+):
+    """Represents a column expression from any textual string.
+
+    The :class:`.ColumnClause`, a lightweight analogue to the
+    :class:`_schema.Column` class, is typically invoked using the
+    :func:`_expression.column` function, as in::
+
+        from sqlalchemy import column
+
+        id, name = column("id"), column("name")
+        stmt = select(id, name).select_from("user")
+
+    The above statement would produce SQL like:
+
+    .. sourcecode:: sql
+
+        SELECT id, name FROM user
+
+    :class:`.ColumnClause` is the immediate superclass of the schema-specific
+    :class:`_schema.Column` object.  While the :class:`_schema.Column`
+    class has all the
+    same capabilities as :class:`.ColumnClause`, the :class:`.ColumnClause`
+    class is usable by itself in those cases where behavioral requirements
+    are limited to simple SQL expression generation.  The object has none of
+    the associations with schema-level metadata or with execution-time
+    behavior that :class:`_schema.Column` does,
+    so in that sense is a "lightweight"
+    version of :class:`_schema.Column`.
+
+    Full details on :class:`.ColumnClause` usage is at
+    :func:`_expression.column`.
+
+    .. seealso::
+
+        :func:`_expression.column`
+
+        :class:`_schema.Column`
+
+    """
+
+    table: Optional[FromClause]
+    is_literal: bool
+
+    __visit_name__ = "column"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("name", InternalTraversal.dp_anon_name),
+        ("type", InternalTraversal.dp_type),
+        ("table", InternalTraversal.dp_clauseelement),
+        ("is_literal", InternalTraversal.dp_boolean),
+    ]
+
+    onupdate: Optional[DefaultGenerator] = None
+    default: Optional[DefaultGenerator] = None
+    server_default: Optional[FetchedValue] = None
+    server_onupdate: Optional[FetchedValue] = None
+
+    _is_multiparam_column = False
+
+    @property
+    def _is_star(self):
+        return self.is_literal and self.name == "*"
+
+    def __init__(
+        self,
+        text: str,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        is_literal: bool = False,
+        _selectable: Optional[FromClause] = None,
+    ):
+        self.key = self.name = text
+        self.table = _selectable
+
+        # if type is None, we get NULLTYPE, which is our _T.  But I don't
+        # know how to get the overloads to express that correctly
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+        self.is_literal = is_literal
+
+    def get_children(self, *, column_tables=False, **kw):
+        # override base get_children() to not return the Table
+        # or selectable that is parent to this column.  Traversals
+        # expect the columns of tables and subqueries to be leaf nodes.
+        return []
+
+    @property
+    def entity_namespace(self):
+        if self.table is not None:
+            return self.table.entity_namespace
+        else:
+            return super().entity_namespace
+
+    def _clone(self, detect_subquery_cols=False, **kw):
+        if (
+            detect_subquery_cols
+            and self.table is not None
+            and self.table._is_subquery
+        ):
+            clone = kw.pop("clone")
+            table = clone(self.table, **kw)
+            new = table.c.corresponding_column(self)
+            return new
+
+        return super()._clone(**kw)
+
+    @HasMemoized_ro_memoized_attribute
+    def _from_objects(self) -> List[FromClause]:
+        t = self.table
+        if t is not None:
+            return [t]
+        else:
+            return []
+
+    @HasMemoized.memoized_attribute
+    def _render_label_in_columns_clause(self):
+        return self.table is not None
+
+    @property
+    def _ddl_label(self):
+        return self._gen_tq_label(self.name, dedupe_on_key=False)
+
+    def _compare_name_for_result(self, other):
+        if (
+            self.is_literal
+            or self.table is None
+            or self.table._is_textual
+            or not hasattr(other, "proxy_set")
+            or (
+                isinstance(other, ColumnClause)
+                and (
+                    other.is_literal
+                    or other.table is None
+                    or other.table._is_textual
+                )
+            )
+        ):
+            return (hasattr(other, "name") and self.name == other.name) or (
+                hasattr(other, "_tq_label")
+                and self._tq_label == other._tq_label
+            )
+        else:
+            return other.proxy_set.intersection(self.proxy_set)
+
+    def _gen_tq_label(
+        self, name: str, dedupe_on_key: bool = True
+    ) -> Optional[str]:
+        """generate table-qualified label
+
+        for a table-bound column this is <tablename>_<columnname>.
+
+        used primarily for LABEL_STYLE_TABLENAME_PLUS_COL
+        as well as the .columns collection on a Join object.
+
+        """
+        label: str
+        t = self.table
+        if self.is_literal:
+            return None
+        elif t is not None and is_named_from_clause(t):
+            if has_schema_attr(t) and t.schema:
+                label = t.schema.replace(".", "_") + "_" + t.name + "_" + name
+            else:
+                assert not TYPE_CHECKING or isinstance(t, NamedFromClause)
+                label = t.name + "_" + name
+
+            # propagate name quoting rules for labels.
+            if is_quoted_name(name) and name.quote is not None:
+                if is_quoted_name(label):
+                    label.quote = name.quote
+                else:
+                    label = quoted_name(label, name.quote)
+            elif is_quoted_name(t.name) and t.name.quote is not None:
+                # can't get this situation to occur, so let's
+                # assert false on it for now
+                assert not isinstance(label, quoted_name)
+                label = quoted_name(label, t.name.quote)
+
+            if dedupe_on_key:
+                # ensure the label name doesn't conflict with that of an
+                # existing column.   note that this implies that any Column
+                # must **not** set up its _label before its parent table has
+                # all of its other Column objects set up.  There are several
+                # tables in the test suite which will fail otherwise; example:
+                # table "owner" has columns "name" and "owner_name".  Therefore
+                # column owner.name cannot use the label "owner_name", it has
+                # to be "owner_name_1".
+                if label in t.c:
+                    _label = label
+                    counter = 1
+                    while _label in t.c:
+                        _label = label + "_" + str(counter)
+                        counter += 1
+                    label = _label
+
+            return coercions.expect(roles.TruncatedLabelRole, label)
+
+        else:
+            return name
+
+    def _make_proxy(
+        self,
+        selectable: FromClause,
+        *,
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        name: Optional[str] = None,
+        key: Optional[str] = None,
+        name_is_truncatable: bool = False,
+        compound_select_cols: Optional[Sequence[ColumnElement[Any]]] = None,
+        disallow_is_literal: bool = False,
+        **kw: Any,
+    ) -> typing_Tuple[str, ColumnClause[_T]]:
+        # the "is_literal" flag normally should never be propagated; a proxied
+        # column is always a SQL identifier and never the actual expression
+        # being evaluated. however, there is a case where the "is_literal" flag
+        # might be used to allow the given identifier to have a fixed quoting
+        # pattern already, so maintain the flag for the proxy unless a
+        # :class:`.Label` object is creating the proxy.  See [ticket:4730].
+        is_literal = (
+            not disallow_is_literal
+            and self.is_literal
+            and (
+                # note this does not accommodate for quoted_name differences
+                # right now
+                name is None
+                or name == self.name
+            )
+        )
+        c = self._constructor(
+            (
+                coercions.expect(roles.TruncatedLabelRole, name or self.name)
+                if name_is_truncatable
+                else (name or self.name)
+            ),
+            type_=self.type,
+            _selectable=selectable,
+            is_literal=is_literal,
+        )
+        c._propagate_attrs = selectable._propagate_attrs
+        if name is None:
+            c.key = self.key
+        if compound_select_cols:
+            c._proxies = list(compound_select_cols)
+        else:
+            c._proxies = [self]
+
+        if selectable._is_clone_of is not None:
+            c._is_clone_of = selectable._is_clone_of.columns.get(c.key)
+        return c.key, c
+
+
+class TableValuedColumn(NamedColumn[_T]):
+    __visit_name__ = "table_valued_column"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("name", InternalTraversal.dp_anon_name),
+        ("type", InternalTraversal.dp_type),
+        ("scalar_alias", InternalTraversal.dp_clauseelement),
+    ]
+
+    def __init__(self, scalar_alias: NamedFromClause, type_: TypeEngine[_T]):
+        self.scalar_alias = scalar_alias
+        self.key = self.name = scalar_alias.name
+        self.type = type_
+
+    def _copy_internals(
+        self, clone: _CloneCallableType = _clone, **kw: Any
+    ) -> None:
+        self.scalar_alias = clone(self.scalar_alias, **kw)
+        self.key = self.name = self.scalar_alias.name
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return [self.scalar_alias]
+
+
+class CollationClause(ColumnElement[str]):
+    __visit_name__ = "collation"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("collation", InternalTraversal.dp_string)
+    ]
+
+    @classmethod
+    @util.preload_module("sqlalchemy.sql.sqltypes")
+    def _create_collation_expression(
+        cls, expression: _ColumnExpressionArgument[str], collation: str
+    ) -> BinaryExpression[str]:
+
+        sqltypes = util.preloaded.sql_sqltypes
+
+        expr = coercions.expect(roles.ExpressionElementRole[str], expression)
+
+        if expr.type._type_affinity is sqltypes.String:
+            collate_type = expr.type._with_collation(collation)
+        else:
+            collate_type = expr.type
+
+        return BinaryExpression(
+            expr,
+            CollationClause(collation),
+            operators.collate,
+            type_=collate_type,
+        )
+
+    def __init__(self, collation):
+        self.collation = collation
+
+
+class _IdentifiedClause(Executable, ClauseElement):
+    __visit_name__ = "identified"
+
+    def __init__(self, ident):
+        self.ident = ident
+
+
+class SavepointClause(_IdentifiedClause):
+    __visit_name__ = "savepoint"
+    inherit_cache = False
+
+
+class RollbackToSavepointClause(_IdentifiedClause):
+    __visit_name__ = "rollback_to_savepoint"
+    inherit_cache = False
+
+
+class ReleaseSavepointClause(_IdentifiedClause):
+    __visit_name__ = "release_savepoint"
+    inherit_cache = False
+
+
+class quoted_name(util.MemoizedSlots, str):
+    """Represent a SQL identifier combined with quoting preferences.
+
+    :class:`.quoted_name` is a Python unicode/str subclass which
+    represents a particular identifier name along with a
+    ``quote`` flag.  This ``quote`` flag, when set to
+    ``True`` or ``False``, overrides automatic quoting behavior
+    for this identifier in order to either unconditionally quote
+    or to not quote the name.  If left at its default of ``None``,
+    quoting behavior is applied to the identifier on a per-backend basis
+    based on an examination of the token itself.
+
+    A :class:`.quoted_name` object with ``quote=True`` is also
+    prevented from being modified in the case of a so-called
+    "name normalize" option.  Certain database backends, such as
+    Oracle Database, Firebird, and DB2 "normalize" case-insensitive names
+    as uppercase.  The SQLAlchemy dialects for these backends
+    convert from SQLAlchemy's lower-case-means-insensitive convention
+    to the upper-case-means-insensitive conventions of those backends.
+    The ``quote=True`` flag here will prevent this conversion from occurring
+    to support an identifier that's quoted as all lower case against
+    such a backend.
+
+    The :class:`.quoted_name` object is normally created automatically
+    when specifying the name for key schema constructs such as
+    :class:`_schema.Table`, :class:`_schema.Column`, and others.
+    The class can also be
+    passed explicitly as the name to any function that receives a name which
+    can be quoted.  Such as to use the :meth:`_engine.Engine.has_table`
+    method with
+    an unconditionally quoted name::
+
+        from sqlalchemy import create_engine
+        from sqlalchemy import inspect
+        from sqlalchemy.sql import quoted_name
+
+        engine = create_engine("oracle+oracledb://some_dsn")
+        print(inspect(engine).has_table(quoted_name("some_table", True)))
+
+    The above logic will run the "has table" logic against the Oracle Database
+    backend, passing the name exactly as ``"some_table"`` without converting to
+    upper case.
+
+    .. versionchanged:: 1.2 The :class:`.quoted_name` construct is now
+       importable from ``sqlalchemy.sql``, in addition to the previous
+       location of ``sqlalchemy.sql.elements``.
+
+    """
+
+    __slots__ = "quote", "lower", "upper"
+
+    quote: Optional[bool]
+
+    @overload
+    @classmethod
+    def construct(cls, value: str, quote: Optional[bool]) -> quoted_name: ...
+
+    @overload
+    @classmethod
+    def construct(cls, value: None, quote: Optional[bool]) -> None: ...
+
+    @classmethod
+    def construct(
+        cls, value: Optional[str], quote: Optional[bool]
+    ) -> Optional[quoted_name]:
+        if value is None:
+            return None
+        else:
+            return quoted_name(value, quote)
+
+    def __new__(cls, value: str, quote: Optional[bool]) -> quoted_name:
+        assert (
+            value is not None
+        ), "use quoted_name.construct() for None passthrough"
+        if isinstance(value, cls) and (quote is None or value.quote == quote):
+            return value
+        self = super().__new__(cls, value)
+
+        self.quote = quote
+        return self
+
+    def __reduce__(self):
+        return quoted_name, (str(self), self.quote)
+
+    def _memoized_method_lower(self):
+        if self.quote:
+            return self
+        else:
+            return str(self).lower()
+
+    def _memoized_method_upper(self):
+        if self.quote:
+            return self
+        else:
+            return str(self).upper()
+
+
+def _find_columns(clause: ClauseElement) -> Set[ColumnClause[Any]]:
+    """locate Column objects within the given expression."""
+
+    cols: Set[ColumnClause[Any]] = set()
+    traverse(clause, {}, {"column": cols.add})
+    return cols
+
+
+def _type_from_args(args: Sequence[ColumnElement[_T]]) -> TypeEngine[_T]:
+    for a in args:
+        if not a.type._isnull:
+            return a.type
+    else:
+        return type_api.NULLTYPE  # type: ignore
+
+
+def _corresponding_column_or_error(fromclause, column, require_embedded=False):
+    c = fromclause.corresponding_column(
+        column, require_embedded=require_embedded
+    )
+    if c is None:
+        raise exc.InvalidRequestError(
+            "Given column '%s', attached to table '%s', "
+            "failed to locate a corresponding column from table '%s'"
+            % (column, getattr(column, "table", None), fromclause.description)
+        )
+    return c
+
+
+class _memoized_property_but_not_nulltype(
+    util.memoized_property["TypeEngine[_T]"]
+):
+    """memoized property, but dont memoize NullType"""
+
+    def __get__(self, obj, cls):
+        if obj is None:
+            return self
+        result = self.fget(obj)
+        if not result._isnull:
+            obj.__dict__[self.__name__] = result
+        return result
+
+
+class AnnotatedColumnElement(Annotated):
+    _Annotated__element: ColumnElement[Any]
+
+    def __init__(self, element, values):
+        Annotated.__init__(self, element, values)
+        for attr in (
+            "comparator",
+            "_proxy_key",
+            "_tq_key_label",
+            "_tq_label",
+            "_non_anon_label",
+            "type",
+        ):
+            self.__dict__.pop(attr, None)
+        for attr in ("name", "key", "table"):
+            if self.__dict__.get(attr, False) is None:
+                self.__dict__.pop(attr)
+
+    def _with_annotations(self, values):
+        clone = super()._with_annotations(values)
+        for attr in (
+            "comparator",
+            "_proxy_key",
+            "_tq_key_label",
+            "_tq_label",
+            "_non_anon_label",
+        ):
+            clone.__dict__.pop(attr, None)
+        return clone
+
+    @util.memoized_property
+    def name(self):
+        """pull 'name' from parent, if not present"""
+        return self._Annotated__element.name
+
+    @_memoized_property_but_not_nulltype
+    def type(self):
+        """pull 'type' from parent and don't cache if null.
+
+        type is routinely changed on existing columns within the
+        mapped_column() initialization process, and "type" is also consulted
+        during the creation of SQL expressions.  Therefore it can change after
+        it was already retrieved.  At the same time we don't want annotated
+        objects having overhead when expressions are produced, so continue
+        to memoize, but only when we have a non-null type.
+
+        """
+        return self._Annotated__element.type
+
+    @util.memoized_property
+    def table(self):
+        """pull 'table' from parent, if not present"""
+        return self._Annotated__element.table
+
+    @util.memoized_property
+    def key(self):
+        """pull 'key' from parent, if not present"""
+        return self._Annotated__element.key
+
+    @util.memoized_property
+    def info(self) -> _InfoType:
+        if TYPE_CHECKING:
+            assert isinstance(self._Annotated__element, Column)
+        return self._Annotated__element.info
+
+    @util.memoized_property
+    def _anon_name_label(self) -> str:
+        return self._Annotated__element._anon_name_label
+
+
+class _truncated_label(quoted_name):
+    """A unicode subclass used to identify symbolic "
+    "names that may require truncation."""
+
+    __slots__ = ()
+
+    def __new__(cls, value: str, quote: Optional[bool] = None) -> Any:
+        quote = getattr(value, "quote", quote)
+        # return super(_truncated_label, cls).__new__(cls, value, quote, True)
+        return super().__new__(cls, value, quote)
+
+    def __reduce__(self) -> Any:
+        return self.__class__, (str(self), self.quote)
+
+    def apply_map(self, map_: Mapping[str, Any]) -> str:
+        return self
+
+
+class conv(_truncated_label):
+    """Mark a string indicating that a name has already been converted
+    by a naming convention.
+
+    This is a string subclass that indicates a name that should not be
+    subject to any further naming conventions.
+
+    E.g. when we create a :class:`.Constraint` using a naming convention
+    as follows::
+
+        m = MetaData(
+            naming_convention={"ck": "ck_%(table_name)s_%(constraint_name)s"}
+        )
+        t = Table(
+            "t", m, Column("x", Integer), CheckConstraint("x > 5", name="x5")
+        )
+
+    The name of the above constraint will be rendered as ``"ck_t_x5"``.
+    That is, the existing name ``x5`` is used in the naming convention as the
+    ``constraint_name`` token.
+
+    In some situations, such as in migration scripts, we may be rendering
+    the above :class:`.CheckConstraint` with a name that's already been
+    converted.  In order to make sure the name isn't double-modified, the
+    new name is applied using the :func:`_schema.conv` marker.  We can
+    use this explicitly as follows::
+
+
+        m = MetaData(
+            naming_convention={"ck": "ck_%(table_name)s_%(constraint_name)s"}
+        )
+        t = Table(
+            "t",
+            m,
+            Column("x", Integer),
+            CheckConstraint("x > 5", name=conv("ck_t_x5")),
+        )
+
+    Where above, the :func:`_schema.conv` marker indicates that the constraint
+    name here is final, and the name will render as ``"ck_t_x5"`` and not
+    ``"ck_t_ck_t_x5"``
+
+    .. seealso::
+
+        :ref:`constraint_naming_conventions`
+
+    """
+
+    __slots__ = ()
+
+
+# for backwards compatibility in case
+# someone is re-implementing the
+# _truncated_identifier() sequence in a custom
+# compiler
+_generated_label = _truncated_label
+
+
+class _anonymous_label(_truncated_label):
+    """A unicode subclass used to identify anonymously
+    generated names."""
+
+    __slots__ = ()
+
+    @classmethod
+    def safe_construct(
+        cls,
+        seed: int,
+        body: str,
+        enclosing_label: Optional[str] = None,
+        sanitize_key: bool = False,
+    ) -> _anonymous_label:
+        # need to escape chars that interfere with format
+        # strings in any case, issue #8724
+        body = re.sub(r"[%\(\) \$]+", "_", body)
+
+        if sanitize_key:
+            # sanitize_key is then an extra step used by BindParameter
+            body = body.strip("_")
+
+        label = "%%(%d %s)s" % (seed, body.replace("%", "%%"))
+        if enclosing_label:
+            label = "%s%s" % (enclosing_label, label)
+
+        return _anonymous_label(label)
+
+    def __add__(self, other):
+        if "%" in other and not isinstance(other, _anonymous_label):
+            other = str(other).replace("%", "%%")
+        else:
+            other = str(other)
+
+        return _anonymous_label(
+            quoted_name(
+                str.__add__(self, other),
+                self.quote,
+            )
+        )
+
+    def __radd__(self, other):
+        if "%" in other and not isinstance(other, _anonymous_label):
+            other = str(other).replace("%", "%%")
+        else:
+            other = str(other)
+
+        return _anonymous_label(
+            quoted_name(
+                str.__add__(other, self),
+                self.quote,
+            )
+        )
+
+    def apply_map(self, map_):
+        if self.quote is not None:
+            # preserve quoting only if necessary
+            return quoted_name(self % map_, self.quote)
+        else:
+            # else skip the constructor call
+            return self % map_
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/events.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/events.py
new file mode 100644
index 00000000..601092fd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/events.py
@@ -0,0 +1,458 @@
+# sql/events.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+from __future__ import annotations
+
+from typing import Any
+from typing import TYPE_CHECKING
+
+from .base import SchemaEventTarget
+from .. import event
+
+if TYPE_CHECKING:
+    from .schema import Column
+    from .schema import Constraint
+    from .schema import SchemaItem
+    from .schema import Table
+    from ..engine.base import Connection
+    from ..engine.interfaces import ReflectedColumn
+    from ..engine.reflection import Inspector
+
+
+class DDLEvents(event.Events[SchemaEventTarget]):
+    """
+    Define event listeners for schema objects,
+    that is, :class:`.SchemaItem` and other :class:`.SchemaEventTarget`
+    subclasses, including :class:`_schema.MetaData`, :class:`_schema.Table`,
+    :class:`_schema.Column`, etc.
+
+    **Create / Drop Events**
+
+    Events emitted when CREATE and DROP commands are emitted to the database.
+    The event hooks in this category include :meth:`.DDLEvents.before_create`,
+    :meth:`.DDLEvents.after_create`, :meth:`.DDLEvents.before_drop`, and
+    :meth:`.DDLEvents.after_drop`.
+
+    These events are emitted when using schema-level methods such as
+    :meth:`.MetaData.create_all` and :meth:`.MetaData.drop_all`. Per-object
+    create/drop methods such as :meth:`.Table.create`, :meth:`.Table.drop`,
+    :meth:`.Index.create` are also included, as well as dialect-specific
+    methods such as :meth:`_postgresql.ENUM.create`.
+
+    .. versionadded:: 2.0 :class:`.DDLEvents` event hooks now take place
+       for non-table objects including constraints, indexes, and
+       dialect-specific schema types.
+
+    Event hooks may be attached directly to a :class:`_schema.Table` object or
+    to a :class:`_schema.MetaData` collection, as well as to any
+    :class:`.SchemaItem` class or object that can be individually created and
+    dropped using a distinct SQL command. Such classes include :class:`.Index`,
+    :class:`.Sequence`, and dialect-specific classes such as
+    :class:`_postgresql.ENUM`.
+
+    Example using the :meth:`.DDLEvents.after_create` event, where a custom
+    event hook will emit an ``ALTER TABLE`` command on the current connection,
+    after ``CREATE TABLE`` is emitted::
+
+        from sqlalchemy import create_engine
+        from sqlalchemy import event
+        from sqlalchemy import Table, Column, Metadata, Integer
+
+        m = MetaData()
+        some_table = Table("some_table", m, Column("data", Integer))
+
+
+        @event.listens_for(some_table, "after_create")
+        def after_create(target, connection, **kw):
+            connection.execute(
+                text("ALTER TABLE %s SET name=foo_%s" % (target.name, target.name))
+            )
+
+
+        some_engine = create_engine("postgresql://scott:tiger@host/test")
+
+        # will emit "CREATE TABLE some_table" as well as the above
+        # "ALTER TABLE" statement afterwards
+        m.create_all(some_engine)
+
+    Constraint objects such as :class:`.ForeignKeyConstraint`,
+    :class:`.UniqueConstraint`, :class:`.CheckConstraint` may also be
+    subscribed to these events, however they will **not** normally produce
+    events as these objects are usually rendered inline within an
+    enclosing ``CREATE TABLE`` statement and implicitly dropped from a
+    ``DROP TABLE`` statement.
+
+    For the :class:`.Index` construct, the event hook will be emitted
+    for ``CREATE INDEX``, however SQLAlchemy does not normally emit
+    ``DROP INDEX`` when dropping tables as this is again implicit within the
+    ``DROP TABLE`` statement.
+
+    .. versionadded:: 2.0 Support for :class:`.SchemaItem` objects
+       for create/drop events was expanded from its previous support for
+       :class:`.MetaData` and :class:`.Table` to also include
+       :class:`.Constraint` and all subclasses, :class:`.Index`,
+       :class:`.Sequence` and some type-related constructs such as
+       :class:`_postgresql.ENUM`.
+
+    .. note:: These event hooks are only emitted within the scope of
+       SQLAlchemy's create/drop methods; they are not necessarily supported
+       by tools such as `alembic <https://alembic.sqlalchemy.org>`_.
+
+
+    **Attachment Events**
+
+    Attachment events are provided to customize
+    behavior whenever a child schema element is associated
+    with a parent, such as when a :class:`_schema.Column` is associated
+    with its :class:`_schema.Table`, when a
+    :class:`_schema.ForeignKeyConstraint`
+    is associated with a :class:`_schema.Table`, etc.  These events include
+    :meth:`.DDLEvents.before_parent_attach` and
+    :meth:`.DDLEvents.after_parent_attach`.
+
+    **Reflection Events**
+
+    The :meth:`.DDLEvents.column_reflect` event is used to intercept
+    and modify the in-Python definition of database columns when
+    :term:`reflection` of database tables proceeds.
+
+    **Use with Generic DDL**
+
+    DDL events integrate closely with the
+    :class:`.DDL` class and the :class:`.ExecutableDDLElement` hierarchy
+    of DDL clause constructs, which are themselves appropriate
+    as listener callables::
+
+        from sqlalchemy import DDL
+
+        event.listen(
+            some_table,
+            "after_create",
+            DDL("ALTER TABLE %(table)s SET name=foo_%(table)s"),
+        )
+
+    **Event Propagation to MetaData Copies**
+
+    For all :class:`.DDLEvent` events, the ``propagate=True`` keyword argument
+    will ensure that a given event handler is propagated to copies of the
+    object, which are made when using the :meth:`_schema.Table.to_metadata`
+    method::
+
+        from sqlalchemy import DDL
+
+        metadata = MetaData()
+        some_table = Table("some_table", metadata, Column("data", Integer))
+
+        event.listen(
+            some_table,
+            "after_create",
+            DDL("ALTER TABLE %(table)s SET name=foo_%(table)s"),
+            propagate=True,
+        )
+
+        new_metadata = MetaData()
+        new_table = some_table.to_metadata(new_metadata)
+
+    The above :class:`.DDL` object will be associated with the
+    :meth:`.DDLEvents.after_create` event for both the ``some_table`` and
+    the ``new_table`` :class:`.Table` objects.
+
+    .. seealso::
+
+        :ref:`event_toplevel`
+
+        :class:`.ExecutableDDLElement`
+
+        :class:`.DDL`
+
+        :ref:`schema_ddl_sequences`
+
+    """  # noqa: E501
+
+    _target_class_doc = "SomeSchemaClassOrObject"
+    _dispatch_target = SchemaEventTarget
+
+    def before_create(
+        self, target: SchemaEventTarget, connection: Connection, **kw: Any
+    ) -> None:
+        r"""Called before CREATE statements are emitted.
+
+        :param target: the :class:`.SchemaObject`, such as a
+         :class:`_schema.MetaData` or :class:`_schema.Table`
+         but also including all create/drop objects such as
+         :class:`.Index`, :class:`.Sequence`, etc.,
+         object which is the target of the event.
+
+         .. versionadded:: 2.0 Support for all :class:`.SchemaItem` objects
+            was added.
+
+        :param connection: the :class:`_engine.Connection` where the
+         CREATE statement or statements will be emitted.
+        :param \**kw: additional keyword arguments relevant
+         to the event.  The contents of this dictionary
+         may vary across releases, and include the
+         list of tables being generated for a metadata-level
+         event, the checkfirst flag, and other
+         elements used by internal events.
+
+        :func:`.event.listen` accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        :func:`.event.listen` accepts the ``insert=True``
+        modifier for this event; when True, the listener function will
+        be prepended to the internal list of events upon discovery, and execute
+        before registered listener functions that do not pass this argument.
+
+        """
+
+    def after_create(
+        self, target: SchemaEventTarget, connection: Connection, **kw: Any
+    ) -> None:
+        r"""Called after CREATE statements are emitted.
+
+        :param target: the :class:`.SchemaObject`, such as a
+         :class:`_schema.MetaData` or :class:`_schema.Table`
+         but also including all create/drop objects such as
+         :class:`.Index`, :class:`.Sequence`, etc.,
+         object which is the target of the event.
+
+         .. versionadded:: 2.0 Support for all :class:`.SchemaItem` objects
+            was added.
+
+        :param connection: the :class:`_engine.Connection` where the
+         CREATE statement or statements have been emitted.
+        :param \**kw: additional keyword arguments relevant
+         to the event.  The contents of this dictionary
+         may vary across releases, and include the
+         list of tables being generated for a metadata-level
+         event, the checkfirst flag, and other
+         elements used by internal events.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        """
+
+    def before_drop(
+        self, target: SchemaEventTarget, connection: Connection, **kw: Any
+    ) -> None:
+        r"""Called before DROP statements are emitted.
+
+        :param target: the :class:`.SchemaObject`, such as a
+         :class:`_schema.MetaData` or :class:`_schema.Table`
+         but also including all create/drop objects such as
+         :class:`.Index`, :class:`.Sequence`, etc.,
+         object which is the target of the event.
+
+         .. versionadded:: 2.0 Support for all :class:`.SchemaItem` objects
+            was added.
+
+        :param connection: the :class:`_engine.Connection` where the
+         DROP statement or statements will be emitted.
+        :param \**kw: additional keyword arguments relevant
+         to the event.  The contents of this dictionary
+         may vary across releases, and include the
+         list of tables being generated for a metadata-level
+         event, the checkfirst flag, and other
+         elements used by internal events.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        """
+
+    def after_drop(
+        self, target: SchemaEventTarget, connection: Connection, **kw: Any
+    ) -> None:
+        r"""Called after DROP statements are emitted.
+
+        :param target: the :class:`.SchemaObject`, such as a
+         :class:`_schema.MetaData` or :class:`_schema.Table`
+         but also including all create/drop objects such as
+         :class:`.Index`, :class:`.Sequence`, etc.,
+         object which is the target of the event.
+
+         .. versionadded:: 2.0 Support for all :class:`.SchemaItem` objects
+            was added.
+
+        :param connection: the :class:`_engine.Connection` where the
+         DROP statement or statements have been emitted.
+        :param \**kw: additional keyword arguments relevant
+         to the event.  The contents of this dictionary
+         may vary across releases, and include the
+         list of tables being generated for a metadata-level
+         event, the checkfirst flag, and other
+         elements used by internal events.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        """
+
+    def before_parent_attach(
+        self, target: SchemaEventTarget, parent: SchemaItem
+    ) -> None:
+        """Called before a :class:`.SchemaItem` is associated with
+        a parent :class:`.SchemaItem`.
+
+        :param target: the target object
+        :param parent: the parent to which the target is being attached.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        """
+
+    def after_parent_attach(
+        self, target: SchemaEventTarget, parent: SchemaItem
+    ) -> None:
+        """Called after a :class:`.SchemaItem` is associated with
+        a parent :class:`.SchemaItem`.
+
+        :param target: the target object
+        :param parent: the parent to which the target is being attached.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        """
+
+    def _sa_event_column_added_to_pk_constraint(
+        self, const: Constraint, col: Column[Any]
+    ) -> None:
+        """internal event hook used for primary key naming convention
+        updates.
+
+        """
+
+    def column_reflect(
+        self, inspector: Inspector, table: Table, column_info: ReflectedColumn
+    ) -> None:
+        """Called for each unit of 'column info' retrieved when
+        a :class:`_schema.Table` is being reflected.
+
+        This event is most easily used by applying it to a specific
+        :class:`_schema.MetaData` instance, where it will take effect for
+        all :class:`_schema.Table` objects within that
+        :class:`_schema.MetaData` that undergo reflection::
+
+            metadata = MetaData()
+
+
+            @event.listens_for(metadata, "column_reflect")
+            def receive_column_reflect(inspector, table, column_info):
+                # receives for all Table objects that are reflected
+                # under this MetaData
+                ...
+
+
+            # will use the above event hook
+            my_table = Table("my_table", metadata, autoload_with=some_engine)
+
+        .. versionadded:: 1.4.0b2 The :meth:`_events.DDLEvents.column_reflect`
+           hook may now be applied to a :class:`_schema.MetaData` object as
+           well as the :class:`_schema.MetaData` class itself where it will
+           take place for all :class:`_schema.Table` objects associated with
+           the targeted :class:`_schema.MetaData`.
+
+        It may also be applied to the :class:`_schema.Table` class across
+        the board::
+
+            from sqlalchemy import Table
+
+
+            @event.listens_for(Table, "column_reflect")
+            def receive_column_reflect(inspector, table, column_info):
+                # receives for all Table objects that are reflected
+                ...
+
+        It can also be applied to a specific :class:`_schema.Table` at the
+        point that one is being reflected using the
+        :paramref:`_schema.Table.listeners` parameter::
+
+            t1 = Table(
+                "my_table",
+                autoload_with=some_engine,
+                listeners=[("column_reflect", receive_column_reflect)],
+            )
+
+        The dictionary of column information as returned by the
+        dialect is passed, and can be modified.  The dictionary
+        is that returned in each element of the list returned
+        by :meth:`.reflection.Inspector.get_columns`:
+
+            * ``name`` - the column's name, is applied to the
+              :paramref:`_schema.Column.name` parameter
+
+            * ``type`` - the type of this column, which should be an instance
+              of :class:`~sqlalchemy.types.TypeEngine`, is applied to the
+              :paramref:`_schema.Column.type` parameter
+
+            * ``nullable`` - boolean flag if the column is NULL or NOT NULL,
+              is applied to the :paramref:`_schema.Column.nullable` parameter
+
+            * ``default`` - the column's server default value.  This is
+              normally specified as a plain string SQL expression, however the
+              event can pass a :class:`.FetchedValue`, :class:`.DefaultClause`,
+              or :func:`_expression.text` object as well.  Is applied to the
+              :paramref:`_schema.Column.server_default` parameter
+
+        The event is called before any action is taken against
+        this dictionary, and the contents can be modified; the following
+        additional keys may be added to the dictionary to further modify
+        how the :class:`_schema.Column` is constructed:
+
+
+            * ``key`` - the string key that will be used to access this
+              :class:`_schema.Column` in the ``.c`` collection; will be applied
+              to the :paramref:`_schema.Column.key` parameter. Is also used
+              for ORM mapping.  See the section
+              :ref:`mapper_automated_reflection_schemes` for an example.
+
+            * ``quote`` - force or un-force quoting on the column name;
+              is applied to the :paramref:`_schema.Column.quote` parameter.
+
+            * ``info`` - a dictionary of arbitrary data to follow along with
+              the :class:`_schema.Column`, is applied to the
+              :paramref:`_schema.Column.info` parameter.
+
+        :func:`.event.listen` also accepts the ``propagate=True``
+        modifier for this event; when True, the listener function will
+        be established for any copies made of the target object,
+        i.e. those copies that are generated when
+        :meth:`_schema.Table.to_metadata` is used.
+
+        .. seealso::
+
+            :ref:`mapper_automated_reflection_schemes` -
+            in the ORM mapping documentation
+
+            :ref:`automap_intercepting_columns` -
+            in the :ref:`automap_toplevel` documentation
+
+            :ref:`metadata_reflection_dbagnostic_types` - in
+            the :ref:`metadata_reflection_toplevel` documentation
+
+        """
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/expression.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/expression.py
new file mode 100644
index 00000000..f8ac3a9e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/expression.py
@@ -0,0 +1,162 @@
+# sql/expression.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Defines the public namespace for SQL expression constructs.
+
+
+"""
+
+
+from __future__ import annotations
+
+from ._dml_constructors import delete as delete
+from ._dml_constructors import insert as insert
+from ._dml_constructors import update as update
+from ._elements_constructors import all_ as all_
+from ._elements_constructors import and_ as and_
+from ._elements_constructors import any_ as any_
+from ._elements_constructors import asc as asc
+from ._elements_constructors import between as between
+from ._elements_constructors import bindparam as bindparam
+from ._elements_constructors import bitwise_not as bitwise_not
+from ._elements_constructors import case as case
+from ._elements_constructors import cast as cast
+from ._elements_constructors import collate as collate
+from ._elements_constructors import column as column
+from ._elements_constructors import desc as desc
+from ._elements_constructors import distinct as distinct
+from ._elements_constructors import extract as extract
+from ._elements_constructors import false as false
+from ._elements_constructors import funcfilter as funcfilter
+from ._elements_constructors import label as label
+from ._elements_constructors import not_ as not_
+from ._elements_constructors import null as null
+from ._elements_constructors import nulls_first as nulls_first
+from ._elements_constructors import nulls_last as nulls_last
+from ._elements_constructors import or_ as or_
+from ._elements_constructors import outparam as outparam
+from ._elements_constructors import over as over
+from ._elements_constructors import text as text
+from ._elements_constructors import true as true
+from ._elements_constructors import try_cast as try_cast
+from ._elements_constructors import tuple_ as tuple_
+from ._elements_constructors import type_coerce as type_coerce
+from ._elements_constructors import within_group as within_group
+from ._selectable_constructors import alias as alias
+from ._selectable_constructors import cte as cte
+from ._selectable_constructors import except_ as except_
+from ._selectable_constructors import except_all as except_all
+from ._selectable_constructors import exists as exists
+from ._selectable_constructors import intersect as intersect
+from ._selectable_constructors import intersect_all as intersect_all
+from ._selectable_constructors import join as join
+from ._selectable_constructors import lateral as lateral
+from ._selectable_constructors import outerjoin as outerjoin
+from ._selectable_constructors import select as select
+from ._selectable_constructors import table as table
+from ._selectable_constructors import tablesample as tablesample
+from ._selectable_constructors import union as union
+from ._selectable_constructors import union_all as union_all
+from ._selectable_constructors import values as values
+from ._typing import ColumnExpressionArgument as ColumnExpressionArgument
+from .base import _from_objects as _from_objects
+from .base import _select_iterables as _select_iterables
+from .base import ColumnCollection as ColumnCollection
+from .base import Executable as Executable
+from .cache_key import CacheKey as CacheKey
+from .dml import Delete as Delete
+from .dml import Insert as Insert
+from .dml import Update as Update
+from .dml import UpdateBase as UpdateBase
+from .dml import ValuesBase as ValuesBase
+from .elements import _truncated_label as _truncated_label
+from .elements import BinaryExpression as BinaryExpression
+from .elements import BindParameter as BindParameter
+from .elements import BooleanClauseList as BooleanClauseList
+from .elements import Case as Case
+from .elements import Cast as Cast
+from .elements import ClauseElement as ClauseElement
+from .elements import ClauseList as ClauseList
+from .elements import CollectionAggregate as CollectionAggregate
+from .elements import ColumnClause as ColumnClause
+from .elements import ColumnElement as ColumnElement
+from .elements import ExpressionClauseList as ExpressionClauseList
+from .elements import Extract as Extract
+from .elements import False_ as False_
+from .elements import FunctionFilter as FunctionFilter
+from .elements import Grouping as Grouping
+from .elements import Label as Label
+from .elements import literal as literal
+from .elements import literal_column as literal_column
+from .elements import Null as Null
+from .elements import Over as Over
+from .elements import quoted_name as quoted_name
+from .elements import ReleaseSavepointClause as ReleaseSavepointClause
+from .elements import RollbackToSavepointClause as RollbackToSavepointClause
+from .elements import SavepointClause as SavepointClause
+from .elements import SQLColumnExpression as SQLColumnExpression
+from .elements import TextClause as TextClause
+from .elements import True_ as True_
+from .elements import TryCast as TryCast
+from .elements import Tuple as Tuple
+from .elements import TypeClause as TypeClause
+from .elements import TypeCoerce as TypeCoerce
+from .elements import UnaryExpression as UnaryExpression
+from .elements import WithinGroup as WithinGroup
+from .functions import func as func
+from .functions import Function as Function
+from .functions import FunctionElement as FunctionElement
+from .functions import modifier as modifier
+from .lambdas import lambda_stmt as lambda_stmt
+from .lambdas import LambdaElement as LambdaElement
+from .lambdas import StatementLambdaElement as StatementLambdaElement
+from .operators import ColumnOperators as ColumnOperators
+from .operators import custom_op as custom_op
+from .operators import Operators as Operators
+from .selectable import Alias as Alias
+from .selectable import AliasedReturnsRows as AliasedReturnsRows
+from .selectable import CompoundSelect as CompoundSelect
+from .selectable import CTE as CTE
+from .selectable import Exists as Exists
+from .selectable import FromClause as FromClause
+from .selectable import FromGrouping as FromGrouping
+from .selectable import GenerativeSelect as GenerativeSelect
+from .selectable import HasCTE as HasCTE
+from .selectable import HasPrefixes as HasPrefixes
+from .selectable import HasSuffixes as HasSuffixes
+from .selectable import Join as Join
+from .selectable import LABEL_STYLE_DEFAULT as LABEL_STYLE_DEFAULT
+from .selectable import (
+    LABEL_STYLE_DISAMBIGUATE_ONLY as LABEL_STYLE_DISAMBIGUATE_ONLY,
+)
+from .selectable import LABEL_STYLE_NONE as LABEL_STYLE_NONE
+from .selectable import (
+    LABEL_STYLE_TABLENAME_PLUS_COL as LABEL_STYLE_TABLENAME_PLUS_COL,
+)
+from .selectable import Lateral as Lateral
+from .selectable import ReturnsRows as ReturnsRows
+from .selectable import ScalarSelect as ScalarSelect
+from .selectable import ScalarValues as ScalarValues
+from .selectable import Select as Select
+from .selectable import Selectable as Selectable
+from .selectable import SelectBase as SelectBase
+from .selectable import SelectLabelStyle as SelectLabelStyle
+from .selectable import Subquery as Subquery
+from .selectable import TableClause as TableClause
+from .selectable import TableSample as TableSample
+from .selectable import TableValuedAlias as TableValuedAlias
+from .selectable import TextAsFrom as TextAsFrom
+from .selectable import TextualSelect as TextualSelect
+from .selectable import Values as Values
+from .visitors import Visitable as Visitable
+
+nullsfirst = nulls_first
+"""Synonym for the :func:`.nulls_first` function."""
+
+
+nullslast = nulls_last
+"""Synonym for the :func:`.nulls_last` function."""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/functions.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/functions.py
new file mode 100644
index 00000000..ea02279d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/functions.py
@@ -0,0 +1,2064 @@
+# sql/functions.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+"""SQL function API, factories, and built-in functions.
+
+"""
+
+from __future__ import annotations
+
+import datetime
+import decimal
+from typing import Any
+from typing import cast
+from typing import Dict
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import annotation
+from . import coercions
+from . import operators
+from . import roles
+from . import schema
+from . import sqltypes
+from . import type_api
+from . import util as sqlutil
+from ._typing import is_table_value_type
+from .base import _entity_namespace
+from .base import ColumnCollection
+from .base import Executable
+from .base import Generative
+from .base import HasMemoized
+from .elements import _type_from_args
+from .elements import BinaryExpression
+from .elements import BindParameter
+from .elements import Cast
+from .elements import ClauseList
+from .elements import ColumnElement
+from .elements import Extract
+from .elements import FunctionFilter
+from .elements import Grouping
+from .elements import literal_column
+from .elements import NamedColumn
+from .elements import Over
+from .elements import WithinGroup
+from .selectable import FromClause
+from .selectable import Select
+from .selectable import TableValuedAlias
+from .sqltypes import TableValueType
+from .type_api import TypeEngine
+from .visitors import InternalTraversal
+from .. import util
+
+
+if TYPE_CHECKING:
+    from ._typing import _ByArgument
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnExpressionOrLiteralArgument
+    from ._typing import _ColumnExpressionOrStrLabelArgument
+    from ._typing import _StarOrOne
+    from ._typing import _TypeEngineArgument
+    from .base import _EntityNamespace
+    from .elements import ClauseElement
+    from .elements import KeyedColumnElement
+    from .elements import TableValuedColumn
+    from .operators import OperatorType
+    from ..engine.base import Connection
+    from ..engine.cursor import CursorResult
+    from ..engine.interfaces import _CoreMultiExecuteParams
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..util.typing import Self
+
+_T = TypeVar("_T", bound=Any)
+_S = TypeVar("_S", bound=Any)
+
+_registry: util.defaultdict[str, Dict[str, Type[Function[Any]]]] = (
+    util.defaultdict(dict)
+)
+
+
+def register_function(
+    identifier: str, fn: Type[Function[Any]], package: str = "_default"
+) -> None:
+    """Associate a callable with a particular func. name.
+
+    This is normally called by GenericFunction, but is also
+    available by itself so that a non-Function construct
+    can be associated with the :data:`.func` accessor (i.e.
+    CAST, EXTRACT).
+
+    """
+    reg = _registry[package]
+
+    identifier = str(identifier).lower()
+
+    # Check if a function with the same identifier is registered.
+    if identifier in reg:
+        util.warn(
+            "The GenericFunction '{}' is already registered and "
+            "is going to be overridden.".format(identifier)
+        )
+    reg[identifier] = fn
+
+
+class FunctionElement(Executable, ColumnElement[_T], FromClause, Generative):
+    """Base for SQL function-oriented constructs.
+
+    This is a `generic type <https://peps.python.org/pep-0484/#generics>`_,
+    meaning that type checkers and IDEs can be instructed on the types to
+    expect in a :class:`_engine.Result` for this function. See
+    :class:`.GenericFunction` for an example of how this is done.
+
+    .. seealso::
+
+        :ref:`tutorial_functions` - in the :ref:`unified_tutorial`
+
+        :class:`.Function` - named SQL function.
+
+        :data:`.func` - namespace which produces registered or ad-hoc
+        :class:`.Function` instances.
+
+        :class:`.GenericFunction` - allows creation of registered function
+        types.
+
+    """
+
+    _traverse_internals = [
+        ("clause_expr", InternalTraversal.dp_clauseelement),
+        ("_with_ordinality", InternalTraversal.dp_boolean),
+        ("_table_value_type", InternalTraversal.dp_has_cache_key),
+    ]
+
+    packagenames: Tuple[str, ...] = ()
+
+    _has_args = False
+    _with_ordinality = False
+    _table_value_type: Optional[TableValueType] = None
+
+    # some attributes that are defined between both ColumnElement and
+    # FromClause are set to Any here to avoid typing errors
+    primary_key: Any
+    _is_clone_of: Any
+
+    clause_expr: Grouping[Any]
+
+    def __init__(self, *clauses: _ColumnExpressionOrLiteralArgument[Any]):
+        r"""Construct a :class:`.FunctionElement`.
+
+        :param \*clauses: list of column expressions that form the arguments
+         of the SQL function call.
+
+        :param \**kwargs:  additional kwargs are typically consumed by
+         subclasses.
+
+        .. seealso::
+
+            :data:`.func`
+
+            :class:`.Function`
+
+        """
+        args: Sequence[_ColumnExpressionArgument[Any]] = [
+            coercions.expect(
+                roles.ExpressionElementRole,
+                c,
+                name=getattr(self, "name", None),
+                apply_propagate_attrs=self,
+            )
+            for c in clauses
+        ]
+        self._has_args = self._has_args or bool(args)
+        self.clause_expr = Grouping(
+            ClauseList(operator=operators.comma_op, group_contents=True, *args)
+        )
+
+    _non_anon_label = None
+
+    @property
+    def _proxy_key(self) -> Any:
+        return super()._proxy_key or getattr(self, "name", None)
+
+    def _execute_on_connection(
+        self,
+        connection: Connection,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> CursorResult[Any]:
+        return connection._execute_function(
+            self, distilled_params, execution_options
+        )
+
+    def scalar_table_valued(
+        self, name: str, type_: Optional[_TypeEngineArgument[_T]] = None
+    ) -> ScalarFunctionColumn[_T]:
+        """Return a column expression that's against this
+        :class:`_functions.FunctionElement` as a scalar
+        table-valued expression.
+
+        The returned expression is similar to that returned by a single column
+        accessed off of a :meth:`_functions.FunctionElement.table_valued`
+        construct, except no FROM clause is generated; the function is rendered
+        in the similar way as a scalar subquery.
+
+        E.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import func, select
+            >>> fn = func.jsonb_each("{'k', 'v'}").scalar_table_valued("key")
+            >>> print(select(fn))
+            {printsql}SELECT (jsonb_each(:jsonb_each_1)).key
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :meth:`_functions.FunctionElement.table_valued`
+
+            :meth:`_functions.FunctionElement.alias`
+
+            :meth:`_functions.FunctionElement.column_valued`
+
+        """  # noqa: E501
+
+        return ScalarFunctionColumn(self, name, type_)
+
+    def table_valued(
+        self, *expr: _ColumnExpressionOrStrLabelArgument[Any], **kw: Any
+    ) -> TableValuedAlias:
+        r"""Return a :class:`_sql.TableValuedAlias` representation of this
+        :class:`_functions.FunctionElement` with table-valued expressions added.
+
+        e.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> fn = func.generate_series(1, 5).table_valued(
+            ...     "value", "start", "stop", "step"
+            ... )
+
+            >>> print(select(fn))
+            {printsql}SELECT anon_1.value, anon_1.start, anon_1.stop, anon_1.step
+            FROM generate_series(:generate_series_1, :generate_series_2) AS anon_1{stop}
+
+            >>> print(select(fn.c.value, fn.c.stop).where(fn.c.value > 2))
+            {printsql}SELECT anon_1.value, anon_1.stop
+            FROM generate_series(:generate_series_1, :generate_series_2) AS anon_1
+            WHERE anon_1.value > :value_1{stop}
+
+        A WITH ORDINALITY expression may be generated by passing the keyword
+        argument "with_ordinality":
+
+        .. sourcecode:: pycon+sql
+
+            >>> fn = func.generate_series(4, 1, -1).table_valued(
+            ...     "gen", with_ordinality="ordinality"
+            ... )
+            >>> print(select(fn))
+            {printsql}SELECT anon_1.gen, anon_1.ordinality
+            FROM generate_series(:generate_series_1, :generate_series_2, :generate_series_3) WITH ORDINALITY AS anon_1
+
+        :param \*expr: A series of string column names that will be added to the
+         ``.c`` collection of the resulting :class:`_sql.TableValuedAlias`
+         construct as columns.  :func:`_sql.column` objects with or without
+         datatypes may also be used.
+
+        :param name: optional name to assign to the alias name that's generated.
+         If omitted, a unique anonymizing name is used.
+
+        :param with_ordinality: string name that when present results in the
+         ``WITH ORDINALITY`` clause being added to the alias, and the given
+         string name will be added as a column to the .c collection
+         of the resulting :class:`_sql.TableValuedAlias`.
+
+        :param joins_implicitly: when True, the table valued function may be
+         used in the FROM clause without any explicit JOIN to other tables
+         in the SQL query, and no "cartesian product" warning will be generated.
+         May be useful for SQL functions such as ``func.json_each()``.
+
+         .. versionadded:: 1.4.33
+
+        .. versionadded:: 1.4.0b2
+
+
+        .. seealso::
+
+            :ref:`tutorial_functions_table_valued` - in the :ref:`unified_tutorial`
+
+            :ref:`postgresql_table_valued` - in the :ref:`postgresql_toplevel` documentation
+
+            :meth:`_functions.FunctionElement.scalar_table_valued` - variant of
+            :meth:`_functions.FunctionElement.table_valued` which delivers the
+            complete table valued expression as a scalar column expression
+
+            :meth:`_functions.FunctionElement.column_valued`
+
+            :meth:`_sql.TableValuedAlias.render_derived` - renders the alias
+            using a derived column clause, e.g. ``AS name(col1, col2, ...)``
+
+        """  # noqa: 501
+
+        new_func = self._generate()
+
+        with_ordinality = kw.pop("with_ordinality", None)
+        joins_implicitly = kw.pop("joins_implicitly", None)
+        name = kw.pop("name", None)
+
+        if with_ordinality:
+            expr += (with_ordinality,)
+            new_func._with_ordinality = True
+
+        new_func.type = new_func._table_value_type = TableValueType(*expr)
+
+        return new_func.alias(name=name, joins_implicitly=joins_implicitly)
+
+    def column_valued(
+        self, name: Optional[str] = None, joins_implicitly: bool = False
+    ) -> TableValuedColumn[_T]:
+        """Return this :class:`_functions.FunctionElement` as a column expression that
+        selects from itself as a FROM clause.
+
+        E.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import select, func
+            >>> gs = func.generate_series(1, 5, -1).column_valued()
+            >>> print(select(gs))
+            {printsql}SELECT anon_1
+            FROM generate_series(:generate_series_1, :generate_series_2, :generate_series_3) AS anon_1
+
+        This is shorthand for::
+
+            gs = func.generate_series(1, 5, -1).alias().column
+
+        :param name: optional name to assign to the alias name that's generated.
+         If omitted, a unique anonymizing name is used.
+
+        :param joins_implicitly: when True, the "table" portion of the column
+         valued function may be a member of the FROM clause without any
+         explicit JOIN to other tables in the SQL query, and no "cartesian
+         product" warning will be generated. May be useful for SQL functions
+         such as ``func.json_array_elements()``.
+
+         .. versionadded:: 1.4.46
+
+        .. seealso::
+
+            :ref:`tutorial_functions_column_valued` - in the :ref:`unified_tutorial`
+
+            :ref:`postgresql_column_valued` - in the :ref:`postgresql_toplevel` documentation
+
+            :meth:`_functions.FunctionElement.table_valued`
+
+        """  # noqa: 501
+
+        return self.alias(name=name, joins_implicitly=joins_implicitly).column
+
+    @util.ro_non_memoized_property
+    def columns(self) -> ColumnCollection[str, KeyedColumnElement[Any]]:  # type: ignore[override]  # noqa: E501
+        r"""The set of columns exported by this :class:`.FunctionElement`.
+
+        This is a placeholder collection that allows the function to be
+        placed in the FROM clause of a statement:
+
+        .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import column, select, func
+            >>> stmt = select(column("x"), column("y")).select_from(func.myfunction())
+            >>> print(stmt)
+            {printsql}SELECT x, y FROM myfunction()
+
+        The above form is a legacy feature that is now superseded by the
+        fully capable :meth:`_functions.FunctionElement.table_valued`
+        method; see that method for details.
+
+        .. seealso::
+
+            :meth:`_functions.FunctionElement.table_valued` - generates table-valued
+            SQL function expressions.
+
+        """  # noqa: E501
+        return self.c
+
+    @util.ro_memoized_property
+    def c(self) -> ColumnCollection[str, KeyedColumnElement[Any]]:  # type: ignore[override]  # noqa: E501
+        """synonym for :attr:`.FunctionElement.columns`."""
+
+        return ColumnCollection(
+            columns=[(col.key, col) for col in self._all_selected_columns]
+        )
+
+    @property
+    def _all_selected_columns(self) -> Sequence[KeyedColumnElement[Any]]:
+        if is_table_value_type(self.type):
+            # TODO: this might not be fully accurate
+            cols = cast(
+                "Sequence[KeyedColumnElement[Any]]", self.type._elements
+            )
+        else:
+            cols = [self.label(None)]
+
+        return cols
+
+    @property
+    def exported_columns(  # type: ignore[override]
+        self,
+    ) -> ColumnCollection[str, KeyedColumnElement[Any]]:
+        return self.columns
+
+    @HasMemoized.memoized_attribute
+    def clauses(self) -> ClauseList:
+        """Return the underlying :class:`.ClauseList` which contains
+        the arguments for this :class:`.FunctionElement`.
+
+        """
+        return cast(ClauseList, self.clause_expr.element)
+
+    def over(
+        self,
+        *,
+        partition_by: Optional[_ByArgument] = None,
+        order_by: Optional[_ByArgument] = None,
+        rows: Optional[Tuple[Optional[int], Optional[int]]] = None,
+        range_: Optional[Tuple[Optional[int], Optional[int]]] = None,
+    ) -> Over[_T]:
+        """Produce an OVER clause against this function.
+
+        Used against aggregate or so-called "window" functions,
+        for database backends that support window functions.
+
+        The expression::
+
+            func.row_number().over(order_by="x")
+
+        is shorthand for::
+
+            from sqlalchemy import over
+
+            over(func.row_number(), order_by="x")
+
+        See :func:`_expression.over` for a full description.
+
+        .. seealso::
+
+            :func:`_expression.over`
+
+            :ref:`tutorial_window_functions` - in the :ref:`unified_tutorial`
+
+        """
+        return Over(
+            self,
+            partition_by=partition_by,
+            order_by=order_by,
+            rows=rows,
+            range_=range_,
+        )
+
+    def within_group(
+        self, *order_by: _ColumnExpressionArgument[Any]
+    ) -> WithinGroup[_T]:
+        """Produce a WITHIN GROUP (ORDER BY expr) clause against this function.
+
+        Used against so-called "ordered set aggregate" and "hypothetical
+        set aggregate" functions, including :class:`.percentile_cont`,
+        :class:`.rank`, :class:`.dense_rank`, etc.
+
+        See :func:`_expression.within_group` for a full description.
+
+        .. seealso::
+
+            :ref:`tutorial_functions_within_group` -
+            in the :ref:`unified_tutorial`
+
+
+        """
+        return WithinGroup(self, *order_by)
+
+    @overload
+    def filter(self) -> Self: ...
+
+    @overload
+    def filter(
+        self,
+        __criterion0: _ColumnExpressionArgument[bool],
+        *criterion: _ColumnExpressionArgument[bool],
+    ) -> FunctionFilter[_T]: ...
+
+    def filter(
+        self, *criterion: _ColumnExpressionArgument[bool]
+    ) -> Union[Self, FunctionFilter[_T]]:
+        """Produce a FILTER clause against this function.
+
+        Used against aggregate and window functions,
+        for database backends that support the "FILTER" clause.
+
+        The expression::
+
+            func.count(1).filter(True)
+
+        is shorthand for::
+
+            from sqlalchemy import funcfilter
+
+            funcfilter(func.count(1), True)
+
+        .. seealso::
+
+            :ref:`tutorial_functions_within_group` -
+            in the :ref:`unified_tutorial`
+
+            :class:`.FunctionFilter`
+
+            :func:`.funcfilter`
+
+
+        """
+        if not criterion:
+            return self
+        return FunctionFilter(self, *criterion)
+
+    def as_comparison(
+        self, left_index: int, right_index: int
+    ) -> FunctionAsBinary:
+        """Interpret this expression as a boolean comparison between two
+        values.
+
+        This method is used for an ORM use case described at
+        :ref:`relationship_custom_operator_sql_function`.
+
+        A hypothetical SQL function "is_equal()" which compares to values
+        for equality would be written in the Core expression language as::
+
+            expr = func.is_equal("a", "b")
+
+        If "is_equal()" above is comparing "a" and "b" for equality, the
+        :meth:`.FunctionElement.as_comparison` method would be invoked as::
+
+            expr = func.is_equal("a", "b").as_comparison(1, 2)
+
+        Where above, the integer value "1" refers to the first argument of the
+        "is_equal()" function and the integer value "2" refers to the second.
+
+        This would create a :class:`.BinaryExpression` that is equivalent to::
+
+            BinaryExpression("a", "b", operator=op.eq)
+
+        However, at the SQL level it would still render as
+        "is_equal('a', 'b')".
+
+        The ORM, when it loads a related object or collection, needs to be able
+        to manipulate the "left" and "right" sides of the ON clause of a JOIN
+        expression. The purpose of this method is to provide a SQL function
+        construct that can also supply this information to the ORM, when used
+        with the :paramref:`_orm.relationship.primaryjoin` parameter. The
+        return value is a containment object called :class:`.FunctionAsBinary`.
+
+        An ORM example is as follows::
+
+            class Venue(Base):
+                __tablename__ = "venue"
+                id = Column(Integer, primary_key=True)
+                name = Column(String)
+
+                descendants = relationship(
+                    "Venue",
+                    primaryjoin=func.instr(
+                        remote(foreign(name)), name + "/"
+                    ).as_comparison(1, 2)
+                    == 1,
+                    viewonly=True,
+                    order_by=name,
+                )
+
+        Above, the "Venue" class can load descendant "Venue" objects by
+        determining if the name of the parent Venue is contained within the
+        start of the hypothetical descendant value's name, e.g. "parent1" would
+        match up to "parent1/child1", but not to "parent2/child1".
+
+        Possible use cases include the "materialized path" example given above,
+        as well as making use of special SQL functions such as geometric
+        functions to create join conditions.
+
+        :param left_index: the integer 1-based index of the function argument
+         that serves as the "left" side of the expression.
+        :param right_index: the integer 1-based index of the function argument
+         that serves as the "right" side of the expression.
+
+        .. versionadded:: 1.3
+
+        .. seealso::
+
+            :ref:`relationship_custom_operator_sql_function` -
+            example use within the ORM
+
+        """
+        return FunctionAsBinary(self, left_index, right_index)
+
+    @property
+    def _from_objects(self) -> Any:
+        return self.clauses._from_objects
+
+    def within_group_type(
+        self, within_group: WithinGroup[_S]
+    ) -> Optional[TypeEngine[_S]]:
+        """For types that define their return type as based on the criteria
+        within a WITHIN GROUP (ORDER BY) expression, called by the
+        :class:`.WithinGroup` construct.
+
+        Returns None by default, in which case the function's normal ``.type``
+        is used.
+
+        """
+
+        return None
+
+    def alias(
+        self, name: Optional[str] = None, joins_implicitly: bool = False
+    ) -> TableValuedAlias:
+        r"""Produce a :class:`_expression.Alias` construct against this
+        :class:`.FunctionElement`.
+
+        .. tip::
+
+            The :meth:`_functions.FunctionElement.alias` method is part of the
+            mechanism by which "table valued" SQL functions are created.
+            However, most use cases are covered by higher level methods on
+            :class:`_functions.FunctionElement` including
+            :meth:`_functions.FunctionElement.table_valued`, and
+            :meth:`_functions.FunctionElement.column_valued`.
+
+        This construct wraps the function in a named alias which
+        is suitable for the FROM clause, in the style accepted for example
+        by PostgreSQL.  A column expression is also provided using the
+        special ``.column`` attribute, which may
+        be used to refer to the output of the function as a scalar value
+        in the columns or where clause, for a backend such as PostgreSQL.
+
+        For a full table-valued expression, use the
+        :meth:`_functions.FunctionElement.table_valued` method first to
+        establish named columns.
+
+        e.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import func, select, column
+            >>> data_view = func.unnest([1, 2, 3]).alias("data_view")
+            >>> print(select(data_view.column))
+            {printsql}SELECT data_view
+            FROM unnest(:unnest_1) AS data_view
+
+        The :meth:`_functions.FunctionElement.column_valued` method provides
+        a shortcut for the above pattern:
+
+        .. sourcecode:: pycon+sql
+
+            >>> data_view = func.unnest([1, 2, 3]).column_valued("data_view")
+            >>> print(select(data_view))
+            {printsql}SELECT data_view
+            FROM unnest(:unnest_1) AS data_view
+
+        .. versionadded:: 1.4.0b2  Added the ``.column`` accessor
+
+        :param name: alias name, will be rendered as ``AS <name>`` in the
+         FROM clause
+
+        :param joins_implicitly: when True, the table valued function may be
+         used in the FROM clause without any explicit JOIN to other tables
+         in the SQL query, and no "cartesian product" warning will be
+         generated.  May be useful for SQL functions such as
+         ``func.json_each()``.
+
+         .. versionadded:: 1.4.33
+
+        .. seealso::
+
+            :ref:`tutorial_functions_table_valued` -
+            in the :ref:`unified_tutorial`
+
+            :meth:`_functions.FunctionElement.table_valued`
+
+            :meth:`_functions.FunctionElement.scalar_table_valued`
+
+            :meth:`_functions.FunctionElement.column_valued`
+
+
+        """
+
+        return TableValuedAlias._construct(
+            self,
+            name=name,
+            table_value_type=self.type,
+            joins_implicitly=joins_implicitly,
+        )
+
+    def select(self) -> Select[Tuple[_T]]:
+        """Produce a :func:`_expression.select` construct
+        against this :class:`.FunctionElement`.
+
+        This is shorthand for::
+
+            s = select(function_element)
+
+        """
+        s: Select[Any] = Select(self)
+        if self._execution_options:
+            s = s.execution_options(**self._execution_options)
+        return s
+
+    def _bind_param(
+        self,
+        operator: OperatorType,
+        obj: Any,
+        type_: Optional[TypeEngine[_T]] = None,
+        expanding: bool = False,
+        **kw: Any,
+    ) -> BindParameter[_T]:
+        return BindParameter(
+            None,
+            obj,
+            _compared_to_operator=operator,
+            _compared_to_type=self.type,
+            unique=True,
+            type_=type_,
+            expanding=expanding,
+            **kw,
+        )
+
+    def self_group(self, against: Optional[OperatorType] = None) -> ClauseElement:  # type: ignore[override]  # noqa E501
+        # for the moment, we are parenthesizing all array-returning
+        # expressions against getitem.  This may need to be made
+        # more portable if in the future we support other DBs
+        # besides postgresql.
+        if against is operators.getitem and isinstance(
+            self.type, sqltypes.ARRAY
+        ):
+            return Grouping(self)
+        else:
+            return super().self_group(against=against)
+
+    @property
+    def entity_namespace(self) -> _EntityNamespace:
+        """overrides FromClause.entity_namespace as functions are generally
+        column expressions and not FromClauses.
+
+        """
+        # ideally functions would not be fromclauses but we failed to make
+        # this adjustment in 1.4
+        return _entity_namespace(self.clause_expr)
+
+
+class FunctionAsBinary(BinaryExpression[Any]):
+    _traverse_internals = [
+        ("sql_function", InternalTraversal.dp_clauseelement),
+        ("left_index", InternalTraversal.dp_plain_obj),
+        ("right_index", InternalTraversal.dp_plain_obj),
+        ("modifiers", InternalTraversal.dp_plain_dict),
+    ]
+
+    sql_function: FunctionElement[Any]
+    left_index: int
+    right_index: int
+
+    def _gen_cache_key(self, anon_map: Any, bindparams: Any) -> Any:
+        return ColumnElement._gen_cache_key(self, anon_map, bindparams)
+
+    def __init__(
+        self, fn: FunctionElement[Any], left_index: int, right_index: int
+    ):
+        self.sql_function = fn
+        self.left_index = left_index
+        self.right_index = right_index
+
+        self.operator = operators.function_as_comparison_op
+        self.type = sqltypes.BOOLEANTYPE
+        self.negate = None
+        self._is_implicitly_boolean = True
+        self.modifiers = {}
+
+    @property
+    def left_expr(self) -> ColumnElement[Any]:
+        return self.sql_function.clauses.clauses[self.left_index - 1]
+
+    @left_expr.setter
+    def left_expr(self, value: ColumnElement[Any]) -> None:
+        self.sql_function.clauses.clauses[self.left_index - 1] = value
+
+    @property
+    def right_expr(self) -> ColumnElement[Any]:
+        return self.sql_function.clauses.clauses[self.right_index - 1]
+
+    @right_expr.setter
+    def right_expr(self, value: ColumnElement[Any]) -> None:
+        self.sql_function.clauses.clauses[self.right_index - 1] = value
+
+    if not TYPE_CHECKING:
+        # mypy can't accommodate @property to replace an instance
+        # variable
+
+        left = left_expr
+        right = right_expr
+
+
+class ScalarFunctionColumn(NamedColumn[_T]):
+    __visit_name__ = "scalar_function_column"
+
+    _traverse_internals = [
+        ("name", InternalTraversal.dp_anon_name),
+        ("type", InternalTraversal.dp_type),
+        ("fn", InternalTraversal.dp_clauseelement),
+    ]
+
+    is_literal = False
+    table = None
+
+    def __init__(
+        self,
+        fn: FunctionElement[_T],
+        name: str,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+    ):
+        self.fn = fn
+        self.name = name
+
+        # if type is None, we get NULLTYPE, which is our _T.  But I don't
+        # know how to get the overloads to express that correctly
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+
+class _FunctionGenerator:
+    """Generate SQL function expressions.
+
+    :data:`.func` is a special object instance which generates SQL
+    functions based on name-based attributes, e.g.:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(func.count(1))
+        {printsql}count(:param_1)
+
+    The returned object is an instance of :class:`.Function`, and  is a
+    column-oriented SQL element like any other, and is used in that way:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(func.count(table.c.id)))
+        {printsql}SELECT count(sometable.id) FROM sometable
+
+    Any name can be given to :data:`.func`. If the function name is unknown to
+    SQLAlchemy, it will be rendered exactly as is. For common SQL functions
+    which SQLAlchemy is aware of, the name may be interpreted as a *generic
+    function* which will be compiled appropriately to the target database:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(func.current_timestamp())
+        {printsql}CURRENT_TIMESTAMP
+
+    To call functions which are present in dot-separated packages,
+    specify them in the same manner:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(func.stats.yield_curve(5, 10))
+        {printsql}stats.yield_curve(:yield_curve_1, :yield_curve_2)
+
+    SQLAlchemy can be made aware of the return type of functions to enable
+    type-specific lexical and result-based behavior. For example, to ensure
+    that a string-based function returns a Unicode value and is similarly
+    treated as a string in expressions, specify
+    :class:`~sqlalchemy.types.Unicode` as the type:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(
+        ...     func.my_string("hi", type_=Unicode)
+        ...     + " "
+        ...     + func.my_string("there", type_=Unicode)
+        ... )
+        {printsql}my_string(:my_string_1) || :my_string_2 || my_string(:my_string_3)
+
+    The object returned by a :data:`.func` call is usually an instance of
+    :class:`.Function`.
+    This object meets the "column" interface, including comparison and labeling
+    functions.  The object can also be passed the :meth:`~.Connectable.execute`
+    method of a :class:`_engine.Connection` or :class:`_engine.Engine`,
+    where it will be
+    wrapped inside of a SELECT statement first::
+
+        print(connection.execute(func.current_timestamp()).scalar())
+
+    In a few exception cases, the :data:`.func` accessor
+    will redirect a name to a built-in expression such as :func:`.cast`
+    or :func:`.extract`, as these names have well-known meaning
+    but are not exactly the same as "functions" from a SQLAlchemy
+    perspective.
+
+    Functions which are interpreted as "generic" functions know how to
+    calculate their return type automatically. For a listing of known generic
+    functions, see :ref:`generic_functions`.
+
+    .. note::
+
+        The :data:`.func` construct has only limited support for calling
+        standalone "stored procedures", especially those with special
+        parameterization concerns.
+
+        See the section :ref:`stored_procedures` for details on how to use
+        the DBAPI-level ``callproc()`` method for fully traditional stored
+        procedures.
+
+    .. seealso::
+
+        :ref:`tutorial_functions` - in the :ref:`unified_tutorial`
+
+        :class:`.Function`
+
+    """  # noqa
+
+    def __init__(self, **opts: Any):
+        self.__names: List[str] = []
+        self.opts = opts
+
+    def __getattr__(self, name: str) -> _FunctionGenerator:
+        # passthru __ attributes; fixes pydoc
+        if name.startswith("__"):
+            try:
+                return self.__dict__[name]  # type: ignore
+            except KeyError:
+                raise AttributeError(name)
+
+        elif name.endswith("_"):
+            name = name[0:-1]
+        f = _FunctionGenerator(**self.opts)
+        f.__names = list(self.__names) + [name]
+        return f
+
+    @overload
+    def __call__(
+        self, *c: Any, type_: _TypeEngineArgument[_T], **kwargs: Any
+    ) -> Function[_T]: ...
+
+    @overload
+    def __call__(self, *c: Any, **kwargs: Any) -> Function[Any]: ...
+
+    def __call__(self, *c: Any, **kwargs: Any) -> Function[Any]:
+        o = self.opts.copy()
+        o.update(kwargs)
+
+        tokens = len(self.__names)
+
+        if tokens == 2:
+            package, fname = self.__names
+        elif tokens == 1:
+            package, fname = "_default", self.__names[0]
+        else:
+            package = None
+
+        if package is not None:
+            func = _registry[package].get(fname.lower())
+            if func is not None:
+                return func(*c, **o)
+
+        return Function(
+            self.__names[-1], packagenames=tuple(self.__names[0:-1]), *c, **o
+        )
+
+    if TYPE_CHECKING:
+        # START GENERATED FUNCTION ACCESSORS
+
+        # code within this block is **programmatically,
+        # statically generated** by tools/generate_sql_functions.py
+
+        @property
+        def aggregate_strings(self) -> Type[aggregate_strings]: ...
+
+        @property
+        def ansifunction(self) -> Type[AnsiFunction[Any]]: ...
+
+        @property
+        def array_agg(self) -> Type[array_agg[Any]]: ...
+
+        @property
+        def cast(self) -> Type[Cast[Any]]: ...
+
+        @property
+        def char_length(self) -> Type[char_length]: ...
+
+        # set ColumnElement[_T] as a separate overload, to appease mypy
+        # which seems to not want to accept _T from _ColumnExpressionArgument.
+        # this is even if all non-generic types are removed from it, so
+        # reasons remain unclear for why this does not work
+
+        @overload
+        def coalesce(
+            self,
+            col: ColumnElement[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> coalesce[_T]: ...
+
+        @overload
+        def coalesce(
+            self,
+            col: _ColumnExpressionArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> coalesce[_T]: ...
+
+        @overload
+        def coalesce(
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> coalesce[_T]: ...
+
+        def coalesce(
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> coalesce[_T]: ...
+
+        @property
+        def concat(self) -> Type[concat]: ...
+
+        @property
+        def count(self) -> Type[count]: ...
+
+        @property
+        def cube(self) -> Type[cube[Any]]: ...
+
+        @property
+        def cume_dist(self) -> Type[cume_dist]: ...
+
+        @property
+        def current_date(self) -> Type[current_date]: ...
+
+        @property
+        def current_time(self) -> Type[current_time]: ...
+
+        @property
+        def current_timestamp(self) -> Type[current_timestamp]: ...
+
+        @property
+        def current_user(self) -> Type[current_user]: ...
+
+        @property
+        def dense_rank(self) -> Type[dense_rank]: ...
+
+        @property
+        def extract(self) -> Type[Extract]: ...
+
+        @property
+        def grouping_sets(self) -> Type[grouping_sets[Any]]: ...
+
+        @property
+        def localtime(self) -> Type[localtime]: ...
+
+        @property
+        def localtimestamp(self) -> Type[localtimestamp]: ...
+
+        # set ColumnElement[_T] as a separate overload, to appease mypy
+        # which seems to not want to accept _T from _ColumnExpressionArgument.
+        # this is even if all non-generic types are removed from it, so
+        # reasons remain unclear for why this does not work
+
+        @overload
+        def max(  # noqa: A001
+            self,
+            col: ColumnElement[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> max[_T]: ...
+
+        @overload
+        def max(  # noqa: A001
+            self,
+            col: _ColumnExpressionArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> max[_T]: ...
+
+        @overload
+        def max(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> max[_T]: ...
+
+        def max(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> max[_T]: ...
+
+        # set ColumnElement[_T] as a separate overload, to appease mypy
+        # which seems to not want to accept _T from _ColumnExpressionArgument.
+        # this is even if all non-generic types are removed from it, so
+        # reasons remain unclear for why this does not work
+
+        @overload
+        def min(  # noqa: A001
+            self,
+            col: ColumnElement[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> min[_T]: ...
+
+        @overload
+        def min(  # noqa: A001
+            self,
+            col: _ColumnExpressionArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> min[_T]: ...
+
+        @overload
+        def min(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> min[_T]: ...
+
+        def min(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> min[_T]: ...
+
+        @property
+        def mode(self) -> Type[mode[Any]]: ...
+
+        @property
+        def next_value(self) -> Type[next_value]: ...
+
+        @property
+        def now(self) -> Type[now]: ...
+
+        @property
+        def orderedsetagg(self) -> Type[OrderedSetAgg[Any]]: ...
+
+        @property
+        def percent_rank(self) -> Type[percent_rank]: ...
+
+        @property
+        def percentile_cont(self) -> Type[percentile_cont[Any]]: ...
+
+        @property
+        def percentile_disc(self) -> Type[percentile_disc[Any]]: ...
+
+        @property
+        def random(self) -> Type[random]: ...
+
+        @property
+        def rank(self) -> Type[rank]: ...
+
+        @property
+        def rollup(self) -> Type[rollup[Any]]: ...
+
+        @property
+        def session_user(self) -> Type[session_user]: ...
+
+        # set ColumnElement[_T] as a separate overload, to appease mypy
+        # which seems to not want to accept _T from _ColumnExpressionArgument.
+        # this is even if all non-generic types are removed from it, so
+        # reasons remain unclear for why this does not work
+
+        @overload
+        def sum(  # noqa: A001
+            self,
+            col: ColumnElement[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> sum[_T]: ...
+
+        @overload
+        def sum(  # noqa: A001
+            self,
+            col: _ColumnExpressionArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> sum[_T]: ...
+
+        @overload
+        def sum(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> sum[_T]: ...
+
+        def sum(  # noqa: A001
+            self,
+            col: _ColumnExpressionOrLiteralArgument[_T],
+            *args: _ColumnExpressionOrLiteralArgument[Any],
+            **kwargs: Any,
+        ) -> sum[_T]: ...
+
+        @property
+        def sysdate(self) -> Type[sysdate]: ...
+
+        @property
+        def user(self) -> Type[user]: ...
+
+        # END GENERATED FUNCTION ACCESSORS
+
+
+func = _FunctionGenerator()
+func.__doc__ = _FunctionGenerator.__doc__
+
+modifier = _FunctionGenerator(group=False)
+
+
+class Function(FunctionElement[_T]):
+    r"""Describe a named SQL function.
+
+    The :class:`.Function` object is typically generated from the
+    :data:`.func` generation object.
+
+
+    :param \*clauses: list of column expressions that form the arguments
+     of the SQL function call.
+
+    :param type\_: optional :class:`.TypeEngine` datatype object that will be
+     used as the return value of the column expression generated by this
+     function call.
+
+    :param packagenames: a string which indicates package prefix names
+     to be prepended to the function name when the SQL is generated.
+     The :data:`.func` generator creates these when it is called using
+     dotted format, e.g.::
+
+        func.mypackage.some_function(col1, col2)
+
+    .. seealso::
+
+        :ref:`tutorial_functions` - in the :ref:`unified_tutorial`
+
+        :data:`.func` - namespace which produces registered or ad-hoc
+        :class:`.Function` instances.
+
+        :class:`.GenericFunction` - allows creation of registered function
+        types.
+
+    """
+
+    __visit_name__ = "function"
+
+    _traverse_internals = FunctionElement._traverse_internals + [
+        ("packagenames", InternalTraversal.dp_plain_obj),
+        ("name", InternalTraversal.dp_string),
+        ("type", InternalTraversal.dp_type),
+    ]
+
+    name: str
+
+    identifier: str
+
+    type: TypeEngine[_T]
+    """A :class:`_types.TypeEngine` object which refers to the SQL return
+    type represented by this SQL function.
+
+    This datatype may be configured when generating a
+    :class:`_functions.Function` object by passing the
+    :paramref:`_functions.Function.type_` parameter, e.g.::
+
+        >>> select(func.lower("some VALUE", type_=String))
+
+    The small number of built-in classes of :class:`_functions.Function` come
+    with a built-in datatype that's appropriate to the class of function and
+    its arguments. For functions that aren't known, the type defaults to the
+    "null type".
+
+    """
+
+    @overload
+    def __init__(
+        self,
+        name: str,
+        *clauses: _ColumnExpressionOrLiteralArgument[_T],
+        type_: None = ...,
+        packagenames: Optional[Tuple[str, ...]] = ...,
+    ): ...
+
+    @overload
+    def __init__(
+        self,
+        name: str,
+        *clauses: _ColumnExpressionOrLiteralArgument[Any],
+        type_: _TypeEngineArgument[_T] = ...,
+        packagenames: Optional[Tuple[str, ...]] = ...,
+    ): ...
+
+    def __init__(
+        self,
+        name: str,
+        *clauses: _ColumnExpressionOrLiteralArgument[Any],
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        packagenames: Optional[Tuple[str, ...]] = None,
+    ):
+        """Construct a :class:`.Function`.
+
+        The :data:`.func` construct is normally used to construct
+        new :class:`.Function` instances.
+
+        """
+        self.packagenames = packagenames or ()
+        self.name = name
+
+        # if type is None, we get NULLTYPE, which is our _T.  But I don't
+        # know how to get the overloads to express that correctly
+        self.type = type_api.to_instance(type_)  # type: ignore
+
+        FunctionElement.__init__(self, *clauses)
+
+    def _bind_param(
+        self,
+        operator: OperatorType,
+        obj: Any,
+        type_: Optional[TypeEngine[_T]] = None,
+        expanding: bool = False,
+        **kw: Any,
+    ) -> BindParameter[_T]:
+        return BindParameter(
+            self.name,
+            obj,
+            _compared_to_operator=operator,
+            _compared_to_type=self.type,
+            type_=type_,
+            unique=True,
+            expanding=expanding,
+            **kw,
+        )
+
+
+class GenericFunction(Function[_T]):
+    """Define a 'generic' function.
+
+    A generic function is a pre-established :class:`.Function`
+    class that is instantiated automatically when called
+    by name from the :data:`.func` attribute.    Note that
+    calling any name from :data:`.func` has the effect that
+    a new :class:`.Function` instance is created automatically,
+    given that name.  The primary use case for defining
+    a :class:`.GenericFunction` class is so that a function
+    of a particular name may be given a fixed return type.
+    It can also include custom argument parsing schemes as well
+    as additional methods.
+
+    Subclasses of :class:`.GenericFunction` are automatically
+    registered under the name of the class.  For
+    example, a user-defined function ``as_utc()`` would
+    be available immediately::
+
+        from sqlalchemy.sql.functions import GenericFunction
+        from sqlalchemy.types import DateTime
+
+
+        class as_utc(GenericFunction):
+            type = DateTime()
+            inherit_cache = True
+
+
+        print(select(func.as_utc()))
+
+    User-defined generic functions can be organized into
+    packages by specifying the "package" attribute when defining
+    :class:`.GenericFunction`.   Third party libraries
+    containing many functions may want to use this in order
+    to avoid name conflicts with other systems.   For example,
+    if our ``as_utc()`` function were part of a package
+    "time"::
+
+        class as_utc(GenericFunction):
+            type = DateTime()
+            package = "time"
+            inherit_cache = True
+
+    The above function would be available from :data:`.func`
+    using the package name ``time``::
+
+        print(select(func.time.as_utc()))
+
+    A final option is to allow the function to be accessed
+    from one name in :data:`.func` but to render as a different name.
+    The ``identifier`` attribute will override the name used to
+    access the function as loaded from :data:`.func`, but will retain
+    the usage of ``name`` as the rendered name::
+
+        class GeoBuffer(GenericFunction):
+            type = Geometry()
+            package = "geo"
+            name = "ST_Buffer"
+            identifier = "buffer"
+            inherit_cache = True
+
+    The above function will render as follows:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(func.geo.buffer())
+        {printsql}ST_Buffer()
+
+    The name will be rendered as is, however without quoting unless the name
+    contains special characters that require quoting.  To force quoting
+    on or off for the name, use the :class:`.sqlalchemy.sql.quoted_name`
+    construct::
+
+        from sqlalchemy.sql import quoted_name
+
+
+        class GeoBuffer(GenericFunction):
+            type = Geometry()
+            package = "geo"
+            name = quoted_name("ST_Buffer", True)
+            identifier = "buffer"
+            inherit_cache = True
+
+    The above function will render as:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(func.geo.buffer())
+        {printsql}"ST_Buffer"()
+
+    Type parameters for this class as a
+    `generic type <https://peps.python.org/pep-0484/#generics>`_ can be passed
+    and should match the type seen in a :class:`_engine.Result`. For example::
+
+        class as_utc(GenericFunction[datetime.datetime]):
+            type = DateTime()
+            inherit_cache = True
+
+    The above indicates that the following expression returns a ``datetime``
+    object::
+
+        connection.scalar(select(func.as_utc()))
+
+    .. versionadded:: 1.3.13  The :class:`.quoted_name` construct is now
+       recognized for quoting when used with the "name" attribute of the
+       object, so that quoting can be forced on or off for the function
+       name.
+
+
+    """
+
+    coerce_arguments = True
+    inherit_cache = True
+
+    _register: bool
+
+    name = "GenericFunction"
+
+    def __init_subclass__(cls) -> None:
+        if annotation.Annotated not in cls.__mro__:
+            cls._register_generic_function(cls.__name__, cls.__dict__)
+        super().__init_subclass__()
+
+    @classmethod
+    def _register_generic_function(
+        cls, clsname: str, clsdict: Mapping[str, Any]
+    ) -> None:
+        cls.name = name = clsdict.get("name", clsname)
+        cls.identifier = identifier = clsdict.get("identifier", name)
+        package = clsdict.get("package", "_default")
+        # legacy
+        if "__return_type__" in clsdict:
+            cls.type = clsdict["__return_type__"]
+
+        # Check _register attribute status
+        cls._register = getattr(cls, "_register", True)
+
+        # Register the function if required
+        if cls._register:
+            register_function(identifier, cls, package)
+        else:
+            # Set _register to True to register child classes by default
+            cls._register = True
+
+    def __init__(
+        self, *args: _ColumnExpressionOrLiteralArgument[Any], **kwargs: Any
+    ):
+        parsed_args = kwargs.pop("_parsed_args", None)
+        if parsed_args is None:
+            parsed_args = [
+                coercions.expect(
+                    roles.ExpressionElementRole,
+                    c,
+                    name=self.name,
+                    apply_propagate_attrs=self,
+                )
+                for c in args
+            ]
+        self._has_args = self._has_args or bool(parsed_args)
+        self.packagenames = ()
+
+        self.clause_expr = Grouping(
+            ClauseList(
+                operator=operators.comma_op, group_contents=True, *parsed_args
+            )
+        )
+
+        self.type = type_api.to_instance(  # type: ignore
+            kwargs.pop("type_", None) or getattr(self, "type", None)
+        )
+
+
+register_function("cast", Cast)  # type: ignore
+register_function("extract", Extract)  # type: ignore
+
+
+class next_value(GenericFunction[int]):
+    """Represent the 'next value', given a :class:`.Sequence`
+    as its single argument.
+
+    Compiles into the appropriate function on each backend,
+    or will raise NotImplementedError if used on a backend
+    that does not provide support for sequences.
+
+    """
+
+    type = sqltypes.Integer()
+    name = "next_value"
+
+    _traverse_internals = [
+        ("sequence", InternalTraversal.dp_named_ddl_element)
+    ]
+
+    def __init__(self, seq: schema.Sequence, **kw: Any):
+        assert isinstance(
+            seq, schema.Sequence
+        ), "next_value() accepts a Sequence object as input."
+        self.sequence = seq
+        self.type = sqltypes.to_instance(  # type: ignore
+            seq.data_type or getattr(self, "type", None)
+        )
+
+    def compare(self, other: Any, **kw: Any) -> bool:
+        return (
+            isinstance(other, next_value)
+            and self.sequence.name == other.sequence.name
+        )
+
+    @property
+    def _from_objects(self) -> Any:
+        return []
+
+
+class AnsiFunction(GenericFunction[_T]):
+    """Define a function in "ansi" format, which doesn't render parenthesis."""
+
+    inherit_cache = True
+
+    def __init__(self, *args: _ColumnExpressionArgument[Any], **kwargs: Any):
+        GenericFunction.__init__(self, *args, **kwargs)
+
+
+class ReturnTypeFromArgs(GenericFunction[_T]):
+    """Define a function whose return type is the same as its arguments."""
+
+    inherit_cache = True
+
+    # set ColumnElement[_T] as a separate overload, to appease mypy which seems
+    # to not want to accept _T from _ColumnExpressionArgument.  this is even if
+    # all non-generic types are removed from it, so reasons remain unclear for
+    # why this does not work
+
+    @overload
+    def __init__(
+        self,
+        col: ColumnElement[_T],
+        *args: _ColumnExpressionOrLiteralArgument[Any],
+        **kwargs: Any,
+    ): ...
+
+    @overload
+    def __init__(
+        self,
+        col: _ColumnExpressionArgument[_T],
+        *args: _ColumnExpressionOrLiteralArgument[Any],
+        **kwargs: Any,
+    ): ...
+
+    @overload
+    def __init__(
+        self,
+        col: _ColumnExpressionOrLiteralArgument[_T],
+        *args: _ColumnExpressionOrLiteralArgument[Any],
+        **kwargs: Any,
+    ): ...
+
+    def __init__(
+        self, *args: _ColumnExpressionOrLiteralArgument[Any], **kwargs: Any
+    ):
+        fn_args: Sequence[ColumnElement[Any]] = [
+            coercions.expect(
+                roles.ExpressionElementRole,
+                c,
+                name=self.name,
+                apply_propagate_attrs=self,
+            )
+            for c in args
+        ]
+        kwargs.setdefault("type_", _type_from_args(fn_args))
+        kwargs["_parsed_args"] = fn_args
+        super().__init__(*fn_args, **kwargs)
+
+
+class coalesce(ReturnTypeFromArgs[_T]):
+    _has_args = True
+    inherit_cache = True
+
+
+class max(ReturnTypeFromArgs[_T]):  # noqa:  A001
+    """The SQL MAX() aggregate function."""
+
+    inherit_cache = True
+
+
+class min(ReturnTypeFromArgs[_T]):  # noqa: A001
+    """The SQL MIN() aggregate function."""
+
+    inherit_cache = True
+
+
+class sum(ReturnTypeFromArgs[_T]):  # noqa: A001
+    """The SQL SUM() aggregate function."""
+
+    inherit_cache = True
+
+
+class now(GenericFunction[datetime.datetime]):
+    """The SQL now() datetime function.
+
+    SQLAlchemy dialects will usually render this particular function
+    in a backend-specific way, such as rendering it as ``CURRENT_TIMESTAMP``.
+
+    """
+
+    type = sqltypes.DateTime()
+    inherit_cache = True
+
+
+class concat(GenericFunction[str]):
+    """The SQL CONCAT() function, which concatenates strings.
+
+    E.g.:
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(func.concat("a", "b")))
+        {printsql}SELECT concat(:concat_2, :concat_3) AS concat_1
+
+    String concatenation in SQLAlchemy is more commonly available using the
+    Python ``+`` operator with string datatypes, which will render a
+    backend-specific concatenation operator, such as :
+
+    .. sourcecode:: pycon+sql
+
+        >>> print(select(literal("a") + "b"))
+        {printsql}SELECT :param_1 || :param_2 AS anon_1
+
+
+    """
+
+    type = sqltypes.String()
+    inherit_cache = True
+
+
+class char_length(GenericFunction[int]):
+    """The CHAR_LENGTH() SQL function."""
+
+    type = sqltypes.Integer()
+    inherit_cache = True
+
+    def __init__(self, arg: _ColumnExpressionArgument[str], **kw: Any):
+        # slight hack to limit to just one positional argument
+        # not sure why this one function has this special treatment
+        super().__init__(arg, **kw)
+
+
+class random(GenericFunction[float]):
+    """The RANDOM() SQL function."""
+
+    _has_args = True
+    inherit_cache = True
+
+
+class count(GenericFunction[int]):
+    r"""The ANSI COUNT aggregate function.  With no arguments,
+    emits COUNT \*.
+
+    E.g.::
+
+        from sqlalchemy import func
+        from sqlalchemy import select
+        from sqlalchemy import table, column
+
+        my_table = table("some_table", column("id"))
+
+        stmt = select(func.count()).select_from(my_table)
+
+    Executing ``stmt`` would emit:
+
+    .. sourcecode:: sql
+
+        SELECT count(*) AS count_1
+        FROM some_table
+
+
+    """
+
+    type = sqltypes.Integer()
+    inherit_cache = True
+
+    def __init__(
+        self,
+        expression: Union[
+            _ColumnExpressionArgument[Any], _StarOrOne, None
+        ] = None,
+        **kwargs: Any,
+    ):
+        if expression is None:
+            expression = literal_column("*")
+        super().__init__(expression, **kwargs)
+
+
+class current_date(AnsiFunction[datetime.date]):
+    """The CURRENT_DATE() SQL function."""
+
+    type = sqltypes.Date()
+    inherit_cache = True
+
+
+class current_time(AnsiFunction[datetime.time]):
+    """The CURRENT_TIME() SQL function."""
+
+    type = sqltypes.Time()
+    inherit_cache = True
+
+
+class current_timestamp(AnsiFunction[datetime.datetime]):
+    """The CURRENT_TIMESTAMP() SQL function."""
+
+    type = sqltypes.DateTime()
+    inherit_cache = True
+
+
+class current_user(AnsiFunction[str]):
+    """The CURRENT_USER() SQL function."""
+
+    type = sqltypes.String()
+    inherit_cache = True
+
+
+class localtime(AnsiFunction[datetime.datetime]):
+    """The localtime() SQL function."""
+
+    type = sqltypes.DateTime()
+    inherit_cache = True
+
+
+class localtimestamp(AnsiFunction[datetime.datetime]):
+    """The localtimestamp() SQL function."""
+
+    type = sqltypes.DateTime()
+    inherit_cache = True
+
+
+class session_user(AnsiFunction[str]):
+    """The SESSION_USER() SQL function."""
+
+    type = sqltypes.String()
+    inherit_cache = True
+
+
+class sysdate(AnsiFunction[datetime.datetime]):
+    """The SYSDATE() SQL function."""
+
+    type = sqltypes.DateTime()
+    inherit_cache = True
+
+
+class user(AnsiFunction[str]):
+    """The USER() SQL function."""
+
+    type = sqltypes.String()
+    inherit_cache = True
+
+
+class array_agg(GenericFunction[_T]):
+    """Support for the ARRAY_AGG function.
+
+    The ``func.array_agg(expr)`` construct returns an expression of
+    type :class:`_types.ARRAY`.
+
+    e.g.::
+
+        stmt = select(func.array_agg(table.c.values)[2:5])
+
+    .. seealso::
+
+        :func:`_postgresql.array_agg` - PostgreSQL-specific version that
+        returns :class:`_postgresql.ARRAY`, which has PG-specific operators
+        added.
+
+    """
+
+    inherit_cache = True
+
+    def __init__(self, *args: _ColumnExpressionArgument[Any], **kwargs: Any):
+        fn_args: Sequence[ColumnElement[Any]] = [
+            coercions.expect(
+                roles.ExpressionElementRole, c, apply_propagate_attrs=self
+            )
+            for c in args
+        ]
+
+        default_array_type = kwargs.pop("_default_array_type", sqltypes.ARRAY)
+        if "type_" not in kwargs:
+            type_from_args = _type_from_args(fn_args)
+            if isinstance(type_from_args, sqltypes.ARRAY):
+                kwargs["type_"] = type_from_args
+            else:
+                kwargs["type_"] = default_array_type(
+                    type_from_args, dimensions=1
+                )
+        kwargs["_parsed_args"] = fn_args
+        super().__init__(*fn_args, **kwargs)
+
+
+class OrderedSetAgg(GenericFunction[_T]):
+    """Define a function where the return type is based on the sort
+    expression type as defined by the expression passed to the
+    :meth:`.FunctionElement.within_group` method."""
+
+    array_for_multi_clause = False
+    inherit_cache = True
+
+    def within_group_type(
+        self, within_group: WithinGroup[Any]
+    ) -> TypeEngine[Any]:
+        func_clauses = cast(ClauseList, self.clause_expr.element)
+        order_by: Sequence[ColumnElement[Any]] = sqlutil.unwrap_order_by(
+            within_group.order_by
+        )
+        if self.array_for_multi_clause and len(func_clauses.clauses) > 1:
+            return sqltypes.ARRAY(order_by[0].type)
+        else:
+            return order_by[0].type
+
+
+class mode(OrderedSetAgg[_T]):
+    """Implement the ``mode`` ordered-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is the same as the sort expression.
+
+    """
+
+    inherit_cache = True
+
+
+class percentile_cont(OrderedSetAgg[_T]):
+    """Implement the ``percentile_cont`` ordered-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is the same as the sort expression,
+    or if the arguments are an array, an :class:`_types.ARRAY` of the sort
+    expression's type.
+
+    """
+
+    array_for_multi_clause = True
+    inherit_cache = True
+
+
+class percentile_disc(OrderedSetAgg[_T]):
+    """Implement the ``percentile_disc`` ordered-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is the same as the sort expression,
+    or if the arguments are an array, an :class:`_types.ARRAY` of the sort
+    expression's type.
+
+    """
+
+    array_for_multi_clause = True
+    inherit_cache = True
+
+
+class rank(GenericFunction[int]):
+    """Implement the ``rank`` hypothetical-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is :class:`.Integer`.
+
+    """
+
+    type = sqltypes.Integer()
+    inherit_cache = True
+
+
+class dense_rank(GenericFunction[int]):
+    """Implement the ``dense_rank`` hypothetical-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is :class:`.Integer`.
+
+    """
+
+    type = sqltypes.Integer()
+    inherit_cache = True
+
+
+class percent_rank(GenericFunction[decimal.Decimal]):
+    """Implement the ``percent_rank`` hypothetical-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is :class:`.Numeric`.
+
+    """
+
+    type: sqltypes.Numeric[decimal.Decimal] = sqltypes.Numeric()
+    inherit_cache = True
+
+
+class cume_dist(GenericFunction[decimal.Decimal]):
+    """Implement the ``cume_dist`` hypothetical-set aggregate function.
+
+    This function must be used with the :meth:`.FunctionElement.within_group`
+    modifier to supply a sort expression to operate upon.
+
+    The return type of this function is :class:`.Numeric`.
+
+    """
+
+    type: sqltypes.Numeric[decimal.Decimal] = sqltypes.Numeric()
+    inherit_cache = True
+
+
+class cube(GenericFunction[_T]):
+    r"""Implement the ``CUBE`` grouping operation.
+
+    This function is used as part of the GROUP BY of a statement,
+    e.g. :meth:`_expression.Select.group_by`::
+
+        stmt = select(
+            func.sum(table.c.value), table.c.col_1, table.c.col_2
+        ).group_by(func.cube(table.c.col_1, table.c.col_2))
+
+    .. versionadded:: 1.2
+
+    """
+
+    _has_args = True
+    inherit_cache = True
+
+
+class rollup(GenericFunction[_T]):
+    r"""Implement the ``ROLLUP`` grouping operation.
+
+    This function is used as part of the GROUP BY of a statement,
+    e.g. :meth:`_expression.Select.group_by`::
+
+        stmt = select(
+            func.sum(table.c.value), table.c.col_1, table.c.col_2
+        ).group_by(func.rollup(table.c.col_1, table.c.col_2))
+
+    .. versionadded:: 1.2
+
+    """
+
+    _has_args = True
+    inherit_cache = True
+
+
+class grouping_sets(GenericFunction[_T]):
+    r"""Implement the ``GROUPING SETS`` grouping operation.
+
+    This function is used as part of the GROUP BY of a statement,
+    e.g. :meth:`_expression.Select.group_by`::
+
+        stmt = select(
+            func.sum(table.c.value), table.c.col_1, table.c.col_2
+        ).group_by(func.grouping_sets(table.c.col_1, table.c.col_2))
+
+    In order to group by multiple sets, use the :func:`.tuple_` construct::
+
+        from sqlalchemy import tuple_
+
+        stmt = select(
+            func.sum(table.c.value), table.c.col_1, table.c.col_2, table.c.col_3
+        ).group_by(
+            func.grouping_sets(
+                tuple_(table.c.col_1, table.c.col_2),
+                tuple_(table.c.value, table.c.col_3),
+            )
+        )
+
+    .. versionadded:: 1.2
+
+    """  # noqa: E501
+
+    _has_args = True
+    inherit_cache = True
+
+
+class aggregate_strings(GenericFunction[str]):
+    """Implement a generic string aggregation function.
+
+    This function will concatenate non-null values into a string and
+    separate the values by a delimiter.
+
+    This function is compiled on a per-backend basis, into functions
+    such as ``group_concat()``, ``string_agg()``, or ``LISTAGG()``.
+
+    e.g. Example usage with delimiter '.'::
+
+        stmt = select(func.aggregate_strings(table.c.str_col, "."))
+
+    The return type of this function is :class:`.String`.
+
+    .. versionadded: 2.0.21
+
+    """
+
+    type = sqltypes.String()
+    _has_args = True
+    inherit_cache = True
+
+    def __init__(self, clause: _ColumnExpressionArgument[Any], separator: str):
+        super().__init__(clause, separator)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/lambdas.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/lambdas.py
new file mode 100644
index 00000000..8d70f800
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/lambdas.py
@@ -0,0 +1,1443 @@
+# sql/lambdas.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import inspect
+import itertools
+import operator
+import threading
+import types
+from types import CodeType
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import List
+from typing import MutableMapping
+from typing import Optional
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import weakref
+
+from . import cache_key as _cache_key
+from . import coercions
+from . import elements
+from . import roles
+from . import schema
+from . import visitors
+from .base import _clone
+from .base import Executable
+from .base import Options
+from .cache_key import CacheConst
+from .operators import ColumnOperators
+from .. import exc
+from .. import inspection
+from .. import util
+from ..util.typing import Literal
+
+
+if TYPE_CHECKING:
+    from .elements import BindParameter
+    from .elements import ClauseElement
+    from .roles import SQLRole
+    from .visitors import _CloneCallableType
+
+_LambdaCacheType = MutableMapping[
+    Tuple[Any, ...], Union["NonAnalyzedFunction", "AnalyzedFunction"]
+]
+_BoundParameterGetter = Callable[..., Any]
+
+_closure_per_cache_key: _LambdaCacheType = util.LRUCache(1000)
+
+
+_LambdaType = Callable[[], Any]
+
+_AnyLambdaType = Callable[..., Any]
+
+_StmtLambdaType = Callable[[], Any]
+
+_E = TypeVar("_E", bound=Executable)
+_StmtLambdaElementType = Callable[[_E], Any]
+
+
+class LambdaOptions(Options):
+    enable_tracking = True
+    track_closure_variables = True
+    track_on: Optional[object] = None
+    global_track_bound_values = True
+    track_bound_values = True
+    lambda_cache: Optional[_LambdaCacheType] = None
+
+
+def lambda_stmt(
+    lmb: _StmtLambdaType,
+    enable_tracking: bool = True,
+    track_closure_variables: bool = True,
+    track_on: Optional[object] = None,
+    global_track_bound_values: bool = True,
+    track_bound_values: bool = True,
+    lambda_cache: Optional[_LambdaCacheType] = None,
+) -> StatementLambdaElement:
+    """Produce a SQL statement that is cached as a lambda.
+
+    The Python code object within the lambda is scanned for both Python
+    literals that will become bound parameters as well as closure variables
+    that refer to Core or ORM constructs that may vary.   The lambda itself
+    will be invoked only once per particular set of constructs detected.
+
+    E.g.::
+
+        from sqlalchemy import lambda_stmt
+
+        stmt = lambda_stmt(lambda: table.select())
+        stmt += lambda s: s.where(table.c.id == 5)
+
+        result = connection.execute(stmt)
+
+    The object returned is an instance of :class:`_sql.StatementLambdaElement`.
+
+    .. versionadded:: 1.4
+
+    :param lmb: a Python function, typically a lambda, which takes no arguments
+     and returns a SQL expression construct
+    :param enable_tracking: when False, all scanning of the given lambda for
+     changes in closure variables or bound parameters is disabled.  Use for
+     a lambda that produces the identical results in all cases with no
+     parameterization.
+    :param track_closure_variables: when False, changes in closure variables
+     within the lambda will not be scanned.   Use for a lambda where the
+     state of its closure variables will never change the SQL structure
+     returned by the lambda.
+    :param track_bound_values: when False, bound parameter tracking will
+     be disabled for the given lambda.  Use for a lambda that either does
+     not produce any bound values, or where the initial bound values never
+     change.
+    :param global_track_bound_values: when False, bound parameter tracking
+     will be disabled for the entire statement including additional links
+     added via the :meth:`_sql.StatementLambdaElement.add_criteria` method.
+    :param lambda_cache: a dictionary or other mapping-like object where
+     information about the lambda's Python code as well as the tracked closure
+     variables in the lambda itself will be stored.   Defaults
+     to a global LRU cache.  This cache is independent of the "compiled_cache"
+     used by the :class:`_engine.Connection` object.
+
+    .. seealso::
+
+        :ref:`engine_lambda_caching`
+
+
+    """
+
+    return StatementLambdaElement(
+        lmb,
+        roles.StatementRole,
+        LambdaOptions(
+            enable_tracking=enable_tracking,
+            track_on=track_on,
+            track_closure_variables=track_closure_variables,
+            global_track_bound_values=global_track_bound_values,
+            track_bound_values=track_bound_values,
+            lambda_cache=lambda_cache,
+        ),
+    )
+
+
+class LambdaElement(elements.ClauseElement):
+    """A SQL construct where the state is stored as an un-invoked lambda.
+
+    The :class:`_sql.LambdaElement` is produced transparently whenever
+    passing lambda expressions into SQL constructs, such as::
+
+        stmt = select(table).where(lambda: table.c.col == parameter)
+
+    The :class:`_sql.LambdaElement` is the base of the
+    :class:`_sql.StatementLambdaElement` which represents a full statement
+    within a lambda.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`engine_lambda_caching`
+
+    """
+
+    __visit_name__ = "lambda_element"
+
+    _is_lambda_element = True
+
+    _traverse_internals = [
+        ("_resolved", visitors.InternalTraversal.dp_clauseelement)
+    ]
+
+    _transforms: Tuple[_CloneCallableType, ...] = ()
+
+    _resolved_bindparams: List[BindParameter[Any]]
+    parent_lambda: Optional[StatementLambdaElement] = None
+    closure_cache_key: Union[Tuple[Any, ...], Literal[CacheConst.NO_CACHE]]
+    role: Type[SQLRole]
+    _rec: Union[AnalyzedFunction, NonAnalyzedFunction]
+    fn: _AnyLambdaType
+    tracker_key: Tuple[CodeType, ...]
+
+    def __repr__(self):
+        return "%s(%r)" % (
+            self.__class__.__name__,
+            self.fn.__code__,
+        )
+
+    def __init__(
+        self,
+        fn: _LambdaType,
+        role: Type[SQLRole],
+        opts: Union[Type[LambdaOptions], LambdaOptions] = LambdaOptions,
+        apply_propagate_attrs: Optional[ClauseElement] = None,
+    ):
+        self.fn = fn
+        self.role = role
+        self.tracker_key = (fn.__code__,)
+        self.opts = opts
+
+        if apply_propagate_attrs is None and (role is roles.StatementRole):
+            apply_propagate_attrs = self
+
+        rec = self._retrieve_tracker_rec(fn, apply_propagate_attrs, opts)
+
+        if apply_propagate_attrs is not None:
+            propagate_attrs = rec.propagate_attrs
+            if propagate_attrs:
+                apply_propagate_attrs._propagate_attrs = propagate_attrs
+
+    def _retrieve_tracker_rec(self, fn, apply_propagate_attrs, opts):
+        lambda_cache = opts.lambda_cache
+        if lambda_cache is None:
+            lambda_cache = _closure_per_cache_key
+
+        tracker_key = self.tracker_key
+
+        fn = self.fn
+        closure = fn.__closure__
+        tracker = AnalyzedCode.get(
+            fn,
+            self,
+            opts,
+        )
+
+        bindparams: List[BindParameter[Any]]
+        self._resolved_bindparams = bindparams = []
+
+        if self.parent_lambda is not None:
+            parent_closure_cache_key = self.parent_lambda.closure_cache_key
+        else:
+            parent_closure_cache_key = ()
+
+        cache_key: Union[Tuple[Any, ...], Literal[CacheConst.NO_CACHE]]
+
+        if parent_closure_cache_key is not _cache_key.NO_CACHE:
+            anon_map = visitors.anon_map()
+            cache_key = tuple(
+                [
+                    getter(closure, opts, anon_map, bindparams)
+                    for getter in tracker.closure_trackers
+                ]
+            )
+
+            if _cache_key.NO_CACHE not in anon_map:
+                cache_key = parent_closure_cache_key + cache_key
+
+                self.closure_cache_key = cache_key
+
+                try:
+                    rec = lambda_cache[tracker_key + cache_key]
+                except KeyError:
+                    rec = None
+            else:
+                cache_key = _cache_key.NO_CACHE
+                rec = None
+
+        else:
+            cache_key = _cache_key.NO_CACHE
+            rec = None
+
+        self.closure_cache_key = cache_key
+
+        if rec is None:
+            if cache_key is not _cache_key.NO_CACHE:
+                with AnalyzedCode._generation_mutex:
+                    key = tracker_key + cache_key
+                    if key not in lambda_cache:
+                        rec = AnalyzedFunction(
+                            tracker, self, apply_propagate_attrs, fn
+                        )
+                        rec.closure_bindparams = list(bindparams)
+                        lambda_cache[key] = rec
+                    else:
+                        rec = lambda_cache[key]
+            else:
+                rec = NonAnalyzedFunction(self._invoke_user_fn(fn))
+
+        else:
+            bindparams[:] = [
+                orig_bind._with_value(new_bind.value, maintain_key=True)
+                for orig_bind, new_bind in zip(
+                    rec.closure_bindparams, bindparams
+                )
+            ]
+
+        self._rec = rec
+
+        if cache_key is not _cache_key.NO_CACHE:
+            if self.parent_lambda is not None:
+                bindparams[:0] = self.parent_lambda._resolved_bindparams
+
+            lambda_element: Optional[LambdaElement] = self
+            while lambda_element is not None:
+                rec = lambda_element._rec
+                if rec.bindparam_trackers:
+                    tracker_instrumented_fn = rec.tracker_instrumented_fn
+                    for tracker in rec.bindparam_trackers:
+                        tracker(
+                            lambda_element.fn,
+                            tracker_instrumented_fn,
+                            bindparams,
+                        )
+                lambda_element = lambda_element.parent_lambda
+
+        return rec
+
+    def __getattr__(self, key):
+        return getattr(self._rec.expected_expr, key)
+
+    @property
+    def _is_sequence(self):
+        return self._rec.is_sequence
+
+    @property
+    def _select_iterable(self):
+        if self._is_sequence:
+            return itertools.chain.from_iterable(
+                [element._select_iterable for element in self._resolved]
+            )
+
+        else:
+            return self._resolved._select_iterable
+
+    @property
+    def _from_objects(self):
+        if self._is_sequence:
+            return itertools.chain.from_iterable(
+                [element._from_objects for element in self._resolved]
+            )
+
+        else:
+            return self._resolved._from_objects
+
+    def _param_dict(self):
+        return {b.key: b.value for b in self._resolved_bindparams}
+
+    def _setup_binds_for_tracked_expr(self, expr):
+        bindparam_lookup = {b.key: b for b in self._resolved_bindparams}
+
+        def replace(
+            element: Optional[visitors.ExternallyTraversible], **kw: Any
+        ) -> Optional[visitors.ExternallyTraversible]:
+            if isinstance(element, elements.BindParameter):
+                if element.key in bindparam_lookup:
+                    bind = bindparam_lookup[element.key]
+                    if element.expanding:
+                        bind.expanding = True
+                        bind.expand_op = element.expand_op
+                        bind.type = element.type
+                    return bind
+
+            return None
+
+        if self._rec.is_sequence:
+            expr = [
+                visitors.replacement_traverse(sub_expr, {}, replace)
+                for sub_expr in expr
+            ]
+        elif getattr(expr, "is_clause_element", False):
+            expr = visitors.replacement_traverse(expr, {}, replace)
+
+        return expr
+
+    def _copy_internals(
+        self,
+        clone: _CloneCallableType = _clone,
+        deferred_copy_internals: Optional[_CloneCallableType] = None,
+        **kw: Any,
+    ) -> None:
+        # TODO: this needs A LOT of tests
+        self._resolved = clone(
+            self._resolved,
+            deferred_copy_internals=deferred_copy_internals,
+            **kw,
+        )
+
+    @util.memoized_property
+    def _resolved(self):
+        expr = self._rec.expected_expr
+
+        if self._resolved_bindparams:
+            expr = self._setup_binds_for_tracked_expr(expr)
+
+        return expr
+
+    def _gen_cache_key(self, anon_map, bindparams):
+        if self.closure_cache_key is _cache_key.NO_CACHE:
+            anon_map[_cache_key.NO_CACHE] = True
+            return None
+
+        cache_key = (
+            self.fn.__code__,
+            self.__class__,
+        ) + self.closure_cache_key
+
+        parent = self.parent_lambda
+
+        while parent is not None:
+            assert parent.closure_cache_key is not CacheConst.NO_CACHE
+            parent_closure_cache_key: Tuple[Any, ...] = (
+                parent.closure_cache_key
+            )
+
+            cache_key = (
+                (parent.fn.__code__,) + parent_closure_cache_key + cache_key
+            )
+
+            parent = parent.parent_lambda
+
+        if self._resolved_bindparams:
+            bindparams.extend(self._resolved_bindparams)
+        return cache_key
+
+    def _invoke_user_fn(self, fn: _AnyLambdaType, *arg: Any) -> ClauseElement:
+        return fn()  # type: ignore[no-any-return]
+
+
+class DeferredLambdaElement(LambdaElement):
+    """A LambdaElement where the lambda accepts arguments and is
+    invoked within the compile phase with special context.
+
+    This lambda doesn't normally produce its real SQL expression outside of the
+    compile phase.  It is passed a fixed set of initial arguments
+    so that it can generate a sample expression.
+
+    """
+
+    def __init__(
+        self,
+        fn: _AnyLambdaType,
+        role: Type[roles.SQLRole],
+        opts: Union[Type[LambdaOptions], LambdaOptions] = LambdaOptions,
+        lambda_args: Tuple[Any, ...] = (),
+    ):
+        self.lambda_args = lambda_args
+        super().__init__(fn, role, opts)
+
+    def _invoke_user_fn(self, fn, *arg):
+        return fn(*self.lambda_args)
+
+    def _resolve_with_args(self, *lambda_args: Any) -> ClauseElement:
+        assert isinstance(self._rec, AnalyzedFunction)
+        tracker_fn = self._rec.tracker_instrumented_fn
+        expr = tracker_fn(*lambda_args)
+
+        expr = coercions.expect(self.role, expr)
+
+        expr = self._setup_binds_for_tracked_expr(expr)
+
+        # this validation is getting very close, but not quite, to achieving
+        # #5767.  The problem is if the base lambda uses an unnamed column
+        # as is very common with mixins, the parameter name is different
+        # and it produces a false positive; that is, for the documented case
+        # that is exactly what people will be doing, it doesn't work, so
+        # I'm not really sure how to handle this right now.
+        # expected_binds = [
+        #    b._orig_key
+        #    for b in self._rec.expr._generate_cache_key()[1]
+        #    if b.required
+        # ]
+        # got_binds = [
+        #    b._orig_key for b in expr._generate_cache_key()[1] if b.required
+        # ]
+        # if expected_binds != got_binds:
+        #    raise exc.InvalidRequestError(
+        #        "Lambda callable at %s produced a different set of bound "
+        #        "parameters than its original run: %s"
+        #        % (self.fn.__code__, ", ".join(got_binds))
+        #    )
+
+        # TODO: TEST TEST TEST, this is very out there
+        for deferred_copy_internals in self._transforms:
+            expr = deferred_copy_internals(expr)
+
+        return expr  # type: ignore
+
+    def _copy_internals(
+        self, clone=_clone, deferred_copy_internals=None, **kw
+    ):
+        super()._copy_internals(
+            clone=clone,
+            deferred_copy_internals=deferred_copy_internals,  # **kw
+            opts=kw,
+        )
+
+        # TODO: A LOT A LOT of tests.   for _resolve_with_args, we don't know
+        # our expression yet.   so hold onto the replacement
+        if deferred_copy_internals:
+            self._transforms += (deferred_copy_internals,)
+
+
+class StatementLambdaElement(
+    roles.AllowsLambdaRole, LambdaElement, Executable
+):
+    """Represent a composable SQL statement as a :class:`_sql.LambdaElement`.
+
+    The :class:`_sql.StatementLambdaElement` is constructed using the
+    :func:`_sql.lambda_stmt` function::
+
+
+        from sqlalchemy import lambda_stmt
+
+        stmt = lambda_stmt(lambda: select(table))
+
+    Once constructed, additional criteria can be built onto the statement
+    by adding subsequent lambdas, which accept the existing statement
+    object as a single parameter::
+
+        stmt += lambda s: s.where(table.c.col == parameter)
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`engine_lambda_caching`
+
+    """
+
+    if TYPE_CHECKING:
+
+        def __init__(
+            self,
+            fn: _StmtLambdaType,
+            role: Type[SQLRole],
+            opts: Union[Type[LambdaOptions], LambdaOptions] = LambdaOptions,
+            apply_propagate_attrs: Optional[ClauseElement] = None,
+        ): ...
+
+    def __add__(
+        self, other: _StmtLambdaElementType[Any]
+    ) -> StatementLambdaElement:
+        return self.add_criteria(other)
+
+    def add_criteria(
+        self,
+        other: _StmtLambdaElementType[Any],
+        enable_tracking: bool = True,
+        track_on: Optional[Any] = None,
+        track_closure_variables: bool = True,
+        track_bound_values: bool = True,
+    ) -> StatementLambdaElement:
+        """Add new criteria to this :class:`_sql.StatementLambdaElement`.
+
+        E.g.::
+
+            >>> def my_stmt(parameter):
+            ...     stmt = lambda_stmt(
+            ...         lambda: select(table.c.x, table.c.y),
+            ...     )
+            ...     stmt = stmt.add_criteria(lambda: table.c.x > parameter)
+            ...     return stmt
+
+        The :meth:`_sql.StatementLambdaElement.add_criteria` method is
+        equivalent to using the Python addition operator to add a new
+        lambda, except that additional arguments may be added including
+        ``track_closure_values`` and ``track_on``::
+
+            >>> def my_stmt(self, foo):
+            ...     stmt = lambda_stmt(
+            ...         lambda: select(func.max(foo.x, foo.y)),
+            ...         track_closure_variables=False,
+            ...     )
+            ...     stmt = stmt.add_criteria(lambda: self.where_criteria, track_on=[self])
+            ...     return stmt
+
+        See :func:`_sql.lambda_stmt` for a description of the parameters
+        accepted.
+
+        """  # noqa: E501
+
+        opts = self.opts + dict(
+            enable_tracking=enable_tracking,
+            track_closure_variables=track_closure_variables,
+            global_track_bound_values=self.opts.global_track_bound_values,
+            track_on=track_on,
+            track_bound_values=track_bound_values,
+        )
+
+        return LinkedLambdaElement(other, parent_lambda=self, opts=opts)
+
+    def _execute_on_connection(
+        self, connection, distilled_params, execution_options
+    ):
+        if TYPE_CHECKING:
+            assert isinstance(self._rec.expected_expr, ClauseElement)
+        if self._rec.expected_expr.supports_execution:
+            return connection._execute_clauseelement(
+                self, distilled_params, execution_options
+            )
+        else:
+            raise exc.ObjectNotExecutableError(self)
+
+    @property
+    def _proxied(self) -> Any:
+        return self._rec_expected_expr
+
+    @property
+    def _with_options(self):
+        return self._proxied._with_options
+
+    @property
+    def _effective_plugin_target(self):
+        return self._proxied._effective_plugin_target
+
+    @property
+    def _execution_options(self):
+        return self._proxied._execution_options
+
+    @property
+    def _all_selected_columns(self):
+        return self._proxied._all_selected_columns
+
+    @property
+    def is_select(self):
+        return self._proxied.is_select
+
+    @property
+    def is_update(self):
+        return self._proxied.is_update
+
+    @property
+    def is_insert(self):
+        return self._proxied.is_insert
+
+    @property
+    def is_text(self):
+        return self._proxied.is_text
+
+    @property
+    def is_delete(self):
+        return self._proxied.is_delete
+
+    @property
+    def is_dml(self):
+        return self._proxied.is_dml
+
+    def spoil(self) -> NullLambdaStatement:
+        """Return a new :class:`.StatementLambdaElement` that will run
+        all lambdas unconditionally each time.
+
+        """
+        return NullLambdaStatement(self.fn())
+
+
+class NullLambdaStatement(roles.AllowsLambdaRole, elements.ClauseElement):
+    """Provides the :class:`.StatementLambdaElement` API but does not
+    cache or analyze lambdas.
+
+    the lambdas are instead invoked immediately.
+
+    The intended use is to isolate issues that may arise when using
+    lambda statements.
+
+    """
+
+    __visit_name__ = "lambda_element"
+
+    _is_lambda_element = True
+
+    _traverse_internals = [
+        ("_resolved", visitors.InternalTraversal.dp_clauseelement)
+    ]
+
+    def __init__(self, statement):
+        self._resolved = statement
+        self._propagate_attrs = statement._propagate_attrs
+
+    def __getattr__(self, key):
+        return getattr(self._resolved, key)
+
+    def __add__(self, other):
+        statement = other(self._resolved)
+
+        return NullLambdaStatement(statement)
+
+    def add_criteria(self, other, **kw):
+        statement = other(self._resolved)
+
+        return NullLambdaStatement(statement)
+
+    def _execute_on_connection(
+        self, connection, distilled_params, execution_options
+    ):
+        if self._resolved.supports_execution:
+            return connection._execute_clauseelement(
+                self, distilled_params, execution_options
+            )
+        else:
+            raise exc.ObjectNotExecutableError(self)
+
+
+class LinkedLambdaElement(StatementLambdaElement):
+    """Represent subsequent links of a :class:`.StatementLambdaElement`."""
+
+    parent_lambda: StatementLambdaElement
+
+    def __init__(
+        self,
+        fn: _StmtLambdaElementType[Any],
+        parent_lambda: StatementLambdaElement,
+        opts: Union[Type[LambdaOptions], LambdaOptions],
+    ):
+        self.opts = opts
+        self.fn = fn
+        self.parent_lambda = parent_lambda
+
+        self.tracker_key = parent_lambda.tracker_key + (fn.__code__,)
+        self._retrieve_tracker_rec(fn, self, opts)
+        self._propagate_attrs = parent_lambda._propagate_attrs
+
+    def _invoke_user_fn(self, fn, *arg):
+        return fn(self.parent_lambda._resolved)
+
+
+class AnalyzedCode:
+    __slots__ = (
+        "track_closure_variables",
+        "track_bound_values",
+        "bindparam_trackers",
+        "closure_trackers",
+        "build_py_wrappers",
+    )
+    _fns: weakref.WeakKeyDictionary[CodeType, AnalyzedCode] = (
+        weakref.WeakKeyDictionary()
+    )
+
+    _generation_mutex = threading.RLock()
+
+    @classmethod
+    def get(cls, fn, lambda_element, lambda_kw, **kw):
+        try:
+            # TODO: validate kw haven't changed?
+            return cls._fns[fn.__code__]
+        except KeyError:
+            pass
+
+        with cls._generation_mutex:
+            # check for other thread already created object
+            if fn.__code__ in cls._fns:
+                return cls._fns[fn.__code__]
+
+            analyzed: AnalyzedCode
+            cls._fns[fn.__code__] = analyzed = AnalyzedCode(
+                fn, lambda_element, lambda_kw, **kw
+            )
+            return analyzed
+
+    def __init__(self, fn, lambda_element, opts):
+        if inspect.ismethod(fn):
+            raise exc.ArgumentError(
+                "Method %s may not be passed as a SQL expression" % fn
+            )
+        closure = fn.__closure__
+
+        self.track_bound_values = (
+            opts.track_bound_values and opts.global_track_bound_values
+        )
+        enable_tracking = opts.enable_tracking
+        track_on = opts.track_on
+        track_closure_variables = opts.track_closure_variables
+
+        self.track_closure_variables = track_closure_variables and not track_on
+
+        # a list of callables generated from _bound_parameter_getter_*
+        # functions.  Each of these uses a PyWrapper object to retrieve
+        # a parameter value
+        self.bindparam_trackers = []
+
+        # a list of callables generated from _cache_key_getter_* functions
+        # these callables work to generate a cache key for the lambda
+        # based on what's inside its closure variables.
+        self.closure_trackers = []
+
+        self.build_py_wrappers = []
+
+        if enable_tracking:
+            if track_on:
+                self._init_track_on(track_on)
+
+            self._init_globals(fn)
+
+            if closure:
+                self._init_closure(fn)
+
+        self._setup_additional_closure_trackers(fn, lambda_element, opts)
+
+    def _init_track_on(self, track_on):
+        self.closure_trackers.extend(
+            self._cache_key_getter_track_on(idx, elem)
+            for idx, elem in enumerate(track_on)
+        )
+
+    def _init_globals(self, fn):
+        build_py_wrappers = self.build_py_wrappers
+        bindparam_trackers = self.bindparam_trackers
+        track_bound_values = self.track_bound_values
+
+        for name in fn.__code__.co_names:
+            if name not in fn.__globals__:
+                continue
+
+            _bound_value = self._roll_down_to_literal(fn.__globals__[name])
+
+            if coercions._deep_is_literal(_bound_value):
+                build_py_wrappers.append((name, None))
+                if track_bound_values:
+                    bindparam_trackers.append(
+                        self._bound_parameter_getter_func_globals(name)
+                    )
+
+    def _init_closure(self, fn):
+        build_py_wrappers = self.build_py_wrappers
+        closure = fn.__closure__
+
+        track_bound_values = self.track_bound_values
+        track_closure_variables = self.track_closure_variables
+        bindparam_trackers = self.bindparam_trackers
+        closure_trackers = self.closure_trackers
+
+        for closure_index, (fv, cell) in enumerate(
+            zip(fn.__code__.co_freevars, closure)
+        ):
+            _bound_value = self._roll_down_to_literal(cell.cell_contents)
+
+            if coercions._deep_is_literal(_bound_value):
+                build_py_wrappers.append((fv, closure_index))
+                if track_bound_values:
+                    bindparam_trackers.append(
+                        self._bound_parameter_getter_func_closure(
+                            fv, closure_index
+                        )
+                    )
+            else:
+                # for normal cell contents, add them to a list that
+                # we can compare later when we get new lambdas.  if
+                # any identities have changed, then we will
+                # recalculate the whole lambda and run it again.
+
+                if track_closure_variables:
+                    closure_trackers.append(
+                        self._cache_key_getter_closure_variable(
+                            fn, fv, closure_index, cell.cell_contents
+                        )
+                    )
+
+    def _setup_additional_closure_trackers(self, fn, lambda_element, opts):
+        # an additional step is to actually run the function, then
+        # go through the PyWrapper objects that were set up to catch a bound
+        # parameter.   then if they *didn't* make a param, oh they're another
+        # object in the closure we have to track for our cache key.  so
+        # create trackers to catch those.
+
+        analyzed_function = AnalyzedFunction(
+            self,
+            lambda_element,
+            None,
+            fn,
+        )
+
+        closure_trackers = self.closure_trackers
+
+        for pywrapper in analyzed_function.closure_pywrappers:
+            if not pywrapper._sa__has_param:
+                closure_trackers.append(
+                    self._cache_key_getter_tracked_literal(fn, pywrapper)
+                )
+
+    @classmethod
+    def _roll_down_to_literal(cls, element):
+        is_clause_element = hasattr(element, "__clause_element__")
+
+        if is_clause_element:
+            while not isinstance(
+                element, (elements.ClauseElement, schema.SchemaItem, type)
+            ):
+                try:
+                    element = element.__clause_element__()
+                except AttributeError:
+                    break
+
+        if not is_clause_element:
+            insp = inspection.inspect(element, raiseerr=False)
+            if insp is not None:
+                try:
+                    return insp.__clause_element__()
+                except AttributeError:
+                    return insp
+
+            # TODO: should we coerce consts None/True/False here?
+            return element
+        else:
+            return element
+
+    def _bound_parameter_getter_func_globals(self, name):
+        """Return a getter that will extend a list of bound parameters
+        with new entries from the ``__globals__`` collection of a particular
+        lambda.
+
+        """
+
+        def extract_parameter_value(
+            current_fn, tracker_instrumented_fn, result
+        ):
+            wrapper = tracker_instrumented_fn.__globals__[name]
+            object.__getattribute__(wrapper, "_extract_bound_parameters")(
+                current_fn.__globals__[name], result
+            )
+
+        return extract_parameter_value
+
+    def _bound_parameter_getter_func_closure(self, name, closure_index):
+        """Return a getter that will extend a list of bound parameters
+        with new entries from the ``__closure__`` collection of a particular
+        lambda.
+
+        """
+
+        def extract_parameter_value(
+            current_fn, tracker_instrumented_fn, result
+        ):
+            wrapper = tracker_instrumented_fn.__closure__[
+                closure_index
+            ].cell_contents
+            object.__getattribute__(wrapper, "_extract_bound_parameters")(
+                current_fn.__closure__[closure_index].cell_contents, result
+            )
+
+        return extract_parameter_value
+
+    def _cache_key_getter_track_on(self, idx, elem):
+        """Return a getter that will extend a cache key with new entries
+        from the "track_on" parameter passed to a :class:`.LambdaElement`.
+
+        """
+
+        if isinstance(elem, tuple):
+            # tuple must contain hascachekey elements
+            def get(closure, opts, anon_map, bindparams):
+                return tuple(
+                    tup_elem._gen_cache_key(anon_map, bindparams)
+                    for tup_elem in opts.track_on[idx]
+                )
+
+        elif isinstance(elem, _cache_key.HasCacheKey):
+
+            def get(closure, opts, anon_map, bindparams):
+                return opts.track_on[idx]._gen_cache_key(anon_map, bindparams)
+
+        else:
+
+            def get(closure, opts, anon_map, bindparams):
+                return opts.track_on[idx]
+
+        return get
+
+    def _cache_key_getter_closure_variable(
+        self,
+        fn,
+        variable_name,
+        idx,
+        cell_contents,
+        use_clause_element=False,
+        use_inspect=False,
+    ):
+        """Return a getter that will extend a cache key with new entries
+        from the ``__closure__`` collection of a particular lambda.
+
+        """
+
+        if isinstance(cell_contents, _cache_key.HasCacheKey):
+
+            def get(closure, opts, anon_map, bindparams):
+                obj = closure[idx].cell_contents
+                if use_inspect:
+                    obj = inspection.inspect(obj)
+                elif use_clause_element:
+                    while hasattr(obj, "__clause_element__"):
+                        if not getattr(obj, "is_clause_element", False):
+                            obj = obj.__clause_element__()
+
+                return obj._gen_cache_key(anon_map, bindparams)
+
+        elif isinstance(cell_contents, types.FunctionType):
+
+            def get(closure, opts, anon_map, bindparams):
+                return closure[idx].cell_contents.__code__
+
+        elif isinstance(cell_contents, collections_abc.Sequence):
+
+            def get(closure, opts, anon_map, bindparams):
+                contents = closure[idx].cell_contents
+
+                try:
+                    return tuple(
+                        elem._gen_cache_key(anon_map, bindparams)
+                        for elem in contents
+                    )
+                except AttributeError as ae:
+                    self._raise_for_uncacheable_closure_variable(
+                        variable_name, fn, from_=ae
+                    )
+
+        else:
+            # if the object is a mapped class or aliased class, or some
+            # other object in the ORM realm of things like that, imitate
+            # the logic used in coercions.expect() to roll it down to the
+            # SQL element
+            element = cell_contents
+            is_clause_element = False
+            while hasattr(element, "__clause_element__"):
+                is_clause_element = True
+                if not getattr(element, "is_clause_element", False):
+                    element = element.__clause_element__()
+                else:
+                    break
+
+            if not is_clause_element:
+                insp = inspection.inspect(element, raiseerr=False)
+                if insp is not None:
+                    return self._cache_key_getter_closure_variable(
+                        fn, variable_name, idx, insp, use_inspect=True
+                    )
+            else:
+                return self._cache_key_getter_closure_variable(
+                    fn, variable_name, idx, element, use_clause_element=True
+                )
+
+            self._raise_for_uncacheable_closure_variable(variable_name, fn)
+
+        return get
+
+    def _raise_for_uncacheable_closure_variable(
+        self, variable_name, fn, from_=None
+    ):
+        raise exc.InvalidRequestError(
+            "Closure variable named '%s' inside of lambda callable %s "
+            "does not refer to a cacheable SQL element, and also does not "
+            "appear to be serving as a SQL literal bound value based on "
+            "the default "
+            "SQL expression returned by the function.   This variable "
+            "needs to remain outside the scope of a SQL-generating lambda "
+            "so that a proper cache key may be generated from the "
+            "lambda's state.  Evaluate this variable outside of the "
+            "lambda, set track_on=[<elements>] to explicitly select "
+            "closure elements to track, or set "
+            "track_closure_variables=False to exclude "
+            "closure variables from being part of the cache key."
+            % (variable_name, fn.__code__),
+        ) from from_
+
+    def _cache_key_getter_tracked_literal(self, fn, pytracker):
+        """Return a getter that will extend a cache key with new entries
+        from the ``__closure__`` collection of a particular lambda.
+
+        this getter differs from _cache_key_getter_closure_variable
+        in that these are detected after the function is run, and PyWrapper
+        objects have recorded that a particular literal value is in fact
+        not being interpreted as a bound parameter.
+
+        """
+
+        elem = pytracker._sa__to_evaluate
+        closure_index = pytracker._sa__closure_index
+        variable_name = pytracker._sa__name
+
+        return self._cache_key_getter_closure_variable(
+            fn, variable_name, closure_index, elem
+        )
+
+
+class NonAnalyzedFunction:
+    __slots__ = ("expr",)
+
+    closure_bindparams: Optional[List[BindParameter[Any]]] = None
+    bindparam_trackers: Optional[List[_BoundParameterGetter]] = None
+
+    is_sequence = False
+
+    expr: ClauseElement
+
+    def __init__(self, expr: ClauseElement):
+        self.expr = expr
+
+    @property
+    def expected_expr(self) -> ClauseElement:
+        return self.expr
+
+
+class AnalyzedFunction:
+    __slots__ = (
+        "analyzed_code",
+        "fn",
+        "closure_pywrappers",
+        "tracker_instrumented_fn",
+        "expr",
+        "bindparam_trackers",
+        "expected_expr",
+        "is_sequence",
+        "propagate_attrs",
+        "closure_bindparams",
+    )
+
+    closure_bindparams: Optional[List[BindParameter[Any]]]
+    expected_expr: Union[ClauseElement, List[ClauseElement]]
+    bindparam_trackers: Optional[List[_BoundParameterGetter]]
+
+    def __init__(
+        self,
+        analyzed_code,
+        lambda_element,
+        apply_propagate_attrs,
+        fn,
+    ):
+        self.analyzed_code = analyzed_code
+        self.fn = fn
+
+        self.bindparam_trackers = analyzed_code.bindparam_trackers
+
+        self._instrument_and_run_function(lambda_element)
+
+        self._coerce_expression(lambda_element, apply_propagate_attrs)
+
+    def _instrument_and_run_function(self, lambda_element):
+        analyzed_code = self.analyzed_code
+
+        fn = self.fn
+        self.closure_pywrappers = closure_pywrappers = []
+
+        build_py_wrappers = analyzed_code.build_py_wrappers
+
+        if not build_py_wrappers:
+            self.tracker_instrumented_fn = tracker_instrumented_fn = fn
+            self.expr = lambda_element._invoke_user_fn(tracker_instrumented_fn)
+        else:
+            track_closure_variables = analyzed_code.track_closure_variables
+            closure = fn.__closure__
+
+            # will form the __closure__ of the function when we rebuild it
+            if closure:
+                new_closure = {
+                    fv: cell.cell_contents
+                    for fv, cell in zip(fn.__code__.co_freevars, closure)
+                }
+            else:
+                new_closure = {}
+
+            # will form the __globals__ of the function when we rebuild it
+            new_globals = fn.__globals__.copy()
+
+            for name, closure_index in build_py_wrappers:
+                if closure_index is not None:
+                    value = closure[closure_index].cell_contents
+                    new_closure[name] = bind = PyWrapper(
+                        fn,
+                        name,
+                        value,
+                        closure_index=closure_index,
+                        track_bound_values=(
+                            self.analyzed_code.track_bound_values
+                        ),
+                    )
+                    if track_closure_variables:
+                        closure_pywrappers.append(bind)
+                else:
+                    value = fn.__globals__[name]
+                    new_globals[name] = bind = PyWrapper(fn, name, value)
+
+            # rewrite the original fn.   things that look like they will
+            # become bound parameters are wrapped in a PyWrapper.
+            self.tracker_instrumented_fn = tracker_instrumented_fn = (
+                self._rewrite_code_obj(
+                    fn,
+                    [new_closure[name] for name in fn.__code__.co_freevars],
+                    new_globals,
+                )
+            )
+
+            # now invoke the function.  This will give us a new SQL
+            # expression, but all the places that there would be a bound
+            # parameter, the PyWrapper in its place will give us a bind
+            # with a predictable name we can match up later.
+
+            # additionally, each PyWrapper will log that it did in fact
+            # create a parameter, otherwise, it's some kind of Python
+            # object in the closure and we want to track that, to make
+            # sure it doesn't change to something else, or if it does,
+            # that we create a different tracked function with that
+            # variable.
+            self.expr = lambda_element._invoke_user_fn(tracker_instrumented_fn)
+
+    def _coerce_expression(self, lambda_element, apply_propagate_attrs):
+        """Run the tracker-generated expression through coercion rules.
+
+        After the user-defined lambda has been invoked to produce a statement
+        for re-use, run it through coercion rules to both check that it's the
+        correct type of object and also to coerce it to its useful form.
+
+        """
+
+        parent_lambda = lambda_element.parent_lambda
+        expr = self.expr
+
+        if parent_lambda is None:
+            if isinstance(expr, collections_abc.Sequence):
+                self.expected_expr = [
+                    cast(
+                        "ClauseElement",
+                        coercions.expect(
+                            lambda_element.role,
+                            sub_expr,
+                            apply_propagate_attrs=apply_propagate_attrs,
+                        ),
+                    )
+                    for sub_expr in expr
+                ]
+                self.is_sequence = True
+            else:
+                self.expected_expr = cast(
+                    "ClauseElement",
+                    coercions.expect(
+                        lambda_element.role,
+                        expr,
+                        apply_propagate_attrs=apply_propagate_attrs,
+                    ),
+                )
+                self.is_sequence = False
+        else:
+            self.expected_expr = expr
+            self.is_sequence = False
+
+        if apply_propagate_attrs is not None:
+            self.propagate_attrs = apply_propagate_attrs._propagate_attrs
+        else:
+            self.propagate_attrs = util.EMPTY_DICT
+
+    def _rewrite_code_obj(self, f, cell_values, globals_):
+        """Return a copy of f, with a new closure and new globals
+
+        yes it works in pypy :P
+
+        """
+
+        argrange = range(len(cell_values))
+
+        code = "def make_cells():\n"
+        if cell_values:
+            code += "    (%s) = (%s)\n" % (
+                ", ".join("i%d" % i for i in argrange),
+                ", ".join("o%d" % i for i in argrange),
+            )
+        code += "    def closure():\n"
+        code += "        return %s\n" % ", ".join("i%d" % i for i in argrange)
+        code += "    return closure.__closure__"
+        vars_ = {"o%d" % i: cell_values[i] for i in argrange}
+        exec(code, vars_, vars_)
+        closure = vars_["make_cells"]()
+
+        func = type(f)(
+            f.__code__, globals_, f.__name__, f.__defaults__, closure
+        )
+        func.__annotations__ = f.__annotations__
+        func.__kwdefaults__ = f.__kwdefaults__
+        func.__doc__ = f.__doc__
+        func.__module__ = f.__module__
+
+        return func
+
+
+class PyWrapper(ColumnOperators):
+    """A wrapper object that is injected into the ``__globals__`` and
+    ``__closure__`` of a Python function.
+
+    When the function is instrumented with :class:`.PyWrapper` objects, it is
+    then invoked just once in order to set up the wrappers.  We look through
+    all the :class:`.PyWrapper` objects we made to find the ones that generated
+    a :class:`.BindParameter` object, e.g. the expression system interpreted
+    something as a literal.   Those positions in the globals/closure are then
+    ones that we will look at, each time a new lambda comes in that refers to
+    the same ``__code__`` object.   In this way, we keep a single version of
+    the SQL expression that this lambda produced, without calling upon the
+    Python function that created it more than once, unless its other closure
+    variables have changed.   The expression is then transformed to have the
+    new bound values embedded into it.
+
+    """
+
+    def __init__(
+        self,
+        fn,
+        name,
+        to_evaluate,
+        closure_index=None,
+        getter=None,
+        track_bound_values=True,
+    ):
+        self.fn = fn
+        self._name = name
+        self._to_evaluate = to_evaluate
+        self._param = None
+        self._has_param = False
+        self._bind_paths = {}
+        self._getter = getter
+        self._closure_index = closure_index
+        self.track_bound_values = track_bound_values
+
+    def __call__(self, *arg, **kw):
+        elem = object.__getattribute__(self, "_to_evaluate")
+        value = elem(*arg, **kw)
+        if (
+            self._sa_track_bound_values
+            and coercions._deep_is_literal(value)
+            and not isinstance(
+                # TODO: coverage where an ORM option or similar is here
+                value,
+                _cache_key.HasCacheKey,
+            )
+        ):
+            name = object.__getattribute__(self, "_name")
+            raise exc.InvalidRequestError(
+                "Can't invoke Python callable %s() inside of lambda "
+                "expression argument at %s; lambda SQL constructs should "
+                "not invoke functions from closure variables to produce "
+                "literal values since the "
+                "lambda SQL system normally extracts bound values without "
+                "actually "
+                "invoking the lambda or any functions within it.  Call the "
+                "function outside of the "
+                "lambda and assign to a local variable that is used in the "
+                "lambda as a closure variable, or set "
+                "track_bound_values=False if the return value of this "
+                "function is used in some other way other than a SQL bound "
+                "value." % (name, self._sa_fn.__code__)
+            )
+        else:
+            return value
+
+    def operate(self, op, *other, **kwargs):
+        elem = object.__getattribute__(self, "_py_wrapper_literal")()
+        return op(elem, *other, **kwargs)
+
+    def reverse_operate(self, op, other, **kwargs):
+        elem = object.__getattribute__(self, "_py_wrapper_literal")()
+        return op(other, elem, **kwargs)
+
+    def _extract_bound_parameters(self, starting_point, result_list):
+        param = object.__getattribute__(self, "_param")
+        if param is not None:
+            param = param._with_value(starting_point, maintain_key=True)
+            result_list.append(param)
+        for pywrapper in object.__getattribute__(self, "_bind_paths").values():
+            getter = object.__getattribute__(pywrapper, "_getter")
+            element = getter(starting_point)
+            pywrapper._sa__extract_bound_parameters(element, result_list)
+
+    def _py_wrapper_literal(self, expr=None, operator=None, **kw):
+        param = object.__getattribute__(self, "_param")
+        to_evaluate = object.__getattribute__(self, "_to_evaluate")
+        if param is None:
+            name = object.__getattribute__(self, "_name")
+            self._param = param = elements.BindParameter(
+                name,
+                required=False,
+                unique=True,
+                _compared_to_operator=operator,
+                _compared_to_type=expr.type if expr is not None else None,
+            )
+            self._has_param = True
+        return param._with_value(to_evaluate, maintain_key=True)
+
+    def __bool__(self):
+        to_evaluate = object.__getattribute__(self, "_to_evaluate")
+        return bool(to_evaluate)
+
+    def __getattribute__(self, key):
+        if key.startswith("_sa_"):
+            return object.__getattribute__(self, key[4:])
+        elif key in (
+            "__clause_element__",
+            "operate",
+            "reverse_operate",
+            "_py_wrapper_literal",
+            "__class__",
+            "__dict__",
+        ):
+            return object.__getattribute__(self, key)
+
+        if key.startswith("__"):
+            elem = object.__getattribute__(self, "_to_evaluate")
+            return getattr(elem, key)
+        else:
+            return self._sa__add_getter(key, operator.attrgetter)
+
+    def __iter__(self):
+        elem = object.__getattribute__(self, "_to_evaluate")
+        return iter(elem)
+
+    def __getitem__(self, key):
+        elem = object.__getattribute__(self, "_to_evaluate")
+        if not hasattr(elem, "__getitem__"):
+            raise AttributeError("__getitem__")
+
+        if isinstance(key, PyWrapper):
+            # TODO: coverage
+            raise exc.InvalidRequestError(
+                "Dictionary keys / list indexes inside of a cached "
+                "lambda must be Python literals only"
+            )
+        return self._sa__add_getter(key, operator.itemgetter)
+
+    def _add_getter(self, key, getter_fn):
+        bind_paths = object.__getattribute__(self, "_bind_paths")
+
+        bind_path_key = (key, getter_fn)
+        if bind_path_key in bind_paths:
+            return bind_paths[bind_path_key]
+
+        getter = getter_fn(key)
+        elem = object.__getattribute__(self, "_to_evaluate")
+        value = getter(elem)
+
+        rolled_down_value = AnalyzedCode._roll_down_to_literal(value)
+
+        if coercions._deep_is_literal(rolled_down_value):
+            wrapper = PyWrapper(self._sa_fn, key, value, getter=getter)
+            bind_paths[bind_path_key] = wrapper
+            return wrapper
+        else:
+            return value
+
+
+@inspection._inspects(LambdaElement)
+def insp(lmb):
+    return inspection.inspect(lmb._resolved)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/naming.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/naming.py
new file mode 100644
index 00000000..58203e4b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/naming.py
@@ -0,0 +1,212 @@
+# sql/naming.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Establish constraint and index naming conventions.
+
+
+"""
+
+from __future__ import annotations
+
+import re
+
+from . import events  # noqa
+from .base import _NONE_NAME
+from .elements import conv as conv
+from .schema import CheckConstraint
+from .schema import Column
+from .schema import Constraint
+from .schema import ForeignKeyConstraint
+from .schema import Index
+from .schema import PrimaryKeyConstraint
+from .schema import Table
+from .schema import UniqueConstraint
+from .. import event
+from .. import exc
+
+
+class ConventionDict:
+    def __init__(self, const, table, convention):
+        self.const = const
+        self._is_fk = isinstance(const, ForeignKeyConstraint)
+        self.table = table
+        self.convention = convention
+        self._const_name = const.name
+
+    def _key_table_name(self):
+        return self.table.name
+
+    def _column_X(self, idx, attrname):
+        if self._is_fk:
+            try:
+                fk = self.const.elements[idx]
+            except IndexError:
+                return ""
+            else:
+                return getattr(fk.parent, attrname)
+        else:
+            cols = list(self.const.columns)
+            try:
+                col = cols[idx]
+            except IndexError:
+                return ""
+            else:
+                return getattr(col, attrname)
+
+    def _key_constraint_name(self):
+        if self._const_name in (None, _NONE_NAME):
+            raise exc.InvalidRequestError(
+                "Naming convention including "
+                "%(constraint_name)s token requires that "
+                "constraint is explicitly named."
+            )
+        if not isinstance(self._const_name, conv):
+            self.const.name = None
+        return self._const_name
+
+    def _key_column_X_key(self, idx):
+        # note this method was missing before
+        # [ticket:3989], meaning tokens like ``%(column_0_key)s`` weren't
+        # working even though documented.
+        return self._column_X(idx, "key")
+
+    def _key_column_X_name(self, idx):
+        return self._column_X(idx, "name")
+
+    def _key_column_X_label(self, idx):
+        return self._column_X(idx, "_ddl_label")
+
+    def _key_referred_table_name(self):
+        fk = self.const.elements[0]
+        refs = fk.target_fullname.split(".")
+        if len(refs) == 3:
+            refschema, reftable, refcol = refs
+        else:
+            reftable, refcol = refs
+        return reftable
+
+    def _key_referred_column_X_name(self, idx):
+        fk = self.const.elements[idx]
+        # note that before [ticket:3989], this method was returning
+        # the specification for the :class:`.ForeignKey` itself, which normally
+        # would be using the ``.key`` of the column, not the name.
+        return fk.column.name
+
+    def __getitem__(self, key):
+        if key in self.convention:
+            return self.convention[key](self.const, self.table)
+        elif hasattr(self, "_key_%s" % key):
+            return getattr(self, "_key_%s" % key)()
+        else:
+            col_template = re.match(r".*_?column_(\d+)(_?N)?_.+", key)
+            if col_template:
+                idx = col_template.group(1)
+                multiples = col_template.group(2)
+
+                if multiples:
+                    if self._is_fk:
+                        elems = self.const.elements
+                    else:
+                        elems = list(self.const.columns)
+                    tokens = []
+                    for idx, elem in enumerate(elems):
+                        attr = "_key_" + key.replace("0" + multiples, "X")
+                        try:
+                            tokens.append(getattr(self, attr)(idx))
+                        except AttributeError:
+                            raise KeyError(key)
+                    sep = "_" if multiples.startswith("_") else ""
+                    return sep.join(tokens)
+                else:
+                    attr = "_key_" + key.replace(idx, "X")
+                    idx = int(idx)
+                    if hasattr(self, attr):
+                        return getattr(self, attr)(idx)
+        raise KeyError(key)
+
+
+_prefix_dict = {
+    Index: "ix",
+    PrimaryKeyConstraint: "pk",
+    CheckConstraint: "ck",
+    UniqueConstraint: "uq",
+    ForeignKeyConstraint: "fk",
+}
+
+
+def _get_convention(dict_, key):
+    for super_ in key.__mro__:
+        if super_ in _prefix_dict and _prefix_dict[super_] in dict_:
+            return dict_[_prefix_dict[super_]]
+        elif super_ in dict_:
+            return dict_[super_]
+    else:
+        return None
+
+
+def _constraint_name_for_table(const, table):
+    metadata = table.metadata
+    convention = _get_convention(metadata.naming_convention, type(const))
+
+    if isinstance(const.name, conv):
+        return const.name
+    elif (
+        convention is not None
+        and not isinstance(const.name, conv)
+        and (
+            const.name is None
+            or "constraint_name" in convention
+            or const.name is _NONE_NAME
+        )
+    ):
+        return conv(
+            convention
+            % ConventionDict(const, table, metadata.naming_convention)
+        )
+    elif convention is _NONE_NAME:
+        return None
+
+
+@event.listens_for(
+    PrimaryKeyConstraint, "_sa_event_column_added_to_pk_constraint"
+)
+def _column_added_to_pk_constraint(pk_constraint, col):
+    if pk_constraint._implicit_generated:
+        # only operate upon the "implicit" pk constraint for now,
+        # as we have to force the name to None to reset it.  the
+        # "implicit" constraint will only have a naming convention name
+        # if at all.
+        table = pk_constraint.table
+        pk_constraint.name = None
+        newname = _constraint_name_for_table(pk_constraint, table)
+        if newname:
+            pk_constraint.name = newname
+
+
+@event.listens_for(Constraint, "after_parent_attach")
+@event.listens_for(Index, "after_parent_attach")
+def _constraint_name(const, table):
+    if isinstance(table, Column):
+        # this path occurs for a CheckConstraint linked to a Column
+
+        # for column-attached constraint, set another event
+        # to link the column attached to the table as this constraint
+        # associated with the table.
+        event.listen(
+            table,
+            "after_parent_attach",
+            lambda col, table: _constraint_name(const, table),
+        )
+
+    elif isinstance(table, Table):
+        if isinstance(const.name, conv) or const.name is _NONE_NAME:
+            return
+
+        newname = _constraint_name_for_table(const, table)
+        if newname:
+            const.name = newname
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/operators.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/operators.py
new file mode 100644
index 00000000..d5f876cb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/operators.py
@@ -0,0 +1,2623 @@
+# sql/operators.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Defines operators used in SQL expressions."""
+
+from __future__ import annotations
+
+from enum import IntEnum
+from operator import add as _uncast_add
+from operator import and_ as _uncast_and_
+from operator import contains as _uncast_contains
+from operator import eq as _uncast_eq
+from operator import floordiv as _uncast_floordiv
+from operator import ge as _uncast_ge
+from operator import getitem as _uncast_getitem
+from operator import gt as _uncast_gt
+from operator import inv as _uncast_inv
+from operator import le as _uncast_le
+from operator import lshift as _uncast_lshift
+from operator import lt as _uncast_lt
+from operator import mod as _uncast_mod
+from operator import mul as _uncast_mul
+from operator import ne as _uncast_ne
+from operator import neg as _uncast_neg
+from operator import or_ as _uncast_or_
+from operator import rshift as _uncast_rshift
+from operator import sub as _uncast_sub
+from operator import truediv as _uncast_truediv
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import exc
+from .. import util
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from ._typing import ColumnExpressionArgument
+    from .cache_key import CacheConst
+    from .elements import ColumnElement
+    from .type_api import TypeEngine
+
+_T = TypeVar("_T", bound=Any)
+_FN = TypeVar("_FN", bound=Callable[..., Any])
+
+
+class OperatorType(Protocol):
+    """describe an op() function."""
+
+    __slots__ = ()
+
+    __name__: str
+
+    @overload
+    def __call__(
+        self,
+        left: ColumnExpressionArgument[Any],
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> ColumnElement[Any]: ...
+
+    @overload
+    def __call__(
+        self,
+        left: Operators,
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> Operators: ...
+
+    def __call__(
+        self,
+        left: Any,
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> Operators: ...
+
+
+add = cast(OperatorType, _uncast_add)
+and_ = cast(OperatorType, _uncast_and_)
+contains = cast(OperatorType, _uncast_contains)
+eq = cast(OperatorType, _uncast_eq)
+floordiv = cast(OperatorType, _uncast_floordiv)
+ge = cast(OperatorType, _uncast_ge)
+getitem = cast(OperatorType, _uncast_getitem)
+gt = cast(OperatorType, _uncast_gt)
+inv = cast(OperatorType, _uncast_inv)
+le = cast(OperatorType, _uncast_le)
+lshift = cast(OperatorType, _uncast_lshift)
+lt = cast(OperatorType, _uncast_lt)
+mod = cast(OperatorType, _uncast_mod)
+mul = cast(OperatorType, _uncast_mul)
+ne = cast(OperatorType, _uncast_ne)
+neg = cast(OperatorType, _uncast_neg)
+or_ = cast(OperatorType, _uncast_or_)
+rshift = cast(OperatorType, _uncast_rshift)
+sub = cast(OperatorType, _uncast_sub)
+truediv = cast(OperatorType, _uncast_truediv)
+
+
+class Operators:
+    """Base of comparison and logical operators.
+
+    Implements base methods
+    :meth:`~sqlalchemy.sql.operators.Operators.operate` and
+    :meth:`~sqlalchemy.sql.operators.Operators.reverse_operate`, as well as
+    :meth:`~sqlalchemy.sql.operators.Operators.__and__`,
+    :meth:`~sqlalchemy.sql.operators.Operators.__or__`,
+    :meth:`~sqlalchemy.sql.operators.Operators.__invert__`.
+
+    Usually is used via its most common subclass
+    :class:`.ColumnOperators`.
+
+    """
+
+    __slots__ = ()
+
+    def __and__(self, other: Any) -> Operators:
+        """Implement the ``&`` operator.
+
+        When used with SQL expressions, results in an
+        AND operation, equivalent to
+        :func:`_expression.and_`, that is::
+
+            a & b
+
+        is equivalent to::
+
+            from sqlalchemy import and_
+
+            and_(a, b)
+
+        Care should be taken when using ``&`` regarding
+        operator precedence; the ``&`` operator has the highest precedence.
+        The operands should be enclosed in parenthesis if they contain
+        further sub expressions::
+
+            (a == 2) & (b == 4)
+
+        """
+        return self.operate(and_, other)
+
+    def __or__(self, other: Any) -> Operators:
+        """Implement the ``|`` operator.
+
+        When used with SQL expressions, results in an
+        OR operation, equivalent to
+        :func:`_expression.or_`, that is::
+
+            a | b
+
+        is equivalent to::
+
+            from sqlalchemy import or_
+
+            or_(a, b)
+
+        Care should be taken when using ``|`` regarding
+        operator precedence; the ``|`` operator has the highest precedence.
+        The operands should be enclosed in parenthesis if they contain
+        further sub expressions::
+
+            (a == 2) | (b == 4)
+
+        """
+        return self.operate(or_, other)
+
+    def __invert__(self) -> Operators:
+        """Implement the ``~`` operator.
+
+        When used with SQL expressions, results in a
+        NOT operation, equivalent to
+        :func:`_expression.not_`, that is::
+
+            ~a
+
+        is equivalent to::
+
+            from sqlalchemy import not_
+
+            not_(a)
+
+        """
+        return self.operate(inv)
+
+    def op(
+        self,
+        opstring: str,
+        precedence: int = 0,
+        is_comparison: bool = False,
+        return_type: Optional[
+            Union[Type[TypeEngine[Any]], TypeEngine[Any]]
+        ] = None,
+        python_impl: Optional[Callable[..., Any]] = None,
+    ) -> Callable[[Any], Operators]:
+        """Produce a generic operator function.
+
+        e.g.::
+
+          somecolumn.op("*")(5)
+
+        produces::
+
+          somecolumn * 5
+
+        This function can also be used to make bitwise operators explicit. For
+        example::
+
+          somecolumn.op("&")(0xFF)
+
+        is a bitwise AND of the value in ``somecolumn``.
+
+        :param opstring: a string which will be output as the infix operator
+          between this element and the expression passed to the
+          generated function.
+
+        :param precedence: precedence which the database is expected to apply
+         to the operator in SQL expressions. This integer value acts as a hint
+         for the SQL compiler to know when explicit parenthesis should be
+         rendered around a particular operation. A lower number will cause the
+         expression to be parenthesized when applied against another operator
+         with higher precedence. The default value of ``0`` is lower than all
+         operators except for the comma (``,``) and ``AS`` operators. A value
+         of 100 will be higher or equal to all operators, and -100 will be
+         lower than or equal to all operators.
+
+         .. seealso::
+
+            :ref:`faq_sql_expression_op_parenthesis` - detailed description
+            of how the SQLAlchemy SQL compiler renders parenthesis
+
+        :param is_comparison: legacy; if True, the operator will be considered
+         as a "comparison" operator, that is which evaluates to a boolean
+         true/false value, like ``==``, ``>``, etc.  This flag is provided
+         so that ORM relationships can establish that the operator is a
+         comparison operator when used in a custom join condition.
+
+         Using the ``is_comparison`` parameter is superseded by using the
+         :meth:`.Operators.bool_op` method instead;  this more succinct
+         operator sets this parameter automatically, but also provides
+         correct :pep:`484` typing support as the returned object will
+         express a "boolean" datatype, i.e. ``BinaryExpression[bool]``.
+
+        :param return_type: a :class:`.TypeEngine` class or object that will
+          force the return type of an expression produced by this operator
+          to be of that type.   By default, operators that specify
+          :paramref:`.Operators.op.is_comparison` will resolve to
+          :class:`.Boolean`, and those that do not will be of the same
+          type as the left-hand operand.
+
+        :param python_impl: an optional Python function that can evaluate
+         two Python values in the same way as this operator works when
+         run on the database server.  Useful for in-Python SQL expression
+         evaluation functions, such as for ORM hybrid attributes, and the
+         ORM "evaluator" used to match objects in a session after a multi-row
+         update or delete.
+
+         e.g.::
+
+            >>> expr = column("x").op("+", python_impl=lambda a, b: a + b)("y")
+
+         The operator for the above expression will also work for non-SQL
+         left and right objects::
+
+            >>> expr.operator(5, 10)
+            15
+
+         .. versionadded:: 2.0
+
+
+        .. seealso::
+
+            :meth:`.Operators.bool_op`
+
+            :ref:`types_operators`
+
+            :ref:`relationship_custom_operator`
+
+        """
+        operator = custom_op(
+            opstring,
+            precedence,
+            is_comparison,
+            return_type,
+            python_impl=python_impl,
+        )
+
+        def against(other: Any) -> Operators:
+            return operator(self, other)
+
+        return against
+
+    def bool_op(
+        self,
+        opstring: str,
+        precedence: int = 0,
+        python_impl: Optional[Callable[..., Any]] = None,
+    ) -> Callable[[Any], Operators]:
+        """Return a custom boolean operator.
+
+        This method is shorthand for calling
+        :meth:`.Operators.op` and passing the
+        :paramref:`.Operators.op.is_comparison`
+        flag with True.    A key advantage to using :meth:`.Operators.bool_op`
+        is that when using column constructs, the "boolean" nature of the
+        returned expression will be present for :pep:`484` purposes.
+
+        .. seealso::
+
+            :meth:`.Operators.op`
+
+        """
+        return self.op(
+            opstring,
+            precedence=precedence,
+            is_comparison=True,
+            python_impl=python_impl,
+        )
+
+    def operate(
+        self, op: OperatorType, *other: Any, **kwargs: Any
+    ) -> Operators:
+        r"""Operate on an argument.
+
+        This is the lowest level of operation, raises
+        :class:`NotImplementedError` by default.
+
+        Overriding this on a subclass can allow common
+        behavior to be applied to all operations.
+        For example, overriding :class:`.ColumnOperators`
+        to apply ``func.lower()`` to the left and right
+        side::
+
+            class MyComparator(ColumnOperators):
+                def operate(self, op, other, **kwargs):
+                    return op(func.lower(self), func.lower(other), **kwargs)
+
+        :param op:  Operator callable.
+        :param \*other: the 'other' side of the operation. Will
+         be a single scalar for most operations.
+        :param \**kwargs: modifiers.  These may be passed by special
+         operators such as :meth:`ColumnOperators.contains`.
+
+
+        """
+        raise NotImplementedError(str(op))
+
+    __sa_operate__ = operate
+
+    def reverse_operate(
+        self, op: OperatorType, other: Any, **kwargs: Any
+    ) -> Operators:
+        """Reverse operate on an argument.
+
+        Usage is the same as :meth:`operate`.
+
+        """
+        raise NotImplementedError(str(op))
+
+
+class custom_op(OperatorType, Generic[_T]):
+    """Represent a 'custom' operator.
+
+    :class:`.custom_op` is normally instantiated when the
+    :meth:`.Operators.op` or :meth:`.Operators.bool_op` methods
+    are used to create a custom operator callable.  The class can also be
+    used directly when programmatically constructing expressions.   E.g.
+    to represent the "factorial" operation::
+
+        from sqlalchemy.sql import UnaryExpression
+        from sqlalchemy.sql import operators
+        from sqlalchemy import Numeric
+
+        unary = UnaryExpression(
+            table.c.somecolumn, modifier=operators.custom_op("!"), type_=Numeric
+        )
+
+    .. seealso::
+
+        :meth:`.Operators.op`
+
+        :meth:`.Operators.bool_op`
+
+    """  # noqa: E501
+
+    __name__ = "custom_op"
+
+    __slots__ = (
+        "opstring",
+        "precedence",
+        "is_comparison",
+        "natural_self_precedent",
+        "eager_grouping",
+        "return_type",
+        "python_impl",
+    )
+
+    def __init__(
+        self,
+        opstring: str,
+        precedence: int = 0,
+        is_comparison: bool = False,
+        return_type: Optional[
+            Union[Type[TypeEngine[_T]], TypeEngine[_T]]
+        ] = None,
+        natural_self_precedent: bool = False,
+        eager_grouping: bool = False,
+        python_impl: Optional[Callable[..., Any]] = None,
+    ):
+        self.opstring = opstring
+        self.precedence = precedence
+        self.is_comparison = is_comparison
+        self.natural_self_precedent = natural_self_precedent
+        self.eager_grouping = eager_grouping
+        self.return_type = (
+            return_type._to_instance(return_type) if return_type else None
+        )
+        self.python_impl = python_impl
+
+    def __eq__(self, other: Any) -> bool:
+        return (
+            isinstance(other, custom_op)
+            and other._hash_key() == self._hash_key()
+        )
+
+    def __hash__(self) -> int:
+        return hash(self._hash_key())
+
+    def _hash_key(self) -> Union[CacheConst, Tuple[Any, ...]]:
+        return (
+            self.__class__,
+            self.opstring,
+            self.precedence,
+            self.is_comparison,
+            self.natural_self_precedent,
+            self.eager_grouping,
+            self.return_type._static_cache_key if self.return_type else None,
+        )
+
+    @overload
+    def __call__(
+        self,
+        left: ColumnExpressionArgument[Any],
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> ColumnElement[Any]: ...
+
+    @overload
+    def __call__(
+        self,
+        left: Operators,
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> Operators: ...
+
+    def __call__(
+        self,
+        left: Any,
+        right: Optional[Any] = None,
+        *other: Any,
+        **kwargs: Any,
+    ) -> Operators:
+        if hasattr(left, "__sa_operate__"):
+            return left.operate(self, right, *other, **kwargs)  # type: ignore
+        elif self.python_impl:
+            return self.python_impl(left, right, *other, **kwargs)  # type: ignore  # noqa: E501
+        else:
+            raise exc.InvalidRequestError(
+                f"Custom operator {self.opstring!r} can't be used with "
+                "plain Python objects unless it includes the "
+                "'python_impl' parameter."
+            )
+
+
+class ColumnOperators(Operators):
+    """Defines boolean, comparison, and other operators for
+    :class:`_expression.ColumnElement` expressions.
+
+    By default, all methods call down to
+    :meth:`.operate` or :meth:`.reverse_operate`,
+    passing in the appropriate operator function from the
+    Python builtin ``operator`` module or
+    a SQLAlchemy-specific operator function from
+    :mod:`sqlalchemy.expression.operators`.   For example
+    the ``__eq__`` function::
+
+        def __eq__(self, other):
+            return self.operate(operators.eq, other)
+
+    Where ``operators.eq`` is essentially::
+
+        def eq(a, b):
+            return a == b
+
+    The core column expression unit :class:`_expression.ColumnElement`
+    overrides :meth:`.Operators.operate` and others
+    to return further :class:`_expression.ColumnElement` constructs,
+    so that the ``==`` operation above is replaced by a clause
+    construct.
+
+    .. seealso::
+
+        :ref:`types_operators`
+
+        :attr:`.TypeEngine.comparator_factory`
+
+        :class:`.ColumnOperators`
+
+        :class:`.PropComparator`
+
+    """
+
+    __slots__ = ()
+
+    timetuple: Literal[None] = None
+    """Hack, allows datetime objects to be compared on the LHS."""
+
+    if typing.TYPE_CHECKING:
+
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnOperators: ...
+
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnOperators: ...
+
+    def __lt__(self, other: Any) -> ColumnOperators:
+        """Implement the ``<`` operator.
+
+        In a column context, produces the clause ``a < b``.
+
+        """
+        return self.operate(lt, other)
+
+    def __le__(self, other: Any) -> ColumnOperators:
+        """Implement the ``<=`` operator.
+
+        In a column context, produces the clause ``a <= b``.
+
+        """
+        return self.operate(le, other)
+
+    # ColumnOperators defines an __eq__ so it must explicitly declare also
+    # an hash or it's set to None by python:
+    # https://docs.python.org/3/reference/datamodel.html#object.__hash__
+    if TYPE_CHECKING:
+
+        def __hash__(self) -> int: ...
+
+    else:
+        __hash__ = Operators.__hash__
+
+    def __eq__(self, other: Any) -> ColumnOperators:  # type: ignore[override]
+        """Implement the ``==`` operator.
+
+        In a column context, produces the clause ``a = b``.
+        If the target is ``None``, produces ``a IS NULL``.
+
+        """
+        return self.operate(eq, other)
+
+    def __ne__(self, other: Any) -> ColumnOperators:  # type: ignore[override]
+        """Implement the ``!=`` operator.
+
+        In a column context, produces the clause ``a != b``.
+        If the target is ``None``, produces ``a IS NOT NULL``.
+
+        """
+        return self.operate(ne, other)
+
+    def is_distinct_from(self, other: Any) -> ColumnOperators:
+        """Implement the ``IS DISTINCT FROM`` operator.
+
+        Renders "a IS DISTINCT FROM b" on most platforms;
+        on some such as SQLite may render "a IS NOT b".
+
+        """
+        return self.operate(is_distinct_from, other)
+
+    def is_not_distinct_from(self, other: Any) -> ColumnOperators:
+        """Implement the ``IS NOT DISTINCT FROM`` operator.
+
+        Renders "a IS NOT DISTINCT FROM b" on most platforms;
+        on some such as SQLite may render "a IS b".
+
+        .. versionchanged:: 1.4 The ``is_not_distinct_from()`` operator is
+           renamed from ``isnot_distinct_from()`` in previous releases.
+           The previous name remains available for backwards compatibility.
+
+        """
+        return self.operate(is_not_distinct_from, other)
+
+    # deprecated 1.4; see #5435
+    if TYPE_CHECKING:
+
+        def isnot_distinct_from(self, other: Any) -> ColumnOperators: ...
+
+    else:
+        isnot_distinct_from = is_not_distinct_from
+
+    def __gt__(self, other: Any) -> ColumnOperators:
+        """Implement the ``>`` operator.
+
+        In a column context, produces the clause ``a > b``.
+
+        """
+        return self.operate(gt, other)
+
+    def __ge__(self, other: Any) -> ColumnOperators:
+        """Implement the ``>=`` operator.
+
+        In a column context, produces the clause ``a >= b``.
+
+        """
+        return self.operate(ge, other)
+
+    def __neg__(self) -> ColumnOperators:
+        """Implement the ``-`` operator.
+
+        In a column context, produces the clause ``-a``.
+
+        """
+        return self.operate(neg)
+
+    def __contains__(self, other: Any) -> ColumnOperators:
+        return self.operate(contains, other)
+
+    def __getitem__(self, index: Any) -> ColumnOperators:
+        """Implement the [] operator.
+
+        This can be used by some database-specific types
+        such as PostgreSQL ARRAY and HSTORE.
+
+        """
+        return self.operate(getitem, index)
+
+    def __lshift__(self, other: Any) -> ColumnOperators:
+        """implement the << operator.
+
+        Not used by SQLAlchemy core, this is provided
+        for custom operator systems which want to use
+        << as an extension point.
+        """
+        return self.operate(lshift, other)
+
+    def __rshift__(self, other: Any) -> ColumnOperators:
+        """implement the >> operator.
+
+        Not used by SQLAlchemy core, this is provided
+        for custom operator systems which want to use
+        >> as an extension point.
+        """
+        return self.operate(rshift, other)
+
+    def concat(self, other: Any) -> ColumnOperators:
+        """Implement the 'concat' operator.
+
+        In a column context, produces the clause ``a || b``,
+        or uses the ``concat()`` operator on MySQL.
+
+        """
+        return self.operate(concat_op, other)
+
+    def _rconcat(self, other: Any) -> ColumnOperators:
+        """Implement an 'rconcat' operator.
+
+        this is for internal use at the moment
+
+        .. versionadded:: 1.4.40
+
+        """
+        return self.reverse_operate(concat_op, other)
+
+    def like(
+        self, other: Any, escape: Optional[str] = None
+    ) -> ColumnOperators:
+        r"""Implement the ``like`` operator.
+
+        In a column context, produces the expression:
+
+        .. sourcecode:: sql
+
+            a LIKE other
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.like("%foobar%"))
+
+        :param other: expression to be compared
+        :param escape: optional escape character, renders the ``ESCAPE``
+          keyword, e.g.::
+
+            somecolumn.like("foo/%bar", escape="/")
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.ilike`
+
+        """
+        return self.operate(like_op, other, escape=escape)
+
+    def ilike(
+        self, other: Any, escape: Optional[str] = None
+    ) -> ColumnOperators:
+        r"""Implement the ``ilike`` operator, e.g. case insensitive LIKE.
+
+        In a column context, produces an expression either of the form:
+
+        .. sourcecode:: sql
+
+            lower(a) LIKE lower(other)
+
+        Or on backends that support the ILIKE operator:
+
+        .. sourcecode:: sql
+
+            a ILIKE other
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.ilike("%foobar%"))
+
+        :param other: expression to be compared
+        :param escape: optional escape character, renders the ``ESCAPE``
+          keyword, e.g.::
+
+            somecolumn.ilike("foo/%bar", escape="/")
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.like`
+
+        """  # noqa: E501
+        return self.operate(ilike_op, other, escape=escape)
+
+    def bitwise_xor(self, other: Any) -> ColumnOperators:
+        """Produce a bitwise XOR operation, typically via the ``^``
+        operator, or ``#`` for PostgreSQL.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_xor_op, other)
+
+    def bitwise_or(self, other: Any) -> ColumnOperators:
+        """Produce a bitwise OR operation, typically via the ``|``
+        operator.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_or_op, other)
+
+    def bitwise_and(self, other: Any) -> ColumnOperators:
+        """Produce a bitwise AND operation, typically via the ``&``
+        operator.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_and_op, other)
+
+    def bitwise_not(self) -> ColumnOperators:
+        """Produce a bitwise NOT operation, typically via the ``~``
+        operator.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_not_op)
+
+    def bitwise_lshift(self, other: Any) -> ColumnOperators:
+        """Produce a bitwise LSHIFT operation, typically via the ``<<``
+        operator.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_lshift_op, other)
+
+    def bitwise_rshift(self, other: Any) -> ColumnOperators:
+        """Produce a bitwise RSHIFT operation, typically via the ``>>``
+        operator.
+
+        .. versionadded:: 2.0.2
+
+        .. seealso::
+
+            :ref:`operators_bitwise`
+
+        """
+
+        return self.operate(bitwise_rshift_op, other)
+
+    def in_(self, other: Any) -> ColumnOperators:
+        """Implement the ``in`` operator.
+
+        In a column context, produces the clause ``column IN <other>``.
+
+        The given parameter ``other`` may be:
+
+        * A list of literal values,
+          e.g.::
+
+            stmt.where(column.in_([1, 2, 3]))
+
+          In this calling form, the list of items is converted to a set of
+          bound parameters the same length as the list given:
+
+          .. sourcecode:: sql
+
+            WHERE COL IN (?, ?, ?)
+
+        * A list of tuples may be provided if the comparison is against a
+          :func:`.tuple_` containing multiple expressions::
+
+            from sqlalchemy import tuple_
+
+            stmt.where(tuple_(col1, col2).in_([(1, 10), (2, 20), (3, 30)]))
+
+        * An empty list,
+          e.g.::
+
+            stmt.where(column.in_([]))
+
+          In this calling form, the expression renders an "empty set"
+          expression.  These expressions are tailored to individual backends
+          and are generally trying to get an empty SELECT statement as a
+          subquery.  Such as on SQLite, the expression is:
+
+          .. sourcecode:: sql
+
+            WHERE col IN (SELECT 1 FROM (SELECT 1) WHERE 1!=1)
+
+          .. versionchanged:: 1.4  empty IN expressions now use an
+             execution-time generated SELECT subquery in all cases.
+
+        * A bound parameter, e.g. :func:`.bindparam`, may be used if it
+          includes the :paramref:`.bindparam.expanding` flag::
+
+            stmt.where(column.in_(bindparam("value", expanding=True)))
+
+          In this calling form, the expression renders a special non-SQL
+          placeholder expression that looks like:
+
+          .. sourcecode:: sql
+
+            WHERE COL IN ([EXPANDING_value])
+
+          This placeholder expression is intercepted at statement execution
+          time to be converted into the variable number of bound parameter
+          form illustrated earlier.   If the statement were executed as::
+
+            connection.execute(stmt, {"value": [1, 2, 3]})
+
+          The database would be passed a bound parameter for each value:
+
+          .. sourcecode:: sql
+
+            WHERE COL IN (?, ?, ?)
+
+          .. versionadded:: 1.2 added "expanding" bound parameters
+
+          If an empty list is passed, a special "empty list" expression,
+          which is specific to the database in use, is rendered.  On
+          SQLite this would be:
+
+          .. sourcecode:: sql
+
+            WHERE COL IN (SELECT 1 FROM (SELECT 1) WHERE 1!=1)
+
+          .. versionadded:: 1.3 "expanding" bound parameters now support
+             empty lists
+
+        * a :func:`_expression.select` construct, which is usually a
+          correlated scalar select::
+
+            stmt.where(
+                column.in_(select(othertable.c.y).where(table.c.x == othertable.c.x))
+            )
+
+          In this calling form, :meth:`.ColumnOperators.in_` renders as given:
+
+          .. sourcecode:: sql
+
+            WHERE COL IN (SELECT othertable.y
+            FROM othertable WHERE othertable.x = table.x)
+
+        :param other: a list of literals, a :func:`_expression.select`
+         construct, or a :func:`.bindparam` construct that includes the
+         :paramref:`.bindparam.expanding` flag set to True.
+
+        """  # noqa: E501
+        return self.operate(in_op, other)
+
+    def not_in(self, other: Any) -> ColumnOperators:
+        """implement the ``NOT IN`` operator.
+
+        This is equivalent to using negation with
+        :meth:`.ColumnOperators.in_`, i.e. ``~x.in_(y)``.
+
+        In the case that ``other`` is an empty sequence, the compiler
+        produces an "empty not in" expression.   This defaults to the
+        expression "1 = 1" to produce true in all cases.  The
+        :paramref:`_sa.create_engine.empty_in_strategy` may be used to
+        alter this behavior.
+
+        .. versionchanged:: 1.4 The ``not_in()`` operator is renamed from
+           ``notin_()`` in previous releases.  The previous name remains
+           available for backwards compatibility.
+
+        .. versionchanged:: 1.2  The :meth:`.ColumnOperators.in_` and
+           :meth:`.ColumnOperators.not_in` operators
+           now produce a "static" expression for an empty IN sequence
+           by default.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.in_`
+
+        """
+        return self.operate(not_in_op, other)
+
+    # deprecated 1.4; see #5429
+    if TYPE_CHECKING:
+
+        def notin_(self, other: Any) -> ColumnOperators: ...
+
+    else:
+        notin_ = not_in
+
+    def not_like(
+        self, other: Any, escape: Optional[str] = None
+    ) -> ColumnOperators:
+        """implement the ``NOT LIKE`` operator.
+
+        This is equivalent to using negation with
+        :meth:`.ColumnOperators.like`, i.e. ``~x.like(y)``.
+
+        .. versionchanged:: 1.4 The ``not_like()`` operator is renamed from
+           ``notlike()`` in previous releases.  The previous name remains
+           available for backwards compatibility.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.like`
+
+        """
+        return self.operate(not_like_op, other, escape=escape)
+
+    # deprecated 1.4; see #5435
+    if TYPE_CHECKING:
+
+        def notlike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> ColumnOperators: ...
+
+    else:
+        notlike = not_like
+
+    def not_ilike(
+        self, other: Any, escape: Optional[str] = None
+    ) -> ColumnOperators:
+        """implement the ``NOT ILIKE`` operator.
+
+        This is equivalent to using negation with
+        :meth:`.ColumnOperators.ilike`, i.e. ``~x.ilike(y)``.
+
+        .. versionchanged:: 1.4 The ``not_ilike()`` operator is renamed from
+           ``notilike()`` in previous releases.  The previous name remains
+           available for backwards compatibility.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.ilike`
+
+        """
+        return self.operate(not_ilike_op, other, escape=escape)
+
+    # deprecated 1.4; see #5435
+    if TYPE_CHECKING:
+
+        def notilike(
+            self, other: Any, escape: Optional[str] = None
+        ) -> ColumnOperators: ...
+
+    else:
+        notilike = not_ilike
+
+    def is_(self, other: Any) -> ColumnOperators:
+        """Implement the ``IS`` operator.
+
+        Normally, ``IS`` is generated automatically when comparing to a
+        value of ``None``, which resolves to ``NULL``.  However, explicit
+        usage of ``IS`` may be desirable if comparing to boolean values
+        on certain platforms.
+
+        .. seealso:: :meth:`.ColumnOperators.is_not`
+
+        """
+        return self.operate(is_, other)
+
+    def is_not(self, other: Any) -> ColumnOperators:
+        """Implement the ``IS NOT`` operator.
+
+        Normally, ``IS NOT`` is generated automatically when comparing to a
+        value of ``None``, which resolves to ``NULL``.  However, explicit
+        usage of ``IS NOT`` may be desirable if comparing to boolean values
+        on certain platforms.
+
+        .. versionchanged:: 1.4 The ``is_not()`` operator is renamed from
+           ``isnot()`` in previous releases.  The previous name remains
+           available for backwards compatibility.
+
+        .. seealso:: :meth:`.ColumnOperators.is_`
+
+        """
+        return self.operate(is_not, other)
+
+    # deprecated 1.4; see #5429
+    if TYPE_CHECKING:
+
+        def isnot(self, other: Any) -> ColumnOperators: ...
+
+    else:
+        isnot = is_not
+
+    def startswith(
+        self,
+        other: Any,
+        escape: Optional[str] = None,
+        autoescape: bool = False,
+    ) -> ColumnOperators:
+        r"""Implement the ``startswith`` operator.
+
+        Produces a LIKE expression that tests against a match for the start
+        of a string value:
+
+        .. sourcecode:: sql
+
+            column LIKE <other> || '%'
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.startswith("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.startswith.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.startswith.escape` parameter will establish
+        a given character as an escape character which can be of use when
+        the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.startswith.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.startswith("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE :param || '%' ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.startswith("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE :param || '%' ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.startswith.autoescape`::
+
+            somecolumn.startswith("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.endswith`
+
+            :meth:`.ColumnOperators.contains`
+
+            :meth:`.ColumnOperators.like`
+
+        """  # noqa: E501
+        return self.operate(
+            startswith_op, other, escape=escape, autoescape=autoescape
+        )
+
+    def istartswith(
+        self,
+        other: Any,
+        escape: Optional[str] = None,
+        autoescape: bool = False,
+    ) -> ColumnOperators:
+        r"""Implement the ``istartswith`` operator, e.g. case insensitive
+        version of :meth:`.ColumnOperators.startswith`.
+
+        Produces a LIKE expression that tests against an insensitive
+        match for the start of a string value:
+
+        .. sourcecode:: sql
+
+            lower(column) LIKE lower(<other>) || '%'
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.istartswith("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.istartswith.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.istartswith.escape` parameter will
+        establish a given character as an escape character which can be of
+        use when the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.istartswith.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.istartswith("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE lower(:param) || '%' ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.istartswith("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE lower(:param) || '%' ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.istartswith.autoescape`::
+
+            somecolumn.istartswith("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.startswith`
+        """  # noqa: E501
+        return self.operate(
+            istartswith_op, other, escape=escape, autoescape=autoescape
+        )
+
+    def endswith(
+        self,
+        other: Any,
+        escape: Optional[str] = None,
+        autoescape: bool = False,
+    ) -> ColumnOperators:
+        r"""Implement the 'endswith' operator.
+
+        Produces a LIKE expression that tests against a match for the end
+        of a string value:
+
+        .. sourcecode:: sql
+
+            column LIKE '%' || <other>
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.endswith("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.endswith.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.endswith.escape` parameter will establish
+        a given character as an escape character which can be of use when
+        the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.endswith.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.endswith("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE '%' || :param ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.endswith("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE '%' || :param ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.endswith.autoescape`::
+
+            somecolumn.endswith("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.startswith`
+
+            :meth:`.ColumnOperators.contains`
+
+            :meth:`.ColumnOperators.like`
+
+        """  # noqa: E501
+        return self.operate(
+            endswith_op, other, escape=escape, autoescape=autoescape
+        )
+
+    def iendswith(
+        self,
+        other: Any,
+        escape: Optional[str] = None,
+        autoescape: bool = False,
+    ) -> ColumnOperators:
+        r"""Implement the ``iendswith`` operator, e.g. case insensitive
+        version of :meth:`.ColumnOperators.endswith`.
+
+        Produces a LIKE expression that tests against an insensitive match
+        for the end of a string value:
+
+        .. sourcecode:: sql
+
+            lower(column) LIKE '%' || lower(<other>)
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.iendswith("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.iendswith.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.iendswith.escape` parameter will establish
+        a given character as an escape character which can be of use when
+        the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.iendswith.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.iendswith("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE '%' || lower(:param) ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.iendswith("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE '%' || lower(:param) ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.iendswith.autoescape`::
+
+            somecolumn.endswith("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.endswith`
+        """  # noqa: E501
+        return self.operate(
+            iendswith_op, other, escape=escape, autoescape=autoescape
+        )
+
+    def contains(self, other: Any, **kw: Any) -> ColumnOperators:
+        r"""Implement the 'contains' operator.
+
+        Produces a LIKE expression that tests against a match for the middle
+        of a string value:
+
+        .. sourcecode:: sql
+
+            column LIKE '%' || <other> || '%'
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.contains("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.contains.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.contains.escape` parameter will establish
+        a given character as an escape character which can be of use when
+        the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.contains.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.contains("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE '%' || :param || '%' ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.contains("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            somecolumn LIKE '%' || :param || '%' ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.contains.autoescape`::
+
+            somecolumn.contains("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.startswith`
+
+            :meth:`.ColumnOperators.endswith`
+
+            :meth:`.ColumnOperators.like`
+
+
+        """  # noqa: E501
+        return self.operate(contains_op, other, **kw)
+
+    def icontains(self, other: Any, **kw: Any) -> ColumnOperators:
+        r"""Implement the ``icontains`` operator, e.g. case insensitive
+        version of :meth:`.ColumnOperators.contains`.
+
+        Produces a LIKE expression that tests against an insensitive match
+        for the middle of a string value:
+
+        .. sourcecode:: sql
+
+            lower(column) LIKE '%' || lower(<other>) || '%'
+
+        E.g.::
+
+            stmt = select(sometable).where(sometable.c.column.icontains("foobar"))
+
+        Since the operator uses ``LIKE``, wildcard characters
+        ``"%"`` and ``"_"`` that are present inside the <other> expression
+        will behave like wildcards as well.   For literal string
+        values, the :paramref:`.ColumnOperators.icontains.autoescape` flag
+        may be set to ``True`` to apply escaping to occurrences of these
+        characters within the string value so that they match as themselves
+        and not as wildcard characters.  Alternatively, the
+        :paramref:`.ColumnOperators.icontains.escape` parameter will establish
+        a given character as an escape character which can be of use when
+        the target expression is not a literal string.
+
+        :param other: expression to be compared.   This is usually a plain
+          string value, but can also be an arbitrary SQL expression.  LIKE
+          wildcard characters ``%`` and ``_`` are not escaped by default unless
+          the :paramref:`.ColumnOperators.icontains.autoescape` flag is
+          set to True.
+
+        :param autoescape: boolean; when True, establishes an escape character
+          within the LIKE expression, then applies it to all occurrences of
+          ``"%"``, ``"_"`` and the escape character itself within the
+          comparison value, which is assumed to be a literal string and not a
+          SQL expression.
+
+          An expression such as::
+
+            somecolumn.icontains("foo%bar", autoescape=True)
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE '%' || lower(:param) || '%' ESCAPE '/'
+
+          With the value of ``:param`` as ``"foo/%bar"``.
+
+        :param escape: a character which when given will render with the
+          ``ESCAPE`` keyword to establish that character as the escape
+          character.  This character can then be placed preceding occurrences
+          of ``%`` and ``_`` to allow them to act as themselves and not
+          wildcard characters.
+
+          An expression such as::
+
+            somecolumn.icontains("foo/%bar", escape="^")
+
+          Will render as:
+
+          .. sourcecode:: sql
+
+            lower(somecolumn) LIKE '%' || lower(:param) || '%' ESCAPE '^'
+
+          The parameter may also be combined with
+          :paramref:`.ColumnOperators.contains.autoescape`::
+
+            somecolumn.icontains("foo%bar^bat", escape="^", autoescape=True)
+
+          Where above, the given literal parameter will be converted to
+          ``"foo^%bar^^bat"`` before being passed to the database.
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.contains`
+
+        """  # noqa: E501
+        return self.operate(icontains_op, other, **kw)
+
+    def match(self, other: Any, **kwargs: Any) -> ColumnOperators:
+        """Implements a database-specific 'match' operator.
+
+        :meth:`_sql.ColumnOperators.match` attempts to resolve to
+        a MATCH-like function or operator provided by the backend.
+        Examples include:
+
+        * PostgreSQL - renders ``x @@ plainto_tsquery(y)``
+
+            .. versionchanged:: 2.0  ``plainto_tsquery()`` is used instead
+               of ``to_tsquery()`` for PostgreSQL now; for compatibility with
+               other forms, see :ref:`postgresql_match`.
+
+
+        * MySQL - renders ``MATCH (x) AGAINST (y IN BOOLEAN MODE)``
+
+          .. seealso::
+
+                :class:`_mysql.match` - MySQL specific construct with
+                additional features.
+
+        * Oracle Database - renders ``CONTAINS(x, y)``
+        * other backends may provide special implementations.
+        * Backends without any special implementation will emit
+          the operator as "MATCH".  This is compatible with SQLite, for
+          example.
+
+        """
+        return self.operate(match_op, other, **kwargs)
+
+    def regexp_match(
+        self, pattern: Any, flags: Optional[str] = None
+    ) -> ColumnOperators:
+        """Implements a database-specific 'regexp match' operator.
+
+        E.g.::
+
+            stmt = select(table.c.some_column).where(
+                table.c.some_column.regexp_match("^(b|c)")
+            )
+
+        :meth:`_sql.ColumnOperators.regexp_match` attempts to resolve to
+        a REGEXP-like function or operator provided by the backend, however
+        the specific regular expression syntax and flags available are
+        **not backend agnostic**.
+
+        Examples include:
+
+        * PostgreSQL - renders ``x ~ y`` or ``x !~ y`` when negated.
+        * Oracle Database - renders ``REGEXP_LIKE(x, y)``
+        * SQLite - uses SQLite's ``REGEXP`` placeholder operator and calls into
+          the Python ``re.match()`` builtin.
+        * other backends may provide special implementations.
+        * Backends without any special implementation will emit
+          the operator as "REGEXP" or "NOT REGEXP".  This is compatible with
+          SQLite and MySQL, for example.
+
+        Regular expression support is currently implemented for Oracle
+        Database, PostgreSQL, MySQL and MariaDB.  Partial support is available
+        for SQLite.  Support among third-party dialects may vary.
+
+        :param pattern: The regular expression pattern string or column
+          clause.
+        :param flags: Any regular expression string flags to apply, passed as
+          plain Python string only.  These flags are backend specific.
+          Some backends, like PostgreSQL and MariaDB, may alternatively
+          specify the flags as part of the pattern.
+          When using the ignore case flag 'i' in PostgreSQL, the ignore case
+          regexp match operator ``~*`` or ``!~*`` will be used.
+
+        .. versionadded:: 1.4
+
+        .. versionchanged:: 1.4.48, 2.0.18  Note that due to an implementation
+           error, the "flags" parameter previously accepted SQL expression
+           objects such as column expressions in addition to plain Python
+           strings.   This implementation did not work correctly with caching
+           and was removed; strings only should be passed for the "flags"
+           parameter, as these flags are rendered as literal inline values
+           within SQL expressions.
+
+        .. seealso::
+
+            :meth:`_sql.ColumnOperators.regexp_replace`
+
+
+        """
+        return self.operate(regexp_match_op, pattern, flags=flags)
+
+    def regexp_replace(
+        self, pattern: Any, replacement: Any, flags: Optional[str] = None
+    ) -> ColumnOperators:
+        """Implements a database-specific 'regexp replace' operator.
+
+        E.g.::
+
+            stmt = select(
+                table.c.some_column.regexp_replace("b(..)", "X\1Y", flags="g")
+            )
+
+        :meth:`_sql.ColumnOperators.regexp_replace` attempts to resolve to
+        a REGEXP_REPLACE-like function provided by the backend, that
+        usually emit the function ``REGEXP_REPLACE()``.  However,
+        the specific regular expression syntax and flags available are
+        **not backend agnostic**.
+
+        Regular expression replacement support is currently implemented for
+        Oracle Database, PostgreSQL, MySQL 8 or greater and MariaDB.  Support
+        among third-party dialects may vary.
+
+        :param pattern: The regular expression pattern string or column
+          clause.
+        :param pattern: The replacement string or column clause.
+        :param flags: Any regular expression string flags to apply, passed as
+          plain Python string only.  These flags are backend specific.
+          Some backends, like PostgreSQL and MariaDB, may alternatively
+          specify the flags as part of the pattern.
+
+        .. versionadded:: 1.4
+
+        .. versionchanged:: 1.4.48, 2.0.18  Note that due to an implementation
+           error, the "flags" parameter previously accepted SQL expression
+           objects such as column expressions in addition to plain Python
+           strings.   This implementation did not work correctly with caching
+           and was removed; strings only should be passed for the "flags"
+           parameter, as these flags are rendered as literal inline values
+           within SQL expressions.
+
+
+        .. seealso::
+
+            :meth:`_sql.ColumnOperators.regexp_match`
+
+        """
+        return self.operate(
+            regexp_replace_op,
+            pattern,
+            replacement=replacement,
+            flags=flags,
+        )
+
+    def desc(self) -> ColumnOperators:
+        """Produce a :func:`_expression.desc` clause against the
+        parent object."""
+        return self.operate(desc_op)
+
+    def asc(self) -> ColumnOperators:
+        """Produce a :func:`_expression.asc` clause against the
+        parent object."""
+        return self.operate(asc_op)
+
+    def nulls_first(self) -> ColumnOperators:
+        """Produce a :func:`_expression.nulls_first` clause against the
+        parent object.
+
+        .. versionchanged:: 1.4 The ``nulls_first()`` operator is
+           renamed from ``nullsfirst()`` in previous releases.
+           The previous name remains available for backwards compatibility.
+        """
+        return self.operate(nulls_first_op)
+
+    # deprecated 1.4; see #5435
+    if TYPE_CHECKING:
+
+        def nullsfirst(self) -> ColumnOperators: ...
+
+    else:
+        nullsfirst = nulls_first
+
+    def nulls_last(self) -> ColumnOperators:
+        """Produce a :func:`_expression.nulls_last` clause against the
+        parent object.
+
+        .. versionchanged:: 1.4 The ``nulls_last()`` operator is
+           renamed from ``nullslast()`` in previous releases.
+           The previous name remains available for backwards compatibility.
+        """
+        return self.operate(nulls_last_op)
+
+    # deprecated 1.4; see #5429
+    if TYPE_CHECKING:
+
+        def nullslast(self) -> ColumnOperators: ...
+
+    else:
+        nullslast = nulls_last
+
+    def collate(self, collation: str) -> ColumnOperators:
+        """Produce a :func:`_expression.collate` clause against
+        the parent object, given the collation string.
+
+        .. seealso::
+
+            :func:`_expression.collate`
+
+        """
+        return self.operate(collate, collation)
+
+    def __radd__(self, other: Any) -> ColumnOperators:
+        """Implement the ``+`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__add__`.
+
+        """
+        return self.reverse_operate(add, other)
+
+    def __rsub__(self, other: Any) -> ColumnOperators:
+        """Implement the ``-`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__sub__`.
+
+        """
+        return self.reverse_operate(sub, other)
+
+    def __rmul__(self, other: Any) -> ColumnOperators:
+        """Implement the ``*`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__mul__`.
+
+        """
+        return self.reverse_operate(mul, other)
+
+    def __rmod__(self, other: Any) -> ColumnOperators:
+        """Implement the ``%`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__mod__`.
+
+        """
+        return self.reverse_operate(mod, other)
+
+    def between(
+        self, cleft: Any, cright: Any, symmetric: bool = False
+    ) -> ColumnOperators:
+        """Produce a :func:`_expression.between` clause against
+        the parent object, given the lower and upper range.
+
+        """
+        return self.operate(between_op, cleft, cright, symmetric=symmetric)
+
+    def distinct(self) -> ColumnOperators:
+        """Produce a :func:`_expression.distinct` clause against the
+        parent object.
+
+        """
+        return self.operate(distinct_op)
+
+    def any_(self) -> ColumnOperators:
+        """Produce an :func:`_expression.any_` clause against the
+        parent object.
+
+        See the documentation for :func:`_sql.any_` for examples.
+
+        .. note:: be sure to not confuse the newer
+            :meth:`_sql.ColumnOperators.any_` method with the **legacy**
+            version of this method, the :meth:`_types.ARRAY.Comparator.any`
+            method that's specific to :class:`_types.ARRAY`, which uses a
+            different calling style.
+
+        """
+        return self.operate(any_op)
+
+    def all_(self) -> ColumnOperators:
+        """Produce an :func:`_expression.all_` clause against the
+        parent object.
+
+        See the documentation for :func:`_sql.all_` for examples.
+
+        .. note:: be sure to not confuse the newer
+            :meth:`_sql.ColumnOperators.all_` method with the **legacy**
+            version of this method, the :meth:`_types.ARRAY.Comparator.all`
+            method that's specific to :class:`_types.ARRAY`, which uses a
+            different calling style.
+
+        """
+        return self.operate(all_op)
+
+    def __add__(self, other: Any) -> ColumnOperators:
+        """Implement the ``+`` operator.
+
+        In a column context, produces the clause ``a + b``
+        if the parent object has non-string affinity.
+        If the parent object has a string affinity,
+        produces the concatenation operator, ``a || b`` -
+        see :meth:`.ColumnOperators.concat`.
+
+        """
+        return self.operate(add, other)
+
+    def __sub__(self, other: Any) -> ColumnOperators:
+        """Implement the ``-`` operator.
+
+        In a column context, produces the clause ``a - b``.
+
+        """
+        return self.operate(sub, other)
+
+    def __mul__(self, other: Any) -> ColumnOperators:
+        """Implement the ``*`` operator.
+
+        In a column context, produces the clause ``a * b``.
+
+        """
+        return self.operate(mul, other)
+
+    def __mod__(self, other: Any) -> ColumnOperators:
+        """Implement the ``%`` operator.
+
+        In a column context, produces the clause ``a % b``.
+
+        """
+        return self.operate(mod, other)
+
+    def __truediv__(self, other: Any) -> ColumnOperators:
+        """Implement the ``/`` operator.
+
+        In a column context, produces the clause ``a / b``, and
+        considers the result type to be numeric.
+
+        .. versionchanged:: 2.0  The truediv operator against two integers
+           is now considered to return a numeric value.    Behavior on specific
+           backends may vary.
+
+        """
+        return self.operate(truediv, other)
+
+    def __rtruediv__(self, other: Any) -> ColumnOperators:
+        """Implement the ``/`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__truediv__`.
+
+        """
+        return self.reverse_operate(truediv, other)
+
+    def __floordiv__(self, other: Any) -> ColumnOperators:
+        """Implement the ``//`` operator.
+
+        In a column context, produces the clause ``a / b``,
+        which is the same as "truediv", but considers the result
+        type to be integer.
+
+        .. versionadded:: 2.0
+
+        """
+        return self.operate(floordiv, other)
+
+    def __rfloordiv__(self, other: Any) -> ColumnOperators:
+        """Implement the ``//`` operator in reverse.
+
+        See :meth:`.ColumnOperators.__floordiv__`.
+
+        """
+        return self.reverse_operate(floordiv, other)
+
+
+_commutative: Set[Any] = {eq, ne, add, mul}
+_comparison: Set[Any] = {eq, ne, lt, gt, ge, le}
+
+
+def _operator_fn(fn: Callable[..., Any]) -> OperatorType:
+    return cast(OperatorType, fn)
+
+
+def commutative_op(fn: _FN) -> _FN:
+    _commutative.add(fn)
+    return fn
+
+
+def comparison_op(fn: _FN) -> _FN:
+    _comparison.add(fn)
+    return fn
+
+
+@_operator_fn
+def from_() -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+@comparison_op
+def function_as_comparison_op() -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def as_() -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def exists() -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def is_true(a: Any) -> Any:
+    raise NotImplementedError()
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def istrue(a: Any) -> Any: ...
+
+else:
+    istrue = is_true
+
+
+@_operator_fn
+def is_false(a: Any) -> Any:
+    raise NotImplementedError()
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def isfalse(a: Any) -> Any: ...
+
+else:
+    isfalse = is_false
+
+
+@comparison_op
+@_operator_fn
+def is_distinct_from(a: Any, b: Any) -> Any:
+    return a.is_distinct_from(b)
+
+
+@comparison_op
+@_operator_fn
+def is_not_distinct_from(a: Any, b: Any) -> Any:
+    return a.is_not_distinct_from(b)
+
+
+# deprecated 1.4; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def isnot_distinct_from(a: Any, b: Any) -> Any: ...
+
+else:
+    isnot_distinct_from = is_not_distinct_from
+
+
+@comparison_op
+@_operator_fn
+def is_(a: Any, b: Any) -> Any:
+    return a.is_(b)
+
+
+@comparison_op
+@_operator_fn
+def is_not(a: Any, b: Any) -> Any:
+    return a.is_not(b)
+
+
+# 1.4 deprecated; see #5429
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def isnot(a: Any, b: Any) -> Any: ...
+
+else:
+    isnot = is_not
+
+
+@_operator_fn
+def collate(a: Any, b: Any) -> Any:
+    return a.collate(b)
+
+
+@_operator_fn
+def op(a: Any, opstring: str, b: Any) -> Any:
+    return a.op(opstring)(b)
+
+
+@comparison_op
+@_operator_fn
+def like_op(a: Any, b: Any, escape: Optional[str] = None) -> Any:
+    return a.like(b, escape=escape)
+
+
+@comparison_op
+@_operator_fn
+def not_like_op(a: Any, b: Any, escape: Optional[str] = None) -> Any:
+    return a.notlike(b, escape=escape)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notlike_op(a: Any, b: Any, escape: Optional[str] = None) -> Any: ...
+
+else:
+    notlike_op = not_like_op
+
+
+@comparison_op
+@_operator_fn
+def ilike_op(a: Any, b: Any, escape: Optional[str] = None) -> Any:
+    return a.ilike(b, escape=escape)
+
+
+@comparison_op
+@_operator_fn
+def not_ilike_op(a: Any, b: Any, escape: Optional[str] = None) -> Any:
+    return a.not_ilike(b, escape=escape)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notilike_op(a: Any, b: Any, escape: Optional[str] = None) -> Any: ...
+
+else:
+    notilike_op = not_ilike_op
+
+
+@comparison_op
+@_operator_fn
+def between_op(a: Any, b: Any, c: Any, symmetric: bool = False) -> Any:
+    return a.between(b, c, symmetric=symmetric)
+
+
+@comparison_op
+@_operator_fn
+def not_between_op(a: Any, b: Any, c: Any, symmetric: bool = False) -> Any:
+    return ~a.between(b, c, symmetric=symmetric)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notbetween_op(
+        a: Any, b: Any, c: Any, symmetric: bool = False
+    ) -> Any: ...
+
+else:
+    notbetween_op = not_between_op
+
+
+@comparison_op
+@_operator_fn
+def in_op(a: Any, b: Any) -> Any:
+    return a.in_(b)
+
+
+@comparison_op
+@_operator_fn
+def not_in_op(a: Any, b: Any) -> Any:
+    return a.not_in(b)
+
+
+# 1.4 deprecated; see #5429
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notin_op(a: Any, b: Any) -> Any: ...
+
+else:
+    notin_op = not_in_op
+
+
+@_operator_fn
+def distinct_op(a: Any) -> Any:
+    return a.distinct()
+
+
+@_operator_fn
+def any_op(a: Any) -> Any:
+    return a.any_()
+
+
+@_operator_fn
+def all_op(a: Any) -> Any:
+    return a.all_()
+
+
+def _escaped_like_impl(
+    fn: Callable[..., Any], other: Any, escape: Optional[str], autoescape: bool
+) -> Any:
+    if autoescape:
+        if autoescape is not True:
+            util.warn(
+                "The autoescape parameter is now a simple boolean True/False"
+            )
+        if escape is None:
+            escape = "/"
+
+        if not isinstance(other, str):
+            raise TypeError("String value expected when autoescape=True")
+
+        if escape not in ("%", "_"):
+            other = other.replace(escape, escape + escape)
+
+        other = other.replace("%", escape + "%").replace("_", escape + "_")
+
+    return fn(other, escape=escape)
+
+
+@comparison_op
+@_operator_fn
+def startswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.startswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_startswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.startswith, b, escape, autoescape)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notstartswith_op(
+        a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+    ) -> Any: ...
+
+else:
+    notstartswith_op = not_startswith_op
+
+
+@comparison_op
+@_operator_fn
+def istartswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.istartswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_istartswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.istartswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def endswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.endswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_endswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.endswith, b, escape, autoescape)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notendswith_op(
+        a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+    ) -> Any: ...
+
+else:
+    notendswith_op = not_endswith_op
+
+
+@comparison_op
+@_operator_fn
+def iendswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.iendswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_iendswith_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.iendswith, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def contains_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.contains, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_contains_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.contains, b, escape, autoescape)
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notcontains_op(
+        a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+    ) -> Any: ...
+
+else:
+    notcontains_op = not_contains_op
+
+
+@comparison_op
+@_operator_fn
+def icontains_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return _escaped_like_impl(a.icontains, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def not_icontains_op(
+    a: Any, b: Any, escape: Optional[str] = None, autoescape: bool = False
+) -> Any:
+    return ~_escaped_like_impl(a.icontains, b, escape, autoescape)
+
+
+@comparison_op
+@_operator_fn
+def match_op(a: Any, b: Any, **kw: Any) -> Any:
+    return a.match(b, **kw)
+
+
+@comparison_op
+@_operator_fn
+def regexp_match_op(a: Any, b: Any, flags: Optional[str] = None) -> Any:
+    return a.regexp_match(b, flags=flags)
+
+
+@comparison_op
+@_operator_fn
+def not_regexp_match_op(a: Any, b: Any, flags: Optional[str] = None) -> Any:
+    return ~a.regexp_match(b, flags=flags)
+
+
+@_operator_fn
+def regexp_replace_op(
+    a: Any, b: Any, replacement: Any, flags: Optional[str] = None
+) -> Any:
+    return a.regexp_replace(b, replacement=replacement, flags=flags)
+
+
+@comparison_op
+@_operator_fn
+def not_match_op(a: Any, b: Any, **kw: Any) -> Any:
+    return ~a.match(b, **kw)
+
+
+# 1.4 deprecated; see #5429
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def notmatch_op(a: Any, b: Any, **kw: Any) -> Any: ...
+
+else:
+    notmatch_op = not_match_op
+
+
+@_operator_fn
+def comma_op(a: Any, b: Any) -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def filter_op(a: Any, b: Any) -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def concat_op(a: Any, b: Any) -> Any:
+    try:
+        concat = a.concat
+    except AttributeError:
+        return b._rconcat(a)
+    else:
+        return concat(b)
+
+
+@_operator_fn
+def desc_op(a: Any) -> Any:
+    return a.desc()
+
+
+@_operator_fn
+def asc_op(a: Any) -> Any:
+    return a.asc()
+
+
+@_operator_fn
+def nulls_first_op(a: Any) -> Any:
+    return a.nulls_first()
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def nullsfirst_op(a: Any) -> Any: ...
+
+else:
+    nullsfirst_op = nulls_first_op
+
+
+@_operator_fn
+def nulls_last_op(a: Any) -> Any:
+    return a.nulls_last()
+
+
+# 1.4 deprecated; see #5435
+if TYPE_CHECKING:
+
+    @_operator_fn
+    def nullslast_op(a: Any) -> Any: ...
+
+else:
+    nullslast_op = nulls_last_op
+
+
+@_operator_fn
+def json_getitem_op(a: Any, b: Any) -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def json_path_getitem_op(a: Any, b: Any) -> Any:
+    raise NotImplementedError()
+
+
+@_operator_fn
+def bitwise_xor_op(a: Any, b: Any) -> Any:
+    return a.bitwise_xor(b)
+
+
+@_operator_fn
+def bitwise_or_op(a: Any, b: Any) -> Any:
+    return a.bitwise_or(b)
+
+
+@_operator_fn
+def bitwise_and_op(a: Any, b: Any) -> Any:
+    return a.bitwise_and(b)
+
+
+@_operator_fn
+def bitwise_not_op(a: Any) -> Any:
+    return a.bitwise_not()
+
+
+@_operator_fn
+def bitwise_lshift_op(a: Any, b: Any) -> Any:
+    return a.bitwise_lshift(b)
+
+
+@_operator_fn
+def bitwise_rshift_op(a: Any, b: Any) -> Any:
+    return a.bitwise_rshift(b)
+
+
+def is_comparison(op: OperatorType) -> bool:
+    return op in _comparison or isinstance(op, custom_op) and op.is_comparison
+
+
+def is_commutative(op: OperatorType) -> bool:
+    return op in _commutative
+
+
+def is_ordering_modifier(op: OperatorType) -> bool:
+    return op in (asc_op, desc_op, nulls_first_op, nulls_last_op)
+
+
+def is_natural_self_precedent(op: OperatorType) -> bool:
+    return (
+        op in _natural_self_precedent
+        or isinstance(op, custom_op)
+        and op.natural_self_precedent
+    )
+
+
+_booleans = (inv, is_true, is_false, and_, or_)
+
+
+def is_boolean(op: OperatorType) -> bool:
+    return is_comparison(op) or op in _booleans
+
+
+_mirror = {gt: lt, ge: le, lt: gt, le: ge}
+
+
+def mirror(op: OperatorType) -> OperatorType:
+    """rotate a comparison operator 180 degrees.
+
+    Note this is not the same as negation.
+
+    """
+    return _mirror.get(op, op)
+
+
+_associative = _commutative.union([concat_op, and_, or_]).difference([eq, ne])
+
+
+def is_associative(op: OperatorType) -> bool:
+    return op in _associative
+
+
+def is_order_by_modifier(op: Optional[OperatorType]) -> bool:
+    return op in _order_by_modifier
+
+
+_order_by_modifier = {desc_op, asc_op, nulls_first_op, nulls_last_op}
+
+_natural_self_precedent = _associative.union(
+    [getitem, json_getitem_op, json_path_getitem_op]
+)
+"""Operators where if we have (a op b) op c, we don't want to
+parenthesize (a op b).
+
+"""
+
+
+@_operator_fn
+def _asbool(a: Any) -> Any:
+    raise NotImplementedError()
+
+
+class _OpLimit(IntEnum):
+    _smallest = -100
+    _largest = 100
+
+
+_PRECEDENCE: Dict[OperatorType, int] = {
+    from_: 15,
+    function_as_comparison_op: 15,
+    any_op: 15,
+    all_op: 15,
+    getitem: 15,
+    json_getitem_op: 15,
+    json_path_getitem_op: 15,
+    mul: 8,
+    truediv: 8,
+    floordiv: 8,
+    mod: 8,
+    neg: 8,
+    bitwise_not_op: 8,
+    add: 7,
+    sub: 7,
+    bitwise_xor_op: 7,
+    bitwise_or_op: 7,
+    bitwise_and_op: 7,
+    bitwise_lshift_op: 7,
+    bitwise_rshift_op: 7,
+    filter_op: 6,
+    concat_op: 5,
+    match_op: 5,
+    not_match_op: 5,
+    regexp_match_op: 5,
+    not_regexp_match_op: 5,
+    regexp_replace_op: 5,
+    ilike_op: 5,
+    not_ilike_op: 5,
+    like_op: 5,
+    not_like_op: 5,
+    in_op: 5,
+    not_in_op: 5,
+    is_: 5,
+    is_not: 5,
+    eq: 5,
+    ne: 5,
+    is_distinct_from: 5,
+    is_not_distinct_from: 5,
+    gt: 5,
+    lt: 5,
+    ge: 5,
+    le: 5,
+    between_op: 5,
+    not_between_op: 5,
+    distinct_op: 5,
+    inv: 5,
+    is_true: 5,
+    is_false: 5,
+    and_: 3,
+    or_: 2,
+    comma_op: -1,
+    desc_op: 3,
+    asc_op: 3,
+    collate: 4,
+    as_: -1,
+    exists: 0,
+    _asbool: -10,
+}
+
+
+def is_precedent(
+    operator: OperatorType, against: Optional[OperatorType]
+) -> bool:
+    if operator is against and is_natural_self_precedent(operator):
+        return False
+    elif against is None:
+        return True
+    else:
+        return bool(
+            _PRECEDENCE.get(
+                operator, getattr(operator, "precedence", _OpLimit._smallest)
+            )
+            <= _PRECEDENCE.get(
+                against, getattr(against, "precedence", _OpLimit._largest)
+            )
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/roles.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/roles.py
new file mode 100644
index 00000000..da69616d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/roles.py
@@ -0,0 +1,323 @@
+# sql/roles.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from __future__ import annotations
+
+from typing import Any
+from typing import Generic
+from typing import Optional
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+from .. import util
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from ._typing import _PropagateAttrsType
+    from .elements import Label
+    from .selectable import _SelectIterable
+    from .selectable import FromClause
+    from .selectable import Subquery
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+
+
+class SQLRole:
+    """Define a "role" within a SQL statement structure.
+
+    Classes within SQL Core participate within SQLRole hierarchies in order
+    to more accurately indicate where they may be used within SQL statements
+    of all types.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __slots__ = ()
+    allows_lambda = False
+    uses_inspection = False
+
+
+class UsesInspection:
+    __slots__ = ()
+    _post_inspect: Literal[None] = None
+    uses_inspection = True
+
+
+class AllowsLambdaRole:
+    __slots__ = ()
+    allows_lambda = True
+
+
+class HasCacheKeyRole(SQLRole):
+    __slots__ = ()
+    _role_name = "Cacheable Core or ORM object"
+
+
+class ExecutableOptionRole(SQLRole):
+    __slots__ = ()
+    _role_name = "ExecutionOption Core or ORM object"
+
+
+class LiteralValueRole(SQLRole):
+    __slots__ = ()
+    _role_name = "Literal Python value"
+
+
+class ColumnArgumentRole(SQLRole):
+    __slots__ = ()
+    _role_name = "Column expression"
+
+
+class ColumnArgumentOrKeyRole(ColumnArgumentRole):
+    __slots__ = ()
+    _role_name = "Column expression or string key"
+
+
+class StrAsPlainColumnRole(ColumnArgumentRole):
+    __slots__ = ()
+    _role_name = "Column expression or string key"
+
+
+class ColumnListRole(SQLRole):
+    """Elements suitable for forming comma separated lists of expressions."""
+
+    __slots__ = ()
+
+
+class StringRole(SQLRole):
+    """mixin indicating a role that results in strings"""
+
+    __slots__ = ()
+
+
+class TruncatedLabelRole(StringRole, SQLRole):
+    __slots__ = ()
+    _role_name = "String SQL identifier"
+
+
+class ColumnsClauseRole(AllowsLambdaRole, UsesInspection, ColumnListRole):
+    __slots__ = ()
+    _role_name = (
+        "Column expression, FROM clause, or other columns clause element"
+    )
+
+    @property
+    def _select_iterable(self) -> _SelectIterable:
+        raise NotImplementedError()
+
+
+class TypedColumnsClauseRole(Generic[_T_co], SQLRole):
+    """element-typed form of ColumnsClauseRole"""
+
+    __slots__ = ()
+
+
+class LimitOffsetRole(SQLRole):
+    __slots__ = ()
+    _role_name = "LIMIT / OFFSET expression"
+
+
+class ByOfRole(ColumnListRole):
+    __slots__ = ()
+    _role_name = "GROUP BY / OF / etc. expression"
+
+
+class GroupByRole(AllowsLambdaRole, UsesInspection, ByOfRole):
+    __slots__ = ()
+    # note there's a special case right now where you can pass a whole
+    # ORM entity to group_by() and it splits out.   we may not want to keep
+    # this around
+
+    _role_name = "GROUP BY expression"
+
+
+class OrderByRole(AllowsLambdaRole, ByOfRole):
+    __slots__ = ()
+    _role_name = "ORDER BY expression"
+
+
+class StructuralRole(SQLRole):
+    __slots__ = ()
+
+
+class StatementOptionRole(StructuralRole):
+    __slots__ = ()
+    _role_name = "statement sub-expression element"
+
+
+class OnClauseRole(AllowsLambdaRole, StructuralRole):
+    __slots__ = ()
+    _role_name = (
+        "ON clause, typically a SQL expression or "
+        "ORM relationship attribute"
+    )
+
+
+class WhereHavingRole(OnClauseRole):
+    __slots__ = ()
+    _role_name = "SQL expression for WHERE/HAVING role"
+
+
+class ExpressionElementRole(TypedColumnsClauseRole[_T_co]):
+    # note when using generics for ExpressionElementRole,
+    # the generic type needs to be in
+    # sqlalchemy.sql.coercions._impl_lookup mapping also.
+    # these are set up for basic types like int, bool, str, float
+    # right now
+
+    __slots__ = ()
+    _role_name = "SQL expression element"
+
+    def label(self, name: Optional[str]) -> Label[_T]:
+        raise NotImplementedError()
+
+
+class ConstExprRole(ExpressionElementRole[_T]):
+    __slots__ = ()
+    _role_name = "Constant True/False/None expression"
+
+
+class LabeledColumnExprRole(ExpressionElementRole[_T]):
+    __slots__ = ()
+
+
+class BinaryElementRole(ExpressionElementRole[_T]):
+    __slots__ = ()
+    _role_name = "SQL expression element or literal value"
+
+
+class InElementRole(SQLRole):
+    __slots__ = ()
+    _role_name = (
+        "IN expression list, SELECT construct, or bound parameter object"
+    )
+
+
+class JoinTargetRole(AllowsLambdaRole, UsesInspection, StructuralRole):
+    __slots__ = ()
+    _role_name = (
+        "Join target, typically a FROM expression, or ORM "
+        "relationship attribute"
+    )
+
+
+class FromClauseRole(ColumnsClauseRole, JoinTargetRole):
+    __slots__ = ()
+    _role_name = "FROM expression, such as a Table or alias() object"
+
+    _is_subquery = False
+
+    named_with_column: bool
+
+
+class StrictFromClauseRole(FromClauseRole):
+    __slots__ = ()
+    # does not allow text() or select() objects
+
+
+class AnonymizedFromClauseRole(StrictFromClauseRole):
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        def _anonymous_fromclause(
+            self, *, name: Optional[str] = None, flat: bool = False
+        ) -> FromClause: ...
+
+
+class ReturnsRowsRole(SQLRole):
+    __slots__ = ()
+    _role_name = (
+        "Row returning expression such as a SELECT, a FROM clause, or an "
+        "INSERT/UPDATE/DELETE with RETURNING"
+    )
+
+
+class StatementRole(SQLRole):
+    __slots__ = ()
+    _role_name = "Executable SQL or text() construct"
+
+    if TYPE_CHECKING:
+
+        @util.memoized_property
+        def _propagate_attrs(self) -> _PropagateAttrsType: ...
+
+    else:
+        _propagate_attrs = util.EMPTY_DICT
+
+
+class SelectStatementRole(StatementRole, ReturnsRowsRole):
+    __slots__ = ()
+    _role_name = "SELECT construct or equivalent text() construct"
+
+    def subquery(self) -> Subquery:
+        raise NotImplementedError(
+            "All SelectStatementRole objects should implement a "
+            ".subquery() method."
+        )
+
+
+class HasCTERole(ReturnsRowsRole):
+    __slots__ = ()
+
+
+class IsCTERole(SQLRole):
+    __slots__ = ()
+    _role_name = "CTE object"
+
+
+class CompoundElementRole(AllowsLambdaRole, SQLRole):
+    """SELECT statements inside a CompoundSelect, e.g. UNION, EXTRACT, etc."""
+
+    __slots__ = ()
+    _role_name = (
+        "SELECT construct for inclusion in a UNION or other set construct"
+    )
+
+
+# TODO: are we using this?
+class DMLRole(StatementRole):
+    __slots__ = ()
+
+
+class DMLTableRole(FromClauseRole):
+    __slots__ = ()
+    _role_name = "subject table for an INSERT, UPDATE or DELETE"
+
+
+class DMLColumnRole(SQLRole):
+    __slots__ = ()
+    _role_name = "SET/VALUES column expression or string key"
+
+
+class DMLSelectRole(SQLRole):
+    """A SELECT statement embedded in DML, typically INSERT from SELECT"""
+
+    __slots__ = ()
+    _role_name = "SELECT statement or equivalent textual object"
+
+
+class DDLRole(StatementRole):
+    __slots__ = ()
+
+
+class DDLExpressionRole(StructuralRole):
+    __slots__ = ()
+    _role_name = "SQL expression element for DDL constraint"
+
+
+class DDLConstraintColumnRole(SQLRole):
+    __slots__ = ()
+    _role_name = "String column name or column expression for DDL constraint"
+
+
+class DDLReferredColumnRole(DDLConstraintColumnRole):
+    __slots__ = ()
+    _role_name = (
+        "String column name or Column object for DDL foreign key constraint"
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/schema.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/schema.py
new file mode 100644
index 00000000..8daa8864
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/schema.py
@@ -0,0 +1,6201 @@
+# sql/schema.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""The schema module provides the building blocks for database metadata.
+
+Each element within this module describes a database entity which can be
+created and dropped, or is otherwise part of such an entity.  Examples include
+tables, columns, sequences, and indexes.
+
+All entities are subclasses of :class:`~sqlalchemy.schema.SchemaItem`, and as
+defined in this module they are intended to be agnostic of any vendor-specific
+constructs.
+
+A collection of entities are grouped into a unit called
+:class:`~sqlalchemy.schema.MetaData`. MetaData serves as a logical grouping of
+schema elements, and can also be associated with an actual database connection
+such that operations involving the contained elements can contact the database
+as needed.
+
+Two of the elements here also build upon their "syntactic" counterparts, which
+are defined in :class:`~sqlalchemy.sql.expression.`, specifically
+:class:`~sqlalchemy.schema.Table` and :class:`~sqlalchemy.schema.Column`.
+Since these objects are part of the SQL expression language, they are usable
+as components in SQL expressions.
+
+"""
+from __future__ import annotations
+
+from abc import ABC
+import collections
+from enum import Enum
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence as _typing_Sequence
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import ddl
+from . import roles
+from . import type_api
+from . import visitors
+from .base import _DefaultDescriptionTuple
+from .base import _NoArg
+from .base import _NoneName
+from .base import _SentinelColumnCharacterization
+from .base import _SentinelDefaultCharacterization
+from .base import DedupeColumnCollection
+from .base import DialectKWArgs
+from .base import Executable
+from .base import SchemaEventTarget as SchemaEventTarget
+from .coercions import _document_text_coercion
+from .elements import ClauseElement
+from .elements import ColumnClause
+from .elements import ColumnElement
+from .elements import quoted_name
+from .elements import TextClause
+from .selectable import TableClause
+from .type_api import to_instance
+from .visitors import ExternallyTraversible
+from .. import event
+from .. import exc
+from .. import inspection
+from .. import util
+from ..util import HasMemoized
+from ..util.typing import Final
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import Self
+from ..util.typing import TypedDict
+from ..util.typing import TypeGuard
+
+if typing.TYPE_CHECKING:
+    from ._typing import _AutoIncrementType
+    from ._typing import _DDLColumnArgument
+    from ._typing import _InfoType
+    from ._typing import _TextCoercedExpressionArgument
+    from ._typing import _TypeEngineArgument
+    from .base import ColumnSet
+    from .base import ReadOnlyColumnCollection
+    from .compiler import DDLCompiler
+    from .elements import BindParameter
+    from .elements import KeyedColumnElement
+    from .functions import Function
+    from .type_api import TypeEngine
+    from .visitors import anon_map
+    from ..engine import Connection
+    from ..engine import Engine
+    from ..engine.interfaces import _CoreMultiExecuteParams
+    from ..engine.interfaces import CoreExecuteOptionsParameter
+    from ..engine.interfaces import ExecutionContext
+    from ..engine.mock import MockConnection
+    from ..engine.reflection import _ReflectionInfo
+    from ..sql.selectable import FromClause
+
+_T = TypeVar("_T", bound="Any")
+_SI = TypeVar("_SI", bound="SchemaItem")
+_TAB = TypeVar("_TAB", bound="Table")
+
+
+_CreateDropBind = Union["Engine", "Connection", "MockConnection"]
+
+_ConstraintNameArgument = Optional[Union[str, _NoneName]]
+
+_ServerDefaultArgument = Union[
+    "FetchedValue", str, TextClause, ColumnElement[Any]
+]
+
+_ServerOnUpdateArgument = _ServerDefaultArgument
+
+
+class SchemaConst(Enum):
+    RETAIN_SCHEMA = 1
+    """Symbol indicating that a :class:`_schema.Table`, :class:`.Sequence`
+    or in some cases a :class:`_schema.ForeignKey` object, in situations
+    where the object is being copied for a :meth:`.Table.to_metadata`
+    operation, should retain the schema name that it already has.
+
+    """
+
+    BLANK_SCHEMA = 2
+    """Symbol indicating that a :class:`_schema.Table` or :class:`.Sequence`
+    should have 'None' for its schema, even if the parent
+    :class:`_schema.MetaData` has specified a schema.
+
+    .. seealso::
+
+        :paramref:`_schema.MetaData.schema`
+
+        :paramref:`_schema.Table.schema`
+
+        :paramref:`.Sequence.schema`
+
+    """
+
+    NULL_UNSPECIFIED = 3
+    """Symbol indicating the "nullable" keyword was not passed to a Column.
+
+    This is used to distinguish between the use case of passing
+    ``nullable=None`` to a :class:`.Column`, which has special meaning
+    on some backends such as SQL Server.
+
+    """
+
+
+RETAIN_SCHEMA: Final[Literal[SchemaConst.RETAIN_SCHEMA]] = (
+    SchemaConst.RETAIN_SCHEMA
+)
+BLANK_SCHEMA: Final[Literal[SchemaConst.BLANK_SCHEMA]] = (
+    SchemaConst.BLANK_SCHEMA
+)
+NULL_UNSPECIFIED: Final[Literal[SchemaConst.NULL_UNSPECIFIED]] = (
+    SchemaConst.NULL_UNSPECIFIED
+)
+
+
+def _get_table_key(name: str, schema: Optional[str]) -> str:
+    if schema is None:
+        return name
+    else:
+        return schema + "." + name
+
+
+# this should really be in sql/util.py but we'd have to
+# break an import cycle
+def _copy_expression(
+    expression: ColumnElement[Any],
+    source_table: Optional[Table],
+    target_table: Optional[Table],
+) -> ColumnElement[Any]:
+    if source_table is None or target_table is None:
+        return expression
+
+    fixed_source_table = source_table
+    fixed_target_table = target_table
+
+    def replace(
+        element: ExternallyTraversible, **kw: Any
+    ) -> Optional[ExternallyTraversible]:
+        if (
+            isinstance(element, Column)
+            and element.table is fixed_source_table
+            and element.key in fixed_source_table.c
+        ):
+            return fixed_target_table.c[element.key]
+        else:
+            return None
+
+    return cast(
+        ColumnElement[Any],
+        visitors.replacement_traverse(expression, {}, replace),
+    )
+
+
+@inspection._self_inspects
+class SchemaItem(SchemaEventTarget, visitors.Visitable):
+    """Base class for items that define a database schema."""
+
+    __visit_name__ = "schema_item"
+
+    create_drop_stringify_dialect = "default"
+
+    def _init_items(self, *args: SchemaItem, **kw: Any) -> None:
+        """Initialize the list of child items for this SchemaItem."""
+        for item in args:
+            if item is not None:
+                try:
+                    spwd = item._set_parent_with_dispatch
+                except AttributeError as err:
+                    raise exc.ArgumentError(
+                        "'SchemaItem' object, such as a 'Column' or a "
+                        f"'Constraint' expected, got {item!r}"
+                    ) from err
+                else:
+                    spwd(self, **kw)
+
+    def __repr__(self) -> str:
+        return util.generic_repr(self, omit_kwarg=["info"])
+
+    @util.memoized_property
+    def info(self) -> _InfoType:
+        """Info dictionary associated with the object, allowing user-defined
+        data to be associated with this :class:`.SchemaItem`.
+
+        The dictionary is automatically generated when first accessed.
+        It can also be specified in the constructor of some objects,
+        such as :class:`_schema.Table` and :class:`_schema.Column`.
+
+        """
+        return {}
+
+    def _schema_item_copy(self, schema_item: _SI) -> _SI:
+        if "info" in self.__dict__:
+            schema_item.info = self.info.copy()
+        schema_item.dispatch._update(self.dispatch)
+        return schema_item
+
+    _use_schema_map = True
+
+
+class HasConditionalDDL:
+    """define a class that includes the :meth:`.HasConditionalDDL.ddl_if`
+    method, allowing for conditional rendering of DDL.
+
+    Currently applies to constraints and indexes.
+
+    .. versionadded:: 2.0
+
+
+    """
+
+    _ddl_if: Optional[ddl.DDLIf] = None
+
+    def ddl_if(
+        self,
+        dialect: Optional[str] = None,
+        callable_: Optional[ddl.DDLIfCallable] = None,
+        state: Optional[Any] = None,
+    ) -> Self:
+        r"""apply a conditional DDL rule to this schema item.
+
+        These rules work in a similar manner to the
+        :meth:`.ExecutableDDLElement.execute_if` callable, with the added
+        feature that the criteria may be checked within the DDL compilation
+        phase for a construct such as :class:`.CreateTable`.
+        :meth:`.HasConditionalDDL.ddl_if` currently applies towards the
+        :class:`.Index` construct as well as all :class:`.Constraint`
+        constructs.
+
+        :param dialect: string name of a dialect, or a tuple of string names
+         to indicate multiple dialect types.
+
+        :param callable\_: a callable that is constructed using the same form
+         as that described in
+         :paramref:`.ExecutableDDLElement.execute_if.callable_`.
+
+        :param state: any arbitrary object that will be passed to the
+         callable, if present.
+
+        .. versionadded:: 2.0
+
+        .. seealso::
+
+            :ref:`schema_ddl_ddl_if` - background and usage examples
+
+
+        """
+        self._ddl_if = ddl.DDLIf(dialect, callable_, state)
+        return self
+
+
+class HasSchemaAttr(SchemaItem):
+    """schema item that includes a top-level schema name"""
+
+    schema: Optional[str]
+
+
+class Table(
+    DialectKWArgs, HasSchemaAttr, TableClause, inspection.Inspectable["Table"]
+):
+    r"""Represent a table in a database.
+
+    e.g.::
+
+        mytable = Table(
+            "mytable",
+            metadata,
+            Column("mytable_id", Integer, primary_key=True),
+            Column("value", String(50)),
+        )
+
+    The :class:`_schema.Table`
+    object constructs a unique instance of itself based
+    on its name and optional schema name within the given
+    :class:`_schema.MetaData` object. Calling the :class:`_schema.Table`
+    constructor with the same name and same :class:`_schema.MetaData` argument
+    a second time will return the *same* :class:`_schema.Table`
+    object - in this way
+    the :class:`_schema.Table` constructor acts as a registry function.
+
+    .. seealso::
+
+        :ref:`metadata_describing` - Introduction to database metadata
+
+    """
+
+    __visit_name__ = "table"
+
+    if TYPE_CHECKING:
+
+        @util.ro_non_memoized_property
+        def primary_key(self) -> PrimaryKeyConstraint: ...
+
+        @util.ro_non_memoized_property
+        def foreign_keys(self) -> Set[ForeignKey]: ...
+
+    _columns: DedupeColumnCollection[Column[Any]]
+
+    _sentinel_column: Optional[Column[Any]]
+
+    constraints: Set[Constraint]
+    """A collection of all :class:`_schema.Constraint` objects associated with
+      this :class:`_schema.Table`.
+
+      Includes :class:`_schema.PrimaryKeyConstraint`,
+      :class:`_schema.ForeignKeyConstraint`, :class:`_schema.UniqueConstraint`,
+      :class:`_schema.CheckConstraint`.  A separate collection
+      :attr:`_schema.Table.foreign_key_constraints` refers to the collection
+      of all :class:`_schema.ForeignKeyConstraint` objects, and the
+      :attr:`_schema.Table.primary_key` attribute refers to the single
+      :class:`_schema.PrimaryKeyConstraint` associated with the
+      :class:`_schema.Table`.
+
+      .. seealso::
+
+            :attr:`_schema.Table.constraints`
+
+            :attr:`_schema.Table.primary_key`
+
+            :attr:`_schema.Table.foreign_key_constraints`
+
+            :attr:`_schema.Table.indexes`
+
+            :class:`_reflection.Inspector`
+
+
+    """
+
+    indexes: Set[Index]
+    """A collection of all :class:`_schema.Index` objects associated with this
+      :class:`_schema.Table`.
+
+      .. seealso::
+
+            :meth:`_reflection.Inspector.get_indexes`
+
+    """
+
+    if TYPE_CHECKING:
+
+        @util.ro_non_memoized_property
+        def columns(self) -> ReadOnlyColumnCollection[str, Column[Any]]: ...
+
+        @util.ro_non_memoized_property
+        def exported_columns(
+            self,
+        ) -> ReadOnlyColumnCollection[str, Column[Any]]: ...
+
+        @util.ro_non_memoized_property
+        def c(self) -> ReadOnlyColumnCollection[str, Column[Any]]: ...
+
+    def _gen_cache_key(
+        self, anon_map: anon_map, bindparams: List[BindParameter[Any]]
+    ) -> Tuple[Any, ...]:
+        if self._annotations:
+            return (self,) + self._annotations_cache_key
+        else:
+            return (self,)
+
+    if not typing.TYPE_CHECKING:
+        # typing tools seem to be inconsistent in how they handle
+        # __new__, so suggest this pattern for classes that use
+        # __new__.  apply typing to the __init__ method normally
+        @util.deprecated_params(
+            mustexist=(
+                "1.4",
+                "Deprecated alias of :paramref:`_schema.Table.must_exist`",
+            ),
+        )
+        def __new__(cls, *args: Any, **kw: Any) -> Any:
+            return cls._new(*args, **kw)
+
+    @classmethod
+    def _new(cls, *args: Any, **kw: Any) -> Any:
+        if not args and not kw:
+            # python3k pickle seems to call this
+            return object.__new__(cls)
+
+        try:
+            name, metadata, args = args[0], args[1], args[2:]
+        except IndexError:
+            raise TypeError(
+                "Table() takes at least two positional-only "
+                "arguments 'name' and 'metadata'"
+            )
+
+        schema = kw.get("schema", None)
+        if schema is None:
+            schema = metadata.schema
+        elif schema is BLANK_SCHEMA:
+            schema = None
+        keep_existing = kw.get("keep_existing", False)
+        extend_existing = kw.get("extend_existing", False)
+
+        if keep_existing and extend_existing:
+            msg = "keep_existing and extend_existing are mutually exclusive."
+            raise exc.ArgumentError(msg)
+
+        must_exist = kw.pop("must_exist", kw.pop("mustexist", False))
+        key = _get_table_key(name, schema)
+        if key in metadata.tables:
+            if not keep_existing and not extend_existing and bool(args):
+                raise exc.InvalidRequestError(
+                    f"Table '{key}' is already defined for this MetaData "
+                    "instance.  Specify 'extend_existing=True' "
+                    "to redefine "
+                    "options and columns on an "
+                    "existing Table object."
+                )
+            table = metadata.tables[key]
+            if extend_existing:
+                table._init_existing(*args, **kw)
+            return table
+        else:
+            if must_exist:
+                raise exc.InvalidRequestError(f"Table '{key}' not defined")
+            table = object.__new__(cls)
+            table.dispatch.before_parent_attach(table, metadata)
+            metadata._add_table(name, schema, table)
+            try:
+                table.__init__(name, metadata, *args, _no_init=False, **kw)
+                table.dispatch.after_parent_attach(table, metadata)
+                return table
+            except Exception:
+                with util.safe_reraise():
+                    metadata._remove_table(name, schema)
+
+    def __init__(
+        self,
+        name: str,
+        metadata: MetaData,
+        *args: SchemaItem,
+        schema: Optional[Union[str, Literal[SchemaConst.BLANK_SCHEMA]]] = None,
+        quote: Optional[bool] = None,
+        quote_schema: Optional[bool] = None,
+        autoload_with: Optional[Union[Engine, Connection]] = None,
+        autoload_replace: bool = True,
+        keep_existing: bool = False,
+        extend_existing: bool = False,
+        resolve_fks: bool = True,
+        include_columns: Optional[Collection[str]] = None,
+        implicit_returning: bool = True,
+        comment: Optional[str] = None,
+        info: Optional[Dict[Any, Any]] = None,
+        listeners: Optional[
+            _typing_Sequence[Tuple[str, Callable[..., Any]]]
+        ] = None,
+        prefixes: Optional[_typing_Sequence[str]] = None,
+        # used internally in the metadata.reflect() process
+        _extend_on: Optional[Set[Table]] = None,
+        # used by __new__ to bypass __init__
+        _no_init: bool = True,
+        # dialect-specific keyword args
+        **kw: Any,
+    ) -> None:
+        r"""Constructor for :class:`_schema.Table`.
+
+
+        :param name: The name of this table as represented in the database.
+
+            The table name, along with the value of the ``schema`` parameter,
+            forms a key which uniquely identifies this :class:`_schema.Table`
+            within
+            the owning :class:`_schema.MetaData` collection.
+            Additional calls to :class:`_schema.Table` with the same name,
+            metadata,
+            and schema name will return the same :class:`_schema.Table` object.
+
+            Names which contain no upper case characters
+            will be treated as case insensitive names, and will not be quoted
+            unless they are a reserved word or contain special characters.
+            A name with any number of upper case characters is considered
+            to be case sensitive, and will be sent as quoted.
+
+            To enable unconditional quoting for the table name, specify the flag
+            ``quote=True`` to the constructor, or use the :class:`.quoted_name`
+            construct to specify the name.
+
+        :param metadata: a :class:`_schema.MetaData`
+            object which will contain this
+            table.  The metadata is used as a point of association of this table
+            with other tables which are referenced via foreign key.  It also
+            may be used to associate this table with a particular
+            :class:`.Connection` or :class:`.Engine`.
+
+        :param \*args: Additional positional arguments are used primarily
+            to add the list of :class:`_schema.Column`
+            objects contained within this
+            table. Similar to the style of a CREATE TABLE statement, other
+            :class:`.SchemaItem` constructs may be added here, including
+            :class:`.PrimaryKeyConstraint`, and
+            :class:`_schema.ForeignKeyConstraint`.
+
+        :param autoload_replace: Defaults to ``True``; when using
+            :paramref:`_schema.Table.autoload_with`
+            in conjunction with :paramref:`_schema.Table.extend_existing`,
+            indicates
+            that :class:`_schema.Column` objects present in the already-existing
+            :class:`_schema.Table`
+            object should be replaced with columns of the same
+            name retrieved from the autoload process.   When ``False``, columns
+            already present under existing names will be omitted from the
+            reflection process.
+
+            Note that this setting does not impact :class:`_schema.Column` objects
+            specified programmatically within the call to :class:`_schema.Table`
+            that
+            also is autoloading; those :class:`_schema.Column` objects will always
+            replace existing columns of the same name when
+            :paramref:`_schema.Table.extend_existing` is ``True``.
+
+            .. seealso::
+
+                :paramref:`_schema.Table.autoload_with`
+
+                :paramref:`_schema.Table.extend_existing`
+
+        :param autoload_with: An :class:`_engine.Engine` or
+            :class:`_engine.Connection` object,
+            or a :class:`_reflection.Inspector` object as returned by
+            :func:`_sa.inspect`
+            against one, with which this :class:`_schema.Table`
+            object will be reflected.
+            When set to a non-None value, the autoload process will take place
+            for this table against the given engine or connection.
+
+            .. seealso::
+
+                :ref:`metadata_reflection_toplevel`
+
+                :meth:`_events.DDLEvents.column_reflect`
+
+                :ref:`metadata_reflection_dbagnostic_types`
+
+        :param extend_existing: When ``True``, indicates that if this
+            :class:`_schema.Table` is already present in the given
+            :class:`_schema.MetaData`,
+            apply further arguments within the constructor to the existing
+            :class:`_schema.Table`.
+
+            If :paramref:`_schema.Table.extend_existing` or
+            :paramref:`_schema.Table.keep_existing` are not set,
+            and the given name
+            of the new :class:`_schema.Table` refers to a :class:`_schema.Table`
+            that is
+            already present in the target :class:`_schema.MetaData` collection,
+            and
+            this :class:`_schema.Table`
+            specifies additional columns or other constructs
+            or flags that modify the table's state, an
+            error is raised.  The purpose of these two mutually-exclusive flags
+            is to specify what action should be taken when a
+            :class:`_schema.Table`
+            is specified that matches an existing :class:`_schema.Table`,
+            yet specifies
+            additional constructs.
+
+            :paramref:`_schema.Table.extend_existing`
+            will also work in conjunction
+            with :paramref:`_schema.Table.autoload_with` to run a new reflection
+            operation against the database, even if a :class:`_schema.Table`
+            of the same name is already present in the target
+            :class:`_schema.MetaData`; newly reflected :class:`_schema.Column`
+            objects
+            and other options will be added into the state of the
+            :class:`_schema.Table`, potentially overwriting existing columns
+            and options of the same name.
+
+            As is always the case with :paramref:`_schema.Table.autoload_with`,
+            :class:`_schema.Column` objects can be specified in the same
+            :class:`_schema.Table`
+            constructor, which will take precedence.  Below, the existing
+            table ``mytable`` will be augmented with :class:`_schema.Column`
+            objects
+            both reflected from the database, as well as the given
+            :class:`_schema.Column`
+            named "y"::
+
+                Table(
+                    "mytable",
+                    metadata,
+                    Column("y", Integer),
+                    extend_existing=True,
+                    autoload_with=engine,
+                )
+
+            .. seealso::
+
+                :paramref:`_schema.Table.autoload_with`
+
+                :paramref:`_schema.Table.autoload_replace`
+
+                :paramref:`_schema.Table.keep_existing`
+
+
+        :param implicit_returning: True by default - indicates that
+            RETURNING can be used, typically by the ORM, in order to fetch
+            server-generated values such as primary key values and
+            server side defaults, on those backends which support RETURNING.
+
+            In modern SQLAlchemy there is generally no reason to alter this
+            setting, except for some backend specific cases
+            (see :ref:`mssql_triggers` in the SQL Server dialect documentation
+            for one such example).
+
+        :param include_columns: A list of strings indicating a subset of
+            columns to be loaded via the ``autoload`` operation; table columns who
+            aren't present in this list will not be represented on the resulting
+            ``Table`` object. Defaults to ``None`` which indicates all columns
+            should be reflected.
+
+        :param resolve_fks: Whether or not to reflect :class:`_schema.Table`
+            objects
+            related to this one via :class:`_schema.ForeignKey` objects, when
+            :paramref:`_schema.Table.autoload_with` is
+            specified.   Defaults to True.  Set to False to disable reflection of
+            related tables as :class:`_schema.ForeignKey`
+            objects are encountered; may be
+            used either to save on SQL calls or to avoid issues with related tables
+            that can't be accessed. Note that if a related table is already present
+            in the :class:`_schema.MetaData` collection, or becomes present later,
+            a
+            :class:`_schema.ForeignKey` object associated with this
+            :class:`_schema.Table` will
+            resolve to that table normally.
+
+            .. versionadded:: 1.3
+
+            .. seealso::
+
+                :paramref:`.MetaData.reflect.resolve_fks`
+
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param keep_existing: When ``True``, indicates that if this Table
+            is already present in the given :class:`_schema.MetaData`, ignore
+            further arguments within the constructor to the existing
+            :class:`_schema.Table`, and return the :class:`_schema.Table`
+            object as
+            originally created. This is to allow a function that wishes
+            to define a new :class:`_schema.Table` on first call, but on
+            subsequent calls will return the same :class:`_schema.Table`,
+            without any of the declarations (particularly constraints)
+            being applied a second time.
+
+            If :paramref:`_schema.Table.extend_existing` or
+            :paramref:`_schema.Table.keep_existing` are not set,
+            and the given name
+            of the new :class:`_schema.Table` refers to a :class:`_schema.Table`
+            that is
+            already present in the target :class:`_schema.MetaData` collection,
+            and
+            this :class:`_schema.Table`
+            specifies additional columns or other constructs
+            or flags that modify the table's state, an
+            error is raised.  The purpose of these two mutually-exclusive flags
+            is to specify what action should be taken when a
+            :class:`_schema.Table`
+            is specified that matches an existing :class:`_schema.Table`,
+            yet specifies
+            additional constructs.
+
+            .. seealso::
+
+                :paramref:`_schema.Table.extend_existing`
+
+        :param listeners: A list of tuples of the form ``(<eventname>, <fn>)``
+            which will be passed to :func:`.event.listen` upon construction.
+            This alternate hook to :func:`.event.listen` allows the establishment
+            of a listener function specific to this :class:`_schema.Table` before
+            the "autoload" process begins.  Historically this has been intended
+            for use with the :meth:`.DDLEvents.column_reflect` event, however
+            note that this event hook may now be associated with the
+            :class:`_schema.MetaData` object directly::
+
+                def listen_for_reflect(table, column_info):
+                    "handle the column reflection event"
+                    # ...
+
+
+                t = Table(
+                    "sometable",
+                    autoload_with=engine,
+                    listeners=[("column_reflect", listen_for_reflect)],
+                )
+
+            .. seealso::
+
+                :meth:`_events.DDLEvents.column_reflect`
+
+        :param must_exist: When ``True``, indicates that this Table must already
+            be present in the given :class:`_schema.MetaData` collection, else
+            an exception is raised.
+
+        :param prefixes:
+            A list of strings to insert after CREATE in the CREATE TABLE
+            statement.  They will be separated by spaces.
+
+        :param quote: Force quoting of this table's name on or off, corresponding
+            to ``True`` or ``False``.  When left at its default of ``None``,
+            the column identifier will be quoted according to whether the name is
+            case sensitive (identifiers with at least one upper case character are
+            treated as case sensitive), or if it's a reserved word.  This flag
+            is only needed to force quoting of a reserved word which is not known
+            by the SQLAlchemy dialect.
+
+            .. note:: setting this flag to ``False`` will not provide
+              case-insensitive behavior for table reflection; table reflection
+              will always search for a mixed-case name in a case sensitive
+              fashion.  Case insensitive names are specified in SQLAlchemy only
+              by stating the name with all lower case characters.
+
+        :param quote_schema: same as 'quote' but applies to the schema identifier.
+
+        :param schema: The schema name for this table, which is required if
+            the table resides in a schema other than the default selected schema
+            for the engine's database connection.  Defaults to ``None``.
+
+            If the owning :class:`_schema.MetaData` of this :class:`_schema.Table`
+            specifies its
+            own :paramref:`_schema.MetaData.schema` parameter,
+            then that schema name will
+            be applied to this :class:`_schema.Table`
+            if the schema parameter here is set
+            to ``None``.  To set a blank schema name on a :class:`_schema.Table`
+            that
+            would otherwise use the schema set on the owning
+            :class:`_schema.MetaData`,
+            specify the special symbol :attr:`.BLANK_SCHEMA`.
+
+            The quoting rules for the schema name are the same as those for the
+            ``name`` parameter, in that quoting is applied for reserved words or
+            case-sensitive names; to enable unconditional quoting for the schema
+            name, specify the flag ``quote_schema=True`` to the constructor, or use
+            the :class:`.quoted_name` construct to specify the name.
+
+        :param comment: Optional string that will render an SQL comment on table
+            creation.
+
+            .. versionadded:: 1.2 Added the :paramref:`_schema.Table.comment`
+                parameter
+                to :class:`_schema.Table`.
+
+        :param \**kw: Additional keyword arguments not mentioned above are
+            dialect specific, and passed in the form ``<dialectname>_<argname>``.
+            See the documentation regarding an individual dialect at
+            :ref:`dialect_toplevel` for detail on documented arguments.
+
+        """  # noqa: E501
+        if _no_init:
+            # don't run __init__ from __new__ by default;
+            # __new__ has a specific place that __init__ is called
+            return
+
+        super().__init__(quoted_name(name, quote))
+        self.metadata = metadata
+
+        if schema is None:
+            self.schema = metadata.schema
+        elif schema is BLANK_SCHEMA:
+            self.schema = None
+        else:
+            quote_schema = quote_schema
+            assert isinstance(schema, str)
+            self.schema = quoted_name(schema, quote_schema)
+
+        self._sentinel_column = None
+
+        self.indexes = set()
+        self.constraints = set()
+        PrimaryKeyConstraint(
+            _implicit_generated=True
+        )._set_parent_with_dispatch(self)
+        self.foreign_keys = set()  # type: ignore
+        self._extra_dependencies: Set[Table] = set()
+        if self.schema is not None:
+            self.fullname = "%s.%s" % (self.schema, self.name)
+        else:
+            self.fullname = self.name
+
+        self.implicit_returning = implicit_returning
+        _reflect_info = kw.pop("_reflect_info", None)
+
+        self.comment = comment
+
+        if info is not None:
+            self.info = info
+
+        if listeners is not None:
+            for evt, fn in listeners:
+                event.listen(self, evt, fn)
+
+        self._prefixes = prefixes if prefixes else []
+
+        self._extra_kwargs(**kw)
+
+        # load column definitions from the database if 'autoload' is defined
+        # we do it after the table is in the singleton dictionary to support
+        # circular foreign keys
+        if autoload_with is not None:
+            self._autoload(
+                metadata,
+                autoload_with,
+                include_columns,
+                _extend_on=_extend_on,
+                _reflect_info=_reflect_info,
+                resolve_fks=resolve_fks,
+            )
+
+        # initialize all the column, etc. objects.  done after reflection to
+        # allow user-overrides
+
+        self._init_items(
+            *args,
+            allow_replacements=extend_existing
+            or keep_existing
+            or autoload_with,
+            all_names={},
+        )
+
+    def _autoload(
+        self,
+        metadata: MetaData,
+        autoload_with: Union[Engine, Connection],
+        include_columns: Optional[Collection[str]],
+        exclude_columns: Collection[str] = (),
+        resolve_fks: bool = True,
+        _extend_on: Optional[Set[Table]] = None,
+        _reflect_info: _ReflectionInfo | None = None,
+    ) -> None:
+        insp = inspection.inspect(autoload_with)
+        with insp._inspection_context() as conn_insp:
+            conn_insp.reflect_table(
+                self,
+                include_columns,
+                exclude_columns,
+                resolve_fks,
+                _extend_on=_extend_on,
+                _reflect_info=_reflect_info,
+            )
+
+    @property
+    def _sorted_constraints(self) -> List[Constraint]:
+        """Return the set of constraints as a list, sorted by creation
+        order.
+
+        """
+
+        return sorted(self.constraints, key=lambda c: c._creation_order)
+
+    @property
+    def foreign_key_constraints(self) -> Set[ForeignKeyConstraint]:
+        """:class:`_schema.ForeignKeyConstraint` objects referred to by this
+        :class:`_schema.Table`.
+
+        This list is produced from the collection of
+        :class:`_schema.ForeignKey`
+        objects currently associated.
+
+
+        .. seealso::
+
+            :attr:`_schema.Table.constraints`
+
+            :attr:`_schema.Table.foreign_keys`
+
+            :attr:`_schema.Table.indexes`
+
+        """
+        return {
+            fkc.constraint
+            for fkc in self.foreign_keys
+            if fkc.constraint is not None
+        }
+
+    def _init_existing(self, *args: Any, **kwargs: Any) -> None:
+        autoload_with = kwargs.pop("autoload_with", None)
+        autoload = kwargs.pop("autoload", autoload_with is not None)
+        autoload_replace = kwargs.pop("autoload_replace", True)
+        schema = kwargs.pop("schema", None)
+        _extend_on = kwargs.pop("_extend_on", None)
+        _reflect_info = kwargs.pop("_reflect_info", None)
+
+        # these arguments are only used with _init()
+        extend_existing = kwargs.pop("extend_existing", False)
+        keep_existing = kwargs.pop("keep_existing", False)
+
+        assert extend_existing
+        assert not keep_existing
+
+        if schema and schema != self.schema:
+            raise exc.ArgumentError(
+                f"Can't change schema of existing table "
+                f"from '{self.schema}' to '{schema}'",
+            )
+
+        include_columns = kwargs.pop("include_columns", None)
+        if include_columns is not None:
+            for c in self.c:
+                if c.name not in include_columns:
+                    self._columns.remove(c)
+
+        resolve_fks = kwargs.pop("resolve_fks", True)
+
+        for key in ("quote", "quote_schema"):
+            if key in kwargs:
+                raise exc.ArgumentError(
+                    "Can't redefine 'quote' or 'quote_schema' arguments"
+                )
+
+        # update `self` with these kwargs, if provided
+        self.comment = kwargs.pop("comment", self.comment)
+        self.implicit_returning = kwargs.pop(
+            "implicit_returning", self.implicit_returning
+        )
+        self.info = kwargs.pop("info", self.info)
+
+        exclude_columns: _typing_Sequence[str]
+
+        if autoload:
+            if not autoload_replace:
+                # don't replace columns already present.
+                # we'd like to do this for constraints also however we don't
+                # have simple de-duping for unnamed constraints.
+                exclude_columns = [c.name for c in self.c]
+            else:
+                exclude_columns = ()
+            self._autoload(
+                self.metadata,
+                autoload_with,
+                include_columns,
+                exclude_columns,
+                resolve_fks,
+                _extend_on=_extend_on,
+                _reflect_info=_reflect_info,
+            )
+
+        all_names = {c.name: c for c in self.c}
+        self._extra_kwargs(**kwargs)
+        self._init_items(*args, allow_replacements=True, all_names=all_names)
+
+    def _extra_kwargs(self, **kwargs: Any) -> None:
+        self._validate_dialect_kwargs(kwargs)
+
+    def _init_collections(self) -> None:
+        pass
+
+    def _reset_exported(self) -> None:
+        pass
+
+    @util.ro_non_memoized_property
+    def _autoincrement_column(self) -> Optional[Column[int]]:
+        return self.primary_key._autoincrement_column
+
+    @util.ro_memoized_property
+    def _sentinel_column_characteristics(
+        self,
+    ) -> _SentinelColumnCharacterization:
+        """determine a candidate column (or columns, in case of a client
+        generated composite primary key) which can be used as an
+        "insert sentinel" for an INSERT statement.
+
+        The returned structure, :class:`_SentinelColumnCharacterization`,
+        includes all the details needed by :class:`.Dialect` and
+        :class:`.SQLCompiler` to determine if these column(s) can be used
+        as an INSERT..RETURNING sentinel for a particular database
+        dialect.
+
+        .. versionadded:: 2.0.10
+
+        """
+
+        sentinel_is_explicit = False
+        sentinel_is_autoinc = False
+        the_sentinel: Optional[_typing_Sequence[Column[Any]]] = None
+
+        # see if a column was explicitly marked "insert_sentinel=True".
+        explicit_sentinel_col = self._sentinel_column
+
+        if explicit_sentinel_col is not None:
+            the_sentinel = (explicit_sentinel_col,)
+            sentinel_is_explicit = True
+
+        autoinc_col = self._autoincrement_column
+        if sentinel_is_explicit and explicit_sentinel_col is autoinc_col:
+            assert autoinc_col is not None
+            sentinel_is_autoinc = True
+        elif explicit_sentinel_col is None and autoinc_col is not None:
+            the_sentinel = (autoinc_col,)
+            sentinel_is_autoinc = True
+
+        default_characterization = _SentinelDefaultCharacterization.UNKNOWN
+
+        if the_sentinel:
+            the_sentinel_zero = the_sentinel[0]
+            if the_sentinel_zero.identity:
+                if the_sentinel_zero.identity._increment_is_negative:
+                    if sentinel_is_explicit:
+                        raise exc.InvalidRequestError(
+                            "Can't use IDENTITY default with negative "
+                            "increment as an explicit sentinel column"
+                        )
+                    else:
+                        if sentinel_is_autoinc:
+                            autoinc_col = None
+                            sentinel_is_autoinc = False
+                        the_sentinel = None
+                else:
+                    default_characterization = (
+                        _SentinelDefaultCharacterization.IDENTITY
+                    )
+            elif (
+                the_sentinel_zero.default is None
+                and the_sentinel_zero.server_default is None
+            ):
+                if the_sentinel_zero.nullable:
+                    raise exc.InvalidRequestError(
+                        f"Column {the_sentinel_zero} has been marked as a "
+                        "sentinel "
+                        "column with no default generation function; it "
+                        "at least needs to be marked nullable=False assuming "
+                        "user-populated sentinel values will be used."
+                    )
+                default_characterization = (
+                    _SentinelDefaultCharacterization.NONE
+                )
+            elif the_sentinel_zero.default is not None:
+                if the_sentinel_zero.default.is_sentinel:
+                    default_characterization = (
+                        _SentinelDefaultCharacterization.SENTINEL_DEFAULT
+                    )
+                elif default_is_sequence(the_sentinel_zero.default):
+                    if the_sentinel_zero.default._increment_is_negative:
+                        if sentinel_is_explicit:
+                            raise exc.InvalidRequestError(
+                                "Can't use SEQUENCE default with negative "
+                                "increment as an explicit sentinel column"
+                            )
+                        else:
+                            if sentinel_is_autoinc:
+                                autoinc_col = None
+                                sentinel_is_autoinc = False
+                            the_sentinel = None
+
+                    default_characterization = (
+                        _SentinelDefaultCharacterization.SEQUENCE
+                    )
+                elif the_sentinel_zero.default.is_callable:
+                    default_characterization = (
+                        _SentinelDefaultCharacterization.CLIENTSIDE
+                    )
+            elif the_sentinel_zero.server_default is not None:
+                if sentinel_is_explicit:
+                    raise exc.InvalidRequestError(
+                        f"Column {the_sentinel[0]} can't be a sentinel column "
+                        "because it uses an explicit server side default "
+                        "that's not the Identity() default."
+                    )
+
+                default_characterization = (
+                    _SentinelDefaultCharacterization.SERVERSIDE
+                )
+
+        if the_sentinel is None and self.primary_key:
+            assert autoinc_col is None
+
+            # determine for non-autoincrement pk if all elements are
+            # client side
+            for _pkc in self.primary_key:
+                if _pkc.server_default is not None or (
+                    _pkc.default and not _pkc.default.is_callable
+                ):
+                    break
+            else:
+                the_sentinel = tuple(self.primary_key)
+                default_characterization = (
+                    _SentinelDefaultCharacterization.CLIENTSIDE
+                )
+
+        return _SentinelColumnCharacterization(
+            the_sentinel,
+            sentinel_is_explicit,
+            sentinel_is_autoinc,
+            default_characterization,
+        )
+
+    @property
+    def autoincrement_column(self) -> Optional[Column[int]]:
+        """Returns the :class:`.Column` object which currently represents
+        the "auto increment" column, if any, else returns None.
+
+        This is based on the rules for :class:`.Column` as defined by the
+        :paramref:`.Column.autoincrement` parameter, which generally means the
+        column within a single integer column primary key constraint that is
+        not constrained by a foreign key.   If the table does not have such
+        a primary key constraint, then there's no "autoincrement" column.
+        A :class:`.Table` may have only one column defined as the
+        "autoincrement" column.
+
+        .. versionadded:: 2.0.4
+
+        .. seealso::
+
+            :paramref:`.Column.autoincrement`
+
+        """
+        return self._autoincrement_column
+
+    @property
+    def key(self) -> str:
+        """Return the 'key' for this :class:`_schema.Table`.
+
+        This value is used as the dictionary key within the
+        :attr:`_schema.MetaData.tables` collection.   It is typically the same
+        as that of :attr:`_schema.Table.name` for a table with no
+        :attr:`_schema.Table.schema`
+        set; otherwise it is typically of the form
+        ``schemaname.tablename``.
+
+        """
+        return _get_table_key(self.name, self.schema)
+
+    def __repr__(self) -> str:
+        return "Table(%s)" % ", ".join(
+            [repr(self.name)]
+            + [repr(self.metadata)]
+            + [repr(x) for x in self.columns]
+            + ["%s=%s" % (k, repr(getattr(self, k))) for k in ["schema"]]
+        )
+
+    def __str__(self) -> str:
+        return _get_table_key(self.description, self.schema)
+
+    def add_is_dependent_on(self, table: Table) -> None:
+        """Add a 'dependency' for this Table.
+
+        This is another Table object which must be created
+        first before this one can, or dropped after this one.
+
+        Usually, dependencies between tables are determined via
+        ForeignKey objects.   However, for other situations that
+        create dependencies outside of foreign keys (rules, inheriting),
+        this method can manually establish such a link.
+
+        """
+        self._extra_dependencies.add(table)
+
+    def append_column(
+        self, column: ColumnClause[Any], replace_existing: bool = False
+    ) -> None:
+        """Append a :class:`_schema.Column` to this :class:`_schema.Table`.
+
+        The "key" of the newly added :class:`_schema.Column`, i.e. the
+        value of its ``.key`` attribute, will then be available
+        in the ``.c`` collection of this :class:`_schema.Table`, and the
+        column definition will be included in any CREATE TABLE, SELECT,
+        UPDATE, etc. statements generated from this :class:`_schema.Table`
+        construct.
+
+        Note that this does **not** change the definition of the table
+        as it exists within any underlying database, assuming that
+        table has already been created in the database.   Relational
+        databases support the addition of columns to existing tables
+        using the SQL ALTER command, which would need to be
+        emitted for an already-existing table that doesn't contain
+        the newly added column.
+
+        :param replace_existing: When ``True``, allows replacing existing
+            columns. When ``False``, the default, an warning will be raised
+            if a column with the same ``.key`` already exists. A future
+            version of sqlalchemy will instead rise a warning.
+
+            .. versionadded:: 1.4.0
+        """
+
+        try:
+            column._set_parent_with_dispatch(
+                self,
+                allow_replacements=replace_existing,
+                all_names={c.name: c for c in self.c},
+            )
+        except exc.DuplicateColumnError as de:
+            raise exc.DuplicateColumnError(
+                f"{de.args[0]} Specify replace_existing=True to "
+                "Table.append_column() to replace an "
+                "existing column."
+            ) from de
+
+    def append_constraint(self, constraint: Union[Index, Constraint]) -> None:
+        """Append a :class:`_schema.Constraint` to this
+        :class:`_schema.Table`.
+
+        This has the effect of the constraint being included in any
+        future CREATE TABLE statement, assuming specific DDL creation
+        events have not been associated with the given
+        :class:`_schema.Constraint` object.
+
+        Note that this does **not** produce the constraint within the
+        relational database automatically, for a table that already exists
+        in the database.   To add a constraint to an
+        existing relational database table, the SQL ALTER command must
+        be used.  SQLAlchemy also provides the
+        :class:`.AddConstraint` construct which can produce this SQL when
+        invoked as an executable clause.
+
+        """
+
+        constraint._set_parent_with_dispatch(self)
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        metadata = parent
+        assert isinstance(metadata, MetaData)
+        metadata._add_table(self.name, self.schema, self)
+        self.metadata = metadata
+
+    def create(self, bind: _CreateDropBind, checkfirst: bool = False) -> None:
+        """Issue a ``CREATE`` statement for this
+        :class:`_schema.Table`, using the given
+        :class:`.Connection` or :class:`.Engine`
+        for connectivity.
+
+        .. seealso::
+
+            :meth:`_schema.MetaData.create_all`.
+
+        """
+
+        bind._run_ddl_visitor(ddl.SchemaGenerator, self, checkfirst=checkfirst)
+
+    def drop(self, bind: _CreateDropBind, checkfirst: bool = False) -> None:
+        """Issue a ``DROP`` statement for this
+        :class:`_schema.Table`, using the given
+        :class:`.Connection` or :class:`.Engine` for connectivity.
+
+        .. seealso::
+
+            :meth:`_schema.MetaData.drop_all`.
+
+        """
+        bind._run_ddl_visitor(ddl.SchemaDropper, self, checkfirst=checkfirst)
+
+    @util.deprecated(
+        "1.4",
+        ":meth:`_schema.Table.tometadata` is renamed to "
+        ":meth:`_schema.Table.to_metadata`",
+    )
+    def tometadata(
+        self,
+        metadata: MetaData,
+        schema: Union[str, Literal[SchemaConst.RETAIN_SCHEMA]] = RETAIN_SCHEMA,
+        referred_schema_fn: Optional[
+            Callable[
+                [Table, Optional[str], ForeignKeyConstraint, Optional[str]],
+                Optional[str],
+            ]
+        ] = None,
+        name: Optional[str] = None,
+    ) -> Table:
+        """Return a copy of this :class:`_schema.Table`
+        associated with a different
+        :class:`_schema.MetaData`.
+
+        See :meth:`_schema.Table.to_metadata` for a full description.
+
+        """
+        return self.to_metadata(
+            metadata,
+            schema=schema,
+            referred_schema_fn=referred_schema_fn,
+            name=name,
+        )
+
+    def to_metadata(
+        self,
+        metadata: MetaData,
+        schema: Union[str, Literal[SchemaConst.RETAIN_SCHEMA]] = RETAIN_SCHEMA,
+        referred_schema_fn: Optional[
+            Callable[
+                [Table, Optional[str], ForeignKeyConstraint, Optional[str]],
+                Optional[str],
+            ]
+        ] = None,
+        name: Optional[str] = None,
+    ) -> Table:
+        """Return a copy of this :class:`_schema.Table` associated with a
+        different :class:`_schema.MetaData`.
+
+        E.g.::
+
+            m1 = MetaData()
+
+            user = Table("user", m1, Column("id", Integer, primary_key=True))
+
+            m2 = MetaData()
+            user_copy = user.to_metadata(m2)
+
+        .. versionchanged:: 1.4  The :meth:`_schema.Table.to_metadata` function
+           was renamed from :meth:`_schema.Table.tometadata`.
+
+
+        :param metadata: Target :class:`_schema.MetaData` object,
+         into which the
+         new :class:`_schema.Table` object will be created.
+
+        :param schema: optional string name indicating the target schema.
+         Defaults to the special symbol :attr:`.RETAIN_SCHEMA` which indicates
+         that no change to the schema name should be made in the new
+         :class:`_schema.Table`.  If set to a string name, the new
+         :class:`_schema.Table`
+         will have this new name as the ``.schema``.  If set to ``None``, the
+         schema will be set to that of the schema set on the target
+         :class:`_schema.MetaData`, which is typically ``None`` as well,
+         unless
+         set explicitly::
+
+            m2 = MetaData(schema="newschema")
+
+            # user_copy_one will have "newschema" as the schema name
+            user_copy_one = user.to_metadata(m2, schema=None)
+
+            m3 = MetaData()  # schema defaults to None
+
+            # user_copy_two will have None as the schema name
+            user_copy_two = user.to_metadata(m3, schema=None)
+
+        :param referred_schema_fn: optional callable which can be supplied
+         in order to provide for the schema name that should be assigned
+         to the referenced table of a :class:`_schema.ForeignKeyConstraint`.
+         The callable accepts this parent :class:`_schema.Table`, the
+         target schema that we are changing to, the
+         :class:`_schema.ForeignKeyConstraint` object, and the existing
+         "target schema" of that constraint.  The function should return the
+         string schema name that should be applied.    To reset the schema
+         to "none", return the symbol :data:`.BLANK_SCHEMA`.  To effect no
+         change, return ``None`` or :data:`.RETAIN_SCHEMA`.
+
+         .. versionchanged:: 1.4.33  The ``referred_schema_fn`` function
+            may return the :data:`.BLANK_SCHEMA` or :data:`.RETAIN_SCHEMA`
+            symbols.
+
+         E.g.::
+
+                def referred_schema_fn(table, to_schema, constraint, referred_schema):
+                    if referred_schema == "base_tables":
+                        return referred_schema
+                    else:
+                        return to_schema
+
+
+                new_table = table.to_metadata(
+                    m2, schema="alt_schema", referred_schema_fn=referred_schema_fn
+                )
+
+        :param name: optional string name indicating the target table name.
+         If not specified or None, the table name is retained.  This allows
+         a :class:`_schema.Table` to be copied to the same
+         :class:`_schema.MetaData` target
+         with a new name.
+
+        """  # noqa: E501
+        if name is None:
+            name = self.name
+
+        actual_schema: Optional[str]
+
+        if schema is RETAIN_SCHEMA:
+            actual_schema = self.schema
+        elif schema is None:
+            actual_schema = metadata.schema
+        else:
+            actual_schema = schema
+        key = _get_table_key(name, actual_schema)
+        if key in metadata.tables:
+            util.warn(
+                f"Table '{self.description}' already exists within the given "
+                "MetaData - not copying."
+            )
+            return metadata.tables[key]
+
+        args = []
+        for col in self.columns:
+            args.append(col._copy(schema=actual_schema, _to_metadata=metadata))
+        table = Table(
+            name,
+            metadata,
+            schema=actual_schema,
+            comment=self.comment,
+            *args,
+            **self.kwargs,
+        )
+        for const in self.constraints:
+            if isinstance(const, ForeignKeyConstraint):
+                referred_schema = const._referred_schema
+                if referred_schema_fn:
+                    fk_constraint_schema = referred_schema_fn(
+                        self, actual_schema, const, referred_schema
+                    )
+                else:
+                    fk_constraint_schema = (
+                        actual_schema
+                        if referred_schema == self.schema
+                        else None
+                    )
+                table.append_constraint(
+                    const._copy(
+                        schema=fk_constraint_schema, target_table=table
+                    )
+                )
+            elif not const._type_bound:
+                # skip unique constraints that would be generated
+                # by the 'unique' flag on Column
+                if const._column_flag:
+                    continue
+
+                table.append_constraint(
+                    const._copy(schema=actual_schema, target_table=table)
+                )
+        for index in self.indexes:
+            # skip indexes that would be generated
+            # by the 'index' flag on Column
+            if index._column_flag:
+                continue
+            Index(
+                index.name,
+                unique=index.unique,
+                *[
+                    _copy_expression(expr, self, table)
+                    for expr in index._table_bound_expressions
+                ],
+                _table=table,
+                **index.kwargs,
+            )
+        return self._schema_item_copy(table)
+
+
+class Column(DialectKWArgs, SchemaItem, ColumnClause[_T]):
+    """Represents a column in a database table."""
+
+    __visit_name__ = "column"
+
+    inherit_cache = True
+    key: str
+
+    server_default: Optional[FetchedValue]
+
+    def __init__(
+        self,
+        __name_pos: Optional[
+            Union[str, _TypeEngineArgument[_T], SchemaEventTarget]
+        ] = None,
+        __type_pos: Optional[
+            Union[_TypeEngineArgument[_T], SchemaEventTarget]
+        ] = None,
+        *args: SchemaEventTarget,
+        name: Optional[str] = None,
+        type_: Optional[_TypeEngineArgument[_T]] = None,
+        autoincrement: _AutoIncrementType = "auto",
+        default: Optional[Any] = _NoArg.NO_ARG,
+        insert_default: Optional[Any] = _NoArg.NO_ARG,
+        doc: Optional[str] = None,
+        key: Optional[str] = None,
+        index: Optional[bool] = None,
+        unique: Optional[bool] = None,
+        info: Optional[_InfoType] = None,
+        nullable: Optional[
+            Union[bool, Literal[SchemaConst.NULL_UNSPECIFIED]]
+        ] = SchemaConst.NULL_UNSPECIFIED,
+        onupdate: Optional[Any] = None,
+        primary_key: bool = False,
+        server_default: Optional[_ServerDefaultArgument] = None,
+        server_onupdate: Optional[_ServerOnUpdateArgument] = None,
+        quote: Optional[bool] = None,
+        system: bool = False,
+        comment: Optional[str] = None,
+        insert_sentinel: bool = False,
+        _omit_from_statements: bool = False,
+        _proxies: Optional[Any] = None,
+        **dialect_kwargs: Any,
+    ):
+        r"""
+        Construct a new ``Column`` object.
+
+        :param name: The name of this column as represented in the database.
+          This argument may be the first positional argument, or specified
+          via keyword.
+
+          Names which contain no upper case characters
+          will be treated as case insensitive names, and will not be quoted
+          unless they are a reserved word.  Names with any number of upper
+          case characters will be quoted and sent exactly.  Note that this
+          behavior applies even for databases which standardize upper
+          case names as case insensitive such as Oracle Database.
+
+          The name field may be omitted at construction time and applied
+          later, at any time before the Column is associated with a
+          :class:`_schema.Table`.  This is to support convenient
+          usage within the :mod:`~sqlalchemy.ext.declarative` extension.
+
+        :param type\_: The column's type, indicated using an instance which
+          subclasses :class:`~sqlalchemy.types.TypeEngine`.  If no arguments
+          are required for the type, the class of the type can be sent
+          as well, e.g.::
+
+            # use a type with arguments
+            Column("data", String(50))
+
+            # use no arguments
+            Column("level", Integer)
+
+          The ``type`` argument may be the second positional argument
+          or specified by keyword.
+
+          If the ``type`` is ``None`` or is omitted, it will first default to
+          the special type :class:`.NullType`.  If and when this
+          :class:`_schema.Column` is made to refer to another column using
+          :class:`_schema.ForeignKey` and/or
+          :class:`_schema.ForeignKeyConstraint`, the type
+          of the remote-referenced column will be copied to this column as
+          well, at the moment that the foreign key is resolved against that
+          remote :class:`_schema.Column` object.
+
+        :param \*args: Additional positional arguments include various
+          :class:`.SchemaItem` derived constructs which will be applied
+          as options to the column.  These include instances of
+          :class:`.Constraint`, :class:`_schema.ForeignKey`,
+          :class:`.ColumnDefault`, :class:`.Sequence`, :class:`.Computed`
+          :class:`.Identity`.  In some cases an
+          equivalent keyword argument is available such as ``server_default``,
+          ``default`` and ``unique``.
+
+        :param autoincrement: Set up "auto increment" semantics for an
+          **integer primary key column with no foreign key dependencies**
+          (see later in this docstring for a more specific definition).
+          This may influence the :term:`DDL` that will be emitted for
+          this column during a table create, as well as how the column
+          will be considered when INSERT statements are compiled and
+          executed.
+
+          The default value is the string ``"auto"``,
+          which indicates that a single-column (i.e. non-composite) primary key
+          that is of an INTEGER type with no other client-side or server-side
+          default constructs indicated should receive auto increment semantics
+          automatically. Other values include ``True`` (force this column to
+          have auto-increment semantics for a :term:`composite primary key` as
+          well), ``False`` (this column should never have auto-increment
+          semantics), and the string ``"ignore_fk"`` (special-case for foreign
+          key columns, see below).
+
+          The term "auto increment semantics" refers both to the kind of DDL
+          that will be emitted for the column within a CREATE TABLE statement,
+          when methods such as :meth:`.MetaData.create_all` and
+          :meth:`.Table.create` are invoked, as well as how the column will be
+          considered when an INSERT statement is compiled and emitted to the
+          database:
+
+          * **DDL rendering** (i.e. :meth:`.MetaData.create_all`,
+            :meth:`.Table.create`): When used on a :class:`.Column` that has
+            no other
+            default-generating construct associated with it (such as a
+            :class:`.Sequence` or :class:`.Identity` construct), the parameter
+            will imply that database-specific keywords such as PostgreSQL
+            ``SERIAL``, MySQL ``AUTO_INCREMENT``, or ``IDENTITY`` on SQL Server
+            should also be rendered.  Not every database backend has an
+            "implied" default generator available; for example the Oracle Database
+            backends alway needs an explicit construct such as
+            :class:`.Identity` to be included with a :class:`.Column` in order
+            for the DDL rendered to include auto-generating constructs to also
+            be produced in the database.
+
+          * **INSERT semantics** (i.e. when a :func:`_sql.insert` construct is
+            compiled into a SQL string and is then executed on a database using
+            :meth:`_engine.Connection.execute` or equivalent): A single-row
+            INSERT statement will be known to produce a new integer primary key
+            value automatically for this column, which will be accessible
+            after the statement is invoked via the
+            :attr:`.CursorResult.inserted_primary_key` attribute upon the
+            :class:`_result.Result` object.   This also applies towards use of the
+            ORM when ORM-mapped objects are persisted to the database,
+            indicating that a new integer primary key will be available to
+            become part of the :term:`identity key` for that object.  This
+            behavior takes place regardless of what DDL constructs are
+            associated with the :class:`_schema.Column` and is independent
+            of the "DDL Rendering" behavior discussed in the previous note
+            above.
+
+          The parameter may be set to ``True`` to indicate that a column which
+          is part of a composite (i.e. multi-column) primary key should
+          have autoincrement semantics, though note that only one column
+          within a primary key may have this setting.    It can also
+          be set to ``True`` to indicate autoincrement semantics on a
+          column that has a client-side or server-side default configured,
+          however note that not all dialects can accommodate all styles
+          of default as an "autoincrement".  It can also be
+          set to ``False`` on a single-column primary key that has a
+          datatype of INTEGER in order to disable auto increment semantics
+          for that column.
+
+          The setting *only* has an effect for columns which are:
+
+          * Integer derived (i.e. INT, SMALLINT, BIGINT).
+
+          * Part of the primary key
+
+          * Not referring to another column via :class:`_schema.ForeignKey`,
+            unless
+            the value is specified as ``'ignore_fk'``::
+
+                # turn on autoincrement for this column despite
+                # the ForeignKey()
+                Column(
+                    "id",
+                    ForeignKey("other.id"),
+                    primary_key=True,
+                    autoincrement="ignore_fk",
+                )
+
+          It is typically not desirable to have "autoincrement" enabled on a
+          column that refers to another via foreign key, as such a column is
+          required to refer to a value that originates from elsewhere.
+
+          The setting has these effects on columns that meet the
+          above criteria:
+
+          * DDL issued for the column, if the column does not already include
+            a default generating construct supported by the backend such as
+            :class:`.Identity`, will include database-specific
+            keywords intended to signify this column as an
+            "autoincrement" column for specific backends.   Behavior for
+            primary SQLAlchemy dialects includes:
+
+            * AUTO INCREMENT on MySQL and MariaDB
+            * SERIAL on PostgreSQL
+            * IDENTITY on MS-SQL - this occurs even without the
+              :class:`.Identity` construct as the
+              :paramref:`.Column.autoincrement` parameter pre-dates this
+              construct.
+            * SQLite - SQLite integer primary key columns are implicitly
+              "auto incrementing" and no additional keywords are rendered;
+              to render the special SQLite keyword ``AUTOINCREMENT``
+              is not included as this is unnecessary and not recommended
+              by the database vendor.  See the section
+              :ref:`sqlite_autoincrement` for more background.
+            * Oracle Database - The Oracle Database dialects have no default "autoincrement"
+              feature available at this time, instead the :class:`.Identity`
+              construct is recommended to achieve this (the :class:`.Sequence`
+              construct may also be used).
+            * Third-party dialects - consult those dialects' documentation
+              for details on their specific behaviors.
+
+          * When a single-row :func:`_sql.insert` construct is compiled and
+            executed, which does not set the :meth:`_sql.Insert.inline`
+            modifier, newly generated primary key values for this column
+            will be automatically retrieved upon statement execution
+            using a method specific to the database driver in use:
+
+            * MySQL, SQLite - calling upon ``cursor.lastrowid()``
+              (see
+              `https://www.python.org/dev/peps/pep-0249/#lastrowid
+              <https://www.python.org/dev/peps/pep-0249/#lastrowid>`_)
+            * PostgreSQL, SQL Server, Oracle Database - use RETURNING or an equivalent
+              construct when rendering an INSERT statement, and then retrieving
+              the newly generated primary key values after execution
+            * PostgreSQL, Oracle Database for :class:`_schema.Table` objects that
+              set :paramref:`_schema.Table.implicit_returning` to False -
+              for a :class:`.Sequence` only, the :class:`.Sequence` is invoked
+              explicitly before the INSERT statement takes place so that the
+              newly generated primary key value is available to the client
+            * SQL Server for :class:`_schema.Table` objects that
+              set :paramref:`_schema.Table.implicit_returning` to False -
+              the ``SELECT scope_identity()`` construct is used after the
+              INSERT statement is invoked to retrieve the newly generated
+              primary key value.
+            * Third-party dialects - consult those dialects' documentation
+              for details on their specific behaviors.
+
+          * For multiple-row :func:`_sql.insert` constructs invoked with
+            a list of parameters (i.e. "executemany" semantics), primary-key
+            retrieving behaviors are generally disabled, however there may
+            be special APIs that may be used to retrieve lists of new
+            primary key values for an "executemany", such as the psycopg2
+            "fast insertmany" feature.  Such features are very new and
+            may not yet be well covered in documentation.
+
+        :param default: A scalar, Python callable, or
+            :class:`_expression.ColumnElement` expression representing the
+            *default value* for this column, which will be invoked upon insert
+            if this column is otherwise not specified in the VALUES clause of
+            the insert. This is a shortcut to using :class:`.ColumnDefault` as
+            a positional argument; see that class for full detail on the
+            structure of the argument.
+
+            Contrast this argument to
+            :paramref:`_schema.Column.server_default`
+            which creates a default generator on the database side.
+
+            .. seealso::
+
+                :ref:`metadata_defaults_toplevel`
+
+        :param insert_default: An alias of :paramref:`.Column.default`
+            for compatibility with :func:`_orm.mapped_column`.
+
+            .. versionadded: 2.0.31
+
+        :param doc: optional String that can be used by the ORM or similar
+            to document attributes on the Python side.   This attribute does
+            **not** render SQL comments; use the
+            :paramref:`_schema.Column.comment`
+            parameter for this purpose.
+
+        :param key: An optional string identifier which will identify this
+            ``Column`` object on the :class:`_schema.Table`.
+            When a key is provided,
+            this is the only identifier referencing the ``Column`` within the
+            application, including ORM attribute mapping; the ``name`` field
+            is used only when rendering SQL.
+
+        :param index: When ``True``, indicates that a :class:`_schema.Index`
+            construct will be automatically generated for this
+            :class:`_schema.Column`, which will result in a "CREATE INDEX"
+            statement being emitted for the :class:`_schema.Table` when the DDL
+            create operation is invoked.
+
+            Using this flag is equivalent to making use of the
+            :class:`_schema.Index` construct explicitly at the level of the
+            :class:`_schema.Table` construct itself::
+
+                Table(
+                    "some_table",
+                    metadata,
+                    Column("x", Integer),
+                    Index("ix_some_table_x", "x"),
+                )
+
+            To add the :paramref:`_schema.Index.unique` flag to the
+            :class:`_schema.Index`, set both the
+            :paramref:`_schema.Column.unique` and
+            :paramref:`_schema.Column.index` flags to True simultaneously,
+            which will have the effect of rendering the "CREATE UNIQUE INDEX"
+            DDL instruction instead of "CREATE INDEX".
+
+            The name of the index is generated using the
+            :ref:`default naming convention <constraint_default_naming_convention>`
+            which for the :class:`_schema.Index` construct is of the form
+            ``ix_<tablename>_<columnname>``.
+
+            As this flag is intended only as a convenience for the common case
+            of adding a single-column, default configured index to a table
+            definition, explicit use of the :class:`_schema.Index` construct
+            should be preferred for most use cases, including composite indexes
+            that encompass more than one column, indexes with SQL expressions
+            or ordering, backend-specific index configuration options, and
+            indexes that use a specific name.
+
+            .. note:: the :attr:`_schema.Column.index` attribute on
+               :class:`_schema.Column`
+               **does not indicate** if this column is indexed or not, only
+               if this flag was explicitly set here.  To view indexes on
+               a column, view the :attr:`_schema.Table.indexes` collection
+               or use :meth:`_reflection.Inspector.get_indexes`.
+
+            .. seealso::
+
+                :ref:`schema_indexes`
+
+                :ref:`constraint_naming_conventions`
+
+                :paramref:`_schema.Column.unique`
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param nullable: When set to ``False``, will cause the "NOT NULL"
+            phrase to be added when generating DDL for the column.   When
+            ``True``, will normally generate nothing (in SQL this defaults to
+            "NULL"), except in some very specific backend-specific edge cases
+            where "NULL" may render explicitly.
+            Defaults to ``True`` unless :paramref:`_schema.Column.primary_key`
+            is also ``True`` or the column specifies a :class:`_sql.Identity`,
+            in which case it defaults to ``False``.
+            This parameter is only used when issuing CREATE TABLE statements.
+
+            .. note::
+
+                When the column specifies a :class:`_sql.Identity` this
+                parameter is in general ignored by the DDL compiler. The
+                PostgreSQL database allows nullable identity column by
+                setting this parameter to ``True`` explicitly.
+
+        :param onupdate: A scalar, Python callable, or
+            :class:`~sqlalchemy.sql.expression.ClauseElement` representing a
+            default value to be applied to the column within UPDATE
+            statements, which will be invoked upon update if this column is not
+            present in the SET clause of the update. This is a shortcut to
+            using :class:`.ColumnDefault` as a positional argument with
+            ``for_update=True``.
+
+            .. seealso::
+
+                :ref:`metadata_defaults` - complete discussion of onupdate
+
+        :param primary_key: If ``True``, marks this column as a primary key
+            column. Multiple columns can have this flag set to specify
+            composite primary keys. As an alternative, the primary key of a
+            :class:`_schema.Table` can be specified via an explicit
+            :class:`.PrimaryKeyConstraint` object.
+
+        :param server_default: A :class:`.FetchedValue` instance, str, Unicode
+            or :func:`~sqlalchemy.sql.expression.text` construct representing
+            the DDL DEFAULT value for the column.
+
+            String types will be emitted as-is, surrounded by single quotes::
+
+                Column("x", Text, server_default="val")
+
+            will render:
+
+            .. sourcecode:: sql
+
+                x TEXT DEFAULT 'val'
+
+            A :func:`~sqlalchemy.sql.expression.text` expression will be
+            rendered as-is, without quotes::
+
+                Column("y", DateTime, server_default=text("NOW()"))
+
+            will render:
+
+            .. sourcecode:: sql
+
+                y DATETIME DEFAULT NOW()
+
+            Strings and text() will be converted into a
+            :class:`.DefaultClause` object upon initialization.
+
+            This parameter can also accept complex combinations of contextually
+            valid SQLAlchemy expressions or constructs::
+
+                from sqlalchemy import create_engine
+                from sqlalchemy import Table, Column, MetaData, ARRAY, Text
+                from sqlalchemy.dialects.postgresql import array
+
+                engine = create_engine(
+                    "postgresql+psycopg2://scott:tiger@localhost/mydatabase"
+                )
+                metadata_obj = MetaData()
+                tbl = Table(
+                    "foo",
+                    metadata_obj,
+                    Column(
+                        "bar", ARRAY(Text), server_default=array(["biz", "bang", "bash"])
+                    ),
+                )
+                metadata_obj.create_all(engine)
+
+            The above results in a table created with the following SQL:
+
+            .. sourcecode:: sql
+
+                CREATE TABLE foo (
+                    bar TEXT[] DEFAULT ARRAY['biz', 'bang', 'bash']
+                )
+
+            Use :class:`.FetchedValue` to indicate that an already-existing
+            column will generate a default value on the database side which
+            will be available to SQLAlchemy for post-fetch after inserts. This
+            construct does not specify any DDL and the implementation is left
+            to the database, such as via a trigger.
+
+            .. seealso::
+
+                :ref:`server_defaults` - complete discussion of server side
+                defaults
+
+        :param server_onupdate: A :class:`.FetchedValue` instance
+            representing a database-side default generation function,
+            such as a trigger. This
+            indicates to SQLAlchemy that a newly generated value will be
+            available after updates. This construct does not actually
+            implement any kind of generation function within the database,
+            which instead must be specified separately.
+
+
+            .. warning:: This directive **does not** currently produce MySQL's
+               "ON UPDATE CURRENT_TIMESTAMP()" clause.  See
+               :ref:`mysql_timestamp_onupdate` for background on how to
+               produce this clause.
+
+            .. seealso::
+
+                :ref:`triggered_columns`
+
+        :param quote: Force quoting of this column's name on or off,
+             corresponding to ``True`` or ``False``. When left at its default
+             of ``None``, the column identifier will be quoted according to
+             whether the name is case sensitive (identifiers with at least one
+             upper case character are treated as case sensitive), or if it's a
+             reserved word. This flag is only needed to force quoting of a
+             reserved word which is not known by the SQLAlchemy dialect.
+
+        :param unique: When ``True``, and the :paramref:`_schema.Column.index`
+            parameter is left at its default value of ``False``,
+            indicates that a :class:`_schema.UniqueConstraint`
+            construct will be automatically generated for this
+            :class:`_schema.Column`,
+            which will result in a "UNIQUE CONSTRAINT" clause referring
+            to this column being included
+            in the ``CREATE TABLE`` statement emitted, when the DDL create
+            operation for the :class:`_schema.Table` object is invoked.
+
+            When this flag is ``True`` while the
+            :paramref:`_schema.Column.index` parameter is simultaneously
+            set to ``True``, the effect instead is that a
+            :class:`_schema.Index` construct which includes the
+            :paramref:`_schema.Index.unique` parameter set to ``True``
+            is generated.  See the documentation for
+            :paramref:`_schema.Column.index` for additional detail.
+
+            Using this flag is equivalent to making use of the
+            :class:`_schema.UniqueConstraint` construct explicitly at the
+            level of the :class:`_schema.Table` construct itself::
+
+                Table("some_table", metadata, Column("x", Integer), UniqueConstraint("x"))
+
+            The :paramref:`_schema.UniqueConstraint.name` parameter
+            of the unique constraint object is left at its default value
+            of ``None``; in the absence of a :ref:`naming convention <constraint_naming_conventions>`
+            for the enclosing :class:`_schema.MetaData`, the UNIQUE CONSTRAINT
+            construct will be emitted as unnamed, which typically invokes
+            a database-specific naming convention to take place.
+
+            As this flag is intended only as a convenience for the common case
+            of adding a single-column, default configured unique constraint to a table
+            definition, explicit use of the :class:`_schema.UniqueConstraint` construct
+            should be preferred for most use cases, including composite constraints
+            that encompass more than one column, backend-specific index configuration options, and
+            constraints that use a specific name.
+
+            .. note:: the :attr:`_schema.Column.unique` attribute on
+                :class:`_schema.Column`
+                **does not indicate** if this column has a unique constraint or
+                not, only if this flag was explicitly set here.  To view
+                indexes and unique constraints that may involve this column,
+                view the
+                :attr:`_schema.Table.indexes` and/or
+                :attr:`_schema.Table.constraints` collections or use
+                :meth:`_reflection.Inspector.get_indexes` and/or
+                :meth:`_reflection.Inspector.get_unique_constraints`
+
+            .. seealso::
+
+                :ref:`schema_unique_constraint`
+
+                :ref:`constraint_naming_conventions`
+
+                :paramref:`_schema.Column.index`
+
+        :param system: When ``True``, indicates this is a "system" column,
+             that is a column which is automatically made available by the
+             database, and should not be included in the columns list for a
+             ``CREATE TABLE`` statement.
+
+             For more elaborate scenarios where columns should be
+             conditionally rendered differently on different backends,
+             consider custom compilation rules for :class:`.CreateColumn`.
+
+        :param comment: Optional string that will render an SQL comment on
+             table creation.
+
+             .. versionadded:: 1.2 Added the
+                :paramref:`_schema.Column.comment`
+                parameter to :class:`_schema.Column`.
+
+        :param insert_sentinel: Marks this :class:`_schema.Column` as an
+         :term:`insert sentinel` used for optimizing the performance of the
+         :term:`insertmanyvalues` feature for tables that don't
+         otherwise have qualifying primary key configurations.
+
+         .. versionadded:: 2.0.10
+
+         .. seealso::
+
+            :func:`_schema.insert_sentinel` - all in one helper for declaring
+            sentinel columns
+
+            :ref:`engine_insertmanyvalues`
+
+            :ref:`engine_insertmanyvalues_sentinel_columns`
+
+
+        """  # noqa: E501, RST201, RST202
+
+        l_args = [__name_pos, __type_pos] + list(args)
+        del args
+
+        if l_args:
+            if isinstance(l_args[0], str):
+                if name is not None:
+                    raise exc.ArgumentError(
+                        "May not pass name positionally and as a keyword."
+                    )
+                name = l_args.pop(0)  # type: ignore
+            elif l_args[0] is None:
+                l_args.pop(0)
+        if l_args:
+            coltype = l_args[0]
+
+            if hasattr(coltype, "_sqla_type"):
+                if type_ is not None:
+                    raise exc.ArgumentError(
+                        "May not pass type_ positionally and as a keyword."
+                    )
+                type_ = l_args.pop(0)  # type: ignore
+            elif l_args[0] is None:
+                l_args.pop(0)
+
+        if name is not None:
+            name = quoted_name(name, quote)
+        elif quote is not None:
+            raise exc.ArgumentError(
+                "Explicit 'name' is required when sending 'quote' argument"
+            )
+
+        # name = None is expected to be an interim state
+        # note this use case is legacy now that ORM declarative has a
+        # dedicated "column" construct local to the ORM
+        super().__init__(name, type_)  # type: ignore
+
+        self.key = key if key is not None else name  # type: ignore
+        self.primary_key = primary_key
+        self._insert_sentinel = insert_sentinel
+        self._omit_from_statements = _omit_from_statements
+        self._user_defined_nullable = udn = nullable
+        if udn is not NULL_UNSPECIFIED:
+            self.nullable = udn
+        else:
+            self.nullable = not primary_key
+
+        # these default to None because .index and .unique is *not*
+        # an informational flag about Column - there can still be an
+        # Index or UniqueConstraint referring to this Column.
+        self.index = index
+        self.unique = unique
+
+        self.system = system
+        self.doc = doc
+        self.autoincrement: _AutoIncrementType = autoincrement
+        self.constraints = set()
+        self.foreign_keys = set()
+        self.comment = comment
+        self.computed = None
+        self.identity = None
+
+        # check if this Column is proxying another column
+
+        if _proxies is not None:
+            self._proxies = _proxies
+        else:
+            # otherwise, add DDL-related events
+            self._set_type(self.type)
+
+        if insert_default is not _NoArg.NO_ARG:
+            resolved_default = insert_default
+        elif default is not _NoArg.NO_ARG:
+            resolved_default = default
+        else:
+            resolved_default = None
+
+        if resolved_default is not None:
+            if not isinstance(resolved_default, (ColumnDefault, Sequence)):
+                resolved_default = ColumnDefault(resolved_default)
+
+            self.default = resolved_default
+            l_args.append(resolved_default)
+        else:
+            self.default = None
+
+        if onupdate is not None:
+            if not isinstance(onupdate, (ColumnDefault, Sequence)):
+                onupdate = ColumnDefault(onupdate, for_update=True)
+
+            self.onupdate = onupdate
+            l_args.append(onupdate)
+        else:
+            self.onupdate = None
+
+        if server_default is not None:
+            if isinstance(server_default, FetchedValue):
+                server_default = server_default._as_for_update(False)
+                l_args.append(server_default)
+            else:
+                server_default = DefaultClause(server_default)
+                l_args.append(server_default)
+        self.server_default = server_default
+
+        if server_onupdate is not None:
+            if isinstance(server_onupdate, FetchedValue):
+                server_onupdate = server_onupdate._as_for_update(True)
+                l_args.append(server_onupdate)
+            else:
+                server_onupdate = DefaultClause(
+                    server_onupdate, for_update=True
+                )
+                l_args.append(server_onupdate)
+        self.server_onupdate = server_onupdate
+
+        self._init_items(*cast(_typing_Sequence[SchemaItem], l_args))
+
+        util.set_creation_order(self)
+
+        if info is not None:
+            self.info = info
+
+        self._extra_kwargs(**dialect_kwargs)
+
+    table: Table
+
+    constraints: Set[Constraint]
+
+    foreign_keys: Set[ForeignKey]
+    """A collection of all :class:`_schema.ForeignKey` marker objects
+       associated with this :class:`_schema.Column`.
+
+       Each object is a member of a :class:`_schema.Table`-wide
+       :class:`_schema.ForeignKeyConstraint`.
+
+       .. seealso::
+
+           :attr:`_schema.Table.foreign_keys`
+
+    """
+
+    index: Optional[bool]
+    """The value of the :paramref:`_schema.Column.index` parameter.
+
+       Does not indicate if this :class:`_schema.Column` is actually indexed
+       or not; use :attr:`_schema.Table.indexes`.
+
+       .. seealso::
+
+           :attr:`_schema.Table.indexes`
+    """
+
+    unique: Optional[bool]
+    """The value of the :paramref:`_schema.Column.unique` parameter.
+
+       Does not indicate if this :class:`_schema.Column` is actually subject to
+       a unique constraint or not; use :attr:`_schema.Table.indexes` and
+       :attr:`_schema.Table.constraints`.
+
+       .. seealso::
+
+           :attr:`_schema.Table.indexes`
+
+           :attr:`_schema.Table.constraints`.
+
+    """
+
+    computed: Optional[Computed]
+
+    identity: Optional[Identity]
+
+    def _set_type(self, type_: TypeEngine[Any]) -> None:
+        assert self.type._isnull or type_ is self.type
+
+        self.type = type_
+        if isinstance(self.type, SchemaEventTarget):
+            self.type._set_parent_with_dispatch(self)
+        for impl in self.type._variant_mapping.values():
+            if isinstance(impl, SchemaEventTarget):
+                impl._set_parent_with_dispatch(self)
+
+    @HasMemoized.memoized_attribute
+    def _default_description_tuple(self) -> _DefaultDescriptionTuple:
+        """used by default.py -> _process_execute_defaults()"""
+
+        return _DefaultDescriptionTuple._from_column_default(self.default)
+
+    @HasMemoized.memoized_attribute
+    def _onupdate_description_tuple(self) -> _DefaultDescriptionTuple:
+        """used by default.py -> _process_execute_defaults()"""
+        return _DefaultDescriptionTuple._from_column_default(self.onupdate)
+
+    @util.memoized_property
+    def _gen_static_annotations_cache_key(self) -> bool:  # type: ignore
+        """special attribute used by cache key gen, if true, we will
+        use a static cache key for the annotations dictionary, else we
+        will generate a new cache key for annotations each time.
+
+        Added for #8790
+
+        """
+        return self.table is not None and self.table._is_table
+
+    def _extra_kwargs(self, **kwargs: Any) -> None:
+        self._validate_dialect_kwargs(kwargs)
+
+    def __str__(self) -> str:
+        if self.name is None:
+            return "(no name)"
+        elif self.table is not None:
+            if self.table.named_with_column:
+                return self.table.description + "." + self.description
+            else:
+                return self.description
+        else:
+            return self.description
+
+    def references(self, column: Column[Any]) -> bool:
+        """Return True if this Column references the given column via foreign
+        key."""
+
+        for fk in self.foreign_keys:
+            if fk.column.proxy_set.intersection(column.proxy_set):
+                return True
+        else:
+            return False
+
+    def append_foreign_key(self, fk: ForeignKey) -> None:
+        fk._set_parent_with_dispatch(self)
+
+    def __repr__(self) -> str:
+        kwarg = []
+        if self.key != self.name:
+            kwarg.append("key")
+        if self.primary_key:
+            kwarg.append("primary_key")
+        if not self.nullable:
+            kwarg.append("nullable")
+        if self.onupdate:
+            kwarg.append("onupdate")
+        if self.default:
+            kwarg.append("default")
+        if self.server_default:
+            kwarg.append("server_default")
+        if self.comment:
+            kwarg.append("comment")
+        return "Column(%s)" % ", ".join(
+            [repr(self.name)]
+            + [repr(self.type)]
+            + [repr(x) for x in self.foreign_keys if x is not None]
+            + [repr(x) for x in self.constraints]
+            + [
+                (
+                    self.table is not None
+                    and "table=<%s>" % self.table.description
+                    or "table=None"
+                )
+            ]
+            + ["%s=%s" % (k, repr(getattr(self, k))) for k in kwarg]
+        )
+
+    def _set_parent(  # type: ignore[override]
+        self,
+        parent: SchemaEventTarget,
+        *,
+        all_names: Dict[str, Column[Any]],
+        allow_replacements: bool,
+        **kw: Any,
+    ) -> None:
+        table = parent
+        assert isinstance(table, Table)
+        if not self.name:
+            raise exc.ArgumentError(
+                "Column must be constructed with a non-blank name or "
+                "assign a non-blank .name before adding to a Table."
+            )
+
+        self._reset_memoizations()
+
+        if self.key is None:
+            self.key = self.name
+
+        existing = getattr(self, "table", None)
+        if existing is not None and existing is not table:
+            raise exc.ArgumentError(
+                f"Column object '{self.key}' already "
+                f"assigned to Table '{existing.description}'"
+            )
+
+        extra_remove = None
+        existing_col = None
+        conflicts_on = ""
+
+        if self.key in table._columns:
+            existing_col = table._columns[self.key]
+            if self.key == self.name:
+                conflicts_on = "name"
+            else:
+                conflicts_on = "key"
+        elif self.name in all_names:
+            existing_col = all_names[self.name]
+            extra_remove = {existing_col}
+            conflicts_on = "name"
+
+        if existing_col is not None:
+            if existing_col is not self:
+                if not allow_replacements:
+                    raise exc.DuplicateColumnError(
+                        f"A column with {conflicts_on} "
+                        f"""'{
+                            self.key if conflicts_on == 'key' else self.name
+                        }' """
+                        f"is already present in table '{table.name}'."
+                    )
+                for fk in existing_col.foreign_keys:
+                    table.foreign_keys.remove(fk)
+                    if fk.constraint in table.constraints:
+                        # this might have been removed
+                        # already, if it's a composite constraint
+                        # and more than one col being replaced
+                        table.constraints.remove(fk.constraint)
+
+        if extra_remove and existing_col is not None and self.key == self.name:
+            util.warn(
+                f'Column with user-specified key "{existing_col.key}" is '
+                "being replaced with "
+                f'plain named column "{self.name}", '
+                f'key "{existing_col.key}" is being removed.  If this is a '
+                "reflection operation, specify autoload_replace=False to "
+                "prevent this replacement."
+            )
+        table._columns.replace(self, extra_remove=extra_remove)
+        all_names[self.name] = self
+        self.table = table
+
+        if self._insert_sentinel:
+            if self.table._sentinel_column is not None:
+                raise exc.ArgumentError(
+                    "a Table may have only one explicit sentinel column"
+                )
+            self.table._sentinel_column = self
+
+        if self.primary_key:
+            table.primary_key._replace(self)
+        elif self.key in table.primary_key:
+            raise exc.ArgumentError(
+                f"Trying to redefine primary-key column '{self.key}' as a "
+                f"non-primary-key column on table '{table.fullname}'"
+            )
+
+        if self.index:
+            if isinstance(self.index, str):
+                raise exc.ArgumentError(
+                    "The 'index' keyword argument on Column is boolean only. "
+                    "To create indexes with a specific name, create an "
+                    "explicit Index object external to the Table."
+                )
+            table.append_constraint(
+                Index(
+                    None, self.key, unique=bool(self.unique), _column_flag=True
+                )
+            )
+
+        elif self.unique:
+            if isinstance(self.unique, str):
+                raise exc.ArgumentError(
+                    "The 'unique' keyword argument on Column is boolean "
+                    "only. To create unique constraints or indexes with a "
+                    "specific name, append an explicit UniqueConstraint to "
+                    "the Table's list of elements, or create an explicit "
+                    "Index object external to the Table."
+                )
+            table.append_constraint(
+                UniqueConstraint(self.key, _column_flag=True)
+            )
+
+        self._setup_on_memoized_fks(lambda fk: fk._set_remote_table(table))
+
+        if self.identity and (
+            isinstance(self.default, Sequence)
+            or isinstance(self.onupdate, Sequence)
+        ):
+            raise exc.ArgumentError(
+                "An column cannot specify both Identity and Sequence."
+            )
+
+    def _setup_on_memoized_fks(self, fn: Callable[..., Any]) -> None:
+        fk_keys = [
+            ((self.table.key, self.key), False),
+            ((self.table.key, self.name), True),
+        ]
+        for fk_key, link_to_name in fk_keys:
+            if fk_key in self.table.metadata._fk_memos:
+                for fk in self.table.metadata._fk_memos[fk_key]:
+                    if fk.link_to_name is link_to_name:
+                        fn(fk)
+
+    def _on_table_attach(self, fn: Callable[..., Any]) -> None:
+        if self.table is not None:
+            fn(self, self.table)
+        else:
+            event.listen(self, "after_parent_attach", fn)
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.Column.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(self, **kw: Any) -> Column[Any]:
+        return self._copy(**kw)
+
+    def _copy(self, **kw: Any) -> Column[Any]:
+        """Create a copy of this ``Column``, uninitialized.
+
+        This is used in :meth:`_schema.Table.to_metadata` and by the ORM.
+
+        """
+
+        # Constraint objects plus non-constraint-bound ForeignKey objects
+        args: List[SchemaItem] = [
+            c._copy(**kw) for c in self.constraints if not c._type_bound
+        ] + [c._copy(**kw) for c in self.foreign_keys if not c.constraint]
+
+        # ticket #5276
+        column_kwargs = {}
+        for dialect_name in self.dialect_options:
+            dialect_options = self.dialect_options[dialect_name]._non_defaults
+            for (
+                dialect_option_key,
+                dialect_option_value,
+            ) in dialect_options.items():
+                column_kwargs[dialect_name + "_" + dialect_option_key] = (
+                    dialect_option_value
+                )
+
+        server_default = self.server_default
+        server_onupdate = self.server_onupdate
+        if isinstance(server_default, (Computed, Identity)):
+            # TODO: likely should be copied in all cases
+            # TODO: if a Sequence, we would need to transfer the Sequence
+            # .metadata as well
+            args.append(server_default._copy(**kw))
+            server_default = server_onupdate = None
+
+        type_ = self.type
+        if isinstance(type_, SchemaEventTarget):
+            type_ = type_.copy(**kw)
+
+        # TODO: DefaultGenerator is not copied here!  it's just used again
+        # with _set_parent() pointing to the old column.  see the new
+        # use of _copy() in the new _merge() method
+
+        c = self._constructor(
+            name=self.name,
+            type_=type_,
+            key=self.key,
+            primary_key=self.primary_key,
+            unique=self.unique,
+            system=self.system,
+            # quote=self.quote,  # disabled 2013-08-27 (commit 031ef080)
+            index=self.index,
+            autoincrement=self.autoincrement,
+            default=self.default,
+            server_default=server_default,
+            onupdate=self.onupdate,
+            server_onupdate=server_onupdate,
+            doc=self.doc,
+            comment=self.comment,
+            _omit_from_statements=self._omit_from_statements,
+            insert_sentinel=self._insert_sentinel,
+            *args,
+            **column_kwargs,
+        )
+
+        # copy the state of "nullable" exactly, to accommodate for
+        # ORM flipping the .nullable flag directly
+        c.nullable = self.nullable
+        c._user_defined_nullable = self._user_defined_nullable
+
+        return self._schema_item_copy(c)
+
+    def _merge(self, other: Column[Any]) -> None:
+        """merge the elements of another column into this one.
+
+        this is used by ORM pep-593 merge and will likely need a lot
+        of fixes.
+
+
+        """
+
+        if self.primary_key:
+            other.primary_key = True
+
+        if self.autoincrement != "auto" and other.autoincrement == "auto":
+            other.autoincrement = self.autoincrement
+
+        if self.system:
+            other.system = self.system
+
+        if self.info:
+            other.info.update(self.info)
+
+        type_ = self.type
+        if not type_._isnull and other.type._isnull:
+            if isinstance(type_, SchemaEventTarget):
+                type_ = type_.copy()
+
+            other.type = type_
+
+            if isinstance(type_, SchemaEventTarget):
+                type_._set_parent_with_dispatch(other)
+
+            for impl in type_._variant_mapping.values():
+                if isinstance(impl, SchemaEventTarget):
+                    impl._set_parent_with_dispatch(other)
+
+        if (
+            self._user_defined_nullable is not NULL_UNSPECIFIED
+            and other._user_defined_nullable is NULL_UNSPECIFIED
+        ):
+            other.nullable = self.nullable
+            other._user_defined_nullable = self._user_defined_nullable
+
+        if self.default is not None and other.default is None:
+            new_default = self.default._copy()
+            new_default._set_parent(other)
+
+        if self.server_default and other.server_default is None:
+            new_server_default = self.server_default
+            if isinstance(new_server_default, FetchedValue):
+                new_server_default = new_server_default._copy()
+                new_server_default._set_parent(other)
+            else:
+                other.server_default = new_server_default
+
+        if self.server_onupdate and other.server_onupdate is None:
+            new_server_onupdate = self.server_onupdate
+            new_server_onupdate = new_server_onupdate._copy()
+            new_server_onupdate._set_parent(other)
+
+        if self.onupdate and other.onupdate is None:
+            new_onupdate = self.onupdate._copy()
+            new_onupdate._set_parent(other)
+
+        if self.index in (True, False) and other.index is None:
+            other.index = self.index
+
+        if self.unique in (True, False) and other.unique is None:
+            other.unique = self.unique
+
+        if self.doc and other.doc is None:
+            other.doc = self.doc
+
+        if self.comment and other.comment is None:
+            other.comment = self.comment
+
+        for const in self.constraints:
+            if not const._type_bound:
+                new_const = const._copy()
+                new_const._set_parent(other)
+
+        for fk in self.foreign_keys:
+            if not fk.constraint:
+                new_fk = fk._copy()
+                new_fk._set_parent(other)
+
+    def _make_proxy(
+        self,
+        selectable: FromClause,
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        name: Optional[str] = None,
+        key: Optional[str] = None,
+        name_is_truncatable: bool = False,
+        compound_select_cols: Optional[
+            _typing_Sequence[ColumnElement[Any]]
+        ] = None,
+        **kw: Any,
+    ) -> Tuple[str, ColumnClause[_T]]:
+        """Create a *proxy* for this column.
+
+        This is a copy of this ``Column`` referenced by a different parent
+        (such as an alias or select statement).  The column should
+        be used only in select scenarios, as its full DDL/default
+        information is not transferred.
+
+        """
+
+        fk = [
+            ForeignKey(
+                col if col is not None else f._colspec,
+                _unresolvable=col is None,
+                _constraint=f.constraint,
+            )
+            for f, col in [
+                (fk, fk._resolve_column(raiseerr=False))
+                for fk in self.foreign_keys
+            ]
+        ]
+
+        if name is None and self.name is None:
+            raise exc.InvalidRequestError(
+                "Cannot initialize a sub-selectable"
+                " with this Column object until its 'name' has "
+                "been assigned."
+            )
+        try:
+            c = self._constructor(
+                (
+                    coercions.expect(
+                        roles.TruncatedLabelRole, name if name else self.name
+                    )
+                    if name_is_truncatable
+                    else (name or self.name)
+                ),
+                self.type,
+                # this may actually be ._proxy_key when the key is incoming
+                key=key if key else name if name else self.key,
+                primary_key=self.primary_key,
+                nullable=self.nullable,
+                _proxies=(
+                    list(compound_select_cols)
+                    if compound_select_cols
+                    else [self]
+                ),
+                *fk,
+            )
+        except TypeError as err:
+            raise TypeError(
+                "Could not create a copy of this %r object.  "
+                "Ensure the class includes a _constructor() "
+                "attribute or method which accepts the "
+                "standard Column constructor arguments, or "
+                "references the Column class itself." % self.__class__
+            ) from err
+
+        c.table = selectable
+        c._propagate_attrs = selectable._propagate_attrs
+        if selectable._is_clone_of is not None:
+            c._is_clone_of = selectable._is_clone_of.columns.get(c.key)
+
+        if self.primary_key:
+            primary_key.add(c)
+
+        if fk:
+            foreign_keys.update(fk)  # type: ignore
+
+        return c.key, c
+
+
+def insert_sentinel(
+    name: Optional[str] = None,
+    type_: Optional[_TypeEngineArgument[_T]] = None,
+    *,
+    default: Optional[Any] = None,
+    omit_from_statements: bool = True,
+) -> Column[Any]:
+    """Provides a surrogate :class:`_schema.Column` that will act as a
+    dedicated insert :term:`sentinel` column, allowing efficient bulk
+    inserts with deterministic RETURNING sorting for tables that
+    don't otherwise have qualifying primary key configurations.
+
+    Adding this column to a :class:`.Table` object requires that a
+    corresponding database table actually has this column present, so if adding
+    it to an existing model, existing database tables would need to be migrated
+    (e.g. using ALTER TABLE or similar) to include this column.
+
+    For background on how this object is used, see the section
+    :ref:`engine_insertmanyvalues_sentinel_columns` as part of the
+    section :ref:`engine_insertmanyvalues`.
+
+    The :class:`_schema.Column` returned will be a nullable integer column by
+    default and make use of a sentinel-specific default generator used only in
+    "insertmanyvalues" operations.
+
+    .. seealso::
+
+        :func:`_orm.orm_insert_sentinel`
+
+        :paramref:`_schema.Column.insert_sentinel`
+
+        :ref:`engine_insertmanyvalues`
+
+        :ref:`engine_insertmanyvalues_sentinel_columns`
+
+
+    .. versionadded:: 2.0.10
+
+    """
+    return Column(
+        name=name,
+        type_=type_api.INTEGERTYPE if type_ is None else type_,
+        default=(
+            default if default is not None else _InsertSentinelColumnDefault()
+        ),
+        _omit_from_statements=omit_from_statements,
+        insert_sentinel=True,
+    )
+
+
+class ForeignKey(DialectKWArgs, SchemaItem):
+    """Defines a dependency between two columns.
+
+    ``ForeignKey`` is specified as an argument to a :class:`_schema.Column`
+    object,
+    e.g.::
+
+        t = Table(
+            "remote_table",
+            metadata,
+            Column("remote_id", ForeignKey("main_table.id")),
+        )
+
+    Note that ``ForeignKey`` is only a marker object that defines
+    a dependency between two columns.   The actual constraint
+    is in all cases represented by the :class:`_schema.ForeignKeyConstraint`
+    object.   This object will be generated automatically when
+    a ``ForeignKey`` is associated with a :class:`_schema.Column` which
+    in turn is associated with a :class:`_schema.Table`.   Conversely,
+    when :class:`_schema.ForeignKeyConstraint` is applied to a
+    :class:`_schema.Table`,
+    ``ForeignKey`` markers are automatically generated to be
+    present on each associated :class:`_schema.Column`, which are also
+    associated with the constraint object.
+
+    Note that you cannot define a "composite" foreign key constraint,
+    that is a constraint between a grouping of multiple parent/child
+    columns, using ``ForeignKey`` objects.   To define this grouping,
+    the :class:`_schema.ForeignKeyConstraint` object must be used, and applied
+    to the :class:`_schema.Table`.   The associated ``ForeignKey`` objects
+    are created automatically.
+
+    The ``ForeignKey`` objects associated with an individual
+    :class:`_schema.Column`
+    object are available in the `foreign_keys` collection
+    of that column.
+
+    Further examples of foreign key configuration are in
+    :ref:`metadata_foreignkeys`.
+
+    """
+
+    __visit_name__ = "foreign_key"
+
+    parent: Column[Any]
+
+    _table_column: Optional[Column[Any]]
+
+    def __init__(
+        self,
+        column: _DDLColumnArgument,
+        _constraint: Optional[ForeignKeyConstraint] = None,
+        use_alter: bool = False,
+        name: _ConstraintNameArgument = None,
+        onupdate: Optional[str] = None,
+        ondelete: Optional[str] = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        link_to_name: bool = False,
+        match: Optional[str] = None,
+        info: Optional[_InfoType] = None,
+        comment: Optional[str] = None,
+        _unresolvable: bool = False,
+        **dialect_kw: Any,
+    ):
+        r"""
+        Construct a column-level FOREIGN KEY.
+
+        The :class:`_schema.ForeignKey` object when constructed generates a
+        :class:`_schema.ForeignKeyConstraint`
+        which is associated with the parent
+        :class:`_schema.Table` object's collection of constraints.
+
+        :param column: A single target column for the key relationship. A
+            :class:`_schema.Column` object or a column name as a string:
+            ``tablename.columnkey`` or ``schema.tablename.columnkey``.
+            ``columnkey`` is the ``key`` which has been assigned to the column
+            (defaults to the column name itself), unless ``link_to_name`` is
+            ``True`` in which case the rendered name of the column is used.
+
+        :param name: Optional string. An in-database name for the key if
+            `constraint` is not provided.
+
+        :param onupdate: Optional string. If set, emit ON UPDATE <value> when
+            issuing DDL for this constraint. Typical values include CASCADE,
+            DELETE and RESTRICT.
+
+        :param ondelete: Optional string. If set, emit ON DELETE <value> when
+            issuing DDL for this constraint. Typical values include CASCADE,
+            DELETE and RESTRICT.
+
+        :param deferrable: Optional bool. If set, emit DEFERRABLE or NOT
+            DEFERRABLE when issuing DDL for this constraint.
+
+        :param initially: Optional string. If set, emit INITIALLY <value> when
+            issuing DDL for this constraint.
+
+        :param link_to_name: if True, the string name given in ``column`` is
+            the rendered name of the referenced column, not its locally
+            assigned ``key``.
+
+        :param use_alter: passed to the underlying
+            :class:`_schema.ForeignKeyConstraint`
+            to indicate the constraint should
+            be generated/dropped externally from the CREATE TABLE/ DROP TABLE
+            statement.  See :paramref:`_schema.ForeignKeyConstraint.use_alter`
+            for further description.
+
+            .. seealso::
+
+                :paramref:`_schema.ForeignKeyConstraint.use_alter`
+
+                :ref:`use_alter`
+
+        :param match: Optional string. If set, emit MATCH <value> when issuing
+            DDL for this constraint. Typical values include SIMPLE, PARTIAL
+            and FULL.
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param comment: Optional string that will render an SQL comment on
+          foreign key constraint creation.
+
+            .. versionadded:: 2.0
+
+        :param \**dialect_kw:  Additional keyword arguments are dialect
+            specific, and passed in the form ``<dialectname>_<argname>``.  The
+            arguments are ultimately handled by a corresponding
+            :class:`_schema.ForeignKeyConstraint`.
+            See the documentation regarding
+            an individual dialect at :ref:`dialect_toplevel` for detail on
+            documented arguments.
+
+        """
+
+        self._colspec = coercions.expect(roles.DDLReferredColumnRole, column)
+        self._unresolvable = _unresolvable
+
+        if isinstance(self._colspec, str):
+            self._table_column = None
+        else:
+            self._table_column = self._colspec
+
+            if not isinstance(
+                self._table_column.table, (type(None), TableClause)
+            ):
+                raise exc.ArgumentError(
+                    "ForeignKey received Column not bound "
+                    "to a Table, got: %r" % self._table_column.table
+                )
+
+        # the linked ForeignKeyConstraint.
+        # ForeignKey will create this when parent Column
+        # is attached to a Table, *or* ForeignKeyConstraint
+        # object passes itself in when creating ForeignKey
+        # markers.
+        self.constraint = _constraint
+
+        # .parent is not Optional under normal use
+        self.parent = None  # type: ignore
+
+        self.use_alter = use_alter
+        self.name = name
+        self.onupdate = onupdate
+        self.ondelete = ondelete
+        self.deferrable = deferrable
+        self.initially = initially
+        self.link_to_name = link_to_name
+        self.match = match
+        self.comment = comment
+        if info:
+            self.info = info
+        self._unvalidated_dialect_kw = dialect_kw
+
+    def __repr__(self) -> str:
+        return "ForeignKey(%r)" % self._get_colspec()
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.ForeignKey.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(self, *, schema: Optional[str] = None, **kw: Any) -> ForeignKey:
+        return self._copy(schema=schema, **kw)
+
+    def _copy(self, *, schema: Optional[str] = None, **kw: Any) -> ForeignKey:
+        """Produce a copy of this :class:`_schema.ForeignKey` object.
+
+        The new :class:`_schema.ForeignKey` will not be bound
+        to any :class:`_schema.Column`.
+
+        This method is usually used by the internal
+        copy procedures of :class:`_schema.Column`, :class:`_schema.Table`,
+        and :class:`_schema.MetaData`.
+
+        :param schema: The returned :class:`_schema.ForeignKey` will
+          reference the original table and column name, qualified
+          by the given string schema name.
+
+        """
+        fk = ForeignKey(
+            self._get_colspec(schema=schema),
+            use_alter=self.use_alter,
+            name=self.name,
+            onupdate=self.onupdate,
+            ondelete=self.ondelete,
+            deferrable=self.deferrable,
+            initially=self.initially,
+            link_to_name=self.link_to_name,
+            match=self.match,
+            comment=self.comment,
+            **self._unvalidated_dialect_kw,
+        )
+        return self._schema_item_copy(fk)
+
+    def _get_colspec(
+        self,
+        schema: Optional[
+            Union[
+                str,
+                Literal[SchemaConst.RETAIN_SCHEMA, SchemaConst.BLANK_SCHEMA],
+            ]
+        ] = None,
+        table_name: Optional[str] = None,
+        _is_copy: bool = False,
+    ) -> str:
+        """Return a string based 'column specification' for this
+        :class:`_schema.ForeignKey`.
+
+        This is usually the equivalent of the string-based "tablename.colname"
+        argument first passed to the object's constructor.
+
+        """
+        if schema not in (None, RETAIN_SCHEMA):
+            _schema, tname, colname = self._column_tokens
+            if table_name is not None:
+                tname = table_name
+            if schema is BLANK_SCHEMA:
+                return "%s.%s" % (tname, colname)
+            else:
+                return "%s.%s.%s" % (schema, tname, colname)
+        elif table_name:
+            schema, tname, colname = self._column_tokens
+            if schema:
+                return "%s.%s.%s" % (schema, table_name, colname)
+            else:
+                return "%s.%s" % (table_name, colname)
+        elif self._table_column is not None:
+            if self._table_column.table is None:
+                if _is_copy:
+                    raise exc.InvalidRequestError(
+                        f"Can't copy ForeignKey object which refers to "
+                        f"non-table bound Column {self._table_column!r}"
+                    )
+                else:
+                    return self._table_column.key
+            return "%s.%s" % (
+                self._table_column.table.fullname,
+                self._table_column.key,
+            )
+        else:
+            assert isinstance(self._colspec, str)
+            return self._colspec
+
+    @property
+    def _referred_schema(self) -> Optional[str]:
+        return self._column_tokens[0]
+
+    def _table_key(self) -> Any:
+        if self._table_column is not None:
+            if self._table_column.table is None:
+                return None
+            else:
+                return self._table_column.table.key
+        else:
+            schema, tname, colname = self._column_tokens
+            return _get_table_key(tname, schema)
+
+    target_fullname = property(_get_colspec)
+
+    def references(self, table: Table) -> bool:
+        """Return True if the given :class:`_schema.Table`
+        is referenced by this
+        :class:`_schema.ForeignKey`."""
+
+        return table.corresponding_column(self.column) is not None
+
+    def get_referent(self, table: FromClause) -> Optional[Column[Any]]:
+        """Return the :class:`_schema.Column` in the given
+        :class:`_schema.Table` (or any :class:`.FromClause`)
+        referenced by this :class:`_schema.ForeignKey`.
+
+        Returns None if this :class:`_schema.ForeignKey`
+        does not reference the given
+        :class:`_schema.Table`.
+
+        """
+        # our column is a Column, and any subquery etc. proxying us
+        # would be doing so via another Column, so that's what would
+        # be returned here
+        return table.columns.corresponding_column(self.column)  # type: ignore
+
+    @util.memoized_property
+    def _column_tokens(self) -> Tuple[Optional[str], str, Optional[str]]:
+        """parse a string-based _colspec into its component parts."""
+
+        m = self._get_colspec().split(".")
+        if m is None:
+            raise exc.ArgumentError(
+                f"Invalid foreign key column specification: {self._colspec}"
+            )
+        if len(m) == 1:
+            tname = m.pop()
+            colname = None
+        else:
+            colname = m.pop()
+            tname = m.pop()
+
+        # A FK between column 'bar' and table 'foo' can be
+        # specified as 'foo', 'foo.bar', 'dbo.foo.bar',
+        # 'otherdb.dbo.foo.bar'. Once we have the column name and
+        # the table name, treat everything else as the schema
+        # name. Some databases (e.g. Sybase) support
+        # inter-database foreign keys. See tickets#1341 and --
+        # indirectly related -- Ticket #594. This assumes that '.'
+        # will never appear *within* any component of the FK.
+
+        if len(m) > 0:
+            schema = ".".join(m)
+        else:
+            schema = None
+        return schema, tname, colname
+
+    def _resolve_col_tokens(self) -> Tuple[Table, str, Optional[str]]:
+        if self.parent is None:
+            raise exc.InvalidRequestError(
+                "this ForeignKey object does not yet have a "
+                "parent Column associated with it."
+            )
+
+        elif self.parent.table is None:
+            raise exc.InvalidRequestError(
+                "this ForeignKey's parent column is not yet associated "
+                "with a Table."
+            )
+
+        parenttable = self.parent.table
+
+        if self._unresolvable:
+            schema, tname, colname = self._column_tokens
+            tablekey = _get_table_key(tname, schema)
+            return parenttable, tablekey, colname
+
+        # assertion
+        # basically Column._make_proxy() sends the actual
+        # target Column to the ForeignKey object, so the
+        # string resolution here is never called.
+        for c in self.parent.base_columns:
+            if isinstance(c, Column):
+                assert c.table is parenttable
+                break
+        else:
+            assert False
+        ######################
+
+        schema, tname, colname = self._column_tokens
+
+        if schema is None and parenttable.metadata.schema is not None:
+            schema = parenttable.metadata.schema
+
+        tablekey = _get_table_key(tname, schema)
+        return parenttable, tablekey, colname
+
+    def _link_to_col_by_colstring(
+        self, parenttable: Table, table: Table, colname: Optional[str]
+    ) -> Column[Any]:
+        _column = None
+        if colname is None:
+            # colname is None in the case that ForeignKey argument
+            # was specified as table name only, in which case we
+            # match the column name to the same column on the
+            # parent.
+            # this use case wasn't working in later 1.x series
+            # as it had no test coverage; fixed in 2.0
+            parent = self.parent
+            assert parent is not None
+            key = parent.key
+            _column = table.c.get(key, None)
+        elif self.link_to_name:
+            key = colname
+            for c in table.c:
+                if c.name == colname:
+                    _column = c
+        else:
+            key = colname
+            _column = table.c.get(colname, None)
+
+        if _column is None:
+            raise exc.NoReferencedColumnError(
+                "Could not initialize target column "
+                f"for ForeignKey '{self._colspec}' "
+                f"on table '{parenttable.name}': "
+                f"table '{table.name}' has no column named '{key}'",
+                table.name,
+                key,
+            )
+
+        return _column
+
+    def _set_target_column(self, column: Column[Any]) -> None:
+        assert self.parent is not None
+
+        # propagate TypeEngine to parent if it didn't have one
+        if self.parent.type._isnull:
+            self.parent.type = column.type
+
+        # super-edgy case, if other FKs point to our column,
+        # they'd get the type propagated out also.
+
+        def set_type(fk: ForeignKey) -> None:
+            if fk.parent.type._isnull:
+                fk.parent.type = column.type
+
+        self.parent._setup_on_memoized_fks(set_type)
+
+        self.column = column  # type: ignore
+
+    @util.ro_memoized_property
+    def column(self) -> Column[Any]:
+        """Return the target :class:`_schema.Column` referenced by this
+        :class:`_schema.ForeignKey`.
+
+        If no target column has been established, an exception
+        is raised.
+
+        """
+
+        return self._resolve_column()
+
+    @overload
+    def _resolve_column(
+        self, *, raiseerr: Literal[True] = ...
+    ) -> Column[Any]: ...
+
+    @overload
+    def _resolve_column(
+        self, *, raiseerr: bool = ...
+    ) -> Optional[Column[Any]]: ...
+
+    def _resolve_column(
+        self, *, raiseerr: bool = True
+    ) -> Optional[Column[Any]]:
+        _column: Column[Any]
+
+        if isinstance(self._colspec, str):
+            parenttable, tablekey, colname = self._resolve_col_tokens()
+
+            if self._unresolvable or tablekey not in parenttable.metadata:
+                if not raiseerr:
+                    return None
+                raise exc.NoReferencedTableError(
+                    f"Foreign key associated with column "
+                    f"'{self.parent}' could not find "
+                    f"table '{tablekey}' with which to generate a "
+                    f"foreign key to target column '{colname}'",
+                    tablekey,
+                )
+            elif parenttable.key not in parenttable.metadata:
+                if not raiseerr:
+                    return None
+                raise exc.InvalidRequestError(
+                    f"Table {parenttable} is no longer associated with its "
+                    "parent MetaData"
+                )
+            else:
+                table = parenttable.metadata.tables[tablekey]
+                return self._link_to_col_by_colstring(
+                    parenttable, table, colname
+                )
+
+        elif hasattr(self._colspec, "__clause_element__"):
+            _column = self._colspec.__clause_element__()
+            return _column
+        else:
+            _column = self._colspec
+            return _column
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, Column)
+
+        if self.parent is not None and self.parent is not parent:
+            raise exc.InvalidRequestError(
+                "This ForeignKey already has a parent !"
+            )
+        self.parent = parent
+        self.parent.foreign_keys.add(self)
+        self.parent._on_table_attach(self._set_table)
+
+    def _set_remote_table(self, table: Table) -> None:
+        parenttable, _, colname = self._resolve_col_tokens()
+        _column = self._link_to_col_by_colstring(parenttable, table, colname)
+        self._set_target_column(_column)
+        assert self.constraint is not None
+        self.constraint._validate_dest_table(table)
+
+    def _remove_from_metadata(self, metadata: MetaData) -> None:
+        parenttable, table_key, colname = self._resolve_col_tokens()
+        fk_key = (table_key, colname)
+
+        if self in metadata._fk_memos[fk_key]:
+            # TODO: no test coverage for self not in memos
+            metadata._fk_memos[fk_key].remove(self)
+
+    def _set_table(self, column: Column[Any], table: Table) -> None:
+        # standalone ForeignKey - create ForeignKeyConstraint
+        # on the hosting Table when attached to the Table.
+        assert isinstance(table, Table)
+        if self.constraint is None:
+            self.constraint = ForeignKeyConstraint(
+                [],
+                [],
+                use_alter=self.use_alter,
+                name=self.name,
+                onupdate=self.onupdate,
+                ondelete=self.ondelete,
+                deferrable=self.deferrable,
+                initially=self.initially,
+                match=self.match,
+                comment=self.comment,
+                **self._unvalidated_dialect_kw,
+            )
+            self.constraint._append_element(column, self)
+            self.constraint._set_parent_with_dispatch(table)
+        table.foreign_keys.add(self)
+        # set up remote ".column" attribute, or a note to pick it
+        # up when the other Table/Column shows up
+        if isinstance(self._colspec, str):
+            parenttable, table_key, colname = self._resolve_col_tokens()
+            fk_key = (table_key, colname)
+            if table_key in parenttable.metadata.tables:
+                table = parenttable.metadata.tables[table_key]
+                try:
+                    _column = self._link_to_col_by_colstring(
+                        parenttable, table, colname
+                    )
+                except exc.NoReferencedColumnError:
+                    # this is OK, we'll try later
+                    pass
+                else:
+                    self._set_target_column(_column)
+
+            parenttable.metadata._fk_memos[fk_key].append(self)
+        elif hasattr(self._colspec, "__clause_element__"):
+            _column = self._colspec.__clause_element__()
+            self._set_target_column(_column)
+        else:
+            _column = self._colspec
+            self._set_target_column(_column)
+
+
+if TYPE_CHECKING:
+
+    def default_is_sequence(
+        obj: Optional[DefaultGenerator],
+    ) -> TypeGuard[Sequence]: ...
+
+    def default_is_clause_element(
+        obj: Optional[DefaultGenerator],
+    ) -> TypeGuard[ColumnElementColumnDefault]: ...
+
+    def default_is_scalar(
+        obj: Optional[DefaultGenerator],
+    ) -> TypeGuard[ScalarElementColumnDefault]: ...
+
+else:
+    default_is_sequence = operator.attrgetter("is_sequence")
+
+    default_is_clause_element = operator.attrgetter("is_clause_element")
+
+    default_is_scalar = operator.attrgetter("is_scalar")
+
+
+class DefaultGenerator(Executable, SchemaItem):
+    """Base class for column *default* values.
+
+    This object is only present on column.default or column.onupdate.
+    It's not valid as a server default.
+
+    """
+
+    __visit_name__ = "default_generator"
+
+    _is_default_generator = True
+    is_sequence = False
+    is_identity = False
+    is_server_default = False
+    is_clause_element = False
+    is_callable = False
+    is_scalar = False
+    has_arg = False
+    is_sentinel = False
+    column: Optional[Column[Any]]
+
+    def __init__(self, for_update: bool = False) -> None:
+        self.for_update = for_update
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        if TYPE_CHECKING:
+            assert isinstance(parent, Column)
+        self.column = parent
+        if self.for_update:
+            self.column.onupdate = self
+        else:
+            self.column.default = self
+
+    def _copy(self) -> DefaultGenerator:
+        raise NotImplementedError()
+
+    def _execute_on_connection(
+        self,
+        connection: Connection,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> Any:
+        util.warn_deprecated(
+            "Using the .execute() method to invoke a "
+            "DefaultGenerator object is deprecated; please use "
+            "the .scalar() method.",
+            "2.0",
+        )
+        return self._execute_on_scalar(
+            connection, distilled_params, execution_options
+        )
+
+    def _execute_on_scalar(
+        self,
+        connection: Connection,
+        distilled_params: _CoreMultiExecuteParams,
+        execution_options: CoreExecuteOptionsParameter,
+    ) -> Any:
+        return connection._execute_default(
+            self, distilled_params, execution_options
+        )
+
+
+class ColumnDefault(DefaultGenerator, ABC):
+    """A plain default value on a column.
+
+    This could correspond to a constant, a callable function,
+    or a SQL clause.
+
+    :class:`.ColumnDefault` is generated automatically
+    whenever the ``default``, ``onupdate`` arguments of
+    :class:`_schema.Column` are used.  A :class:`.ColumnDefault`
+    can be passed positionally as well.
+
+    For example, the following::
+
+        Column("foo", Integer, default=50)
+
+    Is equivalent to::
+
+        Column("foo", Integer, ColumnDefault(50))
+
+    """
+
+    arg: Any
+
+    @overload
+    def __new__(
+        cls, arg: Callable[..., Any], for_update: bool = ...
+    ) -> CallableColumnDefault: ...
+
+    @overload
+    def __new__(
+        cls, arg: ColumnElement[Any], for_update: bool = ...
+    ) -> ColumnElementColumnDefault: ...
+
+    # if I return ScalarElementColumnDefault here, which is what's actually
+    # returned, mypy complains that
+    # overloads overlap w/ incompatible return types.
+    @overload
+    def __new__(cls, arg: object, for_update: bool = ...) -> ColumnDefault: ...
+
+    def __new__(
+        cls, arg: Any = None, for_update: bool = False
+    ) -> ColumnDefault:
+        """Construct a new :class:`.ColumnDefault`.
+
+
+        :param arg: argument representing the default value.
+         May be one of the following:
+
+         * a plain non-callable Python value, such as a
+           string, integer, boolean, or other simple type.
+           The default value will be used as is each time.
+         * a SQL expression, that is one which derives from
+           :class:`_expression.ColumnElement`.  The SQL expression will
+           be rendered into the INSERT or UPDATE statement,
+           or in the case of a primary key column when
+           RETURNING is not used may be
+           pre-executed before an INSERT within a SELECT.
+         * A Python callable.  The function will be invoked for each
+           new row subject to an INSERT or UPDATE.
+           The callable must accept exactly
+           zero or one positional arguments.  The one-argument form
+           will receive an instance of the :class:`.ExecutionContext`,
+           which provides contextual information as to the current
+           :class:`_engine.Connection` in use as well as the current
+           statement and parameters.
+
+        """
+
+        if isinstance(arg, FetchedValue):
+            raise exc.ArgumentError(
+                "ColumnDefault may not be a server-side default type."
+            )
+        elif callable(arg):
+            cls = CallableColumnDefault
+        elif isinstance(arg, ClauseElement):
+            cls = ColumnElementColumnDefault
+        elif arg is not None:
+            cls = ScalarElementColumnDefault
+
+        return object.__new__(cls)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.arg!r})"
+
+
+class ScalarElementColumnDefault(ColumnDefault):
+    """default generator for a fixed scalar Python value
+
+    .. versionadded: 2.0
+
+    """
+
+    is_scalar = True
+    has_arg = True
+
+    def __init__(self, arg: Any, for_update: bool = False) -> None:
+        self.for_update = for_update
+        self.arg = arg
+
+    def _copy(self) -> ScalarElementColumnDefault:
+        return ScalarElementColumnDefault(
+            arg=self.arg, for_update=self.for_update
+        )
+
+
+class _InsertSentinelColumnDefault(ColumnDefault):
+    """Default generator that's specific to the use of a "sentinel" column
+    when using the insertmanyvalues feature.
+
+    This default is used as part of the :func:`_schema.insert_sentinel`
+    construct.
+
+    """
+
+    is_sentinel = True
+    for_update = False
+    arg = None
+
+    def __new__(cls) -> _InsertSentinelColumnDefault:
+        return object.__new__(cls)
+
+    def __init__(self) -> None:
+        pass
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        col = cast("Column[Any]", parent)
+        if not col._insert_sentinel:
+            raise exc.ArgumentError(
+                "The _InsertSentinelColumnDefault may only be applied to a "
+                "Column marked as insert_sentinel=True"
+            )
+        elif not col.nullable:
+            raise exc.ArgumentError(
+                "The _InsertSentinelColumnDefault may only be applied to a "
+                "Column that is nullable"
+            )
+
+        super()._set_parent(parent, **kw)
+
+    def _copy(self) -> _InsertSentinelColumnDefault:
+        return _InsertSentinelColumnDefault()
+
+
+_SQLExprDefault = Union["ColumnElement[Any]", "TextClause"]
+
+
+class ColumnElementColumnDefault(ColumnDefault):
+    """default generator for a SQL expression
+
+    .. versionadded:: 2.0
+
+    """
+
+    is_clause_element = True
+    has_arg = True
+    arg: _SQLExprDefault
+
+    def __init__(
+        self,
+        arg: _SQLExprDefault,
+        for_update: bool = False,
+    ) -> None:
+        self.for_update = for_update
+        self.arg = arg
+
+    def _copy(self) -> ColumnElementColumnDefault:
+        return ColumnElementColumnDefault(
+            arg=self.arg, for_update=self.for_update
+        )
+
+    @util.memoized_property
+    @util.preload_module("sqlalchemy.sql.sqltypes")
+    def _arg_is_typed(self) -> bool:
+        sqltypes = util.preloaded.sql_sqltypes
+
+        return not isinstance(self.arg.type, sqltypes.NullType)
+
+
+class _CallableColumnDefaultProtocol(Protocol):
+    def __call__(self, context: ExecutionContext) -> Any: ...
+
+
+class CallableColumnDefault(ColumnDefault):
+    """default generator for a callable Python function
+
+    .. versionadded:: 2.0
+
+    """
+
+    is_callable = True
+    arg: _CallableColumnDefaultProtocol
+    has_arg = True
+
+    def __init__(
+        self,
+        arg: Union[_CallableColumnDefaultProtocol, Callable[[], Any]],
+        for_update: bool = False,
+    ) -> None:
+        self.for_update = for_update
+        self.arg = self._maybe_wrap_callable(arg)
+
+    def _copy(self) -> CallableColumnDefault:
+        return CallableColumnDefault(arg=self.arg, for_update=self.for_update)
+
+    def _maybe_wrap_callable(
+        self, fn: Union[_CallableColumnDefaultProtocol, Callable[[], Any]]
+    ) -> _CallableColumnDefaultProtocol:
+        """Wrap callables that don't accept a context.
+
+        This is to allow easy compatibility with default callables
+        that aren't specific to accepting of a context.
+
+        """
+
+        try:
+            argspec = util.get_callable_argspec(fn, no_self=True)
+        except TypeError:
+            return util.wrap_callable(lambda ctx: fn(), fn)  # type: ignore
+
+        defaulted = argspec[3] is not None and len(argspec[3]) or 0
+        positionals = len(argspec[0]) - defaulted
+
+        if positionals == 0:
+            return util.wrap_callable(lambda ctx: fn(), fn)  # type: ignore
+
+        elif positionals == 1:
+            return fn  # type: ignore
+        else:
+            raise exc.ArgumentError(
+                "ColumnDefault Python function takes zero or one "
+                "positional arguments"
+            )
+
+
+class IdentityOptions:
+    """Defines options for a named database sequence or an identity column.
+
+    .. versionadded:: 1.3.18
+
+    .. seealso::
+
+        :class:`.Sequence`
+
+    """
+
+    def __init__(
+        self,
+        start: Optional[int] = None,
+        increment: Optional[int] = None,
+        minvalue: Optional[int] = None,
+        maxvalue: Optional[int] = None,
+        nominvalue: Optional[bool] = None,
+        nomaxvalue: Optional[bool] = None,
+        cycle: Optional[bool] = None,
+        cache: Optional[int] = None,
+        order: Optional[bool] = None,
+    ) -> None:
+        """Construct a :class:`.IdentityOptions` object.
+
+        See the :class:`.Sequence` documentation for a complete description
+        of the parameters.
+
+        :param start: the starting index of the sequence.
+        :param increment: the increment value of the sequence.
+        :param minvalue: the minimum value of the sequence.
+        :param maxvalue: the maximum value of the sequence.
+        :param nominvalue: no minimum value of the sequence.
+        :param nomaxvalue: no maximum value of the sequence.
+        :param cycle: allows the sequence to wrap around when the maxvalue
+         or minvalue has been reached.
+        :param cache: optional integer value; number of future values in the
+         sequence which are calculated in advance.
+        :param order: optional boolean value; if ``True``, renders the
+         ORDER keyword.
+
+        """
+        self.start = start
+        self.increment = increment
+        self.minvalue = minvalue
+        self.maxvalue = maxvalue
+        self.nominvalue = nominvalue
+        self.nomaxvalue = nomaxvalue
+        self.cycle = cycle
+        self.cache = cache
+        self.order = order
+
+    @property
+    def _increment_is_negative(self) -> bool:
+        return self.increment is not None and self.increment < 0
+
+
+class Sequence(HasSchemaAttr, IdentityOptions, DefaultGenerator):
+    """Represents a named database sequence.
+
+    The :class:`.Sequence` object represents the name and configurational
+    parameters of a database sequence.   It also represents
+    a construct that can be "executed" by a SQLAlchemy :class:`_engine.Engine`
+    or :class:`_engine.Connection`,
+    rendering the appropriate "next value" function
+    for the target database and returning a result.
+
+    The :class:`.Sequence` is typically associated with a primary key column::
+
+        some_table = Table(
+            "some_table",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                Sequence("some_table_seq", start=1),
+                primary_key=True,
+            ),
+        )
+
+    When CREATE TABLE is emitted for the above :class:`_schema.Table`, if the
+    target platform supports sequences, a CREATE SEQUENCE statement will
+    be emitted as well.   For platforms that don't support sequences,
+    the :class:`.Sequence` construct is ignored.
+
+    .. seealso::
+
+        :ref:`defaults_sequences`
+
+        :class:`.CreateSequence`
+
+        :class:`.DropSequence`
+
+    """
+
+    __visit_name__ = "sequence"
+
+    is_sequence = True
+
+    column: Optional[Column[Any]]
+    data_type: Optional[TypeEngine[int]]
+
+    def __init__(
+        self,
+        name: str,
+        start: Optional[int] = None,
+        increment: Optional[int] = None,
+        minvalue: Optional[int] = None,
+        maxvalue: Optional[int] = None,
+        nominvalue: Optional[bool] = None,
+        nomaxvalue: Optional[bool] = None,
+        cycle: Optional[bool] = None,
+        schema: Optional[Union[str, Literal[SchemaConst.BLANK_SCHEMA]]] = None,
+        cache: Optional[int] = None,
+        order: Optional[bool] = None,
+        data_type: Optional[_TypeEngineArgument[int]] = None,
+        optional: bool = False,
+        quote: Optional[bool] = None,
+        metadata: Optional[MetaData] = None,
+        quote_schema: Optional[bool] = None,
+        for_update: bool = False,
+    ) -> None:
+        """Construct a :class:`.Sequence` object.
+
+        :param name: the name of the sequence.
+
+        :param start: the starting index of the sequence.  This value is
+         used when the CREATE SEQUENCE command is emitted to the database
+         as the value of the "START WITH" clause. If ``None``, the
+         clause is omitted, which on most platforms indicates a starting
+         value of 1.
+
+         .. versionchanged:: 2.0 The :paramref:`.Sequence.start` parameter
+            is required in order to have DDL emit "START WITH".  This is a
+            reversal of a change made in version 1.4 which would implicitly
+            render "START WITH 1" if the :paramref:`.Sequence.start` were
+            not included.  See :ref:`change_7211` for more detail.
+
+        :param increment: the increment value of the sequence.  This
+         value is used when the CREATE SEQUENCE command is emitted to
+         the database as the value of the "INCREMENT BY" clause.  If ``None``,
+         the clause is omitted, which on most platforms indicates an
+         increment of 1.
+        :param minvalue: the minimum value of the sequence.  This
+         value is used when the CREATE SEQUENCE command is emitted to
+         the database as the value of the "MINVALUE" clause.  If ``None``,
+         the clause is omitted, which on most platforms indicates a
+         minvalue of 1 and -2^63-1 for ascending and descending sequences,
+         respectively.
+
+        :param maxvalue: the maximum value of the sequence.  This
+         value is used when the CREATE SEQUENCE command is emitted to
+         the database as the value of the "MAXVALUE" clause.  If ``None``,
+         the clause is omitted, which on most platforms indicates a
+         maxvalue of 2^63-1 and -1 for ascending and descending sequences,
+         respectively.
+
+        :param nominvalue: no minimum value of the sequence.  This
+         value is used when the CREATE SEQUENCE command is emitted to
+         the database as the value of the "NO MINVALUE" clause.  If ``None``,
+         the clause is omitted, which on most platforms indicates a
+         minvalue of 1 and -2^63-1 for ascending and descending sequences,
+         respectively.
+
+        :param nomaxvalue: no maximum value of the sequence.  This
+         value is used when the CREATE SEQUENCE command is emitted to
+         the database as the value of the "NO MAXVALUE" clause.  If ``None``,
+         the clause is omitted, which on most platforms indicates a
+         maxvalue of 2^63-1 and -1 for ascending and descending sequences,
+         respectively.
+
+        :param cycle: allows the sequence to wrap around when the maxvalue
+         or minvalue has been reached by an ascending or descending sequence
+         respectively.  This value is used when the CREATE SEQUENCE command
+         is emitted to the database as the "CYCLE" clause.  If the limit is
+         reached, the next number generated will be the minvalue or maxvalue,
+         respectively.  If cycle=False (the default) any calls to nextval
+         after the sequence has reached its maximum value will return an
+         error.
+
+        :param schema: optional schema name for the sequence, if located
+         in a schema other than the default.  The rules for selecting the
+         schema name when a :class:`_schema.MetaData`
+         is also present are the same
+         as that of :paramref:`_schema.Table.schema`.
+
+        :param cache: optional integer value; number of future values in the
+         sequence which are calculated in advance.  Renders the CACHE keyword
+         understood by Oracle Database and PostgreSQL.
+
+        :param order: optional boolean value; if ``True``, renders the
+         ORDER keyword, understood by Oracle Database, indicating the sequence
+         is definitively ordered.   May be necessary to provide deterministic
+         ordering using Oracle RAC.
+
+        :param data_type: The type to be returned by the sequence, for
+         dialects that allow us to choose between INTEGER, BIGINT, etc.
+         (e.g., mssql).
+
+         .. versionadded:: 1.4.0
+
+        :param optional: boolean value, when ``True``, indicates that this
+         :class:`.Sequence` object only needs to be explicitly generated
+         on backends that don't provide another way to generate primary
+         key identifiers.  Currently, it essentially means, "don't create
+         this sequence on the PostgreSQL backend, where the SERIAL keyword
+         creates a sequence for us automatically".
+        :param quote: boolean value, when ``True`` or ``False``, explicitly
+         forces quoting of the :paramref:`_schema.Sequence.name` on or off.
+         When left at its default of ``None``, normal quoting rules based
+         on casing and reserved words take place.
+        :param quote_schema: Set the quoting preferences for the ``schema``
+         name.
+
+        :param metadata: optional :class:`_schema.MetaData` object which this
+         :class:`.Sequence` will be associated with.  A :class:`.Sequence`
+         that is associated with a :class:`_schema.MetaData`
+         gains the following
+         capabilities:
+
+         * The :class:`.Sequence` will inherit the
+           :paramref:`_schema.MetaData.schema`
+           parameter specified to the target :class:`_schema.MetaData`, which
+           affects the production of CREATE / DROP DDL, if any.
+
+         * The :meth:`.Sequence.create` and :meth:`.Sequence.drop` methods
+           automatically use the engine bound to the :class:`_schema.MetaData`
+           object, if any.
+
+         * The :meth:`_schema.MetaData.create_all` and
+           :meth:`_schema.MetaData.drop_all`
+           methods will emit CREATE / DROP for this :class:`.Sequence`,
+           even if the :class:`.Sequence` is not associated with any
+           :class:`_schema.Table` / :class:`_schema.Column`
+           that's a member of this
+           :class:`_schema.MetaData`.
+
+         The above behaviors can only occur if the :class:`.Sequence` is
+         explicitly associated with the :class:`_schema.MetaData`
+         via this parameter.
+
+         .. seealso::
+
+            :ref:`sequence_metadata` - full discussion of the
+            :paramref:`.Sequence.metadata` parameter.
+
+        :param for_update: Indicates this :class:`.Sequence`, when associated
+         with a :class:`_schema.Column`,
+         should be invoked for UPDATE statements
+         on that column's table, rather than for INSERT statements, when
+         no value is otherwise present for that column in the statement.
+
+        """
+        DefaultGenerator.__init__(self, for_update=for_update)
+        IdentityOptions.__init__(
+            self,
+            start=start,
+            increment=increment,
+            minvalue=minvalue,
+            maxvalue=maxvalue,
+            nominvalue=nominvalue,
+            nomaxvalue=nomaxvalue,
+            cycle=cycle,
+            cache=cache,
+            order=order,
+        )
+        self.column = None
+        self.name = quoted_name(name, quote)
+        self.optional = optional
+        if schema is BLANK_SCHEMA:
+            self.schema = schema = None
+        elif metadata is not None and schema is None and metadata.schema:
+            self.schema = schema = metadata.schema
+        else:
+            self.schema = quoted_name.construct(schema, quote_schema)
+        self.metadata = metadata
+        self._key = _get_table_key(name, schema)
+        if metadata:
+            self._set_metadata(metadata)
+        if data_type is not None:
+            self.data_type = to_instance(data_type)
+        else:
+            self.data_type = None
+
+    @util.preload_module("sqlalchemy.sql.functions")
+    def next_value(self) -> Function[int]:
+        """Return a :class:`.next_value` function element
+        which will render the appropriate increment function
+        for this :class:`.Sequence` within any SQL expression.
+
+        """
+        return util.preloaded.sql_functions.func.next_value(self)
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        column = parent
+        assert isinstance(column, Column)
+        super()._set_parent(column)
+        column._on_table_attach(self._set_table)
+
+    def _copy(self) -> Sequence:
+        return Sequence(
+            name=self.name,
+            start=self.start,
+            increment=self.increment,
+            minvalue=self.minvalue,
+            maxvalue=self.maxvalue,
+            nominvalue=self.nominvalue,
+            nomaxvalue=self.nomaxvalue,
+            cycle=self.cycle,
+            schema=self.schema,
+            cache=self.cache,
+            order=self.order,
+            data_type=self.data_type,
+            optional=self.optional,
+            metadata=self.metadata,
+            for_update=self.for_update,
+        )
+
+    def _set_table(self, column: Column[Any], table: Table) -> None:
+        self._set_metadata(table.metadata)
+
+    def _set_metadata(self, metadata: MetaData) -> None:
+        self.metadata = metadata
+        self.metadata._sequences[self._key] = self
+
+    def create(self, bind: _CreateDropBind, checkfirst: bool = True) -> None:
+        """Creates this sequence in the database."""
+
+        bind._run_ddl_visitor(ddl.SchemaGenerator, self, checkfirst=checkfirst)
+
+    def drop(self, bind: _CreateDropBind, checkfirst: bool = True) -> None:
+        """Drops this sequence from the database."""
+
+        bind._run_ddl_visitor(ddl.SchemaDropper, self, checkfirst=checkfirst)
+
+    def _not_a_column_expr(self) -> NoReturn:
+        raise exc.InvalidRequestError(
+            f"This {self.__class__.__name__} cannot be used directly "
+            "as a column expression.  Use func.next_value(sequence) "
+            "to produce a 'next value' function that's usable "
+            "as a column element."
+        )
+
+
+@inspection._self_inspects
+class FetchedValue(SchemaEventTarget):
+    """A marker for a transparent database-side default.
+
+    Use :class:`.FetchedValue` when the database is configured
+    to provide some automatic default for a column.
+
+    E.g.::
+
+        Column("foo", Integer, FetchedValue())
+
+    Would indicate that some trigger or default generator
+    will create a new value for the ``foo`` column during an
+    INSERT.
+
+    .. seealso::
+
+        :ref:`triggered_columns`
+
+    """
+
+    is_server_default = True
+    reflected = False
+    has_argument = False
+    is_clause_element = False
+    is_identity = False
+
+    column: Optional[Column[Any]]
+
+    def __init__(self, for_update: bool = False) -> None:
+        self.for_update = for_update
+
+    def _as_for_update(self, for_update: bool) -> FetchedValue:
+        if for_update == self.for_update:
+            return self
+        else:
+            return self._clone(for_update)
+
+    def _copy(self) -> FetchedValue:
+        return FetchedValue(self.for_update)
+
+    def _clone(self, for_update: bool) -> Self:
+        n = self.__class__.__new__(self.__class__)
+        n.__dict__.update(self.__dict__)
+        n.__dict__.pop("column", None)
+        n.for_update = for_update
+        return n
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        column = parent
+        assert isinstance(column, Column)
+        self.column = column
+        if self.for_update:
+            self.column.server_onupdate = self
+        else:
+            self.column.server_default = self
+
+    def __repr__(self) -> str:
+        return util.generic_repr(self)
+
+
+class DefaultClause(FetchedValue):
+    """A DDL-specified DEFAULT column value.
+
+    :class:`.DefaultClause` is a :class:`.FetchedValue`
+    that also generates a "DEFAULT" clause when
+    "CREATE TABLE" is emitted.
+
+    :class:`.DefaultClause` is generated automatically
+    whenever the ``server_default``, ``server_onupdate`` arguments of
+    :class:`_schema.Column` are used.  A :class:`.DefaultClause`
+    can be passed positionally as well.
+
+    For example, the following::
+
+        Column("foo", Integer, server_default="50")
+
+    Is equivalent to::
+
+        Column("foo", Integer, DefaultClause("50"))
+
+    """
+
+    has_argument = True
+
+    def __init__(
+        self,
+        arg: Union[str, ClauseElement, TextClause],
+        for_update: bool = False,
+        _reflected: bool = False,
+    ) -> None:
+        util.assert_arg_type(arg, (str, ClauseElement, TextClause), "arg")
+        super().__init__(for_update)
+        self.arg = arg
+        self.reflected = _reflected
+
+    def _copy(self) -> DefaultClause:
+        return DefaultClause(
+            arg=self.arg, for_update=self.for_update, _reflected=self.reflected
+        )
+
+    def __repr__(self) -> str:
+        return "DefaultClause(%r, for_update=%r)" % (self.arg, self.for_update)
+
+
+class Constraint(DialectKWArgs, HasConditionalDDL, SchemaItem):
+    """A table-level SQL constraint.
+
+    :class:`_schema.Constraint` serves as the base class for the series of
+    constraint objects that can be associated with :class:`_schema.Table`
+    objects, including :class:`_schema.PrimaryKeyConstraint`,
+    :class:`_schema.ForeignKeyConstraint`
+    :class:`_schema.UniqueConstraint`, and
+    :class:`_schema.CheckConstraint`.
+
+    """
+
+    __visit_name__ = "constraint"
+
+    _creation_order: int
+    _column_flag: bool
+
+    def __init__(
+        self,
+        name: _ConstraintNameArgument = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        info: Optional[_InfoType] = None,
+        comment: Optional[str] = None,
+        _create_rule: Optional[Any] = None,
+        _type_bound: bool = False,
+        **dialect_kw: Any,
+    ) -> None:
+        r"""Create a SQL constraint.
+
+        :param name:
+          Optional, the in-database name of this ``Constraint``.
+
+        :param deferrable:
+          Optional bool.  If set, emit DEFERRABLE or NOT DEFERRABLE when
+          issuing DDL for this constraint.
+
+        :param initially:
+          Optional string.  If set, emit INITIALLY <value> when issuing DDL
+          for this constraint.
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param comment: Optional string that will render an SQL comment on
+          foreign key constraint creation.
+
+            .. versionadded:: 2.0
+
+        :param \**dialect_kw:  Additional keyword arguments are dialect
+            specific, and passed in the form ``<dialectname>_<argname>``.  See
+            the documentation regarding an individual dialect at
+            :ref:`dialect_toplevel` for detail on documented arguments.
+
+        :param _create_rule:
+          used internally by some datatypes that also create constraints.
+
+        :param _type_bound:
+          used internally to indicate that this constraint is associated with
+          a specific datatype.
+
+        """
+
+        self.name = name
+        self.deferrable = deferrable
+        self.initially = initially
+        if info:
+            self.info = info
+        self._create_rule = _create_rule
+        self._type_bound = _type_bound
+        util.set_creation_order(self)
+        self._validate_dialect_kwargs(dialect_kw)
+        self.comment = comment
+
+    def _should_create_for_compiler(
+        self, compiler: DDLCompiler, **kw: Any
+    ) -> bool:
+        if self._create_rule is not None and not self._create_rule(compiler):
+            return False
+        elif self._ddl_if is not None:
+            return self._ddl_if._should_execute(
+                ddl.CreateConstraint(self), self, None, compiler=compiler, **kw
+            )
+        else:
+            return True
+
+    @property
+    def table(self) -> Table:
+        try:
+            if isinstance(self.parent, Table):
+                return self.parent
+        except AttributeError:
+            pass
+        raise exc.InvalidRequestError(
+            "This constraint is not bound to a table.  Did you "
+            "mean to call table.append_constraint(constraint) ?"
+        )
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, (Table, Column))
+        self.parent = parent
+        parent.constraints.add(self)
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.Constraint.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(self, **kw: Any) -> Self:
+        return self._copy(**kw)
+
+    def _copy(self, **kw: Any) -> Self:
+        raise NotImplementedError()
+
+
+class ColumnCollectionMixin:
+    """A :class:`_expression.ColumnCollection` of :class:`_schema.Column`
+    objects.
+
+    This collection represents the columns which are referred to by
+    this object.
+
+    """
+
+    _columns: DedupeColumnCollection[Column[Any]]
+
+    _allow_multiple_tables = False
+
+    _pending_colargs: List[Optional[Union[str, Column[Any]]]]
+
+    if TYPE_CHECKING:
+
+        def _set_parent_with_dispatch(
+            self, parent: SchemaEventTarget, **kw: Any
+        ) -> None: ...
+
+    def __init__(
+        self,
+        *columns: _DDLColumnArgument,
+        _autoattach: bool = True,
+        _column_flag: bool = False,
+        _gather_expressions: Optional[
+            List[Union[str, ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        self._column_flag = _column_flag
+        self._columns = DedupeColumnCollection()
+
+        processed_expressions: Optional[
+            List[Union[ColumnElement[Any], str]]
+        ] = _gather_expressions
+
+        if processed_expressions is not None:
+
+            # this is expected to be an empty list
+            assert not processed_expressions
+
+            self._pending_colargs = []
+            for (
+                expr,
+                _,
+                _,
+                add_element,
+            ) in coercions.expect_col_expression_collection(
+                roles.DDLConstraintColumnRole, columns
+            ):
+                self._pending_colargs.append(add_element)
+                processed_expressions.append(expr)
+        else:
+            self._pending_colargs = [
+                coercions.expect(roles.DDLConstraintColumnRole, column)
+                for column in columns
+            ]
+
+        if _autoattach and self._pending_colargs:
+            self._check_attach()
+
+    def _check_attach(self, evt: bool = False) -> None:
+        col_objs = [c for c in self._pending_colargs if isinstance(c, Column)]
+
+        cols_w_table = [c for c in col_objs if isinstance(c.table, Table)]
+
+        cols_wo_table = set(col_objs).difference(cols_w_table)
+        if cols_wo_table:
+            # feature #3341 - place event listeners for Column objects
+            # such that when all those cols are attached, we autoattach.
+            assert not evt, "Should not reach here on event call"
+
+            # issue #3411 - don't do the per-column auto-attach if some of the
+            # columns are specified as strings.
+            has_string_cols = {
+                c for c in self._pending_colargs if c is not None
+            }.difference(col_objs)
+            if not has_string_cols:
+
+                def _col_attached(column: Column[Any], table: Table) -> None:
+                    # this isinstance() corresponds with the
+                    # isinstance() above; only want to count Table-bound
+                    # columns
+                    if isinstance(table, Table):
+                        cols_wo_table.discard(column)
+                        if not cols_wo_table:
+                            self._check_attach(evt=True)
+
+                self._cols_wo_table = cols_wo_table
+                for col in cols_wo_table:
+                    col._on_table_attach(_col_attached)
+                return
+
+        columns = cols_w_table
+
+        tables = {c.table for c in columns}
+        if len(tables) == 1:
+            self._set_parent_with_dispatch(tables.pop())
+        elif len(tables) > 1 and not self._allow_multiple_tables:
+            table = columns[0].table
+            others = [c for c in columns[1:] if c.table is not table]
+            if others:
+                # black could not format this inline
+                other_str = ", ".join("'%s'" % c for c in others)
+                raise exc.ArgumentError(
+                    f"Column(s) {other_str} "
+                    f"are not part of table '{table.description}'."
+                )
+
+    @util.ro_memoized_property
+    def columns(self) -> ReadOnlyColumnCollection[str, Column[Any]]:
+        return self._columns.as_readonly()
+
+    @util.ro_memoized_property
+    def c(self) -> ReadOnlyColumnCollection[str, Column[Any]]:
+        return self._columns.as_readonly()
+
+    def _col_expressions(
+        self, parent: Union[Table, Column[Any]]
+    ) -> List[Optional[Column[Any]]]:
+        if isinstance(parent, Column):
+            result: List[Optional[Column[Any]]] = [
+                c for c in self._pending_colargs if isinstance(c, Column)
+            ]
+            assert len(result) == len(self._pending_colargs)
+            return result
+        else:
+            try:
+                return [
+                    parent.c[col] if isinstance(col, str) else col
+                    for col in self._pending_colargs
+                ]
+            except KeyError as ke:
+                raise exc.ConstraintColumnNotFoundError(
+                    f"Can't create {self.__class__.__name__} "
+                    f"on table '{parent.description}': no column "
+                    f"named '{ke.args[0]}' is present."
+                ) from ke
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, (Table, Column))
+
+        for col in self._col_expressions(parent):
+            if col is not None:
+                self._columns.add(col)
+
+
+class ColumnCollectionConstraint(ColumnCollectionMixin, Constraint):
+    """A constraint that proxies a ColumnCollection."""
+
+    def __init__(
+        self,
+        *columns: _DDLColumnArgument,
+        name: _ConstraintNameArgument = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        info: Optional[_InfoType] = None,
+        _autoattach: bool = True,
+        _column_flag: bool = False,
+        _gather_expressions: Optional[List[_DDLColumnArgument]] = None,
+        **dialect_kw: Any,
+    ) -> None:
+        r"""
+        :param \*columns:
+          A sequence of column names or Column objects.
+
+        :param name:
+          Optional, the in-database name of this constraint.
+
+        :param deferrable:
+          Optional bool.  If set, emit DEFERRABLE or NOT DEFERRABLE when
+          issuing DDL for this constraint.
+
+        :param initially:
+          Optional string.  If set, emit INITIALLY <value> when issuing DDL
+          for this constraint.
+
+        :param \**dialect_kw: other keyword arguments including
+          dialect-specific arguments are propagated to the :class:`.Constraint`
+          superclass.
+
+        """
+        Constraint.__init__(
+            self,
+            name=name,
+            deferrable=deferrable,
+            initially=initially,
+            info=info,
+            **dialect_kw,
+        )
+        ColumnCollectionMixin.__init__(
+            self, *columns, _autoattach=_autoattach, _column_flag=_column_flag
+        )
+
+    columns: ReadOnlyColumnCollection[str, Column[Any]]
+    """A :class:`_expression.ColumnCollection` representing the set of columns
+    for this constraint.
+
+    """
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, (Column, Table))
+        Constraint._set_parent(self, parent)
+        ColumnCollectionMixin._set_parent(self, parent)
+
+    def __contains__(self, x: Any) -> bool:
+        return x in self._columns
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.ColumnCollectionConstraint.copy` method "
+        "is deprecated and will be removed in a future release.",
+    )
+    def copy(
+        self,
+        *,
+        target_table: Optional[Table] = None,
+        **kw: Any,
+    ) -> ColumnCollectionConstraint:
+        return self._copy(target_table=target_table, **kw)
+
+    def _copy(
+        self,
+        *,
+        target_table: Optional[Table] = None,
+        **kw: Any,
+    ) -> ColumnCollectionConstraint:
+        # ticket #5276
+        constraint_kwargs = {}
+        for dialect_name in self.dialect_options:
+            dialect_options = self.dialect_options[dialect_name]._non_defaults
+            for (
+                dialect_option_key,
+                dialect_option_value,
+            ) in dialect_options.items():
+                constraint_kwargs[dialect_name + "_" + dialect_option_key] = (
+                    dialect_option_value
+                )
+
+        assert isinstance(self.parent, Table)
+        c = self.__class__(
+            name=self.name,
+            deferrable=self.deferrable,
+            initially=self.initially,
+            *[
+                _copy_expression(expr, self.parent, target_table)
+                for expr in self._columns
+            ],
+            comment=self.comment,
+            **constraint_kwargs,
+        )
+        return self._schema_item_copy(c)
+
+    def contains_column(self, col: Column[Any]) -> bool:
+        """Return True if this constraint contains the given column.
+
+        Note that this object also contains an attribute ``.columns``
+        which is a :class:`_expression.ColumnCollection` of
+        :class:`_schema.Column` objects.
+
+        """
+
+        return self._columns.contains_column(col)
+
+    def __iter__(self) -> Iterator[Column[Any]]:
+        return iter(self._columns)
+
+    def __len__(self) -> int:
+        return len(self._columns)
+
+
+class CheckConstraint(ColumnCollectionConstraint):
+    """A table- or column-level CHECK constraint.
+
+    Can be included in the definition of a Table or Column.
+    """
+
+    _allow_multiple_tables = True
+
+    __visit_name__ = "table_or_column_check_constraint"
+
+    @_document_text_coercion(
+        "sqltext",
+        ":class:`.CheckConstraint`",
+        ":paramref:`.CheckConstraint.sqltext`",
+    )
+    def __init__(
+        self,
+        sqltext: _TextCoercedExpressionArgument[Any],
+        name: _ConstraintNameArgument = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        table: Optional[Table] = None,
+        info: Optional[_InfoType] = None,
+        _create_rule: Optional[Any] = None,
+        _autoattach: bool = True,
+        _type_bound: bool = False,
+        **dialect_kw: Any,
+    ) -> None:
+        r"""Construct a CHECK constraint.
+
+        :param sqltext:
+         A string containing the constraint definition, which will be used
+         verbatim, or a SQL expression construct.   If given as a string,
+         the object is converted to a :func:`_expression.text` object.
+         If the textual
+         string includes a colon character, escape this using a backslash::
+
+           CheckConstraint(r"foo ~ E'a(?\:b|c)d")
+
+        :param name:
+          Optional, the in-database name of the constraint.
+
+        :param deferrable:
+          Optional bool.  If set, emit DEFERRABLE or NOT DEFERRABLE when
+          issuing DDL for this constraint.
+
+        :param initially:
+          Optional string.  If set, emit INITIALLY <value> when issuing DDL
+          for this constraint.
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        """
+
+        self.sqltext = coercions.expect(roles.DDLExpressionRole, sqltext)
+        columns: List[Column[Any]] = []
+        visitors.traverse(self.sqltext, {}, {"column": columns.append})
+
+        super().__init__(
+            name=name,
+            deferrable=deferrable,
+            initially=initially,
+            _create_rule=_create_rule,
+            info=info,
+            _type_bound=_type_bound,
+            _autoattach=_autoattach,
+            *columns,
+            **dialect_kw,
+        )
+        if table is not None:
+            self._set_parent_with_dispatch(table)
+
+    @property
+    def is_column_level(self) -> bool:
+        return not isinstance(self.parent, Table)
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.CheckConstraint.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(
+        self, *, target_table: Optional[Table] = None, **kw: Any
+    ) -> CheckConstraint:
+        return self._copy(target_table=target_table, **kw)
+
+    def _copy(
+        self, *, target_table: Optional[Table] = None, **kw: Any
+    ) -> CheckConstraint:
+        if target_table is not None:
+            # note that target_table is None for the copy process of
+            # a column-bound CheckConstraint, so this path is not reached
+            # in that case.
+            sqltext = _copy_expression(self.sqltext, self.table, target_table)
+        else:
+            sqltext = self.sqltext
+        c = CheckConstraint(
+            sqltext,
+            name=self.name,
+            initially=self.initially,
+            deferrable=self.deferrable,
+            _create_rule=self._create_rule,
+            table=target_table,
+            comment=self.comment,
+            _autoattach=False,
+            _type_bound=self._type_bound,
+        )
+        return self._schema_item_copy(c)
+
+
+class ForeignKeyConstraint(ColumnCollectionConstraint):
+    """A table-level FOREIGN KEY constraint.
+
+    Defines a single column or composite FOREIGN KEY ... REFERENCES
+    constraint. For a no-frills, single column foreign key, adding a
+    :class:`_schema.ForeignKey` to the definition of a :class:`_schema.Column`
+    is a
+    shorthand equivalent for an unnamed, single column
+    :class:`_schema.ForeignKeyConstraint`.
+
+    Examples of foreign key configuration are in :ref:`metadata_foreignkeys`.
+
+    """
+
+    __visit_name__ = "foreign_key_constraint"
+
+    def __init__(
+        self,
+        columns: _typing_Sequence[_DDLColumnArgument],
+        refcolumns: _typing_Sequence[_DDLColumnArgument],
+        name: _ConstraintNameArgument = None,
+        onupdate: Optional[str] = None,
+        ondelete: Optional[str] = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        use_alter: bool = False,
+        link_to_name: bool = False,
+        match: Optional[str] = None,
+        table: Optional[Table] = None,
+        info: Optional[_InfoType] = None,
+        comment: Optional[str] = None,
+        **dialect_kw: Any,
+    ) -> None:
+        r"""Construct a composite-capable FOREIGN KEY.
+
+        :param columns: A sequence of local column names. The named columns
+          must be defined and present in the parent Table. The names should
+          match the ``key`` given to each column (defaults to the name) unless
+          ``link_to_name`` is True.
+
+        :param refcolumns: A sequence of foreign column names or Column
+          objects. The columns must all be located within the same Table.
+
+        :param name: Optional, the in-database name of the key.
+
+        :param onupdate: Optional string. If set, emit ON UPDATE <value> when
+          issuing DDL for this constraint. Typical values include CASCADE,
+          DELETE and RESTRICT.
+
+        :param ondelete: Optional string. If set, emit ON DELETE <value> when
+          issuing DDL for this constraint. Typical values include CASCADE,
+          DELETE and RESTRICT.
+
+        :param deferrable: Optional bool. If set, emit DEFERRABLE or NOT
+          DEFERRABLE when issuing DDL for this constraint.
+
+        :param initially: Optional string. If set, emit INITIALLY <value> when
+          issuing DDL for this constraint.
+
+        :param link_to_name: if True, the string name given in ``column`` is
+          the rendered name of the referenced column, not its locally assigned
+          ``key``.
+
+        :param use_alter: If True, do not emit the DDL for this constraint as
+          part of the CREATE TABLE definition. Instead, generate it via an
+          ALTER TABLE statement issued after the full collection of tables
+          have been created, and drop it via an ALTER TABLE statement before
+          the full collection of tables are dropped.
+
+          The use of :paramref:`_schema.ForeignKeyConstraint.use_alter` is
+          particularly geared towards the case where two or more tables
+          are established within a mutually-dependent foreign key constraint
+          relationship; however, the :meth:`_schema.MetaData.create_all` and
+          :meth:`_schema.MetaData.drop_all`
+          methods will perform this resolution
+          automatically, so the flag is normally not needed.
+
+          .. seealso::
+
+                :ref:`use_alter`
+
+        :param match: Optional string. If set, emit MATCH <value> when issuing
+          DDL for this constraint. Typical values include SIMPLE, PARTIAL
+          and FULL.
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param comment: Optional string that will render an SQL comment on
+          foreign key constraint creation.
+
+            .. versionadded:: 2.0
+
+        :param \**dialect_kw:  Additional keyword arguments are dialect
+          specific, and passed in the form ``<dialectname>_<argname>``.  See
+          the documentation regarding an individual dialect at
+          :ref:`dialect_toplevel` for detail on documented arguments.
+
+        """
+
+        Constraint.__init__(
+            self,
+            name=name,
+            deferrable=deferrable,
+            initially=initially,
+            info=info,
+            comment=comment,
+            **dialect_kw,
+        )
+        self.onupdate = onupdate
+        self.ondelete = ondelete
+        self.link_to_name = link_to_name
+        self.use_alter = use_alter
+        self.match = match
+
+        if len(set(columns)) != len(refcolumns):
+            if len(set(columns)) != len(columns):
+                # e.g. FOREIGN KEY (a, a) REFERENCES r (b, c)
+                raise exc.ArgumentError(
+                    "ForeignKeyConstraint with duplicate source column "
+                    "references are not supported."
+                )
+            else:
+                # e.g. FOREIGN KEY (a) REFERENCES r (b, c)
+                # paraphrasing
+                # https://www.postgresql.org/docs/current/static/ddl-constraints.html
+                raise exc.ArgumentError(
+                    "ForeignKeyConstraint number "
+                    "of constrained columns must match the number of "
+                    "referenced columns."
+                )
+
+        # standalone ForeignKeyConstraint - create
+        # associated ForeignKey objects which will be applied to hosted
+        # Column objects (in col.foreign_keys), either now or when attached
+        # to the Table for string-specified names
+        self.elements = [
+            ForeignKey(
+                refcol,
+                _constraint=self,
+                name=self.name,
+                onupdate=self.onupdate,
+                ondelete=self.ondelete,
+                use_alter=self.use_alter,
+                link_to_name=self.link_to_name,
+                match=self.match,
+                deferrable=self.deferrable,
+                initially=self.initially,
+                **self.dialect_kwargs,
+            )
+            for refcol in refcolumns
+        ]
+
+        ColumnCollectionMixin.__init__(self, *columns)
+        if table is not None:
+            if hasattr(self, "parent"):
+                assert table is self.parent
+            self._set_parent_with_dispatch(table)
+
+    def _append_element(self, column: Column[Any], fk: ForeignKey) -> None:
+        self._columns.add(column)
+        self.elements.append(fk)
+
+    columns: ReadOnlyColumnCollection[str, Column[Any]]
+    """A :class:`_expression.ColumnCollection` representing the set of columns
+    for this constraint.
+
+    """
+
+    elements: List[ForeignKey]
+    """A sequence of :class:`_schema.ForeignKey` objects.
+
+    Each :class:`_schema.ForeignKey`
+    represents a single referring column/referred
+    column pair.
+
+    This collection is intended to be read-only.
+
+    """
+
+    @property
+    def _elements(self) -> util.OrderedDict[str, ForeignKey]:
+        # legacy - provide a dictionary view of (column_key, fk)
+        return util.OrderedDict(zip(self.column_keys, self.elements))
+
+    @property
+    def _referred_schema(self) -> Optional[str]:
+        for elem in self.elements:
+            return elem._referred_schema
+        else:
+            return None
+
+    @property
+    def referred_table(self) -> Table:
+        """The :class:`_schema.Table` object to which this
+        :class:`_schema.ForeignKeyConstraint` references.
+
+        This is a dynamically calculated attribute which may not be available
+        if the constraint and/or parent table is not yet associated with
+        a metadata collection that contains the referred table.
+
+        """
+        return self.elements[0].column.table
+
+    def _validate_dest_table(self, table: Table) -> None:
+        table_keys = {elem._table_key() for elem in self.elements}
+        if None not in table_keys and len(table_keys) > 1:
+            elem0, elem1 = sorted(table_keys)[0:2]
+            raise exc.ArgumentError(
+                f"ForeignKeyConstraint on "
+                f"{table.fullname}({self._col_description}) refers to "
+                f"multiple remote tables: {elem0} and {elem1}"
+            )
+
+    @property
+    def column_keys(self) -> _typing_Sequence[str]:
+        """Return a list of string keys representing the local
+        columns in this :class:`_schema.ForeignKeyConstraint`.
+
+        This list is either the original string arguments sent
+        to the constructor of the :class:`_schema.ForeignKeyConstraint`,
+        or if the constraint has been initialized with :class:`_schema.Column`
+        objects, is the string ``.key`` of each element.
+
+        """
+        if hasattr(self, "parent"):
+            return self._columns.keys()
+        else:
+            return [
+                col.key if isinstance(col, ColumnElement) else str(col)
+                for col in self._pending_colargs
+            ]
+
+    @property
+    def _col_description(self) -> str:
+        return ", ".join(self.column_keys)
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        table = parent
+        assert isinstance(table, Table)
+        Constraint._set_parent(self, table)
+
+        ColumnCollectionConstraint._set_parent(self, table)
+
+        for col, fk in zip(self._columns, self.elements):
+            if not hasattr(fk, "parent") or fk.parent is not col:
+                fk._set_parent_with_dispatch(col)
+
+        self._validate_dest_table(table)
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.ForeignKeyConstraint.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(
+        self,
+        *,
+        schema: Optional[str] = None,
+        target_table: Optional[Table] = None,
+        **kw: Any,
+    ) -> ForeignKeyConstraint:
+        return self._copy(schema=schema, target_table=target_table, **kw)
+
+    def _copy(
+        self,
+        *,
+        schema: Optional[str] = None,
+        target_table: Optional[Table] = None,
+        **kw: Any,
+    ) -> ForeignKeyConstraint:
+        fkc = ForeignKeyConstraint(
+            [x.parent.key for x in self.elements],
+            [
+                x._get_colspec(
+                    schema=schema,
+                    table_name=(
+                        target_table.name
+                        if target_table is not None
+                        and x._table_key() == x.parent.table.key
+                        else None
+                    ),
+                    _is_copy=True,
+                )
+                for x in self.elements
+            ],
+            name=self.name,
+            onupdate=self.onupdate,
+            ondelete=self.ondelete,
+            use_alter=self.use_alter,
+            deferrable=self.deferrable,
+            initially=self.initially,
+            link_to_name=self.link_to_name,
+            match=self.match,
+            comment=self.comment,
+        )
+        for self_fk, other_fk in zip(self.elements, fkc.elements):
+            self_fk._schema_item_copy(other_fk)
+        return self._schema_item_copy(fkc)
+
+
+class PrimaryKeyConstraint(ColumnCollectionConstraint):
+    """A table-level PRIMARY KEY constraint.
+
+    The :class:`.PrimaryKeyConstraint` object is present automatically
+    on any :class:`_schema.Table` object; it is assigned a set of
+    :class:`_schema.Column` objects corresponding to those marked with
+    the :paramref:`_schema.Column.primary_key` flag::
+
+        >>> my_table = Table(
+        ...     "mytable",
+        ...     metadata,
+        ...     Column("id", Integer, primary_key=True),
+        ...     Column("version_id", Integer, primary_key=True),
+        ...     Column("data", String(50)),
+        ... )
+        >>> my_table.primary_key
+        PrimaryKeyConstraint(
+            Column('id', Integer(), table=<mytable>,
+                   primary_key=True, nullable=False),
+            Column('version_id', Integer(), table=<mytable>,
+                   primary_key=True, nullable=False)
+        )
+
+    The primary key of a :class:`_schema.Table` can also be specified by using
+    a :class:`.PrimaryKeyConstraint` object explicitly; in this mode of usage,
+    the "name" of the constraint can also be specified, as well as other
+    options which may be recognized by dialects::
+
+        my_table = Table(
+            "mytable",
+            metadata,
+            Column("id", Integer),
+            Column("version_id", Integer),
+            Column("data", String(50)),
+            PrimaryKeyConstraint("id", "version_id", name="mytable_pk"),
+        )
+
+    The two styles of column-specification should generally not be mixed.
+    An warning is emitted if the columns present in the
+    :class:`.PrimaryKeyConstraint`
+    don't match the columns that were marked as ``primary_key=True``, if both
+    are present; in this case, the columns are taken strictly from the
+    :class:`.PrimaryKeyConstraint` declaration, and those columns otherwise
+    marked as ``primary_key=True`` are ignored.  This behavior is intended to
+    be backwards compatible with previous behavior.
+
+    For the use case where specific options are to be specified on the
+    :class:`.PrimaryKeyConstraint`, but the usual style of using
+    ``primary_key=True`` flags is still desirable, an empty
+    :class:`.PrimaryKeyConstraint` may be specified, which will take on the
+    primary key column collection from the :class:`_schema.Table` based on the
+    flags::
+
+        my_table = Table(
+            "mytable",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("version_id", Integer, primary_key=True),
+            Column("data", String(50)),
+            PrimaryKeyConstraint(name="mytable_pk", mssql_clustered=True),
+        )
+
+    """
+
+    __visit_name__ = "primary_key_constraint"
+
+    def __init__(
+        self,
+        *columns: _DDLColumnArgument,
+        name: Optional[str] = None,
+        deferrable: Optional[bool] = None,
+        initially: Optional[str] = None,
+        info: Optional[_InfoType] = None,
+        _implicit_generated: bool = False,
+        **dialect_kw: Any,
+    ) -> None:
+        self._implicit_generated = _implicit_generated
+        super().__init__(
+            *columns,
+            name=name,
+            deferrable=deferrable,
+            initially=initially,
+            info=info,
+            **dialect_kw,
+        )
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        table = parent
+        assert isinstance(table, Table)
+        super()._set_parent(table)
+
+        if table.primary_key is not self:
+            table.constraints.discard(table.primary_key)
+            table.primary_key = self  # type: ignore
+            table.constraints.add(self)
+
+        table_pks = [c for c in table.c if c.primary_key]
+        if (
+            self._columns
+            and table_pks
+            and set(table_pks) != set(self._columns)
+        ):
+            # black could not format these inline
+            table_pk_str = ", ".join("'%s'" % c.name for c in table_pks)
+            col_str = ", ".join("'%s'" % c.name for c in self._columns)
+
+            util.warn(
+                f"Table '{table.name}' specifies columns "
+                f"{table_pk_str} as "
+                f"primary_key=True, "
+                f"not matching locally specified columns {col_str}; "
+                f"setting the "
+                f"current primary key columns to "
+                f"{col_str}. "
+                f"This warning "
+                f"may become an exception in a future release"
+            )
+            table_pks[:] = []
+
+        for c in self._columns:
+            c.primary_key = True
+            if c._user_defined_nullable is NULL_UNSPECIFIED:
+                c.nullable = False
+        if table_pks:
+            self._columns.extend(table_pks)
+
+    def _reload(self, columns: Iterable[Column[Any]]) -> None:
+        """repopulate this :class:`.PrimaryKeyConstraint` given
+        a set of columns.
+
+        Existing columns in the table that are marked as primary_key=True
+        are maintained.
+
+        Also fires a new event.
+
+        This is basically like putting a whole new
+        :class:`.PrimaryKeyConstraint` object on the parent
+        :class:`_schema.Table` object without actually replacing the object.
+
+        The ordering of the given list of columns is also maintained; these
+        columns will be appended to the list of columns after any which
+        are already present.
+
+        """
+        # set the primary key flag on new columns.
+        # note any existing PK cols on the table also have their
+        # flag still set.
+        for col in columns:
+            col.primary_key = True
+
+        self._columns.extend(columns)
+
+        PrimaryKeyConstraint._autoincrement_column._reset(self)  # type: ignore
+        self._set_parent_with_dispatch(self.table)
+
+    def _replace(self, col: Column[Any]) -> None:
+        PrimaryKeyConstraint._autoincrement_column._reset(self)  # type: ignore
+        self._columns.replace(col)
+
+        self.dispatch._sa_event_column_added_to_pk_constraint(self, col)
+
+    @property
+    def columns_autoinc_first(self) -> List[Column[Any]]:
+        autoinc = self._autoincrement_column
+
+        if autoinc is not None:
+            return [autoinc] + [c for c in self._columns if c is not autoinc]
+        else:
+            return list(self._columns)
+
+    @util.ro_memoized_property
+    def _autoincrement_column(self) -> Optional[Column[int]]:
+        def _validate_autoinc(col: Column[Any], autoinc_true: bool) -> bool:
+            if col.type._type_affinity is None or not issubclass(
+                col.type._type_affinity,
+                (
+                    type_api.INTEGERTYPE._type_affinity,
+                    type_api.NUMERICTYPE._type_affinity,
+                ),
+            ):
+                if autoinc_true:
+                    raise exc.ArgumentError(
+                        f"Column type {col.type} on column '{col}' is not "
+                        f"compatible with autoincrement=True"
+                    )
+                else:
+                    return False
+            elif (
+                not isinstance(col.default, (type(None), Sequence))
+                and not autoinc_true
+            ):
+                return False
+            elif (
+                col.server_default is not None
+                and not isinstance(col.server_default, Identity)
+                and not autoinc_true
+            ):
+                return False
+            elif col.foreign_keys and col.autoincrement not in (
+                True,
+                "ignore_fk",
+            ):
+                return False
+            return True
+
+        if len(self._columns) == 1:
+            col = list(self._columns)[0]
+
+            if col.autoincrement is True:
+                _validate_autoinc(col, True)
+                return col
+            elif col.autoincrement in (
+                "auto",
+                "ignore_fk",
+            ) and _validate_autoinc(col, False):
+                return col
+            else:
+                return None
+
+        else:
+            autoinc = None
+            for col in self._columns:
+                if col.autoincrement is True:
+                    _validate_autoinc(col, True)
+                    if autoinc is not None:
+                        raise exc.ArgumentError(
+                            f"Only one Column may be marked "
+                            f"autoincrement=True, found both "
+                            f"{col.name} and {autoinc.name}."
+                        )
+                    else:
+                        autoinc = col
+
+            return autoinc
+
+
+class UniqueConstraint(ColumnCollectionConstraint):
+    """A table-level UNIQUE constraint.
+
+    Defines a single column or composite UNIQUE constraint. For a no-frills,
+    single column constraint, adding ``unique=True`` to the ``Column``
+    definition is a shorthand equivalent for an unnamed, single column
+    UniqueConstraint.
+    """
+
+    __visit_name__ = "unique_constraint"
+
+
+class Index(
+    DialectKWArgs, ColumnCollectionMixin, HasConditionalDDL, SchemaItem
+):
+    """A table-level INDEX.
+
+    Defines a composite (one or more column) INDEX.
+
+    E.g.::
+
+        sometable = Table(
+            "sometable",
+            metadata,
+            Column("name", String(50)),
+            Column("address", String(100)),
+        )
+
+        Index("some_index", sometable.c.name)
+
+    For a no-frills, single column index, adding
+    :class:`_schema.Column` also supports ``index=True``::
+
+        sometable = Table(
+            "sometable", metadata, Column("name", String(50), index=True)
+        )
+
+    For a composite index, multiple columns can be specified::
+
+        Index("some_index", sometable.c.name, sometable.c.address)
+
+    Functional indexes are supported as well, typically by using the
+    :data:`.func` construct in conjunction with table-bound
+    :class:`_schema.Column` objects::
+
+        Index("some_index", func.lower(sometable.c.name))
+
+    An :class:`.Index` can also be manually associated with a
+    :class:`_schema.Table`,
+    either through inline declaration or using
+    :meth:`_schema.Table.append_constraint`.  When this approach is used,
+    the names
+    of the indexed columns can be specified as strings::
+
+        Table(
+            "sometable",
+            metadata,
+            Column("name", String(50)),
+            Column("address", String(100)),
+            Index("some_index", "name", "address"),
+        )
+
+    To support functional or expression-based indexes in this form, the
+    :func:`_expression.text` construct may be used::
+
+        from sqlalchemy import text
+
+        Table(
+            "sometable",
+            metadata,
+            Column("name", String(50)),
+            Column("address", String(100)),
+            Index("some_index", text("lower(name)")),
+        )
+
+    .. seealso::
+
+        :ref:`schema_indexes` - General information on :class:`.Index`.
+
+        :ref:`postgresql_indexes` - PostgreSQL-specific options available for
+        the :class:`.Index` construct.
+
+        :ref:`mysql_indexes` - MySQL-specific options available for the
+        :class:`.Index` construct.
+
+        :ref:`mssql_indexes` - MSSQL-specific options available for the
+        :class:`.Index` construct.
+
+    """
+
+    __visit_name__ = "index"
+
+    table: Optional[Table]
+    expressions: _typing_Sequence[Union[str, ColumnElement[Any]]]
+    _table_bound_expressions: _typing_Sequence[ColumnElement[Any]]
+
+    def __init__(
+        self,
+        name: Optional[str],
+        *expressions: _DDLColumnArgument,
+        unique: bool = False,
+        quote: Optional[bool] = None,
+        info: Optional[_InfoType] = None,
+        _table: Optional[Table] = None,
+        _column_flag: bool = False,
+        **dialect_kw: Any,
+    ) -> None:
+        r"""Construct an index object.
+
+        :param name:
+          The name of the index
+
+        :param \*expressions:
+          Column expressions to include in the index.   The expressions
+          are normally instances of :class:`_schema.Column`, but may also
+          be arbitrary SQL expressions which ultimately refer to a
+          :class:`_schema.Column`.
+
+        :param unique=False:
+            Keyword only argument; if True, create a unique index.
+
+        :param quote=None:
+            Keyword only argument; whether to apply quoting to the name of
+            the index.  Works in the same manner as that of
+            :paramref:`_schema.Column.quote`.
+
+        :param info=None: Optional data dictionary which will be populated
+            into the :attr:`.SchemaItem.info` attribute of this object.
+
+        :param \**dialect_kw: Additional keyword arguments not mentioned above
+            are dialect specific, and passed in the form
+            ``<dialectname>_<argname>``. See the documentation regarding an
+            individual dialect at :ref:`dialect_toplevel` for detail on
+            documented arguments.
+
+        """
+        self.table = table = None
+
+        self.name = quoted_name.construct(name, quote)
+        self.unique = unique
+        if info is not None:
+            self.info = info
+
+        # TODO: consider "table" argument being public, but for
+        # the purpose of the fix here, it starts as private.
+        if _table is not None:
+            table = _table
+
+        self._validate_dialect_kwargs(dialect_kw)
+
+        self.expressions = []
+        # will call _set_parent() if table-bound column
+        # objects are present
+        ColumnCollectionMixin.__init__(
+            self,
+            *expressions,
+            _column_flag=_column_flag,
+            _gather_expressions=self.expressions,
+        )
+        if table is not None:
+            self._set_parent(table)
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        table = parent
+        assert isinstance(table, Table)
+        ColumnCollectionMixin._set_parent(self, table)
+
+        if self.table is not None and table is not self.table:
+            raise exc.ArgumentError(
+                f"Index '{self.name}' is against table "
+                f"'{self.table.description}', and "
+                f"cannot be associated with table '{table.description}'."
+            )
+        self.table = table
+        table.indexes.add(self)
+
+        expressions = self.expressions
+        col_expressions = self._col_expressions(table)
+        assert len(expressions) == len(col_expressions)
+
+        exprs = []
+        for expr, colexpr in zip(expressions, col_expressions):
+            if isinstance(expr, ClauseElement):
+                exprs.append(expr)
+            elif colexpr is not None:
+                exprs.append(colexpr)
+            else:
+                assert False
+        self.expressions = self._table_bound_expressions = exprs
+
+    def create(self, bind: _CreateDropBind, checkfirst: bool = False) -> None:
+        """Issue a ``CREATE`` statement for this
+        :class:`.Index`, using the given
+        :class:`.Connection` or :class:`.Engine`` for connectivity.
+
+        .. seealso::
+
+            :meth:`_schema.MetaData.create_all`.
+
+        """
+        bind._run_ddl_visitor(ddl.SchemaGenerator, self, checkfirst=checkfirst)
+
+    def drop(self, bind: _CreateDropBind, checkfirst: bool = False) -> None:
+        """Issue a ``DROP`` statement for this
+        :class:`.Index`, using the given
+        :class:`.Connection` or :class:`.Engine` for connectivity.
+
+        .. seealso::
+
+            :meth:`_schema.MetaData.drop_all`.
+
+        """
+        bind._run_ddl_visitor(ddl.SchemaDropper, self, checkfirst=checkfirst)
+
+    def __repr__(self) -> str:
+        exprs: _typing_Sequence[Any]  # noqa: F842
+
+        return "Index(%s)" % (
+            ", ".join(
+                [repr(self.name)]
+                + [repr(e) for e in self.expressions]
+                + (self.unique and ["unique=True"] or [])
+            )
+        )
+
+
+_NamingSchemaCallable = Callable[[Constraint, Table], str]
+_NamingSchemaDirective = Union[str, _NamingSchemaCallable]
+
+
+class _NamingSchemaTD(TypedDict, total=False):
+    fk: _NamingSchemaDirective
+    pk: _NamingSchemaDirective
+    ix: _NamingSchemaDirective
+    ck: _NamingSchemaDirective
+    uq: _NamingSchemaDirective
+
+
+_NamingSchemaParameter = Union[
+    # it seems like the TypedDict here is useful for pylance typeahead,
+    # and not much else
+    _NamingSchemaTD,
+    # there is no form that allows Union[Type[Any], str] to work in all
+    # cases, including breaking out Mapping[] entries for each combination
+    # even, therefore keys must be `Any` (see #10264)
+    Mapping[Any, _NamingSchemaDirective],
+]
+
+
+DEFAULT_NAMING_CONVENTION: _NamingSchemaParameter = util.immutabledict(
+    {"ix": "ix_%(column_0_label)s"}
+)
+
+
+class MetaData(HasSchemaAttr):
+    """A collection of :class:`_schema.Table`
+    objects and their associated schema
+    constructs.
+
+    Holds a collection of :class:`_schema.Table` objects as well as
+    an optional binding to an :class:`_engine.Engine` or
+    :class:`_engine.Connection`.  If bound, the :class:`_schema.Table` objects
+    in the collection and their columns may participate in implicit SQL
+    execution.
+
+    The :class:`_schema.Table` objects themselves are stored in the
+    :attr:`_schema.MetaData.tables` dictionary.
+
+    :class:`_schema.MetaData` is a thread-safe object for read operations.
+    Construction of new tables within a single :class:`_schema.MetaData`
+    object,
+    either explicitly or via reflection, may not be completely thread-safe.
+
+    .. seealso::
+
+        :ref:`metadata_describing` - Introduction to database metadata
+
+    """
+
+    __visit_name__ = "metadata"
+
+    def __init__(
+        self,
+        schema: Optional[str] = None,
+        quote_schema: Optional[bool] = None,
+        naming_convention: Optional[_NamingSchemaParameter] = None,
+        info: Optional[_InfoType] = None,
+    ) -> None:
+        """Create a new MetaData object.
+
+        :param schema:
+           The default schema to use for the :class:`_schema.Table`,
+           :class:`.Sequence`, and potentially other objects associated with
+           this :class:`_schema.MetaData`. Defaults to ``None``.
+
+           .. seealso::
+
+                :ref:`schema_metadata_schema_name` - details on how the
+                :paramref:`_schema.MetaData.schema` parameter is used.
+
+                :paramref:`_schema.Table.schema`
+
+                :paramref:`.Sequence.schema`
+
+        :param quote_schema:
+            Sets the ``quote_schema`` flag for those :class:`_schema.Table`,
+            :class:`.Sequence`, and other objects which make usage of the
+            local ``schema`` name.
+
+        :param info: Optional data dictionary which will be populated into the
+            :attr:`.SchemaItem.info` attribute of this object.
+
+        :param naming_convention: a dictionary referring to values which
+          will establish default naming conventions for :class:`.Constraint`
+          and :class:`.Index` objects, for those objects which are not given
+          a name explicitly.
+
+          The keys of this dictionary may be:
+
+          * a constraint or Index class, e.g. the :class:`.UniqueConstraint`,
+            :class:`_schema.ForeignKeyConstraint` class, the :class:`.Index`
+            class
+
+          * a string mnemonic for one of the known constraint classes;
+            ``"fk"``, ``"pk"``, ``"ix"``, ``"ck"``, ``"uq"`` for foreign key,
+            primary key, index, check, and unique constraint, respectively.
+
+          * the string name of a user-defined "token" that can be used
+            to define new naming tokens.
+
+          The values associated with each "constraint class" or "constraint
+          mnemonic" key are string naming templates, such as
+          ``"uq_%(table_name)s_%(column_0_name)s"``,
+          which describe how the name should be composed.  The values
+          associated with user-defined "token" keys should be callables of the
+          form ``fn(constraint, table)``, which accepts the constraint/index
+          object and :class:`_schema.Table` as arguments, returning a string
+          result.
+
+          The built-in names are as follows, some of which may only be
+          available for certain types of constraint:
+
+            * ``%(table_name)s`` - the name of the :class:`_schema.Table`
+              object
+              associated with the constraint.
+
+            * ``%(referred_table_name)s`` - the name of the
+              :class:`_schema.Table`
+              object associated with the referencing target of a
+              :class:`_schema.ForeignKeyConstraint`.
+
+            * ``%(column_0_name)s`` - the name of the :class:`_schema.Column`
+              at
+              index position "0" within the constraint.
+
+            * ``%(column_0N_name)s`` - the name of all :class:`_schema.Column`
+              objects in order within the constraint, joined without a
+              separator.
+
+            * ``%(column_0_N_name)s`` - the name of all
+              :class:`_schema.Column`
+              objects in order within the constraint, joined with an
+              underscore as a separator.
+
+            * ``%(column_0_label)s``, ``%(column_0N_label)s``,
+              ``%(column_0_N_label)s`` - the label of either the zeroth
+              :class:`_schema.Column` or all :class:`.Columns`, separated with
+              or without an underscore
+
+            * ``%(column_0_key)s``, ``%(column_0N_key)s``,
+              ``%(column_0_N_key)s`` - the key of either the zeroth
+              :class:`_schema.Column` or all :class:`.Columns`, separated with
+              or without an underscore
+
+            * ``%(referred_column_0_name)s``, ``%(referred_column_0N_name)s``
+              ``%(referred_column_0_N_name)s``,  ``%(referred_column_0_key)s``,
+              ``%(referred_column_0N_key)s``, ...  column tokens which
+              render the names/keys/labels of columns that are referenced
+              by a  :class:`_schema.ForeignKeyConstraint`.
+
+            * ``%(constraint_name)s`` - a special key that refers to the
+              existing name given to the constraint.  When this key is
+              present, the :class:`.Constraint` object's existing name will be
+              replaced with one that is composed from template string that
+              uses this token. When this token is present, it is required that
+              the :class:`.Constraint` is given an explicit name ahead of time.
+
+            * user-defined: any additional token may be implemented by passing
+              it along with a ``fn(constraint, table)`` callable to the
+              naming_convention dictionary.
+
+          .. versionadded:: 1.3.0 - added new ``%(column_0N_name)s``,
+             ``%(column_0_N_name)s``, and related tokens that produce
+             concatenations of names, keys, or labels for all columns referred
+             to by a given constraint.
+
+          .. seealso::
+
+                :ref:`constraint_naming_conventions` - for detailed usage
+                examples.
+
+        """
+        if schema is not None and not isinstance(schema, str):
+            raise exc.ArgumentError(
+                "expected schema argument to be a string, "
+                f"got {type(schema)}."
+            )
+        self.tables = util.FacadeDict()
+        self.schema = quoted_name.construct(schema, quote_schema)
+        self.naming_convention = (
+            naming_convention
+            if naming_convention
+            else DEFAULT_NAMING_CONVENTION
+        )
+        if info:
+            self.info = info
+        self._schemas: Set[str] = set()
+        self._sequences: Dict[str, Sequence] = {}
+        self._fk_memos: Dict[Tuple[str, Optional[str]], List[ForeignKey]] = (
+            collections.defaultdict(list)
+        )
+
+    tables: util.FacadeDict[str, Table]
+    """A dictionary of :class:`_schema.Table`
+    objects keyed to their name or "table key".
+
+    The exact key is that determined by the :attr:`_schema.Table.key`
+    attribute;
+    for a table with no :attr:`_schema.Table.schema` attribute,
+    this is the same
+    as :attr:`_schema.Table.name`.  For a table with a schema,
+    it is typically of the
+    form ``schemaname.tablename``.
+
+    .. seealso::
+
+        :attr:`_schema.MetaData.sorted_tables`
+
+    """
+
+    def __repr__(self) -> str:
+        return "MetaData()"
+
+    def __contains__(self, table_or_key: Union[str, Table]) -> bool:
+        if not isinstance(table_or_key, str):
+            table_or_key = table_or_key.key
+        return table_or_key in self.tables
+
+    def _add_table(
+        self, name: str, schema: Optional[str], table: Table
+    ) -> None:
+        key = _get_table_key(name, schema)
+        self.tables._insert_item(key, table)
+        if schema:
+            self._schemas.add(schema)
+
+    def _remove_table(self, name: str, schema: Optional[str]) -> None:
+        key = _get_table_key(name, schema)
+        removed = dict.pop(self.tables, key, None)
+        if removed is not None:
+            for fk in removed.foreign_keys:
+                fk._remove_from_metadata(self)
+        if self._schemas:
+            self._schemas = {
+                t.schema for t in self.tables.values() if t.schema is not None
+            }
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {
+            "tables": self.tables,
+            "schema": self.schema,
+            "schemas": self._schemas,
+            "sequences": self._sequences,
+            "fk_memos": self._fk_memos,
+            "naming_convention": self.naming_convention,
+        }
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        self.tables = state["tables"]
+        self.schema = state["schema"]
+        self.naming_convention = state["naming_convention"]
+        self._sequences = state["sequences"]
+        self._schemas = state["schemas"]
+        self._fk_memos = state["fk_memos"]
+
+    def clear(self) -> None:
+        """Clear all Table objects from this MetaData."""
+
+        dict.clear(self.tables)  # type: ignore
+        self._schemas.clear()
+        self._fk_memos.clear()
+
+    def remove(self, table: Table) -> None:
+        """Remove the given Table object from this MetaData."""
+
+        self._remove_table(table.name, table.schema)
+
+    @property
+    def sorted_tables(self) -> List[Table]:
+        """Returns a list of :class:`_schema.Table` objects sorted in order of
+        foreign key dependency.
+
+        The sorting will place :class:`_schema.Table`
+        objects that have dependencies
+        first, before the dependencies themselves, representing the
+        order in which they can be created.   To get the order in which
+        the tables would be dropped, use the ``reversed()`` Python built-in.
+
+        .. warning::
+
+            The :attr:`.MetaData.sorted_tables` attribute cannot by itself
+            accommodate automatic resolution of dependency cycles between
+            tables, which are usually caused by mutually dependent foreign key
+            constraints. When these cycles are detected, the foreign keys
+            of these tables are omitted from consideration in the sort.
+            A warning is emitted when this condition occurs, which will be an
+            exception raise in a future release.   Tables which are not part
+            of the cycle will still be returned in dependency order.
+
+            To resolve these cycles, the
+            :paramref:`_schema.ForeignKeyConstraint.use_alter` parameter may be
+            applied to those constraints which create a cycle.  Alternatively,
+            the :func:`_schema.sort_tables_and_constraints` function will
+            automatically return foreign key constraints in a separate
+            collection when cycles are detected so that they may be applied
+            to a schema separately.
+
+            .. versionchanged:: 1.3.17 - a warning is emitted when
+               :attr:`.MetaData.sorted_tables` cannot perform a proper sort
+               due to cyclical dependencies.  This will be an exception in a
+               future release.  Additionally, the sort will continue to return
+               other tables not involved in the cycle in dependency order which
+               was not the case previously.
+
+        .. seealso::
+
+            :func:`_schema.sort_tables`
+
+            :func:`_schema.sort_tables_and_constraints`
+
+            :attr:`_schema.MetaData.tables`
+
+            :meth:`_reflection.Inspector.get_table_names`
+
+            :meth:`_reflection.Inspector.get_sorted_table_and_fkc_names`
+
+
+        """
+        return ddl.sort_tables(
+            sorted(self.tables.values(), key=lambda t: t.key)  # type: ignore
+        )
+
+    # overload needed to work around mypy this mypy
+    # https://github.com/python/mypy/issues/17093
+    @overload
+    def reflect(
+        self,
+        bind: Engine,
+        schema: Optional[str] = ...,
+        views: bool = ...,
+        only: Union[
+            _typing_Sequence[str], Callable[[str, MetaData], bool], None
+        ] = ...,
+        extend_existing: bool = ...,
+        autoload_replace: bool = ...,
+        resolve_fks: bool = ...,
+        **dialect_kwargs: Any,
+    ) -> None: ...
+
+    @overload
+    def reflect(
+        self,
+        bind: Connection,
+        schema: Optional[str] = ...,
+        views: bool = ...,
+        only: Union[
+            _typing_Sequence[str], Callable[[str, MetaData], bool], None
+        ] = ...,
+        extend_existing: bool = ...,
+        autoload_replace: bool = ...,
+        resolve_fks: bool = ...,
+        **dialect_kwargs: Any,
+    ) -> None: ...
+
+    @util.preload_module("sqlalchemy.engine.reflection")
+    def reflect(
+        self,
+        bind: Union[Engine, Connection],
+        schema: Optional[str] = None,
+        views: bool = False,
+        only: Union[
+            _typing_Sequence[str], Callable[[str, MetaData], bool], None
+        ] = None,
+        extend_existing: bool = False,
+        autoload_replace: bool = True,
+        resolve_fks: bool = True,
+        **dialect_kwargs: Any,
+    ) -> None:
+        r"""Load all available table definitions from the database.
+
+        Automatically creates ``Table`` entries in this ``MetaData`` for any
+        table available in the database but not yet present in the
+        ``MetaData``.  May be called multiple times to pick up tables recently
+        added to the database, however no special action is taken if a table
+        in this ``MetaData`` no longer exists in the database.
+
+        :param bind:
+          A :class:`.Connection` or :class:`.Engine` used to access the
+          database.
+
+        :param schema:
+          Optional, query and reflect tables from an alternate schema.
+          If None, the schema associated with this :class:`_schema.MetaData`
+          is used, if any.
+
+        :param views:
+          If True, also reflect views (materialized and plain).
+
+        :param only:
+          Optional.  Load only a sub-set of available named tables.  May be
+          specified as a sequence of names or a callable.
+
+          If a sequence of names is provided, only those tables will be
+          reflected.  An error is raised if a table is requested but not
+          available.  Named tables already present in this ``MetaData`` are
+          ignored.
+
+          If a callable is provided, it will be used as a boolean predicate to
+          filter the list of potential table names.  The callable is called
+          with a table name and this ``MetaData`` instance as positional
+          arguments and should return a true value for any table to reflect.
+
+        :param extend_existing: Passed along to each :class:`_schema.Table` as
+          :paramref:`_schema.Table.extend_existing`.
+
+        :param autoload_replace: Passed along to each :class:`_schema.Table`
+          as
+          :paramref:`_schema.Table.autoload_replace`.
+
+        :param resolve_fks: if True, reflect :class:`_schema.Table`
+         objects linked
+         to :class:`_schema.ForeignKey` objects located in each
+         :class:`_schema.Table`.
+         For :meth:`_schema.MetaData.reflect`,
+         this has the effect of reflecting
+         related tables that might otherwise not be in the list of tables
+         being reflected, for example if the referenced table is in a
+         different schema or is omitted via the
+         :paramref:`.MetaData.reflect.only` parameter.  When False,
+         :class:`_schema.ForeignKey` objects are not followed to the
+         :class:`_schema.Table`
+         in which they link, however if the related table is also part of the
+         list of tables that would be reflected in any case, the
+         :class:`_schema.ForeignKey` object will still resolve to its related
+         :class:`_schema.Table` after the :meth:`_schema.MetaData.reflect`
+         operation is
+         complete.   Defaults to True.
+
+         .. versionadded:: 1.3.0
+
+         .. seealso::
+
+            :paramref:`_schema.Table.resolve_fks`
+
+        :param \**dialect_kwargs: Additional keyword arguments not mentioned
+         above are dialect specific, and passed in the form
+         ``<dialectname>_<argname>``.  See the documentation regarding an
+         individual dialect at :ref:`dialect_toplevel` for detail on
+         documented arguments.
+
+        .. seealso::
+
+            :ref:`metadata_reflection_toplevel`
+
+            :meth:`_events.DDLEvents.column_reflect` - Event used to customize
+            the reflected columns. Usually used to generalize the types using
+            :meth:`_types.TypeEngine.as_generic`
+
+            :ref:`metadata_reflection_dbagnostic_types` - describes how to
+            reflect tables using general types.
+
+        """
+
+        with inspection.inspect(bind)._inspection_context() as insp:
+            reflect_opts: Any = {
+                "autoload_with": insp,
+                "extend_existing": extend_existing,
+                "autoload_replace": autoload_replace,
+                "resolve_fks": resolve_fks,
+                "_extend_on": set(),
+            }
+
+            reflect_opts.update(dialect_kwargs)
+
+            if schema is None:
+                schema = self.schema
+
+            if schema is not None:
+                reflect_opts["schema"] = schema
+
+            kind = util.preloaded.engine_reflection.ObjectKind.TABLE
+            available: util.OrderedSet[str] = util.OrderedSet(
+                insp.get_table_names(schema)
+            )
+            if views:
+                kind = util.preloaded.engine_reflection.ObjectKind.ANY
+                available.update(insp.get_view_names(schema))
+                try:
+                    available.update(insp.get_materialized_view_names(schema))
+                except NotImplementedError:
+                    pass
+
+            if schema is not None:
+                available_w_schema: util.OrderedSet[str] = util.OrderedSet(
+                    [f"{schema}.{name}" for name in available]
+                )
+            else:
+                available_w_schema = available
+
+            current = set(self.tables)
+
+            if only is None:
+                load = [
+                    name
+                    for name, schname in zip(available, available_w_schema)
+                    if extend_existing or schname not in current
+                ]
+            elif callable(only):
+                load = [
+                    name
+                    for name, schname in zip(available, available_w_schema)
+                    if (extend_existing or schname not in current)
+                    and only(name, self)
+                ]
+            else:
+                missing = [name for name in only if name not in available]
+                if missing:
+                    s = schema and (" schema '%s'" % schema) or ""
+                    missing_str = ", ".join(missing)
+                    raise exc.InvalidRequestError(
+                        f"Could not reflect: requested table(s) not available "
+                        f"in {bind.engine!r}{s}: ({missing_str})"
+                    )
+                load = [
+                    name
+                    for name in only
+                    if extend_existing or name not in current
+                ]
+            # pass the available tables so the inspector can
+            # choose to ignore the filter_names
+            _reflect_info = insp._get_reflection_info(
+                schema=schema,
+                filter_names=load,
+                available=available,
+                kind=kind,
+                scope=util.preloaded.engine_reflection.ObjectScope.ANY,
+                **dialect_kwargs,
+            )
+            reflect_opts["_reflect_info"] = _reflect_info
+
+            for name in load:
+                try:
+                    Table(name, self, **reflect_opts)
+                except exc.UnreflectableTableError as uerr:
+                    util.warn(f"Skipping table {name}: {uerr}")
+
+    def create_all(
+        self,
+        bind: _CreateDropBind,
+        tables: Optional[_typing_Sequence[Table]] = None,
+        checkfirst: bool = True,
+    ) -> None:
+        """Create all tables stored in this metadata.
+
+        Conditional by default, will not attempt to recreate tables already
+        present in the target database.
+
+        :param bind:
+          A :class:`.Connection` or :class:`.Engine` used to access the
+          database.
+
+        :param tables:
+          Optional list of ``Table`` objects, which is a subset of the total
+          tables in the ``MetaData`` (others are ignored).
+
+        :param checkfirst:
+          Defaults to True, don't issue CREATEs for tables already present
+          in the target database.
+
+        """
+        bind._run_ddl_visitor(
+            ddl.SchemaGenerator, self, checkfirst=checkfirst, tables=tables
+        )
+
+    def drop_all(
+        self,
+        bind: _CreateDropBind,
+        tables: Optional[_typing_Sequence[Table]] = None,
+        checkfirst: bool = True,
+    ) -> None:
+        """Drop all tables stored in this metadata.
+
+        Conditional by default, will not attempt to drop tables not present in
+        the target database.
+
+        :param bind:
+          A :class:`.Connection` or :class:`.Engine` used to access the
+          database.
+
+        :param tables:
+          Optional list of ``Table`` objects, which is a subset of the
+          total tables in the ``MetaData`` (others are ignored).
+
+        :param checkfirst:
+          Defaults to True, only issue DROPs for tables confirmed to be
+          present in the target database.
+
+        """
+        bind._run_ddl_visitor(
+            ddl.SchemaDropper, self, checkfirst=checkfirst, tables=tables
+        )
+
+
+class Computed(FetchedValue, SchemaItem):
+    """Defines a generated column, i.e. "GENERATED ALWAYS AS" syntax.
+
+    The :class:`.Computed` construct is an inline construct added to the
+    argument list of a :class:`_schema.Column` object::
+
+        from sqlalchemy import Computed
+
+        Table(
+            "square",
+            metadata_obj,
+            Column("side", Float, nullable=False),
+            Column("area", Float, Computed("side * side")),
+        )
+
+    See the linked documentation below for complete details.
+
+    .. versionadded:: 1.3.11
+
+    .. seealso::
+
+        :ref:`computed_ddl`
+
+    """
+
+    __visit_name__ = "computed_column"
+
+    column: Optional[Column[Any]]
+
+    @_document_text_coercion(
+        "sqltext", ":class:`.Computed`", ":paramref:`.Computed.sqltext`"
+    )
+    def __init__(
+        self, sqltext: _DDLColumnArgument, persisted: Optional[bool] = None
+    ) -> None:
+        """Construct a GENERATED ALWAYS AS DDL construct to accompany a
+        :class:`_schema.Column`.
+
+        :param sqltext:
+          A string containing the column generation expression, which will be
+          used verbatim, or a SQL expression construct, such as a
+          :func:`_expression.text`
+          object.   If given as a string, the object is converted to a
+          :func:`_expression.text` object.
+
+        :param persisted:
+          Optional, controls how this column should be persisted by the
+          database.   Possible values are:
+
+          * ``None``, the default, it will use the default persistence
+            defined by the database.
+          * ``True``, will render ``GENERATED ALWAYS AS ... STORED``, or the
+            equivalent for the target database if supported.
+          * ``False``, will render ``GENERATED ALWAYS AS ... VIRTUAL``, or
+            the equivalent for the target database if supported.
+
+          Specifying ``True`` or ``False`` may raise an error when the DDL
+          is emitted to the target database if the database does not support
+          that persistence option.   Leaving this parameter at its default
+          of ``None`` is guaranteed to succeed for all databases that support
+          ``GENERATED ALWAYS AS``.
+
+        """
+        self.sqltext = coercions.expect(roles.DDLExpressionRole, sqltext)
+        self.persisted = persisted
+        self.column = None
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, Column)
+
+        if not isinstance(
+            parent.server_default, (type(None), Computed)
+        ) or not isinstance(parent.server_onupdate, (type(None), Computed)):
+            raise exc.ArgumentError(
+                "A generated column cannot specify a server_default or a "
+                "server_onupdate argument"
+            )
+        self.column = parent
+        parent.computed = self
+        self.column.server_onupdate = self
+        self.column.server_default = self
+
+    def _as_for_update(self, for_update: bool) -> FetchedValue:
+        return self
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.Computed.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(
+        self, *, target_table: Optional[Table] = None, **kw: Any
+    ) -> Computed:
+        return self._copy(target_table=target_table, **kw)
+
+    def _copy(
+        self, *, target_table: Optional[Table] = None, **kw: Any
+    ) -> Computed:
+        sqltext = _copy_expression(
+            self.sqltext,
+            self.column.table if self.column is not None else None,
+            target_table,
+        )
+        g = Computed(sqltext, persisted=self.persisted)
+
+        return self._schema_item_copy(g)
+
+
+class Identity(IdentityOptions, FetchedValue, SchemaItem):
+    """Defines an identity column, i.e. "GENERATED { ALWAYS | BY DEFAULT }
+    AS IDENTITY" syntax.
+
+    The :class:`.Identity` construct is an inline construct added to the
+    argument list of a :class:`_schema.Column` object::
+
+        from sqlalchemy import Identity
+
+        Table(
+            "foo",
+            metadata_obj,
+            Column("id", Integer, Identity()),
+            Column("description", Text),
+        )
+
+    See the linked documentation below for complete details.
+
+    .. versionadded:: 1.4
+
+    .. seealso::
+
+        :ref:`identity_ddl`
+
+    """
+
+    __visit_name__ = "identity_column"
+
+    is_identity = True
+
+    def __init__(
+        self,
+        always: bool = False,
+        on_null: Optional[bool] = None,
+        start: Optional[int] = None,
+        increment: Optional[int] = None,
+        minvalue: Optional[int] = None,
+        maxvalue: Optional[int] = None,
+        nominvalue: Optional[bool] = None,
+        nomaxvalue: Optional[bool] = None,
+        cycle: Optional[bool] = None,
+        cache: Optional[int] = None,
+        order: Optional[bool] = None,
+    ) -> None:
+        """Construct a GENERATED { ALWAYS | BY DEFAULT } AS IDENTITY DDL
+        construct to accompany a :class:`_schema.Column`.
+
+        See the :class:`.Sequence` documentation for a complete description
+        of most parameters.
+
+        .. note::
+            MSSQL supports this construct as the preferred alternative to
+            generate an IDENTITY on a column, but it uses non standard
+            syntax that only support :paramref:`_schema.Identity.start`
+            and :paramref:`_schema.Identity.increment`.
+            All other parameters are ignored.
+
+        :param always:
+          A boolean, that indicates the type of identity column.
+          If ``False`` is specified, the default, then the user-specified
+          value takes precedence.
+          If ``True`` is specified, a user-specified value is not accepted (
+          on some backends, like PostgreSQL, OVERRIDING SYSTEM VALUE, or
+          similar, may be specified in an INSERT to override the sequence
+          value).
+          Some backends also have a default value for this parameter,
+          ``None`` can be used to omit rendering this part in the DDL. It
+          will be treated as ``False`` if a backend does not have a default
+          value.
+
+        :param on_null:
+          Set to ``True`` to specify ON NULL in conjunction with a
+          ``always=False`` identity column. This option is only supported on
+          some backends, like Oracle Database.
+
+        :param start: the starting index of the sequence.
+        :param increment: the increment value of the sequence.
+        :param minvalue: the minimum value of the sequence.
+        :param maxvalue: the maximum value of the sequence.
+        :param nominvalue: no minimum value of the sequence.
+        :param nomaxvalue: no maximum value of the sequence.
+        :param cycle: allows the sequence to wrap around when the maxvalue
+         or minvalue has been reached.
+        :param cache: optional integer value; number of future values in the
+         sequence which are calculated in advance.
+        :param order: optional boolean value; if true, renders the
+         ORDER keyword.
+
+        """
+        IdentityOptions.__init__(
+            self,
+            start=start,
+            increment=increment,
+            minvalue=minvalue,
+            maxvalue=maxvalue,
+            nominvalue=nominvalue,
+            nomaxvalue=nomaxvalue,
+            cycle=cycle,
+            cache=cache,
+            order=order,
+        )
+        self.always = always
+        self.on_null = on_null
+        self.column = None
+
+    def _set_parent(self, parent: SchemaEventTarget, **kw: Any) -> None:
+        assert isinstance(parent, Column)
+        if not isinstance(
+            parent.server_default, (type(None), Identity)
+        ) or not isinstance(parent.server_onupdate, type(None)):
+            raise exc.ArgumentError(
+                "A column with an Identity object cannot specify a "
+                "server_default or a server_onupdate argument"
+            )
+        if parent.autoincrement is False:
+            raise exc.ArgumentError(
+                "A column with an Identity object cannot specify "
+                "autoincrement=False"
+            )
+        self.column = parent
+
+        parent.identity = self
+        if parent._user_defined_nullable is NULL_UNSPECIFIED:
+            parent.nullable = False
+
+        parent.server_default = self
+
+    def _as_for_update(self, for_update: bool) -> FetchedValue:
+        return self
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_schema.Identity.copy` method is deprecated "
+        "and will be removed in a future release.",
+    )
+    def copy(self, **kw: Any) -> Identity:
+        return self._copy(**kw)
+
+    def _copy(self, **kw: Any) -> Identity:
+        i = Identity(
+            always=self.always,
+            on_null=self.on_null,
+            start=self.start,
+            increment=self.increment,
+            minvalue=self.minvalue,
+            maxvalue=self.maxvalue,
+            nominvalue=self.nominvalue,
+            nomaxvalue=self.nomaxvalue,
+            cycle=self.cycle,
+            cache=self.cache,
+            order=self.order,
+        )
+
+        return self._schema_item_copy(i)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/selectable.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/selectable.py
new file mode 100644
index 00000000..8aa9f41e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/selectable.py
@@ -0,0 +1,7179 @@
+# sql/selectable.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""The :class:`_expression.FromClause` class of SQL expression elements,
+representing
+SQL tables and derived rowsets.
+
+"""
+
+from __future__ import annotations
+
+import collections
+from enum import Enum
+import itertools
+from typing import AbstractSet
+from typing import Any as TODO_Any
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import NamedTuple
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import cache_key
+from . import coercions
+from . import operators
+from . import roles
+from . import traversals
+from . import type_api
+from . import visitors
+from ._typing import _ColumnsClauseArgument
+from ._typing import _no_kw
+from ._typing import _TP
+from ._typing import is_column_element
+from ._typing import is_select_statement
+from ._typing import is_subquery
+from ._typing import is_table
+from ._typing import is_text_clause
+from .annotation import Annotated
+from .annotation import SupportsCloneAnnotations
+from .base import _clone
+from .base import _cloned_difference
+from .base import _cloned_intersection
+from .base import _entity_namespace_key
+from .base import _EntityNamespace
+from .base import _expand_cloned
+from .base import _from_objects
+from .base import _generative
+from .base import _never_select_column
+from .base import _NoArg
+from .base import _select_iterables
+from .base import CacheableOptions
+from .base import ColumnCollection
+from .base import ColumnSet
+from .base import CompileState
+from .base import DedupeColumnCollection
+from .base import Executable
+from .base import Generative
+from .base import HasCompileState
+from .base import HasMemoized
+from .base import Immutable
+from .coercions import _document_text_coercion
+from .elements import _anonymous_label
+from .elements import BindParameter
+from .elements import BooleanClauseList
+from .elements import ClauseElement
+from .elements import ClauseList
+from .elements import ColumnClause
+from .elements import ColumnElement
+from .elements import DQLDMLClauseElement
+from .elements import GroupedElement
+from .elements import literal_column
+from .elements import TableValuedColumn
+from .elements import UnaryExpression
+from .operators import OperatorType
+from .sqltypes import NULLTYPE
+from .visitors import _TraverseInternalsType
+from .visitors import InternalTraversal
+from .visitors import prefix_anon_map
+from .. import exc
+from .. import util
+from ..util import HasMemoized_ro_memoized_attribute
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+and_ = BooleanClauseList.and_
+
+_T = TypeVar("_T", bound=Any)
+
+if TYPE_CHECKING:
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _ColumnExpressionOrStrLabelArgument
+    from ._typing import _FromClauseArgument
+    from ._typing import _JoinTargetArgument
+    from ._typing import _LimitOffsetType
+    from ._typing import _MAYBE_ENTITY
+    from ._typing import _NOT_ENTITY
+    from ._typing import _OnClauseArgument
+    from ._typing import _SelectStatementForCompoundArgument
+    from ._typing import _T0
+    from ._typing import _T1
+    from ._typing import _T2
+    from ._typing import _T3
+    from ._typing import _T4
+    from ._typing import _T5
+    from ._typing import _T6
+    from ._typing import _T7
+    from ._typing import _TextCoercedExpressionArgument
+    from ._typing import _TypedColumnClauseArgument as _TCCA
+    from ._typing import _TypeEngineArgument
+    from .base import _AmbiguousTableNameMap
+    from .base import ExecutableOption
+    from .base import ReadOnlyColumnCollection
+    from .cache_key import _CacheKeyTraversalType
+    from .compiler import SQLCompiler
+    from .dml import Delete
+    from .dml import Update
+    from .elements import BinaryExpression
+    from .elements import KeyedColumnElement
+    from .elements import Label
+    from .elements import NamedColumn
+    from .elements import TextClause
+    from .functions import Function
+    from .schema import ForeignKey
+    from .schema import ForeignKeyConstraint
+    from .sqltypes import TableValueType
+    from .type_api import TypeEngine
+    from .visitors import _CloneCallableType
+
+
+_ColumnsClauseElement = Union["FromClause", ColumnElement[Any], "TextClause"]
+_LabelConventionCallable = Callable[
+    [Union["ColumnElement[Any]", "TextClause"]], Optional[str]
+]
+
+
+class _JoinTargetProtocol(Protocol):
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]: ...
+
+    @util.ro_non_memoized_property
+    def entity_namespace(self) -> _EntityNamespace: ...
+
+
+_JoinTargetElement = Union["FromClause", _JoinTargetProtocol]
+_OnClauseElement = Union["ColumnElement[bool]", _JoinTargetProtocol]
+
+_ForUpdateOfArgument = Union[
+    # single column, Table, ORM Entity
+    Union[
+        "_ColumnExpressionArgument[Any]",
+        "_FromClauseArgument",
+    ],
+    # or sequence of single column elements
+    Sequence["_ColumnExpressionArgument[Any]"],
+]
+
+
+_SetupJoinsElement = Tuple[
+    _JoinTargetElement,
+    Optional[_OnClauseElement],
+    Optional["FromClause"],
+    Dict[str, Any],
+]
+
+
+_SelectIterable = Iterable[Union["ColumnElement[Any]", "TextClause"]]
+
+
+class _OffsetLimitParam(BindParameter[int]):
+    inherit_cache = True
+
+    @property
+    def _limit_offset_value(self) -> Optional[int]:
+        return self.effective_value
+
+
+class ReturnsRows(roles.ReturnsRowsRole, DQLDMLClauseElement):
+    """The base-most class for Core constructs that have some concept of
+    columns that can represent rows.
+
+    While the SELECT statement and TABLE are the primary things we think
+    of in this category,  DML like INSERT, UPDATE and DELETE can also specify
+    RETURNING which means they can be used in CTEs and other forms, and
+    PostgreSQL has functions that return rows also.
+
+    .. versionadded:: 1.4
+
+    """
+
+    _is_returns_rows = True
+
+    # sub-elements of returns_rows
+    _is_from_clause = False
+    _is_select_base = False
+    _is_select_statement = False
+    _is_lateral = False
+
+    @property
+    def selectable(self) -> ReturnsRows:
+        return self
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        """A sequence of column expression objects that represents the
+        "selected" columns of this :class:`_expression.ReturnsRows`.
+
+        This is typically equivalent to .exported_columns except it is
+        delivered in the form of a straight sequence and not  keyed
+        :class:`_expression.ColumnCollection`.
+
+        """
+        raise NotImplementedError()
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        """Return ``True`` if this :class:`.ReturnsRows` is
+        'derived' from the given :class:`.FromClause`.
+
+        An example would be an Alias of a Table is derived from that Table.
+
+        """
+        raise NotImplementedError()
+
+    def _generate_fromclause_column_proxies(
+        self,
+        fromclause: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        """Populate columns into an :class:`.AliasedReturnsRows` object."""
+
+        raise NotImplementedError()
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        """reset internal collections for an incoming column being added."""
+        raise NotImplementedError()
+
+    @property
+    def exported_columns(self) -> ReadOnlyColumnCollection[Any, Any]:
+        """A :class:`_expression.ColumnCollection`
+        that represents the "exported"
+        columns of this :class:`_expression.ReturnsRows`.
+
+        The "exported" columns represent the collection of
+        :class:`_expression.ColumnElement`
+        expressions that are rendered by this SQL
+        construct.   There are primary varieties which are the
+        "FROM clause columns" of a FROM clause, such as a table, join,
+        or subquery, the "SELECTed columns", which are the columns in
+        the "columns clause" of a SELECT statement, and the RETURNING
+        columns in a DML statement..
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_expression.FromClause.exported_columns`
+
+            :attr:`_expression.SelectBase.exported_columns`
+        """
+
+        raise NotImplementedError()
+
+
+class ExecutableReturnsRows(Executable, ReturnsRows):
+    """base for executable statements that return rows."""
+
+
+class TypedReturnsRows(ExecutableReturnsRows, Generic[_TP]):
+    """base for executable statements that return rows."""
+
+
+class Selectable(ReturnsRows):
+    """Mark a class as being selectable."""
+
+    __visit_name__ = "selectable"
+
+    is_selectable = True
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        raise NotImplementedError()
+
+    def lateral(self, name: Optional[str] = None) -> LateralFromClause:
+        """Return a LATERAL alias of this :class:`_expression.Selectable`.
+
+        The return value is the :class:`_expression.Lateral` construct also
+        provided by the top-level :func:`_expression.lateral` function.
+
+        .. seealso::
+
+            :ref:`tutorial_lateral_correlation` -  overview of usage.
+
+        """
+        return Lateral._construct(self, name=name)
+
+    @util.deprecated(
+        "1.4",
+        message="The :meth:`.Selectable.replace_selectable` method is "
+        "deprecated, and will be removed in a future release.  Similar "
+        "functionality is available via the sqlalchemy.sql.visitors module.",
+    )
+    @util.preload_module("sqlalchemy.sql.util")
+    def replace_selectable(self, old: FromClause, alias: Alias) -> Self:
+        """Replace all occurrences of :class:`_expression.FromClause`
+        'old' with the given :class:`_expression.Alias`
+        object, returning a copy of this :class:`_expression.FromClause`.
+
+        """
+        return util.preloaded.sql_util.ClauseAdapter(alias).traverse(self)
+
+    def corresponding_column(
+        self, column: KeyedColumnElement[Any], require_embedded: bool = False
+    ) -> Optional[KeyedColumnElement[Any]]:
+        """Given a :class:`_expression.ColumnElement`, return the exported
+        :class:`_expression.ColumnElement` object from the
+        :attr:`_expression.Selectable.exported_columns`
+        collection of this :class:`_expression.Selectable`
+        which corresponds to that
+        original :class:`_expression.ColumnElement` via a common ancestor
+        column.
+
+        :param column: the target :class:`_expression.ColumnElement`
+                      to be matched.
+
+        :param require_embedded: only return corresponding columns for
+         the given :class:`_expression.ColumnElement`, if the given
+         :class:`_expression.ColumnElement`
+         is actually present within a sub-element
+         of this :class:`_expression.Selectable`.
+         Normally the column will match if
+         it merely shares a common ancestor with one of the exported
+         columns of this :class:`_expression.Selectable`.
+
+        .. seealso::
+
+            :attr:`_expression.Selectable.exported_columns` - the
+            :class:`_expression.ColumnCollection`
+            that is used for the operation.
+
+            :meth:`_expression.ColumnCollection.corresponding_column`
+            - implementation
+            method.
+
+        """
+
+        return self.exported_columns.corresponding_column(
+            column, require_embedded
+        )
+
+
+class HasPrefixes:
+    _prefixes: Tuple[Tuple[DQLDMLClauseElement, str], ...] = ()
+
+    _has_prefixes_traverse_internals: _TraverseInternalsType = [
+        ("_prefixes", InternalTraversal.dp_prefix_sequence)
+    ]
+
+    @_generative
+    @_document_text_coercion(
+        "prefixes",
+        ":meth:`_expression.HasPrefixes.prefix_with`",
+        ":paramref:`.HasPrefixes.prefix_with.*prefixes`",
+    )
+    def prefix_with(
+        self,
+        *prefixes: _TextCoercedExpressionArgument[Any],
+        dialect: str = "*",
+    ) -> Self:
+        r"""Add one or more expressions following the statement keyword, i.e.
+        SELECT, INSERT, UPDATE, or DELETE. Generative.
+
+        This is used to support backend-specific prefix keywords such as those
+        provided by MySQL.
+
+        E.g.::
+
+            stmt = table.insert().prefix_with("LOW_PRIORITY", dialect="mysql")
+
+            # MySQL 5.7 optimizer hints
+            stmt = select(table).prefix_with("/*+ BKA(t1) */", dialect="mysql")
+
+        Multiple prefixes can be specified by multiple calls
+        to :meth:`_expression.HasPrefixes.prefix_with`.
+
+        :param \*prefixes: textual or :class:`_expression.ClauseElement`
+         construct which
+         will be rendered following the INSERT, UPDATE, or DELETE
+         keyword.
+        :param dialect: optional string dialect name which will
+         limit rendering of this prefix to only that dialect.
+
+        """
+        self._prefixes = self._prefixes + tuple(
+            [
+                (coercions.expect(roles.StatementOptionRole, p), dialect)
+                for p in prefixes
+            ]
+        )
+        return self
+
+
+class HasSuffixes:
+    _suffixes: Tuple[Tuple[DQLDMLClauseElement, str], ...] = ()
+
+    _has_suffixes_traverse_internals: _TraverseInternalsType = [
+        ("_suffixes", InternalTraversal.dp_prefix_sequence)
+    ]
+
+    @_generative
+    @_document_text_coercion(
+        "suffixes",
+        ":meth:`_expression.HasSuffixes.suffix_with`",
+        ":paramref:`.HasSuffixes.suffix_with.*suffixes`",
+    )
+    def suffix_with(
+        self,
+        *suffixes: _TextCoercedExpressionArgument[Any],
+        dialect: str = "*",
+    ) -> Self:
+        r"""Add one or more expressions following the statement as a whole.
+
+        This is used to support backend-specific suffix keywords on
+        certain constructs.
+
+        E.g.::
+
+            stmt = (
+                select(col1, col2)
+                .cte()
+                .suffix_with(
+                    "cycle empno set y_cycle to 1 default 0", dialect="oracle"
+                )
+            )
+
+        Multiple suffixes can be specified by multiple calls
+        to :meth:`_expression.HasSuffixes.suffix_with`.
+
+        :param \*suffixes: textual or :class:`_expression.ClauseElement`
+         construct which
+         will be rendered following the target clause.
+        :param dialect: Optional string dialect name which will
+         limit rendering of this suffix to only that dialect.
+
+        """
+        self._suffixes = self._suffixes + tuple(
+            [
+                (coercions.expect(roles.StatementOptionRole, p), dialect)
+                for p in suffixes
+            ]
+        )
+        return self
+
+
+class HasHints:
+    _hints: util.immutabledict[Tuple[FromClause, str], str] = (
+        util.immutabledict()
+    )
+    _statement_hints: Tuple[Tuple[str, str], ...] = ()
+
+    _has_hints_traverse_internals: _TraverseInternalsType = [
+        ("_statement_hints", InternalTraversal.dp_statement_hint_list),
+        ("_hints", InternalTraversal.dp_table_hint_list),
+    ]
+
+    @_generative
+    def with_statement_hint(self, text: str, dialect_name: str = "*") -> Self:
+        """Add a statement hint to this :class:`_expression.Select` or
+        other selectable object.
+
+        .. tip::
+
+            :meth:`_expression.Select.with_statement_hint` generally adds hints
+            **at the trailing end** of a SELECT statement.  To place
+            dialect-specific hints such as optimizer hints at the **front** of
+            the SELECT statement after the SELECT keyword, use the
+            :meth:`_expression.Select.prefix_with` method for an open-ended
+            space, or for table-specific hints the
+            :meth:`_expression.Select.with_hint` may be used, which places
+            hints in a dialect-specific location.
+
+        This method is similar to :meth:`_expression.Select.with_hint` except
+        that it does not require an individual table, and instead applies to
+        the statement as a whole.
+
+        Hints here are specific to the backend database and may include
+        directives such as isolation levels, file directives, fetch directives,
+        etc.
+
+        .. seealso::
+
+            :meth:`_expression.Select.with_hint`
+
+            :meth:`_expression.Select.prefix_with` - generic SELECT prefixing
+            which also can suit some database-specific HINT syntaxes such as
+            MySQL or Oracle Database optimizer hints
+
+        """
+        return self._with_hint(None, text, dialect_name)
+
+    @_generative
+    def with_hint(
+        self,
+        selectable: _FromClauseArgument,
+        text: str,
+        dialect_name: str = "*",
+    ) -> Self:
+        r"""Add an indexing or other executional context hint for the given
+        selectable to this :class:`_expression.Select` or other selectable
+        object.
+
+        .. tip::
+
+            The :meth:`_expression.Select.with_hint` method adds hints that are
+            **specific to a single table** to a statement, in a location that
+            is **dialect-specific**.  To add generic optimizer hints to the
+            **beginning** of a statement ahead of the SELECT keyword such as
+            for MySQL or Oracle Database, use the
+            :meth:`_expression.Select.prefix_with` method.  To add optimizer
+            hints to the **end** of a statement such as for PostgreSQL, use the
+            :meth:`_expression.Select.with_statement_hint` method.
+
+        The text of the hint is rendered in the appropriate
+        location for the database backend in use, relative
+        to the given :class:`_schema.Table` or :class:`_expression.Alias`
+        passed as the
+        ``selectable`` argument. The dialect implementation
+        typically uses Python string substitution syntax
+        with the token ``%(name)s`` to render the name of
+        the table or alias. E.g. when using Oracle Database, the
+        following::
+
+            select(mytable).with_hint(mytable, "index(%(name)s ix_mytable)")
+
+        Would render SQL as:
+
+        .. sourcecode:: sql
+
+            select /*+ index(mytable ix_mytable) */ ... from mytable
+
+        The ``dialect_name`` option will limit the rendering of a particular
+        hint to a particular backend. Such as, to add hints for both Oracle
+        Database and MSSql simultaneously::
+
+            select(mytable).with_hint(
+                mytable, "index(%(name)s ix_mytable)", "oracle"
+            ).with_hint(mytable, "WITH INDEX ix_mytable", "mssql")
+
+        .. seealso::
+
+            :meth:`_expression.Select.with_statement_hint`
+
+            :meth:`_expression.Select.prefix_with` - generic SELECT prefixing
+            which also can suit some database-specific HINT syntaxes such as
+            MySQL or Oracle Database optimizer hints
+
+        """
+
+        return self._with_hint(selectable, text, dialect_name)
+
+    def _with_hint(
+        self,
+        selectable: Optional[_FromClauseArgument],
+        text: str,
+        dialect_name: str,
+    ) -> Self:
+        if selectable is None:
+            self._statement_hints += ((dialect_name, text),)
+        else:
+            self._hints = self._hints.union(
+                {
+                    (
+                        coercions.expect(roles.FromClauseRole, selectable),
+                        dialect_name,
+                    ): text
+                }
+            )
+        return self
+
+
+class FromClause(roles.AnonymizedFromClauseRole, Selectable):
+    """Represent an element that can be used within the ``FROM``
+    clause of a ``SELECT`` statement.
+
+    The most common forms of :class:`_expression.FromClause` are the
+    :class:`_schema.Table` and the :func:`_expression.select` constructs.  Key
+    features common to all :class:`_expression.FromClause` objects include:
+
+    * a :attr:`.c` collection, which provides per-name access to a collection
+      of :class:`_expression.ColumnElement` objects.
+    * a :attr:`.primary_key` attribute, which is a collection of all those
+      :class:`_expression.ColumnElement`
+      objects that indicate the ``primary_key`` flag.
+    * Methods to generate various derivations of a "from" clause, including
+      :meth:`_expression.FromClause.alias`,
+      :meth:`_expression.FromClause.join`,
+      :meth:`_expression.FromClause.select`.
+
+
+    """
+
+    __visit_name__ = "fromclause"
+    named_with_column = False
+
+    @util.ro_non_memoized_property
+    def _hide_froms(self) -> Iterable[FromClause]:
+        return ()
+
+    _is_clone_of: Optional[FromClause]
+
+    _columns: ColumnCollection[Any, Any]
+
+    schema: Optional[str] = None
+    """Define the 'schema' attribute for this :class:`_expression.FromClause`.
+
+    This is typically ``None`` for most objects except that of
+    :class:`_schema.Table`, where it is taken as the value of the
+    :paramref:`_schema.Table.schema` argument.
+
+    """
+
+    is_selectable = True
+    _is_from_clause = True
+    _is_join = False
+
+    _use_schema_map = False
+
+    def select(self) -> Select[Any]:
+        r"""Return a SELECT of this :class:`_expression.FromClause`.
+
+
+        e.g.::
+
+            stmt = some_table.select().where(some_table.c.id == 5)
+
+        .. seealso::
+
+            :func:`_expression.select` - general purpose
+            method which allows for arbitrary column lists.
+
+        """
+        return Select(self)
+
+    def join(
+        self,
+        right: _FromClauseArgument,
+        onclause: Optional[_ColumnExpressionArgument[bool]] = None,
+        isouter: bool = False,
+        full: bool = False,
+    ) -> Join:
+        """Return a :class:`_expression.Join` from this
+        :class:`_expression.FromClause`
+        to another :class:`FromClause`.
+
+        E.g.::
+
+            from sqlalchemy import join
+
+            j = user_table.join(
+                address_table, user_table.c.id == address_table.c.user_id
+            )
+            stmt = select(user_table).select_from(j)
+
+        would emit SQL along the lines of:
+
+        .. sourcecode:: sql
+
+            SELECT user.id, user.name FROM user
+            JOIN address ON user.id = address.user_id
+
+        :param right: the right side of the join; this is any
+         :class:`_expression.FromClause` object such as a
+         :class:`_schema.Table` object, and
+         may also be a selectable-compatible object such as an ORM-mapped
+         class.
+
+        :param onclause: a SQL expression representing the ON clause of the
+         join.  If left at ``None``, :meth:`_expression.FromClause.join`
+         will attempt to
+         join the two tables based on a foreign key relationship.
+
+        :param isouter: if True, render a LEFT OUTER JOIN, instead of JOIN.
+
+        :param full: if True, render a FULL OUTER JOIN, instead of LEFT OUTER
+         JOIN.  Implies :paramref:`.FromClause.join.isouter`.
+
+        .. seealso::
+
+            :func:`_expression.join` - standalone function
+
+            :class:`_expression.Join` - the type of object produced
+
+        """
+
+        return Join(self, right, onclause, isouter, full)
+
+    def outerjoin(
+        self,
+        right: _FromClauseArgument,
+        onclause: Optional[_ColumnExpressionArgument[bool]] = None,
+        full: bool = False,
+    ) -> Join:
+        """Return a :class:`_expression.Join` from this
+        :class:`_expression.FromClause`
+        to another :class:`FromClause`, with the "isouter" flag set to
+        True.
+
+        E.g.::
+
+            from sqlalchemy import outerjoin
+
+            j = user_table.outerjoin(
+                address_table, user_table.c.id == address_table.c.user_id
+            )
+
+        The above is equivalent to::
+
+            j = user_table.join(
+                address_table, user_table.c.id == address_table.c.user_id, isouter=True
+            )
+
+        :param right: the right side of the join; this is any
+         :class:`_expression.FromClause` object such as a
+         :class:`_schema.Table` object, and
+         may also be a selectable-compatible object such as an ORM-mapped
+         class.
+
+        :param onclause: a SQL expression representing the ON clause of the
+         join.  If left at ``None``, :meth:`_expression.FromClause.join`
+         will attempt to
+         join the two tables based on a foreign key relationship.
+
+        :param full: if True, render a FULL OUTER JOIN, instead of
+         LEFT OUTER JOIN.
+
+        .. seealso::
+
+            :meth:`_expression.FromClause.join`
+
+            :class:`_expression.Join`
+
+        """  # noqa: E501
+
+        return Join(self, right, onclause, True, full)
+
+    def alias(
+        self, name: Optional[str] = None, flat: bool = False
+    ) -> NamedFromClause:
+        """Return an alias of this :class:`_expression.FromClause`.
+
+        E.g.::
+
+            a2 = some_table.alias("a2")
+
+        The above code creates an :class:`_expression.Alias`
+        object which can be used
+        as a FROM clause in any SELECT statement.
+
+        .. seealso::
+
+            :ref:`tutorial_using_aliases`
+
+            :func:`_expression.alias`
+
+        """
+
+        return Alias._construct(self, name=name)
+
+    def tablesample(
+        self,
+        sampling: Union[float, Function[Any]],
+        name: Optional[str] = None,
+        seed: Optional[roles.ExpressionElementRole[Any]] = None,
+    ) -> TableSample:
+        """Return a TABLESAMPLE alias of this :class:`_expression.FromClause`.
+
+        The return value is the :class:`_expression.TableSample`
+        construct also
+        provided by the top-level :func:`_expression.tablesample` function.
+
+        .. seealso::
+
+            :func:`_expression.tablesample` - usage guidelines and parameters
+
+        """
+        return TableSample._construct(
+            self, sampling=sampling, name=name, seed=seed
+        )
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        """Return ``True`` if this :class:`_expression.FromClause` is
+        'derived' from the given ``FromClause``.
+
+        An example would be an Alias of a Table is derived from that Table.
+
+        """
+        # this is essentially an "identity" check in the base class.
+        # Other constructs override this to traverse through
+        # contained elements.
+        return fromclause in self._cloned_set
+
+    def _is_lexical_equivalent(self, other: FromClause) -> bool:
+        """Return ``True`` if this :class:`_expression.FromClause` and
+        the other represent the same lexical identity.
+
+        This tests if either one is a copy of the other, or
+        if they are the same via annotation identity.
+
+        """
+        return bool(self._cloned_set.intersection(other._cloned_set))
+
+    @util.ro_non_memoized_property
+    def description(self) -> str:
+        """A brief description of this :class:`_expression.FromClause`.
+
+        Used primarily for error message formatting.
+
+        """
+        return getattr(self, "name", self.__class__.__name__ + " object")
+
+    def _generate_fromclause_column_proxies(
+        self,
+        fromclause: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        columns._populate_separate_keys(
+            col._make_proxy(
+                fromclause, primary_key=primary_key, foreign_keys=foreign_keys
+            )
+            for col in self.c
+        )
+
+    @util.ro_non_memoized_property
+    def exported_columns(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        that represents the "exported"
+        columns of this :class:`_expression.Selectable`.
+
+        The "exported" columns for a :class:`_expression.FromClause`
+        object are synonymous
+        with the :attr:`_expression.FromClause.columns` collection.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_expression.Selectable.exported_columns`
+
+            :attr:`_expression.SelectBase.exported_columns`
+
+
+        """
+        return self.c
+
+    @util.ro_non_memoized_property
+    def columns(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """A named-based collection of :class:`_expression.ColumnElement`
+        objects maintained by this :class:`_expression.FromClause`.
+
+        The :attr:`.columns`, or :attr:`.c` collection, is the gateway
+        to the construction of SQL expressions using table-bound or
+        other selectable-bound columns::
+
+            select(mytable).where(mytable.c.somecolumn == 5)
+
+        :return: a :class:`.ColumnCollection` object.
+
+        """
+        return self.c
+
+    @util.ro_memoized_property
+    def c(self) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """
+        A synonym for :attr:`.FromClause.columns`
+
+        :return: a :class:`.ColumnCollection`
+
+        """
+        if "_columns" not in self.__dict__:
+            self._setup_collections()
+        return self._columns.as_readonly()
+
+    def _setup_collections(self) -> None:
+        assert "_columns" not in self.__dict__
+        assert "primary_key" not in self.__dict__
+        assert "foreign_keys" not in self.__dict__
+
+        _columns: ColumnCollection[Any, Any] = ColumnCollection()
+        primary_key = ColumnSet()
+        foreign_keys: Set[KeyedColumnElement[Any]] = set()
+
+        self._populate_column_collection(
+            columns=_columns,
+            primary_key=primary_key,
+            foreign_keys=foreign_keys,
+        )
+
+        # assigning these three collections separately is not itself atomic,
+        # but greatly reduces the surface for problems
+        self._columns = _columns
+        self.primary_key = primary_key  # type: ignore
+        self.foreign_keys = foreign_keys  # type: ignore
+
+    @util.ro_non_memoized_property
+    def entity_namespace(self) -> _EntityNamespace:
+        """Return a namespace used for name-based access in SQL expressions.
+
+        This is the namespace that is used to resolve "filter_by()" type
+        expressions, such as::
+
+            stmt.filter_by(address="some address")
+
+        It defaults to the ``.c`` collection, however internally it can
+        be overridden using the "entity_namespace" annotation to deliver
+        alternative results.
+
+        """
+        return self.c
+
+    @util.ro_memoized_property
+    def primary_key(self) -> Iterable[NamedColumn[Any]]:
+        """Return the iterable collection of :class:`_schema.Column` objects
+        which comprise the primary key of this :class:`_selectable.FromClause`.
+
+        For a :class:`_schema.Table` object, this collection is represented
+        by the :class:`_schema.PrimaryKeyConstraint` which itself is an
+        iterable collection of :class:`_schema.Column` objects.
+
+        """
+        self._setup_collections()
+        return self.primary_key
+
+    @util.ro_memoized_property
+    def foreign_keys(self) -> Iterable[ForeignKey]:
+        """Return the collection of :class:`_schema.ForeignKey` marker objects
+        which this FromClause references.
+
+        Each :class:`_schema.ForeignKey` is a member of a
+        :class:`_schema.Table`-wide
+        :class:`_schema.ForeignKeyConstraint`.
+
+        .. seealso::
+
+            :attr:`_schema.Table.foreign_key_constraints`
+
+        """
+        self._setup_collections()
+        return self.foreign_keys
+
+    def _reset_column_collection(self) -> None:
+        """Reset the attributes linked to the ``FromClause.c`` attribute.
+
+        This collection is separate from all the other memoized things
+        as it has shown to be sensitive to being cleared out in situations
+        where enclosing code, typically in a replacement traversal scenario,
+        has already established strong relationships
+        with the exported columns.
+
+        The collection is cleared for the case where a table is having a
+        column added to it as well as within a Join during copy internals.
+
+        """
+
+        for key in ["_columns", "columns", "c", "primary_key", "foreign_keys"]:
+            self.__dict__.pop(key, None)
+
+    @util.ro_non_memoized_property
+    def _select_iterable(self) -> _SelectIterable:
+        return (c for c in self.c if not _never_select_column(c))
+
+    @property
+    def _cols_populated(self) -> bool:
+        return "_columns" in self.__dict__
+
+    def _populate_column_collection(
+        self,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        """Called on subclasses to establish the .c collection.
+
+        Each implementation has a different way of establishing
+        this collection.
+
+        """
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        """Given a column added to the .c collection of an underlying
+        selectable, produce the local version of that column, assuming this
+        selectable ultimately should proxy this column.
+
+        this is used to "ping" a derived selectable to add a new column
+        to its .c. collection when a Column has been added to one of the
+        Table objects it ultimately derives from.
+
+        If the given selectable hasn't populated its .c. collection yet,
+        it should at least pass on the message to the contained selectables,
+        but it will return None.
+
+        This method is currently used by Declarative to allow Table
+        columns to be added to a partially constructed inheritance
+        mapping that may have already produced joins.  The method
+        isn't public right now, as the full span of implications
+        and/or caveats aren't yet clear.
+
+        It's also possible that this functionality could be invoked by
+        default via an event, which would require that
+        selectables maintain a weak referencing collection of all
+        derivations.
+
+        """
+        self._reset_column_collection()
+
+    def _anonymous_fromclause(
+        self, *, name: Optional[str] = None, flat: bool = False
+    ) -> FromClause:
+        return self.alias(name=name)
+
+    if TYPE_CHECKING:
+
+        def self_group(
+            self, against: Optional[OperatorType] = None
+        ) -> Union[FromGrouping, Self]: ...
+
+
+class NamedFromClause(FromClause):
+    """A :class:`.FromClause` that has a name.
+
+    Examples include tables, subqueries, CTEs, aliased tables.
+
+    .. versionadded:: 2.0
+
+    """
+
+    named_with_column = True
+
+    name: str
+
+    @util.preload_module("sqlalchemy.sql.sqltypes")
+    def table_valued(self) -> TableValuedColumn[Any]:
+        """Return a :class:`_sql.TableValuedColumn` object for this
+        :class:`_expression.FromClause`.
+
+        A :class:`_sql.TableValuedColumn` is a :class:`_sql.ColumnElement` that
+        represents a complete row in a table. Support for this construct is
+        backend dependent, and is supported in various forms by backends
+        such as PostgreSQL, Oracle Database and SQL Server.
+
+        E.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import select, column, func, table
+            >>> a = table("a", column("id"), column("x"), column("y"))
+            >>> stmt = select(func.row_to_json(a.table_valued()))
+            >>> print(stmt)
+            {printsql}SELECT row_to_json(a) AS row_to_json_1
+            FROM a
+
+        .. versionadded:: 1.4.0b2
+
+        .. seealso::
+
+            :ref:`tutorial_functions` - in the :ref:`unified_tutorial`
+
+        """
+        return TableValuedColumn(self, type_api.TABLEVALUE)
+
+
+class SelectLabelStyle(Enum):
+    """Label style constants that may be passed to
+    :meth:`_sql.Select.set_label_style`."""
+
+    LABEL_STYLE_NONE = 0
+    """Label style indicating no automatic labeling should be applied to the
+    columns clause of a SELECT statement.
+
+    Below, the columns named ``columna`` are both rendered as is, meaning that
+    the name ``columna`` can only refer to the first occurrence of this name
+    within a result set, as well as if the statement were used as a subquery:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import table, column, select, true, LABEL_STYLE_NONE
+        >>> table1 = table("table1", column("columna"), column("columnb"))
+        >>> table2 = table("table2", column("columna"), column("columnc"))
+        >>> print(
+        ...     select(table1, table2)
+        ...     .join(table2, true())
+        ...     .set_label_style(LABEL_STYLE_NONE)
+        ... )
+        {printsql}SELECT table1.columna, table1.columnb, table2.columna, table2.columnc
+        FROM table1 JOIN table2 ON true
+
+    Used with the :meth:`_sql.Select.set_label_style` method.
+
+    .. versionadded:: 1.4
+
+    """  # noqa: E501
+
+    LABEL_STYLE_TABLENAME_PLUS_COL = 1
+    """Label style indicating all columns should be labeled as
+    ``<tablename>_<columnname>`` when generating the columns clause of a SELECT
+    statement, to disambiguate same-named columns referenced from different
+    tables, aliases, or subqueries.
+
+    Below, all column names are given a label so that the two same-named
+    columns ``columna`` are disambiguated as ``table1_columna`` and
+    ``table2_columna``:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import (
+        ...     table,
+        ...     column,
+        ...     select,
+        ...     true,
+        ...     LABEL_STYLE_TABLENAME_PLUS_COL,
+        ... )
+        >>> table1 = table("table1", column("columna"), column("columnb"))
+        >>> table2 = table("table2", column("columna"), column("columnc"))
+        >>> print(
+        ...     select(table1, table2)
+        ...     .join(table2, true())
+        ...     .set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+        ... )
+        {printsql}SELECT table1.columna AS table1_columna, table1.columnb AS table1_columnb, table2.columna AS table2_columna, table2.columnc AS table2_columnc
+        FROM table1 JOIN table2 ON true
+
+    Used with the :meth:`_sql.GenerativeSelect.set_label_style` method.
+    Equivalent to the legacy method ``Select.apply_labels()``;
+    :data:`_sql.LABEL_STYLE_TABLENAME_PLUS_COL` is SQLAlchemy's legacy
+    auto-labeling style. :data:`_sql.LABEL_STYLE_DISAMBIGUATE_ONLY` provides a
+    less intrusive approach to disambiguation of same-named column expressions.
+
+
+    .. versionadded:: 1.4
+
+    """  # noqa: E501
+
+    LABEL_STYLE_DISAMBIGUATE_ONLY = 2
+    """Label style indicating that columns with a name that conflicts with
+    an existing name should be labeled with a semi-anonymizing label
+    when generating the columns clause of a SELECT statement.
+
+    Below, most column names are left unaffected, except for the second
+    occurrence of the name ``columna``, which is labeled using the
+    label ``columna_1`` to disambiguate it from that of ``tablea.columna``:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import (
+        ...     table,
+        ...     column,
+        ...     select,
+        ...     true,
+        ...     LABEL_STYLE_DISAMBIGUATE_ONLY,
+        ... )
+        >>> table1 = table("table1", column("columna"), column("columnb"))
+        >>> table2 = table("table2", column("columna"), column("columnc"))
+        >>> print(
+        ...     select(table1, table2)
+        ...     .join(table2, true())
+        ...     .set_label_style(LABEL_STYLE_DISAMBIGUATE_ONLY)
+        ... )
+        {printsql}SELECT table1.columna, table1.columnb, table2.columna AS columna_1, table2.columnc
+        FROM table1 JOIN table2 ON true
+
+    Used with the :meth:`_sql.GenerativeSelect.set_label_style` method,
+    :data:`_sql.LABEL_STYLE_DISAMBIGUATE_ONLY` is the default labeling style
+    for all SELECT statements outside of :term:`1.x style` ORM queries.
+
+    .. versionadded:: 1.4
+
+    """  # noqa: E501
+
+    LABEL_STYLE_DEFAULT = LABEL_STYLE_DISAMBIGUATE_ONLY
+    """The default label style, refers to
+    :data:`_sql.LABEL_STYLE_DISAMBIGUATE_ONLY`.
+
+    .. versionadded:: 1.4
+
+    """
+
+    LABEL_STYLE_LEGACY_ORM = 3
+
+
+(
+    LABEL_STYLE_NONE,
+    LABEL_STYLE_TABLENAME_PLUS_COL,
+    LABEL_STYLE_DISAMBIGUATE_ONLY,
+    _,
+) = list(SelectLabelStyle)
+
+LABEL_STYLE_DEFAULT = LABEL_STYLE_DISAMBIGUATE_ONLY
+
+
+class Join(roles.DMLTableRole, FromClause):
+    """Represent a ``JOIN`` construct between two
+    :class:`_expression.FromClause`
+    elements.
+
+    The public constructor function for :class:`_expression.Join`
+    is the module-level
+    :func:`_expression.join()` function, as well as the
+    :meth:`_expression.FromClause.join` method
+    of any :class:`_expression.FromClause` (e.g. such as
+    :class:`_schema.Table`).
+
+    .. seealso::
+
+        :func:`_expression.join`
+
+        :meth:`_expression.FromClause.join`
+
+    """
+
+    __visit_name__ = "join"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("left", InternalTraversal.dp_clauseelement),
+        ("right", InternalTraversal.dp_clauseelement),
+        ("onclause", InternalTraversal.dp_clauseelement),
+        ("isouter", InternalTraversal.dp_boolean),
+        ("full", InternalTraversal.dp_boolean),
+    ]
+
+    _is_join = True
+
+    left: FromClause
+    right: FromClause
+    onclause: Optional[ColumnElement[bool]]
+    isouter: bool
+    full: bool
+
+    def __init__(
+        self,
+        left: _FromClauseArgument,
+        right: _FromClauseArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        isouter: bool = False,
+        full: bool = False,
+    ):
+        """Construct a new :class:`_expression.Join`.
+
+        The usual entrypoint here is the :func:`_expression.join`
+        function or the :meth:`_expression.FromClause.join` method of any
+        :class:`_expression.FromClause` object.
+
+        """
+
+        # when deannotate was removed here, callcounts went up for ORM
+        # compilation of eager joins, since there were more comparisons of
+        # annotated objects.   test_orm.py -> test_fetch_results
+        # was therefore changed to show a more real-world use case, where the
+        # compilation is cached; there's no change in post-cache callcounts.
+        # callcounts for a single compilation in that particular test
+        # that includes about eight joins about 1100 extra fn calls, from
+        # 29200 -> 30373
+
+        self.left = coercions.expect(
+            roles.FromClauseRole,
+            left,
+        )
+        self.right = coercions.expect(
+            roles.FromClauseRole,
+            right,
+        ).self_group()
+
+        if onclause is None:
+            self.onclause = self._match_primaries(self.left, self.right)
+        else:
+            # note: taken from If91f61527236fd4d7ae3cad1f24c38be921c90ba
+            # not merged yet
+            self.onclause = coercions.expect(
+                roles.OnClauseRole, onclause
+            ).self_group(against=operators._asbool)
+
+        self.isouter = isouter
+        self.full = full
+
+    @util.ro_non_memoized_property
+    def description(self) -> str:
+        return "Join object on %s(%d) and %s(%d)" % (
+            self.left.description,
+            id(self.left),
+            self.right.description,
+            id(self.right),
+        )
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        return (
+            # use hash() to ensure direct comparison to annotated works
+            # as well
+            hash(fromclause) == hash(self)
+            or self.left.is_derived_from(fromclause)
+            or self.right.is_derived_from(fromclause)
+        )
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> FromGrouping:
+        return FromGrouping(self)
+
+    @util.preload_module("sqlalchemy.sql.util")
+    def _populate_column_collection(
+        self,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        sqlutil = util.preloaded.sql_util
+        _columns: List[KeyedColumnElement[Any]] = [c for c in self.left.c] + [
+            c for c in self.right.c
+        ]
+
+        primary_key.extend(  # type: ignore
+            sqlutil.reduce_columns(
+                (c for c in _columns if c.primary_key), self.onclause
+            )
+        )
+        columns._populate_separate_keys(
+            (col._tq_key_label, col) for col in _columns  # type: ignore
+        )
+        foreign_keys.update(
+            itertools.chain(*[col.foreign_keys for col in _columns])  # type: ignore  # noqa: E501
+        )
+
+    def _copy_internals(
+        self, clone: _CloneCallableType = _clone, **kw: Any
+    ) -> None:
+        # see Select._copy_internals() for similar concept
+
+        # here we pre-clone "left" and "right" so that we can
+        # determine the new FROM clauses
+        all_the_froms = set(
+            itertools.chain(
+                _from_objects(self.left),
+                _from_objects(self.right),
+            )
+        )
+
+        # run the clone on those.  these will be placed in the
+        # cache used by the clone function
+        new_froms = {f: clone(f, **kw) for f in all_the_froms}
+
+        # set up a special replace function that will replace for
+        # ColumnClause with parent table referring to those
+        # replaced FromClause objects
+        def replace(
+            obj: Union[BinaryExpression[Any], ColumnClause[Any]],
+            **kw: Any,
+        ) -> Optional[KeyedColumnElement[Any]]:
+            if isinstance(obj, ColumnClause) and obj.table in new_froms:
+                newelem = new_froms[obj.table].corresponding_column(obj)
+                return newelem
+            return None
+
+        kw["replace"] = replace
+
+        # run normal _copy_internals.  the clones for
+        # left and right will come from the clone function's
+        # cache
+        super()._copy_internals(clone=clone, **kw)
+
+        self._reset_memoizations()
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        super()._refresh_for_new_column(column)
+        self.left._refresh_for_new_column(column)
+        self.right._refresh_for_new_column(column)
+
+    def _match_primaries(
+        self,
+        left: FromClause,
+        right: FromClause,
+    ) -> ColumnElement[bool]:
+        if isinstance(left, Join):
+            left_right = left.right
+        else:
+            left_right = None
+        return self._join_condition(left, right, a_subset=left_right)
+
+    @classmethod
+    def _join_condition(
+        cls,
+        a: FromClause,
+        b: FromClause,
+        *,
+        a_subset: Optional[FromClause] = None,
+        consider_as_foreign_keys: Optional[
+            AbstractSet[ColumnClause[Any]]
+        ] = None,
+    ) -> ColumnElement[bool]:
+        """Create a join condition between two tables or selectables.
+
+        See sqlalchemy.sql.util.join_condition() for full docs.
+
+        """
+        constraints = cls._joincond_scan_left_right(
+            a, a_subset, b, consider_as_foreign_keys
+        )
+
+        if len(constraints) > 1:
+            cls._joincond_trim_constraints(
+                a, b, constraints, consider_as_foreign_keys
+            )
+
+        if len(constraints) == 0:
+            if isinstance(b, FromGrouping):
+                hint = (
+                    " Perhaps you meant to convert the right side to a "
+                    "subquery using alias()?"
+                )
+            else:
+                hint = ""
+            raise exc.NoForeignKeysError(
+                "Can't find any foreign key relationships "
+                "between '%s' and '%s'.%s"
+                % (a.description, b.description, hint)
+            )
+
+        crit = [(x == y) for x, y in list(constraints.values())[0]]
+        if len(crit) == 1:
+            return crit[0]
+        else:
+            return and_(*crit)
+
+    @classmethod
+    def _can_join(
+        cls,
+        left: FromClause,
+        right: FromClause,
+        *,
+        consider_as_foreign_keys: Optional[
+            AbstractSet[ColumnClause[Any]]
+        ] = None,
+    ) -> bool:
+        if isinstance(left, Join):
+            left_right = left.right
+        else:
+            left_right = None
+
+        constraints = cls._joincond_scan_left_right(
+            a=left,
+            b=right,
+            a_subset=left_right,
+            consider_as_foreign_keys=consider_as_foreign_keys,
+        )
+
+        return bool(constraints)
+
+    @classmethod
+    @util.preload_module("sqlalchemy.sql.util")
+    def _joincond_scan_left_right(
+        cls,
+        a: FromClause,
+        a_subset: Optional[FromClause],
+        b: FromClause,
+        consider_as_foreign_keys: Optional[AbstractSet[ColumnClause[Any]]],
+    ) -> collections.defaultdict[
+        Optional[ForeignKeyConstraint],
+        List[Tuple[ColumnClause[Any], ColumnClause[Any]]],
+    ]:
+        sql_util = util.preloaded.sql_util
+
+        a = coercions.expect(roles.FromClauseRole, a)
+        b = coercions.expect(roles.FromClauseRole, b)
+
+        constraints: collections.defaultdict[
+            Optional[ForeignKeyConstraint],
+            List[Tuple[ColumnClause[Any], ColumnClause[Any]]],
+        ] = collections.defaultdict(list)
+
+        for left in (a_subset, a):
+            if left is None:
+                continue
+            for fk in sorted(
+                b.foreign_keys,
+                key=lambda fk: fk.parent._creation_order,
+            ):
+                if (
+                    consider_as_foreign_keys is not None
+                    and fk.parent not in consider_as_foreign_keys
+                ):
+                    continue
+                try:
+                    col = fk.get_referent(left)
+                except exc.NoReferenceError as nrte:
+                    table_names = {t.name for t in sql_util.find_tables(left)}
+                    if nrte.table_name in table_names:
+                        raise
+                    else:
+                        continue
+
+                if col is not None:
+                    constraints[fk.constraint].append((col, fk.parent))
+            if left is not b:
+                for fk in sorted(
+                    left.foreign_keys,
+                    key=lambda fk: fk.parent._creation_order,
+                ):
+                    if (
+                        consider_as_foreign_keys is not None
+                        and fk.parent not in consider_as_foreign_keys
+                    ):
+                        continue
+                    try:
+                        col = fk.get_referent(b)
+                    except exc.NoReferenceError as nrte:
+                        table_names = {t.name for t in sql_util.find_tables(b)}
+                        if nrte.table_name in table_names:
+                            raise
+                        else:
+                            continue
+
+                    if col is not None:
+                        constraints[fk.constraint].append((col, fk.parent))
+            if constraints:
+                break
+        return constraints
+
+    @classmethod
+    def _joincond_trim_constraints(
+        cls,
+        a: FromClause,
+        b: FromClause,
+        constraints: Dict[Any, Any],
+        consider_as_foreign_keys: Optional[Any],
+    ) -> None:
+        # more than one constraint matched.  narrow down the list
+        # to include just those FKCs that match exactly to
+        # "consider_as_foreign_keys".
+        if consider_as_foreign_keys:
+            for const in list(constraints):
+                if {f.parent for f in const.elements} != set(
+                    consider_as_foreign_keys
+                ):
+                    del constraints[const]
+
+        # if still multiple constraints, but
+        # they all refer to the exact same end result, use it.
+        if len(constraints) > 1:
+            dedupe = {tuple(crit) for crit in constraints.values()}
+            if len(dedupe) == 1:
+                key = list(constraints)[0]
+                constraints = {key: constraints[key]}
+
+        if len(constraints) != 1:
+            raise exc.AmbiguousForeignKeysError(
+                "Can't determine join between '%s' and '%s'; "
+                "tables have more than one foreign key "
+                "constraint relationship between them. "
+                "Please specify the 'onclause' of this "
+                "join explicitly." % (a.description, b.description)
+            )
+
+    def select(self) -> Select[Any]:
+        r"""Create a :class:`_expression.Select` from this
+        :class:`_expression.Join`.
+
+        E.g.::
+
+            stmt = table_a.join(table_b, table_a.c.id == table_b.c.a_id)
+
+            stmt = stmt.select()
+
+        The above will produce a SQL string resembling:
+
+        .. sourcecode:: sql
+
+            SELECT table_a.id, table_a.col, table_b.id, table_b.a_id
+            FROM table_a JOIN table_b ON table_a.id = table_b.a_id
+
+        """
+        return Select(self.left, self.right).select_from(self)
+
+    @util.preload_module("sqlalchemy.sql.util")
+    def _anonymous_fromclause(
+        self, name: Optional[str] = None, flat: bool = False
+    ) -> TODO_Any:
+        sqlutil = util.preloaded.sql_util
+        if flat:
+            if isinstance(self.left, (FromGrouping, Join)):
+                left_name = name  # will recurse
+            else:
+                if name and isinstance(self.left, NamedFromClause):
+                    left_name = f"{name}_{self.left.name}"
+                else:
+                    left_name = name
+            if isinstance(self.right, (FromGrouping, Join)):
+                right_name = name  # will recurse
+            else:
+                if name and isinstance(self.right, NamedFromClause):
+                    right_name = f"{name}_{self.right.name}"
+                else:
+                    right_name = name
+            left_a, right_a = (
+                self.left._anonymous_fromclause(name=left_name, flat=flat),
+                self.right._anonymous_fromclause(name=right_name, flat=flat),
+            )
+            adapter = sqlutil.ClauseAdapter(left_a).chain(
+                sqlutil.ClauseAdapter(right_a)
+            )
+
+            return left_a.join(
+                right_a,
+                adapter.traverse(self.onclause),
+                isouter=self.isouter,
+                full=self.full,
+            )
+        else:
+            return (
+                self.select()
+                .set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL)
+                .correlate(None)
+                .alias(name)
+            )
+
+    @util.ro_non_memoized_property
+    def _hide_froms(self) -> Iterable[FromClause]:
+        return itertools.chain(
+            *[_from_objects(x.left, x.right) for x in self._cloned_set]
+        )
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        self_list: List[FromClause] = [self]
+        return self_list + self.left._from_objects + self.right._from_objects
+
+
+class NoInit:
+    def __init__(self, *arg: Any, **kw: Any):
+        raise NotImplementedError(
+            "The %s class is not intended to be constructed "
+            "directly.  Please use the %s() standalone "
+            "function or the %s() method available from appropriate "
+            "selectable objects."
+            % (
+                self.__class__.__name__,
+                self.__class__.__name__.lower(),
+                self.__class__.__name__.lower(),
+            )
+        )
+
+
+class LateralFromClause(NamedFromClause):
+    """mark a FROM clause as being able to render directly as LATERAL"""
+
+
+# FromClause ->
+#   AliasedReturnsRows
+#        -> Alias   only for FromClause
+#        -> Subquery  only for SelectBase
+#        -> CTE only for HasCTE -> SelectBase, DML
+#        -> Lateral -> FromClause, but we accept SelectBase
+#           w/ non-deprecated coercion
+#        -> TableSample -> only for FromClause
+
+
+class AliasedReturnsRows(NoInit, NamedFromClause):
+    """Base class of aliases against tables, subqueries, and other
+    selectables."""
+
+    _is_from_container = True
+
+    _supports_derived_columns = False
+
+    element: ReturnsRows
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("name", InternalTraversal.dp_anon_name),
+    ]
+
+    @classmethod
+    def _construct(
+        cls,
+        selectable: Any,
+        *,
+        name: Optional[str] = None,
+        **kw: Any,
+    ) -> Self:
+        obj = cls.__new__(cls)
+        obj._init(selectable, name=name, **kw)
+        return obj
+
+    def _init(self, selectable: Any, *, name: Optional[str] = None) -> None:
+        self.element = coercions.expect(
+            roles.ReturnsRowsRole, selectable, apply_propagate_attrs=self
+        )
+        self.element = selectable
+        self._orig_name = name
+        if name is None:
+            if (
+                isinstance(selectable, FromClause)
+                and selectable.named_with_column
+            ):
+                name = getattr(selectable, "name", None)
+                if isinstance(name, _anonymous_label):
+                    name = None
+            name = _anonymous_label.safe_construct(id(self), name or "anon")
+        self.name = name
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        super()._refresh_for_new_column(column)
+        self.element._refresh_for_new_column(column)
+
+    def _populate_column_collection(
+        self,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        self.element._generate_fromclause_column_proxies(
+            self, columns, primary_key=primary_key, foreign_keys=foreign_keys
+        )
+
+    @util.ro_non_memoized_property
+    def description(self) -> str:
+        name = self.name
+        if isinstance(name, _anonymous_label):
+            name = "anon_1"
+
+        return name
+
+    @util.ro_non_memoized_property
+    def implicit_returning(self) -> bool:
+        return self.element.implicit_returning  # type: ignore
+
+    @property
+    def original(self) -> ReturnsRows:
+        """Legacy for dialects that are referring to Alias.original."""
+        return self.element
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        if fromclause in self._cloned_set:
+            return True
+        return self.element.is_derived_from(fromclause)
+
+    def _copy_internals(
+        self, clone: _CloneCallableType = _clone, **kw: Any
+    ) -> None:
+        existing_element = self.element
+
+        super()._copy_internals(clone=clone, **kw)
+
+        # the element clone is usually against a Table that returns the
+        # same object.  don't reset exported .c. collections and other
+        # memoized details if it was not changed.  this saves a lot on
+        # performance.
+        if existing_element is not self.element:
+            self._reset_column_collection()
+
+    @property
+    def _from_objects(self) -> List[FromClause]:
+        return [self]
+
+
+class FromClauseAlias(AliasedReturnsRows):
+    element: FromClause
+
+
+class Alias(roles.DMLTableRole, FromClauseAlias):
+    """Represents an table or selectable alias (AS).
+
+    Represents an alias, as typically applied to any table or
+    sub-select within a SQL statement using the ``AS`` keyword (or
+    without the keyword on certain databases such as Oracle Database).
+
+    This object is constructed from the :func:`_expression.alias` module
+    level function as well as the :meth:`_expression.FromClause.alias`
+    method available
+    on all :class:`_expression.FromClause` subclasses.
+
+    .. seealso::
+
+        :meth:`_expression.FromClause.alias`
+
+    """
+
+    __visit_name__ = "alias"
+
+    inherit_cache = True
+
+    element: FromClause
+
+    @classmethod
+    def _factory(
+        cls,
+        selectable: FromClause,
+        name: Optional[str] = None,
+        flat: bool = False,
+    ) -> NamedFromClause:
+        return coercions.expect(
+            roles.FromClauseRole, selectable, allow_select=True
+        ).alias(name=name, flat=flat)
+
+
+class TableValuedAlias(LateralFromClause, Alias):
+    """An alias against a "table valued" SQL function.
+
+    This construct provides for a SQL function that returns columns
+    to be used in the FROM clause of a SELECT statement.   The
+    object is generated using the :meth:`_functions.FunctionElement.table_valued`
+    method, e.g.:
+
+    .. sourcecode:: pycon+sql
+
+        >>> from sqlalchemy import select, func
+        >>> fn = func.json_array_elements_text('["one", "two", "three"]').table_valued(
+        ...     "value"
+        ... )
+        >>> print(select(fn.c.value))
+        {printsql}SELECT anon_1.value
+        FROM json_array_elements_text(:json_array_elements_text_1) AS anon_1
+
+    .. versionadded:: 1.4.0b2
+
+    .. seealso::
+
+        :ref:`tutorial_functions_table_valued` - in the :ref:`unified_tutorial`
+
+    """  # noqa: E501
+
+    __visit_name__ = "table_valued_alias"
+
+    _supports_derived_columns = True
+    _render_derived = False
+    _render_derived_w_types = False
+    joins_implicitly = False
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("name", InternalTraversal.dp_anon_name),
+        ("_tableval_type", InternalTraversal.dp_type),
+        ("_render_derived", InternalTraversal.dp_boolean),
+        ("_render_derived_w_types", InternalTraversal.dp_boolean),
+    ]
+
+    def _init(
+        self,
+        selectable: Any,
+        *,
+        name: Optional[str] = None,
+        table_value_type: Optional[TableValueType] = None,
+        joins_implicitly: bool = False,
+    ) -> None:
+        super()._init(selectable, name=name)
+
+        self.joins_implicitly = joins_implicitly
+        self._tableval_type = (
+            type_api.TABLEVALUE
+            if table_value_type is None
+            else table_value_type
+        )
+
+    @HasMemoized.memoized_attribute
+    def column(self) -> TableValuedColumn[Any]:
+        """Return a column expression representing this
+        :class:`_sql.TableValuedAlias`.
+
+        This accessor is used to implement the
+        :meth:`_functions.FunctionElement.column_valued` method. See that
+        method for further details.
+
+        E.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> print(select(func.some_func().table_valued("value").column))
+            {printsql}SELECT anon_1 FROM some_func() AS anon_1
+
+        .. seealso::
+
+            :meth:`_functions.FunctionElement.column_valued`
+
+        """
+
+        return TableValuedColumn(self, self._tableval_type)
+
+    def alias(
+        self, name: Optional[str] = None, flat: bool = False
+    ) -> TableValuedAlias:
+        """Return a new alias of this :class:`_sql.TableValuedAlias`.
+
+        This creates a distinct FROM object that will be distinguished
+        from the original one when used in a SQL statement.
+
+        """
+
+        tva: TableValuedAlias = TableValuedAlias._construct(
+            self,
+            name=name,
+            table_value_type=self._tableval_type,
+            joins_implicitly=self.joins_implicitly,
+        )
+
+        if self._render_derived:
+            tva._render_derived = True
+            tva._render_derived_w_types = self._render_derived_w_types
+
+        return tva
+
+    def lateral(self, name: Optional[str] = None) -> LateralFromClause:
+        """Return a new :class:`_sql.TableValuedAlias` with the lateral flag
+        set, so that it renders as LATERAL.
+
+        .. seealso::
+
+            :func:`_expression.lateral`
+
+        """
+        tva = self.alias(name=name)
+        tva._is_lateral = True
+        return tva
+
+    def render_derived(
+        self,
+        name: Optional[str] = None,
+        with_types: bool = False,
+    ) -> TableValuedAlias:
+        """Apply "render derived" to this :class:`_sql.TableValuedAlias`.
+
+        This has the effect of the individual column names listed out
+        after the alias name in the "AS" sequence, e.g.:
+
+        .. sourcecode:: pycon+sql
+
+            >>> print(
+            ...     select(
+            ...         func.unnest(array(["one", "two", "three"]))
+            ...         .table_valued("x", with_ordinality="o")
+            ...         .render_derived()
+            ...     )
+            ... )
+            {printsql}SELECT anon_1.x, anon_1.o
+            FROM unnest(ARRAY[%(param_1)s, %(param_2)s, %(param_3)s]) WITH ORDINALITY AS anon_1(x, o)
+
+        The ``with_types`` keyword will render column types inline within
+        the alias expression (this syntax currently applies to the
+        PostgreSQL database):
+
+        .. sourcecode:: pycon+sql
+
+            >>> print(
+            ...     select(
+            ...         func.json_to_recordset('[{"a":1,"b":"foo"},{"a":"2","c":"bar"}]')
+            ...         .table_valued(column("a", Integer), column("b", String))
+            ...         .render_derived(with_types=True)
+            ...     )
+            ... )
+            {printsql}SELECT anon_1.a, anon_1.b FROM json_to_recordset(:json_to_recordset_1)
+            AS anon_1(a INTEGER, b VARCHAR)
+
+        :param name: optional string name that will be applied to the alias
+         generated.  If left as None, a unique anonymizing name will be used.
+
+        :param with_types: if True, the derived columns will include the
+         datatype specification with each column. This is a special syntax
+         currently known to be required by PostgreSQL for some SQL functions.
+
+        """  # noqa: E501
+
+        # note: don't use the @_generative system here, keep a reference
+        # to the original object.  otherwise you can have re-use of the
+        # python id() of the original which can cause name conflicts if
+        # a new anon-name grabs the same identifier as the local anon-name
+        # (just saw it happen on CI)
+
+        # construct against original to prevent memory growth
+        # for repeated generations
+        new_alias: TableValuedAlias = TableValuedAlias._construct(
+            self.element,
+            name=name,
+            table_value_type=self._tableval_type,
+            joins_implicitly=self.joins_implicitly,
+        )
+        new_alias._render_derived = True
+        new_alias._render_derived_w_types = with_types
+        return new_alias
+
+
+class Lateral(FromClauseAlias, LateralFromClause):
+    """Represent a LATERAL subquery.
+
+    This object is constructed from the :func:`_expression.lateral` module
+    level function as well as the :meth:`_expression.FromClause.lateral`
+    method available
+    on all :class:`_expression.FromClause` subclasses.
+
+    While LATERAL is part of the SQL standard, currently only more recent
+    PostgreSQL versions provide support for this keyword.
+
+    .. seealso::
+
+        :ref:`tutorial_lateral_correlation` -  overview of usage.
+
+    """
+
+    __visit_name__ = "lateral"
+    _is_lateral = True
+
+    inherit_cache = True
+
+    @classmethod
+    def _factory(
+        cls,
+        selectable: Union[SelectBase, _FromClauseArgument],
+        name: Optional[str] = None,
+    ) -> LateralFromClause:
+        return coercions.expect(
+            roles.FromClauseRole, selectable, explicit_subquery=True
+        ).lateral(name=name)
+
+
+class TableSample(FromClauseAlias):
+    """Represent a TABLESAMPLE clause.
+
+    This object is constructed from the :func:`_expression.tablesample` module
+    level function as well as the :meth:`_expression.FromClause.tablesample`
+    method
+    available on all :class:`_expression.FromClause` subclasses.
+
+    .. seealso::
+
+        :func:`_expression.tablesample`
+
+    """
+
+    __visit_name__ = "tablesample"
+
+    _traverse_internals: _TraverseInternalsType = (
+        AliasedReturnsRows._traverse_internals
+        + [
+            ("sampling", InternalTraversal.dp_clauseelement),
+            ("seed", InternalTraversal.dp_clauseelement),
+        ]
+    )
+
+    @classmethod
+    def _factory(
+        cls,
+        selectable: _FromClauseArgument,
+        sampling: Union[float, Function[Any]],
+        name: Optional[str] = None,
+        seed: Optional[roles.ExpressionElementRole[Any]] = None,
+    ) -> TableSample:
+        return coercions.expect(roles.FromClauseRole, selectable).tablesample(
+            sampling, name=name, seed=seed
+        )
+
+    @util.preload_module("sqlalchemy.sql.functions")
+    def _init(  # type: ignore[override]
+        self,
+        selectable: Any,
+        *,
+        name: Optional[str] = None,
+        sampling: Union[float, Function[Any]],
+        seed: Optional[roles.ExpressionElementRole[Any]] = None,
+    ) -> None:
+        assert sampling is not None
+        functions = util.preloaded.sql_functions
+        if not isinstance(sampling, functions.Function):
+            sampling = functions.func.system(sampling)
+
+        self.sampling: Function[Any] = sampling
+        self.seed = seed
+        super()._init(selectable, name=name)
+
+    def _get_method(self) -> Function[Any]:
+        return self.sampling
+
+
+class CTE(
+    roles.DMLTableRole,
+    roles.IsCTERole,
+    Generative,
+    HasPrefixes,
+    HasSuffixes,
+    AliasedReturnsRows,
+):
+    """Represent a Common Table Expression.
+
+    The :class:`_expression.CTE` object is obtained using the
+    :meth:`_sql.SelectBase.cte` method from any SELECT statement. A less often
+    available syntax also allows use of the :meth:`_sql.HasCTE.cte` method
+    present on :term:`DML` constructs such as :class:`_sql.Insert`,
+    :class:`_sql.Update` and
+    :class:`_sql.Delete`.   See the :meth:`_sql.HasCTE.cte` method for
+    usage details on CTEs.
+
+    .. seealso::
+
+        :ref:`tutorial_subqueries_ctes` - in the 2.0 tutorial
+
+        :meth:`_sql.HasCTE.cte` - examples of calling styles
+
+    """
+
+    __visit_name__ = "cte"
+
+    _traverse_internals: _TraverseInternalsType = (
+        AliasedReturnsRows._traverse_internals
+        + [
+            ("_cte_alias", InternalTraversal.dp_clauseelement),
+            ("_restates", InternalTraversal.dp_clauseelement),
+            ("recursive", InternalTraversal.dp_boolean),
+            ("nesting", InternalTraversal.dp_boolean),
+        ]
+        + HasPrefixes._has_prefixes_traverse_internals
+        + HasSuffixes._has_suffixes_traverse_internals
+    )
+
+    element: HasCTE
+
+    @classmethod
+    def _factory(
+        cls,
+        selectable: HasCTE,
+        name: Optional[str] = None,
+        recursive: bool = False,
+    ) -> CTE:
+        r"""Return a new :class:`_expression.CTE`,
+        or Common Table Expression instance.
+
+        Please see :meth:`_expression.HasCTE.cte` for detail on CTE usage.
+
+        """
+        return coercions.expect(roles.HasCTERole, selectable).cte(
+            name=name, recursive=recursive
+        )
+
+    def _init(
+        self,
+        selectable: Select[Any],
+        *,
+        name: Optional[str] = None,
+        recursive: bool = False,
+        nesting: bool = False,
+        _cte_alias: Optional[CTE] = None,
+        _restates: Optional[CTE] = None,
+        _prefixes: Optional[Tuple[()]] = None,
+        _suffixes: Optional[Tuple[()]] = None,
+    ) -> None:
+        self.recursive = recursive
+        self.nesting = nesting
+        self._cte_alias = _cte_alias
+        # Keep recursivity reference with union/union_all
+        self._restates = _restates
+        if _prefixes:
+            self._prefixes = _prefixes
+        if _suffixes:
+            self._suffixes = _suffixes
+        super()._init(selectable, name=name)
+
+    def _populate_column_collection(
+        self,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        if self._cte_alias is not None:
+            self._cte_alias._generate_fromclause_column_proxies(
+                self,
+                columns,
+                primary_key=primary_key,
+                foreign_keys=foreign_keys,
+            )
+        else:
+            self.element._generate_fromclause_column_proxies(
+                self,
+                columns,
+                primary_key=primary_key,
+                foreign_keys=foreign_keys,
+            )
+
+    def alias(self, name: Optional[str] = None, flat: bool = False) -> CTE:
+        """Return an :class:`_expression.Alias` of this
+        :class:`_expression.CTE`.
+
+        This method is a CTE-specific specialization of the
+        :meth:`_expression.FromClause.alias` method.
+
+        .. seealso::
+
+            :ref:`tutorial_using_aliases`
+
+            :func:`_expression.alias`
+
+        """
+        return CTE._construct(
+            self.element,
+            name=name,
+            recursive=self.recursive,
+            nesting=self.nesting,
+            _cte_alias=self,
+            _prefixes=self._prefixes,
+            _suffixes=self._suffixes,
+        )
+
+    def union(self, *other: _SelectStatementForCompoundArgument) -> CTE:
+        r"""Return a new :class:`_expression.CTE` with a SQL ``UNION``
+        of the original CTE against the given selectables provided
+        as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28 multiple elements are now accepted.
+
+        .. seealso::
+
+            :meth:`_sql.HasCTE.cte` - examples of calling styles
+
+        """
+        assert is_select_statement(
+            self.element
+        ), f"CTE element f{self.element} does not support union()"
+
+        return CTE._construct(
+            self.element.union(*other),
+            name=self.name,
+            recursive=self.recursive,
+            nesting=self.nesting,
+            _restates=self,
+            _prefixes=self._prefixes,
+            _suffixes=self._suffixes,
+        )
+
+    def union_all(self, *other: _SelectStatementForCompoundArgument) -> CTE:
+        r"""Return a new :class:`_expression.CTE` with a SQL ``UNION ALL``
+        of the original CTE against the given selectables provided
+        as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28 multiple elements are now accepted.
+
+        .. seealso::
+
+            :meth:`_sql.HasCTE.cte` - examples of calling styles
+
+        """
+
+        assert is_select_statement(
+            self.element
+        ), f"CTE element f{self.element} does not support union_all()"
+
+        return CTE._construct(
+            self.element.union_all(*other),
+            name=self.name,
+            recursive=self.recursive,
+            nesting=self.nesting,
+            _restates=self,
+            _prefixes=self._prefixes,
+            _suffixes=self._suffixes,
+        )
+
+    def _get_reference_cte(self) -> CTE:
+        """
+        A recursive CTE is updated to attach the recursive part.
+        Updated CTEs should still refer to the original CTE.
+        This function returns this reference identifier.
+        """
+        return self._restates if self._restates is not None else self
+
+
+class _CTEOpts(NamedTuple):
+    nesting: bool
+
+
+class _ColumnsPlusNames(NamedTuple):
+    required_label_name: Optional[str]
+    """
+    string label name, if non-None, must be rendered as a
+    label, i.e. "AS <name>"
+    """
+
+    proxy_key: Optional[str]
+    """
+    proxy_key that is to be part of the result map for this
+    col.  this is also the key in a fromclause.c or
+    select.selected_columns collection
+    """
+
+    fallback_label_name: Optional[str]
+    """
+    name that can be used to render an "AS <name>" when
+    we have to render a label even though
+    required_label_name was not given
+    """
+
+    column: Union[ColumnElement[Any], TextClause]
+    """
+    the ColumnElement itself
+    """
+
+    repeated: bool
+    """
+    True if this is a duplicate of a previous column
+    in the list of columns
+    """
+
+
+class SelectsRows(ReturnsRows):
+    """Sub-base of ReturnsRows for elements that deliver rows
+    directly, namely SELECT and INSERT/UPDATE/DELETE..RETURNING"""
+
+    _label_style: SelectLabelStyle = LABEL_STYLE_NONE
+
+    def _generate_columns_plus_names(
+        self,
+        anon_for_dupe_key: bool,
+        cols: Optional[_SelectIterable] = None,
+    ) -> List[_ColumnsPlusNames]:
+        """Generate column names as rendered in a SELECT statement by
+        the compiler.
+
+        This is distinct from the _column_naming_convention generator that's
+        intended for population of .c collections and similar, which has
+        different rules.   the collection returned here calls upon the
+        _column_naming_convention as well.
+
+        """
+
+        if cols is None:
+            cols = self._all_selected_columns
+
+        key_naming_convention = SelectState._column_naming_convention(
+            self._label_style
+        )
+
+        names = {}
+
+        result: List[_ColumnsPlusNames] = []
+        result_append = result.append
+
+        table_qualified = self._label_style is LABEL_STYLE_TABLENAME_PLUS_COL
+        label_style_none = self._label_style is LABEL_STYLE_NONE
+
+        # a counter used for "dedupe" labels, which have double underscores
+        # in them and are never referred by name; they only act
+        # as positional placeholders.  they need only be unique within
+        # the single columns clause they're rendered within (required by
+        # some dbs such as mysql).  So their anon identity is tracked against
+        # a fixed counter rather than hash() identity.
+        dedupe_hash = 1
+
+        for c in cols:
+            repeated = False
+
+            if not c._render_label_in_columns_clause:
+                effective_name = required_label_name = fallback_label_name = (
+                    None
+                )
+            elif label_style_none:
+                if TYPE_CHECKING:
+                    assert is_column_element(c)
+
+                effective_name = required_label_name = None
+                fallback_label_name = c._non_anon_label or c._anon_name_label
+            else:
+                if TYPE_CHECKING:
+                    assert is_column_element(c)
+
+                if table_qualified:
+                    required_label_name = effective_name = (
+                        fallback_label_name
+                    ) = c._tq_label
+                else:
+                    effective_name = fallback_label_name = c._non_anon_label
+                    required_label_name = None
+
+                if effective_name is None:
+                    # it seems like this could be _proxy_key and we would
+                    # not need _expression_label but it isn't
+                    # giving us a clue when to use anon_label instead
+                    expr_label = c._expression_label
+                    if expr_label is None:
+                        repeated = c._anon_name_label in names
+                        names[c._anon_name_label] = c
+                        effective_name = required_label_name = None
+
+                        if repeated:
+                            # here, "required_label_name" is sent as
+                            # "None" and "fallback_label_name" is sent.
+                            if table_qualified:
+                                fallback_label_name = (
+                                    c._dedupe_anon_tq_label_idx(dedupe_hash)
+                                )
+                                dedupe_hash += 1
+                            else:
+                                fallback_label_name = c._dedupe_anon_label_idx(
+                                    dedupe_hash
+                                )
+                                dedupe_hash += 1
+                        else:
+                            fallback_label_name = c._anon_name_label
+                    else:
+                        required_label_name = effective_name = (
+                            fallback_label_name
+                        ) = expr_label
+
+            if effective_name is not None:
+                if TYPE_CHECKING:
+                    assert is_column_element(c)
+
+                if effective_name in names:
+                    # when looking to see if names[name] is the same column as
+                    # c, use hash(), so that an annotated version of the column
+                    # is seen as the same as the non-annotated
+                    if hash(names[effective_name]) != hash(c):
+                        # different column under the same name.  apply
+                        # disambiguating label
+                        if table_qualified:
+                            required_label_name = fallback_label_name = (
+                                c._anon_tq_label
+                            )
+                        else:
+                            required_label_name = fallback_label_name = (
+                                c._anon_name_label
+                            )
+
+                        if anon_for_dupe_key and required_label_name in names:
+                            # here, c._anon_tq_label is definitely unique to
+                            # that column identity (or annotated version), so
+                            # this should always be true.
+                            # this is also an infrequent codepath because
+                            # you need two levels of duplication to be here
+                            assert hash(names[required_label_name]) == hash(c)
+
+                            # the column under the disambiguating label is
+                            # already present.  apply the "dedupe" label to
+                            # subsequent occurrences of the column so that the
+                            # original stays non-ambiguous
+                            if table_qualified:
+                                required_label_name = fallback_label_name = (
+                                    c._dedupe_anon_tq_label_idx(dedupe_hash)
+                                )
+                                dedupe_hash += 1
+                            else:
+                                required_label_name = fallback_label_name = (
+                                    c._dedupe_anon_label_idx(dedupe_hash)
+                                )
+                                dedupe_hash += 1
+                            repeated = True
+                        else:
+                            names[required_label_name] = c
+                    elif anon_for_dupe_key:
+                        # same column under the same name. apply the "dedupe"
+                        # label so that the original stays non-ambiguous
+                        if table_qualified:
+                            required_label_name = fallback_label_name = (
+                                c._dedupe_anon_tq_label_idx(dedupe_hash)
+                            )
+                            dedupe_hash += 1
+                        else:
+                            required_label_name = fallback_label_name = (
+                                c._dedupe_anon_label_idx(dedupe_hash)
+                            )
+                            dedupe_hash += 1
+                        repeated = True
+                else:
+                    names[effective_name] = c
+
+            result_append(
+                _ColumnsPlusNames(
+                    required_label_name,
+                    key_naming_convention(c),
+                    fallback_label_name,
+                    c,
+                    repeated,
+                )
+            )
+
+        return result
+
+
+class HasCTE(roles.HasCTERole, SelectsRows):
+    """Mixin that declares a class to include CTE support."""
+
+    _has_ctes_traverse_internals: _TraverseInternalsType = [
+        ("_independent_ctes", InternalTraversal.dp_clauseelement_list),
+        ("_independent_ctes_opts", InternalTraversal.dp_plain_obj),
+    ]
+
+    _independent_ctes: Tuple[CTE, ...] = ()
+    _independent_ctes_opts: Tuple[_CTEOpts, ...] = ()
+
+    @_generative
+    def add_cte(self, *ctes: CTE, nest_here: bool = False) -> Self:
+        r"""Add one or more :class:`_sql.CTE` constructs to this statement.
+
+        This method will associate the given :class:`_sql.CTE` constructs with
+        the parent statement such that they will each be unconditionally
+        rendered in the WITH clause of the final statement, even if not
+        referenced elsewhere within the statement or any sub-selects.
+
+        The optional :paramref:`.HasCTE.add_cte.nest_here` parameter when set
+        to True will have the effect that each given :class:`_sql.CTE` will
+        render in a WITH clause rendered directly along with this statement,
+        rather than being moved to the top of the ultimate rendered statement,
+        even if this statement is rendered as a subquery within a larger
+        statement.
+
+        This method has two general uses. One is to embed CTE statements that
+        serve some purpose without being referenced explicitly, such as the use
+        case of embedding a DML statement such as an INSERT or UPDATE as a CTE
+        inline with a primary statement that may draw from its results
+        indirectly.  The other is to provide control over the exact placement
+        of a particular series of CTE constructs that should remain rendered
+        directly in terms of a particular statement that may be nested in a
+        larger statement.
+
+        E.g.::
+
+            from sqlalchemy import table, column, select
+
+            t = table("t", column("c1"), column("c2"))
+
+            ins = t.insert().values({"c1": "x", "c2": "y"}).cte()
+
+            stmt = select(t).add_cte(ins)
+
+        Would render:
+
+        .. sourcecode:: sql
+
+            WITH anon_1 AS (
+                INSERT INTO t (c1, c2) VALUES (:param_1, :param_2)
+            )
+            SELECT t.c1, t.c2
+            FROM t
+
+        Above, the "anon_1" CTE is not referenced in the SELECT
+        statement, however still accomplishes the task of running an INSERT
+        statement.
+
+        Similarly in a DML-related context, using the PostgreSQL
+        :class:`_postgresql.Insert` construct to generate an "upsert"::
+
+            from sqlalchemy import table, column
+            from sqlalchemy.dialects.postgresql import insert
+
+            t = table("t", column("c1"), column("c2"))
+
+            delete_statement_cte = t.delete().where(t.c.c1 < 1).cte("deletions")
+
+            insert_stmt = insert(t).values({"c1": 1, "c2": 2})
+            update_statement = insert_stmt.on_conflict_do_update(
+                index_elements=[t.c.c1],
+                set_={
+                    "c1": insert_stmt.excluded.c1,
+                    "c2": insert_stmt.excluded.c2,
+                },
+            ).add_cte(delete_statement_cte)
+
+            print(update_statement)
+
+        The above statement renders as:
+
+        .. sourcecode:: sql
+
+            WITH deletions AS (
+                DELETE FROM t WHERE t.c1 < %(c1_1)s
+            )
+            INSERT INTO t (c1, c2) VALUES (%(c1)s, %(c2)s)
+            ON CONFLICT (c1) DO UPDATE SET c1 = excluded.c1, c2 = excluded.c2
+
+        .. versionadded:: 1.4.21
+
+        :param \*ctes: zero or more :class:`.CTE` constructs.
+
+         .. versionchanged:: 2.0  Multiple CTE instances are accepted
+
+        :param nest_here: if True, the given CTE or CTEs will be rendered
+         as though they specified the :paramref:`.HasCTE.cte.nesting` flag
+         to ``True`` when they were added to this :class:`.HasCTE`.
+         Assuming the given CTEs are not referenced in an outer-enclosing
+         statement as well, the CTEs given should render at the level of
+         this statement when this flag is given.
+
+         .. versionadded:: 2.0
+
+         .. seealso::
+
+            :paramref:`.HasCTE.cte.nesting`
+
+
+        """  # noqa: E501
+        opt = _CTEOpts(nest_here)
+        for cte in ctes:
+            cte = coercions.expect(roles.IsCTERole, cte)
+            self._independent_ctes += (cte,)
+            self._independent_ctes_opts += (opt,)
+        return self
+
+    def cte(
+        self,
+        name: Optional[str] = None,
+        recursive: bool = False,
+        nesting: bool = False,
+    ) -> CTE:
+        r"""Return a new :class:`_expression.CTE`,
+        or Common Table Expression instance.
+
+        Common table expressions are a SQL standard whereby SELECT
+        statements can draw upon secondary statements specified along
+        with the primary statement, using a clause called "WITH".
+        Special semantics regarding UNION can also be employed to
+        allow "recursive" queries, where a SELECT statement can draw
+        upon the set of rows that have previously been selected.
+
+        CTEs can also be applied to DML constructs UPDATE, INSERT
+        and DELETE on some databases, both as a source of CTE rows
+        when combined with RETURNING, as well as a consumer of
+        CTE rows.
+
+        SQLAlchemy detects :class:`_expression.CTE` objects, which are treated
+        similarly to :class:`_expression.Alias` objects, as special elements
+        to be delivered to the FROM clause of the statement as well
+        as to a WITH clause at the top of the statement.
+
+        For special prefixes such as PostgreSQL "MATERIALIZED" and
+        "NOT MATERIALIZED", the :meth:`_expression.CTE.prefix_with`
+        method may be
+        used to establish these.
+
+        .. versionchanged:: 1.3.13 Added support for prefixes.
+           In particular - MATERIALIZED and NOT MATERIALIZED.
+
+        :param name: name given to the common table expression.  Like
+         :meth:`_expression.FromClause.alias`, the name can be left as
+         ``None`` in which case an anonymous symbol will be used at query
+         compile time.
+        :param recursive: if ``True``, will render ``WITH RECURSIVE``.
+         A recursive common table expression is intended to be used in
+         conjunction with UNION ALL in order to derive rows
+         from those already selected.
+        :param nesting: if ``True``, will render the CTE locally to the
+         statement in which it is referenced.   For more complex scenarios,
+         the :meth:`.HasCTE.add_cte` method using the
+         :paramref:`.HasCTE.add_cte.nest_here`
+         parameter may also be used to more carefully
+         control the exact placement of a particular CTE.
+
+         .. versionadded:: 1.4.24
+
+         .. seealso::
+
+            :meth:`.HasCTE.add_cte`
+
+        The following examples include two from PostgreSQL's documentation at
+        https://www.postgresql.org/docs/current/static/queries-with.html,
+        as well as additional examples.
+
+        Example 1, non recursive::
+
+            from sqlalchemy import (
+                Table,
+                Column,
+                String,
+                Integer,
+                MetaData,
+                select,
+                func,
+            )
+
+            metadata = MetaData()
+
+            orders = Table(
+                "orders",
+                metadata,
+                Column("region", String),
+                Column("amount", Integer),
+                Column("product", String),
+                Column("quantity", Integer),
+            )
+
+            regional_sales = (
+                select(orders.c.region, func.sum(orders.c.amount).label("total_sales"))
+                .group_by(orders.c.region)
+                .cte("regional_sales")
+            )
+
+
+            top_regions = (
+                select(regional_sales.c.region)
+                .where(
+                    regional_sales.c.total_sales
+                    > select(func.sum(regional_sales.c.total_sales) / 10)
+                )
+                .cte("top_regions")
+            )
+
+            statement = (
+                select(
+                    orders.c.region,
+                    orders.c.product,
+                    func.sum(orders.c.quantity).label("product_units"),
+                    func.sum(orders.c.amount).label("product_sales"),
+                )
+                .where(orders.c.region.in_(select(top_regions.c.region)))
+                .group_by(orders.c.region, orders.c.product)
+            )
+
+            result = conn.execute(statement).fetchall()
+
+        Example 2, WITH RECURSIVE::
+
+            from sqlalchemy import (
+                Table,
+                Column,
+                String,
+                Integer,
+                MetaData,
+                select,
+                func,
+            )
+
+            metadata = MetaData()
+
+            parts = Table(
+                "parts",
+                metadata,
+                Column("part", String),
+                Column("sub_part", String),
+                Column("quantity", Integer),
+            )
+
+            included_parts = (
+                select(parts.c.sub_part, parts.c.part, parts.c.quantity)
+                .where(parts.c.part == "our part")
+                .cte(recursive=True)
+            )
+
+
+            incl_alias = included_parts.alias()
+            parts_alias = parts.alias()
+            included_parts = included_parts.union_all(
+                select(
+                    parts_alias.c.sub_part, parts_alias.c.part, parts_alias.c.quantity
+                ).where(parts_alias.c.part == incl_alias.c.sub_part)
+            )
+
+            statement = select(
+                included_parts.c.sub_part,
+                func.sum(included_parts.c.quantity).label("total_quantity"),
+            ).group_by(included_parts.c.sub_part)
+
+            result = conn.execute(statement).fetchall()
+
+        Example 3, an upsert using UPDATE and INSERT with CTEs::
+
+            from datetime import date
+            from sqlalchemy import (
+                MetaData,
+                Table,
+                Column,
+                Integer,
+                Date,
+                select,
+                literal,
+                and_,
+                exists,
+            )
+
+            metadata = MetaData()
+
+            visitors = Table(
+                "visitors",
+                metadata,
+                Column("product_id", Integer, primary_key=True),
+                Column("date", Date, primary_key=True),
+                Column("count", Integer),
+            )
+
+            # add 5 visitors for the product_id == 1
+            product_id = 1
+            day = date.today()
+            count = 5
+
+            update_cte = (
+                visitors.update()
+                .where(
+                    and_(visitors.c.product_id == product_id, visitors.c.date == day)
+                )
+                .values(count=visitors.c.count + count)
+                .returning(literal(1))
+                .cte("update_cte")
+            )
+
+            upsert = visitors.insert().from_select(
+                [visitors.c.product_id, visitors.c.date, visitors.c.count],
+                select(literal(product_id), literal(day), literal(count)).where(
+                    ~exists(update_cte.select())
+                ),
+            )
+
+            connection.execute(upsert)
+
+        Example 4, Nesting CTE (SQLAlchemy 1.4.24 and above)::
+
+            value_a = select(literal("root").label("n")).cte("value_a")
+
+            # A nested CTE with the same name as the root one
+            value_a_nested = select(literal("nesting").label("n")).cte(
+                "value_a", nesting=True
+            )
+
+            # Nesting CTEs takes ascendency locally
+            # over the CTEs at a higher level
+            value_b = select(value_a_nested.c.n).cte("value_b")
+
+            value_ab = select(value_a.c.n.label("a"), value_b.c.n.label("b"))
+
+        The above query will render the second CTE nested inside the first,
+        shown with inline parameters below as:
+
+        .. sourcecode:: sql
+
+            WITH
+                value_a AS
+                    (SELECT 'root' AS n),
+                value_b AS
+                    (WITH value_a AS
+                        (SELECT 'nesting' AS n)
+                    SELECT value_a.n AS n FROM value_a)
+            SELECT value_a.n AS a, value_b.n AS b
+            FROM value_a, value_b
+
+        The same CTE can be set up using the :meth:`.HasCTE.add_cte` method
+        as follows (SQLAlchemy 2.0 and above)::
+
+            value_a = select(literal("root").label("n")).cte("value_a")
+
+            # A nested CTE with the same name as the root one
+            value_a_nested = select(literal("nesting").label("n")).cte("value_a")
+
+            # Nesting CTEs takes ascendency locally
+            # over the CTEs at a higher level
+            value_b = (
+                select(value_a_nested.c.n)
+                .add_cte(value_a_nested, nest_here=True)
+                .cte("value_b")
+            )
+
+            value_ab = select(value_a.c.n.label("a"), value_b.c.n.label("b"))
+
+        Example 5, Non-Linear CTE (SQLAlchemy 1.4.28 and above)::
+
+            edge = Table(
+                "edge",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("left", Integer),
+                Column("right", Integer),
+            )
+
+            root_node = select(literal(1).label("node")).cte("nodes", recursive=True)
+
+            left_edge = select(edge.c.left).join(
+                root_node, edge.c.right == root_node.c.node
+            )
+            right_edge = select(edge.c.right).join(
+                root_node, edge.c.left == root_node.c.node
+            )
+
+            subgraph_cte = root_node.union(left_edge, right_edge)
+
+            subgraph = select(subgraph_cte)
+
+        The above query will render 2 UNIONs inside the recursive CTE:
+
+        .. sourcecode:: sql
+
+            WITH RECURSIVE nodes(node) AS (
+                    SELECT 1 AS node
+                UNION
+                    SELECT edge."left" AS "left"
+                    FROM edge JOIN nodes ON edge."right" = nodes.node
+                UNION
+                    SELECT edge."right" AS "right"
+                    FROM edge JOIN nodes ON edge."left" = nodes.node
+            )
+            SELECT nodes.node FROM nodes
+
+        .. seealso::
+
+            :meth:`_orm.Query.cte` - ORM version of
+            :meth:`_expression.HasCTE.cte`.
+
+        """  # noqa: E501
+        return CTE._construct(
+            self, name=name, recursive=recursive, nesting=nesting
+        )
+
+
+class Subquery(AliasedReturnsRows):
+    """Represent a subquery of a SELECT.
+
+    A :class:`.Subquery` is created by invoking the
+    :meth:`_expression.SelectBase.subquery` method, or for convenience the
+    :meth:`_expression.SelectBase.alias` method, on any
+    :class:`_expression.SelectBase` subclass
+    which includes :class:`_expression.Select`,
+    :class:`_expression.CompoundSelect`, and
+    :class:`_expression.TextualSelect`.  As rendered in a FROM clause,
+    it represents the
+    body of the SELECT statement inside of parenthesis, followed by the usual
+    "AS <somename>" that defines all "alias" objects.
+
+    The :class:`.Subquery` object is very similar to the
+    :class:`_expression.Alias`
+    object and can be used in an equivalent way.    The difference between
+    :class:`_expression.Alias` and :class:`.Subquery` is that
+    :class:`_expression.Alias` always
+    contains a :class:`_expression.FromClause` object whereas
+    :class:`.Subquery`
+    always contains a :class:`_expression.SelectBase` object.
+
+    .. versionadded:: 1.4 The :class:`.Subquery` class was added which now
+       serves the purpose of providing an aliased version of a SELECT
+       statement.
+
+    """
+
+    __visit_name__ = "subquery"
+
+    _is_subquery = True
+
+    inherit_cache = True
+
+    element: SelectBase
+
+    @classmethod
+    def _factory(
+        cls, selectable: SelectBase, name: Optional[str] = None
+    ) -> Subquery:
+        """Return a :class:`.Subquery` object."""
+
+        return coercions.expect(
+            roles.SelectStatementRole, selectable
+        ).subquery(name=name)
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`.Subquery.as_scalar` method, which was previously "
+        "``Alias.as_scalar()`` prior to version 1.4, is deprecated and "
+        "will be removed in a future release; Please use the "
+        ":meth:`_expression.Select.scalar_subquery` method of the "
+        ":func:`_expression.select` "
+        "construct before constructing a subquery object, or with the ORM "
+        "use the :meth:`_query.Query.scalar_subquery` method.",
+    )
+    def as_scalar(self) -> ScalarSelect[Any]:
+        return self.element.set_label_style(LABEL_STYLE_NONE).scalar_subquery()
+
+
+class FromGrouping(GroupedElement, FromClause):
+    """Represent a grouping of a FROM clause"""
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement)
+    ]
+
+    element: FromClause
+
+    def __init__(self, element: FromClause):
+        self.element = coercions.expect(roles.FromClauseRole, element)
+
+    @util.ro_non_memoized_property
+    def columns(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        return self.element.columns
+
+    @util.ro_non_memoized_property
+    def c(self) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        return self.element.columns
+
+    @property
+    def primary_key(self) -> Iterable[NamedColumn[Any]]:
+        return self.element.primary_key
+
+    @property
+    def foreign_keys(self) -> Iterable[ForeignKey]:
+        return self.element.foreign_keys
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        return self.element.is_derived_from(fromclause)
+
+    def alias(
+        self, name: Optional[str] = None, flat: bool = False
+    ) -> NamedFromGrouping:
+        return NamedFromGrouping(self.element.alias(name=name, flat=flat))
+
+    def _anonymous_fromclause(self, **kw: Any) -> FromGrouping:
+        return FromGrouping(self.element._anonymous_fromclause(**kw))
+
+    @util.ro_non_memoized_property
+    def _hide_froms(self) -> Iterable[FromClause]:
+        return self.element._hide_froms
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.element._from_objects
+
+    def __getstate__(self) -> Dict[str, FromClause]:
+        return {"element": self.element}
+
+    def __setstate__(self, state: Dict[str, FromClause]) -> None:
+        self.element = state["element"]
+
+    if TYPE_CHECKING:
+
+        def self_group(
+            self, against: Optional[OperatorType] = None
+        ) -> Self: ...
+
+
+class NamedFromGrouping(FromGrouping, NamedFromClause):
+    """represent a grouping of a named FROM clause
+
+    .. versionadded:: 2.0
+
+    """
+
+    inherit_cache = True
+
+    if TYPE_CHECKING:
+
+        def self_group(
+            self, against: Optional[OperatorType] = None
+        ) -> Self: ...
+
+
+class TableClause(roles.DMLTableRole, Immutable, NamedFromClause):
+    """Represents a minimal "table" construct.
+
+    This is a lightweight table object that has only a name, a
+    collection of columns, which are typically produced
+    by the :func:`_expression.column` function, and a schema::
+
+        from sqlalchemy import table, column
+
+        user = table(
+            "user",
+            column("id"),
+            column("name"),
+            column("description"),
+        )
+
+    The :class:`_expression.TableClause` construct serves as the base for
+    the more commonly used :class:`_schema.Table` object, providing
+    the usual set of :class:`_expression.FromClause` services including
+    the ``.c.`` collection and statement generation methods.
+
+    It does **not** provide all the additional schema-level services
+    of :class:`_schema.Table`, including constraints, references to other
+    tables, or support for :class:`_schema.MetaData`-level services.
+    It's useful
+    on its own as an ad-hoc construct used to generate quick SQL
+    statements when a more fully fledged :class:`_schema.Table`
+    is not on hand.
+
+    """
+
+    __visit_name__ = "table"
+
+    _traverse_internals: _TraverseInternalsType = [
+        (
+            "columns",
+            InternalTraversal.dp_fromclause_canonical_column_collection,
+        ),
+        ("name", InternalTraversal.dp_string),
+        ("schema", InternalTraversal.dp_string),
+    ]
+
+    _is_table = True
+
+    fullname: str
+
+    implicit_returning = False
+    """:class:`_expression.TableClause`
+    doesn't support having a primary key or column
+    -level defaults, so implicit returning doesn't apply."""
+
+    @util.ro_memoized_property
+    def _autoincrement_column(self) -> Optional[ColumnClause[Any]]:
+        """No PK or default support so no autoincrement column."""
+        return None
+
+    def __init__(self, name: str, *columns: ColumnClause[Any], **kw: Any):
+        super().__init__()
+        self.name = name
+        self._columns = DedupeColumnCollection()
+        self.primary_key = ColumnSet()  # type: ignore
+        self.foreign_keys = set()  # type: ignore
+        for c in columns:
+            self.append_column(c)
+
+        schema = kw.pop("schema", None)
+        if schema is not None:
+            self.schema = schema
+        if self.schema is not None:
+            self.fullname = "%s.%s" % (self.schema, self.name)
+        else:
+            self.fullname = self.name
+        if kw:
+            raise exc.ArgumentError("Unsupported argument(s): %s" % list(kw))
+
+    if TYPE_CHECKING:
+
+        @util.ro_non_memoized_property
+        def columns(
+            self,
+        ) -> ReadOnlyColumnCollection[str, ColumnClause[Any]]: ...
+
+        @util.ro_non_memoized_property
+        def c(self) -> ReadOnlyColumnCollection[str, ColumnClause[Any]]: ...
+
+    def __str__(self) -> str:
+        if self.schema is not None:
+            return self.schema + "." + self.name
+        else:
+            return self.name
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        pass
+
+    @util.ro_memoized_property
+    def description(self) -> str:
+        return self.name
+
+    def append_column(self, c: ColumnClause[Any]) -> None:
+        existing = c.table
+        if existing is not None and existing is not self:
+            raise exc.ArgumentError(
+                "column object '%s' already assigned to table '%s'"
+                % (c.key, existing)
+            )
+
+        self._columns.add(c)
+        c.table = self
+
+    @util.preload_module("sqlalchemy.sql.dml")
+    def insert(self) -> util.preloaded.sql_dml.Insert:
+        """Generate an :class:`_sql.Insert` construct against this
+        :class:`_expression.TableClause`.
+
+        E.g.::
+
+            table.insert().values(name="foo")
+
+        See :func:`_expression.insert` for argument and usage information.
+
+        """
+
+        return util.preloaded.sql_dml.Insert(self)
+
+    @util.preload_module("sqlalchemy.sql.dml")
+    def update(self) -> Update:
+        """Generate an :func:`_expression.update` construct against this
+        :class:`_expression.TableClause`.
+
+        E.g.::
+
+            table.update().where(table.c.id == 7).values(name="foo")
+
+        See :func:`_expression.update` for argument and usage information.
+
+        """
+        return util.preloaded.sql_dml.Update(
+            self,
+        )
+
+    @util.preload_module("sqlalchemy.sql.dml")
+    def delete(self) -> Delete:
+        """Generate a :func:`_expression.delete` construct against this
+        :class:`_expression.TableClause`.
+
+        E.g.::
+
+            table.delete().where(table.c.id == 7)
+
+        See :func:`_expression.delete` for argument and usage information.
+
+        """
+        return util.preloaded.sql_dml.Delete(self)
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return [self]
+
+
+ForUpdateParameter = Union["ForUpdateArg", None, bool, Dict[str, Any]]
+
+
+class ForUpdateArg(ClauseElement):
+    _traverse_internals: _TraverseInternalsType = [
+        ("of", InternalTraversal.dp_clauseelement_list),
+        ("nowait", InternalTraversal.dp_boolean),
+        ("read", InternalTraversal.dp_boolean),
+        ("skip_locked", InternalTraversal.dp_boolean),
+        ("key_share", InternalTraversal.dp_boolean),
+    ]
+
+    of: Optional[Sequence[ClauseElement]]
+    nowait: bool
+    read: bool
+    skip_locked: bool
+
+    @classmethod
+    def _from_argument(
+        cls, with_for_update: ForUpdateParameter
+    ) -> Optional[ForUpdateArg]:
+        if isinstance(with_for_update, ForUpdateArg):
+            return with_for_update
+        elif with_for_update in (None, False):
+            return None
+        elif with_for_update is True:
+            return ForUpdateArg()
+        else:
+            return ForUpdateArg(**cast("Dict[str, Any]", with_for_update))
+
+    def __eq__(self, other: Any) -> bool:
+        return (
+            isinstance(other, ForUpdateArg)
+            and other.nowait == self.nowait
+            and other.read == self.read
+            and other.skip_locked == self.skip_locked
+            and other.key_share == self.key_share
+            and other.of is self.of
+        )
+
+    def __ne__(self, other: Any) -> bool:
+        return not self.__eq__(other)
+
+    def __hash__(self) -> int:
+        return id(self)
+
+    def __init__(
+        self,
+        *,
+        nowait: bool = False,
+        read: bool = False,
+        of: Optional[_ForUpdateOfArgument] = None,
+        skip_locked: bool = False,
+        key_share: bool = False,
+    ):
+        """Represents arguments specified to
+        :meth:`_expression.Select.for_update`.
+
+        """
+
+        self.nowait = nowait
+        self.read = read
+        self.skip_locked = skip_locked
+        self.key_share = key_share
+        if of is not None:
+            self.of = [
+                coercions.expect(roles.ColumnsClauseRole, elem)
+                for elem in util.to_list(of)
+            ]
+        else:
+            self.of = None
+
+
+class Values(roles.InElementRole, Generative, LateralFromClause):
+    """Represent a ``VALUES`` construct that can be used as a FROM element
+    in a statement.
+
+    The :class:`_expression.Values` object is created from the
+    :func:`_expression.values` function.
+
+    .. versionadded:: 1.4
+
+    """
+
+    __visit_name__ = "values"
+
+    _data: Tuple[Sequence[Tuple[Any, ...]], ...] = ()
+
+    _unnamed: bool
+    _traverse_internals: _TraverseInternalsType = [
+        ("_column_args", InternalTraversal.dp_clauseelement_list),
+        ("_data", InternalTraversal.dp_dml_multi_values),
+        ("name", InternalTraversal.dp_string),
+        ("literal_binds", InternalTraversal.dp_boolean),
+    ]
+
+    def __init__(
+        self,
+        *columns: ColumnClause[Any],
+        name: Optional[str] = None,
+        literal_binds: bool = False,
+    ):
+        super().__init__()
+        self._column_args = columns
+
+        if name is None:
+            self._unnamed = True
+            self.name = _anonymous_label.safe_construct(id(self), "anon")
+        else:
+            self._unnamed = False
+            self.name = name
+        self.literal_binds = literal_binds
+        self.named_with_column = not self._unnamed
+
+    @property
+    def _column_types(self) -> List[TypeEngine[Any]]:
+        return [col.type for col in self._column_args]
+
+    @_generative
+    def alias(self, name: Optional[str] = None, flat: bool = False) -> Self:
+        """Return a new :class:`_expression.Values`
+        construct that is a copy of this
+        one with the given name.
+
+        This method is a VALUES-specific specialization of the
+        :meth:`_expression.FromClause.alias` method.
+
+        .. seealso::
+
+            :ref:`tutorial_using_aliases`
+
+            :func:`_expression.alias`
+
+        """
+        non_none_name: str
+
+        if name is None:
+            non_none_name = _anonymous_label.safe_construct(id(self), "anon")
+        else:
+            non_none_name = name
+
+        self.name = non_none_name
+        self.named_with_column = True
+        self._unnamed = False
+        return self
+
+    @_generative
+    def lateral(self, name: Optional[str] = None) -> LateralFromClause:
+        """Return a new :class:`_expression.Values` with the lateral flag set,
+        so that
+        it renders as LATERAL.
+
+        .. seealso::
+
+            :func:`_expression.lateral`
+
+        """
+        non_none_name: str
+
+        if name is None:
+            non_none_name = self.name
+        else:
+            non_none_name = name
+
+        self._is_lateral = True
+        self.name = non_none_name
+        self._unnamed = False
+        return self
+
+    @_generative
+    def data(self, values: Sequence[Tuple[Any, ...]]) -> Self:
+        """Return a new :class:`_expression.Values` construct,
+        adding the given data to the data list.
+
+        E.g.::
+
+            my_values = my_values.data([(1, "value 1"), (2, "value2")])
+
+        :param values: a sequence (i.e. list) of tuples that map to the
+         column expressions given in the :class:`_expression.Values`
+         constructor.
+
+        """
+
+        self._data += (values,)
+        return self
+
+    def scalar_values(self) -> ScalarValues:
+        """Returns a scalar ``VALUES`` construct that can be used as a
+        COLUMN element in a statement.
+
+        .. versionadded:: 2.0.0b4
+
+        """
+        return ScalarValues(self._column_args, self._data, self.literal_binds)
+
+    def _populate_column_collection(
+        self,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+    ) -> None:
+        for c in self._column_args:
+            if c.table is not None and c.table is not self:
+                _, c = c._make_proxy(
+                    self, primary_key=primary_key, foreign_keys=foreign_keys
+                )
+            else:
+                # if the column was used in other contexts, ensure
+                # no memoizations of other FROM clauses.
+                # see test_values.py -> test_auto_proxy_select_direct_col
+                c._reset_memoizations()
+            columns.add(c)
+            c.table = self
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return [self]
+
+
+class ScalarValues(roles.InElementRole, GroupedElement, ColumnElement[Any]):
+    """Represent a scalar ``VALUES`` construct that can be used as a
+    COLUMN element in a statement.
+
+    The :class:`_expression.ScalarValues` object is created from the
+    :meth:`_expression.Values.scalar_values` method. It's also
+    automatically generated when a :class:`_expression.Values` is used in
+    an ``IN`` or ``NOT IN`` condition.
+
+    .. versionadded:: 2.0.0b4
+
+    """
+
+    __visit_name__ = "scalar_values"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("_column_args", InternalTraversal.dp_clauseelement_list),
+        ("_data", InternalTraversal.dp_dml_multi_values),
+        ("literal_binds", InternalTraversal.dp_boolean),
+    ]
+
+    def __init__(
+        self,
+        columns: Sequence[ColumnClause[Any]],
+        data: Tuple[Sequence[Tuple[Any, ...]], ...],
+        literal_binds: bool,
+    ):
+        super().__init__()
+        self._column_args = columns
+        self._data = data
+        self.literal_binds = literal_binds
+
+    @property
+    def _column_types(self) -> List[TypeEngine[Any]]:
+        return [col.type for col in self._column_args]
+
+    def __clause_element__(self) -> ScalarValues:
+        return self
+
+    if TYPE_CHECKING:
+
+        def self_group(
+            self, against: Optional[OperatorType] = None
+        ) -> Self: ...
+
+
+class SelectBase(
+    roles.SelectStatementRole,
+    roles.DMLSelectRole,
+    roles.CompoundElementRole,
+    roles.InElementRole,
+    HasCTE,
+    SupportsCloneAnnotations,
+    Selectable,
+):
+    """Base class for SELECT statements.
+
+
+    This includes :class:`_expression.Select`,
+    :class:`_expression.CompoundSelect` and
+    :class:`_expression.TextualSelect`.
+
+
+    """
+
+    _is_select_base = True
+    is_select = True
+
+    _label_style: SelectLabelStyle = LABEL_STYLE_NONE
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        self._reset_memoizations()
+
+    @util.ro_non_memoized_property
+    def selected_columns(
+        self,
+    ) -> ColumnCollection[str, ColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        representing the columns that
+        this SELECT statement or similar construct returns in its result set.
+
+        This collection differs from the :attr:`_expression.FromClause.columns`
+        collection of a :class:`_expression.FromClause` in that the columns
+        within this collection cannot be directly nested inside another SELECT
+        statement; a subquery must be applied first which provides for the
+        necessary parenthesization required by SQL.
+
+        .. note::
+
+            The :attr:`_sql.SelectBase.selected_columns` collection does not
+            include expressions established in the columns clause using the
+            :func:`_sql.text` construct; these are silently omitted from the
+            collection. To use plain textual column expressions inside of a
+            :class:`_sql.Select` construct, use the :func:`_sql.literal_column`
+            construct.
+
+        .. seealso::
+
+            :attr:`_sql.Select.selected_columns`
+
+        .. versionadded:: 1.4
+
+        """
+        raise NotImplementedError()
+
+    def _generate_fromclause_column_proxies(
+        self,
+        subquery: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        *,
+        proxy_compound_columns: Optional[
+            Iterable[Sequence[ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        raise NotImplementedError()
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        """A sequence of expressions that correspond to what is rendered
+        in the columns clause, including :class:`_sql.TextClause`
+        constructs.
+
+        .. versionadded:: 1.4.12
+
+        .. seealso::
+
+            :attr:`_sql.SelectBase.exported_columns`
+
+        """
+        raise NotImplementedError()
+
+    @property
+    def exported_columns(
+        self,
+    ) -> ReadOnlyColumnCollection[str, ColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        that represents the "exported"
+        columns of this :class:`_expression.Selectable`, not including
+        :class:`_sql.TextClause` constructs.
+
+        The "exported" columns for a :class:`_expression.SelectBase`
+        object are synonymous
+        with the :attr:`_expression.SelectBase.selected_columns` collection.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_expression.Select.exported_columns`
+
+            :attr:`_expression.Selectable.exported_columns`
+
+            :attr:`_expression.FromClause.exported_columns`
+
+
+        """
+        return self.selected_columns.as_readonly()
+
+    @property
+    @util.deprecated(
+        "1.4",
+        "The :attr:`_expression.SelectBase.c` and "
+        ":attr:`_expression.SelectBase.columns` attributes "
+        "are deprecated and will be removed in a future release; these "
+        "attributes implicitly create a subquery that should be explicit.  "
+        "Please call :meth:`_expression.SelectBase.subquery` "
+        "first in order to create "
+        "a subquery, which then contains this attribute.  To access the "
+        "columns that this SELECT object SELECTs "
+        "from, use the :attr:`_expression.SelectBase.selected_columns` "
+        "attribute.",
+    )
+    def c(self) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        return self._implicit_subquery.columns
+
+    @property
+    def columns(
+        self,
+    ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        return self.c
+
+    def get_label_style(self) -> SelectLabelStyle:
+        """
+        Retrieve the current label style.
+
+        Implemented by subclasses.
+
+        """
+        raise NotImplementedError()
+
+    def set_label_style(self, style: SelectLabelStyle) -> Self:
+        """Return a new selectable with the specified label style.
+
+        Implemented by subclasses.
+
+        """
+
+        raise NotImplementedError()
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_expression.SelectBase.select` method is deprecated "
+        "and will be removed in a future release; this method implicitly "
+        "creates a subquery that should be explicit.  "
+        "Please call :meth:`_expression.SelectBase.subquery` "
+        "first in order to create "
+        "a subquery, which then can be selected.",
+    )
+    def select(self, *arg: Any, **kw: Any) -> Select[Any]:
+        return self._implicit_subquery.select(*arg, **kw)
+
+    @HasMemoized.memoized_attribute
+    def _implicit_subquery(self) -> Subquery:
+        return self.subquery()
+
+    def _scalar_type(self) -> TypeEngine[Any]:
+        raise NotImplementedError()
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_expression.SelectBase.as_scalar` "
+        "method is deprecated and will be "
+        "removed in a future release.  Please refer to "
+        ":meth:`_expression.SelectBase.scalar_subquery`.",
+    )
+    def as_scalar(self) -> ScalarSelect[Any]:
+        return self.scalar_subquery()
+
+    def exists(self) -> Exists:
+        """Return an :class:`_sql.Exists` representation of this selectable,
+        which can be used as a column expression.
+
+        The returned object is an instance of :class:`_sql.Exists`.
+
+        .. seealso::
+
+            :func:`_sql.exists`
+
+            :ref:`tutorial_exists` - in the :term:`2.0 style` tutorial.
+
+        .. versionadded:: 1.4
+
+        """
+        return Exists(self)
+
+    def scalar_subquery(self) -> ScalarSelect[Any]:
+        """Return a 'scalar' representation of this selectable, which can be
+        used as a column expression.
+
+        The returned object is an instance of :class:`_sql.ScalarSelect`.
+
+        Typically, a select statement which has only one column in its columns
+        clause is eligible to be used as a scalar expression.  The scalar
+        subquery can then be used in the WHERE clause or columns clause of
+        an enclosing SELECT.
+
+        Note that the scalar subquery differentiates from the FROM-level
+        subquery that can be produced using the
+        :meth:`_expression.SelectBase.subquery`
+        method.
+
+        .. versionchanged: 1.4 - the ``.as_scalar()`` method was renamed to
+           :meth:`_expression.SelectBase.scalar_subquery`.
+
+        .. seealso::
+
+            :ref:`tutorial_scalar_subquery` - in the 2.0 tutorial
+
+        """
+        if self._label_style is not LABEL_STYLE_NONE:
+            self = self.set_label_style(LABEL_STYLE_NONE)
+
+        return ScalarSelect(self)
+
+    def label(self, name: Optional[str]) -> Label[Any]:
+        """Return a 'scalar' representation of this selectable, embedded as a
+        subquery with a label.
+
+        .. seealso::
+
+            :meth:`_expression.SelectBase.scalar_subquery`.
+
+        """
+        return self.scalar_subquery().label(name)
+
+    def lateral(self, name: Optional[str] = None) -> LateralFromClause:
+        """Return a LATERAL alias of this :class:`_expression.Selectable`.
+
+        The return value is the :class:`_expression.Lateral` construct also
+        provided by the top-level :func:`_expression.lateral` function.
+
+        .. seealso::
+
+            :ref:`tutorial_lateral_correlation` -  overview of usage.
+
+        """
+        return Lateral._factory(self, name)
+
+    def subquery(self, name: Optional[str] = None) -> Subquery:
+        """Return a subquery of this :class:`_expression.SelectBase`.
+
+        A subquery is from a SQL perspective a parenthesized, named
+        construct that can be placed in the FROM clause of another
+        SELECT statement.
+
+        Given a SELECT statement such as::
+
+            stmt = select(table.c.id, table.c.name)
+
+        The above statement might look like:
+
+        .. sourcecode:: sql
+
+            SELECT table.id, table.name FROM table
+
+        The subquery form by itself renders the same way, however when
+        embedded into the FROM clause of another SELECT statement, it becomes
+        a named sub-element::
+
+            subq = stmt.subquery()
+            new_stmt = select(subq)
+
+        The above renders as:
+
+        .. sourcecode:: sql
+
+            SELECT anon_1.id, anon_1.name
+            FROM (SELECT table.id, table.name FROM table) AS anon_1
+
+        Historically, :meth:`_expression.SelectBase.subquery`
+        is equivalent to calling
+        the :meth:`_expression.FromClause.alias`
+        method on a FROM object; however,
+        as a :class:`_expression.SelectBase`
+        object is not directly  FROM object,
+        the :meth:`_expression.SelectBase.subquery`
+        method provides clearer semantics.
+
+        .. versionadded:: 1.4
+
+        """
+
+        return Subquery._construct(
+            self._ensure_disambiguated_names(), name=name
+        )
+
+    def _ensure_disambiguated_names(self) -> Self:
+        """Ensure that the names generated by this selectbase will be
+        disambiguated in some way, if possible.
+
+        """
+
+        raise NotImplementedError()
+
+    def alias(
+        self, name: Optional[str] = None, flat: bool = False
+    ) -> Subquery:
+        """Return a named subquery against this
+        :class:`_expression.SelectBase`.
+
+        For a :class:`_expression.SelectBase` (as opposed to a
+        :class:`_expression.FromClause`),
+        this returns a :class:`.Subquery` object which behaves mostly the
+        same as the :class:`_expression.Alias` object that is used with a
+        :class:`_expression.FromClause`.
+
+        .. versionchanged:: 1.4 The :meth:`_expression.SelectBase.alias`
+           method is now
+           a synonym for the :meth:`_expression.SelectBase.subquery` method.
+
+        """
+        return self.subquery(name=name)
+
+
+_SB = TypeVar("_SB", bound=SelectBase)
+
+
+class SelectStatementGrouping(GroupedElement, SelectBase, Generic[_SB]):
+    """Represent a grouping of a :class:`_expression.SelectBase`.
+
+    This differs from :class:`.Subquery` in that we are still
+    an "inner" SELECT statement, this is strictly for grouping inside of
+    compound selects.
+
+    """
+
+    __visit_name__ = "select_statement_grouping"
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement)
+    ] + SupportsCloneAnnotations._clone_annotations_traverse_internals
+
+    _is_select_container = True
+
+    element: _SB
+
+    def __init__(self, element: _SB) -> None:
+        self.element = cast(
+            _SB, coercions.expect(roles.SelectStatementRole, element)
+        )
+
+    def _ensure_disambiguated_names(self) -> SelectStatementGrouping[_SB]:
+        new_element = self.element._ensure_disambiguated_names()
+        if new_element is not self.element:
+            return SelectStatementGrouping(new_element)
+        else:
+            return self
+
+    def get_label_style(self) -> SelectLabelStyle:
+        return self.element.get_label_style()
+
+    def set_label_style(
+        self, label_style: SelectLabelStyle
+    ) -> SelectStatementGrouping[_SB]:
+        return SelectStatementGrouping(
+            self.element.set_label_style(label_style)
+        )
+
+    @property
+    def select_statement(self) -> _SB:
+        return self.element
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        return self
+
+    if TYPE_CHECKING:
+
+        def _ungroup(self) -> _SB: ...
+
+    # def _generate_columns_plus_names(
+    #    self, anon_for_dupe_key: bool
+    # ) -> List[Tuple[str, str, str, ColumnElement[Any], bool]]:
+    #    return self.element._generate_columns_plus_names(anon_for_dupe_key)
+
+    def _generate_fromclause_column_proxies(
+        self,
+        subquery: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        *,
+        proxy_compound_columns: Optional[
+            Iterable[Sequence[ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        self.element._generate_fromclause_column_proxies(
+            subquery,
+            columns,
+            proxy_compound_columns=proxy_compound_columns,
+            primary_key=primary_key,
+            foreign_keys=foreign_keys,
+        )
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        return self.element._all_selected_columns
+
+    @util.ro_non_memoized_property
+    def selected_columns(self) -> ColumnCollection[str, ColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        representing the columns that
+        the embedded SELECT statement returns in its result set, not including
+        :class:`_sql.TextClause` constructs.
+
+        .. versionadded:: 1.4
+
+        .. seealso::
+
+            :attr:`_sql.Select.selected_columns`
+
+        """
+        return self.element.selected_columns
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return self.element._from_objects
+
+    def add_cte(self, *ctes: CTE, nest_here: bool = False) -> Self:
+        # SelectStatementGrouping not generative: has no attribute '_generate'
+        raise NotImplementedError
+
+
+class GenerativeSelect(SelectBase, Generative):
+    """Base class for SELECT statements where additional elements can be
+    added.
+
+    This serves as the base for :class:`_expression.Select` and
+    :class:`_expression.CompoundSelect`
+    where elements such as ORDER BY, GROUP BY can be added and column
+    rendering can be controlled.  Compare to
+    :class:`_expression.TextualSelect`, which,
+    while it subclasses :class:`_expression.SelectBase`
+    and is also a SELECT construct,
+    represents a fixed textual string which cannot be altered at this level,
+    only wrapped as a subquery.
+
+    """
+
+    _order_by_clauses: Tuple[ColumnElement[Any], ...] = ()
+    _group_by_clauses: Tuple[ColumnElement[Any], ...] = ()
+    _limit_clause: Optional[ColumnElement[Any]] = None
+    _offset_clause: Optional[ColumnElement[Any]] = None
+    _fetch_clause: Optional[ColumnElement[Any]] = None
+    _fetch_clause_options: Optional[Dict[str, bool]] = None
+    _for_update_arg: Optional[ForUpdateArg] = None
+
+    def __init__(self, _label_style: SelectLabelStyle = LABEL_STYLE_DEFAULT):
+        self._label_style = _label_style
+
+    @_generative
+    def with_for_update(
+        self,
+        *,
+        nowait: bool = False,
+        read: bool = False,
+        of: Optional[_ForUpdateOfArgument] = None,
+        skip_locked: bool = False,
+        key_share: bool = False,
+    ) -> Self:
+        """Specify a ``FOR UPDATE`` clause for this
+        :class:`_expression.GenerativeSelect`.
+
+        E.g.::
+
+            stmt = select(table).with_for_update(nowait=True)
+
+        On a database like PostgreSQL or Oracle Database, the above would
+        render a statement like:
+
+        .. sourcecode:: sql
+
+            SELECT table.a, table.b FROM table FOR UPDATE NOWAIT
+
+        on other backends, the ``nowait`` option is ignored and instead
+        would produce:
+
+        .. sourcecode:: sql
+
+            SELECT table.a, table.b FROM table FOR UPDATE
+
+        When called with no arguments, the statement will render with
+        the suffix ``FOR UPDATE``.   Additional arguments can then be
+        provided which allow for common database-specific
+        variants.
+
+        :param nowait: boolean; will render ``FOR UPDATE NOWAIT`` on Oracle
+         Database and PostgreSQL dialects.
+
+        :param read: boolean; will render ``LOCK IN SHARE MODE`` on MySQL,
+         ``FOR SHARE`` on PostgreSQL.  On PostgreSQL, when combined with
+         ``nowait``, will render ``FOR SHARE NOWAIT``.
+
+        :param of: SQL expression or list of SQL expression elements,
+         (typically :class:`_schema.Column` objects or a compatible expression,
+         for some backends may also be a table expression) which will render
+         into a ``FOR UPDATE OF`` clause; supported by PostgreSQL, Oracle
+         Database, some MySQL versions and possibly others. May render as a
+         table or as a column depending on backend.
+
+        :param skip_locked: boolean, will render ``FOR UPDATE SKIP LOCKED`` on
+         Oracle Database and PostgreSQL dialects or ``FOR SHARE SKIP LOCKED``
+         if ``read=True`` is also specified.
+
+        :param key_share: boolean, will render ``FOR NO KEY UPDATE``,
+         or if combined with ``read=True`` will render ``FOR KEY SHARE``,
+         on the PostgreSQL dialect.
+
+        """
+        self._for_update_arg = ForUpdateArg(
+            nowait=nowait,
+            read=read,
+            of=of,
+            skip_locked=skip_locked,
+            key_share=key_share,
+        )
+        return self
+
+    def get_label_style(self) -> SelectLabelStyle:
+        """
+        Retrieve the current label style.
+
+        .. versionadded:: 1.4
+
+        """
+        return self._label_style
+
+    def set_label_style(self, style: SelectLabelStyle) -> Self:
+        """Return a new selectable with the specified label style.
+
+        There are three "label styles" available,
+        :attr:`_sql.SelectLabelStyle.LABEL_STYLE_DISAMBIGUATE_ONLY`,
+        :attr:`_sql.SelectLabelStyle.LABEL_STYLE_TABLENAME_PLUS_COL`, and
+        :attr:`_sql.SelectLabelStyle.LABEL_STYLE_NONE`.   The default style is
+        :attr:`_sql.SelectLabelStyle.LABEL_STYLE_DISAMBIGUATE_ONLY`.
+
+        In modern SQLAlchemy, there is not generally a need to change the
+        labeling style, as per-expression labels are more effectively used by
+        making use of the :meth:`_sql.ColumnElement.label` method. In past
+        versions, :data:`_sql.LABEL_STYLE_TABLENAME_PLUS_COL` was used to
+        disambiguate same-named columns from different tables, aliases, or
+        subqueries; the newer :data:`_sql.LABEL_STYLE_DISAMBIGUATE_ONLY` now
+        applies labels only to names that conflict with an existing name so
+        that the impact of this labeling is minimal.
+
+        The rationale for disambiguation is mostly so that all column
+        expressions are available from a given :attr:`_sql.FromClause.c`
+        collection when a subquery is created.
+
+        .. versionadded:: 1.4 - the
+            :meth:`_sql.GenerativeSelect.set_label_style` method replaces the
+            previous combination of ``.apply_labels()``, ``.with_labels()`` and
+            ``use_labels=True`` methods and/or parameters.
+
+        .. seealso::
+
+            :data:`_sql.LABEL_STYLE_DISAMBIGUATE_ONLY`
+
+            :data:`_sql.LABEL_STYLE_TABLENAME_PLUS_COL`
+
+            :data:`_sql.LABEL_STYLE_NONE`
+
+            :data:`_sql.LABEL_STYLE_DEFAULT`
+
+        """
+        if self._label_style is not style:
+            self = self._generate()
+            self._label_style = style
+        return self
+
+    @property
+    def _group_by_clause(self) -> ClauseList:
+        """ClauseList access to group_by_clauses for legacy dialects"""
+        return ClauseList._construct_raw(
+            operators.comma_op, self._group_by_clauses
+        )
+
+    @property
+    def _order_by_clause(self) -> ClauseList:
+        """ClauseList access to order_by_clauses for legacy dialects"""
+        return ClauseList._construct_raw(
+            operators.comma_op, self._order_by_clauses
+        )
+
+    def _offset_or_limit_clause(
+        self,
+        element: _LimitOffsetType,
+        name: Optional[str] = None,
+        type_: Optional[_TypeEngineArgument[int]] = None,
+    ) -> ColumnElement[Any]:
+        """Convert the given value to an "offset or limit" clause.
+
+        This handles incoming integers and converts to an expression; if
+        an expression is already given, it is passed through.
+
+        """
+        return coercions.expect(
+            roles.LimitOffsetRole, element, name=name, type_=type_
+        )
+
+    @overload
+    def _offset_or_limit_clause_asint(
+        self, clause: ColumnElement[Any], attrname: str
+    ) -> NoReturn: ...
+
+    @overload
+    def _offset_or_limit_clause_asint(
+        self, clause: Optional[_OffsetLimitParam], attrname: str
+    ) -> Optional[int]: ...
+
+    def _offset_or_limit_clause_asint(
+        self, clause: Optional[ColumnElement[Any]], attrname: str
+    ) -> Union[NoReturn, Optional[int]]:
+        """Convert the "offset or limit" clause of a select construct to an
+        integer.
+
+        This is only possible if the value is stored as a simple bound
+        parameter. Otherwise, a compilation error is raised.
+
+        """
+        if clause is None:
+            return None
+        try:
+            value = clause._limit_offset_value
+        except AttributeError as err:
+            raise exc.CompileError(
+                "This SELECT structure does not use a simple "
+                "integer value for %s" % attrname
+            ) from err
+        else:
+            return util.asint(value)
+
+    @property
+    def _limit(self) -> Optional[int]:
+        """Get an integer value for the limit.  This should only be used
+        by code that cannot support a limit as a BindParameter or
+        other custom clause as it will throw an exception if the limit
+        isn't currently set to an integer.
+
+        """
+        return self._offset_or_limit_clause_asint(self._limit_clause, "limit")
+
+    def _simple_int_clause(self, clause: ClauseElement) -> bool:
+        """True if the clause is a simple integer, False
+        if it is not present or is a SQL expression.
+        """
+        return isinstance(clause, _OffsetLimitParam)
+
+    @property
+    def _offset(self) -> Optional[int]:
+        """Get an integer value for the offset.  This should only be used
+        by code that cannot support an offset as a BindParameter or
+        other custom clause as it will throw an exception if the
+        offset isn't currently set to an integer.
+
+        """
+        return self._offset_or_limit_clause_asint(
+            self._offset_clause, "offset"
+        )
+
+    @property
+    def _has_row_limiting_clause(self) -> bool:
+        return (
+            self._limit_clause is not None
+            or self._offset_clause is not None
+            or self._fetch_clause is not None
+        )
+
+    @_generative
+    def limit(self, limit: _LimitOffsetType) -> Self:
+        """Return a new selectable with the given LIMIT criterion
+        applied.
+
+        This is a numerical value which usually renders as a ``LIMIT``
+        expression in the resulting select.  Backends that don't
+        support ``LIMIT`` will attempt to provide similar
+        functionality.
+
+        .. note::
+
+           The :meth:`_sql.GenerativeSelect.limit` method will replace
+           any clause applied with :meth:`_sql.GenerativeSelect.fetch`.
+
+        :param limit: an integer LIMIT parameter, or a SQL expression
+         that provides an integer result. Pass ``None`` to reset it.
+
+        .. seealso::
+
+           :meth:`_sql.GenerativeSelect.fetch`
+
+           :meth:`_sql.GenerativeSelect.offset`
+
+        """
+
+        self._fetch_clause = self._fetch_clause_options = None
+        self._limit_clause = self._offset_or_limit_clause(limit)
+        return self
+
+    @_generative
+    def fetch(
+        self,
+        count: _LimitOffsetType,
+        with_ties: bool = False,
+        percent: bool = False,
+    ) -> Self:
+        """Return a new selectable with the given FETCH FIRST criterion
+        applied.
+
+        This is a numeric value which usually renders as ``FETCH {FIRST | NEXT}
+        [ count ] {ROW | ROWS} {ONLY | WITH TIES}`` expression in the resulting
+        select. This functionality is is currently implemented for Oracle
+        Database, PostgreSQL, MSSQL.
+
+        Use :meth:`_sql.GenerativeSelect.offset` to specify the offset.
+
+        .. note::
+
+           The :meth:`_sql.GenerativeSelect.fetch` method will replace
+           any clause applied with :meth:`_sql.GenerativeSelect.limit`.
+
+        .. versionadded:: 1.4
+
+        :param count: an integer COUNT parameter, or a SQL expression
+         that provides an integer result. When ``percent=True`` this will
+         represent the percentage of rows to return, not the absolute value.
+         Pass ``None`` to reset it.
+
+        :param with_ties: When ``True``, the WITH TIES option is used
+         to return any additional rows that tie for the last place in the
+         result set according to the ``ORDER BY`` clause. The
+         ``ORDER BY`` may be mandatory in this case. Defaults to ``False``
+
+        :param percent: When ``True``, ``count`` represents the percentage
+         of the total number of selected rows to return. Defaults to ``False``
+
+        .. seealso::
+
+           :meth:`_sql.GenerativeSelect.limit`
+
+           :meth:`_sql.GenerativeSelect.offset`
+
+        """
+
+        self._limit_clause = None
+        if count is None:
+            self._fetch_clause = self._fetch_clause_options = None
+        else:
+            self._fetch_clause = self._offset_or_limit_clause(count)
+            self._fetch_clause_options = {
+                "with_ties": with_ties,
+                "percent": percent,
+            }
+        return self
+
+    @_generative
+    def offset(self, offset: _LimitOffsetType) -> Self:
+        """Return a new selectable with the given OFFSET criterion
+        applied.
+
+
+        This is a numeric value which usually renders as an ``OFFSET``
+        expression in the resulting select.  Backends that don't
+        support ``OFFSET`` will attempt to provide similar
+        functionality.
+
+        :param offset: an integer OFFSET parameter, or a SQL expression
+         that provides an integer result. Pass ``None`` to reset it.
+
+        .. seealso::
+
+           :meth:`_sql.GenerativeSelect.limit`
+
+           :meth:`_sql.GenerativeSelect.fetch`
+
+        """
+
+        self._offset_clause = self._offset_or_limit_clause(offset)
+        return self
+
+    @_generative
+    @util.preload_module("sqlalchemy.sql.util")
+    def slice(
+        self,
+        start: int,
+        stop: int,
+    ) -> Self:
+        """Apply LIMIT / OFFSET to this statement based on a slice.
+
+        The start and stop indices behave like the argument to Python's
+        built-in :func:`range` function. This method provides an
+        alternative to using ``LIMIT``/``OFFSET`` to get a slice of the
+        query.
+
+        For example, ::
+
+            stmt = select(User).order_by(User.id).slice(1, 3)
+
+        renders as
+
+        .. sourcecode:: sql
+
+           SELECT users.id AS users_id,
+                  users.name AS users_name
+           FROM users ORDER BY users.id
+           LIMIT ? OFFSET ?
+           (2, 1)
+
+        .. note::
+
+           The :meth:`_sql.GenerativeSelect.slice` method will replace
+           any clause applied with :meth:`_sql.GenerativeSelect.fetch`.
+
+        .. versionadded:: 1.4  Added the :meth:`_sql.GenerativeSelect.slice`
+           method generalized from the ORM.
+
+        .. seealso::
+
+           :meth:`_sql.GenerativeSelect.limit`
+
+           :meth:`_sql.GenerativeSelect.offset`
+
+           :meth:`_sql.GenerativeSelect.fetch`
+
+        """
+        sql_util = util.preloaded.sql_util
+        self._fetch_clause = self._fetch_clause_options = None
+        self._limit_clause, self._offset_clause = sql_util._make_slice(
+            self._limit_clause, self._offset_clause, start, stop
+        )
+        return self
+
+    @_generative
+    def order_by(
+        self,
+        __first: Union[
+            Literal[None, _NoArg.NO_ARG],
+            _ColumnExpressionOrStrLabelArgument[Any],
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionOrStrLabelArgument[Any],
+    ) -> Self:
+        r"""Return a new selectable with the given list of ORDER BY
+        criteria applied.
+
+        e.g.::
+
+            stmt = select(table).order_by(table.c.id, table.c.name)
+
+        Calling this method multiple times is equivalent to calling it once
+        with all the clauses concatenated. All existing ORDER BY criteria may
+        be cancelled by passing ``None`` by itself.  New ORDER BY criteria may
+        then be added by invoking :meth:`_orm.Query.order_by` again, e.g.::
+
+            # will erase all ORDER BY and ORDER BY new_col alone
+            stmt = stmt.order_by(None).order_by(new_col)
+
+        :param \*clauses: a series of :class:`_expression.ColumnElement`
+         constructs
+         which will be used to generate an ORDER BY clause.
+
+        .. seealso::
+
+            :ref:`tutorial_order_by` - in the :ref:`unified_tutorial`
+
+            :ref:`tutorial_order_by_label` - in the :ref:`unified_tutorial`
+
+        """
+
+        if not clauses and __first is None:
+            self._order_by_clauses = ()
+        elif __first is not _NoArg.NO_ARG:
+            self._order_by_clauses += tuple(
+                coercions.expect(
+                    roles.OrderByRole, clause, apply_propagate_attrs=self
+                )
+                for clause in (__first,) + clauses
+            )
+        return self
+
+    @_generative
+    def group_by(
+        self,
+        __first: Union[
+            Literal[None, _NoArg.NO_ARG],
+            _ColumnExpressionOrStrLabelArgument[Any],
+        ] = _NoArg.NO_ARG,
+        *clauses: _ColumnExpressionOrStrLabelArgument[Any],
+    ) -> Self:
+        r"""Return a new selectable with the given list of GROUP BY
+        criterion applied.
+
+        All existing GROUP BY settings can be suppressed by passing ``None``.
+
+        e.g.::
+
+            stmt = select(table.c.name, func.max(table.c.stat)).group_by(table.c.name)
+
+        :param \*clauses: a series of :class:`_expression.ColumnElement`
+         constructs
+         which will be used to generate an GROUP BY clause.
+
+        .. seealso::
+
+            :ref:`tutorial_group_by_w_aggregates` - in the
+            :ref:`unified_tutorial`
+
+            :ref:`tutorial_order_by_label` - in the :ref:`unified_tutorial`
+
+        """  # noqa: E501
+
+        if not clauses and __first is None:
+            self._group_by_clauses = ()
+        elif __first is not _NoArg.NO_ARG:
+            self._group_by_clauses += tuple(
+                coercions.expect(
+                    roles.GroupByRole, clause, apply_propagate_attrs=self
+                )
+                for clause in (__first,) + clauses
+            )
+        return self
+
+
+@CompileState.plugin_for("default", "compound_select")
+class CompoundSelectState(CompileState):
+    @util.memoized_property
+    def _label_resolve_dict(
+        self,
+    ) -> Tuple[
+        Dict[str, ColumnElement[Any]],
+        Dict[str, ColumnElement[Any]],
+        Dict[str, ColumnElement[Any]],
+    ]:
+        # TODO: this is hacky and slow
+        hacky_subquery = self.statement.subquery()
+        hacky_subquery.named_with_column = False
+        d = {c.key: c for c in hacky_subquery.c}
+        return d, d, d
+
+
+class _CompoundSelectKeyword(Enum):
+    UNION = "UNION"
+    UNION_ALL = "UNION ALL"
+    EXCEPT = "EXCEPT"
+    EXCEPT_ALL = "EXCEPT ALL"
+    INTERSECT = "INTERSECT"
+    INTERSECT_ALL = "INTERSECT ALL"
+
+
+class CompoundSelect(HasCompileState, GenerativeSelect, ExecutableReturnsRows):
+    """Forms the basis of ``UNION``, ``UNION ALL``, and other
+    SELECT-based set operations.
+
+
+    .. seealso::
+
+        :func:`_expression.union`
+
+        :func:`_expression.union_all`
+
+        :func:`_expression.intersect`
+
+        :func:`_expression.intersect_all`
+
+        :func:`_expression.except`
+
+        :func:`_expression.except_all`
+
+    """
+
+    __visit_name__ = "compound_select"
+
+    _traverse_internals: _TraverseInternalsType = (
+        [
+            ("selects", InternalTraversal.dp_clauseelement_list),
+            ("_limit_clause", InternalTraversal.dp_clauseelement),
+            ("_offset_clause", InternalTraversal.dp_clauseelement),
+            ("_fetch_clause", InternalTraversal.dp_clauseelement),
+            ("_fetch_clause_options", InternalTraversal.dp_plain_dict),
+            ("_order_by_clauses", InternalTraversal.dp_clauseelement_list),
+            ("_group_by_clauses", InternalTraversal.dp_clauseelement_list),
+            ("_for_update_arg", InternalTraversal.dp_clauseelement),
+            ("keyword", InternalTraversal.dp_string),
+        ]
+        + SupportsCloneAnnotations._clone_annotations_traverse_internals
+        + HasCTE._has_ctes_traverse_internals
+    )
+
+    selects: List[SelectBase]
+
+    _is_from_container = True
+    _auto_correlate = False
+
+    def __init__(
+        self,
+        keyword: _CompoundSelectKeyword,
+        *selects: _SelectStatementForCompoundArgument,
+    ):
+        self.keyword = keyword
+        self.selects = [
+            coercions.expect(
+                roles.CompoundElementRole, s, apply_propagate_attrs=self
+            ).self_group(against=self)
+            for s in selects
+        ]
+
+        GenerativeSelect.__init__(self)
+
+    @classmethod
+    def _create_union(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.UNION, *selects)
+
+    @classmethod
+    def _create_union_all(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.UNION_ALL, *selects)
+
+    @classmethod
+    def _create_except(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.EXCEPT, *selects)
+
+    @classmethod
+    def _create_except_all(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.EXCEPT_ALL, *selects)
+
+    @classmethod
+    def _create_intersect(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.INTERSECT, *selects)
+
+    @classmethod
+    def _create_intersect_all(
+        cls, *selects: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        return CompoundSelect(_CompoundSelectKeyword.INTERSECT_ALL, *selects)
+
+    def _scalar_type(self) -> TypeEngine[Any]:
+        return self.selects[0]._scalar_type()
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> GroupedElement:
+        return SelectStatementGrouping(self)
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        for s in self.selects:
+            if s.is_derived_from(fromclause):
+                return True
+        return False
+
+    def set_label_style(self, style: SelectLabelStyle) -> CompoundSelect:
+        if self._label_style is not style:
+            self = self._generate()
+            select_0 = self.selects[0].set_label_style(style)
+            self.selects = [select_0] + self.selects[1:]
+
+        return self
+
+    def _ensure_disambiguated_names(self) -> CompoundSelect:
+        new_select = self.selects[0]._ensure_disambiguated_names()
+        if new_select is not self.selects[0]:
+            self = self._generate()
+            self.selects = [new_select] + self.selects[1:]
+
+        return self
+
+    def _generate_fromclause_column_proxies(
+        self,
+        subquery: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        *,
+        proxy_compound_columns: Optional[
+            Iterable[Sequence[ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        # this is a slightly hacky thing - the union exports a
+        # column that resembles just that of the *first* selectable.
+        # to get at a "composite" column, particularly foreign keys,
+        # you have to dig through the proxies collection which we
+        # generate below.
+        select_0 = self.selects[0]
+
+        if self._label_style is not LABEL_STYLE_DEFAULT:
+            select_0 = select_0.set_label_style(self._label_style)
+
+        # hand-construct the "_proxies" collection to include all
+        # derived columns place a 'weight' annotation corresponding
+        # to how low in the list of select()s the column occurs, so
+        # that the corresponding_column() operation can resolve
+        # conflicts
+        extra_col_iterator = zip(
+            *[
+                [
+                    c._annotate(dd)
+                    for c in stmt._all_selected_columns
+                    if is_column_element(c)
+                ]
+                for dd, stmt in [
+                    ({"weight": i + 1}, stmt)
+                    for i, stmt in enumerate(self.selects)
+                ]
+            ]
+        )
+
+        # the incoming proxy_compound_columns can be present also if this is
+        # a compound embedded in a compound.  it's probably more appropriate
+        # that we generate new weights local to this nested compound, though
+        # i haven't tried to think what it means for compound nested in
+        # compound
+        select_0._generate_fromclause_column_proxies(
+            subquery,
+            columns,
+            proxy_compound_columns=extra_col_iterator,
+            primary_key=primary_key,
+            foreign_keys=foreign_keys,
+        )
+
+    def _refresh_for_new_column(self, column: ColumnElement[Any]) -> None:
+        super()._refresh_for_new_column(column)
+        for select in self.selects:
+            select._refresh_for_new_column(column)
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        return self.selects[0]._all_selected_columns
+
+    @util.ro_non_memoized_property
+    def selected_columns(
+        self,
+    ) -> ColumnCollection[str, ColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        representing the columns that
+        this SELECT statement or similar construct returns in its result set,
+        not including :class:`_sql.TextClause` constructs.
+
+        For a :class:`_expression.CompoundSelect`, the
+        :attr:`_expression.CompoundSelect.selected_columns`
+        attribute returns the selected
+        columns of the first SELECT statement contained within the series of
+        statements within the set operation.
+
+        .. seealso::
+
+            :attr:`_sql.Select.selected_columns`
+
+        .. versionadded:: 1.4
+
+        """
+        return self.selects[0].selected_columns
+
+
+# backwards compat
+for elem in _CompoundSelectKeyword:
+    setattr(CompoundSelect, elem.name, elem)
+
+
+@CompileState.plugin_for("default", "select")
+class SelectState(util.MemoizedSlots, CompileState):
+    __slots__ = (
+        "from_clauses",
+        "froms",
+        "columns_plus_names",
+        "_label_resolve_dict",
+    )
+
+    if TYPE_CHECKING:
+        default_select_compile_options: CacheableOptions
+    else:
+
+        class default_select_compile_options(CacheableOptions):
+            _cache_key_traversal = []
+
+    if TYPE_CHECKING:
+
+        @classmethod
+        def get_plugin_class(
+            cls, statement: Executable
+        ) -> Type[SelectState]: ...
+
+    def __init__(
+        self,
+        statement: Select[Any],
+        compiler: Optional[SQLCompiler],
+        **kw: Any,
+    ):
+        self.statement = statement
+        self.from_clauses = statement._from_obj
+
+        for memoized_entities in statement._memoized_select_entities:
+            self._setup_joins(
+                memoized_entities._setup_joins, memoized_entities._raw_columns
+            )
+
+        if statement._setup_joins:
+            self._setup_joins(statement._setup_joins, statement._raw_columns)
+
+        self.froms = self._get_froms(statement)
+
+        self.columns_plus_names = statement._generate_columns_plus_names(True)
+
+    @classmethod
+    def _plugin_not_implemented(cls) -> NoReturn:
+        raise NotImplementedError(
+            "The default SELECT construct without plugins does not "
+            "implement this method."
+        )
+
+    @classmethod
+    def get_column_descriptions(
+        cls, statement: Select[Any]
+    ) -> List[Dict[str, Any]]:
+        return [
+            {
+                "name": name,
+                "type": element.type,
+                "expr": element,
+            }
+            for _, name, _, element, _ in (
+                statement._generate_columns_plus_names(False)
+            )
+        ]
+
+    @classmethod
+    def from_statement(
+        cls, statement: Select[Any], from_statement: roles.ReturnsRowsRole
+    ) -> ExecutableReturnsRows:
+        cls._plugin_not_implemented()
+
+    @classmethod
+    def get_columns_clause_froms(
+        cls, statement: Select[Any]
+    ) -> List[FromClause]:
+        return cls._normalize_froms(
+            itertools.chain.from_iterable(
+                element._from_objects for element in statement._raw_columns
+            )
+        )
+
+    @classmethod
+    def _column_naming_convention(
+        cls, label_style: SelectLabelStyle
+    ) -> _LabelConventionCallable:
+        table_qualified = label_style is LABEL_STYLE_TABLENAME_PLUS_COL
+
+        dedupe = label_style is not LABEL_STYLE_NONE
+
+        pa = prefix_anon_map()
+        names = set()
+
+        def go(
+            c: Union[ColumnElement[Any], TextClause],
+            col_name: Optional[str] = None,
+        ) -> Optional[str]:
+            if is_text_clause(c):
+                return None
+            elif TYPE_CHECKING:
+                assert is_column_element(c)
+
+            if not dedupe:
+                name = c._proxy_key
+                if name is None:
+                    name = "_no_label"
+                return name
+
+            name = c._tq_key_label if table_qualified else c._proxy_key
+
+            if name is None:
+                name = "_no_label"
+                if name in names:
+                    return c._anon_label(name) % pa
+                else:
+                    names.add(name)
+                    return name
+
+            elif name in names:
+                return (
+                    c._anon_tq_key_label % pa
+                    if table_qualified
+                    else c._anon_key_label % pa
+                )
+            else:
+                names.add(name)
+                return name
+
+        return go
+
+    def _get_froms(self, statement: Select[Any]) -> List[FromClause]:
+        ambiguous_table_name_map: _AmbiguousTableNameMap
+        self._ambiguous_table_name_map = ambiguous_table_name_map = {}
+
+        return self._normalize_froms(
+            itertools.chain(
+                self.from_clauses,
+                itertools.chain.from_iterable(
+                    [
+                        element._from_objects
+                        for element in statement._raw_columns
+                    ]
+                ),
+                itertools.chain.from_iterable(
+                    [
+                        element._from_objects
+                        for element in statement._where_criteria
+                    ]
+                ),
+            ),
+            check_statement=statement,
+            ambiguous_table_name_map=ambiguous_table_name_map,
+        )
+
+    @classmethod
+    def _normalize_froms(
+        cls,
+        iterable_of_froms: Iterable[FromClause],
+        check_statement: Optional[Select[Any]] = None,
+        ambiguous_table_name_map: Optional[_AmbiguousTableNameMap] = None,
+    ) -> List[FromClause]:
+        """given an iterable of things to select FROM, reduce them to what
+        would actually render in the FROM clause of a SELECT.
+
+        This does the job of checking for JOINs, tables, etc. that are in fact
+        overlapping due to cloning, adaption, present in overlapping joins,
+        etc.
+
+        """
+        seen: Set[FromClause] = set()
+        froms: List[FromClause] = []
+
+        for item in iterable_of_froms:
+            if is_subquery(item) and item.element is check_statement:
+                raise exc.InvalidRequestError(
+                    "select() construct refers to itself as a FROM"
+                )
+
+            if not seen.intersection(item._cloned_set):
+                froms.append(item)
+                seen.update(item._cloned_set)
+
+        if froms:
+            toremove = set(
+                itertools.chain.from_iterable(
+                    [_expand_cloned(f._hide_froms) for f in froms]
+                )
+            )
+            if toremove:
+                # filter out to FROM clauses not in the list,
+                # using a list to maintain ordering
+                froms = [f for f in froms if f not in toremove]
+
+            if ambiguous_table_name_map is not None:
+                ambiguous_table_name_map.update(
+                    (
+                        fr.name,
+                        _anonymous_label.safe_construct(
+                            hash(fr.name), fr.name
+                        ),
+                    )
+                    for item in froms
+                    for fr in item._from_objects
+                    if is_table(fr)
+                    and fr.schema
+                    and fr.name not in ambiguous_table_name_map
+                )
+
+        return froms
+
+    def _get_display_froms(
+        self,
+        explicit_correlate_froms: Optional[Sequence[FromClause]] = None,
+        implicit_correlate_froms: Optional[Sequence[FromClause]] = None,
+    ) -> List[FromClause]:
+        """Return the full list of 'from' clauses to be displayed.
+
+        Takes into account a set of existing froms which may be
+        rendered in the FROM clause of enclosing selects; this Select
+        may want to leave those absent if it is automatically
+        correlating.
+
+        """
+
+        froms = self.froms
+
+        if self.statement._correlate:
+            to_correlate = self.statement._correlate
+            if to_correlate:
+                froms = [
+                    f
+                    for f in froms
+                    if f
+                    not in _cloned_intersection(
+                        _cloned_intersection(
+                            froms, explicit_correlate_froms or ()
+                        ),
+                        to_correlate,
+                    )
+                ]
+
+        if self.statement._correlate_except is not None:
+            froms = [
+                f
+                for f in froms
+                if f
+                not in _cloned_difference(
+                    _cloned_intersection(
+                        froms, explicit_correlate_froms or ()
+                    ),
+                    self.statement._correlate_except,
+                )
+            ]
+
+        if (
+            self.statement._auto_correlate
+            and implicit_correlate_froms
+            and len(froms) > 1
+        ):
+            froms = [
+                f
+                for f in froms
+                if f
+                not in _cloned_intersection(froms, implicit_correlate_froms)
+            ]
+
+            if not len(froms):
+                raise exc.InvalidRequestError(
+                    "Select statement '%r"
+                    "' returned no FROM clauses "
+                    "due to auto-correlation; "
+                    "specify correlate(<tables>) "
+                    "to control correlation "
+                    "manually." % self.statement
+                )
+
+        return froms
+
+    def _memoized_attr__label_resolve_dict(
+        self,
+    ) -> Tuple[
+        Dict[str, ColumnElement[Any]],
+        Dict[str, ColumnElement[Any]],
+        Dict[str, ColumnElement[Any]],
+    ]:
+        with_cols: Dict[str, ColumnElement[Any]] = {
+            c._tq_label or c.key: c
+            for c in self.statement._all_selected_columns
+            if c._allow_label_resolve
+        }
+        only_froms: Dict[str, ColumnElement[Any]] = {
+            c.key: c  # type: ignore
+            for c in _select_iterables(self.froms)
+            if c._allow_label_resolve
+        }
+        only_cols: Dict[str, ColumnElement[Any]] = with_cols.copy()
+        for key, value in only_froms.items():
+            with_cols.setdefault(key, value)
+
+        return with_cols, only_froms, only_cols
+
+    @classmethod
+    def determine_last_joined_entity(
+        cls, stmt: Select[Any]
+    ) -> Optional[_JoinTargetElement]:
+        if stmt._setup_joins:
+            return stmt._setup_joins[-1][0]
+        else:
+            return None
+
+    @classmethod
+    def all_selected_columns(cls, statement: Select[Any]) -> _SelectIterable:
+        return [c for c in _select_iterables(statement._raw_columns)]
+
+    def _setup_joins(
+        self,
+        args: Tuple[_SetupJoinsElement, ...],
+        raw_columns: List[_ColumnsClauseElement],
+    ) -> None:
+        for right, onclause, left, flags in args:
+            if TYPE_CHECKING:
+                if onclause is not None:
+                    assert isinstance(onclause, ColumnElement)
+
+            isouter = flags["isouter"]
+            full = flags["full"]
+
+            if left is None:
+                (
+                    left,
+                    replace_from_obj_index,
+                ) = self._join_determine_implicit_left_side(
+                    raw_columns, left, right, onclause
+                )
+            else:
+                (replace_from_obj_index) = self._join_place_explicit_left_side(
+                    left
+                )
+
+            # these assertions can be made here, as if the right/onclause
+            # contained ORM elements, the select() statement would have been
+            # upgraded to an ORM select, and this method would not be called;
+            # orm.context.ORMSelectCompileState._join() would be
+            # used instead.
+            if TYPE_CHECKING:
+                assert isinstance(right, FromClause)
+                if onclause is not None:
+                    assert isinstance(onclause, ColumnElement)
+
+            if replace_from_obj_index is not None:
+                # splice into an existing element in the
+                # self._from_obj list
+                left_clause = self.from_clauses[replace_from_obj_index]
+
+                self.from_clauses = (
+                    self.from_clauses[:replace_from_obj_index]
+                    + (
+                        Join(
+                            left_clause,
+                            right,
+                            onclause,
+                            isouter=isouter,
+                            full=full,
+                        ),
+                    )
+                    + self.from_clauses[replace_from_obj_index + 1 :]
+                )
+            else:
+                assert left is not None
+                self.from_clauses = self.from_clauses + (
+                    Join(left, right, onclause, isouter=isouter, full=full),
+                )
+
+    @util.preload_module("sqlalchemy.sql.util")
+    def _join_determine_implicit_left_side(
+        self,
+        raw_columns: List[_ColumnsClauseElement],
+        left: Optional[FromClause],
+        right: _JoinTargetElement,
+        onclause: Optional[ColumnElement[Any]],
+    ) -> Tuple[Optional[FromClause], Optional[int]]:
+        """When join conditions don't express the left side explicitly,
+        determine if an existing FROM or entity in this query
+        can serve as the left hand side.
+
+        """
+
+        sql_util = util.preloaded.sql_util
+
+        replace_from_obj_index: Optional[int] = None
+
+        from_clauses = self.from_clauses
+
+        if from_clauses:
+            indexes: List[int] = sql_util.find_left_clause_to_join_from(
+                from_clauses, right, onclause
+            )
+
+            if len(indexes) == 1:
+                replace_from_obj_index = indexes[0]
+                left = from_clauses[replace_from_obj_index]
+        else:
+            potential = {}
+            statement = self.statement
+
+            for from_clause in itertools.chain(
+                itertools.chain.from_iterable(
+                    [element._from_objects for element in raw_columns]
+                ),
+                itertools.chain.from_iterable(
+                    [
+                        element._from_objects
+                        for element in statement._where_criteria
+                    ]
+                ),
+            ):
+                potential[from_clause] = ()
+
+            all_clauses = list(potential.keys())
+            indexes = sql_util.find_left_clause_to_join_from(
+                all_clauses, right, onclause
+            )
+
+            if len(indexes) == 1:
+                left = all_clauses[indexes[0]]
+
+        if len(indexes) > 1:
+            raise exc.InvalidRequestError(
+                "Can't determine which FROM clause to join "
+                "from, there are multiple FROMS which can "
+                "join to this entity. Please use the .select_from() "
+                "method to establish an explicit left side, as well as "
+                "providing an explicit ON clause if not present already to "
+                "help resolve the ambiguity."
+            )
+        elif not indexes:
+            raise exc.InvalidRequestError(
+                "Don't know how to join to %r. "
+                "Please use the .select_from() "
+                "method to establish an explicit left side, as well as "
+                "providing an explicit ON clause if not present already to "
+                "help resolve the ambiguity." % (right,)
+            )
+        return left, replace_from_obj_index
+
+    @util.preload_module("sqlalchemy.sql.util")
+    def _join_place_explicit_left_side(
+        self, left: FromClause
+    ) -> Optional[int]:
+        replace_from_obj_index: Optional[int] = None
+
+        sql_util = util.preloaded.sql_util
+
+        from_clauses = list(self.statement._iterate_from_elements())
+
+        if from_clauses:
+            indexes: List[int] = sql_util.find_left_clause_that_matches_given(
+                self.from_clauses, left
+            )
+        else:
+            indexes = []
+
+        if len(indexes) > 1:
+            raise exc.InvalidRequestError(
+                "Can't identify which entity in which to assign the "
+                "left side of this join.   Please use a more specific "
+                "ON clause."
+            )
+
+        # have an index, means the left side is already present in
+        # an existing FROM in the self._from_obj tuple
+        if indexes:
+            replace_from_obj_index = indexes[0]
+
+        # no index, means we need to add a new element to the
+        # self._from_obj tuple
+
+        return replace_from_obj_index
+
+
+class _SelectFromElements:
+    __slots__ = ()
+
+    _raw_columns: List[_ColumnsClauseElement]
+    _where_criteria: Tuple[ColumnElement[Any], ...]
+    _from_obj: Tuple[FromClause, ...]
+
+    def _iterate_from_elements(self) -> Iterator[FromClause]:
+        # note this does not include elements
+        # in _setup_joins
+
+        seen = set()
+        for element in self._raw_columns:
+            for fr in element._from_objects:
+                if fr in seen:
+                    continue
+                seen.add(fr)
+                yield fr
+        for element in self._where_criteria:
+            for fr in element._from_objects:
+                if fr in seen:
+                    continue
+                seen.add(fr)
+                yield fr
+        for element in self._from_obj:
+            if element in seen:
+                continue
+            seen.add(element)
+            yield element
+
+
+class _MemoizedSelectEntities(
+    cache_key.HasCacheKey, traversals.HasCopyInternals, visitors.Traversible
+):
+    """represents partial state from a Select object, for the case
+    where Select.columns() has redefined the set of columns/entities the
+    statement will be SELECTing from.  This object represents
+    the entities from the SELECT before that transformation was applied,
+    so that transformations that were made in terms of the SELECT at that
+    time, such as join() as well as options(), can access the correct context.
+
+    In previous SQLAlchemy versions, this wasn't needed because these
+    constructs calculated everything up front, like when you called join()
+    or options(), it did everything to figure out how that would translate
+    into specific SQL constructs that would be ready to send directly to the
+    SQL compiler when needed.  But as of
+    1.4, all of that stuff is done in the compilation phase, during the
+    "compile state" portion of the process, so that the work can all be
+    cached.  So it needs to be able to resolve joins/options2 based on what
+    the list of entities was when those methods were called.
+
+
+    """
+
+    __visit_name__ = "memoized_select_entities"
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("_raw_columns", InternalTraversal.dp_clauseelement_list),
+        ("_setup_joins", InternalTraversal.dp_setup_join_tuple),
+        ("_with_options", InternalTraversal.dp_executable_options),
+    ]
+
+    _is_clone_of: Optional[ClauseElement]
+    _raw_columns: List[_ColumnsClauseElement]
+    _setup_joins: Tuple[_SetupJoinsElement, ...]
+    _with_options: Tuple[ExecutableOption, ...]
+
+    _annotations = util.EMPTY_DICT
+
+    def _clone(self, **kw: Any) -> Self:
+        c = self.__class__.__new__(self.__class__)
+        c.__dict__ = {k: v for k, v in self.__dict__.items()}
+
+        c._is_clone_of = self.__dict__.get("_is_clone_of", self)
+        return c
+
+    @classmethod
+    def _generate_for_statement(cls, select_stmt: Select[Any]) -> None:
+        if select_stmt._setup_joins or select_stmt._with_options:
+            self = _MemoizedSelectEntities()
+            self._raw_columns = select_stmt._raw_columns
+            self._setup_joins = select_stmt._setup_joins
+            self._with_options = select_stmt._with_options
+
+            select_stmt._memoized_select_entities += (self,)
+            select_stmt._raw_columns = []
+            select_stmt._setup_joins = select_stmt._with_options = ()
+
+
+class Select(
+    HasPrefixes,
+    HasSuffixes,
+    HasHints,
+    HasCompileState,
+    _SelectFromElements,
+    GenerativeSelect,
+    TypedReturnsRows[_TP],
+):
+    """Represents a ``SELECT`` statement.
+
+    The :class:`_sql.Select` object is normally constructed using the
+    :func:`_sql.select` function.  See that function for details.
+
+    .. seealso::
+
+        :func:`_sql.select`
+
+        :ref:`tutorial_selecting_data` - in the 2.0 tutorial
+
+    """
+
+    __visit_name__ = "select"
+
+    _setup_joins: Tuple[_SetupJoinsElement, ...] = ()
+    _memoized_select_entities: Tuple[TODO_Any, ...] = ()
+
+    _raw_columns: List[_ColumnsClauseElement]
+
+    _distinct: bool = False
+    _distinct_on: Tuple[ColumnElement[Any], ...] = ()
+    _correlate: Tuple[FromClause, ...] = ()
+    _correlate_except: Optional[Tuple[FromClause, ...]] = None
+    _where_criteria: Tuple[ColumnElement[Any], ...] = ()
+    _having_criteria: Tuple[ColumnElement[Any], ...] = ()
+    _from_obj: Tuple[FromClause, ...] = ()
+    _auto_correlate = True
+    _is_select_statement = True
+    _compile_options: CacheableOptions = (
+        SelectState.default_select_compile_options
+    )
+
+    _traverse_internals: _TraverseInternalsType = (
+        [
+            ("_raw_columns", InternalTraversal.dp_clauseelement_list),
+            (
+                "_memoized_select_entities",
+                InternalTraversal.dp_memoized_select_entities,
+            ),
+            ("_from_obj", InternalTraversal.dp_clauseelement_list),
+            ("_where_criteria", InternalTraversal.dp_clauseelement_tuple),
+            ("_having_criteria", InternalTraversal.dp_clauseelement_tuple),
+            ("_order_by_clauses", InternalTraversal.dp_clauseelement_tuple),
+            ("_group_by_clauses", InternalTraversal.dp_clauseelement_tuple),
+            ("_setup_joins", InternalTraversal.dp_setup_join_tuple),
+            ("_correlate", InternalTraversal.dp_clauseelement_tuple),
+            ("_correlate_except", InternalTraversal.dp_clauseelement_tuple),
+            ("_limit_clause", InternalTraversal.dp_clauseelement),
+            ("_offset_clause", InternalTraversal.dp_clauseelement),
+            ("_fetch_clause", InternalTraversal.dp_clauseelement),
+            ("_fetch_clause_options", InternalTraversal.dp_plain_dict),
+            ("_for_update_arg", InternalTraversal.dp_clauseelement),
+            ("_distinct", InternalTraversal.dp_boolean),
+            ("_distinct_on", InternalTraversal.dp_clauseelement_tuple),
+            ("_label_style", InternalTraversal.dp_plain_obj),
+        ]
+        + HasCTE._has_ctes_traverse_internals
+        + HasPrefixes._has_prefixes_traverse_internals
+        + HasSuffixes._has_suffixes_traverse_internals
+        + HasHints._has_hints_traverse_internals
+        + SupportsCloneAnnotations._clone_annotations_traverse_internals
+        + Executable._executable_traverse_internals
+    )
+
+    _cache_key_traversal: _CacheKeyTraversalType = _traverse_internals + [
+        ("_compile_options", InternalTraversal.dp_has_cache_key)
+    ]
+
+    _compile_state_factory: Type[SelectState]
+
+    @classmethod
+    def _create_raw_select(cls, **kw: Any) -> Select[Any]:
+        """Create a :class:`.Select` using raw ``__new__`` with no coercions.
+
+        Used internally to build up :class:`.Select` constructs with
+        pre-established state.
+
+        """
+
+        stmt = Select.__new__(Select)
+        stmt.__dict__.update(kw)
+        return stmt
+
+    def __init__(self, *entities: _ColumnsClauseArgument[Any]):
+        r"""Construct a new :class:`_expression.Select`.
+
+        The public constructor for :class:`_expression.Select` is the
+        :func:`_sql.select` function.
+
+        """
+        self._raw_columns = [
+            coercions.expect(
+                roles.ColumnsClauseRole, ent, apply_propagate_attrs=self
+            )
+            for ent in entities
+        ]
+
+        GenerativeSelect.__init__(self)
+
+    def _scalar_type(self) -> TypeEngine[Any]:
+        if not self._raw_columns:
+            return NULLTYPE
+        elem = self._raw_columns[0]
+        cols = list(elem._select_iterable)
+        return cols[0].type
+
+    def filter(self, *criteria: _ColumnExpressionArgument[bool]) -> Self:
+        """A synonym for the :meth:`_sql.Select.where` method."""
+
+        return self.where(*criteria)
+
+    def _filter_by_zero(
+        self,
+    ) -> Union[
+        FromClause, _JoinTargetProtocol, ColumnElement[Any], TextClause
+    ]:
+        if self._setup_joins:
+            meth = SelectState.get_plugin_class(
+                self
+            ).determine_last_joined_entity
+            _last_joined_entity = meth(self)
+            if _last_joined_entity is not None:
+                return _last_joined_entity
+
+        if self._from_obj:
+            return self._from_obj[0]
+
+        return self._raw_columns[0]
+
+    if TYPE_CHECKING:
+
+        @overload
+        def scalar_subquery(
+            self: Select[Tuple[_MAYBE_ENTITY]],
+        ) -> ScalarSelect[Any]: ...
+
+        @overload
+        def scalar_subquery(
+            self: Select[Tuple[_NOT_ENTITY]],
+        ) -> ScalarSelect[_NOT_ENTITY]: ...
+
+        @overload
+        def scalar_subquery(self) -> ScalarSelect[Any]: ...
+
+        def scalar_subquery(self) -> ScalarSelect[Any]: ...
+
+    def filter_by(self, **kwargs: Any) -> Self:
+        r"""apply the given filtering criterion as a WHERE clause
+        to this select.
+
+        """
+        from_entity = self._filter_by_zero()
+
+        clauses = [
+            _entity_namespace_key(from_entity, key) == value
+            for key, value in kwargs.items()
+        ]
+        return self.filter(*clauses)
+
+    @property
+    def column_descriptions(self) -> Any:
+        """Return a :term:`plugin-enabled` 'column descriptions' structure
+        referring to the columns which are SELECTed by this statement.
+
+        This attribute is generally useful when using the ORM, as an
+        extended structure which includes information about mapped
+        entities is returned.  The section :ref:`queryguide_inspection`
+        contains more background.
+
+        For a Core-only statement, the structure returned by this accessor
+        is derived from the same objects that are returned by the
+        :attr:`.Select.selected_columns` accessor, formatted as a list of
+        dictionaries which contain the keys ``name``, ``type`` and ``expr``,
+        which indicate the column expressions to be selected::
+
+            >>> stmt = select(user_table)
+            >>> stmt.column_descriptions
+            [
+                {
+                    'name': 'id',
+                    'type': Integer(),
+                    'expr': Column('id', Integer(), ...)},
+                {
+                    'name': 'name',
+                    'type': String(length=30),
+                    'expr': Column('name', String(length=30), ...)}
+            ]
+
+        .. versionchanged:: 1.4.33 The :attr:`.Select.column_descriptions`
+           attribute returns a structure for a Core-only set of entities,
+           not just ORM-only entities.
+
+        .. seealso::
+
+            :attr:`.UpdateBase.entity_description` - entity information for
+            an :func:`.insert`, :func:`.update`, or :func:`.delete`
+
+            :ref:`queryguide_inspection` - ORM background
+
+        """
+        meth = SelectState.get_plugin_class(self).get_column_descriptions
+        return meth(self)
+
+    def from_statement(
+        self, statement: roles.ReturnsRowsRole
+    ) -> ExecutableReturnsRows:
+        """Apply the columns which this :class:`.Select` would select
+        onto another statement.
+
+        This operation is :term:`plugin-specific` and will raise a not
+        supported exception if this :class:`_sql.Select` does not select from
+        plugin-enabled entities.
+
+
+        The statement is typically either a :func:`_expression.text` or
+        :func:`_expression.select` construct, and should return the set of
+        columns appropriate to the entities represented by this
+        :class:`.Select`.
+
+        .. seealso::
+
+            :ref:`orm_queryguide_selecting_text` - usage examples in the
+            ORM Querying Guide
+
+        """
+        meth = SelectState.get_plugin_class(self).from_statement
+        return meth(self, statement)
+
+    @_generative
+    def join(
+        self,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        isouter: bool = False,
+        full: bool = False,
+    ) -> Self:
+        r"""Create a SQL JOIN against this :class:`_expression.Select`
+        object's criterion
+        and apply generatively, returning the newly resulting
+        :class:`_expression.Select`.
+
+        E.g.::
+
+            stmt = select(user_table).join(
+                address_table, user_table.c.id == address_table.c.user_id
+            )
+
+        The above statement generates SQL similar to:
+
+        .. sourcecode:: sql
+
+            SELECT user.id, user.name
+            FROM user
+            JOIN address ON user.id = address.user_id
+
+        .. versionchanged:: 1.4 :meth:`_expression.Select.join` now creates
+           a :class:`_sql.Join` object between a :class:`_sql.FromClause`
+           source that is within the FROM clause of the existing SELECT,
+           and a given target :class:`_sql.FromClause`, and then adds
+           this :class:`_sql.Join` to the FROM clause of the newly generated
+           SELECT statement.    This is completely reworked from the behavior
+           in 1.3, which would instead create a subquery of the entire
+           :class:`_expression.Select` and then join that subquery to the
+           target.
+
+           This is a **backwards incompatible change** as the previous behavior
+           was mostly useless, producing an unnamed subquery rejected by
+           most databases in any case.   The new behavior is modeled after
+           that of the very successful :meth:`_orm.Query.join` method in the
+           ORM, in order to support the functionality of :class:`_orm.Query`
+           being available by using a :class:`_sql.Select` object with an
+           :class:`_orm.Session`.
+
+           See the notes for this change at :ref:`change_select_join`.
+
+
+        :param target: target table to join towards
+
+        :param onclause: ON clause of the join.  If omitted, an ON clause
+         is generated automatically based on the :class:`_schema.ForeignKey`
+         linkages between the two tables, if one can be unambiguously
+         determined, otherwise an error is raised.
+
+        :param isouter: if True, generate LEFT OUTER join.  Same as
+         :meth:`_expression.Select.outerjoin`.
+
+        :param full: if True, generate FULL OUTER join.
+
+        .. seealso::
+
+            :ref:`tutorial_select_join` - in the :doc:`/tutorial/index`
+
+            :ref:`orm_queryguide_joins` - in the :ref:`queryguide_toplevel`
+
+            :meth:`_expression.Select.join_from`
+
+            :meth:`_expression.Select.outerjoin`
+
+        """  # noqa: E501
+        join_target = coercions.expect(
+            roles.JoinTargetRole, target, apply_propagate_attrs=self
+        )
+        if onclause is not None:
+            onclause_element = coercions.expect(roles.OnClauseRole, onclause)
+        else:
+            onclause_element = None
+
+        self._setup_joins += (
+            (
+                join_target,
+                onclause_element,
+                None,
+                {"isouter": isouter, "full": full},
+            ),
+        )
+        return self
+
+    def outerjoin_from(
+        self,
+        from_: _FromClauseArgument,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        full: bool = False,
+    ) -> Self:
+        r"""Create a SQL LEFT OUTER JOIN against this
+        :class:`_expression.Select` object's criterion and apply generatively,
+        returning the newly resulting :class:`_expression.Select`.
+
+        Usage is the same as that of :meth:`_selectable.Select.join_from`.
+
+        """
+        return self.join_from(
+            from_, target, onclause=onclause, isouter=True, full=full
+        )
+
+    @_generative
+    def join_from(
+        self,
+        from_: _FromClauseArgument,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        isouter: bool = False,
+        full: bool = False,
+    ) -> Self:
+        r"""Create a SQL JOIN against this :class:`_expression.Select`
+        object's criterion
+        and apply generatively, returning the newly resulting
+        :class:`_expression.Select`.
+
+        E.g.::
+
+            stmt = select(user_table, address_table).join_from(
+                user_table, address_table, user_table.c.id == address_table.c.user_id
+            )
+
+        The above statement generates SQL similar to:
+
+        .. sourcecode:: sql
+
+            SELECT user.id, user.name, address.id, address.email, address.user_id
+            FROM user JOIN address ON user.id = address.user_id
+
+        .. versionadded:: 1.4
+
+        :param from\_: the left side of the join, will be rendered in the
+         FROM clause and is roughly equivalent to using the
+         :meth:`.Select.select_from` method.
+
+        :param target: target table to join towards
+
+        :param onclause: ON clause of the join.
+
+        :param isouter: if True, generate LEFT OUTER join.  Same as
+         :meth:`_expression.Select.outerjoin`.
+
+        :param full: if True, generate FULL OUTER join.
+
+        .. seealso::
+
+            :ref:`tutorial_select_join` - in the :doc:`/tutorial/index`
+
+            :ref:`orm_queryguide_joins` - in the :ref:`queryguide_toplevel`
+
+            :meth:`_expression.Select.join`
+
+        """  # noqa: E501
+
+        # note the order of parsing from vs. target is important here, as we
+        # are also deriving the source of the plugin (i.e. the subject mapper
+        # in an ORM query) which should favor the "from_" over the "target"
+
+        from_ = coercions.expect(
+            roles.FromClauseRole, from_, apply_propagate_attrs=self
+        )
+        join_target = coercions.expect(
+            roles.JoinTargetRole, target, apply_propagate_attrs=self
+        )
+        if onclause is not None:
+            onclause_element = coercions.expect(roles.OnClauseRole, onclause)
+        else:
+            onclause_element = None
+
+        self._setup_joins += (
+            (
+                join_target,
+                onclause_element,
+                from_,
+                {"isouter": isouter, "full": full},
+            ),
+        )
+        return self
+
+    def outerjoin(
+        self,
+        target: _JoinTargetArgument,
+        onclause: Optional[_OnClauseArgument] = None,
+        *,
+        full: bool = False,
+    ) -> Self:
+        """Create a left outer join.
+
+        Parameters are the same as that of :meth:`_expression.Select.join`.
+
+        .. versionchanged:: 1.4 :meth:`_expression.Select.outerjoin` now
+           creates a :class:`_sql.Join` object between a
+           :class:`_sql.FromClause` source that is within the FROM clause of
+           the existing SELECT, and a given target :class:`_sql.FromClause`,
+           and then adds this :class:`_sql.Join` to the FROM clause of the
+           newly generated SELECT statement.    This is completely reworked
+           from the behavior in 1.3, which would instead create a subquery of
+           the entire
+           :class:`_expression.Select` and then join that subquery to the
+           target.
+
+           This is a **backwards incompatible change** as the previous behavior
+           was mostly useless, producing an unnamed subquery rejected by
+           most databases in any case.   The new behavior is modeled after
+           that of the very successful :meth:`_orm.Query.join` method in the
+           ORM, in order to support the functionality of :class:`_orm.Query`
+           being available by using a :class:`_sql.Select` object with an
+           :class:`_orm.Session`.
+
+           See the notes for this change at :ref:`change_select_join`.
+
+        .. seealso::
+
+            :ref:`tutorial_select_join` - in the :doc:`/tutorial/index`
+
+            :ref:`orm_queryguide_joins` - in the :ref:`queryguide_toplevel`
+
+            :meth:`_expression.Select.join`
+
+        """
+        return self.join(target, onclause=onclause, isouter=True, full=full)
+
+    def get_final_froms(self) -> Sequence[FromClause]:
+        """Compute the final displayed list of :class:`_expression.FromClause`
+        elements.
+
+        This method will run through the full computation required to
+        determine what FROM elements will be displayed in the resulting
+        SELECT statement, including shadowing individual tables with
+        JOIN objects, as well as full computation for ORM use cases including
+        eager loading clauses.
+
+        For ORM use, this accessor returns the **post compilation**
+        list of FROM objects; this collection will include elements such as
+        eagerly loaded tables and joins.  The objects will **not** be
+        ORM enabled and not work as a replacement for the
+        :meth:`_sql.Select.select_froms` collection; additionally, the
+        method is not well performing for an ORM enabled statement as it
+        will incur the full ORM construction process.
+
+        To retrieve the FROM list that's implied by the "columns" collection
+        passed to the :class:`_sql.Select` originally, use the
+        :attr:`_sql.Select.columns_clause_froms` accessor.
+
+        To select from an alternative set of columns while maintaining the
+        FROM list, use the :meth:`_sql.Select.with_only_columns` method and
+        pass the
+        :paramref:`_sql.Select.with_only_columns.maintain_column_froms`
+        parameter.
+
+        .. versionadded:: 1.4.23 - the :meth:`_sql.Select.get_final_froms`
+           method replaces the previous :attr:`_sql.Select.froms` accessor,
+           which is deprecated.
+
+        .. seealso::
+
+            :attr:`_sql.Select.columns_clause_froms`
+
+        """
+
+        return self._compile_state_factory(self, None)._get_display_froms()
+
+    @property
+    @util.deprecated(
+        "1.4.23",
+        "The :attr:`_expression.Select.froms` attribute is moved to "
+        "the :meth:`_expression.Select.get_final_froms` method.",
+    )
+    def froms(self) -> Sequence[FromClause]:
+        """Return the displayed list of :class:`_expression.FromClause`
+        elements.
+
+
+        """
+        return self.get_final_froms()
+
+    @property
+    def columns_clause_froms(self) -> List[FromClause]:
+        """Return the set of :class:`_expression.FromClause` objects implied
+        by the columns clause of this SELECT statement.
+
+        .. versionadded:: 1.4.23
+
+        .. seealso::
+
+            :attr:`_sql.Select.froms` - "final" FROM list taking the full
+            statement into account
+
+            :meth:`_sql.Select.with_only_columns` - makes use of this
+            collection to set up a new FROM list
+
+        """
+
+        return SelectState.get_plugin_class(self).get_columns_clause_froms(
+            self
+        )
+
+    @property
+    def inner_columns(self) -> _SelectIterable:
+        """An iterator of all :class:`_expression.ColumnElement`
+        expressions which would
+        be rendered into the columns clause of the resulting SELECT statement.
+
+        This method is legacy as of 1.4 and is superseded by the
+        :attr:`_expression.Select.exported_columns` collection.
+
+        """
+
+        return iter(self._all_selected_columns)
+
+    def is_derived_from(self, fromclause: Optional[FromClause]) -> bool:
+        if fromclause is not None and self in fromclause._cloned_set:
+            return True
+
+        for f in self._iterate_from_elements():
+            if f.is_derived_from(fromclause):
+                return True
+        return False
+
+    def _copy_internals(
+        self, clone: _CloneCallableType = _clone, **kw: Any
+    ) -> None:
+        # Select() object has been cloned and probably adapted by the
+        # given clone function.  Apply the cloning function to internal
+        # objects
+
+        # 1. keep a dictionary of the froms we've cloned, and what
+        # they've become.  This allows us to ensure the same cloned from
+        # is used when other items such as columns are "cloned"
+
+        all_the_froms = set(
+            itertools.chain(
+                _from_objects(*self._raw_columns),
+                _from_objects(*self._where_criteria),
+                _from_objects(*[elem[0] for elem in self._setup_joins]),
+            )
+        )
+
+        # do a clone for the froms we've gathered.  what is important here
+        # is if any of the things we are selecting from, like tables,
+        # were converted into Join objects.   if so, these need to be
+        # added to _from_obj explicitly, because otherwise they won't be
+        # part of the new state, as they don't associate themselves with
+        # their columns.
+        new_froms = {f: clone(f, **kw) for f in all_the_froms}
+
+        # 2. copy FROM collections, adding in joins that we've created.
+        existing_from_obj = [clone(f, **kw) for f in self._from_obj]
+        add_froms = (
+            {f for f in new_froms.values() if isinstance(f, Join)}
+            .difference(all_the_froms)
+            .difference(existing_from_obj)
+        )
+
+        self._from_obj = tuple(existing_from_obj) + tuple(add_froms)
+
+        # 3. clone everything else, making sure we use columns
+        # corresponding to the froms we just made.
+        def replace(
+            obj: Union[BinaryExpression[Any], ColumnClause[Any]],
+            **kw: Any,
+        ) -> Optional[KeyedColumnElement[Any]]:
+            if isinstance(obj, ColumnClause) and obj.table in new_froms:
+                newelem = new_froms[obj.table].corresponding_column(obj)
+                return newelem
+            return None
+
+        kw["replace"] = replace
+
+        # copy everything else.   for table-ish things like correlate,
+        # correlate_except, setup_joins, these clone normally.  For
+        # column-expression oriented things like raw_columns, where_criteria,
+        # order by, we get this from the new froms.
+        super()._copy_internals(clone=clone, omit_attrs=("_from_obj",), **kw)
+
+        self._reset_memoizations()
+
+    def get_children(self, **kw: Any) -> Iterable[ClauseElement]:
+        return itertools.chain(
+            super().get_children(
+                omit_attrs=("_from_obj", "_correlate", "_correlate_except"),
+                **kw,
+            ),
+            self._iterate_from_elements(),
+        )
+
+    @_generative
+    def add_columns(
+        self, *entities: _ColumnsClauseArgument[Any]
+    ) -> Select[Any]:
+        r"""Return a new :func:`_expression.select` construct with
+        the given entities appended to its columns clause.
+
+        E.g.::
+
+            my_select = my_select.add_columns(table.c.new_column)
+
+        The original expressions in the columns clause remain in place.
+        To replace the original expressions with new ones, see the method
+        :meth:`_expression.Select.with_only_columns`.
+
+        :param \*entities: column, table, or other entity expressions to be
+         added to the columns clause
+
+        .. seealso::
+
+            :meth:`_expression.Select.with_only_columns` - replaces existing
+            expressions rather than appending.
+
+            :ref:`orm_queryguide_select_multiple_entities` - ORM-centric
+            example
+
+        """
+        self._reset_memoizations()
+
+        self._raw_columns = self._raw_columns + [
+            coercions.expect(
+                roles.ColumnsClauseRole, column, apply_propagate_attrs=self
+            )
+            for column in entities
+        ]
+        return self
+
+    def _set_entities(
+        self, entities: Iterable[_ColumnsClauseArgument[Any]]
+    ) -> None:
+        self._raw_columns = [
+            coercions.expect(
+                roles.ColumnsClauseRole, ent, apply_propagate_attrs=self
+            )
+            for ent in util.to_list(entities)
+        ]
+
+    @util.deprecated(
+        "1.4",
+        "The :meth:`_expression.Select.column` method is deprecated and will "
+        "be removed in a future release.  Please use "
+        ":meth:`_expression.Select.add_columns`",
+    )
+    def column(self, column: _ColumnsClauseArgument[Any]) -> Select[Any]:
+        """Return a new :func:`_expression.select` construct with
+        the given column expression added to its columns clause.
+
+        E.g.::
+
+            my_select = my_select.column(table.c.new_column)
+
+        See the documentation for
+        :meth:`_expression.Select.with_only_columns`
+        for guidelines on adding /replacing the columns of a
+        :class:`_expression.Select` object.
+
+        """
+        return self.add_columns(column)
+
+    @util.preload_module("sqlalchemy.sql.util")
+    def reduce_columns(self, only_synonyms: bool = True) -> Select[Any]:
+        """Return a new :func:`_expression.select` construct with redundantly
+        named, equivalently-valued columns removed from the columns clause.
+
+        "Redundant" here means two columns where one refers to the
+        other either based on foreign key, or via a simple equality
+        comparison in the WHERE clause of the statement.   The primary purpose
+        of this method is to automatically construct a select statement
+        with all uniquely-named columns, without the need to use
+        table-qualified labels as
+        :meth:`_expression.Select.set_label_style`
+        does.
+
+        When columns are omitted based on foreign key, the referred-to
+        column is the one that's kept.  When columns are omitted based on
+        WHERE equivalence, the first column in the columns clause is the
+        one that's kept.
+
+        :param only_synonyms: when True, limit the removal of columns
+         to those which have the same name as the equivalent.   Otherwise,
+         all columns that are equivalent to another are removed.
+
+        """
+        woc: Select[Any]
+        woc = self.with_only_columns(
+            *util.preloaded.sql_util.reduce_columns(
+                self._all_selected_columns,
+                only_synonyms=only_synonyms,
+                *(self._where_criteria + self._from_obj),
+            )
+        )
+        return woc
+
+    # START OVERLOADED FUNCTIONS self.with_only_columns Select 1-8 ", *, maintain_column_froms: bool =..." # noqa: E501
+
+    # code within this block is **programmatically,
+    # statically generated** by tools/generate_tuple_map_overloads.py
+
+    @overload
+    def with_only_columns(
+        self, __ent0: _TCCA[_T0], *, maintain_column_froms: bool = ...
+    ) -> Select[Tuple[_T0]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2, _T3]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6]]: ...
+
+    @overload
+    def with_only_columns(
+        self,
+        __ent0: _TCCA[_T0],
+        __ent1: _TCCA[_T1],
+        __ent2: _TCCA[_T2],
+        __ent3: _TCCA[_T3],
+        __ent4: _TCCA[_T4],
+        __ent5: _TCCA[_T5],
+        __ent6: _TCCA[_T6],
+        __ent7: _TCCA[_T7],
+        *,
+        maintain_column_froms: bool = ...,
+    ) -> Select[Tuple[_T0, _T1, _T2, _T3, _T4, _T5, _T6, _T7]]: ...
+
+    # END OVERLOADED FUNCTIONS self.with_only_columns
+
+    @overload
+    def with_only_columns(
+        self,
+        *entities: _ColumnsClauseArgument[Any],
+        maintain_column_froms: bool = False,
+        **__kw: Any,
+    ) -> Select[Any]: ...
+
+    @_generative
+    def with_only_columns(
+        self,
+        *entities: _ColumnsClauseArgument[Any],
+        maintain_column_froms: bool = False,
+        **__kw: Any,
+    ) -> Select[Any]:
+        r"""Return a new :func:`_expression.select` construct with its columns
+        clause replaced with the given entities.
+
+        By default, this method is exactly equivalent to as if the original
+        :func:`_expression.select` had been called with the given entities.
+        E.g. a statement::
+
+            s = select(table1.c.a, table1.c.b)
+            s = s.with_only_columns(table1.c.b)
+
+        should be exactly equivalent to::
+
+            s = select(table1.c.b)
+
+        In this mode of operation, :meth:`_sql.Select.with_only_columns`
+        will also dynamically alter the FROM clause of the
+        statement if it is not explicitly stated.
+        To maintain the existing set of FROMs including those implied by the
+        current columns clause, add the
+        :paramref:`_sql.Select.with_only_columns.maintain_column_froms`
+        parameter::
+
+            s = select(table1.c.a, table2.c.b)
+            s = s.with_only_columns(table1.c.a, maintain_column_froms=True)
+
+        The above parameter performs a transfer of the effective FROMs
+        in the columns collection to the :meth:`_sql.Select.select_from`
+        method, as though the following were invoked::
+
+            s = select(table1.c.a, table2.c.b)
+            s = s.select_from(table1, table2).with_only_columns(table1.c.a)
+
+        The :paramref:`_sql.Select.with_only_columns.maintain_column_froms`
+        parameter makes use of the :attr:`_sql.Select.columns_clause_froms`
+        collection and performs an operation equivalent to the following::
+
+            s = select(table1.c.a, table2.c.b)
+            s = s.select_from(*s.columns_clause_froms).with_only_columns(table1.c.a)
+
+        :param \*entities: column expressions to be used.
+
+        :param maintain_column_froms: boolean parameter that will ensure the
+         FROM list implied from the current columns clause will be transferred
+         to the :meth:`_sql.Select.select_from` method first.
+
+         .. versionadded:: 1.4.23
+
+        """  # noqa: E501
+
+        if __kw:
+            raise _no_kw()
+
+        # memoizations should be cleared here as of
+        # I95c560ffcbfa30b26644999412fb6a385125f663 , asserting this
+        # is the case for now.
+        self._assert_no_memoizations()
+
+        if maintain_column_froms:
+            self.select_from.non_generative(  # type: ignore
+                self, *self.columns_clause_froms
+            )
+
+        # then memoize the FROMs etc.
+        _MemoizedSelectEntities._generate_for_statement(self)
+
+        self._raw_columns = [
+            coercions.expect(roles.ColumnsClauseRole, c)
+            for c in coercions._expression_collection_was_a_list(
+                "entities", "Select.with_only_columns", entities
+            )
+        ]
+        return self
+
+    @property
+    def whereclause(self) -> Optional[ColumnElement[Any]]:
+        """Return the completed WHERE clause for this
+        :class:`_expression.Select` statement.
+
+        This assembles the current collection of WHERE criteria
+        into a single :class:`_expression.BooleanClauseList` construct.
+
+
+        .. versionadded:: 1.4
+
+        """
+
+        return BooleanClauseList._construct_for_whereclause(
+            self._where_criteria
+        )
+
+    _whereclause = whereclause
+
+    @_generative
+    def where(self, *whereclause: _ColumnExpressionArgument[bool]) -> Self:
+        """Return a new :func:`_expression.select` construct with
+        the given expression added to
+        its WHERE clause, joined to the existing clause via AND, if any.
+
+        """
+
+        assert isinstance(self._where_criteria, tuple)
+
+        for criterion in whereclause:
+            where_criteria: ColumnElement[Any] = coercions.expect(
+                roles.WhereHavingRole, criterion, apply_propagate_attrs=self
+            )
+            self._where_criteria += (where_criteria,)
+        return self
+
+    @_generative
+    def having(self, *having: _ColumnExpressionArgument[bool]) -> Self:
+        """Return a new :func:`_expression.select` construct with
+        the given expression added to
+        its HAVING clause, joined to the existing clause via AND, if any.
+
+        """
+
+        for criterion in having:
+            having_criteria = coercions.expect(
+                roles.WhereHavingRole, criterion, apply_propagate_attrs=self
+            )
+            self._having_criteria += (having_criteria,)
+        return self
+
+    @_generative
+    def distinct(self, *expr: _ColumnExpressionArgument[Any]) -> Self:
+        r"""Return a new :func:`_expression.select` construct which
+        will apply DISTINCT to the SELECT statement overall.
+
+        E.g.::
+
+            from sqlalchemy import select
+
+            stmt = select(users_table.c.id, users_table.c.name).distinct()
+
+        The above would produce an statement resembling:
+
+        .. sourcecode:: sql
+
+            SELECT DISTINCT user.id, user.name FROM user
+
+        The method also accepts an ``*expr`` parameter which produces the
+        PostgreSQL dialect-specific ``DISTINCT ON`` expression.  Using this
+        parameter on other backends which don't support this syntax will
+        raise an error.
+
+        :param \*expr: optional column expressions.  When present,
+         the PostgreSQL dialect will render a ``DISTINCT ON (<expressions>)``
+         construct.  A deprecation warning and/or :class:`_exc.CompileError`
+         will be raised on other backends.
+
+         .. deprecated:: 1.4 Using \*expr in other dialects is deprecated
+            and will raise :class:`_exc.CompileError` in a future version.
+
+        """
+        if expr:
+            self._distinct = True
+            self._distinct_on = self._distinct_on + tuple(
+                coercions.expect(roles.ByOfRole, e, apply_propagate_attrs=self)
+                for e in expr
+            )
+        else:
+            self._distinct = True
+        return self
+
+    @_generative
+    def select_from(self, *froms: _FromClauseArgument) -> Self:
+        r"""Return a new :func:`_expression.select` construct with the
+        given FROM expression(s)
+        merged into its list of FROM objects.
+
+        E.g.::
+
+            table1 = table("t1", column("a"))
+            table2 = table("t2", column("b"))
+            s = select(table1.c.a).select_from(
+                table1.join(table2, table1.c.a == table2.c.b)
+            )
+
+        The "from" list is a unique set on the identity of each element,
+        so adding an already present :class:`_schema.Table`
+        or other selectable
+        will have no effect.   Passing a :class:`_expression.Join` that refers
+        to an already present :class:`_schema.Table`
+        or other selectable will have
+        the effect of concealing the presence of that selectable as
+        an individual element in the rendered FROM list, instead
+        rendering it into a JOIN clause.
+
+        While the typical purpose of :meth:`_expression.Select.select_from`
+        is to
+        replace the default, derived FROM clause with a join, it can
+        also be called with individual table elements, multiple times
+        if desired, in the case that the FROM clause cannot be fully
+        derived from the columns clause::
+
+            select(func.count("*")).select_from(table1)
+
+        """
+
+        self._from_obj += tuple(
+            coercions.expect(
+                roles.FromClauseRole, fromclause, apply_propagate_attrs=self
+            )
+            for fromclause in froms
+        )
+        return self
+
+    @_generative
+    def correlate(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        r"""Return a new :class:`_expression.Select`
+        which will correlate the given FROM
+        clauses to that of an enclosing :class:`_expression.Select`.
+
+        Calling this method turns off the :class:`_expression.Select` object's
+        default behavior of "auto-correlation".  Normally, FROM elements
+        which appear in a :class:`_expression.Select`
+        that encloses this one via
+        its :term:`WHERE clause`, ORDER BY, HAVING or
+        :term:`columns clause` will be omitted from this
+        :class:`_expression.Select`
+        object's :term:`FROM clause`.
+        Setting an explicit correlation collection using the
+        :meth:`_expression.Select.correlate`
+        method provides a fixed list of FROM objects
+        that can potentially take place in this process.
+
+        When :meth:`_expression.Select.correlate`
+        is used to apply specific FROM clauses
+        for correlation, the FROM elements become candidates for
+        correlation regardless of how deeply nested this
+        :class:`_expression.Select`
+        object is, relative to an enclosing :class:`_expression.Select`
+        which refers to
+        the same FROM object.  This is in contrast to the behavior of
+        "auto-correlation" which only correlates to an immediate enclosing
+        :class:`_expression.Select`.
+        Multi-level correlation ensures that the link
+        between enclosed and enclosing :class:`_expression.Select`
+        is always via
+        at least one WHERE/ORDER BY/HAVING/columns clause in order for
+        correlation to take place.
+
+        If ``None`` is passed, the :class:`_expression.Select`
+        object will correlate
+        none of its FROM entries, and all will render unconditionally
+        in the local FROM clause.
+
+        :param \*fromclauses: one or more :class:`.FromClause` or other
+         FROM-compatible construct such as an ORM mapped entity to become part
+         of the correlate collection; alternatively pass a single value
+         ``None`` to remove all existing correlations.
+
+        .. seealso::
+
+            :meth:`_expression.Select.correlate_except`
+
+            :ref:`tutorial_scalar_subquery`
+
+        """
+
+        # tests failing when we try to change how these
+        # arguments are passed
+
+        self._auto_correlate = False
+        if not fromclauses or fromclauses[0] in {None, False}:
+            if len(fromclauses) > 1:
+                raise exc.ArgumentError(
+                    "additional FROM objects not accepted when "
+                    "passing None/False to correlate()"
+                )
+            self._correlate = ()
+        else:
+            self._correlate = self._correlate + tuple(
+                coercions.expect(roles.FromClauseRole, f) for f in fromclauses
+            )
+        return self
+
+    @_generative
+    def correlate_except(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        r"""Return a new :class:`_expression.Select`
+        which will omit the given FROM
+        clauses from the auto-correlation process.
+
+        Calling :meth:`_expression.Select.correlate_except` turns off the
+        :class:`_expression.Select` object's default behavior of
+        "auto-correlation" for the given FROM elements.  An element
+        specified here will unconditionally appear in the FROM list, while
+        all other FROM elements remain subject to normal auto-correlation
+        behaviors.
+
+        If ``None`` is passed, or no arguments are passed,
+        the :class:`_expression.Select` object will correlate all of its
+        FROM entries.
+
+        :param \*fromclauses: a list of one or more
+         :class:`_expression.FromClause`
+         constructs, or other compatible constructs (i.e. ORM-mapped
+         classes) to become part of the correlate-exception collection.
+
+        .. seealso::
+
+            :meth:`_expression.Select.correlate`
+
+            :ref:`tutorial_scalar_subquery`
+
+        """
+
+        self._auto_correlate = False
+        if not fromclauses or fromclauses[0] in {None, False}:
+            if len(fromclauses) > 1:
+                raise exc.ArgumentError(
+                    "additional FROM objects not accepted when "
+                    "passing None/False to correlate_except()"
+                )
+            self._correlate_except = ()
+        else:
+            self._correlate_except = (self._correlate_except or ()) + tuple(
+                coercions.expect(roles.FromClauseRole, f) for f in fromclauses
+            )
+
+        return self
+
+    @HasMemoized_ro_memoized_attribute
+    def selected_columns(
+        self,
+    ) -> ColumnCollection[str, ColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        representing the columns that
+        this SELECT statement or similar construct returns in its result set,
+        not including :class:`_sql.TextClause` constructs.
+
+        This collection differs from the :attr:`_expression.FromClause.columns`
+        collection of a :class:`_expression.FromClause` in that the columns
+        within this collection cannot be directly nested inside another SELECT
+        statement; a subquery must be applied first which provides for the
+        necessary parenthesization required by SQL.
+
+        For a :func:`_expression.select` construct, the collection here is
+        exactly what would be rendered inside the "SELECT" statement, and the
+        :class:`_expression.ColumnElement` objects are directly present as they
+        were given, e.g.::
+
+            col1 = column("q", Integer)
+            col2 = column("p", Integer)
+            stmt = select(col1, col2)
+
+        Above, ``stmt.selected_columns`` would be a collection that contains
+        the ``col1`` and ``col2`` objects directly. For a statement that is
+        against a :class:`_schema.Table` or other
+        :class:`_expression.FromClause`, the collection will use the
+        :class:`_expression.ColumnElement` objects that are in the
+        :attr:`_expression.FromClause.c` collection of the from element.
+
+        A use case for the :attr:`_sql.Select.selected_columns` collection is
+        to allow the existing columns to be referenced when adding additional
+        criteria, e.g.::
+
+            def filter_on_id(my_select, id):
+                return my_select.where(my_select.selected_columns["id"] == id)
+
+
+            stmt = select(MyModel)
+
+            # adds "WHERE id=:param" to the statement
+            stmt = filter_on_id(stmt, 42)
+
+        .. note::
+
+            The :attr:`_sql.Select.selected_columns` collection does not
+            include expressions established in the columns clause using the
+            :func:`_sql.text` construct; these are silently omitted from the
+            collection. To use plain textual column expressions inside of a
+            :class:`_sql.Select` construct, use the :func:`_sql.literal_column`
+            construct.
+
+
+        .. versionadded:: 1.4
+
+        """
+
+        # compare to SelectState._generate_columns_plus_names, which
+        # generates the actual names used in the SELECT string.  that
+        # method is more complex because it also renders columns that are
+        # fully ambiguous, e.g. same column more than once.
+        conv = cast(
+            "Callable[[Any], str]",
+            SelectState._column_naming_convention(self._label_style),
+        )
+
+        cc: ColumnCollection[str, ColumnElement[Any]] = ColumnCollection(
+            [
+                (conv(c), c)
+                for c in self._all_selected_columns
+                if is_column_element(c)
+            ]
+        )
+        return cc.as_readonly()
+
+    @HasMemoized_ro_memoized_attribute
+    def _all_selected_columns(self) -> _SelectIterable:
+        meth = SelectState.get_plugin_class(self).all_selected_columns
+        return list(meth(self))
+
+    def _ensure_disambiguated_names(self) -> Select[Any]:
+        if self._label_style is LABEL_STYLE_NONE:
+            self = self.set_label_style(LABEL_STYLE_DISAMBIGUATE_ONLY)
+        return self
+
+    def _generate_fromclause_column_proxies(
+        self,
+        subquery: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        *,
+        proxy_compound_columns: Optional[
+            Iterable[Sequence[ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        """Generate column proxies to place in the exported ``.c``
+        collection of a subquery."""
+
+        if proxy_compound_columns:
+            extra_col_iterator = proxy_compound_columns
+            prox = [
+                c._make_proxy(
+                    subquery,
+                    key=proxy_key,
+                    name=required_label_name,
+                    name_is_truncatable=True,
+                    compound_select_cols=extra_cols,
+                    primary_key=primary_key,
+                    foreign_keys=foreign_keys,
+                )
+                for (
+                    (
+                        required_label_name,
+                        proxy_key,
+                        fallback_label_name,
+                        c,
+                        repeated,
+                    ),
+                    extra_cols,
+                ) in (
+                    zip(
+                        self._generate_columns_plus_names(False),
+                        extra_col_iterator,
+                    )
+                )
+                if is_column_element(c)
+            ]
+        else:
+            prox = [
+                c._make_proxy(
+                    subquery,
+                    key=proxy_key,
+                    name=required_label_name,
+                    name_is_truncatable=True,
+                    primary_key=primary_key,
+                    foreign_keys=foreign_keys,
+                )
+                for (
+                    required_label_name,
+                    proxy_key,
+                    fallback_label_name,
+                    c,
+                    repeated,
+                ) in (self._generate_columns_plus_names(False))
+                if is_column_element(c)
+            ]
+
+        columns._populate_separate_keys(prox)
+
+    def _needs_parens_for_grouping(self) -> bool:
+        return self._has_row_limiting_clause or bool(
+            self._order_by_clause.clauses
+        )
+
+    def self_group(
+        self, against: Optional[OperatorType] = None
+    ) -> Union[SelectStatementGrouping[Self], Self]:
+        """Return a 'grouping' construct as per the
+        :class:`_expression.ClauseElement` specification.
+
+        This produces an element that can be embedded in an expression. Note
+        that this method is called automatically as needed when constructing
+        expressions and should not require explicit use.
+
+        """
+        if (
+            isinstance(against, CompoundSelect)
+            and not self._needs_parens_for_grouping()
+        ):
+            return self
+        else:
+            return SelectStatementGrouping(self)
+
+    def union(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``UNION`` of this select() construct against
+        the given selectables provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        :param \**kwargs: keyword arguments are forwarded to the constructor
+         for the newly created :class:`_sql.CompoundSelect` object.
+
+        """
+        return CompoundSelect._create_union(self, *other)
+
+    def union_all(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``UNION ALL`` of this select() construct against
+        the given selectables provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        :param \**kwargs: keyword arguments are forwarded to the constructor
+         for the newly created :class:`_sql.CompoundSelect` object.
+
+        """
+        return CompoundSelect._create_union_all(self, *other)
+
+    def except_(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``EXCEPT`` of this select() construct against
+        the given selectable provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        """
+        return CompoundSelect._create_except(self, *other)
+
+    def except_all(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``EXCEPT ALL`` of this select() construct against
+        the given selectables provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        """
+        return CompoundSelect._create_except_all(self, *other)
+
+    def intersect(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``INTERSECT`` of this select() construct against
+        the given selectables provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        :param \**kwargs: keyword arguments are forwarded to the constructor
+         for the newly created :class:`_sql.CompoundSelect` object.
+
+        """
+        return CompoundSelect._create_intersect(self, *other)
+
+    def intersect_all(
+        self, *other: _SelectStatementForCompoundArgument
+    ) -> CompoundSelect:
+        r"""Return a SQL ``INTERSECT ALL`` of this select() construct
+        against the given selectables provided as positional arguments.
+
+        :param \*other: one or more elements with which to create a
+         UNION.
+
+         .. versionchanged:: 1.4.28
+
+            multiple elements are now accepted.
+
+        :param \**kwargs: keyword arguments are forwarded to the constructor
+         for the newly created :class:`_sql.CompoundSelect` object.
+
+        """
+        return CompoundSelect._create_intersect_all(self, *other)
+
+
+class ScalarSelect(
+    roles.InElementRole, Generative, GroupedElement, ColumnElement[_T]
+):
+    """Represent a scalar subquery.
+
+
+    A :class:`_sql.ScalarSelect` is created by invoking the
+    :meth:`_sql.SelectBase.scalar_subquery` method.   The object
+    then participates in other SQL expressions as a SQL column expression
+    within the :class:`_sql.ColumnElement` hierarchy.
+
+    .. seealso::
+
+        :meth:`_sql.SelectBase.scalar_subquery`
+
+        :ref:`tutorial_scalar_subquery` - in the 2.0 tutorial
+
+    """
+
+    _traverse_internals: _TraverseInternalsType = [
+        ("element", InternalTraversal.dp_clauseelement),
+        ("type", InternalTraversal.dp_type),
+    ]
+
+    _from_objects: List[FromClause] = []
+    _is_from_container = True
+    if not TYPE_CHECKING:
+        _is_implicitly_boolean = False
+    inherit_cache = True
+
+    element: SelectBase
+
+    def __init__(self, element: SelectBase) -> None:
+        self.element = element
+        self.type = element._scalar_type()
+        self._propagate_attrs = element._propagate_attrs
+
+    def __getattr__(self, attr: str) -> Any:
+        return getattr(self.element, attr)
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {"element": self.element, "type": self.type}
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        self.element = state["element"]
+        self.type = state["type"]
+
+    @property
+    def columns(self) -> NoReturn:
+        raise exc.InvalidRequestError(
+            "Scalar Select expression has no "
+            "columns; use this object directly "
+            "within a column-level expression."
+        )
+
+    c = columns
+
+    @_generative
+    def where(self, crit: _ColumnExpressionArgument[bool]) -> Self:
+        """Apply a WHERE clause to the SELECT statement referred to
+        by this :class:`_expression.ScalarSelect`.
+
+        """
+        self.element = cast("Select[Any]", self.element).where(crit)
+        return self
+
+    def self_group(self, against: Optional[OperatorType] = None) -> Self:
+        return self
+
+    if TYPE_CHECKING:
+
+        def _ungroup(self) -> Select[Any]: ...
+
+    @_generative
+    def correlate(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        r"""Return a new :class:`_expression.ScalarSelect`
+        which will correlate the given FROM
+        clauses to that of an enclosing :class:`_expression.Select`.
+
+        This method is mirrored from the :meth:`_sql.Select.correlate` method
+        of the underlying :class:`_sql.Select`.  The method applies the
+        :meth:_sql.Select.correlate` method, then returns a new
+        :class:`_sql.ScalarSelect` against that statement.
+
+        .. versionadded:: 1.4 Previously, the
+           :meth:`_sql.ScalarSelect.correlate`
+           method was only available from :class:`_sql.Select`.
+
+        :param \*fromclauses: a list of one or more
+         :class:`_expression.FromClause`
+         constructs, or other compatible constructs (i.e. ORM-mapped
+         classes) to become part of the correlate collection.
+
+        .. seealso::
+
+            :meth:`_expression.ScalarSelect.correlate_except`
+
+            :ref:`tutorial_scalar_subquery` - in the 2.0 tutorial
+
+
+        """
+        self.element = cast("Select[Any]", self.element).correlate(
+            *fromclauses
+        )
+        return self
+
+    @_generative
+    def correlate_except(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        r"""Return a new :class:`_expression.ScalarSelect`
+        which will omit the given FROM
+        clauses from the auto-correlation process.
+
+        This method is mirrored from the
+        :meth:`_sql.Select.correlate_except` method of the underlying
+        :class:`_sql.Select`.  The method applies the
+        :meth:_sql.Select.correlate_except` method, then returns a new
+        :class:`_sql.ScalarSelect` against that statement.
+
+        .. versionadded:: 1.4 Previously, the
+           :meth:`_sql.ScalarSelect.correlate_except`
+           method was only available from :class:`_sql.Select`.
+
+        :param \*fromclauses: a list of one or more
+         :class:`_expression.FromClause`
+         constructs, or other compatible constructs (i.e. ORM-mapped
+         classes) to become part of the correlate-exception collection.
+
+        .. seealso::
+
+            :meth:`_expression.ScalarSelect.correlate`
+
+            :ref:`tutorial_scalar_subquery` - in the 2.0 tutorial
+
+
+        """
+
+        self.element = cast("Select[Any]", self.element).correlate_except(
+            *fromclauses
+        )
+        return self
+
+
+class Exists(UnaryExpression[bool]):
+    """Represent an ``EXISTS`` clause.
+
+    See :func:`_sql.exists` for a description of usage.
+
+    An ``EXISTS`` clause can also be constructed from a :func:`_sql.select`
+    instance by calling :meth:`_sql.SelectBase.exists`.
+
+    """
+
+    inherit_cache = True
+    element: Union[SelectStatementGrouping[Select[Any]], ScalarSelect[Any]]
+
+    def __init__(
+        self,
+        __argument: Optional[
+            Union[_ColumnsClauseArgument[Any], SelectBase, ScalarSelect[Any]]
+        ] = None,
+    ):
+        s: ScalarSelect[Any]
+
+        # TODO: this seems like we should be using coercions for this
+        if __argument is None:
+            s = Select(literal_column("*")).scalar_subquery()
+        elif isinstance(__argument, SelectBase):
+            s = __argument.scalar_subquery()
+            s._propagate_attrs = __argument._propagate_attrs
+        elif isinstance(__argument, ScalarSelect):
+            s = __argument
+        else:
+            s = Select(__argument).scalar_subquery()
+
+        UnaryExpression.__init__(
+            self,
+            s,
+            operator=operators.exists,
+            type_=type_api.BOOLEANTYPE,
+            wraps_column_expression=True,
+        )
+
+    @util.ro_non_memoized_property
+    def _from_objects(self) -> List[FromClause]:
+        return []
+
+    def _regroup(
+        self, fn: Callable[[Select[Any]], Select[Any]]
+    ) -> SelectStatementGrouping[Select[Any]]:
+        element = self.element._ungroup()
+        new_element = fn(element)
+
+        return_value = new_element.self_group(against=operators.exists)
+        assert isinstance(return_value, SelectStatementGrouping)
+        return return_value
+
+    def select(self) -> Select[Tuple[bool]]:
+        r"""Return a SELECT of this :class:`_expression.Exists`.
+
+        e.g.::
+
+            stmt = exists(some_table.c.id).where(some_table.c.id == 5).select()
+
+        This will produce a statement resembling:
+
+        .. sourcecode:: sql
+
+            SELECT EXISTS (SELECT id FROM some_table WHERE some_table = :param) AS anon_1
+
+        .. seealso::
+
+            :func:`_expression.select` - general purpose
+            method which allows for arbitrary column lists.
+
+        """  # noqa
+
+        return Select(self)
+
+    def correlate(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        """Apply correlation to the subquery noted by this
+        :class:`_sql.Exists`.
+
+        .. seealso::
+
+            :meth:`_sql.ScalarSelect.correlate`
+
+        """
+        e = self._clone()
+        e.element = self._regroup(
+            lambda element: element.correlate(*fromclauses)
+        )
+        return e
+
+    def correlate_except(
+        self,
+        *fromclauses: Union[Literal[None, False], _FromClauseArgument],
+    ) -> Self:
+        """Apply correlation to the subquery noted by this
+        :class:`_sql.Exists`.
+
+        .. seealso::
+
+            :meth:`_sql.ScalarSelect.correlate_except`
+
+        """
+
+        e = self._clone()
+        e.element = self._regroup(
+            lambda element: element.correlate_except(*fromclauses)
+        )
+        return e
+
+    def select_from(self, *froms: _FromClauseArgument) -> Self:
+        """Return a new :class:`_expression.Exists` construct,
+        applying the given
+        expression to the :meth:`_expression.Select.select_from`
+        method of the select
+        statement contained.
+
+        .. note:: it is typically preferable to build a :class:`_sql.Select`
+           statement first, including the desired WHERE clause, then use the
+           :meth:`_sql.SelectBase.exists` method to produce an
+           :class:`_sql.Exists` object at once.
+
+        """
+        e = self._clone()
+        e.element = self._regroup(lambda element: element.select_from(*froms))
+        return e
+
+    def where(self, *clause: _ColumnExpressionArgument[bool]) -> Self:
+        """Return a new :func:`_expression.exists` construct with the
+        given expression added to
+        its WHERE clause, joined to the existing clause via AND, if any.
+
+
+        .. note:: it is typically preferable to build a :class:`_sql.Select`
+           statement first, including the desired WHERE clause, then use the
+           :meth:`_sql.SelectBase.exists` method to produce an
+           :class:`_sql.Exists` object at once.
+
+        """
+        e = self._clone()
+        e.element = self._regroup(lambda element: element.where(*clause))
+        return e
+
+
+class TextualSelect(SelectBase, ExecutableReturnsRows, Generative):
+    """Wrap a :class:`_expression.TextClause` construct within a
+    :class:`_expression.SelectBase`
+    interface.
+
+    This allows the :class:`_expression.TextClause` object to gain a
+    ``.c`` collection
+    and other FROM-like capabilities such as
+    :meth:`_expression.FromClause.alias`,
+    :meth:`_expression.SelectBase.cte`, etc.
+
+    The :class:`_expression.TextualSelect` construct is produced via the
+    :meth:`_expression.TextClause.columns`
+    method - see that method for details.
+
+    .. versionchanged:: 1.4 the :class:`_expression.TextualSelect`
+       class was renamed
+       from ``TextAsFrom``, to more correctly suit its role as a
+       SELECT-oriented object and not a FROM clause.
+
+    .. seealso::
+
+        :func:`_expression.text`
+
+        :meth:`_expression.TextClause.columns` - primary creation interface.
+
+    """
+
+    __visit_name__ = "textual_select"
+
+    _label_style = LABEL_STYLE_NONE
+
+    _traverse_internals: _TraverseInternalsType = (
+        [
+            ("element", InternalTraversal.dp_clauseelement),
+            ("column_args", InternalTraversal.dp_clauseelement_list),
+        ]
+        + SupportsCloneAnnotations._clone_annotations_traverse_internals
+        + HasCTE._has_ctes_traverse_internals
+    )
+
+    _is_textual = True
+
+    is_text = True
+    is_select = True
+
+    def __init__(
+        self,
+        text: TextClause,
+        columns: List[_ColumnExpressionArgument[Any]],
+        positional: bool = False,
+    ) -> None:
+        self._init(
+            text,
+            # convert for ORM attributes->columns, etc
+            [
+                coercions.expect(roles.LabeledColumnExprRole, c)
+                for c in columns
+            ],
+            positional,
+        )
+
+    def _init(
+        self,
+        text: TextClause,
+        columns: List[NamedColumn[Any]],
+        positional: bool = False,
+    ) -> None:
+        self.element = text
+        self.column_args = columns
+        self.positional = positional
+
+    @HasMemoized_ro_memoized_attribute
+    def selected_columns(
+        self,
+    ) -> ColumnCollection[str, KeyedColumnElement[Any]]:
+        """A :class:`_expression.ColumnCollection`
+        representing the columns that
+        this SELECT statement or similar construct returns in its result set,
+        not including :class:`_sql.TextClause` constructs.
+
+        This collection differs from the :attr:`_expression.FromClause.columns`
+        collection of a :class:`_expression.FromClause` in that the columns
+        within this collection cannot be directly nested inside another SELECT
+        statement; a subquery must be applied first which provides for the
+        necessary parenthesization required by SQL.
+
+        For a :class:`_expression.TextualSelect` construct, the collection
+        contains the :class:`_expression.ColumnElement` objects that were
+        passed to the constructor, typically via the
+        :meth:`_expression.TextClause.columns` method.
+
+
+        .. versionadded:: 1.4
+
+        """
+        return ColumnCollection(
+            (c.key, c) for c in self.column_args
+        ).as_readonly()
+
+    @util.ro_non_memoized_property
+    def _all_selected_columns(self) -> _SelectIterable:
+        return self.column_args
+
+    def set_label_style(self, style: SelectLabelStyle) -> TextualSelect:
+        return self
+
+    def _ensure_disambiguated_names(self) -> TextualSelect:
+        return self
+
+    @_generative
+    def bindparams(
+        self,
+        *binds: BindParameter[Any],
+        **bind_as_values: Any,
+    ) -> Self:
+        self.element = self.element.bindparams(*binds, **bind_as_values)
+        return self
+
+    def _generate_fromclause_column_proxies(
+        self,
+        fromclause: FromClause,
+        columns: ColumnCollection[str, KeyedColumnElement[Any]],
+        primary_key: ColumnSet,
+        foreign_keys: Set[KeyedColumnElement[Any]],
+        *,
+        proxy_compound_columns: Optional[
+            Iterable[Sequence[ColumnElement[Any]]]
+        ] = None,
+    ) -> None:
+        if TYPE_CHECKING:
+            assert isinstance(fromclause, Subquery)
+
+        if proxy_compound_columns:
+            columns._populate_separate_keys(
+                c._make_proxy(
+                    fromclause,
+                    compound_select_cols=extra_cols,
+                    primary_key=primary_key,
+                    foreign_keys=foreign_keys,
+                )
+                for c, extra_cols in zip(
+                    self.column_args, proxy_compound_columns
+                )
+            )
+        else:
+            columns._populate_separate_keys(
+                c._make_proxy(
+                    fromclause,
+                    primary_key=primary_key,
+                    foreign_keys=foreign_keys,
+                )
+                for c in self.column_args
+            )
+
+    def _scalar_type(self) -> Union[TypeEngine[Any], Any]:
+        return self.column_args[0].type
+
+
+TextAsFrom = TextualSelect
+"""Backwards compatibility with the previous name"""
+
+
+class AnnotatedFromClause(Annotated):
+    def _copy_internals(self, **kw: Any) -> None:
+        super()._copy_internals(**kw)
+        if kw.get("ind_cols_on_fromclause", False):
+            ee = self._Annotated__element  # type: ignore
+
+            self.c = ee.__class__.c.fget(self)  # type: ignore
+
+    @util.ro_memoized_property
+    def c(self) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]:
+        """proxy the .c collection of the underlying FromClause.
+
+        Originally implemented in 2008 as a simple load of the .c collection
+        when the annotated construct was created (see d3621ae961a), in modern
+        SQLAlchemy versions this can be expensive for statements constructed
+        with ORM aliases.   So for #8796 SQLAlchemy 2.0 we instead proxy
+        it, which works just as well.
+
+        Two different use cases seem to require the collection either copied
+        from the underlying one, or unique to this AnnotatedFromClause.
+
+        See test_selectable->test_annotated_corresponding_column
+
+        """
+        ee = self._Annotated__element  # type: ignore
+        return ee.c  # type: ignore
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/sqltypes.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/sqltypes.py
new file mode 100644
index 00000000..ee471a6c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/sqltypes.py
@@ -0,0 +1,3835 @@
+# sql/sqltypes.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""SQL specific types.
+
+"""
+from __future__ import annotations
+
+import collections.abc as collections_abc
+import datetime as dt
+import decimal
+import enum
+import json
+import pickle
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+from uuid import UUID as _python_UUID
+
+from . import coercions
+from . import elements
+from . import operators
+from . import roles
+from . import type_api
+from .base import _NONE_NAME
+from .base import NO_ARG
+from .base import SchemaEventTarget
+from .cache_key import HasCacheKey
+from .elements import quoted_name
+from .elements import Slice
+from .elements import TypeCoerce as type_coerce  # noqa
+from .type_api import Emulated
+from .type_api import NativeForEmulated  # noqa
+from .type_api import to_instance as to_instance
+from .type_api import TypeDecorator as TypeDecorator
+from .type_api import TypeEngine as TypeEngine
+from .type_api import TypeEngineMixin
+from .type_api import Variant  # noqa
+from .visitors import InternalTraversal
+from .. import event
+from .. import exc
+from .. import inspection
+from .. import util
+from ..engine import processors
+from ..util import langhelpers
+from ..util import OrderedDict
+from ..util import warn_deprecated
+from ..util.typing import get_args
+from ..util.typing import is_literal
+from ..util.typing import is_pep695
+from ..util.typing import Literal
+
+if TYPE_CHECKING:
+    from ._typing import _ColumnExpressionArgument
+    from ._typing import _TypeEngineArgument
+    from .operators import OperatorType
+    from .schema import MetaData
+    from .type_api import _BindProcessorType
+    from .type_api import _ComparatorFactory
+    from .type_api import _MatchedOnType
+    from .type_api import _ResultProcessorType
+    from ..engine.interfaces import Dialect
+
+_T = TypeVar("_T", bound="Any")
+_CT = TypeVar("_CT", bound=Any)
+_TE = TypeVar("_TE", bound="TypeEngine[Any]")
+
+
+class HasExpressionLookup(TypeEngineMixin):
+    """Mixin expression adaptations based on lookup tables.
+
+    These rules are currently used by the numeric, integer and date types
+    which have detailed cross-expression coercion rules.
+
+    """
+
+    @property
+    def _expression_adaptations(self):
+        raise NotImplementedError()
+
+    class Comparator(TypeEngine.Comparator[_CT]):
+        __slots__ = ()
+
+        _blank_dict = util.EMPTY_DICT
+
+        def _adapt_expression(
+            self,
+            op: OperatorType,
+            other_comparator: TypeEngine.Comparator[Any],
+        ) -> Tuple[OperatorType, TypeEngine[Any]]:
+            othertype = other_comparator.type._type_affinity
+            if TYPE_CHECKING:
+                assert isinstance(self.type, HasExpressionLookup)
+            lookup = self.type._expression_adaptations.get(
+                op, self._blank_dict
+            ).get(othertype, self.type)
+            if lookup is othertype:
+                return (op, other_comparator.type)
+            elif lookup is self.type._type_affinity:
+                return (op, self.type)
+            else:
+                return (op, to_instance(lookup))
+
+    comparator_factory: _ComparatorFactory[Any] = Comparator
+
+
+class Concatenable(TypeEngineMixin):
+    """A mixin that marks a type as supporting 'concatenation',
+    typically strings."""
+
+    class Comparator(TypeEngine.Comparator[_T]):
+        __slots__ = ()
+
+        def _adapt_expression(
+            self,
+            op: OperatorType,
+            other_comparator: TypeEngine.Comparator[Any],
+        ) -> Tuple[OperatorType, TypeEngine[Any]]:
+            if op is operators.add and isinstance(
+                other_comparator,
+                (Concatenable.Comparator, NullType.Comparator),
+            ):
+                return operators.concat_op, self.expr.type
+            else:
+                return super()._adapt_expression(op, other_comparator)
+
+    comparator_factory: _ComparatorFactory[Any] = Comparator
+
+
+class Indexable(TypeEngineMixin):
+    """A mixin that marks a type as supporting indexing operations,
+    such as array or JSON structures.
+
+    """
+
+    class Comparator(TypeEngine.Comparator[_T]):
+        __slots__ = ()
+
+        def _setup_getitem(self, index):
+            raise NotImplementedError()
+
+        def __getitem__(self, index):
+            (
+                adjusted_op,
+                adjusted_right_expr,
+                result_type,
+            ) = self._setup_getitem(index)
+            return self.operate(
+                adjusted_op, adjusted_right_expr, result_type=result_type
+            )
+
+    comparator_factory: _ComparatorFactory[Any] = Comparator
+
+
+class String(Concatenable, TypeEngine[str]):
+    """The base for all string and character types.
+
+    In SQL, corresponds to VARCHAR.
+
+    The `length` field is usually required when the `String` type is
+    used within a CREATE TABLE statement, as VARCHAR requires a length
+    on most databases.
+
+    """
+
+    __visit_name__ = "string"
+
+    def __init__(
+        self,
+        length: Optional[int] = None,
+        collation: Optional[str] = None,
+    ):
+        """
+        Create a string-holding type.
+
+        :param length: optional, a length for the column for use in
+          DDL and CAST expressions.  May be safely omitted if no ``CREATE
+          TABLE`` will be issued.  Certain databases may require a
+          ``length`` for use in DDL, and will raise an exception when
+          the ``CREATE TABLE`` DDL is issued if a ``VARCHAR``
+          with no length is included.  Whether the value is
+          interpreted as bytes or characters is database specific.
+
+        :param collation: Optional, a column-level collation for
+          use in DDL and CAST expressions.  Renders using the
+          COLLATE keyword supported by SQLite, MySQL, and PostgreSQL.
+          E.g.:
+
+          .. sourcecode:: pycon+sql
+
+            >>> from sqlalchemy import cast, select, String
+            >>> print(select(cast("some string", String(collation="utf8"))))
+            {printsql}SELECT CAST(:param_1 AS VARCHAR COLLATE utf8) AS anon_1
+
+          .. note::
+
+            In most cases, the :class:`.Unicode` or :class:`.UnicodeText`
+            datatypes should be used for a :class:`_schema.Column` that expects
+            to store non-ascii data. These datatypes will ensure that the
+            correct types are used on the database.
+
+        """
+
+        self.length = length
+        self.collation = collation
+
+    def _with_collation(self, collation):
+        new_type = self.copy()
+        new_type.collation = collation
+        return new_type
+
+    def _resolve_for_literal(self, value):
+        # I was SO PROUD of my regex trick, but we dont need it.
+        # re.search(r"[^\u0000-\u007F]", value)
+
+        if value.isascii():
+            return _STRING
+        else:
+            return _UNICODE
+
+    def literal_processor(self, dialect):
+        def process(value):
+            value = value.replace("'", "''")
+
+            if dialect.identifier_preparer._double_percents:
+                value = value.replace("%", "%%")
+
+            return "'%s'" % value
+
+        return process
+
+    def bind_processor(self, dialect):
+        return None
+
+    def result_processor(self, dialect, coltype):
+        return None
+
+    @property
+    def python_type(self):
+        return str
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.STRING
+
+
+class Text(String):
+    """A variably sized string type.
+
+    In SQL, usually corresponds to CLOB or TEXT.  In general, TEXT objects
+    do not have a length; while some databases will accept a length
+    argument here, it will be rejected by others.
+
+    """
+
+    __visit_name__ = "text"
+
+
+class Unicode(String):
+    """A variable length Unicode string type.
+
+    The :class:`.Unicode` type is a :class:`.String` subclass that assumes
+    input and output strings that may contain non-ASCII characters, and for
+    some backends implies an underlying column type that is explicitly
+    supporting of non-ASCII data, such as ``NVARCHAR`` on Oracle Database and
+    SQL Server.  This will impact the output of ``CREATE TABLE`` statements and
+    ``CAST`` functions at the dialect level.
+
+    The character encoding used by the :class:`.Unicode` type that is used to
+    transmit and receive data to the database is usually determined by the
+    DBAPI itself. All modern DBAPIs accommodate non-ASCII strings but may have
+    different methods of managing database encodings; if necessary, this
+    encoding should be configured as detailed in the notes for the target DBAPI
+    in the :ref:`dialect_toplevel` section.
+
+    In modern SQLAlchemy, use of the :class:`.Unicode` datatype does not
+    imply any encoding/decoding behavior within SQLAlchemy itself.  In Python
+    3, all string objects are inherently Unicode capable, and SQLAlchemy
+    does not produce bytestring objects nor does it accommodate a DBAPI that
+    does not return Python Unicode objects in result sets for string values.
+
+    .. warning:: Some database backends, particularly SQL Server with pyodbc,
+       are known to have undesirable behaviors regarding data that is noted
+       as being of ``NVARCHAR`` type as opposed to ``VARCHAR``, including
+       datatype mismatch errors and non-use of indexes.  See the section
+       on :meth:`.DialectEvents.do_setinputsizes` for background on working
+       around unicode character issues for backends like SQL Server with
+       pyodbc as well as cx_Oracle.
+
+    .. seealso::
+
+        :class:`.UnicodeText` - unlengthed textual counterpart
+        to :class:`.Unicode`.
+
+        :meth:`.DialectEvents.do_setinputsizes`
+
+    """
+
+    __visit_name__ = "unicode"
+
+
+class UnicodeText(Text):
+    """An unbounded-length Unicode string type.
+
+    See :class:`.Unicode` for details on the unicode
+    behavior of this object.
+
+    Like :class:`.Unicode`, usage the :class:`.UnicodeText` type implies a
+    unicode-capable type being used on the backend, such as
+    ``NCLOB``, ``NTEXT``.
+
+    """
+
+    __visit_name__ = "unicode_text"
+
+
+class Integer(HasExpressionLookup, TypeEngine[int]):
+    """A type for ``int`` integers."""
+
+    __visit_name__ = "integer"
+
+    if TYPE_CHECKING:
+
+        @util.ro_memoized_property
+        def _type_affinity(self) -> Type[Integer]: ...
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.NUMBER
+
+    @property
+    def python_type(self):
+        return int
+
+    def _resolve_for_literal(self, value):
+        if value.bit_length() >= 32:
+            return _BIGINTEGER
+        else:
+            return self
+
+    def literal_processor(self, dialect):
+        def process(value):
+            return str(int(value))
+
+        return process
+
+    @util.memoized_property
+    def _expression_adaptations(self):
+        return {
+            operators.add: {
+                Date: Date,
+                Integer: self.__class__,
+                Numeric: Numeric,
+            },
+            operators.mul: {
+                Interval: Interval,
+                Integer: self.__class__,
+                Numeric: Numeric,
+            },
+            operators.truediv: {Integer: Numeric, Numeric: Numeric},
+            operators.floordiv: {Integer: self.__class__, Numeric: Numeric},
+            operators.sub: {Integer: self.__class__, Numeric: Numeric},
+        }
+
+
+class SmallInteger(Integer):
+    """A type for smaller ``int`` integers.
+
+    Typically generates a ``SMALLINT`` in DDL, and otherwise acts like
+    a normal :class:`.Integer` on the Python side.
+
+    """
+
+    __visit_name__ = "small_integer"
+
+
+class BigInteger(Integer):
+    """A type for bigger ``int`` integers.
+
+    Typically generates a ``BIGINT`` in DDL, and otherwise acts like
+    a normal :class:`.Integer` on the Python side.
+
+    """
+
+    __visit_name__ = "big_integer"
+
+
+_N = TypeVar("_N", bound=Union[decimal.Decimal, float])
+
+
+class Numeric(HasExpressionLookup, TypeEngine[_N]):
+    """Base for non-integer numeric types, such as
+    ``NUMERIC``, ``FLOAT``, ``DECIMAL``, and other variants.
+
+    The :class:`.Numeric` datatype when used directly will render DDL
+    corresponding to precision numerics if available, such as
+    ``NUMERIC(precision, scale)``.  The :class:`.Float` subclass will
+    attempt to render a floating-point datatype such as ``FLOAT(precision)``.
+
+    :class:`.Numeric` returns Python ``decimal.Decimal`` objects by default,
+    based on the default value of ``True`` for the
+    :paramref:`.Numeric.asdecimal` parameter.  If this parameter is set to
+    False, returned values are coerced to Python ``float`` objects.
+
+    The :class:`.Float` subtype, being more specific to floating point,
+    defaults the :paramref:`.Float.asdecimal` flag to False so that the
+    default Python datatype is ``float``.
+
+    .. note::
+
+        When using a :class:`.Numeric` datatype against a database type that
+        returns Python floating point values to the driver, the accuracy of the
+        decimal conversion indicated by :paramref:`.Numeric.asdecimal` may be
+        limited.   The behavior of specific numeric/floating point datatypes
+        is a product of the SQL datatype in use, the Python :term:`DBAPI`
+        in use, as well as strategies that may be present within
+        the SQLAlchemy dialect in use.   Users requiring specific precision/
+        scale are encouraged to experiment with the available datatypes
+        in order to determine the best results.
+
+    """
+
+    __visit_name__ = "numeric"
+
+    if TYPE_CHECKING:
+
+        @util.ro_memoized_property
+        def _type_affinity(self) -> Type[Numeric[_N]]: ...
+
+    _default_decimal_return_scale = 10
+
+    @overload
+    def __init__(
+        self: Numeric[decimal.Decimal],
+        precision: Optional[int] = ...,
+        scale: Optional[int] = ...,
+        decimal_return_scale: Optional[int] = ...,
+        asdecimal: Literal[True] = ...,
+    ): ...
+
+    @overload
+    def __init__(
+        self: Numeric[float],
+        precision: Optional[int] = ...,
+        scale: Optional[int] = ...,
+        decimal_return_scale: Optional[int] = ...,
+        asdecimal: Literal[False] = ...,
+    ): ...
+
+    def __init__(
+        self,
+        precision: Optional[int] = None,
+        scale: Optional[int] = None,
+        decimal_return_scale: Optional[int] = None,
+        asdecimal: bool = True,
+    ):
+        """
+        Construct a Numeric.
+
+        :param precision: the numeric precision for use in DDL ``CREATE
+          TABLE``.
+
+        :param scale: the numeric scale for use in DDL ``CREATE TABLE``.
+
+        :param asdecimal: default True.  Return whether or not
+          values should be sent as Python Decimal objects, or
+          as floats.   Different DBAPIs send one or the other based on
+          datatypes - the Numeric type will ensure that return values
+          are one or the other across DBAPIs consistently.
+
+        :param decimal_return_scale: Default scale to use when converting
+         from floats to Python decimals.  Floating point values will typically
+         be much longer due to decimal inaccuracy, and most floating point
+         database types don't have a notion of "scale", so by default the
+         float type looks for the first ten decimal places when converting.
+         Specifying this value will override that length.  Types which
+         do include an explicit ".scale" value, such as the base
+         :class:`.Numeric` as well as the MySQL float types, will use the
+         value of ".scale" as the default for decimal_return_scale, if not
+         otherwise specified.
+
+        When using the ``Numeric`` type, care should be taken to ensure
+        that the asdecimal setting is appropriate for the DBAPI in use -
+        when Numeric applies a conversion from Decimal->float or float->
+        Decimal, this conversion incurs an additional performance overhead
+        for all result columns received.
+
+        DBAPIs that return Decimal natively (e.g. psycopg2) will have
+        better accuracy and higher performance with a setting of ``True``,
+        as the native translation to Decimal reduces the amount of floating-
+        point issues at play, and the Numeric type itself doesn't need
+        to apply any further conversions.  However, another DBAPI which
+        returns floats natively *will* incur an additional conversion
+        overhead, and is still subject to floating point data loss - in
+        which case ``asdecimal=False`` will at least remove the extra
+        conversion overhead.
+
+        """
+        self.precision = precision
+        self.scale = scale
+        self.decimal_return_scale = decimal_return_scale
+        self.asdecimal = asdecimal
+
+    @property
+    def _effective_decimal_return_scale(self):
+        if self.decimal_return_scale is not None:
+            return self.decimal_return_scale
+        elif getattr(self, "scale", None) is not None:
+            return self.scale
+        else:
+            return self._default_decimal_return_scale
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.NUMBER
+
+    def literal_processor(self, dialect):
+        def process(value):
+            return str(value)
+
+        return process
+
+    @property
+    def python_type(self):
+        if self.asdecimal:
+            return decimal.Decimal
+        else:
+            return float
+
+    def bind_processor(self, dialect):
+        if dialect.supports_native_decimal:
+            return None
+        else:
+            return processors.to_float
+
+    def result_processor(self, dialect, coltype):
+        if self.asdecimal:
+            if dialect.supports_native_decimal:
+                # we're a "numeric", DBAPI will give us Decimal directly
+                return None
+            else:
+                # we're a "numeric", DBAPI returns floats, convert.
+                return processors.to_decimal_processor_factory(
+                    decimal.Decimal,
+                    (
+                        self.scale
+                        if self.scale is not None
+                        else self._default_decimal_return_scale
+                    ),
+                )
+        else:
+            if dialect.supports_native_decimal:
+                return processors.to_float
+            else:
+                return None
+
+    @util.memoized_property
+    def _expression_adaptations(self):
+        return {
+            operators.mul: {
+                Interval: Interval,
+                Numeric: self.__class__,
+                Integer: self.__class__,
+            },
+            operators.truediv: {
+                Numeric: self.__class__,
+                Integer: self.__class__,
+            },
+            operators.add: {Numeric: self.__class__, Integer: self.__class__},
+            operators.sub: {Numeric: self.__class__, Integer: self.__class__},
+        }
+
+
+class Float(Numeric[_N]):
+    """Type representing floating point types, such as ``FLOAT`` or ``REAL``.
+
+    This type returns Python ``float`` objects by default, unless the
+    :paramref:`.Float.asdecimal` flag is set to ``True``, in which case they
+    are coerced to ``decimal.Decimal`` objects.
+
+    When a :paramref:`.Float.precision` is not provided in a
+    :class:`_types.Float` type some backend may compile this type as
+    an 8 bytes / 64 bit float datatype. To use a 4 bytes / 32 bit float
+    datatype a precision <= 24 can usually be provided or the
+    :class:`_types.REAL` type can be used.
+    This is known to be the case in the PostgreSQL and MSSQL dialects
+    that render the type as ``FLOAT`` that's in both an alias of
+    ``DOUBLE PRECISION``. Other third party dialects may have similar
+    behavior.
+    """
+
+    __visit_name__ = "float"
+
+    scale = None
+
+    @overload
+    def __init__(
+        self: Float[float],
+        precision: Optional[int] = ...,
+        asdecimal: Literal[False] = ...,
+        decimal_return_scale: Optional[int] = ...,
+    ): ...
+
+    @overload
+    def __init__(
+        self: Float[decimal.Decimal],
+        precision: Optional[int] = ...,
+        asdecimal: Literal[True] = ...,
+        decimal_return_scale: Optional[int] = ...,
+    ): ...
+
+    def __init__(
+        self: Float[_N],
+        precision: Optional[int] = None,
+        asdecimal: bool = False,
+        decimal_return_scale: Optional[int] = None,
+    ):
+        r"""
+        Construct a Float.
+
+        :param precision: the numeric precision for use in DDL ``CREATE
+           TABLE``. Backends **should** attempt to ensure this precision
+           indicates a number of digits for the generic
+           :class:`_sqltypes.Float` datatype.
+
+           .. note:: For the Oracle Database backend, the
+              :paramref:`_sqltypes.Float.precision` parameter is not accepted
+              when rendering DDL, as Oracle Database does not support float precision
+              specified as a number of decimal places. Instead, use the
+              Oracle Database-specific :class:`_oracle.FLOAT` datatype and specify the
+              :paramref:`_oracle.FLOAT.binary_precision` parameter. This is new
+              in version 2.0 of SQLAlchemy.
+
+              To create a database agnostic :class:`_types.Float` that
+              separately specifies binary precision for Oracle Database, use
+              :meth:`_types.TypeEngine.with_variant` as follows::
+
+                    from sqlalchemy import Column
+                    from sqlalchemy import Float
+                    from sqlalchemy.dialects import oracle
+
+                    Column(
+                        "float_data",
+                        Float(5).with_variant(oracle.FLOAT(binary_precision=16), "oracle"),
+                    )
+
+        :param asdecimal: the same flag as that of :class:`.Numeric`, but
+          defaults to ``False``.   Note that setting this flag to ``True``
+          results in floating point conversion.
+
+        :param decimal_return_scale: Default scale to use when converting
+         from floats to Python decimals.  Floating point values will typically
+         be much longer due to decimal inaccuracy, and most floating point
+         database types don't have a notion of "scale", so by default the
+         float type looks for the first ten decimal places when converting.
+         Specifying this value will override that length.  Note that the
+         MySQL float types, which do include "scale", will use "scale"
+         as the default for decimal_return_scale, if not otherwise specified.
+
+        """  # noqa: E501
+        self.precision = precision
+        self.asdecimal = asdecimal
+        self.decimal_return_scale = decimal_return_scale
+
+    def result_processor(self, dialect, coltype):
+        if self.asdecimal:
+            return processors.to_decimal_processor_factory(
+                decimal.Decimal, self._effective_decimal_return_scale
+            )
+        elif dialect.supports_native_decimal:
+            return processors.to_float
+        else:
+            return None
+
+
+class Double(Float[_N]):
+    """A type for double ``FLOAT`` floating point types.
+
+    Typically generates a ``DOUBLE`` or ``DOUBLE_PRECISION`` in DDL,
+    and otherwise acts like a normal :class:`.Float` on the Python
+    side.
+
+    .. versionadded:: 2.0
+
+    """
+
+    __visit_name__ = "double"
+
+
+class _RenderISO8601NoT:
+    def _literal_processor_datetime(self, dialect):
+        return self._literal_processor_portion(dialect, None)
+
+    def _literal_processor_date(self, dialect):
+        return self._literal_processor_portion(dialect, 0)
+
+    def _literal_processor_time(self, dialect):
+        return self._literal_processor_portion(dialect, -1)
+
+    def _literal_processor_portion(self, dialect, _portion=None):
+        assert _portion in (None, 0, -1)
+        if _portion is not None:
+
+            def process(value):
+                return f"""'{value.isoformat().split("T")[_portion]}'"""
+
+        else:
+
+            def process(value):
+                return f"""'{value.isoformat().replace("T", " ")}'"""
+
+        return process
+
+
+class DateTime(
+    _RenderISO8601NoT, HasExpressionLookup, TypeEngine[dt.datetime]
+):
+    """A type for ``datetime.datetime()`` objects.
+
+    Date and time types return objects from the Python ``datetime``
+    module.  Most DBAPIs have built in support for the datetime
+    module, with the noted exception of SQLite.  In the case of
+    SQLite, date and time types are stored as strings which are then
+    converted back to datetime objects when rows are returned.
+
+    For the time representation within the datetime type, some
+    backends include additional options, such as timezone support and
+    fractional seconds support.  For fractional seconds, use the
+    dialect-specific datatype, such as :class:`.mysql.TIME`.  For
+    timezone support, use at least the :class:`_types.TIMESTAMP` datatype,
+    if not the dialect-specific datatype object.
+
+    """
+
+    __visit_name__ = "datetime"
+
+    def __init__(self, timezone: bool = False):
+        """Construct a new :class:`.DateTime`.
+
+        :param timezone: boolean.  Indicates that the datetime type should
+         enable timezone support, if available on the
+         **base date/time-holding type only**.   It is recommended
+         to make use of the :class:`_types.TIMESTAMP` datatype directly when
+         using this flag, as some databases include separate generic
+         date/time-holding types distinct from the timezone-capable
+         TIMESTAMP datatype, such as Oracle Database.
+
+
+        """
+        self.timezone = timezone
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.DATETIME
+
+    def _resolve_for_literal(self, value):
+        with_timezone = value.tzinfo is not None
+        if with_timezone and not self.timezone:
+            return DATETIME_TIMEZONE
+        else:
+            return self
+
+    def literal_processor(self, dialect):
+        return self._literal_processor_datetime(dialect)
+
+    @property
+    def python_type(self):
+        return dt.datetime
+
+    @util.memoized_property
+    def _expression_adaptations(self):
+        # Based on
+        # https://www.postgresql.org/docs/current/static/functions-datetime.html.
+
+        return {
+            operators.add: {Interval: self.__class__},
+            operators.sub: {Interval: self.__class__, DateTime: Interval},
+        }
+
+
+class Date(_RenderISO8601NoT, HasExpressionLookup, TypeEngine[dt.date]):
+    """A type for ``datetime.date()`` objects."""
+
+    __visit_name__ = "date"
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.DATETIME
+
+    @property
+    def python_type(self):
+        return dt.date
+
+    def literal_processor(self, dialect):
+        return self._literal_processor_date(dialect)
+
+    @util.memoized_property
+    def _expression_adaptations(self):
+        # Based on
+        # https://www.postgresql.org/docs/current/static/functions-datetime.html.
+
+        return {
+            operators.add: {
+                Integer: self.__class__,
+                Interval: DateTime,
+                Time: DateTime,
+            },
+            operators.sub: {
+                # date - integer = date
+                Integer: self.__class__,
+                # date - date = integer.
+                Date: Integer,
+                Interval: DateTime,
+                # date - datetime = interval,
+                # this one is not in the PG docs
+                # but works
+                DateTime: Interval,
+            },
+        }
+
+
+class Time(_RenderISO8601NoT, HasExpressionLookup, TypeEngine[dt.time]):
+    """A type for ``datetime.time()`` objects."""
+
+    __visit_name__ = "time"
+
+    def __init__(self, timezone: bool = False):
+        self.timezone = timezone
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.DATETIME
+
+    @property
+    def python_type(self):
+        return dt.time
+
+    def _resolve_for_literal(self, value):
+        with_timezone = value.tzinfo is not None
+        if with_timezone and not self.timezone:
+            return TIME_TIMEZONE
+        else:
+            return self
+
+    @util.memoized_property
+    def _expression_adaptations(self):
+        # Based on
+        # https://www.postgresql.org/docs/current/static/functions-datetime.html.
+
+        return {
+            operators.add: {Date: DateTime, Interval: self.__class__},
+            operators.sub: {Time: Interval, Interval: self.__class__},
+        }
+
+    def literal_processor(self, dialect):
+        return self._literal_processor_time(dialect)
+
+
+class _Binary(TypeEngine[bytes]):
+    """Define base behavior for binary types."""
+
+    def __init__(self, length: Optional[int] = None):
+        self.length = length
+
+    @util.ro_memoized_property
+    def _generic_type_affinity(
+        self,
+    ) -> Type[TypeEngine[bytes]]:
+        return LargeBinary
+
+    def literal_processor(self, dialect):
+        def process(value):
+            # TODO: this is useless for real world scenarios; implement
+            # real binary literals
+            value = value.decode(
+                dialect._legacy_binary_type_literal_encoding
+            ).replace("'", "''")
+            return "'%s'" % value
+
+        return process
+
+    @property
+    def python_type(self):
+        return bytes
+
+    # Python 3 - sqlite3 doesn't need the `Binary` conversion
+    # here, though pg8000 does to indicate "bytea"
+    def bind_processor(self, dialect):
+        if dialect.dbapi is None:
+            return None
+
+        DBAPIBinary = dialect.dbapi.Binary
+
+        def process(value):
+            if value is not None:
+                return DBAPIBinary(value)
+            else:
+                return None
+
+        return process
+
+    # Python 3 has native bytes() type
+    # both sqlite3 and pg8000 seem to return it,
+    # psycopg2 as of 2.5 returns 'memoryview'
+    def result_processor(self, dialect, coltype):
+        if dialect.returns_native_bytes:
+            return None
+
+        def process(value):
+            if value is not None:
+                value = bytes(value)
+            return value
+
+        return process
+
+    def coerce_compared_value(self, op, value):
+        """See :meth:`.TypeEngine.coerce_compared_value` for a description."""
+
+        if isinstance(value, str):
+            return self
+        else:
+            return super().coerce_compared_value(op, value)
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.BINARY
+
+
+class LargeBinary(_Binary):
+    """A type for large binary byte data.
+
+    The :class:`.LargeBinary` type corresponds to a large and/or unlengthed
+    binary type for the target platform, such as BLOB on MySQL and BYTEA for
+    PostgreSQL.  It also handles the necessary conversions for the DBAPI.
+
+    """
+
+    __visit_name__ = "large_binary"
+
+    def __init__(self, length: Optional[int] = None):
+        """
+        Construct a LargeBinary type.
+
+        :param length: optional, a length for the column for use in
+          DDL statements, for those binary types that accept a length,
+          such as the MySQL BLOB type.
+
+        """
+        _Binary.__init__(self, length=length)
+
+
+class SchemaType(SchemaEventTarget, TypeEngineMixin):
+    """Add capabilities to a type which allow for schema-level DDL to be
+    associated with a type.
+
+    Supports types that must be explicitly created/dropped (i.e. PG ENUM type)
+    as well as types that are complimented by table or schema level
+    constraints, triggers, and other rules.
+
+    :class:`.SchemaType` classes can also be targets for the
+    :meth:`.DDLEvents.before_parent_attach` and
+    :meth:`.DDLEvents.after_parent_attach` events, where the events fire off
+    surrounding the association of the type object with a parent
+    :class:`_schema.Column`.
+
+    .. seealso::
+
+        :class:`.Enum`
+
+        :class:`.Boolean`
+
+
+    """
+
+    _use_schema_map = True
+
+    name: Optional[str]
+
+    def __init__(
+        self,
+        name: Optional[str] = None,
+        schema: Optional[str] = None,
+        metadata: Optional[MetaData] = None,
+        inherit_schema: bool = False,
+        quote: Optional[bool] = None,
+        _create_events: bool = True,
+        _adapted_from: Optional[SchemaType] = None,
+    ):
+        if name is not None:
+            self.name = quoted_name(name, quote)
+        else:
+            self.name = None
+        self.schema = schema
+        self.metadata = metadata
+        self.inherit_schema = inherit_schema
+        self._create_events = _create_events
+
+        if _create_events and self.metadata:
+            event.listen(
+                self.metadata,
+                "before_create",
+                util.portable_instancemethod(self._on_metadata_create),
+            )
+            event.listen(
+                self.metadata,
+                "after_drop",
+                util.portable_instancemethod(self._on_metadata_drop),
+            )
+
+        if _adapted_from:
+            self.dispatch = self.dispatch._join(_adapted_from.dispatch)
+
+    def _set_parent(self, parent, **kw):
+        # set parent hook is when this type is associated with a column.
+        # Column calls it for all SchemaEventTarget instances, either the
+        # base type and/or variants in _variant_mapping.
+
+        # we want to register a second hook to trigger when that column is
+        # associated with a table.  in that event, we and all of our variants
+        # may want to set up some state on the table such as a CheckConstraint
+        # that will conditionally render at DDL render time.
+
+        # the base SchemaType also sets up events for
+        # on_table/metadata_create/drop in this method, which is used by
+        # "native" types with a separate CREATE/DROP e.g. Postgresql.ENUM
+
+        parent._on_table_attach(util.portable_instancemethod(self._set_table))
+
+    def _variant_mapping_for_set_table(self, column):
+        if column.type._variant_mapping:
+            variant_mapping = dict(column.type._variant_mapping)
+            variant_mapping["_default"] = column.type
+        else:
+            variant_mapping = None
+        return variant_mapping
+
+    def _set_table(self, column, table):
+        if self.inherit_schema:
+            self.schema = table.schema
+        elif self.metadata and self.schema is None and self.metadata.schema:
+            self.schema = self.metadata.schema
+
+        if not self._create_events:
+            return
+
+        variant_mapping = self._variant_mapping_for_set_table(column)
+
+        event.listen(
+            table,
+            "before_create",
+            util.portable_instancemethod(
+                self._on_table_create, {"variant_mapping": variant_mapping}
+            ),
+        )
+        event.listen(
+            table,
+            "after_drop",
+            util.portable_instancemethod(
+                self._on_table_drop, {"variant_mapping": variant_mapping}
+            ),
+        )
+        if self.metadata is None:
+            # if SchemaType were created w/ a metadata argument, these
+            # events would already have been associated with that metadata
+            # and would preclude an association with table.metadata
+            event.listen(
+                table.metadata,
+                "before_create",
+                util.portable_instancemethod(
+                    self._on_metadata_create,
+                    {"variant_mapping": variant_mapping},
+                ),
+            )
+            event.listen(
+                table.metadata,
+                "after_drop",
+                util.portable_instancemethod(
+                    self._on_metadata_drop,
+                    {"variant_mapping": variant_mapping},
+                ),
+            )
+
+    def copy(self, **kw):
+        return self.adapt(
+            cast("Type[TypeEngine[Any]]", self.__class__),
+            _create_events=True,
+            metadata=(
+                kw.get("_to_metadata", self.metadata)
+                if self.metadata is not None
+                else None
+            ),
+        )
+
+    @overload
+    def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ...
+
+    @overload
+    def adapt(
+        self, cls: Type[TypeEngineMixin], **kw: Any
+    ) -> TypeEngine[Any]: ...
+
+    def adapt(
+        self, cls: Type[Union[TypeEngine[Any], TypeEngineMixin]], **kw: Any
+    ) -> TypeEngine[Any]:
+        kw.setdefault("_create_events", False)
+        kw.setdefault("_adapted_from", self)
+        return super().adapt(cls, **kw)
+
+    def create(self, bind, checkfirst=False):
+        """Issue CREATE DDL for this type, if applicable."""
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t.create(bind, checkfirst=checkfirst)
+
+    def drop(self, bind, checkfirst=False):
+        """Issue DROP DDL for this type, if applicable."""
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t.drop(bind, checkfirst=checkfirst)
+
+    def _on_table_create(self, target, bind, **kw):
+        if not self._is_impl_for_variant(bind.dialect, kw):
+            return
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t._on_table_create(target, bind, **kw)
+
+    def _on_table_drop(self, target, bind, **kw):
+        if not self._is_impl_for_variant(bind.dialect, kw):
+            return
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t._on_table_drop(target, bind, **kw)
+
+    def _on_metadata_create(self, target, bind, **kw):
+        if not self._is_impl_for_variant(bind.dialect, kw):
+            return
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t._on_metadata_create(target, bind, **kw)
+
+    def _on_metadata_drop(self, target, bind, **kw):
+        if not self._is_impl_for_variant(bind.dialect, kw):
+            return
+
+        t = self.dialect_impl(bind.dialect)
+        if isinstance(t, SchemaType) and t.__class__ is not self.__class__:
+            t._on_metadata_drop(target, bind, **kw)
+
+    def _is_impl_for_variant(self, dialect, kw):
+        variant_mapping = kw.pop("variant_mapping", None)
+
+        if not variant_mapping:
+            return True
+
+        # for types that have _variant_mapping, all the impls in the map
+        # that are SchemaEventTarget subclasses get set up as event holders.
+        # this is so that constructs that need
+        # to be associated with the Table at dialect-agnostic time etc. like
+        # CheckConstraints can be set up with that table.  they then add
+        # to these constraints a DDL check_rule that among other things
+        # will check this _is_impl_for_variant() method to determine when
+        # the dialect is known that we are part of the table's DDL sequence.
+
+        # since PostgreSQL is the only DB that has ARRAY this can only
+        # be integration tested by PG-specific tests
+        def _we_are_the_impl(typ):
+            return (
+                typ is self
+                or isinstance(typ, ARRAY)
+                and typ.item_type is self  # type: ignore[comparison-overlap]
+            )
+
+        if dialect.name in variant_mapping and _we_are_the_impl(
+            variant_mapping[dialect.name]
+        ):
+            return True
+        elif dialect.name not in variant_mapping:
+            return _we_are_the_impl(variant_mapping["_default"])
+
+
+class Enum(String, SchemaType, Emulated, TypeEngine[Union[str, enum.Enum]]):
+    """Generic Enum Type.
+
+    The :class:`.Enum` type provides a set of possible string values
+    which the column is constrained towards.
+
+    The :class:`.Enum` type will make use of the backend's native "ENUM"
+    type if one is available; otherwise, it uses a VARCHAR datatype.
+    An option also exists to automatically produce a CHECK constraint
+    when the VARCHAR (so called "non-native") variant is produced;
+    see the  :paramref:`.Enum.create_constraint` flag.
+
+    The :class:`.Enum` type also provides in-Python validation of string
+    values during both read and write operations.  When reading a value
+    from the database in a result set, the string value is always checked
+    against the list of possible values and a ``LookupError`` is raised
+    if no match is found.  When passing a value to the database as a
+    plain string within a SQL statement, if the
+    :paramref:`.Enum.validate_strings` parameter is
+    set to True, a ``LookupError`` is raised for any string value that's
+    not located in the given list of possible values; note that this
+    impacts usage of LIKE expressions with enumerated values (an unusual
+    use case).
+
+    The source of enumerated values may be a list of string values, or
+    alternatively a PEP-435-compliant enumerated class.  For the purposes
+    of the :class:`.Enum` datatype, this class need only provide a
+    ``__members__`` method.
+
+    When using an enumerated class, the enumerated objects are used
+    both for input and output, rather than strings as is the case with
+    a plain-string enumerated type::
+
+        import enum
+        from sqlalchemy import Enum
+
+
+        class MyEnum(enum.Enum):
+            one = 1
+            two = 2
+            three = 3
+
+
+        t = Table("data", MetaData(), Column("value", Enum(MyEnum)))
+
+        connection.execute(t.insert(), {"value": MyEnum.two})
+        assert connection.scalar(t.select()) is MyEnum.two
+
+    Above, the string names of each element, e.g. "one", "two", "three",
+    are persisted to the database; the values of the Python Enum, here
+    indicated as integers, are **not** used; the value of each enum can
+    therefore be any kind of Python object whether or not it is persistable.
+
+    In order to persist the values and not the names, the
+    :paramref:`.Enum.values_callable` parameter may be used.   The value of
+    this parameter is a user-supplied callable, which  is intended to be used
+    with a PEP-435-compliant enumerated class and  returns a list of string
+    values to be persisted.   For a simple enumeration that uses string values,
+    a callable such as  ``lambda x: [e.value for e in x]`` is sufficient.
+
+    .. seealso::
+
+        :ref:`orm_declarative_mapped_column_enums` - background on using
+        the :class:`_sqltypes.Enum` datatype with the ORM's
+        :ref:`ORM Annotated Declarative <orm_declarative_mapped_column>`
+        feature.
+
+        :class:`_postgresql.ENUM` - PostgreSQL-specific type,
+        which has additional functionality.
+
+        :class:`.mysql.ENUM` - MySQL-specific type
+
+    """
+
+    __visit_name__ = "enum"
+
+    def __init__(self, *enums: object, **kw: Any):
+        r"""Construct an enum.
+
+        Keyword arguments which don't apply to a specific backend are ignored
+        by that backend.
+
+        :param \*enums: either exactly one PEP-435 compliant enumerated type
+           or one or more string labels.
+
+        :param create_constraint: defaults to False.  When creating a
+           non-native enumerated type, also build a CHECK constraint on the
+           database against the valid values.
+
+           .. note:: it is strongly recommended that the CHECK constraint
+              have an explicit name in order to support schema-management
+              concerns.  This can be established either by setting the
+              :paramref:`.Enum.name` parameter or by setting up an
+              appropriate naming convention; see
+              :ref:`constraint_naming_conventions` for background.
+
+           .. versionchanged:: 1.4 - this flag now defaults to False, meaning
+              no CHECK constraint is generated for a non-native enumerated
+              type.
+
+        :param metadata: Associate this type directly with a ``MetaData``
+           object. For types that exist on the target database as an
+           independent schema construct (PostgreSQL), this type will be
+           created and dropped within ``create_all()`` and ``drop_all()``
+           operations. If the type is not associated with any ``MetaData``
+           object, it will associate itself with each ``Table`` in which it is
+           used, and will be created when any of those individual tables are
+           created, after a check is performed for its existence. The type is
+           only dropped when ``drop_all()`` is called for that ``Table``
+           object's metadata, however.
+
+           The value of the :paramref:`_schema.MetaData.schema` parameter of
+           the :class:`_schema.MetaData` object, if set, will be used as the
+           default value of the :paramref:`_types.Enum.schema` on this object
+           if an explicit value is not otherwise supplied.
+
+           .. versionchanged:: 1.4.12 :class:`_types.Enum` inherits the
+              :paramref:`_schema.MetaData.schema` parameter of the
+              :class:`_schema.MetaData` object if present, when passed using
+              the :paramref:`_types.Enum.metadata` parameter.
+
+        :param name: The name of this type. This is required for PostgreSQL
+           and any future supported database which requires an explicitly
+           named type, or an explicitly named constraint in order to generate
+           the type and/or a table that uses it. If a PEP-435 enumerated
+           class was used, its name (converted to lower case) is used by
+           default.
+
+        :param native_enum: Use the database's native ENUM type when
+           available. Defaults to True. When False, uses VARCHAR + check
+           constraint for all backends. When False, the VARCHAR length can be
+           controlled with :paramref:`.Enum.length`; currently "length" is
+           ignored if native_enum=True.
+
+        :param length: Allows specifying a custom length for the VARCHAR
+           when a non-native enumeration datatype is used.  By default it uses
+           the length of the longest value.
+
+           .. versionchanged:: 2.0.0 The :paramref:`.Enum.length` parameter
+              is used unconditionally for ``VARCHAR`` rendering regardless of
+              the :paramref:`.Enum.native_enum` parameter, for those backends
+              where ``VARCHAR`` is used for enumerated datatypes.
+
+
+        :param schema: Schema name of this type. For types that exist on the
+           target database as an independent schema construct (PostgreSQL),
+           this parameter specifies the named schema in which the type is
+           present.
+
+           If not present, the schema name will be taken from the
+           :class:`_schema.MetaData` collection if passed as
+           :paramref:`_types.Enum.metadata`, for a :class:`_schema.MetaData`
+           that includes the :paramref:`_schema.MetaData.schema` parameter.
+
+           .. versionchanged:: 1.4.12 :class:`_types.Enum` inherits the
+              :paramref:`_schema.MetaData.schema` parameter of the
+              :class:`_schema.MetaData` object if present, when passed using
+              the :paramref:`_types.Enum.metadata` parameter.
+
+           Otherwise, if the :paramref:`_types.Enum.inherit_schema` flag is set
+           to ``True``, the schema will be inherited from the associated
+           :class:`_schema.Table` object if any; when
+           :paramref:`_types.Enum.inherit_schema` is at its default of
+           ``False``, the owning table's schema is **not** used.
+
+
+        :param quote: Set explicit quoting preferences for the type's name.
+
+        :param inherit_schema: When ``True``, the "schema" from the owning
+           :class:`_schema.Table`
+           will be copied to the "schema" attribute of this
+           :class:`.Enum`, replacing whatever value was passed for the
+           ``schema`` attribute.   This also takes effect when using the
+           :meth:`_schema.Table.to_metadata` operation.
+
+        :param validate_strings: when True, string values that are being
+           passed to the database in a SQL statement will be checked
+           for validity against the list of enumerated values.  Unrecognized
+           values will result in a ``LookupError`` being raised.
+
+        :param values_callable: A callable which will be passed the PEP-435
+           compliant enumerated type, which should then return a list of string
+           values to be persisted. This allows for alternate usages such as
+           using the string value of an enum to be persisted to the database
+           instead of its name. The callable must return the values to be
+           persisted in the same order as iterating through the Enum's
+           ``__member__`` attribute. For example
+           ``lambda x: [i.value for i in x]``.
+
+           .. versionadded:: 1.2.3
+
+        :param sort_key_function: a Python callable which may be used as the
+           "key" argument in the Python ``sorted()`` built-in.   The SQLAlchemy
+           ORM requires that primary key columns which are mapped must
+           be sortable in some way.  When using an unsortable enumeration
+           object such as a Python 3 ``Enum`` object, this parameter may be
+           used to set a default sort key function for the objects.  By
+           default, the database value of the enumeration is used as the
+           sorting function.
+
+           .. versionadded:: 1.3.8
+
+        :param omit_aliases: A boolean that when true will remove aliases from
+           pep 435 enums. defaults to ``True``.
+
+           .. versionchanged:: 2.0 This parameter now defaults to True.
+
+        """
+        self._enum_init(enums, kw)
+
+    @property
+    def _enums_argument(self):
+        if self.enum_class is not None:
+            return [self.enum_class]
+        else:
+            return self.enums
+
+    def _enum_init(self, enums, kw):
+        """internal init for :class:`.Enum` and subclasses.
+
+        friendly init helper used by subclasses to remove
+        all the Enum-specific keyword arguments from kw.  Allows all
+        other arguments in kw to pass through.
+
+        """
+        self.native_enum = kw.pop("native_enum", True)
+        self.create_constraint = kw.pop("create_constraint", False)
+        self.values_callable = kw.pop("values_callable", None)
+        self._sort_key_function = kw.pop("sort_key_function", NO_ARG)
+        length_arg = kw.pop("length", NO_ARG)
+        self._omit_aliases = kw.pop("omit_aliases", True)
+        _disable_warnings = kw.pop("_disable_warnings", False)
+        values, objects = self._parse_into_values(enums, kw)
+        self._setup_for_values(values, objects, kw)
+
+        self.validate_strings = kw.pop("validate_strings", False)
+
+        if self.enums:
+            self._default_length = length = max(len(x) for x in self.enums)
+        else:
+            self._default_length = length = 0
+
+        if length_arg is not NO_ARG:
+            if (
+                not _disable_warnings
+                and length_arg is not None
+                and length_arg < length
+            ):
+                raise ValueError(
+                    "When provided, length must be larger or equal"
+                    " than the length of the longest enum value. %s < %s"
+                    % (length_arg, length)
+                )
+            length = length_arg
+
+        self._valid_lookup[None] = self._object_lookup[None] = None
+
+        super().__init__(length=length)
+
+        # assign name to the given enum class if no other name, and this
+        # enum is not an "empty" enum.  if the enum is "empty" we assume
+        # this is a template enum that will be used to generate
+        # new Enum classes.
+        if self.enum_class and values:
+            kw.setdefault("name", self.enum_class.__name__.lower())
+        SchemaType.__init__(
+            self,
+            name=kw.pop("name", None),
+            schema=kw.pop("schema", None),
+            metadata=kw.pop("metadata", None),
+            inherit_schema=kw.pop("inherit_schema", False),
+            quote=kw.pop("quote", None),
+            _create_events=kw.pop("_create_events", True),
+            _adapted_from=kw.pop("_adapted_from", None),
+        )
+
+    def _parse_into_values(self, enums, kw):
+        if not enums and "_enums" in kw:
+            enums = kw.pop("_enums")
+
+        if len(enums) == 1 and hasattr(enums[0], "__members__"):
+            self.enum_class = enums[0]
+
+            _members = self.enum_class.__members__
+
+            if self._omit_aliases is True:
+                # remove aliases
+                members = OrderedDict(
+                    (n, v) for n, v in _members.items() if v.name == n
+                )
+            else:
+                members = _members
+            if self.values_callable:
+                values = self.values_callable(self.enum_class)
+            else:
+                values = list(members)
+            objects = [members[k] for k in members]
+            return values, objects
+        else:
+            self.enum_class = None
+            return enums, enums
+
+    def _resolve_for_literal(self, value: Any) -> Enum:
+        tv = type(value)
+        typ = self._resolve_for_python_type(tv, tv, tv)
+        assert typ is not None
+        return typ
+
+    def _resolve_for_python_type(
+        self,
+        python_type: Type[Any],
+        matched_on: _MatchedOnType,
+        matched_on_flattened: Type[Any],
+    ) -> Optional[Enum]:
+        # "generic form" indicates we were placed in a type map
+        # as ``sqlalchemy.Enum(enum.Enum)`` which indicates we need to
+        # get enumerated values from the datatype
+        we_are_generic_form = self._enums_argument == [enum.Enum]
+
+        native_enum = None
+
+        def process_literal(pt):
+            # for a literal, where we need to get its contents, parse it out.
+            enum_args = get_args(pt)
+            bad_args = [arg for arg in enum_args if not isinstance(arg, str)]
+            if bad_args:
+                raise exc.ArgumentError(
+                    f"Can't create string-based Enum datatype from non-string "
+                    f"values: {', '.join(repr(x) for x in bad_args)}.  Please "
+                    f"provide an explicit Enum datatype for this Python type"
+                )
+            native_enum = False
+            return enum_args, native_enum
+
+        if not we_are_generic_form and python_type is matched_on:
+            # if we have enumerated values, and the incoming python
+            # type is exactly the one that matched in the type map,
+            # then we use these enumerated values and dont try to parse
+            # what's incoming
+            enum_args = self._enums_argument
+
+        elif is_literal(python_type):
+            enum_args, native_enum = process_literal(python_type)
+        elif is_pep695(python_type):
+            value = python_type.__value__
+            if is_pep695(value):
+                new_value = value
+                while is_pep695(new_value):
+                    new_value = new_value.__value__
+                if is_literal(new_value):
+                    value = new_value
+                    warn_deprecated(
+                        f"Mapping recursive TypeAliasType '{python_type}' "
+                        "that resolve to literal to generate an Enum is "
+                        "deprecated. SQLAlchemy 2.1 will not support this "
+                        "use case. Please avoid using recursing "
+                        "TypeAliasType.",
+                        "2.0",
+                    )
+            if not is_literal(value):
+                raise exc.ArgumentError(
+                    f"Can't associate TypeAliasType '{python_type}' to an "
+                    "Enum since it's not a direct alias of a Literal. Only "
+                    "aliases in this form `type my_alias = Literal['a', "
+                    "'b']` are supported when generating Enums."
+                )
+            enum_args, native_enum = process_literal(value)
+
+        elif isinstance(python_type, type) and issubclass(
+            python_type, enum.Enum
+        ):
+            # same for an enum.Enum
+            enum_args = [python_type]
+
+        else:
+            enum_args = self._enums_argument
+
+        # make a new Enum that looks like this one.
+        # arguments or other rules
+        kw = self._make_enum_kw({})
+
+        if native_enum is False:
+            kw["native_enum"] = False
+
+        kw["length"] = NO_ARG if self.length == 0 else self.length
+        return cast(
+            Enum,
+            self._generic_type_affinity(_enums=enum_args, **kw),  # type: ignore  # noqa: E501
+        )
+
+    def _setup_for_values(self, values, objects, kw):
+        self.enums = list(values)
+
+        self._valid_lookup = dict(zip(reversed(objects), reversed(values)))
+
+        self._object_lookup = dict(zip(values, objects))
+
+        self._valid_lookup.update(
+            [
+                (value, self._valid_lookup[self._object_lookup[value]])
+                for value in values
+            ]
+        )
+
+    @property
+    def sort_key_function(self):
+        if self._sort_key_function is NO_ARG:
+            return self._db_value_for_elem
+        else:
+            return self._sort_key_function
+
+    @property
+    def native(self):
+        return self.native_enum
+
+    def _db_value_for_elem(self, elem):
+        try:
+            return self._valid_lookup[elem]
+        except KeyError as err:
+            # for unknown string values, we return as is.  While we can
+            # validate these if we wanted, that does not allow for lesser-used
+            # end-user use cases, such as using a LIKE comparison with an enum,
+            # or for an application that wishes to apply string tests to an
+            # ENUM (see [ticket:3725]).  While we can decide to differentiate
+            # here between an INSERT statement and a criteria used in a SELECT,
+            # for now we're staying conservative w/ behavioral changes (perhaps
+            # someone has a trigger that handles strings on INSERT)
+            if not self.validate_strings and isinstance(elem, str):
+                return elem
+            else:
+                raise LookupError(
+                    "'%s' is not among the defined enum values. "
+                    "Enum name: %s. Possible values: %s"
+                    % (
+                        elem,
+                        self.name,
+                        langhelpers.repr_tuple_names(self.enums),
+                    )
+                ) from err
+
+    class Comparator(String.Comparator[str]):
+        __slots__ = ()
+
+        type: String
+
+        def _adapt_expression(
+            self,
+            op: OperatorType,
+            other_comparator: TypeEngine.Comparator[Any],
+        ) -> Tuple[OperatorType, TypeEngine[Any]]:
+            op, typ = super()._adapt_expression(op, other_comparator)
+            if op is operators.concat_op:
+                typ = String(self.type.length)
+            return op, typ
+
+    comparator_factory = Comparator
+
+    def _object_value_for_elem(self, elem):
+        try:
+            return self._object_lookup[elem]
+        except KeyError as err:
+            raise LookupError(
+                "'%s' is not among the defined enum values. "
+                "Enum name: %s. Possible values: %s"
+                % (
+                    elem,
+                    self.name,
+                    langhelpers.repr_tuple_names(self.enums),
+                )
+            ) from err
+
+    def __repr__(self):
+        return util.generic_repr(
+            self,
+            additional_kw=[
+                ("native_enum", True),
+                ("create_constraint", False),
+                ("length", self._default_length),
+            ],
+            to_inspect=[Enum, SchemaType],
+        )
+
+    def as_generic(self, allow_nulltype=False):
+        try:
+            args = self.enums
+        except AttributeError:
+            raise NotImplementedError(
+                "TypeEngine.as_generic() heuristic "
+                "is undefined for types that inherit Enum but do not have "
+                "an `enums` attribute."
+            ) from None
+
+        return util.constructor_copy(
+            self, self._generic_type_affinity, *args, _disable_warnings=True
+        )
+
+    def _make_enum_kw(self, kw):
+        kw.setdefault("validate_strings", self.validate_strings)
+        if self.name:
+            kw.setdefault("name", self.name)
+        kw.setdefault("schema", self.schema)
+        kw.setdefault("inherit_schema", self.inherit_schema)
+        kw.setdefault("metadata", self.metadata)
+        kw.setdefault("native_enum", self.native_enum)
+        kw.setdefault("values_callable", self.values_callable)
+        kw.setdefault("create_constraint", self.create_constraint)
+        kw.setdefault("length", self.length)
+        kw.setdefault("omit_aliases", self._omit_aliases)
+        return kw
+
+    def adapt_to_emulated(self, impltype, **kw):
+        self._make_enum_kw(kw)
+        kw["_disable_warnings"] = True
+        kw.setdefault("_create_events", False)
+        assert "_enums" in kw
+        return impltype(**kw)
+
+    def adapt(self, cls, **kw):
+        kw["_enums"] = self._enums_argument
+        kw["_disable_warnings"] = True
+        return super().adapt(cls, **kw)
+
+    def _should_create_constraint(self, compiler, **kw):
+        if not self._is_impl_for_variant(compiler.dialect, kw):
+            return False
+        return (
+            not self.native_enum or not compiler.dialect.supports_native_enum
+        )
+
+    @util.preload_module("sqlalchemy.sql.schema")
+    def _set_table(self, column, table):
+        schema = util.preloaded.sql_schema
+        SchemaType._set_table(self, column, table)
+
+        if not self.create_constraint:
+            return
+
+        variant_mapping = self._variant_mapping_for_set_table(column)
+
+        e = schema.CheckConstraint(
+            type_coerce(column, String()).in_(self.enums),
+            name=_NONE_NAME if self.name is None else self.name,
+            _create_rule=util.portable_instancemethod(
+                self._should_create_constraint,
+                {"variant_mapping": variant_mapping},
+            ),
+            _type_bound=True,
+        )
+        assert e.table is table
+
+    def literal_processor(self, dialect):
+        parent_processor = super().literal_processor(dialect)
+
+        def process(value):
+            value = self._db_value_for_elem(value)
+            if parent_processor:
+                value = parent_processor(value)
+            return value
+
+        return process
+
+    def bind_processor(self, dialect):
+        parent_processor = super().bind_processor(dialect)
+
+        def process(value):
+            value = self._db_value_for_elem(value)
+            if parent_processor:
+                value = parent_processor(value)
+            return value
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        parent_processor = super().result_processor(dialect, coltype)
+
+        def process(value):
+            if parent_processor:
+                value = parent_processor(value)
+
+            value = self._object_value_for_elem(value)
+            return value
+
+        return process
+
+    def copy(self, **kw):
+        return SchemaType.copy(self, **kw)
+
+    @property
+    def python_type(self):
+        if self.enum_class:
+            return self.enum_class
+        else:
+            return super().python_type
+
+
+class PickleType(TypeDecorator[object]):
+    """Holds Python objects, which are serialized using pickle.
+
+    PickleType builds upon the Binary type to apply Python's
+    ``pickle.dumps()`` to incoming objects, and ``pickle.loads()`` on
+    the way out, allowing any pickleable Python object to be stored as
+    a serialized binary field.
+
+    To allow ORM change events to propagate for elements associated
+    with :class:`.PickleType`, see :ref:`mutable_toplevel`.
+
+    """
+
+    impl = LargeBinary
+    cache_ok = True
+
+    def __init__(
+        self,
+        protocol: int = pickle.HIGHEST_PROTOCOL,
+        pickler: Any = None,
+        comparator: Optional[Callable[[Any, Any], bool]] = None,
+        impl: Optional[_TypeEngineArgument[Any]] = None,
+    ):
+        """
+        Construct a PickleType.
+
+        :param protocol: defaults to ``pickle.HIGHEST_PROTOCOL``.
+
+        :param pickler: defaults to pickle.  May be any object with
+          pickle-compatible ``dumps`` and ``loads`` methods.
+
+        :param comparator: a 2-arg callable predicate used
+          to compare values of this type.  If left as ``None``,
+          the Python "equals" operator is used to compare values.
+
+        :param impl: A binary-storing :class:`_types.TypeEngine` class or
+          instance to use in place of the default :class:`_types.LargeBinary`.
+          For example the :class: `_mysql.LONGBLOB` class may be more effective
+          when using MySQL.
+
+          .. versionadded:: 1.4.20
+
+        """
+        self.protocol = protocol
+        self.pickler = pickler or pickle
+        self.comparator = comparator
+        super().__init__()
+
+        if impl:
+            # custom impl is not necessarily a LargeBinary subclass.
+            # make an exception to typing for this
+            self.impl = to_instance(impl)  # type: ignore
+
+    def __reduce__(self):
+        return PickleType, (self.protocol, None, self.comparator)
+
+    def bind_processor(self, dialect):
+        impl_processor = self.impl_instance.bind_processor(dialect)
+        dumps = self.pickler.dumps
+        protocol = self.protocol
+        if impl_processor:
+            fixed_impl_processor = impl_processor
+
+            def process(value):
+                if value is not None:
+                    value = dumps(value, protocol)
+                return fixed_impl_processor(value)
+
+        else:
+
+            def process(value):
+                if value is not None:
+                    value = dumps(value, protocol)
+                return value
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        impl_processor = self.impl_instance.result_processor(dialect, coltype)
+        loads = self.pickler.loads
+        if impl_processor:
+            fixed_impl_processor = impl_processor
+
+            def process(value):
+                value = fixed_impl_processor(value)
+                if value is None:
+                    return None
+                return loads(value)
+
+        else:
+
+            def process(value):
+                if value is None:
+                    return None
+                return loads(value)
+
+        return process
+
+    def compare_values(self, x, y):
+        if self.comparator:
+            return self.comparator(x, y)
+        else:
+            return x == y
+
+
+class Boolean(SchemaType, Emulated, TypeEngine[bool]):
+    """A bool datatype.
+
+    :class:`.Boolean` typically uses BOOLEAN or SMALLINT on the DDL side,
+    and on the Python side deals in ``True`` or ``False``.
+
+    The :class:`.Boolean` datatype currently has two levels of assertion
+    that the values persisted are simple true/false values.  For all
+    backends, only the Python values ``None``, ``True``, ``False``, ``1``
+    or ``0`` are accepted as parameter values.   For those backends that
+    don't support a "native boolean" datatype, an option exists to
+    also create a CHECK constraint on the target column
+
+    .. versionchanged:: 1.2 the :class:`.Boolean` datatype now asserts that
+       incoming Python values are already in pure boolean form.
+
+
+    """
+
+    __visit_name__ = "boolean"
+    native = True
+
+    def __init__(
+        self,
+        create_constraint: bool = False,
+        name: Optional[str] = None,
+        _create_events: bool = True,
+        _adapted_from: Optional[SchemaType] = None,
+    ):
+        """Construct a Boolean.
+
+        :param create_constraint: defaults to False.  If the boolean
+          is generated as an int/smallint, also create a CHECK constraint
+          on the table that ensures 1 or 0 as a value.
+
+          .. note:: it is strongly recommended that the CHECK constraint
+             have an explicit name in order to support schema-management
+             concerns.  This can be established either by setting the
+             :paramref:`.Boolean.name` parameter or by setting up an
+             appropriate naming convention; see
+             :ref:`constraint_naming_conventions` for background.
+
+          .. versionchanged:: 1.4 - this flag now defaults to False, meaning
+             no CHECK constraint is generated for a non-native enumerated
+             type.
+
+        :param name: if a CHECK constraint is generated, specify
+          the name of the constraint.
+
+        """
+        self.create_constraint = create_constraint
+        self.name = name
+        self._create_events = _create_events
+        if _adapted_from:
+            self.dispatch = self.dispatch._join(_adapted_from.dispatch)
+
+    def copy(self, **kw):
+        # override SchemaType.copy() to not include to_metadata logic
+        return self.adapt(
+            cast("Type[TypeEngine[Any]]", self.__class__),
+            _create_events=True,
+        )
+
+    def _should_create_constraint(self, compiler, **kw):
+        if not self._is_impl_for_variant(compiler.dialect, kw):
+            return False
+        return (
+            not compiler.dialect.supports_native_boolean
+            and compiler.dialect.non_native_boolean_check_constraint
+        )
+
+    @util.preload_module("sqlalchemy.sql.schema")
+    def _set_table(self, column, table):
+        schema = util.preloaded.sql_schema
+        if not self.create_constraint:
+            return
+
+        variant_mapping = self._variant_mapping_for_set_table(column)
+
+        e = schema.CheckConstraint(
+            type_coerce(column, self).in_([0, 1]),
+            name=_NONE_NAME if self.name is None else self.name,
+            _create_rule=util.portable_instancemethod(
+                self._should_create_constraint,
+                {"variant_mapping": variant_mapping},
+            ),
+            _type_bound=True,
+        )
+        assert e.table is table
+
+    @property
+    def python_type(self):
+        return bool
+
+    _strict_bools = frozenset([None, True, False])
+
+    def _strict_as_bool(self, value):
+        if value not in self._strict_bools:
+            if not isinstance(value, int):
+                raise TypeError("Not a boolean value: %r" % (value,))
+            else:
+                raise ValueError(
+                    "Value %r is not None, True, or False" % (value,)
+                )
+        return value
+
+    def literal_processor(self, dialect):
+        compiler = dialect.statement_compiler(dialect, None)
+        true = compiler.visit_true(None)
+        false = compiler.visit_false(None)
+
+        def process(value):
+            return true if self._strict_as_bool(value) else false
+
+        return process
+
+    def bind_processor(self, dialect):
+        _strict_as_bool = self._strict_as_bool
+
+        _coerce: Union[Type[bool], Type[int]]
+
+        if dialect.supports_native_boolean:
+            _coerce = bool
+        else:
+            _coerce = int
+
+        def process(value):
+            value = _strict_as_bool(value)
+            if value is not None:
+                value = _coerce(value)
+            return value
+
+        return process
+
+    def result_processor(self, dialect, coltype):
+        if dialect.supports_native_boolean:
+            return None
+        else:
+            return processors.int_to_boolean
+
+
+class _AbstractInterval(HasExpressionLookup, TypeEngine[dt.timedelta]):
+    @util.memoized_property
+    def _expression_adaptations(self):
+        # Based on
+        # https://www.postgresql.org/docs/current/static/functions-datetime.html.
+
+        return {
+            operators.add: {
+                Date: DateTime,
+                Interval: self.__class__,
+                DateTime: DateTime,
+                Time: Time,
+            },
+            operators.sub: {Interval: self.__class__},
+            operators.mul: {Numeric: self.__class__},
+            operators.truediv: {Numeric: self.__class__},
+        }
+
+    @util.ro_non_memoized_property
+    def _type_affinity(self) -> Type[Interval]:
+        return Interval
+
+
+class Interval(Emulated, _AbstractInterval, TypeDecorator[dt.timedelta]):
+    """A type for ``datetime.timedelta()`` objects.
+
+    The Interval type deals with ``datetime.timedelta`` objects.  In PostgreSQL
+    and Oracle Database, the native ``INTERVAL`` type is used; for others, the
+    value is stored as a date which is relative to the "epoch" (Jan. 1, 1970).
+
+    Note that the ``Interval`` type does not currently provide date arithmetic
+    operations on platforms which do not support interval types natively. Such
+    operations usually require transformation of both sides of the expression
+    (such as, conversion of both sides into integer epoch values first) which
+    currently is a manual procedure (such as via
+    :attr:`~sqlalchemy.sql.expression.func`).
+
+    """
+
+    impl = DateTime
+    epoch = dt.datetime.fromtimestamp(0, dt.timezone.utc).replace(tzinfo=None)
+    cache_ok = True
+
+    def __init__(
+        self,
+        native: bool = True,
+        second_precision: Optional[int] = None,
+        day_precision: Optional[int] = None,
+    ):
+        """Construct an Interval object.
+
+        :param native: when True, use the actual
+          INTERVAL type provided by the database, if
+          supported (currently PostgreSQL, Oracle Database).
+          Otherwise, represent the interval data as
+          an epoch value regardless.
+
+        :param second_precision: For native interval types
+          which support a "fractional seconds precision" parameter,
+          i.e. Oracle Database and PostgreSQL
+
+        :param day_precision: for native interval types which
+          support a "day precision" parameter, i.e. Oracle Database.
+
+        """
+        super().__init__()
+        self.native = native
+        self.second_precision = second_precision
+        self.day_precision = day_precision
+
+    class Comparator(
+        TypeDecorator.Comparator[_CT],
+        _AbstractInterval.Comparator[_CT],
+    ):
+        __slots__ = ()
+
+    comparator_factory = Comparator
+
+    @property
+    def python_type(self):
+        return dt.timedelta
+
+    def adapt_to_emulated(self, impltype, **kw):
+        return _AbstractInterval.adapt(self, impltype, **kw)
+
+    def coerce_compared_value(self, op, value):
+        return self.impl_instance.coerce_compared_value(op, value)
+
+    def bind_processor(
+        self, dialect: Dialect
+    ) -> _BindProcessorType[dt.timedelta]:
+        if TYPE_CHECKING:
+            assert isinstance(self.impl_instance, DateTime)
+        impl_processor = self.impl_instance.bind_processor(dialect)
+        epoch = self.epoch
+        if impl_processor:
+            fixed_impl_processor = impl_processor
+
+            def process(
+                value: Optional[dt.timedelta],
+            ) -> Any:
+                if value is not None:
+                    dt_value = epoch + value
+                else:
+                    dt_value = None
+                return fixed_impl_processor(dt_value)
+
+        else:
+
+            def process(
+                value: Optional[dt.timedelta],
+            ) -> Any:
+                if value is not None:
+                    dt_value = epoch + value
+                else:
+                    dt_value = None
+                return dt_value
+
+        return process
+
+    def result_processor(
+        self, dialect: Dialect, coltype: Any
+    ) -> _ResultProcessorType[dt.timedelta]:
+        if TYPE_CHECKING:
+            assert isinstance(self.impl_instance, DateTime)
+        impl_processor = self.impl_instance.result_processor(dialect, coltype)
+        epoch = self.epoch
+        if impl_processor:
+            fixed_impl_processor = impl_processor
+
+            def process(value: Any) -> Optional[dt.timedelta]:
+                dt_value = fixed_impl_processor(value)
+                if dt_value is None:
+                    return None
+                return dt_value - epoch
+
+        else:
+
+            def process(value: Any) -> Optional[dt.timedelta]:
+                if value is None:
+                    return None
+                return value - epoch  # type: ignore
+
+        return process
+
+
+class JSON(Indexable, TypeEngine[Any]):
+    """Represent a SQL JSON type.
+
+    .. note::  :class:`_types.JSON`
+       is provided as a facade for vendor-specific
+       JSON types.  Since it supports JSON SQL operations, it only
+       works on backends that have an actual JSON type, currently:
+
+       * PostgreSQL - see :class:`sqlalchemy.dialects.postgresql.JSON` and
+         :class:`sqlalchemy.dialects.postgresql.JSONB` for backend-specific
+         notes
+
+       * MySQL - see
+         :class:`sqlalchemy.dialects.mysql.JSON` for backend-specific notes
+
+       * SQLite as of version 3.9 - see
+         :class:`sqlalchemy.dialects.sqlite.JSON` for backend-specific notes
+
+       * Microsoft SQL Server 2016 and later - see
+         :class:`sqlalchemy.dialects.mssql.JSON` for backend-specific notes
+
+    :class:`_types.JSON` is part of the Core in support of the growing
+    popularity of native JSON datatypes.
+
+    The :class:`_types.JSON` type stores arbitrary JSON format data, e.g.::
+
+        data_table = Table(
+            "data_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", JSON),
+        )
+
+        with engine.connect() as conn:
+            conn.execute(
+                data_table.insert(), {"data": {"key1": "value1", "key2": "value2"}}
+            )
+
+    **JSON-Specific Expression Operators**
+
+    The :class:`_types.JSON`
+    datatype provides these additional SQL operations:
+
+    * Keyed index operations::
+
+        data_table.c.data["some key"]
+
+    * Integer index operations::
+
+        data_table.c.data[3]
+
+    * Path index operations::
+
+        data_table.c.data[("key_1", "key_2", 5, ..., "key_n")]
+
+    * Data casters for specific JSON element types, subsequent to an index
+      or path operation being invoked::
+
+        data_table.c.data["some key"].as_integer()
+
+      .. versionadded:: 1.3.11
+
+    Additional operations may be available from the dialect-specific versions
+    of :class:`_types.JSON`, such as
+    :class:`sqlalchemy.dialects.postgresql.JSON` and
+    :class:`sqlalchemy.dialects.postgresql.JSONB` which both offer additional
+    PostgreSQL-specific operations.
+
+    **Casting JSON Elements to Other Types**
+
+    Index operations, i.e. those invoked by calling upon the expression using
+    the Python bracket operator as in ``some_column['some key']``, return an
+    expression object whose type defaults to :class:`_types.JSON` by default,
+    so that
+    further JSON-oriented instructions may be called upon the result type.
+    However, it is likely more common that an index operation is expected
+    to return a specific scalar element, such as a string or integer.  In
+    order to provide access to these elements in a backend-agnostic way,
+    a series of data casters are provided:
+
+    * :meth:`.JSON.Comparator.as_string` - return the element as a string
+
+    * :meth:`.JSON.Comparator.as_boolean` - return the element as a boolean
+
+    * :meth:`.JSON.Comparator.as_float` - return the element as a float
+
+    * :meth:`.JSON.Comparator.as_integer` - return the element as an integer
+
+    These data casters are implemented by supporting dialects in order to
+    assure that comparisons to the above types will work as expected, such as::
+
+        # integer comparison
+        data_table.c.data["some_integer_key"].as_integer() == 5
+
+        # boolean comparison
+        data_table.c.data["some_boolean"].as_boolean() == True
+
+    .. versionadded:: 1.3.11 Added type-specific casters for the basic JSON
+       data element types.
+
+    .. note::
+
+        The data caster functions are new in version 1.3.11, and supersede
+        the previous documented approaches of using CAST; for reference,
+        this looked like::
+
+           from sqlalchemy import cast, type_coerce
+           from sqlalchemy import String, JSON
+
+           cast(data_table.c.data["some_key"], String) == type_coerce(55, JSON)
+
+        The above case now works directly as::
+
+            data_table.c.data["some_key"].as_integer() == 5
+
+        For details on the previous comparison approach within the 1.3.x
+        series, see the documentation for SQLAlchemy 1.2 or the included HTML
+        files in the doc/ directory of the version's distribution.
+
+    **Detecting Changes in JSON columns when using the ORM**
+
+    The :class:`_types.JSON` type, when used with the SQLAlchemy ORM, does not
+    detect in-place mutations to the structure.  In order to detect these, the
+    :mod:`sqlalchemy.ext.mutable` extension must be used, most typically
+    using the :class:`.MutableDict` class.  This extension will
+    allow "in-place" changes to the datastructure to produce events which
+    will be detected by the unit of work.  See the example at :class:`.HSTORE`
+    for a simple example involving a dictionary.
+
+    Alternatively, assigning a JSON structure to an ORM element that
+    replaces the old one will always trigger a change event.
+
+    **Support for JSON null vs. SQL NULL**
+
+    When working with NULL values, the :class:`_types.JSON` type recommends the
+    use of two specific constants in order to differentiate between a column
+    that evaluates to SQL NULL, e.g. no value, vs. the JSON-encoded string of
+    ``"null"``. To insert or select against a value that is SQL NULL, use the
+    constant :func:`.null`. This symbol may be passed as a parameter value
+    specifically when using the :class:`_types.JSON` datatype, which contains
+    special logic that interprets this symbol to mean that the column value
+    should be SQL NULL as opposed to JSON ``"null"``::
+
+        from sqlalchemy import null
+
+        conn.execute(table.insert(), {"json_value": null()})
+
+    To insert or select against a value that is JSON ``"null"``, use the
+    constant :attr:`_types.JSON.NULL`::
+
+        conn.execute(table.insert(), {"json_value": JSON.NULL})
+
+    The :class:`_types.JSON` type supports a flag
+    :paramref:`_types.JSON.none_as_null` which when set to True will result
+    in the Python constant ``None`` evaluating to the value of SQL
+    NULL, and when set to False results in the Python constant
+    ``None`` evaluating to the value of JSON ``"null"``.    The Python
+    value ``None`` may be used in conjunction with either
+    :attr:`_types.JSON.NULL` and :func:`.null` in order to indicate NULL
+    values, but care must be taken as to the value of the
+    :paramref:`_types.JSON.none_as_null` in these cases.
+
+    **Customizing the JSON Serializer**
+
+    The JSON serializer and deserializer used by :class:`_types.JSON`
+    defaults to
+    Python's ``json.dumps`` and ``json.loads`` functions; in the case of the
+    psycopg2 dialect, psycopg2 may be using its own custom loader function.
+
+    In order to affect the serializer / deserializer, they are currently
+    configurable at the :func:`_sa.create_engine` level via the
+    :paramref:`_sa.create_engine.json_serializer` and
+    :paramref:`_sa.create_engine.json_deserializer` parameters.  For example,
+    to turn off ``ensure_ascii``::
+
+        engine = create_engine(
+            "sqlite://",
+            json_serializer=lambda obj: json.dumps(obj, ensure_ascii=False),
+        )
+
+    .. versionchanged:: 1.3.7
+
+        SQLite dialect's ``json_serializer`` and ``json_deserializer``
+        parameters renamed from ``_json_serializer`` and
+        ``_json_deserializer``.
+
+    .. seealso::
+
+        :class:`sqlalchemy.dialects.postgresql.JSON`
+
+        :class:`sqlalchemy.dialects.postgresql.JSONB`
+
+        :class:`sqlalchemy.dialects.mysql.JSON`
+
+        :class:`sqlalchemy.dialects.sqlite.JSON`
+
+    """  # noqa: E501
+
+    __visit_name__ = "JSON"
+
+    hashable = False
+    NULL = util.symbol("JSON_NULL")
+    """Describe the json value of NULL.
+
+    This value is used to force the JSON value of ``"null"`` to be
+    used as the value.   A value of Python ``None`` will be recognized
+    either as SQL NULL or JSON ``"null"``, based on the setting
+    of the :paramref:`_types.JSON.none_as_null` flag; the
+    :attr:`_types.JSON.NULL`
+    constant can be used to always resolve to JSON ``"null"`` regardless
+    of this setting.  This is in contrast to the :func:`_expression.null`
+    construct,
+    which always resolves to SQL NULL.  E.g.::
+
+        from sqlalchemy import null
+        from sqlalchemy.dialects.postgresql import JSON
+
+        # will *always* insert SQL NULL
+        obj1 = MyObject(json_value=null())
+
+        # will *always* insert JSON string "null"
+        obj2 = MyObject(json_value=JSON.NULL)
+
+        session.add_all([obj1, obj2])
+        session.commit()
+
+    In order to set JSON NULL as a default value for a column, the most
+    transparent method is to use :func:`_expression.text`::
+
+        Table(
+            "my_table", metadata, Column("json_data", JSON, default=text("'null'"))
+        )
+
+    While it is possible to use :attr:`_types.JSON.NULL` in this context, the
+    :attr:`_types.JSON.NULL` value will be returned as the value of the
+    column,
+    which in the context of the ORM or other repurposing of the default
+    value, may not be desirable.  Using a SQL expression means the value
+    will be re-fetched from the database within the context of retrieving
+    generated defaults.
+
+
+    """  # noqa: E501
+
+    def __init__(self, none_as_null: bool = False):
+        """Construct a :class:`_types.JSON` type.
+
+        :param none_as_null=False: if True, persist the value ``None`` as a
+         SQL NULL value, not the JSON encoding of ``null``. Note that when this
+         flag is False, the :func:`.null` construct can still be used to
+         persist a NULL value, which may be passed directly as a parameter
+         value that is specially interpreted by the :class:`_types.JSON` type
+         as SQL NULL::
+
+             from sqlalchemy import null
+
+             conn.execute(table.insert(), {"data": null()})
+
+         .. note::
+
+              :paramref:`_types.JSON.none_as_null` does **not** apply to the
+              values passed to :paramref:`_schema.Column.default` and
+              :paramref:`_schema.Column.server_default`; a value of ``None``
+              passed for these parameters means "no default present".
+
+              Additionally, when used in SQL comparison expressions, the
+              Python value ``None`` continues to refer to SQL null, and not
+              JSON NULL.  The :paramref:`_types.JSON.none_as_null` flag refers
+              explicitly to the **persistence** of the value within an
+              INSERT or UPDATE statement.   The :attr:`_types.JSON.NULL`
+              value should be used for SQL expressions that wish to compare to
+              JSON null.
+
+         .. seealso::
+
+              :attr:`.types.JSON.NULL`
+
+        """
+        self.none_as_null = none_as_null
+
+    class JSONElementType(TypeEngine[Any]):
+        """Common function for index / path elements in a JSON expression."""
+
+        _integer = Integer()
+        _string = String()
+
+        def string_bind_processor(self, dialect):
+            return self._string._cached_bind_processor(dialect)
+
+        def string_literal_processor(self, dialect):
+            return self._string._cached_literal_processor(dialect)
+
+        def bind_processor(self, dialect):
+            int_processor = self._integer._cached_bind_processor(dialect)
+            string_processor = self.string_bind_processor(dialect)
+
+            def process(value):
+                if int_processor and isinstance(value, int):
+                    value = int_processor(value)
+                elif string_processor and isinstance(value, str):
+                    value = string_processor(value)
+                return value
+
+            return process
+
+        def literal_processor(self, dialect):
+            int_processor = self._integer._cached_literal_processor(dialect)
+            string_processor = self.string_literal_processor(dialect)
+
+            def process(value):
+                if int_processor and isinstance(value, int):
+                    value = int_processor(value)
+                elif string_processor and isinstance(value, str):
+                    value = string_processor(value)
+                else:
+                    raise NotImplementedError()
+
+                return value
+
+            return process
+
+    class JSONIndexType(JSONElementType):
+        """Placeholder for the datatype of a JSON index value.
+
+        This allows execution-time processing of JSON index values
+        for special syntaxes.
+
+        """
+
+    class JSONIntIndexType(JSONIndexType):
+        """Placeholder for the datatype of a JSON index value.
+
+        This allows execution-time processing of JSON index values
+        for special syntaxes.
+
+        """
+
+    class JSONStrIndexType(JSONIndexType):
+        """Placeholder for the datatype of a JSON index value.
+
+        This allows execution-time processing of JSON index values
+        for special syntaxes.
+
+        """
+
+    class JSONPathType(JSONElementType):
+        """Placeholder type for JSON path operations.
+
+        This allows execution-time processing of a path-based
+        index value into a specific SQL syntax.
+
+        """
+
+        __visit_name__ = "json_path"
+
+    class Comparator(Indexable.Comparator[_T], Concatenable.Comparator[_T]):
+        """Define comparison operations for :class:`_types.JSON`."""
+
+        __slots__ = ()
+
+        def _setup_getitem(self, index):
+            if not isinstance(index, str) and isinstance(
+                index, collections_abc.Sequence
+            ):
+                index = coercions.expect(
+                    roles.BinaryElementRole,
+                    index,
+                    expr=self.expr,
+                    operator=operators.json_path_getitem_op,
+                    bindparam_type=JSON.JSONPathType,
+                )
+
+                operator = operators.json_path_getitem_op
+            else:
+                index = coercions.expect(
+                    roles.BinaryElementRole,
+                    index,
+                    expr=self.expr,
+                    operator=operators.json_getitem_op,
+                    bindparam_type=(
+                        JSON.JSONIntIndexType
+                        if isinstance(index, int)
+                        else JSON.JSONStrIndexType
+                    ),
+                )
+                operator = operators.json_getitem_op
+
+            return operator, index, self.type
+
+        def as_boolean(self):
+            """Consider an indexed value as boolean.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_boolean()).where(
+                    mytable.c.json_column["some_data"].as_boolean() == True
+                )
+
+            .. versionadded:: 1.3.11
+
+            """  # noqa: E501
+            return self._binary_w_type(Boolean(), "as_boolean")
+
+        def as_string(self):
+            """Consider an indexed value as string.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_string()).where(
+                    mytable.c.json_column["some_data"].as_string() == "some string"
+                )
+
+            .. versionadded:: 1.3.11
+
+            """  # noqa: E501
+            return self._binary_w_type(Unicode(), "as_string")
+
+        def as_integer(self):
+            """Consider an indexed value as integer.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_integer()).where(
+                    mytable.c.json_column["some_data"].as_integer() == 5
+                )
+
+            .. versionadded:: 1.3.11
+
+            """  # noqa: E501
+            return self._binary_w_type(Integer(), "as_integer")
+
+        def as_float(self):
+            """Consider an indexed value as float.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_float()).where(
+                    mytable.c.json_column["some_data"].as_float() == 29.75
+                )
+
+            .. versionadded:: 1.3.11
+
+            """  # noqa: E501
+            return self._binary_w_type(Float(), "as_float")
+
+        def as_numeric(self, precision, scale, asdecimal=True):
+            """Consider an indexed value as numeric/decimal.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_numeric(10, 6)).where(
+                    mytable.c.json_column["some_data"].as_numeric(10, 6) == 29.75
+                )
+
+            .. versionadded:: 1.4.0b2
+
+            """  # noqa: E501
+            return self._binary_w_type(
+                Numeric(precision, scale, asdecimal=asdecimal), "as_numeric"
+            )
+
+        def as_json(self):
+            """Consider an indexed value as JSON.
+
+            This is similar to using :class:`_sql.type_coerce`, and will
+            usually not apply a ``CAST()``.
+
+            e.g.::
+
+                stmt = select(mytable.c.json_column["some_data"].as_json())
+
+            This is typically the default behavior of indexed elements in any
+            case.
+
+            Note that comparison of full JSON structures may not be
+            supported by all backends.
+
+            .. versionadded:: 1.3.11
+
+            """
+            return self.expr
+
+        def _binary_w_type(self, typ, method_name):
+            if not isinstance(
+                self.expr, elements.BinaryExpression
+            ) or self.expr.operator not in (
+                operators.json_getitem_op,
+                operators.json_path_getitem_op,
+            ):
+                raise exc.InvalidRequestError(
+                    "The JSON cast operator JSON.%s() only works with a JSON "
+                    "index expression e.g. col['q'].%s()"
+                    % (method_name, method_name)
+                )
+            expr = self.expr._clone()
+            expr.type = typ
+            return expr
+
+    comparator_factory = Comparator
+
+    @property
+    def python_type(self):
+        return dict
+
+    @property  # type: ignore  # mypy property bug
+    def should_evaluate_none(self):
+        """Alias of :attr:`_types.JSON.none_as_null`"""
+        return not self.none_as_null
+
+    @should_evaluate_none.setter
+    def should_evaluate_none(self, value):
+        self.none_as_null = not value
+
+    @util.memoized_property
+    def _str_impl(self):
+        return String()
+
+    def _make_bind_processor(self, string_process, json_serializer):
+        if string_process:
+
+            def process(value):
+                if value is self.NULL:
+                    value = None
+                elif isinstance(value, elements.Null) or (
+                    value is None and self.none_as_null
+                ):
+                    return None
+
+                serialized = json_serializer(value)
+                return string_process(serialized)
+
+        else:
+
+            def process(value):
+                if value is self.NULL:
+                    value = None
+                elif isinstance(value, elements.Null) or (
+                    value is None and self.none_as_null
+                ):
+                    return None
+
+                return json_serializer(value)
+
+        return process
+
+    def bind_processor(self, dialect):
+        string_process = self._str_impl.bind_processor(dialect)
+        json_serializer = dialect._json_serializer or json.dumps
+
+        return self._make_bind_processor(string_process, json_serializer)
+
+    def result_processor(self, dialect, coltype):
+        string_process = self._str_impl.result_processor(dialect, coltype)
+        json_deserializer = dialect._json_deserializer or json.loads
+
+        def process(value):
+            if value is None:
+                return None
+            if string_process:
+                value = string_process(value)
+            return json_deserializer(value)
+
+        return process
+
+
+class ARRAY(
+    SchemaEventTarget, Indexable, Concatenable, TypeEngine[Sequence[Any]]
+):
+    """Represent a SQL Array type.
+
+    .. note::  This type serves as the basis for all ARRAY operations.
+       However, currently **only the PostgreSQL backend has support for SQL
+       arrays in SQLAlchemy**. It is recommended to use the PostgreSQL-specific
+       :class:`sqlalchemy.dialects.postgresql.ARRAY` type directly when using
+       ARRAY types with PostgreSQL, as it provides additional operators
+       specific to that backend.
+
+    :class:`_types.ARRAY` is part of the Core in support of various SQL
+    standard functions such as :class:`_functions.array_agg`
+    which explicitly involve
+    arrays; however, with the exception of the PostgreSQL backend and possibly
+    some third-party dialects, no other SQLAlchemy built-in dialect has support
+    for this type.
+
+    An :class:`_types.ARRAY` type is constructed given the "type"
+    of element::
+
+        mytable = Table("mytable", metadata, Column("data", ARRAY(Integer)))
+
+    The above type represents an N-dimensional array,
+    meaning a supporting backend such as PostgreSQL will interpret values
+    with any number of dimensions automatically.   To produce an INSERT
+    construct that passes in a 1-dimensional array of integers::
+
+        connection.execute(mytable.insert(), {"data": [1, 2, 3]})
+
+    The :class:`_types.ARRAY` type can be constructed given a fixed number
+    of dimensions::
+
+        mytable = Table(
+            "mytable", metadata, Column("data", ARRAY(Integer, dimensions=2))
+        )
+
+    Sending a number of dimensions is optional, but recommended if the
+    datatype is to represent arrays of more than one dimension.  This number
+    is used:
+
+    * When emitting the type declaration itself to the database, e.g.
+      ``INTEGER[][]``
+
+    * When translating Python values to database values, and vice versa, e.g.
+      an ARRAY of :class:`.Unicode` objects uses this number to efficiently
+      access the string values inside of array structures without resorting
+      to per-row type inspection
+
+    * When used with the Python ``getitem`` accessor, the number of dimensions
+      serves to define the kind of type that the ``[]`` operator should
+      return, e.g. for an ARRAY of INTEGER with two dimensions::
+
+          >>> expr = table.c.column[5]  # returns ARRAY(Integer, dimensions=1)
+          >>> expr = expr[6]  # returns Integer
+
+    For 1-dimensional arrays, an :class:`_types.ARRAY` instance with no
+    dimension parameter will generally assume single-dimensional behaviors.
+
+    SQL expressions of type :class:`_types.ARRAY` have support for "index" and
+    "slice" behavior.  The ``[]`` operator produces expression
+    constructs which will produce the appropriate SQL, both for
+    SELECT statements::
+
+        select(mytable.c.data[5], mytable.c.data[2:7])
+
+    as well as UPDATE statements when the :meth:`_expression.Update.values`
+    method is used::
+
+        mytable.update().values(
+            {mytable.c.data[5]: 7, mytable.c.data[2:7]: [1, 2, 3]}
+        )
+
+    Indexed access is one-based by default;
+    for zero-based index conversion, set :paramref:`_types.ARRAY.zero_indexes`.
+
+    The :class:`_types.ARRAY` type also provides for the operators
+    :meth:`.types.ARRAY.Comparator.any` and
+    :meth:`.types.ARRAY.Comparator.all`. The PostgreSQL-specific version of
+    :class:`_types.ARRAY` also provides additional operators.
+
+    .. container:: topic
+
+        **Detecting Changes in ARRAY columns when using the ORM**
+
+        The :class:`_sqltypes.ARRAY` type, when used with the SQLAlchemy ORM,
+        does not detect in-place mutations to the array. In order to detect
+        these, the :mod:`sqlalchemy.ext.mutable` extension must be used, using
+        the :class:`.MutableList` class::
+
+            from sqlalchemy import ARRAY
+            from sqlalchemy.ext.mutable import MutableList
+
+
+            class SomeOrmClass(Base):
+                # ...
+
+                data = Column(MutableList.as_mutable(ARRAY(Integer)))
+
+        This extension will allow "in-place" changes such to the array
+        such as ``.append()`` to produce events which will be detected by the
+        unit of work.  Note that changes to elements **inside** the array,
+        including subarrays that are mutated in place, are **not** detected.
+
+        Alternatively, assigning a new array value to an ORM element that
+        replaces the old one will always trigger a change event.
+
+    .. seealso::
+
+        :class:`sqlalchemy.dialects.postgresql.ARRAY`
+
+    """
+
+    __visit_name__ = "ARRAY"
+
+    _is_array = True
+
+    zero_indexes = False
+    """If True, Python zero-based indexes should be interpreted as one-based
+    on the SQL expression side."""
+
+    def __init__(
+        self,
+        item_type: _TypeEngineArgument[Any],
+        as_tuple: bool = False,
+        dimensions: Optional[int] = None,
+        zero_indexes: bool = False,
+    ):
+        """Construct an :class:`_types.ARRAY`.
+
+        E.g.::
+
+          Column("myarray", ARRAY(Integer))
+
+        Arguments are:
+
+        :param item_type: The data type of items of this array. Note that
+          dimensionality is irrelevant here, so multi-dimensional arrays like
+          ``INTEGER[][]``, are constructed as ``ARRAY(Integer)``, not as
+          ``ARRAY(ARRAY(Integer))`` or such.
+
+        :param as_tuple=False: Specify whether return results
+          should be converted to tuples from lists.  This parameter is
+          not generally needed as a Python list corresponds well
+          to a SQL array.
+
+        :param dimensions: if non-None, the ARRAY will assume a fixed
+         number of dimensions.   This impacts how the array is declared
+         on the database, how it goes about interpreting Python and
+         result values, as well as how expression behavior in conjunction
+         with the "getitem" operator works.  See the description at
+         :class:`_types.ARRAY` for additional detail.
+
+        :param zero_indexes=False: when True, index values will be converted
+         between Python zero-based and SQL one-based indexes, e.g.
+         a value of one will be added to all index values before passing
+         to the database.
+
+        """
+        if isinstance(item_type, ARRAY):
+            raise ValueError(
+                "Do not nest ARRAY types; ARRAY(basetype) "
+                "handles multi-dimensional arrays of basetype"
+            )
+        if isinstance(item_type, type):
+            item_type = item_type()
+        self.item_type = item_type
+        self.as_tuple = as_tuple
+        self.dimensions = dimensions
+        self.zero_indexes = zero_indexes
+
+    class Comparator(
+        Indexable.Comparator[Sequence[Any]],
+        Concatenable.Comparator[Sequence[Any]],
+    ):
+        """Define comparison operations for :class:`_types.ARRAY`.
+
+        More operators are available on the dialect-specific form
+        of this type.  See :class:`.postgresql.ARRAY.Comparator`.
+
+        """
+
+        __slots__ = ()
+
+        type: ARRAY
+
+        def _setup_getitem(self, index):
+            arr_type = self.type
+
+            return_type: TypeEngine[Any]
+
+            if isinstance(index, slice):
+                return_type = arr_type
+                if arr_type.zero_indexes:
+                    index = slice(index.start + 1, index.stop + 1, index.step)
+                slice_ = Slice(
+                    index.start, index.stop, index.step, _name=self.expr.key
+                )
+                return operators.getitem, slice_, return_type
+            else:
+                if arr_type.zero_indexes:
+                    index += 1
+                if arr_type.dimensions is None or arr_type.dimensions == 1:
+                    return_type = arr_type.item_type
+                else:
+                    adapt_kw = {"dimensions": arr_type.dimensions - 1}
+                    return_type = arr_type.adapt(
+                        arr_type.__class__, **adapt_kw
+                    )
+
+                return operators.getitem, index, return_type
+
+        def contains(self, *arg, **kw):
+            """``ARRAY.contains()`` not implemented for the base ARRAY type.
+            Use the dialect-specific ARRAY type.
+
+            .. seealso::
+
+                :class:`_postgresql.ARRAY` - PostgreSQL specific version.
+            """
+            raise NotImplementedError(
+                "ARRAY.contains() not implemented for the base "
+                "ARRAY type; please use the dialect-specific ARRAY type"
+            )
+
+        @util.preload_module("sqlalchemy.sql.elements")
+        def any(self, other, operator=None):
+            """Return ``other operator ANY (array)`` clause.
+
+            .. legacy:: This method is an :class:`_types.ARRAY` - specific
+                construct that is now superseded by the :func:`_sql.any_`
+                function, which features a different calling style. The
+                :func:`_sql.any_` function is also mirrored at the method level
+                via the :meth:`_sql.ColumnOperators.any_` method.
+
+            Usage of array-specific :meth:`_types.ARRAY.Comparator.any`
+            is as follows::
+
+                from sqlalchemy.sql import operators
+
+                conn.execute(
+                    select(table.c.data).where(table.c.data.any(7, operator=operators.lt))
+                )
+
+            :param other: expression to be compared
+            :param operator: an operator object from the
+             :mod:`sqlalchemy.sql.operators`
+             package, defaults to :func:`.operators.eq`.
+
+            .. seealso::
+
+                :func:`_expression.any_`
+
+                :meth:`.types.ARRAY.Comparator.all`
+
+            """  # noqa: E501
+            elements = util.preloaded.sql_elements
+            operator = operator if operator else operators.eq
+
+            arr_type = self.type
+
+            return elements.CollectionAggregate._create_any(self.expr).operate(
+                operators.mirror(operator),
+                coercions.expect(
+                    roles.BinaryElementRole,
+                    element=other,
+                    operator=operator,
+                    expr=self.expr,
+                    bindparam_type=arr_type.item_type,
+                ),
+            )
+
+        @util.preload_module("sqlalchemy.sql.elements")
+        def all(self, other, operator=None):
+            """Return ``other operator ALL (array)`` clause.
+
+            .. legacy:: This method is an :class:`_types.ARRAY` - specific
+                construct that is now superseded by the :func:`_sql.all_`
+                function, which features a different calling style. The
+                :func:`_sql.all_` function is also mirrored at the method level
+                via the :meth:`_sql.ColumnOperators.all_` method.
+
+            Usage of array-specific :meth:`_types.ARRAY.Comparator.all`
+            is as follows::
+
+                from sqlalchemy.sql import operators
+
+                conn.execute(
+                    select(table.c.data).where(table.c.data.all(7, operator=operators.lt))
+                )
+
+            :param other: expression to be compared
+            :param operator: an operator object from the
+             :mod:`sqlalchemy.sql.operators`
+             package, defaults to :func:`.operators.eq`.
+
+            .. seealso::
+
+                :func:`_expression.all_`
+
+                :meth:`.types.ARRAY.Comparator.any`
+
+            """  # noqa: E501
+            elements = util.preloaded.sql_elements
+            operator = operator if operator else operators.eq
+
+            arr_type = self.type
+
+            return elements.CollectionAggregate._create_all(self.expr).operate(
+                operators.mirror(operator),
+                coercions.expect(
+                    roles.BinaryElementRole,
+                    element=other,
+                    operator=operator,
+                    expr=self.expr,
+                    bindparam_type=arr_type.item_type,
+                ),
+            )
+
+    comparator_factory = Comparator
+
+    @property
+    def hashable(self):
+        return self.as_tuple
+
+    @property
+    def python_type(self):
+        return list
+
+    def compare_values(self, x, y):
+        return x == y
+
+    def _set_parent(self, parent, outer=False, **kw):
+        """Support SchemaEventTarget"""
+
+        if not outer and isinstance(self.item_type, SchemaEventTarget):
+            self.item_type._set_parent(parent, **kw)
+
+    def _set_parent_with_dispatch(self, parent, **kw):
+        """Support SchemaEventTarget"""
+
+        super()._set_parent_with_dispatch(parent, outer=True)
+
+        if isinstance(self.item_type, SchemaEventTarget):
+            self.item_type._set_parent_with_dispatch(parent)
+
+    def literal_processor(self, dialect):
+        item_proc = self.item_type.dialect_impl(dialect).literal_processor(
+            dialect
+        )
+        if item_proc is None:
+            return None
+
+        def to_str(elements):
+            return f"[{', '.join(elements)}]"
+
+        def process(value):
+            inner = self._apply_item_processor(
+                value, item_proc, self.dimensions, to_str
+            )
+            return inner
+
+        return process
+
+    def _apply_item_processor(self, arr, itemproc, dim, collection_callable):
+        """Helper method that can be used by bind_processor(),
+        literal_processor(), etc. to apply an item processor to elements of
+        an array value, taking into account the 'dimensions' for this
+        array type.
+
+        See the Postgresql ARRAY datatype for usage examples.
+
+        .. versionadded:: 2.0
+
+        """
+
+        if dim is None:
+            arr = list(arr)
+        if (
+            dim == 1
+            or dim is None
+            and (
+                # this has to be (list, tuple), or at least
+                # not hasattr('__iter__'), since Py3K strings
+                # etc. have __iter__
+                not arr
+                or not isinstance(arr[0], (list, tuple))
+            )
+        ):
+            if itemproc:
+                return collection_callable(itemproc(x) for x in arr)
+            else:
+                return collection_callable(arr)
+        else:
+            return collection_callable(
+                (
+                    self._apply_item_processor(
+                        x,
+                        itemproc,
+                        dim - 1 if dim is not None else None,
+                        collection_callable,
+                    )
+                    if x is not None
+                    else None
+                )
+                for x in arr
+            )
+
+
+class TupleType(TypeEngine[Tuple[Any, ...]]):
+    """represent the composite type of a Tuple."""
+
+    _is_tuple_type = True
+
+    types: List[TypeEngine[Any]]
+
+    def __init__(self, *types: _TypeEngineArgument[Any]):
+        self._fully_typed = NULLTYPE not in types
+        self.types = [
+            item_type() if isinstance(item_type, type) else item_type
+            for item_type in types
+        ]
+
+    def coerce_compared_value(
+        self, op: Optional[OperatorType], value: Any
+    ) -> TypeEngine[Any]:
+        if value is type_api._NO_VALUE_IN_LIST:
+            return super().coerce_compared_value(op, value)
+        else:
+            return TupleType(
+                *[
+                    typ.coerce_compared_value(op, elem)
+                    for typ, elem in zip(self.types, value)
+                ]
+            )
+
+    def _resolve_values_to_types(self, value: Any) -> TupleType:
+        if self._fully_typed:
+            return self
+        else:
+            return TupleType(
+                *[
+                    _resolve_value_to_type(elem) if typ is NULLTYPE else typ
+                    for typ, elem in zip(self.types, value)
+                ]
+            )
+
+    def result_processor(self, dialect, coltype):
+        raise NotImplementedError(
+            "The tuple type does not support being fetched "
+            "as a column in a result row."
+        )
+
+
+class REAL(Float[_N]):
+    """The SQL REAL type.
+
+    .. seealso::
+
+        :class:`_types.Float` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "REAL"
+
+
+class FLOAT(Float[_N]):
+    """The SQL FLOAT type.
+
+    .. seealso::
+
+        :class:`_types.Float` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "FLOAT"
+
+
+class DOUBLE(Double[_N]):
+    """The SQL DOUBLE type.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`_types.Double` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "DOUBLE"
+
+
+class DOUBLE_PRECISION(Double[_N]):
+    """The SQL DOUBLE PRECISION type.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`_types.Double` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "DOUBLE_PRECISION"
+
+
+class NUMERIC(Numeric[_N]):
+    """The SQL NUMERIC type.
+
+    .. seealso::
+
+        :class:`_types.Numeric` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "NUMERIC"
+
+
+class DECIMAL(Numeric[_N]):
+    """The SQL DECIMAL type.
+
+    .. seealso::
+
+        :class:`_types.Numeric` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "DECIMAL"
+
+
+class INTEGER(Integer):
+    """The SQL INT or INTEGER type.
+
+    .. seealso::
+
+        :class:`_types.Integer` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "INTEGER"
+
+
+INT = INTEGER
+
+
+class SMALLINT(SmallInteger):
+    """The SQL SMALLINT type.
+
+    .. seealso::
+
+        :class:`_types.SmallInteger` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "SMALLINT"
+
+
+class BIGINT(BigInteger):
+    """The SQL BIGINT type.
+
+    .. seealso::
+
+        :class:`_types.BigInteger` - documentation for the base type.
+
+    """
+
+    __visit_name__ = "BIGINT"
+
+
+class TIMESTAMP(DateTime):
+    """The SQL TIMESTAMP type.
+
+    :class:`_types.TIMESTAMP` datatypes have support for timezone storage on
+    some backends, such as PostgreSQL and Oracle Database.  Use the
+    :paramref:`~types.TIMESTAMP.timezone` argument in order to enable
+    "TIMESTAMP WITH TIMEZONE" for these backends.
+
+    """
+
+    __visit_name__ = "TIMESTAMP"
+
+    def __init__(self, timezone: bool = False):
+        """Construct a new :class:`_types.TIMESTAMP`.
+
+        :param timezone: boolean.  Indicates that the TIMESTAMP type should
+         enable timezone support, if available on the target database.
+         On a per-dialect basis is similar to "TIMESTAMP WITH TIMEZONE".
+         If the target database does not support timezones, this flag is
+         ignored.
+
+
+        """
+        super().__init__(timezone=timezone)
+
+    def get_dbapi_type(self, dbapi):
+        return dbapi.TIMESTAMP
+
+
+class DATETIME(DateTime):
+    """The SQL DATETIME type."""
+
+    __visit_name__ = "DATETIME"
+
+
+class DATE(Date):
+    """The SQL DATE type."""
+
+    __visit_name__ = "DATE"
+
+
+class TIME(Time):
+    """The SQL TIME type."""
+
+    __visit_name__ = "TIME"
+
+
+class TEXT(Text):
+    """The SQL TEXT type."""
+
+    __visit_name__ = "TEXT"
+
+
+class CLOB(Text):
+    """The CLOB type.
+
+    This type is found in Oracle Database and Informix.
+    """
+
+    __visit_name__ = "CLOB"
+
+
+class VARCHAR(String):
+    """The SQL VARCHAR type."""
+
+    __visit_name__ = "VARCHAR"
+
+
+class NVARCHAR(Unicode):
+    """The SQL NVARCHAR type."""
+
+    __visit_name__ = "NVARCHAR"
+
+
+class CHAR(String):
+    """The SQL CHAR type."""
+
+    __visit_name__ = "CHAR"
+
+
+class NCHAR(Unicode):
+    """The SQL NCHAR type."""
+
+    __visit_name__ = "NCHAR"
+
+
+class BLOB(LargeBinary):
+    """The SQL BLOB type."""
+
+    __visit_name__ = "BLOB"
+
+
+class BINARY(_Binary):
+    """The SQL BINARY type."""
+
+    __visit_name__ = "BINARY"
+
+
+class VARBINARY(_Binary):
+    """The SQL VARBINARY type."""
+
+    __visit_name__ = "VARBINARY"
+
+
+class BOOLEAN(Boolean):
+    """The SQL BOOLEAN type."""
+
+    __visit_name__ = "BOOLEAN"
+
+
+class NullType(TypeEngine[None]):
+    """An unknown type.
+
+    :class:`.NullType` is used as a default type for those cases where
+    a type cannot be determined, including:
+
+    * During table reflection, when the type of a column is not recognized
+      by the :class:`.Dialect`
+    * When constructing SQL expressions using plain Python objects of
+      unknown types (e.g. ``somecolumn == my_special_object``)
+    * When a new :class:`_schema.Column` is created,
+      and the given type is passed
+      as ``None`` or is not passed at all.
+
+    The :class:`.NullType` can be used within SQL expression invocation
+    without issue, it just has no behavior either at the expression
+    construction level or at the bind-parameter/result processing level.
+    :class:`.NullType` will result in a :exc:`.CompileError` if the compiler
+    is asked to render the type itself, such as if it is used in a
+    :func:`.cast` operation or within a schema creation operation such as that
+    invoked by :meth:`_schema.MetaData.create_all` or the
+    :class:`.CreateTable`
+    construct.
+
+    """
+
+    __visit_name__ = "null"
+
+    _isnull = True
+
+    def literal_processor(self, dialect):
+        return None
+
+    class Comparator(TypeEngine.Comparator[_T]):
+        __slots__ = ()
+
+        def _adapt_expression(
+            self,
+            op: OperatorType,
+            other_comparator: TypeEngine.Comparator[Any],
+        ) -> Tuple[OperatorType, TypeEngine[Any]]:
+            if isinstance(
+                other_comparator, NullType.Comparator
+            ) or not operators.is_commutative(op):
+                return op, self.expr.type
+            else:
+                return other_comparator._adapt_expression(op, self)
+
+    comparator_factory = Comparator
+
+
+class TableValueType(HasCacheKey, TypeEngine[Any]):
+    """Refers to a table value type."""
+
+    _is_table_value = True
+
+    _traverse_internals = [
+        ("_elements", InternalTraversal.dp_clauseelement_list),
+    ]
+
+    def __init__(self, *elements: Union[str, _ColumnExpressionArgument[Any]]):
+        self._elements = [
+            coercions.expect(roles.StrAsPlainColumnRole, elem)
+            for elem in elements
+        ]
+
+
+class MatchType(Boolean):
+    """Refers to the return type of the MATCH operator.
+
+    As the :meth:`.ColumnOperators.match` is probably the most open-ended
+    operator in generic SQLAlchemy Core, we can't assume the return type
+    at SQL evaluation time, as MySQL returns a floating point, not a boolean,
+    and other backends might do something different.    So this type
+    acts as a placeholder, currently subclassing :class:`.Boolean`.
+    The type allows dialects to inject result-processing functionality
+    if needed, and on MySQL will return floating-point values.
+
+    """
+
+
+_UUID_RETURN = TypeVar("_UUID_RETURN", str, _python_UUID)
+
+
+class Uuid(Emulated, TypeEngine[_UUID_RETURN]):
+    """Represent a database agnostic UUID datatype.
+
+    For backends that have no "native" UUID datatype, the value will
+    make use of ``CHAR(32)`` and store the UUID as a 32-character alphanumeric
+    hex string.
+
+    For backends which are known to support ``UUID`` directly or a similar
+    uuid-storing datatype such as SQL Server's ``UNIQUEIDENTIFIER``, a
+    "native" mode enabled by default allows these types will be used on those
+    backends.
+
+    In its default mode of use, the :class:`_sqltypes.Uuid` datatype expects
+    **Python uuid objects**, from the Python
+    `uuid <https://docs.python.org/3/library/uuid.html>`_
+    module::
+
+        import uuid
+
+        from sqlalchemy import Uuid
+        from sqlalchemy import Table, Column, MetaData, String
+
+
+        metadata_obj = MetaData()
+
+        t = Table(
+            "t",
+            metadata_obj,
+            Column("uuid_data", Uuid, primary_key=True),
+            Column("other_data", String),
+        )
+
+        with engine.begin() as conn:
+            conn.execute(
+                t.insert(), {"uuid_data": uuid.uuid4(), "other_data": "some data"}
+            )
+
+    To have the :class:`_sqltypes.Uuid` datatype work with string-based
+    Uuids (e.g. 32 character hexadecimal strings), pass the
+    :paramref:`_sqltypes.Uuid.as_uuid` parameter with the value ``False``.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`_sqltypes.UUID` - represents exactly the ``UUID`` datatype
+        without any backend-agnostic behaviors.
+
+    """  # noqa: E501
+
+    __visit_name__ = "uuid"
+
+    collation: Optional[str] = None
+
+    @overload
+    def __init__(
+        self: Uuid[_python_UUID],
+        as_uuid: Literal[True] = ...,
+        native_uuid: bool = ...,
+    ): ...
+
+    @overload
+    def __init__(
+        self: Uuid[str],
+        as_uuid: Literal[False] = ...,
+        native_uuid: bool = ...,
+    ): ...
+
+    def __init__(self, as_uuid: bool = True, native_uuid: bool = True):
+        """Construct a :class:`_sqltypes.Uuid` type.
+
+        :param as_uuid=True: if True, values will be interpreted
+         as Python uuid objects, converting to/from string via the
+         DBAPI.
+
+         .. versionchanged: 2.0 ``as_uuid`` now defaults to ``True``.
+
+        :param native_uuid=True: if True, backends that support either the
+         ``UUID`` datatype directly, or a UUID-storing value
+         (such as SQL Server's ``UNIQUEIDENTIFIER`` will be used by those
+         backends.   If False, a ``CHAR(32)`` datatype will be used for
+         all backends regardless of native support.
+
+        """
+        self.as_uuid = as_uuid
+        self.native_uuid = native_uuid
+
+    @property
+    def python_type(self):
+        return _python_UUID if self.as_uuid else str
+
+    @property
+    def native(self):
+        return self.native_uuid
+
+    def coerce_compared_value(self, op, value):
+        """See :meth:`.TypeEngine.coerce_compared_value` for a description."""
+
+        if isinstance(value, str):
+            return self
+        else:
+            return super().coerce_compared_value(op, value)
+
+    def bind_processor(self, dialect):
+        character_based_uuid = (
+            not dialect.supports_native_uuid or not self.native_uuid
+        )
+
+        if character_based_uuid:
+            if self.as_uuid:
+
+                def process(value):
+                    if value is not None:
+                        value = value.hex
+                    return value
+
+                return process
+            else:
+
+                def process(value):
+                    if value is not None:
+                        value = value.replace("-", "")
+                    return value
+
+                return process
+        else:
+            return None
+
+    def result_processor(self, dialect, coltype):
+        character_based_uuid = (
+            not dialect.supports_native_uuid or not self.native_uuid
+        )
+
+        if character_based_uuid:
+            if self.as_uuid:
+
+                def process(value):
+                    if value is not None:
+                        value = _python_UUID(value)
+                    return value
+
+                return process
+            else:
+
+                def process(value):
+                    if value is not None:
+                        value = str(_python_UUID(value))
+                    return value
+
+                return process
+        else:
+            if not self.as_uuid:
+
+                def process(value):
+                    if value is not None:
+                        value = str(value)
+                    return value
+
+                return process
+            else:
+                return None
+
+    def literal_processor(self, dialect):
+        character_based_uuid = (
+            not dialect.supports_native_uuid or not self.native_uuid
+        )
+
+        if not self.as_uuid:
+
+            def process(value):
+                return f"""'{value.replace("-", "").replace("'", "''")}'"""
+
+            return process
+        else:
+            if character_based_uuid:
+
+                def process(value):
+                    return f"""'{value.hex}'"""
+
+                return process
+            else:
+
+                def process(value):
+                    return f"""'{str(value).replace("'", "''")}'"""
+
+                return process
+
+
+class UUID(Uuid[_UUID_RETURN], type_api.NativeForEmulated):
+    """Represent the SQL UUID type.
+
+    This is the SQL-native form of the :class:`_types.Uuid` database agnostic
+    datatype, and is backwards compatible with the previous PostgreSQL-only
+    version of ``UUID``.
+
+    The :class:`_sqltypes.UUID` datatype only works on databases that have a
+    SQL datatype named ``UUID``. It will not function for backends which don't
+    have this exact-named type, including SQL Server. For backend-agnostic UUID
+    values with native support, including for SQL Server's ``UNIQUEIDENTIFIER``
+    datatype, use the :class:`_sqltypes.Uuid` datatype.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`_sqltypes.Uuid`
+
+    """
+
+    __visit_name__ = "UUID"
+
+    @overload
+    def __init__(self: UUID[_python_UUID], as_uuid: Literal[True] = ...): ...
+
+    @overload
+    def __init__(self: UUID[str], as_uuid: Literal[False] = ...): ...
+
+    def __init__(self, as_uuid: bool = True):
+        """Construct a :class:`_sqltypes.UUID` type.
+
+
+        :param as_uuid=True: if True, values will be interpreted
+         as Python uuid objects, converting to/from string via the
+         DBAPI.
+
+         .. versionchanged: 2.0 ``as_uuid`` now defaults to ``True``.
+
+        """
+        self.as_uuid = as_uuid
+        self.native_uuid = True
+
+    @classmethod
+    def adapt_emulated_to_native(cls, impl, **kw):
+        kw.setdefault("as_uuid", impl.as_uuid)
+        return cls(**kw)
+
+
+NULLTYPE = NullType()
+BOOLEANTYPE = Boolean()
+STRINGTYPE = String()
+INTEGERTYPE = Integer()
+NUMERICTYPE: Numeric[decimal.Decimal] = Numeric()
+MATCHTYPE = MatchType()
+TABLEVALUE = TableValueType()
+DATETIME_TIMEZONE = DateTime(timezone=True)
+TIME_TIMEZONE = Time(timezone=True)
+_BIGINTEGER = BigInteger()
+_DATETIME = DateTime()
+_TIME = Time()
+_STRING = String()
+_UNICODE = Unicode()
+
+_type_map: Dict[Type[Any], TypeEngine[Any]] = {
+    int: Integer(),
+    float: Float(),
+    bool: BOOLEANTYPE,
+    _python_UUID: Uuid(),
+    decimal.Decimal: Numeric(),
+    dt.date: Date(),
+    dt.datetime: _DATETIME,
+    dt.time: _TIME,
+    dt.timedelta: Interval(),
+    type(None): NULLTYPE,
+    bytes: LargeBinary(),
+    str: _STRING,
+    enum.Enum: Enum(enum.Enum),
+    Literal: Enum(enum.Enum),  # type: ignore[dict-item]
+}
+
+
+_type_map_get = _type_map.get
+
+
+def _resolve_value_to_type(value: Any) -> TypeEngine[Any]:
+    _result_type = _type_map_get(type(value), False)
+
+    if _result_type is False:
+        _result_type = getattr(value, "__sa_type_engine__", False)
+
+    if _result_type is False:
+        # use inspect() to detect SQLAlchemy built-in
+        # objects.
+        insp = inspection.inspect(value, False)
+        if (
+            insp is not None
+            and
+            # foil mock.Mock() and other impostors by ensuring
+            # the inspection target itself self-inspects
+            insp.__class__ in inspection._registrars
+        ):
+            raise exc.ArgumentError(
+                "Object %r is not legal as a SQL literal value" % (value,)
+            )
+        return NULLTYPE
+    else:
+        return _result_type._resolve_for_literal(  # type: ignore [union-attr]
+            value
+        )
+
+
+# back-assign to type_api
+type_api.BOOLEANTYPE = BOOLEANTYPE
+type_api.STRINGTYPE = STRINGTYPE
+type_api.INTEGERTYPE = INTEGERTYPE
+type_api.NULLTYPE = NULLTYPE
+type_api.NUMERICTYPE = NUMERICTYPE
+type_api.MATCHTYPE = MATCHTYPE
+type_api.INDEXABLE = INDEXABLE = Indexable
+type_api.TABLEVALUE = TABLEVALUE
+type_api._resolve_value_to_type = _resolve_value_to_type
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/traversals.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/traversals.py
new file mode 100644
index 00000000..13ad2899
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/traversals.py
@@ -0,0 +1,1024 @@
+# sql/traversals.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+from collections import deque
+import collections.abc as collections_abc
+import itertools
+from itertools import zip_longest
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import Deque
+from typing import Dict
+from typing import Iterable
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import Type
+
+from . import operators
+from .cache_key import HasCacheKey
+from .visitors import _TraverseInternalsType
+from .visitors import anon_map
+from .visitors import ExternallyTraversible
+from .visitors import HasTraversalDispatch
+from .visitors import HasTraverseInternals
+from .. import util
+from ..util import langhelpers
+from ..util.typing import Self
+
+
+SKIP_TRAVERSE = util.symbol("skip_traverse")
+COMPARE_FAILED = False
+COMPARE_SUCCEEDED = True
+
+
+def compare(obj1: Any, obj2: Any, **kw: Any) -> bool:
+    strategy: TraversalComparatorStrategy
+    if kw.get("use_proxies", False):
+        strategy = ColIdentityComparatorStrategy()
+    else:
+        strategy = TraversalComparatorStrategy()
+
+    return strategy.compare(obj1, obj2, **kw)
+
+
+def _preconfigure_traversals(target_hierarchy: Type[Any]) -> None:
+    for cls in util.walk_subclasses(target_hierarchy):
+        if hasattr(cls, "_generate_cache_attrs") and hasattr(
+            cls, "_traverse_internals"
+        ):
+            cls._generate_cache_attrs()
+            _copy_internals.generate_dispatch(
+                cls,
+                cls._traverse_internals,
+                "_generated_copy_internals_traversal",
+            )
+            _get_children.generate_dispatch(
+                cls,
+                cls._traverse_internals,
+                "_generated_get_children_traversal",
+            )
+
+
+class HasShallowCopy(HasTraverseInternals):
+    """attribute-wide operations that are useful for classes that use
+    __slots__ and therefore can't operate on their attributes in a dictionary.
+
+
+    """
+
+    __slots__ = ()
+
+    if typing.TYPE_CHECKING:
+
+        def _generated_shallow_copy_traversal(self, other: Self) -> None: ...
+
+        def _generated_shallow_from_dict_traversal(
+            self, d: Dict[str, Any]
+        ) -> None: ...
+
+        def _generated_shallow_to_dict_traversal(self) -> Dict[str, Any]: ...
+
+    @classmethod
+    def _generate_shallow_copy(
+        cls,
+        internal_dispatch: _TraverseInternalsType,
+        method_name: str,
+    ) -> Callable[[Self, Self], None]:
+        code = "\n".join(
+            f"    other.{attrname} = self.{attrname}"
+            for attrname, _ in internal_dispatch
+        )
+        meth_text = f"def {method_name}(self, other):\n{code}\n"
+        return langhelpers._exec_code_in_env(meth_text, {}, method_name)
+
+    @classmethod
+    def _generate_shallow_to_dict(
+        cls,
+        internal_dispatch: _TraverseInternalsType,
+        method_name: str,
+    ) -> Callable[[Self], Dict[str, Any]]:
+        code = ",\n".join(
+            f"    '{attrname}': self.{attrname}"
+            for attrname, _ in internal_dispatch
+        )
+        meth_text = f"def {method_name}(self):\n    return {{{code}}}\n"
+        return langhelpers._exec_code_in_env(meth_text, {}, method_name)
+
+    @classmethod
+    def _generate_shallow_from_dict(
+        cls,
+        internal_dispatch: _TraverseInternalsType,
+        method_name: str,
+    ) -> Callable[[Self, Dict[str, Any]], None]:
+        code = "\n".join(
+            f"    self.{attrname} = d['{attrname}']"
+            for attrname, _ in internal_dispatch
+        )
+        meth_text = f"def {method_name}(self, d):\n{code}\n"
+        return langhelpers._exec_code_in_env(meth_text, {}, method_name)
+
+    def _shallow_from_dict(self, d: Dict[str, Any]) -> None:
+        cls = self.__class__
+
+        shallow_from_dict: Callable[[HasShallowCopy, Dict[str, Any]], None]
+        try:
+            shallow_from_dict = cls.__dict__[
+                "_generated_shallow_from_dict_traversal"
+            ]
+        except KeyError:
+            shallow_from_dict = self._generate_shallow_from_dict(
+                cls._traverse_internals,
+                "_generated_shallow_from_dict_traversal",
+            )
+
+            cls._generated_shallow_from_dict_traversal = shallow_from_dict  # type: ignore  # noqa: E501
+
+        shallow_from_dict(self, d)
+
+    def _shallow_to_dict(self) -> Dict[str, Any]:
+        cls = self.__class__
+
+        shallow_to_dict: Callable[[HasShallowCopy], Dict[str, Any]]
+
+        try:
+            shallow_to_dict = cls.__dict__[
+                "_generated_shallow_to_dict_traversal"
+            ]
+        except KeyError:
+            shallow_to_dict = self._generate_shallow_to_dict(
+                cls._traverse_internals, "_generated_shallow_to_dict_traversal"
+            )
+
+            cls._generated_shallow_to_dict_traversal = shallow_to_dict  # type: ignore  # noqa: E501
+        return shallow_to_dict(self)
+
+    def _shallow_copy_to(self, other: Self) -> None:
+        cls = self.__class__
+
+        shallow_copy: Callable[[Self, Self], None]
+        try:
+            shallow_copy = cls.__dict__["_generated_shallow_copy_traversal"]
+        except KeyError:
+            shallow_copy = self._generate_shallow_copy(
+                cls._traverse_internals, "_generated_shallow_copy_traversal"
+            )
+
+            cls._generated_shallow_copy_traversal = shallow_copy  # type: ignore  # noqa: E501
+        shallow_copy(self, other)
+
+    def _clone(self, **kw: Any) -> Self:
+        """Create a shallow copy"""
+        c = self.__class__.__new__(self.__class__)
+        self._shallow_copy_to(c)
+        return c
+
+
+class GenerativeOnTraversal(HasShallowCopy):
+    """Supplies Generative behavior but making use of traversals to shallow
+    copy.
+
+    .. seealso::
+
+        :class:`sqlalchemy.sql.base.Generative`
+
+
+    """
+
+    __slots__ = ()
+
+    def _generate(self) -> Self:
+        cls = self.__class__
+        s = cls.__new__(cls)
+        self._shallow_copy_to(s)
+        return s
+
+
+def _clone(element, **kw):
+    return element._clone()
+
+
+class HasCopyInternals(HasTraverseInternals):
+    __slots__ = ()
+
+    def _clone(self, **kw):
+        raise NotImplementedError()
+
+    def _copy_internals(
+        self, *, omit_attrs: Iterable[str] = (), **kw: Any
+    ) -> None:
+        """Reassign internal elements to be clones of themselves.
+
+        Called during a copy-and-traverse operation on newly
+        shallow-copied elements to create a deep copy.
+
+        The given clone function should be used, which may be applying
+        additional transformations to the element (i.e. replacement
+        traversal, cloned traversal, annotations).
+
+        """
+
+        try:
+            traverse_internals = self._traverse_internals
+        except AttributeError:
+            # user-defined classes may not have a _traverse_internals
+            return
+
+        for attrname, obj, meth in _copy_internals.run_generated_dispatch(
+            self, traverse_internals, "_generated_copy_internals_traversal"
+        ):
+            if attrname in omit_attrs:
+                continue
+
+            if obj is not None:
+                result = meth(attrname, self, obj, **kw)
+                if result is not None:
+                    setattr(self, attrname, result)
+
+
+class _CopyInternalsTraversal(HasTraversalDispatch):
+    """Generate a _copy_internals internal traversal dispatch for classes
+    with a _traverse_internals collection."""
+
+    def visit_clauseelement(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return clone(element, **kw)
+
+    def visit_clauseelement_list(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return [clone(clause, **kw) for clause in element]
+
+    def visit_clauseelement_tuple(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return tuple([clone(clause, **kw) for clause in element])
+
+    def visit_executable_options(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return tuple([clone(clause, **kw) for clause in element])
+
+    def visit_clauseelement_unordered_set(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return {clone(clause, **kw) for clause in element}
+
+    def visit_clauseelement_tuples(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return [
+            tuple(clone(tup_elem, **kw) for tup_elem in elem)
+            for elem in element
+        ]
+
+    def visit_string_clauseelement_dict(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return {key: clone(value, **kw) for key, value in element.items()}
+
+    def visit_setup_join_tuple(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return tuple(
+            (
+                clone(target, **kw) if target is not None else None,
+                clone(onclause, **kw) if onclause is not None else None,
+                clone(from_, **kw) if from_ is not None else None,
+                flags,
+            )
+            for (target, onclause, from_, flags) in element
+        )
+
+    def visit_memoized_select_entities(self, attrname, parent, element, **kw):
+        return self.visit_clauseelement_tuple(attrname, parent, element, **kw)
+
+    def visit_dml_ordered_values(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        # sequence of 2-tuples
+        return [
+            (
+                (
+                    clone(key, **kw)
+                    if hasattr(key, "__clause_element__")
+                    else key
+                ),
+                clone(value, **kw),
+            )
+            for key, value in element
+        ]
+
+    def visit_dml_values(self, attrname, parent, element, clone=_clone, **kw):
+        return {
+            (
+                clone(key, **kw) if hasattr(key, "__clause_element__") else key
+            ): clone(value, **kw)
+            for key, value in element.items()
+        }
+
+    def visit_dml_multi_values(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        # sequence of sequences, each sequence contains a list/dict/tuple
+
+        def copy(elem):
+            if isinstance(elem, (list, tuple)):
+                return [
+                    (
+                        clone(value, **kw)
+                        if hasattr(value, "__clause_element__")
+                        else value
+                    )
+                    for value in elem
+                ]
+            elif isinstance(elem, dict):
+                return {
+                    (
+                        clone(key, **kw)
+                        if hasattr(key, "__clause_element__")
+                        else key
+                    ): (
+                        clone(value, **kw)
+                        if hasattr(value, "__clause_element__")
+                        else value
+                    )
+                    for key, value in elem.items()
+                }
+            else:
+                # TODO: use abc classes
+                assert False
+
+        return [
+            [copy(sub_element) for sub_element in sequence]
+            for sequence in element
+        ]
+
+    def visit_propagate_attrs(
+        self, attrname, parent, element, clone=_clone, **kw
+    ):
+        return element
+
+
+_copy_internals = _CopyInternalsTraversal()
+
+
+def _flatten_clauseelement(element):
+    while hasattr(element, "__clause_element__") and not getattr(
+        element, "is_clause_element", False
+    ):
+        element = element.__clause_element__()
+
+    return element
+
+
+class _GetChildrenTraversal(HasTraversalDispatch):
+    """Generate a _children_traversal internal traversal dispatch for classes
+    with a _traverse_internals collection."""
+
+    def visit_has_cache_key(self, element, **kw):
+        # the GetChildren traversal refers explicitly to ClauseElement
+        # structures.  Within these, a plain HasCacheKey is not a
+        # ClauseElement, so don't include these.
+        return ()
+
+    def visit_clauseelement(self, element, **kw):
+        return (element,)
+
+    def visit_clauseelement_list(self, element, **kw):
+        return element
+
+    def visit_clauseelement_tuple(self, element, **kw):
+        return element
+
+    def visit_clauseelement_tuples(self, element, **kw):
+        return itertools.chain.from_iterable(element)
+
+    def visit_fromclause_canonical_column_collection(self, element, **kw):
+        return ()
+
+    def visit_string_clauseelement_dict(self, element, **kw):
+        return element.values()
+
+    def visit_fromclause_ordered_set(self, element, **kw):
+        return element
+
+    def visit_clauseelement_unordered_set(self, element, **kw):
+        return element
+
+    def visit_setup_join_tuple(self, element, **kw):
+        for target, onclause, from_, flags in element:
+            if from_ is not None:
+                yield from_
+
+            if not isinstance(target, str):
+                yield _flatten_clauseelement(target)
+
+            if onclause is not None and not isinstance(onclause, str):
+                yield _flatten_clauseelement(onclause)
+
+    def visit_memoized_select_entities(self, element, **kw):
+        return self.visit_clauseelement_tuple(element, **kw)
+
+    def visit_dml_ordered_values(self, element, **kw):
+        for k, v in element:
+            if hasattr(k, "__clause_element__"):
+                yield k
+            yield v
+
+    def visit_dml_values(self, element, **kw):
+        expr_values = {k for k in element if hasattr(k, "__clause_element__")}
+        str_values = expr_values.symmetric_difference(element)
+
+        for k in sorted(str_values):
+            yield element[k]
+        for k in expr_values:
+            yield k
+            yield element[k]
+
+    def visit_dml_multi_values(self, element, **kw):
+        return ()
+
+    def visit_propagate_attrs(self, element, **kw):
+        return ()
+
+
+_get_children = _GetChildrenTraversal()
+
+
+@util.preload_module("sqlalchemy.sql.elements")
+def _resolve_name_for_compare(element, name, anon_map, **kw):
+    if isinstance(name, util.preloaded.sql_elements._anonymous_label):
+        name = name.apply_map(anon_map)
+
+    return name
+
+
+class TraversalComparatorStrategy(HasTraversalDispatch, util.MemoizedSlots):
+    __slots__ = "stack", "cache", "anon_map"
+
+    def __init__(self):
+        self.stack: Deque[
+            Tuple[
+                Optional[ExternallyTraversible],
+                Optional[ExternallyTraversible],
+            ]
+        ] = deque()
+        self.cache = set()
+
+    def _memoized_attr_anon_map(self):
+        return (anon_map(), anon_map())
+
+    def compare(
+        self,
+        obj1: ExternallyTraversible,
+        obj2: ExternallyTraversible,
+        **kw: Any,
+    ) -> bool:
+        stack = self.stack
+        cache = self.cache
+
+        compare_annotations = kw.get("compare_annotations", False)
+
+        stack.append((obj1, obj2))
+
+        while stack:
+            left, right = stack.popleft()
+
+            if left is right:
+                continue
+            elif left is None or right is None:
+                # we know they are different so no match
+                return False
+            elif (left, right) in cache:
+                continue
+            cache.add((left, right))
+
+            visit_name = left.__visit_name__
+            if visit_name != right.__visit_name__:
+                return False
+
+            meth = getattr(self, "compare_%s" % visit_name, None)
+
+            if meth:
+                attributes_compared = meth(left, right, **kw)
+                if attributes_compared is COMPARE_FAILED:
+                    return False
+                elif attributes_compared is SKIP_TRAVERSE:
+                    continue
+
+                # attributes_compared is returned as a list of attribute
+                # names that were "handled" by the comparison method above.
+                # remaining attribute names in the _traverse_internals
+                # will be compared.
+            else:
+                attributes_compared = ()
+
+            for (
+                (left_attrname, left_visit_sym),
+                (right_attrname, right_visit_sym),
+            ) in zip_longest(
+                left._traverse_internals,
+                right._traverse_internals,
+                fillvalue=(None, None),
+            ):
+                if not compare_annotations and (
+                    (left_attrname == "_annotations")
+                    or (right_attrname == "_annotations")
+                ):
+                    continue
+
+                if (
+                    left_attrname != right_attrname
+                    or left_visit_sym is not right_visit_sym
+                ):
+                    return False
+                elif left_attrname in attributes_compared:
+                    continue
+
+                assert left_visit_sym is not None
+                assert left_attrname is not None
+                assert right_attrname is not None
+
+                dispatch = self.dispatch(left_visit_sym)
+                assert dispatch is not None, (
+                    f"{self.__class__} has no dispatch for "
+                    f"'{self._dispatch_lookup[left_visit_sym]}'"
+                )
+                left_child = operator.attrgetter(left_attrname)(left)
+                right_child = operator.attrgetter(right_attrname)(right)
+                if left_child is None:
+                    if right_child is not None:
+                        return False
+                    else:
+                        continue
+                elif right_child is None:
+                    return False
+
+                comparison = dispatch(
+                    left_attrname, left, left_child, right, right_child, **kw
+                )
+                if comparison is COMPARE_FAILED:
+                    return False
+
+        return True
+
+    def compare_inner(self, obj1, obj2, **kw):
+        comparator = self.__class__()
+        return comparator.compare(obj1, obj2, **kw)
+
+    def visit_has_cache_key(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        if left._gen_cache_key(self.anon_map[0], []) != right._gen_cache_key(
+            self.anon_map[1], []
+        ):
+            return COMPARE_FAILED
+
+    def visit_propagate_attrs(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return self.compare_inner(
+            left.get("plugin_subject", None), right.get("plugin_subject", None)
+        )
+
+    def visit_has_cache_key_list(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for l, r in zip_longest(left, right, fillvalue=None):
+            if l is None:
+                if r is not None:
+                    return COMPARE_FAILED
+                else:
+                    continue
+            elif r is None:
+                return COMPARE_FAILED
+
+            if l._gen_cache_key(self.anon_map[0], []) != r._gen_cache_key(
+                self.anon_map[1], []
+            ):
+                return COMPARE_FAILED
+
+    def visit_executable_options(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for l, r in zip_longest(left, right, fillvalue=None):
+            if l is None:
+                if r is not None:
+                    return COMPARE_FAILED
+                else:
+                    continue
+            elif r is None:
+                return COMPARE_FAILED
+
+            if (
+                l._gen_cache_key(self.anon_map[0], [])
+                if l._is_has_cache_key
+                else l
+            ) != (
+                r._gen_cache_key(self.anon_map[1], [])
+                if r._is_has_cache_key
+                else r
+            ):
+                return COMPARE_FAILED
+
+    def visit_clauseelement(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        self.stack.append((left, right))
+
+    def visit_fromclause_canonical_column_collection(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for lcol, rcol in zip_longest(left, right, fillvalue=None):
+            self.stack.append((lcol, rcol))
+
+    def visit_fromclause_derived_column_collection(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        pass
+
+    def visit_string_clauseelement_dict(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for lstr, rstr in zip_longest(
+            sorted(left), sorted(right), fillvalue=None
+        ):
+            if lstr != rstr:
+                return COMPARE_FAILED
+            self.stack.append((left[lstr], right[rstr]))
+
+    def visit_clauseelement_tuples(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for ltup, rtup in zip_longest(left, right, fillvalue=None):
+            if ltup is None or rtup is None:
+                return COMPARE_FAILED
+
+            for l, r in zip_longest(ltup, rtup, fillvalue=None):
+                self.stack.append((l, r))
+
+    def visit_clauseelement_list(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for l, r in zip_longest(left, right, fillvalue=None):
+            self.stack.append((l, r))
+
+    def visit_clauseelement_tuple(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for l, r in zip_longest(left, right, fillvalue=None):
+            self.stack.append((l, r))
+
+    def _compare_unordered_sequences(self, seq1, seq2, **kw):
+        if seq1 is None:
+            return seq2 is None
+
+        completed: Set[object] = set()
+        for clause in seq1:
+            for other_clause in set(seq2).difference(completed):
+                if self.compare_inner(clause, other_clause, **kw):
+                    completed.add(other_clause)
+                    break
+        return len(completed) == len(seq1) == len(seq2)
+
+    def visit_clauseelement_unordered_set(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return self._compare_unordered_sequences(left, right, **kw)
+
+    def visit_fromclause_ordered_set(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for l, r in zip_longest(left, right, fillvalue=None):
+            self.stack.append((l, r))
+
+    def visit_string(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_string_list(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_string_multi_dict(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for lk, rk in zip_longest(
+            sorted(left.keys()), sorted(right.keys()), fillvalue=(None, None)
+        ):
+            if lk != rk:
+                return COMPARE_FAILED
+
+            lv, rv = left[lk], right[rk]
+
+            lhc = isinstance(left, HasCacheKey)
+            rhc = isinstance(right, HasCacheKey)
+            if lhc and rhc:
+                if lv._gen_cache_key(
+                    self.anon_map[0], []
+                ) != rv._gen_cache_key(self.anon_map[1], []):
+                    return COMPARE_FAILED
+            elif lhc != rhc:
+                return COMPARE_FAILED
+            elif lv != rv:
+                return COMPARE_FAILED
+
+    def visit_multi(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        lhc = isinstance(left, HasCacheKey)
+        rhc = isinstance(right, HasCacheKey)
+        if lhc and rhc:
+            if left._gen_cache_key(
+                self.anon_map[0], []
+            ) != right._gen_cache_key(self.anon_map[1], []):
+                return COMPARE_FAILED
+        elif lhc != rhc:
+            return COMPARE_FAILED
+        else:
+            return left == right
+
+    def visit_anon_name(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return _resolve_name_for_compare(
+            left_parent, left, self.anon_map[0], **kw
+        ) == _resolve_name_for_compare(
+            right_parent, right, self.anon_map[1], **kw
+        )
+
+    def visit_boolean(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_operator(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_type(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left._compare_type_affinity(right)
+
+    def visit_plain_dict(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_dialect_options(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_annotations_key(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        if left and right:
+            return (
+                left_parent._annotations_cache_key
+                == right_parent._annotations_cache_key
+            )
+        else:
+            return left == right
+
+    def visit_with_context_options(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return tuple((fn.__code__, c_key) for fn, c_key in left) == tuple(
+            (fn.__code__, c_key) for fn, c_key in right
+        )
+
+    def visit_plain_obj(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_named_ddl_element(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        if left is None:
+            if right is not None:
+                return COMPARE_FAILED
+
+        return left.name == right.name
+
+    def visit_prefix_sequence(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for (l_clause, l_str), (r_clause, r_str) in zip_longest(
+            left, right, fillvalue=(None, None)
+        ):
+            if l_str != r_str:
+                return COMPARE_FAILED
+            else:
+                self.stack.append((l_clause, r_clause))
+
+    def visit_setup_join_tuple(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        # TODO: look at attrname for "legacy_join" and use different structure
+        for (
+            (l_target, l_onclause, l_from, l_flags),
+            (r_target, r_onclause, r_from, r_flags),
+        ) in zip_longest(left, right, fillvalue=(None, None, None, None)):
+            if l_flags != r_flags:
+                return COMPARE_FAILED
+            self.stack.append((l_target, r_target))
+            self.stack.append((l_onclause, r_onclause))
+            self.stack.append((l_from, r_from))
+
+    def visit_memoized_select_entities(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return self.visit_clauseelement_tuple(
+            attrname, left_parent, left, right_parent, right, **kw
+        )
+
+    def visit_table_hint_list(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        left_keys = sorted(left, key=lambda elem: (elem[0].fullname, elem[1]))
+        right_keys = sorted(
+            right, key=lambda elem: (elem[0].fullname, elem[1])
+        )
+        for (ltable, ldialect), (rtable, rdialect) in zip_longest(
+            left_keys, right_keys, fillvalue=(None, None)
+        ):
+            if ldialect != rdialect:
+                return COMPARE_FAILED
+            elif left[(ltable, ldialect)] != right[(rtable, rdialect)]:
+                return COMPARE_FAILED
+            else:
+                self.stack.append((ltable, rtable))
+
+    def visit_statement_hint_list(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        return left == right
+
+    def visit_unknown_structure(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        raise NotImplementedError()
+
+    def visit_dml_ordered_values(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        # sequence of tuple pairs
+
+        for (lk, lv), (rk, rv) in zip_longest(
+            left, right, fillvalue=(None, None)
+        ):
+            if not self._compare_dml_values_or_ce(lk, rk, **kw):
+                return COMPARE_FAILED
+
+    def _compare_dml_values_or_ce(self, lv, rv, **kw):
+        lvce = hasattr(lv, "__clause_element__")
+        rvce = hasattr(rv, "__clause_element__")
+        if lvce != rvce:
+            return False
+        elif lvce and not self.compare_inner(lv, rv, **kw):
+            return False
+        elif not lvce and lv != rv:
+            return False
+        elif not self.compare_inner(lv, rv, **kw):
+            return False
+
+        return True
+
+    def visit_dml_values(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        if left is None or right is None or len(left) != len(right):
+            return COMPARE_FAILED
+
+        if isinstance(left, collections_abc.Sequence):
+            for lv, rv in zip(left, right):
+                if not self._compare_dml_values_or_ce(lv, rv, **kw):
+                    return COMPARE_FAILED
+        elif isinstance(right, collections_abc.Sequence):
+            return COMPARE_FAILED
+        else:
+            # dictionaries guaranteed to support insert ordering in
+            # py37 so that we can compare the keys in order.  without
+            # this, we can't compare SQL expression keys because we don't
+            # know which key is which
+            for (lk, lv), (rk, rv) in zip(left.items(), right.items()):
+                if not self._compare_dml_values_or_ce(lk, rk, **kw):
+                    return COMPARE_FAILED
+                if not self._compare_dml_values_or_ce(lv, rv, **kw):
+                    return COMPARE_FAILED
+
+    def visit_dml_multi_values(
+        self, attrname, left_parent, left, right_parent, right, **kw
+    ):
+        for lseq, rseq in zip_longest(left, right, fillvalue=None):
+            if lseq is None or rseq is None:
+                return COMPARE_FAILED
+
+            for ld, rd in zip_longest(lseq, rseq, fillvalue=None):
+                if (
+                    self.visit_dml_values(
+                        attrname, left_parent, ld, right_parent, rd, **kw
+                    )
+                    is COMPARE_FAILED
+                ):
+                    return COMPARE_FAILED
+
+    def compare_expression_clauselist(self, left, right, **kw):
+        if left.operator is right.operator:
+            if operators.is_associative(left.operator):
+                if self._compare_unordered_sequences(
+                    left.clauses, right.clauses, **kw
+                ):
+                    return ["operator", "clauses"]
+                else:
+                    return COMPARE_FAILED
+            else:
+                return ["operator"]
+        else:
+            return COMPARE_FAILED
+
+    def compare_clauselist(self, left, right, **kw):
+        return self.compare_expression_clauselist(left, right, **kw)
+
+    def compare_binary(self, left, right, **kw):
+        if left.operator == right.operator:
+            if operators.is_commutative(left.operator):
+                if (
+                    self.compare_inner(left.left, right.left, **kw)
+                    and self.compare_inner(left.right, right.right, **kw)
+                ) or (
+                    self.compare_inner(left.left, right.right, **kw)
+                    and self.compare_inner(left.right, right.left, **kw)
+                ):
+                    return ["operator", "negate", "left", "right"]
+                else:
+                    return COMPARE_FAILED
+            else:
+                return ["operator", "negate"]
+        else:
+            return COMPARE_FAILED
+
+    def compare_bindparam(self, left, right, **kw):
+        compare_keys = kw.pop("compare_keys", True)
+        compare_values = kw.pop("compare_values", True)
+
+        if compare_values:
+            omit = []
+        else:
+            # this means, "skip these, we already compared"
+            omit = ["callable", "value"]
+
+        if not compare_keys:
+            omit.append("key")
+
+        return omit
+
+
+class ColIdentityComparatorStrategy(TraversalComparatorStrategy):
+    def compare_column_element(
+        self, left, right, use_proxies=True, equivalents=(), **kw
+    ):
+        """Compare ColumnElements using proxies and equivalent collections.
+
+        This is a comparison strategy specific to the ORM.
+        """
+
+        to_compare = (right,)
+        if equivalents and right in equivalents:
+            to_compare = equivalents[right].union(to_compare)
+
+        for oth in to_compare:
+            if use_proxies and left.shares_lineage(oth):
+                return SKIP_TRAVERSE
+            elif hash(left) == hash(right):
+                return SKIP_TRAVERSE
+        else:
+            return COMPARE_FAILED
+
+    def compare_column(self, left, right, **kw):
+        return self.compare_column_element(left, right, **kw)
+
+    def compare_label(self, left, right, **kw):
+        return self.compare_column_element(left, right, **kw)
+
+    def compare_table(self, left, right, **kw):
+        # tables compare on identity, since it's not really feasible to
+        # compare them column by column with the above rules
+        return SKIP_TRAVERSE if left is right else COMPARE_FAILED
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/type_api.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/type_api.py
new file mode 100644
index 00000000..aeb804d3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/type_api.py
@@ -0,0 +1,2343 @@
+# sql/type_api.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Base types API.
+
+"""
+
+from __future__ import annotations
+
+from enum import Enum
+from types import ModuleType
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import Generic
+from typing import Mapping
+from typing import NewType
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .base import SchemaEventTarget
+from .cache_key import CacheConst
+from .cache_key import NO_CACHE
+from .operators import ColumnOperators
+from .visitors import Visitable
+from .. import exc
+from .. import util
+from ..util.typing import Protocol
+from ..util.typing import Self
+from ..util.typing import TypeAliasType
+from ..util.typing import TypedDict
+from ..util.typing import TypeGuard
+
+# these are back-assigned by sqltypes.
+if typing.TYPE_CHECKING:
+    from ._typing import _TypeEngineArgument
+    from .elements import BindParameter
+    from .elements import ColumnElement
+    from .operators import OperatorType
+    from .sqltypes import _resolve_value_to_type as _resolve_value_to_type
+    from .sqltypes import BOOLEANTYPE as BOOLEANTYPE  # noqa: F401
+    from .sqltypes import INDEXABLE as INDEXABLE  # noqa: F401
+    from .sqltypes import INTEGERTYPE as INTEGERTYPE  # noqa: F401
+    from .sqltypes import MATCHTYPE as MATCHTYPE  # noqa: F401
+    from .sqltypes import NULLTYPE as NULLTYPE
+    from .sqltypes import NUMERICTYPE as NUMERICTYPE  # noqa: F401
+    from .sqltypes import STRINGTYPE as STRINGTYPE  # noqa: F401
+    from .sqltypes import TABLEVALUE as TABLEVALUE  # noqa: F401
+    from ..engine.interfaces import Dialect
+    from ..util.typing import GenericProtocol
+
+_T = TypeVar("_T", bound=Any)
+_T_co = TypeVar("_T_co", bound=Any, covariant=True)
+_T_con = TypeVar("_T_con", bound=Any, contravariant=True)
+_O = TypeVar("_O", bound=object)
+_TE = TypeVar("_TE", bound="TypeEngine[Any]")
+_CT = TypeVar("_CT", bound=Any)
+
+_MatchedOnType = Union[
+    "GenericProtocol[Any]", TypeAliasType, NewType, Type[Any]
+]
+
+
+class _NoValueInList(Enum):
+    NO_VALUE_IN_LIST = 0
+    """indicates we are trying to determine the type of an expression
+    against an empty list."""
+
+
+_NO_VALUE_IN_LIST = _NoValueInList.NO_VALUE_IN_LIST
+
+
+class _LiteralProcessorType(Protocol[_T_co]):
+    def __call__(self, value: Any) -> str: ...
+
+
+class _BindProcessorType(Protocol[_T_con]):
+    def __call__(self, value: Optional[_T_con]) -> Any: ...
+
+
+class _ResultProcessorType(Protocol[_T_co]):
+    def __call__(self, value: Any) -> Optional[_T_co]: ...
+
+
+class _SentinelProcessorType(Protocol[_T_co]):
+    def __call__(self, value: Any) -> Optional[_T_co]: ...
+
+
+class _BaseTypeMemoDict(TypedDict):
+    impl: TypeEngine[Any]
+    result: Dict[Any, Optional[_ResultProcessorType[Any]]]
+
+
+class _TypeMemoDict(_BaseTypeMemoDict, total=False):
+    literal: Optional[_LiteralProcessorType[Any]]
+    bind: Optional[_BindProcessorType[Any]]
+    sentinel: Optional[_SentinelProcessorType[Any]]
+    custom: Dict[Any, object]
+
+
+class _ComparatorFactory(Protocol[_T]):
+    def __call__(
+        self, expr: ColumnElement[_T]
+    ) -> TypeEngine.Comparator[_T]: ...
+
+
+class TypeEngine(Visitable, Generic[_T]):
+    """The ultimate base class for all SQL datatypes.
+
+    Common subclasses of :class:`.TypeEngine` include
+    :class:`.String`, :class:`.Integer`, and :class:`.Boolean`.
+
+    For an overview of the SQLAlchemy typing system, see
+    :ref:`types_toplevel`.
+
+    .. seealso::
+
+        :ref:`types_toplevel`
+
+    """
+
+    _sqla_type = True
+    _isnull = False
+    _is_tuple_type = False
+    _is_table_value = False
+    _is_array = False
+    _is_type_decorator = False
+
+    render_bind_cast = False
+    """Render bind casts for :attr:`.BindTyping.RENDER_CASTS` mode.
+
+    If True, this type (usually a dialect level impl type) signals
+    to the compiler that a cast should be rendered around a bound parameter
+    for this type.
+
+    .. versionadded:: 2.0
+
+    .. seealso::
+
+        :class:`.BindTyping`
+
+    """
+
+    render_literal_cast = False
+    """render casts when rendering a value as an inline literal,
+    e.g. with :meth:`.TypeEngine.literal_processor`.
+
+    .. versionadded:: 2.0
+
+    """
+
+    class Comparator(
+        ColumnOperators,
+        Generic[_CT],
+    ):
+        """Base class for custom comparison operations defined at the
+        type level.  See :attr:`.TypeEngine.comparator_factory`.
+
+
+        """
+
+        __slots__ = "expr", "type"
+
+        expr: ColumnElement[_CT]
+        type: TypeEngine[_CT]
+
+        def __clause_element__(self) -> ColumnElement[_CT]:
+            return self.expr
+
+        def __init__(self, expr: ColumnElement[_CT]):
+            self.expr = expr
+            self.type = expr.type
+
+        def __reduce__(self) -> Any:
+            return self.__class__, (self.expr,)
+
+        @util.preload_module("sqlalchemy.sql.default_comparator")
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnElement[_CT]:
+            default_comparator = util.preloaded.sql_default_comparator
+            op_fn, addtl_kw = default_comparator.operator_lookup[op.__name__]
+            if kwargs:
+                addtl_kw = addtl_kw.union(kwargs)
+            return op_fn(self.expr, op, *other, **addtl_kw)
+
+        @util.preload_module("sqlalchemy.sql.default_comparator")
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnElement[_CT]:
+            default_comparator = util.preloaded.sql_default_comparator
+            op_fn, addtl_kw = default_comparator.operator_lookup[op.__name__]
+            if kwargs:
+                addtl_kw = addtl_kw.union(kwargs)
+            return op_fn(self.expr, op, other, reverse=True, **addtl_kw)
+
+        def _adapt_expression(
+            self,
+            op: OperatorType,
+            other_comparator: TypeEngine.Comparator[Any],
+        ) -> Tuple[OperatorType, TypeEngine[Any]]:
+            """evaluate the return type of <self> <op> <othertype>,
+            and apply any adaptations to the given operator.
+
+            This method determines the type of a resulting binary expression
+            given two source types and an operator.   For example, two
+            :class:`_schema.Column` objects, both of the type
+            :class:`.Integer`, will
+            produce a :class:`.BinaryExpression` that also has the type
+            :class:`.Integer` when compared via the addition (``+``) operator.
+            However, using the addition operator with an :class:`.Integer`
+            and a :class:`.Date` object will produce a :class:`.Date`, assuming
+            "days delta" behavior by the database (in reality, most databases
+            other than PostgreSQL don't accept this particular operation).
+
+            The method returns a tuple of the form <operator>, <type>.
+            The resulting operator and type will be those applied to the
+            resulting :class:`.BinaryExpression` as the final operator and the
+            right-hand side of the expression.
+
+            Note that only a subset of operators make usage of
+            :meth:`._adapt_expression`,
+            including math operators and user-defined operators, but not
+            boolean comparison or special SQL keywords like MATCH or BETWEEN.
+
+            """
+
+            return op, self.type
+
+    hashable = True
+    """Flag, if False, means values from this type aren't hashable.
+
+    Used by the ORM when uniquing result lists.
+
+    """
+
+    comparator_factory: _ComparatorFactory[Any] = Comparator
+    """A :class:`.TypeEngine.Comparator` class which will apply
+    to operations performed by owning :class:`_expression.ColumnElement`
+    objects.
+
+    The :attr:`.comparator_factory` attribute is a hook consulted by
+    the core expression system when column and SQL expression operations
+    are performed.   When a :class:`.TypeEngine.Comparator` class is
+    associated with this attribute, it allows custom re-definition of
+    all existing operators, as well as definition of new operators.
+    Existing operators include those provided by Python operator overloading
+    such as :meth:`.operators.ColumnOperators.__add__` and
+    :meth:`.operators.ColumnOperators.__eq__`,
+    those provided as standard
+    attributes of :class:`.operators.ColumnOperators` such as
+    :meth:`.operators.ColumnOperators.like`
+    and :meth:`.operators.ColumnOperators.in_`.
+
+    Rudimentary usage of this hook is allowed through simple subclassing
+    of existing types, or alternatively by using :class:`.TypeDecorator`.
+    See the documentation section :ref:`types_operators` for examples.
+
+    """
+
+    sort_key_function: Optional[Callable[[Any], Any]] = None
+    """A sorting function that can be passed as the key to sorted.
+
+    The default value of ``None`` indicates that the values stored by
+    this type are self-sorting.
+
+    .. versionadded:: 1.3.8
+
+    """
+
+    should_evaluate_none: bool = False
+    """If True, the Python constant ``None`` is considered to be handled
+    explicitly by this type.
+
+    The ORM uses this flag to indicate that a positive value of ``None``
+    is passed to the column in an INSERT statement, rather than omitting
+    the column from the INSERT statement which has the effect of firing
+    off column-level defaults.   It also allows types which have special
+    behavior for Python None, such as a JSON type, to indicate that
+    they'd like to handle the None value explicitly.
+
+    To set this flag on an existing type, use the
+    :meth:`.TypeEngine.evaluates_none` method.
+
+    .. seealso::
+
+        :meth:`.TypeEngine.evaluates_none`
+
+    """
+
+    _variant_mapping: util.immutabledict[str, TypeEngine[Any]] = (
+        util.EMPTY_DICT
+    )
+
+    def evaluates_none(self) -> Self:
+        """Return a copy of this type which has the
+        :attr:`.should_evaluate_none` flag set to True.
+
+        E.g.::
+
+                Table(
+                    "some_table",
+                    metadata,
+                    Column(
+                        String(50).evaluates_none(),
+                        nullable=True,
+                        server_default="no value",
+                    ),
+                )
+
+        The ORM uses this flag to indicate that a positive value of ``None``
+        is passed to the column in an INSERT statement, rather than omitting
+        the column from the INSERT statement which has the effect of firing
+        off column-level defaults.   It also allows for types which have
+        special behavior associated with the Python None value to indicate
+        that the value doesn't necessarily translate into SQL NULL; a
+        prime example of this is a JSON type which may wish to persist the
+        JSON value ``'null'``.
+
+        In all cases, the actual NULL SQL value can be always be
+        persisted in any column by using
+        the :obj:`_expression.null` SQL construct in an INSERT statement
+        or associated with an ORM-mapped attribute.
+
+        .. note::
+
+            The "evaluates none" flag does **not** apply to a value
+            of ``None`` passed to :paramref:`_schema.Column.default` or
+            :paramref:`_schema.Column.server_default`; in these cases,
+            ``None``
+            still means "no default".
+
+        .. seealso::
+
+            :ref:`session_forcing_null` - in the ORM documentation
+
+            :paramref:`.postgresql.JSON.none_as_null` - PostgreSQL JSON
+            interaction with this flag.
+
+            :attr:`.TypeEngine.should_evaluate_none` - class-level flag
+
+        """
+        typ = self.copy()
+        typ.should_evaluate_none = True
+        return typ
+
+    def copy(self, **kw: Any) -> Self:
+        return self.adapt(self.__class__)
+
+    def copy_value(self, value: Any) -> Any:
+        return value
+
+    def literal_processor(
+        self, dialect: Dialect
+    ) -> Optional[_LiteralProcessorType[_T]]:
+        """Return a conversion function for processing literal values that are
+        to be rendered directly without using binds.
+
+        This function is used when the compiler makes use of the
+        "literal_binds" flag, typically used in DDL generation as well
+        as in certain scenarios where backends don't accept bound parameters.
+
+        Returns a callable which will receive a literal Python value
+        as the sole positional argument and will return a string representation
+        to be rendered in a SQL statement.
+
+        .. note::
+
+            This method is only called relative to a **dialect specific type
+            object**, which is often **private to a dialect in use** and is not
+            the same type object as the public facing one, which means it's not
+            feasible to subclass a :class:`.types.TypeEngine` class in order to
+            provide an alternate :meth:`_types.TypeEngine.literal_processor`
+            method, unless subclassing the :class:`_types.UserDefinedType`
+            class explicitly.
+
+            To provide alternate behavior for
+            :meth:`_types.TypeEngine.literal_processor`, implement a
+            :class:`_types.TypeDecorator` class and provide an implementation
+            of :meth:`_types.TypeDecorator.process_literal_param`.
+
+            .. seealso::
+
+                :ref:`types_typedecorator`
+
+
+        """
+        return None
+
+    def bind_processor(
+        self, dialect: Dialect
+    ) -> Optional[_BindProcessorType[_T]]:
+        """Return a conversion function for processing bind values.
+
+        Returns a callable which will receive a bind parameter value
+        as the sole positional argument and will return a value to
+        send to the DB-API.
+
+        If processing is not necessary, the method should return ``None``.
+
+        .. note::
+
+            This method is only called relative to a **dialect specific type
+            object**, which is often **private to a dialect in use** and is not
+            the same type object as the public facing one, which means it's not
+            feasible to subclass a :class:`.types.TypeEngine` class in order to
+            provide an alternate :meth:`_types.TypeEngine.bind_processor`
+            method, unless subclassing the :class:`_types.UserDefinedType`
+            class explicitly.
+
+            To provide alternate behavior for
+            :meth:`_types.TypeEngine.bind_processor`, implement a
+            :class:`_types.TypeDecorator` class and provide an implementation
+            of :meth:`_types.TypeDecorator.process_bind_param`.
+
+            .. seealso::
+
+                :ref:`types_typedecorator`
+
+
+        :param dialect: Dialect instance in use.
+
+        """
+        return None
+
+    def result_processor(
+        self, dialect: Dialect, coltype: object
+    ) -> Optional[_ResultProcessorType[_T]]:
+        """Return a conversion function for processing result row values.
+
+        Returns a callable which will receive a result row column
+        value as the sole positional argument and will return a value
+        to return to the user.
+
+        If processing is not necessary, the method should return ``None``.
+
+        .. note::
+
+            This method is only called relative to a **dialect specific type
+            object**, which is often **private to a dialect in use** and is not
+            the same type object as the public facing one, which means it's not
+            feasible to subclass a :class:`.types.TypeEngine` class in order to
+            provide an alternate :meth:`_types.TypeEngine.result_processor`
+            method, unless subclassing the :class:`_types.UserDefinedType`
+            class explicitly.
+
+            To provide alternate behavior for
+            :meth:`_types.TypeEngine.result_processor`, implement a
+            :class:`_types.TypeDecorator` class and provide an implementation
+            of :meth:`_types.TypeDecorator.process_result_value`.
+
+            .. seealso::
+
+                :ref:`types_typedecorator`
+
+        :param dialect: Dialect instance in use.
+
+        :param coltype: DBAPI coltype argument received in cursor.description.
+
+        """
+        return None
+
+    def column_expression(
+        self, colexpr: ColumnElement[_T]
+    ) -> Optional[ColumnElement[_T]]:
+        """Given a SELECT column expression, return a wrapping SQL expression.
+
+        This is typically a SQL function that wraps a column expression
+        as rendered in the columns clause of a SELECT statement.
+        It is used for special data types that require
+        columns to be wrapped in some special database function in order
+        to coerce the value before being sent back to the application.
+        It is the SQL analogue of the :meth:`.TypeEngine.result_processor`
+        method.
+
+        This method is called during the **SQL compilation** phase of a
+        statement, when rendering a SQL string. It is **not** called
+        against specific values.
+
+        .. note::
+
+            This method is only called relative to a **dialect specific type
+            object**, which is often **private to a dialect in use** and is not
+            the same type object as the public facing one, which means it's not
+            feasible to subclass a :class:`.types.TypeEngine` class in order to
+            provide an alternate :meth:`_types.TypeEngine.column_expression`
+            method, unless subclassing the :class:`_types.UserDefinedType`
+            class explicitly.
+
+            To provide alternate behavior for
+            :meth:`_types.TypeEngine.column_expression`, implement a
+            :class:`_types.TypeDecorator` class and provide an implementation
+            of :meth:`_types.TypeDecorator.column_expression`.
+
+            .. seealso::
+
+                :ref:`types_typedecorator`
+
+
+        .. seealso::
+
+            :ref:`types_sql_value_processing`
+
+        """
+
+        return None
+
+    @util.memoized_property
+    def _has_column_expression(self) -> bool:
+        """memoized boolean, check if column_expression is implemented.
+
+        Allows the method to be skipped for the vast majority of expression
+        types that don't use this feature.
+
+        """
+
+        return (
+            self.__class__.column_expression.__code__
+            is not TypeEngine.column_expression.__code__
+        )
+
+    def bind_expression(
+        self, bindvalue: BindParameter[_T]
+    ) -> Optional[ColumnElement[_T]]:
+        """Given a bind value (i.e. a :class:`.BindParameter` instance),
+        return a SQL expression in its place.
+
+        This is typically a SQL function that wraps the existing bound
+        parameter within the statement.  It is used for special data types
+        that require literals being wrapped in some special database function
+        in order to coerce an application-level value into a database-specific
+        format.  It is the SQL analogue of the
+        :meth:`.TypeEngine.bind_processor` method.
+
+        This method is called during the **SQL compilation** phase of a
+        statement, when rendering a SQL string. It is **not** called
+        against specific values.
+
+        Note that this method, when implemented, should always return
+        the exact same structure, without any conditional logic, as it
+        may be used in an executemany() call against an arbitrary number
+        of bound parameter sets.
+
+        .. note::
+
+            This method is only called relative to a **dialect specific type
+            object**, which is often **private to a dialect in use** and is not
+            the same type object as the public facing one, which means it's not
+            feasible to subclass a :class:`.types.TypeEngine` class in order to
+            provide an alternate :meth:`_types.TypeEngine.bind_expression`
+            method, unless subclassing the :class:`_types.UserDefinedType`
+            class explicitly.
+
+            To provide alternate behavior for
+            :meth:`_types.TypeEngine.bind_expression`, implement a
+            :class:`_types.TypeDecorator` class and provide an implementation
+            of :meth:`_types.TypeDecorator.bind_expression`.
+
+            .. seealso::
+
+                :ref:`types_typedecorator`
+
+        .. seealso::
+
+            :ref:`types_sql_value_processing`
+
+        """
+        return None
+
+    @util.memoized_property
+    def _has_bind_expression(self) -> bool:
+        """memoized boolean, check if bind_expression is implemented.
+
+        Allows the method to be skipped for the vast majority of expression
+        types that don't use this feature.
+
+        """
+
+        return util.method_is_overridden(self, TypeEngine.bind_expression)
+
+    @staticmethod
+    def _to_instance(cls_or_self: Union[Type[_TE], _TE]) -> _TE:
+        return to_instance(cls_or_self)
+
+    def compare_values(self, x: Any, y: Any) -> bool:
+        """Compare two values for equality."""
+
+        return x == y  # type: ignore[no-any-return]
+
+    def get_dbapi_type(self, dbapi: ModuleType) -> Optional[Any]:
+        """Return the corresponding type object from the underlying DB-API, if
+        any.
+
+        This can be useful for calling ``setinputsizes()``, for example.
+
+        """
+        return None
+
+    @property
+    def python_type(self) -> Type[Any]:
+        """Return the Python type object expected to be returned
+        by instances of this type, if known.
+
+        Basically, for those types which enforce a return type,
+        or are known across the board to do such for all common
+        DBAPIs (like ``int`` for example), will return that type.
+
+        If a return type is not defined, raises
+        ``NotImplementedError``.
+
+        Note that any type also accommodates NULL in SQL which
+        means you can also get back ``None`` from any type
+        in practice.
+
+        """
+        raise NotImplementedError()
+
+    def with_variant(
+        self,
+        type_: _TypeEngineArgument[Any],
+        *dialect_names: str,
+    ) -> Self:
+        r"""Produce a copy of this type object that will utilize the given
+        type when applied to the dialect of the given name.
+
+        e.g.::
+
+            from sqlalchemy.types import String
+            from sqlalchemy.dialects import mysql
+
+            string_type = String()
+
+            string_type = string_type.with_variant(
+                mysql.VARCHAR(collation="foo"), "mysql", "mariadb"
+            )
+
+        The variant mapping indicates that when this type is
+        interpreted by a specific dialect, it will instead be
+        transmuted into the given type, rather than using the
+        primary type.
+
+        .. versionchanged:: 2.0 the :meth:`_types.TypeEngine.with_variant`
+           method now works with a :class:`_types.TypeEngine` object "in
+           place", returning a copy of the original type rather than returning
+           a wrapping object; the ``Variant`` class is no longer used.
+
+        :param type\_: a :class:`.TypeEngine` that will be selected
+         as a variant from the originating type, when a dialect
+         of the given name is in use.
+        :param \*dialect_names: one or more base names of the dialect which
+         uses this type. (i.e. ``'postgresql'``, ``'mysql'``, etc.)
+
+         .. versionchanged:: 2.0 multiple dialect names can be specified
+            for one variant.
+
+        .. seealso::
+
+            :ref:`types_with_variant` - illustrates the use of
+            :meth:`_types.TypeEngine.with_variant`.
+
+        """
+
+        if not dialect_names:
+            raise exc.ArgumentError("At least one dialect name is required")
+        for dialect_name in dialect_names:
+            if dialect_name in self._variant_mapping:
+                raise exc.ArgumentError(
+                    f"Dialect {dialect_name!r} is already present in "
+                    f"the mapping for this {self!r}"
+                )
+        new_type = self.copy()
+        type_ = to_instance(type_)
+        if type_._variant_mapping:
+            raise exc.ArgumentError(
+                "can't pass a type that already has variants as a "
+                "dialect-level type to with_variant()"
+            )
+
+        new_type._variant_mapping = self._variant_mapping.union(
+            {dialect_name: type_ for dialect_name in dialect_names}
+        )
+        return new_type
+
+    def _resolve_for_literal(self, value: Any) -> Self:
+        """adjust this type given a literal Python value that will be
+        stored in a bound parameter.
+
+        Used exclusively by _resolve_value_to_type().
+
+        .. versionadded:: 1.4.30 or 2.0
+
+        TODO: this should be part of public API
+
+        .. seealso::
+
+            :meth:`.TypeEngine._resolve_for_python_type`
+
+        """
+        return self
+
+    def _resolve_for_python_type(
+        self,
+        python_type: Type[Any],
+        matched_on: _MatchedOnType,
+        matched_on_flattened: Type[Any],
+    ) -> Optional[Self]:
+        """given a Python type (e.g. ``int``, ``str``, etc. ) return an
+        instance of this :class:`.TypeEngine` that's appropriate for this type.
+
+        An additional argument ``matched_on`` is passed, which indicates an
+        entry from the ``__mro__`` of the given ``python_type`` that more
+        specifically matches how the caller located this :class:`.TypeEngine`
+        object.   Such as, if a lookup of some kind links the ``int`` Python
+        type to the :class:`.Integer` SQL type, and the original object
+        was some custom subclass of ``int`` such as ``MyInt(int)``, the
+        arguments passed would be ``(MyInt, int)``.
+
+        If the given Python type does not correspond to this
+        :class:`.TypeEngine`, or the Python type is otherwise ambiguous, the
+        method should return None.
+
+        For simple cases, the method checks that the ``python_type``
+        and ``matched_on`` types are the same (i.e. not a subclass), and
+        returns self; for all other cases, it returns ``None``.
+
+        The initial use case here is for the ORM to link user-defined
+        Python standard library ``enum.Enum`` classes to the SQLAlchemy
+        :class:`.Enum` SQL type when constructing ORM Declarative mappings.
+
+        :param python_type: the Python type we want to use
+        :param matched_on: the Python type that led us to choose this
+         particular :class:`.TypeEngine` class, which would be a supertype
+         of ``python_type``.   By default, the request is rejected if
+         ``python_type`` doesn't match ``matched_on`` (None is returned).
+
+        .. versionadded:: 2.0.0b4
+
+        TODO: this should be part of public API
+
+        .. seealso::
+
+            :meth:`.TypeEngine._resolve_for_literal`
+
+        """
+
+        if python_type is not matched_on_flattened:
+            return None
+
+        return self
+
+    def _with_collation(self, collation: str) -> Self:
+        """set up error handling for the collate expression"""
+        raise NotImplementedError("this datatype does not support collation")
+
+    @util.ro_memoized_property
+    def _type_affinity(self) -> Optional[Type[TypeEngine[_T]]]:
+        """Return a rudimental 'affinity' value expressing the general class
+        of type."""
+
+        typ = None
+        for t in self.__class__.__mro__:
+            if t is TypeEngine or TypeEngineMixin in t.__bases__:
+                return typ
+            elif issubclass(t, TypeEngine):
+                typ = t
+        else:
+            return self.__class__
+
+    @util.ro_memoized_property
+    def _generic_type_affinity(
+        self,
+    ) -> Type[TypeEngine[_T]]:
+        best_camelcase = None
+        best_uppercase = None
+
+        if not isinstance(self, TypeEngine):
+            return self.__class__
+
+        for t in self.__class__.__mro__:
+            if (
+                t.__module__
+                in (
+                    "sqlalchemy.sql.sqltypes",
+                    "sqlalchemy.sql.type_api",
+                )
+                and issubclass(t, TypeEngine)
+                and TypeEngineMixin not in t.__bases__
+                and t not in (TypeEngine, TypeEngineMixin)
+                and t.__name__[0] != "_"
+            ):
+                if t.__name__.isupper() and not best_uppercase:
+                    best_uppercase = t
+                elif not t.__name__.isupper() and not best_camelcase:
+                    best_camelcase = t
+
+        return (
+            best_camelcase
+            or best_uppercase
+            or cast("Type[TypeEngine[_T]]", NULLTYPE.__class__)
+        )
+
+    def as_generic(self, allow_nulltype: bool = False) -> TypeEngine[_T]:
+        """
+        Return an instance of the generic type corresponding to this type
+        using heuristic rule. The method may be overridden if this
+        heuristic rule is not sufficient.
+
+        >>> from sqlalchemy.dialects.mysql import INTEGER
+        >>> INTEGER(display_width=4).as_generic()
+        Integer()
+
+        >>> from sqlalchemy.dialects.mysql import NVARCHAR
+        >>> NVARCHAR(length=100).as_generic()
+        Unicode(length=100)
+
+        .. versionadded:: 1.4.0b2
+
+
+        .. seealso::
+
+            :ref:`metadata_reflection_dbagnostic_types` - describes the
+            use of :meth:`_types.TypeEngine.as_generic` in conjunction with
+            the :meth:`_sql.DDLEvents.column_reflect` event, which is its
+            intended use.
+
+        """
+        if (
+            not allow_nulltype
+            and self._generic_type_affinity == NULLTYPE.__class__
+        ):
+            raise NotImplementedError(
+                "Default TypeEngine.as_generic() "
+                "heuristic method was unsuccessful for {}. A custom "
+                "as_generic() method must be implemented for this "
+                "type class.".format(
+                    self.__class__.__module__ + "." + self.__class__.__name__
+                )
+            )
+
+        return util.constructor_copy(self, self._generic_type_affinity)
+
+    def dialect_impl(self, dialect: Dialect) -> TypeEngine[_T]:
+        """Return a dialect-specific implementation for this
+        :class:`.TypeEngine`.
+
+        """
+        try:
+            tm = dialect._type_memos[self]
+        except KeyError:
+            pass
+        else:
+            return tm["impl"]
+        return self._dialect_info(dialect)["impl"]
+
+    def _unwrapped_dialect_impl(self, dialect: Dialect) -> TypeEngine[_T]:
+        """Return the 'unwrapped' dialect impl for this type.
+
+        For a type that applies wrapping logic (e.g. TypeDecorator), give
+        us the real, actual dialect-level type that is used.
+
+        This is used by TypeDecorator itself as well at least one case where
+        dialects need to check that a particular specific dialect-level
+        type is in use, within the :meth:`.DefaultDialect.set_input_sizes`
+        method.
+
+        """
+        return self.dialect_impl(dialect)
+
+    def _cached_literal_processor(
+        self, dialect: Dialect
+    ) -> Optional[_LiteralProcessorType[_T]]:
+        """Return a dialect-specific literal processor for this type."""
+
+        try:
+            return dialect._type_memos[self]["literal"]
+        except KeyError:
+            pass
+
+        # avoid KeyError context coming into literal_processor() function
+        # raises
+        d = self._dialect_info(dialect)
+        d["literal"] = lp = d["impl"].literal_processor(dialect)
+        return lp
+
+    def _cached_bind_processor(
+        self, dialect: Dialect
+    ) -> Optional[_BindProcessorType[_T]]:
+        """Return a dialect-specific bind processor for this type."""
+
+        try:
+            return dialect._type_memos[self]["bind"]
+        except KeyError:
+            pass
+
+        # avoid KeyError context coming into bind_processor() function
+        # raises
+        d = self._dialect_info(dialect)
+        d["bind"] = bp = d["impl"].bind_processor(dialect)
+        return bp
+
+    def _cached_result_processor(
+        self, dialect: Dialect, coltype: Any
+    ) -> Optional[_ResultProcessorType[_T]]:
+        """Return a dialect-specific result processor for this type."""
+
+        try:
+            return dialect._type_memos[self]["result"][coltype]
+        except KeyError:
+            pass
+
+        # avoid KeyError context coming into result_processor() function
+        # raises
+        d = self._dialect_info(dialect)
+        # key assumption: DBAPI type codes are
+        # constants.  Else this dictionary would
+        # grow unbounded.
+        rp = d["impl"].result_processor(dialect, coltype)
+        d["result"][coltype] = rp
+        return rp
+
+    def _cached_custom_processor(
+        self, dialect: Dialect, key: str, fn: Callable[[TypeEngine[_T]], _O]
+    ) -> _O:
+        """return a dialect-specific processing object for
+        custom purposes.
+
+        The cx_Oracle dialect uses this at the moment.
+
+        """
+        try:
+            return cast(_O, dialect._type_memos[self]["custom"][key])
+        except KeyError:
+            pass
+        # avoid KeyError context coming into fn() function
+        # raises
+        d = self._dialect_info(dialect)
+        impl = d["impl"]
+        custom_dict = d.setdefault("custom", {})
+        custom_dict[key] = result = fn(impl)
+        return result
+
+    def _dialect_info(self, dialect: Dialect) -> _TypeMemoDict:
+        """Return a dialect-specific registry which
+        caches a dialect-specific implementation, bind processing
+        function, and one or more result processing functions."""
+
+        if self in dialect._type_memos:
+            return dialect._type_memos[self]
+        else:
+            impl = self._gen_dialect_impl(dialect)
+            if impl is self:
+                impl = self.adapt(type(self))
+            # this can't be self, else we create a cycle
+            assert impl is not self
+            d: _TypeMemoDict = {"impl": impl, "result": {}}
+            dialect._type_memos[self] = d
+            return d
+
+    def _gen_dialect_impl(self, dialect: Dialect) -> TypeEngine[Any]:
+        if dialect.name in self._variant_mapping:
+            return self._variant_mapping[dialect.name]._gen_dialect_impl(
+                dialect
+            )
+        else:
+            return dialect.type_descriptor(self)
+
+    @util.memoized_property
+    def _static_cache_key(
+        self,
+    ) -> Union[CacheConst, Tuple[Any, ...]]:
+        names = util.get_cls_kwargs(self.__class__)
+        return (self.__class__,) + tuple(
+            (
+                k,
+                (
+                    self.__dict__[k]._static_cache_key
+                    if isinstance(self.__dict__[k], TypeEngine)
+                    else self.__dict__[k]
+                ),
+            )
+            for k in names
+            if k in self.__dict__
+            and not k.startswith("_")
+            and self.__dict__[k] is not None
+        )
+
+    @overload
+    def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ...
+
+    @overload
+    def adapt(
+        self, cls: Type[TypeEngineMixin], **kw: Any
+    ) -> TypeEngine[Any]: ...
+
+    def adapt(
+        self, cls: Type[Union[TypeEngine[Any], TypeEngineMixin]], **kw: Any
+    ) -> TypeEngine[Any]:
+        """Produce an "adapted" form of this type, given an "impl" class
+        to work with.
+
+        This method is used internally to associate generic
+        types with "implementation" types that are specific to a particular
+        dialect.
+        """
+        typ = util.constructor_copy(
+            self, cast(Type[TypeEngine[Any]], cls), **kw
+        )
+        typ._variant_mapping = self._variant_mapping
+        return typ
+
+    def coerce_compared_value(
+        self, op: Optional[OperatorType], value: Any
+    ) -> TypeEngine[Any]:
+        """Suggest a type for a 'coerced' Python value in an expression.
+
+        Given an operator and value, gives the type a chance
+        to return a type which the value should be coerced into.
+
+        The default behavior here is conservative; if the right-hand
+        side is already coerced into a SQL type based on its
+        Python type, it is usually left alone.
+
+        End-user functionality extension here should generally be via
+        :class:`.TypeDecorator`, which provides more liberal behavior in that
+        it defaults to coercing the other side of the expression into this
+        type, thus applying special Python conversions above and beyond those
+        needed by the DBAPI to both ides. It also provides the public method
+        :meth:`.TypeDecorator.coerce_compared_value` which is intended for
+        end-user customization of this behavior.
+
+        """
+        _coerced_type = _resolve_value_to_type(value)
+        if (
+            _coerced_type is NULLTYPE
+            or _coerced_type._type_affinity is self._type_affinity
+        ):
+            return self
+        else:
+            return _coerced_type
+
+    def _compare_type_affinity(self, other: TypeEngine[Any]) -> bool:
+        return self._type_affinity is other._type_affinity
+
+    def compile(self, dialect: Optional[Dialect] = None) -> str:
+        """Produce a string-compiled form of this :class:`.TypeEngine`.
+
+        When called with no arguments, uses a "default" dialect
+        to produce a string result.
+
+        :param dialect: a :class:`.Dialect` instance.
+
+        """
+        # arg, return value is inconsistent with
+        # ClauseElement.compile()....this is a mistake.
+
+        if dialect is None:
+            dialect = self._default_dialect()
+
+        return dialect.type_compiler_instance.process(self)
+
+    @util.preload_module("sqlalchemy.engine.default")
+    def _default_dialect(self) -> Dialect:
+        default = util.preloaded.engine_default
+
+        # dmypy / mypy seems to sporadically keep thinking this line is
+        # returning Any, which seems to be caused by the @deprecated_params
+        # decorator on the DefaultDialect constructor
+        return default.StrCompileDialect()  # type: ignore
+
+    def __str__(self) -> str:
+        return str(self.compile())
+
+    def __repr__(self) -> str:
+        return util.generic_repr(self)
+
+
+class TypeEngineMixin:
+    """classes which subclass this can act as "mixin" classes for
+    TypeEngine."""
+
+    __slots__ = ()
+
+    if TYPE_CHECKING:
+
+        @util.memoized_property
+        def _static_cache_key(
+            self,
+        ) -> Union[CacheConst, Tuple[Any, ...]]: ...
+
+        @overload
+        def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ...
+
+        @overload
+        def adapt(
+            self, cls: Type[TypeEngineMixin], **kw: Any
+        ) -> TypeEngine[Any]: ...
+
+        def adapt(
+            self, cls: Type[Union[TypeEngine[Any], TypeEngineMixin]], **kw: Any
+        ) -> TypeEngine[Any]: ...
+
+        def dialect_impl(self, dialect: Dialect) -> TypeEngine[Any]: ...
+
+
+class ExternalType(TypeEngineMixin):
+    """mixin that defines attributes and behaviors specific to third-party
+    datatypes.
+
+    "Third party" refers to datatypes that are defined outside the scope
+    of SQLAlchemy within either end-user application code or within
+    external extensions to SQLAlchemy.
+
+    Subclasses currently include :class:`.TypeDecorator` and
+    :class:`.UserDefinedType`.
+
+    .. versionadded:: 1.4.28
+
+    """
+
+    cache_ok: Optional[bool] = None
+    '''Indicate if statements using this :class:`.ExternalType` are "safe to
+    cache".
+
+    The default value ``None`` will emit a warning and then not allow caching
+    of a statement which includes this type.   Set to ``False`` to disable
+    statements using this type from being cached at all without a warning.
+    When set to ``True``, the object's class and selected elements from its
+    state will be used as part of the cache key.  For example, using a
+    :class:`.TypeDecorator`::
+
+        class MyType(TypeDecorator):
+            impl = String
+
+            cache_ok = True
+
+            def __init__(self, choices):
+                self.choices = tuple(choices)
+                self.internal_only = True
+
+    The cache key for the above type would be equivalent to::
+
+        >>> MyType(["a", "b", "c"])._static_cache_key
+        (<class '__main__.MyType'>, ('choices', ('a', 'b', 'c')))
+
+    The caching scheme will extract attributes from the type that correspond
+    to the names of parameters in the ``__init__()`` method.  Above, the
+    "choices" attribute becomes part of the cache key but "internal_only"
+    does not, because there is no parameter named "internal_only".
+
+    The requirements for cacheable elements is that they are hashable
+    and also that they indicate the same SQL rendered for expressions using
+    this type every time for a given cache value.
+
+    To accommodate for datatypes that refer to unhashable structures such
+    as dictionaries, sets and lists, these objects can be made "cacheable"
+    by assigning hashable structures to the attributes whose names
+    correspond with the names of the arguments.  For example, a datatype
+    which accepts a dictionary of lookup values may publish this as a sorted
+    series of tuples.   Given a previously un-cacheable type as::
+
+        class LookupType(UserDefinedType):
+            """a custom type that accepts a dictionary as a parameter.
+
+            this is the non-cacheable version, as "self.lookup" is not
+            hashable.
+
+            """
+
+            def __init__(self, lookup):
+                self.lookup = lookup
+
+            def get_col_spec(self, **kw):
+                return "VARCHAR(255)"
+
+            def bind_processor(self, dialect): ...  # works with "self.lookup" ...
+
+    Where "lookup" is a dictionary.  The type will not be able to generate
+    a cache key::
+
+        >>> type_ = LookupType({"a": 10, "b": 20})
+        >>> type_._static_cache_key
+        <stdin>:1: SAWarning: UserDefinedType LookupType({'a': 10, 'b': 20}) will not
+        produce a cache key because the ``cache_ok`` flag is not set to True.
+        Set this flag to True if this type object's state is safe to use
+        in a cache key, or False to disable this warning.
+        symbol('no_cache')
+
+    If we **did** set up such a cache key, it wouldn't be usable. We would
+    get a tuple structure that contains a dictionary inside of it, which
+    cannot itself be used as a key in a "cache dictionary" such as SQLAlchemy's
+    statement cache, since Python dictionaries aren't hashable::
+
+        >>> # set cache_ok = True
+        >>> type_.cache_ok = True
+
+        >>> # this is the cache key it would generate
+        >>> key = type_._static_cache_key
+        >>> key
+        (<class '__main__.LookupType'>, ('lookup', {'a': 10, 'b': 20}))
+
+        >>> # however this key is not hashable, will fail when used with
+        >>> # SQLAlchemy statement cache
+        >>> some_cache = {key: "some sql value"}
+        Traceback (most recent call last): File "<stdin>", line 1,
+        in <module> TypeError: unhashable type: 'dict'
+
+    The type may be made cacheable by assigning a sorted tuple of tuples
+    to the ".lookup" attribute::
+
+        class LookupType(UserDefinedType):
+            """a custom type that accepts a dictionary as a parameter.
+
+            The dictionary is stored both as itself in a private variable,
+            and published in a public variable as a sorted tuple of tuples,
+            which is hashable and will also return the same value for any
+            two equivalent dictionaries.  Note it assumes the keys and
+            values of the dictionary are themselves hashable.
+
+            """
+
+            cache_ok = True
+
+            def __init__(self, lookup):
+                self._lookup = lookup
+
+                # assume keys/values of "lookup" are hashable; otherwise
+                # they would also need to be converted in some way here
+                self.lookup = tuple((key, lookup[key]) for key in sorted(lookup))
+
+            def get_col_spec(self, **kw):
+                return "VARCHAR(255)"
+
+            def bind_processor(self, dialect): ...  # works with "self._lookup" ...
+
+    Where above, the cache key for ``LookupType({"a": 10, "b": 20})`` will be::
+
+        >>> LookupType({"a": 10, "b": 20})._static_cache_key
+        (<class '__main__.LookupType'>, ('lookup', (('a', 10), ('b', 20))))
+
+    .. versionadded:: 1.4.14 - added the ``cache_ok`` flag to allow
+       some configurability of caching for :class:`.TypeDecorator` classes.
+
+    .. versionadded:: 1.4.28 - added the :class:`.ExternalType` mixin which
+       generalizes the ``cache_ok`` flag to both the :class:`.TypeDecorator`
+       and :class:`.UserDefinedType` classes.
+
+    .. seealso::
+
+        :ref:`sql_caching`
+
+    '''  # noqa: E501
+
+    @util.non_memoized_property
+    def _static_cache_key(
+        self,
+    ) -> Union[CacheConst, Tuple[Any, ...]]:
+        cache_ok = self.__class__.__dict__.get("cache_ok", None)
+
+        if cache_ok is None:
+            for subtype in self.__class__.__mro__:
+                if ExternalType in subtype.__bases__:
+                    break
+            else:
+                subtype = self.__class__.__mro__[1]
+
+            util.warn(
+                "%s %r will not produce a cache key because "
+                "the ``cache_ok`` attribute is not set to True.  This can "
+                "have significant performance implications including some "
+                "performance degradations in comparison to prior SQLAlchemy "
+                "versions.  Set this attribute to True if this type object's "
+                "state is safe to use in a cache key, or False to "
+                "disable this warning." % (subtype.__name__, self),
+                code="cprf",
+            )
+        elif cache_ok is True:
+            return super()._static_cache_key
+
+        return NO_CACHE
+
+
+class UserDefinedType(
+    ExternalType, TypeEngineMixin, TypeEngine[_T], util.EnsureKWArg
+):
+    """Base for user defined types.
+
+    This should be the base of new types.  Note that
+    for most cases, :class:`.TypeDecorator` is probably
+    more appropriate::
+
+      import sqlalchemy.types as types
+
+
+      class MyType(types.UserDefinedType):
+          cache_ok = True
+
+          def __init__(self, precision=8):
+              self.precision = precision
+
+          def get_col_spec(self, **kw):
+              return "MYTYPE(%s)" % self.precision
+
+          def bind_processor(self, dialect):
+              def process(value):
+                  return value
+
+              return process
+
+          def result_processor(self, dialect, coltype):
+              def process(value):
+                  return value
+
+              return process
+
+    Once the type is made, it's immediately usable::
+
+      table = Table(
+          "foo",
+          metadata_obj,
+          Column("id", Integer, primary_key=True),
+          Column("data", MyType(16)),
+      )
+
+    The ``get_col_spec()`` method will in most cases receive a keyword
+    argument ``type_expression`` which refers to the owning expression
+    of the type as being compiled, such as a :class:`_schema.Column` or
+    :func:`.cast` construct.  This keyword is only sent if the method
+    accepts keyword arguments (e.g. ``**kw``) in its argument signature;
+    introspection is used to check for this in order to support legacy
+    forms of this function.
+
+    The :attr:`.UserDefinedType.cache_ok` class-level flag indicates if this
+    custom :class:`.UserDefinedType` is safe to be used as part of a cache key.
+    This flag defaults to ``None`` which will initially generate a warning
+    when the SQL compiler attempts to generate a cache key for a statement
+    that uses this type.  If the :class:`.UserDefinedType` is not guaranteed
+    to produce the same bind/result behavior and SQL generation
+    every time, this flag should be set to ``False``; otherwise if the
+    class produces the same behavior each time, it may be set to ``True``.
+    See :attr:`.UserDefinedType.cache_ok` for further notes on how this works.
+
+    .. versionadded:: 1.4.28 Generalized the :attr:`.ExternalType.cache_ok`
+       flag so that it is available for both :class:`.TypeDecorator` as well
+       as :class:`.UserDefinedType`.
+
+    """
+
+    __visit_name__ = "user_defined"
+
+    ensure_kwarg = "get_col_spec"
+
+    def coerce_compared_value(
+        self, op: Optional[OperatorType], value: Any
+    ) -> TypeEngine[Any]:
+        """Suggest a type for a 'coerced' Python value in an expression.
+
+        Default behavior for :class:`.UserDefinedType` is the
+        same as that of :class:`.TypeDecorator`; by default it returns
+        ``self``, assuming the compared value should be coerced into
+        the same type as this one.  See
+        :meth:`.TypeDecorator.coerce_compared_value` for more detail.
+
+        """
+
+        return self
+
+
+class Emulated(TypeEngineMixin):
+    """Mixin for base types that emulate the behavior of a DB-native type.
+
+    An :class:`.Emulated` type will use an available database type
+    in conjunction with Python-side routines and/or database constraints
+    in order to approximate the behavior of a database type that is provided
+    natively by some backends.  When a native-providing backend is in
+    use, the native version of the type is used.  This native version
+    should include the :class:`.NativeForEmulated` mixin to allow it to be
+    distinguished from :class:`.Emulated`.
+
+    Current examples of :class:`.Emulated` are:  :class:`.Interval`,
+    :class:`.Enum`, :class:`.Boolean`.
+
+    .. versionadded:: 1.2.0b3
+
+    """
+
+    native: bool
+
+    def adapt_to_emulated(
+        self,
+        impltype: Type[Union[TypeEngine[Any], TypeEngineMixin]],
+        **kw: Any,
+    ) -> TypeEngine[Any]:
+        """Given an impl class, adapt this type to the impl assuming
+        "emulated".
+
+        The impl should also be an "emulated" version of this type,
+        most likely the same class as this type itself.
+
+        e.g.: sqltypes.Enum adapts to the Enum class.
+
+        """
+        return super().adapt(impltype, **kw)
+
+    @overload
+    def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ...
+
+    @overload
+    def adapt(
+        self, cls: Type[TypeEngineMixin], **kw: Any
+    ) -> TypeEngine[Any]: ...
+
+    def adapt(
+        self, cls: Type[Union[TypeEngine[Any], TypeEngineMixin]], **kw: Any
+    ) -> TypeEngine[Any]:
+        if _is_native_for_emulated(cls):
+            if self.native:
+                # native support requested, dialect gave us a native
+                # implementor, pass control over to it
+                return cls.adapt_emulated_to_native(self, **kw)
+            else:
+                # non-native support, let the native implementor
+                # decide also, at the moment this is just to help debugging
+                # as only the default logic is implemented.
+                return cls.adapt_native_to_emulated(self, **kw)
+        else:
+            # this would be, both classes are Enum, or both classes
+            # are postgresql.ENUM
+            if issubclass(cls, self.__class__):
+                return self.adapt_to_emulated(cls, **kw)
+            else:
+                return super().adapt(cls, **kw)
+
+
+def _is_native_for_emulated(
+    typ: Type[Union[TypeEngine[Any], TypeEngineMixin]],
+) -> TypeGuard[Type[NativeForEmulated]]:
+    return hasattr(typ, "adapt_emulated_to_native")
+
+
+class NativeForEmulated(TypeEngineMixin):
+    """Indicates DB-native types supported by an :class:`.Emulated` type.
+
+    .. versionadded:: 1.2.0b3
+
+    """
+
+    @classmethod
+    def adapt_native_to_emulated(
+        cls,
+        impl: Union[TypeEngine[Any], TypeEngineMixin],
+        **kw: Any,
+    ) -> TypeEngine[Any]:
+        """Given an impl, adapt this type's class to the impl assuming
+        "emulated".
+
+
+        """
+        impltype = impl.__class__
+        return impl.adapt(impltype, **kw)
+
+    @classmethod
+    def adapt_emulated_to_native(
+        cls,
+        impl: Union[TypeEngine[Any], TypeEngineMixin],
+        **kw: Any,
+    ) -> TypeEngine[Any]:
+        """Given an impl, adapt this type's class to the impl assuming
+        "native".
+
+        The impl will be an :class:`.Emulated` class but not a
+        :class:`.NativeForEmulated`.
+
+        e.g.: postgresql.ENUM produces a type given an Enum instance.
+
+        """
+
+        # dmypy seems to crash on this
+        return cls(**kw)  # type: ignore
+
+    # dmypy seems to crash with this, on repeated runs with changes
+    # if TYPE_CHECKING:
+    #    def __init__(self, **kw: Any):
+    #        ...
+
+
+class TypeDecorator(SchemaEventTarget, ExternalType, TypeEngine[_T]):
+    '''Allows the creation of types which add additional functionality
+    to an existing type.
+
+    This method is preferred to direct subclassing of SQLAlchemy's
+    built-in types as it ensures that all required functionality of
+    the underlying type is kept in place.
+
+    Typical usage::
+
+      import sqlalchemy.types as types
+
+
+      class MyType(types.TypeDecorator):
+          """Prefixes Unicode values with "PREFIX:" on the way in and
+          strips it off on the way out.
+          """
+
+          impl = types.Unicode
+
+          cache_ok = True
+
+          def process_bind_param(self, value, dialect):
+              return "PREFIX:" + value
+
+          def process_result_value(self, value, dialect):
+              return value[7:]
+
+          def copy(self, **kw):
+              return MyType(self.impl.length)
+
+    The class-level ``impl`` attribute is required, and can reference any
+    :class:`.TypeEngine` class.  Alternatively, the :meth:`load_dialect_impl`
+    method can be used to provide different type classes based on the dialect
+    given; in this case, the ``impl`` variable can reference
+    ``TypeEngine`` as a placeholder.
+
+    The :attr:`.TypeDecorator.cache_ok` class-level flag indicates if this
+    custom :class:`.TypeDecorator` is safe to be used as part of a cache key.
+    This flag defaults to ``None`` which will initially generate a warning
+    when the SQL compiler attempts to generate a cache key for a statement
+    that uses this type.  If the :class:`.TypeDecorator` is not guaranteed
+    to produce the same bind/result behavior and SQL generation
+    every time, this flag should be set to ``False``; otherwise if the
+    class produces the same behavior each time, it may be set to ``True``.
+    See :attr:`.TypeDecorator.cache_ok` for further notes on how this works.
+
+    Types that receive a Python type that isn't similar to the ultimate type
+    used may want to define the :meth:`TypeDecorator.coerce_compared_value`
+    method. This is used to give the expression system a hint when coercing
+    Python objects into bind parameters within expressions. Consider this
+    expression::
+
+        mytable.c.somecol + datetime.date(2009, 5, 15)
+
+    Above, if "somecol" is an ``Integer`` variant, it makes sense that
+    we're doing date arithmetic, where above is usually interpreted
+    by databases as adding a number of days to the given date.
+    The expression system does the right thing by not attempting to
+    coerce the "date()" value into an integer-oriented bind parameter.
+
+    However, in the case of ``TypeDecorator``, we are usually changing an
+    incoming Python type to something new - ``TypeDecorator`` by default will
+    "coerce" the non-typed side to be the same type as itself. Such as below,
+    we define an "epoch" type that stores a date value as an integer::
+
+        class MyEpochType(types.TypeDecorator):
+            impl = types.Integer
+
+            cache_ok = True
+
+            epoch = datetime.date(1970, 1, 1)
+
+            def process_bind_param(self, value, dialect):
+                return (value - self.epoch).days
+
+            def process_result_value(self, value, dialect):
+                return self.epoch + timedelta(days=value)
+
+    Our expression of ``somecol + date`` with the above type will coerce the
+    "date" on the right side to also be treated as ``MyEpochType``.
+
+    This behavior can be overridden via the
+    :meth:`~TypeDecorator.coerce_compared_value` method, which returns a type
+    that should be used for the value of the expression. Below we set it such
+    that an integer value will be treated as an ``Integer``, and any other
+    value is assumed to be a date and will be treated as a ``MyEpochType``::
+
+        def coerce_compared_value(self, op, value):
+            if isinstance(value, int):
+                return Integer()
+            else:
+                return self
+
+    .. warning::
+
+       Note that the **behavior of coerce_compared_value is not inherited
+       by default from that of the base type**.
+       If the :class:`.TypeDecorator` is augmenting a
+       type that requires special logic for certain types of operators,
+       this method **must** be overridden.  A key example is when decorating
+       the :class:`_postgresql.JSON` and :class:`_postgresql.JSONB` types;
+       the default rules of :meth:`.TypeEngine.coerce_compared_value` should
+       be used in order to deal with operators like index operations::
+
+            from sqlalchemy import JSON
+            from sqlalchemy import TypeDecorator
+
+
+            class MyJsonType(TypeDecorator):
+                impl = JSON
+
+                cache_ok = True
+
+                def coerce_compared_value(self, op, value):
+                    return self.impl.coerce_compared_value(op, value)
+
+       Without the above step, index operations such as ``mycol['foo']``
+       will cause the index value ``'foo'`` to be JSON encoded.
+
+       Similarly, when working with the :class:`.ARRAY` datatype, the
+       type coercion for index operations (e.g. ``mycol[5]``) is also
+       handled by :meth:`.TypeDecorator.coerce_compared_value`, where
+       again a simple override is sufficient unless special rules are needed
+       for particular operators::
+
+            from sqlalchemy import ARRAY
+            from sqlalchemy import TypeDecorator
+
+
+            class MyArrayType(TypeDecorator):
+                impl = ARRAY
+
+                cache_ok = True
+
+                def coerce_compared_value(self, op, value):
+                    return self.impl.coerce_compared_value(op, value)
+
+    '''
+
+    __visit_name__ = "type_decorator"
+
+    _is_type_decorator = True
+
+    # this is that pattern I've used in a few places (Dialect.dbapi,
+    # Dialect.type_compiler) where the "cls.attr" is a class to make something,
+    # and "instance.attr" is an instance of that thing.  It's such a nifty,
+    # great pattern, and there is zero chance Python typing tools will ever be
+    # OK with it.  For TypeDecorator.impl, this is a highly public attribute so
+    # we really can't change its behavior without a major deprecation routine.
+    impl: Union[TypeEngine[Any], Type[TypeEngine[Any]]]
+
+    # we are changing its behavior *slightly*, which is that we now consume
+    # the instance level version from this memoized property instead, so you
+    # can't reassign "impl" on an existing TypeDecorator that's already been
+    # used (something one shouldn't do anyway) without also updating
+    # impl_instance.
+    @util.memoized_property
+    def impl_instance(self) -> TypeEngine[Any]:
+        return self.impl  # type: ignore
+
+    def __init__(self, *args: Any, **kwargs: Any):
+        """Construct a :class:`.TypeDecorator`.
+
+        Arguments sent here are passed to the constructor
+        of the class assigned to the ``impl`` class level attribute,
+        assuming the ``impl`` is a callable, and the resulting
+        object is assigned to the ``self.impl`` instance attribute
+        (thus overriding the class attribute of the same name).
+
+        If the class level ``impl`` is not a callable (the unusual case),
+        it will be assigned to the same instance attribute 'as-is',
+        ignoring those arguments passed to the constructor.
+
+        Subclasses can override this to customize the generation
+        of ``self.impl`` entirely.
+
+        """
+
+        if not hasattr(self.__class__, "impl"):
+            raise AssertionError(
+                "TypeDecorator implementations "
+                "require a class-level variable "
+                "'impl' which refers to the class of "
+                "type being decorated"
+            )
+
+        self.impl = to_instance(self.__class__.impl, *args, **kwargs)
+
+    coerce_to_is_types: Sequence[Type[Any]] = (type(None),)
+    """Specify those Python types which should be coerced at the expression
+    level to "IS <constant>" when compared using ``==`` (and same for
+    ``IS NOT`` in conjunction with ``!=``).
+
+    For most SQLAlchemy types, this includes ``NoneType``, as well as
+    ``bool``.
+
+    :class:`.TypeDecorator` modifies this list to only include ``NoneType``,
+    as typedecorator implementations that deal with boolean types are common.
+
+    Custom :class:`.TypeDecorator` classes can override this attribute to
+    return an empty tuple, in which case no values will be coerced to
+    constants.
+
+    """
+
+    class Comparator(TypeEngine.Comparator[_CT]):
+        """A :class:`.TypeEngine.Comparator` that is specific to
+        :class:`.TypeDecorator`.
+
+        User-defined :class:`.TypeDecorator` classes should not typically
+        need to modify this.
+
+
+        """
+
+        __slots__ = ()
+
+        def operate(
+            self, op: OperatorType, *other: Any, **kwargs: Any
+        ) -> ColumnElement[_CT]:
+            if TYPE_CHECKING:
+                assert isinstance(self.expr.type, TypeDecorator)
+            kwargs["_python_is_types"] = self.expr.type.coerce_to_is_types
+            return super().operate(op, *other, **kwargs)
+
+        def reverse_operate(
+            self, op: OperatorType, other: Any, **kwargs: Any
+        ) -> ColumnElement[_CT]:
+            if TYPE_CHECKING:
+                assert isinstance(self.expr.type, TypeDecorator)
+            kwargs["_python_is_types"] = self.expr.type.coerce_to_is_types
+            return super().reverse_operate(op, other, **kwargs)
+
+    @staticmethod
+    def _reduce_td_comparator(
+        impl: TypeEngine[Any], expr: ColumnElement[_T]
+    ) -> Any:
+        return TypeDecorator._create_td_comparator_type(impl)(expr)
+
+    @staticmethod
+    def _create_td_comparator_type(
+        impl: TypeEngine[Any],
+    ) -> _ComparatorFactory[Any]:
+
+        def __reduce__(self: TypeDecorator.Comparator[Any]) -> Any:
+            return (TypeDecorator._reduce_td_comparator, (impl, self.expr))
+
+        return type(
+            "TDComparator",
+            (TypeDecorator.Comparator, impl.comparator_factory),  # type: ignore # noqa: E501
+            {"__reduce__": __reduce__},
+        )
+
+    @property
+    def comparator_factory(  # type: ignore  # mypy properties bug
+        self,
+    ) -> _ComparatorFactory[Any]:
+        if TypeDecorator.Comparator in self.impl.comparator_factory.__mro__:  # type: ignore # noqa: E501
+            return self.impl_instance.comparator_factory
+        else:
+            # reconcile the Comparator class on the impl with that
+            # of TypeDecorator.
+            # the use of multiple staticmethods is to support repeated
+            # pickling of the Comparator itself
+            return TypeDecorator._create_td_comparator_type(self.impl_instance)
+
+    def _copy_with_check(self) -> Self:
+        tt = self.copy()
+        if not isinstance(tt, self.__class__):
+            raise AssertionError(
+                "Type object %s does not properly "
+                "implement the copy() method, it must "
+                "return an object of type %s" % (self, self.__class__)
+            )
+        return tt
+
+    def _gen_dialect_impl(self, dialect: Dialect) -> TypeEngine[_T]:
+        if dialect.name in self._variant_mapping:
+            adapted = dialect.type_descriptor(
+                self._variant_mapping[dialect.name]
+            )
+        else:
+            adapted = dialect.type_descriptor(self)
+        if adapted is not self:
+            return adapted
+
+        # otherwise adapt the impl type, link
+        # to a copy of this TypeDecorator and return
+        # that.
+        typedesc = self.load_dialect_impl(dialect).dialect_impl(dialect)
+        tt = self._copy_with_check()
+        tt.impl = tt.impl_instance = typedesc
+        return tt
+
+    def _with_collation(self, collation: str) -> Self:
+        tt = self._copy_with_check()
+        tt.impl = tt.impl_instance = self.impl_instance._with_collation(
+            collation
+        )
+        return tt
+
+    @util.ro_non_memoized_property
+    def _type_affinity(self) -> Optional[Type[TypeEngine[Any]]]:
+        return self.impl_instance._type_affinity
+
+    def _set_parent(
+        self, parent: SchemaEventTarget, outer: bool = False, **kw: Any
+    ) -> None:
+        """Support SchemaEventTarget"""
+
+        super()._set_parent(parent)
+
+        if not outer and isinstance(self.impl_instance, SchemaEventTarget):
+            self.impl_instance._set_parent(parent, outer=False, **kw)
+
+    def _set_parent_with_dispatch(
+        self, parent: SchemaEventTarget, **kw: Any
+    ) -> None:
+        """Support SchemaEventTarget"""
+
+        super()._set_parent_with_dispatch(parent, outer=True, **kw)
+
+        if isinstance(self.impl_instance, SchemaEventTarget):
+            self.impl_instance._set_parent_with_dispatch(parent)
+
+    def type_engine(self, dialect: Dialect) -> TypeEngine[Any]:
+        """Return a dialect-specific :class:`.TypeEngine` instance
+        for this :class:`.TypeDecorator`.
+
+        In most cases this returns a dialect-adapted form of
+        the :class:`.TypeEngine` type represented by ``self.impl``.
+        Makes usage of :meth:`dialect_impl`.
+        Behavior can be customized here by overriding
+        :meth:`load_dialect_impl`.
+
+        """
+        adapted = dialect.type_descriptor(self)
+        if not isinstance(adapted, type(self)):
+            return adapted
+        else:
+            return self.load_dialect_impl(dialect)
+
+    def load_dialect_impl(self, dialect: Dialect) -> TypeEngine[Any]:
+        """Return a :class:`.TypeEngine` object corresponding to a dialect.
+
+        This is an end-user override hook that can be used to provide
+        differing types depending on the given dialect.  It is used
+        by the :class:`.TypeDecorator` implementation of :meth:`type_engine`
+        to help determine what type should ultimately be returned
+        for a given :class:`.TypeDecorator`.
+
+        By default returns ``self.impl``.
+
+        """
+        return self.impl_instance
+
+    def _unwrapped_dialect_impl(self, dialect: Dialect) -> TypeEngine[Any]:
+        """Return the 'unwrapped' dialect impl for this type.
+
+        This is used by the :meth:`.DefaultDialect.set_input_sizes`
+        method.
+
+        """
+        # some dialects have a lookup for a TypeDecorator subclass directly.
+        # postgresql.INTERVAL being the main example
+        typ = self.dialect_impl(dialect)
+
+        # if we are still a type decorator, load the per-dialect switch
+        # (such as what Variant uses), then get the dialect impl for that.
+        if isinstance(typ, self.__class__):
+            return typ.load_dialect_impl(dialect).dialect_impl(dialect)
+        else:
+            return typ
+
+    def __getattr__(self, key: str) -> Any:
+        """Proxy all other undefined accessors to the underlying
+        implementation."""
+        return getattr(self.impl_instance, key)
+
+    def process_literal_param(
+        self, value: Optional[_T], dialect: Dialect
+    ) -> str:
+        """Receive a literal parameter value to be rendered inline within
+        a statement.
+
+        .. note::
+
+            This method is called during the **SQL compilation** phase of a
+            statement, when rendering a SQL string. Unlike other SQL
+            compilation methods, it is passed a specific Python value to be
+            rendered as a string. However it should not be confused with the
+            :meth:`_types.TypeDecorator.process_bind_param` method, which is
+            the more typical method that processes the actual value passed to a
+            particular parameter at statement execution time.
+
+        Custom subclasses of :class:`_types.TypeDecorator` should override
+        this method to provide custom behaviors for incoming data values
+        that are in the special case of being rendered as literals.
+
+        The returned string will be rendered into the output string.
+
+        """
+        raise NotImplementedError()
+
+    def process_bind_param(self, value: Optional[_T], dialect: Dialect) -> Any:
+        """Receive a bound parameter value to be converted.
+
+        Custom subclasses of :class:`_types.TypeDecorator` should override
+        this method to provide custom behaviors for incoming data values.
+        This method is called at **statement execution time** and is passed
+        the literal Python data value which is to be associated with a bound
+        parameter in the statement.
+
+        The operation could be anything desired to perform custom
+        behavior, such as transforming or serializing data.
+        This could also be used as a hook for validating logic.
+
+        :param value: Data to operate upon, of any type expected by
+         this method in the subclass.  Can be ``None``.
+        :param dialect: the :class:`.Dialect` in use.
+
+        .. seealso::
+
+            :ref:`types_typedecorator`
+
+            :meth:`_types.TypeDecorator.process_result_value`
+
+        """
+
+        raise NotImplementedError()
+
+    def process_result_value(
+        self, value: Optional[Any], dialect: Dialect
+    ) -> Optional[_T]:
+        """Receive a result-row column value to be converted.
+
+        Custom subclasses of :class:`_types.TypeDecorator` should override
+        this method to provide custom behaviors for data values
+        being received in result rows coming from the database.
+        This method is called at **result fetching time** and is passed
+        the literal Python data value that's extracted from a database result
+        row.
+
+        The operation could be anything desired to perform custom
+        behavior, such as transforming or deserializing data.
+
+        :param value: Data to operate upon, of any type expected by
+         this method in the subclass.  Can be ``None``.
+        :param dialect: the :class:`.Dialect` in use.
+
+        .. seealso::
+
+            :ref:`types_typedecorator`
+
+            :meth:`_types.TypeDecorator.process_bind_param`
+
+
+        """
+
+        raise NotImplementedError()
+
+    @util.memoized_property
+    def _has_bind_processor(self) -> bool:
+        """memoized boolean, check if process_bind_param is implemented.
+
+        Allows the base process_bind_param to raise
+        NotImplementedError without needing to test an expensive
+        exception throw.
+
+        """
+
+        return util.method_is_overridden(
+            self, TypeDecorator.process_bind_param
+        )
+
+    @util.memoized_property
+    def _has_literal_processor(self) -> bool:
+        """memoized boolean, check if process_literal_param is implemented."""
+
+        return util.method_is_overridden(
+            self, TypeDecorator.process_literal_param
+        )
+
+    def literal_processor(
+        self, dialect: Dialect
+    ) -> Optional[_LiteralProcessorType[_T]]:
+        """Provide a literal processing function for the given
+        :class:`.Dialect`.
+
+        This is the method that fulfills the :class:`.TypeEngine`
+        contract for literal value conversion which normally occurs via
+        the :meth:`_types.TypeEngine.literal_processor` method.
+
+        .. note::
+
+            User-defined subclasses of :class:`_types.TypeDecorator` should
+            **not** implement this method, and should instead implement
+            :meth:`_types.TypeDecorator.process_literal_param` so that the
+            "inner" processing provided by the implementing type is maintained.
+
+        """
+
+        if self._has_literal_processor:
+            process_literal_param = self.process_literal_param
+            process_bind_param = None
+        elif self._has_bind_processor:
+            # use the bind processor if dont have a literal processor,
+            # but we have an impl literal processor
+            process_literal_param = None
+            process_bind_param = self.process_bind_param
+        else:
+            process_literal_param = None
+            process_bind_param = None
+
+        if process_literal_param is not None:
+            impl_processor = self.impl_instance.literal_processor(dialect)
+            if impl_processor:
+                fixed_impl_processor = impl_processor
+                fixed_process_literal_param = process_literal_param
+
+                def process(value: Any) -> str:
+                    return fixed_impl_processor(
+                        fixed_process_literal_param(value, dialect)
+                    )
+
+            else:
+                fixed_process_literal_param = process_literal_param
+
+                def process(value: Any) -> str:
+                    return fixed_process_literal_param(value, dialect)
+
+            return process
+
+        elif process_bind_param is not None:
+            impl_processor = self.impl_instance.literal_processor(dialect)
+            if not impl_processor:
+                return None
+            else:
+                fixed_impl_processor = impl_processor
+                fixed_process_bind_param = process_bind_param
+
+                def process(value: Any) -> str:
+                    return fixed_impl_processor(
+                        fixed_process_bind_param(value, dialect)
+                    )
+
+                return process
+        else:
+            return self.impl_instance.literal_processor(dialect)
+
+    def bind_processor(
+        self, dialect: Dialect
+    ) -> Optional[_BindProcessorType[_T]]:
+        """Provide a bound value processing function for the
+        given :class:`.Dialect`.
+
+        This is the method that fulfills the :class:`.TypeEngine`
+        contract for bound value conversion which normally occurs via
+        the :meth:`_types.TypeEngine.bind_processor` method.
+
+        .. note::
+
+            User-defined subclasses of :class:`_types.TypeDecorator` should
+            **not** implement this method, and should instead implement
+            :meth:`_types.TypeDecorator.process_bind_param` so that the "inner"
+            processing provided by the implementing type is maintained.
+
+        :param dialect: Dialect instance in use.
+
+        """
+        if self._has_bind_processor:
+            process_param = self.process_bind_param
+            impl_processor = self.impl_instance.bind_processor(dialect)
+            if impl_processor:
+                fixed_impl_processor = impl_processor
+                fixed_process_param = process_param
+
+                def process(value: Optional[_T]) -> Any:
+                    return fixed_impl_processor(
+                        fixed_process_param(value, dialect)
+                    )
+
+            else:
+                fixed_process_param = process_param
+
+                def process(value: Optional[_T]) -> Any:
+                    return fixed_process_param(value, dialect)
+
+            return process
+        else:
+            return self.impl_instance.bind_processor(dialect)
+
+    @util.memoized_property
+    def _has_result_processor(self) -> bool:
+        """memoized boolean, check if process_result_value is implemented.
+
+        Allows the base process_result_value to raise
+        NotImplementedError without needing to test an expensive
+        exception throw.
+
+        """
+
+        return util.method_is_overridden(
+            self, TypeDecorator.process_result_value
+        )
+
+    def result_processor(
+        self, dialect: Dialect, coltype: Any
+    ) -> Optional[_ResultProcessorType[_T]]:
+        """Provide a result value processing function for the given
+        :class:`.Dialect`.
+
+        This is the method that fulfills the :class:`.TypeEngine`
+        contract for bound value conversion which normally occurs via
+        the :meth:`_types.TypeEngine.result_processor` method.
+
+        .. note::
+
+            User-defined subclasses of :class:`_types.TypeDecorator` should
+            **not** implement this method, and should instead implement
+            :meth:`_types.TypeDecorator.process_result_value` so that the
+            "inner" processing provided by the implementing type is maintained.
+
+        :param dialect: Dialect instance in use.
+        :param coltype: A SQLAlchemy data type
+
+        """
+        if self._has_result_processor:
+            process_value = self.process_result_value
+            impl_processor = self.impl_instance.result_processor(
+                dialect, coltype
+            )
+            if impl_processor:
+                fixed_process_value = process_value
+                fixed_impl_processor = impl_processor
+
+                def process(value: Any) -> Optional[_T]:
+                    return fixed_process_value(
+                        fixed_impl_processor(value), dialect
+                    )
+
+            else:
+                fixed_process_value = process_value
+
+                def process(value: Any) -> Optional[_T]:
+                    return fixed_process_value(value, dialect)
+
+            return process
+        else:
+            return self.impl_instance.result_processor(dialect, coltype)
+
+    @util.memoized_property
+    def _has_bind_expression(self) -> bool:
+        return (
+            util.method_is_overridden(self, TypeDecorator.bind_expression)
+            or self.impl_instance._has_bind_expression
+        )
+
+    def bind_expression(
+        self, bindparam: BindParameter[_T]
+    ) -> Optional[ColumnElement[_T]]:
+        """Given a bind value (i.e. a :class:`.BindParameter` instance),
+        return a SQL expression which will typically wrap the given parameter.
+
+        .. note::
+
+            This method is called during the **SQL compilation** phase of a
+            statement, when rendering a SQL string. It is **not** necessarily
+            called against specific values, and should not be confused with the
+            :meth:`_types.TypeDecorator.process_bind_param` method, which is
+            the more typical method that processes the actual value passed to a
+            particular parameter at statement execution time.
+
+        Subclasses of :class:`_types.TypeDecorator` can override this method
+        to provide custom bind expression behavior for the type.  This
+        implementation will **replace** that of the underlying implementation
+        type.
+
+        """
+        return self.impl_instance.bind_expression(bindparam)
+
+    @util.memoized_property
+    def _has_column_expression(self) -> bool:
+        """memoized boolean, check if column_expression is implemented.
+
+        Allows the method to be skipped for the vast majority of expression
+        types that don't use this feature.
+
+        """
+
+        return (
+            util.method_is_overridden(self, TypeDecorator.column_expression)
+            or self.impl_instance._has_column_expression
+        )
+
+    def column_expression(
+        self, column: ColumnElement[_T]
+    ) -> Optional[ColumnElement[_T]]:
+        """Given a SELECT column expression, return a wrapping SQL expression.
+
+        .. note::
+
+            This method is called during the **SQL compilation** phase of a
+            statement, when rendering a SQL string. It is **not** called
+            against specific values, and should not be confused with the
+            :meth:`_types.TypeDecorator.process_result_value` method, which is
+            the more typical method that processes the actual value returned
+            in a result row subsequent to statement execution time.
+
+        Subclasses of :class:`_types.TypeDecorator` can override this method
+        to provide custom column expression behavior for the type.  This
+        implementation will **replace** that of the underlying implementation
+        type.
+
+        See the description of :meth:`_types.TypeEngine.column_expression`
+        for a complete description of the method's use.
+
+        """
+
+        return self.impl_instance.column_expression(column)
+
+    def coerce_compared_value(
+        self, op: Optional[OperatorType], value: Any
+    ) -> Any:
+        """Suggest a type for a 'coerced' Python value in an expression.
+
+        By default, returns self.   This method is called by
+        the expression system when an object using this type is
+        on the left or right side of an expression against a plain Python
+        object which does not yet have a SQLAlchemy type assigned::
+
+            expr = table.c.somecolumn + 35
+
+        Where above, if ``somecolumn`` uses this type, this method will
+        be called with the value ``operator.add``
+        and ``35``.  The return value is whatever SQLAlchemy type should
+        be used for ``35`` for this particular operation.
+
+        """
+        return self
+
+    def copy(self, **kw: Any) -> Self:
+        """Produce a copy of this :class:`.TypeDecorator` instance.
+
+        This is a shallow copy and is provided to fulfill part of
+        the :class:`.TypeEngine` contract.  It usually does not
+        need to be overridden unless the user-defined :class:`.TypeDecorator`
+        has local state that should be deep-copied.
+
+        """
+
+        instance = self.__class__.__new__(self.__class__)
+        instance.__dict__.update(self.__dict__)
+        return instance
+
+    def get_dbapi_type(self, dbapi: ModuleType) -> Optional[Any]:
+        """Return the DBAPI type object represented by this
+        :class:`.TypeDecorator`.
+
+        By default this calls upon :meth:`.TypeEngine.get_dbapi_type` of the
+        underlying "impl".
+        """
+        return self.impl_instance.get_dbapi_type(dbapi)
+
+    def compare_values(self, x: Any, y: Any) -> bool:
+        """Given two values, compare them for equality.
+
+        By default this calls upon :meth:`.TypeEngine.compare_values`
+        of the underlying "impl", which in turn usually
+        uses the Python equals operator ``==``.
+
+        This function is used by the ORM to compare
+        an original-loaded value with an intercepted
+        "changed" value, to determine if a net change
+        has occurred.
+
+        """
+        return self.impl_instance.compare_values(x, y)
+
+    # mypy property bug
+    @property
+    def sort_key_function(self) -> Optional[Callable[[Any], Any]]:  # type: ignore # noqa: E501
+        return self.impl_instance.sort_key_function
+
+    def __repr__(self) -> str:
+        return util.generic_repr(self, to_inspect=self.impl_instance)
+
+
+class Variant(TypeDecorator[_T]):
+    """deprecated.  symbol is present for backwards-compatibility with
+    workaround recipes, however this actual type should not be used.
+
+    """
+
+    def __init__(self, *arg: Any, **kw: Any):
+        raise NotImplementedError(
+            "Variant is no longer used in SQLAlchemy; this is a "
+            "placeholder symbol for backwards compatibility."
+        )
+
+
+@overload
+def to_instance(
+    typeobj: Union[Type[_TE], _TE], *arg: Any, **kw: Any
+) -> _TE: ...
+
+
+@overload
+def to_instance(typeobj: None, *arg: Any, **kw: Any) -> TypeEngine[None]: ...
+
+
+def to_instance(
+    typeobj: Union[Type[_TE], _TE, None], *arg: Any, **kw: Any
+) -> Union[_TE, TypeEngine[None]]:
+    if typeobj is None:
+        return NULLTYPE
+
+    if callable(typeobj):
+        return typeobj(*arg, **kw)
+    else:
+        return typeobj
+
+
+def adapt_type(
+    typeobj: _TypeEngineArgument[Any],
+    colspecs: Mapping[Type[Any], Type[TypeEngine[Any]]],
+) -> TypeEngine[Any]:
+    typeobj = to_instance(typeobj)
+    for t in typeobj.__class__.__mro__[0:-1]:
+        try:
+            impltype = colspecs[t]
+            break
+        except KeyError:
+            pass
+    else:
+        # couldn't adapt - so just return the type itself
+        # (it may be a user-defined type)
+        return typeobj
+    # if we adapted the given generic type to a database-specific type,
+    # but it turns out the originally given "generic" type
+    # is actually a subclass of our resulting type, then we were already
+    # given a more specific type than that required; so use that.
+    if issubclass(typeobj.__class__, impltype):
+        return typeobj
+    return typeobj.adapt(impltype)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/util.py
new file mode 100644
index 00000000..29cd0e2b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/util.py
@@ -0,0 +1,1487 @@
+# sql/util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""High level utilities which build upon other modules here.
+
+"""
+from __future__ import annotations
+
+from collections import deque
+import copy
+from itertools import chain
+import typing
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import coercions
+from . import operators
+from . import roles
+from . import visitors
+from ._typing import is_text_clause
+from .annotation import _deep_annotate as _deep_annotate  # noqa: F401
+from .annotation import _deep_deannotate as _deep_deannotate  # noqa: F401
+from .annotation import _shallow_annotate as _shallow_annotate  # noqa: F401
+from .base import _expand_cloned
+from .base import _from_objects
+from .cache_key import HasCacheKey as HasCacheKey  # noqa: F401
+from .ddl import sort_tables as sort_tables  # noqa: F401
+from .elements import _find_columns as _find_columns
+from .elements import _label_reference
+from .elements import _textual_label_reference
+from .elements import BindParameter
+from .elements import ClauseElement
+from .elements import ColumnClause
+from .elements import ColumnElement
+from .elements import Grouping
+from .elements import KeyedColumnElement
+from .elements import Label
+from .elements import NamedColumn
+from .elements import Null
+from .elements import UnaryExpression
+from .schema import Column
+from .selectable import Alias
+from .selectable import FromClause
+from .selectable import FromGrouping
+from .selectable import Join
+from .selectable import ScalarSelect
+from .selectable import SelectBase
+from .selectable import TableClause
+from .visitors import _ET
+from .. import exc
+from .. import util
+from ..util.typing import Literal
+from ..util.typing import Protocol
+
+if typing.TYPE_CHECKING:
+    from ._typing import _EquivalentColumnMap
+    from ._typing import _LimitOffsetType
+    from ._typing import _TypeEngineArgument
+    from .elements import BinaryExpression
+    from .elements import TextClause
+    from .selectable import _JoinTargetElement
+    from .selectable import _SelectIterable
+    from .selectable import Selectable
+    from .visitors import _TraverseCallableType
+    from .visitors import ExternallyTraversible
+    from .visitors import ExternalTraversal
+    from ..engine.interfaces import _AnyExecuteParams
+    from ..engine.interfaces import _AnyMultiExecuteParams
+    from ..engine.interfaces import _AnySingleExecuteParams
+    from ..engine.interfaces import _CoreSingleExecuteParams
+    from ..engine.row import Row
+
+_CE = TypeVar("_CE", bound="ColumnElement[Any]")
+
+
+def join_condition(
+    a: FromClause,
+    b: FromClause,
+    a_subset: Optional[FromClause] = None,
+    consider_as_foreign_keys: Optional[AbstractSet[ColumnClause[Any]]] = None,
+) -> ColumnElement[bool]:
+    """Create a join condition between two tables or selectables.
+
+    e.g.::
+
+        join_condition(tablea, tableb)
+
+    would produce an expression along the lines of::
+
+        tablea.c.id == tableb.c.tablea_id
+
+    The join is determined based on the foreign key relationships
+    between the two selectables.   If there are multiple ways
+    to join, or no way to join, an error is raised.
+
+    :param a_subset: An optional expression that is a sub-component
+        of ``a``.  An attempt will be made to join to just this sub-component
+        first before looking at the full ``a`` construct, and if found
+        will be successful even if there are other ways to join to ``a``.
+        This allows the "right side" of a join to be passed thereby
+        providing a "natural join".
+
+    """
+    return Join._join_condition(
+        a,
+        b,
+        a_subset=a_subset,
+        consider_as_foreign_keys=consider_as_foreign_keys,
+    )
+
+
+def find_join_source(
+    clauses: List[FromClause], join_to: FromClause
+) -> List[int]:
+    """Given a list of FROM clauses and a selectable,
+    return the first index and element from the list of
+    clauses which can be joined against the selectable.  returns
+    None, None if no match is found.
+
+    e.g.::
+
+        clause1 = table1.join(table2)
+        clause2 = table4.join(table5)
+
+        join_to = table2.join(table3)
+
+        find_join_source([clause1, clause2], join_to) == clause1
+
+    """
+
+    selectables = list(_from_objects(join_to))
+    idx = []
+    for i, f in enumerate(clauses):
+        for s in selectables:
+            if f.is_derived_from(s):
+                idx.append(i)
+    return idx
+
+
+def find_left_clause_that_matches_given(
+    clauses: Sequence[FromClause], join_from: FromClause
+) -> List[int]:
+    """Given a list of FROM clauses and a selectable,
+    return the indexes from the list of
+    clauses which is derived from the selectable.
+
+    """
+
+    selectables = list(_from_objects(join_from))
+    liberal_idx = []
+    for i, f in enumerate(clauses):
+        for s in selectables:
+            # basic check, if f is derived from s.
+            # this can be joins containing a table, or an aliased table
+            # or select statement matching to a table.  This check
+            # will match a table to a selectable that is adapted from
+            # that table.  With Query, this suits the case where a join
+            # is being made to an adapted entity
+            if f.is_derived_from(s):
+                liberal_idx.append(i)
+                break
+
+    # in an extremely small set of use cases, a join is being made where
+    # there are multiple FROM clauses where our target table is represented
+    # in more than one, such as embedded or similar.   in this case, do
+    # another pass where we try to get a more exact match where we aren't
+    # looking at adaption relationships.
+    if len(liberal_idx) > 1:
+        conservative_idx = []
+        for idx in liberal_idx:
+            f = clauses[idx]
+            for s in selectables:
+                if set(surface_selectables(f)).intersection(
+                    surface_selectables(s)
+                ):
+                    conservative_idx.append(idx)
+                    break
+        if conservative_idx:
+            return conservative_idx
+
+    return liberal_idx
+
+
+def find_left_clause_to_join_from(
+    clauses: Sequence[FromClause],
+    join_to: _JoinTargetElement,
+    onclause: Optional[ColumnElement[Any]],
+) -> List[int]:
+    """Given a list of FROM clauses, a selectable,
+    and optional ON clause, return a list of integer indexes from the
+    clauses list indicating the clauses that can be joined from.
+
+    The presence of an "onclause" indicates that at least one clause can
+    definitely be joined from; if the list of clauses is of length one
+    and the onclause is given, returns that index.   If the list of clauses
+    is more than length one, and the onclause is given, attempts to locate
+    which clauses contain the same columns.
+
+    """
+    idx = []
+    selectables = set(_from_objects(join_to))
+
+    # if we are given more than one target clause to join
+    # from, use the onclause to provide a more specific answer.
+    # otherwise, don't try to limit, after all, "ON TRUE" is a valid
+    # on clause
+    if len(clauses) > 1 and onclause is not None:
+        resolve_ambiguity = True
+        cols_in_onclause = _find_columns(onclause)
+    else:
+        resolve_ambiguity = False
+        cols_in_onclause = None
+
+    for i, f in enumerate(clauses):
+        for s in selectables.difference([f]):
+            if resolve_ambiguity:
+                assert cols_in_onclause is not None
+                if set(f.c).union(s.c).issuperset(cols_in_onclause):
+                    idx.append(i)
+                    break
+            elif onclause is not None or Join._can_join(f, s):
+                idx.append(i)
+                break
+
+    if len(idx) > 1:
+        # this is the same "hide froms" logic from
+        # Selectable._get_display_froms
+        toremove = set(
+            chain(*[_expand_cloned(f._hide_froms) for f in clauses])
+        )
+        idx = [i for i in idx if clauses[i] not in toremove]
+
+    # onclause was given and none of them resolved, so assume
+    # all indexes can match
+    if not idx and onclause is not None:
+        return list(range(len(clauses)))
+    else:
+        return idx
+
+
+def visit_binary_product(
+    fn: Callable[
+        [BinaryExpression[Any], ColumnElement[Any], ColumnElement[Any]], None
+    ],
+    expr: ColumnElement[Any],
+) -> None:
+    """Produce a traversal of the given expression, delivering
+    column comparisons to the given function.
+
+    The function is of the form::
+
+        def my_fn(binary, left, right): ...
+
+    For each binary expression located which has a
+    comparison operator, the product of "left" and
+    "right" will be delivered to that function,
+    in terms of that binary.
+
+    Hence an expression like::
+
+        and_((a + b) == q + func.sum(e + f), j == r)
+
+    would have the traversal:
+
+    .. sourcecode:: text
+
+        a <eq> q
+        a <eq> e
+        a <eq> f
+        b <eq> q
+        b <eq> e
+        b <eq> f
+        j <eq> r
+
+    That is, every combination of "left" and
+    "right" that doesn't further contain
+    a binary comparison is passed as pairs.
+
+    """
+    stack: List[BinaryExpression[Any]] = []
+
+    def visit(element: ClauseElement) -> Iterator[ColumnElement[Any]]:
+        if isinstance(element, ScalarSelect):
+            # we don't want to dig into correlated subqueries,
+            # those are just column elements by themselves
+            yield element
+        elif element.__visit_name__ == "binary" and operators.is_comparison(
+            element.operator  # type: ignore
+        ):
+            stack.insert(0, element)  # type: ignore
+            for l in visit(element.left):  # type: ignore
+                for r in visit(element.right):  # type: ignore
+                    fn(stack[0], l, r)
+            stack.pop(0)
+            for elem in element.get_children():
+                visit(elem)
+        else:
+            if isinstance(element, ColumnClause):
+                yield element
+            for elem in element.get_children():
+                yield from visit(elem)
+
+    list(visit(expr))
+    visit = None  # type: ignore  # remove gc cycles
+
+
+def find_tables(
+    clause: ClauseElement,
+    *,
+    check_columns: bool = False,
+    include_aliases: bool = False,
+    include_joins: bool = False,
+    include_selects: bool = False,
+    include_crud: bool = False,
+) -> List[TableClause]:
+    """locate Table objects within the given expression."""
+
+    tables: List[TableClause] = []
+    _visitors: Dict[str, _TraverseCallableType[Any]] = {}
+
+    if include_selects:
+        _visitors["select"] = _visitors["compound_select"] = tables.append
+
+    if include_joins:
+        _visitors["join"] = tables.append
+
+    if include_aliases:
+        _visitors["alias"] = _visitors["subquery"] = _visitors[
+            "tablesample"
+        ] = _visitors["lateral"] = tables.append
+
+    if include_crud:
+        _visitors["insert"] = _visitors["update"] = _visitors["delete"] = (
+            lambda ent: tables.append(ent.table)
+        )
+
+    if check_columns:
+
+        def visit_column(column):
+            tables.append(column.table)
+
+        _visitors["column"] = visit_column
+
+    _visitors["table"] = tables.append
+
+    visitors.traverse(clause, {}, _visitors)
+    return tables
+
+
+def unwrap_order_by(clause: Any) -> Any:
+    """Break up an 'order by' expression into individual column-expressions,
+    without DESC/ASC/NULLS FIRST/NULLS LAST"""
+
+    cols = util.column_set()
+    result = []
+    stack = deque([clause])
+
+    # examples
+    # column -> ASC/DESC == column
+    # column -> ASC/DESC -> label == column
+    # column -> label -> ASC/DESC -> label == column
+    # scalar_select -> label -> ASC/DESC == scalar_select -> label
+
+    while stack:
+        t = stack.popleft()
+        if isinstance(t, ColumnElement) and (
+            not isinstance(t, UnaryExpression)
+            or not operators.is_ordering_modifier(t.modifier)  # type: ignore
+        ):
+            if isinstance(t, Label) and not isinstance(
+                t.element, ScalarSelect
+            ):
+                t = t.element
+
+                if isinstance(t, Grouping):
+                    t = t.element
+
+                stack.append(t)
+                continue
+            elif isinstance(t, _label_reference):
+                t = t.element
+
+                stack.append(t)
+                continue
+            if isinstance(t, (_textual_label_reference)):
+                continue
+            if t not in cols:
+                cols.add(t)
+                result.append(t)
+
+        else:
+            for c in t.get_children():
+                stack.append(c)
+    return result
+
+
+def unwrap_label_reference(element):
+    def replace(
+        element: ExternallyTraversible, **kw: Any
+    ) -> Optional[ExternallyTraversible]:
+        if isinstance(element, _label_reference):
+            return element.element
+        elif isinstance(element, _textual_label_reference):
+            assert False, "can't unwrap a textual label reference"
+        return None
+
+    return visitors.replacement_traverse(element, {}, replace)
+
+
+def expand_column_list_from_order_by(collist, order_by):
+    """Given the columns clause and ORDER BY of a selectable,
+    return a list of column expressions that can be added to the collist
+    corresponding to the ORDER BY, without repeating those already
+    in the collist.
+
+    """
+    cols_already_present = {
+        col.element if col._order_by_label_element is not None else col
+        for col in collist
+    }
+
+    to_look_for = list(chain(*[unwrap_order_by(o) for o in order_by]))
+
+    return [col for col in to_look_for if col not in cols_already_present]
+
+
+def clause_is_present(clause, search):
+    """Given a target clause and a second to search within, return True
+    if the target is plainly present in the search without any
+    subqueries or aliases involved.
+
+    Basically descends through Joins.
+
+    """
+
+    for elem in surface_selectables(search):
+        if clause == elem:  # use == here so that Annotated's compare
+            return True
+    else:
+        return False
+
+
+def tables_from_leftmost(clause: FromClause) -> Iterator[FromClause]:
+    if isinstance(clause, Join):
+        yield from tables_from_leftmost(clause.left)
+        yield from tables_from_leftmost(clause.right)
+    elif isinstance(clause, FromGrouping):
+        yield from tables_from_leftmost(clause.element)
+    else:
+        yield clause
+
+
+def surface_selectables(clause):
+    stack = [clause]
+    while stack:
+        elem = stack.pop()
+        yield elem
+        if isinstance(elem, Join):
+            stack.extend((elem.left, elem.right))
+        elif isinstance(elem, FromGrouping):
+            stack.append(elem.element)
+
+
+def surface_selectables_only(clause):
+    stack = [clause]
+    while stack:
+        elem = stack.pop()
+        if isinstance(elem, (TableClause, Alias)):
+            yield elem
+        if isinstance(elem, Join):
+            stack.extend((elem.left, elem.right))
+        elif isinstance(elem, FromGrouping):
+            stack.append(elem.element)
+        elif isinstance(elem, ColumnClause):
+            if elem.table is not None:
+                stack.append(elem.table)
+            else:
+                yield elem
+        elif elem is not None:
+            yield elem
+
+
+def extract_first_column_annotation(column, annotation_name):
+    filter_ = (FromGrouping, SelectBase)
+
+    stack = deque([column])
+    while stack:
+        elem = stack.popleft()
+        if annotation_name in elem._annotations:
+            return elem._annotations[annotation_name]
+        for sub in elem.get_children():
+            if isinstance(sub, filter_):
+                continue
+            stack.append(sub)
+    return None
+
+
+def selectables_overlap(left: FromClause, right: FromClause) -> bool:
+    """Return True if left/right have some overlapping selectable"""
+
+    return bool(
+        set(surface_selectables(left)).intersection(surface_selectables(right))
+    )
+
+
+def bind_values(clause):
+    """Return an ordered list of "bound" values in the given clause.
+
+    E.g.::
+
+        >>> expr = and_(table.c.foo == 5, table.c.foo == 7)
+        >>> bind_values(expr)
+        [5, 7]
+    """
+
+    v = []
+
+    def visit_bindparam(bind):
+        v.append(bind.effective_value)
+
+    visitors.traverse(clause, {}, {"bindparam": visit_bindparam})
+    return v
+
+
+def _quote_ddl_expr(element):
+    if isinstance(element, str):
+        element = element.replace("'", "''")
+        return "'%s'" % element
+    else:
+        return repr(element)
+
+
+class _repr_base:
+    _LIST: int = 0
+    _TUPLE: int = 1
+    _DICT: int = 2
+
+    __slots__ = ("max_chars",)
+
+    max_chars: int
+
+    def trunc(self, value: Any) -> str:
+        rep = repr(value)
+        lenrep = len(rep)
+        if lenrep > self.max_chars:
+            segment_length = self.max_chars // 2
+            rep = (
+                rep[0:segment_length]
+                + (
+                    " ... (%d characters truncated) ... "
+                    % (lenrep - self.max_chars)
+                )
+                + rep[-segment_length:]
+            )
+        return rep
+
+
+def _repr_single_value(value):
+    rp = _repr_base()
+    rp.max_chars = 300
+    return rp.trunc(value)
+
+
+class _repr_row(_repr_base):
+    """Provide a string view of a row."""
+
+    __slots__ = ("row",)
+
+    def __init__(self, row: Row[Any], max_chars: int = 300):
+        self.row = row
+        self.max_chars = max_chars
+
+    def __repr__(self) -> str:
+        trunc = self.trunc
+        return "(%s%s)" % (
+            ", ".join(trunc(value) for value in self.row),
+            "," if len(self.row) == 1 else "",
+        )
+
+
+class _long_statement(str):
+    def __str__(self) -> str:
+        lself = len(self)
+        if lself > 500:
+            lleft = 250
+            lright = 100
+            trunc = lself - lleft - lright
+            return (
+                f"{self[0:lleft]} ... {trunc} "
+                f"characters truncated ... {self[-lright:]}"
+            )
+        else:
+            return str.__str__(self)
+
+
+class _repr_params(_repr_base):
+    """Provide a string view of bound parameters.
+
+    Truncates display to a given number of 'multi' parameter sets,
+    as well as long values to a given number of characters.
+
+    """
+
+    __slots__ = "params", "batches", "ismulti", "max_params"
+
+    def __init__(
+        self,
+        params: Optional[_AnyExecuteParams],
+        batches: int,
+        max_params: int = 100,
+        max_chars: int = 300,
+        ismulti: Optional[bool] = None,
+    ):
+        self.params = params
+        self.ismulti = ismulti
+        self.batches = batches
+        self.max_chars = max_chars
+        self.max_params = max_params
+
+    def __repr__(self) -> str:
+        if self.ismulti is None:
+            return self.trunc(self.params)
+
+        if isinstance(self.params, list):
+            typ = self._LIST
+
+        elif isinstance(self.params, tuple):
+            typ = self._TUPLE
+        elif isinstance(self.params, dict):
+            typ = self._DICT
+        else:
+            return self.trunc(self.params)
+
+        if self.ismulti:
+            multi_params = cast(
+                "_AnyMultiExecuteParams",
+                self.params,
+            )
+
+            if len(self.params) > self.batches:
+                msg = (
+                    " ... displaying %i of %i total bound parameter sets ... "
+                )
+                return " ".join(
+                    (
+                        self._repr_multi(
+                            multi_params[: self.batches - 2],
+                            typ,
+                        )[0:-1],
+                        msg % (self.batches, len(self.params)),
+                        self._repr_multi(multi_params[-2:], typ)[1:],
+                    )
+                )
+            else:
+                return self._repr_multi(multi_params, typ)
+        else:
+            return self._repr_params(
+                cast(
+                    "_AnySingleExecuteParams",
+                    self.params,
+                ),
+                typ,
+            )
+
+    def _repr_multi(
+        self,
+        multi_params: _AnyMultiExecuteParams,
+        typ: int,
+    ) -> str:
+        if multi_params:
+            if isinstance(multi_params[0], list):
+                elem_type = self._LIST
+            elif isinstance(multi_params[0], tuple):
+                elem_type = self._TUPLE
+            elif isinstance(multi_params[0], dict):
+                elem_type = self._DICT
+            else:
+                assert False, "Unknown parameter type %s" % (
+                    type(multi_params[0])
+                )
+
+            elements = ", ".join(
+                self._repr_params(params, elem_type) for params in multi_params
+            )
+        else:
+            elements = ""
+
+        if typ == self._LIST:
+            return "[%s]" % elements
+        else:
+            return "(%s)" % elements
+
+    def _get_batches(self, params: Iterable[Any]) -> Any:
+        lparams = list(params)
+        lenparams = len(lparams)
+        if lenparams > self.max_params:
+            lleft = self.max_params // 2
+            return (
+                lparams[0:lleft],
+                lparams[-lleft:],
+                lenparams - self.max_params,
+            )
+        else:
+            return lparams, None, None
+
+    def _repr_params(
+        self,
+        params: _AnySingleExecuteParams,
+        typ: int,
+    ) -> str:
+        if typ is self._DICT:
+            return self._repr_param_dict(
+                cast("_CoreSingleExecuteParams", params)
+            )
+        elif typ is self._TUPLE:
+            return self._repr_param_tuple(cast("Sequence[Any]", params))
+        else:
+            return self._repr_param_list(params)
+
+    def _repr_param_dict(self, params: _CoreSingleExecuteParams) -> str:
+        trunc = self.trunc
+        (
+            items_first_batch,
+            items_second_batch,
+            trunclen,
+        ) = self._get_batches(params.items())
+
+        if items_second_batch:
+            text = "{%s" % (
+                ", ".join(
+                    f"{key!r}: {trunc(value)}"
+                    for key, value in items_first_batch
+                )
+            )
+            text += f" ... {trunclen} parameters truncated ... "
+            text += "%s}" % (
+                ", ".join(
+                    f"{key!r}: {trunc(value)}"
+                    for key, value in items_second_batch
+                )
+            )
+        else:
+            text = "{%s}" % (
+                ", ".join(
+                    f"{key!r}: {trunc(value)}"
+                    for key, value in items_first_batch
+                )
+            )
+        return text
+
+    def _repr_param_tuple(self, params: Sequence[Any]) -> str:
+        trunc = self.trunc
+
+        (
+            items_first_batch,
+            items_second_batch,
+            trunclen,
+        ) = self._get_batches(params)
+
+        if items_second_batch:
+            text = "(%s" % (
+                ", ".join(trunc(value) for value in items_first_batch)
+            )
+            text += f" ... {trunclen} parameters truncated ... "
+            text += "%s)" % (
+                ", ".join(trunc(value) for value in items_second_batch),
+            )
+        else:
+            text = "(%s%s)" % (
+                ", ".join(trunc(value) for value in items_first_batch),
+                "," if len(items_first_batch) == 1 else "",
+            )
+        return text
+
+    def _repr_param_list(self, params: _AnySingleExecuteParams) -> str:
+        trunc = self.trunc
+        (
+            items_first_batch,
+            items_second_batch,
+            trunclen,
+        ) = self._get_batches(params)
+
+        if items_second_batch:
+            text = "[%s" % (
+                ", ".join(trunc(value) for value in items_first_batch)
+            )
+            text += f" ... {trunclen} parameters truncated ... "
+            text += "%s]" % (
+                ", ".join(trunc(value) for value in items_second_batch)
+            )
+        else:
+            text = "[%s]" % (
+                ", ".join(trunc(value) for value in items_first_batch)
+            )
+        return text
+
+
+def adapt_criterion_to_null(crit: _CE, nulls: Collection[Any]) -> _CE:
+    """given criterion containing bind params, convert selected elements
+    to IS NULL.
+
+    """
+
+    def visit_binary(binary):
+        if (
+            isinstance(binary.left, BindParameter)
+            and binary.left._identifying_key in nulls
+        ):
+            # reverse order if the NULL is on the left side
+            binary.left = binary.right
+            binary.right = Null()
+            binary.operator = operators.is_
+            binary.negate = operators.is_not
+        elif (
+            isinstance(binary.right, BindParameter)
+            and binary.right._identifying_key in nulls
+        ):
+            binary.right = Null()
+            binary.operator = operators.is_
+            binary.negate = operators.is_not
+
+    return visitors.cloned_traverse(crit, {}, {"binary": visit_binary})
+
+
+def splice_joins(
+    left: Optional[FromClause],
+    right: Optional[FromClause],
+    stop_on: Optional[FromClause] = None,
+) -> Optional[FromClause]:
+    if left is None:
+        return right
+
+    stack: List[Tuple[Optional[FromClause], Optional[Join]]] = [(right, None)]
+
+    adapter = ClauseAdapter(left)
+    ret = None
+    while stack:
+        (right, prevright) = stack.pop()
+        if isinstance(right, Join) and right is not stop_on:
+            right = right._clone()
+            right.onclause = adapter.traverse(right.onclause)
+            stack.append((right.left, right))
+        else:
+            right = adapter.traverse(right)
+        if prevright is not None:
+            assert right is not None
+            prevright.left = right
+        if ret is None:
+            ret = right
+
+    return ret
+
+
+@overload
+def reduce_columns(
+    columns: Iterable[ColumnElement[Any]],
+    *clauses: Optional[ClauseElement],
+    **kw: bool,
+) -> Sequence[ColumnElement[Any]]: ...
+
+
+@overload
+def reduce_columns(
+    columns: _SelectIterable,
+    *clauses: Optional[ClauseElement],
+    **kw: bool,
+) -> Sequence[Union[ColumnElement[Any], TextClause]]: ...
+
+
+def reduce_columns(
+    columns: _SelectIterable,
+    *clauses: Optional[ClauseElement],
+    **kw: bool,
+) -> Collection[Union[ColumnElement[Any], TextClause]]:
+    r"""given a list of columns, return a 'reduced' set based on natural
+    equivalents.
+
+    the set is reduced to the smallest list of columns which have no natural
+    equivalent present in the list.  A "natural equivalent" means that two
+    columns will ultimately represent the same value because they are related
+    by a foreign key.
+
+    \*clauses is an optional list of join clauses which will be traversed
+    to further identify columns that are "equivalent".
+
+    \**kw may specify 'ignore_nonexistent_tables' to ignore foreign keys
+    whose tables are not yet configured, or columns that aren't yet present.
+
+    This function is primarily used to determine the most minimal "primary
+    key" from a selectable, by reducing the set of primary key columns present
+    in the selectable to just those that are not repeated.
+
+    """
+    ignore_nonexistent_tables = kw.pop("ignore_nonexistent_tables", False)
+    only_synonyms = kw.pop("only_synonyms", False)
+
+    column_set = util.OrderedSet(columns)
+    cset_no_text: util.OrderedSet[ColumnElement[Any]] = column_set.difference(
+        c for c in column_set if is_text_clause(c)  # type: ignore
+    )
+
+    omit = util.column_set()
+    for col in cset_no_text:
+        for fk in chain(*[c.foreign_keys for c in col.proxy_set]):
+            for c in cset_no_text:
+                if c is col:
+                    continue
+                try:
+                    fk_col = fk.column
+                except exc.NoReferencedColumnError:
+                    # TODO: add specific coverage here
+                    # to test/sql/test_selectable ReduceTest
+                    if ignore_nonexistent_tables:
+                        continue
+                    else:
+                        raise
+                except exc.NoReferencedTableError:
+                    # TODO: add specific coverage here
+                    # to test/sql/test_selectable ReduceTest
+                    if ignore_nonexistent_tables:
+                        continue
+                    else:
+                        raise
+                if fk_col.shares_lineage(c) and (
+                    not only_synonyms or c.name == col.name
+                ):
+                    omit.add(col)
+                    break
+
+    if clauses:
+
+        def visit_binary(binary):
+            if binary.operator == operators.eq:
+                cols = util.column_set(
+                    chain(
+                        *[c.proxy_set for c in cset_no_text.difference(omit)]
+                    )
+                )
+                if binary.left in cols and binary.right in cols:
+                    for c in reversed(cset_no_text):
+                        if c.shares_lineage(binary.right) and (
+                            not only_synonyms or c.name == binary.left.name
+                        ):
+                            omit.add(c)
+                            break
+
+        for clause in clauses:
+            if clause is not None:
+                visitors.traverse(clause, {}, {"binary": visit_binary})
+
+    return column_set.difference(omit)
+
+
+def criterion_as_pairs(
+    expression,
+    consider_as_foreign_keys=None,
+    consider_as_referenced_keys=None,
+    any_operator=False,
+):
+    """traverse an expression and locate binary criterion pairs."""
+
+    if consider_as_foreign_keys and consider_as_referenced_keys:
+        raise exc.ArgumentError(
+            "Can only specify one of "
+            "'consider_as_foreign_keys' or "
+            "'consider_as_referenced_keys'"
+        )
+
+    def col_is(a, b):
+        # return a is b
+        return a.compare(b)
+
+    def visit_binary(binary):
+        if not any_operator and binary.operator is not operators.eq:
+            return
+        if not isinstance(binary.left, ColumnElement) or not isinstance(
+            binary.right, ColumnElement
+        ):
+            return
+
+        if consider_as_foreign_keys:
+            if binary.left in consider_as_foreign_keys and (
+                col_is(binary.right, binary.left)
+                or binary.right not in consider_as_foreign_keys
+            ):
+                pairs.append((binary.right, binary.left))
+            elif binary.right in consider_as_foreign_keys and (
+                col_is(binary.left, binary.right)
+                or binary.left not in consider_as_foreign_keys
+            ):
+                pairs.append((binary.left, binary.right))
+        elif consider_as_referenced_keys:
+            if binary.left in consider_as_referenced_keys and (
+                col_is(binary.right, binary.left)
+                or binary.right not in consider_as_referenced_keys
+            ):
+                pairs.append((binary.left, binary.right))
+            elif binary.right in consider_as_referenced_keys and (
+                col_is(binary.left, binary.right)
+                or binary.left not in consider_as_referenced_keys
+            ):
+                pairs.append((binary.right, binary.left))
+        else:
+            if isinstance(binary.left, Column) and isinstance(
+                binary.right, Column
+            ):
+                if binary.left.references(binary.right):
+                    pairs.append((binary.right, binary.left))
+                elif binary.right.references(binary.left):
+                    pairs.append((binary.left, binary.right))
+
+    pairs: List[Tuple[ColumnElement[Any], ColumnElement[Any]]] = []
+    visitors.traverse(expression, {}, {"binary": visit_binary})
+    return pairs
+
+
+class ClauseAdapter(visitors.ReplacingExternalTraversal):
+    """Clones and modifies clauses based on column correspondence.
+
+    E.g.::
+
+      table1 = Table(
+          "sometable",
+          metadata,
+          Column("col1", Integer),
+          Column("col2", Integer),
+      )
+      table2 = Table(
+          "someothertable",
+          metadata,
+          Column("col1", Integer),
+          Column("col2", Integer),
+      )
+
+      condition = table1.c.col1 == table2.c.col1
+
+    make an alias of table1::
+
+      s = table1.alias("foo")
+
+    calling ``ClauseAdapter(s).traverse(condition)`` converts
+    condition to read::
+
+      s.c.col1 == table2.c.col1
+
+    """
+
+    __slots__ = (
+        "__traverse_options__",
+        "selectable",
+        "include_fn",
+        "exclude_fn",
+        "equivalents",
+        "adapt_on_names",
+        "adapt_from_selectables",
+    )
+
+    def __init__(
+        self,
+        selectable: Selectable,
+        equivalents: Optional[_EquivalentColumnMap] = None,
+        include_fn: Optional[Callable[[ClauseElement], bool]] = None,
+        exclude_fn: Optional[Callable[[ClauseElement], bool]] = None,
+        adapt_on_names: bool = False,
+        anonymize_labels: bool = False,
+        adapt_from_selectables: Optional[AbstractSet[FromClause]] = None,
+    ):
+        self.__traverse_options__ = {
+            "stop_on": [selectable],
+            "anonymize_labels": anonymize_labels,
+        }
+        self.selectable = selectable
+        self.include_fn = include_fn
+        self.exclude_fn = exclude_fn
+        self.equivalents = util.column_dict(equivalents or {})
+        self.adapt_on_names = adapt_on_names
+        self.adapt_from_selectables = adapt_from_selectables
+
+    if TYPE_CHECKING:
+
+        @overload
+        def traverse(self, obj: Literal[None]) -> None: ...
+
+        # note this specializes the ReplacingExternalTraversal.traverse()
+        # method to state
+        # that we will return the same kind of ExternalTraversal object as
+        # we were given.  This is probably not 100% true, such as it's
+        # possible for us to swap out Alias for Table at the top level.
+        # Ideally there could be overloads specific to ColumnElement and
+        # FromClause but Mypy is not accepting those as compatible with
+        # the base ReplacingExternalTraversal
+        @overload
+        def traverse(self, obj: _ET) -> _ET: ...
+
+        def traverse(
+            self, obj: Optional[ExternallyTraversible]
+        ) -> Optional[ExternallyTraversible]: ...
+
+    def _corresponding_column(
+        self, col, require_embedded, _seen=util.EMPTY_SET
+    ):
+        newcol = self.selectable.corresponding_column(
+            col, require_embedded=require_embedded
+        )
+        if newcol is None and col in self.equivalents and col not in _seen:
+            for equiv in self.equivalents[col]:
+                newcol = self._corresponding_column(
+                    equiv,
+                    require_embedded=require_embedded,
+                    _seen=_seen.union([col]),
+                )
+                if newcol is not None:
+                    return newcol
+
+        if (
+            self.adapt_on_names
+            and newcol is None
+            and isinstance(col, NamedColumn)
+        ):
+            newcol = self.selectable.exported_columns.get(col.name)
+        return newcol
+
+    @util.preload_module("sqlalchemy.sql.functions")
+    def replace(
+        self, col: _ET, _include_singleton_constants: bool = False
+    ) -> Optional[_ET]:
+        functions = util.preloaded.sql_functions
+
+        # TODO: cython candidate
+
+        if self.include_fn and not self.include_fn(col):  # type: ignore
+            return None
+        elif self.exclude_fn and self.exclude_fn(col):  # type: ignore
+            return None
+
+        if isinstance(col, FromClause) and not isinstance(
+            col, functions.FunctionElement
+        ):
+            if self.selectable.is_derived_from(col):
+                if self.adapt_from_selectables:
+                    for adp in self.adapt_from_selectables:
+                        if adp.is_derived_from(col):
+                            break
+                    else:
+                        return None
+                return self.selectable  # type: ignore
+            elif isinstance(col, Alias) and isinstance(
+                col.element, TableClause
+            ):
+                # we are a SELECT statement and not derived from an alias of a
+                # table (which nonetheless may be a table our SELECT derives
+                # from), so return the alias to prevent further traversal
+                # or
+                # we are an alias of a table and we are not derived from an
+                # alias of a table (which nonetheless may be the same table
+                # as ours) so, same thing
+                return col  # type: ignore
+            else:
+                # other cases where we are a selectable and the element
+                # is another join or selectable that contains a table which our
+                # selectable derives from, that we want to process
+                return None
+
+        elif not isinstance(col, ColumnElement):
+            return None
+        elif not _include_singleton_constants and col._is_singleton_constant:
+            # dont swap out NULL, TRUE, FALSE for a label name
+            # in a SQL statement that's being rewritten,
+            # leave them as the constant.  This is first noted in #6259,
+            # however the logic to check this moved here as of #7154 so that
+            # it is made specific to SQL rewriting and not all column
+            # correspondence
+
+            return None
+
+        if "adapt_column" in col._annotations:
+            col = col._annotations["adapt_column"]
+
+        if TYPE_CHECKING:
+            assert isinstance(col, KeyedColumnElement)
+
+        if self.adapt_from_selectables and col not in self.equivalents:
+            for adp in self.adapt_from_selectables:
+                if adp.c.corresponding_column(col, False) is not None:
+                    break
+            else:
+                return None
+
+        if TYPE_CHECKING:
+            assert isinstance(col, KeyedColumnElement)
+
+        return self._corresponding_column(  # type: ignore
+            col, require_embedded=True
+        )
+
+
+class _ColumnLookup(Protocol):
+    @overload
+    def __getitem__(self, key: None) -> None: ...
+
+    @overload
+    def __getitem__(self, key: ColumnClause[Any]) -> ColumnClause[Any]: ...
+
+    @overload
+    def __getitem__(self, key: ColumnElement[Any]) -> ColumnElement[Any]: ...
+
+    @overload
+    def __getitem__(self, key: _ET) -> _ET: ...
+
+    def __getitem__(self, key: Any) -> Any: ...
+
+
+class ColumnAdapter(ClauseAdapter):
+    """Extends ClauseAdapter with extra utility functions.
+
+    Key aspects of ColumnAdapter include:
+
+    * Expressions that are adapted are stored in a persistent
+      .columns collection; so that an expression E adapted into
+      an expression E1, will return the same object E1 when adapted
+      a second time.   This is important in particular for things like
+      Label objects that are anonymized, so that the ColumnAdapter can
+      be used to present a consistent "adapted" view of things.
+
+    * Exclusion of items from the persistent collection based on
+      include/exclude rules, but also independent of hash identity.
+      This because "annotated" items all have the same hash identity as their
+      parent.
+
+    * "wrapping" capability is added, so that the replacement of an expression
+      E can proceed through a series of adapters.  This differs from the
+      visitor's "chaining" feature in that the resulting object is passed
+      through all replacing functions unconditionally, rather than stopping
+      at the first one that returns non-None.
+
+    * An adapt_required option, used by eager loading to indicate that
+      We don't trust a result row column that is not translated.
+      This is to prevent a column from being interpreted as that
+      of the child row in a self-referential scenario, see
+      inheritance/test_basic.py->EagerTargetingTest.test_adapt_stringency
+
+    """
+
+    __slots__ = (
+        "columns",
+        "adapt_required",
+        "allow_label_resolve",
+        "_wrap",
+        "__weakref__",
+    )
+
+    columns: _ColumnLookup
+
+    def __init__(
+        self,
+        selectable: Selectable,
+        equivalents: Optional[_EquivalentColumnMap] = None,
+        adapt_required: bool = False,
+        include_fn: Optional[Callable[[ClauseElement], bool]] = None,
+        exclude_fn: Optional[Callable[[ClauseElement], bool]] = None,
+        adapt_on_names: bool = False,
+        allow_label_resolve: bool = True,
+        anonymize_labels: bool = False,
+        adapt_from_selectables: Optional[AbstractSet[FromClause]] = None,
+    ):
+        super().__init__(
+            selectable,
+            equivalents,
+            include_fn=include_fn,
+            exclude_fn=exclude_fn,
+            adapt_on_names=adapt_on_names,
+            anonymize_labels=anonymize_labels,
+            adapt_from_selectables=adapt_from_selectables,
+        )
+
+        self.columns = util.WeakPopulateDict(self._locate_col)  # type: ignore
+        if self.include_fn or self.exclude_fn:
+            self.columns = self._IncludeExcludeMapping(self, self.columns)
+        self.adapt_required = adapt_required
+        self.allow_label_resolve = allow_label_resolve
+        self._wrap = None
+
+    class _IncludeExcludeMapping:
+        def __init__(self, parent, columns):
+            self.parent = parent
+            self.columns = columns
+
+        def __getitem__(self, key):
+            if (
+                self.parent.include_fn and not self.parent.include_fn(key)
+            ) or (self.parent.exclude_fn and self.parent.exclude_fn(key)):
+                if self.parent._wrap:
+                    return self.parent._wrap.columns[key]
+                else:
+                    return key
+            return self.columns[key]
+
+    def wrap(self, adapter):
+        ac = copy.copy(self)
+        ac._wrap = adapter
+        ac.columns = util.WeakPopulateDict(ac._locate_col)  # type: ignore
+        if ac.include_fn or ac.exclude_fn:
+            ac.columns = self._IncludeExcludeMapping(ac, ac.columns)
+
+        return ac
+
+    @overload
+    def traverse(self, obj: Literal[None]) -> None: ...
+
+    @overload
+    def traverse(self, obj: _ET) -> _ET: ...
+
+    def traverse(
+        self, obj: Optional[ExternallyTraversible]
+    ) -> Optional[ExternallyTraversible]:
+        return self.columns[obj]
+
+    def chain(self, visitor: ExternalTraversal) -> ColumnAdapter:
+        assert isinstance(visitor, ColumnAdapter)
+
+        return super().chain(visitor)
+
+    if TYPE_CHECKING:
+
+        @property
+        def visitor_iterator(self) -> Iterator[ColumnAdapter]: ...
+
+    adapt_clause = traverse
+    adapt_list = ClauseAdapter.copy_and_process
+
+    def adapt_check_present(
+        self, col: ColumnElement[Any]
+    ) -> Optional[ColumnElement[Any]]:
+        newcol = self.columns[col]
+
+        if newcol is col and self._corresponding_column(col, True) is None:
+            return None
+
+        return newcol
+
+    def _locate_col(
+        self, col: ColumnElement[Any]
+    ) -> Optional[ColumnElement[Any]]:
+        # both replace and traverse() are overly complicated for what
+        # we are doing here and we would do better to have an inlined
+        # version that doesn't build up as much overhead.  the issue is that
+        # sometimes the lookup does in fact have to adapt the insides of
+        # say a labeled scalar subquery.   However, if the object is an
+        # Immutable, i.e. Column objects, we can skip the "clone" /
+        # "copy internals" part since those will be no-ops in any case.
+        # additionally we want to catch singleton objects null/true/false
+        # and make sure they are adapted as well here.
+
+        if col._is_immutable:
+            for vis in self.visitor_iterator:
+                c = vis.replace(col, _include_singleton_constants=True)
+                if c is not None:
+                    break
+            else:
+                c = col
+        else:
+            c = ClauseAdapter.traverse(self, col)
+
+        if self._wrap:
+            c2 = self._wrap._locate_col(c)
+            if c2 is not None:
+                c = c2
+
+        if self.adapt_required and c is col:
+            return None
+
+        # allow_label_resolve is consumed by one case for joined eager loading
+        # as part of its logic to prevent its own columns from being affected
+        # by .order_by().  Before full typing were applied to the ORM, this
+        # logic would set this attribute on the incoming object (which is
+        # typically a column, but we have a test for it being a non-column
+        # object) if no column were found.  While this seemed to
+        # have no negative effects, this adjustment should only occur on the
+        # new column which is assumed to be local to an adapted selectable.
+        if c is not col:
+            c._allow_label_resolve = self.allow_label_resolve
+
+        return c
+
+
+def _offset_or_limit_clause(
+    element: _LimitOffsetType,
+    name: Optional[str] = None,
+    type_: Optional[_TypeEngineArgument[int]] = None,
+) -> ColumnElement[int]:
+    """Convert the given value to an "offset or limit" clause.
+
+    This handles incoming integers and converts to an expression; if
+    an expression is already given, it is passed through.
+
+    """
+    return coercions.expect(
+        roles.LimitOffsetRole, element, name=name, type_=type_
+    )
+
+
+def _offset_or_limit_clause_asint_if_possible(
+    clause: _LimitOffsetType,
+) -> _LimitOffsetType:
+    """Return the offset or limit clause as a simple integer if possible,
+    else return the clause.
+
+    """
+    if clause is None:
+        return None
+    if hasattr(clause, "_limit_offset_value"):
+        value = clause._limit_offset_value
+        return util.asint(value)
+    else:
+        return clause
+
+
+def _make_slice(
+    limit_clause: _LimitOffsetType,
+    offset_clause: _LimitOffsetType,
+    start: int,
+    stop: int,
+) -> Tuple[Optional[ColumnElement[int]], Optional[ColumnElement[int]]]:
+    """Compute LIMIT/OFFSET in terms of slice start/end"""
+
+    # for calculated limit/offset, try to do the addition of
+    # values to offset in Python, however if a SQL clause is present
+    # then the addition has to be on the SQL side.
+
+    # TODO: typing is finding a few gaps in here, see if they can be
+    # closed up
+
+    if start is not None and stop is not None:
+        offset_clause = _offset_or_limit_clause_asint_if_possible(
+            offset_clause
+        )
+        if offset_clause is None:
+            offset_clause = 0
+
+        if start != 0:
+            offset_clause = offset_clause + start  # type: ignore
+
+        if offset_clause == 0:
+            offset_clause = None
+        else:
+            assert offset_clause is not None
+            offset_clause = _offset_or_limit_clause(offset_clause)
+
+        limit_clause = _offset_or_limit_clause(stop - start)
+
+    elif start is None and stop is not None:
+        limit_clause = _offset_or_limit_clause(stop)
+    elif start is not None and stop is None:
+        offset_clause = _offset_or_limit_clause_asint_if_possible(
+            offset_clause
+        )
+        if offset_clause is None:
+            offset_clause = 0
+
+        if start != 0:
+            offset_clause = offset_clause + start
+
+        if offset_clause == 0:
+            offset_clause = None
+        else:
+            offset_clause = _offset_or_limit_clause(offset_clause)
+
+    return limit_clause, offset_clause
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/visitors.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/visitors.py
new file mode 100644
index 00000000..e758350a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/sql/visitors.py
@@ -0,0 +1,1167 @@
+# sql/visitors.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Visitor/traversal interface and library functions.
+
+
+"""
+
+from __future__ import annotations
+
+from collections import deque
+from enum import Enum
+import itertools
+import operator
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import ClassVar
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import overload
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .. import exc
+from .. import util
+from ..util import langhelpers
+from ..util._has_cy import HAS_CYEXTENSION
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import Self
+
+if TYPE_CHECKING:
+    from .annotation import _AnnotationDict
+    from .elements import ColumnElement
+
+if typing.TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_util import prefix_anon_map as prefix_anon_map
+    from ._py_util import cache_anon_map as anon_map
+else:
+    from sqlalchemy.cyextension.util import (  # noqa: F401,E501
+        prefix_anon_map as prefix_anon_map,
+    )
+    from sqlalchemy.cyextension.util import (  # noqa: F401,E501
+        cache_anon_map as anon_map,
+    )
+
+
+__all__ = [
+    "iterate",
+    "traverse_using",
+    "traverse",
+    "cloned_traverse",
+    "replacement_traverse",
+    "Visitable",
+    "ExternalTraversal",
+    "InternalTraversal",
+    "anon_map",
+]
+
+
+class _CompilerDispatchType(Protocol):
+    def __call__(_self, self: Visitable, visitor: Any, **kw: Any) -> Any: ...
+
+
+class Visitable:
+    """Base class for visitable objects.
+
+    :class:`.Visitable` is used to implement the SQL compiler dispatch
+    functions.    Other forms of traversal such as for cache key generation
+    are implemented separately using the :class:`.HasTraverseInternals`
+    interface.
+
+    .. versionchanged:: 2.0  The :class:`.Visitable` class was named
+       :class:`.Traversible` in the 1.4 series; the name is changed back
+       to :class:`.Visitable` in 2.0 which is what it was prior to 1.4.
+
+       Both names remain importable in both 1.4 and 2.0 versions.
+
+    """
+
+    __slots__ = ()
+
+    __visit_name__: str
+
+    _original_compiler_dispatch: _CompilerDispatchType
+
+    if typing.TYPE_CHECKING:
+
+        def _compiler_dispatch(self, visitor: Any, **kw: Any) -> str: ...
+
+    def __init_subclass__(cls) -> None:
+        if "__visit_name__" in cls.__dict__:
+            cls._generate_compiler_dispatch()
+        super().__init_subclass__()
+
+    @classmethod
+    def _generate_compiler_dispatch(cls) -> None:
+        visit_name = cls.__visit_name__
+
+        if "_compiler_dispatch" in cls.__dict__:
+            # class has a fixed _compiler_dispatch() method.
+            # copy it to "original" so that we can get it back if
+            # sqlalchemy.ext.compiles overrides it.
+            cls._original_compiler_dispatch = cls._compiler_dispatch
+            return
+
+        if not isinstance(visit_name, str):
+            raise exc.InvalidRequestError(
+                f"__visit_name__ on class {cls.__name__} must be a string "
+                "at the class level"
+            )
+
+        name = "visit_%s" % visit_name
+        getter = operator.attrgetter(name)
+
+        def _compiler_dispatch(
+            self: Visitable, visitor: Any, **kw: Any
+        ) -> str:
+            """Look for an attribute named "visit_<visit_name>" on the
+            visitor, and call it with the same kw params.
+
+            """
+            try:
+                meth = getter(visitor)
+            except AttributeError as err:
+                return visitor.visit_unsupported_compilation(self, err, **kw)  # type: ignore  # noqa: E501
+            else:
+                return meth(self, **kw)  # type: ignore  # noqa: E501
+
+        cls._compiler_dispatch = (  # type: ignore
+            cls._original_compiler_dispatch
+        ) = _compiler_dispatch
+
+    def __class_getitem__(cls, key: Any) -> Any:
+        # allow generic classes in py3.9+
+        return cls
+
+
+class InternalTraversal(Enum):
+    r"""Defines visitor symbols used for internal traversal.
+
+    The :class:`.InternalTraversal` class is used in two ways.  One is that
+    it can serve as the superclass for an object that implements the
+    various visit methods of the class.   The other is that the symbols
+    themselves of :class:`.InternalTraversal` are used within
+    the ``_traverse_internals`` collection.   Such as, the :class:`.Case`
+    object defines ``_traverse_internals`` as ::
+
+        class Case(ColumnElement[_T]):
+            _traverse_internals = [
+                ("value", InternalTraversal.dp_clauseelement),
+                ("whens", InternalTraversal.dp_clauseelement_tuples),
+                ("else_", InternalTraversal.dp_clauseelement),
+            ]
+
+    Above, the :class:`.Case` class indicates its internal state as the
+    attributes named ``value``, ``whens``, and ``else_``.    They each
+    link to an :class:`.InternalTraversal` method which indicates the type
+    of datastructure to which each attribute refers.
+
+    Using the ``_traverse_internals`` structure, objects of type
+    :class:`.InternalTraversible` will have the following methods automatically
+    implemented:
+
+    * :meth:`.HasTraverseInternals.get_children`
+
+    * :meth:`.HasTraverseInternals._copy_internals`
+
+    * :meth:`.HasCacheKey._gen_cache_key`
+
+    Subclasses can also implement these methods directly, particularly for the
+    :meth:`.HasTraverseInternals._copy_internals` method, when special steps
+    are needed.
+
+    .. versionadded:: 1.4
+
+    """
+
+    dp_has_cache_key = "HC"
+    """Visit a :class:`.HasCacheKey` object."""
+
+    dp_has_cache_key_list = "HL"
+    """Visit a list of :class:`.HasCacheKey` objects."""
+
+    dp_clauseelement = "CE"
+    """Visit a :class:`_expression.ClauseElement` object."""
+
+    dp_fromclause_canonical_column_collection = "FC"
+    """Visit a :class:`_expression.FromClause` object in the context of the
+    ``columns`` attribute.
+
+    The column collection is "canonical", meaning it is the originally
+    defined location of the :class:`.ColumnClause` objects.   Right now
+    this means that the object being visited is a
+    :class:`_expression.TableClause`
+    or :class:`_schema.Table` object only.
+
+    """
+
+    dp_clauseelement_tuples = "CTS"
+    """Visit a list of tuples which contain :class:`_expression.ClauseElement`
+    objects.
+
+    """
+
+    dp_clauseelement_list = "CL"
+    """Visit a list of :class:`_expression.ClauseElement` objects.
+
+    """
+
+    dp_clauseelement_tuple = "CT"
+    """Visit a tuple of :class:`_expression.ClauseElement` objects.
+
+    """
+
+    dp_executable_options = "EO"
+
+    dp_with_context_options = "WC"
+
+    dp_fromclause_ordered_set = "CO"
+    """Visit an ordered set of :class:`_expression.FromClause` objects. """
+
+    dp_string = "S"
+    """Visit a plain string value.
+
+    Examples include table and column names, bound parameter keys, special
+    keywords such as "UNION", "UNION ALL".
+
+    The string value is considered to be significant for cache key
+    generation.
+
+    """
+
+    dp_string_list = "SL"
+    """Visit a list of strings."""
+
+    dp_anon_name = "AN"
+    """Visit a potentially "anonymized" string value.
+
+    The string value is considered to be significant for cache key
+    generation.
+
+    """
+
+    dp_boolean = "B"
+    """Visit a boolean value.
+
+    The boolean value is considered to be significant for cache key
+    generation.
+
+    """
+
+    dp_operator = "O"
+    """Visit an operator.
+
+    The operator is a function from the :mod:`sqlalchemy.sql.operators`
+    module.
+
+    The operator value is considered to be significant for cache key
+    generation.
+
+    """
+
+    dp_type = "T"
+    """Visit a :class:`.TypeEngine` object
+
+    The type object is considered to be significant for cache key
+    generation.
+
+    """
+
+    dp_plain_dict = "PD"
+    """Visit a dictionary with string keys.
+
+    The keys of the dictionary should be strings, the values should
+    be immutable and hashable.   The dictionary is considered to be
+    significant for cache key generation.
+
+    """
+
+    dp_dialect_options = "DO"
+    """Visit a dialect options structure."""
+
+    dp_string_clauseelement_dict = "CD"
+    """Visit a dictionary of string keys to :class:`_expression.ClauseElement`
+    objects.
+
+    """
+
+    dp_string_multi_dict = "MD"
+    """Visit a dictionary of string keys to values which may either be
+    plain immutable/hashable or :class:`.HasCacheKey` objects.
+
+    """
+
+    dp_annotations_key = "AK"
+    """Visit the _annotations_cache_key element.
+
+    This is a dictionary of additional information about a ClauseElement
+    that modifies its role.  It should be included when comparing or caching
+    objects, however generating this key is relatively expensive.   Visitors
+    should check the "_annotations" dict for non-None first before creating
+    this key.
+
+    """
+
+    dp_plain_obj = "PO"
+    """Visit a plain python object.
+
+    The value should be immutable and hashable, such as an integer.
+    The value is considered to be significant for cache key generation.
+
+    """
+
+    dp_named_ddl_element = "DD"
+    """Visit a simple named DDL element.
+
+    The current object used by this method is the :class:`.Sequence`.
+
+    The object is only considered to be important for cache key generation
+    as far as its name, but not any other aspects of it.
+
+    """
+
+    dp_prefix_sequence = "PS"
+    """Visit the sequence represented by :class:`_expression.HasPrefixes`
+    or :class:`_expression.HasSuffixes`.
+
+    """
+
+    dp_table_hint_list = "TH"
+    """Visit the ``_hints`` collection of a :class:`_expression.Select`
+    object.
+
+    """
+
+    dp_setup_join_tuple = "SJ"
+
+    dp_memoized_select_entities = "ME"
+
+    dp_statement_hint_list = "SH"
+    """Visit the ``_statement_hints`` collection of a
+    :class:`_expression.Select`
+    object.
+
+    """
+
+    dp_unknown_structure = "UK"
+    """Visit an unknown structure.
+
+    """
+
+    dp_dml_ordered_values = "DML_OV"
+    """Visit the values() ordered tuple list of an
+    :class:`_expression.Update` object."""
+
+    dp_dml_values = "DML_V"
+    """Visit the values() dictionary of a :class:`.ValuesBase`
+    (e.g. Insert or Update) object.
+
+    """
+
+    dp_dml_multi_values = "DML_MV"
+    """Visit the values() multi-valued list of dictionaries of an
+    :class:`_expression.Insert` object.
+
+    """
+
+    dp_propagate_attrs = "PA"
+    """Visit the propagate attrs dict.  This hardcodes to the particular
+    elements we care about right now."""
+
+    """Symbols that follow are additional symbols that are useful in
+    caching applications.
+
+    Traversals for :class:`_expression.ClauseElement` objects only need to use
+    those symbols present in :class:`.InternalTraversal`.  However, for
+    additional caching use cases within the ORM, symbols dealing with the
+    :class:`.HasCacheKey` class are added here.
+
+    """
+
+    dp_ignore = "IG"
+    """Specify an object that should be ignored entirely.
+
+    This currently applies function call argument caching where some
+    arguments should not be considered to be part of a cache key.
+
+    """
+
+    dp_inspectable = "IS"
+    """Visit an inspectable object where the return value is a
+    :class:`.HasCacheKey` object."""
+
+    dp_multi = "M"
+    """Visit an object that may be a :class:`.HasCacheKey` or may be a
+    plain hashable object."""
+
+    dp_multi_list = "MT"
+    """Visit a tuple containing elements that may be :class:`.HasCacheKey` or
+    may be a plain hashable object."""
+
+    dp_has_cache_key_tuples = "HT"
+    """Visit a list of tuples which contain :class:`.HasCacheKey`
+    objects.
+
+    """
+
+    dp_inspectable_list = "IL"
+    """Visit a list of inspectable objects which upon inspection are
+    HasCacheKey objects."""
+
+
+_TraverseInternalsType = List[Tuple[str, InternalTraversal]]
+"""a structure that defines how a HasTraverseInternals should be
+traversed.
+
+This structure consists of a list of (attributename, internaltraversal)
+tuples, where the "attributename" refers to the name of an attribute on an
+instance of the HasTraverseInternals object, and "internaltraversal" refers
+to an :class:`.InternalTraversal` enumeration symbol defining what kind
+of data this attribute stores, which indicates to the traverser how it should
+be handled.
+
+"""
+
+
+class HasTraverseInternals:
+    """base for classes that have a "traverse internals" element,
+    which defines all kinds of ways of traversing the elements of an object.
+
+    Compared to :class:`.Visitable`, which relies upon an external visitor to
+    define how the object is travered (i.e. the :class:`.SQLCompiler`), the
+    :class:`.HasTraverseInternals` interface allows classes to define their own
+    traversal, that is, what attributes are accessed and in what order.
+
+    """
+
+    __slots__ = ()
+
+    _traverse_internals: _TraverseInternalsType
+
+    _is_immutable: bool = False
+
+    @util.preload_module("sqlalchemy.sql.traversals")
+    def get_children(
+        self, *, omit_attrs: Tuple[str, ...] = (), **kw: Any
+    ) -> Iterable[HasTraverseInternals]:
+        r"""Return immediate child :class:`.visitors.HasTraverseInternals`
+        elements of this :class:`.visitors.HasTraverseInternals`.
+
+        This is used for visit traversal.
+
+        \**kw may contain flags that change the collection that is
+        returned, for example to return a subset of items in order to
+        cut down on larger traversals, or to return child items from a
+        different context (such as schema-level collections instead of
+        clause-level).
+
+        """
+
+        traversals = util.preloaded.sql_traversals
+
+        try:
+            traverse_internals = self._traverse_internals
+        except AttributeError:
+            # user-defined classes may not have a _traverse_internals
+            return []
+
+        dispatch = traversals._get_children.run_generated_dispatch
+        return itertools.chain.from_iterable(
+            meth(obj, **kw)
+            for attrname, obj, meth in dispatch(
+                self, traverse_internals, "_generated_get_children_traversal"
+            )
+            if attrname not in omit_attrs and obj is not None
+        )
+
+
+class _InternalTraversalDispatchType(Protocol):
+    def __call__(s, self: object, visitor: HasTraversalDispatch) -> Any: ...
+
+
+class HasTraversalDispatch:
+    r"""Define infrastructure for classes that perform internal traversals
+
+    .. versionadded:: 2.0
+
+    """
+
+    __slots__ = ()
+
+    _dispatch_lookup: ClassVar[Dict[Union[InternalTraversal, str], str]] = {}
+
+    def dispatch(self, visit_symbol: InternalTraversal) -> Callable[..., Any]:
+        """Given a method from :class:`.HasTraversalDispatch`, return the
+        corresponding method on a subclass.
+
+        """
+        name = _dispatch_lookup[visit_symbol]
+        return getattr(self, name, None)  # type: ignore
+
+    def run_generated_dispatch(
+        self,
+        target: object,
+        internal_dispatch: _TraverseInternalsType,
+        generate_dispatcher_name: str,
+    ) -> Any:
+        dispatcher: _InternalTraversalDispatchType
+        try:
+            dispatcher = target.__class__.__dict__[generate_dispatcher_name]
+        except KeyError:
+            # traversals.py -> _preconfigure_traversals()
+            # may be used to run these ahead of time, but
+            # is not enabled right now.
+            # this block will generate any remaining dispatchers.
+            dispatcher = self.generate_dispatch(
+                target.__class__, internal_dispatch, generate_dispatcher_name
+            )
+        return dispatcher(target, self)
+
+    def generate_dispatch(
+        self,
+        target_cls: Type[object],
+        internal_dispatch: _TraverseInternalsType,
+        generate_dispatcher_name: str,
+    ) -> _InternalTraversalDispatchType:
+        dispatcher = self._generate_dispatcher(
+            internal_dispatch, generate_dispatcher_name
+        )
+        # assert isinstance(target_cls, type)
+        setattr(target_cls, generate_dispatcher_name, dispatcher)
+        return dispatcher
+
+    def _generate_dispatcher(
+        self, internal_dispatch: _TraverseInternalsType, method_name: str
+    ) -> _InternalTraversalDispatchType:
+        names = []
+        for attrname, visit_sym in internal_dispatch:
+            meth = self.dispatch(visit_sym)
+            if meth is not None:
+                visit_name = _dispatch_lookup[visit_sym]
+                names.append((attrname, visit_name))
+
+        code = (
+            ("    return [\n")
+            + (
+                ", \n".join(
+                    "        (%r, self.%s, visitor.%s)"
+                    % (attrname, attrname, visit_name)
+                    for attrname, visit_name in names
+                )
+            )
+            + ("\n    ]\n")
+        )
+        meth_text = ("def %s(self, visitor):\n" % method_name) + code + "\n"
+        return cast(
+            _InternalTraversalDispatchType,
+            langhelpers._exec_code_in_env(meth_text, {}, method_name),
+        )
+
+
+ExtendedInternalTraversal = InternalTraversal
+
+
+def _generate_traversal_dispatch() -> None:
+    lookup = _dispatch_lookup
+
+    for sym in InternalTraversal:
+        key = sym.name
+        if key.startswith("dp_"):
+            visit_key = key.replace("dp_", "visit_")
+            sym_name = sym.value
+            assert sym_name not in lookup, sym_name
+            lookup[sym] = lookup[sym_name] = visit_key
+
+
+_dispatch_lookup = HasTraversalDispatch._dispatch_lookup
+_generate_traversal_dispatch()
+
+
+class ExternallyTraversible(HasTraverseInternals, Visitable):
+    __slots__ = ()
+
+    _annotations: Mapping[Any, Any] = util.EMPTY_DICT
+
+    if typing.TYPE_CHECKING:
+
+        def _annotate(self, values: _AnnotationDict) -> Self: ...
+
+        def get_children(
+            self, *, omit_attrs: Tuple[str, ...] = (), **kw: Any
+        ) -> Iterable[ExternallyTraversible]: ...
+
+    def _clone(self, **kw: Any) -> Self:
+        """clone this element"""
+        raise NotImplementedError()
+
+    def _copy_internals(
+        self, *, omit_attrs: Tuple[str, ...] = (), **kw: Any
+    ) -> None:
+        """Reassign internal elements to be clones of themselves.
+
+        Called during a copy-and-traverse operation on newly
+        shallow-copied elements to create a deep copy.
+
+        The given clone function should be used, which may be applying
+        additional transformations to the element (i.e. replacement
+        traversal, cloned traversal, annotations).
+
+        """
+        raise NotImplementedError()
+
+
+_ET = TypeVar("_ET", bound=ExternallyTraversible)
+
+_CE = TypeVar("_CE", bound="ColumnElement[Any]")
+
+_TraverseCallableType = Callable[[_ET], None]
+
+
+class _CloneCallableType(Protocol):
+    def __call__(self, element: _ET, **kw: Any) -> _ET: ...
+
+
+class _TraverseTransformCallableType(Protocol[_ET]):
+    def __call__(self, element: _ET, **kw: Any) -> Optional[_ET]: ...
+
+
+_ExtT = TypeVar("_ExtT", bound="ExternalTraversal")
+
+
+class ExternalTraversal(util.MemoizedSlots):
+    """Base class for visitor objects which can traverse externally using
+    the :func:`.visitors.traverse` function.
+
+    Direct usage of the :func:`.visitors.traverse` function is usually
+    preferred.
+
+    """
+
+    __slots__ = ("_visitor_dict", "_next")
+
+    __traverse_options__: Dict[str, Any] = {}
+    _next: Optional[ExternalTraversal]
+
+    def traverse_single(self, obj: Visitable, **kw: Any) -> Any:
+        for v in self.visitor_iterator:
+            meth = getattr(v, "visit_%s" % obj.__visit_name__, None)
+            if meth:
+                return meth(obj, **kw)
+
+    def iterate(
+        self, obj: Optional[ExternallyTraversible]
+    ) -> Iterator[ExternallyTraversible]:
+        """Traverse the given expression structure, returning an iterator
+        of all elements.
+
+        """
+        return iterate(obj, self.__traverse_options__)
+
+    @overload
+    def traverse(self, obj: Literal[None]) -> None: ...
+
+    @overload
+    def traverse(
+        self, obj: ExternallyTraversible
+    ) -> ExternallyTraversible: ...
+
+    def traverse(
+        self, obj: Optional[ExternallyTraversible]
+    ) -> Optional[ExternallyTraversible]:
+        """Traverse and visit the given expression structure."""
+
+        return traverse(obj, self.__traverse_options__, self._visitor_dict)
+
+    def _memoized_attr__visitor_dict(
+        self,
+    ) -> Dict[str, _TraverseCallableType[Any]]:
+        visitors = {}
+
+        for name in dir(self):
+            if name.startswith("visit_"):
+                visitors[name[6:]] = getattr(self, name)
+        return visitors
+
+    @property
+    def visitor_iterator(self) -> Iterator[ExternalTraversal]:
+        """Iterate through this visitor and each 'chained' visitor."""
+
+        v: Optional[ExternalTraversal] = self
+        while v:
+            yield v
+            v = getattr(v, "_next", None)
+
+    def chain(self: _ExtT, visitor: ExternalTraversal) -> _ExtT:
+        """'Chain' an additional ExternalTraversal onto this ExternalTraversal
+
+        The chained visitor will receive all visit events after this one.
+
+        """
+        tail = list(self.visitor_iterator)[-1]
+        tail._next = visitor
+        return self
+
+
+class CloningExternalTraversal(ExternalTraversal):
+    """Base class for visitor objects which can traverse using
+    the :func:`.visitors.cloned_traverse` function.
+
+    Direct usage of the :func:`.visitors.cloned_traverse` function is usually
+    preferred.
+
+
+    """
+
+    __slots__ = ()
+
+    def copy_and_process(
+        self, list_: List[ExternallyTraversible]
+    ) -> List[ExternallyTraversible]:
+        """Apply cloned traversal to the given list of elements, and return
+        the new list.
+
+        """
+        return [self.traverse(x) for x in list_]
+
+    @overload
+    def traverse(self, obj: Literal[None]) -> None: ...
+
+    @overload
+    def traverse(
+        self, obj: ExternallyTraversible
+    ) -> ExternallyTraversible: ...
+
+    def traverse(
+        self, obj: Optional[ExternallyTraversible]
+    ) -> Optional[ExternallyTraversible]:
+        """Traverse and visit the given expression structure."""
+
+        return cloned_traverse(
+            obj, self.__traverse_options__, self._visitor_dict
+        )
+
+
+class ReplacingExternalTraversal(CloningExternalTraversal):
+    """Base class for visitor objects which can traverse using
+    the :func:`.visitors.replacement_traverse` function.
+
+    Direct usage of the :func:`.visitors.replacement_traverse` function is
+    usually preferred.
+
+    """
+
+    __slots__ = ()
+
+    def replace(
+        self, elem: ExternallyTraversible
+    ) -> Optional[ExternallyTraversible]:
+        """Receive pre-copied elements during a cloning traversal.
+
+        If the method returns a new element, the element is used
+        instead of creating a simple copy of the element.  Traversal
+        will halt on the newly returned element if it is re-encountered.
+        """
+        return None
+
+    @overload
+    def traverse(self, obj: Literal[None]) -> None: ...
+
+    @overload
+    def traverse(
+        self, obj: ExternallyTraversible
+    ) -> ExternallyTraversible: ...
+
+    def traverse(
+        self, obj: Optional[ExternallyTraversible]
+    ) -> Optional[ExternallyTraversible]:
+        """Traverse and visit the given expression structure."""
+
+        def replace(
+            element: ExternallyTraversible,
+            **kw: Any,
+        ) -> Optional[ExternallyTraversible]:
+            for v in self.visitor_iterator:
+                e = cast(ReplacingExternalTraversal, v).replace(element)
+                if e is not None:
+                    return e
+
+            return None
+
+        return replacement_traverse(obj, self.__traverse_options__, replace)
+
+
+# backwards compatibility
+Traversible = Visitable
+
+ClauseVisitor = ExternalTraversal
+CloningVisitor = CloningExternalTraversal
+ReplacingCloningVisitor = ReplacingExternalTraversal
+
+
+def iterate(
+    obj: Optional[ExternallyTraversible],
+    opts: Mapping[str, Any] = util.EMPTY_DICT,
+) -> Iterator[ExternallyTraversible]:
+    r"""Traverse the given expression structure, returning an iterator.
+
+    Traversal is configured to be breadth-first.
+
+    The central API feature used by the :func:`.visitors.iterate`
+    function is the
+    :meth:`_expression.ClauseElement.get_children` method of
+    :class:`_expression.ClauseElement` objects.  This method should return all
+    the :class:`_expression.ClauseElement` objects which are associated with a
+    particular :class:`_expression.ClauseElement` object. For example, a
+    :class:`.Case` structure will refer to a series of
+    :class:`_expression.ColumnElement` objects within its "whens" and "else\_"
+    member variables.
+
+    :param obj: :class:`_expression.ClauseElement` structure to be traversed
+
+    :param opts: dictionary of iteration options.   This dictionary is usually
+     empty in modern usage.
+
+    """
+    if obj is None:
+        return
+
+    yield obj
+    children = obj.get_children(**opts)
+
+    if not children:
+        return
+
+    stack = deque([children])
+    while stack:
+        t_iterator = stack.popleft()
+        for t in t_iterator:
+            yield t
+            stack.append(t.get_children(**opts))
+
+
+@overload
+def traverse_using(
+    iterator: Iterable[ExternallyTraversible],
+    obj: Literal[None],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> None: ...
+
+
+@overload
+def traverse_using(
+    iterator: Iterable[ExternallyTraversible],
+    obj: ExternallyTraversible,
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> ExternallyTraversible: ...
+
+
+def traverse_using(
+    iterator: Iterable[ExternallyTraversible],
+    obj: Optional[ExternallyTraversible],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> Optional[ExternallyTraversible]:
+    """Visit the given expression structure using the given iterator of
+    objects.
+
+    :func:`.visitors.traverse_using` is usually called internally as the result
+    of the :func:`.visitors.traverse` function.
+
+    :param iterator: an iterable or sequence which will yield
+     :class:`_expression.ClauseElement`
+     structures; the iterator is assumed to be the
+     product of the :func:`.visitors.iterate` function.
+
+    :param obj: the :class:`_expression.ClauseElement`
+     that was used as the target of the
+     :func:`.iterate` function.
+
+    :param visitors: dictionary of visit functions.  See :func:`.traverse`
+     for details on this dictionary.
+
+    .. seealso::
+
+        :func:`.traverse`
+
+
+    """
+    for target in iterator:
+        meth = visitors.get(target.__visit_name__, None)
+        if meth:
+            meth(target)
+    return obj
+
+
+@overload
+def traverse(
+    obj: Literal[None],
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> None: ...
+
+
+@overload
+def traverse(
+    obj: ExternallyTraversible,
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> ExternallyTraversible: ...
+
+
+def traverse(
+    obj: Optional[ExternallyTraversible],
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> Optional[ExternallyTraversible]:
+    """Traverse and visit the given expression structure using the default
+    iterator.
+
+     e.g.::
+
+        from sqlalchemy.sql import visitors
+
+        stmt = select(some_table).where(some_table.c.foo == "bar")
+
+
+        def visit_bindparam(bind_param):
+            print("found bound value: %s" % bind_param.value)
+
+
+        visitors.traverse(stmt, {}, {"bindparam": visit_bindparam})
+
+    The iteration of objects uses the :func:`.visitors.iterate` function,
+    which does a breadth-first traversal using a stack.
+
+    :param obj: :class:`_expression.ClauseElement` structure to be traversed
+
+    :param opts: dictionary of iteration options.   This dictionary is usually
+     empty in modern usage.
+
+    :param visitors: dictionary of visit functions.   The dictionary should
+     have strings as keys, each of which would correspond to the
+     ``__visit_name__`` of a particular kind of SQL expression object, and
+     callable functions  as values, each of which represents a visitor function
+     for that kind of object.
+
+    """
+    return traverse_using(iterate(obj, opts), obj, visitors)
+
+
+@overload
+def cloned_traverse(
+    obj: Literal[None],
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> None: ...
+
+
+# a bit of controversy here, as the clone of the lead element
+# *could* in theory replace with an entirely different kind of element.
+# however this is really not how cloned_traverse is ever used internally
+# at least.
+@overload
+def cloned_traverse(
+    obj: _ET,
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> _ET: ...
+
+
+def cloned_traverse(
+    obj: Optional[ExternallyTraversible],
+    opts: Mapping[str, Any],
+    visitors: Mapping[str, _TraverseCallableType[Any]],
+) -> Optional[ExternallyTraversible]:
+    """Clone the given expression structure, allowing modifications by
+    visitors for mutable objects.
+
+    Traversal usage is the same as that of :func:`.visitors.traverse`.
+    The visitor functions present in the ``visitors`` dictionary may also
+    modify the internals of the given structure as the traversal proceeds.
+
+    The :func:`.cloned_traverse` function does **not** provide objects that are
+    part of the :class:`.Immutable` interface to the visit methods (this
+    primarily includes :class:`.ColumnClause`, :class:`.Column`,
+    :class:`.TableClause` and :class:`.Table` objects). As this traversal is
+    only intended to allow in-place mutation of objects, :class:`.Immutable`
+    objects are skipped. The :meth:`.Immutable._clone` method is still called
+    on each object to allow for objects to replace themselves with a different
+    object based on a clone of their sub-internals (e.g. a
+    :class:`.ColumnClause` that clones its subquery to return a new
+    :class:`.ColumnClause`).
+
+    .. versionchanged:: 2.0  The :func:`.cloned_traverse` function omits
+       objects that are part of the :class:`.Immutable` interface.
+
+    The central API feature used by the :func:`.visitors.cloned_traverse`
+    and :func:`.visitors.replacement_traverse` functions, in addition to the
+    :meth:`_expression.ClauseElement.get_children`
+    function that is used to achieve
+    the iteration, is the :meth:`_expression.ClauseElement._copy_internals`
+    method.
+    For a :class:`_expression.ClauseElement`
+    structure to support cloning and replacement
+    traversals correctly, it needs to be able to pass a cloning function into
+    its internal members in order to make copies of them.
+
+    .. seealso::
+
+        :func:`.visitors.traverse`
+
+        :func:`.visitors.replacement_traverse`
+
+    """
+
+    cloned: Dict[int, ExternallyTraversible] = {}
+    stop_on = set(opts.get("stop_on", []))
+
+    def deferred_copy_internals(
+        obj: ExternallyTraversible,
+    ) -> ExternallyTraversible:
+        return cloned_traverse(obj, opts, visitors)
+
+    def clone(elem: ExternallyTraversible, **kw: Any) -> ExternallyTraversible:
+        if elem in stop_on:
+            return elem
+        else:
+            if id(elem) not in cloned:
+                if "replace" in kw:
+                    newelem = cast(
+                        Optional[ExternallyTraversible], kw["replace"](elem)
+                    )
+                    if newelem is not None:
+                        cloned[id(elem)] = newelem
+                        return newelem
+
+                # the _clone method for immutable normally returns "self".
+                # however, the method is still allowed to return a
+                # different object altogether; ColumnClause._clone() will
+                # based on options clone the subquery to which it is associated
+                # and return the new corresponding column.
+                cloned[id(elem)] = newelem = elem._clone(clone=clone, **kw)
+                newelem._copy_internals(clone=clone, **kw)
+
+                # however, visit methods which are tasked with in-place
+                # mutation of the object should not get access to the immutable
+                # object.
+                if not elem._is_immutable:
+                    meth = visitors.get(newelem.__visit_name__, None)
+                    if meth:
+                        meth(newelem)
+            return cloned[id(elem)]
+
+    if obj is not None:
+        obj = clone(
+            obj, deferred_copy_internals=deferred_copy_internals, **opts
+        )
+    clone = None  # type: ignore[assignment]  # remove gc cycles
+    return obj
+
+
+@overload
+def replacement_traverse(
+    obj: Literal[None],
+    opts: Mapping[str, Any],
+    replace: _TraverseTransformCallableType[Any],
+) -> None: ...
+
+
+@overload
+def replacement_traverse(
+    obj: _CE,
+    opts: Mapping[str, Any],
+    replace: _TraverseTransformCallableType[Any],
+) -> _CE: ...
+
+
+@overload
+def replacement_traverse(
+    obj: ExternallyTraversible,
+    opts: Mapping[str, Any],
+    replace: _TraverseTransformCallableType[Any],
+) -> ExternallyTraversible: ...
+
+
+def replacement_traverse(
+    obj: Optional[ExternallyTraversible],
+    opts: Mapping[str, Any],
+    replace: _TraverseTransformCallableType[Any],
+) -> Optional[ExternallyTraversible]:
+    """Clone the given expression structure, allowing element
+    replacement by a given replacement function.
+
+    This function is very similar to the :func:`.visitors.cloned_traverse`
+    function, except instead of being passed a dictionary of visitors, all
+    elements are unconditionally passed into the given replace function.
+    The replace function then has the option to return an entirely new object
+    which will replace the one given.  If it returns ``None``, then the object
+    is kept in place.
+
+    The difference in usage between :func:`.visitors.cloned_traverse` and
+    :func:`.visitors.replacement_traverse` is that in the former case, an
+    already-cloned object is passed to the visitor function, and the visitor
+    function can then manipulate the internal state of the object.
+    In the case of the latter, the visitor function should only return an
+    entirely different object, or do nothing.
+
+    The use case for :func:`.visitors.replacement_traverse` is that of
+    replacing a FROM clause inside of a SQL structure with a different one,
+    as is a common use case within the ORM.
+
+    """
+
+    cloned = {}
+    stop_on = {id(x) for x in opts.get("stop_on", [])}
+
+    def deferred_copy_internals(
+        obj: ExternallyTraversible,
+    ) -> ExternallyTraversible:
+        return replacement_traverse(obj, opts, replace)
+
+    def clone(elem: ExternallyTraversible, **kw: Any) -> ExternallyTraversible:
+        if (
+            id(elem) in stop_on
+            or "no_replacement_traverse" in elem._annotations
+        ):
+            return elem
+        else:
+            newelem = replace(elem)
+            if newelem is not None:
+                stop_on.add(id(newelem))
+                return newelem  # type: ignore
+            else:
+                # base "already seen" on id(), not hash, so that we don't
+                # replace an Annotated element with its non-annotated one, and
+                # vice versa
+                id_elem = id(elem)
+                if id_elem not in cloned:
+                    if "replace" in kw:
+                        newelem = kw["replace"](elem)
+                        if newelem is not None:
+                            cloned[id_elem] = newelem
+                            return newelem  # type: ignore
+
+                    cloned[id_elem] = newelem = elem._clone(**kw)
+                    newelem._copy_internals(clone=clone, **kw)
+                return cloned[id_elem]  # type: ignore
+
+    if obj is not None:
+        obj = clone(
+            obj, deferred_copy_internals=deferred_copy_internals, **opts
+        )
+    clone = None  # type: ignore[assignment]  # remove gc cycles
+    return obj
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__init__.py
new file mode 100644
index 00000000..4e574bbb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__init__.py
@@ -0,0 +1,96 @@
+# testing/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from unittest import mock
+
+from . import config
+from .assertions import assert_raises
+from .assertions import assert_raises_context_ok
+from .assertions import assert_raises_message
+from .assertions import assert_raises_message_context_ok
+from .assertions import assert_warns
+from .assertions import assert_warns_message
+from .assertions import AssertsCompiledSQL
+from .assertions import AssertsExecutionResults
+from .assertions import ComparesIndexes
+from .assertions import ComparesTables
+from .assertions import emits_warning
+from .assertions import emits_warning_on
+from .assertions import eq_
+from .assertions import eq_ignore_whitespace
+from .assertions import eq_regex
+from .assertions import expect_deprecated
+from .assertions import expect_deprecated_20
+from .assertions import expect_raises
+from .assertions import expect_raises_message
+from .assertions import expect_warnings
+from .assertions import in_
+from .assertions import int_within_variance
+from .assertions import is_
+from .assertions import is_false
+from .assertions import is_instance_of
+from .assertions import is_none
+from .assertions import is_not
+from .assertions import is_not_
+from .assertions import is_not_none
+from .assertions import is_true
+from .assertions import le_
+from .assertions import ne_
+from .assertions import not_in
+from .assertions import not_in_
+from .assertions import startswith_
+from .assertions import uses_deprecated
+from .config import add_to_marker
+from .config import async_test
+from .config import combinations
+from .config import combinations_list
+from .config import db
+from .config import fixture
+from .config import requirements as requires
+from .config import skip_test
+from .config import Variation
+from .config import variation
+from .config import variation_fixture
+from .exclusions import _is_excluded
+from .exclusions import _server_version
+from .exclusions import against as _against
+from .exclusions import db_spec
+from .exclusions import exclude
+from .exclusions import fails
+from .exclusions import fails_if
+from .exclusions import fails_on
+from .exclusions import fails_on_everything_except
+from .exclusions import future
+from .exclusions import only_if
+from .exclusions import only_on
+from .exclusions import skip
+from .exclusions import skip_if
+from .schema import eq_clause_element
+from .schema import eq_type_affinity
+from .util import adict
+from .util import fail
+from .util import flag_combinations
+from .util import force_drop_names
+from .util import lambda_combinations
+from .util import metadata_fixture
+from .util import provide_metadata
+from .util import resolve_lambda
+from .util import rowset
+from .util import run_as_contextmanager
+from .util import skip_if_timeout
+from .util import teardown_events
+from .warnings import assert_warnings
+from .warnings import warn_test_suite
+
+
+def against(*queries):
+    return _against(config._current, *queries)
+
+
+crashes = skip
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..22f06e14
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertions.cpython-312.pyc
new file mode 100644
index 00000000..ed38ab58
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertsql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertsql.cpython-312.pyc
new file mode 100644
index 00000000..991ed4df
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/assertsql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/asyncio.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/asyncio.cpython-312.pyc
new file mode 100644
index 00000000..4915ab6d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/asyncio.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/config.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/config.cpython-312.pyc
new file mode 100644
index 00000000..aa18c40b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/config.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/engines.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/engines.cpython-312.pyc
new file mode 100644
index 00000000..0ada3128
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/engines.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/entities.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/entities.cpython-312.pyc
new file mode 100644
index 00000000..71bc905f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/entities.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/exclusions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/exclusions.cpython-312.pyc
new file mode 100644
index 00000000..e727a61a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/exclusions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/pickleable.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/pickleable.cpython-312.pyc
new file mode 100644
index 00000000..d3f870db
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/pickleable.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/profiling.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/profiling.cpython-312.pyc
new file mode 100644
index 00000000..c57669e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/profiling.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/provision.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/provision.cpython-312.pyc
new file mode 100644
index 00000000..1240e476
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/provision.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/requirements.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/requirements.cpython-312.pyc
new file mode 100644
index 00000000..57f8b424
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/requirements.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/schema.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/schema.cpython-312.pyc
new file mode 100644
index 00000000..94c46f75
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/schema.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..9821569e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/warnings.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/warnings.cpython-312.pyc
new file mode 100644
index 00000000..ff26c835
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/__pycache__/warnings.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertions.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertions.py
new file mode 100644
index 00000000..8364c15f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertions.py
@@ -0,0 +1,989 @@
+# testing/assertions.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+from collections import defaultdict
+import contextlib
+from copy import copy
+from itertools import filterfalse
+import re
+import sys
+import warnings
+
+from . import assertsql
+from . import config
+from . import engines
+from . import mock
+from .exclusions import db_spec
+from .util import fail
+from .. import exc as sa_exc
+from .. import schema
+from .. import sql
+from .. import types as sqltypes
+from .. import util
+from ..engine import default
+from ..engine import url
+from ..sql.selectable import LABEL_STYLE_TABLENAME_PLUS_COL
+from ..util import decorator
+
+
+def expect_warnings(*messages, **kw):
+    """Context manager which expects one or more warnings.
+
+    With no arguments, squelches all SAWarning emitted via
+    sqlalchemy.util.warn and sqlalchemy.util.warn_limited.   Otherwise
+    pass string expressions that will match selected warnings via regex;
+    all non-matching warnings are sent through.
+
+    The expect version **asserts** that the warnings were in fact seen.
+
+    Note that the test suite sets SAWarning warnings to raise exceptions.
+
+    """  # noqa
+    return _expect_warnings_sqla_only(sa_exc.SAWarning, messages, **kw)
+
+
+@contextlib.contextmanager
+def expect_warnings_on(db, *messages, **kw):
+    """Context manager which expects one or more warnings on specific
+    dialects.
+
+    The expect version **asserts** that the warnings were in fact seen.
+
+    """
+    spec = db_spec(db)
+
+    if isinstance(db, str) and not spec(config._current):
+        yield
+    else:
+        with expect_warnings(*messages, **kw):
+            yield
+
+
+def emits_warning(*messages):
+    """Decorator form of expect_warnings().
+
+    Note that emits_warning does **not** assert that the warnings
+    were in fact seen.
+
+    """
+
+    @decorator
+    def decorate(fn, *args, **kw):
+        with expect_warnings(assert_=False, *messages):
+            return fn(*args, **kw)
+
+    return decorate
+
+
+def expect_deprecated(*messages, **kw):
+    return _expect_warnings_sqla_only(
+        sa_exc.SADeprecationWarning, messages, **kw
+    )
+
+
+def expect_deprecated_20(*messages, **kw):
+    return _expect_warnings_sqla_only(
+        sa_exc.Base20DeprecationWarning, messages, **kw
+    )
+
+
+def emits_warning_on(db, *messages):
+    """Mark a test as emitting a warning on a specific dialect.
+
+    With no arguments, squelches all SAWarning failures.  Or pass one or more
+    strings; these will be matched to the root of the warning description by
+    warnings.filterwarnings().
+
+    Note that emits_warning_on does **not** assert that the warnings
+    were in fact seen.
+
+    """
+
+    @decorator
+    def decorate(fn, *args, **kw):
+        with expect_warnings_on(db, assert_=False, *messages):
+            return fn(*args, **kw)
+
+    return decorate
+
+
+def uses_deprecated(*messages):
+    """Mark a test as immune from fatal deprecation warnings.
+
+    With no arguments, squelches all SADeprecationWarning failures.
+    Or pass one or more strings; these will be matched to the root
+    of the warning description by warnings.filterwarnings().
+
+    As a special case, you may pass a function name prefixed with //
+    and it will be re-written as needed to match the standard warning
+    verbiage emitted by the sqlalchemy.util.deprecated decorator.
+
+    Note that uses_deprecated does **not** assert that the warnings
+    were in fact seen.
+
+    """
+
+    @decorator
+    def decorate(fn, *args, **kw):
+        with expect_deprecated(*messages, assert_=False):
+            return fn(*args, **kw)
+
+    return decorate
+
+
+_FILTERS = None
+_SEEN = None
+_EXC_CLS = None
+
+
+def _expect_warnings_sqla_only(
+    exc_cls,
+    messages,
+    regex=True,
+    search_msg=False,
+    assert_=True,
+):
+    """SQLAlchemy internal use only _expect_warnings().
+
+    Alembic is using _expect_warnings() directly, and should be updated
+    to use this new interface.
+
+    """
+    return _expect_warnings(
+        exc_cls,
+        messages,
+        regex=regex,
+        search_msg=search_msg,
+        assert_=assert_,
+        raise_on_any_unexpected=True,
+    )
+
+
+@contextlib.contextmanager
+def _expect_warnings(
+    exc_cls,
+    messages,
+    regex=True,
+    search_msg=False,
+    assert_=True,
+    raise_on_any_unexpected=False,
+    squelch_other_warnings=False,
+):
+    global _FILTERS, _SEEN, _EXC_CLS
+
+    if regex or search_msg:
+        filters = [re.compile(msg, re.I | re.S) for msg in messages]
+    else:
+        filters = list(messages)
+
+    if _FILTERS is not None:
+        # nested call; update _FILTERS and _SEEN, return.  outer
+        # block will assert our messages
+        assert _SEEN is not None
+        assert _EXC_CLS is not None
+        _FILTERS.extend(filters)
+        _SEEN.update(filters)
+        _EXC_CLS += (exc_cls,)
+        yield
+    else:
+        seen = _SEEN = set(filters)
+        _FILTERS = filters
+        _EXC_CLS = (exc_cls,)
+
+        if raise_on_any_unexpected:
+
+            def real_warn(msg, *arg, **kw):
+                raise AssertionError("Got unexpected warning: %r" % msg)
+
+        else:
+            real_warn = warnings.warn
+
+        def our_warn(msg, *arg, **kw):
+            if isinstance(msg, _EXC_CLS):
+                exception = type(msg)
+                msg = str(msg)
+            elif arg:
+                exception = arg[0]
+            else:
+                exception = None
+
+            if not exception or not issubclass(exception, _EXC_CLS):
+                if not squelch_other_warnings:
+                    return real_warn(msg, *arg, **kw)
+                else:
+                    return
+
+            if not filters and not raise_on_any_unexpected:
+                return
+
+            for filter_ in filters:
+                if (
+                    (search_msg and filter_.search(msg))
+                    or (regex and filter_.match(msg))
+                    or (not regex and filter_ == msg)
+                ):
+                    seen.discard(filter_)
+                    break
+            else:
+                if not squelch_other_warnings:
+                    real_warn(msg, *arg, **kw)
+
+        with mock.patch("warnings.warn", our_warn):
+            try:
+                yield
+            finally:
+                _SEEN = _FILTERS = _EXC_CLS = None
+
+                if assert_:
+                    assert not seen, "Warnings were not seen: %s" % ", ".join(
+                        "%r" % (s.pattern if regex else s) for s in seen
+                    )
+
+
+def global_cleanup_assertions():
+    """Check things that have to be finalized at the end of a test suite.
+
+    Hardcoded at the moment, a modular system can be built here
+    to support things like PG prepared transactions, tables all
+    dropped, etc.
+
+    """
+    _assert_no_stray_pool_connections()
+
+
+def _assert_no_stray_pool_connections():
+    engines.testing_reaper.assert_all_closed()
+
+
+def int_within_variance(expected, received, variance):
+    deviance = int(expected * variance)
+    assert (
+        abs(received - expected) < deviance
+    ), "Given int value %s is not within %d%% of expected value %s" % (
+        received,
+        variance * 100,
+        expected,
+    )
+
+
+def eq_regex(a, b, msg=None):
+    assert re.match(b, a), msg or "%r !~ %r" % (a, b)
+
+
+def eq_(a, b, msg=None):
+    """Assert a == b, with repr messaging on failure."""
+    assert a == b, msg or "%r != %r" % (a, b)
+
+
+def ne_(a, b, msg=None):
+    """Assert a != b, with repr messaging on failure."""
+    assert a != b, msg or "%r == %r" % (a, b)
+
+
+def le_(a, b, msg=None):
+    """Assert a <= b, with repr messaging on failure."""
+    assert a <= b, msg or "%r != %r" % (a, b)
+
+
+def is_instance_of(a, b, msg=None):
+    assert isinstance(a, b), msg or "%r is not an instance of %r" % (a, b)
+
+
+def is_none(a, msg=None):
+    is_(a, None, msg=msg)
+
+
+def is_not_none(a, msg=None):
+    is_not(a, None, msg=msg)
+
+
+def is_true(a, msg=None):
+    is_(bool(a), True, msg=msg)
+
+
+def is_false(a, msg=None):
+    is_(bool(a), False, msg=msg)
+
+
+def is_(a, b, msg=None):
+    """Assert a is b, with repr messaging on failure."""
+    assert a is b, msg or "%r is not %r" % (a, b)
+
+
+def is_not(a, b, msg=None):
+    """Assert a is not b, with repr messaging on failure."""
+    assert a is not b, msg or "%r is %r" % (a, b)
+
+
+# deprecated.  See #5429
+is_not_ = is_not
+
+
+def in_(a, b, msg=None):
+    """Assert a in b, with repr messaging on failure."""
+    assert a in b, msg or "%r not in %r" % (a, b)
+
+
+def not_in(a, b, msg=None):
+    """Assert a in not b, with repr messaging on failure."""
+    assert a not in b, msg or "%r is in %r" % (a, b)
+
+
+# deprecated.  See #5429
+not_in_ = not_in
+
+
+def startswith_(a, fragment, msg=None):
+    """Assert a.startswith(fragment), with repr messaging on failure."""
+    assert a.startswith(fragment), msg or "%r does not start with %r" % (
+        a,
+        fragment,
+    )
+
+
+def eq_ignore_whitespace(a, b, msg=None):
+    a = re.sub(r"^\s+?|\n", "", a)
+    a = re.sub(r" {2,}", " ", a)
+    a = re.sub(r"\t", "", a)
+    b = re.sub(r"^\s+?|\n", "", b)
+    b = re.sub(r" {2,}", " ", b)
+    b = re.sub(r"\t", "", b)
+
+    assert a == b, msg or "%r != %r" % (a, b)
+
+
+def _assert_proper_exception_context(exception):
+    """assert that any exception we're catching does not have a __context__
+    without a __cause__, and that __suppress_context__ is never set.
+
+    Python 3 will report nested as exceptions as "during the handling of
+    error X, error Y occurred". That's not what we want to do.  we want
+    these exceptions in a cause chain.
+
+    """
+
+    if (
+        exception.__context__ is not exception.__cause__
+        and not exception.__suppress_context__
+    ):
+        assert False, (
+            "Exception %r was correctly raised but did not set a cause, "
+            "within context %r as its cause."
+            % (exception, exception.__context__)
+        )
+
+
+def assert_raises(except_cls, callable_, *args, **kw):
+    return _assert_raises(except_cls, callable_, args, kw, check_context=True)
+
+
+def assert_raises_context_ok(except_cls, callable_, *args, **kw):
+    return _assert_raises(except_cls, callable_, args, kw)
+
+
+def assert_raises_message(except_cls, msg, callable_, *args, **kwargs):
+    return _assert_raises(
+        except_cls, callable_, args, kwargs, msg=msg, check_context=True
+    )
+
+
+def assert_warns(except_cls, callable_, *args, **kwargs):
+    """legacy adapter function for functions that were previously using
+    assert_raises with SAWarning or similar.
+
+    has some workarounds to accommodate the fact that the callable completes
+    with this approach rather than stopping at the exception raise.
+
+
+    """
+    with _expect_warnings_sqla_only(except_cls, [".*"]):
+        return callable_(*args, **kwargs)
+
+
+def assert_warns_message(except_cls, msg, callable_, *args, **kwargs):
+    """legacy adapter function for functions that were previously using
+    assert_raises with SAWarning or similar.
+
+    has some workarounds to accommodate the fact that the callable completes
+    with this approach rather than stopping at the exception raise.
+
+    Also uses regex.search() to match the given message to the error string
+    rather than regex.match().
+
+    """
+    with _expect_warnings_sqla_only(
+        except_cls,
+        [msg],
+        search_msg=True,
+        regex=False,
+    ):
+        return callable_(*args, **kwargs)
+
+
+def assert_raises_message_context_ok(
+    except_cls, msg, callable_, *args, **kwargs
+):
+    return _assert_raises(except_cls, callable_, args, kwargs, msg=msg)
+
+
+def _assert_raises(
+    except_cls, callable_, args, kwargs, msg=None, check_context=False
+):
+    with _expect_raises(except_cls, msg, check_context) as ec:
+        callable_(*args, **kwargs)
+    return ec.error
+
+
+class _ErrorContainer:
+    error = None
+
+
+@contextlib.contextmanager
+def _expect_raises(except_cls, msg=None, check_context=False):
+    if (
+        isinstance(except_cls, type)
+        and issubclass(except_cls, Warning)
+        or isinstance(except_cls, Warning)
+    ):
+        raise TypeError(
+            "Use expect_warnings for warnings, not "
+            "expect_raises / assert_raises"
+        )
+    ec = _ErrorContainer()
+    if check_context:
+        are_we_already_in_a_traceback = sys.exc_info()[0]
+    try:
+        yield ec
+        success = False
+    except except_cls as err:
+        ec.error = err
+        success = True
+        if msg is not None:
+            # I'm often pdbing here, and "err" above isn't
+            # in scope, so assign the string explicitly
+            error_as_string = str(err)
+            assert re.search(msg, error_as_string, re.UNICODE), "%r !~ %s" % (
+                msg,
+                error_as_string,
+            )
+        if check_context and not are_we_already_in_a_traceback:
+            _assert_proper_exception_context(err)
+        print(str(err).encode("utf-8"))
+
+    # it's generally a good idea to not carry traceback objects outside
+    # of the except: block, but in this case especially we seem to have
+    # hit some bug in either python 3.10.0b2 or greenlet or both which
+    # this seems to fix:
+    # https://github.com/python-greenlet/greenlet/issues/242
+    del ec
+
+    # assert outside the block so it works for AssertionError too !
+    assert success, "Callable did not raise an exception"
+
+
+def expect_raises(except_cls, check_context=True):
+    return _expect_raises(except_cls, check_context=check_context)
+
+
+def expect_raises_message(except_cls, msg, check_context=True):
+    return _expect_raises(except_cls, msg=msg, check_context=check_context)
+
+
+class AssertsCompiledSQL:
+    def assert_compile(
+        self,
+        clause,
+        result,
+        params=None,
+        checkparams=None,
+        for_executemany=False,
+        check_literal_execute=None,
+        check_post_param=None,
+        dialect=None,
+        checkpositional=None,
+        check_prefetch=None,
+        use_default_dialect=False,
+        allow_dialect_select=False,
+        supports_default_values=True,
+        supports_default_metavalue=True,
+        literal_binds=False,
+        render_postcompile=False,
+        schema_translate_map=None,
+        render_schema_translate=False,
+        default_schema_name=None,
+        from_linting=False,
+        check_param_order=True,
+        use_literal_execute_for_simple_int=False,
+    ):
+        if use_default_dialect:
+            dialect = default.DefaultDialect()
+            dialect.supports_default_values = supports_default_values
+            dialect.supports_default_metavalue = supports_default_metavalue
+        elif allow_dialect_select:
+            dialect = None
+        else:
+            if dialect is None:
+                dialect = getattr(self, "__dialect__", None)
+
+            if dialect is None:
+                dialect = config.db.dialect
+            elif dialect == "default" or dialect == "default_qmark":
+                if dialect == "default":
+                    dialect = default.DefaultDialect()
+                else:
+                    dialect = default.DefaultDialect("qmark")
+                dialect.supports_default_values = supports_default_values
+                dialect.supports_default_metavalue = supports_default_metavalue
+            elif dialect == "default_enhanced":
+                dialect = default.StrCompileDialect()
+            elif isinstance(dialect, str):
+                dialect = url.URL.create(dialect).get_dialect()()
+
+        if default_schema_name:
+            dialect.default_schema_name = default_schema_name
+
+        kw = {}
+        compile_kwargs = {}
+
+        if schema_translate_map:
+            kw["schema_translate_map"] = schema_translate_map
+
+        if params is not None:
+            kw["column_keys"] = list(params)
+
+        if literal_binds:
+            compile_kwargs["literal_binds"] = True
+
+        if render_postcompile:
+            compile_kwargs["render_postcompile"] = True
+
+        if use_literal_execute_for_simple_int:
+            compile_kwargs["use_literal_execute_for_simple_int"] = True
+
+        if for_executemany:
+            kw["for_executemany"] = True
+
+        if render_schema_translate:
+            kw["render_schema_translate"] = True
+
+        if from_linting or getattr(self, "assert_from_linting", False):
+            kw["linting"] = sql.FROM_LINTING
+
+        from sqlalchemy import orm
+
+        if isinstance(clause, orm.Query):
+            stmt = clause._statement_20()
+            stmt._label_style = LABEL_STYLE_TABLENAME_PLUS_COL
+            clause = stmt
+
+        if compile_kwargs:
+            kw["compile_kwargs"] = compile_kwargs
+
+        class DontAccess:
+            def __getattribute__(self, key):
+                raise NotImplementedError(
+                    "compiler accessed .statement; use "
+                    "compiler.current_executable"
+                )
+
+        class CheckCompilerAccess:
+            def __init__(self, test_statement):
+                self.test_statement = test_statement
+                self._annotations = {}
+                self.supports_execution = getattr(
+                    test_statement, "supports_execution", False
+                )
+
+                if self.supports_execution:
+                    self._execution_options = test_statement._execution_options
+
+                    if hasattr(test_statement, "_returning"):
+                        self._returning = test_statement._returning
+                    if hasattr(test_statement, "_inline"):
+                        self._inline = test_statement._inline
+                    if hasattr(test_statement, "_return_defaults"):
+                        self._return_defaults = test_statement._return_defaults
+
+            @property
+            def _variant_mapping(self):
+                return self.test_statement._variant_mapping
+
+            def _default_dialect(self):
+                return self.test_statement._default_dialect()
+
+            def compile(self, dialect, **kw):
+                return self.test_statement.compile.__func__(
+                    self, dialect=dialect, **kw
+                )
+
+            def _compiler(self, dialect, **kw):
+                return self.test_statement._compiler.__func__(
+                    self, dialect, **kw
+                )
+
+            def _compiler_dispatch(self, compiler, **kwargs):
+                if hasattr(compiler, "statement"):
+                    with mock.patch.object(
+                        compiler, "statement", DontAccess()
+                    ):
+                        return self.test_statement._compiler_dispatch(
+                            compiler, **kwargs
+                        )
+                else:
+                    return self.test_statement._compiler_dispatch(
+                        compiler, **kwargs
+                    )
+
+        # no construct can assume it's the "top level" construct in all cases
+        # as anything can be nested.  ensure constructs don't assume they
+        # are the "self.statement" element
+        c = CheckCompilerAccess(clause).compile(dialect=dialect, **kw)
+
+        if isinstance(clause, sqltypes.TypeEngine):
+            cache_key_no_warnings = clause._static_cache_key
+            if cache_key_no_warnings:
+                hash(cache_key_no_warnings)
+        else:
+            cache_key_no_warnings = clause._generate_cache_key()
+            if cache_key_no_warnings:
+                hash(cache_key_no_warnings[0])
+
+        param_str = repr(getattr(c, "params", {}))
+        param_str = param_str.encode("utf-8").decode("ascii", "ignore")
+        print(("\nSQL String:\n" + str(c) + param_str).encode("utf-8"))
+
+        cc = re.sub(r"[\n\t]", "", str(c))
+
+        eq_(cc, result, "%r != %r on dialect %r" % (cc, result, dialect))
+
+        if checkparams is not None:
+            if render_postcompile:
+                expanded_state = c.construct_expanded_state(
+                    params, escape_names=False
+                )
+                eq_(expanded_state.parameters, checkparams)
+            else:
+                eq_(c.construct_params(params), checkparams)
+        if checkpositional is not None:
+            if render_postcompile:
+                expanded_state = c.construct_expanded_state(
+                    params, escape_names=False
+                )
+                eq_(
+                    tuple(
+                        [
+                            expanded_state.parameters[x]
+                            for x in expanded_state.positiontup
+                        ]
+                    ),
+                    checkpositional,
+                )
+            else:
+                p = c.construct_params(params, escape_names=False)
+                eq_(tuple([p[x] for x in c.positiontup]), checkpositional)
+        if check_prefetch is not None:
+            eq_(c.prefetch, check_prefetch)
+        if check_literal_execute is not None:
+            eq_(
+                {
+                    c.bind_names[b]: b.effective_value
+                    for b in c.literal_execute_params
+                },
+                check_literal_execute,
+            )
+        if check_post_param is not None:
+            eq_(
+                {
+                    c.bind_names[b]: b.effective_value
+                    for b in c.post_compile_params
+                },
+                check_post_param,
+            )
+        if check_param_order and getattr(c, "params", None):
+
+            def get_dialect(paramstyle, positional):
+                cp = copy(dialect)
+                cp.paramstyle = paramstyle
+                cp.positional = positional
+                return cp
+
+            pyformat_dialect = get_dialect("pyformat", False)
+            pyformat_c = clause.compile(dialect=pyformat_dialect, **kw)
+            stmt = re.sub(r"[\n\t]", "", str(pyformat_c))
+
+            qmark_dialect = get_dialect("qmark", True)
+            qmark_c = clause.compile(dialect=qmark_dialect, **kw)
+            values = list(qmark_c.positiontup)
+            escaped = qmark_c.escaped_bind_names
+
+            for post_param in (
+                qmark_c.post_compile_params | qmark_c.literal_execute_params
+            ):
+                name = qmark_c.bind_names[post_param]
+                if name in values:
+                    values = [v for v in values if v != name]
+            positions = []
+            pos_by_value = defaultdict(list)
+            for v in values:
+                try:
+                    if v in pos_by_value:
+                        start = pos_by_value[v][-1]
+                    else:
+                        start = 0
+                    esc = escaped.get(v, v)
+                    pos = stmt.index("%%(%s)s" % (esc,), start) + 2
+                    positions.append(pos)
+                    pos_by_value[v].append(pos)
+                except ValueError:
+                    msg = "Expected to find bindparam %r in %r" % (v, stmt)
+                    assert False, msg
+
+            ordered = all(
+                positions[i - 1] < positions[i]
+                for i in range(1, len(positions))
+            )
+
+            expected = [v for _, v in sorted(zip(positions, values))]
+
+            msg = (
+                "Order of parameters %s does not match the order "
+                "in the statement %s. Statement %r" % (values, expected, stmt)
+            )
+
+            is_true(ordered, msg)
+
+
+class ComparesTables:
+    def assert_tables_equal(
+        self,
+        table,
+        reflected_table,
+        strict_types=False,
+        strict_constraints=True,
+    ):
+        assert len(table.c) == len(reflected_table.c)
+        for c, reflected_c in zip(table.c, reflected_table.c):
+            eq_(c.name, reflected_c.name)
+            assert reflected_c is reflected_table.c[c.name]
+
+            if strict_constraints:
+                eq_(c.primary_key, reflected_c.primary_key)
+                eq_(c.nullable, reflected_c.nullable)
+
+            if strict_types:
+                msg = "Type '%s' doesn't correspond to type '%s'"
+                assert isinstance(reflected_c.type, type(c.type)), msg % (
+                    reflected_c.type,
+                    c.type,
+                )
+            else:
+                self.assert_types_base(reflected_c, c)
+
+            if isinstance(c.type, sqltypes.String):
+                eq_(c.type.length, reflected_c.type.length)
+
+            if strict_constraints:
+                eq_(
+                    {f.column.name for f in c.foreign_keys},
+                    {f.column.name for f in reflected_c.foreign_keys},
+                )
+            if c.server_default:
+                assert isinstance(
+                    reflected_c.server_default, schema.FetchedValue
+                )
+
+        if strict_constraints:
+            assert len(table.primary_key) == len(reflected_table.primary_key)
+            for c in table.primary_key:
+                assert reflected_table.primary_key.columns[c.name] is not None
+
+    def assert_types_base(self, c1, c2):
+        assert c1.type._compare_type_affinity(
+            c2.type
+        ), "On column %r, type '%s' doesn't correspond to type '%s'" % (
+            c1.name,
+            c1.type,
+            c2.type,
+        )
+
+
+class AssertsExecutionResults:
+    def assert_result(self, result, class_, *objects):
+        result = list(result)
+        print(repr(result))
+        self.assert_list(result, class_, objects)
+
+    def assert_list(self, result, class_, list_):
+        self.assert_(
+            len(result) == len(list_),
+            "result list is not the same size as test list, "
+            + "for class "
+            + class_.__name__,
+        )
+        for i in range(0, len(list_)):
+            self.assert_row(class_, result[i], list_[i])
+
+    def assert_row(self, class_, rowobj, desc):
+        self.assert_(
+            rowobj.__class__ is class_, "item class is not " + repr(class_)
+        )
+        for key, value in desc.items():
+            if isinstance(value, tuple):
+                if isinstance(value[1], list):
+                    self.assert_list(getattr(rowobj, key), value[0], value[1])
+                else:
+                    self.assert_row(value[0], getattr(rowobj, key), value[1])
+            else:
+                self.assert_(
+                    getattr(rowobj, key) == value,
+                    "attribute %s value %s does not match %s"
+                    % (key, getattr(rowobj, key), value),
+                )
+
+    def assert_unordered_result(self, result, cls, *expected):
+        """As assert_result, but the order of objects is not considered.
+
+        The algorithm is very expensive but not a big deal for the small
+        numbers of rows that the test suite manipulates.
+        """
+
+        class immutabledict(dict):
+            def __hash__(self):
+                return id(self)
+
+        found = util.IdentitySet(result)
+        expected = {immutabledict(e) for e in expected}
+
+        for wrong in filterfalse(lambda o: isinstance(o, cls), found):
+            fail(
+                'Unexpected type "%s", expected "%s"'
+                % (type(wrong).__name__, cls.__name__)
+            )
+
+        if len(found) != len(expected):
+            fail(
+                'Unexpected object count "%s", expected "%s"'
+                % (len(found), len(expected))
+            )
+
+        NOVALUE = object()
+
+        def _compare_item(obj, spec):
+            for key, value in spec.items():
+                if isinstance(value, tuple):
+                    try:
+                        self.assert_unordered_result(
+                            getattr(obj, key), value[0], *value[1]
+                        )
+                    except AssertionError:
+                        return False
+                else:
+                    if getattr(obj, key, NOVALUE) != value:
+                        return False
+            return True
+
+        for expected_item in expected:
+            for found_item in found:
+                if _compare_item(found_item, expected_item):
+                    found.remove(found_item)
+                    break
+            else:
+                fail(
+                    "Expected %s instance with attributes %s not found."
+                    % (cls.__name__, repr(expected_item))
+                )
+        return True
+
+    def sql_execution_asserter(self, db=None):
+        if db is None:
+            from . import db as db
+
+        return assertsql.assert_engine(db)
+
+    def assert_sql_execution(self, db, callable_, *rules):
+        with self.sql_execution_asserter(db) as asserter:
+            result = callable_()
+        asserter.assert_(*rules)
+        return result
+
+    def assert_sql(self, db, callable_, rules):
+        newrules = []
+        for rule in rules:
+            if isinstance(rule, dict):
+                newrule = assertsql.AllOf(
+                    *[assertsql.CompiledSQL(k, v) for k, v in rule.items()]
+                )
+            else:
+                newrule = assertsql.CompiledSQL(*rule)
+            newrules.append(newrule)
+
+        return self.assert_sql_execution(db, callable_, *newrules)
+
+    def assert_sql_count(self, db, callable_, count):
+        return self.assert_sql_execution(
+            db, callable_, assertsql.CountStatements(count)
+        )
+
+    @contextlib.contextmanager
+    def assert_execution(self, db, *rules):
+        with self.sql_execution_asserter(db) as asserter:
+            yield
+        asserter.assert_(*rules)
+
+    def assert_statement_count(self, db, count):
+        return self.assert_execution(db, assertsql.CountStatements(count))
+
+    @contextlib.contextmanager
+    def assert_statement_count_multi_db(self, dbs, counts):
+        recs = [
+            (self.sql_execution_asserter(db), db, count)
+            for (db, count) in zip(dbs, counts)
+        ]
+        asserters = []
+        for ctx, db, count in recs:
+            asserters.append(ctx.__enter__())
+        try:
+            yield
+        finally:
+            for asserter, (ctx, db, count) in zip(asserters, recs):
+                ctx.__exit__(None, None, None)
+                asserter.assert_(assertsql.CountStatements(count))
+
+
+class ComparesIndexes:
+    def compare_table_index_with_expected(
+        self, table: schema.Table, expected: list, dialect_name: str
+    ):
+        eq_(len(table.indexes), len(expected))
+        idx_dict = {idx.name: idx for idx in table.indexes}
+        for exp in expected:
+            idx = idx_dict[exp["name"]]
+            eq_(idx.unique, exp["unique"])
+            cols = [c for c in exp["column_names"] if c is not None]
+            eq_(len(idx.columns), len(cols))
+            for c in cols:
+                is_true(c in idx.columns)
+            exprs = exp.get("expressions")
+            if exprs:
+                eq_(len(idx.expressions), len(exprs))
+                for idx_exp, expr, col in zip(
+                    idx.expressions, exprs, exp["column_names"]
+                ):
+                    if col is None:
+                        eq_(idx_exp.text, expr)
+            if (
+                exp.get("dialect_options")
+                and f"{dialect_name}_include" in exp["dialect_options"]
+            ):
+                eq_(
+                    idx.dialect_options[dialect_name]["include"],
+                    exp["dialect_options"][f"{dialect_name}_include"],
+                )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertsql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertsql.py
new file mode 100644
index 00000000..81c7138c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/assertsql.py
@@ -0,0 +1,516 @@
+# testing/assertsql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import collections
+import contextlib
+import itertools
+import re
+
+from .. import event
+from ..engine import url
+from ..engine.default import DefaultDialect
+from ..schema import BaseDDLElement
+
+
+class AssertRule:
+    is_consumed = False
+    errormessage = None
+    consume_statement = True
+
+    def process_statement(self, execute_observed):
+        pass
+
+    def no_more_statements(self):
+        assert False, (
+            "All statements are complete, but pending "
+            "assertion rules remain"
+        )
+
+
+class SQLMatchRule(AssertRule):
+    pass
+
+
+class CursorSQL(SQLMatchRule):
+    def __init__(self, statement, params=None, consume_statement=True):
+        self.statement = statement
+        self.params = params
+        self.consume_statement = consume_statement
+
+    def process_statement(self, execute_observed):
+        stmt = execute_observed.statements[0]
+        if self.statement != stmt.statement or (
+            self.params is not None and self.params != stmt.parameters
+        ):
+            self.consume_statement = True
+            self.errormessage = (
+                "Testing for exact SQL %s parameters %s received %s %s"
+                % (
+                    self.statement,
+                    self.params,
+                    stmt.statement,
+                    stmt.parameters,
+                )
+            )
+        else:
+            execute_observed.statements.pop(0)
+            self.is_consumed = True
+            if not execute_observed.statements:
+                self.consume_statement = True
+
+
+class CompiledSQL(SQLMatchRule):
+    def __init__(
+        self, statement, params=None, dialect="default", enable_returning=True
+    ):
+        self.statement = statement
+        self.params = params
+        self.dialect = dialect
+        self.enable_returning = enable_returning
+
+    def _compare_sql(self, execute_observed, received_statement):
+        stmt = re.sub(r"[\n\t]", "", self.statement)
+        return received_statement == stmt
+
+    def _compile_dialect(self, execute_observed):
+        if self.dialect == "default":
+            dialect = DefaultDialect()
+            # this is currently what tests are expecting
+            # dialect.supports_default_values = True
+            dialect.supports_default_metavalue = True
+
+            if self.enable_returning:
+                dialect.insert_returning = dialect.update_returning = (
+                    dialect.delete_returning
+                ) = True
+                dialect.use_insertmanyvalues = True
+                dialect.supports_multivalues_insert = True
+                dialect.update_returning_multifrom = True
+                dialect.delete_returning_multifrom = True
+                # dialect.favor_returning_over_lastrowid = True
+                # dialect.insert_null_pk_still_autoincrements = True
+
+                # this is calculated but we need it to be True for this
+                # to look like all the current RETURNING dialects
+                assert dialect.insert_executemany_returning
+
+            return dialect
+        else:
+            return url.URL.create(self.dialect).get_dialect()()
+
+    def _received_statement(self, execute_observed):
+        """reconstruct the statement and params in terms
+        of a target dialect, which for CompiledSQL is just DefaultDialect."""
+
+        context = execute_observed.context
+        compare_dialect = self._compile_dialect(execute_observed)
+
+        # received_statement runs a full compile().  we should not need to
+        # consider extracted_parameters; if we do this indicates some state
+        # is being sent from a previous cached query, which some misbehaviors
+        # in the ORM can cause, see #6881
+        cache_key = None  # execute_observed.context.compiled.cache_key
+        extracted_parameters = (
+            None  # execute_observed.context.extracted_parameters
+        )
+
+        if "schema_translate_map" in context.execution_options:
+            map_ = context.execution_options["schema_translate_map"]
+        else:
+            map_ = None
+
+        if isinstance(execute_observed.clauseelement, BaseDDLElement):
+            compiled = execute_observed.clauseelement.compile(
+                dialect=compare_dialect,
+                schema_translate_map=map_,
+            )
+        else:
+            compiled = execute_observed.clauseelement.compile(
+                cache_key=cache_key,
+                dialect=compare_dialect,
+                column_keys=context.compiled.column_keys,
+                for_executemany=context.compiled.for_executemany,
+                schema_translate_map=map_,
+            )
+        _received_statement = re.sub(r"[\n\t]", "", str(compiled))
+        parameters = execute_observed.parameters
+
+        if not parameters:
+            _received_parameters = [
+                compiled.construct_params(
+                    extracted_parameters=extracted_parameters
+                )
+            ]
+        else:
+            _received_parameters = [
+                compiled.construct_params(
+                    m, extracted_parameters=extracted_parameters
+                )
+                for m in parameters
+            ]
+
+        return _received_statement, _received_parameters
+
+    def process_statement(self, execute_observed):
+        context = execute_observed.context
+
+        _received_statement, _received_parameters = self._received_statement(
+            execute_observed
+        )
+        params = self._all_params(context)
+
+        equivalent = self._compare_sql(execute_observed, _received_statement)
+
+        if equivalent:
+            if params is not None:
+                all_params = list(params)
+                all_received = list(_received_parameters)
+                while all_params and all_received:
+                    param = dict(all_params.pop(0))
+
+                    for idx, received in enumerate(list(all_received)):
+                        # do a positive compare only
+                        for param_key in param:
+                            # a key in param did not match current
+                            # 'received'
+                            if (
+                                param_key not in received
+                                or received[param_key] != param[param_key]
+                            ):
+                                break
+                        else:
+                            # all keys in param matched 'received';
+                            # onto next param
+                            del all_received[idx]
+                            break
+                    else:
+                        # param did not match any entry
+                        # in all_received
+                        equivalent = False
+                        break
+                if all_params or all_received:
+                    equivalent = False
+
+        if equivalent:
+            self.is_consumed = True
+            self.errormessage = None
+        else:
+            self.errormessage = self._failure_message(
+                execute_observed, params
+            ) % {
+                "received_statement": _received_statement,
+                "received_parameters": _received_parameters,
+            }
+
+    def _all_params(self, context):
+        if self.params:
+            if callable(self.params):
+                params = self.params(context)
+            else:
+                params = self.params
+            if not isinstance(params, list):
+                params = [params]
+            return params
+        else:
+            return None
+
+    def _failure_message(self, execute_observed, expected_params):
+        return (
+            "Testing for compiled statement\n%r partial params %s, "
+            "received\n%%(received_statement)r with params "
+            "%%(received_parameters)r"
+            % (
+                self.statement.replace("%", "%%"),
+                repr(expected_params).replace("%", "%%"),
+            )
+        )
+
+
+class RegexSQL(CompiledSQL):
+    def __init__(
+        self, regex, params=None, dialect="default", enable_returning=False
+    ):
+        SQLMatchRule.__init__(self)
+        self.regex = re.compile(regex)
+        self.orig_regex = regex
+        self.params = params
+        self.dialect = dialect
+        self.enable_returning = enable_returning
+
+    def _failure_message(self, execute_observed, expected_params):
+        return (
+            "Testing for compiled statement ~%r partial params %s, "
+            "received %%(received_statement)r with params "
+            "%%(received_parameters)r"
+            % (
+                self.orig_regex.replace("%", "%%"),
+                repr(expected_params).replace("%", "%%"),
+            )
+        )
+
+    def _compare_sql(self, execute_observed, received_statement):
+        return bool(self.regex.match(received_statement))
+
+
+class DialectSQL(CompiledSQL):
+    def _compile_dialect(self, execute_observed):
+        return execute_observed.context.dialect
+
+    def _compare_no_space(self, real_stmt, received_stmt):
+        stmt = re.sub(r"[\n\t]", "", real_stmt)
+        return received_stmt == stmt
+
+    def _received_statement(self, execute_observed):
+        received_stmt, received_params = super()._received_statement(
+            execute_observed
+        )
+
+        # TODO: why do we need this part?
+        for real_stmt in execute_observed.statements:
+            if self._compare_no_space(real_stmt.statement, received_stmt):
+                break
+        else:
+            raise AssertionError(
+                "Can't locate compiled statement %r in list of "
+                "statements actually invoked" % received_stmt
+            )
+
+        return received_stmt, execute_observed.context.compiled_parameters
+
+    def _dialect_adjusted_statement(self, dialect):
+        paramstyle = dialect.paramstyle
+        stmt = re.sub(r"[\n\t]", "", self.statement)
+
+        # temporarily escape out PG double colons
+        stmt = stmt.replace("::", "!!")
+
+        if paramstyle == "pyformat":
+            stmt = re.sub(r":([\w_]+)", r"%(\1)s", stmt)
+        else:
+            # positional params
+            repl = None
+            if paramstyle == "qmark":
+                repl = "?"
+            elif paramstyle == "format":
+                repl = r"%s"
+            elif paramstyle.startswith("numeric"):
+                counter = itertools.count(1)
+
+                num_identifier = "$" if paramstyle == "numeric_dollar" else ":"
+
+                def repl(m):
+                    return f"{num_identifier}{next(counter)}"
+
+            stmt = re.sub(r":([\w_]+)", repl, stmt)
+
+        # put them back
+        stmt = stmt.replace("!!", "::")
+
+        return stmt
+
+    def _compare_sql(self, execute_observed, received_statement):
+        stmt = self._dialect_adjusted_statement(
+            execute_observed.context.dialect
+        )
+        return received_statement == stmt
+
+    def _failure_message(self, execute_observed, expected_params):
+        return (
+            "Testing for compiled statement\n%r partial params %s, "
+            "received\n%%(received_statement)r with params "
+            "%%(received_parameters)r"
+            % (
+                self._dialect_adjusted_statement(
+                    execute_observed.context.dialect
+                ).replace("%", "%%"),
+                repr(expected_params).replace("%", "%%"),
+            )
+        )
+
+
+class CountStatements(AssertRule):
+    def __init__(self, count):
+        self.count = count
+        self._statement_count = 0
+
+    def process_statement(self, execute_observed):
+        self._statement_count += 1
+
+    def no_more_statements(self):
+        if self.count != self._statement_count:
+            assert False, "desired statement count %d does not match %d" % (
+                self.count,
+                self._statement_count,
+            )
+
+
+class AllOf(AssertRule):
+    def __init__(self, *rules):
+        self.rules = set(rules)
+
+    def process_statement(self, execute_observed):
+        for rule in list(self.rules):
+            rule.errormessage = None
+            rule.process_statement(execute_observed)
+            if rule.is_consumed:
+                self.rules.discard(rule)
+                if not self.rules:
+                    self.is_consumed = True
+                break
+            elif not rule.errormessage:
+                # rule is not done yet
+                self.errormessage = None
+                break
+        else:
+            self.errormessage = list(self.rules)[0].errormessage
+
+
+class EachOf(AssertRule):
+    def __init__(self, *rules):
+        self.rules = list(rules)
+
+    def process_statement(self, execute_observed):
+        if not self.rules:
+            self.is_consumed = True
+            self.consume_statement = False
+
+        while self.rules:
+            rule = self.rules[0]
+            rule.process_statement(execute_observed)
+            if rule.is_consumed:
+                self.rules.pop(0)
+            elif rule.errormessage:
+                self.errormessage = rule.errormessage
+            if rule.consume_statement:
+                break
+
+        if not self.rules:
+            self.is_consumed = True
+
+    def no_more_statements(self):
+        if self.rules and not self.rules[0].is_consumed:
+            self.rules[0].no_more_statements()
+        elif self.rules:
+            super().no_more_statements()
+
+
+class Conditional(EachOf):
+    def __init__(self, condition, rules, else_rules):
+        if condition:
+            super().__init__(*rules)
+        else:
+            super().__init__(*else_rules)
+
+
+class Or(AllOf):
+    def process_statement(self, execute_observed):
+        for rule in self.rules:
+            rule.process_statement(execute_observed)
+            if rule.is_consumed:
+                self.is_consumed = True
+                break
+        else:
+            self.errormessage = list(self.rules)[0].errormessage
+
+
+class SQLExecuteObserved:
+    def __init__(self, context, clauseelement, multiparams, params):
+        self.context = context
+        self.clauseelement = clauseelement
+
+        if multiparams:
+            self.parameters = multiparams
+        elif params:
+            self.parameters = [params]
+        else:
+            self.parameters = []
+        self.statements = []
+
+    def __repr__(self):
+        return str(self.statements)
+
+
+class SQLCursorExecuteObserved(
+    collections.namedtuple(
+        "SQLCursorExecuteObserved",
+        ["statement", "parameters", "context", "executemany"],
+    )
+):
+    pass
+
+
+class SQLAsserter:
+    def __init__(self):
+        self.accumulated = []
+
+    def _close(self):
+        self._final = self.accumulated
+        del self.accumulated
+
+    def assert_(self, *rules):
+        rule = EachOf(*rules)
+
+        observed = list(self._final)
+        while observed:
+            statement = observed.pop(0)
+            rule.process_statement(statement)
+            if rule.is_consumed:
+                break
+            elif rule.errormessage:
+                assert False, rule.errormessage
+        if observed:
+            assert False, "Additional SQL statements remain:\n%s" % observed
+        elif not rule.is_consumed:
+            rule.no_more_statements()
+
+
+@contextlib.contextmanager
+def assert_engine(engine):
+    asserter = SQLAsserter()
+
+    orig = []
+
+    @event.listens_for(engine, "before_execute")
+    def connection_execute(
+        conn, clauseelement, multiparams, params, execution_options
+    ):
+        # grab the original statement + params before any cursor
+        # execution
+        orig[:] = clauseelement, multiparams, params
+
+    @event.listens_for(engine, "after_cursor_execute")
+    def cursor_execute(
+        conn, cursor, statement, parameters, context, executemany
+    ):
+        if not context:
+            return
+        # then grab real cursor statements and associate them all
+        # around a single context
+        if (
+            asserter.accumulated
+            and asserter.accumulated[-1].context is context
+        ):
+            obs = asserter.accumulated[-1]
+        else:
+            obs = SQLExecuteObserved(context, orig[0], orig[1], orig[2])
+            asserter.accumulated.append(obs)
+        obs.statements.append(
+            SQLCursorExecuteObserved(
+                statement, parameters, context, executemany
+            )
+        )
+
+    try:
+        yield asserter
+    finally:
+        event.remove(engine, "after_cursor_execute", cursor_execute)
+        event.remove(engine, "before_execute", connection_execute)
+        asserter._close()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/asyncio.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/asyncio.py
new file mode 100644
index 00000000..28470ba2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/asyncio.py
@@ -0,0 +1,135 @@
+# testing/asyncio.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+# functions and wrappers to run tests, fixtures, provisioning and
+# setup/teardown in an asyncio event loop, conditionally based on the
+# current DB driver being used for a test.
+
+# note that SQLAlchemy's asyncio integration also supports a method
+# of running individual asyncio functions inside of separate event loops
+# using "async_fallback" mode; however running whole functions in the event
+# loop is a more accurate test for how SQLAlchemy's asyncio features
+# would run in the real world.
+
+
+from __future__ import annotations
+
+from functools import wraps
+import inspect
+
+from . import config
+from ..util.concurrency import _AsyncUtil
+
+# may be set to False if the
+# --disable-asyncio flag is passed to the test runner.
+ENABLE_ASYNCIO = True
+_async_util = _AsyncUtil()  # it has lazy init so just always create one
+
+
+def _shutdown():
+    """called when the test finishes"""
+    _async_util.close()
+
+
+def _run_coroutine_function(fn, *args, **kwargs):
+    return _async_util.run(fn, *args, **kwargs)
+
+
+def _assume_async(fn, *args, **kwargs):
+    """Run a function in an asyncio loop unconditionally.
+
+    This function is used for provisioning features like
+    testing a database connection for server info.
+
+    Note that for blocking IO database drivers, this means they block the
+    event loop.
+
+    """
+
+    if not ENABLE_ASYNCIO:
+        return fn(*args, **kwargs)
+
+    return _async_util.run_in_greenlet(fn, *args, **kwargs)
+
+
+def _maybe_async_provisioning(fn, *args, **kwargs):
+    """Run a function in an asyncio loop if any current drivers might need it.
+
+    This function is used for provisioning features that take
+    place outside of a specific database driver being selected, so if the
+    current driver that happens to be used for the provisioning operation
+    is an async driver, it will run in asyncio and not fail.
+
+    Note that for blocking IO database drivers, this means they block the
+    event loop.
+
+    """
+    if not ENABLE_ASYNCIO:
+        return fn(*args, **kwargs)
+
+    if config.any_async:
+        return _async_util.run_in_greenlet(fn, *args, **kwargs)
+    else:
+        return fn(*args, **kwargs)
+
+
+def _maybe_async(fn, *args, **kwargs):
+    """Run a function in an asyncio loop if the current selected driver is
+    async.
+
+    This function is used for test setup/teardown and tests themselves
+    where the current DB driver is known.
+
+
+    """
+    if not ENABLE_ASYNCIO:
+        return fn(*args, **kwargs)
+
+    is_async = config._current.is_async
+
+    if is_async:
+        return _async_util.run_in_greenlet(fn, *args, **kwargs)
+    else:
+        return fn(*args, **kwargs)
+
+
+def _maybe_async_wrapper(fn):
+    """Apply the _maybe_async function to an existing function and return
+    as a wrapped callable, supporting generator functions as well.
+
+    This is currently used for pytest fixtures that support generator use.
+
+    """
+
+    if inspect.isgeneratorfunction(fn):
+        _stop = object()
+
+        def call_next(gen):
+            try:
+                return next(gen)
+                # can't raise StopIteration in an awaitable.
+            except StopIteration:
+                return _stop
+
+        @wraps(fn)
+        def wrap_fixture(*args, **kwargs):
+            gen = fn(*args, **kwargs)
+            while True:
+                value = _maybe_async(call_next, gen)
+                if value is _stop:
+                    break
+                yield value
+
+    else:
+
+        @wraps(fn)
+        def wrap_fixture(*args, **kwargs):
+            return _maybe_async(fn, *args, **kwargs)
+
+    return wrap_fixture
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/config.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/config.py
new file mode 100644
index 00000000..2eec642b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/config.py
@@ -0,0 +1,423 @@
+# testing/config.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+from argparse import Namespace
+import collections
+import inspect
+import typing
+from typing import Any
+from typing import Callable
+from typing import Iterable
+from typing import NoReturn
+from typing import Optional
+from typing import Tuple
+from typing import TypeVar
+from typing import Union
+
+from . import mock
+from . import requirements as _requirements
+from .util import fail
+from .. import util
+
+# default requirements; this is replaced by plugin_base when pytest
+# is run
+requirements = _requirements.SuiteRequirements()
+
+db = None
+db_url = None
+db_opts = None
+file_config = None
+test_schema = None
+test_schema_2 = None
+any_async = False
+_current = None
+ident = "main"
+options: Namespace = None  # type: ignore
+
+if typing.TYPE_CHECKING:
+    from .plugin.plugin_base import FixtureFunctions
+
+    _fixture_functions: FixtureFunctions
+else:
+
+    class _NullFixtureFunctions:
+        def _null_decorator(self):
+            def go(fn):
+                return fn
+
+            return go
+
+        def skip_test_exception(self, *arg, **kw):
+            return Exception()
+
+        @property
+        def add_to_marker(self):
+            return mock.Mock()
+
+        def mark_base_test_class(self):
+            return self._null_decorator()
+
+        def combinations(self, *arg_sets, **kw):
+            return self._null_decorator()
+
+        def param_ident(self, *parameters):
+            return self._null_decorator()
+
+        def fixture(self, *arg, **kw):
+            return self._null_decorator()
+
+        def get_current_test_name(self):
+            return None
+
+        def async_test(self, fn):
+            return fn
+
+    # default fixture functions; these are replaced by plugin_base when
+    # pytest runs
+    _fixture_functions = _NullFixtureFunctions()
+
+
+_FN = TypeVar("_FN", bound=Callable[..., Any])
+
+
+def combinations(
+    *comb: Union[Any, Tuple[Any, ...]],
+    argnames: Optional[str] = None,
+    id_: Optional[str] = None,
+    **kw: str,
+) -> Callable[[_FN], _FN]:
+    r"""Deliver multiple versions of a test based on positional combinations.
+
+    This is a facade over pytest.mark.parametrize.
+
+
+    :param \*comb: argument combinations.  These are tuples that will be passed
+     positionally to the decorated function.
+
+    :param argnames: optional list of argument names.   These are the names
+     of the arguments in the test function that correspond to the entries
+     in each argument tuple.   pytest.mark.parametrize requires this, however
+     the combinations function will derive it automatically if not present
+     by using ``inspect.getfullargspec(fn).args[1:]``.  Note this assumes the
+     first argument is "self" which is discarded.
+
+    :param id\_: optional id template.  This is a string template that
+     describes how the "id" for each parameter set should be defined, if any.
+     The number of characters in the template should match the number of
+     entries in each argument tuple.   Each character describes how the
+     corresponding entry in the argument tuple should be handled, as far as
+     whether or not it is included in the arguments passed to the function, as
+     well as if it is included in the tokens used to create the id of the
+     parameter set.
+
+     If omitted, the argument combinations are passed to parametrize as is.  If
+     passed, each argument combination is turned into a pytest.param() object,
+     mapping the elements of the argument tuple to produce an id based on a
+     character value in the same position within the string template using the
+     following scheme:
+
+     .. sourcecode:: text
+
+        i - the given argument is a string that is part of the id only, don't
+            pass it as an argument
+
+        n - the given argument should be passed and it should be added to the
+            id by calling the .__name__ attribute
+
+        r - the given argument should be passed and it should be added to the
+            id by calling repr()
+
+        s - the given argument should be passed and it should be added to the
+            id by calling str()
+
+        a - (argument) the given argument should be passed and it should not
+            be used to generated the id
+
+     e.g.::
+
+        @testing.combinations(
+            (operator.eq, "eq"),
+            (operator.ne, "ne"),
+            (operator.gt, "gt"),
+            (operator.lt, "lt"),
+            id_="na",
+        )
+        def test_operator(self, opfunc, name):
+            pass
+
+    The above combination will call ``.__name__`` on the first member of
+    each tuple and use that as the "id" to pytest.param().
+
+
+    """
+    return _fixture_functions.combinations(
+        *comb, id_=id_, argnames=argnames, **kw
+    )
+
+
+def combinations_list(arg_iterable: Iterable[Tuple[Any, ...]], **kw):
+    "As combination, but takes a single iterable"
+    return combinations(*arg_iterable, **kw)
+
+
+class Variation:
+    __slots__ = ("_name", "_argname")
+
+    def __init__(self, case, argname, case_names):
+        self._name = case
+        self._argname = argname
+        for casename in case_names:
+            setattr(self, casename, casename == case)
+
+    if typing.TYPE_CHECKING:
+
+        def __getattr__(self, key: str) -> bool: ...
+
+    @property
+    def name(self):
+        return self._name
+
+    def __bool__(self):
+        return self._name == self._argname
+
+    def __nonzero__(self):
+        return not self.__bool__()
+
+    def __str__(self):
+        return f"{self._argname}={self._name!r}"
+
+    def __repr__(self):
+        return str(self)
+
+    def fail(self) -> NoReturn:
+        fail(f"Unknown {self}")
+
+    @classmethod
+    def idfn(cls, variation):
+        return variation.name
+
+    @classmethod
+    def generate_cases(cls, argname, cases):
+        case_names = [
+            argname if c is True else "not_" + argname if c is False else c
+            for c in cases
+        ]
+
+        typ = type(
+            argname,
+            (Variation,),
+            {
+                "__slots__": tuple(case_names),
+            },
+        )
+
+        return [typ(casename, argname, case_names) for casename in case_names]
+
+
+def variation(argname_or_fn, cases=None):
+    """a helper around testing.combinations that provides a single namespace
+    that can be used as a switch.
+
+    e.g.::
+
+        @testing.variation("querytyp", ["select", "subquery", "legacy_query"])
+        @testing.variation("lazy", ["select", "raise", "raise_on_sql"])
+        def test_thing(self, querytyp, lazy, decl_base):
+            class Thing(decl_base):
+                __tablename__ = "thing"
+
+                # use name directly
+                rel = relationship("Rel", lazy=lazy.name)
+
+            # use as a switch
+            if querytyp.select:
+                stmt = select(Thing)
+            elif querytyp.subquery:
+                stmt = select(Thing).subquery()
+            elif querytyp.legacy_query:
+                stmt = Session.query(Thing)
+            else:
+                querytyp.fail()
+
+    The variable provided is a slots object of boolean variables, as well
+    as the name of the case itself under the attribute ".name"
+
+    """
+
+    if inspect.isfunction(argname_or_fn):
+        argname = argname_or_fn.__name__
+        cases = argname_or_fn(None)
+
+        @variation_fixture(argname, cases)
+        def go(self, request):
+            yield request.param
+
+        return go
+    else:
+        argname = argname_or_fn
+    cases_plus_limitations = [
+        (
+            entry
+            if (isinstance(entry, tuple) and len(entry) == 2)
+            else (entry, None)
+        )
+        for entry in cases
+    ]
+
+    variations = Variation.generate_cases(
+        argname, [c for c, l in cases_plus_limitations]
+    )
+    return combinations(
+        *[
+            (
+                (variation._name, variation, limitation)
+                if limitation is not None
+                else (variation._name, variation)
+            )
+            for variation, (case, limitation) in zip(
+                variations, cases_plus_limitations
+            )
+        ],
+        id_="ia",
+        argnames=argname,
+    )
+
+
+def variation_fixture(argname, cases, scope="function"):
+    return fixture(
+        params=Variation.generate_cases(argname, cases),
+        ids=Variation.idfn,
+        scope=scope,
+    )
+
+
+def fixture(*arg: Any, **kw: Any) -> Any:
+    return _fixture_functions.fixture(*arg, **kw)
+
+
+def get_current_test_name() -> str:
+    return _fixture_functions.get_current_test_name()
+
+
+def mark_base_test_class() -> Any:
+    return _fixture_functions.mark_base_test_class()
+
+
+class _AddToMarker:
+    def __getattr__(self, attr: str) -> Any:
+        return getattr(_fixture_functions.add_to_marker, attr)
+
+
+add_to_marker = _AddToMarker()
+
+
+class Config:
+    def __init__(self, db, db_opts, options, file_config):
+        self._set_name(db)
+        self.db = db
+        self.db_opts = db_opts
+        self.options = options
+        self.file_config = file_config
+        self.test_schema = "test_schema"
+        self.test_schema_2 = "test_schema_2"
+
+        self.is_async = db.dialect.is_async and not util.asbool(
+            db.url.query.get("async_fallback", False)
+        )
+
+    _stack = collections.deque()
+    _configs = set()
+
+    def _set_name(self, db):
+        suffix = "_async" if db.dialect.is_async else ""
+        if db.dialect.server_version_info:
+            svi = ".".join(str(tok) for tok in db.dialect.server_version_info)
+            self.name = "%s+%s%s_[%s]" % (db.name, db.driver, suffix, svi)
+        else:
+            self.name = "%s+%s%s" % (db.name, db.driver, suffix)
+
+    @classmethod
+    def register(cls, db, db_opts, options, file_config):
+        """add a config as one of the global configs.
+
+        If there are no configs set up yet, this config also
+        gets set as the "_current".
+        """
+        global any_async
+
+        cfg = Config(db, db_opts, options, file_config)
+
+        # if any backends include an async driver, then ensure
+        # all setup/teardown and tests are wrapped in the maybe_async()
+        # decorator that will set up a greenlet context for async drivers.
+        any_async = any_async or cfg.is_async
+
+        cls._configs.add(cfg)
+        return cfg
+
+    @classmethod
+    def set_as_current(cls, config, namespace):
+        global db, _current, db_url, test_schema, test_schema_2, db_opts
+        _current = config
+        db_url = config.db.url
+        db_opts = config.db_opts
+        test_schema = config.test_schema
+        test_schema_2 = config.test_schema_2
+        namespace.db = db = config.db
+
+    @classmethod
+    def push_engine(cls, db, namespace):
+        assert _current, "Can't push without a default Config set up"
+        cls.push(
+            Config(
+                db, _current.db_opts, _current.options, _current.file_config
+            ),
+            namespace,
+        )
+
+    @classmethod
+    def push(cls, config, namespace):
+        cls._stack.append(_current)
+        cls.set_as_current(config, namespace)
+
+    @classmethod
+    def pop(cls, namespace):
+        if cls._stack:
+            # a failed test w/ -x option can call reset() ahead of time
+            _current = cls._stack[-1]
+            del cls._stack[-1]
+            cls.set_as_current(_current, namespace)
+
+    @classmethod
+    def reset(cls, namespace):
+        if cls._stack:
+            cls.set_as_current(cls._stack[0], namespace)
+            cls._stack.clear()
+
+    @classmethod
+    def all_configs(cls):
+        return cls._configs
+
+    @classmethod
+    def all_dbs(cls):
+        for cfg in cls.all_configs():
+            yield cfg.db
+
+    def skip_test(self, msg):
+        skip_test(msg)
+
+
+def skip_test(msg):
+    raise _fixture_functions.skip_test_exception(msg)
+
+
+def async_test(fn):
+    return _fixture_functions.async_test(fn)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/engines.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/engines.py
new file mode 100644
index 00000000..51beed98
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/engines.py
@@ -0,0 +1,474 @@
+# testing/engines.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import collections
+import re
+import typing
+from typing import Any
+from typing import Dict
+from typing import Optional
+import warnings
+import weakref
+
+from . import config
+from .util import decorator
+from .util import gc_collect
+from .. import event
+from .. import pool
+from ..util import await_only
+from ..util.typing import Literal
+
+
+if typing.TYPE_CHECKING:
+    from ..engine import Engine
+    from ..engine.url import URL
+    from ..ext.asyncio import AsyncEngine
+
+
+class ConnectionKiller:
+    def __init__(self):
+        self.proxy_refs = weakref.WeakKeyDictionary()
+        self.testing_engines = collections.defaultdict(set)
+        self.dbapi_connections = set()
+
+    def add_pool(self, pool):
+        event.listen(pool, "checkout", self._add_conn)
+        event.listen(pool, "checkin", self._remove_conn)
+        event.listen(pool, "close", self._remove_conn)
+        event.listen(pool, "close_detached", self._remove_conn)
+        # note we are keeping "invalidated" here, as those are still
+        # opened connections we would like to roll back
+
+    def _add_conn(self, dbapi_con, con_record, con_proxy):
+        self.dbapi_connections.add(dbapi_con)
+        self.proxy_refs[con_proxy] = True
+
+    def _remove_conn(self, dbapi_conn, *arg):
+        self.dbapi_connections.discard(dbapi_conn)
+
+    def add_engine(self, engine, scope):
+        self.add_pool(engine.pool)
+
+        assert scope in ("class", "global", "function", "fixture")
+        self.testing_engines[scope].add(engine)
+
+    def _safe(self, fn):
+        try:
+            fn()
+        except Exception as e:
+            warnings.warn(
+                "testing_reaper couldn't rollback/close connection: %s" % e
+            )
+
+    def rollback_all(self):
+        for rec in list(self.proxy_refs):
+            if rec is not None and rec.is_valid:
+                self._safe(rec.rollback)
+
+    def checkin_all(self):
+        # run pool.checkin() for all ConnectionFairy instances we have
+        # tracked.
+
+        for rec in list(self.proxy_refs):
+            if rec is not None and rec.is_valid:
+                self.dbapi_connections.discard(rec.dbapi_connection)
+                self._safe(rec._checkin)
+
+        # for fairy refs that were GCed and could not close the connection,
+        # such as asyncio, roll back those remaining connections
+        for con in self.dbapi_connections:
+            self._safe(con.rollback)
+        self.dbapi_connections.clear()
+
+    def close_all(self):
+        self.checkin_all()
+
+    def prepare_for_drop_tables(self, connection):
+        # don't do aggressive checks for third party test suites
+        if not config.bootstrapped_as_sqlalchemy:
+            return
+
+        from . import provision
+
+        provision.prepare_for_drop_tables(connection.engine.url, connection)
+
+    def _drop_testing_engines(self, scope):
+        eng = self.testing_engines[scope]
+        for rec in list(eng):
+            for proxy_ref in list(self.proxy_refs):
+                if proxy_ref is not None and proxy_ref.is_valid:
+                    if (
+                        proxy_ref._pool is not None
+                        and proxy_ref._pool is rec.pool
+                    ):
+                        self._safe(proxy_ref._checkin)
+
+            if hasattr(rec, "sync_engine"):
+                await_only(rec.dispose())
+            else:
+                rec.dispose()
+        eng.clear()
+
+    def after_test(self):
+        self._drop_testing_engines("function")
+
+    def after_test_outside_fixtures(self, test):
+        # don't do aggressive checks for third party test suites
+        if not config.bootstrapped_as_sqlalchemy:
+            return
+
+        if test.__class__.__leave_connections_for_teardown__:
+            return
+
+        self.checkin_all()
+
+        # on PostgreSQL, this will test for any "idle in transaction"
+        # connections.   useful to identify tests with unusual patterns
+        # that can't be cleaned up correctly.
+        from . import provision
+
+        with config.db.connect() as conn:
+            provision.prepare_for_drop_tables(conn.engine.url, conn)
+
+    def stop_test_class_inside_fixtures(self):
+        self.checkin_all()
+        self._drop_testing_engines("function")
+        self._drop_testing_engines("class")
+
+    def stop_test_class_outside_fixtures(self):
+        # ensure no refs to checked out connections at all.
+
+        if pool.base._strong_ref_connection_records:
+            gc_collect()
+
+            if pool.base._strong_ref_connection_records:
+                ln = len(pool.base._strong_ref_connection_records)
+                pool.base._strong_ref_connection_records.clear()
+                assert (
+                    False
+                ), "%d connection recs not cleared after test suite" % (ln)
+
+    def final_cleanup(self):
+        self.checkin_all()
+        for scope in self.testing_engines:
+            self._drop_testing_engines(scope)
+
+    def assert_all_closed(self):
+        for rec in self.proxy_refs:
+            if rec.is_valid:
+                assert False
+
+
+testing_reaper = ConnectionKiller()
+
+
+@decorator
+def assert_conns_closed(fn, *args, **kw):
+    try:
+        fn(*args, **kw)
+    finally:
+        testing_reaper.assert_all_closed()
+
+
+@decorator
+def rollback_open_connections(fn, *args, **kw):
+    """Decorator that rolls back all open connections after fn execution."""
+
+    try:
+        fn(*args, **kw)
+    finally:
+        testing_reaper.rollback_all()
+
+
+@decorator
+def close_first(fn, *args, **kw):
+    """Decorator that closes all connections before fn execution."""
+
+    testing_reaper.checkin_all()
+    fn(*args, **kw)
+
+
+@decorator
+def close_open_connections(fn, *args, **kw):
+    """Decorator that closes all connections after fn execution."""
+    try:
+        fn(*args, **kw)
+    finally:
+        testing_reaper.checkin_all()
+
+
+def all_dialects(exclude=None):
+    import sqlalchemy.dialects as d
+
+    for name in d.__all__:
+        # TEMPORARY
+        if exclude and name in exclude:
+            continue
+        mod = getattr(d, name, None)
+        if not mod:
+            mod = getattr(
+                __import__("sqlalchemy.dialects.%s" % name).dialects, name
+            )
+        yield mod.dialect()
+
+
+class ReconnectFixture:
+    def __init__(self, dbapi):
+        self.dbapi = dbapi
+        self.connections = []
+        self.is_stopped = False
+
+    def __getattr__(self, key):
+        return getattr(self.dbapi, key)
+
+    def connect(self, *args, **kwargs):
+        conn = self.dbapi.connect(*args, **kwargs)
+        if self.is_stopped:
+            self._safe(conn.close)
+            curs = conn.cursor()  # should fail on Oracle etc.
+            # should fail for everything that didn't fail
+            # above, connection is closed
+            curs.execute("select 1")
+            assert False, "simulated connect failure didn't work"
+        else:
+            self.connections.append(conn)
+            return conn
+
+    def _safe(self, fn):
+        try:
+            fn()
+        except Exception as e:
+            warnings.warn("ReconnectFixture couldn't close connection: %s" % e)
+
+    def shutdown(self, stop=False):
+        # TODO: this doesn't cover all cases
+        # as nicely as we'd like, namely MySQLdb.
+        # would need to implement R. Brewer's
+        # proxy server idea to get better
+        # coverage.
+        self.is_stopped = stop
+        for c in list(self.connections):
+            self._safe(c.close)
+        self.connections = []
+
+    def restart(self):
+        self.is_stopped = False
+
+
+def reconnecting_engine(url=None, options=None):
+    url = url or config.db.url
+    dbapi = config.db.dialect.dbapi
+    if not options:
+        options = {}
+    options["module"] = ReconnectFixture(dbapi)
+    engine = testing_engine(url, options)
+    _dispose = engine.dispose
+
+    def dispose():
+        engine.dialect.dbapi.shutdown()
+        engine.dialect.dbapi.is_stopped = False
+        _dispose()
+
+    engine.test_shutdown = engine.dialect.dbapi.shutdown
+    engine.test_restart = engine.dialect.dbapi.restart
+    engine.dispose = dispose
+    return engine
+
+
+@typing.overload
+def testing_engine(
+    url: Optional[URL] = None,
+    options: Optional[Dict[str, Any]] = None,
+    asyncio: Literal[False] = False,
+    transfer_staticpool: bool = False,
+) -> Engine: ...
+
+
+@typing.overload
+def testing_engine(
+    url: Optional[URL] = None,
+    options: Optional[Dict[str, Any]] = None,
+    asyncio: Literal[True] = True,
+    transfer_staticpool: bool = False,
+) -> AsyncEngine: ...
+
+
+def testing_engine(
+    url=None,
+    options=None,
+    asyncio=False,
+    transfer_staticpool=False,
+    share_pool=False,
+    _sqlite_savepoint=False,
+):
+    if asyncio:
+        assert not _sqlite_savepoint
+        from sqlalchemy.ext.asyncio import (
+            create_async_engine as create_engine,
+        )
+    else:
+        from sqlalchemy import create_engine
+    from sqlalchemy.engine.url import make_url
+
+    if not options:
+        use_reaper = True
+        scope = "function"
+        sqlite_savepoint = False
+    else:
+        use_reaper = options.pop("use_reaper", True)
+        scope = options.pop("scope", "function")
+        sqlite_savepoint = options.pop("sqlite_savepoint", False)
+
+    url = url or config.db.url
+
+    url = make_url(url)
+
+    if (
+        config.db is None or url.drivername == config.db.url.drivername
+    ) and config.db_opts:
+        use_options = config.db_opts.copy()
+    else:
+        use_options = {}
+
+    if options is not None:
+        use_options.update(options)
+
+    engine = create_engine(url, **use_options)
+
+    if sqlite_savepoint and engine.name == "sqlite":
+        # apply SQLite savepoint workaround
+        @event.listens_for(engine, "connect")
+        def do_connect(dbapi_connection, connection_record):
+            dbapi_connection.isolation_level = None
+
+        @event.listens_for(engine, "begin")
+        def do_begin(conn):
+            conn.exec_driver_sql("BEGIN")
+
+    if transfer_staticpool:
+        from sqlalchemy.pool import StaticPool
+
+        if config.db is not None and isinstance(config.db.pool, StaticPool):
+            use_reaper = False
+            engine.pool._transfer_from(config.db.pool)
+    elif share_pool:
+        engine.pool = config.db.pool
+
+    if scope == "global":
+        if asyncio:
+            engine.sync_engine._has_events = True
+        else:
+            engine._has_events = (
+                True  # enable event blocks, helps with profiling
+            )
+
+    if (
+        isinstance(engine.pool, pool.QueuePool)
+        and "pool" not in use_options
+        and "pool_timeout" not in use_options
+        and "max_overflow" not in use_options
+    ):
+        engine.pool._timeout = 0
+        engine.pool._max_overflow = 0
+    if use_reaper:
+        testing_reaper.add_engine(engine, scope)
+
+    return engine
+
+
+def mock_engine(dialect_name=None):
+    """Provides a mocking engine based on the current testing.db.
+
+    This is normally used to test DDL generation flow as emitted
+    by an Engine.
+
+    It should not be used in other cases, as assert_compile() and
+    assert_sql_execution() are much better choices with fewer
+    moving parts.
+
+    """
+
+    from sqlalchemy import create_mock_engine
+
+    if not dialect_name:
+        dialect_name = config.db.name
+
+    buffer = []
+
+    def executor(sql, *a, **kw):
+        buffer.append(sql)
+
+    def assert_sql(stmts):
+        recv = [re.sub(r"[\n\t]", "", str(s)) for s in buffer]
+        assert recv == stmts, recv
+
+    def print_sql():
+        d = engine.dialect
+        return "\n".join(str(s.compile(dialect=d)) for s in engine.mock)
+
+    engine = create_mock_engine(dialect_name + "://", executor)
+    assert not hasattr(engine, "mock")
+    engine.mock = buffer
+    engine.assert_sql = assert_sql
+    engine.print_sql = print_sql
+    return engine
+
+
+class DBAPIProxyCursor:
+    """Proxy a DBAPI cursor.
+
+    Tests can provide subclasses of this to intercept
+    DBAPI-level cursor operations.
+
+    """
+
+    def __init__(self, engine, conn, *args, **kwargs):
+        self.engine = engine
+        self.connection = conn
+        self.cursor = conn.cursor(*args, **kwargs)
+
+    def execute(self, stmt, parameters=None, **kw):
+        if parameters:
+            return self.cursor.execute(stmt, parameters, **kw)
+        else:
+            return self.cursor.execute(stmt, **kw)
+
+    def executemany(self, stmt, params, **kw):
+        return self.cursor.executemany(stmt, params, **kw)
+
+    def __iter__(self):
+        return iter(self.cursor)
+
+    def __getattr__(self, key):
+        return getattr(self.cursor, key)
+
+
+class DBAPIProxyConnection:
+    """Proxy a DBAPI connection.
+
+    Tests can provide subclasses of this to intercept
+    DBAPI-level connection operations.
+
+    """
+
+    def __init__(self, engine, conn, cursor_cls):
+        self.conn = conn
+        self.engine = engine
+        self.cursor_cls = cursor_cls
+
+    def cursor(self, *args, **kwargs):
+        return self.cursor_cls(self.engine, self.conn, *args, **kwargs)
+
+    def close(self):
+        self.conn.close()
+
+    def __getattr__(self, key):
+        return getattr(self.conn, key)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/entities.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/entities.py
new file mode 100644
index 00000000..5bd4f7de
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/entities.py
@@ -0,0 +1,117 @@
+# testing/entities.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import sqlalchemy as sa
+from .. import exc as sa_exc
+from ..orm.writeonly import WriteOnlyCollection
+
+_repr_stack = set()
+
+
+class BasicEntity:
+    def __init__(self, **kw):
+        for key, value in kw.items():
+            setattr(self, key, value)
+
+    def __repr__(self):
+        if id(self) in _repr_stack:
+            return object.__repr__(self)
+        _repr_stack.add(id(self))
+        try:
+            return "%s(%s)" % (
+                (self.__class__.__name__),
+                ", ".join(
+                    [
+                        "%s=%r" % (key, getattr(self, key))
+                        for key in sorted(self.__dict__.keys())
+                        if not key.startswith("_")
+                    ]
+                ),
+            )
+        finally:
+            _repr_stack.remove(id(self))
+
+
+_recursion_stack = set()
+
+
+class ComparableMixin:
+    def __ne__(self, other):
+        return not self.__eq__(other)
+
+    def __eq__(self, other):
+        """'Deep, sparse compare.
+
+        Deeply compare two entities, following the non-None attributes of the
+        non-persisted object, if possible.
+
+        """
+        if other is self:
+            return True
+        elif not self.__class__ == other.__class__:
+            return False
+
+        if id(self) in _recursion_stack:
+            return True
+        _recursion_stack.add(id(self))
+
+        try:
+            # pick the entity that's not SA persisted as the source
+            try:
+                self_key = sa.orm.attributes.instance_state(self).key
+            except sa.orm.exc.NO_STATE:
+                self_key = None
+
+            if other is None:
+                a = self
+                b = other
+            elif self_key is not None:
+                a = other
+                b = self
+            else:
+                a = self
+                b = other
+
+            for attr in list(a.__dict__):
+                if attr.startswith("_"):
+                    continue
+
+                value = getattr(a, attr)
+
+                if isinstance(value, WriteOnlyCollection):
+                    continue
+
+                try:
+                    # handle lazy loader errors
+                    battr = getattr(b, attr)
+                except (AttributeError, sa_exc.UnboundExecutionError):
+                    return False
+
+                if hasattr(value, "__iter__") and not isinstance(value, str):
+                    if hasattr(value, "__getitem__") and not hasattr(
+                        value, "keys"
+                    ):
+                        if list(value) != list(battr):
+                            return False
+                    else:
+                        if set(value) != set(battr):
+                            return False
+                else:
+                    if value is not None and value != battr:
+                        return False
+            return True
+        finally:
+            _recursion_stack.remove(id(self))
+
+
+class ComparableEntity(ComparableMixin, BasicEntity):
+    def __hash__(self):
+        return hash(self.__class__)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/exclusions.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/exclusions.py
new file mode 100644
index 00000000..8ff9b644
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/exclusions.py
@@ -0,0 +1,435 @@
+# testing/exclusions.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import contextlib
+import operator
+import re
+import sys
+
+from . import config
+from .. import util
+from ..util import decorator
+from ..util.compat import inspect_getfullargspec
+
+
+def skip_if(predicate, reason=None):
+    rule = compound()
+    pred = _as_predicate(predicate, reason)
+    rule.skips.add(pred)
+    return rule
+
+
+def fails_if(predicate, reason=None):
+    rule = compound()
+    pred = _as_predicate(predicate, reason)
+    rule.fails.add(pred)
+    return rule
+
+
+class compound:
+    def __init__(self):
+        self.fails = set()
+        self.skips = set()
+
+    def __add__(self, other):
+        return self.add(other)
+
+    def as_skips(self):
+        rule = compound()
+        rule.skips.update(self.skips)
+        rule.skips.update(self.fails)
+        return rule
+
+    def add(self, *others):
+        copy = compound()
+        copy.fails.update(self.fails)
+        copy.skips.update(self.skips)
+
+        for other in others:
+            copy.fails.update(other.fails)
+            copy.skips.update(other.skips)
+        return copy
+
+    def not_(self):
+        copy = compound()
+        copy.fails.update(NotPredicate(fail) for fail in self.fails)
+        copy.skips.update(NotPredicate(skip) for skip in self.skips)
+        return copy
+
+    @property
+    def enabled(self):
+        return self.enabled_for_config(config._current)
+
+    def enabled_for_config(self, config):
+        for predicate in self.skips.union(self.fails):
+            if predicate(config):
+                return False
+        else:
+            return True
+
+    def matching_config_reasons(self, config):
+        return [
+            predicate._as_string(config)
+            for predicate in self.skips.union(self.fails)
+            if predicate(config)
+        ]
+
+    def _extend(self, other):
+        self.skips.update(other.skips)
+        self.fails.update(other.fails)
+
+    def __call__(self, fn):
+        if hasattr(fn, "_sa_exclusion_extend"):
+            fn._sa_exclusion_extend._extend(self)
+            return fn
+
+        @decorator
+        def decorate(fn, *args, **kw):
+            return self._do(config._current, fn, *args, **kw)
+
+        decorated = decorate(fn)
+        decorated._sa_exclusion_extend = self
+        return decorated
+
+    @contextlib.contextmanager
+    def fail_if(self):
+        all_fails = compound()
+        all_fails.fails.update(self.skips.union(self.fails))
+
+        try:
+            yield
+        except Exception as ex:
+            all_fails._expect_failure(config._current, ex)
+        else:
+            all_fails._expect_success(config._current)
+
+    def _do(self, cfg, fn, *args, **kw):
+        for skip in self.skips:
+            if skip(cfg):
+                msg = "'%s' : %s" % (
+                    config.get_current_test_name(),
+                    skip._as_string(cfg),
+                )
+                config.skip_test(msg)
+
+        try:
+            return_value = fn(*args, **kw)
+        except Exception as ex:
+            self._expect_failure(cfg, ex, name=fn.__name__)
+        else:
+            self._expect_success(cfg, name=fn.__name__)
+            return return_value
+
+    def _expect_failure(self, config, ex, name="block"):
+        for fail in self.fails:
+            if fail(config):
+                print(
+                    "%s failed as expected (%s): %s "
+                    % (name, fail._as_string(config), ex)
+                )
+                break
+        else:
+            raise ex.with_traceback(sys.exc_info()[2])
+
+    def _expect_success(self, config, name="block"):
+        if not self.fails:
+            return
+
+        for fail in self.fails:
+            if fail(config):
+                raise AssertionError(
+                    "Unexpected success for '%s' (%s)"
+                    % (
+                        name,
+                        " and ".join(
+                            fail._as_string(config) for fail in self.fails
+                        ),
+                    )
+                )
+
+
+def only_if(predicate, reason=None):
+    predicate = _as_predicate(predicate)
+    return skip_if(NotPredicate(predicate), reason)
+
+
+def succeeds_if(predicate, reason=None):
+    predicate = _as_predicate(predicate)
+    return fails_if(NotPredicate(predicate), reason)
+
+
+class Predicate:
+    @classmethod
+    def as_predicate(cls, predicate, description=None):
+        if isinstance(predicate, compound):
+            return cls.as_predicate(predicate.enabled_for_config, description)
+        elif isinstance(predicate, Predicate):
+            if description and predicate.description is None:
+                predicate.description = description
+            return predicate
+        elif isinstance(predicate, (list, set)):
+            return OrPredicate(
+                [cls.as_predicate(pred) for pred in predicate], description
+            )
+        elif isinstance(predicate, tuple):
+            return SpecPredicate(*predicate)
+        elif isinstance(predicate, str):
+            tokens = re.match(
+                r"([\+\w]+)\s*(?:(>=|==|!=|<=|<|>)\s*([\d\.]+))?", predicate
+            )
+            if not tokens:
+                raise ValueError(
+                    "Couldn't locate DB name in predicate: %r" % predicate
+                )
+            db = tokens.group(1)
+            op = tokens.group(2)
+            spec = (
+                tuple(int(d) for d in tokens.group(3).split("."))
+                if tokens.group(3)
+                else None
+            )
+
+            return SpecPredicate(db, op, spec, description=description)
+        elif callable(predicate):
+            return LambdaPredicate(predicate, description)
+        else:
+            assert False, "unknown predicate type: %s" % predicate
+
+    def _format_description(self, config, negate=False):
+        bool_ = self(config)
+        if negate:
+            bool_ = not negate
+        return self.description % {
+            "driver": (
+                config.db.url.get_driver_name() if config else "<no driver>"
+            ),
+            "database": (
+                config.db.url.get_backend_name() if config else "<no database>"
+            ),
+            "doesnt_support": "doesn't support" if bool_ else "does support",
+            "does_support": "does support" if bool_ else "doesn't support",
+        }
+
+    def _as_string(self, config=None, negate=False):
+        raise NotImplementedError()
+
+
+class BooleanPredicate(Predicate):
+    def __init__(self, value, description=None):
+        self.value = value
+        self.description = description or "boolean %s" % value
+
+    def __call__(self, config):
+        return self.value
+
+    def _as_string(self, config, negate=False):
+        return self._format_description(config, negate=negate)
+
+
+class SpecPredicate(Predicate):
+    def __init__(self, db, op=None, spec=None, description=None):
+        self.db = db
+        self.op = op
+        self.spec = spec
+        self.description = description
+
+    _ops = {
+        "<": operator.lt,
+        ">": operator.gt,
+        "==": operator.eq,
+        "!=": operator.ne,
+        "<=": operator.le,
+        ">=": operator.ge,
+        "in": operator.contains,
+        "between": lambda val, pair: val >= pair[0] and val <= pair[1],
+    }
+
+    def __call__(self, config):
+        if config is None:
+            return False
+
+        engine = config.db
+
+        if "+" in self.db:
+            dialect, driver = self.db.split("+")
+        else:
+            dialect, driver = self.db, None
+
+        if dialect and engine.name != dialect:
+            return False
+        if driver is not None and engine.driver != driver:
+            return False
+
+        if self.op is not None:
+            assert driver is None, "DBAPI version specs not supported yet"
+
+            version = _server_version(engine)
+            oper = (
+                hasattr(self.op, "__call__") and self.op or self._ops[self.op]
+            )
+            return oper(version, self.spec)
+        else:
+            return True
+
+    def _as_string(self, config, negate=False):
+        if self.description is not None:
+            return self._format_description(config)
+        elif self.op is None:
+            if negate:
+                return "not %s" % self.db
+            else:
+                return "%s" % self.db
+        else:
+            if negate:
+                return "not %s %s %s" % (self.db, self.op, self.spec)
+            else:
+                return "%s %s %s" % (self.db, self.op, self.spec)
+
+
+class LambdaPredicate(Predicate):
+    def __init__(self, lambda_, description=None, args=None, kw=None):
+        spec = inspect_getfullargspec(lambda_)
+        if not spec[0]:
+            self.lambda_ = lambda db: lambda_()
+        else:
+            self.lambda_ = lambda_
+        self.args = args or ()
+        self.kw = kw or {}
+        if description:
+            self.description = description
+        elif lambda_.__doc__:
+            self.description = lambda_.__doc__
+        else:
+            self.description = "custom function"
+
+    def __call__(self, config):
+        return self.lambda_(config)
+
+    def _as_string(self, config, negate=False):
+        return self._format_description(config)
+
+
+class NotPredicate(Predicate):
+    def __init__(self, predicate, description=None):
+        self.predicate = predicate
+        self.description = description
+
+    def __call__(self, config):
+        return not self.predicate(config)
+
+    def _as_string(self, config, negate=False):
+        if self.description:
+            return self._format_description(config, not negate)
+        else:
+            return self.predicate._as_string(config, not negate)
+
+
+class OrPredicate(Predicate):
+    def __init__(self, predicates, description=None):
+        self.predicates = predicates
+        self.description = description
+
+    def __call__(self, config):
+        for pred in self.predicates:
+            if pred(config):
+                return True
+        return False
+
+    def _eval_str(self, config, negate=False):
+        if negate:
+            conjunction = " and "
+        else:
+            conjunction = " or "
+        return conjunction.join(
+            p._as_string(config, negate=negate) for p in self.predicates
+        )
+
+    def _negation_str(self, config):
+        if self.description is not None:
+            return "Not " + self._format_description(config)
+        else:
+            return self._eval_str(config, negate=True)
+
+    def _as_string(self, config, negate=False):
+        if negate:
+            return self._negation_str(config)
+        else:
+            if self.description is not None:
+                return self._format_description(config)
+            else:
+                return self._eval_str(config)
+
+
+_as_predicate = Predicate.as_predicate
+
+
+def _is_excluded(db, op, spec):
+    return SpecPredicate(db, op, spec)(config._current)
+
+
+def _server_version(engine):
+    """Return a server_version_info tuple."""
+
+    # force metadata to be retrieved
+    conn = engine.connect()
+    version = getattr(engine.dialect, "server_version_info", None)
+    if version is None:
+        version = ()
+    conn.close()
+    return version
+
+
+def db_spec(*dbs):
+    return OrPredicate([Predicate.as_predicate(db) for db in dbs])
+
+
+def open():  # noqa
+    return skip_if(BooleanPredicate(False, "mark as execute"))
+
+
+def closed():
+    return skip_if(BooleanPredicate(True, "marked as skip"))
+
+
+def fails(reason=None):
+    return fails_if(BooleanPredicate(True, reason or "expected to fail"))
+
+
+def future():
+    return fails_if(BooleanPredicate(True, "Future feature"))
+
+
+def fails_on(db, reason=None):
+    return fails_if(db, reason)
+
+
+def fails_on_everything_except(*dbs):
+    return succeeds_if(OrPredicate([Predicate.as_predicate(db) for db in dbs]))
+
+
+def skip(db, reason=None):
+    return skip_if(db, reason)
+
+
+def only_on(dbs, reason=None):
+    return only_if(
+        OrPredicate(
+            [Predicate.as_predicate(db, reason) for db in util.to_list(dbs)]
+        )
+    )
+
+
+def exclude(db, op, spec, reason=None):
+    return skip_if(SpecPredicate(db, op, spec), reason)
+
+
+def against(config, *queries):
+    assert queries, "no queries sent!"
+    return OrPredicate([Predicate.as_predicate(query) for query in queries])(
+        config
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__init__.py
new file mode 100644
index 00000000..f2948dee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__init__.py
@@ -0,0 +1,28 @@
+# testing/fixtures/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from .base import FutureEngineMixin as FutureEngineMixin
+from .base import TestBase as TestBase
+from .mypy import MypyTest as MypyTest
+from .orm import after_test as after_test
+from .orm import close_all_sessions as close_all_sessions
+from .orm import DeclarativeMappedTest as DeclarativeMappedTest
+from .orm import fixture_session as fixture_session
+from .orm import MappedTest as MappedTest
+from .orm import ORMTest as ORMTest
+from .orm import RemoveORMEventsGlobally as RemoveORMEventsGlobally
+from .orm import (
+    stop_test_class_inside_fixtures as stop_test_class_inside_fixtures,
+)
+from .sql import CacheKeyFixture as CacheKeyFixture
+from .sql import (
+    ComputedReflectionFixtureTest as ComputedReflectionFixtureTest,
+)
+from .sql import insertmanyvalues_fixture as insertmanyvalues_fixture
+from .sql import NoCache as NoCache
+from .sql import RemovesEvents as RemovesEvents
+from .sql import TablesTest as TablesTest
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..4c2edc8b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..3b25780b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/mypy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/mypy.cpython-312.pyc
new file mode 100644
index 00000000..e4c3cdf5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/mypy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/orm.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/orm.cpython-312.pyc
new file mode 100644
index 00000000..ec8a14e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/orm.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/sql.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/sql.cpython-312.pyc
new file mode 100644
index 00000000..21e2dce1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/__pycache__/sql.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/base.py
new file mode 100644
index 00000000..09d45a0a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/base.py
@@ -0,0 +1,366 @@
+# testing/fixtures/base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import sqlalchemy as sa
+from .. import assertions
+from .. import config
+from ..assertions import eq_
+from ..util import drop_all_tables_from_metadata
+from ... import Column
+from ... import func
+from ... import Integer
+from ... import select
+from ... import Table
+from ...orm import DeclarativeBase
+from ...orm import MappedAsDataclass
+from ...orm import registry
+
+
+@config.mark_base_test_class()
+class TestBase:
+    # A sequence of requirement names matching testing.requires decorators
+    __requires__ = ()
+
+    # A sequence of dialect names to exclude from the test class.
+    __unsupported_on__ = ()
+
+    # If present, test class is only runnable for the *single* specified
+    # dialect.  If you need multiple, use __unsupported_on__ and invert.
+    __only_on__ = None
+
+    # A sequence of no-arg callables. If any are True, the entire testcase is
+    # skipped.
+    __skip_if__ = None
+
+    # if True, the testing reaper will not attempt to touch connection
+    # state after a test is completed and before the outer teardown
+    # starts
+    __leave_connections_for_teardown__ = False
+
+    def assert_(self, val, msg=None):
+        assert val, msg
+
+    @config.fixture()
+    def nocache(self):
+        _cache = config.db._compiled_cache
+        config.db._compiled_cache = None
+        yield
+        config.db._compiled_cache = _cache
+
+    @config.fixture()
+    def connection_no_trans(self):
+        eng = getattr(self, "bind", None) or config.db
+
+        with eng.connect() as conn:
+            yield conn
+
+    @config.fixture()
+    def connection(self):
+        global _connection_fixture_connection
+
+        eng = getattr(self, "bind", None) or config.db
+
+        conn = eng.connect()
+        trans = conn.begin()
+
+        _connection_fixture_connection = conn
+        yield conn
+
+        _connection_fixture_connection = None
+
+        if trans.is_active:
+            trans.rollback()
+        # trans would not be active here if the test is using
+        # the legacy @provide_metadata decorator still, as it will
+        # run a close all connections.
+        conn.close()
+
+    @config.fixture()
+    def close_result_when_finished(self):
+        to_close = []
+        to_consume = []
+
+        def go(result, consume=False):
+            to_close.append(result)
+            if consume:
+                to_consume.append(result)
+
+        yield go
+        for r in to_consume:
+            try:
+                r.all()
+            except:
+                pass
+        for r in to_close:
+            try:
+                r.close()
+            except:
+                pass
+
+    @config.fixture()
+    def registry(self, metadata):
+        reg = registry(
+            metadata=metadata,
+            type_annotation_map={
+                str: sa.String().with_variant(
+                    sa.String(50), "mysql", "mariadb", "oracle"
+                )
+            },
+        )
+        yield reg
+        reg.dispose()
+
+    @config.fixture
+    def decl_base(self, metadata):
+        _md = metadata
+
+        class Base(DeclarativeBase):
+            metadata = _md
+            type_annotation_map = {
+                str: sa.String().with_variant(
+                    sa.String(50), "mysql", "mariadb", "oracle"
+                )
+            }
+
+        yield Base
+        Base.registry.dispose()
+
+    @config.fixture
+    def dc_decl_base(self, metadata):
+        _md = metadata
+
+        class Base(MappedAsDataclass, DeclarativeBase):
+            metadata = _md
+            type_annotation_map = {
+                str: sa.String().with_variant(
+                    sa.String(50), "mysql", "mariadb"
+                )
+            }
+
+        yield Base
+        Base.registry.dispose()
+
+    @config.fixture()
+    def future_connection(self, future_engine, connection):
+        # integrate the future_engine and connection fixtures so
+        # that users of the "connection" fixture will get at the
+        # "future" connection
+        yield connection
+
+    @config.fixture()
+    def future_engine(self):
+        yield
+
+    @config.fixture()
+    def testing_engine(self):
+        from .. import engines
+
+        def gen_testing_engine(
+            url=None,
+            options=None,
+            future=None,
+            asyncio=False,
+            transfer_staticpool=False,
+            share_pool=False,
+        ):
+            if options is None:
+                options = {}
+            options["scope"] = "fixture"
+            return engines.testing_engine(
+                url=url,
+                options=options,
+                asyncio=asyncio,
+                transfer_staticpool=transfer_staticpool,
+                share_pool=share_pool,
+            )
+
+        yield gen_testing_engine
+
+        engines.testing_reaper._drop_testing_engines("fixture")
+
+    @config.fixture()
+    def async_testing_engine(self, testing_engine):
+        def go(**kw):
+            kw["asyncio"] = True
+            return testing_engine(**kw)
+
+        return go
+
+    @config.fixture()
+    def metadata(self, request):
+        """Provide bound MetaData for a single test, dropping afterwards."""
+
+        from ...sql import schema
+
+        metadata = schema.MetaData()
+        request.instance.metadata = metadata
+        yield metadata
+        del request.instance.metadata
+
+        if (
+            _connection_fixture_connection
+            and _connection_fixture_connection.in_transaction()
+        ):
+            trans = _connection_fixture_connection.get_transaction()
+            trans.rollback()
+            with _connection_fixture_connection.begin():
+                drop_all_tables_from_metadata(
+                    metadata, _connection_fixture_connection
+                )
+        else:
+            drop_all_tables_from_metadata(metadata, config.db)
+
+    @config.fixture(
+        params=[
+            (rollback, second_operation, begin_nested)
+            for rollback in (True, False)
+            for second_operation in ("none", "execute", "begin")
+            for begin_nested in (
+                True,
+                False,
+            )
+        ]
+    )
+    def trans_ctx_manager_fixture(self, request, metadata):
+        rollback, second_operation, begin_nested = request.param
+
+        t = Table("test", metadata, Column("data", Integer))
+        eng = getattr(self, "bind", None) or config.db
+
+        t.create(eng)
+
+        def run_test(subject, trans_on_subject, execute_on_subject):
+            with subject.begin() as trans:
+                if begin_nested:
+                    if not config.requirements.savepoints.enabled:
+                        config.skip_test("savepoints not enabled")
+                    if execute_on_subject:
+                        nested_trans = subject.begin_nested()
+                    else:
+                        nested_trans = trans.begin_nested()
+
+                    with nested_trans:
+                        if execute_on_subject:
+                            subject.execute(t.insert(), {"data": 10})
+                        else:
+                            trans.execute(t.insert(), {"data": 10})
+
+                        # for nested trans, we always commit/rollback on the
+                        # "nested trans" object itself.
+                        # only Session(future=False) will affect savepoint
+                        # transaction for session.commit/rollback
+
+                        if rollback:
+                            nested_trans.rollback()
+                        else:
+                            nested_trans.commit()
+
+                        if second_operation != "none":
+                            with assertions.expect_raises_message(
+                                sa.exc.InvalidRequestError,
+                                "Can't operate on closed transaction "
+                                "inside context "
+                                "manager.  Please complete the context "
+                                "manager "
+                                "before emitting further commands.",
+                            ):
+                                if second_operation == "execute":
+                                    if execute_on_subject:
+                                        subject.execute(
+                                            t.insert(), {"data": 12}
+                                        )
+                                    else:
+                                        trans.execute(t.insert(), {"data": 12})
+                                elif second_operation == "begin":
+                                    if execute_on_subject:
+                                        subject.begin_nested()
+                                    else:
+                                        trans.begin_nested()
+
+                    # outside the nested trans block, but still inside the
+                    # transaction block, we can run SQL, and it will be
+                    # committed
+                    if execute_on_subject:
+                        subject.execute(t.insert(), {"data": 14})
+                    else:
+                        trans.execute(t.insert(), {"data": 14})
+
+                else:
+                    if execute_on_subject:
+                        subject.execute(t.insert(), {"data": 10})
+                    else:
+                        trans.execute(t.insert(), {"data": 10})
+
+                    if trans_on_subject:
+                        if rollback:
+                            subject.rollback()
+                        else:
+                            subject.commit()
+                    else:
+                        if rollback:
+                            trans.rollback()
+                        else:
+                            trans.commit()
+
+                    if second_operation != "none":
+                        with assertions.expect_raises_message(
+                            sa.exc.InvalidRequestError,
+                            "Can't operate on closed transaction inside "
+                            "context "
+                            "manager.  Please complete the context manager "
+                            "before emitting further commands.",
+                        ):
+                            if second_operation == "execute":
+                                if execute_on_subject:
+                                    subject.execute(t.insert(), {"data": 12})
+                                else:
+                                    trans.execute(t.insert(), {"data": 12})
+                            elif second_operation == "begin":
+                                if hasattr(trans, "begin"):
+                                    trans.begin()
+                                else:
+                                    subject.begin()
+                            elif second_operation == "begin_nested":
+                                if execute_on_subject:
+                                    subject.begin_nested()
+                                else:
+                                    trans.begin_nested()
+
+            expected_committed = 0
+            if begin_nested:
+                # begin_nested variant, we inserted a row after the nested
+                # block
+                expected_committed += 1
+            if not rollback:
+                # not rollback variant, our row inserted in the target
+                # block itself would be committed
+                expected_committed += 1
+
+            if execute_on_subject:
+                eq_(
+                    subject.scalar(select(func.count()).select_from(t)),
+                    expected_committed,
+                )
+            else:
+                with subject.connect() as conn:
+                    eq_(
+                        conn.scalar(select(func.count()).select_from(t)),
+                        expected_committed,
+                    )
+
+        return run_test
+
+
+_connection_fixture_connection = None
+
+
+class FutureEngineMixin:
+    """alembic's suite still using this"""
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/mypy.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/mypy.py
new file mode 100644
index 00000000..0832d892
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/mypy.py
@@ -0,0 +1,312 @@
+# testing/fixtures/mypy.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import inspect
+import os
+from pathlib import Path
+import re
+import shutil
+import sys
+import tempfile
+
+from .base import TestBase
+from .. import config
+from ..assertions import eq_
+from ... import util
+
+
+@config.add_to_marker.mypy
+class MypyTest(TestBase):
+    __requires__ = ("no_sqlalchemy2_stubs",)
+
+    @config.fixture(scope="function")
+    def per_func_cachedir(self):
+        yield from self._cachedir()
+
+    @config.fixture(scope="class")
+    def cachedir(self):
+        yield from self._cachedir()
+
+    def _cachedir(self):
+        # as of mypy 0.971 i think we need to keep mypy_path empty
+        mypy_path = ""
+
+        with tempfile.TemporaryDirectory() as cachedir:
+            with open(
+                Path(cachedir) / "sqla_mypy_config.cfg", "w"
+            ) as config_file:
+                config_file.write(
+                    f"""
+                    [mypy]\n
+                    plugins = sqlalchemy.ext.mypy.plugin\n
+                    show_error_codes = True\n
+                    {mypy_path}
+                    disable_error_code = no-untyped-call
+
+                    [mypy-sqlalchemy.*]
+                    ignore_errors = True
+
+                    """
+                )
+            with open(
+                Path(cachedir) / "plain_mypy_config.cfg", "w"
+            ) as config_file:
+                config_file.write(
+                    f"""
+                    [mypy]\n
+                    show_error_codes = True\n
+                    {mypy_path}
+                    disable_error_code = var-annotated,no-untyped-call
+                    [mypy-sqlalchemy.*]
+                    ignore_errors = True
+
+                    """
+                )
+            yield cachedir
+
+    @config.fixture()
+    def mypy_runner(self, cachedir):
+        from mypy import api
+
+        def run(path, use_plugin=False, use_cachedir=None):
+            if use_cachedir is None:
+                use_cachedir = cachedir
+            args = [
+                "--strict",
+                "--raise-exceptions",
+                "--cache-dir",
+                use_cachedir,
+                "--config-file",
+                os.path.join(
+                    use_cachedir,
+                    (
+                        "sqla_mypy_config.cfg"
+                        if use_plugin
+                        else "plain_mypy_config.cfg"
+                    ),
+                ),
+            ]
+
+            # mypy as of 0.990 is more aggressively blocking messaging
+            # for paths that are in sys.path, and as pytest puts currdir,
+            # test/ etc in sys.path, just copy the source file to the
+            # tempdir we are working in so that we don't have to try to
+            # manipulate sys.path and/or guess what mypy is doing
+            filename = os.path.basename(path)
+            test_program = os.path.join(use_cachedir, filename)
+            if path != test_program:
+                shutil.copyfile(path, test_program)
+            args.append(test_program)
+
+            # I set this locally but for the suite here needs to be
+            # disabled
+            os.environ.pop("MYPY_FORCE_COLOR", None)
+
+            stdout, stderr, exitcode = api.run(args)
+            return stdout, stderr, exitcode
+
+        return run
+
+    @config.fixture
+    def mypy_typecheck_file(self, mypy_runner):
+        def run(path, use_plugin=False):
+            expected_messages = self._collect_messages(path)
+            stdout, stderr, exitcode = mypy_runner(path, use_plugin=use_plugin)
+            self._check_output(
+                path, expected_messages, stdout, stderr, exitcode
+            )
+
+        return run
+
+    @staticmethod
+    def file_combinations(dirname):
+        if os.path.isabs(dirname):
+            path = dirname
+        else:
+            caller_path = inspect.stack()[1].filename
+            path = os.path.join(os.path.dirname(caller_path), dirname)
+        files = list(Path(path).glob("**/*.py"))
+
+        for extra_dir in config.options.mypy_extra_test_paths:
+            if extra_dir and os.path.isdir(extra_dir):
+                files.extend((Path(extra_dir) / dirname).glob("**/*.py"))
+        return files
+
+    def _collect_messages(self, path):
+        from sqlalchemy.ext.mypy.util import mypy_14
+
+        expected_messages = []
+        expected_re = re.compile(r"\s*# EXPECTED(_MYPY)?(_RE)?(_TYPE)?: (.+)")
+        py_ver_re = re.compile(r"^#\s*PYTHON_VERSION\s?>=\s?(\d+\.\d+)")
+        with open(path) as file_:
+            current_assert_messages = []
+            for num, line in enumerate(file_, 1):
+                m = py_ver_re.match(line)
+                if m:
+                    major, _, minor = m.group(1).partition(".")
+                    if sys.version_info < (int(major), int(minor)):
+                        config.skip_test(
+                            "Requires python >= %s" % (m.group(1))
+                        )
+                    continue
+
+                m = expected_re.match(line)
+                if m:
+                    is_mypy = bool(m.group(1))
+                    is_re = bool(m.group(2))
+                    is_type = bool(m.group(3))
+
+                    expected_msg = re.sub(r"# noqa[:]? ?.*", "", m.group(4))
+                    if is_type:
+                        if not is_re:
+                            # the goal here is that we can cut-and-paste
+                            # from vscode -> pylance into the
+                            # EXPECTED_TYPE: line, then the test suite will
+                            # validate that line against what mypy produces
+                            expected_msg = re.sub(
+                                r"([\[\]])",
+                                lambda m: rf"\{m.group(0)}",
+                                expected_msg,
+                            )
+
+                            # note making sure preceding text matches
+                            # with a dot, so that an expect for "Select"
+                            # does not match "TypedSelect"
+                            expected_msg = re.sub(
+                                r"([\w_]+)",
+                                lambda m: rf"(?:.*\.)?{m.group(1)}\*?",
+                                expected_msg,
+                            )
+
+                            expected_msg = re.sub(
+                                "List", "builtins.list", expected_msg
+                            )
+
+                            expected_msg = re.sub(
+                                r"\b(int|str|float|bool)\b",
+                                lambda m: rf"builtins.{m.group(0)}\*?",
+                                expected_msg,
+                            )
+                            # expected_msg = re.sub(
+                            #     r"(Sequence|Tuple|List|Union)",
+                            #     lambda m: fr"typing.{m.group(0)}\*?",
+                            #     expected_msg,
+                            # )
+
+                        is_mypy = is_re = True
+                        expected_msg = f'Revealed type is "{expected_msg}"'
+
+                    if mypy_14 and util.py39:
+                        # use_lowercase_names, py39 and above
+                        # https://github.com/python/mypy/blob/304997bfb85200fb521ac727ee0ce3e6085e5278/mypy/options.py#L363  # noqa: E501
+
+                        # skip first character which could be capitalized
+                        # "List item x not found" type of message
+                        expected_msg = expected_msg[0] + re.sub(
+                            (
+                                r"\b(List|Tuple|Dict|Set)\b"
+                                if is_type
+                                else r"\b(List|Tuple|Dict|Set|Type)\b"
+                            ),
+                            lambda m: m.group(1).lower(),
+                            expected_msg[1:],
+                        )
+
+                    if mypy_14 and util.py310:
+                        # use_or_syntax, py310 and above
+                        # https://github.com/python/mypy/blob/304997bfb85200fb521ac727ee0ce3e6085e5278/mypy/options.py#L368  # noqa: E501
+                        expected_msg = re.sub(
+                            r"Optional\[(.*?)\]",
+                            lambda m: f"{m.group(1)} | None",
+                            expected_msg,
+                        )
+                    current_assert_messages.append(
+                        (is_mypy, is_re, expected_msg.strip())
+                    )
+                elif current_assert_messages:
+                    expected_messages.extend(
+                        (num, is_mypy, is_re, expected_msg)
+                        for (
+                            is_mypy,
+                            is_re,
+                            expected_msg,
+                        ) in current_assert_messages
+                    )
+                    current_assert_messages[:] = []
+
+        return expected_messages
+
+    def _check_output(self, path, expected_messages, stdout, stderr, exitcode):
+        not_located = []
+        filename = os.path.basename(path)
+        if expected_messages:
+            # mypy 0.990 changed how return codes work, so don't assume a
+            # 1 or a 0 return code here, could be either depending on if
+            # errors were generated or not
+
+            output = []
+
+            raw_lines = stdout.split("\n")
+            while raw_lines:
+                e = raw_lines.pop(0)
+                if re.match(r".+\.py:\d+: error: .*", e):
+                    output.append(("error", e))
+                elif re.match(
+                    r".+\.py:\d+: note: +(?:Possible overload|def ).*", e
+                ):
+                    while raw_lines:
+                        ol = raw_lines.pop(0)
+                        if not re.match(r".+\.py:\d+: note: +def \[.*", ol):
+                            break
+                elif re.match(
+                    r".+\.py:\d+: note: .*(?:perhaps|suggestion)", e, re.I
+                ):
+                    pass
+                elif re.match(r".+\.py:\d+: note: .*", e):
+                    output.append(("note", e))
+
+            for num, is_mypy, is_re, msg in expected_messages:
+                msg = msg.replace("'", '"')
+                prefix = "[SQLAlchemy Mypy plugin] " if not is_mypy else ""
+                for idx, (typ, errmsg) in enumerate(output):
+                    if is_re:
+                        if re.match(
+                            rf".*{filename}\:{num}\: {typ}\: {prefix}{msg}",
+                            errmsg,
+                        ):
+                            break
+                    elif (
+                        f"{filename}:{num}: {typ}: {prefix}{msg}"
+                        in errmsg.replace("'", '"')
+                    ):
+                        break
+                else:
+                    not_located.append(msg)
+                    continue
+                del output[idx]
+
+            if not_located:
+                missing = "\n".join(not_located)
+                print("Couldn't locate expected messages:", missing, sep="\n")
+                if output:
+                    extra = "\n".join(msg for _, msg in output)
+                    print("Remaining messages:", extra, sep="\n")
+                assert False, "expected messages not found, see stdout"
+
+            if output:
+                print(f"{len(output)} messages from mypy were not consumed:")
+                print("\n".join(msg for _, msg in output))
+                assert False, "errors and/or notes remain, see stdout"
+
+        else:
+            if exitcode != 0:
+                print(stdout, stderr, sep="\n")
+
+            eq_(exitcode, 0, msg=stdout)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/orm.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/orm.py
new file mode 100644
index 00000000..77cb243a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/orm.py
@@ -0,0 +1,227 @@
+# testing/fixtures/orm.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+from typing import Any
+
+import sqlalchemy as sa
+from .base import TestBase
+from .sql import TablesTest
+from .. import assertions
+from .. import config
+from .. import schema
+from ..entities import BasicEntity
+from ..entities import ComparableEntity
+from ..util import adict
+from ... import orm
+from ...orm import DeclarativeBase
+from ...orm import events as orm_events
+from ...orm import registry
+
+
+class ORMTest(TestBase):
+    @config.fixture
+    def fixture_session(self):
+        return fixture_session()
+
+
+class MappedTest(ORMTest, TablesTest, assertions.AssertsExecutionResults):
+    # 'once', 'each', None
+    run_setup_classes = "once"
+
+    # 'once', 'each', None
+    run_setup_mappers = "each"
+
+    classes: Any = None
+
+    @config.fixture(autouse=True, scope="class")
+    def _setup_tables_test_class(self):
+        cls = self.__class__
+        cls._init_class()
+
+        if cls.classes is None:
+            cls.classes = adict()
+
+        cls._setup_once_tables()
+        cls._setup_once_classes()
+        cls._setup_once_mappers()
+        cls._setup_once_inserts()
+
+        yield
+
+        cls._teardown_once_class()
+        cls._teardown_once_metadata_bind()
+
+    @config.fixture(autouse=True, scope="function")
+    def _setup_tables_test_instance(self):
+        self._setup_each_tables()
+        self._setup_each_classes()
+        self._setup_each_mappers()
+        self._setup_each_inserts()
+
+        yield
+
+        orm.session.close_all_sessions()
+        self._teardown_each_mappers()
+        self._teardown_each_classes()
+        self._teardown_each_tables()
+
+    @classmethod
+    def _teardown_once_class(cls):
+        cls.classes.clear()
+
+    @classmethod
+    def _setup_once_classes(cls):
+        if cls.run_setup_classes == "once":
+            cls._with_register_classes(cls.setup_classes)
+
+    @classmethod
+    def _setup_once_mappers(cls):
+        if cls.run_setup_mappers == "once":
+            cls.mapper_registry, cls.mapper = cls._generate_registry()
+            cls._with_register_classes(cls.setup_mappers)
+
+    def _setup_each_mappers(self):
+        if self.run_setup_mappers != "once":
+            (
+                self.__class__.mapper_registry,
+                self.__class__.mapper,
+            ) = self._generate_registry()
+
+        if self.run_setup_mappers == "each":
+            self._with_register_classes(self.setup_mappers)
+
+    def _setup_each_classes(self):
+        if self.run_setup_classes == "each":
+            self._with_register_classes(self.setup_classes)
+
+    @classmethod
+    def _generate_registry(cls):
+        decl = registry(metadata=cls._tables_metadata)
+        return decl, decl.map_imperatively
+
+    @classmethod
+    def _with_register_classes(cls, fn):
+        """Run a setup method, framing the operation with a Base class
+        that will catch new subclasses to be established within
+        the "classes" registry.
+
+        """
+        cls_registry = cls.classes
+
+        class _Base:
+            def __init_subclass__(cls) -> None:
+                assert cls_registry is not None
+                cls_registry[cls.__name__] = cls
+                super().__init_subclass__()
+
+        class Basic(BasicEntity, _Base):
+            pass
+
+        class Comparable(ComparableEntity, _Base):
+            pass
+
+        cls.Basic = Basic
+        cls.Comparable = Comparable
+        fn()
+
+    def _teardown_each_mappers(self):
+        # some tests create mappers in the test bodies
+        # and will define setup_mappers as None -
+        # clear mappers in any case
+        if self.run_setup_mappers != "once":
+            orm.clear_mappers()
+
+    def _teardown_each_classes(self):
+        if self.run_setup_classes != "once":
+            self.classes.clear()
+
+    @classmethod
+    def setup_classes(cls):
+        pass
+
+    @classmethod
+    def setup_mappers(cls):
+        pass
+
+
+class DeclarativeMappedTest(MappedTest):
+    run_setup_classes = "once"
+    run_setup_mappers = "once"
+
+    @classmethod
+    def _setup_once_tables(cls):
+        pass
+
+    @classmethod
+    def _with_register_classes(cls, fn):
+        cls_registry = cls.classes
+
+        class _DeclBase(DeclarativeBase):
+            __table_cls__ = schema.Table
+            metadata = cls._tables_metadata
+            type_annotation_map = {
+                str: sa.String().with_variant(
+                    sa.String(50), "mysql", "mariadb", "oracle"
+                )
+            }
+
+            def __init_subclass__(cls, **kw) -> None:
+                assert cls_registry is not None
+                cls_registry[cls.__name__] = cls
+                super().__init_subclass__(**kw)
+
+        cls.DeclarativeBasic = _DeclBase
+
+        # sets up cls.Basic which is helpful for things like composite
+        # classes
+        super()._with_register_classes(fn)
+
+        if cls._tables_metadata.tables and cls.run_create_tables:
+            cls._tables_metadata.create_all(config.db)
+
+
+class RemoveORMEventsGlobally:
+    @config.fixture(autouse=True)
+    def _remove_listeners(self):
+        yield
+        orm_events.MapperEvents._clear()
+        orm_events.InstanceEvents._clear()
+        orm_events.SessionEvents._clear()
+        orm_events.InstrumentationEvents._clear()
+        orm_events.QueryEvents._clear()
+
+
+_fixture_sessions = set()
+
+
+def fixture_session(**kw):
+    kw.setdefault("autoflush", True)
+    kw.setdefault("expire_on_commit", True)
+
+    bind = kw.pop("bind", config.db)
+
+    sess = orm.Session(bind, **kw)
+    _fixture_sessions.add(sess)
+    return sess
+
+
+def close_all_sessions():
+    # will close all still-referenced sessions
+    orm.close_all_sessions()
+    _fixture_sessions.clear()
+
+
+def stop_test_class_inside_fixtures(cls):
+    close_all_sessions()
+    orm.clear_mappers()
+
+
+def after_test():
+    if _fixture_sessions:
+        close_all_sessions()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/sql.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/sql.py
new file mode 100644
index 00000000..44cf21c2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/fixtures/sql.py
@@ -0,0 +1,503 @@
+# testing/fixtures/sql.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+from __future__ import annotations
+
+import itertools
+import random
+import re
+import sys
+
+import sqlalchemy as sa
+from .base import TestBase
+from .. import config
+from .. import mock
+from ..assertions import eq_
+from ..assertions import ne_
+from ..util import adict
+from ..util import drop_all_tables_from_metadata
+from ... import event
+from ... import util
+from ...schema import sort_tables_and_constraints
+from ...sql import visitors
+from ...sql.elements import ClauseElement
+
+
+class TablesTest(TestBase):
+    # 'once', None
+    run_setup_bind = "once"
+
+    # 'once', 'each', None
+    run_define_tables = "once"
+
+    # 'once', 'each', None
+    run_create_tables = "once"
+
+    # 'once', 'each', None
+    run_inserts = "each"
+
+    # 'each', None
+    run_deletes = "each"
+
+    # 'once', None
+    run_dispose_bind = None
+
+    bind = None
+    _tables_metadata = None
+    tables = None
+    other = None
+    sequences = None
+
+    @config.fixture(autouse=True, scope="class")
+    def _setup_tables_test_class(self):
+        cls = self.__class__
+        cls._init_class()
+
+        cls._setup_once_tables()
+
+        cls._setup_once_inserts()
+
+        yield
+
+        cls._teardown_once_metadata_bind()
+
+    @config.fixture(autouse=True, scope="function")
+    def _setup_tables_test_instance(self):
+        self._setup_each_tables()
+        self._setup_each_inserts()
+
+        yield
+
+        self._teardown_each_tables()
+
+    @property
+    def tables_test_metadata(self):
+        return self._tables_metadata
+
+    @classmethod
+    def _init_class(cls):
+        if cls.run_define_tables == "each":
+            if cls.run_create_tables == "once":
+                cls.run_create_tables = "each"
+            assert cls.run_inserts in ("each", None)
+
+        cls.other = adict()
+        cls.tables = adict()
+        cls.sequences = adict()
+
+        cls.bind = cls.setup_bind()
+        cls._tables_metadata = sa.MetaData()
+
+    @classmethod
+    def _setup_once_inserts(cls):
+        if cls.run_inserts == "once":
+            cls._load_fixtures()
+            with cls.bind.begin() as conn:
+                cls.insert_data(conn)
+
+    @classmethod
+    def _setup_once_tables(cls):
+        if cls.run_define_tables == "once":
+            cls.define_tables(cls._tables_metadata)
+            if cls.run_create_tables == "once":
+                cls._tables_metadata.create_all(cls.bind)
+            cls.tables.update(cls._tables_metadata.tables)
+            cls.sequences.update(cls._tables_metadata._sequences)
+
+    def _setup_each_tables(self):
+        if self.run_define_tables == "each":
+            self.define_tables(self._tables_metadata)
+            if self.run_create_tables == "each":
+                self._tables_metadata.create_all(self.bind)
+            self.tables.update(self._tables_metadata.tables)
+            self.sequences.update(self._tables_metadata._sequences)
+        elif self.run_create_tables == "each":
+            self._tables_metadata.create_all(self.bind)
+
+    def _setup_each_inserts(self):
+        if self.run_inserts == "each":
+            self._load_fixtures()
+            with self.bind.begin() as conn:
+                self.insert_data(conn)
+
+    def _teardown_each_tables(self):
+        if self.run_define_tables == "each":
+            self.tables.clear()
+            if self.run_create_tables == "each":
+                drop_all_tables_from_metadata(self._tables_metadata, self.bind)
+            self._tables_metadata.clear()
+        elif self.run_create_tables == "each":
+            drop_all_tables_from_metadata(self._tables_metadata, self.bind)
+
+        savepoints = getattr(config.requirements, "savepoints", False)
+        if savepoints:
+            savepoints = savepoints.enabled
+
+        # no need to run deletes if tables are recreated on setup
+        if (
+            self.run_define_tables != "each"
+            and self.run_create_tables != "each"
+            and self.run_deletes == "each"
+        ):
+            with self.bind.begin() as conn:
+                for table in reversed(
+                    [
+                        t
+                        for (t, fks) in sort_tables_and_constraints(
+                            self._tables_metadata.tables.values()
+                        )
+                        if t is not None
+                    ]
+                ):
+                    try:
+                        if savepoints:
+                            with conn.begin_nested():
+                                conn.execute(table.delete())
+                        else:
+                            conn.execute(table.delete())
+                    except sa.exc.DBAPIError as ex:
+                        print(
+                            ("Error emptying table %s: %r" % (table, ex)),
+                            file=sys.stderr,
+                        )
+
+    @classmethod
+    def _teardown_once_metadata_bind(cls):
+        if cls.run_create_tables:
+            drop_all_tables_from_metadata(cls._tables_metadata, cls.bind)
+
+        if cls.run_dispose_bind == "once":
+            cls.dispose_bind(cls.bind)
+
+        cls._tables_metadata.bind = None
+
+        if cls.run_setup_bind is not None:
+            cls.bind = None
+
+    @classmethod
+    def setup_bind(cls):
+        return config.db
+
+    @classmethod
+    def dispose_bind(cls, bind):
+        if hasattr(bind, "dispose"):
+            bind.dispose()
+        elif hasattr(bind, "close"):
+            bind.close()
+
+    @classmethod
+    def define_tables(cls, metadata):
+        pass
+
+    @classmethod
+    def fixtures(cls):
+        return {}
+
+    @classmethod
+    def insert_data(cls, connection):
+        pass
+
+    def sql_count_(self, count, fn):
+        self.assert_sql_count(self.bind, fn, count)
+
+    def sql_eq_(self, callable_, statements):
+        self.assert_sql(self.bind, callable_, statements)
+
+    @classmethod
+    def _load_fixtures(cls):
+        """Insert rows as represented by the fixtures() method."""
+        headers, rows = {}, {}
+        for table, data in cls.fixtures().items():
+            if len(data) < 2:
+                continue
+            if isinstance(table, str):
+                table = cls.tables[table]
+            headers[table] = data[0]
+            rows[table] = data[1:]
+        for table, fks in sort_tables_and_constraints(
+            cls._tables_metadata.tables.values()
+        ):
+            if table is None:
+                continue
+            if table not in headers:
+                continue
+            with cls.bind.begin() as conn:
+                conn.execute(
+                    table.insert(),
+                    [
+                        dict(zip(headers[table], column_values))
+                        for column_values in rows[table]
+                    ],
+                )
+
+
+class NoCache:
+    @config.fixture(autouse=True, scope="function")
+    def _disable_cache(self):
+        _cache = config.db._compiled_cache
+        config.db._compiled_cache = None
+        yield
+        config.db._compiled_cache = _cache
+
+
+class RemovesEvents:
+    @util.memoized_property
+    def _event_fns(self):
+        return set()
+
+    def event_listen(self, target, name, fn, **kw):
+        self._event_fns.add((target, name, fn))
+        event.listen(target, name, fn, **kw)
+
+    @config.fixture(autouse=True, scope="function")
+    def _remove_events(self):
+        yield
+        for key in self._event_fns:
+            event.remove(*key)
+
+
+class ComputedReflectionFixtureTest(TablesTest):
+    run_inserts = run_deletes = None
+
+    __backend__ = True
+    __requires__ = ("computed_columns", "table_reflection")
+
+    regexp = re.compile(r"[\[\]\(\)\s`'\"]*")
+
+    def normalize(self, text):
+        return self.regexp.sub("", text).lower()
+
+    @classmethod
+    def define_tables(cls, metadata):
+        from ... import Integer
+        from ... import testing
+        from ...schema import Column
+        from ...schema import Computed
+        from ...schema import Table
+
+        Table(
+            "computed_default_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("normal", Integer),
+            Column("computed_col", Integer, Computed("normal + 42")),
+            Column("with_default", Integer, server_default="42"),
+        )
+
+        t = Table(
+            "computed_column_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("normal", Integer),
+            Column("computed_no_flag", Integer, Computed("normal + 42")),
+        )
+
+        if testing.requires.schemas.enabled:
+            t2 = Table(
+                "computed_column_table",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("normal", Integer),
+                Column("computed_no_flag", Integer, Computed("normal / 42")),
+                schema=config.test_schema,
+            )
+
+        if testing.requires.computed_columns_virtual.enabled:
+            t.append_column(
+                Column(
+                    "computed_virtual",
+                    Integer,
+                    Computed("normal + 2", persisted=False),
+                )
+            )
+            if testing.requires.schemas.enabled:
+                t2.append_column(
+                    Column(
+                        "computed_virtual",
+                        Integer,
+                        Computed("normal / 2", persisted=False),
+                    )
+                )
+        if testing.requires.computed_columns_stored.enabled:
+            t.append_column(
+                Column(
+                    "computed_stored",
+                    Integer,
+                    Computed("normal - 42", persisted=True),
+                )
+            )
+            if testing.requires.schemas.enabled:
+                t2.append_column(
+                    Column(
+                        "computed_stored",
+                        Integer,
+                        Computed("normal * 42", persisted=True),
+                    )
+                )
+
+
+class CacheKeyFixture:
+    def _compare_equal(self, a, b, compare_values):
+        a_key = a._generate_cache_key()
+        b_key = b._generate_cache_key()
+
+        if a_key is None:
+            assert a._annotations.get("nocache")
+
+            assert b_key is None
+        else:
+            eq_(a_key.key, b_key.key)
+            eq_(hash(a_key.key), hash(b_key.key))
+
+            for a_param, b_param in zip(a_key.bindparams, b_key.bindparams):
+                assert a_param.compare(b_param, compare_values=compare_values)
+        return a_key, b_key
+
+    def _run_cache_key_fixture(self, fixture, compare_values):
+        case_a = fixture()
+        case_b = fixture()
+
+        for a, b in itertools.combinations_with_replacement(
+            range(len(case_a)), 2
+        ):
+            if a == b:
+                a_key, b_key = self._compare_equal(
+                    case_a[a], case_b[b], compare_values
+                )
+                if a_key is None:
+                    continue
+            else:
+                a_key = case_a[a]._generate_cache_key()
+                b_key = case_b[b]._generate_cache_key()
+
+                if a_key is None or b_key is None:
+                    if a_key is None:
+                        assert case_a[a]._annotations.get("nocache")
+                    if b_key is None:
+                        assert case_b[b]._annotations.get("nocache")
+                    continue
+
+                if a_key.key == b_key.key:
+                    for a_param, b_param in zip(
+                        a_key.bindparams, b_key.bindparams
+                    ):
+                        if not a_param.compare(
+                            b_param, compare_values=compare_values
+                        ):
+                            break
+                    else:
+                        # this fails unconditionally since we could not
+                        # find bound parameter values that differed.
+                        # Usually we intended to get two distinct keys here
+                        # so the failure will be more descriptive using the
+                        # ne_() assertion.
+                        ne_(a_key.key, b_key.key)
+                else:
+                    ne_(a_key.key, b_key.key)
+
+            # ClauseElement-specific test to ensure the cache key
+            # collected all the bound parameters that aren't marked
+            # as "literal execute"
+            if isinstance(case_a[a], ClauseElement) and isinstance(
+                case_b[b], ClauseElement
+            ):
+                assert_a_params = []
+                assert_b_params = []
+
+                for elem in visitors.iterate(case_a[a]):
+                    if elem.__visit_name__ == "bindparam":
+                        assert_a_params.append(elem)
+
+                for elem in visitors.iterate(case_b[b]):
+                    if elem.__visit_name__ == "bindparam":
+                        assert_b_params.append(elem)
+
+                # note we're asserting the order of the params as well as
+                # if there are dupes or not.  ordering has to be
+                # deterministic and matches what a traversal would provide.
+                eq_(
+                    sorted(a_key.bindparams, key=lambda b: b.key),
+                    sorted(
+                        util.unique_list(assert_a_params), key=lambda b: b.key
+                    ),
+                )
+                eq_(
+                    sorted(b_key.bindparams, key=lambda b: b.key),
+                    sorted(
+                        util.unique_list(assert_b_params), key=lambda b: b.key
+                    ),
+                )
+
+    def _run_cache_key_equal_fixture(self, fixture, compare_values):
+        case_a = fixture()
+        case_b = fixture()
+
+        for a, b in itertools.combinations_with_replacement(
+            range(len(case_a)), 2
+        ):
+            self._compare_equal(case_a[a], case_b[b], compare_values)
+
+
+def insertmanyvalues_fixture(
+    connection, randomize_rows=False, warn_on_downgraded=False
+):
+    dialect = connection.dialect
+    orig_dialect = dialect._deliver_insertmanyvalues_batches
+    orig_conn = connection._exec_insertmany_context
+
+    class RandomCursor:
+        __slots__ = ("cursor",)
+
+        def __init__(self, cursor):
+            self.cursor = cursor
+
+        # only this method is called by the deliver method.
+        # by not having the other methods we assert that those aren't being
+        # used
+
+        @property
+        def description(self):
+            return self.cursor.description
+
+        def fetchall(self):
+            rows = self.cursor.fetchall()
+            rows = list(rows)
+            random.shuffle(rows)
+            return rows
+
+    def _deliver_insertmanyvalues_batches(
+        connection,
+        cursor,
+        statement,
+        parameters,
+        generic_setinputsizes,
+        context,
+    ):
+        if randomize_rows:
+            cursor = RandomCursor(cursor)
+        for batch in orig_dialect(
+            connection,
+            cursor,
+            statement,
+            parameters,
+            generic_setinputsizes,
+            context,
+        ):
+            if warn_on_downgraded and batch.is_downgraded:
+                util.warn("Batches were downgraded for sorted INSERT")
+
+            yield batch
+
+    def _exec_insertmany_context(dialect, context):
+        with mock.patch.object(
+            dialect,
+            "_deliver_insertmanyvalues_batches",
+            new=_deliver_insertmanyvalues_batches,
+        ):
+            return orig_conn(dialect, context)
+
+    connection._exec_insertmany_context = _exec_insertmany_context
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/pickleable.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/pickleable.py
new file mode 100644
index 00000000..9317be63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/pickleable.py
@@ -0,0 +1,155 @@
+# testing/pickleable.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Classes used in pickling tests, need to be at the module level for
+unpickling.
+"""
+
+from __future__ import annotations
+
+from .entities import ComparableEntity
+from ..schema import Column
+from ..types import String
+
+
+class User(ComparableEntity):
+    pass
+
+
+class Order(ComparableEntity):
+    pass
+
+
+class Dingaling(ComparableEntity):
+    pass
+
+
+class EmailUser(User):
+    pass
+
+
+class Address(ComparableEntity):
+    pass
+
+
+# TODO: these are kind of arbitrary....
+class Child1(ComparableEntity):
+    pass
+
+
+class Child2(ComparableEntity):
+    pass
+
+
+class Parent(ComparableEntity):
+    pass
+
+
+class Screen:
+    def __init__(self, obj, parent=None):
+        self.obj = obj
+        self.parent = parent
+
+
+class Mixin:
+    email_address = Column(String)
+
+
+class AddressWMixin(Mixin, ComparableEntity):
+    pass
+
+
+class Foo:
+    def __init__(self, moredata, stuff="im stuff"):
+        self.data = "im data"
+        self.stuff = stuff
+        self.moredata = moredata
+
+    __hash__ = object.__hash__
+
+    def __eq__(self, other):
+        return (
+            other.data == self.data
+            and other.stuff == self.stuff
+            and other.moredata == self.moredata
+        )
+
+
+class Bar:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+
+    __hash__ = object.__hash__
+
+    def __eq__(self, other):
+        return (
+            other.__class__ is self.__class__
+            and other.x == self.x
+            and other.y == self.y
+        )
+
+    def __str__(self):
+        return "Bar(%d, %d)" % (self.x, self.y)
+
+
+class OldSchool:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+
+    def __eq__(self, other):
+        return (
+            other.__class__ is self.__class__
+            and other.x == self.x
+            and other.y == self.y
+        )
+
+
+class OldSchoolWithoutCompare:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+
+
+class BarWithoutCompare:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+
+    def __str__(self):
+        return "Bar(%d, %d)" % (self.x, self.y)
+
+
+class NotComparable:
+    def __init__(self, data):
+        self.data = data
+
+    def __hash__(self):
+        return id(self)
+
+    def __eq__(self, other):
+        return NotImplemented
+
+    def __ne__(self, other):
+        return NotImplemented
+
+
+class BrokenComparable:
+    def __init__(self, data):
+        self.data = data
+
+    def __hash__(self):
+        return id(self)
+
+    def __eq__(self, other):
+        raise NotImplementedError
+
+    def __ne__(self, other):
+        raise NotImplementedError
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__init__.py
new file mode 100644
index 00000000..ce960be9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__init__.py
@@ -0,0 +1,6 @@
+# testing/plugin/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..541e46ee
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/bootstrap.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/bootstrap.cpython-312.pyc
new file mode 100644
index 00000000..80892003
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/bootstrap.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/plugin_base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/plugin_base.cpython-312.pyc
new file mode 100644
index 00000000..c73ceb2a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/plugin_base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/pytestplugin.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/pytestplugin.cpython-312.pyc
new file mode 100644
index 00000000..8436e53b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/__pycache__/pytestplugin.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/bootstrap.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/bootstrap.py
new file mode 100644
index 00000000..2ad4d991
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/bootstrap.py
@@ -0,0 +1,51 @@
+# testing/plugin/bootstrap.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+"""
+Bootstrapper for test framework plugins.
+
+The entire rationale for this system is to get the modules in plugin/
+imported without importing all of the supporting library, so that we can
+set up things for testing before coverage starts.
+
+The rationale for all of plugin/ being *in* the supporting library in the
+first place is so that the testing and plugin suite is available to other
+libraries, mainly external SQLAlchemy and Alembic dialects, to make use
+of the same test environment and standard suites available to
+SQLAlchemy/Alembic themselves without the need to ship/install a separate
+package outside of SQLAlchemy.
+
+
+"""
+
+import importlib.util
+import os
+import sys
+
+
+bootstrap_file = locals()["bootstrap_file"]
+to_bootstrap = locals()["to_bootstrap"]
+
+
+def load_file_as_module(name):
+    path = os.path.join(os.path.dirname(bootstrap_file), "%s.py" % name)
+
+    spec = importlib.util.spec_from_file_location(name, path)
+    assert spec is not None
+    assert spec.loader is not None
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    return mod
+
+
+if to_bootstrap == "pytest":
+    sys.modules["sqla_plugin_base"] = load_file_as_module("plugin_base")
+    sys.modules["sqla_plugin_base"].bootstrapped_as_sqlalchemy = True
+    sys.modules["sqla_pytestplugin"] = load_file_as_module("pytestplugin")
+else:
+    raise Exception("unknown bootstrap: %s" % to_bootstrap)  # noqa
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/plugin_base.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/plugin_base.py
new file mode 100644
index 00000000..2dfa4414
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/plugin_base.py
@@ -0,0 +1,779 @@
+# testing/plugin/plugin_base.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+import abc
+from argparse import Namespace
+import configparser
+import logging
+import os
+from pathlib import Path
+import re
+import sys
+from typing import Any
+
+from sqlalchemy.testing import asyncio
+
+"""Testing extensions.
+
+this module is designed to work as a testing-framework-agnostic library,
+created so that multiple test frameworks can be supported at once
+(mostly so that we can migrate to new ones). The current target
+is pytest.
+
+"""
+
+# flag which indicates we are in the SQLAlchemy testing suite,
+# and not that of Alembic or a third party dialect.
+bootstrapped_as_sqlalchemy = False
+
+log = logging.getLogger("sqlalchemy.testing.plugin_base")
+
+# late imports
+fixtures = None
+engines = None
+exclusions = None
+warnings = None
+profiling = None
+provision = None
+assertions = None
+requirements = None
+config = None
+testing = None
+util = None
+file_config = None
+
+logging = None
+include_tags = set()
+exclude_tags = set()
+options: Namespace = None  # type: ignore
+
+
+def setup_options(make_option):
+    make_option(
+        "--log-info",
+        action="callback",
+        type=str,
+        callback=_log,
+        help="turn on info logging for <LOG> (multiple OK)",
+    )
+    make_option(
+        "--log-debug",
+        action="callback",
+        type=str,
+        callback=_log,
+        help="turn on debug logging for <LOG> (multiple OK)",
+    )
+    make_option(
+        "--db",
+        action="append",
+        type=str,
+        dest="db",
+        help="Use prefab database uri. Multiple OK, "
+        "first one is run by default.",
+    )
+    make_option(
+        "--dbs",
+        action="callback",
+        zeroarg_callback=_list_dbs,
+        help="List available prefab dbs",
+    )
+    make_option(
+        "--dburi",
+        action="append",
+        type=str,
+        dest="dburi",
+        help="Database uri.  Multiple OK, first one is run by default.",
+    )
+    make_option(
+        "--dbdriver",
+        action="append",
+        type=str,
+        dest="dbdriver",
+        help="Additional database drivers to include in tests.  "
+        "These are linked to the existing database URLs by the "
+        "provisioning system.",
+    )
+    make_option(
+        "--dropfirst",
+        action="store_true",
+        dest="dropfirst",
+        help="Drop all tables in the target database first",
+    )
+    make_option(
+        "--disable-asyncio",
+        action="store_true",
+        help="disable test / fixtures / provisoning running in asyncio",
+    )
+    make_option(
+        "--backend-only",
+        action="callback",
+        zeroarg_callback=_set_tag_include("backend"),
+        help=(
+            "Run only tests marked with __backend__ or __sparse_backend__; "
+            "this is now equivalent to the pytest -m backend mark expression"
+        ),
+    )
+    make_option(
+        "--nomemory",
+        action="callback",
+        zeroarg_callback=_set_tag_exclude("memory_intensive"),
+        help="Don't run memory profiling tests; "
+        "this is now equivalent to the pytest -m 'not memory_intensive' "
+        "mark expression",
+    )
+    make_option(
+        "--notimingintensive",
+        action="callback",
+        zeroarg_callback=_set_tag_exclude("timing_intensive"),
+        help="Don't run timing intensive tests; "
+        "this is now equivalent to the pytest -m 'not timing_intensive' "
+        "mark expression",
+    )
+    make_option(
+        "--nomypy",
+        action="callback",
+        zeroarg_callback=_set_tag_exclude("mypy"),
+        help="Don't run mypy typing tests; "
+        "this is now equivalent to the pytest -m 'not mypy' mark expression",
+    )
+    make_option(
+        "--profile-sort",
+        type=str,
+        default="cumulative",
+        dest="profilesort",
+        help="Type of sort for profiling standard output",
+    )
+    make_option(
+        "--profile-dump",
+        type=str,
+        dest="profiledump",
+        help="Filename where a single profile run will be dumped",
+    )
+    make_option(
+        "--low-connections",
+        action="store_true",
+        dest="low_connections",
+        help="Use a low number of distinct connections - "
+        "i.e. for Oracle TNS",
+    )
+    make_option(
+        "--write-idents",
+        type=str,
+        dest="write_idents",
+        help="write out generated follower idents to <file>, "
+        "when -n<num> is used",
+    )
+    make_option(
+        "--requirements",
+        action="callback",
+        type=str,
+        callback=_requirements_opt,
+        help="requirements class for testing, overrides setup.cfg",
+    )
+    make_option(
+        "--include-tag",
+        action="callback",
+        callback=_include_tag,
+        type=str,
+        help="Include tests with tag <tag>; "
+        "legacy, use pytest -m 'tag' instead",
+    )
+    make_option(
+        "--exclude-tag",
+        action="callback",
+        callback=_exclude_tag,
+        type=str,
+        help="Exclude tests with tag <tag>; "
+        "legacy, use pytest -m 'not tag' instead",
+    )
+    make_option(
+        "--write-profiles",
+        action="store_true",
+        dest="write_profiles",
+        default=False,
+        help="Write/update failing profiling data.",
+    )
+    make_option(
+        "--force-write-profiles",
+        action="store_true",
+        dest="force_write_profiles",
+        default=False,
+        help="Unconditionally write/update profiling data.",
+    )
+    make_option(
+        "--dump-pyannotate",
+        type=str,
+        dest="dump_pyannotate",
+        help="Run pyannotate and dump json info to given file",
+    )
+    make_option(
+        "--mypy-extra-test-path",
+        type=str,
+        action="append",
+        default=[],
+        dest="mypy_extra_test_paths",
+        help="Additional test directories to add to the mypy tests. "
+        "This is used only when running mypy tests. Multiple OK",
+    )
+    # db specific options
+    make_option(
+        "--postgresql-templatedb",
+        type=str,
+        help="name of template database to use for PostgreSQL "
+        "CREATE DATABASE (defaults to current database)",
+    )
+    make_option(
+        "--oracledb-thick-mode",
+        action="store_true",
+        help="enables the 'thick mode' when testing with oracle+oracledb",
+    )
+
+
+def configure_follower(follower_ident):
+    """Configure required state for a follower.
+
+    This invokes in the parent process and typically includes
+    database creation.
+
+    """
+    from sqlalchemy.testing import provision
+
+    provision.FOLLOWER_IDENT = follower_ident
+
+
+def memoize_important_follower_config(dict_):
+    """Store important configuration we will need to send to a follower.
+
+    This invokes in the parent process after normal config is set up.
+
+    Hook is currently not used.
+
+    """
+
+
+def restore_important_follower_config(dict_):
+    """Restore important configuration needed by a follower.
+
+    This invokes in the follower process.
+
+    Hook is currently not used.
+
+    """
+
+
+def read_config(root_path):
+    global file_config
+    file_config = configparser.ConfigParser()
+    file_config.read(
+        [str(root_path / "setup.cfg"), str(root_path / "test.cfg")]
+    )
+
+
+def pre_begin(opt):
+    """things to set up early, before coverage might be setup."""
+    global options
+    options = opt
+    for fn in pre_configure:
+        fn(options, file_config)
+
+
+def set_coverage_flag(value):
+    options.has_coverage = value
+
+
+def post_begin():
+    """things to set up later, once we know coverage is running."""
+    # Lazy setup of other options (post coverage)
+    for fn in post_configure:
+        fn(options, file_config)
+
+    # late imports, has to happen after config.
+    global util, fixtures, engines, exclusions, assertions, provision
+    global warnings, profiling, config, testing
+    from sqlalchemy import testing  # noqa
+    from sqlalchemy.testing import fixtures, engines, exclusions  # noqa
+    from sqlalchemy.testing import assertions, warnings, profiling  # noqa
+    from sqlalchemy.testing import config, provision  # noqa
+    from sqlalchemy import util  # noqa
+
+    warnings.setup_filters()
+
+
+def _log(opt_str, value, parser):
+    global logging
+    if not logging:
+        import logging
+
+        logging.basicConfig()
+
+    if opt_str.endswith("-info"):
+        logging.getLogger(value).setLevel(logging.INFO)
+    elif opt_str.endswith("-debug"):
+        logging.getLogger(value).setLevel(logging.DEBUG)
+
+
+def _list_dbs(*args):
+    if file_config is None:
+        # assume the current working directory is the one containing the
+        # setup file
+        read_config(Path.cwd())
+    print("Available --db options (use --dburi to override)")
+    for macro in sorted(file_config.options("db")):
+        print("%20s\t%s" % (macro, file_config.get("db", macro)))
+    sys.exit(0)
+
+
+def _requirements_opt(opt_str, value, parser):
+    _setup_requirements(value)
+
+
+def _set_tag_include(tag):
+    def _do_include_tag(opt_str, value, parser):
+        _include_tag(opt_str, tag, parser)
+
+    return _do_include_tag
+
+
+def _set_tag_exclude(tag):
+    def _do_exclude_tag(opt_str, value, parser):
+        _exclude_tag(opt_str, tag, parser)
+
+    return _do_exclude_tag
+
+
+def _exclude_tag(opt_str, value, parser):
+    exclude_tags.add(value.replace("-", "_"))
+
+
+def _include_tag(opt_str, value, parser):
+    include_tags.add(value.replace("-", "_"))
+
+
+pre_configure = []
+post_configure = []
+
+
+def pre(fn):
+    pre_configure.append(fn)
+    return fn
+
+
+def post(fn):
+    post_configure.append(fn)
+    return fn
+
+
+@pre
+def _setup_options(opt, file_config):
+    global options
+    options = opt
+
+
+@pre
+def _register_sqlite_numeric_dialect(opt, file_config):
+    from sqlalchemy.dialects import registry
+
+    registry.register(
+        "sqlite.pysqlite_numeric",
+        "sqlalchemy.dialects.sqlite.pysqlite",
+        "_SQLiteDialect_pysqlite_numeric",
+    )
+    registry.register(
+        "sqlite.pysqlite_dollar",
+        "sqlalchemy.dialects.sqlite.pysqlite",
+        "_SQLiteDialect_pysqlite_dollar",
+    )
+
+
+@post
+def __ensure_cext(opt, file_config):
+    if os.environ.get("REQUIRE_SQLALCHEMY_CEXT", "0") == "1":
+        from sqlalchemy.util import has_compiled_ext
+
+        try:
+            has_compiled_ext(raise_=True)
+        except ImportError as err:
+            raise AssertionError(
+                "REQUIRE_SQLALCHEMY_CEXT is set but can't import the "
+                "cython extensions"
+            ) from err
+
+
+@post
+def _init_symbols(options, file_config):
+    from sqlalchemy.testing import config
+
+    config._fixture_functions = _fixture_fn_class()
+
+
+@pre
+def _set_disable_asyncio(opt, file_config):
+    if opt.disable_asyncio:
+        asyncio.ENABLE_ASYNCIO = False
+
+
+@post
+def _engine_uri(options, file_config):
+    from sqlalchemy import testing
+    from sqlalchemy.testing import config
+    from sqlalchemy.testing import provision
+    from sqlalchemy.engine import url as sa_url
+
+    if options.dburi:
+        db_urls = list(options.dburi)
+    else:
+        db_urls = []
+
+    extra_drivers = options.dbdriver or []
+
+    if options.db:
+        for db_token in options.db:
+            for db in re.split(r"[,\s]+", db_token):
+                if db not in file_config.options("db"):
+                    raise RuntimeError(
+                        "Unknown URI specifier '%s'.  "
+                        "Specify --dbs for known uris." % db
+                    )
+                else:
+                    db_urls.append(file_config.get("db", db))
+
+    if not db_urls:
+        db_urls.append(file_config.get("db", "default"))
+
+    config._current = None
+
+    if options.write_idents and provision.FOLLOWER_IDENT:
+        for db_url in [sa_url.make_url(db_url) for db_url in db_urls]:
+            with open(options.write_idents, "a") as file_:
+                file_.write(
+                    f"{provision.FOLLOWER_IDENT} "
+                    f"{db_url.render_as_string(hide_password=False)}\n"
+                )
+
+    expanded_urls = list(provision.generate_db_urls(db_urls, extra_drivers))
+
+    for db_url in expanded_urls:
+        log.info("Adding database URL: %s", db_url)
+
+        cfg = provision.setup_config(
+            db_url, options, file_config, provision.FOLLOWER_IDENT
+        )
+        if not config._current:
+            cfg.set_as_current(cfg, testing)
+
+
+@post
+def _requirements(options, file_config):
+    requirement_cls = file_config.get("sqla_testing", "requirement_cls")
+    _setup_requirements(requirement_cls)
+
+
+def _setup_requirements(argument):
+    from sqlalchemy.testing import config
+    from sqlalchemy import testing
+
+    modname, clsname = argument.split(":")
+
+    # importlib.import_module() only introduced in 2.7, a little
+    # late
+    mod = __import__(modname)
+    for component in modname.split(".")[1:]:
+        mod = getattr(mod, component)
+    req_cls = getattr(mod, clsname)
+
+    config.requirements = testing.requires = req_cls()
+
+    config.bootstrapped_as_sqlalchemy = bootstrapped_as_sqlalchemy
+
+
+@post
+def _prep_testing_database(options, file_config):
+    from sqlalchemy.testing import config
+
+    if options.dropfirst:
+        from sqlalchemy.testing import provision
+
+        for cfg in config.Config.all_configs():
+            provision.drop_all_schema_objects(cfg, cfg.db)
+
+
+@post
+def _post_setup_options(opt, file_config):
+    from sqlalchemy.testing import config
+
+    config.options = options
+    config.file_config = file_config
+
+
+@post
+def _setup_profiling(options, file_config):
+    from sqlalchemy.testing import profiling
+
+    profiling._profile_stats = profiling.ProfileStatsFile(
+        file_config.get("sqla_testing", "profile_file"),
+        sort=options.profilesort,
+        dump=options.profiledump,
+    )
+
+
+def want_class(name, cls):
+    if not issubclass(cls, fixtures.TestBase):
+        return False
+    elif name.startswith("_"):
+        return False
+    else:
+        return True
+
+
+def want_method(cls, fn):
+    if not fn.__name__.startswith("test_"):
+        return False
+    elif fn.__module__ is None:
+        return False
+    else:
+        return True
+
+
+def generate_sub_tests(cls, module, markers):
+    if "backend" in markers or "sparse_backend" in markers:
+        sparse = "sparse_backend" in markers
+        for cfg in _possible_configs_for_cls(cls, sparse=sparse):
+            orig_name = cls.__name__
+
+            # we can have special chars in these names except for the
+            # pytest junit plugin, which is tripped up by the brackets
+            # and periods, so sanitize
+
+            alpha_name = re.sub(r"[_\[\]\.]+", "_", cfg.name)
+            alpha_name = re.sub(r"_+$", "", alpha_name)
+            name = "%s_%s" % (cls.__name__, alpha_name)
+            subcls = type(
+                name,
+                (cls,),
+                {"_sa_orig_cls_name": orig_name, "__only_on_config__": cfg},
+            )
+            setattr(module, name, subcls)
+            yield subcls
+    else:
+        yield cls
+
+
+def start_test_class_outside_fixtures(cls):
+    _do_skips(cls)
+    _setup_engine(cls)
+
+
+def stop_test_class(cls):
+    # close sessions, immediate connections, etc.
+    fixtures.stop_test_class_inside_fixtures(cls)
+
+    # close outstanding connection pool connections, dispose of
+    # additional engines
+    engines.testing_reaper.stop_test_class_inside_fixtures()
+
+
+def stop_test_class_outside_fixtures(cls):
+    engines.testing_reaper.stop_test_class_outside_fixtures()
+    provision.stop_test_class_outside_fixtures(config, config.db, cls)
+    try:
+        if not options.low_connections:
+            assertions.global_cleanup_assertions()
+    finally:
+        _restore_engine()
+
+
+def _restore_engine():
+    if config._current:
+        config._current.reset(testing)
+
+
+def final_process_cleanup():
+    engines.testing_reaper.final_cleanup()
+    assertions.global_cleanup_assertions()
+    _restore_engine()
+
+
+def _setup_engine(cls):
+    if getattr(cls, "__engine_options__", None):
+        opts = dict(cls.__engine_options__)
+        opts["scope"] = "class"
+        eng = engines.testing_engine(options=opts)
+        config._current.push_engine(eng, testing)
+
+
+def before_test(test, test_module_name, test_class, test_name):
+    # format looks like:
+    # "test.aaa_profiling.test_compiler.CompileTest.test_update_whereclause"
+
+    name = getattr(test_class, "_sa_orig_cls_name", test_class.__name__)
+
+    id_ = "%s.%s.%s" % (test_module_name, name, test_name)
+
+    profiling._start_current_test(id_)
+
+
+def after_test(test):
+    fixtures.after_test()
+    engines.testing_reaper.after_test()
+
+
+def after_test_fixtures(test):
+    engines.testing_reaper.after_test_outside_fixtures(test)
+
+
+def _possible_configs_for_cls(cls, reasons=None, sparse=False):
+    all_configs = set(config.Config.all_configs())
+
+    if cls.__unsupported_on__:
+        spec = exclusions.db_spec(*cls.__unsupported_on__)
+        for config_obj in list(all_configs):
+            if spec(config_obj):
+                all_configs.remove(config_obj)
+
+    if getattr(cls, "__only_on__", None):
+        spec = exclusions.db_spec(*util.to_list(cls.__only_on__))
+        for config_obj in list(all_configs):
+            if not spec(config_obj):
+                all_configs.remove(config_obj)
+
+    if getattr(cls, "__only_on_config__", None):
+        all_configs.intersection_update([cls.__only_on_config__])
+
+    if hasattr(cls, "__requires__"):
+        requirements = config.requirements
+        for config_obj in list(all_configs):
+            for requirement in cls.__requires__:
+                check = getattr(requirements, requirement)
+
+                skip_reasons = check.matching_config_reasons(config_obj)
+                if skip_reasons:
+                    all_configs.remove(config_obj)
+                    if reasons is not None:
+                        reasons.extend(skip_reasons)
+                    break
+
+    if hasattr(cls, "__prefer_requires__"):
+        non_preferred = set()
+        requirements = config.requirements
+        for config_obj in list(all_configs):
+            for requirement in cls.__prefer_requires__:
+                check = getattr(requirements, requirement)
+
+                if not check.enabled_for_config(config_obj):
+                    non_preferred.add(config_obj)
+        if all_configs.difference(non_preferred):
+            all_configs.difference_update(non_preferred)
+
+    if sparse:
+        # pick only one config from each base dialect
+        # sorted so we get the same backend each time selecting the highest
+        # server version info.
+        per_dialect = {}
+        for cfg in reversed(
+            sorted(
+                all_configs,
+                key=lambda cfg: (
+                    cfg.db.name,
+                    cfg.db.driver,
+                    cfg.db.dialect.server_version_info,
+                ),
+            )
+        ):
+            db = cfg.db.name
+            if db not in per_dialect:
+                per_dialect[db] = cfg
+        return per_dialect.values()
+
+    return all_configs
+
+
+def _do_skips(cls):
+    reasons = []
+    all_configs = _possible_configs_for_cls(cls, reasons)
+
+    if getattr(cls, "__skip_if__", False):
+        for c in getattr(cls, "__skip_if__"):
+            if c():
+                config.skip_test(
+                    "'%s' skipped by %s" % (cls.__name__, c.__name__)
+                )
+
+    if not all_configs:
+        msg = "'%s.%s' unsupported on any DB implementation %s%s" % (
+            cls.__module__,
+            cls.__name__,
+            ", ".join(
+                "'%s(%s)+%s'"
+                % (
+                    config_obj.db.name,
+                    ".".join(
+                        str(dig)
+                        for dig in exclusions._server_version(config_obj.db)
+                    ),
+                    config_obj.db.driver,
+                )
+                for config_obj in config.Config.all_configs()
+            ),
+            ", ".join(reasons),
+        )
+        config.skip_test(msg)
+    elif hasattr(cls, "__prefer_backends__"):
+        non_preferred = set()
+        spec = exclusions.db_spec(*util.to_list(cls.__prefer_backends__))
+        for config_obj in all_configs:
+            if not spec(config_obj):
+                non_preferred.add(config_obj)
+        if all_configs.difference(non_preferred):
+            all_configs.difference_update(non_preferred)
+
+    if config._current not in all_configs:
+        _setup_config(all_configs.pop(), cls)
+
+
+def _setup_config(config_obj, ctx):
+    config._current.push(config_obj, testing)
+
+
+class FixtureFunctions(abc.ABC):
+    @abc.abstractmethod
+    def skip_test_exception(self, *arg, **kw):
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def combinations(self, *args, **kw):
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def param_ident(self, *args, **kw):
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def fixture(self, *arg, **kw):
+        raise NotImplementedError()
+
+    def get_current_test_name(self):
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def mark_base_test_class(self) -> Any:
+        raise NotImplementedError()
+
+    @abc.abstractproperty
+    def add_to_marker(self):
+        raise NotImplementedError()
+
+
+_fixture_fn_class = None
+
+
+def set_fixture_functions(fixture_fn_class):
+    global _fixture_fn_class
+    _fixture_fn_class = fixture_fn_class
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/pytestplugin.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/pytestplugin.py
new file mode 100644
index 00000000..f6d47c63
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/plugin/pytestplugin.py
@@ -0,0 +1,868 @@
+# testing/plugin/pytestplugin.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import argparse
+import collections
+from functools import update_wrapper
+import inspect
+import itertools
+import operator
+import os
+import re
+import sys
+from typing import TYPE_CHECKING
+import uuid
+
+import pytest
+
+try:
+    # installed by bootstrap.py
+    if not TYPE_CHECKING:
+        import sqla_plugin_base as plugin_base
+except ImportError:
+    # assume we're a package, use traditional import
+    from . import plugin_base
+
+
+def pytest_addoption(parser):
+    group = parser.getgroup("sqlalchemy")
+
+    def make_option(name, **kw):
+        callback_ = kw.pop("callback", None)
+        if callback_:
+
+            class CallableAction(argparse.Action):
+                def __call__(
+                    self, parser, namespace, values, option_string=None
+                ):
+                    callback_(option_string, values, parser)
+
+            kw["action"] = CallableAction
+
+        zeroarg_callback = kw.pop("zeroarg_callback", None)
+        if zeroarg_callback:
+
+            class CallableAction(argparse.Action):
+                def __init__(
+                    self,
+                    option_strings,
+                    dest,
+                    default=False,
+                    required=False,
+                    help=None,  # noqa
+                ):
+                    super().__init__(
+                        option_strings=option_strings,
+                        dest=dest,
+                        nargs=0,
+                        const=True,
+                        default=default,
+                        required=required,
+                        help=help,
+                    )
+
+                def __call__(
+                    self, parser, namespace, values, option_string=None
+                ):
+                    zeroarg_callback(option_string, values, parser)
+
+            kw["action"] = CallableAction
+
+        group.addoption(name, **kw)
+
+    plugin_base.setup_options(make_option)
+
+
+def pytest_configure(config: pytest.Config):
+    plugin_base.read_config(config.rootpath)
+    if plugin_base.exclude_tags or plugin_base.include_tags:
+        new_expr = " and ".join(
+            list(plugin_base.include_tags)
+            + [f"not {tag}" for tag in plugin_base.exclude_tags]
+        )
+
+        if config.option.markexpr:
+            config.option.markexpr += f" and {new_expr}"
+        else:
+            config.option.markexpr = new_expr
+
+    if config.pluginmanager.hasplugin("xdist"):
+        config.pluginmanager.register(XDistHooks())
+
+    if hasattr(config, "workerinput"):
+        plugin_base.restore_important_follower_config(config.workerinput)
+        plugin_base.configure_follower(config.workerinput["follower_ident"])
+    else:
+        if config.option.write_idents and os.path.exists(
+            config.option.write_idents
+        ):
+            os.remove(config.option.write_idents)
+
+    plugin_base.pre_begin(config.option)
+
+    plugin_base.set_coverage_flag(
+        bool(getattr(config.option, "cov_source", False))
+    )
+
+    plugin_base.set_fixture_functions(PytestFixtureFunctions)
+
+    if config.option.dump_pyannotate:
+        global DUMP_PYANNOTATE
+        DUMP_PYANNOTATE = True
+
+
+DUMP_PYANNOTATE = False
+
+
+@pytest.fixture(autouse=True)
+def collect_types_fixture():
+    if DUMP_PYANNOTATE:
+        from pyannotate_runtime import collect_types
+
+        collect_types.start()
+    yield
+    if DUMP_PYANNOTATE:
+        collect_types.stop()
+
+
+def _log_sqlalchemy_info(session):
+    import sqlalchemy
+    from sqlalchemy import __version__
+    from sqlalchemy.util import has_compiled_ext
+    from sqlalchemy.util._has_cy import _CYEXTENSION_MSG
+
+    greet = "sqlalchemy installation"
+    site = "no user site" if sys.flags.no_user_site else "user site loaded"
+    msgs = [
+        f"SQLAlchemy {__version__} ({site})",
+        f"Path: {sqlalchemy.__file__}",
+    ]
+
+    if has_compiled_ext():
+        from sqlalchemy.cyextension import util
+
+        msgs.append(f"compiled extension enabled, e.g. {util.__file__} ")
+    else:
+        msgs.append(f"compiled extension not enabled; {_CYEXTENSION_MSG}")
+
+    pm = session.config.pluginmanager.get_plugin("terminalreporter")
+    if pm:
+        pm.write_sep("=", greet)
+        for m in msgs:
+            pm.write_line(m)
+    else:
+        # fancy pants reporter not found, fallback to plain print
+        print("=" * 25, greet, "=" * 25)
+        for m in msgs:
+            print(m)
+
+
+def pytest_sessionstart(session):
+    from sqlalchemy.testing import asyncio
+
+    _log_sqlalchemy_info(session)
+    asyncio._assume_async(plugin_base.post_begin)
+
+
+def pytest_sessionfinish(session):
+    from sqlalchemy.testing import asyncio
+
+    asyncio._maybe_async_provisioning(plugin_base.final_process_cleanup)
+
+    if session.config.option.dump_pyannotate:
+        from pyannotate_runtime import collect_types
+
+        collect_types.dump_stats(session.config.option.dump_pyannotate)
+
+
+def pytest_unconfigure(config):
+    from sqlalchemy.testing import asyncio
+
+    asyncio._shutdown()
+
+
+def pytest_collection_finish(session):
+    if session.config.option.dump_pyannotate:
+        from pyannotate_runtime import collect_types
+
+        lib_sqlalchemy = os.path.abspath("lib/sqlalchemy")
+
+        def _filter(filename):
+            filename = os.path.normpath(os.path.abspath(filename))
+            if "lib/sqlalchemy" not in os.path.commonpath(
+                [filename, lib_sqlalchemy]
+            ):
+                return None
+            if "testing" in filename:
+                return None
+
+            return filename
+
+        collect_types.init_types_collection(filter_filename=_filter)
+
+
+class XDistHooks:
+    def pytest_configure_node(self, node):
+        from sqlalchemy.testing import provision
+        from sqlalchemy.testing import asyncio
+
+        # the master for each node fills workerinput dictionary
+        # which pytest-xdist will transfer to the subprocess
+
+        plugin_base.memoize_important_follower_config(node.workerinput)
+
+        node.workerinput["follower_ident"] = "test_%s" % uuid.uuid4().hex[0:12]
+
+        asyncio._maybe_async_provisioning(
+            provision.create_follower_db, node.workerinput["follower_ident"]
+        )
+
+    def pytest_testnodedown(self, node, error):
+        from sqlalchemy.testing import provision
+        from sqlalchemy.testing import asyncio
+
+        asyncio._maybe_async_provisioning(
+            provision.drop_follower_db, node.workerinput["follower_ident"]
+        )
+
+
+def pytest_collection_modifyitems(session, config, items):
+    # look for all those classes that specify __backend__ and
+    # expand them out into per-database test cases.
+
+    # this is much easier to do within pytest_pycollect_makeitem, however
+    # pytest is iterating through cls.__dict__ as makeitem is
+    # called which causes a "dictionary changed size" error on py3k.
+    # I'd submit a pullreq for them to turn it into a list first, but
+    # it's to suit the rather odd use case here which is that we are adding
+    # new classes to a module on the fly.
+
+    from sqlalchemy.testing import asyncio
+
+    rebuilt_items = collections.defaultdict(
+        lambda: collections.defaultdict(list)
+    )
+
+    items[:] = [
+        item
+        for item in items
+        if item.getparent(pytest.Class) is not None
+        and not item.getparent(pytest.Class).name.startswith("_")
+    ]
+
+    test_classes = {item.getparent(pytest.Class) for item in items}
+
+    def collect(element):
+        for inst_or_fn in element.collect():
+            if isinstance(inst_or_fn, pytest.Collector):
+                yield from collect(inst_or_fn)
+            else:
+                yield inst_or_fn
+
+    def setup_test_classes():
+        for test_class in test_classes:
+            # transfer legacy __backend__ and __sparse_backend__ symbols
+            # to be markers
+            add_markers = set()
+            if getattr(test_class.cls, "__backend__", False) or getattr(
+                test_class.cls, "__only_on__", False
+            ):
+                add_markers = {"backend"}
+            elif getattr(test_class.cls, "__sparse_backend__", False):
+                add_markers = {"sparse_backend"}
+            else:
+                add_markers = frozenset()
+
+            existing_markers = {
+                mark.name for mark in test_class.iter_markers()
+            }
+            add_markers = add_markers - existing_markers
+            all_markers = existing_markers.union(add_markers)
+
+            for marker in add_markers:
+                test_class.add_marker(marker)
+
+            for sub_cls in plugin_base.generate_sub_tests(
+                test_class.cls, test_class.module, all_markers
+            ):
+                if sub_cls is not test_class.cls:
+                    per_cls_dict = rebuilt_items[test_class.cls]
+
+                    module = test_class.getparent(pytest.Module)
+
+                    new_cls = pytest.Class.from_parent(
+                        name=sub_cls.__name__, parent=module
+                    )
+                    for marker in add_markers:
+                        new_cls.add_marker(marker)
+
+                    for fn in collect(new_cls):
+                        per_cls_dict[fn.name].append(fn)
+
+    # class requirements will sometimes need to access the DB to check
+    # capabilities, so need to do this for async
+    asyncio._maybe_async_provisioning(setup_test_classes)
+
+    newitems = []
+    for item in items:
+        cls_ = item.cls
+        if cls_ in rebuilt_items:
+            newitems.extend(rebuilt_items[cls_][item.name])
+        else:
+            newitems.append(item)
+
+    # seems like the functions attached to a test class aren't sorted already?
+    # is that true and why's that? (when using unittest, they're sorted)
+    items[:] = sorted(
+        newitems,
+        key=lambda item: (
+            item.getparent(pytest.Module).name,
+            item.getparent(pytest.Class).name,
+            item.name,
+        ),
+    )
+
+
+def pytest_pycollect_makeitem(collector, name, obj):
+    if inspect.isclass(obj) and plugin_base.want_class(name, obj):
+        from sqlalchemy.testing import config
+
+        if config.any_async:
+            obj = _apply_maybe_async(obj)
+
+        return [
+            pytest.Class.from_parent(
+                name=parametrize_cls.__name__, parent=collector
+            )
+            for parametrize_cls in _parametrize_cls(collector.module, obj)
+        ]
+    elif (
+        inspect.isfunction(obj)
+        and collector.cls is not None
+        and plugin_base.want_method(collector.cls, obj)
+    ):
+        # None means, fall back to default logic, which includes
+        # method-level parametrize
+        return None
+    else:
+        # empty list means skip this item
+        return []
+
+
+def _is_wrapped_coroutine_function(fn):
+    while hasattr(fn, "__wrapped__"):
+        fn = fn.__wrapped__
+
+    return inspect.iscoroutinefunction(fn)
+
+
+def _apply_maybe_async(obj, recurse=True):
+    from sqlalchemy.testing import asyncio
+
+    for name, value in vars(obj).items():
+        if (
+            (callable(value) or isinstance(value, classmethod))
+            and not getattr(value, "_maybe_async_applied", False)
+            and (name.startswith("test_"))
+            and not _is_wrapped_coroutine_function(value)
+        ):
+            is_classmethod = False
+            if isinstance(value, classmethod):
+                value = value.__func__
+                is_classmethod = True
+
+            @_pytest_fn_decorator
+            def make_async(fn, *args, **kwargs):
+                return asyncio._maybe_async(fn, *args, **kwargs)
+
+            do_async = make_async(value)
+            if is_classmethod:
+                do_async = classmethod(do_async)
+            do_async._maybe_async_applied = True
+
+            setattr(obj, name, do_async)
+    if recurse:
+        for cls in obj.mro()[1:]:
+            if cls != object:
+                _apply_maybe_async(cls, False)
+    return obj
+
+
+def _parametrize_cls(module, cls):
+    """implement a class-based version of pytest parametrize."""
+
+    if "_sa_parametrize" not in cls.__dict__:
+        return [cls]
+
+    _sa_parametrize = cls._sa_parametrize
+    classes = []
+    for full_param_set in itertools.product(
+        *[params for argname, params in _sa_parametrize]
+    ):
+        cls_variables = {}
+
+        for argname, param in zip(
+            [_sa_param[0] for _sa_param in _sa_parametrize], full_param_set
+        ):
+            if not argname:
+                raise TypeError("need argnames for class-based combinations")
+            argname_split = re.split(r",\s*", argname)
+            for arg, val in zip(argname_split, param.values):
+                cls_variables[arg] = val
+        parametrized_name = "_".join(
+            re.sub(r"\W", "", token)
+            for param in full_param_set
+            for token in param.id.split("-")
+        )
+        name = "%s_%s" % (cls.__name__, parametrized_name)
+        newcls = type.__new__(type, name, (cls,), cls_variables)
+        setattr(module, name, newcls)
+        classes.append(newcls)
+    return classes
+
+
+_current_class = None
+
+
+def pytest_runtest_setup(item):
+    from sqlalchemy.testing import asyncio
+
+    # pytest_runtest_setup runs *before* pytest fixtures with scope="class".
+    # plugin_base.start_test_class_outside_fixtures may opt to raise SkipTest
+    # for the whole class and has to run things that are across all current
+    # databases, so we run this outside of the pytest fixture system altogether
+    # and ensure asyncio greenlet if any engines are async
+
+    global _current_class
+
+    if isinstance(item, pytest.Function) and _current_class is None:
+        asyncio._maybe_async_provisioning(
+            plugin_base.start_test_class_outside_fixtures,
+            item.cls,
+        )
+        _current_class = item.getparent(pytest.Class)
+
+
+@pytest.hookimpl(hookwrapper=True)
+def pytest_runtest_teardown(item, nextitem):
+    # runs inside of pytest function fixture scope
+    # after test function runs
+
+    from sqlalchemy.testing import asyncio
+
+    asyncio._maybe_async(plugin_base.after_test, item)
+
+    yield
+    # this is now after all the fixture teardown have run, the class can be
+    # finalized. Since pytest v7 this finalizer can no longer be added in
+    # pytest_runtest_setup since the class has not yet been setup at that
+    # time.
+    # See https://github.com/pytest-dev/pytest/issues/9343
+    global _current_class, _current_report
+
+    if _current_class is not None and (
+        # last test or a new class
+        nextitem is None
+        or nextitem.getparent(pytest.Class) is not _current_class
+    ):
+        _current_class = None
+
+        try:
+            asyncio._maybe_async_provisioning(
+                plugin_base.stop_test_class_outside_fixtures, item.cls
+            )
+        except Exception as e:
+            # in case of an exception during teardown attach the original
+            # error to the exception message, otherwise it will get lost
+            if _current_report.failed:
+                if not e.args:
+                    e.args = (
+                        "__Original test failure__:\n"
+                        + _current_report.longreprtext,
+                    )
+                elif e.args[-1] and isinstance(e.args[-1], str):
+                    args = list(e.args)
+                    args[-1] += (
+                        "\n__Original test failure__:\n"
+                        + _current_report.longreprtext
+                    )
+                    e.args = tuple(args)
+                else:
+                    e.args += (
+                        "__Original test failure__",
+                        _current_report.longreprtext,
+                    )
+            raise
+        finally:
+            _current_report = None
+
+
+def pytest_runtest_call(item):
+    # runs inside of pytest function fixture scope
+    # before test function runs
+
+    from sqlalchemy.testing import asyncio
+
+    asyncio._maybe_async(
+        plugin_base.before_test,
+        item,
+        item.module.__name__,
+        item.cls,
+        item.name,
+    )
+
+
+_current_report = None
+
+
+def pytest_runtest_logreport(report):
+    global _current_report
+    if report.when == "call":
+        _current_report = report
+
+
+@pytest.fixture(scope="class")
+def setup_class_methods(request):
+    from sqlalchemy.testing import asyncio
+
+    cls = request.cls
+
+    if hasattr(cls, "setup_test_class"):
+        asyncio._maybe_async(cls.setup_test_class)
+
+    yield
+
+    if hasattr(cls, "teardown_test_class"):
+        asyncio._maybe_async(cls.teardown_test_class)
+
+    asyncio._maybe_async(plugin_base.stop_test_class, cls)
+
+
+@pytest.fixture(scope="function")
+def setup_test_methods(request):
+    from sqlalchemy.testing import asyncio
+
+    # called for each test
+
+    self = request.instance
+
+    # before this fixture runs:
+
+    # 1. function level "autouse" fixtures under py3k (examples: TablesTest
+    #    define tables / data, MappedTest define tables / mappers / data)
+
+    # 2. was for p2k. no longer applies
+
+    # 3. run outer xdist-style setup
+    if hasattr(self, "setup_test"):
+        asyncio._maybe_async(self.setup_test)
+
+    # alembic test suite is using setUp and tearDown
+    # xdist methods; support these in the test suite
+    # for the near term
+    if hasattr(self, "setUp"):
+        asyncio._maybe_async(self.setUp)
+
+    # inside the yield:
+    # 4. function level fixtures defined on test functions themselves,
+    #    e.g. "connection", "metadata" run next
+
+    # 5. pytest hook pytest_runtest_call then runs
+
+    # 6. test itself runs
+
+    yield
+
+    # yield finishes:
+
+    # 7. function level fixtures defined on test functions
+    #    themselves, e.g. "connection" rolls back the transaction, "metadata"
+    #    emits drop all
+
+    # 8. pytest hook pytest_runtest_teardown hook runs, this is associated
+    #    with fixtures close all sessions, provisioning.stop_test_class(),
+    #    engines.testing_reaper -> ensure all connection pool connections
+    #    are returned, engines created by testing_engine that aren't the
+    #    config engine are disposed
+
+    asyncio._maybe_async(plugin_base.after_test_fixtures, self)
+
+    # 10. run xdist-style teardown
+    if hasattr(self, "tearDown"):
+        asyncio._maybe_async(self.tearDown)
+
+    if hasattr(self, "teardown_test"):
+        asyncio._maybe_async(self.teardown_test)
+
+    # 11. was for p2k. no longer applies
+
+    # 12. function level "autouse" fixtures under py3k (examples: TablesTest /
+    #    MappedTest delete table data, possibly drop tables and clear mappers
+    #    depending on the flags defined by the test class)
+
+
+def _pytest_fn_decorator(target):
+    """Port of langhelpers.decorator with pytest-specific tricks."""
+
+    from sqlalchemy.util.langhelpers import format_argspec_plus
+    from sqlalchemy.util.compat import inspect_getfullargspec
+
+    def _exec_code_in_env(code, env, fn_name):
+        # note this is affected by "from __future__ import annotations" at
+        # the top; exec'ed code will use non-evaluated annotations
+        # which allows us to be more flexible with code rendering
+        # in format_argpsec_plus()
+        exec(code, env)
+        return env[fn_name]
+
+    def decorate(fn, add_positional_parameters=()):
+        spec = inspect_getfullargspec(fn)
+        if add_positional_parameters:
+            spec.args.extend(add_positional_parameters)
+
+        metadata = dict(
+            __target_fn="__target_fn", __orig_fn="__orig_fn", name=fn.__name__
+        )
+        metadata.update(format_argspec_plus(spec, grouped=False))
+        code = (
+            """\
+def %(name)s%(grouped_args)s:
+    return %(__target_fn)s(%(__orig_fn)s, %(apply_kw)s)
+"""
+            % metadata
+        )
+        decorated = _exec_code_in_env(
+            code, {"__target_fn": target, "__orig_fn": fn}, fn.__name__
+        )
+        if not add_positional_parameters:
+            decorated.__defaults__ = getattr(fn, "__func__", fn).__defaults__
+            decorated.__wrapped__ = fn
+            return update_wrapper(decorated, fn)
+        else:
+            # this is the pytest hacky part.  don't do a full update wrapper
+            # because pytest is really being sneaky about finding the args
+            # for the wrapped function
+            decorated.__module__ = fn.__module__
+            decorated.__name__ = fn.__name__
+            if hasattr(fn, "pytestmark"):
+                decorated.pytestmark = fn.pytestmark
+            return decorated
+
+    return decorate
+
+
+class PytestFixtureFunctions(plugin_base.FixtureFunctions):
+    def skip_test_exception(self, *arg, **kw):
+        return pytest.skip.Exception(*arg, **kw)
+
+    @property
+    def add_to_marker(self):
+        return pytest.mark
+
+    def mark_base_test_class(self):
+        return pytest.mark.usefixtures(
+            "setup_class_methods", "setup_test_methods"
+        )
+
+    _combination_id_fns = {
+        "i": lambda obj: obj,
+        "r": repr,
+        "s": str,
+        "n": lambda obj: (
+            obj.__name__ if hasattr(obj, "__name__") else type(obj).__name__
+        ),
+    }
+
+    def combinations(self, *arg_sets, **kw):
+        """Facade for pytest.mark.parametrize.
+
+        Automatically derives argument names from the callable which in our
+        case is always a method on a class with positional arguments.
+
+        ids for parameter sets are derived using an optional template.
+
+        """
+        from sqlalchemy.testing import exclusions
+
+        if len(arg_sets) == 1 and hasattr(arg_sets[0], "__next__"):
+            arg_sets = list(arg_sets[0])
+
+        argnames = kw.pop("argnames", None)
+
+        def _filter_exclusions(args):
+            result = []
+            gathered_exclusions = []
+            for a in args:
+                if isinstance(a, exclusions.compound):
+                    gathered_exclusions.append(a)
+                else:
+                    result.append(a)
+
+            return result, gathered_exclusions
+
+        id_ = kw.pop("id_", None)
+
+        tobuild_pytest_params = []
+        has_exclusions = False
+        if id_:
+            _combination_id_fns = self._combination_id_fns
+
+            # because itemgetter is not consistent for one argument vs.
+            # multiple, make it multiple in all cases and use a slice
+            # to omit the first argument
+            _arg_getter = operator.itemgetter(
+                0,
+                *[
+                    idx
+                    for idx, char in enumerate(id_)
+                    if char in ("n", "r", "s", "a")
+                ],
+            )
+            fns = [
+                (operator.itemgetter(idx), _combination_id_fns[char])
+                for idx, char in enumerate(id_)
+                if char in _combination_id_fns
+            ]
+
+            for arg in arg_sets:
+                if not isinstance(arg, tuple):
+                    arg = (arg,)
+
+                fn_params, param_exclusions = _filter_exclusions(arg)
+
+                parameters = _arg_getter(fn_params)[1:]
+
+                if param_exclusions:
+                    has_exclusions = True
+
+                tobuild_pytest_params.append(
+                    (
+                        parameters,
+                        param_exclusions,
+                        "-".join(
+                            comb_fn(getter(arg)) for getter, comb_fn in fns
+                        ),
+                    )
+                )
+
+        else:
+            for arg in arg_sets:
+                if not isinstance(arg, tuple):
+                    arg = (arg,)
+
+                fn_params, param_exclusions = _filter_exclusions(arg)
+
+                if param_exclusions:
+                    has_exclusions = True
+
+                tobuild_pytest_params.append(
+                    (fn_params, param_exclusions, None)
+                )
+
+        pytest_params = []
+        for parameters, param_exclusions, id_ in tobuild_pytest_params:
+            if has_exclusions:
+                parameters += (param_exclusions,)
+
+            param = pytest.param(*parameters, id=id_)
+            pytest_params.append(param)
+
+        def decorate(fn):
+            if inspect.isclass(fn):
+                if has_exclusions:
+                    raise NotImplementedError(
+                        "exclusions not supported for class level combinations"
+                    )
+                if "_sa_parametrize" not in fn.__dict__:
+                    fn._sa_parametrize = []
+                fn._sa_parametrize.append((argnames, pytest_params))
+                return fn
+            else:
+                _fn_argnames = inspect.getfullargspec(fn).args[1:]
+                if argnames is None:
+                    _argnames = _fn_argnames
+                else:
+                    _argnames = re.split(r", *", argnames)
+
+                if has_exclusions:
+                    existing_exl = sum(
+                        1 for n in _fn_argnames if n.startswith("_exclusions")
+                    )
+                    current_exclusion_name = f"_exclusions_{existing_exl}"
+                    _argnames += [current_exclusion_name]
+
+                    @_pytest_fn_decorator
+                    def check_exclusions(fn, *args, **kw):
+                        _exclusions = args[-1]
+                        if _exclusions:
+                            exlu = exclusions.compound().add(*_exclusions)
+                            fn = exlu(fn)
+                        return fn(*args[:-1], **kw)
+
+                    fn = check_exclusions(
+                        fn, add_positional_parameters=(current_exclusion_name,)
+                    )
+
+                return pytest.mark.parametrize(_argnames, pytest_params)(fn)
+
+        return decorate
+
+    def param_ident(self, *parameters):
+        ident = parameters[0]
+        return pytest.param(*parameters[1:], id=ident)
+
+    def fixture(self, *arg, **kw):
+        from sqlalchemy.testing import config
+        from sqlalchemy.testing import asyncio
+
+        # wrapping pytest.fixture function.  determine if
+        # decorator was called as @fixture or @fixture().
+        if len(arg) > 0 and callable(arg[0]):
+            # was called as @fixture(), we have the function to wrap.
+            fn = arg[0]
+            arg = arg[1:]
+        else:
+            # was called as @fixture, don't have the function yet.
+            fn = None
+
+        # create a pytest.fixture marker.  because the fn is not being
+        # passed, this is always a pytest.FixtureFunctionMarker()
+        # object (or whatever pytest is calling it when you read this)
+        # that is waiting for a function.
+        fixture = pytest.fixture(*arg, **kw)
+
+        # now apply wrappers to the function, including fixture itself
+
+        def wrap(fn):
+            if config.any_async:
+                fn = asyncio._maybe_async_wrapper(fn)
+            # other wrappers may be added here
+
+            # now apply FixtureFunctionMarker
+            fn = fixture(fn)
+
+            return fn
+
+        if fn:
+            return wrap(fn)
+        else:
+            return wrap
+
+    def get_current_test_name(self):
+        return os.environ.get("PYTEST_CURRENT_TEST")
+
+    def async_test(self, fn):
+        from sqlalchemy.testing import asyncio
+
+        @_pytest_fn_decorator
+        def decorate(fn, *args, **kwargs):
+            asyncio._run_coroutine_function(fn, *args, **kwargs)
+
+        return decorate(fn)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/profiling.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/profiling.py
new file mode 100644
index 00000000..0d90947e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/profiling.py
@@ -0,0 +1,324 @@
+# testing/profiling.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Profiling support for unit and performance tests.
+
+These are special purpose profiling methods which operate
+in a more fine-grained way than nose's profiling plugin.
+
+"""
+
+from __future__ import annotations
+
+import collections
+import contextlib
+import os
+import platform
+import pstats
+import re
+import sys
+
+from . import config
+from .util import gc_collect
+from ..util import has_compiled_ext
+
+
+try:
+    import cProfile
+except ImportError:
+    cProfile = None
+
+_profile_stats = None
+"""global ProfileStatsFileInstance.
+
+plugin_base assigns this at the start of all tests.
+
+"""
+
+
+_current_test = None
+"""String id of current test.
+
+plugin_base assigns this at the start of each test using
+_start_current_test.
+
+"""
+
+
+def _start_current_test(id_):
+    global _current_test
+    _current_test = id_
+
+    if _profile_stats.force_write:
+        _profile_stats.reset_count()
+
+
+class ProfileStatsFile:
+    """Store per-platform/fn profiling results in a file.
+
+    There was no json module available when this was written, but now
+    the file format which is very deterministically line oriented is kind of
+    handy in any case for diffs and merges.
+
+    """
+
+    def __init__(self, filename, sort="cumulative", dump=None):
+        self.force_write = (
+            config.options is not None and config.options.force_write_profiles
+        )
+        self.write = self.force_write or (
+            config.options is not None and config.options.write_profiles
+        )
+        self.fname = os.path.abspath(filename)
+        self.short_fname = os.path.split(self.fname)[-1]
+        self.data = collections.defaultdict(
+            lambda: collections.defaultdict(dict)
+        )
+        self.dump = dump
+        self.sort = sort
+        self._read()
+        if self.write:
+            # rewrite for the case where features changed,
+            # etc.
+            self._write()
+
+    @property
+    def platform_key(self):
+        dbapi_key = config.db.name + "_" + config.db.driver
+
+        if config.db.name == "sqlite" and config.db.dialect._is_url_file_db(
+            config.db.url
+        ):
+            dbapi_key += "_file"
+
+        # keep it at 2.7, 3.1, 3.2, etc. for now.
+        py_version = ".".join([str(v) for v in sys.version_info[0:2]])
+
+        platform_tokens = [
+            platform.machine(),
+            platform.system().lower(),
+            platform.python_implementation().lower(),
+            py_version,
+            dbapi_key,
+        ]
+
+        platform_tokens.append("dbapiunicode")
+        _has_cext = has_compiled_ext()
+        platform_tokens.append(_has_cext and "cextensions" or "nocextensions")
+        return "_".join(platform_tokens)
+
+    def has_stats(self):
+        test_key = _current_test
+        return (
+            test_key in self.data and self.platform_key in self.data[test_key]
+        )
+
+    def result(self, callcount):
+        test_key = _current_test
+        per_fn = self.data[test_key]
+        per_platform = per_fn[self.platform_key]
+
+        if "counts" not in per_platform:
+            per_platform["counts"] = counts = []
+        else:
+            counts = per_platform["counts"]
+
+        if "current_count" not in per_platform:
+            per_platform["current_count"] = current_count = 0
+        else:
+            current_count = per_platform["current_count"]
+
+        has_count = len(counts) > current_count
+
+        if not has_count:
+            counts.append(callcount)
+            if self.write:
+                self._write()
+            result = None
+        else:
+            result = per_platform["lineno"], counts[current_count]
+        per_platform["current_count"] += 1
+        return result
+
+    def reset_count(self):
+        test_key = _current_test
+        # since self.data is a defaultdict, don't access a key
+        # if we don't know it's there first.
+        if test_key not in self.data:
+            return
+        per_fn = self.data[test_key]
+        if self.platform_key not in per_fn:
+            return
+        per_platform = per_fn[self.platform_key]
+        if "counts" in per_platform:
+            per_platform["counts"][:] = []
+
+    def replace(self, callcount):
+        test_key = _current_test
+        per_fn = self.data[test_key]
+        per_platform = per_fn[self.platform_key]
+        counts = per_platform["counts"]
+        current_count = per_platform["current_count"]
+        if current_count < len(counts):
+            counts[current_count - 1] = callcount
+        else:
+            counts[-1] = callcount
+        if self.write:
+            self._write()
+
+    def _header(self):
+        return (
+            "# %s\n"
+            "# This file is written out on a per-environment basis.\n"
+            "# For each test in aaa_profiling, the corresponding "
+            "function and \n"
+            "# environment is located within this file.  "
+            "If it doesn't exist,\n"
+            "# the test is skipped.\n"
+            "# If a callcount does exist, it is compared "
+            "to what we received. \n"
+            "# assertions are raised if the counts do not match.\n"
+            "# \n"
+            "# To add a new callcount test, apply the function_call_count \n"
+            "# decorator and re-run the tests using the --write-profiles \n"
+            "# option - this file will be rewritten including the new count.\n"
+            "# \n"
+        ) % (self.fname)
+
+    def _read(self):
+        try:
+            profile_f = open(self.fname)
+        except OSError:
+            return
+        for lineno, line in enumerate(profile_f):
+            line = line.strip()
+            if not line or line.startswith("#"):
+                continue
+
+            test_key, platform_key, counts = line.split()
+            per_fn = self.data[test_key]
+            per_platform = per_fn[platform_key]
+            c = [int(count) for count in counts.split(",")]
+            per_platform["counts"] = c
+            per_platform["lineno"] = lineno + 1
+            per_platform["current_count"] = 0
+        profile_f.close()
+
+    def _write(self):
+        print("Writing profile file %s" % self.fname)
+        profile_f = open(self.fname, "w")
+        profile_f.write(self._header())
+        for test_key in sorted(self.data):
+            per_fn = self.data[test_key]
+            profile_f.write("\n# TEST: %s\n\n" % test_key)
+            for platform_key in sorted(per_fn):
+                per_platform = per_fn[platform_key]
+                c = ",".join(str(count) for count in per_platform["counts"])
+                profile_f.write("%s %s %s\n" % (test_key, platform_key, c))
+        profile_f.close()
+
+
+def function_call_count(variance=0.05, times=1, warmup=0):
+    """Assert a target for a test case's function call count.
+
+    The main purpose of this assertion is to detect changes in
+    callcounts for various functions - the actual number is not as important.
+    Callcounts are stored in a file keyed to Python version and OS platform
+    information.  This file is generated automatically for new tests,
+    and versioned so that unexpected changes in callcounts will be detected.
+
+    """
+
+    # use signature-rewriting decorator function so that pytest fixtures
+    # still work on py27.  In Py3, update_wrapper() alone is good enough,
+    # likely due to the introduction of __signature__.
+
+    from sqlalchemy.util import decorator
+
+    @decorator
+    def wrap(fn, *args, **kw):
+        for warm in range(warmup):
+            fn(*args, **kw)
+
+        timerange = range(times)
+        with count_functions(variance=variance):
+            for time in timerange:
+                rv = fn(*args, **kw)
+            return rv
+
+    return wrap
+
+
+@contextlib.contextmanager
+def count_functions(variance=0.05):
+    if cProfile is None:
+        raise config._skip_test_exception("cProfile is not installed")
+
+    if not _profile_stats.has_stats() and not _profile_stats.write:
+        config.skip_test(
+            "No profiling stats available on this "
+            "platform for this function.  Run tests with "
+            "--write-profiles to add statistics to %s for "
+            "this platform." % _profile_stats.short_fname
+        )
+
+    gc_collect()
+
+    pr = cProfile.Profile()
+    pr.enable()
+    # began = time.time()
+    yield
+    # ended = time.time()
+    pr.disable()
+
+    # s = StringIO()
+    stats = pstats.Stats(pr, stream=sys.stdout)
+
+    # timespent = ended - began
+    callcount = stats.total_calls
+
+    expected = _profile_stats.result(callcount)
+
+    if expected is None:
+        expected_count = None
+    else:
+        line_no, expected_count = expected
+
+    print("Pstats calls: %d Expected %s" % (callcount, expected_count))
+    stats.sort_stats(*re.split(r"[, ]", _profile_stats.sort))
+    stats.print_stats()
+    if _profile_stats.dump:
+        base, ext = os.path.splitext(_profile_stats.dump)
+        test_name = _current_test.split(".")[-1]
+        dumpfile = "%s_%s%s" % (base, test_name, ext or ".profile")
+        stats.dump_stats(dumpfile)
+        print("Dumped stats to file %s" % dumpfile)
+    # stats.print_callers()
+    if _profile_stats.force_write:
+        _profile_stats.replace(callcount)
+    elif expected_count:
+        deviance = int(callcount * variance)
+        failed = abs(callcount - expected_count) > deviance
+
+        if failed:
+            if _profile_stats.write:
+                _profile_stats.replace(callcount)
+            else:
+                raise AssertionError(
+                    "Adjusted function call count %s not within %s%% "
+                    "of expected %s, platform %s. Rerun with "
+                    "--write-profiles to "
+                    "regenerate this callcount."
+                    % (
+                        callcount,
+                        (variance * 100),
+                        expected_count,
+                        _profile_stats.platform_key,
+                    )
+                )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/provision.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/provision.py
new file mode 100644
index 00000000..3afcf119
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/provision.py
@@ -0,0 +1,502 @@
+# testing/provision.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import collections
+import logging
+
+from . import config
+from . import engines
+from . import util
+from .. import exc
+from .. import inspect
+from ..engine import url as sa_url
+from ..sql import ddl
+from ..sql import schema
+
+
+log = logging.getLogger(__name__)
+
+FOLLOWER_IDENT = None
+
+
+class register:
+    def __init__(self, decorator=None):
+        self.fns = {}
+        self.decorator = decorator
+
+    @classmethod
+    def init(cls, fn):
+        return register().for_db("*")(fn)
+
+    @classmethod
+    def init_decorator(cls, decorator):
+        return register(decorator).for_db("*")
+
+    def for_db(self, *dbnames):
+        def decorate(fn):
+            if self.decorator:
+                fn = self.decorator(fn)
+            for dbname in dbnames:
+                self.fns[dbname] = fn
+            return self
+
+        return decorate
+
+    def __call__(self, cfg, *arg, **kw):
+        if isinstance(cfg, str):
+            url = sa_url.make_url(cfg)
+        elif isinstance(cfg, sa_url.URL):
+            url = cfg
+        else:
+            url = cfg.db.url
+        backend = url.get_backend_name()
+        if backend in self.fns:
+            return self.fns[backend](cfg, *arg, **kw)
+        else:
+            return self.fns["*"](cfg, *arg, **kw)
+
+
+def create_follower_db(follower_ident):
+    for cfg in _configs_for_db_operation():
+        log.info("CREATE database %s, URI %r", follower_ident, cfg.db.url)
+        create_db(cfg, cfg.db, follower_ident)
+
+
+def setup_config(db_url, options, file_config, follower_ident):
+    # load the dialect, which should also have it set up its provision
+    # hooks
+
+    dialect = sa_url.make_url(db_url).get_dialect()
+
+    dialect.load_provisioning()
+
+    if follower_ident:
+        db_url = follower_url_from_main(db_url, follower_ident)
+    db_opts = {}
+    update_db_opts(db_url, db_opts, options)
+    db_opts["scope"] = "global"
+    eng = engines.testing_engine(db_url, db_opts)
+    post_configure_engine(db_url, eng, follower_ident)
+    eng.connect().close()
+
+    cfg = config.Config.register(eng, db_opts, options, file_config)
+
+    # a symbolic name that tests can use if they need to disambiguate
+    # names across databases
+    if follower_ident:
+        config.ident = follower_ident
+
+    if follower_ident:
+        configure_follower(cfg, follower_ident)
+    return cfg
+
+
+def drop_follower_db(follower_ident):
+    for cfg in _configs_for_db_operation():
+        log.info("DROP database %s, URI %r", follower_ident, cfg.db.url)
+        drop_db(cfg, cfg.db, follower_ident)
+
+
+def generate_db_urls(db_urls, extra_drivers):
+    """Generate a set of URLs to test given configured URLs plus additional
+    driver names.
+
+    Given:
+
+    .. sourcecode:: text
+
+        --dburi postgresql://db1  \
+        --dburi postgresql://db2  \
+        --dburi postgresql://db2  \
+        --dbdriver=psycopg2 --dbdriver=asyncpg?async_fallback=true
+
+    Noting that the default postgresql driver is psycopg2,  the output
+    would be:
+
+    .. sourcecode:: text
+
+        postgresql+psycopg2://db1
+        postgresql+asyncpg://db1
+        postgresql+psycopg2://db2
+        postgresql+psycopg2://db3
+
+    That is, for the driver in a --dburi, we want to keep that and use that
+    driver for each URL it's part of .   For a driver that is only
+    in --dbdrivers, we want to use it just once for one of the URLs.
+    for a driver that is both coming from --dburi as well as --dbdrivers,
+    we want to keep it in that dburi.
+
+    Driver specific query options can be specified by added them to the
+    driver name. For example, to enable the async fallback option for
+    asyncpg::
+
+    .. sourcecode:: text
+
+        --dburi postgresql://db1  \
+        --dbdriver=asyncpg?async_fallback=true
+
+    """
+    urls = set()
+
+    backend_to_driver_we_already_have = collections.defaultdict(set)
+
+    urls_plus_dialects = [
+        (url_obj, url_obj.get_dialect())
+        for url_obj in [sa_url.make_url(db_url) for db_url in db_urls]
+    ]
+
+    for url_obj, dialect in urls_plus_dialects:
+        # use get_driver_name instead of dialect.driver to account for
+        # "_async" virtual drivers like oracledb and psycopg
+        driver_name = url_obj.get_driver_name()
+        backend_to_driver_we_already_have[dialect.name].add(driver_name)
+
+    backend_to_driver_we_need = {}
+
+    for url_obj, dialect in urls_plus_dialects:
+        backend = dialect.name
+        dialect.load_provisioning()
+
+        if backend not in backend_to_driver_we_need:
+            backend_to_driver_we_need[backend] = extra_per_backend = set(
+                extra_drivers
+            ).difference(backend_to_driver_we_already_have[backend])
+        else:
+            extra_per_backend = backend_to_driver_we_need[backend]
+
+        for driver_url in _generate_driver_urls(url_obj, extra_per_backend):
+            if driver_url in urls:
+                continue
+            urls.add(driver_url)
+            yield driver_url
+
+
+def _generate_driver_urls(url, extra_drivers):
+    main_driver = url.get_driver_name()
+    extra_drivers.discard(main_driver)
+
+    url = generate_driver_url(url, main_driver, "")
+    yield url
+
+    for drv in list(extra_drivers):
+        if "?" in drv:
+            driver_only, query_str = drv.split("?", 1)
+
+        else:
+            driver_only = drv
+            query_str = None
+
+        new_url = generate_driver_url(url, driver_only, query_str)
+        if new_url:
+            extra_drivers.remove(drv)
+
+            yield new_url
+
+
+@register.init
+def generate_driver_url(url, driver, query_str):
+    backend = url.get_backend_name()
+
+    new_url = url.set(
+        drivername="%s+%s" % (backend, driver),
+    )
+    if query_str:
+        new_url = new_url.update_query_string(query_str)
+
+    try:
+        new_url.get_dialect()
+    except exc.NoSuchModuleError:
+        return None
+    else:
+        return new_url
+
+
+def _configs_for_db_operation():
+    hosts = set()
+
+    for cfg in config.Config.all_configs():
+        cfg.db.dispose()
+
+    for cfg in config.Config.all_configs():
+        url = cfg.db.url
+        backend = url.get_backend_name()
+        host_conf = (backend, url.username, url.host, url.database)
+
+        if host_conf not in hosts:
+            yield cfg
+            hosts.add(host_conf)
+
+    for cfg in config.Config.all_configs():
+        cfg.db.dispose()
+
+
+@register.init
+def drop_all_schema_objects_pre_tables(cfg, eng):
+    pass
+
+
+@register.init
+def drop_all_schema_objects_post_tables(cfg, eng):
+    pass
+
+
+def drop_all_schema_objects(cfg, eng):
+    drop_all_schema_objects_pre_tables(cfg, eng)
+
+    drop_views(cfg, eng)
+
+    if config.requirements.materialized_views.enabled:
+        drop_materialized_views(cfg, eng)
+
+    inspector = inspect(eng)
+
+    consider_schemas = (None,)
+    if config.requirements.schemas.enabled_for_config(cfg):
+        consider_schemas += (cfg.test_schema, cfg.test_schema_2)
+    util.drop_all_tables(eng, inspector, consider_schemas=consider_schemas)
+
+    drop_all_schema_objects_post_tables(cfg, eng)
+
+    if config.requirements.sequences.enabled_for_config(cfg):
+        with eng.begin() as conn:
+            for seq in inspector.get_sequence_names():
+                conn.execute(ddl.DropSequence(schema.Sequence(seq)))
+            if config.requirements.schemas.enabled_for_config(cfg):
+                for schema_name in [cfg.test_schema, cfg.test_schema_2]:
+                    for seq in inspector.get_sequence_names(
+                        schema=schema_name
+                    ):
+                        conn.execute(
+                            ddl.DropSequence(
+                                schema.Sequence(seq, schema=schema_name)
+                            )
+                        )
+
+
+def drop_views(cfg, eng):
+    inspector = inspect(eng)
+
+    try:
+        view_names = inspector.get_view_names()
+    except NotImplementedError:
+        pass
+    else:
+        with eng.begin() as conn:
+            for vname in view_names:
+                conn.execute(
+                    ddl._DropView(schema.Table(vname, schema.MetaData()))
+                )
+
+    if config.requirements.schemas.enabled_for_config(cfg):
+        try:
+            view_names = inspector.get_view_names(schema=cfg.test_schema)
+        except NotImplementedError:
+            pass
+        else:
+            with eng.begin() as conn:
+                for vname in view_names:
+                    conn.execute(
+                        ddl._DropView(
+                            schema.Table(
+                                vname,
+                                schema.MetaData(),
+                                schema=cfg.test_schema,
+                            )
+                        )
+                    )
+
+
+def drop_materialized_views(cfg, eng):
+    inspector = inspect(eng)
+
+    mview_names = inspector.get_materialized_view_names()
+
+    with eng.begin() as conn:
+        for vname in mview_names:
+            conn.exec_driver_sql(f"DROP MATERIALIZED VIEW {vname}")
+
+    if config.requirements.schemas.enabled_for_config(cfg):
+        mview_names = inspector.get_materialized_view_names(
+            schema=cfg.test_schema
+        )
+        with eng.begin() as conn:
+            for vname in mview_names:
+                conn.exec_driver_sql(
+                    f"DROP MATERIALIZED VIEW {cfg.test_schema}.{vname}"
+                )
+
+
+@register.init
+def create_db(cfg, eng, ident):
+    """Dynamically create a database for testing.
+
+    Used when a test run will employ multiple processes, e.g., when run
+    via `tox` or `pytest -n4`.
+    """
+    raise NotImplementedError(
+        "no DB creation routine for cfg: %s" % (eng.url,)
+    )
+
+
+@register.init
+def drop_db(cfg, eng, ident):
+    """Drop a database that we dynamically created for testing."""
+    raise NotImplementedError("no DB drop routine for cfg: %s" % (eng.url,))
+
+
+def _adapt_update_db_opts(fn):
+    insp = util.inspect_getfullargspec(fn)
+    if len(insp.args) == 3:
+        return fn
+    else:
+        return lambda db_url, db_opts, _options: fn(db_url, db_opts)
+
+
+@register.init_decorator(_adapt_update_db_opts)
+def update_db_opts(db_url, db_opts, options):
+    """Set database options (db_opts) for a test database that we created."""
+
+
+@register.init
+def post_configure_engine(url, engine, follower_ident):
+    """Perform extra steps after configuring an engine for testing.
+
+    (For the internal dialects, currently only used by sqlite, oracle, mssql)
+    """
+
+
+@register.init
+def follower_url_from_main(url, ident):
+    """Create a connection URL for a dynamically-created test database.
+
+    :param url: the connection URL specified when the test run was invoked
+    :param ident: the pytest-xdist "worker identifier" to be used as the
+                  database name
+    """
+    url = sa_url.make_url(url)
+    return url.set(database=ident)
+
+
+@register.init
+def configure_follower(cfg, ident):
+    """Create dialect-specific config settings for a follower database."""
+    pass
+
+
+@register.init
+def run_reap_dbs(url, ident):
+    """Remove databases that were created during the test process, after the
+    process has ended.
+
+    This is an optional step that is invoked for certain backends that do not
+    reliably release locks on the database as long as a process is still in
+    use. For the internal dialects, this is currently only necessary for
+    mssql and oracle.
+    """
+
+
+def reap_dbs(idents_file):
+    log.info("Reaping databases...")
+
+    urls = collections.defaultdict(set)
+    idents = collections.defaultdict(set)
+    dialects = {}
+
+    with open(idents_file) as file_:
+        for line in file_:
+            line = line.strip()
+            db_name, db_url = line.split(" ")
+            url_obj = sa_url.make_url(db_url)
+            if db_name not in dialects:
+                dialects[db_name] = url_obj.get_dialect()
+                dialects[db_name].load_provisioning()
+            url_key = (url_obj.get_backend_name(), url_obj.host)
+            urls[url_key].add(db_url)
+            idents[url_key].add(db_name)
+
+    for url_key in urls:
+        url = list(urls[url_key])[0]
+        ident = idents[url_key]
+        run_reap_dbs(url, ident)
+
+
+@register.init
+def temp_table_keyword_args(cfg, eng):
+    """Specify keyword arguments for creating a temporary Table.
+
+    Dialect-specific implementations of this method will return the
+    kwargs that are passed to the Table method when creating a temporary
+    table for testing, e.g., in the define_temp_tables method of the
+    ComponentReflectionTest class in suite/test_reflection.py
+    """
+    raise NotImplementedError(
+        "no temp table keyword args routine for cfg: %s" % (eng.url,)
+    )
+
+
+@register.init
+def prepare_for_drop_tables(config, connection):
+    pass
+
+
+@register.init
+def stop_test_class_outside_fixtures(config, db, testcls):
+    pass
+
+
+@register.init
+def get_temp_table_name(cfg, eng, base_name):
+    """Specify table name for creating a temporary Table.
+
+    Dialect-specific implementations of this method will return the
+    name to use when creating a temporary table for testing,
+    e.g., in the define_temp_tables method of the
+    ComponentReflectionTest class in suite/test_reflection.py
+
+    Default to just the base name since that's what most dialects will
+    use. The mssql dialect's implementation will need a "#" prepended.
+    """
+    return base_name
+
+
+@register.init
+def set_default_schema_on_connection(cfg, dbapi_connection, schema_name):
+    raise NotImplementedError(
+        "backend does not implement a schema name set function: %s"
+        % (cfg.db.url,)
+    )
+
+
+@register.init
+def upsert(
+    cfg, table, returning, *, set_lambda=None, sort_by_parameter_order=False
+):
+    """return the backends insert..on conflict / on dupe etc. construct.
+
+    while we should add a backend-neutral upsert construct as well, such as
+    insert().upsert(), it's important that we continue to test the
+    backend-specific insert() constructs since if we do implement
+    insert().upsert(), that would be using a different codepath for the things
+    we need to test like insertmanyvalues, etc.
+
+    """
+    raise NotImplementedError(
+        f"backend does not include an upsert implementation: {cfg.db.url}"
+    )
+
+
+@register.init
+def normalize_sequence(cfg, sequence):
+    """Normalize sequence parameters for dialect that don't start with 1
+    by default.
+
+    The default implementation does nothing
+    """
+    return sequence
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/requirements.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/requirements.py
new file mode 100644
index 00000000..bbf56a05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/requirements.py
@@ -0,0 +1,1828 @@
+# testing/requirements.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+"""Global database feature support policy.
+
+Provides decorators to mark tests requiring specific feature support from the
+target database.
+
+External dialect test suites should subclass SuiteRequirements
+to provide specific inclusion/exclusions.
+
+"""
+
+from __future__ import annotations
+
+import platform
+
+from . import asyncio as _test_asyncio
+from . import exclusions
+from .exclusions import only_on
+from .. import create_engine
+from .. import util
+from ..pool import QueuePool
+
+
+class Requirements:
+    pass
+
+
+class SuiteRequirements(Requirements):
+    @property
+    def create_table(self):
+        """target platform can emit basic CreateTable DDL."""
+
+        return exclusions.open()
+
+    @property
+    def drop_table(self):
+        """target platform can emit basic DropTable DDL."""
+
+        return exclusions.open()
+
+    @property
+    def table_ddl_if_exists(self):
+        """target platform supports IF NOT EXISTS / IF EXISTS for tables."""
+
+        return exclusions.closed()
+
+    @property
+    def index_ddl_if_exists(self):
+        """target platform supports IF NOT EXISTS / IF EXISTS for indexes."""
+
+        return exclusions.closed()
+
+    @property
+    def uuid_data_type(self):
+        """Return databases that support the UUID datatype."""
+
+        return exclusions.closed()
+
+    @property
+    def foreign_keys(self):
+        """Target database must support foreign keys."""
+
+        return exclusions.open()
+
+    @property
+    def foreign_keys_reflect_as_index(self):
+        """Target database creates an index that's reflected for
+        foreign keys."""
+
+        return exclusions.closed()
+
+    @property
+    def unique_index_reflect_as_unique_constraints(self):
+        """Target database reflects unique indexes as unique constrains."""
+
+        return exclusions.closed()
+
+    @property
+    def unique_constraints_reflect_as_index(self):
+        """Target database reflects unique constraints as indexes."""
+
+        return exclusions.closed()
+
+    @property
+    def table_value_constructor(self):
+        """Database / dialect supports a query like:
+
+        .. sourcecode:: sql
+
+             SELECT * FROM VALUES ( (c1, c2), (c1, c2), ...)
+             AS some_table(col1, col2)
+
+        SQLAlchemy generates this with the :func:`_sql.values` function.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def standard_cursor_sql(self):
+        """Target database passes SQL-92 style statements to cursor.execute()
+        when a statement like select() or insert() is run.
+
+        A very small portion of dialect-level tests will ensure that certain
+        conditions are present in SQL strings, and these tests use very basic
+        SQL that will work on any SQL-like platform in order to assert results.
+
+        It's normally a given for any pep-249 DBAPI that a statement like
+        "SELECT id, name FROM table WHERE some_table.id=5" will work.
+        However, there are dialects that don't actually produce SQL Strings
+        and instead may work with symbolic objects instead, or dialects that
+        aren't working with SQL, so for those this requirement can be marked
+        as excluded.
+
+        """
+
+        return exclusions.open()
+
+    @property
+    def on_update_cascade(self):
+        """target database must support ON UPDATE..CASCADE behavior in
+        foreign keys."""
+
+        return exclusions.open()
+
+    @property
+    def non_updating_cascade(self):
+        """target database must *not* support ON UPDATE..CASCADE behavior in
+        foreign keys."""
+        return exclusions.closed()
+
+    @property
+    def deferrable_fks(self):
+        return exclusions.closed()
+
+    @property
+    def on_update_or_deferrable_fks(self):
+        # TODO: exclusions should be composable,
+        # somehow only_if([x, y]) isn't working here, negation/conjunctions
+        # getting confused.
+        return exclusions.only_if(
+            lambda: self.on_update_cascade.enabled
+            or self.deferrable_fks.enabled
+        )
+
+    @property
+    def queue_pool(self):
+        """target database is using QueuePool"""
+
+        def go(config):
+            return isinstance(config.db.pool, QueuePool)
+
+        return exclusions.only_if(go)
+
+    @property
+    def self_referential_foreign_keys(self):
+        """Target database must support self-referential foreign keys."""
+
+        return exclusions.open()
+
+    @property
+    def foreign_key_ddl(self):
+        """Target database must support the DDL phrases for FOREIGN KEY."""
+
+        return exclusions.open()
+
+    @property
+    def named_constraints(self):
+        """target database must support names for constraints."""
+
+        return exclusions.open()
+
+    @property
+    def implicitly_named_constraints(self):
+        """target database must apply names to unnamed constraints."""
+
+        return exclusions.open()
+
+    @property
+    def unusual_column_name_characters(self):
+        """target database allows column names that have unusual characters
+        in them, such as dots, spaces, slashes, or percent signs.
+
+        The column names are as always in such a case quoted, however the
+        DB still needs to support those characters in the name somehow.
+
+        """
+        return exclusions.open()
+
+    @property
+    def subqueries(self):
+        """Target database must support subqueries."""
+
+        return exclusions.open()
+
+    @property
+    def offset(self):
+        """target database can render OFFSET, or an equivalent, in a
+        SELECT.
+        """
+
+        return exclusions.open()
+
+    @property
+    def bound_limit_offset(self):
+        """target database can render LIMIT and/or OFFSET using a bound
+        parameter
+        """
+
+        return exclusions.open()
+
+    @property
+    def sql_expression_limit_offset(self):
+        """target database can render LIMIT and/or OFFSET with a complete
+        SQL expression, such as one that uses the addition operator.
+        parameter
+        """
+
+        return exclusions.open()
+
+    @property
+    def parens_in_union_contained_select_w_limit_offset(self):
+        """Target database must support parenthesized SELECT in UNION
+        when LIMIT/OFFSET is specifically present.
+
+        E.g. (SELECT ...) UNION (SELECT ..)
+
+        This is known to fail on SQLite.
+
+        """
+        return exclusions.open()
+
+    @property
+    def parens_in_union_contained_select_wo_limit_offset(self):
+        """Target database must support parenthesized SELECT in UNION
+        when OFFSET/LIMIT is specifically not present.
+
+        E.g. (SELECT ... LIMIT ..) UNION (SELECT .. OFFSET ..)
+
+        This is known to fail on SQLite.  It also fails on Oracle
+        because without LIMIT/OFFSET, there is currently no step that
+        creates an additional subquery.
+
+        """
+        return exclusions.open()
+
+    @property
+    def boolean_col_expressions(self):
+        """Target database must support boolean expressions as columns"""
+
+        return exclusions.closed()
+
+    @property
+    def nullable_booleans(self):
+        """Target database allows boolean columns to store NULL."""
+
+        return exclusions.open()
+
+    @property
+    def nullsordering(self):
+        """Target backends that support nulls ordering."""
+
+        return exclusions.closed()
+
+    @property
+    def standalone_binds(self):
+        """target database/driver supports bound parameters as column
+        expressions without being in the context of a typed column.
+        """
+        return exclusions.open()
+
+    @property
+    def standalone_null_binds_whereclause(self):
+        """target database/driver supports bound parameters with NULL in the
+        WHERE clause, in situations where it has to be typed.
+
+        """
+        return exclusions.open()
+
+    @property
+    def intersect(self):
+        """Target database must support INTERSECT or equivalent."""
+        return exclusions.closed()
+
+    @property
+    def except_(self):
+        """Target database must support EXCEPT or equivalent (i.e. MINUS)."""
+        return exclusions.closed()
+
+    @property
+    def window_functions(self):
+        """Target database must support window functions."""
+        return exclusions.closed()
+
+    @property
+    def ctes(self):
+        """Target database supports CTEs"""
+
+        return exclusions.closed()
+
+    @property
+    def ctes_with_update_delete(self):
+        """target database supports CTES that ride on top of a normal UPDATE
+        or DELETE statement which refers to the CTE in a correlated subquery.
+
+        """
+
+        return exclusions.closed()
+
+    @property
+    def ctes_on_dml(self):
+        """target database supports CTES which consist of INSERT, UPDATE
+        or DELETE *within* the CTE, e.g. WITH x AS (UPDATE....)"""
+
+        return exclusions.closed()
+
+    @property
+    def autoincrement_insert(self):
+        """target platform generates new surrogate integer primary key values
+        when insert() is executed, excluding the pk column."""
+
+        return exclusions.open()
+
+    @property
+    def fetch_rows_post_commit(self):
+        """target platform will allow cursor.fetchone() to proceed after a
+        COMMIT.
+
+        Typically this refers to an INSERT statement with RETURNING which
+        is invoked within "autocommit".   If the row can be returned
+        after the autocommit, then this rule can be open.
+
+        """
+
+        return exclusions.open()
+
+    @property
+    def group_by_complex_expression(self):
+        """target platform supports SQL expressions in GROUP BY
+
+        e.g.
+
+        SELECT x + y AS somelabel FROM table GROUP BY x + y
+
+        """
+
+        return exclusions.open()
+
+    @property
+    def sane_rowcount(self):
+        return exclusions.skip_if(
+            lambda config: not config.db.dialect.supports_sane_rowcount,
+            "driver doesn't support 'sane' rowcount",
+        )
+
+    @property
+    def sane_multi_rowcount(self):
+        return exclusions.fails_if(
+            lambda config: not config.db.dialect.supports_sane_multi_rowcount,
+            "driver %(driver)s %(doesnt_support)s 'sane' multi row count",
+        )
+
+    @property
+    def sane_rowcount_w_returning(self):
+        return exclusions.fails_if(
+            lambda config: not (
+                config.db.dialect.supports_sane_rowcount_returning
+            ),
+            "driver doesn't support 'sane' rowcount when returning is on",
+        )
+
+    @property
+    def empty_inserts(self):
+        """target platform supports INSERT with no values, i.e.
+        INSERT DEFAULT VALUES or equivalent."""
+
+        return exclusions.only_if(
+            lambda config: config.db.dialect.supports_empty_insert
+            or config.db.dialect.supports_default_values
+            or config.db.dialect.supports_default_metavalue,
+            "empty inserts not supported",
+        )
+
+    @property
+    def empty_inserts_executemany(self):
+        """target platform supports INSERT with no values, i.e.
+        INSERT DEFAULT VALUES or equivalent, within executemany()"""
+
+        return self.empty_inserts
+
+    @property
+    def insert_from_select(self):
+        """target platform supports INSERT from a SELECT."""
+
+        return exclusions.open()
+
+    @property
+    def delete_returning(self):
+        """target platform supports DELETE ... RETURNING."""
+
+        return exclusions.only_if(
+            lambda config: config.db.dialect.delete_returning,
+            "%(database)s %(does_support)s 'DELETE ... RETURNING'",
+        )
+
+    @property
+    def insert_returning(self):
+        """target platform supports INSERT ... RETURNING."""
+
+        return exclusions.only_if(
+            lambda config: config.db.dialect.insert_returning,
+            "%(database)s %(does_support)s 'INSERT ... RETURNING'",
+        )
+
+    @property
+    def update_returning(self):
+        """target platform supports UPDATE ... RETURNING."""
+
+        return exclusions.only_if(
+            lambda config: config.db.dialect.update_returning,
+            "%(database)s %(does_support)s 'UPDATE ... RETURNING'",
+        )
+
+    @property
+    def insert_executemany_returning(self):
+        """target platform supports RETURNING when INSERT is used with
+        executemany(), e.g. multiple parameter sets, indicating
+        as many rows come back as do parameter sets were passed.
+
+        """
+
+        return exclusions.only_if(
+            lambda config: config.db.dialect.insert_executemany_returning,
+            "%(database)s %(does_support)s 'RETURNING of "
+            "multiple rows with INSERT executemany'",
+        )
+
+    @property
+    def insertmanyvalues(self):
+        return exclusions.only_if(
+            lambda config: config.db.dialect.supports_multivalues_insert
+            and config.db.dialect.insert_returning
+            and config.db.dialect.use_insertmanyvalues,
+            "%(database)s %(does_support)s 'insertmanyvalues functionality",
+        )
+
+    @property
+    def tuple_in(self):
+        """Target platform supports the syntax
+        "(x, y) IN ((x1, y1), (x2, y2), ...)"
+        """
+
+        return exclusions.closed()
+
+    @property
+    def tuple_in_w_empty(self):
+        """Target platform tuple IN w/ empty set"""
+        return self.tuple_in
+
+    @property
+    def duplicate_names_in_cursor_description(self):
+        """target platform supports a SELECT statement that has
+        the same name repeated more than once in the columns list."""
+
+        return exclusions.open()
+
+    @property
+    def denormalized_names(self):
+        """Target database must have 'denormalized', i.e.
+        UPPERCASE as case insensitive names."""
+
+        return exclusions.skip_if(
+            lambda config: not config.db.dialect.requires_name_normalize,
+            "Backend does not require denormalized names.",
+        )
+
+    @property
+    def multivalues_inserts(self):
+        """target database must support multiple VALUES clauses in an
+        INSERT statement."""
+
+        return exclusions.skip_if(
+            lambda config: not config.db.dialect.supports_multivalues_insert,
+            "Backend does not support multirow inserts.",
+        )
+
+    @property
+    def implements_get_lastrowid(self):
+        """target dialect implements the executioncontext.get_lastrowid()
+        method without reliance on RETURNING.
+
+        """
+        return exclusions.open()
+
+    @property
+    def arraysize(self):
+        """dialect includes the required pep-249 attribute
+        ``cursor.arraysize``"""
+
+        return exclusions.open()
+
+    @property
+    def emulated_lastrowid(self):
+        """target dialect retrieves cursor.lastrowid, or fetches
+        from a database-side function after an insert() construct executes,
+        within the get_lastrowid() method.
+
+        Only dialects that "pre-execute", or need RETURNING to get last
+        inserted id, would return closed/fail/skip for this.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def emulated_lastrowid_even_with_sequences(self):
+        """target dialect retrieves cursor.lastrowid or an equivalent
+        after an insert() construct executes, even if the table has a
+        Sequence on it.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def dbapi_lastrowid(self):
+        """target platform includes a 'lastrowid' accessor on the DBAPI
+        cursor object.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def views(self):
+        """Target database must support VIEWs."""
+
+        return exclusions.closed()
+
+    @property
+    def schemas(self):
+        """Target database must support external schemas, and have one
+        named 'test_schema'."""
+
+        return only_on(lambda config: config.db.dialect.supports_schemas)
+
+    @property
+    def cross_schema_fk_reflection(self):
+        """target system must support reflection of inter-schema
+        foreign keys"""
+        return exclusions.closed()
+
+    @property
+    def foreign_key_constraint_name_reflection(self):
+        """Target supports reflection of FOREIGN KEY constraints and
+        will return the name of the constraint that was used in the
+        "CONSTRAINT <name> FOREIGN KEY" DDL.
+
+        MySQL prior to version 8 and MariaDB prior to version 10.5
+        don't support this.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def implicit_default_schema(self):
+        """target system has a strong concept of 'default' schema that can
+        be referred to implicitly.
+
+        basically, PostgreSQL.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def default_schema_name_switch(self):
+        """target dialect implements provisioning module including
+        set_default_schema_on_connection"""
+
+        return exclusions.closed()
+
+    @property
+    def server_side_cursors(self):
+        """Target dialect must support server side cursors."""
+
+        return exclusions.only_if(
+            [lambda config: config.db.dialect.supports_server_side_cursors],
+            "no server side cursors support",
+        )
+
+    @property
+    def sequences(self):
+        """Target database must support SEQUENCEs."""
+
+        return exclusions.only_if(
+            [lambda config: config.db.dialect.supports_sequences],
+            "no sequence support",
+        )
+
+    @property
+    def no_sequences(self):
+        """the opposite of "sequences", DB does not support sequences at
+        all."""
+
+        return exclusions.NotPredicate(self.sequences)
+
+    @property
+    def sequences_optional(self):
+        """Target database supports sequences, but also optionally
+        as a means of generating new PK values."""
+
+        return exclusions.only_if(
+            [
+                lambda config: config.db.dialect.supports_sequences
+                and config.db.dialect.sequences_optional
+            ],
+            "no sequence support, or sequences not optional",
+        )
+
+    @property
+    def supports_lastrowid(self):
+        """target database / driver supports cursor.lastrowid as a means
+        of retrieving the last inserted primary key value.
+
+        note that if the target DB supports sequences also, this is still
+        assumed to work.  This is a new use case brought on by MariaDB 10.3.
+
+        """
+        return exclusions.only_if(
+            [lambda config: config.db.dialect.postfetch_lastrowid]
+        )
+
+    @property
+    def no_lastrowid_support(self):
+        """the opposite of supports_lastrowid"""
+        return exclusions.only_if(
+            [lambda config: not config.db.dialect.postfetch_lastrowid]
+        )
+
+    @property
+    def reflects_pk_names(self):
+        return exclusions.closed()
+
+    @property
+    def table_reflection(self):
+        """target database has general support for table reflection"""
+        return exclusions.open()
+
+    @property
+    def reflect_tables_no_columns(self):
+        """target database supports creation and reflection of tables with no
+        columns, or at least tables that seem to have no columns."""
+
+        return exclusions.closed()
+
+    @property
+    def comment_reflection(self):
+        """Indicates if the database support table comment reflection"""
+        return exclusions.closed()
+
+    @property
+    def comment_reflection_full_unicode(self):
+        """Indicates if the database support table comment reflection in the
+        full unicode range, including emoji etc.
+        """
+        return exclusions.closed()
+
+    @property
+    def constraint_comment_reflection(self):
+        """indicates if the database support comments on constraints
+        and their reflection"""
+        return exclusions.closed()
+
+    @property
+    def view_column_reflection(self):
+        """target database must support retrieval of the columns in a view,
+        similarly to how a table is inspected.
+
+        This does not include the full CREATE VIEW definition.
+
+        """
+        return self.views
+
+    @property
+    def view_reflection(self):
+        """target database must support inspection of the full CREATE VIEW
+        definition."""
+        return self.views
+
+    @property
+    def schema_reflection(self):
+        return self.schemas
+
+    @property
+    def schema_create_delete(self):
+        """target database supports schema create and dropped with
+        'CREATE SCHEMA' and 'DROP SCHEMA'"""
+        return exclusions.closed()
+
+    @property
+    def primary_key_constraint_reflection(self):
+        return exclusions.open()
+
+    @property
+    def foreign_key_constraint_reflection(self):
+        return exclusions.open()
+
+    @property
+    def foreign_key_constraint_option_reflection_ondelete(self):
+        return exclusions.closed()
+
+    @property
+    def fk_constraint_option_reflection_ondelete_restrict(self):
+        return exclusions.closed()
+
+    @property
+    def fk_constraint_option_reflection_ondelete_noaction(self):
+        return exclusions.closed()
+
+    @property
+    def foreign_key_constraint_option_reflection_onupdate(self):
+        return exclusions.closed()
+
+    @property
+    def fk_constraint_option_reflection_onupdate_restrict(self):
+        return exclusions.closed()
+
+    @property
+    def temp_table_reflection(self):
+        return exclusions.open()
+
+    @property
+    def temp_table_reflect_indexes(self):
+        return self.temp_table_reflection
+
+    @property
+    def temp_table_names(self):
+        """target dialect supports listing of temporary table names"""
+        return exclusions.closed()
+
+    @property
+    def has_temp_table(self):
+        """target dialect supports checking a single temp table name"""
+        return exclusions.closed()
+
+    @property
+    def temporary_tables(self):
+        """target database supports temporary tables"""
+        return exclusions.open()
+
+    @property
+    def temporary_views(self):
+        """target database supports temporary views"""
+        return exclusions.closed()
+
+    @property
+    def index_reflection(self):
+        return exclusions.open()
+
+    @property
+    def index_reflects_included_columns(self):
+        return exclusions.closed()
+
+    @property
+    def indexes_with_ascdesc(self):
+        """target database supports CREATE INDEX with per-column ASC/DESC."""
+        return exclusions.open()
+
+    @property
+    def reflect_indexes_with_ascdesc(self):
+        """target database supports reflecting INDEX with per-column
+        ASC/DESC."""
+        return exclusions.open()
+
+    @property
+    def reflect_indexes_with_ascdesc_as_expression(self):
+        """target database supports reflecting INDEX with per-column
+        ASC/DESC but reflects them as expressions (like oracle)."""
+        return exclusions.closed()
+
+    @property
+    def indexes_with_expressions(self):
+        """target database supports CREATE INDEX against SQL expressions."""
+        return exclusions.closed()
+
+    @property
+    def reflect_indexes_with_expressions(self):
+        """target database supports reflection of indexes with
+        SQL expressions."""
+        return exclusions.closed()
+
+    @property
+    def unique_constraint_reflection(self):
+        """target dialect supports reflection of unique constraints"""
+        return exclusions.open()
+
+    @property
+    def inline_check_constraint_reflection(self):
+        """target dialect supports reflection of inline check constraints"""
+        return exclusions.closed()
+
+    @property
+    def check_constraint_reflection(self):
+        """target dialect supports reflection of check constraints"""
+        return exclusions.closed()
+
+    @property
+    def duplicate_key_raises_integrity_error(self):
+        """target dialect raises IntegrityError when reporting an INSERT
+        with a primary key violation.  (hint: it should)
+
+        """
+        return exclusions.open()
+
+    @property
+    def unbounded_varchar(self):
+        """Target database must support VARCHAR with no length"""
+
+        return exclusions.open()
+
+    @property
+    def unicode_data_no_special_types(self):
+        """Target database/dialect can receive / deliver / compare data with
+        non-ASCII characters in plain VARCHAR, TEXT columns, without the need
+        for special "national" datatypes like NVARCHAR or similar.
+
+        """
+        return exclusions.open()
+
+    @property
+    def unicode_data(self):
+        """Target database/dialect must support Python unicode objects with
+        non-ASCII characters represented, delivered as bound parameters
+        as well as in result rows.
+
+        """
+        return exclusions.open()
+
+    @property
+    def unicode_ddl(self):
+        """Target driver must support some degree of non-ascii symbol
+        names.
+        """
+        return exclusions.closed()
+
+    @property
+    def symbol_names_w_double_quote(self):
+        """Target driver can create tables with a name like 'some " table'"""
+        return exclusions.open()
+
+    @property
+    def datetime_interval(self):
+        """target dialect supports rendering of a datetime.timedelta as a
+        literal string, e.g. via the TypeEngine.literal_processor() method.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def datetime_literals(self):
+        """target dialect supports rendering of a date, time, or datetime as a
+        literal string, e.g. via the TypeEngine.literal_processor() method.
+
+        """
+
+        return exclusions.closed()
+
+    @property
+    def datetime(self):
+        """target dialect supports representation of Python
+        datetime.datetime() objects."""
+
+        return exclusions.open()
+
+    @property
+    def datetime_timezone(self):
+        """target dialect supports representation of Python
+        datetime.datetime() with tzinfo with DateTime(timezone=True)."""
+
+        return exclusions.closed()
+
+    @property
+    def time_timezone(self):
+        """target dialect supports representation of Python
+        datetime.time() with tzinfo with Time(timezone=True)."""
+
+        return exclusions.closed()
+
+    @property
+    def date_implicit_bound(self):
+        """target dialect when given a date object will bind it such
+        that the database server knows the object is a date, and not
+        a plain string.
+
+        """
+        return exclusions.open()
+
+    @property
+    def time_implicit_bound(self):
+        """target dialect when given a time object will bind it such
+        that the database server knows the object is a time, and not
+        a plain string.
+
+        """
+        return exclusions.open()
+
+    @property
+    def datetime_implicit_bound(self):
+        """target dialect when given a datetime object will bind it such
+        that the database server knows the object is a datetime, and not
+        a plain string.
+
+        """
+        return exclusions.open()
+
+    @property
+    def datetime_microseconds(self):
+        """target dialect supports representation of Python
+        datetime.datetime() with microsecond objects."""
+
+        return exclusions.open()
+
+    @property
+    def timestamp_microseconds(self):
+        """target dialect supports representation of Python
+        datetime.datetime() with microsecond objects but only
+        if TIMESTAMP is used."""
+        return exclusions.closed()
+
+    @property
+    def timestamp_microseconds_implicit_bound(self):
+        """target dialect when given a datetime object which also includes
+        a microseconds portion when using the TIMESTAMP data type
+        will bind it such that the database server knows
+        the object is a datetime with microseconds, and not a plain string.
+
+        """
+        return self.timestamp_microseconds
+
+    @property
+    def datetime_historic(self):
+        """target dialect supports representation of Python
+        datetime.datetime() objects with historic (pre 1970) values."""
+
+        return exclusions.closed()
+
+    @property
+    def date(self):
+        """target dialect supports representation of Python
+        datetime.date() objects."""
+
+        return exclusions.open()
+
+    @property
+    def date_coerces_from_datetime(self):
+        """target dialect accepts a datetime object as the target
+        of a date column."""
+
+        return exclusions.open()
+
+    @property
+    def date_historic(self):
+        """target dialect supports representation of Python
+        datetime.datetime() objects with historic (pre 1970) values."""
+
+        return exclusions.closed()
+
+    @property
+    def time(self):
+        """target dialect supports representation of Python
+        datetime.time() objects."""
+
+        return exclusions.open()
+
+    @property
+    def time_microseconds(self):
+        """target dialect supports representation of Python
+        datetime.time() with microsecond objects."""
+
+        return exclusions.open()
+
+    @property
+    def binary_comparisons(self):
+        """target database/driver can allow BLOB/BINARY fields to be compared
+        against a bound parameter value.
+        """
+
+        return exclusions.open()
+
+    @property
+    def binary_literals(self):
+        """target backend supports simple binary literals, e.g. an
+        expression like:
+
+        .. sourcecode:: sql
+
+            SELECT CAST('foo' AS BINARY)
+
+        Where ``BINARY`` is the type emitted from :class:`.LargeBinary`,
+        e.g. it could be ``BLOB`` or similar.
+
+        Basically fails on Oracle.
+
+        """
+
+        return exclusions.open()
+
+    @property
+    def autocommit(self):
+        """target dialect supports 'AUTOCOMMIT' as an isolation_level"""
+        return exclusions.closed()
+
+    @property
+    def isolation_level(self):
+        """target dialect supports general isolation level settings.
+
+        Note that this requirement, when enabled, also requires that
+        the get_isolation_levels() method be implemented.
+
+        """
+        return exclusions.closed()
+
+    def get_isolation_levels(self, config):
+        """Return a structure of supported isolation levels for the current
+        testing dialect.
+
+        The structure indicates to the testing suite what the expected
+        "default" isolation should be, as well as the other values that
+        are accepted.  The dictionary has two keys, "default" and "supported".
+        The "supported" key refers to a list of all supported levels and
+        it should include AUTOCOMMIT if the dialect supports it.
+
+        If the :meth:`.DefaultRequirements.isolation_level` requirement is
+        not open, then this method has no return value.
+
+        E.g.::
+
+            >>> testing.requirements.get_isolation_levels()
+            {
+                "default": "READ_COMMITTED",
+                "supported": [
+                    "SERIALIZABLE", "READ UNCOMMITTED",
+                    "READ COMMITTED", "REPEATABLE READ",
+                    "AUTOCOMMIT"
+                ]
+            }
+        """
+        with config.db.connect() as conn:
+            try:
+                supported = conn.dialect.get_isolation_level_values(
+                    conn.connection.dbapi_connection
+                )
+            except NotImplementedError:
+                return None
+            else:
+                return {
+                    "default": conn.dialect.default_isolation_level,
+                    "supported": supported,
+                }
+
+    @property
+    def get_isolation_level_values(self):
+        """target dialect supports the
+        :meth:`_engine.Dialect.get_isolation_level_values`
+        method added in SQLAlchemy 2.0.
+
+        """
+
+        def go(config):
+            with config.db.connect() as conn:
+                try:
+                    conn.dialect.get_isolation_level_values(
+                        conn.connection.dbapi_connection
+                    )
+                except NotImplementedError:
+                    return False
+                else:
+                    return True
+
+        return exclusions.only_if(go)
+
+    @property
+    def dialect_level_isolation_level_param(self):
+        """test that the dialect allows the 'isolation_level' argument
+        to be handled by DefaultDialect"""
+
+        def go(config):
+            try:
+                e = create_engine(
+                    config.db.url, isolation_level="READ COMMITTED"
+                )
+            except:
+                return False
+            else:
+                return (
+                    e.dialect._on_connect_isolation_level == "READ COMMITTED"
+                )
+
+        return exclusions.only_if(go)
+
+    @property
+    def array_type(self):
+        """Target platform implements a native ARRAY type"""
+        return exclusions.closed()
+
+    @property
+    def json_type(self):
+        """target platform implements a native JSON type."""
+
+        return exclusions.closed()
+
+    @property
+    def json_array_indexes(self):
+        """target platform supports numeric array indexes
+        within a JSON structure"""
+
+        return self.json_type
+
+    @property
+    def json_index_supplementary_unicode_element(self):
+        return exclusions.open()
+
+    @property
+    def legacy_unconditional_json_extract(self):
+        """Backend has a JSON_EXTRACT or similar function that returns a
+        valid JSON string in all cases.
+
+        Used to test a legacy feature and is not needed.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def precision_numerics_general(self):
+        """target backend has general support for moderately high-precision
+        numerics."""
+        return exclusions.open()
+
+    @property
+    def precision_numerics_enotation_small(self):
+        """target backend supports Decimal() objects using E notation
+        to represent very small values."""
+        return exclusions.closed()
+
+    @property
+    def precision_numerics_enotation_large(self):
+        """target backend supports Decimal() objects using E notation
+        to represent very large values."""
+        return exclusions.open()
+
+    @property
+    def precision_numerics_many_significant_digits(self):
+        """target backend supports values with many digits on both sides,
+        such as 319438950232418390.273596, 87673.594069654243
+
+        """
+        return exclusions.closed()
+
+    @property
+    def cast_precision_numerics_many_significant_digits(self):
+        """same as precision_numerics_many_significant_digits but within the
+        context of a CAST statement (hello MySQL)
+
+        """
+        return self.precision_numerics_many_significant_digits
+
+    @property
+    def implicit_decimal_binds(self):
+        """target backend will return a selected Decimal as a Decimal, not
+        a string.
+
+        e.g.::
+
+            expr = decimal.Decimal("15.7563")
+
+            value = e.scalar(select(literal(expr)))
+
+            assert value == expr
+
+        See :ticket:`4036`
+
+        """
+
+        return exclusions.open()
+
+    @property
+    def numeric_received_as_decimal_untyped(self):
+        """target backend will return result columns that are explicitly
+        against NUMERIC or similar precision-numeric datatypes (not including
+        FLOAT or INT types) as Python Decimal objects, and not as floats
+        or ints, including when no SQLAlchemy-side typing information is
+        associated with the statement (e.g. such as a raw SQL string).
+
+        This should be enabled if either the DBAPI itself returns Decimal
+        objects, or if the dialect has set up DBAPI-specific return type
+        handlers such that Decimal objects come back automatically.
+
+        """
+        return exclusions.open()
+
+    @property
+    def nested_aggregates(self):
+        """target database can select an aggregate from a subquery that's
+        also using an aggregate
+
+        """
+        return exclusions.open()
+
+    @property
+    def recursive_fk_cascade(self):
+        """target database must support ON DELETE CASCADE on a self-referential
+        foreign key
+
+        """
+        return exclusions.open()
+
+    @property
+    def precision_numerics_retains_significant_digits(self):
+        """A precision numeric type will return empty significant digits,
+        i.e. a value such as 10.000 will come back in Decimal form with
+        the .000 maintained."""
+
+        return exclusions.closed()
+
+    @property
+    def infinity_floats(self):
+        """The Float type can persist and load float('inf'), float('-inf')."""
+
+        return exclusions.closed()
+
+    @property
+    def float_or_double_precision_behaves_generically(self):
+        return exclusions.closed()
+
+    @property
+    def precision_generic_float_type(self):
+        """target backend will return native floating point numbers with at
+        least seven decimal places when using the generic Float type.
+
+        """
+        return exclusions.open()
+
+    @property
+    def literal_float_coercion(self):
+        """target backend will return the exact float value 15.7563
+        with only four significant digits from this statement:
+
+        SELECT :param
+
+        where :param is the Python float 15.7563
+
+        i.e. it does not return 15.75629997253418
+
+        """
+        return exclusions.open()
+
+    @property
+    def floats_to_four_decimals(self):
+        """target backend can return a floating-point number with four
+        significant digits (such as 15.7563) accurately
+        (i.e. without FP inaccuracies, such as 15.75629997253418).
+
+        """
+        return exclusions.open()
+
+    @property
+    def fetch_null_from_numeric(self):
+        """target backend doesn't crash when you try to select a NUMERIC
+        value that has a value of NULL.
+
+        Added to support Pyodbc bug #351.
+        """
+
+        return exclusions.open()
+
+    @property
+    def float_is_numeric(self):
+        """target backend uses Numeric for Float/Dual"""
+
+        return exclusions.open()
+
+    @property
+    def text_type(self):
+        """Target database must support an unbounded Text() "
+        "type such as TEXT or CLOB"""
+
+        return exclusions.open()
+
+    @property
+    def empty_strings_varchar(self):
+        """target database can persist/return an empty string with a
+        varchar.
+
+        """
+        return exclusions.open()
+
+    @property
+    def empty_strings_text(self):
+        """target database can persist/return an empty string with an
+        unbounded text."""
+
+        return exclusions.open()
+
+    @property
+    def expressions_against_unbounded_text(self):
+        """target database supports use of an unbounded textual field in a
+        WHERE clause."""
+
+        return exclusions.open()
+
+    @property
+    def selectone(self):
+        """target driver must support the literal statement 'select 1'"""
+        return exclusions.open()
+
+    @property
+    def savepoints(self):
+        """Target database must support savepoints."""
+
+        return exclusions.closed()
+
+    @property
+    def two_phase_transactions(self):
+        """Target database must support two-phase transactions."""
+
+        return exclusions.closed()
+
+    @property
+    def update_from(self):
+        """Target must support UPDATE..FROM syntax"""
+        return exclusions.closed()
+
+    @property
+    def delete_from(self):
+        """Target must support DELETE FROM..FROM or DELETE..USING syntax"""
+        return exclusions.closed()
+
+    @property
+    def update_where_target_in_subquery(self):
+        """Target must support UPDATE (or DELETE) where the same table is
+        present in a subquery in the WHERE clause.
+
+        This is an ANSI-standard syntax that apparently MySQL can't handle,
+        such as:
+
+        .. sourcecode:: sql
+
+            UPDATE documents SET flag=1 WHERE documents.title IN
+                (SELECT max(documents.title) AS title
+                    FROM documents GROUP BY documents.user_id
+                )
+
+        """
+        return exclusions.open()
+
+    @property
+    def mod_operator_as_percent_sign(self):
+        """target database must use a plain percent '%' as the 'modulus'
+        operator."""
+        return exclusions.closed()
+
+    @property
+    def percent_schema_names(self):
+        """target backend supports weird identifiers with percent signs
+        in them, e.g. 'some % column'.
+
+        this is a very weird use case but often has problems because of
+        DBAPIs that use python formatting.  It's not a critical use
+        case either.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def order_by_col_from_union(self):
+        """target database supports ordering by a column from a SELECT
+        inside of a UNION
+
+        E.g.:
+
+        .. sourcecode:: sql
+
+            (SELECT id, ...) UNION (SELECT id, ...) ORDER BY id
+
+        """
+        return exclusions.open()
+
+    @property
+    def order_by_label_with_expression(self):
+        """target backend supports ORDER BY a column label within an
+        expression.
+
+        Basically this:
+
+        .. sourcecode:: sql
+
+            select data as foo from test order by foo || 'bar'
+
+        Lots of databases including PostgreSQL don't support this,
+        so this is off by default.
+
+        """
+        return exclusions.closed()
+
+    @property
+    def order_by_collation(self):
+        def check(config):
+            try:
+                self.get_order_by_collation(config)
+                return False
+            except NotImplementedError:
+                return True
+
+        return exclusions.skip_if(check)
+
+    def get_order_by_collation(self, config):
+        raise NotImplementedError()
+
+    @property
+    def unicode_connections(self):
+        """Target driver must support non-ASCII characters being passed at
+        all.
+        """
+        return exclusions.open()
+
+    @property
+    def graceful_disconnects(self):
+        """Target driver must raise a DBAPI-level exception, such as
+        InterfaceError, when the underlying connection has been closed
+        and the execute() method is called.
+        """
+        return exclusions.open()
+
+    @property
+    def independent_connections(self):
+        """
+        Target must support simultaneous, independent database connections.
+        """
+        return exclusions.open()
+
+    @property
+    def independent_readonly_connections(self):
+        """
+        Target must support simultaneous, independent database connections
+        that will be used in a readonly fashion.
+
+        """
+        return exclusions.open()
+
+    @property
+    def skip_mysql_on_windows(self):
+        """Catchall for a large variety of MySQL on Windows failures"""
+        return exclusions.open()
+
+    @property
+    def ad_hoc_engines(self):
+        """Test environment must allow ad-hoc engine/connection creation.
+
+        DBs that scale poorly for many connections, even when closed, i.e.
+        Oracle, may use the "--low-connections" option which flags this
+        requirement as not present.
+
+        """
+        return exclusions.skip_if(
+            lambda config: config.options.low_connections
+        )
+
+    @property
+    def no_windows(self):
+        return exclusions.skip_if(self._running_on_windows())
+
+    def _running_on_windows(self):
+        return exclusions.LambdaPredicate(
+            lambda: platform.system() == "Windows",
+            description="running on Windows",
+        )
+
+    @property
+    def timing_intensive(self):
+        from . import config
+
+        return config.add_to_marker.timing_intensive
+
+    @property
+    def memory_intensive(self):
+        from . import config
+
+        return config.add_to_marker.memory_intensive
+
+    @property
+    def threading_with_mock(self):
+        """Mark tests that use threading and mock at the same time - stability
+        issues have been observed with coverage
+
+        """
+        return exclusions.skip_if(
+            lambda config: config.options.has_coverage,
+            "Stability issues with coverage",
+        )
+
+    @property
+    def sqlalchemy2_stubs(self):
+        def check(config):
+            try:
+                __import__("sqlalchemy-stubs.ext.mypy")
+            except ImportError:
+                return False
+            else:
+                return True
+
+        return exclusions.only_if(check)
+
+    @property
+    def no_sqlalchemy2_stubs(self):
+        def check(config):
+            try:
+                __import__("sqlalchemy-stubs.ext.mypy")
+            except ImportError:
+                return False
+            else:
+                return True
+
+        return exclusions.skip_if(check)
+
+    @property
+    def python38(self):
+        return exclusions.only_if(
+            lambda: util.py38, "Python 3.8 or above required"
+        )
+
+    @property
+    def python39(self):
+        return exclusions.only_if(
+            lambda: util.py39, "Python 3.9 or above required"
+        )
+
+    @property
+    def python310(self):
+        return exclusions.only_if(
+            lambda: util.py310, "Python 3.10 or above required"
+        )
+
+    @property
+    def python311(self):
+        return exclusions.only_if(
+            lambda: util.py311, "Python 3.11 or above required"
+        )
+
+    @property
+    def python312(self):
+        return exclusions.only_if(
+            lambda: util.py312, "Python 3.12 or above required"
+        )
+
+    @property
+    def cpython(self):
+        return exclusions.only_if(
+            lambda: util.cpython, "cPython interpreter needed"
+        )
+
+    @property
+    def is64bit(self):
+        return exclusions.only_if(lambda: util.is64bit, "64bit required")
+
+    @property
+    def patch_library(self):
+        def check_lib():
+            try:
+                __import__("patch")
+            except ImportError:
+                return False
+            else:
+                return True
+
+        return exclusions.only_if(check_lib, "patch library needed")
+
+    @property
+    def predictable_gc(self):
+        """target platform must remove all cycles unconditionally when
+        gc.collect() is called, as well as clean out unreferenced subclasses.
+
+        """
+        return self.cpython
+
+    @property
+    def no_coverage(self):
+        """Test should be skipped if coverage is enabled.
+
+        This is to block tests that exercise libraries that seem to be
+        sensitive to coverage, such as PostgreSQL notice logging.
+
+        """
+        return exclusions.skip_if(
+            lambda config: config.options.has_coverage,
+            "Issues observed when coverage is enabled",
+        )
+
+    def _has_mysql_on_windows(self, config):
+        return False
+
+    def _has_mysql_fully_case_sensitive(self, config):
+        return False
+
+    @property
+    def sqlite(self):
+        return exclusions.skip_if(lambda: not self._has_sqlite())
+
+    @property
+    def cextensions(self):
+        return exclusions.skip_if(
+            lambda: not util.has_compiled_ext(),
+            "Cython extensions not installed",
+        )
+
+    def _has_sqlite(self):
+        from sqlalchemy import create_engine
+
+        try:
+            create_engine("sqlite://")
+            return True
+        except ImportError:
+            return False
+
+    @property
+    def async_dialect(self):
+        """dialect makes use of await_() to invoke operations on the DBAPI."""
+
+        return exclusions.closed()
+
+    @property
+    def asyncio(self):
+        return self.greenlet
+
+    @property
+    def no_greenlet(self):
+        def go(config):
+            try:
+                import greenlet  # noqa: F401
+            except ImportError:
+                return True
+            else:
+                return False
+
+        return exclusions.only_if(go)
+
+    @property
+    def greenlet(self):
+        def go(config):
+            if not _test_asyncio.ENABLE_ASYNCIO:
+                return False
+
+            try:
+                import greenlet  # noqa: F401
+            except ImportError:
+                return False
+            else:
+                return True
+
+        return exclusions.only_if(go)
+
+    @property
+    def computed_columns(self):
+        "Supports computed columns"
+        return exclusions.closed()
+
+    @property
+    def computed_columns_stored(self):
+        "Supports computed columns with `persisted=True`"
+        return exclusions.closed()
+
+    @property
+    def computed_columns_virtual(self):
+        "Supports computed columns with `persisted=False`"
+        return exclusions.closed()
+
+    @property
+    def computed_columns_default_persisted(self):
+        """If the default persistence is virtual or stored when `persisted`
+        is omitted"""
+        return exclusions.closed()
+
+    @property
+    def computed_columns_reflect_persisted(self):
+        """If persistence information is returned by the reflection of
+        computed columns"""
+        return exclusions.closed()
+
+    @property
+    def supports_distinct_on(self):
+        """If a backend supports the DISTINCT ON in a select"""
+        return exclusions.closed()
+
+    @property
+    def supports_is_distinct_from(self):
+        """Supports some form of "x IS [NOT] DISTINCT FROM y" construct.
+        Different dialects will implement their own flavour, e.g.,
+        sqlite will emit "x IS NOT y" instead of "x IS DISTINCT FROM y".
+
+        .. seealso::
+
+            :meth:`.ColumnOperators.is_distinct_from`
+
+        """
+        return exclusions.skip_if(
+            lambda config: not config.db.dialect.supports_is_distinct_from,
+            "driver doesn't support an IS DISTINCT FROM construct",
+        )
+
+    @property
+    def identity_columns(self):
+        """If a backend supports GENERATED { ALWAYS | BY DEFAULT }
+        AS IDENTITY"""
+        return exclusions.closed()
+
+    @property
+    def identity_columns_standard(self):
+        """If a backend supports GENERATED { ALWAYS | BY DEFAULT }
+        AS IDENTITY with a standard syntax.
+        This is mainly to exclude MSSql.
+        """
+        return exclusions.closed()
+
+    @property
+    def regexp_match(self):
+        """backend supports the regexp_match operator."""
+        return exclusions.closed()
+
+    @property
+    def regexp_replace(self):
+        """backend supports the regexp_replace operator."""
+        return exclusions.closed()
+
+    @property
+    def fetch_first(self):
+        """backend supports the fetch first clause."""
+        return exclusions.closed()
+
+    @property
+    def fetch_percent(self):
+        """backend supports the fetch first clause with percent."""
+        return exclusions.closed()
+
+    @property
+    def fetch_ties(self):
+        """backend supports the fetch first clause with ties."""
+        return exclusions.closed()
+
+    @property
+    def fetch_no_order_by(self):
+        """backend supports the fetch first without order by"""
+        return exclusions.closed()
+
+    @property
+    def fetch_offset_with_options(self):
+        """backend supports the offset when using fetch first with percent
+        or ties. basically this is "not mssql"
+        """
+        return exclusions.closed()
+
+    @property
+    def fetch_expression(self):
+        """backend supports fetch / offset with expression in them, like
+
+        SELECT * FROM some_table
+        OFFSET 1 + 1 ROWS FETCH FIRST 1 + 1 ROWS ONLY
+        """
+        return exclusions.closed()
+
+    @property
+    def autoincrement_without_sequence(self):
+        """If autoincrement=True on a column does not require an explicit
+        sequence. This should be false only for oracle.
+        """
+        return exclusions.open()
+
+    @property
+    def generic_classes(self):
+        "If X[Y] can be implemented with ``__class_getitem__``. py3.7+"
+        return exclusions.open()
+
+    @property
+    def json_deserializer_binary(self):
+        "indicates if the json_deserializer function is called with bytes"
+        return exclusions.closed()
+
+    @property
+    def reflect_table_options(self):
+        """Target database must support reflecting table_options."""
+        return exclusions.closed()
+
+    @property
+    def materialized_views(self):
+        """Target database must support MATERIALIZED VIEWs."""
+        return exclusions.closed()
+
+    @property
+    def materialized_views_reflect_pk(self):
+        """Target database reflect MATERIALIZED VIEWs pks."""
+        return exclusions.closed()
+
+    @property
+    def supports_bitwise_or(self):
+        """Target database supports bitwise or"""
+        return exclusions.closed()
+
+    @property
+    def supports_bitwise_and(self):
+        """Target database supports bitwise and"""
+        return exclusions.closed()
+
+    @property
+    def supports_bitwise_not(self):
+        """Target database supports bitwise not"""
+        return exclusions.closed()
+
+    @property
+    def supports_bitwise_xor(self):
+        """Target database supports bitwise xor"""
+        return exclusions.closed()
+
+    @property
+    def supports_bitwise_shift(self):
+        """Target database supports bitwise left or right shift"""
+        return exclusions.closed()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/schema.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/schema.py
new file mode 100644
index 00000000..0dd7de20
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/schema.py
@@ -0,0 +1,224 @@
+# testing/schema.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import sys
+
+from . import config
+from . import exclusions
+from .. import event
+from .. import schema
+from .. import types as sqltypes
+from ..orm import mapped_column as _orm_mapped_column
+from ..util import OrderedDict
+
+__all__ = ["Table", "Column"]
+
+table_options = {}
+
+
+def Table(*args, **kw) -> schema.Table:
+    """A schema.Table wrapper/hook for dialect-specific tweaks."""
+
+    test_opts = {k: kw.pop(k) for k in list(kw) if k.startswith("test_")}
+
+    kw.update(table_options)
+
+    if exclusions.against(config._current, "mysql"):
+        if (
+            "mysql_engine" not in kw
+            and "mysql_type" not in kw
+            and "autoload_with" not in kw
+        ):
+            if "test_needs_fk" in test_opts or "test_needs_acid" in test_opts:
+                kw["mysql_engine"] = "InnoDB"
+            else:
+                # there are in fact test fixtures that rely upon MyISAM,
+                # due to MySQL / MariaDB having poor FK behavior under innodb,
+                # such as a self-referential table can't be deleted from at
+                # once without attending to per-row dependencies.  We'd need to
+                # add special steps to some fixtures if we want to not
+                # explicitly state MyISAM here
+                kw["mysql_engine"] = "MyISAM"
+    elif exclusions.against(config._current, "mariadb"):
+        if (
+            "mariadb_engine" not in kw
+            and "mariadb_type" not in kw
+            and "autoload_with" not in kw
+        ):
+            if "test_needs_fk" in test_opts or "test_needs_acid" in test_opts:
+                kw["mariadb_engine"] = "InnoDB"
+            else:
+                kw["mariadb_engine"] = "MyISAM"
+
+    return schema.Table(*args, **kw)
+
+
+def mapped_column(*args, **kw):
+    """An orm.mapped_column wrapper/hook for dialect-specific tweaks."""
+
+    return _schema_column(_orm_mapped_column, args, kw)
+
+
+def Column(*args, **kw):
+    """A schema.Column wrapper/hook for dialect-specific tweaks."""
+
+    return _schema_column(schema.Column, args, kw)
+
+
+def _schema_column(factory, args, kw):
+    test_opts = {k: kw.pop(k) for k in list(kw) if k.startswith("test_")}
+
+    if not config.requirements.foreign_key_ddl.enabled_for_config(config):
+        args = [arg for arg in args if not isinstance(arg, schema.ForeignKey)]
+
+    construct = factory(*args, **kw)
+
+    if factory is schema.Column:
+        col = construct
+    else:
+        col = construct.column
+
+    if test_opts.get("test_needs_autoincrement", False) and kw.get(
+        "primary_key", False
+    ):
+        if col.default is None and col.server_default is None:
+            col.autoincrement = True
+
+        # allow any test suite to pick up on this
+        col.info["test_needs_autoincrement"] = True
+
+        # hardcoded rule for oracle; this should
+        # be moved out
+        if exclusions.against(config._current, "oracle"):
+
+            def add_seq(c, tbl):
+                c._init_items(
+                    schema.Sequence(
+                        _truncate_name(
+                            config.db.dialect, tbl.name + "_" + c.name + "_seq"
+                        ),
+                        optional=True,
+                    )
+                )
+
+            event.listen(col, "after_parent_attach", add_seq, propagate=True)
+    return construct
+
+
+class eq_type_affinity:
+    """Helper to compare types inside of datastructures based on affinity.
+
+    E.g.::
+
+        eq_(
+            inspect(connection).get_columns("foo"),
+            [
+                {
+                    "name": "id",
+                    "type": testing.eq_type_affinity(sqltypes.INTEGER),
+                    "nullable": False,
+                    "default": None,
+                    "autoincrement": False,
+                },
+                {
+                    "name": "data",
+                    "type": testing.eq_type_affinity(sqltypes.NullType),
+                    "nullable": True,
+                    "default": None,
+                    "autoincrement": False,
+                },
+            ],
+        )
+
+    """
+
+    def __init__(self, target):
+        self.target = sqltypes.to_instance(target)
+
+    def __eq__(self, other):
+        return self.target._type_affinity is other._type_affinity
+
+    def __ne__(self, other):
+        return self.target._type_affinity is not other._type_affinity
+
+
+class eq_compile_type:
+    """similar to eq_type_affinity but uses compile"""
+
+    def __init__(self, target):
+        self.target = target
+
+    def __eq__(self, other):
+        return self.target == other.compile()
+
+    def __ne__(self, other):
+        return self.target != other.compile()
+
+
+class eq_clause_element:
+    """Helper to compare SQL structures based on compare()"""
+
+    def __init__(self, target):
+        self.target = target
+
+    def __eq__(self, other):
+        return self.target.compare(other)
+
+    def __ne__(self, other):
+        return not self.target.compare(other)
+
+
+def _truncate_name(dialect, name):
+    if len(name) > dialect.max_identifier_length:
+        return (
+            name[0 : max(dialect.max_identifier_length - 6, 0)]
+            + "_"
+            + hex(hash(name) % 64)[2:]
+        )
+    else:
+        return name
+
+
+def pep435_enum(name):
+    # Implements PEP 435 in the minimal fashion needed by SQLAlchemy
+    __members__ = OrderedDict()
+
+    def __init__(self, name, value, alias=None):
+        self.name = name
+        self.value = value
+        self.__members__[name] = self
+        value_to_member[value] = self
+        setattr(self.__class__, name, self)
+        if alias:
+            self.__members__[alias] = self
+            setattr(self.__class__, alias, self)
+
+    value_to_member = {}
+
+    @classmethod
+    def get(cls, value):
+        return value_to_member[value]
+
+    someenum = type(
+        name,
+        (object,),
+        {"__members__": __members__, "__init__": __init__, "get": get},
+    )
+
+    # getframe() trick for pickling I don't understand courtesy
+    # Python namedtuple()
+    try:
+        module = sys._getframe(1).f_globals.get("__name__", "__main__")
+    except (AttributeError, ValueError):
+        pass
+    if module is not None:
+        someenum.__module__ = module
+
+    return someenum
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__init__.py
new file mode 100644
index 00000000..8435aa00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__init__.py
@@ -0,0 +1,19 @@
+# testing/suite/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+from .test_cte import *  # noqa
+from .test_ddl import *  # noqa
+from .test_deprecations import *  # noqa
+from .test_dialect import *  # noqa
+from .test_insert import *  # noqa
+from .test_reflection import *  # noqa
+from .test_results import *  # noqa
+from .test_rowcount import *  # noqa
+from .test_select import *  # noqa
+from .test_sequence import *  # noqa
+from .test_types import *  # noqa
+from .test_unicode_ddl import *  # noqa
+from .test_update_delete import *  # noqa
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7cdebf09
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_cte.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_cte.cpython-312.pyc
new file mode 100644
index 00000000..fc8ac506
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_cte.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_ddl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_ddl.cpython-312.pyc
new file mode 100644
index 00000000..92e2efcc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_ddl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_deprecations.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_deprecations.cpython-312.pyc
new file mode 100644
index 00000000..d499662e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_deprecations.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_dialect.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_dialect.cpython-312.pyc
new file mode 100644
index 00000000..4a7e76c0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_dialect.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_insert.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_insert.cpython-312.pyc
new file mode 100644
index 00000000..bdcdefbe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_insert.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_reflection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_reflection.cpython-312.pyc
new file mode 100644
index 00000000..9d4d4c26
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_reflection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_results.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_results.cpython-312.pyc
new file mode 100644
index 00000000..50a60a1e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_results.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_rowcount.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_rowcount.cpython-312.pyc
new file mode 100644
index 00000000..3b3ffae6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_rowcount.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_select.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_select.cpython-312.pyc
new file mode 100644
index 00000000..54df6c54
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_select.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_sequence.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_sequence.cpython-312.pyc
new file mode 100644
index 00000000..65141ffe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_sequence.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_types.cpython-312.pyc
new file mode 100644
index 00000000..e70f3fa1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_unicode_ddl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_unicode_ddl.cpython-312.pyc
new file mode 100644
index 00000000..44b1b56e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_unicode_ddl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_update_delete.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_update_delete.cpython-312.pyc
new file mode 100644
index 00000000..901a8fd8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/__pycache__/test_update_delete.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_cte.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_cte.py
new file mode 100644
index 00000000..4e4d420f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_cte.py
@@ -0,0 +1,211 @@
+# testing/suite/test_cte.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .. import fixtures
+from ..assertions import eq_
+from ..schema import Column
+from ..schema import Table
+from ... import ForeignKey
+from ... import Integer
+from ... import select
+from ... import String
+from ... import testing
+
+
+class CTETest(fixtures.TablesTest):
+    __backend__ = True
+    __requires__ = ("ctes",)
+
+    run_inserts = "each"
+    run_deletes = "each"
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+            Column("parent_id", ForeignKey("some_table.id")),
+        )
+
+        Table(
+            "some_other_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+            Column("parent_id", Integer),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "data": "d1", "parent_id": None},
+                {"id": 2, "data": "d2", "parent_id": 1},
+                {"id": 3, "data": "d3", "parent_id": 1},
+                {"id": 4, "data": "d4", "parent_id": 3},
+                {"id": 5, "data": "d5", "parent_id": 3},
+            ],
+        )
+
+    def test_select_nonrecursive_round_trip(self, connection):
+        some_table = self.tables.some_table
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte")
+        )
+        result = connection.execute(
+            select(cte.c.data).where(cte.c.data.in_(["d4", "d5"]))
+        )
+        eq_(result.fetchall(), [("d4",)])
+
+    def test_select_recursive_round_trip(self, connection):
+        some_table = self.tables.some_table
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte", recursive=True)
+        )
+
+        cte_alias = cte.alias("c1")
+        st1 = some_table.alias()
+        # note that SQL Server requires this to be UNION ALL,
+        # can't be UNION
+        cte = cte.union_all(
+            select(st1).where(st1.c.id == cte_alias.c.parent_id)
+        )
+        result = connection.execute(
+            select(cte.c.data)
+            .where(cte.c.data != "d2")
+            .order_by(cte.c.data.desc())
+        )
+        eq_(
+            result.fetchall(),
+            [("d4",), ("d3",), ("d3",), ("d1",), ("d1",), ("d1",)],
+        )
+
+    def test_insert_from_select_round_trip(self, connection):
+        some_table = self.tables.some_table
+        some_other_table = self.tables.some_other_table
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte")
+        )
+        connection.execute(
+            some_other_table.insert().from_select(
+                ["id", "data", "parent_id"], select(cte)
+            )
+        )
+        eq_(
+            connection.execute(
+                select(some_other_table).order_by(some_other_table.c.id)
+            ).fetchall(),
+            [(2, "d2", 1), (3, "d3", 1), (4, "d4", 3)],
+        )
+
+    @testing.requires.ctes_with_update_delete
+    @testing.requires.update_from
+    def test_update_from_round_trip(self, connection):
+        some_table = self.tables.some_table
+        some_other_table = self.tables.some_other_table
+
+        connection.execute(
+            some_other_table.insert().from_select(
+                ["id", "data", "parent_id"], select(some_table)
+            )
+        )
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte")
+        )
+        connection.execute(
+            some_other_table.update()
+            .values(parent_id=5)
+            .where(some_other_table.c.data == cte.c.data)
+        )
+        eq_(
+            connection.execute(
+                select(some_other_table).order_by(some_other_table.c.id)
+            ).fetchall(),
+            [
+                (1, "d1", None),
+                (2, "d2", 5),
+                (3, "d3", 5),
+                (4, "d4", 5),
+                (5, "d5", 3),
+            ],
+        )
+
+    @testing.requires.ctes_with_update_delete
+    @testing.requires.delete_from
+    def test_delete_from_round_trip(self, connection):
+        some_table = self.tables.some_table
+        some_other_table = self.tables.some_other_table
+
+        connection.execute(
+            some_other_table.insert().from_select(
+                ["id", "data", "parent_id"], select(some_table)
+            )
+        )
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte")
+        )
+        connection.execute(
+            some_other_table.delete().where(
+                some_other_table.c.data == cte.c.data
+            )
+        )
+        eq_(
+            connection.execute(
+                select(some_other_table).order_by(some_other_table.c.id)
+            ).fetchall(),
+            [(1, "d1", None), (5, "d5", 3)],
+        )
+
+    @testing.requires.ctes_with_update_delete
+    def test_delete_scalar_subq_round_trip(self, connection):
+        some_table = self.tables.some_table
+        some_other_table = self.tables.some_other_table
+
+        connection.execute(
+            some_other_table.insert().from_select(
+                ["id", "data", "parent_id"], select(some_table)
+            )
+        )
+
+        cte = (
+            select(some_table)
+            .where(some_table.c.data.in_(["d2", "d3", "d4"]))
+            .cte("some_cte")
+        )
+        connection.execute(
+            some_other_table.delete().where(
+                some_other_table.c.data
+                == select(cte.c.data)
+                .where(cte.c.id == some_other_table.c.id)
+                .scalar_subquery()
+            )
+        )
+        eq_(
+            connection.execute(
+                select(some_other_table).order_by(some_other_table.c.id)
+            ).fetchall(),
+            [(1, "d1", None), (5, "d5", 3)],
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_ddl.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_ddl.py
new file mode 100644
index 00000000..c7e7d817
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_ddl.py
@@ -0,0 +1,389 @@
+# testing/suite/test_ddl.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import random
+
+from . import testing
+from .. import config
+from .. import fixtures
+from .. import util
+from ..assertions import eq_
+from ..assertions import is_false
+from ..assertions import is_true
+from ..config import requirements
+from ..schema import Table
+from ... import CheckConstraint
+from ... import Column
+from ... import ForeignKeyConstraint
+from ... import Index
+from ... import inspect
+from ... import Integer
+from ... import schema
+from ... import String
+from ... import UniqueConstraint
+
+
+class TableDDLTest(fixtures.TestBase):
+    __backend__ = True
+
+    def _simple_fixture(self, schema=None):
+        return Table(
+            "test_table",
+            self.metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+            schema=schema,
+        )
+
+    def _underscore_fixture(self):
+        return Table(
+            "_test_table",
+            self.metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("_data", String(50)),
+        )
+
+    def _table_index_fixture(self, schema=None):
+        table = self._simple_fixture(schema=schema)
+        idx = Index("test_index", table.c.data)
+        return table, idx
+
+    def _simple_roundtrip(self, table):
+        with config.db.begin() as conn:
+            conn.execute(table.insert().values((1, "some data")))
+            result = conn.execute(table.select())
+            eq_(result.first(), (1, "some data"))
+
+    @requirements.create_table
+    @util.provide_metadata
+    def test_create_table(self):
+        table = self._simple_fixture()
+        table.create(config.db, checkfirst=False)
+        self._simple_roundtrip(table)
+
+    @requirements.create_table
+    @requirements.schemas
+    @util.provide_metadata
+    def test_create_table_schema(self):
+        table = self._simple_fixture(schema=config.test_schema)
+        table.create(config.db, checkfirst=False)
+        self._simple_roundtrip(table)
+
+    @requirements.drop_table
+    @util.provide_metadata
+    def test_drop_table(self):
+        table = self._simple_fixture()
+        table.create(config.db, checkfirst=False)
+        table.drop(config.db, checkfirst=False)
+
+    @requirements.create_table
+    @util.provide_metadata
+    def test_underscore_names(self):
+        table = self._underscore_fixture()
+        table.create(config.db, checkfirst=False)
+        self._simple_roundtrip(table)
+
+    @requirements.comment_reflection
+    @util.provide_metadata
+    def test_add_table_comment(self, connection):
+        table = self._simple_fixture()
+        table.create(connection, checkfirst=False)
+        table.comment = "a comment"
+        connection.execute(schema.SetTableComment(table))
+        eq_(
+            inspect(connection).get_table_comment("test_table"),
+            {"text": "a comment"},
+        )
+
+    @requirements.comment_reflection
+    @util.provide_metadata
+    def test_drop_table_comment(self, connection):
+        table = self._simple_fixture()
+        table.create(connection, checkfirst=False)
+        table.comment = "a comment"
+        connection.execute(schema.SetTableComment(table))
+        connection.execute(schema.DropTableComment(table))
+        eq_(
+            inspect(connection).get_table_comment("test_table"), {"text": None}
+        )
+
+    @requirements.table_ddl_if_exists
+    @util.provide_metadata
+    def test_create_table_if_not_exists(self, connection):
+        table = self._simple_fixture()
+
+        connection.execute(schema.CreateTable(table, if_not_exists=True))
+
+        is_true(inspect(connection).has_table("test_table"))
+        connection.execute(schema.CreateTable(table, if_not_exists=True))
+
+    @requirements.index_ddl_if_exists
+    @util.provide_metadata
+    def test_create_index_if_not_exists(self, connection):
+        table, idx = self._table_index_fixture()
+
+        connection.execute(schema.CreateTable(table, if_not_exists=True))
+        is_true(inspect(connection).has_table("test_table"))
+        is_false(
+            "test_index"
+            in [
+                ix["name"]
+                for ix in inspect(connection).get_indexes("test_table")
+            ]
+        )
+
+        connection.execute(schema.CreateIndex(idx, if_not_exists=True))
+
+        is_true(
+            "test_index"
+            in [
+                ix["name"]
+                for ix in inspect(connection).get_indexes("test_table")
+            ]
+        )
+
+        connection.execute(schema.CreateIndex(idx, if_not_exists=True))
+
+    @requirements.table_ddl_if_exists
+    @util.provide_metadata
+    def test_drop_table_if_exists(self, connection):
+        table = self._simple_fixture()
+
+        table.create(connection)
+
+        is_true(inspect(connection).has_table("test_table"))
+
+        connection.execute(schema.DropTable(table, if_exists=True))
+
+        is_false(inspect(connection).has_table("test_table"))
+
+        connection.execute(schema.DropTable(table, if_exists=True))
+
+    @requirements.index_ddl_if_exists
+    @util.provide_metadata
+    def test_drop_index_if_exists(self, connection):
+        table, idx = self._table_index_fixture()
+
+        table.create(connection)
+
+        is_true(
+            "test_index"
+            in [
+                ix["name"]
+                for ix in inspect(connection).get_indexes("test_table")
+            ]
+        )
+
+        connection.execute(schema.DropIndex(idx, if_exists=True))
+
+        is_false(
+            "test_index"
+            in [
+                ix["name"]
+                for ix in inspect(connection).get_indexes("test_table")
+            ]
+        )
+
+        connection.execute(schema.DropIndex(idx, if_exists=True))
+
+
+class FutureTableDDLTest(fixtures.FutureEngineMixin, TableDDLTest):
+    pass
+
+
+class LongNameBlowoutTest(fixtures.TestBase):
+    """test the creation of a variety of DDL structures and ensure
+    label length limits pass on backends
+
+    """
+
+    __backend__ = True
+
+    def fk(self, metadata, connection):
+        convention = {
+            "fk": "foreign_key_%(table_name)s_"
+            "%(column_0_N_name)s_"
+            "%(referred_table_name)s_"
+            + (
+                "_".join(
+                    "".join(random.choice("abcdef") for j in range(20))
+                    for i in range(10)
+                )
+            ),
+        }
+        metadata.naming_convention = convention
+
+        Table(
+            "a_things_with_stuff",
+            metadata,
+            Column("id_long_column_name", Integer, primary_key=True),
+            test_needs_fk=True,
+        )
+
+        cons = ForeignKeyConstraint(
+            ["aid"], ["a_things_with_stuff.id_long_column_name"]
+        )
+        Table(
+            "b_related_things_of_value",
+            metadata,
+            Column(
+                "aid",
+            ),
+            cons,
+            test_needs_fk=True,
+        )
+        actual_name = cons.name
+
+        metadata.create_all(connection)
+
+        if testing.requires.foreign_key_constraint_name_reflection.enabled:
+            insp = inspect(connection)
+            fks = insp.get_foreign_keys("b_related_things_of_value")
+            reflected_name = fks[0]["name"]
+
+            return actual_name, reflected_name
+        else:
+            return actual_name, None
+
+    def pk(self, metadata, connection):
+        convention = {
+            "pk": "primary_key_%(table_name)s_"
+            "%(column_0_N_name)s"
+            + (
+                "_".join(
+                    "".join(random.choice("abcdef") for j in range(30))
+                    for i in range(10)
+                )
+            ),
+        }
+        metadata.naming_convention = convention
+
+        a = Table(
+            "a_things_with_stuff",
+            metadata,
+            Column("id_long_column_name", Integer, primary_key=True),
+            Column("id_another_long_name", Integer, primary_key=True),
+        )
+        cons = a.primary_key
+        actual_name = cons.name
+
+        metadata.create_all(connection)
+        insp = inspect(connection)
+        pk = insp.get_pk_constraint("a_things_with_stuff")
+        reflected_name = pk["name"]
+        return actual_name, reflected_name
+
+    def ix(self, metadata, connection):
+        convention = {
+            "ix": "index_%(table_name)s_"
+            "%(column_0_N_name)s"
+            + (
+                "_".join(
+                    "".join(random.choice("abcdef") for j in range(30))
+                    for i in range(10)
+                )
+            ),
+        }
+        metadata.naming_convention = convention
+
+        a = Table(
+            "a_things_with_stuff",
+            metadata,
+            Column("id_long_column_name", Integer, primary_key=True),
+            Column("id_another_long_name", Integer),
+        )
+        cons = Index(None, a.c.id_long_column_name, a.c.id_another_long_name)
+        actual_name = cons.name
+
+        metadata.create_all(connection)
+        insp = inspect(connection)
+        ix = insp.get_indexes("a_things_with_stuff")
+        reflected_name = ix[0]["name"]
+        return actual_name, reflected_name
+
+    def uq(self, metadata, connection):
+        convention = {
+            "uq": "unique_constraint_%(table_name)s_"
+            "%(column_0_N_name)s"
+            + (
+                "_".join(
+                    "".join(random.choice("abcdef") for j in range(30))
+                    for i in range(10)
+                )
+            ),
+        }
+        metadata.naming_convention = convention
+
+        cons = UniqueConstraint("id_long_column_name", "id_another_long_name")
+        Table(
+            "a_things_with_stuff",
+            metadata,
+            Column("id_long_column_name", Integer, primary_key=True),
+            Column("id_another_long_name", Integer),
+            cons,
+        )
+        actual_name = cons.name
+
+        metadata.create_all(connection)
+        insp = inspect(connection)
+        uq = insp.get_unique_constraints("a_things_with_stuff")
+        reflected_name = uq[0]["name"]
+        return actual_name, reflected_name
+
+    def ck(self, metadata, connection):
+        convention = {
+            "ck": "check_constraint_%(table_name)s"
+            + (
+                "_".join(
+                    "".join(random.choice("abcdef") for j in range(30))
+                    for i in range(10)
+                )
+            ),
+        }
+        metadata.naming_convention = convention
+
+        cons = CheckConstraint("some_long_column_name > 5")
+        Table(
+            "a_things_with_stuff",
+            metadata,
+            Column("id_long_column_name", Integer, primary_key=True),
+            Column("some_long_column_name", Integer),
+            cons,
+        )
+        actual_name = cons.name
+
+        metadata.create_all(connection)
+        insp = inspect(connection)
+        ck = insp.get_check_constraints("a_things_with_stuff")
+        reflected_name = ck[0]["name"]
+        return actual_name, reflected_name
+
+    @testing.combinations(
+        ("fk",),
+        ("pk",),
+        ("ix",),
+        ("ck", testing.requires.check_constraint_reflection.as_skips()),
+        ("uq", testing.requires.unique_constraint_reflection.as_skips()),
+        argnames="type_",
+    )
+    def test_long_convention_name(self, type_, metadata, connection):
+        actual_name, reflected_name = getattr(self, type_)(
+            metadata, connection
+        )
+
+        assert len(actual_name) > 255
+
+        if reflected_name is not None:
+            overlap = actual_name[0 : len(reflected_name)]
+            if len(overlap) < len(actual_name):
+                eq_(overlap[0:-5], reflected_name[0 : len(overlap) - 5])
+            else:
+                eq_(overlap, reflected_name)
+
+
+__all__ = ("TableDDLTest", "FutureTableDDLTest", "LongNameBlowoutTest")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_deprecations.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_deprecations.py
new file mode 100644
index 00000000..db0a9fc4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_deprecations.py
@@ -0,0 +1,153 @@
+# testing/suite/test_deprecations.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .. import fixtures
+from ..assertions import eq_
+from ..schema import Column
+from ..schema import Table
+from ... import Integer
+from ... import select
+from ... import testing
+from ... import union
+
+
+class DeprecatedCompoundSelectTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2},
+                {"id": 2, "x": 2, "y": 3},
+                {"id": 3, "x": 3, "y": 4},
+                {"id": 4, "x": 4, "y": 5},
+            ],
+        )
+
+    def _assert_result(self, conn, select, result, params=()):
+        eq_(conn.execute(select, params).fetchall(), result)
+
+    def test_plain_union(self, connection):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2)
+        s2 = select(table).where(table.c.id == 3)
+
+        u1 = union(s1, s2)
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
+
+    # note we've had to remove one use case entirely, which is this
+    # one.   the Select gets its FROMS from the WHERE clause and the
+    # columns clause, but not the ORDER BY, which means the old ".c" system
+    # allowed you to "order_by(s.c.foo)" to get an unnamed column in the
+    # ORDER BY without adding the SELECT into the FROM and breaking the
+    # query.  Users will have to adjust for this use case if they were doing
+    # it before.
+    def _dont_test_select_from_plain_union(self, connection):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2)
+        s2 = select(table).where(table.c.id == 3)
+
+        u1 = union(s1, s2).alias().select()
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
+
+    @testing.requires.order_by_col_from_union
+    @testing.requires.parens_in_union_contained_select_w_limit_offset
+    def test_limit_offset_selectable_in_unions(self, connection):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).limit(1).order_by(table.c.id)
+        s2 = select(table).where(table.c.id == 3).limit(1).order_by(table.c.id)
+
+        u1 = union(s1, s2).limit(2)
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
+
+    @testing.requires.parens_in_union_contained_select_wo_limit_offset
+    def test_order_by_selectable_in_unions(self, connection):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).order_by(table.c.id)
+        s2 = select(table).where(table.c.id == 3).order_by(table.c.id)
+
+        u1 = union(s1, s2).limit(2)
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
+
+    def test_distinct_selectable_in_unions(self, connection):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).distinct()
+        s2 = select(table).where(table.c.id == 3).distinct()
+
+        u1 = union(s1, s2).limit(2)
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
+
+    def test_limit_offset_aliased_selectable_in_unions(self, connection):
+        table = self.tables.some_table
+        s1 = (
+            select(table)
+            .where(table.c.id == 2)
+            .limit(1)
+            .order_by(table.c.id)
+            .alias()
+            .select()
+        )
+        s2 = (
+            select(table)
+            .where(table.c.id == 3)
+            .limit(1)
+            .order_by(table.c.id)
+            .alias()
+            .select()
+        )
+
+        u1 = union(s1, s2).limit(2)
+        with testing.expect_deprecated(
+            "The SelectBase.c and SelectBase.columns "
+            "attributes are deprecated"
+        ):
+            self._assert_result(
+                connection, u1.order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_dialect.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_dialect.py
new file mode 100644
index 00000000..ae67cc10
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_dialect.py
@@ -0,0 +1,740 @@
+# testing/suite/test_dialect.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import importlib
+
+from . import testing
+from .. import assert_raises
+from .. import config
+from .. import engines
+from .. import eq_
+from .. import fixtures
+from .. import is_not_none
+from .. import is_true
+from .. import ne_
+from .. import provide_metadata
+from ..assertions import expect_raises
+from ..assertions import expect_raises_message
+from ..config import requirements
+from ..provision import set_default_schema_on_connection
+from ..schema import Column
+from ..schema import Table
+from ... import bindparam
+from ... import dialects
+from ... import event
+from ... import exc
+from ... import Integer
+from ... import literal_column
+from ... import select
+from ... import String
+from ...sql.compiler import Compiled
+from ...util import inspect_getfullargspec
+
+
+class PingTest(fixtures.TestBase):
+    __backend__ = True
+
+    def test_do_ping(self):
+        with testing.db.connect() as conn:
+            is_true(
+                testing.db.dialect.do_ping(conn.connection.dbapi_connection)
+            )
+
+
+class ArgSignatureTest(fixtures.TestBase):
+    """test that all visit_XYZ() in :class:`_sql.Compiler` subclasses have
+    ``**kw``, for #8988.
+
+    This test uses runtime code inspection.   Does not need to be a
+    ``__backend__`` test as it only needs to run once provided all target
+    dialects have been imported.
+
+    For third party dialects, the suite would be run with that third
+    party as a "--dburi", which means its compiler classes will have been
+    imported by the time this test runs.
+
+    """
+
+    def _all_subclasses():  # type: ignore  # noqa
+        for d in dialects.__all__:
+            if not d.startswith("_"):
+                importlib.import_module("sqlalchemy.dialects.%s" % d)
+
+        stack = [Compiled]
+
+        while stack:
+            cls = stack.pop(0)
+            stack.extend(cls.__subclasses__())
+            yield cls
+
+    @testing.fixture(params=list(_all_subclasses()))
+    def all_subclasses(self, request):
+        yield request.param
+
+    def test_all_visit_methods_accept_kw(self, all_subclasses):
+        cls = all_subclasses
+
+        for k in cls.__dict__:
+            if k.startswith("visit_"):
+                meth = getattr(cls, k)
+
+                insp = inspect_getfullargspec(meth)
+                is_not_none(
+                    insp.varkw,
+                    f"Compiler visit method {cls.__name__}.{k}() does "
+                    "not accommodate for **kw in its argument signature",
+                )
+
+
+class ExceptionTest(fixtures.TablesTest):
+    """Test basic exception wrapping.
+
+    DBAPIs vary a lot in exception behavior so to actually anticipate
+    specific exceptions from real round trips, we need to be conservative.
+
+    """
+
+    run_deletes = "each"
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "manual_pk",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+        )
+
+    @requirements.duplicate_key_raises_integrity_error
+    def test_integrity_error(self):
+        with config.db.connect() as conn:
+            trans = conn.begin()
+            conn.execute(
+                self.tables.manual_pk.insert(), {"id": 1, "data": "d1"}
+            )
+
+            assert_raises(
+                exc.IntegrityError,
+                conn.execute,
+                self.tables.manual_pk.insert(),
+                {"id": 1, "data": "d1"},
+            )
+
+            trans.rollback()
+
+    def test_exception_with_non_ascii(self):
+        with config.db.connect() as conn:
+            try:
+                # try to create an error message that likely has non-ascii
+                # characters in the DBAPI's message string.  unfortunately
+                # there's no way to make this happen with some drivers like
+                # mysqlclient, pymysql.  this at least does produce a non-
+                # ascii error message for cx_oracle, psycopg2
+                conn.execute(select(literal_column("méil")))
+                assert False
+            except exc.DBAPIError as err:
+                err_str = str(err)
+
+                assert str(err.orig) in str(err)
+
+            assert isinstance(err_str, str)
+
+
+class IsolationLevelTest(fixtures.TestBase):
+    __backend__ = True
+
+    __requires__ = ("isolation_level",)
+
+    def _get_non_default_isolation_level(self):
+        levels = requirements.get_isolation_levels(config)
+
+        default = levels["default"]
+        supported = levels["supported"]
+
+        s = set(supported).difference(["AUTOCOMMIT", default])
+        if s:
+            return s.pop()
+        else:
+            config.skip_test("no non-default isolation level available")
+
+    def test_default_isolation_level(self):
+        eq_(
+            config.db.dialect.default_isolation_level,
+            requirements.get_isolation_levels(config)["default"],
+        )
+
+    def test_non_default_isolation_level(self):
+        non_default = self._get_non_default_isolation_level()
+
+        with config.db.connect() as conn:
+            existing = conn.get_isolation_level()
+
+            ne_(existing, non_default)
+
+            conn.execution_options(isolation_level=non_default)
+
+            eq_(conn.get_isolation_level(), non_default)
+
+            conn.dialect.reset_isolation_level(
+                conn.connection.dbapi_connection
+            )
+
+            eq_(conn.get_isolation_level(), existing)
+
+    def test_all_levels(self):
+        levels = requirements.get_isolation_levels(config)
+
+        all_levels = levels["supported"]
+
+        for level in set(all_levels).difference(["AUTOCOMMIT"]):
+            with config.db.connect() as conn:
+                conn.execution_options(isolation_level=level)
+
+                eq_(conn.get_isolation_level(), level)
+
+                trans = conn.begin()
+                trans.rollback()
+
+                eq_(conn.get_isolation_level(), level)
+
+            with config.db.connect() as conn:
+                eq_(
+                    conn.get_isolation_level(),
+                    levels["default"],
+                )
+
+    @testing.requires.get_isolation_level_values
+    def test_invalid_level_execution_option(self, connection_no_trans):
+        """test for the new get_isolation_level_values() method"""
+
+        connection = connection_no_trans
+        with expect_raises_message(
+            exc.ArgumentError,
+            "Invalid value '%s' for isolation_level. "
+            "Valid isolation levels for '%s' are %s"
+            % (
+                "FOO",
+                connection.dialect.name,
+                ", ".join(
+                    requirements.get_isolation_levels(config)["supported"]
+                ),
+            ),
+        ):
+            connection.execution_options(isolation_level="FOO")
+
+    @testing.requires.get_isolation_level_values
+    @testing.requires.dialect_level_isolation_level_param
+    def test_invalid_level_engine_param(self, testing_engine):
+        """test for the new get_isolation_level_values() method
+        and support for the dialect-level 'isolation_level' parameter.
+
+        """
+
+        eng = testing_engine(options=dict(isolation_level="FOO"))
+        with expect_raises_message(
+            exc.ArgumentError,
+            "Invalid value '%s' for isolation_level. "
+            "Valid isolation levels for '%s' are %s"
+            % (
+                "FOO",
+                eng.dialect.name,
+                ", ".join(
+                    requirements.get_isolation_levels(config)["supported"]
+                ),
+            ),
+        ):
+            eng.connect()
+
+    @testing.requires.independent_readonly_connections
+    def test_dialect_user_setting_is_restored(self, testing_engine):
+        levels = requirements.get_isolation_levels(config)
+        default = levels["default"]
+        supported = (
+            sorted(
+                set(levels["supported"]).difference([default, "AUTOCOMMIT"])
+            )
+        )[0]
+
+        e = testing_engine(options={"isolation_level": supported})
+
+        with e.connect() as conn:
+            eq_(conn.get_isolation_level(), supported)
+
+        with e.connect() as conn:
+            conn.execution_options(isolation_level=default)
+            eq_(conn.get_isolation_level(), default)
+
+        with e.connect() as conn:
+            eq_(conn.get_isolation_level(), supported)
+
+
+class AutocommitIsolationTest(fixtures.TablesTest):
+    run_deletes = "each"
+
+    __requires__ = ("autocommit",)
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+            test_needs_acid=True,
+        )
+
+    def _test_conn_autocommits(self, conn, autocommit):
+        trans = conn.begin()
+        conn.execute(
+            self.tables.some_table.insert(), {"id": 1, "data": "some data"}
+        )
+        trans.rollback()
+
+        eq_(
+            conn.scalar(select(self.tables.some_table.c.id)),
+            1 if autocommit else None,
+        )
+        conn.rollback()
+
+        with conn.begin():
+            conn.execute(self.tables.some_table.delete())
+
+    def test_autocommit_on(self, connection_no_trans):
+        conn = connection_no_trans
+        c2 = conn.execution_options(isolation_level="AUTOCOMMIT")
+        self._test_conn_autocommits(c2, True)
+
+        c2.dialect.reset_isolation_level(c2.connection.dbapi_connection)
+
+        self._test_conn_autocommits(conn, False)
+
+    def test_autocommit_off(self, connection_no_trans):
+        conn = connection_no_trans
+        self._test_conn_autocommits(conn, False)
+
+    def test_turn_autocommit_off_via_default_iso_level(
+        self, connection_no_trans
+    ):
+        conn = connection_no_trans
+        conn = conn.execution_options(isolation_level="AUTOCOMMIT")
+        self._test_conn_autocommits(conn, True)
+
+        conn.execution_options(
+            isolation_level=requirements.get_isolation_levels(config)[
+                "default"
+            ]
+        )
+        self._test_conn_autocommits(conn, False)
+
+    @testing.requires.independent_readonly_connections
+    @testing.variation("use_dialect_setting", [True, False])
+    def test_dialect_autocommit_is_restored(
+        self, testing_engine, use_dialect_setting
+    ):
+        """test #10147"""
+
+        if use_dialect_setting:
+            e = testing_engine(options={"isolation_level": "AUTOCOMMIT"})
+        else:
+            e = testing_engine().execution_options(
+                isolation_level="AUTOCOMMIT"
+            )
+
+        levels = requirements.get_isolation_levels(config)
+
+        default = levels["default"]
+
+        with e.connect() as conn:
+            self._test_conn_autocommits(conn, True)
+
+        with e.connect() as conn:
+            conn.execution_options(isolation_level=default)
+            self._test_conn_autocommits(conn, False)
+
+        with e.connect() as conn:
+            self._test_conn_autocommits(conn, True)
+
+
+class EscapingTest(fixtures.TestBase):
+    @provide_metadata
+    def test_percent_sign_round_trip(self):
+        """test that the DBAPI accommodates for escaped / nonescaped
+        percent signs in a way that matches the compiler
+
+        """
+        m = self.metadata
+        t = Table("t", m, Column("data", String(50)))
+        t.create(config.db)
+        with config.db.begin() as conn:
+            conn.execute(t.insert(), dict(data="some % value"))
+            conn.execute(t.insert(), dict(data="some %% other value"))
+
+            eq_(
+                conn.scalar(
+                    select(t.c.data).where(
+                        t.c.data == literal_column("'some % value'")
+                    )
+                ),
+                "some % value",
+            )
+
+            eq_(
+                conn.scalar(
+                    select(t.c.data).where(
+                        t.c.data == literal_column("'some %% other value'")
+                    )
+                ),
+                "some %% other value",
+            )
+
+
+class WeCanSetDefaultSchemaWEventsTest(fixtures.TestBase):
+    __backend__ = True
+
+    __requires__ = ("default_schema_name_switch",)
+
+    def test_control_case(self):
+        default_schema_name = config.db.dialect.default_schema_name
+
+        eng = engines.testing_engine()
+        with eng.connect():
+            pass
+
+        eq_(eng.dialect.default_schema_name, default_schema_name)
+
+    def test_wont_work_wo_insert(self):
+        default_schema_name = config.db.dialect.default_schema_name
+
+        eng = engines.testing_engine()
+
+        @event.listens_for(eng, "connect")
+        def on_connect(dbapi_connection, connection_record):
+            set_default_schema_on_connection(
+                config, dbapi_connection, config.test_schema
+            )
+
+        with eng.connect() as conn:
+            what_it_should_be = eng.dialect._get_default_schema_name(conn)
+            eq_(what_it_should_be, config.test_schema)
+
+        eq_(eng.dialect.default_schema_name, default_schema_name)
+
+    def test_schema_change_on_connect(self):
+        eng = engines.testing_engine()
+
+        @event.listens_for(eng, "connect", insert=True)
+        def on_connect(dbapi_connection, connection_record):
+            set_default_schema_on_connection(
+                config, dbapi_connection, config.test_schema
+            )
+
+        with eng.connect() as conn:
+            what_it_should_be = eng.dialect._get_default_schema_name(conn)
+            eq_(what_it_should_be, config.test_schema)
+
+        eq_(eng.dialect.default_schema_name, config.test_schema)
+
+    def test_schema_change_works_w_transactions(self):
+        eng = engines.testing_engine()
+
+        @event.listens_for(eng, "connect", insert=True)
+        def on_connect(dbapi_connection, *arg):
+            set_default_schema_on_connection(
+                config, dbapi_connection, config.test_schema
+            )
+
+        with eng.connect() as conn:
+            trans = conn.begin()
+            what_it_should_be = eng.dialect._get_default_schema_name(conn)
+            eq_(what_it_should_be, config.test_schema)
+            trans.rollback()
+
+            what_it_should_be = eng.dialect._get_default_schema_name(conn)
+            eq_(what_it_should_be, config.test_schema)
+
+        eq_(eng.dialect.default_schema_name, config.test_schema)
+
+
+class FutureWeCanSetDefaultSchemaWEventsTest(
+    fixtures.FutureEngineMixin, WeCanSetDefaultSchemaWEventsTest
+):
+    pass
+
+
+class DifficultParametersTest(fixtures.TestBase):
+    __backend__ = True
+
+    tough_parameters = testing.combinations(
+        ("boring",),
+        ("per cent",),
+        ("per % cent",),
+        ("%percent",),
+        ("par(ens)",),
+        ("percent%(ens)yah",),
+        ("col:ons",),
+        ("_starts_with_underscore",),
+        ("dot.s",),
+        ("more :: %colons%",),
+        ("_name",),
+        ("___name",),
+        ("[BracketsAndCase]",),
+        ("42numbers",),
+        ("percent%signs",),
+        ("has spaces",),
+        ("/slashes/",),
+        ("more/slashes",),
+        ("q?marks",),
+        ("1param",),
+        ("1col:on",),
+        argnames="paramname",
+    )
+
+    @tough_parameters
+    @config.requirements.unusual_column_name_characters
+    def test_round_trip_same_named_column(
+        self, paramname, connection, metadata
+    ):
+        name = paramname
+
+        t = Table(
+            "t",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column(name, String(50), nullable=False),
+        )
+
+        # table is created
+        t.create(connection)
+
+        # automatic param generated by insert
+        connection.execute(t.insert().values({"id": 1, name: "some name"}))
+
+        # automatic param generated by criteria, plus selecting the column
+        stmt = select(t.c[name]).where(t.c[name] == "some name")
+
+        eq_(connection.scalar(stmt), "some name")
+
+        # use the name in a param explicitly
+        stmt = select(t.c[name]).where(t.c[name] == bindparam(name))
+
+        row = connection.execute(stmt, {name: "some name"}).first()
+
+        # name works as the key from cursor.description
+        eq_(row._mapping[name], "some name")
+
+        # use expanding IN
+        stmt = select(t.c[name]).where(
+            t.c[name].in_(["some name", "some other_name"])
+        )
+
+        row = connection.execute(stmt).first()
+
+    @testing.fixture
+    def multirow_fixture(self, metadata, connection):
+        mytable = Table(
+            "mytable",
+            metadata,
+            Column("myid", Integer),
+            Column("name", String(50)),
+            Column("desc", String(50)),
+        )
+
+        mytable.create(connection)
+
+        connection.execute(
+            mytable.insert(),
+            [
+                {"myid": 1, "name": "a", "desc": "a_desc"},
+                {"myid": 2, "name": "b", "desc": "b_desc"},
+                {"myid": 3, "name": "c", "desc": "c_desc"},
+                {"myid": 4, "name": "d", "desc": "d_desc"},
+            ],
+        )
+        yield mytable
+
+    @tough_parameters
+    def test_standalone_bindparam_escape(
+        self, paramname, connection, multirow_fixture
+    ):
+        tbl1 = multirow_fixture
+        stmt = select(tbl1.c.myid).where(
+            tbl1.c.name == bindparam(paramname, value="x")
+        )
+        res = connection.scalar(stmt, {paramname: "c"})
+        eq_(res, 3)
+
+    @tough_parameters
+    def test_standalone_bindparam_escape_expanding(
+        self, paramname, connection, multirow_fixture
+    ):
+        tbl1 = multirow_fixture
+        stmt = (
+            select(tbl1.c.myid)
+            .where(tbl1.c.name.in_(bindparam(paramname, value=["a", "b"])))
+            .order_by(tbl1.c.myid)
+        )
+
+        res = connection.scalars(stmt, {paramname: ["d", "a"]}).all()
+        eq_(res, [1, 4])
+
+
+class ReturningGuardsTest(fixtures.TablesTest):
+    """test that the various 'returning' flags are set appropriately"""
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "t",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+        )
+
+    @testing.fixture
+    def run_stmt(self, connection):
+        t = self.tables.t
+
+        def go(stmt, executemany, id_param_name, expect_success):
+            stmt = stmt.returning(t.c.id)
+
+            if executemany:
+                if not expect_success:
+                    # for RETURNING executemany(), we raise our own
+                    # error as this is independent of general RETURNING
+                    # support
+                    with expect_raises_message(
+                        exc.StatementError,
+                        rf"Dialect {connection.dialect.name}\+"
+                        f"{connection.dialect.driver} with "
+                        f"current server capabilities does not support "
+                        f".*RETURNING when executemany is used",
+                    ):
+                        result = connection.execute(
+                            stmt,
+                            [
+                                {id_param_name: 1, "data": "d1"},
+                                {id_param_name: 2, "data": "d2"},
+                                {id_param_name: 3, "data": "d3"},
+                            ],
+                        )
+                else:
+                    result = connection.execute(
+                        stmt,
+                        [
+                            {id_param_name: 1, "data": "d1"},
+                            {id_param_name: 2, "data": "d2"},
+                            {id_param_name: 3, "data": "d3"},
+                        ],
+                    )
+                    eq_(result.all(), [(1,), (2,), (3,)])
+            else:
+                if not expect_success:
+                    # for RETURNING execute(), we pass all the way to the DB
+                    # and let it fail
+                    with expect_raises(exc.DBAPIError):
+                        connection.execute(
+                            stmt, {id_param_name: 1, "data": "d1"}
+                        )
+                else:
+                    result = connection.execute(
+                        stmt, {id_param_name: 1, "data": "d1"}
+                    )
+                    eq_(result.all(), [(1,)])
+
+        return go
+
+    def test_insert_single(self, connection, run_stmt):
+        t = self.tables.t
+
+        stmt = t.insert()
+
+        run_stmt(stmt, False, "id", connection.dialect.insert_returning)
+
+    def test_insert_many(self, connection, run_stmt):
+        t = self.tables.t
+
+        stmt = t.insert()
+
+        run_stmt(
+            stmt, True, "id", connection.dialect.insert_executemany_returning
+        )
+
+    def test_update_single(self, connection, run_stmt):
+        t = self.tables.t
+
+        connection.execute(
+            t.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+        stmt = t.update().where(t.c.id == bindparam("b_id"))
+
+        run_stmt(stmt, False, "b_id", connection.dialect.update_returning)
+
+    def test_update_many(self, connection, run_stmt):
+        t = self.tables.t
+
+        connection.execute(
+            t.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+        stmt = t.update().where(t.c.id == bindparam("b_id"))
+
+        run_stmt(
+            stmt, True, "b_id", connection.dialect.update_executemany_returning
+        )
+
+    def test_delete_single(self, connection, run_stmt):
+        t = self.tables.t
+
+        connection.execute(
+            t.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+        stmt = t.delete().where(t.c.id == bindparam("b_id"))
+
+        run_stmt(stmt, False, "b_id", connection.dialect.delete_returning)
+
+    def test_delete_many(self, connection, run_stmt):
+        t = self.tables.t
+
+        connection.execute(
+            t.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+        stmt = t.delete().where(t.c.id == bindparam("b_id"))
+
+        run_stmt(
+            stmt, True, "b_id", connection.dialect.delete_executemany_returning
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_insert.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_insert.py
new file mode 100644
index 00000000..8467c351
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_insert.py
@@ -0,0 +1,630 @@
+# testing/suite/test_insert.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from decimal import Decimal
+import uuid
+
+from . import testing
+from .. import fixtures
+from ..assertions import eq_
+from ..config import requirements
+from ..schema import Column
+from ..schema import Table
+from ... import Double
+from ... import Float
+from ... import Identity
+from ... import Integer
+from ... import literal
+from ... import literal_column
+from ... import Numeric
+from ... import select
+from ... import String
+from ...types import LargeBinary
+from ...types import UUID
+from ...types import Uuid
+
+
+class LastrowidTest(fixtures.TablesTest):
+    run_deletes = "each"
+
+    __backend__ = True
+
+    __requires__ = "implements_get_lastrowid", "autoincrement_insert"
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "autoinc_pk",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+            implicit_returning=False,
+        )
+
+        Table(
+            "manual_pk",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+            implicit_returning=False,
+        )
+
+    def _assert_round_trip(self, table, conn):
+        row = conn.execute(table.select()).first()
+        eq_(
+            row,
+            (
+                conn.dialect.default_sequence_base,
+                "some data",
+            ),
+        )
+
+    def test_autoincrement_on_insert(self, connection):
+        connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        self._assert_round_trip(self.tables.autoinc_pk, connection)
+
+    def test_last_inserted_id(self, connection):
+        r = connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        pk = connection.scalar(select(self.tables.autoinc_pk.c.id))
+        eq_(r.inserted_primary_key, (pk,))
+
+    @requirements.dbapi_lastrowid
+    def test_native_lastrowid_autoinc(self, connection):
+        r = connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        lastrowid = r.lastrowid
+        pk = connection.scalar(select(self.tables.autoinc_pk.c.id))
+        eq_(lastrowid, pk)
+
+
+class InsertBehaviorTest(fixtures.TablesTest):
+    run_deletes = "each"
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "autoinc_pk",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+        )
+        Table(
+            "manual_pk",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("data", String(50)),
+        )
+        Table(
+            "no_implicit_returning",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+            implicit_returning=False,
+        )
+        Table(
+            "includes_defaults",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+            Column("x", Integer, default=5),
+            Column(
+                "y",
+                Integer,
+                default=literal_column("2", type_=Integer) + literal(2),
+            ),
+        )
+
+    @testing.variation("style", ["plain", "return_defaults"])
+    @testing.variation("executemany", [True, False])
+    def test_no_results_for_non_returning_insert(
+        self, connection, style, executemany
+    ):
+        """test another INSERT issue found during #10453"""
+
+        table = self.tables.no_implicit_returning
+
+        stmt = table.insert()
+        if style.return_defaults:
+            stmt = stmt.return_defaults()
+
+        if executemany:
+            data = [
+                {"data": "d1"},
+                {"data": "d2"},
+                {"data": "d3"},
+                {"data": "d4"},
+                {"data": "d5"},
+            ]
+        else:
+            data = {"data": "d1"}
+
+        r = connection.execute(stmt, data)
+        assert not r.returns_rows
+
+    @requirements.autoincrement_insert
+    def test_autoclose_on_insert(self, connection):
+        r = connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        assert r._soft_closed
+        assert not r.closed
+        assert r.is_insert
+
+        # new as of I8091919d45421e3f53029b8660427f844fee0228; for the moment
+        # an insert where the PK was taken from a row that the dialect
+        # selected, as is the case for mssql/pyodbc, will still report
+        # returns_rows as true because there's a cursor description.  in that
+        # case, the row had to have been consumed at least.
+        assert not r.returns_rows or r.fetchone() is None
+
+    @requirements.insert_returning
+    def test_autoclose_on_insert_implicit_returning(self, connection):
+        r = connection.execute(
+            # return_defaults() ensures RETURNING will be used,
+            # new in 2.0 as sqlite/mariadb offer both RETURNING and
+            # cursor.lastrowid
+            self.tables.autoinc_pk.insert().return_defaults(),
+            dict(data="some data"),
+        )
+        assert r._soft_closed
+        assert not r.closed
+        assert r.is_insert
+
+        # note we are experimenting with having this be True
+        # as of I8091919d45421e3f53029b8660427f844fee0228 .
+        # implicit returning has fetched the row, but it still is a
+        # "returns rows"
+        assert r.returns_rows
+
+        # and we should be able to fetchone() on it, we just get no row
+        eq_(r.fetchone(), None)
+
+        # and the keys, etc.
+        eq_(r.keys(), ["id"])
+
+        # but the dialect took in the row already.   not really sure
+        # what the best behavior is.
+
+    @requirements.empty_inserts
+    def test_empty_insert(self, connection):
+        r = connection.execute(self.tables.autoinc_pk.insert())
+        assert r._soft_closed
+        assert not r.closed
+
+        r = connection.execute(
+            self.tables.autoinc_pk.select().where(
+                self.tables.autoinc_pk.c.id != None
+            )
+        )
+        eq_(len(r.all()), 1)
+
+    @requirements.empty_inserts_executemany
+    def test_empty_insert_multiple(self, connection):
+        r = connection.execute(self.tables.autoinc_pk.insert(), [{}, {}, {}])
+        assert r._soft_closed
+        assert not r.closed
+
+        r = connection.execute(
+            self.tables.autoinc_pk.select().where(
+                self.tables.autoinc_pk.c.id != None
+            )
+        )
+
+        eq_(len(r.all()), 3)
+
+    @requirements.insert_from_select
+    def test_insert_from_select_autoinc(self, connection):
+        src_table = self.tables.manual_pk
+        dest_table = self.tables.autoinc_pk
+        connection.execute(
+            src_table.insert(),
+            [
+                dict(id=1, data="data1"),
+                dict(id=2, data="data2"),
+                dict(id=3, data="data3"),
+            ],
+        )
+
+        result = connection.execute(
+            dest_table.insert().from_select(
+                ("data",),
+                select(src_table.c.data).where(
+                    src_table.c.data.in_(["data2", "data3"])
+                ),
+            )
+        )
+
+        eq_(result.inserted_primary_key, (None,))
+
+        result = connection.execute(
+            select(dest_table.c.data).order_by(dest_table.c.data)
+        )
+        eq_(result.fetchall(), [("data2",), ("data3",)])
+
+    @requirements.insert_from_select
+    def test_insert_from_select_autoinc_no_rows(self, connection):
+        src_table = self.tables.manual_pk
+        dest_table = self.tables.autoinc_pk
+
+        result = connection.execute(
+            dest_table.insert().from_select(
+                ("data",),
+                select(src_table.c.data).where(
+                    src_table.c.data.in_(["data2", "data3"])
+                ),
+            )
+        )
+        eq_(result.inserted_primary_key, (None,))
+
+        result = connection.execute(
+            select(dest_table.c.data).order_by(dest_table.c.data)
+        )
+
+        eq_(result.fetchall(), [])
+
+    @requirements.insert_from_select
+    def test_insert_from_select(self, connection):
+        table = self.tables.manual_pk
+        connection.execute(
+            table.insert(),
+            [
+                dict(id=1, data="data1"),
+                dict(id=2, data="data2"),
+                dict(id=3, data="data3"),
+            ],
+        )
+
+        connection.execute(
+            table.insert()
+            .inline()
+            .from_select(
+                ("id", "data"),
+                select(table.c.id + 5, table.c.data).where(
+                    table.c.data.in_(["data2", "data3"])
+                ),
+            )
+        )
+
+        eq_(
+            connection.execute(
+                select(table.c.data).order_by(table.c.data)
+            ).fetchall(),
+            [("data1",), ("data2",), ("data2",), ("data3",), ("data3",)],
+        )
+
+    @requirements.insert_from_select
+    def test_insert_from_select_with_defaults(self, connection):
+        table = self.tables.includes_defaults
+        connection.execute(
+            table.insert(),
+            [
+                dict(id=1, data="data1"),
+                dict(id=2, data="data2"),
+                dict(id=3, data="data3"),
+            ],
+        )
+
+        connection.execute(
+            table.insert()
+            .inline()
+            .from_select(
+                ("id", "data"),
+                select(table.c.id + 5, table.c.data).where(
+                    table.c.data.in_(["data2", "data3"])
+                ),
+            )
+        )
+
+        eq_(
+            connection.execute(
+                select(table).order_by(table.c.data, table.c.id)
+            ).fetchall(),
+            [
+                (1, "data1", 5, 4),
+                (2, "data2", 5, 4),
+                (7, "data2", 5, 4),
+                (3, "data3", 5, 4),
+                (8, "data3", 5, 4),
+            ],
+        )
+
+
+class ReturningTest(fixtures.TablesTest):
+    run_create_tables = "each"
+    __requires__ = "insert_returning", "autoincrement_insert"
+    __backend__ = True
+
+    def _assert_round_trip(self, table, conn):
+        row = conn.execute(table.select()).first()
+        eq_(
+            row,
+            (
+                conn.dialect.default_sequence_base,
+                "some data",
+            ),
+        )
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "autoinc_pk",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+        )
+
+    @requirements.fetch_rows_post_commit
+    def test_explicit_returning_pk_autocommit(self, connection):
+        table = self.tables.autoinc_pk
+        r = connection.execute(
+            table.insert().returning(table.c.id), dict(data="some data")
+        )
+        pk = r.first()[0]
+        fetched_pk = connection.scalar(select(table.c.id))
+        eq_(fetched_pk, pk)
+
+    def test_explicit_returning_pk_no_autocommit(self, connection):
+        table = self.tables.autoinc_pk
+        r = connection.execute(
+            table.insert().returning(table.c.id), dict(data="some data")
+        )
+
+        pk = r.first()[0]
+        fetched_pk = connection.scalar(select(table.c.id))
+        eq_(fetched_pk, pk)
+
+    def test_autoincrement_on_insert_implicit_returning(self, connection):
+        connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        self._assert_round_trip(self.tables.autoinc_pk, connection)
+
+    def test_last_inserted_id_implicit_returning(self, connection):
+        r = connection.execute(
+            self.tables.autoinc_pk.insert(), dict(data="some data")
+        )
+        pk = connection.scalar(select(self.tables.autoinc_pk.c.id))
+        eq_(r.inserted_primary_key, (pk,))
+
+    @requirements.insert_executemany_returning
+    def test_insertmanyvalues_returning(self, connection):
+        r = connection.execute(
+            self.tables.autoinc_pk.insert().returning(
+                self.tables.autoinc_pk.c.id
+            ),
+            [
+                {"data": "d1"},
+                {"data": "d2"},
+                {"data": "d3"},
+                {"data": "d4"},
+                {"data": "d5"},
+            ],
+        )
+        rall = r.all()
+
+        pks = connection.execute(select(self.tables.autoinc_pk.c.id))
+
+        eq_(rall, pks.all())
+
+    @testing.combinations(
+        (Double(), 8.5514716, True),
+        (
+            Double(53),
+            8.5514716,
+            True,
+            testing.requires.float_or_double_precision_behaves_generically,
+        ),
+        (Float(), 8.5514, True),
+        (
+            Float(8),
+            8.5514,
+            True,
+            testing.requires.float_or_double_precision_behaves_generically,
+        ),
+        (
+            Numeric(precision=15, scale=12, asdecimal=False),
+            8.5514716,
+            True,
+            testing.requires.literal_float_coercion,
+        ),
+        (
+            Numeric(precision=15, scale=12, asdecimal=True),
+            Decimal("8.5514716"),
+            False,
+        ),
+        argnames="type_,value,do_rounding",
+    )
+    @testing.variation("sort_by_parameter_order", [True, False])
+    @testing.variation("multiple_rows", [True, False])
+    def test_insert_w_floats(
+        self,
+        connection,
+        metadata,
+        sort_by_parameter_order,
+        type_,
+        value,
+        do_rounding,
+        multiple_rows,
+    ):
+        """test #9701.
+
+        this tests insertmanyvalues as well as decimal / floating point
+        RETURNING types
+
+        """
+
+        t = Table(
+            # Oracle backends seems to be getting confused if
+            # this table is named the same as the one
+            # in test_imv_returning_datatypes.  use a different name
+            "f_t",
+            metadata,
+            Column("id", Integer, Identity(), primary_key=True),
+            Column("value", type_),
+        )
+
+        t.create(connection)
+
+        result = connection.execute(
+            t.insert().returning(
+                t.c.id,
+                t.c.value,
+                sort_by_parameter_order=bool(sort_by_parameter_order),
+            ),
+            (
+                [{"value": value} for i in range(10)]
+                if multiple_rows
+                else {"value": value}
+            ),
+        )
+
+        if multiple_rows:
+            i_range = range(1, 11)
+        else:
+            i_range = range(1, 2)
+
+        # we want to test only that we are getting floating points back
+        # with some degree of the original value maintained, that it is not
+        # being truncated to an integer.  there's too much variation in how
+        # drivers return floats, which should not be relied upon to be
+        # exact, for us to just compare as is (works for PG drivers but not
+        # others) so we use rounding here.  There's precedent for this
+        # in suite/test_types.py::NumericTest as well
+
+        if do_rounding:
+            eq_(
+                {(id_, round(val_, 5)) for id_, val_ in result},
+                {(id_, round(value, 5)) for id_ in i_range},
+            )
+
+            eq_(
+                {
+                    round(val_, 5)
+                    for val_ in connection.scalars(select(t.c.value))
+                },
+                {round(value, 5)},
+            )
+        else:
+            eq_(
+                set(result),
+                {(id_, value) for id_ in i_range},
+            )
+
+            eq_(
+                set(connection.scalars(select(t.c.value))),
+                {value},
+            )
+
+    @testing.combinations(
+        (
+            "non_native_uuid",
+            Uuid(native_uuid=False),
+            uuid.uuid4(),
+        ),
+        (
+            "non_native_uuid_str",
+            Uuid(as_uuid=False, native_uuid=False),
+            str(uuid.uuid4()),
+        ),
+        (
+            "generic_native_uuid",
+            Uuid(native_uuid=True),
+            uuid.uuid4(),
+            testing.requires.uuid_data_type,
+        ),
+        (
+            "generic_native_uuid_str",
+            Uuid(as_uuid=False, native_uuid=True),
+            str(uuid.uuid4()),
+            testing.requires.uuid_data_type,
+        ),
+        ("UUID", UUID(), uuid.uuid4(), testing.requires.uuid_data_type),
+        (
+            "LargeBinary1",
+            LargeBinary(),
+            b"this is binary",
+        ),
+        ("LargeBinary2", LargeBinary(), b"7\xe7\x9f"),
+        argnames="type_,value",
+        id_="iaa",
+    )
+    @testing.variation("sort_by_parameter_order", [True, False])
+    @testing.variation("multiple_rows", [True, False])
+    @testing.requires.insert_returning
+    def test_imv_returning_datatypes(
+        self,
+        connection,
+        metadata,
+        sort_by_parameter_order,
+        type_,
+        value,
+        multiple_rows,
+    ):
+        """test #9739, #9808 (similar to #9701).
+
+        this tests insertmanyvalues in conjunction with various datatypes.
+
+        These tests are particularly for the asyncpg driver which needs
+        most types to be explicitly cast for the new IMV format
+
+        """
+        t = Table(
+            "d_t",
+            metadata,
+            Column("id", Integer, Identity(), primary_key=True),
+            Column("value", type_),
+        )
+
+        t.create(connection)
+
+        result = connection.execute(
+            t.insert().returning(
+                t.c.id,
+                t.c.value,
+                sort_by_parameter_order=bool(sort_by_parameter_order),
+            ),
+            (
+                [{"value": value} for i in range(10)]
+                if multiple_rows
+                else {"value": value}
+            ),
+        )
+
+        if multiple_rows:
+            i_range = range(1, 11)
+        else:
+            i_range = range(1, 2)
+
+        eq_(
+            set(result),
+            {(id_, value) for id_ in i_range},
+        )
+
+        eq_(
+            set(connection.scalars(select(t.c.value))),
+            {value},
+        )
+
+
+__all__ = ("LastrowidTest", "InsertBehaviorTest", "ReturningTest")
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_reflection.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_reflection.py
new file mode 100644
index 00000000..54d0d449
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_reflection.py
@@ -0,0 +1,3225 @@
+# testing/suite/test_reflection.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import contextlib
+import operator
+import re
+
+import sqlalchemy as sa
+from .. import config
+from .. import engines
+from .. import eq_
+from .. import expect_raises
+from .. import expect_raises_message
+from .. import expect_warnings
+from .. import fixtures
+from .. import is_
+from ..provision import get_temp_table_name
+from ..provision import temp_table_keyword_args
+from ..schema import Column
+from ..schema import Table
+from ... import event
+from ... import ForeignKey
+from ... import func
+from ... import Identity
+from ... import inspect
+from ... import Integer
+from ... import MetaData
+from ... import String
+from ... import testing
+from ... import types as sql_types
+from ...engine import Inspector
+from ...engine import ObjectKind
+from ...engine import ObjectScope
+from ...exc import NoSuchTableError
+from ...exc import UnreflectableTableError
+from ...schema import DDL
+from ...schema import Index
+from ...sql.elements import quoted_name
+from ...sql.schema import BLANK_SCHEMA
+from ...testing import ComparesIndexes
+from ...testing import ComparesTables
+from ...testing import is_false
+from ...testing import is_true
+from ...testing import mock
+
+
+metadata, users = None, None
+
+
+class OneConnectionTablesTest(fixtures.TablesTest):
+    @classmethod
+    def setup_bind(cls):
+        # TODO: when temp tables are subject to server reset,
+        # this will also have to disable that server reset from
+        # happening
+        if config.requirements.independent_connections.enabled:
+            from sqlalchemy import pool
+
+            return engines.testing_engine(
+                options=dict(poolclass=pool.StaticPool, scope="class"),
+            )
+        else:
+            return config.db
+
+
+class HasTableTest(OneConnectionTablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "test_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+        )
+        if testing.requires.schemas.enabled:
+            Table(
+                "test_table_s",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("data", String(50)),
+                schema=config.test_schema,
+            )
+
+        if testing.requires.view_reflection:
+            cls.define_views(metadata)
+        if testing.requires.has_temp_table.enabled:
+            cls.define_temp_tables(metadata)
+
+    @classmethod
+    def define_views(cls, metadata):
+        query = "CREATE VIEW vv AS SELECT id, data FROM test_table"
+
+        event.listen(metadata, "after_create", DDL(query))
+        event.listen(metadata, "before_drop", DDL("DROP VIEW vv"))
+
+        if testing.requires.schemas.enabled:
+            query = (
+                "CREATE VIEW %s.vv AS SELECT id, data FROM %s.test_table_s"
+                % (
+                    config.test_schema,
+                    config.test_schema,
+                )
+            )
+            event.listen(metadata, "after_create", DDL(query))
+            event.listen(
+                metadata,
+                "before_drop",
+                DDL("DROP VIEW %s.vv" % (config.test_schema)),
+            )
+
+    @classmethod
+    def temp_table_name(cls):
+        return get_temp_table_name(
+            config, config.db, f"user_tmp_{config.ident}"
+        )
+
+    @classmethod
+    def define_temp_tables(cls, metadata):
+        kw = temp_table_keyword_args(config, config.db)
+        table_name = cls.temp_table_name()
+        user_tmp = Table(
+            table_name,
+            metadata,
+            Column("id", sa.INT, primary_key=True),
+            Column("name", sa.VARCHAR(50)),
+            **kw,
+        )
+        if (
+            testing.requires.view_reflection.enabled
+            and testing.requires.temporary_views.enabled
+        ):
+            event.listen(
+                user_tmp,
+                "after_create",
+                DDL(
+                    "create temporary view user_tmp_v as "
+                    "select * from user_tmp_%s" % config.ident
+                ),
+            )
+            event.listen(user_tmp, "before_drop", DDL("drop view user_tmp_v"))
+
+    def test_has_table(self):
+        with config.db.begin() as conn:
+            is_true(config.db.dialect.has_table(conn, "test_table"))
+            is_false(config.db.dialect.has_table(conn, "test_table_s"))
+            is_false(config.db.dialect.has_table(conn, "nonexistent_table"))
+
+    def test_has_table_cache(self, metadata):
+        insp = inspect(config.db)
+        is_true(insp.has_table("test_table"))
+        nt = Table("new_table", metadata, Column("col", Integer))
+        is_false(insp.has_table("new_table"))
+        nt.create(config.db)
+        try:
+            is_false(insp.has_table("new_table"))
+            insp.clear_cache()
+            is_true(insp.has_table("new_table"))
+        finally:
+            nt.drop(config.db)
+
+    @testing.requires.schemas
+    def test_has_table_schema(self):
+        with config.db.begin() as conn:
+            is_false(
+                config.db.dialect.has_table(
+                    conn, "test_table", schema=config.test_schema
+                )
+            )
+            is_true(
+                config.db.dialect.has_table(
+                    conn, "test_table_s", schema=config.test_schema
+                )
+            )
+            is_false(
+                config.db.dialect.has_table(
+                    conn, "nonexistent_table", schema=config.test_schema
+                )
+            )
+
+    @testing.requires.schemas
+    def test_has_table_nonexistent_schema(self):
+        with config.db.begin() as conn:
+            is_false(
+                config.db.dialect.has_table(
+                    conn, "test_table", schema="nonexistent_schema"
+                )
+            )
+
+    @testing.requires.views
+    def test_has_table_view(self, connection):
+        insp = inspect(connection)
+        is_true(insp.has_table("vv"))
+
+    @testing.requires.has_temp_table
+    def test_has_table_temp_table(self, connection):
+        insp = inspect(connection)
+        temp_table_name = self.temp_table_name()
+        is_true(insp.has_table(temp_table_name))
+
+    @testing.requires.has_temp_table
+    @testing.requires.view_reflection
+    @testing.requires.temporary_views
+    def test_has_table_temp_view(self, connection):
+        insp = inspect(connection)
+        is_true(insp.has_table("user_tmp_v"))
+
+    @testing.requires.views
+    @testing.requires.schemas
+    def test_has_table_view_schema(self, connection):
+        insp = inspect(connection)
+        is_true(insp.has_table("vv", config.test_schema))
+
+
+class HasIndexTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        tt = Table(
+            "test_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+            Column("data2", String(50)),
+        )
+        Index("my_idx", tt.c.data)
+
+        if testing.requires.schemas.enabled:
+            tt = Table(
+                "test_table",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("data", String(50)),
+                schema=config.test_schema,
+            )
+            Index("my_idx_s", tt.c.data)
+
+    kind = testing.combinations("dialect", "inspector", argnames="kind")
+
+    def _has_index(self, kind, conn):
+        if kind == "dialect":
+            return lambda *a, **k: config.db.dialect.has_index(conn, *a, **k)
+        else:
+            return inspect(conn).has_index
+
+    @kind
+    def test_has_index(self, kind, connection, metadata):
+        meth = self._has_index(kind, connection)
+        assert meth("test_table", "my_idx")
+        assert not meth("test_table", "my_idx_s")
+        assert not meth("nonexistent_table", "my_idx")
+        assert not meth("test_table", "nonexistent_idx")
+
+        assert not meth("test_table", "my_idx_2")
+        assert not meth("test_table_2", "my_idx_3")
+        idx = Index("my_idx_2", self.tables.test_table.c.data2)
+        tbl = Table(
+            "test_table_2",
+            metadata,
+            Column("foo", Integer),
+            Index("my_idx_3", "foo"),
+        )
+        idx.create(connection)
+        tbl.create(connection)
+        try:
+            if kind == "inspector":
+                assert not meth("test_table", "my_idx_2")
+                assert not meth("test_table_2", "my_idx_3")
+                meth.__self__.clear_cache()
+            assert meth("test_table", "my_idx_2") is True
+            assert meth("test_table_2", "my_idx_3") is True
+        finally:
+            tbl.drop(connection)
+            idx.drop(connection)
+
+    @testing.requires.schemas
+    @kind
+    def test_has_index_schema(self, kind, connection):
+        meth = self._has_index(kind, connection)
+        assert meth("test_table", "my_idx_s", schema=config.test_schema)
+        assert not meth("test_table", "my_idx", schema=config.test_schema)
+        assert not meth(
+            "nonexistent_table", "my_idx_s", schema=config.test_schema
+        )
+        assert not meth(
+            "test_table", "nonexistent_idx_s", schema=config.test_schema
+        )
+
+
+class BizarroCharacterFKResolutionTest(fixtures.TestBase):
+    """tests for #10275"""
+
+    __backend__ = True
+
+    @testing.combinations(
+        ("id",), ("(3)",), ("col%p",), ("[brack]",), argnames="columnname"
+    )
+    @testing.variation("use_composite", [True, False])
+    @testing.combinations(
+        ("plain",),
+        ("(2)",),
+        ("per % cent",),
+        ("[brackets]",),
+        argnames="tablename",
+    )
+    def test_fk_ref(
+        self, connection, metadata, use_composite, tablename, columnname
+    ):
+        tt = Table(
+            tablename,
+            metadata,
+            Column(columnname, Integer, key="id", primary_key=True),
+            test_needs_fk=True,
+        )
+        if use_composite:
+            tt.append_column(Column("id2", Integer, primary_key=True))
+
+        if use_composite:
+            Table(
+                "other",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("ref", Integer),
+                Column("ref2", Integer),
+                sa.ForeignKeyConstraint(["ref", "ref2"], [tt.c.id, tt.c.id2]),
+                test_needs_fk=True,
+            )
+        else:
+            Table(
+                "other",
+                metadata,
+                Column("id", Integer, primary_key=True),
+                Column("ref", ForeignKey(tt.c.id)),
+                test_needs_fk=True,
+            )
+
+        metadata.create_all(connection)
+
+        m2 = MetaData()
+
+        o2 = Table("other", m2, autoload_with=connection)
+        t1 = m2.tables[tablename]
+
+        assert o2.c.ref.references(t1.c[0])
+        if use_composite:
+            assert o2.c.ref2.references(t1.c[1])
+
+
+class QuotedNameArgumentTest(fixtures.TablesTest):
+    run_create_tables = "once"
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "quote ' one",
+            metadata,
+            Column("id", Integer),
+            Column("name", String(50)),
+            Column("data", String(50)),
+            Column("related_id", Integer),
+            sa.PrimaryKeyConstraint("id", name="pk quote ' one"),
+            sa.Index("ix quote ' one", "name"),
+            sa.UniqueConstraint(
+                "data",
+                name="uq quote' one",
+            ),
+            sa.ForeignKeyConstraint(
+                ["id"], ["related.id"], name="fk quote ' one"
+            ),
+            sa.CheckConstraint("name != 'foo'", name="ck quote ' one"),
+            comment=r"""quote ' one comment""",
+            test_needs_fk=True,
+        )
+
+        if testing.requires.symbol_names_w_double_quote.enabled:
+            Table(
+                'quote " two',
+                metadata,
+                Column("id", Integer),
+                Column("name", String(50)),
+                Column("data", String(50)),
+                Column("related_id", Integer),
+                sa.PrimaryKeyConstraint("id", name='pk quote " two'),
+                sa.Index('ix quote " two', "name"),
+                sa.UniqueConstraint(
+                    "data",
+                    name='uq quote" two',
+                ),
+                sa.ForeignKeyConstraint(
+                    ["id"], ["related.id"], name='fk quote " two'
+                ),
+                sa.CheckConstraint("name != 'foo'", name='ck quote " two '),
+                comment=r"""quote " two comment""",
+                test_needs_fk=True,
+            )
+
+        Table(
+            "related",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("related", Integer),
+            test_needs_fk=True,
+        )
+
+        if testing.requires.view_column_reflection.enabled:
+            if testing.requires.symbol_names_w_double_quote.enabled:
+                names = [
+                    "quote ' one",
+                    'quote " two',
+                ]
+            else:
+                names = [
+                    "quote ' one",
+                ]
+            for name in names:
+                query = "CREATE VIEW %s AS SELECT * FROM %s" % (
+                    config.db.dialect.identifier_preparer.quote(
+                        "view %s" % name
+                    ),
+                    config.db.dialect.identifier_preparer.quote(name),
+                )
+
+                event.listen(metadata, "after_create", DDL(query))
+                event.listen(
+                    metadata,
+                    "before_drop",
+                    DDL(
+                        "DROP VIEW %s"
+                        % config.db.dialect.identifier_preparer.quote(
+                            "view %s" % name
+                        )
+                    ),
+                )
+
+    def quote_fixtures(fn):
+        return testing.combinations(
+            ("quote ' one",),
+            ('quote " two', testing.requires.symbol_names_w_double_quote),
+        )(fn)
+
+    @quote_fixtures
+    def test_get_table_options(self, name):
+        insp = inspect(config.db)
+
+        if testing.requires.reflect_table_options.enabled:
+            res = insp.get_table_options(name)
+            is_true(isinstance(res, dict))
+        else:
+            with expect_raises(NotImplementedError):
+                res = insp.get_table_options(name)
+
+    @quote_fixtures
+    @testing.requires.view_column_reflection
+    def test_get_view_definition(self, name):
+        insp = inspect(config.db)
+        assert insp.get_view_definition("view %s" % name)
+
+    @quote_fixtures
+    def test_get_columns(self, name):
+        insp = inspect(config.db)
+        assert insp.get_columns(name)
+
+    @quote_fixtures
+    def test_get_pk_constraint(self, name):
+        insp = inspect(config.db)
+        assert insp.get_pk_constraint(name)
+
+    @quote_fixtures
+    def test_get_foreign_keys(self, name):
+        insp = inspect(config.db)
+        assert insp.get_foreign_keys(name)
+
+    @quote_fixtures
+    def test_get_indexes(self, name):
+        insp = inspect(config.db)
+        assert insp.get_indexes(name)
+
+    @quote_fixtures
+    @testing.requires.unique_constraint_reflection
+    def test_get_unique_constraints(self, name):
+        insp = inspect(config.db)
+        assert insp.get_unique_constraints(name)
+
+    @quote_fixtures
+    @testing.requires.comment_reflection
+    def test_get_table_comment(self, name):
+        insp = inspect(config.db)
+        assert insp.get_table_comment(name)
+
+    @quote_fixtures
+    @testing.requires.check_constraint_reflection
+    def test_get_check_constraints(self, name):
+        insp = inspect(config.db)
+        assert insp.get_check_constraints(name)
+
+
+def _multi_combination(fn):
+    schema = testing.combinations(
+        None,
+        (
+            lambda: config.test_schema,
+            testing.requires.schemas,
+        ),
+        argnames="schema",
+    )
+    scope = testing.combinations(
+        ObjectScope.DEFAULT,
+        ObjectScope.TEMPORARY,
+        ObjectScope.ANY,
+        argnames="scope",
+    )
+    kind = testing.combinations(
+        ObjectKind.TABLE,
+        ObjectKind.VIEW,
+        ObjectKind.MATERIALIZED_VIEW,
+        ObjectKind.ANY,
+        ObjectKind.ANY_VIEW,
+        ObjectKind.TABLE | ObjectKind.VIEW,
+        ObjectKind.TABLE | ObjectKind.MATERIALIZED_VIEW,
+        argnames="kind",
+    )
+    filter_names = testing.combinations(True, False, argnames="use_filter")
+
+    return schema(scope(kind(filter_names(fn))))
+
+
+class ComponentReflectionTest(ComparesTables, OneConnectionTablesTest):
+    run_inserts = run_deletes = None
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        cls.define_reflected_tables(metadata, None)
+        if testing.requires.schemas.enabled:
+            cls.define_reflected_tables(metadata, testing.config.test_schema)
+
+    @classmethod
+    def define_reflected_tables(cls, metadata, schema):
+        if schema:
+            schema_prefix = schema + "."
+        else:
+            schema_prefix = ""
+
+        if testing.requires.self_referential_foreign_keys.enabled:
+            parent_id_args = (
+                ForeignKey(
+                    "%susers.user_id" % schema_prefix, name="user_id_fk"
+                ),
+            )
+        else:
+            parent_id_args = ()
+        users = Table(
+            "users",
+            metadata,
+            Column("user_id", sa.INT, primary_key=True),
+            Column("test1", sa.CHAR(5), nullable=False),
+            Column("test2", sa.Float(), nullable=False),
+            Column("parent_user_id", sa.Integer, *parent_id_args),
+            sa.CheckConstraint(
+                "test2 > 0",
+                name="zz_test2_gt_zero",
+                comment="users check constraint",
+            ),
+            sa.CheckConstraint("test2 <= 1000"),
+            schema=schema,
+            test_needs_fk=True,
+        )
+
+        Table(
+            "dingalings",
+            metadata,
+            Column("dingaling_id", sa.Integer, primary_key=True),
+            Column(
+                "address_id",
+                sa.Integer,
+                ForeignKey(
+                    "%semail_addresses.address_id" % schema_prefix,
+                    name="zz_email_add_id_fg",
+                    comment="di fk comment",
+                ),
+            ),
+            Column(
+                "id_user",
+                sa.Integer,
+                ForeignKey("%susers.user_id" % schema_prefix),
+            ),
+            Column("data", sa.String(30), unique=True),
+            sa.CheckConstraint(
+                "address_id > 0 AND address_id < 1000",
+                name="address_id_gt_zero",
+            ),
+            sa.UniqueConstraint(
+                "address_id",
+                "dingaling_id",
+                name="zz_dingalings_multiple",
+                comment="di unique comment",
+            ),
+            schema=schema,
+            test_needs_fk=True,
+        )
+        Table(
+            "email_addresses",
+            metadata,
+            Column("address_id", sa.Integer),
+            Column("remote_user_id", sa.Integer, ForeignKey(users.c.user_id)),
+            Column("email_address", sa.String(20), index=True),
+            sa.PrimaryKeyConstraint(
+                "address_id", name="email_ad_pk", comment="ea pk comment"
+            ),
+            schema=schema,
+            test_needs_fk=True,
+        )
+        Table(
+            "comment_test",
+            metadata,
+            Column("id", sa.Integer, primary_key=True, comment="id comment"),
+            Column("data", sa.String(20), comment="data % comment"),
+            Column(
+                "d2",
+                sa.String(20),
+                comment=r"""Comment types type speedily ' " \ '' Fun!""",
+            ),
+            Column("d3", sa.String(42), comment="Comment\nwith\rescapes"),
+            schema=schema,
+            comment=r"""the test % ' " \ table comment""",
+        )
+        Table(
+            "no_constraints",
+            metadata,
+            Column("data", sa.String(20)),
+            schema=schema,
+            comment="no\nconstraints\rhas\fescaped\vcomment",
+        )
+
+        if testing.requires.cross_schema_fk_reflection.enabled:
+            if schema is None:
+                Table(
+                    "local_table",
+                    metadata,
+                    Column("id", sa.Integer, primary_key=True),
+                    Column("data", sa.String(20)),
+                    Column(
+                        "remote_id",
+                        ForeignKey(
+                            "%s.remote_table_2.id" % testing.config.test_schema
+                        ),
+                    ),
+                    test_needs_fk=True,
+                    schema=config.db.dialect.default_schema_name,
+                )
+            else:
+                Table(
+                    "remote_table",
+                    metadata,
+                    Column("id", sa.Integer, primary_key=True),
+                    Column(
+                        "local_id",
+                        ForeignKey(
+                            "%s.local_table.id"
+                            % config.db.dialect.default_schema_name
+                        ),
+                    ),
+                    Column("data", sa.String(20)),
+                    schema=schema,
+                    test_needs_fk=True,
+                )
+                Table(
+                    "remote_table_2",
+                    metadata,
+                    Column("id", sa.Integer, primary_key=True),
+                    Column("data", sa.String(20)),
+                    schema=schema,
+                    test_needs_fk=True,
+                )
+
+        if testing.requires.index_reflection.enabled:
+            Index("users_t_idx", users.c.test1, users.c.test2, unique=True)
+            Index(
+                "users_all_idx", users.c.user_id, users.c.test2, users.c.test1
+            )
+
+            if not schema:
+                # test_needs_fk is at the moment to force MySQL InnoDB
+                noncol_idx_test_nopk = Table(
+                    "noncol_idx_test_nopk",
+                    metadata,
+                    Column("q", sa.String(5)),
+                    test_needs_fk=True,
+                )
+
+                noncol_idx_test_pk = Table(
+                    "noncol_idx_test_pk",
+                    metadata,
+                    Column("id", sa.Integer, primary_key=True),
+                    Column("q", sa.String(5)),
+                    test_needs_fk=True,
+                )
+
+                if (
+                    testing.requires.indexes_with_ascdesc.enabled
+                    and testing.requires.reflect_indexes_with_ascdesc.enabled
+                ):
+                    Index("noncol_idx_nopk", noncol_idx_test_nopk.c.q.desc())
+                    Index("noncol_idx_pk", noncol_idx_test_pk.c.q.desc())
+
+        if testing.requires.view_column_reflection.enabled:
+            cls.define_views(metadata, schema)
+        if not schema and testing.requires.temp_table_reflection.enabled:
+            cls.define_temp_tables(metadata)
+
+    @classmethod
+    def temp_table_name(cls):
+        return get_temp_table_name(
+            config, config.db, f"user_tmp_{config.ident}"
+        )
+
+    @classmethod
+    def define_temp_tables(cls, metadata):
+        kw = temp_table_keyword_args(config, config.db)
+        table_name = cls.temp_table_name()
+        user_tmp = Table(
+            table_name,
+            metadata,
+            Column("id", sa.INT, primary_key=True),
+            Column("name", sa.VARCHAR(50)),
+            Column("foo", sa.INT),
+            # disambiguate temp table unique constraint names.  this is
+            # pretty arbitrary for a generic dialect however we are doing
+            # it to suit SQL Server which will produce name conflicts for
+            # unique constraints created against temp tables in different
+            # databases.
+            # https://www.arbinada.com/en/node/1645
+            sa.UniqueConstraint("name", name=f"user_tmp_uq_{config.ident}"),
+            sa.Index("user_tmp_ix", "foo"),
+            **kw,
+        )
+        if (
+            testing.requires.view_reflection.enabled
+            and testing.requires.temporary_views.enabled
+        ):
+            event.listen(
+                user_tmp,
+                "after_create",
+                DDL(
+                    "create temporary view user_tmp_v as "
+                    "select * from user_tmp_%s" % config.ident
+                ),
+            )
+            event.listen(user_tmp, "before_drop", DDL("drop view user_tmp_v"))
+
+    @classmethod
+    def define_views(cls, metadata, schema):
+        if testing.requires.materialized_views.enabled:
+            materialized = {"dingalings"}
+        else:
+            materialized = set()
+        for table_name in ("users", "email_addresses", "dingalings"):
+            fullname = table_name
+            if schema:
+                fullname = f"{schema}.{table_name}"
+            view_name = fullname + "_v"
+            prefix = "MATERIALIZED " if table_name in materialized else ""
+            query = (
+                f"CREATE {prefix}VIEW {view_name} AS SELECT * FROM {fullname}"
+            )
+
+            event.listen(metadata, "after_create", DDL(query))
+            if table_name in materialized:
+                index_name = "mat_index"
+                if schema and testing.against("oracle"):
+                    index_name = f"{schema}.{index_name}"
+                idx = f"CREATE INDEX {index_name} ON {view_name}(data)"
+                event.listen(metadata, "after_create", DDL(idx))
+            event.listen(
+                metadata, "before_drop", DDL(f"DROP {prefix}VIEW {view_name}")
+            )
+
+    def _resolve_kind(self, kind, tables, views, materialized):
+        res = {}
+        if ObjectKind.TABLE in kind:
+            res.update(tables)
+        if ObjectKind.VIEW in kind:
+            res.update(views)
+        if ObjectKind.MATERIALIZED_VIEW in kind:
+            res.update(materialized)
+        return res
+
+    def _resolve_views(self, views, materialized):
+        if not testing.requires.view_column_reflection.enabled:
+            materialized.clear()
+            views.clear()
+        elif not testing.requires.materialized_views.enabled:
+            views.update(materialized)
+            materialized.clear()
+
+    def _resolve_names(self, schema, scope, filter_names, values):
+        scope_filter = lambda _: True  # noqa: E731
+        if scope is ObjectScope.DEFAULT:
+            scope_filter = lambda k: "tmp" not in k[1]  # noqa: E731
+        if scope is ObjectScope.TEMPORARY:
+            scope_filter = lambda k: "tmp" in k[1]  # noqa: E731
+
+        removed = {
+            None: {"remote_table", "remote_table_2"},
+            testing.config.test_schema: {
+                "local_table",
+                "noncol_idx_test_nopk",
+                "noncol_idx_test_pk",
+                "user_tmp_v",
+                self.temp_table_name(),
+            },
+        }
+        if not testing.requires.cross_schema_fk_reflection.enabled:
+            removed[None].add("local_table")
+            removed[testing.config.test_schema].update(
+                ["remote_table", "remote_table_2"]
+            )
+        if not testing.requires.index_reflection.enabled:
+            removed[None].update(
+                ["noncol_idx_test_nopk", "noncol_idx_test_pk"]
+            )
+        if (
+            not testing.requires.temp_table_reflection.enabled
+            or not testing.requires.temp_table_names.enabled
+        ):
+            removed[None].update(["user_tmp_v", self.temp_table_name()])
+        if not testing.requires.temporary_views.enabled:
+            removed[None].update(["user_tmp_v"])
+
+        res = {
+            k: v
+            for k, v in values.items()
+            if scope_filter(k)
+            and k[1] not in removed[schema]
+            and (not filter_names or k[1] in filter_names)
+        }
+        return res
+
+    def exp_options(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        materialized = {(schema, "dingalings_v"): mock.ANY}
+        views = {
+            (schema, "email_addresses_v"): mock.ANY,
+            (schema, "users_v"): mock.ANY,
+            (schema, "user_tmp_v"): mock.ANY,
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): mock.ANY,
+            (schema, "dingalings"): mock.ANY,
+            (schema, "email_addresses"): mock.ANY,
+            (schema, "comment_test"): mock.ANY,
+            (schema, "no_constraints"): mock.ANY,
+            (schema, "local_table"): mock.ANY,
+            (schema, "remote_table"): mock.ANY,
+            (schema, "remote_table_2"): mock.ANY,
+            (schema, "noncol_idx_test_nopk"): mock.ANY,
+            (schema, "noncol_idx_test_pk"): mock.ANY,
+            (schema, self.temp_table_name()): mock.ANY,
+        }
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    def exp_comments(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        empty = {"text": None}
+        materialized = {(schema, "dingalings_v"): empty}
+        views = {
+            (schema, "email_addresses_v"): empty,
+            (schema, "users_v"): empty,
+            (schema, "user_tmp_v"): empty,
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): empty,
+            (schema, "dingalings"): empty,
+            (schema, "email_addresses"): empty,
+            (schema, "comment_test"): {
+                "text": r"""the test % ' " \ table comment"""
+            },
+            (schema, "no_constraints"): {
+                "text": "no\nconstraints\rhas\fescaped\vcomment"
+            },
+            (schema, "local_table"): empty,
+            (schema, "remote_table"): empty,
+            (schema, "remote_table_2"): empty,
+            (schema, "noncol_idx_test_nopk"): empty,
+            (schema, "noncol_idx_test_pk"): empty,
+            (schema, self.temp_table_name()): empty,
+        }
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    def exp_columns(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        def col(
+            name, auto=False, default=mock.ANY, comment=None, nullable=True
+        ):
+            res = {
+                "name": name,
+                "autoincrement": auto,
+                "type": mock.ANY,
+                "default": default,
+                "comment": comment,
+                "nullable": nullable,
+            }
+            if auto == "omit":
+                res.pop("autoincrement")
+            return res
+
+        def pk(name, **kw):
+            kw = {"auto": True, "default": mock.ANY, "nullable": False, **kw}
+            return col(name, **kw)
+
+        materialized = {
+            (schema, "dingalings_v"): [
+                col("dingaling_id", auto="omit", nullable=mock.ANY),
+                col("address_id"),
+                col("id_user"),
+                col("data"),
+            ]
+        }
+        views = {
+            (schema, "email_addresses_v"): [
+                col("address_id", auto="omit", nullable=mock.ANY),
+                col("remote_user_id"),
+                col("email_address"),
+            ],
+            (schema, "users_v"): [
+                col("user_id", auto="omit", nullable=mock.ANY),
+                col("test1", nullable=mock.ANY),
+                col("test2", nullable=mock.ANY),
+                col("parent_user_id"),
+            ],
+            (schema, "user_tmp_v"): [
+                col("id", auto="omit", nullable=mock.ANY),
+                col("name"),
+                col("foo"),
+            ],
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): [
+                pk("user_id"),
+                col("test1", nullable=False),
+                col("test2", nullable=False),
+                col("parent_user_id"),
+            ],
+            (schema, "dingalings"): [
+                pk("dingaling_id"),
+                col("address_id"),
+                col("id_user"),
+                col("data"),
+            ],
+            (schema, "email_addresses"): [
+                pk("address_id"),
+                col("remote_user_id"),
+                col("email_address"),
+            ],
+            (schema, "comment_test"): [
+                pk("id", comment="id comment"),
+                col("data", comment="data % comment"),
+                col(
+                    "d2",
+                    comment=r"""Comment types type speedily ' " \ '' Fun!""",
+                ),
+                col("d3", comment="Comment\nwith\rescapes"),
+            ],
+            (schema, "no_constraints"): [col("data")],
+            (schema, "local_table"): [pk("id"), col("data"), col("remote_id")],
+            (schema, "remote_table"): [pk("id"), col("local_id"), col("data")],
+            (schema, "remote_table_2"): [pk("id"), col("data")],
+            (schema, "noncol_idx_test_nopk"): [col("q")],
+            (schema, "noncol_idx_test_pk"): [pk("id"), col("q")],
+            (schema, self.temp_table_name()): [
+                pk("id"),
+                col("name"),
+                col("foo"),
+            ],
+        }
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    @property
+    def _required_column_keys(self):
+        return {"name", "type", "nullable", "default"}
+
+    def exp_pks(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        def pk(*cols, name=mock.ANY, comment=None):
+            return {
+                "constrained_columns": list(cols),
+                "name": name,
+                "comment": comment,
+            }
+
+        empty = pk(name=None)
+        if testing.requires.materialized_views_reflect_pk.enabled:
+            materialized = {(schema, "dingalings_v"): pk("dingaling_id")}
+        else:
+            materialized = {(schema, "dingalings_v"): empty}
+        views = {
+            (schema, "email_addresses_v"): empty,
+            (schema, "users_v"): empty,
+            (schema, "user_tmp_v"): empty,
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): pk("user_id"),
+            (schema, "dingalings"): pk("dingaling_id"),
+            (schema, "email_addresses"): pk(
+                "address_id", name="email_ad_pk", comment="ea pk comment"
+            ),
+            (schema, "comment_test"): pk("id"),
+            (schema, "no_constraints"): empty,
+            (schema, "local_table"): pk("id"),
+            (schema, "remote_table"): pk("id"),
+            (schema, "remote_table_2"): pk("id"),
+            (schema, "noncol_idx_test_nopk"): empty,
+            (schema, "noncol_idx_test_pk"): pk("id"),
+            (schema, self.temp_table_name()): pk("id"),
+        }
+        if not testing.requires.reflects_pk_names.enabled:
+            for val in tables.values():
+                if val["name"] is not None:
+                    val["name"] = mock.ANY
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    @property
+    def _required_pk_keys(self):
+        return {"name", "constrained_columns"}
+
+    def exp_fks(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        class tt:
+            def __eq__(self, other):
+                return (
+                    other is None
+                    or config.db.dialect.default_schema_name == other
+                )
+
+        def fk(
+            cols,
+            ref_col,
+            ref_table,
+            ref_schema=schema,
+            name=mock.ANY,
+            comment=None,
+        ):
+            return {
+                "constrained_columns": cols,
+                "referred_columns": ref_col,
+                "name": name,
+                "options": mock.ANY,
+                "referred_schema": (
+                    ref_schema if ref_schema is not None else tt()
+                ),
+                "referred_table": ref_table,
+                "comment": comment,
+            }
+
+        materialized = {(schema, "dingalings_v"): []}
+        views = {
+            (schema, "email_addresses_v"): [],
+            (schema, "users_v"): [],
+            (schema, "user_tmp_v"): [],
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): [
+                fk(["parent_user_id"], ["user_id"], "users", name="user_id_fk")
+            ],
+            (schema, "dingalings"): [
+                fk(["id_user"], ["user_id"], "users"),
+                fk(
+                    ["address_id"],
+                    ["address_id"],
+                    "email_addresses",
+                    name="zz_email_add_id_fg",
+                    comment="di fk comment",
+                ),
+            ],
+            (schema, "email_addresses"): [
+                fk(["remote_user_id"], ["user_id"], "users")
+            ],
+            (schema, "comment_test"): [],
+            (schema, "no_constraints"): [],
+            (schema, "local_table"): [
+                fk(
+                    ["remote_id"],
+                    ["id"],
+                    "remote_table_2",
+                    ref_schema=config.test_schema,
+                )
+            ],
+            (schema, "remote_table"): [
+                fk(["local_id"], ["id"], "local_table", ref_schema=None)
+            ],
+            (schema, "remote_table_2"): [],
+            (schema, "noncol_idx_test_nopk"): [],
+            (schema, "noncol_idx_test_pk"): [],
+            (schema, self.temp_table_name()): [],
+        }
+        if not testing.requires.self_referential_foreign_keys.enabled:
+            tables[(schema, "users")].clear()
+        if not testing.requires.named_constraints.enabled:
+            for vals in tables.values():
+                for val in vals:
+                    if val["name"] is not mock.ANY:
+                        val["name"] = mock.ANY
+
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    @property
+    def _required_fk_keys(self):
+        return {
+            "name",
+            "constrained_columns",
+            "referred_schema",
+            "referred_table",
+            "referred_columns",
+        }
+
+    def exp_indexes(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        def idx(
+            *cols,
+            name,
+            unique=False,
+            column_sorting=None,
+            duplicates=False,
+            fk=False,
+        ):
+            fk_req = testing.requires.foreign_keys_reflect_as_index
+            dup_req = testing.requires.unique_constraints_reflect_as_index
+            sorting_expression = (
+                testing.requires.reflect_indexes_with_ascdesc_as_expression
+            )
+
+            if (fk and not fk_req.enabled) or (
+                duplicates and not dup_req.enabled
+            ):
+                return ()
+            res = {
+                "unique": unique,
+                "column_names": list(cols),
+                "name": name,
+                "dialect_options": mock.ANY,
+                "include_columns": [],
+            }
+            if column_sorting:
+                res["column_sorting"] = column_sorting
+                if sorting_expression.enabled:
+                    res["expressions"] = orig = res["column_names"]
+                    res["column_names"] = [
+                        None if c in column_sorting else c for c in orig
+                    ]
+
+            if duplicates:
+                res["duplicates_constraint"] = name
+            return [res]
+
+        materialized = {(schema, "dingalings_v"): []}
+        views = {
+            (schema, "email_addresses_v"): [],
+            (schema, "users_v"): [],
+            (schema, "user_tmp_v"): [],
+        }
+        self._resolve_views(views, materialized)
+        if materialized:
+            materialized[(schema, "dingalings_v")].extend(
+                idx("data", name="mat_index")
+            )
+        tables = {
+            (schema, "users"): [
+                *idx("parent_user_id", name="user_id_fk", fk=True),
+                *idx("user_id", "test2", "test1", name="users_all_idx"),
+                *idx("test1", "test2", name="users_t_idx", unique=True),
+            ],
+            (schema, "dingalings"): [
+                *idx("data", name=mock.ANY, unique=True, duplicates=True),
+                *idx("id_user", name=mock.ANY, fk=True),
+                *idx(
+                    "address_id",
+                    "dingaling_id",
+                    name="zz_dingalings_multiple",
+                    unique=True,
+                    duplicates=True,
+                ),
+            ],
+            (schema, "email_addresses"): [
+                *idx("email_address", name=mock.ANY),
+                *idx("remote_user_id", name=mock.ANY, fk=True),
+            ],
+            (schema, "comment_test"): [],
+            (schema, "no_constraints"): [],
+            (schema, "local_table"): [
+                *idx("remote_id", name=mock.ANY, fk=True)
+            ],
+            (schema, "remote_table"): [
+                *idx("local_id", name=mock.ANY, fk=True)
+            ],
+            (schema, "remote_table_2"): [],
+            (schema, "noncol_idx_test_nopk"): [
+                *idx(
+                    "q",
+                    name="noncol_idx_nopk",
+                    column_sorting={"q": ("desc",)},
+                )
+            ],
+            (schema, "noncol_idx_test_pk"): [
+                *idx(
+                    "q", name="noncol_idx_pk", column_sorting={"q": ("desc",)}
+                )
+            ],
+            (schema, self.temp_table_name()): [
+                *idx("foo", name="user_tmp_ix"),
+                *idx(
+                    "name",
+                    name=f"user_tmp_uq_{config.ident}",
+                    duplicates=True,
+                    unique=True,
+                ),
+            ],
+        }
+        if (
+            not testing.requires.indexes_with_ascdesc.enabled
+            or not testing.requires.reflect_indexes_with_ascdesc.enabled
+        ):
+            tables[(schema, "noncol_idx_test_nopk")].clear()
+            tables[(schema, "noncol_idx_test_pk")].clear()
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    @property
+    def _required_index_keys(self):
+        return {"name", "column_names", "unique"}
+
+    def exp_ucs(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+        all_=False,
+    ):
+        def uc(
+            *cols, name, duplicates_index=None, is_index=False, comment=None
+        ):
+            req = testing.requires.unique_index_reflect_as_unique_constraints
+            if is_index and not req.enabled:
+                return ()
+            res = {
+                "column_names": list(cols),
+                "name": name,
+                "comment": comment,
+            }
+            if duplicates_index:
+                res["duplicates_index"] = duplicates_index
+            return [res]
+
+        materialized = {(schema, "dingalings_v"): []}
+        views = {
+            (schema, "email_addresses_v"): [],
+            (schema, "users_v"): [],
+            (schema, "user_tmp_v"): [],
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): [
+                *uc(
+                    "test1",
+                    "test2",
+                    name="users_t_idx",
+                    duplicates_index="users_t_idx",
+                    is_index=True,
+                )
+            ],
+            (schema, "dingalings"): [
+                *uc("data", name=mock.ANY, duplicates_index=mock.ANY),
+                *uc(
+                    "address_id",
+                    "dingaling_id",
+                    name="zz_dingalings_multiple",
+                    duplicates_index="zz_dingalings_multiple",
+                    comment="di unique comment",
+                ),
+            ],
+            (schema, "email_addresses"): [],
+            (schema, "comment_test"): [],
+            (schema, "no_constraints"): [],
+            (schema, "local_table"): [],
+            (schema, "remote_table"): [],
+            (schema, "remote_table_2"): [],
+            (schema, "noncol_idx_test_nopk"): [],
+            (schema, "noncol_idx_test_pk"): [],
+            (schema, self.temp_table_name()): [
+                *uc("name", name=f"user_tmp_uq_{config.ident}")
+            ],
+        }
+        if all_:
+            return {**materialized, **views, **tables}
+        else:
+            res = self._resolve_kind(kind, tables, views, materialized)
+            res = self._resolve_names(schema, scope, filter_names, res)
+            return res
+
+    @property
+    def _required_unique_cst_keys(self):
+        return {"name", "column_names"}
+
+    def exp_ccs(
+        self,
+        schema=None,
+        scope=ObjectScope.ANY,
+        kind=ObjectKind.ANY,
+        filter_names=None,
+    ):
+        class tt(str):
+            def __eq__(self, other):
+                res = (
+                    other.lower()
+                    .replace("(", "")
+                    .replace(")", "")
+                    .replace("`", "")
+                )
+                return self in res
+
+        def cc(text, name, comment=None):
+            return {"sqltext": tt(text), "name": name, "comment": comment}
+
+        # print({1: "test2 > (0)::double precision"} == {1: tt("test2 > 0")})
+        # assert 0
+        materialized = {(schema, "dingalings_v"): []}
+        views = {
+            (schema, "email_addresses_v"): [],
+            (schema, "users_v"): [],
+            (schema, "user_tmp_v"): [],
+        }
+        self._resolve_views(views, materialized)
+        tables = {
+            (schema, "users"): [
+                cc("test2 <= 1000", mock.ANY),
+                cc(
+                    "test2 > 0",
+                    "zz_test2_gt_zero",
+                    comment="users check constraint",
+                ),
+            ],
+            (schema, "dingalings"): [
+                cc(
+                    "address_id > 0 and address_id < 1000",
+                    name="address_id_gt_zero",
+                ),
+            ],
+            (schema, "email_addresses"): [],
+            (schema, "comment_test"): [],
+            (schema, "no_constraints"): [],
+            (schema, "local_table"): [],
+            (schema, "remote_table"): [],
+            (schema, "remote_table_2"): [],
+            (schema, "noncol_idx_test_nopk"): [],
+            (schema, "noncol_idx_test_pk"): [],
+            (schema, self.temp_table_name()): [],
+        }
+        res = self._resolve_kind(kind, tables, views, materialized)
+        res = self._resolve_names(schema, scope, filter_names, res)
+        return res
+
+    @property
+    def _required_cc_keys(self):
+        return {"name", "sqltext"}
+
+    @testing.requires.schema_reflection
+    def test_get_schema_names(self, connection):
+        insp = inspect(connection)
+
+        is_true(testing.config.test_schema in insp.get_schema_names())
+
+    @testing.requires.schema_reflection
+    def test_has_schema(self, connection):
+        insp = inspect(connection)
+
+        is_true(insp.has_schema(testing.config.test_schema))
+        is_false(insp.has_schema("sa_fake_schema_foo"))
+
+    @testing.requires.schema_reflection
+    def test_get_schema_names_w_translate_map(self, connection):
+        """test #7300"""
+
+        connection = connection.execution_options(
+            schema_translate_map={
+                "foo": "bar",
+                BLANK_SCHEMA: testing.config.test_schema,
+            }
+        )
+        insp = inspect(connection)
+
+        is_true(testing.config.test_schema in insp.get_schema_names())
+
+    @testing.requires.schema_reflection
+    def test_has_schema_w_translate_map(self, connection):
+        connection = connection.execution_options(
+            schema_translate_map={
+                "foo": "bar",
+                BLANK_SCHEMA: testing.config.test_schema,
+            }
+        )
+        insp = inspect(connection)
+
+        is_true(insp.has_schema(testing.config.test_schema))
+        is_false(insp.has_schema("sa_fake_schema_foo"))
+
+    @testing.requires.schema_reflection
+    @testing.requires.schema_create_delete
+    def test_schema_cache(self, connection):
+        insp = inspect(connection)
+
+        is_false("foo_bar" in insp.get_schema_names())
+        is_false(insp.has_schema("foo_bar"))
+        connection.execute(DDL("CREATE SCHEMA foo_bar"))
+        try:
+            is_false("foo_bar" in insp.get_schema_names())
+            is_false(insp.has_schema("foo_bar"))
+            insp.clear_cache()
+            is_true("foo_bar" in insp.get_schema_names())
+            is_true(insp.has_schema("foo_bar"))
+        finally:
+            connection.execute(DDL("DROP SCHEMA foo_bar"))
+
+    @testing.requires.schema_reflection
+    def test_dialect_initialize(self):
+        engine = engines.testing_engine()
+        inspect(engine)
+        assert hasattr(engine.dialect, "default_schema_name")
+
+    @testing.requires.schema_reflection
+    def test_get_default_schema_name(self, connection):
+        insp = inspect(connection)
+        eq_(insp.default_schema_name, connection.dialect.default_schema_name)
+
+    @testing.combinations(
+        None,
+        ("foreign_key", testing.requires.foreign_key_constraint_reflection),
+        argnames="order_by",
+    )
+    @testing.combinations(
+        (True, testing.requires.schemas), False, argnames="use_schema"
+    )
+    def test_get_table_names(self, connection, order_by, use_schema):
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+
+        _ignore_tables = {
+            "comment_test",
+            "noncol_idx_test_pk",
+            "noncol_idx_test_nopk",
+            "local_table",
+            "remote_table",
+            "remote_table_2",
+            "no_constraints",
+        }
+
+        insp = inspect(connection)
+
+        if order_by:
+            tables = [
+                rec[0]
+                for rec in insp.get_sorted_table_and_fkc_names(schema)
+                if rec[0]
+            ]
+        else:
+            tables = insp.get_table_names(schema)
+        table_names = [t for t in tables if t not in _ignore_tables]
+
+        if order_by == "foreign_key":
+            answer = ["users", "email_addresses", "dingalings"]
+            eq_(table_names, answer)
+        else:
+            answer = ["dingalings", "email_addresses", "users"]
+            eq_(sorted(table_names), answer)
+
+    @testing.combinations(
+        (True, testing.requires.schemas), False, argnames="use_schema"
+    )
+    def test_get_view_names(self, connection, use_schema):
+        insp = inspect(connection)
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+        table_names = insp.get_view_names(schema)
+        if testing.requires.materialized_views.enabled:
+            eq_(sorted(table_names), ["email_addresses_v", "users_v"])
+            eq_(insp.get_materialized_view_names(schema), ["dingalings_v"])
+        else:
+            answer = ["dingalings_v", "email_addresses_v", "users_v"]
+            eq_(sorted(table_names), answer)
+
+    @testing.requires.temp_table_names
+    def test_get_temp_table_names(self, connection):
+        insp = inspect(connection)
+        temp_table_names = insp.get_temp_table_names()
+        eq_(sorted(temp_table_names), [f"user_tmp_{config.ident}"])
+
+    @testing.requires.view_reflection
+    @testing.requires.temporary_views
+    def test_get_temp_view_names(self, connection):
+        insp = inspect(connection)
+        temp_table_names = insp.get_temp_view_names()
+        eq_(sorted(temp_table_names), ["user_tmp_v"])
+
+    @testing.requires.comment_reflection
+    def test_get_comments(self, connection):
+        self._test_get_comments(connection)
+
+    @testing.requires.comment_reflection
+    @testing.requires.schemas
+    def test_get_comments_with_schema(self, connection):
+        self._test_get_comments(connection, testing.config.test_schema)
+
+    def _test_get_comments(self, connection, schema=None):
+        insp = inspect(connection)
+        exp = self.exp_comments(schema=schema)
+        eq_(
+            insp.get_table_comment("comment_test", schema=schema),
+            exp[(schema, "comment_test")],
+        )
+
+        eq_(
+            insp.get_table_comment("users", schema=schema),
+            exp[(schema, "users")],
+        )
+
+        eq_(
+            insp.get_table_comment("comment_test", schema=schema),
+            exp[(schema, "comment_test")],
+        )
+
+        no_cst = self.tables.no_constraints.name
+        eq_(
+            insp.get_table_comment(no_cst, schema=schema),
+            exp[(schema, no_cst)],
+        )
+
+    @testing.combinations(
+        (False, False),
+        (False, True, testing.requires.schemas),
+        (True, False, testing.requires.view_reflection),
+        (
+            True,
+            True,
+            testing.requires.schemas + testing.requires.view_reflection,
+        ),
+        argnames="use_views,use_schema",
+    )
+    def test_get_columns(self, connection, use_views, use_schema):
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+
+        users, addresses = (self.tables.users, self.tables.email_addresses)
+        if use_views:
+            table_names = ["users_v", "email_addresses_v", "dingalings_v"]
+        else:
+            table_names = ["users", "email_addresses"]
+
+        insp = inspect(connection)
+        for table_name, table in zip(table_names, (users, addresses)):
+            schema_name = schema
+            cols = insp.get_columns(table_name, schema=schema_name)
+            is_true(len(cols) > 0, len(cols))
+
+            # should be in order
+
+            for i, col in enumerate(table.columns):
+                eq_(col.name, cols[i]["name"])
+                ctype = cols[i]["type"].__class__
+                ctype_def = col.type
+                if isinstance(ctype_def, sa.types.TypeEngine):
+                    ctype_def = ctype_def.__class__
+
+                # Oracle returns Date for DateTime.
+
+                if testing.against("oracle") and ctype_def in (
+                    sql_types.Date,
+                    sql_types.DateTime,
+                ):
+                    ctype_def = sql_types.Date
+
+                # assert that the desired type and return type share
+                # a base within one of the generic types.
+
+                is_true(
+                    len(
+                        set(ctype.__mro__)
+                        .intersection(ctype_def.__mro__)
+                        .intersection(
+                            [
+                                sql_types.Integer,
+                                sql_types.Numeric,
+                                sql_types.DateTime,
+                                sql_types.Date,
+                                sql_types.Time,
+                                sql_types.String,
+                                sql_types._Binary,
+                            ]
+                        )
+                    )
+                    > 0,
+                    "%s(%s), %s(%s)"
+                    % (col.name, col.type, cols[i]["name"], ctype),
+                )
+
+                if not col.primary_key:
+                    assert cols[i]["default"] is None
+
+        # The case of a table with no column
+        # is tested below in TableNoColumnsTest
+
+    @testing.requires.temp_table_reflection
+    def test_reflect_table_temp_table(self, connection):
+        table_name = self.temp_table_name()
+        user_tmp = self.tables[table_name]
+
+        reflected_user_tmp = Table(
+            table_name, MetaData(), autoload_with=connection
+        )
+        self.assert_tables_equal(
+            user_tmp, reflected_user_tmp, strict_constraints=False
+        )
+
+    @testing.requires.temp_table_reflection
+    def test_get_temp_table_columns(self, connection):
+        table_name = self.temp_table_name()
+        user_tmp = self.tables[table_name]
+        insp = inspect(connection)
+        cols = insp.get_columns(table_name)
+        is_true(len(cols) > 0, len(cols))
+
+        for i, col in enumerate(user_tmp.columns):
+            eq_(col.name, cols[i]["name"])
+
+    @testing.requires.temp_table_reflection
+    @testing.requires.view_column_reflection
+    @testing.requires.temporary_views
+    def test_get_temp_view_columns(self, connection):
+        insp = inspect(connection)
+        cols = insp.get_columns("user_tmp_v")
+        eq_([col["name"] for col in cols], ["id", "name", "foo"])
+
+    @testing.combinations(
+        (False,), (True, testing.requires.schemas), argnames="use_schema"
+    )
+    @testing.requires.primary_key_constraint_reflection
+    def test_get_pk_constraint(self, connection, use_schema):
+        if use_schema:
+            schema = testing.config.test_schema
+        else:
+            schema = None
+
+        users, addresses = self.tables.users, self.tables.email_addresses
+        insp = inspect(connection)
+        exp = self.exp_pks(schema=schema)
+
+        users_cons = insp.get_pk_constraint(users.name, schema=schema)
+        self._check_list(
+            [users_cons], [exp[(schema, users.name)]], self._required_pk_keys
+        )
+
+        addr_cons = insp.get_pk_constraint(addresses.name, schema=schema)
+        exp_cols = exp[(schema, addresses.name)]["constrained_columns"]
+        eq_(addr_cons["constrained_columns"], exp_cols)
+
+        with testing.requires.reflects_pk_names.fail_if():
+            eq_(addr_cons["name"], "email_ad_pk")
+
+        no_cst = self.tables.no_constraints.name
+        self._check_list(
+            [insp.get_pk_constraint(no_cst, schema=schema)],
+            [exp[(schema, no_cst)]],
+            self._required_pk_keys,
+        )
+
+    @testing.combinations(
+        (False,), (True, testing.requires.schemas), argnames="use_schema"
+    )
+    @testing.requires.foreign_key_constraint_reflection
+    def test_get_foreign_keys(self, connection, use_schema):
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+
+        users, addresses = (self.tables.users, self.tables.email_addresses)
+        insp = inspect(connection)
+        expected_schema = schema
+        # users
+
+        if testing.requires.self_referential_foreign_keys.enabled:
+            users_fkeys = insp.get_foreign_keys(users.name, schema=schema)
+            fkey1 = users_fkeys[0]
+
+            with testing.requires.named_constraints.fail_if():
+                eq_(fkey1["name"], "user_id_fk")
+
+            eq_(fkey1["referred_schema"], expected_schema)
+            eq_(fkey1["referred_table"], users.name)
+            eq_(fkey1["referred_columns"], ["user_id"])
+            eq_(fkey1["constrained_columns"], ["parent_user_id"])
+
+        # addresses
+        addr_fkeys = insp.get_foreign_keys(addresses.name, schema=schema)
+        fkey1 = addr_fkeys[0]
+
+        with testing.requires.implicitly_named_constraints.fail_if():
+            is_true(fkey1["name"] is not None)
+
+        eq_(fkey1["referred_schema"], expected_schema)
+        eq_(fkey1["referred_table"], users.name)
+        eq_(fkey1["referred_columns"], ["user_id"])
+        eq_(fkey1["constrained_columns"], ["remote_user_id"])
+
+        no_cst = self.tables.no_constraints.name
+        eq_(insp.get_foreign_keys(no_cst, schema=schema), [])
+
+    @testing.requires.cross_schema_fk_reflection
+    @testing.requires.schemas
+    def test_get_inter_schema_foreign_keys(self, connection):
+        local_table, remote_table, remote_table_2 = self.tables(
+            "%s.local_table" % connection.dialect.default_schema_name,
+            "%s.remote_table" % testing.config.test_schema,
+            "%s.remote_table_2" % testing.config.test_schema,
+        )
+
+        insp = inspect(connection)
+
+        local_fkeys = insp.get_foreign_keys(local_table.name)
+        eq_(len(local_fkeys), 1)
+
+        fkey1 = local_fkeys[0]
+        eq_(fkey1["referred_schema"], testing.config.test_schema)
+        eq_(fkey1["referred_table"], remote_table_2.name)
+        eq_(fkey1["referred_columns"], ["id"])
+        eq_(fkey1["constrained_columns"], ["remote_id"])
+
+        remote_fkeys = insp.get_foreign_keys(
+            remote_table.name, schema=testing.config.test_schema
+        )
+        eq_(len(remote_fkeys), 1)
+
+        fkey2 = remote_fkeys[0]
+
+        is_true(
+            fkey2["referred_schema"]
+            in (
+                None,
+                connection.dialect.default_schema_name,
+            )
+        )
+        eq_(fkey2["referred_table"], local_table.name)
+        eq_(fkey2["referred_columns"], ["id"])
+        eq_(fkey2["constrained_columns"], ["local_id"])
+
+    @testing.combinations(
+        (False,), (True, testing.requires.schemas), argnames="use_schema"
+    )
+    @testing.requires.index_reflection
+    def test_get_indexes(self, connection, use_schema):
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+
+        # The database may decide to create indexes for foreign keys, etc.
+        # so there may be more indexes than expected.
+        insp = inspect(connection)
+        indexes = insp.get_indexes("users", schema=schema)
+        exp = self.exp_indexes(schema=schema)
+        self._check_list(
+            indexes, exp[(schema, "users")], self._required_index_keys
+        )
+
+        no_cst = self.tables.no_constraints.name
+        self._check_list(
+            insp.get_indexes(no_cst, schema=schema),
+            exp[(schema, no_cst)],
+            self._required_index_keys,
+        )
+
+    @testing.combinations(
+        ("noncol_idx_test_nopk", "noncol_idx_nopk"),
+        ("noncol_idx_test_pk", "noncol_idx_pk"),
+        argnames="tname,ixname",
+    )
+    @testing.requires.index_reflection
+    @testing.requires.indexes_with_ascdesc
+    @testing.requires.reflect_indexes_with_ascdesc
+    def test_get_noncol_index(self, connection, tname, ixname):
+        insp = inspect(connection)
+        indexes = insp.get_indexes(tname)
+        # reflecting an index that has "x DESC" in it as the column.
+        # the DB may or may not give us "x", but make sure we get the index
+        # back, it has a name, it's connected to the table.
+        expected_indexes = self.exp_indexes()[(None, tname)]
+        self._check_list(indexes, expected_indexes, self._required_index_keys)
+
+        t = Table(tname, MetaData(), autoload_with=connection)
+        eq_(len(t.indexes), 1)
+        is_(list(t.indexes)[0].table, t)
+        eq_(list(t.indexes)[0].name, ixname)
+
+    @testing.requires.temp_table_reflection
+    @testing.requires.unique_constraint_reflection
+    def test_get_temp_table_unique_constraints(self, connection):
+        insp = inspect(connection)
+        name = self.temp_table_name()
+        reflected = insp.get_unique_constraints(name)
+        exp = self.exp_ucs(all_=True)[(None, name)]
+        self._check_list(reflected, exp, self._required_index_keys)
+
+    @testing.requires.temp_table_reflect_indexes
+    def test_get_temp_table_indexes(self, connection):
+        insp = inspect(connection)
+        table_name = self.temp_table_name()
+        indexes = insp.get_indexes(table_name)
+        for ind in indexes:
+            ind.pop("dialect_options", None)
+        expected = [
+            {"unique": False, "column_names": ["foo"], "name": "user_tmp_ix"}
+        ]
+        if testing.requires.index_reflects_included_columns.enabled:
+            expected[0]["include_columns"] = []
+        eq_(
+            [idx for idx in indexes if idx["name"] == "user_tmp_ix"],
+            expected,
+        )
+
+    @testing.combinations(
+        (True, testing.requires.schemas), (False,), argnames="use_schema"
+    )
+    @testing.requires.unique_constraint_reflection
+    def test_get_unique_constraints(self, metadata, connection, use_schema):
+        # SQLite dialect needs to parse the names of the constraints
+        # separately from what it gets from PRAGMA index_list(), and
+        # then matches them up.  so same set of column_names in two
+        # constraints will confuse it.    Perhaps we should no longer
+        # bother with index_list() here since we have the whole
+        # CREATE TABLE?
+
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+        uniques = sorted(
+            [
+                {"name": "unique_a", "column_names": ["a"]},
+                {"name": "unique_a_b_c", "column_names": ["a", "b", "c"]},
+                {"name": "unique_c_a_b", "column_names": ["c", "a", "b"]},
+                {"name": "unique_asc_key", "column_names": ["asc", "key"]},
+                {"name": "i.have.dots", "column_names": ["b"]},
+                {"name": "i have spaces", "column_names": ["c"]},
+            ],
+            key=operator.itemgetter("name"),
+        )
+        table = Table(
+            "testtbl",
+            metadata,
+            Column("a", sa.String(20)),
+            Column("b", sa.String(30)),
+            Column("c", sa.Integer),
+            # reserved identifiers
+            Column("asc", sa.String(30)),
+            Column("key", sa.String(30)),
+            schema=schema,
+        )
+        for uc in uniques:
+            table.append_constraint(
+                sa.UniqueConstraint(*uc["column_names"], name=uc["name"])
+            )
+        table.create(connection)
+
+        insp = inspect(connection)
+        reflected = sorted(
+            insp.get_unique_constraints("testtbl", schema=schema),
+            key=operator.itemgetter("name"),
+        )
+
+        names_that_duplicate_index = set()
+
+        eq_(len(uniques), len(reflected))
+
+        for orig, refl in zip(uniques, reflected):
+            # Different dialects handle duplicate index and constraints
+            # differently, so ignore this flag
+            dupe = refl.pop("duplicates_index", None)
+            if dupe:
+                names_that_duplicate_index.add(dupe)
+            eq_(refl.pop("comment", None), None)
+            eq_(orig, refl)
+
+        reflected_metadata = MetaData()
+        reflected = Table(
+            "testtbl",
+            reflected_metadata,
+            autoload_with=connection,
+            schema=schema,
+        )
+
+        # test "deduplicates for index" logic.   MySQL and Oracle
+        # "unique constraints" are actually unique indexes (with possible
+        # exception of a unique that is a dupe of another one in the case
+        # of Oracle).  make sure # they aren't duplicated.
+        idx_names = {idx.name for idx in reflected.indexes}
+        uq_names = {
+            uq.name
+            for uq in reflected.constraints
+            if isinstance(uq, sa.UniqueConstraint)
+        }.difference(["unique_c_a_b"])
+
+        assert not idx_names.intersection(uq_names)
+        if names_that_duplicate_index:
+            eq_(names_that_duplicate_index, idx_names)
+            eq_(uq_names, set())
+
+        no_cst = self.tables.no_constraints.name
+        eq_(insp.get_unique_constraints(no_cst, schema=schema), [])
+
+    @testing.requires.view_reflection
+    @testing.combinations(
+        (False,), (True, testing.requires.schemas), argnames="use_schema"
+    )
+    def test_get_view_definition(self, connection, use_schema):
+        if use_schema:
+            schema = config.test_schema
+        else:
+            schema = None
+        insp = inspect(connection)
+        for view in ["users_v", "email_addresses_v", "dingalings_v"]:
+            v = insp.get_view_definition(view, schema=schema)
+            is_true(bool(v))
+
+    @testing.requires.view_reflection
+    def test_get_view_definition_does_not_exist(self, connection):
+        insp = inspect(connection)
+        with expect_raises(NoSuchTableError):
+            insp.get_view_definition("view_does_not_exist")
+        with expect_raises(NoSuchTableError):
+            insp.get_view_definition("users")  # a table
+
+    @testing.requires.table_reflection
+    def test_autoincrement_col(self, connection):
+        """test that 'autoincrement' is reflected according to sqla's policy.
+
+        Don't mark this test as unsupported for any backend !
+
+        (technically it fails with MySQL InnoDB since "id" comes before "id2")
+
+        A backend is better off not returning "autoincrement" at all,
+        instead of potentially returning "False" for an auto-incrementing
+        primary key column.
+
+        """
+
+        insp = inspect(connection)
+
+        for tname, cname in [
+            ("users", "user_id"),
+            ("email_addresses", "address_id"),
+            ("dingalings", "dingaling_id"),
+        ]:
+            cols = insp.get_columns(tname)
+            id_ = {c["name"]: c for c in cols}[cname]
+            assert id_.get("autoincrement", True)
+
+    @testing.combinations(
+        (True, testing.requires.schemas), (False,), argnames="use_schema"
+    )
+    def test_get_table_options(self, use_schema):
+        insp = inspect(config.db)
+        schema = config.test_schema if use_schema else None
+
+        if testing.requires.reflect_table_options.enabled:
+            res = insp.get_table_options("users", schema=schema)
+            is_true(isinstance(res, dict))
+            # NOTE: can't really create a table with no option
+            res = insp.get_table_options("no_constraints", schema=schema)
+            is_true(isinstance(res, dict))
+        else:
+            with expect_raises(NotImplementedError):
+                res = insp.get_table_options("users", schema=schema)
+
+    @testing.combinations((True, testing.requires.schemas), False)
+    def test_multi_get_table_options(self, use_schema):
+        insp = inspect(config.db)
+        if testing.requires.reflect_table_options.enabled:
+            schema = config.test_schema if use_schema else None
+            res = insp.get_multi_table_options(schema=schema)
+
+            exp = {
+                (schema, table): insp.get_table_options(table, schema=schema)
+                for table in insp.get_table_names(schema=schema)
+            }
+            eq_(res, exp)
+        else:
+            with expect_raises(NotImplementedError):
+                res = insp.get_multi_table_options()
+
+    @testing.fixture
+    def get_multi_exp(self, connection):
+        def provide_fixture(
+            schema, scope, kind, use_filter, single_reflect_fn, exp_method
+        ):
+            insp = inspect(connection)
+            # call the reflection function at least once to avoid
+            # "Unexpected success" errors if the result is actually empty
+            # and NotImplementedError is not raised
+            single_reflect_fn(insp, "email_addresses")
+            kw = {"scope": scope, "kind": kind}
+            if schema:
+                schema = schema()
+
+            filter_names = []
+
+            if ObjectKind.TABLE in kind:
+                filter_names.extend(
+                    ["comment_test", "users", "does-not-exist"]
+                )
+            if ObjectKind.VIEW in kind:
+                filter_names.extend(["email_addresses_v", "does-not-exist"])
+            if ObjectKind.MATERIALIZED_VIEW in kind:
+                filter_names.extend(["dingalings_v", "does-not-exist"])
+
+            if schema:
+                kw["schema"] = schema
+            if use_filter:
+                kw["filter_names"] = filter_names
+
+            exp = exp_method(
+                schema=schema,
+                scope=scope,
+                kind=kind,
+                filter_names=kw.get("filter_names"),
+            )
+            kws = [kw]
+            if scope == ObjectScope.DEFAULT:
+                nkw = kw.copy()
+                nkw.pop("scope")
+                kws.append(nkw)
+            if kind == ObjectKind.TABLE:
+                nkw = kw.copy()
+                nkw.pop("kind")
+                kws.append(nkw)
+
+            return inspect(connection), kws, exp
+
+        return provide_fixture
+
+    @testing.requires.reflect_table_options
+    @_multi_combination
+    def test_multi_get_table_options_tables(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_table_options,
+            self.exp_options,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_table_options(**kw)
+            eq_(result, exp)
+
+    @testing.requires.comment_reflection
+    @_multi_combination
+    def test_get_multi_table_comment(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_table_comment,
+            self.exp_comments,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            eq_(insp.get_multi_table_comment(**kw), exp)
+
+    def _check_expressions(self, result, exp, err_msg):
+        def _clean(text: str):
+            return re.sub(r"['\" ]", "", text).lower()
+
+        if isinstance(exp, dict):
+            eq_({_clean(e): v for e, v in result.items()}, exp, err_msg)
+        else:
+            eq_([_clean(e) for e in result], exp, err_msg)
+
+    def _check_list(self, result, exp, req_keys=None, msg=None):
+        if req_keys is None:
+            eq_(result, exp, msg)
+        else:
+            eq_(len(result), len(exp), msg)
+            for r, e in zip(result, exp):
+                for k in set(r) | set(e):
+                    if k in req_keys or (k in r and k in e):
+                        err_msg = f"{msg} - {k} - {r}"
+                        if k in ("expressions", "column_sorting"):
+                            self._check_expressions(r[k], e[k], err_msg)
+                        else:
+                            eq_(r[k], e[k], err_msg)
+
+    def _check_table_dict(self, result, exp, req_keys=None, make_lists=False):
+        eq_(set(result.keys()), set(exp.keys()))
+        for k in result:
+            r, e = result[k], exp[k]
+            if make_lists:
+                r, e = [r], [e]
+            self._check_list(r, e, req_keys, k)
+
+    @_multi_combination
+    def test_get_multi_columns(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_columns,
+            self.exp_columns,
+        )
+
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_columns(**kw)
+            self._check_table_dict(result, exp, self._required_column_keys)
+
+    @testing.requires.primary_key_constraint_reflection
+    @_multi_combination
+    def test_get_multi_pk_constraint(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_pk_constraint,
+            self.exp_pks,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_pk_constraint(**kw)
+            self._check_table_dict(
+                result, exp, self._required_pk_keys, make_lists=True
+            )
+
+    def _adjust_sort(self, result, expected, key):
+        if not testing.requires.implicitly_named_constraints.enabled:
+            for obj in [result, expected]:
+                for val in obj.values():
+                    if len(val) > 1 and any(
+                        v.get("name") in (None, mock.ANY) for v in val
+                    ):
+                        val.sort(key=key)
+
+    @testing.requires.foreign_key_constraint_reflection
+    @_multi_combination
+    def test_get_multi_foreign_keys(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_foreign_keys,
+            self.exp_fks,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_foreign_keys(**kw)
+            self._adjust_sort(
+                result, exp, lambda d: tuple(d["constrained_columns"])
+            )
+            self._check_table_dict(result, exp, self._required_fk_keys)
+
+    @testing.requires.index_reflection
+    @_multi_combination
+    def test_get_multi_indexes(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_indexes,
+            self.exp_indexes,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_indexes(**kw)
+            self._check_table_dict(result, exp, self._required_index_keys)
+
+    @testing.requires.unique_constraint_reflection
+    @_multi_combination
+    def test_get_multi_unique_constraints(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_unique_constraints,
+            self.exp_ucs,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_unique_constraints(**kw)
+            self._adjust_sort(result, exp, lambda d: tuple(d["column_names"]))
+            self._check_table_dict(result, exp, self._required_unique_cst_keys)
+
+    @testing.requires.check_constraint_reflection
+    @_multi_combination
+    def test_get_multi_check_constraints(
+        self, get_multi_exp, schema, scope, kind, use_filter
+    ):
+        insp, kws, exp = get_multi_exp(
+            schema,
+            scope,
+            kind,
+            use_filter,
+            Inspector.get_check_constraints,
+            self.exp_ccs,
+        )
+        for kw in kws:
+            insp.clear_cache()
+            result = insp.get_multi_check_constraints(**kw)
+            self._adjust_sort(result, exp, lambda d: tuple(d["sqltext"]))
+            self._check_table_dict(result, exp, self._required_cc_keys)
+
+    @testing.combinations(
+        ("get_table_options", testing.requires.reflect_table_options),
+        "get_columns",
+        (
+            "get_pk_constraint",
+            testing.requires.primary_key_constraint_reflection,
+        ),
+        (
+            "get_foreign_keys",
+            testing.requires.foreign_key_constraint_reflection,
+        ),
+        ("get_indexes", testing.requires.index_reflection),
+        (
+            "get_unique_constraints",
+            testing.requires.unique_constraint_reflection,
+        ),
+        (
+            "get_check_constraints",
+            testing.requires.check_constraint_reflection,
+        ),
+        ("get_table_comment", testing.requires.comment_reflection),
+        argnames="method",
+    )
+    def test_not_existing_table(self, method, connection):
+        insp = inspect(connection)
+        meth = getattr(insp, method)
+        with expect_raises(NoSuchTableError):
+            meth("table_does_not_exists")
+
+    def test_unreflectable(self, connection):
+        mc = Inspector.get_multi_columns
+
+        def patched(*a, **k):
+            ur = k.setdefault("unreflectable", {})
+            ur[(None, "some_table")] = UnreflectableTableError("err")
+            return mc(*a, **k)
+
+        with mock.patch.object(Inspector, "get_multi_columns", patched):
+            with expect_raises_message(UnreflectableTableError, "err"):
+                inspect(connection).reflect_table(
+                    Table("some_table", MetaData()), None
+                )
+
+    @testing.combinations(True, False, argnames="use_schema")
+    @testing.combinations(
+        (True, testing.requires.views), False, argnames="views"
+    )
+    def test_metadata(self, connection, use_schema, views):
+        m = MetaData()
+        schema = config.test_schema if use_schema else None
+        m.reflect(connection, schema=schema, views=views, resolve_fks=False)
+
+        insp = inspect(connection)
+        tables = insp.get_table_names(schema)
+        if views:
+            tables += insp.get_view_names(schema)
+            try:
+                tables += insp.get_materialized_view_names(schema)
+            except NotImplementedError:
+                pass
+        if schema:
+            tables = [f"{schema}.{t}" for t in tables]
+        eq_(sorted(m.tables), sorted(tables))
+
+    @testing.requires.comment_reflection
+    def test_comments_unicode(self, connection, metadata):
+        Table(
+            "unicode_comments",
+            metadata,
+            Column("unicode", Integer, comment="é試蛇ẟΩ"),
+            Column("emoji", Integer, comment="☁️✨"),
+            comment="試蛇ẟΩ✨",
+        )
+
+        metadata.create_all(connection)
+
+        insp = inspect(connection)
+        tc = insp.get_table_comment("unicode_comments")
+        eq_(tc, {"text": "試蛇ẟΩ✨"})
+
+        cols = insp.get_columns("unicode_comments")
+        value = {c["name"]: c["comment"] for c in cols}
+        exp = {"unicode": "é試蛇ẟΩ", "emoji": "☁️✨"}
+        eq_(value, exp)
+
+    @testing.requires.comment_reflection_full_unicode
+    def test_comments_unicode_full(self, connection, metadata):
+        Table(
+            "unicode_comments",
+            metadata,
+            Column("emoji", Integer, comment="🐍🧙🝝🧙‍♂️🧙‍♀️"),
+            comment="🎩🁰🝑🤷‍♀️🤷‍♂️",
+        )
+
+        metadata.create_all(connection)
+
+        insp = inspect(connection)
+        tc = insp.get_table_comment("unicode_comments")
+        eq_(tc, {"text": "🎩🁰🝑🤷‍♀️🤷‍♂️"})
+        c = insp.get_columns("unicode_comments")[0]
+        eq_({c["name"]: c["comment"]}, {"emoji": "🐍🧙🝝🧙‍♂️🧙‍♀️"})
+
+
+class TableNoColumnsTest(fixtures.TestBase):
+    __requires__ = ("reflect_tables_no_columns",)
+    __backend__ = True
+
+    @testing.fixture
+    def table_no_columns(self, connection, metadata):
+        Table("empty", metadata)
+        metadata.create_all(connection)
+
+    @testing.fixture
+    def view_no_columns(self, connection, metadata):
+        Table("empty", metadata)
+        event.listen(
+            metadata,
+            "after_create",
+            DDL("CREATE VIEW empty_v AS SELECT * FROM empty"),
+        )
+
+        # for transactional DDL the transaction is rolled back before this
+        # drop statement is invoked
+        event.listen(
+            metadata, "before_drop", DDL("DROP VIEW IF EXISTS empty_v")
+        )
+        metadata.create_all(connection)
+
+    def test_reflect_table_no_columns(self, connection, table_no_columns):
+        t2 = Table("empty", MetaData(), autoload_with=connection)
+        eq_(list(t2.c), [])
+
+    def test_get_columns_table_no_columns(self, connection, table_no_columns):
+        insp = inspect(connection)
+        eq_(insp.get_columns("empty"), [])
+        multi = insp.get_multi_columns()
+        eq_(multi, {(None, "empty"): []})
+
+    def test_reflect_incl_table_no_columns(self, connection, table_no_columns):
+        m = MetaData()
+        m.reflect(connection)
+        assert set(m.tables).intersection(["empty"])
+
+    @testing.requires.views
+    def test_reflect_view_no_columns(self, connection, view_no_columns):
+        t2 = Table("empty_v", MetaData(), autoload_with=connection)
+        eq_(list(t2.c), [])
+
+    @testing.requires.views
+    def test_get_columns_view_no_columns(self, connection, view_no_columns):
+        insp = inspect(connection)
+        eq_(insp.get_columns("empty_v"), [])
+        multi = insp.get_multi_columns(kind=ObjectKind.VIEW)
+        eq_(multi, {(None, "empty_v"): []})
+
+
+class ComponentReflectionTestExtra(ComparesIndexes, fixtures.TestBase):
+    __backend__ = True
+
+    @testing.fixture(params=[True, False])
+    def use_schema_fixture(self, request):
+        if request.param:
+            return config.test_schema
+        else:
+            return None
+
+    @testing.fixture()
+    def inspect_for_table(self, metadata, connection, use_schema_fixture):
+        @contextlib.contextmanager
+        def go(tablename):
+            yield use_schema_fixture, inspect(connection)
+
+            metadata.create_all(connection)
+
+        return go
+
+    def ck_eq(self, reflected, expected):
+        # trying to minimize effect of quoting, parenthesis, etc.
+        # may need to add more to this as new dialects get CHECK
+        # constraint reflection support
+        def normalize(sqltext):
+            return " ".join(
+                re.findall(r"and|\d|=|a|b|c|or|<|>", sqltext.lower(), re.I)
+            )
+
+        reflected = sorted(
+            [
+                {"name": item["name"], "sqltext": normalize(item["sqltext"])}
+                for item in reflected
+            ],
+            key=lambda item: (item["sqltext"]),
+        )
+
+        expected = sorted(
+            expected,
+            key=lambda item: (item["sqltext"]),
+        )
+        eq_(reflected, expected)
+
+    @testing.requires.check_constraint_reflection
+    def test_check_constraint_no_constraint(self, metadata, inspect_for_table):
+        with inspect_for_table("no_constraints") as (schema, inspector):
+            Table(
+                "no_constraints",
+                metadata,
+                Column("data", sa.String(20)),
+                schema=schema,
+            )
+
+        self.ck_eq(
+            inspector.get_check_constraints("no_constraints", schema=schema),
+            [],
+        )
+
+    @testing.requires.inline_check_constraint_reflection
+    @testing.combinations(
+        "my_inline", "MyInline", None, argnames="constraint_name"
+    )
+    def test_check_constraint_inline(
+        self, metadata, inspect_for_table, constraint_name
+    ):
+
+        with inspect_for_table("sa_cc") as (schema, inspector):
+            Table(
+                "sa_cc",
+                metadata,
+                Column("id", Integer(), primary_key=True),
+                Column(
+                    "a",
+                    Integer(),
+                    sa.CheckConstraint(
+                        "a > 1 AND a < 5", name=constraint_name
+                    ),
+                ),
+                Column("data", String(50)),
+                schema=schema,
+            )
+
+        reflected = inspector.get_check_constraints("sa_cc", schema=schema)
+
+        self.ck_eq(
+            reflected,
+            [
+                {
+                    "name": constraint_name or mock.ANY,
+                    "sqltext": "a > 1 and a < 5",
+                },
+            ],
+        )
+
+    @testing.requires.check_constraint_reflection
+    @testing.combinations(
+        "my_ck_const", "MyCkConst", None, argnames="constraint_name"
+    )
+    def test_check_constraint_standalone(
+        self, metadata, inspect_for_table, constraint_name
+    ):
+        with inspect_for_table("sa_cc") as (schema, inspector):
+            Table(
+                "sa_cc",
+                metadata,
+                Column("a", Integer()),
+                sa.CheckConstraint(
+                    "a = 1 OR (a > 2 AND a < 5)", name=constraint_name
+                ),
+                schema=schema,
+            )
+
+        reflected = inspector.get_check_constraints("sa_cc", schema=schema)
+
+        self.ck_eq(
+            reflected,
+            [
+                {
+                    "name": constraint_name or mock.ANY,
+                    "sqltext": "a = 1 or a > 2 and a < 5",
+                },
+            ],
+        )
+
+    @testing.requires.inline_check_constraint_reflection
+    def test_check_constraint_mixed(self, metadata, inspect_for_table):
+        with inspect_for_table("sa_cc") as (schema, inspector):
+            Table(
+                "sa_cc",
+                metadata,
+                Column("id", Integer(), primary_key=True),
+                Column("a", Integer(), sa.CheckConstraint("a > 1 AND a < 5")),
+                Column(
+                    "b",
+                    Integer(),
+                    sa.CheckConstraint("b > 1 AND b < 5", name="my_inline"),
+                ),
+                Column("c", Integer()),
+                Column("data", String(50)),
+                sa.UniqueConstraint("data", name="some_uq"),
+                sa.CheckConstraint("c > 1 AND c < 5", name="cc1"),
+                sa.UniqueConstraint("c", name="some_c_uq"),
+                schema=schema,
+            )
+
+        reflected = inspector.get_check_constraints("sa_cc", schema=schema)
+
+        self.ck_eq(
+            reflected,
+            [
+                {"name": "cc1", "sqltext": "c > 1 and c < 5"},
+                {"name": "my_inline", "sqltext": "b > 1 and b < 5"},
+                {"name": mock.ANY, "sqltext": "a > 1 and a < 5"},
+            ],
+        )
+
+    @testing.requires.indexes_with_expressions
+    def test_reflect_expression_based_indexes(self, metadata, connection):
+        t = Table(
+            "t",
+            metadata,
+            Column("x", String(30)),
+            Column("y", String(30)),
+            Column("z", String(30)),
+        )
+
+        Index("t_idx", func.lower(t.c.x), t.c.z, func.lower(t.c.y))
+        long_str = "long string " * 100
+        Index("t_idx_long", func.coalesce(t.c.x, long_str))
+        Index("t_idx_2", t.c.x)
+
+        metadata.create_all(connection)
+
+        insp = inspect(connection)
+
+        expected = [
+            {
+                "name": "t_idx_2",
+                "column_names": ["x"],
+                "unique": False,
+                "dialect_options": {},
+            }
+        ]
+
+        def completeIndex(entry):
+            if testing.requires.index_reflects_included_columns.enabled:
+                entry["include_columns"] = []
+                entry["dialect_options"] = {
+                    f"{connection.engine.name}_include": []
+                }
+            else:
+                entry.setdefault("dialect_options", {})
+
+        completeIndex(expected[0])
+
+        class lower_index_str(str):
+            def __eq__(self, other):
+                ol = other.lower()
+                # test that lower and x or y are in the string
+                return "lower" in ol and ("x" in ol or "y" in ol)
+
+        class coalesce_index_str(str):
+            def __eq__(self, other):
+                # test that coalesce and the string is in other
+                return "coalesce" in other.lower() and long_str in other
+
+        if testing.requires.reflect_indexes_with_expressions.enabled:
+            expr_index = {
+                "name": "t_idx",
+                "column_names": [None, "z", None],
+                "expressions": [
+                    lower_index_str("lower(x)"),
+                    "z",
+                    lower_index_str("lower(y)"),
+                ],
+                "unique": False,
+            }
+            completeIndex(expr_index)
+            expected.insert(0, expr_index)
+
+            expr_index_long = {
+                "name": "t_idx_long",
+                "column_names": [None],
+                "expressions": [
+                    coalesce_index_str(f"coalesce(x, '{long_str}')")
+                ],
+                "unique": False,
+            }
+            completeIndex(expr_index_long)
+            expected.append(expr_index_long)
+
+            eq_(insp.get_indexes("t"), expected)
+            m2 = MetaData()
+            t2 = Table("t", m2, autoload_with=connection)
+        else:
+            with expect_warnings(
+                "Skipped unsupported reflection of expression-based "
+                "index t_idx"
+            ):
+                eq_(insp.get_indexes("t"), expected)
+                m2 = MetaData()
+                t2 = Table("t", m2, autoload_with=connection)
+
+        self.compare_table_index_with_expected(
+            t2, expected, connection.engine.name
+        )
+
+    @testing.requires.index_reflects_included_columns
+    def test_reflect_covering_index(self, metadata, connection):
+        t = Table(
+            "t",
+            metadata,
+            Column("x", String(30)),
+            Column("y", String(30)),
+        )
+        idx = Index("t_idx", t.c.x)
+        idx.dialect_options[connection.engine.name]["include"] = ["y"]
+
+        metadata.create_all(connection)
+
+        insp = inspect(connection)
+
+        get_indexes = insp.get_indexes("t")
+        eq_(
+            get_indexes,
+            [
+                {
+                    "name": "t_idx",
+                    "column_names": ["x"],
+                    "include_columns": ["y"],
+                    "unique": False,
+                    "dialect_options": mock.ANY,
+                }
+            ],
+        )
+        eq_(
+            get_indexes[0]["dialect_options"][
+                "%s_include" % connection.engine.name
+            ],
+            ["y"],
+        )
+
+        t2 = Table("t", MetaData(), autoload_with=connection)
+        eq_(
+            list(t2.indexes)[0].dialect_options[connection.engine.name][
+                "include"
+            ],
+            ["y"],
+        )
+
+    def _type_round_trip(self, connection, metadata, *types):
+        t = Table(
+            "t",
+            metadata,
+            *[Column("t%d" % i, type_) for i, type_ in enumerate(types)],
+        )
+        t.create(connection)
+
+        return [c["type"] for c in inspect(connection).get_columns("t")]
+
+    @testing.requires.table_reflection
+    def test_numeric_reflection(self, connection, metadata):
+        for typ in self._type_round_trip(
+            connection, metadata, sql_types.Numeric(18, 5)
+        ):
+            assert isinstance(typ, sql_types.Numeric)
+            eq_(typ.precision, 18)
+            eq_(typ.scale, 5)
+
+    @testing.requires.table_reflection
+    def test_varchar_reflection(self, connection, metadata):
+        typ = self._type_round_trip(
+            connection, metadata, sql_types.String(52)
+        )[0]
+        assert isinstance(typ, sql_types.String)
+        eq_(typ.length, 52)
+
+    @testing.requires.table_reflection
+    def test_nullable_reflection(self, connection, metadata):
+        t = Table(
+            "t",
+            metadata,
+            Column("a", Integer, nullable=True),
+            Column("b", Integer, nullable=False),
+        )
+        t.create(connection)
+        eq_(
+            {
+                col["name"]: col["nullable"]
+                for col in inspect(connection).get_columns("t")
+            },
+            {"a": True, "b": False},
+        )
+
+    @testing.combinations(
+        (
+            None,
+            "CASCADE",
+            None,
+            testing.requires.foreign_key_constraint_option_reflection_ondelete,
+        ),
+        (
+            None,
+            None,
+            "SET NULL",
+            testing.requires.foreign_key_constraint_option_reflection_onupdate,
+        ),
+        (
+            {},
+            None,
+            "NO ACTION",
+            testing.requires.foreign_key_constraint_option_reflection_onupdate,
+        ),
+        (
+            {},
+            "NO ACTION",
+            None,
+            testing.requires.fk_constraint_option_reflection_ondelete_noaction,
+        ),
+        (
+            None,
+            None,
+            "RESTRICT",
+            testing.requires.fk_constraint_option_reflection_onupdate_restrict,
+        ),
+        (
+            None,
+            "RESTRICT",
+            None,
+            testing.requires.fk_constraint_option_reflection_ondelete_restrict,
+        ),
+        argnames="expected,ondelete,onupdate",
+    )
+    def test_get_foreign_key_options(
+        self, connection, metadata, expected, ondelete, onupdate
+    ):
+        options = {}
+        if ondelete:
+            options["ondelete"] = ondelete
+        if onupdate:
+            options["onupdate"] = onupdate
+
+        if expected is None:
+            expected = options
+
+        Table(
+            "x",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            test_needs_fk=True,
+        )
+
+        Table(
+            "table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x_id", Integer, ForeignKey("x.id", name="xid")),
+            Column("test", String(10)),
+            test_needs_fk=True,
+        )
+
+        Table(
+            "user",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("name", String(50), nullable=False),
+            Column("tid", Integer),
+            sa.ForeignKeyConstraint(
+                ["tid"], ["table.id"], name="myfk", **options
+            ),
+            test_needs_fk=True,
+        )
+
+        metadata.create_all(connection)
+
+        insp = inspect(connection)
+
+        # test 'options' is always present for a backend
+        # that can reflect these, since alembic looks for this
+        opts = insp.get_foreign_keys("table")[0]["options"]
+
+        eq_({k: opts[k] for k in opts if opts[k]}, {})
+
+        opts = insp.get_foreign_keys("user")[0]["options"]
+        eq_(opts, expected)
+        # eq_(dict((k, opts[k]) for k in opts if opts[k]), expected)
+
+
+class NormalizedNameTest(fixtures.TablesTest):
+    __requires__ = ("denormalized_names",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            quoted_name("t1", quote=True),
+            metadata,
+            Column("id", Integer, primary_key=True),
+        )
+        Table(
+            quoted_name("t2", quote=True),
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("t1id", ForeignKey("t1.id")),
+        )
+
+    def test_reflect_lowercase_forced_tables(self):
+        m2 = MetaData()
+        t2_ref = Table(
+            quoted_name("t2", quote=True), m2, autoload_with=config.db
+        )
+        t1_ref = m2.tables["t1"]
+        assert t2_ref.c.t1id.references(t1_ref.c.id)
+
+        m3 = MetaData()
+        m3.reflect(
+            config.db, only=lambda name, m: name.lower() in ("t1", "t2")
+        )
+        assert m3.tables["t2"].c.t1id.references(m3.tables["t1"].c.id)
+
+    def test_get_table_names(self):
+        tablenames = [
+            t
+            for t in inspect(config.db).get_table_names()
+            if t.lower() in ("t1", "t2")
+        ]
+
+        eq_(tablenames[0].upper(), tablenames[0].lower())
+        eq_(tablenames[1].upper(), tablenames[1].lower())
+
+
+class ComputedReflectionTest(fixtures.ComputedReflectionFixtureTest):
+    def test_computed_col_default_not_set(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns("computed_default_table")
+        col_data = {c["name"]: c for c in cols}
+        is_true("42" in col_data["with_default"]["default"])
+        is_(col_data["normal"]["default"], None)
+        is_(col_data["computed_col"]["default"], None)
+
+    def test_get_column_returns_computed(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns("computed_default_table")
+        data = {c["name"]: c for c in cols}
+        for key in ("id", "normal", "with_default"):
+            is_true("computed" not in data[key])
+        compData = data["computed_col"]
+        is_true("computed" in compData)
+        is_true("sqltext" in compData["computed"])
+        eq_(self.normalize(compData["computed"]["sqltext"]), "normal+42")
+        eq_(
+            "persisted" in compData["computed"],
+            testing.requires.computed_columns_reflect_persisted.enabled,
+        )
+        if testing.requires.computed_columns_reflect_persisted.enabled:
+            eq_(
+                compData["computed"]["persisted"],
+                testing.requires.computed_columns_default_persisted.enabled,
+            )
+
+    def check_column(self, data, column, sqltext, persisted):
+        is_true("computed" in data[column])
+        compData = data[column]["computed"]
+        eq_(self.normalize(compData["sqltext"]), sqltext)
+        if testing.requires.computed_columns_reflect_persisted.enabled:
+            is_true("persisted" in compData)
+            is_(compData["persisted"], persisted)
+
+    def test_get_column_returns_persisted(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns("computed_column_table")
+        data = {c["name"]: c for c in cols}
+
+        self.check_column(
+            data,
+            "computed_no_flag",
+            "normal+42",
+            testing.requires.computed_columns_default_persisted.enabled,
+        )
+        if testing.requires.computed_columns_virtual.enabled:
+            self.check_column(
+                data,
+                "computed_virtual",
+                "normal+2",
+                False,
+            )
+        if testing.requires.computed_columns_stored.enabled:
+            self.check_column(
+                data,
+                "computed_stored",
+                "normal-42",
+                True,
+            )
+
+    @testing.requires.schemas
+    def test_get_column_returns_persisted_with_schema(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns(
+            "computed_column_table", schema=config.test_schema
+        )
+        data = {c["name"]: c for c in cols}
+
+        self.check_column(
+            data,
+            "computed_no_flag",
+            "normal/42",
+            testing.requires.computed_columns_default_persisted.enabled,
+        )
+        if testing.requires.computed_columns_virtual.enabled:
+            self.check_column(
+                data,
+                "computed_virtual",
+                "normal/2",
+                False,
+            )
+        if testing.requires.computed_columns_stored.enabled:
+            self.check_column(
+                data,
+                "computed_stored",
+                "normal*42",
+                True,
+            )
+
+
+class IdentityReflectionTest(fixtures.TablesTest):
+    run_inserts = run_deletes = None
+
+    __backend__ = True
+    __requires__ = ("identity_columns", "table_reflection")
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "t1",
+            metadata,
+            Column("normal", Integer),
+            Column("id1", Integer, Identity()),
+        )
+        Table(
+            "t2",
+            metadata,
+            Column(
+                "id2",
+                Integer,
+                Identity(
+                    always=True,
+                    start=2,
+                    increment=3,
+                    minvalue=-2,
+                    maxvalue=42,
+                    cycle=True,
+                    cache=4,
+                ),
+            ),
+        )
+        if testing.requires.schemas.enabled:
+            Table(
+                "t1",
+                metadata,
+                Column("normal", Integer),
+                Column("id1", Integer, Identity(always=True, start=20)),
+                schema=config.test_schema,
+            )
+
+    def check(self, value, exp, approx):
+        if testing.requires.identity_columns_standard.enabled:
+            common_keys = (
+                "always",
+                "start",
+                "increment",
+                "minvalue",
+                "maxvalue",
+                "cycle",
+                "cache",
+            )
+            for k in list(value):
+                if k not in common_keys:
+                    value.pop(k)
+            if approx:
+                eq_(len(value), len(exp))
+                for k in value:
+                    if k == "minvalue":
+                        is_true(value[k] <= exp[k])
+                    elif k in {"maxvalue", "cache"}:
+                        is_true(value[k] >= exp[k])
+                    else:
+                        eq_(value[k], exp[k], k)
+            else:
+                eq_(value, exp)
+        else:
+            eq_(value["start"], exp["start"])
+            eq_(value["increment"], exp["increment"])
+
+    def test_reflect_identity(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns("t1") + insp.get_columns("t2")
+        for col in cols:
+            if col["name"] == "normal":
+                is_false("identity" in col)
+            elif col["name"] == "id1":
+                if "autoincrement" in col:
+                    is_true(col["autoincrement"])
+                eq_(col["default"], None)
+                is_true("identity" in col)
+                self.check(
+                    col["identity"],
+                    dict(
+                        always=False,
+                        start=1,
+                        increment=1,
+                        minvalue=1,
+                        maxvalue=2147483647,
+                        cycle=False,
+                        cache=1,
+                    ),
+                    approx=True,
+                )
+            elif col["name"] == "id2":
+                if "autoincrement" in col:
+                    is_true(col["autoincrement"])
+                eq_(col["default"], None)
+                is_true("identity" in col)
+                self.check(
+                    col["identity"],
+                    dict(
+                        always=True,
+                        start=2,
+                        increment=3,
+                        minvalue=-2,
+                        maxvalue=42,
+                        cycle=True,
+                        cache=4,
+                    ),
+                    approx=False,
+                )
+
+    @testing.requires.schemas
+    def test_reflect_identity_schema(self):
+        insp = inspect(config.db)
+
+        cols = insp.get_columns("t1", schema=config.test_schema)
+        for col in cols:
+            if col["name"] == "normal":
+                is_false("identity" in col)
+            elif col["name"] == "id1":
+                if "autoincrement" in col:
+                    is_true(col["autoincrement"])
+                eq_(col["default"], None)
+                is_true("identity" in col)
+                self.check(
+                    col["identity"],
+                    dict(
+                        always=True,
+                        start=20,
+                        increment=1,
+                        minvalue=1,
+                        maxvalue=2147483647,
+                        cycle=False,
+                        cache=1,
+                    ),
+                    approx=True,
+                )
+
+
+class CompositeKeyReflectionTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        tb1 = Table(
+            "tb1",
+            metadata,
+            Column("id", Integer),
+            Column("attr", Integer),
+            Column("name", sql_types.VARCHAR(20)),
+            sa.PrimaryKeyConstraint("name", "id", "attr", name="pk_tb1"),
+            schema=None,
+            test_needs_fk=True,
+        )
+        Table(
+            "tb2",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("pid", Integer),
+            Column("pattr", Integer),
+            Column("pname", sql_types.VARCHAR(20)),
+            sa.ForeignKeyConstraint(
+                ["pname", "pid", "pattr"],
+                [tb1.c.name, tb1.c.id, tb1.c.attr],
+                name="fk_tb1_name_id_attr",
+            ),
+            schema=None,
+            test_needs_fk=True,
+        )
+
+    @testing.requires.primary_key_constraint_reflection
+    def test_pk_column_order(self, connection):
+        # test for issue #5661
+        insp = inspect(connection)
+        primary_key = insp.get_pk_constraint(self.tables.tb1.name)
+        eq_(primary_key.get("constrained_columns"), ["name", "id", "attr"])
+
+    @testing.requires.foreign_key_constraint_reflection
+    def test_fk_column_order(self, connection):
+        # test for issue #5661
+        insp = inspect(connection)
+        foreign_keys = insp.get_foreign_keys(self.tables.tb2.name)
+        eq_(len(foreign_keys), 1)
+        fkey1 = foreign_keys[0]
+        eq_(fkey1.get("referred_columns"), ["name", "id", "attr"])
+        eq_(fkey1.get("constrained_columns"), ["pname", "pid", "pattr"])
+
+
+__all__ = (
+    "ComponentReflectionTest",
+    "ComponentReflectionTestExtra",
+    "TableNoColumnsTest",
+    "QuotedNameArgumentTest",
+    "BizarroCharacterFKResolutionTest",
+    "HasTableTest",
+    "HasIndexTest",
+    "NormalizedNameTest",
+    "ComputedReflectionTest",
+    "IdentityReflectionTest",
+    "CompositeKeyReflectionTest",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_results.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_results.py
new file mode 100644
index 00000000..a6179d85
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_results.py
@@ -0,0 +1,502 @@
+# testing/suite/test_results.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import datetime
+import re
+
+from .. import engines
+from .. import fixtures
+from ..assertions import eq_
+from ..config import requirements
+from ..schema import Column
+from ..schema import Table
+from ... import DateTime
+from ... import func
+from ... import Integer
+from ... import select
+from ... import sql
+from ... import String
+from ... import testing
+from ... import text
+
+
+class RowFetchTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "plain_pk",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+        )
+        Table(
+            "has_dates",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("today", DateTime),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.plain_pk.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+        connection.execute(
+            cls.tables.has_dates.insert(),
+            [{"id": 1, "today": datetime.datetime(2006, 5, 12, 12, 0, 0)}],
+        )
+
+    def test_via_attr(self, connection):
+        row = connection.execute(
+            self.tables.plain_pk.select().order_by(self.tables.plain_pk.c.id)
+        ).first()
+
+        eq_(row.id, 1)
+        eq_(row.data, "d1")
+
+    def test_via_string(self, connection):
+        row = connection.execute(
+            self.tables.plain_pk.select().order_by(self.tables.plain_pk.c.id)
+        ).first()
+
+        eq_(row._mapping["id"], 1)
+        eq_(row._mapping["data"], "d1")
+
+    def test_via_int(self, connection):
+        row = connection.execute(
+            self.tables.plain_pk.select().order_by(self.tables.plain_pk.c.id)
+        ).first()
+
+        eq_(row[0], 1)
+        eq_(row[1], "d1")
+
+    def test_via_col_object(self, connection):
+        row = connection.execute(
+            self.tables.plain_pk.select().order_by(self.tables.plain_pk.c.id)
+        ).first()
+
+        eq_(row._mapping[self.tables.plain_pk.c.id], 1)
+        eq_(row._mapping[self.tables.plain_pk.c.data], "d1")
+
+    @requirements.duplicate_names_in_cursor_description
+    def test_row_with_dupe_names(self, connection):
+        result = connection.execute(
+            select(
+                self.tables.plain_pk.c.data,
+                self.tables.plain_pk.c.data.label("data"),
+            ).order_by(self.tables.plain_pk.c.id)
+        )
+        row = result.first()
+        eq_(result.keys(), ["data", "data"])
+        eq_(row, ("d1", "d1"))
+
+    def test_row_w_scalar_select(self, connection):
+        """test that a scalar select as a column is returned as such
+        and that type conversion works OK.
+
+        (this is half a SQLAlchemy Core test and half to catch database
+        backends that may have unusual behavior with scalar selects.)
+
+        """
+        datetable = self.tables.has_dates
+        s = select(datetable.alias("x").c.today).scalar_subquery()
+        s2 = select(datetable.c.id, s.label("somelabel"))
+        row = connection.execute(s2).first()
+
+        eq_(row.somelabel, datetime.datetime(2006, 5, 12, 12, 0, 0))
+
+
+class PercentSchemaNamesTest(fixtures.TablesTest):
+    """tests using percent signs, spaces in table and column names.
+
+    This didn't work for PostgreSQL / MySQL drivers for a long time
+    but is now supported.
+
+    """
+
+    __requires__ = ("percent_schema_names",)
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        cls.tables.percent_table = Table(
+            "percent%table",
+            metadata,
+            Column("percent%", Integer),
+            Column("spaces % more spaces", Integer),
+        )
+        cls.tables.lightweight_percent_table = sql.table(
+            "percent%table",
+            sql.column("percent%"),
+            sql.column("spaces % more spaces"),
+        )
+
+    def test_single_roundtrip(self, connection):
+        percent_table = self.tables.percent_table
+        for params in [
+            {"percent%": 5, "spaces % more spaces": 12},
+            {"percent%": 7, "spaces % more spaces": 11},
+            {"percent%": 9, "spaces % more spaces": 10},
+            {"percent%": 11, "spaces % more spaces": 9},
+        ]:
+            connection.execute(percent_table.insert(), params)
+        self._assert_table(connection)
+
+    def test_executemany_roundtrip(self, connection):
+        percent_table = self.tables.percent_table
+        connection.execute(
+            percent_table.insert(), {"percent%": 5, "spaces % more spaces": 12}
+        )
+        connection.execute(
+            percent_table.insert(),
+            [
+                {"percent%": 7, "spaces % more spaces": 11},
+                {"percent%": 9, "spaces % more spaces": 10},
+                {"percent%": 11, "spaces % more spaces": 9},
+            ],
+        )
+        self._assert_table(connection)
+
+    @requirements.insert_executemany_returning
+    def test_executemany_returning_roundtrip(self, connection):
+        percent_table = self.tables.percent_table
+        connection.execute(
+            percent_table.insert(), {"percent%": 5, "spaces % more spaces": 12}
+        )
+        result = connection.execute(
+            percent_table.insert().returning(
+                percent_table.c["percent%"],
+                percent_table.c["spaces % more spaces"],
+            ),
+            [
+                {"percent%": 7, "spaces % more spaces": 11},
+                {"percent%": 9, "spaces % more spaces": 10},
+                {"percent%": 11, "spaces % more spaces": 9},
+            ],
+        )
+        eq_(result.all(), [(7, 11), (9, 10), (11, 9)])
+        self._assert_table(connection)
+
+    def _assert_table(self, conn):
+        percent_table = self.tables.percent_table
+        lightweight_percent_table = self.tables.lightweight_percent_table
+
+        for table in (
+            percent_table,
+            percent_table.alias(),
+            lightweight_percent_table,
+            lightweight_percent_table.alias(),
+        ):
+            eq_(
+                list(
+                    conn.execute(table.select().order_by(table.c["percent%"]))
+                ),
+                [(5, 12), (7, 11), (9, 10), (11, 9)],
+            )
+
+            eq_(
+                list(
+                    conn.execute(
+                        table.select()
+                        .where(table.c["spaces % more spaces"].in_([9, 10]))
+                        .order_by(table.c["percent%"])
+                    )
+                ),
+                [(9, 10), (11, 9)],
+            )
+
+            row = conn.execute(
+                table.select().order_by(table.c["percent%"])
+            ).first()
+            eq_(row._mapping["percent%"], 5)
+            eq_(row._mapping["spaces % more spaces"], 12)
+
+            eq_(row._mapping[table.c["percent%"]], 5)
+            eq_(row._mapping[table.c["spaces % more spaces"]], 12)
+
+        conn.execute(
+            percent_table.update().values(
+                {percent_table.c["spaces % more spaces"]: 15}
+            )
+        )
+
+        eq_(
+            list(
+                conn.execute(
+                    percent_table.select().order_by(
+                        percent_table.c["percent%"]
+                    )
+                )
+            ),
+            [(5, 15), (7, 15), (9, 15), (11, 15)],
+        )
+
+
+class ServerSideCursorsTest(
+    fixtures.TestBase, testing.AssertsExecutionResults
+):
+    __requires__ = ("server_side_cursors",)
+
+    __backend__ = True
+
+    def _is_server_side(self, cursor):
+        # TODO: this is a huge issue as it prevents these tests from being
+        # usable by third party dialects.
+        if self.engine.dialect.driver == "psycopg2":
+            return bool(cursor.name)
+        elif self.engine.dialect.driver == "pymysql":
+            sscursor = __import__("pymysql.cursors").cursors.SSCursor
+            return isinstance(cursor, sscursor)
+        elif self.engine.dialect.driver in ("aiomysql", "asyncmy", "aioodbc"):
+            return cursor.server_side
+        elif self.engine.dialect.driver == "mysqldb":
+            sscursor = __import__("MySQLdb.cursors").cursors.SSCursor
+            return isinstance(cursor, sscursor)
+        elif self.engine.dialect.driver == "mariadbconnector":
+            return not cursor.buffered
+        elif self.engine.dialect.driver in ("asyncpg", "aiosqlite"):
+            return cursor.server_side
+        elif self.engine.dialect.driver == "pg8000":
+            return getattr(cursor, "server_side", False)
+        elif self.engine.dialect.driver == "psycopg":
+            return bool(getattr(cursor, "name", False))
+        elif self.engine.dialect.driver == "oracledb":
+            return getattr(cursor, "server_side", False)
+        else:
+            return False
+
+    def _fixture(self, server_side_cursors):
+        if server_side_cursors:
+            with testing.expect_deprecated(
+                "The create_engine.server_side_cursors parameter is "
+                "deprecated and will be removed in a future release.  "
+                "Please use the Connection.execution_options.stream_results "
+                "parameter."
+            ):
+                self.engine = engines.testing_engine(
+                    options={"server_side_cursors": server_side_cursors}
+                )
+        else:
+            self.engine = engines.testing_engine(
+                options={"server_side_cursors": server_side_cursors}
+            )
+        return self.engine
+
+    def stringify(self, str_):
+        return re.compile(r"SELECT (\d+)", re.I).sub(
+            lambda m: str(select(int(m.group(1))).compile(testing.db)), str_
+        )
+
+    @testing.combinations(
+        ("global_string", True, lambda stringify: stringify("select 1"), True),
+        (
+            "global_text",
+            True,
+            lambda stringify: text(stringify("select 1")),
+            True,
+        ),
+        ("global_expr", True, select(1), True),
+        (
+            "global_off_explicit",
+            False,
+            lambda stringify: text(stringify("select 1")),
+            False,
+        ),
+        (
+            "stmt_option",
+            False,
+            select(1).execution_options(stream_results=True),
+            True,
+        ),
+        (
+            "stmt_option_disabled",
+            True,
+            select(1).execution_options(stream_results=False),
+            False,
+        ),
+        ("for_update_expr", True, select(1).with_for_update(), True),
+        # TODO: need a real requirement for this, or dont use this test
+        (
+            "for_update_string",
+            True,
+            lambda stringify: stringify("SELECT 1 FOR UPDATE"),
+            True,
+            testing.skip_if(["sqlite", "mssql"]),
+        ),
+        (
+            "text_no_ss",
+            False,
+            lambda stringify: text(stringify("select 42")),
+            False,
+        ),
+        (
+            "text_ss_option",
+            False,
+            lambda stringify: text(stringify("select 42")).execution_options(
+                stream_results=True
+            ),
+            True,
+        ),
+        id_="iaaa",
+        argnames="engine_ss_arg, statement, cursor_ss_status",
+    )
+    def test_ss_cursor_status(
+        self, engine_ss_arg, statement, cursor_ss_status
+    ):
+        engine = self._fixture(engine_ss_arg)
+        with engine.begin() as conn:
+            if callable(statement):
+                statement = testing.resolve_lambda(
+                    statement, stringify=self.stringify
+                )
+
+            if isinstance(statement, str):
+                result = conn.exec_driver_sql(statement)
+            else:
+                result = conn.execute(statement)
+            eq_(self._is_server_side(result.cursor), cursor_ss_status)
+            result.close()
+
+    def test_conn_option(self):
+        engine = self._fixture(False)
+
+        with engine.connect() as conn:
+            # should be enabled for this one
+            result = conn.execution_options(
+                stream_results=True
+            ).exec_driver_sql(self.stringify("select 1"))
+            assert self._is_server_side(result.cursor)
+
+            # the connection has autobegun, which means at the end of the
+            # block, we will roll back, which on MySQL at least will fail
+            # with "Commands out of sync" if the result set
+            # is not closed, so we close it first.
+            #
+            # fun fact!  why did we not have this result.close() in this test
+            # before 2.0? don't we roll back in the connection pool
+            # unconditionally? yes!  and in fact if you run this test in 1.4
+            # with stdout shown, there is in fact "Exception during reset or
+            # similar" with "Commands out sync" emitted a warning!  2.0's
+            # architecture finds and fixes what was previously an expensive
+            # silent error condition.
+            result.close()
+
+    def test_stmt_enabled_conn_option_disabled(self):
+        engine = self._fixture(False)
+
+        s = select(1).execution_options(stream_results=True)
+
+        with engine.connect() as conn:
+            # not this one
+            result = conn.execution_options(stream_results=False).execute(s)
+            assert not self._is_server_side(result.cursor)
+
+    def test_aliases_and_ss(self):
+        engine = self._fixture(False)
+        s1 = (
+            select(sql.literal_column("1").label("x"))
+            .execution_options(stream_results=True)
+            .subquery()
+        )
+
+        # options don't propagate out when subquery is used as a FROM clause
+        with engine.begin() as conn:
+            result = conn.execute(s1.select())
+            assert not self._is_server_side(result.cursor)
+            result.close()
+
+        s2 = select(1).select_from(s1)
+        with engine.begin() as conn:
+            result = conn.execute(s2)
+            assert not self._is_server_side(result.cursor)
+            result.close()
+
+    def test_roundtrip_fetchall(self, metadata):
+        md = self.metadata
+
+        engine = self._fixture(True)
+        test_table = Table(
+            "test_table",
+            md,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+        )
+
+        with engine.begin() as connection:
+            test_table.create(connection, checkfirst=True)
+            connection.execute(test_table.insert(), dict(data="data1"))
+            connection.execute(test_table.insert(), dict(data="data2"))
+            eq_(
+                connection.execute(
+                    test_table.select().order_by(test_table.c.id)
+                ).fetchall(),
+                [(1, "data1"), (2, "data2")],
+            )
+            connection.execute(
+                test_table.update()
+                .where(test_table.c.id == 2)
+                .values(data=test_table.c.data + " updated")
+            )
+            eq_(
+                connection.execute(
+                    test_table.select().order_by(test_table.c.id)
+                ).fetchall(),
+                [(1, "data1"), (2, "data2 updated")],
+            )
+            connection.execute(test_table.delete())
+            eq_(
+                connection.scalar(
+                    select(func.count("*")).select_from(test_table)
+                ),
+                0,
+            )
+
+    def test_roundtrip_fetchmany(self, metadata):
+        md = self.metadata
+
+        engine = self._fixture(True)
+        test_table = Table(
+            "test_table",
+            md,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("data", String(50)),
+        )
+
+        with engine.begin() as connection:
+            test_table.create(connection, checkfirst=True)
+            connection.execute(
+                test_table.insert(),
+                [dict(data="data%d" % i) for i in range(1, 20)],
+            )
+
+            result = connection.execute(
+                test_table.select().order_by(test_table.c.id)
+            )
+
+            eq_(
+                result.fetchmany(5),
+                [(i, "data%d" % i) for i in range(1, 6)],
+            )
+            eq_(
+                result.fetchmany(10),
+                [(i, "data%d" % i) for i in range(6, 16)],
+            )
+            eq_(result.fetchall(), [(i, "data%d" % i) for i in range(16, 20)])
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_rowcount.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_rowcount.py
new file mode 100644
index 00000000..59953fff
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_rowcount.py
@@ -0,0 +1,258 @@
+# testing/suite/test_rowcount.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from sqlalchemy import bindparam
+from sqlalchemy import Column
+from sqlalchemy import Integer
+from sqlalchemy import MetaData
+from sqlalchemy import select
+from sqlalchemy import String
+from sqlalchemy import Table
+from sqlalchemy import testing
+from sqlalchemy import text
+from sqlalchemy.testing import eq_
+from sqlalchemy.testing import fixtures
+
+
+class RowCountTest(fixtures.TablesTest):
+    """test rowcount functionality"""
+
+    __requires__ = ("sane_rowcount",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "employees",
+            metadata,
+            Column(
+                "employee_id",
+                Integer,
+                autoincrement=False,
+                primary_key=True,
+            ),
+            Column("name", String(50)),
+            Column("department", String(1)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        cls.data = data = [
+            ("Angela", "A"),
+            ("Andrew", "A"),
+            ("Anand", "A"),
+            ("Bob", "B"),
+            ("Bobette", "B"),
+            ("Buffy", "B"),
+            ("Charlie", "C"),
+            ("Cynthia", "C"),
+            ("Chris", "C"),
+        ]
+
+        employees_table = cls.tables.employees
+        connection.execute(
+            employees_table.insert(),
+            [
+                {"employee_id": i, "name": n, "department": d}
+                for i, (n, d) in enumerate(data)
+            ],
+        )
+
+    def test_basic(self, connection):
+        employees_table = self.tables.employees
+        s = select(
+            employees_table.c.name, employees_table.c.department
+        ).order_by(employees_table.c.employee_id)
+        rows = connection.execute(s).fetchall()
+
+        eq_(rows, self.data)
+
+    @testing.variation("statement", ["update", "delete", "insert", "select"])
+    @testing.variation("close_first", [True, False])
+    def test_non_rowcount_scenarios_no_raise(
+        self, connection, statement, close_first
+    ):
+        employees_table = self.tables.employees
+
+        # WHERE matches 3, 3 rows changed
+        department = employees_table.c.department
+
+        if statement.update:
+            r = connection.execute(
+                employees_table.update().where(department == "C"),
+                {"department": "Z"},
+            )
+        elif statement.delete:
+            r = connection.execute(
+                employees_table.delete().where(department == "C"),
+                {"department": "Z"},
+            )
+        elif statement.insert:
+            r = connection.execute(
+                employees_table.insert(),
+                [
+                    {"employee_id": 25, "name": "none 1", "department": "X"},
+                    {"employee_id": 26, "name": "none 2", "department": "Z"},
+                    {"employee_id": 27, "name": "none 3", "department": "Z"},
+                ],
+            )
+        elif statement.select:
+            s = select(
+                employees_table.c.name, employees_table.c.department
+            ).where(employees_table.c.department == "C")
+            r = connection.execute(s)
+            r.all()
+        else:
+            statement.fail()
+
+        if close_first:
+            r.close()
+
+        assert r.rowcount in (-1, 3)
+
+    def test_update_rowcount1(self, connection):
+        employees_table = self.tables.employees
+
+        # WHERE matches 3, 3 rows changed
+        department = employees_table.c.department
+        r = connection.execute(
+            employees_table.update().where(department == "C"),
+            {"department": "Z"},
+        )
+        assert r.rowcount == 3
+
+    def test_update_rowcount2(self, connection):
+        employees_table = self.tables.employees
+
+        # WHERE matches 3, 0 rows changed
+        department = employees_table.c.department
+
+        r = connection.execute(
+            employees_table.update().where(department == "C"),
+            {"department": "C"},
+        )
+        eq_(r.rowcount, 3)
+
+    @testing.variation("implicit_returning", [True, False])
+    @testing.variation(
+        "dml",
+        [
+            ("update", testing.requires.update_returning),
+            ("delete", testing.requires.delete_returning),
+        ],
+    )
+    def test_update_delete_rowcount_return_defaults(
+        self, connection, implicit_returning, dml
+    ):
+        """note this test should succeed for all RETURNING backends
+        as of 2.0.  In
+        Idf28379f8705e403a3c6a937f6a798a042ef2540 we changed rowcount to use
+        len(rows) when we have implicit returning
+
+        """
+
+        if implicit_returning:
+            employees_table = self.tables.employees
+        else:
+            employees_table = Table(
+                "employees",
+                MetaData(),
+                Column(
+                    "employee_id",
+                    Integer,
+                    autoincrement=False,
+                    primary_key=True,
+                ),
+                Column("name", String(50)),
+                Column("department", String(1)),
+                implicit_returning=False,
+            )
+
+        department = employees_table.c.department
+
+        if dml.update:
+            stmt = (
+                employees_table.update()
+                .where(department == "C")
+                .values(name=employees_table.c.department + "Z")
+                .return_defaults()
+            )
+        elif dml.delete:
+            stmt = (
+                employees_table.delete()
+                .where(department == "C")
+                .return_defaults()
+            )
+        else:
+            dml.fail()
+
+        r = connection.execute(stmt)
+        eq_(r.rowcount, 3)
+
+    def test_raw_sql_rowcount(self, connection):
+        # test issue #3622, make sure eager rowcount is called for text
+        result = connection.exec_driver_sql(
+            "update employees set department='Z' where department='C'"
+        )
+        eq_(result.rowcount, 3)
+
+    def test_text_rowcount(self, connection):
+        # test issue #3622, make sure eager rowcount is called for text
+        result = connection.execute(
+            text("update employees set department='Z' where department='C'")
+        )
+        eq_(result.rowcount, 3)
+
+    def test_delete_rowcount(self, connection):
+        employees_table = self.tables.employees
+
+        # WHERE matches 3, 3 rows deleted
+        department = employees_table.c.department
+        r = connection.execute(
+            employees_table.delete().where(department == "C")
+        )
+        eq_(r.rowcount, 3)
+
+    @testing.requires.sane_multi_rowcount
+    def test_multi_update_rowcount(self, connection):
+        employees_table = self.tables.employees
+        stmt = (
+            employees_table.update()
+            .where(employees_table.c.name == bindparam("emp_name"))
+            .values(department="C")
+        )
+
+        r = connection.execute(
+            stmt,
+            [
+                {"emp_name": "Bob"},
+                {"emp_name": "Cynthia"},
+                {"emp_name": "nonexistent"},
+            ],
+        )
+
+        eq_(r.rowcount, 2)
+
+    @testing.requires.sane_multi_rowcount
+    def test_multi_delete_rowcount(self, connection):
+        employees_table = self.tables.employees
+
+        stmt = employees_table.delete().where(
+            employees_table.c.name == bindparam("emp_name")
+        )
+
+        r = connection.execute(
+            stmt,
+            [
+                {"emp_name": "Bob"},
+                {"emp_name": "Cynthia"},
+                {"emp_name": "nonexistent"},
+            ],
+        )
+
+        eq_(r.rowcount, 2)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_select.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_select.py
new file mode 100644
index 00000000..7eb5cd00
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_select.py
@@ -0,0 +1,1999 @@
+# testing/suite/test_select.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import collections.abc as collections_abc
+import itertools
+
+from .. import AssertsCompiledSQL
+from .. import AssertsExecutionResults
+from .. import config
+from .. import fixtures
+from ..assertions import assert_raises
+from ..assertions import eq_
+from ..assertions import in_
+from ..assertsql import CursorSQL
+from ..schema import Column
+from ..schema import Table
+from ... import bindparam
+from ... import case
+from ... import column
+from ... import Computed
+from ... import exists
+from ... import false
+from ... import ForeignKey
+from ... import func
+from ... import Identity
+from ... import Integer
+from ... import literal
+from ... import literal_column
+from ... import null
+from ... import select
+from ... import String
+from ... import table
+from ... import testing
+from ... import text
+from ... import true
+from ... import tuple_
+from ... import TupleType
+from ... import union
+from ... import values
+from ...exc import DatabaseError
+from ...exc import ProgrammingError
+
+
+class CollateTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(100)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "data": "collate data1"},
+                {"id": 2, "data": "collate data2"},
+            ],
+        )
+
+    def _assert_result(self, select, result):
+        with config.db.connect() as conn:
+            eq_(conn.execute(select).fetchall(), result)
+
+    @testing.requires.order_by_collation
+    def test_collate_order_by(self):
+        collation = testing.requires.get_order_by_collation(testing.config)
+
+        self._assert_result(
+            select(self.tables.some_table).order_by(
+                self.tables.some_table.c.data.collate(collation).asc()
+            ),
+            [(1, "collate data1"), (2, "collate data2")],
+        )
+
+
+class OrderByLabelTest(fixtures.TablesTest):
+    """Test the dialect sends appropriate ORDER BY expressions when
+    labels are used.
+
+    This essentially exercises the "supports_simple_order_by_label"
+    setting.
+
+    """
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+            Column("q", String(50)),
+            Column("p", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2, "q": "q1", "p": "p3"},
+                {"id": 2, "x": 2, "y": 3, "q": "q2", "p": "p2"},
+                {"id": 3, "x": 3, "y": 4, "q": "q3", "p": "p1"},
+            ],
+        )
+
+    def _assert_result(self, select, result):
+        with config.db.connect() as conn:
+            eq_(conn.execute(select).fetchall(), result)
+
+    def test_plain(self):
+        table = self.tables.some_table
+        lx = table.c.x.label("lx")
+        self._assert_result(select(lx).order_by(lx), [(1,), (2,), (3,)])
+
+    def test_composed_int(self):
+        table = self.tables.some_table
+        lx = (table.c.x + table.c.y).label("lx")
+        self._assert_result(select(lx).order_by(lx), [(3,), (5,), (7,)])
+
+    def test_composed_multiple(self):
+        table = self.tables.some_table
+        lx = (table.c.x + table.c.y).label("lx")
+        ly = (func.lower(table.c.q) + table.c.p).label("ly")
+        self._assert_result(
+            select(lx, ly).order_by(lx, ly.desc()),
+            [(3, "q1p3"), (5, "q2p2"), (7, "q3p1")],
+        )
+
+    def test_plain_desc(self):
+        table = self.tables.some_table
+        lx = table.c.x.label("lx")
+        self._assert_result(select(lx).order_by(lx.desc()), [(3,), (2,), (1,)])
+
+    def test_composed_int_desc(self):
+        table = self.tables.some_table
+        lx = (table.c.x + table.c.y).label("lx")
+        self._assert_result(select(lx).order_by(lx.desc()), [(7,), (5,), (3,)])
+
+    @testing.requires.group_by_complex_expression
+    def test_group_by_composed(self):
+        table = self.tables.some_table
+        expr = (table.c.x + table.c.y).label("lx")
+        stmt = (
+            select(func.count(table.c.id), expr).group_by(expr).order_by(expr)
+        )
+        self._assert_result(stmt, [(1, 3), (1, 5), (1, 7)])
+
+
+class ValuesExpressionTest(fixtures.TestBase):
+    __requires__ = ("table_value_constructor",)
+
+    __backend__ = True
+
+    def test_tuples(self, connection):
+        value_expr = values(
+            column("id", Integer), column("name", String), name="my_values"
+        ).data([(1, "name1"), (2, "name2"), (3, "name3")])
+
+        eq_(
+            connection.execute(select(value_expr)).all(),
+            [(1, "name1"), (2, "name2"), (3, "name3")],
+        )
+
+
+class FetchLimitOffsetTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2},
+                {"id": 2, "x": 2, "y": 3},
+                {"id": 3, "x": 3, "y": 4},
+                {"id": 4, "x": 4, "y": 5},
+                {"id": 5, "x": 4, "y": 6},
+            ],
+        )
+
+    def _assert_result(
+        self, connection, select, result, params=(), set_=False
+    ):
+        if set_:
+            query_res = connection.execute(select, params).fetchall()
+            eq_(len(query_res), len(result))
+            eq_(set(query_res), set(result))
+
+        else:
+            eq_(connection.execute(select, params).fetchall(), result)
+
+    def _assert_result_str(self, select, result, params=()):
+        with config.db.connect() as conn:
+            eq_(conn.exec_driver_sql(select, params).fetchall(), result)
+
+    def test_simple_limit(self, connection):
+        table = self.tables.some_table
+        stmt = select(table).order_by(table.c.id)
+        self._assert_result(
+            connection,
+            stmt.limit(2),
+            [(1, 1, 2), (2, 2, 3)],
+        )
+        self._assert_result(
+            connection,
+            stmt.limit(3),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4)],
+        )
+
+    def test_limit_render_multiple_times(self, connection):
+        table = self.tables.some_table
+        stmt = select(table.c.id).limit(1).scalar_subquery()
+
+        u = union(select(stmt), select(stmt)).subquery().select()
+
+        self._assert_result(
+            connection,
+            u,
+            [
+                (1,),
+            ],
+        )
+
+    @testing.requires.fetch_first
+    def test_simple_fetch(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).fetch(2),
+            [(1, 1, 2), (2, 2, 3)],
+        )
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).fetch(3),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4)],
+        )
+
+    @testing.requires.offset
+    def test_simple_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(2),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+        )
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(3),
+            [(4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.combinations(
+        ([(2, 0), (2, 1), (3, 2)]),
+        ([(2, 1), (2, 0), (3, 2)]),
+        ([(3, 1), (2, 1), (3, 1)]),
+        argnames="cases",
+    )
+    @testing.requires.offset
+    def test_simple_limit_offset(self, connection, cases):
+        table = self.tables.some_table
+        connection = connection.execution_options(compiled_cache={})
+
+        assert_data = [(1, 1, 2), (2, 2, 3), (3, 3, 4), (4, 4, 5), (5, 4, 6)]
+
+        for limit, offset in cases:
+            expected = assert_data[offset : offset + limit]
+            self._assert_result(
+                connection,
+                select(table).order_by(table.c.id).limit(limit).offset(offset),
+                expected,
+            )
+
+    @testing.requires.fetch_first
+    def test_simple_fetch_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).fetch(2).offset(1),
+            [(2, 2, 3), (3, 3, 4)],
+        )
+
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).fetch(3).offset(2),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.requires.fetch_no_order_by
+    def test_fetch_offset_no_order(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).fetch(10),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            set_=True,
+        )
+
+    @testing.requires.offset
+    def test_simple_offset_zero(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(0),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4), (4, 4, 5), (5, 4, 6)],
+        )
+
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(1),
+            [(2, 2, 3), (3, 3, 4), (4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.requires.offset
+    def test_limit_offset_nobinds(self):
+        """test that 'literal binds' mode works - no bound params."""
+
+        table = self.tables.some_table
+        stmt = select(table).order_by(table.c.id).limit(2).offset(1)
+        sql = stmt.compile(
+            dialect=config.db.dialect, compile_kwargs={"literal_binds": True}
+        )
+        sql = str(sql)
+
+        self._assert_result_str(sql, [(2, 2, 3), (3, 3, 4)])
+
+    @testing.requires.fetch_first
+    def test_fetch_offset_nobinds(self):
+        """test that 'literal binds' mode works - no bound params."""
+
+        table = self.tables.some_table
+        stmt = select(table).order_by(table.c.id).fetch(2).offset(1)
+        sql = stmt.compile(
+            dialect=config.db.dialect, compile_kwargs={"literal_binds": True}
+        )
+        sql = str(sql)
+
+        self._assert_result_str(sql, [(2, 2, 3), (3, 3, 4)])
+
+    @testing.requires.bound_limit_offset
+    def test_bound_limit(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).limit(bindparam("l")),
+            [(1, 1, 2), (2, 2, 3)],
+            params={"l": 2},
+        )
+
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).limit(bindparam("l")),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4)],
+            params={"l": 3},
+        )
+
+    @testing.requires.bound_limit_offset
+    def test_bound_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(bindparam("o")),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            params={"o": 2},
+        )
+
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).offset(bindparam("o")),
+            [(2, 2, 3), (3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            params={"o": 1},
+        )
+
+    @testing.requires.bound_limit_offset
+    def test_bound_limit_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(bindparam("l"))
+            .offset(bindparam("o")),
+            [(2, 2, 3), (3, 3, 4)],
+            params={"l": 2, "o": 1},
+        )
+
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(bindparam("l"))
+            .offset(bindparam("o")),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            params={"l": 3, "o": 2},
+        )
+
+    @testing.requires.fetch_first
+    def test_bound_fetch_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .fetch(bindparam("f"))
+            .offset(bindparam("o")),
+            [(2, 2, 3), (3, 3, 4)],
+            params={"f": 2, "o": 1},
+        )
+
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .fetch(bindparam("f"))
+            .offset(bindparam("o")),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            params={"f": 3, "o": 2},
+        )
+
+    @testing.requires.sql_expression_limit_offset
+    def test_expr_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .offset(literal_column("1") + literal_column("2")),
+            [(4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.requires.sql_expression_limit_offset
+    def test_expr_limit(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(literal_column("1") + literal_column("2")),
+            [(1, 1, 2), (2, 2, 3), (3, 3, 4)],
+        )
+
+    @testing.requires.sql_expression_limit_offset
+    def test_expr_limit_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(literal_column("1") + literal_column("1"))
+            .offset(literal_column("1") + literal_column("1")),
+            [(3, 3, 4), (4, 4, 5)],
+        )
+
+    @testing.requires.fetch_first
+    @testing.requires.fetch_expression
+    def test_expr_fetch_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .fetch(literal_column("1") + literal_column("1"))
+            .offset(literal_column("1") + literal_column("1")),
+            [(3, 3, 4), (4, 4, 5)],
+        )
+
+    @testing.requires.sql_expression_limit_offset
+    def test_simple_limit_expr_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(2)
+            .offset(literal_column("1") + literal_column("1")),
+            [(3, 3, 4), (4, 4, 5)],
+        )
+
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(3)
+            .offset(literal_column("1") + literal_column("1")),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.requires.sql_expression_limit_offset
+    def test_expr_limit_simple_offset(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(literal_column("1") + literal_column("1"))
+            .offset(2),
+            [(3, 3, 4), (4, 4, 5)],
+        )
+
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .limit(literal_column("1") + literal_column("1"))
+            .offset(1),
+            [(2, 2, 3), (3, 3, 4)],
+        )
+
+    @testing.requires.fetch_ties
+    def test_simple_fetch_ties(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.x.desc()).fetch(1, with_ties=True),
+            [(4, 4, 5), (5, 4, 6)],
+            set_=True,
+        )
+
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.x.desc()).fetch(3, with_ties=True),
+            [(3, 3, 4), (4, 4, 5), (5, 4, 6)],
+            set_=True,
+        )
+
+    @testing.requires.fetch_ties
+    @testing.requires.fetch_offset_with_options
+    def test_fetch_offset_ties(self, connection):
+        table = self.tables.some_table
+        fa = connection.execute(
+            select(table)
+            .order_by(table.c.x)
+            .fetch(2, with_ties=True)
+            .offset(2)
+        ).fetchall()
+        eq_(fa[0], (3, 3, 4))
+        eq_(set(fa), {(3, 3, 4), (4, 4, 5), (5, 4, 6)})
+
+    @testing.requires.fetch_ties
+    @testing.requires.fetch_offset_with_options
+    def test_fetch_offset_ties_exact_number(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.x)
+            .fetch(2, with_ties=True)
+            .offset(1),
+            [(2, 2, 3), (3, 3, 4)],
+        )
+
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.x)
+            .fetch(3, with_ties=True)
+            .offset(3),
+            [(4, 4, 5), (5, 4, 6)],
+        )
+
+    @testing.requires.fetch_percent
+    def test_simple_fetch_percent(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table).order_by(table.c.id).fetch(20, percent=True),
+            [(1, 1, 2)],
+        )
+
+    @testing.requires.fetch_percent
+    @testing.requires.fetch_offset_with_options
+    def test_fetch_offset_percent(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.id)
+            .fetch(40, percent=True)
+            .offset(1),
+            [(2, 2, 3), (3, 3, 4)],
+        )
+
+    @testing.requires.fetch_ties
+    @testing.requires.fetch_percent
+    def test_simple_fetch_percent_ties(self, connection):
+        table = self.tables.some_table
+        self._assert_result(
+            connection,
+            select(table)
+            .order_by(table.c.x.desc())
+            .fetch(20, percent=True, with_ties=True),
+            [(4, 4, 5), (5, 4, 6)],
+            set_=True,
+        )
+
+    @testing.requires.fetch_ties
+    @testing.requires.fetch_percent
+    @testing.requires.fetch_offset_with_options
+    def test_fetch_offset_percent_ties(self, connection):
+        table = self.tables.some_table
+        fa = connection.execute(
+            select(table)
+            .order_by(table.c.x)
+            .fetch(40, percent=True, with_ties=True)
+            .offset(2)
+        ).fetchall()
+        eq_(fa[0], (3, 3, 4))
+        eq_(set(fa), {(3, 3, 4), (4, 4, 5), (5, 4, 6)})
+
+
+class SameNamedSchemaTableTest(fixtures.TablesTest):
+    """tests for #7471"""
+
+    __backend__ = True
+
+    __requires__ = ("schemas",)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            schema=config.test_schema,
+        )
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column(
+                "some_table_id",
+                Integer,
+                # ForeignKey("%s.some_table.id" % config.test_schema),
+                nullable=False,
+            ),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        some_table, some_table_schema = cls.tables(
+            "some_table", "%s.some_table" % config.test_schema
+        )
+        connection.execute(some_table_schema.insert(), {"id": 1})
+        connection.execute(some_table.insert(), {"id": 1, "some_table_id": 1})
+
+    def test_simple_join_both_tables(self, connection):
+        some_table, some_table_schema = self.tables(
+            "some_table", "%s.some_table" % config.test_schema
+        )
+
+        eq_(
+            connection.execute(
+                select(some_table, some_table_schema).join_from(
+                    some_table,
+                    some_table_schema,
+                    some_table.c.some_table_id == some_table_schema.c.id,
+                )
+            ).first(),
+            (1, 1, 1),
+        )
+
+    def test_simple_join_whereclause_only(self, connection):
+        some_table, some_table_schema = self.tables(
+            "some_table", "%s.some_table" % config.test_schema
+        )
+
+        eq_(
+            connection.execute(
+                select(some_table)
+                .join_from(
+                    some_table,
+                    some_table_schema,
+                    some_table.c.some_table_id == some_table_schema.c.id,
+                )
+                .where(some_table.c.id == 1)
+            ).first(),
+            (1, 1),
+        )
+
+    def test_subquery(self, connection):
+        some_table, some_table_schema = self.tables(
+            "some_table", "%s.some_table" % config.test_schema
+        )
+
+        subq = (
+            select(some_table)
+            .join_from(
+                some_table,
+                some_table_schema,
+                some_table.c.some_table_id == some_table_schema.c.id,
+            )
+            .where(some_table.c.id == 1)
+            .subquery()
+        )
+
+        eq_(
+            connection.execute(
+                select(some_table, subq.c.id)
+                .join_from(
+                    some_table,
+                    subq,
+                    some_table.c.some_table_id == subq.c.id,
+                )
+                .where(some_table.c.id == 1)
+            ).first(),
+            (1, 1, 1),
+        )
+
+
+class JoinTest(fixtures.TablesTest):
+    __backend__ = True
+
+    def _assert_result(self, select, result, params=()):
+        with config.db.connect() as conn:
+            eq_(conn.execute(select, params).fetchall(), result)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table("a", metadata, Column("id", Integer, primary_key=True))
+        Table(
+            "b",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("a_id", ForeignKey("a.id"), nullable=False),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.a.insert(),
+            [{"id": 1}, {"id": 2}, {"id": 3}, {"id": 4}, {"id": 5}],
+        )
+
+        connection.execute(
+            cls.tables.b.insert(),
+            [
+                {"id": 1, "a_id": 1},
+                {"id": 2, "a_id": 1},
+                {"id": 4, "a_id": 2},
+                {"id": 5, "a_id": 3},
+            ],
+        )
+
+    def test_inner_join_fk(self):
+        a, b = self.tables("a", "b")
+
+        stmt = select(a, b).select_from(a.join(b)).order_by(a.c.id, b.c.id)
+
+        self._assert_result(stmt, [(1, 1, 1), (1, 2, 1), (2, 4, 2), (3, 5, 3)])
+
+    def test_inner_join_true(self):
+        a, b = self.tables("a", "b")
+
+        stmt = (
+            select(a, b)
+            .select_from(a.join(b, true()))
+            .order_by(a.c.id, b.c.id)
+        )
+
+        self._assert_result(
+            stmt,
+            [
+                (a, b, c)
+                for (a,), (b, c) in itertools.product(
+                    [(1,), (2,), (3,), (4,), (5,)],
+                    [(1, 1), (2, 1), (4, 2), (5, 3)],
+                )
+            ],
+        )
+
+    def test_inner_join_false(self):
+        a, b = self.tables("a", "b")
+
+        stmt = (
+            select(a, b)
+            .select_from(a.join(b, false()))
+            .order_by(a.c.id, b.c.id)
+        )
+
+        self._assert_result(stmt, [])
+
+    def test_outer_join_false(self):
+        a, b = self.tables("a", "b")
+
+        stmt = (
+            select(a, b)
+            .select_from(a.outerjoin(b, false()))
+            .order_by(a.c.id, b.c.id)
+        )
+
+        self._assert_result(
+            stmt,
+            [
+                (1, None, None),
+                (2, None, None),
+                (3, None, None),
+                (4, None, None),
+                (5, None, None),
+            ],
+        )
+
+    def test_outer_join_fk(self):
+        a, b = self.tables("a", "b")
+
+        stmt = select(a, b).select_from(a.join(b)).order_by(a.c.id, b.c.id)
+
+        self._assert_result(stmt, [(1, 1, 1), (1, 2, 1), (2, 4, 2), (3, 5, 3)])
+
+
+class CompoundSelectTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2},
+                {"id": 2, "x": 2, "y": 3},
+                {"id": 3, "x": 3, "y": 4},
+                {"id": 4, "x": 4, "y": 5},
+            ],
+        )
+
+    def _assert_result(self, select, result, params=()):
+        with config.db.connect() as conn:
+            eq_(conn.execute(select, params).fetchall(), result)
+
+    def test_plain_union(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2)
+        s2 = select(table).where(table.c.id == 3)
+
+        u1 = union(s1, s2)
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    def test_select_from_plain_union(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2)
+        s2 = select(table).where(table.c.id == 3)
+
+        u1 = union(s1, s2).alias().select()
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    @testing.requires.order_by_col_from_union
+    @testing.requires.parens_in_union_contained_select_w_limit_offset
+    def test_limit_offset_selectable_in_unions(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).limit(1).order_by(table.c.id)
+        s2 = select(table).where(table.c.id == 3).limit(1).order_by(table.c.id)
+
+        u1 = union(s1, s2).limit(2)
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    @testing.requires.parens_in_union_contained_select_wo_limit_offset
+    def test_order_by_selectable_in_unions(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).order_by(table.c.id)
+        s2 = select(table).where(table.c.id == 3).order_by(table.c.id)
+
+        u1 = union(s1, s2).limit(2)
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    def test_distinct_selectable_in_unions(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).distinct()
+        s2 = select(table).where(table.c.id == 3).distinct()
+
+        u1 = union(s1, s2).limit(2)
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    @testing.requires.parens_in_union_contained_select_w_limit_offset
+    def test_limit_offset_in_unions_from_alias(self):
+        table = self.tables.some_table
+        s1 = select(table).where(table.c.id == 2).limit(1).order_by(table.c.id)
+        s2 = select(table).where(table.c.id == 3).limit(1).order_by(table.c.id)
+
+        # this necessarily has double parens
+        u1 = union(s1, s2).alias()
+        self._assert_result(
+            u1.select().limit(2).order_by(u1.c.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+    def test_limit_offset_aliased_selectable_in_unions(self):
+        table = self.tables.some_table
+        s1 = (
+            select(table)
+            .where(table.c.id == 2)
+            .limit(1)
+            .order_by(table.c.id)
+            .alias()
+            .select()
+        )
+        s2 = (
+            select(table)
+            .where(table.c.id == 3)
+            .limit(1)
+            .order_by(table.c.id)
+            .alias()
+            .select()
+        )
+
+        u1 = union(s1, s2).limit(2)
+        self._assert_result(
+            u1.order_by(u1.selected_columns.id), [(2, 2, 3), (3, 3, 4)]
+        )
+
+
+class PostCompileParamsTest(
+    AssertsExecutionResults, AssertsCompiledSQL, fixtures.TablesTest
+):
+    __backend__ = True
+
+    __requires__ = ("standard_cursor_sql",)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+            Column("z", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2, "z": "z1"},
+                {"id": 2, "x": 2, "y": 3, "z": "z2"},
+                {"id": 3, "x": 3, "y": 4, "z": "z3"},
+                {"id": 4, "x": 4, "y": 5, "z": "z4"},
+            ],
+        )
+
+    def test_compile(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            table.c.x == bindparam("q", literal_execute=True)
+        )
+
+        self.assert_compile(
+            stmt,
+            "SELECT some_table.id FROM some_table "
+            "WHERE some_table.x = __[POSTCOMPILE_q]",
+            {},
+        )
+
+    def test_compile_literal_binds(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            table.c.x == bindparam("q", 10, literal_execute=True)
+        )
+
+        self.assert_compile(
+            stmt,
+            "SELECT some_table.id FROM some_table WHERE some_table.x = 10",
+            {},
+            literal_binds=True,
+        )
+
+    def test_execute(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            table.c.x == bindparam("q", literal_execute=True)
+        )
+
+        with self.sql_execution_asserter() as asserter:
+            with config.db.connect() as conn:
+                conn.execute(stmt, dict(q=10))
+
+        asserter.assert_(
+            CursorSQL(
+                "SELECT some_table.id \nFROM some_table "
+                "\nWHERE some_table.x = 10",
+                () if config.db.dialect.positional else {},
+            )
+        )
+
+    def test_execute_expanding_plus_literal_execute(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            table.c.x.in_(bindparam("q", expanding=True, literal_execute=True))
+        )
+
+        with self.sql_execution_asserter() as asserter:
+            with config.db.connect() as conn:
+                conn.execute(stmt, dict(q=[5, 6, 7]))
+
+        asserter.assert_(
+            CursorSQL(
+                "SELECT some_table.id \nFROM some_table "
+                "\nWHERE some_table.x IN (5, 6, 7)",
+                () if config.db.dialect.positional else {},
+            )
+        )
+
+    @testing.requires.tuple_in
+    def test_execute_tuple_expanding_plus_literal_execute(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            tuple_(table.c.x, table.c.y).in_(
+                bindparam("q", expanding=True, literal_execute=True)
+            )
+        )
+
+        with self.sql_execution_asserter() as asserter:
+            with config.db.connect() as conn:
+                conn.execute(stmt, dict(q=[(5, 10), (12, 18)]))
+
+        asserter.assert_(
+            CursorSQL(
+                "SELECT some_table.id \nFROM some_table "
+                "\nWHERE (some_table.x, some_table.y) "
+                "IN (%s(5, 10), (12, 18))"
+                % ("VALUES " if config.db.dialect.tuple_in_values else ""),
+                () if config.db.dialect.positional else {},
+            )
+        )
+
+    @testing.requires.tuple_in
+    def test_execute_tuple_expanding_plus_literal_heterogeneous_execute(self):
+        table = self.tables.some_table
+
+        stmt = select(table.c.id).where(
+            tuple_(table.c.x, table.c.z).in_(
+                bindparam("q", expanding=True, literal_execute=True)
+            )
+        )
+
+        with self.sql_execution_asserter() as asserter:
+            with config.db.connect() as conn:
+                conn.execute(stmt, dict(q=[(5, "z1"), (12, "z3")]))
+
+        asserter.assert_(
+            CursorSQL(
+                "SELECT some_table.id \nFROM some_table "
+                "\nWHERE (some_table.x, some_table.z) "
+                "IN (%s(5, 'z1'), (12, 'z3'))"
+                % ("VALUES " if config.db.dialect.tuple_in_values else ""),
+                () if config.db.dialect.positional else {},
+            )
+        )
+
+
+class ExpandingBoundInTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("x", Integer),
+            Column("y", Integer),
+            Column("z", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "x": 1, "y": 2, "z": "z1"},
+                {"id": 2, "x": 2, "y": 3, "z": "z2"},
+                {"id": 3, "x": 3, "y": 4, "z": "z3"},
+                {"id": 4, "x": 4, "y": 5, "z": "z4"},
+            ],
+        )
+
+    def _assert_result(self, select, result, params=()):
+        with config.db.connect() as conn:
+            eq_(conn.execute(select, params).fetchall(), result)
+
+    def test_multiple_empty_sets_bindparam(self):
+        # test that any anonymous aliasing used by the dialect
+        # is fine with duplicates
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_(bindparam("q")))
+            .where(table.c.y.in_(bindparam("p")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [], params={"q": [], "p": []})
+
+    def test_multiple_empty_sets_direct(self):
+        # test that any anonymous aliasing used by the dialect
+        # is fine with duplicates
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_([]))
+            .where(table.c.y.in_([]))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [])
+
+    @testing.requires.tuple_in_w_empty
+    def test_empty_heterogeneous_tuples_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(tuple_(table.c.x, table.c.z).in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [], params={"q": []})
+
+    @testing.requires.tuple_in_w_empty
+    def test_empty_heterogeneous_tuples_direct(self):
+        table = self.tables.some_table
+
+        def go(val, expected):
+            stmt = (
+                select(table.c.id)
+                .where(tuple_(table.c.x, table.c.z).in_(val))
+                .order_by(table.c.id)
+            )
+            self._assert_result(stmt, expected)
+
+        go([], [])
+        go([(2, "z2"), (3, "z3"), (4, "z4")], [(2,), (3,), (4,)])
+        go([], [])
+
+    @testing.requires.tuple_in_w_empty
+    def test_empty_homogeneous_tuples_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(tuple_(table.c.x, table.c.y).in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [], params={"q": []})
+
+    @testing.requires.tuple_in_w_empty
+    def test_empty_homogeneous_tuples_direct(self):
+        table = self.tables.some_table
+
+        def go(val, expected):
+            stmt = (
+                select(table.c.id)
+                .where(tuple_(table.c.x, table.c.y).in_(val))
+                .order_by(table.c.id)
+            )
+            self._assert_result(stmt, expected)
+
+        go([], [])
+        go([(1, 2), (2, 3), (3, 4)], [(1,), (2,), (3,)])
+        go([], [])
+
+    def test_bound_in_scalar_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(2,), (3,), (4,)], params={"q": [2, 3, 4]})
+
+    def test_bound_in_scalar_direct(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_([2, 3, 4]))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(2,), (3,), (4,)])
+
+    def test_nonempty_in_plus_empty_notin(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_([2, 3]))
+            .where(table.c.id.not_in([]))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(2,), (3,)])
+
+    def test_empty_in_plus_notempty_notin(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_([]))
+            .where(table.c.id.not_in([2, 3]))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [])
+
+    def test_typed_str_in(self):
+        """test related to #7292.
+
+        as a type is given to the bound param, there is no ambiguity
+        to the type of element.
+
+        """
+
+        stmt = text(
+            "select id FROM some_table WHERE z IN :q ORDER BY id"
+        ).bindparams(bindparam("q", type_=String, expanding=True))
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={"q": ["z2", "z3", "z4"]},
+        )
+
+    def test_untyped_str_in(self):
+        """test related to #7292.
+
+        for untyped expression, we look at the types of elements.
+        Test for Sequence to detect tuple in.  but not strings or bytes!
+        as always....
+
+        """
+
+        stmt = text(
+            "select id FROM some_table WHERE z IN :q ORDER BY id"
+        ).bindparams(bindparam("q", expanding=True))
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={"q": ["z2", "z3", "z4"]},
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_two_tuple_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(tuple_(table.c.x, table.c.y).in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(
+            stmt, [(2,), (3,), (4,)], params={"q": [(2, 3), (3, 4), (4, 5)]}
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_two_tuple_direct(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(tuple_(table.c.x, table.c.y).in_([(2, 3), (3, 4), (4, 5)]))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(2,), (3,), (4,)])
+
+    @testing.requires.tuple_in
+    def test_bound_in_heterogeneous_two_tuple_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(tuple_(table.c.x, table.c.z).in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={"q": [(2, "z2"), (3, "z3"), (4, "z4")]},
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_heterogeneous_two_tuple_direct(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(
+                tuple_(table.c.x, table.c.z).in_(
+                    [(2, "z2"), (3, "z3"), (4, "z4")]
+                )
+            )
+            .order_by(table.c.id)
+        )
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_heterogeneous_two_tuple_text_bindparam(self):
+        # note this becomes ARRAY if we dont use expanding
+        # explicitly right now
+        stmt = text(
+            "select id FROM some_table WHERE (x, z) IN :q ORDER BY id"
+        ).bindparams(bindparam("q", expanding=True))
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={"q": [(2, "z2"), (3, "z3"), (4, "z4")]},
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_heterogeneous_two_tuple_typed_bindparam_non_tuple(self):
+        class LikeATuple(collections_abc.Sequence):
+            def __init__(self, *data):
+                self._data = data
+
+            def __iter__(self):
+                return iter(self._data)
+
+            def __getitem__(self, idx):
+                return self._data[idx]
+
+            def __len__(self):
+                return len(self._data)
+
+        stmt = text(
+            "select id FROM some_table WHERE (x, z) IN :q ORDER BY id"
+        ).bindparams(
+            bindparam(
+                "q", type_=TupleType(Integer(), String()), expanding=True
+            )
+        )
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={
+                "q": [
+                    LikeATuple(2, "z2"),
+                    LikeATuple(3, "z3"),
+                    LikeATuple(4, "z4"),
+                ]
+            },
+        )
+
+    @testing.requires.tuple_in
+    def test_bound_in_heterogeneous_two_tuple_text_bindparam_non_tuple(self):
+        # note this becomes ARRAY if we dont use expanding
+        # explicitly right now
+
+        class LikeATuple(collections_abc.Sequence):
+            def __init__(self, *data):
+                self._data = data
+
+            def __iter__(self):
+                return iter(self._data)
+
+            def __getitem__(self, idx):
+                return self._data[idx]
+
+            def __len__(self):
+                return len(self._data)
+
+        stmt = text(
+            "select id FROM some_table WHERE (x, z) IN :q ORDER BY id"
+        ).bindparams(bindparam("q", expanding=True))
+        self._assert_result(
+            stmt,
+            [(2,), (3,), (4,)],
+            params={
+                "q": [
+                    LikeATuple(2, "z2"),
+                    LikeATuple(3, "z3"),
+                    LikeATuple(4, "z4"),
+                ]
+            },
+        )
+
+    def test_empty_set_against_integer_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [], params={"q": []})
+
+    def test_empty_set_against_integer_direct(self):
+        table = self.tables.some_table
+        stmt = select(table.c.id).where(table.c.x.in_([])).order_by(table.c.id)
+        self._assert_result(stmt, [])
+
+    def test_empty_set_against_integer_negation_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.x.not_in(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(1,), (2,), (3,), (4,)], params={"q": []})
+
+    def test_empty_set_against_integer_negation_direct(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id).where(table.c.x.not_in([])).order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(1,), (2,), (3,), (4,)])
+
+    def test_empty_set_against_string_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.z.in_(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [], params={"q": []})
+
+    def test_empty_set_against_string_direct(self):
+        table = self.tables.some_table
+        stmt = select(table.c.id).where(table.c.z.in_([])).order_by(table.c.id)
+        self._assert_result(stmt, [])
+
+    def test_empty_set_against_string_negation_bindparam(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id)
+            .where(table.c.z.not_in(bindparam("q")))
+            .order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(1,), (2,), (3,), (4,)], params={"q": []})
+
+    def test_empty_set_against_string_negation_direct(self):
+        table = self.tables.some_table
+        stmt = (
+            select(table.c.id).where(table.c.z.not_in([])).order_by(table.c.id)
+        )
+        self._assert_result(stmt, [(1,), (2,), (3,), (4,)])
+
+    def test_null_in_empty_set_is_false_bindparam(self, connection):
+        stmt = select(
+            case(
+                (
+                    null().in_(bindparam("foo", value=())),
+                    true(),
+                ),
+                else_=false(),
+            )
+        )
+        in_(connection.execute(stmt).fetchone()[0], (False, 0))
+
+    def test_null_in_empty_set_is_false_direct(self, connection):
+        stmt = select(
+            case(
+                (
+                    null().in_([]),
+                    true(),
+                ),
+                else_=false(),
+            )
+        )
+        in_(connection.execute(stmt).fetchone()[0], (False, 0))
+
+
+class LikeFunctionsTest(fixtures.TablesTest):
+    __backend__ = True
+
+    run_inserts = "once"
+    run_deletes = None
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [
+                {"id": 1, "data": "abcdefg"},
+                {"id": 2, "data": "ab/cdefg"},
+                {"id": 3, "data": "ab%cdefg"},
+                {"id": 4, "data": "ab_cdefg"},
+                {"id": 5, "data": "abcde/fg"},
+                {"id": 6, "data": "abcde%fg"},
+                {"id": 7, "data": "ab#cdefg"},
+                {"id": 8, "data": "ab9cdefg"},
+                {"id": 9, "data": "abcde#fg"},
+                {"id": 10, "data": "abcd9fg"},
+                {"id": 11, "data": None},
+            ],
+        )
+
+    def _test(self, expr, expected):
+        some_table = self.tables.some_table
+
+        with config.db.connect() as conn:
+            rows = {
+                value
+                for value, in conn.execute(select(some_table.c.id).where(expr))
+            }
+
+        eq_(rows, expected)
+
+    def test_startswith_unescaped(self):
+        col = self.tables.some_table.c.data
+        self._test(col.startswith("ab%c"), {1, 2, 3, 4, 5, 6, 7, 8, 9, 10})
+
+    def test_startswith_autoescape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.startswith("ab%c", autoescape=True), {3})
+
+    def test_startswith_sqlexpr(self):
+        col = self.tables.some_table.c.data
+        self._test(
+            col.startswith(literal_column("'ab%c'")),
+            {1, 2, 3, 4, 5, 6, 7, 8, 9, 10},
+        )
+
+    def test_startswith_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.startswith("ab##c", escape="#"), {7})
+
+    def test_startswith_autoescape_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.startswith("ab%c", autoescape=True, escape="#"), {3})
+        self._test(col.startswith("ab#c", autoescape=True, escape="#"), {7})
+
+    def test_endswith_unescaped(self):
+        col = self.tables.some_table.c.data
+        self._test(col.endswith("e%fg"), {1, 2, 3, 4, 5, 6, 7, 8, 9})
+
+    def test_endswith_sqlexpr(self):
+        col = self.tables.some_table.c.data
+        self._test(
+            col.endswith(literal_column("'e%fg'")), {1, 2, 3, 4, 5, 6, 7, 8, 9}
+        )
+
+    def test_endswith_autoescape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.endswith("e%fg", autoescape=True), {6})
+
+    def test_endswith_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.endswith("e##fg", escape="#"), {9})
+
+    def test_endswith_autoescape_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.endswith("e%fg", autoescape=True, escape="#"), {6})
+        self._test(col.endswith("e#fg", autoescape=True, escape="#"), {9})
+
+    def test_contains_unescaped(self):
+        col = self.tables.some_table.c.data
+        self._test(col.contains("b%cde"), {1, 2, 3, 4, 5, 6, 7, 8, 9})
+
+    def test_contains_autoescape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.contains("b%cde", autoescape=True), {3})
+
+    def test_contains_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.contains("b##cde", escape="#"), {7})
+
+    def test_contains_autoescape_escape(self):
+        col = self.tables.some_table.c.data
+        self._test(col.contains("b%cd", autoescape=True, escape="#"), {3})
+        self._test(col.contains("b#cd", autoescape=True, escape="#"), {7})
+
+    @testing.requires.regexp_match
+    def test_not_regexp_match(self):
+        col = self.tables.some_table.c.data
+        self._test(~col.regexp_match("a.cde"), {2, 3, 4, 7, 8, 10})
+
+    @testing.requires.regexp_replace
+    def test_regexp_replace(self):
+        col = self.tables.some_table.c.data
+        self._test(
+            col.regexp_replace("a.cde", "FOO").contains("FOO"), {1, 5, 6, 9}
+        )
+
+    @testing.requires.regexp_match
+    @testing.combinations(
+        ("a.cde", {1, 5, 6, 9}),
+        ("abc", {1, 5, 6, 9, 10}),
+        ("^abc", {1, 5, 6, 9, 10}),
+        ("9cde", {8}),
+        ("^a", set(range(1, 11))),
+        ("(b|c)", set(range(1, 11))),
+        ("^(b|c)", set()),
+    )
+    def test_regexp_match(self, text, expected):
+        col = self.tables.some_table.c.data
+        self._test(col.regexp_match(text), expected)
+
+
+class ComputedColumnTest(fixtures.TablesTest):
+    __backend__ = True
+    __requires__ = ("computed_columns",)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "square",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("side", Integer),
+            Column("area", Integer, Computed("side * side")),
+            Column("perimeter", Integer, Computed("4 * side")),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.square.insert(),
+            [{"id": 1, "side": 10}, {"id": 10, "side": 42}],
+        )
+
+    def test_select_all(self):
+        with config.db.connect() as conn:
+            res = conn.execute(
+                select(text("*"))
+                .select_from(self.tables.square)
+                .order_by(self.tables.square.c.id)
+            ).fetchall()
+            eq_(res, [(1, 10, 100, 40), (10, 42, 1764, 168)])
+
+    def test_select_columns(self):
+        with config.db.connect() as conn:
+            res = conn.execute(
+                select(
+                    self.tables.square.c.area, self.tables.square.c.perimeter
+                )
+                .select_from(self.tables.square)
+                .order_by(self.tables.square.c.id)
+            ).fetchall()
+            eq_(res, [(100, 40), (1764, 168)])
+
+
+class IdentityColumnTest(fixtures.TablesTest):
+    __backend__ = True
+    __requires__ = ("identity_columns",)
+    run_inserts = "once"
+    run_deletes = "once"
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "tbl_a",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                Identity(
+                    always=True, start=42, nominvalue=True, nomaxvalue=True
+                ),
+                primary_key=True,
+            ),
+            Column("desc", String(100)),
+        )
+        Table(
+            "tbl_b",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                Identity(increment=-5, start=0, minvalue=-1000, maxvalue=0),
+                primary_key=True,
+            ),
+            Column("desc", String(100)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.tbl_a.insert(),
+            [{"desc": "a"}, {"desc": "b"}],
+        )
+        connection.execute(
+            cls.tables.tbl_b.insert(),
+            [{"desc": "a"}, {"desc": "b"}],
+        )
+        connection.execute(
+            cls.tables.tbl_b.insert(),
+            [{"id": 42, "desc": "c"}],
+        )
+
+    def test_select_all(self, connection):
+        res = connection.execute(
+            select(text("*"))
+            .select_from(self.tables.tbl_a)
+            .order_by(self.tables.tbl_a.c.id)
+        ).fetchall()
+        eq_(res, [(42, "a"), (43, "b")])
+
+        res = connection.execute(
+            select(text("*"))
+            .select_from(self.tables.tbl_b)
+            .order_by(self.tables.tbl_b.c.id)
+        ).fetchall()
+        eq_(res, [(-5, "b"), (0, "a"), (42, "c")])
+
+    def test_select_columns(self, connection):
+        res = connection.execute(
+            select(self.tables.tbl_a.c.id).order_by(self.tables.tbl_a.c.id)
+        ).fetchall()
+        eq_(res, [(42,), (43,)])
+
+    @testing.requires.identity_columns_standard
+    def test_insert_always_error(self, connection):
+        def fn():
+            connection.execute(
+                self.tables.tbl_a.insert(),
+                [{"id": 200, "desc": "a"}],
+            )
+
+        assert_raises((DatabaseError, ProgrammingError), fn)
+
+
+class IdentityAutoincrementTest(fixtures.TablesTest):
+    __backend__ = True
+    __requires__ = ("autoincrement_without_sequence",)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "tbl",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                Identity(),
+                primary_key=True,
+                autoincrement=True,
+            ),
+            Column("desc", String(100)),
+        )
+
+    def test_autoincrement_with_identity(self, connection):
+        res = connection.execute(self.tables.tbl.insert(), {"desc": "row"})
+        res = connection.execute(self.tables.tbl.select()).first()
+        eq_(res, (1, "row"))
+
+
+class ExistsTest(fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "stuff",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.stuff.insert(),
+            [
+                {"id": 1, "data": "some data"},
+                {"id": 2, "data": "some data"},
+                {"id": 3, "data": "some data"},
+                {"id": 4, "data": "some other data"},
+            ],
+        )
+
+    def test_select_exists(self, connection):
+        stuff = self.tables.stuff
+        eq_(
+            connection.execute(
+                select(literal(1)).where(
+                    exists().where(stuff.c.data == "some data")
+                )
+            ).fetchall(),
+            [(1,)],
+        )
+
+    def test_select_exists_false(self, connection):
+        stuff = self.tables.stuff
+        eq_(
+            connection.execute(
+                select(literal(1)).where(
+                    exists().where(stuff.c.data == "no data")
+                )
+            ).fetchall(),
+            [],
+        )
+
+
+class DistinctOnTest(AssertsCompiledSQL, fixtures.TablesTest):
+    __backend__ = True
+
+    @testing.fails_if(testing.requires.supports_distinct_on)
+    def test_distinct_on(self):
+        stm = select("*").distinct(column("q")).select_from(table("foo"))
+        with testing.expect_deprecated(
+            "DISTINCT ON is currently supported only by the PostgreSQL "
+        ):
+            self.assert_compile(stm, "SELECT DISTINCT * FROM foo")
+
+
+class IsOrIsNotDistinctFromTest(fixtures.TablesTest):
+    __backend__ = True
+    __requires__ = ("supports_is_distinct_from",)
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "is_distinct_test",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("col_a", Integer, nullable=True),
+            Column("col_b", Integer, nullable=True),
+        )
+
+    @testing.combinations(
+        ("both_int_different", 0, 1, 1),
+        ("both_int_same", 1, 1, 0),
+        ("one_null_first", None, 1, 1),
+        ("one_null_second", 0, None, 1),
+        ("both_null", None, None, 0),
+        id_="iaaa",
+        argnames="col_a_value, col_b_value, expected_row_count_for_is",
+    )
+    def test_is_or_is_not_distinct_from(
+        self, col_a_value, col_b_value, expected_row_count_for_is, connection
+    ):
+        tbl = self.tables.is_distinct_test
+
+        connection.execute(
+            tbl.insert(),
+            [{"id": 1, "col_a": col_a_value, "col_b": col_b_value}],
+        )
+
+        result = connection.execute(
+            tbl.select().where(tbl.c.col_a.is_distinct_from(tbl.c.col_b))
+        ).fetchall()
+        eq_(
+            len(result),
+            expected_row_count_for_is,
+        )
+
+        expected_row_count_for_is_not = (
+            1 if expected_row_count_for_is == 0 else 0
+        )
+        result = connection.execute(
+            tbl.select().where(tbl.c.col_a.is_not_distinct_from(tbl.c.col_b))
+        ).fetchall()
+        eq_(
+            len(result),
+            expected_row_count_for_is_not,
+        )
+
+
+class WindowFunctionTest(fixtures.TablesTest):
+    __requires__ = ("window_functions",)
+
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "some_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("col1", Integer),
+            Column("col2", Integer),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.some_table.insert(),
+            [{"id": i, "col1": i, "col2": i * 5} for i in range(1, 50)],
+        )
+
+    def test_window(self, connection):
+        some_table = self.tables.some_table
+        rows = connection.execute(
+            select(
+                func.max(some_table.c.col2).over(
+                    order_by=[some_table.c.col1.desc()]
+                )
+            ).where(some_table.c.col1 < 20)
+        ).all()
+
+        eq_(rows, [(95,) for i in range(19)])
+
+    def test_window_rows_between(self, connection):
+        some_table = self.tables.some_table
+
+        # note the rows are part of the cache key right now, not handled
+        # as binds.  this is issue #11515
+        rows = connection.execute(
+            select(
+                func.max(some_table.c.col2).over(
+                    order_by=[some_table.c.col1],
+                    rows=(-5, 0),
+                )
+            )
+        ).all()
+
+        eq_(rows, [(i,) for i in range(5, 250, 5)])
+
+
+class BitwiseTest(fixtures.TablesTest):
+    __backend__ = True
+    run_inserts = run_deletes = "once"
+
+    inserted_data = [{"a": i, "b": i + 1} for i in range(10)]
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table("bitwise", metadata, Column("a", Integer), Column("b", Integer))
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(cls.tables.bitwise.insert(), cls.inserted_data)
+
+    @testing.combinations(
+        (
+            lambda a: a.bitwise_xor(5),
+            [i for i in range(10) if i != 5],
+            testing.requires.supports_bitwise_xor,
+        ),
+        (
+            lambda a: a.bitwise_or(1),
+            list(range(10)),
+            testing.requires.supports_bitwise_or,
+        ),
+        (
+            lambda a: a.bitwise_and(4),
+            list(range(4, 8)),
+            testing.requires.supports_bitwise_and,
+        ),
+        (
+            lambda a: (a - 2).bitwise_not(),
+            [0],
+            testing.requires.supports_bitwise_not,
+        ),
+        (
+            lambda a: a.bitwise_lshift(1),
+            list(range(1, 10)),
+            testing.requires.supports_bitwise_shift,
+        ),
+        (
+            lambda a: a.bitwise_rshift(2),
+            list(range(4, 10)),
+            testing.requires.supports_bitwise_shift,
+        ),
+        argnames="case, expected",
+    )
+    def test_bitwise(self, case, expected, connection):
+        tbl = self.tables.bitwise
+
+        a = tbl.c.a
+
+        op = testing.resolve_lambda(case, a=a)
+
+        stmt = select(tbl).where(op > 0).order_by(a)
+
+        res = connection.execute(stmt).mappings().all()
+        eq_(res, [self.inserted_data[i] for i in expected])
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_sequence.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_sequence.py
new file mode 100644
index 00000000..f0e65753
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_sequence.py
@@ -0,0 +1,317 @@
+# testing/suite/test_sequence.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .. import config
+from .. import fixtures
+from ..assertions import eq_
+from ..assertions import is_true
+from ..config import requirements
+from ..provision import normalize_sequence
+from ..schema import Column
+from ..schema import Table
+from ... import inspect
+from ... import Integer
+from ... import MetaData
+from ... import Sequence
+from ... import String
+from ... import testing
+
+
+class SequenceTest(fixtures.TablesTest):
+    __requires__ = ("sequences",)
+    __backend__ = True
+
+    run_create_tables = "each"
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "seq_pk",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                normalize_sequence(config, Sequence("tab_id_seq")),
+                primary_key=True,
+            ),
+            Column("data", String(50)),
+        )
+
+        Table(
+            "seq_opt_pk",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                normalize_sequence(
+                    config,
+                    Sequence("tab_id_seq", data_type=Integer, optional=True),
+                ),
+                primary_key=True,
+            ),
+            Column("data", String(50)),
+        )
+
+        Table(
+            "seq_no_returning",
+            metadata,
+            Column(
+                "id",
+                Integer,
+                normalize_sequence(config, Sequence("noret_id_seq")),
+                primary_key=True,
+            ),
+            Column("data", String(50)),
+            implicit_returning=False,
+        )
+
+        if testing.requires.schemas.enabled:
+            Table(
+                "seq_no_returning_sch",
+                metadata,
+                Column(
+                    "id",
+                    Integer,
+                    normalize_sequence(
+                        config,
+                        Sequence(
+                            "noret_sch_id_seq", schema=config.test_schema
+                        ),
+                    ),
+                    primary_key=True,
+                ),
+                Column("data", String(50)),
+                implicit_returning=False,
+                schema=config.test_schema,
+            )
+
+    def test_insert_roundtrip(self, connection):
+        connection.execute(self.tables.seq_pk.insert(), dict(data="some data"))
+        self._assert_round_trip(self.tables.seq_pk, connection)
+
+    def test_insert_lastrowid(self, connection):
+        r = connection.execute(
+            self.tables.seq_pk.insert(), dict(data="some data")
+        )
+        eq_(
+            r.inserted_primary_key, (testing.db.dialect.default_sequence_base,)
+        )
+
+    def test_nextval_direct(self, connection):
+        r = connection.scalar(self.tables.seq_pk.c.id.default)
+        eq_(r, testing.db.dialect.default_sequence_base)
+
+    @requirements.sequences_optional
+    def test_optional_seq(self, connection):
+        r = connection.execute(
+            self.tables.seq_opt_pk.insert(), dict(data="some data")
+        )
+        eq_(r.inserted_primary_key, (1,))
+
+    def _assert_round_trip(self, table, conn):
+        row = conn.execute(table.select()).first()
+        eq_(row, (testing.db.dialect.default_sequence_base, "some data"))
+
+    def test_insert_roundtrip_no_implicit_returning(self, connection):
+        connection.execute(
+            self.tables.seq_no_returning.insert(), dict(data="some data")
+        )
+        self._assert_round_trip(self.tables.seq_no_returning, connection)
+
+    @testing.combinations((True,), (False,), argnames="implicit_returning")
+    @testing.requires.schemas
+    def test_insert_roundtrip_translate(self, connection, implicit_returning):
+        seq_no_returning = Table(
+            "seq_no_returning_sch",
+            MetaData(),
+            Column(
+                "id",
+                Integer,
+                normalize_sequence(
+                    config, Sequence("noret_sch_id_seq", schema="alt_schema")
+                ),
+                primary_key=True,
+            ),
+            Column("data", String(50)),
+            implicit_returning=implicit_returning,
+            schema="alt_schema",
+        )
+
+        connection = connection.execution_options(
+            schema_translate_map={"alt_schema": config.test_schema}
+        )
+        connection.execute(seq_no_returning.insert(), dict(data="some data"))
+        self._assert_round_trip(seq_no_returning, connection)
+
+    @testing.requires.schemas
+    def test_nextval_direct_schema_translate(self, connection):
+        seq = normalize_sequence(
+            config, Sequence("noret_sch_id_seq", schema="alt_schema")
+        )
+        connection = connection.execution_options(
+            schema_translate_map={"alt_schema": config.test_schema}
+        )
+
+        r = connection.scalar(seq)
+        eq_(r, testing.db.dialect.default_sequence_base)
+
+
+class SequenceCompilerTest(testing.AssertsCompiledSQL, fixtures.TestBase):
+    __requires__ = ("sequences",)
+    __backend__ = True
+
+    def test_literal_binds_inline_compile(self, connection):
+        table = Table(
+            "x",
+            MetaData(),
+            Column(
+                "y", Integer, normalize_sequence(config, Sequence("y_seq"))
+            ),
+            Column("q", Integer),
+        )
+
+        stmt = table.insert().values(q=5)
+
+        seq_nextval = connection.dialect.statement_compiler(
+            statement=None, dialect=connection.dialect
+        ).visit_sequence(normalize_sequence(config, Sequence("y_seq")))
+        self.assert_compile(
+            stmt,
+            "INSERT INTO x (y, q) VALUES (%s, 5)" % (seq_nextval,),
+            literal_binds=True,
+            dialect=connection.dialect,
+        )
+
+
+class HasSequenceTest(fixtures.TablesTest):
+    run_deletes = None
+
+    __requires__ = ("sequences",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        normalize_sequence(config, Sequence("user_id_seq", metadata=metadata))
+        normalize_sequence(
+            config,
+            Sequence(
+                "other_seq",
+                metadata=metadata,
+                nomaxvalue=True,
+                nominvalue=True,
+            ),
+        )
+        if testing.requires.schemas.enabled:
+            normalize_sequence(
+                config,
+                Sequence(
+                    "user_id_seq", schema=config.test_schema, metadata=metadata
+                ),
+            )
+            normalize_sequence(
+                config,
+                Sequence(
+                    "schema_seq", schema=config.test_schema, metadata=metadata
+                ),
+            )
+        Table(
+            "user_id_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+        )
+
+    def test_has_sequence(self, connection):
+        eq_(inspect(connection).has_sequence("user_id_seq"), True)
+
+    def test_has_sequence_cache(self, connection, metadata):
+        insp = inspect(connection)
+        eq_(insp.has_sequence("user_id_seq"), True)
+        ss = normalize_sequence(config, Sequence("new_seq", metadata=metadata))
+        eq_(insp.has_sequence("new_seq"), False)
+        ss.create(connection)
+        try:
+            eq_(insp.has_sequence("new_seq"), False)
+            insp.clear_cache()
+            eq_(insp.has_sequence("new_seq"), True)
+        finally:
+            ss.drop(connection)
+
+    def test_has_sequence_other_object(self, connection):
+        eq_(inspect(connection).has_sequence("user_id_table"), False)
+
+    @testing.requires.schemas
+    def test_has_sequence_schema(self, connection):
+        eq_(
+            inspect(connection).has_sequence(
+                "user_id_seq", schema=config.test_schema
+            ),
+            True,
+        )
+
+    def test_has_sequence_neg(self, connection):
+        eq_(inspect(connection).has_sequence("some_sequence"), False)
+
+    @testing.requires.schemas
+    def test_has_sequence_schemas_neg(self, connection):
+        eq_(
+            inspect(connection).has_sequence(
+                "some_sequence", schema=config.test_schema
+            ),
+            False,
+        )
+
+    @testing.requires.schemas
+    def test_has_sequence_default_not_in_remote(self, connection):
+        eq_(
+            inspect(connection).has_sequence(
+                "other_sequence", schema=config.test_schema
+            ),
+            False,
+        )
+
+    @testing.requires.schemas
+    def test_has_sequence_remote_not_in_default(self, connection):
+        eq_(inspect(connection).has_sequence("schema_seq"), False)
+
+    def test_get_sequence_names(self, connection):
+        exp = {"other_seq", "user_id_seq"}
+
+        res = set(inspect(connection).get_sequence_names())
+        is_true(res.intersection(exp) == exp)
+        is_true("schema_seq" not in res)
+
+    @testing.requires.schemas
+    def test_get_sequence_names_no_sequence_schema(self, connection):
+        eq_(
+            inspect(connection).get_sequence_names(
+                schema=config.test_schema_2
+            ),
+            [],
+        )
+
+    @testing.requires.schemas
+    def test_get_sequence_names_sequences_schema(self, connection):
+        eq_(
+            sorted(
+                inspect(connection).get_sequence_names(
+                    schema=config.test_schema
+                )
+            ),
+            ["schema_seq", "user_id_seq"],
+        )
+
+
+class HasSequenceTestEmpty(fixtures.TestBase):
+    __requires__ = ("sequences",)
+    __backend__ = True
+
+    def test_get_sequence_names_no_sequence(self, connection):
+        eq_(
+            inspect(connection).get_sequence_names(),
+            [],
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_types.py
new file mode 100644
index 00000000..de3cd53e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_types.py
@@ -0,0 +1,2141 @@
+# testing/suite/test_types.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+import datetime
+import decimal
+import json
+import re
+import uuid
+
+from .. import config
+from .. import engines
+from .. import fixtures
+from .. import mock
+from ..assertions import eq_
+from ..assertions import is_
+from ..assertions import ne_
+from ..config import requirements
+from ..schema import Column
+from ..schema import Table
+from ... import and_
+from ... import ARRAY
+from ... import BigInteger
+from ... import bindparam
+from ... import Boolean
+from ... import case
+from ... import cast
+from ... import Date
+from ... import DateTime
+from ... import Enum
+from ... import Float
+from ... import Integer
+from ... import Interval
+from ... import JSON
+from ... import literal
+from ... import literal_column
+from ... import MetaData
+from ... import null
+from ... import Numeric
+from ... import select
+from ... import String
+from ... import testing
+from ... import Text
+from ... import Time
+from ... import TIMESTAMP
+from ... import type_coerce
+from ... import TypeDecorator
+from ... import Unicode
+from ... import UnicodeText
+from ... import UUID
+from ... import Uuid
+from ...orm import declarative_base
+from ...orm import Session
+from ...sql import sqltypes
+from ...sql.sqltypes import LargeBinary
+from ...sql.sqltypes import PickleType
+
+
+class _LiteralRoundTripFixture:
+    supports_whereclause = True
+
+    @testing.fixture
+    def literal_round_trip(self, metadata, connection):
+        """test literal rendering"""
+
+        # for literal, we test the literal render in an INSERT
+        # into a typed column.  we can then SELECT it back as its
+        # official type; ideally we'd be able to use CAST here
+        # but MySQL in particular can't CAST fully
+
+        def run(
+            type_,
+            input_,
+            output,
+            filter_=None,
+            compare=None,
+            support_whereclause=True,
+        ):
+            t = Table("t", metadata, Column("x", type_))
+            t.create(connection)
+
+            for value in input_:
+                ins = t.insert().values(
+                    x=literal(value, type_, literal_execute=True)
+                )
+                connection.execute(ins)
+
+            ins = t.insert().values(
+                x=literal(None, type_, literal_execute=True)
+            )
+            connection.execute(ins)
+
+            if support_whereclause and self.supports_whereclause:
+                if compare:
+                    stmt = t.select().where(
+                        t.c.x
+                        == literal(
+                            compare,
+                            type_,
+                            literal_execute=True,
+                        ),
+                        t.c.x
+                        == literal(
+                            input_[0],
+                            type_,
+                            literal_execute=True,
+                        ),
+                    )
+                else:
+                    stmt = t.select().where(
+                        t.c.x
+                        == literal(
+                            compare if compare is not None else input_[0],
+                            type_,
+                            literal_execute=True,
+                        )
+                    )
+            else:
+                stmt = t.select().where(t.c.x.is_not(None))
+
+            rows = connection.execute(stmt).all()
+            assert rows, "No rows returned"
+            for row in rows:
+                value = row[0]
+                if filter_ is not None:
+                    value = filter_(value)
+                assert value in output
+
+            stmt = t.select().where(t.c.x.is_(None))
+            rows = connection.execute(stmt).all()
+            eq_(rows, [(None,)])
+
+        return run
+
+
+class _UnicodeFixture(_LiteralRoundTripFixture, fixtures.TestBase):
+    __requires__ = ("unicode_data",)
+
+    data = (
+        "Alors vous imaginez ma 🐍 surprise, au lever du jour, "
+        "quand une drôle de petite 🐍 voix m’a réveillé. Elle "
+        "disait: « S’il vous plaît… dessine-moi 🐍 un mouton! »"
+    )
+
+    @property
+    def supports_whereclause(self):
+        return config.requirements.expressions_against_unbounded_text.enabled
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "unicode_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("unicode_data", cls.datatype),
+        )
+
+    def test_round_trip(self, connection):
+        unicode_table = self.tables.unicode_table
+
+        connection.execute(
+            unicode_table.insert(), {"id": 1, "unicode_data": self.data}
+        )
+
+        row = connection.execute(select(unicode_table.c.unicode_data)).first()
+
+        eq_(row, (self.data,))
+        assert isinstance(row[0], str)
+
+    def test_round_trip_executemany(self, connection):
+        unicode_table = self.tables.unicode_table
+
+        connection.execute(
+            unicode_table.insert(),
+            [{"id": i, "unicode_data": self.data} for i in range(1, 4)],
+        )
+
+        rows = connection.execute(
+            select(unicode_table.c.unicode_data)
+        ).fetchall()
+        eq_(rows, [(self.data,) for i in range(1, 4)])
+        for row in rows:
+            assert isinstance(row[0], str)
+
+    def _test_null_strings(self, connection):
+        unicode_table = self.tables.unicode_table
+
+        connection.execute(
+            unicode_table.insert(), {"id": 1, "unicode_data": None}
+        )
+        row = connection.execute(select(unicode_table.c.unicode_data)).first()
+        eq_(row, (None,))
+
+    def _test_empty_strings(self, connection):
+        unicode_table = self.tables.unicode_table
+
+        connection.execute(
+            unicode_table.insert(), {"id": 1, "unicode_data": ""}
+        )
+        row = connection.execute(select(unicode_table.c.unicode_data)).first()
+        eq_(row, ("",))
+
+    def test_literal(self, literal_round_trip):
+        literal_round_trip(self.datatype, [self.data], [self.data])
+
+    def test_literal_non_ascii(self, literal_round_trip):
+        literal_round_trip(self.datatype, ["réve🐍 illé"], ["réve🐍 illé"])
+
+
+class UnicodeVarcharTest(_UnicodeFixture, fixtures.TablesTest):
+    __requires__ = ("unicode_data",)
+    __backend__ = True
+
+    datatype = Unicode(255)
+
+    @requirements.empty_strings_varchar
+    def test_empty_strings_varchar(self, connection):
+        self._test_empty_strings(connection)
+
+    def test_null_strings_varchar(self, connection):
+        self._test_null_strings(connection)
+
+
+class UnicodeTextTest(_UnicodeFixture, fixtures.TablesTest):
+    __requires__ = "unicode_data", "text_type"
+    __backend__ = True
+
+    datatype = UnicodeText()
+
+    @requirements.empty_strings_text
+    def test_empty_strings_text(self, connection):
+        self._test_empty_strings(connection)
+
+    def test_null_strings_text(self, connection):
+        self._test_null_strings(connection)
+
+
+class ArrayTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    """Add ARRAY test suite, #8138.
+
+    This only works on PostgreSQL right now.
+
+    """
+
+    __requires__ = ("array_type",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "array_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("single_dim", ARRAY(Integer)),
+            Column("multi_dim", ARRAY(String, dimensions=2)),
+        )
+
+    def test_array_roundtrip(self, connection):
+        array_table = self.tables.array_table
+
+        connection.execute(
+            array_table.insert(),
+            {
+                "id": 1,
+                "single_dim": [1, 2, 3],
+                "multi_dim": [["one", "two"], ["thr'ee", "réve🐍 illé"]],
+            },
+        )
+        row = connection.execute(
+            select(array_table.c.single_dim, array_table.c.multi_dim)
+        ).first()
+        eq_(row, ([1, 2, 3], [["one", "two"], ["thr'ee", "réve🐍 illé"]]))
+
+    def test_literal_simple(self, literal_round_trip):
+        literal_round_trip(
+            ARRAY(Integer),
+            ([1, 2, 3],),
+            ([1, 2, 3],),
+            support_whereclause=False,
+        )
+
+    def test_literal_complex(self, literal_round_trip):
+        literal_round_trip(
+            ARRAY(String, dimensions=2),
+            ([["one", "two"], ["thr'ee", "réve🐍 illé"]],),
+            ([["one", "two"], ["thr'ee", "réve🐍 illé"]],),
+            support_whereclause=False,
+        )
+
+
+class BinaryTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "binary_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("binary_data", LargeBinary),
+            Column("pickle_data", PickleType),
+        )
+
+    @testing.combinations(b"this is binary", b"7\xe7\x9f", argnames="data")
+    def test_binary_roundtrip(self, connection, data):
+        binary_table = self.tables.binary_table
+
+        connection.execute(
+            binary_table.insert(), {"id": 1, "binary_data": data}
+        )
+        row = connection.execute(select(binary_table.c.binary_data)).first()
+        eq_(row, (data,))
+
+    def test_pickle_roundtrip(self, connection):
+        binary_table = self.tables.binary_table
+
+        connection.execute(
+            binary_table.insert(),
+            {"id": 1, "pickle_data": {"foo": [1, 2, 3], "bar": "bat"}},
+        )
+        row = connection.execute(select(binary_table.c.pickle_data)).first()
+        eq_(row, ({"foo": [1, 2, 3], "bar": "bat"},))
+
+
+class TextTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __requires__ = ("text_type",)
+    __backend__ = True
+
+    @property
+    def supports_whereclause(self):
+        return config.requirements.expressions_against_unbounded_text.enabled
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "text_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("text_data", Text),
+        )
+
+    def test_text_roundtrip(self, connection):
+        text_table = self.tables.text_table
+
+        connection.execute(
+            text_table.insert(), {"id": 1, "text_data": "some text"}
+        )
+        row = connection.execute(select(text_table.c.text_data)).first()
+        eq_(row, ("some text",))
+
+    @testing.requires.empty_strings_text
+    def test_text_empty_strings(self, connection):
+        text_table = self.tables.text_table
+
+        connection.execute(text_table.insert(), {"id": 1, "text_data": ""})
+        row = connection.execute(select(text_table.c.text_data)).first()
+        eq_(row, ("",))
+
+    def test_text_null_strings(self, connection):
+        text_table = self.tables.text_table
+
+        connection.execute(text_table.insert(), {"id": 1, "text_data": None})
+        row = connection.execute(select(text_table.c.text_data)).first()
+        eq_(row, (None,))
+
+    def test_literal(self, literal_round_trip):
+        literal_round_trip(Text, ["some text"], ["some text"])
+
+    @requirements.unicode_data_no_special_types
+    def test_literal_non_ascii(self, literal_round_trip):
+        literal_round_trip(Text, ["réve🐍 illé"], ["réve🐍 illé"])
+
+    def test_literal_quoting(self, literal_round_trip):
+        data = """some 'text' hey "hi there" that's text"""
+        literal_round_trip(Text, [data], [data])
+
+    def test_literal_backslashes(self, literal_round_trip):
+        data = r"backslash one \ backslash two \\ end"
+        literal_round_trip(Text, [data], [data])
+
+    def test_literal_percentsigns(self, literal_round_trip):
+        data = r"percent % signs %% percent"
+        literal_round_trip(Text, [data], [data])
+
+
+class StringTest(_LiteralRoundTripFixture, fixtures.TestBase):
+    __backend__ = True
+
+    @requirements.unbounded_varchar
+    def test_nolength_string(self):
+        metadata = MetaData()
+        foo = Table("foo", metadata, Column("one", String))
+
+        foo.create(config.db)
+        foo.drop(config.db)
+
+    def test_literal(self, literal_round_trip):
+        # note that in Python 3, this invokes the Unicode
+        # datatype for the literal part because all strings are unicode
+        literal_round_trip(String(40), ["some text"], ["some text"])
+
+    @requirements.unicode_data_no_special_types
+    def test_literal_non_ascii(self, literal_round_trip):
+        literal_round_trip(String(40), ["réve🐍 illé"], ["réve🐍 illé"])
+
+    @testing.combinations(
+        ("%B%", ["AB", "BC"]),
+        ("A%C", ["AC"]),
+        ("A%C%Z", []),
+        argnames="expr, expected",
+    )
+    def test_dont_truncate_rightside(
+        self, metadata, connection, expr, expected
+    ):
+        t = Table("t", metadata, Column("x", String(2)))
+        t.create(connection)
+
+        connection.execute(t.insert(), [{"x": "AB"}, {"x": "BC"}, {"x": "AC"}])
+
+        eq_(
+            connection.scalars(select(t.c.x).where(t.c.x.like(expr))).all(),
+            expected,
+        )
+
+    def test_literal_quoting(self, literal_round_trip):
+        data = """some 'text' hey "hi there" that's text"""
+        literal_round_trip(String(40), [data], [data])
+
+    def test_literal_backslashes(self, literal_round_trip):
+        data = r"backslash one \ backslash two \\ end"
+        literal_round_trip(String(40), [data], [data])
+
+    def test_concatenate_binary(self, connection):
+        """dialects with special string concatenation operators should
+        implement visit_concat_op_binary() and visit_concat_op_clauselist()
+        in their compiler.
+
+        .. versionchanged:: 2.0  visit_concat_op_clauselist() is also needed
+           for dialects to override the string concatenation operator.
+
+        """
+        eq_(connection.scalar(select(literal("a") + "b")), "ab")
+
+    def test_concatenate_clauselist(self, connection):
+        """dialects with special string concatenation operators should
+        implement visit_concat_op_binary() and visit_concat_op_clauselist()
+        in their compiler.
+
+        .. versionchanged:: 2.0  visit_concat_op_clauselist() is also needed
+           for dialects to override the string concatenation operator.
+
+        """
+        eq_(
+            connection.scalar(select(literal("a") + "b" + "c" + "d" + "e")),
+            "abcde",
+        )
+
+
+class IntervalTest(_LiteralRoundTripFixture, fixtures.TestBase):
+    __requires__ = ("datetime_interval",)
+    __backend__ = True
+
+    datatype = Interval
+    data = datetime.timedelta(days=1, seconds=4)
+
+    def test_literal(self, literal_round_trip):
+        literal_round_trip(self.datatype, [self.data], [self.data])
+
+    def test_select_direct_literal_interval(self, connection):
+        row = connection.execute(select(literal(self.data))).first()
+        eq_(row, (self.data,))
+
+    def test_arithmetic_operation_literal_interval(self, connection):
+        now = datetime.datetime.now().replace(microsecond=0)
+        # Able to subtract
+        row = connection.execute(
+            select(literal(now) - literal(self.data))
+        ).scalar()
+        eq_(row, now - self.data)
+
+        # Able to Add
+        row = connection.execute(
+            select(literal(now) + literal(self.data))
+        ).scalar()
+        eq_(row, now + self.data)
+
+    @testing.fixture
+    def arithmetic_table_fixture(cls, metadata, connection):
+        class Decorated(TypeDecorator):
+            impl = cls.datatype
+            cache_ok = True
+
+        it = Table(
+            "interval_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("interval_data", cls.datatype),
+            Column("date_data", DateTime),
+            Column("decorated_interval_data", Decorated),
+        )
+        it.create(connection)
+        return it
+
+    def test_arithmetic_operation_table_interval_and_literal_interval(
+        self, connection, arithmetic_table_fixture
+    ):
+        interval_table = arithmetic_table_fixture
+        data = datetime.timedelta(days=2, seconds=5)
+        connection.execute(
+            interval_table.insert(), {"id": 1, "interval_data": data}
+        )
+        # Subtraction Operation
+        value = connection.execute(
+            select(interval_table.c.interval_data - literal(self.data))
+        ).scalar()
+        eq_(value, data - self.data)
+
+        # Addition Operation
+        value = connection.execute(
+            select(interval_table.c.interval_data + literal(self.data))
+        ).scalar()
+        eq_(value, data + self.data)
+
+    def test_arithmetic_operation_table_date_and_literal_interval(
+        self, connection, arithmetic_table_fixture
+    ):
+        interval_table = arithmetic_table_fixture
+        now = datetime.datetime.now().replace(microsecond=0)
+        connection.execute(
+            interval_table.insert(), {"id": 1, "date_data": now}
+        )
+        # Subtraction Operation
+        value = connection.execute(
+            select(interval_table.c.date_data - literal(self.data))
+        ).scalar()
+        eq_(value, (now - self.data))
+
+        # Addition Operation
+        value = connection.execute(
+            select(interval_table.c.date_data + literal(self.data))
+        ).scalar()
+        eq_(value, (now + self.data))
+
+
+class PrecisionIntervalTest(IntervalTest):
+    __requires__ = ("datetime_interval",)
+    __backend__ = True
+
+    datatype = Interval(day_precision=9, second_precision=9)
+    data = datetime.timedelta(days=103, seconds=4)
+
+
+class _DateFixture(_LiteralRoundTripFixture, fixtures.TestBase):
+    compare = None
+
+    @classmethod
+    def define_tables(cls, metadata):
+        class Decorated(TypeDecorator):
+            impl = cls.datatype
+            cache_ok = True
+
+        Table(
+            "date_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("date_data", cls.datatype),
+            Column("decorated_date_data", Decorated),
+        )
+
+    def test_round_trip(self, connection):
+        date_table = self.tables.date_table
+
+        connection.execute(
+            date_table.insert(), {"id": 1, "date_data": self.data}
+        )
+
+        row = connection.execute(select(date_table.c.date_data)).first()
+
+        compare = self.compare or self.data
+        eq_(row, (compare,))
+        assert isinstance(row[0], type(compare))
+
+    def test_round_trip_decorated(self, connection):
+        date_table = self.tables.date_table
+
+        connection.execute(
+            date_table.insert(), {"id": 1, "decorated_date_data": self.data}
+        )
+
+        row = connection.execute(
+            select(date_table.c.decorated_date_data)
+        ).first()
+
+        compare = self.compare or self.data
+        eq_(row, (compare,))
+        assert isinstance(row[0], type(compare))
+
+    def test_null(self, connection):
+        date_table = self.tables.date_table
+
+        connection.execute(date_table.insert(), {"id": 1, "date_data": None})
+
+        row = connection.execute(select(date_table.c.date_data)).first()
+        eq_(row, (None,))
+
+    @testing.requires.datetime_literals
+    def test_literal(self, literal_round_trip):
+        compare = self.compare or self.data
+
+        literal_round_trip(
+            self.datatype, [self.data], [compare], compare=compare
+        )
+
+    @testing.requires.standalone_null_binds_whereclause
+    def test_null_bound_comparison(self):
+        # this test is based on an Oracle issue observed in #4886.
+        # passing NULL for an expression that needs to be interpreted as
+        # a certain type, does the DBAPI have the info it needs to do this.
+        date_table = self.tables.date_table
+        with config.db.begin() as conn:
+            result = conn.execute(
+                date_table.insert(), {"id": 1, "date_data": self.data}
+            )
+            id_ = result.inserted_primary_key[0]
+            stmt = select(date_table.c.id).where(
+                case(
+                    (
+                        bindparam("foo", type_=self.datatype) != None,
+                        bindparam("foo", type_=self.datatype),
+                    ),
+                    else_=date_table.c.date_data,
+                )
+                == date_table.c.date_data
+            )
+
+            row = conn.execute(stmt, {"foo": None}).first()
+            eq_(row[0], id_)
+
+
+class DateTimeTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("datetime",)
+    __backend__ = True
+    datatype = DateTime
+    data = datetime.datetime(2012, 10, 15, 12, 57, 18)
+
+    @testing.requires.datetime_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateTimeTZTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("datetime_timezone",)
+    __backend__ = True
+    datatype = DateTime(timezone=True)
+    data = datetime.datetime(
+        2012, 10, 15, 12, 57, 18, tzinfo=datetime.timezone.utc
+    )
+
+    @testing.requires.datetime_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateTimeMicrosecondsTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("datetime_microseconds",)
+    __backend__ = True
+    datatype = DateTime
+    data = datetime.datetime(2012, 10, 15, 12, 57, 18, 39642)
+
+
+class TimestampMicrosecondsTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("timestamp_microseconds",)
+    __backend__ = True
+    datatype = TIMESTAMP
+    data = datetime.datetime(2012, 10, 15, 12, 57, 18, 396)
+
+    @testing.requires.timestamp_microseconds_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class TimeTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("time",)
+    __backend__ = True
+    datatype = Time
+    data = datetime.time(12, 57, 18)
+
+    @testing.requires.time_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class TimeTZTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("time_timezone",)
+    __backend__ = True
+    datatype = Time(timezone=True)
+    data = datetime.time(12, 57, 18, tzinfo=datetime.timezone.utc)
+
+    @testing.requires.time_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class TimeMicrosecondsTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("time_microseconds",)
+    __backend__ = True
+    datatype = Time
+    data = datetime.time(12, 57, 18, 396)
+
+    @testing.requires.time_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("date",)
+    __backend__ = True
+    datatype = Date
+    data = datetime.date(2012, 10, 15)
+
+    @testing.requires.date_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateTimeCoercedToDateTimeTest(_DateFixture, fixtures.TablesTest):
+    """this particular suite is testing that datetime parameters get
+    coerced to dates, which tends to be something DBAPIs do.
+
+    """
+
+    __requires__ = "date", "date_coerces_from_datetime"
+    __backend__ = True
+    datatype = Date
+    data = datetime.datetime(2012, 10, 15, 12, 57, 18)
+    compare = datetime.date(2012, 10, 15)
+
+    @testing.requires.datetime_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateTimeHistoricTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("datetime_historic",)
+    __backend__ = True
+    datatype = DateTime
+    data = datetime.datetime(1850, 11, 10, 11, 52, 35)
+
+    @testing.requires.date_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class DateHistoricTest(_DateFixture, fixtures.TablesTest):
+    __requires__ = ("date_historic",)
+    __backend__ = True
+    datatype = Date
+    data = datetime.date(1727, 4, 1)
+
+    @testing.requires.date_implicit_bound
+    def test_select_direct(self, connection):
+        result = connection.scalar(select(literal(self.data)))
+        eq_(result, self.data)
+
+
+class IntegerTest(_LiteralRoundTripFixture, fixtures.TestBase):
+    __backend__ = True
+
+    def test_literal(self, literal_round_trip):
+        literal_round_trip(Integer, [5], [5])
+
+    def _huge_ints():
+        return testing.combinations(
+            2147483649,  # 32 bits
+            2147483648,  # 32 bits
+            2147483647,  # 31 bits
+            2147483646,  # 31 bits
+            -2147483649,  # 32 bits
+            -2147483648,  # 32 interestingly, asyncpg accepts this one as int32
+            -2147483647,  # 31
+            -2147483646,  # 31
+            0,
+            1376537018368127,
+            -1376537018368127,
+            argnames="intvalue",
+        )
+
+    @_huge_ints()
+    def test_huge_int_auto_accommodation(self, connection, intvalue):
+        """test #7909"""
+
+        eq_(
+            connection.scalar(
+                select(intvalue).where(literal(intvalue) == intvalue)
+            ),
+            intvalue,
+        )
+
+    @_huge_ints()
+    def test_huge_int(self, integer_round_trip, intvalue):
+        integer_round_trip(BigInteger, intvalue)
+
+    @testing.fixture
+    def integer_round_trip(self, metadata, connection):
+        def run(datatype, data):
+            int_table = Table(
+                "integer_table",
+                metadata,
+                Column(
+                    "id",
+                    Integer,
+                    primary_key=True,
+                    test_needs_autoincrement=True,
+                ),
+                Column("integer_data", datatype),
+            )
+
+            metadata.create_all(config.db)
+
+            connection.execute(
+                int_table.insert(), {"id": 1, "integer_data": data}
+            )
+
+            row = connection.execute(select(int_table.c.integer_data)).first()
+
+            eq_(row, (data,))
+
+            assert isinstance(row[0], int)
+
+        return run
+
+
+class CastTypeDecoratorTest(_LiteralRoundTripFixture, fixtures.TestBase):
+    __backend__ = True
+
+    @testing.fixture
+    def string_as_int(self):
+        class StringAsInt(TypeDecorator):
+            impl = String(50)
+            cache_ok = True
+
+            def column_expression(self, col):
+                return cast(col, Integer)
+
+            def bind_expression(self, col):
+                return cast(type_coerce(col, Integer), String(50))
+
+        return StringAsInt()
+
+    def test_special_type(self, metadata, connection, string_as_int):
+        type_ = string_as_int
+
+        t = Table("t", metadata, Column("x", type_))
+        t.create(connection)
+
+        connection.execute(t.insert(), [{"x": x} for x in [1, 2, 3]])
+
+        result = {row[0] for row in connection.execute(t.select())}
+        eq_(result, {1, 2, 3})
+
+        result = {
+            row[0] for row in connection.execute(t.select().where(t.c.x == 2))
+        }
+        eq_(result, {2})
+
+
+class TrueDivTest(fixtures.TestBase):
+    __backend__ = True
+
+    @testing.combinations(
+        ("15", "10", 1.5),
+        ("-15", "10", -1.5),
+        argnames="left, right, expected",
+    )
+    def test_truediv_integer(self, connection, left, right, expected):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(
+                select(
+                    literal_column(left, type_=Integer())
+                    / literal_column(right, type_=Integer())
+                )
+            ),
+            expected,
+        )
+
+    @testing.combinations(
+        ("15", "10", 1), ("-15", "5", -3), argnames="left, right, expected"
+    )
+    def test_floordiv_integer(self, connection, left, right, expected):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(
+                select(
+                    literal_column(left, type_=Integer())
+                    // literal_column(right, type_=Integer())
+                )
+            ),
+            expected,
+        )
+
+    @testing.combinations(
+        ("5.52", "2.4", "2.3"), argnames="left, right, expected"
+    )
+    def test_truediv_numeric(self, connection, left, right, expected):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(
+                select(
+                    literal_column(left, type_=Numeric(10, 2))
+                    / literal_column(right, type_=Numeric(10, 2))
+                )
+            ),
+            decimal.Decimal(expected),
+        )
+
+    @testing.combinations(
+        ("5.52", "2.4", 2.3), argnames="left, right, expected"
+    )
+    def test_truediv_float(self, connection, left, right, expected):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(
+                select(
+                    literal_column(left, type_=Float())
+                    / literal_column(right, type_=Float())
+                )
+            ),
+            expected,
+        )
+
+    @testing.combinations(
+        ("5.52", "2.4", "2.0"), argnames="left, right, expected"
+    )
+    def test_floordiv_numeric(self, connection, left, right, expected):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(
+                select(
+                    literal_column(left, type_=Numeric())
+                    // literal_column(right, type_=Numeric())
+                )
+            ),
+            decimal.Decimal(expected),
+        )
+
+    def test_truediv_integer_bound(self, connection):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(select(literal(15) / literal(10))),
+            1.5,
+        )
+
+    def test_floordiv_integer_bound(self, connection):
+        """test #4926"""
+
+        eq_(
+            connection.scalar(select(literal(15) // literal(10))),
+            1,
+        )
+
+
+class NumericTest(_LiteralRoundTripFixture, fixtures.TestBase):
+    __backend__ = True
+
+    @testing.fixture
+    def do_numeric_test(self, metadata, connection):
+        def run(type_, input_, output, filter_=None, check_scale=False):
+            t = Table("t", metadata, Column("x", type_))
+            t.create(connection)
+            connection.execute(t.insert(), [{"x": x} for x in input_])
+
+            result = {row[0] for row in connection.execute(t.select())}
+            output = set(output)
+            if filter_:
+                result = {filter_(x) for x in result}
+                output = {filter_(x) for x in output}
+            eq_(result, output)
+            if check_scale:
+                eq_([str(x) for x in result], [str(x) for x in output])
+
+            connection.execute(t.delete())
+
+            # test that this is actually a number!
+            # note we have tiny scale here as we have tests with very
+            # small scale Numeric types.  PostgreSQL will raise an error
+            # if you use values outside the available scale.
+            if type_.asdecimal:
+                test_value = decimal.Decimal("2.9")
+                add_value = decimal.Decimal("37.12")
+            else:
+                test_value = 2.9
+                add_value = 37.12
+
+            connection.execute(t.insert(), {"x": test_value})
+            assert_we_are_a_number = connection.scalar(
+                select(type_coerce(t.c.x + add_value, type_))
+            )
+            eq_(
+                round(assert_we_are_a_number, 3),
+                round(test_value + add_value, 3),
+            )
+
+        return run
+
+    def test_render_literal_numeric(self, literal_round_trip):
+        literal_round_trip(
+            Numeric(precision=8, scale=4),
+            [15.7563, decimal.Decimal("15.7563")],
+            [decimal.Decimal("15.7563")],
+        )
+
+    def test_render_literal_numeric_asfloat(self, literal_round_trip):
+        literal_round_trip(
+            Numeric(precision=8, scale=4, asdecimal=False),
+            [15.7563, decimal.Decimal("15.7563")],
+            [15.7563],
+        )
+
+    def test_render_literal_float(self, literal_round_trip):
+        literal_round_trip(
+            Float(),
+            [15.7563, decimal.Decimal("15.7563")],
+            [15.7563],
+            filter_=lambda n: n is not None and round(n, 5) or None,
+            support_whereclause=False,
+        )
+
+    @testing.requires.precision_generic_float_type
+    def test_float_custom_scale(self, do_numeric_test):
+        do_numeric_test(
+            Float(None, decimal_return_scale=7, asdecimal=True),
+            [15.7563827, decimal.Decimal("15.7563827")],
+            [decimal.Decimal("15.7563827")],
+            check_scale=True,
+        )
+
+    def test_numeric_as_decimal(self, do_numeric_test):
+        do_numeric_test(
+            Numeric(precision=8, scale=4),
+            [15.7563, decimal.Decimal("15.7563")],
+            [decimal.Decimal("15.7563")],
+        )
+
+    def test_numeric_as_float(self, do_numeric_test):
+        do_numeric_test(
+            Numeric(precision=8, scale=4, asdecimal=False),
+            [15.7563, decimal.Decimal("15.7563")],
+            [15.7563],
+        )
+
+    @testing.requires.infinity_floats
+    def test_infinity_floats(self, do_numeric_test):
+        """test for #977, #7283"""
+
+        do_numeric_test(
+            Float(None),
+            [float("inf")],
+            [float("inf")],
+        )
+
+    @testing.requires.fetch_null_from_numeric
+    def test_numeric_null_as_decimal(self, do_numeric_test):
+        do_numeric_test(Numeric(precision=8, scale=4), [None], [None])
+
+    @testing.requires.fetch_null_from_numeric
+    def test_numeric_null_as_float(self, do_numeric_test):
+        do_numeric_test(
+            Numeric(precision=8, scale=4, asdecimal=False), [None], [None]
+        )
+
+    @testing.requires.floats_to_four_decimals
+    def test_float_as_decimal(self, do_numeric_test):
+        do_numeric_test(
+            Float(asdecimal=True),
+            [15.756, decimal.Decimal("15.756"), None],
+            [decimal.Decimal("15.756"), None],
+            filter_=lambda n: n is not None and round(n, 4) or None,
+        )
+
+    def test_float_as_float(self, do_numeric_test):
+        do_numeric_test(
+            Float(),
+            [15.756, decimal.Decimal("15.756")],
+            [15.756],
+            filter_=lambda n: n is not None and round(n, 5) or None,
+        )
+
+    @testing.requires.literal_float_coercion
+    def test_float_coerce_round_trip(self, connection):
+        expr = 15.7563
+
+        val = connection.scalar(select(literal(expr)))
+        eq_(val, expr)
+
+    # this does not work in MySQL, see #4036, however we choose not
+    # to render CAST unconditionally since this is kind of an edge case.
+
+    @testing.requires.implicit_decimal_binds
+    def test_decimal_coerce_round_trip(self, connection):
+        expr = decimal.Decimal("15.7563")
+
+        val = connection.scalar(select(literal(expr)))
+        eq_(val, expr)
+
+    def test_decimal_coerce_round_trip_w_cast(self, connection):
+        expr = decimal.Decimal("15.7563")
+
+        val = connection.scalar(select(cast(expr, Numeric(10, 4))))
+        eq_(val, expr)
+
+    @testing.requires.precision_numerics_general
+    def test_precision_decimal(self, do_numeric_test):
+        numbers = {
+            decimal.Decimal("54.234246451650"),
+            decimal.Decimal("0.004354"),
+            decimal.Decimal("900.0"),
+        }
+
+        do_numeric_test(Numeric(precision=18, scale=12), numbers, numbers)
+
+    @testing.requires.precision_numerics_enotation_large
+    def test_enotation_decimal(self, do_numeric_test):
+        """test exceedingly small decimals.
+
+        Decimal reports values with E notation when the exponent
+        is greater than 6.
+
+        """
+
+        numbers = {
+            decimal.Decimal("1E-2"),
+            decimal.Decimal("1E-3"),
+            decimal.Decimal("1E-4"),
+            decimal.Decimal("1E-5"),
+            decimal.Decimal("1E-6"),
+            decimal.Decimal("1E-7"),
+            decimal.Decimal("1E-8"),
+            decimal.Decimal("0.01000005940696"),
+            decimal.Decimal("0.00000005940696"),
+            decimal.Decimal("0.00000000000696"),
+            decimal.Decimal("0.70000000000696"),
+            decimal.Decimal("696E-12"),
+        }
+        do_numeric_test(Numeric(precision=18, scale=14), numbers, numbers)
+
+    @testing.requires.precision_numerics_enotation_large
+    def test_enotation_decimal_large(self, do_numeric_test):
+        """test exceedingly large decimals."""
+
+        numbers = {
+            decimal.Decimal("4E+8"),
+            decimal.Decimal("5748E+15"),
+            decimal.Decimal("1.521E+15"),
+            decimal.Decimal("00000000000000.1E+12"),
+        }
+        do_numeric_test(Numeric(precision=25, scale=2), numbers, numbers)
+
+    @testing.requires.precision_numerics_many_significant_digits
+    def test_many_significant_digits(self, do_numeric_test):
+        numbers = {
+            decimal.Decimal("31943874831932418390.01"),
+            decimal.Decimal("319438950232418390.273596"),
+            decimal.Decimal("87673.594069654243"),
+        }
+        do_numeric_test(Numeric(precision=38, scale=12), numbers, numbers)
+
+    @testing.requires.precision_numerics_retains_significant_digits
+    def test_numeric_no_decimal(self, do_numeric_test):
+        numbers = {decimal.Decimal("1.000")}
+        do_numeric_test(
+            Numeric(precision=5, scale=3), numbers, numbers, check_scale=True
+        )
+
+    @testing.combinations(sqltypes.Float, sqltypes.Double, argnames="cls_")
+    @testing.requires.float_is_numeric
+    def test_float_is_not_numeric(self, connection, cls_):
+        target_type = cls_().dialect_impl(connection.dialect)
+        numeric_type = sqltypes.Numeric().dialect_impl(connection.dialect)
+
+        ne_(target_type.__visit_name__, numeric_type.__visit_name__)
+        ne_(target_type.__class__, numeric_type.__class__)
+
+
+class BooleanTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "boolean_table",
+            metadata,
+            Column("id", Integer, primary_key=True, autoincrement=False),
+            Column("value", Boolean),
+            Column("unconstrained_value", Boolean(create_constraint=False)),
+        )
+
+    def test_render_literal_bool(self, literal_round_trip):
+        literal_round_trip(Boolean(), [True, False], [True, False])
+
+    def test_round_trip(self, connection):
+        boolean_table = self.tables.boolean_table
+
+        connection.execute(
+            boolean_table.insert(),
+            {"id": 1, "value": True, "unconstrained_value": False},
+        )
+
+        row = connection.execute(
+            select(boolean_table.c.value, boolean_table.c.unconstrained_value)
+        ).first()
+
+        eq_(row, (True, False))
+        assert isinstance(row[0], bool)
+
+    @testing.requires.nullable_booleans
+    def test_null(self, connection):
+        boolean_table = self.tables.boolean_table
+
+        connection.execute(
+            boolean_table.insert(),
+            {"id": 1, "value": None, "unconstrained_value": None},
+        )
+
+        row = connection.execute(
+            select(boolean_table.c.value, boolean_table.c.unconstrained_value)
+        ).first()
+
+        eq_(row, (None, None))
+
+    def test_whereclause(self):
+        # testing "WHERE <column>" renders a compatible expression
+        boolean_table = self.tables.boolean_table
+
+        with config.db.begin() as conn:
+            conn.execute(
+                boolean_table.insert(),
+                [
+                    {"id": 1, "value": True, "unconstrained_value": True},
+                    {"id": 2, "value": False, "unconstrained_value": False},
+                ],
+            )
+
+            eq_(
+                conn.scalar(
+                    select(boolean_table.c.id).where(boolean_table.c.value)
+                ),
+                1,
+            )
+            eq_(
+                conn.scalar(
+                    select(boolean_table.c.id).where(
+                        boolean_table.c.unconstrained_value
+                    )
+                ),
+                1,
+            )
+            eq_(
+                conn.scalar(
+                    select(boolean_table.c.id).where(~boolean_table.c.value)
+                ),
+                2,
+            )
+            eq_(
+                conn.scalar(
+                    select(boolean_table.c.id).where(
+                        ~boolean_table.c.unconstrained_value
+                    )
+                ),
+                2,
+            )
+
+
+class JSONTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __requires__ = ("json_type",)
+    __backend__ = True
+
+    datatype = JSON
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "data_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("name", String(30), nullable=False),
+            Column("data", cls.datatype, nullable=False),
+            Column("nulldata", cls.datatype(none_as_null=True)),
+        )
+
+    def test_round_trip_data1(self, connection):
+        self._test_round_trip({"key1": "value1", "key2": "value2"}, connection)
+
+    @testing.combinations(
+        ("unicode", True), ("ascii", False), argnames="unicode_", id_="ia"
+    )
+    @testing.combinations(100, 1999, 3000, 4000, 5000, 9000, argnames="length")
+    def test_round_trip_pretty_large_data(self, connection, unicode_, length):
+        if unicode_:
+            data = "réve🐍illé" * ((length // 9) + 1)
+            data = data[0 : (length // 2)]
+        else:
+            data = "abcdefg" * ((length // 7) + 1)
+            data = data[0:length]
+
+        self._test_round_trip({"key1": data, "key2": data}, connection)
+
+    def _test_round_trip(self, data_element, connection):
+        data_table = self.tables.data_table
+
+        connection.execute(
+            data_table.insert(),
+            {"id": 1, "name": "row1", "data": data_element},
+        )
+
+        row = connection.execute(select(data_table.c.data)).first()
+
+        eq_(row, (data_element,))
+
+    def _index_fixtures(include_comparison):
+        if include_comparison:
+            # basically SQL Server and MariaDB can kind of do json
+            # comparison, MySQL, PG and SQLite can't.  not worth it.
+            json_elements = []
+        else:
+            json_elements = [
+                ("json", {"foo": "bar"}),
+                ("json", ["one", "two", "three"]),
+                (None, {"foo": "bar"}),
+                (None, ["one", "two", "three"]),
+            ]
+
+        elements = [
+            ("boolean", True),
+            ("boolean", False),
+            ("boolean", None),
+            ("string", "some string"),
+            ("string", None),
+            ("string", "réve illé"),
+            (
+                "string",
+                "réve🐍 illé",
+                testing.requires.json_index_supplementary_unicode_element,
+            ),
+            ("integer", 15),
+            ("integer", 1),
+            ("integer", 0),
+            ("integer", None),
+            ("float", 28.5),
+            ("float", None),
+            ("float", 1234567.89, testing.requires.literal_float_coercion),
+            ("numeric", 1234567.89),
+            # this one "works" because the float value you see here is
+            # lost immediately to floating point stuff
+            (
+                "numeric",
+                99998969694839.983485848,
+            ),
+            ("numeric", 99939.983485848),
+            ("_decimal", decimal.Decimal("1234567.89")),
+            (
+                "_decimal",
+                decimal.Decimal("99998969694839.983485848"),
+                # fails on SQLite and MySQL (non-mariadb)
+                requirements.cast_precision_numerics_many_significant_digits,
+            ),
+            (
+                "_decimal",
+                decimal.Decimal("99939.983485848"),
+            ),
+        ] + json_elements
+
+        def decorate(fn):
+            fn = testing.combinations(id_="sa", *elements)(fn)
+
+            return fn
+
+        return decorate
+
+    def _json_value_insert(self, connection, datatype, value, data_element):
+        data_table = self.tables.data_table
+        if datatype == "_decimal":
+            # Python's builtin json serializer basically doesn't support
+            # Decimal objects without implicit float conversion period.
+            # users can otherwise use simplejson which supports
+            # precision decimals
+
+            # https://bugs.python.org/issue16535
+
+            # inserting as strings to avoid a new fixture around the
+            # dialect which would have idiosyncrasies for different
+            # backends.
+
+            class DecimalEncoder(json.JSONEncoder):
+                def default(self, o):
+                    if isinstance(o, decimal.Decimal):
+                        return str(o)
+                    return super().default(o)
+
+            json_data = json.dumps(data_element, cls=DecimalEncoder)
+
+            # take the quotes out.  yup, there is *literally* no other
+            # way to get Python's json.dumps() to put all the digits in
+            # the string
+            json_data = re.sub(r'"(%s)"' % str(value), str(value), json_data)
+
+            datatype = "numeric"
+
+            connection.execute(
+                data_table.insert().values(
+                    name="row1",
+                    # to pass the string directly to every backend, including
+                    # PostgreSQL which needs the value to be CAST as JSON
+                    # both in the SQL as well as at the prepared statement
+                    # level for asyncpg, while at the same time MySQL
+                    # doesn't even support CAST for JSON, here we are
+                    # sending the string embedded in the SQL without using
+                    # a parameter.
+                    data=bindparam(None, json_data, literal_execute=True),
+                    nulldata=bindparam(None, json_data, literal_execute=True),
+                ),
+            )
+        else:
+            connection.execute(
+                data_table.insert(),
+                {
+                    "name": "row1",
+                    "data": data_element,
+                    "nulldata": data_element,
+                },
+            )
+
+        p_s = None
+
+        if datatype:
+            if datatype == "numeric":
+                a, b = str(value).split(".")
+                s = len(b)
+                p = len(a) + s
+
+                if isinstance(value, decimal.Decimal):
+                    compare_value = value
+                else:
+                    compare_value = decimal.Decimal(str(value))
+
+                p_s = (p, s)
+            else:
+                compare_value = value
+        else:
+            compare_value = value
+
+        return datatype, compare_value, p_s
+
+    @_index_fixtures(False)
+    def test_index_typed_access(self, datatype, value):
+        data_table = self.tables.data_table
+        data_element = {"key1": value}
+
+        with config.db.begin() as conn:
+            datatype, compare_value, p_s = self._json_value_insert(
+                conn, datatype, value, data_element
+            )
+
+            expr = data_table.c.data["key1"]
+            if datatype:
+                if datatype == "numeric" and p_s:
+                    expr = expr.as_numeric(*p_s)
+                else:
+                    expr = getattr(expr, "as_%s" % datatype)()
+
+            roundtrip = conn.scalar(select(expr))
+            eq_(roundtrip, compare_value)
+            is_(type(roundtrip), type(compare_value))
+
+    @_index_fixtures(True)
+    def test_index_typed_comparison(self, datatype, value):
+        data_table = self.tables.data_table
+        data_element = {"key1": value}
+
+        with config.db.begin() as conn:
+            datatype, compare_value, p_s = self._json_value_insert(
+                conn, datatype, value, data_element
+            )
+
+            expr = data_table.c.data["key1"]
+            if datatype:
+                if datatype == "numeric" and p_s:
+                    expr = expr.as_numeric(*p_s)
+                else:
+                    expr = getattr(expr, "as_%s" % datatype)()
+
+            row = conn.execute(
+                select(expr).where(expr == compare_value)
+            ).first()
+
+            # make sure we get a row even if value is None
+            eq_(row, (compare_value,))
+
+    @_index_fixtures(True)
+    def test_path_typed_comparison(self, datatype, value):
+        data_table = self.tables.data_table
+        data_element = {"key1": {"subkey1": value}}
+        with config.db.begin() as conn:
+            datatype, compare_value, p_s = self._json_value_insert(
+                conn, datatype, value, data_element
+            )
+
+            expr = data_table.c.data[("key1", "subkey1")]
+
+            if datatype:
+                if datatype == "numeric" and p_s:
+                    expr = expr.as_numeric(*p_s)
+                else:
+                    expr = getattr(expr, "as_%s" % datatype)()
+
+            row = conn.execute(
+                select(expr).where(expr == compare_value)
+            ).first()
+
+            # make sure we get a row even if value is None
+            eq_(row, (compare_value,))
+
+    @testing.combinations(
+        (True,),
+        (False,),
+        (None,),
+        (15,),
+        (0,),
+        (-1,),
+        (-1.0,),
+        (15.052,),
+        ("a string",),
+        ("réve illé",),
+        ("réve🐍 illé",),
+    )
+    def test_single_element_round_trip(self, element):
+        data_table = self.tables.data_table
+        data_element = element
+        with config.db.begin() as conn:
+            conn.execute(
+                data_table.insert(),
+                {
+                    "name": "row1",
+                    "data": data_element,
+                    "nulldata": data_element,
+                },
+            )
+
+            row = conn.execute(
+                select(data_table.c.data, data_table.c.nulldata)
+            ).first()
+
+            eq_(row, (data_element, data_element))
+
+    def test_round_trip_custom_json(self):
+        data_table = self.tables.data_table
+        data_element = {"key1": "data1"}
+
+        js = mock.Mock(side_effect=json.dumps)
+        jd = mock.Mock(side_effect=json.loads)
+        engine = engines.testing_engine(
+            options=dict(json_serializer=js, json_deserializer=jd)
+        )
+
+        # support sqlite :memory: database...
+        data_table.create(engine, checkfirst=True)
+        with engine.begin() as conn:
+            conn.execute(
+                data_table.insert(), {"name": "row1", "data": data_element}
+            )
+            row = conn.execute(select(data_table.c.data)).first()
+
+            eq_(row, (data_element,))
+            eq_(js.mock_calls, [mock.call(data_element)])
+            if testing.requires.json_deserializer_binary.enabled:
+                eq_(
+                    jd.mock_calls,
+                    [mock.call(json.dumps(data_element).encode())],
+                )
+            else:
+                eq_(jd.mock_calls, [mock.call(json.dumps(data_element))])
+
+    @testing.combinations(
+        ("parameters",),
+        ("multiparameters",),
+        ("values",),
+        ("omit",),
+        argnames="insert_type",
+    )
+    def test_round_trip_none_as_sql_null(self, connection, insert_type):
+        col = self.tables.data_table.c["nulldata"]
+
+        conn = connection
+
+        if insert_type == "parameters":
+            stmt, params = self.tables.data_table.insert(), {
+                "name": "r1",
+                "nulldata": None,
+                "data": None,
+            }
+        elif insert_type == "multiparameters":
+            stmt, params = self.tables.data_table.insert(), [
+                {"name": "r1", "nulldata": None, "data": None}
+            ]
+        elif insert_type == "values":
+            stmt, params = (
+                self.tables.data_table.insert().values(
+                    name="r1",
+                    nulldata=None,
+                    data=None,
+                ),
+                {},
+            )
+        elif insert_type == "omit":
+            stmt, params = (
+                self.tables.data_table.insert(),
+                {"name": "r1", "data": None},
+            )
+
+        else:
+            assert False
+
+        conn.execute(stmt, params)
+
+        eq_(
+            conn.scalar(
+                select(self.tables.data_table.c.name).where(col.is_(null()))
+            ),
+            "r1",
+        )
+
+        eq_(conn.scalar(select(col)), None)
+
+    def test_round_trip_json_null_as_json_null(self, connection):
+        col = self.tables.data_table.c["data"]
+
+        conn = connection
+        conn.execute(
+            self.tables.data_table.insert(),
+            {"name": "r1", "data": JSON.NULL},
+        )
+
+        eq_(
+            conn.scalar(
+                select(self.tables.data_table.c.name).where(
+                    cast(col, String) == "null"
+                )
+            ),
+            "r1",
+        )
+
+        eq_(conn.scalar(select(col)), None)
+
+    @testing.combinations(
+        ("parameters",),
+        ("multiparameters",),
+        ("values",),
+        argnames="insert_type",
+    )
+    def test_round_trip_none_as_json_null(self, connection, insert_type):
+        col = self.tables.data_table.c["data"]
+
+        if insert_type == "parameters":
+            stmt, params = self.tables.data_table.insert(), {
+                "name": "r1",
+                "data": None,
+            }
+        elif insert_type == "multiparameters":
+            stmt, params = self.tables.data_table.insert(), [
+                {"name": "r1", "data": None}
+            ]
+        elif insert_type == "values":
+            stmt, params = (
+                self.tables.data_table.insert().values(name="r1", data=None),
+                {},
+            )
+        else:
+            assert False
+
+        conn = connection
+        conn.execute(stmt, params)
+
+        eq_(
+            conn.scalar(
+                select(self.tables.data_table.c.name).where(
+                    cast(col, String) == "null"
+                )
+            ),
+            "r1",
+        )
+
+        eq_(conn.scalar(select(col)), None)
+
+    def test_unicode_round_trip(self):
+        # note we include Unicode supplementary characters as well
+        with config.db.begin() as conn:
+            conn.execute(
+                self.tables.data_table.insert(),
+                {
+                    "name": "r1",
+                    "data": {
+                        "réve🐍 illé": "réve🐍 illé",
+                        "data": {"k1": "drôl🐍e"},
+                    },
+                },
+            )
+
+            eq_(
+                conn.scalar(select(self.tables.data_table.c.data)),
+                {
+                    "réve🐍 illé": "réve🐍 illé",
+                    "data": {"k1": "drôl🐍e"},
+                },
+            )
+
+    def test_eval_none_flag_orm(self, connection):
+        Base = declarative_base()
+
+        class Data(Base):
+            __table__ = self.tables.data_table
+
+        with Session(connection) as s:
+            d1 = Data(name="d1", data=None, nulldata=None)
+            s.add(d1)
+            s.commit()
+
+            s.bulk_insert_mappings(
+                Data, [{"name": "d2", "data": None, "nulldata": None}]
+            )
+            eq_(
+                s.query(
+                    cast(self.tables.data_table.c.data, String()),
+                    cast(self.tables.data_table.c.nulldata, String),
+                )
+                .filter(self.tables.data_table.c.name == "d1")
+                .first(),
+                ("null", None),
+            )
+            eq_(
+                s.query(
+                    cast(self.tables.data_table.c.data, String()),
+                    cast(self.tables.data_table.c.nulldata, String),
+                )
+                .filter(self.tables.data_table.c.name == "d2")
+                .first(),
+                ("null", None),
+            )
+
+
+class JSONLegacyStringCastIndexTest(
+    _LiteralRoundTripFixture, fixtures.TablesTest
+):
+    """test JSON index access with "cast to string", which we have documented
+    for a long time as how to compare JSON values, but is ultimately not
+    reliable in all cases.   The "as_XYZ()" comparators should be used
+    instead.
+
+    """
+
+    __requires__ = ("json_type", "legacy_unconditional_json_extract")
+    __backend__ = True
+
+    datatype = JSON
+
+    data1 = {"key1": "value1", "key2": "value2"}
+
+    data2 = {
+        "Key 'One'": "value1",
+        "key two": "value2",
+        "key three": "value ' three '",
+    }
+
+    data3 = {
+        "key1": [1, 2, 3],
+        "key2": ["one", "two", "three"],
+        "key3": [{"four": "five"}, {"six": "seven"}],
+    }
+
+    data4 = ["one", "two", "three"]
+
+    data5 = {
+        "nested": {
+            "elem1": [{"a": "b", "c": "d"}, {"e": "f", "g": "h"}],
+            "elem2": {"elem3": {"elem4": "elem5"}},
+        }
+    }
+
+    data6 = {"a": 5, "b": "some value", "c": {"foo": "bar"}}
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "data_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("name", String(30), nullable=False),
+            Column("data", cls.datatype),
+            Column("nulldata", cls.datatype(none_as_null=True)),
+        )
+
+    def _criteria_fixture(self):
+        with config.db.begin() as conn:
+            conn.execute(
+                self.tables.data_table.insert(),
+                [
+                    {"name": "r1", "data": self.data1},
+                    {"name": "r2", "data": self.data2},
+                    {"name": "r3", "data": self.data3},
+                    {"name": "r4", "data": self.data4},
+                    {"name": "r5", "data": self.data5},
+                    {"name": "r6", "data": self.data6},
+                ],
+            )
+
+    def _test_index_criteria(self, crit, expected, test_literal=True):
+        self._criteria_fixture()
+        with config.db.connect() as conn:
+            stmt = select(self.tables.data_table.c.name).where(crit)
+
+            eq_(conn.scalar(stmt), expected)
+
+            if test_literal:
+                literal_sql = str(
+                    stmt.compile(
+                        config.db, compile_kwargs={"literal_binds": True}
+                    )
+                )
+
+                eq_(conn.exec_driver_sql(literal_sql).scalar(), expected)
+
+    def test_string_cast_crit_spaces_in_key(self):
+        name = self.tables.data_table.c.name
+        col = self.tables.data_table.c["data"]
+
+        # limit the rows here to avoid PG error
+        # "cannot extract field from a non-object", which is
+        # fixed in 9.4 but may exist in 9.3
+        self._test_index_criteria(
+            and_(
+                name.in_(["r1", "r2", "r3"]),
+                cast(col["key two"], String) == '"value2"',
+            ),
+            "r2",
+        )
+
+    @config.requirements.json_array_indexes
+    def test_string_cast_crit_simple_int(self):
+        name = self.tables.data_table.c.name
+        col = self.tables.data_table.c["data"]
+
+        # limit the rows here to avoid PG error
+        # "cannot extract array element from a non-array", which is
+        # fixed in 9.4 but may exist in 9.3
+        self._test_index_criteria(
+            and_(
+                name == "r4",
+                cast(col[1], String) == '"two"',
+            ),
+            "r4",
+        )
+
+    def test_string_cast_crit_mixed_path(self):
+        col = self.tables.data_table.c["data"]
+        self._test_index_criteria(
+            cast(col[("key3", 1, "six")], String) == '"seven"',
+            "r3",
+        )
+
+    def test_string_cast_crit_string_path(self):
+        col = self.tables.data_table.c["data"]
+        self._test_index_criteria(
+            cast(col[("nested", "elem2", "elem3", "elem4")], String)
+            == '"elem5"',
+            "r5",
+        )
+
+    def test_string_cast_crit_against_string_basic(self):
+        name = self.tables.data_table.c.name
+        col = self.tables.data_table.c["data"]
+
+        self._test_index_criteria(
+            and_(
+                name == "r6",
+                cast(col["b"], String) == '"some value"',
+            ),
+            "r6",
+        )
+
+
+class EnumTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __backend__ = True
+
+    enum_values = "a", "b", "a%", "b%percent", "réveillé"
+
+    datatype = Enum(*enum_values, name="myenum")
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "enum_table",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("enum_data", cls.datatype),
+        )
+
+    @testing.combinations(*enum_values, argnames="data")
+    def test_round_trip(self, data, connection):
+        connection.execute(
+            self.tables.enum_table.insert(), {"id": 1, "enum_data": data}
+        )
+
+        eq_(
+            connection.scalar(
+                select(self.tables.enum_table.c.enum_data).where(
+                    self.tables.enum_table.c.id == 1
+                )
+            ),
+            data,
+        )
+
+    def test_round_trip_executemany(self, connection):
+        connection.execute(
+            self.tables.enum_table.insert(),
+            [
+                {"id": 1, "enum_data": "b%percent"},
+                {"id": 2, "enum_data": "réveillé"},
+                {"id": 3, "enum_data": "b"},
+                {"id": 4, "enum_data": "a%"},
+            ],
+        )
+
+        eq_(
+            connection.scalars(
+                select(self.tables.enum_table.c.enum_data).order_by(
+                    self.tables.enum_table.c.id
+                )
+            ).all(),
+            ["b%percent", "réveillé", "b", "a%"],
+        )
+
+    @testing.requires.insert_executemany_returning
+    def test_round_trip_executemany_returning(self, connection):
+        result = connection.execute(
+            self.tables.enum_table.insert().returning(
+                self.tables.enum_table.c.enum_data
+            ),
+            [
+                {"id": 1, "enum_data": "b%percent"},
+                {"id": 2, "enum_data": "réveillé"},
+                {"id": 3, "enum_data": "b"},
+                {"id": 4, "enum_data": "a%"},
+            ],
+        )
+
+        eq_(result.scalars().all(), ["b%percent", "réveillé", "b", "a%"])
+
+
+class UuidTest(_LiteralRoundTripFixture, fixtures.TablesTest):
+    __backend__ = True
+
+    datatype = Uuid
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "uuid_table",
+            metadata,
+            Column(
+                "id", Integer, primary_key=True, test_needs_autoincrement=True
+            ),
+            Column("uuid_data", cls.datatype),
+            Column("uuid_text_data", cls.datatype(as_uuid=False)),
+            Column("uuid_data_nonnative", Uuid(native_uuid=False)),
+            Column(
+                "uuid_text_data_nonnative",
+                Uuid(as_uuid=False, native_uuid=False),
+            ),
+        )
+
+    def test_uuid_round_trip(self, connection):
+        data = uuid.uuid4()
+        uuid_table = self.tables.uuid_table
+
+        connection.execute(
+            uuid_table.insert(),
+            {"id": 1, "uuid_data": data, "uuid_data_nonnative": data},
+        )
+        row = connection.execute(
+            select(
+                uuid_table.c.uuid_data, uuid_table.c.uuid_data_nonnative
+            ).where(
+                uuid_table.c.uuid_data == data,
+                uuid_table.c.uuid_data_nonnative == data,
+            )
+        ).first()
+        eq_(row, (data, data))
+
+    def test_uuid_text_round_trip(self, connection):
+        data = str(uuid.uuid4())
+        uuid_table = self.tables.uuid_table
+
+        connection.execute(
+            uuid_table.insert(),
+            {
+                "id": 1,
+                "uuid_text_data": data,
+                "uuid_text_data_nonnative": data,
+            },
+        )
+        row = connection.execute(
+            select(
+                uuid_table.c.uuid_text_data,
+                uuid_table.c.uuid_text_data_nonnative,
+            ).where(
+                uuid_table.c.uuid_text_data == data,
+                uuid_table.c.uuid_text_data_nonnative == data,
+            )
+        ).first()
+        eq_((row[0].lower(), row[1].lower()), (data, data))
+
+    def test_literal_uuid(self, literal_round_trip):
+        data = uuid.uuid4()
+        literal_round_trip(self.datatype, [data], [data])
+
+    def test_literal_text(self, literal_round_trip):
+        data = str(uuid.uuid4())
+        literal_round_trip(
+            self.datatype(as_uuid=False),
+            [data],
+            [data],
+            filter_=lambda x: x.lower(),
+        )
+
+    def test_literal_nonnative_uuid(self, literal_round_trip):
+        data = uuid.uuid4()
+        literal_round_trip(Uuid(native_uuid=False), [data], [data])
+
+    def test_literal_nonnative_text(self, literal_round_trip):
+        data = str(uuid.uuid4())
+        literal_round_trip(
+            Uuid(as_uuid=False, native_uuid=False),
+            [data],
+            [data],
+            filter_=lambda x: x.lower(),
+        )
+
+    @testing.requires.insert_returning
+    def test_uuid_returning(self, connection):
+        data = uuid.uuid4()
+        str_data = str(data)
+        uuid_table = self.tables.uuid_table
+
+        result = connection.execute(
+            uuid_table.insert().returning(
+                uuid_table.c.uuid_data,
+                uuid_table.c.uuid_text_data,
+                uuid_table.c.uuid_data_nonnative,
+                uuid_table.c.uuid_text_data_nonnative,
+            ),
+            {
+                "id": 1,
+                "uuid_data": data,
+                "uuid_text_data": str_data,
+                "uuid_data_nonnative": data,
+                "uuid_text_data_nonnative": str_data,
+            },
+        )
+        row = result.first()
+
+        eq_(row, (data, str_data, data, str_data))
+
+
+class NativeUUIDTest(UuidTest):
+    __requires__ = ("uuid_data_type",)
+
+    datatype = UUID
+
+
+__all__ = (
+    "ArrayTest",
+    "BinaryTest",
+    "UnicodeVarcharTest",
+    "UnicodeTextTest",
+    "JSONTest",
+    "JSONLegacyStringCastIndexTest",
+    "DateTest",
+    "DateTimeTest",
+    "DateTimeTZTest",
+    "TextTest",
+    "NumericTest",
+    "IntegerTest",
+    "IntervalTest",
+    "PrecisionIntervalTest",
+    "CastTypeDecoratorTest",
+    "DateTimeHistoricTest",
+    "DateTimeCoercedToDateTimeTest",
+    "TimeMicrosecondsTest",
+    "TimestampMicrosecondsTest",
+    "TimeTest",
+    "TimeTZTest",
+    "TrueDivTest",
+    "DateTimeMicrosecondsTest",
+    "DateHistoricTest",
+    "StringTest",
+    "BooleanTest",
+    "EnumTest",
+    "UuidTest",
+    "NativeUUIDTest",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_unicode_ddl.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_unicode_ddl.py
new file mode 100644
index 00000000..c8dd3350
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_unicode_ddl.py
@@ -0,0 +1,189 @@
+# testing/suite/test_unicode_ddl.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from sqlalchemy import desc
+from sqlalchemy import ForeignKey
+from sqlalchemy import Integer
+from sqlalchemy import MetaData
+from sqlalchemy import testing
+from sqlalchemy.testing import eq_
+from sqlalchemy.testing import fixtures
+from sqlalchemy.testing.schema import Column
+from sqlalchemy.testing.schema import Table
+
+
+class UnicodeSchemaTest(fixtures.TablesTest):
+    __requires__ = ("unicode_ddl",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        global t1, t2, t3
+
+        t1 = Table(
+            "unitable1",
+            metadata,
+            Column("méil", Integer, primary_key=True),
+            Column("\u6e2c\u8a66", Integer),
+            test_needs_fk=True,
+        )
+        t2 = Table(
+            "Unitéble2",
+            metadata,
+            Column("méil", Integer, primary_key=True, key="a"),
+            Column(
+                "\u6e2c\u8a66",
+                Integer,
+                ForeignKey("unitable1.méil"),
+                key="b",
+            ),
+            test_needs_fk=True,
+        )
+
+        # Few DBs support Unicode foreign keys
+        if testing.against("sqlite"):
+            t3 = Table(
+                "\u6e2c\u8a66",
+                metadata,
+                Column(
+                    "\u6e2c\u8a66_id",
+                    Integer,
+                    primary_key=True,
+                    autoincrement=False,
+                ),
+                Column(
+                    "unitable1_\u6e2c\u8a66",
+                    Integer,
+                    ForeignKey("unitable1.\u6e2c\u8a66"),
+                ),
+                Column("Unitéble2_b", Integer, ForeignKey("Unitéble2.b")),
+                Column(
+                    "\u6e2c\u8a66_self",
+                    Integer,
+                    ForeignKey("\u6e2c\u8a66.\u6e2c\u8a66_id"),
+                ),
+                test_needs_fk=True,
+            )
+        else:
+            t3 = Table(
+                "\u6e2c\u8a66",
+                metadata,
+                Column(
+                    "\u6e2c\u8a66_id",
+                    Integer,
+                    primary_key=True,
+                    autoincrement=False,
+                ),
+                Column("unitable1_\u6e2c\u8a66", Integer),
+                Column("Unitéble2_b", Integer),
+                Column("\u6e2c\u8a66_self", Integer),
+                test_needs_fk=True,
+            )
+
+    def test_insert(self, connection):
+        connection.execute(t1.insert(), {"méil": 1, "\u6e2c\u8a66": 5})
+        connection.execute(t2.insert(), {"a": 1, "b": 1})
+        connection.execute(
+            t3.insert(),
+            {
+                "\u6e2c\u8a66_id": 1,
+                "unitable1_\u6e2c\u8a66": 5,
+                "Unitéble2_b": 1,
+                "\u6e2c\u8a66_self": 1,
+            },
+        )
+
+        eq_(connection.execute(t1.select()).fetchall(), [(1, 5)])
+        eq_(connection.execute(t2.select()).fetchall(), [(1, 1)])
+        eq_(connection.execute(t3.select()).fetchall(), [(1, 5, 1, 1)])
+
+    def test_col_targeting(self, connection):
+        connection.execute(t1.insert(), {"méil": 1, "\u6e2c\u8a66": 5})
+        connection.execute(t2.insert(), {"a": 1, "b": 1})
+        connection.execute(
+            t3.insert(),
+            {
+                "\u6e2c\u8a66_id": 1,
+                "unitable1_\u6e2c\u8a66": 5,
+                "Unitéble2_b": 1,
+                "\u6e2c\u8a66_self": 1,
+            },
+        )
+
+        row = connection.execute(t1.select()).first()
+        eq_(row._mapping[t1.c["méil"]], 1)
+        eq_(row._mapping[t1.c["\u6e2c\u8a66"]], 5)
+
+        row = connection.execute(t2.select()).first()
+        eq_(row._mapping[t2.c["a"]], 1)
+        eq_(row._mapping[t2.c["b"]], 1)
+
+        row = connection.execute(t3.select()).first()
+        eq_(row._mapping[t3.c["\u6e2c\u8a66_id"]], 1)
+        eq_(row._mapping[t3.c["unitable1_\u6e2c\u8a66"]], 5)
+        eq_(row._mapping[t3.c["Unitéble2_b"]], 1)
+        eq_(row._mapping[t3.c["\u6e2c\u8a66_self"]], 1)
+
+    def test_reflect(self, connection):
+        connection.execute(t1.insert(), {"méil": 2, "\u6e2c\u8a66": 7})
+        connection.execute(t2.insert(), {"a": 2, "b": 2})
+        connection.execute(
+            t3.insert(),
+            {
+                "\u6e2c\u8a66_id": 2,
+                "unitable1_\u6e2c\u8a66": 7,
+                "Unitéble2_b": 2,
+                "\u6e2c\u8a66_self": 2,
+            },
+        )
+
+        meta = MetaData()
+        tt1 = Table(t1.name, meta, autoload_with=connection)
+        tt2 = Table(t2.name, meta, autoload_with=connection)
+        tt3 = Table(t3.name, meta, autoload_with=connection)
+
+        connection.execute(tt1.insert(), {"méil": 1, "\u6e2c\u8a66": 5})
+        connection.execute(tt2.insert(), {"méil": 1, "\u6e2c\u8a66": 1})
+        connection.execute(
+            tt3.insert(),
+            {
+                "\u6e2c\u8a66_id": 1,
+                "unitable1_\u6e2c\u8a66": 5,
+                "Unitéble2_b": 1,
+                "\u6e2c\u8a66_self": 1,
+            },
+        )
+
+        eq_(
+            connection.execute(tt1.select().order_by(desc("méil"))).fetchall(),
+            [(2, 7), (1, 5)],
+        )
+        eq_(
+            connection.execute(tt2.select().order_by(desc("méil"))).fetchall(),
+            [(2, 2), (1, 1)],
+        )
+        eq_(
+            connection.execute(
+                tt3.select().order_by(desc("\u6e2c\u8a66_id"))
+            ).fetchall(),
+            [(2, 7, 2, 2), (1, 5, 1, 1)],
+        )
+
+    def test_repr(self):
+        meta = MetaData()
+        t = Table("\u6e2c\u8a66", meta, Column("\u6e2c\u8a66_id", Integer))
+        eq_(
+            repr(t),
+            (
+                "Table('測試', MetaData(), "
+                "Column('測試_id', Integer(), "
+                "table=<測試>), "
+                "schema=None)"
+            ),
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_update_delete.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_update_delete.py
new file mode 100644
index 00000000..85a8d393
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/suite/test_update_delete.py
@@ -0,0 +1,139 @@
+# testing/suite/test_update_delete.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from .. import fixtures
+from ..assertions import eq_
+from ..schema import Column
+from ..schema import Table
+from ... import Integer
+from ... import String
+from ... import testing
+
+
+class SimpleUpdateDeleteTest(fixtures.TablesTest):
+    run_deletes = "each"
+    __requires__ = ("sane_rowcount",)
+    __backend__ = True
+
+    @classmethod
+    def define_tables(cls, metadata):
+        Table(
+            "plain_pk",
+            metadata,
+            Column("id", Integer, primary_key=True),
+            Column("data", String(50)),
+        )
+
+    @classmethod
+    def insert_data(cls, connection):
+        connection.execute(
+            cls.tables.plain_pk.insert(),
+            [
+                {"id": 1, "data": "d1"},
+                {"id": 2, "data": "d2"},
+                {"id": 3, "data": "d3"},
+            ],
+        )
+
+    def test_update(self, connection):
+        t = self.tables.plain_pk
+        r = connection.execute(
+            t.update().where(t.c.id == 2), dict(data="d2_new")
+        )
+        assert not r.is_insert
+        assert not r.returns_rows
+        assert r.rowcount == 1
+
+        eq_(
+            connection.execute(t.select().order_by(t.c.id)).fetchall(),
+            [(1, "d1"), (2, "d2_new"), (3, "d3")],
+        )
+
+    def test_delete(self, connection):
+        t = self.tables.plain_pk
+        r = connection.execute(t.delete().where(t.c.id == 2))
+        assert not r.is_insert
+        assert not r.returns_rows
+        assert r.rowcount == 1
+        eq_(
+            connection.execute(t.select().order_by(t.c.id)).fetchall(),
+            [(1, "d1"), (3, "d3")],
+        )
+
+    @testing.variation("criteria", ["rows", "norows", "emptyin"])
+    @testing.requires.update_returning
+    def test_update_returning(self, connection, criteria):
+        t = self.tables.plain_pk
+
+        stmt = t.update().returning(t.c.id, t.c.data)
+
+        if criteria.norows:
+            stmt = stmt.where(t.c.id == 10)
+        elif criteria.rows:
+            stmt = stmt.where(t.c.id == 2)
+        elif criteria.emptyin:
+            stmt = stmt.where(t.c.id.in_([]))
+        else:
+            criteria.fail()
+
+        r = connection.execute(stmt, dict(data="d2_new"))
+        assert not r.is_insert
+        assert r.returns_rows
+        eq_(r.keys(), ["id", "data"])
+
+        if criteria.rows:
+            eq_(r.all(), [(2, "d2_new")])
+        else:
+            eq_(r.all(), [])
+
+        eq_(
+            connection.execute(t.select().order_by(t.c.id)).fetchall(),
+            (
+                [(1, "d1"), (2, "d2_new"), (3, "d3")]
+                if criteria.rows
+                else [(1, "d1"), (2, "d2"), (3, "d3")]
+            ),
+        )
+
+    @testing.variation("criteria", ["rows", "norows", "emptyin"])
+    @testing.requires.delete_returning
+    def test_delete_returning(self, connection, criteria):
+        t = self.tables.plain_pk
+
+        stmt = t.delete().returning(t.c.id, t.c.data)
+
+        if criteria.norows:
+            stmt = stmt.where(t.c.id == 10)
+        elif criteria.rows:
+            stmt = stmt.where(t.c.id == 2)
+        elif criteria.emptyin:
+            stmt = stmt.where(t.c.id.in_([]))
+        else:
+            criteria.fail()
+
+        r = connection.execute(stmt)
+        assert not r.is_insert
+        assert r.returns_rows
+        eq_(r.keys(), ["id", "data"])
+
+        if criteria.rows:
+            eq_(r.all(), [(2, "d2")])
+        else:
+            eq_(r.all(), [])
+
+        eq_(
+            connection.execute(t.select().order_by(t.c.id)).fetchall(),
+            (
+                [(1, "d1"), (3, "d3")]
+                if criteria.rows
+                else [(1, "d1"), (2, "d2"), (3, "d3")]
+            ),
+        )
+
+
+__all__ = ("SimpleUpdateDeleteTest",)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/util.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/util.py
new file mode 100644
index 00000000..42f07710
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/util.py
@@ -0,0 +1,538 @@
+# testing/util.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+
+from __future__ import annotations
+
+from collections import deque
+import contextlib
+import decimal
+import gc
+from itertools import chain
+import random
+import sys
+from sys import getsizeof
+import time
+import types
+from typing import Any
+
+from . import config
+from . import mock
+from .. import inspect
+from ..engine import Connection
+from ..schema import Column
+from ..schema import DropConstraint
+from ..schema import DropTable
+from ..schema import ForeignKeyConstraint
+from ..schema import MetaData
+from ..schema import Table
+from ..sql import schema
+from ..sql.sqltypes import Integer
+from ..util import decorator
+from ..util import defaultdict
+from ..util import has_refcount_gc
+from ..util import inspect_getfullargspec
+
+
+if not has_refcount_gc:
+
+    def non_refcount_gc_collect(*args):
+        gc.collect()
+        gc.collect()
+
+    gc_collect = lazy_gc = non_refcount_gc_collect
+else:
+    # assume CPython - straight gc.collect, lazy_gc() is a pass
+    gc_collect = gc.collect
+
+    def lazy_gc():
+        pass
+
+
+def picklers():
+    picklers = set()
+    import pickle
+
+    picklers.add(pickle)
+
+    # yes, this thing needs this much testing
+    for pickle_ in picklers:
+        for protocol in range(-2, pickle.HIGHEST_PROTOCOL + 1):
+            yield pickle_.loads, lambda d: pickle_.dumps(d, protocol)
+
+
+def random_choices(population, k=1):
+    return random.choices(population, k=k)
+
+
+def round_decimal(value, prec):
+    if isinstance(value, float):
+        return round(value, prec)
+
+    # can also use shift() here but that is 2.6 only
+    return (value * decimal.Decimal("1" + "0" * prec)).to_integral(
+        decimal.ROUND_FLOOR
+    ) / pow(10, prec)
+
+
+class RandomSet(set):
+    def __iter__(self):
+        l = list(set.__iter__(self))
+        random.shuffle(l)
+        return iter(l)
+
+    def pop(self):
+        index = random.randint(0, len(self) - 1)
+        item = list(set.__iter__(self))[index]
+        self.remove(item)
+        return item
+
+    def union(self, other):
+        return RandomSet(set.union(self, other))
+
+    def difference(self, other):
+        return RandomSet(set.difference(self, other))
+
+    def intersection(self, other):
+        return RandomSet(set.intersection(self, other))
+
+    def copy(self):
+        return RandomSet(self)
+
+
+def conforms_partial_ordering(tuples, sorted_elements):
+    """True if the given sorting conforms to the given partial ordering."""
+
+    deps = defaultdict(set)
+    for parent, child in tuples:
+        deps[parent].add(child)
+    for i, node in enumerate(sorted_elements):
+        for n in sorted_elements[i:]:
+            if node in deps[n]:
+                return False
+    else:
+        return True
+
+
+def all_partial_orderings(tuples, elements):
+    edges = defaultdict(set)
+    for parent, child in tuples:
+        edges[child].add(parent)
+
+    def _all_orderings(elements):
+        if len(elements) == 1:
+            yield list(elements)
+        else:
+            for elem in elements:
+                subset = set(elements).difference([elem])
+                if not subset.intersection(edges[elem]):
+                    for sub_ordering in _all_orderings(subset):
+                        yield [elem] + sub_ordering
+
+    return iter(_all_orderings(elements))
+
+
+def function_named(fn, name):
+    """Return a function with a given __name__.
+
+    Will assign to __name__ and return the original function if possible on
+    the Python implementation, otherwise a new function will be constructed.
+
+    This function should be phased out as much as possible
+    in favor of @decorator.   Tests that "generate" many named tests
+    should be modernized.
+
+    """
+    try:
+        fn.__name__ = name
+    except TypeError:
+        fn = types.FunctionType(
+            fn.__code__, fn.__globals__, name, fn.__defaults__, fn.__closure__
+        )
+    return fn
+
+
+def run_as_contextmanager(ctx, fn, *arg, **kw):
+    """Run the given function under the given contextmanager,
+    simulating the behavior of 'with' to support older
+    Python versions.
+
+    This is not necessary anymore as we have placed 2.6
+    as minimum Python version, however some tests are still using
+    this structure.
+
+    """
+
+    obj = ctx.__enter__()
+    try:
+        result = fn(obj, *arg, **kw)
+        ctx.__exit__(None, None, None)
+        return result
+    except:
+        exc_info = sys.exc_info()
+        raise_ = ctx.__exit__(*exc_info)
+        if not raise_:
+            raise
+        else:
+            return raise_
+
+
+def rowset(results):
+    """Converts the results of sql execution into a plain set of column tuples.
+
+    Useful for asserting the results of an unordered query.
+    """
+
+    return {tuple(row) for row in results}
+
+
+def fail(msg):
+    assert False, msg
+
+
+@decorator
+def provide_metadata(fn, *args, **kw):
+    """Provide bound MetaData for a single test, dropping afterwards.
+
+    Legacy; use the "metadata" pytest fixture.
+
+    """
+
+    from . import fixtures
+
+    metadata = schema.MetaData()
+    self = args[0]
+    prev_meta = getattr(self, "metadata", None)
+    self.metadata = metadata
+    try:
+        return fn(*args, **kw)
+    finally:
+        # close out some things that get in the way of dropping tables.
+        # when using the "metadata" fixture, there is a set ordering
+        # of things that makes sure things are cleaned up in order, however
+        # the simple "decorator" nature of this legacy function means
+        # we have to hardcode some of that cleanup ahead of time.
+
+        # close ORM sessions
+        fixtures.close_all_sessions()
+
+        # integrate with the "connection" fixture as there are many
+        # tests where it is used along with provide_metadata
+        cfc = fixtures.base._connection_fixture_connection
+        if cfc:
+            # TODO: this warning can be used to find all the places
+            # this is used with connection fixture
+            # warn("mixing legacy provide metadata with connection fixture")
+            drop_all_tables_from_metadata(metadata, cfc)
+            # as the provide_metadata fixture is often used with "testing.db",
+            # when we do the drop we have to commit the transaction so that
+            # the DB is actually updated as the CREATE would have been
+            # committed
+            cfc.get_transaction().commit()
+        else:
+            drop_all_tables_from_metadata(metadata, config.db)
+        self.metadata = prev_meta
+
+
+def flag_combinations(*combinations):
+    """A facade around @testing.combinations() oriented towards boolean
+    keyword-based arguments.
+
+    Basically generates a nice looking identifier based on the keywords
+    and also sets up the argument names.
+
+    E.g.::
+
+        @testing.flag_combinations(
+            dict(lazy=False, passive=False),
+            dict(lazy=True, passive=False),
+            dict(lazy=False, passive=True),
+            dict(lazy=False, passive=True, raiseload=True),
+        )
+        def test_fn(lazy, passive, raiseload): ...
+
+    would result in::
+
+        @testing.combinations(
+            ("", False, False, False),
+            ("lazy", True, False, False),
+            ("lazy_passive", True, True, False),
+            ("lazy_passive", True, True, True),
+            id_="iaaa",
+            argnames="lazy,passive,raiseload",
+        )
+        def test_fn(lazy, passive, raiseload): ...
+
+    """
+
+    keys = set()
+
+    for d in combinations:
+        keys.update(d)
+
+    keys = sorted(keys)
+
+    return config.combinations(
+        *[
+            ("_".join(k for k in keys if d.get(k, False)),)
+            + tuple(d.get(k, False) for k in keys)
+            for d in combinations
+        ],
+        id_="i" + ("a" * len(keys)),
+        argnames=",".join(keys),
+    )
+
+
+def lambda_combinations(lambda_arg_sets, **kw):
+    args = inspect_getfullargspec(lambda_arg_sets)
+
+    arg_sets = lambda_arg_sets(*[mock.Mock() for arg in args[0]])
+
+    def create_fixture(pos):
+        def fixture(**kw):
+            return lambda_arg_sets(**kw)[pos]
+
+        fixture.__name__ = "fixture_%3.3d" % pos
+        return fixture
+
+    return config.combinations(
+        *[(create_fixture(i),) for i in range(len(arg_sets))], **kw
+    )
+
+
+def resolve_lambda(__fn, **kw):
+    """Given a no-arg lambda and a namespace, return a new lambda that
+    has all the values filled in.
+
+    This is used so that we can have module-level fixtures that
+    refer to instance-level variables using lambdas.
+
+    """
+
+    pos_args = inspect_getfullargspec(__fn)[0]
+    pass_pos_args = {arg: kw.pop(arg) for arg in pos_args}
+    glb = dict(__fn.__globals__)
+    glb.update(kw)
+    new_fn = types.FunctionType(__fn.__code__, glb)
+    return new_fn(**pass_pos_args)
+
+
+def metadata_fixture(ddl="function"):
+    """Provide MetaData for a pytest fixture."""
+
+    def decorate(fn):
+        def run_ddl(self):
+            metadata = self.metadata = schema.MetaData()
+            try:
+                result = fn(self, metadata)
+                metadata.create_all(config.db)
+                # TODO:
+                # somehow get a per-function dml erase fixture here
+                yield result
+            finally:
+                metadata.drop_all(config.db)
+
+        return config.fixture(scope=ddl)(run_ddl)
+
+    return decorate
+
+
+def force_drop_names(*names):
+    """Force the given table names to be dropped after test complete,
+    isolating for foreign key cycles
+
+    """
+
+    @decorator
+    def go(fn, *args, **kw):
+        try:
+            return fn(*args, **kw)
+        finally:
+            drop_all_tables(config.db, inspect(config.db), include_names=names)
+
+    return go
+
+
+class adict(dict):
+    """Dict keys available as attributes.  Shadows."""
+
+    def __getattribute__(self, key):
+        try:
+            return self[key]
+        except KeyError:
+            return dict.__getattribute__(self, key)
+
+    def __call__(self, *keys):
+        return tuple([self[key] for key in keys])
+
+    get_all = __call__
+
+
+def drop_all_tables_from_metadata(metadata, engine_or_connection):
+    from . import engines
+
+    def go(connection):
+        engines.testing_reaper.prepare_for_drop_tables(connection)
+
+        if not connection.dialect.supports_alter:
+            from . import assertions
+
+            with assertions.expect_warnings(
+                "Can't sort tables", assert_=False
+            ):
+                metadata.drop_all(connection)
+        else:
+            metadata.drop_all(connection)
+
+    if not isinstance(engine_or_connection, Connection):
+        with engine_or_connection.begin() as connection:
+            go(connection)
+    else:
+        go(engine_or_connection)
+
+
+def drop_all_tables(
+    engine,
+    inspector,
+    schema=None,
+    consider_schemas=(None,),
+    include_names=None,
+):
+    if include_names is not None:
+        include_names = set(include_names)
+
+    if schema is not None:
+        assert consider_schemas == (
+            None,
+        ), "consider_schemas and schema are mutually exclusive"
+        consider_schemas = (schema,)
+
+    with engine.begin() as conn:
+        for table_key, fkcs in reversed(
+            inspector.sort_tables_on_foreign_key_dependency(
+                consider_schemas=consider_schemas
+            )
+        ):
+            if table_key:
+                if (
+                    include_names is not None
+                    and table_key[1] not in include_names
+                ):
+                    continue
+                conn.execute(
+                    DropTable(
+                        Table(table_key[1], MetaData(), schema=table_key[0])
+                    )
+                )
+            elif fkcs:
+                if not engine.dialect.supports_alter:
+                    continue
+                for t_key, fkc in fkcs:
+                    if (
+                        include_names is not None
+                        and t_key[1] not in include_names
+                    ):
+                        continue
+                    tb = Table(
+                        t_key[1],
+                        MetaData(),
+                        Column("x", Integer),
+                        Column("y", Integer),
+                        schema=t_key[0],
+                    )
+                    conn.execute(
+                        DropConstraint(
+                            ForeignKeyConstraint([tb.c.x], [tb.c.y], name=fkc)
+                        )
+                    )
+
+
+def teardown_events(event_cls):
+    @decorator
+    def decorate(fn, *arg, **kw):
+        try:
+            return fn(*arg, **kw)
+        finally:
+            event_cls._clear()
+
+    return decorate
+
+
+def total_size(o):
+    """Returns the approximate memory footprint an object and all of its
+    contents.
+
+    source: https://code.activestate.com/recipes/577504/
+
+
+    """
+
+    def dict_handler(d):
+        return chain.from_iterable(d.items())
+
+    all_handlers = {
+        tuple: iter,
+        list: iter,
+        deque: iter,
+        dict: dict_handler,
+        set: iter,
+        frozenset: iter,
+    }
+    seen = set()  # track which object id's have already been seen
+    default_size = getsizeof(0)  # estimate sizeof object without __sizeof__
+
+    def sizeof(o):
+        if id(o) in seen:  # do not double count the same object
+            return 0
+        seen.add(id(o))
+        s = getsizeof(o, default_size)
+
+        for typ, handler in all_handlers.items():
+            if isinstance(o, typ):
+                s += sum(map(sizeof, handler(o)))
+                break
+        return s
+
+    return sizeof(o)
+
+
+def count_cache_key_tuples(tup):
+    """given a cache key tuple, counts how many instances of actual
+    tuples are found.
+
+    used to alert large jumps in cache key complexity.
+
+    """
+    stack = [tup]
+
+    sentinel = object()
+    num_elements = 0
+
+    while stack:
+        elem = stack.pop(0)
+        if elem is sentinel:
+            num_elements += 1
+        elif isinstance(elem, tuple):
+            if elem:
+                stack = list(elem) + [sentinel] + stack
+    return num_elements
+
+
+@contextlib.contextmanager
+def skip_if_timeout(seconds: float, cleanup: Any = None):
+
+    now = time.time()
+    yield
+    sec = time.time() - now
+    if sec > seconds:
+        try:
+            cleanup()
+        finally:
+            config.skip_test(
+                f"test took too long ({sec:.4f} seconds > {seconds})"
+            )
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/warnings.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/warnings.py
new file mode 100644
index 00000000..9be0813b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/testing/warnings.py
@@ -0,0 +1,52 @@
+# testing/warnings.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+from __future__ import annotations
+
+import warnings
+
+from . import assertions
+from .. import exc
+from .. import exc as sa_exc
+from ..exc import SATestSuiteWarning
+from ..util.langhelpers import _warnings_warn
+
+
+def warn_test_suite(message):
+    _warnings_warn(message, category=SATestSuiteWarning)
+
+
+def setup_filters():
+    """hook for setting up warnings filters.
+
+    SQLAlchemy-specific classes must only be here and not in pytest config,
+    as we need to delay importing SQLAlchemy until conftest.py has been
+    processed.
+
+    NOTE: filters on subclasses of DeprecationWarning or
+    PendingDeprecationWarning have no effect if added here, since pytest
+    will add at each test the following filters
+    ``always::PendingDeprecationWarning`` and ``always::DeprecationWarning``
+    that will take precedence over any added here.
+
+    """
+    warnings.filterwarnings("error", category=exc.SAWarning)
+    warnings.filterwarnings("always", category=exc.SATestSuiteWarning)
+
+
+def assert_warnings(fn, warning_msgs, regex=False):
+    """Assert that each of the given warnings are emitted by fn.
+
+    Deprecated.  Please use assertions.expect_warnings().
+
+    """
+
+    with assertions._expect_warnings(
+        sa_exc.SAWarning, warning_msgs, regex=regex
+    ):
+        return fn()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/types.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/types.py
new file mode 100644
index 00000000..c2b1ab19
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/types.py
@@ -0,0 +1,76 @@
+# types.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Compatibility namespace for sqlalchemy.sql.types.
+
+"""
+
+
+from __future__ import annotations
+
+from .sql.sqltypes import _Binary as _Binary
+from .sql.sqltypes import ARRAY as ARRAY
+from .sql.sqltypes import BIGINT as BIGINT
+from .sql.sqltypes import BigInteger as BigInteger
+from .sql.sqltypes import BINARY as BINARY
+from .sql.sqltypes import BLOB as BLOB
+from .sql.sqltypes import BOOLEAN as BOOLEAN
+from .sql.sqltypes import Boolean as Boolean
+from .sql.sqltypes import CHAR as CHAR
+from .sql.sqltypes import CLOB as CLOB
+from .sql.sqltypes import Concatenable as Concatenable
+from .sql.sqltypes import DATE as DATE
+from .sql.sqltypes import Date as Date
+from .sql.sqltypes import DATETIME as DATETIME
+from .sql.sqltypes import DateTime as DateTime
+from .sql.sqltypes import DECIMAL as DECIMAL
+from .sql.sqltypes import DOUBLE as DOUBLE
+from .sql.sqltypes import Double as Double
+from .sql.sqltypes import DOUBLE_PRECISION as DOUBLE_PRECISION
+from .sql.sqltypes import Enum as Enum
+from .sql.sqltypes import FLOAT as FLOAT
+from .sql.sqltypes import Float as Float
+from .sql.sqltypes import Indexable as Indexable
+from .sql.sqltypes import INT as INT
+from .sql.sqltypes import INTEGER as INTEGER
+from .sql.sqltypes import Integer as Integer
+from .sql.sqltypes import Interval as Interval
+from .sql.sqltypes import JSON as JSON
+from .sql.sqltypes import LargeBinary as LargeBinary
+from .sql.sqltypes import MatchType as MatchType
+from .sql.sqltypes import NCHAR as NCHAR
+from .sql.sqltypes import NULLTYPE as NULLTYPE
+from .sql.sqltypes import NullType as NullType
+from .sql.sqltypes import NUMERIC as NUMERIC
+from .sql.sqltypes import Numeric as Numeric
+from .sql.sqltypes import NVARCHAR as NVARCHAR
+from .sql.sqltypes import PickleType as PickleType
+from .sql.sqltypes import REAL as REAL
+from .sql.sqltypes import SchemaType as SchemaType
+from .sql.sqltypes import SMALLINT as SMALLINT
+from .sql.sqltypes import SmallInteger as SmallInteger
+from .sql.sqltypes import String as String
+from .sql.sqltypes import STRINGTYPE as STRINGTYPE
+from .sql.sqltypes import TEXT as TEXT
+from .sql.sqltypes import Text as Text
+from .sql.sqltypes import TIME as TIME
+from .sql.sqltypes import Time as Time
+from .sql.sqltypes import TIMESTAMP as TIMESTAMP
+from .sql.sqltypes import TupleType as TupleType
+from .sql.sqltypes import Unicode as Unicode
+from .sql.sqltypes import UnicodeText as UnicodeText
+from .sql.sqltypes import UUID as UUID
+from .sql.sqltypes import Uuid as Uuid
+from .sql.sqltypes import VARBINARY as VARBINARY
+from .sql.sqltypes import VARCHAR as VARCHAR
+from .sql.type_api import adapt_type as adapt_type
+from .sql.type_api import ExternalType as ExternalType
+from .sql.type_api import to_instance as to_instance
+from .sql.type_api import TypeDecorator as TypeDecorator
+from .sql.type_api import TypeEngine as TypeEngine
+from .sql.type_api import UserDefinedType as UserDefinedType
+from .sql.type_api import Variant as Variant
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__init__.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__init__.py
new file mode 100644
index 00000000..a9b4c3b1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__init__.py
@@ -0,0 +1,160 @@
+# util/__init__.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+
+from collections import defaultdict as defaultdict
+from functools import partial as partial
+from functools import update_wrapper as update_wrapper
+from typing import TYPE_CHECKING
+
+from . import preloaded as preloaded
+from ._collections import coerce_generator_arg as coerce_generator_arg
+from ._collections import coerce_to_immutabledict as coerce_to_immutabledict
+from ._collections import column_dict as column_dict
+from ._collections import column_set as column_set
+from ._collections import EMPTY_DICT as EMPTY_DICT
+from ._collections import EMPTY_SET as EMPTY_SET
+from ._collections import FacadeDict as FacadeDict
+from ._collections import flatten_iterator as flatten_iterator
+from ._collections import has_dupes as has_dupes
+from ._collections import has_intersection as has_intersection
+from ._collections import IdentitySet as IdentitySet
+from ._collections import immutabledict as immutabledict
+from ._collections import LRUCache as LRUCache
+from ._collections import merge_lists_w_ordering as merge_lists_w_ordering
+from ._collections import NONE_SET as NONE_SET
+from ._collections import ordered_column_set as ordered_column_set
+from ._collections import OrderedDict as OrderedDict
+from ._collections import OrderedIdentitySet as OrderedIdentitySet
+from ._collections import OrderedProperties as OrderedProperties
+from ._collections import OrderedSet as OrderedSet
+from ._collections import PopulateDict as PopulateDict
+from ._collections import Properties as Properties
+from ._collections import ReadOnlyContainer as ReadOnlyContainer
+from ._collections import ReadOnlyProperties as ReadOnlyProperties
+from ._collections import ScopedRegistry as ScopedRegistry
+from ._collections import sort_dictionary as sort_dictionary
+from ._collections import ThreadLocalRegistry as ThreadLocalRegistry
+from ._collections import to_column_set as to_column_set
+from ._collections import to_list as to_list
+from ._collections import to_set as to_set
+from ._collections import unique_list as unique_list
+from ._collections import UniqueAppender as UniqueAppender
+from ._collections import update_copy as update_copy
+from ._collections import WeakPopulateDict as WeakPopulateDict
+from ._collections import WeakSequence as WeakSequence
+from .compat import anext_ as anext_
+from .compat import arm as arm
+from .compat import b as b
+from .compat import b64decode as b64decode
+from .compat import b64encode as b64encode
+from .compat import cmp as cmp
+from .compat import cpython as cpython
+from .compat import dataclass_fields as dataclass_fields
+from .compat import decode_backslashreplace as decode_backslashreplace
+from .compat import dottedgetter as dottedgetter
+from .compat import has_refcount_gc as has_refcount_gc
+from .compat import inspect_getfullargspec as inspect_getfullargspec
+from .compat import is64bit as is64bit
+from .compat import local_dataclass_fields as local_dataclass_fields
+from .compat import osx as osx
+from .compat import py310 as py310
+from .compat import py311 as py311
+from .compat import py312 as py312
+from .compat import py313 as py313
+from .compat import py38 as py38
+from .compat import py39 as py39
+from .compat import pypy as pypy
+from .compat import win32 as win32
+from .concurrency import await_fallback as await_fallback
+from .concurrency import await_only as await_only
+from .concurrency import greenlet_spawn as greenlet_spawn
+from .concurrency import is_exit_exception as is_exit_exception
+from .deprecations import became_legacy_20 as became_legacy_20
+from .deprecations import deprecated as deprecated
+from .deprecations import deprecated_cls as deprecated_cls
+from .deprecations import deprecated_params as deprecated_params
+from .deprecations import moved_20 as moved_20
+from .deprecations import warn_deprecated as warn_deprecated
+from .langhelpers import add_parameter_text as add_parameter_text
+from .langhelpers import as_interface as as_interface
+from .langhelpers import asbool as asbool
+from .langhelpers import asint as asint
+from .langhelpers import assert_arg_type as assert_arg_type
+from .langhelpers import attrsetter as attrsetter
+from .langhelpers import bool_or_str as bool_or_str
+from .langhelpers import chop_traceback as chop_traceback
+from .langhelpers import class_hierarchy as class_hierarchy
+from .langhelpers import classproperty as classproperty
+from .langhelpers import clsname_as_plain_name as clsname_as_plain_name
+from .langhelpers import coerce_kw_type as coerce_kw_type
+from .langhelpers import constructor_copy as constructor_copy
+from .langhelpers import constructor_key as constructor_key
+from .langhelpers import counter as counter
+from .langhelpers import create_proxy_methods as create_proxy_methods
+from .langhelpers import decode_slice as decode_slice
+from .langhelpers import decorator as decorator
+from .langhelpers import dictlike_iteritems as dictlike_iteritems
+from .langhelpers import duck_type_collection as duck_type_collection
+from .langhelpers import ellipses_string as ellipses_string
+from .langhelpers import EnsureKWArg as EnsureKWArg
+from .langhelpers import FastIntFlag as FastIntFlag
+from .langhelpers import format_argspec_init as format_argspec_init
+from .langhelpers import format_argspec_plus as format_argspec_plus
+from .langhelpers import generic_fn_descriptor as generic_fn_descriptor
+from .langhelpers import generic_repr as generic_repr
+from .langhelpers import get_annotations as get_annotations
+from .langhelpers import get_callable_argspec as get_callable_argspec
+from .langhelpers import get_cls_kwargs as get_cls_kwargs
+from .langhelpers import get_func_kwargs as get_func_kwargs
+from .langhelpers import getargspec_init as getargspec_init
+from .langhelpers import has_compiled_ext as has_compiled_ext
+from .langhelpers import HasMemoized as HasMemoized
+from .langhelpers import (
+    HasMemoized_ro_memoized_attribute as HasMemoized_ro_memoized_attribute,
+)
+from .langhelpers import hybridmethod as hybridmethod
+from .langhelpers import hybridproperty as hybridproperty
+from .langhelpers import inject_docstring_text as inject_docstring_text
+from .langhelpers import iterate_attributes as iterate_attributes
+from .langhelpers import map_bits as map_bits
+from .langhelpers import md5_hex as md5_hex
+from .langhelpers import memoized_instancemethod as memoized_instancemethod
+from .langhelpers import memoized_property as memoized_property
+from .langhelpers import MemoizedSlots as MemoizedSlots
+from .langhelpers import method_is_overridden as method_is_overridden
+from .langhelpers import methods_equivalent as methods_equivalent
+from .langhelpers import (
+    monkeypatch_proxied_specials as monkeypatch_proxied_specials,
+)
+from .langhelpers import non_memoized_property as non_memoized_property
+from .langhelpers import NoneType as NoneType
+from .langhelpers import only_once as only_once
+from .langhelpers import (
+    parse_user_argument_for_enum as parse_user_argument_for_enum,
+)
+from .langhelpers import PluginLoader as PluginLoader
+from .langhelpers import portable_instancemethod as portable_instancemethod
+from .langhelpers import quoted_token_parser as quoted_token_parser
+from .langhelpers import ro_memoized_property as ro_memoized_property
+from .langhelpers import ro_non_memoized_property as ro_non_memoized_property
+from .langhelpers import rw_hybridproperty as rw_hybridproperty
+from .langhelpers import safe_reraise as safe_reraise
+from .langhelpers import set_creation_order as set_creation_order
+from .langhelpers import string_or_unprintable as string_or_unprintable
+from .langhelpers import symbol as symbol
+from .langhelpers import TypingOnly as TypingOnly
+from .langhelpers import (
+    unbound_method_to_callable as unbound_method_to_callable,
+)
+from .langhelpers import walk_subclasses as walk_subclasses
+from .langhelpers import warn as warn
+from .langhelpers import warn_exception as warn_exception
+from .langhelpers import warn_limited as warn_limited
+from .langhelpers import wrap_callable as wrap_callable
+from .preloaded import preload_module as preload_module
+from .typing import is_non_string_iterable as is_non_string_iterable
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b48daf39
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_collections.cpython-312.pyc
new file mode 100644
index 00000000..6d9ae50f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_concurrency_py3k.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_concurrency_py3k.cpython-312.pyc
new file mode 100644
index 00000000..908941b2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_concurrency_py3k.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_has_cy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_has_cy.cpython-312.pyc
new file mode 100644
index 00000000..c87df937
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_has_cy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_py_collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_py_collections.cpython-312.pyc
new file mode 100644
index 00000000..6ca016ae
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/_py_collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/compat.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/compat.cpython-312.pyc
new file mode 100644
index 00000000..aeeccc45
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/compat.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/concurrency.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/concurrency.cpython-312.pyc
new file mode 100644
index 00000000..68c26364
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/concurrency.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/deprecations.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/deprecations.cpython-312.pyc
new file mode 100644
index 00000000..f219b50e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/deprecations.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/langhelpers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/langhelpers.cpython-312.pyc
new file mode 100644
index 00000000..e55da2bb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/langhelpers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/preloaded.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/preloaded.cpython-312.pyc
new file mode 100644
index 00000000..041f5838
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/preloaded.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/queue.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/queue.cpython-312.pyc
new file mode 100644
index 00000000..fcb83dc2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/queue.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/tool_support.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/tool_support.cpython-312.pyc
new file mode 100644
index 00000000..2c476182
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/tool_support.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/topological.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/topological.cpython-312.pyc
new file mode 100644
index 00000000..e99f37a3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/topological.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/typing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/typing.cpython-312.pyc
new file mode 100644
index 00000000..904d05c8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/__pycache__/typing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_collections.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_collections.py
new file mode 100644
index 00000000..12283eba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_collections.py
@@ -0,0 +1,715 @@
+# util/_collections.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Collection classes and helpers."""
+from __future__ import annotations
+
+import operator
+import threading
+import types
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Container
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import TypeVar
+from typing import Union
+from typing import ValuesView
+import weakref
+
+from ._has_cy import HAS_CYEXTENSION
+from .typing import is_non_string_iterable
+from .typing import Literal
+from .typing import Protocol
+
+if typing.TYPE_CHECKING or not HAS_CYEXTENSION:
+    from ._py_collections import immutabledict as immutabledict
+    from ._py_collections import IdentitySet as IdentitySet
+    from ._py_collections import ReadOnlyContainer as ReadOnlyContainer
+    from ._py_collections import ImmutableDictBase as ImmutableDictBase
+    from ._py_collections import OrderedSet as OrderedSet
+    from ._py_collections import unique_list as unique_list
+else:
+    from sqlalchemy.cyextension.immutabledict import (
+        ReadOnlyContainer as ReadOnlyContainer,
+    )
+    from sqlalchemy.cyextension.immutabledict import (
+        ImmutableDictBase as ImmutableDictBase,
+    )
+    from sqlalchemy.cyextension.immutabledict import (
+        immutabledict as immutabledict,
+    )
+    from sqlalchemy.cyextension.collections import IdentitySet as IdentitySet
+    from sqlalchemy.cyextension.collections import OrderedSet as OrderedSet
+    from sqlalchemy.cyextension.collections import (  # noqa
+        unique_list as unique_list,
+    )
+
+
+_T = TypeVar("_T", bound=Any)
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+_T_co = TypeVar("_T_co", covariant=True)
+
+EMPTY_SET: FrozenSet[Any] = frozenset()
+NONE_SET: FrozenSet[Any] = frozenset([None])
+
+
+def merge_lists_w_ordering(a: List[Any], b: List[Any]) -> List[Any]:
+    """merge two lists, maintaining ordering as much as possible.
+
+    this is to reconcile vars(cls) with cls.__annotations__.
+
+    Example::
+
+        >>> a = ["__tablename__", "id", "x", "created_at"]
+        >>> b = ["id", "name", "data", "y", "created_at"]
+        >>> merge_lists_w_ordering(a, b)
+        ['__tablename__', 'id', 'name', 'data', 'y', 'x', 'created_at']
+
+    This is not necessarily the ordering that things had on the class,
+    in this case the class is::
+
+        class User(Base):
+            __tablename__ = "users"
+
+            id: Mapped[int] = mapped_column(primary_key=True)
+            name: Mapped[str]
+            data: Mapped[Optional[str]]
+            x = Column(Integer)
+            y: Mapped[int]
+            created_at: Mapped[datetime.datetime] = mapped_column()
+
+    But things are *mostly* ordered.
+
+    The algorithm could also be done by creating a partial ordering for
+    all items in both lists and then using topological_sort(), but that
+    is too much overhead.
+
+    Background on how I came up with this is at:
+    https://gist.github.com/zzzeek/89de958cf0803d148e74861bd682ebae
+
+    """
+    overlap = set(a).intersection(b)
+
+    result = []
+
+    current, other = iter(a), iter(b)
+
+    while True:
+        for element in current:
+            if element in overlap:
+                overlap.discard(element)
+                other, current = current, other
+                break
+
+            result.append(element)
+        else:
+            result.extend(other)
+            break
+
+    return result
+
+
+def coerce_to_immutabledict(d: Mapping[_KT, _VT]) -> immutabledict[_KT, _VT]:
+    if not d:
+        return EMPTY_DICT
+    elif isinstance(d, immutabledict):
+        return d
+    else:
+        return immutabledict(d)
+
+
+EMPTY_DICT: immutabledict[Any, Any] = immutabledict()
+
+
+class FacadeDict(ImmutableDictBase[_KT, _VT]):
+    """A dictionary that is not publicly mutable."""
+
+    def __new__(cls, *args: Any) -> FacadeDict[Any, Any]:
+        new = ImmutableDictBase.__new__(cls)
+        return new
+
+    def copy(self) -> NoReturn:
+        raise NotImplementedError(
+            "an immutabledict shouldn't need to be copied.  use dict(d) "
+            "if you need a mutable dictionary."
+        )
+
+    def __reduce__(self) -> Any:
+        return FacadeDict, (dict(self),)
+
+    def _insert_item(self, key: _KT, value: _VT) -> None:
+        """insert an item into the dictionary directly."""
+        dict.__setitem__(self, key, value)
+
+    def __repr__(self) -> str:
+        return "FacadeDict(%s)" % dict.__repr__(self)
+
+
+_DT = TypeVar("_DT", bound=Any)
+
+_F = TypeVar("_F", bound=Any)
+
+
+class Properties(Generic[_T]):
+    """Provide a __getattr__/__setattr__ interface over a dict."""
+
+    __slots__ = ("_data",)
+
+    _data: Dict[str, _T]
+
+    def __init__(self, data: Dict[str, _T]):
+        object.__setattr__(self, "_data", data)
+
+    def __len__(self) -> int:
+        return len(self._data)
+
+    def __iter__(self) -> Iterator[_T]:
+        return iter(list(self._data.values()))
+
+    def __dir__(self) -> List[str]:
+        return dir(super()) + [str(k) for k in self._data.keys()]
+
+    def __add__(self, other: Properties[_F]) -> List[Union[_T, _F]]:
+        return list(self) + list(other)
+
+    def __setitem__(self, key: str, obj: _T) -> None:
+        self._data[key] = obj
+
+    def __getitem__(self, key: str) -> _T:
+        return self._data[key]
+
+    def __delitem__(self, key: str) -> None:
+        del self._data[key]
+
+    def __setattr__(self, key: str, obj: _T) -> None:
+        self._data[key] = obj
+
+    def __getstate__(self) -> Dict[str, Any]:
+        return {"_data": self._data}
+
+    def __setstate__(self, state: Dict[str, Any]) -> None:
+        object.__setattr__(self, "_data", state["_data"])
+
+    def __getattr__(self, key: str) -> _T:
+        try:
+            return self._data[key]
+        except KeyError:
+            raise AttributeError(key)
+
+    def __contains__(self, key: str) -> bool:
+        return key in self._data
+
+    def as_readonly(self) -> ReadOnlyProperties[_T]:
+        """Return an immutable proxy for this :class:`.Properties`."""
+
+        return ReadOnlyProperties(self._data)
+
+    def update(self, value: Dict[str, _T]) -> None:
+        self._data.update(value)
+
+    @overload
+    def get(self, key: str) -> Optional[_T]: ...
+
+    @overload
+    def get(self, key: str, default: Union[_DT, _T]) -> Union[_DT, _T]: ...
+
+    def get(
+        self, key: str, default: Optional[Union[_DT, _T]] = None
+    ) -> Optional[Union[_T, _DT]]:
+        if key in self:
+            return self[key]
+        else:
+            return default
+
+    def keys(self) -> List[str]:
+        return list(self._data)
+
+    def values(self) -> List[_T]:
+        return list(self._data.values())
+
+    def items(self) -> List[Tuple[str, _T]]:
+        return list(self._data.items())
+
+    def has_key(self, key: str) -> bool:
+        return key in self._data
+
+    def clear(self) -> None:
+        self._data.clear()
+
+
+class OrderedProperties(Properties[_T]):
+    """Provide a __getattr__/__setattr__ interface with an OrderedDict
+    as backing store."""
+
+    __slots__ = ()
+
+    def __init__(self):
+        Properties.__init__(self, OrderedDict())
+
+
+class ReadOnlyProperties(ReadOnlyContainer, Properties[_T]):
+    """Provide immutable dict/object attribute to an underlying dictionary."""
+
+    __slots__ = ()
+
+
+def _ordered_dictionary_sort(d, key=None):
+    """Sort an OrderedDict in-place."""
+
+    items = [(k, d[k]) for k in sorted(d, key=key)]
+
+    d.clear()
+
+    d.update(items)
+
+
+OrderedDict = dict
+sort_dictionary = _ordered_dictionary_sort
+
+
+class WeakSequence(Sequence[_T]):
+    def __init__(self, __elements: Sequence[_T] = ()):
+        # adapted from weakref.WeakKeyDictionary, prevent reference
+        # cycles in the collection itself
+        def _remove(item, selfref=weakref.ref(self)):
+            self = selfref()
+            if self is not None:
+                self._storage.remove(item)
+
+        self._remove = _remove
+        self._storage = [
+            weakref.ref(element, _remove) for element in __elements
+        ]
+
+    def append(self, item):
+        self._storage.append(weakref.ref(item, self._remove))
+
+    def __len__(self):
+        return len(self._storage)
+
+    def __iter__(self):
+        return (
+            obj for obj in (ref() for ref in self._storage) if obj is not None
+        )
+
+    def __getitem__(self, index):
+        try:
+            obj = self._storage[index]
+        except KeyError:
+            raise IndexError("Index %s out of range" % index)
+        else:
+            return obj()
+
+
+class OrderedIdentitySet(IdentitySet):
+    def __init__(self, iterable: Optional[Iterable[Any]] = None):
+        IdentitySet.__init__(self)
+        self._members = OrderedDict()
+        if iterable:
+            for o in iterable:
+                self.add(o)
+
+
+class PopulateDict(Dict[_KT, _VT]):
+    """A dict which populates missing values via a creation function.
+
+    Note the creation function takes a key, unlike
+    collections.defaultdict.
+
+    """
+
+    def __init__(self, creator: Callable[[_KT], _VT]):
+        self.creator = creator
+
+    def __missing__(self, key: Any) -> Any:
+        self[key] = val = self.creator(key)
+        return val
+
+
+class WeakPopulateDict(Dict[_KT, _VT]):
+    """Like PopulateDict, but assumes a self + a method and does not create
+    a reference cycle.
+
+    """
+
+    def __init__(self, creator_method: types.MethodType):
+        self.creator = creator_method.__func__
+        weakself = creator_method.__self__
+        self.weakself = weakref.ref(weakself)
+
+    def __missing__(self, key: Any) -> Any:
+        self[key] = val = self.creator(self.weakself(), key)
+        return val
+
+
+# Define collections that are capable of storing
+# ColumnElement objects as hashable keys/elements.
+# At this point, these are mostly historical, things
+# used to be more complicated.
+column_set = set
+column_dict = dict
+ordered_column_set = OrderedSet
+
+
+class UniqueAppender(Generic[_T]):
+    """Appends items to a collection ensuring uniqueness.
+
+    Additional appends() of the same object are ignored.  Membership is
+    determined by identity (``is a``) not equality (``==``).
+    """
+
+    __slots__ = "data", "_data_appender", "_unique"
+
+    data: Union[Iterable[_T], Set[_T], List[_T]]
+    _data_appender: Callable[[_T], None]
+    _unique: Dict[int, Literal[True]]
+
+    def __init__(
+        self,
+        data: Union[Iterable[_T], Set[_T], List[_T]],
+        via: Optional[str] = None,
+    ):
+        self.data = data
+        self._unique = {}
+        if via:
+            self._data_appender = getattr(data, via)
+        elif hasattr(data, "append"):
+            self._data_appender = cast("List[_T]", data).append
+        elif hasattr(data, "add"):
+            self._data_appender = cast("Set[_T]", data).add
+
+    def append(self, item: _T) -> None:
+        id_ = id(item)
+        if id_ not in self._unique:
+            self._data_appender(item)
+            self._unique[id_] = True
+
+    def __iter__(self) -> Iterator[_T]:
+        return iter(self.data)
+
+
+def coerce_generator_arg(arg: Any) -> List[Any]:
+    if len(arg) == 1 and isinstance(arg[0], types.GeneratorType):
+        return list(arg[0])
+    else:
+        return cast("List[Any]", arg)
+
+
+def to_list(x: Any, default: Optional[List[Any]] = None) -> List[Any]:
+    if x is None:
+        return default  # type: ignore
+    if not is_non_string_iterable(x):
+        return [x]
+    elif isinstance(x, list):
+        return x
+    else:
+        return list(x)
+
+
+def has_intersection(set_: Container[Any], iterable: Iterable[Any]) -> bool:
+    r"""return True if any items of set\_ are present in iterable.
+
+    Goes through special effort to ensure __hash__ is not called
+    on items in iterable that don't support it.
+
+    """
+    return any(i in set_ for i in iterable if i.__hash__)
+
+
+def to_set(x):
+    if x is None:
+        return set()
+    if not isinstance(x, set):
+        return set(to_list(x))
+    else:
+        return x
+
+
+def to_column_set(x: Any) -> Set[Any]:
+    if x is None:
+        return column_set()
+    if not isinstance(x, column_set):
+        return column_set(to_list(x))
+    else:
+        return x
+
+
+def update_copy(d, _new=None, **kw):
+    """Copy the given dict and update with the given values."""
+
+    d = d.copy()
+    if _new:
+        d.update(_new)
+    d.update(**kw)
+    return d
+
+
+def flatten_iterator(x: Iterable[_T]) -> Iterator[_T]:
+    """Given an iterator of which further sub-elements may also be
+    iterators, flatten the sub-elements into a single iterator.
+
+    """
+    elem: _T
+    for elem in x:
+        if not isinstance(elem, str) and hasattr(elem, "__iter__"):
+            yield from flatten_iterator(elem)
+        else:
+            yield elem
+
+
+class LRUCache(typing.MutableMapping[_KT, _VT]):
+    """Dictionary with 'squishy' removal of least
+    recently used items.
+
+    Note that either get() or [] should be used here, but
+    generally its not safe to do an "in" check first as the dictionary
+    can change subsequent to that call.
+
+    """
+
+    __slots__ = (
+        "capacity",
+        "threshold",
+        "size_alert",
+        "_data",
+        "_counter",
+        "_mutex",
+    )
+
+    capacity: int
+    threshold: float
+    size_alert: Optional[Callable[[LRUCache[_KT, _VT]], None]]
+
+    def __init__(
+        self,
+        capacity: int = 100,
+        threshold: float = 0.5,
+        size_alert: Optional[Callable[..., None]] = None,
+    ):
+        self.capacity = capacity
+        self.threshold = threshold
+        self.size_alert = size_alert
+        self._counter = 0
+        self._mutex = threading.Lock()
+        self._data: Dict[_KT, Tuple[_KT, _VT, List[int]]] = {}
+
+    def _inc_counter(self):
+        self._counter += 1
+        return self._counter
+
+    @overload
+    def get(self, key: _KT) -> Optional[_VT]: ...
+
+    @overload
+    def get(self, key: _KT, default: Union[_VT, _T]) -> Union[_VT, _T]: ...
+
+    def get(
+        self, key: _KT, default: Optional[Union[_VT, _T]] = None
+    ) -> Optional[Union[_VT, _T]]:
+        item = self._data.get(key)
+        if item is not None:
+            item[2][0] = self._inc_counter()
+            return item[1]
+        else:
+            return default
+
+    def __getitem__(self, key: _KT) -> _VT:
+        item = self._data[key]
+        item[2][0] = self._inc_counter()
+        return item[1]
+
+    def __iter__(self) -> Iterator[_KT]:
+        return iter(self._data)
+
+    def __len__(self) -> int:
+        return len(self._data)
+
+    def values(self) -> ValuesView[_VT]:
+        return typing.ValuesView({k: i[1] for k, i in self._data.items()})
+
+    def __setitem__(self, key: _KT, value: _VT) -> None:
+        self._data[key] = (key, value, [self._inc_counter()])
+        self._manage_size()
+
+    def __delitem__(self, __v: _KT) -> None:
+        del self._data[__v]
+
+    @property
+    def size_threshold(self) -> float:
+        return self.capacity + self.capacity * self.threshold
+
+    def _manage_size(self) -> None:
+        if not self._mutex.acquire(False):
+            return
+        try:
+            size_alert = bool(self.size_alert)
+            while len(self) > self.capacity + self.capacity * self.threshold:
+                if size_alert:
+                    size_alert = False
+                    self.size_alert(self)  # type: ignore
+                by_counter = sorted(
+                    self._data.values(),
+                    key=operator.itemgetter(2),
+                    reverse=True,
+                )
+                for item in by_counter[self.capacity :]:
+                    try:
+                        del self._data[item[0]]
+                    except KeyError:
+                        # deleted elsewhere; skip
+                        continue
+        finally:
+            self._mutex.release()
+
+
+class _CreateFuncType(Protocol[_T_co]):
+    def __call__(self) -> _T_co: ...
+
+
+class _ScopeFuncType(Protocol):
+    def __call__(self) -> Any: ...
+
+
+class ScopedRegistry(Generic[_T]):
+    """A Registry that can store one or multiple instances of a single
+    class on the basis of a "scope" function.
+
+    The object implements ``__call__`` as the "getter", so by
+    calling ``myregistry()`` the contained object is returned
+    for the current scope.
+
+    :param createfunc:
+      a callable that returns a new object to be placed in the registry
+
+    :param scopefunc:
+      a callable that will return a key to store/retrieve an object.
+    """
+
+    __slots__ = "createfunc", "scopefunc", "registry"
+
+    createfunc: _CreateFuncType[_T]
+    scopefunc: _ScopeFuncType
+    registry: Any
+
+    def __init__(
+        self, createfunc: Callable[[], _T], scopefunc: Callable[[], Any]
+    ):
+        """Construct a new :class:`.ScopedRegistry`.
+
+        :param createfunc:  A creation function that will generate
+          a new value for the current scope, if none is present.
+
+        :param scopefunc:  A function that returns a hashable
+          token representing the current scope (such as, current
+          thread identifier).
+
+        """
+        self.createfunc = createfunc
+        self.scopefunc = scopefunc
+        self.registry = {}
+
+    def __call__(self) -> _T:
+        key = self.scopefunc()
+        try:
+            return self.registry[key]  # type: ignore[no-any-return]
+        except KeyError:
+            return self.registry.setdefault(key, self.createfunc())  # type: ignore[no-any-return] # noqa: E501
+
+    def has(self) -> bool:
+        """Return True if an object is present in the current scope."""
+
+        return self.scopefunc() in self.registry
+
+    def set(self, obj: _T) -> None:
+        """Set the value for the current scope."""
+
+        self.registry[self.scopefunc()] = obj
+
+    def clear(self) -> None:
+        """Clear the current scope, if any."""
+
+        try:
+            del self.registry[self.scopefunc()]
+        except KeyError:
+            pass
+
+
+class ThreadLocalRegistry(ScopedRegistry[_T]):
+    """A :class:`.ScopedRegistry` that uses a ``threading.local()``
+    variable for storage.
+
+    """
+
+    def __init__(self, createfunc: Callable[[], _T]):
+        self.createfunc = createfunc
+        self.registry = threading.local()
+
+    def __call__(self) -> _T:
+        try:
+            return self.registry.value  # type: ignore[no-any-return]
+        except AttributeError:
+            val = self.registry.value = self.createfunc()
+            return val
+
+    def has(self) -> bool:
+        return hasattr(self.registry, "value")
+
+    def set(self, obj: _T) -> None:
+        self.registry.value = obj
+
+    def clear(self) -> None:
+        try:
+            del self.registry.value
+        except AttributeError:
+            pass
+
+
+def has_dupes(sequence, target):
+    """Given a sequence and search object, return True if there's more
+    than one, False if zero or one of them.
+
+
+    """
+    # compare to .index version below, this version introduces less function
+    # overhead and is usually the same speed.  At 15000 items (way bigger than
+    # a relationship-bound collection in memory usually is) it begins to
+    # fall behind the other version only by microseconds.
+    c = 0
+    for item in sequence:
+        if item is target:
+            c += 1
+            if c > 1:
+                return True
+    return False
+
+
+# .index version.  the two __contains__ calls as well
+# as .index() and isinstance() slow this down.
+# def has_dupes(sequence, target):
+#    if target not in sequence:
+#        return False
+#    elif not isinstance(sequence, collections_abc.Sequence):
+#        return False
+#
+#    idx = sequence.index(target)
+#    return target in sequence[idx + 1:]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_concurrency_py3k.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_concurrency_py3k.py
new file mode 100644
index 00000000..718c077c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_concurrency_py3k.py
@@ -0,0 +1,288 @@
+# util/_concurrency_py3k.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import asyncio
+from contextvars import Context
+import sys
+import typing
+from typing import Any
+from typing import Awaitable
+from typing import Callable
+from typing import Coroutine
+from typing import Optional
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from .langhelpers import memoized_property
+from .. import exc
+from ..util import py311
+from ..util.typing import Literal
+from ..util.typing import Protocol
+from ..util.typing import Self
+from ..util.typing import TypeGuard
+
+_T = TypeVar("_T")
+
+if typing.TYPE_CHECKING:
+
+    class greenlet(Protocol):
+        dead: bool
+        gr_context: Optional[Context]
+
+        def __init__(self, fn: Callable[..., Any], driver: greenlet): ...
+
+        def throw(self, *arg: Any) -> Any:
+            return None
+
+        def switch(self, value: Any) -> Any:
+            return None
+
+    def getcurrent() -> greenlet: ...
+
+else:
+    from greenlet import getcurrent
+    from greenlet import greenlet
+
+
+# If greenlet.gr_context is present in current version of greenlet,
+# it will be set with the current context on creation.
+# Refs: https://github.com/python-greenlet/greenlet/pull/198
+_has_gr_context = hasattr(getcurrent(), "gr_context")
+
+
+def is_exit_exception(e: BaseException) -> bool:
+    # note asyncio.CancelledError is already BaseException
+    # so was an exit exception in any case
+    return not isinstance(e, Exception) or isinstance(
+        e, (asyncio.TimeoutError, asyncio.CancelledError)
+    )
+
+
+# implementation based on snaury gist at
+# https://gist.github.com/snaury/202bf4f22c41ca34e56297bae5f33fef
+# Issue for context: https://github.com/python-greenlet/greenlet/issues/173
+
+
+class _AsyncIoGreenlet(greenlet):
+    dead: bool
+
+    __sqlalchemy_greenlet_provider__ = True
+
+    def __init__(self, fn: Callable[..., Any], driver: greenlet):
+        greenlet.__init__(self, fn, driver)
+        if _has_gr_context:
+            self.gr_context = driver.gr_context
+
+
+_T_co = TypeVar("_T_co", covariant=True)
+
+if TYPE_CHECKING:
+
+    def iscoroutine(
+        awaitable: Awaitable[_T_co],
+    ) -> TypeGuard[Coroutine[Any, Any, _T_co]]: ...
+
+else:
+    iscoroutine = asyncio.iscoroutine
+
+
+def _safe_cancel_awaitable(awaitable: Awaitable[Any]) -> None:
+    # https://docs.python.org/3/reference/datamodel.html#coroutine.close
+
+    if iscoroutine(awaitable):
+        awaitable.close()
+
+
+def in_greenlet() -> bool:
+    current = getcurrent()
+    return getattr(current, "__sqlalchemy_greenlet_provider__", False)
+
+
+def await_only(awaitable: Awaitable[_T]) -> _T:
+    """Awaits an async function in a sync method.
+
+    The sync method must be inside a :func:`greenlet_spawn` context.
+    :func:`await_only` calls cannot be nested.
+
+    :param awaitable: The coroutine to call.
+
+    """
+    # this is called in the context greenlet while running fn
+    current = getcurrent()
+    if not getattr(current, "__sqlalchemy_greenlet_provider__", False):
+        _safe_cancel_awaitable(awaitable)
+
+        raise exc.MissingGreenlet(
+            "greenlet_spawn has not been called; can't call await_only() "
+            "here. Was IO attempted in an unexpected place?"
+        )
+
+    # returns the control to the driver greenlet passing it
+    # a coroutine to run. Once the awaitable is done, the driver greenlet
+    # switches back to this greenlet with the result of awaitable that is
+    # then returned to the caller (or raised as error)
+    return current.parent.switch(awaitable)  # type: ignore[no-any-return,attr-defined] # noqa: E501
+
+
+def await_fallback(awaitable: Awaitable[_T]) -> _T:
+    """Awaits an async function in a sync method.
+
+    The sync method must be inside a :func:`greenlet_spawn` context.
+    :func:`await_fallback` calls cannot be nested.
+
+    :param awaitable: The coroutine to call.
+
+    .. deprecated:: 2.0.24 The ``await_fallback()`` function will be removed
+       in SQLAlchemy 2.1.  Use :func:`_util.await_only` instead, running the
+       function / program / etc. within a top-level greenlet that is set up
+       using :func:`_util.greenlet_spawn`.
+
+    """
+
+    # this is called in the context greenlet while running fn
+    current = getcurrent()
+    if not getattr(current, "__sqlalchemy_greenlet_provider__", False):
+        loop = get_event_loop()
+        if loop.is_running():
+            _safe_cancel_awaitable(awaitable)
+
+            raise exc.MissingGreenlet(
+                "greenlet_spawn has not been called and asyncio event "
+                "loop is already running; can't call await_fallback() here. "
+                "Was IO attempted in an unexpected place?"
+            )
+        return loop.run_until_complete(awaitable)
+
+    return current.parent.switch(awaitable)  # type: ignore[no-any-return,attr-defined] # noqa: E501
+
+
+async def greenlet_spawn(
+    fn: Callable[..., _T],
+    *args: Any,
+    _require_await: bool = False,
+    **kwargs: Any,
+) -> _T:
+    """Runs a sync function ``fn`` in a new greenlet.
+
+    The sync function can then use :func:`await_only` to wait for async
+    functions.
+
+    :param fn: The sync callable to call.
+    :param \\*args: Positional arguments to pass to the ``fn`` callable.
+    :param \\*\\*kwargs: Keyword arguments to pass to the ``fn`` callable.
+    """
+
+    result: Any
+    context = _AsyncIoGreenlet(fn, getcurrent())
+    # runs the function synchronously in gl greenlet. If the execution
+    # is interrupted by await_only, context is not dead and result is a
+    # coroutine to wait. If the context is dead the function has
+    # returned, and its result can be returned.
+    switch_occurred = False
+    result = context.switch(*args, **kwargs)
+    while not context.dead:
+        switch_occurred = True
+        try:
+            # wait for a coroutine from await_only and then return its
+            # result back to it.
+            value = await result
+        except BaseException:
+            # this allows an exception to be raised within
+            # the moderated greenlet so that it can continue
+            # its expected flow.
+            result = context.throw(*sys.exc_info())
+        else:
+            result = context.switch(value)
+
+    if _require_await and not switch_occurred:
+        raise exc.AwaitRequired(
+            "The current operation required an async execution but none was "
+            "detected. This will usually happen when using a non compatible "
+            "DBAPI driver. Please ensure that an async DBAPI is used."
+        )
+    return result  # type: ignore[no-any-return]
+
+
+class AsyncAdaptedLock:
+    @memoized_property
+    def mutex(self) -> asyncio.Lock:
+        # there should not be a race here for coroutines creating the
+        # new lock as we are not using await, so therefore no concurrency
+        return asyncio.Lock()
+
+    def __enter__(self) -> bool:
+        # await is used to acquire the lock only after the first calling
+        # coroutine has created the mutex.
+        return await_fallback(self.mutex.acquire())
+
+    def __exit__(self, *arg: Any, **kw: Any) -> None:
+        self.mutex.release()
+
+
+def get_event_loop() -> asyncio.AbstractEventLoop:
+    """vendor asyncio.get_event_loop() for python 3.7 and above.
+
+    Python 3.10 deprecates get_event_loop() as a standalone.
+
+    """
+    try:
+        return asyncio.get_running_loop()
+    except RuntimeError:
+        # avoid "During handling of the above exception, another exception..."
+        pass
+    return asyncio.get_event_loop_policy().get_event_loop()
+
+
+if not TYPE_CHECKING and py311:
+    _Runner = asyncio.Runner
+else:
+
+    class _Runner:
+        """Runner implementation for test only"""
+
+        _loop: Union[None, asyncio.AbstractEventLoop, Literal[False]]
+
+        def __init__(self) -> None:
+            self._loop = None
+
+        def __enter__(self) -> Self:
+            self._lazy_init()
+            return self
+
+        def __exit__(self, exc_type: Any, exc_val: Any, exc_tb: Any) -> None:
+            self.close()
+
+        def close(self) -> None:
+            if self._loop:
+                try:
+                    self._loop.run_until_complete(
+                        self._loop.shutdown_asyncgens()
+                    )
+                finally:
+                    self._loop.close()
+                    self._loop = False
+
+        def get_loop(self) -> asyncio.AbstractEventLoop:
+            """Return embedded event loop."""
+            self._lazy_init()
+            assert self._loop
+            return self._loop
+
+        def run(self, coro: Coroutine[Any, Any, _T]) -> _T:
+            self._lazy_init()
+            assert self._loop
+            return self._loop.run_until_complete(coro)
+
+        def _lazy_init(self) -> None:
+            if self._loop is False:
+                raise RuntimeError("Runner is closed")
+            if self._loop is None:
+                self._loop = asyncio.new_event_loop()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_has_cy.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_has_cy.py
new file mode 100644
index 00000000..21faed04
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_has_cy.py
@@ -0,0 +1,40 @@
+# util/_has_cy.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: ignore-errors
+
+import os
+import typing
+
+
+def _import_cy_extensions():
+    # all cython extension extension modules are treated as optional by the
+    # setup, so to ensure that all are compiled, all should be imported here
+    from ..cyextension import collections
+    from ..cyextension import immutabledict
+    from ..cyextension import processors
+    from ..cyextension import resultproxy
+    from ..cyextension import util
+
+    return (collections, immutabledict, processors, resultproxy, util)
+
+
+_CYEXTENSION_MSG: str
+if not typing.TYPE_CHECKING:
+    if os.environ.get("DISABLE_SQLALCHEMY_CEXT_RUNTIME"):
+        HAS_CYEXTENSION = False
+        _CYEXTENSION_MSG = "DISABLE_SQLALCHEMY_CEXT_RUNTIME is set"
+    else:
+        try:
+            _import_cy_extensions()
+        except ImportError as err:
+            HAS_CYEXTENSION = False
+            _CYEXTENSION_MSG = str(err)
+        else:
+            _CYEXTENSION_MSG = "Loaded"
+            HAS_CYEXTENSION = True
+else:
+    HAS_CYEXTENSION = False
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_py_collections.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_py_collections.py
new file mode 100644
index 00000000..f6aefcf6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/_py_collections.py
@@ -0,0 +1,541 @@
+# util/_py_collections.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+from itertools import filterfalse
+from typing import AbstractSet
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Collection
+from typing import Dict
+from typing import Iterable
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import Set
+from typing import Tuple
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from ..util.typing import Self
+
+_T = TypeVar("_T", bound=Any)
+_S = TypeVar("_S", bound=Any)
+_KT = TypeVar("_KT", bound=Any)
+_VT = TypeVar("_VT", bound=Any)
+
+
+class ReadOnlyContainer:
+    __slots__ = ()
+
+    def _readonly(self, *arg: Any, **kw: Any) -> NoReturn:
+        raise TypeError(
+            "%s object is immutable and/or readonly" % self.__class__.__name__
+        )
+
+    def _immutable(self, *arg: Any, **kw: Any) -> NoReturn:
+        raise TypeError("%s object is immutable" % self.__class__.__name__)
+
+    def __delitem__(self, key: Any) -> NoReturn:
+        self._readonly()
+
+    def __setitem__(self, key: Any, value: Any) -> NoReturn:
+        self._readonly()
+
+    def __setattr__(self, key: str, value: Any) -> NoReturn:
+        self._readonly()
+
+
+class ImmutableDictBase(ReadOnlyContainer, Dict[_KT, _VT]):
+    if TYPE_CHECKING:
+
+        def __new__(cls, *args: Any) -> Self: ...
+
+        def __init__(cls, *args: Any): ...
+
+    def _readonly(self, *arg: Any, **kw: Any) -> NoReturn:
+        self._immutable()
+
+    def clear(self) -> NoReturn:
+        self._readonly()
+
+    def pop(self, key: Any, default: Optional[Any] = None) -> NoReturn:
+        self._readonly()
+
+    def popitem(self) -> NoReturn:
+        self._readonly()
+
+    def setdefault(self, key: Any, default: Optional[Any] = None) -> NoReturn:
+        self._readonly()
+
+    def update(self, *arg: Any, **kw: Any) -> NoReturn:
+        self._readonly()
+
+
+class immutabledict(ImmutableDictBase[_KT, _VT]):
+    def __new__(cls, *args):
+        new = ImmutableDictBase.__new__(cls)
+        dict.__init__(new, *args)
+        return new
+
+    def __init__(
+        self, *args: Union[Mapping[_KT, _VT], Iterable[Tuple[_KT, _VT]]]
+    ):
+        pass
+
+    def __reduce__(self):
+        return immutabledict, (dict(self),)
+
+    def union(
+        self, __d: Optional[Mapping[_KT, _VT]] = None
+    ) -> immutabledict[_KT, _VT]:
+        if not __d:
+            return self
+
+        new = ImmutableDictBase.__new__(self.__class__)
+        dict.__init__(new, self)
+        dict.update(new, __d)  # type: ignore
+        return new
+
+    def _union_w_kw(
+        self, __d: Optional[Mapping[_KT, _VT]] = None, **kw: _VT
+    ) -> immutabledict[_KT, _VT]:
+        # not sure if C version works correctly w/ this yet
+        if not __d and not kw:
+            return self
+
+        new = ImmutableDictBase.__new__(self.__class__)
+        dict.__init__(new, self)
+        if __d:
+            dict.update(new, __d)  # type: ignore
+        dict.update(new, kw)  # type: ignore
+        return new
+
+    def merge_with(
+        self, *dicts: Optional[Mapping[_KT, _VT]]
+    ) -> immutabledict[_KT, _VT]:
+        new = None
+        for d in dicts:
+            if d:
+                if new is None:
+                    new = ImmutableDictBase.__new__(self.__class__)
+                    dict.__init__(new, self)
+                dict.update(new, d)  # type: ignore
+        if new is None:
+            return self
+
+        return new
+
+    def __repr__(self) -> str:
+        return "immutabledict(%s)" % dict.__repr__(self)
+
+    # PEP 584
+    def __ior__(self, __value: Any) -> NoReturn:  # type: ignore
+        self._readonly()
+
+    def __or__(  # type: ignore[override]
+        self, __value: Mapping[_KT, _VT]
+    ) -> immutabledict[_KT, _VT]:
+        return immutabledict(
+            super().__or__(__value),  # type: ignore[call-overload]
+        )
+
+    def __ror__(  # type: ignore[override]
+        self, __value: Mapping[_KT, _VT]
+    ) -> immutabledict[_KT, _VT]:
+        return immutabledict(
+            super().__ror__(__value),  # type: ignore[call-overload]
+        )
+
+
+class OrderedSet(Set[_T]):
+    __slots__ = ("_list",)
+
+    _list: List[_T]
+
+    def __init__(self, d: Optional[Iterable[_T]] = None) -> None:
+        if d is not None:
+            self._list = unique_list(d)
+            super().update(self._list)
+        else:
+            self._list = []
+
+    def copy(self) -> OrderedSet[_T]:
+        cp = self.__class__()
+        cp._list = self._list.copy()
+        set.update(cp, cp._list)
+        return cp
+
+    def add(self, element: _T) -> None:
+        if element not in self:
+            self._list.append(element)
+        super().add(element)
+
+    def remove(self, element: _T) -> None:
+        super().remove(element)
+        self._list.remove(element)
+
+    def pop(self) -> _T:
+        try:
+            value = self._list.pop()
+        except IndexError:
+            raise KeyError("pop from an empty set") from None
+        super().remove(value)
+        return value
+
+    def insert(self, pos: int, element: _T) -> None:
+        if element not in self:
+            self._list.insert(pos, element)
+        super().add(element)
+
+    def discard(self, element: _T) -> None:
+        if element in self:
+            self._list.remove(element)
+            super().remove(element)
+
+    def clear(self) -> None:
+        super().clear()
+        self._list = []
+
+    def __getitem__(self, key: int) -> _T:
+        return self._list[key]
+
+    def __iter__(self) -> Iterator[_T]:
+        return iter(self._list)
+
+    def __add__(self, other: Iterator[_T]) -> OrderedSet[_T]:
+        return self.union(other)
+
+    def __repr__(self) -> str:
+        return "%s(%r)" % (self.__class__.__name__, self._list)
+
+    __str__ = __repr__
+
+    def update(self, *iterables: Iterable[_T]) -> None:
+        for iterable in iterables:
+            for e in iterable:
+                if e not in self:
+                    self._list.append(e)
+                    super().add(e)
+
+    def __ior__(self, other: AbstractSet[_S]) -> OrderedSet[Union[_T, _S]]:
+        self.update(other)
+        return self
+
+    def union(self, *other: Iterable[_S]) -> OrderedSet[Union[_T, _S]]:
+        result: OrderedSet[Union[_T, _S]] = self.copy()
+        result.update(*other)
+        return result
+
+    def __or__(self, other: AbstractSet[_S]) -> OrderedSet[Union[_T, _S]]:
+        return self.union(other)
+
+    def intersection(self, *other: Iterable[Any]) -> OrderedSet[_T]:
+        other_set: Set[Any] = set()
+        other_set.update(*other)
+        return self.__class__(a for a in self if a in other_set)
+
+    def __and__(self, other: AbstractSet[object]) -> OrderedSet[_T]:
+        return self.intersection(other)
+
+    def symmetric_difference(self, other: Iterable[_T]) -> OrderedSet[_T]:
+        collection: Collection[_T]
+        if isinstance(other, set):
+            collection = other_set = other
+        elif isinstance(other, Collection):
+            collection = other
+            other_set = set(other)
+        else:
+            collection = list(other)
+            other_set = set(collection)
+        result = self.__class__(a for a in self if a not in other_set)
+        result.update(a for a in collection if a not in self)
+        return result
+
+    def __xor__(self, other: AbstractSet[_S]) -> OrderedSet[Union[_T, _S]]:
+        return cast(OrderedSet[Union[_T, _S]], self).symmetric_difference(
+            other
+        )
+
+    def difference(self, *other: Iterable[Any]) -> OrderedSet[_T]:
+        other_set = super().difference(*other)
+        return self.__class__(a for a in self._list if a in other_set)
+
+    def __sub__(self, other: AbstractSet[Optional[_T]]) -> OrderedSet[_T]:
+        return self.difference(other)
+
+    def intersection_update(self, *other: Iterable[Any]) -> None:
+        super().intersection_update(*other)
+        self._list = [a for a in self._list if a in self]
+
+    def __iand__(self, other: AbstractSet[object]) -> OrderedSet[_T]:
+        self.intersection_update(other)
+        return self
+
+    def symmetric_difference_update(self, other: Iterable[Any]) -> None:
+        collection = other if isinstance(other, Collection) else list(other)
+        super().symmetric_difference_update(collection)
+        self._list = [a for a in self._list if a in self]
+        self._list += [a for a in collection if a in self]
+
+    def __ixor__(self, other: AbstractSet[_S]) -> OrderedSet[Union[_T, _S]]:
+        self.symmetric_difference_update(other)
+        return cast(OrderedSet[Union[_T, _S]], self)
+
+    def difference_update(self, *other: Iterable[Any]) -> None:
+        super().difference_update(*other)
+        self._list = [a for a in self._list if a in self]
+
+    def __isub__(self, other: AbstractSet[Optional[_T]]) -> OrderedSet[_T]:  # type: ignore  # noqa: E501
+        self.difference_update(other)
+        return self
+
+
+class IdentitySet:
+    """A set that considers only object id() for uniqueness.
+
+    This strategy has edge cases for builtin types- it's possible to have
+    two 'foo' strings in one of these sets, for example.  Use sparingly.
+
+    """
+
+    _members: Dict[int, Any]
+
+    def __init__(self, iterable: Optional[Iterable[Any]] = None):
+        self._members = dict()
+        if iterable:
+            self.update(iterable)
+
+    def add(self, value: Any) -> None:
+        self._members[id(value)] = value
+
+    def __contains__(self, value: Any) -> bool:
+        return id(value) in self._members
+
+    def remove(self, value: Any) -> None:
+        del self._members[id(value)]
+
+    def discard(self, value: Any) -> None:
+        try:
+            self.remove(value)
+        except KeyError:
+            pass
+
+    def pop(self) -> Any:
+        try:
+            pair = self._members.popitem()
+            return pair[1]
+        except KeyError:
+            raise KeyError("pop from an empty set")
+
+    def clear(self) -> None:
+        self._members.clear()
+
+    def __eq__(self, other: Any) -> bool:
+        if isinstance(other, IdentitySet):
+            return self._members == other._members
+        else:
+            return False
+
+    def __ne__(self, other: Any) -> bool:
+        if isinstance(other, IdentitySet):
+            return self._members != other._members
+        else:
+            return True
+
+    def issubset(self, iterable: Iterable[Any]) -> bool:
+        if isinstance(iterable, self.__class__):
+            other = iterable
+        else:
+            other = self.__class__(iterable)
+
+        if len(self) > len(other):
+            return False
+        for m in filterfalse(
+            other._members.__contains__, iter(self._members.keys())
+        ):
+            return False
+        return True
+
+    def __le__(self, other: Any) -> bool:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.issubset(other)
+
+    def __lt__(self, other: Any) -> bool:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return len(self) < len(other) and self.issubset(other)
+
+    def issuperset(self, iterable: Iterable[Any]) -> bool:
+        if isinstance(iterable, self.__class__):
+            other = iterable
+        else:
+            other = self.__class__(iterable)
+
+        if len(self) < len(other):
+            return False
+
+        for m in filterfalse(
+            self._members.__contains__, iter(other._members.keys())
+        ):
+            return False
+        return True
+
+    def __ge__(self, other: Any) -> bool:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.issuperset(other)
+
+    def __gt__(self, other: Any) -> bool:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return len(self) > len(other) and self.issuperset(other)
+
+    def union(self, iterable: Iterable[Any]) -> IdentitySet:
+        result = self.__class__()
+        members = self._members
+        result._members.update(members)
+        result._members.update((id(obj), obj) for obj in iterable)
+        return result
+
+    def __or__(self, other: Any) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.union(other)
+
+    def update(self, iterable: Iterable[Any]) -> None:
+        self._members.update((id(obj), obj) for obj in iterable)
+
+    def __ior__(self, other: Any) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.update(other)
+        return self
+
+    def difference(self, iterable: Iterable[Any]) -> IdentitySet:
+        result = self.__new__(self.__class__)
+        other: Collection[Any]
+
+        if isinstance(iterable, self.__class__):
+            other = iterable._members
+        else:
+            other = {id(obj) for obj in iterable}
+        result._members = {
+            k: v for k, v in self._members.items() if k not in other
+        }
+        return result
+
+    def __sub__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.difference(other)
+
+    def difference_update(self, iterable: Iterable[Any]) -> None:
+        self._members = self.difference(iterable)._members
+
+    def __isub__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.difference_update(other)
+        return self
+
+    def intersection(self, iterable: Iterable[Any]) -> IdentitySet:
+        result = self.__new__(self.__class__)
+
+        other: Collection[Any]
+
+        if isinstance(iterable, self.__class__):
+            other = iterable._members
+        else:
+            other = {id(obj) for obj in iterable}
+        result._members = {
+            k: v for k, v in self._members.items() if k in other
+        }
+        return result
+
+    def __and__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.intersection(other)
+
+    def intersection_update(self, iterable: Iterable[Any]) -> None:
+        self._members = self.intersection(iterable)._members
+
+    def __iand__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.intersection_update(other)
+        return self
+
+    def symmetric_difference(self, iterable: Iterable[Any]) -> IdentitySet:
+        result = self.__new__(self.__class__)
+        if isinstance(iterable, self.__class__):
+            other = iterable._members
+        else:
+            other = {id(obj): obj for obj in iterable}
+        result._members = {
+            k: v for k, v in self._members.items() if k not in other
+        }
+        result._members.update(
+            (k, v) for k, v in other.items() if k not in self._members
+        )
+        return result
+
+    def __xor__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        return self.symmetric_difference(other)
+
+    def symmetric_difference_update(self, iterable: Iterable[Any]) -> None:
+        self._members = self.symmetric_difference(iterable)._members
+
+    def __ixor__(self, other: IdentitySet) -> IdentitySet:
+        if not isinstance(other, IdentitySet):
+            return NotImplemented
+        self.symmetric_difference(other)
+        return self
+
+    def copy(self) -> IdentitySet:
+        result = self.__new__(self.__class__)
+        result._members = self._members.copy()
+        return result
+
+    __copy__ = copy
+
+    def __len__(self) -> int:
+        return len(self._members)
+
+    def __iter__(self) -> Iterator[Any]:
+        return iter(self._members.values())
+
+    def __hash__(self) -> NoReturn:
+        raise TypeError("set objects are unhashable")
+
+    def __repr__(self) -> str:
+        return "%s(%r)" % (type(self).__name__, list(self._members.values()))
+
+
+def unique_list(
+    seq: Iterable[_T], hashfunc: Optional[Callable[[_T], int]] = None
+) -> List[_T]:
+    seen: Set[Any] = set()
+    seen_add = seen.add
+    if not hashfunc:
+        return [x for x in seq if x not in seen and not seen_add(x)]
+    else:
+        return [
+            x
+            for x in seq
+            if hashfunc(x) not in seen and not seen_add(hashfunc(x))
+        ]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/compat.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/compat.py
new file mode 100644
index 00000000..c8b5e7a2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/compat.py
@@ -0,0 +1,301 @@
+# util/compat.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Handle Python version/platform incompatibilities."""
+
+from __future__ import annotations
+
+import base64
+import dataclasses
+import hashlib
+import inspect
+import operator
+import platform
+import sys
+import typing
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Iterable
+from typing import List
+from typing import Mapping
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+
+
+py313 = sys.version_info >= (3, 13)
+py312 = sys.version_info >= (3, 12)
+py311 = sys.version_info >= (3, 11)
+py310 = sys.version_info >= (3, 10)
+py39 = sys.version_info >= (3, 9)
+py38 = sys.version_info >= (3, 8)
+pypy = platform.python_implementation() == "PyPy"
+cpython = platform.python_implementation() == "CPython"
+
+win32 = sys.platform.startswith("win")
+osx = sys.platform.startswith("darwin")
+arm = "aarch" in platform.machine().lower()
+is64bit = sys.maxsize > 2**32
+
+has_refcount_gc = bool(cpython)
+
+dottedgetter = operator.attrgetter
+
+_T_co = TypeVar("_T_co", covariant=True)
+
+
+class FullArgSpec(typing.NamedTuple):
+    args: List[str]
+    varargs: Optional[str]
+    varkw: Optional[str]
+    defaults: Optional[Tuple[Any, ...]]
+    kwonlyargs: List[str]
+    kwonlydefaults: Optional[Dict[str, Any]]
+    annotations: Dict[str, Any]
+
+
+def inspect_getfullargspec(func: Callable[..., Any]) -> FullArgSpec:
+    """Fully vendored version of getfullargspec from Python 3.3."""
+
+    if inspect.ismethod(func):
+        func = func.__func__
+    if not inspect.isfunction(func):
+        raise TypeError(f"{func!r} is not a Python function")
+
+    co = func.__code__
+    if not inspect.iscode(co):
+        raise TypeError(f"{co!r} is not a code object")
+
+    nargs = co.co_argcount
+    names = co.co_varnames
+    nkwargs = co.co_kwonlyargcount
+    args = list(names[:nargs])
+    kwonlyargs = list(names[nargs : nargs + nkwargs])
+
+    nargs += nkwargs
+    varargs = None
+    if co.co_flags & inspect.CO_VARARGS:
+        varargs = co.co_varnames[nargs]
+        nargs = nargs + 1
+    varkw = None
+    if co.co_flags & inspect.CO_VARKEYWORDS:
+        varkw = co.co_varnames[nargs]
+
+    return FullArgSpec(
+        args,
+        varargs,
+        varkw,
+        func.__defaults__,
+        kwonlyargs,
+        func.__kwdefaults__,
+        func.__annotations__,
+    )
+
+
+if py39:
+    # python stubs don't have a public type for this. not worth
+    # making a protocol
+    def md5_not_for_security() -> Any:
+        return hashlib.md5(usedforsecurity=False)
+
+else:
+
+    def md5_not_for_security() -> Any:
+        return hashlib.md5()
+
+
+if typing.TYPE_CHECKING or py38:
+    from importlib import metadata as importlib_metadata
+else:
+    import importlib_metadata  # noqa
+
+
+if typing.TYPE_CHECKING or py39:
+    # pep 584 dict union
+    dict_union = operator.or_  # noqa
+else:
+
+    def dict_union(a: dict, b: dict) -> dict:
+        a = a.copy()
+        a.update(b)
+        return a
+
+
+if py310:
+    anext_ = anext
+else:
+    _NOT_PROVIDED = object()
+    from collections.abc import AsyncIterator
+
+    async def anext_(async_iterator, default=_NOT_PROVIDED):
+        """vendored from https://github.com/python/cpython/pull/8895"""
+
+        if not isinstance(async_iterator, AsyncIterator):
+            raise TypeError(
+                f"anext expected an AsyncIterator, got {type(async_iterator)}"
+            )
+        anxt = type(async_iterator).__anext__
+        try:
+            return await anxt(async_iterator)
+        except StopAsyncIteration:
+            if default is _NOT_PROVIDED:
+                raise
+            return default
+
+
+def importlib_metadata_get(group):
+    ep = importlib_metadata.entry_points()
+    if typing.TYPE_CHECKING or hasattr(ep, "select"):
+        return ep.select(group=group)
+    else:
+        return ep.get(group, ())
+
+
+def b(s):
+    return s.encode("latin-1")
+
+
+def b64decode(x: str) -> bytes:
+    return base64.b64decode(x.encode("ascii"))
+
+
+def b64encode(x: bytes) -> str:
+    return base64.b64encode(x).decode("ascii")
+
+
+def decode_backslashreplace(text: bytes, encoding: str) -> str:
+    return text.decode(encoding, errors="backslashreplace")
+
+
+def cmp(a, b):
+    return (a > b) - (a < b)
+
+
+def _formatannotation(annotation, base_module=None):
+    """vendored from python 3.7"""
+
+    if isinstance(annotation, str):
+        return annotation
+
+    if getattr(annotation, "__module__", None) == "typing":
+        return repr(annotation).replace("typing.", "").replace("~", "")
+    if isinstance(annotation, type):
+        if annotation.__module__ in ("builtins", base_module):
+            return repr(annotation.__qualname__)
+        return annotation.__module__ + "." + annotation.__qualname__
+    elif isinstance(annotation, typing.TypeVar):
+        return repr(annotation).replace("~", "")
+    return repr(annotation).replace("~", "")
+
+
+def inspect_formatargspec(
+    args: List[str],
+    varargs: Optional[str] = None,
+    varkw: Optional[str] = None,
+    defaults: Optional[Sequence[Any]] = None,
+    kwonlyargs: Optional[Sequence[str]] = (),
+    kwonlydefaults: Optional[Mapping[str, Any]] = {},
+    annotations: Mapping[str, Any] = {},
+    formatarg: Callable[[str], str] = str,
+    formatvarargs: Callable[[str], str] = lambda name: "*" + name,
+    formatvarkw: Callable[[str], str] = lambda name: "**" + name,
+    formatvalue: Callable[[Any], str] = lambda value: "=" + repr(value),
+    formatreturns: Callable[[Any], str] = lambda text: " -> " + str(text),
+    formatannotation: Callable[[Any], str] = _formatannotation,
+) -> str:
+    """Copy formatargspec from python 3.7 standard library.
+
+    Python 3 has deprecated formatargspec and requested that Signature
+    be used instead, however this requires a full reimplementation
+    of formatargspec() in terms of creating Parameter objects and such.
+    Instead of introducing all the object-creation overhead and having
+    to reinvent from scratch, just copy their compatibility routine.
+
+    Ultimately we would need to rewrite our "decorator" routine completely
+    which is not really worth it right now, until all Python 2.x support
+    is dropped.
+
+    """
+
+    kwonlydefaults = kwonlydefaults or {}
+    annotations = annotations or {}
+
+    def formatargandannotation(arg):
+        result = formatarg(arg)
+        if arg in annotations:
+            result += ": " + formatannotation(annotations[arg])
+        return result
+
+    specs = []
+    if defaults:
+        firstdefault = len(args) - len(defaults)
+    else:
+        firstdefault = -1
+
+    for i, arg in enumerate(args):
+        spec = formatargandannotation(arg)
+        if defaults and i >= firstdefault:
+            spec = spec + formatvalue(defaults[i - firstdefault])
+        specs.append(spec)
+
+    if varargs is not None:
+        specs.append(formatvarargs(formatargandannotation(varargs)))
+    else:
+        if kwonlyargs:
+            specs.append("*")
+
+    if kwonlyargs:
+        for kwonlyarg in kwonlyargs:
+            spec = formatargandannotation(kwonlyarg)
+            if kwonlydefaults and kwonlyarg in kwonlydefaults:
+                spec += formatvalue(kwonlydefaults[kwonlyarg])
+            specs.append(spec)
+
+    if varkw is not None:
+        specs.append(formatvarkw(formatargandannotation(varkw)))
+
+    result = "(" + ", ".join(specs) + ")"
+    if "return" in annotations:
+        result += formatreturns(formatannotation(annotations["return"]))
+    return result
+
+
+def dataclass_fields(cls: Type[Any]) -> Iterable[dataclasses.Field[Any]]:
+    """Return a sequence of all dataclasses.Field objects associated
+    with a class as an already processed dataclass.
+
+    The class must **already be a dataclass** for Field objects to be returned.
+
+    """
+
+    if dataclasses.is_dataclass(cls):
+        return dataclasses.fields(cls)
+    else:
+        return []
+
+
+def local_dataclass_fields(cls: Type[Any]) -> Iterable[dataclasses.Field[Any]]:
+    """Return a sequence of all dataclasses.Field objects associated with
+    an already processed dataclass, excluding those that originate from a
+    superclass.
+
+    The class must **already be a dataclass** for Field objects to be returned.
+
+    """
+
+    if dataclasses.is_dataclass(cls):
+        super_fields: Set[dataclasses.Field[Any]] = set()
+        for sup in cls.__bases__:
+            super_fields.update(dataclass_fields(sup))
+        return [f for f in dataclasses.fields(cls) if f not in super_fields]
+    else:
+        return []
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/concurrency.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/concurrency.py
new file mode 100644
index 00000000..006340f5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/concurrency.py
@@ -0,0 +1,108 @@
+# util/concurrency.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import asyncio  # noqa
+import typing
+from typing import Any
+from typing import Callable
+from typing import Coroutine
+from typing import TypeVar
+
+have_greenlet = False
+greenlet_error = None
+try:
+    import greenlet  # type: ignore[import-untyped,unused-ignore]  # noqa: F401,E501
+except ImportError as e:
+    greenlet_error = str(e)
+    pass
+else:
+    have_greenlet = True
+    from ._concurrency_py3k import await_only as await_only
+    from ._concurrency_py3k import await_fallback as await_fallback
+    from ._concurrency_py3k import in_greenlet as in_greenlet
+    from ._concurrency_py3k import greenlet_spawn as greenlet_spawn
+    from ._concurrency_py3k import is_exit_exception as is_exit_exception
+    from ._concurrency_py3k import AsyncAdaptedLock as AsyncAdaptedLock
+    from ._concurrency_py3k import _Runner
+
+_T = TypeVar("_T")
+
+
+class _AsyncUtil:
+    """Asyncio util for test suite/ util only"""
+
+    def __init__(self) -> None:
+        if have_greenlet:
+            self.runner = _Runner()
+
+    def run(
+        self,
+        fn: Callable[..., Coroutine[Any, Any, _T]],
+        *args: Any,
+        **kwargs: Any,
+    ) -> _T:
+        """Run coroutine on the loop"""
+        return self.runner.run(fn(*args, **kwargs))
+
+    def run_in_greenlet(
+        self, fn: Callable[..., _T], *args: Any, **kwargs: Any
+    ) -> _T:
+        """Run sync function in greenlet. Support nested calls"""
+        if have_greenlet:
+            if self.runner.get_loop().is_running():
+                return fn(*args, **kwargs)
+            else:
+                return self.runner.run(greenlet_spawn(fn, *args, **kwargs))
+        else:
+            return fn(*args, **kwargs)
+
+    def close(self) -> None:
+        if have_greenlet:
+            self.runner.close()
+
+
+if not typing.TYPE_CHECKING and not have_greenlet:
+
+    def _not_implemented():
+        # this conditional is to prevent pylance from considering
+        # greenlet_spawn() etc as "no return" and dimming out code below it
+        if have_greenlet:
+            return None
+
+        raise ValueError(
+            "the greenlet library is required to use this function."
+            " %s" % greenlet_error
+            if greenlet_error
+            else ""
+        )
+
+    def is_exit_exception(e):  # noqa: F811
+        return not isinstance(e, Exception)
+
+    def await_only(thing):  # type: ignore  # noqa: F811
+        _not_implemented()
+
+    def await_fallback(thing):  # type: ignore  # noqa: F811
+        return thing
+
+    def in_greenlet():  # type: ignore  # noqa: F811
+        _not_implemented()
+
+    def greenlet_spawn(fn, *args, **kw):  # type: ignore  # noqa: F811
+        _not_implemented()
+
+    def AsyncAdaptedLock(*args, **kw):  # type: ignore  # noqa: F811
+        _not_implemented()
+
+    def _util_async_run(fn, *arg, **kw):  # type: ignore  # noqa: F811
+        return fn(*arg, **kw)
+
+    def _util_async_run_coroutine_function(fn, *arg, **kw):  # type: ignore  # noqa: F811,E501
+        _not_implemented()
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/deprecations.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/deprecations.py
new file mode 100644
index 00000000..88b68724
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/deprecations.py
@@ -0,0 +1,401 @@
+# util/deprecations.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Helpers related to deprecation of functions, methods, classes, other
+functionality."""
+
+from __future__ import annotations
+
+import re
+from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Match
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TypeVar
+from typing import Union
+
+from . import compat
+from .langhelpers import _hash_limit_string
+from .langhelpers import _warnings_warn
+from .langhelpers import decorator
+from .langhelpers import inject_docstring_text
+from .langhelpers import inject_param_text
+from .. import exc
+
+_T = TypeVar("_T", bound=Any)
+
+
+# https://mypy.readthedocs.io/en/stable/generics.html#declaring-decorators
+_F = TypeVar("_F", bound="Callable[..., Any]")
+
+
+def _warn_with_version(
+    msg: str,
+    version: str,
+    type_: Type[exc.SADeprecationWarning],
+    stacklevel: int,
+    code: Optional[str] = None,
+) -> None:
+    warn = type_(msg, code=code)
+    warn.deprecated_since = version
+
+    _warnings_warn(warn, stacklevel=stacklevel + 1)
+
+
+def warn_deprecated(
+    msg: str, version: str, stacklevel: int = 3, code: Optional[str] = None
+) -> None:
+    _warn_with_version(
+        msg, version, exc.SADeprecationWarning, stacklevel, code=code
+    )
+
+
+def warn_deprecated_limited(
+    msg: str,
+    args: Sequence[Any],
+    version: str,
+    stacklevel: int = 3,
+    code: Optional[str] = None,
+) -> None:
+    """Issue a deprecation warning with a parameterized string,
+    limiting the number of registrations.
+
+    """
+    if args:
+        msg = _hash_limit_string(msg, 10, args)
+    _warn_with_version(
+        msg, version, exc.SADeprecationWarning, stacklevel, code=code
+    )
+
+
+def deprecated_cls(
+    version: str, message: str, constructor: Optional[str] = "__init__"
+) -> Callable[[Type[_T]], Type[_T]]:
+    header = ".. deprecated:: %s %s" % (version, (message or ""))
+
+    def decorate(cls: Type[_T]) -> Type[_T]:
+        return _decorate_cls_with_warning(
+            cls,
+            constructor,
+            exc.SADeprecationWarning,
+            message % dict(func=constructor),
+            version,
+            header,
+        )
+
+    return decorate
+
+
+def deprecated(
+    version: str,
+    message: Optional[str] = None,
+    add_deprecation_to_docstring: bool = True,
+    warning: Optional[Type[exc.SADeprecationWarning]] = None,
+    enable_warnings: bool = True,
+) -> Callable[[_F], _F]:
+    """Decorates a function and issues a deprecation warning on use.
+
+    :param version:
+      Issue version in the warning.
+
+    :param message:
+      If provided, issue message in the warning.  A sensible default
+      is used if not provided.
+
+    :param add_deprecation_to_docstring:
+      Default True.  If False, the wrapped function's __doc__ is left
+      as-is.  If True, the 'message' is prepended to the docs if
+      provided, or sensible default if message is omitted.
+
+    """
+
+    if add_deprecation_to_docstring:
+        header = ".. deprecated:: %s %s" % (
+            version,
+            (message or ""),
+        )
+    else:
+        header = None
+
+    if message is None:
+        message = "Call to deprecated function %(func)s"
+
+    if warning is None:
+        warning = exc.SADeprecationWarning
+
+    message += " (deprecated since: %s)" % version
+
+    def decorate(fn: _F) -> _F:
+        assert message is not None
+        assert warning is not None
+        return _decorate_with_warning(
+            fn,
+            warning,
+            message % dict(func=fn.__name__),
+            version,
+            header,
+            enable_warnings=enable_warnings,
+        )
+
+    return decorate
+
+
+def moved_20(
+    message: str, **kw: Any
+) -> Callable[[Callable[..., _T]], Callable[..., _T]]:
+    return deprecated(
+        "2.0", message=message, warning=exc.MovedIn20Warning, **kw
+    )
+
+
+def became_legacy_20(
+    api_name: str, alternative: Optional[str] = None, **kw: Any
+) -> Callable[[_F], _F]:
+    type_reg = re.match("^:(attr|func|meth):", api_name)
+    if type_reg:
+        type_ = {"attr": "attribute", "func": "function", "meth": "method"}[
+            type_reg.group(1)
+        ]
+    else:
+        type_ = "construct"
+    message = (
+        "The %s %s is considered legacy as of the "
+        "1.x series of SQLAlchemy and %s in 2.0."
+        % (
+            api_name,
+            type_,
+            "becomes a legacy construct",
+        )
+    )
+
+    if ":attr:" in api_name:
+        attribute_ok = kw.pop("warn_on_attribute_access", False)
+        if not attribute_ok:
+            assert kw.get("enable_warnings") is False, (
+                "attribute %s will emit a warning on read access.  "
+                "If you *really* want this, "
+                "add warn_on_attribute_access=True.  Otherwise please add "
+                "enable_warnings=False." % api_name
+            )
+
+    if alternative:
+        message += " " + alternative
+
+    warning_cls = exc.LegacyAPIWarning
+
+    return deprecated("2.0", message=message, warning=warning_cls, **kw)
+
+
+def deprecated_params(**specs: Tuple[str, str]) -> Callable[[_F], _F]:
+    """Decorates a function to warn on use of certain parameters.
+
+    e.g. ::
+
+        @deprecated_params(
+            weak_identity_map=(
+                "0.7",
+                "the :paramref:`.Session.weak_identity_map parameter "
+                "is deprecated.",
+            )
+        )
+        def some_function(**kwargs): ...
+
+    """
+
+    messages: Dict[str, str] = {}
+    versions: Dict[str, str] = {}
+    version_warnings: Dict[str, Type[exc.SADeprecationWarning]] = {}
+
+    for param, (version, message) in specs.items():
+        versions[param] = version
+        messages[param] = _sanitize_restructured_text(message)
+        version_warnings[param] = exc.SADeprecationWarning
+
+    def decorate(fn: _F) -> _F:
+        spec = compat.inspect_getfullargspec(fn)
+
+        check_defaults: Union[Set[str], Tuple[()]]
+        if spec.defaults is not None:
+            defaults = dict(
+                zip(
+                    spec.args[(len(spec.args) - len(spec.defaults)) :],
+                    spec.defaults,
+                )
+            )
+            check_defaults = set(defaults).intersection(messages)
+            check_kw = set(messages).difference(defaults)
+        elif spec.kwonlydefaults is not None:
+            defaults = spec.kwonlydefaults
+            check_defaults = set(defaults).intersection(messages)
+            check_kw = set(messages).difference(defaults)
+        else:
+            check_defaults = ()
+            check_kw = set(messages)
+
+        check_any_kw = spec.varkw
+
+        # latest mypy has opinions here, not sure if they implemented
+        # Concatenate or something
+        @decorator
+        def warned(fn: _F, *args: Any, **kwargs: Any) -> _F:
+            for m in check_defaults:
+                if (defaults[m] is None and kwargs[m] is not None) or (
+                    defaults[m] is not None and kwargs[m] != defaults[m]
+                ):
+                    _warn_with_version(
+                        messages[m],
+                        versions[m],
+                        version_warnings[m],
+                        stacklevel=3,
+                    )
+
+            if check_any_kw in messages and set(kwargs).difference(
+                check_defaults
+            ):
+                assert check_any_kw is not None
+                _warn_with_version(
+                    messages[check_any_kw],
+                    versions[check_any_kw],
+                    version_warnings[check_any_kw],
+                    stacklevel=3,
+                )
+
+            for m in check_kw:
+                if m in kwargs:
+                    _warn_with_version(
+                        messages[m],
+                        versions[m],
+                        version_warnings[m],
+                        stacklevel=3,
+                    )
+            return fn(*args, **kwargs)  # type: ignore[no-any-return]
+
+        doc = fn.__doc__ is not None and fn.__doc__ or ""
+        if doc:
+            doc = inject_param_text(
+                doc,
+                {
+                    param: ".. deprecated:: %s %s"
+                    % ("1.4" if version == "2.0" else version, (message or ""))
+                    for param, (version, message) in specs.items()
+                },
+            )
+        decorated = warned(fn)
+        decorated.__doc__ = doc
+        return decorated
+
+    return decorate
+
+
+def _sanitize_restructured_text(text: str) -> str:
+    def repl(m: Match[str]) -> str:
+        type_, name = m.group(1, 2)
+        if type_ in ("func", "meth"):
+            name += "()"
+        return name
+
+    text = re.sub(r":ref:`(.+) <.*>`", lambda m: '"%s"' % m.group(1), text)
+    return re.sub(r"\:(\w+)\:`~?(?:_\w+)?\.?(.+?)`", repl, text)
+
+
+def _decorate_cls_with_warning(
+    cls: Type[_T],
+    constructor: Optional[str],
+    wtype: Type[exc.SADeprecationWarning],
+    message: str,
+    version: str,
+    docstring_header: Optional[str] = None,
+) -> Type[_T]:
+    doc = cls.__doc__ is not None and cls.__doc__ or ""
+    if docstring_header is not None:
+        if constructor is not None:
+            docstring_header %= dict(func=constructor)
+
+        if issubclass(wtype, exc.Base20DeprecationWarning):
+            docstring_header += (
+                " (Background on SQLAlchemy 2.0 at: "
+                ":ref:`migration_20_toplevel`)"
+            )
+        doc = inject_docstring_text(doc, docstring_header, 1)
+
+        constructor_fn = None
+        if type(cls) is type:
+            clsdict = dict(cls.__dict__)
+            clsdict["__doc__"] = doc
+            clsdict.pop("__dict__", None)
+            clsdict.pop("__weakref__", None)
+            cls = type(cls.__name__, cls.__bases__, clsdict)
+            if constructor is not None:
+                constructor_fn = clsdict[constructor]
+
+        else:
+            cls.__doc__ = doc
+            if constructor is not None:
+                constructor_fn = getattr(cls, constructor)
+
+        if constructor is not None:
+            assert constructor_fn is not None
+            assert wtype is not None
+            setattr(
+                cls,
+                constructor,
+                _decorate_with_warning(
+                    constructor_fn, wtype, message, version, None
+                ),
+            )
+    return cls
+
+
+def _decorate_with_warning(
+    func: _F,
+    wtype: Type[exc.SADeprecationWarning],
+    message: str,
+    version: str,
+    docstring_header: Optional[str] = None,
+    enable_warnings: bool = True,
+) -> _F:
+    """Wrap a function with a warnings.warn and augmented docstring."""
+
+    message = _sanitize_restructured_text(message)
+
+    if issubclass(wtype, exc.Base20DeprecationWarning):
+        doc_only = (
+            " (Background on SQLAlchemy 2.0 at: "
+            ":ref:`migration_20_toplevel`)"
+        )
+    else:
+        doc_only = ""
+
+    @decorator
+    def warned(fn: _F, *args: Any, **kwargs: Any) -> _F:
+        skip_warning = not enable_warnings or kwargs.pop(
+            "_sa_skip_warning", False
+        )
+        if not skip_warning:
+            _warn_with_version(message, version, wtype, stacklevel=3)
+        return fn(*args, **kwargs)  # type: ignore[no-any-return]
+
+    doc = func.__doc__ is not None and func.__doc__ or ""
+    if docstring_header is not None:
+        docstring_header %= dict(func=func.__name__)
+
+        docstring_header += doc_only
+
+        doc = inject_docstring_text(doc, docstring_header, 1)
+
+    decorated = warned(func)
+    decorated.__doc__ = doc
+    decorated._sa_warn = lambda: _warn_with_version(  # type: ignore
+        message, version, wtype, stacklevel=3
+    )
+    return decorated
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/langhelpers.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/langhelpers.py
new file mode 100644
index 00000000..b4086f1d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/langhelpers.py
@@ -0,0 +1,2222 @@
+# util/langhelpers.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""Routines to help with the creation, loading and introspection of
+modules, classes, hierarchies, attributes, functions, and methods.
+
+"""
+from __future__ import annotations
+
+import collections
+import enum
+from functools import update_wrapper
+import inspect
+import itertools
+import operator
+import re
+import sys
+import textwrap
+import threading
+import types
+from types import CodeType
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import FrozenSet
+from typing import Generic
+from typing import Iterator
+from typing import List
+from typing import Mapping
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+import warnings
+
+from . import _collections
+from . import compat
+from ._has_cy import HAS_CYEXTENSION
+from .typing import Literal
+from .. import exc
+
+_T = TypeVar("_T")
+_T_co = TypeVar("_T_co", covariant=True)
+_F = TypeVar("_F", bound=Callable[..., Any])
+_MP = TypeVar("_MP", bound="memoized_property[Any]")
+_MA = TypeVar("_MA", bound="HasMemoized.memoized_attribute[Any]")
+_HP = TypeVar("_HP", bound="hybridproperty[Any]")
+_HM = TypeVar("_HM", bound="hybridmethod[Any]")
+
+
+if compat.py310:
+
+    def get_annotations(obj: Any) -> Mapping[str, Any]:
+        return inspect.get_annotations(obj)
+
+else:
+
+    def get_annotations(obj: Any) -> Mapping[str, Any]:
+        # it's been observed that cls.__annotations__ can be non present.
+        # it's not clear what causes this, running under tox py37/38 it
+        # happens, running straight pytest it doesnt
+
+        # https://docs.python.org/3/howto/annotations.html#annotations-howto
+        if isinstance(obj, type):
+            ann = obj.__dict__.get("__annotations__", None)
+        else:
+            ann = getattr(obj, "__annotations__", None)
+
+        if ann is None:
+            return _collections.EMPTY_DICT
+        else:
+            return cast("Mapping[str, Any]", ann)
+
+
+def md5_hex(x: Any) -> str:
+    x = x.encode("utf-8")
+    m = compat.md5_not_for_security()
+    m.update(x)
+    return cast(str, m.hexdigest())
+
+
+class safe_reraise:
+    """Reraise an exception after invoking some
+    handler code.
+
+    Stores the existing exception info before
+    invoking so that it is maintained across a potential
+    coroutine context switch.
+
+    e.g.::
+
+        try:
+            sess.commit()
+        except:
+            with safe_reraise():
+                sess.rollback()
+
+    TODO: we should at some point evaluate current behaviors in this regard
+    based on current greenlet, gevent/eventlet implementations in Python 3, and
+    also see the degree to which our own asyncio (based on greenlet also) is
+    impacted by this. .rollback() will cause IO / context switch to occur in
+    all these scenarios; what happens to the exception context from an
+    "except:" block if we don't explicitly store it? Original issue was #2703.
+
+    """
+
+    __slots__ = ("_exc_info",)
+
+    _exc_info: Union[
+        None,
+        Tuple[
+            Type[BaseException],
+            BaseException,
+            types.TracebackType,
+        ],
+        Tuple[None, None, None],
+    ]
+
+    def __enter__(self) -> None:
+        self._exc_info = sys.exc_info()
+
+    def __exit__(
+        self,
+        type_: Optional[Type[BaseException]],
+        value: Optional[BaseException],
+        traceback: Optional[types.TracebackType],
+    ) -> NoReturn:
+        assert self._exc_info is not None
+        # see #2703 for notes
+        if type_ is None:
+            exc_type, exc_value, exc_tb = self._exc_info
+            assert exc_value is not None
+            self._exc_info = None  # remove potential circular references
+            raise exc_value.with_traceback(exc_tb)
+        else:
+            self._exc_info = None  # remove potential circular references
+            assert value is not None
+            raise value.with_traceback(traceback)
+
+
+def walk_subclasses(cls: Type[_T]) -> Iterator[Type[_T]]:
+    seen: Set[Any] = set()
+
+    stack = [cls]
+    while stack:
+        cls = stack.pop()
+        if cls in seen:
+            continue
+        else:
+            seen.add(cls)
+        stack.extend(cls.__subclasses__())
+        yield cls
+
+
+def string_or_unprintable(element: Any) -> str:
+    if isinstance(element, str):
+        return element
+    else:
+        try:
+            return str(element)
+        except Exception:
+            return "unprintable element %r" % element
+
+
+def clsname_as_plain_name(
+    cls: Type[Any], use_name: Optional[str] = None
+) -> str:
+    name = use_name or cls.__name__
+    return " ".join(n.lower() for n in re.findall(r"([A-Z][a-z]+|SQL)", name))
+
+
+def method_is_overridden(
+    instance_or_cls: Union[Type[Any], object],
+    against_method: Callable[..., Any],
+) -> bool:
+    """Return True if the two class methods don't match."""
+
+    if not isinstance(instance_or_cls, type):
+        current_cls = instance_or_cls.__class__
+    else:
+        current_cls = instance_or_cls
+
+    method_name = against_method.__name__
+
+    current_method: types.MethodType = getattr(current_cls, method_name)
+
+    return current_method != against_method
+
+
+def decode_slice(slc: slice) -> Tuple[Any, ...]:
+    """decode a slice object as sent to __getitem__.
+
+    takes into account the 2.5 __index__() method, basically.
+
+    """
+    ret: List[Any] = []
+    for x in slc.start, slc.stop, slc.step:
+        if hasattr(x, "__index__"):
+            x = x.__index__()
+        ret.append(x)
+    return tuple(ret)
+
+
+def _unique_symbols(used: Sequence[str], *bases: str) -> Iterator[str]:
+    used_set = set(used)
+    for base in bases:
+        pool = itertools.chain(
+            (base,),
+            map(lambda i: base + str(i), range(1000)),
+        )
+        for sym in pool:
+            if sym not in used_set:
+                used_set.add(sym)
+                yield sym
+                break
+        else:
+            raise NameError("exhausted namespace for symbol base %s" % base)
+
+
+def map_bits(fn: Callable[[int], Any], n: int) -> Iterator[Any]:
+    """Call the given function given each nonzero bit from n."""
+
+    while n:
+        b = n & (~n + 1)
+        yield fn(b)
+        n ^= b
+
+
+_Fn = TypeVar("_Fn", bound="Callable[..., Any]")
+
+# this seems to be in flux in recent mypy versions
+
+
+def decorator(target: Callable[..., Any]) -> Callable[[_Fn], _Fn]:
+    """A signature-matching decorator factory."""
+
+    def decorate(fn: _Fn) -> _Fn:
+        if not inspect.isfunction(fn) and not inspect.ismethod(fn):
+            raise Exception("not a decoratable function")
+
+        spec = compat.inspect_getfullargspec(fn)
+        env: Dict[str, Any] = {}
+
+        spec = _update_argspec_defaults_into_env(spec, env)
+
+        names = (
+            tuple(cast("Tuple[str, ...]", spec[0]))
+            + cast("Tuple[str, ...]", spec[1:3])
+            + (fn.__name__,)
+        )
+        targ_name, fn_name = _unique_symbols(names, "target", "fn")
+
+        metadata: Dict[str, Optional[str]] = dict(target=targ_name, fn=fn_name)
+        metadata.update(format_argspec_plus(spec, grouped=False))
+        metadata["name"] = fn.__name__
+
+        if inspect.iscoroutinefunction(fn):
+            metadata["prefix"] = "async "
+            metadata["target_prefix"] = "await "
+        else:
+            metadata["prefix"] = ""
+            metadata["target_prefix"] = ""
+
+        # look for __ positional arguments.  This is a convention in
+        # SQLAlchemy that arguments should be passed positionally
+        # rather than as keyword
+        # arguments.  note that apply_pos doesn't currently work in all cases
+        # such as when a kw-only indicator "*" is present, which is why
+        # we limit the use of this to just that case we can detect.  As we add
+        # more kinds of methods that use @decorator, things may have to
+        # be further improved in this area
+        if "__" in repr(spec[0]):
+            code = (
+                """\
+%(prefix)sdef %(name)s%(grouped_args)s:
+    return %(target_prefix)s%(target)s(%(fn)s, %(apply_pos)s)
+"""
+                % metadata
+            )
+        else:
+            code = (
+                """\
+%(prefix)sdef %(name)s%(grouped_args)s:
+    return %(target_prefix)s%(target)s(%(fn)s, %(apply_kw)s)
+"""
+                % metadata
+            )
+
+        mod = sys.modules[fn.__module__]
+        env.update(vars(mod))
+        env.update({targ_name: target, fn_name: fn, "__name__": fn.__module__})
+
+        decorated = cast(
+            types.FunctionType,
+            _exec_code_in_env(code, env, fn.__name__),
+        )
+        decorated.__defaults__ = getattr(fn, "__func__", fn).__defaults__
+
+        decorated.__wrapped__ = fn  # type: ignore[attr-defined]
+        return update_wrapper(decorated, fn)  # type: ignore[return-value]
+
+    return update_wrapper(decorate, target)  # type: ignore[return-value]
+
+
+def _update_argspec_defaults_into_env(spec, env):
+    """given a FullArgSpec, convert defaults to be symbol names in an env."""
+
+    if spec.defaults:
+        new_defaults = []
+        i = 0
+        for arg in spec.defaults:
+            if type(arg).__module__ not in ("builtins", "__builtin__"):
+                name = "x%d" % i
+                env[name] = arg
+                new_defaults.append(name)
+                i += 1
+            else:
+                new_defaults.append(arg)
+        elem = list(spec)
+        elem[3] = tuple(new_defaults)
+        return compat.FullArgSpec(*elem)
+    else:
+        return spec
+
+
+def _exec_code_in_env(
+    code: Union[str, types.CodeType], env: Dict[str, Any], fn_name: str
+) -> Callable[..., Any]:
+    exec(code, env)
+    return env[fn_name]  # type: ignore[no-any-return]
+
+
+_PF = TypeVar("_PF")
+_TE = TypeVar("_TE")
+
+
+class PluginLoader:
+    def __init__(
+        self, group: str, auto_fn: Optional[Callable[..., Any]] = None
+    ):
+        self.group = group
+        self.impls: Dict[str, Any] = {}
+        self.auto_fn = auto_fn
+
+    def clear(self):
+        self.impls.clear()
+
+    def load(self, name: str) -> Any:
+        if name in self.impls:
+            return self.impls[name]()
+
+        if self.auto_fn:
+            loader = self.auto_fn(name)
+            if loader:
+                self.impls[name] = loader
+                return loader()
+
+        for impl in compat.importlib_metadata_get(self.group):
+            if impl.name == name:
+                self.impls[name] = impl.load
+                return impl.load()
+
+        raise exc.NoSuchModuleError(
+            "Can't load plugin: %s:%s" % (self.group, name)
+        )
+
+    def register(self, name: str, modulepath: str, objname: str) -> None:
+        def load():
+            mod = __import__(modulepath)
+            for token in modulepath.split(".")[1:]:
+                mod = getattr(mod, token)
+            return getattr(mod, objname)
+
+        self.impls[name] = load
+
+
+def _inspect_func_args(fn):
+    try:
+        co_varkeywords = inspect.CO_VARKEYWORDS
+    except AttributeError:
+        # https://docs.python.org/3/library/inspect.html
+        # The flags are specific to CPython, and may not be defined in other
+        # Python implementations. Furthermore, the flags are an implementation
+        # detail, and can be removed or deprecated in future Python releases.
+        spec = compat.inspect_getfullargspec(fn)
+        return spec[0], bool(spec[2])
+    else:
+        # use fn.__code__ plus flags to reduce method call overhead
+        co = fn.__code__
+        nargs = co.co_argcount
+        return (
+            list(co.co_varnames[:nargs]),
+            bool(co.co_flags & co_varkeywords),
+        )
+
+
+@overload
+def get_cls_kwargs(
+    cls: type,
+    *,
+    _set: Optional[Set[str]] = None,
+    raiseerr: Literal[True] = ...,
+) -> Set[str]: ...
+
+
+@overload
+def get_cls_kwargs(
+    cls: type, *, _set: Optional[Set[str]] = None, raiseerr: bool = False
+) -> Optional[Set[str]]: ...
+
+
+def get_cls_kwargs(
+    cls: type, *, _set: Optional[Set[str]] = None, raiseerr: bool = False
+) -> Optional[Set[str]]:
+    r"""Return the full set of inherited kwargs for the given `cls`.
+
+    Probes a class's __init__ method, collecting all named arguments.  If the
+    __init__ defines a \**kwargs catch-all, then the constructor is presumed
+    to pass along unrecognized keywords to its base classes, and the
+    collection process is repeated recursively on each of the bases.
+
+    Uses a subset of inspect.getfullargspec() to cut down on method overhead,
+    as this is used within the Core typing system to create copies of type
+    objects which is a performance-sensitive operation.
+
+    No anonymous tuple arguments please !
+
+    """
+    toplevel = _set is None
+    if toplevel:
+        _set = set()
+    assert _set is not None
+
+    ctr = cls.__dict__.get("__init__", False)
+
+    has_init = (
+        ctr
+        and isinstance(ctr, types.FunctionType)
+        and isinstance(ctr.__code__, types.CodeType)
+    )
+
+    if has_init:
+        names, has_kw = _inspect_func_args(ctr)
+        _set.update(names)
+
+        if not has_kw and not toplevel:
+            if raiseerr:
+                raise TypeError(
+                    f"given cls {cls} doesn't have an __init__ method"
+                )
+            else:
+                return None
+    else:
+        has_kw = False
+
+    if not has_init or has_kw:
+        for c in cls.__bases__:
+            if get_cls_kwargs(c, _set=_set) is None:
+                break
+
+    _set.discard("self")
+    return _set
+
+
+def get_func_kwargs(func: Callable[..., Any]) -> List[str]:
+    """Return the set of legal kwargs for the given `func`.
+
+    Uses getargspec so is safe to call for methods, functions,
+    etc.
+
+    """
+
+    return compat.inspect_getfullargspec(func)[0]
+
+
+def get_callable_argspec(
+    fn: Callable[..., Any], no_self: bool = False, _is_init: bool = False
+) -> compat.FullArgSpec:
+    """Return the argument signature for any callable.
+
+    All pure-Python callables are accepted, including
+    functions, methods, classes, objects with __call__;
+    builtins and other edge cases like functools.partial() objects
+    raise a TypeError.
+
+    """
+    if inspect.isbuiltin(fn):
+        raise TypeError("Can't inspect builtin: %s" % fn)
+    elif inspect.isfunction(fn):
+        if _is_init and no_self:
+            spec = compat.inspect_getfullargspec(fn)
+            return compat.FullArgSpec(
+                spec.args[1:],
+                spec.varargs,
+                spec.varkw,
+                spec.defaults,
+                spec.kwonlyargs,
+                spec.kwonlydefaults,
+                spec.annotations,
+            )
+        else:
+            return compat.inspect_getfullargspec(fn)
+    elif inspect.ismethod(fn):
+        if no_self and (_is_init or fn.__self__):
+            spec = compat.inspect_getfullargspec(fn.__func__)
+            return compat.FullArgSpec(
+                spec.args[1:],
+                spec.varargs,
+                spec.varkw,
+                spec.defaults,
+                spec.kwonlyargs,
+                spec.kwonlydefaults,
+                spec.annotations,
+            )
+        else:
+            return compat.inspect_getfullargspec(fn.__func__)
+    elif inspect.isclass(fn):
+        return get_callable_argspec(
+            fn.__init__, no_self=no_self, _is_init=True
+        )
+    elif hasattr(fn, "__func__"):
+        return compat.inspect_getfullargspec(fn.__func__)
+    elif hasattr(fn, "__call__"):
+        if inspect.ismethod(fn.__call__):
+            return get_callable_argspec(fn.__call__, no_self=no_self)
+        else:
+            raise TypeError("Can't inspect callable: %s" % fn)
+    else:
+        raise TypeError("Can't inspect callable: %s" % fn)
+
+
+def format_argspec_plus(
+    fn: Union[Callable[..., Any], compat.FullArgSpec], grouped: bool = True
+) -> Dict[str, Optional[str]]:
+    """Returns a dictionary of formatted, introspected function arguments.
+
+    A enhanced variant of inspect.formatargspec to support code generation.
+
+    fn
+       An inspectable callable or tuple of inspect getargspec() results.
+    grouped
+      Defaults to True; include (parens, around, argument) lists
+
+    Returns:
+
+    args
+      Full inspect.formatargspec for fn
+    self_arg
+      The name of the first positional argument, varargs[0], or None
+      if the function defines no positional arguments.
+    apply_pos
+      args, re-written in calling rather than receiving syntax.  Arguments are
+      passed positionally.
+    apply_kw
+      Like apply_pos, except keyword-ish args are passed as keywords.
+    apply_pos_proxied
+      Like apply_pos but omits the self/cls argument
+
+    Example::
+
+      >>> format_argspec_plus(lambda self, a, b, c=3, **d: 123)
+      {'grouped_args': '(self, a, b, c=3, **d)',
+       'self_arg': 'self',
+       'apply_kw': '(self, a, b, c=c, **d)',
+       'apply_pos': '(self, a, b, c, **d)'}
+
+    """
+    if callable(fn):
+        spec = compat.inspect_getfullargspec(fn)
+    else:
+        spec = fn
+
+    args = compat.inspect_formatargspec(*spec)
+
+    apply_pos = compat.inspect_formatargspec(
+        spec[0], spec[1], spec[2], None, spec[4]
+    )
+
+    if spec[0]:
+        self_arg = spec[0][0]
+
+        apply_pos_proxied = compat.inspect_formatargspec(
+            spec[0][1:], spec[1], spec[2], None, spec[4]
+        )
+
+    elif spec[1]:
+        # I'm not sure what this is
+        self_arg = "%s[0]" % spec[1]
+
+        apply_pos_proxied = apply_pos
+    else:
+        self_arg = None
+        apply_pos_proxied = apply_pos
+
+    num_defaults = 0
+    if spec[3]:
+        num_defaults += len(cast(Tuple[Any], spec[3]))
+    if spec[4]:
+        num_defaults += len(spec[4])
+
+    name_args = spec[0] + spec[4]
+
+    defaulted_vals: Union[List[str], Tuple[()]]
+
+    if num_defaults:
+        defaulted_vals = name_args[0 - num_defaults :]
+    else:
+        defaulted_vals = ()
+
+    apply_kw = compat.inspect_formatargspec(
+        name_args,
+        spec[1],
+        spec[2],
+        defaulted_vals,
+        formatvalue=lambda x: "=" + str(x),
+    )
+
+    if spec[0]:
+        apply_kw_proxied = compat.inspect_formatargspec(
+            name_args[1:],
+            spec[1],
+            spec[2],
+            defaulted_vals,
+            formatvalue=lambda x: "=" + str(x),
+        )
+    else:
+        apply_kw_proxied = apply_kw
+
+    if grouped:
+        return dict(
+            grouped_args=args,
+            self_arg=self_arg,
+            apply_pos=apply_pos,
+            apply_kw=apply_kw,
+            apply_pos_proxied=apply_pos_proxied,
+            apply_kw_proxied=apply_kw_proxied,
+        )
+    else:
+        return dict(
+            grouped_args=args,
+            self_arg=self_arg,
+            apply_pos=apply_pos[1:-1],
+            apply_kw=apply_kw[1:-1],
+            apply_pos_proxied=apply_pos_proxied[1:-1],
+            apply_kw_proxied=apply_kw_proxied[1:-1],
+        )
+
+
+def format_argspec_init(method, grouped=True):
+    """format_argspec_plus with considerations for typical __init__ methods
+
+    Wraps format_argspec_plus with error handling strategies for typical
+    __init__ cases:
+
+    .. sourcecode:: text
+
+      object.__init__ -> (self)
+      other unreflectable (usually C) -> (self, *args, **kwargs)
+
+    """
+    if method is object.__init__:
+        grouped_args = "(self)"
+        args = "(self)" if grouped else "self"
+        proxied = "()" if grouped else ""
+    else:
+        try:
+            return format_argspec_plus(method, grouped=grouped)
+        except TypeError:
+            grouped_args = "(self, *args, **kwargs)"
+            args = grouped_args if grouped else "self, *args, **kwargs"
+            proxied = "(*args, **kwargs)" if grouped else "*args, **kwargs"
+    return dict(
+        self_arg="self",
+        grouped_args=grouped_args,
+        apply_pos=args,
+        apply_kw=args,
+        apply_pos_proxied=proxied,
+        apply_kw_proxied=proxied,
+    )
+
+
+def create_proxy_methods(
+    target_cls: Type[Any],
+    target_cls_sphinx_name: str,
+    proxy_cls_sphinx_name: str,
+    classmethods: Sequence[str] = (),
+    methods: Sequence[str] = (),
+    attributes: Sequence[str] = (),
+    use_intermediate_variable: Sequence[str] = (),
+) -> Callable[[_T], _T]:
+    """A class decorator indicating attributes should refer to a proxy
+    class.
+
+    This decorator is now a "marker" that does nothing at runtime.  Instead,
+    it is consumed by the tools/generate_proxy_methods.py script to
+    statically generate proxy methods and attributes that are fully
+    recognized by typing tools such as mypy.
+
+    """
+
+    def decorate(cls):
+        return cls
+
+    return decorate
+
+
+def getargspec_init(method):
+    """inspect.getargspec with considerations for typical __init__ methods
+
+    Wraps inspect.getargspec with error handling for typical __init__ cases:
+
+    .. sourcecode:: text
+
+      object.__init__ -> (self)
+      other unreflectable (usually C) -> (self, *args, **kwargs)
+
+    """
+    try:
+        return compat.inspect_getfullargspec(method)
+    except TypeError:
+        if method is object.__init__:
+            return (["self"], None, None, None)
+        else:
+            return (["self"], "args", "kwargs", None)
+
+
+def unbound_method_to_callable(func_or_cls):
+    """Adjust the incoming callable such that a 'self' argument is not
+    required.
+
+    """
+
+    if isinstance(func_or_cls, types.MethodType) and not func_or_cls.__self__:
+        return func_or_cls.__func__
+    else:
+        return func_or_cls
+
+
+def generic_repr(
+    obj: Any,
+    additional_kw: Sequence[Tuple[str, Any]] = (),
+    to_inspect: Optional[Union[object, List[object]]] = None,
+    omit_kwarg: Sequence[str] = (),
+) -> str:
+    """Produce a __repr__() based on direct association of the __init__()
+    specification vs. same-named attributes present.
+
+    """
+    if to_inspect is None:
+        to_inspect = [obj]
+    else:
+        to_inspect = _collections.to_list(to_inspect)
+
+    missing = object()
+
+    pos_args = []
+    kw_args: _collections.OrderedDict[str, Any] = _collections.OrderedDict()
+    vargs = None
+    for i, insp in enumerate(to_inspect):
+        try:
+            spec = compat.inspect_getfullargspec(insp.__init__)
+        except TypeError:
+            continue
+        else:
+            default_len = len(spec.defaults) if spec.defaults else 0
+            if i == 0:
+                if spec.varargs:
+                    vargs = spec.varargs
+                if default_len:
+                    pos_args.extend(spec.args[1:-default_len])
+                else:
+                    pos_args.extend(spec.args[1:])
+            else:
+                kw_args.update(
+                    [(arg, missing) for arg in spec.args[1:-default_len]]
+                )
+
+            if default_len:
+                assert spec.defaults
+                kw_args.update(
+                    [
+                        (arg, default)
+                        for arg, default in zip(
+                            spec.args[-default_len:], spec.defaults
+                        )
+                    ]
+                )
+    output: List[str] = []
+
+    output.extend(repr(getattr(obj, arg, None)) for arg in pos_args)
+
+    if vargs is not None and hasattr(obj, vargs):
+        output.extend([repr(val) for val in getattr(obj, vargs)])
+
+    for arg, defval in kw_args.items():
+        if arg in omit_kwarg:
+            continue
+        try:
+            val = getattr(obj, arg, missing)
+            if val is not missing and val != defval:
+                output.append("%s=%r" % (arg, val))
+        except Exception:
+            pass
+
+    if additional_kw:
+        for arg, defval in additional_kw:
+            try:
+                val = getattr(obj, arg, missing)
+                if val is not missing and val != defval:
+                    output.append("%s=%r" % (arg, val))
+            except Exception:
+                pass
+
+    return "%s(%s)" % (obj.__class__.__name__, ", ".join(output))
+
+
+class portable_instancemethod:
+    """Turn an instancemethod into a (parent, name) pair
+    to produce a serializable callable.
+
+    """
+
+    __slots__ = "target", "name", "kwargs", "__weakref__"
+
+    def __getstate__(self):
+        return {
+            "target": self.target,
+            "name": self.name,
+            "kwargs": self.kwargs,
+        }
+
+    def __setstate__(self, state):
+        self.target = state["target"]
+        self.name = state["name"]
+        self.kwargs = state.get("kwargs", ())
+
+    def __init__(self, meth, kwargs=()):
+        self.target = meth.__self__
+        self.name = meth.__name__
+        self.kwargs = kwargs
+
+    def __call__(self, *arg, **kw):
+        kw.update(self.kwargs)
+        return getattr(self.target, self.name)(*arg, **kw)
+
+
+def class_hierarchy(cls):
+    """Return an unordered sequence of all classes related to cls.
+
+    Traverses diamond hierarchies.
+
+    Fibs slightly: subclasses of builtin types are not returned.  Thus
+    class_hierarchy(class A(object)) returns (A, object), not A plus every
+    class systemwide that derives from object.
+
+    """
+
+    hier = {cls}
+    process = list(cls.__mro__)
+    while process:
+        c = process.pop()
+        bases = (_ for _ in c.__bases__ if _ not in hier)
+
+        for b in bases:
+            process.append(b)
+            hier.add(b)
+
+        if c.__module__ == "builtins" or not hasattr(c, "__subclasses__"):
+            continue
+
+        for s in [
+            _
+            for _ in (
+                c.__subclasses__()
+                if not issubclass(c, type)
+                else c.__subclasses__(c)
+            )
+            if _ not in hier
+        ]:
+            process.append(s)
+            hier.add(s)
+    return list(hier)
+
+
+def iterate_attributes(cls):
+    """iterate all the keys and attributes associated
+    with a class, without using getattr().
+
+    Does not use getattr() so that class-sensitive
+    descriptors (i.e. property.__get__()) are not called.
+
+    """
+    keys = dir(cls)
+    for key in keys:
+        for c in cls.__mro__:
+            if key in c.__dict__:
+                yield (key, c.__dict__[key])
+                break
+
+
+def monkeypatch_proxied_specials(
+    into_cls,
+    from_cls,
+    skip=None,
+    only=None,
+    name="self.proxy",
+    from_instance=None,
+):
+    """Automates delegation of __specials__ for a proxying type."""
+
+    if only:
+        dunders = only
+    else:
+        if skip is None:
+            skip = (
+                "__slots__",
+                "__del__",
+                "__getattribute__",
+                "__metaclass__",
+                "__getstate__",
+                "__setstate__",
+            )
+        dunders = [
+            m
+            for m in dir(from_cls)
+            if (
+                m.startswith("__")
+                and m.endswith("__")
+                and not hasattr(into_cls, m)
+                and m not in skip
+            )
+        ]
+
+    for method in dunders:
+        try:
+            maybe_fn = getattr(from_cls, method)
+            if not hasattr(maybe_fn, "__call__"):
+                continue
+            maybe_fn = getattr(maybe_fn, "__func__", maybe_fn)
+            fn = cast(types.FunctionType, maybe_fn)
+
+        except AttributeError:
+            continue
+        try:
+            spec = compat.inspect_getfullargspec(fn)
+            fn_args = compat.inspect_formatargspec(spec[0])
+            d_args = compat.inspect_formatargspec(spec[0][1:])
+        except TypeError:
+            fn_args = "(self, *args, **kw)"
+            d_args = "(*args, **kw)"
+
+        py = (
+            "def %(method)s%(fn_args)s: "
+            "return %(name)s.%(method)s%(d_args)s" % locals()
+        )
+
+        env: Dict[str, types.FunctionType] = (
+            from_instance is not None and {name: from_instance} or {}
+        )
+        exec(py, env)
+        try:
+            env[method].__defaults__ = fn.__defaults__
+        except AttributeError:
+            pass
+        setattr(into_cls, method, env[method])
+
+
+def methods_equivalent(meth1, meth2):
+    """Return True if the two methods are the same implementation."""
+
+    return getattr(meth1, "__func__", meth1) is getattr(
+        meth2, "__func__", meth2
+    )
+
+
+def as_interface(obj, cls=None, methods=None, required=None):
+    """Ensure basic interface compliance for an instance or dict of callables.
+
+    Checks that ``obj`` implements public methods of ``cls`` or has members
+    listed in ``methods``. If ``required`` is not supplied, implementing at
+    least one interface method is sufficient. Methods present on ``obj`` that
+    are not in the interface are ignored.
+
+    If ``obj`` is a dict and ``dict`` does not meet the interface
+    requirements, the keys of the dictionary are inspected. Keys present in
+    ``obj`` that are not in the interface will raise TypeErrors.
+
+    Raises TypeError if ``obj`` does not meet the interface criteria.
+
+    In all passing cases, an object with callable members is returned.  In the
+    simple case, ``obj`` is returned as-is; if dict processing kicks in then
+    an anonymous class is returned.
+
+    obj
+      A type, instance, or dictionary of callables.
+    cls
+      Optional, a type.  All public methods of cls are considered the
+      interface.  An ``obj`` instance of cls will always pass, ignoring
+      ``required``..
+    methods
+      Optional, a sequence of method names to consider as the interface.
+    required
+      Optional, a sequence of mandatory implementations. If omitted, an
+      ``obj`` that provides at least one interface method is considered
+      sufficient.  As a convenience, required may be a type, in which case
+      all public methods of the type are required.
+
+    """
+    if not cls and not methods:
+        raise TypeError("a class or collection of method names are required")
+
+    if isinstance(cls, type) and isinstance(obj, cls):
+        return obj
+
+    interface = set(methods or [m for m in dir(cls) if not m.startswith("_")])
+    implemented = set(dir(obj))
+
+    complies = operator.ge
+    if isinstance(required, type):
+        required = interface
+    elif not required:
+        required = set()
+        complies = operator.gt
+    else:
+        required = set(required)
+
+    if complies(implemented.intersection(interface), required):
+        return obj
+
+    # No dict duck typing here.
+    if not isinstance(obj, dict):
+        qualifier = complies is operator.gt and "any of" or "all of"
+        raise TypeError(
+            "%r does not implement %s: %s"
+            % (obj, qualifier, ", ".join(interface))
+        )
+
+    class AnonymousInterface:
+        """A callable-holding shell."""
+
+    if cls:
+        AnonymousInterface.__name__ = "Anonymous" + cls.__name__
+    found = set()
+
+    for method, impl in dictlike_iteritems(obj):
+        if method not in interface:
+            raise TypeError("%r: unknown in this interface" % method)
+        if not callable(impl):
+            raise TypeError("%r=%r is not callable" % (method, impl))
+        setattr(AnonymousInterface, method, staticmethod(impl))
+        found.add(method)
+
+    if complies(found, required):
+        return AnonymousInterface
+
+    raise TypeError(
+        "dictionary does not contain required keys %s"
+        % ", ".join(required - found)
+    )
+
+
+_GFD = TypeVar("_GFD", bound="generic_fn_descriptor[Any]")
+
+
+class generic_fn_descriptor(Generic[_T_co]):
+    """Descriptor which proxies a function when the attribute is not
+    present in dict
+
+    This superclass is organized in a particular way with "memoized" and
+    "non-memoized" implementation classes that are hidden from type checkers,
+    as Mypy seems to not be able to handle seeing multiple kinds of descriptor
+    classes used for the same attribute.
+
+    """
+
+    fget: Callable[..., _T_co]
+    __doc__: Optional[str]
+    __name__: str
+
+    def __init__(self, fget: Callable[..., _T_co], doc: Optional[str] = None):
+        self.fget = fget
+        self.__doc__ = doc or fget.__doc__
+        self.__name__ = fget.__name__
+
+    @overload
+    def __get__(self: _GFD, obj: None, cls: Any) -> _GFD: ...
+
+    @overload
+    def __get__(self, obj: object, cls: Any) -> _T_co: ...
+
+    def __get__(self: _GFD, obj: Any, cls: Any) -> Union[_GFD, _T_co]:
+        raise NotImplementedError()
+
+    if TYPE_CHECKING:
+
+        def __set__(self, instance: Any, value: Any) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
+
+    def _reset(self, obj: Any) -> None:
+        raise NotImplementedError()
+
+    @classmethod
+    def reset(cls, obj: Any, name: str) -> None:
+        raise NotImplementedError()
+
+
+class _non_memoized_property(generic_fn_descriptor[_T_co]):
+    """a plain descriptor that proxies a function.
+
+    primary rationale is to provide a plain attribute that's
+    compatible with memoized_property which is also recognized as equivalent
+    by mypy.
+
+    """
+
+    if not TYPE_CHECKING:
+
+        def __get__(self, obj, cls):
+            if obj is None:
+                return self
+            return self.fget(obj)
+
+
+class _memoized_property(generic_fn_descriptor[_T_co]):
+    """A read-only @property that is only evaluated once."""
+
+    if not TYPE_CHECKING:
+
+        def __get__(self, obj, cls):
+            if obj is None:
+                return self
+            obj.__dict__[self.__name__] = result = self.fget(obj)
+            return result
+
+    def _reset(self, obj):
+        _memoized_property.reset(obj, self.__name__)
+
+    @classmethod
+    def reset(cls, obj, name):
+        obj.__dict__.pop(name, None)
+
+
+# despite many attempts to get Mypy to recognize an overridden descriptor
+# where one is memoized and the other isn't, there seems to be no reliable
+# way other than completely deceiving the type checker into thinking there
+# is just one single descriptor type everywhere.  Otherwise, if a superclass
+# has non-memoized and subclass has memoized, that requires
+# "class memoized(non_memoized)".  but then if a superclass has memoized and
+# superclass has non-memoized, the class hierarchy of the descriptors
+# would need to be reversed; "class non_memoized(memoized)".  so there's no
+# way to achieve this.
+# additional issues, RO properties:
+# https://github.com/python/mypy/issues/12440
+if TYPE_CHECKING:
+    # allow memoized and non-memoized to be freely mixed by having them
+    # be the same class
+    memoized_property = generic_fn_descriptor
+    non_memoized_property = generic_fn_descriptor
+
+    # for read only situations, mypy only sees @property as read only.
+    # read only is needed when a subtype specializes the return type
+    # of a property, meaning assignment needs to be disallowed
+    ro_memoized_property = property
+    ro_non_memoized_property = property
+
+else:
+    memoized_property = ro_memoized_property = _memoized_property
+    non_memoized_property = ro_non_memoized_property = _non_memoized_property
+
+
+def memoized_instancemethod(fn: _F) -> _F:
+    """Decorate a method memoize its return value.
+
+    Best applied to no-arg methods: memoization is not sensitive to
+    argument values, and will always return the same value even when
+    called with different arguments.
+
+    """
+
+    def oneshot(self, *args, **kw):
+        result = fn(self, *args, **kw)
+
+        def memo(*a, **kw):
+            return result
+
+        memo.__name__ = fn.__name__
+        memo.__doc__ = fn.__doc__
+        self.__dict__[fn.__name__] = memo
+        return result
+
+    return update_wrapper(oneshot, fn)  # type: ignore
+
+
+class HasMemoized:
+    """A mixin class that maintains the names of memoized elements in a
+    collection for easy cache clearing, generative, etc.
+
+    """
+
+    if not TYPE_CHECKING:
+        # support classes that want to have __slots__ with an explicit
+        # slot for __dict__.  not sure if that requires base __slots__ here.
+        __slots__ = ()
+
+    _memoized_keys: FrozenSet[str] = frozenset()
+
+    def _reset_memoizations(self) -> None:
+        for elem in self._memoized_keys:
+            self.__dict__.pop(elem, None)
+
+    def _assert_no_memoizations(self) -> None:
+        for elem in self._memoized_keys:
+            assert elem not in self.__dict__
+
+    def _set_memoized_attribute(self, key: str, value: Any) -> None:
+        self.__dict__[key] = value
+        self._memoized_keys |= {key}
+
+    class memoized_attribute(memoized_property[_T]):
+        """A read-only @property that is only evaluated once.
+
+        :meta private:
+
+        """
+
+        fget: Callable[..., _T]
+        __doc__: Optional[str]
+        __name__: str
+
+        def __init__(self, fget: Callable[..., _T], doc: Optional[str] = None):
+            self.fget = fget
+            self.__doc__ = doc or fget.__doc__
+            self.__name__ = fget.__name__
+
+        @overload
+        def __get__(self: _MA, obj: None, cls: Any) -> _MA: ...
+
+        @overload
+        def __get__(self, obj: Any, cls: Any) -> _T: ...
+
+        def __get__(self, obj, cls):
+            if obj is None:
+                return self
+            obj.__dict__[self.__name__] = result = self.fget(obj)
+            obj._memoized_keys |= {self.__name__}
+            return result
+
+    @classmethod
+    def memoized_instancemethod(cls, fn: _F) -> _F:
+        """Decorate a method memoize its return value.
+
+        :meta private:
+
+        """
+
+        def oneshot(self: Any, *args: Any, **kw: Any) -> Any:
+            result = fn(self, *args, **kw)
+
+            def memo(*a, **kw):
+                return result
+
+            memo.__name__ = fn.__name__
+            memo.__doc__ = fn.__doc__
+            self.__dict__[fn.__name__] = memo
+            self._memoized_keys |= {fn.__name__}
+            return result
+
+        return update_wrapper(oneshot, fn)  # type: ignore
+
+
+if TYPE_CHECKING:
+    HasMemoized_ro_memoized_attribute = property
+else:
+    HasMemoized_ro_memoized_attribute = HasMemoized.memoized_attribute
+
+
+class MemoizedSlots:
+    """Apply memoized items to an object using a __getattr__ scheme.
+
+    This allows the functionality of memoized_property and
+    memoized_instancemethod to be available to a class using __slots__.
+
+    """
+
+    __slots__ = ()
+
+    def _fallback_getattr(self, key):
+        raise AttributeError(key)
+
+    def __getattr__(self, key: str) -> Any:
+        if key.startswith("_memoized_attr_") or key.startswith(
+            "_memoized_method_"
+        ):
+            raise AttributeError(key)
+        # to avoid recursion errors when interacting with other __getattr__
+        # schemes that refer to this one, when testing for memoized method
+        # look at __class__ only rather than going into __getattr__ again.
+        elif hasattr(self.__class__, f"_memoized_attr_{key}"):
+            value = getattr(self, f"_memoized_attr_{key}")()
+            setattr(self, key, value)
+            return value
+        elif hasattr(self.__class__, f"_memoized_method_{key}"):
+            fn = getattr(self, f"_memoized_method_{key}")
+
+            def oneshot(*args, **kw):
+                result = fn(*args, **kw)
+
+                def memo(*a, **kw):
+                    return result
+
+                memo.__name__ = fn.__name__
+                memo.__doc__ = fn.__doc__
+                setattr(self, key, memo)
+                return result
+
+            oneshot.__doc__ = fn.__doc__
+            return oneshot
+        else:
+            return self._fallback_getattr(key)
+
+
+# from paste.deploy.converters
+def asbool(obj: Any) -> bool:
+    if isinstance(obj, str):
+        obj = obj.strip().lower()
+        if obj in ["true", "yes", "on", "y", "t", "1"]:
+            return True
+        elif obj in ["false", "no", "off", "n", "f", "0"]:
+            return False
+        else:
+            raise ValueError("String is not true/false: %r" % obj)
+    return bool(obj)
+
+
+def bool_or_str(*text: str) -> Callable[[str], Union[str, bool]]:
+    """Return a callable that will evaluate a string as
+    boolean, or one of a set of "alternate" string values.
+
+    """
+
+    def bool_or_value(obj: str) -> Union[str, bool]:
+        if obj in text:
+            return obj
+        else:
+            return asbool(obj)
+
+    return bool_or_value
+
+
+def asint(value: Any) -> Optional[int]:
+    """Coerce to integer."""
+
+    if value is None:
+        return value
+    return int(value)
+
+
+def coerce_kw_type(
+    kw: Dict[str, Any],
+    key: str,
+    type_: Type[Any],
+    flexi_bool: bool = True,
+    dest: Optional[Dict[str, Any]] = None,
+) -> None:
+    r"""If 'key' is present in dict 'kw', coerce its value to type 'type\_' if
+    necessary.  If 'flexi_bool' is True, the string '0' is considered false
+    when coercing to boolean.
+    """
+
+    if dest is None:
+        dest = kw
+
+    if (
+        key in kw
+        and (not isinstance(type_, type) or not isinstance(kw[key], type_))
+        and kw[key] is not None
+    ):
+        if type_ is bool and flexi_bool:
+            dest[key] = asbool(kw[key])
+        else:
+            dest[key] = type_(kw[key])
+
+
+def constructor_key(obj: Any, cls: Type[Any]) -> Tuple[Any, ...]:
+    """Produce a tuple structure that is cacheable using the __dict__ of
+    obj to retrieve values
+
+    """
+    names = get_cls_kwargs(cls)
+    return (cls,) + tuple(
+        (k, obj.__dict__[k]) for k in names if k in obj.__dict__
+    )
+
+
+def constructor_copy(obj: _T, cls: Type[_T], *args: Any, **kw: Any) -> _T:
+    """Instantiate cls using the __dict__ of obj as constructor arguments.
+
+    Uses inspect to match the named arguments of ``cls``.
+
+    """
+
+    names = get_cls_kwargs(cls)
+    kw.update(
+        (k, obj.__dict__[k]) for k in names.difference(kw) if k in obj.__dict__
+    )
+    return cls(*args, **kw)
+
+
+def counter() -> Callable[[], int]:
+    """Return a threadsafe counter function."""
+
+    lock = threading.Lock()
+    counter = itertools.count(1)
+
+    # avoid the 2to3 "next" transformation...
+    def _next():
+        with lock:
+            return next(counter)
+
+    return _next
+
+
+def duck_type_collection(
+    specimen: Any, default: Optional[Type[Any]] = None
+) -> Optional[Type[Any]]:
+    """Given an instance or class, guess if it is or is acting as one of
+    the basic collection types: list, set and dict.  If the __emulates__
+    property is present, return that preferentially.
+    """
+
+    if hasattr(specimen, "__emulates__"):
+        # canonicalize set vs sets.Set to a standard: the builtin set
+        if specimen.__emulates__ is not None and issubclass(
+            specimen.__emulates__, set
+        ):
+            return set
+        else:
+            return specimen.__emulates__  # type: ignore
+
+    isa = issubclass if isinstance(specimen, type) else isinstance
+    if isa(specimen, list):
+        return list
+    elif isa(specimen, set):
+        return set
+    elif isa(specimen, dict):
+        return dict
+
+    if hasattr(specimen, "append"):
+        return list
+    elif hasattr(specimen, "add"):
+        return set
+    elif hasattr(specimen, "set"):
+        return dict
+    else:
+        return default
+
+
+def assert_arg_type(
+    arg: Any, argtype: Union[Tuple[Type[Any], ...], Type[Any]], name: str
+) -> Any:
+    if isinstance(arg, argtype):
+        return arg
+    else:
+        if isinstance(argtype, tuple):
+            raise exc.ArgumentError(
+                "Argument '%s' is expected to be one of type %s, got '%s'"
+                % (name, " or ".join("'%s'" % a for a in argtype), type(arg))
+            )
+        else:
+            raise exc.ArgumentError(
+                "Argument '%s' is expected to be of type '%s', got '%s'"
+                % (name, argtype, type(arg))
+            )
+
+
+def dictlike_iteritems(dictlike):
+    """Return a (key, value) iterator for almost any dict-like object."""
+
+    if hasattr(dictlike, "items"):
+        return list(dictlike.items())
+
+    getter = getattr(dictlike, "__getitem__", getattr(dictlike, "get", None))
+    if getter is None:
+        raise TypeError("Object '%r' is not dict-like" % dictlike)
+
+    if hasattr(dictlike, "iterkeys"):
+
+        def iterator():
+            for key in dictlike.iterkeys():
+                assert getter is not None
+                yield key, getter(key)
+
+        return iterator()
+    elif hasattr(dictlike, "keys"):
+        return iter((key, getter(key)) for key in dictlike.keys())
+    else:
+        raise TypeError("Object '%r' is not dict-like" % dictlike)
+
+
+class classproperty(property):
+    """A decorator that behaves like @property except that operates
+    on classes rather than instances.
+
+    The decorator is currently special when using the declarative
+    module, but note that the
+    :class:`~.sqlalchemy.ext.declarative.declared_attr`
+    decorator should be used for this purpose with declarative.
+
+    """
+
+    fget: Callable[[Any], Any]
+
+    def __init__(self, fget: Callable[[Any], Any], *arg: Any, **kw: Any):
+        super().__init__(fget, *arg, **kw)
+        self.__doc__ = fget.__doc__
+
+    def __get__(self, obj: Any, cls: Optional[type] = None) -> Any:
+        return self.fget(cls)
+
+
+class hybridproperty(Generic[_T]):
+    def __init__(self, func: Callable[..., _T]):
+        self.func = func
+        self.clslevel = func
+
+    def __get__(self, instance: Any, owner: Any) -> _T:
+        if instance is None:
+            clsval = self.clslevel(owner)
+            return clsval
+        else:
+            return self.func(instance)
+
+    def classlevel(self, func: Callable[..., Any]) -> hybridproperty[_T]:
+        self.clslevel = func
+        return self
+
+
+class rw_hybridproperty(Generic[_T]):
+    def __init__(self, func: Callable[..., _T]):
+        self.func = func
+        self.clslevel = func
+        self.setfn: Optional[Callable[..., Any]] = None
+
+    def __get__(self, instance: Any, owner: Any) -> _T:
+        if instance is None:
+            clsval = self.clslevel(owner)
+            return clsval
+        else:
+            return self.func(instance)
+
+    def __set__(self, instance: Any, value: Any) -> None:
+        assert self.setfn is not None
+        self.setfn(instance, value)
+
+    def setter(self, func: Callable[..., Any]) -> rw_hybridproperty[_T]:
+        self.setfn = func
+        return self
+
+    def classlevel(self, func: Callable[..., Any]) -> rw_hybridproperty[_T]:
+        self.clslevel = func
+        return self
+
+
+class hybridmethod(Generic[_T]):
+    """Decorate a function as cls- or instance- level."""
+
+    def __init__(self, func: Callable[..., _T]):
+        self.func = self.__func__ = func
+        self.clslevel = func
+
+    def __get__(self, instance: Any, owner: Any) -> Callable[..., _T]:
+        if instance is None:
+            return self.clslevel.__get__(owner, owner.__class__)  # type:ignore
+        else:
+            return self.func.__get__(instance, owner)  # type:ignore
+
+    def classlevel(self, func: Callable[..., Any]) -> hybridmethod[_T]:
+        self.clslevel = func
+        return self
+
+
+class symbol(int):
+    """A constant symbol.
+
+    >>> symbol("foo") is symbol("foo")
+    True
+    >>> symbol("foo")
+    <symbol 'foo>
+
+    A slight refinement of the MAGICCOOKIE=object() pattern.  The primary
+    advantage of symbol() is its repr().  They are also singletons.
+
+    Repeated calls of symbol('name') will all return the same instance.
+
+    """
+
+    name: str
+
+    symbols: Dict[str, symbol] = {}
+    _lock = threading.Lock()
+
+    def __new__(
+        cls,
+        name: str,
+        doc: Optional[str] = None,
+        canonical: Optional[int] = None,
+    ) -> symbol:
+        with cls._lock:
+            sym = cls.symbols.get(name)
+            if sym is None:
+                assert isinstance(name, str)
+                if canonical is None:
+                    canonical = hash(name)
+                sym = int.__new__(symbol, canonical)
+                sym.name = name
+                if doc:
+                    sym.__doc__ = doc
+
+                # NOTE: we should ultimately get rid of this global thing,
+                # however, currently it is to support pickling.  The best
+                # change would be when we are on py3.11 at a minimum, we
+                # switch to stdlib enum.IntFlag.
+                cls.symbols[name] = sym
+            else:
+                if canonical and canonical != sym:
+                    raise TypeError(
+                        f"Can't replace canonical symbol for {name!r} "
+                        f"with new int value {canonical}"
+                    )
+            return sym
+
+    def __reduce__(self):
+        return symbol, (self.name, "x", int(self))
+
+    def __str__(self):
+        return repr(self)
+
+    def __repr__(self):
+        return f"symbol({self.name!r})"
+
+
+class _IntFlagMeta(type):
+    def __init__(
+        cls,
+        classname: str,
+        bases: Tuple[Type[Any], ...],
+        dict_: Dict[str, Any],
+        **kw: Any,
+    ) -> None:
+        items: List[symbol]
+        cls._items = items = []
+        for k, v in dict_.items():
+            if re.match(r"^__.*__$", k):
+                continue
+            if isinstance(v, int):
+                sym = symbol(k, canonical=v)
+            elif not k.startswith("_"):
+                raise TypeError("Expected integer values for IntFlag")
+            else:
+                continue
+            setattr(cls, k, sym)
+            items.append(sym)
+
+        cls.__members__ = _collections.immutabledict(
+            {sym.name: sym for sym in items}
+        )
+
+    def __iter__(self) -> Iterator[symbol]:
+        raise NotImplementedError(
+            "iter not implemented to ensure compatibility with "
+            "Python 3.11 IntFlag.  Please use __members__.  See "
+            "https://github.com/python/cpython/issues/99304"
+        )
+
+
+class _FastIntFlag(metaclass=_IntFlagMeta):
+    """An 'IntFlag' copycat that isn't slow when performing bitwise
+    operations.
+
+    the ``FastIntFlag`` class will return ``enum.IntFlag`` under TYPE_CHECKING
+    and ``_FastIntFlag`` otherwise.
+
+    """
+
+
+if TYPE_CHECKING:
+    from enum import IntFlag
+
+    FastIntFlag = IntFlag
+else:
+    FastIntFlag = _FastIntFlag
+
+
+_E = TypeVar("_E", bound=enum.Enum)
+
+
+def parse_user_argument_for_enum(
+    arg: Any,
+    choices: Dict[_E, List[Any]],
+    name: str,
+    resolve_symbol_names: bool = False,
+) -> Optional[_E]:
+    """Given a user parameter, parse the parameter into a chosen value
+    from a list of choice objects, typically Enum values.
+
+    The user argument can be a string name that matches the name of a
+    symbol, or the symbol object itself, or any number of alternate choices
+    such as True/False/ None etc.
+
+    :param arg: the user argument.
+    :param choices: dictionary of enum values to lists of possible
+        entries for each.
+    :param name: name of the argument.   Used in an :class:`.ArgumentError`
+        that is raised if the parameter doesn't match any available argument.
+
+    """
+    for enum_value, choice in choices.items():
+        if arg is enum_value:
+            return enum_value
+        elif resolve_symbol_names and arg == enum_value.name:
+            return enum_value
+        elif arg in choice:
+            return enum_value
+
+    if arg is None:
+        return None
+
+    raise exc.ArgumentError(f"Invalid value for '{name}': {arg!r}")
+
+
+_creation_order = 1
+
+
+def set_creation_order(instance: Any) -> None:
+    """Assign a '_creation_order' sequence to the given instance.
+
+    This allows multiple instances to be sorted in order of creation
+    (typically within a single thread; the counter is not particularly
+    threadsafe).
+
+    """
+    global _creation_order
+    instance._creation_order = _creation_order
+    _creation_order += 1
+
+
+def warn_exception(func: Callable[..., Any], *args: Any, **kwargs: Any) -> Any:
+    """executes the given function, catches all exceptions and converts to
+    a warning.
+
+    """
+    try:
+        return func(*args, **kwargs)
+    except Exception:
+        warn("%s('%s') ignored" % sys.exc_info()[0:2])
+
+
+def ellipses_string(value, len_=25):
+    try:
+        if len(value) > len_:
+            return "%s..." % value[0:len_]
+        else:
+            return value
+    except TypeError:
+        return value
+
+
+class _hash_limit_string(str):
+    """A string subclass that can only be hashed on a maximum amount
+    of unique values.
+
+    This is used for warnings so that we can send out parameterized warnings
+    without the __warningregistry__ of the module,  or the non-overridable
+    "once" registry within warnings.py, overloading memory,
+
+
+    """
+
+    _hash: int
+
+    def __new__(
+        cls, value: str, num: int, args: Sequence[Any]
+    ) -> _hash_limit_string:
+        interpolated = (value % args) + (
+            " (this warning may be suppressed after %d occurrences)" % num
+        )
+        self = super().__new__(cls, interpolated)
+        self._hash = hash("%s_%d" % (value, hash(interpolated) % num))
+        return self
+
+    def __hash__(self) -> int:
+        return self._hash
+
+    def __eq__(self, other: Any) -> bool:
+        return hash(self) == hash(other)
+
+
+def warn(msg: str, code: Optional[str] = None) -> None:
+    """Issue a warning.
+
+    If msg is a string, :class:`.exc.SAWarning` is used as
+    the category.
+
+    """
+    if code:
+        _warnings_warn(exc.SAWarning(msg, code=code))
+    else:
+        _warnings_warn(msg, exc.SAWarning)
+
+
+def warn_limited(msg: str, args: Sequence[Any]) -> None:
+    """Issue a warning with a parameterized string, limiting the number
+    of registrations.
+
+    """
+    if args:
+        msg = _hash_limit_string(msg, 10, args)
+    _warnings_warn(msg, exc.SAWarning)
+
+
+_warning_tags: Dict[CodeType, Tuple[str, Type[Warning]]] = {}
+
+
+def tag_method_for_warnings(
+    message: str, category: Type[Warning]
+) -> Callable[[_F], _F]:
+    def go(fn):
+        _warning_tags[fn.__code__] = (message, category)
+        return fn
+
+    return go
+
+
+_not_sa_pattern = re.compile(r"^(?:sqlalchemy\.(?!testing)|alembic\.)")
+
+
+def _warnings_warn(
+    message: Union[str, Warning],
+    category: Optional[Type[Warning]] = None,
+    stacklevel: int = 2,
+) -> None:
+    # adjust the given stacklevel to be outside of SQLAlchemy
+    try:
+        frame = sys._getframe(stacklevel)
+    except ValueError:
+        # being called from less than 3 (or given) stacklevels, weird,
+        # but don't crash
+        stacklevel = 0
+    except:
+        # _getframe() doesn't work, weird interpreter issue, weird,
+        # ok, but don't crash
+        stacklevel = 0
+    else:
+        stacklevel_found = warning_tag_found = False
+        while frame is not None:
+            # using __name__ here requires that we have __name__ in the
+            # __globals__ of the decorated string functions we make also.
+            # we generate this using {"__name__": fn.__module__}
+            if not stacklevel_found and not re.match(
+                _not_sa_pattern, frame.f_globals.get("__name__", "")
+            ):
+                # stop incrementing stack level if an out-of-SQLA line
+                # were found.
+                stacklevel_found = True
+
+                # however, for the warning tag thing, we have to keep
+                # scanning up the whole traceback
+
+            if frame.f_code in _warning_tags:
+                warning_tag_found = True
+                (_suffix, _category) = _warning_tags[frame.f_code]
+                category = category or _category
+                message = f"{message} ({_suffix})"
+
+            frame = frame.f_back  # type: ignore[assignment]
+
+            if not stacklevel_found:
+                stacklevel += 1
+            elif stacklevel_found and warning_tag_found:
+                break
+
+    if category is not None:
+        warnings.warn(message, category, stacklevel=stacklevel + 1)
+    else:
+        warnings.warn(message, stacklevel=stacklevel + 1)
+
+
+def only_once(
+    fn: Callable[..., _T], retry_on_exception: bool
+) -> Callable[..., Optional[_T]]:
+    """Decorate the given function to be a no-op after it is called exactly
+    once."""
+
+    once = [fn]
+
+    def go(*arg: Any, **kw: Any) -> Optional[_T]:
+        # strong reference fn so that it isn't garbage collected,
+        # which interferes with the event system's expectations
+        strong_fn = fn  # noqa
+        if once:
+            once_fn = once.pop()
+            try:
+                return once_fn(*arg, **kw)
+            except:
+                if retry_on_exception:
+                    once.insert(0, once_fn)
+                raise
+
+        return None
+
+    return go
+
+
+_SQLA_RE = re.compile(r"sqlalchemy/([a-z_]+/){0,2}[a-z_]+\.py")
+_UNITTEST_RE = re.compile(r"unit(?:2|test2?/)")
+
+
+def chop_traceback(
+    tb: List[str],
+    exclude_prefix: re.Pattern[str] = _UNITTEST_RE,
+    exclude_suffix: re.Pattern[str] = _SQLA_RE,
+) -> List[str]:
+    """Chop extraneous lines off beginning and end of a traceback.
+
+    :param tb:
+      a list of traceback lines as returned by ``traceback.format_stack()``
+
+    :param exclude_prefix:
+      a regular expression object matching lines to skip at beginning of
+      ``tb``
+
+    :param exclude_suffix:
+      a regular expression object matching lines to skip at end of ``tb``
+    """
+    start = 0
+    end = len(tb) - 1
+    while start <= end and exclude_prefix.search(tb[start]):
+        start += 1
+    while start <= end and exclude_suffix.search(tb[end]):
+        end -= 1
+    return tb[start : end + 1]
+
+
+NoneType = type(None)
+
+
+def attrsetter(attrname):
+    code = "def set(obj, value):    obj.%s = value" % attrname
+    env = locals().copy()
+    exec(code, env)
+    return env["set"]
+
+
+_dunders = re.compile("^__.+__$")
+
+
+class TypingOnly:
+    """A mixin class that marks a class as 'typing only', meaning it has
+    absolutely no methods, attributes, or runtime functionality whatsoever.
+
+    """
+
+    __slots__ = ()
+
+    def __init_subclass__(cls) -> None:
+        if TypingOnly in cls.__bases__:
+            remaining = {
+                name for name in cls.__dict__ if not _dunders.match(name)
+            }
+            if remaining:
+                raise AssertionError(
+                    f"Class {cls} directly inherits TypingOnly but has "
+                    f"additional attributes {remaining}."
+                )
+        super().__init_subclass__()
+
+
+class EnsureKWArg:
+    r"""Apply translation of functions to accept \**kw arguments if they
+    don't already.
+
+    Used to ensure cross-compatibility with third party legacy code, for things
+    like compiler visit methods that need to accept ``**kw`` arguments,
+    but may have been copied from old code that didn't accept them.
+
+    """
+
+    ensure_kwarg: str
+    """a regular expression that indicates method names for which the method
+    should accept ``**kw`` arguments.
+
+    The class will scan for methods matching the name template and decorate
+    them if necessary to ensure ``**kw`` parameters are accepted.
+
+    """
+
+    def __init_subclass__(cls) -> None:
+        fn_reg = cls.ensure_kwarg
+        clsdict = cls.__dict__
+        if fn_reg:
+            for key in clsdict:
+                m = re.match(fn_reg, key)
+                if m:
+                    fn = clsdict[key]
+                    spec = compat.inspect_getfullargspec(fn)
+                    if not spec.varkw:
+                        wrapped = cls._wrap_w_kw(fn)
+                        setattr(cls, key, wrapped)
+        super().__init_subclass__()
+
+    @classmethod
+    def _wrap_w_kw(cls, fn: Callable[..., Any]) -> Callable[..., Any]:
+        def wrap(*arg: Any, **kw: Any) -> Any:
+            return fn(*arg)
+
+        return update_wrapper(wrap, fn)
+
+
+def wrap_callable(wrapper, fn):
+    """Augment functools.update_wrapper() to work with objects with
+    a ``__call__()`` method.
+
+    :param fn:
+      object with __call__ method
+
+    """
+    if hasattr(fn, "__name__"):
+        return update_wrapper(wrapper, fn)
+    else:
+        _f = wrapper
+        _f.__name__ = fn.__class__.__name__
+        if hasattr(fn, "__module__"):
+            _f.__module__ = fn.__module__
+
+        if hasattr(fn.__call__, "__doc__") and fn.__call__.__doc__:
+            _f.__doc__ = fn.__call__.__doc__
+        elif fn.__doc__:
+            _f.__doc__ = fn.__doc__
+
+        return _f
+
+
+def quoted_token_parser(value):
+    """Parse a dotted identifier with accommodation for quoted names.
+
+    Includes support for SQL-style double quotes as a literal character.
+
+    E.g.::
+
+        >>> quoted_token_parser("name")
+        ["name"]
+        >>> quoted_token_parser("schema.name")
+        ["schema", "name"]
+        >>> quoted_token_parser('"Schema"."Name"')
+        ['Schema', 'Name']
+        >>> quoted_token_parser('"Schema"."Name""Foo"')
+        ['Schema', 'Name""Foo']
+
+    """
+
+    if '"' not in value:
+        return value.split(".")
+
+    # 0 = outside of quotes
+    # 1 = inside of quotes
+    state = 0
+    result: List[List[str]] = [[]]
+    idx = 0
+    lv = len(value)
+    while idx < lv:
+        char = value[idx]
+        if char == '"':
+            if state == 1 and idx < lv - 1 and value[idx + 1] == '"':
+                result[-1].append('"')
+                idx += 1
+            else:
+                state ^= 1
+        elif char == "." and state == 0:
+            result.append([])
+        else:
+            result[-1].append(char)
+        idx += 1
+
+    return ["".join(token) for token in result]
+
+
+def add_parameter_text(params: Any, text: str) -> Callable[[_F], _F]:
+    params = _collections.to_list(params)
+
+    def decorate(fn):
+        doc = fn.__doc__ is not None and fn.__doc__ or ""
+        if doc:
+            doc = inject_param_text(doc, {param: text for param in params})
+        fn.__doc__ = doc
+        return fn
+
+    return decorate
+
+
+def _dedent_docstring(text: str) -> str:
+    split_text = text.split("\n", 1)
+    if len(split_text) == 1:
+        return text
+    else:
+        firstline, remaining = split_text
+    if not firstline.startswith(" "):
+        return firstline + "\n" + textwrap.dedent(remaining)
+    else:
+        return textwrap.dedent(text)
+
+
+def inject_docstring_text(
+    given_doctext: Optional[str], injecttext: str, pos: int
+) -> str:
+    doctext: str = _dedent_docstring(given_doctext or "")
+    lines = doctext.split("\n")
+    if len(lines) == 1:
+        lines.append("")
+    injectlines = textwrap.dedent(injecttext).split("\n")
+    if injectlines[0]:
+        injectlines.insert(0, "")
+
+    blanks = [num for num, line in enumerate(lines) if not line.strip()]
+    blanks.insert(0, 0)
+
+    inject_pos = blanks[min(pos, len(blanks) - 1)]
+
+    lines = lines[0:inject_pos] + injectlines + lines[inject_pos:]
+    return "\n".join(lines)
+
+
+_param_reg = re.compile(r"(\s+):param (.+?):")
+
+
+def inject_param_text(doctext: str, inject_params: Dict[str, str]) -> str:
+    doclines = collections.deque(doctext.splitlines())
+    lines = []
+
+    # TODO: this is not working for params like ":param case_sensitive=True:"
+
+    to_inject = None
+    while doclines:
+        line = doclines.popleft()
+
+        m = _param_reg.match(line)
+
+        if to_inject is None:
+            if m:
+                param = m.group(2).lstrip("*")
+                if param in inject_params:
+                    # default indent to that of :param: plus one
+                    indent = " " * len(m.group(1)) + " "
+
+                    # but if the next line has text, use that line's
+                    # indentation
+                    if doclines:
+                        m2 = re.match(r"(\s+)\S", doclines[0])
+                        if m2:
+                            indent = " " * len(m2.group(1))
+
+                    to_inject = indent + inject_params[param]
+        elif m:
+            lines.extend(["\n", to_inject, "\n"])
+            to_inject = None
+        elif not line.rstrip():
+            lines.extend([line, to_inject, "\n"])
+            to_inject = None
+        elif line.endswith("::"):
+            # TODO: this still won't cover if the code example itself has
+            # blank lines in it, need to detect those via indentation.
+            lines.extend([line, doclines.popleft()])
+            continue
+        lines.append(line)
+
+    return "\n".join(lines)
+
+
+def repr_tuple_names(names: List[str]) -> Optional[str]:
+    """Trims a list of strings from the middle and return a string of up to
+    four elements. Strings greater than 11 characters will be truncated"""
+    if len(names) == 0:
+        return None
+    flag = len(names) <= 4
+    names = names[0:4] if flag else names[0:3] + names[-1:]
+    res = ["%s.." % name[:11] if len(name) > 11 else name for name in names]
+    if flag:
+        return ", ".join(res)
+    else:
+        return "%s, ..., %s" % (", ".join(res[0:3]), res[-1])
+
+
+def has_compiled_ext(raise_=False):
+    if HAS_CYEXTENSION:
+        return True
+    elif raise_:
+        raise ImportError(
+            "cython extensions were expected to be installed, "
+            "but are not present"
+        )
+    else:
+        return False
+
+
+class _Missing(enum.Enum):
+    Missing = enum.auto()
+
+
+Missing = _Missing.Missing
+MissingOr = Union[_T, Literal[_Missing.Missing]]
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/preloaded.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/preloaded.py
new file mode 100644
index 00000000..4ea9aa90
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/preloaded.py
@@ -0,0 +1,150 @@
+# util/preloaded.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""supplies the "preloaded" registry to resolve circular module imports at
+runtime.
+
+"""
+from __future__ import annotations
+
+import sys
+from typing import Any
+from typing import Callable
+from typing import TYPE_CHECKING
+from typing import TypeVar
+
+_FN = TypeVar("_FN", bound=Callable[..., Any])
+
+
+if TYPE_CHECKING:
+    from sqlalchemy import dialects as _dialects
+    from sqlalchemy import orm as _orm
+    from sqlalchemy.engine import cursor as _engine_cursor
+    from sqlalchemy.engine import default as _engine_default
+    from sqlalchemy.engine import reflection as _engine_reflection
+    from sqlalchemy.engine import result as _engine_result
+    from sqlalchemy.engine import url as _engine_url
+    from sqlalchemy.orm import attributes as _orm_attributes
+    from sqlalchemy.orm import base as _orm_base
+    from sqlalchemy.orm import clsregistry as _orm_clsregistry
+    from sqlalchemy.orm import decl_api as _orm_decl_api
+    from sqlalchemy.orm import decl_base as _orm_decl_base
+    from sqlalchemy.orm import dependency as _orm_dependency
+    from sqlalchemy.orm import descriptor_props as _orm_descriptor_props
+    from sqlalchemy.orm import mapperlib as _orm_mapper
+    from sqlalchemy.orm import properties as _orm_properties
+    from sqlalchemy.orm import relationships as _orm_relationships
+    from sqlalchemy.orm import session as _orm_session
+    from sqlalchemy.orm import state as _orm_state
+    from sqlalchemy.orm import strategies as _orm_strategies
+    from sqlalchemy.orm import strategy_options as _orm_strategy_options
+    from sqlalchemy.orm import util as _orm_util
+    from sqlalchemy.sql import default_comparator as _sql_default_comparator
+    from sqlalchemy.sql import dml as _sql_dml
+    from sqlalchemy.sql import elements as _sql_elements
+    from sqlalchemy.sql import functions as _sql_functions
+    from sqlalchemy.sql import naming as _sql_naming
+    from sqlalchemy.sql import schema as _sql_schema
+    from sqlalchemy.sql import selectable as _sql_selectable
+    from sqlalchemy.sql import sqltypes as _sql_sqltypes
+    from sqlalchemy.sql import traversals as _sql_traversals
+    from sqlalchemy.sql import util as _sql_util
+
+    # sigh, appease mypy 0.971 which does not accept imports as instance
+    # variables of a module
+    dialects = _dialects
+    engine_cursor = _engine_cursor
+    engine_default = _engine_default
+    engine_reflection = _engine_reflection
+    engine_result = _engine_result
+    engine_url = _engine_url
+    orm_clsregistry = _orm_clsregistry
+    orm_base = _orm_base
+    orm = _orm
+    orm_attributes = _orm_attributes
+    orm_decl_api = _orm_decl_api
+    orm_decl_base = _orm_decl_base
+    orm_descriptor_props = _orm_descriptor_props
+    orm_dependency = _orm_dependency
+    orm_mapper = _orm_mapper
+    orm_properties = _orm_properties
+    orm_relationships = _orm_relationships
+    orm_session = _orm_session
+    orm_strategies = _orm_strategies
+    orm_strategy_options = _orm_strategy_options
+    orm_state = _orm_state
+    orm_util = _orm_util
+    sql_default_comparator = _sql_default_comparator
+    sql_dml = _sql_dml
+    sql_elements = _sql_elements
+    sql_functions = _sql_functions
+    sql_naming = _sql_naming
+    sql_selectable = _sql_selectable
+    sql_traversals = _sql_traversals
+    sql_schema = _sql_schema
+    sql_sqltypes = _sql_sqltypes
+    sql_util = _sql_util
+
+
+class _ModuleRegistry:
+    """Registry of modules to load in a package init file.
+
+    To avoid potential thread safety issues for imports that are deferred
+    in a function, like https://bugs.python.org/issue38884, these modules
+    are added to the system module cache by importing them after the packages
+    has finished initialization.
+
+    A global instance is provided under the name :attr:`.preloaded`. Use
+    the function :func:`.preload_module` to register modules to load and
+    :meth:`.import_prefix` to load all the modules that start with the
+    given path.
+
+    While the modules are loaded in the global module cache, it's advisable
+    to access them using :attr:`.preloaded` to ensure that it was actually
+    registered. Each registered module is added to the instance ``__dict__``
+    in the form `<package>_<module>`, omitting ``sqlalchemy`` from the package
+    name. Example: ``sqlalchemy.sql.util`` becomes ``preloaded.sql_util``.
+    """
+
+    def __init__(self, prefix="sqlalchemy."):
+        self.module_registry = set()
+        self.prefix = prefix
+
+    def preload_module(self, *deps: str) -> Callable[[_FN], _FN]:
+        """Adds the specified modules to the list to load.
+
+        This method can be used both as a normal function and as a decorator.
+        No change is performed to the decorated object.
+        """
+        self.module_registry.update(deps)
+        return lambda fn: fn
+
+    def import_prefix(self, path: str) -> None:
+        """Resolve all the modules in the registry that start with the
+        specified path.
+        """
+        for module in self.module_registry:
+            if self.prefix:
+                key = module.split(self.prefix)[-1].replace(".", "_")
+            else:
+                key = module
+            if (
+                not path or module.startswith(path)
+            ) and key not in self.__dict__:
+                __import__(module, globals(), locals())
+                self.__dict__[key] = globals()[key] = sys.modules[module]
+
+
+_reg = _ModuleRegistry()
+preload_module = _reg.preload_module
+import_prefix = _reg.import_prefix
+
+# this appears to do absolutely nothing for any version of mypy
+# if TYPE_CHECKING:
+#    def __getattr__(key: str) -> ModuleType:
+#        ...
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/queue.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/queue.py
new file mode 100644
index 00000000..3fb01a9a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/queue.py
@@ -0,0 +1,322 @@
+# util/queue.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+"""An adaptation of Py2.3/2.4's Queue module which supports reentrant
+behavior, using RLock instead of Lock for its mutex object.  The
+Queue object is used exclusively by the sqlalchemy.pool.QueuePool
+class.
+
+This is to support the connection pool's usage of weakref callbacks to return
+connections to the underlying Queue, which can in extremely
+rare cases be invoked within the ``get()`` method of the Queue itself,
+producing a ``put()`` inside the ``get()`` and therefore a reentrant
+condition.
+
+"""
+from __future__ import annotations
+
+import asyncio
+from collections import deque
+import threading
+from time import time as _time
+import typing
+from typing import Any
+from typing import Awaitable
+from typing import Deque
+from typing import Generic
+from typing import Optional
+from typing import TypeVar
+
+from .concurrency import await_fallback
+from .concurrency import await_only
+from .langhelpers import memoized_property
+
+
+_T = TypeVar("_T", bound=Any)
+__all__ = ["Empty", "Full", "Queue"]
+
+
+class Empty(Exception):
+    "Exception raised by Queue.get(block=0)/get_nowait()."
+
+    pass
+
+
+class Full(Exception):
+    "Exception raised by Queue.put(block=0)/put_nowait()."
+
+    pass
+
+
+class QueueCommon(Generic[_T]):
+    maxsize: int
+    use_lifo: bool
+
+    def __init__(self, maxsize: int = 0, use_lifo: bool = False): ...
+
+    def empty(self) -> bool:
+        raise NotImplementedError()
+
+    def full(self) -> bool:
+        raise NotImplementedError()
+
+    def qsize(self) -> int:
+        raise NotImplementedError()
+
+    def put_nowait(self, item: _T) -> None:
+        raise NotImplementedError()
+
+    def put(
+        self, item: _T, block: bool = True, timeout: Optional[float] = None
+    ) -> None:
+        raise NotImplementedError()
+
+    def get_nowait(self) -> _T:
+        raise NotImplementedError()
+
+    def get(self, block: bool = True, timeout: Optional[float] = None) -> _T:
+        raise NotImplementedError()
+
+
+class Queue(QueueCommon[_T]):
+    queue: Deque[_T]
+
+    def __init__(self, maxsize: int = 0, use_lifo: bool = False):
+        """Initialize a queue object with a given maximum size.
+
+        If `maxsize` is <= 0, the queue size is infinite.
+
+        If `use_lifo` is True, this Queue acts like a Stack (LIFO).
+        """
+
+        self._init(maxsize)
+        # mutex must be held whenever the queue is mutating.  All methods
+        # that acquire mutex must release it before returning.  mutex
+        # is shared between the two conditions, so acquiring and
+        # releasing the conditions also acquires and releases mutex.
+        self.mutex = threading.RLock()
+        # Notify not_empty whenever an item is added to the queue; a
+        # thread waiting to get is notified then.
+        self.not_empty = threading.Condition(self.mutex)
+        # Notify not_full whenever an item is removed from the queue;
+        # a thread waiting to put is notified then.
+        self.not_full = threading.Condition(self.mutex)
+        # If this queue uses LIFO or FIFO
+        self.use_lifo = use_lifo
+
+    def qsize(self) -> int:
+        """Return the approximate size of the queue (not reliable!)."""
+
+        with self.mutex:
+            return self._qsize()
+
+    def empty(self) -> bool:
+        """Return True if the queue is empty, False otherwise (not
+        reliable!)."""
+
+        with self.mutex:
+            return self._empty()
+
+    def full(self) -> bool:
+        """Return True if the queue is full, False otherwise (not
+        reliable!)."""
+
+        with self.mutex:
+            return self._full()
+
+    def put(
+        self, item: _T, block: bool = True, timeout: Optional[float] = None
+    ) -> None:
+        """Put an item into the queue.
+
+        If optional args `block` is True and `timeout` is None (the
+        default), block if necessary until a free slot is
+        available. If `timeout` is a positive number, it blocks at
+        most `timeout` seconds and raises the ``Full`` exception if no
+        free slot was available within that time.  Otherwise (`block`
+        is false), put an item on the queue if a free slot is
+        immediately available, else raise the ``Full`` exception
+        (`timeout` is ignored in that case).
+        """
+
+        with self.not_full:
+            if not block:
+                if self._full():
+                    raise Full
+            elif timeout is None:
+                while self._full():
+                    self.not_full.wait()
+            else:
+                if timeout < 0:
+                    raise ValueError("'timeout' must be a positive number")
+                endtime = _time() + timeout
+                while self._full():
+                    remaining = endtime - _time()
+                    if remaining <= 0.0:
+                        raise Full
+                    self.not_full.wait(remaining)
+            self._put(item)
+            self.not_empty.notify()
+
+    def put_nowait(self, item: _T) -> None:
+        """Put an item into the queue without blocking.
+
+        Only enqueue the item if a free slot is immediately available.
+        Otherwise raise the ``Full`` exception.
+        """
+        return self.put(item, False)
+
+    def get(self, block: bool = True, timeout: Optional[float] = None) -> _T:
+        """Remove and return an item from the queue.
+
+        If optional args `block` is True and `timeout` is None (the
+        default), block if necessary until an item is available. If
+        `timeout` is a positive number, it blocks at most `timeout`
+        seconds and raises the ``Empty`` exception if no item was
+        available within that time.  Otherwise (`block` is false),
+        return an item if one is immediately available, else raise the
+        ``Empty`` exception (`timeout` is ignored in that case).
+
+        """
+        with self.not_empty:
+            if not block:
+                if self._empty():
+                    raise Empty
+            elif timeout is None:
+                while self._empty():
+                    self.not_empty.wait()
+            else:
+                if timeout < 0:
+                    raise ValueError("'timeout' must be a positive number")
+                endtime = _time() + timeout
+                while self._empty():
+                    remaining = endtime - _time()
+                    if remaining <= 0.0:
+                        raise Empty
+                    self.not_empty.wait(remaining)
+            item = self._get()
+            self.not_full.notify()
+            return item
+
+    def get_nowait(self) -> _T:
+        """Remove and return an item from the queue without blocking.
+
+        Only get an item if one is immediately available. Otherwise
+        raise the ``Empty`` exception.
+        """
+
+        return self.get(False)
+
+    def _init(self, maxsize: int) -> None:
+        self.maxsize = maxsize
+        self.queue = deque()
+
+    def _qsize(self) -> int:
+        return len(self.queue)
+
+    def _empty(self) -> bool:
+        return not self.queue
+
+    def _full(self) -> bool:
+        return self.maxsize > 0 and len(self.queue) == self.maxsize
+
+    def _put(self, item: _T) -> None:
+        self.queue.append(item)
+
+    def _get(self) -> _T:
+        if self.use_lifo:
+            # LIFO
+            return self.queue.pop()
+        else:
+            # FIFO
+            return self.queue.popleft()
+
+
+class AsyncAdaptedQueue(QueueCommon[_T]):
+    if typing.TYPE_CHECKING:
+
+        @staticmethod
+        def await_(coroutine: Awaitable[Any]) -> _T: ...
+
+    else:
+        await_ = staticmethod(await_only)
+
+    def __init__(self, maxsize: int = 0, use_lifo: bool = False):
+        self.use_lifo = use_lifo
+        self.maxsize = maxsize
+
+    def empty(self) -> bool:
+        return self._queue.empty()
+
+    def full(self):
+        return self._queue.full()
+
+    def qsize(self):
+        return self._queue.qsize()
+
+    @memoized_property
+    def _queue(self) -> asyncio.Queue[_T]:
+        # Delay creation of the queue until it is first used, to avoid
+        # binding it to a possibly wrong event loop.
+        # By delaying the creation of the pool we accommodate the common
+        # usage pattern of instantiating the engine at module level, where a
+        # different event loop is in present compared to when the application
+        # is actually run.
+
+        queue: asyncio.Queue[_T]
+
+        if self.use_lifo:
+            queue = asyncio.LifoQueue(maxsize=self.maxsize)
+        else:
+            queue = asyncio.Queue(maxsize=self.maxsize)
+        return queue
+
+    def put_nowait(self, item: _T) -> None:
+        try:
+            self._queue.put_nowait(item)
+        except asyncio.QueueFull as err:
+            raise Full() from err
+
+    def put(
+        self, item: _T, block: bool = True, timeout: Optional[float] = None
+    ) -> None:
+        if not block:
+            return self.put_nowait(item)
+
+        try:
+            if timeout is not None:
+                self.await_(asyncio.wait_for(self._queue.put(item), timeout))
+            else:
+                self.await_(self._queue.put(item))
+        except (asyncio.QueueFull, asyncio.TimeoutError) as err:
+            raise Full() from err
+
+    def get_nowait(self) -> _T:
+        try:
+            return self._queue.get_nowait()
+        except asyncio.QueueEmpty as err:
+            raise Empty() from err
+
+    def get(self, block: bool = True, timeout: Optional[float] = None) -> _T:
+        if not block:
+            return self.get_nowait()
+
+        try:
+            if timeout is not None:
+                return self.await_(
+                    asyncio.wait_for(self._queue.get(), timeout)
+                )
+            else:
+                return self.await_(self._queue.get())
+        except (asyncio.QueueEmpty, asyncio.TimeoutError) as err:
+            raise Empty() from err
+
+
+class FallbackAsyncAdaptedQueue(AsyncAdaptedQueue[_T]):
+    if not typing.TYPE_CHECKING:
+        await_ = staticmethod(await_fallback)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/tool_support.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/tool_support.py
new file mode 100644
index 00000000..407c2d45
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/tool_support.py
@@ -0,0 +1,201 @@
+# util/tool_support.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+"""support routines for the helpers in tools/.
+
+These aren't imported by the enclosing util package as the are not
+needed for normal library use.
+
+"""
+from __future__ import annotations
+
+from argparse import ArgumentParser
+from argparse import Namespace
+import contextlib
+import difflib
+import os
+from pathlib import Path
+import shlex
+import shutil
+import subprocess
+import sys
+from typing import Any
+from typing import Dict
+from typing import Iterator
+from typing import Optional
+from typing import Union
+
+from . import compat
+
+
+class code_writer_cmd:
+    parser: ArgumentParser
+    args: Namespace
+    suppress_output: bool
+    diffs_detected: bool
+    source_root: Path
+    pyproject_toml_path: Path
+
+    def __init__(self, tool_script: str):
+        self.source_root = Path(tool_script).parent.parent
+        self.pyproject_toml_path = self.source_root / Path("pyproject.toml")
+        assert self.pyproject_toml_path.exists()
+
+        self.parser = ArgumentParser()
+        self.parser.add_argument(
+            "--stdout",
+            action="store_true",
+            help="Write to stdout instead of saving to file",
+        )
+        self.parser.add_argument(
+            "-c",
+            "--check",
+            help="Don't write the files back, just return the "
+            "status. Return code 0 means nothing would change. "
+            "Return code 1 means some files would be reformatted",
+            action="store_true",
+        )
+
+    def run_zimports(self, tempfile: str) -> None:
+        self._run_console_script(
+            str(tempfile),
+            {
+                "entrypoint": "zimports",
+                "options": f"--toml-config {self.pyproject_toml_path}",
+            },
+        )
+
+    def run_black(self, tempfile: str) -> None:
+        self._run_console_script(
+            str(tempfile),
+            {
+                "entrypoint": "black",
+                "options": f"--config {self.pyproject_toml_path}",
+            },
+        )
+
+    def _run_console_script(self, path: str, options: Dict[str, Any]) -> None:
+        """Run a Python console application from within the process.
+
+        Used for black, zimports
+
+        """
+
+        is_posix = os.name == "posix"
+
+        entrypoint_name = options["entrypoint"]
+
+        for entry in compat.importlib_metadata_get("console_scripts"):
+            if entry.name == entrypoint_name:
+                impl = entry
+                break
+        else:
+            raise Exception(
+                f"Could not find entrypoint console_scripts.{entrypoint_name}"
+            )
+        cmdline_options_str = options.get("options", "")
+        cmdline_options_list = shlex.split(
+            cmdline_options_str, posix=is_posix
+        ) + [path]
+
+        kw: Dict[str, Any] = {}
+        if self.suppress_output:
+            kw["stdout"] = kw["stderr"] = subprocess.DEVNULL
+
+        subprocess.run(
+            [
+                sys.executable,
+                "-c",
+                "import %s; %s.%s()" % (impl.module, impl.module, impl.attr),
+            ]
+            + cmdline_options_list,
+            cwd=str(self.source_root),
+            **kw,
+        )
+
+    def write_status(self, *text: str) -> None:
+        if not self.suppress_output:
+            sys.stderr.write(" ".join(text))
+
+    def write_output_file_from_text(
+        self, text: str, destination_path: Union[str, Path]
+    ) -> None:
+        if self.args.check:
+            self._run_diff(destination_path, source=text)
+        elif self.args.stdout:
+            print(text)
+        else:
+            self.write_status(f"Writing {destination_path}...")
+            Path(destination_path).write_text(
+                text, encoding="utf-8", newline="\n"
+            )
+            self.write_status("done\n")
+
+    def write_output_file_from_tempfile(
+        self, tempfile: str, destination_path: str
+    ) -> None:
+        if self.args.check:
+            self._run_diff(destination_path, source_file=tempfile)
+            os.unlink(tempfile)
+        elif self.args.stdout:
+            with open(tempfile) as tf:
+                print(tf.read())
+            os.unlink(tempfile)
+        else:
+            self.write_status(f"Writing {destination_path}...")
+            shutil.move(tempfile, destination_path)
+            self.write_status("done\n")
+
+    def _run_diff(
+        self,
+        destination_path: Union[str, Path],
+        *,
+        source: Optional[str] = None,
+        source_file: Optional[str] = None,
+    ) -> None:
+        if source_file:
+            with open(source_file, encoding="utf-8") as tf:
+                source_lines = list(tf)
+        elif source is not None:
+            source_lines = source.splitlines(keepends=True)
+        else:
+            assert False, "source or source_file is required"
+
+        with open(destination_path, encoding="utf-8") as dp:
+            d = difflib.unified_diff(
+                list(dp),
+                source_lines,
+                fromfile=Path(destination_path).as_posix(),
+                tofile="<proposed changes>",
+                n=3,
+                lineterm="\n",
+            )
+            d_as_list = list(d)
+            if d_as_list:
+                self.diffs_detected = True
+                print("".join(d_as_list))
+
+    @contextlib.contextmanager
+    def add_arguments(self) -> Iterator[ArgumentParser]:
+        yield self.parser
+
+    @contextlib.contextmanager
+    def run_program(self) -> Iterator[None]:
+        self.args = self.parser.parse_args()
+        if self.args.check:
+            self.diffs_detected = False
+            self.suppress_output = True
+        elif self.args.stdout:
+            self.suppress_output = True
+        else:
+            self.suppress_output = False
+        yield
+
+        if self.args.check and self.diffs_detected:
+            sys.exit(1)
+        else:
+            sys.exit(0)
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/topological.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/topological.py
new file mode 100644
index 00000000..393c855a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/topological.py
@@ -0,0 +1,120 @@
+# util/topological.py
+# Copyright (C) 2005-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+
+"""Topological sorting algorithms."""
+
+from __future__ import annotations
+
+from typing import Any
+from typing import Collection
+from typing import DefaultDict
+from typing import Iterable
+from typing import Iterator
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+from typing import TypeVar
+
+from .. import util
+from ..exc import CircularDependencyError
+
+_T = TypeVar("_T", bound=Any)
+
+__all__ = ["sort", "sort_as_subsets", "find_cycles"]
+
+
+def sort_as_subsets(
+    tuples: Collection[Tuple[_T, _T]], allitems: Collection[_T]
+) -> Iterator[Sequence[_T]]:
+    edges: DefaultDict[_T, Set[_T]] = util.defaultdict(set)
+    for parent, child in tuples:
+        edges[child].add(parent)
+
+    todo = list(allitems)
+    todo_set = set(allitems)
+
+    while todo_set:
+        output = []
+        for node in todo:
+            if todo_set.isdisjoint(edges[node]):
+                output.append(node)
+
+        if not output:
+            raise CircularDependencyError(
+                "Circular dependency detected.",
+                find_cycles(tuples, allitems),
+                _gen_edges(edges),
+            )
+
+        todo_set.difference_update(output)
+        todo = [t for t in todo if t in todo_set]
+        yield output
+
+
+def sort(
+    tuples: Collection[Tuple[_T, _T]],
+    allitems: Collection[_T],
+    deterministic_order: bool = True,
+) -> Iterator[_T]:
+    """sort the given list of items by dependency.
+
+    'tuples' is a list of tuples representing a partial ordering.
+
+    deterministic_order is no longer used, the order is now always
+    deterministic given the order of "allitems".    the flag is there
+    for backwards compatibility with Alembic.
+
+    """
+
+    for set_ in sort_as_subsets(tuples, allitems):
+        yield from set_
+
+
+def find_cycles(
+    tuples: Iterable[Tuple[_T, _T]], allitems: Iterable[_T]
+) -> Set[_T]:
+    # adapted from:
+    # https://neopythonic.blogspot.com/2009/01/detecting-cycles-in-directed-graph.html
+
+    edges: DefaultDict[_T, Set[_T]] = util.defaultdict(set)
+    for parent, child in tuples:
+        edges[parent].add(child)
+    nodes_to_test = set(edges)
+
+    output = set()
+
+    # we'd like to find all nodes that are
+    # involved in cycles, so we do the full
+    # pass through the whole thing for each
+    # node in the original list.
+
+    # we can go just through parent edge nodes.
+    # if a node is only a child and never a parent,
+    # by definition it can't be part of a cycle.  same
+    # if it's not in the edges at all.
+    for node in nodes_to_test:
+        stack = [node]
+        todo = nodes_to_test.difference(stack)
+        while stack:
+            top = stack[-1]
+            for node in edges[top]:
+                if node in stack:
+                    cyc = stack[stack.index(node) :]
+                    todo.difference_update(cyc)
+                    output.update(cyc)
+
+                if node in todo:
+                    stack.append(node)
+                    todo.remove(node)
+                    break
+            else:
+                node = stack.pop()
+    return output
+
+
+def _gen_edges(edges: DefaultDict[_T, Set[_T]]) -> Set[Tuple[_T, _T]]:
+    return {(right, left) for left in edges for right in edges[left]}
diff --git a/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/typing.py b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/typing.py
new file mode 100644
index 00000000..62fd47c6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/sqlalchemy/util/typing.py
@@ -0,0 +1,689 @@
+# util/typing.py
+# Copyright (C) 2022-2025 the SQLAlchemy authors and contributors
+# <see AUTHORS file>
+#
+# This module is part of SQLAlchemy and is released under
+# the MIT License: https://www.opensource.org/licenses/mit-license.php
+# mypy: allow-untyped-defs, allow-untyped-calls
+
+from __future__ import annotations
+
+import builtins
+from collections import deque
+import collections.abc as collections_abc
+import re
+import sys
+import typing
+from typing import Any
+from typing import Callable
+from typing import cast
+from typing import Dict
+from typing import ForwardRef
+from typing import Generic
+from typing import Iterable
+from typing import Mapping
+from typing import NewType
+from typing import NoReturn
+from typing import Optional
+from typing import overload
+from typing import Set
+from typing import Tuple
+from typing import Type
+from typing import TYPE_CHECKING
+from typing import TypeVar
+from typing import Union
+
+from . import compat
+
+if True:  # zimports removes the tailing comments
+    from typing_extensions import Annotated as Annotated  # 3.8
+    from typing_extensions import Concatenate as Concatenate  # 3.10
+    from typing_extensions import (
+        dataclass_transform as dataclass_transform,  # 3.11,
+    )
+    from typing_extensions import Final as Final  # 3.8
+    from typing_extensions import final as final  # 3.8
+    from typing_extensions import get_args as get_args  # 3.10
+    from typing_extensions import get_origin as get_origin  # 3.10
+    from typing_extensions import Literal as Literal  # 3.8
+    from typing_extensions import NotRequired as NotRequired  # 3.11
+    from typing_extensions import ParamSpec as ParamSpec  # 3.10
+    from typing_extensions import Protocol as Protocol  # 3.8
+    from typing_extensions import SupportsIndex as SupportsIndex  # 3.8
+    from typing_extensions import TypeAlias as TypeAlias  # 3.10
+    from typing_extensions import TypedDict as TypedDict  # 3.8
+    from typing_extensions import TypeGuard as TypeGuard  # 3.10
+    from typing_extensions import Self as Self  # 3.11
+    from typing_extensions import TypeAliasType as TypeAliasType  # 3.12
+    from typing_extensions import Never as Never  # 3.11
+
+_T = TypeVar("_T", bound=Any)
+_KT = TypeVar("_KT")
+_KT_co = TypeVar("_KT_co", covariant=True)
+_KT_contra = TypeVar("_KT_contra", contravariant=True)
+_VT = TypeVar("_VT")
+_VT_co = TypeVar("_VT_co", covariant=True)
+
+if compat.py38:
+    # typing_extensions.Literal is different from typing.Literal until
+    # Python 3.10.1
+    LITERAL_TYPES = frozenset([typing.Literal, Literal])
+else:
+    LITERAL_TYPES = frozenset([Literal])
+
+
+if compat.py310:
+    # why they took until py310 to put this in stdlib is beyond me,
+    # I've been wanting it since py27
+    from types import NoneType as NoneType
+else:
+    NoneType = type(None)  # type: ignore
+
+NoneFwd = ForwardRef("None")
+
+
+_AnnotationScanType = Union[
+    Type[Any], str, ForwardRef, NewType, TypeAliasType, "GenericProtocol[Any]"
+]
+
+
+class ArgsTypeProtocol(Protocol):
+    """protocol for types that have ``__args__``
+
+    there's no public interface for this AFAIK
+
+    """
+
+    __args__: Tuple[_AnnotationScanType, ...]
+
+
+class GenericProtocol(Protocol[_T]):
+    """protocol for generic types.
+
+    this since Python.typing _GenericAlias is private
+
+    """
+
+    __args__: Tuple[_AnnotationScanType, ...]
+    __origin__: Type[_T]
+
+    # Python's builtin _GenericAlias has this method, however builtins like
+    # list, dict, etc. do not, even though they have ``__origin__`` and
+    # ``__args__``
+    #
+    # def copy_with(self, params: Tuple[_AnnotationScanType, ...]) -> Type[_T]:
+    #     ...
+
+
+# copied from TypeShed, required in order to implement
+# MutableMapping.update()
+class SupportsKeysAndGetItem(Protocol[_KT, _VT_co]):
+    def keys(self) -> Iterable[_KT]: ...
+
+    def __getitem__(self, __k: _KT) -> _VT_co: ...
+
+
+# work around https://github.com/microsoft/pyright/issues/3025
+_LiteralStar = Literal["*"]
+
+
+def de_stringify_annotation(
+    cls: Type[Any],
+    annotation: _AnnotationScanType,
+    originating_module: str,
+    locals_: Mapping[str, Any],
+    *,
+    str_cleanup_fn: Optional[Callable[[str, str], str]] = None,
+    include_generic: bool = False,
+    _already_seen: Optional[Set[Any]] = None,
+) -> Type[Any]:
+    """Resolve annotations that may be string based into real objects.
+
+    This is particularly important if a module defines "from __future__ import
+    annotations", as everything inside of __annotations__ is a string. We want
+    to at least have generic containers like ``Mapped``, ``Union``, ``List``,
+    etc.
+
+    """
+    # looked at typing.get_type_hints(), looked at pydantic.  We need much
+    # less here, and we here try to not use any private typing internals
+    # or construct ForwardRef objects which is documented as something
+    # that should be avoided.
+
+    original_annotation = annotation
+
+    if is_fwd_ref(annotation):
+        annotation = annotation.__forward_arg__
+
+    if isinstance(annotation, str):
+        if str_cleanup_fn:
+            annotation = str_cleanup_fn(annotation, originating_module)
+
+        annotation = eval_expression(
+            annotation, originating_module, locals_=locals_, in_class=cls
+        )
+
+    if (
+        include_generic
+        and is_generic(annotation)
+        and not is_literal(annotation)
+    ):
+        if _already_seen is None:
+            _already_seen = set()
+
+        if annotation in _already_seen:
+            # only occurs recursively.  outermost return type
+            # will always be Type.
+            # the element here will be either ForwardRef or
+            # Optional[ForwardRef]
+            return original_annotation  # type: ignore
+        else:
+            _already_seen.add(annotation)
+
+        elements = tuple(
+            de_stringify_annotation(
+                cls,
+                elem,
+                originating_module,
+                locals_,
+                str_cleanup_fn=str_cleanup_fn,
+                include_generic=include_generic,
+                _already_seen=_already_seen,
+            )
+            for elem in annotation.__args__
+        )
+
+        return _copy_generic_annotation_with(annotation, elements)
+
+    return annotation  # type: ignore
+
+
+def fixup_container_fwd_refs(
+    type_: _AnnotationScanType,
+) -> _AnnotationScanType:
+    """Correct dict['x', 'y'] into dict[ForwardRef('x'), ForwardRef('y')]
+    and similar for list, set
+
+    """
+
+    if (
+        is_generic(type_)
+        and get_origin(type_)
+        in (
+            dict,
+            set,
+            list,
+            collections_abc.MutableSet,
+            collections_abc.MutableMapping,
+            collections_abc.MutableSequence,
+            collections_abc.Mapping,
+            collections_abc.Sequence,
+        )
+        # fight, kick and scream to struggle to tell the difference between
+        # dict[] and typing.Dict[] which DO NOT compare the same and DO NOT
+        # behave the same yet there is NO WAY to distinguish between which type
+        # it is using public attributes
+        and not re.match(
+            "typing.(?:Dict|List|Set|.*Mapping|.*Sequence|.*Set)", repr(type_)
+        )
+    ):
+        # compat with py3.10 and earlier
+        return get_origin(type_).__class_getitem__(  # type: ignore
+            tuple(
+                [
+                    ForwardRef(elem) if isinstance(elem, str) else elem
+                    for elem in get_args(type_)
+                ]
+            )
+        )
+    return type_
+
+
+def _copy_generic_annotation_with(
+    annotation: GenericProtocol[_T], elements: Tuple[_AnnotationScanType, ...]
+) -> Type[_T]:
+    if hasattr(annotation, "copy_with"):
+        # List, Dict, etc. real generics
+        return annotation.copy_with(elements)  # type: ignore
+    else:
+        # Python builtins list, dict, etc.
+        return annotation.__origin__[elements]  # type: ignore
+
+
+def eval_expression(
+    expression: str,
+    module_name: str,
+    *,
+    locals_: Optional[Mapping[str, Any]] = None,
+    in_class: Optional[Type[Any]] = None,
+) -> Any:
+    try:
+        base_globals: Dict[str, Any] = sys.modules[module_name].__dict__
+    except KeyError as ke:
+        raise NameError(
+            f"Module {module_name} isn't present in sys.modules; can't "
+            f"evaluate expression {expression}"
+        ) from ke
+
+    try:
+        if in_class is not None:
+            cls_namespace = dict(in_class.__dict__)
+            cls_namespace.setdefault(in_class.__name__, in_class)
+
+            # see #10899.  We want the locals/globals to take precedence
+            # over the class namespace in this context, even though this
+            # is not the usual way variables would resolve.
+            cls_namespace.update(base_globals)
+
+            annotation = eval(expression, cls_namespace, locals_)
+        else:
+            annotation = eval(expression, base_globals, locals_)
+    except Exception as err:
+        raise NameError(
+            f"Could not de-stringify annotation {expression!r}"
+        ) from err
+    else:
+        return annotation
+
+
+def eval_name_only(
+    name: str,
+    module_name: str,
+    *,
+    locals_: Optional[Mapping[str, Any]] = None,
+) -> Any:
+    if "." in name:
+        return eval_expression(name, module_name, locals_=locals_)
+
+    try:
+        base_globals: Dict[str, Any] = sys.modules[module_name].__dict__
+    except KeyError as ke:
+        raise NameError(
+            f"Module {module_name} isn't present in sys.modules; can't "
+            f"resolve name {name}"
+        ) from ke
+
+    # name only, just look in globals.  eval() works perfectly fine here,
+    # however we are seeking to have this be faster, as this occurs for
+    # every Mapper[] keyword, etc. depending on configuration
+    try:
+        return base_globals[name]
+    except KeyError as ke:
+        # check in builtins as well to handle `list`, `set` or `dict`, etc.
+        try:
+            return builtins.__dict__[name]
+        except KeyError:
+            pass
+
+        raise NameError(
+            f"Could not locate name {name} in module {module_name}"
+        ) from ke
+
+
+def resolve_name_to_real_class_name(name: str, module_name: str) -> str:
+    try:
+        obj = eval_name_only(name, module_name)
+    except NameError:
+        return name
+    else:
+        return getattr(obj, "__name__", name)
+
+
+def is_pep593(type_: Optional[Any]) -> bool:
+    return type_ is not None and get_origin(type_) is Annotated
+
+
+def is_non_string_iterable(obj: Any) -> TypeGuard[Iterable[Any]]:
+    return isinstance(obj, collections_abc.Iterable) and not isinstance(
+        obj, (str, bytes)
+    )
+
+
+def is_literal(type_: Any) -> bool:
+    return get_origin(type_) in LITERAL_TYPES
+
+
+def is_newtype(type_: Optional[_AnnotationScanType]) -> TypeGuard[NewType]:
+    return hasattr(type_, "__supertype__")
+
+    # doesn't work in 3.8, 3.7 as it passes a closure, not an
+    # object instance
+    # return isinstance(type_, NewType)
+
+
+def is_generic(type_: _AnnotationScanType) -> TypeGuard[GenericProtocol[Any]]:
+    return hasattr(type_, "__args__") and hasattr(type_, "__origin__")
+
+
+def is_pep695(type_: _AnnotationScanType) -> TypeGuard[TypeAliasType]:
+    return isinstance(type_, TypeAliasType)
+
+
+def flatten_newtype(type_: NewType) -> Type[Any]:
+    super_type = type_.__supertype__
+    while is_newtype(super_type):
+        super_type = super_type.__supertype__
+    return super_type  # type: ignore[return-value]
+
+
+def pep695_values(type_: _AnnotationScanType) -> Set[Any]:
+    """Extracts the value from a TypeAliasType, recursively exploring unions
+    and inner TypeAliasType to flatten them into a single set.
+
+    Forward references are not evaluated, so no recursive exploration happens
+    into them.
+    """
+    _seen = set()
+
+    def recursive_value(type_):
+        if type_ in _seen:
+            # recursion are not supported (at least it's flagged as
+            # an error by pyright). Just avoid infinite loop
+            return type_
+        _seen.add(type_)
+        if not is_pep695(type_):
+            return type_
+        value = type_.__value__
+        if not is_union(value):
+            return value
+        return [recursive_value(t) for t in value.__args__]
+
+    res = recursive_value(type_)
+    if isinstance(res, list):
+        types = set()
+        stack = deque(res)
+        while stack:
+            t = stack.popleft()
+            if isinstance(t, list):
+                stack.extend(t)
+            else:
+                types.add(None if t in {NoneType, NoneFwd} else t)
+        return types
+    else:
+        return {res}
+
+
+def is_fwd_ref(
+    type_: _AnnotationScanType,
+    check_generic: bool = False,
+    check_for_plain_string: bool = False,
+) -> TypeGuard[ForwardRef]:
+    if check_for_plain_string and isinstance(type_, str):
+        return True
+    elif isinstance(type_, ForwardRef):
+        return True
+    elif check_generic and is_generic(type_):
+        return any(
+            is_fwd_ref(
+                arg, True, check_for_plain_string=check_for_plain_string
+            )
+            for arg in type_.__args__
+        )
+    else:
+        return False
+
+
+@overload
+def de_optionalize_union_types(type_: str) -> str: ...
+
+
+@overload
+def de_optionalize_union_types(type_: Type[Any]) -> Type[Any]: ...
+
+
+@overload
+def de_optionalize_union_types(
+    type_: _AnnotationScanType,
+) -> _AnnotationScanType: ...
+
+
+def de_optionalize_union_types(
+    type_: _AnnotationScanType,
+) -> _AnnotationScanType:
+    """Given a type, filter out ``Union`` types that include ``NoneType``
+    to not include the ``NoneType``.
+
+    Contains extra logic to work on non-flattened unions, unions that contain
+    ``None`` (seen in py38, 37)
+
+    """
+
+    if is_fwd_ref(type_):
+        return _de_optionalize_fwd_ref_union_types(type_, False)
+
+    elif is_union(type_) and includes_none(type_):
+        if compat.py39:
+            typ = set(type_.__args__)
+        else:
+            # py38, 37 - unions are not automatically flattened, can contain
+            # None rather than NoneType
+            stack_of_unions = deque([type_])
+            typ = set()
+            while stack_of_unions:
+                u_typ = stack_of_unions.popleft()
+                for elem in u_typ.__args__:
+                    if is_union(elem):
+                        stack_of_unions.append(elem)
+                    else:
+                        typ.add(elem)
+
+            typ.discard(None)  # type: ignore
+
+        typ.discard(NoneType)
+        typ.discard(NoneFwd)
+
+        return make_union_type(*typ)
+
+    else:
+        return type_
+
+
+@overload
+def _de_optionalize_fwd_ref_union_types(
+    type_: ForwardRef, return_has_none: Literal[True]
+) -> bool: ...
+
+
+@overload
+def _de_optionalize_fwd_ref_union_types(
+    type_: ForwardRef, return_has_none: Literal[False]
+) -> _AnnotationScanType: ...
+
+
+def _de_optionalize_fwd_ref_union_types(
+    type_: ForwardRef, return_has_none: bool
+) -> Union[_AnnotationScanType, bool]:
+    """return the non-optional type for Optional[], Union[None, ...], x|None,
+    etc. without de-stringifying forward refs.
+
+    unfortunately this seems to require lots of hardcoded heuristics
+
+    """
+
+    annotation = type_.__forward_arg__
+
+    mm = re.match(r"^(.+?)\[(.+)\]$", annotation)
+    if mm:
+        g1 = mm.group(1).split(".")[-1]
+        if g1 == "Optional":
+            return True if return_has_none else ForwardRef(mm.group(2))
+        elif g1 == "Union":
+            if "[" in mm.group(2):
+                # cases like "Union[Dict[str, int], int, None]"
+                elements: list[str] = []
+                current: list[str] = []
+                ignore_comma = 0
+                for char in mm.group(2):
+                    if char == "[":
+                        ignore_comma += 1
+                    elif char == "]":
+                        ignore_comma -= 1
+                    elif ignore_comma == 0 and char == ",":
+                        elements.append("".join(current).strip())
+                        current.clear()
+                        continue
+                    current.append(char)
+            else:
+                elements = re.split(r",\s*", mm.group(2))
+            parts = [ForwardRef(elem) for elem in elements if elem != "None"]
+            if return_has_none:
+                return len(elements) != len(parts)
+            else:
+                return make_union_type(*parts) if parts else Never  # type: ignore[return-value] # noqa: E501
+        else:
+            return False if return_has_none else type_
+
+    pipe_tokens = re.split(r"\s*\|\s*", annotation)
+    has_none = "None" in pipe_tokens
+    if return_has_none:
+        return has_none
+    if has_none:
+        anno_str = "|".join(p for p in pipe_tokens if p != "None")
+        return ForwardRef(anno_str) if anno_str else Never  # type: ignore[return-value] # noqa: E501
+
+    return type_
+
+
+def make_union_type(*types: _AnnotationScanType) -> Type[Any]:
+    """Make a Union type."""
+    return Union.__getitem__(types)  # type: ignore
+
+
+def includes_none(type_: Any) -> bool:
+    """Returns if the type annotation ``type_`` allows ``None``.
+
+    This function supports:
+    * forward refs
+    * unions
+    * pep593 - Annotated
+    * pep695 - TypeAliasType (does not support looking into
+    fw reference of other pep695)
+    * NewType
+    * plain types like ``int``, ``None``, etc
+    """
+    if is_fwd_ref(type_):
+        return _de_optionalize_fwd_ref_union_types(type_, True)
+    if is_union(type_):
+        return any(includes_none(t) for t in get_args(type_))
+    if is_pep593(type_):
+        return includes_none(get_args(type_)[0])
+    if is_pep695(type_):
+        return any(includes_none(t) for t in pep695_values(type_))
+    if is_newtype(type_):
+        return includes_none(type_.__supertype__)
+    return type_ in (NoneFwd, NoneType, None)
+
+
+def is_union(type_: Any) -> TypeGuard[ArgsTypeProtocol]:
+    return is_origin_of(type_, "Union", "UnionType")
+
+
+def is_origin_of_cls(
+    type_: Any, class_obj: Union[Tuple[Type[Any], ...], Type[Any]]
+) -> bool:
+    """return True if the given type has an __origin__ that shares a base
+    with the given class"""
+
+    origin = get_origin(type_)
+    if origin is None:
+        return False
+
+    return isinstance(origin, type) and issubclass(origin, class_obj)
+
+
+def is_origin_of(
+    type_: Any, *names: str, module: Optional[str] = None
+) -> bool:
+    """return True if the given type has an __origin__ with the given name
+    and optional module."""
+
+    origin = get_origin(type_)
+    if origin is None:
+        return False
+
+    return _get_type_name(origin) in names and (
+        module is None or origin.__module__.startswith(module)
+    )
+
+
+def _get_type_name(type_: Type[Any]) -> str:
+    if compat.py310:
+        return type_.__name__
+    else:
+        typ_name = getattr(type_, "__name__", None)
+        if typ_name is None:
+            typ_name = getattr(type_, "_name", None)
+
+        return typ_name  # type: ignore
+
+
+class DescriptorProto(Protocol):
+    def __get__(self, instance: object, owner: Any) -> Any: ...
+
+    def __set__(self, instance: Any, value: Any) -> None: ...
+
+    def __delete__(self, instance: Any) -> None: ...
+
+
+_DESC = TypeVar("_DESC", bound=DescriptorProto)
+
+
+class DescriptorReference(Generic[_DESC]):
+    """a descriptor that refers to a descriptor.
+
+    used for cases where we need to have an instance variable referring to an
+    object that is itself a descriptor, which typically confuses typing tools
+    as they don't know when they should use ``__get__`` or not when referring
+    to the descriptor assignment as an instance variable. See
+    sqlalchemy.orm.interfaces.PropComparator.prop
+
+    """
+
+    if TYPE_CHECKING:
+
+        def __get__(self, instance: object, owner: Any) -> _DESC: ...
+
+        def __set__(self, instance: Any, value: _DESC) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
+
+
+_DESC_co = TypeVar("_DESC_co", bound=DescriptorProto, covariant=True)
+
+
+class RODescriptorReference(Generic[_DESC_co]):
+    """a descriptor that refers to a descriptor.
+
+    same as :class:`.DescriptorReference` but is read-only, so that subclasses
+    can define a subtype as the generically contained element
+
+    """
+
+    if TYPE_CHECKING:
+
+        def __get__(self, instance: object, owner: Any) -> _DESC_co: ...
+
+        def __set__(self, instance: Any, value: Any) -> NoReturn: ...
+
+        def __delete__(self, instance: Any) -> NoReturn: ...
+
+
+_FN = TypeVar("_FN", bound=Optional[Callable[..., Any]])
+
+
+class CallableReference(Generic[_FN]):
+    """a descriptor that refers to a callable.
+
+    works around mypy's limitation of not allowing callables assigned
+    as instance variables
+
+
+    """
+
+    if TYPE_CHECKING:
+
+        def __get__(self, instance: object, owner: Any) -> _FN: ...
+
+        def __set__(self, instance: Any, value: _FN) -> None: ...
+
+        def __delete__(self, instance: Any) -> None: ...
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/LICENSE.TXT b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/LICENSE.TXT
new file mode 100644
index 00000000..eac792c2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/LICENSE.TXT
@@ -0,0 +1,45 @@
+Copyright (c) 2012, Manfred Moitzi
+
+Permission is hereby granted, free of charge, to any person obtaining a 
+copy of this software and associated documentation files (the 
+"Software"), to deal in the Software without restriction, including 
+without limitation the rights to use, copy, modify, merge, publish, 
+distribute, sublicense, and/or sell copies of the Software, and to 
+permit persons to whom the Software is furnished to do so, subject to 
+the following conditions: 
+
+The above copyright notice and this permission notice shall be included 
+in all copies or substantial portions of the Software. 
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS 
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. 
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY 
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, 
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE 
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 
+
+Deutsche Übersetzung:
+
+Copyright (c) 2012, Manfred Moitzi
+
+Hiermit wird unentgeltlich, jeder Person, die eine Kopie der Software 
+und der zugehörigen Dokumentationen (die "Software") erhält, die 
+Erlaubnis erteilt, uneingeschränkt zu benutzen, inklusive und ohne 
+Ausnahme, dem Recht, sie zu verwenden, kopieren, ändern, fusionieren, 
+verlegen, verbreiten, unterlizenzieren und/oder zu verkaufen, und 
+Personen, die diese Software erhalten, diese Rechte zu geben, unter den 
+folgenden Bedingungen: 
+
+Der obige Urheberrechtsvermerk und dieser Erlaubnisvermerk sind in allen 
+Kopien oder Teilkopien der Software beizulegen. 
+
+DIE SOFTWARE WIRD OHNE JEDE AUSDRÜCKLICHE ODER IMPLIZIERTE GARANTIE 
+BEREITGESTELLT, EINSCHLIESSLICH DER GARANTIE ZUR BENUTZUNG FÜR DEN 
+VORGESEHENEN ODER EINEM BESTIMMTEN ZWECK SOWIE JEGLICHER 
+RECHTSVERLETZUNG, JEDOCH NICHT DARAUF BESCHRÄNKT. IN KEINEM FALL SIND 
+DIE AUTOREN ODER COPYRIGHTINHABER FÜR JEGLICHEN SCHADEN ODER SONSTIGE 
+ANSPRÜCHE HAFTBAR ZU MACHEN, OB INFOLGE DER ERFÜLLUNG EINES VERTRAGES, 
+EINES DELIKTES ODER ANDERS IM ZUSAMMENHANG MIT DER SOFTWARE ODER 
+SONSTIGER VERWENDUNG DER SOFTWARE ENTSTANDEN. 
+
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/METADATA
new file mode 100644
index 00000000..24241123
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/METADATA
@@ -0,0 +1,285 @@
+Metadata-Version: 2.1
+Name: svgwrite
+Version: 1.4.3
+Summary: A Python library to create SVG drawings.
+Home-page: http://github.com/mozman/svgwrite.git
+Author: Manfred Moitzi
+Author-email: me@mozman.at
+License: MIT License
+Download-URL: http://github.com/mozman/svgwrite/releases
+Platform: OS Independent
+Classifier: Development Status :: 7 - Inactive
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Intended Audience :: Developers
+Classifier: Topic :: Multimedia :: Graphics
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Provides: svgwrite
+Requires-Python: >=3.6
+License-File: LICENSE.TXT
+
+svgwrite
+========
+
+This package is inactive! No new features will be added, there will
+be no change of behavior, just bugfixes will be merged.
+
+Abstract
+--------
+
+A Python library to create SVG drawings.
+
+a simple example::
+
+    import svgwrite
+
+    dwg = svgwrite.Drawing('test.svg', profile='tiny')
+    dwg.add(dwg.line((0, 0), (10, 0), stroke=svgwrite.rgb(10, 10, 16, '%')))
+    dwg.add(dwg.text('Test', insert=(0, 0.2), fill='red'))
+    dwg.save()
+
+for more examples see: examples.py
+
+As the name `svgwrite` implies, `svgwrite` creates new SVG drawings, it does not read existing drawings and also does
+not import existing drawings, but you can always include other SVG drawings by the <image> entity.
+
+`svgwrite` is a pure Python package and has no external dependencies.
+
+Installation
+------------
+
+with pip::
+
+    pip install svgwrite
+
+or from source::
+
+    python setup.py install
+
+
+Documentation
+-------------
+
+http://readthedocs.org/docs/svgwrite/
+
+svgwrite can be found on GitHub.com at:
+
+http://github.com/mozman/svgwrite.git
+
+Contact
+-------
+
+svgwrite@mozman.at
+
+NEWS
+====
+
+Version 1.4.3 - 2022-07-14
+--------------------------
+
+* BUGFIX: `#114 <https://github.com/mozman/svgwrite/pull/114>`_ Polyline validation by @normanrz
+
+Version 1.4.2 - 2022-03-23
+--------------------------
+
+* This package is INACTIVE!
+
+  * No new features will be added.
+  * There will be no change of behavior.
+  * Only bugfixes will be merged.
+
+* Merged some contributions and bugfixes.
+
+Version 1.4.1 - 2021-01-15
+--------------------------
+
+* This package is in maintenance mode, no new features will be added, there will
+  be no change of behavior, just bugfixes will be merged.
+* Merged some contributions and bugfixes.
+
+Version 1.4 - 2020-03-28
+------------------------
+
+* Requires Python 3.6
+* Removed Python2 support
+* Optimized/faster ``svgparser.py`` module by Florian Festi
+* Removed batch files to run tests on Windows, use ``pytest`` or ``tox`` instead
+
+Version 1.3.1 - 2019-06-28
+--------------------------
+
+* BUGFIX: changed license of module shapes.py to MIT license
+
+Version 1.3.0 - 2019-06-23
+--------------------------
+
+* Dropping Python 2 support in v1.4
+* NEW: easy stylesheet embedding: `Drawing.embed_stylesheet(content)`, see example `using_fonts.py <https://github.com/mozman/svgwrite/blob/9d509fe1842e519b8d8475f83543a7589b7c1879/examples/using_fonts.py#L21>`_
+* NEW: embedded local fonts: `Drawing.embed_font(fontname, filename)`, see example using_fonts.py
+* NEW: embedded web fonts: `Drawing.embed_google_web_font(fontname, uri)`, see example using_fonts.py
+* NEW: shapes extension by `Christof Hanke <https://www.induhviduals.de/>`_
+* NEW: for write(), save() and saveas(), indentation for pretty print is configurable by `Mitar <http://mitar.tnode.com/>`_
+* BUGFIX: converted regex strings in tests to raw strings by `tirkarthi <http://tirkarthi.github.io/>`_
+
+Version 1.2.1 - 2018-11-16
+--------------------------
+
+* BUGFIX: Python 2 pretty print encoding issue
+
+Version 1.2.0 - 2018-10-27
+--------------------------
+
+* NEW: Inkscape extension - support for Inkscape layers
+* BUGFIX: compatibility with Python 3.7 by fixing re.split() usage
+
+Version 1.1.12 - 2017-12-27
+---------------------------
+
+* BUGFIX: values attribute for element animate, accepts any semicolon separated values
+
+Version 1.1.11 - 2017-05-31
+---------------------------
+
+* CHANGED: e.set_markers((sm, mm, em)), accepts None/False values to set markers individually, like just set end
+  marker: s.set_markers((None, False, em))
+* BUGFIX: text-decoration attribute accepts multiple values like 'underline overline'
+
+Version 1.1.10 - 2017-01-28
+---------------------------
+
+* NEW: added `pretty` argument to `Drawing.write`, `Drawing.save` and `Drawing.saveas` to produce easy to read XML
+  output, svgwrite reparses the output string with minidom, this reduces the performance because this process requires
+  memory and runtime, default value for `pretty` is False.
+* Moved repository to GitHub.com: http://github.com/mozman/svgwrite.git
+
+Version 1.1.9 - 2016-10-26
+--------------------------
+
+* NEW: added solidColor paint server (only valid in the Tiny 1.2 profile)
+
+Version 1.1.8 - 2016-05-31
+--------------------------
+
+* BUGFIX: None checks: 'if value:' -> 'if value is not None:'
+
+Version 1.1.7 - 2016-05-22
+--------------------------
+
+* BUGFIX: color accepts percentage values as floats like "rgb(10.2%, 3.78%, 20%)"
+
+Version 1.1.6 - 2014-05-30
+--------------------------
+
+* BUGFIX: sign for offset-value wasn't optional
+
+Version 1.1.5 - 2014-03-26
+--------------------------
+
+* BUGFIX: xml serialization for CPython 3.4.0
+
+Version 1.1.4 - 2014-03-16
+--------------------------
+
+* simplified path parser
+* pyparsing as external dependency (by jenselme)
+
+Version 1.1.3 - 2013-10-01
+--------------------------
+
+* updated pyparsing for Python 3 to version 2.0.1 (prior version caused memory leaks)
+* BUGFIX: utf8 to unicode encoding error for Python 2.7
+* Tests for Python 3 require CPython3.3 or newer, using the 'u' prefix.
+
+Version 1.1.2 - 2013-01-08
+--------------------------
+
+* prevent setup.py from compiling all modules - error with 'pyparsing_py2.py' and Python3
+* BUGFIX: all tests run with CPython3.3
+
+Version 1.1.1 - 2012-08-15
+--------------------------
+
+* License changed to MIT License
+* tested with CPython2.7, CPython3.2, CPython3.3 and pypy-1.9 on Win7 Pro 32-bit
+* BUGFIX: dwg.animateTranform() -> dwg.animateTransform()
+* BUGFIX: in examples, replaced width and height params by size parameter
+* added examples
+* edit docs
+
+Version 1.0.1 - 2012-06-08
+--------------------------
+
+* added inline stylesheets
+* added examples created by Lawrence Tattrie
+
+Version 1.0.0 - 2012-05-27
+--------------------------
+
+* stable
+* tested with CPython 2.7, CPython 3.2, pypy-1.8
+* added script tag - thx to jmahmood
+* docs also available at: http://readthedocs.org/docs/svgwrite
+
+Version 0.2.4 - 2011-12-30
+--------------------------
+
+* beta version
+* Python 2.7: all strings will be converted by the unicode() function, for strings containing none-ascii-characters use
+  prefix ``u""`` or better use ``from __future__ import unicode_literals``, because this is Python 3 compatible.
+* tested with CPython 2.7, CPython 3.2, and PyPy 1.7
+* BUGFIX: color parsing accepts white spaces in ``rgb()`` like ``rgb(0, 0, 0)``
+
+Version 0.2.3 - 2010-11-13
+--------------------------
+
+* beta version
+* Python 3.1 support
+* splitted examples.py into several files and moved them to the subdir 'examples'
+
+Version 0.2.2 - 2010-11-05
+--------------------------
+
+* alpha version
+* removed 'attribs' parameter from all constructors
+* new elements: Set, Animate, AnimateMotion, AnimateColor, AnimateTransform, all filter elements
+* added set_desc(title, desc), set_metadata(xmldata) to BaseElement class
+* moved content of interfaces.py to mixins.py, (ITransform -> Transform and so on)
+
+Version 0.2.1 - 2010-10-31
+--------------------------
+
+* alpha version
+* new elements: Marker, ClipPath, Mask
+* paint service: LinearGradient, RadialGradient, Pattern
+
+Version 0.2.0 - 2010-10-24
+--------------------------
+
+* alpha version
+* validator rewritten as validator2.py
+* debug and profile options separated for each drawing object
+* important change: create objects with factory functions of the *Drawing* class: drawing.<svg-elementname>(...)
+* added mixins for setting stroke and fill properties
+* new elements: Hyperlink, Image, TextArea,
+
+Version 0.1.0 - 2010-09-26
+--------------------------
+
+* alpha version
+* new elements:
+
+    * basic shapes: Line, Rect, Circle, Ellipse, Polyline, Polygon, Path
+    * text elements: Text, TSpan, TRef, TextPath
+    * container elements: Group, Symbol, SVG, Use, Defs
+
+* for examples see: examples.py
+
+
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/RECORD
new file mode 100644
index 00000000..91eb575c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/RECORD
@@ -0,0 +1,70 @@
+svgwrite-1.4.3.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+svgwrite-1.4.3.dist-info/LICENSE.TXT,sha256=GBOnVT3pJ-iGU0WxBRw210G3DQ6fyrJXLfZzoY-TZik,2282
+svgwrite-1.4.3.dist-info/METADATA,sha256=03ifbSvQep7d3u1O3kmEV3FB1IY2n8kJfaPI60LT83U,8843
+svgwrite-1.4.3.dist-info/RECORD,,
+svgwrite-1.4.3.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
+svgwrite-1.4.3.dist-info/top_level.txt,sha256=zz1R0XrwfTXSn2upAtibS7jzihSwVNRcynsFEvrqhJI,43
+svgwrite/__init__.py,sha256=FVp7zscdyoByyETIUnz6YJ8MYFG0VqxKDa4BHlbu_QI,2506
+svgwrite/__pycache__/__init__.cpython-312.pyc,,
+svgwrite/__pycache__/animate.cpython-312.pyc,,
+svgwrite/__pycache__/base.cpython-312.pyc,,
+svgwrite/__pycache__/container.cpython-312.pyc,,
+svgwrite/__pycache__/drawing.cpython-312.pyc,,
+svgwrite/__pycache__/elementfactory.cpython-312.pyc,,
+svgwrite/__pycache__/etree.cpython-312.pyc,,
+svgwrite/__pycache__/filters.cpython-312.pyc,,
+svgwrite/__pycache__/gradients.cpython-312.pyc,,
+svgwrite/__pycache__/image.cpython-312.pyc,,
+svgwrite/__pycache__/masking.cpython-312.pyc,,
+svgwrite/__pycache__/mixins.cpython-312.pyc,,
+svgwrite/__pycache__/params.cpython-312.pyc,,
+svgwrite/__pycache__/path.cpython-312.pyc,,
+svgwrite/__pycache__/pattern.cpython-312.pyc,,
+svgwrite/__pycache__/shapes.cpython-312.pyc,,
+svgwrite/__pycache__/solidcolor.cpython-312.pyc,,
+svgwrite/__pycache__/text.cpython-312.pyc,,
+svgwrite/__pycache__/utils.cpython-312.pyc,,
+svgwrite/__pycache__/validator2.cpython-312.pyc,,
+svgwrite/__pycache__/version.cpython-312.pyc,,
+svgwrite/animate.py,sha256=S9Gezc9yOawAZ4PZ1LgcRH4DZftWDUb0K6S63PE6sqQ,6663
+svgwrite/base.py,sha256=w2DYG_cbigZiOYjNMlKI3ZkoVBY7DOX0HU7flqDfi0U,8624
+svgwrite/container.py,sha256=qVnz-lWRgaQERdKsOatsBUJB6Fo_dT_BO0uj3qyp7cM,10741
+svgwrite/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+svgwrite/data/__pycache__/__init__.cpython-312.pyc,,
+svgwrite/data/__pycache__/colors.cpython-312.pyc,,
+svgwrite/data/__pycache__/full11.cpython-312.pyc,,
+svgwrite/data/__pycache__/pattern.cpython-312.pyc,,
+svgwrite/data/__pycache__/svgparser.cpython-312.pyc,,
+svgwrite/data/__pycache__/tiny12.cpython-312.pyc,,
+svgwrite/data/__pycache__/typechecker.cpython-312.pyc,,
+svgwrite/data/__pycache__/types.cpython-312.pyc,,
+svgwrite/data/colors.py,sha256=Y8WAYb9_lAaBJjtk5QFm_RIoVShgASw579h10gJurFo,2845
+svgwrite/data/full11.py,sha256=RUu9Biz9FyjN6WVfe-kyEQSJyCWcX5PL5_FmyRgG32A,76415
+svgwrite/data/pattern.py,sha256=N5mZ7X1YKW_T1NjRtsYZgm3xIyOgTYJDFJ_2xfuj0K0,1064
+svgwrite/data/svgparser.py,sha256=fL55kB3fIJ4eY7iis2otAUvKM6xuCUCUpwTk__3hj7Y,5036
+svgwrite/data/tiny12.py,sha256=HIbUKIGF-SdabNS8GKFPsy0ElJn4o8smEs1qRx7bV14,58868
+svgwrite/data/typechecker.py,sha256=vWcHGp-qqwpqneYRlAQAZJq1pBNfmHQcu7CHQVw0yOw,11944
+svgwrite/data/types.py,sha256=78ETW6fsNqkTDp0V5yHgNGz3xrHbz_oKZb9iBGIL7FY,2807
+svgwrite/drawing.py,sha256=PorqOz93xr1GBi_2zQKqPinNGEr3c-SsVqx-KSIR5zo,5291
+svgwrite/elementfactory.py,sha256=onqBpxGNW6bi-uyjK_9FbGQplpdlprEcxDr2WZtUw_Y,2297
+svgwrite/etree.py,sha256=CbWxoUynhLc1mvDUh528JDVD16pXZ-YMT8Vohx6u9ts,1361
+svgwrite/extensions/__init__.py,sha256=Wh7mopLfv5YiM_W4XQPCa8WfBjLcipCQdojYKZ2UvgA,93
+svgwrite/extensions/__pycache__/__init__.cpython-312.pyc,,
+svgwrite/extensions/__pycache__/inkscape.cpython-312.pyc,,
+svgwrite/extensions/__pycache__/shapes.cpython-312.pyc,,
+svgwrite/extensions/inkscape.py,sha256=OItSjLMhbGGxJNjZQqvCdoa90t8ffvvUD7bVn5AspCw,3073
+svgwrite/extensions/shapes.py,sha256=_IXOvaSj05EdAsS4iRT1JGjdPOvL7oIMFdbrgJfWbjI,3728
+svgwrite/filters.py,sha256=0HBiWiOBwIOB_VXShvsKZCs9IS4y18c-INoNlQYBUmE,8002
+svgwrite/gradients.py,sha256=OpdjLDrltwsNM7y2lEdVSd1HRnvrXOO7sKUV-JYPrRw,4702
+svgwrite/image.py,sha256=gjoenfpS2a1NqxzE3QWQybl6V7Nu3g-m6XlWhUqMPWo,2493
+svgwrite/masking.py,sha256=uV6d9L6DDZ1XFlRsN6k9TS8N7ItXckHltqp0_LDY4ys,1856
+svgwrite/mixins.py,sha256=PqVuKFpT9yfx0EgVe0DD9IdA8qKqCPIzwZgYMy8SBJ8,10749
+svgwrite/params.py,sha256=Z_oIjKHWIpsox5JOfP7J0_xtLsZqKJX0XWwHhmyE0AQ,1903
+svgwrite/path.py,sha256=gr8YNCcLbczxAMSf7V6_3yrQAXa8c96tU8twWpJCpKs,2848
+svgwrite/pattern.py,sha256=FF_KSTPQflfLhFbjNMcLLmNvfMLk-jRcLGkG3crvus0,1973
+svgwrite/shapes.py,sha256=JiIvhn2v7s0bgJbdPB1pFUQ2z6IFiObwiAwk1Azsrw0,5310
+svgwrite/solidcolor.py,sha256=wjE_109Gqk1x3I6djo8bSrmG2SmKQ4sE63kXldkWYN4,1734
+svgwrite/text.py,sha256=UXWUbBCKUUlqX06ClKxlHYDb2X5KIa8vpUHtp74q-mY,8138
+svgwrite/utils.py,sha256=-nWhA1VpQPdrm8US9g-IQRwEdnDBlH_BAzrLY1vpcpk,7421
+svgwrite/validator2.py,sha256=vY49WBrteUtKRbgE8RVJhawF8O49Jo0BfTgmU-Vi5FU,6205
+svgwrite/version.py,sha256=Bxd0Ytp8hQjNE2o2xa6T9zfL0eBdnQDAUCprGz1jlvs,1119
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/WHEEL
new file mode 100644
index 00000000..becc9a66
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/WHEEL
@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: bdist_wheel (0.37.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/top_level.txt b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/top_level.txt
new file mode 100644
index 00000000..7da3e2d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite-1.4.3.dist-info/top_level.txt
@@ -0,0 +1,3 @@
+svgwrite
+svgwrite/data
+svgwrite/extensions
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__init__.py b/backend/venv/lib/python3.12/site-packages/svgwrite/__init__.py
new file mode 100644
index 00000000..240d4d04
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/__init__.py
@@ -0,0 +1,93 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman
+# Purpose: package definition file
+# Created: 08.09.2010
+# License: MIT License
+# Copyright (c) 2010-2018  Manfred Moitzi
+
+"""
+A Python library to create SVG drawings.
+
+SVG is a language for describing two-dimensional graphics in XML. SVG allows
+for three types of graphic objects: vector graphic shapes (e.g., paths
+consisting of straight lines and curves), images and text. Graphical objects
+can be grouped, styled, transformed and composed into previously rendered
+objects. The feature set includes nested transformations, clipping paths,
+alpha masks, filter effects and template objects.
+
+SVG drawings can be interactive and dynamic. Animations can be defined and
+triggered either declarative (i.e., by embedding SVG animation elements in
+SVG content) or via scripting.
+
+.. seealso:: http://www.w3.org/TR/SVG11/intro.html#AboutSVG
+
+a simple example::
+
+    import svgwrite
+
+    dwg = svgwrite.Drawing('test.svg', profile='tiny')
+    dwg.add(dwg.line((0, 0), (10, 0), stroke=svgwrite.rgb(10, 10, 16, '%')))
+    dwg.add(dwg.text('Test', insert=(0, 0.2)))
+    dwg.save()
+
+SVG Version
+-----------
+
+You can only create two types of SVG drawings:
+
+* *SVG 1.2 Tiny Profile*, use Drawing(profile= ``'tiny'``)
+* *SVG 1.1 Full Profile*, use Drawing(profile= ``'full'``)
+
+"""
+from .version import __version__, version
+VERSION = __version__
+
+__author__ = "mozman <me@mozman.at>"
+
+AUTHOR_NAME = 'Manfred Moitzi'
+AUTHOR_EMAIL = 'me@mozman.at'
+CYEAR = '2014-2019'
+
+
+from svgwrite.drawing import Drawing
+from svgwrite.utils import rgb
+
+
+class Unit(object):
+    """ Add units to values.
+    """
+    def __init__(self, unit='cm'):
+        """ Unit constructor
+
+        :param str unit: specify the unit string
+        """
+        self._unit = unit
+
+    def __rmul__(self, other):
+        """ add unit-string to 'other'. (e.g. 5*cm => '5cm') """
+        return "%s%s" % (other, self._unit)
+
+    def __call__(self, *args):
+        """ Add unit-strings to all arguments.
+
+        :param args: list of values
+            e.g.: cm(1,2,3) => '1cm,2cm,3cm'
+        """
+        return ','.join(["%s%s" % (arg, self._unit) for arg in args])
+
+
+cm = Unit('cm')
+mm = Unit('mm')
+em = Unit('em')
+ex = Unit('ex')
+px = Unit('px')
+inch = Unit('in')
+pc = Unit('pc')
+pt = Unit('pt')
+percent = Unit('%')
+deg = Unit('deg')
+grad = Unit('grad')
+rad = Unit('rad')
+Hz = Unit('Hz')
+kHz = Unit('kHz')
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0d1a4a39
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/animate.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/animate.cpython-312.pyc
new file mode 100644
index 00000000..69169b02
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/animate.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/base.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/base.cpython-312.pyc
new file mode 100644
index 00000000..510cbf69
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/base.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/container.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/container.cpython-312.pyc
new file mode 100644
index 00000000..ef6d11a8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/container.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/drawing.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/drawing.cpython-312.pyc
new file mode 100644
index 00000000..a127d4bd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/drawing.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/elementfactory.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/elementfactory.cpython-312.pyc
new file mode 100644
index 00000000..7444a685
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/elementfactory.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/etree.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/etree.cpython-312.pyc
new file mode 100644
index 00000000..9323021d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/etree.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/filters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/filters.cpython-312.pyc
new file mode 100644
index 00000000..1a6689b6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/filters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/gradients.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/gradients.cpython-312.pyc
new file mode 100644
index 00000000..0634aa2f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/gradients.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/image.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/image.cpython-312.pyc
new file mode 100644
index 00000000..69907e8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/image.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/masking.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/masking.cpython-312.pyc
new file mode 100644
index 00000000..35a29ea1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/masking.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/mixins.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/mixins.cpython-312.pyc
new file mode 100644
index 00000000..321cdc71
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/mixins.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/params.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/params.cpython-312.pyc
new file mode 100644
index 00000000..31bbb498
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/params.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/path.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/path.cpython-312.pyc
new file mode 100644
index 00000000..67a40cde
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/path.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/pattern.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/pattern.cpython-312.pyc
new file mode 100644
index 00000000..2cfa1236
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/pattern.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/shapes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/shapes.cpython-312.pyc
new file mode 100644
index 00000000..f6be15e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/shapes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/solidcolor.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/solidcolor.cpython-312.pyc
new file mode 100644
index 00000000..df5f275a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/solidcolor.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/text.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/text.cpython-312.pyc
new file mode 100644
index 00000000..f3a722fa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/text.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..7f757347
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/validator2.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/validator2.cpython-312.pyc
new file mode 100644
index 00000000..ebc96b1d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/validator2.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/version.cpython-312.pyc
new file mode 100644
index 00000000..e5cfa963
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/__pycache__/version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/animate.py b/backend/venv/lib/python3.12/site-packages/svgwrite/animate.py
new file mode 100644
index 00000000..5908ec1c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/animate.py
@@ -0,0 +1,183 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: animate elements
+# Created: 31.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import XLink
+from svgwrite.utils import strlist, is_string
+
+
+class Set(BaseElement, XLink):
+    """ The **set** element provides a simple means of just setting the value
+    of an attribute for a specified duration. It supports all attribute types,
+    including those that cannot reasonably be interpolated, such as string
+    and boolean values. The **set** element is non-additive. The additive and
+    accumulate attributes are not allowed, and will be ignored if specified.
+    """
+    elementname = 'set'
+
+    def __init__(self, href=None, **extra):
+        """ Set constructor.
+
+        :param href: target svg element, if **href** is not `None`; else
+            the target SVG Element is the parent SVG Element.
+        """
+        super(Set, self).__init__(**extra)
+        if href is not None:
+            self.set_href(href)
+
+    def get_xml(self):
+        self.update_id() # if href is an object - 'id' - attribute may be changed!
+        return super(Set, self).get_xml()
+
+    def set_target(self, attributeName, attributeType=None):
+        """
+        Set animation attributes :ref:`attributeName` and :ref:`attributeType`.
+        """
+        self['attributeName'] = attributeName
+        if attributeType is not None:
+            self['attributeType'] = attributeType
+
+
+    def set_event(self, onbegin=None, onend=None, onrepeat=None, onload=None):
+        """
+        Set animation attributes :ref:`onbegin`, :ref:`onend`, :ref:`onrepeat`
+        and :ref:`onload`.
+        """
+        if onbegin is not None:
+            self['onbegin'] = onbegin
+        if onend is not None:
+            self['onend'] = onend
+        if onrepeat is not None:
+            self['onrepeat'] = onrepeat
+        if onload is not None:
+            self['onload'] = onload
+
+    def set_timing(self, begin=None, end=None, dur=None, min=None, max=None,
+                   restart=None, repeatCount=None, repeatDur=None):
+        """
+        Set animation attributes :ref:`begin`, :ref:`end`, :ref:`dur`,
+        :ref:`min`, :ref:`max`, :ref:`restart`, :ref:`repeatCount` and
+        :ref:`repeatDur`.
+        """
+        if begin is not None:
+            self['begin'] = begin
+        if end is not None:
+            self['end'] = end
+        if dur is not None:
+            self['dur'] = dur
+        if min is not None:
+            self['min'] = min
+        if max is not None:
+            self['max'] = max
+        if restart is not None:
+            self['restart'] = restart
+        if repeatCount is not None:
+            self['repeatCount'] = repeatCount
+        if repeatDur is not None:
+            self['repeatDur'] = repeatDur
+
+    def freeze(self):
+        """ Freeze the animation effect. (see also :ref:`fill <animateFill>`)
+        """
+        self['fill'] = 'freeze'
+
+class AnimateMotion(Set):
+    """ The **animateMotion** element causes a referenced element to move
+    along a motion path.
+    """
+    elementname = 'animateMotion'
+
+    def __init__(self, path=None, href=None, **extra):
+        """
+        :param path: the motion path
+        :param href: target svg element, if **href** is not `None`; else
+          the target SVG Element is the parent SVG Element.
+        """
+        super(AnimateMotion, self).__init__(href=href, **extra)
+        if path is not None:
+            self['path'] = path
+
+    def set_value(self, path=None, calcMode=None, keyPoints=None, rotate=None):
+        """
+        Set animation attributes `path`, `calcMode`, `keyPoints` and `rotate`.
+        """
+        if path is not None:
+            self['path'] = path
+        if calcMode is not None:
+            self['calcMode'] = calcMode
+        if keyPoints is not None:
+            self['keyPoints'] = keyPoints
+        if rotate is not None:
+            self['rotate'] = rotate
+
+
+class Animate(Set):
+    """ The **animate** element allows scalar attributes and properties to be
+    assigned different values over time .
+    """
+    elementname = 'animate'
+
+    def __init__(self, attributeName=None, values=None, href=None, **extra):
+        """
+        :param attributeName: name of the SVG Attribute to animate
+        :param values: interpolation values, `string` as `<semicolon-list>` or a python `list`
+        :param href: target svg element, if **href** is not `None`; else
+          the target SVG Element is the parent SVG Element.
+        """
+        super(Animate, self).__init__(href=href, **extra)
+        if values is not None:
+            self.set_value(values)
+        if attributeName is not None:
+            self.set_target(attributeName)
+
+    def set_value(self, values, calcMode=None, keyTimes=None, keySplines=None,
+                  from_=None, to=None, by=None):
+        """
+        Set animation attributes :ref:`values`, :ref:`calcMode`, :ref:`keyTimes`,
+        :ref:`keySplines`, :ref:`from`, :ref:`to` and :ref:`by`.
+        """
+        if values is not None:
+            if not is_string(values):
+                values = strlist(values, ';')
+            self['values'] = values
+
+        if calcMode is not None:
+            self['calcMode'] = calcMode
+        if keyTimes is not None:
+            self['keyTimes'] = keyTimes
+        if keySplines is not None:
+            self['keySplines'] = keySplines
+        if from_ is not None:
+            self['from'] = from_
+        if to is not None:
+            self['to'] = to
+        if by is not None:
+            self['by'] = by
+
+
+class AnimateColor(Animate):
+    """ The **animateColor** element specifies a color transformation over
+    time.
+    """
+    elementname = 'animateColor'
+
+
+class AnimateTransform(Animate):
+    """ The **animateTransform** element animates a transformation attribute
+    on a target element, thereby allowing animations to control translation,
+    scaling, rotation and/or skewing.
+    """
+    elementname = 'animateTransform'
+    def __init__(self, transform, element=None, **extra):
+        """
+        :param element: target svg element, if element is not `None`; else
+          the target svg element is the parent svg element.
+        :param string transform: ``'translate | scale | rotate | skewX | skewY'``
+        """
+        super(AnimateTransform, self).__init__(element, **extra)
+        self['type'] = transform
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/base.py b/backend/venv/lib/python3.12/site-packages/svgwrite/base.py
new file mode 100644
index 00000000..f38cb38b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/base.py
@@ -0,0 +1,285 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman
+# Purpose: svg base element
+# Created: 08.09.2010
+# Copyright (c) 2010-2020, Manfred Moitzi
+# License: MIT License
+"""
+The **BaseElement** is the root for all SVG elements.
+"""
+
+from svgwrite.etree import etree
+
+import copy
+
+from svgwrite.params import Parameter
+from svgwrite.utils import AutoID
+
+
+class BaseElement(object):
+    """
+    The **BaseElement** is the root for all SVG elements. The SVG attributes
+    are stored in **attribs**, and the SVG subelements are stored in
+    **elements**.
+
+    """
+    elementname = 'baseElement'
+
+    def __init__(self, **extra):
+        """
+        :param extra: extra SVG attributes (keyword arguments)
+
+          * add trailing '_' to reserved keywords: ``'class_'``, ``'from_'``
+          * replace inner '-' by '_': ``'stroke_width'``
+
+
+        SVG attribute names will be checked, if **debug** is `True`.
+
+        workaround for removed **attribs** parameter in Version 0.2.2::
+
+            # replace
+            element = BaseElement(attribs=adict)
+
+            #by
+            element = BaseElement()
+            element.update(adict)
+
+        """
+        # the keyword 'factory' specifies the object creator
+        factory = extra.pop('factory', None)
+        if factory is not None:
+            # take parameter from 'factory'
+            self._parameter = factory._parameter
+        else:
+            # default parameter debug=True profile='full'
+            self._parameter = Parameter()
+
+        # override debug setting
+        debug = extra.pop('debug', None)
+        if debug is not None:
+            self._parameter.debug = debug
+
+        # override profile setting
+        profile = extra.pop('profile', None)
+        if profile is not None:
+            self._parameter.profile = profile
+
+        self.attribs = dict()
+        self.update(extra)
+        self.elements = list()
+
+    def update(self, attribs):
+        """ Update SVG Attributes from `dict` attribs.
+
+        Rules for keys:
+
+        1. trailing '_' will be removed (``'class_'`` -> ``'class'``)
+        2. inner '_' will be replaced by '-' (``'stroke_width'`` -> ``'stroke-width'``)
+
+        """
+        for key, value in attribs.items():
+            # remove trailing underscores
+            # and replace inner underscores
+            key = key.rstrip('_').replace('_', '-')
+            self.__setitem__(key, value)
+
+    def copy(self):
+        newobj = copy.copy(self)  # shallow copy of object
+        newobj.attribs = copy.copy(self.attribs)  # shallow copy of attributes
+        newobj.elements = copy.copy(self.elements)  # shallow copy of subelements
+        if 'id' in newobj.attribs:  # create a new 'id'
+            newobj['id'] = newobj.next_id()
+        return newobj
+
+    @property
+    def debug(self):
+        return self._parameter.debug
+
+    @property
+    def profile(self):
+        return self._parameter.profile
+
+    @property
+    def validator(self):
+        return self._parameter.validator
+
+    @validator.setter
+    def validator(self, value):
+        self._parameter.validator = value
+
+    @property
+    def version(self):
+        return self._parameter.get_version()
+
+    def set_parameter(self, parameter):
+        self._parameter = parameter
+
+    def next_id(self, value=None):
+        return AutoID.next_id(value)
+
+    def get_id(self):
+        """ Get the object `id` string, if the object does not have an `id`,
+        a new `id` will be created.
+
+        :returns: `string`
+        """
+        if 'id' not in self.attribs:
+            self.attribs['id'] = self.next_id()
+        return self.attribs['id']
+
+    def get_iri(self):
+        """
+        Get the `IRI` reference string of the object. (i.e., ``'#id'``).
+
+        :returns: `string`
+        """
+        return "#%s" % self.get_id()
+
+    def get_funciri(self):
+        """
+        Get the `FuncIRI` reference string of the object. (i.e. ``'url(#id)'``).
+
+        :returns: `string`
+        """
+        return "url(%s)" % self.get_iri()
+
+    def __getitem__(self, key):
+        """ Get SVG attribute by `key`.
+
+        :param string key: SVG attribute name
+        :return: SVG attribute value
+
+        """
+        return self.attribs[key]
+
+    def __setitem__(self, key, value):
+        """ Set SVG attribute by `key` to `value`.
+
+        :param string key: SVG attribute name
+        :param object value: SVG attribute value
+
+        """
+        # Attribute checking is only done by using the __setitem__() method or
+        # by self['attribute'] = value
+        if self.debug:
+            self.validator.check_svg_attribute_value(self.elementname, key, value)
+        self.attribs[key] = value
+
+    def add(self, element):
+        """ Add an SVG element as subelement.
+
+        :param element: append this SVG element
+        :returns: the added element
+
+        """
+        if self.debug:
+            self.validator.check_valid_children(self.elementname, element.elementname)
+        self.elements.append(element)
+        return element
+
+    def tostring(self):
+        """ Get the XML representation as unicode `string`.
+
+        :return: unicode XML string of this object and all its subelements
+
+        """
+        xml = self.get_xml()
+        # required for Python 2 support
+        xml_utf8_str = etree.tostring(xml, encoding='utf-8')
+        return xml_utf8_str.decode('utf-8')
+        # just Python 3: return etree.tostring(xml, encoding='unicode')
+    
+    def _repr_svg_(self):
+        """ Show SVG in IPython, Jupyter Notebook, and Jupyter Lab
+
+        :return: unicode XML string of this object and all its subelements
+
+        """
+        return self.tostring()
+
+    def get_xml(self):
+        """ Get the XML representation as `ElementTree` object.
+
+        :return: XML `ElementTree` of this object and all its subelements
+
+        """
+        xml = etree.Element(self.elementname)
+        if self.debug:
+            self.validator.check_all_svg_attribute_values(self.elementname, self.attribs)
+        for attribute, value in sorted(self.attribs.items()):
+            # filter 'None' values
+            if value is not None:
+                value = self.value_to_string(value)
+                if value:  # just add not empty attributes
+                    xml.set(attribute, value)
+
+        for element in self.elements:
+            xml.append(element.get_xml())
+        return xml
+
+    def value_to_string(self, value):
+        """
+        Converts *value* into a <string> includes a value check, depending
+        on :attr:`self.debug` and :attr:`self.profile`.
+
+        """
+        if isinstance(value, (int, float)):
+            if self.debug:
+                self.validator.check_svg_type(value, 'number')
+            if isinstance(value, float) and self.profile == 'tiny':
+                value = round(value, 4)
+        return str(value)
+
+    def set_desc(self, title=None, desc=None):
+        """ Insert a **title** and/or a **desc** element as first subelement.
+        """
+        if desc is not None:
+            self.elements.insert(0, Desc(desc))
+        if title is not None:
+            self.elements.insert(0, Title(title))
+
+    def set_metadata(self, xmldata):
+        """
+        :param xmldata: an xml.etree.ElementTree - Element() object.
+        """
+        metadata = Metadata(xmldata)
+        if len(self.elements) == 0:
+            self.elements.append(metadata)
+        else:
+            pos = 0
+            while self.elements[pos].elementname in ('title', 'desc'):
+                pos += 1
+                if pos == len(self.elements):
+                    self.elements.append(metadata)
+                    return
+            if self.elements[pos].elementname == 'metadata':
+                self.elements[pos].xml.append(xmldata)
+            else:
+                self.elements.insert(pos, metadata)
+
+
+class Title(object):
+    elementname = 'title'
+
+    def __init__(self, text):
+        self.xml = etree.Element(self.elementname)
+        self.xml.text = str(text)
+
+    def get_xml(self):
+        return self.xml
+
+
+class Desc(Title):
+    elementname = 'desc'
+
+
+class Metadata(Title):
+    elementname = 'metadata'
+
+    def __init__(self, xmldata):
+        """
+        :param xmldata: an xml.etree.ElementTree - Element() object.
+        """
+        self.xml = etree.Element('metadata')
+        self.xml.append(xmldata)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/container.py b/backend/venv/lib/python3.12/site-packages/svgwrite/container.py
new file mode 100644
index 00000000..0e215dc0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/container.py
@@ -0,0 +1,287 @@
+#coding:utf-8
+# Author:  mozman
+# Purpose: svg container classes
+# Created: 15.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+"""
+The **container** module provides following structural objects:
+
+* :class:`svgwrite.Group`
+* :class:`svgwrite.SVG`
+* :class:`svgwrite.Defs`
+* :class:`svgwrite.Symbol`
+* :class:`svgwrite.Marker`
+* :class:`svgwrite.Use`
+* :class:`svgwrite.Hyperlink`
+* :class:`svgwrite.Script`
+* :class:`svgwrite.Style`
+
+set/get SVG attributes::
+
+    element['attribute'] = value
+    value = element['attribute']
+
+"""
+from urllib.request import urlopen
+from svgwrite.utils import font_mimetype, base64_data, find_first_url
+from svgwrite.base import BaseElement
+from svgwrite.mixins import ViewBox, Transform, XLink
+from svgwrite.mixins import Presentation, Clipping
+from svgwrite.etree import CDATA
+
+
+class Group(BaseElement, Transform, Presentation):
+    """ The **Group** (SVG **g**) element is a container element for grouping
+    together related graphics elements.
+
+    Grouping constructs, when used in conjunction with the **desc** and **title**
+    elements, provide information about document structure and semantics.
+    Documents that are rich in structure may be rendered graphically, as speech,
+    or as braille, and thus promote accessibility.
+
+    A group of elements, as well as individual objects, can be given a name using
+    the **id** attribute. Named groups are needed for several purposes such as
+    animation and re-usable objects.
+
+    """
+    elementname = 'g'
+
+
+class Defs(Group):
+    """ The **defs** element is a container element for referenced elements. For
+    understandability and accessibility reasons, it is recommended that, whenever
+    possible, referenced elements be defined inside of a **defs**.
+    """
+    elementname = 'defs'
+
+
+class Symbol(BaseElement, ViewBox, Presentation, Clipping):
+    """ The **symbol** element is used to define graphical template objects which
+    can be instantiated by a **use** element. The use of **symbol** elements for
+    graphics that are used multiple times in the same document adds structure and
+    semantics. Documents that are rich in structure may be rendered graphically,
+    as speech, or as braille, and thus promote accessibility.
+    """
+    # ITransform interface is not valid for Symbol -> do not inherit from Group
+    elementname = 'symbol'
+
+
+class Marker(BaseElement, ViewBox, Presentation):
+    """ The **marker** element defines the graphics that is to be used for
+    drawing arrowheads or polymarkers on a given **path**, **line**, **polyline**
+    or **polygon** element.
+
+    Add Marker definitions to a **defs** section, preferred to the **defs** section
+    of the **main drawing**.
+
+    """
+    elementname = 'marker'
+
+    def __init__(self, insert=None, size=None, orient=None, **extra):
+        """
+        :param 2-tuple insert: reference point (**refX**, **refY**)
+        :param 2-tuple size: (**markerWidth**, **markerHeight**)
+        :param orient: ``'auto'`` | `angle`
+        :param extra: additional SVG attributes as keyword-arguments
+        """
+        super(Marker, self).__init__(**extra)
+        if insert is not None:
+            self['refX'] = insert[0]
+            self['refY'] = insert[1]
+        if size is not None:
+            self['markerWidth'] = size[0]
+            self['markerHeight'] = size[1]
+        if orient is not None:
+            self['orient'] = orient
+        if 'id' not in self.attribs:  # an 'id' is necessary
+            self['id'] = self.next_id()
+
+
+FONT_TEMPLATE = """@font-face{{ 
+    font-family: "{name}"; 
+    src: url("{data}"); 
+}}
+"""
+
+
+class SVG(Symbol):
+    """ A SVG document fragment consists of any number of SVG elements contained
+    within an **svg** element.
+
+    An SVG document fragment can range from an empty fragment (i.e., no content
+    inside of the **svg** element), to a very simple SVG document fragment containing
+    a single SVG graphics element such as a **rect**, to a complex, deeply nested
+    collection of container elements and graphics elements.
+    """
+    elementname = 'svg'
+
+    def __init__(self, insert=None, size=None, **extra):
+        """
+        :param 2-tuple insert: insert position (**x**, **y**)
+        :param 2-tuple size: (**width**, **height**)
+        :param extra: additional SVG attributes as keyword-arguments
+        """
+        super(SVG, self).__init__(**extra)
+        if insert is not None:
+            self['x'] = insert[0]
+            self['y'] = insert[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+
+        self.defs = Defs(factory=self)  # defs container
+        self.add(self.defs)  # add defs as first element
+
+    def embed_stylesheet(self, content):
+        """ Add <style> tag to the defs section.
+
+        :param content: style sheet content as string
+        :return: :class:`~svgwrite.container.Style` object
+        """
+        return self.defs.add(Style(content))
+
+    def embed_font(self, name, filename):
+        """ Embed font as base64 encoded data from font file.
+
+        :param name: font name
+        :param filename: file name of local stored font
+        """
+        data = open(filename, 'rb').read()
+        self._embed_font_data(name, data, font_mimetype(filename))
+
+    def embed_google_web_font(self, name, uri):
+        """ Embed font as base64 encoded data acquired from google fonts.
+
+        :param name: font name
+        :param uri: google fonts request uri like 'http://fonts.googleapis.com/css?family=Indie+Flower'
+        """
+        font_info = urlopen(uri).read()
+        font_url = find_first_url(font_info.decode())
+        if font_url is None:
+            raise ValueError("Got no font data from uri: '{}'".format(uri))
+        else:
+            data = urlopen(font_url).read()
+            self._embed_font_data(name, data, font_mimetype(font_url))
+
+    def _embed_font_data(self, name, data, mimetype):
+        content = FONT_TEMPLATE.format(name=name, data=base64_data(data, mimetype))
+        self.embed_stylesheet(content)
+
+
+class Use(BaseElement, Transform, XLink, Presentation):
+    """ The **use** element references another element and indicates that the graphical
+    contents of that element is included/drawn at that given point in the document.
+
+    Link to objects by href = ``'#object-id'`` or use the object itself as
+    href-argument, if the given element has no **id** attribute it gets an
+    automatic generated id.
+
+    """
+    elementname = 'use'
+
+    def __init__(self, href, insert=None, size=None, **extra):
+        """
+        :param string href: object link (id-string) or an object with an id-attribute
+        :param 2-tuple insert: insert point (**x**, **y**)
+        :param 2-tuple size: (**width**, **height**)
+        :param extra: additional SVG attributes as keyword-arguments
+        """
+        super(Use, self).__init__(**extra)
+        self.set_href(href)
+        if insert is not None:
+            self['x'] = insert[0]
+            self['y'] = insert[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+
+    def get_xml(self):
+        self.update_id()  # if href is an object - 'id' - attribute may be changed!
+        return super(Use, self).get_xml()
+
+
+class Hyperlink(BaseElement, Transform, Presentation):
+    """ The **a** element indicate links (also known as Hyperlinks or Web links).
+
+    The remote resource (the destination for the link) is defined by a `<URI>`
+    specified by the XLink **xlink:href** attribute. The remote resource may be
+    any Web resource (e.g., an image, a video clip, a sound bite, a program,
+    another SVG document, an HTML document, an element within the current
+    document, an element within a different document, etc.). By activating
+    these links (by clicking with the mouse, through keyboard input, voice
+    commands, etc.), users may visit these resources.
+
+    A **Hyperlink** is defined for each separate rendered element
+    contained within the **Hyperlink** class; add sublements as usual with
+    the `add` method.
+
+    """
+    elementname = 'a'
+
+    def __init__(self, href, target='_blank', **extra):
+        """
+        :param string href: hyperlink to the target resource
+        :param string target: ``'_blank|_replace|_self|_parent|_top|<XML-name>'``
+        :param extra: additional SVG attributes as keyword-arguments
+        """
+        super(Hyperlink, self).__init__(**extra)
+        self['xlink:href'] = href
+        if target is not None:
+            self['target'] = target
+
+
+class Script(BaseElement):
+    """ The **script** element indicate links to a client-side language.  This
+    is normally a  (also known as Hyperlinks or Web links).
+
+    The remote resource (the source of the script) is defined by a `<URI>`
+    specified by the XLink **xlink:href** attribute. The remote resource must
+    be a text-file that contains the script contents.  This script can be used
+    within the SVG file by catching events or adding the mouseover/mousedown/
+    mouseup elements to the markup.
+
+    """
+    elementname = 'script'
+
+    def __init__(self, href=None, content="", **extra):
+        """
+        :param string href: hyperlink to the target resource or *None* if using *content*
+        :param string content: script content
+        :param extra: additional attributes as keyword-arguments
+
+        Use *href* **or** *content*, but not both at the same time.
+
+        """
+        # removed type parameter, default is "application/ecmascript"
+        super(Script, self).__init__(**extra)
+        if href is not None:
+            self['xlink:href'] = href
+        self._content = content
+
+    def get_xml(self):
+        xml = super(Script, self).get_xml()
+        if self._content:
+            xml.append(CDATA(self._content))
+        return xml
+
+    def append(self, content):
+        """ Append content to the existing element-content. """
+        self._content += content
+
+
+class Style(Script):
+    """ The *style* element allows style sheets to be embedded directly within
+    SVG content. SVG's *style* element has the same attributes as the
+    corresponding element in HTML.
+
+    """
+    elementname = 'style'
+
+    def __init__(self, content="", **extra):
+        """
+        :param string content: stylesheet content
+        """
+        super(Style, self).__init__(content=content, **extra)
+        self['type'] = "text/css"
+
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__init__.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..e76016d1
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/colors.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/colors.cpython-312.pyc
new file mode 100644
index 00000000..c6fc7ab6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/colors.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/full11.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/full11.cpython-312.pyc
new file mode 100644
index 00000000..d6fcaa47
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/full11.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/pattern.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/pattern.cpython-312.pyc
new file mode 100644
index 00000000..cfc09bb5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/pattern.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/svgparser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/svgparser.cpython-312.pyc
new file mode 100644
index 00000000..a7acc3b2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/svgparser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/tiny12.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/tiny12.cpython-312.pyc
new file mode 100644
index 00000000..66802345
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/tiny12.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/typechecker.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/typechecker.cpython-312.pyc
new file mode 100644
index 00000000..90c87b1b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/typechecker.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/types.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/types.cpython-312.pyc
new file mode 100644
index 00000000..b8016f2b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/data/__pycache__/types.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/colors.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/colors.py
new file mode 100644
index 00000000..7672194e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/colors.py
@@ -0,0 +1,156 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: colornames
+# Created: 06.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+colornames = frozenset([
+    'aliceblue',
+    'antiquewhite',
+    'aqua',
+    'aquamarine',
+    'azure',
+    'beige',
+    'bisque',
+    'black',
+    'blanchedalmond',
+    'blue',
+    'blueviolet',
+    'brown',
+    'burlywood',
+    'cadetblue',
+    'chartreuse',
+    'chocolate',
+    'coral',
+    'cornflowerblue',
+    'cornsilk',
+    'crimson',
+    'cyan',
+    'darkblue',
+    'darkcyan',
+    'darkgoldenrod',
+    'darkgray',
+    'darkgreen',
+    'darkgrey',
+    'darkkhaki',
+    'darkmagenta',
+    'darkolivegreen',
+    'darkorange',
+    'darkorchid',
+    'darkred',
+    'darksalmon',
+    'darkseagreen',
+    'darkslateblue',
+    'darkslategray',
+    'darkslategrey',
+    'darkturquoise',
+    'darkviolet',
+    'deeppink',
+    'deepskyblue',
+    'dimgray',
+    'dimgrey',
+    'dodgerblue',
+    'firebrick',
+    'floralwhite',
+    'forestgreen',
+    'fuchsia',
+    'gainsboro',
+    'ghostwhite',
+    'gold',
+    'goldenrod',
+    'gray',
+    'green',
+    'greenyellow',
+    'grey',
+    'honeydew',
+    'hotpink',
+    'indianred',
+    'indigo',
+    'ivory',
+    'khaki',
+    'lavender',
+    'lavenderblush',
+    'lawngreen',
+    'lemonchiffon',
+    'lightblue',
+    'lightcoral',
+    'lightcyan',
+    'lightgoldenrodyellow',
+    'lightgray',
+    'lightgreen',
+    'lightgrey',
+    'lightpink',
+    'lightsalmon',
+    'lightseagreen',
+    'lightskyblue',
+    'lightslategray',
+    'lightslategrey',
+    'lightsteelblue',
+    'lightyellow',
+    'lime',
+    'limegreen',
+    'linen',
+    'magenta',
+    'maroon',
+    'mediumaquamarine',
+    'mediumblue',
+    'mediumorchid',
+    'mediumpurple',
+    'mediumseagreen',
+    'mediumslateblue',
+    'mediumspringgreen',
+    'mediumturquoise',
+    'mediumvioletred',
+    'midnightblue',
+    'mintcream',
+    'mistyrose',
+    'moccasin',
+    'navajowhite',
+    'navy',
+    'oldlace',
+    'olive',
+    'olivedrab',
+    'orange',
+    'orangered',
+    'orchid',
+    'palegoldenrod',
+    'palegreen',
+    'paleturquoise',
+    'palevioletred',
+    'papayawhip',
+    'peachpuff',
+    'peru',
+    'pink',
+    'plum',
+    'powderblue',
+    'purple',
+    'red',
+    'rosybrown',
+    'royalblue',
+    'saddlebrown',
+    'salmon',
+    'sandybrown',
+    'seagreen',
+    'seashell',
+    'sienna',
+    'silver',
+    'skyblue',
+    'slateblue',
+    'slategray',
+    'slategrey',
+    'snow',
+    'springgreen',
+    'steelblue',
+    'tan',
+    'teal',
+    'thistle',
+    'tomato',
+    'turquoise',
+    'violet',
+    'wheat',
+    'white',
+    'whitesmoke',
+    'yellow',
+    'yellowgreen',])
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/full11.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/full11.py
new file mode 100644
index 00000000..fe07995a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/full11.py
@@ -0,0 +1,1405 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: full11 data
+# Created: 15.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.data.types import SVGAttribute, SVGMultiAttribute
+from svgwrite.data.types import SVGElement
+from svgwrite.data.typechecker import Full11TypeChecker as TypeChecker
+
+empty_list = []
+
+attributes = {
+'accent-height': SVGAttribute('accent-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'accumulate': SVGAttribute('accumulate', anim=False,
+    types=empty_list,
+    const=frozenset(['none', 'sum'])),
+'additive': SVGAttribute('additive', anim=False,
+    types=empty_list,
+    const=frozenset(['sum', 'replace'])),
+'alignment-baseline': SVGAttribute('alignment-baseline', anim=True,
+    types=empty_list,
+    const=frozenset(['mathematical', 'before-edge', 'central', 'baseline',
+                     'auto', 'hanging', 'ideographic', 'inherit', 'middle',
+                     'alphabetic', 'text-before-edge', 'text-after-edge',
+                     'after-edge'])),
+'alphabetic': SVGAttribute('alphabetic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'amplitude': SVGAttribute('amplitude', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'arabic-form': SVGAttribute('arabic-form', anim=False,
+    types=empty_list,
+    const=frozenset(['terminal', 'initial', 'isolated', 'medial'])),
+'ascent': SVGAttribute('ascent', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'attributeName': SVGAttribute('attributeName', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'attributeType': SVGAttribute('attributeType', anim=False,
+    types=empty_list,
+    const=frozenset(['XML', 'auto', 'CSS'])),
+'azimuth': SVGAttribute('azimuth', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'baseFrequency': SVGAttribute('baseFrequency', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'baseline-shift': SVGAttribute('baseline-shift', anim=True,
+    types=frozenset(['percentage', 'length']),
+    const=frozenset(['super', 'baseline', 'inherit', 'sub'])),
+'baseProfile': SVGAttribute('baseProfile', anim=False,
+    types=empty_list,
+    const=frozenset(['full', 'tiny', 'basic', 'none'])),
+'bbox': SVGAttribute('bbox', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'begin': SVGAttribute('begin', anim=True,
+    types=frozenset(['timing-value-list']),
+    const=frozenset(['indefinite'])),
+'bias': SVGAttribute('bias', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'by': SVGAttribute('by', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'calcMode': SVGAttribute('calcMode', anim=False,
+    types=empty_list,
+    const=frozenset(['discrete', 'linear', 'paced', 'spline'])),
+'cap-height': SVGAttribute('cap-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'class': SVGAttribute('class', anim=True,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'clip': SVGAttribute('clip', anim=True,
+    types=frozenset(['shape']),
+    const=frozenset(['auto', 'inherit'])),
+'clip-path': SVGAttribute('clip-path', anim=True,
+    types=frozenset(['IRI']),
+    const=frozenset(['none', 'inherit'])),
+'clip-rule': SVGAttribute('clip-rule', anim=True,
+    types=empty_list,
+    const=frozenset(['nonzero', 'evenodd', 'inherit'])),
+'clipPathUnits': SVGAttribute('clipPathUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'color': SVGAttribute('color', anim=True,
+    types=frozenset(['color']),
+    const=frozenset(['inherit'])),
+'color-interpolation': SVGAttribute('color-interpolation', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'sRGB', 'inherit', 'linearRGB'])),
+'color-interpolation-filters': SVGAttribute('color-interpolation-filters', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'sRGB', 'inherit', 'linearRGB'])),
+'color-profile': SVGAttribute('color-profile', anim=True,
+    types=frozenset(['FuncIRI', 'name']),
+    const=frozenset(['auto', 'sRGB', 'inherit'])),
+'color-rendering': SVGAttribute('color-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeQuality', 'optimizeSpeed', 'inherit'])),
+'contentScriptType': SVGAttribute('contentScriptType', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'contentStyleType': SVGAttribute('contentStyleType', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'cursor': SVGAttribute('cursor', anim=True,
+    types=frozenset(['list-of-FuncIRI']),
+    const=frozenset(['sw-resize', 'n-resize', 'help', 'text', 'move', 'auto',
+                     'w-resize', 'pointer', 'wait', 's-resize', 'e-resize',
+                     'default', 'inherit', 'nw-resize', 'ne-resize',
+                     'crosshair', 'se-resize'])),
+'cx': SVGAttribute('cx', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'cy': SVGAttribute('cy', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'d': SVGMultiAttribute({
+    '* path': SVGAttribute( # '*' means default attribute
+        'd', anim=True,
+        types=frozenset(['path-data']),
+        const=empty_list
+        ),
+    'glyph missing-glyph': SVGAttribute(
+        'd', anim=False,
+        types=frozenset(['path-data']),
+        const=empty_list
+        ),
+    }),
+'descent': SVGAttribute('descent', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'diffuseConstant': SVGAttribute('diffuseConstant', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'direction': SVGAttribute('direction', anim=False,
+    types=empty_list,
+    const=frozenset(['ltr', 'inherit', 'rtl'])),
+'display': SVGAttribute('display', anim=True,
+    types=empty_list,
+    const=frozenset(['inline-table', 'table-header-group', 'table-footer-group',
+                     'none', 'table-row', 'table-caption', 'table-column',
+                     'marker', 'table', 'compact', 'table-row-group', 'run-in',
+                     'inherit', 'list-item', 'table-cell', 'inline',
+                     'block', 'table-column-group'])),
+'divisor': SVGAttribute('divisor', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'dominant-baseline': SVGAttribute('dominant-baseline', anim=True,
+    types=empty_list,
+    const=frozenset(['mathematical', 'use-script', 'ideographic', 'central',
+                     'reset-size', 'auto', 'hanging', 'inherit', 'middle',
+                     'alphabetic', 'text-before-edge', 'text-after-edge',
+                     'no-change'])),
+'dur': SVGAttribute('dur', anim=True,
+    types=frozenset(['time']),
+    const=frozenset(['media', 'indefinite'])),
+'dx': SVGMultiAttribute({
+    '* altGlyph text tref tspan': SVGAttribute(
+        'dx', anim=True,
+        types=frozenset(['list-of-length']),
+        const=empty_list),
+    'feOffset': SVGAttribute(
+        'dx', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    'glyphRef': SVGAttribute(
+        'dx', anim=False,
+        types=frozenset(['number']),
+        const=empty_list),
+    }),
+'dy': SVGMultiAttribute({
+    '* altGlyph text tref tspan': SVGAttribute(
+        'dy', anim=True,
+        types=frozenset(['list-of-length']),
+        const=empty_list),
+    'feOffset': SVGAttribute(
+        'dy', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    'glyphRef': SVGAttribute(
+        'dy', anim=False,
+        types=frozenset(['number']),
+        const=empty_list),
+    }),
+'edgeMode': SVGAttribute('edgeMode', anim=True,
+    types=empty_list,
+    const=frozenset(['wrap', 'duplicate', 'none'])),
+'elevation': SVGAttribute('elevation', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'enable-background': SVGAttribute('enable-background', anim=True,
+    types=frozenset(['string']),
+    const=frozenset(['accummulate', 'new', 'inherit'])),
+'end': SVGAttribute('end', anim=False,
+    types=frozenset(['timing-value-list']),
+    const=frozenset(['indefinite'])),
+'exponent': SVGAttribute('exponent', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'externalResourcesRequired': SVGAttribute('externalResourcesRequired', anim=True,
+    types=empty_list,
+    const=frozenset(['true', 'false'])),
+'fill': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'fill', anim=True,
+        types=frozenset(['paint']),
+        const=empty_list),
+    'set animateMotion animate animateColor animateTransform': SVGAttribute(
+        'fill', anim=False,
+        types=empty_list,
+        const=frozenset(['freeze', 'remove']))
+    }),
+'fill-opacity': SVGAttribute('fill-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'fill-rule': SVGAttribute('fill-rule', anim=True,
+    types=empty_list,
+    const=frozenset(['nonzero', 'evenodd', 'inherit'])),
+'filter': SVGAttribute('filter', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'filterRes': SVGAttribute('filterRes', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'filterUnits': SVGAttribute('filterUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'flood-color': SVGAttribute('flood-color', anim=True,
+    types=frozenset(['color', 'icccolor']),
+    const=frozenset(['currentColor', 'inherit'])),
+'flood-opacity': SVGAttribute('flood-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'font': SVGAttribute('font', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'font-family': SVGAttribute('font-family', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'font-size': SVGAttribute('font-size', anim=False,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'font-size-adjust': SVGAttribute('font-size-adjust', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['none', 'inherit'])),
+'font-stretch': SVGAttribute('font-stretch', anim=False,
+    types=empty_list,
+    const=frozenset(['condensed', 'normal', 'ultra-condensed', 'expanded',
+                     'narrower', 'inherit', 'semi-condensed', 'extra-condensed',
+                     'ultra-expanded', 'wider', 'semi-expanded', 'extra-expanded'])),
+'font-style': SVGAttribute('font-style', anim=False,
+    types=empty_list,
+    const=frozenset(['oblique', 'inherit', 'italic', 'normal'])),
+'font-variant': SVGAttribute('font-variant', anim=False,
+    types=empty_list,
+    const=frozenset(['small-caps', 'inherit', 'normal'])),
+'font-weight': SVGAttribute('font-weight', anim=False,
+    types=empty_list,
+    const=frozenset(['200', '900', 'bold', 'bolder', 'normal', '300', '700',
+                     'inherit', 'lighter', '400', '100', '800', '500', '600'])),
+'format': SVGAttribute('format', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'from': SVGAttribute('from', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'fx': SVGAttribute('fx', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'fy': SVGAttribute('fy', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'g1': SVGAttribute('g1', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'g2': SVGAttribute('g2', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'glyph-name': SVGAttribute('glyph-name', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'glyph-orientation-horizontal': SVGAttribute('glyph-orientation-horizontal', anim=True,
+    types=frozenset(['angle']),
+    const=frozenset(['inherit'])),
+'glyph-orientation-vertical': SVGAttribute('glyph-orientation-vertical', anim=True,
+    types=frozenset(['angle']),
+    const=frozenset(['auto', 'inherit'])),
+'glyphRef': SVGAttribute('glyphRef', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'gradientTransform': SVGAttribute('gradientTransform', anim=True,
+    types=frozenset(['transform-list']),
+    const=empty_list),
+'gradientUnits': SVGAttribute('gradientUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'hanging': SVGAttribute('hanging', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'height': SVGAttribute('height', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'horiz-adv-x': SVGAttribute('horiz-adv-x', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'horiz-origin-x': SVGAttribute('horiz-origin-x', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'horiz-origin-y': SVGAttribute('horiz-origin-y', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'id': SVGAttribute('id', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'ideographic': SVGAttribute('ideographic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'image-rendering': SVGAttribute('image-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeQuality', 'optimizeSpeed', 'inherit'])),
+'in': SVGAttribute('in', anim=True,
+    types=frozenset(['name']),
+    const=frozenset(['SourceAlpha', 'SourceGraphic', 'BackgroundAlpha',
+                     'BackgroundImage', 'StrokePaint', 'FillPaint'])),
+'in2': SVGAttribute('in2', anim=True,
+    types=frozenset(['name']),
+    const=frozenset(['SourceAlpha', 'SourceGraphic', 'BackgroundAlpha',
+                     'BackgroundImage', 'StrokePaint', 'FillPaint'])),
+'intercept': SVGAttribute('intercept', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'k': SVGAttribute('k', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'k1': SVGAttribute('k1', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'k2': SVGAttribute('k2', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'k3': SVGAttribute('k3', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'k4': SVGAttribute('k4', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'kernelMatrix': SVGAttribute('kernelMatrix', anim=True,
+    types=frozenset(['list-of-number']),
+    const=empty_list),
+'kernelUnitLength': SVGAttribute('kernelUnitLength', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'kerning': SVGAttribute('kerning', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['auto', 'inherit'])),
+'keyPoints': SVGAttribute('keyPoints', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'keySplines': SVGAttribute('keySplines', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'keyTimes': SVGAttribute('keyTimes', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'lang': SVGAttribute('lang', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'lengthAdjust': SVGAttribute('lengthAdjust', anim=True,
+    types=empty_list,
+    const=frozenset(['spacingAndGlyphs', 'spacing'])),
+'letter-spacing': SVGAttribute('letter-spacing', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit', 'normal'])),
+'lighting-color': SVGAttribute('lighting-color', anim=True,
+    types=frozenset(['color', 'icccolor']),
+    const=frozenset(['currentColor', 'inherit'])),
+'limitingConeAngle': SVGAttribute('limitingConeAngle', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'local': SVGAttribute('local', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'marker': SVGAttribute('marker', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'marker-end': SVGAttribute('marker-end', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'marker-mid': SVGAttribute('marker-mid', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'marker-start': SVGAttribute('marker-start', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'markerHeight': SVGAttribute('markerHeight', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'markerUnits': SVGAttribute('markerUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'strokeWidth'])),
+'markerWidth': SVGAttribute('markerWidth', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'mask': SVGAttribute('mask', anim=True,
+    types=frozenset(['FuncIRI']),
+    const=frozenset(['none', 'inherit'])),
+'maskContentUnits': SVGAttribute('maskContentUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'maskUnits': SVGAttribute('maskUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'mathematical': SVGAttribute('mathematical', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'max': SVGAttribute('max', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['media'])),
+'media': SVGAttribute('media', anim=False,
+    types=empty_list,
+    const=frozenset(['all', 'aureal', 'braille', 'embossed', 'handheld', 'print',
+                     'projection', 'screen', 'tty', 'tv'])),
+'method': SVGAttribute('method', anim=True,
+    types=empty_list,
+    const=frozenset(['stretch', 'align'])),
+'min': SVGAttribute('min', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['media'])),
+'mode': SVGAttribute('mode', anim=True,
+    types=empty_list,
+    const=frozenset(['multiply', 'screen', 'darken', 'lighten', 'normal'])),
+'name': SVGMultiAttribute({
+    '* font-face-name': SVGAttribute(
+        'name', anim=False,
+        types=frozenset(['anything']),
+        const=empty_list),
+    'color-profile': SVGAttribute(
+        'name', anim=False,
+        types=frozenset(['name']),
+        const=empty_list),
+    }),
+'numOctaves': SVGAttribute('numOctaves', anim=True,
+    types=frozenset(['integer']),
+    const=empty_list),
+'offset': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'offset', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    'stop': SVGAttribute(
+        'offset', anim=True,
+        types=frozenset(['number', 'percentage']),
+        const=empty_list),
+    }),
+'onabort': SVGAttribute('onabort', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onactivate': SVGAttribute('onactivate', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onbegin': SVGAttribute('onbegin', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onclick': SVGAttribute('onclick', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onend': SVGAttribute('onend', anim=True,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onerror': SVGAttribute('onerror', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onfocusin': SVGAttribute('onfocusin', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onfocusout': SVGAttribute('onfocusout', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onload': SVGAttribute('onload', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onmousedown': SVGAttribute('onmousedown', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onmousemove': SVGAttribute('onmousemove', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onmouseout': SVGAttribute('onmouseout', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onmouseover': SVGAttribute('onmouseover', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onmouseup': SVGAttribute('onmouseup', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onrepeat': SVGAttribute('onrepeat', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onresize': SVGAttribute('onresize', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onscroll': SVGAttribute('onscroll', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onunload': SVGAttribute('onunload', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'onzoom': SVGAttribute('onzoom', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'opacity': SVGAttribute('opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'operator': SVGMultiAttribute({
+    '* feComposite': SVGAttribute(
+        'operator', anim=True,
+        types=empty_list,
+        const=frozenset(['xor', 'in', 'over', 'atop', 'arithmetic', 'out'])),
+    'feMorphology': SVGAttribute(
+        'operator', anim=True,
+        types=empty_list,
+        const=frozenset(['erode', 'dilate'])),
+    }),
+'order': SVGAttribute('order', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'orient': SVGAttribute('orient', anim=True,
+    types=frozenset(['angle']),
+    const=frozenset(['auto'])),
+'orientation': SVGAttribute('orientation', anim=False,
+    types=empty_list,
+    const=frozenset(['h', 'v'])),
+'origin': SVGAttribute('origin', anim=False,
+    types=empty_list,
+    const=frozenset(['default'])),
+'overflow': SVGAttribute('overflow', anim=True,
+    types=empty_list,
+    const=frozenset(['visible', 'hidden', 'scroll', 'inherit', 'auto'])),
+'overline-position': SVGAttribute('overline-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'overline-thickness': SVGAttribute('overline-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'panose-1': SVGAttribute('panose-1', anim=False,
+    types=frozenset(['list-of-integer']),
+    const=empty_list),
+'path': SVGAttribute('path', anim=False,
+    types=frozenset(['path-data']),
+    const=empty_list),
+'pathLength': SVGAttribute('pathLength', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'patternContentUnits': SVGAttribute('patternContentUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'patternTransform': SVGAttribute('patternTransform', anim=True,
+    types=frozenset(['transform-list']),
+    const=empty_list),
+'patternUnits': SVGAttribute('patternUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'pointer-events': SVGAttribute('pointer-events', anim=True,
+    types=empty_list,
+    const=frozenset(['all', 'visibleStroke', 'painted', 'none', 'visibleFill',
+                     'visible', 'stroke', 'inherit', 'visiblePainted', 'fill'])),
+'points': SVGAttribute('points', anim=True,
+    types=frozenset(['list-of-points']),
+    const=empty_list),
+'pointsAtX': SVGAttribute('pointsAtX', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'pointsAtY': SVGAttribute('pointsAtY', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'pointsAtZ': SVGAttribute('pointsAtZ', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'preserveAlpha': SVGAttribute('preserveAlpha', anim=True,
+    types=empty_list,
+    const=frozenset(['true', 'false'])),
+'preserveAspectRatio': SVGAttribute('preserveAspectRatio', anim=True,
+    types=frozenset('string'),
+    const=empty_list),
+'primitiveUnits': SVGAttribute('primitiveUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'r': SVGAttribute('r', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'radius': SVGAttribute('radius', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'refX': SVGAttribute('refX', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'refY': SVGAttribute('refY', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'rendering-intent': SVGAttribute('rendering-intent', anim=False,
+    types=empty_list,
+    const=frozenset(['auto', 'saturation', 'perceptual', 'relative-colorimetric',
+                     'absolute-colorimetric'])),
+'repeatCount': SVGAttribute('repeatCount', anim=False,
+    types=frozenset(['number']),
+    const=frozenset(['indefinite'])),
+'repeatDur': SVGAttribute('repeatDur', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['indefinite'])),
+'requiredExtensions': SVGAttribute('requiredExtensions', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'requiredFeatures': SVGAttribute('requiredFeatures', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'restart': SVGAttribute('restart', anim=False,
+    types=empty_list,
+    const=frozenset(['always', 'never', 'whenNotActive'])),
+'result': SVGAttribute('result', anim=True,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'rotate': SVGMultiAttribute({
+    '* altGlyph text tref tspan': SVGAttribute(
+        'rotate', anim=True,
+        types=frozenset(['list-of-number']),
+        const=empty_list),
+    'animateMotion': SVGAttribute(
+        'rotate', anim=False,
+        types=frozenset(['number']),
+        const=frozenset(['auto', 'auto-reverse'])),
+    }),
+'rx': SVGAttribute('rx', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'ry': SVGAttribute('ry', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'scale': SVGAttribute('scale', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'seed': SVGAttribute('seed', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'shape-rendering': SVGAttribute('shape-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'inherit', 'geometricPrecision',
+                     'crispEdges'])),
+'slope': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'slope', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    'font-face': SVGAttribute(
+        'slope', anim=False,
+        types=frozenset(['number']),
+        const=empty_list),
+    }),
+'spacing': SVGAttribute('spacing', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'exact'])),
+'specularConstant': SVGAttribute('specularConstant', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'specularExponent': SVGAttribute('specularExponent', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'spreadMethod': SVGAttribute('spreadMethod', anim=True,
+    types=empty_list,
+    const=frozenset(['reflect', 'repeat', 'pad'])),
+'startOffset': SVGAttribute('startOffset', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'stdDeviation': SVGAttribute('stdDeviation', anim=True,
+    types=frozenset(['number-optional-number']),
+    const=empty_list),
+'stemh': SVGAttribute('stemh', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'stemv': SVGAttribute('stemv', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'stitchTiles': SVGAttribute('stitchTiles', anim=True,
+    types=empty_list,
+    const=frozenset(['noStitch', 'stitch'])),
+'stop-color': SVGAttribute('stop-color', anim=True,
+    types=frozenset(['color', 'icccolor']),
+    const=frozenset(['currentColor', 'inherit'])),
+'stop-opacity': SVGAttribute('stop-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'strikethrough-position': SVGAttribute('strikethrough-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'strikethrough-thickness': SVGAttribute('strikethrough-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'string': SVGAttribute('string', anim=False,
+    types=frozenset('anything'),
+    const=empty_list),
+'stroke': SVGAttribute('stroke', anim=True,
+    types=frozenset(['paint']),
+    const=empty_list),
+'stroke-dasharray': SVGAttribute('stroke-dasharray', anim=True,
+    types=frozenset(['list-of-length']),
+    const=frozenset(['none', 'inherit'])),
+'stroke-dashoffset': SVGAttribute('stroke-dashoffset', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'stroke-linecap': SVGAttribute('stroke-linecap', anim=True,
+    types=empty_list,
+    const=frozenset(['square', 'round', 'inherit', 'butt'])),
+'stroke-linejoin': SVGAttribute('stroke-linejoin', anim=True,
+    types=empty_list,
+    const=frozenset(['bevel', 'miter', 'round', 'inherit'])),
+'stroke-miterlimit': SVGAttribute('stroke-miterlimit', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'stroke-opacity': SVGAttribute('stroke-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'stroke-width': SVGAttribute('stroke-width', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'style': SVGAttribute('style', anim=False,
+    types=frozenset('anything'),
+    const=empty_list),
+'surfaceScale': SVGAttribute('surfaceScale', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'systemLanguage': SVGAttribute('systemLanguage', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'tableValues': SVGAttribute('tableValues', anim=True,
+    types=frozenset(['list-of-number']),
+    const=empty_list),
+'target': SVGAttribute('target', anim=True,
+    types=frozenset(['XML-Name']),
+    const=frozenset(['_replace', '_self', '_parent', '_top', '_blank'])),
+'targetX': SVGAttribute('targetX', anim=True,
+    types=frozenset(['integer']),
+    const=empty_list),
+'targetY': SVGAttribute('targetY', anim=True,
+    types=frozenset(['integer']),
+    const=empty_list),
+'text-anchor': SVGAttribute('text-anchor', anim=True,
+    types=empty_list,
+    const=frozenset(['start', 'end', 'inherit', 'middle'])),
+'text-decoration': SVGAttribute('text-decoration', anim=True,
+    types=frozenset(['list-of-text-decoration-style']),
+    const=frozenset(['', 'none', 'inherit'])),
+'text-rendering': SVGAttribute('text-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'optimizeLegibility',
+                     'geometricPrecision', 'inherit'])),
+'textLength': SVGAttribute('textLength', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'title': SVGAttribute('title', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'to': SVGAttribute('to', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'transform': SVGAttribute('transform', anim=True,
+    types=frozenset(['transform-list']),
+    const=empty_list),
+'type': SVGMultiAttribute({
+    '* feColorMatrix': SVGAttribute(
+        'type', anim=True,
+        types=empty_list,
+        const=frozenset(['matrix', 'saturate', 'hueRotate', 'luminanceToAlpha'])),
+    'feTurbulence': SVGAttribute(
+        'type', anim=True,
+        types=empty_list,
+        const=frozenset(['fractalNoise', 'turbulence'])),
+    'feFuncR feFuncG feFuncB feFuncA': SVGAttribute(
+        'type', anim=True,
+        types=empty_list,
+        const=frozenset(['identity', 'table', 'discrete', 'linear', 'gamma'])),
+    'script style': SVGAttribute(
+        'type', anim=False,
+        types=frozenset(['content-type']),
+        const=empty_list),
+    'animateTransform': SVGAttribute(
+        'type', anim=False,
+        types=empty_list,
+        const=frozenset(['translate', 'scale', 'rotate', 'skewX', 'skewY'])),
+    }),
+'u1': SVGAttribute('u1', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'u2': SVGAttribute('u2', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'underline-position': SVGAttribute('underline-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'underline-thickness': SVGAttribute('underline-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'unicode': SVGAttribute('unicode', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'unicode-bidi': SVGAttribute('unicode-bidi', anim=False,
+    types=empty_list,
+    const=frozenset(['embed', 'inherit', 'bidi-override', 'normal'])),
+'unicode-range': SVGAttribute('unicode-range', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'units-per-em': SVGAttribute('units-per-em', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'v-alphabetic': SVGAttribute('v-alphabetic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'v-hanging': SVGAttribute('v-hanging', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'v-ideographic': SVGAttribute('v-ideographic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'v-mathematical': SVGAttribute('v-mathematical', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'values': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'values', anim=False,
+        types=frozenset(['semicolon-list']),
+        const=empty_list),
+    'feColorMatrix': SVGAttribute(
+        'values', anim=True,
+        types=frozenset(['list-of-number']),
+        const=empty_list),
+    }),
+'version': SVGAttribute('version', anim=False,
+    types=empty_list,
+    const=frozenset(['1.1', '1.2'])),
+'vert-adv-y': SVGAttribute('vert-adv-y', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'vert-origin-x': SVGAttribute('vert-origin-x', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'vert-origin-y': SVGAttribute('vert-origin-y', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'viewBox': SVGAttribute('viewBox', anim=True,
+    types=frozenset(['four-numbers']),
+    const=empty_list),
+'viewTarget': SVGAttribute('viewTarget', anim=False,
+    types=frozenset(['list-of-XML-Name']),
+    const=empty_list),
+'visibility': SVGAttribute('visibility', anim=True,
+    types=empty_list,
+    const=frozenset(['visible', 'hidden', 'collapse', 'inherit'])),
+'width': SVGAttribute('width', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'widths': SVGAttribute('widths', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'word-spacing': SVGAttribute('word-spacing', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit', 'normal'])),
+'writing-mode': SVGAttribute('writing-mode', anim=False,
+    types=empty_list,
+    const=frozenset(['rl-tb', 'lr', 'rl', 'tb-rl', 'lr-tb', 'inherit', 'tb'])),
+'x': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'x', anim=True,
+        types=frozenset(['coordinate']),
+        const=empty_list),
+    'altGlyph text tref tspan': SVGAttribute(
+        'x', anim=True,
+        types=frozenset(['list-of-coordinate']),
+        const=empty_list),
+    'fePointLight feSpotLight glyphRef': SVGAttribute(
+        'x', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    }),
+'x-height': SVGAttribute('x-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'x1': SVGAttribute('x1', anim=True,
+    types=frozenset(['list-of-coordinate']),
+    const=empty_list),
+'x2': SVGAttribute('x2', anim=True,
+    types=frozenset(['list-of-coordinate']),
+    const=empty_list),
+'xChannelSelector': SVGAttribute('xChannelSelector', anim=True,
+    types=empty_list,
+    const=frozenset(['A', 'B', 'R', 'G'])),
+'xlink:actuate': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'xlink:actuate', anim=False,
+        types=empty_list,
+        const=frozenset(['onLoad'])),
+    'a': SVGAttribute(
+        'xlink:actuate', anim=False,
+        types=empty_list,
+        const=frozenset(['onRequest'])),
+    }),
+'xlink:arcrole': SVGAttribute('xlink:arcrole', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:href': SVGAttribute('xlink:href', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:role': SVGAttribute('xlink:role', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:show': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'xlink:show', anim=False,
+        types=empty_list,
+        const=frozenset(['other', 'new', 'replace', 'none', 'embed'])),
+    'a': SVGAttribute(
+        'xlink:show', anim=False,
+        types=empty_list,
+        const=frozenset(['new', 'replace'])),
+    }),
+'xlink:title': SVGAttribute('xlink:title', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'xlink:type': SVGAttribute('xlink:type', anim=False,
+    types=empty_list,
+    const=frozenset(['simple'])),
+'xmlns': SVGAttribute('xmlns', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xmlns:xlink': SVGAttribute('xmlns:xlink', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xmlns:ev': SVGAttribute('xmlns:ev', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xml:base': SVGAttribute('xml:base', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xml:lang': SVGAttribute('xml:lang', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'xml:space': SVGAttribute('xml:space', anim=False,
+    types=empty_list,
+    const=frozenset(['default', 'preserve'])),
+'y': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'y', anim=True,
+        types=frozenset(['coordinate']),
+        const=empty_list),
+    'altGlyph text tref tspan': SVGAttribute(
+        'y', anim=True,
+        types=frozenset(['list-of-coordinate']),
+        const=empty_list),
+    'fePointLight feSpotLight glyphRef': SVGAttribute(
+        'y', anim=True,
+        types=frozenset(['number']),
+        const=empty_list),
+    }),
+'y1': SVGAttribute('y1', anim=True,
+    types=frozenset(['list-of-coordinate']),
+    const=empty_list),
+'y2': SVGAttribute('y2', anim=True,
+    types=frozenset(['list-of-coordinate']),
+    const=empty_list),
+'yChannelSelector': SVGAttribute('yChannelSelector', anim=True,
+    types=empty_list,
+    const=frozenset(['A', 'B', 'R', 'G'])),
+'z': SVGAttribute('z', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'zoomAndPan': SVGAttribute('zoomAndPan', anim=False,
+    types=empty_list,
+    const=frozenset(['disable', 'magnify'])),
+}
+
+presentation_attributes = frozenset([ "alignment-baseline", "baseline-shift",
+    "clip", "clip-path", "clip-rule", "color", "color-interpolation",
+    "color-interpolation-filters", "color-profile", "color-rendering", "cursor",
+    "direction", "display", "dominant-baseline", "enable-background",
+    "fill", "fill-opacity", "fill-rule", "filter", "flood-color",
+    "flood-opacity", "font-family", "font-size", "font-size-adjust",
+    "font-stretch", "font-style", "font-variant", "font-weight",
+    "glyph-orientation-horizontal", "glyph-orientation-vertical",
+    "image-rendering", "kerning", "letter-spacing", "lighting-color", "marker",
+    "marker-end", "marker-mid", "marker-start", "mask", "opacity",
+    "overflow", "pointer-events", "shape-rendering", "stop-color",
+    "stop-opacity", "stroke", "stroke-dasharray", "stroke-dashoffset",
+    "stroke-linecap", "stroke-linejoin", "stroke-miterlimit",
+    "stroke-opacity", "stroke-width", "text-anchor", "text-decoration",
+    "text-rendering", "unicode-bidi", "visibility", "word-spacing",
+    "writing-mode"])
+
+elements = {
+    'a': SVGElement('a',
+    attributes=frozenset(['xlink:title', 'xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'systemLanguage', 'onmouseover', 'xlink:type', 'externalResourcesRequired', 'id', 'xlink:actuate', 'onload', 'style', 'xlink:show', 'target', 'onactivate', 'onmousedown', 'transform', 'class', 'xlink:role', 'requiredFeatures', 'xml:lang', 'onmousemove', 'xmlns:xlink', 'onclick', 'xlink:arcrole', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'altGlyph': SVGElement('altGlyph',
+    attributes=frozenset(['requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'id', 'onload', 'style', 'onmousedown', 'onmousemove', 'onclick', 'xlink:arcrole', 'onfocusin', 'xml:base', 'onmouseup', 'onmouseout', 'format', 'xlink:title', 'systemLanguage', 'onmouseover', 'dx', 'dy', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'xlink:show', 'onactivate', 'glyphRef', 'xlink:role', 'requiredFeatures', 'xml:lang', 'y', 'x', 'rotate']),
+    properties=presentation_attributes,
+    children=empty_list),
+
+    'altGlyphDef': SVGElement('altGlyphDef',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'altGlyphItem': SVGElement('altGlyphItem',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'animate': SVGElement('animate',
+    attributes=frozenset(['requiredExtensions', 'from', 'repeatCount', 'xml:space', 'xlink:href', 'xlink:type', 'attributeType', 'repeatDur', 'id', 'fill', 'onload', 'additive', 'calcMode', 'min', 'keySplines', 'to', 'dur', 'xlink:arcrole', 'onend', 'begin', 'xml:base', 'max', 'xlink:title', 'attributeName', 'onbegin', 'systemLanguage', 'accumulate', 'end', 'externalResourcesRequired', 'by', 'restart', 'xlink:actuate', 'xlink:show', 'xlink:role', 'requiredFeatures', 'xml:lang', 'values', 'keyTimes', 'onrepeat']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'animateColor': SVGElement('animateColor',
+    attributes=frozenset(['requiredExtensions', 'from', 'repeatCount', 'xml:space', 'xlink:href', 'xlink:type', 'attributeType', 'repeatDur', 'id', 'fill', 'onload', 'additive', 'calcMode', 'min', 'keySplines', 'to', 'dur', 'xlink:arcrole', 'onend', 'begin', 'xml:base', 'max', 'xlink:title', 'attributeName', 'onbegin', 'systemLanguage', 'accumulate', 'end', 'externalResourcesRequired', 'by', 'restart', 'xlink:actuate', 'xlink:show', 'xlink:role', 'requiredFeatures', 'xml:lang', 'values', 'keyTimes', 'onrepeat']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'animateMotion': SVGElement('animateMotion',
+    attributes=frozenset(['origin', 'requiredExtensions', 'from', 'repeatCount', 'xml:space', 'xlink:href', 'xlink:type', 'repeatDur', 'id', 'fill', 'onload', 'additive', 'calcMode', 'min', 'keySplines', 'to', 'dur', 'xlink:arcrole', 'onend', 'begin', 'xlink:title', 'xml:base', 'max', 'end', 'keyPoints', 'onbegin', 'systemLanguage', 'accumulate', 'path', 'externalResourcesRequired', 'by', 'restart', 'xlink:actuate', 'xlink:show', 'xlink:role', 'requiredFeatures', 'xml:lang', 'values', 'keyTimes', 'onrepeat', 'rotate']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'mpath', 'title'])),
+
+    'animateTransform': SVGElement('animateTransform',
+    attributes=frozenset(['requiredExtensions', 'from', 'repeatCount', 'xml:space', 'xlink:href', 'xlink:type', 'attributeType', 'repeatDur', 'id', 'fill', 'onload', 'additive', 'calcMode', 'min', 'keySplines', 'to', 'dur', 'xlink:arcrole', 'type', 'onend', 'begin', 'xml:base', 'max', 'xlink:title', 'attributeName', 'onbegin', 'systemLanguage', 'accumulate', 'end', 'externalResourcesRequired', 'by', 'restart', 'xlink:actuate', 'xlink:show', 'xlink:role', 'requiredFeatures', 'xml:lang', 'values', 'keyTimes', 'onrepeat']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'circle': SVGElement('circle',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'cy', 'cx', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'class', 'requiredFeatures', 'r', 'onmousemove', 'onclick', 'xml:lang', 'onfocusin', 'systemLanguage']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'metadata', 'desc'])),
+
+    'clipPath': SVGElement('clipPath',
+    attributes=frozenset(['clipPathUnits', 'style', 'xml:base', 'requiredExtensions', 'xml:space', 'transform', 'id', 'requiredFeatures', 'xml:lang', 'externalResourcesRequired', 'class', 'systemLanguage']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'animate', 'text', 'use', 'animateColor', 'polyline', 'path', 'line', 'ellipse', 'rect', 'desc', 'animateMotion', 'polygon', 'title', 'animateTransform', 'circle', 'metadata'])),
+
+    'color-profile': SVGElement('color-profile',
+    attributes=frozenset(['xlink:actuate', 'xlink:show', 'xml:base', 'name', 'rendering-intent', 'xml:space', 'xlink:href', 'xlink:role', 'xml:lang', 'xlink:type', 'xlink:title', 'xlink:arcrole', 'local', 'id']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'cursor': SVGElement('cursor',
+    attributes=frozenset(['xlink:title', 'xml:base', 'requiredExtensions', 'xml:space', 'xlink:href', 'systemLanguage', 'xlink:type', 'externalResourcesRequired', 'id', 'xlink:actuate', 'xlink:show', 'xlink:role', 'requiredFeatures', 'xml:lang', 'y', 'x', 'xlink:arcrole']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'defs': SVGElement('defs',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'class', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'id', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'desc': SVGElement('desc',
+    attributes=frozenset(['style', 'xml:lang', 'xml:base', 'xml:space', 'class', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'ellipse': SVGElement('ellipse',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'ry', 'cy', 'cx', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'onactivate', 'onmousedown', 'rx', 'transform', 'class', 'requiredFeatures', 'systemLanguage', 'onmousemove', 'onclick', 'xml:lang', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'feBlend': SVGElement('feBlend',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'in2', 'height', 'width', 'xml:lang', 'id', 'result', 'in', 'y', 'x', 'class', 'mode']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feColorMatrix': SVGElement('feColorMatrix',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'id', 'height', 'width', 'xml:lang', 'values', 'result', 'in', 'y', 'x', 'type', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feComponentTransfer': SVGElement('feComponentTransfer',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'height', 'width', 'xml:lang', 'id', 'result', 'in', 'y', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['feFuncA', 'feFuncR', 'feFuncB', 'feFuncG'])),
+
+    'feComposite': SVGElement('feComposite',
+    attributes=frozenset(['xml:base', 'xml:space', 'in2', 'height', 'result', 'in', 'operator', 'class', 'style', 'width', 'id', 'k3', 'k2', 'k1', 'xml:lang', 'k4', 'y', 'x']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feConvolveMatrix': SVGElement('feConvolveMatrix',
+    attributes=frozenset(['xml:base', 'xml:space', 'kernelUnitLength', 'edgeMode', 'height', 'bias', 'result', 'in', 'preserveAlpha', 'id', 'style', 'divisor', 'kernelMatrix', 'width', 'xml:lang', 'targetX', 'targetY', 'y', 'x', 'class2', 'order']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feDiffuseLighting': SVGElement('feDiffuseLighting',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'diffuseConstant', 'height', 'kernelUnitLength', 'width', 'xml:lang', 'id', 'result', 'in', 'y', 'x', 'class', 'surfaceScale']),
+    properties=presentation_attributes,
+    children=frozenset(['fePointLight', 'feSpotLight', 'title', 'metadata', 'feDistantLight', 'desc'])),
+
+    'feDisplacementMap': SVGElement('feDisplacementMap',
+    attributes=frozenset(['xml:base', 'xml:space', 'yChannelSelector', 'in2', 'height', 'result', 'in', 'class', 'style', 'scale', 'id', 'width', 'xml:lang', 'xChannelSelector', 'y', 'x']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feDistantLight': SVGElement('feDistantLight',
+    attributes=frozenset(['xml:lang', 'elevation', 'azimuth', 'xml:base', 'xml:space', 'id']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feFlood': SVGElement('feFlood',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'height', 'width', 'xml:lang', 'id', 'result', 'y', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set', 'animateColor'])),
+
+    'feFuncA': SVGElement('feFuncA',
+    attributes=frozenset(['slope', 'xml:base', 'tableValues', 'xml:space', 'xml:lang', 'intercept', 'amplitude', 'offset', 'type', 'id', 'exponent']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feFuncB': SVGElement('feFuncB',
+    attributes=frozenset(['slope', 'xml:base', 'tableValues', 'xml:space', 'xml:lang', 'intercept', 'amplitude', 'offset', 'type', 'id', 'exponent']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feFuncG': SVGElement('feFuncG',
+    attributes=frozenset(['slope', 'xml:base', 'tableValues', 'xml:space', 'xml:lang', 'intercept', 'amplitude', 'offset', 'type', 'id', 'exponent']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feFuncR': SVGElement('feFuncR',
+    attributes=frozenset(['slope', 'xml:base', 'tableValues', 'xml:space', 'xml:lang', 'intercept', 'amplitude', 'offset', 'type', 'id', 'exponent']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feGaussianBlur': SVGElement('feGaussianBlur',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'height', 'width', 'xml:lang', 'id', 'result', 'in', 'y', 'x', 'stdDeviation', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feImage': SVGElement('feImage',
+    attributes=frozenset(['xlink:title', 'xml:base', 'xml:space', 'xlink:href', 'height', 'result', 'xlink:type', 'externalResourcesRequired', 'preserveAsectRatio', 'class', 'xlink:actuate', 'style', 'xlink:show', 'id', 'xlink:role', 'width', 'xml:lang', 'y', 'x', 'xlink:arcrole']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set', 'animateColor'])),
+
+    'feMerge': SVGElement('feMerge',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'height', 'width', 'xml:lang', 'id', 'result', 'y', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set', 'feMergeNode'])),
+
+    'feMergeNode': SVGElement('feMergeNode',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id', 'in']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feMorphology': SVGElement('feMorphology',
+    attributes=frozenset(['style', 'xml:base', 'y', 'xml:space', 'id', 'height', 'width', 'xml:lang', 'radius', 'result', 'in', 'operator', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feOffset': SVGElement('feOffset',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'in', 'height', 'width', 'xml:lang', 'id', 'result', 'dx', 'dy', 'y', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'fePointLight': SVGElement('fePointLight',
+    attributes=frozenset(['xml:lang', 'xml:base', 'y', 'x', 'xml:space', 'z', 'id']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feSpecularLighting': SVGElement('feSpecularLighting',
+    attributes=frozenset(['specularConstant', 'xml:base', 'xml:space', 'kernelUnitLength', 'height', 'result', 'in', 'class', 'style', 'id', 'width', 'xml:lang', 'specularExponent', 'y', 'x', 'surfaceScale']),
+    properties=presentation_attributes,
+    children=frozenset(['fePointLight', 'feSpotLight', 'title', 'metadata', 'feDistantLight', 'desc'])),
+
+    'feSpotLight': SVGElement('feSpotLight',
+    attributes=frozenset(['pointsAtX', 'xml:base', 'xml:space', 'limitingConeAngle', 'xml:lang', 'specularExponent', 'pointsAtZ', 'y', 'x', 'pointsAtY', 'z', 'id']),
+    properties=empty_list,
+    children=frozenset(['animate', 'set'])),
+
+    'feTile': SVGElement('feTile',
+    attributes=frozenset(['style', 'xml:base', 'xml:space', 'height', 'width', 'xml:lang', 'id', 'result', 'in', 'y', 'x', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set'])),
+
+    'feTurbulence': SVGElement('feTurbulence',
+    attributes=frozenset(['xml:base', 'baseFrequency', 'xml:space', 'stitchTiles', 'height', 'width', 'xml:lang', 'id', 'result', 'x', 'y', 'numOctaves', 'type', 'seed']),
+    properties=presentation_attributes,
+    children=empty_list),
+
+    'filter': SVGElement('filter',
+    attributes=frozenset(['xlink:title', 'xml:base', 'xml:space', 'xlink:href', 'height', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'style', 'xlink:show', 'filterRes', 'primitiveUnits', 'id', 'xlink:role', 'width', 'xml:lang', 'y', 'x', 'xlink:arcrole', 'filterUnits']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'animate', 'metadata', 'desc', 'feBlend', 'feColorMatrix', 'feComponentTransfer',
+                               'feComposite', 'feConvolveMatrix', 'feDiffuseLighting',
+                               'feDisplacementMap', 'feFlood', 'feGaussianBlur',
+                               'feImage', 'feMerge', 'feMorphology', 'feOffset',
+                               'feSpecularLighting', 'feTile', 'feTurbulence'])),
+    'font': SVGElement('font',
+    attributes=frozenset(['xml:space', 'id', 'xml:lang', 'xml:base', 'class', 'style', 'externalResourcesRequired', 'horiz-origin-x', 'horiz-origin-y', 'horiz-adv-x', 'vert-origin-x', 'vert-origin-y', 'vert-adv-y']),
+    properties=presentation_attributes,
+    children=frozenset(['title', 'metadata', 'missing-glyph', 'font-face', 'vkern', 'hkern', 'glyph', 'desc'])),
+
+    'font-face': SVGElement('font-face',
+    attributes=frozenset(['mathematical', 'slope', 'font-size', 'xml:space', 'v-hanging', 'hanging', 'overline-thickness', 'ascent', 'id', 'strikethrough-position', 'underline-position', 'descent', 'cap-height', 'units-per-em', 'font-style', 'unicode-range', 'font-stretch', 'font-variant', 'x-height', 'font-weight', 'xml:base', 'panose-1', 'strikethrough-thickness', 'stemh', 'v-alphabetic', 'stemv', 'bbox', 'underline-thickness', 'font-family', 'v-mathematical', 'v-ideographic', 'ideographic', 'overline-position', 'widths', 'xml:lang', 'accent-height', 'alphabetic']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'font-face-src', 'title'])),
+
+    'font-face-format': SVGElement('font-face-format',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id']),
+    properties=empty_list,
+    children=empty_list),
+
+    'font-face-name': SVGElement('font-face-name',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id', 'name']),
+    properties=empty_list,
+    children=empty_list),
+
+    'font-face-uri': SVGElement('font-face-uri',
+    attributes=frozenset(['xlink:actuate', 'xlink:show', 'xml:base', 'xml:space', 'xlink:href', 'xlink:role', 'xml:lang', 'xlink:type', 'xlink:title', 'xlink:arcrole', 'id']),
+    properties=empty_list,
+    children=frozenset(['font-face-format'])),
+
+    'foreignObject': SVGElement('foreignObject',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'height', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'class', 'width', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'y', 'x', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['*'])),
+
+    'g': SVGElement('g',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'class', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'id', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'glyph': SVGElement('glyph',
+    attributes=frozenset(['xml:base', 'xml:space', 'id', 'xml:lang', 'class', 'style','d','horiz-adv-x', 'vert-origin-x', 'vert-origin-y', 'vert-adv-y', 'unicode', 'glyph-name', 'orientation', 'arabic-form', 'lang']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'glyphRef': SVGElement('glyphRef',
+    attributes=frozenset(['xlink:title', 'xml:base', 'format', 'xml:space', 'xlink:href', 'dx', 'dy', 'xlink:type', 'class', 'xlink:actuate', 'style', 'xlink:show', 'id', 'xlink:role', 'xml:lang', 'y', 'x', 'xlink:arcrole', 'glyphRef']),
+    properties=presentation_attributes,
+    children=empty_list),
+
+    'hkern': SVGElement('hkern',
+    attributes=frozenset(['xml:base', 'g2', 'g1', 'xml:space', 'u1', 'u2', 'xml:lang', 'id', 'k']),
+    properties=empty_list,
+    children=empty_list),
+
+    'image': SVGElement('image',
+    attributes=frozenset(['requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'height', 'id', 'onload', 'style', 'onmousedown', 'transform', 'width', 'onmousemove', 'onclick', 'xlink:arcrole', 'onfocusin', 'xml:base', 'onmouseup', 'onmouseout', 'xlink:title', 'systemLanguage', 'onmouseover', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'xlink:show', 'onactivate', 'xlink:role', 'requiredFeatures', 'xml:lang', 'y', 'x', 'preserveAspectRatio']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'line': SVGElement('line',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'x2', 'systemLanguage', 'onmouseover', 'y1', 'externalResourcesRequired', 'y2', 'id', 'onload', 'style', 'x1', 'onactivate', 'onmousedown', 'transform', 'class', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'linearGradient': SVGElement('linearGradient',
+    attributes=frozenset(['xlink:title', 'xml:base', 'xml:space', 'xlink:href', 'x2', 'y1', 'externalResourcesRequired', 'y2', 'class', 'xlink:actuate', 'xlink:role', 'style', 'xlink:show', 'spreadMethod', 'id', 'gradientUnits', 'xml:lang', 'gradientTransform', 'xlink:type', 'xlink:arcrole', 'x1']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'title', 'animate', 'animateTransform', 'stop', 'metadata', 'desc'])),
+
+    'marker': SVGElement('marker',
+    attributes=frozenset(['xml:space', 'id', 'xml:lang', 'xml:base', 'class', 'style', 'externalResourcesRequired', 'viewBox', 'preserveAspectRatio', 'refX', 'refY', 'markerUnits', 'markerWidth',  'markerHeight', 'orient']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'mask': SVGElement('mask',
+    attributes=frozenset(['xml:base', 'requiredExtensions', 'xml:space', 'height', 'systemLanguage', 'externalResourcesRequired', 'maskContentUnits', 'class', 'style', 'id', 'width', 'requiredFeatures', 'xml:lang', 'y', 'x', 'maskUnits']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'metadata': SVGElement('metadata',
+    attributes=frozenset(['xml:space', 'xml:lang', 'xml:base', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'missing-glyph': SVGElement('missing-glyph',
+    attributes=frozenset(['xml:base', 'xml:space', 'id', 'xml:lang', 'class', 'style', 'd', 'horiz-adv-x', 'vert-origin-x', 'vert-origin-y', 'vert-adv-y']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'mpath': SVGElement('mpath',
+    attributes=frozenset(['xlink:actuate', 'xlink:show', 'xml:base', 'xml:space', 'xlink:href', 'id', 'xlink:role', 'xml:lang', 'xlink:type', 'xlink:title', 'xlink:arcrole', 'externalResourcesRequired']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'path': SVGElement('path',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'pathLength', 'externalResourcesRequired', 'id', 'onload', 'style', 'd', 'onactivate', 'onmousedown', 'transform', 'class', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'pattern': SVGElement('pattern',
+    attributes=frozenset(['xlink:title', 'xml:base', 'requiredExtensions', 'xml:space', 'xlink:href', 'height', 'class', 'systemLanguage', 'patternContentUnits', 'xlink:type', 'externalResourcesRequired', 'id', 'xlink:actuate', 'style', 'xlink:show', 'viewBox', 'xlink:role', 'width', 'requiredFeatures', 'patternUnits', 'patternTransform', 'y', 'x', 'preserveAspectRatio', 'xlink:arcrole', 'xml:lang']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'polygon': SVGElement('polygon',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'class', 'requiredFeatures', 'points', 'onmousemove', 'onclick', 'xml:lang', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'polyline': SVGElement('polyline',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'class', 'requiredFeatures', 'points', 'onmousemove', 'onclick', 'xml:lang', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'radialGradient': SVGElement('radialGradient',
+    attributes=frozenset(['xlink:title', 'xml:base', 'fx', 'fy', 'xml:space', 'xlink:href', 'cy', 'cx', 'xlink:type', 'externalResourcesRequired', 'r', 'id', 'xlink:actuate', 'gradientUnits', 'style', 'xlink:show', 'spreadMethod', 'class', 'xlink:role', 'xml:lang', 'gradientTransform', 'xlink:arcrole']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'title', 'animate', 'animateTransform', 'stop', 'metadata', 'desc'])),
+
+    'rect': SVGElement('rect',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'height', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'id', 'onload', 'style', 'ry', 'onactivate', 'onmousedown', 'rx', 'transform', 'class', 'width', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'y', 'x', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'metadata', 'desc'])),
+
+    'script': SVGElement('script',
+    attributes=frozenset(['xlink:actuate', 'xlink:show', 'xml:base', 'xml:space', 'xlink:href', 'id', 'xlink:role', 'xml:lang', 'xlink:type', 'xlink:title', 'xlink:arcrole', 'type', 'externalResourcesRequired']),
+    properties=empty_list,
+    children=empty_list),
+
+    'set': SVGElement('set',
+    attributes=frozenset(['begin', 'xlink:title', 'xml:base', 'requiredExtensions', 'repeatCount', 'xml:space', 'xlink:href', 'attributeName', 'onbegin', 'systemLanguage', 'attributeType', 'xlink:type', 'externalResourcesRequired', 'id', 'restart', 'fill', 'xlink:actuate', 'onload', 'xlink:show', 'end', 'min', 'repeatDur', 'xlink:role', 'to', 'requiredFeatures', 'xml:lang', 'max', 'dur', 'xlink:arcrole', 'onrepeat', 'onend']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'stop': SVGElement('stop',
+    attributes=frozenset(['xml:base', 'xml:space', 'xml:lang', 'offset', 'style', 'class']),
+    properties=presentation_attributes,
+    children=frozenset(['animate', 'set', 'animateColor'])),
+
+    'style': SVGElement('style',
+    attributes=frozenset(['xml:lang', 'xml:base', 'title', 'media', 'xml:space', 'type', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'svg': SVGElement('svg',
+    attributes=frozenset(['requiredExtensions', 'onerror', 'onfocusout', 'xml:space', 'height', 'onscroll', 'baseProfile', 'contentStyleType', 'id', 'onabort', 'onload', 'style', 'onmousedown', 'onzoom', 'onresize', 'width', 'version', 'onmousemove', 'onmouseup', 'xmlns:xlink', 'xmlns:ev', 'onfocusin', 'xml:base', 'onclick', 'onmouseout', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'zoomAndPan', 'class', 'onunload', 'xmlns', 'onactivate', 'viewBox', 'requiredFeatures', 'xml:lang', 'y', 'x', 'preserveAspectRatio', 'contentScriptType']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'switch': SVGElement('switch',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'systemLanguage', 'onmouseover', 'externalResourcesRequired', 'class', 'onload', 'style', 'onactivate', 'onmousedown', 'transform', 'id', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'line', 'use', 'animateColor', 'polyline', 'path', 'animate', 'ellipse', 'rect', 'desc', 'a', 'animateMotion', 'polygon', 'g', 'title', 'svg', 'switch', 'animateTransform', 'foreignObject', 'circle', 'metadata'])),
+
+    'symbol': SVGElement('symbol',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'onfocusout', 'xml:space', 'onmouseover', 'id', 'externalResourcesRequired', 'viewBox', 'onload', 'style', 'onactivate', 'onmousedown', 'class', 'xml:lang', 'onmousemove', 'onclick', 'preserveAspectRatio', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'text', 'image', 'font-face', 'polyline', 'marker', 'animate', 'font', 'color-profile', 'ellipse', 'cursor', 'style', 'polygon', 'title', 'pattern', 'circle', 'radialGradient', 'metadata', 'defs', 'symbol', 'use', 'animateMotion', 'animateColor', 'path', 'line', 'rect', 'desc', 'a', 'g', 'svg', 'script', 'mask', 'altGlyphDef', 'filter', 'switch', 'animateTransform', 'linearGradient', 'clipPath', 'foreignObject', 'view'])),
+
+    'text': SVGElement('text',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'class', 'systemLanguage', 'onmouseover', 'dx', 'dy', 'externalResourcesRequired', 'lengthAdjust', 'onload', 'style', 'rotate', 'onactivate', 'onmousedown', 'textLength', 'transform', 'id', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'y', 'x', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['a', 'animateMotion', 'set', 'title', 'textPath', 'tspan', 'animateColor', 'tref', 'animateTransform', 'altGlyph', 'animate', 'desc', 'metadata'])),
+
+    'textPath': SVGElement('textPath',
+    attributes=frozenset(['requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'startOffset', 'id', 'onload', 'style', 'onmousedown', 'lengthAdjust', 'onmousemove', 'onclick', 'xlink:arcrole', 'onfocusin', 'xml:base', 'onmouseup', 'onmouseout', 'xlink:title', 'spacing', 'systemLanguage', 'onmouseover', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'xlink:show', 'onactivate', 'textLength', 'method', 'xlink:role', 'requiredFeatures', 'xml:lang']),
+    properties=presentation_attributes,
+    children=frozenset(['a', 'set', 'title', 'tspan', 'animateColor', 'tref', 'altGlyph', 'animate', 'metadata', 'desc'])),
+
+    'title': SVGElement('title',
+    attributes=frozenset(['style', 'xml:lang', 'xml:base', 'xml:space', 'class', 'id']),
+    properties=empty_list,
+    children=frozenset(['*'])),
+
+    'tref': SVGElement('tref',
+    attributes=frozenset(['requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'id', 'onload', 'style', 'onmousedown', 'lengthAdjust', 'onmousemove', 'onclick', 'xlink:arcrole', 'onfocusin', 'xml:base', 'onmouseup', 'onmouseout', 'xlink:title', 'systemLanguage', 'onmouseover', 'dx', 'dy', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'xlink:show', 'onactivate', 'textLength', 'xlink:role', 'requiredFeatures', 'xml:lang', 'y', 'x', 'rotate']),
+    properties=presentation_attributes,
+    children=frozenset(['set', 'title', 'animate', 'metadata', 'animateColor', 'desc'])),
+
+    'tspan': SVGElement('tspan',
+    attributes=frozenset(['xml:base', 'onmouseup', 'onmouseout', 'requiredExtensions', 'onfocusout', 'xml:space', 'class', 'systemLanguage', 'onmouseover', 'dx', 'dy', 'externalResourcesRequired', 'lengthAdjust', 'onload', 'style', 'rotate', 'onactivate', 'onmousedown', 'textLength', 'id', 'requiredFeatures', 'xml:lang', 'onmousemove', 'onclick', 'y', 'x', 'onfocusin']),
+    properties=presentation_attributes,
+    children=frozenset(['a', 'set', 'title', 'tspan', 'animateColor', 'tref', 'altGlyph', 'animate', 'metadata', 'desc'])),
+
+    'use': SVGElement('use',
+    attributes=frozenset(['requiredExtensions', 'onfocusout', 'xml:space', 'xlink:href', 'height', 'id', 'onload', 'style', 'onmousedown', 'transform', 'width', 'onmousemove', 'onclick', 'xlink:arcrole', 'onfocusin', 'xml:base', 'onmouseup', 'onmouseout', 'xlink:title', 'systemLanguage', 'onmouseover', 'xlink:type', 'externalResourcesRequired', 'class', 'xlink:actuate', 'xlink:show', 'onactivate', 'xlink:role', 'requiredFeatures', 'xml:lang', 'y', 'x']),
+    properties=presentation_attributes,
+    children=frozenset(['animateMotion', 'set', 'title', 'animateColor', 'animateTransform', 'animate', 'desc', 'metadata'])),
+
+    'view': SVGElement('view',
+    attributes=frozenset(['xml:base', 'viewTarget', 'xml:space', 'viewBox', 'xml:lang', 'preserveAspectRatio', 'externalResourcesRequired', 'zoomAndPan', 'id']),
+    properties=empty_list,
+    children=frozenset(['desc', 'metadata', 'title'])),
+
+    'vkern': SVGElement('vkern',
+    attributes=frozenset(['xml:base', 'g2', 'g1', 'xml:space', 'u1', 'u2', 'xml:lang', 'id', 'k']),
+    properties=empty_list,
+    children=empty_list),
+
+}
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/pattern.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/pattern.py
new file mode 100644
index 00000000..233722d2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/pattern.py
@@ -0,0 +1,30 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: pattern module
+# Created: 27.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+import re
+
+#coordinate ::= number ("em" | "ex" | "px" | "in" | "cm" | "mm" | "pt" | "pc" | "%")?
+coordinate = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)(cm|em|ex|in|mm|pc|pt|px|%)?$")
+
+#length ::= number ("em" | "ex" | "px" | "in" | "cm" | "mm" | "pt" | "pc" | "%")?
+length = coordinate
+
+#angle ::= number (~"deg" | ~"grad" | ~"rad")?
+angle = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)(deg|rad|grad)?$")
+
+# numbers without units
+number = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)$")
+
+# number as percentage value '###%'
+percentage = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)%$")
+
+#frequency ::= number (~"Hz" | ~"kHz")
+frequency = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)(Hz|kHz)?$")
+
+#time ::= number (~"s" | ~"ms")
+time = re.compile(r"(^[-+]?[0-9]*\.?[0-9]+([eE][-+]?[0-9]+)?)(s|ms)?$")
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/svgparser.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/svgparser.py
new file mode 100644
index 00000000..1e6a82fa
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/svgparser.py
@@ -0,0 +1,132 @@
+#!/usr/bin/env python
+# coding:utf-8
+# Authors:  mozman <me@mozman.at>, Florian Festi
+# Purpose: svgparser using re module
+# Created: 16.10.2010
+# Copyright (c) 2010, Manfred Moitzi & 2020, Florian Festi
+# License: MIT License
+
+__all__ = ["is_valid_transferlist", "is_valid_pathdata", "is_valid_animation_timing"]
+
+import re
+
+event_names = [
+    "focusin", "focusout", "activate", "click", "mousedown", "mouseup", "mouseover",
+    "mousemove", "mouseout", "DOMSubtreeModified", "DOMNodeInserted", "DOMNodeRemoved",
+    "DOMNodeRemovedFromDocument", "DOMNodeInsertedtoDocument", "DOMAttrModified",
+    "DOMCharacterDataModified", "SVGLoad", "SVGUnload", "SVGAbort", "SVGError",
+    "SVGResize", "SVGScroll", "SVGZoom", "beginEvent", "endEvent", "repeatEvent",
+]
+
+c = r"\s*[, ]\s*"
+s = r"\s*[; ]\s*"
+integer_constant = r"\d+"
+exponent = r"([eE][+-]?\d+)"
+nonnegative_number = fr"(\d+\.?\d*|\.\d+){exponent}?"
+number = r"[+-]?" + nonnegative_number
+flag = r"[01]"
+
+comma_delimited_coordinates = fr"\s*{number}({c}{number})*\s*"
+two_comma_delimited_numbers = fr"\s*{number}({c}{number}\s*)" "{1}"
+four_comma_delimited_numbers = fr"\s*{number}\s*({c}{number}\s*)" "{3}"
+six_comma_delimited_numbers = fr"\s*{number}\s*({c}{number}\s*)" "{5}"
+comma_delimited_coordinate_pairs = fr"{two_comma_delimited_numbers}({c}{two_comma_delimited_numbers})*"
+
+
+def is_valid(regex):
+    reg = re.compile(regex)
+
+    def f(term):
+        return bool(reg.fullmatch(term))
+
+    return f
+
+
+def build_transferlist_parser():
+    matrix = fr"matrix\s*\(\s*{six_comma_delimited_numbers}\s*\)"
+    translate = fr"translate\s*\(\s*{number}({c}{number})?\s*\)"
+    scale = fr"scale\s*\(\s*{number}({c}{number})?\s*\)"
+    rotate = fr"rotate\s*\(\s*{number}({c}{number}{c}{number})?\s*\)"
+    skewX = fr"skewX\s*\(\s*{number}\s*\)"
+    skewY = fr"skewY\s*\(\s*{number}\s*\)"
+
+    tl_re = "|".join((fr"(\s*{cmd}\s*)" for cmd in (
+        matrix, translate, scale, rotate, skewX, skewY)))
+    return fr"({tl_re})({c}({tl_re}))*"
+
+
+is_valid_transferlist = is_valid(build_transferlist_parser())
+
+
+def build_pathdata_parser():
+    moveto = fr"[mM]\s*{comma_delimited_coordinate_pairs}"
+    lineto = fr"[lL]\s*{comma_delimited_coordinate_pairs}"
+    horizontal_lineto = fr"[hH]\s*{comma_delimited_coordinates}"
+    vertical_lineto = fr"[vV]\s*{comma_delimited_coordinates}"
+    curveto = fr"[cC]\s*({six_comma_delimited_numbers})({c}{six_comma_delimited_numbers})*"
+    smooth_curveto = fr"[sS]{four_comma_delimited_numbers}({c}{four_comma_delimited_numbers})*"
+    quadratic_bezier_curveto = fr"[qQ]{four_comma_delimited_numbers}({c}{four_comma_delimited_numbers})*"
+    smooth_quadratic_bezier_curveto = fr"[tT]\s*{comma_delimited_coordinate_pairs}"
+
+    elliptical_arc_argument = fr"{c}".join((
+        fr"{nonnegative_number}",
+        fr"{nonnegative_number}",
+        fr"{number}",
+        fr"{flag}",
+        fr"{flag}",
+        fr"{number}",
+        fr"{number}",))
+    elliptical_arc_argument = r"\s*" + elliptical_arc_argument + r"\s*"
+    elliptical_arc = fr"[aA]({elliptical_arc_argument})({c}{elliptical_arc_argument})*"
+
+    drawto_command = "|".join((fr"(\s*{cmd}\s*)" for cmd in (
+        moveto, lineto, horizontal_lineto, vertical_lineto, "[zZ]",
+        curveto, smooth_curveto, quadratic_bezier_curveto,
+        smooth_quadratic_bezier_curveto, elliptical_arc)))
+
+    return f"{moveto}({drawto_command})*"
+
+
+is_valid_pathdata = is_valid(build_pathdata_parser())
+
+digit2 = r"\d{2}"
+digit4 = r"\d{4}"
+seconds = fr"\d+(\.\d+)?"
+seconds2 = fr"{digit2}(\.\d+)?"
+metric = "(h|min|s|ms)"
+
+
+def clock_val_re():
+    timecount_val = fr"{seconds}\s*({metric})?"
+    clock_val = fr"{digit2}:({digit2}:)?{seconds2}"
+    return fr"({timecount_val}|{clock_val})"
+
+
+def wall_clock_val_re():
+    hhmmss = fr"{digit2}:{digit2}(:{seconds2})?"
+    walltime = fr"{hhmmss}(Z|[+-]?{digit2}:{digit2})?"
+    date = fr"{digit4}-{digit2}-{digit2}"
+    datetime = fr"{date}(T{walltime})?"
+    return "(" + "|".join((walltime, datetime)) + ")"
+
+
+def build_animation_timing_parser():
+    clock_val = clock_val_re()
+    wallclock_val = wall_clock_val_re()
+
+    event_ref = "(" + "|".join(event_names) + ")"
+    id_value = "#?[-_a-zA-Z0-9]+"
+
+    wallclock_sync_value = fr"wallclock\(\s*{wallclock_val}\s*\)"
+    accesskey_value = fr"accessKey\(\s*[a-zA-Z]\s*\)\s*([+-]?{clock_val})?"
+    repeat_value = fr"({id_value}\.)?repeat\s*\(\s*\d+\s*\)\s*([+-?]{clock_val})?"
+    event_value = fr"({id_value}\.)?{event_ref}([+-]?{clock_val})?"
+    offset_value = fr"[-+]?{clock_val}"
+    syncbase_value = fr"{id_value}\.(begin|end)({offset_value})?"
+    begin_value = "(" + "|".join((f"({reg})" for reg in (
+        offset_value, syncbase_value, event_value, repeat_value,
+        accesskey_value, wallclock_sync_value, "indefinite"))) + ")"
+    return fr"{begin_value}({s}{begin_value})*"
+
+
+is_valid_animation_timing = is_valid(build_animation_timing_parser())
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/tiny12.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/tiny12.py
new file mode 100644
index 00000000..53108f90
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/tiny12.py
@@ -0,0 +1,966 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: tiny12 data
+# Created: 15.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.data.types import SVGAttribute, SVGMultiAttribute
+from svgwrite.data.types import SVGElement
+from svgwrite.data.typechecker import Tiny12TypeChecker as TypeChecker
+
+empty_list = []
+
+attributes = {
+'about': SVGAttribute('about', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'accent-height': SVGAttribute('accent-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'accumulate': SVGAttribute('accumulate', anim=False,
+    types=empty_list,
+    const=frozenset(['none', 'sum'])),
+'additive': SVGAttribute('additive', anim=False,
+    types=empty_list,
+    const=frozenset(['replace', 'sum'])),
+'alphabetic': SVGAttribute('alphabetic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'arabic-form': SVGAttribute('arabic-form', anim=False,
+    types=empty_list,
+    const=frozenset(['terminal', 'initial', 'isolated', 'medial'])),
+'ascent': SVGAttribute('ascent', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'attributeName': SVGAttribute('attributeName', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'attributeType': SVGAttribute('attributeType', anim=False,
+    types=empty_list,
+    const=frozenset(['XML', 'CSS', 'auto'])),
+'audio-level': SVGAttribute('audio-level', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'bandwidth': SVGAttribute('bandwidth', anim=False,
+    types=frozenset(['number']),
+    const=frozenset(['auto'])),
+'baseProfile': SVGAttribute('baseProfile', anim=False,
+    types=empty_list,
+    const=frozenset(['none', 'tiny', 'basic', 'full'])),
+'bbox': SVGAttribute('bbox', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'begin': SVGAttribute('begin', anim=False,
+    types=frozenset(['timing-value-list']),
+    const=frozenset(['indefinite'])),
+'buffered-rendering': SVGAttribute('buffered-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'dynamic', 'static', 'inherit'])),
+'by': SVGAttribute('by', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'calcMode': SVGAttribute('calcMode', anim=False,
+    types=empty_list,
+    const=frozenset(['discrete', 'linear', 'paced', 'spline'])),
+'cap-height': SVGAttribute('cap-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'class': SVGAttribute('class', anim=True,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'color': SVGAttribute('color', anim=True,
+    types=frozenset(['color']),
+    const=frozenset(['inherit'])),
+'color-rendering': SVGAttribute('color-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'optimizeQuality', 'inherit'])),
+'content': SVGAttribute('content', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'contentScriptType': SVGAttribute('contentScriptType', anim=False,
+    types=frozenset(['content-type']),
+    const=empty_list),
+'cx': SVGAttribute('cx', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'cy': SVGAttribute('cy', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'d': SVGAttribute('d', anim=True,
+    types=frozenset(['path-data']),
+    const=empty_list),
+'datatype': SVGAttribute('datatype', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'defaultAction': SVGAttribute('defaultAction', anim=False,
+    types=empty_list,
+    const=frozenset(['perform', 'cancel'])),
+'descent': SVGAttribute('descent', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'direction': SVGAttribute('direction', anim=False,
+    types=empty_list,
+    const=frozenset(['ltr', 'rtl', 'inherit'])),
+'display': SVGAttribute('display', anim=True,
+    types=empty_list,
+    const=frozenset(['inline', 'block', 'list-item', 'run-in', 'compact',
+                     'marker', 'table', 'inline-table', 'table-row-group',
+                     'table-header-group', 'table-footer-group', 'table-row',
+                     'table-column-group', 'table-column', 'table-cell',
+                     'table-caption', 'none', 'inherit'])),
+'display-align': SVGAttribute('display-align', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'before', 'center', 'after', 'inherit'])),
+'dur': SVGAttribute('dur', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['media', 'indefinite'])),
+'editable': SVGAttribute('editable', anim=True,
+    types=empty_list,
+    const=frozenset(['none', 'simple'])),
+'end': SVGAttribute('end', anim=False,
+    types=frozenset(['timing-value-list']),
+    const=frozenset(['indefinite'])),
+'ev:event': SVGAttribute('ev:event', anim=False,
+    types=frozenset(['XML-Name']),
+    const=empty_list),
+'event': SVGAttribute('event', anim=False,
+    types=frozenset(['XML-Name']),
+    const=empty_list),
+'externalResourcesRequired': SVGAttribute('externalResourcesRequired', anim=False,
+    types=frozenset(['boolean']),
+    const=empty_list),
+'fill': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'fill', anim=True,
+        types=frozenset(['paint']),
+        const=frozenset(['inherit'])),
+    'set animateMotion animate animateColor animateTransform': SVGAttribute(
+        'fill', anim=False,
+        types=empty_list,
+        const=frozenset(['freeze', 'remove']))
+    }),
+'fill-opacity': SVGAttribute('fill-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'fill-rule': SVGAttribute('fill-rule', anim=True,
+    types=empty_list,
+    const=frozenset(['nonzero', 'evenodd', 'inherit'])),
+'focusHighlight': SVGAttribute('focusHighlight', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'none'])),
+'focusable': SVGAttribute('focusable', anim=True,
+    types=frozenset(['boolean']),
+    const=frozenset(['auto'])),
+'font-family': SVGAttribute('font-family', anim=True,
+    types=frozenset(['string']),
+    const=frozenset(['inherit'])),
+'font-size': SVGAttribute('font-size', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'font-stretch': SVGAttribute('font-stretch', anim=False,
+    types=empty_list,
+    const=frozenset(['condensed', 'normal', 'ultra-condensed', 'expanded',
+                     'narrower', 'inherit', 'semi-condensed', 'extra-condensed',
+                     'ultra-expanded', 'wider', 'semi-expanded', 'extra-expanded'])),
+'font-style': SVGAttribute('font-style', anim=True,
+    types=empty_list,
+    const=frozenset(['normal', 'italic', 'oblique', 'inherit'])),
+'font-variant': SVGAttribute('font-variant', anim=True,
+    types=empty_list,
+    const=frozenset(['normal', 'small-caps', 'inherit'])),
+'font-weight': SVGAttribute('font-weight', anim=True,
+    types=empty_list,
+    const=frozenset(['normal', 'bold', 'bolder', 'lighter', '100', '200',
+                     '300', '400', '500', '600', '700', '800', '900',
+                     'inherit'])),
+'from': SVGAttribute('from', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'g1': SVGAttribute('g1', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'g2': SVGAttribute('g2', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'glyph-name': SVGAttribute('glyph-name', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'gradientUnits': SVGAttribute('gradientUnits', anim=True,
+    types=empty_list,
+    const=frozenset(['userSpaceOnUse', 'objectBoundingBox'])),
+'handler': SVGAttribute('handler', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'hanging': SVGAttribute('hanging', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'height': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'height', anim=True,
+        types=frozenset(['length']),
+        const=empty_list),
+    'textArea': SVGAttribute(
+        'height', anim=True,
+        types=frozenset(['length']),
+        const=frozenset(['auto'])),
+    }),
+'horiz-adv-x': SVGAttribute('horiz-adv-x', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'horiz-origin-x': SVGAttribute('horiz-origin-x', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'id': SVGAttribute('id', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'ideographic': SVGAttribute('ideographic', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'image-rendering': SVGAttribute('image-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'optimizeQuality', 'inherit'])),
+'initialVisibility': SVGAttribute('initialVisibility', anim=False,
+    types=empty_list,
+    const=frozenset(['whenStarted', 'always'])),
+'k': SVGAttribute('k', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'keyPoints': SVGAttribute('keyPoints', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'keySplines': SVGAttribute('keySplines', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'keyTimes': SVGAttribute('keyTimes', anim=False,
+    types=frozenset(['semicolon-list']),
+    const=empty_list),
+'lang': SVGAttribute('lang', anim=False,
+    types=frozenset(['list-of-name']),
+    const=empty_list),
+'line-increment': SVGAttribute('line-increment', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['auto', 'inherit'])),
+'mathematical': SVGAttribute('mathematical', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'max': SVGAttribute('max', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['media'])),
+'mediaCharacterEncoding': SVGAttribute('mediaCharacterEncoding', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'mediaContentEncodings': SVGAttribute('mediaContentEncodings', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'mediaSize': SVGAttribute('mediaSize', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'mediaTime': SVGAttribute('mediaTime', anim=False,
+    types=frozenset(['time']),
+    const=empty_list),
+'min': SVGAttribute('min', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['media'])),
+'nav-down': SVGAttribute('nav-down', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-down-left': SVGAttribute('nav-down-left', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-down-right': SVGAttribute('nav-down-right', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-left': SVGAttribute('nav-left', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-next': SVGAttribute('nav-next', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-prev': SVGAttribute('nav-prev', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-right': SVGAttribute('nav-right', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-up': SVGAttribute('nav-up', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-up-left': SVGAttribute('nav-up-left', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'nav-up-right': SVGAttribute('nav-up-right', anim=True,
+    types=frozenset(['focus']),
+    const=empty_list),
+'observer': SVGAttribute('observer', anim=False,
+    types=frozenset(['IDREF']),
+    const=empty_list),
+'offset': SVGAttribute('offset', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'opacity': SVGAttribute('opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'origin': SVGAttribute('origin', anim=False,
+    types=empty_list,
+    const=frozenset(['default'])),
+'overlay': SVGAttribute('overlay', anim=False,
+    types=empty_list,
+    const=frozenset(['none', 'top'])),
+'overline-position': SVGAttribute('overline-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'overline-thickness': SVGAttribute('overline-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'panose-1': SVGAttribute('panose-1', anim=False,
+    types=frozenset(['list-of-integer']),
+    const=empty_list),
+'path': SVGAttribute('path', anim=False,
+    types=frozenset(['path-data']),
+    const=empty_list),
+'pathLength': SVGAttribute('pathLength', anim=True,
+    types=frozenset(['number']),
+    const=empty_list),
+'phase': SVGAttribute('phase', anim=False,
+    types=empty_list,
+    const=frozenset(['default', 'capture'])),
+'playbackOrder': SVGAttribute('playbackOrder', anim=False,
+    types=empty_list,
+    const=frozenset(['all', 'forwardOnly'])),
+'pointer-events': SVGAttribute('pointer-events', anim=True,
+    types=empty_list,
+    const=frozenset(['visiblePainted', 'visibleFill', 'visibleStroke', 'visible',
+                     'painted', 'fill', 'stroke', 'all', 'none', 'inherit'])),
+'points': SVGAttribute('points', anim=True,
+    types=frozenset(['list-of-number']),
+    const=empty_list),
+'preserveAspectRatio': SVGAttribute('preserveAspectRatio', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'propagate': SVGAttribute('propagate', anim=False,
+    types=empty_list,
+    const=frozenset(['continue', 'stop'])),
+'property': SVGAttribute('property', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'r': SVGAttribute('r', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'rel': SVGAttribute('rel', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'repeatCount': SVGAttribute('repeatCount', anim=False,
+    types=frozenset(['number']),
+    const=frozenset(['indefinite'])),
+'repeatDur': SVGAttribute('repeatDur', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['indefinite'])),
+'requiredExtensions': SVGAttribute('requiredExtensions', anim=False,
+    types=frozenset(['list-of-IRI']),
+    const=empty_list),
+'requiredFeatures': SVGAttribute('requiredFeatures', anim=False,
+    types=frozenset(['list-of-IRI']),
+    const=empty_list),
+'requiredFonts': SVGAttribute('requiredFonts', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'requiredFormats': SVGAttribute('requiredFormats', anim=False,
+    types=frozenset(['anything']),
+    const=empty_list),
+'resource': SVGAttribute('resource', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'restart': SVGAttribute('restart', anim=False,
+    types=empty_list,
+    const=frozenset(['always', 'never', 'whenNotActive'])),
+'rev': SVGAttribute('rev', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'role': SVGAttribute('role', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'rotate': SVGAttribute('rotate', anim=True,
+    types=frozenset(['list-of-number']),
+    const=empty_list),
+'rx': SVGAttribute('rx', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'ry': SVGAttribute('ry', anim=True,
+    types=frozenset(['length']),
+    const=empty_list),
+'shape-rendering': SVGAttribute('shape-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'crispEdges', 'geometricPrecision', 'inherit'])),
+'slope': SVGAttribute('slope', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'snapshotTime': SVGAttribute('snapshotTime', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['none'])),
+'solid-color': SVGAttribute('solid-color', anim=True,
+    types=frozenset(['color']),
+    const=frozenset(['currentColor', 'inherit'])),
+'solid-opacity': SVGAttribute('solid-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'stemh': SVGAttribute('stemh', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'stemv': SVGAttribute('stemv', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'stop-color': SVGAttribute('stop-color', anim=True,
+    types=frozenset(['color']),
+    const=frozenset(['inherit'])),
+'stop-opacity': SVGAttribute('stop-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'strikethrough-position': SVGAttribute('strikethrough-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'strikethrough-thickness': SVGAttribute('strikethrough-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'stroke': SVGAttribute('stroke', anim=True,
+    types=frozenset(['paint']),
+    const=frozenset(['inherit'])),
+'stroke-dasharray': SVGAttribute('stroke-dasharray', anim=True,
+    types=frozenset(['list-of-length']),
+    const=frozenset(['none', 'inherit'])),
+'stroke-dashoffset': SVGAttribute('stroke-dashoffset', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'stroke-linecap': SVGAttribute('stroke-linecap', anim=True,
+    types=empty_list,
+    const=frozenset(['butt', 'round', 'square', 'inherit'])),
+'stroke-linejoin': SVGAttribute('stroke-linejoin', anim=True,
+    types=empty_list,
+    const=frozenset(['miter', 'round', 'bevel', 'inherit'])),
+'stroke-miterlimit': SVGAttribute('stroke-miterlimit', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'stroke-opacity': SVGAttribute('stroke-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'stroke-width': SVGAttribute('stroke-width', anim=True,
+    types=frozenset(['length']),
+    const=frozenset(['inherit'])),
+'syncBehavior': SVGAttribute('syncBehavior', anim=False,
+    types=empty_list,
+    const=frozenset(['canSlip', 'locked', 'independent', 'default'])),
+'syncBehaviorDefault': SVGAttribute('syncBehaviorDefault', anim=False,
+    types=empty_list,
+    const=frozenset(['canSlip', 'locked', 'independent', 'inherit'])),
+'syncMaster': SVGAttribute('syncMaster', anim=False,
+    types=frozenset(['boolean']),
+    const=empty_list),
+'syncTolerance': SVGAttribute('syncTolerance', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['default'])),
+'syncToleranceDefault': SVGAttribute('syncToleranceDefault', anim=False,
+    types=frozenset(['time']),
+    const=frozenset(['inherit'])),
+'systemLanguage': SVGAttribute('systemLanguage', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'target': SVGMultiAttribute({
+    '* a': SVGAttribute(
+        'target', anim=True,
+        types=frozenset(['XML-Name']),
+        const=frozenset(['_replace', '_self', '_parent', '_top', '_blank'])),
+    'listener': SVGAttribute(
+        'target', anim=False,
+        types=frozenset(['XML-Name']),
+        const=empty_list),
+    }),
+'text-align': SVGAttribute('text-align', anim=True,
+    types=empty_list,
+    const=frozenset(['start', 'center', 'end', 'inherit'])),
+'text-anchor': SVGAttribute('text-anchor', anim=True,
+    types=empty_list,
+    const=frozenset(['start', 'middle', 'end', 'inherit'])),
+'text-rendering': SVGAttribute('text-rendering', anim=True,
+    types=empty_list,
+    const=frozenset(['auto', 'optimizeSpeed', 'optimizeLegibility', 'geometricPrecision', 'inherit'])),
+'timelineBegin': SVGAttribute('timelineBegin', anim=False,
+    types=empty_list,
+    const=frozenset(['onLoad', 'onStart'])),
+'to': SVGAttribute('to', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'transform': SVGAttribute('transform', anim=True,
+    types=frozenset(['transform-list']),
+    const=frozenset(['none'])),
+'transformBehavior': SVGAttribute('transformBehavior', anim=False,
+    types=empty_list,
+    const=frozenset(['geometric', 'pinned', 'pinned90', 'pinned180', 'pinned270'])),
+'type': SVGMultiAttribute({
+    '* audio image video': SVGAttribute(
+        'type', anim=True,
+        types=frozenset(['content-type']),
+        const=empty_list),
+    'handler script': SVGAttribute(
+        'type', anim=False,
+        types=frozenset(['content-type']),
+        const=empty_list),
+    'animateTransform': SVGAttribute(
+        'type', anim=False,
+        types=empty_list,
+        const=frozenset(['translate', 'scale', 'rotate', 'skewX', 'skewY'])),
+    }),
+'typeof': SVGAttribute('typeof', anim=True,
+    types=frozenset(['string']),
+    const=empty_list),
+'u1': SVGAttribute('u1', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'u2': SVGAttribute('u2', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'underline-position': SVGAttribute('underline-position', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'underline-thickness': SVGAttribute('underline-thickness', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'unicode': SVGAttribute('unicode', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'unicode-bidi': SVGAttribute('unicode-bidi', anim=False,
+    types=empty_list,
+    const=frozenset(['normal', 'embed', 'bidi-override', 'inherit'])),
+'unicode-range': SVGAttribute('unicode-range', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'units-per-em': SVGAttribute('units-per-em', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'values': SVGAttribute('values', anim=False,
+    types=frozenset(['list-of-number']),
+    const=empty_list),
+'vector-effect': SVGAttribute('vector-effect', anim=True,
+    types=empty_list,
+    const=frozenset(['none', 'non-scaling-stroke', 'inherit'])),
+'version': SVGAttribute('version', anim=False,
+    types=empty_list,
+    const=frozenset(['1.1', '1.2'])),
+'viewBox': SVGAttribute('viewBox', anim=True,
+    types=frozenset(['four-numbers']),
+    const=empty_list),
+'viewport-fill': SVGAttribute('viewport-fill', anim=True,
+    types=frozenset(['color']),
+    const=frozenset(['none', 'inherit'])),
+'viewport-fill-opacity': SVGAttribute('viewport-fill-opacity', anim=True,
+    types=frozenset(['number']),
+    const=frozenset(['inherit'])),
+'visibility': SVGAttribute('visibility', anim=True,
+    types=empty_list,
+    const=frozenset(['visible', 'hidden', 'inherit'])),
+'width': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'width', anim=True,
+        types=frozenset(['length']),
+        const=empty_list),
+    'textArea': SVGAttribute(
+        'width', anim=True,
+        types=frozenset(['length']),
+        const=frozenset(['auto'])),
+    }),
+'widths': SVGAttribute('widths', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'x': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'x', anim=True,
+        types=frozenset(['coordinate']),
+        const=empty_list),
+    'text': SVGAttribute(
+        'x', anim=True,
+        types=frozenset(['list-of-coordinate']),
+        const=empty_list),
+    }),
+'x-height': SVGAttribute('x-height', anim=False,
+    types=frozenset(['number']),
+    const=empty_list),
+'x1': SVGAttribute('x1', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'x2': SVGAttribute('x2', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'xlink:actuate': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'xlink:actuate', anim=False,
+        types=empty_list,
+        const=frozenset(['onLoad'])),
+    'a': SVGAttribute(
+        'xlink:actuate', anim=False,
+        types=empty_list,
+        const=frozenset(['onRequest'])),
+    }),
+'xlink:arcrole': SVGAttribute('xlink:arcrole', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:href': SVGAttribute('xlink:href', anim=True,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:role': SVGAttribute('xlink:role', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xlink:show': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'xlink:show', anim=False,
+        types=empty_list,
+        const=frozenset(['other'])),
+    'animation audio foreignObject image use video': SVGAttribute(
+        'xlink:show', anim=False,
+        types=empty_list,
+        const=frozenset(['embed'])),
+    'a': SVGAttribute(
+        'xlink:show', anim=False,
+        types=empty_list,
+        const=frozenset(['new', 'replace'])),
+    }),
+'xlink:title': SVGAttribute('xlink:title', anim=False,
+    types=frozenset(['string']),
+    const=empty_list),
+'xlink:type': SVGAttribute('xlink:type', anim=True,
+    types=empty_list,
+    const=frozenset(['simple'])),
+'xmlns': SVGAttribute('xmlns', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xmlns:xlink': SVGAttribute('xmlns:xlink', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xmlns:ev': SVGAttribute('xmlns:ev', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xml:base': SVGAttribute('xml:base', anim=False,
+    types=frozenset(['IRI']),
+    const=empty_list),
+'xml:id': SVGAttribute('xml:id', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'xml:lang': SVGAttribute('xml:lang', anim=False,
+    types=frozenset(['name']),
+    const=empty_list),
+'xml:space': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'xml:space', anim=False,
+        types=empty_list,
+        const=frozenset(['preserve', 'default'])),
+    'handler script': SVGAttribute(
+        'xml:space', anim=False,
+        types=empty_list,
+        const=frozenset(['preserve'])),
+    }),
+'y': SVGMultiAttribute({
+    '*': SVGAttribute(
+        'y', anim=True,
+        types=frozenset(['coordinate']),
+        const=empty_list),
+    'text': SVGAttribute(
+        'y', anim=True,
+        types=frozenset(['list-of-coordinate']),
+        const=empty_list),
+    }),
+'y1': SVGAttribute('y1', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'y2': SVGAttribute('y2', anim=True,
+    types=frozenset(['coordinate']),
+    const=empty_list),
+'zoomAndPan': SVGAttribute('zoomAndPan', anim=False,
+    types=empty_list,
+    const=frozenset(['disable', 'magnify'])),
+}
+
+attribute_names = ['slope', 'keySplines', 'rx', 'accumulate', 'bandwidth',
+                   'attributeType', 'unicode', 'nav-right', 'arabic-form',
+                   'y2', 'horiz-origin-x', 'underline-position', 'zoomAndPan',
+                   'cap-height', 'defaultAction', 'to', 'syncBehavior',
+                   'alphabetic', 'g2', 'g1', 'panose-1', 'strikethrough-thickness',
+                   'attributeName', 'bbox', 'nav-up-left', 'nav-left',
+                   'restart', 'target', 'xlink:actuate', 'rotate', 'resource',
+                   'd', 'syncToleranceDefault', 'initialVisibility',
+                   'transformBehavior', 'nav-up-right', 'keyTimes',
+                   'x', 'requiredFormats', 'nav-next', 'glyph-name',
+                   'xml:lang', 'mathematical', 'observer', 'repeatDur',
+                   'hanging', 'y1', 'xml:base', 'ascent', 'event',
+                   'strikethrough-position', 'overlay', 'rev', 'ry',
+                   'overline-thickness', 'content', 'version', 'rel',
+                   'focusable', 'requiredFonts', 'nav-down-right', 'xml:id',
+                   'offset', 'additive', 'underline-thickness', 'font-family',
+                   'by', 'mediaTime', 'timelineBegin', 'about', 'horiz-adv-x',
+                   'widths', 'k', 'requiredFeatures', 'preserveAspectRatio',
+                   'contentScriptType', 'origin', 'xml:space', 'xlink:href',
+                   'height', 'baseProfile', 'cy', 'cx', 'path', 'xlink:role',
+                   'from', 'u1', 'transform', 'units-per-em', 'u2', 'width',
+                   'handler', 'font-variant', 'x-height', 'dur', 'xlink:arcrole',
+                   'type', 'focusHighlight', 'mediaCharacterEncoding',
+                   'xlink:title', 'editable', 'stemv', 'systemLanguage',
+                   'x2', 'x1', 'ideographic', 'xlink:show', 'overline-position',
+                   'syncTolerance', 'gradientUnits', 'r', 'values', 'typeof',
+                   'mediaContentEncodings', 'property', 'requiredExtensions',
+                   'repeatCount', 'ev:event', 'nav-down', 'mediaSize', 'pathLength',
+                   'syncMaster', 'font-style', 'fill', 'end', 'descent',
+                   'calcMode', 'min', 'stemh', 'id', 'unicode-range',
+                   'nav-up', 'font-stretch', 'role', 'font-weight', 'begin',
+                   'xlink:type', 'syncBehaviorDefault', 'max', 'snapshotTime',
+                   'playbackOrder', 'keyPoints', 'nav-prev', 'propagate',
+                   'phase', 'externalResourcesRequired', 'nav-down-left',
+                   'class', 'lang', 'datatype', 'viewBox', 'points',
+                   'accent-height', 'y']
+property_names = ['stroke-linejoin', 'font-size', 'text-rendering', 'color-rendering',
+                  'fill-opacity', 'color', 'shape-rendering', 'solid-color', 'stroke',
+                  'stroke-linecap', 'vector-effect', 'stroke-width', 'font-style',
+                  'fill', 'solid-opacity', 'fill-rule', 'viewport-fill-opacity',
+                  'display-align', 'buffered-rendering', 'stroke-miterlimit',
+                  'font-variant', 'stop-opacity', 'font-weight', 'opacity', 'direction',
+                  'audio-level', 'visibility', 'unicode-bidi', 'line-increment',
+                  'image-rendering', 'font-family', 'viewport-fill', 'text-align',
+                  'stroke-opacity', 'stroke-dashoffset', 'text-anchor', 'stop-color',
+                  'pointer-events', 'stroke-dasharray', 'display']
+media_group_names = ['audio-level', 'buffered-rendering', 'display', 'image-rendering',
+                     'pointer-events', 'shape-rendering', 'text-rendering',
+                     'viewport-fill', 'viewport-fill-opacity', 'visibility']
+
+elements = {
+    'a': SVGElement('a',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'target', 'transform', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['likeparent', 'defs', 'text', 'g', 'textArea', 'svg']),
+
+    'animate': SVGElement('animate',
+    attributes=['about', 'accumulate', 'additive', 'attributeName', 'attributeType', 'begin', 'by', 'calcMode', 'class', 'content', 'datatype', 'dur', 'end', 'fill', 'from', 'id', 'keySplines', 'keyTimes', 'max', 'min', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'systemLanguage', 'to', 'typeof', 'values', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'switch', 'title']),
+
+    'animateColor': SVGElement('animateColor',
+    attributes=['about', 'accumulate', 'additive', 'attributeName', 'attributeType', 'begin', 'by', 'calcMode', 'class', 'content', 'datatype', 'dur', 'end', 'fill', 'from', 'id', 'keySplines', 'keyTimes', 'max', 'min', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'systemLanguage', 'to', 'typeof', 'values', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'switch', 'title']),
+
+    'animateMotion': SVGElement('animateMotion',
+    attributes=['about', 'accumulate', 'additive', 'begin', 'by', 'calcMode', 'class', 'content', 'datatype', 'dur', 'end', 'fill', 'from', 'id', 'keyPoints', 'keySplines', 'keyTimes', 'max', 'min', 'origin', 'path', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'rotate', 'systemLanguage', 'to', 'typeof', 'values', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'mpath', 'switch', 'title']),
+
+    'animateTransform': SVGElement('animateTransform',
+    attributes=['about', 'accumulate', 'additive', 'attributeName', 'attributeType', 'begin', 'by', 'calcMode', 'class', 'content', 'datatype', 'dur', 'end', 'fill', 'from', 'id', 'keySplines', 'keyTimes', 'max', 'min', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'systemLanguage', 'to', 'type', 'typeof', 'values', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'switch', 'title']),
+
+    'animation': SVGElement('animation',
+    attributes=['about', 'begin', 'class', 'content', 'datatype', 'dur', 'end', 'externalResourcesRequired', 'fill', 'focusHighlight', 'focusable', 'height', 'id', 'initialVisibility', 'max', 'min', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'preserveAspectRatio', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'syncBehavior', 'syncMaster', 'syncTolerance', 'systemLanguage', 'transform', 'typeof', 'width', 'x', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=media_group_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'audio': SVGElement('audio',
+    attributes=['about', 'begin', 'class', 'content', 'datatype', 'dur', 'end', 'externalResourcesRequired', 'fill', 'id', 'max', 'min', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'syncBehavior', 'syncMaster', 'syncTolerance', 'systemLanguage', 'type', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=media_group_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'circle': SVGElement('circle',
+    attributes=['about', 'class', 'content', 'cx', 'cy', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'r', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'defs': SVGElement('defs',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['a', 'animate', 'animateColor', 'animateMotion', 'animateTransform', 'animation', 'audio', 'circle', 'defs', 'desc', 'discard', 'ellipse', 'font', 'font-face', 'foreignObject', 'g', 'handler', 'image', 'line', 'linearGradient', 'listener', 'metadata', 'path', 'polygon', 'polyline', 'prefetch', 'radialGradient', 'rect', 'script', 'set', 'solidColor', 'switch', 'text', 'textArea', 'title', 'use', 'video']),
+
+    'desc': SVGElement('desc',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=media_group_names,
+    children=empty_list),
+
+    'discard': SVGElement('discard',
+    attributes=['about', 'begin', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'switch', 'title']),
+
+    'ellipse': SVGElement('ellipse',
+    attributes=['about', 'class', 'content', 'cx', 'cy', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'rx', 'ry', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'font': SVGElement('font',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'horiz-adv-x', 'horiz-origin-x', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'font-face', 'glyph', 'hkern', 'metadata', 'missing-glyph', 'switch', 'title']),
+
+    'font-face': SVGElement('font-face',
+    attributes=['about', 'accent-height', 'alphabetic', 'ascent', 'bbox', 'cap-height', 'class', 'content', 'datatype', 'descent', 'externalResourcesRequired', 'font-family', 'font-stretch', 'font-style', 'font-variant', 'font-weight', 'hanging', 'id', 'ideographic', 'mathematical', 'overline-position', 'overline-thickness', 'panose-1', 'property', 'rel', 'resource', 'rev', 'role', 'slope', 'stemh', 'stemv', 'strikethrough-position', 'strikethrough-thickness', 'typeof', 'underline-position', 'underline-thickness', 'unicode-range', 'units-per-em', 'widths', 'x-height', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'font-face-src', 'metadata', 'switch', 'title']),
+
+    'font-face-src': SVGElement('font-face-src',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'font-face-uri', 'metadata', 'switch', 'title']),
+
+    'font-face-uri': SVGElement('font-face-uri',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'foreignObject': SVGElement('foreignObject',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'height', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'width', 'x', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=property_names,
+    children=['desc', 'metadata', 'svg', 'switch', 'title']),
+
+    'g': SVGElement('g',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['a', 'animate', 'animateColor', 'animateMotion', 'animateTransform', 'animation', 'audio', 'circle', 'defs', 'desc', 'discard', 'ellipse', 'font', 'font-face', 'foreignObject', 'g', 'handler', 'image', 'line', 'linearGradient', 'listener', 'metadata', 'path', 'polygon', 'polyline', 'prefetch', 'radialGradient', 'rect', 'script', 'set', 'solidColor', 'switch', 'text', 'textArea', 'title', 'use', 'video']),
+
+    'glyph': SVGElement('glyph',
+    attributes=['about', 'arabic-form', 'class', 'content', 'd', 'datatype', 'glyph-name', 'horiz-adv-x', 'id', 'lang', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'unicode', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'handler': SVGElement('handler',
+    attributes=['about', 'class', 'content', 'datatype', 'ev:event', 'externalResourcesRequired', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'type', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'hkern': SVGElement('hkern',
+    attributes=['about', 'class', 'content', 'datatype', 'g1', 'g2', 'id', 'k', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'u1', 'u2', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'image': SVGElement('image',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'height', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'opacity', 'preserveAspectRatio', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'type', 'typeof', 'width', 'x', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=media_group_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'line': SVGElement('line',
+    attributes=['about', 'class', 'content', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'x1', 'x2', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y1', 'y2'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'linearGradient': SVGElement('linearGradient',
+    attributes=['about', 'class', 'content', 'datatype', 'gradientUnits', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'x1', 'x2', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y1', 'y2'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'metadata', 'set', 'stop', 'switch', 'title']),
+
+    'listener': SVGElement('listener',
+    attributes=['about', 'class', 'content', 'datatype', 'defaultAction', 'event', 'handler', 'id', 'observer', 'phase', 'propagate', 'property', 'rel', 'resource', 'rev', 'role', 'target', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=empty_list),
+
+    'metadata': SVGElement('metadata',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=media_group_names,
+    children=empty_list),
+
+    'missing-glyph': SVGElement('missing-glyph',
+    attributes=['about', 'class', 'content', 'd', 'datatype', 'horiz-adv-x', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'mpath': SVGElement('mpath',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'path': SVGElement('path',
+    attributes=['about', 'class', 'content', 'd', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'pathLength', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'polygon': SVGElement('polygon',
+    attributes=['about', 'class', 'content', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'points', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'polyline': SVGElement('polyline',
+    attributes=['about', 'class', 'content', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'points', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'prefetch': SVGElement('prefetch',
+    attributes=['about', 'bandwidth', 'class', 'content', 'datatype', 'id', 'mediaCharacterEncoding', 'mediaContentEncodings', 'mediaSize', 'mediaTime', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'radialGradient': SVGElement('radialGradient',
+    attributes=['about', 'class', 'content', 'cx', 'cy', 'datatype', 'gradientUnits', 'id', 'property', 'r', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'metadata', 'set', 'stop', 'switch', 'title']),
+
+    'rect': SVGElement('rect',
+    attributes=['about', 'class', 'content', 'datatype', 'focusHighlight', 'focusable', 'height', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'rx', 'ry', 'systemLanguage', 'transform', 'typeof', 'width', 'x', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'script': SVGElement('script',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'type', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'metadata', 'switch', 'title']),
+
+    'set': SVGElement('set',
+    attributes=['about', 'attributeName', 'attributeType', 'begin', 'class', 'content', 'datatype', 'dur', 'end', 'fill', 'id', 'max', 'min', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'systemLanguage', 'to', 'typeof', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=['desc', 'handler', 'metadata', 'switch', 'title']),
+
+    'solidColor': SVGElement('solidColor',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'stop': SVGElement('stop',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'offset', 'property', 'rel', 'resource', 'rev', 'role', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'metadata', 'set', 'switch', 'title']),
+
+    'svg': SVGElement('svg',
+    attributes=['about', 'baseProfile', 'class', 'content', 'contentScriptType', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'height', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'playbackOrder', 'preserveAspectRatio', 'property', 'rel', 'resource', 'rev', 'role', 'snapshotTime', 'syncBehaviorDefault', 'syncToleranceDefault', 'timelineBegin', 'transform', 'typeof', 'version', 'viewBox', 'width', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'xmlns', 'xmlns:xlink', 'xmlns:ev', 'zoomAndPan'],
+    properties=property_names,
+    children=['a', 'animate', 'animateColor', 'animateMotion', 'animateTransform', 'animation', 'audio', 'circle', 'defs', 'desc', 'discard', 'ellipse', 'font', 'font-face', 'foreignObject', 'g', 'handler', 'image', 'line', 'linearGradient', 'listener', 'metadata', 'path', 'polygon', 'polyline', 'prefetch', 'radialGradient', 'rect', 'script', 'set', 'solidColor', 'switch', 'text', 'textArea', 'title', 'use', 'video']),
+
+    'switch': SVGElement('switch',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['likeparent', 'set', 'textArea', 'text', 'image', 'missing-glyph', 'font-face', 'video', 'path', 'animate', 'font', 'ellipse', 'glyph', 'use', 'font-face-src', 'polygon', 'script', 'handler', 'circle', 'radialGradient', 'prefetch', 'defs', 'mpath', 'stop', 'animateMotion', 'animateColor', 'discard', 'solidColor', 'hkern', 'line', 'animation', 'rect', 'g', 'svg', 'animateTransform', 'linearGradient', 'font-face-uri', 'foreignObject', 'polyline', 'audio']),
+
+    'tbreak': SVGElement('tbreak',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=empty_list,
+    children=empty_list),
+
+    'text': SVGElement('text',
+    attributes=['about', 'class', 'content', 'datatype', 'editable', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'rotate', 'systemLanguage', 'transform', 'typeof', 'x', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=property_names,
+    children=['a', 'animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title', 'tspan']),
+
+    'textArea': SVGElement('textArea',
+    attributes=['about', 'class', 'content', 'datatype', 'editable', 'focusHighlight', 'focusable', 'height', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'width', 'x', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=property_names,
+    children=['a', 'animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'tbreak', 'title', 'tspan']),
+
+    'title': SVGElement('title',
+    attributes=['about', 'class', 'content', 'datatype', 'id', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=media_group_names,
+    children=empty_list),
+
+    'tspan': SVGElement('tspan',
+    attributes=['about', 'class', 'content', 'datatype', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'typeof', 'xml:base', 'xml:id', 'xml:lang', 'xml:space'],
+    properties=property_names,
+    children=['likeparent', 'text', 'textArea']),
+
+    'use': SVGElement('use',
+    attributes=['about', 'class', 'content', 'datatype', 'externalResourcesRequired', 'focusHighlight', 'focusable', 'id', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'property', 'rel', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'rev', 'role', 'systemLanguage', 'transform', 'typeof', 'x', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=property_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+
+    'video': SVGElement('video',
+    attributes=['about', 'begin', 'class', 'content', 'datatype', 'dur', 'end', 'externalResourcesRequired', 'fill', 'focusHighlight', 'focusable', 'height', 'id', 'initialVisibility', 'max', 'min', 'nav-down', 'nav-down-left', 'nav-down-right', 'nav-left', 'nav-next', 'nav-prev', 'nav-right', 'nav-up', 'nav-up-left', 'nav-up-right', 'overlay', 'preserveAspectRatio', 'property', 'rel', 'repeatCount', 'repeatDur', 'requiredExtensions', 'requiredFeatures', 'requiredFonts', 'requiredFormats', 'resource', 'restart', 'rev', 'role', 'syncBehavior', 'syncMaster', 'syncTolerance', 'systemLanguage', 'transform', 'transformBehavior', 'type', 'typeof', 'width', 'x', 'xlink:actuate', 'xlink:arcrole', 'xlink:href', 'xlink:role', 'xlink:show', 'xlink:title', 'xlink:type', 'xml:base', 'xml:id', 'xml:lang', 'xml:space', 'y'],
+    properties=media_group_names,
+    children=['animate', 'animateColor', 'animateMotion', 'animateTransform', 'desc', 'discard', 'handler', 'metadata', 'set', 'switch', 'title']),
+}
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/typechecker.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/typechecker.py
new file mode 100644
index 00000000..8cace93e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/typechecker.py
@@ -0,0 +1,342 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: typechecker
+# Created: 15.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+import re
+
+from svgwrite.data import pattern
+from svgwrite.data.colors import colornames
+from svgwrite.data.svgparser import is_valid_transferlist, is_valid_pathdata, is_valid_animation_timing
+from svgwrite.utils import is_string
+
+def iterflatlist(values):
+    """ Flatten nested *values*, returns an *iterator*. """
+    for element in values:
+        if hasattr(element, "__iter__") and not is_string(element):
+            for item in iterflatlist(element):
+                yield item
+        else:
+            yield element
+
+INVALID_NAME_CHARS = frozenset([' ', '\t', '\r', '\n', ',', '(', ')'])
+WHITESPACE = frozenset([' ', '\t', '\r', '\n'])
+SHAPE_PATTERN = re.compile(r"^rect\((.*),(.*),(.*),(.*)\)$")
+FUNCIRI_PATTERN = re.compile(r"^url\((.*)\)$")
+ICCCOLOR_PATTERN = re.compile(r"^icc-color\((.*)\)$")
+COLOR_HEXDIGIT_PATTERN = re.compile(r"^#[a-fA-F0-9]{3}([a-fA-F0-9]{3})?$")
+COLOR_RGB_INTEGER_PATTERN = re.compile(r"^rgb\( *\d+ *, *\d+ *, *\d+ *\)$")
+COLOR_RGB_PERCENTAGE_PATTERN = re.compile(r"^rgb\( *\d+(\.\d*)?% *, *\d+(\.\d*)?% *, *\d+(\.\d*)?% *\)$")
+NMTOKEN_PATTERN = re.compile(r"^[a-zA-Z_:][\w\-\.:]*$")
+
+
+class Full11TypeChecker(object):
+    def get_version(self):
+        return '1.1', 'full'
+
+    def is_angle(self, value):
+        #angle ::= number (~"deg" | ~"grad" | ~"rad")?
+        if self.is_number(value):
+            return True
+        elif is_string(value):
+            return pattern.angle.match(value.strip()) is not None
+        return False
+
+    def is_anything(self, value):
+        #anything ::= Char*
+        return bool(str(value).strip())
+    is_string = is_anything
+    is_content_type = is_anything
+
+    def is_color(self, value):
+        #color    ::= "#" hexdigit hexdigit hexdigit (hexdigit hexdigit hexdigit)?
+        #             | "rgb(" wsp* integer comma integer comma integer wsp* ")"
+        #             | "rgb(" wsp* number "%" comma number "%" comma number "%" wsp* ")"
+        #             | color-keyword
+        #hexdigit ::= [0-9A-Fa-f]
+        #comma    ::= wsp* "," wsp*
+        value = str(value).strip()
+        if value.startswith('#'):
+            if COLOR_HEXDIGIT_PATTERN.match(value):
+                return True
+            else:
+                return False
+        elif value.startswith('rgb('):
+            if COLOR_RGB_INTEGER_PATTERN.match(value):
+                return True
+            elif COLOR_RGB_PERCENTAGE_PATTERN.match(value):
+                return True
+            return False
+        return self.is_color_keyword(value)
+
+    def is_color_keyword(self, value):
+        return value.strip() in colornames
+
+    def is_frequency(self, value):
+        # frequency ::= number (~"Hz" | ~"kHz")
+        if self.is_number(value):
+            return True
+        elif is_string(value):
+            return pattern.frequency.match(value.strip()) is not None
+        return False
+
+    def is_FuncIRI(self, value):
+        # FuncIRI ::= "url(" <IRI> ")"
+        res = FUNCIRI_PATTERN.match(str(value).strip())
+        if res:
+            return self.is_IRI(res.group(1))
+        return False
+
+    def is_icccolor(self, value):
+        # icccolor ::= "icc-color(" name (comma-wsp number)+ ")"
+        res = ICCCOLOR_PATTERN.match(str(value).strip())
+        if res:
+            return self.is_list_of_T(res.group(1), 'name')
+        return False
+
+    def is_integer(self, value):
+        if isinstance(value, float):
+            return False
+        try:
+            number = int(value)
+            return True
+        except:
+            return False
+
+    def is_IRI(self, value):
+        # Internationalized Resource Identifiers
+        # a more generalized complement to Uniform Resource Identifiers (URIs)
+        # nearly everything can be a valid <IRI>
+        # only a none-empty string ist a valid input
+        if is_string(value):
+            return bool(value.strip())
+        else:
+            return False
+
+    def is_length(self, value):
+        # length ::= number ("em" | "ex" | "px" | "in" | "cm" | "mm" | "pt" | "pc" | "%")?
+        if value is None:
+            return False
+        if isinstance(value, (int, float)):
+            return self.is_number(value)
+        elif is_string(value):
+            result = pattern.length.match(value.strip())
+            if result:
+                number, tmp, unit = result.groups()
+                return self.is_number(number) # for tiny check!
+        return False
+
+    is_coordinate = is_length
+
+    def is_list_of_T(self, value, t='string'):
+        def split(value):
+            #TODO: improve split function!!!!
+            if isinstance(value, (int, float)):
+                return (value, )
+            if is_string(value):
+                return iterflatlist(v.split(',') for v in value.split(' '))
+            return value
+        #list-of-Ts ::= T
+        #               | T comma-wsp list-of-Ts
+        #comma-wsp  ::= (wsp+ ","? wsp*) | ("," wsp*)
+        #wsp        ::= (#x20 | #x9 | #xD | #xA)
+        checker = self.get_func_by_name(t)
+        for v in split(value):
+            if not checker(v):
+                return False
+        return True
+
+    def is_four_numbers(self, value):
+        def split(value):
+            if is_string(value):
+                values = iterflatlist( (v.strip().split(' ') for v in value.split(',')) )
+                return (v for v in values if v)
+            else:
+                return iterflatlist(value)
+
+        values = list(split(value))
+        if len(values) != 4:
+            return False
+        checker = self.get_func_by_name('number')
+        for v in values:
+            if not checker(v):
+                return False
+        return True
+
+    def is_semicolon_list(self, value):
+        #a semicolon-separated list of values
+        #               | value comma-wsp list-of-values
+        #comma-wsp  ::= (wsp+ ";" wsp*) | ("," wsp*)
+        #wsp        ::= (#x20 | #x9 | #xD | #xA)
+        return self.is_list_of_T(value.replace(';', ' '), 'string')
+
+    def is_name(self, value):
+        # name  ::= [^,()#x20#x9#xD#xA] /* any char except ",", "(", ")" or wsp */
+        chars = frozenset(str(value).strip())
+        if not chars or INVALID_NAME_CHARS.intersection(chars):
+            return False
+        else:
+            return True
+
+    def is_number(self, value):
+        try:
+            number = float(value)
+            return True
+        except:
+            return False
+
+    def is_number_optional_number(self, value):
+        #number-optional-number ::= number
+        #                           | number comma-wsp number
+        if is_string(value):
+            values = re.split('[ ,]+', value.strip())
+            if 0 < len(values) < 3: # 1 or 2 numbers
+                for v in values:
+                    if not self.is_number(v):
+                        return False
+                return True
+        else:
+            try: # is it a 2-tuple
+                n1, n2 = value
+                if self.is_number(n1) and \
+                   self.is_number(n2):
+                    return True
+            except TypeError: # just one value
+                return self.is_number(value)
+            except ValueError: # more than 2 values
+                pass
+        return False
+
+    def is_paint(self, value):
+        #paint ::=	"none" |
+        #           "currentColor" |
+        #           <color> [<icccolor>] |
+        #           <funciri> [ "none" | "currentColor" | <color> [<icccolor>] |
+        #           "inherit"
+        def split_values(value):
+            try:
+                funcIRI, value = value.split(")", 1)
+                values = [funcIRI+")"]
+                values.extend(split_values(value))
+                return values
+            except ValueError:
+                return value.split()
+
+        values = split_values(str(value).strip())
+        for value in [v.strip() for v in values]:
+            if value in ('none', 'currentColor', 'inherit'):
+                continue
+            elif self.is_color(value):
+                continue
+            elif self.is_icccolor(value):
+                continue
+            elif self.is_FuncIRI(value):
+                continue
+            return False
+        return True
+
+    def is_percentage(self, value):
+        #percentage ::= number "%"
+        if self.is_number(value):
+            return True
+        elif is_string(value):
+            return pattern.percentage.match(value.strip()) is not None
+        return False
+
+    def is_time(self, value):
+        #time ::= <number> (~"ms" | ~"s")?
+        if self.is_number(value):
+            return True
+        elif is_string(value):
+            return pattern.time.match(value.strip()) is not None
+        return False
+
+    def is_transform_list(self, value):
+        if is_string(value):
+            return is_valid_transferlist(value)
+        else:
+            return False
+
+    def is_path_data(self, value):
+        if is_string(value):
+            return is_valid_pathdata(value)
+        else:
+            return False
+
+    def is_XML_Name(self, value):
+        # http://www.w3.org/TR/2006/REC-xml-20060816/#NT-Name
+        # Nmtoken
+        return bool(NMTOKEN_PATTERN.match(str(value).strip()))
+
+    def is_shape(self, value):
+        # shape ::= (<top> <right> <bottom> <left>)
+        # where <top>, <bottom> <right>, and <left> specify offsets from the
+        # respective sides of the box.
+        # <top>, <right>, <bottom>, and <left> are <length> values
+        # i.e. 'rect(5px, 10px, 10px, 5px)'
+        res = SHAPE_PATTERN.match(value.strip())
+        if res:
+            for arg in res.groups():
+                if arg.strip() == 'auto':
+                    continue
+                if not self.is_length(arg):
+                    return False
+        else:
+            return False
+        return True
+
+    def is_timing_value_list(self, value):
+        if is_string(value):
+            return is_valid_animation_timing(value)
+        else:
+            return False
+
+    def is_list_of_text_decoration_style(self, value):
+        return self.is_list_of_T(value, t='text_decoration_style')
+
+    def is_text_decoration_style(self, value):
+        return value in ('overline', 'underline', 'line-through', 'blink')
+
+    def get_func_by_name(self, funcname):
+        return getattr(self,
+                       'is_'+funcname.replace('-', '_'),
+                       self.is_anything)
+
+    def check(self, typename, value):
+        if typename.startswith('list-of-'):
+            t = typename[8:]
+            return self.is_list_of_T(value, t)
+        return self.get_func_by_name(typename)(value)
+
+
+FOCUS_CONST = frozenset(['nav-next', 'nav-prev', 'nav-up', 'nav-down', 'nav-left',
+                         'nav-right', 'nav-up-left', 'nav-up-right', 'nav-down-left',
+                         'nav-down-right'])
+
+
+class Tiny12TypeChecker(Full11TypeChecker):
+    def get_version(self):
+        return '1.2', 'tiny'
+
+    def is_boolean(self, value):
+        if isinstance(value, bool):
+            return True
+        if is_string(value):
+            return value.strip().lower() in ('true', 'false')
+        return False
+
+    def is_number(self, value):
+        try:
+            number = float(value)
+            if -32767.9999 <= number <= 32767.9999:
+                return True
+            else:
+                return False
+        except:
+            return False
+
+    def is_focus(self, value):
+        return str(value).strip() in FOCUS_CONST
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/data/types.py b/backend/venv/lib/python3.12/site-packages/svgwrite/data/types.py
new file mode 100644
index 00000000..5c6f3025
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/data/types.py
@@ -0,0 +1,81 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: svg types
+# Created: 30.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+
+class SVGAttribute(object):
+    def __init__(self, name, anim, types, const):
+        self.name = name
+        self._anim = anim
+        self._types = types
+        self._const = const
+
+    # 'elementname' is ignored, but necessary because of the signatures of
+    # the SVGMultiAttribute class methods get_...()
+
+    def get_anim(self, elementname='*'):
+        return self._anim
+
+    def get_types(self, elementname='*'):
+        return self._types
+
+    def get_const(self, elementname='*'):
+        return self._const
+
+
+class SVGMultiAttribute(object):
+    # example: SVGMultiAttribute({'*':SVGAttribute(...), 'text tref':SVGAttribute(...)} )
+    # parametr is a dict-like object
+    # '*' is the default attribute definition
+    # 'text' and 'tref' share the same attribute definition
+
+    def __init__(self, attributes):
+        self.name = None
+        self._attributes = {}
+
+        for names, attribute in attributes.items():
+            for name in names.split():
+                name = name.strip()
+                self._attributes[name] = attribute
+                if not self.name:
+                    self.name = attribute.name
+                elif self.name != attribute.name:
+                    raise ValueError("Different attribute-names for SVGMultiAttribute "\
+                                     "(%s != %s)." % (self.name, attribute.name))
+
+        if '*' not in self._attributes and len(self._attributes):
+            # if no default attribute definition were given
+            # set the first attribute definition as the default attribute definition
+            firstkey = sorted(self._attributes.keys())[0]
+            self._attributes['*'] = self._attributes[firstkey]
+
+    def get_attribute(self, elementname):
+        if elementname in self._attributes:
+            return self._attributes[elementname]
+        else:
+            return self._attributes['*']
+
+    def get_anim(self, elementname='*'):
+        attribute = self.get_attribute(elementname)
+        return attribute.get_anim()
+
+    def get_types(self, elementname='*'):
+        attribute = self.get_attribute(elementname)
+        return attribute.get_types()
+
+    def get_const(self, elementname='*'):
+        attribute = self.get_attribute(elementname)
+        return attribute.get_const()
+
+
+class SVGElement(object):
+    def __init__(self, name, attributes, properties, children):
+        self.name = name
+        s = set(attributes)
+        s.update(properties)
+        self.valid_attributes = frozenset(s)
+        self.valid_children = frozenset(children)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/drawing.py b/backend/venv/lib/python3.12/site-packages/svgwrite/drawing.py
new file mode 100644
index 00000000..26e04502
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/drawing.py
@@ -0,0 +1,140 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman
+# Purpose: drawing
+# Created: 10.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+"""
+The *Drawing* object is the overall container for all SVG
+elements. It provides the methods to store the drawing into a file or a
+file-like object. If you want to use stylesheets, the reference links
+to this stylesheets were also stored (`add_stylesheet`)
+in the *Drawing* object.
+
+set/get SVG attributes::
+
+    element['attribute'] = value
+    value = element['attribute']
+
+The Drawing object also includes a defs section, add elements to the defs
+section by::
+
+    drawing.defs.add(element)
+
+"""
+import io
+
+from svgwrite.container import SVG, Defs
+from svgwrite.elementfactory import ElementFactory
+from svgwrite.utils import pretty_xml
+
+
+class Drawing(SVG, ElementFactory):
+    """ This is the SVG drawing represented by the top level *svg* element.
+
+    A drawing consists of any number of SVG elements contained within the drawing
+    element, stored in the *elements* attribute.
+
+    A drawing can range from an empty drawing (i.e., no content inside of the drawing),
+    to a very simple drawing containing a single SVG element such as a *rect*,
+    to a complex, deeply nested collection of container elements and graphics elements.
+    """
+    def __init__(self, filename="noname.svg", size=('100%', '100%'), **extra):
+        """
+        :param string filename: filesystem filename valid for :func:`open`
+        :param 2-tuple size: width, height
+        :param keywords extra: additional svg-attributes for the *SVG* object
+
+        Important (and not SVG Attributes) **extra** parameters:
+
+        :param string profile: ``'tiny | full'`` - define the SVG baseProfile
+        :param bool debug: switch validation on/off
+
+        """
+        super(Drawing, self).__init__(size=size, **extra)
+        self.filename = filename
+        self._stylesheets = []  # list of stylesheets appended
+
+    def get_xml(self):
+        """ Get the XML representation as `ElementTree` object.
+
+        :return: XML `ElementTree` of this object and all its subelements
+
+        """
+        profile = self.profile
+        version = self.version
+        self.attribs['xmlns'] = "http://www.w3.org/2000/svg"
+        self.attribs['xmlns:xlink'] = "http://www.w3.org/1999/xlink"
+        self.attribs['xmlns:ev'] = "http://www.w3.org/2001/xml-events"
+
+        self.attribs['baseProfile'] = profile
+        self.attribs['version'] = version
+        return super(Drawing, self).get_xml()
+
+    def add_stylesheet(self, href, title, alternate="no", media="screen"):
+        """ Add a stylesheet reference.
+
+        :param string href: link to stylesheet <URI>
+        :param string title: name of stylesheet
+        :param string alternate: ``'yes'|'no'``
+        :param string media: ``'all | aureal | braille | embossed | handheld | print | projection | screen | tty | tv'``
+
+        """
+        self._stylesheets.append((href, title, alternate, media))
+
+    def write(self, fileobj, pretty=False, indent=2):
+        """ Write XML string to `fileobj`.
+
+        :param fileobj: a file-like object
+        :param pretty: True for easy readable output
+        :param indent: how much to indent if pretty is enabled, by default 2 spaces
+
+        Python 3.x - set encoding at the open command::
+
+            open('filename', 'w', encoding='utf-8')
+        """
+        # write xml header
+        fileobj.write('<?xml version="1.0" encoding="utf-8" ?>\n')
+
+        # don't use DOCTYPE. It's useless. see also:
+        # http://tech.groups.yahoo.com/group/svg-developers/message/48562
+        # write stylesheets
+        stylesheet_template = '<?xml-stylesheet href="%s" type="text/css" ' \
+            'title="%s" alternate="%s" media="%s"?>\n'
+        # removed map(), does not work with Python 3
+        for stylesheet in self._stylesheets:
+            fileobj.write(stylesheet_template % stylesheet)
+
+        xml_string = self.tostring()
+        if pretty:  # write easy readable XML file
+            xml_string = pretty_xml(xml_string, indent=indent)
+        fileobj.write(xml_string)
+
+    def save(self, pretty=False, indent=2):
+        """ Write the XML string to `self.filename`.
+
+        :param pretty: True for easy readable output
+        :param indent: how much to indent if pretty is enabled, by default 2 spaces
+        """
+        fileobj = io.open(self.filename, mode='w', encoding='utf-8')
+        self.write(fileobj, pretty=pretty, indent=indent)
+        fileobj.close()
+
+    def saveas(self, filename, pretty=False, indent=2):
+        """ Write the XML string to `filename`.
+
+        :param string filename: filesystem filename valid for :func:`open`
+        :param pretty: True for easy readable output
+        :param indent: how much to indent if pretty is enabled, by default 2 spaces
+        """
+        self.filename = filename
+        self.save(pretty=pretty, indent=indent)
+
+    def _repr_svg_(self):
+        """ Show SVG in IPython, Jupyter Notebook, and Jupyter Lab
+
+        :return: unicode XML string of this object and all its subelements
+
+        """
+        return self.tostring()
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/elementfactory.py b/backend/venv/lib/python3.12/site-packages/svgwrite/elementfactory.py
new file mode 100644
index 00000000..c7f60765
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/elementfactory.py
@@ -0,0 +1,76 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: element factory
+# Created: 15.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite import container
+from svgwrite import shapes
+from svgwrite import path
+from svgwrite import image
+from svgwrite import text
+from svgwrite import gradients
+from svgwrite import pattern
+from svgwrite import masking
+from svgwrite import animate
+from svgwrite import filters
+from svgwrite import solidcolor
+
+factoryelements = {
+    'g': container.Group,
+    'svg': container.SVG,
+    'defs': container.Defs,
+    'symbol': container.Symbol,
+    'marker': container.Marker,
+    'use': container.Use,
+    'a': container.Hyperlink,
+    'script': container.Script,
+    'style': container.Style,
+    'line': shapes.Line,
+    'rect': shapes.Rect,
+    'circle': shapes.Circle,
+    'ellipse': shapes.Ellipse,
+    'polyline': shapes.Polyline,
+    'polygon': shapes.Polygon,
+    'path': path.Path,
+    'image': image.Image,
+    'text': text.Text,
+    'tspan': text.TSpan,
+    'tref': text.TRef,
+    'textPath': text.TextPath,
+    'textArea': text.TextArea,
+    'linearGradient': gradients.LinearGradient,
+    'radialGradient': gradients.RadialGradient,
+    'pattern': pattern.Pattern,
+    'solidColor': solidcolor.SolidColor,
+    'clipPath': masking.ClipPath,
+    'mask': masking.Mask,
+    'animate': animate.Animate,
+    'set': animate.Set,
+    'animateColor': animate.AnimateColor,
+    'animateMotion': animate.AnimateMotion,
+    'animateTransform': animate.AnimateTransform,
+    'filter': filters.Filter,
+}
+
+
+class ElementBuilder(object):
+    def __init__(self, cls, factory):
+        self.cls = cls
+        self.factory = factory
+
+    def __call__(self, *args, **kwargs):
+        # inject creator object - inherit _parameter from factory
+        kwargs['factory'] = self.factory
+        # create an object of type 'cls'
+        return self.cls(*args, **kwargs)
+
+
+class ElementFactory(object):
+    def __getattr__(self, name):
+        if name in factoryelements:
+            return ElementBuilder(factoryelements[name], self)
+        else:
+            raise AttributeError("'%s' has no attribute '%s'" % (self.__class__.__name__, name))
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/etree.py b/backend/venv/lib/python3.12/site-packages/svgwrite/etree.py
new file mode 100644
index 00000000..4e897b69
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/etree.py
@@ -0,0 +1,49 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman
+# Purpose: a hack to generate XML containing CDATA by ElementTree
+# Created: 26.05.2012
+# Copyright (C) 2012, Manfred Moitzi
+# License: MIT License
+
+# usage:
+#
+# from svgwrite.etree import etree, CDATA
+#
+# element = etree.Element('myTag')
+# element.append(CDATA("< and >"))
+#
+# assert etree.tostring(element) == "<myTag><![CDATA[< and >]]></myTag>"
+
+
+import sys
+PY3 = sys.version_info[0] > 2
+
+import xml.etree.ElementTree as etree
+
+CDATA_TPL = "<![CDATA[%s]]>"
+CDATA_TAG = CDATA_TPL
+
+
+def CDATA(text):
+    element = etree.Element(CDATA_TAG)
+    element.text = text
+    return element
+
+original_serialize_xml = etree._serialize_xml
+
+if PY3:
+    def _serialize_xml_with_CDATA_support(write, elem, qnames, namespaces, **kwargs):
+        if elem.tag == CDATA_TAG:
+            write(CDATA_TPL % elem.text)
+        else:
+            original_serialize_xml(write, elem, qnames, namespaces, **kwargs)
+else:
+    def _serialize_xml_with_CDATA_support(write, elem, encoding, qnames, namespaces):
+        if elem.tag == CDATA_TAG:
+            write(CDATA_TPL % elem.text.encode(encoding))
+        else:
+            original_serialize_xml(write, elem, encoding, qnames, namespaces)
+
+# ugly, ugly, ugly patching
+etree._serialize_xml = _serialize_xml_with_CDATA_support
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__init__.py b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__init__.py
new file mode 100644
index 00000000..0481e963
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__init__.py
@@ -0,0 +1,3 @@
+# Copyright (c) 2018 Manfred Moitzi
+# License: MIT License
+from .inkscape import Inkscape
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..044ee207
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/inkscape.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/inkscape.cpython-312.pyc
new file mode 100644
index 00000000..ad2c8966
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/inkscape.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/shapes.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/shapes.cpython-312.pyc
new file mode 100644
index 00000000..a2d31cf5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/__pycache__/shapes.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/inkscape.py b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/inkscape.py
new file mode 100644
index 00000000..7ea5c029
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/inkscape.py
@@ -0,0 +1,89 @@
+# Copyright (c) 2018 Manfred Moitzi
+# License: MIT License
+# based on work of:
+# Copyright (c) 2018 Antonio Ospite <ao2@ao2.it>
+
+from svgwrite.data.types import SVGAttribute
+
+INKSCAPE_NAMESPACE = 'http://www.inkscape.org/namespaces/inkscape'
+SODIPODI_NAMESPACE = 'http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd'
+INKSCAPE_ATTRIBUTES = {
+    'xmlns:inkscape': SVGAttribute('xmlns:inkscape',
+                                   anim=False,
+                                   types=[],
+                                   const=frozenset([INKSCAPE_NAMESPACE])),
+    'xmlns:sodipodi': SVGAttribute('xmlns:sodipodi',
+                                   anim=False,
+                                   types=[],
+                                   const=frozenset([SODIPODI_NAMESPACE])),
+    'inkscape:groupmode': SVGAttribute('inkscape:groupmode',
+                                       anim=False,
+                                       types=[],
+                                       const=frozenset(['layer'])),
+    'inkscape:label': SVGAttribute('inkscape:label',
+                                   anim=False,
+                                   types=frozenset(['string']),
+                                   const=[]),
+    'sodipodi:insensitive': SVGAttribute('sodipodi:insensitive',
+                                         anim=False,
+                                         types=[],
+                                         const=frozenset(['true', 'false', '0', '1']))
+}
+
+
+def _setup_validator(validator):
+    # setup already done?
+    if 'xmlns:inkscape' in validator.attributes:
+        return
+
+    validator.attributes.update(INKSCAPE_ATTRIBUTES)
+    elements = validator.elements
+
+    # extend SVG attributes
+    elements['svg'].valid_attributes = \
+        {
+            'xmlns:inkscape',
+            'xmlns:sodipodi',
+        } | elements['svg'].valid_attributes
+
+    # extend group attributes
+    elements['g'].valid_attributes = \
+        {
+            'inkscape:groupmode',
+            'inkscape:label',
+            'sodipodi:insensitive',
+        } | elements['g'].valid_attributes
+
+
+GROUP_MODE = 'inkscape:groupmode'
+LABEL = 'inkscape:label'
+INSENSITIVE = 'sodipodi:insensitive'
+
+
+class Inkscape(object):
+    """
+    Extension to support SOME Inkscape features.
+
+    """
+    def __init__(self, drawing):
+        self.svg = drawing
+        _setup_validator(drawing.validator)
+        drawing['xmlns:inkscape'] = INKSCAPE_NAMESPACE
+        drawing['xmlns:sodipodi'] = SODIPODI_NAMESPACE
+
+    def layer(self, label=None, locked=False, **kwargs):
+        """
+        Create new Inkscape layer.
+
+        Args:
+            label: layer name as string
+            locked: when set to True, make objects at this layer unselectable
+
+        """
+        new_layer = self.svg.g(**kwargs)
+        new_layer[GROUP_MODE] = 'layer'
+        if label is not None:
+            new_layer[LABEL] = label
+        if locked:
+            new_layer[INSENSITIVE] = 1
+        return new_layer
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/shapes.py b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/shapes.py
new file mode 100644
index 00000000..d2714580
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/extensions/shapes.py
@@ -0,0 +1,129 @@
+"""
+Extension to create and manipulate shapes
+"""
+# Copyright (c) 2019 Christof Hanke (christof.hanke@induhviduals.de)
+# License: MIT License
+import math
+
+
+def ngon(num_corners, edge_length=None, radius=None, rotation=0.):
+    """
+    Returns the corners of a regular polygon as iterable of (x, y) tuples. The polygon size is determined by the
+    `edge_length` or the `radius` argument. If both are given `edge_length` will be taken.
+
+    Args:
+        num_corners: count of polygon corners
+        edge_length: length of polygon side
+        radius: circum radius
+        rotation: rotation angle in radians
+
+    Returns: iterable of (x, y) tuples
+
+    """
+    if num_corners < 3:
+        raise ValueError('Argument `num_corners` has to be greater than 2.')
+    if edge_length is not None:
+        radius = edge_length / 2 / math.sin(math.pi / num_corners)
+    elif radius is not None:
+        if radius <= 0.:
+            raise ValueError('Argument `radius` has to be greater than 0.')
+    else:
+        raise ValueError('Argument `edge_length` or `radius` required.')
+
+    delta = 2 * math.pi / num_corners
+    angle = rotation
+    for _ in range(num_corners):
+        yield (radius * math.cos(angle), radius * math.sin(angle))
+        angle += delta
+
+
+def star(spikes, r1, r2, rotation=0.):
+    """
+    Create a star shape as iterable of (x, y) vertices.
+
+    Argument `spikes` defines the count of star spikes, `r1` defines the radius of the "outer" vertices and `r2`
+    defines the radius of the "inner" vertices, but this does not mean that `r1` has to greater than `r2`.
+
+    Args:
+        spikes: spike count
+        r1: radius 1
+        r2: radius 2
+        rotation: rotation angle in radians
+
+    Returns: iterable of (x, y) tuples
+
+    """
+    if spikes < 3:
+        raise ValueError('Argument `spikes` has to be greater than 2.')
+    if r1 <= 0.:
+        raise ValueError('Argument `r1` has to be greater than 0.')
+    if r2 <= 0.:
+        raise ValueError('Argument `r2` has to be greater than 0.')
+
+    corners1 = ngon(spikes, radius=r1, rotation=rotation)
+    corners2 = ngon(spikes, radius=r2, rotation=math.pi/spikes+rotation)
+    for s1, s2 in zip(corners1, corners2):
+        yield s1
+        yield s2
+
+
+def translate(vertices, delta_x, delta_y):
+    """
+    Translates `vertices` about `delta_x` and `delta_y`
+
+    Args:
+         vertices: iterable of (x, y) tuples
+         delta_x: translation in x axis
+         delta_y: translation in y axis
+
+    Returns: iterable of (x, y) tuples
+
+    """
+    for x, y in vertices:
+        yield (x + delta_x, y + delta_y)
+
+
+def scale(vertices, scale_x, scale_y):
+    """
+    Scales `vertices` about `scale_x` and `scale_y`
+
+    Args:
+         vertices: iterable of (x, y) tuples
+         scale_x: scaling factor in x axis direction
+         scale_y: scaling factor in y axis direction
+
+    Returns: iterable of (x, y) tuples
+
+    """
+    for x, y in vertices:
+        yield (x * scale_x, y * scale_y)
+
+
+def rotate(vertices, delta):
+    """
+    Rotates `vertices` about `delta` degrees around the origin (0, 0).
+
+    Args:
+         vertices: iterable of (x, y) tuples
+         delta: rotation angle in radians
+
+    Returns: iterable of (x, y) tuples
+
+    """
+    for x, y in vertices:
+        r = math.hypot(x, y)
+        angle = math.atan2(y, x) + delta
+        yield (r * math.cos(angle), r * math.sin(angle))
+
+
+def centroid(vertices):
+    """
+    Returns the centroid of a series of `vertices`.
+
+    """
+    k, c_x, c_y = 0, 0, 0
+    for x, y in vertices:
+        c_x += x
+        c_y += y
+        k += 1
+    return c_x / k, c_y / k
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/filters.py b/backend/venv/lib/python3.12/site-packages/svgwrite/filters.py
new file mode 100644
index 00000000..1d3e632a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/filters.py
@@ -0,0 +1,273 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: filters module
+# Created: 03.11.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import XLink, Presentation
+from svgwrite.utils import strlist, is_string
+
+__all__ = ['Filter']
+
+
+class _feDistantLight(BaseElement):
+    elementname = 'feDistantLight'
+
+    def __init__(self, azimuth=0, elevation=0, **extra):
+        super(_feDistantLight, self).__init__(**extra)
+        if azimuth != 0:
+            self['azimuth'] = azimuth
+        if elevation != 0:
+            self['elevation'] = elevation
+
+
+class _fePointLight(BaseElement):
+    elementname = 'fePointLight'
+
+    def __init__(self, source=(0, 0, 0), **extra):
+        super(_fePointLight, self).__init__(**extra)
+        x, y, z = source
+        if x != 0:
+            self['x'] = x
+        if y != 0:
+            self['y'] = y
+        if z != 0:
+            self['z'] = z
+
+
+class _feSpotLight(_fePointLight):
+    elementname = 'feSpotLight'
+
+    def __init__(self, source=(0, 0, 0), target=(0, 0, 0), **extra):
+        super(_feSpotLight, self).__init__(source, **extra)
+        x, y, z = target
+        if x != 0:
+            self['pointsAtX'] = x
+        if y != 0:
+            self['pointsAtY'] = y
+        if z != 0:
+            self['pointsAtZ'] = z
+
+
+class _FilterPrimitive(BaseElement, Presentation):
+    pass
+
+
+class _FilterNoInput(_FilterPrimitive):
+    def __init__(self, start=None, size=None, **extra):
+        super(_FilterNoInput, self).__init__(**extra)
+        if start is not None:
+            self['x'] = start[0]
+            self['y'] = start[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+
+
+class _FilterRequireInput(_FilterNoInput):
+    def __init__(self, in_='SourceGraphic', **extra):
+        super(_FilterRequireInput, self).__init__(**extra)
+        self['in'] = in_
+
+
+class _feBlend(_FilterRequireInput):
+    elementname = 'feBlend'
+
+
+class _feColorMatrix(_FilterRequireInput):
+    elementname = 'feColorMatrix'
+
+
+class _feComponentTransfer(_FilterRequireInput):
+    elementname = 'feComponentTransfer'
+
+    def feFuncR(self, type_, **extra):
+        return self.add(_feFuncR(type_, factory=self, **extra))
+
+    def feFuncG(self, type_, **extra):
+        return self.add(_feFuncG(type_, factory=self, **extra))
+
+    def feFuncB(self, type_, **extra):
+        return self.add(_feFuncB(type_, factory=self, **extra))
+
+    def feFuncA(self, type_, **extra):
+        return self.add(_feFuncA(type_, factory=self, **extra))
+
+
+class _feFuncR(_FilterPrimitive):
+    elementname = 'feFuncR'
+
+    def __init__(self, type_, **extra):
+        super(_feFuncR, self).__init__(**extra)
+        self['type'] = type_
+
+
+class _feFuncG(_feFuncR):
+    elementname = 'feFuncG'
+
+
+class _feFuncB(_feFuncR):
+    elementname = 'feFuncB'
+
+
+class _feFuncA(_feFuncR):
+    elementname = 'feFuncA'
+
+
+class _feComposite(_FilterRequireInput):
+    elementname = 'feComposite'
+
+
+class _feConvolveMatrix(_FilterRequireInput):
+    elementname = 'feConvolveMatrix'
+
+
+class _feDiffuseLighting(_FilterRequireInput):
+    elementname = 'feDiffuseLighting'
+
+    def feDistantLight(self, azimuth=0, elevation=0, **extra):
+        return self.add(_feDistantLight(azimuth, elevation, **extra))
+
+    def fePointLight(self, source=(0, 0, 0), **extra):
+        return self.add(_fePointLight(source, **extra))
+
+    def feSpotLight(self, source=(0, 0, 0), target=(0, 0, 0), **extra):
+        return self.add(_feSpotLight(source, target, **extra))
+
+
+class _feDisplacementMap(_FilterRequireInput):
+    elementname = 'feDisplacementMap'
+
+
+class _feFlood(_FilterNoInput):
+    elementname = 'feFlood'
+
+
+class _feGaussianBlur(_FilterRequireInput):
+    elementname = 'feGaussianBlur'
+
+
+class _feImage(_FilterNoInput, XLink):
+    elementname = 'feImage'
+
+    def __init__(self, href, start=None, size=None, **extra):
+        super(_feImage, self).__init__(start, size, **extra)
+        self.set_href(href)
+
+
+class _feMergeNode(_FilterPrimitive):
+    elementname = 'feMergeNode'
+
+
+class _feMerge(_FilterNoInput):
+    elementname = 'feMerge'
+    def __init__(self, layernames, **extra):
+        super(_feMerge, self).__init__(**extra)
+        self.feMergeNode(layernames)
+
+    def feMergeNode(self, layernames):
+        for layername in layernames:
+            self.add(_feMergeNode(in_=layername, factory=self))
+
+
+class _feMorphology(_FilterRequireInput):
+    elementname = 'feMorphology'
+
+
+class _feOffset(_FilterRequireInput):
+    elementname = 'feOffset'
+
+
+class _feSpecularLighting(_feDiffuseLighting):
+    elementname = 'feSpecularLighting'
+
+
+class _feTile(_FilterRequireInput):
+    elementname = 'feTile'
+
+
+class _feTurbulence(_FilterNoInput):
+    elementname = 'feTurbulence'
+
+
+filter_factory = {
+    'feBlend': _feBlend,
+    'feColorMatrix': _feColorMatrix,
+    'feComponentTransfer': _feComponentTransfer,
+    'feComposite': _feComposite,
+    'feConvolveMatrix': _feConvolveMatrix,
+    'feDiffuseLighting': _feDiffuseLighting,
+    'feDisplacementMap': _feDisplacementMap,
+    'feFlood': _feFlood,
+    'feGaussianBlur': _feGaussianBlur,
+    'feImage': _feImage,
+    'feMerge': _feMerge,
+    'feMorphology': _feMorphology,
+    'feOffset': _feOffset,
+    'feSpecularLighting': _feSpecularLighting,
+    'feTile': _feTile,
+    'feTurbulence': _feTurbulence,
+}
+
+
+class _FilterBuilder(object):
+    def __init__(self, cls, parent):
+        self.cls = cls # primitive filter class to build
+        self.parent = parent # the parent Filter() object
+
+    def __call__(self, *args, **kwargs):
+        kwargs['factory'] = self.parent # to get the _paramters object
+        obj = self.cls(*args, **kwargs) # create an object of type 'cls'
+        self.parent.add(obj) # add primitive filter to parent Filter()
+        return obj
+
+
+class Filter(BaseElement, XLink, Presentation):
+    """
+    The filter element is a container element for filter primitives, and
+    also a **factory** for filter primitives.
+    """
+    elementname = 'filter'
+
+    def __init__(self, start=None, size=None, resolution=None, inherit=None, **extra):
+        """
+        :param 2-tuple start: defines the start point of the filter effects region (**x**, **y**)
+        :param 2-tuple size: defines the size of the filter effects region (**width**, **height**)
+        :param resolution: takes the form ``'x-pixels [y-pixels]'``, and indicates
+            the width and height of the intermediate images in pixels.
+        :param inherit: inherits properties from Filter `inherit` see: **xlink:href**
+
+        """
+        super(Filter, self).__init__(**extra)
+        if start is not None:
+            self['x'] = start[0]
+            self['y'] = start[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+        if resolution is not None:
+            if is_string(resolution):
+                self['filterRes'] = resolution
+            elif hasattr(resolution, '__iter__'):
+                self['filterRes'] = strlist(resolution, ' ')
+            else:
+                self['filterRes'] = str(resolution)
+
+        if inherit is not None:
+            self.href = inherit
+            self.update_id()
+
+    def get_xml(self):
+        self.update_id()
+        return super(Filter, self).get_xml()
+
+    def __getattr__(self, name):
+        # create primitive filters by Filter.<filtername>(...)
+        # and auto-add the new filter as subelement of Filter()
+        if name in filter_factory:
+            return _FilterBuilder(filter_factory[name], self)
+        else:
+            raise AttributeError("'%s' has no attribute '%s'" % (self.__class__.__name__, name))
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/gradients.py b/backend/venv/lib/python3.12/site-packages/svgwrite/gradients.py
new file mode 100644
index 00000000..0857af3d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/gradients.py
@@ -0,0 +1,128 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: gradients module
+# Created: 26.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+"""
+Gradients consist of continuously smooth color transitions along a vector
+from one color to another, possibly followed by additional transitions along
+the same vector to other colors. SVG provides for two types of gradients:
+linear gradients and radial gradients.
+"""
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import Transform, XLink
+from svgwrite.utils import is_string
+
+
+class _GradientStop(BaseElement):
+    elementname = 'stop'
+
+    def __init__(self, offset=None, color=None, opacity=None, **extra):
+        super(_GradientStop, self).__init__(**extra)
+
+        if offset is not None:
+            self['offset'] = offset
+        if color is not None:
+            self['stop-color'] = color
+        if opacity is not None:
+            self['stop-opacity'] = opacity
+
+
+class _AbstractGradient(BaseElement, Transform, XLink):
+    transformname = 'gradientTransform'
+
+    def __init__(self, inherit=None, **extra):
+        super(_AbstractGradient, self).__init__(**extra)
+        if inherit is not None:
+            if is_string(inherit):
+                self.set_href(inherit)
+            else:
+                self.set_href(inherit.get_iri())
+
+    def get_paint_server(self, default='none'):
+        """ Returns the <FuncIRI> of the gradient. """
+        return "%s %s" % (self.get_funciri(), default)
+
+    def add_stop_color(self, offset=None, color=None, opacity=None):
+        """ Adds a stop-color to the gradient.
+
+        :param offset: is either a <number> (usually ranging from 0 to 1) or
+          a `<percentage>` (usually ranging from 0% to 100%) which indicates where
+          the gradient stop is placed. Represents a location along the gradient
+          vector. For radial gradients, it represents a percentage distance from
+          (fx,fy) to the edge of the outermost/largest circle.
+        :param color: indicates what color to use at that gradient stop
+        :param opacity: defines the opacity of a given gradient stop
+        """
+        self.add(_GradientStop(offset, color, opacity, factory=self))
+        return self
+
+    def add_colors(self, colors, sweep=(0., 1.), opacity=None):
+        """ Add stop-colors from colors with linear offset distributuion
+        from sweep[0] to sweep[1].
+
+        i.e. colors=['white', 'red', 'blue']
+          'white': offset = 0.0
+          'red': offset = 0.5
+          'blue': offset = 1.0
+        """
+        delta = (sweep[1] - sweep[0]) / (len(colors) - 1)
+        offset = sweep[0]
+        for color in colors:
+            self.add_stop_color(round(offset, 3), color, opacity)
+            offset += delta
+        return self
+
+    def get_xml(self):
+        if hasattr(self, 'href'):
+            self.update_id()
+        return super(_AbstractGradient, self).get_xml()
+
+
+class LinearGradient(_AbstractGradient):
+    """ Linear gradients are defined by a SVG <linearGradient> element.
+    """
+    elementname = 'linearGradient'
+
+    def __init__(self, start=None, end=None, inherit=None, **extra):
+        """
+        :param 2-tuple start: start point of the gradient (**x1**, **y1**)
+        :param 2-tuple end: end point of the gradient (**x2**, **y2**)
+        :param inherit: gradient inherits properties from `inherit` see: **xlink:href**
+
+        """
+        super(LinearGradient, self).__init__(inherit=inherit, **extra)
+        if start is not None:
+            self['x1'] = start[0]
+            self['y1'] = start[1]
+        if end is not None:
+            self['x2'] = end[0]
+            self['y2'] = end[1]
+
+
+class RadialGradient(_AbstractGradient):
+    """ Radial gradients are defined by a SVG <radialGradient> element.
+    """
+    elementname = 'radialGradient'
+
+    def __init__(self, center=None, r=None, focal=None, inherit=None, **extra):
+        """
+        :param 2-tuple center: center point for the gradient (**cx**, **cy**)
+        :param r: radius for the gradient
+        :param 2-tuple focal: focal point for the radial gradient (**fx**, **fy**)
+        :param inherit: gradient inherits properties from `inherit` see: **xlink:href**
+
+        """
+
+        super(RadialGradient, self).__init__(inherit=inherit, **extra)
+        if center is not None:
+            self['cx'] = center[0]
+            self['cy'] = center[1]
+        if r is not None:
+            self['r'] = r
+        if focal is not None:
+            self['fx'] = focal[0]
+            self['fy'] = focal[1]
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/image.py b/backend/venv/lib/python3.12/site-packages/svgwrite/image.py
new file mode 100644
index 00000000..64c9710e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/image.py
@@ -0,0 +1,61 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: svg image element
+# Created: 09.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import Transform, _vert, _horiz, Clipping
+
+class Image(BaseElement, Transform, Clipping):
+    """ The **image** element indicates that the contents of a complete file are
+    to be rendered into a given rectangle within the current user coordinate
+    system. The **image** element can refer to raster image files such as PNG
+    or JPEG or to files with MIME type of "image/svg+xml".
+
+    """
+    elementname = 'image'
+
+    def __init__(self, href, insert=None, size=None, **extra):
+        """
+        :param string href: hyperlink to the image resource
+        :param 2-tuple insert: insert point (**x**, **y**)
+        :param 2-tuple size: (**width**, **height**)
+        :param dict attribs: additional SVG attributes
+        :param extra: additional SVG attributes as keyword-arguments
+        """
+        super(Image, self).__init__(**extra)
+        self['xlink:href'] = href
+        if insert is not None:
+            self['x'] = insert[0]
+            self['y'] = insert[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+
+    def stretch(self):
+        """ Stretch viewBox in x and y direction to fill viewport, does not
+        preserve aspect ratio.
+        """
+        self['preserveAspectRatio'] = 'none'
+
+    def fit(self, horiz="center", vert="middle", scale="meet"):
+        """ Set the preserveAspectRatio attribute.
+
+        :param string horiz: horizontal alignment ``'left'|'center'|'right'``
+        :param string vert: vertical alignment ``'top'|'middle'|'bottom'``
+        :param string scale: scale method ``'meet'|'slice'``
+
+        ============= ===========
+        Scale methods Description
+        ============= ===========
+        ``meet``      preserve aspect ration and zoom to limits of viewBox
+        ``slice``     preserve aspect ration and viewBox touch viewport on all bounds, viewBox will extend beyond the bounds of the viewport
+        ============= ===========
+
+        """
+        if self.debug and scale not in ('meet', 'slice'):
+            raise ValueError("Invalid scale parameter '%s'" % scale)
+        self.attribs['preserveAspectRatio'] = "%s%s %s" % (_horiz[horiz],_vert[vert], scale)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/masking.py b/backend/venv/lib/python3.12/site-packages/svgwrite/masking.py
new file mode 100644
index 00000000..c6c344bf
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/masking.py
@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: masking module
+# Created: 30.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import Transform
+
+
+class ClipPath(BaseElement, Transform):
+    """
+    The clipping path restricts the region to which paint can be applied.
+    Conceptually, any parts of the drawing that lie outside of the region
+    bounded by the currently active clipping path are not drawn. A clipping
+    path can be thought of as a mask wherein those pixels outside the clipping
+    path are black with an alpha value of zero and those pixels inside the
+    clipping path are white with an alpha value of one (with the possible
+    exception of anti-aliasing along the edge of the silhouette).
+
+    A **clipPath** element can contain **path** elements, **text** elements,
+    basic shapes (such as **circle**) or a **use** element. If a **use**
+    element is a child of a **clipPath** element, it must directly reference
+    **path**, **text** or basic shape elements. Indirect references are an
+    error.
+    """
+    elementname = 'clipPath'
+
+
+class Mask(BaseElement):
+    """
+    In SVG, you can specify that any other graphics object or **g** element
+    can be used as an alpha mask for compositing the current object into the
+    background.
+
+    A **mask** can contain any graphical elements or container elements such
+    as a **g**.
+    """
+    elementname = 'mask'
+
+    def __init__(self, start=None, size=None, **extra):
+        super(Mask, self).__init__(**extra)
+        if start is not None:
+            self['x'] = start[0]
+            self['y'] = start[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/mixins.py b/backend/venv/lib/python3.12/site-packages/svgwrite/mixins.py
new file mode 100644
index 00000000..b4f70cae
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/mixins.py
@@ -0,0 +1,298 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: mixins
+# Created: 19.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.utils import strlist
+from svgwrite.utils import is_string
+
+_horiz = {'center': 'xMid', 'left': 'xMin', 'right': 'xMax'}
+_vert  = {'middle': 'YMid', 'top': 'YMin', 'bottom':'YMax'}
+
+class ViewBox(object):
+    """ The **ViewBox** mixin provides the ability to specify that a
+    given set of graphics stretch to fit a particular container element.
+
+    The value of the **viewBox** attribute is a list of four numbers
+    **min-x**, **min-y**, **width** and **height**, separated by whitespace
+    and/or a comma, which specify a rectangle in **user space** which should
+    be mapped to the bounds of the viewport established by the given element,
+    taking into account attribute **preserveAspectRatio**.
+
+    """
+    def viewbox(self, minx=0, miny=0, width=0, height=0):
+        """ Specify a rectangle in **user space** (no units allowed) which
+        should be mapped to the bounds of the viewport established by the
+        given element.
+
+        :param number minx: left border of the viewBox
+        :param number miny: top border of the viewBox
+        :param number width: width of the viewBox
+        :param number height: height of the viewBox
+
+        """
+        self['viewBox'] = strlist( [minx, miny, width, height] )
+
+    def stretch(self):
+        """ Stretch viewBox in x and y direction to fill viewport, does not
+        preserve aspect ratio.
+        """
+        self['preserveAspectRatio'] = 'none'
+
+    def fit(self, horiz="center", vert="middle", scale="meet"):
+        """ Set the **preserveAspectRatio** attribute.
+
+        :param string horiz: horizontal alignment ``'left | center | right'``
+        :param string vert: vertical alignment ``'top | middle | bottom'``
+        :param string scale: scale method ``'meet | slice'``
+
+        ============= =======================================================
+        Scale methods Description
+        ============= =======================================================
+        ``'meet'``    preserve aspect ration and zoom to limits of viewBox
+        ``'slice'``   preserve aspect ration and viewBox touch viewport on
+                      all bounds, viewBox will extend beyond the bounds of
+                      the viewport
+        ============= =======================================================
+
+        """
+        if self.debug and scale not in ('meet', 'slice'):
+            raise ValueError("Invalid scale parameter '%s'" % scale)
+        self['preserveAspectRatio'] = "%s%s %s" % (_horiz[horiz],_vert[vert], scale)
+
+class Transform(object):
+    """ The **Transform** mixin operates on the **transform** attribute.
+    The value of the **transform** attribute is a `<transform-list>`, which
+    is defined as a list of transform definitions, which are applied in the
+    order provided. The individual transform definitions are separated by
+    whitespace and/or a comma. All coordinates are **user
+    space coordinates**.
+
+    """
+    transformname = 'transform'
+    def translate(self, tx, ty=None):
+        """
+        Specifies a translation by **tx** and **ty**. If **ty** is not provided,
+        it is assumed to be zero.
+
+        :param number tx: user coordinate - no units allowed
+        :param number ty: user coordinate - no units allowed
+        """
+        self._add_transformation("translate(%s)" % strlist( [tx, ty] ))
+
+    def rotate(self, angle, center=None):
+        """
+        Specifies a rotation by **angle** degrees about a given point.
+        If optional parameter **center** are not supplied, the rotate is
+        about the origin of the current user coordinate system.
+
+        :param number angle: rotate-angle in degrees
+        :param 2-tuple center: rotate-center as user coordinate - no units allowed
+
+        """
+        self._add_transformation("rotate(%s)" % strlist( [angle, center] ))
+
+    def scale(self, sx, sy=None):
+        """
+        Specifies a scale operation by **sx** and **sy**. If **sy** is not
+        provided, it is assumed to be equal to **sx**.
+
+        :param number sx: scalar factor x-axis, no units allowed
+        :param number sy: scalar factor y-axis, no units allowed
+
+        """
+        self._add_transformation("scale(%s)" % strlist([sx, sy]))
+
+    def skewX(self, angle):
+        """ Specifies a skew transformation along the x-axis.
+
+        :param number angle: skew-angle in degrees, no units allowed
+
+        """
+        self._add_transformation("skewX(%s)" % angle)
+
+    def skewY(self, angle):
+        """ Specifies a skew transformation along the y-axis.
+
+        :param number angle: skew-angle in degrees, no units allowed
+
+        """
+        self._add_transformation("skewY(%s)" % angle)
+
+    def matrix(self, a, b, c, d, e, f):
+        self._add_transformation("matrix(%s)" % strlist( [a, b, c, d, e, f] ))
+
+    def _add_transformation(self, new_transform):
+        old_transform = self.attribs.get(self.transformname, '')
+        self[self.transformname] = ("%s %s" % (old_transform, new_transform)).strip()
+
+
+class XLink(object):
+    """ XLink mixin """
+    def set_href(self, element):
+        """
+        Create a reference to **element**.
+
+        :param element: if element is a `string` its the **id** name of the
+          referenced element, if element is a **BaseElement** class the **id**
+          SVG Attribute is used to create the reference.
+
+        """
+        self.href = element
+        self.update_id()
+
+    def set_xlink(self, title=None, show=None, role=None, arcrole=None):
+        """ Set XLink attributes (for `href` use :meth:`set_href`).
+        """
+        if role is not None:
+            self['xlink:role'] = role
+        if arcrole is not None:
+            self['xlink:arcrole'] = arcrole
+        if title is not None:
+            self['xlink:title'] = title
+        if show is not None:
+            self['xlink:show'] = show
+
+    def update_id(self):
+        if not hasattr(self, 'href'):
+            return
+        if is_string(self.href):
+            idstr = self.href
+        else:
+            idstr = self.href.get_iri()
+        self.attribs['xlink:href'] = idstr
+
+
+class Presentation(object):
+    """
+    Helper methods to set presentation attributes.
+    """
+    def fill(self, color=None, rule=None, opacity=None):
+        """
+        Set SVG Properties **fill**, **fill-rule** and **fill-opacity**.
+
+        """
+        if color is not None:
+            if is_string(color):
+                self['fill'] = color
+            else:
+                self['fill'] = color.get_paint_server()
+        if rule is not None:
+            self['fill-rule'] = rule
+        if opacity is not None:
+            self['fill-opacity'] = opacity
+        return self
+
+    def stroke(self, color=None, width=None, opacity=None, linecap=None,
+               linejoin=None, miterlimit=None):
+        """
+        Set SVG Properties **stroke**, **stroke-width**, **stroke-opacity**,
+        **stroke-linecap** and **stroke-miterlimit**.
+
+        """
+
+        if color is not None:
+            if is_string(color):
+                self['stroke'] = color
+            else:
+                self['stroke'] = color.get_paint_server()
+        if width is not None:
+            self['stroke-width'] = width
+        if opacity is not None:
+            self['stroke-opacity'] = opacity
+        if linecap is not None:
+            self['stroke-linecap'] = linecap
+        if linejoin is not None:
+            self['stroke-linejoin'] = linejoin
+        if miterlimit is not None:
+            self['stroke-miterlimit'] = miterlimit
+        return self
+
+    def dasharray(self, dasharray=None, offset=None):
+        """
+        Set SVG Properties **stroke-dashoffset** and **stroke-dasharray**.
+
+        Where *dasharray* specify the lengths of alternating dashes and gaps as
+        <list> of <int> or <float> values or a <string> of comma and/or white
+        space separated <lengths> or <percentages>. (e.g. as <list> dasharray=[1, 0.5]
+        or as <string> dasharray='1 0.5')
+        """
+        if dasharray is not None:
+            self['stroke-dasharray'] = strlist(dasharray, ' ')
+        if offset is not None:
+            self['stroke-dashoffset'] = offset
+        return self
+
+
+class MediaGroup(object):
+    """
+    Helper methods to set media group attributes.
+
+    """
+
+    def viewport_fill(self, color=None, opacity=None):
+        """
+        Set SVG Properties **viewport-fill** and **viewport-fill-opacity**.
+
+        """
+        if color is not None:
+            self['viewport-fill'] = color
+        if opacity is not None:
+            self['viewport-fill-opacity'] = opacity
+        return self
+
+
+class Markers(object):
+    """
+    Helper methods to set marker attributes.
+
+    """
+    def set_markers(self, markers):
+        """
+        Set markers for line elements (line, polygon, polyline, path) to
+        values specified by  `markers`.
+
+        * if `markers` is a 3-tuple:
+
+          * attribute 'marker-start' = markers[0]
+          * attribute 'marker-mid' = markers[1]
+          * attribute 'marker-end' = markers[2]
+
+        * `markers` is a `string` or a `Marker` class:
+
+          * attribute 'marker' = `FuncIRI` of markers
+
+        """
+        def get_funciri(value):
+            if is_string(value):
+                # strings has to be a valid reference including the '#'
+                return 'url(%s)' % value
+            else:
+                # else create a reference to the object '#id'
+                return 'url(#%s)' % value['id']
+
+        if is_string(markers):
+            self['marker'] = get_funciri(markers)
+        else:
+            try:
+                start_marker, mid_marker, end_marker = markers
+                if start_marker:
+                    self['marker-start'] = get_funciri(start_marker)
+                if mid_marker:
+                    self['marker-mid'] = get_funciri(mid_marker)
+                if end_marker:
+                    self['marker-end'] = get_funciri(end_marker)
+            except (TypeError, KeyError):
+                self['marker'] = get_funciri(markers)
+
+
+class Clipping(object):
+    def clip_rect(self, top='auto', right='auto', bottom='auto', left='auto'):
+        """
+        Set SVG Property **clip**.
+
+        """
+        self['clip'] = "rect(%s,%s,%s,%s)" % (top, right, bottom, left)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/params.py b/backend/venv/lib/python3.12/site-packages/svgwrite/params.py
new file mode 100644
index 00000000..62c257b6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/params.py
@@ -0,0 +1,71 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman
+# Purpose: svgwrite package parameter
+# Created: 10.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.validator2 import get_validator
+
+
+class Parameter(object):
+    """
+    .. attribute:: Parameter.debug
+
+       *read/write* property
+
+       * *True* : debug mode is on, all SVG attributes are checked if valid
+         in the element context. Also the included SVG subelements will be
+         checked if they are valid for the parent SVG element.
+
+       * *False*: no validation checks will be done, but program execution is
+         faster.
+
+    .. attribute:: Parameter.profile
+
+       *read/write* property
+
+       name of the SVG profile, valid profiles are: ``'full|basic|tiny'``
+    """
+    __slots__ = ['_debug', 'validator', '_profile']
+
+    def __init__(self, debug=True, profile='full'):
+        self._debug = debug
+        self.profile = profile
+
+    def _init_validator(self):
+        self.validator = get_validator(self.profile,  self.debug)
+
+    @property
+    def debug(self):
+        return self._debug
+
+    @debug.setter
+    def debug(self, debug):
+        self._debug = debug
+        self._init_validator()
+
+    def get_version(self):
+        if self._profile == 'tiny':
+            return '1.2'
+        else:
+            return '1.1'
+
+    @property
+    def profile(self):
+        return self._profile
+
+    @profile.setter
+    def profile(self, profile):
+        """
+        :param string profile: name of the SVG profile, valid profiles are:
+        ``'full|basic|tiny'``
+
+        """
+        profile = profile.lower()
+        if profile in ('tiny', 'basic', 'full'):
+            self._profile = profile
+            self._init_validator()
+        else:
+            raise ValueError("'%s' is not a valid profile." % profile)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/path.py b/backend/venv/lib/python3.12/site-packages/svgwrite/path.py
new file mode 100644
index 00000000..150ff688
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/path.py
@@ -0,0 +1,75 @@
+#coding:utf-8
+# Author:  mozman
+# Purpose: svg path element
+# Created: 08.09.2010
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.utils import strlist
+from svgwrite.mixins import Presentation, Markers, Transform
+
+
+class Path(BaseElement, Transform, Presentation, Markers):
+    """ The <path> element represent the outline of a shape which can be filled,
+    stroked, used as a clipping path, or any combination of the three.
+
+    """
+    elementname = 'path'
+
+    def __init__(self, d=None, **extra):
+        """
+        :param `iterable` d: *coordinates*, *length* and *commands*
+        :param dict attribs: additional SVG attributes
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Path, self).__init__(**extra)
+        self.commands = []
+        self.push(d)
+        if self.debug:
+            self.validator.check_all_svg_attribute_values(self.elementname, self.attribs)
+
+    def push(self, *elements):
+        """ Push commands and coordinates onto the command stack.
+
+        :param `iterable` elements: *coordinates*, *length* and *commands*
+
+        """
+        self.commands.extend(elements)
+
+    @staticmethod
+    def arc_flags(large_arc=True, angle_dir='+'):
+        large_arc_flag = int(large_arc)
+        sweep_flag = {'+': 1, '-': 0}[angle_dir]
+        return "%d,%d" % (large_arc_flag, sweep_flag)
+
+    def push_arc(self, target, rotation, r, large_arc=True, angle_dir='+', absolute=False):
+        """ Helper function for the elliptical-arc command.
+
+        see SVG-Reference: http://www.w3.org/TR/SVG11/paths.html#PathData
+
+        :param 2-tuple target: *coordinate* of the arc end point
+        :param number rotation: x-axis-rotation of the ellipse in degrees
+        :param number|2-tuple r: radii rx, ry when r is a *2-tuple* or rx=ry=r if r is a *number*
+        :param bool large_arc: draw the arc sweep of greater than or equal to 180 degrees (**large-arc-flag**)
+        :param angle_dir: ``'+|-'`` ``'+'`` means the arc will be drawn in a "positive-angle" direction (**sweep-flag**)
+        :param bool absolute: indicates that target *coordinates* are absolute else they are relative to the current point
+
+        """
+        self.push({True: 'A', False: 'a'}[absolute])
+        if isinstance(r, (float, int)):
+            self.push(r, r)
+        else:
+            self.push(r)
+        self.push(rotation)
+        self.push(Path.arc_flags(large_arc, angle_dir))
+        self.push(target)
+
+    def get_xml(self):
+        """ Get the XML representation as `ElementTree` object.
+
+        :return: XML `ElementTree` of this object and all its subelements
+
+        """
+        self.attribs['d'] = str(strlist(self.commands, ' '))
+        return super(Path, self).get_xml()
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/pattern.py b/backend/venv/lib/python3.12/site-packages/svgwrite/pattern.py
new file mode 100644
index 00000000..239bc28c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/pattern.py
@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: pattern module
+# Created: 29.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import XLink, ViewBox, Transform, Presentation
+from svgwrite.utils import is_string
+
+class Pattern(BaseElement, XLink, ViewBox, Transform, Presentation):
+    """
+    A pattern is used to fill or stroke an object using a pre-defined graphic
+    object which can be replicated ("tiled") at fixed intervals in x and y to
+    cover the areas to be painted. Patterns are defined using a `pattern` element
+    and then referenced by properties `fill` and `stroke` on a given graphics
+    element to indicate that the given element shall be filled or stroked with
+    the referenced pattern.
+    """
+    elementname = 'pattern'
+    transformname = 'patternTransform'
+
+    def __init__(self, insert=None, size=None, inherit=None, **extra):
+        """
+        :param 2-tuple insert: base point of the pattern (**x**, **y**)
+        :param 2-tuple size: size of the pattern (**width**, **height**)
+        :param inherit: pattern inherits properties from `inherit` see: **xlink:href**
+
+        """
+        super(Pattern, self).__init__(**extra)
+        if insert is not None:
+            self['x'] = insert[0]
+            self['y'] = insert[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+        if inherit is not None:
+            if is_string(inherit):
+                self.set_href(inherit)
+            else:
+                self.set_href(inherit.get_iri())
+
+        if self.debug:
+            self.validator.check_all_svg_attribute_values(self.elementname, self.attribs)
+
+    def get_paint_server(self, default='none'):
+        """ Returns the <FuncIRI> of the gradient. """
+        return "%s %s" % (self.get_funciri(), default)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/shapes.py b/backend/venv/lib/python3.12/site-packages/svgwrite/shapes.py
new file mode 100644
index 00000000..dcff9690
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/shapes.py
@@ -0,0 +1,159 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman <me@mozman.at>
+# Purpose: svg shapes
+# Created: 08.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import Presentation, Markers, Transform
+
+
+class Line(BaseElement, Transform, Presentation, Markers):
+    """ The **line** element defines a line segment that starts at one point
+    and ends at another.
+    """
+    elementname = 'line'
+
+    def __init__(self, start=(0, 0), end=(0, 0), **extra):
+        """
+        :param 2-tuple start: start point (**x1**, **y1**)
+        :param 2-tuple end: end point (**x2**, **y2**)
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Line, self).__init__(**extra)
+        x1, y1 = start
+        x2, y2 = end
+        self['x1'] = x1
+        self['y1'] = y1
+        self['x2'] = x2
+        self['y2'] = y2
+
+
+class Rect(BaseElement, Transform, Presentation):
+    """ The **rect** element defines a rectangle which is axis-aligned with the current
+    user coordinate system. Rounded rectangles can be achieved by setting appropriate
+    values for attributes **rx** and **ry**.
+    """
+    elementname = 'rect'
+
+    def __init__(self, insert=(0, 0), size=(1, 1), rx=None, ry=None, **extra):
+        """
+        :param 2-tuple insert: insert point (**x**, **y**), left-upper point
+        :param 2-tuple size: (**width**, **height**)
+        :param <length> rx: corner x-radius
+        :param <length> ry: corner y-radius
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Rect, self).__init__(**extra)
+        x, y = insert
+        width, height = size
+        self['x'] = x
+        self['y'] = y
+        self['width'] = width
+        self['height'] = height
+        if rx is not None:
+            self['rx'] = rx
+        if ry is not None:
+            self['ry'] = ry
+
+
+class Circle(BaseElement, Transform, Presentation):
+    """ The **circle** element defines a circle based on a center point and a radius.
+    """
+    elementname = 'circle'
+
+    def __init__(self, center=(0, 0), r=1, **extra):
+        """
+        :param 2-tuple center: circle center point (**cx**, **cy**)
+        :param length r: circle-radius **r**
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Circle, self).__init__(**extra)
+        cx, cy = center
+        self['cx'] = cx
+        self['cy'] = cy
+        self['r'] = r
+
+
+class Ellipse(BaseElement, Transform, Presentation):
+    """ The **ellipse** element defines an ellipse which is axis-aligned with the
+    current user coordinate system based on a center point and two radii.
+    """
+    elementname = 'ellipse'
+
+    def __init__(self, center=(0, 0), r=(1, 1), **extra):
+        """
+        :param 2-tuple center: ellipse center point (**cx**, **cy**)
+        :param 2-tuple r: ellipse radii (**rx**, **ry**)
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Ellipse, self).__init__(**extra)
+        cx, cy = center
+        rx, ry = r
+        self['cx'] = cx
+        self['cy'] = cy
+        self['rx'] = rx
+        self['ry'] = ry
+
+
+class Polyline(BaseElement, Transform, Presentation, Markers):
+    """ The **polyline** element defines a set of connected straight line
+    segments. Typically, **polyline** elements define open shapes.
+    """
+    elementname = 'polyline'
+
+    def __init__(self, points=[], **extra):
+        """
+        :param `iterable` points: `iterable` of points (points are `2-tuples`)
+        :param extra: additional SVG attributes as keyword-arguments
+
+        """
+        super(Polyline, self).__init__(**extra)
+        self.points = list(points)
+        if self.debug:
+            for point in self.points:
+                x, y = point
+                self.validator.check_svg_type(x, 'coordinate')
+                self.validator.check_svg_type(y, 'coordinate')
+
+    def get_xml(self):
+        self.attribs['points'] = self.points_to_string(self.points)
+        return super(Polyline, self).get_xml()
+
+    def points_to_string(self, points):
+        """
+        Convert a `list` of points `2-tuples` to a `string` ``'p1x,p1y p2x,p2y ...'``.
+
+        """
+        strings = []
+        for point in points:
+            if len(point) != 2:
+                raise TypeError('got %s values, but expected 2 values.' % len(point))
+            x, y = point
+            if self.debug:
+                self.validator.check_svg_type(x, 'coordinate')
+                self.validator.check_svg_type(y, 'coordinate')
+            if self.profile == 'tiny':
+                if isinstance(x, float):
+                    x = round(x, 4)
+                if isinstance(y, float):
+                    y = round(y, 4)
+            point = "%s,%s" % (x, y)
+            strings.append(point)
+        return ' '.join(strings)
+
+
+class Polygon(Polyline):
+    """ The **polygon** element defines a closed shape consisting of a set of
+    connected straight line segments.
+
+    Same as :class:`~svgwrite.shapes.Polyline` but closed.
+    """
+    elementname = 'polygon'
+
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/solidcolor.py b/backend/venv/lib/python3.12/site-packages/svgwrite/solidcolor.py
new file mode 100644
index 00000000..148821b8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/solidcolor.py
@@ -0,0 +1,41 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: solidColor paint serve (Tiny 1.2 profile)
+# Created: 26.10.2016
+# Copyright (C) 2016, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import XLink
+
+
+class SolidColor(BaseElement, XLink):
+    """
+    The `solidColor` element is a paint server that provides a single color with opacity.
+    It can be referenced like the other paint servers (i.e. gradients).
+    The `color` parameter specifies the color that shall be used for this `solidColor` element.
+    The keyword ``"currentColor"`` can be specified in the same manner as within a <paint> specification for the `fill`
+    and `stroke` properties. The `opacity` parameter defines the opacity of the `solidColor`.
+    """
+    elementname = 'solidColor'
+
+    def __init__(self, color="currentColor", opacity=None, **extra):
+        """
+        :param color: solid color like the other paint servers (i.e. gradients).
+        :param float opacity: opacity of the solid color in the range `0.0` (fully transparent) to `1.0` (fully opaque)
+
+        """
+        super(SolidColor, self).__init__(**extra)
+        if self.profile != 'tiny':
+            raise TypeError("Paint server 'solidColor' requires the Tiny SVG profile.")
+        self['solid-color'] = color
+        if opacity is not None:
+            self['solid-opacity'] = opacity
+
+        if self.debug:
+            self.validator.check_all_svg_attribute_values(self.elementname, self.attribs)
+
+    def get_paint_server(self, default='none'):
+        """ Returns the <FuncIRI> of the gradient. """
+        return "%s %s" % (self.get_funciri(), default)
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/text.py b/backend/venv/lib/python3.12/site-packages/svgwrite/text.py
new file mode 100644
index 00000000..0b3d3b5a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/text.py
@@ -0,0 +1,218 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: text objects
+# Created: 20.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+"""
+Text that is to be rendered as part of an SVG document fragment is specified
+using the **text** element. The characters to be drawn are expressed as XML
+character data inside the **text** element.
+
+"""
+
+from svgwrite.base import BaseElement
+from svgwrite.mixins import Presentation, Transform, XLink
+from svgwrite.utils import iterflatlist, strlist, is_string
+
+
+class TSpan(BaseElement, Presentation):
+    """
+    Within a **Text** element, text and font properties
+    and the current text position can be adjusted with absolute or relative
+    coordinate values by using the **TSpan** element.
+    The characters to be drawn are expressed as XML character data inside the
+    **TSpan** element.
+
+    """
+    elementname = 'tspan'
+
+    def __init__(self, text, insert=None, x=None, y=None, dx=None, dy=None, rotate=None,
+                 **extra):
+        """
+        :param string text: **tspan** content
+        :param 2-tuple insert: The **insert** parameter is the absolute insert point
+                               of the text, don't use this parameter in combination
+                               with the **x** or the **y** parameter.
+        :param list x: list of absolute x-axis values for characters
+        :param list y: list of absolute y-axis values for characters
+        :param list dx: list of relative x-axis values for characters
+        :param list dy: list of relative y-axis values for characters
+        :param list rotate: list of rotation-values for characters (in degrees)
+
+        """
+        super(TSpan, self).__init__(**extra)
+        self.text = text
+        if insert is not None:
+            if is_string(insert):
+                raise TypeError("'insert' should be a <tuple> or a <list>  with"
+                                " at least two elements.")
+            if x or y:
+                raise ValueError("Use 'insert' and 'x' or 'y' parameter not"
+                                 " at the same time!")
+            x = [insert[0]]
+            y = [insert[1]]
+
+        if x is not None:
+            self['x'] = strlist(list(iterflatlist(x)), ' ')
+        if y is not None:
+            self['y'] = strlist(list(iterflatlist(y)), ' ')
+        if dx is not None:
+            self['dx'] = strlist(list(iterflatlist(dx)), ' ')
+        if dy is not None:
+            self['dy'] = strlist(list(iterflatlist(dy)), ' ')
+        if rotate is not None:
+            self['rotate'] = strlist(list(iterflatlist(rotate)), ' ')
+
+    def get_xml(self):
+        xml = super(TSpan, self).get_xml()
+        xml.text = str(self.text)
+        return xml
+
+
+class Text(TSpan, Transform):
+    """
+    The **Text** element defines a graphics element consisting of text.
+    The characters to be drawn are expressed as XML character data inside the
+    **Text** element.
+
+    """
+    elementname = 'text'
+
+
+class TRef(BaseElement, XLink, Presentation):
+    """
+    The textual content for a **Text** can be either character data directly
+    embedded within the <text> element or the character data content of a
+    referenced element, where the referencing is specified with a **TRef**
+    element.
+
+    """
+    elementname = 'tref'
+
+    def __init__(self, element, **extra):
+        """
+        :param element: create a reference this element, if element is a \
+                        `string` its the **id** name of the referenced element, \
+                        if element is a :class:`~svgwrite.base.BaseElement` \
+                        the **id** SVG Attribute is used to create the reference.
+
+        """
+        super(TRef, self).__init__(**extra)
+        self.set_href(element)
+
+    def get_xml(self):
+        self.update_id()  # if href is an object - 'id' - attribute may be changed!
+        return super(TRef, self).get_xml()
+
+
+class TextPath(BaseElement, XLink, Presentation):
+    """
+    In addition to text drawn in a straight line, SVG also includes the
+    ability to place text along the shape of a **path** element. To specify that
+    a block of text is to be rendered along the shape of a **path**, include
+    the given text within a **textPath** element which includes an **xlink:href**
+    attribute with a IRI reference to a **path** element.
+
+    """
+    elementname = 'textPath'
+
+    def __init__(self, path, text, startOffset=None, method='align', spacing='exact',
+                 **extra):
+        """
+        :param path: link to **path**, **id** string or **Path** object
+        :param string text: **textPath** content
+        :param number startOffset: text starts with offset from begin of path.
+        :param string method: ``align|stretch``
+        :param string spacing: ``exact|auto``
+
+        """
+        super(TextPath, self).__init__(**extra)
+        self.text = text
+        if method == 'stretch':
+            self['method'] = method
+        if spacing == 'auto':
+            self['spacing'] = spacing
+        if startOffset is not None:
+            self['startOffset'] = startOffset
+        self.set_href(path)
+
+    def get_xml(self):
+        self.update_id() # if href is an object - 'id' - attribute may be changed!
+        xml = super(TextPath, self).get_xml()
+        xml.text = str(self.text)
+        return xml
+
+
+class TBreak(BaseElement):
+    elementname = 'tbreak'
+
+    def __init__(self, **extra):
+        super(TBreak, self).__init__(**extra)
+
+    def __getitem__(self, key):
+        raise NotImplementedError("__getitem__() not supported by TBreak class.")
+
+    def __setitem__(self, key, value):
+        raise NotImplementedError("__setitem__() not supported by TBreak class.")
+
+    def add(self, element):
+        raise NotImplementedError("add() not supported by TBreak class.")
+
+
+class TextArea(BaseElement, Transform, Presentation):
+    """
+    At this time **textArea** is only available for SVG 1.2 Tiny profile.
+
+    The **textArea**  element allows simplistic wrapping of text content within a
+    given region. The `tiny` profile of SVG specifies a single rectangular region.
+    Other profiles may allow a sequence of arbitrary shapes.
+
+    Text wrapping via the **textArea** element is available as a lightweight and
+    convenient facility for simple text wrapping where a complete box model layout
+    engine is not required.
+
+    The layout of wrapped text is user agent dependent; thus, content developers
+    need to be aware that there might be different results, particularly with
+    regard to where line breaks occur.
+
+    The TextArea class wraps every text added by write() or writeline() as
+    **tspan** element.
+
+    """
+    elementname = 'textArea'
+
+    def __init__(self, text=None, insert=None, size=None, **extra):
+        super(TextArea, self).__init__(**extra)
+        if text is not None:
+            self.write(text)
+        if insert is not None:
+            self['x'] = insert[0]
+            self['y'] = insert[1]
+        if size is not None:
+            self['width'] = size[0]
+            self['height'] = size[1]
+
+    def line_increment(self, value):
+        """ Set the line-spacing to *value*. """
+        self['line-increment'] = value
+
+    def write(self, text, **extra):
+        """
+        Add text as **tspan** elements, with extra-params for the **tspan** element.
+
+        Use the '\\\\n' character for line breaks.
+        """
+        if '\n' not in text:
+            self.add(TSpan(text, **extra))
+        else:
+            lines = text.split('\n')
+            for line in lines[:-1]:
+                if line:  # no text between '\n'+
+                    self.add(TSpan(line, **extra))
+                self.add(TBreak())
+            # case "text\n" : last element is ''
+            # case "texta\ntextb : last element is 'textb'
+            if lines[-1]:
+                self.add(TSpan(lines[-1], **extra))
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/utils.py b/backend/venv/lib/python3.12/site-packages/svgwrite/utils.py
new file mode 100644
index 00000000..ab709581
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/utils.py
@@ -0,0 +1,266 @@
+#!/usr/bin/env python
+# coding:utf-8
+# Author:  mozman
+# Purpose: svg util functions and classes
+# Created: 08.09.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+"""
+
+.. autofunction:: rgb
+
+.. autofunction:: iterflatlist
+
+.. autofunction:: strlist
+
+.. autofunction:: get_unit
+
+.. autofunction:: split_coordinate
+
+.. autofunction:: split_angle
+
+.. autofunction:: rect_top_left_corner
+
+.. autofunction:: pretty_xml
+
+"""
+import base64
+from pathlib import Path
+from svgwrite.data import pattern
+
+
+def is_string(value):
+    return isinstance(value, str)
+
+
+def rgb(r=0, g=0, b=0, mode='RGB'):
+    """
+    Convert **r**, **g**, **b** values to a `string`.
+
+    :param r: red part
+    :param g: green part
+    :param b: blue part
+    :param string mode: ``'RGB | %'``
+
+    :rtype: string
+
+    ========= =============================================================
+    mode      Description
+    ========= =============================================================
+    ``'RGB'`` returns a rgb-string format: ``'rgb(r, g, b)'``
+    ``'%'``   returns percent-values as rgb-string format: ``'rgb(r%, g%, b%)'``
+    ========= =============================================================
+
+    """
+
+    def percent(value):
+        value = float(value)
+        if value < 0:
+            value = 0
+        if value > 100:
+            value = 100
+        return value
+
+    if mode.upper() == 'RGB':
+        return "rgb(%d,%d,%d)" % (int(r) & 255, int(g) & 255, int(b) & 255)
+    elif mode == "%":
+        # see http://www.w3.org/TR/SVG11/types.html#DataTypeColor
+        # percentage is an 'number' value
+        return "rgb(%d%%,%d%%,%d%%)" % (percent(r), percent(g), percent(b))
+    else:
+        raise ValueError("Invalid mode '%s'" % mode)
+
+
+def iterflatlist(values):
+    """
+    Flatten nested *values*, returns an `iterator`.
+
+    """
+    for element in values:
+        if hasattr(element, "__iter__") and not is_string(element):
+            for item in iterflatlist(element):
+                yield item
+        else:
+            yield element
+
+
+def strlist(values, seperator=","):
+    """
+    Concatenate **values** with **sepertator**, `None` values will be excluded.
+
+    :param values: `iterable` object
+    :returns: `string`
+
+    """
+    if is_string(values):
+        return values
+    else:
+        return seperator.join([str(value) for value in iterflatlist(values) if value is not None])
+
+
+def get_unit(coordinate):
+    """
+    Get the `unit` identifier of **coordinate**, if **coordinate** has a valid
+    `unit` identifier appended, else returns `None`.
+
+    """
+    if isinstance(coordinate, (int, float)):
+        return None
+    result = pattern.coordinate.match(coordinate)
+    if result:
+        return result.group(3)
+    else:
+        raise ValueError("Invalid format: '%s'" % coordinate)
+
+
+def split_coordinate(coordinate):
+    """
+    Split coordinate into `<number>` and 'unit` identifier.
+
+    :returns: <2-tuple> (number, unit-identifier) or (number, None) if no unit-identifier
+      is present or coordinate is an int or float.
+
+    """
+    if isinstance(coordinate, (int, float)):
+        return (float(coordinate), None)
+    result = pattern.coordinate.match(coordinate)
+    if result:
+        return (float(result.group(1)), result.group(3))
+    else:
+        raise ValueError("Invalid format: '%s'" % coordinate)
+
+
+def split_angle(angle):
+    """
+    Split angle into `<number>` and `<angle>` identifier.
+
+    :returns: <2-tuple> (number, angle-identifier) or (number, None) if no angle-identifier
+      is present or angle is an int or float.
+
+    """
+
+    if isinstance(angle, (int, float)):
+        return (float(angle), None)
+    result = pattern.angle.match(angle)
+    if result:
+        return (float(result.group(1)), result.group(3))
+    else:
+        raise ValueError("Invalid format: '%s'" % angle)
+
+
+def rect_top_left_corner(insert, size, pos='top-left'):
+    """
+    Calculate top-left corner of a rectangle.
+
+    **insert** and **size** must have the same units.
+
+    :param 2-tuple insert: insert point
+    :param 2-tuple size: (width, height)
+    :param string pos: insert position ``'vert-horiz'``
+    :return: ``'top-left'`` corner of the rect
+    :rtype: 2-tuple
+
+    ========== ==============================
+    pos        valid values
+    ========== ==============================
+    **vert**   ``'top | middle | bottom'``
+    **horiz**  ``'left'|'center'|'right'``
+    ========== ==============================
+    """
+    vert, horiz = pos.lower().split('-')
+    x, xunit = split_coordinate(insert[0])
+    y, yunit = split_coordinate(insert[1])
+    width, wunit = split_coordinate(size[0])
+    height, hunit = split_coordinate(size[1])
+
+    if xunit != wunit:
+        raise ValueError("x-coordinate and width has to have the same unit")
+    if yunit != hunit:
+        raise ValueError("y-coordinate and height has to have the same unit")
+
+    if horiz == 'center':
+        x = x - width / 2.
+    elif horiz == 'right':
+        x = x - width
+    elif horiz != 'left':
+        raise ValueError("Invalid horizontal position: '%s'" % horiz)
+
+    if vert == 'middle':
+        y = y - height / 2.
+    elif vert == 'bottom':
+        y = y - height
+    elif vert != 'top':
+        raise ValueError("Invalid vertical position: '%s'" % vert)
+
+    if xunit:
+        x = "%s%s" % (x, xunit)
+    if yunit:
+        y = "%s%s" % (y, yunit)
+    return x, y
+
+
+class AutoID(object):
+    _nextid = 1
+
+    def __init__(self, value=None):
+        self._set_value(value)
+
+    @classmethod
+    def _set_value(cls, value=None):
+        if value is not None:
+            cls._nextid = value
+
+    @classmethod
+    def next_id(cls, value=None):
+        cls._set_value(value)
+        retval = "id%d" % cls._nextid
+        cls._nextid += 1
+        return retval
+
+
+def pretty_xml(xml_string, indent=2):
+    """
+    Create human readable XML string.
+
+    :param xml_string: input xml string without line breaks and indentation
+    :indent int: how much to indent, by default 2 spaces
+    :return: xml_string with linebreaks and indentation
+
+    """
+    import xml.dom.minidom as minidom
+
+    # check for empty string, len check avoids unnecessary string manipulation with large XML strings
+    if len(xml_string) < 20 and xml_string.strip() == "":
+        return ""
+    xml_tree = minidom.parseString(xml_string)
+    lines = xml_tree.toprettyxml(indent=' ' * indent).split('\n')
+    # remove 1. line = xml declaration
+    return '\n'.join(lines[1:])
+
+
+FONT_MIMETYPES = {
+    'ttf': "application/x-font-ttf",
+    'otf': "application/x-font-opentype",
+    'woff': "application/font-woff",
+    'woff2': "application/font-woff2",
+    'eot': "application/vnd.ms-fontobject",
+    'sfnt': "application/font-sfnt",
+}
+
+
+def font_mimetype(name):
+    return FONT_MIMETYPES[Path(name.lower()).suffix[1:]]
+
+
+def base64_data(data, mimetype):
+    data = base64.b64encode(data).decode()
+    return "data:{mimetype};charset=utf-8;base64,{data}".format(mimetype=mimetype, data=data)
+
+
+def find_first_url(text):
+    import re
+    result = re.findall(r"url\((.*?)\)", text)
+    if result:
+        return result[0]
+    else:
+        return None
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/validator2.py b/backend/venv/lib/python3.12/site-packages/svgwrite/validator2.py
new file mode 100644
index 00000000..0121ff08
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/validator2.py
@@ -0,0 +1,170 @@
+#!/usr/bin/env python
+#coding:utf-8
+# Author:  mozman --<mozman@gmx.at>
+# Purpose: validator2 module - new validator module
+# Created: 01.10.2010
+# Copyright (C) 2010, Manfred Moitzi
+# License: MIT License
+
+from svgwrite.data import full11
+from svgwrite.data import tiny12
+from svgwrite.data import pattern
+
+validator_cache = {}
+
+
+def cache_key(profile, debug):
+    return str(profile) + str(debug)
+
+
+def get_validator(profile, debug=True):
+    """ Validator factory """
+    try:
+        return validator_cache[cache_key(profile, debug)]
+    except KeyError:
+        if profile == 'tiny':
+            validator = Tiny12Validator(debug)
+        elif profile in ('full', 'basic', 'none'):
+            validator = Full11Validator(debug)
+        else:
+            raise ValueError("Unsupported profile: '%s'" % profile)
+        validator_cache[cache_key(profile, debug)] = validator
+        return validator
+
+
+class Tiny12Validator(object):
+    profilename = "Tiny 1.2"
+
+    def __init__(self, debug=True):
+        self.debug = debug
+        self.attributes = tiny12.attributes
+        self.elements = tiny12.elements
+        self.typechecker = tiny12.TypeChecker()
+
+    def check_all_svg_attribute_values(self, elementname, attributes):
+        """
+        Check if attributes are valid for object 'elementname' and all svg
+        attributes have valid types and values.
+
+        Raises ValueError.
+        """
+        for attributename, value in attributes.items():
+            self.check_svg_attribute_value(elementname, attributename, value)
+
+    def check_svg_attribute_value(self, elementname, attributename, value):
+        """
+        Check if 'attributename' is valid for object 'elementname' and 'value'
+        is a valid svg type and value.
+
+        Raises ValueError.
+        """
+        self._check_valid_svg_attribute_name(elementname, attributename)
+        self._check_svg_value(elementname, attributename, value)
+
+    def _check_svg_value(self, elementname, attributename, value):
+        """
+        Checks if 'value' is a valid svg-type for svg-attribute
+        'attributename' at svg-element 'elementname'.
+
+        Raises TypeError.
+        """
+        attribute = self.attributes[attributename]
+        # check if 'value' match a valid datatype
+        for typename in attribute.get_types(elementname):
+            if self.typechecker.check(typename, value):
+                return
+        # check if 'value' is a valid constant
+        valuestr = str(value)
+        if not valuestr in attribute.get_const(elementname):
+            raise TypeError("'%s' is not a valid value for attribute '%s' at svg-element <%s>." % (value, attributename, elementname))
+
+    def _check_valid_svg_attribute_name(self, elementname, attributename):
+        """ Check if 'attributename' is a valid svg-attribute for svg-element
+        'elementname'.
+
+        Raises ValueError.
+        """
+        if not self.is_valid_svg_attribute(elementname, attributename):
+            raise ValueError("Invalid attribute '%s' for svg-element <%s>." % (attributename, elementname))
+
+    def _get_element(self, elementname):
+        try:
+            return self.elements[elementname]
+        except KeyError:
+            raise KeyError("<%s> is not valid for selected profile: '%s'." % (elementname, self.profilename))
+
+    def check_svg_type(self, value, typename='string'):
+        """
+        Check if 'value' matches svg type 'typename'.
+
+        Raises TypeError.
+        """
+        if self.typechecker.check(typename, value):
+            return value
+        else:
+            raise TypeError("%s is not of type '%s'." % (value, typename))
+
+    def is_valid_svg_type(self, value, typename):
+        return self.typechecker.check(typename, value)
+
+    def is_valid_elementname(self, elementname):
+        """ True if 'elementname' is a valid svg-element name. """
+        return elementname in self.elements
+
+    def is_valid_svg_attribute(self, elementname, attributename):
+        """ True if 'attributename' is a valid svg-attribute for svg-element
+        'elementname'.
+        """
+        element = self._get_element(elementname)
+        return attributename in element.valid_attributes
+
+    def is_valid_children(self, elementname, childrenname):
+        """ True if svg-element 'childrenname' is a valid children of
+        svg-element 'elementname'.
+        """
+        element = self._get_element(elementname)
+        return childrenname in element.valid_children
+
+    def check_valid_children(self, elementname, childrenname):
+        """ Checks if svg-element 'childrenname' is a valid children of
+        svg-element 'elementname'.
+
+        Raises ValueError.
+        """
+        if not self.is_valid_children(elementname, childrenname):
+            raise ValueError("Invalid children '%s' for svg-element <%s>." % (childrenname, elementname))
+
+    def get_coordinate(self, value):
+        """ Split value in (number, unit) if value has an unit or (number, None).
+
+        Raises ValueError.
+        """
+
+        if value is None:
+            raise TypeError("Invalid type 'None'.")
+        if isinstance(value, (int, float)):
+            result = (value, None)
+        else:
+            result = pattern.coordinate.match(value.strip())
+            if result:
+                number, tmp, unit = result.groups()
+                number = float(number)
+            else:
+                raise ValueError("'%s' is not a valid svg-coordinate." % value)
+            result = (number, unit)
+        if self.typechecker.is_number(result[0]):
+            return result
+        else:
+            version = "SVG %s %s" % self.typechecker.get_version()
+            raise ValueError("%s is not a valid number for: %s." % (value, version))
+    get_length = get_coordinate
+
+
+class Full11Validator(Tiny12Validator):
+    profilename = "Full 1.1"
+
+    def __init__(self, debug=True):
+        self.debug = debug
+        self.attributes = full11.attributes
+        self.elements = full11.elements
+        self.typechecker = full11.TypeChecker()
diff --git a/backend/venv/lib/python3.12/site-packages/svgwrite/version.py b/backend/venv/lib/python3.12/site-packages/svgwrite/version.py
new file mode 100644
index 00000000..7ff61f27
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/svgwrite/version.py
@@ -0,0 +1,31 @@
+# 2018-11-29: future consistent version numbers
+# ---------------------------------------------
+#
+# version scheme for version: (major, minor, micro, release_level)
+#
+# major:
+#   0 .. not all planned features done
+#   1 .. all features available
+#   2 .. if significant API change (2, 3, ...)
+#
+# minor:
+#   changes with new features or minor API changes
+#
+# micro:
+#   changes with bug fixes, maybe also minor changes
+#
+# release_state:
+#   a .. alpha: adding new features - non public development state
+#   b .. beta: testing new features - public development state
+#   rc .. release candidate: testing release - public testing
+#   release: public release
+# examples
+#
+# major pre release alpha 2: VERSION = "0.9a2"; version = (0, 9, 0, 'a2')
+# major release candidate 0: VERSION = "0.9rc0"; version = (0, 9, 0, 'rc0')
+# major release: VERSION = "0.9"; version = (0, 9, 0, 'release')
+# 1. bug fix release beta0: VERSION = "0.9.1b0"; version = (0, 9, 1, 'b0')
+# 2. bug fix release: VERSION = "0.9.2"; version = (0, 9, 2, 'release')
+
+version = (1, 4, 3, 'release')
+__version__ = "1.4.3"
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/LICENSE b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/LICENSE
new file mode 100644
index 00000000..f26bcf4d
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/LICENSE
@@ -0,0 +1,279 @@
+A. HISTORY OF THE SOFTWARE
+==========================
+
+Python was created in the early 1990s by Guido van Rossum at Stichting
+Mathematisch Centrum (CWI, see https://www.cwi.nl) in the Netherlands
+as a successor of a language called ABC.  Guido remains Python's
+principal author, although it includes many contributions from others.
+
+In 1995, Guido continued his work on Python at the Corporation for
+National Research Initiatives (CNRI, see https://www.cnri.reston.va.us)
+in Reston, Virginia where he released several versions of the
+software.
+
+In May 2000, Guido and the Python core development team moved to
+BeOpen.com to form the BeOpen PythonLabs team.  In October of the same
+year, the PythonLabs team moved to Digital Creations, which became
+Zope Corporation.  In 2001, the Python Software Foundation (PSF, see
+https://www.python.org/psf/) was formed, a non-profit organization
+created specifically to own Python-related Intellectual Property.
+Zope Corporation was a sponsoring member of the PSF.
+
+All Python releases are Open Source (see https://opensource.org for
+the Open Source Definition).  Historically, most, but not all, Python
+releases have also been GPL-compatible; the table below summarizes
+the various releases.
+
+    Release         Derived     Year        Owner       GPL-
+                    from                                compatible? (1)
+
+    0.9.0 thru 1.2              1991-1995   CWI         yes
+    1.3 thru 1.5.2  1.2         1995-1999   CNRI        yes
+    1.6             1.5.2       2000        CNRI        no
+    2.0             1.6         2000        BeOpen.com  no
+    1.6.1           1.6         2001        CNRI        yes (2)
+    2.1             2.0+1.6.1   2001        PSF         no
+    2.0.1           2.0+1.6.1   2001        PSF         yes
+    2.1.1           2.1+2.0.1   2001        PSF         yes
+    2.1.2           2.1.1       2002        PSF         yes
+    2.1.3           2.1.2       2002        PSF         yes
+    2.2 and above   2.1.1       2001-now    PSF         yes
+
+Footnotes:
+
+(1) GPL-compatible doesn't mean that we're distributing Python under
+    the GPL.  All Python licenses, unlike the GPL, let you distribute
+    a modified version without making your changes open source.  The
+    GPL-compatible licenses make it possible to combine Python with
+    other software that is released under the GPL; the others don't.
+
+(2) According to Richard Stallman, 1.6.1 is not GPL-compatible,
+    because its license has a choice of law clause.  According to
+    CNRI, however, Stallman's lawyer has told CNRI's lawyer that 1.6.1
+    is "not incompatible" with the GPL.
+
+Thanks to the many outside volunteers who have worked under Guido's
+direction to make these releases possible.
+
+
+B. TERMS AND CONDITIONS FOR ACCESSING OR OTHERWISE USING PYTHON
+===============================================================
+
+Python software and documentation are licensed under the
+Python Software Foundation License Version 2.
+
+Starting with Python 3.8.6, examples, recipes, and other code in
+the documentation are dual licensed under the PSF License Version 2
+and the Zero-Clause BSD license.
+
+Some software incorporated into Python is under different licenses.
+The licenses are listed with code falling under that license.
+
+
+PYTHON SOFTWARE FOUNDATION LICENSE VERSION 2
+--------------------------------------------
+
+1. This LICENSE AGREEMENT is between the Python Software Foundation
+("PSF"), and the Individual or Organization ("Licensee") accessing and
+otherwise using this software ("Python") in source or binary form and
+its associated documentation.
+
+2. Subject to the terms and conditions of this License Agreement, PSF hereby
+grants Licensee a nonexclusive, royalty-free, world-wide license to reproduce,
+analyze, test, perform and/or display publicly, prepare derivative works,
+distribute, and otherwise use Python alone or in any derivative version,
+provided, however, that PSF's License Agreement and PSF's notice of copyright,
+i.e., "Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
+2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023 Python Software Foundation;
+All Rights Reserved" are retained in Python alone or in any derivative version
+prepared by Licensee.
+
+3. In the event Licensee prepares a derivative work that is based on
+or incorporates Python or any part thereof, and wants to make
+the derivative work available to others as provided herein, then
+Licensee hereby agrees to include in any such work a brief summary of
+the changes made to Python.
+
+4. PSF is making Python available to Licensee on an "AS IS"
+basis.  PSF MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, PSF MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF PYTHON WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+5. PSF SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF PYTHON
+FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS AS
+A RESULT OF MODIFYING, DISTRIBUTING, OR OTHERWISE USING PYTHON,
+OR ANY DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+6. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+7. Nothing in this License Agreement shall be deemed to create any
+relationship of agency, partnership, or joint venture between PSF and
+Licensee.  This License Agreement does not grant permission to use PSF
+trademarks or trade name in a trademark sense to endorse or promote
+products or services of Licensee, or any third party.
+
+8. By copying, installing or otherwise using Python, Licensee
+agrees to be bound by the terms and conditions of this License
+Agreement.
+
+
+BEOPEN.COM LICENSE AGREEMENT FOR PYTHON 2.0
+-------------------------------------------
+
+BEOPEN PYTHON OPEN SOURCE LICENSE AGREEMENT VERSION 1
+
+1. This LICENSE AGREEMENT is between BeOpen.com ("BeOpen"), having an
+office at 160 Saratoga Avenue, Santa Clara, CA 95051, and the
+Individual or Organization ("Licensee") accessing and otherwise using
+this software in source or binary form and its associated
+documentation ("the Software").
+
+2. Subject to the terms and conditions of this BeOpen Python License
+Agreement, BeOpen hereby grants Licensee a non-exclusive,
+royalty-free, world-wide license to reproduce, analyze, test, perform
+and/or display publicly, prepare derivative works, distribute, and
+otherwise use the Software alone or in any derivative version,
+provided, however, that the BeOpen Python License is retained in the
+Software, alone or in any derivative version prepared by Licensee.
+
+3. BeOpen is making the Software available to Licensee on an "AS IS"
+basis.  BEOPEN MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, BEOPEN MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF THE SOFTWARE WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+4. BEOPEN SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF THE
+SOFTWARE FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS
+AS A RESULT OF USING, MODIFYING OR DISTRIBUTING THE SOFTWARE, OR ANY
+DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+5. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+6. This License Agreement shall be governed by and interpreted in all
+respects by the law of the State of California, excluding conflict of
+law provisions.  Nothing in this License Agreement shall be deemed to
+create any relationship of agency, partnership, or joint venture
+between BeOpen and Licensee.  This License Agreement does not grant
+permission to use BeOpen trademarks or trade names in a trademark
+sense to endorse or promote products or services of Licensee, or any
+third party.  As an exception, the "BeOpen Python" logos available at
+http://www.pythonlabs.com/logos.html may be used according to the
+permissions granted on that web page.
+
+7. By copying, installing or otherwise using the software, Licensee
+agrees to be bound by the terms and conditions of this License
+Agreement.
+
+
+CNRI LICENSE AGREEMENT FOR PYTHON 1.6.1
+---------------------------------------
+
+1. This LICENSE AGREEMENT is between the Corporation for National
+Research Initiatives, having an office at 1895 Preston White Drive,
+Reston, VA 20191 ("CNRI"), and the Individual or Organization
+("Licensee") accessing and otherwise using Python 1.6.1 software in
+source or binary form and its associated documentation.
+
+2. Subject to the terms and conditions of this License Agreement, CNRI
+hereby grants Licensee a nonexclusive, royalty-free, world-wide
+license to reproduce, analyze, test, perform and/or display publicly,
+prepare derivative works, distribute, and otherwise use Python 1.6.1
+alone or in any derivative version, provided, however, that CNRI's
+License Agreement and CNRI's notice of copyright, i.e., "Copyright (c)
+1995-2001 Corporation for National Research Initiatives; All Rights
+Reserved" are retained in Python 1.6.1 alone or in any derivative
+version prepared by Licensee.  Alternately, in lieu of CNRI's License
+Agreement, Licensee may substitute the following text (omitting the
+quotes): "Python 1.6.1 is made available subject to the terms and
+conditions in CNRI's License Agreement.  This Agreement together with
+Python 1.6.1 may be located on the internet using the following
+unique, persistent identifier (known as a handle): 1895.22/1013.  This
+Agreement may also be obtained from a proxy server on the internet
+using the following URL: http://hdl.handle.net/1895.22/1013".
+
+3. In the event Licensee prepares a derivative work that is based on
+or incorporates Python 1.6.1 or any part thereof, and wants to make
+the derivative work available to others as provided herein, then
+Licensee hereby agrees to include in any such work a brief summary of
+the changes made to Python 1.6.1.
+
+4. CNRI is making Python 1.6.1 available to Licensee on an "AS IS"
+basis.  CNRI MAKES NO REPRESENTATIONS OR WARRANTIES, EXPRESS OR
+IMPLIED.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, CNRI MAKES NO AND
+DISCLAIMS ANY REPRESENTATION OR WARRANTY OF MERCHANTABILITY OR FITNESS
+FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF PYTHON 1.6.1 WILL NOT
+INFRINGE ANY THIRD PARTY RIGHTS.
+
+5. CNRI SHALL NOT BE LIABLE TO LICENSEE OR ANY OTHER USERS OF PYTHON
+1.6.1 FOR ANY INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES OR LOSS AS
+A RESULT OF MODIFYING, DISTRIBUTING, OR OTHERWISE USING PYTHON 1.6.1,
+OR ANY DERIVATIVE THEREOF, EVEN IF ADVISED OF THE POSSIBILITY THEREOF.
+
+6. This License Agreement will automatically terminate upon a material
+breach of its terms and conditions.
+
+7. This License Agreement shall be governed by the federal
+intellectual property law of the United States, including without
+limitation the federal copyright law, and, to the extent such
+U.S. federal law does not apply, by the law of the Commonwealth of
+Virginia, excluding Virginia's conflict of law provisions.
+Notwithstanding the foregoing, with regard to derivative works based
+on Python 1.6.1 that incorporate non-separable material that was
+previously distributed under the GNU General Public License (GPL), the
+law of the Commonwealth of Virginia shall govern this License
+Agreement only as to issues arising under or with respect to
+Paragraphs 4, 5, and 7 of this License Agreement.  Nothing in this
+License Agreement shall be deemed to create any relationship of
+agency, partnership, or joint venture between CNRI and Licensee.  This
+License Agreement does not grant permission to use CNRI trademarks or
+trade name in a trademark sense to endorse or promote products or
+services of Licensee, or any third party.
+
+8. By clicking on the "ACCEPT" button where indicated, or by copying,
+installing or otherwise using Python 1.6.1, Licensee agrees to be
+bound by the terms and conditions of this License Agreement.
+
+        ACCEPT
+
+
+CWI LICENSE AGREEMENT FOR PYTHON 0.9.0 THROUGH 1.2
+--------------------------------------------------
+
+Copyright (c) 1991 - 1995, Stichting Mathematisch Centrum Amsterdam,
+The Netherlands.  All rights reserved.
+
+Permission to use, copy, modify, and distribute this software and its
+documentation for any purpose and without fee is hereby granted,
+provided that the above copyright notice appear in all copies and that
+both that copyright notice and this permission notice appear in
+supporting documentation, and that the name of Stichting Mathematisch
+Centrum or CWI not be used in advertising or publicity pertaining to
+distribution of the software without specific, written prior
+permission.
+
+STICHTING MATHEMATISCH CENTRUM DISCLAIMS ALL WARRANTIES WITH REGARD TO
+THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
+FITNESS, IN NO EVENT SHALL STICHTING MATHEMATISCH CENTRUM BE LIABLE
+FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
+OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+
+ZERO-CLAUSE BSD LICENSE FOR CODE IN THE PYTHON DOCUMENTATION
+----------------------------------------------------------------------
+
+Permission to use, copy, modify, and/or distribute this software for any
+purpose with or without fee is hereby granted.
+
+THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH
+REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
+AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
+INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+PERFORMANCE OF THIS SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/METADATA
new file mode 100644
index 00000000..f15e2b38
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/METADATA
@@ -0,0 +1,67 @@
+Metadata-Version: 2.1
+Name: typing_extensions
+Version: 4.12.2
+Summary: Backported and Experimental Type Hints for Python 3.8+
+Keywords: annotations,backport,checker,checking,function,hinting,hints,type,typechecking,typehinting,typehints,typing
+Author-email: "Guido van Rossum, Jukka Lehtosalo, Łukasz Langa, Michael Lee" <levkivskyi@gmail.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Python Software Foundation License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development
+Project-URL: Bug Tracker, https://github.com/python/typing_extensions/issues
+Project-URL: Changes, https://github.com/python/typing_extensions/blob/main/CHANGELOG.md
+Project-URL: Documentation, https://typing-extensions.readthedocs.io/
+Project-URL: Home, https://github.com/python/typing_extensions
+Project-URL: Q & A, https://github.com/python/typing/discussions
+Project-URL: Repository, https://github.com/python/typing_extensions
+
+# Typing Extensions
+
+[![Chat at https://gitter.im/python/typing](https://badges.gitter.im/python/typing.svg)](https://gitter.im/python/typing)
+
+[Documentation](https://typing-extensions.readthedocs.io/en/latest/#) –
+[PyPI](https://pypi.org/project/typing-extensions/)
+
+## Overview
+
+The `typing_extensions` module serves two related purposes:
+
+- Enable use of new type system features on older Python versions. For example,
+  `typing.TypeGuard` is new in Python 3.10, but `typing_extensions` allows
+  users on previous Python versions to use it too.
+- Enable experimentation with new type system PEPs before they are accepted and
+  added to the `typing` module.
+
+`typing_extensions` is treated specially by static type checkers such as
+mypy and pyright. Objects defined in `typing_extensions` are treated the same
+way as equivalent forms in `typing`.
+
+`typing_extensions` uses
+[Semantic Versioning](https://semver.org/). The
+major version will be incremented only for backwards-incompatible changes.
+Therefore, it's safe to depend
+on `typing_extensions` like this: `typing_extensions >=x.y, <(x+1)`,
+where `x.y` is the first version that includes all features you need.
+
+## Included items
+
+See [the documentation](https://typing-extensions.readthedocs.io/en/latest/#) for a
+complete listing of module contents.
+
+## Contributing
+
+See [CONTRIBUTING.md](https://github.com/python/typing_extensions/blob/main/CONTRIBUTING.md)
+for how to contribute to `typing_extensions`.
+
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/RECORD
new file mode 100644
index 00000000..6ba81a0c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/RECORD
@@ -0,0 +1,8 @@
+__pycache__/typing_extensions.cpython-312.pyc,,
+typing_extensions-4.12.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+typing_extensions-4.12.2.dist-info/LICENSE,sha256=Oy-B_iHRgcSZxZolbI4ZaEVdZonSaaqFNzv7avQdo78,13936
+typing_extensions-4.12.2.dist-info/METADATA,sha256=BeUQIa8cnYbrjWx-N8TOznM9UGW5Gm2DicVpDtRA8W0,3018
+typing_extensions-4.12.2.dist-info/RECORD,,
+typing_extensions-4.12.2.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+typing_extensions-4.12.2.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+typing_extensions.py,sha256=gwekpyG9DVG3lxWKX4ni8u7nk3We5slG98mA9F3DJQw,134451
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/WHEEL
new file mode 100644
index 00000000..3b5e64b5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions-4.12.2.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.9.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/typing_extensions.py b/backend/venv/lib/python3.12/site-packages/typing_extensions.py
new file mode 100644
index 00000000..dec429ca
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/typing_extensions.py
@@ -0,0 +1,3641 @@
+import abc
+import collections
+import collections.abc
+import contextlib
+import functools
+import inspect
+import operator
+import sys
+import types as _types
+import typing
+import warnings
+
+__all__ = [
+    # Super-special typing primitives.
+    'Any',
+    'ClassVar',
+    'Concatenate',
+    'Final',
+    'LiteralString',
+    'ParamSpec',
+    'ParamSpecArgs',
+    'ParamSpecKwargs',
+    'Self',
+    'Type',
+    'TypeVar',
+    'TypeVarTuple',
+    'Unpack',
+
+    # ABCs (from collections.abc).
+    'Awaitable',
+    'AsyncIterator',
+    'AsyncIterable',
+    'Coroutine',
+    'AsyncGenerator',
+    'AsyncContextManager',
+    'Buffer',
+    'ChainMap',
+
+    # Concrete collection types.
+    'ContextManager',
+    'Counter',
+    'Deque',
+    'DefaultDict',
+    'NamedTuple',
+    'OrderedDict',
+    'TypedDict',
+
+    # Structural checks, a.k.a. protocols.
+    'SupportsAbs',
+    'SupportsBytes',
+    'SupportsComplex',
+    'SupportsFloat',
+    'SupportsIndex',
+    'SupportsInt',
+    'SupportsRound',
+
+    # One-off things.
+    'Annotated',
+    'assert_never',
+    'assert_type',
+    'clear_overloads',
+    'dataclass_transform',
+    'deprecated',
+    'Doc',
+    'get_overloads',
+    'final',
+    'get_args',
+    'get_origin',
+    'get_original_bases',
+    'get_protocol_members',
+    'get_type_hints',
+    'IntVar',
+    'is_protocol',
+    'is_typeddict',
+    'Literal',
+    'NewType',
+    'overload',
+    'override',
+    'Protocol',
+    'reveal_type',
+    'runtime',
+    'runtime_checkable',
+    'Text',
+    'TypeAlias',
+    'TypeAliasType',
+    'TypeGuard',
+    'TypeIs',
+    'TYPE_CHECKING',
+    'Never',
+    'NoReturn',
+    'ReadOnly',
+    'Required',
+    'NotRequired',
+
+    # Pure aliases, have always been in typing
+    'AbstractSet',
+    'AnyStr',
+    'BinaryIO',
+    'Callable',
+    'Collection',
+    'Container',
+    'Dict',
+    'ForwardRef',
+    'FrozenSet',
+    'Generator',
+    'Generic',
+    'Hashable',
+    'IO',
+    'ItemsView',
+    'Iterable',
+    'Iterator',
+    'KeysView',
+    'List',
+    'Mapping',
+    'MappingView',
+    'Match',
+    'MutableMapping',
+    'MutableSequence',
+    'MutableSet',
+    'NoDefault',
+    'Optional',
+    'Pattern',
+    'Reversible',
+    'Sequence',
+    'Set',
+    'Sized',
+    'TextIO',
+    'Tuple',
+    'Union',
+    'ValuesView',
+    'cast',
+    'no_type_check',
+    'no_type_check_decorator',
+]
+
+# for backward compatibility
+PEP_560 = True
+GenericMeta = type
+_PEP_696_IMPLEMENTED = sys.version_info >= (3, 13, 0, "beta")
+
+# The functions below are modified copies of typing internal helpers.
+# They are needed by _ProtocolMeta and they provide support for PEP 646.
+
+
+class _Sentinel:
+    def __repr__(self):
+        return "<sentinel>"
+
+
+_marker = _Sentinel()
+
+
+if sys.version_info >= (3, 10):
+    def _should_collect_from_parameters(t):
+        return isinstance(
+            t, (typing._GenericAlias, _types.GenericAlias, _types.UnionType)
+        )
+elif sys.version_info >= (3, 9):
+    def _should_collect_from_parameters(t):
+        return isinstance(t, (typing._GenericAlias, _types.GenericAlias))
+else:
+    def _should_collect_from_parameters(t):
+        return isinstance(t, typing._GenericAlias) and not t._special
+
+
+NoReturn = typing.NoReturn
+
+# Some unconstrained type variables.  These are used by the container types.
+# (These are not for export.)
+T = typing.TypeVar('T')  # Any type.
+KT = typing.TypeVar('KT')  # Key type.
+VT = typing.TypeVar('VT')  # Value type.
+T_co = typing.TypeVar('T_co', covariant=True)  # Any type covariant containers.
+T_contra = typing.TypeVar('T_contra', contravariant=True)  # Ditto contravariant.
+
+
+if sys.version_info >= (3, 11):
+    from typing import Any
+else:
+
+    class _AnyMeta(type):
+        def __instancecheck__(self, obj):
+            if self is Any:
+                raise TypeError("typing_extensions.Any cannot be used with isinstance()")
+            return super().__instancecheck__(obj)
+
+        def __repr__(self):
+            if self is Any:
+                return "typing_extensions.Any"
+            return super().__repr__()
+
+    class Any(metaclass=_AnyMeta):
+        """Special type indicating an unconstrained type.
+        - Any is compatible with every type.
+        - Any assumed to have all methods.
+        - All values assumed to be instances of Any.
+        Note that all the above statements are true from the point of view of
+        static type checkers. At runtime, Any should not be used with instance
+        checks.
+        """
+        def __new__(cls, *args, **kwargs):
+            if cls is Any:
+                raise TypeError("Any cannot be instantiated")
+            return super().__new__(cls, *args, **kwargs)
+
+
+ClassVar = typing.ClassVar
+
+
+class _ExtensionsSpecialForm(typing._SpecialForm, _root=True):
+    def __repr__(self):
+        return 'typing_extensions.' + self._name
+
+
+Final = typing.Final
+
+if sys.version_info >= (3, 11):
+    final = typing.final
+else:
+    # @final exists in 3.8+, but we backport it for all versions
+    # before 3.11 to keep support for the __final__ attribute.
+    # See https://bugs.python.org/issue46342
+    def final(f):
+        """This decorator can be used to indicate to type checkers that
+        the decorated method cannot be overridden, and decorated class
+        cannot be subclassed. For example:
+
+            class Base:
+                @final
+                def done(self) -> None:
+                    ...
+            class Sub(Base):
+                def done(self) -> None:  # Error reported by type checker
+                    ...
+            @final
+            class Leaf:
+                ...
+            class Other(Leaf):  # Error reported by type checker
+                ...
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__final__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+        """
+        try:
+            f.__final__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return f
+
+
+def IntVar(name):
+    return typing.TypeVar(name)
+
+
+# A Literal bug was fixed in 3.11.0, 3.10.1 and 3.9.8
+if sys.version_info >= (3, 10, 1):
+    Literal = typing.Literal
+else:
+    def _flatten_literal_params(parameters):
+        """An internal helper for Literal creation: flatten Literals among parameters"""
+        params = []
+        for p in parameters:
+            if isinstance(p, _LiteralGenericAlias):
+                params.extend(p.__args__)
+            else:
+                params.append(p)
+        return tuple(params)
+
+    def _value_and_type_iter(params):
+        for p in params:
+            yield p, type(p)
+
+    class _LiteralGenericAlias(typing._GenericAlias, _root=True):
+        def __eq__(self, other):
+            if not isinstance(other, _LiteralGenericAlias):
+                return NotImplemented
+            these_args_deduped = set(_value_and_type_iter(self.__args__))
+            other_args_deduped = set(_value_and_type_iter(other.__args__))
+            return these_args_deduped == other_args_deduped
+
+        def __hash__(self):
+            return hash(frozenset(_value_and_type_iter(self.__args__)))
+
+    class _LiteralForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, doc: str):
+            self._name = 'Literal'
+            self._doc = self.__doc__ = doc
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+
+            parameters = _flatten_literal_params(parameters)
+
+            val_type_pairs = list(_value_and_type_iter(parameters))
+            try:
+                deduped_pairs = set(val_type_pairs)
+            except TypeError:
+                # unhashable parameters
+                pass
+            else:
+                # similar logic to typing._deduplicate on Python 3.9+
+                if len(deduped_pairs) < len(val_type_pairs):
+                    new_parameters = []
+                    for pair in val_type_pairs:
+                        if pair in deduped_pairs:
+                            new_parameters.append(pair[0])
+                            deduped_pairs.remove(pair)
+                    assert not deduped_pairs, deduped_pairs
+                    parameters = tuple(new_parameters)
+
+            return _LiteralGenericAlias(self, parameters)
+
+    Literal = _LiteralForm(doc="""\
+                           A type that can be used to indicate to type checkers
+                           that the corresponding value has a value literally equivalent
+                           to the provided parameter. For example:
+
+                               var: Literal[4] = 4
+
+                           The type checker understands that 'var' is literally equal to
+                           the value 4 and no other value.
+
+                           Literal[...] cannot be subclassed. There is no runtime
+                           checking verifying that the parameter is actually a value
+                           instead of a type.""")
+
+
+_overload_dummy = typing._overload_dummy
+
+
+if hasattr(typing, "get_overloads"):  # 3.11+
+    overload = typing.overload
+    get_overloads = typing.get_overloads
+    clear_overloads = typing.clear_overloads
+else:
+    # {module: {qualname: {firstlineno: func}}}
+    _overload_registry = collections.defaultdict(
+        functools.partial(collections.defaultdict, dict)
+    )
+
+    def overload(func):
+        """Decorator for overloaded functions/methods.
+
+        In a stub file, place two or more stub definitions for the same
+        function in a row, each decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+
+        In a non-stub file (i.e. a regular .py file), do the same but
+        follow it with an implementation.  The implementation should *not*
+        be decorated with @overload.  For example:
+
+        @overload
+        def utf8(value: None) -> None: ...
+        @overload
+        def utf8(value: bytes) -> bytes: ...
+        @overload
+        def utf8(value: str) -> bytes: ...
+        def utf8(value):
+            # implementation goes here
+
+        The overloads for a function can be retrieved at runtime using the
+        get_overloads() function.
+        """
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        try:
+            _overload_registry[f.__module__][f.__qualname__][
+                f.__code__.co_firstlineno
+            ] = func
+        except AttributeError:
+            # Not a normal function; ignore.
+            pass
+        return _overload_dummy
+
+    def get_overloads(func):
+        """Return all defined overloads for *func* as a sequence."""
+        # classmethod and staticmethod
+        f = getattr(func, "__func__", func)
+        if f.__module__ not in _overload_registry:
+            return []
+        mod_dict = _overload_registry[f.__module__]
+        if f.__qualname__ not in mod_dict:
+            return []
+        return list(mod_dict[f.__qualname__].values())
+
+    def clear_overloads():
+        """Clear all overloads in the registry."""
+        _overload_registry.clear()
+
+
+# This is not a real generic class.  Don't use outside annotations.
+Type = typing.Type
+
+# Various ABCs mimicking those in collections.abc.
+# A few are simply re-exported for completeness.
+Awaitable = typing.Awaitable
+Coroutine = typing.Coroutine
+AsyncIterable = typing.AsyncIterable
+AsyncIterator = typing.AsyncIterator
+Deque = typing.Deque
+DefaultDict = typing.DefaultDict
+OrderedDict = typing.OrderedDict
+Counter = typing.Counter
+ChainMap = typing.ChainMap
+Text = typing.Text
+TYPE_CHECKING = typing.TYPE_CHECKING
+
+
+if sys.version_info >= (3, 13, 0, "beta"):
+    from typing import AsyncContextManager, AsyncGenerator, ContextManager, Generator
+else:
+    def _is_dunder(attr):
+        return attr.startswith('__') and attr.endswith('__')
+
+    # Python <3.9 doesn't have typing._SpecialGenericAlias
+    _special_generic_alias_base = getattr(
+        typing, "_SpecialGenericAlias", typing._GenericAlias
+    )
+
+    class _SpecialGenericAlias(_special_generic_alias_base, _root=True):
+        def __init__(self, origin, nparams, *, inst=True, name=None, defaults=()):
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                self.__origin__ = origin
+                self._nparams = nparams
+                super().__init__(origin, nparams, special=True, inst=inst, name=name)
+            else:
+                # Python >= 3.9
+                super().__init__(origin, nparams, inst=inst, name=name)
+            self._defaults = defaults
+
+        def __setattr__(self, attr, val):
+            allowed_attrs = {'_name', '_inst', '_nparams', '_defaults'}
+            if _special_generic_alias_base is typing._GenericAlias:
+                # Python <3.9
+                allowed_attrs.add("__origin__")
+            if _is_dunder(attr) or attr in allowed_attrs:
+                object.__setattr__(self, attr, val)
+            else:
+                setattr(self.__origin__, attr, val)
+
+        @typing._tp_cache
+        def __getitem__(self, params):
+            if not isinstance(params, tuple):
+                params = (params,)
+            msg = "Parameters to generic types must be types."
+            params = tuple(typing._type_check(p, msg) for p in params)
+            if (
+                self._defaults
+                and len(params) < self._nparams
+                and len(params) + len(self._defaults) >= self._nparams
+            ):
+                params = (*params, *self._defaults[len(params) - self._nparams:])
+            actual_len = len(params)
+
+            if actual_len != self._nparams:
+                if self._defaults:
+                    expected = f"at least {self._nparams - len(self._defaults)}"
+                else:
+                    expected = str(self._nparams)
+                if not self._nparams:
+                    raise TypeError(f"{self} is not a generic class")
+                raise TypeError(
+                    f"Too {'many' if actual_len > self._nparams else 'few'}"
+                    f" arguments for {self};"
+                    f" actual {actual_len}, expected {expected}"
+                )
+            return self.copy_with(params)
+
+    _NoneType = type(None)
+    Generator = _SpecialGenericAlias(
+        collections.abc.Generator, 3, defaults=(_NoneType, _NoneType)
+    )
+    AsyncGenerator = _SpecialGenericAlias(
+        collections.abc.AsyncGenerator, 2, defaults=(_NoneType,)
+    )
+    ContextManager = _SpecialGenericAlias(
+        contextlib.AbstractContextManager,
+        2,
+        name="ContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+    AsyncContextManager = _SpecialGenericAlias(
+        contextlib.AbstractAsyncContextManager,
+        2,
+        name="AsyncContextManager",
+        defaults=(typing.Optional[bool],)
+    )
+
+
+_PROTO_ALLOWLIST = {
+    'collections.abc': [
+        'Callable', 'Awaitable', 'Iterable', 'Iterator', 'AsyncIterable',
+        'Hashable', 'Sized', 'Container', 'Collection', 'Reversible', 'Buffer',
+    ],
+    'contextlib': ['AbstractContextManager', 'AbstractAsyncContextManager'],
+    'typing_extensions': ['Buffer'],
+}
+
+
+_EXCLUDED_ATTRS = frozenset(typing.EXCLUDED_ATTRIBUTES) | {
+    "__match_args__", "__protocol_attrs__", "__non_callable_proto_members__",
+    "__final__",
+}
+
+
+def _get_protocol_attrs(cls):
+    attrs = set()
+    for base in cls.__mro__[:-1]:  # without object
+        if base.__name__ in {'Protocol', 'Generic'}:
+            continue
+        annotations = getattr(base, '__annotations__', {})
+        for attr in (*base.__dict__, *annotations):
+            if (not attr.startswith('_abc_') and attr not in _EXCLUDED_ATTRS):
+                attrs.add(attr)
+    return attrs
+
+
+def _caller(depth=2):
+    try:
+        return sys._getframe(depth).f_globals.get('__name__', '__main__')
+    except (AttributeError, ValueError):  # For platforms without _getframe()
+        return None
+
+
+# `__match_args__` attribute was removed from protocol members in 3.13,
+# we want to backport this change to older Python versions.
+if sys.version_info >= (3, 13):
+    Protocol = typing.Protocol
+else:
+    def _allow_reckless_class_checks(depth=3):
+        """Allow instance and class checks for special stdlib modules.
+        The abc and functools modules indiscriminately call isinstance() and
+        issubclass() on the whole MRO of a user class, which may contain protocols.
+        """
+        return _caller(depth) in {'abc', 'functools', None}
+
+    def _no_init(self, *args, **kwargs):
+        if type(self)._is_protocol:
+            raise TypeError('Protocols cannot be instantiated')
+
+    def _type_check_issubclass_arg_1(arg):
+        """Raise TypeError if `arg` is not an instance of `type`
+        in `issubclass(arg, <protocol>)`.
+
+        In most cases, this is verified by type.__subclasscheck__.
+        Checking it again unnecessarily would slow down issubclass() checks,
+        so, we don't perform this check unless we absolutely have to.
+
+        For various error paths, however,
+        we want to ensure that *this* error message is shown to the user
+        where relevant, rather than a typing.py-specific error message.
+        """
+        if not isinstance(arg, type):
+            # Same error message as for issubclass(1, int).
+            raise TypeError('issubclass() arg 1 must be a class')
+
+    # Inheriting from typing._ProtocolMeta isn't actually desirable,
+    # but is necessary to allow typing.Protocol and typing_extensions.Protocol
+    # to mix without getting TypeErrors about "metaclass conflict"
+    class _ProtocolMeta(type(typing.Protocol)):
+        # This metaclass is somewhat unfortunate,
+        # but is necessary for several reasons...
+        #
+        # NOTE: DO NOT call super() in any methods in this class
+        # That would call the methods on typing._ProtocolMeta on Python 3.8-3.11
+        # and those are slow
+        def __new__(mcls, name, bases, namespace, **kwargs):
+            if name == "Protocol" and len(bases) < 2:
+                pass
+            elif {Protocol, typing.Protocol} & set(bases):
+                for base in bases:
+                    if not (
+                        base in {object, typing.Generic, Protocol, typing.Protocol}
+                        or base.__name__ in _PROTO_ALLOWLIST.get(base.__module__, [])
+                        or is_protocol(base)
+                    ):
+                        raise TypeError(
+                            f"Protocols can only inherit from other protocols, "
+                            f"got {base!r}"
+                        )
+            return abc.ABCMeta.__new__(mcls, name, bases, namespace, **kwargs)
+
+        def __init__(cls, *args, **kwargs):
+            abc.ABCMeta.__init__(cls, *args, **kwargs)
+            if getattr(cls, "_is_protocol", False):
+                cls.__protocol_attrs__ = _get_protocol_attrs(cls)
+
+        def __subclasscheck__(cls, other):
+            if cls is Protocol:
+                return type.__subclasscheck__(cls, other)
+            if (
+                getattr(cls, '_is_protocol', False)
+                and not _allow_reckless_class_checks()
+            ):
+                if not getattr(cls, '_is_runtime_protocol', False):
+                    _type_check_issubclass_arg_1(other)
+                    raise TypeError(
+                        "Instance and class checks can only be used with "
+                        "@runtime_checkable protocols"
+                    )
+                if (
+                    # this attribute is set by @runtime_checkable:
+                    cls.__non_callable_proto_members__
+                    and cls.__dict__.get("__subclasshook__") is _proto_hook
+                ):
+                    _type_check_issubclass_arg_1(other)
+                    non_method_attrs = sorted(cls.__non_callable_proto_members__)
+                    raise TypeError(
+                        "Protocols with non-method members don't support issubclass()."
+                        f" Non-method members: {str(non_method_attrs)[1:-1]}."
+                    )
+            return abc.ABCMeta.__subclasscheck__(cls, other)
+
+        def __instancecheck__(cls, instance):
+            # We need this method for situations where attributes are
+            # assigned in __init__.
+            if cls is Protocol:
+                return type.__instancecheck__(cls, instance)
+            if not getattr(cls, "_is_protocol", False):
+                # i.e., it's a concrete subclass of a protocol
+                return abc.ABCMeta.__instancecheck__(cls, instance)
+
+            if (
+                not getattr(cls, '_is_runtime_protocol', False) and
+                not _allow_reckless_class_checks()
+            ):
+                raise TypeError("Instance and class checks can only be used with"
+                                " @runtime_checkable protocols")
+
+            if abc.ABCMeta.__instancecheck__(cls, instance):
+                return True
+
+            for attr in cls.__protocol_attrs__:
+                try:
+                    val = inspect.getattr_static(instance, attr)
+                except AttributeError:
+                    break
+                # this attribute is set by @runtime_checkable:
+                if val is None and attr not in cls.__non_callable_proto_members__:
+                    break
+            else:
+                return True
+
+            return False
+
+        def __eq__(cls, other):
+            # Hack so that typing.Generic.__class_getitem__
+            # treats typing_extensions.Protocol
+            # as equivalent to typing.Protocol
+            if abc.ABCMeta.__eq__(cls, other) is True:
+                return True
+            return cls is Protocol and other is typing.Protocol
+
+        # This has to be defined, or the abc-module cache
+        # complains about classes with this metaclass being unhashable,
+        # if we define only __eq__!
+        def __hash__(cls) -> int:
+            return type.__hash__(cls)
+
+    @classmethod
+    def _proto_hook(cls, other):
+        if not cls.__dict__.get('_is_protocol', False):
+            return NotImplemented
+
+        for attr in cls.__protocol_attrs__:
+            for base in other.__mro__:
+                # Check if the members appears in the class dictionary...
+                if attr in base.__dict__:
+                    if base.__dict__[attr] is None:
+                        return NotImplemented
+                    break
+
+                # ...or in annotations, if it is a sub-protocol.
+                annotations = getattr(base, '__annotations__', {})
+                if (
+                    isinstance(annotations, collections.abc.Mapping)
+                    and attr in annotations
+                    and is_protocol(other)
+                ):
+                    break
+            else:
+                return NotImplemented
+        return True
+
+    class Protocol(typing.Generic, metaclass=_ProtocolMeta):
+        __doc__ = typing.Protocol.__doc__
+        __slots__ = ()
+        _is_protocol = True
+        _is_runtime_protocol = False
+
+        def __init_subclass__(cls, *args, **kwargs):
+            super().__init_subclass__(*args, **kwargs)
+
+            # Determine if this is a protocol or a concrete subclass.
+            if not cls.__dict__.get('_is_protocol', False):
+                cls._is_protocol = any(b is Protocol for b in cls.__bases__)
+
+            # Set (or override) the protocol subclass hook.
+            if '__subclasshook__' not in cls.__dict__:
+                cls.__subclasshook__ = _proto_hook
+
+            # Prohibit instantiation for protocol classes
+            if cls._is_protocol and cls.__init__ is Protocol.__init__:
+                cls.__init__ = _no_init
+
+
+if sys.version_info >= (3, 13):
+    runtime_checkable = typing.runtime_checkable
+else:
+    def runtime_checkable(cls):
+        """Mark a protocol class as a runtime protocol.
+
+        Such protocol can be used with isinstance() and issubclass().
+        Raise TypeError if applied to a non-protocol class.
+        This allows a simple-minded structural check very similar to
+        one trick ponies in collections.abc such as Iterable.
+
+        For example::
+
+            @runtime_checkable
+            class Closable(Protocol):
+                def close(self): ...
+
+            assert isinstance(open('/some/file'), Closable)
+
+        Warning: this will check only the presence of the required methods,
+        not their type signatures!
+        """
+        if not issubclass(cls, typing.Generic) or not getattr(cls, '_is_protocol', False):
+            raise TypeError(f'@runtime_checkable can be only applied to protocol classes,'
+                            f' got {cls!r}')
+        cls._is_runtime_protocol = True
+
+        # typing.Protocol classes on <=3.11 break if we execute this block,
+        # because typing.Protocol classes on <=3.11 don't have a
+        # `__protocol_attrs__` attribute, and this block relies on the
+        # `__protocol_attrs__` attribute. Meanwhile, typing.Protocol classes on 3.12.2+
+        # break if we *don't* execute this block, because *they* assume that all
+        # protocol classes have a `__non_callable_proto_members__` attribute
+        # (which this block sets)
+        if isinstance(cls, _ProtocolMeta) or sys.version_info >= (3, 12, 2):
+            # PEP 544 prohibits using issubclass()
+            # with protocols that have non-method members.
+            # See gh-113320 for why we compute this attribute here,
+            # rather than in `_ProtocolMeta.__init__`
+            cls.__non_callable_proto_members__ = set()
+            for attr in cls.__protocol_attrs__:
+                try:
+                    is_callable = callable(getattr(cls, attr, None))
+                except Exception as e:
+                    raise TypeError(
+                        f"Failed to determine whether protocol member {attr!r} "
+                        "is a method member"
+                    ) from e
+                else:
+                    if not is_callable:
+                        cls.__non_callable_proto_members__.add(attr)
+
+        return cls
+
+
+# The "runtime" alias exists for backwards compatibility.
+runtime = runtime_checkable
+
+
+# Our version of runtime-checkable protocols is faster on Python 3.8-3.11
+if sys.version_info >= (3, 12):
+    SupportsInt = typing.SupportsInt
+    SupportsFloat = typing.SupportsFloat
+    SupportsComplex = typing.SupportsComplex
+    SupportsBytes = typing.SupportsBytes
+    SupportsIndex = typing.SupportsIndex
+    SupportsAbs = typing.SupportsAbs
+    SupportsRound = typing.SupportsRound
+else:
+    @runtime_checkable
+    class SupportsInt(Protocol):
+        """An ABC with one abstract method __int__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __int__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsFloat(Protocol):
+        """An ABC with one abstract method __float__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __float__(self) -> float:
+            pass
+
+    @runtime_checkable
+    class SupportsComplex(Protocol):
+        """An ABC with one abstract method __complex__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __complex__(self) -> complex:
+            pass
+
+    @runtime_checkable
+    class SupportsBytes(Protocol):
+        """An ABC with one abstract method __bytes__."""
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __bytes__(self) -> bytes:
+            pass
+
+    @runtime_checkable
+    class SupportsIndex(Protocol):
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __index__(self) -> int:
+            pass
+
+    @runtime_checkable
+    class SupportsAbs(Protocol[T_co]):
+        """
+        An ABC with one abstract method __abs__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __abs__(self) -> T_co:
+            pass
+
+    @runtime_checkable
+    class SupportsRound(Protocol[T_co]):
+        """
+        An ABC with one abstract method __round__ that is covariant in its return type.
+        """
+        __slots__ = ()
+
+        @abc.abstractmethod
+        def __round__(self, ndigits: int = 0) -> T_co:
+            pass
+
+
+def _ensure_subclassable(mro_entries):
+    def inner(func):
+        if sys.implementation.name == "pypy" and sys.version_info < (3, 9):
+            cls_dict = {
+                "__call__": staticmethod(func),
+                "__mro_entries__": staticmethod(mro_entries)
+            }
+            t = type(func.__name__, (), cls_dict)
+            return functools.update_wrapper(t(), func)
+        else:
+            func.__mro_entries__ = mro_entries
+            return func
+    return inner
+
+
+# Update this to something like >=3.13.0b1 if and when
+# PEP 728 is implemented in CPython
+_PEP_728_IMPLEMENTED = False
+
+if _PEP_728_IMPLEMENTED:
+    # The standard library TypedDict in Python 3.8 does not store runtime information
+    # about which (if any) keys are optional.  See https://bugs.python.org/issue38834
+    # The standard library TypedDict in Python 3.9.0/1 does not honour the "total"
+    # keyword with old-style TypedDict().  See https://bugs.python.org/issue42059
+    # The standard library TypedDict below Python 3.11 does not store runtime
+    # information about optional and required keys when using Required or NotRequired.
+    # Generic TypedDicts are also impossible using typing.TypedDict on Python <3.11.
+    # Aaaand on 3.12 we add __orig_bases__ to TypedDict
+    # to enable better runtime introspection.
+    # On 3.13 we deprecate some odd ways of creating TypedDicts.
+    # Also on 3.13, PEP 705 adds the ReadOnly[] qualifier.
+    # PEP 728 (still pending) makes more changes.
+    TypedDict = typing.TypedDict
+    _TypedDictMeta = typing._TypedDictMeta
+    is_typeddict = typing.is_typeddict
+else:
+    # 3.10.0 and later
+    _TAKES_MODULE = "module" in inspect.signature(typing._type_check).parameters
+
+    def _get_typeddict_qualifiers(annotation_type):
+        while True:
+            annotation_origin = get_origin(annotation_type)
+            if annotation_origin is Annotated:
+                annotation_args = get_args(annotation_type)
+                if annotation_args:
+                    annotation_type = annotation_args[0]
+                else:
+                    break
+            elif annotation_origin is Required:
+                yield Required
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is NotRequired:
+                yield NotRequired
+                annotation_type, = get_args(annotation_type)
+            elif annotation_origin is ReadOnly:
+                yield ReadOnly
+                annotation_type, = get_args(annotation_type)
+            else:
+                break
+
+    class _TypedDictMeta(type):
+        def __new__(cls, name, bases, ns, *, total=True, closed=False):
+            """Create new typed dict class object.
+
+            This method is called when TypedDict is subclassed,
+            or when TypedDict is instantiated. This way
+            TypedDict supports all three syntax forms described in its docstring.
+            Subclasses and instances of TypedDict return actual dictionaries.
+            """
+            for base in bases:
+                if type(base) is not _TypedDictMeta and base is not typing.Generic:
+                    raise TypeError('cannot inherit from both a TypedDict type '
+                                    'and a non-TypedDict base class')
+
+            if any(issubclass(b, typing.Generic) for b in bases):
+                generic_base = (typing.Generic,)
+            else:
+                generic_base = ()
+
+            # typing.py generally doesn't let you inherit from plain Generic, unless
+            # the name of the class happens to be "Protocol"
+            tp_dict = type.__new__(_TypedDictMeta, "Protocol", (*generic_base, dict), ns)
+            tp_dict.__name__ = name
+            if tp_dict.__qualname__ == "Protocol":
+                tp_dict.__qualname__ = name
+
+            if not hasattr(tp_dict, '__orig_bases__'):
+                tp_dict.__orig_bases__ = bases
+
+            annotations = {}
+            if "__annotations__" in ns:
+                own_annotations = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                own_annotations = ns["__annotate__"](1)
+            else:
+                own_annotations = {}
+            msg = "TypedDict('Name', {f0: t0, f1: t1, ...}); each t must be a type"
+            if _TAKES_MODULE:
+                own_annotations = {
+                    n: typing._type_check(tp, msg, module=tp_dict.__module__)
+                    for n, tp in own_annotations.items()
+                }
+            else:
+                own_annotations = {
+                    n: typing._type_check(tp, msg)
+                    for n, tp in own_annotations.items()
+                }
+            required_keys = set()
+            optional_keys = set()
+            readonly_keys = set()
+            mutable_keys = set()
+            extra_items_type = None
+
+            for base in bases:
+                base_dict = base.__dict__
+
+                annotations.update(base_dict.get('__annotations__', {}))
+                required_keys.update(base_dict.get('__required_keys__', ()))
+                optional_keys.update(base_dict.get('__optional_keys__', ()))
+                readonly_keys.update(base_dict.get('__readonly_keys__', ()))
+                mutable_keys.update(base_dict.get('__mutable_keys__', ()))
+                base_extra_items_type = base_dict.get('__extra_items__', None)
+                if base_extra_items_type is not None:
+                    extra_items_type = base_extra_items_type
+
+            if closed and extra_items_type is None:
+                extra_items_type = Never
+            if closed and "__extra_items__" in own_annotations:
+                annotation_type = own_annotations.pop("__extra_items__")
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+                if Required in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "Required"
+                    )
+                if NotRequired in qualifiers:
+                    raise TypeError(
+                        "Special key __extra_items__ does not support "
+                        "NotRequired"
+                    )
+                extra_items_type = annotation_type
+
+            annotations.update(own_annotations)
+            for annotation_key, annotation_type in own_annotations.items():
+                qualifiers = set(_get_typeddict_qualifiers(annotation_type))
+
+                if Required in qualifiers:
+                    required_keys.add(annotation_key)
+                elif NotRequired in qualifiers:
+                    optional_keys.add(annotation_key)
+                elif total:
+                    required_keys.add(annotation_key)
+                else:
+                    optional_keys.add(annotation_key)
+                if ReadOnly in qualifiers:
+                    mutable_keys.discard(annotation_key)
+                    readonly_keys.add(annotation_key)
+                else:
+                    mutable_keys.add(annotation_key)
+                    readonly_keys.discard(annotation_key)
+
+            tp_dict.__annotations__ = annotations
+            tp_dict.__required_keys__ = frozenset(required_keys)
+            tp_dict.__optional_keys__ = frozenset(optional_keys)
+            tp_dict.__readonly_keys__ = frozenset(readonly_keys)
+            tp_dict.__mutable_keys__ = frozenset(mutable_keys)
+            if not hasattr(tp_dict, '__total__'):
+                tp_dict.__total__ = total
+            tp_dict.__closed__ = closed
+            tp_dict.__extra_items__ = extra_items_type
+            return tp_dict
+
+        __call__ = dict  # static method
+
+        def __subclasscheck__(cls, other):
+            # Typed dicts are only for static structural subtyping.
+            raise TypeError('TypedDict does not support instance and class checks')
+
+        __instancecheck__ = __subclasscheck__
+
+    _TypedDict = type.__new__(_TypedDictMeta, 'TypedDict', (), {})
+
+    @_ensure_subclassable(lambda bases: (_TypedDict,))
+    def TypedDict(typename, fields=_marker, /, *, total=True, closed=False, **kwargs):
+        """A simple typed namespace. At runtime it is equivalent to a plain dict.
+
+        TypedDict creates a dictionary type such that a type checker will expect all
+        instances to have a certain set of keys, where each key is
+        associated with a value of a consistent type. This expectation
+        is not checked at runtime.
+
+        Usage::
+
+            class Point2D(TypedDict):
+                x: int
+                y: int
+                label: str
+
+            a: Point2D = {'x': 1, 'y': 2, 'label': 'good'}  # OK
+            b: Point2D = {'z': 3, 'label': 'bad'}           # Fails type check
+
+            assert Point2D(x=1, y=2, label='first') == dict(x=1, y=2, label='first')
+
+        The type info can be accessed via the Point2D.__annotations__ dict, and
+        the Point2D.__required_keys__ and Point2D.__optional_keys__ frozensets.
+        TypedDict supports an additional equivalent form::
+
+            Point2D = TypedDict('Point2D', {'x': int, 'y': int, 'label': str})
+
+        By default, all keys must be present in a TypedDict. It is possible
+        to override this by specifying totality::
+
+            class Point2D(TypedDict, total=False):
+                x: int
+                y: int
+
+        This means that a Point2D TypedDict can have any of the keys omitted. A type
+        checker is only expected to support a literal False or True as the value of
+        the total argument. True is the default, and makes all items defined in the
+        class body be required.
+
+        The Required and NotRequired special forms can also be used to mark
+        individual keys as being required or not required::
+
+            class Point2D(TypedDict):
+                x: int  # the "x" key must always be present (Required is the default)
+                y: NotRequired[int]  # the "y" key can be omitted
+
+        See PEP 655 for more details on Required and NotRequired.
+        """
+        if fields is _marker or fields is None:
+            if fields is _marker:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+
+            example = f"`{typename} = TypedDict({typename!r}, {{}})`"
+            deprecation_msg = (
+                f"{deprecated_thing} is deprecated and will be disallowed in "
+                "Python 3.15. To create a TypedDict class with 0 fields "
+                "using the functional syntax, pass an empty dictionary, e.g. "
+            ) + example + "."
+            warnings.warn(deprecation_msg, DeprecationWarning, stacklevel=2)
+            if closed is not False and closed is not True:
+                kwargs["closed"] = closed
+                closed = False
+            fields = kwargs
+        elif kwargs:
+            raise TypeError("TypedDict takes either a dict or keyword arguments,"
+                            " but not both")
+        if kwargs:
+            if sys.version_info >= (3, 13):
+                raise TypeError("TypedDict takes no keyword arguments")
+            warnings.warn(
+                "The kwargs-based syntax for TypedDict definitions is deprecated "
+                "in Python 3.11, will be removed in Python 3.13, and may not be "
+                "understood by third-party type checkers.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+
+        ns = {'__annotations__': dict(fields)}
+        module = _caller()
+        if module is not None:
+            # Setting correct module is necessary to make typed dict classes pickleable.
+            ns['__module__'] = module
+
+        td = _TypedDictMeta(typename, (), ns, total=total, closed=closed)
+        td.__orig_bases__ = (TypedDict,)
+        return td
+
+    if hasattr(typing, "_TypedDictMeta"):
+        _TYPEDDICT_TYPES = (typing._TypedDictMeta, _TypedDictMeta)
+    else:
+        _TYPEDDICT_TYPES = (_TypedDictMeta,)
+
+    def is_typeddict(tp):
+        """Check if an annotation is a TypedDict class
+
+        For example::
+            class Film(TypedDict):
+                title: str
+                year: int
+
+            is_typeddict(Film)  # => True
+            is_typeddict(Union[list, str])  # => False
+        """
+        # On 3.8, this would otherwise return True
+        if hasattr(typing, "TypedDict") and tp is typing.TypedDict:
+            return False
+        return isinstance(tp, _TYPEDDICT_TYPES)
+
+
+if hasattr(typing, "assert_type"):
+    assert_type = typing.assert_type
+
+else:
+    def assert_type(val, typ, /):
+        """Assert (to the type checker) that the value is of the given type.
+
+        When the type checker encounters a call to assert_type(), it
+        emits an error if the value is not of the specified type::
+
+            def greet(name: str) -> None:
+                assert_type(name, str)  # ok
+                assert_type(name, int)  # type checker error
+
+        At runtime this returns the first argument unchanged and otherwise
+        does nothing.
+        """
+        return val
+
+
+if hasattr(typing, "ReadOnly"):  # 3.13+
+    get_type_hints = typing.get_type_hints
+else:  # <=3.13
+    # replaces _strip_annotations()
+    def _strip_extras(t):
+        """Strips Annotated, Required and NotRequired from a given type."""
+        if isinstance(t, _AnnotatedAlias):
+            return _strip_extras(t.__origin__)
+        if hasattr(t, "__origin__") and t.__origin__ in (Required, NotRequired, ReadOnly):
+            return _strip_extras(t.__args__[0])
+        if isinstance(t, typing._GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return t.copy_with(stripped_args)
+        if hasattr(_types, "GenericAlias") and isinstance(t, _types.GenericAlias):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return _types.GenericAlias(t.__origin__, stripped_args)
+        if hasattr(_types, "UnionType") and isinstance(t, _types.UnionType):
+            stripped_args = tuple(_strip_extras(a) for a in t.__args__)
+            if stripped_args == t.__args__:
+                return t
+            return functools.reduce(operator.or_, stripped_args)
+
+        return t
+
+    def get_type_hints(obj, globalns=None, localns=None, include_extras=False):
+        """Return type hints for an object.
+
+        This is often the same as obj.__annotations__, but it handles
+        forward references encoded as string literals, adds Optional[t] if a
+        default value equal to None is set and recursively replaces all
+        'Annotated[T, ...]', 'Required[T]' or 'NotRequired[T]' with 'T'
+        (unless 'include_extras=True').
+
+        The argument may be a module, class, method, or function. The annotations
+        are returned as a dictionary. For classes, annotations include also
+        inherited members.
+
+        TypeError is raised if the argument is not of a type that can contain
+        annotations, and an empty dictionary is returned if no annotations are
+        present.
+
+        BEWARE -- the behavior of globalns and localns is counterintuitive
+        (unless you are familiar with how eval() and exec() work).  The
+        search order is locals first, then globals.
+
+        - If no dict arguments are passed, an attempt is made to use the
+          globals from obj (or the respective module's globals for classes),
+          and these are also used as the locals.  If the object does not appear
+          to have globals, an empty dictionary is used.
+
+        - If one dict argument is passed, it is used for both globals and
+          locals.
+
+        - If two dict arguments are passed, they specify globals and
+          locals, respectively.
+        """
+        if hasattr(typing, "Annotated"):  # 3.9+
+            hint = typing.get_type_hints(
+                obj, globalns=globalns, localns=localns, include_extras=True
+            )
+        else:  # 3.8
+            hint = typing.get_type_hints(obj, globalns=globalns, localns=localns)
+        if include_extras:
+            return hint
+        return {k: _strip_extras(t) for k, t in hint.items()}
+
+
+# Python 3.9+ has PEP 593 (Annotated)
+if hasattr(typing, 'Annotated'):
+    Annotated = typing.Annotated
+    # Not exported and not a public API, but needed for get_origin() and get_args()
+    # to work.
+    _AnnotatedAlias = typing._AnnotatedAlias
+# 3.8
+else:
+    class _AnnotatedAlias(typing._GenericAlias, _root=True):
+        """Runtime representation of an annotated type.
+
+        At its core 'Annotated[t, dec1, dec2, ...]' is an alias for the type 't'
+        with extra annotations. The alias behaves like a normal typing alias,
+        instantiating is the same as instantiating the underlying type, binding
+        it to types is also the same.
+        """
+        def __init__(self, origin, metadata):
+            if isinstance(origin, _AnnotatedAlias):
+                metadata = origin.__metadata__ + metadata
+                origin = origin.__origin__
+            super().__init__(origin, origin)
+            self.__metadata__ = metadata
+
+        def copy_with(self, params):
+            assert len(params) == 1
+            new_type = params[0]
+            return _AnnotatedAlias(new_type, self.__metadata__)
+
+        def __repr__(self):
+            return (f"typing_extensions.Annotated[{typing._type_repr(self.__origin__)}, "
+                    f"{', '.join(repr(a) for a in self.__metadata__)}]")
+
+        def __reduce__(self):
+            return operator.getitem, (
+                Annotated, (self.__origin__, *self.__metadata__)
+            )
+
+        def __eq__(self, other):
+            if not isinstance(other, _AnnotatedAlias):
+                return NotImplemented
+            if self.__origin__ != other.__origin__:
+                return False
+            return self.__metadata__ == other.__metadata__
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__metadata__))
+
+    class Annotated:
+        """Add context specific metadata to a type.
+
+        Example: Annotated[int, runtime_check.Unsigned] indicates to the
+        hypothetical runtime_check module that this type is an unsigned int.
+        Every other consumer of this type can ignore this metadata and treat
+        this type as int.
+
+        The first argument to Annotated must be a valid type (and will be in
+        the __origin__ field), the remaining arguments are kept as a tuple in
+        the __extra__ field.
+
+        Details:
+
+        - It's an error to call `Annotated` with less than two arguments.
+        - Nested Annotated are flattened::
+
+            Annotated[Annotated[T, Ann1, Ann2], Ann3] == Annotated[T, Ann1, Ann2, Ann3]
+
+        - Instantiating an annotated type is equivalent to instantiating the
+        underlying type::
+
+            Annotated[C, Ann1](5) == C(5)
+
+        - Annotated can be used as a generic type alias::
+
+            Optimized = Annotated[T, runtime.Optimize()]
+            Optimized[int] == Annotated[int, runtime.Optimize()]
+
+            OptimizedList = Annotated[List[T], runtime.Optimize()]
+            OptimizedList[int] == Annotated[List[int], runtime.Optimize()]
+        """
+
+        __slots__ = ()
+
+        def __new__(cls, *args, **kwargs):
+            raise TypeError("Type Annotated cannot be instantiated.")
+
+        @typing._tp_cache
+        def __class_getitem__(cls, params):
+            if not isinstance(params, tuple) or len(params) < 2:
+                raise TypeError("Annotated[...] should be used "
+                                "with at least two arguments (a type and an "
+                                "annotation).")
+            allowed_special_forms = (ClassVar, Final)
+            if get_origin(params[0]) in allowed_special_forms:
+                origin = params[0]
+            else:
+                msg = "Annotated[t, ...]: t must be a type."
+                origin = typing._type_check(params[0], msg)
+            metadata = tuple(params[1:])
+            return _AnnotatedAlias(origin, metadata)
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                f"Cannot subclass {cls.__module__}.Annotated"
+            )
+
+# Python 3.8 has get_origin() and get_args() but those implementations aren't
+# Annotated-aware, so we can't use those. Python 3.9's versions don't support
+# ParamSpecArgs and ParamSpecKwargs, so only Python 3.10's versions will do.
+if sys.version_info[:2] >= (3, 10):
+    get_origin = typing.get_origin
+    get_args = typing.get_args
+# 3.8-3.9
+else:
+    try:
+        # 3.9+
+        from typing import _BaseGenericAlias
+    except ImportError:
+        _BaseGenericAlias = typing._GenericAlias
+    try:
+        # 3.9+
+        from typing import GenericAlias as _typing_GenericAlias
+    except ImportError:
+        _typing_GenericAlias = typing._GenericAlias
+
+    def get_origin(tp):
+        """Get the unsubscripted version of a type.
+
+        This supports generic types, Callable, Tuple, Union, Literal, Final, ClassVar
+        and Annotated. Return None for unsupported types. Examples::
+
+            get_origin(Literal[42]) is Literal
+            get_origin(int) is None
+            get_origin(ClassVar[int]) is ClassVar
+            get_origin(Generic) is Generic
+            get_origin(Generic[T]) is Generic
+            get_origin(Union[T, int]) is Union
+            get_origin(List[Tuple[T, T]][int]) == list
+            get_origin(P.args) is P
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return Annotated
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias, _BaseGenericAlias,
+                           ParamSpecArgs, ParamSpecKwargs)):
+            return tp.__origin__
+        if tp is typing.Generic:
+            return typing.Generic
+        return None
+
+    def get_args(tp):
+        """Get type arguments with all substitutions performed.
+
+        For unions, basic simplifications used by Union constructor are performed.
+        Examples::
+            get_args(Dict[str, int]) == (str, int)
+            get_args(int) == ()
+            get_args(Union[int, Union[T, int], str][int]) == (int, str)
+            get_args(Union[int, Tuple[T, int]][str]) == (int, Tuple[str, int])
+            get_args(Callable[[], T][int]) == ([], int)
+        """
+        if isinstance(tp, _AnnotatedAlias):
+            return (tp.__origin__, *tp.__metadata__)
+        if isinstance(tp, (typing._GenericAlias, _typing_GenericAlias)):
+            if getattr(tp, "_special", False):
+                return ()
+            res = tp.__args__
+            if get_origin(tp) is collections.abc.Callable and res[0] is not Ellipsis:
+                res = (list(res[:-1]), res[-1])
+            return res
+        return ()
+
+
+# 3.10+
+if hasattr(typing, 'TypeAlias'):
+    TypeAlias = typing.TypeAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeAlias(self, parameters):
+        """Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example above.
+        """
+        raise TypeError(f"{self} is not subscriptable")
+# 3.8
+else:
+    TypeAlias = _ExtensionsSpecialForm(
+        'TypeAlias',
+        doc="""Special marker indicating that an assignment should
+        be recognized as a proper type alias definition by type
+        checkers.
+
+        For example::
+
+            Predicate: TypeAlias = Callable[..., bool]
+
+        It's invalid when used anywhere except as in the example
+        above."""
+    )
+
+
+if hasattr(typing, "NoDefault"):
+    NoDefault = typing.NoDefault
+else:
+    class NoDefaultTypeMeta(type):
+        def __setattr__(cls, attr, value):
+            # TypeError is consistent with the behavior of NoneType
+            raise TypeError(
+                f"cannot set {attr!r} attribute of immutable type {cls.__name__!r}"
+            )
+
+    class NoDefaultType(metaclass=NoDefaultTypeMeta):
+        """The type of the NoDefault singleton."""
+
+        __slots__ = ()
+
+        def __new__(cls):
+            return globals().get("NoDefault") or object.__new__(cls)
+
+        def __repr__(self):
+            return "typing_extensions.NoDefault"
+
+        def __reduce__(self):
+            return "NoDefault"
+
+    NoDefault = NoDefaultType()
+    del NoDefaultType, NoDefaultTypeMeta
+
+
+def _set_default(type_param, default):
+    type_param.has_default = lambda: default is not NoDefault
+    type_param.__default__ = default
+
+
+def _set_module(typevarlike):
+    # for pickling:
+    def_mod = _caller(depth=3)
+    if def_mod != 'typing_extensions':
+        typevarlike.__module__ = def_mod
+
+
+class _DefaultMixin:
+    """Mixin for TypeVarLike defaults."""
+
+    __slots__ = ()
+    __init__ = _set_default
+
+
+# Classes using this metaclass must provide a _backported_typevarlike ClassVar
+class _TypeVarLikeMeta(type):
+    def __instancecheck__(cls, __instance: Any) -> bool:
+        return isinstance(__instance, cls._backported_typevarlike)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVar
+else:
+    # Add default and infer_variance parameters from PEP 696 and 695
+    class TypeVar(metaclass=_TypeVarLikeMeta):
+        """Type variable."""
+
+        _backported_typevarlike = typing.TypeVar
+
+        def __new__(cls, name, *constraints, bound=None,
+                    covariant=False, contravariant=False,
+                    default=NoDefault, infer_variance=False):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented (3.12+), can pass infer_variance to typing.TypeVar
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant,
+                                         infer_variance=infer_variance)
+            else:
+                typevar = typing.TypeVar(name, *constraints, bound=bound,
+                                         covariant=covariant, contravariant=contravariant)
+                if infer_variance and (covariant or contravariant):
+                    raise ValueError("Variance cannot be specified with infer_variance.")
+                typevar.__infer_variance__ = infer_variance
+
+            _set_default(typevar, default)
+            _set_module(typevar)
+
+            def _tvar_prepare_subst(alias, args):
+                if (
+                    typevar.has_default()
+                    and alias.__parameters__.index(typevar) == len(args)
+                ):
+                    args += (typevar.__default__,)
+                return args
+
+            typevar.__typing_prepare_subst__ = _tvar_prepare_subst
+            return typevar
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.TypeVar' is not an acceptable base type")
+
+
+# Python 3.10+ has PEP 612
+if hasattr(typing, 'ParamSpecArgs'):
+    ParamSpecArgs = typing.ParamSpecArgs
+    ParamSpecKwargs = typing.ParamSpecKwargs
+# 3.8-3.9
+else:
+    class _Immutable:
+        """Mixin to indicate that object should not be copied."""
+        __slots__ = ()
+
+        def __copy__(self):
+            return self
+
+        def __deepcopy__(self, memo):
+            return self
+
+    class ParamSpecArgs(_Immutable):
+        """The args for a ParamSpec object.
+
+        Given a ParamSpec object P, P.args is an instance of ParamSpecArgs.
+
+        ParamSpecArgs objects have a reference back to their ParamSpec:
+
+        P.args.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.args"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecArgs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+    class ParamSpecKwargs(_Immutable):
+        """The kwargs for a ParamSpec object.
+
+        Given a ParamSpec object P, P.kwargs is an instance of ParamSpecKwargs.
+
+        ParamSpecKwargs objects have a reference back to their ParamSpec:
+
+        P.kwargs.__origin__ is P
+
+        This type is meant for runtime introspection and has no special meaning to
+        static type checkers.
+        """
+        def __init__(self, origin):
+            self.__origin__ = origin
+
+        def __repr__(self):
+            return f"{self.__origin__.__name__}.kwargs"
+
+        def __eq__(self, other):
+            if not isinstance(other, ParamSpecKwargs):
+                return NotImplemented
+            return self.__origin__ == other.__origin__
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import ParamSpec
+
+# 3.10+
+elif hasattr(typing, 'ParamSpec'):
+
+    # Add default parameter - PEP 696
+    class ParamSpec(metaclass=_TypeVarLikeMeta):
+        """Parameter specification."""
+
+        _backported_typevarlike = typing.ParamSpec
+
+        def __new__(cls, name, *, bound=None,
+                    covariant=False, contravariant=False,
+                    infer_variance=False, default=NoDefault):
+            if hasattr(typing, "TypeAliasType"):
+                # PEP 695 implemented, can pass infer_variance to typing.TypeVar
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant,
+                                             infer_variance=infer_variance)
+            else:
+                paramspec = typing.ParamSpec(name, bound=bound,
+                                             covariant=covariant,
+                                             contravariant=contravariant)
+                paramspec.__infer_variance__ = infer_variance
+
+            _set_default(paramspec, default)
+            _set_module(paramspec)
+
+            def _paramspec_prepare_subst(alias, args):
+                params = alias.__parameters__
+                i = params.index(paramspec)
+                if i == len(args) and paramspec.has_default():
+                    args = [*args, paramspec.__default__]
+                if i >= len(args):
+                    raise TypeError(f"Too few arguments for {alias}")
+                # Special case where Z[[int, str, bool]] == Z[int, str, bool] in PEP 612.
+                if len(params) == 1 and not typing._is_param_expr(args[0]):
+                    assert i == 0
+                    args = (args,)
+                # Convert lists to tuples to help other libraries cache the results.
+                elif isinstance(args[i], list):
+                    args = (*args[:i], tuple(args[i]), *args[i + 1:])
+                return args
+
+            paramspec.__typing_prepare_subst__ = _paramspec_prepare_subst
+            return paramspec
+
+        def __init_subclass__(cls) -> None:
+            raise TypeError(f"type '{__name__}.ParamSpec' is not an acceptable base type")
+
+# 3.8-3.9
+else:
+
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class ParamSpec(list, _DefaultMixin):
+        """Parameter specification variable.
+
+        Usage::
+
+           P = ParamSpec('P')
+
+        Parameter specification variables exist primarily for the benefit of static
+        type checkers.  They are used to forward the parameter types of one
+        callable to another callable, a pattern commonly found in higher order
+        functions and decorators.  They are only valid when used in ``Concatenate``,
+        or s the first argument to ``Callable``. In Python 3.10 and higher,
+        they are also supported in user-defined Generics at runtime.
+        See class Generic for more information on generic types.  An
+        example for annotating a decorator::
+
+           T = TypeVar('T')
+           P = ParamSpec('P')
+
+           def add_logging(f: Callable[P, T]) -> Callable[P, T]:
+               '''A type-safe decorator to add logging to a function.'''
+               def inner(*args: P.args, **kwargs: P.kwargs) -> T:
+                   logging.info(f'{f.__name__} was called')
+                   return f(*args, **kwargs)
+               return inner
+
+           @add_logging
+           def add_two(x: float, y: float) -> float:
+               '''Add two numbers together.'''
+               return x + y
+
+        Parameter specification variables defined with covariant=True or
+        contravariant=True can be used to declare covariant or contravariant
+        generic types.  These keyword arguments are valid, but their actual semantics
+        are yet to be decided.  See PEP 612 for details.
+
+        Parameter specification variables can be introspected. e.g.:
+
+           P.__name__ == 'T'
+           P.__bound__ == None
+           P.__covariant__ == False
+           P.__contravariant__ == False
+
+        Note that only parameter specification variables defined in global scope can
+        be pickled.
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        @property
+        def args(self):
+            return ParamSpecArgs(self)
+
+        @property
+        def kwargs(self):
+            return ParamSpecKwargs(self)
+
+        def __init__(self, name, *, bound=None, covariant=False, contravariant=False,
+                     infer_variance=False, default=NoDefault):
+            list.__init__(self, [self])
+            self.__name__ = name
+            self.__covariant__ = bool(covariant)
+            self.__contravariant__ = bool(contravariant)
+            self.__infer_variance__ = bool(infer_variance)
+            if bound:
+                self.__bound__ = typing._type_check(bound, 'Bound must be a type.')
+            else:
+                self.__bound__ = None
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __repr__(self):
+            if self.__infer_variance__:
+                prefix = ''
+            elif self.__covariant__:
+                prefix = '+'
+            elif self.__contravariant__:
+                prefix = '-'
+            else:
+                prefix = '~'
+            return prefix + self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        # Hack to get typing._type_check to pass.
+        def __call__(self, *args, **kwargs):
+            pass
+
+
+# 3.8-3.9
+if not hasattr(typing, 'Concatenate'):
+    # Inherits from list as a workaround for Callable checks in Python < 3.9.2.
+    class _ConcatenateGenericAlias(list):
+
+        # Trick Generic into looking into this for __parameters__.
+        __class__ = typing._GenericAlias
+
+        # Flag in 3.8.
+        _special = False
+
+        def __init__(self, origin, args):
+            super().__init__(args)
+            self.__origin__ = origin
+            self.__args__ = args
+
+        def __repr__(self):
+            _type_repr = typing._type_repr
+            return (f'{_type_repr(self.__origin__)}'
+                    f'[{", ".join(_type_repr(arg) for arg in self.__args__)}]')
+
+        def __hash__(self):
+            return hash((self.__origin__, self.__args__))
+
+        # Hack to get typing._type_check to pass in Generic.
+        def __call__(self, *args, **kwargs):
+            pass
+
+        @property
+        def __parameters__(self):
+            return tuple(
+                tp for tp in self.__args__ if isinstance(tp, (typing.TypeVar, ParamSpec))
+            )
+
+
+# 3.8-3.9
+@typing._tp_cache
+def _concatenate_getitem(self, parameters):
+    if parameters == ():
+        raise TypeError("Cannot take a Concatenate of no types.")
+    if not isinstance(parameters, tuple):
+        parameters = (parameters,)
+    if not isinstance(parameters[-1], ParamSpec):
+        raise TypeError("The last parameter to Concatenate should be a "
+                        "ParamSpec variable.")
+    msg = "Concatenate[arg, ...]: each arg must be a type."
+    parameters = tuple(typing._type_check(p, msg) for p in parameters)
+    return _ConcatenateGenericAlias(self, parameters)
+
+
+# 3.10+
+if hasattr(typing, 'Concatenate'):
+    Concatenate = typing.Concatenate
+    _ConcatenateGenericAlias = typing._ConcatenateGenericAlias
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def Concatenate(self, parameters):
+        """Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """
+        return _concatenate_getitem(self, parameters)
+# 3.8
+else:
+    class _ConcatenateForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            return _concatenate_getitem(self, parameters)
+
+    Concatenate = _ConcatenateForm(
+        'Concatenate',
+        doc="""Used in conjunction with ``ParamSpec`` and ``Callable`` to represent a
+        higher order function which adds, removes or transforms parameters of a
+        callable.
+
+        For example::
+
+           Callable[Concatenate[int, P], int]
+
+        See PEP 612 for detailed information.
+        """)
+
+# 3.10+
+if hasattr(typing, 'TypeGuard'):
+    TypeGuard = typing.TypeGuard
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeGuard(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeGuardForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeGuard = _TypeGuardForm(
+        'TypeGuard',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type guard function.  ``TypeGuard`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeGuard`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeGuard[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeGuard`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the type inside ``TypeGuard``.
+
+        For example::
+
+            def is_str(val: Union[str, float]):
+                # "isinstance" type guard
+                if isinstance(val, str):
+                    # Type of ``val`` is narrowed to ``str``
+                    ...
+                else:
+                    # Else, type of ``val`` is narrowed to ``float``.
+                    ...
+
+        Strict type narrowing is not enforced -- ``TypeB`` need not be a narrower
+        form of ``TypeA`` (it can even be a wider form) and this may lead to
+        type-unsafe results.  The main reason is to allow for things like
+        narrowing ``List[object]`` to ``List[str]`` even though the latter is not
+        a subtype of the former, since ``List`` is invariant.  The responsibility of
+        writing type-safe type guards is left to the user.
+
+        ``TypeGuard`` also works with type variables.  For more information, see
+        PEP 647 (User-Defined Type Guards).
+        """)
+
+# 3.13+
+if hasattr(typing, 'TypeIs'):
+    TypeIs = typing.TypeIs
+# 3.9
+elif sys.version_info[:2] >= (3, 9):
+    @_ExtensionsSpecialForm
+    def TypeIs(self, parameters):
+        """Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """
+        item = typing._type_check(parameters, f'{self} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+# 3.8
+else:
+    class _TypeIsForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type')
+            return typing._GenericAlias(self, (item,))
+
+    TypeIs = _TypeIsForm(
+        'TypeIs',
+        doc="""Special typing form used to annotate the return type of a user-defined
+        type narrower function.  ``TypeIs`` only accepts a single type argument.
+        At runtime, functions marked this way should return a boolean.
+
+        ``TypeIs`` aims to benefit *type narrowing* -- a technique used by static
+        type checkers to determine a more precise type of an expression within a
+        program's code flow.  Usually type narrowing is done by analyzing
+        conditional code flow and applying the narrowing to a block of code.  The
+        conditional expression here is sometimes referred to as a "type guard".
+
+        Sometimes it would be convenient to use a user-defined boolean function
+        as a type guard.  Such a function should use ``TypeIs[...]`` as its
+        return type to alert static type checkers to this intention.
+
+        Using  ``-> TypeIs`` tells the static type checker that for a given
+        function:
+
+        1. The return value is a boolean.
+        2. If the return value is ``True``, the type of its argument
+        is the intersection of the type inside ``TypeGuard`` and the argument's
+        previously known type.
+
+        For example::
+
+            def is_awaitable(val: object) -> TypeIs[Awaitable[Any]]:
+                return hasattr(val, '__await__')
+
+            def f(val: Union[int, Awaitable[int]]) -> int:
+                if is_awaitable(val):
+                    assert_type(val, Awaitable[int])
+                else:
+                    assert_type(val, int)
+
+        ``TypeIs`` also works with type variables.  For more information, see
+        PEP 742 (Narrowing types with TypeIs).
+        """)
+
+
+# Vendored from cpython typing._SpecialFrom
+class _SpecialForm(typing._Final, _root=True):
+    __slots__ = ('_name', '__doc__', '_getitem')
+
+    def __init__(self, getitem):
+        self._getitem = getitem
+        self._name = getitem.__name__
+        self.__doc__ = getitem.__doc__
+
+    def __getattr__(self, item):
+        if item in {'__name__', '__qualname__'}:
+            return self._name
+
+        raise AttributeError(item)
+
+    def __mro_entries__(self, bases):
+        raise TypeError(f"Cannot subclass {self!r}")
+
+    def __repr__(self):
+        return f'typing_extensions.{self._name}'
+
+    def __reduce__(self):
+        return self._name
+
+    def __call__(self, *args, **kwds):
+        raise TypeError(f"Cannot instantiate {self!r}")
+
+    def __or__(self, other):
+        return typing.Union[self, other]
+
+    def __ror__(self, other):
+        return typing.Union[other, self]
+
+    def __instancecheck__(self, obj):
+        raise TypeError(f"{self} cannot be used with isinstance()")
+
+    def __subclasscheck__(self, cls):
+        raise TypeError(f"{self} cannot be used with issubclass()")
+
+    @typing._tp_cache
+    def __getitem__(self, parameters):
+        return self._getitem(self, parameters)
+
+
+if hasattr(typing, "LiteralString"):  # 3.11+
+    LiteralString = typing.LiteralString
+else:
+    @_SpecialForm
+    def LiteralString(self, params):
+        """Represents an arbitrary literal string.
+
+        Example::
+
+          from typing_extensions import LiteralString
+
+          def query(sql: LiteralString) -> ...:
+              ...
+
+          query("SELECT * FROM table")  # ok
+          query(f"SELECT * FROM {input()}")  # not ok
+
+        See PEP 675 for details.
+
+        """
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Self"):  # 3.11+
+    Self = typing.Self
+else:
+    @_SpecialForm
+    def Self(self, params):
+        """Used to spell the type of "self" in classes.
+
+        Example::
+
+          from typing import Self
+
+          class ReturnsSelf:
+              def parse(self, data: bytes) -> Self:
+                  ...
+                  return self
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, "Never"):  # 3.11+
+    Never = typing.Never
+else:
+    @_SpecialForm
+    def Never(self, params):
+        """The bottom type, a type that has no members.
+
+        This can be used to define a function that should never be
+        called, or a function that never returns::
+
+            from typing_extensions import Never
+
+            def never_call_me(arg: Never) -> None:
+                pass
+
+            def int_or_str(arg: int | str) -> None:
+                never_call_me(arg)  # type checker error
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        never_call_me(arg)  # ok, arg is of type Never
+
+        """
+
+        raise TypeError(f"{self} is not subscriptable")
+
+
+if hasattr(typing, 'Required'):  # 3.11+
+    Required = typing.Required
+    NotRequired = typing.NotRequired
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.10
+    @_ExtensionsSpecialForm
+    def Required(self, parameters):
+        """A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+    @_ExtensionsSpecialForm
+    def NotRequired(self, parameters):
+        """A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _RequiredForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    Required = _RequiredForm(
+        'Required',
+        doc="""A special typing construct to mark a key of a total=False TypedDict
+        as required. For example:
+
+            class Movie(TypedDict, total=False):
+                title: Required[str]
+                year: int
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+
+        There is no runtime checking that a required key is actually provided
+        when instantiating a related TypedDict.
+        """)
+    NotRequired = _RequiredForm(
+        'NotRequired',
+        doc="""A special typing construct to mark a key of a TypedDict as
+        potentially missing. For example:
+
+            class Movie(TypedDict):
+                title: str
+                year: NotRequired[int]
+
+            m = Movie(
+                title='The Matrix',  # typechecker error if key is omitted
+                year=1999,
+            )
+        """)
+
+
+if hasattr(typing, 'ReadOnly'):
+    ReadOnly = typing.ReadOnly
+elif sys.version_info[:2] >= (3, 9):  # 3.9-3.12
+    @_ExtensionsSpecialForm
+    def ReadOnly(self, parameters):
+        """A special typing construct to mark an item of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this property.
+        """
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return typing._GenericAlias(self, (item,))
+
+else:  # 3.8
+    class _ReadOnlyForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return typing._GenericAlias(self, (item,))
+
+    ReadOnly = _ReadOnlyForm(
+        'ReadOnly',
+        doc="""A special typing construct to mark a key of a TypedDict as read-only.
+
+        For example:
+
+            class Movie(TypedDict):
+                title: ReadOnly[str]
+                year: int
+
+            def mutate_movie(m: Movie) -> None:
+                m["year"] = 1992  # allowed
+                m["title"] = "The Matrix"  # typechecker error
+
+        There is no runtime checking for this propery.
+        """)
+
+
+_UNPACK_DOC = """\
+Type unpack operator.
+
+The type unpack operator takes the child types from some container type,
+such as `tuple[int, str]` or a `TypeVarTuple`, and 'pulls them out'. For
+example:
+
+  # For some generic class `Foo`:
+  Foo[Unpack[tuple[int, str]]]  # Equivalent to Foo[int, str]
+
+  Ts = TypeVarTuple('Ts')
+  # Specifies that `Bar` is generic in an arbitrary number of types.
+  # (Think of `Ts` as a tuple of an arbitrary number of individual
+  #  `TypeVar`s, which the `Unpack` is 'pulling out' directly into the
+  #  `Generic[]`.)
+  class Bar(Generic[Unpack[Ts]]): ...
+  Bar[int]  # Valid
+  Bar[int, str]  # Also valid
+
+From Python 3.11, this can also be done using the `*` operator:
+
+    Foo[*tuple[int, str]]
+    class Bar(Generic[*Ts]): ...
+
+The operator can also be used along with a `TypedDict` to annotate
+`**kwargs` in a function signature. For instance:
+
+  class Movie(TypedDict):
+    name: str
+    year: int
+
+  # This function expects two keyword arguments - *name* of type `str` and
+  # *year* of type `int`.
+  def foo(**kwargs: Unpack[Movie]): ...
+
+Note that there is only some runtime checking of this operator. Not
+everything the runtime allows may be accepted by static type checkers.
+
+For more information, see PEP 646 and PEP 692.
+"""
+
+
+if sys.version_info >= (3, 12):  # PEP 692 changed the repr of Unpack[]
+    Unpack = typing.Unpack
+
+    def _is_unpack(obj):
+        return get_origin(obj) is Unpack
+
+elif sys.version_info[:2] >= (3, 9):  # 3.9+
+    class _UnpackSpecialForm(_ExtensionsSpecialForm, _root=True):
+        def __init__(self, getitem):
+            super().__init__(getitem)
+            self.__doc__ = _UNPACK_DOC
+
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+        @property
+        def __typing_unpacked_tuple_args__(self):
+            assert self.__origin__ is Unpack
+            assert len(self.__args__) == 1
+            arg, = self.__args__
+            if isinstance(arg, (typing._GenericAlias, _types.GenericAlias)):
+                if arg.__origin__ is not tuple:
+                    raise TypeError("Unpack[...] must be used with a tuple type")
+                return arg.__args__
+            return None
+
+    @_UnpackSpecialForm
+    def Unpack(self, parameters):
+        item = typing._type_check(parameters, f'{self._name} accepts only a single type.')
+        return _UnpackAlias(self, (item,))
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+else:  # 3.8
+    class _UnpackAlias(typing._GenericAlias, _root=True):
+        __class__ = typing.TypeVar
+
+    class _UnpackForm(_ExtensionsSpecialForm, _root=True):
+        def __getitem__(self, parameters):
+            item = typing._type_check(parameters,
+                                      f'{self._name} accepts only a single type.')
+            return _UnpackAlias(self, (item,))
+
+    Unpack = _UnpackForm('Unpack', doc=_UNPACK_DOC)
+
+    def _is_unpack(obj):
+        return isinstance(obj, _UnpackAlias)
+
+
+if _PEP_696_IMPLEMENTED:
+    from typing import TypeVarTuple
+
+elif hasattr(typing, "TypeVarTuple"):  # 3.11+
+
+    def _unpack_args(*args):
+        newargs = []
+        for arg in args:
+            subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+            if subargs is not None and not (subargs and subargs[-1] is ...):
+                newargs.extend(subargs)
+            else:
+                newargs.append(arg)
+        return newargs
+
+    # Add default parameter - PEP 696
+    class TypeVarTuple(metaclass=_TypeVarLikeMeta):
+        """Type variable tuple."""
+
+        _backported_typevarlike = typing.TypeVarTuple
+
+        def __new__(cls, name, *, default=NoDefault):
+            tvt = typing.TypeVarTuple(name)
+            _set_default(tvt, default)
+            _set_module(tvt)
+
+            def _typevartuple_prepare_subst(alias, args):
+                params = alias.__parameters__
+                typevartuple_index = params.index(tvt)
+                for param in params[typevartuple_index + 1:]:
+                    if isinstance(param, TypeVarTuple):
+                        raise TypeError(
+                            f"More than one TypeVarTuple parameter in {alias}"
+                        )
+
+                alen = len(args)
+                plen = len(params)
+                left = typevartuple_index
+                right = plen - typevartuple_index - 1
+                var_tuple_index = None
+                fillarg = None
+                for k, arg in enumerate(args):
+                    if not isinstance(arg, type):
+                        subargs = getattr(arg, '__typing_unpacked_tuple_args__', None)
+                        if subargs and len(subargs) == 2 and subargs[-1] is ...:
+                            if var_tuple_index is not None:
+                                raise TypeError(
+                                    "More than one unpacked "
+                                    "arbitrary-length tuple argument"
+                                )
+                            var_tuple_index = k
+                            fillarg = subargs[0]
+                if var_tuple_index is not None:
+                    left = min(left, var_tuple_index)
+                    right = min(right, alen - var_tuple_index - 1)
+                elif left + right > alen:
+                    raise TypeError(f"Too few arguments for {alias};"
+                                    f" actual {alen}, expected at least {plen - 1}")
+                if left == alen - right and tvt.has_default():
+                    replacement = _unpack_args(tvt.__default__)
+                else:
+                    replacement = args[left: alen - right]
+
+                return (
+                    *args[:left],
+                    *([fillarg] * (typevartuple_index - left)),
+                    replacement,
+                    *([fillarg] * (plen - right - left - typevartuple_index - 1)),
+                    *args[alen - right:],
+                )
+
+            tvt.__typing_prepare_subst__ = _typevartuple_prepare_subst
+            return tvt
+
+        def __init_subclass__(self, *args, **kwds):
+            raise TypeError("Cannot subclass special typing classes")
+
+else:  # <=3.10
+    class TypeVarTuple(_DefaultMixin):
+        """Type variable tuple.
+
+        Usage::
+
+            Ts = TypeVarTuple('Ts')
+
+        In the same way that a normal type variable is a stand-in for a single
+        type such as ``int``, a type variable *tuple* is a stand-in for a *tuple*
+        type such as ``Tuple[int, str]``.
+
+        Type variable tuples can be used in ``Generic`` declarations.
+        Consider the following example::
+
+            class Array(Generic[*Ts]): ...
+
+        The ``Ts`` type variable tuple here behaves like ``tuple[T1, T2]``,
+        where ``T1`` and ``T2`` are type variables. To use these type variables
+        as type parameters of ``Array``, we must *unpack* the type variable tuple using
+        the star operator: ``*Ts``. The signature of ``Array`` then behaves
+        as if we had simply written ``class Array(Generic[T1, T2]): ...``.
+        In contrast to ``Generic[T1, T2]``, however, ``Generic[*Shape]`` allows
+        us to parameterise the class with an *arbitrary* number of type parameters.
+
+        Type variable tuples can be used anywhere a normal ``TypeVar`` can.
+        This includes class definitions, as shown above, as well as function
+        signatures and variable annotations::
+
+            class Array(Generic[*Ts]):
+
+                def __init__(self, shape: Tuple[*Ts]):
+                    self._shape: Tuple[*Ts] = shape
+
+                def get_shape(self) -> Tuple[*Ts]:
+                    return self._shape
+
+            shape = (Height(480), Width(640))
+            x: Array[Height, Width] = Array(shape)
+            y = abs(x)  # Inferred type is Array[Height, Width]
+            z = x + x   #        ...    is Array[Height, Width]
+            x.get_shape()  #     ...    is tuple[Height, Width]
+
+        """
+
+        # Trick Generic __parameters__.
+        __class__ = typing.TypeVar
+
+        def __iter__(self):
+            yield self.__unpacked__
+
+        def __init__(self, name, *, default=NoDefault):
+            self.__name__ = name
+            _DefaultMixin.__init__(self, default)
+
+            # for pickling:
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+            self.__unpacked__ = Unpack[self]
+
+        def __repr__(self):
+            return self.__name__
+
+        def __hash__(self):
+            return object.__hash__(self)
+
+        def __eq__(self, other):
+            return self is other
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(self, *args, **kwds):
+            if '_root' not in kwds:
+                raise TypeError("Cannot subclass special typing classes")
+
+
+if hasattr(typing, "reveal_type"):  # 3.11+
+    reveal_type = typing.reveal_type
+else:  # <=3.10
+    def reveal_type(obj: T, /) -> T:
+        """Reveal the inferred type of a variable.
+
+        When a static type checker encounters a call to ``reveal_type()``,
+        it will emit the inferred type of the argument::
+
+            x: int = 1
+            reveal_type(x)
+
+        Running a static type checker (e.g., ``mypy``) on this example
+        will produce output similar to 'Revealed type is "builtins.int"'.
+
+        At runtime, the function prints the runtime type of the
+        argument and returns it unchanged.
+
+        """
+        print(f"Runtime type is {type(obj).__name__!r}", file=sys.stderr)
+        return obj
+
+
+if hasattr(typing, "_ASSERT_NEVER_REPR_MAX_LENGTH"):  # 3.11+
+    _ASSERT_NEVER_REPR_MAX_LENGTH = typing._ASSERT_NEVER_REPR_MAX_LENGTH
+else:  # <=3.10
+    _ASSERT_NEVER_REPR_MAX_LENGTH = 100
+
+
+if hasattr(typing, "assert_never"):  # 3.11+
+    assert_never = typing.assert_never
+else:  # <=3.10
+    def assert_never(arg: Never, /) -> Never:
+        """Assert to the type checker that a line of code is unreachable.
+
+        Example::
+
+            def int_or_str(arg: int | str) -> None:
+                match arg:
+                    case int():
+                        print("It's an int")
+                    case str():
+                        print("It's a str")
+                    case _:
+                        assert_never(arg)
+
+        If a type checker finds that a call to assert_never() is
+        reachable, it will emit an error.
+
+        At runtime, this throws an exception when called.
+
+        """
+        value = repr(arg)
+        if len(value) > _ASSERT_NEVER_REPR_MAX_LENGTH:
+            value = value[:_ASSERT_NEVER_REPR_MAX_LENGTH] + '...'
+        raise AssertionError(f"Expected code to be unreachable, but got: {value}")
+
+
+if sys.version_info >= (3, 12):  # 3.12+
+    # dataclass_transform exists in 3.11 but lacks the frozen_default parameter
+    dataclass_transform = typing.dataclass_transform
+else:  # <=3.11
+    def dataclass_transform(
+        *,
+        eq_default: bool = True,
+        order_default: bool = False,
+        kw_only_default: bool = False,
+        frozen_default: bool = False,
+        field_specifiers: typing.Tuple[
+            typing.Union[typing.Type[typing.Any], typing.Callable[..., typing.Any]],
+            ...
+        ] = (),
+        **kwargs: typing.Any,
+    ) -> typing.Callable[[T], T]:
+        """Decorator that marks a function, class, or metaclass as providing
+        dataclass-like behavior.
+
+        Example:
+
+            from typing_extensions import dataclass_transform
+
+            _T = TypeVar("_T")
+
+            # Used on a decorator function
+            @dataclass_transform()
+            def create_model(cls: type[_T]) -> type[_T]:
+                ...
+                return cls
+
+            @create_model
+            class CustomerModel:
+                id: int
+                name: str
+
+            # Used on a base class
+            @dataclass_transform()
+            class ModelBase: ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+            # Used on a metaclass
+            @dataclass_transform()
+            class ModelMeta(type): ...
+
+            class ModelBase(metaclass=ModelMeta): ...
+
+            class CustomerModel(ModelBase):
+                id: int
+                name: str
+
+        Each of the ``CustomerModel`` classes defined in this example will now
+        behave similarly to a dataclass created with the ``@dataclasses.dataclass``
+        decorator. For example, the type checker will synthesize an ``__init__``
+        method.
+
+        The arguments to this decorator can be used to customize this behavior:
+        - ``eq_default`` indicates whether the ``eq`` parameter is assumed to be
+          True or False if it is omitted by the caller.
+        - ``order_default`` indicates whether the ``order`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``kw_only_default`` indicates whether the ``kw_only`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``frozen_default`` indicates whether the ``frozen`` parameter is
+          assumed to be True or False if it is omitted by the caller.
+        - ``field_specifiers`` specifies a static list of supported classes
+          or functions that describe fields, similar to ``dataclasses.field()``.
+
+        At runtime, this decorator records its arguments in the
+        ``__dataclass_transform__`` attribute on the decorated object.
+
+        See PEP 681 for details.
+
+        """
+        def decorator(cls_or_fn):
+            cls_or_fn.__dataclass_transform__ = {
+                "eq_default": eq_default,
+                "order_default": order_default,
+                "kw_only_default": kw_only_default,
+                "frozen_default": frozen_default,
+                "field_specifiers": field_specifiers,
+                "kwargs": kwargs,
+            }
+            return cls_or_fn
+        return decorator
+
+
+if hasattr(typing, "override"):  # 3.12+
+    override = typing.override
+else:  # <=3.11
+    _F = typing.TypeVar("_F", bound=typing.Callable[..., typing.Any])
+
+    def override(arg: _F, /) -> _F:
+        """Indicate that a method is intended to override a method in a base class.
+
+        Usage:
+
+            class Base:
+                def method(self) -> None:
+                    pass
+
+            class Child(Base):
+                @override
+                def method(self) -> None:
+                    super().method()
+
+        When this decorator is applied to a method, the type checker will
+        validate that it overrides a method with the same name on a base class.
+        This helps prevent bugs that may occur when a base class is changed
+        without an equivalent change to a child class.
+
+        There is no runtime checking of these properties. The decorator
+        sets the ``__override__`` attribute to ``True`` on the decorated object
+        to allow runtime introspection.
+
+        See PEP 698 for details.
+
+        """
+        try:
+            arg.__override__ = True
+        except (AttributeError, TypeError):
+            # Skip the attribute silently if it is not writable.
+            # AttributeError happens if the object has __slots__ or a
+            # read-only property, TypeError if it's a builtin class.
+            pass
+        return arg
+
+
+if hasattr(warnings, "deprecated"):
+    deprecated = warnings.deprecated
+else:
+    _T = typing.TypeVar("_T")
+
+    class deprecated:
+        """Indicate that a class, function or overload is deprecated.
+
+        When this decorator is applied to an object, the type checker
+        will generate a diagnostic on usage of the deprecated object.
+
+        Usage:
+
+            @deprecated("Use B instead")
+            class A:
+                pass
+
+            @deprecated("Use g instead")
+            def f():
+                pass
+
+            @overload
+            @deprecated("int support is deprecated")
+            def g(x: int) -> int: ...
+            @overload
+            def g(x: str) -> int: ...
+
+        The warning specified by *category* will be emitted at runtime
+        on use of deprecated objects. For functions, that happens on calls;
+        for classes, on instantiation and on creation of subclasses.
+        If the *category* is ``None``, no warning is emitted at runtime.
+        The *stacklevel* determines where the
+        warning is emitted. If it is ``1`` (the default), the warning
+        is emitted at the direct caller of the deprecated object; if it
+        is higher, it is emitted further up the stack.
+        Static type checker behavior is not affected by the *category*
+        and *stacklevel* arguments.
+
+        The deprecation message passed to the decorator is saved in the
+        ``__deprecated__`` attribute on the decorated object.
+        If applied to an overload, the decorator
+        must be after the ``@overload`` decorator for the attribute to
+        exist on the overload as returned by ``get_overloads()``.
+
+        See PEP 702 for details.
+
+        """
+        def __init__(
+            self,
+            message: str,
+            /,
+            *,
+            category: typing.Optional[typing.Type[Warning]] = DeprecationWarning,
+            stacklevel: int = 1,
+        ) -> None:
+            if not isinstance(message, str):
+                raise TypeError(
+                    "Expected an object of type str for 'message', not "
+                    f"{type(message).__name__!r}"
+                )
+            self.message = message
+            self.category = category
+            self.stacklevel = stacklevel
+
+        def __call__(self, arg: _T, /) -> _T:
+            # Make sure the inner functions created below don't
+            # retain a reference to self.
+            msg = self.message
+            category = self.category
+            stacklevel = self.stacklevel
+            if category is None:
+                arg.__deprecated__ = msg
+                return arg
+            elif isinstance(arg, type):
+                import functools
+                from types import MethodType
+
+                original_new = arg.__new__
+
+                @functools.wraps(original_new)
+                def __new__(cls, *args, **kwargs):
+                    if cls is arg:
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    if original_new is not object.__new__:
+                        return original_new(cls, *args, **kwargs)
+                    # Mirrors a similar check in object.__new__.
+                    elif cls.__init__ is object.__init__ and (args or kwargs):
+                        raise TypeError(f"{cls.__name__}() takes no arguments")
+                    else:
+                        return original_new(cls)
+
+                arg.__new__ = staticmethod(__new__)
+
+                original_init_subclass = arg.__init_subclass__
+                # We need slightly different behavior if __init_subclass__
+                # is a bound method (likely if it was implemented in Python)
+                if isinstance(original_init_subclass, MethodType):
+                    original_init_subclass = original_init_subclass.__func__
+
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = classmethod(__init_subclass__)
+                # Or otherwise, which likely means it's a builtin such as
+                # object's implementation of __init_subclass__.
+                else:
+                    @functools.wraps(original_init_subclass)
+                    def __init_subclass__(*args, **kwargs):
+                        warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                        return original_init_subclass(*args, **kwargs)
+
+                    arg.__init_subclass__ = __init_subclass__
+
+                arg.__deprecated__ = __new__.__deprecated__ = msg
+                __init_subclass__.__deprecated__ = msg
+                return arg
+            elif callable(arg):
+                import functools
+
+                @functools.wraps(arg)
+                def wrapper(*args, **kwargs):
+                    warnings.warn(msg, category=category, stacklevel=stacklevel + 1)
+                    return arg(*args, **kwargs)
+
+                arg.__deprecated__ = wrapper.__deprecated__ = msg
+                return wrapper
+            else:
+                raise TypeError(
+                    "@deprecated decorator with non-None category must be applied to "
+                    f"a class or callable, not {arg!r}"
+                )
+
+
+# We have to do some monkey patching to deal with the dual nature of
+# Unpack/TypeVarTuple:
+# - We want Unpack to be a kind of TypeVar so it gets accepted in
+#   Generic[Unpack[Ts]]
+# - We want it to *not* be treated as a TypeVar for the purposes of
+#   counting generic parameters, so that when we subscript a generic,
+#   the runtime doesn't try to substitute the Unpack with the subscripted type.
+if not hasattr(typing, "TypeVarTuple"):
+    def _check_generic(cls, parameters, elen=_marker):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        if elen is _marker:
+            if not hasattr(cls, "__parameters__") or not cls.__parameters__:
+                raise TypeError(f"{cls} is not a generic class")
+            elen = len(cls.__parameters__)
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+                num_tv_tuples = sum(isinstance(p, TypeVarTuple) for p in parameters)
+                if (num_tv_tuples > 0) and (alen >= elen - num_tv_tuples):
+                    return
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            things = "arguments" if sys.version_info >= (3, 10) else "parameters"
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} {things}"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+else:
+    # Python 3.11+
+
+    def _check_generic(cls, parameters, elen):
+        """Check correct count for parameters of a generic cls (internal helper).
+
+        This gives a nice error message in case of count mismatch.
+        """
+        if not elen:
+            raise TypeError(f"{cls} is not a generic class")
+        alen = len(parameters)
+        if alen != elen:
+            expect_val = elen
+            if hasattr(cls, "__parameters__"):
+                parameters = [p for p in cls.__parameters__ if not _is_unpack(p)]
+
+                # deal with TypeVarLike defaults
+                # required TypeVarLikes cannot appear after a defaulted one.
+                if alen < elen:
+                    # since we validate TypeVarLike default in _collect_type_vars
+                    # or _collect_parameters we can safely check parameters[alen]
+                    if (
+                        getattr(parameters[alen], '__default__', NoDefault)
+                        is not NoDefault
+                    ):
+                        return
+
+                    num_default_tv = sum(getattr(p, '__default__', NoDefault)
+                                         is not NoDefault for p in parameters)
+
+                    elen -= num_default_tv
+
+                    expect_val = f"at least {elen}"
+
+            raise TypeError(f"Too {'many' if alen > elen else 'few'} arguments"
+                            f" for {cls}; actual {alen}, expected {expect_val}")
+
+if not _PEP_696_IMPLEMENTED:
+    typing._check_generic = _check_generic
+
+
+def _has_generic_or_protocol_as_origin() -> bool:
+    try:
+        frame = sys._getframe(2)
+    # - Catch AttributeError: not all Python implementations have sys._getframe()
+    # - Catch ValueError: maybe we're called from an unexpected module
+    #   and the call stack isn't deep enough
+    except (AttributeError, ValueError):
+        return False  # err on the side of leniency
+    else:
+        # If we somehow get invoked from outside typing.py,
+        # also err on the side of leniency
+        if frame.f_globals.get("__name__") != "typing":
+            return False
+        origin = frame.f_locals.get("origin")
+        # Cannot use "in" because origin may be an object with a buggy __eq__ that
+        # throws an error.
+        return origin is typing.Generic or origin is Protocol or origin is typing.Protocol
+
+
+_TYPEVARTUPLE_TYPES = {TypeVarTuple, getattr(typing, "TypeVarTuple", None)}
+
+
+def _is_unpacked_typevartuple(x) -> bool:
+    if get_origin(x) is not Unpack:
+        return False
+    args = get_args(x)
+    return (
+        bool(args)
+        and len(args) == 1
+        and type(args[0]) in _TYPEVARTUPLE_TYPES
+    )
+
+
+# Python 3.11+ _collect_type_vars was renamed to _collect_parameters
+if hasattr(typing, '_collect_type_vars'):
+    def _collect_type_vars(types, typevar_types=None):
+        """Collect all type variable contained in types in order of
+        first appearance (lexicographic order). For example::
+
+            _collect_type_vars((T, List[S, T])) == (T, S)
+        """
+        if typevar_types is None:
+            typevar_types = typing.TypeVar
+        tvars = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with a default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in types:
+            if _is_unpacked_typevartuple(t):
+                type_var_tuple_encountered = True
+            elif isinstance(t, typevar_types) and t not in tvars:
+                if enforce_default_ordering:
+                    has_default = getattr(t, '__default__', NoDefault) is not NoDefault
+                    if has_default:
+                        if type_var_tuple_encountered:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+                        default_encountered = True
+                    elif default_encountered:
+                        raise TypeError(f'Type parameter {t!r} without a default'
+                                        ' follows type parameter with a default')
+
+                tvars.append(t)
+            if _should_collect_from_parameters(t):
+                tvars.extend([t for t in t.__parameters__ if t not in tvars])
+        return tuple(tvars)
+
+    typing._collect_type_vars = _collect_type_vars
+else:
+    def _collect_parameters(args):
+        """Collect all type variables and parameter specifications in args
+        in order of first appearance (lexicographic order).
+
+        For example::
+
+            assert _collect_parameters((T, Callable[P, T])) == (T, P)
+        """
+        parameters = []
+
+        # A required TypeVarLike cannot appear after a TypeVarLike with default
+        # if it was a direct call to `Generic[]` or `Protocol[]`
+        enforce_default_ordering = _has_generic_or_protocol_as_origin()
+        default_encountered = False
+
+        # Also, a TypeVarLike with a default cannot appear after a TypeVarTuple
+        type_var_tuple_encountered = False
+
+        for t in args:
+            if isinstance(t, type):
+                # We don't want __parameters__ descriptor of a bare Python class.
+                pass
+            elif isinstance(t, tuple):
+                # `t` might be a tuple, when `ParamSpec` is substituted with
+                # `[T, int]`, or `[int, *Ts]`, etc.
+                for x in t:
+                    for collected in _collect_parameters([x]):
+                        if collected not in parameters:
+                            parameters.append(collected)
+            elif hasattr(t, '__typing_subst__'):
+                if t not in parameters:
+                    if enforce_default_ordering:
+                        has_default = (
+                            getattr(t, '__default__', NoDefault) is not NoDefault
+                        )
+
+                        if type_var_tuple_encountered and has_default:
+                            raise TypeError('Type parameter with a default'
+                                            ' follows TypeVarTuple')
+
+                        if has_default:
+                            default_encountered = True
+                        elif default_encountered:
+                            raise TypeError(f'Type parameter {t!r} without a default'
+                                            ' follows type parameter with a default')
+
+                    parameters.append(t)
+            else:
+                if _is_unpacked_typevartuple(t):
+                    type_var_tuple_encountered = True
+                for x in getattr(t, '__parameters__', ()):
+                    if x not in parameters:
+                        parameters.append(x)
+
+        return tuple(parameters)
+
+    if not _PEP_696_IMPLEMENTED:
+        typing._collect_parameters = _collect_parameters
+
+# Backport typing.NamedTuple as it exists in Python 3.13.
+# In 3.11, the ability to define generic `NamedTuple`s was supported.
+# This was explicitly disallowed in 3.9-3.10, and only half-worked in <=3.8.
+# On 3.12, we added __orig_bases__ to call-based NamedTuples
+# On 3.13, we deprecated kwargs-based NamedTuples
+if sys.version_info >= (3, 13):
+    NamedTuple = typing.NamedTuple
+else:
+    def _make_nmtuple(name, types, module, defaults=()):
+        fields = [n for n, t in types]
+        annotations = {n: typing._type_check(t, f"field {n} annotation must be a type")
+                       for n, t in types}
+        nm_tpl = collections.namedtuple(name, fields,
+                                        defaults=defaults, module=module)
+        nm_tpl.__annotations__ = nm_tpl.__new__.__annotations__ = annotations
+        # The `_field_types` attribute was removed in 3.9;
+        # in earlier versions, it is the same as the `__annotations__` attribute
+        if sys.version_info < (3, 9):
+            nm_tpl._field_types = annotations
+        return nm_tpl
+
+    _prohibited_namedtuple_fields = typing._prohibited
+    _special_namedtuple_fields = frozenset({'__module__', '__name__', '__annotations__'})
+
+    class _NamedTupleMeta(type):
+        def __new__(cls, typename, bases, ns):
+            assert _NamedTuple in bases
+            for base in bases:
+                if base is not _NamedTuple and base is not typing.Generic:
+                    raise TypeError(
+                        'can only inherit from a NamedTuple type and Generic')
+            bases = tuple(tuple if base is _NamedTuple else base for base in bases)
+            if "__annotations__" in ns:
+                types = ns["__annotations__"]
+            elif "__annotate__" in ns:
+                # TODO: Use inspect.VALUE here, and make the annotations lazily evaluated
+                types = ns["__annotate__"](1)
+            else:
+                types = {}
+            default_names = []
+            for field_name in types:
+                if field_name in ns:
+                    default_names.append(field_name)
+                elif default_names:
+                    raise TypeError(f"Non-default namedtuple field {field_name} "
+                                    f"cannot follow default field"
+                                    f"{'s' if len(default_names) > 1 else ''} "
+                                    f"{', '.join(default_names)}")
+            nm_tpl = _make_nmtuple(
+                typename, types.items(),
+                defaults=[ns[n] for n in default_names],
+                module=ns['__module__']
+            )
+            nm_tpl.__bases__ = bases
+            if typing.Generic in bases:
+                if hasattr(typing, '_generic_class_getitem'):  # 3.12+
+                    nm_tpl.__class_getitem__ = classmethod(typing._generic_class_getitem)
+                else:
+                    class_getitem = typing.Generic.__class_getitem__.__func__
+                    nm_tpl.__class_getitem__ = classmethod(class_getitem)
+            # update from user namespace without overriding special namedtuple attributes
+            for key, val in ns.items():
+                if key in _prohibited_namedtuple_fields:
+                    raise AttributeError("Cannot overwrite NamedTuple attribute " + key)
+                elif key not in _special_namedtuple_fields:
+                    if key not in nm_tpl._fields:
+                        setattr(nm_tpl, key, ns[key])
+                    try:
+                        set_name = type(val).__set_name__
+                    except AttributeError:
+                        pass
+                    else:
+                        try:
+                            set_name(val, nm_tpl, key)
+                        except BaseException as e:
+                            msg = (
+                                f"Error calling __set_name__ on {type(val).__name__!r} "
+                                f"instance {key!r} in {typename!r}"
+                            )
+                            # BaseException.add_note() existed on py311,
+                            # but the __set_name__ machinery didn't start
+                            # using add_note() until py312.
+                            # Making sure exceptions are raised in the same way
+                            # as in "normal" classes seems most important here.
+                            if sys.version_info >= (3, 12):
+                                e.add_note(msg)
+                                raise
+                            else:
+                                raise RuntimeError(msg) from e
+
+            if typing.Generic in bases:
+                nm_tpl.__init_subclass__()
+            return nm_tpl
+
+    _NamedTuple = type.__new__(_NamedTupleMeta, 'NamedTuple', (), {})
+
+    def _namedtuple_mro_entries(bases):
+        assert NamedTuple in bases
+        return (_NamedTuple,)
+
+    @_ensure_subclassable(_namedtuple_mro_entries)
+    def NamedTuple(typename, fields=_marker, /, **kwargs):
+        """Typed version of namedtuple.
+
+        Usage::
+
+            class Employee(NamedTuple):
+                name: str
+                id: int
+
+        This is equivalent to::
+
+            Employee = collections.namedtuple('Employee', ['name', 'id'])
+
+        The resulting class has an extra __annotations__ attribute, giving a
+        dict that maps field names to types.  (The field names are also in
+        the _fields attribute, which is part of the namedtuple API.)
+        An alternative equivalent functional syntax is also accepted::
+
+            Employee = NamedTuple('Employee', [('name', str), ('id', int)])
+        """
+        if fields is _marker:
+            if kwargs:
+                deprecated_thing = "Creating NamedTuple classes using keyword arguments"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "Use the class-based or functional syntax instead."
+                )
+            else:
+                deprecated_thing = "Failing to pass a value for the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif fields is None:
+            if kwargs:
+                raise TypeError(
+                    "Cannot pass `None` as the 'fields' parameter "
+                    "and also specify fields using keyword arguments"
+                )
+            else:
+                deprecated_thing = "Passing `None` as the 'fields' parameter"
+                example = f"`{typename} = NamedTuple({typename!r}, [])`"
+                deprecation_msg = (
+                    "{name} is deprecated and will be disallowed in Python {remove}. "
+                    "To create a NamedTuple class with 0 fields "
+                    "using the functional syntax, "
+                    "pass an empty list, e.g. "
+                ) + example + "."
+        elif kwargs:
+            raise TypeError("Either list of fields or keywords"
+                            " can be provided to NamedTuple, not both")
+        if fields is _marker or fields is None:
+            warnings.warn(
+                deprecation_msg.format(name=deprecated_thing, remove="3.15"),
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            fields = kwargs.items()
+        nt = _make_nmtuple(typename, fields, module=_caller())
+        nt.__orig_bases__ = (NamedTuple,)
+        return nt
+
+
+if hasattr(collections.abc, "Buffer"):
+    Buffer = collections.abc.Buffer
+else:
+    class Buffer(abc.ABC):  # noqa: B024
+        """Base class for classes that implement the buffer protocol.
+
+        The buffer protocol allows Python objects to expose a low-level
+        memory buffer interface. Before Python 3.12, it is not possible
+        to implement the buffer protocol in pure Python code, or even
+        to check whether a class implements the buffer protocol. In
+        Python 3.12 and higher, the ``__buffer__`` method allows access
+        to the buffer protocol from Python code, and the
+        ``collections.abc.Buffer`` ABC allows checking whether a class
+        implements the buffer protocol.
+
+        To indicate support for the buffer protocol in earlier versions,
+        inherit from this ABC, either in a stub file or at runtime,
+        or use ABC registration. This ABC provides no methods, because
+        there is no Python-accessible methods shared by pre-3.12 buffer
+        classes. It is useful primarily for static checks.
+
+        """
+
+    # As a courtesy, register the most common stdlib buffer classes.
+    Buffer.register(memoryview)
+    Buffer.register(bytearray)
+    Buffer.register(bytes)
+
+
+# Backport of types.get_original_bases, available on 3.12+ in CPython
+if hasattr(_types, "get_original_bases"):
+    get_original_bases = _types.get_original_bases
+else:
+    def get_original_bases(cls, /):
+        """Return the class's "original" bases prior to modification by `__mro_entries__`.
+
+        Examples::
+
+            from typing import TypeVar, Generic
+            from typing_extensions import NamedTuple, TypedDict
+
+            T = TypeVar("T")
+            class Foo(Generic[T]): ...
+            class Bar(Foo[int], float): ...
+            class Baz(list[str]): ...
+            Eggs = NamedTuple("Eggs", [("a", int), ("b", str)])
+            Spam = TypedDict("Spam", {"a": int, "b": str})
+
+            assert get_original_bases(Bar) == (Foo[int], float)
+            assert get_original_bases(Baz) == (list[str],)
+            assert get_original_bases(Eggs) == (NamedTuple,)
+            assert get_original_bases(Spam) == (TypedDict,)
+            assert get_original_bases(int) == (object,)
+        """
+        try:
+            return cls.__dict__.get("__orig_bases__", cls.__bases__)
+        except AttributeError:
+            raise TypeError(
+                f'Expected an instance of type, not {type(cls).__name__!r}'
+            ) from None
+
+
+# NewType is a class on Python 3.10+, making it pickleable
+# The error message for subclassing instances of NewType was improved on 3.11+
+if sys.version_info >= (3, 11):
+    NewType = typing.NewType
+else:
+    class NewType:
+        """NewType creates simple unique types with almost zero
+        runtime overhead. NewType(name, tp) is considered a subtype of tp
+        by static type checkers. At runtime, NewType(name, tp) returns
+        a dummy callable that simply returns its argument. Usage::
+            UserId = NewType('UserId', int)
+            def name_by_id(user_id: UserId) -> str:
+                ...
+            UserId('user')          # Fails type check
+            name_by_id(42)          # Fails type check
+            name_by_id(UserId(42))  # OK
+            num = UserId(5) + 1     # type: int
+        """
+
+        def __call__(self, obj, /):
+            return obj
+
+        def __init__(self, name, tp):
+            self.__qualname__ = name
+            if '.' in name:
+                name = name.rpartition('.')[-1]
+            self.__name__ = name
+            self.__supertype__ = tp
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+
+        def __mro_entries__(self, bases):
+            # We defined __mro_entries__ to get a better error message
+            # if a user attempts to subclass a NewType instance. bpo-46170
+            supercls_name = self.__name__
+
+            class Dummy:
+                def __init_subclass__(cls):
+                    subcls_name = cls.__name__
+                    raise TypeError(
+                        f"Cannot subclass an instance of NewType. "
+                        f"Perhaps you were looking for: "
+                        f"`{subcls_name} = NewType({subcls_name!r}, {supercls_name})`"
+                    )
+
+            return (Dummy,)
+
+        def __repr__(self):
+            return f'{self.__module__}.{self.__qualname__}'
+
+        def __reduce__(self):
+            return self.__qualname__
+
+        if sys.version_info >= (3, 10):
+            # PEP 604 methods
+            # It doesn't make sense to have these methods on Python <3.10
+
+            def __or__(self, other):
+                return typing.Union[self, other]
+
+            def __ror__(self, other):
+                return typing.Union[other, self]
+
+
+if hasattr(typing, "TypeAliasType"):
+    TypeAliasType = typing.TypeAliasType
+else:
+    def _is_unionable(obj):
+        """Corresponds to is_unionable() in unionobject.c in CPython."""
+        return obj is None or isinstance(obj, (
+            type,
+            _types.GenericAlias,
+            _types.UnionType,
+            TypeAliasType,
+        ))
+
+    class TypeAliasType:
+        """Create named, parameterized type aliases.
+
+        This provides a backport of the new `type` statement in Python 3.12:
+
+            type ListOrSet[T] = list[T] | set[T]
+
+        is equivalent to:
+
+            T = TypeVar("T")
+            ListOrSet = TypeAliasType("ListOrSet", list[T] | set[T], type_params=(T,))
+
+        The name ListOrSet can then be used as an alias for the type it refers to.
+
+        The type_params argument should contain all the type parameters used
+        in the value of the type alias. If the alias is not generic, this
+        argument is omitted.
+
+        Static type checkers should only support type aliases declared using
+        TypeAliasType that follow these rules:
+
+        - The first argument (the name) must be a string literal.
+        - The TypeAliasType instance must be immediately assigned to a variable
+          of the same name. (For example, 'X = TypeAliasType("Y", int)' is invalid,
+          as is 'X, Y = TypeAliasType("X", int), TypeAliasType("Y", int)').
+
+        """
+
+        def __init__(self, name: str, value, *, type_params=()):
+            if not isinstance(name, str):
+                raise TypeError("TypeAliasType name must be a string")
+            self.__value__ = value
+            self.__type_params__ = type_params
+
+            parameters = []
+            for type_param in type_params:
+                if isinstance(type_param, TypeVarTuple):
+                    parameters.extend(type_param)
+                else:
+                    parameters.append(type_param)
+            self.__parameters__ = tuple(parameters)
+            def_mod = _caller()
+            if def_mod != 'typing_extensions':
+                self.__module__ = def_mod
+            # Setting this attribute closes the TypeAliasType from further modification
+            self.__name__ = name
+
+        def __setattr__(self, name: str, value: object, /) -> None:
+            if hasattr(self, "__name__"):
+                self._raise_attribute_error(name)
+            super().__setattr__(name, value)
+
+        def __delattr__(self, name: str, /) -> Never:
+            self._raise_attribute_error(name)
+
+        def _raise_attribute_error(self, name: str) -> Never:
+            # Match the Python 3.12 error messages exactly
+            if name == "__name__":
+                raise AttributeError("readonly attribute")
+            elif name in {"__value__", "__type_params__", "__parameters__", "__module__"}:
+                raise AttributeError(
+                    f"attribute '{name}' of 'typing.TypeAliasType' objects "
+                    "is not writable"
+                )
+            else:
+                raise AttributeError(
+                    f"'typing.TypeAliasType' object has no attribute '{name}'"
+                )
+
+        def __repr__(self) -> str:
+            return self.__name__
+
+        def __getitem__(self, parameters):
+            if not isinstance(parameters, tuple):
+                parameters = (parameters,)
+            parameters = [
+                typing._type_check(
+                    item, f'Subscripting {self.__name__} requires a type.'
+                )
+                for item in parameters
+            ]
+            return typing._GenericAlias(self, tuple(parameters))
+
+        def __reduce__(self):
+            return self.__name__
+
+        def __init_subclass__(cls, *args, **kwargs):
+            raise TypeError(
+                "type 'typing_extensions.TypeAliasType' is not an acceptable base type"
+            )
+
+        # The presence of this method convinces typing._type_check
+        # that TypeAliasTypes are types.
+        def __call__(self):
+            raise TypeError("Type alias is not callable")
+
+        if sys.version_info >= (3, 10):
+            def __or__(self, right):
+                # For forward compatibility with 3.12, reject Unions
+                # that are not accepted by the built-in Union.
+                if not _is_unionable(right):
+                    return NotImplemented
+                return typing.Union[self, right]
+
+            def __ror__(self, left):
+                if not _is_unionable(left):
+                    return NotImplemented
+                return typing.Union[left, self]
+
+
+if hasattr(typing, "is_protocol"):
+    is_protocol = typing.is_protocol
+    get_protocol_members = typing.get_protocol_members
+else:
+    def is_protocol(tp: type, /) -> bool:
+        """Return True if the given type is a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, is_protocol
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> is_protocol(P)
+            True
+            >>> is_protocol(int)
+            False
+        """
+        return (
+            isinstance(tp, type)
+            and getattr(tp, '_is_protocol', False)
+            and tp is not Protocol
+            and tp is not typing.Protocol
+        )
+
+    def get_protocol_members(tp: type, /) -> typing.FrozenSet[str]:
+        """Return the set of members defined in a Protocol.
+
+        Example::
+
+            >>> from typing_extensions import Protocol, get_protocol_members
+            >>> class P(Protocol):
+            ...     def a(self) -> str: ...
+            ...     b: int
+            >>> get_protocol_members(P)
+            frozenset({'a', 'b'})
+
+        Raise a TypeError for arguments that are not Protocols.
+        """
+        if not is_protocol(tp):
+            raise TypeError(f'{tp!r} is not a Protocol')
+        if hasattr(tp, '__protocol_attrs__'):
+            return frozenset(tp.__protocol_attrs__)
+        return frozenset(_get_protocol_attrs(tp))
+
+
+if hasattr(typing, "Doc"):
+    Doc = typing.Doc
+else:
+    class Doc:
+        """Define the documentation of a type annotation using ``Annotated``, to be
+         used in class attributes, function and method parameters, return values,
+         and variables.
+
+        The value should be a positional-only string literal to allow static tools
+        like editors and documentation generators to use it.
+
+        This complements docstrings.
+
+        The string value passed is available in the attribute ``documentation``.
+
+        Example::
+
+            >>> from typing_extensions import Annotated, Doc
+            >>> def hi(to: Annotated[str, Doc("Who to say hi to")]) -> None: ...
+        """
+        def __init__(self, documentation: str, /) -> None:
+            self.documentation = documentation
+
+        def __repr__(self) -> str:
+            return f"Doc({self.documentation!r})"
+
+        def __hash__(self) -> int:
+            return hash(self.documentation)
+
+        def __eq__(self, other: object) -> bool:
+            if not isinstance(other, Doc):
+                return NotImplemented
+            return self.documentation == other.documentation
+
+
+_CapsuleType = getattr(_types, "CapsuleType", None)
+
+if _CapsuleType is None:
+    try:
+        import _socket
+    except ImportError:
+        pass
+    else:
+        _CAPI = getattr(_socket, "CAPI", None)
+        if _CAPI is not None:
+            _CapsuleType = type(_CAPI)
+
+if _CapsuleType is not None:
+    CapsuleType = _CapsuleType
+    __all__.append("CapsuleType")
+
+
+# Aliases for items that have always been in typing.
+# Explicitly assign these (rather than using `from typing import *` at the top),
+# so that we get a CI error if one of these is deleted from typing.py
+# in a future version of Python
+AbstractSet = typing.AbstractSet
+AnyStr = typing.AnyStr
+BinaryIO = typing.BinaryIO
+Callable = typing.Callable
+Collection = typing.Collection
+Container = typing.Container
+Dict = typing.Dict
+ForwardRef = typing.ForwardRef
+FrozenSet = typing.FrozenSet
+Generic = typing.Generic
+Hashable = typing.Hashable
+IO = typing.IO
+ItemsView = typing.ItemsView
+Iterable = typing.Iterable
+Iterator = typing.Iterator
+KeysView = typing.KeysView
+List = typing.List
+Mapping = typing.Mapping
+MappingView = typing.MappingView
+Match = typing.Match
+MutableMapping = typing.MutableMapping
+MutableSequence = typing.MutableSequence
+MutableSet = typing.MutableSet
+Optional = typing.Optional
+Pattern = typing.Pattern
+Reversible = typing.Reversible
+Sequence = typing.Sequence
+Set = typing.Set
+Sized = typing.Sized
+TextIO = typing.TextIO
+Tuple = typing.Tuple
+Union = typing.Union
+ValuesView = typing.ValuesView
+cast = typing.cast
+no_type_check = typing.no_type_check
+no_type_check_decorator = typing.no_type_check_decorator
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/METADATA
new file mode 100644
index 00000000..d2064b92
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/METADATA
@@ -0,0 +1,154 @@
+Metadata-Version: 2.4
+Name: urllib3
+Version: 2.3.0
+Summary: HTTP library with thread-safe connection pooling, file post, and more.
+Project-URL: Changelog, https://github.com/urllib3/urllib3/blob/main/CHANGES.rst
+Project-URL: Documentation, https://urllib3.readthedocs.io
+Project-URL: Code, https://github.com/urllib3/urllib3
+Project-URL: Issue tracker, https://github.com/urllib3/urllib3/issues
+Author-email: Andrey Petrov <andrey.petrov@shazow.net>
+Maintainer-email: Seth Michael Larson <sethmichaellarson@gmail.com>, Quentin Pradet <quentin@pradet.me>, Illia Volochii <illia.volochii@gmail.com>
+License-File: LICENSE.txt
+Keywords: filepost,http,httplib,https,pooling,ssl,threadsafe,urllib
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: Implementation :: PyPy
+Classifier: Topic :: Internet :: WWW/HTTP
+Classifier: Topic :: Software Development :: Libraries
+Requires-Python: >=3.9
+Provides-Extra: brotli
+Requires-Dist: brotli>=1.0.9; (platform_python_implementation == 'CPython') and extra == 'brotli'
+Requires-Dist: brotlicffi>=0.8.0; (platform_python_implementation != 'CPython') and extra == 'brotli'
+Provides-Extra: h2
+Requires-Dist: h2<5,>=4; extra == 'h2'
+Provides-Extra: socks
+Requires-Dist: pysocks!=1.5.7,<2.0,>=1.5.6; extra == 'socks'
+Provides-Extra: zstd
+Requires-Dist: zstandard>=0.18.0; extra == 'zstd'
+Description-Content-Type: text/markdown
+
+<h1 align="center">
+
+![urllib3](https://github.com/urllib3/urllib3/raw/main/docs/_static/banner_github.svg)
+
+</h1>
+
+<p align="center">
+  <a href="https://pypi.org/project/urllib3"><img alt="PyPI Version" src="https://img.shields.io/pypi/v/urllib3.svg?maxAge=86400" /></a>
+  <a href="https://pypi.org/project/urllib3"><img alt="Python Versions" src="https://img.shields.io/pypi/pyversions/urllib3.svg?maxAge=86400" /></a>
+  <a href="https://discord.gg/urllib3"><img alt="Join our Discord" src="https://img.shields.io/discord/756342717725933608?color=%237289da&label=discord" /></a>
+  <a href="https://github.com/urllib3/urllib3/actions?query=workflow%3ACI"><img alt="Coverage Status" src="https://img.shields.io/badge/coverage-100%25-success" /></a>
+  <a href="https://github.com/urllib3/urllib3/actions/workflows/ci.yml?query=branch%3Amain"><img alt="Build Status on GitHub" src="https://github.com/urllib3/urllib3/actions/workflows/ci.yml/badge.svg?branch:main&workflow:CI" /></a>
+  <a href="https://urllib3.readthedocs.io"><img alt="Documentation Status" src="https://readthedocs.org/projects/urllib3/badge/?version=latest" /></a><br>
+  <a href="https://deps.dev/pypi/urllib3"><img alt="OpenSSF Scorecard" src="https://api.securityscorecards.dev/projects/github.com/urllib3/urllib3/badge" /></a>
+  <a href="https://slsa.dev"><img alt="SLSA 3" src="https://slsa.dev/images/gh-badge-level3.svg" /></a>
+  <a href="https://bestpractices.coreinfrastructure.org/projects/6227"><img alt="CII Best Practices" src="https://bestpractices.coreinfrastructure.org/projects/6227/badge" /></a>
+</p>
+
+urllib3 is a powerful, *user-friendly* HTTP client for Python. Much of the
+Python ecosystem already uses urllib3 and you should too.
+urllib3 brings many critical features that are missing from the Python
+standard libraries:
+
+- Thread safety.
+- Connection pooling.
+- Client-side SSL/TLS verification.
+- File uploads with multipart encoding.
+- Helpers for retrying requests and dealing with HTTP redirects.
+- Support for gzip, deflate, brotli, and zstd encoding.
+- Proxy support for HTTP and SOCKS.
+- 100% test coverage.
+
+urllib3 is powerful and easy to use:
+
+```python3
+>>> import urllib3
+>>> resp = urllib3.request("GET", "http://httpbin.org/robots.txt")
+>>> resp.status
+200
+>>> resp.data
+b"User-agent: *\nDisallow: /deny\n"
+```
+
+## Installing
+
+urllib3 can be installed with [pip](https://pip.pypa.io):
+
+```bash
+$ python -m pip install urllib3
+```
+
+Alternatively, you can grab the latest source code from [GitHub](https://github.com/urllib3/urllib3):
+
+```bash
+$ git clone https://github.com/urllib3/urllib3.git
+$ cd urllib3
+$ pip install .
+```
+
+
+## Documentation
+
+urllib3 has usage and reference documentation at [urllib3.readthedocs.io](https://urllib3.readthedocs.io).
+
+
+## Community
+
+urllib3 has a [community Discord channel](https://discord.gg/urllib3) for asking questions and
+collaborating with other contributors. Drop by and say hello 👋
+
+
+## Contributing
+
+urllib3 happily accepts contributions. Please see our
+[contributing documentation](https://urllib3.readthedocs.io/en/latest/contributing.html)
+for some tips on getting started.
+
+
+## Security Disclosures
+
+To report a security vulnerability, please use the
+[Tidelift security contact](https://tidelift.com/security).
+Tidelift will coordinate the fix and disclosure with maintainers.
+
+
+## Maintainers
+
+- [@sethmlarson](https://github.com/sethmlarson) (Seth M. Larson)
+- [@pquentin](https://github.com/pquentin) (Quentin Pradet)
+- [@illia-v](https://github.com/illia-v) (Illia Volochii)
+- [@theacodes](https://github.com/theacodes) (Thea Flowers)
+- [@haikuginger](https://github.com/haikuginger) (Jess Shapiro)
+- [@lukasa](https://github.com/lukasa) (Cory Benfield)
+- [@sigmavirus24](https://github.com/sigmavirus24) (Ian Stapleton Cordasco)
+- [@shazow](https://github.com/shazow) (Andrey Petrov)
+
+👋
+
+
+## Sponsorship
+
+If your company benefits from this library, please consider [sponsoring its
+development](https://urllib3.readthedocs.io/en/latest/sponsors.html).
+
+
+## For Enterprise
+
+Professional support for urllib3 is available as part of the [Tidelift
+Subscription][1].  Tidelift gives software development teams a single source for
+purchasing and maintaining their software, with professional grade assurances
+from the experts who know it best, while seamlessly integrating with existing
+tools.
+
+[1]: https://tidelift.com/subscription/pkg/pypi-urllib3?utm_source=pypi-urllib3&utm_medium=referral&utm_campaign=readme
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/RECORD
new file mode 100644
index 00000000..e5b69cdb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/RECORD
@@ -0,0 +1,79 @@
+urllib3-2.3.0.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+urllib3-2.3.0.dist-info/METADATA,sha256=quSdNDoWoaG7idKpqr1_2N52fUBfxTtifpP3sofNwAY,6488
+urllib3-2.3.0.dist-info/RECORD,,
+urllib3-2.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+urllib3-2.3.0.dist-info/licenses/LICENSE.txt,sha256=Ew46ZNX91dCWp1JpRjSn2d8oRGnehuVzIQAmgEHj1oY,1093
+urllib3/__init__.py,sha256=JMo1tg1nIV1AeJ2vENC_Txfl0e5h6Gzl9DGVk1rWRbo,6979
+urllib3/__pycache__/__init__.cpython-312.pyc,,
+urllib3/__pycache__/_base_connection.cpython-312.pyc,,
+urllib3/__pycache__/_collections.cpython-312.pyc,,
+urllib3/__pycache__/_request_methods.cpython-312.pyc,,
+urllib3/__pycache__/_version.cpython-312.pyc,,
+urllib3/__pycache__/connection.cpython-312.pyc,,
+urllib3/__pycache__/connectionpool.cpython-312.pyc,,
+urllib3/__pycache__/exceptions.cpython-312.pyc,,
+urllib3/__pycache__/fields.cpython-312.pyc,,
+urllib3/__pycache__/filepost.cpython-312.pyc,,
+urllib3/__pycache__/poolmanager.cpython-312.pyc,,
+urllib3/__pycache__/response.cpython-312.pyc,,
+urllib3/_base_connection.py,sha256=T1cwH3RhzsrBh6Bz3AOGVDboRsE7veijqZPXXQTR2Rg,5568
+urllib3/_collections.py,sha256=tM7c6J1iKtWZYV_QGYb8-r7Nr1524Dehnsa0Ufh6_mU,17295
+urllib3/_request_methods.py,sha256=gCeF85SO_UU4WoPwYHIoz_tw-eM_EVOkLFp8OFsC7DA,9931
+urllib3/_version.py,sha256=ChsIHG8bRc-eXUbXOgv4Fm4DstSKLq9FpsTAsaMeR08,411
+urllib3/connection.py,sha256=dsVIUaPrOdATuO9OGnF5GzMhlVKlAw-2qH9FWYuic4s,39875
+urllib3/connectionpool.py,sha256=ZEhudsa8BIubD2M0XoxBBsjxbsXwMgUScH7oQ9i-j1Y,43371
+urllib3/contrib/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+urllib3/contrib/__pycache__/__init__.cpython-312.pyc,,
+urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc,,
+urllib3/contrib/__pycache__/socks.cpython-312.pyc,,
+urllib3/contrib/emscripten/__init__.py,sha256=u6KNgzjlFZbuAAXa_ybCR7gQ71VJESnF-IIdDA73brw,733
+urllib3/contrib/emscripten/__pycache__/__init__.cpython-312.pyc,,
+urllib3/contrib/emscripten/__pycache__/connection.cpython-312.pyc,,
+urllib3/contrib/emscripten/__pycache__/fetch.cpython-312.pyc,,
+urllib3/contrib/emscripten/__pycache__/request.cpython-312.pyc,,
+urllib3/contrib/emscripten/__pycache__/response.cpython-312.pyc,,
+urllib3/contrib/emscripten/connection.py,sha256=j8DR_flE7hsoFhNfiqHLiaPaCsVbzG44jgahwvsQ52A,8771
+urllib3/contrib/emscripten/emscripten_fetch_worker.js,sha256=CDfYF_9CDobtx2lGidyJ1zjDEvwNT5F-dchmVWXDh0E,3655
+urllib3/contrib/emscripten/fetch.py,sha256=Li6sUnFuFog0fBiahk1Y0C0tdn5fxmj4ucDofPWFiXc,22867
+urllib3/contrib/emscripten/request.py,sha256=mL28szy1KvE3NJhWor5jNmarp8gwplDU-7gwGZY5g0Q,566
+urllib3/contrib/emscripten/response.py,sha256=sc0CJCEV_3t_HTadOmvWNuO-WkYbI1YfQWnBBmaeriE,9981
+urllib3/contrib/pyopenssl.py,sha256=uMNfy0e-wT-WNrTF-4oOQ17-I3w0NPrREm1t5AC9wxg,19398
+urllib3/contrib/socks.py,sha256=-iardc61GypsJzD6W6yuRS7KVCyfowcQrl_719H7lIM,7549
+urllib3/exceptions.py,sha256=VSkjQkvw3iolMKP_ZWfiiXumAj2VCvhmz2B3W-MP4BA,9633
+urllib3/fields.py,sha256=FCf7UULSkf10cuTRUWTQESzxgl1WT8e2aCy3kfyZins,10829
+urllib3/filepost.py,sha256=U8eNZ-mpKKHhrlbHEEiTxxgK16IejhEa7uz42yqA_dI,2388
+urllib3/http2/__init__.py,sha256=xzrASH7R5ANRkPJOot5lGnATOq3KKuyXzI42rcnwmqs,1741
+urllib3/http2/__pycache__/__init__.cpython-312.pyc,,
+urllib3/http2/__pycache__/connection.cpython-312.pyc,,
+urllib3/http2/__pycache__/probe.cpython-312.pyc,,
+urllib3/http2/connection.py,sha256=GNlp9BjI3DmfSKe1W0b9IqRBeM8Q13xd2MA3ROcJ3dY,12668
+urllib3/http2/probe.py,sha256=nnAkqbhAakOiF75rz7W0udZ38Eeh_uD8fjV74N73FEI,3014
+urllib3/poolmanager.py,sha256=2_L2AjVDgoQ0qBmYbX9u9QqyU1u5J37zQbtv_-ueZQA,22913
+urllib3/py.typed,sha256=UaCuPFa3H8UAakbt-5G8SPacldTOGvJv18pPjUJ5gDY,93
+urllib3/response.py,sha256=PBW5ZnK3kYeq0fqeUYywyASKQWK7uRzSa-HgDBzZedU,45190
+urllib3/util/__init__.py,sha256=-qeS0QceivazvBEKDNFCAI-6ACcdDOE4TMvo7SLNlAQ,1001
+urllib3/util/__pycache__/__init__.cpython-312.pyc,,
+urllib3/util/__pycache__/connection.cpython-312.pyc,,
+urllib3/util/__pycache__/proxy.cpython-312.pyc,,
+urllib3/util/__pycache__/request.cpython-312.pyc,,
+urllib3/util/__pycache__/response.cpython-312.pyc,,
+urllib3/util/__pycache__/retry.cpython-312.pyc,,
+urllib3/util/__pycache__/ssl_.cpython-312.pyc,,
+urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc,,
+urllib3/util/__pycache__/ssltransport.cpython-312.pyc,,
+urllib3/util/__pycache__/timeout.cpython-312.pyc,,
+urllib3/util/__pycache__/url.cpython-312.pyc,,
+urllib3/util/__pycache__/util.cpython-312.pyc,,
+urllib3/util/__pycache__/wait.cpython-312.pyc,,
+urllib3/util/connection.py,sha256=JjO722lzHlzLXPTkr9ZWBdhseXnMVjMSb1DJLVrXSnQ,4444
+urllib3/util/proxy.py,sha256=seP8-Q5B6bB0dMtwPj-YcZZQ30vHuLqRu-tI0JZ2fzs,1148
+urllib3/util/request.py,sha256=qSwxEsJJ-9DUfFDEAZIgQe8sgyywKY0o3faH3ixi3i8,8218
+urllib3/util/response.py,sha256=vQE639uoEhj1vpjEdxu5lNIhJCSUZkd7pqllUI0BZOA,3374
+urllib3/util/retry.py,sha256=bj-2YUqblxLlv8THg5fxww-DM54XCbjgZXIQ71XioCY,18459
+urllib3/util/ssl_.py,sha256=CcYPnFKqJDn58Us2J5AFBQLMthIVVMjb69HwvdmPgoQ,18884
+urllib3/util/ssl_match_hostname.py,sha256=gaWqixoYtQ_GKO8fcRGFj3VXeMoqyxQQuUTPgWeiL_M,5812
+urllib3/util/ssltransport.py,sha256=Ez4O8pR_vT8dan_FvqBYS6dgDfBXEMfVfrzcdUoWfi4,8847
+urllib3/util/timeout.py,sha256=4eT1FVeZZU7h7mYD1Jq2OXNe4fxekdNvhoWUkZusRpA,10346
+urllib3/util/url.py,sha256=WRh-TMYXosmgp8m8lT4H5spoHw5yUjlcMCfU53AkoAs,15205
+urllib3/util/util.py,sha256=j3lbZK1jPyiwD34T8IgJzdWEZVT-4E-0vYIJi9UjeNA,1146
+urllib3/util/wait.py,sha256=_ph8IrUR3sqPqi0OopQgJUlH4wzkGeM5CiyA7XGGtmI,4423
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/WHEEL
new file mode 100644
index 00000000..12228d41
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.27.0
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/licenses/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/licenses/LICENSE.txt
new file mode 100644
index 00000000..e6183d02
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3-2.3.0.dist-info/licenses/LICENSE.txt
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2008-2020 Andrey Petrov and contributors.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__init__.py b/backend/venv/lib/python3.12/site-packages/urllib3/__init__.py
new file mode 100644
index 00000000..3fe782c8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/__init__.py
@@ -0,0 +1,211 @@
+"""
+Python HTTP library with thread-safe connection pooling, file post support, user friendly, and more
+"""
+
+from __future__ import annotations
+
+# Set default logging handler to avoid "No handler found" warnings.
+import logging
+import sys
+import typing
+import warnings
+from logging import NullHandler
+
+from . import exceptions
+from ._base_connection import _TYPE_BODY
+from ._collections import HTTPHeaderDict
+from ._version import __version__
+from .connectionpool import HTTPConnectionPool, HTTPSConnectionPool, connection_from_url
+from .filepost import _TYPE_FIELDS, encode_multipart_formdata
+from .poolmanager import PoolManager, ProxyManager, proxy_from_url
+from .response import BaseHTTPResponse, HTTPResponse
+from .util.request import make_headers
+from .util.retry import Retry
+from .util.timeout import Timeout
+
+# Ensure that Python is compiled with OpenSSL 1.1.1+
+# If the 'ssl' module isn't available at all that's
+# fine, we only care if the module is available.
+try:
+    import ssl
+except ImportError:
+    pass
+else:
+    if not ssl.OPENSSL_VERSION.startswith("OpenSSL "):  # Defensive:
+        warnings.warn(
+            "urllib3 v2 only supports OpenSSL 1.1.1+, currently "
+            f"the 'ssl' module is compiled with {ssl.OPENSSL_VERSION!r}. "
+            "See: https://github.com/urllib3/urllib3/issues/3020",
+            exceptions.NotOpenSSLWarning,
+        )
+    elif ssl.OPENSSL_VERSION_INFO < (1, 1, 1):  # Defensive:
+        raise ImportError(
+            "urllib3 v2 only supports OpenSSL 1.1.1+, currently "
+            f"the 'ssl' module is compiled with {ssl.OPENSSL_VERSION!r}. "
+            "See: https://github.com/urllib3/urllib3/issues/2168"
+        )
+
+__author__ = "Andrey Petrov (andrey.petrov@shazow.net)"
+__license__ = "MIT"
+__version__ = __version__
+
+__all__ = (
+    "HTTPConnectionPool",
+    "HTTPHeaderDict",
+    "HTTPSConnectionPool",
+    "PoolManager",
+    "ProxyManager",
+    "HTTPResponse",
+    "Retry",
+    "Timeout",
+    "add_stderr_logger",
+    "connection_from_url",
+    "disable_warnings",
+    "encode_multipart_formdata",
+    "make_headers",
+    "proxy_from_url",
+    "request",
+    "BaseHTTPResponse",
+)
+
+logging.getLogger(__name__).addHandler(NullHandler())
+
+
+def add_stderr_logger(
+    level: int = logging.DEBUG,
+) -> logging.StreamHandler[typing.TextIO]:
+    """
+    Helper for quickly adding a StreamHandler to the logger. Useful for
+    debugging.
+
+    Returns the handler after adding it.
+    """
+    # This method needs to be in this __init__.py to get the __name__ correct
+    # even if urllib3 is vendored within another package.
+    logger = logging.getLogger(__name__)
+    handler = logging.StreamHandler()
+    handler.setFormatter(logging.Formatter("%(asctime)s %(levelname)s %(message)s"))
+    logger.addHandler(handler)
+    logger.setLevel(level)
+    logger.debug("Added a stderr logging handler to logger: %s", __name__)
+    return handler
+
+
+# ... Clean up.
+del NullHandler
+
+
+# All warning filters *must* be appended unless you're really certain that they
+# shouldn't be: otherwise, it's very hard for users to use most Python
+# mechanisms to silence them.
+# SecurityWarning's always go off by default.
+warnings.simplefilter("always", exceptions.SecurityWarning, append=True)
+# InsecurePlatformWarning's don't vary between requests, so we keep it default.
+warnings.simplefilter("default", exceptions.InsecurePlatformWarning, append=True)
+
+
+def disable_warnings(category: type[Warning] = exceptions.HTTPWarning) -> None:
+    """
+    Helper for quickly disabling all urllib3 warnings.
+    """
+    warnings.simplefilter("ignore", category)
+
+
+_DEFAULT_POOL = PoolManager()
+
+
+def request(
+    method: str,
+    url: str,
+    *,
+    body: _TYPE_BODY | None = None,
+    fields: _TYPE_FIELDS | None = None,
+    headers: typing.Mapping[str, str] | None = None,
+    preload_content: bool | None = True,
+    decode_content: bool | None = True,
+    redirect: bool | None = True,
+    retries: Retry | bool | int | None = None,
+    timeout: Timeout | float | int | None = 3,
+    json: typing.Any | None = None,
+) -> BaseHTTPResponse:
+    """
+    A convenience, top-level request method. It uses a module-global ``PoolManager`` instance.
+    Therefore, its side effects could be shared across dependencies relying on it.
+    To avoid side effects create a new ``PoolManager`` instance and use it instead.
+    The method does not accept low-level ``**urlopen_kw`` keyword arguments.
+
+    :param method:
+        HTTP request method (such as GET, POST, PUT, etc.)
+
+    :param url:
+        The URL to perform the request on.
+
+    :param body:
+        Data to send in the request body, either :class:`str`, :class:`bytes`,
+        an iterable of :class:`str`/:class:`bytes`, or a file-like object.
+
+    :param fields:
+        Data to encode and send in the request body.
+
+    :param headers:
+        Dictionary of custom headers to send, such as User-Agent,
+        If-None-Match, etc.
+
+    :param bool preload_content:
+        If True, the response's body will be preloaded into memory.
+
+    :param bool decode_content:
+        If True, will attempt to decode the body based on the
+        'content-encoding' header.
+
+    :param redirect:
+        If True, automatically handle redirects (status codes 301, 302,
+        303, 307, 308). Each redirect counts as a retry. Disabling retries
+        will disable redirect, too.
+
+    :param retries:
+        Configure the number of retries to allow before raising a
+        :class:`~urllib3.exceptions.MaxRetryError` exception.
+
+        If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a
+        :class:`~urllib3.util.retry.Retry` object for fine-grained control
+        over different types of retries.
+        Pass an integer number to retry connection errors that many times,
+        but no other types of errors. Pass zero to never retry.
+
+        If ``False``, then retries are disabled and any exception is raised
+        immediately. Also, instead of raising a MaxRetryError on redirects,
+        the redirect response will be returned.
+
+    :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int.
+
+    :param timeout:
+        If specified, overrides the default timeout for this one
+        request. It may be a float (in seconds) or an instance of
+        :class:`urllib3.util.Timeout`.
+
+    :param json:
+        Data to encode and send as JSON with UTF-encoded in the request body.
+        The ``"Content-Type"`` header will be set to ``"application/json"``
+        unless specified otherwise.
+    """
+
+    return _DEFAULT_POOL.request(
+        method,
+        url,
+        body=body,
+        fields=fields,
+        headers=headers,
+        preload_content=preload_content,
+        decode_content=decode_content,
+        redirect=redirect,
+        retries=retries,
+        timeout=timeout,
+        json=json,
+    )
+
+
+if sys.platform == "emscripten":
+    from .contrib.emscripten import inject_into_urllib3  # noqa: 401
+
+    inject_into_urllib3()
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..f42d2770
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_base_connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_base_connection.cpython-312.pyc
new file mode 100644
index 00000000..06894517
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_base_connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_collections.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_collections.cpython-312.pyc
new file mode 100644
index 00000000..0a61e7af
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_collections.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_request_methods.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_request_methods.cpython-312.pyc
new file mode 100644
index 00000000..6ec9ab94
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_request_methods.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_version.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_version.cpython-312.pyc
new file mode 100644
index 00000000..355a9b71
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/_version.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..17f4c6fe
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connectionpool.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connectionpool.cpython-312.pyc
new file mode 100644
index 00000000..35d55ccc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/connectionpool.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..2312c442
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/fields.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/fields.cpython-312.pyc
new file mode 100644
index 00000000..1d57dea5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/fields.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/filepost.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/filepost.cpython-312.pyc
new file mode 100644
index 00000000..669ce98c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/filepost.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/poolmanager.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/poolmanager.cpython-312.pyc
new file mode 100644
index 00000000..50f512ae
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/poolmanager.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..5a84f949
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/_base_connection.py b/backend/venv/lib/python3.12/site-packages/urllib3/_base_connection.py
new file mode 100644
index 00000000..dc0f318c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/_base_connection.py
@@ -0,0 +1,165 @@
+from __future__ import annotations
+
+import typing
+
+from .util.connection import _TYPE_SOCKET_OPTIONS
+from .util.timeout import _DEFAULT_TIMEOUT, _TYPE_TIMEOUT
+from .util.url import Url
+
+_TYPE_BODY = typing.Union[bytes, typing.IO[typing.Any], typing.Iterable[bytes], str]
+
+
+class ProxyConfig(typing.NamedTuple):
+    ssl_context: ssl.SSLContext | None
+    use_forwarding_for_https: bool
+    assert_hostname: None | str | typing.Literal[False]
+    assert_fingerprint: str | None
+
+
+class _ResponseOptions(typing.NamedTuple):
+    # TODO: Remove this in favor of a better
+    # HTTP request/response lifecycle tracking.
+    request_method: str
+    request_url: str
+    preload_content: bool
+    decode_content: bool
+    enforce_content_length: bool
+
+
+if typing.TYPE_CHECKING:
+    import ssl
+    from typing import Protocol
+
+    from .response import BaseHTTPResponse
+
+    class BaseHTTPConnection(Protocol):
+        default_port: typing.ClassVar[int]
+        default_socket_options: typing.ClassVar[_TYPE_SOCKET_OPTIONS]
+
+        host: str
+        port: int
+        timeout: None | (
+            float
+        )  # Instance doesn't store _DEFAULT_TIMEOUT, must be resolved.
+        blocksize: int
+        source_address: tuple[str, int] | None
+        socket_options: _TYPE_SOCKET_OPTIONS | None
+
+        proxy: Url | None
+        proxy_config: ProxyConfig | None
+
+        is_verified: bool
+        proxy_is_verified: bool | None
+
+        def __init__(
+            self,
+            host: str,
+            port: int | None = None,
+            *,
+            timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+            source_address: tuple[str, int] | None = None,
+            blocksize: int = 8192,
+            socket_options: _TYPE_SOCKET_OPTIONS | None = ...,
+            proxy: Url | None = None,
+            proxy_config: ProxyConfig | None = None,
+        ) -> None: ...
+
+        def set_tunnel(
+            self,
+            host: str,
+            port: int | None = None,
+            headers: typing.Mapping[str, str] | None = None,
+            scheme: str = "http",
+        ) -> None: ...
+
+        def connect(self) -> None: ...
+
+        def request(
+            self,
+            method: str,
+            url: str,
+            body: _TYPE_BODY | None = None,
+            headers: typing.Mapping[str, str] | None = None,
+            # We know *at least* botocore is depending on the order of the
+            # first 3 parameters so to be safe we only mark the later ones
+            # as keyword-only to ensure we have space to extend.
+            *,
+            chunked: bool = False,
+            preload_content: bool = True,
+            decode_content: bool = True,
+            enforce_content_length: bool = True,
+        ) -> None: ...
+
+        def getresponse(self) -> BaseHTTPResponse: ...
+
+        def close(self) -> None: ...
+
+        @property
+        def is_closed(self) -> bool:
+            """Whether the connection either is brand new or has been previously closed.
+            If this property is True then both ``is_connected`` and ``has_connected_to_proxy``
+            properties must be False.
+            """
+
+        @property
+        def is_connected(self) -> bool:
+            """Whether the connection is actively connected to any origin (proxy or target)"""
+
+        @property
+        def has_connected_to_proxy(self) -> bool:
+            """Whether the connection has successfully connected to its proxy.
+            This returns False if no proxy is in use. Used to determine whether
+            errors are coming from the proxy layer or from tunnelling to the target origin.
+            """
+
+    class BaseHTTPSConnection(BaseHTTPConnection, Protocol):
+        default_port: typing.ClassVar[int]
+        default_socket_options: typing.ClassVar[_TYPE_SOCKET_OPTIONS]
+
+        # Certificate verification methods
+        cert_reqs: int | str | None
+        assert_hostname: None | str | typing.Literal[False]
+        assert_fingerprint: str | None
+        ssl_context: ssl.SSLContext | None
+
+        # Trusted CAs
+        ca_certs: str | None
+        ca_cert_dir: str | None
+        ca_cert_data: None | str | bytes
+
+        # TLS version
+        ssl_minimum_version: int | None
+        ssl_maximum_version: int | None
+        ssl_version: int | str | None  # Deprecated
+
+        # Client certificates
+        cert_file: str | None
+        key_file: str | None
+        key_password: str | None
+
+        def __init__(
+            self,
+            host: str,
+            port: int | None = None,
+            *,
+            timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+            source_address: tuple[str, int] | None = None,
+            blocksize: int = 16384,
+            socket_options: _TYPE_SOCKET_OPTIONS | None = ...,
+            proxy: Url | None = None,
+            proxy_config: ProxyConfig | None = None,
+            cert_reqs: int | str | None = None,
+            assert_hostname: None | str | typing.Literal[False] = None,
+            assert_fingerprint: str | None = None,
+            server_hostname: str | None = None,
+            ssl_context: ssl.SSLContext | None = None,
+            ca_certs: str | None = None,
+            ca_cert_dir: str | None = None,
+            ca_cert_data: None | str | bytes = None,
+            ssl_minimum_version: int | None = None,
+            ssl_maximum_version: int | None = None,
+            ssl_version: int | str | None = None,  # Deprecated
+            cert_file: str | None = None,
+            key_file: str | None = None,
+            key_password: str | None = None,
+        ) -> None: ...
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/_collections.py b/backend/venv/lib/python3.12/site-packages/urllib3/_collections.py
new file mode 100644
index 00000000..1b6c1364
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/_collections.py
@@ -0,0 +1,479 @@
+from __future__ import annotations
+
+import typing
+from collections import OrderedDict
+from enum import Enum, auto
+from threading import RLock
+
+if typing.TYPE_CHECKING:
+    # We can only import Protocol if TYPE_CHECKING because it's a development
+    # dependency, and is not available at runtime.
+    from typing import Protocol
+
+    from typing_extensions import Self
+
+    class HasGettableStringKeys(Protocol):
+        def keys(self) -> typing.Iterator[str]: ...
+
+        def __getitem__(self, key: str) -> str: ...
+
+
+__all__ = ["RecentlyUsedContainer", "HTTPHeaderDict"]
+
+
+# Key type
+_KT = typing.TypeVar("_KT")
+# Value type
+_VT = typing.TypeVar("_VT")
+# Default type
+_DT = typing.TypeVar("_DT")
+
+ValidHTTPHeaderSource = typing.Union[
+    "HTTPHeaderDict",
+    typing.Mapping[str, str],
+    typing.Iterable[tuple[str, str]],
+    "HasGettableStringKeys",
+]
+
+
+class _Sentinel(Enum):
+    not_passed = auto()
+
+
+def ensure_can_construct_http_header_dict(
+    potential: object,
+) -> ValidHTTPHeaderSource | None:
+    if isinstance(potential, HTTPHeaderDict):
+        return potential
+    elif isinstance(potential, typing.Mapping):
+        # Full runtime checking of the contents of a Mapping is expensive, so for the
+        # purposes of typechecking, we assume that any Mapping is the right shape.
+        return typing.cast(typing.Mapping[str, str], potential)
+    elif isinstance(potential, typing.Iterable):
+        # Similarly to Mapping, full runtime checking of the contents of an Iterable is
+        # expensive, so for the purposes of typechecking, we assume that any Iterable
+        # is the right shape.
+        return typing.cast(typing.Iterable[tuple[str, str]], potential)
+    elif hasattr(potential, "keys") and hasattr(potential, "__getitem__"):
+        return typing.cast("HasGettableStringKeys", potential)
+    else:
+        return None
+
+
+class RecentlyUsedContainer(typing.Generic[_KT, _VT], typing.MutableMapping[_KT, _VT]):
+    """
+    Provides a thread-safe dict-like container which maintains up to
+    ``maxsize`` keys while throwing away the least-recently-used keys beyond
+    ``maxsize``.
+
+    :param maxsize:
+        Maximum number of recent elements to retain.
+
+    :param dispose_func:
+        Every time an item is evicted from the container,
+        ``dispose_func(value)`` is called.  Callback which will get called
+    """
+
+    _container: typing.OrderedDict[_KT, _VT]
+    _maxsize: int
+    dispose_func: typing.Callable[[_VT], None] | None
+    lock: RLock
+
+    def __init__(
+        self,
+        maxsize: int = 10,
+        dispose_func: typing.Callable[[_VT], None] | None = None,
+    ) -> None:
+        super().__init__()
+        self._maxsize = maxsize
+        self.dispose_func = dispose_func
+        self._container = OrderedDict()
+        self.lock = RLock()
+
+    def __getitem__(self, key: _KT) -> _VT:
+        # Re-insert the item, moving it to the end of the eviction line.
+        with self.lock:
+            item = self._container.pop(key)
+            self._container[key] = item
+            return item
+
+    def __setitem__(self, key: _KT, value: _VT) -> None:
+        evicted_item = None
+        with self.lock:
+            # Possibly evict the existing value of 'key'
+            try:
+                # If the key exists, we'll overwrite it, which won't change the
+                # size of the pool. Because accessing a key should move it to
+                # the end of the eviction line, we pop it out first.
+                evicted_item = key, self._container.pop(key)
+                self._container[key] = value
+            except KeyError:
+                # When the key does not exist, we insert the value first so that
+                # evicting works in all cases, including when self._maxsize is 0
+                self._container[key] = value
+                if len(self._container) > self._maxsize:
+                    # If we didn't evict an existing value, and we've hit our maximum
+                    # size, then we have to evict the least recently used item from
+                    # the beginning of the container.
+                    evicted_item = self._container.popitem(last=False)
+
+        # After releasing the lock on the pool, dispose of any evicted value.
+        if evicted_item is not None and self.dispose_func:
+            _, evicted_value = evicted_item
+            self.dispose_func(evicted_value)
+
+    def __delitem__(self, key: _KT) -> None:
+        with self.lock:
+            value = self._container.pop(key)
+
+        if self.dispose_func:
+            self.dispose_func(value)
+
+    def __len__(self) -> int:
+        with self.lock:
+            return len(self._container)
+
+    def __iter__(self) -> typing.NoReturn:
+        raise NotImplementedError(
+            "Iteration over this class is unlikely to be threadsafe."
+        )
+
+    def clear(self) -> None:
+        with self.lock:
+            # Copy pointers to all values, then wipe the mapping
+            values = list(self._container.values())
+            self._container.clear()
+
+        if self.dispose_func:
+            for value in values:
+                self.dispose_func(value)
+
+    def keys(self) -> set[_KT]:  # type: ignore[override]
+        with self.lock:
+            return set(self._container.keys())
+
+
+class HTTPHeaderDictItemView(set[tuple[str, str]]):
+    """
+    HTTPHeaderDict is unusual for a Mapping[str, str] in that it has two modes of
+    address.
+
+    If we directly try to get an item with a particular name, we will get a string
+    back that is the concatenated version of all the values:
+
+    >>> d['X-Header-Name']
+    'Value1, Value2, Value3'
+
+    However, if we iterate over an HTTPHeaderDict's items, we will optionally combine
+    these values based on whether combine=True was called when building up the dictionary
+
+    >>> d = HTTPHeaderDict({"A": "1", "B": "foo"})
+    >>> d.add("A", "2", combine=True)
+    >>> d.add("B", "bar")
+    >>> list(d.items())
+    [
+        ('A', '1, 2'),
+        ('B', 'foo'),
+        ('B', 'bar'),
+    ]
+
+    This class conforms to the interface required by the MutableMapping ABC while
+    also giving us the nonstandard iteration behavior we want; items with duplicate
+    keys, ordered by time of first insertion.
+    """
+
+    _headers: HTTPHeaderDict
+
+    def __init__(self, headers: HTTPHeaderDict) -> None:
+        self._headers = headers
+
+    def __len__(self) -> int:
+        return len(list(self._headers.iteritems()))
+
+    def __iter__(self) -> typing.Iterator[tuple[str, str]]:
+        return self._headers.iteritems()
+
+    def __contains__(self, item: object) -> bool:
+        if isinstance(item, tuple) and len(item) == 2:
+            passed_key, passed_val = item
+            if isinstance(passed_key, str) and isinstance(passed_val, str):
+                return self._headers._has_value_for_header(passed_key, passed_val)
+        return False
+
+
+class HTTPHeaderDict(typing.MutableMapping[str, str]):
+    """
+    :param headers:
+        An iterable of field-value pairs. Must not contain multiple field names
+        when compared case-insensitively.
+
+    :param kwargs:
+        Additional field-value pairs to pass in to ``dict.update``.
+
+    A ``dict`` like container for storing HTTP Headers.
+
+    Field names are stored and compared case-insensitively in compliance with
+    RFC 7230. Iteration provides the first case-sensitive key seen for each
+    case-insensitive pair.
+
+    Using ``__setitem__`` syntax overwrites fields that compare equal
+    case-insensitively in order to maintain ``dict``'s api. For fields that
+    compare equal, instead create a new ``HTTPHeaderDict`` and use ``.add``
+    in a loop.
+
+    If multiple fields that are equal case-insensitively are passed to the
+    constructor or ``.update``, the behavior is undefined and some will be
+    lost.
+
+    >>> headers = HTTPHeaderDict()
+    >>> headers.add('Set-Cookie', 'foo=bar')
+    >>> headers.add('set-cookie', 'baz=quxx')
+    >>> headers['content-length'] = '7'
+    >>> headers['SET-cookie']
+    'foo=bar, baz=quxx'
+    >>> headers['Content-Length']
+    '7'
+    """
+
+    _container: typing.MutableMapping[str, list[str]]
+
+    def __init__(self, headers: ValidHTTPHeaderSource | None = None, **kwargs: str):
+        super().__init__()
+        self._container = {}  # 'dict' is insert-ordered
+        if headers is not None:
+            if isinstance(headers, HTTPHeaderDict):
+                self._copy_from(headers)
+            else:
+                self.extend(headers)
+        if kwargs:
+            self.extend(kwargs)
+
+    def __setitem__(self, key: str, val: str) -> None:
+        # avoid a bytes/str comparison by decoding before httplib
+        if isinstance(key, bytes):
+            key = key.decode("latin-1")
+        self._container[key.lower()] = [key, val]
+
+    def __getitem__(self, key: str) -> str:
+        val = self._container[key.lower()]
+        return ", ".join(val[1:])
+
+    def __delitem__(self, key: str) -> None:
+        del self._container[key.lower()]
+
+    def __contains__(self, key: object) -> bool:
+        if isinstance(key, str):
+            return key.lower() in self._container
+        return False
+
+    def setdefault(self, key: str, default: str = "") -> str:
+        return super().setdefault(key, default)
+
+    def __eq__(self, other: object) -> bool:
+        maybe_constructable = ensure_can_construct_http_header_dict(other)
+        if maybe_constructable is None:
+            return False
+        else:
+            other_as_http_header_dict = type(self)(maybe_constructable)
+
+        return {k.lower(): v for k, v in self.itermerged()} == {
+            k.lower(): v for k, v in other_as_http_header_dict.itermerged()
+        }
+
+    def __ne__(self, other: object) -> bool:
+        return not self.__eq__(other)
+
+    def __len__(self) -> int:
+        return len(self._container)
+
+    def __iter__(self) -> typing.Iterator[str]:
+        # Only provide the originally cased names
+        for vals in self._container.values():
+            yield vals[0]
+
+    def discard(self, key: str) -> None:
+        try:
+            del self[key]
+        except KeyError:
+            pass
+
+    def add(self, key: str, val: str, *, combine: bool = False) -> None:
+        """Adds a (name, value) pair, doesn't overwrite the value if it already
+        exists.
+
+        If this is called with combine=True, instead of adding a new header value
+        as a distinct item during iteration, this will instead append the value to
+        any existing header value with a comma. If no existing header value exists
+        for the key, then the value will simply be added, ignoring the combine parameter.
+
+        >>> headers = HTTPHeaderDict(foo='bar')
+        >>> headers.add('Foo', 'baz')
+        >>> headers['foo']
+        'bar, baz'
+        >>> list(headers.items())
+        [('foo', 'bar'), ('foo', 'baz')]
+        >>> headers.add('foo', 'quz', combine=True)
+        >>> list(headers.items())
+        [('foo', 'bar, baz, quz')]
+        """
+        # avoid a bytes/str comparison by decoding before httplib
+        if isinstance(key, bytes):
+            key = key.decode("latin-1")
+        key_lower = key.lower()
+        new_vals = [key, val]
+        # Keep the common case aka no item present as fast as possible
+        vals = self._container.setdefault(key_lower, new_vals)
+        if new_vals is not vals:
+            # if there are values here, then there is at least the initial
+            # key/value pair
+            assert len(vals) >= 2
+            if combine:
+                vals[-1] = vals[-1] + ", " + val
+            else:
+                vals.append(val)
+
+    def extend(self, *args: ValidHTTPHeaderSource, **kwargs: str) -> None:
+        """Generic import function for any type of header-like object.
+        Adapted version of MutableMapping.update in order to insert items
+        with self.add instead of self.__setitem__
+        """
+        if len(args) > 1:
+            raise TypeError(
+                f"extend() takes at most 1 positional arguments ({len(args)} given)"
+            )
+        other = args[0] if len(args) >= 1 else ()
+
+        if isinstance(other, HTTPHeaderDict):
+            for key, val in other.iteritems():
+                self.add(key, val)
+        elif isinstance(other, typing.Mapping):
+            for key, val in other.items():
+                self.add(key, val)
+        elif isinstance(other, typing.Iterable):
+            other = typing.cast(typing.Iterable[tuple[str, str]], other)
+            for key, value in other:
+                self.add(key, value)
+        elif hasattr(other, "keys") and hasattr(other, "__getitem__"):
+            # THIS IS NOT A TYPESAFE BRANCH
+            # In this branch, the object has a `keys` attr but is not a Mapping or any of
+            # the other types indicated in the method signature. We do some stuff with
+            # it as though it partially implements the Mapping interface, but we're not
+            # doing that stuff safely AT ALL.
+            for key in other.keys():
+                self.add(key, other[key])
+
+        for key, value in kwargs.items():
+            self.add(key, value)
+
+    @typing.overload
+    def getlist(self, key: str) -> list[str]: ...
+
+    @typing.overload
+    def getlist(self, key: str, default: _DT) -> list[str] | _DT: ...
+
+    def getlist(
+        self, key: str, default: _Sentinel | _DT = _Sentinel.not_passed
+    ) -> list[str] | _DT:
+        """Returns a list of all the values for the named field. Returns an
+        empty list if the key doesn't exist."""
+        try:
+            vals = self._container[key.lower()]
+        except KeyError:
+            if default is _Sentinel.not_passed:
+                # _DT is unbound; empty list is instance of List[str]
+                return []
+            # _DT is bound; default is instance of _DT
+            return default
+        else:
+            # _DT may or may not be bound; vals[1:] is instance of List[str], which
+            # meets our external interface requirement of `Union[List[str], _DT]`.
+            return vals[1:]
+
+    def _prepare_for_method_change(self) -> Self:
+        """
+        Remove content-specific header fields before changing the request
+        method to GET or HEAD according to RFC 9110, Section 15.4.
+        """
+        content_specific_headers = [
+            "Content-Encoding",
+            "Content-Language",
+            "Content-Location",
+            "Content-Type",
+            "Content-Length",
+            "Digest",
+            "Last-Modified",
+        ]
+        for header in content_specific_headers:
+            self.discard(header)
+        return self
+
+    # Backwards compatibility for httplib
+    getheaders = getlist
+    getallmatchingheaders = getlist
+    iget = getlist
+
+    # Backwards compatibility for http.cookiejar
+    get_all = getlist
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({dict(self.itermerged())})"
+
+    def _copy_from(self, other: HTTPHeaderDict) -> None:
+        for key in other:
+            val = other.getlist(key)
+            self._container[key.lower()] = [key, *val]
+
+    def copy(self) -> Self:
+        clone = type(self)()
+        clone._copy_from(self)
+        return clone
+
+    def iteritems(self) -> typing.Iterator[tuple[str, str]]:
+        """Iterate over all header lines, including duplicate ones."""
+        for key in self:
+            vals = self._container[key.lower()]
+            for val in vals[1:]:
+                yield vals[0], val
+
+    def itermerged(self) -> typing.Iterator[tuple[str, str]]:
+        """Iterate over all headers, merging duplicate ones together."""
+        for key in self:
+            val = self._container[key.lower()]
+            yield val[0], ", ".join(val[1:])
+
+    def items(self) -> HTTPHeaderDictItemView:  # type: ignore[override]
+        return HTTPHeaderDictItemView(self)
+
+    def _has_value_for_header(self, header_name: str, potential_value: str) -> bool:
+        if header_name in self:
+            return potential_value in self._container[header_name.lower()][1:]
+        return False
+
+    def __ior__(self, other: object) -> HTTPHeaderDict:
+        # Supports extending a header dict in-place using operator |=
+        # combining items with add instead of __setitem__
+        maybe_constructable = ensure_can_construct_http_header_dict(other)
+        if maybe_constructable is None:
+            return NotImplemented
+        self.extend(maybe_constructable)
+        return self
+
+    def __or__(self, other: object) -> Self:
+        # Supports merging header dicts using operator |
+        # combining items with add instead of __setitem__
+        maybe_constructable = ensure_can_construct_http_header_dict(other)
+        if maybe_constructable is None:
+            return NotImplemented
+        result = self.copy()
+        result.extend(maybe_constructable)
+        return result
+
+    def __ror__(self, other: object) -> Self:
+        # Supports merging header dicts using operator | when other is on left side
+        # combining items with add instead of __setitem__
+        maybe_constructable = ensure_can_construct_http_header_dict(other)
+        if maybe_constructable is None:
+            return NotImplemented
+        result = type(self)(maybe_constructable)
+        result.extend(self)
+        return result
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/_request_methods.py b/backend/venv/lib/python3.12/site-packages/urllib3/_request_methods.py
new file mode 100644
index 00000000..297c271b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/_request_methods.py
@@ -0,0 +1,278 @@
+from __future__ import annotations
+
+import json as _json
+import typing
+from urllib.parse import urlencode
+
+from ._base_connection import _TYPE_BODY
+from ._collections import HTTPHeaderDict
+from .filepost import _TYPE_FIELDS, encode_multipart_formdata
+from .response import BaseHTTPResponse
+
+__all__ = ["RequestMethods"]
+
+_TYPE_ENCODE_URL_FIELDS = typing.Union[
+    typing.Sequence[tuple[str, typing.Union[str, bytes]]],
+    typing.Mapping[str, typing.Union[str, bytes]],
+]
+
+
+class RequestMethods:
+    """
+    Convenience mixin for classes who implement a :meth:`urlopen` method, such
+    as :class:`urllib3.HTTPConnectionPool` and
+    :class:`urllib3.PoolManager`.
+
+    Provides behavior for making common types of HTTP request methods and
+    decides which type of request field encoding to use.
+
+    Specifically,
+
+    :meth:`.request_encode_url` is for sending requests whose fields are
+    encoded in the URL (such as GET, HEAD, DELETE).
+
+    :meth:`.request_encode_body` is for sending requests whose fields are
+    encoded in the *body* of the request using multipart or www-form-urlencoded
+    (such as for POST, PUT, PATCH).
+
+    :meth:`.request` is for making any kind of request, it will look up the
+    appropriate encoding format and use one of the above two methods to make
+    the request.
+
+    Initializer parameters:
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+    """
+
+    _encode_url_methods = {"DELETE", "GET", "HEAD", "OPTIONS"}
+
+    def __init__(self, headers: typing.Mapping[str, str] | None = None) -> None:
+        self.headers = headers or {}
+
+    def urlopen(
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        encode_multipart: bool = True,
+        multipart_boundary: str | None = None,
+        **kw: typing.Any,
+    ) -> BaseHTTPResponse:  # Abstract
+        raise NotImplementedError(
+            "Classes extending RequestMethods must implement "
+            "their own ``urlopen`` method."
+        )
+
+    def request(
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        fields: _TYPE_FIELDS | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        json: typing.Any | None = None,
+        **urlopen_kw: typing.Any,
+    ) -> BaseHTTPResponse:
+        """
+        Make a request using :meth:`urlopen` with the appropriate encoding of
+        ``fields`` based on the ``method`` used.
+
+        This is a convenience method that requires the least amount of manual
+        effort. It can be used in most situations, while still having the
+        option to drop down to more specific methods when necessary, such as
+        :meth:`request_encode_url`, :meth:`request_encode_body`,
+        or even the lowest level :meth:`urlopen`.
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param body:
+            Data to send in the request body, either :class:`str`, :class:`bytes`,
+            an iterable of :class:`str`/:class:`bytes`, or a file-like object.
+
+        :param fields:
+            Data to encode and send in the URL or request body, depending on ``method``.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+
+        :param json:
+            Data to encode and send as JSON with UTF-encoded in the request body.
+            The ``"Content-Type"`` header will be set to ``"application/json"``
+            unless specified otherwise.
+        """
+        method = method.upper()
+
+        if json is not None and body is not None:
+            raise TypeError(
+                "request got values for both 'body' and 'json' parameters which are mutually exclusive"
+            )
+
+        if json is not None:
+            if headers is None:
+                headers = self.headers
+
+            if not ("content-type" in map(str.lower, headers.keys())):
+                headers = HTTPHeaderDict(headers)
+                headers["Content-Type"] = "application/json"
+
+            body = _json.dumps(json, separators=(",", ":"), ensure_ascii=False).encode(
+                "utf-8"
+            )
+
+        if body is not None:
+            urlopen_kw["body"] = body
+
+        if method in self._encode_url_methods:
+            return self.request_encode_url(
+                method,
+                url,
+                fields=fields,  # type: ignore[arg-type]
+                headers=headers,
+                **urlopen_kw,
+            )
+        else:
+            return self.request_encode_body(
+                method, url, fields=fields, headers=headers, **urlopen_kw
+            )
+
+    def request_encode_url(
+        self,
+        method: str,
+        url: str,
+        fields: _TYPE_ENCODE_URL_FIELDS | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        **urlopen_kw: str,
+    ) -> BaseHTTPResponse:
+        """
+        Make a request using :meth:`urlopen` with the ``fields`` encoded in
+        the url. This is useful for request methods like GET, HEAD, DELETE, etc.
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param fields:
+            Data to encode and send in the URL.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+        """
+        if headers is None:
+            headers = self.headers
+
+        extra_kw: dict[str, typing.Any] = {"headers": headers}
+        extra_kw.update(urlopen_kw)
+
+        if fields:
+            url += "?" + urlencode(fields)
+
+        return self.urlopen(method, url, **extra_kw)
+
+    def request_encode_body(
+        self,
+        method: str,
+        url: str,
+        fields: _TYPE_FIELDS | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        encode_multipart: bool = True,
+        multipart_boundary: str | None = None,
+        **urlopen_kw: str,
+    ) -> BaseHTTPResponse:
+        """
+        Make a request using :meth:`urlopen` with the ``fields`` encoded in
+        the body. This is useful for request methods like POST, PUT, PATCH, etc.
+
+        When ``encode_multipart=True`` (default), then
+        :func:`urllib3.encode_multipart_formdata` is used to encode
+        the payload with the appropriate content type. Otherwise
+        :func:`urllib.parse.urlencode` is used with the
+        'application/x-www-form-urlencoded' content type.
+
+        Multipart encoding must be used when posting files, and it's reasonably
+        safe to use it in other times too. However, it may break request
+        signing, such as with OAuth.
+
+        Supports an optional ``fields`` parameter of key/value strings AND
+        key/filetuple. A filetuple is a (filename, data, MIME type) tuple where
+        the MIME type is optional. For example::
+
+            fields = {
+                'foo': 'bar',
+                'fakefile': ('foofile.txt', 'contents of foofile'),
+                'realfile': ('barfile.txt', open('realfile').read()),
+                'typedfile': ('bazfile.bin', open('bazfile').read(),
+                              'image/jpeg'),
+                'nonamefile': 'contents of nonamefile field',
+            }
+
+        When uploading a file, providing a filename (the first parameter of the
+        tuple) is optional but recommended to best mimic behavior of browsers.
+
+        Note that if ``headers`` are supplied, the 'Content-Type' header will
+        be overwritten because it depends on the dynamic random boundary string
+        which is used to compose the body of the request. The random boundary
+        string can be explicitly set with the ``multipart_boundary`` parameter.
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param fields:
+            Data to encode and send in the request body.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+
+        :param encode_multipart:
+            If True, encode the ``fields`` using the multipart/form-data MIME
+            format.
+
+        :param multipart_boundary:
+            If not specified, then a random boundary will be generated using
+            :func:`urllib3.filepost.choose_boundary`.
+        """
+        if headers is None:
+            headers = self.headers
+
+        extra_kw: dict[str, typing.Any] = {"headers": HTTPHeaderDict(headers)}
+        body: bytes | str
+
+        if fields:
+            if "body" in urlopen_kw:
+                raise TypeError(
+                    "request got values for both 'fields' and 'body', can only specify one."
+                )
+
+            if encode_multipart:
+                body, content_type = encode_multipart_formdata(
+                    fields, boundary=multipart_boundary
+                )
+            else:
+                body, content_type = (
+                    urlencode(fields),  # type: ignore[arg-type]
+                    "application/x-www-form-urlencoded",
+                )
+
+            extra_kw["body"] = body
+            extra_kw["headers"].setdefault("Content-Type", content_type)
+
+        extra_kw.update(urlopen_kw)
+
+        return self.urlopen(method, url, **extra_kw)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/_version.py b/backend/venv/lib/python3.12/site-packages/urllib3/_version.py
new file mode 100644
index 00000000..af6144bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/_version.py
@@ -0,0 +1,16 @@
+# file generated by setuptools_scm
+# don't change, don't track in version control
+TYPE_CHECKING = False
+if TYPE_CHECKING:
+    from typing import Tuple, Union
+    VERSION_TUPLE = Tuple[Union[int, str], ...]
+else:
+    VERSION_TUPLE = object
+
+version: str
+__version__: str
+__version_tuple__: VERSION_TUPLE
+version_tuple: VERSION_TUPLE
+
+__version__ = version = '2.3.0'
+__version_tuple__ = version_tuple = (2, 3, 0)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/connection.py b/backend/venv/lib/python3.12/site-packages/urllib3/connection.py
new file mode 100644
index 00000000..591ac407
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/connection.py
@@ -0,0 +1,1044 @@
+from __future__ import annotations
+
+import datetime
+import http.client
+import logging
+import os
+import re
+import socket
+import sys
+import threading
+import typing
+import warnings
+from http.client import HTTPConnection as _HTTPConnection
+from http.client import HTTPException as HTTPException  # noqa: F401
+from http.client import ResponseNotReady
+from socket import timeout as SocketTimeout
+
+if typing.TYPE_CHECKING:
+    from .response import HTTPResponse
+    from .util.ssl_ import _TYPE_PEER_CERT_RET_DICT
+    from .util.ssltransport import SSLTransport
+
+from ._collections import HTTPHeaderDict
+from .http2 import probe as http2_probe
+from .util.response import assert_header_parsing
+from .util.timeout import _DEFAULT_TIMEOUT, _TYPE_TIMEOUT, Timeout
+from .util.util import to_str
+from .util.wait import wait_for_read
+
+try:  # Compiled with SSL?
+    import ssl
+
+    BaseSSLError = ssl.SSLError
+except (ImportError, AttributeError):
+    ssl = None  # type: ignore[assignment]
+
+    class BaseSSLError(BaseException):  # type: ignore[no-redef]
+        pass
+
+
+from ._base_connection import _TYPE_BODY
+from ._base_connection import ProxyConfig as ProxyConfig
+from ._base_connection import _ResponseOptions as _ResponseOptions
+from ._version import __version__
+from .exceptions import (
+    ConnectTimeoutError,
+    HeaderParsingError,
+    NameResolutionError,
+    NewConnectionError,
+    ProxyError,
+    SystemTimeWarning,
+)
+from .util import SKIP_HEADER, SKIPPABLE_HEADERS, connection, ssl_
+from .util.request import body_to_chunks
+from .util.ssl_ import assert_fingerprint as _assert_fingerprint
+from .util.ssl_ import (
+    create_urllib3_context,
+    is_ipaddress,
+    resolve_cert_reqs,
+    resolve_ssl_version,
+    ssl_wrap_socket,
+)
+from .util.ssl_match_hostname import CertificateError, match_hostname
+from .util.url import Url
+
+# Not a no-op, we're adding this to the namespace so it can be imported.
+ConnectionError = ConnectionError
+BrokenPipeError = BrokenPipeError
+
+
+log = logging.getLogger(__name__)
+
+port_by_scheme = {"http": 80, "https": 443}
+
+# When it comes time to update this value as a part of regular maintenance
+# (ie test_recent_date is failing) update it to ~6 months before the current date.
+RECENT_DATE = datetime.date(2023, 6, 1)
+
+_CONTAINS_CONTROL_CHAR_RE = re.compile(r"[^-!#$%&'*+.^_`|~0-9a-zA-Z]")
+
+
+class HTTPConnection(_HTTPConnection):
+    """
+    Based on :class:`http.client.HTTPConnection` but provides an extra constructor
+    backwards-compatibility layer between older and newer Pythons.
+
+    Additional keyword parameters are used to configure attributes of the connection.
+    Accepted parameters include:
+
+    - ``source_address``: Set the source address for the current connection.
+    - ``socket_options``: Set specific options on the underlying socket. If not specified, then
+      defaults are loaded from ``HTTPConnection.default_socket_options`` which includes disabling
+      Nagle's algorithm (sets TCP_NODELAY to 1) unless the connection is behind a proxy.
+
+      For example, if you wish to enable TCP Keep Alive in addition to the defaults,
+      you might pass:
+
+      .. code-block:: python
+
+         HTTPConnection.default_socket_options + [
+             (socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1),
+         ]
+
+      Or you may want to disable the defaults by passing an empty list (e.g., ``[]``).
+    """
+
+    default_port: typing.ClassVar[int] = port_by_scheme["http"]  # type: ignore[misc]
+
+    #: Disable Nagle's algorithm by default.
+    #: ``[(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)]``
+    default_socket_options: typing.ClassVar[connection._TYPE_SOCKET_OPTIONS] = [
+        (socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)
+    ]
+
+    #: Whether this connection verifies the host's certificate.
+    is_verified: bool = False
+
+    #: Whether this proxy connection verified the proxy host's certificate.
+    # If no proxy is currently connected to the value will be ``None``.
+    proxy_is_verified: bool | None = None
+
+    blocksize: int
+    source_address: tuple[str, int] | None
+    socket_options: connection._TYPE_SOCKET_OPTIONS | None
+
+    _has_connected_to_proxy: bool
+    _response_options: _ResponseOptions | None
+    _tunnel_host: str | None
+    _tunnel_port: int | None
+    _tunnel_scheme: str | None
+
+    def __init__(
+        self,
+        host: str,
+        port: int | None = None,
+        *,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        source_address: tuple[str, int] | None = None,
+        blocksize: int = 16384,
+        socket_options: None | (
+            connection._TYPE_SOCKET_OPTIONS
+        ) = default_socket_options,
+        proxy: Url | None = None,
+        proxy_config: ProxyConfig | None = None,
+    ) -> None:
+        super().__init__(
+            host=host,
+            port=port,
+            timeout=Timeout.resolve_default_timeout(timeout),
+            source_address=source_address,
+            blocksize=blocksize,
+        )
+        self.socket_options = socket_options
+        self.proxy = proxy
+        self.proxy_config = proxy_config
+
+        self._has_connected_to_proxy = False
+        self._response_options = None
+        self._tunnel_host: str | None = None
+        self._tunnel_port: int | None = None
+        self._tunnel_scheme: str | None = None
+
+    @property
+    def host(self) -> str:
+        """
+        Getter method to remove any trailing dots that indicate the hostname is an FQDN.
+
+        In general, SSL certificates don't include the trailing dot indicating a
+        fully-qualified domain name, and thus, they don't validate properly when
+        checked against a domain name that includes the dot. In addition, some
+        servers may not expect to receive the trailing dot when provided.
+
+        However, the hostname with trailing dot is critical to DNS resolution; doing a
+        lookup with the trailing dot will properly only resolve the appropriate FQDN,
+        whereas a lookup without a trailing dot will search the system's search domain
+        list. Thus, it's important to keep the original host around for use only in
+        those cases where it's appropriate (i.e., when doing DNS lookup to establish the
+        actual TCP connection across which we're going to send HTTP requests).
+        """
+        return self._dns_host.rstrip(".")
+
+    @host.setter
+    def host(self, value: str) -> None:
+        """
+        Setter for the `host` property.
+
+        We assume that only urllib3 uses the _dns_host attribute; httplib itself
+        only uses `host`, and it seems reasonable that other libraries follow suit.
+        """
+        self._dns_host = value
+
+    def _new_conn(self) -> socket.socket:
+        """Establish a socket connection and set nodelay settings on it.
+
+        :return: New socket connection.
+        """
+        try:
+            sock = connection.create_connection(
+                (self._dns_host, self.port),
+                self.timeout,
+                source_address=self.source_address,
+                socket_options=self.socket_options,
+            )
+        except socket.gaierror as e:
+            raise NameResolutionError(self.host, self, e) from e
+        except SocketTimeout as e:
+            raise ConnectTimeoutError(
+                self,
+                f"Connection to {self.host} timed out. (connect timeout={self.timeout})",
+            ) from e
+
+        except OSError as e:
+            raise NewConnectionError(
+                self, f"Failed to establish a new connection: {e}"
+            ) from e
+
+        sys.audit("http.client.connect", self, self.host, self.port)
+
+        return sock
+
+    def set_tunnel(
+        self,
+        host: str,
+        port: int | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        scheme: str = "http",
+    ) -> None:
+        if scheme not in ("http", "https"):
+            raise ValueError(
+                f"Invalid proxy scheme for tunneling: {scheme!r}, must be either 'http' or 'https'"
+            )
+        super().set_tunnel(host, port=port, headers=headers)
+        self._tunnel_scheme = scheme
+
+    if sys.version_info < (3, 11, 4):
+
+        def _tunnel(self) -> None:
+            _MAXLINE = http.client._MAXLINE  # type: ignore[attr-defined]
+            connect = b"CONNECT %s:%d HTTP/1.0\r\n" % (  # type: ignore[str-format]
+                self._tunnel_host.encode("ascii"),  # type: ignore[union-attr]
+                self._tunnel_port,
+            )
+            headers = [connect]
+            for header, value in self._tunnel_headers.items():  # type: ignore[attr-defined]
+                headers.append(f"{header}: {value}\r\n".encode("latin-1"))
+            headers.append(b"\r\n")
+            # Making a single send() call instead of one per line encourages
+            # the host OS to use a more optimal packet size instead of
+            # potentially emitting a series of small packets.
+            self.send(b"".join(headers))
+            del headers
+
+            response = self.response_class(self.sock, method=self._method)  # type: ignore[attr-defined]
+            try:
+                (version, code, message) = response._read_status()  # type: ignore[attr-defined]
+
+                if code != http.HTTPStatus.OK:
+                    self.close()
+                    raise OSError(f"Tunnel connection failed: {code} {message.strip()}")
+                while True:
+                    line = response.fp.readline(_MAXLINE + 1)
+                    if len(line) > _MAXLINE:
+                        raise http.client.LineTooLong("header line")
+                    if not line:
+                        # for sites which EOF without sending a trailer
+                        break
+                    if line in (b"\r\n", b"\n", b""):
+                        break
+
+                    if self.debuglevel > 0:
+                        print("header:", line.decode())
+            finally:
+                response.close()
+
+    def connect(self) -> None:
+        self.sock = self._new_conn()
+        if self._tunnel_host:
+            # If we're tunneling it means we're connected to our proxy.
+            self._has_connected_to_proxy = True
+
+            # TODO: Fix tunnel so it doesn't depend on self.sock state.
+            self._tunnel()
+
+        # If there's a proxy to be connected to we are fully connected.
+        # This is set twice (once above and here) due to forwarding proxies
+        # not using tunnelling.
+        self._has_connected_to_proxy = bool(self.proxy)
+
+        if self._has_connected_to_proxy:
+            self.proxy_is_verified = False
+
+    @property
+    def is_closed(self) -> bool:
+        return self.sock is None
+
+    @property
+    def is_connected(self) -> bool:
+        if self.sock is None:
+            return False
+        return not wait_for_read(self.sock, timeout=0.0)
+
+    @property
+    def has_connected_to_proxy(self) -> bool:
+        return self._has_connected_to_proxy
+
+    @property
+    def proxy_is_forwarding(self) -> bool:
+        """
+        Return True if a forwarding proxy is configured, else return False
+        """
+        return bool(self.proxy) and self._tunnel_host is None
+
+    @property
+    def proxy_is_tunneling(self) -> bool:
+        """
+        Return True if a tunneling proxy is configured, else return False
+        """
+        return self._tunnel_host is not None
+
+    def close(self) -> None:
+        try:
+            super().close()
+        finally:
+            # Reset all stateful properties so connection
+            # can be re-used without leaking prior configs.
+            self.sock = None
+            self.is_verified = False
+            self.proxy_is_verified = None
+            self._has_connected_to_proxy = False
+            self._response_options = None
+            self._tunnel_host = None
+            self._tunnel_port = None
+            self._tunnel_scheme = None
+
+    def putrequest(
+        self,
+        method: str,
+        url: str,
+        skip_host: bool = False,
+        skip_accept_encoding: bool = False,
+    ) -> None:
+        """"""
+        # Empty docstring because the indentation of CPython's implementation
+        # is broken but we don't want this method in our documentation.
+        match = _CONTAINS_CONTROL_CHAR_RE.search(method)
+        if match:
+            raise ValueError(
+                f"Method cannot contain non-token characters {method!r} (found at least {match.group()!r})"
+            )
+
+        return super().putrequest(
+            method, url, skip_host=skip_host, skip_accept_encoding=skip_accept_encoding
+        )
+
+    def putheader(self, header: str, *values: str) -> None:  # type: ignore[override]
+        """"""
+        if not any(isinstance(v, str) and v == SKIP_HEADER for v in values):
+            super().putheader(header, *values)
+        elif to_str(header.lower()) not in SKIPPABLE_HEADERS:
+            skippable_headers = "', '".join(
+                [str.title(header) for header in sorted(SKIPPABLE_HEADERS)]
+            )
+            raise ValueError(
+                f"urllib3.util.SKIP_HEADER only supports '{skippable_headers}'"
+            )
+
+    # `request` method's signature intentionally violates LSP.
+    # urllib3's API is different from `http.client.HTTPConnection` and the subclassing is only incidental.
+    def request(  # type: ignore[override]
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        *,
+        chunked: bool = False,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        enforce_content_length: bool = True,
+    ) -> None:
+        # Update the inner socket's timeout value to send the request.
+        # This only triggers if the connection is re-used.
+        if self.sock is not None:
+            self.sock.settimeout(self.timeout)
+
+        # Store these values to be fed into the HTTPResponse
+        # object later. TODO: Remove this in favor of a real
+        # HTTP lifecycle mechanism.
+
+        # We have to store these before we call .request()
+        # because sometimes we can still salvage a response
+        # off the wire even if we aren't able to completely
+        # send the request body.
+        self._response_options = _ResponseOptions(
+            request_method=method,
+            request_url=url,
+            preload_content=preload_content,
+            decode_content=decode_content,
+            enforce_content_length=enforce_content_length,
+        )
+
+        if headers is None:
+            headers = {}
+        header_keys = frozenset(to_str(k.lower()) for k in headers)
+        skip_accept_encoding = "accept-encoding" in header_keys
+        skip_host = "host" in header_keys
+        self.putrequest(
+            method, url, skip_accept_encoding=skip_accept_encoding, skip_host=skip_host
+        )
+
+        # Transform the body into an iterable of sendall()-able chunks
+        # and detect if an explicit Content-Length is doable.
+        chunks_and_cl = body_to_chunks(body, method=method, blocksize=self.blocksize)
+        chunks = chunks_and_cl.chunks
+        content_length = chunks_and_cl.content_length
+
+        # When chunked is explicit set to 'True' we respect that.
+        if chunked:
+            if "transfer-encoding" not in header_keys:
+                self.putheader("Transfer-Encoding", "chunked")
+        else:
+            # Detect whether a framing mechanism is already in use. If so
+            # we respect that value, otherwise we pick chunked vs content-length
+            # depending on the type of 'body'.
+            if "content-length" in header_keys:
+                chunked = False
+            elif "transfer-encoding" in header_keys:
+                chunked = True
+
+            # Otherwise we go off the recommendation of 'body_to_chunks()'.
+            else:
+                chunked = False
+                if content_length is None:
+                    if chunks is not None:
+                        chunked = True
+                        self.putheader("Transfer-Encoding", "chunked")
+                else:
+                    self.putheader("Content-Length", str(content_length))
+
+        # Now that framing headers are out of the way we send all the other headers.
+        if "user-agent" not in header_keys:
+            self.putheader("User-Agent", _get_default_user_agent())
+        for header, value in headers.items():
+            self.putheader(header, value)
+        self.endheaders()
+
+        # If we're given a body we start sending that in chunks.
+        if chunks is not None:
+            for chunk in chunks:
+                # Sending empty chunks isn't allowed for TE: chunked
+                # as it indicates the end of the body.
+                if not chunk:
+                    continue
+                if isinstance(chunk, str):
+                    chunk = chunk.encode("utf-8")
+                if chunked:
+                    self.send(b"%x\r\n%b\r\n" % (len(chunk), chunk))
+                else:
+                    self.send(chunk)
+
+        # Regardless of whether we have a body or not, if we're in
+        # chunked mode we want to send an explicit empty chunk.
+        if chunked:
+            self.send(b"0\r\n\r\n")
+
+    def request_chunked(
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+    ) -> None:
+        """
+        Alternative to the common request method, which sends the
+        body with chunked encoding and not as one block
+        """
+        warnings.warn(
+            "HTTPConnection.request_chunked() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead use HTTPConnection.request(..., chunked=True).",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+        self.request(method, url, body=body, headers=headers, chunked=True)
+
+    def getresponse(  # type: ignore[override]
+        self,
+    ) -> HTTPResponse:
+        """
+        Get the response from the server.
+
+        If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable.
+
+        If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed.
+        """
+        # Raise the same error as http.client.HTTPConnection
+        if self._response_options is None:
+            raise ResponseNotReady()
+
+        # Reset this attribute for being used again.
+        resp_options = self._response_options
+        self._response_options = None
+
+        # Since the connection's timeout value may have been updated
+        # we need to set the timeout on the socket.
+        self.sock.settimeout(self.timeout)
+
+        # This is needed here to avoid circular import errors
+        from .response import HTTPResponse
+
+        # Save a reference to the shutdown function before ownership is passed
+        # to httplib_response
+        # TODO should we implement it everywhere?
+        _shutdown = getattr(self.sock, "shutdown", None)
+
+        # Get the response from http.client.HTTPConnection
+        httplib_response = super().getresponse()
+
+        try:
+            assert_header_parsing(httplib_response.msg)
+        except (HeaderParsingError, TypeError) as hpe:
+            log.warning(
+                "Failed to parse headers (url=%s): %s",
+                _url_from_connection(self, resp_options.request_url),
+                hpe,
+                exc_info=True,
+            )
+
+        headers = HTTPHeaderDict(httplib_response.msg.items())
+
+        response = HTTPResponse(
+            body=httplib_response,
+            headers=headers,
+            status=httplib_response.status,
+            version=httplib_response.version,
+            version_string=getattr(self, "_http_vsn_str", "HTTP/?"),
+            reason=httplib_response.reason,
+            preload_content=resp_options.preload_content,
+            decode_content=resp_options.decode_content,
+            original_response=httplib_response,
+            enforce_content_length=resp_options.enforce_content_length,
+            request_method=resp_options.request_method,
+            request_url=resp_options.request_url,
+            sock_shutdown=_shutdown,
+        )
+        return response
+
+
+class HTTPSConnection(HTTPConnection):
+    """
+    Many of the parameters to this constructor are passed to the underlying SSL
+    socket by means of :py:func:`urllib3.util.ssl_wrap_socket`.
+    """
+
+    default_port = port_by_scheme["https"]  # type: ignore[misc]
+
+    cert_reqs: int | str | None = None
+    ca_certs: str | None = None
+    ca_cert_dir: str | None = None
+    ca_cert_data: None | str | bytes = None
+    ssl_version: int | str | None = None
+    ssl_minimum_version: int | None = None
+    ssl_maximum_version: int | None = None
+    assert_fingerprint: str | None = None
+    _connect_callback: typing.Callable[..., None] | None = None
+
+    def __init__(
+        self,
+        host: str,
+        port: int | None = None,
+        *,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        source_address: tuple[str, int] | None = None,
+        blocksize: int = 16384,
+        socket_options: None | (
+            connection._TYPE_SOCKET_OPTIONS
+        ) = HTTPConnection.default_socket_options,
+        proxy: Url | None = None,
+        proxy_config: ProxyConfig | None = None,
+        cert_reqs: int | str | None = None,
+        assert_hostname: None | str | typing.Literal[False] = None,
+        assert_fingerprint: str | None = None,
+        server_hostname: str | None = None,
+        ssl_context: ssl.SSLContext | None = None,
+        ca_certs: str | None = None,
+        ca_cert_dir: str | None = None,
+        ca_cert_data: None | str | bytes = None,
+        ssl_minimum_version: int | None = None,
+        ssl_maximum_version: int | None = None,
+        ssl_version: int | str | None = None,  # Deprecated
+        cert_file: str | None = None,
+        key_file: str | None = None,
+        key_password: str | None = None,
+    ) -> None:
+        super().__init__(
+            host,
+            port=port,
+            timeout=timeout,
+            source_address=source_address,
+            blocksize=blocksize,
+            socket_options=socket_options,
+            proxy=proxy,
+            proxy_config=proxy_config,
+        )
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.key_password = key_password
+        self.ssl_context = ssl_context
+        self.server_hostname = server_hostname
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+        self.ssl_version = ssl_version
+        self.ssl_minimum_version = ssl_minimum_version
+        self.ssl_maximum_version = ssl_maximum_version
+        self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
+        self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
+        self.ca_cert_data = ca_cert_data
+
+        # cert_reqs depends on ssl_context so calculate last.
+        if cert_reqs is None:
+            if self.ssl_context is not None:
+                cert_reqs = self.ssl_context.verify_mode
+            else:
+                cert_reqs = resolve_cert_reqs(None)
+        self.cert_reqs = cert_reqs
+        self._connect_callback = None
+
+    def set_cert(
+        self,
+        key_file: str | None = None,
+        cert_file: str | None = None,
+        cert_reqs: int | str | None = None,
+        key_password: str | None = None,
+        ca_certs: str | None = None,
+        assert_hostname: None | str | typing.Literal[False] = None,
+        assert_fingerprint: str | None = None,
+        ca_cert_dir: str | None = None,
+        ca_cert_data: None | str | bytes = None,
+    ) -> None:
+        """
+        This method should only be called once, before the connection is used.
+        """
+        warnings.warn(
+            "HTTPSConnection.set_cert() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead provide the parameters to the "
+            "HTTPSConnection constructor.",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+
+        # If cert_reqs is not provided we'll assume CERT_REQUIRED unless we also
+        # have an SSLContext object in which case we'll use its verify_mode.
+        if cert_reqs is None:
+            if self.ssl_context is not None:
+                cert_reqs = self.ssl_context.verify_mode
+            else:
+                cert_reqs = resolve_cert_reqs(None)
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.cert_reqs = cert_reqs
+        self.key_password = key_password
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+        self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
+        self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
+        self.ca_cert_data = ca_cert_data
+
+    def connect(self) -> None:
+        # Today we don't need to be doing this step before the /actual/ socket
+        # connection, however in the future we'll need to decide whether to
+        # create a new socket or re-use an existing "shared" socket as a part
+        # of the HTTP/2 handshake dance.
+        if self._tunnel_host is not None and self._tunnel_port is not None:
+            probe_http2_host = self._tunnel_host
+            probe_http2_port = self._tunnel_port
+        else:
+            probe_http2_host = self.host
+            probe_http2_port = self.port
+
+        # Check if the target origin supports HTTP/2.
+        # If the value comes back as 'None' it means that the current thread
+        # is probing for HTTP/2 support. Otherwise, we're waiting for another
+        # probe to complete, or we get a value right away.
+        target_supports_http2: bool | None
+        if "h2" in ssl_.ALPN_PROTOCOLS:
+            target_supports_http2 = http2_probe.acquire_and_get(
+                host=probe_http2_host, port=probe_http2_port
+            )
+        else:
+            # If HTTP/2 isn't going to be offered it doesn't matter if
+            # the target supports HTTP/2. Don't want to make a probe.
+            target_supports_http2 = False
+
+        if self._connect_callback is not None:
+            self._connect_callback(
+                "before connect",
+                thread_id=threading.get_ident(),
+                target_supports_http2=target_supports_http2,
+            )
+
+        try:
+            sock: socket.socket | ssl.SSLSocket
+            self.sock = sock = self._new_conn()
+            server_hostname: str = self.host
+            tls_in_tls = False
+
+            # Do we need to establish a tunnel?
+            if self.proxy_is_tunneling:
+                # We're tunneling to an HTTPS origin so need to do TLS-in-TLS.
+                if self._tunnel_scheme == "https":
+                    # _connect_tls_proxy will verify and assign proxy_is_verified
+                    self.sock = sock = self._connect_tls_proxy(self.host, sock)
+                    tls_in_tls = True
+                elif self._tunnel_scheme == "http":
+                    self.proxy_is_verified = False
+
+                # If we're tunneling it means we're connected to our proxy.
+                self._has_connected_to_proxy = True
+
+                self._tunnel()
+                # Override the host with the one we're requesting data from.
+                server_hostname = typing.cast(str, self._tunnel_host)
+
+            if self.server_hostname is not None:
+                server_hostname = self.server_hostname
+
+            is_time_off = datetime.date.today() < RECENT_DATE
+            if is_time_off:
+                warnings.warn(
+                    (
+                        f"System time is way off (before {RECENT_DATE}). This will probably "
+                        "lead to SSL verification errors"
+                    ),
+                    SystemTimeWarning,
+                )
+
+            # Remove trailing '.' from fqdn hostnames to allow certificate validation
+            server_hostname_rm_dot = server_hostname.rstrip(".")
+
+            sock_and_verified = _ssl_wrap_socket_and_match_hostname(
+                sock=sock,
+                cert_reqs=self.cert_reqs,
+                ssl_version=self.ssl_version,
+                ssl_minimum_version=self.ssl_minimum_version,
+                ssl_maximum_version=self.ssl_maximum_version,
+                ca_certs=self.ca_certs,
+                ca_cert_dir=self.ca_cert_dir,
+                ca_cert_data=self.ca_cert_data,
+                cert_file=self.cert_file,
+                key_file=self.key_file,
+                key_password=self.key_password,
+                server_hostname=server_hostname_rm_dot,
+                ssl_context=self.ssl_context,
+                tls_in_tls=tls_in_tls,
+                assert_hostname=self.assert_hostname,
+                assert_fingerprint=self.assert_fingerprint,
+            )
+            self.sock = sock_and_verified.socket
+
+        # If an error occurs during connection/handshake we may need to release
+        # our lock so another connection can probe the origin.
+        except BaseException:
+            if self._connect_callback is not None:
+                self._connect_callback(
+                    "after connect failure",
+                    thread_id=threading.get_ident(),
+                    target_supports_http2=target_supports_http2,
+                )
+
+            if target_supports_http2 is None:
+                http2_probe.set_and_release(
+                    host=probe_http2_host, port=probe_http2_port, supports_http2=None
+                )
+            raise
+
+        # If this connection doesn't know if the origin supports HTTP/2
+        # we report back to the HTTP/2 probe our result.
+        if target_supports_http2 is None:
+            supports_http2 = sock_and_verified.socket.selected_alpn_protocol() == "h2"
+            http2_probe.set_and_release(
+                host=probe_http2_host,
+                port=probe_http2_port,
+                supports_http2=supports_http2,
+            )
+
+        # Forwarding proxies can never have a verified target since
+        # the proxy is the one doing the verification. Should instead
+        # use a CONNECT tunnel in order to verify the target.
+        # See: https://github.com/urllib3/urllib3/issues/3267.
+        if self.proxy_is_forwarding:
+            self.is_verified = False
+        else:
+            self.is_verified = sock_and_verified.is_verified
+
+        # If there's a proxy to be connected to we are fully connected.
+        # This is set twice (once above and here) due to forwarding proxies
+        # not using tunnelling.
+        self._has_connected_to_proxy = bool(self.proxy)
+
+        # Set `self.proxy_is_verified` unless it's already set while
+        # establishing a tunnel.
+        if self._has_connected_to_proxy and self.proxy_is_verified is None:
+            self.proxy_is_verified = sock_and_verified.is_verified
+
+    def _connect_tls_proxy(self, hostname: str, sock: socket.socket) -> ssl.SSLSocket:
+        """
+        Establish a TLS connection to the proxy using the provided SSL context.
+        """
+        # `_connect_tls_proxy` is called when self._tunnel_host is truthy.
+        proxy_config = typing.cast(ProxyConfig, self.proxy_config)
+        ssl_context = proxy_config.ssl_context
+        sock_and_verified = _ssl_wrap_socket_and_match_hostname(
+            sock,
+            cert_reqs=self.cert_reqs,
+            ssl_version=self.ssl_version,
+            ssl_minimum_version=self.ssl_minimum_version,
+            ssl_maximum_version=self.ssl_maximum_version,
+            ca_certs=self.ca_certs,
+            ca_cert_dir=self.ca_cert_dir,
+            ca_cert_data=self.ca_cert_data,
+            server_hostname=hostname,
+            ssl_context=ssl_context,
+            assert_hostname=proxy_config.assert_hostname,
+            assert_fingerprint=proxy_config.assert_fingerprint,
+            # Features that aren't implemented for proxies yet:
+            cert_file=None,
+            key_file=None,
+            key_password=None,
+            tls_in_tls=False,
+        )
+        self.proxy_is_verified = sock_and_verified.is_verified
+        return sock_and_verified.socket  # type: ignore[return-value]
+
+
+class _WrappedAndVerifiedSocket(typing.NamedTuple):
+    """
+    Wrapped socket and whether the connection is
+    verified after the TLS handshake
+    """
+
+    socket: ssl.SSLSocket | SSLTransport
+    is_verified: bool
+
+
+def _ssl_wrap_socket_and_match_hostname(
+    sock: socket.socket,
+    *,
+    cert_reqs: None | str | int,
+    ssl_version: None | str | int,
+    ssl_minimum_version: int | None,
+    ssl_maximum_version: int | None,
+    cert_file: str | None,
+    key_file: str | None,
+    key_password: str | None,
+    ca_certs: str | None,
+    ca_cert_dir: str | None,
+    ca_cert_data: None | str | bytes,
+    assert_hostname: None | str | typing.Literal[False],
+    assert_fingerprint: str | None,
+    server_hostname: str | None,
+    ssl_context: ssl.SSLContext | None,
+    tls_in_tls: bool = False,
+) -> _WrappedAndVerifiedSocket:
+    """Logic for constructing an SSLContext from all TLS parameters, passing
+    that down into ssl_wrap_socket, and then doing certificate verification
+    either via hostname or fingerprint. This function exists to guarantee
+    that both proxies and targets have the same behavior when connecting via TLS.
+    """
+    default_ssl_context = False
+    if ssl_context is None:
+        default_ssl_context = True
+        context = create_urllib3_context(
+            ssl_version=resolve_ssl_version(ssl_version),
+            ssl_minimum_version=ssl_minimum_version,
+            ssl_maximum_version=ssl_maximum_version,
+            cert_reqs=resolve_cert_reqs(cert_reqs),
+        )
+    else:
+        context = ssl_context
+
+    context.verify_mode = resolve_cert_reqs(cert_reqs)
+
+    # In some cases, we want to verify hostnames ourselves
+    if (
+        # `ssl` can't verify fingerprints or alternate hostnames
+        assert_fingerprint
+        or assert_hostname
+        # assert_hostname can be set to False to disable hostname checking
+        or assert_hostname is False
+        # We still support OpenSSL 1.0.2, which prevents us from verifying
+        # hostnames easily: https://github.com/pyca/pyopenssl/pull/933
+        or ssl_.IS_PYOPENSSL
+        or not ssl_.HAS_NEVER_CHECK_COMMON_NAME
+    ):
+        context.check_hostname = False
+
+    # Try to load OS default certs if none are given. We need to do the hasattr() check
+    # for custom pyOpenSSL SSLContext objects because they don't support
+    # load_default_certs().
+    if (
+        not ca_certs
+        and not ca_cert_dir
+        and not ca_cert_data
+        and default_ssl_context
+        and hasattr(context, "load_default_certs")
+    ):
+        context.load_default_certs()
+
+    # Ensure that IPv6 addresses are in the proper format and don't have a
+    # scope ID. Python's SSL module fails to recognize scoped IPv6 addresses
+    # and interprets them as DNS hostnames.
+    if server_hostname is not None:
+        normalized = server_hostname.strip("[]")
+        if "%" in normalized:
+            normalized = normalized[: normalized.rfind("%")]
+        if is_ipaddress(normalized):
+            server_hostname = normalized
+
+    ssl_sock = ssl_wrap_socket(
+        sock=sock,
+        keyfile=key_file,
+        certfile=cert_file,
+        key_password=key_password,
+        ca_certs=ca_certs,
+        ca_cert_dir=ca_cert_dir,
+        ca_cert_data=ca_cert_data,
+        server_hostname=server_hostname,
+        ssl_context=context,
+        tls_in_tls=tls_in_tls,
+    )
+
+    try:
+        if assert_fingerprint:
+            _assert_fingerprint(
+                ssl_sock.getpeercert(binary_form=True), assert_fingerprint
+            )
+        elif (
+            context.verify_mode != ssl.CERT_NONE
+            and not context.check_hostname
+            and assert_hostname is not False
+        ):
+            cert: _TYPE_PEER_CERT_RET_DICT = ssl_sock.getpeercert()  # type: ignore[assignment]
+
+            # Need to signal to our match_hostname whether to use 'commonName' or not.
+            # If we're using our own constructed SSLContext we explicitly set 'False'
+            # because PyPy hard-codes 'True' from SSLContext.hostname_checks_common_name.
+            if default_ssl_context:
+                hostname_checks_common_name = False
+            else:
+                hostname_checks_common_name = (
+                    getattr(context, "hostname_checks_common_name", False) or False
+                )
+
+            _match_hostname(
+                cert,
+                assert_hostname or server_hostname,  # type: ignore[arg-type]
+                hostname_checks_common_name,
+            )
+
+        return _WrappedAndVerifiedSocket(
+            socket=ssl_sock,
+            is_verified=context.verify_mode == ssl.CERT_REQUIRED
+            or bool(assert_fingerprint),
+        )
+    except BaseException:
+        ssl_sock.close()
+        raise
+
+
+def _match_hostname(
+    cert: _TYPE_PEER_CERT_RET_DICT | None,
+    asserted_hostname: str,
+    hostname_checks_common_name: bool = False,
+) -> None:
+    # Our upstream implementation of ssl.match_hostname()
+    # only applies this normalization to IP addresses so it doesn't
+    # match DNS SANs so we do the same thing!
+    stripped_hostname = asserted_hostname.strip("[]")
+    if is_ipaddress(stripped_hostname):
+        asserted_hostname = stripped_hostname
+
+    try:
+        match_hostname(cert, asserted_hostname, hostname_checks_common_name)
+    except CertificateError as e:
+        log.warning(
+            "Certificate did not match expected hostname: %s. Certificate: %s",
+            asserted_hostname,
+            cert,
+        )
+        # Add cert to exception and reraise so client code can inspect
+        # the cert when catching the exception, if they want to
+        e._peer_cert = cert  # type: ignore[attr-defined]
+        raise
+
+
+def _wrap_proxy_error(err: Exception, proxy_scheme: str | None) -> ProxyError:
+    # Look for the phrase 'wrong version number', if found
+    # then we should warn the user that we're very sure that
+    # this proxy is HTTP-only and they have a configuration issue.
+    error_normalized = " ".join(re.split("[^a-z]", str(err).lower()))
+    is_likely_http_proxy = (
+        "wrong version number" in error_normalized
+        or "unknown protocol" in error_normalized
+        or "record layer failure" in error_normalized
+    )
+    http_proxy_warning = (
+        ". Your proxy appears to only use HTTP and not HTTPS, "
+        "try changing your proxy URL to be HTTP. See: "
+        "https://urllib3.readthedocs.io/en/latest/advanced-usage.html"
+        "#https-proxy-error-http-proxy"
+    )
+    new_err = ProxyError(
+        f"Unable to connect to proxy"
+        f"{http_proxy_warning if is_likely_http_proxy and proxy_scheme == 'https' else ''}",
+        err,
+    )
+    new_err.__cause__ = err
+    return new_err
+
+
+def _get_default_user_agent() -> str:
+    return f"python-urllib3/{__version__}"
+
+
+class DummyConnection:
+    """Used to detect a failed ConnectionCls import."""
+
+
+if not ssl:
+    HTTPSConnection = DummyConnection  # type: ignore[misc, assignment] # noqa: F811
+
+
+VerifiedHTTPSConnection = HTTPSConnection
+
+
+def _url_from_connection(
+    conn: HTTPConnection | HTTPSConnection, path: str | None = None
+) -> str:
+    """Returns the URL from a given connection. This is mainly used for testing and logging."""
+
+    scheme = "https" if isinstance(conn, HTTPSConnection) else "http"
+
+    return Url(scheme=scheme, host=conn.host, port=conn.port, path=path).url
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/connectionpool.py b/backend/venv/lib/python3.12/site-packages/urllib3/connectionpool.py
new file mode 100644
index 00000000..3a0685b4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/connectionpool.py
@@ -0,0 +1,1178 @@
+from __future__ import annotations
+
+import errno
+import logging
+import queue
+import sys
+import typing
+import warnings
+import weakref
+from socket import timeout as SocketTimeout
+from types import TracebackType
+
+from ._base_connection import _TYPE_BODY
+from ._collections import HTTPHeaderDict
+from ._request_methods import RequestMethods
+from .connection import (
+    BaseSSLError,
+    BrokenPipeError,
+    DummyConnection,
+    HTTPConnection,
+    HTTPException,
+    HTTPSConnection,
+    ProxyConfig,
+    _wrap_proxy_error,
+)
+from .connection import port_by_scheme as port_by_scheme
+from .exceptions import (
+    ClosedPoolError,
+    EmptyPoolError,
+    FullPoolError,
+    HostChangedError,
+    InsecureRequestWarning,
+    LocationValueError,
+    MaxRetryError,
+    NewConnectionError,
+    ProtocolError,
+    ProxyError,
+    ReadTimeoutError,
+    SSLError,
+    TimeoutError,
+)
+from .response import BaseHTTPResponse
+from .util.connection import is_connection_dropped
+from .util.proxy import connection_requires_http_tunnel
+from .util.request import _TYPE_BODY_POSITION, set_file_position
+from .util.retry import Retry
+from .util.ssl_match_hostname import CertificateError
+from .util.timeout import _DEFAULT_TIMEOUT, _TYPE_DEFAULT, Timeout
+from .util.url import Url, _encode_target
+from .util.url import _normalize_host as normalize_host
+from .util.url import parse_url
+from .util.util import to_str
+
+if typing.TYPE_CHECKING:
+    import ssl
+
+    from typing_extensions import Self
+
+    from ._base_connection import BaseHTTPConnection, BaseHTTPSConnection
+
+log = logging.getLogger(__name__)
+
+_TYPE_TIMEOUT = typing.Union[Timeout, float, _TYPE_DEFAULT, None]
+
+
+# Pool objects
+class ConnectionPool:
+    """
+    Base class for all connection pools, such as
+    :class:`.HTTPConnectionPool` and :class:`.HTTPSConnectionPool`.
+
+    .. note::
+       ConnectionPool.urlopen() does not normalize or percent-encode target URIs
+       which is useful if your target server doesn't support percent-encoded
+       target URIs.
+    """
+
+    scheme: str | None = None
+    QueueCls = queue.LifoQueue
+
+    def __init__(self, host: str, port: int | None = None) -> None:
+        if not host:
+            raise LocationValueError("No host specified.")
+
+        self.host = _normalize_host(host, scheme=self.scheme)
+        self.port = port
+
+        # This property uses 'normalize_host()' (not '_normalize_host()')
+        # to avoid removing square braces around IPv6 addresses.
+        # This value is sent to `HTTPConnection.set_tunnel()` if called
+        # because square braces are required for HTTP CONNECT tunneling.
+        self._tunnel_host = normalize_host(host, scheme=self.scheme).lower()
+
+    def __str__(self) -> str:
+        return f"{type(self).__name__}(host={self.host!r}, port={self.port!r})"
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> typing.Literal[False]:
+        self.close()
+        # Return False to re-raise any potential exceptions
+        return False
+
+    def close(self) -> None:
+        """
+        Close all pooled connections and disable the pool.
+        """
+
+
+# This is taken from http://hg.python.org/cpython/file/7aaba721ebc0/Lib/socket.py#l252
+_blocking_errnos = {errno.EAGAIN, errno.EWOULDBLOCK}
+
+
+class HTTPConnectionPool(ConnectionPool, RequestMethods):
+    """
+    Thread-safe connection pool for one host.
+
+    :param host:
+        Host used for this HTTP Connection (e.g. "localhost"), passed into
+        :class:`http.client.HTTPConnection`.
+
+    :param port:
+        Port used for this HTTP Connection (None is equivalent to 80), passed
+        into :class:`http.client.HTTPConnection`.
+
+    :param timeout:
+        Socket timeout in seconds for each individual connection. This can
+        be a float or integer, which sets the timeout for the HTTP request,
+        or an instance of :class:`urllib3.util.Timeout` which gives you more
+        fine-grained control over request timeouts. After the constructor has
+        been parsed, this is always a `urllib3.util.Timeout` object.
+
+    :param maxsize:
+        Number of connections to save that can be reused. More than 1 is useful
+        in multithreaded situations. If ``block`` is set to False, more
+        connections will be created but they will not be saved once they've
+        been used.
+
+    :param block:
+        If set to True, no more than ``maxsize`` connections will be used at
+        a time. When no free connections are available, the call will block
+        until a connection has been released. This is a useful side effect for
+        particular multithreaded situations where one does not want to use more
+        than maxsize connections per host to prevent flooding.
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+
+    :param retries:
+        Retry configuration to use by default with requests in this pool.
+
+    :param _proxy:
+        Parsed proxy URL, should not be used directly, instead, see
+        :class:`urllib3.ProxyManager`
+
+    :param _proxy_headers:
+        A dictionary with proxy headers, should not be used directly,
+        instead, see :class:`urllib3.ProxyManager`
+
+    :param \\**conn_kw:
+        Additional parameters are used to create fresh :class:`urllib3.connection.HTTPConnection`,
+        :class:`urllib3.connection.HTTPSConnection` instances.
+    """
+
+    scheme = "http"
+    ConnectionCls: type[BaseHTTPConnection] | type[BaseHTTPSConnection] = HTTPConnection
+
+    def __init__(
+        self,
+        host: str,
+        port: int | None = None,
+        timeout: _TYPE_TIMEOUT | None = _DEFAULT_TIMEOUT,
+        maxsize: int = 1,
+        block: bool = False,
+        headers: typing.Mapping[str, str] | None = None,
+        retries: Retry | bool | int | None = None,
+        _proxy: Url | None = None,
+        _proxy_headers: typing.Mapping[str, str] | None = None,
+        _proxy_config: ProxyConfig | None = None,
+        **conn_kw: typing.Any,
+    ):
+        ConnectionPool.__init__(self, host, port)
+        RequestMethods.__init__(self, headers)
+
+        if not isinstance(timeout, Timeout):
+            timeout = Timeout.from_float(timeout)
+
+        if retries is None:
+            retries = Retry.DEFAULT
+
+        self.timeout = timeout
+        self.retries = retries
+
+        self.pool: queue.LifoQueue[typing.Any] | None = self.QueueCls(maxsize)
+        self.block = block
+
+        self.proxy = _proxy
+        self.proxy_headers = _proxy_headers or {}
+        self.proxy_config = _proxy_config
+
+        # Fill the queue up so that doing get() on it will block properly
+        for _ in range(maxsize):
+            self.pool.put(None)
+
+        # These are mostly for testing and debugging purposes.
+        self.num_connections = 0
+        self.num_requests = 0
+        self.conn_kw = conn_kw
+
+        if self.proxy:
+            # Enable Nagle's algorithm for proxies, to avoid packet fragmentation.
+            # We cannot know if the user has added default socket options, so we cannot replace the
+            # list.
+            self.conn_kw.setdefault("socket_options", [])
+
+            self.conn_kw["proxy"] = self.proxy
+            self.conn_kw["proxy_config"] = self.proxy_config
+
+        # Do not pass 'self' as callback to 'finalize'.
+        # Then the 'finalize' would keep an endless living (leak) to self.
+        # By just passing a reference to the pool allows the garbage collector
+        # to free self if nobody else has a reference to it.
+        pool = self.pool
+
+        # Close all the HTTPConnections in the pool before the
+        # HTTPConnectionPool object is garbage collected.
+        weakref.finalize(self, _close_pool_connections, pool)
+
+    def _new_conn(self) -> BaseHTTPConnection:
+        """
+        Return a fresh :class:`HTTPConnection`.
+        """
+        self.num_connections += 1
+        log.debug(
+            "Starting new HTTP connection (%d): %s:%s",
+            self.num_connections,
+            self.host,
+            self.port or "80",
+        )
+
+        conn = self.ConnectionCls(
+            host=self.host,
+            port=self.port,
+            timeout=self.timeout.connect_timeout,
+            **self.conn_kw,
+        )
+        return conn
+
+    def _get_conn(self, timeout: float | None = None) -> BaseHTTPConnection:
+        """
+        Get a connection. Will return a pooled connection if one is available.
+
+        If no connections are available and :prop:`.block` is ``False``, then a
+        fresh connection is returned.
+
+        :param timeout:
+            Seconds to wait before giving up and raising
+            :class:`urllib3.exceptions.EmptyPoolError` if the pool is empty and
+            :prop:`.block` is ``True``.
+        """
+        conn = None
+
+        if self.pool is None:
+            raise ClosedPoolError(self, "Pool is closed.")
+
+        try:
+            conn = self.pool.get(block=self.block, timeout=timeout)
+
+        except AttributeError:  # self.pool is None
+            raise ClosedPoolError(self, "Pool is closed.") from None  # Defensive:
+
+        except queue.Empty:
+            if self.block:
+                raise EmptyPoolError(
+                    self,
+                    "Pool is empty and a new connection can't be opened due to blocking mode.",
+                ) from None
+            pass  # Oh well, we'll create a new connection then
+
+        # If this is a persistent connection, check if it got disconnected
+        if conn and is_connection_dropped(conn):
+            log.debug("Resetting dropped connection: %s", self.host)
+            conn.close()
+
+        return conn or self._new_conn()
+
+    def _put_conn(self, conn: BaseHTTPConnection | None) -> None:
+        """
+        Put a connection back into the pool.
+
+        :param conn:
+            Connection object for the current host and port as returned by
+            :meth:`._new_conn` or :meth:`._get_conn`.
+
+        If the pool is already full, the connection is closed and discarded
+        because we exceeded maxsize. If connections are discarded frequently,
+        then maxsize should be increased.
+
+        If the pool is closed, then the connection will be closed and discarded.
+        """
+        if self.pool is not None:
+            try:
+                self.pool.put(conn, block=False)
+                return  # Everything is dandy, done.
+            except AttributeError:
+                # self.pool is None.
+                pass
+            except queue.Full:
+                # Connection never got put back into the pool, close it.
+                if conn:
+                    conn.close()
+
+                if self.block:
+                    # This should never happen if you got the conn from self._get_conn
+                    raise FullPoolError(
+                        self,
+                        "Pool reached maximum size and no more connections are allowed.",
+                    ) from None
+
+                log.warning(
+                    "Connection pool is full, discarding connection: %s. Connection pool size: %s",
+                    self.host,
+                    self.pool.qsize(),
+                )
+
+        # Connection never got put back into the pool, close it.
+        if conn:
+            conn.close()
+
+    def _validate_conn(self, conn: BaseHTTPConnection) -> None:
+        """
+        Called right before a request is made, after the socket is created.
+        """
+
+    def _prepare_proxy(self, conn: BaseHTTPConnection) -> None:
+        # Nothing to do for HTTP connections.
+        pass
+
+    def _get_timeout(self, timeout: _TYPE_TIMEOUT) -> Timeout:
+        """Helper that always returns a :class:`urllib3.util.Timeout`"""
+        if timeout is _DEFAULT_TIMEOUT:
+            return self.timeout.clone()
+
+        if isinstance(timeout, Timeout):
+            return timeout.clone()
+        else:
+            # User passed us an int/float. This is for backwards compatibility,
+            # can be removed later
+            return Timeout.from_float(timeout)
+
+    def _raise_timeout(
+        self,
+        err: BaseSSLError | OSError | SocketTimeout,
+        url: str,
+        timeout_value: _TYPE_TIMEOUT | None,
+    ) -> None:
+        """Is the error actually a timeout? Will raise a ReadTimeout or pass"""
+
+        if isinstance(err, SocketTimeout):
+            raise ReadTimeoutError(
+                self, url, f"Read timed out. (read timeout={timeout_value})"
+            ) from err
+
+        # See the above comment about EAGAIN in Python 3.
+        if hasattr(err, "errno") and err.errno in _blocking_errnos:
+            raise ReadTimeoutError(
+                self, url, f"Read timed out. (read timeout={timeout_value})"
+            ) from err
+
+    def _make_request(
+        self,
+        conn: BaseHTTPConnection,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        retries: Retry | None = None,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        chunked: bool = False,
+        response_conn: BaseHTTPConnection | None = None,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        enforce_content_length: bool = True,
+    ) -> BaseHTTPResponse:
+        """
+        Perform a request on a given urllib connection object taken from our
+        pool.
+
+        :param conn:
+            a connection from one of our connection pools
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param body:
+            Data to send in the request body, either :class:`str`, :class:`bytes`,
+            an iterable of :class:`str`/:class:`bytes`, or a file-like object.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+
+        :param retries:
+            Configure the number of retries to allow before raising a
+            :class:`~urllib3.exceptions.MaxRetryError` exception.
+
+            Pass ``None`` to retry until you receive a response. Pass a
+            :class:`~urllib3.util.retry.Retry` object for fine-grained control
+            over different types of retries.
+            Pass an integer number to retry connection errors that many times,
+            but no other types of errors. Pass zero to never retry.
+
+            If ``False``, then retries are disabled and any exception is raised
+            immediately. Also, instead of raising a MaxRetryError on redirects,
+            the redirect response will be returned.
+
+        :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int.
+
+        :param timeout:
+            If specified, overrides the default timeout for this one
+            request. It may be a float (in seconds) or an instance of
+            :class:`urllib3.util.Timeout`.
+
+        :param chunked:
+            If True, urllib3 will send the body using chunked transfer
+            encoding. Otherwise, urllib3 will send the body using the standard
+            content-length form. Defaults to False.
+
+        :param response_conn:
+            Set this to ``None`` if you will handle releasing the connection or
+            set the connection to have the response release it.
+
+        :param preload_content:
+          If True, the response's body will be preloaded during construction.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+
+        :param enforce_content_length:
+            Enforce content length checking. Body returned by server must match
+            value of Content-Length header, if present. Otherwise, raise error.
+        """
+        self.num_requests += 1
+
+        timeout_obj = self._get_timeout(timeout)
+        timeout_obj.start_connect()
+        conn.timeout = Timeout.resolve_default_timeout(timeout_obj.connect_timeout)
+
+        try:
+            # Trigger any extra validation we need to do.
+            try:
+                self._validate_conn(conn)
+            except (SocketTimeout, BaseSSLError) as e:
+                self._raise_timeout(err=e, url=url, timeout_value=conn.timeout)
+                raise
+
+        # _validate_conn() starts the connection to an HTTPS proxy
+        # so we need to wrap errors with 'ProxyError' here too.
+        except (
+            OSError,
+            NewConnectionError,
+            TimeoutError,
+            BaseSSLError,
+            CertificateError,
+            SSLError,
+        ) as e:
+            new_e: Exception = e
+            if isinstance(e, (BaseSSLError, CertificateError)):
+                new_e = SSLError(e)
+            # If the connection didn't successfully connect to it's proxy
+            # then there
+            if isinstance(
+                new_e, (OSError, NewConnectionError, TimeoutError, SSLError)
+            ) and (conn and conn.proxy and not conn.has_connected_to_proxy):
+                new_e = _wrap_proxy_error(new_e, conn.proxy.scheme)
+            raise new_e
+
+        # conn.request() calls http.client.*.request, not the method in
+        # urllib3.request. It also calls makefile (recv) on the socket.
+        try:
+            conn.request(
+                method,
+                url,
+                body=body,
+                headers=headers,
+                chunked=chunked,
+                preload_content=preload_content,
+                decode_content=decode_content,
+                enforce_content_length=enforce_content_length,
+            )
+
+        # We are swallowing BrokenPipeError (errno.EPIPE) since the server is
+        # legitimately able to close the connection after sending a valid response.
+        # With this behaviour, the received response is still readable.
+        except BrokenPipeError:
+            pass
+        except OSError as e:
+            # MacOS/Linux
+            # EPROTOTYPE and ECONNRESET are needed on macOS
+            # https://erickt.github.io/blog/2014/11/19/adventures-in-debugging-a-potential-osx-kernel-bug/
+            # Condition changed later to emit ECONNRESET instead of only EPROTOTYPE.
+            if e.errno != errno.EPROTOTYPE and e.errno != errno.ECONNRESET:
+                raise
+
+        # Reset the timeout for the recv() on the socket
+        read_timeout = timeout_obj.read_timeout
+
+        if not conn.is_closed:
+            # In Python 3 socket.py will catch EAGAIN and return None when you
+            # try and read into the file pointer created by http.client, which
+            # instead raises a BadStatusLine exception. Instead of catching
+            # the exception and assuming all BadStatusLine exceptions are read
+            # timeouts, check for a zero timeout before making the request.
+            if read_timeout == 0:
+                raise ReadTimeoutError(
+                    self, url, f"Read timed out. (read timeout={read_timeout})"
+                )
+            conn.timeout = read_timeout
+
+        # Receive the response from the server
+        try:
+            response = conn.getresponse()
+        except (BaseSSLError, OSError) as e:
+            self._raise_timeout(err=e, url=url, timeout_value=read_timeout)
+            raise
+
+        # Set properties that are used by the pooling layer.
+        response.retries = retries
+        response._connection = response_conn  # type: ignore[attr-defined]
+        response._pool = self  # type: ignore[attr-defined]
+
+        log.debug(
+            '%s://%s:%s "%s %s %s" %s %s',
+            self.scheme,
+            self.host,
+            self.port,
+            method,
+            url,
+            response.version_string,
+            response.status,
+            response.length_remaining,
+        )
+
+        return response
+
+    def close(self) -> None:
+        """
+        Close all pooled connections and disable the pool.
+        """
+        if self.pool is None:
+            return
+        # Disable access to the pool
+        old_pool, self.pool = self.pool, None
+
+        # Close all the HTTPConnections in the pool.
+        _close_pool_connections(old_pool)
+
+    def is_same_host(self, url: str) -> bool:
+        """
+        Check if the given ``url`` is a member of the same host as this
+        connection pool.
+        """
+        if url.startswith("/"):
+            return True
+
+        # TODO: Add optional support for socket.gethostbyname checking.
+        scheme, _, host, port, *_ = parse_url(url)
+        scheme = scheme or "http"
+        if host is not None:
+            host = _normalize_host(host, scheme=scheme)
+
+        # Use explicit default port for comparison when none is given
+        if self.port and not port:
+            port = port_by_scheme.get(scheme)
+        elif not self.port and port == port_by_scheme.get(scheme):
+            port = None
+
+        return (scheme, host, port) == (self.scheme, self.host, self.port)
+
+    def urlopen(  # type: ignore[override]
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        retries: Retry | bool | int | None = None,
+        redirect: bool = True,
+        assert_same_host: bool = True,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        pool_timeout: int | None = None,
+        release_conn: bool | None = None,
+        chunked: bool = False,
+        body_pos: _TYPE_BODY_POSITION | None = None,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        **response_kw: typing.Any,
+    ) -> BaseHTTPResponse:
+        """
+        Get a connection from the pool and perform an HTTP request. This is the
+        lowest level call for making a request, so you'll need to specify all
+        the raw details.
+
+        .. note::
+
+           More commonly, it's appropriate to use a convenience method
+           such as :meth:`request`.
+
+        .. note::
+
+           `release_conn` will only behave as expected if
+           `preload_content=False` because we want to make
+           `preload_content=False` the default behaviour someday soon without
+           breaking backwards compatibility.
+
+        :param method:
+            HTTP request method (such as GET, POST, PUT, etc.)
+
+        :param url:
+            The URL to perform the request on.
+
+        :param body:
+            Data to send in the request body, either :class:`str`, :class:`bytes`,
+            an iterable of :class:`str`/:class:`bytes`, or a file-like object.
+
+        :param headers:
+            Dictionary of custom headers to send, such as User-Agent,
+            If-None-Match, etc. If None, pool headers are used. If provided,
+            these headers completely replace any pool-specific headers.
+
+        :param retries:
+            Configure the number of retries to allow before raising a
+            :class:`~urllib3.exceptions.MaxRetryError` exception.
+
+            If ``None`` (default) will retry 3 times, see ``Retry.DEFAULT``. Pass a
+            :class:`~urllib3.util.retry.Retry` object for fine-grained control
+            over different types of retries.
+            Pass an integer number to retry connection errors that many times,
+            but no other types of errors. Pass zero to never retry.
+
+            If ``False``, then retries are disabled and any exception is raised
+            immediately. Also, instead of raising a MaxRetryError on redirects,
+            the redirect response will be returned.
+
+        :type retries: :class:`~urllib3.util.retry.Retry`, False, or an int.
+
+        :param redirect:
+            If True, automatically handle redirects (status codes 301, 302,
+            303, 307, 308). Each redirect counts as a retry. Disabling retries
+            will disable redirect, too.
+
+        :param assert_same_host:
+            If ``True``, will make sure that the host of the pool requests is
+            consistent else will raise HostChangedError. When ``False``, you can
+            use the pool on an HTTP proxy and request foreign hosts.
+
+        :param timeout:
+            If specified, overrides the default timeout for this one
+            request. It may be a float (in seconds) or an instance of
+            :class:`urllib3.util.Timeout`.
+
+        :param pool_timeout:
+            If set and the pool is set to block=True, then this method will
+            block for ``pool_timeout`` seconds and raise EmptyPoolError if no
+            connection is available within the time period.
+
+        :param bool preload_content:
+            If True, the response's body will be preloaded into memory.
+
+        :param bool decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+
+        :param release_conn:
+            If False, then the urlopen call will not release the connection
+            back into the pool once a response is received (but will release if
+            you read the entire contents of the response such as when
+            `preload_content=True`). This is useful if you're not preloading
+            the response's content immediately. You will need to call
+            ``r.release_conn()`` on the response ``r`` to return the connection
+            back into the pool. If None, it takes the value of ``preload_content``
+            which defaults to ``True``.
+
+        :param bool chunked:
+            If True, urllib3 will send the body using chunked transfer
+            encoding. Otherwise, urllib3 will send the body using the standard
+            content-length form. Defaults to False.
+
+        :param int body_pos:
+            Position to seek to in file-like body in the event of a retry or
+            redirect. Typically this won't need to be set because urllib3 will
+            auto-populate the value when needed.
+        """
+        parsed_url = parse_url(url)
+        destination_scheme = parsed_url.scheme
+
+        if headers is None:
+            headers = self.headers
+
+        if not isinstance(retries, Retry):
+            retries = Retry.from_int(retries, redirect=redirect, default=self.retries)
+
+        if release_conn is None:
+            release_conn = preload_content
+
+        # Check host
+        if assert_same_host and not self.is_same_host(url):
+            raise HostChangedError(self, url, retries)
+
+        # Ensure that the URL we're connecting to is properly encoded
+        if url.startswith("/"):
+            url = to_str(_encode_target(url))
+        else:
+            url = to_str(parsed_url.url)
+
+        conn = None
+
+        # Track whether `conn` needs to be released before
+        # returning/raising/recursing. Update this variable if necessary, and
+        # leave `release_conn` constant throughout the function. That way, if
+        # the function recurses, the original value of `release_conn` will be
+        # passed down into the recursive call, and its value will be respected.
+        #
+        # See issue #651 [1] for details.
+        #
+        # [1] <https://github.com/urllib3/urllib3/issues/651>
+        release_this_conn = release_conn
+
+        http_tunnel_required = connection_requires_http_tunnel(
+            self.proxy, self.proxy_config, destination_scheme
+        )
+
+        # Merge the proxy headers. Only done when not using HTTP CONNECT. We
+        # have to copy the headers dict so we can safely change it without those
+        # changes being reflected in anyone else's copy.
+        if not http_tunnel_required:
+            headers = headers.copy()  # type: ignore[attr-defined]
+            headers.update(self.proxy_headers)  # type: ignore[union-attr]
+
+        # Must keep the exception bound to a separate variable or else Python 3
+        # complains about UnboundLocalError.
+        err = None
+
+        # Keep track of whether we cleanly exited the except block. This
+        # ensures we do proper cleanup in finally.
+        clean_exit = False
+
+        # Rewind body position, if needed. Record current position
+        # for future rewinds in the event of a redirect/retry.
+        body_pos = set_file_position(body, body_pos)
+
+        try:
+            # Request a connection from the queue.
+            timeout_obj = self._get_timeout(timeout)
+            conn = self._get_conn(timeout=pool_timeout)
+
+            conn.timeout = timeout_obj.connect_timeout  # type: ignore[assignment]
+
+            # Is this a closed/new connection that requires CONNECT tunnelling?
+            if self.proxy is not None and http_tunnel_required and conn.is_closed:
+                try:
+                    self._prepare_proxy(conn)
+                except (BaseSSLError, OSError, SocketTimeout) as e:
+                    self._raise_timeout(
+                        err=e, url=self.proxy.url, timeout_value=conn.timeout
+                    )
+                    raise
+
+            # If we're going to release the connection in ``finally:``, then
+            # the response doesn't need to know about the connection. Otherwise
+            # it will also try to release it and we'll have a double-release
+            # mess.
+            response_conn = conn if not release_conn else None
+
+            # Make the request on the HTTPConnection object
+            response = self._make_request(
+                conn,
+                method,
+                url,
+                timeout=timeout_obj,
+                body=body,
+                headers=headers,
+                chunked=chunked,
+                retries=retries,
+                response_conn=response_conn,
+                preload_content=preload_content,
+                decode_content=decode_content,
+                **response_kw,
+            )
+
+            # Everything went great!
+            clean_exit = True
+
+        except EmptyPoolError:
+            # Didn't get a connection from the pool, no need to clean up
+            clean_exit = True
+            release_this_conn = False
+            raise
+
+        except (
+            TimeoutError,
+            HTTPException,
+            OSError,
+            ProtocolError,
+            BaseSSLError,
+            SSLError,
+            CertificateError,
+            ProxyError,
+        ) as e:
+            # Discard the connection for these exceptions. It will be
+            # replaced during the next _get_conn() call.
+            clean_exit = False
+            new_e: Exception = e
+            if isinstance(e, (BaseSSLError, CertificateError)):
+                new_e = SSLError(e)
+            if isinstance(
+                new_e,
+                (
+                    OSError,
+                    NewConnectionError,
+                    TimeoutError,
+                    SSLError,
+                    HTTPException,
+                ),
+            ) and (conn and conn.proxy and not conn.has_connected_to_proxy):
+                new_e = _wrap_proxy_error(new_e, conn.proxy.scheme)
+            elif isinstance(new_e, (OSError, HTTPException)):
+                new_e = ProtocolError("Connection aborted.", new_e)
+
+            retries = retries.increment(
+                method, url, error=new_e, _pool=self, _stacktrace=sys.exc_info()[2]
+            )
+            retries.sleep()
+
+            # Keep track of the error for the retry warning.
+            err = e
+
+        finally:
+            if not clean_exit:
+                # We hit some kind of exception, handled or otherwise. We need
+                # to throw the connection away unless explicitly told not to.
+                # Close the connection, set the variable to None, and make sure
+                # we put the None back in the pool to avoid leaking it.
+                if conn:
+                    conn.close()
+                    conn = None
+                release_this_conn = True
+
+            if release_this_conn:
+                # Put the connection back to be reused. If the connection is
+                # expired then it will be None, which will get replaced with a
+                # fresh connection during _get_conn.
+                self._put_conn(conn)
+
+        if not conn:
+            # Try again
+            log.warning(
+                "Retrying (%r) after connection broken by '%r': %s", retries, err, url
+            )
+            return self.urlopen(
+                method,
+                url,
+                body,
+                headers,
+                retries,
+                redirect,
+                assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                preload_content=preload_content,
+                decode_content=decode_content,
+                **response_kw,
+            )
+
+        # Handle redirect?
+        redirect_location = redirect and response.get_redirect_location()
+        if redirect_location:
+            if response.status == 303:
+                # Change the method according to RFC 9110, Section 15.4.4.
+                method = "GET"
+                # And lose the body not to transfer anything sensitive.
+                body = None
+                headers = HTTPHeaderDict(headers)._prepare_for_method_change()
+
+            try:
+                retries = retries.increment(method, url, response=response, _pool=self)
+            except MaxRetryError:
+                if retries.raise_on_redirect:
+                    response.drain_conn()
+                    raise
+                return response
+
+            response.drain_conn()
+            retries.sleep_for_retry(response)
+            log.debug("Redirecting %s -> %s", url, redirect_location)
+            return self.urlopen(
+                method,
+                redirect_location,
+                body,
+                headers,
+                retries=retries,
+                redirect=redirect,
+                assert_same_host=assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                preload_content=preload_content,
+                decode_content=decode_content,
+                **response_kw,
+            )
+
+        # Check if we should retry the HTTP response.
+        has_retry_after = bool(response.headers.get("Retry-After"))
+        if retries.is_retry(method, response.status, has_retry_after):
+            try:
+                retries = retries.increment(method, url, response=response, _pool=self)
+            except MaxRetryError:
+                if retries.raise_on_status:
+                    response.drain_conn()
+                    raise
+                return response
+
+            response.drain_conn()
+            retries.sleep(response)
+            log.debug("Retry: %s", url)
+            return self.urlopen(
+                method,
+                url,
+                body,
+                headers,
+                retries=retries,
+                redirect=redirect,
+                assert_same_host=assert_same_host,
+                timeout=timeout,
+                pool_timeout=pool_timeout,
+                release_conn=release_conn,
+                chunked=chunked,
+                body_pos=body_pos,
+                preload_content=preload_content,
+                decode_content=decode_content,
+                **response_kw,
+            )
+
+        return response
+
+
+class HTTPSConnectionPool(HTTPConnectionPool):
+    """
+    Same as :class:`.HTTPConnectionPool`, but HTTPS.
+
+    :class:`.HTTPSConnection` uses one of ``assert_fingerprint``,
+    ``assert_hostname`` and ``host`` in this order to verify connections.
+    If ``assert_hostname`` is False, no verification is done.
+
+    The ``key_file``, ``cert_file``, ``cert_reqs``, ``ca_certs``,
+    ``ca_cert_dir``, ``ssl_version``, ``key_password`` are only used if :mod:`ssl`
+    is available and are fed into :meth:`urllib3.util.ssl_wrap_socket` to upgrade
+    the connection socket into an SSL socket.
+    """
+
+    scheme = "https"
+    ConnectionCls: type[BaseHTTPSConnection] = HTTPSConnection
+
+    def __init__(
+        self,
+        host: str,
+        port: int | None = None,
+        timeout: _TYPE_TIMEOUT | None = _DEFAULT_TIMEOUT,
+        maxsize: int = 1,
+        block: bool = False,
+        headers: typing.Mapping[str, str] | None = None,
+        retries: Retry | bool | int | None = None,
+        _proxy: Url | None = None,
+        _proxy_headers: typing.Mapping[str, str] | None = None,
+        key_file: str | None = None,
+        cert_file: str | None = None,
+        cert_reqs: int | str | None = None,
+        key_password: str | None = None,
+        ca_certs: str | None = None,
+        ssl_version: int | str | None = None,
+        ssl_minimum_version: ssl.TLSVersion | None = None,
+        ssl_maximum_version: ssl.TLSVersion | None = None,
+        assert_hostname: str | typing.Literal[False] | None = None,
+        assert_fingerprint: str | None = None,
+        ca_cert_dir: str | None = None,
+        **conn_kw: typing.Any,
+    ) -> None:
+        super().__init__(
+            host,
+            port,
+            timeout,
+            maxsize,
+            block,
+            headers,
+            retries,
+            _proxy,
+            _proxy_headers,
+            **conn_kw,
+        )
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.cert_reqs = cert_reqs
+        self.key_password = key_password
+        self.ca_certs = ca_certs
+        self.ca_cert_dir = ca_cert_dir
+        self.ssl_version = ssl_version
+        self.ssl_minimum_version = ssl_minimum_version
+        self.ssl_maximum_version = ssl_maximum_version
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+
+    def _prepare_proxy(self, conn: HTTPSConnection) -> None:  # type: ignore[override]
+        """Establishes a tunnel connection through HTTP CONNECT."""
+        if self.proxy and self.proxy.scheme == "https":
+            tunnel_scheme = "https"
+        else:
+            tunnel_scheme = "http"
+
+        conn.set_tunnel(
+            scheme=tunnel_scheme,
+            host=self._tunnel_host,
+            port=self.port,
+            headers=self.proxy_headers,
+        )
+        conn.connect()
+
+    def _new_conn(self) -> BaseHTTPSConnection:
+        """
+        Return a fresh :class:`urllib3.connection.HTTPConnection`.
+        """
+        self.num_connections += 1
+        log.debug(
+            "Starting new HTTPS connection (%d): %s:%s",
+            self.num_connections,
+            self.host,
+            self.port or "443",
+        )
+
+        if not self.ConnectionCls or self.ConnectionCls is DummyConnection:  # type: ignore[comparison-overlap]
+            raise ImportError(
+                "Can't connect to HTTPS URL because the SSL module is not available."
+            )
+
+        actual_host: str = self.host
+        actual_port = self.port
+        if self.proxy is not None and self.proxy.host is not None:
+            actual_host = self.proxy.host
+            actual_port = self.proxy.port
+
+        return self.ConnectionCls(
+            host=actual_host,
+            port=actual_port,
+            timeout=self.timeout.connect_timeout,
+            cert_file=self.cert_file,
+            key_file=self.key_file,
+            key_password=self.key_password,
+            cert_reqs=self.cert_reqs,
+            ca_certs=self.ca_certs,
+            ca_cert_dir=self.ca_cert_dir,
+            assert_hostname=self.assert_hostname,
+            assert_fingerprint=self.assert_fingerprint,
+            ssl_version=self.ssl_version,
+            ssl_minimum_version=self.ssl_minimum_version,
+            ssl_maximum_version=self.ssl_maximum_version,
+            **self.conn_kw,
+        )
+
+    def _validate_conn(self, conn: BaseHTTPConnection) -> None:
+        """
+        Called right before a request is made, after the socket is created.
+        """
+        super()._validate_conn(conn)
+
+        # Force connect early to allow us to validate the connection.
+        if conn.is_closed:
+            conn.connect()
+
+        # TODO revise this, see https://github.com/urllib3/urllib3/issues/2791
+        if not conn.is_verified and not conn.proxy_is_verified:
+            warnings.warn(
+                (
+                    f"Unverified HTTPS request is being made to host '{conn.host}'. "
+                    "Adding certificate verification is strongly advised. See: "
+                    "https://urllib3.readthedocs.io/en/latest/advanced-usage.html"
+                    "#tls-warnings"
+                ),
+                InsecureRequestWarning,
+            )
+
+
+def connection_from_url(url: str, **kw: typing.Any) -> HTTPConnectionPool:
+    """
+    Given a url, return an :class:`.ConnectionPool` instance of its host.
+
+    This is a shortcut for not having to parse out the scheme, host, and port
+    of the url before creating an :class:`.ConnectionPool` instance.
+
+    :param url:
+        Absolute URL string that must include the scheme. Port is optional.
+
+    :param \\**kw:
+        Passes additional parameters to the constructor of the appropriate
+        :class:`.ConnectionPool`. Useful for specifying things like
+        timeout, maxsize, headers, etc.
+
+    Example::
+
+        >>> conn = connection_from_url('http://google.com/')
+        >>> r = conn.request('GET', '/')
+    """
+    scheme, _, host, port, *_ = parse_url(url)
+    scheme = scheme or "http"
+    port = port or port_by_scheme.get(scheme, 80)
+    if scheme == "https":
+        return HTTPSConnectionPool(host, port=port, **kw)  # type: ignore[arg-type]
+    else:
+        return HTTPConnectionPool(host, port=port, **kw)  # type: ignore[arg-type]
+
+
+@typing.overload
+def _normalize_host(host: None, scheme: str | None) -> None: ...
+
+
+@typing.overload
+def _normalize_host(host: str, scheme: str | None) -> str: ...
+
+
+def _normalize_host(host: str | None, scheme: str | None) -> str | None:
+    """
+    Normalize hosts for comparisons and use with sockets.
+    """
+
+    host = normalize_host(host, scheme)
+
+    # httplib doesn't like it when we include brackets in IPv6 addresses
+    # Specifically, if we include brackets but also pass the port then
+    # httplib crazily doubles up the square brackets on the Host header.
+    # Instead, we need to make sure we never pass ``None`` as the port.
+    # However, for backward compatibility reasons we can't actually
+    # *assert* that.  See http://bugs.python.org/issue28539
+    if host and host.startswith("[") and host.endswith("]"):
+        host = host[1:-1]
+    return host
+
+
+def _url_from_pool(
+    pool: HTTPConnectionPool | HTTPSConnectionPool, path: str | None = None
+) -> str:
+    """Returns the URL from a given connection pool. This is mainly used for testing and logging."""
+    return Url(scheme=pool.scheme, host=pool.host, port=pool.port, path=path).url
+
+
+def _close_pool_connections(pool: queue.LifoQueue[typing.Any]) -> None:
+    """Drains a queue of connections and closes each one."""
+    try:
+        while True:
+            conn = pool.get(block=False)
+            if conn:
+                conn.close()
+    except queue.Empty:
+        pass  # Done.
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__init__.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b9e46eca
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc
new file mode 100644
index 00000000..673fb1de
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/pyopenssl.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/socks.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/socks.cpython-312.pyc
new file mode 100644
index 00000000..f0ff46e8
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/__pycache__/socks.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__init__.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__init__.py
new file mode 100644
index 00000000..8a3c5beb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__init__.py
@@ -0,0 +1,16 @@
+from __future__ import annotations
+
+import urllib3.connection
+
+from ...connectionpool import HTTPConnectionPool, HTTPSConnectionPool
+from .connection import EmscriptenHTTPConnection, EmscriptenHTTPSConnection
+
+
+def inject_into_urllib3() -> None:
+    # override connection classes to use emscripten specific classes
+    # n.b. mypy complains about the overriding of classes below
+    # if it isn't ignored
+    HTTPConnectionPool.ConnectionCls = EmscriptenHTTPConnection
+    HTTPSConnectionPool.ConnectionCls = EmscriptenHTTPSConnection
+    urllib3.connection.HTTPConnection = EmscriptenHTTPConnection  # type: ignore[misc,assignment]
+    urllib3.connection.HTTPSConnection = EmscriptenHTTPSConnection  # type: ignore[misc,assignment]
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..8ab0289c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..7116eb56
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/fetch.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/fetch.cpython-312.pyc
new file mode 100644
index 00000000..3c4759fd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/fetch.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..8deadc6e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..03764f32
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/connection.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/connection.py
new file mode 100644
index 00000000..41bfd279
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/connection.py
@@ -0,0 +1,255 @@
+from __future__ import annotations
+
+import os
+import typing
+
+# use http.client.HTTPException for consistency with non-emscripten
+from http.client import HTTPException as HTTPException  # noqa: F401
+from http.client import ResponseNotReady
+
+from ..._base_connection import _TYPE_BODY
+from ...connection import HTTPConnection, ProxyConfig, port_by_scheme
+from ...exceptions import TimeoutError
+from ...response import BaseHTTPResponse
+from ...util.connection import _TYPE_SOCKET_OPTIONS
+from ...util.timeout import _DEFAULT_TIMEOUT, _TYPE_TIMEOUT
+from ...util.url import Url
+from .fetch import _RequestError, _TimeoutError, send_request, send_streaming_request
+from .request import EmscriptenRequest
+from .response import EmscriptenHttpResponseWrapper, EmscriptenResponse
+
+if typing.TYPE_CHECKING:
+    from ..._base_connection import BaseHTTPConnection, BaseHTTPSConnection
+
+
+class EmscriptenHTTPConnection:
+    default_port: typing.ClassVar[int] = port_by_scheme["http"]
+    default_socket_options: typing.ClassVar[_TYPE_SOCKET_OPTIONS]
+
+    timeout: None | (float)
+
+    host: str
+    port: int
+    blocksize: int
+    source_address: tuple[str, int] | None
+    socket_options: _TYPE_SOCKET_OPTIONS | None
+
+    proxy: Url | None
+    proxy_config: ProxyConfig | None
+
+    is_verified: bool = False
+    proxy_is_verified: bool | None = None
+
+    _response: EmscriptenResponse | None
+
+    def __init__(
+        self,
+        host: str,
+        port: int = 0,
+        *,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        source_address: tuple[str, int] | None = None,
+        blocksize: int = 8192,
+        socket_options: _TYPE_SOCKET_OPTIONS | None = None,
+        proxy: Url | None = None,
+        proxy_config: ProxyConfig | None = None,
+    ) -> None:
+        self.host = host
+        self.port = port
+        self.timeout = timeout if isinstance(timeout, float) else 0.0
+        self.scheme = "http"
+        self._closed = True
+        self._response = None
+        # ignore these things because we don't
+        # have control over that stuff
+        self.proxy = None
+        self.proxy_config = None
+        self.blocksize = blocksize
+        self.source_address = None
+        self.socket_options = None
+        self.is_verified = False
+
+    def set_tunnel(
+        self,
+        host: str,
+        port: int | None = 0,
+        headers: typing.Mapping[str, str] | None = None,
+        scheme: str = "http",
+    ) -> None:
+        pass
+
+    def connect(self) -> None:
+        pass
+
+    def request(
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        # We know *at least* botocore is depending on the order of the
+        # first 3 parameters so to be safe we only mark the later ones
+        # as keyword-only to ensure we have space to extend.
+        *,
+        chunked: bool = False,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        enforce_content_length: bool = True,
+    ) -> None:
+        self._closed = False
+        if url.startswith("/"):
+            # no scheme / host / port included, make a full url
+            url = f"{self.scheme}://{self.host}:{self.port}" + url
+        request = EmscriptenRequest(
+            url=url,
+            method=method,
+            timeout=self.timeout if self.timeout else 0,
+            decode_content=decode_content,
+        )
+        request.set_body(body)
+        if headers:
+            for k, v in headers.items():
+                request.set_header(k, v)
+        self._response = None
+        try:
+            if not preload_content:
+                self._response = send_streaming_request(request)
+            if self._response is None:
+                self._response = send_request(request)
+        except _TimeoutError as e:
+            raise TimeoutError(e.message) from e
+        except _RequestError as e:
+            raise HTTPException(e.message) from e
+
+    def getresponse(self) -> BaseHTTPResponse:
+        if self._response is not None:
+            return EmscriptenHttpResponseWrapper(
+                internal_response=self._response,
+                url=self._response.request.url,
+                connection=self,
+            )
+        else:
+            raise ResponseNotReady()
+
+    def close(self) -> None:
+        self._closed = True
+        self._response = None
+
+    @property
+    def is_closed(self) -> bool:
+        """Whether the connection either is brand new or has been previously closed.
+        If this property is True then both ``is_connected`` and ``has_connected_to_proxy``
+        properties must be False.
+        """
+        return self._closed
+
+    @property
+    def is_connected(self) -> bool:
+        """Whether the connection is actively connected to any origin (proxy or target)"""
+        return True
+
+    @property
+    def has_connected_to_proxy(self) -> bool:
+        """Whether the connection has successfully connected to its proxy.
+        This returns False if no proxy is in use. Used to determine whether
+        errors are coming from the proxy layer or from tunnelling to the target origin.
+        """
+        return False
+
+
+class EmscriptenHTTPSConnection(EmscriptenHTTPConnection):
+    default_port = port_by_scheme["https"]
+    # all this is basically ignored, as browser handles https
+    cert_reqs: int | str | None = None
+    ca_certs: str | None = None
+    ca_cert_dir: str | None = None
+    ca_cert_data: None | str | bytes = None
+    cert_file: str | None
+    key_file: str | None
+    key_password: str | None
+    ssl_context: typing.Any | None
+    ssl_version: int | str | None = None
+    ssl_minimum_version: int | None = None
+    ssl_maximum_version: int | None = None
+    assert_hostname: None | str | typing.Literal[False]
+    assert_fingerprint: str | None = None
+
+    def __init__(
+        self,
+        host: str,
+        port: int = 0,
+        *,
+        timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        source_address: tuple[str, int] | None = None,
+        blocksize: int = 16384,
+        socket_options: (
+            None | _TYPE_SOCKET_OPTIONS
+        ) = HTTPConnection.default_socket_options,
+        proxy: Url | None = None,
+        proxy_config: ProxyConfig | None = None,
+        cert_reqs: int | str | None = None,
+        assert_hostname: None | str | typing.Literal[False] = None,
+        assert_fingerprint: str | None = None,
+        server_hostname: str | None = None,
+        ssl_context: typing.Any | None = None,
+        ca_certs: str | None = None,
+        ca_cert_dir: str | None = None,
+        ca_cert_data: None | str | bytes = None,
+        ssl_minimum_version: int | None = None,
+        ssl_maximum_version: int | None = None,
+        ssl_version: int | str | None = None,  # Deprecated
+        cert_file: str | None = None,
+        key_file: str | None = None,
+        key_password: str | None = None,
+    ) -> None:
+        super().__init__(
+            host,
+            port=port,
+            timeout=timeout,
+            source_address=source_address,
+            blocksize=blocksize,
+            socket_options=socket_options,
+            proxy=proxy,
+            proxy_config=proxy_config,
+        )
+        self.scheme = "https"
+
+        self.key_file = key_file
+        self.cert_file = cert_file
+        self.key_password = key_password
+        self.ssl_context = ssl_context
+        self.server_hostname = server_hostname
+        self.assert_hostname = assert_hostname
+        self.assert_fingerprint = assert_fingerprint
+        self.ssl_version = ssl_version
+        self.ssl_minimum_version = ssl_minimum_version
+        self.ssl_maximum_version = ssl_maximum_version
+        self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
+        self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
+        self.ca_cert_data = ca_cert_data
+
+        self.cert_reqs = None
+
+        # The browser will automatically verify all requests.
+        # We have no control over that setting.
+        self.is_verified = True
+
+    def set_cert(
+        self,
+        key_file: str | None = None,
+        cert_file: str | None = None,
+        cert_reqs: int | str | None = None,
+        key_password: str | None = None,
+        ca_certs: str | None = None,
+        assert_hostname: None | str | typing.Literal[False] = None,
+        assert_fingerprint: str | None = None,
+        ca_cert_dir: str | None = None,
+        ca_cert_data: None | str | bytes = None,
+    ) -> None:
+        pass
+
+
+# verify that this class implements BaseHTTP(s) connection correctly
+if typing.TYPE_CHECKING:
+    _supports_http_protocol: BaseHTTPConnection = EmscriptenHTTPConnection("", 0)
+    _supports_https_protocol: BaseHTTPSConnection = EmscriptenHTTPSConnection("", 0)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/emscripten_fetch_worker.js b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/emscripten_fetch_worker.js
new file mode 100644
index 00000000..243b8622
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/emscripten_fetch_worker.js
@@ -0,0 +1,110 @@
+let Status = {
+  SUCCESS_HEADER: -1,
+  SUCCESS_EOF: -2,
+  ERROR_TIMEOUT: -3,
+  ERROR_EXCEPTION: -4,
+};
+
+let connections = {};
+let nextConnectionID = 1;
+const encoder = new TextEncoder();
+
+self.addEventListener("message", async function (event) {
+  if (event.data.close) {
+    let connectionID = event.data.close;
+    delete connections[connectionID];
+    return;
+  } else if (event.data.getMore) {
+    let connectionID = event.data.getMore;
+    let { curOffset, value, reader, intBuffer, byteBuffer } =
+      connections[connectionID];
+    // if we still have some in buffer, then just send it back straight away
+    if (!value || curOffset >= value.length) {
+      // read another buffer if required
+      try {
+        let readResponse = await reader.read();
+
+        if (readResponse.done) {
+          // read everything - clear connection and return
+          delete connections[connectionID];
+          Atomics.store(intBuffer, 0, Status.SUCCESS_EOF);
+          Atomics.notify(intBuffer, 0);
+          // finished reading successfully
+          // return from event handler
+          return;
+        }
+        curOffset = 0;
+        connections[connectionID].value = readResponse.value;
+        value = readResponse.value;
+      } catch (error) {
+        console.log("Request exception:", error);
+        let errorBytes = encoder.encode(error.message);
+        let written = errorBytes.length;
+        byteBuffer.set(errorBytes);
+        intBuffer[1] = written;
+        Atomics.store(intBuffer, 0, Status.ERROR_EXCEPTION);
+        Atomics.notify(intBuffer, 0);
+      }
+    }
+
+    // send as much buffer as we can
+    let curLen = value.length - curOffset;
+    if (curLen > byteBuffer.length) {
+      curLen = byteBuffer.length;
+    }
+    byteBuffer.set(value.subarray(curOffset, curOffset + curLen), 0);
+
+    Atomics.store(intBuffer, 0, curLen); // store current length in bytes
+    Atomics.notify(intBuffer, 0);
+    curOffset += curLen;
+    connections[connectionID].curOffset = curOffset;
+
+    return;
+  } else {
+    // start fetch
+    let connectionID = nextConnectionID;
+    nextConnectionID += 1;
+    const intBuffer = new Int32Array(event.data.buffer);
+    const byteBuffer = new Uint8Array(event.data.buffer, 8);
+    try {
+      const response = await fetch(event.data.url, event.data.fetchParams);
+      // return the headers first via textencoder
+      var headers = [];
+      for (const pair of response.headers.entries()) {
+        headers.push([pair[0], pair[1]]);
+      }
+      let headerObj = {
+        headers: headers,
+        status: response.status,
+        connectionID,
+      };
+      const headerText = JSON.stringify(headerObj);
+      let headerBytes = encoder.encode(headerText);
+      let written = headerBytes.length;
+      byteBuffer.set(headerBytes);
+      intBuffer[1] = written;
+      // make a connection
+      connections[connectionID] = {
+        reader: response.body.getReader(),
+        intBuffer: intBuffer,
+        byteBuffer: byteBuffer,
+        value: undefined,
+        curOffset: 0,
+      };
+      // set header ready
+      Atomics.store(intBuffer, 0, Status.SUCCESS_HEADER);
+      Atomics.notify(intBuffer, 0);
+      // all fetching after this goes through a new postmessage call with getMore
+      // this allows for parallel requests
+    } catch (error) {
+      console.log("Request exception:", error);
+      let errorBytes = encoder.encode(error.message);
+      let written = errorBytes.length;
+      byteBuffer.set(errorBytes);
+      intBuffer[1] = written;
+      Atomics.store(intBuffer, 0, Status.ERROR_EXCEPTION);
+      Atomics.notify(intBuffer, 0);
+    }
+  }
+});
+self.postMessage({ inited: true });
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/fetch.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/fetch.py
new file mode 100644
index 00000000..a514306e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/fetch.py
@@ -0,0 +1,708 @@
+"""
+Support for streaming http requests in emscripten.
+
+A few caveats -
+
+If your browser (or Node.js) has WebAssembly JavaScript Promise Integration enabled
+https://github.com/WebAssembly/js-promise-integration/blob/main/proposals/js-promise-integration/Overview.md
+*and* you launch pyodide using `pyodide.runPythonAsync`, this will fetch data using the
+JavaScript asynchronous fetch api (wrapped via `pyodide.ffi.call_sync`). In this case
+timeouts and streaming should just work.
+
+Otherwise, it uses a combination of XMLHttpRequest and a web-worker for streaming.
+
+This approach has several caveats:
+
+Firstly, you can't do streaming http in the main UI thread, because atomics.wait isn't allowed.
+Streaming only works if you're running pyodide in a web worker.
+
+Secondly, this uses an extra web worker and SharedArrayBuffer to do the asynchronous fetch
+operation, so it requires that you have crossOriginIsolation enabled, by serving over https
+(or from localhost) with the two headers below set:
+
+    Cross-Origin-Opener-Policy: same-origin
+    Cross-Origin-Embedder-Policy: require-corp
+
+You can tell if cross origin isolation is successfully enabled by looking at the global crossOriginIsolated variable in
+JavaScript console. If it isn't, streaming requests will fallback to XMLHttpRequest, i.e. getting the whole
+request into a buffer and then returning it. it shows a warning in the JavaScript console in this case.
+
+Finally, the webworker which does the streaming fetch is created on initial import, but will only be started once
+control is returned to javascript. Call `await wait_for_streaming_ready()` to wait for streaming fetch.
+
+NB: in this code, there are a lot of JavaScript objects. They are named js_*
+to make it clear what type of object they are.
+"""
+
+from __future__ import annotations
+
+import io
+import json
+from email.parser import Parser
+from importlib.resources import files
+from typing import TYPE_CHECKING, Any
+
+import js  # type: ignore[import-not-found]
+from pyodide.ffi import (  # type: ignore[import-not-found]
+    JsArray,
+    JsException,
+    JsProxy,
+    to_js,
+)
+
+if TYPE_CHECKING:
+    from typing_extensions import Buffer
+
+from .request import EmscriptenRequest
+from .response import EmscriptenResponse
+
+"""
+There are some headers that trigger unintended CORS preflight requests.
+See also https://github.com/koenvo/pyodide-http/issues/22
+"""
+HEADERS_TO_IGNORE = ("user-agent",)
+
+SUCCESS_HEADER = -1
+SUCCESS_EOF = -2
+ERROR_TIMEOUT = -3
+ERROR_EXCEPTION = -4
+
+_STREAMING_WORKER_CODE = (
+    files(__package__)
+    .joinpath("emscripten_fetch_worker.js")
+    .read_text(encoding="utf-8")
+)
+
+
+class _RequestError(Exception):
+    def __init__(
+        self,
+        message: str | None = None,
+        *,
+        request: EmscriptenRequest | None = None,
+        response: EmscriptenResponse | None = None,
+    ):
+        self.request = request
+        self.response = response
+        self.message = message
+        super().__init__(self.message)
+
+
+class _StreamingError(_RequestError):
+    pass
+
+
+class _TimeoutError(_RequestError):
+    pass
+
+
+def _obj_from_dict(dict_val: dict[str, Any]) -> JsProxy:
+    return to_js(dict_val, dict_converter=js.Object.fromEntries)
+
+
+class _ReadStream(io.RawIOBase):
+    def __init__(
+        self,
+        int_buffer: JsArray,
+        byte_buffer: JsArray,
+        timeout: float,
+        worker: JsProxy,
+        connection_id: int,
+        request: EmscriptenRequest,
+    ):
+        self.int_buffer = int_buffer
+        self.byte_buffer = byte_buffer
+        self.read_pos = 0
+        self.read_len = 0
+        self.connection_id = connection_id
+        self.worker = worker
+        self.timeout = int(1000 * timeout) if timeout > 0 else None
+        self.is_live = True
+        self._is_closed = False
+        self.request: EmscriptenRequest | None = request
+
+    def __del__(self) -> None:
+        self.close()
+
+    # this is compatible with _base_connection
+    def is_closed(self) -> bool:
+        return self._is_closed
+
+    # for compatibility with RawIOBase
+    @property
+    def closed(self) -> bool:
+        return self.is_closed()
+
+    def close(self) -> None:
+        if self.is_closed():
+            return
+        self.read_len = 0
+        self.read_pos = 0
+        self.int_buffer = None
+        self.byte_buffer = None
+        self._is_closed = True
+        self.request = None
+        if self.is_live:
+            self.worker.postMessage(_obj_from_dict({"close": self.connection_id}))
+            self.is_live = False
+        super().close()
+
+    def readable(self) -> bool:
+        return True
+
+    def writable(self) -> bool:
+        return False
+
+    def seekable(self) -> bool:
+        return False
+
+    def readinto(self, byte_obj: Buffer) -> int:
+        if not self.int_buffer:
+            raise _StreamingError(
+                "No buffer for stream in _ReadStream.readinto",
+                request=self.request,
+                response=None,
+            )
+        if self.read_len == 0:
+            # wait for the worker to send something
+            js.Atomics.store(self.int_buffer, 0, ERROR_TIMEOUT)
+            self.worker.postMessage(_obj_from_dict({"getMore": self.connection_id}))
+            if (
+                js.Atomics.wait(self.int_buffer, 0, ERROR_TIMEOUT, self.timeout)
+                == "timed-out"
+            ):
+                raise _TimeoutError
+            data_len = self.int_buffer[0]
+            if data_len > 0:
+                self.read_len = data_len
+                self.read_pos = 0
+            elif data_len == ERROR_EXCEPTION:
+                string_len = self.int_buffer[1]
+                # decode the error string
+                js_decoder = js.TextDecoder.new()
+                json_str = js_decoder.decode(self.byte_buffer.slice(0, string_len))
+                raise _StreamingError(
+                    f"Exception thrown in fetch: {json_str}",
+                    request=self.request,
+                    response=None,
+                )
+            else:
+                # EOF, free the buffers and return zero
+                # and free the request
+                self.is_live = False
+                self.close()
+                return 0
+        # copy from int32array to python bytes
+        ret_length = min(self.read_len, len(memoryview(byte_obj)))
+        subarray = self.byte_buffer.subarray(
+            self.read_pos, self.read_pos + ret_length
+        ).to_py()
+        memoryview(byte_obj)[0:ret_length] = subarray
+        self.read_len -= ret_length
+        self.read_pos += ret_length
+        return ret_length
+
+
+class _StreamingFetcher:
+    def __init__(self) -> None:
+        # make web-worker and data buffer on startup
+        self.streaming_ready = False
+
+        js_data_blob = js.Blob.new(
+            to_js([_STREAMING_WORKER_CODE], create_pyproxies=False),
+            _obj_from_dict({"type": "application/javascript"}),
+        )
+
+        def promise_resolver(js_resolve_fn: JsProxy, js_reject_fn: JsProxy) -> None:
+            def onMsg(e: JsProxy) -> None:
+                self.streaming_ready = True
+                js_resolve_fn(e)
+
+            def onErr(e: JsProxy) -> None:
+                js_reject_fn(e)  # Defensive: never happens in ci
+
+            self.js_worker.onmessage = onMsg
+            self.js_worker.onerror = onErr
+
+        js_data_url = js.URL.createObjectURL(js_data_blob)
+        self.js_worker = js.globalThis.Worker.new(js_data_url)
+        self.js_worker_ready_promise = js.globalThis.Promise.new(promise_resolver)
+
+    def send(self, request: EmscriptenRequest) -> EmscriptenResponse:
+        headers = {
+            k: v for k, v in request.headers.items() if k not in HEADERS_TO_IGNORE
+        }
+
+        body = request.body
+        fetch_data = {"headers": headers, "body": to_js(body), "method": request.method}
+        # start the request off in the worker
+        timeout = int(1000 * request.timeout) if request.timeout > 0 else None
+        js_shared_buffer = js.SharedArrayBuffer.new(1048576)
+        js_int_buffer = js.Int32Array.new(js_shared_buffer)
+        js_byte_buffer = js.Uint8Array.new(js_shared_buffer, 8)
+
+        js.Atomics.store(js_int_buffer, 0, ERROR_TIMEOUT)
+        js.Atomics.notify(js_int_buffer, 0)
+        js_absolute_url = js.URL.new(request.url, js.location).href
+        self.js_worker.postMessage(
+            _obj_from_dict(
+                {
+                    "buffer": js_shared_buffer,
+                    "url": js_absolute_url,
+                    "fetchParams": fetch_data,
+                }
+            )
+        )
+        # wait for the worker to send something
+        js.Atomics.wait(js_int_buffer, 0, ERROR_TIMEOUT, timeout)
+        if js_int_buffer[0] == ERROR_TIMEOUT:
+            raise _TimeoutError(
+                "Timeout connecting to streaming request",
+                request=request,
+                response=None,
+            )
+        elif js_int_buffer[0] == SUCCESS_HEADER:
+            # got response
+            # header length is in second int of intBuffer
+            string_len = js_int_buffer[1]
+            # decode the rest to a JSON string
+            js_decoder = js.TextDecoder.new()
+            # this does a copy (the slice) because decode can't work on shared array
+            # for some silly reason
+            json_str = js_decoder.decode(js_byte_buffer.slice(0, string_len))
+            # get it as an object
+            response_obj = json.loads(json_str)
+            return EmscriptenResponse(
+                request=request,
+                status_code=response_obj["status"],
+                headers=response_obj["headers"],
+                body=_ReadStream(
+                    js_int_buffer,
+                    js_byte_buffer,
+                    request.timeout,
+                    self.js_worker,
+                    response_obj["connectionID"],
+                    request,
+                ),
+            )
+        elif js_int_buffer[0] == ERROR_EXCEPTION:
+            string_len = js_int_buffer[1]
+            # decode the error string
+            js_decoder = js.TextDecoder.new()
+            json_str = js_decoder.decode(js_byte_buffer.slice(0, string_len))
+            raise _StreamingError(
+                f"Exception thrown in fetch: {json_str}", request=request, response=None
+            )
+        else:
+            raise _StreamingError(
+                f"Unknown status from worker in fetch: {js_int_buffer[0]}",
+                request=request,
+                response=None,
+            )
+
+
+class _JSPIReadStream(io.RawIOBase):
+    """
+    A read stream that uses pyodide.ffi.run_sync to read from a JavaScript fetch
+    response. This requires support for WebAssembly JavaScript Promise Integration
+    in the containing browser, and for pyodide to be launched via runPythonAsync.
+
+    :param js_read_stream:
+        The JavaScript stream reader
+
+    :param timeout:
+        Timeout in seconds
+
+    :param request:
+        The request we're handling
+
+    :param response:
+        The response this stream relates to
+
+    :param js_abort_controller:
+        A JavaScript AbortController object, used for timeouts
+    """
+
+    def __init__(
+        self,
+        js_read_stream: Any,
+        timeout: float,
+        request: EmscriptenRequest,
+        response: EmscriptenResponse,
+        js_abort_controller: Any,  # JavaScript AbortController for timeouts
+    ):
+        self.js_read_stream = js_read_stream
+        self.timeout = timeout
+        self._is_closed = False
+        self._is_done = False
+        self.request: EmscriptenRequest | None = request
+        self.response: EmscriptenResponse | None = response
+        self.current_buffer = None
+        self.current_buffer_pos = 0
+        self.js_abort_controller = js_abort_controller
+
+    def __del__(self) -> None:
+        self.close()
+
+    # this is compatible with _base_connection
+    def is_closed(self) -> bool:
+        return self._is_closed
+
+    # for compatibility with RawIOBase
+    @property
+    def closed(self) -> bool:
+        return self.is_closed()
+
+    def close(self) -> None:
+        if self.is_closed():
+            return
+        self.read_len = 0
+        self.read_pos = 0
+        self.js_read_stream.cancel()
+        self.js_read_stream = None
+        self._is_closed = True
+        self._is_done = True
+        self.request = None
+        self.response = None
+        super().close()
+
+    def readable(self) -> bool:
+        return True
+
+    def writable(self) -> bool:
+        return False
+
+    def seekable(self) -> bool:
+        return False
+
+    def _get_next_buffer(self) -> bool:
+        result_js = _run_sync_with_timeout(
+            self.js_read_stream.read(),
+            self.timeout,
+            self.js_abort_controller,
+            request=self.request,
+            response=self.response,
+        )
+        if result_js.done:
+            self._is_done = True
+            return False
+        else:
+            self.current_buffer = result_js.value.to_py()
+            self.current_buffer_pos = 0
+            return True
+
+    def readinto(self, byte_obj: Buffer) -> int:
+        if self.current_buffer is None:
+            if not self._get_next_buffer() or self.current_buffer is None:
+                self.close()
+                return 0
+        ret_length = min(
+            len(byte_obj), len(self.current_buffer) - self.current_buffer_pos
+        )
+        byte_obj[0:ret_length] = self.current_buffer[
+            self.current_buffer_pos : self.current_buffer_pos + ret_length
+        ]
+        self.current_buffer_pos += ret_length
+        if self.current_buffer_pos == len(self.current_buffer):
+            self.current_buffer = None
+        return ret_length
+
+
+# check if we are in a worker or not
+def is_in_browser_main_thread() -> bool:
+    return hasattr(js, "window") and hasattr(js, "self") and js.self == js.window
+
+
+def is_cross_origin_isolated() -> bool:
+    return hasattr(js, "crossOriginIsolated") and js.crossOriginIsolated
+
+
+def is_in_node() -> bool:
+    return (
+        hasattr(js, "process")
+        and hasattr(js.process, "release")
+        and hasattr(js.process.release, "name")
+        and js.process.release.name == "node"
+    )
+
+
+def is_worker_available() -> bool:
+    return hasattr(js, "Worker") and hasattr(js, "Blob")
+
+
+_fetcher: _StreamingFetcher | None = None
+
+if is_worker_available() and (
+    (is_cross_origin_isolated() and not is_in_browser_main_thread())
+    and (not is_in_node())
+):
+    _fetcher = _StreamingFetcher()
+else:
+    _fetcher = None
+
+
+NODE_JSPI_ERROR = (
+    "urllib3 only works in Node.js with pyodide.runPythonAsync"
+    " and requires the flag --experimental-wasm-stack-switching in "
+    " versions of node <24."
+)
+
+
+def send_streaming_request(request: EmscriptenRequest) -> EmscriptenResponse | None:
+    if has_jspi():
+        return send_jspi_request(request, True)
+    elif is_in_node():
+        raise _RequestError(
+            message=NODE_JSPI_ERROR,
+            request=request,
+            response=None,
+        )
+
+    if _fetcher and streaming_ready():
+        return _fetcher.send(request)
+    else:
+        _show_streaming_warning()
+        return None
+
+
+_SHOWN_TIMEOUT_WARNING = False
+
+
+def _show_timeout_warning() -> None:
+    global _SHOWN_TIMEOUT_WARNING
+    if not _SHOWN_TIMEOUT_WARNING:
+        _SHOWN_TIMEOUT_WARNING = True
+        message = "Warning: Timeout is not available on main browser thread"
+        js.console.warn(message)
+
+
+_SHOWN_STREAMING_WARNING = False
+
+
+def _show_streaming_warning() -> None:
+    global _SHOWN_STREAMING_WARNING
+    if not _SHOWN_STREAMING_WARNING:
+        _SHOWN_STREAMING_WARNING = True
+        message = "Can't stream HTTP requests because: \n"
+        if not is_cross_origin_isolated():
+            message += "  Page is not cross-origin isolated\n"
+        if is_in_browser_main_thread():
+            message += "  Python is running in main browser thread\n"
+        if not is_worker_available():
+            message += " Worker or Blob classes are not available in this environment."  # Defensive: this is always False in browsers that we test in
+        if streaming_ready() is False:
+            message += """ Streaming fetch worker isn't ready. If you want to be sure that streaming fetch
+is working, you need to call: 'await urllib3.contrib.emscripten.fetch.wait_for_streaming_ready()`"""
+        from js import console
+
+        console.warn(message)
+
+
+def send_request(request: EmscriptenRequest) -> EmscriptenResponse:
+    if has_jspi():
+        return send_jspi_request(request, False)
+    elif is_in_node():
+        raise _RequestError(
+            message=NODE_JSPI_ERROR,
+            request=request,
+            response=None,
+        )
+    try:
+        js_xhr = js.XMLHttpRequest.new()
+
+        if not is_in_browser_main_thread():
+            js_xhr.responseType = "arraybuffer"
+            if request.timeout:
+                js_xhr.timeout = int(request.timeout * 1000)
+        else:
+            js_xhr.overrideMimeType("text/plain; charset=ISO-8859-15")
+            if request.timeout:
+                # timeout isn't available on the main thread - show a warning in console
+                # if it is set
+                _show_timeout_warning()
+
+        js_xhr.open(request.method, request.url, False)
+        for name, value in request.headers.items():
+            if name.lower() not in HEADERS_TO_IGNORE:
+                js_xhr.setRequestHeader(name, value)
+
+        js_xhr.send(to_js(request.body))
+
+        headers = dict(Parser().parsestr(js_xhr.getAllResponseHeaders()))
+
+        if not is_in_browser_main_thread():
+            body = js_xhr.response.to_py().tobytes()
+        else:
+            body = js_xhr.response.encode("ISO-8859-15")
+        return EmscriptenResponse(
+            status_code=js_xhr.status, headers=headers, body=body, request=request
+        )
+    except JsException as err:
+        if err.name == "TimeoutError":
+            raise _TimeoutError(err.message, request=request)
+        elif err.name == "NetworkError":
+            raise _RequestError(err.message, request=request)
+        else:
+            # general http error
+            raise _RequestError(err.message, request=request)
+
+
+def send_jspi_request(
+    request: EmscriptenRequest, streaming: bool
+) -> EmscriptenResponse:
+    """
+    Send a request using WebAssembly JavaScript Promise Integration
+    to wrap the asynchronous JavaScript fetch api (experimental).
+
+    :param request:
+        Request to send
+
+    :param streaming:
+        Whether to stream the response
+
+    :return: The response object
+    :rtype: EmscriptenResponse
+    """
+    timeout = request.timeout
+    js_abort_controller = js.AbortController.new()
+    headers = {k: v for k, v in request.headers.items() if k not in HEADERS_TO_IGNORE}
+    req_body = request.body
+    fetch_data = {
+        "headers": headers,
+        "body": to_js(req_body),
+        "method": request.method,
+        "signal": js_abort_controller.signal,
+    }
+    # Call JavaScript fetch (async api, returns a promise)
+    fetcher_promise_js = js.fetch(request.url, _obj_from_dict(fetch_data))
+    # Now suspend WebAssembly until we resolve that promise
+    # or time out.
+    response_js = _run_sync_with_timeout(
+        fetcher_promise_js,
+        timeout,
+        js_abort_controller,
+        request=request,
+        response=None,
+    )
+    headers = {}
+    header_iter = response_js.headers.entries()
+    while True:
+        iter_value_js = header_iter.next()
+        if getattr(iter_value_js, "done", False):
+            break
+        else:
+            headers[str(iter_value_js.value[0])] = str(iter_value_js.value[1])
+    status_code = response_js.status
+    body: bytes | io.RawIOBase = b""
+
+    response = EmscriptenResponse(
+        status_code=status_code, headers=headers, body=b"", request=request
+    )
+    if streaming:
+        # get via inputstream
+        if response_js.body is not None:
+            # get a reader from the fetch response
+            body_stream_js = response_js.body.getReader()
+            body = _JSPIReadStream(
+                body_stream_js, timeout, request, response, js_abort_controller
+            )
+    else:
+        # get directly via arraybuffer
+        # n.b. this is another async JavaScript call.
+        body = _run_sync_with_timeout(
+            response_js.arrayBuffer(),
+            timeout,
+            js_abort_controller,
+            request=request,
+            response=response,
+        ).to_py()
+    response.body = body
+    return response
+
+
+def _run_sync_with_timeout(
+    promise: Any,
+    timeout: float,
+    js_abort_controller: Any,
+    request: EmscriptenRequest | None,
+    response: EmscriptenResponse | None,
+) -> Any:
+    """
+    Await a JavaScript promise synchronously with a timeout which is implemented
+    via the AbortController
+
+    :param promise:
+        Javascript promise to await
+
+    :param timeout:
+        Timeout in seconds
+
+    :param js_abort_controller:
+        A JavaScript AbortController object, used on timeout
+
+    :param request:
+        The request being handled
+
+    :param response:
+        The response being handled (if it exists yet)
+
+    :raises _TimeoutError: If the request times out
+    :raises _RequestError: If the request raises a JavaScript exception
+
+    :return: The result of awaiting the promise.
+    """
+    timer_id = None
+    if timeout > 0:
+        timer_id = js.setTimeout(
+            js_abort_controller.abort.bind(js_abort_controller), int(timeout * 1000)
+        )
+    try:
+        from pyodide.ffi import run_sync
+
+        # run_sync here uses WebAssembly JavaScript Promise Integration to
+        # suspend python until the JavaScript promise resolves.
+        return run_sync(promise)
+    except JsException as err:
+        if err.name == "AbortError":
+            raise _TimeoutError(
+                message="Request timed out", request=request, response=response
+            )
+        else:
+            raise _RequestError(message=err.message, request=request, response=response)
+    finally:
+        if timer_id is not None:
+            js.clearTimeout(timer_id)
+
+
+def has_jspi() -> bool:
+    """
+    Return true if jspi can be used.
+
+    This requires both browser support and also WebAssembly
+    to be in the correct state - i.e. that the javascript
+    call into python was async not sync.
+
+    :return: True if jspi can be used.
+    :rtype: bool
+    """
+    try:
+        from pyodide.ffi import can_run_sync, run_sync  # noqa: F401
+
+        return bool(can_run_sync())
+    except ImportError:
+        return False
+
+
+def streaming_ready() -> bool | None:
+    if _fetcher:
+        return _fetcher.streaming_ready
+    else:
+        return None  # no fetcher, return None to signify that
+
+
+async def wait_for_streaming_ready() -> bool:
+    if _fetcher:
+        await _fetcher.js_worker_ready_promise
+        return True
+    else:
+        return False
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/request.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/request.py
new file mode 100644
index 00000000..e692e692
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/request.py
@@ -0,0 +1,22 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+
+from ..._base_connection import _TYPE_BODY
+
+
+@dataclass
+class EmscriptenRequest:
+    method: str
+    url: str
+    params: dict[str, str] | None = None
+    body: _TYPE_BODY | None = None
+    headers: dict[str, str] = field(default_factory=dict)
+    timeout: float = 0
+    decode_content: bool = True
+
+    def set_header(self, name: str, value: str) -> None:
+        self.headers[name.capitalize()] = value
+
+    def set_body(self, body: _TYPE_BODY | None) -> None:
+        self.body = body
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/response.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/response.py
new file mode 100644
index 00000000..b32b4023
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/emscripten/response.py
@@ -0,0 +1,285 @@
+from __future__ import annotations
+
+import json as _json
+import logging
+import typing
+from contextlib import contextmanager
+from dataclasses import dataclass
+from http.client import HTTPException as HTTPException
+from io import BytesIO, IOBase
+
+from ...exceptions import InvalidHeader, TimeoutError
+from ...response import BaseHTTPResponse
+from ...util.retry import Retry
+from .request import EmscriptenRequest
+
+if typing.TYPE_CHECKING:
+    from ..._base_connection import BaseHTTPConnection, BaseHTTPSConnection
+
+log = logging.getLogger(__name__)
+
+
+@dataclass
+class EmscriptenResponse:
+    status_code: int
+    headers: dict[str, str]
+    body: IOBase | bytes
+    request: EmscriptenRequest
+
+
+class EmscriptenHttpResponseWrapper(BaseHTTPResponse):
+    def __init__(
+        self,
+        internal_response: EmscriptenResponse,
+        url: str | None = None,
+        connection: BaseHTTPConnection | BaseHTTPSConnection | None = None,
+    ):
+        self._pool = None  # set by pool class
+        self._body = None
+        self._response = internal_response
+        self._url = url
+        self._connection = connection
+        self._closed = False
+        super().__init__(
+            headers=internal_response.headers,
+            status=internal_response.status_code,
+            request_url=url,
+            version=0,
+            version_string="HTTP/?",
+            reason="",
+            decode_content=True,
+        )
+        self.length_remaining = self._init_length(self._response.request.method)
+        self.length_is_certain = False
+
+    @property
+    def url(self) -> str | None:
+        return self._url
+
+    @url.setter
+    def url(self, url: str | None) -> None:
+        self._url = url
+
+    @property
+    def connection(self) -> BaseHTTPConnection | BaseHTTPSConnection | None:
+        return self._connection
+
+    @property
+    def retries(self) -> Retry | None:
+        return self._retries
+
+    @retries.setter
+    def retries(self, retries: Retry | None) -> None:
+        # Override the request_url if retries has a redirect location.
+        self._retries = retries
+
+    def stream(
+        self, amt: int | None = 2**16, decode_content: bool | None = None
+    ) -> typing.Generator[bytes]:
+        """
+        A generator wrapper for the read() method. A call will block until
+        ``amt`` bytes have been read from the connection or until the
+        connection is closed.
+
+        :param amt:
+            How much of the content to read. The generator will return up to
+            much data per iteration, but may return less. This is particularly
+            likely when using compressed data. However, the empty string will
+            never be returned.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        while True:
+            data = self.read(amt=amt, decode_content=decode_content)
+
+            if data:
+                yield data
+            else:
+                break
+
+    def _init_length(self, request_method: str | None) -> int | None:
+        length: int | None
+        content_length: str | None = self.headers.get("content-length")
+
+        if content_length is not None:
+            try:
+                # RFC 7230 section 3.3.2 specifies multiple content lengths can
+                # be sent in a single Content-Length header
+                # (e.g. Content-Length: 42, 42). This line ensures the values
+                # are all valid ints and that as long as the `set` length is 1,
+                # all values are the same. Otherwise, the header is invalid.
+                lengths = {int(val) for val in content_length.split(",")}
+                if len(lengths) > 1:
+                    raise InvalidHeader(
+                        "Content-Length contained multiple "
+                        "unmatching values (%s)" % content_length
+                    )
+                length = lengths.pop()
+            except ValueError:
+                length = None
+            else:
+                if length < 0:
+                    length = None
+
+        else:  # if content_length is None
+            length = None
+
+        # Check for responses that shouldn't include a body
+        if (
+            self.status in (204, 304)
+            or 100 <= self.status < 200
+            or request_method == "HEAD"
+        ):
+            length = 0
+
+        return length
+
+    def read(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,  # ignored because browser decodes always
+        cache_content: bool = False,
+    ) -> bytes:
+        if (
+            self._closed
+            or self._response is None
+            or (isinstance(self._response.body, IOBase) and self._response.body.closed)
+        ):
+            return b""
+
+        with self._error_catcher():
+            # body has been preloaded as a string by XmlHttpRequest
+            if not isinstance(self._response.body, IOBase):
+                self.length_remaining = len(self._response.body)
+                self.length_is_certain = True
+                # wrap body in IOStream
+                self._response.body = BytesIO(self._response.body)
+            if amt is not None and amt >= 0:
+                # don't cache partial content
+                cache_content = False
+                data = self._response.body.read(amt)
+                if self.length_remaining is not None:
+                    self.length_remaining = max(self.length_remaining - len(data), 0)
+                if (self.length_is_certain and self.length_remaining == 0) or len(
+                    data
+                ) < amt:
+                    # definitely finished reading, close response stream
+                    self._response.body.close()
+                return typing.cast(bytes, data)
+            else:  # read all we can (and cache it)
+                data = self._response.body.read()
+                if cache_content:
+                    self._body = data
+                if self.length_remaining is not None:
+                    self.length_remaining = max(self.length_remaining - len(data), 0)
+                if len(data) == 0 or (
+                    self.length_is_certain and self.length_remaining == 0
+                ):
+                    # definitely finished reading, close response stream
+                    self._response.body.close()
+                return typing.cast(bytes, data)
+
+    def read_chunked(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+    ) -> typing.Generator[bytes]:
+        # chunked is handled by browser
+        while True:
+            bytes = self.read(amt, decode_content)
+            if not bytes:
+                break
+            yield bytes
+
+    def release_conn(self) -> None:
+        if not self._pool or not self._connection:
+            return None
+
+        self._pool._put_conn(self._connection)
+        self._connection = None
+
+    def drain_conn(self) -> None:
+        self.close()
+
+    @property
+    def data(self) -> bytes:
+        if self._body:
+            return self._body
+        else:
+            return self.read(cache_content=True)
+
+    def json(self) -> typing.Any:
+        """
+        Deserializes the body of the HTTP response as a Python object.
+
+        The body of the HTTP response must be encoded using UTF-8, as per
+        `RFC 8529 Section 8.1 <https://www.rfc-editor.org/rfc/rfc8259#section-8.1>`_.
+
+        To use a custom JSON decoder pass the result of :attr:`HTTPResponse.data` to
+        your custom decoder instead.
+
+        If the body of the HTTP response is not decodable to UTF-8, a
+        `UnicodeDecodeError` will be raised. If the body of the HTTP response is not a
+        valid JSON document, a `json.JSONDecodeError` will be raised.
+
+        Read more :ref:`here <json_content>`.
+
+        :returns: The body of the HTTP response as a Python object.
+        """
+        data = self.data.decode("utf-8")
+        return _json.loads(data)
+
+    def close(self) -> None:
+        if not self._closed:
+            if isinstance(self._response.body, IOBase):
+                self._response.body.close()
+            if self._connection:
+                self._connection.close()
+                self._connection = None
+            self._closed = True
+
+    @contextmanager
+    def _error_catcher(self) -> typing.Generator[None]:
+        """
+        Catch Emscripten specific exceptions thrown by fetch.py,
+        instead re-raising urllib3 variants, so that low-level exceptions
+        are not leaked in the high-level api.
+
+        On exit, release the connection back to the pool.
+        """
+        from .fetch import _RequestError, _TimeoutError  # avoid circular import
+
+        clean_exit = False
+
+        try:
+            yield
+            # If no exception is thrown, we should avoid cleaning up
+            # unnecessarily.
+            clean_exit = True
+        except _TimeoutError as e:
+            raise TimeoutError(str(e))
+        except _RequestError as e:
+            raise HTTPException(str(e))
+        finally:
+            # If we didn't terminate cleanly, we need to throw away our
+            # connection.
+            if not clean_exit:
+                # The response may not be closed but we're not going to use it
+                # anymore so close it now
+                if (
+                    isinstance(self._response.body, IOBase)
+                    and not self._response.body.closed
+                ):
+                    self._response.body.close()
+                # release the connection back to the pool
+                self.release_conn()
+            else:
+                # If we have read everything from the response stream,
+                # return the connection back to the pool.
+                if (
+                    isinstance(self._response.body, IOBase)
+                    and self._response.body.closed
+                ):
+                    self.release_conn()
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/pyopenssl.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/pyopenssl.py
new file mode 100644
index 00000000..ed654306
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/pyopenssl.py
@@ -0,0 +1,554 @@
+"""
+Module for using pyOpenSSL as a TLS backend. This module was relevant before
+the standard library ``ssl`` module supported SNI, but now that we've dropped
+support for Python 2.7 all relevant Python versions support SNI so
+**this module is no longer recommended**.
+
+This needs the following packages installed:
+
+* `pyOpenSSL`_ (tested with 16.0.0)
+* `cryptography`_ (minimum 1.3.4, from pyopenssl)
+* `idna`_ (minimum 2.0)
+
+However, pyOpenSSL depends on cryptography, so while we use all three directly here we
+end up having relatively few packages required.
+
+You can install them with the following command:
+
+.. code-block:: bash
+
+    $ python -m pip install pyopenssl cryptography idna
+
+To activate certificate checking, call
+:func:`~urllib3.contrib.pyopenssl.inject_into_urllib3` from your Python code
+before you begin making HTTP requests. This can be done in a ``sitecustomize``
+module, or at any other time before your application begins using ``urllib3``,
+like this:
+
+.. code-block:: python
+
+    try:
+        import urllib3.contrib.pyopenssl
+        urllib3.contrib.pyopenssl.inject_into_urllib3()
+    except ImportError:
+        pass
+
+.. _pyopenssl: https://www.pyopenssl.org
+.. _cryptography: https://cryptography.io
+.. _idna: https://github.com/kjd/idna
+"""
+
+from __future__ import annotations
+
+import OpenSSL.SSL  # type: ignore[import-untyped]
+from cryptography import x509
+
+try:
+    from cryptography.x509 import UnsupportedExtension  # type: ignore[attr-defined]
+except ImportError:
+    # UnsupportedExtension is gone in cryptography >= 2.1.0
+    class UnsupportedExtension(Exception):  # type: ignore[no-redef]
+        pass
+
+
+import logging
+import ssl
+import typing
+from io import BytesIO
+from socket import socket as socket_cls
+from socket import timeout
+
+from .. import util
+
+if typing.TYPE_CHECKING:
+    from OpenSSL.crypto import X509  # type: ignore[import-untyped]
+
+
+__all__ = ["inject_into_urllib3", "extract_from_urllib3"]
+
+# Map from urllib3 to PyOpenSSL compatible parameter-values.
+_openssl_versions: dict[int, int] = {
+    util.ssl_.PROTOCOL_TLS: OpenSSL.SSL.SSLv23_METHOD,  # type: ignore[attr-defined]
+    util.ssl_.PROTOCOL_TLS_CLIENT: OpenSSL.SSL.SSLv23_METHOD,  # type: ignore[attr-defined]
+    ssl.PROTOCOL_TLSv1: OpenSSL.SSL.TLSv1_METHOD,
+}
+
+if hasattr(ssl, "PROTOCOL_TLSv1_1") and hasattr(OpenSSL.SSL, "TLSv1_1_METHOD"):
+    _openssl_versions[ssl.PROTOCOL_TLSv1_1] = OpenSSL.SSL.TLSv1_1_METHOD
+
+if hasattr(ssl, "PROTOCOL_TLSv1_2") and hasattr(OpenSSL.SSL, "TLSv1_2_METHOD"):
+    _openssl_versions[ssl.PROTOCOL_TLSv1_2] = OpenSSL.SSL.TLSv1_2_METHOD
+
+
+_stdlib_to_openssl_verify = {
+    ssl.CERT_NONE: OpenSSL.SSL.VERIFY_NONE,
+    ssl.CERT_OPTIONAL: OpenSSL.SSL.VERIFY_PEER,
+    ssl.CERT_REQUIRED: OpenSSL.SSL.VERIFY_PEER
+    + OpenSSL.SSL.VERIFY_FAIL_IF_NO_PEER_CERT,
+}
+_openssl_to_stdlib_verify = {v: k for k, v in _stdlib_to_openssl_verify.items()}
+
+# The SSLvX values are the most likely to be missing in the future
+# but we check them all just to be sure.
+_OP_NO_SSLv2_OR_SSLv3: int = getattr(OpenSSL.SSL, "OP_NO_SSLv2", 0) | getattr(
+    OpenSSL.SSL, "OP_NO_SSLv3", 0
+)
+_OP_NO_TLSv1: int = getattr(OpenSSL.SSL, "OP_NO_TLSv1", 0)
+_OP_NO_TLSv1_1: int = getattr(OpenSSL.SSL, "OP_NO_TLSv1_1", 0)
+_OP_NO_TLSv1_2: int = getattr(OpenSSL.SSL, "OP_NO_TLSv1_2", 0)
+_OP_NO_TLSv1_3: int = getattr(OpenSSL.SSL, "OP_NO_TLSv1_3", 0)
+
+_openssl_to_ssl_minimum_version: dict[int, int] = {
+    ssl.TLSVersion.MINIMUM_SUPPORTED: _OP_NO_SSLv2_OR_SSLv3,
+    ssl.TLSVersion.TLSv1: _OP_NO_SSLv2_OR_SSLv3,
+    ssl.TLSVersion.TLSv1_1: _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1,
+    ssl.TLSVersion.TLSv1_2: _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1 | _OP_NO_TLSv1_1,
+    ssl.TLSVersion.TLSv1_3: (
+        _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1 | _OP_NO_TLSv1_1 | _OP_NO_TLSv1_2
+    ),
+    ssl.TLSVersion.MAXIMUM_SUPPORTED: (
+        _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1 | _OP_NO_TLSv1_1 | _OP_NO_TLSv1_2
+    ),
+}
+_openssl_to_ssl_maximum_version: dict[int, int] = {
+    ssl.TLSVersion.MINIMUM_SUPPORTED: (
+        _OP_NO_SSLv2_OR_SSLv3
+        | _OP_NO_TLSv1
+        | _OP_NO_TLSv1_1
+        | _OP_NO_TLSv1_2
+        | _OP_NO_TLSv1_3
+    ),
+    ssl.TLSVersion.TLSv1: (
+        _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1_1 | _OP_NO_TLSv1_2 | _OP_NO_TLSv1_3
+    ),
+    ssl.TLSVersion.TLSv1_1: _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1_2 | _OP_NO_TLSv1_3,
+    ssl.TLSVersion.TLSv1_2: _OP_NO_SSLv2_OR_SSLv3 | _OP_NO_TLSv1_3,
+    ssl.TLSVersion.TLSv1_3: _OP_NO_SSLv2_OR_SSLv3,
+    ssl.TLSVersion.MAXIMUM_SUPPORTED: _OP_NO_SSLv2_OR_SSLv3,
+}
+
+# OpenSSL will only write 16K at a time
+SSL_WRITE_BLOCKSIZE = 16384
+
+orig_util_SSLContext = util.ssl_.SSLContext
+
+
+log = logging.getLogger(__name__)
+
+
+def inject_into_urllib3() -> None:
+    "Monkey-patch urllib3 with PyOpenSSL-backed SSL-support."
+
+    _validate_dependencies_met()
+
+    util.SSLContext = PyOpenSSLContext  # type: ignore[assignment]
+    util.ssl_.SSLContext = PyOpenSSLContext  # type: ignore[assignment]
+    util.IS_PYOPENSSL = True
+    util.ssl_.IS_PYOPENSSL = True
+
+
+def extract_from_urllib3() -> None:
+    "Undo monkey-patching by :func:`inject_into_urllib3`."
+
+    util.SSLContext = orig_util_SSLContext
+    util.ssl_.SSLContext = orig_util_SSLContext
+    util.IS_PYOPENSSL = False
+    util.ssl_.IS_PYOPENSSL = False
+
+
+def _validate_dependencies_met() -> None:
+    """
+    Verifies that PyOpenSSL's package-level dependencies have been met.
+    Throws `ImportError` if they are not met.
+    """
+    # Method added in `cryptography==1.1`; not available in older versions
+    from cryptography.x509.extensions import Extensions
+
+    if getattr(Extensions, "get_extension_for_class", None) is None:
+        raise ImportError(
+            "'cryptography' module missing required functionality.  "
+            "Try upgrading to v1.3.4 or newer."
+        )
+
+    # pyOpenSSL 0.14 and above use cryptography for OpenSSL bindings. The _x509
+    # attribute is only present on those versions.
+    from OpenSSL.crypto import X509
+
+    x509 = X509()
+    if getattr(x509, "_x509", None) is None:
+        raise ImportError(
+            "'pyOpenSSL' module missing required functionality. "
+            "Try upgrading to v0.14 or newer."
+        )
+
+
+def _dnsname_to_stdlib(name: str) -> str | None:
+    """
+    Converts a dNSName SubjectAlternativeName field to the form used by the
+    standard library on the given Python version.
+
+    Cryptography produces a dNSName as a unicode string that was idna-decoded
+    from ASCII bytes. We need to idna-encode that string to get it back, and
+    then on Python 3 we also need to convert to unicode via UTF-8 (the stdlib
+    uses PyUnicode_FromStringAndSize on it, which decodes via UTF-8).
+
+    If the name cannot be idna-encoded then we return None signalling that
+    the name given should be skipped.
+    """
+
+    def idna_encode(name: str) -> bytes | None:
+        """
+        Borrowed wholesale from the Python Cryptography Project. It turns out
+        that we can't just safely call `idna.encode`: it can explode for
+        wildcard names. This avoids that problem.
+        """
+        import idna
+
+        try:
+            for prefix in ["*.", "."]:
+                if name.startswith(prefix):
+                    name = name[len(prefix) :]
+                    return prefix.encode("ascii") + idna.encode(name)
+            return idna.encode(name)
+        except idna.core.IDNAError:
+            return None
+
+    # Don't send IPv6 addresses through the IDNA encoder.
+    if ":" in name:
+        return name
+
+    encoded_name = idna_encode(name)
+    if encoded_name is None:
+        return None
+    return encoded_name.decode("utf-8")
+
+
+def get_subj_alt_name(peer_cert: X509) -> list[tuple[str, str]]:
+    """
+    Given an PyOpenSSL certificate, provides all the subject alternative names.
+    """
+    cert = peer_cert.to_cryptography()
+
+    # We want to find the SAN extension. Ask Cryptography to locate it (it's
+    # faster than looping in Python)
+    try:
+        ext = cert.extensions.get_extension_for_class(x509.SubjectAlternativeName).value
+    except x509.ExtensionNotFound:
+        # No such extension, return the empty list.
+        return []
+    except (
+        x509.DuplicateExtension,
+        UnsupportedExtension,
+        x509.UnsupportedGeneralNameType,
+        UnicodeError,
+    ) as e:
+        # A problem has been found with the quality of the certificate. Assume
+        # no SAN field is present.
+        log.warning(
+            "A problem was encountered with the certificate that prevented "
+            "urllib3 from finding the SubjectAlternativeName field. This can "
+            "affect certificate validation. The error was %s",
+            e,
+        )
+        return []
+
+    # We want to return dNSName and iPAddress fields. We need to cast the IPs
+    # back to strings because the match_hostname function wants them as
+    # strings.
+    # Sadly the DNS names need to be idna encoded and then, on Python 3, UTF-8
+    # decoded. This is pretty frustrating, but that's what the standard library
+    # does with certificates, and so we need to attempt to do the same.
+    # We also want to skip over names which cannot be idna encoded.
+    names = [
+        ("DNS", name)
+        for name in map(_dnsname_to_stdlib, ext.get_values_for_type(x509.DNSName))
+        if name is not None
+    ]
+    names.extend(
+        ("IP Address", str(name)) for name in ext.get_values_for_type(x509.IPAddress)
+    )
+
+    return names
+
+
+class WrappedSocket:
+    """API-compatibility wrapper for Python OpenSSL's Connection-class."""
+
+    def __init__(
+        self,
+        connection: OpenSSL.SSL.Connection,
+        socket: socket_cls,
+        suppress_ragged_eofs: bool = True,
+    ) -> None:
+        self.connection = connection
+        self.socket = socket
+        self.suppress_ragged_eofs = suppress_ragged_eofs
+        self._io_refs = 0
+        self._closed = False
+
+    def fileno(self) -> int:
+        return self.socket.fileno()
+
+    # Copy-pasted from Python 3.5 source code
+    def _decref_socketios(self) -> None:
+        if self._io_refs > 0:
+            self._io_refs -= 1
+        if self._closed:
+            self.close()
+
+    def recv(self, *args: typing.Any, **kwargs: typing.Any) -> bytes:
+        try:
+            data = self.connection.recv(*args, **kwargs)
+        except OpenSSL.SSL.SysCallError as e:
+            if self.suppress_ragged_eofs and e.args == (-1, "Unexpected EOF"):
+                return b""
+            else:
+                raise OSError(e.args[0], str(e)) from e
+        except OpenSSL.SSL.ZeroReturnError:
+            if self.connection.get_shutdown() == OpenSSL.SSL.RECEIVED_SHUTDOWN:
+                return b""
+            else:
+                raise
+        except OpenSSL.SSL.WantReadError as e:
+            if not util.wait_for_read(self.socket, self.socket.gettimeout()):
+                raise timeout("The read operation timed out") from e
+            else:
+                return self.recv(*args, **kwargs)
+
+        # TLS 1.3 post-handshake authentication
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError(f"read error: {e!r}") from e
+        else:
+            return data  # type: ignore[no-any-return]
+
+    def recv_into(self, *args: typing.Any, **kwargs: typing.Any) -> int:
+        try:
+            return self.connection.recv_into(*args, **kwargs)  # type: ignore[no-any-return]
+        except OpenSSL.SSL.SysCallError as e:
+            if self.suppress_ragged_eofs and e.args == (-1, "Unexpected EOF"):
+                return 0
+            else:
+                raise OSError(e.args[0], str(e)) from e
+        except OpenSSL.SSL.ZeroReturnError:
+            if self.connection.get_shutdown() == OpenSSL.SSL.RECEIVED_SHUTDOWN:
+                return 0
+            else:
+                raise
+        except OpenSSL.SSL.WantReadError as e:
+            if not util.wait_for_read(self.socket, self.socket.gettimeout()):
+                raise timeout("The read operation timed out") from e
+            else:
+                return self.recv_into(*args, **kwargs)
+
+        # TLS 1.3 post-handshake authentication
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError(f"read error: {e!r}") from e
+
+    def settimeout(self, timeout: float) -> None:
+        return self.socket.settimeout(timeout)
+
+    def _send_until_done(self, data: bytes) -> int:
+        while True:
+            try:
+                return self.connection.send(data)  # type: ignore[no-any-return]
+            except OpenSSL.SSL.WantWriteError as e:
+                if not util.wait_for_write(self.socket, self.socket.gettimeout()):
+                    raise timeout() from e
+                continue
+            except OpenSSL.SSL.SysCallError as e:
+                raise OSError(e.args[0], str(e)) from e
+
+    def sendall(self, data: bytes) -> None:
+        total_sent = 0
+        while total_sent < len(data):
+            sent = self._send_until_done(
+                data[total_sent : total_sent + SSL_WRITE_BLOCKSIZE]
+            )
+            total_sent += sent
+
+    def shutdown(self, how: int) -> None:
+        try:
+            self.connection.shutdown()
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError(f"shutdown error: {e!r}") from e
+
+    def close(self) -> None:
+        self._closed = True
+        if self._io_refs <= 0:
+            self._real_close()
+
+    def _real_close(self) -> None:
+        try:
+            return self.connection.close()  # type: ignore[no-any-return]
+        except OpenSSL.SSL.Error:
+            return
+
+    def getpeercert(
+        self, binary_form: bool = False
+    ) -> dict[str, list[typing.Any]] | None:
+        x509 = self.connection.get_peer_certificate()
+
+        if not x509:
+            return x509  # type: ignore[no-any-return]
+
+        if binary_form:
+            return OpenSSL.crypto.dump_certificate(OpenSSL.crypto.FILETYPE_ASN1, x509)  # type: ignore[no-any-return]
+
+        return {
+            "subject": ((("commonName", x509.get_subject().CN),),),  # type: ignore[dict-item]
+            "subjectAltName": get_subj_alt_name(x509),
+        }
+
+    def version(self) -> str:
+        return self.connection.get_protocol_version_name()  # type: ignore[no-any-return]
+
+    def selected_alpn_protocol(self) -> str | None:
+        alpn_proto = self.connection.get_alpn_proto_negotiated()
+        return alpn_proto.decode() if alpn_proto else None
+
+
+WrappedSocket.makefile = socket_cls.makefile  # type: ignore[attr-defined]
+
+
+class PyOpenSSLContext:
+    """
+    I am a wrapper class for the PyOpenSSL ``Context`` object. I am responsible
+    for translating the interface of the standard library ``SSLContext`` object
+    to calls into PyOpenSSL.
+    """
+
+    def __init__(self, protocol: int) -> None:
+        self.protocol = _openssl_versions[protocol]
+        self._ctx = OpenSSL.SSL.Context(self.protocol)
+        self._options = 0
+        self.check_hostname = False
+        self._minimum_version: int = ssl.TLSVersion.MINIMUM_SUPPORTED
+        self._maximum_version: int = ssl.TLSVersion.MAXIMUM_SUPPORTED
+
+    @property
+    def options(self) -> int:
+        return self._options
+
+    @options.setter
+    def options(self, value: int) -> None:
+        self._options = value
+        self._set_ctx_options()
+
+    @property
+    def verify_mode(self) -> int:
+        return _openssl_to_stdlib_verify[self._ctx.get_verify_mode()]
+
+    @verify_mode.setter
+    def verify_mode(self, value: ssl.VerifyMode) -> None:
+        self._ctx.set_verify(_stdlib_to_openssl_verify[value], _verify_callback)
+
+    def set_default_verify_paths(self) -> None:
+        self._ctx.set_default_verify_paths()
+
+    def set_ciphers(self, ciphers: bytes | str) -> None:
+        if isinstance(ciphers, str):
+            ciphers = ciphers.encode("utf-8")
+        self._ctx.set_cipher_list(ciphers)
+
+    def load_verify_locations(
+        self,
+        cafile: str | None = None,
+        capath: str | None = None,
+        cadata: bytes | None = None,
+    ) -> None:
+        if cafile is not None:
+            cafile = cafile.encode("utf-8")  # type: ignore[assignment]
+        if capath is not None:
+            capath = capath.encode("utf-8")  # type: ignore[assignment]
+        try:
+            self._ctx.load_verify_locations(cafile, capath)
+            if cadata is not None:
+                self._ctx.load_verify_locations(BytesIO(cadata))
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError(f"unable to load trusted certificates: {e!r}") from e
+
+    def load_cert_chain(
+        self,
+        certfile: str,
+        keyfile: str | None = None,
+        password: str | None = None,
+    ) -> None:
+        try:
+            self._ctx.use_certificate_chain_file(certfile)
+            if password is not None:
+                if not isinstance(password, bytes):
+                    password = password.encode("utf-8")  # type: ignore[assignment]
+                self._ctx.set_passwd_cb(lambda *_: password)
+            self._ctx.use_privatekey_file(keyfile or certfile)
+        except OpenSSL.SSL.Error as e:
+            raise ssl.SSLError(f"Unable to load certificate chain: {e!r}") from e
+
+    def set_alpn_protocols(self, protocols: list[bytes | str]) -> None:
+        protocols = [util.util.to_bytes(p, "ascii") for p in protocols]
+        return self._ctx.set_alpn_protos(protocols)  # type: ignore[no-any-return]
+
+    def wrap_socket(
+        self,
+        sock: socket_cls,
+        server_side: bool = False,
+        do_handshake_on_connect: bool = True,
+        suppress_ragged_eofs: bool = True,
+        server_hostname: bytes | str | None = None,
+    ) -> WrappedSocket:
+        cnx = OpenSSL.SSL.Connection(self._ctx, sock)
+
+        # If server_hostname is an IP, don't use it for SNI, per RFC6066 Section 3
+        if server_hostname and not util.ssl_.is_ipaddress(server_hostname):
+            if isinstance(server_hostname, str):
+                server_hostname = server_hostname.encode("utf-8")
+            cnx.set_tlsext_host_name(server_hostname)
+
+        cnx.set_connect_state()
+
+        while True:
+            try:
+                cnx.do_handshake()
+            except OpenSSL.SSL.WantReadError as e:
+                if not util.wait_for_read(sock, sock.gettimeout()):
+                    raise timeout("select timed out") from e
+                continue
+            except OpenSSL.SSL.Error as e:
+                raise ssl.SSLError(f"bad handshake: {e!r}") from e
+            break
+
+        return WrappedSocket(cnx, sock)
+
+    def _set_ctx_options(self) -> None:
+        self._ctx.set_options(
+            self._options
+            | _openssl_to_ssl_minimum_version[self._minimum_version]
+            | _openssl_to_ssl_maximum_version[self._maximum_version]
+        )
+
+    @property
+    def minimum_version(self) -> int:
+        return self._minimum_version
+
+    @minimum_version.setter
+    def minimum_version(self, minimum_version: int) -> None:
+        self._minimum_version = minimum_version
+        self._set_ctx_options()
+
+    @property
+    def maximum_version(self) -> int:
+        return self._maximum_version
+
+    @maximum_version.setter
+    def maximum_version(self, maximum_version: int) -> None:
+        self._maximum_version = maximum_version
+        self._set_ctx_options()
+
+
+def _verify_callback(
+    cnx: OpenSSL.SSL.Connection,
+    x509: X509,
+    err_no: int,
+    err_depth: int,
+    return_code: int,
+) -> bool:
+    return err_no == 0
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/contrib/socks.py b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/socks.py
new file mode 100644
index 00000000..c62b5e03
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/contrib/socks.py
@@ -0,0 +1,228 @@
+"""
+This module contains provisional support for SOCKS proxies from within
+urllib3. This module supports SOCKS4, SOCKS4A (an extension of SOCKS4), and
+SOCKS5. To enable its functionality, either install PySocks or install this
+module with the ``socks`` extra.
+
+The SOCKS implementation supports the full range of urllib3 features. It also
+supports the following SOCKS features:
+
+- SOCKS4A (``proxy_url='socks4a://...``)
+- SOCKS4 (``proxy_url='socks4://...``)
+- SOCKS5 with remote DNS (``proxy_url='socks5h://...``)
+- SOCKS5 with local DNS (``proxy_url='socks5://...``)
+- Usernames and passwords for the SOCKS proxy
+
+.. note::
+   It is recommended to use ``socks5h://`` or ``socks4a://`` schemes in
+   your ``proxy_url`` to ensure that DNS resolution is done from the remote
+   server instead of client-side when connecting to a domain name.
+
+SOCKS4 supports IPv4 and domain names with the SOCKS4A extension. SOCKS5
+supports IPv4, IPv6, and domain names.
+
+When connecting to a SOCKS4 proxy the ``username`` portion of the ``proxy_url``
+will be sent as the ``userid`` section of the SOCKS request:
+
+.. code-block:: python
+
+    proxy_url="socks4a://<userid>@proxy-host"
+
+When connecting to a SOCKS5 proxy the ``username`` and ``password`` portion
+of the ``proxy_url`` will be sent as the username/password to authenticate
+with the proxy:
+
+.. code-block:: python
+
+    proxy_url="socks5h://<username>:<password>@proxy-host"
+
+"""
+
+from __future__ import annotations
+
+try:
+    import socks  # type: ignore[import-not-found]
+except ImportError:
+    import warnings
+
+    from ..exceptions import DependencyWarning
+
+    warnings.warn(
+        (
+            "SOCKS support in urllib3 requires the installation of optional "
+            "dependencies: specifically, PySocks.  For more information, see "
+            "https://urllib3.readthedocs.io/en/latest/advanced-usage.html#socks-proxies"
+        ),
+        DependencyWarning,
+    )
+    raise
+
+import typing
+from socket import timeout as SocketTimeout
+
+from ..connection import HTTPConnection, HTTPSConnection
+from ..connectionpool import HTTPConnectionPool, HTTPSConnectionPool
+from ..exceptions import ConnectTimeoutError, NewConnectionError
+from ..poolmanager import PoolManager
+from ..util.url import parse_url
+
+try:
+    import ssl
+except ImportError:
+    ssl = None  # type: ignore[assignment]
+
+
+class _TYPE_SOCKS_OPTIONS(typing.TypedDict):
+    socks_version: int
+    proxy_host: str | None
+    proxy_port: str | None
+    username: str | None
+    password: str | None
+    rdns: bool
+
+
+class SOCKSConnection(HTTPConnection):
+    """
+    A plain-text HTTP connection that connects via a SOCKS proxy.
+    """
+
+    def __init__(
+        self,
+        _socks_options: _TYPE_SOCKS_OPTIONS,
+        *args: typing.Any,
+        **kwargs: typing.Any,
+    ) -> None:
+        self._socks_options = _socks_options
+        super().__init__(*args, **kwargs)
+
+    def _new_conn(self) -> socks.socksocket:
+        """
+        Establish a new connection via the SOCKS proxy.
+        """
+        extra_kw: dict[str, typing.Any] = {}
+        if self.source_address:
+            extra_kw["source_address"] = self.source_address
+
+        if self.socket_options:
+            extra_kw["socket_options"] = self.socket_options
+
+        try:
+            conn = socks.create_connection(
+                (self.host, self.port),
+                proxy_type=self._socks_options["socks_version"],
+                proxy_addr=self._socks_options["proxy_host"],
+                proxy_port=self._socks_options["proxy_port"],
+                proxy_username=self._socks_options["username"],
+                proxy_password=self._socks_options["password"],
+                proxy_rdns=self._socks_options["rdns"],
+                timeout=self.timeout,
+                **extra_kw,
+            )
+
+        except SocketTimeout as e:
+            raise ConnectTimeoutError(
+                self,
+                f"Connection to {self.host} timed out. (connect timeout={self.timeout})",
+            ) from e
+
+        except socks.ProxyError as e:
+            # This is fragile as hell, but it seems to be the only way to raise
+            # useful errors here.
+            if e.socket_err:
+                error = e.socket_err
+                if isinstance(error, SocketTimeout):
+                    raise ConnectTimeoutError(
+                        self,
+                        f"Connection to {self.host} timed out. (connect timeout={self.timeout})",
+                    ) from e
+                else:
+                    # Adding `from e` messes with coverage somehow, so it's omitted.
+                    # See #2386.
+                    raise NewConnectionError(
+                        self, f"Failed to establish a new connection: {error}"
+                    )
+            else:
+                raise NewConnectionError(
+                    self, f"Failed to establish a new connection: {e}"
+                ) from e
+
+        except OSError as e:  # Defensive: PySocks should catch all these.
+            raise NewConnectionError(
+                self, f"Failed to establish a new connection: {e}"
+            ) from e
+
+        return conn
+
+
+# We don't need to duplicate the Verified/Unverified distinction from
+# urllib3/connection.py here because the HTTPSConnection will already have been
+# correctly set to either the Verified or Unverified form by that module. This
+# means the SOCKSHTTPSConnection will automatically be the correct type.
+class SOCKSHTTPSConnection(SOCKSConnection, HTTPSConnection):
+    pass
+
+
+class SOCKSHTTPConnectionPool(HTTPConnectionPool):
+    ConnectionCls = SOCKSConnection
+
+
+class SOCKSHTTPSConnectionPool(HTTPSConnectionPool):
+    ConnectionCls = SOCKSHTTPSConnection
+
+
+class SOCKSProxyManager(PoolManager):
+    """
+    A version of the urllib3 ProxyManager that routes connections via the
+    defined SOCKS proxy.
+    """
+
+    pool_classes_by_scheme = {
+        "http": SOCKSHTTPConnectionPool,
+        "https": SOCKSHTTPSConnectionPool,
+    }
+
+    def __init__(
+        self,
+        proxy_url: str,
+        username: str | None = None,
+        password: str | None = None,
+        num_pools: int = 10,
+        headers: typing.Mapping[str, str] | None = None,
+        **connection_pool_kw: typing.Any,
+    ):
+        parsed = parse_url(proxy_url)
+
+        if username is None and password is None and parsed.auth is not None:
+            split = parsed.auth.split(":")
+            if len(split) == 2:
+                username, password = split
+        if parsed.scheme == "socks5":
+            socks_version = socks.PROXY_TYPE_SOCKS5
+            rdns = False
+        elif parsed.scheme == "socks5h":
+            socks_version = socks.PROXY_TYPE_SOCKS5
+            rdns = True
+        elif parsed.scheme == "socks4":
+            socks_version = socks.PROXY_TYPE_SOCKS4
+            rdns = False
+        elif parsed.scheme == "socks4a":
+            socks_version = socks.PROXY_TYPE_SOCKS4
+            rdns = True
+        else:
+            raise ValueError(f"Unable to determine SOCKS version from {proxy_url}")
+
+        self.proxy_url = proxy_url
+
+        socks_options = {
+            "socks_version": socks_version,
+            "proxy_host": parsed.host,
+            "proxy_port": parsed.port,
+            "username": username,
+            "password": password,
+            "rdns": rdns,
+        }
+        connection_pool_kw["_socks_options"] = socks_options
+
+        super().__init__(num_pools, headers, **connection_pool_kw)
+
+        self.pool_classes_by_scheme = SOCKSProxyManager.pool_classes_by_scheme
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/exceptions.py b/backend/venv/lib/python3.12/site-packages/urllib3/exceptions.py
new file mode 100644
index 00000000..03945782
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/exceptions.py
@@ -0,0 +1,327 @@
+from __future__ import annotations
+
+import socket
+import typing
+import warnings
+from email.errors import MessageDefect
+from http.client import IncompleteRead as httplib_IncompleteRead
+
+if typing.TYPE_CHECKING:
+    from .connection import HTTPConnection
+    from .connectionpool import ConnectionPool
+    from .response import HTTPResponse
+    from .util.retry import Retry
+
+# Base Exceptions
+
+
+class HTTPError(Exception):
+    """Base exception used by this module."""
+
+
+class HTTPWarning(Warning):
+    """Base warning used by this module."""
+
+
+_TYPE_REDUCE_RESULT = tuple[typing.Callable[..., object], tuple[object, ...]]
+
+
+class PoolError(HTTPError):
+    """Base exception for errors caused within a pool."""
+
+    def __init__(self, pool: ConnectionPool, message: str) -> None:
+        self.pool = pool
+        super().__init__(f"{pool}: {message}")
+
+    def __reduce__(self) -> _TYPE_REDUCE_RESULT:
+        # For pickling purposes.
+        return self.__class__, (None, None)
+
+
+class RequestError(PoolError):
+    """Base exception for PoolErrors that have associated URLs."""
+
+    def __init__(self, pool: ConnectionPool, url: str, message: str) -> None:
+        self.url = url
+        super().__init__(pool, message)
+
+    def __reduce__(self) -> _TYPE_REDUCE_RESULT:
+        # For pickling purposes.
+        return self.__class__, (None, self.url, None)
+
+
+class SSLError(HTTPError):
+    """Raised when SSL certificate fails in an HTTPS connection."""
+
+
+class ProxyError(HTTPError):
+    """Raised when the connection to a proxy fails."""
+
+    # The original error is also available as __cause__.
+    original_error: Exception
+
+    def __init__(self, message: str, error: Exception) -> None:
+        super().__init__(message, error)
+        self.original_error = error
+
+
+class DecodeError(HTTPError):
+    """Raised when automatic decoding based on Content-Type fails."""
+
+
+class ProtocolError(HTTPError):
+    """Raised when something unexpected happens mid-request/response."""
+
+
+#: Renamed to ProtocolError but aliased for backwards compatibility.
+ConnectionError = ProtocolError
+
+
+# Leaf Exceptions
+
+
+class MaxRetryError(RequestError):
+    """Raised when the maximum number of retries is exceeded.
+
+    :param pool: The connection pool
+    :type pool: :class:`~urllib3.connectionpool.HTTPConnectionPool`
+    :param str url: The requested Url
+    :param reason: The underlying error
+    :type reason: :class:`Exception`
+
+    """
+
+    def __init__(
+        self, pool: ConnectionPool, url: str, reason: Exception | None = None
+    ) -> None:
+        self.reason = reason
+
+        message = f"Max retries exceeded with url: {url} (Caused by {reason!r})"
+
+        super().__init__(pool, url, message)
+
+
+class HostChangedError(RequestError):
+    """Raised when an existing pool gets a request for a foreign host."""
+
+    def __init__(
+        self, pool: ConnectionPool, url: str, retries: Retry | int = 3
+    ) -> None:
+        message = f"Tried to open a foreign host with url: {url}"
+        super().__init__(pool, url, message)
+        self.retries = retries
+
+
+class TimeoutStateError(HTTPError):
+    """Raised when passing an invalid state to a timeout"""
+
+
+class TimeoutError(HTTPError):
+    """Raised when a socket timeout error occurs.
+
+    Catching this error will catch both :exc:`ReadTimeoutErrors
+    <ReadTimeoutError>` and :exc:`ConnectTimeoutErrors <ConnectTimeoutError>`.
+    """
+
+
+class ReadTimeoutError(TimeoutError, RequestError):
+    """Raised when a socket timeout occurs while receiving data from a server"""
+
+
+# This timeout error does not have a URL attached and needs to inherit from the
+# base HTTPError
+class ConnectTimeoutError(TimeoutError):
+    """Raised when a socket timeout occurs while connecting to a server"""
+
+
+class NewConnectionError(ConnectTimeoutError, HTTPError):
+    """Raised when we fail to establish a new connection. Usually ECONNREFUSED."""
+
+    def __init__(self, conn: HTTPConnection, message: str) -> None:
+        self.conn = conn
+        super().__init__(f"{conn}: {message}")
+
+    def __reduce__(self) -> _TYPE_REDUCE_RESULT:
+        # For pickling purposes.
+        return self.__class__, (None, None)
+
+    @property
+    def pool(self) -> HTTPConnection:
+        warnings.warn(
+            "The 'pool' property is deprecated and will be removed "
+            "in urllib3 v2.1.0. Use 'conn' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        return self.conn
+
+
+class NameResolutionError(NewConnectionError):
+    """Raised when host name resolution fails."""
+
+    def __init__(self, host: str, conn: HTTPConnection, reason: socket.gaierror):
+        message = f"Failed to resolve '{host}' ({reason})"
+        super().__init__(conn, message)
+
+    def __reduce__(self) -> _TYPE_REDUCE_RESULT:
+        # For pickling purposes.
+        return self.__class__, (None, None, None)
+
+
+class EmptyPoolError(PoolError):
+    """Raised when a pool runs out of connections and no more are allowed."""
+
+
+class FullPoolError(PoolError):
+    """Raised when we try to add a connection to a full pool in blocking mode."""
+
+
+class ClosedPoolError(PoolError):
+    """Raised when a request enters a pool after the pool has been closed."""
+
+
+class LocationValueError(ValueError, HTTPError):
+    """Raised when there is something wrong with a given URL input."""
+
+
+class LocationParseError(LocationValueError):
+    """Raised when get_host or similar fails to parse the URL input."""
+
+    def __init__(self, location: str) -> None:
+        message = f"Failed to parse: {location}"
+        super().__init__(message)
+
+        self.location = location
+
+
+class URLSchemeUnknown(LocationValueError):
+    """Raised when a URL input has an unsupported scheme."""
+
+    def __init__(self, scheme: str):
+        message = f"Not supported URL scheme {scheme}"
+        super().__init__(message)
+
+        self.scheme = scheme
+
+
+class ResponseError(HTTPError):
+    """Used as a container for an error reason supplied in a MaxRetryError."""
+
+    GENERIC_ERROR = "too many error responses"
+    SPECIFIC_ERROR = "too many {status_code} error responses"
+
+
+class SecurityWarning(HTTPWarning):
+    """Warned when performing security reducing actions"""
+
+
+class InsecureRequestWarning(SecurityWarning):
+    """Warned when making an unverified HTTPS request."""
+
+
+class NotOpenSSLWarning(SecurityWarning):
+    """Warned when using unsupported SSL library"""
+
+
+class SystemTimeWarning(SecurityWarning):
+    """Warned when system time is suspected to be wrong"""
+
+
+class InsecurePlatformWarning(SecurityWarning):
+    """Warned when certain TLS/SSL configuration is not available on a platform."""
+
+
+class DependencyWarning(HTTPWarning):
+    """
+    Warned when an attempt is made to import a module with missing optional
+    dependencies.
+    """
+
+
+class ResponseNotChunked(ProtocolError, ValueError):
+    """Response needs to be chunked in order to read it as chunks."""
+
+
+class BodyNotHttplibCompatible(HTTPError):
+    """
+    Body should be :class:`http.client.HTTPResponse` like
+    (have an fp attribute which returns raw chunks) for read_chunked().
+    """
+
+
+class IncompleteRead(HTTPError, httplib_IncompleteRead):
+    """
+    Response length doesn't match expected Content-Length
+
+    Subclass of :class:`http.client.IncompleteRead` to allow int value
+    for ``partial`` to avoid creating large objects on streamed reads.
+    """
+
+    partial: int  # type: ignore[assignment]
+    expected: int
+
+    def __init__(self, partial: int, expected: int) -> None:
+        self.partial = partial
+        self.expected = expected
+
+    def __repr__(self) -> str:
+        return "IncompleteRead(%i bytes read, %i more expected)" % (
+            self.partial,
+            self.expected,
+        )
+
+
+class InvalidChunkLength(HTTPError, httplib_IncompleteRead):
+    """Invalid chunk length in a chunked response."""
+
+    def __init__(self, response: HTTPResponse, length: bytes) -> None:
+        self.partial: int = response.tell()  # type: ignore[assignment]
+        self.expected: int | None = response.length_remaining
+        self.response = response
+        self.length = length
+
+    def __repr__(self) -> str:
+        return "InvalidChunkLength(got length %r, %i bytes read)" % (
+            self.length,
+            self.partial,
+        )
+
+
+class InvalidHeader(HTTPError):
+    """The header provided was somehow invalid."""
+
+
+class ProxySchemeUnknown(AssertionError, URLSchemeUnknown):
+    """ProxyManager does not support the supplied scheme"""
+
+    # TODO(t-8ch): Stop inheriting from AssertionError in v2.0.
+
+    def __init__(self, scheme: str | None) -> None:
+        # 'localhost' is here because our URL parser parses
+        # localhost:8080 -> scheme=localhost, remove if we fix this.
+        if scheme == "localhost":
+            scheme = None
+        if scheme is None:
+            message = "Proxy URL had no scheme, should start with http:// or https://"
+        else:
+            message = f"Proxy URL had unsupported scheme {scheme}, should use http:// or https://"
+        super().__init__(message)
+
+
+class ProxySchemeUnsupported(ValueError):
+    """Fetching HTTPS resources through HTTPS proxies is unsupported"""
+
+
+class HeaderParsingError(HTTPError):
+    """Raised by assert_header_parsing, but we convert it to a log.warning statement."""
+
+    def __init__(
+        self, defects: list[MessageDefect], unparsed_data: bytes | str | None
+    ) -> None:
+        message = f"{defects or 'Unknown'}, unparsed data: {unparsed_data!r}"
+        super().__init__(message)
+
+
+class UnrewindableBodyError(HTTPError):
+    """urllib3 encountered an error when trying to rewind a body"""
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/fields.py b/backend/venv/lib/python3.12/site-packages/urllib3/fields.py
new file mode 100644
index 00000000..97c4730c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/fields.py
@@ -0,0 +1,341 @@
+from __future__ import annotations
+
+import email.utils
+import mimetypes
+import typing
+
+_TYPE_FIELD_VALUE = typing.Union[str, bytes]
+_TYPE_FIELD_VALUE_TUPLE = typing.Union[
+    _TYPE_FIELD_VALUE,
+    tuple[str, _TYPE_FIELD_VALUE],
+    tuple[str, _TYPE_FIELD_VALUE, str],
+]
+
+
+def guess_content_type(
+    filename: str | None, default: str = "application/octet-stream"
+) -> str:
+    """
+    Guess the "Content-Type" of a file.
+
+    :param filename:
+        The filename to guess the "Content-Type" of using :mod:`mimetypes`.
+    :param default:
+        If no "Content-Type" can be guessed, default to `default`.
+    """
+    if filename:
+        return mimetypes.guess_type(filename)[0] or default
+    return default
+
+
+def format_header_param_rfc2231(name: str, value: _TYPE_FIELD_VALUE) -> str:
+    """
+    Helper function to format and quote a single header parameter using the
+    strategy defined in RFC 2231.
+
+    Particularly useful for header parameters which might contain
+    non-ASCII values, like file names. This follows
+    `RFC 2388 Section 4.4 <https://tools.ietf.org/html/rfc2388#section-4.4>`_.
+
+    :param name:
+        The name of the parameter, a string expected to be ASCII only.
+    :param value:
+        The value of the parameter, provided as ``bytes`` or `str``.
+    :returns:
+        An RFC-2231-formatted unicode string.
+
+    .. deprecated:: 2.0.0
+        Will be removed in urllib3 v2.1.0. This is not valid for
+        ``multipart/form-data`` header parameters.
+    """
+    import warnings
+
+    warnings.warn(
+        "'format_header_param_rfc2231' is deprecated and will be "
+        "removed in urllib3 v2.1.0. This is not valid for "
+        "multipart/form-data header parameters.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+
+    if isinstance(value, bytes):
+        value = value.decode("utf-8")
+
+    if not any(ch in value for ch in '"\\\r\n'):
+        result = f'{name}="{value}"'
+        try:
+            result.encode("ascii")
+        except (UnicodeEncodeError, UnicodeDecodeError):
+            pass
+        else:
+            return result
+
+    value = email.utils.encode_rfc2231(value, "utf-8")
+    value = f"{name}*={value}"
+
+    return value
+
+
+def format_multipart_header_param(name: str, value: _TYPE_FIELD_VALUE) -> str:
+    """
+    Format and quote a single multipart header parameter.
+
+    This follows the `WHATWG HTML Standard`_ as of 2021/06/10, matching
+    the behavior of current browser and curl versions. Values are
+    assumed to be UTF-8. The ``\\n``, ``\\r``, and ``"`` characters are
+    percent encoded.
+
+    .. _WHATWG HTML Standard:
+        https://html.spec.whatwg.org/multipage/
+        form-control-infrastructure.html#multipart-form-data
+
+    :param name:
+        The name of the parameter, an ASCII-only ``str``.
+    :param value:
+        The value of the parameter, a ``str`` or UTF-8 encoded
+        ``bytes``.
+    :returns:
+        A string ``name="value"`` with the escaped value.
+
+    .. versionchanged:: 2.0.0
+        Matches the WHATWG HTML Standard as of 2021/06/10. Control
+        characters are no longer percent encoded.
+
+    .. versionchanged:: 2.0.0
+        Renamed from ``format_header_param_html5`` and
+        ``format_header_param``. The old names will be removed in
+        urllib3 v2.1.0.
+    """
+    if isinstance(value, bytes):
+        value = value.decode("utf-8")
+
+    # percent encode \n \r "
+    value = value.translate({10: "%0A", 13: "%0D", 34: "%22"})
+    return f'{name}="{value}"'
+
+
+def format_header_param_html5(name: str, value: _TYPE_FIELD_VALUE) -> str:
+    """
+    .. deprecated:: 2.0.0
+        Renamed to :func:`format_multipart_header_param`. Will be
+        removed in urllib3 v2.1.0.
+    """
+    import warnings
+
+    warnings.warn(
+        "'format_header_param_html5' has been renamed to "
+        "'format_multipart_header_param'. The old name will be "
+        "removed in urllib3 v2.1.0.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    return format_multipart_header_param(name, value)
+
+
+def format_header_param(name: str, value: _TYPE_FIELD_VALUE) -> str:
+    """
+    .. deprecated:: 2.0.0
+        Renamed to :func:`format_multipart_header_param`. Will be
+        removed in urllib3 v2.1.0.
+    """
+    import warnings
+
+    warnings.warn(
+        "'format_header_param' has been renamed to "
+        "'format_multipart_header_param'. The old name will be "
+        "removed in urllib3 v2.1.0.",
+        DeprecationWarning,
+        stacklevel=2,
+    )
+    return format_multipart_header_param(name, value)
+
+
+class RequestField:
+    """
+    A data container for request body parameters.
+
+    :param name:
+        The name of this request field. Must be unicode.
+    :param data:
+        The data/value body.
+    :param filename:
+        An optional filename of the request field. Must be unicode.
+    :param headers:
+        An optional dict-like object of headers to initially use for the field.
+
+    .. versionchanged:: 2.0.0
+        The ``header_formatter`` parameter is deprecated and will
+        be removed in urllib3 v2.1.0.
+    """
+
+    def __init__(
+        self,
+        name: str,
+        data: _TYPE_FIELD_VALUE,
+        filename: str | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        header_formatter: typing.Callable[[str, _TYPE_FIELD_VALUE], str] | None = None,
+    ):
+        self._name = name
+        self._filename = filename
+        self.data = data
+        self.headers: dict[str, str | None] = {}
+        if headers:
+            self.headers = dict(headers)
+
+        if header_formatter is not None:
+            import warnings
+
+            warnings.warn(
+                "The 'header_formatter' parameter is deprecated and "
+                "will be removed in urllib3 v2.1.0.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            self.header_formatter = header_formatter
+        else:
+            self.header_formatter = format_multipart_header_param
+
+    @classmethod
+    def from_tuples(
+        cls,
+        fieldname: str,
+        value: _TYPE_FIELD_VALUE_TUPLE,
+        header_formatter: typing.Callable[[str, _TYPE_FIELD_VALUE], str] | None = None,
+    ) -> RequestField:
+        """
+        A :class:`~urllib3.fields.RequestField` factory from old-style tuple parameters.
+
+        Supports constructing :class:`~urllib3.fields.RequestField` from
+        parameter of key/value strings AND key/filetuple. A filetuple is a
+        (filename, data, MIME type) tuple where the MIME type is optional.
+        For example::
+
+            'foo': 'bar',
+            'fakefile': ('foofile.txt', 'contents of foofile'),
+            'realfile': ('barfile.txt', open('realfile').read()),
+            'typedfile': ('bazfile.bin', open('bazfile').read(), 'image/jpeg'),
+            'nonamefile': 'contents of nonamefile field',
+
+        Field names and filenames must be unicode.
+        """
+        filename: str | None
+        content_type: str | None
+        data: _TYPE_FIELD_VALUE
+
+        if isinstance(value, tuple):
+            if len(value) == 3:
+                filename, data, content_type = value
+            else:
+                filename, data = value
+                content_type = guess_content_type(filename)
+        else:
+            filename = None
+            content_type = None
+            data = value
+
+        request_param = cls(
+            fieldname, data, filename=filename, header_formatter=header_formatter
+        )
+        request_param.make_multipart(content_type=content_type)
+
+        return request_param
+
+    def _render_part(self, name: str, value: _TYPE_FIELD_VALUE) -> str:
+        """
+        Override this method to change how each multipart header
+        parameter is formatted. By default, this calls
+        :func:`format_multipart_header_param`.
+
+        :param name:
+            The name of the parameter, an ASCII-only ``str``.
+        :param value:
+            The value of the parameter, a ``str`` or UTF-8 encoded
+            ``bytes``.
+
+        :meta public:
+        """
+        return self.header_formatter(name, value)
+
+    def _render_parts(
+        self,
+        header_parts: (
+            dict[str, _TYPE_FIELD_VALUE | None]
+            | typing.Sequence[tuple[str, _TYPE_FIELD_VALUE | None]]
+        ),
+    ) -> str:
+        """
+        Helper function to format and quote a single header.
+
+        Useful for single headers that are composed of multiple items. E.g.,
+        'Content-Disposition' fields.
+
+        :param header_parts:
+            A sequence of (k, v) tuples or a :class:`dict` of (k, v) to format
+            as `k1="v1"; k2="v2"; ...`.
+        """
+        iterable: typing.Iterable[tuple[str, _TYPE_FIELD_VALUE | None]]
+
+        parts = []
+        if isinstance(header_parts, dict):
+            iterable = header_parts.items()
+        else:
+            iterable = header_parts
+
+        for name, value in iterable:
+            if value is not None:
+                parts.append(self._render_part(name, value))
+
+        return "; ".join(parts)
+
+    def render_headers(self) -> str:
+        """
+        Renders the headers for this request field.
+        """
+        lines = []
+
+        sort_keys = ["Content-Disposition", "Content-Type", "Content-Location"]
+        for sort_key in sort_keys:
+            if self.headers.get(sort_key, False):
+                lines.append(f"{sort_key}: {self.headers[sort_key]}")
+
+        for header_name, header_value in self.headers.items():
+            if header_name not in sort_keys:
+                if header_value:
+                    lines.append(f"{header_name}: {header_value}")
+
+        lines.append("\r\n")
+        return "\r\n".join(lines)
+
+    def make_multipart(
+        self,
+        content_disposition: str | None = None,
+        content_type: str | None = None,
+        content_location: str | None = None,
+    ) -> None:
+        """
+        Makes this request field into a multipart request field.
+
+        This method overrides "Content-Disposition", "Content-Type" and
+        "Content-Location" headers to the request parameter.
+
+        :param content_disposition:
+            The 'Content-Disposition' of the request body. Defaults to 'form-data'
+        :param content_type:
+            The 'Content-Type' of the request body.
+        :param content_location:
+            The 'Content-Location' of the request body.
+
+        """
+        content_disposition = (content_disposition or "form-data") + "; ".join(
+            [
+                "",
+                self._render_parts(
+                    (("name", self._name), ("filename", self._filename))
+                ),
+            ]
+        )
+
+        self.headers["Content-Disposition"] = content_disposition
+        self.headers["Content-Type"] = content_type
+        self.headers["Content-Location"] = content_location
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/filepost.py b/backend/venv/lib/python3.12/site-packages/urllib3/filepost.py
new file mode 100644
index 00000000..14f70b05
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/filepost.py
@@ -0,0 +1,89 @@
+from __future__ import annotations
+
+import binascii
+import codecs
+import os
+import typing
+from io import BytesIO
+
+from .fields import _TYPE_FIELD_VALUE_TUPLE, RequestField
+
+writer = codecs.lookup("utf-8")[3]
+
+_TYPE_FIELDS_SEQUENCE = typing.Sequence[
+    typing.Union[tuple[str, _TYPE_FIELD_VALUE_TUPLE], RequestField]
+]
+_TYPE_FIELDS = typing.Union[
+    _TYPE_FIELDS_SEQUENCE,
+    typing.Mapping[str, _TYPE_FIELD_VALUE_TUPLE],
+]
+
+
+def choose_boundary() -> str:
+    """
+    Our embarrassingly-simple replacement for mimetools.choose_boundary.
+    """
+    return binascii.hexlify(os.urandom(16)).decode()
+
+
+def iter_field_objects(fields: _TYPE_FIELDS) -> typing.Iterable[RequestField]:
+    """
+    Iterate over fields.
+
+    Supports list of (k, v) tuples and dicts, and lists of
+    :class:`~urllib3.fields.RequestField`.
+
+    """
+    iterable: typing.Iterable[RequestField | tuple[str, _TYPE_FIELD_VALUE_TUPLE]]
+
+    if isinstance(fields, typing.Mapping):
+        iterable = fields.items()
+    else:
+        iterable = fields
+
+    for field in iterable:
+        if isinstance(field, RequestField):
+            yield field
+        else:
+            yield RequestField.from_tuples(*field)
+
+
+def encode_multipart_formdata(
+    fields: _TYPE_FIELDS, boundary: str | None = None
+) -> tuple[bytes, str]:
+    """
+    Encode a dictionary of ``fields`` using the multipart/form-data MIME format.
+
+    :param fields:
+        Dictionary of fields or list of (key, :class:`~urllib3.fields.RequestField`).
+        Values are processed by :func:`urllib3.fields.RequestField.from_tuples`.
+
+    :param boundary:
+        If not specified, then a random boundary will be generated using
+        :func:`urllib3.filepost.choose_boundary`.
+    """
+    body = BytesIO()
+    if boundary is None:
+        boundary = choose_boundary()
+
+    for field in iter_field_objects(fields):
+        body.write(f"--{boundary}\r\n".encode("latin-1"))
+
+        writer(body).write(field.render_headers())
+        data = field.data
+
+        if isinstance(data, int):
+            data = str(data)  # Backwards compatibility
+
+        if isinstance(data, str):
+            writer(body).write(data)
+        else:
+            body.write(data)
+
+        body.write(b"\r\n")
+
+    body.write(f"--{boundary}--\r\n".encode("latin-1"))
+
+    content_type = f"multipart/form-data; boundary={boundary}"
+
+    return body.getvalue(), content_type
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/__init__.py b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__init__.py
new file mode 100644
index 00000000..133e1d8f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__init__.py
@@ -0,0 +1,53 @@
+from __future__ import annotations
+
+from importlib.metadata import version
+
+__all__ = [
+    "inject_into_urllib3",
+    "extract_from_urllib3",
+]
+
+import typing
+
+orig_HTTPSConnection: typing.Any = None
+
+
+def inject_into_urllib3() -> None:
+    # First check if h2 version is valid
+    h2_version = version("h2")
+    if not h2_version.startswith("4."):
+        raise ImportError(
+            "urllib3 v2 supports h2 version 4.x.x, currently "
+            f"the 'h2' module is compiled with {h2_version!r}. "
+            "See: https://github.com/urllib3/urllib3/issues/3290"
+        )
+
+    # Import here to avoid circular dependencies.
+    from .. import connection as urllib3_connection
+    from .. import util as urllib3_util
+    from ..connectionpool import HTTPSConnectionPool
+    from ..util import ssl_ as urllib3_util_ssl
+    from .connection import HTTP2Connection
+
+    global orig_HTTPSConnection
+    orig_HTTPSConnection = urllib3_connection.HTTPSConnection
+
+    HTTPSConnectionPool.ConnectionCls = HTTP2Connection
+    urllib3_connection.HTTPSConnection = HTTP2Connection  # type: ignore[misc]
+
+    # TODO: Offer 'http/1.1' as well, but for testing purposes this is handy.
+    urllib3_util.ALPN_PROTOCOLS = ["h2"]
+    urllib3_util_ssl.ALPN_PROTOCOLS = ["h2"]
+
+
+def extract_from_urllib3() -> None:
+    from .. import connection as urllib3_connection
+    from .. import util as urllib3_util
+    from ..connectionpool import HTTPSConnectionPool
+    from ..util import ssl_ as urllib3_util_ssl
+
+    HTTPSConnectionPool.ConnectionCls = orig_HTTPSConnection
+    urllib3_connection.HTTPSConnection = orig_HTTPSConnection  # type: ignore[misc]
+
+    urllib3_util.ALPN_PROTOCOLS = ["http/1.1"]
+    urllib3_util_ssl.ALPN_PROTOCOLS = ["http/1.1"]
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..b3f312a9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..739287a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/probe.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/probe.cpython-312.pyc
new file mode 100644
index 00000000..4896f055
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/http2/__pycache__/probe.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/connection.py b/backend/venv/lib/python3.12/site-packages/urllib3/http2/connection.py
new file mode 100644
index 00000000..f4861452
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/http2/connection.py
@@ -0,0 +1,356 @@
+from __future__ import annotations
+
+import logging
+import re
+import threading
+import types
+import typing
+
+import h2.config  # type: ignore[import-untyped]
+import h2.connection  # type: ignore[import-untyped]
+import h2.events  # type: ignore[import-untyped]
+
+from .._base_connection import _TYPE_BODY
+from .._collections import HTTPHeaderDict
+from ..connection import HTTPSConnection, _get_default_user_agent
+from ..exceptions import ConnectionError
+from ..response import BaseHTTPResponse
+
+orig_HTTPSConnection = HTTPSConnection
+
+T = typing.TypeVar("T")
+
+log = logging.getLogger(__name__)
+
+RE_IS_LEGAL_HEADER_NAME = re.compile(rb"^[!#$%&'*+\-.^_`|~0-9a-z]+$")
+RE_IS_ILLEGAL_HEADER_VALUE = re.compile(rb"[\0\x00\x0a\x0d\r\n]|^[ \r\n\t]|[ \r\n\t]$")
+
+
+def _is_legal_header_name(name: bytes) -> bool:
+    """
+    "An implementation that validates fields according to the definitions in Sections
+    5.1 and 5.5 of [HTTP] only needs an additional check that field names do not
+    include uppercase characters." (https://httpwg.org/specs/rfc9113.html#n-field-validity)
+
+    `http.client._is_legal_header_name` does not validate the field name according to the
+    HTTP 1.1 spec, so we do that here, in addition to checking for uppercase characters.
+
+    This does not allow for the `:` character in the header name, so should not
+    be used to validate pseudo-headers.
+    """
+    return bool(RE_IS_LEGAL_HEADER_NAME.match(name))
+
+
+def _is_illegal_header_value(value: bytes) -> bool:
+    """
+    "A field value MUST NOT contain the zero value (ASCII NUL, 0x00), line feed
+    (ASCII LF, 0x0a), or carriage return (ASCII CR, 0x0d) at any position. A field
+    value MUST NOT start or end with an ASCII whitespace character (ASCII SP or HTAB,
+    0x20 or 0x09)." (https://httpwg.org/specs/rfc9113.html#n-field-validity)
+    """
+    return bool(RE_IS_ILLEGAL_HEADER_VALUE.search(value))
+
+
+class _LockedObject(typing.Generic[T]):
+    """
+    A wrapper class that hides a specific object behind a lock.
+    The goal here is to provide a simple way to protect access to an object
+    that cannot safely be simultaneously accessed from multiple threads. The
+    intended use of this class is simple: take hold of it with a context
+    manager, which returns the protected object.
+    """
+
+    __slots__ = (
+        "lock",
+        "_obj",
+    )
+
+    def __init__(self, obj: T):
+        self.lock = threading.RLock()
+        self._obj = obj
+
+    def __enter__(self) -> T:
+        self.lock.acquire()
+        return self._obj
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: types.TracebackType | None,
+    ) -> None:
+        self.lock.release()
+
+
+class HTTP2Connection(HTTPSConnection):
+    def __init__(
+        self, host: str, port: int | None = None, **kwargs: typing.Any
+    ) -> None:
+        self._h2_conn = self._new_h2_conn()
+        self._h2_stream: int | None = None
+        self._headers: list[tuple[bytes, bytes]] = []
+
+        if "proxy" in kwargs or "proxy_config" in kwargs:  # Defensive:
+            raise NotImplementedError("Proxies aren't supported with HTTP/2")
+
+        super().__init__(host, port, **kwargs)
+
+        if self._tunnel_host is not None:
+            raise NotImplementedError("Tunneling isn't supported with HTTP/2")
+
+    def _new_h2_conn(self) -> _LockedObject[h2.connection.H2Connection]:
+        config = h2.config.H2Configuration(client_side=True)
+        return _LockedObject(h2.connection.H2Connection(config=config))
+
+    def connect(self) -> None:
+        super().connect()
+        with self._h2_conn as conn:
+            conn.initiate_connection()
+            if data_to_send := conn.data_to_send():
+                self.sock.sendall(data_to_send)
+
+    def putrequest(  # type: ignore[override]
+        self,
+        method: str,
+        url: str,
+        **kwargs: typing.Any,
+    ) -> None:
+        """putrequest
+        This deviates from the HTTPConnection method signature since we never need to override
+        sending accept-encoding headers or the host header.
+        """
+        if "skip_host" in kwargs:
+            raise NotImplementedError("`skip_host` isn't supported")
+        if "skip_accept_encoding" in kwargs:
+            raise NotImplementedError("`skip_accept_encoding` isn't supported")
+
+        self._request_url = url or "/"
+        self._validate_path(url)  # type: ignore[attr-defined]
+
+        if ":" in self.host:
+            authority = f"[{self.host}]:{self.port or 443}"
+        else:
+            authority = f"{self.host}:{self.port or 443}"
+
+        self._headers.append((b":scheme", b"https"))
+        self._headers.append((b":method", method.encode()))
+        self._headers.append((b":authority", authority.encode()))
+        self._headers.append((b":path", url.encode()))
+
+        with self._h2_conn as conn:
+            self._h2_stream = conn.get_next_available_stream_id()
+
+    def putheader(self, header: str | bytes, *values: str | bytes) -> None:
+        # TODO SKIPPABLE_HEADERS from urllib3 are ignored.
+        header = header.encode() if isinstance(header, str) else header
+        header = header.lower()  # A lot of upstream code uses capitalized headers.
+        if not _is_legal_header_name(header):
+            raise ValueError(f"Illegal header name {str(header)}")
+
+        for value in values:
+            value = value.encode() if isinstance(value, str) else value
+            if _is_illegal_header_value(value):
+                raise ValueError(f"Illegal header value {str(value)}")
+            self._headers.append((header, value))
+
+    def endheaders(self, message_body: typing.Any = None) -> None:  # type: ignore[override]
+        if self._h2_stream is None:
+            raise ConnectionError("Must call `putrequest` first.")
+
+        with self._h2_conn as conn:
+            conn.send_headers(
+                stream_id=self._h2_stream,
+                headers=self._headers,
+                end_stream=(message_body is None),
+            )
+            if data_to_send := conn.data_to_send():
+                self.sock.sendall(data_to_send)
+        self._headers = []  # Reset headers for the next request.
+
+    def send(self, data: typing.Any) -> None:
+        """Send data to the server.
+        `data` can be: `str`, `bytes`, an iterable, or file-like objects
+        that support a .read() method.
+        """
+        if self._h2_stream is None:
+            raise ConnectionError("Must call `putrequest` first.")
+
+        with self._h2_conn as conn:
+            if data_to_send := conn.data_to_send():
+                self.sock.sendall(data_to_send)
+
+            if hasattr(data, "read"):  # file-like objects
+                while True:
+                    chunk = data.read(self.blocksize)
+                    if not chunk:
+                        break
+                    if isinstance(chunk, str):
+                        chunk = chunk.encode()  # pragma: no cover
+                    conn.send_data(self._h2_stream, chunk, end_stream=False)
+                    if data_to_send := conn.data_to_send():
+                        self.sock.sendall(data_to_send)
+                conn.end_stream(self._h2_stream)
+                return
+
+            if isinstance(data, str):  # str -> bytes
+                data = data.encode()
+
+            try:
+                if isinstance(data, bytes):
+                    conn.send_data(self._h2_stream, data, end_stream=True)
+                    if data_to_send := conn.data_to_send():
+                        self.sock.sendall(data_to_send)
+                else:
+                    for chunk in data:
+                        conn.send_data(self._h2_stream, chunk, end_stream=False)
+                        if data_to_send := conn.data_to_send():
+                            self.sock.sendall(data_to_send)
+                    conn.end_stream(self._h2_stream)
+            except TypeError:
+                raise TypeError(
+                    "`data` should be str, bytes, iterable, or file. got %r"
+                    % type(data)
+                )
+
+    def set_tunnel(
+        self,
+        host: str,
+        port: int | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        scheme: str = "http",
+    ) -> None:
+        raise NotImplementedError(
+            "HTTP/2 does not support setting up a tunnel through a proxy"
+        )
+
+    def getresponse(  # type: ignore[override]
+        self,
+    ) -> HTTP2Response:
+        status = None
+        data = bytearray()
+        with self._h2_conn as conn:
+            end_stream = False
+            while not end_stream:
+                # TODO: Arbitrary read value.
+                if received_data := self.sock.recv(65535):
+                    events = conn.receive_data(received_data)
+                    for event in events:
+                        if isinstance(event, h2.events.ResponseReceived):
+                            headers = HTTPHeaderDict()
+                            for header, value in event.headers:
+                                if header == b":status":
+                                    status = int(value.decode())
+                                else:
+                                    headers.add(
+                                        header.decode("ascii"), value.decode("ascii")
+                                    )
+
+                        elif isinstance(event, h2.events.DataReceived):
+                            data += event.data
+                            conn.acknowledge_received_data(
+                                event.flow_controlled_length, event.stream_id
+                            )
+
+                        elif isinstance(event, h2.events.StreamEnded):
+                            end_stream = True
+
+                if data_to_send := conn.data_to_send():
+                    self.sock.sendall(data_to_send)
+
+        assert status is not None
+        return HTTP2Response(
+            status=status,
+            headers=headers,
+            request_url=self._request_url,
+            data=bytes(data),
+        )
+
+    def request(  # type: ignore[override]
+        self,
+        method: str,
+        url: str,
+        body: _TYPE_BODY | None = None,
+        headers: typing.Mapping[str, str] | None = None,
+        *,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        enforce_content_length: bool = True,
+        **kwargs: typing.Any,
+    ) -> None:
+        """Send an HTTP/2 request"""
+        if "chunked" in kwargs:
+            # TODO this is often present from upstream.
+            # raise NotImplementedError("`chunked` isn't supported with HTTP/2")
+            pass
+
+        if self.sock is not None:
+            self.sock.settimeout(self.timeout)
+
+        self.putrequest(method, url)
+
+        headers = headers or {}
+        for k, v in headers.items():
+            if k.lower() == "transfer-encoding" and v == "chunked":
+                continue
+            else:
+                self.putheader(k, v)
+
+        if b"user-agent" not in dict(self._headers):
+            self.putheader(b"user-agent", _get_default_user_agent())
+
+        if body:
+            self.endheaders(message_body=body)
+            self.send(body)
+        else:
+            self.endheaders()
+
+    def close(self) -> None:
+        with self._h2_conn as conn:
+            try:
+                conn.close_connection()
+                if data := conn.data_to_send():
+                    self.sock.sendall(data)
+            except Exception:
+                pass
+
+        # Reset all our HTTP/2 connection state.
+        self._h2_conn = self._new_h2_conn()
+        self._h2_stream = None
+        self._headers = []
+
+        super().close()
+
+
+class HTTP2Response(BaseHTTPResponse):
+    # TODO: This is a woefully incomplete response object, but works for non-streaming.
+    def __init__(
+        self,
+        status: int,
+        headers: HTTPHeaderDict,
+        request_url: str,
+        data: bytes,
+        decode_content: bool = False,  # TODO: support decoding
+    ) -> None:
+        super().__init__(
+            status=status,
+            headers=headers,
+            # Following CPython, we map HTTP versions to major * 10 + minor integers
+            version=20,
+            version_string="HTTP/2",
+            # No reason phrase in HTTP/2
+            reason=None,
+            decode_content=decode_content,
+            request_url=request_url,
+        )
+        self._data = data
+        self.length_remaining = 0
+
+    @property
+    def data(self) -> bytes:
+        return self._data
+
+    def get_redirect_location(self) -> None:
+        return None
+
+    def close(self) -> None:
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/http2/probe.py b/backend/venv/lib/python3.12/site-packages/urllib3/http2/probe.py
new file mode 100644
index 00000000..9ea90076
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/http2/probe.py
@@ -0,0 +1,87 @@
+from __future__ import annotations
+
+import threading
+
+
+class _HTTP2ProbeCache:
+    __slots__ = (
+        "_lock",
+        "_cache_locks",
+        "_cache_values",
+    )
+
+    def __init__(self) -> None:
+        self._lock = threading.Lock()
+        self._cache_locks: dict[tuple[str, int], threading.RLock] = {}
+        self._cache_values: dict[tuple[str, int], bool | None] = {}
+
+    def acquire_and_get(self, host: str, port: int) -> bool | None:
+        # By the end of this block we know that
+        # _cache_[values,locks] is available.
+        value = None
+        with self._lock:
+            key = (host, port)
+            try:
+                value = self._cache_values[key]
+                # If it's a known value we return right away.
+                if value is not None:
+                    return value
+            except KeyError:
+                self._cache_locks[key] = threading.RLock()
+                self._cache_values[key] = None
+
+        # If the value is unknown, we acquire the lock to signal
+        # to the requesting thread that the probe is in progress
+        # or that the current thread needs to return their findings.
+        key_lock = self._cache_locks[key]
+        key_lock.acquire()
+        try:
+            # If the by the time we get the lock the value has been
+            # updated we want to return the updated value.
+            value = self._cache_values[key]
+
+        # In case an exception like KeyboardInterrupt is raised here.
+        except BaseException as e:  # Defensive:
+            assert not isinstance(e, KeyError)  # KeyError shouldn't be possible.
+            key_lock.release()
+            raise
+
+        return value
+
+    def set_and_release(
+        self, host: str, port: int, supports_http2: bool | None
+    ) -> None:
+        key = (host, port)
+        key_lock = self._cache_locks[key]
+        with key_lock:  # Uses an RLock, so can be locked again from same thread.
+            if supports_http2 is None and self._cache_values[key] is not None:
+                raise ValueError(
+                    "Cannot reset HTTP/2 support for origin after value has been set."
+                )  # Defensive: not expected in normal usage
+
+        self._cache_values[key] = supports_http2
+        key_lock.release()
+
+    def _values(self) -> dict[tuple[str, int], bool | None]:
+        """This function is for testing purposes only. Gets the current state of the probe cache"""
+        with self._lock:
+            return {k: v for k, v in self._cache_values.items()}
+
+    def _reset(self) -> None:
+        """This function is for testing purposes only. Reset the cache values"""
+        with self._lock:
+            self._cache_locks = {}
+            self._cache_values = {}
+
+
+_HTTP2_PROBE_CACHE = _HTTP2ProbeCache()
+
+set_and_release = _HTTP2_PROBE_CACHE.set_and_release
+acquire_and_get = _HTTP2_PROBE_CACHE.acquire_and_get
+_values = _HTTP2_PROBE_CACHE._values
+_reset = _HTTP2_PROBE_CACHE._reset
+
+__all__ = [
+    "set_and_release",
+    "acquire_and_get",
+]
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/poolmanager.py b/backend/venv/lib/python3.12/site-packages/urllib3/poolmanager.py
new file mode 100644
index 00000000..085d1dba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/poolmanager.py
@@ -0,0 +1,637 @@
+from __future__ import annotations
+
+import functools
+import logging
+import typing
+import warnings
+from types import TracebackType
+from urllib.parse import urljoin
+
+from ._collections import HTTPHeaderDict, RecentlyUsedContainer
+from ._request_methods import RequestMethods
+from .connection import ProxyConfig
+from .connectionpool import HTTPConnectionPool, HTTPSConnectionPool, port_by_scheme
+from .exceptions import (
+    LocationValueError,
+    MaxRetryError,
+    ProxySchemeUnknown,
+    URLSchemeUnknown,
+)
+from .response import BaseHTTPResponse
+from .util.connection import _TYPE_SOCKET_OPTIONS
+from .util.proxy import connection_requires_http_tunnel
+from .util.retry import Retry
+from .util.timeout import Timeout
+from .util.url import Url, parse_url
+
+if typing.TYPE_CHECKING:
+    import ssl
+
+    from typing_extensions import Self
+
+__all__ = ["PoolManager", "ProxyManager", "proxy_from_url"]
+
+
+log = logging.getLogger(__name__)
+
+SSL_KEYWORDS = (
+    "key_file",
+    "cert_file",
+    "cert_reqs",
+    "ca_certs",
+    "ca_cert_data",
+    "ssl_version",
+    "ssl_minimum_version",
+    "ssl_maximum_version",
+    "ca_cert_dir",
+    "ssl_context",
+    "key_password",
+    "server_hostname",
+)
+# Default value for `blocksize` - a new parameter introduced to
+# http.client.HTTPConnection & http.client.HTTPSConnection in Python 3.7
+_DEFAULT_BLOCKSIZE = 16384
+
+
+class PoolKey(typing.NamedTuple):
+    """
+    All known keyword arguments that could be provided to the pool manager, its
+    pools, or the underlying connections.
+
+    All custom key schemes should include the fields in this key at a minimum.
+    """
+
+    key_scheme: str
+    key_host: str
+    key_port: int | None
+    key_timeout: Timeout | float | int | None
+    key_retries: Retry | bool | int | None
+    key_block: bool | None
+    key_source_address: tuple[str, int] | None
+    key_key_file: str | None
+    key_key_password: str | None
+    key_cert_file: str | None
+    key_cert_reqs: str | None
+    key_ca_certs: str | None
+    key_ca_cert_data: str | bytes | None
+    key_ssl_version: int | str | None
+    key_ssl_minimum_version: ssl.TLSVersion | None
+    key_ssl_maximum_version: ssl.TLSVersion | None
+    key_ca_cert_dir: str | None
+    key_ssl_context: ssl.SSLContext | None
+    key_maxsize: int | None
+    key_headers: frozenset[tuple[str, str]] | None
+    key__proxy: Url | None
+    key__proxy_headers: frozenset[tuple[str, str]] | None
+    key__proxy_config: ProxyConfig | None
+    key_socket_options: _TYPE_SOCKET_OPTIONS | None
+    key__socks_options: frozenset[tuple[str, str]] | None
+    key_assert_hostname: bool | str | None
+    key_assert_fingerprint: str | None
+    key_server_hostname: str | None
+    key_blocksize: int | None
+
+
+def _default_key_normalizer(
+    key_class: type[PoolKey], request_context: dict[str, typing.Any]
+) -> PoolKey:
+    """
+    Create a pool key out of a request context dictionary.
+
+    According to RFC 3986, both the scheme and host are case-insensitive.
+    Therefore, this function normalizes both before constructing the pool
+    key for an HTTPS request. If you wish to change this behaviour, provide
+    alternate callables to ``key_fn_by_scheme``.
+
+    :param key_class:
+        The class to use when constructing the key. This should be a namedtuple
+        with the ``scheme`` and ``host`` keys at a minimum.
+    :type  key_class: namedtuple
+    :param request_context:
+        A dictionary-like object that contain the context for a request.
+    :type  request_context: dict
+
+    :return: A namedtuple that can be used as a connection pool key.
+    :rtype:  PoolKey
+    """
+    # Since we mutate the dictionary, make a copy first
+    context = request_context.copy()
+    context["scheme"] = context["scheme"].lower()
+    context["host"] = context["host"].lower()
+
+    # These are both dictionaries and need to be transformed into frozensets
+    for key in ("headers", "_proxy_headers", "_socks_options"):
+        if key in context and context[key] is not None:
+            context[key] = frozenset(context[key].items())
+
+    # The socket_options key may be a list and needs to be transformed into a
+    # tuple.
+    socket_opts = context.get("socket_options")
+    if socket_opts is not None:
+        context["socket_options"] = tuple(socket_opts)
+
+    # Map the kwargs to the names in the namedtuple - this is necessary since
+    # namedtuples can't have fields starting with '_'.
+    for key in list(context.keys()):
+        context["key_" + key] = context.pop(key)
+
+    # Default to ``None`` for keys missing from the context
+    for field in key_class._fields:
+        if field not in context:
+            context[field] = None
+
+    # Default key_blocksize to _DEFAULT_BLOCKSIZE if missing from the context
+    if context.get("key_blocksize") is None:
+        context["key_blocksize"] = _DEFAULT_BLOCKSIZE
+
+    return key_class(**context)
+
+
+#: A dictionary that maps a scheme to a callable that creates a pool key.
+#: This can be used to alter the way pool keys are constructed, if desired.
+#: Each PoolManager makes a copy of this dictionary so they can be configured
+#: globally here, or individually on the instance.
+key_fn_by_scheme = {
+    "http": functools.partial(_default_key_normalizer, PoolKey),
+    "https": functools.partial(_default_key_normalizer, PoolKey),
+}
+
+pool_classes_by_scheme = {"http": HTTPConnectionPool, "https": HTTPSConnectionPool}
+
+
+class PoolManager(RequestMethods):
+    """
+    Allows for arbitrary requests while transparently keeping track of
+    necessary connection pools for you.
+
+    :param num_pools:
+        Number of connection pools to cache before discarding the least
+        recently used pool.
+
+    :param headers:
+        Headers to include with all requests, unless other headers are given
+        explicitly.
+
+    :param \\**connection_pool_kw:
+        Additional parameters are used to create fresh
+        :class:`urllib3.connectionpool.ConnectionPool` instances.
+
+    Example:
+
+    .. code-block:: python
+
+        import urllib3
+
+        http = urllib3.PoolManager(num_pools=2)
+
+        resp1 = http.request("GET", "https://google.com/")
+        resp2 = http.request("GET", "https://google.com/mail")
+        resp3 = http.request("GET", "https://yahoo.com/")
+
+        print(len(http.pools))
+        # 2
+
+    """
+
+    proxy: Url | None = None
+    proxy_config: ProxyConfig | None = None
+
+    def __init__(
+        self,
+        num_pools: int = 10,
+        headers: typing.Mapping[str, str] | None = None,
+        **connection_pool_kw: typing.Any,
+    ) -> None:
+        super().__init__(headers)
+        self.connection_pool_kw = connection_pool_kw
+
+        self.pools: RecentlyUsedContainer[PoolKey, HTTPConnectionPool]
+        self.pools = RecentlyUsedContainer(num_pools)
+
+        # Locally set the pool classes and keys so other PoolManagers can
+        # override them.
+        self.pool_classes_by_scheme = pool_classes_by_scheme
+        self.key_fn_by_scheme = key_fn_by_scheme.copy()
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: TracebackType | None,
+    ) -> typing.Literal[False]:
+        self.clear()
+        # Return False to re-raise any potential exceptions
+        return False
+
+    def _new_pool(
+        self,
+        scheme: str,
+        host: str,
+        port: int,
+        request_context: dict[str, typing.Any] | None = None,
+    ) -> HTTPConnectionPool:
+        """
+        Create a new :class:`urllib3.connectionpool.ConnectionPool` based on host, port, scheme, and
+        any additional pool keyword arguments.
+
+        If ``request_context`` is provided, it is provided as keyword arguments
+        to the pool class used. This method is used to actually create the
+        connection pools handed out by :meth:`connection_from_url` and
+        companion methods. It is intended to be overridden for customization.
+        """
+        pool_cls: type[HTTPConnectionPool] = self.pool_classes_by_scheme[scheme]
+        if request_context is None:
+            request_context = self.connection_pool_kw.copy()
+
+        # Default blocksize to _DEFAULT_BLOCKSIZE if missing or explicitly
+        # set to 'None' in the request_context.
+        if request_context.get("blocksize") is None:
+            request_context["blocksize"] = _DEFAULT_BLOCKSIZE
+
+        # Although the context has everything necessary to create the pool,
+        # this function has historically only used the scheme, host, and port
+        # in the positional args. When an API change is acceptable these can
+        # be removed.
+        for key in ("scheme", "host", "port"):
+            request_context.pop(key, None)
+
+        if scheme == "http":
+            for kw in SSL_KEYWORDS:
+                request_context.pop(kw, None)
+
+        return pool_cls(host, port, **request_context)
+
+    def clear(self) -> None:
+        """
+        Empty our store of pools and direct them all to close.
+
+        This will not affect in-flight connections, but they will not be
+        re-used after completion.
+        """
+        self.pools.clear()
+
+    def connection_from_host(
+        self,
+        host: str | None,
+        port: int | None = None,
+        scheme: str | None = "http",
+        pool_kwargs: dict[str, typing.Any] | None = None,
+    ) -> HTTPConnectionPool:
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the host, port, and scheme.
+
+        If ``port`` isn't given, it will be derived from the ``scheme`` using
+        ``urllib3.connectionpool.port_by_scheme``. If ``pool_kwargs`` is
+        provided, it is merged with the instance's ``connection_pool_kw``
+        variable and used to create the new connection pool, if one is
+        needed.
+        """
+
+        if not host:
+            raise LocationValueError("No host specified.")
+
+        request_context = self._merge_pool_kwargs(pool_kwargs)
+        request_context["scheme"] = scheme or "http"
+        if not port:
+            port = port_by_scheme.get(request_context["scheme"].lower(), 80)
+        request_context["port"] = port
+        request_context["host"] = host
+
+        return self.connection_from_context(request_context)
+
+    def connection_from_context(
+        self, request_context: dict[str, typing.Any]
+    ) -> HTTPConnectionPool:
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the request context.
+
+        ``request_context`` must at least contain the ``scheme`` key and its
+        value must be a key in ``key_fn_by_scheme`` instance variable.
+        """
+        if "strict" in request_context:
+            warnings.warn(
+                "The 'strict' parameter is no longer needed on Python 3+. "
+                "This will raise an error in urllib3 v2.1.0.",
+                DeprecationWarning,
+            )
+            request_context.pop("strict")
+
+        scheme = request_context["scheme"].lower()
+        pool_key_constructor = self.key_fn_by_scheme.get(scheme)
+        if not pool_key_constructor:
+            raise URLSchemeUnknown(scheme)
+        pool_key = pool_key_constructor(request_context)
+
+        return self.connection_from_pool_key(pool_key, request_context=request_context)
+
+    def connection_from_pool_key(
+        self, pool_key: PoolKey, request_context: dict[str, typing.Any]
+    ) -> HTTPConnectionPool:
+        """
+        Get a :class:`urllib3.connectionpool.ConnectionPool` based on the provided pool key.
+
+        ``pool_key`` should be a namedtuple that only contains immutable
+        objects. At a minimum it must have the ``scheme``, ``host``, and
+        ``port`` fields.
+        """
+        with self.pools.lock:
+            # If the scheme, host, or port doesn't match existing open
+            # connections, open a new ConnectionPool.
+            pool = self.pools.get(pool_key)
+            if pool:
+                return pool
+
+            # Make a fresh ConnectionPool of the desired type
+            scheme = request_context["scheme"]
+            host = request_context["host"]
+            port = request_context["port"]
+            pool = self._new_pool(scheme, host, port, request_context=request_context)
+            self.pools[pool_key] = pool
+
+        return pool
+
+    def connection_from_url(
+        self, url: str, pool_kwargs: dict[str, typing.Any] | None = None
+    ) -> HTTPConnectionPool:
+        """
+        Similar to :func:`urllib3.connectionpool.connection_from_url`.
+
+        If ``pool_kwargs`` is not provided and a new pool needs to be
+        constructed, ``self.connection_pool_kw`` is used to initialize
+        the :class:`urllib3.connectionpool.ConnectionPool`. If ``pool_kwargs``
+        is provided, it is used instead. Note that if a new pool does not
+        need to be created for the request, the provided ``pool_kwargs`` are
+        not used.
+        """
+        u = parse_url(url)
+        return self.connection_from_host(
+            u.host, port=u.port, scheme=u.scheme, pool_kwargs=pool_kwargs
+        )
+
+    def _merge_pool_kwargs(
+        self, override: dict[str, typing.Any] | None
+    ) -> dict[str, typing.Any]:
+        """
+        Merge a dictionary of override values for self.connection_pool_kw.
+
+        This does not modify self.connection_pool_kw and returns a new dict.
+        Any keys in the override dictionary with a value of ``None`` are
+        removed from the merged dictionary.
+        """
+        base_pool_kwargs = self.connection_pool_kw.copy()
+        if override:
+            for key, value in override.items():
+                if value is None:
+                    try:
+                        del base_pool_kwargs[key]
+                    except KeyError:
+                        pass
+                else:
+                    base_pool_kwargs[key] = value
+        return base_pool_kwargs
+
+    def _proxy_requires_url_absolute_form(self, parsed_url: Url) -> bool:
+        """
+        Indicates if the proxy requires the complete destination URL in the
+        request.  Normally this is only needed when not using an HTTP CONNECT
+        tunnel.
+        """
+        if self.proxy is None:
+            return False
+
+        return not connection_requires_http_tunnel(
+            self.proxy, self.proxy_config, parsed_url.scheme
+        )
+
+    def urlopen(  # type: ignore[override]
+        self, method: str, url: str, redirect: bool = True, **kw: typing.Any
+    ) -> BaseHTTPResponse:
+        """
+        Same as :meth:`urllib3.HTTPConnectionPool.urlopen`
+        with custom cross-host redirect logic and only sends the request-uri
+        portion of the ``url``.
+
+        The given ``url`` parameter must be absolute, such that an appropriate
+        :class:`urllib3.connectionpool.ConnectionPool` can be chosen for it.
+        """
+        u = parse_url(url)
+
+        if u.scheme is None:
+            warnings.warn(
+                "URLs without a scheme (ie 'https://') are deprecated and will raise an error "
+                "in a future version of urllib3. To avoid this DeprecationWarning ensure all URLs "
+                "start with 'https://' or 'http://'. Read more in this issue: "
+                "https://github.com/urllib3/urllib3/issues/2920",
+                category=DeprecationWarning,
+                stacklevel=2,
+            )
+
+        conn = self.connection_from_host(u.host, port=u.port, scheme=u.scheme)
+
+        kw["assert_same_host"] = False
+        kw["redirect"] = False
+
+        if "headers" not in kw:
+            kw["headers"] = self.headers
+
+        if self._proxy_requires_url_absolute_form(u):
+            response = conn.urlopen(method, url, **kw)
+        else:
+            response = conn.urlopen(method, u.request_uri, **kw)
+
+        redirect_location = redirect and response.get_redirect_location()
+        if not redirect_location:
+            return response
+
+        # Support relative URLs for redirecting.
+        redirect_location = urljoin(url, redirect_location)
+
+        if response.status == 303:
+            # Change the method according to RFC 9110, Section 15.4.4.
+            method = "GET"
+            # And lose the body not to transfer anything sensitive.
+            kw["body"] = None
+            kw["headers"] = HTTPHeaderDict(kw["headers"])._prepare_for_method_change()
+
+        retries = kw.get("retries")
+        if not isinstance(retries, Retry):
+            retries = Retry.from_int(retries, redirect=redirect)
+
+        # Strip headers marked as unsafe to forward to the redirected location.
+        # Check remove_headers_on_redirect to avoid a potential network call within
+        # conn.is_same_host() which may use socket.gethostbyname() in the future.
+        if retries.remove_headers_on_redirect and not conn.is_same_host(
+            redirect_location
+        ):
+            new_headers = kw["headers"].copy()
+            for header in kw["headers"]:
+                if header.lower() in retries.remove_headers_on_redirect:
+                    new_headers.pop(header, None)
+            kw["headers"] = new_headers
+
+        try:
+            retries = retries.increment(method, url, response=response, _pool=conn)
+        except MaxRetryError:
+            if retries.raise_on_redirect:
+                response.drain_conn()
+                raise
+            return response
+
+        kw["retries"] = retries
+        kw["redirect"] = redirect
+
+        log.info("Redirecting %s -> %s", url, redirect_location)
+
+        response.drain_conn()
+        return self.urlopen(method, redirect_location, **kw)
+
+
+class ProxyManager(PoolManager):
+    """
+    Behaves just like :class:`PoolManager`, but sends all requests through
+    the defined proxy, using the CONNECT method for HTTPS URLs.
+
+    :param proxy_url:
+        The URL of the proxy to be used.
+
+    :param proxy_headers:
+        A dictionary containing headers that will be sent to the proxy. In case
+        of HTTP they are being sent with each request, while in the
+        HTTPS/CONNECT case they are sent only once. Could be used for proxy
+        authentication.
+
+    :param proxy_ssl_context:
+        The proxy SSL context is used to establish the TLS connection to the
+        proxy when using HTTPS proxies.
+
+    :param use_forwarding_for_https:
+        (Defaults to False) If set to True will forward requests to the HTTPS
+        proxy to be made on behalf of the client instead of creating a TLS
+        tunnel via the CONNECT method. **Enabling this flag means that request
+        and response headers and content will be visible from the HTTPS proxy**
+        whereas tunneling keeps request and response headers and content
+        private.  IP address, target hostname, SNI, and port are always visible
+        to an HTTPS proxy even when this flag is disabled.
+
+    :param proxy_assert_hostname:
+        The hostname of the certificate to verify against.
+
+    :param proxy_assert_fingerprint:
+        The fingerprint of the certificate to verify against.
+
+    Example:
+
+    .. code-block:: python
+
+        import urllib3
+
+        proxy = urllib3.ProxyManager("https://localhost:3128/")
+
+        resp1 = proxy.request("GET", "https://google.com/")
+        resp2 = proxy.request("GET", "https://httpbin.org/")
+
+        print(len(proxy.pools))
+        # 1
+
+        resp3 = proxy.request("GET", "https://httpbin.org/")
+        resp4 = proxy.request("GET", "https://twitter.com/")
+
+        print(len(proxy.pools))
+        # 3
+
+    """
+
+    def __init__(
+        self,
+        proxy_url: str,
+        num_pools: int = 10,
+        headers: typing.Mapping[str, str] | None = None,
+        proxy_headers: typing.Mapping[str, str] | None = None,
+        proxy_ssl_context: ssl.SSLContext | None = None,
+        use_forwarding_for_https: bool = False,
+        proxy_assert_hostname: None | str | typing.Literal[False] = None,
+        proxy_assert_fingerprint: str | None = None,
+        **connection_pool_kw: typing.Any,
+    ) -> None:
+        if isinstance(proxy_url, HTTPConnectionPool):
+            str_proxy_url = f"{proxy_url.scheme}://{proxy_url.host}:{proxy_url.port}"
+        else:
+            str_proxy_url = proxy_url
+        proxy = parse_url(str_proxy_url)
+
+        if proxy.scheme not in ("http", "https"):
+            raise ProxySchemeUnknown(proxy.scheme)
+
+        if not proxy.port:
+            port = port_by_scheme.get(proxy.scheme, 80)
+            proxy = proxy._replace(port=port)
+
+        self.proxy = proxy
+        self.proxy_headers = proxy_headers or {}
+        self.proxy_ssl_context = proxy_ssl_context
+        self.proxy_config = ProxyConfig(
+            proxy_ssl_context,
+            use_forwarding_for_https,
+            proxy_assert_hostname,
+            proxy_assert_fingerprint,
+        )
+
+        connection_pool_kw["_proxy"] = self.proxy
+        connection_pool_kw["_proxy_headers"] = self.proxy_headers
+        connection_pool_kw["_proxy_config"] = self.proxy_config
+
+        super().__init__(num_pools, headers, **connection_pool_kw)
+
+    def connection_from_host(
+        self,
+        host: str | None,
+        port: int | None = None,
+        scheme: str | None = "http",
+        pool_kwargs: dict[str, typing.Any] | None = None,
+    ) -> HTTPConnectionPool:
+        if scheme == "https":
+            return super().connection_from_host(
+                host, port, scheme, pool_kwargs=pool_kwargs
+            )
+
+        return super().connection_from_host(
+            self.proxy.host, self.proxy.port, self.proxy.scheme, pool_kwargs=pool_kwargs  # type: ignore[union-attr]
+        )
+
+    def _set_proxy_headers(
+        self, url: str, headers: typing.Mapping[str, str] | None = None
+    ) -> typing.Mapping[str, str]:
+        """
+        Sets headers needed by proxies: specifically, the Accept and Host
+        headers. Only sets headers not provided by the user.
+        """
+        headers_ = {"Accept": "*/*"}
+
+        netloc = parse_url(url).netloc
+        if netloc:
+            headers_["Host"] = netloc
+
+        if headers:
+            headers_.update(headers)
+        return headers_
+
+    def urlopen(  # type: ignore[override]
+        self, method: str, url: str, redirect: bool = True, **kw: typing.Any
+    ) -> BaseHTTPResponse:
+        "Same as HTTP(S)ConnectionPool.urlopen, ``url`` must be absolute."
+        u = parse_url(url)
+        if not connection_requires_http_tunnel(self.proxy, self.proxy_config, u.scheme):
+            # For connections using HTTP CONNECT, httplib sets the necessary
+            # headers on the CONNECT to the proxy. If we're not using CONNECT,
+            # we'll definitely need to set 'Host' at the very least.
+            headers = kw.get("headers", self.headers)
+            kw["headers"] = self._set_proxy_headers(url, headers)
+
+        return super().urlopen(method, url, redirect=redirect, **kw)
+
+
+def proxy_from_url(url: str, **kw: typing.Any) -> ProxyManager:
+    return ProxyManager(proxy_url=url, **kw)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/py.typed b/backend/venv/lib/python3.12/site-packages/urllib3/py.typed
new file mode 100644
index 00000000..5f3ea3d9
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/py.typed
@@ -0,0 +1,2 @@
+# Instruct type checkers to look for inline type annotations in this package.
+# See PEP 561.
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/response.py b/backend/venv/lib/python3.12/site-packages/urllib3/response.py
new file mode 100644
index 00000000..66c6a687
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/response.py
@@ -0,0 +1,1278 @@
+from __future__ import annotations
+
+import collections
+import io
+import json as _json
+import logging
+import re
+import socket
+import sys
+import typing
+import warnings
+import zlib
+from contextlib import contextmanager
+from http.client import HTTPMessage as _HttplibHTTPMessage
+from http.client import HTTPResponse as _HttplibHTTPResponse
+from socket import timeout as SocketTimeout
+
+if typing.TYPE_CHECKING:
+    from ._base_connection import BaseHTTPConnection
+
+try:
+    try:
+        import brotlicffi as brotli  # type: ignore[import-not-found]
+    except ImportError:
+        import brotli  # type: ignore[import-not-found]
+except ImportError:
+    brotli = None
+
+try:
+    import zstandard as zstd
+except (AttributeError, ImportError, ValueError):  # Defensive:
+    HAS_ZSTD = False
+else:
+    # The package 'zstandard' added the 'eof' property starting
+    # in v0.18.0 which we require to ensure a complete and
+    # valid zstd stream was fed into the ZstdDecoder.
+    # See: https://github.com/urllib3/urllib3/pull/2624
+    _zstd_version = tuple(
+        map(int, re.search(r"^([0-9]+)\.([0-9]+)", zstd.__version__).groups())  # type: ignore[union-attr]
+    )
+    if _zstd_version < (0, 18):  # Defensive:
+        HAS_ZSTD = False
+    else:
+        HAS_ZSTD = True
+
+from . import util
+from ._base_connection import _TYPE_BODY
+from ._collections import HTTPHeaderDict
+from .connection import BaseSSLError, HTTPConnection, HTTPException
+from .exceptions import (
+    BodyNotHttplibCompatible,
+    DecodeError,
+    HTTPError,
+    IncompleteRead,
+    InvalidChunkLength,
+    InvalidHeader,
+    ProtocolError,
+    ReadTimeoutError,
+    ResponseNotChunked,
+    SSLError,
+)
+from .util.response import is_fp_closed, is_response_to_head
+from .util.retry import Retry
+
+if typing.TYPE_CHECKING:
+    from .connectionpool import HTTPConnectionPool
+
+log = logging.getLogger(__name__)
+
+
+class ContentDecoder:
+    def decompress(self, data: bytes) -> bytes:
+        raise NotImplementedError()
+
+    def flush(self) -> bytes:
+        raise NotImplementedError()
+
+
+class DeflateDecoder(ContentDecoder):
+    def __init__(self) -> None:
+        self._first_try = True
+        self._data = b""
+        self._obj = zlib.decompressobj()
+
+    def decompress(self, data: bytes) -> bytes:
+        if not data:
+            return data
+
+        if not self._first_try:
+            return self._obj.decompress(data)
+
+        self._data += data
+        try:
+            decompressed = self._obj.decompress(data)
+            if decompressed:
+                self._first_try = False
+                self._data = None  # type: ignore[assignment]
+            return decompressed
+        except zlib.error:
+            self._first_try = False
+            self._obj = zlib.decompressobj(-zlib.MAX_WBITS)
+            try:
+                return self.decompress(self._data)
+            finally:
+                self._data = None  # type: ignore[assignment]
+
+    def flush(self) -> bytes:
+        return self._obj.flush()
+
+
+class GzipDecoderState:
+    FIRST_MEMBER = 0
+    OTHER_MEMBERS = 1
+    SWALLOW_DATA = 2
+
+
+class GzipDecoder(ContentDecoder):
+    def __init__(self) -> None:
+        self._obj = zlib.decompressobj(16 + zlib.MAX_WBITS)
+        self._state = GzipDecoderState.FIRST_MEMBER
+
+    def decompress(self, data: bytes) -> bytes:
+        ret = bytearray()
+        if self._state == GzipDecoderState.SWALLOW_DATA or not data:
+            return bytes(ret)
+        while True:
+            try:
+                ret += self._obj.decompress(data)
+            except zlib.error:
+                previous_state = self._state
+                # Ignore data after the first error
+                self._state = GzipDecoderState.SWALLOW_DATA
+                if previous_state == GzipDecoderState.OTHER_MEMBERS:
+                    # Allow trailing garbage acceptable in other gzip clients
+                    return bytes(ret)
+                raise
+            data = self._obj.unused_data
+            if not data:
+                return bytes(ret)
+            self._state = GzipDecoderState.OTHER_MEMBERS
+            self._obj = zlib.decompressobj(16 + zlib.MAX_WBITS)
+
+    def flush(self) -> bytes:
+        return self._obj.flush()
+
+
+if brotli is not None:
+
+    class BrotliDecoder(ContentDecoder):
+        # Supports both 'brotlipy' and 'Brotli' packages
+        # since they share an import name. The top branches
+        # are for 'brotlipy' and bottom branches for 'Brotli'
+        def __init__(self) -> None:
+            self._obj = brotli.Decompressor()
+            if hasattr(self._obj, "decompress"):
+                setattr(self, "decompress", self._obj.decompress)
+            else:
+                setattr(self, "decompress", self._obj.process)
+
+        def flush(self) -> bytes:
+            if hasattr(self._obj, "flush"):
+                return self._obj.flush()  # type: ignore[no-any-return]
+            return b""
+
+
+if HAS_ZSTD:
+
+    class ZstdDecoder(ContentDecoder):
+        def __init__(self) -> None:
+            self._obj = zstd.ZstdDecompressor().decompressobj()
+
+        def decompress(self, data: bytes) -> bytes:
+            if not data:
+                return b""
+            data_parts = [self._obj.decompress(data)]
+            while self._obj.eof and self._obj.unused_data:
+                unused_data = self._obj.unused_data
+                self._obj = zstd.ZstdDecompressor().decompressobj()
+                data_parts.append(self._obj.decompress(unused_data))
+            return b"".join(data_parts)
+
+        def flush(self) -> bytes:
+            ret = self._obj.flush()  # note: this is a no-op
+            if not self._obj.eof:
+                raise DecodeError("Zstandard data is incomplete")
+            return ret
+
+
+class MultiDecoder(ContentDecoder):
+    """
+    From RFC7231:
+        If one or more encodings have been applied to a representation, the
+        sender that applied the encodings MUST generate a Content-Encoding
+        header field that lists the content codings in the order in which
+        they were applied.
+    """
+
+    def __init__(self, modes: str) -> None:
+        self._decoders = [_get_decoder(m.strip()) for m in modes.split(",")]
+
+    def flush(self) -> bytes:
+        return self._decoders[0].flush()
+
+    def decompress(self, data: bytes) -> bytes:
+        for d in reversed(self._decoders):
+            data = d.decompress(data)
+        return data
+
+
+def _get_decoder(mode: str) -> ContentDecoder:
+    if "," in mode:
+        return MultiDecoder(mode)
+
+    # According to RFC 9110 section 8.4.1.3, recipients should
+    # consider x-gzip equivalent to gzip
+    if mode in ("gzip", "x-gzip"):
+        return GzipDecoder()
+
+    if brotli is not None and mode == "br":
+        return BrotliDecoder()
+
+    if HAS_ZSTD and mode == "zstd":
+        return ZstdDecoder()
+
+    return DeflateDecoder()
+
+
+class BytesQueueBuffer:
+    """Memory-efficient bytes buffer
+
+    To return decoded data in read() and still follow the BufferedIOBase API, we need a
+    buffer to always return the correct amount of bytes.
+
+    This buffer should be filled using calls to put()
+
+    Our maximum memory usage is determined by the sum of the size of:
+
+     * self.buffer, which contains the full data
+     * the largest chunk that we will copy in get()
+
+    The worst case scenario is a single chunk, in which case we'll make a full copy of
+    the data inside get().
+    """
+
+    def __init__(self) -> None:
+        self.buffer: typing.Deque[bytes] = collections.deque()
+        self._size: int = 0
+
+    def __len__(self) -> int:
+        return self._size
+
+    def put(self, data: bytes) -> None:
+        self.buffer.append(data)
+        self._size += len(data)
+
+    def get(self, n: int) -> bytes:
+        if n == 0:
+            return b""
+        elif not self.buffer:
+            raise RuntimeError("buffer is empty")
+        elif n < 0:
+            raise ValueError("n should be > 0")
+
+        fetched = 0
+        ret = io.BytesIO()
+        while fetched < n:
+            remaining = n - fetched
+            chunk = self.buffer.popleft()
+            chunk_length = len(chunk)
+            if remaining < chunk_length:
+                left_chunk, right_chunk = chunk[:remaining], chunk[remaining:]
+                ret.write(left_chunk)
+                self.buffer.appendleft(right_chunk)
+                self._size -= remaining
+                break
+            else:
+                ret.write(chunk)
+                self._size -= chunk_length
+            fetched += chunk_length
+
+            if not self.buffer:
+                break
+
+        return ret.getvalue()
+
+    def get_all(self) -> bytes:
+        buffer = self.buffer
+        if not buffer:
+            assert self._size == 0
+            return b""
+        if len(buffer) == 1:
+            result = buffer.pop()
+        else:
+            ret = io.BytesIO()
+            ret.writelines(buffer.popleft() for _ in range(len(buffer)))
+            result = ret.getvalue()
+        self._size = 0
+        return result
+
+
+class BaseHTTPResponse(io.IOBase):
+    CONTENT_DECODERS = ["gzip", "x-gzip", "deflate"]
+    if brotli is not None:
+        CONTENT_DECODERS += ["br"]
+    if HAS_ZSTD:
+        CONTENT_DECODERS += ["zstd"]
+    REDIRECT_STATUSES = [301, 302, 303, 307, 308]
+
+    DECODER_ERROR_CLASSES: tuple[type[Exception], ...] = (IOError, zlib.error)
+    if brotli is not None:
+        DECODER_ERROR_CLASSES += (brotli.error,)
+
+    if HAS_ZSTD:
+        DECODER_ERROR_CLASSES += (zstd.ZstdError,)
+
+    def __init__(
+        self,
+        *,
+        headers: typing.Mapping[str, str] | typing.Mapping[bytes, bytes] | None = None,
+        status: int,
+        version: int,
+        version_string: str,
+        reason: str | None,
+        decode_content: bool,
+        request_url: str | None,
+        retries: Retry | None = None,
+    ) -> None:
+        if isinstance(headers, HTTPHeaderDict):
+            self.headers = headers
+        else:
+            self.headers = HTTPHeaderDict(headers)  # type: ignore[arg-type]
+        self.status = status
+        self.version = version
+        self.version_string = version_string
+        self.reason = reason
+        self.decode_content = decode_content
+        self._has_decoded_content = False
+        self._request_url: str | None = request_url
+        self.retries = retries
+
+        self.chunked = False
+        tr_enc = self.headers.get("transfer-encoding", "").lower()
+        # Don't incur the penalty of creating a list and then discarding it
+        encodings = (enc.strip() for enc in tr_enc.split(","))
+        if "chunked" in encodings:
+            self.chunked = True
+
+        self._decoder: ContentDecoder | None = None
+        self.length_remaining: int | None
+
+    def get_redirect_location(self) -> str | None | typing.Literal[False]:
+        """
+        Should we redirect and where to?
+
+        :returns: Truthy redirect location string if we got a redirect status
+            code and valid location. ``None`` if redirect status and no
+            location. ``False`` if not a redirect status code.
+        """
+        if self.status in self.REDIRECT_STATUSES:
+            return self.headers.get("location")
+        return False
+
+    @property
+    def data(self) -> bytes:
+        raise NotImplementedError()
+
+    def json(self) -> typing.Any:
+        """
+        Deserializes the body of the HTTP response as a Python object.
+
+        The body of the HTTP response must be encoded using UTF-8, as per
+        `RFC 8529 Section 8.1 <https://www.rfc-editor.org/rfc/rfc8259#section-8.1>`_.
+
+        To use a custom JSON decoder pass the result of :attr:`HTTPResponse.data` to
+        your custom decoder instead.
+
+        If the body of the HTTP response is not decodable to UTF-8, a
+        `UnicodeDecodeError` will be raised. If the body of the HTTP response is not a
+        valid JSON document, a `json.JSONDecodeError` will be raised.
+
+        Read more :ref:`here <json_content>`.
+
+        :returns: The body of the HTTP response as a Python object.
+        """
+        data = self.data.decode("utf-8")
+        return _json.loads(data)
+
+    @property
+    def url(self) -> str | None:
+        raise NotImplementedError()
+
+    @url.setter
+    def url(self, url: str | None) -> None:
+        raise NotImplementedError()
+
+    @property
+    def connection(self) -> BaseHTTPConnection | None:
+        raise NotImplementedError()
+
+    @property
+    def retries(self) -> Retry | None:
+        return self._retries
+
+    @retries.setter
+    def retries(self, retries: Retry | None) -> None:
+        # Override the request_url if retries has a redirect location.
+        if retries is not None and retries.history:
+            self.url = retries.history[-1].redirect_location
+        self._retries = retries
+
+    def stream(
+        self, amt: int | None = 2**16, decode_content: bool | None = None
+    ) -> typing.Iterator[bytes]:
+        raise NotImplementedError()
+
+    def read(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+        cache_content: bool = False,
+    ) -> bytes:
+        raise NotImplementedError()
+
+    def read1(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+    ) -> bytes:
+        raise NotImplementedError()
+
+    def read_chunked(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+    ) -> typing.Iterator[bytes]:
+        raise NotImplementedError()
+
+    def release_conn(self) -> None:
+        raise NotImplementedError()
+
+    def drain_conn(self) -> None:
+        raise NotImplementedError()
+
+    def shutdown(self) -> None:
+        raise NotImplementedError()
+
+    def close(self) -> None:
+        raise NotImplementedError()
+
+    def _init_decoder(self) -> None:
+        """
+        Set-up the _decoder attribute if necessary.
+        """
+        # Note: content-encoding value should be case-insensitive, per RFC 7230
+        # Section 3.2
+        content_encoding = self.headers.get("content-encoding", "").lower()
+        if self._decoder is None:
+            if content_encoding in self.CONTENT_DECODERS:
+                self._decoder = _get_decoder(content_encoding)
+            elif "," in content_encoding:
+                encodings = [
+                    e.strip()
+                    for e in content_encoding.split(",")
+                    if e.strip() in self.CONTENT_DECODERS
+                ]
+                if encodings:
+                    self._decoder = _get_decoder(content_encoding)
+
+    def _decode(
+        self, data: bytes, decode_content: bool | None, flush_decoder: bool
+    ) -> bytes:
+        """
+        Decode the data passed in and potentially flush the decoder.
+        """
+        if not decode_content:
+            if self._has_decoded_content:
+                raise RuntimeError(
+                    "Calling read(decode_content=False) is not supported after "
+                    "read(decode_content=True) was called."
+                )
+            return data
+
+        try:
+            if self._decoder:
+                data = self._decoder.decompress(data)
+                self._has_decoded_content = True
+        except self.DECODER_ERROR_CLASSES as e:
+            content_encoding = self.headers.get("content-encoding", "").lower()
+            raise DecodeError(
+                "Received response with content-encoding: %s, but "
+                "failed to decode it." % content_encoding,
+                e,
+            ) from e
+        if flush_decoder:
+            data += self._flush_decoder()
+
+        return data
+
+    def _flush_decoder(self) -> bytes:
+        """
+        Flushes the decoder. Should only be called if the decoder is actually
+        being used.
+        """
+        if self._decoder:
+            return self._decoder.decompress(b"") + self._decoder.flush()
+        return b""
+
+    # Compatibility methods for `io` module
+    def readinto(self, b: bytearray) -> int:
+        temp = self.read(len(b))
+        if len(temp) == 0:
+            return 0
+        else:
+            b[: len(temp)] = temp
+            return len(temp)
+
+    # Compatibility methods for http.client.HTTPResponse
+    def getheaders(self) -> HTTPHeaderDict:
+        warnings.warn(
+            "HTTPResponse.getheaders() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead access HTTPResponse.headers directly.",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.headers
+
+    def getheader(self, name: str, default: str | None = None) -> str | None:
+        warnings.warn(
+            "HTTPResponse.getheader() is deprecated and will be removed "
+            "in urllib3 v2.1.0. Instead use HTTPResponse.headers.get(name, default).",
+            category=DeprecationWarning,
+            stacklevel=2,
+        )
+        return self.headers.get(name, default)
+
+    # Compatibility method for http.cookiejar
+    def info(self) -> HTTPHeaderDict:
+        return self.headers
+
+    def geturl(self) -> str | None:
+        return self.url
+
+
+class HTTPResponse(BaseHTTPResponse):
+    """
+    HTTP Response container.
+
+    Backwards-compatible with :class:`http.client.HTTPResponse` but the response ``body`` is
+    loaded and decoded on-demand when the ``data`` property is accessed.  This
+    class is also compatible with the Python standard library's :mod:`io`
+    module, and can hence be treated as a readable object in the context of that
+    framework.
+
+    Extra parameters for behaviour not present in :class:`http.client.HTTPResponse`:
+
+    :param preload_content:
+        If True, the response's body will be preloaded during construction.
+
+    :param decode_content:
+        If True, will attempt to decode the body based on the
+        'content-encoding' header.
+
+    :param original_response:
+        When this HTTPResponse wrapper is generated from an :class:`http.client.HTTPResponse`
+        object, it's convenient to include the original for debug purposes. It's
+        otherwise unused.
+
+    :param retries:
+        The retries contains the last :class:`~urllib3.util.retry.Retry` that
+        was used during the request.
+
+    :param enforce_content_length:
+        Enforce content length checking. Body returned by server must match
+        value of Content-Length header, if present. Otherwise, raise error.
+    """
+
+    def __init__(
+        self,
+        body: _TYPE_BODY = "",
+        headers: typing.Mapping[str, str] | typing.Mapping[bytes, bytes] | None = None,
+        status: int = 0,
+        version: int = 0,
+        version_string: str = "HTTP/?",
+        reason: str | None = None,
+        preload_content: bool = True,
+        decode_content: bool = True,
+        original_response: _HttplibHTTPResponse | None = None,
+        pool: HTTPConnectionPool | None = None,
+        connection: HTTPConnection | None = None,
+        msg: _HttplibHTTPMessage | None = None,
+        retries: Retry | None = None,
+        enforce_content_length: bool = True,
+        request_method: str | None = None,
+        request_url: str | None = None,
+        auto_close: bool = True,
+        sock_shutdown: typing.Callable[[int], None] | None = None,
+    ) -> None:
+        super().__init__(
+            headers=headers,
+            status=status,
+            version=version,
+            version_string=version_string,
+            reason=reason,
+            decode_content=decode_content,
+            request_url=request_url,
+            retries=retries,
+        )
+
+        self.enforce_content_length = enforce_content_length
+        self.auto_close = auto_close
+
+        self._body = None
+        self._fp: _HttplibHTTPResponse | None = None
+        self._original_response = original_response
+        self._fp_bytes_read = 0
+        self.msg = msg
+
+        if body and isinstance(body, (str, bytes)):
+            self._body = body
+
+        self._pool = pool
+        self._connection = connection
+
+        if hasattr(body, "read"):
+            self._fp = body  # type: ignore[assignment]
+        self._sock_shutdown = sock_shutdown
+
+        # Are we using the chunked-style of transfer encoding?
+        self.chunk_left: int | None = None
+
+        # Determine length of response
+        self.length_remaining = self._init_length(request_method)
+
+        # Used to return the correct amount of bytes for partial read()s
+        self._decoded_buffer = BytesQueueBuffer()
+
+        # If requested, preload the body.
+        if preload_content and not self._body:
+            self._body = self.read(decode_content=decode_content)
+
+    def release_conn(self) -> None:
+        if not self._pool or not self._connection:
+            return None
+
+        self._pool._put_conn(self._connection)
+        self._connection = None
+
+    def drain_conn(self) -> None:
+        """
+        Read and discard any remaining HTTP response data in the response connection.
+
+        Unread data in the HTTPResponse connection blocks the connection from being released back to the pool.
+        """
+        try:
+            self.read()
+        except (HTTPError, OSError, BaseSSLError, HTTPException):
+            pass
+
+    @property
+    def data(self) -> bytes:
+        # For backwards-compat with earlier urllib3 0.4 and earlier.
+        if self._body:
+            return self._body  # type: ignore[return-value]
+
+        if self._fp:
+            return self.read(cache_content=True)
+
+        return None  # type: ignore[return-value]
+
+    @property
+    def connection(self) -> HTTPConnection | None:
+        return self._connection
+
+    def isclosed(self) -> bool:
+        return is_fp_closed(self._fp)
+
+    def tell(self) -> int:
+        """
+        Obtain the number of bytes pulled over the wire so far. May differ from
+        the amount of content returned by :meth:``urllib3.response.HTTPResponse.read``
+        if bytes are encoded on the wire (e.g, compressed).
+        """
+        return self._fp_bytes_read
+
+    def _init_length(self, request_method: str | None) -> int | None:
+        """
+        Set initial length value for Response content if available.
+        """
+        length: int | None
+        content_length: str | None = self.headers.get("content-length")
+
+        if content_length is not None:
+            if self.chunked:
+                # This Response will fail with an IncompleteRead if it can't be
+                # received as chunked. This method falls back to attempt reading
+                # the response before raising an exception.
+                log.warning(
+                    "Received response with both Content-Length and "
+                    "Transfer-Encoding set. This is expressly forbidden "
+                    "by RFC 7230 sec 3.3.2. Ignoring Content-Length and "
+                    "attempting to process response as Transfer-Encoding: "
+                    "chunked."
+                )
+                return None
+
+            try:
+                # RFC 7230 section 3.3.2 specifies multiple content lengths can
+                # be sent in a single Content-Length header
+                # (e.g. Content-Length: 42, 42). This line ensures the values
+                # are all valid ints and that as long as the `set` length is 1,
+                # all values are the same. Otherwise, the header is invalid.
+                lengths = {int(val) for val in content_length.split(",")}
+                if len(lengths) > 1:
+                    raise InvalidHeader(
+                        "Content-Length contained multiple "
+                        "unmatching values (%s)" % content_length
+                    )
+                length = lengths.pop()
+            except ValueError:
+                length = None
+            else:
+                if length < 0:
+                    length = None
+
+        else:  # if content_length is None
+            length = None
+
+        # Convert status to int for comparison
+        # In some cases, httplib returns a status of "_UNKNOWN"
+        try:
+            status = int(self.status)
+        except ValueError:
+            status = 0
+
+        # Check for responses that shouldn't include a body
+        if status in (204, 304) or 100 <= status < 200 or request_method == "HEAD":
+            length = 0
+
+        return length
+
+    @contextmanager
+    def _error_catcher(self) -> typing.Generator[None]:
+        """
+        Catch low-level python exceptions, instead re-raising urllib3
+        variants, so that low-level exceptions are not leaked in the
+        high-level api.
+
+        On exit, release the connection back to the pool.
+        """
+        clean_exit = False
+
+        try:
+            try:
+                yield
+
+            except SocketTimeout as e:
+                # FIXME: Ideally we'd like to include the url in the ReadTimeoutError but
+                # there is yet no clean way to get at it from this context.
+                raise ReadTimeoutError(self._pool, None, "Read timed out.") from e  # type: ignore[arg-type]
+
+            except BaseSSLError as e:
+                # FIXME: Is there a better way to differentiate between SSLErrors?
+                if "read operation timed out" not in str(e):
+                    # SSL errors related to framing/MAC get wrapped and reraised here
+                    raise SSLError(e) from e
+
+                raise ReadTimeoutError(self._pool, None, "Read timed out.") from e  # type: ignore[arg-type]
+
+            except IncompleteRead as e:
+                if (
+                    e.expected is not None
+                    and e.partial is not None
+                    and e.expected == -e.partial
+                ):
+                    arg = "Response may not contain content."
+                else:
+                    arg = f"Connection broken: {e!r}"
+                raise ProtocolError(arg, e) from e
+
+            except (HTTPException, OSError) as e:
+                raise ProtocolError(f"Connection broken: {e!r}", e) from e
+
+            # If no exception is thrown, we should avoid cleaning up
+            # unnecessarily.
+            clean_exit = True
+        finally:
+            # If we didn't terminate cleanly, we need to throw away our
+            # connection.
+            if not clean_exit:
+                # The response may not be closed but we're not going to use it
+                # anymore so close it now to ensure that the connection is
+                # released back to the pool.
+                if self._original_response:
+                    self._original_response.close()
+
+                # Closing the response may not actually be sufficient to close
+                # everything, so if we have a hold of the connection close that
+                # too.
+                if self._connection:
+                    self._connection.close()
+
+            # If we hold the original response but it's closed now, we should
+            # return the connection back to the pool.
+            if self._original_response and self._original_response.isclosed():
+                self.release_conn()
+
+    def _fp_read(
+        self,
+        amt: int | None = None,
+        *,
+        read1: bool = False,
+    ) -> bytes:
+        """
+        Read a response with the thought that reading the number of bytes
+        larger than can fit in a 32-bit int at a time via SSL in some
+        known cases leads to an overflow error that has to be prevented
+        if `amt` or `self.length_remaining` indicate that a problem may
+        happen.
+
+        The known cases:
+          * CPython < 3.9.7 because of a bug
+            https://github.com/urllib3/urllib3/issues/2513#issuecomment-1152559900.
+          * urllib3 injected with pyOpenSSL-backed SSL-support.
+          * CPython < 3.10 only when `amt` does not fit 32-bit int.
+        """
+        assert self._fp
+        c_int_max = 2**31 - 1
+        if (
+            (amt and amt > c_int_max)
+            or (
+                amt is None
+                and self.length_remaining
+                and self.length_remaining > c_int_max
+            )
+        ) and (util.IS_PYOPENSSL or sys.version_info < (3, 10)):
+            if read1:
+                return self._fp.read1(c_int_max)
+            buffer = io.BytesIO()
+            # Besides `max_chunk_amt` being a maximum chunk size, it
+            # affects memory overhead of reading a response by this
+            # method in CPython.
+            # `c_int_max` equal to 2 GiB - 1 byte is the actual maximum
+            # chunk size that does not lead to an overflow error, but
+            # 256 MiB is a compromise.
+            max_chunk_amt = 2**28
+            while amt is None or amt != 0:
+                if amt is not None:
+                    chunk_amt = min(amt, max_chunk_amt)
+                    amt -= chunk_amt
+                else:
+                    chunk_amt = max_chunk_amt
+                data = self._fp.read(chunk_amt)
+                if not data:
+                    break
+                buffer.write(data)
+                del data  # to reduce peak memory usage by `max_chunk_amt`.
+            return buffer.getvalue()
+        elif read1:
+            return self._fp.read1(amt) if amt is not None else self._fp.read1()
+        else:
+            # StringIO doesn't like amt=None
+            return self._fp.read(amt) if amt is not None else self._fp.read()
+
+    def _raw_read(
+        self,
+        amt: int | None = None,
+        *,
+        read1: bool = False,
+    ) -> bytes:
+        """
+        Reads `amt` of bytes from the socket.
+        """
+        if self._fp is None:
+            return None  # type: ignore[return-value]
+
+        fp_closed = getattr(self._fp, "closed", False)
+
+        with self._error_catcher():
+            data = self._fp_read(amt, read1=read1) if not fp_closed else b""
+            if amt is not None and amt != 0 and not data:
+                # Platform-specific: Buggy versions of Python.
+                # Close the connection when no data is returned
+                #
+                # This is redundant to what httplib/http.client _should_
+                # already do.  However, versions of python released before
+                # December 15, 2012 (http://bugs.python.org/issue16298) do
+                # not properly close the connection in all cases. There is
+                # no harm in redundantly calling close.
+                self._fp.close()
+                if (
+                    self.enforce_content_length
+                    and self.length_remaining is not None
+                    and self.length_remaining != 0
+                ):
+                    # This is an edge case that httplib failed to cover due
+                    # to concerns of backward compatibility. We're
+                    # addressing it here to make sure IncompleteRead is
+                    # raised during streaming, so all calls with incorrect
+                    # Content-Length are caught.
+                    raise IncompleteRead(self._fp_bytes_read, self.length_remaining)
+            elif read1 and (
+                (amt != 0 and not data) or self.length_remaining == len(data)
+            ):
+                # All data has been read, but `self._fp.read1` in
+                # CPython 3.12 and older doesn't always close
+                # `http.client.HTTPResponse`, so we close it here.
+                # See https://github.com/python/cpython/issues/113199
+                self._fp.close()
+
+        if data:
+            self._fp_bytes_read += len(data)
+            if self.length_remaining is not None:
+                self.length_remaining -= len(data)
+        return data
+
+    def read(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+        cache_content: bool = False,
+    ) -> bytes:
+        """
+        Similar to :meth:`http.client.HTTPResponse.read`, but with two additional
+        parameters: ``decode_content`` and ``cache_content``.
+
+        :param amt:
+            How much of the content to read. If specified, caching is skipped
+            because it doesn't make sense to cache partial content as the full
+            response.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+
+        :param cache_content:
+            If True, will save the returned data such that the same result is
+            returned despite of the state of the underlying file object. This
+            is useful if you want the ``.data`` property to continue working
+            after having ``.read()`` the file object. (Overridden if ``amt`` is
+            set.)
+        """
+        self._init_decoder()
+        if decode_content is None:
+            decode_content = self.decode_content
+
+        if amt and amt < 0:
+            # Negative numbers and `None` should be treated the same.
+            amt = None
+        elif amt is not None:
+            cache_content = False
+
+            if len(self._decoded_buffer) >= amt:
+                return self._decoded_buffer.get(amt)
+
+        data = self._raw_read(amt)
+
+        flush_decoder = amt is None or (amt != 0 and not data)
+
+        if not data and len(self._decoded_buffer) == 0:
+            return data
+
+        if amt is None:
+            data = self._decode(data, decode_content, flush_decoder)
+            if cache_content:
+                self._body = data
+        else:
+            # do not waste memory on buffer when not decoding
+            if not decode_content:
+                if self._has_decoded_content:
+                    raise RuntimeError(
+                        "Calling read(decode_content=False) is not supported after "
+                        "read(decode_content=True) was called."
+                    )
+                return data
+
+            decoded_data = self._decode(data, decode_content, flush_decoder)
+            self._decoded_buffer.put(decoded_data)
+
+            while len(self._decoded_buffer) < amt and data:
+                # TODO make sure to initially read enough data to get past the headers
+                # For example, the GZ file header takes 10 bytes, we don't want to read
+                # it one byte at a time
+                data = self._raw_read(amt)
+                decoded_data = self._decode(data, decode_content, flush_decoder)
+                self._decoded_buffer.put(decoded_data)
+            data = self._decoded_buffer.get(amt)
+
+        return data
+
+    def read1(
+        self,
+        amt: int | None = None,
+        decode_content: bool | None = None,
+    ) -> bytes:
+        """
+        Similar to ``http.client.HTTPResponse.read1`` and documented
+        in :meth:`io.BufferedReader.read1`, but with an additional parameter:
+        ``decode_content``.
+
+        :param amt:
+            How much of the content to read.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        if decode_content is None:
+            decode_content = self.decode_content
+        if amt and amt < 0:
+            # Negative numbers and `None` should be treated the same.
+            amt = None
+        # try and respond without going to the network
+        if self._has_decoded_content:
+            if not decode_content:
+                raise RuntimeError(
+                    "Calling read1(decode_content=False) is not supported after "
+                    "read1(decode_content=True) was called."
+                )
+            if len(self._decoded_buffer) > 0:
+                if amt is None:
+                    return self._decoded_buffer.get_all()
+                return self._decoded_buffer.get(amt)
+        if amt == 0:
+            return b""
+
+        # FIXME, this method's type doesn't say returning None is possible
+        data = self._raw_read(amt, read1=True)
+        if not decode_content or data is None:
+            return data
+
+        self._init_decoder()
+        while True:
+            flush_decoder = not data
+            decoded_data = self._decode(data, decode_content, flush_decoder)
+            self._decoded_buffer.put(decoded_data)
+            if decoded_data or flush_decoder:
+                break
+            data = self._raw_read(8192, read1=True)
+
+        if amt is None:
+            return self._decoded_buffer.get_all()
+        return self._decoded_buffer.get(amt)
+
+    def stream(
+        self, amt: int | None = 2**16, decode_content: bool | None = None
+    ) -> typing.Generator[bytes]:
+        """
+        A generator wrapper for the read() method. A call will block until
+        ``amt`` bytes have been read from the connection or until the
+        connection is closed.
+
+        :param amt:
+            How much of the content to read. The generator will return up to
+            much data per iteration, but may return less. This is particularly
+            likely when using compressed data. However, the empty string will
+            never be returned.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        if self.chunked and self.supports_chunked_reads():
+            yield from self.read_chunked(amt, decode_content=decode_content)
+        else:
+            while not is_fp_closed(self._fp) or len(self._decoded_buffer) > 0:
+                data = self.read(amt=amt, decode_content=decode_content)
+
+                if data:
+                    yield data
+
+    # Overrides from io.IOBase
+    def readable(self) -> bool:
+        return True
+
+    def shutdown(self) -> None:
+        if not self._sock_shutdown:
+            raise ValueError("Cannot shutdown socket as self._sock_shutdown is not set")
+        self._sock_shutdown(socket.SHUT_RD)
+
+    def close(self) -> None:
+        self._sock_shutdown = None
+
+        if not self.closed and self._fp:
+            self._fp.close()
+
+        if self._connection:
+            self._connection.close()
+
+        if not self.auto_close:
+            io.IOBase.close(self)
+
+    @property
+    def closed(self) -> bool:
+        if not self.auto_close:
+            return io.IOBase.closed.__get__(self)  # type: ignore[no-any-return]
+        elif self._fp is None:
+            return True
+        elif hasattr(self._fp, "isclosed"):
+            return self._fp.isclosed()
+        elif hasattr(self._fp, "closed"):
+            return self._fp.closed
+        else:
+            return True
+
+    def fileno(self) -> int:
+        if self._fp is None:
+            raise OSError("HTTPResponse has no file to get a fileno from")
+        elif hasattr(self._fp, "fileno"):
+            return self._fp.fileno()
+        else:
+            raise OSError(
+                "The file-like object this HTTPResponse is wrapped "
+                "around has no file descriptor"
+            )
+
+    def flush(self) -> None:
+        if (
+            self._fp is not None
+            and hasattr(self._fp, "flush")
+            and not getattr(self._fp, "closed", False)
+        ):
+            return self._fp.flush()
+
+    def supports_chunked_reads(self) -> bool:
+        """
+        Checks if the underlying file-like object looks like a
+        :class:`http.client.HTTPResponse` object. We do this by testing for
+        the fp attribute. If it is present we assume it returns raw chunks as
+        processed by read_chunked().
+        """
+        return hasattr(self._fp, "fp")
+
+    def _update_chunk_length(self) -> None:
+        # First, we'll figure out length of a chunk and then
+        # we'll try to read it from socket.
+        if self.chunk_left is not None:
+            return None
+        line = self._fp.fp.readline()  # type: ignore[union-attr]
+        line = line.split(b";", 1)[0]
+        try:
+            self.chunk_left = int(line, 16)
+        except ValueError:
+            self.close()
+            if line:
+                # Invalid chunked protocol response, abort.
+                raise InvalidChunkLength(self, line) from None
+            else:
+                # Truncated at start of next chunk
+                raise ProtocolError("Response ended prematurely") from None
+
+    def _handle_chunk(self, amt: int | None) -> bytes:
+        returned_chunk = None
+        if amt is None:
+            chunk = self._fp._safe_read(self.chunk_left)  # type: ignore[union-attr]
+            returned_chunk = chunk
+            self._fp._safe_read(2)  # type: ignore[union-attr] # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+        elif self.chunk_left is not None and amt < self.chunk_left:
+            value = self._fp._safe_read(amt)  # type: ignore[union-attr]
+            self.chunk_left = self.chunk_left - amt
+            returned_chunk = value
+        elif amt == self.chunk_left:
+            value = self._fp._safe_read(amt)  # type: ignore[union-attr]
+            self._fp._safe_read(2)  # type: ignore[union-attr] # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+            returned_chunk = value
+        else:  # amt > self.chunk_left
+            returned_chunk = self._fp._safe_read(self.chunk_left)  # type: ignore[union-attr]
+            self._fp._safe_read(2)  # type: ignore[union-attr] # Toss the CRLF at the end of the chunk.
+            self.chunk_left = None
+        return returned_chunk  # type: ignore[no-any-return]
+
+    def read_chunked(
+        self, amt: int | None = None, decode_content: bool | None = None
+    ) -> typing.Generator[bytes]:
+        """
+        Similar to :meth:`HTTPResponse.read`, but with an additional
+        parameter: ``decode_content``.
+
+        :param amt:
+            How much of the content to read. If specified, caching is skipped
+            because it doesn't make sense to cache partial content as the full
+            response.
+
+        :param decode_content:
+            If True, will attempt to decode the body based on the
+            'content-encoding' header.
+        """
+        self._init_decoder()
+        # FIXME: Rewrite this method and make it a class with a better structured logic.
+        if not self.chunked:
+            raise ResponseNotChunked(
+                "Response is not chunked. "
+                "Header 'transfer-encoding: chunked' is missing."
+            )
+        if not self.supports_chunked_reads():
+            raise BodyNotHttplibCompatible(
+                "Body should be http.client.HTTPResponse like. "
+                "It should have have an fp attribute which returns raw chunks."
+            )
+
+        with self._error_catcher():
+            # Don't bother reading the body of a HEAD request.
+            if self._original_response and is_response_to_head(self._original_response):
+                self._original_response.close()
+                return None
+
+            # If a response is already read and closed
+            # then return immediately.
+            if self._fp.fp is None:  # type: ignore[union-attr]
+                return None
+
+            if amt and amt < 0:
+                # Negative numbers and `None` should be treated the same,
+                # but httplib handles only `None` correctly.
+                amt = None
+
+            while True:
+                self._update_chunk_length()
+                if self.chunk_left == 0:
+                    break
+                chunk = self._handle_chunk(amt)
+                decoded = self._decode(
+                    chunk, decode_content=decode_content, flush_decoder=False
+                )
+                if decoded:
+                    yield decoded
+
+            if decode_content:
+                # On CPython and PyPy, we should never need to flush the
+                # decoder. However, on Jython we *might* need to, so
+                # lets defensively do it anyway.
+                decoded = self._flush_decoder()
+                if decoded:  # Platform-specific: Jython.
+                    yield decoded
+
+            # Chunk content ends with \r\n: discard it.
+            while self._fp is not None:
+                line = self._fp.fp.readline()
+                if not line:
+                    # Some sites may not end with '\r\n'.
+                    break
+                if line == b"\r\n":
+                    break
+
+            # We read everything; close the "file".
+            if self._original_response:
+                self._original_response.close()
+
+    @property
+    def url(self) -> str | None:
+        """
+        Returns the URL that was the source of this response.
+        If the request that generated this response redirected, this method
+        will return the final redirect location.
+        """
+        return self._request_url
+
+    @url.setter
+    def url(self, url: str) -> None:
+        self._request_url = url
+
+    def __iter__(self) -> typing.Iterator[bytes]:
+        buffer: list[bytes] = []
+        for chunk in self.stream(decode_content=True):
+            if b"\n" in chunk:
+                chunks = chunk.split(b"\n")
+                yield b"".join(buffer) + chunks[0] + b"\n"
+                for x in chunks[1:-1]:
+                    yield x + b"\n"
+                if chunks[-1]:
+                    buffer = [chunks[-1]]
+                else:
+                    buffer = []
+            else:
+                buffer.append(chunk)
+        if buffer:
+            yield b"".join(buffer)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__init__.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/__init__.py
new file mode 100644
index 00000000..53412603
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/__init__.py
@@ -0,0 +1,42 @@
+# For backwards compatibility, provide imports that used to be here.
+from __future__ import annotations
+
+from .connection import is_connection_dropped
+from .request import SKIP_HEADER, SKIPPABLE_HEADERS, make_headers
+from .response import is_fp_closed
+from .retry import Retry
+from .ssl_ import (
+    ALPN_PROTOCOLS,
+    IS_PYOPENSSL,
+    SSLContext,
+    assert_fingerprint,
+    create_urllib3_context,
+    resolve_cert_reqs,
+    resolve_ssl_version,
+    ssl_wrap_socket,
+)
+from .timeout import Timeout
+from .url import Url, parse_url
+from .wait import wait_for_read, wait_for_write
+
+__all__ = (
+    "IS_PYOPENSSL",
+    "SSLContext",
+    "ALPN_PROTOCOLS",
+    "Retry",
+    "Timeout",
+    "Url",
+    "assert_fingerprint",
+    "create_urllib3_context",
+    "is_connection_dropped",
+    "is_fp_closed",
+    "parse_url",
+    "make_headers",
+    "resolve_cert_reqs",
+    "resolve_ssl_version",
+    "ssl_wrap_socket",
+    "wait_for_read",
+    "wait_for_write",
+    "SKIP_HEADER",
+    "SKIPPABLE_HEADERS",
+)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..878c5925
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/connection.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/connection.cpython-312.pyc
new file mode 100644
index 00000000..93468980
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/connection.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/proxy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/proxy.cpython-312.pyc
new file mode 100644
index 00000000..36f933e4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/proxy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..822fedfa
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..17110a88
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/retry.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/retry.cpython-312.pyc
new file mode 100644
index 00000000..071ece0f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/retry.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_.cpython-312.pyc
new file mode 100644
index 00000000..4517f4d6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc
new file mode 100644
index 00000000..a09a81a5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssl_match_hostname.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssltransport.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssltransport.cpython-312.pyc
new file mode 100644
index 00000000..a82e7ecd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/ssltransport.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/timeout.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/timeout.cpython-312.pyc
new file mode 100644
index 00000000..2d7ad6f3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/timeout.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/url.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/url.cpython-312.pyc
new file mode 100644
index 00000000..89a71ab5
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/url.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/util.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/util.cpython-312.pyc
new file mode 100644
index 00000000..9fd0d5e9
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/util.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/wait.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/wait.cpython-312.pyc
new file mode 100644
index 00000000..6bcaff34
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/urllib3/util/__pycache__/wait.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/connection.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/connection.py
new file mode 100644
index 00000000..f92519ee
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/connection.py
@@ -0,0 +1,137 @@
+from __future__ import annotations
+
+import socket
+import typing
+
+from ..exceptions import LocationParseError
+from .timeout import _DEFAULT_TIMEOUT, _TYPE_TIMEOUT
+
+_TYPE_SOCKET_OPTIONS = list[tuple[int, int, typing.Union[int, bytes]]]
+
+if typing.TYPE_CHECKING:
+    from .._base_connection import BaseHTTPConnection
+
+
+def is_connection_dropped(conn: BaseHTTPConnection) -> bool:  # Platform-specific
+    """
+    Returns True if the connection is dropped and should be closed.
+    :param conn: :class:`urllib3.connection.HTTPConnection` object.
+    """
+    return not conn.is_connected
+
+
+# This function is copied from socket.py in the Python 2.7 standard
+# library test suite. Added to its signature is only `socket_options`.
+# One additional modification is that we avoid binding to IPv6 servers
+# discovered in DNS if the system doesn't have IPv6 functionality.
+def create_connection(
+    address: tuple[str, int],
+    timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+    source_address: tuple[str, int] | None = None,
+    socket_options: _TYPE_SOCKET_OPTIONS | None = None,
+) -> socket.socket:
+    """Connect to *address* and return the socket object.
+
+    Convenience function.  Connect to *address* (a 2-tuple ``(host,
+    port)``) and return the socket object.  Passing the optional
+    *timeout* parameter will set the timeout on the socket instance
+    before attempting to connect.  If no *timeout* is supplied, the
+    global default timeout setting returned by :func:`socket.getdefaulttimeout`
+    is used.  If *source_address* is set it must be a tuple of (host, port)
+    for the socket to bind as a source address before making the connection.
+    An host of '' or port 0 tells the OS to use the default.
+    """
+
+    host, port = address
+    if host.startswith("["):
+        host = host.strip("[]")
+    err = None
+
+    # Using the value from allowed_gai_family() in the context of getaddrinfo lets
+    # us select whether to work with IPv4 DNS records, IPv6 records, or both.
+    # The original create_connection function always returns all records.
+    family = allowed_gai_family()
+
+    try:
+        host.encode("idna")
+    except UnicodeError:
+        raise LocationParseError(f"'{host}', label empty or too long") from None
+
+    for res in socket.getaddrinfo(host, port, family, socket.SOCK_STREAM):
+        af, socktype, proto, canonname, sa = res
+        sock = None
+        try:
+            sock = socket.socket(af, socktype, proto)
+
+            # If provided, set socket level options before connecting.
+            _set_socket_options(sock, socket_options)
+
+            if timeout is not _DEFAULT_TIMEOUT:
+                sock.settimeout(timeout)
+            if source_address:
+                sock.bind(source_address)
+            sock.connect(sa)
+            # Break explicitly a reference cycle
+            err = None
+            return sock
+
+        except OSError as _:
+            err = _
+            if sock is not None:
+                sock.close()
+
+    if err is not None:
+        try:
+            raise err
+        finally:
+            # Break explicitly a reference cycle
+            err = None
+    else:
+        raise OSError("getaddrinfo returns an empty list")
+
+
+def _set_socket_options(
+    sock: socket.socket, options: _TYPE_SOCKET_OPTIONS | None
+) -> None:
+    if options is None:
+        return
+
+    for opt in options:
+        sock.setsockopt(*opt)
+
+
+def allowed_gai_family() -> socket.AddressFamily:
+    """This function is designed to work in the context of
+    getaddrinfo, where family=socket.AF_UNSPEC is the default and
+    will perform a DNS search for both IPv6 and IPv4 records."""
+
+    family = socket.AF_INET
+    if HAS_IPV6:
+        family = socket.AF_UNSPEC
+    return family
+
+
+def _has_ipv6(host: str) -> bool:
+    """Returns True if the system can bind an IPv6 address."""
+    sock = None
+    has_ipv6 = False
+
+    if socket.has_ipv6:
+        # has_ipv6 returns true if cPython was compiled with IPv6 support.
+        # It does not tell us if the system has IPv6 support enabled. To
+        # determine that we must bind to an IPv6 address.
+        # https://github.com/urllib3/urllib3/pull/611
+        # https://bugs.python.org/issue658327
+        try:
+            sock = socket.socket(socket.AF_INET6)
+            sock.bind((host, 0))
+            has_ipv6 = True
+        except Exception:
+            pass
+
+    if sock:
+        sock.close()
+    return has_ipv6
+
+
+HAS_IPV6 = _has_ipv6("::1")
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/proxy.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/proxy.py
new file mode 100644
index 00000000..908fc662
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/proxy.py
@@ -0,0 +1,43 @@
+from __future__ import annotations
+
+import typing
+
+from .url import Url
+
+if typing.TYPE_CHECKING:
+    from ..connection import ProxyConfig
+
+
+def connection_requires_http_tunnel(
+    proxy_url: Url | None = None,
+    proxy_config: ProxyConfig | None = None,
+    destination_scheme: str | None = None,
+) -> bool:
+    """
+    Returns True if the connection requires an HTTP CONNECT through the proxy.
+
+    :param URL proxy_url:
+        URL of the proxy.
+    :param ProxyConfig proxy_config:
+        Proxy configuration from poolmanager.py
+    :param str destination_scheme:
+        The scheme of the destination. (i.e https, http, etc)
+    """
+    # If we're not using a proxy, no way to use a tunnel.
+    if proxy_url is None:
+        return False
+
+    # HTTP destinations never require tunneling, we always forward.
+    if destination_scheme == "http":
+        return False
+
+    # Support for forwarding with HTTPS proxies and HTTPS destinations.
+    if (
+        proxy_url.scheme == "https"
+        and proxy_config
+        and proxy_config.use_forwarding_for_https
+    ):
+        return False
+
+    # Otherwise always use a tunnel.
+    return True
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/request.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/request.py
new file mode 100644
index 00000000..94392a13
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/request.py
@@ -0,0 +1,258 @@
+from __future__ import annotations
+
+import io
+import typing
+from base64 import b64encode
+from enum import Enum
+
+from ..exceptions import UnrewindableBodyError
+from .util import to_bytes
+
+if typing.TYPE_CHECKING:
+    from typing import Final
+
+# Pass as a value within ``headers`` to skip
+# emitting some HTTP headers that are added automatically.
+# The only headers that are supported are ``Accept-Encoding``,
+# ``Host``, and ``User-Agent``.
+SKIP_HEADER = "@@@SKIP_HEADER@@@"
+SKIPPABLE_HEADERS = frozenset(["accept-encoding", "host", "user-agent"])
+
+ACCEPT_ENCODING = "gzip,deflate"
+try:
+    try:
+        import brotlicffi as _unused_module_brotli  # type: ignore[import-not-found] # noqa: F401
+    except ImportError:
+        import brotli as _unused_module_brotli  # type: ignore[import-not-found] # noqa: F401
+except ImportError:
+    pass
+else:
+    ACCEPT_ENCODING += ",br"
+try:
+    import zstandard as _unused_module_zstd  # noqa: F401
+except ImportError:
+    pass
+else:
+    ACCEPT_ENCODING += ",zstd"
+
+
+class _TYPE_FAILEDTELL(Enum):
+    token = 0
+
+
+_FAILEDTELL: Final[_TYPE_FAILEDTELL] = _TYPE_FAILEDTELL.token
+
+_TYPE_BODY_POSITION = typing.Union[int, _TYPE_FAILEDTELL]
+
+# When sending a request with these methods we aren't expecting
+# a body so don't need to set an explicit 'Content-Length: 0'
+# The reason we do this in the negative instead of tracking methods
+# which 'should' have a body is because unknown methods should be
+# treated as if they were 'POST' which *does* expect a body.
+_METHODS_NOT_EXPECTING_BODY = {"GET", "HEAD", "DELETE", "TRACE", "OPTIONS", "CONNECT"}
+
+
+def make_headers(
+    keep_alive: bool | None = None,
+    accept_encoding: bool | list[str] | str | None = None,
+    user_agent: str | None = None,
+    basic_auth: str | None = None,
+    proxy_basic_auth: str | None = None,
+    disable_cache: bool | None = None,
+) -> dict[str, str]:
+    """
+    Shortcuts for generating request headers.
+
+    :param keep_alive:
+        If ``True``, adds 'connection: keep-alive' header.
+
+    :param accept_encoding:
+        Can be a boolean, list, or string.
+        ``True`` translates to 'gzip,deflate'.  If the dependencies for
+        Brotli (either the ``brotli`` or ``brotlicffi`` package) and/or Zstandard
+        (the ``zstandard`` package) algorithms are installed, then their encodings are
+        included in the string ('br' and 'zstd', respectively).
+        List will get joined by comma.
+        String will be used as provided.
+
+    :param user_agent:
+        String representing the user-agent you want, such as
+        "python-urllib3/0.6"
+
+    :param basic_auth:
+        Colon-separated username:password string for 'authorization: basic ...'
+        auth header.
+
+    :param proxy_basic_auth:
+        Colon-separated username:password string for 'proxy-authorization: basic ...'
+        auth header.
+
+    :param disable_cache:
+        If ``True``, adds 'cache-control: no-cache' header.
+
+    Example:
+
+    .. code-block:: python
+
+        import urllib3
+
+        print(urllib3.util.make_headers(keep_alive=True, user_agent="Batman/1.0"))
+        # {'connection': 'keep-alive', 'user-agent': 'Batman/1.0'}
+        print(urllib3.util.make_headers(accept_encoding=True))
+        # {'accept-encoding': 'gzip,deflate'}
+    """
+    headers: dict[str, str] = {}
+    if accept_encoding:
+        if isinstance(accept_encoding, str):
+            pass
+        elif isinstance(accept_encoding, list):
+            accept_encoding = ",".join(accept_encoding)
+        else:
+            accept_encoding = ACCEPT_ENCODING
+        headers["accept-encoding"] = accept_encoding
+
+    if user_agent:
+        headers["user-agent"] = user_agent
+
+    if keep_alive:
+        headers["connection"] = "keep-alive"
+
+    if basic_auth:
+        headers["authorization"] = (
+            f"Basic {b64encode(basic_auth.encode('latin-1')).decode()}"
+        )
+
+    if proxy_basic_auth:
+        headers["proxy-authorization"] = (
+            f"Basic {b64encode(proxy_basic_auth.encode('latin-1')).decode()}"
+        )
+
+    if disable_cache:
+        headers["cache-control"] = "no-cache"
+
+    return headers
+
+
+def set_file_position(
+    body: typing.Any, pos: _TYPE_BODY_POSITION | None
+) -> _TYPE_BODY_POSITION | None:
+    """
+    If a position is provided, move file to that point.
+    Otherwise, we'll attempt to record a position for future use.
+    """
+    if pos is not None:
+        rewind_body(body, pos)
+    elif getattr(body, "tell", None) is not None:
+        try:
+            pos = body.tell()
+        except OSError:
+            # This differentiates from None, allowing us to catch
+            # a failed `tell()` later when trying to rewind the body.
+            pos = _FAILEDTELL
+
+    return pos
+
+
+def rewind_body(body: typing.IO[typing.AnyStr], body_pos: _TYPE_BODY_POSITION) -> None:
+    """
+    Attempt to rewind body to a certain position.
+    Primarily used for request redirects and retries.
+
+    :param body:
+        File-like object that supports seek.
+
+    :param int pos:
+        Position to seek to in file.
+    """
+    body_seek = getattr(body, "seek", None)
+    if body_seek is not None and isinstance(body_pos, int):
+        try:
+            body_seek(body_pos)
+        except OSError as e:
+            raise UnrewindableBodyError(
+                "An error occurred when rewinding request body for redirect/retry."
+            ) from e
+    elif body_pos is _FAILEDTELL:
+        raise UnrewindableBodyError(
+            "Unable to record file position for rewinding "
+            "request body during a redirect/retry."
+        )
+    else:
+        raise ValueError(
+            f"body_pos must be of type integer, instead it was {type(body_pos)}."
+        )
+
+
+class ChunksAndContentLength(typing.NamedTuple):
+    chunks: typing.Iterable[bytes] | None
+    content_length: int | None
+
+
+def body_to_chunks(
+    body: typing.Any | None, method: str, blocksize: int
+) -> ChunksAndContentLength:
+    """Takes the HTTP request method, body, and blocksize and
+    transforms them into an iterable of chunks to pass to
+    socket.sendall() and an optional 'Content-Length' header.
+
+    A 'Content-Length' of 'None' indicates the length of the body
+    can't be determined so should use 'Transfer-Encoding: chunked'
+    for framing instead.
+    """
+
+    chunks: typing.Iterable[bytes] | None
+    content_length: int | None
+
+    # No body, we need to make a recommendation on 'Content-Length'
+    # based on whether that request method is expected to have
+    # a body or not.
+    if body is None:
+        chunks = None
+        if method.upper() not in _METHODS_NOT_EXPECTING_BODY:
+            content_length = 0
+        else:
+            content_length = None
+
+    # Bytes or strings become bytes
+    elif isinstance(body, (str, bytes)):
+        chunks = (to_bytes(body),)
+        content_length = len(chunks[0])
+
+    # File-like object, TODO: use seek() and tell() for length?
+    elif hasattr(body, "read"):
+
+        def chunk_readable() -> typing.Iterable[bytes]:
+            nonlocal body, blocksize
+            encode = isinstance(body, io.TextIOBase)
+            while True:
+                datablock = body.read(blocksize)
+                if not datablock:
+                    break
+                if encode:
+                    datablock = datablock.encode("utf-8")
+                yield datablock
+
+        chunks = chunk_readable()
+        content_length = None
+
+    # Otherwise we need to start checking via duck-typing.
+    else:
+        try:
+            # Check if the body implements the buffer API.
+            mv = memoryview(body)
+        except TypeError:
+            try:
+                # Check if the body is an iterable
+                chunks = iter(body)
+                content_length = None
+            except TypeError:
+                raise TypeError(
+                    f"'body' must be a bytes-like object, file-like "
+                    f"object, or iterable. Instead was {body!r}"
+                ) from None
+        else:
+            # Since it implements the buffer API can be passed directly to socket.sendall()
+            chunks = (body,)
+            content_length = mv.nbytes
+
+    return ChunksAndContentLength(chunks=chunks, content_length=content_length)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/response.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/response.py
new file mode 100644
index 00000000..0f457869
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/response.py
@@ -0,0 +1,101 @@
+from __future__ import annotations
+
+import http.client as httplib
+from email.errors import MultipartInvariantViolationDefect, StartBoundaryNotFoundDefect
+
+from ..exceptions import HeaderParsingError
+
+
+def is_fp_closed(obj: object) -> bool:
+    """
+    Checks whether a given file-like object is closed.
+
+    :param obj:
+        The file-like object to check.
+    """
+
+    try:
+        # Check `isclosed()` first, in case Python3 doesn't set `closed`.
+        # GH Issue #928
+        return obj.isclosed()  # type: ignore[no-any-return, attr-defined]
+    except AttributeError:
+        pass
+
+    try:
+        # Check via the official file-like-object way.
+        return obj.closed  # type: ignore[no-any-return, attr-defined]
+    except AttributeError:
+        pass
+
+    try:
+        # Check if the object is a container for another file-like object that
+        # gets released on exhaustion (e.g. HTTPResponse).
+        return obj.fp is None  # type: ignore[attr-defined]
+    except AttributeError:
+        pass
+
+    raise ValueError("Unable to determine whether fp is closed.")
+
+
+def assert_header_parsing(headers: httplib.HTTPMessage) -> None:
+    """
+    Asserts whether all headers have been successfully parsed.
+    Extracts encountered errors from the result of parsing headers.
+
+    Only works on Python 3.
+
+    :param http.client.HTTPMessage headers: Headers to verify.
+
+    :raises urllib3.exceptions.HeaderParsingError:
+        If parsing errors are found.
+    """
+
+    # This will fail silently if we pass in the wrong kind of parameter.
+    # To make debugging easier add an explicit check.
+    if not isinstance(headers, httplib.HTTPMessage):
+        raise TypeError(f"expected httplib.Message, got {type(headers)}.")
+
+    unparsed_data = None
+
+    # get_payload is actually email.message.Message.get_payload;
+    # we're only interested in the result if it's not a multipart message
+    if not headers.is_multipart():
+        payload = headers.get_payload()
+
+        if isinstance(payload, (bytes, str)):
+            unparsed_data = payload
+
+    # httplib is assuming a response body is available
+    # when parsing headers even when httplib only sends
+    # header data to parse_headers() This results in
+    # defects on multipart responses in particular.
+    # See: https://github.com/urllib3/urllib3/issues/800
+
+    # So we ignore the following defects:
+    # - StartBoundaryNotFoundDefect:
+    #     The claimed start boundary was never found.
+    # - MultipartInvariantViolationDefect:
+    #     A message claimed to be a multipart but no subparts were found.
+    defects = [
+        defect
+        for defect in headers.defects
+        if not isinstance(
+            defect, (StartBoundaryNotFoundDefect, MultipartInvariantViolationDefect)
+        )
+    ]
+
+    if defects or unparsed_data:
+        raise HeaderParsingError(defects=defects, unparsed_data=unparsed_data)
+
+
+def is_response_to_head(response: httplib.HTTPResponse) -> bool:
+    """
+    Checks whether the request of a response has been a HEAD-request.
+
+    :param http.client.HTTPResponse response:
+        Response to check if the originating request
+        used 'HEAD' as a method.
+    """
+    # FIXME: Can we do this somehow without accessing private httplib _method?
+    method_str = response._method  # type: str  # type: ignore[attr-defined]
+    return method_str.upper() == "HEAD"
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/retry.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/retry.py
new file mode 100644
index 00000000..0456cceb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/retry.py
@@ -0,0 +1,533 @@
+from __future__ import annotations
+
+import email
+import logging
+import random
+import re
+import time
+import typing
+from itertools import takewhile
+from types import TracebackType
+
+from ..exceptions import (
+    ConnectTimeoutError,
+    InvalidHeader,
+    MaxRetryError,
+    ProtocolError,
+    ProxyError,
+    ReadTimeoutError,
+    ResponseError,
+)
+from .util import reraise
+
+if typing.TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from ..connectionpool import ConnectionPool
+    from ..response import BaseHTTPResponse
+
+log = logging.getLogger(__name__)
+
+
+# Data structure for representing the metadata of requests that result in a retry.
+class RequestHistory(typing.NamedTuple):
+    method: str | None
+    url: str | None
+    error: Exception | None
+    status: int | None
+    redirect_location: str | None
+
+
+class Retry:
+    """Retry configuration.
+
+    Each retry attempt will create a new Retry object with updated values, so
+    they can be safely reused.
+
+    Retries can be defined as a default for a pool:
+
+    .. code-block:: python
+
+        retries = Retry(connect=5, read=2, redirect=5)
+        http = PoolManager(retries=retries)
+        response = http.request("GET", "https://example.com/")
+
+    Or per-request (which overrides the default for the pool):
+
+    .. code-block:: python
+
+        response = http.request("GET", "https://example.com/", retries=Retry(10))
+
+    Retries can be disabled by passing ``False``:
+
+    .. code-block:: python
+
+        response = http.request("GET", "https://example.com/", retries=False)
+
+    Errors will be wrapped in :class:`~urllib3.exceptions.MaxRetryError` unless
+    retries are disabled, in which case the causing exception will be raised.
+
+    :param int total:
+        Total number of retries to allow. Takes precedence over other counts.
+
+        Set to ``None`` to remove this constraint and fall back on other
+        counts.
+
+        Set to ``0`` to fail on the first retry.
+
+        Set to ``False`` to disable and imply ``raise_on_redirect=False``.
+
+    :param int connect:
+        How many connection-related errors to retry on.
+
+        These are errors raised before the request is sent to the remote server,
+        which we assume has not triggered the server to process the request.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int read:
+        How many times to retry on read errors.
+
+        These errors are raised after the request was sent to the server, so the
+        request may have side-effects.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int redirect:
+        How many redirects to perform. Limit this to avoid infinite redirect
+        loops.
+
+        A redirect is a HTTP response with a status code 301, 302, 303, 307 or
+        308.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+        Set to ``False`` to disable and imply ``raise_on_redirect=False``.
+
+    :param int status:
+        How many times to retry on bad status codes.
+
+        These are retries made on responses, where status code matches
+        ``status_forcelist``.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+    :param int other:
+        How many times to retry on other errors.
+
+        Other errors are errors that are not connect, read, redirect or status errors.
+        These errors might be raised after the request was sent to the server, so the
+        request might have side-effects.
+
+        Set to ``0`` to fail on the first retry of this type.
+
+        If ``total`` is not set, it's a good idea to set this to 0 to account
+        for unexpected edge cases and avoid infinite retry loops.
+
+    :param Collection allowed_methods:
+        Set of uppercased HTTP method verbs that we should retry on.
+
+        By default, we only retry on methods which are considered to be
+        idempotent (multiple requests with the same parameters end with the
+        same state). See :attr:`Retry.DEFAULT_ALLOWED_METHODS`.
+
+        Set to a ``None`` value to retry on any verb.
+
+    :param Collection status_forcelist:
+        A set of integer HTTP status codes that we should force a retry on.
+        A retry is initiated if the request method is in ``allowed_methods``
+        and the response status code is in ``status_forcelist``.
+
+        By default, this is disabled with ``None``.
+
+    :param float backoff_factor:
+        A backoff factor to apply between attempts after the second try
+        (most errors are resolved immediately by a second try without a
+        delay). urllib3 will sleep for::
+
+            {backoff factor} * (2 ** ({number of previous retries}))
+
+        seconds. If `backoff_jitter` is non-zero, this sleep is extended by::
+
+            random.uniform(0, {backoff jitter})
+
+        seconds. For example, if the backoff_factor is 0.1, then :func:`Retry.sleep` will
+        sleep for [0.0s, 0.2s, 0.4s, 0.8s, ...] between retries. No backoff will ever
+        be longer than `backoff_max`.
+
+        By default, backoff is disabled (factor set to 0).
+
+    :param bool raise_on_redirect: Whether, if the number of redirects is
+        exhausted, to raise a MaxRetryError, or to return a response with a
+        response code in the 3xx range.
+
+    :param bool raise_on_status: Similar meaning to ``raise_on_redirect``:
+        whether we should raise an exception, or return a response,
+        if status falls in ``status_forcelist`` range and retries have
+        been exhausted.
+
+    :param tuple history: The history of the request encountered during
+        each call to :meth:`~Retry.increment`. The list is in the order
+        the requests occurred. Each list item is of class :class:`RequestHistory`.
+
+    :param bool respect_retry_after_header:
+        Whether to respect Retry-After header on status codes defined as
+        :attr:`Retry.RETRY_AFTER_STATUS_CODES` or not.
+
+    :param Collection remove_headers_on_redirect:
+        Sequence of headers to remove from the request when a response
+        indicating a redirect is returned before firing off the redirected
+        request.
+    """
+
+    #: Default methods to be used for ``allowed_methods``
+    DEFAULT_ALLOWED_METHODS = frozenset(
+        ["HEAD", "GET", "PUT", "DELETE", "OPTIONS", "TRACE"]
+    )
+
+    #: Default status codes to be used for ``status_forcelist``
+    RETRY_AFTER_STATUS_CODES = frozenset([413, 429, 503])
+
+    #: Default headers to be used for ``remove_headers_on_redirect``
+    DEFAULT_REMOVE_HEADERS_ON_REDIRECT = frozenset(
+        ["Cookie", "Authorization", "Proxy-Authorization"]
+    )
+
+    #: Default maximum backoff time.
+    DEFAULT_BACKOFF_MAX = 120
+
+    # Backward compatibility; assigned outside of the class.
+    DEFAULT: typing.ClassVar[Retry]
+
+    def __init__(
+        self,
+        total: bool | int | None = 10,
+        connect: int | None = None,
+        read: int | None = None,
+        redirect: bool | int | None = None,
+        status: int | None = None,
+        other: int | None = None,
+        allowed_methods: typing.Collection[str] | None = DEFAULT_ALLOWED_METHODS,
+        status_forcelist: typing.Collection[int] | None = None,
+        backoff_factor: float = 0,
+        backoff_max: float = DEFAULT_BACKOFF_MAX,
+        raise_on_redirect: bool = True,
+        raise_on_status: bool = True,
+        history: tuple[RequestHistory, ...] | None = None,
+        respect_retry_after_header: bool = True,
+        remove_headers_on_redirect: typing.Collection[
+            str
+        ] = DEFAULT_REMOVE_HEADERS_ON_REDIRECT,
+        backoff_jitter: float = 0.0,
+    ) -> None:
+        self.total = total
+        self.connect = connect
+        self.read = read
+        self.status = status
+        self.other = other
+
+        if redirect is False or total is False:
+            redirect = 0
+            raise_on_redirect = False
+
+        self.redirect = redirect
+        self.status_forcelist = status_forcelist or set()
+        self.allowed_methods = allowed_methods
+        self.backoff_factor = backoff_factor
+        self.backoff_max = backoff_max
+        self.raise_on_redirect = raise_on_redirect
+        self.raise_on_status = raise_on_status
+        self.history = history or ()
+        self.respect_retry_after_header = respect_retry_after_header
+        self.remove_headers_on_redirect = frozenset(
+            h.lower() for h in remove_headers_on_redirect
+        )
+        self.backoff_jitter = backoff_jitter
+
+    def new(self, **kw: typing.Any) -> Self:
+        params = dict(
+            total=self.total,
+            connect=self.connect,
+            read=self.read,
+            redirect=self.redirect,
+            status=self.status,
+            other=self.other,
+            allowed_methods=self.allowed_methods,
+            status_forcelist=self.status_forcelist,
+            backoff_factor=self.backoff_factor,
+            backoff_max=self.backoff_max,
+            raise_on_redirect=self.raise_on_redirect,
+            raise_on_status=self.raise_on_status,
+            history=self.history,
+            remove_headers_on_redirect=self.remove_headers_on_redirect,
+            respect_retry_after_header=self.respect_retry_after_header,
+            backoff_jitter=self.backoff_jitter,
+        )
+
+        params.update(kw)
+        return type(self)(**params)  # type: ignore[arg-type]
+
+    @classmethod
+    def from_int(
+        cls,
+        retries: Retry | bool | int | None,
+        redirect: bool | int | None = True,
+        default: Retry | bool | int | None = None,
+    ) -> Retry:
+        """Backwards-compatibility for the old retries format."""
+        if retries is None:
+            retries = default if default is not None else cls.DEFAULT
+
+        if isinstance(retries, Retry):
+            return retries
+
+        redirect = bool(redirect) and None
+        new_retries = cls(retries, redirect=redirect)
+        log.debug("Converted retries value: %r -> %r", retries, new_retries)
+        return new_retries
+
+    def get_backoff_time(self) -> float:
+        """Formula for computing the current backoff
+
+        :rtype: float
+        """
+        # We want to consider only the last consecutive errors sequence (Ignore redirects).
+        consecutive_errors_len = len(
+            list(
+                takewhile(lambda x: x.redirect_location is None, reversed(self.history))
+            )
+        )
+        if consecutive_errors_len <= 1:
+            return 0
+
+        backoff_value = self.backoff_factor * (2 ** (consecutive_errors_len - 1))
+        if self.backoff_jitter != 0.0:
+            backoff_value += random.random() * self.backoff_jitter
+        return float(max(0, min(self.backoff_max, backoff_value)))
+
+    def parse_retry_after(self, retry_after: str) -> float:
+        seconds: float
+        # Whitespace: https://tools.ietf.org/html/rfc7230#section-3.2.4
+        if re.match(r"^\s*[0-9]+\s*$", retry_after):
+            seconds = int(retry_after)
+        else:
+            retry_date_tuple = email.utils.parsedate_tz(retry_after)
+            if retry_date_tuple is None:
+                raise InvalidHeader(f"Invalid Retry-After header: {retry_after}")
+
+            retry_date = email.utils.mktime_tz(retry_date_tuple)
+            seconds = retry_date - time.time()
+
+        seconds = max(seconds, 0)
+
+        return seconds
+
+    def get_retry_after(self, response: BaseHTTPResponse) -> float | None:
+        """Get the value of Retry-After in seconds."""
+
+        retry_after = response.headers.get("Retry-After")
+
+        if retry_after is None:
+            return None
+
+        return self.parse_retry_after(retry_after)
+
+    def sleep_for_retry(self, response: BaseHTTPResponse) -> bool:
+        retry_after = self.get_retry_after(response)
+        if retry_after:
+            time.sleep(retry_after)
+            return True
+
+        return False
+
+    def _sleep_backoff(self) -> None:
+        backoff = self.get_backoff_time()
+        if backoff <= 0:
+            return
+        time.sleep(backoff)
+
+    def sleep(self, response: BaseHTTPResponse | None = None) -> None:
+        """Sleep between retry attempts.
+
+        This method will respect a server's ``Retry-After`` response header
+        and sleep the duration of the time requested. If that is not present, it
+        will use an exponential backoff. By default, the backoff factor is 0 and
+        this method will return immediately.
+        """
+
+        if self.respect_retry_after_header and response:
+            slept = self.sleep_for_retry(response)
+            if slept:
+                return
+
+        self._sleep_backoff()
+
+    def _is_connection_error(self, err: Exception) -> bool:
+        """Errors when we're fairly sure that the server did not receive the
+        request, so it should be safe to retry.
+        """
+        if isinstance(err, ProxyError):
+            err = err.original_error
+        return isinstance(err, ConnectTimeoutError)
+
+    def _is_read_error(self, err: Exception) -> bool:
+        """Errors that occur after the request has been started, so we should
+        assume that the server began processing it.
+        """
+        return isinstance(err, (ReadTimeoutError, ProtocolError))
+
+    def _is_method_retryable(self, method: str) -> bool:
+        """Checks if a given HTTP method should be retried upon, depending if
+        it is included in the allowed_methods
+        """
+        if self.allowed_methods and method.upper() not in self.allowed_methods:
+            return False
+        return True
+
+    def is_retry(
+        self, method: str, status_code: int, has_retry_after: bool = False
+    ) -> bool:
+        """Is this method/status code retryable? (Based on allowlists and control
+        variables such as the number of total retries to allow, whether to
+        respect the Retry-After header, whether this header is present, and
+        whether the returned status code is on the list of status codes to
+        be retried upon on the presence of the aforementioned header)
+        """
+        if not self._is_method_retryable(method):
+            return False
+
+        if self.status_forcelist and status_code in self.status_forcelist:
+            return True
+
+        return bool(
+            self.total
+            and self.respect_retry_after_header
+            and has_retry_after
+            and (status_code in self.RETRY_AFTER_STATUS_CODES)
+        )
+
+    def is_exhausted(self) -> bool:
+        """Are we out of retries?"""
+        retry_counts = [
+            x
+            for x in (
+                self.total,
+                self.connect,
+                self.read,
+                self.redirect,
+                self.status,
+                self.other,
+            )
+            if x
+        ]
+        if not retry_counts:
+            return False
+
+        return min(retry_counts) < 0
+
+    def increment(
+        self,
+        method: str | None = None,
+        url: str | None = None,
+        response: BaseHTTPResponse | None = None,
+        error: Exception | None = None,
+        _pool: ConnectionPool | None = None,
+        _stacktrace: TracebackType | None = None,
+    ) -> Self:
+        """Return a new Retry object with incremented retry counters.
+
+        :param response: A response object, or None, if the server did not
+            return a response.
+        :type response: :class:`~urllib3.response.BaseHTTPResponse`
+        :param Exception error: An error encountered during the request, or
+            None if the response was received successfully.
+
+        :return: A new ``Retry`` object.
+        """
+        if self.total is False and error:
+            # Disabled, indicate to re-raise the error.
+            raise reraise(type(error), error, _stacktrace)
+
+        total = self.total
+        if total is not None:
+            total -= 1
+
+        connect = self.connect
+        read = self.read
+        redirect = self.redirect
+        status_count = self.status
+        other = self.other
+        cause = "unknown"
+        status = None
+        redirect_location = None
+
+        if error and self._is_connection_error(error):
+            # Connect retry?
+            if connect is False:
+                raise reraise(type(error), error, _stacktrace)
+            elif connect is not None:
+                connect -= 1
+
+        elif error and self._is_read_error(error):
+            # Read retry?
+            if read is False or method is None or not self._is_method_retryable(method):
+                raise reraise(type(error), error, _stacktrace)
+            elif read is not None:
+                read -= 1
+
+        elif error:
+            # Other retry?
+            if other is not None:
+                other -= 1
+
+        elif response and response.get_redirect_location():
+            # Redirect retry?
+            if redirect is not None:
+                redirect -= 1
+            cause = "too many redirects"
+            response_redirect_location = response.get_redirect_location()
+            if response_redirect_location:
+                redirect_location = response_redirect_location
+            status = response.status
+
+        else:
+            # Incrementing because of a server error like a 500 in
+            # status_forcelist and the given method is in the allowed_methods
+            cause = ResponseError.GENERIC_ERROR
+            if response and response.status:
+                if status_count is not None:
+                    status_count -= 1
+                cause = ResponseError.SPECIFIC_ERROR.format(status_code=response.status)
+                status = response.status
+
+        history = self.history + (
+            RequestHistory(method, url, error, status, redirect_location),
+        )
+
+        new_retry = self.new(
+            total=total,
+            connect=connect,
+            read=read,
+            redirect=redirect,
+            status=status_count,
+            other=other,
+            history=history,
+        )
+
+        if new_retry.is_exhausted():
+            reason = error or ResponseError(cause)
+            raise MaxRetryError(_pool, url, reason) from reason  # type: ignore[arg-type]
+
+        log.debug("Incremented Retry for (url='%s'): %r", url, new_retry)
+
+        return new_retry
+
+    def __repr__(self) -> str:
+        return (
+            f"{type(self).__name__}(total={self.total}, connect={self.connect}, "
+            f"read={self.read}, redirect={self.redirect}, status={self.status})"
+        )
+
+
+# For backwards compatibility (equivalent to pre-v1.9):
+Retry.DEFAULT = Retry(3)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_.py
new file mode 100644
index 00000000..278128eb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_.py
@@ -0,0 +1,504 @@
+from __future__ import annotations
+
+import hashlib
+import hmac
+import os
+import socket
+import sys
+import typing
+import warnings
+from binascii import unhexlify
+
+from ..exceptions import ProxySchemeUnsupported, SSLError
+from .url import _BRACELESS_IPV6_ADDRZ_RE, _IPV4_RE
+
+SSLContext = None
+SSLTransport = None
+HAS_NEVER_CHECK_COMMON_NAME = False
+IS_PYOPENSSL = False
+ALPN_PROTOCOLS = ["http/1.1"]
+
+_TYPE_VERSION_INFO = tuple[int, int, int, str, int]
+
+# Maps the length of a digest to a possible hash function producing this digest
+HASHFUNC_MAP = {
+    length: getattr(hashlib, algorithm, None)
+    for length, algorithm in ((32, "md5"), (40, "sha1"), (64, "sha256"))
+}
+
+
+def _is_bpo_43522_fixed(
+    implementation_name: str,
+    version_info: _TYPE_VERSION_INFO,
+    pypy_version_info: _TYPE_VERSION_INFO | None,
+) -> bool:
+    """Return True for CPython 3.9.3+ or 3.10+ and PyPy 7.3.8+ where
+    setting SSLContext.hostname_checks_common_name to False works.
+
+    Outside of CPython and PyPy we don't know which implementations work
+    or not so we conservatively use our hostname matching as we know that works
+    on all implementations.
+
+    https://github.com/urllib3/urllib3/issues/2192#issuecomment-821832963
+    https://foss.heptapod.net/pypy/pypy/-/issues/3539
+    """
+    if implementation_name == "pypy":
+        # https://foss.heptapod.net/pypy/pypy/-/issues/3129
+        return pypy_version_info >= (7, 3, 8)  # type: ignore[operator]
+    elif implementation_name == "cpython":
+        major_minor = version_info[:2]
+        micro = version_info[2]
+        return (major_minor == (3, 9) and micro >= 3) or major_minor >= (3, 10)
+    else:  # Defensive:
+        return False
+
+
+def _is_has_never_check_common_name_reliable(
+    openssl_version: str,
+    openssl_version_number: int,
+    implementation_name: str,
+    version_info: _TYPE_VERSION_INFO,
+    pypy_version_info: _TYPE_VERSION_INFO | None,
+) -> bool:
+    # As of May 2023, all released versions of LibreSSL fail to reject certificates with
+    # only common names, see https://github.com/urllib3/urllib3/pull/3024
+    is_openssl = openssl_version.startswith("OpenSSL ")
+    # Before fixing OpenSSL issue #14579, the SSL_new() API was not copying hostflags
+    # like X509_CHECK_FLAG_NEVER_CHECK_SUBJECT, which tripped up CPython.
+    # https://github.com/openssl/openssl/issues/14579
+    # This was released in OpenSSL 1.1.1l+ (>=0x101010cf)
+    is_openssl_issue_14579_fixed = openssl_version_number >= 0x101010CF
+
+    return is_openssl and (
+        is_openssl_issue_14579_fixed
+        or _is_bpo_43522_fixed(implementation_name, version_info, pypy_version_info)
+    )
+
+
+if typing.TYPE_CHECKING:
+    from ssl import VerifyMode
+    from typing import TypedDict
+
+    from .ssltransport import SSLTransport as SSLTransportType
+
+    class _TYPE_PEER_CERT_RET_DICT(TypedDict, total=False):
+        subjectAltName: tuple[tuple[str, str], ...]
+        subject: tuple[tuple[tuple[str, str], ...], ...]
+        serialNumber: str
+
+
+# Mapping from 'ssl.PROTOCOL_TLSX' to 'TLSVersion.X'
+_SSL_VERSION_TO_TLS_VERSION: dict[int, int] = {}
+
+try:  # Do we have ssl at all?
+    import ssl
+    from ssl import (  # type: ignore[assignment]
+        CERT_REQUIRED,
+        HAS_NEVER_CHECK_COMMON_NAME,
+        OP_NO_COMPRESSION,
+        OP_NO_TICKET,
+        OPENSSL_VERSION,
+        OPENSSL_VERSION_NUMBER,
+        PROTOCOL_TLS,
+        PROTOCOL_TLS_CLIENT,
+        OP_NO_SSLv2,
+        OP_NO_SSLv3,
+        SSLContext,
+        TLSVersion,
+    )
+
+    PROTOCOL_SSLv23 = PROTOCOL_TLS
+
+    # Setting SSLContext.hostname_checks_common_name = False didn't work before CPython
+    # 3.9.3, and 3.10 (but OK on PyPy) or OpenSSL 1.1.1l+
+    if HAS_NEVER_CHECK_COMMON_NAME and not _is_has_never_check_common_name_reliable(
+        OPENSSL_VERSION,
+        OPENSSL_VERSION_NUMBER,
+        sys.implementation.name,
+        sys.version_info,
+        sys.pypy_version_info if sys.implementation.name == "pypy" else None,  # type: ignore[attr-defined]
+    ):
+        HAS_NEVER_CHECK_COMMON_NAME = False
+
+    # Need to be careful here in case old TLS versions get
+    # removed in future 'ssl' module implementations.
+    for attr in ("TLSv1", "TLSv1_1", "TLSv1_2"):
+        try:
+            _SSL_VERSION_TO_TLS_VERSION[getattr(ssl, f"PROTOCOL_{attr}")] = getattr(
+                TLSVersion, attr
+            )
+        except AttributeError:  # Defensive:
+            continue
+
+    from .ssltransport import SSLTransport  # type: ignore[assignment]
+except ImportError:
+    OP_NO_COMPRESSION = 0x20000  # type: ignore[assignment]
+    OP_NO_TICKET = 0x4000  # type: ignore[assignment]
+    OP_NO_SSLv2 = 0x1000000  # type: ignore[assignment]
+    OP_NO_SSLv3 = 0x2000000  # type: ignore[assignment]
+    PROTOCOL_SSLv23 = PROTOCOL_TLS = 2  # type: ignore[assignment]
+    PROTOCOL_TLS_CLIENT = 16  # type: ignore[assignment]
+
+
+_TYPE_PEER_CERT_RET = typing.Union["_TYPE_PEER_CERT_RET_DICT", bytes, None]
+
+
+def assert_fingerprint(cert: bytes | None, fingerprint: str) -> None:
+    """
+    Checks if given fingerprint matches the supplied certificate.
+
+    :param cert:
+        Certificate as bytes object.
+    :param fingerprint:
+        Fingerprint as string of hexdigits, can be interspersed by colons.
+    """
+
+    if cert is None:
+        raise SSLError("No certificate for the peer.")
+
+    fingerprint = fingerprint.replace(":", "").lower()
+    digest_length = len(fingerprint)
+    if digest_length not in HASHFUNC_MAP:
+        raise SSLError(f"Fingerprint of invalid length: {fingerprint}")
+    hashfunc = HASHFUNC_MAP.get(digest_length)
+    if hashfunc is None:
+        raise SSLError(
+            f"Hash function implementation unavailable for fingerprint length: {digest_length}"
+        )
+
+    # We need encode() here for py32; works on py2 and p33.
+    fingerprint_bytes = unhexlify(fingerprint.encode())
+
+    cert_digest = hashfunc(cert).digest()
+
+    if not hmac.compare_digest(cert_digest, fingerprint_bytes):
+        raise SSLError(
+            f'Fingerprints did not match. Expected "{fingerprint}", got "{cert_digest.hex()}"'
+        )
+
+
+def resolve_cert_reqs(candidate: None | int | str) -> VerifyMode:
+    """
+    Resolves the argument to a numeric constant, which can be passed to
+    the wrap_socket function/method from the ssl module.
+    Defaults to :data:`ssl.CERT_REQUIRED`.
+    If given a string it is assumed to be the name of the constant in the
+    :mod:`ssl` module or its abbreviation.
+    (So you can specify `REQUIRED` instead of `CERT_REQUIRED`.
+    If it's neither `None` nor a string we assume it is already the numeric
+    constant which can directly be passed to wrap_socket.
+    """
+    if candidate is None:
+        return CERT_REQUIRED
+
+    if isinstance(candidate, str):
+        res = getattr(ssl, candidate, None)
+        if res is None:
+            res = getattr(ssl, "CERT_" + candidate)
+        return res  # type: ignore[no-any-return]
+
+    return candidate  # type: ignore[return-value]
+
+
+def resolve_ssl_version(candidate: None | int | str) -> int:
+    """
+    like resolve_cert_reqs
+    """
+    if candidate is None:
+        return PROTOCOL_TLS
+
+    if isinstance(candidate, str):
+        res = getattr(ssl, candidate, None)
+        if res is None:
+            res = getattr(ssl, "PROTOCOL_" + candidate)
+        return typing.cast(int, res)
+
+    return candidate
+
+
+def create_urllib3_context(
+    ssl_version: int | None = None,
+    cert_reqs: int | None = None,
+    options: int | None = None,
+    ciphers: str | None = None,
+    ssl_minimum_version: int | None = None,
+    ssl_maximum_version: int | None = None,
+) -> ssl.SSLContext:
+    """Creates and configures an :class:`ssl.SSLContext` instance for use with urllib3.
+
+    :param ssl_version:
+        The desired protocol version to use. This will default to
+        PROTOCOL_SSLv23 which will negotiate the highest protocol that both
+        the server and your installation of OpenSSL support.
+
+        This parameter is deprecated instead use 'ssl_minimum_version'.
+    :param ssl_minimum_version:
+        The minimum version of TLS to be used. Use the 'ssl.TLSVersion' enum for specifying the value.
+    :param ssl_maximum_version:
+        The maximum version of TLS to be used. Use the 'ssl.TLSVersion' enum for specifying the value.
+        Not recommended to set to anything other than 'ssl.TLSVersion.MAXIMUM_SUPPORTED' which is the
+        default value.
+    :param cert_reqs:
+        Whether to require the certificate verification. This defaults to
+        ``ssl.CERT_REQUIRED``.
+    :param options:
+        Specific OpenSSL options. These default to ``ssl.OP_NO_SSLv2``,
+        ``ssl.OP_NO_SSLv3``, ``ssl.OP_NO_COMPRESSION``, and ``ssl.OP_NO_TICKET``.
+    :param ciphers:
+        Which cipher suites to allow the server to select. Defaults to either system configured
+        ciphers if OpenSSL 1.1.1+, otherwise uses a secure default set of ciphers.
+    :returns:
+        Constructed SSLContext object with specified options
+    :rtype: SSLContext
+    """
+    if SSLContext is None:
+        raise TypeError("Can't create an SSLContext object without an ssl module")
+
+    # This means 'ssl_version' was specified as an exact value.
+    if ssl_version not in (None, PROTOCOL_TLS, PROTOCOL_TLS_CLIENT):
+        # Disallow setting 'ssl_version' and 'ssl_minimum|maximum_version'
+        # to avoid conflicts.
+        if ssl_minimum_version is not None or ssl_maximum_version is not None:
+            raise ValueError(
+                "Can't specify both 'ssl_version' and either "
+                "'ssl_minimum_version' or 'ssl_maximum_version'"
+            )
+
+        # 'ssl_version' is deprecated and will be removed in the future.
+        else:
+            # Use 'ssl_minimum_version' and 'ssl_maximum_version' instead.
+            ssl_minimum_version = _SSL_VERSION_TO_TLS_VERSION.get(
+                ssl_version, TLSVersion.MINIMUM_SUPPORTED
+            )
+            ssl_maximum_version = _SSL_VERSION_TO_TLS_VERSION.get(
+                ssl_version, TLSVersion.MAXIMUM_SUPPORTED
+            )
+
+            # This warning message is pushing users to use 'ssl_minimum_version'
+            # instead of both min/max. Best practice is to only set the minimum version and
+            # keep the maximum version to be it's default value: 'TLSVersion.MAXIMUM_SUPPORTED'
+            warnings.warn(
+                "'ssl_version' option is deprecated and will be "
+                "removed in urllib3 v2.1.0. Instead use 'ssl_minimum_version'",
+                category=DeprecationWarning,
+                stacklevel=2,
+            )
+
+    # PROTOCOL_TLS is deprecated in Python 3.10 so we always use PROTOCOL_TLS_CLIENT
+    context = SSLContext(PROTOCOL_TLS_CLIENT)
+
+    if ssl_minimum_version is not None:
+        context.minimum_version = ssl_minimum_version
+    else:  # Python <3.10 defaults to 'MINIMUM_SUPPORTED' so explicitly set TLSv1.2 here
+        context.minimum_version = TLSVersion.TLSv1_2
+
+    if ssl_maximum_version is not None:
+        context.maximum_version = ssl_maximum_version
+
+    # Unless we're given ciphers defer to either system ciphers in
+    # the case of OpenSSL 1.1.1+ or use our own secure default ciphers.
+    if ciphers:
+        context.set_ciphers(ciphers)
+
+    # Setting the default here, as we may have no ssl module on import
+    cert_reqs = ssl.CERT_REQUIRED if cert_reqs is None else cert_reqs
+
+    if options is None:
+        options = 0
+        # SSLv2 is easily broken and is considered harmful and dangerous
+        options |= OP_NO_SSLv2
+        # SSLv3 has several problems and is now dangerous
+        options |= OP_NO_SSLv3
+        # Disable compression to prevent CRIME attacks for OpenSSL 1.0+
+        # (issue #309)
+        options |= OP_NO_COMPRESSION
+        # TLSv1.2 only. Unless set explicitly, do not request tickets.
+        # This may save some bandwidth on wire, and although the ticket is encrypted,
+        # there is a risk associated with it being on wire,
+        # if the server is not rotating its ticketing keys properly.
+        options |= OP_NO_TICKET
+
+    context.options |= options
+
+    # Enable post-handshake authentication for TLS 1.3, see GH #1634. PHA is
+    # necessary for conditional client cert authentication with TLS 1.3.
+    # The attribute is None for OpenSSL <= 1.1.0 or does not exist when using
+    # an SSLContext created by pyOpenSSL.
+    if getattr(context, "post_handshake_auth", None) is not None:
+        context.post_handshake_auth = True
+
+    # The order of the below lines setting verify_mode and check_hostname
+    # matter due to safe-guards SSLContext has to prevent an SSLContext with
+    # check_hostname=True, verify_mode=NONE/OPTIONAL.
+    # We always set 'check_hostname=False' for pyOpenSSL so we rely on our own
+    # 'ssl.match_hostname()' implementation.
+    if cert_reqs == ssl.CERT_REQUIRED and not IS_PYOPENSSL:
+        context.verify_mode = cert_reqs
+        context.check_hostname = True
+    else:
+        context.check_hostname = False
+        context.verify_mode = cert_reqs
+
+    try:
+        context.hostname_checks_common_name = False
+    except AttributeError:  # Defensive: for CPython < 3.9.3; for PyPy < 7.3.8
+        pass
+
+    sslkeylogfile = os.environ.get("SSLKEYLOGFILE")
+    if sslkeylogfile:
+        context.keylog_filename = sslkeylogfile
+
+    return context
+
+
+@typing.overload
+def ssl_wrap_socket(
+    sock: socket.socket,
+    keyfile: str | None = ...,
+    certfile: str | None = ...,
+    cert_reqs: int | None = ...,
+    ca_certs: str | None = ...,
+    server_hostname: str | None = ...,
+    ssl_version: int | None = ...,
+    ciphers: str | None = ...,
+    ssl_context: ssl.SSLContext | None = ...,
+    ca_cert_dir: str | None = ...,
+    key_password: str | None = ...,
+    ca_cert_data: None | str | bytes = ...,
+    tls_in_tls: typing.Literal[False] = ...,
+) -> ssl.SSLSocket: ...
+
+
+@typing.overload
+def ssl_wrap_socket(
+    sock: socket.socket,
+    keyfile: str | None = ...,
+    certfile: str | None = ...,
+    cert_reqs: int | None = ...,
+    ca_certs: str | None = ...,
+    server_hostname: str | None = ...,
+    ssl_version: int | None = ...,
+    ciphers: str | None = ...,
+    ssl_context: ssl.SSLContext | None = ...,
+    ca_cert_dir: str | None = ...,
+    key_password: str | None = ...,
+    ca_cert_data: None | str | bytes = ...,
+    tls_in_tls: bool = ...,
+) -> ssl.SSLSocket | SSLTransportType: ...
+
+
+def ssl_wrap_socket(
+    sock: socket.socket,
+    keyfile: str | None = None,
+    certfile: str | None = None,
+    cert_reqs: int | None = None,
+    ca_certs: str | None = None,
+    server_hostname: str | None = None,
+    ssl_version: int | None = None,
+    ciphers: str | None = None,
+    ssl_context: ssl.SSLContext | None = None,
+    ca_cert_dir: str | None = None,
+    key_password: str | None = None,
+    ca_cert_data: None | str | bytes = None,
+    tls_in_tls: bool = False,
+) -> ssl.SSLSocket | SSLTransportType:
+    """
+    All arguments except for server_hostname, ssl_context, tls_in_tls, ca_cert_data and
+    ca_cert_dir have the same meaning as they do when using
+    :func:`ssl.create_default_context`, :meth:`ssl.SSLContext.load_cert_chain`,
+    :meth:`ssl.SSLContext.set_ciphers` and :meth:`ssl.SSLContext.wrap_socket`.
+
+    :param server_hostname:
+        When SNI is supported, the expected hostname of the certificate
+    :param ssl_context:
+        A pre-made :class:`SSLContext` object. If none is provided, one will
+        be created using :func:`create_urllib3_context`.
+    :param ciphers:
+        A string of ciphers we wish the client to support.
+    :param ca_cert_dir:
+        A directory containing CA certificates in multiple separate files, as
+        supported by OpenSSL's -CApath flag or the capath argument to
+        SSLContext.load_verify_locations().
+    :param key_password:
+        Optional password if the keyfile is encrypted.
+    :param ca_cert_data:
+        Optional string containing CA certificates in PEM format suitable for
+        passing as the cadata parameter to SSLContext.load_verify_locations()
+    :param tls_in_tls:
+        Use SSLTransport to wrap the existing socket.
+    """
+    context = ssl_context
+    if context is None:
+        # Note: This branch of code and all the variables in it are only used in tests.
+        # We should consider deprecating and removing this code.
+        context = create_urllib3_context(ssl_version, cert_reqs, ciphers=ciphers)
+
+    if ca_certs or ca_cert_dir or ca_cert_data:
+        try:
+            context.load_verify_locations(ca_certs, ca_cert_dir, ca_cert_data)
+        except OSError as e:
+            raise SSLError(e) from e
+
+    elif ssl_context is None and hasattr(context, "load_default_certs"):
+        # try to load OS default certs; works well on Windows.
+        context.load_default_certs()
+
+    # Attempt to detect if we get the goofy behavior of the
+    # keyfile being encrypted and OpenSSL asking for the
+    # passphrase via the terminal and instead error out.
+    if keyfile and key_password is None and _is_key_file_encrypted(keyfile):
+        raise SSLError("Client private key is encrypted, password is required")
+
+    if certfile:
+        if key_password is None:
+            context.load_cert_chain(certfile, keyfile)
+        else:
+            context.load_cert_chain(certfile, keyfile, key_password)
+
+    context.set_alpn_protocols(ALPN_PROTOCOLS)
+
+    ssl_sock = _ssl_wrap_socket_impl(sock, context, tls_in_tls, server_hostname)
+    return ssl_sock
+
+
+def is_ipaddress(hostname: str | bytes) -> bool:
+    """Detects whether the hostname given is an IPv4 or IPv6 address.
+    Also detects IPv6 addresses with Zone IDs.
+
+    :param str hostname: Hostname to examine.
+    :return: True if the hostname is an IP address, False otherwise.
+    """
+    if isinstance(hostname, bytes):
+        # IDN A-label bytes are ASCII compatible.
+        hostname = hostname.decode("ascii")
+    return bool(_IPV4_RE.match(hostname) or _BRACELESS_IPV6_ADDRZ_RE.match(hostname))
+
+
+def _is_key_file_encrypted(key_file: str) -> bool:
+    """Detects if a key file is encrypted or not."""
+    with open(key_file) as f:
+        for line in f:
+            # Look for Proc-Type: 4,ENCRYPTED
+            if "ENCRYPTED" in line:
+                return True
+
+    return False
+
+
+def _ssl_wrap_socket_impl(
+    sock: socket.socket,
+    ssl_context: ssl.SSLContext,
+    tls_in_tls: bool,
+    server_hostname: str | None = None,
+) -> ssl.SSLSocket | SSLTransportType:
+    if tls_in_tls:
+        if not SSLTransport:
+            # Import error, ssl is not available.
+            raise ProxySchemeUnsupported(
+                "TLS in TLS requires support for the 'ssl' module"
+            )
+
+        SSLTransport._validate_ssl_context_for_tls_in_tls(ssl_context)
+        return SSLTransport(sock, ssl_context, server_hostname)
+
+    return ssl_context.wrap_socket(sock, server_hostname=server_hostname)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_match_hostname.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_match_hostname.py
new file mode 100644
index 00000000..453cfd42
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssl_match_hostname.py
@@ -0,0 +1,159 @@
+"""The match_hostname() function from Python 3.5, essential when using SSL."""
+
+# Note: This file is under the PSF license as the code comes from the python
+# stdlib.   http://docs.python.org/3/license.html
+# It is modified to remove commonName support.
+
+from __future__ import annotations
+
+import ipaddress
+import re
+import typing
+from ipaddress import IPv4Address, IPv6Address
+
+if typing.TYPE_CHECKING:
+    from .ssl_ import _TYPE_PEER_CERT_RET_DICT
+
+__version__ = "3.5.0.1"
+
+
+class CertificateError(ValueError):
+    pass
+
+
+def _dnsname_match(
+    dn: typing.Any, hostname: str, max_wildcards: int = 1
+) -> typing.Match[str] | None | bool:
+    """Matching according to RFC 6125, section 6.4.3
+
+    http://tools.ietf.org/html/rfc6125#section-6.4.3
+    """
+    pats = []
+    if not dn:
+        return False
+
+    # Ported from python3-syntax:
+    # leftmost, *remainder = dn.split(r'.')
+    parts = dn.split(r".")
+    leftmost = parts[0]
+    remainder = parts[1:]
+
+    wildcards = leftmost.count("*")
+    if wildcards > max_wildcards:
+        # Issue #17980: avoid denials of service by refusing more
+        # than one wildcard per fragment.  A survey of established
+        # policy among SSL implementations showed it to be a
+        # reasonable choice.
+        raise CertificateError(
+            "too many wildcards in certificate DNS name: " + repr(dn)
+        )
+
+    # speed up common case w/o wildcards
+    if not wildcards:
+        return bool(dn.lower() == hostname.lower())
+
+    # RFC 6125, section 6.4.3, subitem 1.
+    # The client SHOULD NOT attempt to match a presented identifier in which
+    # the wildcard character comprises a label other than the left-most label.
+    if leftmost == "*":
+        # When '*' is a fragment by itself, it matches a non-empty dotless
+        # fragment.
+        pats.append("[^.]+")
+    elif leftmost.startswith("xn--") or hostname.startswith("xn--"):
+        # RFC 6125, section 6.4.3, subitem 3.
+        # The client SHOULD NOT attempt to match a presented identifier
+        # where the wildcard character is embedded within an A-label or
+        # U-label of an internationalized domain name.
+        pats.append(re.escape(leftmost))
+    else:
+        # Otherwise, '*' matches any dotless string, e.g. www*
+        pats.append(re.escape(leftmost).replace(r"\*", "[^.]*"))
+
+    # add the remaining fragments, ignore any wildcards
+    for frag in remainder:
+        pats.append(re.escape(frag))
+
+    pat = re.compile(r"\A" + r"\.".join(pats) + r"\Z", re.IGNORECASE)
+    return pat.match(hostname)
+
+
+def _ipaddress_match(ipname: str, host_ip: IPv4Address | IPv6Address) -> bool:
+    """Exact matching of IP addresses.
+
+    RFC 9110 section 4.3.5: "A reference identity of IP-ID contains the decoded
+    bytes of the IP address. An IP version 4 address is 4 octets, and an IP
+    version 6 address is 16 octets. [...] A reference identity of type IP-ID
+    matches if the address is identical to an iPAddress value of the
+    subjectAltName extension of the certificate."
+    """
+    # OpenSSL may add a trailing newline to a subjectAltName's IP address
+    # Divergence from upstream: ipaddress can't handle byte str
+    ip = ipaddress.ip_address(ipname.rstrip())
+    return bool(ip.packed == host_ip.packed)
+
+
+def match_hostname(
+    cert: _TYPE_PEER_CERT_RET_DICT | None,
+    hostname: str,
+    hostname_checks_common_name: bool = False,
+) -> None:
+    """Verify that *cert* (in decoded format as returned by
+    SSLSocket.getpeercert()) matches the *hostname*.  RFC 2818 and RFC 6125
+    rules are followed, but IP addresses are not accepted for *hostname*.
+
+    CertificateError is raised on failure. On success, the function
+    returns nothing.
+    """
+    if not cert:
+        raise ValueError(
+            "empty or no certificate, match_hostname needs a "
+            "SSL socket or SSL context with either "
+            "CERT_OPTIONAL or CERT_REQUIRED"
+        )
+    try:
+        # Divergence from upstream: ipaddress can't handle byte str
+        #
+        # The ipaddress module shipped with Python < 3.9 does not support
+        # scoped IPv6 addresses so we unconditionally strip the Zone IDs for
+        # now. Once we drop support for Python 3.9 we can remove this branch.
+        if "%" in hostname:
+            host_ip = ipaddress.ip_address(hostname[: hostname.rfind("%")])
+        else:
+            host_ip = ipaddress.ip_address(hostname)
+
+    except ValueError:
+        # Not an IP address (common case)
+        host_ip = None
+    dnsnames = []
+    san: tuple[tuple[str, str], ...] = cert.get("subjectAltName", ())
+    key: str
+    value: str
+    for key, value in san:
+        if key == "DNS":
+            if host_ip is None and _dnsname_match(value, hostname):
+                return
+            dnsnames.append(value)
+        elif key == "IP Address":
+            if host_ip is not None and _ipaddress_match(value, host_ip):
+                return
+            dnsnames.append(value)
+
+    # We only check 'commonName' if it's enabled and we're not verifying
+    # an IP address. IP addresses aren't valid within 'commonName'.
+    if hostname_checks_common_name and host_ip is None and not dnsnames:
+        for sub in cert.get("subject", ()):
+            for key, value in sub:
+                if key == "commonName":
+                    if _dnsname_match(value, hostname):
+                        return
+                    dnsnames.append(value)
+
+    if len(dnsnames) > 1:
+        raise CertificateError(
+            "hostname %r "
+            "doesn't match either of %s" % (hostname, ", ".join(map(repr, dnsnames)))
+        )
+    elif len(dnsnames) == 1:
+        raise CertificateError(f"hostname {hostname!r} doesn't match {dnsnames[0]!r}")
+    else:
+        raise CertificateError("no appropriate subjectAltName fields were found")
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/ssltransport.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssltransport.py
new file mode 100644
index 00000000..6d59bc3b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/ssltransport.py
@@ -0,0 +1,271 @@
+from __future__ import annotations
+
+import io
+import socket
+import ssl
+import typing
+
+from ..exceptions import ProxySchemeUnsupported
+
+if typing.TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from .ssl_ import _TYPE_PEER_CERT_RET, _TYPE_PEER_CERT_RET_DICT
+
+
+_WriteBuffer = typing.Union[bytearray, memoryview]
+_ReturnValue = typing.TypeVar("_ReturnValue")
+
+SSL_BLOCKSIZE = 16384
+
+
+class SSLTransport:
+    """
+    The SSLTransport wraps an existing socket and establishes an SSL connection.
+
+    Contrary to Python's implementation of SSLSocket, it allows you to chain
+    multiple TLS connections together. It's particularly useful if you need to
+    implement TLS within TLS.
+
+    The class supports most of the socket API operations.
+    """
+
+    @staticmethod
+    def _validate_ssl_context_for_tls_in_tls(ssl_context: ssl.SSLContext) -> None:
+        """
+        Raises a ProxySchemeUnsupported if the provided ssl_context can't be used
+        for TLS in TLS.
+
+        The only requirement is that the ssl_context provides the 'wrap_bio'
+        methods.
+        """
+
+        if not hasattr(ssl_context, "wrap_bio"):
+            raise ProxySchemeUnsupported(
+                "TLS in TLS requires SSLContext.wrap_bio() which isn't "
+                "available on non-native SSLContext"
+            )
+
+    def __init__(
+        self,
+        socket: socket.socket,
+        ssl_context: ssl.SSLContext,
+        server_hostname: str | None = None,
+        suppress_ragged_eofs: bool = True,
+    ) -> None:
+        """
+        Create an SSLTransport around socket using the provided ssl_context.
+        """
+        self.incoming = ssl.MemoryBIO()
+        self.outgoing = ssl.MemoryBIO()
+
+        self.suppress_ragged_eofs = suppress_ragged_eofs
+        self.socket = socket
+
+        self.sslobj = ssl_context.wrap_bio(
+            self.incoming, self.outgoing, server_hostname=server_hostname
+        )
+
+        # Perform initial handshake.
+        self._ssl_io_loop(self.sslobj.do_handshake)
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(self, *_: typing.Any) -> None:
+        self.close()
+
+    def fileno(self) -> int:
+        return self.socket.fileno()
+
+    def read(self, len: int = 1024, buffer: typing.Any | None = None) -> int | bytes:
+        return self._wrap_ssl_read(len, buffer)
+
+    def recv(self, buflen: int = 1024, flags: int = 0) -> int | bytes:
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to recv")
+        return self._wrap_ssl_read(buflen)
+
+    def recv_into(
+        self,
+        buffer: _WriteBuffer,
+        nbytes: int | None = None,
+        flags: int = 0,
+    ) -> None | int | bytes:
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to recv_into")
+        if nbytes is None:
+            nbytes = len(buffer)
+        return self.read(nbytes, buffer)
+
+    def sendall(self, data: bytes, flags: int = 0) -> None:
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to sendall")
+        count = 0
+        with memoryview(data) as view, view.cast("B") as byte_view:
+            amount = len(byte_view)
+            while count < amount:
+                v = self.send(byte_view[count:])
+                count += v
+
+    def send(self, data: bytes, flags: int = 0) -> int:
+        if flags != 0:
+            raise ValueError("non-zero flags not allowed in calls to send")
+        return self._ssl_io_loop(self.sslobj.write, data)
+
+    def makefile(
+        self,
+        mode: str,
+        buffering: int | None = None,
+        *,
+        encoding: str | None = None,
+        errors: str | None = None,
+        newline: str | None = None,
+    ) -> typing.BinaryIO | typing.TextIO | socket.SocketIO:
+        """
+        Python's httpclient uses makefile and buffered io when reading HTTP
+        messages and we need to support it.
+
+        This is unfortunately a copy and paste of socket.py makefile with small
+        changes to point to the socket directly.
+        """
+        if not set(mode) <= {"r", "w", "b"}:
+            raise ValueError(f"invalid mode {mode!r} (only r, w, b allowed)")
+
+        writing = "w" in mode
+        reading = "r" in mode or not writing
+        assert reading or writing
+        binary = "b" in mode
+        rawmode = ""
+        if reading:
+            rawmode += "r"
+        if writing:
+            rawmode += "w"
+        raw = socket.SocketIO(self, rawmode)  # type: ignore[arg-type]
+        self.socket._io_refs += 1  # type: ignore[attr-defined]
+        if buffering is None:
+            buffering = -1
+        if buffering < 0:
+            buffering = io.DEFAULT_BUFFER_SIZE
+        if buffering == 0:
+            if not binary:
+                raise ValueError("unbuffered streams must be binary")
+            return raw
+        buffer: typing.BinaryIO
+        if reading and writing:
+            buffer = io.BufferedRWPair(raw, raw, buffering)  # type: ignore[assignment]
+        elif reading:
+            buffer = io.BufferedReader(raw, buffering)
+        else:
+            assert writing
+            buffer = io.BufferedWriter(raw, buffering)
+        if binary:
+            return buffer
+        text = io.TextIOWrapper(buffer, encoding, errors, newline)
+        text.mode = mode  # type: ignore[misc]
+        return text
+
+    def unwrap(self) -> None:
+        self._ssl_io_loop(self.sslobj.unwrap)
+
+    def close(self) -> None:
+        self.socket.close()
+
+    @typing.overload
+    def getpeercert(
+        self, binary_form: typing.Literal[False] = ...
+    ) -> _TYPE_PEER_CERT_RET_DICT | None: ...
+
+    @typing.overload
+    def getpeercert(self, binary_form: typing.Literal[True]) -> bytes | None: ...
+
+    def getpeercert(self, binary_form: bool = False) -> _TYPE_PEER_CERT_RET:
+        return self.sslobj.getpeercert(binary_form)  # type: ignore[return-value]
+
+    def version(self) -> str | None:
+        return self.sslobj.version()
+
+    def cipher(self) -> tuple[str, str, int] | None:
+        return self.sslobj.cipher()
+
+    def selected_alpn_protocol(self) -> str | None:
+        return self.sslobj.selected_alpn_protocol()
+
+    def shared_ciphers(self) -> list[tuple[str, str, int]] | None:
+        return self.sslobj.shared_ciphers()
+
+    def compression(self) -> str | None:
+        return self.sslobj.compression()
+
+    def settimeout(self, value: float | None) -> None:
+        self.socket.settimeout(value)
+
+    def gettimeout(self) -> float | None:
+        return self.socket.gettimeout()
+
+    def _decref_socketios(self) -> None:
+        self.socket._decref_socketios()  # type: ignore[attr-defined]
+
+    def _wrap_ssl_read(self, len: int, buffer: bytearray | None = None) -> int | bytes:
+        try:
+            return self._ssl_io_loop(self.sslobj.read, len, buffer)
+        except ssl.SSLError as e:
+            if e.errno == ssl.SSL_ERROR_EOF and self.suppress_ragged_eofs:
+                return 0  # eof, return 0.
+            else:
+                raise
+
+    # func is sslobj.do_handshake or sslobj.unwrap
+    @typing.overload
+    def _ssl_io_loop(self, func: typing.Callable[[], None]) -> None: ...
+
+    # func is sslobj.write, arg1 is data
+    @typing.overload
+    def _ssl_io_loop(self, func: typing.Callable[[bytes], int], arg1: bytes) -> int: ...
+
+    # func is sslobj.read, arg1 is len, arg2 is buffer
+    @typing.overload
+    def _ssl_io_loop(
+        self,
+        func: typing.Callable[[int, bytearray | None], bytes],
+        arg1: int,
+        arg2: bytearray | None,
+    ) -> bytes: ...
+
+    def _ssl_io_loop(
+        self,
+        func: typing.Callable[..., _ReturnValue],
+        arg1: None | bytes | int = None,
+        arg2: bytearray | None = None,
+    ) -> _ReturnValue:
+        """Performs an I/O loop between incoming/outgoing and the socket."""
+        should_loop = True
+        ret = None
+
+        while should_loop:
+            errno = None
+            try:
+                if arg1 is None and arg2 is None:
+                    ret = func()
+                elif arg2 is None:
+                    ret = func(arg1)
+                else:
+                    ret = func(arg1, arg2)
+            except ssl.SSLError as e:
+                if e.errno not in (ssl.SSL_ERROR_WANT_READ, ssl.SSL_ERROR_WANT_WRITE):
+                    # WANT_READ, and WANT_WRITE are expected, others are not.
+                    raise e
+                errno = e.errno
+
+            buf = self.outgoing.read()
+            self.socket.sendall(buf)
+
+            if errno is None:
+                should_loop = False
+            elif errno == ssl.SSL_ERROR_WANT_READ:
+                buf = self.socket.recv(SSL_BLOCKSIZE)
+                if buf:
+                    self.incoming.write(buf)
+                else:
+                    self.incoming.write_eof()
+        return typing.cast(_ReturnValue, ret)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/timeout.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/timeout.py
new file mode 100644
index 00000000..4bb1be11
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/timeout.py
@@ -0,0 +1,275 @@
+from __future__ import annotations
+
+import time
+import typing
+from enum import Enum
+from socket import getdefaulttimeout
+
+from ..exceptions import TimeoutStateError
+
+if typing.TYPE_CHECKING:
+    from typing import Final
+
+
+class _TYPE_DEFAULT(Enum):
+    # This value should never be passed to socket.settimeout() so for safety we use a -1.
+    # socket.settimout() raises a ValueError for negative values.
+    token = -1
+
+
+_DEFAULT_TIMEOUT: Final[_TYPE_DEFAULT] = _TYPE_DEFAULT.token
+
+_TYPE_TIMEOUT = typing.Optional[typing.Union[float, _TYPE_DEFAULT]]
+
+
+class Timeout:
+    """Timeout configuration.
+
+    Timeouts can be defined as a default for a pool:
+
+    .. code-block:: python
+
+        import urllib3
+
+        timeout = urllib3.util.Timeout(connect=2.0, read=7.0)
+
+        http = urllib3.PoolManager(timeout=timeout)
+
+        resp = http.request("GET", "https://example.com/")
+
+        print(resp.status)
+
+    Or per-request (which overrides the default for the pool):
+
+    .. code-block:: python
+
+       response = http.request("GET", "https://example.com/", timeout=Timeout(10))
+
+    Timeouts can be disabled by setting all the parameters to ``None``:
+
+    .. code-block:: python
+
+       no_timeout = Timeout(connect=None, read=None)
+       response = http.request("GET", "https://example.com/", timeout=no_timeout)
+
+
+    :param total:
+        This combines the connect and read timeouts into one; the read timeout
+        will be set to the time leftover from the connect attempt. In the
+        event that both a connect timeout and a total are specified, or a read
+        timeout and a total are specified, the shorter timeout will be applied.
+
+        Defaults to None.
+
+    :type total: int, float, or None
+
+    :param connect:
+        The maximum amount of time (in seconds) to wait for a connection
+        attempt to a server to succeed. Omitting the parameter will default the
+        connect timeout to the system default, probably `the global default
+        timeout in socket.py
+        <http://hg.python.org/cpython/file/603b4d593758/Lib/socket.py#l535>`_.
+        None will set an infinite timeout for connection attempts.
+
+    :type connect: int, float, or None
+
+    :param read:
+        The maximum amount of time (in seconds) to wait between consecutive
+        read operations for a response from the server. Omitting the parameter
+        will default the read timeout to the system default, probably `the
+        global default timeout in socket.py
+        <http://hg.python.org/cpython/file/603b4d593758/Lib/socket.py#l535>`_.
+        None will set an infinite timeout.
+
+    :type read: int, float, or None
+
+    .. note::
+
+        Many factors can affect the total amount of time for urllib3 to return
+        an HTTP response.
+
+        For example, Python's DNS resolver does not obey the timeout specified
+        on the socket. Other factors that can affect total request time include
+        high CPU load, high swap, the program running at a low priority level,
+        or other behaviors.
+
+        In addition, the read and total timeouts only measure the time between
+        read operations on the socket connecting the client and the server,
+        not the total amount of time for the request to return a complete
+        response. For most requests, the timeout is raised because the server
+        has not sent the first byte in the specified time. This is not always
+        the case; if a server streams one byte every fifteen seconds, a timeout
+        of 20 seconds will not trigger, even though the request will take
+        several minutes to complete.
+    """
+
+    #: A sentinel object representing the default timeout value
+    DEFAULT_TIMEOUT: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT
+
+    def __init__(
+        self,
+        total: _TYPE_TIMEOUT = None,
+        connect: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+        read: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
+    ) -> None:
+        self._connect = self._validate_timeout(connect, "connect")
+        self._read = self._validate_timeout(read, "read")
+        self.total = self._validate_timeout(total, "total")
+        self._start_connect: float | None = None
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}(connect={self._connect!r}, read={self._read!r}, total={self.total!r})"
+
+    # __str__ provided for backwards compatibility
+    __str__ = __repr__
+
+    @staticmethod
+    def resolve_default_timeout(timeout: _TYPE_TIMEOUT) -> float | None:
+        return getdefaulttimeout() if timeout is _DEFAULT_TIMEOUT else timeout
+
+    @classmethod
+    def _validate_timeout(cls, value: _TYPE_TIMEOUT, name: str) -> _TYPE_TIMEOUT:
+        """Check that a timeout attribute is valid.
+
+        :param value: The timeout value to validate
+        :param name: The name of the timeout attribute to validate. This is
+            used to specify in error messages.
+        :return: The validated and casted version of the given value.
+        :raises ValueError: If it is a numeric value less than or equal to
+            zero, or the type is not an integer, float, or None.
+        """
+        if value is None or value is _DEFAULT_TIMEOUT:
+            return value
+
+        if isinstance(value, bool):
+            raise ValueError(
+                "Timeout cannot be a boolean value. It must "
+                "be an int, float or None."
+            )
+        try:
+            float(value)
+        except (TypeError, ValueError):
+            raise ValueError(
+                "Timeout value %s was %s, but it must be an "
+                "int, float or None." % (name, value)
+            ) from None
+
+        try:
+            if value <= 0:
+                raise ValueError(
+                    "Attempted to set %s timeout to %s, but the "
+                    "timeout cannot be set to a value less "
+                    "than or equal to 0." % (name, value)
+                )
+        except TypeError:
+            raise ValueError(
+                "Timeout value %s was %s, but it must be an "
+                "int, float or None." % (name, value)
+            ) from None
+
+        return value
+
+    @classmethod
+    def from_float(cls, timeout: _TYPE_TIMEOUT) -> Timeout:
+        """Create a new Timeout from a legacy timeout value.
+
+        The timeout value used by httplib.py sets the same timeout on the
+        connect(), and recv() socket requests. This creates a :class:`Timeout`
+        object that sets the individual timeouts to the ``timeout`` value
+        passed to this function.
+
+        :param timeout: The legacy timeout value.
+        :type timeout: integer, float, :attr:`urllib3.util.Timeout.DEFAULT_TIMEOUT`, or None
+        :return: Timeout object
+        :rtype: :class:`Timeout`
+        """
+        return Timeout(read=timeout, connect=timeout)
+
+    def clone(self) -> Timeout:
+        """Create a copy of the timeout object
+
+        Timeout properties are stored per-pool but each request needs a fresh
+        Timeout object to ensure each one has its own start/stop configured.
+
+        :return: a copy of the timeout object
+        :rtype: :class:`Timeout`
+        """
+        # We can't use copy.deepcopy because that will also create a new object
+        # for _GLOBAL_DEFAULT_TIMEOUT, which socket.py uses as a sentinel to
+        # detect the user default.
+        return Timeout(connect=self._connect, read=self._read, total=self.total)
+
+    def start_connect(self) -> float:
+        """Start the timeout clock, used during a connect() attempt
+
+        :raises urllib3.exceptions.TimeoutStateError: if you attempt
+            to start a timer that has been started already.
+        """
+        if self._start_connect is not None:
+            raise TimeoutStateError("Timeout timer has already been started.")
+        self._start_connect = time.monotonic()
+        return self._start_connect
+
+    def get_connect_duration(self) -> float:
+        """Gets the time elapsed since the call to :meth:`start_connect`.
+
+        :return: Elapsed time in seconds.
+        :rtype: float
+        :raises urllib3.exceptions.TimeoutStateError: if you attempt
+            to get duration for a timer that hasn't been started.
+        """
+        if self._start_connect is None:
+            raise TimeoutStateError(
+                "Can't get connect duration for timer that has not started."
+            )
+        return time.monotonic() - self._start_connect
+
+    @property
+    def connect_timeout(self) -> _TYPE_TIMEOUT:
+        """Get the value to use when setting a connection timeout.
+
+        This will be a positive float or integer, the value None
+        (never timeout), or the default system timeout.
+
+        :return: Connect timeout.
+        :rtype: int, float, :attr:`Timeout.DEFAULT_TIMEOUT` or None
+        """
+        if self.total is None:
+            return self._connect
+
+        if self._connect is None or self._connect is _DEFAULT_TIMEOUT:
+            return self.total
+
+        return min(self._connect, self.total)  # type: ignore[type-var]
+
+    @property
+    def read_timeout(self) -> float | None:
+        """Get the value for the read timeout.
+
+        This assumes some time has elapsed in the connection timeout and
+        computes the read timeout appropriately.
+
+        If self.total is set, the read timeout is dependent on the amount of
+        time taken by the connect timeout. If the connection time has not been
+        established, a :exc:`~urllib3.exceptions.TimeoutStateError` will be
+        raised.
+
+        :return: Value to use for the read timeout.
+        :rtype: int, float or None
+        :raises urllib3.exceptions.TimeoutStateError: If :meth:`start_connect`
+            has not yet been called on this object.
+        """
+        if (
+            self.total is not None
+            and self.total is not _DEFAULT_TIMEOUT
+            and self._read is not None
+            and self._read is not _DEFAULT_TIMEOUT
+        ):
+            # In case the connect timeout has not yet been established.
+            if self._start_connect is None:
+                return self._read
+            return max(0, min(self.total - self.get_connect_duration(), self._read))
+        elif self.total is not None and self.total is not _DEFAULT_TIMEOUT:
+            return max(0, self.total - self.get_connect_duration())
+        else:
+            return self.resolve_default_timeout(self._read)
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/url.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/url.py
new file mode 100644
index 00000000..db057f17
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/url.py
@@ -0,0 +1,469 @@
+from __future__ import annotations
+
+import re
+import typing
+
+from ..exceptions import LocationParseError
+from .util import to_str
+
+# We only want to normalize urls with an HTTP(S) scheme.
+# urllib3 infers URLs without a scheme (None) to be http.
+_NORMALIZABLE_SCHEMES = ("http", "https", None)
+
+# Almost all of these patterns were derived from the
+# 'rfc3986' module: https://github.com/python-hyper/rfc3986
+_PERCENT_RE = re.compile(r"%[a-fA-F0-9]{2}")
+_SCHEME_RE = re.compile(r"^(?:[a-zA-Z][a-zA-Z0-9+-]*:|/)")
+_URI_RE = re.compile(
+    r"^(?:([a-zA-Z][a-zA-Z0-9+.-]*):)?"
+    r"(?://([^\\/?#]*))?"
+    r"([^?#]*)"
+    r"(?:\?([^#]*))?"
+    r"(?:#(.*))?$",
+    re.UNICODE | re.DOTALL,
+)
+
+_IPV4_PAT = r"(?:[0-9]{1,3}\.){3}[0-9]{1,3}"
+_HEX_PAT = "[0-9A-Fa-f]{1,4}"
+_LS32_PAT = "(?:{hex}:{hex}|{ipv4})".format(hex=_HEX_PAT, ipv4=_IPV4_PAT)
+_subs = {"hex": _HEX_PAT, "ls32": _LS32_PAT}
+_variations = [
+    #                            6( h16 ":" ) ls32
+    "(?:%(hex)s:){6}%(ls32)s",
+    #                       "::" 5( h16 ":" ) ls32
+    "::(?:%(hex)s:){5}%(ls32)s",
+    # [               h16 ] "::" 4( h16 ":" ) ls32
+    "(?:%(hex)s)?::(?:%(hex)s:){4}%(ls32)s",
+    # [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
+    "(?:(?:%(hex)s:)?%(hex)s)?::(?:%(hex)s:){3}%(ls32)s",
+    # [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
+    "(?:(?:%(hex)s:){0,2}%(hex)s)?::(?:%(hex)s:){2}%(ls32)s",
+    # [ *3( h16 ":" ) h16 ] "::"    h16 ":"   ls32
+    "(?:(?:%(hex)s:){0,3}%(hex)s)?::%(hex)s:%(ls32)s",
+    # [ *4( h16 ":" ) h16 ] "::"              ls32
+    "(?:(?:%(hex)s:){0,4}%(hex)s)?::%(ls32)s",
+    # [ *5( h16 ":" ) h16 ] "::"              h16
+    "(?:(?:%(hex)s:){0,5}%(hex)s)?::%(hex)s",
+    # [ *6( h16 ":" ) h16 ] "::"
+    "(?:(?:%(hex)s:){0,6}%(hex)s)?::",
+]
+
+_UNRESERVED_PAT = r"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789._\-~"
+_IPV6_PAT = "(?:" + "|".join([x % _subs for x in _variations]) + ")"
+_ZONE_ID_PAT = "(?:%25|%)(?:[" + _UNRESERVED_PAT + "]|%[a-fA-F0-9]{2})+"
+_IPV6_ADDRZ_PAT = r"\[" + _IPV6_PAT + r"(?:" + _ZONE_ID_PAT + r")?\]"
+_REG_NAME_PAT = r"(?:[^\[\]%:/?#]|%[a-fA-F0-9]{2})*"
+_TARGET_RE = re.compile(r"^(/[^?#]*)(?:\?([^#]*))?(?:#.*)?$")
+
+_IPV4_RE = re.compile("^" + _IPV4_PAT + "$")
+_IPV6_RE = re.compile("^" + _IPV6_PAT + "$")
+_IPV6_ADDRZ_RE = re.compile("^" + _IPV6_ADDRZ_PAT + "$")
+_BRACELESS_IPV6_ADDRZ_RE = re.compile("^" + _IPV6_ADDRZ_PAT[2:-2] + "$")
+_ZONE_ID_RE = re.compile("(" + _ZONE_ID_PAT + r")\]$")
+
+_HOST_PORT_PAT = ("^(%s|%s|%s)(?::0*?(|0|[1-9][0-9]{0,4}))?$") % (
+    _REG_NAME_PAT,
+    _IPV4_PAT,
+    _IPV6_ADDRZ_PAT,
+)
+_HOST_PORT_RE = re.compile(_HOST_PORT_PAT, re.UNICODE | re.DOTALL)
+
+_UNRESERVED_CHARS = set(
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789._-~"
+)
+_SUB_DELIM_CHARS = set("!$&'()*+,;=")
+_USERINFO_CHARS = _UNRESERVED_CHARS | _SUB_DELIM_CHARS | {":"}
+_PATH_CHARS = _USERINFO_CHARS | {"@", "/"}
+_QUERY_CHARS = _FRAGMENT_CHARS = _PATH_CHARS | {"?"}
+
+
+class Url(
+    typing.NamedTuple(
+        "Url",
+        [
+            ("scheme", typing.Optional[str]),
+            ("auth", typing.Optional[str]),
+            ("host", typing.Optional[str]),
+            ("port", typing.Optional[int]),
+            ("path", typing.Optional[str]),
+            ("query", typing.Optional[str]),
+            ("fragment", typing.Optional[str]),
+        ],
+    )
+):
+    """
+    Data structure for representing an HTTP URL. Used as a return value for
+    :func:`parse_url`. Both the scheme and host are normalized as they are
+    both case-insensitive according to RFC 3986.
+    """
+
+    def __new__(  # type: ignore[no-untyped-def]
+        cls,
+        scheme: str | None = None,
+        auth: str | None = None,
+        host: str | None = None,
+        port: int | None = None,
+        path: str | None = None,
+        query: str | None = None,
+        fragment: str | None = None,
+    ):
+        if path and not path.startswith("/"):
+            path = "/" + path
+        if scheme is not None:
+            scheme = scheme.lower()
+        return super().__new__(cls, scheme, auth, host, port, path, query, fragment)
+
+    @property
+    def hostname(self) -> str | None:
+        """For backwards-compatibility with urlparse. We're nice like that."""
+        return self.host
+
+    @property
+    def request_uri(self) -> str:
+        """Absolute path including the query string."""
+        uri = self.path or "/"
+
+        if self.query is not None:
+            uri += "?" + self.query
+
+        return uri
+
+    @property
+    def authority(self) -> str | None:
+        """
+        Authority component as defined in RFC 3986 3.2.
+        This includes userinfo (auth), host and port.
+
+        i.e.
+            userinfo@host:port
+        """
+        userinfo = self.auth
+        netloc = self.netloc
+        if netloc is None or userinfo is None:
+            return netloc
+        else:
+            return f"{userinfo}@{netloc}"
+
+    @property
+    def netloc(self) -> str | None:
+        """
+        Network location including host and port.
+
+        If you need the equivalent of urllib.parse's ``netloc``,
+        use the ``authority`` property instead.
+        """
+        if self.host is None:
+            return None
+        if self.port:
+            return f"{self.host}:{self.port}"
+        return self.host
+
+    @property
+    def url(self) -> str:
+        """
+        Convert self into a url
+
+        This function should more or less round-trip with :func:`.parse_url`. The
+        returned url may not be exactly the same as the url inputted to
+        :func:`.parse_url`, but it should be equivalent by the RFC (e.g., urls
+        with a blank port will have : removed).
+
+        Example:
+
+        .. code-block:: python
+
+            import urllib3
+
+            U = urllib3.util.parse_url("https://google.com/mail/")
+
+            print(U.url)
+            # "https://google.com/mail/"
+
+            print( urllib3.util.Url("https", "username:password",
+                                    "host.com", 80, "/path", "query", "fragment"
+                                    ).url
+                )
+            # "https://username:password@host.com:80/path?query#fragment"
+        """
+        scheme, auth, host, port, path, query, fragment = self
+        url = ""
+
+        # We use "is not None" we want things to happen with empty strings (or 0 port)
+        if scheme is not None:
+            url += scheme + "://"
+        if auth is not None:
+            url += auth + "@"
+        if host is not None:
+            url += host
+        if port is not None:
+            url += ":" + str(port)
+        if path is not None:
+            url += path
+        if query is not None:
+            url += "?" + query
+        if fragment is not None:
+            url += "#" + fragment
+
+        return url
+
+    def __str__(self) -> str:
+        return self.url
+
+
+@typing.overload
+def _encode_invalid_chars(
+    component: str, allowed_chars: typing.Container[str]
+) -> str:  # Abstract
+    ...
+
+
+@typing.overload
+def _encode_invalid_chars(
+    component: None, allowed_chars: typing.Container[str]
+) -> None:  # Abstract
+    ...
+
+
+def _encode_invalid_chars(
+    component: str | None, allowed_chars: typing.Container[str]
+) -> str | None:
+    """Percent-encodes a URI component without reapplying
+    onto an already percent-encoded component.
+    """
+    if component is None:
+        return component
+
+    component = to_str(component)
+
+    # Normalize existing percent-encoded bytes.
+    # Try to see if the component we're encoding is already percent-encoded
+    # so we can skip all '%' characters but still encode all others.
+    component, percent_encodings = _PERCENT_RE.subn(
+        lambda match: match.group(0).upper(), component
+    )
+
+    uri_bytes = component.encode("utf-8", "surrogatepass")
+    is_percent_encoded = percent_encodings == uri_bytes.count(b"%")
+    encoded_component = bytearray()
+
+    for i in range(0, len(uri_bytes)):
+        # Will return a single character bytestring
+        byte = uri_bytes[i : i + 1]
+        byte_ord = ord(byte)
+        if (is_percent_encoded and byte == b"%") or (
+            byte_ord < 128 and byte.decode() in allowed_chars
+        ):
+            encoded_component += byte
+            continue
+        encoded_component.extend(b"%" + (hex(byte_ord)[2:].encode().zfill(2).upper()))
+
+    return encoded_component.decode()
+
+
+def _remove_path_dot_segments(path: str) -> str:
+    # See http://tools.ietf.org/html/rfc3986#section-5.2.4 for pseudo-code
+    segments = path.split("/")  # Turn the path into a list of segments
+    output = []  # Initialize the variable to use to store output
+
+    for segment in segments:
+        # '.' is the current directory, so ignore it, it is superfluous
+        if segment == ".":
+            continue
+        # Anything other than '..', should be appended to the output
+        if segment != "..":
+            output.append(segment)
+        # In this case segment == '..', if we can, we should pop the last
+        # element
+        elif output:
+            output.pop()
+
+    # If the path starts with '/' and the output is empty or the first string
+    # is non-empty
+    if path.startswith("/") and (not output or output[0]):
+        output.insert(0, "")
+
+    # If the path starts with '/.' or '/..' ensure we add one more empty
+    # string to add a trailing '/'
+    if path.endswith(("/.", "/..")):
+        output.append("")
+
+    return "/".join(output)
+
+
+@typing.overload
+def _normalize_host(host: None, scheme: str | None) -> None: ...
+
+
+@typing.overload
+def _normalize_host(host: str, scheme: str | None) -> str: ...
+
+
+def _normalize_host(host: str | None, scheme: str | None) -> str | None:
+    if host:
+        if scheme in _NORMALIZABLE_SCHEMES:
+            is_ipv6 = _IPV6_ADDRZ_RE.match(host)
+            if is_ipv6:
+                # IPv6 hosts of the form 'a::b%zone' are encoded in a URL as
+                # such per RFC 6874: 'a::b%25zone'. Unquote the ZoneID
+                # separator as necessary to return a valid RFC 4007 scoped IP.
+                match = _ZONE_ID_RE.search(host)
+                if match:
+                    start, end = match.span(1)
+                    zone_id = host[start:end]
+
+                    if zone_id.startswith("%25") and zone_id != "%25":
+                        zone_id = zone_id[3:]
+                    else:
+                        zone_id = zone_id[1:]
+                    zone_id = _encode_invalid_chars(zone_id, _UNRESERVED_CHARS)
+                    return f"{host[:start].lower()}%{zone_id}{host[end:]}"
+                else:
+                    return host.lower()
+            elif not _IPV4_RE.match(host):
+                return to_str(
+                    b".".join([_idna_encode(label) for label in host.split(".")]),
+                    "ascii",
+                )
+    return host
+
+
+def _idna_encode(name: str) -> bytes:
+    if not name.isascii():
+        try:
+            import idna
+        except ImportError:
+            raise LocationParseError(
+                "Unable to parse URL without the 'idna' module"
+            ) from None
+
+        try:
+            return idna.encode(name.lower(), strict=True, std3_rules=True)
+        except idna.IDNAError:
+            raise LocationParseError(
+                f"Name '{name}' is not a valid IDNA label"
+            ) from None
+
+    return name.lower().encode("ascii")
+
+
+def _encode_target(target: str) -> str:
+    """Percent-encodes a request target so that there are no invalid characters
+
+    Pre-condition for this function is that 'target' must start with '/'.
+    If that is the case then _TARGET_RE will always produce a match.
+    """
+    match = _TARGET_RE.match(target)
+    if not match:  # Defensive:
+        raise LocationParseError(f"{target!r} is not a valid request URI")
+
+    path, query = match.groups()
+    encoded_target = _encode_invalid_chars(path, _PATH_CHARS)
+    if query is not None:
+        query = _encode_invalid_chars(query, _QUERY_CHARS)
+        encoded_target += "?" + query
+    return encoded_target
+
+
+def parse_url(url: str) -> Url:
+    """
+    Given a url, return a parsed :class:`.Url` namedtuple. Best-effort is
+    performed to parse incomplete urls. Fields not provided will be None.
+    This parser is RFC 3986 and RFC 6874 compliant.
+
+    The parser logic and helper functions are based heavily on
+    work done in the ``rfc3986`` module.
+
+    :param str url: URL to parse into a :class:`.Url` namedtuple.
+
+    Partly backwards-compatible with :mod:`urllib.parse`.
+
+    Example:
+
+    .. code-block:: python
+
+        import urllib3
+
+        print( urllib3.util.parse_url('http://google.com/mail/'))
+        # Url(scheme='http', host='google.com', port=None, path='/mail/', ...)
+
+        print( urllib3.util.parse_url('google.com:80'))
+        # Url(scheme=None, host='google.com', port=80, path=None, ...)
+
+        print( urllib3.util.parse_url('/foo?bar'))
+        # Url(scheme=None, host=None, port=None, path='/foo', query='bar', ...)
+    """
+    if not url:
+        # Empty
+        return Url()
+
+    source_url = url
+    if not _SCHEME_RE.search(url):
+        url = "//" + url
+
+    scheme: str | None
+    authority: str | None
+    auth: str | None
+    host: str | None
+    port: str | None
+    port_int: int | None
+    path: str | None
+    query: str | None
+    fragment: str | None
+
+    try:
+        scheme, authority, path, query, fragment = _URI_RE.match(url).groups()  # type: ignore[union-attr]
+        normalize_uri = scheme is None or scheme.lower() in _NORMALIZABLE_SCHEMES
+
+        if scheme:
+            scheme = scheme.lower()
+
+        if authority:
+            auth, _, host_port = authority.rpartition("@")
+            auth = auth or None
+            host, port = _HOST_PORT_RE.match(host_port).groups()  # type: ignore[union-attr]
+            if auth and normalize_uri:
+                auth = _encode_invalid_chars(auth, _USERINFO_CHARS)
+            if port == "":
+                port = None
+        else:
+            auth, host, port = None, None, None
+
+        if port is not None:
+            port_int = int(port)
+            if not (0 <= port_int <= 65535):
+                raise LocationParseError(url)
+        else:
+            port_int = None
+
+        host = _normalize_host(host, scheme)
+
+        if normalize_uri and path:
+            path = _remove_path_dot_segments(path)
+            path = _encode_invalid_chars(path, _PATH_CHARS)
+        if normalize_uri and query:
+            query = _encode_invalid_chars(query, _QUERY_CHARS)
+        if normalize_uri and fragment:
+            fragment = _encode_invalid_chars(fragment, _FRAGMENT_CHARS)
+
+    except (ValueError, AttributeError) as e:
+        raise LocationParseError(source_url) from e
+
+    # For the sake of backwards compatibility we put empty
+    # string values for path if there are any defined values
+    # beyond the path in the URL.
+    # TODO: Remove this when we break backwards compatibility.
+    if not path:
+        if query is not None or fragment is not None:
+            path = ""
+        else:
+            path = None
+
+    return Url(
+        scheme=scheme,
+        auth=auth,
+        host=host,
+        port=port_int,
+        path=path,
+        query=query,
+        fragment=fragment,
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/util.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/util.py
new file mode 100644
index 00000000..35c77e40
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/util.py
@@ -0,0 +1,42 @@
+from __future__ import annotations
+
+import typing
+from types import TracebackType
+
+
+def to_bytes(
+    x: str | bytes, encoding: str | None = None, errors: str | None = None
+) -> bytes:
+    if isinstance(x, bytes):
+        return x
+    elif not isinstance(x, str):
+        raise TypeError(f"not expecting type {type(x).__name__}")
+    if encoding or errors:
+        return x.encode(encoding or "utf-8", errors=errors or "strict")
+    return x.encode()
+
+
+def to_str(
+    x: str | bytes, encoding: str | None = None, errors: str | None = None
+) -> str:
+    if isinstance(x, str):
+        return x
+    elif not isinstance(x, bytes):
+        raise TypeError(f"not expecting type {type(x).__name__}")
+    if encoding or errors:
+        return x.decode(encoding or "utf-8", errors=errors or "strict")
+    return x.decode()
+
+
+def reraise(
+    tp: type[BaseException] | None,
+    value: BaseException,
+    tb: TracebackType | None = None,
+) -> typing.NoReturn:
+    try:
+        if value.__traceback__ is not tb:
+            raise value.with_traceback(tb)
+        raise value
+    finally:
+        value = None  # type: ignore[assignment]
+        tb = None
diff --git a/backend/venv/lib/python3.12/site-packages/urllib3/util/wait.py b/backend/venv/lib/python3.12/site-packages/urllib3/util/wait.py
new file mode 100644
index 00000000..aeca0c7a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/urllib3/util/wait.py
@@ -0,0 +1,124 @@
+from __future__ import annotations
+
+import select
+import socket
+from functools import partial
+
+__all__ = ["wait_for_read", "wait_for_write"]
+
+
+# How should we wait on sockets?
+#
+# There are two types of APIs you can use for waiting on sockets: the fancy
+# modern stateful APIs like epoll/kqueue, and the older stateless APIs like
+# select/poll. The stateful APIs are more efficient when you have a lots of
+# sockets to keep track of, because you can set them up once and then use them
+# lots of times. But we only ever want to wait on a single socket at a time
+# and don't want to keep track of state, so the stateless APIs are actually
+# more efficient. So we want to use select() or poll().
+#
+# Now, how do we choose between select() and poll()? On traditional Unixes,
+# select() has a strange calling convention that makes it slow, or fail
+# altogether, for high-numbered file descriptors. The point of poll() is to fix
+# that, so on Unixes, we prefer poll().
+#
+# On Windows, there is no poll() (or at least Python doesn't provide a wrapper
+# for it), but that's OK, because on Windows, select() doesn't have this
+# strange calling convention; plain select() works fine.
+#
+# So: on Windows we use select(), and everywhere else we use poll(). We also
+# fall back to select() in case poll() is somehow broken or missing.
+
+
+def select_wait_for_socket(
+    sock: socket.socket,
+    read: bool = False,
+    write: bool = False,
+    timeout: float | None = None,
+) -> bool:
+    if not read and not write:
+        raise RuntimeError("must specify at least one of read=True, write=True")
+    rcheck = []
+    wcheck = []
+    if read:
+        rcheck.append(sock)
+    if write:
+        wcheck.append(sock)
+    # When doing a non-blocking connect, most systems signal success by
+    # marking the socket writable. Windows, though, signals success by marked
+    # it as "exceptional". We paper over the difference by checking the write
+    # sockets for both conditions. (The stdlib selectors module does the same
+    # thing.)
+    fn = partial(select.select, rcheck, wcheck, wcheck)
+    rready, wready, xready = fn(timeout)
+    return bool(rready or wready or xready)
+
+
+def poll_wait_for_socket(
+    sock: socket.socket,
+    read: bool = False,
+    write: bool = False,
+    timeout: float | None = None,
+) -> bool:
+    if not read and not write:
+        raise RuntimeError("must specify at least one of read=True, write=True")
+    mask = 0
+    if read:
+        mask |= select.POLLIN
+    if write:
+        mask |= select.POLLOUT
+    poll_obj = select.poll()
+    poll_obj.register(sock, mask)
+
+    # For some reason, poll() takes timeout in milliseconds
+    def do_poll(t: float | None) -> list[tuple[int, int]]:
+        if t is not None:
+            t *= 1000
+        return poll_obj.poll(t)
+
+    return bool(do_poll(timeout))
+
+
+def _have_working_poll() -> bool:
+    # Apparently some systems have a select.poll that fails as soon as you try
+    # to use it, either due to strange configuration or broken monkeypatching
+    # from libraries like eventlet/greenlet.
+    try:
+        poll_obj = select.poll()
+        poll_obj.poll(0)
+    except (AttributeError, OSError):
+        return False
+    else:
+        return True
+
+
+def wait_for_socket(
+    sock: socket.socket,
+    read: bool = False,
+    write: bool = False,
+    timeout: float | None = None,
+) -> bool:
+    # We delay choosing which implementation to use until the first time we're
+    # called. We could do it at import time, but then we might make the wrong
+    # decision if someone goes wild with monkeypatching select.poll after
+    # we're imported.
+    global wait_for_socket
+    if _have_working_poll():
+        wait_for_socket = poll_wait_for_socket
+    elif hasattr(select, "select"):
+        wait_for_socket = select_wait_for_socket
+    return wait_for_socket(sock, read, write, timeout)
+
+
+def wait_for_read(sock: socket.socket, timeout: float | None = None) -> bool:
+    """Waits for reading to be available on a given socket.
+    Returns True if the socket is readable, or False if the timeout expired.
+    """
+    return wait_for_socket(sock, read=True, timeout=timeout)
+
+
+def wait_for_write(sock: socket.socket, timeout: float | None = None) -> bool:
+    """Waits for writing to be available on a given socket.
+    Returns True if the socket is readable, or False if the timeout expired.
+    """
+    return wait_for_socket(sock, write=True, timeout=timeout)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/INSTALLER b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/INSTALLER
new file mode 100644
index 00000000..a1b589e3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/INSTALLER
@@ -0,0 +1 @@
+pip
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/LICENSE.txt b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/LICENSE.txt
new file mode 100644
index 00000000..c37cae49
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/LICENSE.txt
@@ -0,0 +1,28 @@
+Copyright 2007 Pallets
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+1.  Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+
+2.  Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+
+3.  Neither the name of the copyright holder nor the names of its
+    contributors may be used to endorse or promote products derived from
+    this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
+TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/METADATA b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/METADATA
new file mode 100644
index 00000000..e495b3d2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/METADATA
@@ -0,0 +1,99 @@
+Metadata-Version: 2.3
+Name: Werkzeug
+Version: 3.1.3
+Summary: The comprehensive WSGI web application library.
+Maintainer-email: Pallets <contact@palletsprojects.com>
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Environment :: Web Environment
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python
+Classifier: Topic :: Internet :: WWW/HTTP :: Dynamic Content
+Classifier: Topic :: Internet :: WWW/HTTP :: WSGI
+Classifier: Topic :: Internet :: WWW/HTTP :: WSGI :: Application
+Classifier: Topic :: Internet :: WWW/HTTP :: WSGI :: Middleware
+Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
+Classifier: Typing :: Typed
+Requires-Dist: MarkupSafe>=2.1.1
+Requires-Dist: watchdog>=2.3 ; extra == "watchdog"
+Project-URL: Changes, https://werkzeug.palletsprojects.com/changes/
+Project-URL: Chat, https://discord.gg/pallets
+Project-URL: Documentation, https://werkzeug.palletsprojects.com/
+Project-URL: Donate, https://palletsprojects.com/donate
+Project-URL: Issue Tracker, https://github.com/pallets/werkzeug/issues/
+Project-URL: Source Code, https://github.com/pallets/werkzeug/
+Provides-Extra: watchdog
+
+# Werkzeug
+
+*werkzeug* German noun: "tool". Etymology: *werk* ("work"), *zeug* ("stuff")
+
+Werkzeug is a comprehensive [WSGI][] web application library. It began as
+a simple collection of various utilities for WSGI applications and has
+become one of the most advanced WSGI utility libraries.
+
+It includes:
+
+-   An interactive debugger that allows inspecting stack traces and
+    source code in the browser with an interactive interpreter for any
+    frame in the stack.
+-   A full-featured request object with objects to interact with
+    headers, query args, form data, files, and cookies.
+-   A response object that can wrap other WSGI applications and handle
+    streaming data.
+-   A routing system for matching URLs to endpoints and generating URLs
+    for endpoints, with an extensible system for capturing variables
+    from URLs.
+-   HTTP utilities to handle entity tags, cache control, dates, user
+    agents, cookies, files, and more.
+-   A threaded WSGI server for use while developing applications
+    locally.
+-   A test client for simulating HTTP requests during testing without
+    requiring running a server.
+
+Werkzeug doesn't enforce any dependencies. It is up to the developer to
+choose a template engine, database adapter, and even how to handle
+requests. It can be used to build all sorts of end user applications
+such as blogs, wikis, or bulletin boards.
+
+[Flask][] wraps Werkzeug, using it to handle the details of WSGI while
+providing more structure and patterns for defining powerful
+applications.
+
+[WSGI]: https://wsgi.readthedocs.io/en/latest/
+[Flask]: https://www.palletsprojects.com/p/flask/
+
+
+## A Simple Example
+
+```python
+# save this as app.py
+from werkzeug.wrappers import Request, Response
+
+@Request.application
+def application(request: Request) -> Response:
+    return Response("Hello, World!")
+
+if __name__ == "__main__":
+    from werkzeug.serving import run_simple
+    run_simple("127.0.0.1", 5000, application)
+```
+
+```
+$ python -m app
+  * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
+```
+
+
+## Donate
+
+The Pallets organization develops and supports Werkzeug and other
+popular packages. In order to grow the community of contributors and
+users, and allow the maintainers to devote more time to the projects,
+[please donate today][].
+
+[please donate today]: https://palletsprojects.com/donate
+
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/RECORD b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/RECORD
new file mode 100644
index 00000000..74e05610
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/RECORD
@@ -0,0 +1,117 @@
+werkzeug-3.1.3.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+werkzeug-3.1.3.dist-info/LICENSE.txt,sha256=O0nc7kEF6ze6wQ-vG-JgQI_oXSUrjp3y4JefweCUQ3s,1475
+werkzeug-3.1.3.dist-info/METADATA,sha256=9d8uO6J-ck8mHyFjxFUs5nCVy5iN3XdTYcJfz2QBr0I,3682
+werkzeug-3.1.3.dist-info/RECORD,,
+werkzeug-3.1.3.dist-info/REQUESTED,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+werkzeug-3.1.3.dist-info/WHEEL,sha256=CpUCUxeHQbRN5UGRQHYRJorO5Af-Qy_fHMctcQ8DSGI,82
+werkzeug/__init__.py,sha256=CejNWfCZKDaJ_FmshZFw8dFgjNAlu6q15ec7DwWJlM8,165
+werkzeug/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/__pycache__/_internal.cpython-312.pyc,,
+werkzeug/__pycache__/_reloader.cpython-312.pyc,,
+werkzeug/__pycache__/exceptions.cpython-312.pyc,,
+werkzeug/__pycache__/formparser.cpython-312.pyc,,
+werkzeug/__pycache__/http.cpython-312.pyc,,
+werkzeug/__pycache__/local.cpython-312.pyc,,
+werkzeug/__pycache__/security.cpython-312.pyc,,
+werkzeug/__pycache__/serving.cpython-312.pyc,,
+werkzeug/__pycache__/test.cpython-312.pyc,,
+werkzeug/__pycache__/testapp.cpython-312.pyc,,
+werkzeug/__pycache__/urls.cpython-312.pyc,,
+werkzeug/__pycache__/user_agent.cpython-312.pyc,,
+werkzeug/__pycache__/utils.cpython-312.pyc,,
+werkzeug/__pycache__/wsgi.cpython-312.pyc,,
+werkzeug/_internal.py,sha256=su1olkbHMkzt0VKcEkPLCha8sdVzXNBuqW6YVpp8GHg,5545
+werkzeug/_reloader.py,sha256=QuMO-UwuD-cHsUpwSEUr7iYNbv_ziHFhgNDAiv25J80,15429
+werkzeug/datastructures/__init__.py,sha256=0Vt8Lt8KUYSoAe8ADt1jIE-w3Ib9tkrk0HEhBOktqUw,2615
+werkzeug/datastructures/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/accept.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/auth.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/cache_control.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/csp.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/etag.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/file_storage.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/headers.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/mixins.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/range.cpython-312.pyc,,
+werkzeug/datastructures/__pycache__/structures.cpython-312.pyc,,
+werkzeug/datastructures/accept.py,sha256=xoDD4qwZeTjew_1Ze0gyyKWkbhtowwR5tHTWnobPeGI,12177
+werkzeug/datastructures/auth.py,sha256=Zcqdr8-uUpq6_czg1672S2XthG0UakhUHOuummzqz64,10055
+werkzeug/datastructures/cache_control.py,sha256=hh-0nlj8CWDTrWcDvprMkmHZ8HW_Afo4853vl0LuFU4,9697
+werkzeug/datastructures/csp.py,sha256=2UDhFNqBzjhpA45m11HDg6LnhBSSWWdkfFzPq7ikfho,3816
+werkzeug/datastructures/etag.py,sha256=N2dFqhFsLR0k5hgjol3QTGpcSltJhpgLWeyz6Va80G8,3278
+werkzeug/datastructures/file_storage.py,sha256=2P-4PEkR_kY37g3hc1S30YH_15VrwRnQ13sBFreZyc4,6701
+werkzeug/datastructures/headers.py,sha256=_VjcQCrLxjNKxPL8xSoFnEJnZ5tRc1zdL9spQf4c758,21554
+werkzeug/datastructures/mixins.py,sha256=sYJeA8kZuAOvzlfrRWRN2dyuNrBev4yFxWH8coZyrec,9027
+werkzeug/datastructures/range.py,sha256=5c1Mo5jBkkVhQyLoD9lPdjETKo0iw9g-lFYOqF1ZQvY,7020
+werkzeug/datastructures/structures.py,sha256=IG5WpAj5RZSUfyxYXXHvIgtlOx5PQeZ9YmQO87FJr08,41359
+werkzeug/debug/__init__.py,sha256=80vaavVZYwV21ItrsOLQudEUi2uKDlxcjm5cCmeU1u4,20016
+werkzeug/debug/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/debug/__pycache__/console.cpython-312.pyc,,
+werkzeug/debug/__pycache__/repr.cpython-312.pyc,,
+werkzeug/debug/__pycache__/tbtools.cpython-312.pyc,,
+werkzeug/debug/console.py,sha256=t4hZ0Qg1p6Uu2MWimqoMDi7S3WYZvLMjnc8v_dPaxAo,6089
+werkzeug/debug/repr.py,sha256=iHMYny8whiiMDasvUqj0nm4-1VHVvwe697KleiZVK1s,9303
+werkzeug/debug/shared/ICON_LICENSE.md,sha256=DhA6Y1gUl5Jwfg0NFN9Rj4VWITt8tUx0IvdGf0ux9-s,222
+werkzeug/debug/shared/console.png,sha256=bxax6RXXlvOij_KeqvSNX0ojJf83YbnZ7my-3Gx9w2A,507
+werkzeug/debug/shared/debugger.js,sha256=SRL9YZ9FTVngaYD-INQNincEVdZ-kBHa_-VJx0U7-rg,10068
+werkzeug/debug/shared/less.png,sha256=-4-kNRaXJSONVLahrQKUxMwXGm9R4OnZ9SxDGpHlIR4,191
+werkzeug/debug/shared/more.png,sha256=GngN7CioHQoV58rH6ojnkYi8c_qED2Aka5FO5UXrReY,200
+werkzeug/debug/shared/style.css,sha256=-xSxzUEZGw_IqlDR5iZxitNl8LQUjBM-_Y4UAvXVH8g,6078
+werkzeug/debug/tbtools.py,sha256=4JDhZT_2gm9sMgfiMieEyxRsV196i3kp_0b3MOjy1KY,13560
+werkzeug/exceptions.py,sha256=cpwKKTfe1_yFrBpRHjp2bqwDhtUdHnGoDKcWOXR_SIE,26455
+werkzeug/formparser.py,sha256=v0-F9OcmNHuPLqXePW_b0asWzpH-Z_ZlGj8ao0CMmuo,15847
+werkzeug/http.py,sha256=PRpjIrbvIIiAMlmxwVIo9MMxaNX8vSOSn58rwAThEWM,44880
+werkzeug/local.py,sha256=KUFuAm8BAayQouzVg0MGqW_hiwY8Z_lY5l7d1Scvsx8,22492
+werkzeug/middleware/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+werkzeug/middleware/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/dispatcher.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/http_proxy.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/lint.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/profiler.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/proxy_fix.cpython-312.pyc,,
+werkzeug/middleware/__pycache__/shared_data.cpython-312.pyc,,
+werkzeug/middleware/dispatcher.py,sha256=zWN5_lqJr_sc9UDv-PPoSlDHN_zR33z6B74F_4Cxpo8,2602
+werkzeug/middleware/http_proxy.py,sha256=sdk-V6GoZ6aMny-D0QNKNf5MWD2OTO3AGbBg6upp4Hc,7834
+werkzeug/middleware/lint.py,sha256=jVdyljRxpkz18MiMR8N7YdiElna8DJVvO4CCjZsnSQI,14476
+werkzeug/middleware/profiler.py,sha256=1ZAHlDeYNdhgp8THOXkV5lgmcLl307phAr2Ufy30-lY,5562
+werkzeug/middleware/proxy_fix.py,sha256=n-HW-MRWJquCIhmqiZKoGdbbEeHuWJqPRHhFpuj4pzY,6755
+werkzeug/middleware/shared_data.py,sha256=gsjmgeEFLzDdKsgd19pMhI7BfU_TbL6AyEyZPG2ykfQ,9542
+werkzeug/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+werkzeug/routing/__init__.py,sha256=d8TRxsk24IWu2BdoOYUfL--deolHwiGVCBJqLoEe3YM,4820
+werkzeug/routing/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/routing/__pycache__/converters.cpython-312.pyc,,
+werkzeug/routing/__pycache__/exceptions.cpython-312.pyc,,
+werkzeug/routing/__pycache__/map.cpython-312.pyc,,
+werkzeug/routing/__pycache__/matcher.cpython-312.pyc,,
+werkzeug/routing/__pycache__/rules.cpython-312.pyc,,
+werkzeug/routing/converters.py,sha256=iqpee_mAjr1oGbq0etujYF9PiDv5U7DgNkARHXnMId0,7297
+werkzeug/routing/exceptions.py,sha256=wNBiUmUk4OtFOpbdDSr7KKKUjH7yn84JqwBicUup8p8,4846
+werkzeug/routing/map.py,sha256=mEXlHOyinkg1Jtx5L0UDYsvoX4eVLiEuEVQzD5LVAz8,36515
+werkzeug/routing/matcher.py,sha256=nfBbl37eGAkZ1dQlumshFcPuyfggmFjPuSSQOE6GuYs,7849
+werkzeug/routing/rules.py,sha256=eGi6PD-COG2As_HY0nAw-nxYxLTH0FsuqRaSy8d9FjQ,32510
+werkzeug/sansio/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+werkzeug/sansio/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/sansio/__pycache__/http.cpython-312.pyc,,
+werkzeug/sansio/__pycache__/multipart.cpython-312.pyc,,
+werkzeug/sansio/__pycache__/request.cpython-312.pyc,,
+werkzeug/sansio/__pycache__/response.cpython-312.pyc,,
+werkzeug/sansio/__pycache__/utils.cpython-312.pyc,,
+werkzeug/sansio/http.py,sha256=OYlTJn8fJhHA0CDOOmerekpsyGyIQVm9LvVRO3lH0TA,5337
+werkzeug/sansio/multipart.py,sha256=fMGgY9CvMDRgNgKozz2cBIp0iN2EME2IIOBh306_Ecs,11637
+werkzeug/sansio/request.py,sha256=ijwestWQS-dpb61uKl0NRntzNtH6C72OXCNomx-XKxc,19832
+werkzeug/sansio/response.py,sha256=8pE2fiHDroOOCtBWJWig4Lnv3VztD5szkSzeQELIlcQ,27965
+werkzeug/sansio/utils.py,sha256=3-Gr0Qsa_Rx48mUm7zdFJZNNKMzHfPAknywujHnkhPs,5256
+werkzeug/security.py,sha256=WctngEBHBo8e-gXf6lMURhs1t0tmEjJjBYZi2MXoWss,5581
+werkzeug/serving.py,sha256=k30NzIKq6mkNONk7B5rTyEuLrLZo__Kz2_XGXLjVUic,39857
+werkzeug/test.py,sha256=2G08IExd_x3T1P69mz57LYYQrqk31nxmkgHY5M5HYZU,52795
+werkzeug/testapp.py,sha256=5_IS5Dh_WfWfNcTLmbydj01lomgcKA_4l9PPCNZnmdI,6332
+werkzeug/urls.py,sha256=XyNKwHvK5IC37-wuIDMYWkiCJ3yLTLGv7wn2GF3ndqI,6430
+werkzeug/user_agent.py,sha256=lSlLYKCcbzCUSkbdAoO8zPk2UR-8Mdn6iu_iA2kYPBA,1416
+werkzeug/utils.py,sha256=1StUqI45bcxJx4sj81qVPQH5X8VuWgtICViJI_BKg-c,24725
+werkzeug/wrappers/__init__.py,sha256=b78jCM8x96kJUGLZ5FYFR3zlK-3pnFAmP9RJIGU0ses,138
+werkzeug/wrappers/__pycache__/__init__.cpython-312.pyc,,
+werkzeug/wrappers/__pycache__/request.cpython-312.pyc,,
+werkzeug/wrappers/__pycache__/response.cpython-312.pyc,,
+werkzeug/wrappers/request.py,sha256=2eJKd8yBU4zc0FIm2T6SROeuZERkP9m_Yx1Z7MeMs-Y,24737
+werkzeug/wrappers/response.py,sha256=u6zg7VpNYrCeEjpIgf8VqgfaSi9yR_9wi9ly2uudglg,32459
+werkzeug/wsgi.py,sha256=P7jB0VpG6X6miies4uk7Zgm7NVm4Yz8Ra6Inr5q_FMs,20894
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/REQUESTED b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/REQUESTED
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/WHEEL b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/WHEEL
new file mode 100644
index 00000000..e3c6feef
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug-3.1.3.dist-info/WHEEL
@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: flit 3.10.1
+Root-Is-Purelib: true
+Tag: py3-none-any
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/__init__.py
new file mode 100644
index 00000000..0b248fd8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/__init__.py
@@ -0,0 +1,4 @@
+from .serving import run_simple as run_simple
+from .test import Client as Client
+from .wrappers import Request as Request
+from .wrappers import Response as Response
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..d4b16df3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_internal.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_internal.cpython-312.pyc
new file mode 100644
index 00000000..9a5b00cb
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_internal.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_reloader.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_reloader.cpython-312.pyc
new file mode 100644
index 00000000..a9e500a2
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/_reloader.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..ba13ec3a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/formparser.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/formparser.cpython-312.pyc
new file mode 100644
index 00000000..735d3540
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/formparser.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/http.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/http.cpython-312.pyc
new file mode 100644
index 00000000..c15353ea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/http.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/local.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/local.cpython-312.pyc
new file mode 100644
index 00000000..86481e4e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/local.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/security.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/security.cpython-312.pyc
new file mode 100644
index 00000000..75a2e955
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/security.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/serving.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/serving.cpython-312.pyc
new file mode 100644
index 00000000..4e679009
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/serving.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/test.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/test.cpython-312.pyc
new file mode 100644
index 00000000..ddd77a65
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/test.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/testapp.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/testapp.cpython-312.pyc
new file mode 100644
index 00000000..372ee210
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/testapp.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/urls.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/urls.cpython-312.pyc
new file mode 100644
index 00000000..db3f2d83
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/urls.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/user_agent.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/user_agent.cpython-312.pyc
new file mode 100644
index 00000000..1ece4cea
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/user_agent.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..2811029e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/wsgi.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/wsgi.cpython-312.pyc
new file mode 100644
index 00000000..1cbb400b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/__pycache__/wsgi.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/_internal.py b/backend/venv/lib/python3.12/site-packages/werkzeug/_internal.py
new file mode 100644
index 00000000..7dd2fbcc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/_internal.py
@@ -0,0 +1,211 @@
+from __future__ import annotations
+
+import logging
+import re
+import sys
+import typing as t
+from datetime import datetime
+from datetime import timezone
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .wrappers.request import Request
+
+_logger: logging.Logger | None = None
+
+
+class _Missing:
+    def __repr__(self) -> str:
+        return "no value"
+
+    def __reduce__(self) -> str:
+        return "_missing"
+
+
+_missing = _Missing()
+
+
+def _wsgi_decoding_dance(s: str) -> str:
+    return s.encode("latin1").decode(errors="replace")
+
+
+def _wsgi_encoding_dance(s: str) -> str:
+    return s.encode().decode("latin1")
+
+
+def _get_environ(obj: WSGIEnvironment | Request) -> WSGIEnvironment:
+    env = getattr(obj, "environ", obj)
+    assert isinstance(
+        env, dict
+    ), f"{type(obj).__name__!r} is not a WSGI environment (has to be a dict)"
+    return env
+
+
+def _has_level_handler(logger: logging.Logger) -> bool:
+    """Check if there is a handler in the logging chain that will handle
+    the given logger's effective level.
+    """
+    level = logger.getEffectiveLevel()
+    current = logger
+
+    while current:
+        if any(handler.level <= level for handler in current.handlers):
+            return True
+
+        if not current.propagate:
+            break
+
+        current = current.parent  # type: ignore
+
+    return False
+
+
+class _ColorStreamHandler(logging.StreamHandler):  # type: ignore[type-arg]
+    """On Windows, wrap stream with Colorama for ANSI style support."""
+
+    def __init__(self) -> None:
+        try:
+            import colorama
+        except ImportError:
+            stream = None
+        else:
+            stream = colorama.AnsiToWin32(sys.stderr)
+
+        super().__init__(stream)
+
+
+def _log(type: str, message: str, *args: t.Any, **kwargs: t.Any) -> None:
+    """Log a message to the 'werkzeug' logger.
+
+    The logger is created the first time it is needed. If there is no
+    level set, it is set to :data:`logging.INFO`. If there is no handler
+    for the logger's effective level, a :class:`logging.StreamHandler`
+    is added.
+    """
+    global _logger
+
+    if _logger is None:
+        _logger = logging.getLogger("werkzeug")
+
+        if _logger.level == logging.NOTSET:
+            _logger.setLevel(logging.INFO)
+
+        if not _has_level_handler(_logger):
+            _logger.addHandler(_ColorStreamHandler())
+
+    getattr(_logger, type)(message.rstrip(), *args, **kwargs)
+
+
+@t.overload
+def _dt_as_utc(dt: None) -> None: ...
+
+
+@t.overload
+def _dt_as_utc(dt: datetime) -> datetime: ...
+
+
+def _dt_as_utc(dt: datetime | None) -> datetime | None:
+    if dt is None:
+        return dt
+
+    if dt.tzinfo is None:
+        return dt.replace(tzinfo=timezone.utc)
+    elif dt.tzinfo != timezone.utc:
+        return dt.astimezone(timezone.utc)
+
+    return dt
+
+
+_TAccessorValue = t.TypeVar("_TAccessorValue")
+
+
+class _DictAccessorProperty(t.Generic[_TAccessorValue]):
+    """Baseclass for `environ_property` and `header_property`."""
+
+    read_only = False
+
+    def __init__(
+        self,
+        name: str,
+        default: _TAccessorValue | None = None,
+        load_func: t.Callable[[str], _TAccessorValue] | None = None,
+        dump_func: t.Callable[[_TAccessorValue], str] | None = None,
+        read_only: bool | None = None,
+        doc: str | None = None,
+    ) -> None:
+        self.name = name
+        self.default = default
+        self.load_func = load_func
+        self.dump_func = dump_func
+        if read_only is not None:
+            self.read_only = read_only
+        self.__doc__ = doc
+
+    def lookup(self, instance: t.Any) -> t.MutableMapping[str, t.Any]:
+        raise NotImplementedError
+
+    @t.overload
+    def __get__(
+        self, instance: None, owner: type
+    ) -> _DictAccessorProperty[_TAccessorValue]: ...
+
+    @t.overload
+    def __get__(self, instance: t.Any, owner: type) -> _TAccessorValue: ...
+
+    def __get__(
+        self, instance: t.Any | None, owner: type
+    ) -> _TAccessorValue | _DictAccessorProperty[_TAccessorValue]:
+        if instance is None:
+            return self
+
+        storage = self.lookup(instance)
+
+        if self.name not in storage:
+            return self.default  # type: ignore
+
+        value = storage[self.name]
+
+        if self.load_func is not None:
+            try:
+                return self.load_func(value)
+            except (ValueError, TypeError):
+                return self.default  # type: ignore
+
+        return value  # type: ignore
+
+    def __set__(self, instance: t.Any, value: _TAccessorValue) -> None:
+        if self.read_only:
+            raise AttributeError("read only property")
+
+        if self.dump_func is not None:
+            self.lookup(instance)[self.name] = self.dump_func(value)
+        else:
+            self.lookup(instance)[self.name] = value
+
+    def __delete__(self, instance: t.Any) -> None:
+        if self.read_only:
+            raise AttributeError("read only property")
+
+        self.lookup(instance).pop(self.name, None)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.name}>"
+
+
+_plain_int_re = re.compile(r"-?\d+", re.ASCII)
+
+
+def _plain_int(value: str) -> int:
+    """Parse an int only if it is only ASCII digits and ``-``.
+
+    This disallows ``+``, ``_``, and non-ASCII digits, which are accepted by ``int`` but
+    are not allowed in HTTP header values.
+
+    Any leading or trailing whitespace is stripped
+    """
+    value = value.strip()
+    if _plain_int_re.fullmatch(value) is None:
+        raise ValueError
+
+    return int(value)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/_reloader.py b/backend/venv/lib/python3.12/site-packages/werkzeug/_reloader.py
new file mode 100644
index 00000000..8fd50b96
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/_reloader.py
@@ -0,0 +1,471 @@
+from __future__ import annotations
+
+import fnmatch
+import os
+import subprocess
+import sys
+import threading
+import time
+import typing as t
+from itertools import chain
+from pathlib import PurePath
+
+from ._internal import _log
+
+# The various system prefixes where imports are found. Base values are
+# different when running in a virtualenv. All reloaders will ignore the
+# base paths (usually the system installation). The stat reloader won't
+# scan the virtualenv paths, it will only include modules that are
+# already imported.
+_ignore_always = tuple({sys.base_prefix, sys.base_exec_prefix})
+prefix = {*_ignore_always, sys.prefix, sys.exec_prefix}
+
+if hasattr(sys, "real_prefix"):
+    # virtualenv < 20
+    prefix.add(sys.real_prefix)
+
+_stat_ignore_scan = tuple(prefix)
+del prefix
+_ignore_common_dirs = {
+    "__pycache__",
+    ".git",
+    ".hg",
+    ".tox",
+    ".nox",
+    ".pytest_cache",
+    ".mypy_cache",
+}
+
+
+def _iter_module_paths() -> t.Iterator[str]:
+    """Find the filesystem paths associated with imported modules."""
+    # List is in case the value is modified by the app while updating.
+    for module in list(sys.modules.values()):
+        name = getattr(module, "__file__", None)
+
+        if name is None or name.startswith(_ignore_always):
+            continue
+
+        while not os.path.isfile(name):
+            # Zip file, find the base file without the module path.
+            old = name
+            name = os.path.dirname(name)
+
+            if name == old:  # skip if it was all directories somehow
+                break
+        else:
+            yield name
+
+
+def _remove_by_pattern(paths: set[str], exclude_patterns: set[str]) -> None:
+    for pattern in exclude_patterns:
+        paths.difference_update(fnmatch.filter(paths, pattern))
+
+
+def _find_stat_paths(
+    extra_files: set[str], exclude_patterns: set[str]
+) -> t.Iterable[str]:
+    """Find paths for the stat reloader to watch. Returns imported
+    module files, Python files under non-system paths. Extra files and
+    Python files under extra directories can also be scanned.
+
+    System paths have to be excluded for efficiency. Non-system paths,
+    such as a project root or ``sys.path.insert``, should be the paths
+    of interest to the user anyway.
+    """
+    paths = set()
+
+    for path in chain(list(sys.path), extra_files):
+        path = os.path.abspath(path)
+
+        if os.path.isfile(path):
+            # zip file on sys.path, or extra file
+            paths.add(path)
+            continue
+
+        parent_has_py = {os.path.dirname(path): True}
+
+        for root, dirs, files in os.walk(path):
+            # Optimizations: ignore system prefixes, __pycache__ will
+            # have a py or pyc module at the import path, ignore some
+            # common known dirs such as version control and tool caches.
+            if (
+                root.startswith(_stat_ignore_scan)
+                or os.path.basename(root) in _ignore_common_dirs
+            ):
+                dirs.clear()
+                continue
+
+            has_py = False
+
+            for name in files:
+                if name.endswith((".py", ".pyc")):
+                    has_py = True
+                    paths.add(os.path.join(root, name))
+
+            # Optimization: stop scanning a directory if neither it nor
+            # its parent contained Python files.
+            if not (has_py or parent_has_py[os.path.dirname(root)]):
+                dirs.clear()
+                continue
+
+            parent_has_py[root] = has_py
+
+    paths.update(_iter_module_paths())
+    _remove_by_pattern(paths, exclude_patterns)
+    return paths
+
+
+def _find_watchdog_paths(
+    extra_files: set[str], exclude_patterns: set[str]
+) -> t.Iterable[str]:
+    """Find paths for the stat reloader to watch. Looks at the same
+    sources as the stat reloader, but watches everything under
+    directories instead of individual files.
+    """
+    dirs = set()
+
+    for name in chain(list(sys.path), extra_files):
+        name = os.path.abspath(name)
+
+        if os.path.isfile(name):
+            name = os.path.dirname(name)
+
+        dirs.add(name)
+
+    for name in _iter_module_paths():
+        dirs.add(os.path.dirname(name))
+
+    _remove_by_pattern(dirs, exclude_patterns)
+    return _find_common_roots(dirs)
+
+
+def _find_common_roots(paths: t.Iterable[str]) -> t.Iterable[str]:
+    root: dict[str, dict[str, t.Any]] = {}
+
+    for chunks in sorted((PurePath(x).parts for x in paths), key=len, reverse=True):
+        node = root
+
+        for chunk in chunks:
+            node = node.setdefault(chunk, {})
+
+        node.clear()
+
+    rv = set()
+
+    def _walk(node: t.Mapping[str, dict[str, t.Any]], path: tuple[str, ...]) -> None:
+        for prefix, child in node.items():
+            _walk(child, path + (prefix,))
+
+        # If there are no more nodes, and a path has been accumulated, add it.
+        # Path may be empty if the "" entry is in sys.path.
+        if not node and path:
+            rv.add(os.path.join(*path))
+
+    _walk(root, ())
+    return rv
+
+
+def _get_args_for_reloading() -> list[str]:
+    """Determine how the script was executed, and return the args needed
+    to execute it again in a new process.
+    """
+    if sys.version_info >= (3, 10):
+        # sys.orig_argv, added in Python 3.10, contains the exact args used to invoke
+        # Python. Still replace argv[0] with sys.executable for accuracy.
+        return [sys.executable, *sys.orig_argv[1:]]
+
+    rv = [sys.executable]
+    py_script = sys.argv[0]
+    args = sys.argv[1:]
+    # Need to look at main module to determine how it was executed.
+    __main__ = sys.modules["__main__"]
+
+    # The value of __package__ indicates how Python was called. It may
+    # not exist if a setuptools script is installed as an egg. It may be
+    # set incorrectly for entry points created with pip on Windows.
+    if getattr(__main__, "__package__", None) is None or (
+        os.name == "nt"
+        and __main__.__package__ == ""
+        and not os.path.exists(py_script)
+        and os.path.exists(f"{py_script}.exe")
+    ):
+        # Executed a file, like "python app.py".
+        py_script = os.path.abspath(py_script)
+
+        if os.name == "nt":
+            # Windows entry points have ".exe" extension and should be
+            # called directly.
+            if not os.path.exists(py_script) and os.path.exists(f"{py_script}.exe"):
+                py_script += ".exe"
+
+            if (
+                os.path.splitext(sys.executable)[1] == ".exe"
+                and os.path.splitext(py_script)[1] == ".exe"
+            ):
+                rv.pop(0)
+
+        rv.append(py_script)
+    else:
+        # Executed a module, like "python -m werkzeug.serving".
+        if os.path.isfile(py_script):
+            # Rewritten by Python from "-m script" to "/path/to/script.py".
+            py_module = t.cast(str, __main__.__package__)
+            name = os.path.splitext(os.path.basename(py_script))[0]
+
+            if name != "__main__":
+                py_module += f".{name}"
+        else:
+            # Incorrectly rewritten by pydevd debugger from "-m script" to "script".
+            py_module = py_script
+
+        rv.extend(("-m", py_module.lstrip(".")))
+
+    rv.extend(args)
+    return rv
+
+
+class ReloaderLoop:
+    name = ""
+
+    def __init__(
+        self,
+        extra_files: t.Iterable[str] | None = None,
+        exclude_patterns: t.Iterable[str] | None = None,
+        interval: int | float = 1,
+    ) -> None:
+        self.extra_files: set[str] = {os.path.abspath(x) for x in extra_files or ()}
+        self.exclude_patterns: set[str] = set(exclude_patterns or ())
+        self.interval = interval
+
+    def __enter__(self) -> ReloaderLoop:
+        """Do any setup, then run one step of the watch to populate the
+        initial filesystem state.
+        """
+        self.run_step()
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):  # type: ignore
+        """Clean up any resources associated with the reloader."""
+        pass
+
+    def run(self) -> None:
+        """Continually run the watch step, sleeping for the configured
+        interval after each step.
+        """
+        while True:
+            self.run_step()
+            time.sleep(self.interval)
+
+    def run_step(self) -> None:
+        """Run one step for watching the filesystem. Called once to set
+        up initial state, then repeatedly to update it.
+        """
+        pass
+
+    def restart_with_reloader(self) -> int:
+        """Spawn a new Python interpreter with the same arguments as the
+        current one, but running the reloader thread.
+        """
+        while True:
+            _log("info", f" * Restarting with {self.name}")
+            args = _get_args_for_reloading()
+            new_environ = os.environ.copy()
+            new_environ["WERKZEUG_RUN_MAIN"] = "true"
+            exit_code = subprocess.call(args, env=new_environ, close_fds=False)
+
+            if exit_code != 3:
+                return exit_code
+
+    def trigger_reload(self, filename: str) -> None:
+        self.log_reload(filename)
+        sys.exit(3)
+
+    def log_reload(self, filename: str | bytes) -> None:
+        filename = os.path.abspath(filename)
+        _log("info", f" * Detected change in {filename!r}, reloading")
+
+
+class StatReloaderLoop(ReloaderLoop):
+    name = "stat"
+
+    def __enter__(self) -> ReloaderLoop:
+        self.mtimes: dict[str, float] = {}
+        return super().__enter__()
+
+    def run_step(self) -> None:
+        for name in _find_stat_paths(self.extra_files, self.exclude_patterns):
+            try:
+                mtime = os.stat(name).st_mtime
+            except OSError:
+                continue
+
+            old_time = self.mtimes.get(name)
+
+            if old_time is None:
+                self.mtimes[name] = mtime
+                continue
+
+            if mtime > old_time:
+                self.trigger_reload(name)
+
+
+class WatchdogReloaderLoop(ReloaderLoop):
+    def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:
+        from watchdog.events import EVENT_TYPE_CLOSED
+        from watchdog.events import EVENT_TYPE_CREATED
+        from watchdog.events import EVENT_TYPE_DELETED
+        from watchdog.events import EVENT_TYPE_MODIFIED
+        from watchdog.events import EVENT_TYPE_MOVED
+        from watchdog.events import FileModifiedEvent
+        from watchdog.events import PatternMatchingEventHandler
+        from watchdog.observers import Observer
+
+        super().__init__(*args, **kwargs)
+        trigger_reload = self.trigger_reload
+
+        class EventHandler(PatternMatchingEventHandler):
+            def on_any_event(self, event: FileModifiedEvent):  # type: ignore
+                if event.event_type not in {
+                    EVENT_TYPE_CLOSED,
+                    EVENT_TYPE_CREATED,
+                    EVENT_TYPE_DELETED,
+                    EVENT_TYPE_MODIFIED,
+                    EVENT_TYPE_MOVED,
+                }:
+                    # skip events that don't involve changes to the file
+                    return
+
+                trigger_reload(event.src_path)
+
+        reloader_name = Observer.__name__.lower()  # type: ignore[attr-defined]
+
+        if reloader_name.endswith("observer"):
+            reloader_name = reloader_name[:-8]
+
+        self.name = f"watchdog ({reloader_name})"
+        self.observer = Observer()
+        # Extra patterns can be non-Python files, match them in addition
+        # to all Python files in default and extra directories. Ignore
+        # __pycache__ since a change there will always have a change to
+        # the source file (or initial pyc file) as well. Ignore Git and
+        # Mercurial internal changes.
+        extra_patterns = [p for p in self.extra_files if not os.path.isdir(p)]
+        self.event_handler = EventHandler(
+            patterns=["*.py", "*.pyc", "*.zip", *extra_patterns],
+            ignore_patterns=[
+                *[f"*/{d}/*" for d in _ignore_common_dirs],
+                *self.exclude_patterns,
+            ],
+        )
+        self.should_reload = False
+
+    def trigger_reload(self, filename: str | bytes) -> None:
+        # This is called inside an event handler, which means throwing
+        # SystemExit has no effect.
+        # https://github.com/gorakhargosh/watchdog/issues/294
+        self.should_reload = True
+        self.log_reload(filename)
+
+    def __enter__(self) -> ReloaderLoop:
+        self.watches: dict[str, t.Any] = {}
+        self.observer.start()
+        return super().__enter__()
+
+    def __exit__(self, exc_type, exc_val, exc_tb):  # type: ignore
+        self.observer.stop()
+        self.observer.join()
+
+    def run(self) -> None:
+        while not self.should_reload:
+            self.run_step()
+            time.sleep(self.interval)
+
+        sys.exit(3)
+
+    def run_step(self) -> None:
+        to_delete = set(self.watches)
+
+        for path in _find_watchdog_paths(self.extra_files, self.exclude_patterns):
+            if path not in self.watches:
+                try:
+                    self.watches[path] = self.observer.schedule(
+                        self.event_handler, path, recursive=True
+                    )
+                except OSError:
+                    # Clear this path from list of watches We don't want
+                    # the same error message showing again in the next
+                    # iteration.
+                    self.watches[path] = None
+
+            to_delete.discard(path)
+
+        for path in to_delete:
+            watch = self.watches.pop(path, None)
+
+            if watch is not None:
+                self.observer.unschedule(watch)
+
+
+reloader_loops: dict[str, type[ReloaderLoop]] = {
+    "stat": StatReloaderLoop,
+    "watchdog": WatchdogReloaderLoop,
+}
+
+try:
+    __import__("watchdog.observers")
+except ImportError:
+    reloader_loops["auto"] = reloader_loops["stat"]
+else:
+    reloader_loops["auto"] = reloader_loops["watchdog"]
+
+
+def ensure_echo_on() -> None:
+    """Ensure that echo mode is enabled. Some tools such as PDB disable
+    it which causes usability issues after a reload."""
+    # tcgetattr will fail if stdin isn't a tty
+    if sys.stdin is None or not sys.stdin.isatty():
+        return
+
+    try:
+        import termios
+    except ImportError:
+        return
+
+    attributes = termios.tcgetattr(sys.stdin)
+
+    if not attributes[3] & termios.ECHO:
+        attributes[3] |= termios.ECHO
+        termios.tcsetattr(sys.stdin, termios.TCSANOW, attributes)
+
+
+def run_with_reloader(
+    main_func: t.Callable[[], None],
+    extra_files: t.Iterable[str] | None = None,
+    exclude_patterns: t.Iterable[str] | None = None,
+    interval: int | float = 1,
+    reloader_type: str = "auto",
+) -> None:
+    """Run the given function in an independent Python interpreter."""
+    import signal
+
+    signal.signal(signal.SIGTERM, lambda *args: sys.exit(0))
+    reloader = reloader_loops[reloader_type](
+        extra_files=extra_files, exclude_patterns=exclude_patterns, interval=interval
+    )
+
+    try:
+        if os.environ.get("WERKZEUG_RUN_MAIN") == "true":
+            ensure_echo_on()
+            t = threading.Thread(target=main_func, args=())
+            t.daemon = True
+
+            # Enter the reloader to set up initial state, then start
+            # the app thread and reloader update loop.
+            with reloader:
+                t.start()
+                reloader.run()
+        else:
+            sys.exit(reloader.restart_with_reloader())
+    except KeyboardInterrupt:
+        pass
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__init__.py
new file mode 100644
index 00000000..6582de02
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__init__.py
@@ -0,0 +1,64 @@
+from __future__ import annotations
+
+import typing as t
+
+from .accept import Accept as Accept
+from .accept import CharsetAccept as CharsetAccept
+from .accept import LanguageAccept as LanguageAccept
+from .accept import MIMEAccept as MIMEAccept
+from .auth import Authorization as Authorization
+from .auth import WWWAuthenticate as WWWAuthenticate
+from .cache_control import RequestCacheControl as RequestCacheControl
+from .cache_control import ResponseCacheControl as ResponseCacheControl
+from .csp import ContentSecurityPolicy as ContentSecurityPolicy
+from .etag import ETags as ETags
+from .file_storage import FileMultiDict as FileMultiDict
+from .file_storage import FileStorage as FileStorage
+from .headers import EnvironHeaders as EnvironHeaders
+from .headers import Headers as Headers
+from .mixins import ImmutableDictMixin as ImmutableDictMixin
+from .mixins import ImmutableHeadersMixin as ImmutableHeadersMixin
+from .mixins import ImmutableListMixin as ImmutableListMixin
+from .mixins import ImmutableMultiDictMixin as ImmutableMultiDictMixin
+from .mixins import UpdateDictMixin as UpdateDictMixin
+from .range import ContentRange as ContentRange
+from .range import IfRange as IfRange
+from .range import Range as Range
+from .structures import CallbackDict as CallbackDict
+from .structures import CombinedMultiDict as CombinedMultiDict
+from .structures import HeaderSet as HeaderSet
+from .structures import ImmutableDict as ImmutableDict
+from .structures import ImmutableList as ImmutableList
+from .structures import ImmutableMultiDict as ImmutableMultiDict
+from .structures import ImmutableTypeConversionDict as ImmutableTypeConversionDict
+from .structures import iter_multi_items as iter_multi_items
+from .structures import MultiDict as MultiDict
+from .structures import TypeConversionDict as TypeConversionDict
+
+
+def __getattr__(name: str) -> t.Any:
+    import warnings
+
+    if name == "OrderedMultiDict":
+        from .structures import _OrderedMultiDict
+
+        warnings.warn(
+            "'OrderedMultiDict' is deprecated and will be removed in Werkzeug"
+            " 3.2. Use 'MultiDict' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return _OrderedMultiDict
+
+    if name == "ImmutableOrderedMultiDict":
+        from .structures import _ImmutableOrderedMultiDict
+
+        warnings.warn(
+            "'OrderedMultiDict' is deprecated and will be removed in Werkzeug"
+            " 3.2. Use 'ImmutableMultiDict' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return _ImmutableOrderedMultiDict
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..04fe5449
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/accept.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/accept.cpython-312.pyc
new file mode 100644
index 00000000..95c46b28
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/accept.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/auth.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/auth.cpython-312.pyc
new file mode 100644
index 00000000..ee3c2b2b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/auth.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/cache_control.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/cache_control.cpython-312.pyc
new file mode 100644
index 00000000..64bfb2ab
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/cache_control.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/csp.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/csp.cpython-312.pyc
new file mode 100644
index 00000000..2376051f
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/csp.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/etag.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/etag.cpython-312.pyc
new file mode 100644
index 00000000..4bc50800
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/etag.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/file_storage.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/file_storage.cpython-312.pyc
new file mode 100644
index 00000000..07565f1b
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/file_storage.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/headers.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/headers.cpython-312.pyc
new file mode 100644
index 00000000..abef781a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/headers.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/mixins.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/mixins.cpython-312.pyc
new file mode 100644
index 00000000..a440bbad
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/mixins.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/range.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/range.cpython-312.pyc
new file mode 100644
index 00000000..cbf24526
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/range.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/structures.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/structures.cpython-312.pyc
new file mode 100644
index 00000000..86725373
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/__pycache__/structures.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/accept.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/accept.py
new file mode 100644
index 00000000..44179a93
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/accept.py
@@ -0,0 +1,350 @@
+from __future__ import annotations
+
+import codecs
+import collections.abc as cabc
+import re
+import typing as t
+
+from .structures import ImmutableList
+
+
+class Accept(ImmutableList[tuple[str, float]]):
+    """An :class:`Accept` object is just a list subclass for lists of
+    ``(value, quality)`` tuples.  It is automatically sorted by specificity
+    and quality.
+
+    All :class:`Accept` objects work similar to a list but provide extra
+    functionality for working with the data.  Containment checks are
+    normalized to the rules of that header:
+
+    >>> a = CharsetAccept([('ISO-8859-1', 1), ('utf-8', 0.7)])
+    >>> a.best
+    'ISO-8859-1'
+    >>> 'iso-8859-1' in a
+    True
+    >>> 'UTF8' in a
+    True
+    >>> 'utf7' in a
+    False
+
+    To get the quality for an item you can use normal item lookup:
+
+    >>> print a['utf-8']
+    0.7
+    >>> a['utf7']
+    0
+
+    .. versionchanged:: 0.5
+       :class:`Accept` objects are forced immutable now.
+
+    .. versionchanged:: 1.0.0
+       :class:`Accept` internal values are no longer ordered
+       alphabetically for equal quality tags. Instead the initial
+       order is preserved.
+
+    """
+
+    def __init__(
+        self, values: Accept | cabc.Iterable[tuple[str, float]] | None = ()
+    ) -> None:
+        if values is None:
+            super().__init__()
+            self.provided = False
+        elif isinstance(values, Accept):
+            self.provided = values.provided
+            super().__init__(values)
+        else:
+            self.provided = True
+            values = sorted(
+                values, key=lambda x: (self._specificity(x[0]), x[1]), reverse=True
+            )
+            super().__init__(values)
+
+    def _specificity(self, value: str) -> tuple[bool, ...]:
+        """Returns a tuple describing the value's specificity."""
+        return (value != "*",)
+
+    def _value_matches(self, value: str, item: str) -> bool:
+        """Check if a value matches a given accept item."""
+        return item == "*" or item.lower() == value.lower()
+
+    @t.overload
+    def __getitem__(self, key: str) -> float: ...
+    @t.overload
+    def __getitem__(self, key: t.SupportsIndex) -> tuple[str, float]: ...
+    @t.overload
+    def __getitem__(self, key: slice) -> list[tuple[str, float]]: ...
+    def __getitem__(
+        self, key: str | t.SupportsIndex | slice
+    ) -> float | tuple[str, float] | list[tuple[str, float]]:
+        """Besides index lookup (getting item n) you can also pass it a string
+        to get the quality for the item.  If the item is not in the list, the
+        returned quality is ``0``.
+        """
+        if isinstance(key, str):
+            return self.quality(key)
+        return list.__getitem__(self, key)
+
+    def quality(self, key: str) -> float:
+        """Returns the quality of the key.
+
+        .. versionadded:: 0.6
+           In previous versions you had to use the item-lookup syntax
+           (eg: ``obj[key]`` instead of ``obj.quality(key)``)
+        """
+        for item, quality in self:
+            if self._value_matches(key, item):
+                return quality
+        return 0
+
+    def __contains__(self, value: str) -> bool:  # type: ignore[override]
+        for item, _quality in self:
+            if self._value_matches(value, item):
+                return True
+        return False
+
+    def __repr__(self) -> str:
+        pairs_str = ", ".join(f"({x!r}, {y})" for x, y in self)
+        return f"{type(self).__name__}([{pairs_str}])"
+
+    def index(self, key: str | tuple[str, float]) -> int:  # type: ignore[override]
+        """Get the position of an entry or raise :exc:`ValueError`.
+
+        :param key: The key to be looked up.
+
+        .. versionchanged:: 0.5
+           This used to raise :exc:`IndexError`, which was inconsistent
+           with the list API.
+        """
+        if isinstance(key, str):
+            for idx, (item, _quality) in enumerate(self):
+                if self._value_matches(key, item):
+                    return idx
+            raise ValueError(key)
+        return list.index(self, key)
+
+    def find(self, key: str | tuple[str, float]) -> int:
+        """Get the position of an entry or return -1.
+
+        :param key: The key to be looked up.
+        """
+        try:
+            return self.index(key)
+        except ValueError:
+            return -1
+
+    def values(self) -> cabc.Iterator[str]:
+        """Iterate over all values."""
+        for item in self:
+            yield item[0]
+
+    def to_header(self) -> str:
+        """Convert the header set into an HTTP header string."""
+        result = []
+        for value, quality in self:
+            if quality != 1:
+                value = f"{value};q={quality}"
+            result.append(value)
+        return ",".join(result)
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def _best_single_match(self, match: str) -> tuple[str, float] | None:
+        for client_item, quality in self:
+            if self._value_matches(match, client_item):
+                # self is sorted by specificity descending, we can exit
+                return client_item, quality
+        return None
+
+    @t.overload
+    def best_match(self, matches: cabc.Iterable[str]) -> str | None: ...
+    @t.overload
+    def best_match(self, matches: cabc.Iterable[str], default: str = ...) -> str: ...
+    def best_match(
+        self, matches: cabc.Iterable[str], default: str | None = None
+    ) -> str | None:
+        """Returns the best match from a list of possible matches based
+        on the specificity and quality of the client. If two items have the
+        same quality and specificity, the one is returned that comes first.
+
+        :param matches: a list of matches to check for
+        :param default: the value that is returned if none match
+        """
+        result = default
+        best_quality: float = -1
+        best_specificity: tuple[float, ...] = (-1,)
+        for server_item in matches:
+            match = self._best_single_match(server_item)
+            if not match:
+                continue
+            client_item, quality = match
+            specificity = self._specificity(client_item)
+            if quality <= 0 or quality < best_quality:
+                continue
+            # better quality or same quality but more specific => better match
+            if quality > best_quality or specificity > best_specificity:
+                result = server_item
+                best_quality = quality
+                best_specificity = specificity
+        return result
+
+    @property
+    def best(self) -> str | None:
+        """The best match as value."""
+        if self:
+            return self[0][0]
+
+        return None
+
+
+_mime_split_re = re.compile(r"/|(?:\s*;\s*)")
+
+
+def _normalize_mime(value: str) -> list[str]:
+    return _mime_split_re.split(value.lower())
+
+
+class MIMEAccept(Accept):
+    """Like :class:`Accept` but with special methods and behavior for
+    mimetypes.
+    """
+
+    def _specificity(self, value: str) -> tuple[bool, ...]:
+        return tuple(x != "*" for x in _mime_split_re.split(value))
+
+    def _value_matches(self, value: str, item: str) -> bool:
+        # item comes from the client, can't match if it's invalid.
+        if "/" not in item:
+            return False
+
+        # value comes from the application, tell the developer when it
+        # doesn't look valid.
+        if "/" not in value:
+            raise ValueError(f"invalid mimetype {value!r}")
+
+        # Split the match value into type, subtype, and a sorted list of parameters.
+        normalized_value = _normalize_mime(value)
+        value_type, value_subtype = normalized_value[:2]
+        value_params = sorted(normalized_value[2:])
+
+        # "*/*" is the only valid value that can start with "*".
+        if value_type == "*" and value_subtype != "*":
+            raise ValueError(f"invalid mimetype {value!r}")
+
+        # Split the accept item into type, subtype, and parameters.
+        normalized_item = _normalize_mime(item)
+        item_type, item_subtype = normalized_item[:2]
+        item_params = sorted(normalized_item[2:])
+
+        # "*/not-*" from the client is invalid, can't match.
+        if item_type == "*" and item_subtype != "*":
+            return False
+
+        return (
+            (item_type == "*" and item_subtype == "*")
+            or (value_type == "*" and value_subtype == "*")
+        ) or (
+            item_type == value_type
+            and (
+                item_subtype == "*"
+                or value_subtype == "*"
+                or (item_subtype == value_subtype and item_params == value_params)
+            )
+        )
+
+    @property
+    def accept_html(self) -> bool:
+        """True if this object accepts HTML."""
+        return "text/html" in self or self.accept_xhtml  # type: ignore[comparison-overlap]
+
+    @property
+    def accept_xhtml(self) -> bool:
+        """True if this object accepts XHTML."""
+        return "application/xhtml+xml" in self or "application/xml" in self  # type: ignore[comparison-overlap]
+
+    @property
+    def accept_json(self) -> bool:
+        """True if this object accepts JSON."""
+        return "application/json" in self  # type: ignore[comparison-overlap]
+
+
+_locale_delim_re = re.compile(r"[_-]")
+
+
+def _normalize_lang(value: str) -> list[str]:
+    """Process a language tag for matching."""
+    return _locale_delim_re.split(value.lower())
+
+
+class LanguageAccept(Accept):
+    """Like :class:`Accept` but with normalization for language tags."""
+
+    def _value_matches(self, value: str, item: str) -> bool:
+        return item == "*" or _normalize_lang(value) == _normalize_lang(item)
+
+    @t.overload
+    def best_match(self, matches: cabc.Iterable[str]) -> str | None: ...
+    @t.overload
+    def best_match(self, matches: cabc.Iterable[str], default: str = ...) -> str: ...
+    def best_match(
+        self, matches: cabc.Iterable[str], default: str | None = None
+    ) -> str | None:
+        """Given a list of supported values, finds the best match from
+        the list of accepted values.
+
+        Language tags are normalized for the purpose of matching, but
+        are returned unchanged.
+
+        If no exact match is found, this will fall back to matching
+        the first subtag (primary language only), first with the
+        accepted values then with the match values. This partial is not
+        applied to any other language subtags.
+
+        The default is returned if no exact or fallback match is found.
+
+        :param matches: A list of supported languages to find a match.
+        :param default: The value that is returned if none match.
+        """
+        # Look for an exact match first. If a client accepts "en-US",
+        # "en-US" is a valid match at this point.
+        result = super().best_match(matches)
+
+        if result is not None:
+            return result
+
+        # Fall back to accepting primary tags. If a client accepts
+        # "en-US", "en" is a valid match at this point. Need to use
+        # re.split to account for 2 or 3 letter codes.
+        fallback = Accept(
+            [(_locale_delim_re.split(item[0], 1)[0], item[1]) for item in self]
+        )
+        result = fallback.best_match(matches)
+
+        if result is not None:
+            return result
+
+        # Fall back to matching primary tags. If the client accepts
+        # "en", "en-US" is a valid match at this point.
+        fallback_matches = [_locale_delim_re.split(item, 1)[0] for item in matches]
+        result = super().best_match(fallback_matches)
+
+        # Return a value from the original match list. Find the first
+        # original value that starts with the matched primary tag.
+        if result is not None:
+            return next(item for item in matches if item.startswith(result))
+
+        return default
+
+
+class CharsetAccept(Accept):
+    """Like :class:`Accept` but with normalization for charsets."""
+
+    def _value_matches(self, value: str, item: str) -> bool:
+        def _normalize(name: str) -> str:
+            try:
+                return codecs.lookup(name).name
+            except LookupError:
+                return name.lower()
+
+        return item == "*" or _normalize(value) == _normalize(item)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/auth.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/auth.py
new file mode 100644
index 00000000..42f7aa46
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/auth.py
@@ -0,0 +1,317 @@
+from __future__ import annotations
+
+import base64
+import binascii
+import collections.abc as cabc
+import typing as t
+
+from ..http import dump_header
+from ..http import parse_dict_header
+from ..http import quote_header_value
+from .structures import CallbackDict
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+
+class Authorization:
+    """Represents the parts of an ``Authorization`` request header.
+
+    :attr:`.Request.authorization` returns an instance if the header is set.
+
+    An instance can be used with the test :class:`.Client` request methods' ``auth``
+    parameter to send the header in test requests.
+
+    Depending on the auth scheme, either :attr:`parameters` or :attr:`token` will be
+    set. The ``Basic`` scheme's token is decoded into the ``username`` and ``password``
+    parameters.
+
+    For convenience, ``auth["key"]`` and ``auth.key`` both access the key in the
+    :attr:`parameters` dict, along with ``auth.get("key")`` and ``"key" in auth``.
+
+    .. versionchanged:: 2.3
+        The ``token`` parameter and attribute was added to support auth schemes that use
+        a token instead of parameters, such as ``Bearer``.
+
+    .. versionchanged:: 2.3
+        The object is no longer a ``dict``.
+
+    .. versionchanged:: 0.5
+        The object is an immutable dict.
+    """
+
+    def __init__(
+        self,
+        auth_type: str,
+        data: dict[str, str | None] | None = None,
+        token: str | None = None,
+    ) -> None:
+        self.type = auth_type
+        """The authorization scheme, like ``basic``, ``digest``, or ``bearer``."""
+
+        if data is None:
+            data = {}
+
+        self.parameters = data
+        """A dict of parameters parsed from the header. Either this or :attr:`token`
+        will have a value for a given scheme.
+        """
+
+        self.token = token
+        """A token parsed from the header. Either this or :attr:`parameters` will have a
+        value for a given scheme.
+
+        .. versionadded:: 2.3
+        """
+
+    def __getattr__(self, name: str) -> str | None:
+        return self.parameters.get(name)
+
+    def __getitem__(self, name: str) -> str | None:
+        return self.parameters.get(name)
+
+    def get(self, key: str, default: str | None = None) -> str | None:
+        return self.parameters.get(key, default)
+
+    def __contains__(self, key: str) -> bool:
+        return key in self.parameters
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Authorization):
+            return NotImplemented
+
+        return (
+            other.type == self.type
+            and other.token == self.token
+            and other.parameters == self.parameters
+        )
+
+    @classmethod
+    def from_header(cls, value: str | None) -> te.Self | None:
+        """Parse an ``Authorization`` header value and return an instance, or ``None``
+        if the value is empty.
+
+        :param value: The header value to parse.
+
+        .. versionadded:: 2.3
+        """
+        if not value:
+            return None
+
+        scheme, _, rest = value.partition(" ")
+        scheme = scheme.lower()
+        rest = rest.strip()
+
+        if scheme == "basic":
+            try:
+                username, _, password = base64.b64decode(rest).decode().partition(":")
+            except (binascii.Error, UnicodeError):
+                return None
+
+            return cls(scheme, {"username": username, "password": password})
+
+        if "=" in rest.rstrip("="):
+            # = that is not trailing, this is parameters.
+            return cls(scheme, parse_dict_header(rest), None)
+
+        # No = or only trailing =, this is a token.
+        return cls(scheme, None, rest)
+
+    def to_header(self) -> str:
+        """Produce an ``Authorization`` header value representing this data.
+
+        .. versionadded:: 2.0
+        """
+        if self.type == "basic":
+            value = base64.b64encode(
+                f"{self.username}:{self.password}".encode()
+            ).decode("ascii")
+            return f"Basic {value}"
+
+        if self.token is not None:
+            return f"{self.type.title()} {self.token}"
+
+        return f"{self.type.title()} {dump_header(self.parameters)}"
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.to_header()}>"
+
+
+class WWWAuthenticate:
+    """Represents the parts of a ``WWW-Authenticate`` response header.
+
+    Set :attr:`.Response.www_authenticate` to an instance of list of instances to set
+    values for this header in the response. Modifying this instance will modify the
+    header value.
+
+    Depending on the auth scheme, either :attr:`parameters` or :attr:`token` should be
+    set. The ``Basic`` scheme will encode ``username`` and ``password`` parameters to a
+    token.
+
+    For convenience, ``auth["key"]`` and ``auth.key`` both act on the :attr:`parameters`
+    dict, and can be used to get, set, or delete parameters. ``auth.get("key")`` and
+    ``"key" in auth`` are also provided.
+
+    .. versionchanged:: 2.3
+        The ``token`` parameter and attribute was added to support auth schemes that use
+        a token instead of parameters, such as ``Bearer``.
+
+    .. versionchanged:: 2.3
+        The object is no longer a ``dict``.
+
+    .. versionchanged:: 2.3
+        The ``on_update`` parameter was removed.
+    """
+
+    def __init__(
+        self,
+        auth_type: str,
+        values: dict[str, str | None] | None = None,
+        token: str | None = None,
+    ):
+        self._type = auth_type.lower()
+        self._parameters: dict[str, str | None] = CallbackDict(
+            values, lambda _: self._trigger_on_update()
+        )
+        self._token = token
+        self._on_update: cabc.Callable[[WWWAuthenticate], None] | None = None
+
+    def _trigger_on_update(self) -> None:
+        if self._on_update is not None:
+            self._on_update(self)
+
+    @property
+    def type(self) -> str:
+        """The authorization scheme, like ``basic``, ``digest``, or ``bearer``."""
+        return self._type
+
+    @type.setter
+    def type(self, value: str) -> None:
+        self._type = value
+        self._trigger_on_update()
+
+    @property
+    def parameters(self) -> dict[str, str | None]:
+        """A dict of parameters for the header. Only one of this or :attr:`token` should
+        have a value for a given scheme.
+        """
+        return self._parameters
+
+    @parameters.setter
+    def parameters(self, value: dict[str, str]) -> None:
+        self._parameters = CallbackDict(value, lambda _: self._trigger_on_update())
+        self._trigger_on_update()
+
+    @property
+    def token(self) -> str | None:
+        """A dict of parameters for the header. Only one of this or :attr:`token` should
+        have a value for a given scheme.
+        """
+        return self._token
+
+    @token.setter
+    def token(self, value: str | None) -> None:
+        """A token for the header. Only one of this or :attr:`parameters` should have a
+        value for a given scheme.
+
+        .. versionadded:: 2.3
+        """
+        self._token = value
+        self._trigger_on_update()
+
+    def __getitem__(self, key: str) -> str | None:
+        return self.parameters.get(key)
+
+    def __setitem__(self, key: str, value: str | None) -> None:
+        if value is None:
+            if key in self.parameters:
+                del self.parameters[key]
+        else:
+            self.parameters[key] = value
+
+        self._trigger_on_update()
+
+    def __delitem__(self, key: str) -> None:
+        if key in self.parameters:
+            del self.parameters[key]
+            self._trigger_on_update()
+
+    def __getattr__(self, name: str) -> str | None:
+        return self[name]
+
+    def __setattr__(self, name: str, value: str | None) -> None:
+        if name in {"_type", "_parameters", "_token", "_on_update"}:
+            super().__setattr__(name, value)
+        else:
+            self[name] = value
+
+    def __delattr__(self, name: str) -> None:
+        del self[name]
+
+    def __contains__(self, key: str) -> bool:
+        return key in self.parameters
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, WWWAuthenticate):
+            return NotImplemented
+
+        return (
+            other.type == self.type
+            and other.token == self.token
+            and other.parameters == self.parameters
+        )
+
+    def get(self, key: str, default: str | None = None) -> str | None:
+        return self.parameters.get(key, default)
+
+    @classmethod
+    def from_header(cls, value: str | None) -> te.Self | None:
+        """Parse a ``WWW-Authenticate`` header value and return an instance, or ``None``
+        if the value is empty.
+
+        :param value: The header value to parse.
+
+        .. versionadded:: 2.3
+        """
+        if not value:
+            return None
+
+        scheme, _, rest = value.partition(" ")
+        scheme = scheme.lower()
+        rest = rest.strip()
+
+        if "=" in rest.rstrip("="):
+            # = that is not trailing, this is parameters.
+            return cls(scheme, parse_dict_header(rest), None)
+
+        # No = or only trailing =, this is a token.
+        return cls(scheme, None, rest)
+
+    def to_header(self) -> str:
+        """Produce a ``WWW-Authenticate`` header value representing this data."""
+        if self.token is not None:
+            return f"{self.type.title()} {self.token}"
+
+        if self.type == "digest":
+            items = []
+
+            for key, value in self.parameters.items():
+                if key in {"realm", "domain", "nonce", "opaque", "qop"}:
+                    value = quote_header_value(value, allow_token=False)
+                else:
+                    value = quote_header_value(value)
+
+                items.append(f"{key}={value}")
+
+            return f"Digest {', '.join(items)}"
+
+        return f"{self.type.title()} {dump_header(self.parameters)}"
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.to_header()}>"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/cache_control.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/cache_control.py
new file mode 100644
index 00000000..8d700ab6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/cache_control.py
@@ -0,0 +1,273 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import typing as t
+from inspect import cleandoc
+
+from .mixins import ImmutableDictMixin
+from .structures import CallbackDict
+
+
+def cache_control_property(
+    key: str, empty: t.Any, type: type[t.Any] | None, *, doc: str | None = None
+) -> t.Any:
+    """Return a new property object for a cache header. Useful if you
+    want to add support for a cache extension in a subclass.
+
+    :param key: The attribute name present in the parsed cache-control header dict.
+    :param empty: The value to use if the key is present without a value.
+    :param type: The type to convert the string value to instead of a string. If
+        conversion raises a ``ValueError``, the returned value is ``None``.
+    :param doc: The docstring for the property. If not given, it is generated
+        based on the other params.
+
+    .. versionchanged:: 3.1
+        Added the ``doc`` param.
+
+    .. versionchanged:: 2.0
+        Renamed from ``cache_property``.
+    """
+    if doc is None:
+        parts = [f"The ``{key}`` attribute."]
+
+        if type is bool:
+            parts.append("A ``bool``, either present or not.")
+        else:
+            if type is None:
+                parts.append("A ``str``,")
+            else:
+                parts.append(f"A ``{type.__name__}``,")
+
+            if empty is not None:
+                parts.append(f"``{empty!r}`` if present with no value,")
+
+            parts.append("or ``None`` if not present.")
+
+        doc = " ".join(parts)
+
+    return property(
+        lambda x: x._get_cache_value(key, empty, type),
+        lambda x, v: x._set_cache_value(key, v, type),
+        lambda x: x._del_cache_value(key),
+        doc=cleandoc(doc),
+    )
+
+
+class _CacheControl(CallbackDict[str, t.Optional[str]]):
+    """Subclass of a dict that stores values for a Cache-Control header.  It
+    has accessors for all the cache-control directives specified in RFC 2616.
+    The class does not differentiate between request and response directives.
+
+    Because the cache-control directives in the HTTP header use dashes the
+    python descriptors use underscores for that.
+
+    To get a header of the :class:`CacheControl` object again you can convert
+    the object into a string or call the :meth:`to_header` method.  If you plan
+    to subclass it and add your own items have a look at the sourcecode for
+    that class.
+
+    .. versionchanged:: 3.1
+        Dict values are always ``str | None``. Setting properties will
+        convert the value to a string. Setting a non-bool property to
+        ``False`` is equivalent to setting it to ``None``. Getting typed
+        properties will return ``None`` if conversion raises
+        ``ValueError``, rather than the string.
+
+    .. versionchanged:: 2.1
+        Setting int properties such as ``max_age`` will convert the
+        value to an int.
+
+    .. versionchanged:: 0.4
+       Setting ``no_cache`` or ``private`` to ``True`` will set the
+       implicit value ``"*"``.
+    """
+
+    no_store: bool = cache_control_property("no-store", None, bool)
+    max_age: int | None = cache_control_property("max-age", None, int)
+    no_transform: bool = cache_control_property("no-transform", None, bool)
+    stale_if_error: int | None = cache_control_property("stale-if-error", None, int)
+
+    def __init__(
+        self,
+        values: cabc.Mapping[str, t.Any] | cabc.Iterable[tuple[str, t.Any]] | None = (),
+        on_update: cabc.Callable[[_CacheControl], None] | None = None,
+    ):
+        super().__init__(values, on_update)
+        self.provided = values is not None
+
+    def _get_cache_value(
+        self, key: str, empty: t.Any, type: type[t.Any] | None
+    ) -> t.Any:
+        """Used internally by the accessor properties."""
+        if type is bool:
+            return key in self
+
+        if key not in self:
+            return None
+
+        if (value := self[key]) is None:
+            return empty
+
+        if type is not None:
+            try:
+                value = type(value)
+            except ValueError:
+                return None
+
+        return value
+
+    def _set_cache_value(
+        self, key: str, value: t.Any, type: type[t.Any] | None
+    ) -> None:
+        """Used internally by the accessor properties."""
+        if type is bool:
+            if value:
+                self[key] = None
+            else:
+                self.pop(key, None)
+        elif value is None or value is False:
+            self.pop(key, None)
+        elif value is True:
+            self[key] = None
+        else:
+            if type is not None:
+                value = type(value)
+
+            self[key] = str(value)
+
+    def _del_cache_value(self, key: str) -> None:
+        """Used internally by the accessor properties."""
+        if key in self:
+            del self[key]
+
+    def to_header(self) -> str:
+        """Convert the stored values into a cache control header."""
+        return http.dump_header(self)
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        kv_str = " ".join(f"{k}={v!r}" for k, v in sorted(self.items()))
+        return f"<{type(self).__name__} {kv_str}>"
+
+    cache_property = staticmethod(cache_control_property)
+
+
+class RequestCacheControl(ImmutableDictMixin[str, t.Optional[str]], _CacheControl):  # type: ignore[misc]
+    """A cache control for requests.  This is immutable and gives access
+    to all the request-relevant cache control headers.
+
+    To get a header of the :class:`RequestCacheControl` object again you can
+    convert the object into a string or call the :meth:`to_header` method.  If
+    you plan to subclass it and add your own items have a look at the sourcecode
+    for that class.
+
+    .. versionchanged:: 3.1
+        Dict values are always ``str | None``. Setting properties will
+        convert the value to a string. Setting a non-bool property to
+        ``False`` is equivalent to setting it to ``None``. Getting typed
+        properties will return ``None`` if conversion raises
+        ``ValueError``, rather than the string.
+
+    .. versionchanged:: 3.1
+       ``max_age`` is ``None`` if present without a value, rather
+       than ``-1``.
+
+    .. versionchanged:: 3.1
+        ``no_cache`` is a boolean, it is ``True`` instead of ``"*"``
+        when present.
+
+    .. versionchanged:: 3.1
+        ``max_stale`` is ``True`` if present without a value, rather
+        than ``"*"``.
+
+    .. versionchanged:: 3.1
+       ``no_transform`` is a boolean. Previously it was mistakenly
+       always ``None``.
+
+    .. versionchanged:: 3.1
+       ``min_fresh`` is ``None`` if present without a value, rather
+       than ``"*"``.
+
+    .. versionchanged:: 2.1
+        Setting int properties such as ``max_age`` will convert the
+        value to an int.
+
+    .. versionadded:: 0.5
+        Response-only properties are not present on this request class.
+    """
+
+    no_cache: bool = cache_control_property("no-cache", None, bool)
+    max_stale: int | t.Literal[True] | None = cache_control_property(
+        "max-stale",
+        True,
+        int,
+    )
+    min_fresh: int | None = cache_control_property("min-fresh", None, int)
+    only_if_cached: bool = cache_control_property("only-if-cached", None, bool)
+
+
+class ResponseCacheControl(_CacheControl):
+    """A cache control for responses.  Unlike :class:`RequestCacheControl`
+    this is mutable and gives access to response-relevant cache control
+    headers.
+
+    To get a header of the :class:`ResponseCacheControl` object again you can
+    convert the object into a string or call the :meth:`to_header` method.  If
+    you plan to subclass it and add your own items have a look at the sourcecode
+    for that class.
+
+    .. versionchanged:: 3.1
+        Dict values are always ``str | None``. Setting properties will
+        convert the value to a string. Setting a non-bool property to
+        ``False`` is equivalent to setting it to ``None``. Getting typed
+        properties will return ``None`` if conversion raises
+        ``ValueError``, rather than the string.
+
+    .. versionchanged:: 3.1
+        ``no_cache`` is ``True`` if present without a value, rather than
+        ``"*"``.
+
+    .. versionchanged:: 3.1
+        ``private`` is ``True`` if present without a value, rather than
+        ``"*"``.
+
+    .. versionchanged:: 3.1
+       ``no_transform`` is a boolean. Previously it was mistakenly
+       always ``None``.
+
+    .. versionchanged:: 3.1
+        Added the ``must_understand``, ``stale_while_revalidate``, and
+        ``stale_if_error`` properties.
+
+    .. versionchanged:: 2.1.1
+        ``s_maxage`` converts the value to an int.
+
+    .. versionchanged:: 2.1
+        Setting int properties such as ``max_age`` will convert the
+        value to an int.
+
+    .. versionadded:: 0.5
+       Request-only properties are not present on this response class.
+    """
+
+    no_cache: str | t.Literal[True] | None = cache_control_property(
+        "no-cache", True, None
+    )
+    public: bool = cache_control_property("public", None, bool)
+    private: str | t.Literal[True] | None = cache_control_property(
+        "private", True, None
+    )
+    must_revalidate: bool = cache_control_property("must-revalidate", None, bool)
+    proxy_revalidate: bool = cache_control_property("proxy-revalidate", None, bool)
+    s_maxage: int | None = cache_control_property("s-maxage", None, int)
+    immutable: bool = cache_control_property("immutable", None, bool)
+    must_understand: bool = cache_control_property("must-understand", None, bool)
+    stale_while_revalidate: int | None = cache_control_property(
+        "stale-while-revalidate", None, int
+    )
+
+
+# circular dependencies
+from .. import http
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/csp.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/csp.py
new file mode 100644
index 00000000..0353eebe
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/csp.py
@@ -0,0 +1,100 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import typing as t
+
+from .structures import CallbackDict
+
+
+def csp_property(key: str) -> t.Any:
+    """Return a new property object for a content security policy header.
+    Useful if you want to add support for a csp extension in a
+    subclass.
+    """
+    return property(
+        lambda x: x._get_value(key),
+        lambda x, v: x._set_value(key, v),
+        lambda x: x._del_value(key),
+        f"accessor for {key!r}",
+    )
+
+
+class ContentSecurityPolicy(CallbackDict[str, str]):
+    """Subclass of a dict that stores values for a Content Security Policy
+    header. It has accessors for all the level 3 policies.
+
+    Because the csp directives in the HTTP header use dashes the
+    python descriptors use underscores for that.
+
+    To get a header of the :class:`ContentSecuirtyPolicy` object again
+    you can convert the object into a string or call the
+    :meth:`to_header` method.  If you plan to subclass it and add your
+    own items have a look at the sourcecode for that class.
+
+    .. versionadded:: 1.0.0
+       Support for Content Security Policy headers was added.
+
+    """
+
+    base_uri: str | None = csp_property("base-uri")
+    child_src: str | None = csp_property("child-src")
+    connect_src: str | None = csp_property("connect-src")
+    default_src: str | None = csp_property("default-src")
+    font_src: str | None = csp_property("font-src")
+    form_action: str | None = csp_property("form-action")
+    frame_ancestors: str | None = csp_property("frame-ancestors")
+    frame_src: str | None = csp_property("frame-src")
+    img_src: str | None = csp_property("img-src")
+    manifest_src: str | None = csp_property("manifest-src")
+    media_src: str | None = csp_property("media-src")
+    navigate_to: str | None = csp_property("navigate-to")
+    object_src: str | None = csp_property("object-src")
+    prefetch_src: str | None = csp_property("prefetch-src")
+    plugin_types: str | None = csp_property("plugin-types")
+    report_to: str | None = csp_property("report-to")
+    report_uri: str | None = csp_property("report-uri")
+    sandbox: str | None = csp_property("sandbox")
+    script_src: str | None = csp_property("script-src")
+    script_src_attr: str | None = csp_property("script-src-attr")
+    script_src_elem: str | None = csp_property("script-src-elem")
+    style_src: str | None = csp_property("style-src")
+    style_src_attr: str | None = csp_property("style-src-attr")
+    style_src_elem: str | None = csp_property("style-src-elem")
+    worker_src: str | None = csp_property("worker-src")
+
+    def __init__(
+        self,
+        values: cabc.Mapping[str, str] | cabc.Iterable[tuple[str, str]] | None = (),
+        on_update: cabc.Callable[[ContentSecurityPolicy], None] | None = None,
+    ) -> None:
+        super().__init__(values, on_update)
+        self.provided = values is not None
+
+    def _get_value(self, key: str) -> str | None:
+        """Used internally by the accessor properties."""
+        return self.get(key)
+
+    def _set_value(self, key: str, value: str | None) -> None:
+        """Used internally by the accessor properties."""
+        if value is None:
+            self.pop(key, None)
+        else:
+            self[key] = value
+
+    def _del_value(self, key: str) -> None:
+        """Used internally by the accessor properties."""
+        if key in self:
+            del self[key]
+
+    def to_header(self) -> str:
+        """Convert the stored values into a cache control header."""
+        from ..http import dump_csp_header
+
+        return dump_csp_header(self)
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        kv_str = " ".join(f"{k}={v!r}" for k, v in sorted(self.items()))
+        return f"<{type(self).__name__} {kv_str}>"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/etag.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/etag.py
new file mode 100644
index 00000000..a4ef3424
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/etag.py
@@ -0,0 +1,106 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+
+
+class ETags(cabc.Collection[str]):
+    """A set that can be used to check if one etag is present in a collection
+    of etags.
+    """
+
+    def __init__(
+        self,
+        strong_etags: cabc.Iterable[str] | None = None,
+        weak_etags: cabc.Iterable[str] | None = None,
+        star_tag: bool = False,
+    ):
+        if not star_tag and strong_etags:
+            self._strong = frozenset(strong_etags)
+        else:
+            self._strong = frozenset()
+
+        self._weak = frozenset(weak_etags or ())
+        self.star_tag = star_tag
+
+    def as_set(self, include_weak: bool = False) -> set[str]:
+        """Convert the `ETags` object into a python set.  Per default all the
+        weak etags are not part of this set."""
+        rv = set(self._strong)
+        if include_weak:
+            rv.update(self._weak)
+        return rv
+
+    def is_weak(self, etag: str) -> bool:
+        """Check if an etag is weak."""
+        return etag in self._weak
+
+    def is_strong(self, etag: str) -> bool:
+        """Check if an etag is strong."""
+        return etag in self._strong
+
+    def contains_weak(self, etag: str) -> bool:
+        """Check if an etag is part of the set including weak and strong tags."""
+        return self.is_weak(etag) or self.contains(etag)
+
+    def contains(self, etag: str) -> bool:
+        """Check if an etag is part of the set ignoring weak tags.
+        It is also possible to use the ``in`` operator.
+        """
+        if self.star_tag:
+            return True
+        return self.is_strong(etag)
+
+    def contains_raw(self, etag: str) -> bool:
+        """When passed a quoted tag it will check if this tag is part of the
+        set.  If the tag is weak it is checked against weak and strong tags,
+        otherwise strong only."""
+        from ..http import unquote_etag
+
+        etag, weak = unquote_etag(etag)
+        if weak:
+            return self.contains_weak(etag)
+        return self.contains(etag)
+
+    def to_header(self) -> str:
+        """Convert the etags set into a HTTP header string."""
+        if self.star_tag:
+            return "*"
+        return ", ".join(
+            [f'"{x}"' for x in self._strong] + [f'W/"{x}"' for x in self._weak]
+        )
+
+    def __call__(
+        self,
+        etag: str | None = None,
+        data: bytes | None = None,
+        include_weak: bool = False,
+    ) -> bool:
+        if etag is None:
+            if data is None:
+                raise TypeError("'data' is required when 'etag' is not given.")
+
+            from ..http import generate_etag
+
+            etag = generate_etag(data)
+        if include_weak:
+            if etag in self._weak:
+                return True
+        return etag in self._strong
+
+    def __bool__(self) -> bool:
+        return bool(self.star_tag or self._strong or self._weak)
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __len__(self) -> int:
+        return len(self._strong)
+
+    def __iter__(self) -> cabc.Iterator[str]:
+        return iter(self._strong)
+
+    def __contains__(self, etag: str) -> bool:  # type: ignore[override]
+        return self.contains(etag)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {str(self)!r}>"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/file_storage.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/file_storage.py
new file mode 100644
index 00000000..12342447
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/file_storage.py
@@ -0,0 +1,209 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import mimetypes
+import os
+import typing as t
+from io import BytesIO
+from os import fsdecode
+from os import fspath
+
+from .._internal import _plain_int
+from .headers import Headers
+from .structures import MultiDict
+
+
+class FileStorage:
+    """The :class:`FileStorage` class is a thin wrapper over incoming files.
+    It is used by the request object to represent uploaded files.  All the
+    attributes of the wrapper stream are proxied by the file storage so
+    it's possible to do ``storage.read()`` instead of the long form
+    ``storage.stream.read()``.
+    """
+
+    def __init__(
+        self,
+        stream: t.IO[bytes] | None = None,
+        filename: str | None = None,
+        name: str | None = None,
+        content_type: str | None = None,
+        content_length: int | None = None,
+        headers: Headers | None = None,
+    ):
+        self.name = name
+        self.stream = stream or BytesIO()
+
+        # If no filename is provided, attempt to get the filename from
+        # the stream object. Python names special streams like
+        # ``<stderr>`` with angular brackets, skip these streams.
+        if filename is None:
+            filename = getattr(stream, "name", None)
+
+            if filename is not None:
+                filename = fsdecode(filename)
+
+            if filename and filename[0] == "<" and filename[-1] == ">":
+                filename = None
+        else:
+            filename = fsdecode(filename)
+
+        self.filename = filename
+
+        if headers is None:
+            headers = Headers()
+        self.headers = headers
+        if content_type is not None:
+            headers["Content-Type"] = content_type
+        if content_length is not None:
+            headers["Content-Length"] = str(content_length)
+
+    def _parse_content_type(self) -> None:
+        if not hasattr(self, "_parsed_content_type"):
+            self._parsed_content_type = http.parse_options_header(self.content_type)
+
+    @property
+    def content_type(self) -> str | None:
+        """The content-type sent in the header.  Usually not available"""
+        return self.headers.get("content-type")
+
+    @property
+    def content_length(self) -> int:
+        """The content-length sent in the header.  Usually not available"""
+        if "content-length" in self.headers:
+            try:
+                return _plain_int(self.headers["content-length"])
+            except ValueError:
+                pass
+
+        return 0
+
+    @property
+    def mimetype(self) -> str:
+        """Like :attr:`content_type`, but without parameters (eg, without
+        charset, type etc.) and always lowercase.  For example if the content
+        type is ``text/HTML; charset=utf-8`` the mimetype would be
+        ``'text/html'``.
+
+        .. versionadded:: 0.7
+        """
+        self._parse_content_type()
+        return self._parsed_content_type[0].lower()
+
+    @property
+    def mimetype_params(self) -> dict[str, str]:
+        """The mimetype parameters as dict.  For example if the content
+        type is ``text/html; charset=utf-8`` the params would be
+        ``{'charset': 'utf-8'}``.
+
+        .. versionadded:: 0.7
+        """
+        self._parse_content_type()
+        return self._parsed_content_type[1]
+
+    def save(
+        self, dst: str | os.PathLike[str] | t.IO[bytes], buffer_size: int = 16384
+    ) -> None:
+        """Save the file to a destination path or file object.  If the
+        destination is a file object you have to close it yourself after the
+        call.  The buffer size is the number of bytes held in memory during
+        the copy process.  It defaults to 16KB.
+
+        For secure file saving also have a look at :func:`secure_filename`.
+
+        :param dst: a filename, :class:`os.PathLike`, or open file
+            object to write to.
+        :param buffer_size: Passed as the ``length`` parameter of
+            :func:`shutil.copyfileobj`.
+
+        .. versionchanged:: 1.0
+            Supports :mod:`pathlib`.
+        """
+        from shutil import copyfileobj
+
+        close_dst = False
+
+        if hasattr(dst, "__fspath__"):
+            dst = fspath(dst)
+
+        if isinstance(dst, str):
+            dst = open(dst, "wb")
+            close_dst = True
+
+        try:
+            copyfileobj(self.stream, dst, buffer_size)
+        finally:
+            if close_dst:
+                dst.close()
+
+    def close(self) -> None:
+        """Close the underlying file if possible."""
+        try:
+            self.stream.close()
+        except Exception:
+            pass
+
+    def __bool__(self) -> bool:
+        return bool(self.filename)
+
+    def __getattr__(self, name: str) -> t.Any:
+        try:
+            return getattr(self.stream, name)
+        except AttributeError:
+            # SpooledTemporaryFile on Python < 3.11 doesn't implement IOBase,
+            # get the attribute from its backing file instead.
+            if hasattr(self.stream, "_file"):
+                return getattr(self.stream._file, name)
+            raise
+
+    def __iter__(self) -> cabc.Iterator[bytes]:
+        return iter(self.stream)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__}: {self.filename!r} ({self.content_type!r})>"
+
+
+class FileMultiDict(MultiDict[str, FileStorage]):
+    """A special :class:`MultiDict` that has convenience methods to add
+    files to it.  This is used for :class:`EnvironBuilder` and generally
+    useful for unittesting.
+
+    .. versionadded:: 0.5
+    """
+
+    def add_file(
+        self,
+        name: str,
+        file: str | os.PathLike[str] | t.IO[bytes] | FileStorage,
+        filename: str | None = None,
+        content_type: str | None = None,
+    ) -> None:
+        """Adds a new file to the dict.  `file` can be a file name or
+        a :class:`file`-like or a :class:`FileStorage` object.
+
+        :param name: the name of the field.
+        :param file: a filename or :class:`file`-like object
+        :param filename: an optional filename
+        :param content_type: an optional content type
+        """
+        if isinstance(file, FileStorage):
+            self.add(name, file)
+            return
+
+        if isinstance(file, (str, os.PathLike)):
+            if filename is None:
+                filename = os.fspath(file)
+
+            file_obj: t.IO[bytes] = open(file, "rb")
+        else:
+            file_obj = file  # type: ignore[assignment]
+
+        if filename and content_type is None:
+            content_type = (
+                mimetypes.guess_type(filename)[0] or "application/octet-stream"
+            )
+
+        self.add(name, FileStorage(file_obj, filename, name, content_type))
+
+
+# circular dependencies
+from .. import http
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/headers.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/headers.py
new file mode 100644
index 00000000..1088e3bc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/headers.py
@@ -0,0 +1,662 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import re
+import typing as t
+
+from .._internal import _missing
+from ..exceptions import BadRequestKeyError
+from .mixins import ImmutableHeadersMixin
+from .structures import iter_multi_items
+from .structures import MultiDict
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+    from _typeshed.wsgi import WSGIEnvironment
+
+T = t.TypeVar("T")
+
+
+class Headers:
+    """An object that stores some headers. It has a dict-like interface,
+    but is ordered, can store the same key multiple times, and iterating
+    yields ``(key, value)`` pairs instead of only keys.
+
+    This data structure is useful if you want a nicer way to handle WSGI
+    headers which are stored as tuples in a list.
+
+    From Werkzeug 0.3 onwards, the :exc:`KeyError` raised by this class is
+    also a subclass of the :class:`~exceptions.BadRequest` HTTP exception
+    and will render a page for a ``400 BAD REQUEST`` if caught in a
+    catch-all for HTTP exceptions.
+
+    Headers is mostly compatible with the Python :class:`wsgiref.headers.Headers`
+    class, with the exception of `__getitem__`.  :mod:`wsgiref` will return
+    `None` for ``headers['missing']``, whereas :class:`Headers` will raise
+    a :class:`KeyError`.
+
+    To create a new ``Headers`` object, pass it a list, dict, or
+    other ``Headers`` object with default values. These values are
+    validated the same way values added later are.
+
+    :param defaults: The list of default values for the :class:`Headers`.
+
+    .. versionchanged:: 3.1
+        Implement ``|`` and ``|=`` operators.
+
+    .. versionchanged:: 2.1.0
+        Default values are validated the same as values added later.
+
+    .. versionchanged:: 0.9
+       This data structure now stores unicode values similar to how the
+       multi dicts do it.  The main difference is that bytes can be set as
+       well which will automatically be latin1 decoded.
+
+    .. versionchanged:: 0.9
+       The :meth:`linked` function was removed without replacement as it
+       was an API that does not support the changes to the encoding model.
+    """
+
+    def __init__(
+        self,
+        defaults: (
+            Headers
+            | MultiDict[str, t.Any]
+            | cabc.Mapping[str, t.Any | list[t.Any] | tuple[t.Any, ...] | set[t.Any]]
+            | cabc.Iterable[tuple[str, t.Any]]
+            | None
+        ) = None,
+    ) -> None:
+        self._list: list[tuple[str, str]] = []
+
+        if defaults is not None:
+            self.extend(defaults)
+
+    @t.overload
+    def __getitem__(self, key: str) -> str: ...
+    @t.overload
+    def __getitem__(self, key: int) -> tuple[str, str]: ...
+    @t.overload
+    def __getitem__(self, key: slice) -> te.Self: ...
+    def __getitem__(self, key: str | int | slice) -> str | tuple[str, str] | te.Self:
+        if isinstance(key, str):
+            return self._get_key(key)
+
+        if isinstance(key, int):
+            return self._list[key]
+
+        return self.__class__(self._list[key])
+
+    def _get_key(self, key: str) -> str:
+        ikey = key.lower()
+
+        for k, v in self._list:
+            if k.lower() == ikey:
+                return v
+
+        raise BadRequestKeyError(key)
+
+    def __eq__(self, other: object) -> bool:
+        if other.__class__ is not self.__class__:
+            return NotImplemented
+
+        def lowered(item: tuple[str, ...]) -> tuple[str, ...]:
+            return item[0].lower(), *item[1:]
+
+        return set(map(lowered, other._list)) == set(map(lowered, self._list))  # type: ignore[attr-defined]
+
+    __hash__ = None  # type: ignore[assignment]
+
+    @t.overload
+    def get(self, key: str) -> str | None: ...
+    @t.overload
+    def get(self, key: str, default: str) -> str: ...
+    @t.overload
+    def get(self, key: str, default: T) -> str | T: ...
+    @t.overload
+    def get(self, key: str, type: cabc.Callable[[str], T]) -> T | None: ...
+    @t.overload
+    def get(self, key: str, default: T, type: cabc.Callable[[str], T]) -> T: ...
+    def get(  # type: ignore[misc]
+        self,
+        key: str,
+        default: str | T | None = None,
+        type: cabc.Callable[[str], T] | None = None,
+    ) -> str | T | None:
+        """Return the default value if the requested data doesn't exist.
+        If `type` is provided and is a callable it should convert the value,
+        return it or raise a :exc:`ValueError` if that is not possible.  In
+        this case the function will return the default as if the value was not
+        found:
+
+        >>> d = Headers([('Content-Length', '42')])
+        >>> d.get('Content-Length', type=int)
+        42
+
+        :param key: The key to be looked up.
+        :param default: The default value to be returned if the key can't
+                        be looked up.  If not further specified `None` is
+                        returned.
+        :param type: A callable that is used to cast the value in the
+                     :class:`Headers`.  If a :exc:`ValueError` is raised
+                     by this callable the default value is returned.
+
+        .. versionchanged:: 3.0
+            The ``as_bytes`` parameter was removed.
+
+        .. versionchanged:: 0.9
+            The ``as_bytes`` parameter was added.
+        """
+        try:
+            rv = self._get_key(key)
+        except KeyError:
+            return default
+
+        if type is None:
+            return rv
+
+        try:
+            return type(rv)
+        except ValueError:
+            return default
+
+    @t.overload
+    def getlist(self, key: str) -> list[str]: ...
+    @t.overload
+    def getlist(self, key: str, type: cabc.Callable[[str], T]) -> list[T]: ...
+    def getlist(
+        self, key: str, type: cabc.Callable[[str], T] | None = None
+    ) -> list[str] | list[T]:
+        """Return the list of items for a given key. If that key is not in the
+        :class:`Headers`, the return value will be an empty list.  Just like
+        :meth:`get`, :meth:`getlist` accepts a `type` parameter.  All items will
+        be converted with the callable defined there.
+
+        :param key: The key to be looked up.
+        :param type: A callable that is used to cast the value in the
+                     :class:`Headers`.  If a :exc:`ValueError` is raised
+                     by this callable the value will be removed from the list.
+        :return: a :class:`list` of all the values for the key.
+
+        .. versionchanged:: 3.0
+            The ``as_bytes`` parameter was removed.
+
+        .. versionchanged:: 0.9
+            The ``as_bytes`` parameter was added.
+        """
+        ikey = key.lower()
+
+        if type is not None:
+            result = []
+
+            for k, v in self:
+                if k.lower() == ikey:
+                    try:
+                        result.append(type(v))
+                    except ValueError:
+                        continue
+
+            return result
+
+        return [v for k, v in self if k.lower() == ikey]
+
+    def get_all(self, name: str) -> list[str]:
+        """Return a list of all the values for the named field.
+
+        This method is compatible with the :mod:`wsgiref`
+        :meth:`~wsgiref.headers.Headers.get_all` method.
+        """
+        return self.getlist(name)
+
+    def items(self, lower: bool = False) -> t.Iterable[tuple[str, str]]:
+        for key, value in self:
+            if lower:
+                key = key.lower()
+            yield key, value
+
+    def keys(self, lower: bool = False) -> t.Iterable[str]:
+        for key, _ in self.items(lower):
+            yield key
+
+    def values(self) -> t.Iterable[str]:
+        for _, value in self.items():
+            yield value
+
+    def extend(
+        self,
+        arg: (
+            Headers
+            | MultiDict[str, t.Any]
+            | cabc.Mapping[str, t.Any | list[t.Any] | tuple[t.Any, ...] | set[t.Any]]
+            | cabc.Iterable[tuple[str, t.Any]]
+            | None
+        ) = None,
+        /,
+        **kwargs: str,
+    ) -> None:
+        """Extend headers in this object with items from another object
+        containing header items as well as keyword arguments.
+
+        To replace existing keys instead of extending, use
+        :meth:`update` instead.
+
+        If provided, the first argument can be another :class:`Headers`
+        object, a :class:`MultiDict`, :class:`dict`, or iterable of
+        pairs.
+
+        .. versionchanged:: 1.0
+            Support :class:`MultiDict`. Allow passing ``kwargs``.
+        """
+        if arg is not None:
+            for key, value in iter_multi_items(arg):
+                self.add(key, value)
+
+        for key, value in iter_multi_items(kwargs):
+            self.add(key, value)
+
+    def __delitem__(self, key: str | int | slice) -> None:
+        if isinstance(key, str):
+            self._del_key(key)
+            return
+
+        del self._list[key]
+
+    def _del_key(self, key: str) -> None:
+        key = key.lower()
+        new = []
+
+        for k, v in self._list:
+            if k.lower() != key:
+                new.append((k, v))
+
+        self._list[:] = new
+
+    def remove(self, key: str) -> None:
+        """Remove a key.
+
+        :param key: The key to be removed.
+        """
+        return self._del_key(key)
+
+    @t.overload
+    def pop(self) -> tuple[str, str]: ...
+    @t.overload
+    def pop(self, key: str) -> str: ...
+    @t.overload
+    def pop(self, key: int | None = ...) -> tuple[str, str]: ...
+    @t.overload
+    def pop(self, key: str, default: str) -> str: ...
+    @t.overload
+    def pop(self, key: str, default: T) -> str | T: ...
+    def pop(
+        self,
+        key: str | int | None = None,
+        default: str | T = _missing,  # type: ignore[assignment]
+    ) -> str | tuple[str, str] | T:
+        """Removes and returns a key or index.
+
+        :param key: The key to be popped.  If this is an integer the item at
+                    that position is removed, if it's a string the value for
+                    that key is.  If the key is omitted or `None` the last
+                    item is removed.
+        :return: an item.
+        """
+        if key is None:
+            return self._list.pop()
+
+        if isinstance(key, int):
+            return self._list.pop(key)
+
+        try:
+            rv = self._get_key(key)
+        except KeyError:
+            if default is not _missing:
+                return default
+
+            raise
+
+        self.remove(key)
+        return rv
+
+    def popitem(self) -> tuple[str, str]:
+        """Removes a key or index and returns a (key, value) item."""
+        return self._list.pop()
+
+    def __contains__(self, key: str) -> bool:
+        """Check if a key is present."""
+        try:
+            self._get_key(key)
+        except KeyError:
+            return False
+
+        return True
+
+    def __iter__(self) -> t.Iterator[tuple[str, str]]:
+        """Yield ``(key, value)`` tuples."""
+        return iter(self._list)
+
+    def __len__(self) -> int:
+        return len(self._list)
+
+    def add(self, key: str, value: t.Any, /, **kwargs: t.Any) -> None:
+        """Add a new header tuple to the list.
+
+        Keyword arguments can specify additional parameters for the header
+        value, with underscores converted to dashes::
+
+        >>> d = Headers()
+        >>> d.add('Content-Type', 'text/plain')
+        >>> d.add('Content-Disposition', 'attachment', filename='foo.png')
+
+        The keyword argument dumping uses :func:`dump_options_header`
+        behind the scenes.
+
+        .. versionchanged:: 0.4.1
+            keyword arguments were added for :mod:`wsgiref` compatibility.
+        """
+        if kwargs:
+            value = _options_header_vkw(value, kwargs)
+
+        value_str = _str_header_value(value)
+        self._list.append((key, value_str))
+
+    def add_header(self, key: str, value: t.Any, /, **kwargs: t.Any) -> None:
+        """Add a new header tuple to the list.
+
+        An alias for :meth:`add` for compatibility with the :mod:`wsgiref`
+        :meth:`~wsgiref.headers.Headers.add_header` method.
+        """
+        self.add(key, value, **kwargs)
+
+    def clear(self) -> None:
+        """Clears all headers."""
+        self._list.clear()
+
+    def set(self, key: str, value: t.Any, /, **kwargs: t.Any) -> None:
+        """Remove all header tuples for `key` and add a new one.  The newly
+        added key either appears at the end of the list if there was no
+        entry or replaces the first one.
+
+        Keyword arguments can specify additional parameters for the header
+        value, with underscores converted to dashes.  See :meth:`add` for
+        more information.
+
+        .. versionchanged:: 0.6.1
+           :meth:`set` now accepts the same arguments as :meth:`add`.
+
+        :param key: The key to be inserted.
+        :param value: The value to be inserted.
+        """
+        if kwargs:
+            value = _options_header_vkw(value, kwargs)
+
+        value_str = _str_header_value(value)
+
+        if not self._list:
+            self._list.append((key, value_str))
+            return
+
+        iter_list = iter(self._list)
+        ikey = key.lower()
+
+        for idx, (old_key, _) in enumerate(iter_list):
+            if old_key.lower() == ikey:
+                # replace first occurrence
+                self._list[idx] = (key, value_str)
+                break
+        else:
+            # no existing occurrences
+            self._list.append((key, value_str))
+            return
+
+        # remove remaining occurrences
+        self._list[idx + 1 :] = [t for t in iter_list if t[0].lower() != ikey]
+
+    def setlist(self, key: str, values: cabc.Iterable[t.Any]) -> None:
+        """Remove any existing values for a header and add new ones.
+
+        :param key: The header key to set.
+        :param values: An iterable of values to set for the key.
+
+        .. versionadded:: 1.0
+        """
+        if values:
+            values_iter = iter(values)
+            self.set(key, next(values_iter))
+
+            for value in values_iter:
+                self.add(key, value)
+        else:
+            self.remove(key)
+
+    def setdefault(self, key: str, default: t.Any) -> str:
+        """Return the first value for the key if it is in the headers,
+        otherwise set the header to the value given by ``default`` and
+        return that.
+
+        :param key: The header key to get.
+        :param default: The value to set for the key if it is not in the
+            headers.
+        """
+        try:
+            return self._get_key(key)
+        except KeyError:
+            pass
+
+        self.set(key, default)
+        return self._get_key(key)
+
+    def setlistdefault(self, key: str, default: cabc.Iterable[t.Any]) -> list[str]:
+        """Return the list of values for the key if it is in the
+        headers, otherwise set the header to the list of values given
+        by ``default`` and return that.
+
+        Unlike :meth:`MultiDict.setlistdefault`, modifying the returned
+        list will not affect the headers.
+
+        :param key: The header key to get.
+        :param default: An iterable of values to set for the key if it
+            is not in the headers.
+
+        .. versionadded:: 1.0
+        """
+        if key not in self:
+            self.setlist(key, default)
+
+        return self.getlist(key)
+
+    @t.overload
+    def __setitem__(self, key: str, value: t.Any) -> None: ...
+    @t.overload
+    def __setitem__(self, key: int, value: tuple[str, t.Any]) -> None: ...
+    @t.overload
+    def __setitem__(
+        self, key: slice, value: cabc.Iterable[tuple[str, t.Any]]
+    ) -> None: ...
+    def __setitem__(
+        self,
+        key: str | int | slice,
+        value: t.Any | tuple[str, t.Any] | cabc.Iterable[tuple[str, t.Any]],
+    ) -> None:
+        """Like :meth:`set` but also supports index/slice based setting."""
+        if isinstance(key, str):
+            self.set(key, value)
+        elif isinstance(key, int):
+            self._list[key] = value[0], _str_header_value(value[1])  # type: ignore[index]
+        else:
+            self._list[key] = [(k, _str_header_value(v)) for k, v in value]  # type: ignore[misc]
+
+    def update(
+        self,
+        arg: (
+            Headers
+            | MultiDict[str, t.Any]
+            | cabc.Mapping[
+                str, t.Any | list[t.Any] | tuple[t.Any, ...] | cabc.Set[t.Any]
+            ]
+            | cabc.Iterable[tuple[str, t.Any]]
+            | None
+        ) = None,
+        /,
+        **kwargs: t.Any | list[t.Any] | tuple[t.Any, ...] | cabc.Set[t.Any],
+    ) -> None:
+        """Replace headers in this object with items from another
+        headers object and keyword arguments.
+
+        To extend existing keys instead of replacing, use :meth:`extend`
+        instead.
+
+        If provided, the first argument can be another :class:`Headers`
+        object, a :class:`MultiDict`, :class:`dict`, or iterable of
+        pairs.
+
+        .. versionadded:: 1.0
+        """
+        if arg is not None:
+            if isinstance(arg, (Headers, MultiDict)):
+                for key in arg.keys():
+                    self.setlist(key, arg.getlist(key))
+            elif isinstance(arg, cabc.Mapping):
+                for key, value in arg.items():
+                    if isinstance(value, (list, tuple, set)):
+                        self.setlist(key, value)
+                    else:
+                        self.set(key, value)
+            else:
+                for key, value in arg:
+                    self.set(key, value)
+
+        for key, value in kwargs.items():
+            if isinstance(value, (list, tuple, set)):
+                self.setlist(key, value)
+            else:
+                self.set(key, value)
+
+    def __or__(
+        self,
+        other: cabc.Mapping[
+            str, t.Any | list[t.Any] | tuple[t.Any, ...] | cabc.Set[t.Any]
+        ],
+    ) -> te.Self:
+        if not isinstance(other, cabc.Mapping):
+            return NotImplemented
+
+        rv = self.copy()
+        rv.update(other)
+        return rv
+
+    def __ior__(
+        self,
+        other: (
+            cabc.Mapping[str, t.Any | list[t.Any] | tuple[t.Any, ...] | cabc.Set[t.Any]]
+            | cabc.Iterable[tuple[str, t.Any]]
+        ),
+    ) -> te.Self:
+        if not isinstance(other, (cabc.Mapping, cabc.Iterable)):
+            return NotImplemented
+
+        self.update(other)
+        return self
+
+    def to_wsgi_list(self) -> list[tuple[str, str]]:
+        """Convert the headers into a list suitable for WSGI.
+
+        :return: list
+        """
+        return list(self)
+
+    def copy(self) -> te.Self:
+        return self.__class__(self._list)
+
+    def __copy__(self) -> te.Self:
+        return self.copy()
+
+    def __str__(self) -> str:
+        """Returns formatted headers suitable for HTTP transmission."""
+        strs = []
+        for key, value in self.to_wsgi_list():
+            strs.append(f"{key}: {value}")
+        strs.append("\r\n")
+        return "\r\n".join(strs)
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({list(self)!r})"
+
+
+def _options_header_vkw(value: str, kw: dict[str, t.Any]) -> str:
+    return http.dump_options_header(
+        value, {k.replace("_", "-"): v for k, v in kw.items()}
+    )
+
+
+_newline_re = re.compile(r"[\r\n]")
+
+
+def _str_header_value(value: t.Any) -> str:
+    if not isinstance(value, str):
+        value = str(value)
+
+    if _newline_re.search(value) is not None:
+        raise ValueError("Header values must not contain newline characters.")
+
+    return value  # type: ignore[no-any-return]
+
+
+class EnvironHeaders(ImmutableHeadersMixin, Headers):  # type: ignore[misc]
+    """Read only version of the headers from a WSGI environment.  This
+    provides the same interface as `Headers` and is constructed from
+    a WSGI environment.
+    From Werkzeug 0.3 onwards, the `KeyError` raised by this class is also a
+    subclass of the :exc:`~exceptions.BadRequest` HTTP exception and will
+    render a page for a ``400 BAD REQUEST`` if caught in a catch-all for
+    HTTP exceptions.
+    """
+
+    def __init__(self, environ: WSGIEnvironment) -> None:
+        super().__init__()
+        self.environ = environ
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, EnvironHeaders):
+            return NotImplemented
+
+        return self.environ is other.environ
+
+    __hash__ = None  # type: ignore[assignment]
+
+    def __getitem__(self, key: str) -> str:  # type: ignore[override]
+        return self._get_key(key)
+
+    def _get_key(self, key: str) -> str:
+        if not isinstance(key, str):
+            raise BadRequestKeyError(key)
+
+        key = key.upper().replace("-", "_")
+
+        if key in {"CONTENT_TYPE", "CONTENT_LENGTH"}:
+            return self.environ[key]  # type: ignore[no-any-return]
+
+        return self.environ[f"HTTP_{key}"]  # type: ignore[no-any-return]
+
+    def __len__(self) -> int:
+        return sum(1 for _ in self)
+
+    def __iter__(self) -> cabc.Iterator[tuple[str, str]]:
+        for key, value in self.environ.items():
+            if key.startswith("HTTP_") and key not in {
+                "HTTP_CONTENT_TYPE",
+                "HTTP_CONTENT_LENGTH",
+            }:
+                yield key[5:].replace("_", "-").title(), value
+            elif key in {"CONTENT_TYPE", "CONTENT_LENGTH"} and value:
+                yield key.replace("_", "-").title(), value
+
+    def copy(self) -> t.NoReturn:
+        raise TypeError(f"cannot create {type(self).__name__!r} copies")
+
+    def __or__(self, other: t.Any) -> t.NoReturn:
+        raise TypeError(f"cannot create {type(self).__name__!r} copies")
+
+
+# circular dependencies
+from .. import http
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/mixins.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/mixins.py
new file mode 100644
index 00000000..03d461ad
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/mixins.py
@@ -0,0 +1,317 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import typing as t
+from functools import update_wrapper
+from itertools import repeat
+
+from .._internal import _missing
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+K = t.TypeVar("K")
+V = t.TypeVar("V")
+T = t.TypeVar("T")
+F = t.TypeVar("F", bound=cabc.Callable[..., t.Any])
+
+
+def _immutable_error(self: t.Any) -> t.NoReturn:
+    raise TypeError(f"{type(self).__name__!r} objects are immutable")
+
+
+class ImmutableListMixin:
+    """Makes a :class:`list` immutable.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    _hash_cache: int | None = None
+
+    def __hash__(self) -> int:
+        if self._hash_cache is not None:
+            return self._hash_cache
+        rv = self._hash_cache = hash(tuple(self))  # type: ignore[arg-type]
+        return rv
+
+    def __reduce_ex__(self, protocol: t.SupportsIndex) -> t.Any:
+        return type(self), (list(self),)  # type: ignore[call-overload]
+
+    def __delitem__(self, key: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __iadd__(self, other: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __imul__(self, other: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __setitem__(self, key: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def append(self, item: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def remove(self, item: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def extend(self, iterable: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def insert(self, pos: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def pop(self, index: t.Any = -1) -> t.NoReturn:
+        _immutable_error(self)
+
+    def reverse(self: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def sort(self, key: t.Any = None, reverse: t.Any = False) -> t.NoReturn:
+        _immutable_error(self)
+
+
+class ImmutableDictMixin(t.Generic[K, V]):
+    """Makes a :class:`dict` immutable.
+
+    .. versionchanged:: 3.1
+        Disallow ``|=`` operator.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    _hash_cache: int | None = None
+
+    @classmethod
+    @t.overload
+    def fromkeys(
+        cls, keys: cabc.Iterable[K], value: None
+    ) -> ImmutableDictMixin[K, t.Any | None]: ...
+    @classmethod
+    @t.overload
+    def fromkeys(cls, keys: cabc.Iterable[K], value: V) -> ImmutableDictMixin[K, V]: ...
+    @classmethod
+    def fromkeys(
+        cls, keys: cabc.Iterable[K], value: V | None = None
+    ) -> ImmutableDictMixin[K, t.Any | None] | ImmutableDictMixin[K, V]:
+        instance = super().__new__(cls)
+        instance.__init__(zip(keys, repeat(value)))  # type: ignore[misc]
+        return instance
+
+    def __reduce_ex__(self, protocol: t.SupportsIndex) -> t.Any:
+        return type(self), (dict(self),)  # type: ignore[call-overload]
+
+    def _iter_hashitems(self) -> t.Iterable[t.Any]:
+        return self.items()  # type: ignore[attr-defined,no-any-return]
+
+    def __hash__(self) -> int:
+        if self._hash_cache is not None:
+            return self._hash_cache
+        rv = self._hash_cache = hash(frozenset(self._iter_hashitems()))
+        return rv
+
+    def setdefault(self, key: t.Any, default: t.Any = None) -> t.NoReturn:
+        _immutable_error(self)
+
+    def update(self, arg: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __ior__(self, other: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def pop(self, key: t.Any, default: t.Any = None) -> t.NoReturn:
+        _immutable_error(self)
+
+    def popitem(self) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __setitem__(self, key: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __delitem__(self, key: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def clear(self) -> t.NoReturn:
+        _immutable_error(self)
+
+
+class ImmutableMultiDictMixin(ImmutableDictMixin[K, V]):
+    """Makes a :class:`MultiDict` immutable.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    def __reduce_ex__(self, protocol: t.SupportsIndex) -> t.Any:
+        return type(self), (list(self.items(multi=True)),)  # type: ignore[attr-defined]
+
+    def _iter_hashitems(self) -> t.Iterable[t.Any]:
+        return self.items(multi=True)  # type: ignore[attr-defined,no-any-return]
+
+    def add(self, key: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def popitemlist(self) -> t.NoReturn:
+        _immutable_error(self)
+
+    def poplist(self, key: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def setlist(self, key: t.Any, new_list: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def setlistdefault(self, key: t.Any, default_list: t.Any = None) -> t.NoReturn:
+        _immutable_error(self)
+
+
+class ImmutableHeadersMixin:
+    """Makes a :class:`Headers` immutable.  We do not mark them as
+    hashable though since the only usecase for this datastructure
+    in Werkzeug is a view on a mutable structure.
+
+    .. versionchanged:: 3.1
+        Disallow ``|=`` operator.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    def __delitem__(self, key: t.Any, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __setitem__(self, key: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def set(self, key: t.Any, value: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def setlist(self, key: t.Any, values: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def add(self, key: t.Any, value: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def add_header(self, key: t.Any, value: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def remove(self, key: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def extend(self, arg: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def update(self, arg: t.Any, /, **kwargs: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def __ior__(self, other: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def insert(self, pos: t.Any, value: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def pop(self, key: t.Any = None, default: t.Any = _missing) -> t.NoReturn:
+        _immutable_error(self)
+
+    def popitem(self) -> t.NoReturn:
+        _immutable_error(self)
+
+    def setdefault(self, key: t.Any, default: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+    def setlistdefault(self, key: t.Any, default: t.Any) -> t.NoReturn:
+        _immutable_error(self)
+
+
+def _always_update(f: F) -> F:
+    def wrapper(
+        self: UpdateDictMixin[t.Any, t.Any], /, *args: t.Any, **kwargs: t.Any
+    ) -> t.Any:
+        rv = f(self, *args, **kwargs)
+
+        if self.on_update is not None:
+            self.on_update(self)
+
+        return rv
+
+    return update_wrapper(wrapper, f)  # type: ignore[return-value]
+
+
+class UpdateDictMixin(dict[K, V]):
+    """Makes dicts call `self.on_update` on modifications.
+
+    .. versionchanged:: 3.1
+        Implement ``|=`` operator.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    on_update: cabc.Callable[[te.Self], None] | None = None
+
+    def setdefault(self: te.Self, key: K, default: V | None = None) -> V:
+        modified = key not in self
+        rv = super().setdefault(key, default)  # type: ignore[arg-type]
+        if modified and self.on_update is not None:
+            self.on_update(self)
+        return rv
+
+    @t.overload
+    def pop(self: te.Self, key: K) -> V: ...
+    @t.overload
+    def pop(self: te.Self, key: K, default: V) -> V: ...
+    @t.overload
+    def pop(self: te.Self, key: K, default: T) -> T: ...
+    def pop(
+        self: te.Self,
+        key: K,
+        default: V | T = _missing,  # type: ignore[assignment]
+    ) -> V | T:
+        modified = key in self
+        if default is _missing:
+            rv = super().pop(key)
+        else:
+            rv = super().pop(key, default)  # type: ignore[arg-type]
+        if modified and self.on_update is not None:
+            self.on_update(self)
+        return rv
+
+    @_always_update
+    def __setitem__(self, key: K, value: V) -> None:
+        super().__setitem__(key, value)
+
+    @_always_update
+    def __delitem__(self, key: K) -> None:
+        super().__delitem__(key)
+
+    @_always_update
+    def clear(self) -> None:
+        super().clear()
+
+    @_always_update
+    def popitem(self) -> tuple[K, V]:
+        return super().popitem()
+
+    @_always_update
+    def update(  # type: ignore[override]
+        self,
+        arg: cabc.Mapping[K, V] | cabc.Iterable[tuple[K, V]] | None = None,
+        /,
+        **kwargs: V,
+    ) -> None:
+        if arg is None:
+            super().update(**kwargs)
+        else:
+            super().update(arg, **kwargs)
+
+    @_always_update
+    def __ior__(  # type: ignore[override]
+        self, other: cabc.Mapping[K, V] | cabc.Iterable[tuple[K, V]]
+    ) -> te.Self:
+        return super().__ior__(other)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/range.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/range.py
new file mode 100644
index 00000000..4c9f67d4
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/range.py
@@ -0,0 +1,214 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import typing as t
+from datetime import datetime
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+T = t.TypeVar("T")
+
+
+class IfRange:
+    """Very simple object that represents the `If-Range` header in parsed
+    form.  It will either have neither a etag or date or one of either but
+    never both.
+
+    .. versionadded:: 0.7
+    """
+
+    def __init__(self, etag: str | None = None, date: datetime | None = None):
+        #: The etag parsed and unquoted.  Ranges always operate on strong
+        #: etags so the weakness information is not necessary.
+        self.etag = etag
+        #: The date in parsed format or `None`.
+        self.date = date
+
+    def to_header(self) -> str:
+        """Converts the object back into an HTTP header."""
+        if self.date is not None:
+            return http.http_date(self.date)
+        if self.etag is not None:
+            return http.quote_etag(self.etag)
+        return ""
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {str(self)!r}>"
+
+
+class Range:
+    """Represents a ``Range`` header. All methods only support only
+    bytes as the unit. Stores a list of ranges if given, but the methods
+    only work if only one range is provided.
+
+    :raise ValueError: If the ranges provided are invalid.
+
+    .. versionchanged:: 0.15
+        The ranges passed in are validated.
+
+    .. versionadded:: 0.7
+    """
+
+    def __init__(
+        self, units: str, ranges: cabc.Sequence[tuple[int, int | None]]
+    ) -> None:
+        #: The units of this range.  Usually "bytes".
+        self.units = units
+        #: A list of ``(begin, end)`` tuples for the range header provided.
+        #: The ranges are non-inclusive.
+        self.ranges = ranges
+
+        for start, end in ranges:
+            if start is None or (end is not None and (start < 0 or start >= end)):
+                raise ValueError(f"{(start, end)} is not a valid range.")
+
+    def range_for_length(self, length: int | None) -> tuple[int, int] | None:
+        """If the range is for bytes, the length is not None and there is
+        exactly one range and it is satisfiable it returns a ``(start, stop)``
+        tuple, otherwise `None`.
+        """
+        if self.units != "bytes" or length is None or len(self.ranges) != 1:
+            return None
+        start, end = self.ranges[0]
+        if end is None:
+            end = length
+            if start < 0:
+                start += length
+        if http.is_byte_range_valid(start, end, length):
+            return start, min(end, length)
+        return None
+
+    def make_content_range(self, length: int | None) -> ContentRange | None:
+        """Creates a :class:`~werkzeug.datastructures.ContentRange` object
+        from the current range and given content length.
+        """
+        rng = self.range_for_length(length)
+        if rng is not None:
+            return ContentRange(self.units, rng[0], rng[1], length)
+        return None
+
+    def to_header(self) -> str:
+        """Converts the object back into an HTTP header."""
+        ranges = []
+        for begin, end in self.ranges:
+            if end is None:
+                ranges.append(f"{begin}-" if begin >= 0 else str(begin))
+            else:
+                ranges.append(f"{begin}-{end - 1}")
+        return f"{self.units}={','.join(ranges)}"
+
+    def to_content_range_header(self, length: int | None) -> str | None:
+        """Converts the object into `Content-Range` HTTP header,
+        based on given length
+        """
+        range = self.range_for_length(length)
+        if range is not None:
+            return f"{self.units} {range[0]}-{range[1] - 1}/{length}"
+        return None
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {str(self)!r}>"
+
+
+class _CallbackProperty(t.Generic[T]):
+    def __set_name__(self, owner: type[ContentRange], name: str) -> None:
+        self.attr = f"_{name}"
+
+    @t.overload
+    def __get__(self, instance: None, owner: None) -> te.Self: ...
+    @t.overload
+    def __get__(self, instance: ContentRange, owner: type[ContentRange]) -> T: ...
+    def __get__(
+        self, instance: ContentRange | None, owner: type[ContentRange] | None
+    ) -> te.Self | T:
+        if instance is None:
+            return self
+
+        return instance.__dict__[self.attr]  # type: ignore[no-any-return]
+
+    def __set__(self, instance: ContentRange, value: T) -> None:
+        instance.__dict__[self.attr] = value
+
+        if instance.on_update is not None:
+            instance.on_update(instance)
+
+
+class ContentRange:
+    """Represents the content range header.
+
+    .. versionadded:: 0.7
+    """
+
+    def __init__(
+        self,
+        units: str | None,
+        start: int | None,
+        stop: int | None,
+        length: int | None = None,
+        on_update: cabc.Callable[[ContentRange], None] | None = None,
+    ) -> None:
+        self.on_update = on_update
+        self.set(start, stop, length, units)
+
+    #: The units to use, usually "bytes"
+    units: str | None = _CallbackProperty()  # type: ignore[assignment]
+    #: The start point of the range or `None`.
+    start: int | None = _CallbackProperty()  # type: ignore[assignment]
+    #: The stop point of the range (non-inclusive) or `None`.  Can only be
+    #: `None` if also start is `None`.
+    stop: int | None = _CallbackProperty()  # type: ignore[assignment]
+    #: The length of the range or `None`.
+    length: int | None = _CallbackProperty()  # type: ignore[assignment]
+
+    def set(
+        self,
+        start: int | None,
+        stop: int | None,
+        length: int | None = None,
+        units: str | None = "bytes",
+    ) -> None:
+        """Simple method to update the ranges."""
+        assert http.is_byte_range_valid(start, stop, length), "Bad range provided"
+        self._units: str | None = units
+        self._start: int | None = start
+        self._stop: int | None = stop
+        self._length: int | None = length
+        if self.on_update is not None:
+            self.on_update(self)
+
+    def unset(self) -> None:
+        """Sets the units to `None` which indicates that the header should
+        no longer be used.
+        """
+        self.set(None, None, units=None)
+
+    def to_header(self) -> str:
+        if self._units is None:
+            return ""
+        if self._length is None:
+            length: str | int = "*"
+        else:
+            length = self._length
+        if self._start is None:
+            return f"{self._units} */{length}"
+        return f"{self._units} {self._start}-{self._stop - 1}/{length}"  # type: ignore[operator]
+
+    def __bool__(self) -> bool:
+        return self._units is not None
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {str(self)!r}>"
+
+
+# circular dependencies
+from .. import http
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/structures.py b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/structures.py
new file mode 100644
index 00000000..dbb7e804
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/datastructures/structures.py
@@ -0,0 +1,1239 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import typing as t
+from copy import deepcopy
+
+from .. import exceptions
+from .._internal import _missing
+from .mixins import ImmutableDictMixin
+from .mixins import ImmutableListMixin
+from .mixins import ImmutableMultiDictMixin
+from .mixins import UpdateDictMixin
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+
+K = t.TypeVar("K")
+V = t.TypeVar("V")
+T = t.TypeVar("T")
+
+
+def iter_multi_items(
+    mapping: (
+        MultiDict[K, V]
+        | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+        | cabc.Iterable[tuple[K, V]]
+    ),
+) -> cabc.Iterator[tuple[K, V]]:
+    """Iterates over the items of a mapping yielding keys and values
+    without dropping any from more complex structures.
+    """
+    if isinstance(mapping, MultiDict):
+        yield from mapping.items(multi=True)
+    elif isinstance(mapping, cabc.Mapping):
+        for key, value in mapping.items():
+            if isinstance(value, (list, tuple, set)):
+                for v in value:
+                    yield key, v
+            else:
+                yield key, value
+    else:
+        yield from mapping
+
+
+class ImmutableList(ImmutableListMixin, list[V]):  # type: ignore[misc]
+    """An immutable :class:`list`.
+
+    .. versionadded:: 0.5
+
+    :private:
+    """
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({list.__repr__(self)})"
+
+
+class TypeConversionDict(dict[K, V]):
+    """Works like a regular dict but the :meth:`get` method can perform
+    type conversions.  :class:`MultiDict` and :class:`CombinedMultiDict`
+    are subclasses of this class and provide the same feature.
+
+    .. versionadded:: 0.5
+    """
+
+    @t.overload  # type: ignore[override]
+    def get(self, key: K) -> V | None: ...
+    @t.overload
+    def get(self, key: K, default: V) -> V: ...
+    @t.overload
+    def get(self, key: K, default: T) -> V | T: ...
+    @t.overload
+    def get(self, key: str, type: cabc.Callable[[V], T]) -> T | None: ...
+    @t.overload
+    def get(self, key: str, default: T, type: cabc.Callable[[V], T]) -> T: ...
+    def get(  # type: ignore[misc]
+        self,
+        key: K,
+        default: V | T | None = None,
+        type: cabc.Callable[[V], T] | None = None,
+    ) -> V | T | None:
+        """Return the default value if the requested data doesn't exist.
+        If `type` is provided and is a callable it should convert the value,
+        return it or raise a :exc:`ValueError` if that is not possible.  In
+        this case the function will return the default as if the value was not
+        found:
+
+        >>> d = TypeConversionDict(foo='42', bar='blub')
+        >>> d.get('foo', type=int)
+        42
+        >>> d.get('bar', -1, type=int)
+        -1
+
+        :param key: The key to be looked up.
+        :param default: The default value to be returned if the key can't
+                        be looked up.  If not further specified `None` is
+                        returned.
+        :param type: A callable that is used to cast the value in the
+                     :class:`MultiDict`.  If a :exc:`ValueError` or a
+                     :exc:`TypeError` is raised by this callable the default
+                     value is returned.
+
+        .. versionchanged:: 3.0.2
+           Returns the default value on :exc:`TypeError`, too.
+        """
+        try:
+            rv = self[key]
+        except KeyError:
+            return default
+
+        if type is None:
+            return rv
+
+        try:
+            return type(rv)
+        except (ValueError, TypeError):
+            return default
+
+
+class ImmutableTypeConversionDict(ImmutableDictMixin[K, V], TypeConversionDict[K, V]):  # type: ignore[misc]
+    """Works like a :class:`TypeConversionDict` but does not support
+    modifications.
+
+    .. versionadded:: 0.5
+    """
+
+    def copy(self) -> TypeConversionDict[K, V]:
+        """Return a shallow mutable copy of this object.  Keep in mind that
+        the standard library's :func:`copy` function is a no-op for this class
+        like for any other python immutable type (eg: :class:`tuple`).
+        """
+        return TypeConversionDict(self)
+
+    def __copy__(self) -> te.Self:
+        return self
+
+
+class MultiDict(TypeConversionDict[K, V]):
+    """A :class:`MultiDict` is a dictionary subclass customized to deal with
+    multiple values for the same key which is for example used by the parsing
+    functions in the wrappers.  This is necessary because some HTML form
+    elements pass multiple values for the same key.
+
+    :class:`MultiDict` implements all standard dictionary methods.
+    Internally, it saves all values for a key as a list, but the standard dict
+    access methods will only return the first value for a key. If you want to
+    gain access to the other values, too, you have to use the `list` methods as
+    explained below.
+
+    Basic Usage:
+
+    >>> d = MultiDict([('a', 'b'), ('a', 'c')])
+    >>> d
+    MultiDict([('a', 'b'), ('a', 'c')])
+    >>> d['a']
+    'b'
+    >>> d.getlist('a')
+    ['b', 'c']
+    >>> 'a' in d
+    True
+
+    It behaves like a normal dict thus all dict functions will only return the
+    first value when multiple values for one key are found.
+
+    From Werkzeug 0.3 onwards, the `KeyError` raised by this class is also a
+    subclass of the :exc:`~exceptions.BadRequest` HTTP exception and will
+    render a page for a ``400 BAD REQUEST`` if caught in a catch-all for HTTP
+    exceptions.
+
+    A :class:`MultiDict` can be constructed from an iterable of
+    ``(key, value)`` tuples, a dict, a :class:`MultiDict` or from Werkzeug 0.2
+    onwards some keyword parameters.
+
+    :param mapping: the initial value for the :class:`MultiDict`.  Either a
+                    regular dict, an iterable of ``(key, value)`` tuples
+                    or `None`.
+
+    .. versionchanged:: 3.1
+        Implement ``|`` and ``|=`` operators.
+    """
+
+    def __init__(
+        self,
+        mapping: (
+            MultiDict[K, V]
+            | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+            | None
+        ) = None,
+    ) -> None:
+        if mapping is None:
+            super().__init__()
+        elif isinstance(mapping, MultiDict):
+            super().__init__((k, vs[:]) for k, vs in mapping.lists())
+        elif isinstance(mapping, cabc.Mapping):
+            tmp = {}
+            for key, value in mapping.items():
+                if isinstance(value, (list, tuple, set)):
+                    value = list(value)
+
+                    if not value:
+                        continue
+                else:
+                    value = [value]
+                tmp[key] = value
+            super().__init__(tmp)  # type: ignore[arg-type]
+        else:
+            tmp = {}
+            for key, value in mapping:
+                tmp.setdefault(key, []).append(value)
+            super().__init__(tmp)  # type: ignore[arg-type]
+
+    def __getstate__(self) -> t.Any:
+        return dict(self.lists())
+
+    def __setstate__(self, value: t.Any) -> None:
+        super().clear()
+        super().update(value)
+
+    def __iter__(self) -> cabc.Iterator[K]:
+        # https://github.com/python/cpython/issues/87412
+        # If __iter__ is not overridden, Python uses a fast path for dict(md),
+        # taking the data directly and getting lists of values, rather than
+        # calling __getitem__ and getting only the first value.
+        return super().__iter__()
+
+    def __getitem__(self, key: K) -> V:
+        """Return the first data value for this key;
+        raises KeyError if not found.
+
+        :param key: The key to be looked up.
+        :raise KeyError: if the key does not exist.
+        """
+
+        if key in self:
+            lst = super().__getitem__(key)
+            if len(lst) > 0:  # type: ignore[arg-type]
+                return lst[0]  # type: ignore[index,no-any-return]
+        raise exceptions.BadRequestKeyError(key)
+
+    def __setitem__(self, key: K, value: V) -> None:
+        """Like :meth:`add` but removes an existing key first.
+
+        :param key: the key for the value.
+        :param value: the value to set.
+        """
+        super().__setitem__(key, [value])  # type: ignore[assignment]
+
+    def add(self, key: K, value: V) -> None:
+        """Adds a new value for the key.
+
+        .. versionadded:: 0.6
+
+        :param key: the key for the value.
+        :param value: the value to add.
+        """
+        super().setdefault(key, []).append(value)  # type: ignore[arg-type,attr-defined]
+
+    @t.overload
+    def getlist(self, key: K) -> list[V]: ...
+    @t.overload
+    def getlist(self, key: K, type: cabc.Callable[[V], T]) -> list[T]: ...
+    def getlist(
+        self, key: K, type: cabc.Callable[[V], T] | None = None
+    ) -> list[V] | list[T]:
+        """Return the list of items for a given key. If that key is not in the
+        `MultiDict`, the return value will be an empty list.  Just like `get`,
+        `getlist` accepts a `type` parameter.  All items will be converted
+        with the callable defined there.
+
+        :param key: The key to be looked up.
+        :param type: Callable to convert each value. If a ``ValueError`` or
+            ``TypeError`` is raised, the value is omitted.
+        :return: a :class:`list` of all the values for the key.
+
+        .. versionchanged:: 3.1
+            Catches ``TypeError`` in addition to ``ValueError``.
+        """
+        try:
+            rv: list[V] = super().__getitem__(key)  # type: ignore[assignment]
+        except KeyError:
+            return []
+        if type is None:
+            return list(rv)
+        result = []
+        for item in rv:
+            try:
+                result.append(type(item))
+            except (ValueError, TypeError):
+                pass
+        return result
+
+    def setlist(self, key: K, new_list: cabc.Iterable[V]) -> None:
+        """Remove the old values for a key and add new ones.  Note that the list
+        you pass the values in will be shallow-copied before it is inserted in
+        the dictionary.
+
+        >>> d = MultiDict()
+        >>> d.setlist('foo', ['1', '2'])
+        >>> d['foo']
+        '1'
+        >>> d.getlist('foo')
+        ['1', '2']
+
+        :param key: The key for which the values are set.
+        :param new_list: An iterable with the new values for the key.  Old values
+                         are removed first.
+        """
+        super().__setitem__(key, list(new_list))  # type: ignore[assignment]
+
+    @t.overload
+    def setdefault(self, key: K) -> None: ...
+    @t.overload
+    def setdefault(self, key: K, default: V) -> V: ...
+    def setdefault(self, key: K, default: V | None = None) -> V | None:
+        """Returns the value for the key if it is in the dict, otherwise it
+        returns `default` and sets that value for `key`.
+
+        :param key: The key to be looked up.
+        :param default: The default value to be returned if the key is not
+                        in the dict.  If not further specified it's `None`.
+        """
+        if key not in self:
+            self[key] = default  # type: ignore[assignment]
+
+        return self[key]
+
+    def setlistdefault(
+        self, key: K, default_list: cabc.Iterable[V] | None = None
+    ) -> list[V]:
+        """Like `setdefault` but sets multiple values.  The list returned
+        is not a copy, but the list that is actually used internally.  This
+        means that you can put new values into the dict by appending items
+        to the list:
+
+        >>> d = MultiDict({"foo": 1})
+        >>> d.setlistdefault("foo").extend([2, 3])
+        >>> d.getlist("foo")
+        [1, 2, 3]
+
+        :param key: The key to be looked up.
+        :param default_list: An iterable of default values.  It is either copied
+                             (in case it was a list) or converted into a list
+                             before returned.
+        :return: a :class:`list`
+        """
+        if key not in self:
+            super().__setitem__(key, list(default_list or ()))  # type: ignore[assignment]
+
+        return super().__getitem__(key)  # type: ignore[return-value]
+
+    def items(self, multi: bool = False) -> cabc.Iterable[tuple[K, V]]:  # type: ignore[override]
+        """Return an iterator of ``(key, value)`` pairs.
+
+        :param multi: If set to `True` the iterator returned will have a pair
+                      for each value of each key.  Otherwise it will only
+                      contain pairs for the first value of each key.
+        """
+        values: list[V]
+
+        for key, values in super().items():  # type: ignore[assignment]
+            if multi:
+                for value in values:
+                    yield key, value
+            else:
+                yield key, values[0]
+
+    def lists(self) -> cabc.Iterable[tuple[K, list[V]]]:
+        """Return a iterator of ``(key, values)`` pairs, where values is the list
+        of all values associated with the key."""
+        values: list[V]
+
+        for key, values in super().items():  # type: ignore[assignment]
+            yield key, list(values)
+
+    def values(self) -> cabc.Iterable[V]:  # type: ignore[override]
+        """Returns an iterator of the first value on every key's value list."""
+        values: list[V]
+
+        for values in super().values():  # type: ignore[assignment]
+            yield values[0]
+
+    def listvalues(self) -> cabc.Iterable[list[V]]:
+        """Return an iterator of all values associated with a key.  Zipping
+        :meth:`keys` and this is the same as calling :meth:`lists`:
+
+        >>> d = MultiDict({"foo": [1, 2, 3]})
+        >>> zip(d.keys(), d.listvalues()) == d.lists()
+        True
+        """
+        return super().values()  # type: ignore[return-value]
+
+    def copy(self) -> te.Self:
+        """Return a shallow copy of this object."""
+        return self.__class__(self)
+
+    def deepcopy(self, memo: t.Any = None) -> te.Self:
+        """Return a deep copy of this object."""
+        return self.__class__(deepcopy(self.to_dict(flat=False), memo))
+
+    @t.overload
+    def to_dict(self) -> dict[K, V]: ...
+    @t.overload
+    def to_dict(self, flat: t.Literal[False]) -> dict[K, list[V]]: ...
+    def to_dict(self, flat: bool = True) -> dict[K, V] | dict[K, list[V]]:
+        """Return the contents as regular dict.  If `flat` is `True` the
+        returned dict will only have the first item present, if `flat` is
+        `False` all values will be returned as lists.
+
+        :param flat: If set to `False` the dict returned will have lists
+                     with all the values in it.  Otherwise it will only
+                     contain the first value for each key.
+        :return: a :class:`dict`
+        """
+        if flat:
+            return dict(self.items())
+        return dict(self.lists())
+
+    def update(  # type: ignore[override]
+        self,
+        mapping: (
+            MultiDict[K, V]
+            | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+        ),
+    ) -> None:
+        """update() extends rather than replaces existing key lists:
+
+        >>> a = MultiDict({'x': 1})
+        >>> b = MultiDict({'x': 2, 'y': 3})
+        >>> a.update(b)
+        >>> a
+        MultiDict([('y', 3), ('x', 1), ('x', 2)])
+
+        If the value list for a key in ``other_dict`` is empty, no new values
+        will be added to the dict and the key will not be created:
+
+        >>> x = {'empty_list': []}
+        >>> y = MultiDict()
+        >>> y.update(x)
+        >>> y
+        MultiDict([])
+        """
+        for key, value in iter_multi_items(mapping):
+            self.add(key, value)
+
+    def __or__(  # type: ignore[override]
+        self, other: cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+    ) -> MultiDict[K, V]:
+        if not isinstance(other, cabc.Mapping):
+            return NotImplemented
+
+        rv = self.copy()
+        rv.update(other)
+        return rv
+
+    def __ior__(  # type: ignore[override]
+        self,
+        other: (
+            cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+        ),
+    ) -> te.Self:
+        if not isinstance(other, (cabc.Mapping, cabc.Iterable)):
+            return NotImplemented
+
+        self.update(other)
+        return self
+
+    @t.overload
+    def pop(self, key: K) -> V: ...
+    @t.overload
+    def pop(self, key: K, default: V) -> V: ...
+    @t.overload
+    def pop(self, key: K, default: T) -> V | T: ...
+    def pop(
+        self,
+        key: K,
+        default: V | T = _missing,  # type: ignore[assignment]
+    ) -> V | T:
+        """Pop the first item for a list on the dict.  Afterwards the
+        key is removed from the dict, so additional values are discarded:
+
+        >>> d = MultiDict({"foo": [1, 2, 3]})
+        >>> d.pop("foo")
+        1
+        >>> "foo" in d
+        False
+
+        :param key: the key to pop.
+        :param default: if provided the value to return if the key was
+                        not in the dictionary.
+        """
+        lst: list[V]
+
+        try:
+            lst = super().pop(key)  # type: ignore[assignment]
+
+            if len(lst) == 0:
+                raise exceptions.BadRequestKeyError(key)
+
+            return lst[0]
+        except KeyError:
+            if default is not _missing:
+                return default
+
+            raise exceptions.BadRequestKeyError(key) from None
+
+    def popitem(self) -> tuple[K, V]:
+        """Pop an item from the dict."""
+        item: tuple[K, list[V]]
+
+        try:
+            item = super().popitem()  # type: ignore[assignment]
+
+            if len(item[1]) == 0:
+                raise exceptions.BadRequestKeyError(item[0])
+
+            return item[0], item[1][0]
+        except KeyError as e:
+            raise exceptions.BadRequestKeyError(e.args[0]) from None
+
+    def poplist(self, key: K) -> list[V]:
+        """Pop the list for a key from the dict.  If the key is not in the dict
+        an empty list is returned.
+
+        .. versionchanged:: 0.5
+           If the key does no longer exist a list is returned instead of
+           raising an error.
+        """
+        return super().pop(key, [])  # type: ignore[return-value]
+
+    def popitemlist(self) -> tuple[K, list[V]]:
+        """Pop a ``(key, list)`` tuple from the dict."""
+        try:
+            return super().popitem()  # type: ignore[return-value]
+        except KeyError as e:
+            raise exceptions.BadRequestKeyError(e.args[0]) from None
+
+    def __copy__(self) -> te.Self:
+        return self.copy()
+
+    def __deepcopy__(self, memo: t.Any) -> te.Self:
+        return self.deepcopy(memo=memo)
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({list(self.items(multi=True))!r})"
+
+
+class _omd_bucket(t.Generic[K, V]):
+    """Wraps values in the :class:`OrderedMultiDict`.  This makes it
+    possible to keep an order over multiple different keys.  It requires
+    a lot of extra memory and slows down access a lot, but makes it
+    possible to access elements in O(1) and iterate in O(n).
+    """
+
+    __slots__ = ("prev", "key", "value", "next")
+
+    def __init__(self, omd: _OrderedMultiDict[K, V], key: K, value: V) -> None:
+        self.prev: _omd_bucket[K, V] | None = omd._last_bucket
+        self.key: K = key
+        self.value: V = value
+        self.next: _omd_bucket[K, V] | None = None
+
+        if omd._first_bucket is None:
+            omd._first_bucket = self
+        if omd._last_bucket is not None:
+            omd._last_bucket.next = self
+        omd._last_bucket = self
+
+    def unlink(self, omd: _OrderedMultiDict[K, V]) -> None:
+        if self.prev:
+            self.prev.next = self.next
+        if self.next:
+            self.next.prev = self.prev
+        if omd._first_bucket is self:
+            omd._first_bucket = self.next
+        if omd._last_bucket is self:
+            omd._last_bucket = self.prev
+
+
+class _OrderedMultiDict(MultiDict[K, V]):
+    """Works like a regular :class:`MultiDict` but preserves the
+    order of the fields.  To convert the ordered multi dict into a
+    list you can use the :meth:`items` method and pass it ``multi=True``.
+
+    In general an :class:`OrderedMultiDict` is an order of magnitude
+    slower than a :class:`MultiDict`.
+
+    .. admonition:: note
+
+       Due to a limitation in Python you cannot convert an ordered
+       multi dict into a regular dict by using ``dict(multidict)``.
+       Instead you have to use the :meth:`to_dict` method, otherwise
+       the internal bucket objects are exposed.
+
+    .. deprecated:: 3.1
+        Will be removed in Werkzeug 3.2. Use ``MultiDict`` instead.
+    """
+
+    def __init__(
+        self,
+        mapping: (
+            MultiDict[K, V]
+            | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+            | None
+        ) = None,
+    ) -> None:
+        import warnings
+
+        warnings.warn(
+            "'OrderedMultiDict' is deprecated and will be removed in Werkzeug"
+            " 3.2. Use 'MultiDict' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        super().__init__()
+        self._first_bucket: _omd_bucket[K, V] | None = None
+        self._last_bucket: _omd_bucket[K, V] | None = None
+        if mapping is not None:
+            self.update(mapping)
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, MultiDict):
+            return NotImplemented
+        if isinstance(other, _OrderedMultiDict):
+            iter1 = iter(self.items(multi=True))
+            iter2 = iter(other.items(multi=True))
+            try:
+                for k1, v1 in iter1:
+                    k2, v2 = next(iter2)
+                    if k1 != k2 or v1 != v2:
+                        return False
+            except StopIteration:
+                return False
+            try:
+                next(iter2)
+            except StopIteration:
+                return True
+            return False
+        if len(self) != len(other):
+            return False
+        for key, values in self.lists():
+            if other.getlist(key) != values:
+                return False
+        return True
+
+    __hash__ = None  # type: ignore[assignment]
+
+    def __reduce_ex__(self, protocol: t.SupportsIndex) -> t.Any:
+        return type(self), (list(self.items(multi=True)),)
+
+    def __getstate__(self) -> t.Any:
+        return list(self.items(multi=True))
+
+    def __setstate__(self, values: t.Any) -> None:
+        self.clear()
+
+        for key, value in values:
+            self.add(key, value)
+
+    def __getitem__(self, key: K) -> V:
+        if key in self:
+            return dict.__getitem__(self, key)[0].value  # type: ignore[index,no-any-return]
+        raise exceptions.BadRequestKeyError(key)
+
+    def __setitem__(self, key: K, value: V) -> None:
+        self.poplist(key)
+        self.add(key, value)
+
+    def __delitem__(self, key: K) -> None:
+        self.pop(key)
+
+    def keys(self) -> cabc.Iterable[K]:  # type: ignore[override]
+        return (key for key, _ in self.items())
+
+    def __iter__(self) -> cabc.Iterator[K]:
+        return iter(self.keys())
+
+    def values(self) -> cabc.Iterable[V]:  # type: ignore[override]
+        return (value for key, value in self.items())
+
+    def items(self, multi: bool = False) -> cabc.Iterable[tuple[K, V]]:  # type: ignore[override]
+        ptr = self._first_bucket
+        if multi:
+            while ptr is not None:
+                yield ptr.key, ptr.value
+                ptr = ptr.next
+        else:
+            returned_keys = set()
+            while ptr is not None:
+                if ptr.key not in returned_keys:
+                    returned_keys.add(ptr.key)
+                    yield ptr.key, ptr.value
+                ptr = ptr.next
+
+    def lists(self) -> cabc.Iterable[tuple[K, list[V]]]:
+        returned_keys = set()
+        ptr = self._first_bucket
+        while ptr is not None:
+            if ptr.key not in returned_keys:
+                yield ptr.key, self.getlist(ptr.key)
+                returned_keys.add(ptr.key)
+            ptr = ptr.next
+
+    def listvalues(self) -> cabc.Iterable[list[V]]:
+        for _key, values in self.lists():
+            yield values
+
+    def add(self, key: K, value: V) -> None:
+        dict.setdefault(self, key, []).append(_omd_bucket(self, key, value))  # type: ignore[arg-type,attr-defined]
+
+    @t.overload
+    def getlist(self, key: K) -> list[V]: ...
+    @t.overload
+    def getlist(self, key: K, type: cabc.Callable[[V], T]) -> list[T]: ...
+    def getlist(
+        self, key: K, type: cabc.Callable[[V], T] | None = None
+    ) -> list[V] | list[T]:
+        rv: list[_omd_bucket[K, V]]
+
+        try:
+            rv = dict.__getitem__(self, key)  # type: ignore[index]
+        except KeyError:
+            return []
+        if type is None:
+            return [x.value for x in rv]
+        result = []
+        for item in rv:
+            try:
+                result.append(type(item.value))
+            except (ValueError, TypeError):
+                pass
+        return result
+
+    def setlist(self, key: K, new_list: cabc.Iterable[V]) -> None:
+        self.poplist(key)
+        for value in new_list:
+            self.add(key, value)
+
+    def setlistdefault(self, key: t.Any, default_list: t.Any = None) -> t.NoReturn:
+        raise TypeError("setlistdefault is unsupported for ordered multi dicts")
+
+    def update(  # type: ignore[override]
+        self,
+        mapping: (
+            MultiDict[K, V]
+            | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+        ),
+    ) -> None:
+        for key, value in iter_multi_items(mapping):
+            self.add(key, value)
+
+    def poplist(self, key: K) -> list[V]:
+        buckets: cabc.Iterable[_omd_bucket[K, V]] = dict.pop(self, key, ())  # type: ignore[arg-type]
+        for bucket in buckets:
+            bucket.unlink(self)
+        return [x.value for x in buckets]
+
+    @t.overload
+    def pop(self, key: K) -> V: ...
+    @t.overload
+    def pop(self, key: K, default: V) -> V: ...
+    @t.overload
+    def pop(self, key: K, default: T) -> V | T: ...
+    def pop(
+        self,
+        key: K,
+        default: V | T = _missing,  # type: ignore[assignment]
+    ) -> V | T:
+        buckets: list[_omd_bucket[K, V]]
+
+        try:
+            buckets = dict.pop(self, key)  # type: ignore[arg-type]
+        except KeyError:
+            if default is not _missing:
+                return default
+
+            raise exceptions.BadRequestKeyError(key) from None
+
+        for bucket in buckets:
+            bucket.unlink(self)
+
+        return buckets[0].value
+
+    def popitem(self) -> tuple[K, V]:
+        key: K
+        buckets: list[_omd_bucket[K, V]]
+
+        try:
+            key, buckets = dict.popitem(self)  # type: ignore[arg-type,assignment]
+        except KeyError as e:
+            raise exceptions.BadRequestKeyError(e.args[0]) from None
+
+        for bucket in buckets:
+            bucket.unlink(self)
+
+        return key, buckets[0].value
+
+    def popitemlist(self) -> tuple[K, list[V]]:
+        key: K
+        buckets: list[_omd_bucket[K, V]]
+
+        try:
+            key, buckets = dict.popitem(self)  # type: ignore[arg-type,assignment]
+        except KeyError as e:
+            raise exceptions.BadRequestKeyError(e.args[0]) from None
+
+        for bucket in buckets:
+            bucket.unlink(self)
+
+        return key, [x.value for x in buckets]
+
+
+class CombinedMultiDict(ImmutableMultiDictMixin[K, V], MultiDict[K, V]):  # type: ignore[misc]
+    """A read only :class:`MultiDict` that you can pass multiple :class:`MultiDict`
+    instances as sequence and it will combine the return values of all wrapped
+    dicts:
+
+    >>> from werkzeug.datastructures import CombinedMultiDict, MultiDict
+    >>> post = MultiDict([('foo', 'bar')])
+    >>> get = MultiDict([('blub', 'blah')])
+    >>> combined = CombinedMultiDict([get, post])
+    >>> combined['foo']
+    'bar'
+    >>> combined['blub']
+    'blah'
+
+    This works for all read operations and will raise a `TypeError` for
+    methods that usually change data which isn't possible.
+
+    From Werkzeug 0.3 onwards, the `KeyError` raised by this class is also a
+    subclass of the :exc:`~exceptions.BadRequest` HTTP exception and will
+    render a page for a ``400 BAD REQUEST`` if caught in a catch-all for HTTP
+    exceptions.
+    """
+
+    def __reduce_ex__(self, protocol: t.SupportsIndex) -> t.Any:
+        return type(self), (self.dicts,)
+
+    def __init__(self, dicts: cabc.Iterable[MultiDict[K, V]] | None = None) -> None:
+        super().__init__()
+        self.dicts: list[MultiDict[K, V]] = list(dicts or ())
+
+    @classmethod
+    def fromkeys(cls, keys: t.Any, value: t.Any = None) -> t.NoReturn:
+        raise TypeError(f"cannot create {cls.__name__!r} instances by fromkeys")
+
+    def __getitem__(self, key: K) -> V:
+        for d in self.dicts:
+            if key in d:
+                return d[key]
+        raise exceptions.BadRequestKeyError(key)
+
+    @t.overload  # type: ignore[override]
+    def get(self, key: K) -> V | None: ...
+    @t.overload
+    def get(self, key: K, default: V) -> V: ...
+    @t.overload
+    def get(self, key: K, default: T) -> V | T: ...
+    @t.overload
+    def get(self, key: str, type: cabc.Callable[[V], T]) -> T | None: ...
+    @t.overload
+    def get(self, key: str, default: T, type: cabc.Callable[[V], T]) -> T: ...
+    def get(  # type: ignore[misc]
+        self,
+        key: K,
+        default: V | T | None = None,
+        type: cabc.Callable[[V], T] | None = None,
+    ) -> V | T | None:
+        for d in self.dicts:
+            if key in d:
+                if type is not None:
+                    try:
+                        return type(d[key])
+                    except (ValueError, TypeError):
+                        continue
+                return d[key]
+        return default
+
+    @t.overload
+    def getlist(self, key: K) -> list[V]: ...
+    @t.overload
+    def getlist(self, key: K, type: cabc.Callable[[V], T]) -> list[T]: ...
+    def getlist(
+        self, key: K, type: cabc.Callable[[V], T] | None = None
+    ) -> list[V] | list[T]:
+        rv = []
+        for d in self.dicts:
+            rv.extend(d.getlist(key, type))  # type: ignore[arg-type]
+        return rv
+
+    def _keys_impl(self) -> set[K]:
+        """This function exists so __len__ can be implemented more efficiently,
+        saving one list creation from an iterator.
+        """
+        return set(k for d in self.dicts for k in d)
+
+    def keys(self) -> cabc.Iterable[K]:  # type: ignore[override]
+        return self._keys_impl()
+
+    def __iter__(self) -> cabc.Iterator[K]:
+        return iter(self._keys_impl())
+
+    @t.overload  # type: ignore[override]
+    def items(self) -> cabc.Iterable[tuple[K, V]]: ...
+    @t.overload
+    def items(self, multi: t.Literal[True]) -> cabc.Iterable[tuple[K, list[V]]]: ...
+    def items(
+        self, multi: bool = False
+    ) -> cabc.Iterable[tuple[K, V]] | cabc.Iterable[tuple[K, list[V]]]:
+        found = set()
+        for d in self.dicts:
+            for key, value in d.items(multi):
+                if multi:
+                    yield key, value
+                elif key not in found:
+                    found.add(key)
+                    yield key, value
+
+    def values(self) -> cabc.Iterable[V]:  # type: ignore[override]
+        for _, value in self.items():
+            yield value
+
+    def lists(self) -> cabc.Iterable[tuple[K, list[V]]]:
+        rv: dict[K, list[V]] = {}
+        for d in self.dicts:
+            for key, values in d.lists():
+                rv.setdefault(key, []).extend(values)
+        return rv.items()
+
+    def listvalues(self) -> cabc.Iterable[list[V]]:
+        return (x[1] for x in self.lists())
+
+    def copy(self) -> MultiDict[K, V]:  # type: ignore[override]
+        """Return a shallow mutable copy of this object.
+
+        This returns a :class:`MultiDict` representing the data at the
+        time of copying. The copy will no longer reflect changes to the
+        wrapped dicts.
+
+        .. versionchanged:: 0.15
+            Return a mutable :class:`MultiDict`.
+        """
+        return MultiDict(self)
+
+    def __len__(self) -> int:
+        return len(self._keys_impl())
+
+    def __contains__(self, key: K) -> bool:  # type: ignore[override]
+        for d in self.dicts:
+            if key in d:
+                return True
+        return False
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({self.dicts!r})"
+
+
+class ImmutableDict(ImmutableDictMixin[K, V], dict[K, V]):  # type: ignore[misc]
+    """An immutable :class:`dict`.
+
+    .. versionadded:: 0.5
+    """
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({dict.__repr__(self)})"
+
+    def copy(self) -> dict[K, V]:
+        """Return a shallow mutable copy of this object.  Keep in mind that
+        the standard library's :func:`copy` function is a no-op for this class
+        like for any other python immutable type (eg: :class:`tuple`).
+        """
+        return dict(self)
+
+    def __copy__(self) -> te.Self:
+        return self
+
+
+class ImmutableMultiDict(ImmutableMultiDictMixin[K, V], MultiDict[K, V]):  # type: ignore[misc]
+    """An immutable :class:`MultiDict`.
+
+    .. versionadded:: 0.5
+    """
+
+    def copy(self) -> MultiDict[K, V]:  # type: ignore[override]
+        """Return a shallow mutable copy of this object.  Keep in mind that
+        the standard library's :func:`copy` function is a no-op for this class
+        like for any other python immutable type (eg: :class:`tuple`).
+        """
+        return MultiDict(self)
+
+    def __copy__(self) -> te.Self:
+        return self
+
+
+class _ImmutableOrderedMultiDict(  # type: ignore[misc]
+    ImmutableMultiDictMixin[K, V], _OrderedMultiDict[K, V]
+):
+    """An immutable :class:`OrderedMultiDict`.
+
+    .. deprecated:: 3.1
+        Will be removed in Werkzeug 3.2. Use ``ImmutableMultiDict`` instead.
+
+    .. versionadded:: 0.6
+    """
+
+    def __init__(
+        self,
+        mapping: (
+            MultiDict[K, V]
+            | cabc.Mapping[K, V | list[V] | tuple[V, ...] | set[V]]
+            | cabc.Iterable[tuple[K, V]]
+            | None
+        ) = None,
+    ) -> None:
+        super().__init__()
+
+        if mapping is not None:
+            for k, v in iter_multi_items(mapping):
+                _OrderedMultiDict.add(self, k, v)
+
+    def _iter_hashitems(self) -> cabc.Iterable[t.Any]:
+        return enumerate(self.items(multi=True))
+
+    def copy(self) -> _OrderedMultiDict[K, V]:  # type: ignore[override]
+        """Return a shallow mutable copy of this object.  Keep in mind that
+        the standard library's :func:`copy` function is a no-op for this class
+        like for any other python immutable type (eg: :class:`tuple`).
+        """
+        return _OrderedMultiDict(self)
+
+    def __copy__(self) -> te.Self:
+        return self
+
+
+class CallbackDict(UpdateDictMixin[K, V], dict[K, V]):
+    """A dict that calls a function passed every time something is changed.
+    The function is passed the dict instance.
+    """
+
+    def __init__(
+        self,
+        initial: cabc.Mapping[K, V] | cabc.Iterable[tuple[K, V]] | None = None,
+        on_update: cabc.Callable[[te.Self], None] | None = None,
+    ) -> None:
+        if initial is None:
+            super().__init__()
+        else:
+            super().__init__(initial)
+
+        self.on_update = on_update
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {super().__repr__()}>"
+
+
+class HeaderSet(cabc.MutableSet[str]):
+    """Similar to the :class:`ETags` class this implements a set-like structure.
+    Unlike :class:`ETags` this is case insensitive and used for vary, allow, and
+    content-language headers.
+
+    If not constructed using the :func:`parse_set_header` function the
+    instantiation works like this:
+
+    >>> hs = HeaderSet(['foo', 'bar', 'baz'])
+    >>> hs
+    HeaderSet(['foo', 'bar', 'baz'])
+    """
+
+    def __init__(
+        self,
+        headers: cabc.Iterable[str] | None = None,
+        on_update: cabc.Callable[[te.Self], None] | None = None,
+    ) -> None:
+        self._headers = list(headers or ())
+        self._set = {x.lower() for x in self._headers}
+        self.on_update = on_update
+
+    def add(self, header: str) -> None:
+        """Add a new header to the set."""
+        self.update((header,))
+
+    def remove(self: te.Self, header: str) -> None:
+        """Remove a header from the set.  This raises an :exc:`KeyError` if the
+        header is not in the set.
+
+        .. versionchanged:: 0.5
+            In older versions a :exc:`IndexError` was raised instead of a
+            :exc:`KeyError` if the object was missing.
+
+        :param header: the header to be removed.
+        """
+        key = header.lower()
+        if key not in self._set:
+            raise KeyError(header)
+        self._set.remove(key)
+        for idx, key in enumerate(self._headers):
+            if key.lower() == header:
+                del self._headers[idx]
+                break
+        if self.on_update is not None:
+            self.on_update(self)
+
+    def update(self: te.Self, iterable: cabc.Iterable[str]) -> None:
+        """Add all the headers from the iterable to the set.
+
+        :param iterable: updates the set with the items from the iterable.
+        """
+        inserted_any = False
+        for header in iterable:
+            key = header.lower()
+            if key not in self._set:
+                self._headers.append(header)
+                self._set.add(key)
+                inserted_any = True
+        if inserted_any and self.on_update is not None:
+            self.on_update(self)
+
+    def discard(self, header: str) -> None:
+        """Like :meth:`remove` but ignores errors.
+
+        :param header: the header to be discarded.
+        """
+        try:
+            self.remove(header)
+        except KeyError:
+            pass
+
+    def find(self, header: str) -> int:
+        """Return the index of the header in the set or return -1 if not found.
+
+        :param header: the header to be looked up.
+        """
+        header = header.lower()
+        for idx, item in enumerate(self._headers):
+            if item.lower() == header:
+                return idx
+        return -1
+
+    def index(self, header: str) -> int:
+        """Return the index of the header in the set or raise an
+        :exc:`IndexError`.
+
+        :param header: the header to be looked up.
+        """
+        rv = self.find(header)
+        if rv < 0:
+            raise IndexError(header)
+        return rv
+
+    def clear(self: te.Self) -> None:
+        """Clear the set."""
+        self._set.clear()
+        self._headers.clear()
+
+        if self.on_update is not None:
+            self.on_update(self)
+
+    def as_set(self, preserve_casing: bool = False) -> set[str]:
+        """Return the set as real python set type.  When calling this, all
+        the items are converted to lowercase and the ordering is lost.
+
+        :param preserve_casing: if set to `True` the items in the set returned
+                                will have the original case like in the
+                                :class:`HeaderSet`, otherwise they will
+                                be lowercase.
+        """
+        if preserve_casing:
+            return set(self._headers)
+        return set(self._set)
+
+    def to_header(self) -> str:
+        """Convert the header set into an HTTP header string."""
+        return ", ".join(map(http.quote_header_value, self._headers))
+
+    def __getitem__(self, idx: t.SupportsIndex) -> str:
+        return self._headers[idx]
+
+    def __delitem__(self: te.Self, idx: t.SupportsIndex) -> None:
+        rv = self._headers.pop(idx)
+        self._set.remove(rv.lower())
+        if self.on_update is not None:
+            self.on_update(self)
+
+    def __setitem__(self: te.Self, idx: t.SupportsIndex, value: str) -> None:
+        old = self._headers[idx]
+        self._set.remove(old.lower())
+        self._headers[idx] = value
+        self._set.add(value.lower())
+        if self.on_update is not None:
+            self.on_update(self)
+
+    def __contains__(self, header: str) -> bool:  # type: ignore[override]
+        return header.lower() in self._set
+
+    def __len__(self) -> int:
+        return len(self._set)
+
+    def __iter__(self) -> cabc.Iterator[str]:
+        return iter(self._headers)
+
+    def __bool__(self) -> bool:
+        return bool(self._set)
+
+    def __str__(self) -> str:
+        return self.to_header()
+
+    def __repr__(self) -> str:
+        return f"{type(self).__name__}({self._headers!r})"
+
+
+# circular dependencies
+from .. import http
+
+
+def __getattr__(name: str) -> t.Any:
+    import warnings
+
+    if name == "OrderedMultiDict":
+        warnings.warn(
+            "'OrderedMultiDict' is deprecated and will be removed in Werkzeug"
+            " 3.2. Use 'MultiDict' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return _OrderedMultiDict
+
+    if name == "ImmutableOrderedMultiDict":
+        warnings.warn(
+            "'ImmutableOrderedMultiDict' is deprecated and will be removed in"
+            " Werkzeug 3.2. Use 'ImmutableMultiDict' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        return _ImmutableOrderedMultiDict
+
+    raise AttributeError(name)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__init__.py
new file mode 100644
index 00000000..0c4cabd8
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__init__.py
@@ -0,0 +1,565 @@
+from __future__ import annotations
+
+import getpass
+import hashlib
+import json
+import os
+import pkgutil
+import re
+import sys
+import time
+import typing as t
+import uuid
+from contextlib import ExitStack
+from io import BytesIO
+from itertools import chain
+from multiprocessing import Value
+from os.path import basename
+from os.path import join
+from zlib import adler32
+
+from .._internal import _log
+from ..exceptions import NotFound
+from ..exceptions import SecurityError
+from ..http import parse_cookie
+from ..sansio.utils import host_is_trusted
+from ..security import gen_salt
+from ..utils import send_file
+from ..wrappers.request import Request
+from ..wrappers.response import Response
+from .console import Console
+from .tbtools import DebugFrameSummary
+from .tbtools import DebugTraceback
+from .tbtools import render_console_html
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+# A week
+PIN_TIME = 60 * 60 * 24 * 7
+
+
+def hash_pin(pin: str) -> str:
+    return hashlib.sha1(f"{pin} added salt".encode("utf-8", "replace")).hexdigest()[:12]
+
+
+_machine_id: str | bytes | None = None
+
+
+def get_machine_id() -> str | bytes | None:
+    global _machine_id
+
+    if _machine_id is not None:
+        return _machine_id
+
+    def _generate() -> str | bytes | None:
+        linux = b""
+
+        # machine-id is stable across boots, boot_id is not.
+        for filename in "/etc/machine-id", "/proc/sys/kernel/random/boot_id":
+            try:
+                with open(filename, "rb") as f:
+                    value = f.readline().strip()
+            except OSError:
+                continue
+
+            if value:
+                linux += value
+                break
+
+        # Containers share the same machine id, add some cgroup
+        # information. This is used outside containers too but should be
+        # relatively stable across boots.
+        try:
+            with open("/proc/self/cgroup", "rb") as f:
+                linux += f.readline().strip().rpartition(b"/")[2]
+        except OSError:
+            pass
+
+        if linux:
+            return linux
+
+        # On OS X, use ioreg to get the computer's serial number.
+        try:
+            # subprocess may not be available, e.g. Google App Engine
+            # https://github.com/pallets/werkzeug/issues/925
+            from subprocess import PIPE
+            from subprocess import Popen
+
+            dump = Popen(
+                ["ioreg", "-c", "IOPlatformExpertDevice", "-d", "2"], stdout=PIPE
+            ).communicate()[0]
+            match = re.search(b'"serial-number" = <([^>]+)', dump)
+
+            if match is not None:
+                return match.group(1)
+        except (OSError, ImportError):
+            pass
+
+        # On Windows, use winreg to get the machine guid.
+        if sys.platform == "win32":
+            import winreg
+
+            try:
+                with winreg.OpenKey(
+                    winreg.HKEY_LOCAL_MACHINE,
+                    "SOFTWARE\\Microsoft\\Cryptography",
+                    0,
+                    winreg.KEY_READ | winreg.KEY_WOW64_64KEY,
+                ) as rk:
+                    guid: str | bytes
+                    guid_type: int
+                    guid, guid_type = winreg.QueryValueEx(rk, "MachineGuid")
+
+                    if guid_type == winreg.REG_SZ:
+                        return guid.encode()
+
+                    return guid
+            except OSError:
+                pass
+
+        return None
+
+    _machine_id = _generate()
+    return _machine_id
+
+
+class _ConsoleFrame:
+    """Helper class so that we can reuse the frame console code for the
+    standalone console.
+    """
+
+    def __init__(self, namespace: dict[str, t.Any]):
+        self.console = Console(namespace)
+        self.id = 0
+
+    def eval(self, code: str) -> t.Any:
+        return self.console.eval(code)
+
+
+def get_pin_and_cookie_name(
+    app: WSGIApplication,
+) -> tuple[str, str] | tuple[None, None]:
+    """Given an application object this returns a semi-stable 9 digit pin
+    code and a random key.  The hope is that this is stable between
+    restarts to not make debugging particularly frustrating.  If the pin
+    was forcefully disabled this returns `None`.
+
+    Second item in the resulting tuple is the cookie name for remembering.
+    """
+    pin = os.environ.get("WERKZEUG_DEBUG_PIN")
+    rv = None
+    num = None
+
+    # Pin was explicitly disabled
+    if pin == "off":
+        return None, None
+
+    # Pin was provided explicitly
+    if pin is not None and pin.replace("-", "").isdecimal():
+        # If there are separators in the pin, return it directly
+        if "-" in pin:
+            rv = pin
+        else:
+            num = pin
+
+    modname = getattr(app, "__module__", t.cast(object, app).__class__.__module__)
+    username: str | None
+
+    try:
+        # getuser imports the pwd module, which does not exist in Google
+        # App Engine. It may also raise a KeyError if the UID does not
+        # have a username, such as in Docker.
+        username = getpass.getuser()
+    # Python >= 3.13 only raises OSError
+    except (ImportError, KeyError, OSError):
+        username = None
+
+    mod = sys.modules.get(modname)
+
+    # This information only exists to make the cookie unique on the
+    # computer, not as a security feature.
+    probably_public_bits = [
+        username,
+        modname,
+        getattr(app, "__name__", type(app).__name__),
+        getattr(mod, "__file__", None),
+    ]
+
+    # This information is here to make it harder for an attacker to
+    # guess the cookie name.  They are unlikely to be contained anywhere
+    # within the unauthenticated debug page.
+    private_bits = [str(uuid.getnode()), get_machine_id()]
+
+    h = hashlib.sha1()
+    for bit in chain(probably_public_bits, private_bits):
+        if not bit:
+            continue
+        if isinstance(bit, str):
+            bit = bit.encode()
+        h.update(bit)
+    h.update(b"cookiesalt")
+
+    cookie_name = f"__wzd{h.hexdigest()[:20]}"
+
+    # If we need to generate a pin we salt it a bit more so that we don't
+    # end up with the same value and generate out 9 digits
+    if num is None:
+        h.update(b"pinsalt")
+        num = f"{int(h.hexdigest(), 16):09d}"[:9]
+
+    # Format the pincode in groups of digits for easier remembering if
+    # we don't have a result yet.
+    if rv is None:
+        for group_size in 5, 4, 3:
+            if len(num) % group_size == 0:
+                rv = "-".join(
+                    num[x : x + group_size].rjust(group_size, "0")
+                    for x in range(0, len(num), group_size)
+                )
+                break
+        else:
+            rv = num
+
+    return rv, cookie_name
+
+
+class DebuggedApplication:
+    """Enables debugging support for a given application::
+
+        from werkzeug.debug import DebuggedApplication
+        from myapp import app
+        app = DebuggedApplication(app, evalex=True)
+
+    The ``evalex`` argument allows evaluating expressions in any frame
+    of a traceback. This works by preserving each frame with its local
+    state. Some state, such as context globals, cannot be restored with
+    the frame by default. When ``evalex`` is enabled,
+    ``environ["werkzeug.debug.preserve_context"]`` will be a callable
+    that takes a context manager, and can be called multiple times.
+    Each context manager will be entered before evaluating code in the
+    frame, then exited again, so they can perform setup and cleanup for
+    each call.
+
+    :param app: the WSGI application to run debugged.
+    :param evalex: enable exception evaluation feature (interactive
+                   debugging).  This requires a non-forking server.
+    :param request_key: The key that points to the request object in this
+                        environment.  This parameter is ignored in current
+                        versions.
+    :param console_path: the URL for a general purpose console.
+    :param console_init_func: the function that is executed before starting
+                              the general purpose console.  The return value
+                              is used as initial namespace.
+    :param show_hidden_frames: by default hidden traceback frames are skipped.
+                               You can show them by setting this parameter
+                               to `True`.
+    :param pin_security: can be used to disable the pin based security system.
+    :param pin_logging: enables the logging of the pin system.
+
+    .. versionchanged:: 2.2
+        Added the ``werkzeug.debug.preserve_context`` environ key.
+    """
+
+    _pin: str
+    _pin_cookie: str
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        evalex: bool = False,
+        request_key: str = "werkzeug.request",
+        console_path: str = "/console",
+        console_init_func: t.Callable[[], dict[str, t.Any]] | None = None,
+        show_hidden_frames: bool = False,
+        pin_security: bool = True,
+        pin_logging: bool = True,
+    ) -> None:
+        if not console_init_func:
+            console_init_func = None
+        self.app = app
+        self.evalex = evalex
+        self.frames: dict[int, DebugFrameSummary | _ConsoleFrame] = {}
+        self.frame_contexts: dict[int, list[t.ContextManager[None]]] = {}
+        self.request_key = request_key
+        self.console_path = console_path
+        self.console_init_func = console_init_func
+        self.show_hidden_frames = show_hidden_frames
+        self.secret = gen_salt(20)
+        self._failed_pin_auth = Value("B")
+
+        self.pin_logging = pin_logging
+        if pin_security:
+            # Print out the pin for the debugger on standard out.
+            if os.environ.get("WERKZEUG_RUN_MAIN") == "true" and pin_logging:
+                _log("warning", " * Debugger is active!")
+                if self.pin is None:
+                    _log("warning", " * Debugger PIN disabled. DEBUGGER UNSECURED!")
+                else:
+                    _log("info", " * Debugger PIN: %s", self.pin)
+        else:
+            self.pin = None
+
+        self.trusted_hosts: list[str] = [".localhost", "127.0.0.1"]
+        """List of domains to allow requests to the debugger from. A leading dot
+        allows all subdomains. This only allows ``".localhost"`` domains by
+        default.
+
+        .. versionadded:: 3.0.3
+        """
+
+    @property
+    def pin(self) -> str | None:
+        if not hasattr(self, "_pin"):
+            pin_cookie = get_pin_and_cookie_name(self.app)
+            self._pin, self._pin_cookie = pin_cookie  # type: ignore
+        return self._pin
+
+    @pin.setter
+    def pin(self, value: str) -> None:
+        self._pin = value
+
+    @property
+    def pin_cookie_name(self) -> str:
+        """The name of the pin cookie."""
+        if not hasattr(self, "_pin_cookie"):
+            pin_cookie = get_pin_and_cookie_name(self.app)
+            self._pin, self._pin_cookie = pin_cookie  # type: ignore
+        return self._pin_cookie
+
+    def debug_application(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterator[bytes]:
+        """Run the application and conserve the traceback frames."""
+        contexts: list[t.ContextManager[t.Any]] = []
+
+        if self.evalex:
+            environ["werkzeug.debug.preserve_context"] = contexts.append
+
+        app_iter = None
+        try:
+            app_iter = self.app(environ, start_response)
+            yield from app_iter
+            if hasattr(app_iter, "close"):
+                app_iter.close()
+        except Exception as e:
+            if hasattr(app_iter, "close"):
+                app_iter.close()  # type: ignore
+
+            tb = DebugTraceback(e, skip=1, hide=not self.show_hidden_frames)
+
+            for frame in tb.all_frames:
+                self.frames[id(frame)] = frame
+                self.frame_contexts[id(frame)] = contexts
+
+            is_trusted = bool(self.check_pin_trust(environ))
+            html = tb.render_debugger_html(
+                evalex=self.evalex and self.check_host_trust(environ),
+                secret=self.secret,
+                evalex_trusted=is_trusted,
+            )
+            response = Response(html, status=500, mimetype="text/html")
+
+            try:
+                yield from response(environ, start_response)
+            except Exception:
+                # if we end up here there has been output but an error
+                # occurred.  in that situation we can do nothing fancy any
+                # more, better log something into the error log and fall
+                # back gracefully.
+                environ["wsgi.errors"].write(
+                    "Debugging middleware caught exception in streamed "
+                    "response at a point where response headers were already "
+                    "sent.\n"
+                )
+
+            environ["wsgi.errors"].write("".join(tb.render_traceback_text()))
+
+    def execute_command(
+        self,
+        request: Request,
+        command: str,
+        frame: DebugFrameSummary | _ConsoleFrame,
+    ) -> Response:
+        """Execute a command in a console."""
+        if not self.check_host_trust(request.environ):
+            return SecurityError()  # type: ignore[return-value]
+
+        contexts = self.frame_contexts.get(id(frame), [])
+
+        with ExitStack() as exit_stack:
+            for cm in contexts:
+                exit_stack.enter_context(cm)
+
+            return Response(frame.eval(command), mimetype="text/html")
+
+    def display_console(self, request: Request) -> Response:
+        """Display a standalone shell."""
+        if not self.check_host_trust(request.environ):
+            return SecurityError()  # type: ignore[return-value]
+
+        if 0 not in self.frames:
+            if self.console_init_func is None:
+                ns = {}
+            else:
+                ns = dict(self.console_init_func())
+            ns.setdefault("app", self.app)
+            self.frames[0] = _ConsoleFrame(ns)
+        is_trusted = bool(self.check_pin_trust(request.environ))
+        return Response(
+            render_console_html(secret=self.secret, evalex_trusted=is_trusted),
+            mimetype="text/html",
+        )
+
+    def get_resource(self, request: Request, filename: str) -> Response:
+        """Return a static resource from the shared folder."""
+        path = join("shared", basename(filename))
+
+        try:
+            data = pkgutil.get_data(__package__, path)
+        except OSError:
+            return NotFound()  # type: ignore[return-value]
+        else:
+            if data is None:
+                return NotFound()  # type: ignore[return-value]
+
+            etag = str(adler32(data) & 0xFFFFFFFF)
+            return send_file(
+                BytesIO(data), request.environ, download_name=filename, etag=etag
+            )
+
+    def check_pin_trust(self, environ: WSGIEnvironment) -> bool | None:
+        """Checks if the request passed the pin test.  This returns `True` if the
+        request is trusted on a pin/cookie basis and returns `False` if not.
+        Additionally if the cookie's stored pin hash is wrong it will return
+        `None` so that appropriate action can be taken.
+        """
+        if self.pin is None:
+            return True
+        val = parse_cookie(environ).get(self.pin_cookie_name)
+        if not val or "|" not in val:
+            return False
+        ts_str, pin_hash = val.split("|", 1)
+
+        try:
+            ts = int(ts_str)
+        except ValueError:
+            return False
+
+        if pin_hash != hash_pin(self.pin):
+            return None
+        return (time.time() - PIN_TIME) < ts
+
+    def check_host_trust(self, environ: WSGIEnvironment) -> bool:
+        return host_is_trusted(environ.get("HTTP_HOST"), self.trusted_hosts)
+
+    def _fail_pin_auth(self) -> None:
+        with self._failed_pin_auth.get_lock():
+            count = self._failed_pin_auth.value
+            self._failed_pin_auth.value = count + 1
+
+        time.sleep(5.0 if count > 5 else 0.5)
+
+    def pin_auth(self, request: Request) -> Response:
+        """Authenticates with the pin."""
+        if not self.check_host_trust(request.environ):
+            return SecurityError()  # type: ignore[return-value]
+
+        exhausted = False
+        auth = False
+        trust = self.check_pin_trust(request.environ)
+        pin = t.cast(str, self.pin)
+
+        # If the trust return value is `None` it means that the cookie is
+        # set but the stored pin hash value is bad.  This means that the
+        # pin was changed.  In this case we count a bad auth and unset the
+        # cookie.  This way it becomes harder to guess the cookie name
+        # instead of the pin as we still count up failures.
+        bad_cookie = False
+        if trust is None:
+            self._fail_pin_auth()
+            bad_cookie = True
+
+        # If we're trusted, we're authenticated.
+        elif trust:
+            auth = True
+
+        # If we failed too many times, then we're locked out.
+        elif self._failed_pin_auth.value > 10:
+            exhausted = True
+
+        # Otherwise go through pin based authentication
+        else:
+            entered_pin = request.args["pin"]
+
+            if entered_pin.strip().replace("-", "") == pin.replace("-", ""):
+                self._failed_pin_auth.value = 0
+                auth = True
+            else:
+                self._fail_pin_auth()
+
+        rv = Response(
+            json.dumps({"auth": auth, "exhausted": exhausted}),
+            mimetype="application/json",
+        )
+        if auth:
+            rv.set_cookie(
+                self.pin_cookie_name,
+                f"{int(time.time())}|{hash_pin(pin)}",
+                httponly=True,
+                samesite="Strict",
+                secure=request.is_secure,
+            )
+        elif bad_cookie:
+            rv.delete_cookie(self.pin_cookie_name)
+        return rv
+
+    def log_pin_request(self, request: Request) -> Response:
+        """Log the pin if needed."""
+        if not self.check_host_trust(request.environ):
+            return SecurityError()  # type: ignore[return-value]
+
+        if self.pin_logging and self.pin is not None:
+            _log(
+                "info", " * To enable the debugger you need to enter the security pin:"
+            )
+            _log("info", " * Debugger pin code: %s", self.pin)
+        return Response("")
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        """Dispatch the requests."""
+        # important: don't ever access a function here that reads the incoming
+        # form data!  Otherwise the application won't have access to that data
+        # any more!
+        request = Request(environ)
+        response = self.debug_application
+        if request.args.get("__debugger__") == "yes":
+            cmd = request.args.get("cmd")
+            arg = request.args.get("f")
+            secret = request.args.get("s")
+            frame = self.frames.get(request.args.get("frm", type=int))  # type: ignore
+            if cmd == "resource" and arg:
+                response = self.get_resource(request, arg)  # type: ignore
+            elif cmd == "pinauth" and secret == self.secret:
+                response = self.pin_auth(request)  # type: ignore
+            elif cmd == "printpin" and secret == self.secret:
+                response = self.log_pin_request(request)  # type: ignore
+            elif (
+                self.evalex
+                and cmd is not None
+                and frame is not None
+                and self.secret == secret
+                and self.check_pin_trust(environ)
+            ):
+                response = self.execute_command(request, cmd, frame)  # type: ignore
+        elif (
+            self.evalex
+            and self.console_path is not None
+            and request.path == self.console_path
+        ):
+            response = self.display_console(request)  # type: ignore
+        return response(environ, start_response)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..0ad3c796
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/console.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/console.cpython-312.pyc
new file mode 100644
index 00000000..9d55967e
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/console.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/repr.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/repr.cpython-312.pyc
new file mode 100644
index 00000000..72af2761
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/repr.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/tbtools.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/tbtools.cpython-312.pyc
new file mode 100644
index 00000000..adfff6a0
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/__pycache__/tbtools.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/console.py b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/console.py
new file mode 100644
index 00000000..4e40475a
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/console.py
@@ -0,0 +1,219 @@
+from __future__ import annotations
+
+import code
+import sys
+import typing as t
+from contextvars import ContextVar
+from types import CodeType
+
+from markupsafe import escape
+
+from .repr import debug_repr
+from .repr import dump
+from .repr import helper
+
+_stream: ContextVar[HTMLStringO] = ContextVar("werkzeug.debug.console.stream")
+_ipy: ContextVar[_InteractiveConsole] = ContextVar("werkzeug.debug.console.ipy")
+
+
+class HTMLStringO:
+    """A StringO version that HTML escapes on write."""
+
+    def __init__(self) -> None:
+        self._buffer: list[str] = []
+
+    def isatty(self) -> bool:
+        return False
+
+    def close(self) -> None:
+        pass
+
+    def flush(self) -> None:
+        pass
+
+    def seek(self, n: int, mode: int = 0) -> None:
+        pass
+
+    def readline(self) -> str:
+        if len(self._buffer) == 0:
+            return ""
+        ret = self._buffer[0]
+        del self._buffer[0]
+        return ret
+
+    def reset(self) -> str:
+        val = "".join(self._buffer)
+        del self._buffer[:]
+        return val
+
+    def _write(self, x: str) -> None:
+        self._buffer.append(x)
+
+    def write(self, x: str) -> None:
+        self._write(escape(x))
+
+    def writelines(self, x: t.Iterable[str]) -> None:
+        self._write(escape("".join(x)))
+
+
+class ThreadedStream:
+    """Thread-local wrapper for sys.stdout for the interactive console."""
+
+    @staticmethod
+    def push() -> None:
+        if not isinstance(sys.stdout, ThreadedStream):
+            sys.stdout = t.cast(t.TextIO, ThreadedStream())
+
+        _stream.set(HTMLStringO())
+
+    @staticmethod
+    def fetch() -> str:
+        try:
+            stream = _stream.get()
+        except LookupError:
+            return ""
+
+        return stream.reset()
+
+    @staticmethod
+    def displayhook(obj: object) -> None:
+        try:
+            stream = _stream.get()
+        except LookupError:
+            return _displayhook(obj)  # type: ignore
+
+        # stream._write bypasses escaping as debug_repr is
+        # already generating HTML for us.
+        if obj is not None:
+            _ipy.get().locals["_"] = obj
+            stream._write(debug_repr(obj))
+
+    def __setattr__(self, name: str, value: t.Any) -> None:
+        raise AttributeError(f"read only attribute {name}")
+
+    def __dir__(self) -> list[str]:
+        return dir(sys.__stdout__)
+
+    def __getattribute__(self, name: str) -> t.Any:
+        try:
+            stream = _stream.get()
+        except LookupError:
+            stream = sys.__stdout__  # type: ignore[assignment]
+
+        return getattr(stream, name)
+
+    def __repr__(self) -> str:
+        return repr(sys.__stdout__)
+
+
+# add the threaded stream as display hook
+_displayhook = sys.displayhook
+sys.displayhook = ThreadedStream.displayhook
+
+
+class _ConsoleLoader:
+    def __init__(self) -> None:
+        self._storage: dict[int, str] = {}
+
+    def register(self, code: CodeType, source: str) -> None:
+        self._storage[id(code)] = source
+        # register code objects of wrapped functions too.
+        for var in code.co_consts:
+            if isinstance(var, CodeType):
+                self._storage[id(var)] = source
+
+    def get_source_by_code(self, code: CodeType) -> str | None:
+        try:
+            return self._storage[id(code)]
+        except KeyError:
+            return None
+
+
+class _InteractiveConsole(code.InteractiveInterpreter):
+    locals: dict[str, t.Any]
+
+    def __init__(self, globals: dict[str, t.Any], locals: dict[str, t.Any]) -> None:
+        self.loader = _ConsoleLoader()
+        locals = {
+            **globals,
+            **locals,
+            "dump": dump,
+            "help": helper,
+            "__loader__": self.loader,
+        }
+        super().__init__(locals)
+        original_compile = self.compile
+
+        def compile(source: str, filename: str, symbol: str) -> CodeType | None:
+            code = original_compile(source, filename, symbol)
+
+            if code is not None:
+                self.loader.register(code, source)
+
+            return code
+
+        self.compile = compile  # type: ignore[assignment]
+        self.more = False
+        self.buffer: list[str] = []
+
+    def runsource(self, source: str, **kwargs: t.Any) -> str:  # type: ignore
+        source = f"{source.rstrip()}\n"
+        ThreadedStream.push()
+        prompt = "... " if self.more else ">>> "
+        try:
+            source_to_eval = "".join(self.buffer + [source])
+            if super().runsource(source_to_eval, "<debugger>", "single"):
+                self.more = True
+                self.buffer.append(source)
+            else:
+                self.more = False
+                del self.buffer[:]
+        finally:
+            output = ThreadedStream.fetch()
+        return f"{prompt}{escape(source)}{output}"
+
+    def runcode(self, code: CodeType) -> None:
+        try:
+            exec(code, self.locals)
+        except Exception:
+            self.showtraceback()
+
+    def showtraceback(self) -> None:
+        from .tbtools import DebugTraceback
+
+        exc = t.cast(BaseException, sys.exc_info()[1])
+        te = DebugTraceback(exc, skip=1)
+        sys.stdout._write(te.render_traceback_html())  # type: ignore
+
+    def showsyntaxerror(self, filename: str | None = None) -> None:
+        from .tbtools import DebugTraceback
+
+        exc = t.cast(BaseException, sys.exc_info()[1])
+        te = DebugTraceback(exc, skip=4)
+        sys.stdout._write(te.render_traceback_html())  # type: ignore
+
+    def write(self, data: str) -> None:
+        sys.stdout.write(data)
+
+
+class Console:
+    """An interactive console."""
+
+    def __init__(
+        self,
+        globals: dict[str, t.Any] | None = None,
+        locals: dict[str, t.Any] | None = None,
+    ) -> None:
+        if locals is None:
+            locals = {}
+        if globals is None:
+            globals = {}
+        self._ipy = _InteractiveConsole(globals, locals)
+
+    def eval(self, code: str) -> str:
+        _ipy.set(self._ipy)
+        old_sys_stdout = sys.stdout
+        try:
+            return self._ipy.runsource(code)
+        finally:
+            sys.stdout = old_sys_stdout
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/repr.py b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/repr.py
new file mode 100644
index 00000000..2bbd9d54
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/repr.py
@@ -0,0 +1,282 @@
+"""Object representations for debugging purposes. Unlike the default
+repr, these expose more information and produce HTML instead of ASCII.
+
+Together with the CSS and JavaScript of the debugger this gives a
+colorful and more compact output.
+"""
+
+from __future__ import annotations
+
+import codecs
+import re
+import sys
+import typing as t
+from collections import deque
+from traceback import format_exception_only
+
+from markupsafe import escape
+
+missing = object()
+_paragraph_re = re.compile(r"(?:\r\n|\r|\n){2,}")
+RegexType = type(_paragraph_re)
+
+HELP_HTML = """\
+<div class=box>
+  <h3>%(title)s</h3>
+  <pre class=help>%(text)s</pre>
+</div>\
+"""
+OBJECT_DUMP_HTML = """\
+<div class=box>
+  <h3>%(title)s</h3>
+  %(repr)s
+  <table>%(items)s</table>
+</div>\
+"""
+
+
+def debug_repr(obj: object) -> str:
+    """Creates a debug repr of an object as HTML string."""
+    return DebugReprGenerator().repr(obj)
+
+
+def dump(obj: object = missing) -> None:
+    """Print the object details to stdout._write (for the interactive
+    console of the web debugger.
+    """
+    gen = DebugReprGenerator()
+    if obj is missing:
+        rv = gen.dump_locals(sys._getframe(1).f_locals)
+    else:
+        rv = gen.dump_object(obj)
+    sys.stdout._write(rv)  # type: ignore
+
+
+class _Helper:
+    """Displays an HTML version of the normal help, for the interactive
+    debugger only because it requires a patched sys.stdout.
+    """
+
+    def __repr__(self) -> str:
+        return "Type help(object) for help about object."
+
+    def __call__(self, topic: t.Any | None = None) -> None:
+        if topic is None:
+            sys.stdout._write(f"<span class=help>{self!r}</span>")  # type: ignore
+            return
+        import pydoc
+
+        pydoc.help(topic)
+        rv = sys.stdout.reset()  # type: ignore
+        paragraphs = _paragraph_re.split(rv)
+        if len(paragraphs) > 1:
+            title = paragraphs[0]
+            text = "\n\n".join(paragraphs[1:])
+        else:
+            title = "Help"
+            text = paragraphs[0]
+        sys.stdout._write(HELP_HTML % {"title": title, "text": text})  # type: ignore
+
+
+helper = _Helper()
+
+
+def _add_subclass_info(inner: str, obj: object, base: type | tuple[type, ...]) -> str:
+    if isinstance(base, tuple):
+        for cls in base:
+            if type(obj) is cls:
+                return inner
+    elif type(obj) is base:
+        return inner
+    module = ""
+    if obj.__class__.__module__ not in ("__builtin__", "exceptions"):
+        module = f'<span class="module">{obj.__class__.__module__}.</span>'
+    return f"{module}{type(obj).__name__}({inner})"
+
+
+def _sequence_repr_maker(
+    left: str, right: str, base: type, limit: int = 8
+) -> t.Callable[[DebugReprGenerator, t.Iterable[t.Any], bool], str]:
+    def proxy(self: DebugReprGenerator, obj: t.Iterable[t.Any], recursive: bool) -> str:
+        if recursive:
+            return _add_subclass_info(f"{left}...{right}", obj, base)
+        buf = [left]
+        have_extended_section = False
+        for idx, item in enumerate(obj):
+            if idx:
+                buf.append(", ")
+            if idx == limit:
+                buf.append('<span class="extended">')
+                have_extended_section = True
+            buf.append(self.repr(item))
+        if have_extended_section:
+            buf.append("</span>")
+        buf.append(right)
+        return _add_subclass_info("".join(buf), obj, base)
+
+    return proxy
+
+
+class DebugReprGenerator:
+    def __init__(self) -> None:
+        self._stack: list[t.Any] = []
+
+    list_repr = _sequence_repr_maker("[", "]", list)
+    tuple_repr = _sequence_repr_maker("(", ")", tuple)
+    set_repr = _sequence_repr_maker("set([", "])", set)
+    frozenset_repr = _sequence_repr_maker("frozenset([", "])", frozenset)
+    deque_repr = _sequence_repr_maker(
+        '<span class="module">collections.</span>deque([', "])", deque
+    )
+
+    def regex_repr(self, obj: t.Pattern[t.AnyStr]) -> str:
+        pattern = repr(obj.pattern)
+        pattern = codecs.decode(pattern, "unicode-escape", "ignore")
+        pattern = f"r{pattern}"
+        return f're.compile(<span class="string regex">{pattern}</span>)'
+
+    def string_repr(self, obj: str | bytes, limit: int = 70) -> str:
+        buf = ['<span class="string">']
+        r = repr(obj)
+
+        # shorten the repr when the hidden part would be at least 3 chars
+        if len(r) - limit > 2:
+            buf.extend(
+                (
+                    escape(r[:limit]),
+                    '<span class="extended">',
+                    escape(r[limit:]),
+                    "</span>",
+                )
+            )
+        else:
+            buf.append(escape(r))
+
+        buf.append("</span>")
+        out = "".join(buf)
+
+        # if the repr looks like a standard string, add subclass info if needed
+        if r[0] in "'\"" or (r[0] == "b" and r[1] in "'\""):
+            return _add_subclass_info(out, obj, (bytes, str))
+
+        # otherwise, assume the repr distinguishes the subclass already
+        return out
+
+    def dict_repr(
+        self,
+        d: dict[int, None] | dict[str, int] | dict[str | int, int],
+        recursive: bool,
+        limit: int = 5,
+    ) -> str:
+        if recursive:
+            return _add_subclass_info("{...}", d, dict)
+        buf = ["{"]
+        have_extended_section = False
+        for idx, (key, value) in enumerate(d.items()):
+            if idx:
+                buf.append(", ")
+            if idx == limit - 1:
+                buf.append('<span class="extended">')
+                have_extended_section = True
+            buf.append(
+                f'<span class="pair"><span class="key">{self.repr(key)}</span>:'
+                f' <span class="value">{self.repr(value)}</span></span>'
+            )
+        if have_extended_section:
+            buf.append("</span>")
+        buf.append("}")
+        return _add_subclass_info("".join(buf), d, dict)
+
+    def object_repr(self, obj: t.Any) -> str:
+        r = repr(obj)
+        return f'<span class="object">{escape(r)}</span>'
+
+    def dispatch_repr(self, obj: t.Any, recursive: bool) -> str:
+        if obj is helper:
+            return f'<span class="help">{helper!r}</span>'
+        if isinstance(obj, (int, float, complex)):
+            return f'<span class="number">{obj!r}</span>'
+        if isinstance(obj, str) or isinstance(obj, bytes):
+            return self.string_repr(obj)
+        if isinstance(obj, RegexType):
+            return self.regex_repr(obj)
+        if isinstance(obj, list):
+            return self.list_repr(obj, recursive)
+        if isinstance(obj, tuple):
+            return self.tuple_repr(obj, recursive)
+        if isinstance(obj, set):
+            return self.set_repr(obj, recursive)
+        if isinstance(obj, frozenset):
+            return self.frozenset_repr(obj, recursive)
+        if isinstance(obj, dict):
+            return self.dict_repr(obj, recursive)
+        if isinstance(obj, deque):
+            return self.deque_repr(obj, recursive)
+        return self.object_repr(obj)
+
+    def fallback_repr(self) -> str:
+        try:
+            info = "".join(format_exception_only(*sys.exc_info()[:2]))
+        except Exception:
+            info = "?"
+        return (
+            '<span class="brokenrepr">'
+            f"&lt;broken repr ({escape(info.strip())})&gt;</span>"
+        )
+
+    def repr(self, obj: object) -> str:
+        recursive = False
+        for item in self._stack:
+            if item is obj:
+                recursive = True
+                break
+        self._stack.append(obj)
+        try:
+            try:
+                return self.dispatch_repr(obj, recursive)
+            except Exception:
+                return self.fallback_repr()
+        finally:
+            self._stack.pop()
+
+    def dump_object(self, obj: object) -> str:
+        repr = None
+        items: list[tuple[str, str]] | None = None
+
+        if isinstance(obj, dict):
+            title = "Contents of"
+            items = []
+            for key, value in obj.items():
+                if not isinstance(key, str):
+                    items = None
+                    break
+                items.append((key, self.repr(value)))
+        if items is None:
+            items = []
+            repr = self.repr(obj)
+            for key in dir(obj):
+                try:
+                    items.append((key, self.repr(getattr(obj, key))))
+                except Exception:
+                    pass
+            title = "Details for"
+        title += f" {object.__repr__(obj)[1:-1]}"
+        return self.render_object_dump(items, title, repr)
+
+    def dump_locals(self, d: dict[str, t.Any]) -> str:
+        items = [(key, self.repr(value)) for key, value in d.items()]
+        return self.render_object_dump(items, "Local variables in frame")
+
+    def render_object_dump(
+        self, items: list[tuple[str, str]], title: str, repr: str | None = None
+    ) -> str:
+        html_items = []
+        for key, value in items:
+            html_items.append(f"<tr><th>{escape(key)}<td><pre class=repr>{value}</pre>")
+        if not html_items:
+            html_items.append("<tr><td><em>Nothing</em>")
+        return OBJECT_DUMP_HTML % {
+            "title": escape(title),
+            "repr": f"<pre class=repr>{repr if repr else ''}</pre>",
+            "items": "\n".join(html_items),
+        }
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/ICON_LICENSE.md b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/ICON_LICENSE.md
new file mode 100644
index 00000000..3bdbfc73
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/ICON_LICENSE.md
@@ -0,0 +1,6 @@
+Silk icon set 1.3 by Mark James <mjames@gmail.com>
+
+http://www.famfamfam.com/lab/icons/silk/
+
+License: [CC-BY-2.5](https://creativecommons.org/licenses/by/2.5/)
+or [CC-BY-3.0](https://creativecommons.org/licenses/by/3.0/)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/console.png b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/console.png
new file mode 100644
index 00000000..c28dd638
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/console.png differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/debugger.js b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/debugger.js
new file mode 100644
index 00000000..809b14a6
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/debugger.js
@@ -0,0 +1,344 @@
+docReady(() => {
+  if (!EVALEX_TRUSTED) {
+    initPinBox();
+  }
+  // if we are in console mode, show the console.
+  if (CONSOLE_MODE && EVALEX) {
+    createInteractiveConsole();
+  }
+
+  const frames = document.querySelectorAll("div.traceback div.frame");
+  if (EVALEX) {
+    addConsoleIconToFrames(frames);
+  }
+  addEventListenersToElements(document.querySelectorAll("div.detail"), "click", () =>
+    document.querySelector("div.traceback").scrollIntoView(false)
+  );
+  addToggleFrameTraceback(frames);
+  addToggleTraceTypesOnClick(document.querySelectorAll("h2.traceback"));
+  addInfoPrompt(document.querySelectorAll("span.nojavascript"));
+  wrapPlainTraceback();
+});
+
+function addToggleFrameTraceback(frames) {
+  frames.forEach((frame) => {
+    frame.addEventListener("click", () => {
+      frame.getElementsByTagName("pre")[0].parentElement.classList.toggle("expanded");
+    });
+  })
+}
+
+
+function wrapPlainTraceback() {
+  const plainTraceback = document.querySelector("div.plain textarea");
+  const wrapper = document.createElement("pre");
+  const textNode = document.createTextNode(plainTraceback.textContent);
+  wrapper.appendChild(textNode);
+  plainTraceback.replaceWith(wrapper);
+}
+
+function makeDebugURL(args) {
+  const params = new URLSearchParams(args)
+  params.set("s", SECRET)
+  return `?__debugger__=yes&${params}`
+}
+
+function initPinBox() {
+  document.querySelector(".pin-prompt form").addEventListener(
+    "submit",
+    function (event) {
+      event.preventDefault();
+      const btn = this.btn;
+      btn.disabled = true;
+
+      fetch(
+        makeDebugURL({cmd: "pinauth", pin: this.pin.value})
+      )
+        .then((res) => res.json())
+        .then(({auth, exhausted}) => {
+          if (auth) {
+            EVALEX_TRUSTED = true;
+            fadeOut(document.getElementsByClassName("pin-prompt")[0]);
+          } else {
+            alert(
+              `Error: ${
+                exhausted
+                  ? "too many attempts.  Restart server to retry."
+                  : "incorrect pin"
+              }`
+            );
+          }
+        })
+        .catch((err) => {
+          alert("Error: Could not verify PIN.  Network error?");
+          console.error(err);
+        })
+        .finally(() => (btn.disabled = false));
+    },
+    false
+  );
+}
+
+function promptForPin() {
+  if (!EVALEX_TRUSTED) {
+    fetch(makeDebugURL({cmd: "printpin"}));
+    const pinPrompt = document.getElementsByClassName("pin-prompt")[0];
+    fadeIn(pinPrompt);
+    document.querySelector('.pin-prompt input[name="pin"]').focus();
+  }
+}
+
+/**
+ * Helper function for shell initialization
+ */
+function openShell(consoleNode, target, frameID) {
+  promptForPin();
+  if (consoleNode) {
+    slideToggle(consoleNode);
+    return consoleNode;
+  }
+  let historyPos = 0;
+  const history = [""];
+  const consoleElement = createConsole();
+  const output = createConsoleOutput();
+  const form = createConsoleInputForm();
+  const command = createConsoleInput();
+
+  target.parentNode.appendChild(consoleElement);
+  consoleElement.append(output);
+  consoleElement.append(form);
+  form.append(command);
+  command.focus();
+  slideToggle(consoleElement);
+
+  form.addEventListener("submit", (e) => {
+    handleConsoleSubmit(e, command, frameID).then((consoleOutput) => {
+      output.append(consoleOutput);
+      command.focus();
+      consoleElement.scrollTo(0, consoleElement.scrollHeight);
+      const old = history.pop();
+      history.push(command.value);
+      if (typeof old !== "undefined") {
+        history.push(old);
+      }
+      historyPos = history.length - 1;
+      command.value = "";
+    });
+  });
+
+  command.addEventListener("keydown", (e) => {
+    if (e.key === "l" && e.ctrlKey) {
+      output.innerText = "--- screen cleared ---";
+    } else if (e.key === "ArrowUp" || e.key === "ArrowDown") {
+      // Handle up arrow and down arrow.
+      if (e.key === "ArrowUp" && historyPos > 0) {
+        e.preventDefault();
+        historyPos--;
+      } else if (e.key === "ArrowDown" && historyPos < history.length - 1) {
+        historyPos++;
+      }
+      command.value = history[historyPos];
+    }
+    return false;
+  });
+
+  return consoleElement;
+}
+
+function addEventListenersToElements(elements, event, listener) {
+  elements.forEach((el) => el.addEventListener(event, listener));
+}
+
+/**
+ * Add extra info
+ */
+function addInfoPrompt(elements) {
+  for (let i = 0; i < elements.length; i++) {
+    elements[i].innerHTML =
+      "<p>To switch between the interactive traceback and the plaintext " +
+      'one, you can click on the "Traceback" headline. From the text ' +
+      "traceback you can also create a paste of it. " +
+      (!EVALEX
+        ? ""
+        : "For code execution mouse-over the frame you want to debug and " +
+          "click on the console icon on the right side." +
+          "<p>You can execute arbitrary Python code in the stack frames and " +
+          "there are some extra helpers available for introspection:" +
+          "<ul><li><code>dump()</code> shows all variables in the frame" +
+          "<li><code>dump(obj)</code> dumps all that's known about the object</ul>");
+    elements[i].classList.remove("nojavascript");
+  }
+}
+
+function addConsoleIconToFrames(frames) {
+  for (let i = 0; i < frames.length; i++) {
+    let consoleNode = null;
+    const target = frames[i];
+    const frameID = frames[i].id.substring(6);
+
+    for (let j = 0; j < target.getElementsByTagName("pre").length; j++) {
+      const img = createIconForConsole();
+      img.addEventListener("click", (e) => {
+        e.stopPropagation();
+        consoleNode = openShell(consoleNode, target, frameID);
+        return false;
+      });
+      target.getElementsByTagName("pre")[j].append(img);
+    }
+  }
+}
+
+function slideToggle(target) {
+  target.classList.toggle("active");
+}
+
+/**
+ * toggle traceback types on click.
+ */
+function addToggleTraceTypesOnClick(elements) {
+  for (let i = 0; i < elements.length; i++) {
+    elements[i].addEventListener("click", () => {
+      document.querySelector("div.traceback").classList.toggle("hidden");
+      document.querySelector("div.plain").classList.toggle("hidden");
+    });
+    elements[i].style.cursor = "pointer";
+    document.querySelector("div.plain").classList.toggle("hidden");
+  }
+}
+
+function createConsole() {
+  const consoleNode = document.createElement("pre");
+  consoleNode.classList.add("console");
+  consoleNode.classList.add("active");
+  return consoleNode;
+}
+
+function createConsoleOutput() {
+  const output = document.createElement("div");
+  output.classList.add("output");
+  output.innerHTML = "[console ready]";
+  return output;
+}
+
+function createConsoleInputForm() {
+  const form = document.createElement("form");
+  form.innerHTML = "&gt;&gt;&gt; ";
+  return form;
+}
+
+function createConsoleInput() {
+  const command = document.createElement("input");
+  command.type = "text";
+  command.setAttribute("autocomplete", "off");
+  command.setAttribute("spellcheck", false);
+  command.setAttribute("autocapitalize", "off");
+  command.setAttribute("autocorrect", "off");
+  return command;
+}
+
+function createIconForConsole() {
+  const img = document.createElement("img");
+  img.setAttribute("src", makeDebugURL({cmd: "resource", f: "console.png"}));
+  img.setAttribute("title", "Open an interactive python shell in this frame");
+  return img;
+}
+
+function createExpansionButtonForConsole() {
+  const expansionButton = document.createElement("a");
+  expansionButton.setAttribute("href", "#");
+  expansionButton.setAttribute("class", "toggle");
+  expansionButton.innerHTML = "&nbsp;&nbsp;";
+  return expansionButton;
+}
+
+function createInteractiveConsole() {
+  const target = document.querySelector("div.console div.inner");
+  while (target.firstChild) {
+    target.removeChild(target.firstChild);
+  }
+  openShell(null, target, 0);
+}
+
+function handleConsoleSubmit(e, command, frameID) {
+  // Prevent page from refreshing.
+  e.preventDefault();
+
+  return new Promise((resolve) => {
+    fetch(makeDebugURL({cmd: command.value, frm: frameID}))
+      .then((res) => {
+        return res.text();
+      })
+      .then((data) => {
+        const tmp = document.createElement("div");
+        tmp.innerHTML = data;
+        resolve(tmp);
+
+        // Handle expandable span for long list outputs.
+        // Example to test: list(range(13))
+        let wrapperAdded = false;
+        const wrapperSpan = document.createElement("span");
+        const expansionButton = createExpansionButtonForConsole();
+
+        tmp.querySelectorAll("span.extended").forEach((spanToWrap) => {
+          const parentDiv = spanToWrap.parentNode;
+          if (!wrapperAdded) {
+            parentDiv.insertBefore(wrapperSpan, spanToWrap);
+            wrapperAdded = true;
+          }
+          parentDiv.removeChild(spanToWrap);
+          wrapperSpan.append(spanToWrap);
+          spanToWrap.hidden = true;
+
+          expansionButton.addEventListener("click", (event) => {
+            event.preventDefault();
+            spanToWrap.hidden = !spanToWrap.hidden;
+            expansionButton.classList.toggle("open");
+            return false;
+          });
+        });
+
+        // Add expansion button at end of wrapper.
+        if (wrapperAdded) {
+          wrapperSpan.append(expansionButton);
+        }
+      })
+      .catch((err) => {
+        console.error(err);
+      });
+    return false;
+  });
+}
+
+function fadeOut(element) {
+  element.style.opacity = 1;
+
+  (function fade() {
+    element.style.opacity -= 0.1;
+    if (element.style.opacity < 0) {
+      element.style.display = "none";
+    } else {
+      requestAnimationFrame(fade);
+    }
+  })();
+}
+
+function fadeIn(element, display) {
+  element.style.opacity = 0;
+  element.style.display = display || "block";
+
+  (function fade() {
+    let val = parseFloat(element.style.opacity) + 0.1;
+    if (val <= 1) {
+      element.style.opacity = val;
+      requestAnimationFrame(fade);
+    }
+  })();
+}
+
+function docReady(fn) {
+  if (document.readyState === "complete" || document.readyState === "interactive") {
+    setTimeout(fn, 1);
+  } else {
+    document.addEventListener("DOMContentLoaded", fn);
+  }
+}
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/less.png b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/less.png
new file mode 100644
index 00000000..5efefd62
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/less.png differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/more.png b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/more.png
new file mode 100644
index 00000000..804fa226
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/more.png differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/style.css b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/style.css
new file mode 100644
index 00000000..e9397ca0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/shared/style.css
@@ -0,0 +1,150 @@
+body, input  { font-family: sans-serif; color: #000; text-align: center;
+               margin: 1em; padding: 0; font-size: 15px; }
+h1, h2, h3   { font-weight: normal; }
+
+input        { background-color: #fff; margin: 0; text-align: left;
+               outline: none !important; }
+input[type="submit"] { padding: 3px 6px; }
+a            { color: #11557C; }
+a:hover      { color: #177199; }
+pre, code,
+textarea     { font-family: monospace; font-size: 14px; }
+
+div.debugger { text-align: left; padding: 12px; margin: auto;
+               background-color: white; }
+h1           { font-size: 36px; margin: 0 0 0.3em 0; }
+div.detail { cursor: pointer; }
+div.detail p { margin: 0 0 8px 13px; font-size: 14px; white-space: pre-wrap;
+               font-family: monospace; }
+div.explanation { margin: 20px 13px; font-size: 15px; color: #555; }
+div.footer   { font-size: 13px; text-align: right; margin: 30px 0;
+               color: #86989B; }
+
+h2           { font-size: 16px; margin: 1.3em 0 0.0 0; padding: 9px;
+               background-color: #11557C; color: white; }
+h2 em, h3 em { font-style: normal; color: #A5D6D9; font-weight: normal; }
+
+div.traceback, div.plain { border: 1px solid #ddd; margin: 0 0 1em 0; padding: 10px; }
+div.plain p      { margin: 0; }
+div.plain textarea,
+div.plain pre { margin: 10px 0 0 0; padding: 4px;
+                background-color: #E8EFF0; border: 1px solid #D3E7E9; }
+div.plain textarea { width: 99%; height: 300px; }
+div.traceback h3 { font-size: 1em; margin: 0 0 0.8em 0; }
+div.traceback ul { list-style: none; margin: 0; padding: 0 0 0 1em; }
+div.traceback h4 { font-size: 13px; font-weight: normal; margin: 0.7em 0 0.1em 0; }
+div.traceback pre { margin: 0; padding: 5px 0 3px 15px;
+                    background-color: #E8EFF0; border: 1px solid #D3E7E9; }
+div.traceback .library .current { background: white; color: #555; }
+div.traceback .expanded .current { background: #E8EFF0; color: black; }
+div.traceback pre:hover { background-color: #DDECEE; color: black; cursor: pointer; }
+div.traceback div.source.expanded pre + pre { border-top: none; }
+
+div.traceback span.ws { display: none; }
+div.traceback pre.before, div.traceback pre.after { display: none; background: white; }
+div.traceback div.source.expanded pre.before,
+div.traceback div.source.expanded pre.after {
+    display: block;
+}
+
+div.traceback div.source.expanded span.ws {
+    display: inline;
+}
+
+div.traceback blockquote { margin: 1em 0 0 0; padding: 0; white-space: pre-line; }
+div.traceback img { float: right; padding: 2px; margin: -3px 2px 0 0; display: none; }
+div.traceback img:hover { background-color: #ddd; cursor: pointer;
+                          border-color: #BFDDE0; }
+div.traceback pre:hover img { display: block; }
+div.traceback cite.filename { font-style: normal; color: #3B666B; }
+
+pre.console { border: 1px solid #ccc; background: white!important;
+              color: black; padding: 5px!important;
+              margin: 3px 0 0 0!important; cursor: default!important;
+              max-height: 400px; overflow: auto; }
+pre.console form { color: #555; }
+pre.console input { background-color: transparent; color: #555;
+                    width: 90%; font-family: monospace; font-size: 14px;
+                     border: none!important; }
+
+span.string { color: #30799B; }
+span.number { color: #9C1A1C; }
+span.help   { color: #3A7734; }
+span.object { color: #485F6E; }
+span.extended { opacity: 0.5; }
+span.extended:hover { opacity: 1; }
+a.toggle { text-decoration: none; background-repeat: no-repeat;
+           background-position: center center;
+           background-image: url(?__debugger__=yes&cmd=resource&f=more.png); }
+a.toggle:hover { background-color: #444; }
+a.open { background-image: url(?__debugger__=yes&cmd=resource&f=less.png); }
+
+pre.console div.traceback,
+pre.console div.box { margin: 5px 10px; white-space: normal;
+                      border: 1px solid #11557C; padding: 10px;
+                      font-family: sans-serif;  }
+pre.console div.box h3,
+pre.console div.traceback h3 { margin: -10px -10px 10px -10px; padding: 5px;
+                               background: #11557C; color: white; }
+
+pre.console div.traceback pre:hover { cursor: default; background: #E8EFF0; }
+pre.console div.traceback pre.syntaxerror { background: inherit; border: none;
+                                            margin: 20px -10px -10px -10px;
+                                            padding: 10px; border-top: 1px solid #BFDDE0;
+                                            background: #E8EFF0; }
+pre.console div.noframe-traceback pre.syntaxerror { margin-top: -10px; border: none; }
+
+pre.console div.box pre.repr { padding: 0; margin: 0; background-color: white; border: none; }
+pre.console div.box table { margin-top: 6px; }
+pre.console div.box pre { border: none; }
+pre.console div.box pre.help { background-color: white; }
+pre.console div.box pre.help:hover { cursor: default; }
+pre.console table tr { vertical-align: top; }
+div.console { border: 1px solid #ccc; padding: 4px; background-color: #fafafa; }
+
+div.traceback pre, div.console pre {
+    white-space: pre-wrap;       /* css-3 should we be so lucky... */
+    white-space: -moz-pre-wrap;  /* Mozilla, since 1999 */
+    white-space: -pre-wrap;      /* Opera 4-6 ?? */
+    white-space: -o-pre-wrap;    /* Opera 7 ?? */
+    word-wrap: break-word;       /* Internet Explorer 5.5+ */
+    _white-space: pre;           /* IE only hack to re-specify in
+                                 addition to word-wrap  */
+}
+
+
+div.pin-prompt {
+    position: absolute;
+    display: none;
+    top: 0;
+    bottom: 0;
+    left: 0;
+    right: 0;
+    background: rgba(255, 255, 255, 0.8);
+}
+
+div.pin-prompt .inner {
+    background: #eee;
+    padding: 10px 50px;
+    width: 350px;
+    margin: 10% auto 0 auto;
+    border: 1px solid #ccc;
+    border-radius: 2px;
+}
+
+div.exc-divider {
+    margin: 0.7em 0 0 -1em;
+    padding: 0.5em;
+    background: #11557C;
+    color: #ddd;
+    border: 1px solid #ddd;
+}
+
+.console.active {
+    max-height: 0!important;
+    display: none;
+}
+
+.hidden {
+    display: none;
+}
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/debug/tbtools.py b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/tbtools.py
new file mode 100644
index 00000000..d922893e
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/debug/tbtools.py
@@ -0,0 +1,450 @@
+from __future__ import annotations
+
+import itertools
+import linecache
+import os
+import re
+import sys
+import sysconfig
+import traceback
+import typing as t
+
+from markupsafe import escape
+
+from ..utils import cached_property
+from .console import Console
+
+HEADER = """\
+<!doctype html>
+<html lang=en>
+  <head>
+    <title>%(title)s // Werkzeug Debugger</title>
+    <link rel="stylesheet" href="?__debugger__=yes&amp;cmd=resource&amp;f=style.css">
+    <link rel="shortcut icon"
+        href="?__debugger__=yes&amp;cmd=resource&amp;f=console.png">
+    <script src="?__debugger__=yes&amp;cmd=resource&amp;f=debugger.js"></script>
+    <script>
+      var CONSOLE_MODE = %(console)s,
+          EVALEX = %(evalex)s,
+          EVALEX_TRUSTED = %(evalex_trusted)s,
+          SECRET = "%(secret)s";
+    </script>
+  </head>
+  <body style="background-color: #fff">
+    <div class="debugger">
+"""
+
+FOOTER = """\
+      <div class="footer">
+        Brought to you by <strong class="arthur">DON'T PANIC</strong>, your
+        friendly Werkzeug powered traceback interpreter.
+      </div>
+    </div>
+
+    <div class="pin-prompt">
+      <div class="inner">
+        <h3>Console Locked</h3>
+        <p>
+          The console is locked and needs to be unlocked by entering the PIN.
+          You can find the PIN printed out on the standard output of your
+          shell that runs the server.
+        <form>
+          <p>PIN:
+            <input type=text name=pin size=14>
+            <input type=submit name=btn value="Confirm Pin">
+        </form>
+      </div>
+    </div>
+  </body>
+</html>
+"""
+
+PAGE_HTML = (
+    HEADER
+    + """\
+<h1>%(exception_type)s</h1>
+<div class="detail">
+  <p class="errormsg">%(exception)s</p>
+</div>
+<h2 class="traceback">Traceback <em>(most recent call last)</em></h2>
+%(summary)s
+<div class="plain">
+    <p>
+      This is the Copy/Paste friendly version of the traceback.
+    </p>
+    <textarea cols="50" rows="10" name="code" readonly>%(plaintext)s</textarea>
+</div>
+<div class="explanation">
+  The debugger caught an exception in your WSGI application.  You can now
+  look at the traceback which led to the error.  <span class="nojavascript">
+  If you enable JavaScript you can also use additional features such as code
+  execution (if the evalex feature is enabled), automatic pasting of the
+  exceptions and much more.</span>
+</div>
+"""
+    + FOOTER
+    + """
+<!--
+
+%(plaintext_cs)s
+
+-->
+"""
+)
+
+CONSOLE_HTML = (
+    HEADER
+    + """\
+<h1>Interactive Console</h1>
+<div class="explanation">
+In this console you can execute Python expressions in the context of the
+application.  The initial namespace was created by the debugger automatically.
+</div>
+<div class="console"><div class="inner">The Console requires JavaScript.</div></div>
+"""
+    + FOOTER
+)
+
+SUMMARY_HTML = """\
+<div class="%(classes)s">
+  %(title)s
+  <ul>%(frames)s</ul>
+  %(description)s
+</div>
+"""
+
+FRAME_HTML = """\
+<div class="frame" id="frame-%(id)d">
+  <h4>File <cite class="filename">"%(filename)s"</cite>,
+      line <em class="line">%(lineno)s</em>,
+      in <code class="function">%(function_name)s</code></h4>
+  <div class="source %(library)s">%(lines)s</div>
+</div>
+"""
+
+
+def _process_traceback(
+    exc: BaseException,
+    te: traceback.TracebackException | None = None,
+    *,
+    skip: int = 0,
+    hide: bool = True,
+) -> traceback.TracebackException:
+    if te is None:
+        te = traceback.TracebackException.from_exception(exc, lookup_lines=False)
+
+    # Get the frames the same way StackSummary.extract did, in order
+    # to match each frame with the FrameSummary to augment.
+    frame_gen = traceback.walk_tb(exc.__traceback__)
+    limit = getattr(sys, "tracebacklimit", None)
+
+    if limit is not None:
+        if limit < 0:
+            limit = 0
+
+        frame_gen = itertools.islice(frame_gen, limit)
+
+    if skip:
+        frame_gen = itertools.islice(frame_gen, skip, None)
+        del te.stack[:skip]
+
+    new_stack: list[DebugFrameSummary] = []
+    hidden = False
+
+    # Match each frame with the FrameSummary that was generated.
+    # Hide frames using Paste's __traceback_hide__ rules. Replace
+    # all visible FrameSummary with DebugFrameSummary.
+    for (f, _), fs in zip(frame_gen, te.stack):
+        if hide:
+            hide_value = f.f_locals.get("__traceback_hide__", False)
+
+            if hide_value in {"before", "before_and_this"}:
+                new_stack = []
+                hidden = False
+
+                if hide_value == "before_and_this":
+                    continue
+            elif hide_value in {"reset", "reset_and_this"}:
+                hidden = False
+
+                if hide_value == "reset_and_this":
+                    continue
+            elif hide_value in {"after", "after_and_this"}:
+                hidden = True
+
+                if hide_value == "after_and_this":
+                    continue
+            elif hide_value or hidden:
+                continue
+
+        frame_args: dict[str, t.Any] = {
+            "filename": fs.filename,
+            "lineno": fs.lineno,
+            "name": fs.name,
+            "locals": f.f_locals,
+            "globals": f.f_globals,
+        }
+
+        if sys.version_info >= (3, 11):
+            frame_args["colno"] = fs.colno
+            frame_args["end_colno"] = fs.end_colno
+
+        new_stack.append(DebugFrameSummary(**frame_args))
+
+    # The codeop module is used to compile code from the interactive
+    # debugger. Hide any codeop frames from the bottom of the traceback.
+    while new_stack:
+        module = new_stack[0].global_ns.get("__name__")
+
+        if module is None:
+            module = new_stack[0].local_ns.get("__name__")
+
+        if module == "codeop":
+            del new_stack[0]
+        else:
+            break
+
+    te.stack[:] = new_stack
+
+    if te.__context__:
+        context_exc = t.cast(BaseException, exc.__context__)
+        te.__context__ = _process_traceback(context_exc, te.__context__, hide=hide)
+
+    if te.__cause__:
+        cause_exc = t.cast(BaseException, exc.__cause__)
+        te.__cause__ = _process_traceback(cause_exc, te.__cause__, hide=hide)
+
+    return te
+
+
+class DebugTraceback:
+    __slots__ = ("_te", "_cache_all_tracebacks", "_cache_all_frames")
+
+    def __init__(
+        self,
+        exc: BaseException,
+        te: traceback.TracebackException | None = None,
+        *,
+        skip: int = 0,
+        hide: bool = True,
+    ) -> None:
+        self._te = _process_traceback(exc, te, skip=skip, hide=hide)
+
+    def __str__(self) -> str:
+        return f"<{type(self).__name__} {self._te}>"
+
+    @cached_property
+    def all_tracebacks(
+        self,
+    ) -> list[tuple[str | None, traceback.TracebackException]]:
+        out = []
+        current = self._te
+
+        while current is not None:
+            if current.__cause__ is not None:
+                chained_msg = (
+                    "The above exception was the direct cause of the"
+                    " following exception"
+                )
+                chained_exc = current.__cause__
+            elif current.__context__ is not None and not current.__suppress_context__:
+                chained_msg = (
+                    "During handling of the above exception, another"
+                    " exception occurred"
+                )
+                chained_exc = current.__context__
+            else:
+                chained_msg = None
+                chained_exc = None
+
+            out.append((chained_msg, current))
+            current = chained_exc
+
+        return out
+
+    @cached_property
+    def all_frames(self) -> list[DebugFrameSummary]:
+        return [
+            f  # type: ignore[misc]
+            for _, te in self.all_tracebacks
+            for f in te.stack
+        ]
+
+    def render_traceback_text(self) -> str:
+        return "".join(self._te.format())
+
+    def render_traceback_html(self, include_title: bool = True) -> str:
+        library_frames = [f.is_library for f in self.all_frames]
+        mark_library = 0 < sum(library_frames) < len(library_frames)
+        rows = []
+
+        if not library_frames:
+            classes = "traceback noframe-traceback"
+        else:
+            classes = "traceback"
+
+            for msg, current in reversed(self.all_tracebacks):
+                row_parts = []
+
+                if msg is not None:
+                    row_parts.append(f'<li><div class="exc-divider">{msg}:</div>')
+
+                for frame in current.stack:
+                    frame = t.cast(DebugFrameSummary, frame)
+                    info = f' title="{escape(frame.info)}"' if frame.info else ""
+                    row_parts.append(f"<li{info}>{frame.render_html(mark_library)}")
+
+                rows.append("\n".join(row_parts))
+
+        if sys.version_info < (3, 13):
+            exc_type_str = self._te.exc_type.__name__
+        else:
+            exc_type_str = self._te.exc_type_str
+
+        is_syntax_error = exc_type_str == "SyntaxError"
+
+        if include_title:
+            if is_syntax_error:
+                title = "Syntax Error"
+            else:
+                title = "Traceback <em>(most recent call last)</em>:"
+        else:
+            title = ""
+
+        exc_full = escape("".join(self._te.format_exception_only()))
+
+        if is_syntax_error:
+            description = f"<pre class=syntaxerror>{exc_full}</pre>"
+        else:
+            description = f"<blockquote>{exc_full}</blockquote>"
+
+        return SUMMARY_HTML % {
+            "classes": classes,
+            "title": f"<h3>{title}</h3>",
+            "frames": "\n".join(rows),
+            "description": description,
+        }
+
+    def render_debugger_html(
+        self, evalex: bool, secret: str, evalex_trusted: bool
+    ) -> str:
+        exc_lines = list(self._te.format_exception_only())
+        plaintext = "".join(self._te.format())
+
+        if sys.version_info < (3, 13):
+            exc_type_str = self._te.exc_type.__name__
+        else:
+            exc_type_str = self._te.exc_type_str
+
+        return PAGE_HTML % {
+            "evalex": "true" if evalex else "false",
+            "evalex_trusted": "true" if evalex_trusted else "false",
+            "console": "false",
+            "title": escape(exc_lines[0]),
+            "exception": escape("".join(exc_lines)),
+            "exception_type": escape(exc_type_str),
+            "summary": self.render_traceback_html(include_title=False),
+            "plaintext": escape(plaintext),
+            "plaintext_cs": re.sub("-{2,}", "-", plaintext),
+            "secret": secret,
+        }
+
+
+class DebugFrameSummary(traceback.FrameSummary):
+    """A :class:`traceback.FrameSummary` that can evaluate code in the
+    frame's namespace.
+    """
+
+    __slots__ = (
+        "local_ns",
+        "global_ns",
+        "_cache_info",
+        "_cache_is_library",
+        "_cache_console",
+    )
+
+    def __init__(
+        self,
+        *,
+        locals: dict[str, t.Any],
+        globals: dict[str, t.Any],
+        **kwargs: t.Any,
+    ) -> None:
+        super().__init__(locals=None, **kwargs)
+        self.local_ns = locals
+        self.global_ns = globals
+
+    @cached_property
+    def info(self) -> str | None:
+        return self.local_ns.get("__traceback_info__")
+
+    @cached_property
+    def is_library(self) -> bool:
+        return any(
+            self.filename.startswith((path, os.path.realpath(path)))
+            for path in sysconfig.get_paths().values()
+        )
+
+    @cached_property
+    def console(self) -> Console:
+        return Console(self.global_ns, self.local_ns)
+
+    def eval(self, code: str) -> t.Any:
+        return self.console.eval(code)
+
+    def render_html(self, mark_library: bool) -> str:
+        context = 5
+        lines = linecache.getlines(self.filename)
+        line_idx = self.lineno - 1  # type: ignore[operator]
+        start_idx = max(0, line_idx - context)
+        stop_idx = min(len(lines), line_idx + context + 1)
+        rendered_lines = []
+
+        def render_line(line: str, cls: str) -> None:
+            line = line.expandtabs().rstrip()
+            stripped_line = line.strip()
+            prefix = len(line) - len(stripped_line)
+            colno = getattr(self, "colno", 0)
+            end_colno = getattr(self, "end_colno", 0)
+
+            if cls == "current" and colno and end_colno:
+                arrow = (
+                    f'\n<span class="ws">{" " * prefix}</span>'
+                    f'{" " * (colno - prefix)}{"^" * (end_colno - colno)}'
+                )
+            else:
+                arrow = ""
+
+            rendered_lines.append(
+                f'<pre class="line {cls}"><span class="ws">{" " * prefix}</span>'
+                f"{escape(stripped_line) if stripped_line else ' '}"
+                f"{arrow if arrow else ''}</pre>"
+            )
+
+        if lines:
+            for line in lines[start_idx:line_idx]:
+                render_line(line, "before")
+
+            render_line(lines[line_idx], "current")
+
+            for line in lines[line_idx + 1 : stop_idx]:
+                render_line(line, "after")
+
+        return FRAME_HTML % {
+            "id": id(self),
+            "filename": escape(self.filename),
+            "lineno": self.lineno,
+            "function_name": escape(self.name),
+            "lines": "\n".join(rendered_lines),
+            "library": "library" if mark_library and self.is_library else "",
+        }
+
+
+def render_console_html(secret: str, evalex_trusted: bool) -> str:
+    return CONSOLE_HTML % {
+        "evalex": "true",
+        "evalex_trusted": "true" if evalex_trusted else "false",
+        "console": "true",
+        "title": "Console",
+        "secret": secret,
+    }
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/exceptions.py b/backend/venv/lib/python3.12/site-packages/werkzeug/exceptions.py
new file mode 100644
index 00000000..1cd99977
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/exceptions.py
@@ -0,0 +1,894 @@
+"""Implements a number of Python exceptions which can be raised from within
+a view to trigger a standard HTTP non-200 response.
+
+Usage Example
+-------------
+
+.. code-block:: python
+
+    from werkzeug.wrappers.request import Request
+    from werkzeug.exceptions import HTTPException, NotFound
+
+    def view(request):
+        raise NotFound()
+
+    @Request.application
+    def application(request):
+        try:
+            return view(request)
+        except HTTPException as e:
+            return e
+
+As you can see from this example those exceptions are callable WSGI
+applications. However, they are not Werkzeug response objects. You
+can get a response object by calling ``get_response()`` on a HTTP
+exception.
+
+Keep in mind that you may have to pass an environ (WSGI) or scope
+(ASGI) to ``get_response()`` because some errors fetch additional
+information relating to the request.
+
+If you want to hook in a different exception page to say, a 404 status
+code, you can add a second except for a specific subclass of an error:
+
+.. code-block:: python
+
+    @Request.application
+    def application(request):
+        try:
+            return view(request)
+        except NotFound as e:
+            return not_found(request)
+        except HTTPException as e:
+            return e
+
+"""
+
+from __future__ import annotations
+
+import typing as t
+from datetime import datetime
+
+from markupsafe import escape
+from markupsafe import Markup
+
+from ._internal import _get_environ
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .datastructures import WWWAuthenticate
+    from .sansio.response import Response
+    from .wrappers.request import Request as WSGIRequest
+    from .wrappers.response import Response as WSGIResponse
+
+
+class HTTPException(Exception):
+    """The base class for all HTTP exceptions. This exception can be called as a WSGI
+    application to render a default error page or you can catch the subclasses
+    of it independently and render nicer error messages.
+
+    .. versionchanged:: 2.1
+        Removed the ``wrap`` class method.
+    """
+
+    code: int | None = None
+    description: str | None = None
+
+    def __init__(
+        self,
+        description: str | None = None,
+        response: Response | None = None,
+    ) -> None:
+        super().__init__()
+        if description is not None:
+            self.description = description
+        self.response = response
+
+    @property
+    def name(self) -> str:
+        """The status name."""
+        from .http import HTTP_STATUS_CODES
+
+        return HTTP_STATUS_CODES.get(self.code, "Unknown Error")  # type: ignore
+
+    def get_description(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> str:
+        """Get the description."""
+        if self.description is None:
+            description = ""
+        else:
+            description = self.description
+
+        description = escape(description).replace("\n", Markup("<br>"))
+        return f"<p>{description}</p>"
+
+    def get_body(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> str:
+        """Get the HTML body."""
+        return (
+            "<!doctype html>\n"
+            "<html lang=en>\n"
+            f"<title>{self.code} {escape(self.name)}</title>\n"
+            f"<h1>{escape(self.name)}</h1>\n"
+            f"{self.get_description(environ)}\n"
+        )
+
+    def get_headers(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> list[tuple[str, str]]:
+        """Get a list of headers."""
+        return [("Content-Type", "text/html; charset=utf-8")]
+
+    def get_response(
+        self,
+        environ: WSGIEnvironment | WSGIRequest | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> Response:
+        """Get a response object.  If one was passed to the exception
+        it's returned directly.
+
+        :param environ: the optional environ for the request.  This
+                        can be used to modify the response depending
+                        on how the request looked like.
+        :return: a :class:`Response` object or a subclass thereof.
+        """
+        from .wrappers.response import Response as WSGIResponse  # noqa: F811
+
+        if self.response is not None:
+            return self.response
+        if environ is not None:
+            environ = _get_environ(environ)
+        headers = self.get_headers(environ, scope)
+        return WSGIResponse(self.get_body(environ, scope), self.code, headers)
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        """Call the exception as WSGI application.
+
+        :param environ: the WSGI environment.
+        :param start_response: the response callable provided by the WSGI
+                               server.
+        """
+        response = t.cast("WSGIResponse", self.get_response(environ))
+        return response(environ, start_response)
+
+    def __str__(self) -> str:
+        code = self.code if self.code is not None else "???"
+        return f"{code} {self.name}: {self.description}"
+
+    def __repr__(self) -> str:
+        code = self.code if self.code is not None else "???"
+        return f"<{type(self).__name__} '{code}: {self.name}'>"
+
+
+class BadRequest(HTTPException):
+    """*400* `Bad Request`
+
+    Raise if the browser sends something to the application the application
+    or server cannot handle.
+    """
+
+    code = 400
+    description = (
+        "The browser (or proxy) sent a request that this server could "
+        "not understand."
+    )
+
+
+class BadRequestKeyError(BadRequest, KeyError):
+    """An exception that is used to signal both a :exc:`KeyError` and a
+    :exc:`BadRequest`. Used by many of the datastructures.
+    """
+
+    _description = BadRequest.description
+    #: Show the KeyError along with the HTTP error message in the
+    #: response. This should be disabled in production, but can be
+    #: useful in a debug mode.
+    show_exception = False
+
+    def __init__(self, arg: object | None = None, *args: t.Any, **kwargs: t.Any):
+        super().__init__(*args, **kwargs)
+
+        if arg is None:
+            KeyError.__init__(self)
+        else:
+            KeyError.__init__(self, arg)
+
+    @property  # type: ignore
+    def description(self) -> str:
+        if self.show_exception:
+            return (
+                f"{self._description}\n"
+                f"{KeyError.__name__}: {KeyError.__str__(self)}"
+            )
+
+        return self._description
+
+    @description.setter
+    def description(self, value: str) -> None:
+        self._description = value
+
+
+class ClientDisconnected(BadRequest):
+    """Internal exception that is raised if Werkzeug detects a disconnected
+    client.  Since the client is already gone at that point attempting to
+    send the error message to the client might not work and might ultimately
+    result in another exception in the server.  Mainly this is here so that
+    it is silenced by default as far as Werkzeug is concerned.
+
+    Since disconnections cannot be reliably detected and are unspecified
+    by WSGI to a large extent this might or might not be raised if a client
+    is gone.
+
+    .. versionadded:: 0.8
+    """
+
+
+class SecurityError(BadRequest):
+    """Raised if something triggers a security error.  This is otherwise
+    exactly like a bad request error.
+
+    .. versionadded:: 0.9
+    """
+
+
+class BadHost(BadRequest):
+    """Raised if the submitted host is badly formatted.
+
+    .. versionadded:: 0.11.2
+    """
+
+
+class Unauthorized(HTTPException):
+    """*401* ``Unauthorized``
+
+    Raise if the user is not authorized to access a resource.
+
+    The ``www_authenticate`` argument should be used to set the
+    ``WWW-Authenticate`` header. This is used for HTTP basic auth and
+    other schemes. Use :class:`~werkzeug.datastructures.WWWAuthenticate`
+    to create correctly formatted values. Strictly speaking a 401
+    response is invalid if it doesn't provide at least one value for
+    this header, although real clients typically don't care.
+
+    :param description: Override the default message used for the body
+        of the response.
+    :param www-authenticate: A single value, or list of values, for the
+        WWW-Authenticate header(s).
+
+    .. versionchanged:: 2.0
+        Serialize multiple ``www_authenticate`` items into multiple
+        ``WWW-Authenticate`` headers, rather than joining them
+        into a single value, for better interoperability.
+
+    .. versionchanged:: 0.15.3
+        If the ``www_authenticate`` argument is not set, the
+        ``WWW-Authenticate`` header is not set.
+
+    .. versionchanged:: 0.15.3
+        The ``response`` argument was restored.
+
+    .. versionchanged:: 0.15.1
+        ``description`` was moved back as the first argument, restoring
+         its previous position.
+
+    .. versionchanged:: 0.15.0
+        ``www_authenticate`` was added as the first argument, ahead of
+        ``description``.
+    """
+
+    code = 401
+    description = (
+        "The server could not verify that you are authorized to access"
+        " the URL requested. You either supplied the wrong credentials"
+        " (e.g. a bad password), or your browser doesn't understand"
+        " how to supply the credentials required."
+    )
+
+    def __init__(
+        self,
+        description: str | None = None,
+        response: Response | None = None,
+        www_authenticate: None | (WWWAuthenticate | t.Iterable[WWWAuthenticate]) = None,
+    ) -> None:
+        super().__init__(description, response)
+
+        from .datastructures import WWWAuthenticate
+
+        if isinstance(www_authenticate, WWWAuthenticate):
+            www_authenticate = (www_authenticate,)
+
+        self.www_authenticate = www_authenticate
+
+    def get_headers(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> list[tuple[str, str]]:
+        headers = super().get_headers(environ, scope)
+        if self.www_authenticate:
+            headers.extend(("WWW-Authenticate", str(x)) for x in self.www_authenticate)
+        return headers
+
+
+class Forbidden(HTTPException):
+    """*403* `Forbidden`
+
+    Raise if the user doesn't have the permission for the requested resource
+    but was authenticated.
+    """
+
+    code = 403
+    description = (
+        "You don't have the permission to access the requested"
+        " resource. It is either read-protected or not readable by the"
+        " server."
+    )
+
+
+class NotFound(HTTPException):
+    """*404* `Not Found`
+
+    Raise if a resource does not exist and never existed.
+    """
+
+    code = 404
+    description = (
+        "The requested URL was not found on the server. If you entered"
+        " the URL manually please check your spelling and try again."
+    )
+
+
+class MethodNotAllowed(HTTPException):
+    """*405* `Method Not Allowed`
+
+    Raise if the server used a method the resource does not handle.  For
+    example `POST` if the resource is view only.  Especially useful for REST.
+
+    The first argument for this exception should be a list of allowed methods.
+    Strictly speaking the response would be invalid if you don't provide valid
+    methods in the header which you can do with that list.
+    """
+
+    code = 405
+    description = "The method is not allowed for the requested URL."
+
+    def __init__(
+        self,
+        valid_methods: t.Iterable[str] | None = None,
+        description: str | None = None,
+        response: Response | None = None,
+    ) -> None:
+        """Takes an optional list of valid http methods
+        starting with werkzeug 0.3 the list will be mandatory."""
+        super().__init__(description=description, response=response)
+        self.valid_methods = valid_methods
+
+    def get_headers(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> list[tuple[str, str]]:
+        headers = super().get_headers(environ, scope)
+        if self.valid_methods:
+            headers.append(("Allow", ", ".join(self.valid_methods)))
+        return headers
+
+
+class NotAcceptable(HTTPException):
+    """*406* `Not Acceptable`
+
+    Raise if the server can't return any content conforming to the
+    `Accept` headers of the client.
+    """
+
+    code = 406
+    description = (
+        "The resource identified by the request is only capable of"
+        " generating response entities which have content"
+        " characteristics not acceptable according to the accept"
+        " headers sent in the request."
+    )
+
+
+class RequestTimeout(HTTPException):
+    """*408* `Request Timeout`
+
+    Raise to signalize a timeout.
+    """
+
+    code = 408
+    description = (
+        "The server closed the network connection because the browser"
+        " didn't finish the request within the specified time."
+    )
+
+
+class Conflict(HTTPException):
+    """*409* `Conflict`
+
+    Raise to signal that a request cannot be completed because it conflicts
+    with the current state on the server.
+
+    .. versionadded:: 0.7
+    """
+
+    code = 409
+    description = (
+        "A conflict happened while processing the request. The"
+        " resource might have been modified while the request was being"
+        " processed."
+    )
+
+
+class Gone(HTTPException):
+    """*410* `Gone`
+
+    Raise if a resource existed previously and went away without new location.
+    """
+
+    code = 410
+    description = (
+        "The requested URL is no longer available on this server and"
+        " there is no forwarding address. If you followed a link from a"
+        " foreign page, please contact the author of this page."
+    )
+
+
+class LengthRequired(HTTPException):
+    """*411* `Length Required`
+
+    Raise if the browser submitted data but no ``Content-Length`` header which
+    is required for the kind of processing the server does.
+    """
+
+    code = 411
+    description = (
+        "A request with this method requires a valid <code>Content-"
+        "Length</code> header."
+    )
+
+
+class PreconditionFailed(HTTPException):
+    """*412* `Precondition Failed`
+
+    Status code used in combination with ``If-Match``, ``If-None-Match``, or
+    ``If-Unmodified-Since``.
+    """
+
+    code = 412
+    description = (
+        "The precondition on the request for the URL failed positive evaluation."
+    )
+
+
+class RequestEntityTooLarge(HTTPException):
+    """*413* `Request Entity Too Large`
+
+    The status code one should return if the data submitted exceeded a given
+    limit.
+    """
+
+    code = 413
+    description = "The data value transmitted exceeds the capacity limit."
+
+
+class RequestURITooLarge(HTTPException):
+    """*414* `Request URI Too Large`
+
+    Like *413* but for too long URLs.
+    """
+
+    code = 414
+    description = (
+        "The length of the requested URL exceeds the capacity limit for"
+        " this server. The request cannot be processed."
+    )
+
+
+class UnsupportedMediaType(HTTPException):
+    """*415* `Unsupported Media Type`
+
+    The status code returned if the server is unable to handle the media type
+    the client transmitted.
+    """
+
+    code = 415
+    description = (
+        "The server does not support the media type transmitted in the request."
+    )
+
+
+class RequestedRangeNotSatisfiable(HTTPException):
+    """*416* `Requested Range Not Satisfiable`
+
+    The client asked for an invalid part of the file.
+
+    .. versionadded:: 0.7
+    """
+
+    code = 416
+    description = "The server cannot provide the requested range."
+
+    def __init__(
+        self,
+        length: int | None = None,
+        units: str = "bytes",
+        description: str | None = None,
+        response: Response | None = None,
+    ) -> None:
+        """Takes an optional `Content-Range` header value based on ``length``
+        parameter.
+        """
+        super().__init__(description=description, response=response)
+        self.length = length
+        self.units = units
+
+    def get_headers(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> list[tuple[str, str]]:
+        headers = super().get_headers(environ, scope)
+        if self.length is not None:
+            headers.append(("Content-Range", f"{self.units} */{self.length}"))
+        return headers
+
+
+class ExpectationFailed(HTTPException):
+    """*417* `Expectation Failed`
+
+    The server cannot meet the requirements of the Expect request-header.
+
+    .. versionadded:: 0.7
+    """
+
+    code = 417
+    description = "The server could not meet the requirements of the Expect header"
+
+
+class ImATeapot(HTTPException):
+    """*418* `I'm a teapot`
+
+    The server should return this if it is a teapot and someone attempted
+    to brew coffee with it.
+
+    .. versionadded:: 0.7
+    """
+
+    code = 418
+    description = "This server is a teapot, not a coffee machine"
+
+
+class MisdirectedRequest(HTTPException):
+    """421 Misdirected Request
+
+    Indicates that the request was directed to a server that is not able to
+    produce a response.
+
+    .. versionadded:: 3.1
+    """
+
+    code = 421
+    description = "The server is not able to produce a response."
+
+
+class UnprocessableEntity(HTTPException):
+    """*422* `Unprocessable Entity`
+
+    Used if the request is well formed, but the instructions are otherwise
+    incorrect.
+    """
+
+    code = 422
+    description = (
+        "The request was well-formed but was unable to be followed due"
+        " to semantic errors."
+    )
+
+
+class Locked(HTTPException):
+    """*423* `Locked`
+
+    Used if the resource that is being accessed is locked.
+    """
+
+    code = 423
+    description = "The resource that is being accessed is locked."
+
+
+class FailedDependency(HTTPException):
+    """*424* `Failed Dependency`
+
+    Used if the method could not be performed on the resource
+    because the requested action depended on another action and that action failed.
+    """
+
+    code = 424
+    description = (
+        "The method could not be performed on the resource because the"
+        " requested action depended on another action and that action"
+        " failed."
+    )
+
+
+class PreconditionRequired(HTTPException):
+    """*428* `Precondition Required`
+
+    The server requires this request to be conditional, typically to prevent
+    the lost update problem, which is a race condition between two or more
+    clients attempting to update a resource through PUT or DELETE. By requiring
+    each client to include a conditional header ("If-Match" or "If-Unmodified-
+    Since") with the proper value retained from a recent GET request, the
+    server ensures that each client has at least seen the previous revision of
+    the resource.
+    """
+
+    code = 428
+    description = (
+        "This request is required to be conditional; try using"
+        ' "If-Match" or "If-Unmodified-Since".'
+    )
+
+
+class _RetryAfter(HTTPException):
+    """Adds an optional ``retry_after`` parameter which will set the
+    ``Retry-After`` header. May be an :class:`int` number of seconds or
+    a :class:`~datetime.datetime`.
+    """
+
+    def __init__(
+        self,
+        description: str | None = None,
+        response: Response | None = None,
+        retry_after: datetime | int | None = None,
+    ) -> None:
+        super().__init__(description, response)
+        self.retry_after = retry_after
+
+    def get_headers(
+        self,
+        environ: WSGIEnvironment | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> list[tuple[str, str]]:
+        headers = super().get_headers(environ, scope)
+
+        if self.retry_after:
+            if isinstance(self.retry_after, datetime):
+                from .http import http_date
+
+                value = http_date(self.retry_after)
+            else:
+                value = str(self.retry_after)
+
+            headers.append(("Retry-After", value))
+
+        return headers
+
+
+class TooManyRequests(_RetryAfter):
+    """*429* `Too Many Requests`
+
+    The server is limiting the rate at which this user receives
+    responses, and this request exceeds that rate. (The server may use
+    any convenient method to identify users and their request rates).
+    The server may include a "Retry-After" header to indicate how long
+    the user should wait before retrying.
+
+    :param retry_after: If given, set the ``Retry-After`` header to this
+        value. May be an :class:`int` number of seconds or a
+        :class:`~datetime.datetime`.
+
+    .. versionchanged:: 1.0
+        Added ``retry_after`` parameter.
+    """
+
+    code = 429
+    description = "This user has exceeded an allotted request count. Try again later."
+
+
+class RequestHeaderFieldsTooLarge(HTTPException):
+    """*431* `Request Header Fields Too Large`
+
+    The server refuses to process the request because the header fields are too
+    large. One or more individual fields may be too large, or the set of all
+    headers is too large.
+    """
+
+    code = 431
+    description = "One or more header fields exceeds the maximum size."
+
+
+class UnavailableForLegalReasons(HTTPException):
+    """*451* `Unavailable For Legal Reasons`
+
+    This status code indicates that the server is denying access to the
+    resource as a consequence of a legal demand.
+    """
+
+    code = 451
+    description = "Unavailable for legal reasons."
+
+
+class InternalServerError(HTTPException):
+    """*500* `Internal Server Error`
+
+    Raise if an internal server error occurred.  This is a good fallback if an
+    unknown error occurred in the dispatcher.
+
+    .. versionchanged:: 1.0.0
+        Added the :attr:`original_exception` attribute.
+    """
+
+    code = 500
+    description = (
+        "The server encountered an internal error and was unable to"
+        " complete your request. Either the server is overloaded or"
+        " there is an error in the application."
+    )
+
+    def __init__(
+        self,
+        description: str | None = None,
+        response: Response | None = None,
+        original_exception: BaseException | None = None,
+    ) -> None:
+        #: The original exception that caused this 500 error. Can be
+        #: used by frameworks to provide context when handling
+        #: unexpected errors.
+        self.original_exception = original_exception
+        super().__init__(description=description, response=response)
+
+
+class NotImplemented(HTTPException):
+    """*501* `Not Implemented`
+
+    Raise if the application does not support the action requested by the
+    browser.
+    """
+
+    code = 501
+    description = "The server does not support the action requested by the browser."
+
+
+class BadGateway(HTTPException):
+    """*502* `Bad Gateway`
+
+    If you do proxying in your application you should return this status code
+    if you received an invalid response from the upstream server it accessed
+    in attempting to fulfill the request.
+    """
+
+    code = 502
+    description = (
+        "The proxy server received an invalid response from an upstream server."
+    )
+
+
+class ServiceUnavailable(_RetryAfter):
+    """*503* `Service Unavailable`
+
+    Status code you should return if a service is temporarily
+    unavailable.
+
+    :param retry_after: If given, set the ``Retry-After`` header to this
+        value. May be an :class:`int` number of seconds or a
+        :class:`~datetime.datetime`.
+
+    .. versionchanged:: 1.0
+        Added ``retry_after`` parameter.
+    """
+
+    code = 503
+    description = (
+        "The server is temporarily unable to service your request due"
+        " to maintenance downtime or capacity problems. Please try"
+        " again later."
+    )
+
+
+class GatewayTimeout(HTTPException):
+    """*504* `Gateway Timeout`
+
+    Status code you should return if a connection to an upstream server
+    times out.
+    """
+
+    code = 504
+    description = "The connection to an upstream server timed out."
+
+
+class HTTPVersionNotSupported(HTTPException):
+    """*505* `HTTP Version Not Supported`
+
+    The server does not support the HTTP protocol version used in the request.
+    """
+
+    code = 505
+    description = (
+        "The server does not support the HTTP protocol version used in the request."
+    )
+
+
+default_exceptions: dict[int, type[HTTPException]] = {}
+
+
+def _find_exceptions() -> None:
+    for obj in globals().values():
+        try:
+            is_http_exception = issubclass(obj, HTTPException)
+        except TypeError:
+            is_http_exception = False
+        if not is_http_exception or obj.code is None:
+            continue
+        old_obj = default_exceptions.get(obj.code, None)
+        if old_obj is not None and issubclass(obj, old_obj):
+            continue
+        default_exceptions[obj.code] = obj
+
+
+_find_exceptions()
+del _find_exceptions
+
+
+class Aborter:
+    """When passed a dict of code -> exception items it can be used as
+    callable that raises exceptions.  If the first argument to the
+    callable is an integer it will be looked up in the mapping, if it's
+    a WSGI application it will be raised in a proxy exception.
+
+    The rest of the arguments are forwarded to the exception constructor.
+    """
+
+    def __init__(
+        self,
+        mapping: dict[int, type[HTTPException]] | None = None,
+        extra: dict[int, type[HTTPException]] | None = None,
+    ) -> None:
+        if mapping is None:
+            mapping = default_exceptions
+        self.mapping = dict(mapping)
+        if extra is not None:
+            self.mapping.update(extra)
+
+    def __call__(
+        self, code: int | Response, *args: t.Any, **kwargs: t.Any
+    ) -> t.NoReturn:
+        from .sansio.response import Response
+
+        if isinstance(code, Response):
+            raise HTTPException(response=code)
+
+        if code not in self.mapping:
+            raise LookupError(f"no exception for {code!r}")
+
+        raise self.mapping[code](*args, **kwargs)
+
+
+def abort(status: int | Response, *args: t.Any, **kwargs: t.Any) -> t.NoReturn:
+    """Raises an :py:exc:`HTTPException` for the given status code or WSGI
+    application.
+
+    If a status code is given, it will be looked up in the list of
+    exceptions and will raise that exception.  If passed a WSGI application,
+    it will wrap it in a proxy WSGI exception and raise that::
+
+       abort(404)  # 404 Not Found
+       abort(Response('Hello World'))
+
+    """
+    _aborter(status, *args, **kwargs)
+
+
+_aborter: Aborter = Aborter()
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/formparser.py b/backend/venv/lib/python3.12/site-packages/werkzeug/formparser.py
new file mode 100644
index 00000000..01034149
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/formparser.py
@@ -0,0 +1,430 @@
+from __future__ import annotations
+
+import typing as t
+from io import BytesIO
+from urllib.parse import parse_qsl
+
+from ._internal import _plain_int
+from .datastructures import FileStorage
+from .datastructures import Headers
+from .datastructures import MultiDict
+from .exceptions import RequestEntityTooLarge
+from .http import parse_options_header
+from .sansio.multipart import Data
+from .sansio.multipart import Epilogue
+from .sansio.multipart import Field
+from .sansio.multipart import File
+from .sansio.multipart import MultipartDecoder
+from .sansio.multipart import NeedData
+from .wsgi import get_content_length
+from .wsgi import get_input_stream
+
+# there are some platforms where SpooledTemporaryFile is not available.
+# In that case we need to provide a fallback.
+try:
+    from tempfile import SpooledTemporaryFile
+except ImportError:
+    from tempfile import TemporaryFile
+
+    SpooledTemporaryFile = None  # type: ignore
+
+if t.TYPE_CHECKING:
+    import typing as te
+
+    from _typeshed.wsgi import WSGIEnvironment
+
+    t_parse_result = tuple[
+        t.IO[bytes], MultiDict[str, str], MultiDict[str, FileStorage]
+    ]
+
+    class TStreamFactory(te.Protocol):
+        def __call__(
+            self,
+            total_content_length: int | None,
+            content_type: str | None,
+            filename: str | None,
+            content_length: int | None = None,
+        ) -> t.IO[bytes]: ...
+
+
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+
+def default_stream_factory(
+    total_content_length: int | None,
+    content_type: str | None,
+    filename: str | None,
+    content_length: int | None = None,
+) -> t.IO[bytes]:
+    max_size = 1024 * 500
+
+    if SpooledTemporaryFile is not None:
+        return t.cast(t.IO[bytes], SpooledTemporaryFile(max_size=max_size, mode="rb+"))
+    elif total_content_length is None or total_content_length > max_size:
+        return t.cast(t.IO[bytes], TemporaryFile("rb+"))
+
+    return BytesIO()
+
+
+def parse_form_data(
+    environ: WSGIEnvironment,
+    stream_factory: TStreamFactory | None = None,
+    max_form_memory_size: int | None = None,
+    max_content_length: int | None = None,
+    cls: type[MultiDict[str, t.Any]] | None = None,
+    silent: bool = True,
+    *,
+    max_form_parts: int | None = None,
+) -> t_parse_result:
+    """Parse the form data in the environ and return it as tuple in the form
+    ``(stream, form, files)``.  You should only call this method if the
+    transport method is `POST`, `PUT`, or `PATCH`.
+
+    If the mimetype of the data transmitted is `multipart/form-data` the
+    files multidict will be filled with `FileStorage` objects.  If the
+    mimetype is unknown the input stream is wrapped and returned as first
+    argument, else the stream is empty.
+
+    This is a shortcut for the common usage of :class:`FormDataParser`.
+
+    :param environ: the WSGI environment to be used for parsing.
+    :param stream_factory: An optional callable that returns a new read and
+                           writeable file descriptor.  This callable works
+                           the same as :meth:`Response._get_file_stream`.
+    :param max_form_memory_size: the maximum number of bytes to be accepted for
+                           in-memory stored form data.  If the data
+                           exceeds the value specified an
+                           :exc:`~exceptions.RequestEntityTooLarge`
+                           exception is raised.
+    :param max_content_length: If this is provided and the transmitted data
+                               is longer than this value an
+                               :exc:`~exceptions.RequestEntityTooLarge`
+                               exception is raised.
+    :param cls: an optional dict class to use.  If this is not specified
+                       or `None` the default :class:`MultiDict` is used.
+    :param silent: If set to False parsing errors will not be caught.
+    :param max_form_parts: The maximum number of multipart parts to be parsed. If this
+        is exceeded, a :exc:`~exceptions.RequestEntityTooLarge` exception is raised.
+    :return: A tuple in the form ``(stream, form, files)``.
+
+    .. versionchanged:: 3.0
+        The ``charset`` and ``errors`` parameters were removed.
+
+    .. versionchanged:: 2.3
+        Added the ``max_form_parts`` parameter.
+
+    .. versionadded:: 0.5.1
+       Added the ``silent`` parameter.
+
+    .. versionadded:: 0.5
+       Added the ``max_form_memory_size``, ``max_content_length``, and ``cls``
+       parameters.
+    """
+    return FormDataParser(
+        stream_factory=stream_factory,
+        max_form_memory_size=max_form_memory_size,
+        max_content_length=max_content_length,
+        max_form_parts=max_form_parts,
+        silent=silent,
+        cls=cls,
+    ).parse_from_environ(environ)
+
+
+class FormDataParser:
+    """This class implements parsing of form data for Werkzeug.  By itself
+    it can parse multipart and url encoded form data.  It can be subclassed
+    and extended but for most mimetypes it is a better idea to use the
+    untouched stream and expose it as separate attributes on a request
+    object.
+
+    :param stream_factory: An optional callable that returns a new read and
+                           writeable file descriptor.  This callable works
+                           the same as :meth:`Response._get_file_stream`.
+    :param max_form_memory_size: the maximum number of bytes to be accepted for
+                           in-memory stored form data.  If the data
+                           exceeds the value specified an
+                           :exc:`~exceptions.RequestEntityTooLarge`
+                           exception is raised.
+    :param max_content_length: If this is provided and the transmitted data
+                               is longer than this value an
+                               :exc:`~exceptions.RequestEntityTooLarge`
+                               exception is raised.
+    :param cls: an optional dict class to use.  If this is not specified
+                       or `None` the default :class:`MultiDict` is used.
+    :param silent: If set to False parsing errors will not be caught.
+    :param max_form_parts: The maximum number of multipart parts to be parsed. If this
+        is exceeded, a :exc:`~exceptions.RequestEntityTooLarge` exception is raised.
+
+    .. versionchanged:: 3.0
+        The ``charset`` and ``errors`` parameters were removed.
+
+    .. versionchanged:: 3.0
+        The ``parse_functions`` attribute and ``get_parse_func`` methods were removed.
+
+    .. versionchanged:: 2.2.3
+        Added the ``max_form_parts`` parameter.
+
+    .. versionadded:: 0.8
+    """
+
+    def __init__(
+        self,
+        stream_factory: TStreamFactory | None = None,
+        max_form_memory_size: int | None = None,
+        max_content_length: int | None = None,
+        cls: type[MultiDict[str, t.Any]] | None = None,
+        silent: bool = True,
+        *,
+        max_form_parts: int | None = None,
+    ) -> None:
+        if stream_factory is None:
+            stream_factory = default_stream_factory
+
+        self.stream_factory = stream_factory
+        self.max_form_memory_size = max_form_memory_size
+        self.max_content_length = max_content_length
+        self.max_form_parts = max_form_parts
+
+        if cls is None:
+            cls = t.cast("type[MultiDict[str, t.Any]]", MultiDict)
+
+        self.cls = cls
+        self.silent = silent
+
+    def parse_from_environ(self, environ: WSGIEnvironment) -> t_parse_result:
+        """Parses the information from the environment as form data.
+
+        :param environ: the WSGI environment to be used for parsing.
+        :return: A tuple in the form ``(stream, form, files)``.
+        """
+        stream = get_input_stream(environ, max_content_length=self.max_content_length)
+        content_length = get_content_length(environ)
+        mimetype, options = parse_options_header(environ.get("CONTENT_TYPE"))
+        return self.parse(
+            stream,
+            content_length=content_length,
+            mimetype=mimetype,
+            options=options,
+        )
+
+    def parse(
+        self,
+        stream: t.IO[bytes],
+        mimetype: str,
+        content_length: int | None,
+        options: dict[str, str] | None = None,
+    ) -> t_parse_result:
+        """Parses the information from the given stream, mimetype,
+        content length and mimetype parameters.
+
+        :param stream: an input stream
+        :param mimetype: the mimetype of the data
+        :param content_length: the content length of the incoming data
+        :param options: optional mimetype parameters (used for
+                        the multipart boundary for instance)
+        :return: A tuple in the form ``(stream, form, files)``.
+
+        .. versionchanged:: 3.0
+            The invalid ``application/x-url-encoded`` content type is not
+            treated as ``application/x-www-form-urlencoded``.
+        """
+        if mimetype == "multipart/form-data":
+            parse_func = self._parse_multipart
+        elif mimetype == "application/x-www-form-urlencoded":
+            parse_func = self._parse_urlencoded
+        else:
+            return stream, self.cls(), self.cls()
+
+        if options is None:
+            options = {}
+
+        try:
+            return parse_func(stream, mimetype, content_length, options)
+        except ValueError:
+            if not self.silent:
+                raise
+
+        return stream, self.cls(), self.cls()
+
+    def _parse_multipart(
+        self,
+        stream: t.IO[bytes],
+        mimetype: str,
+        content_length: int | None,
+        options: dict[str, str],
+    ) -> t_parse_result:
+        parser = MultiPartParser(
+            stream_factory=self.stream_factory,
+            max_form_memory_size=self.max_form_memory_size,
+            max_form_parts=self.max_form_parts,
+            cls=self.cls,
+        )
+        boundary = options.get("boundary", "").encode("ascii")
+
+        if not boundary:
+            raise ValueError("Missing boundary")
+
+        form, files = parser.parse(stream, boundary, content_length)
+        return stream, form, files
+
+    def _parse_urlencoded(
+        self,
+        stream: t.IO[bytes],
+        mimetype: str,
+        content_length: int | None,
+        options: dict[str, str],
+    ) -> t_parse_result:
+        if (
+            self.max_form_memory_size is not None
+            and content_length is not None
+            and content_length > self.max_form_memory_size
+        ):
+            raise RequestEntityTooLarge()
+
+        items = parse_qsl(
+            stream.read().decode(),
+            keep_blank_values=True,
+            errors="werkzeug.url_quote",
+        )
+        return stream, self.cls(items), self.cls()
+
+
+class MultiPartParser:
+    def __init__(
+        self,
+        stream_factory: TStreamFactory | None = None,
+        max_form_memory_size: int | None = None,
+        cls: type[MultiDict[str, t.Any]] | None = None,
+        buffer_size: int = 64 * 1024,
+        max_form_parts: int | None = None,
+    ) -> None:
+        self.max_form_memory_size = max_form_memory_size
+        self.max_form_parts = max_form_parts
+
+        if stream_factory is None:
+            stream_factory = default_stream_factory
+
+        self.stream_factory = stream_factory
+
+        if cls is None:
+            cls = t.cast("type[MultiDict[str, t.Any]]", MultiDict)
+
+        self.cls = cls
+        self.buffer_size = buffer_size
+
+    def fail(self, message: str) -> te.NoReturn:
+        raise ValueError(message)
+
+    def get_part_charset(self, headers: Headers) -> str:
+        # Figure out input charset for current part
+        content_type = headers.get("content-type")
+
+        if content_type:
+            parameters = parse_options_header(content_type)[1]
+            ct_charset = parameters.get("charset", "").lower()
+
+            # A safe list of encodings. Modern clients should only send ASCII or UTF-8.
+            # This list will not be extended further.
+            if ct_charset in {"ascii", "us-ascii", "utf-8", "iso-8859-1"}:
+                return ct_charset
+
+        return "utf-8"
+
+    def start_file_streaming(
+        self, event: File, total_content_length: int | None
+    ) -> t.IO[bytes]:
+        content_type = event.headers.get("content-type")
+
+        try:
+            content_length = _plain_int(event.headers["content-length"])
+        except (KeyError, ValueError):
+            content_length = 0
+
+        container = self.stream_factory(
+            total_content_length=total_content_length,
+            filename=event.filename,
+            content_type=content_type,
+            content_length=content_length,
+        )
+        return container
+
+    def parse(
+        self, stream: t.IO[bytes], boundary: bytes, content_length: int | None
+    ) -> tuple[MultiDict[str, str], MultiDict[str, FileStorage]]:
+        current_part: Field | File
+        field_size: int | None = None
+        container: t.IO[bytes] | list[bytes]
+        _write: t.Callable[[bytes], t.Any]
+
+        parser = MultipartDecoder(
+            boundary,
+            max_form_memory_size=self.max_form_memory_size,
+            max_parts=self.max_form_parts,
+        )
+
+        fields = []
+        files = []
+
+        for data in _chunk_iter(stream.read, self.buffer_size):
+            parser.receive_data(data)
+            event = parser.next_event()
+            while not isinstance(event, (Epilogue, NeedData)):
+                if isinstance(event, Field):
+                    current_part = event
+                    field_size = 0
+                    container = []
+                    _write = container.append
+                elif isinstance(event, File):
+                    current_part = event
+                    field_size = None
+                    container = self.start_file_streaming(event, content_length)
+                    _write = container.write
+                elif isinstance(event, Data):
+                    if self.max_form_memory_size is not None and field_size is not None:
+                        # Ensure that accumulated data events do not exceed limit.
+                        # Also checked within single event in MultipartDecoder.
+                        field_size += len(event.data)
+
+                        if field_size > self.max_form_memory_size:
+                            raise RequestEntityTooLarge()
+
+                    _write(event.data)
+                    if not event.more_data:
+                        if isinstance(current_part, Field):
+                            value = b"".join(container).decode(
+                                self.get_part_charset(current_part.headers), "replace"
+                            )
+                            fields.append((current_part.name, value))
+                        else:
+                            container = t.cast(t.IO[bytes], container)
+                            container.seek(0)
+                            files.append(
+                                (
+                                    current_part.name,
+                                    FileStorage(
+                                        container,
+                                        current_part.filename,
+                                        current_part.name,
+                                        headers=current_part.headers,
+                                    ),
+                                )
+                            )
+
+                event = parser.next_event()
+
+        return self.cls(fields), self.cls(files)
+
+
+def _chunk_iter(read: t.Callable[[int], bytes], size: int) -> t.Iterator[bytes | None]:
+    """Read data in chunks for multipart/form-data parsing. Stop if no data is read.
+    Yield ``None`` at the end to signal end of parsing.
+    """
+    while True:
+        data = read(size)
+
+        if not data:
+            break
+
+        yield data
+
+    yield None
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/http.py b/backend/venv/lib/python3.12/site-packages/werkzeug/http.py
new file mode 100644
index 00000000..f1dbb850
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/http.py
@@ -0,0 +1,1405 @@
+from __future__ import annotations
+
+import email.utils
+import re
+import typing as t
+import warnings
+from datetime import date
+from datetime import datetime
+from datetime import time
+from datetime import timedelta
+from datetime import timezone
+from enum import Enum
+from hashlib import sha1
+from time import mktime
+from time import struct_time
+from urllib.parse import quote
+from urllib.parse import unquote
+from urllib.request import parse_http_list as _parse_list_header
+
+from ._internal import _dt_as_utc
+from ._internal import _plain_int
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIEnvironment
+
+_token_chars = frozenset(
+    "!#$%&'*+-.0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ^_`abcdefghijklmnopqrstuvwxyz|~"
+)
+_etag_re = re.compile(r'([Ww]/)?(?:"(.*?)"|(.*?))(?:\s*,\s*|$)')
+_entity_headers = frozenset(
+    [
+        "allow",
+        "content-encoding",
+        "content-language",
+        "content-length",
+        "content-location",
+        "content-md5",
+        "content-range",
+        "content-type",
+        "expires",
+        "last-modified",
+    ]
+)
+_hop_by_hop_headers = frozenset(
+    [
+        "connection",
+        "keep-alive",
+        "proxy-authenticate",
+        "proxy-authorization",
+        "te",
+        "trailer",
+        "transfer-encoding",
+        "upgrade",
+    ]
+)
+HTTP_STATUS_CODES = {
+    100: "Continue",
+    101: "Switching Protocols",
+    102: "Processing",
+    103: "Early Hints",  # see RFC 8297
+    200: "OK",
+    201: "Created",
+    202: "Accepted",
+    203: "Non Authoritative Information",
+    204: "No Content",
+    205: "Reset Content",
+    206: "Partial Content",
+    207: "Multi Status",
+    208: "Already Reported",  # see RFC 5842
+    226: "IM Used",  # see RFC 3229
+    300: "Multiple Choices",
+    301: "Moved Permanently",
+    302: "Found",
+    303: "See Other",
+    304: "Not Modified",
+    305: "Use Proxy",
+    306: "Switch Proxy",  # unused
+    307: "Temporary Redirect",
+    308: "Permanent Redirect",
+    400: "Bad Request",
+    401: "Unauthorized",
+    402: "Payment Required",  # unused
+    403: "Forbidden",
+    404: "Not Found",
+    405: "Method Not Allowed",
+    406: "Not Acceptable",
+    407: "Proxy Authentication Required",
+    408: "Request Timeout",
+    409: "Conflict",
+    410: "Gone",
+    411: "Length Required",
+    412: "Precondition Failed",
+    413: "Request Entity Too Large",
+    414: "Request URI Too Long",
+    415: "Unsupported Media Type",
+    416: "Requested Range Not Satisfiable",
+    417: "Expectation Failed",
+    418: "I'm a teapot",  # see RFC 2324
+    421: "Misdirected Request",  # see RFC 7540
+    422: "Unprocessable Entity",
+    423: "Locked",
+    424: "Failed Dependency",
+    425: "Too Early",  # see RFC 8470
+    426: "Upgrade Required",
+    428: "Precondition Required",  # see RFC 6585
+    429: "Too Many Requests",
+    431: "Request Header Fields Too Large",
+    449: "Retry With",  # proprietary MS extension
+    451: "Unavailable For Legal Reasons",
+    500: "Internal Server Error",
+    501: "Not Implemented",
+    502: "Bad Gateway",
+    503: "Service Unavailable",
+    504: "Gateway Timeout",
+    505: "HTTP Version Not Supported",
+    506: "Variant Also Negotiates",  # see RFC 2295
+    507: "Insufficient Storage",
+    508: "Loop Detected",  # see RFC 5842
+    510: "Not Extended",
+    511: "Network Authentication Failed",
+}
+
+
+class COEP(Enum):
+    """Cross Origin Embedder Policies"""
+
+    UNSAFE_NONE = "unsafe-none"
+    REQUIRE_CORP = "require-corp"
+
+
+class COOP(Enum):
+    """Cross Origin Opener Policies"""
+
+    UNSAFE_NONE = "unsafe-none"
+    SAME_ORIGIN_ALLOW_POPUPS = "same-origin-allow-popups"
+    SAME_ORIGIN = "same-origin"
+
+
+def quote_header_value(value: t.Any, allow_token: bool = True) -> str:
+    """Add double quotes around a header value. If the header contains only ASCII token
+    characters, it will be returned unchanged. If the header contains ``"`` or ``\\``
+    characters, they will be escaped with an additional ``\\`` character.
+
+    This is the reverse of :func:`unquote_header_value`.
+
+    :param value: The value to quote. Will be converted to a string.
+    :param allow_token: Disable to quote the value even if it only has token characters.
+
+    .. versionchanged:: 3.0
+        Passing bytes is not supported.
+
+    .. versionchanged:: 3.0
+        The ``extra_chars`` parameter is removed.
+
+    .. versionchanged:: 2.3
+        The value is quoted if it is the empty string.
+
+    .. versionadded:: 0.5
+    """
+    value_str = str(value)
+
+    if not value_str:
+        return '""'
+
+    if allow_token:
+        token_chars = _token_chars
+
+        if token_chars.issuperset(value_str):
+            return value_str
+
+    value_str = value_str.replace("\\", "\\\\").replace('"', '\\"')
+    return f'"{value_str}"'
+
+
+def unquote_header_value(value: str) -> str:
+    """Remove double quotes and decode slash-escaped ``"`` and ``\\`` characters in a
+    header value.
+
+    This is the reverse of :func:`quote_header_value`.
+
+    :param value: The header value to unquote.
+
+    .. versionchanged:: 3.0
+        The ``is_filename`` parameter is removed.
+    """
+    if len(value) >= 2 and value[0] == value[-1] == '"':
+        value = value[1:-1]
+        return value.replace("\\\\", "\\").replace('\\"', '"')
+
+    return value
+
+
+def dump_options_header(header: str | None, options: t.Mapping[str, t.Any]) -> str:
+    """Produce a header value and ``key=value`` parameters separated by semicolons
+    ``;``. For example, the ``Content-Type`` header.
+
+    .. code-block:: python
+
+        dump_options_header("text/html", {"charset": "UTF-8"})
+        'text/html; charset=UTF-8'
+
+    This is the reverse of :func:`parse_options_header`.
+
+    If a value contains non-token characters, it will be quoted.
+
+    If a value is ``None``, the parameter is skipped.
+
+    In some keys for some headers, a UTF-8 value can be encoded using a special
+    ``key*=UTF-8''value`` form, where ``value`` is percent encoded. This function will
+    not produce that format automatically, but if a given key ends with an asterisk
+    ``*``, the value is assumed to have that form and will not be quoted further.
+
+    :param header: The primary header value.
+    :param options: Parameters to encode as ``key=value`` pairs.
+
+    .. versionchanged:: 2.3
+        Keys with ``None`` values are skipped rather than treated as a bare key.
+
+    .. versionchanged:: 2.2.3
+        If a key ends with ``*``, its value will not be quoted.
+    """
+    segments = []
+
+    if header is not None:
+        segments.append(header)
+
+    for key, value in options.items():
+        if value is None:
+            continue
+
+        if key[-1] == "*":
+            segments.append(f"{key}={value}")
+        else:
+            segments.append(f"{key}={quote_header_value(value)}")
+
+    return "; ".join(segments)
+
+
+def dump_header(iterable: dict[str, t.Any] | t.Iterable[t.Any]) -> str:
+    """Produce a header value from a list of items or ``key=value`` pairs, separated by
+    commas ``,``.
+
+    This is the reverse of :func:`parse_list_header`, :func:`parse_dict_header`, and
+    :func:`parse_set_header`.
+
+    If a value contains non-token characters, it will be quoted.
+
+    If a value is ``None``, the key is output alone.
+
+    In some keys for some headers, a UTF-8 value can be encoded using a special
+    ``key*=UTF-8''value`` form, where ``value`` is percent encoded. This function will
+    not produce that format automatically, but if a given key ends with an asterisk
+    ``*``, the value is assumed to have that form and will not be quoted further.
+
+    .. code-block:: python
+
+        dump_header(["foo", "bar baz"])
+        'foo, "bar baz"'
+
+        dump_header({"foo": "bar baz"})
+        'foo="bar baz"'
+
+    :param iterable: The items to create a header from.
+
+    .. versionchanged:: 3.0
+        The ``allow_token`` parameter is removed.
+
+    .. versionchanged:: 2.2.3
+        If a key ends with ``*``, its value will not be quoted.
+    """
+    if isinstance(iterable, dict):
+        items = []
+
+        for key, value in iterable.items():
+            if value is None:
+                items.append(key)
+            elif key[-1] == "*":
+                items.append(f"{key}={value}")
+            else:
+                items.append(f"{key}={quote_header_value(value)}")
+    else:
+        items = [quote_header_value(x) for x in iterable]
+
+    return ", ".join(items)
+
+
+def dump_csp_header(header: ds.ContentSecurityPolicy) -> str:
+    """Dump a Content Security Policy header.
+
+    These are structured into policies such as "default-src 'self';
+    script-src 'self'".
+
+    .. versionadded:: 1.0.0
+       Support for Content Security Policy headers was added.
+
+    """
+    return "; ".join(f"{key} {value}" for key, value in header.items())
+
+
+def parse_list_header(value: str) -> list[str]:
+    """Parse a header value that consists of a list of comma separated items according
+    to `RFC 9110 <https://httpwg.org/specs/rfc9110.html#abnf.extension>`__.
+
+    This extends :func:`urllib.request.parse_http_list` to remove surrounding quotes
+    from values.
+
+    .. code-block:: python
+
+        parse_list_header('token, "quoted value"')
+        ['token', 'quoted value']
+
+    This is the reverse of :func:`dump_header`.
+
+    :param value: The header value to parse.
+    """
+    result = []
+
+    for item in _parse_list_header(value):
+        if len(item) >= 2 and item[0] == item[-1] == '"':
+            item = item[1:-1]
+
+        result.append(item)
+
+    return result
+
+
+def parse_dict_header(value: str) -> dict[str, str | None]:
+    """Parse a list header using :func:`parse_list_header`, then parse each item as a
+    ``key=value`` pair.
+
+    .. code-block:: python
+
+        parse_dict_header('a=b, c="d, e", f')
+        {"a": "b", "c": "d, e", "f": None}
+
+    This is the reverse of :func:`dump_header`.
+
+    If a key does not have a value, it is ``None``.
+
+    This handles charsets for values as described in
+    `RFC 2231 <https://www.rfc-editor.org/rfc/rfc2231#section-3>`__. Only ASCII, UTF-8,
+    and ISO-8859-1 charsets are accepted, otherwise the value remains quoted.
+
+    :param value: The header value to parse.
+
+    .. versionchanged:: 3.0
+        Passing bytes is not supported.
+
+    .. versionchanged:: 3.0
+        The ``cls`` argument is removed.
+
+    .. versionchanged:: 2.3
+        Added support for ``key*=charset''value`` encoded items.
+
+    .. versionchanged:: 0.9
+       The ``cls`` argument was added.
+    """
+    result: dict[str, str | None] = {}
+
+    for item in parse_list_header(value):
+        key, has_value, value = item.partition("=")
+        key = key.strip()
+
+        if not key:
+            # =value is not valid
+            continue
+
+        if not has_value:
+            result[key] = None
+            continue
+
+        value = value.strip()
+        encoding: str | None = None
+
+        if key[-1] == "*":
+            # key*=charset''value becomes key=value, where value is percent encoded
+            # adapted from parse_options_header, without the continuation handling
+            key = key[:-1]
+            match = _charset_value_re.match(value)
+
+            if match:
+                # If there is a charset marker in the value, split it off.
+                encoding, value = match.groups()
+                encoding = encoding.lower()
+
+            # A safe list of encodings. Modern clients should only send ASCII or UTF-8.
+            # This list will not be extended further. An invalid encoding will leave the
+            # value quoted.
+            if encoding in {"ascii", "us-ascii", "utf-8", "iso-8859-1"}:
+                # invalid bytes are replaced during unquoting
+                value = unquote(value, encoding=encoding)
+
+        if len(value) >= 2 and value[0] == value[-1] == '"':
+            value = value[1:-1]
+
+        result[key] = value
+
+    return result
+
+
+# https://httpwg.org/specs/rfc9110.html#parameter
+_parameter_key_re = re.compile(r"([\w!#$%&'*+\-.^`|~]+)=", flags=re.ASCII)
+_parameter_token_value_re = re.compile(r"[\w!#$%&'*+\-.^`|~]+", flags=re.ASCII)
+# https://www.rfc-editor.org/rfc/rfc2231#section-4
+_charset_value_re = re.compile(
+    r"""
+    ([\w!#$%&*+\-.^`|~]*)'  # charset part, could be empty
+    [\w!#$%&*+\-.^`|~]*'  # don't care about language part, usually empty
+    ([\w!#$%&'*+\-.^`|~]+)  # one or more token chars with percent encoding
+    """,
+    re.ASCII | re.VERBOSE,
+)
+# https://www.rfc-editor.org/rfc/rfc2231#section-3
+_continuation_re = re.compile(r"\*(\d+)$", re.ASCII)
+
+
+def parse_options_header(value: str | None) -> tuple[str, dict[str, str]]:
+    """Parse a header that consists of a value with ``key=value`` parameters separated
+    by semicolons ``;``. For example, the ``Content-Type`` header.
+
+    .. code-block:: python
+
+        parse_options_header("text/html; charset=UTF-8")
+        ('text/html', {'charset': 'UTF-8'})
+
+        parse_options_header("")
+        ("", {})
+
+    This is the reverse of :func:`dump_options_header`.
+
+    This parses valid parameter parts as described in
+    `RFC 9110 <https://httpwg.org/specs/rfc9110.html#parameter>`__. Invalid parts are
+    skipped.
+
+    This handles continuations and charsets as described in
+    `RFC 2231 <https://www.rfc-editor.org/rfc/rfc2231#section-3>`__, although not as
+    strictly as the RFC. Only ASCII, UTF-8, and ISO-8859-1 charsets are accepted,
+    otherwise the value remains quoted.
+
+    Clients may not be consistent in how they handle a quote character within a quoted
+    value. The `HTML Standard <https://html.spec.whatwg.org/#multipart-form-data>`__
+    replaces it with ``%22`` in multipart form data.
+    `RFC 9110 <https://httpwg.org/specs/rfc9110.html#quoted.strings>`__ uses backslash
+    escapes in HTTP headers. Both are decoded to the ``"`` character.
+
+    Clients may not be consistent in how they handle non-ASCII characters. HTML
+    documents must declare ``<meta charset=UTF-8>``, otherwise browsers may replace with
+    HTML character references, which can be decoded using :func:`html.unescape`.
+
+    :param value: The header value to parse.
+    :return: ``(value, options)``, where ``options`` is a dict
+
+    .. versionchanged:: 2.3
+        Invalid parts, such as keys with no value, quoted keys, and incorrectly quoted
+        values, are discarded instead of treating as ``None``.
+
+    .. versionchanged:: 2.3
+        Only ASCII, UTF-8, and ISO-8859-1 are accepted for charset values.
+
+    .. versionchanged:: 2.3
+        Escaped quotes in quoted values, like ``%22`` and ``\\"``, are handled.
+
+    .. versionchanged:: 2.2
+        Option names are always converted to lowercase.
+
+    .. versionchanged:: 2.2
+        The ``multiple`` parameter was removed.
+
+    .. versionchanged:: 0.15
+        :rfc:`2231` parameter continuations are handled.
+
+    .. versionadded:: 0.5
+    """
+    if value is None:
+        return "", {}
+
+    value, _, rest = value.partition(";")
+    value = value.strip()
+    rest = rest.strip()
+
+    if not value or not rest:
+        # empty (invalid) value, or value without options
+        return value, {}
+
+    # Collect all valid key=value parts without processing the value.
+    parts: list[tuple[str, str]] = []
+
+    while True:
+        if (m := _parameter_key_re.match(rest)) is not None:
+            pk = m.group(1).lower()
+            rest = rest[m.end() :]
+
+            # Value may be a token.
+            if (m := _parameter_token_value_re.match(rest)) is not None:
+                parts.append((pk, m.group()))
+
+            # Value may be a quoted string, find the closing quote.
+            elif rest[:1] == '"':
+                pos = 1
+                length = len(rest)
+
+                while pos < length:
+                    if rest[pos : pos + 2] in {"\\\\", '\\"'}:
+                        # Consume escaped slashes and quotes.
+                        pos += 2
+                    elif rest[pos] == '"':
+                        # Stop at an unescaped quote.
+                        parts.append((pk, rest[: pos + 1]))
+                        rest = rest[pos + 1 :]
+                        break
+                    else:
+                        # Consume any other character.
+                        pos += 1
+
+        # Find the next section delimited by `;`, if any.
+        if (end := rest.find(";")) == -1:
+            break
+
+        rest = rest[end + 1 :].lstrip()
+
+    options: dict[str, str] = {}
+    encoding: str | None = None
+    continued_encoding: str | None = None
+
+    # For each collected part, process optional charset and continuation,
+    # unquote quoted values.
+    for pk, pv in parts:
+        if pk[-1] == "*":
+            # key*=charset''value becomes key=value, where value is percent encoded
+            pk = pk[:-1]
+            match = _charset_value_re.match(pv)
+
+            if match:
+                # If there is a valid charset marker in the value, split it off.
+                encoding, pv = match.groups()
+                # This might be the empty string, handled next.
+                encoding = encoding.lower()
+
+            # No charset marker, or marker with empty charset value.
+            if not encoding:
+                encoding = continued_encoding
+
+            # A safe list of encodings. Modern clients should only send ASCII or UTF-8.
+            # This list will not be extended further. An invalid encoding will leave the
+            # value quoted.
+            if encoding in {"ascii", "us-ascii", "utf-8", "iso-8859-1"}:
+                # Continuation parts don't require their own charset marker. This is
+                # looser than the RFC, it will persist across different keys and allows
+                # changing the charset during a continuation. But this implementation is
+                # much simpler than tracking the full state.
+                continued_encoding = encoding
+                # invalid bytes are replaced during unquoting
+                pv = unquote(pv, encoding=encoding)
+
+        # Remove quotes. At this point the value cannot be empty or a single quote.
+        if pv[0] == pv[-1] == '"':
+            # HTTP headers use slash, multipart form data uses percent
+            pv = pv[1:-1].replace("\\\\", "\\").replace('\\"', '"').replace("%22", '"')
+
+        match = _continuation_re.search(pk)
+
+        if match:
+            # key*0=a; key*1=b becomes key=ab
+            pk = pk[: match.start()]
+            options[pk] = options.get(pk, "") + pv
+        else:
+            options[pk] = pv
+
+    return value, options
+
+
+_q_value_re = re.compile(r"-?\d+(\.\d+)?", re.ASCII)
+_TAnyAccept = t.TypeVar("_TAnyAccept", bound="ds.Accept")
+
+
+@t.overload
+def parse_accept_header(value: str | None) -> ds.Accept: ...
+
+
+@t.overload
+def parse_accept_header(value: str | None, cls: type[_TAnyAccept]) -> _TAnyAccept: ...
+
+
+def parse_accept_header(
+    value: str | None, cls: type[_TAnyAccept] | None = None
+) -> _TAnyAccept:
+    """Parse an ``Accept`` header according to
+    `RFC 9110 <https://httpwg.org/specs/rfc9110.html#field.accept>`__.
+
+    Returns an :class:`.Accept` instance, which can sort and inspect items based on
+    their quality parameter. When parsing ``Accept-Charset``, ``Accept-Encoding``, or
+    ``Accept-Language``, pass the appropriate :class:`.Accept` subclass.
+
+    :param value: The header value to parse.
+    :param cls: The :class:`.Accept` class to wrap the result in.
+    :return: An instance of ``cls``.
+
+    .. versionchanged:: 2.3
+        Parse according to RFC 9110. Items with invalid ``q`` values are skipped.
+    """
+    if cls is None:
+        cls = t.cast(type[_TAnyAccept], ds.Accept)
+
+    if not value:
+        return cls(None)
+
+    result = []
+
+    for item in parse_list_header(value):
+        item, options = parse_options_header(item)
+
+        if "q" in options:
+            # pop q, remaining options are reconstructed
+            q_str = options.pop("q").strip()
+
+            if _q_value_re.fullmatch(q_str) is None:
+                # ignore an invalid q
+                continue
+
+            q = float(q_str)
+
+            if q < 0 or q > 1:
+                # ignore an invalid q
+                continue
+        else:
+            q = 1
+
+        if options:
+            # reconstruct the media type with any options
+            item = dump_options_header(item, options)
+
+        result.append((item, q))
+
+    return cls(result)
+
+
+_TAnyCC = t.TypeVar("_TAnyCC", bound="ds.cache_control._CacheControl")
+
+
+@t.overload
+def parse_cache_control_header(
+    value: str | None,
+    on_update: t.Callable[[ds.cache_control._CacheControl], None] | None = None,
+) -> ds.RequestCacheControl: ...
+
+
+@t.overload
+def parse_cache_control_header(
+    value: str | None,
+    on_update: t.Callable[[ds.cache_control._CacheControl], None] | None = None,
+    cls: type[_TAnyCC] = ...,
+) -> _TAnyCC: ...
+
+
+def parse_cache_control_header(
+    value: str | None,
+    on_update: t.Callable[[ds.cache_control._CacheControl], None] | None = None,
+    cls: type[_TAnyCC] | None = None,
+) -> _TAnyCC:
+    """Parse a cache control header.  The RFC differs between response and
+    request cache control, this method does not.  It's your responsibility
+    to not use the wrong control statements.
+
+    .. versionadded:: 0.5
+       The `cls` was added.  If not specified an immutable
+       :class:`~werkzeug.datastructures.RequestCacheControl` is returned.
+
+    :param value: a cache control header to be parsed.
+    :param on_update: an optional callable that is called every time a value
+                      on the :class:`~werkzeug.datastructures.CacheControl`
+                      object is changed.
+    :param cls: the class for the returned object.  By default
+                :class:`~werkzeug.datastructures.RequestCacheControl` is used.
+    :return: a `cls` object.
+    """
+    if cls is None:
+        cls = t.cast("type[_TAnyCC]", ds.RequestCacheControl)
+
+    if not value:
+        return cls((), on_update)
+
+    return cls(parse_dict_header(value), on_update)
+
+
+_TAnyCSP = t.TypeVar("_TAnyCSP", bound="ds.ContentSecurityPolicy")
+
+
+@t.overload
+def parse_csp_header(
+    value: str | None,
+    on_update: t.Callable[[ds.ContentSecurityPolicy], None] | None = None,
+) -> ds.ContentSecurityPolicy: ...
+
+
+@t.overload
+def parse_csp_header(
+    value: str | None,
+    on_update: t.Callable[[ds.ContentSecurityPolicy], None] | None = None,
+    cls: type[_TAnyCSP] = ...,
+) -> _TAnyCSP: ...
+
+
+def parse_csp_header(
+    value: str | None,
+    on_update: t.Callable[[ds.ContentSecurityPolicy], None] | None = None,
+    cls: type[_TAnyCSP] | None = None,
+) -> _TAnyCSP:
+    """Parse a Content Security Policy header.
+
+    .. versionadded:: 1.0.0
+       Support for Content Security Policy headers was added.
+
+    :param value: a csp header to be parsed.
+    :param on_update: an optional callable that is called every time a value
+                      on the object is changed.
+    :param cls: the class for the returned object.  By default
+                :class:`~werkzeug.datastructures.ContentSecurityPolicy` is used.
+    :return: a `cls` object.
+    """
+    if cls is None:
+        cls = t.cast("type[_TAnyCSP]", ds.ContentSecurityPolicy)
+
+    if value is None:
+        return cls((), on_update)
+
+    items = []
+
+    for policy in value.split(";"):
+        policy = policy.strip()
+
+        # Ignore badly formatted policies (no space)
+        if " " in policy:
+            directive, value = policy.strip().split(" ", 1)
+            items.append((directive.strip(), value.strip()))
+
+    return cls(items, on_update)
+
+
+def parse_set_header(
+    value: str | None,
+    on_update: t.Callable[[ds.HeaderSet], None] | None = None,
+) -> ds.HeaderSet:
+    """Parse a set-like header and return a
+    :class:`~werkzeug.datastructures.HeaderSet` object:
+
+    >>> hs = parse_set_header('token, "quoted value"')
+
+    The return value is an object that treats the items case-insensitively
+    and keeps the order of the items:
+
+    >>> 'TOKEN' in hs
+    True
+    >>> hs.index('quoted value')
+    1
+    >>> hs
+    HeaderSet(['token', 'quoted value'])
+
+    To create a header from the :class:`HeaderSet` again, use the
+    :func:`dump_header` function.
+
+    :param value: a set header to be parsed.
+    :param on_update: an optional callable that is called every time a
+                      value on the :class:`~werkzeug.datastructures.HeaderSet`
+                      object is changed.
+    :return: a :class:`~werkzeug.datastructures.HeaderSet`
+    """
+    if not value:
+        return ds.HeaderSet(None, on_update)
+    return ds.HeaderSet(parse_list_header(value), on_update)
+
+
+def parse_if_range_header(value: str | None) -> ds.IfRange:
+    """Parses an if-range header which can be an etag or a date.  Returns
+    a :class:`~werkzeug.datastructures.IfRange` object.
+
+    .. versionchanged:: 2.0
+        If the value represents a datetime, it is timezone-aware.
+
+    .. versionadded:: 0.7
+    """
+    if not value:
+        return ds.IfRange()
+    date = parse_date(value)
+    if date is not None:
+        return ds.IfRange(date=date)
+    # drop weakness information
+    return ds.IfRange(unquote_etag(value)[0])
+
+
+def parse_range_header(
+    value: str | None, make_inclusive: bool = True
+) -> ds.Range | None:
+    """Parses a range header into a :class:`~werkzeug.datastructures.Range`
+    object.  If the header is missing or malformed `None` is returned.
+    `ranges` is a list of ``(start, stop)`` tuples where the ranges are
+    non-inclusive.
+
+    .. versionadded:: 0.7
+    """
+    if not value or "=" not in value:
+        return None
+
+    ranges = []
+    last_end = 0
+    units, rng = value.split("=", 1)
+    units = units.strip().lower()
+
+    for item in rng.split(","):
+        item = item.strip()
+        if "-" not in item:
+            return None
+        if item.startswith("-"):
+            if last_end < 0:
+                return None
+            try:
+                begin = _plain_int(item)
+            except ValueError:
+                return None
+            end = None
+            last_end = -1
+        elif "-" in item:
+            begin_str, end_str = item.split("-", 1)
+            begin_str = begin_str.strip()
+            end_str = end_str.strip()
+
+            try:
+                begin = _plain_int(begin_str)
+            except ValueError:
+                return None
+
+            if begin < last_end or last_end < 0:
+                return None
+            if end_str:
+                try:
+                    end = _plain_int(end_str) + 1
+                except ValueError:
+                    return None
+
+                if begin >= end:
+                    return None
+            else:
+                end = None
+            last_end = end if end is not None else -1
+        ranges.append((begin, end))
+
+    return ds.Range(units, ranges)
+
+
+def parse_content_range_header(
+    value: str | None,
+    on_update: t.Callable[[ds.ContentRange], None] | None = None,
+) -> ds.ContentRange | None:
+    """Parses a range header into a
+    :class:`~werkzeug.datastructures.ContentRange` object or `None` if
+    parsing is not possible.
+
+    .. versionadded:: 0.7
+
+    :param value: a content range header to be parsed.
+    :param on_update: an optional callable that is called every time a value
+                      on the :class:`~werkzeug.datastructures.ContentRange`
+                      object is changed.
+    """
+    if value is None:
+        return None
+    try:
+        units, rangedef = (value or "").strip().split(None, 1)
+    except ValueError:
+        return None
+
+    if "/" not in rangedef:
+        return None
+    rng, length_str = rangedef.split("/", 1)
+    if length_str == "*":
+        length = None
+    else:
+        try:
+            length = _plain_int(length_str)
+        except ValueError:
+            return None
+
+    if rng == "*":
+        if not is_byte_range_valid(None, None, length):
+            return None
+
+        return ds.ContentRange(units, None, None, length, on_update=on_update)
+    elif "-" not in rng:
+        return None
+
+    start_str, stop_str = rng.split("-", 1)
+    try:
+        start = _plain_int(start_str)
+        stop = _plain_int(stop_str) + 1
+    except ValueError:
+        return None
+
+    if is_byte_range_valid(start, stop, length):
+        return ds.ContentRange(units, start, stop, length, on_update=on_update)
+
+    return None
+
+
+def quote_etag(etag: str, weak: bool = False) -> str:
+    """Quote an etag.
+
+    :param etag: the etag to quote.
+    :param weak: set to `True` to tag it "weak".
+    """
+    if '"' in etag:
+        raise ValueError("invalid etag")
+    etag = f'"{etag}"'
+    if weak:
+        etag = f"W/{etag}"
+    return etag
+
+
+@t.overload
+def unquote_etag(etag: str) -> tuple[str, bool]: ...
+@t.overload
+def unquote_etag(etag: None) -> tuple[None, None]: ...
+def unquote_etag(
+    etag: str | None,
+) -> tuple[str, bool] | tuple[None, None]:
+    """Unquote a single etag:
+
+    >>> unquote_etag('W/"bar"')
+    ('bar', True)
+    >>> unquote_etag('"bar"')
+    ('bar', False)
+
+    :param etag: the etag identifier to unquote.
+    :return: a ``(etag, weak)`` tuple.
+    """
+    if not etag:
+        return None, None
+    etag = etag.strip()
+    weak = False
+    if etag.startswith(("W/", "w/")):
+        weak = True
+        etag = etag[2:]
+    if etag[:1] == etag[-1:] == '"':
+        etag = etag[1:-1]
+    return etag, weak
+
+
+def parse_etags(value: str | None) -> ds.ETags:
+    """Parse an etag header.
+
+    :param value: the tag header to parse
+    :return: an :class:`~werkzeug.datastructures.ETags` object.
+    """
+    if not value:
+        return ds.ETags()
+    strong = []
+    weak = []
+    end = len(value)
+    pos = 0
+    while pos < end:
+        match = _etag_re.match(value, pos)
+        if match is None:
+            break
+        is_weak, quoted, raw = match.groups()
+        if raw == "*":
+            return ds.ETags(star_tag=True)
+        elif quoted:
+            raw = quoted
+        if is_weak:
+            weak.append(raw)
+        else:
+            strong.append(raw)
+        pos = match.end()
+    return ds.ETags(strong, weak)
+
+
+def generate_etag(data: bytes) -> str:
+    """Generate an etag for some data.
+
+    .. versionchanged:: 2.0
+        Use SHA-1. MD5 may not be available in some environments.
+    """
+    return sha1(data).hexdigest()
+
+
+def parse_date(value: str | None) -> datetime | None:
+    """Parse an :rfc:`2822` date into a timezone-aware
+    :class:`datetime.datetime` object, or ``None`` if parsing fails.
+
+    This is a wrapper for :func:`email.utils.parsedate_to_datetime`. It
+    returns ``None`` if parsing fails instead of raising an exception,
+    and always returns a timezone-aware datetime object. If the string
+    doesn't have timezone information, it is assumed to be UTC.
+
+    :param value: A string with a supported date format.
+
+    .. versionchanged:: 2.0
+        Return a timezone-aware datetime object. Use
+        ``email.utils.parsedate_to_datetime``.
+    """
+    if value is None:
+        return None
+
+    try:
+        dt = email.utils.parsedate_to_datetime(value)
+    except (TypeError, ValueError):
+        return None
+
+    if dt.tzinfo is None:
+        return dt.replace(tzinfo=timezone.utc)
+
+    return dt
+
+
+def http_date(
+    timestamp: datetime | date | int | float | struct_time | None = None,
+) -> str:
+    """Format a datetime object or timestamp into an :rfc:`2822` date
+    string.
+
+    This is a wrapper for :func:`email.utils.format_datetime`. It
+    assumes naive datetime objects are in UTC instead of raising an
+    exception.
+
+    :param timestamp: The datetime or timestamp to format. Defaults to
+        the current time.
+
+    .. versionchanged:: 2.0
+        Use ``email.utils.format_datetime``. Accept ``date`` objects.
+    """
+    if isinstance(timestamp, date):
+        if not isinstance(timestamp, datetime):
+            # Assume plain date is midnight UTC.
+            timestamp = datetime.combine(timestamp, time(), tzinfo=timezone.utc)
+        else:
+            # Ensure datetime is timezone-aware.
+            timestamp = _dt_as_utc(timestamp)
+
+        return email.utils.format_datetime(timestamp, usegmt=True)
+
+    if isinstance(timestamp, struct_time):
+        timestamp = mktime(timestamp)
+
+    return email.utils.formatdate(timestamp, usegmt=True)
+
+
+def parse_age(value: str | None = None) -> timedelta | None:
+    """Parses a base-10 integer count of seconds into a timedelta.
+
+    If parsing fails, the return value is `None`.
+
+    :param value: a string consisting of an integer represented in base-10
+    :return: a :class:`datetime.timedelta` object or `None`.
+    """
+    if not value:
+        return None
+    try:
+        seconds = int(value)
+    except ValueError:
+        return None
+    if seconds < 0:
+        return None
+    try:
+        return timedelta(seconds=seconds)
+    except OverflowError:
+        return None
+
+
+def dump_age(age: timedelta | int | None = None) -> str | None:
+    """Formats the duration as a base-10 integer.
+
+    :param age: should be an integer number of seconds,
+                a :class:`datetime.timedelta` object, or,
+                if the age is unknown, `None` (default).
+    """
+    if age is None:
+        return None
+    if isinstance(age, timedelta):
+        age = int(age.total_seconds())
+    else:
+        age = int(age)
+
+    if age < 0:
+        raise ValueError("age cannot be negative")
+
+    return str(age)
+
+
+def is_resource_modified(
+    environ: WSGIEnvironment,
+    etag: str | None = None,
+    data: bytes | None = None,
+    last_modified: datetime | str | None = None,
+    ignore_if_range: bool = True,
+) -> bool:
+    """Convenience method for conditional requests.
+
+    :param environ: the WSGI environment of the request to be checked.
+    :param etag: the etag for the response for comparison.
+    :param data: or alternatively the data of the response to automatically
+                 generate an etag using :func:`generate_etag`.
+    :param last_modified: an optional date of the last modification.
+    :param ignore_if_range: If `False`, `If-Range` header will be taken into
+                            account.
+    :return: `True` if the resource was modified, otherwise `False`.
+
+    .. versionchanged:: 2.0
+        SHA-1 is used to generate an etag value for the data. MD5 may
+        not be available in some environments.
+
+    .. versionchanged:: 1.0.0
+        The check is run for methods other than ``GET`` and ``HEAD``.
+    """
+    return _sansio_http.is_resource_modified(
+        http_range=environ.get("HTTP_RANGE"),
+        http_if_range=environ.get("HTTP_IF_RANGE"),
+        http_if_modified_since=environ.get("HTTP_IF_MODIFIED_SINCE"),
+        http_if_none_match=environ.get("HTTP_IF_NONE_MATCH"),
+        http_if_match=environ.get("HTTP_IF_MATCH"),
+        etag=etag,
+        data=data,
+        last_modified=last_modified,
+        ignore_if_range=ignore_if_range,
+    )
+
+
+def remove_entity_headers(
+    headers: ds.Headers | list[tuple[str, str]],
+    allowed: t.Iterable[str] = ("expires", "content-location"),
+) -> None:
+    """Remove all entity headers from a list or :class:`Headers` object.  This
+    operation works in-place.  `Expires` and `Content-Location` headers are
+    by default not removed.  The reason for this is :rfc:`2616` section
+    10.3.5 which specifies some entity headers that should be sent.
+
+    .. versionchanged:: 0.5
+       added `allowed` parameter.
+
+    :param headers: a list or :class:`Headers` object.
+    :param allowed: a list of headers that should still be allowed even though
+                    they are entity headers.
+    """
+    allowed = {x.lower() for x in allowed}
+    headers[:] = [
+        (key, value)
+        for key, value in headers
+        if not is_entity_header(key) or key.lower() in allowed
+    ]
+
+
+def remove_hop_by_hop_headers(headers: ds.Headers | list[tuple[str, str]]) -> None:
+    """Remove all HTTP/1.1 "Hop-by-Hop" headers from a list or
+    :class:`Headers` object.  This operation works in-place.
+
+    .. versionadded:: 0.5
+
+    :param headers: a list or :class:`Headers` object.
+    """
+    headers[:] = [
+        (key, value) for key, value in headers if not is_hop_by_hop_header(key)
+    ]
+
+
+def is_entity_header(header: str) -> bool:
+    """Check if a header is an entity header.
+
+    .. versionadded:: 0.5
+
+    :param header: the header to test.
+    :return: `True` if it's an entity header, `False` otherwise.
+    """
+    return header.lower() in _entity_headers
+
+
+def is_hop_by_hop_header(header: str) -> bool:
+    """Check if a header is an HTTP/1.1 "Hop-by-Hop" header.
+
+    .. versionadded:: 0.5
+
+    :param header: the header to test.
+    :return: `True` if it's an HTTP/1.1 "Hop-by-Hop" header, `False` otherwise.
+    """
+    return header.lower() in _hop_by_hop_headers
+
+
+def parse_cookie(
+    header: WSGIEnvironment | str | None,
+    cls: type[ds.MultiDict[str, str]] | None = None,
+) -> ds.MultiDict[str, str]:
+    """Parse a cookie from a string or WSGI environ.
+
+    The same key can be provided multiple times, the values are stored
+    in-order. The default :class:`MultiDict` will have the first value
+    first, and all values can be retrieved with
+    :meth:`MultiDict.getlist`.
+
+    :param header: The cookie header as a string, or a WSGI environ dict
+        with a ``HTTP_COOKIE`` key.
+    :param cls: A dict-like class to store the parsed cookies in.
+        Defaults to :class:`MultiDict`.
+
+    .. versionchanged:: 3.0
+        Passing bytes, and the ``charset`` and ``errors`` parameters, were removed.
+
+    .. versionchanged:: 1.0
+        Returns a :class:`MultiDict` instead of a ``TypeConversionDict``.
+
+    .. versionchanged:: 0.5
+        Returns a :class:`TypeConversionDict` instead of a regular dict. The ``cls``
+        parameter was added.
+    """
+    if isinstance(header, dict):
+        cookie = header.get("HTTP_COOKIE")
+    else:
+        cookie = header
+
+    if cookie:
+        cookie = cookie.encode("latin1").decode()
+
+    return _sansio_http.parse_cookie(cookie=cookie, cls=cls)
+
+
+_cookie_no_quote_re = re.compile(r"[\w!#$%&'()*+\-./:<=>?@\[\]^`{|}~]*", re.A)
+_cookie_slash_re = re.compile(rb"[\x00-\x19\",;\\\x7f-\xff]", re.A)
+_cookie_slash_map = {b'"': b'\\"', b"\\": b"\\\\"}
+_cookie_slash_map.update(
+    (v.to_bytes(1, "big"), b"\\%03o" % v)
+    for v in [*range(0x20), *b",;", *range(0x7F, 256)]
+)
+
+
+def dump_cookie(
+    key: str,
+    value: str = "",
+    max_age: timedelta | int | None = None,
+    expires: str | datetime | int | float | None = None,
+    path: str | None = "/",
+    domain: str | None = None,
+    secure: bool = False,
+    httponly: bool = False,
+    sync_expires: bool = True,
+    max_size: int = 4093,
+    samesite: str | None = None,
+    partitioned: bool = False,
+) -> str:
+    """Create a Set-Cookie header without the ``Set-Cookie`` prefix.
+
+    The return value is usually restricted to ascii as the vast majority
+    of values are properly escaped, but that is no guarantee. It's
+    tunneled through latin1 as required by :pep:`3333`.
+
+    The return value is not ASCII safe if the key contains unicode
+    characters.  This is technically against the specification but
+    happens in the wild.  It's strongly recommended to not use
+    non-ASCII values for the keys.
+
+    :param max_age: should be a number of seconds, or `None` (default) if
+                    the cookie should last only as long as the client's
+                    browser session.  Additionally `timedelta` objects
+                    are accepted, too.
+    :param expires: should be a `datetime` object or unix timestamp.
+    :param path: limits the cookie to a given path, per default it will
+                 span the whole domain.
+    :param domain: Use this if you want to set a cross-domain cookie. For
+                   example, ``domain="example.com"`` will set a cookie
+                   that is readable by the domain ``www.example.com``,
+                   ``foo.example.com`` etc. Otherwise, a cookie will only
+                   be readable by the domain that set it.
+    :param secure: The cookie will only be available via HTTPS
+    :param httponly: disallow JavaScript to access the cookie.  This is an
+                     extension to the cookie standard and probably not
+                     supported by all browsers.
+    :param charset: the encoding for string values.
+    :param sync_expires: automatically set expires if max_age is defined
+                         but expires not.
+    :param max_size: Warn if the final header value exceeds this size. The
+        default, 4093, should be safely `supported by most browsers
+        <cookie_>`_. Set to 0 to disable this check.
+    :param samesite: Limits the scope of the cookie such that it will
+        only be attached to requests if those requests are same-site.
+    :param partitioned: Opts the cookie into partitioned storage. This
+        will also set secure to True
+
+    .. _`cookie`: http://browsercookielimits.squawky.net/
+
+    .. versionchanged:: 3.1
+        The ``partitioned`` parameter was added.
+
+    .. versionchanged:: 3.0
+        Passing bytes, and the ``charset`` parameter, were removed.
+
+    .. versionchanged:: 2.3.3
+        The ``path`` parameter is ``/`` by default.
+
+    .. versionchanged:: 2.3.1
+        The value allows more characters without quoting.
+
+    .. versionchanged:: 2.3
+        ``localhost`` and other names without a dot are allowed for the domain. A
+        leading dot is ignored.
+
+    .. versionchanged:: 2.3
+        The ``path`` parameter is ``None`` by default.
+
+    .. versionchanged:: 1.0.0
+        The string ``'None'`` is accepted for ``samesite``.
+    """
+    if path is not None:
+        # safe = https://url.spec.whatwg.org/#url-path-segment-string
+        # as well as percent for things that are already quoted
+        # excluding semicolon since it's part of the header syntax
+        path = quote(path, safe="%!$&'()*+,/:=@")
+
+    if domain:
+        domain = domain.partition(":")[0].lstrip(".").encode("idna").decode("ascii")
+
+    if isinstance(max_age, timedelta):
+        max_age = int(max_age.total_seconds())
+
+    if expires is not None:
+        if not isinstance(expires, str):
+            expires = http_date(expires)
+    elif max_age is not None and sync_expires:
+        expires = http_date(datetime.now(tz=timezone.utc).timestamp() + max_age)
+
+    if samesite is not None:
+        samesite = samesite.title()
+
+        if samesite not in {"Strict", "Lax", "None"}:
+            raise ValueError("SameSite must be 'Strict', 'Lax', or 'None'.")
+
+    if partitioned:
+        secure = True
+
+    # Quote value if it contains characters not allowed by RFC 6265. Slash-escape with
+    # three octal digits, which matches http.cookies, although the RFC suggests base64.
+    if not _cookie_no_quote_re.fullmatch(value):
+        # Work with bytes here, since a UTF-8 character could be multiple bytes.
+        value = _cookie_slash_re.sub(
+            lambda m: _cookie_slash_map[m.group()], value.encode()
+        ).decode("ascii")
+        value = f'"{value}"'
+
+    # Send a non-ASCII key as mojibake. Everything else should already be ASCII.
+    # TODO Remove encoding dance, it seems like clients accept UTF-8 keys
+    buf = [f"{key.encode().decode('latin1')}={value}"]
+
+    for k, v in (
+        ("Domain", domain),
+        ("Expires", expires),
+        ("Max-Age", max_age),
+        ("Secure", secure),
+        ("HttpOnly", httponly),
+        ("Path", path),
+        ("SameSite", samesite),
+        ("Partitioned", partitioned),
+    ):
+        if v is None or v is False:
+            continue
+
+        if v is True:
+            buf.append(k)
+            continue
+
+        buf.append(f"{k}={v}")
+
+    rv = "; ".join(buf)
+
+    # Warn if the final value of the cookie is larger than the limit. If the cookie is
+    # too large, then it may be silently ignored by the browser, which can be quite hard
+    # to debug.
+    cookie_size = len(rv)
+
+    if max_size and cookie_size > max_size:
+        value_size = len(value)
+        warnings.warn(
+            f"The '{key}' cookie is too large: the value was {value_size} bytes but the"
+            f" header required {cookie_size - value_size} extra bytes. The final size"
+            f" was {cookie_size} bytes but the limit is {max_size} bytes. Browsers may"
+            " silently ignore cookies larger than this.",
+            stacklevel=2,
+        )
+
+    return rv
+
+
+def is_byte_range_valid(
+    start: int | None, stop: int | None, length: int | None
+) -> bool:
+    """Checks if a given byte content range is valid for the given length.
+
+    .. versionadded:: 0.7
+    """
+    if (start is None) != (stop is None):
+        return False
+    elif start is None:
+        return length is None or length >= 0
+    elif length is None:
+        return 0 <= start < stop  # type: ignore
+    elif start >= stop:  # type: ignore
+        return False
+    return 0 <= start < length
+
+
+# circular dependencies
+from . import datastructures as ds
+from .sansio import http as _sansio_http
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/local.py b/backend/venv/lib/python3.12/site-packages/werkzeug/local.py
new file mode 100644
index 00000000..302589bb
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/local.py
@@ -0,0 +1,653 @@
+from __future__ import annotations
+
+import copy
+import math
+import operator
+import typing as t
+from contextvars import ContextVar
+from functools import partial
+from functools import update_wrapper
+from operator import attrgetter
+
+from .wsgi import ClosingIterator
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+T = t.TypeVar("T")
+F = t.TypeVar("F", bound=t.Callable[..., t.Any])
+
+
+def release_local(local: Local | LocalStack[t.Any]) -> None:
+    """Release the data for the current context in a :class:`Local` or
+    :class:`LocalStack` without using a :class:`LocalManager`.
+
+    This should not be needed for modern use cases, and may be removed
+    in the future.
+
+    .. versionadded:: 0.6.1
+    """
+    local.__release_local__()
+
+
+class Local:
+    """Create a namespace of context-local data. This wraps a
+    :class:`ContextVar` containing a :class:`dict` value.
+
+    This may incur a performance penalty compared to using individual
+    context vars, as it has to copy data to avoid mutating the dict
+    between nested contexts.
+
+    :param context_var: The :class:`~contextvars.ContextVar` to use as
+        storage for this local. If not given, one will be created.
+        Context vars not created at the global scope may interfere with
+        garbage collection.
+
+    .. versionchanged:: 2.0
+        Uses ``ContextVar`` instead of a custom storage implementation.
+    """
+
+    __slots__ = ("__storage",)
+
+    def __init__(self, context_var: ContextVar[dict[str, t.Any]] | None = None) -> None:
+        if context_var is None:
+            # A ContextVar not created at global scope interferes with
+            # Python's garbage collection. However, a local only makes
+            # sense defined at the global scope as well, in which case
+            # the GC issue doesn't seem relevant.
+            context_var = ContextVar(f"werkzeug.Local<{id(self)}>.storage")
+
+        object.__setattr__(self, "_Local__storage", context_var)
+
+    def __iter__(self) -> t.Iterator[tuple[str, t.Any]]:
+        return iter(self.__storage.get({}).items())
+
+    def __call__(
+        self, name: str, *, unbound_message: str | None = None
+    ) -> LocalProxy[t.Any]:
+        """Create a :class:`LocalProxy` that access an attribute on this
+        local namespace.
+
+        :param name: Proxy this attribute.
+        :param unbound_message: The error message that the proxy will
+            show if the attribute isn't set.
+        """
+        return LocalProxy(self, name, unbound_message=unbound_message)
+
+    def __release_local__(self) -> None:
+        self.__storage.set({})
+
+    def __getattr__(self, name: str) -> t.Any:
+        values = self.__storage.get({})
+
+        if name in values:
+            return values[name]
+
+        raise AttributeError(name)
+
+    def __setattr__(self, name: str, value: t.Any) -> None:
+        values = self.__storage.get({}).copy()
+        values[name] = value
+        self.__storage.set(values)
+
+    def __delattr__(self, name: str) -> None:
+        values = self.__storage.get({})
+
+        if name in values:
+            values = values.copy()
+            del values[name]
+            self.__storage.set(values)
+        else:
+            raise AttributeError(name)
+
+
+class LocalStack(t.Generic[T]):
+    """Create a stack of context-local data. This wraps a
+    :class:`ContextVar` containing a :class:`list` value.
+
+    This may incur a performance penalty compared to using individual
+    context vars, as it has to copy data to avoid mutating the list
+    between nested contexts.
+
+    :param context_var: The :class:`~contextvars.ContextVar` to use as
+        storage for this local. If not given, one will be created.
+        Context vars not created at the global scope may interfere with
+        garbage collection.
+
+    .. versionchanged:: 2.0
+        Uses ``ContextVar`` instead of a custom storage implementation.
+
+    .. versionadded:: 0.6.1
+    """
+
+    __slots__ = ("_storage",)
+
+    def __init__(self, context_var: ContextVar[list[T]] | None = None) -> None:
+        if context_var is None:
+            # A ContextVar not created at global scope interferes with
+            # Python's garbage collection. However, a local only makes
+            # sense defined at the global scope as well, in which case
+            # the GC issue doesn't seem relevant.
+            context_var = ContextVar(f"werkzeug.LocalStack<{id(self)}>.storage")
+
+        self._storage = context_var
+
+    def __release_local__(self) -> None:
+        self._storage.set([])
+
+    def push(self, obj: T) -> list[T]:
+        """Add a new item to the top of the stack."""
+        stack = self._storage.get([]).copy()
+        stack.append(obj)
+        self._storage.set(stack)
+        return stack
+
+    def pop(self) -> T | None:
+        """Remove the top item from the stack and return it. If the
+        stack is empty, return ``None``.
+        """
+        stack = self._storage.get([])
+
+        if len(stack) == 0:
+            return None
+
+        rv = stack[-1]
+        self._storage.set(stack[:-1])
+        return rv
+
+    @property
+    def top(self) -> T | None:
+        """The topmost item on the stack.  If the stack is empty,
+        `None` is returned.
+        """
+        stack = self._storage.get([])
+
+        if len(stack) == 0:
+            return None
+
+        return stack[-1]
+
+    def __call__(
+        self, name: str | None = None, *, unbound_message: str | None = None
+    ) -> LocalProxy[t.Any]:
+        """Create a :class:`LocalProxy` that accesses the top of this
+        local stack.
+
+        :param name: If given, the proxy access this attribute of the
+            top item, rather than the item itself.
+        :param unbound_message: The error message that the proxy will
+            show if the stack is empty.
+        """
+        return LocalProxy(self, name, unbound_message=unbound_message)
+
+
+class LocalManager:
+    """Manage releasing the data for the current context in one or more
+    :class:`Local` and :class:`LocalStack` objects.
+
+    This should not be needed for modern use cases, and may be removed
+    in the future.
+
+    :param locals: A local or list of locals to manage.
+
+    .. versionchanged:: 2.1
+        The ``ident_func`` was removed.
+
+    .. versionchanged:: 0.7
+        The ``ident_func`` parameter was added.
+
+    .. versionchanged:: 0.6.1
+        The :func:`release_local` function can be used instead of a
+        manager.
+    """
+
+    __slots__ = ("locals",)
+
+    def __init__(
+        self,
+        locals: None
+        | (Local | LocalStack[t.Any] | t.Iterable[Local | LocalStack[t.Any]]) = None,
+    ) -> None:
+        if locals is None:
+            self.locals = []
+        elif isinstance(locals, Local):
+            self.locals = [locals]
+        else:
+            self.locals = list(locals)  # type: ignore[arg-type]
+
+    def cleanup(self) -> None:
+        """Release the data in the locals for this context. Call this at
+        the end of each request or use :meth:`make_middleware`.
+        """
+        for local in self.locals:
+            release_local(local)
+
+    def make_middleware(self, app: WSGIApplication) -> WSGIApplication:
+        """Wrap a WSGI application so that local data is released
+        automatically after the response has been sent for a request.
+        """
+
+        def application(
+            environ: WSGIEnvironment, start_response: StartResponse
+        ) -> t.Iterable[bytes]:
+            return ClosingIterator(app(environ, start_response), self.cleanup)
+
+        return application
+
+    def middleware(self, func: WSGIApplication) -> WSGIApplication:
+        """Like :meth:`make_middleware` but used as a decorator on the
+        WSGI application function.
+
+        .. code-block:: python
+
+            @manager.middleware
+            def application(environ, start_response):
+                ...
+        """
+        return update_wrapper(self.make_middleware(func), func)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} storages: {len(self.locals)}>"
+
+
+class _ProxyLookup:
+    """Descriptor that handles proxied attribute lookup for
+    :class:`LocalProxy`.
+
+    :param f: The built-in function this attribute is accessed through.
+        Instead of looking up the special method, the function call
+        is redone on the object.
+    :param fallback: Return this function if the proxy is unbound
+        instead of raising a :exc:`RuntimeError`.
+    :param is_attr: This proxied name is an attribute, not a function.
+        Call the fallback immediately to get the value.
+    :param class_value: Value to return when accessed from the
+        ``LocalProxy`` class directly. Used for ``__doc__`` so building
+        docs still works.
+    """
+
+    __slots__ = ("bind_f", "fallback", "is_attr", "class_value", "name")
+
+    def __init__(
+        self,
+        f: t.Callable[..., t.Any] | None = None,
+        fallback: t.Callable[[LocalProxy[t.Any]], t.Any] | None = None,
+        class_value: t.Any | None = None,
+        is_attr: bool = False,
+    ) -> None:
+        bind_f: t.Callable[[LocalProxy[t.Any], t.Any], t.Callable[..., t.Any]] | None
+
+        if hasattr(f, "__get__"):
+            # A Python function, can be turned into a bound method.
+
+            def bind_f(
+                instance: LocalProxy[t.Any], obj: t.Any
+            ) -> t.Callable[..., t.Any]:
+                return f.__get__(obj, type(obj))  # type: ignore
+
+        elif f is not None:
+            # A C function, use partial to bind the first argument.
+
+            def bind_f(
+                instance: LocalProxy[t.Any], obj: t.Any
+            ) -> t.Callable[..., t.Any]:
+                return partial(f, obj)
+
+        else:
+            # Use getattr, which will produce a bound method.
+            bind_f = None
+
+        self.bind_f = bind_f
+        self.fallback = fallback
+        self.class_value = class_value
+        self.is_attr = is_attr
+
+    def __set_name__(self, owner: LocalProxy[t.Any], name: str) -> None:
+        self.name = name
+
+    def __get__(self, instance: LocalProxy[t.Any], owner: type | None = None) -> t.Any:
+        if instance is None:
+            if self.class_value is not None:
+                return self.class_value
+
+            return self
+
+        try:
+            obj = instance._get_current_object()
+        except RuntimeError:
+            if self.fallback is None:
+                raise
+
+            fallback = self.fallback.__get__(instance, owner)
+
+            if self.is_attr:
+                # __class__ and __doc__ are attributes, not methods.
+                # Call the fallback to get the value.
+                return fallback()
+
+            return fallback
+
+        if self.bind_f is not None:
+            return self.bind_f(instance, obj)
+
+        return getattr(obj, self.name)
+
+    def __repr__(self) -> str:
+        return f"proxy {self.name}"
+
+    def __call__(
+        self, instance: LocalProxy[t.Any], *args: t.Any, **kwargs: t.Any
+    ) -> t.Any:
+        """Support calling unbound methods from the class. For example,
+        this happens with ``copy.copy``, which does
+        ``type(x).__copy__(x)``. ``type(x)`` can't be proxied, so it
+        returns the proxy type and descriptor.
+        """
+        return self.__get__(instance, type(instance))(*args, **kwargs)
+
+
+class _ProxyIOp(_ProxyLookup):
+    """Look up an augmented assignment method on a proxied object. The
+    method is wrapped to return the proxy instead of the object.
+    """
+
+    __slots__ = ()
+
+    def __init__(
+        self,
+        f: t.Callable[..., t.Any] | None = None,
+        fallback: t.Callable[[LocalProxy[t.Any]], t.Any] | None = None,
+    ) -> None:
+        super().__init__(f, fallback)
+
+        def bind_f(instance: LocalProxy[t.Any], obj: t.Any) -> t.Callable[..., t.Any]:
+            def i_op(self: t.Any, other: t.Any) -> LocalProxy[t.Any]:
+                f(self, other)  # type: ignore
+                return instance
+
+            return i_op.__get__(obj, type(obj))  # type: ignore
+
+        self.bind_f = bind_f
+
+
+def _l_to_r_op(op: F) -> F:
+    """Swap the argument order to turn an l-op into an r-op."""
+
+    def r_op(obj: t.Any, other: t.Any) -> t.Any:
+        return op(other, obj)
+
+    return t.cast(F, r_op)
+
+
+def _identity(o: T) -> T:
+    return o
+
+
+class LocalProxy(t.Generic[T]):
+    """A proxy to the object bound to a context-local object. All
+    operations on the proxy are forwarded to the bound object. If no
+    object is bound, a ``RuntimeError`` is raised.
+
+    :param local: The context-local object that provides the proxied
+        object.
+    :param name: Proxy this attribute from the proxied object.
+    :param unbound_message: The error message to show if the
+        context-local object is unbound.
+
+    Proxy a :class:`~contextvars.ContextVar` to make it easier to
+    access. Pass a name to proxy that attribute.
+
+    .. code-block:: python
+
+        _request_var = ContextVar("request")
+        request = LocalProxy(_request_var)
+        session = LocalProxy(_request_var, "session")
+
+    Proxy an attribute on a :class:`Local` namespace by calling the
+    local with the attribute name:
+
+    .. code-block:: python
+
+        data = Local()
+        user = data("user")
+
+    Proxy the top item on a :class:`LocalStack` by calling the local.
+    Pass a name to proxy that attribute.
+
+    .. code-block::
+
+        app_stack = LocalStack()
+        current_app = app_stack()
+        g = app_stack("g")
+
+    Pass a function to proxy the return value from that function. This
+    was previously used to access attributes of local objects before
+    that was supported directly.
+
+    .. code-block:: python
+
+        session = LocalProxy(lambda: request.session)
+
+    ``__repr__`` and ``__class__`` are proxied, so ``repr(x)`` and
+    ``isinstance(x, cls)`` will look like the proxied object. Use
+    ``issubclass(type(x), LocalProxy)`` to check if an object is a
+    proxy.
+
+    .. code-block:: python
+
+        repr(user)  # <User admin>
+        isinstance(user, User)  # True
+        issubclass(type(user), LocalProxy)  # True
+
+    .. versionchanged:: 2.2.2
+        ``__wrapped__`` is set when wrapping an object, not only when
+        wrapping a function, to prevent doctest from failing.
+
+    .. versionchanged:: 2.2
+        Can proxy a ``ContextVar`` or ``LocalStack`` directly.
+
+    .. versionchanged:: 2.2
+        The ``name`` parameter can be used with any proxied object, not
+        only ``Local``.
+
+    .. versionchanged:: 2.2
+        Added the ``unbound_message`` parameter.
+
+    .. versionchanged:: 2.0
+        Updated proxied attributes and methods to reflect the current
+        data model.
+
+    .. versionchanged:: 0.6.1
+        The class can be instantiated with a callable.
+    """
+
+    __slots__ = ("__wrapped", "_get_current_object")
+
+    _get_current_object: t.Callable[[], T]
+    """Return the current object this proxy is bound to. If the proxy is
+    unbound, this raises a ``RuntimeError``.
+
+    This should be used if you need to pass the object to something that
+    doesn't understand the proxy. It can also be useful for performance
+    if you are accessing the object multiple times in a function, rather
+    than going through the proxy multiple times.
+    """
+
+    def __init__(
+        self,
+        local: ContextVar[T] | Local | LocalStack[T] | t.Callable[[], T],
+        name: str | None = None,
+        *,
+        unbound_message: str | None = None,
+    ) -> None:
+        if name is None:
+            get_name = _identity
+        else:
+            get_name = attrgetter(name)  # type: ignore[assignment]
+
+        if unbound_message is None:
+            unbound_message = "object is not bound"
+
+        if isinstance(local, Local):
+            if name is None:
+                raise TypeError("'name' is required when proxying a 'Local' object.")
+
+            def _get_current_object() -> T:
+                try:
+                    return get_name(local)  # type: ignore[return-value]
+                except AttributeError:
+                    raise RuntimeError(unbound_message) from None
+
+        elif isinstance(local, LocalStack):
+
+            def _get_current_object() -> T:
+                obj = local.top
+
+                if obj is None:
+                    raise RuntimeError(unbound_message)
+
+                return get_name(obj)
+
+        elif isinstance(local, ContextVar):
+
+            def _get_current_object() -> T:
+                try:
+                    obj = local.get()
+                except LookupError:
+                    raise RuntimeError(unbound_message) from None
+
+                return get_name(obj)
+
+        elif callable(local):
+
+            def _get_current_object() -> T:
+                return get_name(local())
+
+        else:
+            raise TypeError(f"Don't know how to proxy '{type(local)}'.")
+
+        object.__setattr__(self, "_LocalProxy__wrapped", local)
+        object.__setattr__(self, "_get_current_object", _get_current_object)
+
+    __doc__ = _ProxyLookup(  # type: ignore[assignment]
+        class_value=__doc__, fallback=lambda self: type(self).__doc__, is_attr=True
+    )
+    __wrapped__ = _ProxyLookup(
+        fallback=lambda self: self._LocalProxy__wrapped,  # type: ignore[attr-defined]
+        is_attr=True,
+    )
+    # __del__ should only delete the proxy
+    __repr__ = _ProxyLookup(  # type: ignore[assignment]
+        repr, fallback=lambda self: f"<{type(self).__name__} unbound>"
+    )
+    __str__ = _ProxyLookup(str)  # type: ignore[assignment]
+    __bytes__ = _ProxyLookup(bytes)
+    __format__ = _ProxyLookup()  # type: ignore[assignment]
+    __lt__ = _ProxyLookup(operator.lt)
+    __le__ = _ProxyLookup(operator.le)
+    __eq__ = _ProxyLookup(operator.eq)  # type: ignore[assignment]
+    __ne__ = _ProxyLookup(operator.ne)  # type: ignore[assignment]
+    __gt__ = _ProxyLookup(operator.gt)
+    __ge__ = _ProxyLookup(operator.ge)
+    __hash__ = _ProxyLookup(hash)  # type: ignore[assignment]
+    __bool__ = _ProxyLookup(bool, fallback=lambda self: False)
+    __getattr__ = _ProxyLookup(getattr)
+    # __getattribute__ triggered through __getattr__
+    __setattr__ = _ProxyLookup(setattr)  # type: ignore[assignment]
+    __delattr__ = _ProxyLookup(delattr)  # type: ignore[assignment]
+    __dir__ = _ProxyLookup(dir, fallback=lambda self: [])  # type: ignore[assignment]
+    # __get__ (proxying descriptor not supported)
+    # __set__ (descriptor)
+    # __delete__ (descriptor)
+    # __set_name__ (descriptor)
+    # __objclass__ (descriptor)
+    # __slots__ used by proxy itself
+    # __dict__ (__getattr__)
+    # __weakref__ (__getattr__)
+    # __init_subclass__ (proxying metaclass not supported)
+    # __prepare__ (metaclass)
+    __class__ = _ProxyLookup(fallback=lambda self: type(self), is_attr=True)  # type: ignore[assignment]
+    __instancecheck__ = _ProxyLookup(lambda self, other: isinstance(other, self))
+    __subclasscheck__ = _ProxyLookup(lambda self, other: issubclass(other, self))
+    # __class_getitem__ triggered through __getitem__
+    __call__ = _ProxyLookup(lambda self, *args, **kwargs: self(*args, **kwargs))
+    __len__ = _ProxyLookup(len)
+    __length_hint__ = _ProxyLookup(operator.length_hint)
+    __getitem__ = _ProxyLookup(operator.getitem)
+    __setitem__ = _ProxyLookup(operator.setitem)
+    __delitem__ = _ProxyLookup(operator.delitem)
+    # __missing__ triggered through __getitem__
+    __iter__ = _ProxyLookup(iter)
+    __next__ = _ProxyLookup(next)
+    __reversed__ = _ProxyLookup(reversed)
+    __contains__ = _ProxyLookup(operator.contains)
+    __add__ = _ProxyLookup(operator.add)
+    __sub__ = _ProxyLookup(operator.sub)
+    __mul__ = _ProxyLookup(operator.mul)
+    __matmul__ = _ProxyLookup(operator.matmul)
+    __truediv__ = _ProxyLookup(operator.truediv)
+    __floordiv__ = _ProxyLookup(operator.floordiv)
+    __mod__ = _ProxyLookup(operator.mod)
+    __divmod__ = _ProxyLookup(divmod)
+    __pow__ = _ProxyLookup(pow)
+    __lshift__ = _ProxyLookup(operator.lshift)
+    __rshift__ = _ProxyLookup(operator.rshift)
+    __and__ = _ProxyLookup(operator.and_)
+    __xor__ = _ProxyLookup(operator.xor)
+    __or__ = _ProxyLookup(operator.or_)
+    __radd__ = _ProxyLookup(_l_to_r_op(operator.add))
+    __rsub__ = _ProxyLookup(_l_to_r_op(operator.sub))
+    __rmul__ = _ProxyLookup(_l_to_r_op(operator.mul))
+    __rmatmul__ = _ProxyLookup(_l_to_r_op(operator.matmul))
+    __rtruediv__ = _ProxyLookup(_l_to_r_op(operator.truediv))
+    __rfloordiv__ = _ProxyLookup(_l_to_r_op(operator.floordiv))
+    __rmod__ = _ProxyLookup(_l_to_r_op(operator.mod))
+    __rdivmod__ = _ProxyLookup(_l_to_r_op(divmod))
+    __rpow__ = _ProxyLookup(_l_to_r_op(pow))
+    __rlshift__ = _ProxyLookup(_l_to_r_op(operator.lshift))
+    __rrshift__ = _ProxyLookup(_l_to_r_op(operator.rshift))
+    __rand__ = _ProxyLookup(_l_to_r_op(operator.and_))
+    __rxor__ = _ProxyLookup(_l_to_r_op(operator.xor))
+    __ror__ = _ProxyLookup(_l_to_r_op(operator.or_))
+    __iadd__ = _ProxyIOp(operator.iadd)
+    __isub__ = _ProxyIOp(operator.isub)
+    __imul__ = _ProxyIOp(operator.imul)
+    __imatmul__ = _ProxyIOp(operator.imatmul)
+    __itruediv__ = _ProxyIOp(operator.itruediv)
+    __ifloordiv__ = _ProxyIOp(operator.ifloordiv)
+    __imod__ = _ProxyIOp(operator.imod)
+    __ipow__ = _ProxyIOp(operator.ipow)
+    __ilshift__ = _ProxyIOp(operator.ilshift)
+    __irshift__ = _ProxyIOp(operator.irshift)
+    __iand__ = _ProxyIOp(operator.iand)
+    __ixor__ = _ProxyIOp(operator.ixor)
+    __ior__ = _ProxyIOp(operator.ior)
+    __neg__ = _ProxyLookup(operator.neg)
+    __pos__ = _ProxyLookup(operator.pos)
+    __abs__ = _ProxyLookup(abs)
+    __invert__ = _ProxyLookup(operator.invert)
+    __complex__ = _ProxyLookup(complex)
+    __int__ = _ProxyLookup(int)
+    __float__ = _ProxyLookup(float)
+    __index__ = _ProxyLookup(operator.index)
+    __round__ = _ProxyLookup(round)
+    __trunc__ = _ProxyLookup(math.trunc)
+    __floor__ = _ProxyLookup(math.floor)
+    __ceil__ = _ProxyLookup(math.ceil)
+    __enter__ = _ProxyLookup()
+    __exit__ = _ProxyLookup()
+    __await__ = _ProxyLookup()
+    __aiter__ = _ProxyLookup()
+    __anext__ = _ProxyLookup()
+    __aenter__ = _ProxyLookup()
+    __aexit__ = _ProxyLookup()
+    __copy__ = _ProxyLookup(copy.copy)
+    __deepcopy__ = _ProxyLookup(copy.deepcopy)
+    # __getnewargs_ex__ (pickle through proxy not supported)
+    # __getnewargs__ (pickle)
+    # __getstate__ (pickle)
+    # __setstate__ (pickle)
+    # __reduce__ (pickle)
+    # __reduce_ex__ (pickle)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..9ad72c51
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/dispatcher.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/dispatcher.cpython-312.pyc
new file mode 100644
index 00000000..1ae3d2ab
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/dispatcher.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/http_proxy.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/http_proxy.cpython-312.pyc
new file mode 100644
index 00000000..f3f2cfcf
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/http_proxy.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/lint.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/lint.cpython-312.pyc
new file mode 100644
index 00000000..e8b5a0ab
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/lint.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/profiler.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/profiler.cpython-312.pyc
new file mode 100644
index 00000000..8f8b6948
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/profiler.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/proxy_fix.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/proxy_fix.cpython-312.pyc
new file mode 100644
index 00000000..8f2dfa9d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/proxy_fix.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/shared_data.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/shared_data.cpython-312.pyc
new file mode 100644
index 00000000..f5156580
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/__pycache__/shared_data.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/dispatcher.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/dispatcher.py
new file mode 100644
index 00000000..e11bacc5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/dispatcher.py
@@ -0,0 +1,81 @@
+"""
+Application Dispatcher
+======================
+
+This middleware creates a single WSGI application that dispatches to
+multiple other WSGI applications mounted at different URL paths.
+
+A common example is writing a Single Page Application, where you have a
+backend API and a frontend written in JavaScript that does the routing
+in the browser rather than requesting different pages from the server.
+The frontend is a single HTML and JS file that should be served for any
+path besides "/api".
+
+This example dispatches to an API app under "/api", an admin app
+under "/admin", and an app that serves frontend files for all other
+requests::
+
+    app = DispatcherMiddleware(serve_frontend, {
+        '/api': api_app,
+        '/admin': admin_app,
+    })
+
+In production, you might instead handle this at the HTTP server level,
+serving files or proxying to application servers based on location. The
+API and admin apps would each be deployed with a separate WSGI server,
+and the static files would be served directly by the HTTP server.
+
+.. autoclass:: DispatcherMiddleware
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import typing as t
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class DispatcherMiddleware:
+    """Combine multiple applications as a single WSGI application.
+    Requests are dispatched to an application based on the path it is
+    mounted under.
+
+    :param app: The WSGI application to dispatch to if the request
+        doesn't match a mounted path.
+    :param mounts: Maps path prefixes to applications for dispatching.
+    """
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        mounts: dict[str, WSGIApplication] | None = None,
+    ) -> None:
+        self.app = app
+        self.mounts = mounts or {}
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        script = environ.get("PATH_INFO", "")
+        path_info = ""
+
+        while "/" in script:
+            if script in self.mounts:
+                app = self.mounts[script]
+                break
+
+            script, last_item = script.rsplit("/", 1)
+            path_info = f"/{last_item}{path_info}"
+        else:
+            app = self.mounts.get(script, self.app)
+
+        original_script_name = environ.get("SCRIPT_NAME", "")
+        environ["SCRIPT_NAME"] = original_script_name + script
+        environ["PATH_INFO"] = path_info
+        return app(environ, start_response)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/http_proxy.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/http_proxy.py
new file mode 100644
index 00000000..5e239156
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/http_proxy.py
@@ -0,0 +1,236 @@
+"""
+Basic HTTP Proxy
+================
+
+.. autoclass:: ProxyMiddleware
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import typing as t
+from http import client
+from urllib.parse import quote
+from urllib.parse import urlsplit
+
+from ..datastructures import EnvironHeaders
+from ..http import is_hop_by_hop_header
+from ..wsgi import get_input_stream
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class ProxyMiddleware:
+    """Proxy requests under a path to an external server, routing other
+    requests to the app.
+
+    This middleware can only proxy HTTP requests, as HTTP is the only
+    protocol handled by the WSGI server. Other protocols, such as
+    WebSocket requests, cannot be proxied at this layer. This should
+    only be used for development, in production a real proxy server
+    should be used.
+
+    The middleware takes a dict mapping a path prefix to a dict
+    describing the host to be proxied to::
+
+        app = ProxyMiddleware(app, {
+            "/static/": {
+                "target": "http://127.0.0.1:5001/",
+            }
+        })
+
+    Each host has the following options:
+
+    ``target``:
+        The target URL to dispatch to. This is required.
+    ``remove_prefix``:
+        Whether to remove the prefix from the URL before dispatching it
+        to the target. The default is ``False``.
+    ``host``:
+        ``"<auto>"`` (default):
+            The host header is automatically rewritten to the URL of the
+            target.
+        ``None``:
+            The host header is unmodified from the client request.
+        Any other value:
+            The host header is overwritten with the value.
+    ``headers``:
+        A dictionary of headers to be sent with the request to the
+        target. The default is ``{}``.
+    ``ssl_context``:
+        A :class:`ssl.SSLContext` defining how to verify requests if the
+        target is HTTPS. The default is ``None``.
+
+    In the example above, everything under ``"/static/"`` is proxied to
+    the server on port 5001. The host header is rewritten to the target,
+    and the ``"/static/"`` prefix is removed from the URLs.
+
+    :param app: The WSGI application to wrap.
+    :param targets: Proxy target configurations. See description above.
+    :param chunk_size: Size of chunks to read from input stream and
+        write to target.
+    :param timeout: Seconds before an operation to a target fails.
+
+    .. versionadded:: 0.14
+    """
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        targets: t.Mapping[str, dict[str, t.Any]],
+        chunk_size: int = 2 << 13,
+        timeout: int = 10,
+    ) -> None:
+        def _set_defaults(opts: dict[str, t.Any]) -> dict[str, t.Any]:
+            opts.setdefault("remove_prefix", False)
+            opts.setdefault("host", "<auto>")
+            opts.setdefault("headers", {})
+            opts.setdefault("ssl_context", None)
+            return opts
+
+        self.app = app
+        self.targets = {
+            f"/{k.strip('/')}/": _set_defaults(v) for k, v in targets.items()
+        }
+        self.chunk_size = chunk_size
+        self.timeout = timeout
+
+    def proxy_to(
+        self, opts: dict[str, t.Any], path: str, prefix: str
+    ) -> WSGIApplication:
+        target = urlsplit(opts["target"])
+        # socket can handle unicode host, but header must be ascii
+        host = target.hostname.encode("idna").decode("ascii")
+
+        def application(
+            environ: WSGIEnvironment, start_response: StartResponse
+        ) -> t.Iterable[bytes]:
+            headers = list(EnvironHeaders(environ).items())
+            headers[:] = [
+                (k, v)
+                for k, v in headers
+                if not is_hop_by_hop_header(k)
+                and k.lower() not in ("content-length", "host")
+            ]
+            headers.append(("Connection", "close"))
+
+            if opts["host"] == "<auto>":
+                headers.append(("Host", host))
+            elif opts["host"] is None:
+                headers.append(("Host", environ["HTTP_HOST"]))
+            else:
+                headers.append(("Host", opts["host"]))
+
+            headers.extend(opts["headers"].items())
+            remote_path = path
+
+            if opts["remove_prefix"]:
+                remote_path = remote_path[len(prefix) :].lstrip("/")
+                remote_path = f"{target.path.rstrip('/')}/{remote_path}"
+
+            content_length = environ.get("CONTENT_LENGTH")
+            chunked = False
+
+            if content_length not in ("", None):
+                headers.append(("Content-Length", content_length))  # type: ignore
+            elif content_length is not None:
+                headers.append(("Transfer-Encoding", "chunked"))
+                chunked = True
+
+            try:
+                if target.scheme == "http":
+                    con = client.HTTPConnection(
+                        host, target.port or 80, timeout=self.timeout
+                    )
+                elif target.scheme == "https":
+                    con = client.HTTPSConnection(
+                        host,
+                        target.port or 443,
+                        timeout=self.timeout,
+                        context=opts["ssl_context"],
+                    )
+                else:
+                    raise RuntimeError(
+                        "Target scheme must be 'http' or 'https', got"
+                        f" {target.scheme!r}."
+                    )
+
+                con.connect()
+                # safe = https://url.spec.whatwg.org/#url-path-segment-string
+                # as well as percent for things that are already quoted
+                remote_url = quote(remote_path, safe="!$&'()*+,/:;=@%")
+                querystring = environ["QUERY_STRING"]
+
+                if querystring:
+                    remote_url = f"{remote_url}?{querystring}"
+
+                con.putrequest(environ["REQUEST_METHOD"], remote_url, skip_host=True)
+
+                for k, v in headers:
+                    if k.lower() == "connection":
+                        v = "close"
+
+                    con.putheader(k, v)
+
+                con.endheaders()
+                stream = get_input_stream(environ)
+
+                while True:
+                    data = stream.read(self.chunk_size)
+
+                    if not data:
+                        break
+
+                    if chunked:
+                        con.send(b"%x\r\n%s\r\n" % (len(data), data))
+                    else:
+                        con.send(data)
+
+                resp = con.getresponse()
+            except OSError:
+                from ..exceptions import BadGateway
+
+                return BadGateway()(environ, start_response)
+
+            start_response(
+                f"{resp.status} {resp.reason}",
+                [
+                    (k.title(), v)
+                    for k, v in resp.getheaders()
+                    if not is_hop_by_hop_header(k)
+                ],
+            )
+
+            def read() -> t.Iterator[bytes]:
+                while True:
+                    try:
+                        data = resp.read(self.chunk_size)
+                    except OSError:
+                        break
+
+                    if not data:
+                        break
+
+                    yield data
+
+            return read()
+
+        return application
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        path = environ["PATH_INFO"]
+        app = self.app
+
+        for prefix, opts in self.targets.items():
+            if path.startswith(prefix):
+                app = self.proxy_to(opts, path, prefix)
+                break
+
+        return app(environ, start_response)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/lint.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/lint.py
new file mode 100644
index 00000000..3714271b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/lint.py
@@ -0,0 +1,439 @@
+"""
+WSGI Protocol Linter
+====================
+
+This module provides a middleware that performs sanity checks on the
+behavior of the WSGI server and application. It checks that the
+:pep:`3333` WSGI spec is properly implemented. It also warns on some
+common HTTP errors such as non-empty responses for 304 status codes.
+
+.. autoclass:: LintMiddleware
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import typing as t
+from types import TracebackType
+from urllib.parse import urlparse
+from warnings import warn
+
+from ..datastructures import Headers
+from ..http import is_entity_header
+from ..wsgi import FileWrapper
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class WSGIWarning(Warning):
+    """Warning class for WSGI warnings."""
+
+
+class HTTPWarning(Warning):
+    """Warning class for HTTP warnings."""
+
+
+def check_type(context: str, obj: object, need: type = str) -> None:
+    if type(obj) is not need:
+        warn(
+            f"{context!r} requires {need.__name__!r}, got {type(obj).__name__!r}.",
+            WSGIWarning,
+            stacklevel=3,
+        )
+
+
+class InputStream:
+    def __init__(self, stream: t.IO[bytes]) -> None:
+        self._stream = stream
+
+    def read(self, *args: t.Any) -> bytes:
+        if len(args) == 0:
+            warn(
+                "WSGI does not guarantee an EOF marker on the input stream, thus making"
+                " calls to 'wsgi.input.read()' unsafe. Conforming servers may never"
+                " return from this call.",
+                WSGIWarning,
+                stacklevel=2,
+            )
+        elif len(args) != 1:
+            warn(
+                "Too many parameters passed to 'wsgi.input.read()'.",
+                WSGIWarning,
+                stacklevel=2,
+            )
+        return self._stream.read(*args)
+
+    def readline(self, *args: t.Any) -> bytes:
+        if len(args) == 0:
+            warn(
+                "Calls to 'wsgi.input.readline()' without arguments are unsafe. Use"
+                " 'wsgi.input.read()' instead.",
+                WSGIWarning,
+                stacklevel=2,
+            )
+        elif len(args) == 1:
+            warn(
+                "'wsgi.input.readline()' was called with a size hint. WSGI does not"
+                " support this, although it's available on all major servers.",
+                WSGIWarning,
+                stacklevel=2,
+            )
+        else:
+            raise TypeError("Too many arguments passed to 'wsgi.input.readline()'.")
+        return self._stream.readline(*args)
+
+    def __iter__(self) -> t.Iterator[bytes]:
+        try:
+            return iter(self._stream)
+        except TypeError:
+            warn("'wsgi.input' is not iterable.", WSGIWarning, stacklevel=2)
+            return iter(())
+
+    def close(self) -> None:
+        warn("The application closed the input stream!", WSGIWarning, stacklevel=2)
+        self._stream.close()
+
+
+class ErrorStream:
+    def __init__(self, stream: t.IO[str]) -> None:
+        self._stream = stream
+
+    def write(self, s: str) -> None:
+        check_type("wsgi.error.write()", s, str)
+        self._stream.write(s)
+
+    def flush(self) -> None:
+        self._stream.flush()
+
+    def writelines(self, seq: t.Iterable[str]) -> None:
+        for line in seq:
+            self.write(line)
+
+    def close(self) -> None:
+        warn("The application closed the error stream!", WSGIWarning, stacklevel=2)
+        self._stream.close()
+
+
+class GuardedWrite:
+    def __init__(self, write: t.Callable[[bytes], object], chunks: list[int]) -> None:
+        self._write = write
+        self._chunks = chunks
+
+    def __call__(self, s: bytes) -> None:
+        check_type("write()", s, bytes)
+        self._write(s)
+        self._chunks.append(len(s))
+
+
+class GuardedIterator:
+    def __init__(
+        self,
+        iterator: t.Iterable[bytes],
+        headers_set: tuple[int, Headers],
+        chunks: list[int],
+    ) -> None:
+        self._iterator = iterator
+        self._next = iter(iterator).__next__
+        self.closed = False
+        self.headers_set = headers_set
+        self.chunks = chunks
+
+    def __iter__(self) -> GuardedIterator:
+        return self
+
+    def __next__(self) -> bytes:
+        if self.closed:
+            warn("Iterated over closed 'app_iter'.", WSGIWarning, stacklevel=2)
+
+        rv = self._next()
+
+        if not self.headers_set:
+            warn(
+                "The application returned before it started the response.",
+                WSGIWarning,
+                stacklevel=2,
+            )
+
+        check_type("application iterator items", rv, bytes)
+        self.chunks.append(len(rv))
+        return rv
+
+    def close(self) -> None:
+        self.closed = True
+
+        if hasattr(self._iterator, "close"):
+            self._iterator.close()
+
+        if self.headers_set:
+            status_code, headers = self.headers_set
+            bytes_sent = sum(self.chunks)
+            content_length = headers.get("content-length", type=int)
+
+            if status_code == 304:
+                for key, _value in headers:
+                    key = key.lower()
+                    if key not in ("expires", "content-location") and is_entity_header(
+                        key
+                    ):
+                        warn(
+                            f"Entity header {key!r} found in 304 response.",
+                            HTTPWarning,
+                            stacklevel=2,
+                        )
+                if bytes_sent:
+                    warn(
+                        "304 responses must not have a body.",
+                        HTTPWarning,
+                        stacklevel=2,
+                    )
+            elif 100 <= status_code < 200 or status_code == 204:
+                if content_length != 0:
+                    warn(
+                        f"{status_code} responses must have an empty content length.",
+                        HTTPWarning,
+                        stacklevel=2,
+                    )
+                if bytes_sent:
+                    warn(
+                        f"{status_code} responses must not have a body.",
+                        HTTPWarning,
+                        stacklevel=2,
+                    )
+            elif content_length is not None and content_length != bytes_sent:
+                warn(
+                    "Content-Length and the number of bytes sent to the"
+                    " client do not match.",
+                    WSGIWarning,
+                    stacklevel=2,
+                )
+
+    def __del__(self) -> None:
+        if not self.closed:
+            try:
+                warn(
+                    "Iterator was garbage collected before it was closed.",
+                    WSGIWarning,
+                    stacklevel=2,
+                )
+            except Exception:
+                pass
+
+
+class LintMiddleware:
+    """Warns about common errors in the WSGI and HTTP behavior of the
+    server and wrapped application. Some of the issues it checks are:
+
+    -   invalid status codes
+    -   non-bytes sent to the WSGI server
+    -   strings returned from the WSGI application
+    -   non-empty conditional responses
+    -   unquoted etags
+    -   relative URLs in the Location header
+    -   unsafe calls to wsgi.input
+    -   unclosed iterators
+
+    Error information is emitted using the :mod:`warnings` module.
+
+    :param app: The WSGI application to wrap.
+
+    .. code-block:: python
+
+        from werkzeug.middleware.lint import LintMiddleware
+        app = LintMiddleware(app)
+    """
+
+    def __init__(self, app: WSGIApplication) -> None:
+        self.app = app
+
+    def check_environ(self, environ: WSGIEnvironment) -> None:
+        if type(environ) is not dict:  # noqa: E721
+            warn(
+                "WSGI environment is not a standard Python dict.",
+                WSGIWarning,
+                stacklevel=4,
+            )
+        for key in (
+            "REQUEST_METHOD",
+            "SERVER_NAME",
+            "SERVER_PORT",
+            "wsgi.version",
+            "wsgi.input",
+            "wsgi.errors",
+            "wsgi.multithread",
+            "wsgi.multiprocess",
+            "wsgi.run_once",
+        ):
+            if key not in environ:
+                warn(
+                    f"Required environment key {key!r} not found",
+                    WSGIWarning,
+                    stacklevel=3,
+                )
+        if environ["wsgi.version"] != (1, 0):
+            warn("Environ is not a WSGI 1.0 environ.", WSGIWarning, stacklevel=3)
+
+        script_name = environ.get("SCRIPT_NAME", "")
+        path_info = environ.get("PATH_INFO", "")
+
+        if script_name and script_name[0] != "/":
+            warn(
+                f"'SCRIPT_NAME' does not start with a slash: {script_name!r}",
+                WSGIWarning,
+                stacklevel=3,
+            )
+
+        if path_info and path_info[0] != "/":
+            warn(
+                f"'PATH_INFO' does not start with a slash: {path_info!r}",
+                WSGIWarning,
+                stacklevel=3,
+            )
+
+    def check_start_response(
+        self,
+        status: str,
+        headers: list[tuple[str, str]],
+        exc_info: None | (tuple[type[BaseException], BaseException, TracebackType]),
+    ) -> tuple[int, Headers]:
+        check_type("status", status, str)
+        status_code_str = status.split(None, 1)[0]
+
+        if len(status_code_str) != 3 or not status_code_str.isdecimal():
+            warn("Status code must be three digits.", WSGIWarning, stacklevel=3)
+
+        if len(status) < 4 or status[3] != " ":
+            warn(
+                f"Invalid value for status {status!r}. Valid status strings are three"
+                " digits, a space and a status explanation.",
+                WSGIWarning,
+                stacklevel=3,
+            )
+
+        status_code = int(status_code_str)
+
+        if status_code < 100:
+            warn("Status code < 100 detected.", WSGIWarning, stacklevel=3)
+
+        if type(headers) is not list:  # noqa: E721
+            warn("Header list is not a list.", WSGIWarning, stacklevel=3)
+
+        for item in headers:
+            if type(item) is not tuple or len(item) != 2:
+                warn("Header items must be 2-item tuples.", WSGIWarning, stacklevel=3)
+            name, value = item
+            if type(name) is not str or type(value) is not str:  # noqa: E721
+                warn(
+                    "Header keys and values must be strings.", WSGIWarning, stacklevel=3
+                )
+            if name.lower() == "status":
+                warn(
+                    "The status header is not supported due to"
+                    " conflicts with the CGI spec.",
+                    WSGIWarning,
+                    stacklevel=3,
+                )
+
+        if exc_info is not None and not isinstance(exc_info, tuple):
+            warn("Invalid value for exc_info.", WSGIWarning, stacklevel=3)
+
+        headers_obj = Headers(headers)
+        self.check_headers(headers_obj)
+
+        return status_code, headers_obj
+
+    def check_headers(self, headers: Headers) -> None:
+        etag = headers.get("etag")
+
+        if etag is not None:
+            if etag.startswith(("W/", "w/")):
+                if etag.startswith("w/"):
+                    warn(
+                        "Weak etag indicator should be upper case.",
+                        HTTPWarning,
+                        stacklevel=4,
+                    )
+
+                etag = etag[2:]
+
+            if not (etag[:1] == etag[-1:] == '"'):
+                warn("Unquoted etag emitted.", HTTPWarning, stacklevel=4)
+
+        location = headers.get("location")
+
+        if location is not None:
+            if not urlparse(location).netloc:
+                warn(
+                    "Absolute URLs required for location header.",
+                    HTTPWarning,
+                    stacklevel=4,
+                )
+
+    def check_iterator(self, app_iter: t.Iterable[bytes]) -> None:
+        if isinstance(app_iter, str):
+            warn(
+                "The application returned a string. The response will send one"
+                " character at a time to the client, which will kill performance."
+                " Return a list or iterable instead.",
+                WSGIWarning,
+                stacklevel=3,
+            )
+
+    def __call__(self, *args: t.Any, **kwargs: t.Any) -> t.Iterable[bytes]:
+        if len(args) != 2:
+            warn("A WSGI app takes two arguments.", WSGIWarning, stacklevel=2)
+
+        if kwargs:
+            warn(
+                "A WSGI app does not take keyword arguments.", WSGIWarning, stacklevel=2
+            )
+
+        environ: WSGIEnvironment = args[0]
+        start_response: StartResponse = args[1]
+
+        self.check_environ(environ)
+        environ["wsgi.input"] = InputStream(environ["wsgi.input"])
+        environ["wsgi.errors"] = ErrorStream(environ["wsgi.errors"])
+
+        # Hook our own file wrapper in so that applications will always
+        # iterate to the end and we can check the content length.
+        environ["wsgi.file_wrapper"] = FileWrapper
+
+        headers_set: list[t.Any] = []
+        chunks: list[int] = []
+
+        def checking_start_response(
+            *args: t.Any, **kwargs: t.Any
+        ) -> t.Callable[[bytes], None]:
+            if len(args) not in {2, 3}:
+                warn(
+                    f"Invalid number of arguments: {len(args)}, expected 2 or 3.",
+                    WSGIWarning,
+                    stacklevel=2,
+                )
+
+            if kwargs:
+                warn(
+                    "'start_response' does not take keyword arguments.",
+                    WSGIWarning,
+                    stacklevel=2,
+                )
+
+            status: str = args[0]
+            headers: list[tuple[str, str]] = args[1]
+            exc_info: (
+                None | (tuple[type[BaseException], BaseException, TracebackType])
+            ) = args[2] if len(args) == 3 else None
+
+            headers_set[:] = self.check_start_response(status, headers, exc_info)
+            return GuardedWrite(start_response(status, headers, exc_info), chunks)
+
+        app_iter = self.app(environ, t.cast("StartResponse", checking_start_response))
+        self.check_iterator(app_iter)
+        return GuardedIterator(
+            app_iter, t.cast(tuple[int, Headers], headers_set), chunks
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/profiler.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/profiler.py
new file mode 100644
index 00000000..112b8777
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/profiler.py
@@ -0,0 +1,155 @@
+"""
+Application Profiler
+====================
+
+This module provides a middleware that profiles each request with the
+:mod:`cProfile` module. This can help identify bottlenecks in your code
+that may be slowing down your application.
+
+.. autoclass:: ProfilerMiddleware
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import os.path
+import sys
+import time
+import typing as t
+from pstats import Stats
+
+try:
+    from cProfile import Profile
+except ImportError:
+    from profile import Profile  # type: ignore
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class ProfilerMiddleware:
+    """Wrap a WSGI application and profile the execution of each
+    request. Responses are buffered so that timings are more exact.
+
+    If ``stream`` is given, :class:`pstats.Stats` are written to it
+    after each request. If ``profile_dir`` is given, :mod:`cProfile`
+    data files are saved to that directory, one file per request.
+
+    The filename can be customized by passing ``filename_format``. If
+    it is a string, it will be formatted using :meth:`str.format` with
+    the following fields available:
+
+    -   ``{method}`` - The request method; GET, POST, etc.
+    -   ``{path}`` - The request path or 'root' should one not exist.
+    -   ``{elapsed}`` - The elapsed time of the request in milliseconds.
+    -   ``{time}`` - The time of the request.
+
+    If it is a callable, it will be called with the WSGI ``environ`` and
+    be expected to return a filename string. The ``environ`` dictionary
+    will also have the ``"werkzeug.profiler"`` key populated with a
+    dictionary containing the following fields (more may be added in the
+    future):
+    -   ``{elapsed}`` - The elapsed time of the request in milliseconds.
+    -   ``{time}`` - The time of the request.
+
+    :param app: The WSGI application to wrap.
+    :param stream: Write stats to this stream. Disable with ``None``.
+    :param sort_by: A tuple of columns to sort stats by. See
+        :meth:`pstats.Stats.sort_stats`.
+    :param restrictions: A tuple of restrictions to filter stats by. See
+        :meth:`pstats.Stats.print_stats`.
+    :param profile_dir: Save profile data files to this directory.
+    :param filename_format: Format string for profile data file names,
+        or a callable returning a name. See explanation above.
+
+    .. code-block:: python
+
+        from werkzeug.middleware.profiler import ProfilerMiddleware
+        app = ProfilerMiddleware(app)
+
+    .. versionchanged:: 3.0
+        Added the ``"werkzeug.profiler"`` key to the ``filename_format(environ)``
+        parameter with the  ``elapsed`` and ``time`` fields.
+
+    .. versionchanged:: 0.15
+        Stats are written even if ``profile_dir`` is given, and can be
+        disable by passing ``stream=None``.
+
+    .. versionadded:: 0.15
+        Added ``filename_format``.
+
+    .. versionadded:: 0.9
+        Added ``restrictions`` and ``profile_dir``.
+    """
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        stream: t.IO[str] | None = sys.stdout,
+        sort_by: t.Iterable[str] = ("time", "calls"),
+        restrictions: t.Iterable[str | int | float] = (),
+        profile_dir: str | None = None,
+        filename_format: str = "{method}.{path}.{elapsed:.0f}ms.{time:.0f}.prof",
+    ) -> None:
+        self._app = app
+        self._stream = stream
+        self._sort_by = sort_by
+        self._restrictions = restrictions
+        self._profile_dir = profile_dir
+        self._filename_format = filename_format
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        response_body: list[bytes] = []
+
+        def catching_start_response(status, headers, exc_info=None):  # type: ignore
+            start_response(status, headers, exc_info)
+            return response_body.append
+
+        def runapp() -> None:
+            app_iter = self._app(
+                environ, t.cast("StartResponse", catching_start_response)
+            )
+            response_body.extend(app_iter)
+
+            if hasattr(app_iter, "close"):
+                app_iter.close()
+
+        profile = Profile()
+        start = time.time()
+        profile.runcall(runapp)
+        body = b"".join(response_body)
+        elapsed = time.time() - start
+
+        if self._profile_dir is not None:
+            if callable(self._filename_format):
+                environ["werkzeug.profiler"] = {
+                    "elapsed": elapsed * 1000.0,
+                    "time": time.time(),
+                }
+                filename = self._filename_format(environ)
+            else:
+                filename = self._filename_format.format(
+                    method=environ["REQUEST_METHOD"],
+                    path=environ["PATH_INFO"].strip("/").replace("/", ".") or "root",
+                    elapsed=elapsed * 1000.0,
+                    time=time.time(),
+                )
+            filename = os.path.join(self._profile_dir, filename)
+            profile.dump_stats(filename)
+
+        if self._stream is not None:
+            stats = Stats(profile, stream=self._stream)
+            stats.sort_stats(*self._sort_by)
+            print("-" * 80, file=self._stream)
+            path_info = environ.get("PATH_INFO", "")
+            print(f"PATH: {path_info!r}", file=self._stream)
+            stats.print_stats(*self._restrictions)
+            print(f"{'-' * 80}\n", file=self._stream)
+
+        return [body]
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/proxy_fix.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/proxy_fix.py
new file mode 100644
index 00000000..cbf4e0ba
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/proxy_fix.py
@@ -0,0 +1,183 @@
+"""
+X-Forwarded-For Proxy Fix
+=========================
+
+This module provides a middleware that adjusts the WSGI environ based on
+``X-Forwarded-`` headers that proxies in front of an application may
+set.
+
+When an application is running behind a proxy server, WSGI may see the
+request as coming from that server rather than the real client. Proxies
+set various headers to track where the request actually came from.
+
+This middleware should only be used if the application is actually
+behind such a proxy, and should be configured with the number of proxies
+that are chained in front of it. Not all proxies set all the headers.
+Since incoming headers can be faked, you must set how many proxies are
+setting each header so the middleware knows what to trust.
+
+.. autoclass:: ProxyFix
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import typing as t
+
+from ..http import parse_list_header
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class ProxyFix:
+    """Adjust the WSGI environ based on ``X-Forwarded-`` that proxies in
+    front of the application may set.
+
+    -   ``X-Forwarded-For`` sets ``REMOTE_ADDR``.
+    -   ``X-Forwarded-Proto`` sets ``wsgi.url_scheme``.
+    -   ``X-Forwarded-Host`` sets ``HTTP_HOST``, ``SERVER_NAME``, and
+        ``SERVER_PORT``.
+    -   ``X-Forwarded-Port`` sets ``HTTP_HOST`` and ``SERVER_PORT``.
+    -   ``X-Forwarded-Prefix`` sets ``SCRIPT_NAME``.
+
+    You must tell the middleware how many proxies set each header so it
+    knows what values to trust. It is a security issue to trust values
+    that came from the client rather than a proxy.
+
+    The original values of the headers are stored in the WSGI
+    environ as ``werkzeug.proxy_fix.orig``, a dict.
+
+    :param app: The WSGI application to wrap.
+    :param x_for: Number of values to trust for ``X-Forwarded-For``.
+    :param x_proto: Number of values to trust for ``X-Forwarded-Proto``.
+    :param x_host: Number of values to trust for ``X-Forwarded-Host``.
+    :param x_port: Number of values to trust for ``X-Forwarded-Port``.
+    :param x_prefix: Number of values to trust for
+        ``X-Forwarded-Prefix``.
+
+    .. code-block:: python
+
+        from werkzeug.middleware.proxy_fix import ProxyFix
+        # App is behind one proxy that sets the -For and -Host headers.
+        app = ProxyFix(app, x_for=1, x_host=1)
+
+    .. versionchanged:: 1.0
+        The ``num_proxies`` argument and attribute; the ``get_remote_addr`` method; and
+        the environ keys ``orig_remote_addr``, ``orig_wsgi_url_scheme``, and
+        ``orig_http_host`` were removed.
+
+    .. versionchanged:: 0.15
+        All headers support multiple values. Each header is configured with a separate
+        number of trusted proxies.
+
+    .. versionchanged:: 0.15
+        Original WSGI environ values are stored in the ``werkzeug.proxy_fix.orig`` dict.
+
+    .. versionchanged:: 0.15
+        Support ``X-Forwarded-Port`` and ``X-Forwarded-Prefix``.
+
+    .. versionchanged:: 0.15
+        ``X-Forwarded-Host`` and ``X-Forwarded-Port`` modify
+        ``SERVER_NAME`` and ``SERVER_PORT``.
+    """
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        x_for: int = 1,
+        x_proto: int = 1,
+        x_host: int = 0,
+        x_port: int = 0,
+        x_prefix: int = 0,
+    ) -> None:
+        self.app = app
+        self.x_for = x_for
+        self.x_proto = x_proto
+        self.x_host = x_host
+        self.x_port = x_port
+        self.x_prefix = x_prefix
+
+    def _get_real_value(self, trusted: int, value: str | None) -> str | None:
+        """Get the real value from a list header based on the configured
+        number of trusted proxies.
+
+        :param trusted: Number of values to trust in the header.
+        :param value: Comma separated list header value to parse.
+        :return: The real value, or ``None`` if there are fewer values
+            than the number of trusted proxies.
+
+        .. versionchanged:: 1.0
+            Renamed from ``_get_trusted_comma``.
+
+        .. versionadded:: 0.15
+        """
+        if not (trusted and value):
+            return None
+        values = parse_list_header(value)
+        if len(values) >= trusted:
+            return values[-trusted]
+        return None
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        """Modify the WSGI environ based on the various ``Forwarded``
+        headers before calling the wrapped application. Store the
+        original environ values in ``werkzeug.proxy_fix.orig_{key}``.
+        """
+        environ_get = environ.get
+        orig_remote_addr = environ_get("REMOTE_ADDR")
+        orig_wsgi_url_scheme = environ_get("wsgi.url_scheme")
+        orig_http_host = environ_get("HTTP_HOST")
+        environ.update(
+            {
+                "werkzeug.proxy_fix.orig": {
+                    "REMOTE_ADDR": orig_remote_addr,
+                    "wsgi.url_scheme": orig_wsgi_url_scheme,
+                    "HTTP_HOST": orig_http_host,
+                    "SERVER_NAME": environ_get("SERVER_NAME"),
+                    "SERVER_PORT": environ_get("SERVER_PORT"),
+                    "SCRIPT_NAME": environ_get("SCRIPT_NAME"),
+                }
+            }
+        )
+
+        x_for = self._get_real_value(self.x_for, environ_get("HTTP_X_FORWARDED_FOR"))
+        if x_for:
+            environ["REMOTE_ADDR"] = x_for
+
+        x_proto = self._get_real_value(
+            self.x_proto, environ_get("HTTP_X_FORWARDED_PROTO")
+        )
+        if x_proto:
+            environ["wsgi.url_scheme"] = x_proto
+
+        x_host = self._get_real_value(self.x_host, environ_get("HTTP_X_FORWARDED_HOST"))
+        if x_host:
+            environ["HTTP_HOST"] = environ["SERVER_NAME"] = x_host
+            # "]" to check for IPv6 address without port
+            if ":" in x_host and not x_host.endswith("]"):
+                environ["SERVER_NAME"], environ["SERVER_PORT"] = x_host.rsplit(":", 1)
+
+        x_port = self._get_real_value(self.x_port, environ_get("HTTP_X_FORWARDED_PORT"))
+        if x_port:
+            host = environ.get("HTTP_HOST")
+            if host:
+                # "]" to check for IPv6 address without port
+                if ":" in host and not host.endswith("]"):
+                    host = host.rsplit(":", 1)[0]
+                environ["HTTP_HOST"] = f"{host}:{x_port}"
+            environ["SERVER_PORT"] = x_port
+
+        x_prefix = self._get_real_value(
+            self.x_prefix, environ_get("HTTP_X_FORWARDED_PREFIX")
+        )
+        if x_prefix:
+            environ["SCRIPT_NAME"] = x_prefix
+
+        return self.app(environ, start_response)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/shared_data.py b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/shared_data.py
new file mode 100644
index 00000000..c7c06df5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/middleware/shared_data.py
@@ -0,0 +1,283 @@
+"""
+Serve Shared Static Files
+=========================
+
+.. autoclass:: SharedDataMiddleware
+    :members: is_allowed
+
+:copyright: 2007 Pallets
+:license: BSD-3-Clause
+"""
+
+from __future__ import annotations
+
+import collections.abc as cabc
+import importlib.util
+import mimetypes
+import os
+import posixpath
+import typing as t
+from datetime import datetime
+from datetime import timezone
+from io import BytesIO
+from time import time
+from zlib import adler32
+
+from ..http import http_date
+from ..http import is_resource_modified
+from ..security import safe_join
+from ..utils import get_content_type
+from ..wsgi import get_path_info
+from ..wsgi import wrap_file
+
+_TOpener = t.Callable[[], tuple[t.IO[bytes], datetime, int]]
+_TLoader = t.Callable[[t.Optional[str]], tuple[t.Optional[str], t.Optional[_TOpener]]]
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class SharedDataMiddleware:
+    """A WSGI middleware which provides static content for development
+    environments or simple server setups. Its usage is quite simple::
+
+        import os
+        from werkzeug.middleware.shared_data import SharedDataMiddleware
+
+        app = SharedDataMiddleware(app, {
+            '/shared': os.path.join(os.path.dirname(__file__), 'shared')
+        })
+
+    The contents of the folder ``./shared`` will now be available on
+    ``http://example.com/shared/``.  This is pretty useful during development
+    because a standalone media server is not required. Files can also be
+    mounted on the root folder and still continue to use the application because
+    the shared data middleware forwards all unhandled requests to the
+    application, even if the requests are below one of the shared folders.
+
+    If `pkg_resources` is available you can also tell the middleware to serve
+    files from package data::
+
+        app = SharedDataMiddleware(app, {
+            '/static': ('myapplication', 'static')
+        })
+
+    This will then serve the ``static`` folder in the `myapplication`
+    Python package.
+
+    The optional `disallow` parameter can be a list of :func:`~fnmatch.fnmatch`
+    rules for files that are not accessible from the web.  If `cache` is set to
+    `False` no caching headers are sent.
+
+    Currently the middleware does not support non-ASCII filenames. If the
+    encoding on the file system happens to match the encoding of the URI it may
+    work but this could also be by accident. We strongly suggest using ASCII
+    only file names for static files.
+
+    The middleware will guess the mimetype using the Python `mimetype`
+    module.  If it's unable to figure out the charset it will fall back
+    to `fallback_mimetype`.
+
+    :param app: the application to wrap.  If you don't want to wrap an
+                application you can pass it :exc:`NotFound`.
+    :param exports: a list or dict of exported files and folders.
+    :param disallow: a list of :func:`~fnmatch.fnmatch` rules.
+    :param cache: enable or disable caching headers.
+    :param cache_timeout: the cache timeout in seconds for the headers.
+    :param fallback_mimetype: The fallback mimetype for unknown files.
+
+    .. versionchanged:: 1.0
+        The default ``fallback_mimetype`` is
+        ``application/octet-stream``. If a filename looks like a text
+        mimetype, the ``utf-8`` charset is added to it.
+
+    .. versionadded:: 0.6
+        Added ``fallback_mimetype``.
+
+    .. versionchanged:: 0.5
+        Added ``cache_timeout``.
+    """
+
+    def __init__(
+        self,
+        app: WSGIApplication,
+        exports: (
+            cabc.Mapping[str, str | tuple[str, str]]
+            | t.Iterable[tuple[str, str | tuple[str, str]]]
+        ),
+        disallow: None = None,
+        cache: bool = True,
+        cache_timeout: int = 60 * 60 * 12,
+        fallback_mimetype: str = "application/octet-stream",
+    ) -> None:
+        self.app = app
+        self.exports: list[tuple[str, _TLoader]] = []
+        self.cache = cache
+        self.cache_timeout = cache_timeout
+
+        if isinstance(exports, cabc.Mapping):
+            exports = exports.items()
+
+        for key, value in exports:
+            if isinstance(value, tuple):
+                loader = self.get_package_loader(*value)
+            elif isinstance(value, str):
+                if os.path.isfile(value):
+                    loader = self.get_file_loader(value)
+                else:
+                    loader = self.get_directory_loader(value)
+            else:
+                raise TypeError(f"unknown def {value!r}")
+
+            self.exports.append((key, loader))
+
+        if disallow is not None:
+            from fnmatch import fnmatch
+
+            self.is_allowed = lambda x: not fnmatch(x, disallow)
+
+        self.fallback_mimetype = fallback_mimetype
+
+    def is_allowed(self, filename: str) -> bool:
+        """Subclasses can override this method to disallow the access to
+        certain files.  However by providing `disallow` in the constructor
+        this method is overwritten.
+        """
+        return True
+
+    def _opener(self, filename: str) -> _TOpener:
+        return lambda: (
+            open(filename, "rb"),
+            datetime.fromtimestamp(os.path.getmtime(filename), tz=timezone.utc),
+            int(os.path.getsize(filename)),
+        )
+
+    def get_file_loader(self, filename: str) -> _TLoader:
+        return lambda x: (os.path.basename(filename), self._opener(filename))
+
+    def get_package_loader(self, package: str, package_path: str) -> _TLoader:
+        load_time = datetime.now(timezone.utc)
+        spec = importlib.util.find_spec(package)
+        reader = spec.loader.get_resource_reader(package)  # type: ignore[union-attr]
+
+        def loader(
+            path: str | None,
+        ) -> tuple[str | None, _TOpener | None]:
+            if path is None:
+                return None, None
+
+            path = safe_join(package_path, path)
+
+            if path is None:
+                return None, None
+
+            basename = posixpath.basename(path)
+
+            try:
+                resource = reader.open_resource(path)
+            except OSError:
+                return None, None
+
+            if isinstance(resource, BytesIO):
+                return (
+                    basename,
+                    lambda: (resource, load_time, len(resource.getvalue())),
+                )
+
+            return (
+                basename,
+                lambda: (
+                    resource,
+                    datetime.fromtimestamp(
+                        os.path.getmtime(resource.name), tz=timezone.utc
+                    ),
+                    os.path.getsize(resource.name),
+                ),
+            )
+
+        return loader
+
+    def get_directory_loader(self, directory: str) -> _TLoader:
+        def loader(
+            path: str | None,
+        ) -> tuple[str | None, _TOpener | None]:
+            if path is not None:
+                path = safe_join(directory, path)
+
+                if path is None:
+                    return None, None
+            else:
+                path = directory
+
+            if os.path.isfile(path):
+                return os.path.basename(path), self._opener(path)
+
+            return None, None
+
+        return loader
+
+    def generate_etag(self, mtime: datetime, file_size: int, real_filename: str) -> str:
+        fn_str = os.fsencode(real_filename)
+        timestamp = mtime.timestamp()
+        checksum = adler32(fn_str) & 0xFFFFFFFF
+        return f"wzsdm-{timestamp}-{file_size}-{checksum}"
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        path = get_path_info(environ)
+        file_loader = None
+
+        for search_path, loader in self.exports:
+            if search_path == path:
+                real_filename, file_loader = loader(None)
+
+                if file_loader is not None:
+                    break
+
+            if not search_path.endswith("/"):
+                search_path += "/"
+
+            if path.startswith(search_path):
+                real_filename, file_loader = loader(path[len(search_path) :])
+
+                if file_loader is not None:
+                    break
+
+        if file_loader is None or not self.is_allowed(real_filename):  # type: ignore
+            return self.app(environ, start_response)
+
+        guessed_type = mimetypes.guess_type(real_filename)  # type: ignore
+        mime_type = get_content_type(guessed_type[0] or self.fallback_mimetype, "utf-8")
+        f, mtime, file_size = file_loader()
+
+        headers = [("Date", http_date())]
+
+        if self.cache:
+            timeout = self.cache_timeout
+            etag = self.generate_etag(mtime, file_size, real_filename)  # type: ignore
+            headers += [
+                ("Etag", f'"{etag}"'),
+                ("Cache-Control", f"max-age={timeout}, public"),
+            ]
+
+            if not is_resource_modified(environ, etag, last_modified=mtime):
+                f.close()
+                start_response("304 Not Modified", headers)
+                return []
+
+            headers.append(("Expires", http_date(time() + timeout)))
+        else:
+            headers.append(("Cache-Control", "public"))
+
+        headers.extend(
+            (
+                ("Content-Type", mime_type),
+                ("Content-Length", str(file_size)),
+                ("Last-Modified", http_date(mtime)),
+            )
+        )
+        start_response("200 OK", headers)
+        return wrap_file(environ, f)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/py.typed b/backend/venv/lib/python3.12/site-packages/werkzeug/py.typed
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__init__.py
new file mode 100644
index 00000000..62adc48f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__init__.py
@@ -0,0 +1,134 @@
+"""When it comes to combining multiple controller or view functions
+(however you want to call them) you need a dispatcher. A simple way
+would be applying regular expression tests on the ``PATH_INFO`` and
+calling registered callback functions that return the value then.
+
+This module implements a much more powerful system than simple regular
+expression matching because it can also convert values in the URLs and
+build URLs.
+
+Here a simple example that creates a URL map for an application with
+two subdomains (www and kb) and some URL rules:
+
+.. code-block:: python
+
+    m = Map([
+        # Static URLs
+        Rule('/', endpoint='static/index'),
+        Rule('/about', endpoint='static/about'),
+        Rule('/help', endpoint='static/help'),
+        # Knowledge Base
+        Subdomain('kb', [
+            Rule('/', endpoint='kb/index'),
+            Rule('/browse/', endpoint='kb/browse'),
+            Rule('/browse/<int:id>/', endpoint='kb/browse'),
+            Rule('/browse/<int:id>/<int:page>', endpoint='kb/browse')
+        ])
+    ], default_subdomain='www')
+
+If the application doesn't use subdomains it's perfectly fine to not set
+the default subdomain and not use the `Subdomain` rule factory. The
+endpoint in the rules can be anything, for example import paths or
+unique identifiers. The WSGI application can use those endpoints to get the
+handler for that URL.  It doesn't have to be a string at all but it's
+recommended.
+
+Now it's possible to create a URL adapter for one of the subdomains and
+build URLs:
+
+.. code-block:: python
+
+    c = m.bind('example.com')
+
+    c.build("kb/browse", dict(id=42))
+    'http://kb.example.com/browse/42/'
+
+    c.build("kb/browse", dict())
+    'http://kb.example.com/browse/'
+
+    c.build("kb/browse", dict(id=42, page=3))
+    'http://kb.example.com/browse/42/3'
+
+    c.build("static/about")
+    '/about'
+
+    c.build("static/index", force_external=True)
+    'http://www.example.com/'
+
+    c = m.bind('example.com', subdomain='kb')
+
+    c.build("static/about")
+    'http://www.example.com/about'
+
+The first argument to bind is the server name *without* the subdomain.
+Per default it will assume that the script is mounted on the root, but
+often that's not the case so you can provide the real mount point as
+second argument:
+
+.. code-block:: python
+
+    c = m.bind('example.com', '/applications/example')
+
+The third argument can be the subdomain, if not given the default
+subdomain is used.  For more details about binding have a look at the
+documentation of the `MapAdapter`.
+
+And here is how you can match URLs:
+
+.. code-block:: python
+
+    c = m.bind('example.com')
+
+    c.match("/")
+    ('static/index', {})
+
+    c.match("/about")
+    ('static/about', {})
+
+    c = m.bind('example.com', '/', 'kb')
+
+    c.match("/")
+    ('kb/index', {})
+
+    c.match("/browse/42/23")
+    ('kb/browse', {'id': 42, 'page': 23})
+
+If matching fails you get a ``NotFound`` exception, if the rule thinks
+it's a good idea to redirect (for example because the URL was defined
+to have a slash at the end but the request was missing that slash) it
+will raise a ``RequestRedirect`` exception. Both are subclasses of
+``HTTPException`` so you can use those errors as responses in the
+application.
+
+If matching succeeded but the URL rule was incompatible to the given
+method (for example there were only rules for ``GET`` and ``HEAD`` but
+routing tried to match a ``POST`` request) a ``MethodNotAllowed``
+exception is raised.
+"""
+
+from .converters import AnyConverter as AnyConverter
+from .converters import BaseConverter as BaseConverter
+from .converters import FloatConverter as FloatConverter
+from .converters import IntegerConverter as IntegerConverter
+from .converters import PathConverter as PathConverter
+from .converters import UnicodeConverter as UnicodeConverter
+from .converters import UUIDConverter as UUIDConverter
+from .converters import ValidationError as ValidationError
+from .exceptions import BuildError as BuildError
+from .exceptions import NoMatch as NoMatch
+from .exceptions import RequestAliasRedirect as RequestAliasRedirect
+from .exceptions import RequestPath as RequestPath
+from .exceptions import RequestRedirect as RequestRedirect
+from .exceptions import RoutingException as RoutingException
+from .exceptions import WebsocketMismatch as WebsocketMismatch
+from .map import Map as Map
+from .map import MapAdapter as MapAdapter
+from .matcher import StateMachineMatcher as StateMachineMatcher
+from .rules import EndpointPrefix as EndpointPrefix
+from .rules import parse_converter_args as parse_converter_args
+from .rules import Rule as Rule
+from .rules import RuleFactory as RuleFactory
+from .rules import RuleTemplate as RuleTemplate
+from .rules import RuleTemplateFactory as RuleTemplateFactory
+from .rules import Subdomain as Subdomain
+from .rules import Submount as Submount
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..644a4708
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/converters.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/converters.cpython-312.pyc
new file mode 100644
index 00000000..537ecad3
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/converters.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/exceptions.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/exceptions.cpython-312.pyc
new file mode 100644
index 00000000..731f2597
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/exceptions.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/map.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/map.cpython-312.pyc
new file mode 100644
index 00000000..8b75dda4
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/map.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/matcher.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/matcher.cpython-312.pyc
new file mode 100644
index 00000000..a7ca9bbd
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/matcher.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/rules.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/rules.cpython-312.pyc
new file mode 100644
index 00000000..6ce89590
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/__pycache__/rules.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/converters.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/converters.py
new file mode 100644
index 00000000..6016a975
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/converters.py
@@ -0,0 +1,261 @@
+from __future__ import annotations
+
+import re
+import typing as t
+import uuid
+from urllib.parse import quote
+
+if t.TYPE_CHECKING:
+    from .map import Map
+
+
+class ValidationError(ValueError):
+    """Validation error.  If a rule converter raises this exception the rule
+    does not match the current URL and the next URL is tried.
+    """
+
+
+class BaseConverter:
+    """Base class for all converters.
+
+    .. versionchanged:: 2.3
+        ``part_isolating`` defaults to ``False`` if ``regex`` contains a ``/``.
+    """
+
+    regex = "[^/]+"
+    weight = 100
+    part_isolating = True
+
+    def __init_subclass__(cls, **kwargs: t.Any) -> None:
+        super().__init_subclass__(**kwargs)
+
+        # If the converter isn't inheriting its regex, disable part_isolating by default
+        # if the regex contains a / character.
+        if "regex" in cls.__dict__ and "part_isolating" not in cls.__dict__:
+            cls.part_isolating = "/" not in cls.regex
+
+    def __init__(self, map: Map, *args: t.Any, **kwargs: t.Any) -> None:
+        self.map = map
+
+    def to_python(self, value: str) -> t.Any:
+        return value
+
+    def to_url(self, value: t.Any) -> str:
+        # safe = https://url.spec.whatwg.org/#url-path-segment-string
+        return quote(str(value), safe="!$&'()*+,/:;=@")
+
+
+class UnicodeConverter(BaseConverter):
+    """This converter is the default converter and accepts any string but
+    only one path segment.  Thus the string can not include a slash.
+
+    This is the default validator.
+
+    Example::
+
+        Rule('/pages/<page>'),
+        Rule('/<string(length=2):lang_code>')
+
+    :param map: the :class:`Map`.
+    :param minlength: the minimum length of the string.  Must be greater
+                      or equal 1.
+    :param maxlength: the maximum length of the string.
+    :param length: the exact length of the string.
+    """
+
+    def __init__(
+        self,
+        map: Map,
+        minlength: int = 1,
+        maxlength: int | None = None,
+        length: int | None = None,
+    ) -> None:
+        super().__init__(map)
+        if length is not None:
+            length_regex = f"{{{int(length)}}}"
+        else:
+            if maxlength is None:
+                maxlength_value = ""
+            else:
+                maxlength_value = str(int(maxlength))
+            length_regex = f"{{{int(minlength)},{maxlength_value}}}"
+        self.regex = f"[^/]{length_regex}"
+
+
+class AnyConverter(BaseConverter):
+    """Matches one of the items provided.  Items can either be Python
+    identifiers or strings::
+
+        Rule('/<any(about, help, imprint, class, "foo,bar"):page_name>')
+
+    :param map: the :class:`Map`.
+    :param items: this function accepts the possible items as positional
+                  arguments.
+
+    .. versionchanged:: 2.2
+        Value is validated when building a URL.
+    """
+
+    def __init__(self, map: Map, *items: str) -> None:
+        super().__init__(map)
+        self.items = set(items)
+        self.regex = f"(?:{'|'.join([re.escape(x) for x in items])})"
+
+    def to_url(self, value: t.Any) -> str:
+        if value in self.items:
+            return str(value)
+
+        valid_values = ", ".join(f"'{item}'" for item in sorted(self.items))
+        raise ValueError(f"'{value}' is not one of {valid_values}")
+
+
+class PathConverter(BaseConverter):
+    """Like the default :class:`UnicodeConverter`, but it also matches
+    slashes.  This is useful for wikis and similar applications::
+
+        Rule('/<path:wikipage>')
+        Rule('/<path:wikipage>/edit')
+
+    :param map: the :class:`Map`.
+    """
+
+    part_isolating = False
+    regex = "[^/].*?"
+    weight = 200
+
+
+class NumberConverter(BaseConverter):
+    """Baseclass for `IntegerConverter` and `FloatConverter`.
+
+    :internal:
+    """
+
+    weight = 50
+    num_convert: t.Callable[[t.Any], t.Any] = int
+
+    def __init__(
+        self,
+        map: Map,
+        fixed_digits: int = 0,
+        min: int | None = None,
+        max: int | None = None,
+        signed: bool = False,
+    ) -> None:
+        if signed:
+            self.regex = self.signed_regex
+        super().__init__(map)
+        self.fixed_digits = fixed_digits
+        self.min = min
+        self.max = max
+        self.signed = signed
+
+    def to_python(self, value: str) -> t.Any:
+        if self.fixed_digits and len(value) != self.fixed_digits:
+            raise ValidationError()
+        value_num = self.num_convert(value)
+        if (self.min is not None and value_num < self.min) or (
+            self.max is not None and value_num > self.max
+        ):
+            raise ValidationError()
+        return value_num
+
+    def to_url(self, value: t.Any) -> str:
+        value_str = str(self.num_convert(value))
+        if self.fixed_digits:
+            value_str = value_str.zfill(self.fixed_digits)
+        return value_str
+
+    @property
+    def signed_regex(self) -> str:
+        return f"-?{self.regex}"
+
+
+class IntegerConverter(NumberConverter):
+    """This converter only accepts integer values::
+
+        Rule("/page/<int:page>")
+
+    By default it only accepts unsigned, positive values. The ``signed``
+    parameter will enable signed, negative values. ::
+
+        Rule("/page/<int(signed=True):page>")
+
+    :param map: The :class:`Map`.
+    :param fixed_digits: The number of fixed digits in the URL. If you
+        set this to ``4`` for example, the rule will only match if the
+        URL looks like ``/0001/``. The default is variable length.
+    :param min: The minimal value.
+    :param max: The maximal value.
+    :param signed: Allow signed (negative) values.
+
+    .. versionadded:: 0.15
+        The ``signed`` parameter.
+    """
+
+    regex = r"\d+"
+
+
+class FloatConverter(NumberConverter):
+    """This converter only accepts floating point values::
+
+        Rule("/probability/<float:probability>")
+
+    By default it only accepts unsigned, positive values. The ``signed``
+    parameter will enable signed, negative values. ::
+
+        Rule("/offset/<float(signed=True):offset>")
+
+    :param map: The :class:`Map`.
+    :param min: The minimal value.
+    :param max: The maximal value.
+    :param signed: Allow signed (negative) values.
+
+    .. versionadded:: 0.15
+        The ``signed`` parameter.
+    """
+
+    regex = r"\d+\.\d+"
+    num_convert = float
+
+    def __init__(
+        self,
+        map: Map,
+        min: float | None = None,
+        max: float | None = None,
+        signed: bool = False,
+    ) -> None:
+        super().__init__(map, min=min, max=max, signed=signed)  # type: ignore
+
+
+class UUIDConverter(BaseConverter):
+    """This converter only accepts UUID strings::
+
+        Rule('/object/<uuid:identifier>')
+
+    .. versionadded:: 0.10
+
+    :param map: the :class:`Map`.
+    """
+
+    regex = (
+        r"[A-Fa-f0-9]{8}-[A-Fa-f0-9]{4}-"
+        r"[A-Fa-f0-9]{4}-[A-Fa-f0-9]{4}-[A-Fa-f0-9]{12}"
+    )
+
+    def to_python(self, value: str) -> uuid.UUID:
+        return uuid.UUID(value)
+
+    def to_url(self, value: uuid.UUID) -> str:
+        return str(value)
+
+
+#: the default converter mapping for the map.
+DEFAULT_CONVERTERS: t.Mapping[str, type[BaseConverter]] = {
+    "default": UnicodeConverter,
+    "string": UnicodeConverter,
+    "any": AnyConverter,
+    "path": PathConverter,
+    "int": IntegerConverter,
+    "float": FloatConverter,
+    "uuid": UUIDConverter,
+}
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/exceptions.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/exceptions.py
new file mode 100644
index 00000000..eeabd4ed
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/exceptions.py
@@ -0,0 +1,152 @@
+from __future__ import annotations
+
+import difflib
+import typing as t
+
+from ..exceptions import BadRequest
+from ..exceptions import HTTPException
+from ..utils import cached_property
+from ..utils import redirect
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from ..wrappers.request import Request
+    from ..wrappers.response import Response
+    from .map import MapAdapter
+    from .rules import Rule
+
+
+class RoutingException(Exception):
+    """Special exceptions that require the application to redirect, notifying
+    about missing urls, etc.
+
+    :internal:
+    """
+
+
+class RequestRedirect(HTTPException, RoutingException):
+    """Raise if the map requests a redirect. This is for example the case if
+    `strict_slashes` are activated and an url that requires a trailing slash.
+
+    The attribute `new_url` contains the absolute destination url.
+    """
+
+    code = 308
+
+    def __init__(self, new_url: str) -> None:
+        super().__init__(new_url)
+        self.new_url = new_url
+
+    def get_response(
+        self,
+        environ: WSGIEnvironment | Request | None = None,
+        scope: dict[str, t.Any] | None = None,
+    ) -> Response:
+        return redirect(self.new_url, self.code)
+
+
+class RequestPath(RoutingException):
+    """Internal exception."""
+
+    __slots__ = ("path_info",)
+
+    def __init__(self, path_info: str) -> None:
+        super().__init__()
+        self.path_info = path_info
+
+
+class RequestAliasRedirect(RoutingException):  # noqa: B903
+    """This rule is an alias and wants to redirect to the canonical URL."""
+
+    def __init__(self, matched_values: t.Mapping[str, t.Any], endpoint: t.Any) -> None:
+        super().__init__()
+        self.matched_values = matched_values
+        self.endpoint = endpoint
+
+
+class BuildError(RoutingException, LookupError):
+    """Raised if the build system cannot find a URL for an endpoint with the
+    values provided.
+    """
+
+    def __init__(
+        self,
+        endpoint: t.Any,
+        values: t.Mapping[str, t.Any],
+        method: str | None,
+        adapter: MapAdapter | None = None,
+    ) -> None:
+        super().__init__(endpoint, values, method)
+        self.endpoint = endpoint
+        self.values = values
+        self.method = method
+        self.adapter = adapter
+
+    @cached_property
+    def suggested(self) -> Rule | None:
+        return self.closest_rule(self.adapter)
+
+    def closest_rule(self, adapter: MapAdapter | None) -> Rule | None:
+        def _score_rule(rule: Rule) -> float:
+            return sum(
+                [
+                    0.98
+                    * difflib.SequenceMatcher(
+                        # endpoints can be any type, compare as strings
+                        None,
+                        str(rule.endpoint),
+                        str(self.endpoint),
+                    ).ratio(),
+                    0.01 * bool(set(self.values or ()).issubset(rule.arguments)),
+                    0.01 * bool(rule.methods and self.method in rule.methods),
+                ]
+            )
+
+        if adapter and adapter.map._rules:
+            return max(adapter.map._rules, key=_score_rule)
+
+        return None
+
+    def __str__(self) -> str:
+        message = [f"Could not build url for endpoint {self.endpoint!r}"]
+        if self.method:
+            message.append(f" ({self.method!r})")
+        if self.values:
+            message.append(f" with values {sorted(self.values)!r}")
+        message.append(".")
+        if self.suggested:
+            if self.endpoint == self.suggested.endpoint:
+                if (
+                    self.method
+                    and self.suggested.methods is not None
+                    and self.method not in self.suggested.methods
+                ):
+                    message.append(
+                        " Did you mean to use methods"
+                        f" {sorted(self.suggested.methods)!r}?"
+                    )
+                missing_values = self.suggested.arguments.union(
+                    set(self.suggested.defaults or ())
+                ) - set(self.values.keys())
+                if missing_values:
+                    message.append(
+                        f" Did you forget to specify values {sorted(missing_values)!r}?"
+                    )
+            else:
+                message.append(f" Did you mean {self.suggested.endpoint!r} instead?")
+        return "".join(message)
+
+
+class WebsocketMismatch(BadRequest):
+    """The only matched rule is either a WebSocket and the request is
+    HTTP, or the rule is HTTP and the request is a WebSocket.
+    """
+
+
+class NoMatch(Exception):
+    __slots__ = ("have_match_for", "websocket_mismatch")
+
+    def __init__(self, have_match_for: set[str], websocket_mismatch: bool) -> None:
+        self.have_match_for = have_match_for
+        self.websocket_mismatch = websocket_mismatch
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/map.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/map.py
new file mode 100644
index 00000000..4d15e882
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/map.py
@@ -0,0 +1,951 @@
+from __future__ import annotations
+
+import typing as t
+import warnings
+from pprint import pformat
+from threading import Lock
+from urllib.parse import quote
+from urllib.parse import urljoin
+from urllib.parse import urlunsplit
+
+from .._internal import _get_environ
+from .._internal import _wsgi_decoding_dance
+from ..datastructures import ImmutableDict
+from ..datastructures import MultiDict
+from ..exceptions import BadHost
+from ..exceptions import HTTPException
+from ..exceptions import MethodNotAllowed
+from ..exceptions import NotFound
+from ..urls import _urlencode
+from ..wsgi import get_host
+from .converters import DEFAULT_CONVERTERS
+from .exceptions import BuildError
+from .exceptions import NoMatch
+from .exceptions import RequestAliasRedirect
+from .exceptions import RequestPath
+from .exceptions import RequestRedirect
+from .exceptions import WebsocketMismatch
+from .matcher import StateMachineMatcher
+from .rules import _simple_rule_re
+from .rules import Rule
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from ..wrappers.request import Request
+    from .converters import BaseConverter
+    from .rules import RuleFactory
+
+
+class Map:
+    """The map class stores all the URL rules and some configuration
+    parameters.  Some of the configuration values are only stored on the
+    `Map` instance since those affect all rules, others are just defaults
+    and can be overridden for each rule.  Note that you have to specify all
+    arguments besides the `rules` as keyword arguments!
+
+    :param rules: sequence of url rules for this map.
+    :param default_subdomain: The default subdomain for rules without a
+                              subdomain defined.
+    :param strict_slashes: If a rule ends with a slash but the matched
+        URL does not, redirect to the URL with a trailing slash.
+    :param merge_slashes: Merge consecutive slashes when matching or
+        building URLs. Matches will redirect to the normalized URL.
+        Slashes in variable parts are not merged.
+    :param redirect_defaults: This will redirect to the default rule if it
+                              wasn't visited that way. This helps creating
+                              unique URLs.
+    :param converters: A dict of converters that adds additional converters
+                       to the list of converters. If you redefine one
+                       converter this will override the original one.
+    :param sort_parameters: If set to `True` the url parameters are sorted.
+                            See `url_encode` for more details.
+    :param sort_key: The sort key function for `url_encode`.
+    :param host_matching: if set to `True` it enables the host matching
+                          feature and disables the subdomain one.  If
+                          enabled the `host` parameter to rules is used
+                          instead of the `subdomain` one.
+
+    .. versionchanged:: 3.0
+        The ``charset`` and ``encoding_errors`` parameters were removed.
+
+    .. versionchanged:: 1.0
+        If ``url_scheme`` is ``ws`` or ``wss``, only WebSocket rules will match.
+
+    .. versionchanged:: 1.0
+        The ``merge_slashes`` parameter was added.
+
+    .. versionchanged:: 0.7
+        The ``encoding_errors`` and ``host_matching`` parameters were added.
+
+    .. versionchanged:: 0.5
+        The ``sort_parameters`` and ``sort_key``  paramters were added.
+    """
+
+    #: A dict of default converters to be used.
+    default_converters = ImmutableDict(DEFAULT_CONVERTERS)
+
+    #: The type of lock to use when updating.
+    #:
+    #: .. versionadded:: 1.0
+    lock_class = Lock
+
+    def __init__(
+        self,
+        rules: t.Iterable[RuleFactory] | None = None,
+        default_subdomain: str = "",
+        strict_slashes: bool = True,
+        merge_slashes: bool = True,
+        redirect_defaults: bool = True,
+        converters: t.Mapping[str, type[BaseConverter]] | None = None,
+        sort_parameters: bool = False,
+        sort_key: t.Callable[[t.Any], t.Any] | None = None,
+        host_matching: bool = False,
+    ) -> None:
+        self._matcher = StateMachineMatcher(merge_slashes)
+        self._rules_by_endpoint: dict[t.Any, list[Rule]] = {}
+        self._remap = True
+        self._remap_lock = self.lock_class()
+
+        self.default_subdomain = default_subdomain
+        self.strict_slashes = strict_slashes
+        self.redirect_defaults = redirect_defaults
+        self.host_matching = host_matching
+
+        self.converters = self.default_converters.copy()
+        if converters:
+            self.converters.update(converters)
+
+        self.sort_parameters = sort_parameters
+        self.sort_key = sort_key
+
+        for rulefactory in rules or ():
+            self.add(rulefactory)
+
+    @property
+    def merge_slashes(self) -> bool:
+        return self._matcher.merge_slashes
+
+    @merge_slashes.setter
+    def merge_slashes(self, value: bool) -> None:
+        self._matcher.merge_slashes = value
+
+    def is_endpoint_expecting(self, endpoint: t.Any, *arguments: str) -> bool:
+        """Iterate over all rules and check if the endpoint expects
+        the arguments provided.  This is for example useful if you have
+        some URLs that expect a language code and others that do not and
+        you want to wrap the builder a bit so that the current language
+        code is automatically added if not provided but endpoints expect
+        it.
+
+        :param endpoint: the endpoint to check.
+        :param arguments: this function accepts one or more arguments
+                          as positional arguments.  Each one of them is
+                          checked.
+        """
+        self.update()
+        arguments_set = set(arguments)
+        for rule in self._rules_by_endpoint[endpoint]:
+            if arguments_set.issubset(rule.arguments):
+                return True
+        return False
+
+    @property
+    def _rules(self) -> list[Rule]:
+        return [rule for rules in self._rules_by_endpoint.values() for rule in rules]
+
+    def iter_rules(self, endpoint: t.Any | None = None) -> t.Iterator[Rule]:
+        """Iterate over all rules or the rules of an endpoint.
+
+        :param endpoint: if provided only the rules for that endpoint
+                         are returned.
+        :return: an iterator
+        """
+        self.update()
+        if endpoint is not None:
+            return iter(self._rules_by_endpoint[endpoint])
+        return iter(self._rules)
+
+    def add(self, rulefactory: RuleFactory) -> None:
+        """Add a new rule or factory to the map and bind it.  Requires that the
+        rule is not bound to another map.
+
+        :param rulefactory: a :class:`Rule` or :class:`RuleFactory`
+        """
+        for rule in rulefactory.get_rules(self):
+            rule.bind(self)
+            if not rule.build_only:
+                self._matcher.add(rule)
+            self._rules_by_endpoint.setdefault(rule.endpoint, []).append(rule)
+        self._remap = True
+
+    def bind(
+        self,
+        server_name: str,
+        script_name: str | None = None,
+        subdomain: str | None = None,
+        url_scheme: str = "http",
+        default_method: str = "GET",
+        path_info: str | None = None,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+    ) -> MapAdapter:
+        """Return a new :class:`MapAdapter` with the details specified to the
+        call.  Note that `script_name` will default to ``'/'`` if not further
+        specified or `None`.  The `server_name` at least is a requirement
+        because the HTTP RFC requires absolute URLs for redirects and so all
+        redirect exceptions raised by Werkzeug will contain the full canonical
+        URL.
+
+        If no path_info is passed to :meth:`match` it will use the default path
+        info passed to bind.  While this doesn't really make sense for
+        manual bind calls, it's useful if you bind a map to a WSGI
+        environment which already contains the path info.
+
+        `subdomain` will default to the `default_subdomain` for this map if
+        no defined. If there is no `default_subdomain` you cannot use the
+        subdomain feature.
+
+        .. versionchanged:: 1.0
+            If ``url_scheme`` is ``ws`` or ``wss``, only WebSocket rules
+            will match.
+
+        .. versionchanged:: 0.15
+            ``path_info`` defaults to ``'/'`` if ``None``.
+
+        .. versionchanged:: 0.8
+            ``query_args`` can be a string.
+
+        .. versionchanged:: 0.7
+            Added ``query_args``.
+        """
+        server_name = server_name.lower()
+        if self.host_matching:
+            if subdomain is not None:
+                raise RuntimeError("host matching enabled and a subdomain was provided")
+        elif subdomain is None:
+            subdomain = self.default_subdomain
+        if script_name is None:
+            script_name = "/"
+        if path_info is None:
+            path_info = "/"
+
+        # Port isn't part of IDNA, and might push a name over the 63 octet limit.
+        server_name, port_sep, port = server_name.partition(":")
+
+        try:
+            server_name = server_name.encode("idna").decode("ascii")
+        except UnicodeError as e:
+            raise BadHost() from e
+
+        return MapAdapter(
+            self,
+            f"{server_name}{port_sep}{port}",
+            script_name,
+            subdomain,
+            url_scheme,
+            path_info,
+            default_method,
+            query_args,
+        )
+
+    def bind_to_environ(
+        self,
+        environ: WSGIEnvironment | Request,
+        server_name: str | None = None,
+        subdomain: str | None = None,
+    ) -> MapAdapter:
+        """Like :meth:`bind` but you can pass it an WSGI environment and it
+        will fetch the information from that dictionary.  Note that because of
+        limitations in the protocol there is no way to get the current
+        subdomain and real `server_name` from the environment.  If you don't
+        provide it, Werkzeug will use `SERVER_NAME` and `SERVER_PORT` (or
+        `HTTP_HOST` if provided) as used `server_name` with disabled subdomain
+        feature.
+
+        If `subdomain` is `None` but an environment and a server name is
+        provided it will calculate the current subdomain automatically.
+        Example: `server_name` is ``'example.com'`` and the `SERVER_NAME`
+        in the wsgi `environ` is ``'staging.dev.example.com'`` the calculated
+        subdomain will be ``'staging.dev'``.
+
+        If the object passed as environ has an environ attribute, the value of
+        this attribute is used instead.  This allows you to pass request
+        objects.  Additionally `PATH_INFO` added as a default of the
+        :class:`MapAdapter` so that you don't have to pass the path info to
+        the match method.
+
+        .. versionchanged:: 1.0.0
+            If the passed server name specifies port 443, it will match
+            if the incoming scheme is ``https`` without a port.
+
+        .. versionchanged:: 1.0.0
+            A warning is shown when the passed server name does not
+            match the incoming WSGI server name.
+
+        .. versionchanged:: 0.8
+           This will no longer raise a ValueError when an unexpected server
+           name was passed.
+
+        .. versionchanged:: 0.5
+            previously this method accepted a bogus `calculate_subdomain`
+            parameter that did not have any effect.  It was removed because
+            of that.
+
+        :param environ: a WSGI environment.
+        :param server_name: an optional server name hint (see above).
+        :param subdomain: optionally the current subdomain (see above).
+        """
+        env = _get_environ(environ)
+        wsgi_server_name = get_host(env).lower()
+        scheme = env["wsgi.url_scheme"]
+        upgrade = any(
+            v.strip() == "upgrade"
+            for v in env.get("HTTP_CONNECTION", "").lower().split(",")
+        )
+
+        if upgrade and env.get("HTTP_UPGRADE", "").lower() == "websocket":
+            scheme = "wss" if scheme == "https" else "ws"
+
+        if server_name is None:
+            server_name = wsgi_server_name
+        else:
+            server_name = server_name.lower()
+
+            # strip standard port to match get_host()
+            if scheme in {"http", "ws"} and server_name.endswith(":80"):
+                server_name = server_name[:-3]
+            elif scheme in {"https", "wss"} and server_name.endswith(":443"):
+                server_name = server_name[:-4]
+
+        if subdomain is None and not self.host_matching:
+            cur_server_name = wsgi_server_name.split(".")
+            real_server_name = server_name.split(".")
+            offset = -len(real_server_name)
+
+            if cur_server_name[offset:] != real_server_name:
+                # This can happen even with valid configs if the server was
+                # accessed directly by IP address under some situations.
+                # Instead of raising an exception like in Werkzeug 0.7 or
+                # earlier we go by an invalid subdomain which will result
+                # in a 404 error on matching.
+                warnings.warn(
+                    f"Current server name {wsgi_server_name!r} doesn't match configured"
+                    f" server name {server_name!r}",
+                    stacklevel=2,
+                )
+                subdomain = "<invalid>"
+            else:
+                subdomain = ".".join(filter(None, cur_server_name[:offset]))
+
+        def _get_wsgi_string(name: str) -> str | None:
+            val = env.get(name)
+            if val is not None:
+                return _wsgi_decoding_dance(val)
+            return None
+
+        script_name = _get_wsgi_string("SCRIPT_NAME")
+        path_info = _get_wsgi_string("PATH_INFO")
+        query_args = _get_wsgi_string("QUERY_STRING")
+        return Map.bind(
+            self,
+            server_name,
+            script_name,
+            subdomain,
+            scheme,
+            env["REQUEST_METHOD"],
+            path_info,
+            query_args=query_args,
+        )
+
+    def update(self) -> None:
+        """Called before matching and building to keep the compiled rules
+        in the correct order after things changed.
+        """
+        if not self._remap:
+            return
+
+        with self._remap_lock:
+            if not self._remap:
+                return
+
+            self._matcher.update()
+            for rules in self._rules_by_endpoint.values():
+                rules.sort(key=lambda x: x.build_compare_key())
+            self._remap = False
+
+    def __repr__(self) -> str:
+        rules = self.iter_rules()
+        return f"{type(self).__name__}({pformat(list(rules))})"
+
+
+class MapAdapter:
+    """Returned by :meth:`Map.bind` or :meth:`Map.bind_to_environ` and does
+    the URL matching and building based on runtime information.
+    """
+
+    def __init__(
+        self,
+        map: Map,
+        server_name: str,
+        script_name: str,
+        subdomain: str | None,
+        url_scheme: str,
+        path_info: str,
+        default_method: str,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+    ):
+        self.map = map
+        self.server_name = server_name
+
+        if not script_name.endswith("/"):
+            script_name += "/"
+
+        self.script_name = script_name
+        self.subdomain = subdomain
+        self.url_scheme = url_scheme
+        self.path_info = path_info
+        self.default_method = default_method
+        self.query_args = query_args
+        self.websocket = self.url_scheme in {"ws", "wss"}
+
+    def dispatch(
+        self,
+        view_func: t.Callable[[str, t.Mapping[str, t.Any]], WSGIApplication],
+        path_info: str | None = None,
+        method: str | None = None,
+        catch_http_exceptions: bool = False,
+    ) -> WSGIApplication:
+        """Does the complete dispatching process.  `view_func` is called with
+        the endpoint and a dict with the values for the view.  It should
+        look up the view function, call it, and return a response object
+        or WSGI application.  http exceptions are not caught by default
+        so that applications can display nicer error messages by just
+        catching them by hand.  If you want to stick with the default
+        error messages you can pass it ``catch_http_exceptions=True`` and
+        it will catch the http exceptions.
+
+        Here a small example for the dispatch usage::
+
+            from werkzeug.wrappers import Request, Response
+            from werkzeug.wsgi import responder
+            from werkzeug.routing import Map, Rule
+
+            def on_index(request):
+                return Response('Hello from the index')
+
+            url_map = Map([Rule('/', endpoint='index')])
+            views = {'index': on_index}
+
+            @responder
+            def application(environ, start_response):
+                request = Request(environ)
+                urls = url_map.bind_to_environ(environ)
+                return urls.dispatch(lambda e, v: views[e](request, **v),
+                                     catch_http_exceptions=True)
+
+        Keep in mind that this method might return exception objects, too, so
+        use :class:`Response.force_type` to get a response object.
+
+        :param view_func: a function that is called with the endpoint as
+                          first argument and the value dict as second.  Has
+                          to dispatch to the actual view function with this
+                          information.  (see above)
+        :param path_info: the path info to use for matching.  Overrides the
+                          path info specified on binding.
+        :param method: the HTTP method used for matching.  Overrides the
+                       method specified on binding.
+        :param catch_http_exceptions: set to `True` to catch any of the
+                                      werkzeug :class:`HTTPException`\\s.
+        """
+        try:
+            try:
+                endpoint, args = self.match(path_info, method)
+            except RequestRedirect as e:
+                return e
+            return view_func(endpoint, args)
+        except HTTPException as e:
+            if catch_http_exceptions:
+                return e
+            raise
+
+    @t.overload
+    def match(
+        self,
+        path_info: str | None = None,
+        method: str | None = None,
+        return_rule: t.Literal[False] = False,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+        websocket: bool | None = None,
+    ) -> tuple[t.Any, t.Mapping[str, t.Any]]: ...
+
+    @t.overload
+    def match(
+        self,
+        path_info: str | None = None,
+        method: str | None = None,
+        return_rule: t.Literal[True] = True,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+        websocket: bool | None = None,
+    ) -> tuple[Rule, t.Mapping[str, t.Any]]: ...
+
+    def match(
+        self,
+        path_info: str | None = None,
+        method: str | None = None,
+        return_rule: bool = False,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+        websocket: bool | None = None,
+    ) -> tuple[t.Any | Rule, t.Mapping[str, t.Any]]:
+        """The usage is simple: you just pass the match method the current
+        path info as well as the method (which defaults to `GET`).  The
+        following things can then happen:
+
+        - you receive a `NotFound` exception that indicates that no URL is
+          matching.  A `NotFound` exception is also a WSGI application you
+          can call to get a default page not found page (happens to be the
+          same object as `werkzeug.exceptions.NotFound`)
+
+        - you receive a `MethodNotAllowed` exception that indicates that there
+          is a match for this URL but not for the current request method.
+          This is useful for RESTful applications.
+
+        - you receive a `RequestRedirect` exception with a `new_url`
+          attribute.  This exception is used to notify you about a request
+          Werkzeug requests from your WSGI application.  This is for example the
+          case if you request ``/foo`` although the correct URL is ``/foo/``
+          You can use the `RequestRedirect` instance as response-like object
+          similar to all other subclasses of `HTTPException`.
+
+        - you receive a ``WebsocketMismatch`` exception if the only
+          match is a WebSocket rule but the bind is an HTTP request, or
+          if the match is an HTTP rule but the bind is a WebSocket
+          request.
+
+        - you get a tuple in the form ``(endpoint, arguments)`` if there is
+          a match (unless `return_rule` is True, in which case you get a tuple
+          in the form ``(rule, arguments)``)
+
+        If the path info is not passed to the match method the default path
+        info of the map is used (defaults to the root URL if not defined
+        explicitly).
+
+        All of the exceptions raised are subclasses of `HTTPException` so they
+        can be used as WSGI responses. They will all render generic error or
+        redirect pages.
+
+        Here is a small example for matching:
+
+        >>> m = Map([
+        ...     Rule('/', endpoint='index'),
+        ...     Rule('/downloads/', endpoint='downloads/index'),
+        ...     Rule('/downloads/<int:id>', endpoint='downloads/show')
+        ... ])
+        >>> urls = m.bind("example.com", "/")
+        >>> urls.match("/", "GET")
+        ('index', {})
+        >>> urls.match("/downloads/42")
+        ('downloads/show', {'id': 42})
+
+        And here is what happens on redirect and missing URLs:
+
+        >>> urls.match("/downloads")
+        Traceback (most recent call last):
+          ...
+        RequestRedirect: http://example.com/downloads/
+        >>> urls.match("/missing")
+        Traceback (most recent call last):
+          ...
+        NotFound: 404 Not Found
+
+        :param path_info: the path info to use for matching.  Overrides the
+                          path info specified on binding.
+        :param method: the HTTP method used for matching.  Overrides the
+                       method specified on binding.
+        :param return_rule: return the rule that matched instead of just the
+                            endpoint (defaults to `False`).
+        :param query_args: optional query arguments that are used for
+                           automatic redirects as string or dictionary.  It's
+                           currently not possible to use the query arguments
+                           for URL matching.
+        :param websocket: Match WebSocket instead of HTTP requests. A
+            websocket request has a ``ws`` or ``wss``
+            :attr:`url_scheme`. This overrides that detection.
+
+        .. versionadded:: 1.0
+            Added ``websocket``.
+
+        .. versionchanged:: 0.8
+            ``query_args`` can be a string.
+
+        .. versionadded:: 0.7
+            Added ``query_args``.
+
+        .. versionadded:: 0.6
+            Added ``return_rule``.
+        """
+        self.map.update()
+        if path_info is None:
+            path_info = self.path_info
+        if query_args is None:
+            query_args = self.query_args or {}
+        method = (method or self.default_method).upper()
+
+        if websocket is None:
+            websocket = self.websocket
+
+        domain_part = self.server_name
+
+        if not self.map.host_matching and self.subdomain is not None:
+            domain_part = self.subdomain
+
+        path_part = f"/{path_info.lstrip('/')}" if path_info else ""
+
+        try:
+            result = self.map._matcher.match(domain_part, path_part, method, websocket)
+        except RequestPath as e:
+            # safe = https://url.spec.whatwg.org/#url-path-segment-string
+            new_path = quote(e.path_info, safe="!$&'()*+,/:;=@")
+            raise RequestRedirect(
+                self.make_redirect_url(new_path, query_args)
+            ) from None
+        except RequestAliasRedirect as e:
+            raise RequestRedirect(
+                self.make_alias_redirect_url(
+                    f"{domain_part}|{path_part}",
+                    e.endpoint,
+                    e.matched_values,
+                    method,
+                    query_args,
+                )
+            ) from None
+        except NoMatch as e:
+            if e.have_match_for:
+                raise MethodNotAllowed(valid_methods=list(e.have_match_for)) from None
+
+            if e.websocket_mismatch:
+                raise WebsocketMismatch() from None
+
+            raise NotFound() from None
+        else:
+            rule, rv = result
+
+            if self.map.redirect_defaults:
+                redirect_url = self.get_default_redirect(rule, method, rv, query_args)
+                if redirect_url is not None:
+                    raise RequestRedirect(redirect_url)
+
+            if rule.redirect_to is not None:
+                if isinstance(rule.redirect_to, str):
+
+                    def _handle_match(match: t.Match[str]) -> str:
+                        value = rv[match.group(1)]
+                        return rule._converters[match.group(1)].to_url(value)
+
+                    redirect_url = _simple_rule_re.sub(_handle_match, rule.redirect_to)
+                else:
+                    redirect_url = rule.redirect_to(self, **rv)
+
+                if self.subdomain:
+                    netloc = f"{self.subdomain}.{self.server_name}"
+                else:
+                    netloc = self.server_name
+
+                raise RequestRedirect(
+                    urljoin(
+                        f"{self.url_scheme or 'http'}://{netloc}{self.script_name}",
+                        redirect_url,
+                    )
+                )
+
+            if return_rule:
+                return rule, rv
+            else:
+                return rule.endpoint, rv
+
+    def test(self, path_info: str | None = None, method: str | None = None) -> bool:
+        """Test if a rule would match.  Works like `match` but returns `True`
+        if the URL matches, or `False` if it does not exist.
+
+        :param path_info: the path info to use for matching.  Overrides the
+                          path info specified on binding.
+        :param method: the HTTP method used for matching.  Overrides the
+                       method specified on binding.
+        """
+        try:
+            self.match(path_info, method)
+        except RequestRedirect:
+            pass
+        except HTTPException:
+            return False
+        return True
+
+    def allowed_methods(self, path_info: str | None = None) -> t.Iterable[str]:
+        """Returns the valid methods that match for a given path.
+
+        .. versionadded:: 0.7
+        """
+        try:
+            self.match(path_info, method="--")
+        except MethodNotAllowed as e:
+            return e.valid_methods  # type: ignore
+        except HTTPException:
+            pass
+        return []
+
+    def get_host(self, domain_part: str | None) -> str:
+        """Figures out the full host name for the given domain part.  The
+        domain part is a subdomain in case host matching is disabled or
+        a full host name.
+        """
+        if self.map.host_matching:
+            if domain_part is None:
+                return self.server_name
+
+            return domain_part
+
+        if domain_part is None:
+            subdomain = self.subdomain
+        else:
+            subdomain = domain_part
+
+        if subdomain:
+            return f"{subdomain}.{self.server_name}"
+        else:
+            return self.server_name
+
+    def get_default_redirect(
+        self,
+        rule: Rule,
+        method: str,
+        values: t.MutableMapping[str, t.Any],
+        query_args: t.Mapping[str, t.Any] | str,
+    ) -> str | None:
+        """A helper that returns the URL to redirect to if it finds one.
+        This is used for default redirecting only.
+
+        :internal:
+        """
+        assert self.map.redirect_defaults
+        for r in self.map._rules_by_endpoint[rule.endpoint]:
+            # every rule that comes after this one, including ourself
+            # has a lower priority for the defaults.  We order the ones
+            # with the highest priority up for building.
+            if r is rule:
+                break
+            if r.provides_defaults_for(rule) and r.suitable_for(values, method):
+                values.update(r.defaults)  # type: ignore
+                domain_part, path = r.build(values)  # type: ignore
+                return self.make_redirect_url(path, query_args, domain_part=domain_part)
+        return None
+
+    def encode_query_args(self, query_args: t.Mapping[str, t.Any] | str) -> str:
+        if not isinstance(query_args, str):
+            return _urlencode(query_args)
+        return query_args
+
+    def make_redirect_url(
+        self,
+        path_info: str,
+        query_args: t.Mapping[str, t.Any] | str | None = None,
+        domain_part: str | None = None,
+    ) -> str:
+        """Creates a redirect URL.
+
+        :internal:
+        """
+        if query_args is None:
+            query_args = self.query_args
+
+        if query_args:
+            query_str = self.encode_query_args(query_args)
+        else:
+            query_str = None
+
+        scheme = self.url_scheme or "http"
+        host = self.get_host(domain_part)
+        path = "/".join((self.script_name.strip("/"), path_info.lstrip("/")))
+        return urlunsplit((scheme, host, path, query_str, None))
+
+    def make_alias_redirect_url(
+        self,
+        path: str,
+        endpoint: t.Any,
+        values: t.Mapping[str, t.Any],
+        method: str,
+        query_args: t.Mapping[str, t.Any] | str,
+    ) -> str:
+        """Internally called to make an alias redirect URL."""
+        url = self.build(
+            endpoint, values, method, append_unknown=False, force_external=True
+        )
+        if query_args:
+            url += f"?{self.encode_query_args(query_args)}"
+        assert url != path, "detected invalid alias setting. No canonical URL found"
+        return url
+
+    def _partial_build(
+        self,
+        endpoint: t.Any,
+        values: t.Mapping[str, t.Any],
+        method: str | None,
+        append_unknown: bool,
+    ) -> tuple[str, str, bool] | None:
+        """Helper for :meth:`build`.  Returns subdomain and path for the
+        rule that accepts this endpoint, values and method.
+
+        :internal:
+        """
+        # in case the method is none, try with the default method first
+        if method is None:
+            rv = self._partial_build(
+                endpoint, values, self.default_method, append_unknown
+            )
+            if rv is not None:
+                return rv
+
+        # Default method did not match or a specific method is passed.
+        # Check all for first match with matching host. If no matching
+        # host is found, go with first result.
+        first_match = None
+
+        for rule in self.map._rules_by_endpoint.get(endpoint, ()):
+            if rule.suitable_for(values, method):
+                build_rv = rule.build(values, append_unknown)
+
+                if build_rv is not None:
+                    rv = (build_rv[0], build_rv[1], rule.websocket)
+                    if self.map.host_matching:
+                        if rv[0] == self.server_name:
+                            return rv
+                        elif first_match is None:
+                            first_match = rv
+                    else:
+                        return rv
+
+        return first_match
+
+    def build(
+        self,
+        endpoint: t.Any,
+        values: t.Mapping[str, t.Any] | None = None,
+        method: str | None = None,
+        force_external: bool = False,
+        append_unknown: bool = True,
+        url_scheme: str | None = None,
+    ) -> str:
+        """Building URLs works pretty much the other way round.  Instead of
+        `match` you call `build` and pass it the endpoint and a dict of
+        arguments for the placeholders.
+
+        The `build` function also accepts an argument called `force_external`
+        which, if you set it to `True` will force external URLs. Per default
+        external URLs (include the server name) will only be used if the
+        target URL is on a different subdomain.
+
+        >>> m = Map([
+        ...     Rule('/', endpoint='index'),
+        ...     Rule('/downloads/', endpoint='downloads/index'),
+        ...     Rule('/downloads/<int:id>', endpoint='downloads/show')
+        ... ])
+        >>> urls = m.bind("example.com", "/")
+        >>> urls.build("index", {})
+        '/'
+        >>> urls.build("downloads/show", {'id': 42})
+        '/downloads/42'
+        >>> urls.build("downloads/show", {'id': 42}, force_external=True)
+        'http://example.com/downloads/42'
+
+        Because URLs cannot contain non ASCII data you will always get
+        bytes back.  Non ASCII characters are urlencoded with the
+        charset defined on the map instance.
+
+        Additional values are converted to strings and appended to the URL as
+        URL querystring parameters:
+
+        >>> urls.build("index", {'q': 'My Searchstring'})
+        '/?q=My+Searchstring'
+
+        When processing those additional values, lists are furthermore
+        interpreted as multiple values (as per
+        :py:class:`werkzeug.datastructures.MultiDict`):
+
+        >>> urls.build("index", {'q': ['a', 'b', 'c']})
+        '/?q=a&q=b&q=c'
+
+        Passing a ``MultiDict`` will also add multiple values:
+
+        >>> urls.build("index", MultiDict((('p', 'z'), ('q', 'a'), ('q', 'b'))))
+        '/?p=z&q=a&q=b'
+
+        If a rule does not exist when building a `BuildError` exception is
+        raised.
+
+        The build method accepts an argument called `method` which allows you
+        to specify the method you want to have an URL built for if you have
+        different methods for the same endpoint specified.
+
+        :param endpoint: the endpoint of the URL to build.
+        :param values: the values for the URL to build.  Unhandled values are
+                       appended to the URL as query parameters.
+        :param method: the HTTP method for the rule if there are different
+                       URLs for different methods on the same endpoint.
+        :param force_external: enforce full canonical external URLs. If the URL
+                               scheme is not provided, this will generate
+                               a protocol-relative URL.
+        :param append_unknown: unknown parameters are appended to the generated
+                               URL as query string argument.  Disable this
+                               if you want the builder to ignore those.
+        :param url_scheme: Scheme to use in place of the bound
+            :attr:`url_scheme`.
+
+        .. versionchanged:: 2.0
+            Added the ``url_scheme`` parameter.
+
+        .. versionadded:: 0.6
+           Added the ``append_unknown`` parameter.
+        """
+        self.map.update()
+
+        if values:
+            if isinstance(values, MultiDict):
+                values = {
+                    k: (v[0] if len(v) == 1 else v)
+                    for k, v in dict.items(values)
+                    if len(v) != 0
+                }
+            else:  # plain dict
+                values = {k: v for k, v in values.items() if v is not None}
+        else:
+            values = {}
+
+        rv = self._partial_build(endpoint, values, method, append_unknown)
+        if rv is None:
+            raise BuildError(endpoint, values, method, self)
+
+        domain_part, path, websocket = rv
+        host = self.get_host(domain_part)
+
+        if url_scheme is None:
+            url_scheme = self.url_scheme
+
+        # Always build WebSocket routes with the scheme (browsers
+        # require full URLs). If bound to a WebSocket, ensure that HTTP
+        # routes are built with an HTTP scheme.
+        secure = url_scheme in {"https", "wss"}
+
+        if websocket:
+            force_external = True
+            url_scheme = "wss" if secure else "ws"
+        elif url_scheme:
+            url_scheme = "https" if secure else "http"
+
+        # shortcut this.
+        if not force_external and (
+            (self.map.host_matching and host == self.server_name)
+            or (not self.map.host_matching and domain_part == self.subdomain)
+        ):
+            return f"{self.script_name.rstrip('/')}/{path.lstrip('/')}"
+
+        scheme = f"{url_scheme}:" if url_scheme else ""
+        return f"{scheme}//{host}{self.script_name[:-1]}/{path.lstrip('/')}"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/matcher.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/matcher.py
new file mode 100644
index 00000000..1fd00efc
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/matcher.py
@@ -0,0 +1,202 @@
+from __future__ import annotations
+
+import re
+import typing as t
+from dataclasses import dataclass
+from dataclasses import field
+
+from .converters import ValidationError
+from .exceptions import NoMatch
+from .exceptions import RequestAliasRedirect
+from .exceptions import RequestPath
+from .rules import Rule
+from .rules import RulePart
+
+
+class SlashRequired(Exception):
+    pass
+
+
+@dataclass
+class State:
+    """A representation of a rule state.
+
+    This includes the *rules* that correspond to the state and the
+    possible *static* and *dynamic* transitions to the next state.
+    """
+
+    dynamic: list[tuple[RulePart, State]] = field(default_factory=list)
+    rules: list[Rule] = field(default_factory=list)
+    static: dict[str, State] = field(default_factory=dict)
+
+
+class StateMachineMatcher:
+    def __init__(self, merge_slashes: bool) -> None:
+        self._root = State()
+        self.merge_slashes = merge_slashes
+
+    def add(self, rule: Rule) -> None:
+        state = self._root
+        for part in rule._parts:
+            if part.static:
+                state.static.setdefault(part.content, State())
+                state = state.static[part.content]
+            else:
+                for test_part, new_state in state.dynamic:
+                    if test_part == part:
+                        state = new_state
+                        break
+                else:
+                    new_state = State()
+                    state.dynamic.append((part, new_state))
+                    state = new_state
+        state.rules.append(rule)
+
+    def update(self) -> None:
+        # For every state the dynamic transitions should be sorted by
+        # the weight of the transition
+        state = self._root
+
+        def _update_state(state: State) -> None:
+            state.dynamic.sort(key=lambda entry: entry[0].weight)
+            for new_state in state.static.values():
+                _update_state(new_state)
+            for _, new_state in state.dynamic:
+                _update_state(new_state)
+
+        _update_state(state)
+
+    def match(
+        self, domain: str, path: str, method: str, websocket: bool
+    ) -> tuple[Rule, t.MutableMapping[str, t.Any]]:
+        # To match to a rule we need to start at the root state and
+        # try to follow the transitions until we find a match, or find
+        # there is no transition to follow.
+
+        have_match_for = set()
+        websocket_mismatch = False
+
+        def _match(
+            state: State, parts: list[str], values: list[str]
+        ) -> tuple[Rule, list[str]] | None:
+            # This function is meant to be called recursively, and will attempt
+            # to match the head part to the state's transitions.
+            nonlocal have_match_for, websocket_mismatch
+
+            # The base case is when all parts have been matched via
+            # transitions. Hence if there is a rule with methods &
+            # websocket that work return it and the dynamic values
+            # extracted.
+            if parts == []:
+                for rule in state.rules:
+                    if rule.methods is not None and method not in rule.methods:
+                        have_match_for.update(rule.methods)
+                    elif rule.websocket != websocket:
+                        websocket_mismatch = True
+                    else:
+                        return rule, values
+
+                # Test if there is a match with this path with a
+                # trailing slash, if so raise an exception to report
+                # that matching is possible with an additional slash
+                if "" in state.static:
+                    for rule in state.static[""].rules:
+                        if websocket == rule.websocket and (
+                            rule.methods is None or method in rule.methods
+                        ):
+                            if rule.strict_slashes:
+                                raise SlashRequired()
+                            else:
+                                return rule, values
+                return None
+
+            part = parts[0]
+            # To match this part try the static transitions first
+            if part in state.static:
+                rv = _match(state.static[part], parts[1:], values)
+                if rv is not None:
+                    return rv
+            # No match via the static transitions, so try the dynamic
+            # ones.
+            for test_part, new_state in state.dynamic:
+                target = part
+                remaining = parts[1:]
+                # A final part indicates a transition that always
+                # consumes the remaining parts i.e. transitions to a
+                # final state.
+                if test_part.final:
+                    target = "/".join(parts)
+                    remaining = []
+                match = re.compile(test_part.content).match(target)
+                if match is not None:
+                    if test_part.suffixed:
+                        # If a part_isolating=False part has a slash suffix, remove the
+                        # suffix from the match and check for the slash redirect next.
+                        suffix = match.groups()[-1]
+                        if suffix == "/":
+                            remaining = [""]
+
+                    converter_groups = sorted(
+                        match.groupdict().items(), key=lambda entry: entry[0]
+                    )
+                    groups = [
+                        value
+                        for key, value in converter_groups
+                        if key[:11] == "__werkzeug_"
+                    ]
+                    rv = _match(new_state, remaining, values + groups)
+                    if rv is not None:
+                        return rv
+
+            # If there is no match and the only part left is a
+            # trailing slash ("") consider rules that aren't
+            # strict-slashes as these should match if there is a final
+            # slash part.
+            if parts == [""]:
+                for rule in state.rules:
+                    if rule.strict_slashes:
+                        continue
+                    if rule.methods is not None and method not in rule.methods:
+                        have_match_for.update(rule.methods)
+                    elif rule.websocket != websocket:
+                        websocket_mismatch = True
+                    else:
+                        return rule, values
+
+            return None
+
+        try:
+            rv = _match(self._root, [domain, *path.split("/")], [])
+        except SlashRequired:
+            raise RequestPath(f"{path}/") from None
+
+        if self.merge_slashes and rv is None:
+            # Try to match again, but with slashes merged
+            path = re.sub("/{2,}?", "/", path)
+            try:
+                rv = _match(self._root, [domain, *path.split("/")], [])
+            except SlashRequired:
+                raise RequestPath(f"{path}/") from None
+            if rv is None or rv[0].merge_slashes is False:
+                raise NoMatch(have_match_for, websocket_mismatch)
+            else:
+                raise RequestPath(f"{path}")
+        elif rv is not None:
+            rule, values = rv
+
+            result = {}
+            for name, value in zip(rule._converters.keys(), values):
+                try:
+                    value = rule._converters[name].to_python(value)
+                except ValidationError:
+                    raise NoMatch(have_match_for, websocket_mismatch) from None
+                result[str(name)] = value
+            if rule.defaults:
+                result.update(rule.defaults)
+
+            if rule.alias and rule.map.redirect_defaults:
+                raise RequestAliasRedirect(result, rule.endpoint)
+
+            return rule, result
+
+        raise NoMatch(have_match_for, websocket_mismatch)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/routing/rules.py b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/rules.py
new file mode 100644
index 00000000..2dad31dd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/routing/rules.py
@@ -0,0 +1,928 @@
+from __future__ import annotations
+
+import ast
+import re
+import typing as t
+from dataclasses import dataclass
+from string import Template
+from types import CodeType
+from urllib.parse import quote
+
+from ..datastructures import iter_multi_items
+from ..urls import _urlencode
+from .converters import ValidationError
+
+if t.TYPE_CHECKING:
+    from .converters import BaseConverter
+    from .map import Map
+
+
+class Weighting(t.NamedTuple):
+    number_static_weights: int
+    static_weights: list[tuple[int, int]]
+    number_argument_weights: int
+    argument_weights: list[int]
+
+
+@dataclass
+class RulePart:
+    """A part of a rule.
+
+    Rules can be represented by parts as delimited by `/` with
+    instances of this class representing those parts. The *content* is
+    either the raw content if *static* or a regex string to match
+    against. The *weight* can be used to order parts when matching.
+
+    """
+
+    content: str
+    final: bool
+    static: bool
+    suffixed: bool
+    weight: Weighting
+
+
+_part_re = re.compile(
+    r"""
+    (?:
+        (?P<slash>/)                                 # a slash
+      |
+        (?P<static>[^</]+)                           # static rule data
+      |
+        (?:
+          <
+            (?:
+              (?P<converter>[a-zA-Z_][a-zA-Z0-9_]*)   # converter name
+              (?:\((?P<arguments>.*?)\))?             # converter arguments
+              :                                       # variable delimiter
+            )?
+            (?P<variable>[a-zA-Z_][a-zA-Z0-9_]*)      # variable name
+           >
+        )
+    )
+    """,
+    re.VERBOSE,
+)
+
+_simple_rule_re = re.compile(r"<([^>]+)>")
+_converter_args_re = re.compile(
+    r"""
+    \s*
+    ((?P<name>\w+)\s*=\s*)?
+    (?P<value>
+        True|False|
+        \d+.\d+|
+        \d+.|
+        \d+|
+        [\w\d_.]+|
+        [urUR]?(?P<stringval>"[^"]*?"|'[^']*')
+    )\s*,
+    """,
+    re.VERBOSE,
+)
+
+
+_PYTHON_CONSTANTS = {"None": None, "True": True, "False": False}
+
+
+def _find(value: str, target: str, pos: int) -> int:
+    """Find the *target* in *value* after *pos*.
+
+    Returns the *value* length if *target* isn't found.
+    """
+    try:
+        return value.index(target, pos)
+    except ValueError:
+        return len(value)
+
+
+def _pythonize(value: str) -> None | bool | int | float | str:
+    if value in _PYTHON_CONSTANTS:
+        return _PYTHON_CONSTANTS[value]
+    for convert in int, float:
+        try:
+            return convert(value)
+        except ValueError:
+            pass
+    if value[:1] == value[-1:] and value[0] in "\"'":
+        value = value[1:-1]
+    return str(value)
+
+
+def parse_converter_args(argstr: str) -> tuple[tuple[t.Any, ...], dict[str, t.Any]]:
+    argstr += ","
+    args = []
+    kwargs = {}
+    position = 0
+
+    for item in _converter_args_re.finditer(argstr):
+        if item.start() != position:
+            raise ValueError(
+                f"Cannot parse converter argument '{argstr[position:item.start()]}'"
+            )
+
+        value = item.group("stringval")
+        if value is None:
+            value = item.group("value")
+        value = _pythonize(value)
+        if not item.group("name"):
+            args.append(value)
+        else:
+            name = item.group("name")
+            kwargs[name] = value
+        position = item.end()
+
+    return tuple(args), kwargs
+
+
+class RuleFactory:
+    """As soon as you have more complex URL setups it's a good idea to use rule
+    factories to avoid repetitive tasks.  Some of them are builtin, others can
+    be added by subclassing `RuleFactory` and overriding `get_rules`.
+    """
+
+    def get_rules(self, map: Map) -> t.Iterable[Rule]:
+        """Subclasses of `RuleFactory` have to override this method and return
+        an iterable of rules."""
+        raise NotImplementedError()
+
+
+class Subdomain(RuleFactory):
+    """All URLs provided by this factory have the subdomain set to a
+    specific domain. For example if you want to use the subdomain for
+    the current language this can be a good setup::
+
+        url_map = Map([
+            Rule('/', endpoint='#select_language'),
+            Subdomain('<string(length=2):lang_code>', [
+                Rule('/', endpoint='index'),
+                Rule('/about', endpoint='about'),
+                Rule('/help', endpoint='help')
+            ])
+        ])
+
+    All the rules except for the ``'#select_language'`` endpoint will now
+    listen on a two letter long subdomain that holds the language code
+    for the current request.
+    """
+
+    def __init__(self, subdomain: str, rules: t.Iterable[RuleFactory]) -> None:
+        self.subdomain = subdomain
+        self.rules = rules
+
+    def get_rules(self, map: Map) -> t.Iterator[Rule]:
+        for rulefactory in self.rules:
+            for rule in rulefactory.get_rules(map):
+                rule = rule.empty()
+                rule.subdomain = self.subdomain
+                yield rule
+
+
+class Submount(RuleFactory):
+    """Like `Subdomain` but prefixes the URL rule with a given string::
+
+        url_map = Map([
+            Rule('/', endpoint='index'),
+            Submount('/blog', [
+                Rule('/', endpoint='blog/index'),
+                Rule('/entry/<entry_slug>', endpoint='blog/show')
+            ])
+        ])
+
+    Now the rule ``'blog/show'`` matches ``/blog/entry/<entry_slug>``.
+    """
+
+    def __init__(self, path: str, rules: t.Iterable[RuleFactory]) -> None:
+        self.path = path.rstrip("/")
+        self.rules = rules
+
+    def get_rules(self, map: Map) -> t.Iterator[Rule]:
+        for rulefactory in self.rules:
+            for rule in rulefactory.get_rules(map):
+                rule = rule.empty()
+                rule.rule = self.path + rule.rule
+                yield rule
+
+
+class EndpointPrefix(RuleFactory):
+    """Prefixes all endpoints (which must be strings for this factory) with
+    another string. This can be useful for sub applications::
+
+        url_map = Map([
+            Rule('/', endpoint='index'),
+            EndpointPrefix('blog/', [Submount('/blog', [
+                Rule('/', endpoint='index'),
+                Rule('/entry/<entry_slug>', endpoint='show')
+            ])])
+        ])
+    """
+
+    def __init__(self, prefix: str, rules: t.Iterable[RuleFactory]) -> None:
+        self.prefix = prefix
+        self.rules = rules
+
+    def get_rules(self, map: Map) -> t.Iterator[Rule]:
+        for rulefactory in self.rules:
+            for rule in rulefactory.get_rules(map):
+                rule = rule.empty()
+                rule.endpoint = self.prefix + rule.endpoint
+                yield rule
+
+
+class RuleTemplate:
+    """Returns copies of the rules wrapped and expands string templates in
+    the endpoint, rule, defaults or subdomain sections.
+
+    Here a small example for such a rule template::
+
+        from werkzeug.routing import Map, Rule, RuleTemplate
+
+        resource = RuleTemplate([
+            Rule('/$name/', endpoint='$name.list'),
+            Rule('/$name/<int:id>', endpoint='$name.show')
+        ])
+
+        url_map = Map([resource(name='user'), resource(name='page')])
+
+    When a rule template is called the keyword arguments are used to
+    replace the placeholders in all the string parameters.
+    """
+
+    def __init__(self, rules: t.Iterable[Rule]) -> None:
+        self.rules = list(rules)
+
+    def __call__(self, *args: t.Any, **kwargs: t.Any) -> RuleTemplateFactory:
+        return RuleTemplateFactory(self.rules, dict(*args, **kwargs))
+
+
+class RuleTemplateFactory(RuleFactory):
+    """A factory that fills in template variables into rules.  Used by
+    `RuleTemplate` internally.
+
+    :internal:
+    """
+
+    def __init__(
+        self, rules: t.Iterable[RuleFactory], context: dict[str, t.Any]
+    ) -> None:
+        self.rules = rules
+        self.context = context
+
+    def get_rules(self, map: Map) -> t.Iterator[Rule]:
+        for rulefactory in self.rules:
+            for rule in rulefactory.get_rules(map):
+                new_defaults = subdomain = None
+                if rule.defaults:
+                    new_defaults = {}
+                    for key, value in rule.defaults.items():
+                        if isinstance(value, str):
+                            value = Template(value).substitute(self.context)
+                        new_defaults[key] = value
+                if rule.subdomain is not None:
+                    subdomain = Template(rule.subdomain).substitute(self.context)
+                new_endpoint = rule.endpoint
+                if isinstance(new_endpoint, str):
+                    new_endpoint = Template(new_endpoint).substitute(self.context)
+                yield Rule(
+                    Template(rule.rule).substitute(self.context),
+                    new_defaults,
+                    subdomain,
+                    rule.methods,
+                    rule.build_only,
+                    new_endpoint,
+                    rule.strict_slashes,
+                )
+
+
+_ASTT = t.TypeVar("_ASTT", bound=ast.AST)
+
+
+def _prefix_names(src: str, expected_type: type[_ASTT]) -> _ASTT:
+    """ast parse and prefix names with `.` to avoid collision with user vars"""
+    tree: ast.AST = ast.parse(src).body[0]
+    if isinstance(tree, ast.Expr):
+        tree = tree.value
+    if not isinstance(tree, expected_type):
+        raise TypeError(
+            f"AST node is of type {type(tree).__name__}, not {expected_type.__name__}"
+        )
+    for node in ast.walk(tree):
+        if isinstance(node, ast.Name):
+            node.id = f".{node.id}"
+    return tree
+
+
+_CALL_CONVERTER_CODE_FMT = "self._converters[{elem!r}].to_url()"
+_IF_KWARGS_URL_ENCODE_CODE = """\
+if kwargs:
+    params = self._encode_query_vars(kwargs)
+    q = "?" if params else ""
+else:
+    q = params = ""
+"""
+_IF_KWARGS_URL_ENCODE_AST = _prefix_names(_IF_KWARGS_URL_ENCODE_CODE, ast.If)
+_URL_ENCODE_AST_NAMES = (
+    _prefix_names("q", ast.Name),
+    _prefix_names("params", ast.Name),
+)
+
+
+class Rule(RuleFactory):
+    """A Rule represents one URL pattern.  There are some options for `Rule`
+    that change the way it behaves and are passed to the `Rule` constructor.
+    Note that besides the rule-string all arguments *must* be keyword arguments
+    in order to not break the application on Werkzeug upgrades.
+
+    `string`
+        Rule strings basically are just normal URL paths with placeholders in
+        the format ``<converter(arguments):name>`` where the converter and the
+        arguments are optional.  If no converter is defined the `default`
+        converter is used which means `string` in the normal configuration.
+
+        URL rules that end with a slash are branch URLs, others are leaves.
+        If you have `strict_slashes` enabled (which is the default), all
+        branch URLs that are matched without a trailing slash will trigger a
+        redirect to the same URL with the missing slash appended.
+
+        The converters are defined on the `Map`.
+
+    `endpoint`
+        The endpoint for this rule. This can be anything. A reference to a
+        function, a string, a number etc.  The preferred way is using a string
+        because the endpoint is used for URL generation.
+
+    `defaults`
+        An optional dict with defaults for other rules with the same endpoint.
+        This is a bit tricky but useful if you want to have unique URLs::
+
+            url_map = Map([
+                Rule('/all/', defaults={'page': 1}, endpoint='all_entries'),
+                Rule('/all/page/<int:page>', endpoint='all_entries')
+            ])
+
+        If a user now visits ``http://example.com/all/page/1`` they will be
+        redirected to ``http://example.com/all/``.  If `redirect_defaults` is
+        disabled on the `Map` instance this will only affect the URL
+        generation.
+
+    `subdomain`
+        The subdomain rule string for this rule. If not specified the rule
+        only matches for the `default_subdomain` of the map.  If the map is
+        not bound to a subdomain this feature is disabled.
+
+        Can be useful if you want to have user profiles on different subdomains
+        and all subdomains are forwarded to your application::
+
+            url_map = Map([
+                Rule('/', subdomain='<username>', endpoint='user/homepage'),
+                Rule('/stats', subdomain='<username>', endpoint='user/stats')
+            ])
+
+    `methods`
+        A sequence of http methods this rule applies to.  If not specified, all
+        methods are allowed. For example this can be useful if you want different
+        endpoints for `POST` and `GET`.  If methods are defined and the path
+        matches but the method matched against is not in this list or in the
+        list of another rule for that path the error raised is of the type
+        `MethodNotAllowed` rather than `NotFound`.  If `GET` is present in the
+        list of methods and `HEAD` is not, `HEAD` is added automatically.
+
+    `strict_slashes`
+        Override the `Map` setting for `strict_slashes` only for this rule. If
+        not specified the `Map` setting is used.
+
+    `merge_slashes`
+        Override :attr:`Map.merge_slashes` for this rule.
+
+    `build_only`
+        Set this to True and the rule will never match but will create a URL
+        that can be build. This is useful if you have resources on a subdomain
+        or folder that are not handled by the WSGI application (like static data)
+
+    `redirect_to`
+        If given this must be either a string or callable.  In case of a
+        callable it's called with the url adapter that triggered the match and
+        the values of the URL as keyword arguments and has to return the target
+        for the redirect, otherwise it has to be a string with placeholders in
+        rule syntax::
+
+            def foo_with_slug(adapter, id):
+                # ask the database for the slug for the old id.  this of
+                # course has nothing to do with werkzeug.
+                return f'foo/{Foo.get_slug_for_id(id)}'
+
+            url_map = Map([
+                Rule('/foo/<slug>', endpoint='foo'),
+                Rule('/some/old/url/<slug>', redirect_to='foo/<slug>'),
+                Rule('/other/old/url/<int:id>', redirect_to=foo_with_slug)
+            ])
+
+        When the rule is matched the routing system will raise a
+        `RequestRedirect` exception with the target for the redirect.
+
+        Keep in mind that the URL will be joined against the URL root of the
+        script so don't use a leading slash on the target URL unless you
+        really mean root of that domain.
+
+    `alias`
+        If enabled this rule serves as an alias for another rule with the same
+        endpoint and arguments.
+
+    `host`
+        If provided and the URL map has host matching enabled this can be
+        used to provide a match rule for the whole host.  This also means
+        that the subdomain feature is disabled.
+
+    `websocket`
+        If ``True``, this rule is only matches for WebSocket (``ws://``,
+        ``wss://``) requests. By default, rules will only match for HTTP
+        requests.
+
+    .. versionchanged:: 2.1
+        Percent-encoded newlines (``%0a``), which are decoded by WSGI
+        servers, are considered when routing instead of terminating the
+        match early.
+
+    .. versionadded:: 1.0
+        Added ``websocket``.
+
+    .. versionadded:: 1.0
+        Added ``merge_slashes``.
+
+    .. versionadded:: 0.7
+        Added ``alias`` and ``host``.
+
+    .. versionchanged:: 0.6.1
+       ``HEAD`` is added to ``methods`` if ``GET`` is present.
+    """
+
+    def __init__(
+        self,
+        string: str,
+        defaults: t.Mapping[str, t.Any] | None = None,
+        subdomain: str | None = None,
+        methods: t.Iterable[str] | None = None,
+        build_only: bool = False,
+        endpoint: t.Any | None = None,
+        strict_slashes: bool | None = None,
+        merge_slashes: bool | None = None,
+        redirect_to: str | t.Callable[..., str] | None = None,
+        alias: bool = False,
+        host: str | None = None,
+        websocket: bool = False,
+    ) -> None:
+        if not string.startswith("/"):
+            raise ValueError(f"URL rule '{string}' must start with a slash.")
+
+        self.rule = string
+        self.is_leaf = not string.endswith("/")
+        self.is_branch = string.endswith("/")
+
+        self.map: Map = None  # type: ignore
+        self.strict_slashes = strict_slashes
+        self.merge_slashes = merge_slashes
+        self.subdomain = subdomain
+        self.host = host
+        self.defaults = defaults
+        self.build_only = build_only
+        self.alias = alias
+        self.websocket = websocket
+
+        if methods is not None:
+            if isinstance(methods, str):
+                raise TypeError("'methods' should be a list of strings.")
+
+            methods = {x.upper() for x in methods}
+
+            if "HEAD" not in methods and "GET" in methods:
+                methods.add("HEAD")
+
+            if websocket and methods - {"GET", "HEAD", "OPTIONS"}:
+                raise ValueError(
+                    "WebSocket rules can only use 'GET', 'HEAD', and 'OPTIONS' methods."
+                )
+
+        self.methods = methods
+        self.endpoint: t.Any = endpoint
+        self.redirect_to = redirect_to
+
+        if defaults:
+            self.arguments = set(map(str, defaults))
+        else:
+            self.arguments = set()
+
+        self._converters: dict[str, BaseConverter] = {}
+        self._trace: list[tuple[bool, str]] = []
+        self._parts: list[RulePart] = []
+
+    def empty(self) -> Rule:
+        """
+        Return an unbound copy of this rule.
+
+        This can be useful if want to reuse an already bound URL for another
+        map.  See ``get_empty_kwargs`` to override what keyword arguments are
+        provided to the new copy.
+        """
+        return type(self)(self.rule, **self.get_empty_kwargs())
+
+    def get_empty_kwargs(self) -> t.Mapping[str, t.Any]:
+        """
+        Provides kwargs for instantiating empty copy with empty()
+
+        Use this method to provide custom keyword arguments to the subclass of
+        ``Rule`` when calling ``some_rule.empty()``.  Helpful when the subclass
+        has custom keyword arguments that are needed at instantiation.
+
+        Must return a ``dict`` that will be provided as kwargs to the new
+        instance of ``Rule``, following the initial ``self.rule`` value which
+        is always provided as the first, required positional argument.
+        """
+        defaults = None
+        if self.defaults:
+            defaults = dict(self.defaults)
+        return dict(
+            defaults=defaults,
+            subdomain=self.subdomain,
+            methods=self.methods,
+            build_only=self.build_only,
+            endpoint=self.endpoint,
+            strict_slashes=self.strict_slashes,
+            redirect_to=self.redirect_to,
+            alias=self.alias,
+            host=self.host,
+        )
+
+    def get_rules(self, map: Map) -> t.Iterator[Rule]:
+        yield self
+
+    def refresh(self) -> None:
+        """Rebinds and refreshes the URL.  Call this if you modified the
+        rule in place.
+
+        :internal:
+        """
+        self.bind(self.map, rebind=True)
+
+    def bind(self, map: Map, rebind: bool = False) -> None:
+        """Bind the url to a map and create a regular expression based on
+        the information from the rule itself and the defaults from the map.
+
+        :internal:
+        """
+        if self.map is not None and not rebind:
+            raise RuntimeError(f"url rule {self!r} already bound to map {self.map!r}")
+        self.map = map
+        if self.strict_slashes is None:
+            self.strict_slashes = map.strict_slashes
+        if self.merge_slashes is None:
+            self.merge_slashes = map.merge_slashes
+        if self.subdomain is None:
+            self.subdomain = map.default_subdomain
+        self.compile()
+
+    def get_converter(
+        self,
+        variable_name: str,
+        converter_name: str,
+        args: tuple[t.Any, ...],
+        kwargs: t.Mapping[str, t.Any],
+    ) -> BaseConverter:
+        """Looks up the converter for the given parameter.
+
+        .. versionadded:: 0.9
+        """
+        if converter_name not in self.map.converters:
+            raise LookupError(f"the converter {converter_name!r} does not exist")
+        return self.map.converters[converter_name](self.map, *args, **kwargs)
+
+    def _encode_query_vars(self, query_vars: t.Mapping[str, t.Any]) -> str:
+        items: t.Iterable[tuple[str, str]] = iter_multi_items(query_vars)
+
+        if self.map.sort_parameters:
+            items = sorted(items, key=self.map.sort_key)
+
+        return _urlencode(items)
+
+    def _parse_rule(self, rule: str) -> t.Iterable[RulePart]:
+        content = ""
+        static = True
+        argument_weights = []
+        static_weights: list[tuple[int, int]] = []
+        final = False
+        convertor_number = 0
+
+        pos = 0
+        while pos < len(rule):
+            match = _part_re.match(rule, pos)
+            if match is None:
+                raise ValueError(f"malformed url rule: {rule!r}")
+
+            data = match.groupdict()
+            if data["static"] is not None:
+                static_weights.append((len(static_weights), -len(data["static"])))
+                self._trace.append((False, data["static"]))
+                content += data["static"] if static else re.escape(data["static"])
+
+            if data["variable"] is not None:
+                if static:
+                    # Switching content to represent regex, hence the need to escape
+                    content = re.escape(content)
+                static = False
+                c_args, c_kwargs = parse_converter_args(data["arguments"] or "")
+                convobj = self.get_converter(
+                    data["variable"], data["converter"] or "default", c_args, c_kwargs
+                )
+                self._converters[data["variable"]] = convobj
+                self.arguments.add(data["variable"])
+                if not convobj.part_isolating:
+                    final = True
+                content += f"(?P<__werkzeug_{convertor_number}>{convobj.regex})"
+                convertor_number += 1
+                argument_weights.append(convobj.weight)
+                self._trace.append((True, data["variable"]))
+
+            if data["slash"] is not None:
+                self._trace.append((False, "/"))
+                if final:
+                    content += "/"
+                else:
+                    if not static:
+                        content += r"\Z"
+                    weight = Weighting(
+                        -len(static_weights),
+                        static_weights,
+                        -len(argument_weights),
+                        argument_weights,
+                    )
+                    yield RulePart(
+                        content=content,
+                        final=final,
+                        static=static,
+                        suffixed=False,
+                        weight=weight,
+                    )
+                    content = ""
+                    static = True
+                    argument_weights = []
+                    static_weights = []
+                    final = False
+                    convertor_number = 0
+
+            pos = match.end()
+
+        suffixed = False
+        if final and content[-1] == "/":
+            # If a converter is part_isolating=False (matches slashes) and ends with a
+            # slash, augment the regex to support slash redirects.
+            suffixed = True
+            content = content[:-1] + "(?<!/)(/?)"
+        if not static:
+            content += r"\Z"
+        weight = Weighting(
+            -len(static_weights),
+            static_weights,
+            -len(argument_weights),
+            argument_weights,
+        )
+        yield RulePart(
+            content=content,
+            final=final,
+            static=static,
+            suffixed=suffixed,
+            weight=weight,
+        )
+        if suffixed:
+            yield RulePart(
+                content="", final=False, static=True, suffixed=False, weight=weight
+            )
+
+    def compile(self) -> None:
+        """Compiles the regular expression and stores it."""
+        assert self.map is not None, "rule not bound"
+
+        if self.map.host_matching:
+            domain_rule = self.host or ""
+        else:
+            domain_rule = self.subdomain or ""
+        self._parts = []
+        self._trace = []
+        self._converters = {}
+        if domain_rule == "":
+            self._parts = [
+                RulePart(
+                    content="",
+                    final=False,
+                    static=True,
+                    suffixed=False,
+                    weight=Weighting(0, [], 0, []),
+                )
+            ]
+        else:
+            self._parts.extend(self._parse_rule(domain_rule))
+        self._trace.append((False, "|"))
+        rule = self.rule
+        if self.merge_slashes:
+            rule = re.sub("/{2,}?", "/", self.rule)
+        self._parts.extend(self._parse_rule(rule))
+
+        self._build: t.Callable[..., tuple[str, str]]
+        self._build = self._compile_builder(False).__get__(self, None)
+        self._build_unknown: t.Callable[..., tuple[str, str]]
+        self._build_unknown = self._compile_builder(True).__get__(self, None)
+
+    @staticmethod
+    def _get_func_code(code: CodeType, name: str) -> t.Callable[..., tuple[str, str]]:
+        globs: dict[str, t.Any] = {}
+        locs: dict[str, t.Any] = {}
+        exec(code, globs, locs)
+        return locs[name]  # type: ignore
+
+    def _compile_builder(
+        self, append_unknown: bool = True
+    ) -> t.Callable[..., tuple[str, str]]:
+        defaults = self.defaults or {}
+        dom_ops: list[tuple[bool, str]] = []
+        url_ops: list[tuple[bool, str]] = []
+
+        opl = dom_ops
+        for is_dynamic, data in self._trace:
+            if data == "|" and opl is dom_ops:
+                opl = url_ops
+                continue
+            # this seems like a silly case to ever come up but:
+            # if a default is given for a value that appears in the rule,
+            # resolve it to a constant ahead of time
+            if is_dynamic and data in defaults:
+                data = self._converters[data].to_url(defaults[data])
+                opl.append((False, data))
+            elif not is_dynamic:
+                # safe = https://url.spec.whatwg.org/#url-path-segment-string
+                opl.append((False, quote(data, safe="!$&'()*+,/:;=@")))
+            else:
+                opl.append((True, data))
+
+        def _convert(elem: str) -> ast.Call:
+            ret = _prefix_names(_CALL_CONVERTER_CODE_FMT.format(elem=elem), ast.Call)
+            ret.args = [ast.Name(elem, ast.Load())]
+            return ret
+
+        def _parts(ops: list[tuple[bool, str]]) -> list[ast.expr]:
+            parts: list[ast.expr] = [
+                _convert(elem) if is_dynamic else ast.Constant(elem)
+                for is_dynamic, elem in ops
+            ]
+            parts = parts or [ast.Constant("")]
+            # constant fold
+            ret = [parts[0]]
+            for p in parts[1:]:
+                if isinstance(p, ast.Constant) and isinstance(ret[-1], ast.Constant):
+                    ret[-1] = ast.Constant(ret[-1].value + p.value)
+                else:
+                    ret.append(p)
+            return ret
+
+        dom_parts = _parts(dom_ops)
+        url_parts = _parts(url_ops)
+        body: list[ast.stmt]
+        if not append_unknown:
+            body = []
+        else:
+            body = [_IF_KWARGS_URL_ENCODE_AST]
+            url_parts.extend(_URL_ENCODE_AST_NAMES)
+
+        def _join(parts: list[ast.expr]) -> ast.expr:
+            if len(parts) == 1:  # shortcut
+                return parts[0]
+            return ast.JoinedStr(parts)
+
+        body.append(
+            ast.Return(ast.Tuple([_join(dom_parts), _join(url_parts)], ast.Load()))
+        )
+
+        pargs = [
+            elem
+            for is_dynamic, elem in dom_ops + url_ops
+            if is_dynamic and elem not in defaults
+        ]
+        kargs = [str(k) for k in defaults]
+
+        func_ast = _prefix_names("def _(): pass", ast.FunctionDef)
+        func_ast.name = f"<builder:{self.rule!r}>"
+        func_ast.args.args.append(ast.arg(".self", None))
+        for arg in pargs + kargs:
+            func_ast.args.args.append(ast.arg(arg, None))
+        func_ast.args.kwarg = ast.arg(".kwargs", None)
+        for _ in kargs:
+            func_ast.args.defaults.append(ast.Constant(""))
+        func_ast.body = body
+
+        # Use `ast.parse` instead of `ast.Module` for better portability, since the
+        # signature of `ast.Module` can change.
+        module = ast.parse("")
+        module.body = [func_ast]
+
+        # mark everything as on line 1, offset 0
+        # less error-prone than `ast.fix_missing_locations`
+        # bad line numbers cause an assert to fail in debug builds
+        for node in ast.walk(module):
+            if "lineno" in node._attributes:
+                node.lineno = 1  # type: ignore[attr-defined]
+            if "end_lineno" in node._attributes:
+                node.end_lineno = node.lineno  # type: ignore[attr-defined]
+            if "col_offset" in node._attributes:
+                node.col_offset = 0  # type: ignore[attr-defined]
+            if "end_col_offset" in node._attributes:
+                node.end_col_offset = node.col_offset  # type: ignore[attr-defined]
+
+        code = compile(module, "<werkzeug routing>", "exec")
+        return self._get_func_code(code, func_ast.name)
+
+    def build(
+        self, values: t.Mapping[str, t.Any], append_unknown: bool = True
+    ) -> tuple[str, str] | None:
+        """Assembles the relative url for that rule and the subdomain.
+        If building doesn't work for some reasons `None` is returned.
+
+        :internal:
+        """
+        try:
+            if append_unknown:
+                return self._build_unknown(**values)
+            else:
+                return self._build(**values)
+        except ValidationError:
+            return None
+
+    def provides_defaults_for(self, rule: Rule) -> bool:
+        """Check if this rule has defaults for a given rule.
+
+        :internal:
+        """
+        return bool(
+            not self.build_only
+            and self.defaults
+            and self.endpoint == rule.endpoint
+            and self != rule
+            and self.arguments == rule.arguments
+        )
+
+    def suitable_for(
+        self, values: t.Mapping[str, t.Any], method: str | None = None
+    ) -> bool:
+        """Check if the dict of values has enough data for url generation.
+
+        :internal:
+        """
+        # if a method was given explicitly and that method is not supported
+        # by this rule, this rule is not suitable.
+        if (
+            method is not None
+            and self.methods is not None
+            and method not in self.methods
+        ):
+            return False
+
+        defaults = self.defaults or ()
+
+        # all arguments required must be either in the defaults dict or
+        # the value dictionary otherwise it's not suitable
+        for key in self.arguments:
+            if key not in defaults and key not in values:
+                return False
+
+        # in case defaults are given we ensure that either the value was
+        # skipped or the value is the same as the default value.
+        if defaults:
+            for key, value in defaults.items():
+                if key in values and value != values[key]:
+                    return False
+
+        return True
+
+    def build_compare_key(self) -> tuple[int, int, int]:
+        """The build compare key for sorting.
+
+        :internal:
+        """
+        return (1 if self.alias else 0, -len(self.arguments), -len(self.defaults or ()))
+
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, type(self)) and self._trace == other._trace
+
+    __hash__ = None  # type: ignore
+
+    def __str__(self) -> str:
+        return self.rule
+
+    def __repr__(self) -> str:
+        if self.map is None:
+            return f"<{type(self).__name__} (unbound)>"
+        parts = []
+        for is_dynamic, data in self._trace:
+            if is_dynamic:
+                parts.append(f"<{data}>")
+            else:
+                parts.append(data)
+        parts_str = "".join(parts).lstrip("|")
+        methods = f" ({', '.join(self.methods)})" if self.methods is not None else ""
+        return f"<{type(self).__name__} {parts_str!r}{methods} -> {self.endpoint}>"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..7fe7754c
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/http.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/http.cpython-312.pyc
new file mode 100644
index 00000000..32437363
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/http.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/multipart.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/multipart.cpython-312.pyc
new file mode 100644
index 00000000..833053e6
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/multipart.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..e5d0d53d
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..56adebed
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/utils.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/utils.cpython-312.pyc
new file mode 100644
index 00000000..b26b4ecc
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/__pycache__/utils.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/http.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/http.py
new file mode 100644
index 00000000..f02d7fd5
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/http.py
@@ -0,0 +1,170 @@
+from __future__ import annotations
+
+import re
+import typing as t
+from datetime import datetime
+
+from .._internal import _dt_as_utc
+from ..http import generate_etag
+from ..http import parse_date
+from ..http import parse_etags
+from ..http import parse_if_range_header
+from ..http import unquote_etag
+
+_etag_re = re.compile(r'([Ww]/)?(?:"(.*?)"|(.*?))(?:\s*,\s*|$)')
+
+
+def is_resource_modified(
+    http_range: str | None = None,
+    http_if_range: str | None = None,
+    http_if_modified_since: str | None = None,
+    http_if_none_match: str | None = None,
+    http_if_match: str | None = None,
+    etag: str | None = None,
+    data: bytes | None = None,
+    last_modified: datetime | str | None = None,
+    ignore_if_range: bool = True,
+) -> bool:
+    """Convenience method for conditional requests.
+    :param http_range: Range HTTP header
+    :param http_if_range: If-Range HTTP header
+    :param http_if_modified_since: If-Modified-Since HTTP header
+    :param http_if_none_match: If-None-Match HTTP header
+    :param http_if_match: If-Match HTTP header
+    :param etag: the etag for the response for comparison.
+    :param data: or alternatively the data of the response to automatically
+                 generate an etag using :func:`generate_etag`.
+    :param last_modified: an optional date of the last modification.
+    :param ignore_if_range: If `False`, `If-Range` header will be taken into
+                            account.
+    :return: `True` if the resource was modified, otherwise `False`.
+
+    .. versionadded:: 2.2
+    """
+    if etag is None and data is not None:
+        etag = generate_etag(data)
+    elif data is not None:
+        raise TypeError("both data and etag given")
+
+    unmodified = False
+    if isinstance(last_modified, str):
+        last_modified = parse_date(last_modified)
+
+    # HTTP doesn't use microsecond, remove it to avoid false positive
+    # comparisons. Mark naive datetimes as UTC.
+    if last_modified is not None:
+        last_modified = _dt_as_utc(last_modified.replace(microsecond=0))
+
+    if_range = None
+    if not ignore_if_range and http_range is not None:
+        # https://tools.ietf.org/html/rfc7233#section-3.2
+        # A server MUST ignore an If-Range header field received in a request
+        # that does not contain a Range header field.
+        if_range = parse_if_range_header(http_if_range)
+
+    if if_range is not None and if_range.date is not None:
+        modified_since: datetime | None = if_range.date
+    else:
+        modified_since = parse_date(http_if_modified_since)
+
+    if modified_since and last_modified and last_modified <= modified_since:
+        unmodified = True
+
+    if etag:
+        etag, _ = unquote_etag(etag)
+
+        if if_range is not None and if_range.etag is not None:
+            unmodified = parse_etags(if_range.etag).contains(etag)
+        else:
+            if_none_match = parse_etags(http_if_none_match)
+            if if_none_match:
+                # https://tools.ietf.org/html/rfc7232#section-3.2
+                # "A recipient MUST use the weak comparison function when comparing
+                # entity-tags for If-None-Match"
+                unmodified = if_none_match.contains_weak(etag)
+
+            # https://tools.ietf.org/html/rfc7232#section-3.1
+            # "Origin server MUST use the strong comparison function when
+            # comparing entity-tags for If-Match"
+            if_match = parse_etags(http_if_match)
+            if if_match:
+                unmodified = not if_match.is_strong(etag)
+
+    return not unmodified
+
+
+_cookie_re = re.compile(
+    r"""
+    ([^=;]*)
+    (?:\s*=\s*
+      (
+        "(?:[^\\"]|\\.)*"
+      |
+        .*?
+      )
+    )?
+    \s*;\s*
+    """,
+    flags=re.ASCII | re.VERBOSE,
+)
+_cookie_unslash_re = re.compile(rb"\\([0-3][0-7]{2}|.)")
+
+
+def _cookie_unslash_replace(m: t.Match[bytes]) -> bytes:
+    v = m.group(1)
+
+    if len(v) == 1:
+        return v
+
+    return int(v, 8).to_bytes(1, "big")
+
+
+def parse_cookie(
+    cookie: str | None = None,
+    cls: type[ds.MultiDict[str, str]] | None = None,
+) -> ds.MultiDict[str, str]:
+    """Parse a cookie from a string.
+
+    The same key can be provided multiple times, the values are stored
+    in-order. The default :class:`MultiDict` will have the first value
+    first, and all values can be retrieved with
+    :meth:`MultiDict.getlist`.
+
+    :param cookie: The cookie header as a string.
+    :param cls: A dict-like class to store the parsed cookies in.
+        Defaults to :class:`MultiDict`.
+
+    .. versionchanged:: 3.0
+        Passing bytes, and the ``charset`` and ``errors`` parameters, were removed.
+
+    .. versionadded:: 2.2
+    """
+    if cls is None:
+        cls = t.cast("type[ds.MultiDict[str, str]]", ds.MultiDict)
+
+    if not cookie:
+        return cls()
+
+    cookie = f"{cookie};"
+    out = []
+
+    for ck, cv in _cookie_re.findall(cookie):
+        ck = ck.strip()
+        cv = cv.strip()
+
+        if not ck:
+            continue
+
+        if len(cv) >= 2 and cv[0] == cv[-1] == '"':
+            # Work with bytes here, since a UTF-8 character could be multiple bytes.
+            cv = _cookie_unslash_re.sub(
+                _cookie_unslash_replace, cv[1:-1].encode()
+            ).decode(errors="replace")
+
+        out.append((ck, cv))
+
+    return cls(out)
+
+
+# circular dependencies
+from .. import datastructures as ds
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/multipart.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/multipart.py
new file mode 100644
index 00000000..731be033
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/multipart.py
@@ -0,0 +1,323 @@
+from __future__ import annotations
+
+import re
+import typing as t
+from dataclasses import dataclass
+from enum import auto
+from enum import Enum
+
+from ..datastructures import Headers
+from ..exceptions import RequestEntityTooLarge
+from ..http import parse_options_header
+
+
+class Event:
+    pass
+
+
+@dataclass(frozen=True)
+class Preamble(Event):
+    data: bytes
+
+
+@dataclass(frozen=True)
+class Field(Event):
+    name: str
+    headers: Headers
+
+
+@dataclass(frozen=True)
+class File(Event):
+    name: str
+    filename: str
+    headers: Headers
+
+
+@dataclass(frozen=True)
+class Data(Event):
+    data: bytes
+    more_data: bool
+
+
+@dataclass(frozen=True)
+class Epilogue(Event):
+    data: bytes
+
+
+class NeedData(Event):
+    pass
+
+
+NEED_DATA = NeedData()
+
+
+class State(Enum):
+    PREAMBLE = auto()
+    PART = auto()
+    DATA = auto()
+    DATA_START = auto()
+    EPILOGUE = auto()
+    COMPLETE = auto()
+
+
+# Multipart line breaks MUST be CRLF (\r\n) by RFC-7578, except that
+# many implementations break this and either use CR or LF alone.
+LINE_BREAK = b"(?:\r\n|\n|\r)"
+BLANK_LINE_RE = re.compile(b"(?:\r\n\r\n|\r\r|\n\n)", re.MULTILINE)
+LINE_BREAK_RE = re.compile(LINE_BREAK, re.MULTILINE)
+# Header values can be continued via a space or tab after the linebreak, as
+# per RFC2231
+HEADER_CONTINUATION_RE = re.compile(b"%s[ \t]" % LINE_BREAK, re.MULTILINE)
+# This must be long enough to contain any line breaks plus any
+# additional boundary markers (--) such that they will be found in a
+# subsequent search
+SEARCH_EXTRA_LENGTH = 8
+
+
+class MultipartDecoder:
+    """Decodes a multipart message as bytes into Python events.
+
+    The part data is returned as available to allow the caller to save
+    the data from memory to disk, if desired.
+    """
+
+    def __init__(
+        self,
+        boundary: bytes,
+        max_form_memory_size: int | None = None,
+        *,
+        max_parts: int | None = None,
+    ) -> None:
+        self.buffer = bytearray()
+        self.complete = False
+        self.max_form_memory_size = max_form_memory_size
+        self.max_parts = max_parts
+        self.state = State.PREAMBLE
+        self.boundary = boundary
+
+        # Note in the below \h i.e. horizontal whitespace is used
+        # as [^\S\n\r] as \h isn't supported in python.
+
+        # The preamble must end with a boundary where the boundary is
+        # prefixed by a line break, RFC2046. Except that many
+        # implementations including Werkzeug's tests omit the line
+        # break prefix. In addition the first boundary could be the
+        # epilogue boundary (for empty form-data) hence the matching
+        # group to understand if it is an epilogue boundary.
+        self.preamble_re = re.compile(
+            rb"%s?--%s(--[^\S\n\r]*%s?|[^\S\n\r]*%s)"
+            % (LINE_BREAK, re.escape(boundary), LINE_BREAK, LINE_BREAK),
+            re.MULTILINE,
+        )
+        # A boundary must include a line break prefix and suffix, and
+        # may include trailing whitespace. In addition the boundary
+        # could be the epilogue boundary hence the matching group to
+        # understand if it is an epilogue boundary.
+        self.boundary_re = re.compile(
+            rb"%s--%s(--[^\S\n\r]*%s?|[^\S\n\r]*%s)"
+            % (LINE_BREAK, re.escape(boundary), LINE_BREAK, LINE_BREAK),
+            re.MULTILINE,
+        )
+        self._search_position = 0
+        self._parts_decoded = 0
+
+    def last_newline(self, data: bytes) -> int:
+        try:
+            last_nl = data.rindex(b"\n")
+        except ValueError:
+            last_nl = len(data)
+        try:
+            last_cr = data.rindex(b"\r")
+        except ValueError:
+            last_cr = len(data)
+
+        return min(last_nl, last_cr)
+
+    def receive_data(self, data: bytes | None) -> None:
+        if data is None:
+            self.complete = True
+        elif (
+            self.max_form_memory_size is not None
+            and len(self.buffer) + len(data) > self.max_form_memory_size
+        ):
+            # Ensure that data within single event does not exceed limit.
+            # Also checked across accumulated events in MultiPartParser.
+            raise RequestEntityTooLarge()
+        else:
+            self.buffer.extend(data)
+
+    def next_event(self) -> Event:
+        event: Event = NEED_DATA
+
+        if self.state == State.PREAMBLE:
+            match = self.preamble_re.search(self.buffer, self._search_position)
+            if match is not None:
+                if match.group(1).startswith(b"--"):
+                    self.state = State.EPILOGUE
+                else:
+                    self.state = State.PART
+                data = bytes(self.buffer[: match.start()])
+                del self.buffer[: match.end()]
+                event = Preamble(data=data)
+                self._search_position = 0
+            else:
+                # Update the search start position to be equal to the
+                # current buffer length (already searched) minus a
+                # safe buffer for part of the search target.
+                self._search_position = max(
+                    0, len(self.buffer) - len(self.boundary) - SEARCH_EXTRA_LENGTH
+                )
+
+        elif self.state == State.PART:
+            match = BLANK_LINE_RE.search(self.buffer, self._search_position)
+            if match is not None:
+                headers = self._parse_headers(self.buffer[: match.start()])
+                # The final header ends with a single CRLF, however a
+                # blank line indicates the start of the
+                # body. Therefore the end is after the first CRLF.
+                headers_end = (match.start() + match.end()) // 2
+                del self.buffer[:headers_end]
+
+                if "content-disposition" not in headers:
+                    raise ValueError("Missing Content-Disposition header")
+
+                disposition, extra = parse_options_header(
+                    headers["content-disposition"]
+                )
+                name = t.cast(str, extra.get("name"))
+                filename = extra.get("filename")
+                if filename is not None:
+                    event = File(
+                        filename=filename,
+                        headers=headers,
+                        name=name,
+                    )
+                else:
+                    event = Field(
+                        headers=headers,
+                        name=name,
+                    )
+                self.state = State.DATA_START
+                self._search_position = 0
+                self._parts_decoded += 1
+
+                if self.max_parts is not None and self._parts_decoded > self.max_parts:
+                    raise RequestEntityTooLarge()
+            else:
+                # Update the search start position to be equal to the
+                # current buffer length (already searched) minus a
+                # safe buffer for part of the search target.
+                self._search_position = max(0, len(self.buffer) - SEARCH_EXTRA_LENGTH)
+
+        elif self.state == State.DATA_START:
+            data, del_index, more_data = self._parse_data(self.buffer, start=True)
+            del self.buffer[:del_index]
+            event = Data(data=data, more_data=more_data)
+            if more_data:
+                self.state = State.DATA
+
+        elif self.state == State.DATA:
+            data, del_index, more_data = self._parse_data(self.buffer, start=False)
+            del self.buffer[:del_index]
+            if data or not more_data:
+                event = Data(data=data, more_data=more_data)
+
+        elif self.state == State.EPILOGUE and self.complete:
+            event = Epilogue(data=bytes(self.buffer))
+            del self.buffer[:]
+            self.state = State.COMPLETE
+
+        if self.complete and isinstance(event, NeedData):
+            raise ValueError(f"Invalid form-data cannot parse beyond {self.state}")
+
+        return event
+
+    def _parse_headers(self, data: bytes) -> Headers:
+        headers: list[tuple[str, str]] = []
+        # Merge the continued headers into one line
+        data = HEADER_CONTINUATION_RE.sub(b" ", data)
+        # Now there is one header per line
+        for line in data.splitlines():
+            line = line.strip()
+
+            if line != b"":
+                name, _, value = line.decode().partition(":")
+                headers.append((name.strip(), value.strip()))
+        return Headers(headers)
+
+    def _parse_data(self, data: bytes, *, start: bool) -> tuple[bytes, int, bool]:
+        # Body parts must start with CRLF (or CR or LF)
+        if start:
+            match = LINE_BREAK_RE.match(data)
+            data_start = t.cast(t.Match[bytes], match).end()
+        else:
+            data_start = 0
+
+        boundary = b"--" + self.boundary
+
+        if self.buffer.find(boundary) == -1:
+            # No complete boundary in the buffer, but there may be
+            # a partial boundary at the end. As the boundary
+            # starts with either a nl or cr find the earliest and
+            # return up to that as data.
+            data_end = del_index = self.last_newline(data[data_start:]) + data_start
+            # If amount of data after last newline is far from
+            # possible length of partial boundary, we should
+            # assume that there is no partial boundary in the buffer
+            # and return all pending data.
+            if (len(data) - data_end) > len(b"\n" + boundary):
+                data_end = del_index = len(data)
+            more_data = True
+        else:
+            match = self.boundary_re.search(data)
+            if match is not None:
+                if match.group(1).startswith(b"--"):
+                    self.state = State.EPILOGUE
+                else:
+                    self.state = State.PART
+                data_end = match.start()
+                del_index = match.end()
+            else:
+                data_end = del_index = self.last_newline(data[data_start:]) + data_start
+            more_data = match is None
+
+        return bytes(data[data_start:data_end]), del_index, more_data
+
+
+class MultipartEncoder:
+    def __init__(self, boundary: bytes) -> None:
+        self.boundary = boundary
+        self.state = State.PREAMBLE
+
+    def send_event(self, event: Event) -> bytes:
+        if isinstance(event, Preamble) and self.state == State.PREAMBLE:
+            self.state = State.PART
+            return event.data
+        elif isinstance(event, (Field, File)) and self.state in {
+            State.PREAMBLE,
+            State.PART,
+            State.DATA,
+        }:
+            data = b"\r\n--" + self.boundary + b"\r\n"
+            data += b'Content-Disposition: form-data; name="%s"' % event.name.encode()
+            if isinstance(event, File):
+                data += b'; filename="%s"' % event.filename.encode()
+            data += b"\r\n"
+            for name, value in t.cast(Field, event).headers:
+                if name.lower() != "content-disposition":
+                    data += f"{name}: {value}\r\n".encode()
+            self.state = State.DATA_START
+            return data
+        elif isinstance(event, Data) and self.state == State.DATA_START:
+            self.state = State.DATA
+            if len(event.data) > 0:
+                return b"\r\n" + event.data
+            else:
+                return event.data
+        elif isinstance(event, Data) and self.state == State.DATA:
+            return event.data
+        elif isinstance(event, Epilogue):
+            self.state = State.COMPLETE
+            return b"\r\n--" + self.boundary + b"--\r\n" + event.data
+        else:
+            raise ValueError(f"Cannot generate {event} in state: {self.state}")
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/request.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/request.py
new file mode 100644
index 00000000..8d5fbd8f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/request.py
@@ -0,0 +1,534 @@
+from __future__ import annotations
+
+import typing as t
+from datetime import datetime
+from urllib.parse import parse_qsl
+
+from ..datastructures import Accept
+from ..datastructures import Authorization
+from ..datastructures import CharsetAccept
+from ..datastructures import ETags
+from ..datastructures import Headers
+from ..datastructures import HeaderSet
+from ..datastructures import IfRange
+from ..datastructures import ImmutableList
+from ..datastructures import ImmutableMultiDict
+from ..datastructures import LanguageAccept
+from ..datastructures import MIMEAccept
+from ..datastructures import MultiDict
+from ..datastructures import Range
+from ..datastructures import RequestCacheControl
+from ..http import parse_accept_header
+from ..http import parse_cache_control_header
+from ..http import parse_date
+from ..http import parse_etags
+from ..http import parse_if_range_header
+from ..http import parse_list_header
+from ..http import parse_options_header
+from ..http import parse_range_header
+from ..http import parse_set_header
+from ..user_agent import UserAgent
+from ..utils import cached_property
+from ..utils import header_property
+from .http import parse_cookie
+from .utils import get_content_length
+from .utils import get_current_url
+from .utils import get_host
+
+
+class Request:
+    """Represents the non-IO parts of a HTTP request, including the
+    method, URL info, and headers.
+
+    This class is not meant for general use. It should only be used when
+    implementing WSGI, ASGI, or another HTTP application spec. Werkzeug
+    provides a WSGI implementation at :cls:`werkzeug.wrappers.Request`.
+
+    :param method: The method the request was made with, such as
+        ``GET``.
+    :param scheme: The URL scheme of the protocol the request used, such
+        as ``https`` or ``wss``.
+    :param server: The address of the server. ``(host, port)``,
+        ``(path, None)`` for unix sockets, or ``None`` if not known.
+    :param root_path: The prefix that the application is mounted under.
+        This is prepended to generated URLs, but is not part of route
+        matching.
+    :param path: The path part of the URL after ``root_path``.
+    :param query_string: The part of the URL after the "?".
+    :param headers: The headers received with the request.
+    :param remote_addr: The address of the client sending the request.
+
+    .. versionchanged:: 3.0
+        The ``charset``, ``url_charset``, and ``encoding_errors`` attributes
+        were removed.
+
+    .. versionadded:: 2.0
+    """
+
+    #: the class to use for `args` and `form`.  The default is an
+    #: :class:`~werkzeug.datastructures.ImmutableMultiDict` which supports
+    #: multiple values per key. A :class:`~werkzeug.datastructures.ImmutableDict`
+    #: is faster but only remembers the last key. It is also
+    #: possible to use mutable structures, but this is not recommended.
+    #:
+    #: .. versionadded:: 0.6
+    parameter_storage_class: type[MultiDict[str, t.Any]] = ImmutableMultiDict
+
+    #: The type to be used for dict values from the incoming WSGI
+    #: environment. (For example for :attr:`cookies`.) By default an
+    #: :class:`~werkzeug.datastructures.ImmutableMultiDict` is used.
+    #:
+    #: .. versionchanged:: 1.0.0
+    #:     Changed to ``ImmutableMultiDict`` to support multiple values.
+    #:
+    #: .. versionadded:: 0.6
+    dict_storage_class: type[MultiDict[str, t.Any]] = ImmutableMultiDict
+
+    #: the type to be used for list values from the incoming WSGI environment.
+    #: By default an :class:`~werkzeug.datastructures.ImmutableList` is used
+    #: (for example for :attr:`access_list`).
+    #:
+    #: .. versionadded:: 0.6
+    list_storage_class: type[list[t.Any]] = ImmutableList
+
+    user_agent_class: type[UserAgent] = UserAgent
+    """The class used and returned by the :attr:`user_agent` property to
+    parse the header. Defaults to
+    :class:`~werkzeug.user_agent.UserAgent`, which does no parsing. An
+    extension can provide a subclass that uses a parser to provide other
+    data.
+
+    .. versionadded:: 2.0
+    """
+
+    #: Valid host names when handling requests. By default all hosts are
+    #: trusted, which means that whatever the client says the host is
+    #: will be accepted.
+    #:
+    #: Because ``Host`` and ``X-Forwarded-Host`` headers can be set to
+    #: any value by a malicious client, it is recommended to either set
+    #: this property or implement similar validation in the proxy (if
+    #: the application is being run behind one).
+    #:
+    #: .. versionadded:: 0.9
+    trusted_hosts: list[str] | None = None
+
+    def __init__(
+        self,
+        method: str,
+        scheme: str,
+        server: tuple[str, int | None] | None,
+        root_path: str,
+        path: str,
+        query_string: bytes,
+        headers: Headers,
+        remote_addr: str | None,
+    ) -> None:
+        #: The method the request was made with, such as ``GET``.
+        self.method = method.upper()
+        #: The URL scheme of the protocol the request used, such as
+        #: ``https`` or ``wss``.
+        self.scheme = scheme
+        #: The address of the server. ``(host, port)``, ``(path, None)``
+        #: for unix sockets, or ``None`` if not known.
+        self.server = server
+        #: The prefix that the application is mounted under, without a
+        #: trailing slash. :attr:`path` comes after this.
+        self.root_path = root_path.rstrip("/")
+        #: The path part of the URL after :attr:`root_path`. This is the
+        #: path used for routing within the application.
+        self.path = "/" + path.lstrip("/")
+        #: The part of the URL after the "?". This is the raw value, use
+        #: :attr:`args` for the parsed values.
+        self.query_string = query_string
+        #: The headers received with the request.
+        self.headers = headers
+        #: The address of the client sending the request.
+        self.remote_addr = remote_addr
+
+    def __repr__(self) -> str:
+        try:
+            url = self.url
+        except Exception as e:
+            url = f"(invalid URL: {e})"
+
+        return f"<{type(self).__name__} {url!r} [{self.method}]>"
+
+    @cached_property
+    def args(self) -> MultiDict[str, str]:
+        """The parsed URL parameters (the part in the URL after the question
+        mark).
+
+        By default an
+        :class:`~werkzeug.datastructures.ImmutableMultiDict`
+        is returned from this function.  This can be changed by setting
+        :attr:`parameter_storage_class` to a different type.  This might
+        be necessary if the order of the form data is important.
+
+        .. versionchanged:: 2.3
+            Invalid bytes remain percent encoded.
+        """
+        return self.parameter_storage_class(
+            parse_qsl(
+                self.query_string.decode(),
+                keep_blank_values=True,
+                errors="werkzeug.url_quote",
+            )
+        )
+
+    @cached_property
+    def access_route(self) -> list[str]:
+        """If a forwarded header exists this is a list of all ip addresses
+        from the client ip to the last proxy server.
+        """
+        if "X-Forwarded-For" in self.headers:
+            return self.list_storage_class(
+                parse_list_header(self.headers["X-Forwarded-For"])
+            )
+        elif self.remote_addr is not None:
+            return self.list_storage_class([self.remote_addr])
+        return self.list_storage_class()
+
+    @cached_property
+    def full_path(self) -> str:
+        """Requested path, including the query string."""
+        return f"{self.path}?{self.query_string.decode()}"
+
+    @property
+    def is_secure(self) -> bool:
+        """``True`` if the request was made with a secure protocol
+        (HTTPS or WSS).
+        """
+        return self.scheme in {"https", "wss"}
+
+    @cached_property
+    def url(self) -> str:
+        """The full request URL with the scheme, host, root path, path,
+        and query string."""
+        return get_current_url(
+            self.scheme, self.host, self.root_path, self.path, self.query_string
+        )
+
+    @cached_property
+    def base_url(self) -> str:
+        """Like :attr:`url` but without the query string."""
+        return get_current_url(self.scheme, self.host, self.root_path, self.path)
+
+    @cached_property
+    def root_url(self) -> str:
+        """The request URL scheme, host, and root path. This is the root
+        that the application is accessed from.
+        """
+        return get_current_url(self.scheme, self.host, self.root_path)
+
+    @cached_property
+    def host_url(self) -> str:
+        """The request URL scheme and host only."""
+        return get_current_url(self.scheme, self.host)
+
+    @cached_property
+    def host(self) -> str:
+        """The host name the request was made to, including the port if
+        it's non-standard. Validated with :attr:`trusted_hosts`.
+        """
+        return get_host(
+            self.scheme, self.headers.get("host"), self.server, self.trusted_hosts
+        )
+
+    @cached_property
+    def cookies(self) -> ImmutableMultiDict[str, str]:
+        """A :class:`dict` with the contents of all cookies transmitted with
+        the request."""
+        wsgi_combined_cookie = ";".join(self.headers.getlist("Cookie"))
+        return parse_cookie(  # type: ignore
+            wsgi_combined_cookie, cls=self.dict_storage_class
+        )
+
+    # Common Descriptors
+
+    content_type = header_property[str](
+        "Content-Type",
+        doc="""The Content-Type entity-header field indicates the media
+        type of the entity-body sent to the recipient or, in the case of
+        the HEAD method, the media type that would have been sent had
+        the request been a GET.""",
+        read_only=True,
+    )
+
+    @cached_property
+    def content_length(self) -> int | None:
+        """The Content-Length entity-header field indicates the size of the
+        entity-body in bytes or, in the case of the HEAD method, the size of
+        the entity-body that would have been sent had the request been a
+        GET.
+        """
+        return get_content_length(
+            http_content_length=self.headers.get("Content-Length"),
+            http_transfer_encoding=self.headers.get("Transfer-Encoding"),
+        )
+
+    content_encoding = header_property[str](
+        "Content-Encoding",
+        doc="""The Content-Encoding entity-header field is used as a
+        modifier to the media-type. When present, its value indicates
+        what additional content codings have been applied to the
+        entity-body, and thus what decoding mechanisms must be applied
+        in order to obtain the media-type referenced by the Content-Type
+        header field.
+
+        .. versionadded:: 0.9""",
+        read_only=True,
+    )
+    content_md5 = header_property[str](
+        "Content-MD5",
+        doc="""The Content-MD5 entity-header field, as defined in
+        RFC 1864, is an MD5 digest of the entity-body for the purpose of
+        providing an end-to-end message integrity check (MIC) of the
+        entity-body. (Note: a MIC is good for detecting accidental
+        modification of the entity-body in transit, but is not proof
+        against malicious attacks.)
+
+        .. versionadded:: 0.9""",
+        read_only=True,
+    )
+    referrer = header_property[str](
+        "Referer",
+        doc="""The Referer[sic] request-header field allows the client
+        to specify, for the server's benefit, the address (URI) of the
+        resource from which the Request-URI was obtained (the
+        "referrer", although the header field is misspelled).""",
+        read_only=True,
+    )
+    date = header_property(
+        "Date",
+        None,
+        parse_date,
+        doc="""The Date general-header field represents the date and
+        time at which the message was originated, having the same
+        semantics as orig-date in RFC 822.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """,
+        read_only=True,
+    )
+    max_forwards = header_property(
+        "Max-Forwards",
+        None,
+        int,
+        doc="""The Max-Forwards request-header field provides a
+        mechanism with the TRACE and OPTIONS methods to limit the number
+        of proxies or gateways that can forward the request to the next
+        inbound server.""",
+        read_only=True,
+    )
+
+    def _parse_content_type(self) -> None:
+        if not hasattr(self, "_parsed_content_type"):
+            self._parsed_content_type = parse_options_header(
+                self.headers.get("Content-Type", "")
+            )
+
+    @property
+    def mimetype(self) -> str:
+        """Like :attr:`content_type`, but without parameters (eg, without
+        charset, type etc.) and always lowercase.  For example if the content
+        type is ``text/HTML; charset=utf-8`` the mimetype would be
+        ``'text/html'``.
+        """
+        self._parse_content_type()
+        return self._parsed_content_type[0].lower()
+
+    @property
+    def mimetype_params(self) -> dict[str, str]:
+        """The mimetype parameters as dict.  For example if the content
+        type is ``text/html; charset=utf-8`` the params would be
+        ``{'charset': 'utf-8'}``.
+        """
+        self._parse_content_type()
+        return self._parsed_content_type[1]
+
+    @cached_property
+    def pragma(self) -> HeaderSet:
+        """The Pragma general-header field is used to include
+        implementation-specific directives that might apply to any recipient
+        along the request/response chain.  All pragma directives specify
+        optional behavior from the viewpoint of the protocol; however, some
+        systems MAY require that behavior be consistent with the directives.
+        """
+        return parse_set_header(self.headers.get("Pragma", ""))
+
+    # Accept
+
+    @cached_property
+    def accept_mimetypes(self) -> MIMEAccept:
+        """List of mimetypes this client supports as
+        :class:`~werkzeug.datastructures.MIMEAccept` object.
+        """
+        return parse_accept_header(self.headers.get("Accept"), MIMEAccept)
+
+    @cached_property
+    def accept_charsets(self) -> CharsetAccept:
+        """List of charsets this client supports as
+        :class:`~werkzeug.datastructures.CharsetAccept` object.
+        """
+        return parse_accept_header(self.headers.get("Accept-Charset"), CharsetAccept)
+
+    @cached_property
+    def accept_encodings(self) -> Accept:
+        """List of encodings this client accepts.  Encodings in a HTTP term
+        are compression encodings such as gzip.  For charsets have a look at
+        :attr:`accept_charset`.
+        """
+        return parse_accept_header(self.headers.get("Accept-Encoding"))
+
+    @cached_property
+    def accept_languages(self) -> LanguageAccept:
+        """List of languages this client accepts as
+        :class:`~werkzeug.datastructures.LanguageAccept` object.
+
+        .. versionchanged 0.5
+           In previous versions this was a regular
+           :class:`~werkzeug.datastructures.Accept` object.
+        """
+        return parse_accept_header(self.headers.get("Accept-Language"), LanguageAccept)
+
+    # ETag
+
+    @cached_property
+    def cache_control(self) -> RequestCacheControl:
+        """A :class:`~werkzeug.datastructures.RequestCacheControl` object
+        for the incoming cache control headers.
+        """
+        cache_control = self.headers.get("Cache-Control")
+        return parse_cache_control_header(cache_control, None, RequestCacheControl)
+
+    @cached_property
+    def if_match(self) -> ETags:
+        """An object containing all the etags in the `If-Match` header.
+
+        :rtype: :class:`~werkzeug.datastructures.ETags`
+        """
+        return parse_etags(self.headers.get("If-Match"))
+
+    @cached_property
+    def if_none_match(self) -> ETags:
+        """An object containing all the etags in the `If-None-Match` header.
+
+        :rtype: :class:`~werkzeug.datastructures.ETags`
+        """
+        return parse_etags(self.headers.get("If-None-Match"))
+
+    @cached_property
+    def if_modified_since(self) -> datetime | None:
+        """The parsed `If-Modified-Since` header as a datetime object.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """
+        return parse_date(self.headers.get("If-Modified-Since"))
+
+    @cached_property
+    def if_unmodified_since(self) -> datetime | None:
+        """The parsed `If-Unmodified-Since` header as a datetime object.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """
+        return parse_date(self.headers.get("If-Unmodified-Since"))
+
+    @cached_property
+    def if_range(self) -> IfRange:
+        """The parsed ``If-Range`` header.
+
+        .. versionchanged:: 2.0
+            ``IfRange.date`` is timezone-aware.
+
+        .. versionadded:: 0.7
+        """
+        return parse_if_range_header(self.headers.get("If-Range"))
+
+    @cached_property
+    def range(self) -> Range | None:
+        """The parsed `Range` header.
+
+        .. versionadded:: 0.7
+
+        :rtype: :class:`~werkzeug.datastructures.Range`
+        """
+        return parse_range_header(self.headers.get("Range"))
+
+    # User Agent
+
+    @cached_property
+    def user_agent(self) -> UserAgent:
+        """The user agent. Use ``user_agent.string`` to get the header
+        value. Set :attr:`user_agent_class` to a subclass of
+        :class:`~werkzeug.user_agent.UserAgent` to provide parsing for
+        the other properties or other extended data.
+
+        .. versionchanged:: 2.1
+            The built-in parser was removed. Set ``user_agent_class`` to a ``UserAgent``
+            subclass to parse data from the string.
+        """
+        return self.user_agent_class(self.headers.get("User-Agent", ""))
+
+    # Authorization
+
+    @cached_property
+    def authorization(self) -> Authorization | None:
+        """The ``Authorization`` header parsed into an :class:`.Authorization` object.
+        ``None`` if the header is not present.
+
+        .. versionchanged:: 2.3
+            :class:`Authorization` is no longer a ``dict``. The ``token`` attribute
+            was added for auth schemes that use a token instead of parameters.
+        """
+        return Authorization.from_header(self.headers.get("Authorization"))
+
+    # CORS
+
+    origin = header_property[str](
+        "Origin",
+        doc=(
+            "The host that the request originated from. Set"
+            " :attr:`~CORSResponseMixin.access_control_allow_origin` on"
+            " the response to indicate which origins are allowed."
+        ),
+        read_only=True,
+    )
+
+    access_control_request_headers = header_property(
+        "Access-Control-Request-Headers",
+        load_func=parse_set_header,
+        doc=(
+            "Sent with a preflight request to indicate which headers"
+            " will be sent with the cross origin request. Set"
+            " :attr:`~CORSResponseMixin.access_control_allow_headers`"
+            " on the response to indicate which headers are allowed."
+        ),
+        read_only=True,
+    )
+
+    access_control_request_method = header_property[str](
+        "Access-Control-Request-Method",
+        doc=(
+            "Sent with a preflight request to indicate which method"
+            " will be used for the cross origin request. Set"
+            " :attr:`~CORSResponseMixin.access_control_allow_methods`"
+            " on the response to indicate which methods are allowed."
+        ),
+        read_only=True,
+    )
+
+    @property
+    def is_json(self) -> bool:
+        """Check if the mimetype indicates JSON data, either
+        :mimetype:`application/json` or :mimetype:`application/*+json`.
+        """
+        mt = self.mimetype
+        return (
+            mt == "application/json"
+            or mt.startswith("application/")
+            and mt.endswith("+json")
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/response.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/response.py
new file mode 100644
index 00000000..9fed0862
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/response.py
@@ -0,0 +1,763 @@
+from __future__ import annotations
+
+import typing as t
+from datetime import datetime
+from datetime import timedelta
+from datetime import timezone
+from http import HTTPStatus
+
+from ..datastructures import CallbackDict
+from ..datastructures import ContentRange
+from ..datastructures import ContentSecurityPolicy
+from ..datastructures import Headers
+from ..datastructures import HeaderSet
+from ..datastructures import ResponseCacheControl
+from ..datastructures import WWWAuthenticate
+from ..http import COEP
+from ..http import COOP
+from ..http import dump_age
+from ..http import dump_cookie
+from ..http import dump_header
+from ..http import dump_options_header
+from ..http import http_date
+from ..http import HTTP_STATUS_CODES
+from ..http import parse_age
+from ..http import parse_cache_control_header
+from ..http import parse_content_range_header
+from ..http import parse_csp_header
+from ..http import parse_date
+from ..http import parse_options_header
+from ..http import parse_set_header
+from ..http import quote_etag
+from ..http import unquote_etag
+from ..utils import get_content_type
+from ..utils import header_property
+
+if t.TYPE_CHECKING:
+    from ..datastructures.cache_control import _CacheControl
+
+
+def _set_property(name: str, doc: str | None = None) -> property:
+    def fget(self: Response) -> HeaderSet:
+        def on_update(header_set: HeaderSet) -> None:
+            if not header_set and name in self.headers:
+                del self.headers[name]
+            elif header_set:
+                self.headers[name] = header_set.to_header()
+
+        return parse_set_header(self.headers.get(name), on_update)
+
+    def fset(
+        self: Response,
+        value: None | (str | dict[str, str | int] | t.Iterable[str]),
+    ) -> None:
+        if not value:
+            del self.headers[name]
+        elif isinstance(value, str):
+            self.headers[name] = value
+        else:
+            self.headers[name] = dump_header(value)
+
+    return property(fget, fset, doc=doc)
+
+
+class Response:
+    """Represents the non-IO parts of an HTTP response, specifically the
+    status and headers but not the body.
+
+    This class is not meant for general use. It should only be used when
+    implementing WSGI, ASGI, or another HTTP application spec. Werkzeug
+    provides a WSGI implementation at :cls:`werkzeug.wrappers.Response`.
+
+    :param status: The status code for the response. Either an int, in
+        which case the default status message is added, or a string in
+        the form ``{code} {message}``, like ``404 Not Found``. Defaults
+        to 200.
+    :param headers: A :class:`~werkzeug.datastructures.Headers` object,
+        or a list of ``(key, value)`` tuples that will be converted to a
+        ``Headers`` object.
+    :param mimetype: The mime type (content type without charset or
+        other parameters) of the response. If the value starts with
+        ``text/`` (or matches some other special cases), the charset
+        will be added to create the ``content_type``.
+    :param content_type: The full content type of the response.
+        Overrides building the value from ``mimetype``.
+
+    .. versionchanged:: 3.0
+        The ``charset`` attribute was removed.
+
+    .. versionadded:: 2.0
+    """
+
+    #: the default status if none is provided.
+    default_status = 200
+
+    #: the default mimetype if none is provided.
+    default_mimetype: str | None = "text/plain"
+
+    #: Warn if a cookie header exceeds this size. The default, 4093, should be
+    #: safely `supported by most browsers <cookie_>`_. A cookie larger than
+    #: this size will still be sent, but it may be ignored or handled
+    #: incorrectly by some browsers. Set to 0 to disable this check.
+    #:
+    #: .. versionadded:: 0.13
+    #:
+    #: .. _`cookie`: http://browsercookielimits.squawky.net/
+    max_cookie_size = 4093
+
+    # A :class:`Headers` object representing the response headers.
+    headers: Headers
+
+    def __init__(
+        self,
+        status: int | str | HTTPStatus | None = None,
+        headers: t.Mapping[str, str | t.Iterable[str]]
+        | t.Iterable[tuple[str, str]]
+        | None = None,
+        mimetype: str | None = None,
+        content_type: str | None = None,
+    ) -> None:
+        if isinstance(headers, Headers):
+            self.headers = headers
+        elif not headers:
+            self.headers = Headers()
+        else:
+            self.headers = Headers(headers)
+
+        if content_type is None:
+            if mimetype is None and "content-type" not in self.headers:
+                mimetype = self.default_mimetype
+            if mimetype is not None:
+                mimetype = get_content_type(mimetype, "utf-8")
+            content_type = mimetype
+        if content_type is not None:
+            self.headers["Content-Type"] = content_type
+        if status is None:
+            status = self.default_status
+        self.status = status  # type: ignore
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} [{self.status}]>"
+
+    @property
+    def status_code(self) -> int:
+        """The HTTP status code as a number."""
+        return self._status_code
+
+    @status_code.setter
+    def status_code(self, code: int) -> None:
+        self.status = code  # type: ignore
+
+    @property
+    def status(self) -> str:
+        """The HTTP status code as a string."""
+        return self._status
+
+    @status.setter
+    def status(self, value: str | int | HTTPStatus) -> None:
+        self._status, self._status_code = self._clean_status(value)
+
+    def _clean_status(self, value: str | int | HTTPStatus) -> tuple[str, int]:
+        if isinstance(value, (int, HTTPStatus)):
+            status_code = int(value)
+        else:
+            value = value.strip()
+
+            if not value:
+                raise ValueError("Empty status argument")
+
+            code_str, sep, _ = value.partition(" ")
+
+            try:
+                status_code = int(code_str)
+            except ValueError:
+                # only message
+                return f"0 {value}", 0
+
+            if sep:
+                # code and message
+                return value, status_code
+
+        # only code, look up message
+        try:
+            status = f"{status_code} {HTTP_STATUS_CODES[status_code].upper()}"
+        except KeyError:
+            status = f"{status_code} UNKNOWN"
+
+        return status, status_code
+
+    def set_cookie(
+        self,
+        key: str,
+        value: str = "",
+        max_age: timedelta | int | None = None,
+        expires: str | datetime | int | float | None = None,
+        path: str | None = "/",
+        domain: str | None = None,
+        secure: bool = False,
+        httponly: bool = False,
+        samesite: str | None = None,
+        partitioned: bool = False,
+    ) -> None:
+        """Sets a cookie.
+
+        A warning is raised if the size of the cookie header exceeds
+        :attr:`max_cookie_size`, but the header will still be set.
+
+        :param key: the key (name) of the cookie to be set.
+        :param value: the value of the cookie.
+        :param max_age: should be a number of seconds, or `None` (default) if
+                        the cookie should last only as long as the client's
+                        browser session.
+        :param expires: should be a `datetime` object or UNIX timestamp.
+        :param path: limits the cookie to a given path, per default it will
+                     span the whole domain.
+        :param domain: if you want to set a cross-domain cookie.  For example,
+                       ``domain="example.com"`` will set a cookie that is
+                       readable by the domain ``www.example.com``,
+                       ``foo.example.com`` etc.  Otherwise, a cookie will only
+                       be readable by the domain that set it.
+        :param secure: If ``True``, the cookie will only be available
+            via HTTPS.
+        :param httponly: Disallow JavaScript access to the cookie.
+        :param samesite: Limit the scope of the cookie to only be
+            attached to requests that are "same-site".
+        :param partitioned: If ``True``, the cookie will be partitioned.
+
+        .. versionchanged:: 3.1
+            The ``partitioned`` parameter was added.
+        """
+        self.headers.add(
+            "Set-Cookie",
+            dump_cookie(
+                key,
+                value=value,
+                max_age=max_age,
+                expires=expires,
+                path=path,
+                domain=domain,
+                secure=secure,
+                httponly=httponly,
+                max_size=self.max_cookie_size,
+                samesite=samesite,
+                partitioned=partitioned,
+            ),
+        )
+
+    def delete_cookie(
+        self,
+        key: str,
+        path: str | None = "/",
+        domain: str | None = None,
+        secure: bool = False,
+        httponly: bool = False,
+        samesite: str | None = None,
+        partitioned: bool = False,
+    ) -> None:
+        """Delete a cookie.  Fails silently if key doesn't exist.
+
+        :param key: the key (name) of the cookie to be deleted.
+        :param path: if the cookie that should be deleted was limited to a
+                     path, the path has to be defined here.
+        :param domain: if the cookie that should be deleted was limited to a
+                       domain, that domain has to be defined here.
+        :param secure: If ``True``, the cookie will only be available
+            via HTTPS.
+        :param httponly: Disallow JavaScript access to the cookie.
+        :param samesite: Limit the scope of the cookie to only be
+            attached to requests that are "same-site".
+        :param partitioned: If ``True``, the cookie will be partitioned.
+        """
+        self.set_cookie(
+            key,
+            expires=0,
+            max_age=0,
+            path=path,
+            domain=domain,
+            secure=secure,
+            httponly=httponly,
+            samesite=samesite,
+            partitioned=partitioned,
+        )
+
+    @property
+    def is_json(self) -> bool:
+        """Check if the mimetype indicates JSON data, either
+        :mimetype:`application/json` or :mimetype:`application/*+json`.
+        """
+        mt = self.mimetype
+        return mt is not None and (
+            mt == "application/json"
+            or mt.startswith("application/")
+            and mt.endswith("+json")
+        )
+
+    # Common Descriptors
+
+    @property
+    def mimetype(self) -> str | None:
+        """The mimetype (content type without charset etc.)"""
+        ct = self.headers.get("content-type")
+
+        if ct:
+            return ct.split(";")[0].strip()
+        else:
+            return None
+
+    @mimetype.setter
+    def mimetype(self, value: str) -> None:
+        self.headers["Content-Type"] = get_content_type(value, "utf-8")
+
+    @property
+    def mimetype_params(self) -> dict[str, str]:
+        """The mimetype parameters as dict. For example if the
+        content type is ``text/html; charset=utf-8`` the params would be
+        ``{'charset': 'utf-8'}``.
+
+        .. versionadded:: 0.5
+        """
+
+        def on_update(d: CallbackDict[str, str]) -> None:
+            self.headers["Content-Type"] = dump_options_header(self.mimetype, d)
+
+        d = parse_options_header(self.headers.get("content-type", ""))[1]
+        return CallbackDict(d, on_update)
+
+    location = header_property[str](
+        "Location",
+        doc="""The Location response-header field is used to redirect
+        the recipient to a location other than the Request-URI for
+        completion of the request or identification of a new
+        resource.""",
+    )
+    age = header_property(
+        "Age",
+        None,
+        parse_age,
+        dump_age,  # type: ignore
+        doc="""The Age response-header field conveys the sender's
+        estimate of the amount of time since the response (or its
+        revalidation) was generated at the origin server.
+
+        Age values are non-negative decimal integers, representing time
+        in seconds.""",
+    )
+    content_type = header_property[str](
+        "Content-Type",
+        doc="""The Content-Type entity-header field indicates the media
+        type of the entity-body sent to the recipient or, in the case of
+        the HEAD method, the media type that would have been sent had
+        the request been a GET.""",
+    )
+    content_length = header_property(
+        "Content-Length",
+        None,
+        int,
+        str,
+        doc="""The Content-Length entity-header field indicates the size
+        of the entity-body, in decimal number of OCTETs, sent to the
+        recipient or, in the case of the HEAD method, the size of the
+        entity-body that would have been sent had the request been a
+        GET.""",
+    )
+    content_location = header_property[str](
+        "Content-Location",
+        doc="""The Content-Location entity-header field MAY be used to
+        supply the resource location for the entity enclosed in the
+        message when that entity is accessible from a location separate
+        from the requested resource's URI.""",
+    )
+    content_encoding = header_property[str](
+        "Content-Encoding",
+        doc="""The Content-Encoding entity-header field is used as a
+        modifier to the media-type. When present, its value indicates
+        what additional content codings have been applied to the
+        entity-body, and thus what decoding mechanisms must be applied
+        in order to obtain the media-type referenced by the Content-Type
+        header field.""",
+    )
+    content_md5 = header_property[str](
+        "Content-MD5",
+        doc="""The Content-MD5 entity-header field, as defined in
+        RFC 1864, is an MD5 digest of the entity-body for the purpose of
+        providing an end-to-end message integrity check (MIC) of the
+        entity-body. (Note: a MIC is good for detecting accidental
+        modification of the entity-body in transit, but is not proof
+        against malicious attacks.)""",
+    )
+    date = header_property(
+        "Date",
+        None,
+        parse_date,
+        http_date,
+        doc="""The Date general-header field represents the date and
+        time at which the message was originated, having the same
+        semantics as orig-date in RFC 822.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """,
+    )
+    expires = header_property(
+        "Expires",
+        None,
+        parse_date,
+        http_date,
+        doc="""The Expires entity-header field gives the date/time after
+        which the response is considered stale. A stale cache entry may
+        not normally be returned by a cache.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """,
+    )
+    last_modified = header_property(
+        "Last-Modified",
+        None,
+        parse_date,
+        http_date,
+        doc="""The Last-Modified entity-header field indicates the date
+        and time at which the origin server believes the variant was
+        last modified.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """,
+    )
+
+    @property
+    def retry_after(self) -> datetime | None:
+        """The Retry-After response-header field can be used with a
+        503 (Service Unavailable) response to indicate how long the
+        service is expected to be unavailable to the requesting client.
+
+        Time in seconds until expiration or date.
+
+        .. versionchanged:: 2.0
+            The datetime object is timezone-aware.
+        """
+        value = self.headers.get("retry-after")
+        if value is None:
+            return None
+
+        try:
+            seconds = int(value)
+        except ValueError:
+            return parse_date(value)
+
+        return datetime.now(timezone.utc) + timedelta(seconds=seconds)
+
+    @retry_after.setter
+    def retry_after(self, value: datetime | int | str | None) -> None:
+        if value is None:
+            if "retry-after" in self.headers:
+                del self.headers["retry-after"]
+            return
+        elif isinstance(value, datetime):
+            value = http_date(value)
+        else:
+            value = str(value)
+        self.headers["Retry-After"] = value
+
+    vary = _set_property(
+        "Vary",
+        doc="""The Vary field value indicates the set of request-header
+        fields that fully determines, while the response is fresh,
+        whether a cache is permitted to use the response to reply to a
+        subsequent request without revalidation.""",
+    )
+    content_language = _set_property(
+        "Content-Language",
+        doc="""The Content-Language entity-header field describes the
+        natural language(s) of the intended audience for the enclosed
+        entity. Note that this might not be equivalent to all the
+        languages used within the entity-body.""",
+    )
+    allow = _set_property(
+        "Allow",
+        doc="""The Allow entity-header field lists the set of methods
+        supported by the resource identified by the Request-URI. The
+        purpose of this field is strictly to inform the recipient of
+        valid methods associated with the resource. An Allow header
+        field MUST be present in a 405 (Method Not Allowed)
+        response.""",
+    )
+
+    # ETag
+
+    @property
+    def cache_control(self) -> ResponseCacheControl:
+        """The Cache-Control general-header field is used to specify
+        directives that MUST be obeyed by all caching mechanisms along the
+        request/response chain.
+        """
+
+        def on_update(cache_control: _CacheControl) -> None:
+            if not cache_control and "cache-control" in self.headers:
+                del self.headers["cache-control"]
+            elif cache_control:
+                self.headers["Cache-Control"] = cache_control.to_header()
+
+        return parse_cache_control_header(
+            self.headers.get("cache-control"), on_update, ResponseCacheControl
+        )
+
+    def set_etag(self, etag: str, weak: bool = False) -> None:
+        """Set the etag, and override the old one if there was one."""
+        self.headers["ETag"] = quote_etag(etag, weak)
+
+    def get_etag(self) -> tuple[str, bool] | tuple[None, None]:
+        """Return a tuple in the form ``(etag, is_weak)``.  If there is no
+        ETag the return value is ``(None, None)``.
+        """
+        return unquote_etag(self.headers.get("ETag"))
+
+    accept_ranges = header_property[str](
+        "Accept-Ranges",
+        doc="""The `Accept-Ranges` header. Even though the name would
+        indicate that multiple values are supported, it must be one
+        string token only.
+
+        The values ``'bytes'`` and ``'none'`` are common.
+
+        .. versionadded:: 0.7""",
+    )
+
+    @property
+    def content_range(self) -> ContentRange:
+        """The ``Content-Range`` header as a
+        :class:`~werkzeug.datastructures.ContentRange` object. Available
+        even if the header is not set.
+
+        .. versionadded:: 0.7
+        """
+
+        def on_update(rng: ContentRange) -> None:
+            if not rng:
+                del self.headers["content-range"]
+            else:
+                self.headers["Content-Range"] = rng.to_header()
+
+        rv = parse_content_range_header(self.headers.get("content-range"), on_update)
+        # always provide a content range object to make the descriptor
+        # more user friendly.  It provides an unset() method that can be
+        # used to remove the header quickly.
+        if rv is None:
+            rv = ContentRange(None, None, None, on_update=on_update)
+        return rv
+
+    @content_range.setter
+    def content_range(self, value: ContentRange | str | None) -> None:
+        if not value:
+            del self.headers["content-range"]
+        elif isinstance(value, str):
+            self.headers["Content-Range"] = value
+        else:
+            self.headers["Content-Range"] = value.to_header()
+
+    # Authorization
+
+    @property
+    def www_authenticate(self) -> WWWAuthenticate:
+        """The ``WWW-Authenticate`` header parsed into a :class:`.WWWAuthenticate`
+        object. Modifying the object will modify the header value.
+
+        This header is not set by default. To set this header, assign an instance of
+        :class:`.WWWAuthenticate` to this attribute.
+
+        .. code-block:: python
+
+            response.www_authenticate = WWWAuthenticate(
+                "basic", {"realm": "Authentication Required"}
+            )
+
+        Multiple values for this header can be sent to give the client multiple options.
+        Assign a list to set multiple headers. However, modifying the items in the list
+        will not automatically update the header values, and accessing this attribute
+        will only ever return the first value.
+
+        To unset this header, assign ``None`` or use ``del``.
+
+        .. versionchanged:: 2.3
+            This attribute can be assigned to to set the header. A list can be assigned
+            to set multiple header values. Use ``del`` to unset the header.
+
+        .. versionchanged:: 2.3
+            :class:`WWWAuthenticate` is no longer a ``dict``. The ``token`` attribute
+            was added for auth challenges that use a token instead of parameters.
+        """
+        value = WWWAuthenticate.from_header(self.headers.get("WWW-Authenticate"))
+
+        if value is None:
+            value = WWWAuthenticate("basic")
+
+        def on_update(value: WWWAuthenticate) -> None:
+            self.www_authenticate = value
+
+        value._on_update = on_update
+        return value
+
+    @www_authenticate.setter
+    def www_authenticate(
+        self, value: WWWAuthenticate | list[WWWAuthenticate] | None
+    ) -> None:
+        if not value:  # None or empty list
+            del self.www_authenticate
+        elif isinstance(value, list):
+            # Clear any existing header by setting the first item.
+            self.headers.set("WWW-Authenticate", value[0].to_header())
+
+            for item in value[1:]:
+                # Add additional header lines for additional items.
+                self.headers.add("WWW-Authenticate", item.to_header())
+        else:
+            self.headers.set("WWW-Authenticate", value.to_header())
+
+            def on_update(value: WWWAuthenticate) -> None:
+                self.www_authenticate = value
+
+            # When setting a single value, allow updating it directly.
+            value._on_update = on_update
+
+    @www_authenticate.deleter
+    def www_authenticate(self) -> None:
+        if "WWW-Authenticate" in self.headers:
+            del self.headers["WWW-Authenticate"]
+
+    # CSP
+
+    @property
+    def content_security_policy(self) -> ContentSecurityPolicy:
+        """The ``Content-Security-Policy`` header as a
+        :class:`~werkzeug.datastructures.ContentSecurityPolicy` object. Available
+        even if the header is not set.
+
+        The Content-Security-Policy header adds an additional layer of
+        security to help detect and mitigate certain types of attacks.
+        """
+
+        def on_update(csp: ContentSecurityPolicy) -> None:
+            if not csp:
+                del self.headers["content-security-policy"]
+            else:
+                self.headers["Content-Security-Policy"] = csp.to_header()
+
+        rv = parse_csp_header(self.headers.get("content-security-policy"), on_update)
+        if rv is None:
+            rv = ContentSecurityPolicy(None, on_update=on_update)
+        return rv
+
+    @content_security_policy.setter
+    def content_security_policy(
+        self, value: ContentSecurityPolicy | str | None
+    ) -> None:
+        if not value:
+            del self.headers["content-security-policy"]
+        elif isinstance(value, str):
+            self.headers["Content-Security-Policy"] = value
+        else:
+            self.headers["Content-Security-Policy"] = value.to_header()
+
+    @property
+    def content_security_policy_report_only(self) -> ContentSecurityPolicy:
+        """The ``Content-Security-policy-report-only`` header as a
+        :class:`~werkzeug.datastructures.ContentSecurityPolicy` object. Available
+        even if the header is not set.
+
+        The Content-Security-Policy-Report-Only header adds a csp policy
+        that is not enforced but is reported thereby helping detect
+        certain types of attacks.
+        """
+
+        def on_update(csp: ContentSecurityPolicy) -> None:
+            if not csp:
+                del self.headers["content-security-policy-report-only"]
+            else:
+                self.headers["Content-Security-policy-report-only"] = csp.to_header()
+
+        rv = parse_csp_header(
+            self.headers.get("content-security-policy-report-only"), on_update
+        )
+        if rv is None:
+            rv = ContentSecurityPolicy(None, on_update=on_update)
+        return rv
+
+    @content_security_policy_report_only.setter
+    def content_security_policy_report_only(
+        self, value: ContentSecurityPolicy | str | None
+    ) -> None:
+        if not value:
+            del self.headers["content-security-policy-report-only"]
+        elif isinstance(value, str):
+            self.headers["Content-Security-policy-report-only"] = value
+        else:
+            self.headers["Content-Security-policy-report-only"] = value.to_header()
+
+    # CORS
+
+    @property
+    def access_control_allow_credentials(self) -> bool:
+        """Whether credentials can be shared by the browser to
+        JavaScript code. As part of the preflight request it indicates
+        whether credentials can be used on the cross origin request.
+        """
+        return "Access-Control-Allow-Credentials" in self.headers
+
+    @access_control_allow_credentials.setter
+    def access_control_allow_credentials(self, value: bool | None) -> None:
+        if value is True:
+            self.headers["Access-Control-Allow-Credentials"] = "true"
+        else:
+            self.headers.pop("Access-Control-Allow-Credentials", None)
+
+    access_control_allow_headers = header_property(
+        "Access-Control-Allow-Headers",
+        load_func=parse_set_header,
+        dump_func=dump_header,
+        doc="Which headers can be sent with the cross origin request.",
+    )
+
+    access_control_allow_methods = header_property(
+        "Access-Control-Allow-Methods",
+        load_func=parse_set_header,
+        dump_func=dump_header,
+        doc="Which methods can be used for the cross origin request.",
+    )
+
+    access_control_allow_origin = header_property[str](
+        "Access-Control-Allow-Origin",
+        doc="The origin or '*' for any origin that may make cross origin requests.",
+    )
+
+    access_control_expose_headers = header_property(
+        "Access-Control-Expose-Headers",
+        load_func=parse_set_header,
+        dump_func=dump_header,
+        doc="Which headers can be shared by the browser to JavaScript code.",
+    )
+
+    access_control_max_age = header_property(
+        "Access-Control-Max-Age",
+        load_func=int,
+        dump_func=str,
+        doc="The maximum age in seconds the access control settings can be cached for.",
+    )
+
+    cross_origin_opener_policy = header_property[COOP](
+        "Cross-Origin-Opener-Policy",
+        load_func=lambda value: COOP(value),
+        dump_func=lambda value: value.value,
+        default=COOP.UNSAFE_NONE,
+        doc="""Allows control over sharing of browsing context group with cross-origin
+        documents. Values must be a member of the :class:`werkzeug.http.COOP` enum.""",
+    )
+
+    cross_origin_embedder_policy = header_property[COEP](
+        "Cross-Origin-Embedder-Policy",
+        load_func=lambda value: COEP(value),
+        dump_func=lambda value: value.value,
+        default=COEP.UNSAFE_NONE,
+        doc="""Prevents a document from loading any cross-origin resources that do not
+        explicitly grant the document permission. Values must be a member of the
+        :class:`werkzeug.http.COEP` enum.""",
+    )
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/utils.py b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/utils.py
new file mode 100644
index 00000000..ff7ceda3
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/sansio/utils.py
@@ -0,0 +1,167 @@
+from __future__ import annotations
+
+import typing as t
+from urllib.parse import quote
+
+from .._internal import _plain_int
+from ..exceptions import SecurityError
+from ..urls import uri_to_iri
+
+
+def host_is_trusted(hostname: str | None, trusted_list: t.Iterable[str]) -> bool:
+    """Check if a host matches a list of trusted names.
+
+    :param hostname: The name to check.
+    :param trusted_list: A list of valid names to match. If a name
+        starts with a dot it will match all subdomains.
+
+    .. versionadded:: 0.9
+    """
+    if not hostname:
+        return False
+
+    try:
+        hostname = hostname.partition(":")[0].encode("idna").decode("ascii")
+    except UnicodeEncodeError:
+        return False
+
+    if isinstance(trusted_list, str):
+        trusted_list = [trusted_list]
+
+    for ref in trusted_list:
+        if ref.startswith("."):
+            ref = ref[1:]
+            suffix_match = True
+        else:
+            suffix_match = False
+
+        try:
+            ref = ref.partition(":")[0].encode("idna").decode("ascii")
+        except UnicodeEncodeError:
+            return False
+
+        if ref == hostname or (suffix_match and hostname.endswith(f".{ref}")):
+            return True
+
+    return False
+
+
+def get_host(
+    scheme: str,
+    host_header: str | None,
+    server: tuple[str, int | None] | None = None,
+    trusted_hosts: t.Iterable[str] | None = None,
+) -> str:
+    """Return the host for the given parameters.
+
+    This first checks the ``host_header``. If it's not present, then
+    ``server`` is used. The host will only contain the port if it is
+    different than the standard port for the protocol.
+
+    Optionally, verify that the host is trusted using
+    :func:`host_is_trusted` and raise a
+    :exc:`~werkzeug.exceptions.SecurityError` if it is not.
+
+    :param scheme: The protocol the request used, like ``"https"``.
+    :param host_header: The ``Host`` header value.
+    :param server: Address of the server. ``(host, port)``, or
+        ``(path, None)`` for unix sockets.
+    :param trusted_hosts: A list of trusted host names.
+
+    :return: Host, with port if necessary.
+    :raise ~werkzeug.exceptions.SecurityError: If the host is not
+        trusted.
+
+    .. versionchanged:: 3.1.3
+        If ``SERVER_NAME`` is IPv6, it is wrapped in ``[]``.
+    """
+    host = ""
+
+    if host_header is not None:
+        host = host_header
+    elif server is not None:
+        host = server[0]
+
+        # If SERVER_NAME is IPv6, wrap it in [] to match Host header.
+        # Check for : because domain or IPv4 can't have that.
+        if ":" in host and host[0] != "[":
+            host = f"[{host}]"
+
+        if server[1] is not None:
+            host = f"{host}:{server[1]}"
+
+    if scheme in {"http", "ws"} and host.endswith(":80"):
+        host = host[:-3]
+    elif scheme in {"https", "wss"} and host.endswith(":443"):
+        host = host[:-4]
+
+    if trusted_hosts is not None:
+        if not host_is_trusted(host, trusted_hosts):
+            raise SecurityError(f"Host {host!r} is not trusted.")
+
+    return host
+
+
+def get_current_url(
+    scheme: str,
+    host: str,
+    root_path: str | None = None,
+    path: str | None = None,
+    query_string: bytes | None = None,
+) -> str:
+    """Recreate the URL for a request. If an optional part isn't
+    provided, it and subsequent parts are not included in the URL.
+
+    The URL is an IRI, not a URI, so it may contain Unicode characters.
+    Use :func:`~werkzeug.urls.iri_to_uri` to convert it to ASCII.
+
+    :param scheme: The protocol the request used, like ``"https"``.
+    :param host: The host the request was made to. See :func:`get_host`.
+    :param root_path: Prefix that the application is mounted under. This
+        is prepended to ``path``.
+    :param path: The path part of the URL after ``root_path``.
+    :param query_string: The portion of the URL after the "?".
+    """
+    url = [scheme, "://", host]
+
+    if root_path is None:
+        url.append("/")
+        return uri_to_iri("".join(url))
+
+    # safe = https://url.spec.whatwg.org/#url-path-segment-string
+    # as well as percent for things that are already quoted
+    url.append(quote(root_path.rstrip("/"), safe="!$&'()*+,/:;=@%"))
+    url.append("/")
+
+    if path is None:
+        return uri_to_iri("".join(url))
+
+    url.append(quote(path.lstrip("/"), safe="!$&'()*+,/:;=@%"))
+
+    if query_string:
+        url.append("?")
+        url.append(quote(query_string, safe="!$&'()*+,/:;=?@%"))
+
+    return uri_to_iri("".join(url))
+
+
+def get_content_length(
+    http_content_length: str | None = None,
+    http_transfer_encoding: str | None = None,
+) -> int | None:
+    """Return the ``Content-Length`` header value as an int. If the header is not given
+    or the ``Transfer-Encoding`` header is ``chunked``, ``None`` is returned to indicate
+    a streaming request. If the value is not an integer, or negative, 0 is returned.
+
+    :param http_content_length: The Content-Length HTTP header.
+    :param http_transfer_encoding: The Transfer-Encoding HTTP header.
+
+    .. versionadded:: 2.2
+    """
+    if http_transfer_encoding == "chunked" or http_content_length is None:
+        return None
+
+    try:
+        return max(0, _plain_int(http_content_length))
+    except ValueError:
+        return 0
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/security.py b/backend/venv/lib/python3.12/site-packages/werkzeug/security.py
new file mode 100644
index 00000000..3f49ad1b
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/security.py
@@ -0,0 +1,166 @@
+from __future__ import annotations
+
+import hashlib
+import hmac
+import os
+import posixpath
+import secrets
+
+SALT_CHARS = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"
+DEFAULT_PBKDF2_ITERATIONS = 1_000_000
+
+_os_alt_seps: list[str] = list(
+    sep for sep in [os.sep, os.path.altsep] if sep is not None and sep != "/"
+)
+
+
+def gen_salt(length: int) -> str:
+    """Generate a random string of SALT_CHARS with specified ``length``."""
+    if length <= 0:
+        raise ValueError("Salt length must be at least 1.")
+
+    return "".join(secrets.choice(SALT_CHARS) for _ in range(length))
+
+
+def _hash_internal(method: str, salt: str, password: str) -> tuple[str, str]:
+    method, *args = method.split(":")
+    salt_bytes = salt.encode()
+    password_bytes = password.encode()
+
+    if method == "scrypt":
+        if not args:
+            n = 2**15
+            r = 8
+            p = 1
+        else:
+            try:
+                n, r, p = map(int, args)
+            except ValueError:
+                raise ValueError("'scrypt' takes 3 arguments.") from None
+
+        maxmem = 132 * n * r * p  # ideally 128, but some extra seems needed
+        return (
+            hashlib.scrypt(
+                password_bytes, salt=salt_bytes, n=n, r=r, p=p, maxmem=maxmem
+            ).hex(),
+            f"scrypt:{n}:{r}:{p}",
+        )
+    elif method == "pbkdf2":
+        len_args = len(args)
+
+        if len_args == 0:
+            hash_name = "sha256"
+            iterations = DEFAULT_PBKDF2_ITERATIONS
+        elif len_args == 1:
+            hash_name = args[0]
+            iterations = DEFAULT_PBKDF2_ITERATIONS
+        elif len_args == 2:
+            hash_name = args[0]
+            iterations = int(args[1])
+        else:
+            raise ValueError("'pbkdf2' takes 2 arguments.")
+
+        return (
+            hashlib.pbkdf2_hmac(
+                hash_name, password_bytes, salt_bytes, iterations
+            ).hex(),
+            f"pbkdf2:{hash_name}:{iterations}",
+        )
+    else:
+        raise ValueError(f"Invalid hash method '{method}'.")
+
+
+def generate_password_hash(
+    password: str, method: str = "scrypt", salt_length: int = 16
+) -> str:
+    """Securely hash a password for storage. A password can be compared to a stored hash
+    using :func:`check_password_hash`.
+
+    The following methods are supported:
+
+    -   ``scrypt``, the default. The parameters are ``n``, ``r``, and ``p``, the default
+        is ``scrypt:32768:8:1``. See :func:`hashlib.scrypt`.
+    -   ``pbkdf2``, less secure. The parameters are ``hash_method`` and ``iterations``,
+        the default is ``pbkdf2:sha256:600000``. See :func:`hashlib.pbkdf2_hmac`.
+
+    Default parameters may be updated to reflect current guidelines, and methods may be
+    deprecated and removed if they are no longer considered secure. To migrate old
+    hashes, you may generate a new hash when checking an old hash, or you may contact
+    users with a link to reset their password.
+
+    :param password: The plaintext password.
+    :param method: The key derivation function and parameters.
+    :param salt_length: The number of characters to generate for the salt.
+
+    .. versionchanged:: 3.1
+        The default iterations for pbkdf2 was increased to 1,000,000.
+
+    .. versionchanged:: 2.3
+        Scrypt support was added.
+
+    .. versionchanged:: 2.3
+        The default iterations for pbkdf2 was increased to 600,000.
+
+    .. versionchanged:: 2.3
+        All plain hashes are deprecated and will not be supported in Werkzeug 3.0.
+    """
+    salt = gen_salt(salt_length)
+    h, actual_method = _hash_internal(method, salt, password)
+    return f"{actual_method}${salt}${h}"
+
+
+def check_password_hash(pwhash: str, password: str) -> bool:
+    """Securely check that the given stored password hash, previously generated using
+    :func:`generate_password_hash`, matches the given password.
+
+    Methods may be deprecated and removed if they are no longer considered secure. To
+    migrate old hashes, you may generate a new hash when checking an old hash, or you
+    may contact users with a link to reset their password.
+
+    :param pwhash: The hashed password.
+    :param password: The plaintext password.
+
+    .. versionchanged:: 2.3
+        All plain hashes are deprecated and will not be supported in Werkzeug 3.0.
+    """
+    try:
+        method, salt, hashval = pwhash.split("$", 2)
+    except ValueError:
+        return False
+
+    return hmac.compare_digest(_hash_internal(method, salt, password)[0], hashval)
+
+
+def safe_join(directory: str, *pathnames: str) -> str | None:
+    """Safely join zero or more untrusted path components to a base
+    directory to avoid escaping the base directory.
+
+    :param directory: The trusted base directory.
+    :param pathnames: The untrusted path components relative to the
+        base directory.
+    :return: A safe path, otherwise ``None``.
+    """
+    if not directory:
+        # Ensure we end up with ./path if directory="" is given,
+        # otherwise the first untrusted part could become trusted.
+        directory = "."
+
+    parts = [directory]
+
+    for filename in pathnames:
+        if filename != "":
+            filename = posixpath.normpath(filename)
+
+        if (
+            any(sep in filename for sep in _os_alt_seps)
+            or os.path.isabs(filename)
+            # ntpath.isabs doesn't catch this on Python < 3.11
+            or filename.startswith("/")
+            or filename == ".."
+            or filename.startswith("../")
+        ):
+            return None
+
+        parts.append(filename)
+
+    return posixpath.join(*parts)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/serving.py b/backend/venv/lib/python3.12/site-packages/werkzeug/serving.py
new file mode 100644
index 00000000..ec166408
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/serving.py
@@ -0,0 +1,1125 @@
+"""A WSGI and HTTP server for use **during development only**. This
+server is convenient to use, but is not designed to be particularly
+stable, secure, or efficient. Use a dedicate WSGI server and HTTP
+server when deploying to production.
+
+It provides features like interactive debugging and code reloading. Use
+``run_simple`` to start the server. Put this in a ``run.py`` script:
+
+.. code-block:: python
+
+    from myapp import create_app
+    from werkzeug import run_simple
+"""
+
+from __future__ import annotations
+
+import errno
+import io
+import os
+import selectors
+import socket
+import socketserver
+import sys
+import typing as t
+from datetime import datetime as dt
+from datetime import timedelta
+from datetime import timezone
+from http.server import BaseHTTPRequestHandler
+from http.server import HTTPServer
+from urllib.parse import unquote
+from urllib.parse import urlsplit
+
+from ._internal import _log
+from ._internal import _wsgi_encoding_dance
+from .exceptions import InternalServerError
+from .urls import uri_to_iri
+
+try:
+    import ssl
+
+    connection_dropped_errors: tuple[type[Exception], ...] = (
+        ConnectionError,
+        socket.timeout,
+        ssl.SSLEOFError,
+    )
+except ImportError:
+
+    class _SslDummy:
+        def __getattr__(self, name: str) -> t.Any:
+            raise RuntimeError(  # noqa: B904
+                "SSL is unavailable because this Python runtime was not"
+                " compiled with SSL/TLS support."
+            )
+
+    ssl = _SslDummy()  # type: ignore
+    connection_dropped_errors = (ConnectionError, socket.timeout)
+
+_log_add_style = True
+
+if os.name == "nt":
+    try:
+        __import__("colorama")
+    except ImportError:
+        _log_add_style = False
+
+can_fork = hasattr(os, "fork")
+
+if can_fork:
+    ForkingMixIn = socketserver.ForkingMixIn
+else:
+
+    class ForkingMixIn:  # type: ignore
+        pass
+
+
+try:
+    af_unix = socket.AF_UNIX
+except AttributeError:
+    af_unix = None  # type: ignore
+
+LISTEN_QUEUE = 128
+
+_TSSLContextArg = t.Optional[
+    t.Union["ssl.SSLContext", tuple[str, t.Optional[str]], t.Literal["adhoc"]]
+]
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+    from cryptography.hazmat.primitives.asymmetric.rsa import (
+        RSAPrivateKeyWithSerialization,
+    )
+    from cryptography.x509 import Certificate
+
+
+class DechunkedInput(io.RawIOBase):
+    """An input stream that handles Transfer-Encoding 'chunked'"""
+
+    def __init__(self, rfile: t.IO[bytes]) -> None:
+        self._rfile = rfile
+        self._done = False
+        self._len = 0
+
+    def readable(self) -> bool:
+        return True
+
+    def read_chunk_len(self) -> int:
+        try:
+            line = self._rfile.readline().decode("latin1")
+            _len = int(line.strip(), 16)
+        except ValueError as e:
+            raise OSError("Invalid chunk header") from e
+        if _len < 0:
+            raise OSError("Negative chunk length not allowed")
+        return _len
+
+    def readinto(self, buf: bytearray) -> int:  # type: ignore
+        read = 0
+        while not self._done and read < len(buf):
+            if self._len == 0:
+                # This is the first chunk or we fully consumed the previous
+                # one. Read the next length of the next chunk
+                self._len = self.read_chunk_len()
+
+            if self._len == 0:
+                # Found the final chunk of size 0. The stream is now exhausted,
+                # but there is still a final newline that should be consumed
+                self._done = True
+
+            if self._len > 0:
+                # There is data (left) in this chunk, so append it to the
+                # buffer. If this operation fully consumes the chunk, this will
+                # reset self._len to 0.
+                n = min(len(buf), self._len)
+
+                # If (read + chunk size) becomes more than len(buf), buf will
+                # grow beyond the original size and read more data than
+                # required. So only read as much data as can fit in buf.
+                if read + n > len(buf):
+                    buf[read:] = self._rfile.read(len(buf) - read)
+                    self._len -= len(buf) - read
+                    read = len(buf)
+                else:
+                    buf[read : read + n] = self._rfile.read(n)
+                    self._len -= n
+                    read += n
+
+            if self._len == 0:
+                # Skip the terminating newline of a chunk that has been fully
+                # consumed. This also applies to the 0-sized final chunk
+                terminator = self._rfile.readline()
+                if terminator not in (b"\n", b"\r\n", b"\r"):
+                    raise OSError("Missing chunk terminating newline")
+
+        return read
+
+
+class WSGIRequestHandler(BaseHTTPRequestHandler):
+    """A request handler that implements WSGI dispatching."""
+
+    server: BaseWSGIServer
+
+    @property
+    def server_version(self) -> str:  # type: ignore
+        return self.server._server_version
+
+    def make_environ(self) -> WSGIEnvironment:
+        request_url = urlsplit(self.path)
+        url_scheme = "http" if self.server.ssl_context is None else "https"
+
+        if not self.client_address:
+            self.client_address = ("<local>", 0)
+        elif isinstance(self.client_address, str):
+            self.client_address = (self.client_address, 0)
+
+        # If there was no scheme but the path started with two slashes,
+        # the first segment may have been incorrectly parsed as the
+        # netloc, prepend it to the path again.
+        if not request_url.scheme and request_url.netloc:
+            path_info = f"/{request_url.netloc}{request_url.path}"
+        else:
+            path_info = request_url.path
+
+        path_info = unquote(path_info)
+
+        environ: WSGIEnvironment = {
+            "wsgi.version": (1, 0),
+            "wsgi.url_scheme": url_scheme,
+            "wsgi.input": self.rfile,
+            "wsgi.errors": sys.stderr,
+            "wsgi.multithread": self.server.multithread,
+            "wsgi.multiprocess": self.server.multiprocess,
+            "wsgi.run_once": False,
+            "werkzeug.socket": self.connection,
+            "SERVER_SOFTWARE": self.server_version,
+            "REQUEST_METHOD": self.command,
+            "SCRIPT_NAME": "",
+            "PATH_INFO": _wsgi_encoding_dance(path_info),
+            "QUERY_STRING": _wsgi_encoding_dance(request_url.query),
+            # Non-standard, added by mod_wsgi, uWSGI
+            "REQUEST_URI": _wsgi_encoding_dance(self.path),
+            # Non-standard, added by gunicorn
+            "RAW_URI": _wsgi_encoding_dance(self.path),
+            "REMOTE_ADDR": self.address_string(),
+            "REMOTE_PORT": self.port_integer(),
+            "SERVER_NAME": self.server.server_address[0],
+            "SERVER_PORT": str(self.server.server_address[1]),
+            "SERVER_PROTOCOL": self.request_version,
+        }
+
+        for key, value in self.headers.items():
+            if "_" in key:
+                continue
+
+            key = key.upper().replace("-", "_")
+            value = value.replace("\r\n", "")
+            if key not in ("CONTENT_TYPE", "CONTENT_LENGTH"):
+                key = f"HTTP_{key}"
+                if key in environ:
+                    value = f"{environ[key]},{value}"
+            environ[key] = value
+
+        if environ.get("HTTP_TRANSFER_ENCODING", "").strip().lower() == "chunked":
+            environ["wsgi.input_terminated"] = True
+            environ["wsgi.input"] = DechunkedInput(environ["wsgi.input"])
+
+        # Per RFC 2616, if the URL is absolute, use that as the host.
+        # We're using "has a scheme" to indicate an absolute URL.
+        if request_url.scheme and request_url.netloc:
+            environ["HTTP_HOST"] = request_url.netloc
+
+        try:
+            # binary_form=False gives nicer information, but wouldn't be compatible with
+            # what Nginx or Apache could return.
+            peer_cert = self.connection.getpeercert(binary_form=True)
+            if peer_cert is not None:
+                # Nginx and Apache use PEM format.
+                environ["SSL_CLIENT_CERT"] = ssl.DER_cert_to_PEM_cert(peer_cert)
+        except ValueError:
+            # SSL handshake hasn't finished.
+            self.server.log("error", "Cannot fetch SSL peer certificate info")
+        except AttributeError:
+            # Not using TLS, the socket will not have getpeercert().
+            pass
+
+        return environ
+
+    def run_wsgi(self) -> None:
+        if self.headers.get("Expect", "").lower().strip() == "100-continue":
+            self.wfile.write(b"HTTP/1.1 100 Continue\r\n\r\n")
+
+        self.environ = environ = self.make_environ()
+        status_set: str | None = None
+        headers_set: list[tuple[str, str]] | None = None
+        status_sent: str | None = None
+        headers_sent: list[tuple[str, str]] | None = None
+        chunk_response: bool = False
+
+        def write(data: bytes) -> None:
+            nonlocal status_sent, headers_sent, chunk_response
+            assert status_set is not None, "write() before start_response"
+            assert headers_set is not None, "write() before start_response"
+            if status_sent is None:
+                status_sent = status_set
+                headers_sent = headers_set
+                try:
+                    code_str, msg = status_sent.split(None, 1)
+                except ValueError:
+                    code_str, msg = status_sent, ""
+                code = int(code_str)
+                self.send_response(code, msg)
+                header_keys = set()
+                for key, value in headers_sent:
+                    self.send_header(key, value)
+                    header_keys.add(key.lower())
+
+                # Use chunked transfer encoding if there is no content
+                # length. Do not use for 1xx and 204 responses. 304
+                # responses and HEAD requests are also excluded, which
+                # is the more conservative behavior and matches other
+                # parts of the code.
+                # https://httpwg.org/specs/rfc7230.html#rfc.section.3.3.1
+                if (
+                    not (
+                        "content-length" in header_keys
+                        or environ["REQUEST_METHOD"] == "HEAD"
+                        or (100 <= code < 200)
+                        or code in {204, 304}
+                    )
+                    and self.protocol_version >= "HTTP/1.1"
+                ):
+                    chunk_response = True
+                    self.send_header("Transfer-Encoding", "chunked")
+
+                # Always close the connection. This disables HTTP/1.1
+                # keep-alive connections. They aren't handled well by
+                # Python's http.server because it doesn't know how to
+                # drain the stream before the next request line.
+                self.send_header("Connection", "close")
+                self.end_headers()
+
+            assert isinstance(data, bytes), "applications must write bytes"
+
+            if data:
+                if chunk_response:
+                    self.wfile.write(hex(len(data))[2:].encode())
+                    self.wfile.write(b"\r\n")
+
+                self.wfile.write(data)
+
+                if chunk_response:
+                    self.wfile.write(b"\r\n")
+
+            self.wfile.flush()
+
+        def start_response(status, headers, exc_info=None):  # type: ignore
+            nonlocal status_set, headers_set
+            if exc_info:
+                try:
+                    if headers_sent:
+                        raise exc_info[1].with_traceback(exc_info[2])
+                finally:
+                    exc_info = None
+            elif headers_set:
+                raise AssertionError("Headers already set")
+            status_set = status
+            headers_set = headers
+            return write
+
+        def execute(app: WSGIApplication) -> None:
+            application_iter = app(environ, start_response)
+            try:
+                for data in application_iter:
+                    write(data)
+                if not headers_sent:
+                    write(b"")
+                if chunk_response:
+                    self.wfile.write(b"0\r\n\r\n")
+            finally:
+                # Check for any remaining data in the read socket, and discard it. This
+                # will read past request.max_content_length, but lets the client see a
+                # 413 response instead of a connection reset failure. If we supported
+                # keep-alive connections, this naive approach would break by reading the
+                # next request line. Since we know that write (above) closes every
+                # connection we can read everything.
+                selector = selectors.DefaultSelector()
+                selector.register(self.connection, selectors.EVENT_READ)
+                total_size = 0
+                total_reads = 0
+
+                # A timeout of 0 tends to fail because a client needs a small amount of
+                # time to continue sending its data.
+                while selector.select(timeout=0.01):
+                    # Only read 10MB into memory at a time.
+                    data = self.rfile.read(10_000_000)
+                    total_size += len(data)
+                    total_reads += 1
+
+                    # Stop reading on no data, >=10GB, or 1000 reads. If a client sends
+                    # more than that, they'll get a connection reset failure.
+                    if not data or total_size >= 10_000_000_000 or total_reads > 1000:
+                        break
+
+                selector.close()
+
+                if hasattr(application_iter, "close"):
+                    application_iter.close()
+
+        try:
+            execute(self.server.app)
+        except connection_dropped_errors as e:
+            self.connection_dropped(e, environ)
+        except Exception as e:
+            if self.server.passthrough_errors:
+                raise
+
+            if status_sent is not None and chunk_response:
+                self.close_connection = True
+
+            try:
+                # if we haven't yet sent the headers but they are set
+                # we roll back to be able to set them again.
+                if status_sent is None:
+                    status_set = None
+                    headers_set = None
+                execute(InternalServerError())
+            except Exception:
+                pass
+
+            from .debug.tbtools import DebugTraceback
+
+            msg = DebugTraceback(e).render_traceback_text()
+            self.server.log("error", f"Error on request:\n{msg}")
+
+    def handle(self) -> None:
+        """Handles a request ignoring dropped connections."""
+        try:
+            super().handle()
+        except (ConnectionError, socket.timeout) as e:
+            self.connection_dropped(e)
+        except Exception as e:
+            if self.server.ssl_context is not None and is_ssl_error(e):
+                self.log_error("SSL error occurred: %s", e)
+            else:
+                raise
+
+    def connection_dropped(
+        self, error: BaseException, environ: WSGIEnvironment | None = None
+    ) -> None:
+        """Called if the connection was closed by the client.  By default
+        nothing happens.
+        """
+
+    def __getattr__(self, name: str) -> t.Any:
+        # All HTTP methods are handled by run_wsgi.
+        if name.startswith("do_"):
+            return self.run_wsgi
+
+        # All other attributes are forwarded to the base class.
+        return getattr(super(), name)
+
+    def address_string(self) -> str:
+        if getattr(self, "environ", None):
+            return self.environ["REMOTE_ADDR"]  # type: ignore
+
+        if not self.client_address:
+            return "<local>"
+
+        return self.client_address[0]
+
+    def port_integer(self) -> int:
+        return self.client_address[1]
+
+    # Escape control characters. This is defined (but private) in Python 3.12.
+    _control_char_table = str.maketrans(
+        {c: rf"\x{c:02x}" for c in [*range(0x20), *range(0x7F, 0xA0)]}
+    )
+    _control_char_table[ord("\\")] = r"\\"
+
+    def log_request(self, code: int | str = "-", size: int | str = "-") -> None:
+        try:
+            path = uri_to_iri(self.path)
+            msg = f"{self.command} {path} {self.request_version}"
+        except AttributeError:
+            # path isn't set if the requestline was bad
+            msg = self.requestline
+
+        # Escape control characters that may be in the decoded path.
+        msg = msg.translate(self._control_char_table)
+        code = str(code)
+
+        if code[0] == "1":  # 1xx - Informational
+            msg = _ansi_style(msg, "bold")
+        elif code == "200":  # 2xx - Success
+            pass
+        elif code == "304":  # 304 - Resource Not Modified
+            msg = _ansi_style(msg, "cyan")
+        elif code[0] == "3":  # 3xx - Redirection
+            msg = _ansi_style(msg, "green")
+        elif code == "404":  # 404 - Resource Not Found
+            msg = _ansi_style(msg, "yellow")
+        elif code[0] == "4":  # 4xx - Client Error
+            msg = _ansi_style(msg, "bold", "red")
+        else:  # 5xx, or any other response
+            msg = _ansi_style(msg, "bold", "magenta")
+
+        self.log("info", '"%s" %s %s', msg, code, size)
+
+    def log_error(self, format: str, *args: t.Any) -> None:
+        self.log("error", format, *args)
+
+    def log_message(self, format: str, *args: t.Any) -> None:
+        self.log("info", format, *args)
+
+    def log(self, type: str, message: str, *args: t.Any) -> None:
+        # an IPv6 scoped address contains "%" which breaks logging
+        address_string = self.address_string().replace("%", "%%")
+        _log(
+            type,
+            f"{address_string} - - [{self.log_date_time_string()}] {message}\n",
+            *args,
+        )
+
+
+def _ansi_style(value: str, *styles: str) -> str:
+    if not _log_add_style:
+        return value
+
+    codes = {
+        "bold": 1,
+        "red": 31,
+        "green": 32,
+        "yellow": 33,
+        "magenta": 35,
+        "cyan": 36,
+    }
+
+    for style in styles:
+        value = f"\x1b[{codes[style]}m{value}"
+
+    return f"{value}\x1b[0m"
+
+
+def generate_adhoc_ssl_pair(
+    cn: str | None = None,
+) -> tuple[Certificate, RSAPrivateKeyWithSerialization]:
+    try:
+        from cryptography import x509
+        from cryptography.hazmat.backends import default_backend
+        from cryptography.hazmat.primitives import hashes
+        from cryptography.hazmat.primitives.asymmetric import rsa
+        from cryptography.x509.oid import NameOID
+    except ImportError:
+        raise TypeError(
+            "Using ad-hoc certificates requires the cryptography library."
+        ) from None
+
+    backend = default_backend()
+    pkey = rsa.generate_private_key(
+        public_exponent=65537, key_size=2048, backend=backend
+    )
+
+    # pretty damn sure that this is not actually accepted by anyone
+    if cn is None:
+        cn = "*"
+
+    subject = x509.Name(
+        [
+            x509.NameAttribute(NameOID.ORGANIZATION_NAME, "Dummy Certificate"),
+            x509.NameAttribute(NameOID.COMMON_NAME, cn),
+        ]
+    )
+
+    backend = default_backend()
+    cert = (
+        x509.CertificateBuilder()
+        .subject_name(subject)
+        .issuer_name(subject)
+        .public_key(pkey.public_key())
+        .serial_number(x509.random_serial_number())
+        .not_valid_before(dt.now(timezone.utc))
+        .not_valid_after(dt.now(timezone.utc) + timedelta(days=365))
+        .add_extension(x509.ExtendedKeyUsage([x509.OID_SERVER_AUTH]), critical=False)
+        .add_extension(
+            x509.SubjectAlternativeName([x509.DNSName(cn), x509.DNSName(f"*.{cn}")]),
+            critical=False,
+        )
+        .sign(pkey, hashes.SHA256(), backend)
+    )
+    return cert, pkey
+
+
+def make_ssl_devcert(
+    base_path: str, host: str | None = None, cn: str | None = None
+) -> tuple[str, str]:
+    """Creates an SSL key for development.  This should be used instead of
+    the ``'adhoc'`` key which generates a new cert on each server start.
+    It accepts a path for where it should store the key and cert and
+    either a host or CN.  If a host is given it will use the CN
+    ``*.host/CN=host``.
+
+    For more information see :func:`run_simple`.
+
+    .. versionadded:: 0.9
+
+    :param base_path: the path to the certificate and key.  The extension
+                      ``.crt`` is added for the certificate, ``.key`` is
+                      added for the key.
+    :param host: the name of the host.  This can be used as an alternative
+                 for the `cn`.
+    :param cn: the `CN` to use.
+    """
+
+    if host is not None:
+        cn = host
+    cert, pkey = generate_adhoc_ssl_pair(cn=cn)
+
+    from cryptography.hazmat.primitives import serialization
+
+    cert_file = f"{base_path}.crt"
+    pkey_file = f"{base_path}.key"
+
+    with open(cert_file, "wb") as f:
+        f.write(cert.public_bytes(serialization.Encoding.PEM))
+    with open(pkey_file, "wb") as f:
+        f.write(
+            pkey.private_bytes(
+                encoding=serialization.Encoding.PEM,
+                format=serialization.PrivateFormat.TraditionalOpenSSL,
+                encryption_algorithm=serialization.NoEncryption(),
+            )
+        )
+
+    return cert_file, pkey_file
+
+
+def generate_adhoc_ssl_context() -> ssl.SSLContext:
+    """Generates an adhoc SSL context for the development server."""
+    import atexit
+    import tempfile
+
+    cert, pkey = generate_adhoc_ssl_pair()
+
+    from cryptography.hazmat.primitives import serialization
+
+    cert_handle, cert_file = tempfile.mkstemp()
+    pkey_handle, pkey_file = tempfile.mkstemp()
+    atexit.register(os.remove, pkey_file)
+    atexit.register(os.remove, cert_file)
+
+    os.write(cert_handle, cert.public_bytes(serialization.Encoding.PEM))
+    os.write(
+        pkey_handle,
+        pkey.private_bytes(
+            encoding=serialization.Encoding.PEM,
+            format=serialization.PrivateFormat.TraditionalOpenSSL,
+            encryption_algorithm=serialization.NoEncryption(),
+        ),
+    )
+
+    os.close(cert_handle)
+    os.close(pkey_handle)
+    ctx = load_ssl_context(cert_file, pkey_file)
+    return ctx
+
+
+def load_ssl_context(
+    cert_file: str, pkey_file: str | None = None, protocol: int | None = None
+) -> ssl.SSLContext:
+    """Loads SSL context from cert/private key files and optional protocol.
+    Many parameters are directly taken from the API of
+    :py:class:`ssl.SSLContext`.
+
+    :param cert_file: Path of the certificate to use.
+    :param pkey_file: Path of the private key to use. If not given, the key
+                      will be obtained from the certificate file.
+    :param protocol: A ``PROTOCOL`` constant from the :mod:`ssl` module.
+        Defaults to :data:`ssl.PROTOCOL_TLS_SERVER`.
+    """
+    if protocol is None:
+        protocol = ssl.PROTOCOL_TLS_SERVER
+
+    ctx = ssl.SSLContext(protocol)
+    ctx.load_cert_chain(cert_file, pkey_file)
+    return ctx
+
+
+def is_ssl_error(error: Exception | None = None) -> bool:
+    """Checks if the given error (or the current one) is an SSL error."""
+    if error is None:
+        error = t.cast(Exception, sys.exc_info()[1])
+    return isinstance(error, ssl.SSLError)
+
+
+def select_address_family(host: str, port: int) -> socket.AddressFamily:
+    """Return ``AF_INET4``, ``AF_INET6``, or ``AF_UNIX`` depending on
+    the host and port."""
+    if host.startswith("unix://"):
+        return socket.AF_UNIX
+    elif ":" in host and hasattr(socket, "AF_INET6"):
+        return socket.AF_INET6
+    return socket.AF_INET
+
+
+def get_sockaddr(
+    host: str, port: int, family: socket.AddressFamily
+) -> tuple[str, int] | str:
+    """Return a fully qualified socket address that can be passed to
+    :func:`socket.bind`."""
+    if family == af_unix:
+        # Absolute path avoids IDNA encoding error when path starts with dot.
+        return os.path.abspath(host.partition("://")[2])
+    try:
+        res = socket.getaddrinfo(
+            host, port, family, socket.SOCK_STREAM, socket.IPPROTO_TCP
+        )
+    except socket.gaierror:
+        return host, port
+    return res[0][4]  # type: ignore
+
+
+def get_interface_ip(family: socket.AddressFamily) -> str:
+    """Get the IP address of an external interface. Used when binding to
+    0.0.0.0 or ::1 to show a more useful URL.
+
+    :meta private:
+    """
+    # arbitrary private address
+    host = "fd31:f903:5ab5:1::1" if family == socket.AF_INET6 else "10.253.155.219"
+
+    with socket.socket(family, socket.SOCK_DGRAM) as s:
+        try:
+            s.connect((host, 58162))
+        except OSError:
+            return "::1" if family == socket.AF_INET6 else "127.0.0.1"
+
+        return s.getsockname()[0]  # type: ignore
+
+
+class BaseWSGIServer(HTTPServer):
+    """A WSGI server that that handles one request at a time.
+
+    Use :func:`make_server` to create a server instance.
+    """
+
+    multithread = False
+    multiprocess = False
+    request_queue_size = LISTEN_QUEUE
+    allow_reuse_address = True
+
+    def __init__(
+        self,
+        host: str,
+        port: int,
+        app: WSGIApplication,
+        handler: type[WSGIRequestHandler] | None = None,
+        passthrough_errors: bool = False,
+        ssl_context: _TSSLContextArg | None = None,
+        fd: int | None = None,
+    ) -> None:
+        if handler is None:
+            handler = WSGIRequestHandler
+
+        # If the handler doesn't directly set a protocol version and
+        # thread or process workers are used, then allow chunked
+        # responses and keep-alive connections by enabling HTTP/1.1.
+        if "protocol_version" not in vars(handler) and (
+            self.multithread or self.multiprocess
+        ):
+            handler.protocol_version = "HTTP/1.1"
+
+        self.host = host
+        self.port = port
+        self.app = app
+        self.passthrough_errors = passthrough_errors
+
+        self.address_family = address_family = select_address_family(host, port)
+        server_address = get_sockaddr(host, int(port), address_family)
+
+        # Remove a leftover Unix socket file from a previous run. Don't
+        # remove a file that was set up by run_simple.
+        if address_family == af_unix and fd is None:
+            server_address = t.cast(str, server_address)
+
+            if os.path.exists(server_address):
+                os.unlink(server_address)
+
+        # Bind and activate will be handled manually, it should only
+        # happen if we're not using a socket that was already set up.
+        super().__init__(
+            server_address,  # type: ignore[arg-type]
+            handler,
+            bind_and_activate=False,
+        )
+
+        if fd is None:
+            # No existing socket descriptor, do bind_and_activate=True.
+            try:
+                self.server_bind()
+                self.server_activate()
+            except OSError as e:
+                # Catch connection issues and show them without the traceback. Show
+                # extra instructions for address not found, and for macOS.
+                self.server_close()
+                print(e.strerror, file=sys.stderr)
+
+                if e.errno == errno.EADDRINUSE:
+                    print(
+                        f"Port {port} is in use by another program. Either identify and"
+                        " stop that program, or start the server with a different"
+                        " port.",
+                        file=sys.stderr,
+                    )
+
+                    if sys.platform == "darwin" and port == 5000:
+                        print(
+                            "On macOS, try disabling the 'AirPlay Receiver' service"
+                            " from System Preferences -> General -> AirDrop & Handoff.",
+                            file=sys.stderr,
+                        )
+
+                sys.exit(1)
+            except BaseException:
+                self.server_close()
+                raise
+        else:
+            # TCPServer automatically opens a socket even if bind_and_activate is False.
+            # Close it to silence a ResourceWarning.
+            self.server_close()
+
+            # Use the passed in socket directly.
+            self.socket = socket.fromfd(fd, address_family, socket.SOCK_STREAM)
+            self.server_address = self.socket.getsockname()
+
+        if address_family != af_unix:
+            # If port was 0, this will record the bound port.
+            self.port = self.server_address[1]
+
+        if ssl_context is not None:
+            if isinstance(ssl_context, tuple):
+                ssl_context = load_ssl_context(*ssl_context)
+            elif ssl_context == "adhoc":
+                ssl_context = generate_adhoc_ssl_context()
+
+            self.socket = ssl_context.wrap_socket(self.socket, server_side=True)
+            self.ssl_context: ssl.SSLContext | None = ssl_context
+        else:
+            self.ssl_context = None
+
+        import importlib.metadata
+
+        self._server_version = f"Werkzeug/{importlib.metadata.version('werkzeug')}"
+
+    def log(self, type: str, message: str, *args: t.Any) -> None:
+        _log(type, message, *args)
+
+    def serve_forever(self, poll_interval: float = 0.5) -> None:
+        try:
+            super().serve_forever(poll_interval=poll_interval)
+        except KeyboardInterrupt:
+            pass
+        finally:
+            self.server_close()
+
+    def handle_error(
+        self, request: t.Any, client_address: tuple[str, int] | str
+    ) -> None:
+        if self.passthrough_errors:
+            raise
+
+        return super().handle_error(request, client_address)
+
+    def log_startup(self) -> None:
+        """Show information about the address when starting the server."""
+        dev_warning = (
+            "WARNING: This is a development server. Do not use it in a production"
+            " deployment. Use a production WSGI server instead."
+        )
+        dev_warning = _ansi_style(dev_warning, "bold", "red")
+        messages = [dev_warning]
+
+        if self.address_family == af_unix:
+            messages.append(f" * Running on {self.host}")
+        else:
+            scheme = "http" if self.ssl_context is None else "https"
+            display_hostname = self.host
+
+            if self.host in {"0.0.0.0", "::"}:
+                messages.append(f" * Running on all addresses ({self.host})")
+
+                if self.host == "0.0.0.0":
+                    localhost = "127.0.0.1"
+                    display_hostname = get_interface_ip(socket.AF_INET)
+                else:
+                    localhost = "[::1]"
+                    display_hostname = get_interface_ip(socket.AF_INET6)
+
+                messages.append(f" * Running on {scheme}://{localhost}:{self.port}")
+
+            if ":" in display_hostname:
+                display_hostname = f"[{display_hostname}]"
+
+            messages.append(f" * Running on {scheme}://{display_hostname}:{self.port}")
+
+        _log("info", "\n".join(messages))
+
+
+class ThreadedWSGIServer(socketserver.ThreadingMixIn, BaseWSGIServer):
+    """A WSGI server that handles concurrent requests in separate
+    threads.
+
+    Use :func:`make_server` to create a server instance.
+    """
+
+    multithread = True
+    daemon_threads = True
+
+
+class ForkingWSGIServer(ForkingMixIn, BaseWSGIServer):
+    """A WSGI server that handles concurrent requests in separate forked
+    processes.
+
+    Use :func:`make_server` to create a server instance.
+    """
+
+    multiprocess = True
+
+    def __init__(
+        self,
+        host: str,
+        port: int,
+        app: WSGIApplication,
+        processes: int = 40,
+        handler: type[WSGIRequestHandler] | None = None,
+        passthrough_errors: bool = False,
+        ssl_context: _TSSLContextArg | None = None,
+        fd: int | None = None,
+    ) -> None:
+        if not can_fork:
+            raise ValueError("Your platform does not support forking.")
+
+        super().__init__(host, port, app, handler, passthrough_errors, ssl_context, fd)
+        self.max_children = processes
+
+
+def make_server(
+    host: str,
+    port: int,
+    app: WSGIApplication,
+    threaded: bool = False,
+    processes: int = 1,
+    request_handler: type[WSGIRequestHandler] | None = None,
+    passthrough_errors: bool = False,
+    ssl_context: _TSSLContextArg | None = None,
+    fd: int | None = None,
+) -> BaseWSGIServer:
+    """Create an appropriate WSGI server instance based on the value of
+    ``threaded`` and ``processes``.
+
+    This is called from :func:`run_simple`, but can be used separately
+    to have access to the server object, such as to run it in a separate
+    thread.
+
+    See :func:`run_simple` for parameter docs.
+    """
+    if threaded and processes > 1:
+        raise ValueError("Cannot have a multi-thread and multi-process server.")
+
+    if threaded:
+        return ThreadedWSGIServer(
+            host, port, app, request_handler, passthrough_errors, ssl_context, fd=fd
+        )
+
+    if processes > 1:
+        return ForkingWSGIServer(
+            host,
+            port,
+            app,
+            processes,
+            request_handler,
+            passthrough_errors,
+            ssl_context,
+            fd=fd,
+        )
+
+    return BaseWSGIServer(
+        host, port, app, request_handler, passthrough_errors, ssl_context, fd=fd
+    )
+
+
+def is_running_from_reloader() -> bool:
+    """Check if the server is running as a subprocess within the
+    Werkzeug reloader.
+
+    .. versionadded:: 0.10
+    """
+    return os.environ.get("WERKZEUG_RUN_MAIN") == "true"
+
+
+def run_simple(
+    hostname: str,
+    port: int,
+    application: WSGIApplication,
+    use_reloader: bool = False,
+    use_debugger: bool = False,
+    use_evalex: bool = True,
+    extra_files: t.Iterable[str] | None = None,
+    exclude_patterns: t.Iterable[str] | None = None,
+    reloader_interval: int = 1,
+    reloader_type: str = "auto",
+    threaded: bool = False,
+    processes: int = 1,
+    request_handler: type[WSGIRequestHandler] | None = None,
+    static_files: dict[str, str | tuple[str, str]] | None = None,
+    passthrough_errors: bool = False,
+    ssl_context: _TSSLContextArg | None = None,
+) -> None:
+    """Start a development server for a WSGI application. Various
+    optional features can be enabled.
+
+    .. warning::
+
+        Do not use the development server when deploying to production.
+        It is intended for use only during local development. It is not
+        designed to be particularly efficient, stable, or secure.
+
+    :param hostname: The host to bind to, for example ``'localhost'``.
+        Can be a domain, IPv4 or IPv6 address, or file path starting
+        with ``unix://`` for a Unix socket.
+    :param port: The port to bind to, for example ``8080``. Using ``0``
+        tells the OS to pick a random free port.
+    :param application: The WSGI application to run.
+    :param use_reloader: Use a reloader process to restart the server
+        process when files are changed.
+    :param use_debugger: Use Werkzeug's debugger, which will show
+        formatted tracebacks on unhandled exceptions.
+    :param use_evalex: Make the debugger interactive. A Python terminal
+        can be opened for any frame in the traceback. Some protection is
+        provided by requiring a PIN, but this should never be enabled
+        on a publicly visible server.
+    :param extra_files: The reloader will watch these files for changes
+        in addition to Python modules. For example, watch a
+        configuration file.
+    :param exclude_patterns: The reloader will ignore changes to any
+        files matching these :mod:`fnmatch` patterns. For example,
+        ignore cache files.
+    :param reloader_interval: How often the reloader tries to check for
+        changes.
+    :param reloader_type: The reloader to use. The ``'stat'`` reloader
+        is built in, but may require significant CPU to watch files. The
+        ``'watchdog'`` reloader is much more efficient but requires
+        installing the ``watchdog`` package first.
+    :param threaded: Handle concurrent requests using threads. Cannot be
+        used with ``processes``.
+    :param processes: Handle concurrent requests using up to this number
+        of processes. Cannot be used with ``threaded``.
+    :param request_handler: Use a different
+        :class:`~BaseHTTPServer.BaseHTTPRequestHandler` subclass to
+        handle requests.
+    :param static_files: A dict mapping URL prefixes to directories to
+        serve static files from using
+        :class:`~werkzeug.middleware.SharedDataMiddleware`.
+    :param passthrough_errors: Don't catch unhandled exceptions at the
+        server level, let the server crash instead. If ``use_debugger``
+        is enabled, the debugger will still catch such errors.
+    :param ssl_context: Configure TLS to serve over HTTPS. Can be an
+        :class:`ssl.SSLContext` object, a ``(cert_file, key_file)``
+        tuple to create a typical context, or the string ``'adhoc'`` to
+        generate a temporary self-signed certificate.
+
+    .. versionchanged:: 2.1
+        Instructions are shown for dealing with an "address already in
+        use" error.
+
+    .. versionchanged:: 2.1
+        Running on ``0.0.0.0`` or ``::`` shows the loopback IP in
+        addition to a real IP.
+
+    .. versionchanged:: 2.1
+        The command-line interface was removed.
+
+    .. versionchanged:: 2.0
+        Running on ``0.0.0.0`` or ``::`` shows a real IP address that
+        was bound as well as a warning not to run the development server
+        in production.
+
+    .. versionchanged:: 2.0
+        The ``exclude_patterns`` parameter was added.
+
+    .. versionchanged:: 0.15
+        Bind to a Unix socket by passing a ``hostname`` that starts with
+        ``unix://``.
+
+    .. versionchanged:: 0.10
+        Improved the reloader and added support for changing the backend
+        through the ``reloader_type`` parameter.
+
+    .. versionchanged:: 0.9
+        A command-line interface was added.
+
+    .. versionchanged:: 0.8
+        ``ssl_context`` can be a tuple of paths to the certificate and
+        private key files.
+
+    .. versionchanged:: 0.6
+        The ``ssl_context`` parameter was added.
+
+    .. versionchanged:: 0.5
+       The ``static_files`` and ``passthrough_errors`` parameters were
+       added.
+    """
+    if not isinstance(port, int):
+        raise TypeError("port must be an integer")
+
+    if static_files:
+        from .middleware.shared_data import SharedDataMiddleware
+
+        application = SharedDataMiddleware(application, static_files)
+
+    if use_debugger:
+        from .debug import DebuggedApplication
+
+        application = DebuggedApplication(application, evalex=use_evalex)
+        # Allow the specified hostname to use the debugger, in addition to
+        # localhost domains.
+        application.trusted_hosts.append(hostname)
+
+    if not is_running_from_reloader():
+        fd = None
+    else:
+        fd = int(os.environ["WERKZEUG_SERVER_FD"])
+
+    srv = make_server(
+        hostname,
+        port,
+        application,
+        threaded,
+        processes,
+        request_handler,
+        passthrough_errors,
+        ssl_context,
+        fd=fd,
+    )
+    srv.socket.set_inheritable(True)
+    os.environ["WERKZEUG_SERVER_FD"] = str(srv.fileno())
+
+    if not is_running_from_reloader():
+        srv.log_startup()
+        _log("info", _ansi_style("Press CTRL+C to quit", "yellow"))
+
+    if use_reloader:
+        from ._reloader import run_with_reloader
+
+        try:
+            run_with_reloader(
+                srv.serve_forever,
+                extra_files=extra_files,
+                exclude_patterns=exclude_patterns,
+                interval=reloader_interval,
+                reloader_type=reloader_type,
+            )
+        finally:
+            srv.server_close()
+    else:
+        srv.serve_forever()
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/test.py b/backend/venv/lib/python3.12/site-packages/werkzeug/test.py
new file mode 100644
index 00000000..5c3c6088
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/test.py
@@ -0,0 +1,1464 @@
+from __future__ import annotations
+
+import dataclasses
+import mimetypes
+import sys
+import typing as t
+from collections import defaultdict
+from datetime import datetime
+from io import BytesIO
+from itertools import chain
+from random import random
+from tempfile import TemporaryFile
+from time import time
+from urllib.parse import unquote
+from urllib.parse import urlsplit
+from urllib.parse import urlunsplit
+
+from ._internal import _get_environ
+from ._internal import _wsgi_decoding_dance
+from ._internal import _wsgi_encoding_dance
+from .datastructures import Authorization
+from .datastructures import CallbackDict
+from .datastructures import CombinedMultiDict
+from .datastructures import EnvironHeaders
+from .datastructures import FileMultiDict
+from .datastructures import Headers
+from .datastructures import MultiDict
+from .http import dump_cookie
+from .http import dump_options_header
+from .http import parse_cookie
+from .http import parse_date
+from .http import parse_options_header
+from .sansio.multipart import Data
+from .sansio.multipart import Epilogue
+from .sansio.multipart import Field
+from .sansio.multipart import File
+from .sansio.multipart import MultipartEncoder
+from .sansio.multipart import Preamble
+from .urls import _urlencode
+from .urls import iri_to_uri
+from .utils import cached_property
+from .utils import get_content_type
+from .wrappers.request import Request
+from .wrappers.response import Response
+from .wsgi import ClosingIterator
+from .wsgi import get_current_url
+
+if t.TYPE_CHECKING:
+    import typing_extensions as te
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+def stream_encode_multipart(
+    data: t.Mapping[str, t.Any],
+    use_tempfile: bool = True,
+    threshold: int = 1024 * 500,
+    boundary: str | None = None,
+) -> tuple[t.IO[bytes], int, str]:
+    """Encode a dict of values (either strings or file descriptors or
+    :class:`FileStorage` objects.) into a multipart encoded string stored
+    in a file descriptor.
+
+    .. versionchanged:: 3.0
+        The ``charset`` parameter was removed.
+    """
+    if boundary is None:
+        boundary = f"---------------WerkzeugFormPart_{time()}{random()}"
+
+    stream: t.IO[bytes] = BytesIO()
+    total_length = 0
+    on_disk = False
+    write_binary: t.Callable[[bytes], int]
+
+    if use_tempfile:
+
+        def write_binary(s: bytes) -> int:
+            nonlocal stream, total_length, on_disk
+
+            if on_disk:
+                return stream.write(s)
+            else:
+                length = len(s)
+
+                if length + total_length <= threshold:
+                    stream.write(s)
+                else:
+                    new_stream = t.cast(t.IO[bytes], TemporaryFile("wb+"))
+                    new_stream.write(stream.getvalue())  # type: ignore
+                    new_stream.write(s)
+                    stream = new_stream
+                    on_disk = True
+
+                total_length += length
+                return length
+
+    else:
+        write_binary = stream.write
+
+    encoder = MultipartEncoder(boundary.encode())
+    write_binary(encoder.send_event(Preamble(data=b"")))
+    for key, value in _iter_data(data):
+        reader = getattr(value, "read", None)
+        if reader is not None:
+            filename = getattr(value, "filename", getattr(value, "name", None))
+            content_type = getattr(value, "content_type", None)
+            if content_type is None:
+                content_type = (
+                    filename
+                    and mimetypes.guess_type(filename)[0]
+                    or "application/octet-stream"
+                )
+            headers = value.headers
+            headers.update([("Content-Type", content_type)])
+            if filename is None:
+                write_binary(encoder.send_event(Field(name=key, headers=headers)))
+            else:
+                write_binary(
+                    encoder.send_event(
+                        File(name=key, filename=filename, headers=headers)
+                    )
+                )
+            while True:
+                chunk = reader(16384)
+
+                if not chunk:
+                    write_binary(encoder.send_event(Data(data=chunk, more_data=False)))
+                    break
+
+                write_binary(encoder.send_event(Data(data=chunk, more_data=True)))
+        else:
+            if not isinstance(value, str):
+                value = str(value)
+            write_binary(encoder.send_event(Field(name=key, headers=Headers())))
+            write_binary(encoder.send_event(Data(data=value.encode(), more_data=False)))
+
+    write_binary(encoder.send_event(Epilogue(data=b"")))
+
+    length = stream.tell()
+    stream.seek(0)
+    return stream, length, boundary
+
+
+def encode_multipart(
+    values: t.Mapping[str, t.Any], boundary: str | None = None
+) -> tuple[str, bytes]:
+    """Like `stream_encode_multipart` but returns a tuple in the form
+    (``boundary``, ``data``) where data is bytes.
+
+    .. versionchanged:: 3.0
+        The ``charset`` parameter was removed.
+    """
+    stream, length, boundary = stream_encode_multipart(
+        values, use_tempfile=False, boundary=boundary
+    )
+    return boundary, stream.read()
+
+
+def _iter_data(data: t.Mapping[str, t.Any]) -> t.Iterator[tuple[str, t.Any]]:
+    """Iterate over a mapping that might have a list of values, yielding
+    all key, value pairs. Almost like iter_multi_items but only allows
+    lists, not tuples, of values so tuples can be used for files.
+    """
+    if isinstance(data, MultiDict):
+        yield from data.items(multi=True)
+    else:
+        for key, value in data.items():
+            if isinstance(value, list):
+                for v in value:
+                    yield key, v
+            else:
+                yield key, value
+
+
+_TAnyMultiDict = t.TypeVar("_TAnyMultiDict", bound="MultiDict[t.Any, t.Any]")
+
+
+class EnvironBuilder:
+    """This class can be used to conveniently create a WSGI environment
+    for testing purposes.  It can be used to quickly create WSGI environments
+    or request objects from arbitrary data.
+
+    The signature of this class is also used in some other places as of
+    Werkzeug 0.5 (:func:`create_environ`, :meth:`Response.from_values`,
+    :meth:`Client.open`).  Because of this most of the functionality is
+    available through the constructor alone.
+
+    Files and regular form data can be manipulated independently of each
+    other with the :attr:`form` and :attr:`files` attributes, but are
+    passed with the same argument to the constructor: `data`.
+
+    `data` can be any of these values:
+
+    -   a `str` or `bytes` object: The object is converted into an
+        :attr:`input_stream`, the :attr:`content_length` is set and you have to
+        provide a :attr:`content_type`.
+    -   a `dict` or :class:`MultiDict`: The keys have to be strings. The values
+        have to be either any of the following objects, or a list of any of the
+        following objects:
+
+        -   a :class:`file`-like object:  These are converted into
+            :class:`FileStorage` objects automatically.
+        -   a `tuple`:  The :meth:`~FileMultiDict.add_file` method is called
+            with the key and the unpacked `tuple` items as positional
+            arguments.
+        -   a `str`:  The string is set as form data for the associated key.
+    -   a file-like object: The object content is loaded in memory and then
+        handled like a regular `str` or a `bytes`.
+
+    :param path: the path of the request.  In the WSGI environment this will
+                 end up as `PATH_INFO`.  If the `query_string` is not defined
+                 and there is a question mark in the `path` everything after
+                 it is used as query string.
+    :param base_url: the base URL is a URL that is used to extract the WSGI
+                     URL scheme, host (server name + server port) and the
+                     script root (`SCRIPT_NAME`).
+    :param query_string: an optional string or dict with URL parameters.
+    :param method: the HTTP method to use, defaults to `GET`.
+    :param input_stream: an optional input stream.  Do not specify this and
+                         `data`.  As soon as an input stream is set you can't
+                         modify :attr:`args` and :attr:`files` unless you
+                         set the :attr:`input_stream` to `None` again.
+    :param content_type: The content type for the request.  As of 0.5 you
+                         don't have to provide this when specifying files
+                         and form data via `data`.
+    :param content_length: The content length for the request.  You don't
+                           have to specify this when providing data via
+                           `data`.
+    :param errors_stream: an optional error stream that is used for
+                          `wsgi.errors`.  Defaults to :data:`stderr`.
+    :param multithread: controls `wsgi.multithread`.  Defaults to `False`.
+    :param multiprocess: controls `wsgi.multiprocess`.  Defaults to `False`.
+    :param run_once: controls `wsgi.run_once`.  Defaults to `False`.
+    :param headers: an optional list or :class:`Headers` object of headers.
+    :param data: a string or dict of form data or a file-object.
+                 See explanation above.
+    :param json: An object to be serialized and assigned to ``data``.
+        Defaults the content type to ``"application/json"``.
+        Serialized with the function assigned to :attr:`json_dumps`.
+    :param environ_base: an optional dict of environment defaults.
+    :param environ_overrides: an optional dict of environment overrides.
+    :param auth: An authorization object to use for the
+        ``Authorization`` header value. A ``(username, password)`` tuple
+        is a shortcut for ``Basic`` authorization.
+
+    .. versionchanged:: 3.0
+        The ``charset`` parameter was removed.
+
+    .. versionchanged:: 2.1
+        ``CONTENT_TYPE`` and ``CONTENT_LENGTH`` are not duplicated as
+        header keys in the environ.
+
+    .. versionchanged:: 2.0
+        ``REQUEST_URI`` and ``RAW_URI`` is the full raw URI including
+        the query string, not only the path.
+
+    .. versionchanged:: 2.0
+        The default :attr:`request_class` is ``Request`` instead of
+        ``BaseRequest``.
+
+    .. versionadded:: 2.0
+       Added the ``auth`` parameter.
+
+    .. versionadded:: 0.15
+        The ``json`` param and :meth:`json_dumps` method.
+
+    .. versionadded:: 0.15
+        The environ has keys ``REQUEST_URI`` and ``RAW_URI`` containing
+        the path before percent-decoding. This is not part of the WSGI
+        PEP, but many WSGI servers include it.
+
+    .. versionchanged:: 0.6
+       ``path`` and ``base_url`` can now be unicode strings that are
+       encoded with :func:`iri_to_uri`.
+    """
+
+    #: the server protocol to use.  defaults to HTTP/1.1
+    server_protocol = "HTTP/1.1"
+
+    #: the wsgi version to use.  defaults to (1, 0)
+    wsgi_version = (1, 0)
+
+    #: The default request class used by :meth:`get_request`.
+    request_class = Request
+
+    import json
+
+    #: The serialization function used when ``json`` is passed.
+    json_dumps = staticmethod(json.dumps)
+    del json
+
+    _args: MultiDict[str, str] | None
+    _query_string: str | None
+    _input_stream: t.IO[bytes] | None
+    _form: MultiDict[str, str] | None
+    _files: FileMultiDict | None
+
+    def __init__(
+        self,
+        path: str = "/",
+        base_url: str | None = None,
+        query_string: t.Mapping[str, str] | str | None = None,
+        method: str = "GET",
+        input_stream: t.IO[bytes] | None = None,
+        content_type: str | None = None,
+        content_length: int | None = None,
+        errors_stream: t.IO[str] | None = None,
+        multithread: bool = False,
+        multiprocess: bool = False,
+        run_once: bool = False,
+        headers: Headers | t.Iterable[tuple[str, str]] | None = None,
+        data: None | (t.IO[bytes] | str | bytes | t.Mapping[str, t.Any]) = None,
+        environ_base: t.Mapping[str, t.Any] | None = None,
+        environ_overrides: t.Mapping[str, t.Any] | None = None,
+        mimetype: str | None = None,
+        json: t.Mapping[str, t.Any] | None = None,
+        auth: Authorization | tuple[str, str] | None = None,
+    ) -> None:
+        if query_string is not None and "?" in path:
+            raise ValueError("Query string is defined in the path and as an argument")
+        request_uri = urlsplit(path)
+        if query_string is None and "?" in path:
+            query_string = request_uri.query
+
+        self.path = iri_to_uri(request_uri.path)
+        self.request_uri = path
+        if base_url is not None:
+            base_url = iri_to_uri(base_url)
+        self.base_url = base_url  # type: ignore
+        if isinstance(query_string, str):
+            self.query_string = query_string
+        else:
+            if query_string is None:
+                query_string = MultiDict()
+            elif not isinstance(query_string, MultiDict):
+                query_string = MultiDict(query_string)
+            self.args = query_string
+        self.method = method
+        if headers is None:
+            headers = Headers()
+        elif not isinstance(headers, Headers):
+            headers = Headers(headers)
+        self.headers = headers
+        if content_type is not None:
+            self.content_type = content_type
+        if errors_stream is None:
+            errors_stream = sys.stderr
+        self.errors_stream = errors_stream
+        self.multithread = multithread
+        self.multiprocess = multiprocess
+        self.run_once = run_once
+        self.environ_base = environ_base
+        self.environ_overrides = environ_overrides
+        self.input_stream = input_stream
+        self.content_length = content_length
+        self.closed = False
+
+        if auth is not None:
+            if isinstance(auth, tuple):
+                auth = Authorization(
+                    "basic", {"username": auth[0], "password": auth[1]}
+                )
+
+            self.headers.set("Authorization", auth.to_header())
+
+        if json is not None:
+            if data is not None:
+                raise TypeError("can't provide both json and data")
+
+            data = self.json_dumps(json)
+
+            if self.content_type is None:
+                self.content_type = "application/json"
+
+        if data:
+            if input_stream is not None:
+                raise TypeError("can't provide input stream and data")
+            if hasattr(data, "read"):
+                data = data.read()
+            if isinstance(data, str):
+                data = data.encode()
+            if isinstance(data, bytes):
+                self.input_stream = BytesIO(data)
+                if self.content_length is None:
+                    self.content_length = len(data)
+            else:
+                for key, value in _iter_data(data):
+                    if isinstance(value, (tuple, dict)) or hasattr(value, "read"):
+                        self._add_file_from_data(key, value)
+                    else:
+                        self.form.setlistdefault(key).append(value)
+
+        if mimetype is not None:
+            self.mimetype = mimetype
+
+    @classmethod
+    def from_environ(cls, environ: WSGIEnvironment, **kwargs: t.Any) -> EnvironBuilder:
+        """Turn an environ dict back into a builder. Any extra kwargs
+        override the args extracted from the environ.
+
+        .. versionchanged:: 2.0
+            Path and query values are passed through the WSGI decoding
+            dance to avoid double encoding.
+
+        .. versionadded:: 0.15
+        """
+        headers = Headers(EnvironHeaders(environ))
+        out = {
+            "path": _wsgi_decoding_dance(environ["PATH_INFO"]),
+            "base_url": cls._make_base_url(
+                environ["wsgi.url_scheme"],
+                headers.pop("Host"),
+                _wsgi_decoding_dance(environ["SCRIPT_NAME"]),
+            ),
+            "query_string": _wsgi_decoding_dance(environ["QUERY_STRING"]),
+            "method": environ["REQUEST_METHOD"],
+            "input_stream": environ["wsgi.input"],
+            "content_type": headers.pop("Content-Type", None),
+            "content_length": headers.pop("Content-Length", None),
+            "errors_stream": environ["wsgi.errors"],
+            "multithread": environ["wsgi.multithread"],
+            "multiprocess": environ["wsgi.multiprocess"],
+            "run_once": environ["wsgi.run_once"],
+            "headers": headers,
+        }
+        out.update(kwargs)
+        return cls(**out)
+
+    def _add_file_from_data(
+        self,
+        key: str,
+        value: (t.IO[bytes] | tuple[t.IO[bytes], str] | tuple[t.IO[bytes], str, str]),
+    ) -> None:
+        """Called in the EnvironBuilder to add files from the data dict."""
+        if isinstance(value, tuple):
+            self.files.add_file(key, *value)
+        else:
+            self.files.add_file(key, value)
+
+    @staticmethod
+    def _make_base_url(scheme: str, host: str, script_root: str) -> str:
+        return urlunsplit((scheme, host, script_root, "", "")).rstrip("/") + "/"
+
+    @property
+    def base_url(self) -> str:
+        """The base URL is used to extract the URL scheme, host name,
+        port, and root path.
+        """
+        return self._make_base_url(self.url_scheme, self.host, self.script_root)
+
+    @base_url.setter
+    def base_url(self, value: str | None) -> None:
+        if value is None:
+            scheme = "http"
+            netloc = "localhost"
+            script_root = ""
+        else:
+            scheme, netloc, script_root, qs, anchor = urlsplit(value)
+            if qs or anchor:
+                raise ValueError("base url must not contain a query string or fragment")
+        self.script_root = script_root.rstrip("/")
+        self.host = netloc
+        self.url_scheme = scheme
+
+    @property
+    def content_type(self) -> str | None:
+        """The content type for the request.  Reflected from and to
+        the :attr:`headers`.  Do not set if you set :attr:`files` or
+        :attr:`form` for auto detection.
+        """
+        ct = self.headers.get("Content-Type")
+        if ct is None and not self._input_stream:
+            if self._files:
+                return "multipart/form-data"
+            if self._form:
+                return "application/x-www-form-urlencoded"
+            return None
+        return ct
+
+    @content_type.setter
+    def content_type(self, value: str | None) -> None:
+        if value is None:
+            self.headers.pop("Content-Type", None)
+        else:
+            self.headers["Content-Type"] = value
+
+    @property
+    def mimetype(self) -> str | None:
+        """The mimetype (content type without charset etc.)
+
+        .. versionadded:: 0.14
+        """
+        ct = self.content_type
+        return ct.split(";")[0].strip() if ct else None
+
+    @mimetype.setter
+    def mimetype(self, value: str) -> None:
+        self.content_type = get_content_type(value, "utf-8")
+
+    @property
+    def mimetype_params(self) -> t.Mapping[str, str]:
+        """The mimetype parameters as dict.  For example if the
+        content type is ``text/html; charset=utf-8`` the params would be
+        ``{'charset': 'utf-8'}``.
+
+        .. versionadded:: 0.14
+        """
+
+        def on_update(d: CallbackDict[str, str]) -> None:
+            self.headers["Content-Type"] = dump_options_header(self.mimetype, d)
+
+        d = parse_options_header(self.headers.get("content-type", ""))[1]
+        return CallbackDict(d, on_update)
+
+    @property
+    def content_length(self) -> int | None:
+        """The content length as integer.  Reflected from and to the
+        :attr:`headers`.  Do not set if you set :attr:`files` or
+        :attr:`form` for auto detection.
+        """
+        return self.headers.get("Content-Length", type=int)
+
+    @content_length.setter
+    def content_length(self, value: int | None) -> None:
+        if value is None:
+            self.headers.pop("Content-Length", None)
+        else:
+            self.headers["Content-Length"] = str(value)
+
+    def _get_form(self, name: str, storage: type[_TAnyMultiDict]) -> _TAnyMultiDict:
+        """Common behavior for getting the :attr:`form` and
+        :attr:`files` properties.
+
+        :param name: Name of the internal cached attribute.
+        :param storage: Storage class used for the data.
+        """
+        if self.input_stream is not None:
+            raise AttributeError("an input stream is defined")
+
+        rv = getattr(self, name)
+
+        if rv is None:
+            rv = storage()
+            setattr(self, name, rv)
+
+        return rv  # type: ignore
+
+    def _set_form(self, name: str, value: MultiDict[str, t.Any]) -> None:
+        """Common behavior for setting the :attr:`form` and
+        :attr:`files` properties.
+
+        :param name: Name of the internal cached attribute.
+        :param value: Value to assign to the attribute.
+        """
+        self._input_stream = None
+        setattr(self, name, value)
+
+    @property
+    def form(self) -> MultiDict[str, str]:
+        """A :class:`MultiDict` of form values."""
+        return self._get_form("_form", MultiDict)
+
+    @form.setter
+    def form(self, value: MultiDict[str, str]) -> None:
+        self._set_form("_form", value)
+
+    @property
+    def files(self) -> FileMultiDict:
+        """A :class:`FileMultiDict` of uploaded files. Use
+        :meth:`~FileMultiDict.add_file` to add new files.
+        """
+        return self._get_form("_files", FileMultiDict)
+
+    @files.setter
+    def files(self, value: FileMultiDict) -> None:
+        self._set_form("_files", value)
+
+    @property
+    def input_stream(self) -> t.IO[bytes] | None:
+        """An optional input stream. This is mutually exclusive with
+        setting :attr:`form` and :attr:`files`, setting it will clear
+        those. Do not provide this if the method is not ``POST`` or
+        another method that has a body.
+        """
+        return self._input_stream
+
+    @input_stream.setter
+    def input_stream(self, value: t.IO[bytes] | None) -> None:
+        self._input_stream = value
+        self._form = None
+        self._files = None
+
+    @property
+    def query_string(self) -> str:
+        """The query string.  If you set this to a string
+        :attr:`args` will no longer be available.
+        """
+        if self._query_string is None:
+            if self._args is not None:
+                return _urlencode(self._args)
+            return ""
+        return self._query_string
+
+    @query_string.setter
+    def query_string(self, value: str | None) -> None:
+        self._query_string = value
+        self._args = None
+
+    @property
+    def args(self) -> MultiDict[str, str]:
+        """The URL arguments as :class:`MultiDict`."""
+        if self._query_string is not None:
+            raise AttributeError("a query string is defined")
+        if self._args is None:
+            self._args = MultiDict()
+        return self._args
+
+    @args.setter
+    def args(self, value: MultiDict[str, str] | None) -> None:
+        self._query_string = None
+        self._args = value
+
+    @property
+    def server_name(self) -> str:
+        """The server name (read-only, use :attr:`host` to set)"""
+        return self.host.split(":", 1)[0]
+
+    @property
+    def server_port(self) -> int:
+        """The server port as integer (read-only, use :attr:`host` to set)"""
+        pieces = self.host.split(":", 1)
+
+        if len(pieces) == 2:
+            try:
+                return int(pieces[1])
+            except ValueError:
+                pass
+
+        if self.url_scheme == "https":
+            return 443
+        return 80
+
+    def __del__(self) -> None:
+        try:
+            self.close()
+        except Exception:
+            pass
+
+    def close(self) -> None:
+        """Closes all files.  If you put real :class:`file` objects into the
+        :attr:`files` dict you can call this method to automatically close
+        them all in one go.
+        """
+        if self.closed:
+            return
+        try:
+            files = self.files.values()
+        except AttributeError:
+            files = ()
+        for f in files:
+            try:
+                f.close()
+            except Exception:
+                pass
+        self.closed = True
+
+    def get_environ(self) -> WSGIEnvironment:
+        """Return the built environ.
+
+        .. versionchanged:: 0.15
+            The content type and length headers are set based on
+            input stream detection. Previously this only set the WSGI
+            keys.
+        """
+        input_stream = self.input_stream
+        content_length = self.content_length
+
+        mimetype = self.mimetype
+        content_type = self.content_type
+
+        if input_stream is not None:
+            start_pos = input_stream.tell()
+            input_stream.seek(0, 2)
+            end_pos = input_stream.tell()
+            input_stream.seek(start_pos)
+            content_length = end_pos - start_pos
+        elif mimetype == "multipart/form-data":
+            input_stream, content_length, boundary = stream_encode_multipart(
+                CombinedMultiDict([self.form, self.files])
+            )
+            content_type = f'{mimetype}; boundary="{boundary}"'
+        elif mimetype == "application/x-www-form-urlencoded":
+            form_encoded = _urlencode(self.form).encode("ascii")
+            content_length = len(form_encoded)
+            input_stream = BytesIO(form_encoded)
+        else:
+            input_stream = BytesIO()
+
+        result: WSGIEnvironment = {}
+        if self.environ_base:
+            result.update(self.environ_base)
+
+        def _path_encode(x: str) -> str:
+            return _wsgi_encoding_dance(unquote(x))
+
+        raw_uri = _wsgi_encoding_dance(self.request_uri)
+        result.update(
+            {
+                "REQUEST_METHOD": self.method,
+                "SCRIPT_NAME": _path_encode(self.script_root),
+                "PATH_INFO": _path_encode(self.path),
+                "QUERY_STRING": _wsgi_encoding_dance(self.query_string),
+                # Non-standard, added by mod_wsgi, uWSGI
+                "REQUEST_URI": raw_uri,
+                # Non-standard, added by gunicorn
+                "RAW_URI": raw_uri,
+                "SERVER_NAME": self.server_name,
+                "SERVER_PORT": str(self.server_port),
+                "HTTP_HOST": self.host,
+                "SERVER_PROTOCOL": self.server_protocol,
+                "wsgi.version": self.wsgi_version,
+                "wsgi.url_scheme": self.url_scheme,
+                "wsgi.input": input_stream,
+                "wsgi.errors": self.errors_stream,
+                "wsgi.multithread": self.multithread,
+                "wsgi.multiprocess": self.multiprocess,
+                "wsgi.run_once": self.run_once,
+            }
+        )
+
+        headers = self.headers.copy()
+        # Don't send these as headers, they're part of the environ.
+        headers.remove("Content-Type")
+        headers.remove("Content-Length")
+
+        if content_type is not None:
+            result["CONTENT_TYPE"] = content_type
+
+        if content_length is not None:
+            result["CONTENT_LENGTH"] = str(content_length)
+
+        combined_headers = defaultdict(list)
+
+        for key, value in headers.to_wsgi_list():
+            combined_headers[f"HTTP_{key.upper().replace('-', '_')}"].append(value)
+
+        for key, values in combined_headers.items():
+            result[key] = ", ".join(values)
+
+        if self.environ_overrides:
+            result.update(self.environ_overrides)
+
+        return result
+
+    def get_request(self, cls: type[Request] | None = None) -> Request:
+        """Returns a request with the data.  If the request class is not
+        specified :attr:`request_class` is used.
+
+        :param cls: The request wrapper to use.
+        """
+        if cls is None:
+            cls = self.request_class
+
+        return cls(self.get_environ())
+
+
+class ClientRedirectError(Exception):
+    """If a redirect loop is detected when using follow_redirects=True with
+    the :cls:`Client`, then this exception is raised.
+    """
+
+
+class Client:
+    """Simulate sending requests to a WSGI application without running a WSGI or HTTP
+    server.
+
+    :param application: The WSGI application to make requests to.
+    :param response_wrapper: A :class:`.Response` class to wrap response data with.
+        Defaults to :class:`.TestResponse`. If it's not a subclass of ``TestResponse``,
+        one will be created.
+    :param use_cookies: Persist cookies from ``Set-Cookie`` response headers to the
+        ``Cookie`` header in subsequent requests. Domain and path matching is supported,
+        but other cookie parameters are ignored.
+    :param allow_subdomain_redirects: Allow requests to follow redirects to subdomains.
+        Enable this if the application handles subdomains and redirects between them.
+
+    .. versionchanged:: 2.3
+        Simplify cookie implementation, support domain and path matching.
+
+    .. versionchanged:: 2.1
+        All data is available as properties on the returned response object. The
+        response cannot be returned as a tuple.
+
+    .. versionchanged:: 2.0
+        ``response_wrapper`` is always a subclass of :class:``TestResponse``.
+
+    .. versionchanged:: 0.5
+        Added the ``use_cookies`` parameter.
+    """
+
+    def __init__(
+        self,
+        application: WSGIApplication,
+        response_wrapper: type[Response] | None = None,
+        use_cookies: bool = True,
+        allow_subdomain_redirects: bool = False,
+    ) -> None:
+        self.application = application
+
+        if response_wrapper in {None, Response}:
+            response_wrapper = TestResponse
+        elif response_wrapper is not None and not issubclass(
+            response_wrapper, TestResponse
+        ):
+            response_wrapper = type(
+                "WrapperTestResponse",
+                (TestResponse, response_wrapper),
+                {},
+            )
+
+        self.response_wrapper = t.cast(type["TestResponse"], response_wrapper)
+
+        if use_cookies:
+            self._cookies: dict[tuple[str, str, str], Cookie] | None = {}
+        else:
+            self._cookies = None
+
+        self.allow_subdomain_redirects = allow_subdomain_redirects
+
+    def get_cookie(
+        self, key: str, domain: str = "localhost", path: str = "/"
+    ) -> Cookie | None:
+        """Return a :class:`.Cookie` if it exists. Cookies are uniquely identified by
+        ``(domain, path, key)``.
+
+        :param key: The decoded form of the key for the cookie.
+        :param domain: The domain the cookie was set for.
+        :param path: The path the cookie was set for.
+
+        .. versionadded:: 2.3
+        """
+        if self._cookies is None:
+            raise TypeError(
+                "Cookies are disabled. Create a client with 'use_cookies=True'."
+            )
+
+        return self._cookies.get((domain, path, key))
+
+    def set_cookie(
+        self,
+        key: str,
+        value: str = "",
+        *,
+        domain: str = "localhost",
+        origin_only: bool = True,
+        path: str = "/",
+        **kwargs: t.Any,
+    ) -> None:
+        """Set a cookie to be sent in subsequent requests.
+
+        This is a convenience to skip making a test request to a route that would set
+        the cookie. To test the cookie, make a test request to a route that uses the
+        cookie value.
+
+        The client uses ``domain``, ``origin_only``, and ``path`` to determine which
+        cookies to send with a request. It does not use other cookie parameters that
+        browsers use, since they're not applicable in tests.
+
+        :param key: The key part of the cookie.
+        :param value: The value part of the cookie.
+        :param domain: Send this cookie with requests that match this domain. If
+            ``origin_only`` is true, it must be an exact match, otherwise it may be a
+            suffix match.
+        :param origin_only: Whether the domain must be an exact match to the request.
+        :param path: Send this cookie with requests that match this path either exactly
+            or as a prefix.
+        :param kwargs: Passed to :func:`.dump_cookie`.
+
+        .. versionchanged:: 3.0
+            The parameter ``server_name`` is removed. The first parameter is
+            ``key``. Use the ``domain`` and ``origin_only`` parameters instead.
+
+        .. versionchanged:: 2.3
+            The ``origin_only`` parameter was added.
+
+        .. versionchanged:: 2.3
+            The ``domain`` parameter defaults to ``localhost``.
+        """
+        if self._cookies is None:
+            raise TypeError(
+                "Cookies are disabled. Create a client with 'use_cookies=True'."
+            )
+
+        cookie = Cookie._from_response_header(
+            domain, "/", dump_cookie(key, value, domain=domain, path=path, **kwargs)
+        )
+        cookie.origin_only = origin_only
+
+        if cookie._should_delete:
+            self._cookies.pop(cookie._storage_key, None)
+        else:
+            self._cookies[cookie._storage_key] = cookie
+
+    def delete_cookie(
+        self,
+        key: str,
+        *,
+        domain: str = "localhost",
+        path: str = "/",
+    ) -> None:
+        """Delete a cookie if it exists. Cookies are uniquely identified by
+        ``(domain, path, key)``.
+
+        :param key: The decoded form of the key for the cookie.
+        :param domain: The domain the cookie was set for.
+        :param path: The path the cookie was set for.
+
+        .. versionchanged:: 3.0
+            The ``server_name`` parameter is removed. The first parameter is
+            ``key``. Use the ``domain`` parameter instead.
+
+        .. versionchanged:: 3.0
+            The ``secure``, ``httponly`` and ``samesite`` parameters are removed.
+
+        .. versionchanged:: 2.3
+            The ``domain`` parameter defaults to ``localhost``.
+        """
+        if self._cookies is None:
+            raise TypeError(
+                "Cookies are disabled. Create a client with 'use_cookies=True'."
+            )
+
+        self._cookies.pop((domain, path, key), None)
+
+    def _add_cookies_to_wsgi(self, environ: WSGIEnvironment) -> None:
+        """If cookies are enabled, set the ``Cookie`` header in the environ to the
+        cookies that are applicable to the request host and path.
+
+        :meta private:
+
+        .. versionadded:: 2.3
+        """
+        if self._cookies is None:
+            return
+
+        url = urlsplit(get_current_url(environ))
+        server_name = url.hostname or "localhost"
+        value = "; ".join(
+            c._to_request_header()
+            for c in self._cookies.values()
+            if c._matches_request(server_name, url.path)
+        )
+
+        if value:
+            environ["HTTP_COOKIE"] = value
+        else:
+            environ.pop("HTTP_COOKIE", None)
+
+    def _update_cookies_from_response(
+        self, server_name: str, path: str, headers: list[str]
+    ) -> None:
+        """If cookies are enabled, update the stored cookies from any ``Set-Cookie``
+        headers in the response.
+
+        :meta private:
+
+        .. versionadded:: 2.3
+        """
+        if self._cookies is None:
+            return
+
+        for header in headers:
+            cookie = Cookie._from_response_header(server_name, path, header)
+
+            if cookie._should_delete:
+                self._cookies.pop(cookie._storage_key, None)
+            else:
+                self._cookies[cookie._storage_key] = cookie
+
+    def run_wsgi_app(
+        self, environ: WSGIEnvironment, buffered: bool = False
+    ) -> tuple[t.Iterable[bytes], str, Headers]:
+        """Runs the wrapped WSGI app with the given environment.
+
+        :meta private:
+        """
+        self._add_cookies_to_wsgi(environ)
+        rv = run_wsgi_app(self.application, environ, buffered=buffered)
+        url = urlsplit(get_current_url(environ))
+        self._update_cookies_from_response(
+            url.hostname or "localhost", url.path, rv[2].getlist("Set-Cookie")
+        )
+        return rv
+
+    def resolve_redirect(
+        self, response: TestResponse, buffered: bool = False
+    ) -> TestResponse:
+        """Perform a new request to the location given by the redirect
+        response to the previous request.
+
+        :meta private:
+        """
+        scheme, netloc, path, qs, anchor = urlsplit(response.location)
+        builder = EnvironBuilder.from_environ(
+            response.request.environ, path=path, query_string=qs
+        )
+
+        to_name_parts = netloc.split(":", 1)[0].split(".")
+        from_name_parts = builder.server_name.split(".")
+
+        if to_name_parts != [""]:
+            # The new location has a host, use it for the base URL.
+            builder.url_scheme = scheme
+            builder.host = netloc
+        else:
+            # A local redirect with autocorrect_location_header=False
+            # doesn't have a host, so use the request's host.
+            to_name_parts = from_name_parts
+
+        # Explain why a redirect to a different server name won't be followed.
+        if to_name_parts != from_name_parts:
+            if to_name_parts[-len(from_name_parts) :] == from_name_parts:
+                if not self.allow_subdomain_redirects:
+                    raise RuntimeError("Following subdomain redirects is not enabled.")
+            else:
+                raise RuntimeError("Following external redirects is not supported.")
+
+        path_parts = path.split("/")
+        root_parts = builder.script_root.split("/")
+
+        if path_parts[: len(root_parts)] == root_parts:
+            # Strip the script root from the path.
+            builder.path = path[len(builder.script_root) :]
+        else:
+            # The new location is not under the script root, so use the
+            # whole path and clear the previous root.
+            builder.path = path
+            builder.script_root = ""
+
+        # Only 307 and 308 preserve all of the original request.
+        if response.status_code not in {307, 308}:
+            # HEAD is preserved, everything else becomes GET.
+            if builder.method != "HEAD":
+                builder.method = "GET"
+
+            # Clear the body and the headers that describe it.
+
+            if builder.input_stream is not None:
+                builder.input_stream.close()
+                builder.input_stream = None
+
+            builder.content_type = None
+            builder.content_length = None
+            builder.headers.pop("Transfer-Encoding", None)
+
+        return self.open(builder, buffered=buffered)
+
+    def open(
+        self,
+        *args: t.Any,
+        buffered: bool = False,
+        follow_redirects: bool = False,
+        **kwargs: t.Any,
+    ) -> TestResponse:
+        """Generate an environ dict from the given arguments, make a
+        request to the application using it, and return the response.
+
+        :param args: Passed to :class:`EnvironBuilder` to create the
+            environ for the request. If a single arg is passed, it can
+            be an existing :class:`EnvironBuilder` or an environ dict.
+        :param buffered: Convert the iterator returned by the app into
+            a list. If the iterator has a ``close()`` method, it is
+            called automatically.
+        :param follow_redirects: Make additional requests to follow HTTP
+            redirects until a non-redirect status is returned.
+            :attr:`TestResponse.history` lists the intermediate
+            responses.
+
+        .. versionchanged:: 2.1
+            Removed the ``as_tuple`` parameter.
+
+        .. versionchanged:: 2.0
+            The request input stream is closed when calling
+            ``response.close()``. Input streams for redirects are
+            automatically closed.
+
+        .. versionchanged:: 0.5
+            If a dict is provided as file in the dict for the ``data``
+            parameter the content type has to be called ``content_type``
+            instead of ``mimetype``. This change was made for
+            consistency with :class:`werkzeug.FileWrapper`.
+
+        .. versionchanged:: 0.5
+            Added the ``follow_redirects`` parameter.
+        """
+        request: Request | None = None
+
+        if not kwargs and len(args) == 1:
+            arg = args[0]
+
+            if isinstance(arg, EnvironBuilder):
+                request = arg.get_request()
+            elif isinstance(arg, dict):
+                request = EnvironBuilder.from_environ(arg).get_request()
+            elif isinstance(arg, Request):
+                request = arg
+
+        if request is None:
+            builder = EnvironBuilder(*args, **kwargs)
+
+            try:
+                request = builder.get_request()
+            finally:
+                builder.close()
+
+        response_parts = self.run_wsgi_app(request.environ, buffered=buffered)
+        response = self.response_wrapper(*response_parts, request=request)
+
+        redirects = set()
+        history: list[TestResponse] = []
+
+        if not follow_redirects:
+            return response
+
+        while response.status_code in {
+            301,
+            302,
+            303,
+            305,
+            307,
+            308,
+        }:
+            # Exhaust intermediate response bodies to ensure middleware
+            # that returns an iterator runs any cleanup code.
+            if not buffered:
+                response.make_sequence()
+                response.close()
+
+            new_redirect_entry = (response.location, response.status_code)
+
+            if new_redirect_entry in redirects:
+                raise ClientRedirectError(
+                    f"Loop detected: A {response.status_code} redirect"
+                    f" to {response.location} was already made."
+                )
+
+            redirects.add(new_redirect_entry)
+            response.history = tuple(history)
+            history.append(response)
+            response = self.resolve_redirect(response, buffered=buffered)
+        else:
+            # This is the final request after redirects.
+            response.history = tuple(history)
+            # Close the input stream when closing the response, in case
+            # the input is an open temporary file.
+            response.call_on_close(request.input_stream.close)
+            return response
+
+    def get(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``GET``."""
+        kw["method"] = "GET"
+        return self.open(*args, **kw)
+
+    def post(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``POST``."""
+        kw["method"] = "POST"
+        return self.open(*args, **kw)
+
+    def put(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``PUT``."""
+        kw["method"] = "PUT"
+        return self.open(*args, **kw)
+
+    def delete(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``DELETE``."""
+        kw["method"] = "DELETE"
+        return self.open(*args, **kw)
+
+    def patch(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``PATCH``."""
+        kw["method"] = "PATCH"
+        return self.open(*args, **kw)
+
+    def options(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``OPTIONS``."""
+        kw["method"] = "OPTIONS"
+        return self.open(*args, **kw)
+
+    def head(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``HEAD``."""
+        kw["method"] = "HEAD"
+        return self.open(*args, **kw)
+
+    def trace(self, *args: t.Any, **kw: t.Any) -> TestResponse:
+        """Call :meth:`open` with ``method`` set to ``TRACE``."""
+        kw["method"] = "TRACE"
+        return self.open(*args, **kw)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.application!r}>"
+
+
+def create_environ(*args: t.Any, **kwargs: t.Any) -> WSGIEnvironment:
+    """Create a new WSGI environ dict based on the values passed.  The first
+    parameter should be the path of the request which defaults to '/'.  The
+    second one can either be an absolute path (in that case the host is
+    localhost:80) or a full path to the request with scheme, netloc port and
+    the path to the script.
+
+    This accepts the same arguments as the :class:`EnvironBuilder`
+    constructor.
+
+    .. versionchanged:: 0.5
+       This function is now a thin wrapper over :class:`EnvironBuilder` which
+       was added in 0.5.  The `headers`, `environ_base`, `environ_overrides`
+       and `charset` parameters were added.
+    """
+    builder = EnvironBuilder(*args, **kwargs)
+
+    try:
+        return builder.get_environ()
+    finally:
+        builder.close()
+
+
+def run_wsgi_app(
+    app: WSGIApplication, environ: WSGIEnvironment, buffered: bool = False
+) -> tuple[t.Iterable[bytes], str, Headers]:
+    """Return a tuple in the form (app_iter, status, headers) of the
+    application output.  This works best if you pass it an application that
+    returns an iterator all the time.
+
+    Sometimes applications may use the `write()` callable returned
+    by the `start_response` function.  This tries to resolve such edge
+    cases automatically.  But if you don't get the expected output you
+    should set `buffered` to `True` which enforces buffering.
+
+    If passed an invalid WSGI application the behavior of this function is
+    undefined.  Never pass non-conforming WSGI applications to this function.
+
+    :param app: the application to execute.
+    :param buffered: set to `True` to enforce buffering.
+    :return: tuple in the form ``(app_iter, status, headers)``
+    """
+    # Copy environ to ensure any mutations by the app (ProxyFix, for
+    # example) don't affect subsequent requests (such as redirects).
+    environ = _get_environ(environ).copy()
+    status: str
+    response: tuple[str, list[tuple[str, str]]] | None = None
+    buffer: list[bytes] = []
+
+    def start_response(status, headers, exc_info=None):  # type: ignore
+        nonlocal response
+
+        if exc_info:
+            try:
+                raise exc_info[1].with_traceback(exc_info[2])
+            finally:
+                exc_info = None
+
+        response = (status, headers)
+        return buffer.append
+
+    app_rv = app(environ, start_response)
+    close_func = getattr(app_rv, "close", None)
+    app_iter: t.Iterable[bytes] = iter(app_rv)
+
+    # when buffering we emit the close call early and convert the
+    # application iterator into a regular list
+    if buffered:
+        try:
+            app_iter = list(app_iter)
+        finally:
+            if close_func is not None:
+                close_func()
+
+    # otherwise we iterate the application iter until we have a response, chain
+    # the already received data with the already collected data and wrap it in
+    # a new `ClosingIterator` if we need to restore a `close` callable from the
+    # original return value.
+    else:
+        for item in app_iter:
+            buffer.append(item)
+
+            if response is not None:
+                break
+
+        if buffer:
+            app_iter = chain(buffer, app_iter)
+
+        if close_func is not None and app_iter is not app_rv:
+            app_iter = ClosingIterator(app_iter, close_func)
+
+    status, headers = response  # type: ignore
+    return app_iter, status, Headers(headers)
+
+
+class TestResponse(Response):
+    """:class:`~werkzeug.wrappers.Response` subclass that provides extra
+    information about requests made with the test :class:`Client`.
+
+    Test client requests will always return an instance of this class.
+    If a custom response class is passed to the client, it is
+    subclassed along with this to support test information.
+
+    If the test request included large files, or if the application is
+    serving a file, call :meth:`close` to close any open files and
+    prevent Python showing a ``ResourceWarning``.
+
+    .. versionchanged:: 2.2
+        Set the ``default_mimetype`` to None to prevent a mimetype being
+        assumed if missing.
+
+    .. versionchanged:: 2.1
+        Response instances cannot be treated as tuples.
+
+    .. versionadded:: 2.0
+        Test client methods always return instances of this class.
+    """
+
+    default_mimetype = None
+    # Don't assume a mimetype, instead use whatever the response provides
+
+    request: Request
+    """A request object with the environ used to make the request that
+    resulted in this response.
+    """
+
+    history: tuple[TestResponse, ...]
+    """A list of intermediate responses. Populated when the test request
+    is made with ``follow_redirects`` enabled.
+    """
+
+    # Tell Pytest to ignore this, it's not a test class.
+    __test__ = False
+
+    def __init__(
+        self,
+        response: t.Iterable[bytes],
+        status: str,
+        headers: Headers,
+        request: Request,
+        history: tuple[TestResponse] = (),  # type: ignore
+        **kwargs: t.Any,
+    ) -> None:
+        super().__init__(response, status, headers, **kwargs)
+        self.request = request
+        self.history = history
+        self._compat_tuple = response, status, headers
+
+    @cached_property
+    def text(self) -> str:
+        """The response data as text. A shortcut for
+        ``response.get_data(as_text=True)``.
+
+        .. versionadded:: 2.1
+        """
+        return self.get_data(as_text=True)
+
+
+@dataclasses.dataclass
+class Cookie:
+    """A cookie key, value, and parameters.
+
+    The class itself is not a public API. Its attributes are documented for inspection
+    with :meth:`.Client.get_cookie` only.
+
+    .. versionadded:: 2.3
+    """
+
+    key: str
+    """The cookie key, encoded as a client would see it."""
+
+    value: str
+    """The cookie key, encoded as a client would see it."""
+
+    decoded_key: str
+    """The cookie key, decoded as the application would set and see it."""
+
+    decoded_value: str
+    """The cookie value, decoded as the application would set and see it."""
+
+    expires: datetime | None
+    """The time at which the cookie is no longer valid."""
+
+    max_age: int | None
+    """The number of seconds from when the cookie was set at which it is
+    no longer valid.
+    """
+
+    domain: str
+    """The domain that the cookie was set for, or the request domain if not set."""
+
+    origin_only: bool
+    """Whether the cookie will be sent for exact domain matches only. This is ``True``
+    if the ``Domain`` parameter was not present.
+    """
+
+    path: str
+    """The path that the cookie was set for."""
+
+    secure: bool | None
+    """The ``Secure`` parameter."""
+
+    http_only: bool | None
+    """The ``HttpOnly`` parameter."""
+
+    same_site: str | None
+    """The ``SameSite`` parameter."""
+
+    def _matches_request(self, server_name: str, path: str) -> bool:
+        return (
+            server_name == self.domain
+            or (
+                not self.origin_only
+                and server_name.endswith(self.domain)
+                and server_name[: -len(self.domain)].endswith(".")
+            )
+        ) and (
+            path == self.path
+            or (
+                path.startswith(self.path)
+                and path[len(self.path) - self.path.endswith("/") :].startswith("/")
+            )
+        )
+
+    def _to_request_header(self) -> str:
+        return f"{self.key}={self.value}"
+
+    @classmethod
+    def _from_response_header(cls, server_name: str, path: str, header: str) -> te.Self:
+        header, _, parameters_str = header.partition(";")
+        key, _, value = header.partition("=")
+        decoded_key, decoded_value = next(parse_cookie(header).items())  # type: ignore[call-overload]
+        params = {}
+
+        for item in parameters_str.split(";"):
+            k, sep, v = item.partition("=")
+            params[k.strip().lower()] = v.strip() if sep else None
+
+        return cls(
+            key=key.strip(),
+            value=value.strip(),
+            decoded_key=decoded_key,
+            decoded_value=decoded_value,
+            expires=parse_date(params.get("expires")),
+            max_age=int(params["max-age"] or 0) if "max-age" in params else None,
+            domain=params.get("domain") or server_name,
+            origin_only="domain" not in params,
+            path=params.get("path") or path.rpartition("/")[0] or "/",
+            secure="secure" in params,
+            http_only="httponly" in params,
+            same_site=params.get("samesite"),
+        )
+
+    @property
+    def _storage_key(self) -> tuple[str, str, str]:
+        return self.domain, self.path, self.decoded_key
+
+    @property
+    def _should_delete(self) -> bool:
+        return self.max_age == 0 or (
+            self.expires is not None and self.expires.timestamp() == 0
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/testapp.py b/backend/venv/lib/python3.12/site-packages/werkzeug/testapp.py
new file mode 100644
index 00000000..cdf7fac1
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/testapp.py
@@ -0,0 +1,194 @@
+"""A small application that can be used to test a WSGI server and check
+it for WSGI compliance.
+"""
+
+from __future__ import annotations
+
+import importlib.metadata
+import os
+import sys
+import typing as t
+from textwrap import wrap
+
+from markupsafe import escape
+
+from .wrappers.request import Request
+from .wrappers.response import Response
+
+TEMPLATE = """\
+<!doctype html>
+<html lang=en>
+<title>WSGI Information</title>
+<style type="text/css">
+  @import url(https://fonts.googleapis.com/css?family=Ubuntu);
+
+  body       { font-family: 'Lucida Grande', 'Lucida Sans Unicode', 'Geneva',
+               'Verdana', sans-serif; background-color: white; color: #000;
+               font-size: 15px; text-align: center; }
+  div.box    { text-align: left; width: 45em; margin: auto; padding: 50px 0;
+               background-color: white; }
+  h1, h2     { font-family: 'Ubuntu', 'Lucida Grande', 'Lucida Sans Unicode',
+               'Geneva', 'Verdana', sans-serif; font-weight: normal; }
+  h1         { margin: 0 0 30px 0; }
+  h2         { font-size: 1.4em; margin: 1em 0 0.5em 0; }
+  table      { width: 100%%; border-collapse: collapse; border: 1px solid #AFC5C9 }
+  table th   { background-color: #AFC1C4; color: white; font-size: 0.72em;
+               font-weight: normal; width: 18em; vertical-align: top;
+               padding: 0.5em 0 0.1em 0.5em; }
+  table td   { border: 1px solid #AFC5C9; padding: 0.1em 0 0.1em 0.5em; }
+  code       { font-family: 'Consolas', 'Monaco', 'Bitstream Vera Sans Mono',
+               monospace; font-size: 0.7em; }
+  ul li      { line-height: 1.5em; }
+  ul.path    { font-size: 0.7em; margin: 0 -30px; padding: 8px 30px;
+               list-style: none; background: #E8EFF0; }
+  ul.path li { line-height: 1.6em; }
+  li.virtual { color: #999; text-decoration: underline; }
+  li.exp     { background: white; }
+</style>
+<div class="box">
+  <h1>WSGI Information</h1>
+  <p>
+    This page displays all available information about the WSGI server and
+    the underlying Python interpreter.
+  <h2 id="python-interpreter">Python Interpreter</h2>
+  <table>
+    <tr>
+      <th>Python Version
+      <td>%(python_version)s
+    <tr>
+      <th>Platform
+      <td>%(platform)s [%(os)s]
+    <tr>
+      <th>API Version
+      <td>%(api_version)s
+    <tr>
+      <th>Byteorder
+      <td>%(byteorder)s
+    <tr>
+      <th>Werkzeug Version
+      <td>%(werkzeug_version)s
+  </table>
+  <h2 id="wsgi-environment">WSGI Environment</h2>
+  <table>%(wsgi_env)s</table>
+  <h2 id="installed-eggs">Installed Eggs</h2>
+  <p>
+    The following python packages were installed on the system as
+    Python eggs:
+  <ul>%(python_eggs)s</ul>
+  <h2 id="sys-path">System Path</h2>
+  <p>
+    The following paths are the current contents of the load path.  The
+    following entries are looked up for Python packages.  Note that not
+    all items in this path are folders.  Gray and underlined items are
+    entries pointing to invalid resources or used by custom import hooks
+    such as the zip importer.
+  <p>
+    Items with a bright background were expanded for display from a relative
+    path.  If you encounter such paths in the output you might want to check
+    your setup as relative paths are usually problematic in multithreaded
+    environments.
+  <ul class="path">%(sys_path)s</ul>
+</div>
+"""
+
+
+def iter_sys_path() -> t.Iterator[tuple[str, bool, bool]]:
+    if os.name == "posix":
+
+        def strip(x: str) -> str:
+            prefix = os.path.expanduser("~")
+            if x.startswith(prefix):
+                x = f"~{x[len(prefix) :]}"
+            return x
+
+    else:
+
+        def strip(x: str) -> str:
+            return x
+
+    cwd = os.path.abspath(os.getcwd())
+    for item in sys.path:
+        path = os.path.join(cwd, item or os.path.curdir)
+        yield strip(os.path.normpath(path)), not os.path.isdir(path), path != item
+
+
+@Request.application
+def test_app(req: Request) -> Response:
+    """Simple test application that dumps the environment.  You can use
+    it to check if Werkzeug is working properly:
+
+    .. sourcecode:: pycon
+
+        >>> from werkzeug.serving import run_simple
+        >>> from werkzeug.testapp import test_app
+        >>> run_simple('localhost', 3000, test_app)
+         * Running on http://localhost:3000/
+
+    The application displays important information from the WSGI environment,
+    the Python interpreter and the installed libraries.
+    """
+    try:
+        import pkg_resources
+    except ImportError:
+        eggs: t.Iterable[t.Any] = ()
+    else:
+        eggs = sorted(
+            pkg_resources.working_set,
+            key=lambda x: x.project_name.lower(),
+        )
+    python_eggs = []
+    for egg in eggs:
+        try:
+            version = egg.version
+        except (ValueError, AttributeError):
+            version = "unknown"
+        python_eggs.append(
+            f"<li>{escape(egg.project_name)} <small>[{escape(version)}]</small>"
+        )
+
+    wsgi_env = []
+    sorted_environ = sorted(req.environ.items(), key=lambda x: repr(x[0]).lower())
+    for key, value in sorted_environ:
+        value = "".join(wrap(str(escape(repr(value)))))
+        wsgi_env.append(f"<tr><th>{escape(key)}<td><code>{value}</code>")
+
+    sys_path = []
+    for item, virtual, expanded in iter_sys_path():
+        css = []
+        if virtual:
+            css.append("virtual")
+        if expanded:
+            css.append("exp")
+        class_str = f' class="{" ".join(css)}"' if css else ""
+        sys_path.append(f"<li{class_str}>{escape(item)}")
+
+    context = {
+        "python_version": "<br>".join(escape(sys.version).splitlines()),
+        "platform": escape(sys.platform),
+        "os": escape(os.name),
+        "api_version": sys.api_version,
+        "byteorder": sys.byteorder,
+        "werkzeug_version": _get_werkzeug_version(),
+        "python_eggs": "\n".join(python_eggs),
+        "wsgi_env": "\n".join(wsgi_env),
+        "sys_path": "\n".join(sys_path),
+    }
+    return Response(TEMPLATE % context, mimetype="text/html")
+
+
+_werkzeug_version = ""
+
+
+def _get_werkzeug_version() -> str:
+    global _werkzeug_version
+
+    if not _werkzeug_version:
+        _werkzeug_version = importlib.metadata.version("werkzeug")
+
+    return _werkzeug_version
+
+
+if __name__ == "__main__":
+    from .serving import run_simple
+
+    run_simple("localhost", 5000, test_app, use_reloader=True)
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/urls.py b/backend/venv/lib/python3.12/site-packages/werkzeug/urls.py
new file mode 100644
index 00000000..5bffe392
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/urls.py
@@ -0,0 +1,203 @@
+from __future__ import annotations
+
+import codecs
+import re
+import typing as t
+import urllib.parse
+from urllib.parse import quote
+from urllib.parse import unquote
+from urllib.parse import urlencode
+from urllib.parse import urlsplit
+from urllib.parse import urlunsplit
+
+from .datastructures import iter_multi_items
+
+
+def _codec_error_url_quote(e: UnicodeError) -> tuple[str, int]:
+    """Used in :func:`uri_to_iri` after unquoting to re-quote any
+    invalid bytes.
+    """
+    # the docs state that UnicodeError does have these attributes,
+    # but mypy isn't picking them up
+    out = quote(e.object[e.start : e.end], safe="")  # type: ignore
+    return out, e.end  # type: ignore
+
+
+codecs.register_error("werkzeug.url_quote", _codec_error_url_quote)
+
+
+def _make_unquote_part(name: str, chars: str) -> t.Callable[[str], str]:
+    """Create a function that unquotes all percent encoded characters except those
+    given. This allows working with unquoted characters if possible while not changing
+    the meaning of a given part of a URL.
+    """
+    choices = "|".join(f"{ord(c):02X}" for c in sorted(chars))
+    pattern = re.compile(f"((?:%(?:{choices}))+)", re.I)
+
+    def _unquote_partial(value: str) -> str:
+        parts = iter(pattern.split(value))
+        out = []
+
+        for part in parts:
+            out.append(unquote(part, "utf-8", "werkzeug.url_quote"))
+            out.append(next(parts, ""))
+
+        return "".join(out)
+
+    _unquote_partial.__name__ = f"_unquote_{name}"
+    return _unquote_partial
+
+
+# characters that should remain quoted in URL parts
+# based on https://url.spec.whatwg.org/#percent-encoded-bytes
+# always keep all controls, space, and % quoted
+_always_unsafe = bytes((*range(0x21), 0x25, 0x7F)).decode()
+_unquote_fragment = _make_unquote_part("fragment", _always_unsafe)
+_unquote_query = _make_unquote_part("query", _always_unsafe + "&=+#")
+_unquote_path = _make_unquote_part("path", _always_unsafe + "/?#")
+_unquote_user = _make_unquote_part("user", _always_unsafe + ":@/?#")
+
+
+def uri_to_iri(uri: str) -> str:
+    """Convert a URI to an IRI. All valid UTF-8 characters are unquoted,
+    leaving all reserved and invalid characters quoted. If the URL has
+    a domain, it is decoded from Punycode.
+
+    >>> uri_to_iri("http://xn--n3h.net/p%C3%A5th?q=%C3%A8ry%DF")
+    'http://\\u2603.net/p\\xe5th?q=\\xe8ry%DF'
+
+    :param uri: The URI to convert.
+
+    .. versionchanged:: 3.0
+        Passing a tuple or bytes, and the ``charset`` and ``errors`` parameters,
+        are removed.
+
+    .. versionchanged:: 2.3
+        Which characters remain quoted is specific to each part of the URL.
+
+    .. versionchanged:: 0.15
+        All reserved and invalid characters remain quoted. Previously,
+        only some reserved characters were preserved, and invalid bytes
+        were replaced instead of left quoted.
+
+    .. versionadded:: 0.6
+    """
+    parts = urlsplit(uri)
+    path = _unquote_path(parts.path)
+    query = _unquote_query(parts.query)
+    fragment = _unquote_fragment(parts.fragment)
+
+    if parts.hostname:
+        netloc = _decode_idna(parts.hostname)
+    else:
+        netloc = ""
+
+    if ":" in netloc:
+        netloc = f"[{netloc}]"
+
+    if parts.port:
+        netloc = f"{netloc}:{parts.port}"
+
+    if parts.username:
+        auth = _unquote_user(parts.username)
+
+        if parts.password:
+            password = _unquote_user(parts.password)
+            auth = f"{auth}:{password}"
+
+        netloc = f"{auth}@{netloc}"
+
+    return urlunsplit((parts.scheme, netloc, path, query, fragment))
+
+
+def iri_to_uri(iri: str) -> str:
+    """Convert an IRI to a URI. All non-ASCII and unsafe characters are
+    quoted. If the URL has a domain, it is encoded to Punycode.
+
+    >>> iri_to_uri('http://\\u2603.net/p\\xe5th?q=\\xe8ry%DF')
+    'http://xn--n3h.net/p%C3%A5th?q=%C3%A8ry%DF'
+
+    :param iri: The IRI to convert.
+
+    .. versionchanged:: 3.0
+        Passing a tuple or bytes, the ``charset`` and ``errors`` parameters,
+        and the ``safe_conversion`` parameter, are removed.
+
+    .. versionchanged:: 2.3
+        Which characters remain unquoted is specific to each part of the URL.
+
+    .. versionchanged:: 0.15
+        All reserved characters remain unquoted. Previously, only some reserved
+        characters were left unquoted.
+
+    .. versionchanged:: 0.9.6
+       The ``safe_conversion`` parameter was added.
+
+    .. versionadded:: 0.6
+    """
+    parts = urlsplit(iri)
+    # safe = https://url.spec.whatwg.org/#url-path-segment-string
+    # as well as percent for things that are already quoted
+    path = quote(parts.path, safe="%!$&'()*+,/:;=@")
+    query = quote(parts.query, safe="%!$&'()*+,/:;=?@")
+    fragment = quote(parts.fragment, safe="%!#$&'()*+,/:;=?@")
+
+    if parts.hostname:
+        netloc = parts.hostname.encode("idna").decode("ascii")
+    else:
+        netloc = ""
+
+    if ":" in netloc:
+        netloc = f"[{netloc}]"
+
+    if parts.port:
+        netloc = f"{netloc}:{parts.port}"
+
+    if parts.username:
+        auth = quote(parts.username, safe="%!$&'()*+,;=")
+
+        if parts.password:
+            password = quote(parts.password, safe="%!$&'()*+,;=")
+            auth = f"{auth}:{password}"
+
+        netloc = f"{auth}@{netloc}"
+
+    return urlunsplit((parts.scheme, netloc, path, query, fragment))
+
+
+# Python < 3.12
+# itms-services was worked around in previous iri_to_uri implementations, but
+# we can tell Python directly that it needs to preserve the //.
+if "itms-services" not in urllib.parse.uses_netloc:
+    urllib.parse.uses_netloc.append("itms-services")
+
+
+def _decode_idna(domain: str) -> str:
+    try:
+        data = domain.encode("ascii")
+    except UnicodeEncodeError:
+        # If the domain is not ASCII, it's decoded already.
+        return domain
+
+    try:
+        # Try decoding in one shot.
+        return data.decode("idna")
+    except UnicodeDecodeError:
+        pass
+
+    # Decode each part separately, leaving invalid parts as punycode.
+    parts = []
+
+    for part in data.split(b"."):
+        try:
+            parts.append(part.decode("idna"))
+        except UnicodeDecodeError:
+            parts.append(part.decode("ascii"))
+
+    return ".".join(parts)
+
+
+def _urlencode(query: t.Mapping[str, str] | t.Iterable[tuple[str, str]]) -> str:
+    items = [x for x in iter_multi_items(query) if x[1] is not None]
+    # safe = https://url.spec.whatwg.org/#percent-encoded-bytes
+    return urlencode(items, safe="!$'()*,/:;?@")
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/user_agent.py b/backend/venv/lib/python3.12/site-packages/werkzeug/user_agent.py
new file mode 100644
index 00000000..17e5d3fd
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/user_agent.py
@@ -0,0 +1,47 @@
+from __future__ import annotations
+
+
+class UserAgent:
+    """Represents a parsed user agent header value.
+
+    The default implementation does no parsing, only the :attr:`string`
+    attribute is set. A subclass may parse the string to set the
+    common attributes or expose other information. Set
+    :attr:`werkzeug.wrappers.Request.user_agent_class` to use a
+    subclass.
+
+    :param string: The header value to parse.
+
+    .. versionadded:: 2.0
+        This replaces the previous ``useragents`` module, but does not
+        provide a built-in parser.
+    """
+
+    platform: str | None = None
+    """The OS name, if it could be parsed from the string."""
+
+    browser: str | None = None
+    """The browser name, if it could be parsed from the string."""
+
+    version: str | None = None
+    """The browser version, if it could be parsed from the string."""
+
+    language: str | None = None
+    """The browser language, if it could be parsed from the string."""
+
+    def __init__(self, string: str) -> None:
+        self.string: str = string
+        """The original header value."""
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__} {self.browser}/{self.version}>"
+
+    def __str__(self) -> str:
+        return self.string
+
+    def __bool__(self) -> bool:
+        return bool(self.browser)
+
+    def to_header(self) -> str:
+        """Convert to a header value."""
+        return self.string
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/utils.py b/backend/venv/lib/python3.12/site-packages/werkzeug/utils.py
new file mode 100644
index 00000000..3d3bbf06
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/utils.py
@@ -0,0 +1,691 @@
+from __future__ import annotations
+
+import io
+import mimetypes
+import os
+import pkgutil
+import re
+import sys
+import typing as t
+import unicodedata
+from datetime import datetime
+from time import time
+from urllib.parse import quote
+from zlib import adler32
+
+from markupsafe import escape
+
+from ._internal import _DictAccessorProperty
+from ._internal import _missing
+from ._internal import _TAccessorValue
+from .datastructures import Headers
+from .exceptions import NotFound
+from .exceptions import RequestedRangeNotSatisfiable
+from .security import safe_join
+from .wsgi import wrap_file
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .wrappers.request import Request
+    from .wrappers.response import Response
+
+_T = t.TypeVar("_T")
+
+_entity_re = re.compile(r"&([^;]+);")
+_filename_ascii_strip_re = re.compile(r"[^A-Za-z0-9_.-]")
+_windows_device_files = {
+    "CON",
+    "PRN",
+    "AUX",
+    "NUL",
+    *(f"COM{i}" for i in range(10)),
+    *(f"LPT{i}" for i in range(10)),
+}
+
+
+class cached_property(property, t.Generic[_T]):
+    """A :func:`property` that is only evaluated once. Subsequent access
+    returns the cached value. Setting the property sets the cached
+    value. Deleting the property clears the cached value, accessing it
+    again will evaluate it again.
+
+    .. code-block:: python
+
+        class Example:
+            @cached_property
+            def value(self):
+                # calculate something important here
+                return 42
+
+        e = Example()
+        e.value  # evaluates
+        e.value  # uses cache
+        e.value = 16  # sets cache
+        del e.value  # clears cache
+
+    If the class defines ``__slots__``, it must add ``_cache_{name}`` as
+    a slot. Alternatively, it can add ``__dict__``, but that's usually
+    not desirable.
+
+    .. versionchanged:: 2.1
+        Works with ``__slots__``.
+
+    .. versionchanged:: 2.0
+        ``del obj.name`` clears the cached value.
+    """
+
+    def __init__(
+        self,
+        fget: t.Callable[[t.Any], _T],
+        name: str | None = None,
+        doc: str | None = None,
+    ) -> None:
+        super().__init__(fget, doc=doc)
+        self.__name__ = name or fget.__name__
+        self.slot_name = f"_cache_{self.__name__}"
+        self.__module__ = fget.__module__
+
+    def __set__(self, obj: object, value: _T) -> None:
+        if hasattr(obj, "__dict__"):
+            obj.__dict__[self.__name__] = value
+        else:
+            setattr(obj, self.slot_name, value)
+
+    def __get__(self, obj: object, type: type = None) -> _T:  # type: ignore
+        if obj is None:
+            return self  # type: ignore
+
+        obj_dict = getattr(obj, "__dict__", None)
+
+        if obj_dict is not None:
+            value: _T = obj_dict.get(self.__name__, _missing)
+        else:
+            value = getattr(obj, self.slot_name, _missing)  # type: ignore[arg-type]
+
+        if value is _missing:
+            value = self.fget(obj)  # type: ignore
+
+            if obj_dict is not None:
+                obj.__dict__[self.__name__] = value
+            else:
+                setattr(obj, self.slot_name, value)
+
+        return value
+
+    def __delete__(self, obj: object) -> None:
+        if hasattr(obj, "__dict__"):
+            del obj.__dict__[self.__name__]
+        else:
+            setattr(obj, self.slot_name, _missing)
+
+
+class environ_property(_DictAccessorProperty[_TAccessorValue]):
+    """Maps request attributes to environment variables. This works not only
+    for the Werkzeug request object, but also any other class with an
+    environ attribute:
+
+    >>> class Test(object):
+    ...     environ = {'key': 'value'}
+    ...     test = environ_property('key')
+    >>> var = Test()
+    >>> var.test
+    'value'
+
+    If you pass it a second value it's used as default if the key does not
+    exist, the third one can be a converter that takes a value and converts
+    it.  If it raises :exc:`ValueError` or :exc:`TypeError` the default value
+    is used. If no default value is provided `None` is used.
+
+    Per default the property is read only.  You have to explicitly enable it
+    by passing ``read_only=False`` to the constructor.
+    """
+
+    read_only = True
+
+    def lookup(self, obj: Request) -> WSGIEnvironment:
+        return obj.environ
+
+
+class header_property(_DictAccessorProperty[_TAccessorValue]):
+    """Like `environ_property` but for headers."""
+
+    def lookup(self, obj: Request | Response) -> Headers:  # type: ignore[override]
+        return obj.headers
+
+
+# https://cgit.freedesktop.org/xdg/shared-mime-info/tree/freedesktop.org.xml.in
+# https://www.iana.org/assignments/media-types/media-types.xhtml
+# Types listed in the XDG mime info that have a charset in the IANA registration.
+_charset_mimetypes = {
+    "application/ecmascript",
+    "application/javascript",
+    "application/sql",
+    "application/xml",
+    "application/xml-dtd",
+    "application/xml-external-parsed-entity",
+}
+
+
+def get_content_type(mimetype: str, charset: str) -> str:
+    """Returns the full content type string with charset for a mimetype.
+
+    If the mimetype represents text, the charset parameter will be
+    appended, otherwise the mimetype is returned unchanged.
+
+    :param mimetype: The mimetype to be used as content type.
+    :param charset: The charset to be appended for text mimetypes.
+    :return: The content type.
+
+    .. versionchanged:: 0.15
+        Any type that ends with ``+xml`` gets a charset, not just those
+        that start with ``application/``. Known text types such as
+        ``application/javascript`` are also given charsets.
+    """
+    if (
+        mimetype.startswith("text/")
+        or mimetype in _charset_mimetypes
+        or mimetype.endswith("+xml")
+    ):
+        mimetype += f"; charset={charset}"
+
+    return mimetype
+
+
+def secure_filename(filename: str) -> str:
+    r"""Pass it a filename and it will return a secure version of it.  This
+    filename can then safely be stored on a regular file system and passed
+    to :func:`os.path.join`.  The filename returned is an ASCII only string
+    for maximum portability.
+
+    On windows systems the function also makes sure that the file is not
+    named after one of the special device files.
+
+    >>> secure_filename("My cool movie.mov")
+    'My_cool_movie.mov'
+    >>> secure_filename("../../../etc/passwd")
+    'etc_passwd'
+    >>> secure_filename('i contain cool \xfcml\xe4uts.txt')
+    'i_contain_cool_umlauts.txt'
+
+    The function might return an empty filename.  It's your responsibility
+    to ensure that the filename is unique and that you abort or
+    generate a random filename if the function returned an empty one.
+
+    .. versionadded:: 0.5
+
+    :param filename: the filename to secure
+    """
+    filename = unicodedata.normalize("NFKD", filename)
+    filename = filename.encode("ascii", "ignore").decode("ascii")
+
+    for sep in os.sep, os.path.altsep:
+        if sep:
+            filename = filename.replace(sep, " ")
+    filename = str(_filename_ascii_strip_re.sub("", "_".join(filename.split()))).strip(
+        "._"
+    )
+
+    # on nt a couple of special files are present in each folder.  We
+    # have to ensure that the target file is not such a filename.  In
+    # this case we prepend an underline
+    if (
+        os.name == "nt"
+        and filename
+        and filename.split(".")[0].upper() in _windows_device_files
+    ):
+        filename = f"_{filename}"
+
+    return filename
+
+
+def redirect(
+    location: str, code: int = 302, Response: type[Response] | None = None
+) -> Response:
+    """Returns a response object (a WSGI application) that, if called,
+    redirects the client to the target location. Supported codes are
+    301, 302, 303, 305, 307, and 308. 300 is not supported because
+    it's not a real redirect and 304 because it's the answer for a
+    request with a request with defined If-Modified-Since headers.
+
+    .. versionadded:: 0.6
+       The location can now be a unicode string that is encoded using
+       the :func:`iri_to_uri` function.
+
+    .. versionadded:: 0.10
+        The class used for the Response object can now be passed in.
+
+    :param location: the location the response should redirect to.
+    :param code: the redirect status code. defaults to 302.
+    :param class Response: a Response class to use when instantiating a
+        response. The default is :class:`werkzeug.wrappers.Response` if
+        unspecified.
+    """
+    if Response is None:
+        from .wrappers import Response
+
+    html_location = escape(location)
+    response = Response(  # type: ignore[misc]
+        "<!doctype html>\n"
+        "<html lang=en>\n"
+        "<title>Redirecting...</title>\n"
+        "<h1>Redirecting...</h1>\n"
+        "<p>You should be redirected automatically to the target URL: "
+        f'<a href="{html_location}">{html_location}</a>. If not, click the link.\n',
+        code,
+        mimetype="text/html",
+    )
+    response.headers["Location"] = location
+    return response
+
+
+def append_slash_redirect(environ: WSGIEnvironment, code: int = 308) -> Response:
+    """Redirect to the current URL with a slash appended.
+
+    If the current URL is ``/user/42``, the redirect URL will be
+    ``42/``. When joined to the current URL during response
+    processing or by the browser, this will produce ``/user/42/``.
+
+    The behavior is undefined if the path ends with a slash already. If
+    called unconditionally on a URL, it may produce a redirect loop.
+
+    :param environ: Use the path and query from this WSGI environment
+        to produce the redirect URL.
+    :param code: the status code for the redirect.
+
+    .. versionchanged:: 2.1
+        Produce a relative URL that only modifies the last segment.
+        Relevant when the current path has multiple segments.
+
+    .. versionchanged:: 2.1
+        The default status code is 308 instead of 301. This preserves
+        the request method and body.
+    """
+    tail = environ["PATH_INFO"].rpartition("/")[2]
+
+    if not tail:
+        new_path = "./"
+    else:
+        new_path = f"{tail}/"
+
+    query_string = environ.get("QUERY_STRING")
+
+    if query_string:
+        new_path = f"{new_path}?{query_string}"
+
+    return redirect(new_path, code)
+
+
+def send_file(
+    path_or_file: os.PathLike[str] | str | t.IO[bytes],
+    environ: WSGIEnvironment,
+    mimetype: str | None = None,
+    as_attachment: bool = False,
+    download_name: str | None = None,
+    conditional: bool = True,
+    etag: bool | str = True,
+    last_modified: datetime | int | float | None = None,
+    max_age: None | (int | t.Callable[[str | None], int | None]) = None,
+    use_x_sendfile: bool = False,
+    response_class: type[Response] | None = None,
+    _root_path: os.PathLike[str] | str | None = None,
+) -> Response:
+    """Send the contents of a file to the client.
+
+    The first argument can be a file path or a file-like object. Paths
+    are preferred in most cases because Werkzeug can manage the file and
+    get extra information from the path. Passing a file-like object
+    requires that the file is opened in binary mode, and is mostly
+    useful when building a file in memory with :class:`io.BytesIO`.
+
+    Never pass file paths provided by a user. The path is assumed to be
+    trusted, so a user could craft a path to access a file you didn't
+    intend. Use :func:`send_from_directory` to safely serve user-provided paths.
+
+    If the WSGI server sets a ``file_wrapper`` in ``environ``, it is
+    used, otherwise Werkzeug's built-in wrapper is used. Alternatively,
+    if the HTTP server supports ``X-Sendfile``, ``use_x_sendfile=True``
+    will tell the server to send the given path, which is much more
+    efficient than reading it in Python.
+
+    :param path_or_file: The path to the file to send, relative to the
+        current working directory if a relative path is given.
+        Alternatively, a file-like object opened in binary mode. Make
+        sure the file pointer is seeked to the start of the data.
+    :param environ: The WSGI environ for the current request.
+    :param mimetype: The MIME type to send for the file. If not
+        provided, it will try to detect it from the file name.
+    :param as_attachment: Indicate to a browser that it should offer to
+        save the file instead of displaying it.
+    :param download_name: The default name browsers will use when saving
+        the file. Defaults to the passed file name.
+    :param conditional: Enable conditional and range responses based on
+        request headers. Requires passing a file path and ``environ``.
+    :param etag: Calculate an ETag for the file, which requires passing
+        a file path. Can also be a string to use instead.
+    :param last_modified: The last modified time to send for the file,
+        in seconds. If not provided, it will try to detect it from the
+        file path.
+    :param max_age: How long the client should cache the file, in
+        seconds. If set, ``Cache-Control`` will be ``public``, otherwise
+        it will be ``no-cache`` to prefer conditional caching.
+    :param use_x_sendfile: Set the ``X-Sendfile`` header to let the
+        server to efficiently send the file. Requires support from the
+        HTTP server. Requires passing a file path.
+    :param response_class: Build the response using this class. Defaults
+        to :class:`~werkzeug.wrappers.Response`.
+    :param _root_path: Do not use. For internal use only. Use
+        :func:`send_from_directory` to safely send files under a path.
+
+    .. versionchanged:: 2.0.2
+        ``send_file`` only sets a detected ``Content-Encoding`` if
+        ``as_attachment`` is disabled.
+
+    .. versionadded:: 2.0
+        Adapted from Flask's implementation.
+
+    .. versionchanged:: 2.0
+        ``download_name`` replaces Flask's ``attachment_filename``
+         parameter. If ``as_attachment=False``, it is passed with
+         ``Content-Disposition: inline`` instead.
+
+    .. versionchanged:: 2.0
+        ``max_age`` replaces Flask's ``cache_timeout`` parameter.
+        ``conditional`` is enabled and ``max_age`` is not set by
+        default.
+
+    .. versionchanged:: 2.0
+        ``etag`` replaces Flask's ``add_etags`` parameter. It can be a
+        string to use instead of generating one.
+
+    .. versionchanged:: 2.0
+        If an encoding is returned when guessing ``mimetype`` from
+        ``download_name``, set the ``Content-Encoding`` header.
+    """
+    if response_class is None:
+        from .wrappers import Response
+
+        response_class = Response
+
+    path: str | None = None
+    file: t.IO[bytes] | None = None
+    size: int | None = None
+    mtime: float | None = None
+    headers = Headers()
+
+    if isinstance(path_or_file, (os.PathLike, str)) or hasattr(
+        path_or_file, "__fspath__"
+    ):
+        path_or_file = t.cast("t.Union[os.PathLike[str], str]", path_or_file)
+
+        # Flask will pass app.root_path, allowing its send_file wrapper
+        # to not have to deal with paths.
+        if _root_path is not None:
+            path = os.path.join(_root_path, path_or_file)
+        else:
+            path = os.path.abspath(path_or_file)
+
+        stat = os.stat(path)
+        size = stat.st_size
+        mtime = stat.st_mtime
+    else:
+        file = path_or_file
+
+    if download_name is None and path is not None:
+        download_name = os.path.basename(path)
+
+    if mimetype is None:
+        if download_name is None:
+            raise TypeError(
+                "Unable to detect the MIME type because a file name is"
+                " not available. Either set 'download_name', pass a"
+                " path instead of a file, or set 'mimetype'."
+            )
+
+        mimetype, encoding = mimetypes.guess_type(download_name)
+
+        if mimetype is None:
+            mimetype = "application/octet-stream"
+
+        # Don't send encoding for attachments, it causes browsers to
+        # save decompress tar.gz files.
+        if encoding is not None and not as_attachment:
+            headers.set("Content-Encoding", encoding)
+
+    if download_name is not None:
+        try:
+            download_name.encode("ascii")
+        except UnicodeEncodeError:
+            simple = unicodedata.normalize("NFKD", download_name)
+            simple = simple.encode("ascii", "ignore").decode("ascii")
+            # safe = RFC 5987 attr-char
+            quoted = quote(download_name, safe="!#$&+-.^_`|~")
+            names = {"filename": simple, "filename*": f"UTF-8''{quoted}"}
+        else:
+            names = {"filename": download_name}
+
+        value = "attachment" if as_attachment else "inline"
+        headers.set("Content-Disposition", value, **names)
+    elif as_attachment:
+        raise TypeError(
+            "No name provided for attachment. Either set"
+            " 'download_name' or pass a path instead of a file."
+        )
+
+    if use_x_sendfile and path is not None:
+        headers["X-Sendfile"] = path
+        data = None
+    else:
+        if file is None:
+            file = open(path, "rb")  # type: ignore
+        elif isinstance(file, io.BytesIO):
+            size = file.getbuffer().nbytes
+        elif isinstance(file, io.TextIOBase):
+            raise ValueError("Files must be opened in binary mode or use BytesIO.")
+
+        data = wrap_file(environ, file)
+
+    rv = response_class(
+        data, mimetype=mimetype, headers=headers, direct_passthrough=True
+    )
+
+    if size is not None:
+        rv.content_length = size
+
+    if last_modified is not None:
+        rv.last_modified = last_modified  # type: ignore
+    elif mtime is not None:
+        rv.last_modified = mtime  # type: ignore
+
+    rv.cache_control.no_cache = True
+
+    # Flask will pass app.get_send_file_max_age, allowing its send_file
+    # wrapper to not have to deal with paths.
+    if callable(max_age):
+        max_age = max_age(path)
+
+    if max_age is not None:
+        if max_age > 0:
+            rv.cache_control.no_cache = None
+            rv.cache_control.public = True
+
+        rv.cache_control.max_age = max_age
+        rv.expires = int(time() + max_age)  # type: ignore
+
+    if isinstance(etag, str):
+        rv.set_etag(etag)
+    elif etag and path is not None:
+        check = adler32(path.encode()) & 0xFFFFFFFF
+        rv.set_etag(f"{mtime}-{size}-{check}")
+
+    if conditional:
+        try:
+            rv = rv.make_conditional(environ, accept_ranges=True, complete_length=size)
+        except RequestedRangeNotSatisfiable:
+            if file is not None:
+                file.close()
+
+            raise
+
+        # Some x-sendfile implementations incorrectly ignore the 304
+        # status code and send the file anyway.
+        if rv.status_code == 304:
+            rv.headers.pop("x-sendfile", None)
+
+    return rv
+
+
+def send_from_directory(
+    directory: os.PathLike[str] | str,
+    path: os.PathLike[str] | str,
+    environ: WSGIEnvironment,
+    **kwargs: t.Any,
+) -> Response:
+    """Send a file from within a directory using :func:`send_file`.
+
+    This is a secure way to serve files from a folder, such as static
+    files or uploads. Uses :func:`~werkzeug.security.safe_join` to
+    ensure the path coming from the client is not maliciously crafted to
+    point outside the specified directory.
+
+    If the final path does not point to an existing regular file,
+    returns a 404 :exc:`~werkzeug.exceptions.NotFound` error.
+
+    :param directory: The directory that ``path`` must be located under. This *must not*
+        be a value provided by the client, otherwise it becomes insecure.
+    :param path: The path to the file to send, relative to ``directory``. This is the
+        part of the path provided by the client, which is checked for security.
+    :param environ: The WSGI environ for the current request.
+    :param kwargs: Arguments to pass to :func:`send_file`.
+
+    .. versionadded:: 2.0
+        Adapted from Flask's implementation.
+    """
+    path_str = safe_join(os.fspath(directory), os.fspath(path))
+
+    if path_str is None:
+        raise NotFound()
+
+    # Flask will pass app.root_path, allowing its send_from_directory
+    # wrapper to not have to deal with paths.
+    if "_root_path" in kwargs:
+        path_str = os.path.join(kwargs["_root_path"], path_str)
+
+    if not os.path.isfile(path_str):
+        raise NotFound()
+
+    return send_file(path_str, environ, **kwargs)
+
+
+def import_string(import_name: str, silent: bool = False) -> t.Any:
+    """Imports an object based on a string.  This is useful if you want to
+    use import paths as endpoints or something similar.  An import path can
+    be specified either in dotted notation (``xml.sax.saxutils.escape``)
+    or with a colon as object delimiter (``xml.sax.saxutils:escape``).
+
+    If `silent` is True the return value will be `None` if the import fails.
+
+    :param import_name: the dotted name for the object to import.
+    :param silent: if set to `True` import errors are ignored and
+                   `None` is returned instead.
+    :return: imported object
+    """
+    import_name = import_name.replace(":", ".")
+    try:
+        try:
+            __import__(import_name)
+        except ImportError:
+            if "." not in import_name:
+                raise
+        else:
+            return sys.modules[import_name]
+
+        module_name, obj_name = import_name.rsplit(".", 1)
+        module = __import__(module_name, globals(), locals(), [obj_name])
+        try:
+            return getattr(module, obj_name)
+        except AttributeError as e:
+            raise ImportError(e) from None
+
+    except ImportError as e:
+        if not silent:
+            raise ImportStringError(import_name, e).with_traceback(
+                sys.exc_info()[2]
+            ) from None
+
+    return None
+
+
+def find_modules(
+    import_path: str, include_packages: bool = False, recursive: bool = False
+) -> t.Iterator[str]:
+    """Finds all the modules below a package.  This can be useful to
+    automatically import all views / controllers so that their metaclasses /
+    function decorators have a chance to register themselves on the
+    application.
+
+    Packages are not returned unless `include_packages` is `True`.  This can
+    also recursively list modules but in that case it will import all the
+    packages to get the correct load path of that module.
+
+    :param import_path: the dotted name for the package to find child modules.
+    :param include_packages: set to `True` if packages should be returned, too.
+    :param recursive: set to `True` if recursion should happen.
+    :return: generator
+    """
+    module = import_string(import_path)
+    path = getattr(module, "__path__", None)
+    if path is None:
+        raise ValueError(f"{import_path!r} is not a package")
+    basename = f"{module.__name__}."
+    for _importer, modname, ispkg in pkgutil.iter_modules(path):
+        modname = basename + modname
+        if ispkg:
+            if include_packages:
+                yield modname
+            if recursive:
+                yield from find_modules(modname, include_packages, True)
+        else:
+            yield modname
+
+
+class ImportStringError(ImportError):
+    """Provides information about a failed :func:`import_string` attempt."""
+
+    #: String in dotted notation that failed to be imported.
+    import_name: str
+    #: Wrapped exception.
+    exception: BaseException
+
+    def __init__(self, import_name: str, exception: BaseException) -> None:
+        self.import_name = import_name
+        self.exception = exception
+        msg = import_name
+        name = ""
+        tracked = []
+        for part in import_name.replace(":", ".").split("."):
+            name = f"{name}.{part}" if name else part
+            imported = import_string(name, silent=True)
+            if imported:
+                tracked.append((name, getattr(imported, "__file__", None)))
+            else:
+                track = [f"- {n!r} found in {i!r}." for n, i in tracked]
+                track.append(f"- {name!r} not found.")
+                track_str = "\n".join(track)
+                msg = (
+                    f"import_string() failed for {import_name!r}. Possible reasons"
+                    f" are:\n\n"
+                    "- missing __init__.py in a package;\n"
+                    "- package or module path not included in sys.path;\n"
+                    "- duplicated package or module name taking precedence in"
+                    " sys.path;\n"
+                    "- missing module, class, function or variable;\n\n"
+                    f"Debugged import:\n\n{track_str}\n\n"
+                    f"Original exception:\n\n{type(exception).__name__}: {exception}"
+                )
+                break
+
+        super().__init__(msg)
+
+    def __repr__(self) -> str:
+        return f"<{type(self).__name__}({self.import_name!r}, {self.exception!r})>"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__init__.py b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__init__.py
new file mode 100644
index 00000000..b36f228f
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__init__.py
@@ -0,0 +1,3 @@
+from .request import Request as Request
+from .response import Response as Response
+from .response import ResponseStream as ResponseStream
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/__init__.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 00000000..a40fc525
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/__init__.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/request.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/request.cpython-312.pyc
new file mode 100644
index 00000000..b5970921
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/request.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/response.cpython-312.pyc b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/response.cpython-312.pyc
new file mode 100644
index 00000000..32fc1b8a
Binary files /dev/null and b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/__pycache__/response.cpython-312.pyc differ
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/request.py b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/request.py
new file mode 100644
index 00000000..719a3bc0
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/request.py
@@ -0,0 +1,650 @@
+from __future__ import annotations
+
+import collections.abc as cabc
+import functools
+import json
+import typing as t
+from io import BytesIO
+
+from .._internal import _wsgi_decoding_dance
+from ..datastructures import CombinedMultiDict
+from ..datastructures import EnvironHeaders
+from ..datastructures import FileStorage
+from ..datastructures import ImmutableMultiDict
+from ..datastructures import iter_multi_items
+from ..datastructures import MultiDict
+from ..exceptions import BadRequest
+from ..exceptions import UnsupportedMediaType
+from ..formparser import default_stream_factory
+from ..formparser import FormDataParser
+from ..sansio.request import Request as _SansIORequest
+from ..utils import cached_property
+from ..utils import environ_property
+from ..wsgi import _get_server
+from ..wsgi import get_input_stream
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+class Request(_SansIORequest):
+    """Represents an incoming WSGI HTTP request, with headers and body
+    taken from the WSGI environment. Has properties and methods for
+    using the functionality defined by various HTTP specs. The data in
+    requests object is read-only.
+
+    Text data is assumed to use UTF-8 encoding, which should be true for
+    the vast majority of modern clients. Using an encoding set by the
+    client is unsafe in Python due to extra encodings it provides, such
+    as ``zip``. To change the assumed encoding, subclass and replace
+    :attr:`charset`.
+
+    :param environ: The WSGI environ is generated by the WSGI server and
+        contains information about the server configuration and client
+        request.
+    :param populate_request: Add this request object to the WSGI environ
+        as ``environ['werkzeug.request']``. Can be useful when
+        debugging.
+    :param shallow: Makes reading from :attr:`stream` (and any method
+        that would read from it) raise a :exc:`RuntimeError`. Useful to
+        prevent consuming the form data in middleware, which would make
+        it unavailable to the final application.
+
+    .. versionchanged:: 3.0
+        The ``charset``, ``url_charset``, and ``encoding_errors`` parameters
+        were removed.
+
+    .. versionchanged:: 2.1
+        Old ``BaseRequest`` and mixin classes were removed.
+
+    .. versionchanged:: 2.1
+        Remove the ``disable_data_descriptor`` attribute.
+
+    .. versionchanged:: 2.0
+        Combine ``BaseRequest`` and mixins into a single ``Request``
+        class.
+
+    .. versionchanged:: 0.5
+        Read-only mode is enforced with immutable classes for all data.
+    """
+
+    #: the maximum content length.  This is forwarded to the form data
+    #: parsing function (:func:`parse_form_data`).  When set and the
+    #: :attr:`form` or :attr:`files` attribute is accessed and the
+    #: parsing fails because more than the specified value is transmitted
+    #: a :exc:`~werkzeug.exceptions.RequestEntityTooLarge` exception is raised.
+    #:
+    #: .. versionadded:: 0.5
+    max_content_length: int | None = None
+
+    #: the maximum form field size.  This is forwarded to the form data
+    #: parsing function (:func:`parse_form_data`).  When set and the
+    #: :attr:`form` or :attr:`files` attribute is accessed and the
+    #: data in memory for post data is longer than the specified value a
+    #: :exc:`~werkzeug.exceptions.RequestEntityTooLarge` exception is raised.
+    #:
+    #: .. versionchanged:: 3.1
+    #:     Defaults to 500kB instead of unlimited.
+    #:
+    #: .. versionadded:: 0.5
+    max_form_memory_size: int | None = 500_000
+
+    #: The maximum number of multipart parts to parse, passed to
+    #: :attr:`form_data_parser_class`. Parsing form data with more than this
+    #: many parts will raise :exc:`~.RequestEntityTooLarge`.
+    #:
+    #: .. versionadded:: 2.2.3
+    max_form_parts = 1000
+
+    #: The form data parser that should be used.  Can be replaced to customize
+    #: the form date parsing.
+    form_data_parser_class: type[FormDataParser] = FormDataParser
+
+    #: The WSGI environment containing HTTP headers and information from
+    #: the WSGI server.
+    environ: WSGIEnvironment
+
+    #: Set when creating the request object. If ``True``, reading from
+    #: the request body will cause a ``RuntimeException``. Useful to
+    #: prevent modifying the stream from middleware.
+    shallow: bool
+
+    def __init__(
+        self,
+        environ: WSGIEnvironment,
+        populate_request: bool = True,
+        shallow: bool = False,
+    ) -> None:
+        super().__init__(
+            method=environ.get("REQUEST_METHOD", "GET"),
+            scheme=environ.get("wsgi.url_scheme", "http"),
+            server=_get_server(environ),
+            root_path=_wsgi_decoding_dance(environ.get("SCRIPT_NAME") or ""),
+            path=_wsgi_decoding_dance(environ.get("PATH_INFO") or ""),
+            query_string=environ.get("QUERY_STRING", "").encode("latin1"),
+            headers=EnvironHeaders(environ),
+            remote_addr=environ.get("REMOTE_ADDR"),
+        )
+        self.environ = environ
+        self.shallow = shallow
+
+        if populate_request and not shallow:
+            self.environ["werkzeug.request"] = self
+
+    @classmethod
+    def from_values(cls, *args: t.Any, **kwargs: t.Any) -> Request:
+        """Create a new request object based on the values provided.  If
+        environ is given missing values are filled from there.  This method is
+        useful for small scripts when you need to simulate a request from an URL.
+        Do not use this method for unittesting, there is a full featured client
+        object (:class:`Client`) that allows to create multipart requests,
+        support for cookies etc.
+
+        This accepts the same options as the
+        :class:`~werkzeug.test.EnvironBuilder`.
+
+        .. versionchanged:: 0.5
+           This method now accepts the same arguments as
+           :class:`~werkzeug.test.EnvironBuilder`.  Because of this the
+           `environ` parameter is now called `environ_overrides`.
+
+        :return: request object
+        """
+        from ..test import EnvironBuilder
+
+        builder = EnvironBuilder(*args, **kwargs)
+        try:
+            return builder.get_request(cls)
+        finally:
+            builder.close()
+
+    @classmethod
+    def application(cls, f: t.Callable[[Request], WSGIApplication]) -> WSGIApplication:
+        """Decorate a function as responder that accepts the request as
+        the last argument.  This works like the :func:`responder`
+        decorator but the function is passed the request object as the
+        last argument and the request object will be closed
+        automatically::
+
+            @Request.application
+            def my_wsgi_app(request):
+                return Response('Hello World!')
+
+        As of Werkzeug 0.14 HTTP exceptions are automatically caught and
+        converted to responses instead of failing.
+
+        :param f: the WSGI callable to decorate
+        :return: a new WSGI callable
+        """
+        #: return a callable that wraps the -2nd argument with the request
+        #: and calls the function with all the arguments up to that one and
+        #: the request.  The return value is then called with the latest
+        #: two arguments.  This makes it possible to use this decorator for
+        #: both standalone WSGI functions as well as bound methods and
+        #: partially applied functions.
+        from ..exceptions import HTTPException
+
+        @functools.wraps(f)
+        def application(*args: t.Any) -> cabc.Iterable[bytes]:
+            request = cls(args[-2])
+            with request:
+                try:
+                    resp = f(*args[:-2] + (request,))
+                except HTTPException as e:
+                    resp = t.cast("WSGIApplication", e.get_response(args[-2]))
+                return resp(*args[-2:])
+
+        return t.cast("WSGIApplication", application)
+
+    def _get_file_stream(
+        self,
+        total_content_length: int | None,
+        content_type: str | None,
+        filename: str | None = None,
+        content_length: int | None = None,
+    ) -> t.IO[bytes]:
+        """Called to get a stream for the file upload.
+
+        This must provide a file-like class with `read()`, `readline()`
+        and `seek()` methods that is both writeable and readable.
+
+        The default implementation returns a temporary file if the total
+        content length is higher than 500KB.  Because many browsers do not
+        provide a content length for the files only the total content
+        length matters.
+
+        :param total_content_length: the total content length of all the
+                                     data in the request combined.  This value
+                                     is guaranteed to be there.
+        :param content_type: the mimetype of the uploaded file.
+        :param filename: the filename of the uploaded file.  May be `None`.
+        :param content_length: the length of this file.  This value is usually
+                               not provided because webbrowsers do not provide
+                               this value.
+        """
+        return default_stream_factory(
+            total_content_length=total_content_length,
+            filename=filename,
+            content_type=content_type,
+            content_length=content_length,
+        )
+
+    @property
+    def want_form_data_parsed(self) -> bool:
+        """``True`` if the request method carries content. By default
+        this is true if a ``Content-Type`` is sent.
+
+        .. versionadded:: 0.8
+        """
+        return bool(self.environ.get("CONTENT_TYPE"))
+
+    def make_form_data_parser(self) -> FormDataParser:
+        """Creates the form data parser. Instantiates the
+        :attr:`form_data_parser_class` with some parameters.
+
+        .. versionadded:: 0.8
+        """
+        return self.form_data_parser_class(
+            stream_factory=self._get_file_stream,
+            max_form_memory_size=self.max_form_memory_size,
+            max_content_length=self.max_content_length,
+            max_form_parts=self.max_form_parts,
+            cls=self.parameter_storage_class,
+        )
+
+    def _load_form_data(self) -> None:
+        """Method used internally to retrieve submitted data.  After calling
+        this sets `form` and `files` on the request object to multi dicts
+        filled with the incoming form data.  As a matter of fact the input
+        stream will be empty afterwards.  You can also call this method to
+        force the parsing of the form data.
+
+        .. versionadded:: 0.8
+        """
+        # abort early if we have already consumed the stream
+        if "form" in self.__dict__:
+            return
+
+        if self.want_form_data_parsed:
+            parser = self.make_form_data_parser()
+            data = parser.parse(
+                self._get_stream_for_parsing(),
+                self.mimetype,
+                self.content_length,
+                self.mimetype_params,
+            )
+        else:
+            data = (
+                self.stream,
+                self.parameter_storage_class(),
+                self.parameter_storage_class(),
+            )
+
+        # inject the values into the instance dict so that we bypass
+        # our cached_property non-data descriptor.
+        d = self.__dict__
+        d["stream"], d["form"], d["files"] = data
+
+    def _get_stream_for_parsing(self) -> t.IO[bytes]:
+        """This is the same as accessing :attr:`stream` with the difference
+        that if it finds cached data from calling :meth:`get_data` first it
+        will create a new stream out of the cached data.
+
+        .. versionadded:: 0.9.3
+        """
+        cached_data = getattr(self, "_cached_data", None)
+        if cached_data is not None:
+            return BytesIO(cached_data)
+        return self.stream
+
+    def close(self) -> None:
+        """Closes associated resources of this request object.  This
+        closes all file handles explicitly.  You can also use the request
+        object in a with statement which will automatically close it.
+
+        .. versionadded:: 0.9
+        """
+        files = self.__dict__.get("files")
+        for _key, value in iter_multi_items(files or ()):
+            value.close()
+
+    def __enter__(self) -> Request:
+        return self
+
+    def __exit__(self, exc_type, exc_value, tb) -> None:  # type: ignore
+        self.close()
+
+    @cached_property
+    def stream(self) -> t.IO[bytes]:
+        """The WSGI input stream, with safety checks. This stream can only be consumed
+        once.
+
+        Use :meth:`get_data` to get the full data as bytes or text. The :attr:`data`
+        attribute will contain the full bytes only if they do not represent form data.
+        The :attr:`form` attribute will contain the parsed form data in that case.
+
+        Unlike :attr:`input_stream`, this stream guards against infinite streams or
+        reading past :attr:`content_length` or :attr:`max_content_length`.
+
+        If ``max_content_length`` is set, it can be enforced on streams if
+        ``wsgi.input_terminated`` is set. Otherwise, an empty stream is returned.
+
+        If the limit is reached before the underlying stream is exhausted (such as a
+        file that is too large, or an infinite stream), the remaining contents of the
+        stream cannot be read safely. Depending on how the server handles this, clients
+        may show a "connection reset" failure instead of seeing the 413 response.
+
+        .. versionchanged:: 2.3
+            Check ``max_content_length`` preemptively and while reading.
+
+        .. versionchanged:: 0.9
+            The stream is always set (but may be consumed) even if form parsing was
+            accessed first.
+        """
+        if self.shallow:
+            raise RuntimeError(
+                "This request was created with 'shallow=True', reading"
+                " from the input stream is disabled."
+            )
+
+        return get_input_stream(
+            self.environ, max_content_length=self.max_content_length
+        )
+
+    input_stream = environ_property[t.IO[bytes]](
+        "wsgi.input",
+        doc="""The raw WSGI input stream, without any safety checks.
+
+        This is dangerous to use. It does not guard against infinite streams or reading
+        past :attr:`content_length` or :attr:`max_content_length`.
+
+        Use :attr:`stream` instead.
+        """,
+    )
+
+    @cached_property
+    def data(self) -> bytes:
+        """The raw data read from :attr:`stream`. Will be empty if the request
+        represents form data.
+
+        To get the raw data even if it represents form data, use :meth:`get_data`.
+        """
+        return self.get_data(parse_form_data=True)
+
+    @t.overload
+    def get_data(
+        self,
+        cache: bool = True,
+        as_text: t.Literal[False] = False,
+        parse_form_data: bool = False,
+    ) -> bytes: ...
+
+    @t.overload
+    def get_data(
+        self,
+        cache: bool = True,
+        as_text: t.Literal[True] = ...,
+        parse_form_data: bool = False,
+    ) -> str: ...
+
+    def get_data(
+        self, cache: bool = True, as_text: bool = False, parse_form_data: bool = False
+    ) -> bytes | str:
+        """This reads the buffered incoming data from the client into one
+        bytes object.  By default this is cached but that behavior can be
+        changed by setting `cache` to `False`.
+
+        Usually it's a bad idea to call this method without checking the
+        content length first as a client could send dozens of megabytes or more
+        to cause memory problems on the server.
+
+        Note that if the form data was already parsed this method will not
+        return anything as form data parsing does not cache the data like
+        this method does.  To implicitly invoke form data parsing function
+        set `parse_form_data` to `True`.  When this is done the return value
+        of this method will be an empty string if the form parser handles
+        the data.  This generally is not necessary as if the whole data is
+        cached (which is the default) the form parser will used the cached
+        data to parse the form data.  Please be generally aware of checking
+        the content length first in any case before calling this method
+        to avoid exhausting server memory.
+
+        If `as_text` is set to `True` the return value will be a decoded
+        string.
+
+        .. versionadded:: 0.9
+        """
+        rv = getattr(self, "_cached_data", None)
+        if rv is None:
+            if parse_form_data:
+                self._load_form_data()
+            rv = self.stream.read()
+            if cache:
+                self._cached_data = rv
+        if as_text:
+            rv = rv.decode(errors="replace")
+        return rv
+
+    @cached_property
+    def form(self) -> ImmutableMultiDict[str, str]:
+        """The form parameters.  By default an
+        :class:`~werkzeug.datastructures.ImmutableMultiDict`
+        is returned from this function.  This can be changed by setting
+        :attr:`parameter_storage_class` to a different type.  This might
+        be necessary if the order of the form data is important.
+
+        Please keep in mind that file uploads will not end up here, but instead
+        in the :attr:`files` attribute.
+
+        .. versionchanged:: 0.9
+
+            Previous to Werkzeug 0.9 this would only contain form data for POST
+            and PUT requests.
+        """
+        self._load_form_data()
+        return self.form
+
+    @cached_property
+    def values(self) -> CombinedMultiDict[str, str]:
+        """A :class:`werkzeug.datastructures.CombinedMultiDict` that
+        combines :attr:`args` and :attr:`form`.
+
+        For GET requests, only ``args`` are present, not ``form``.
+
+        .. versionchanged:: 2.0
+            For GET requests, only ``args`` are present, not ``form``.
+        """
+        sources = [self.args]
+
+        if self.method != "GET":
+            # GET requests can have a body, and some caching proxies
+            # might not treat that differently than a normal GET
+            # request, allowing form data to "invisibly" affect the
+            # cache without indication in the query string / URL.
+            sources.append(self.form)
+
+        args = []
+
+        for d in sources:
+            if not isinstance(d, MultiDict):
+                d = MultiDict(d)
+
+            args.append(d)
+
+        return CombinedMultiDict(args)
+
+    @cached_property
+    def files(self) -> ImmutableMultiDict[str, FileStorage]:
+        """:class:`~werkzeug.datastructures.MultiDict` object containing
+        all uploaded files.  Each key in :attr:`files` is the name from the
+        ``<input type="file" name="">``.  Each value in :attr:`files` is a
+        Werkzeug :class:`~werkzeug.datastructures.FileStorage` object.
+
+        It basically behaves like a standard file object you know from Python,
+        with the difference that it also has a
+        :meth:`~werkzeug.datastructures.FileStorage.save` function that can
+        store the file on the filesystem.
+
+        Note that :attr:`files` will only contain data if the request method was
+        POST, PUT or PATCH and the ``<form>`` that posted to the request had
+        ``enctype="multipart/form-data"``.  It will be empty otherwise.
+
+        See the :class:`~werkzeug.datastructures.MultiDict` /
+        :class:`~werkzeug.datastructures.FileStorage` documentation for
+        more details about the used data structure.
+        """
+        self._load_form_data()
+        return self.files
+
+    @property
+    def script_root(self) -> str:
+        """Alias for :attr:`self.root_path`. ``environ["SCRIPT_ROOT"]``
+        without a trailing slash.
+        """
+        return self.root_path
+
+    @cached_property
+    def url_root(self) -> str:
+        """Alias for :attr:`root_url`. The URL with scheme, host, and
+        root path. For example, ``https://example.com/app/``.
+        """
+        return self.root_url
+
+    remote_user = environ_property[str](
+        "REMOTE_USER",
+        doc="""If the server supports user authentication, and the
+        script is protected, this attribute contains the username the
+        user has authenticated as.""",
+    )
+    is_multithread = environ_property[bool](
+        "wsgi.multithread",
+        doc="""boolean that is `True` if the application is served by a
+        multithreaded WSGI server.""",
+    )
+    is_multiprocess = environ_property[bool](
+        "wsgi.multiprocess",
+        doc="""boolean that is `True` if the application is served by a
+        WSGI server that spawns multiple processes.""",
+    )
+    is_run_once = environ_property[bool](
+        "wsgi.run_once",
+        doc="""boolean that is `True` if the application will be
+        executed only once in a process lifetime.  This is the case for
+        CGI for example, but it's not guaranteed that the execution only
+        happens one time.""",
+    )
+
+    # JSON
+
+    #: A module or other object that has ``dumps`` and ``loads``
+    #: functions that match the API of the built-in :mod:`json` module.
+    json_module = json
+
+    @property
+    def json(self) -> t.Any | None:
+        """The parsed JSON data if :attr:`mimetype` indicates JSON
+        (:mimetype:`application/json`, see :attr:`is_json`).
+
+        Calls :meth:`get_json` with default arguments.
+
+        If the request content type is not ``application/json``, this
+        will raise a 415 Unsupported Media Type error.
+
+        .. versionchanged:: 2.3
+            Raise a 415 error instead of 400.
+
+        .. versionchanged:: 2.1
+            Raise a 400 error if the content type is incorrect.
+        """
+        return self.get_json()
+
+    # Cached values for ``(silent=False, silent=True)``. Initialized
+    # with sentinel values.
+    _cached_json: tuple[t.Any, t.Any] = (Ellipsis, Ellipsis)
+
+    @t.overload
+    def get_json(
+        self, force: bool = ..., silent: t.Literal[False] = ..., cache: bool = ...
+    ) -> t.Any: ...
+
+    @t.overload
+    def get_json(
+        self, force: bool = ..., silent: bool = ..., cache: bool = ...
+    ) -> t.Any | None: ...
+
+    def get_json(
+        self, force: bool = False, silent: bool = False, cache: bool = True
+    ) -> t.Any | None:
+        """Parse :attr:`data` as JSON.
+
+        If the mimetype does not indicate JSON
+        (:mimetype:`application/json`, see :attr:`is_json`), or parsing
+        fails, :meth:`on_json_loading_failed` is called and
+        its return value is used as the return value. By default this
+        raises a 415 Unsupported Media Type resp.
+
+        :param force: Ignore the mimetype and always try to parse JSON.
+        :param silent: Silence mimetype and parsing errors, and
+            return ``None`` instead.
+        :param cache: Store the parsed JSON to return for subsequent
+            calls.
+
+        .. versionchanged:: 2.3
+            Raise a 415 error instead of 400.
+
+        .. versionchanged:: 2.1
+            Raise a 400 error if the content type is incorrect.
+        """
+        if cache and self._cached_json[silent] is not Ellipsis:
+            return self._cached_json[silent]
+
+        if not (force or self.is_json):
+            if not silent:
+                return self.on_json_loading_failed(None)
+            else:
+                return None
+
+        data = self.get_data(cache=cache)
+
+        try:
+            rv = self.json_module.loads(data)
+        except ValueError as e:
+            if silent:
+                rv = None
+
+                if cache:
+                    normal_rv, _ = self._cached_json
+                    self._cached_json = (normal_rv, rv)
+            else:
+                rv = self.on_json_loading_failed(e)
+
+                if cache:
+                    _, silent_rv = self._cached_json
+                    self._cached_json = (rv, silent_rv)
+        else:
+            if cache:
+                self._cached_json = (rv, rv)
+
+        return rv
+
+    def on_json_loading_failed(self, e: ValueError | None) -> t.Any:
+        """Called if :meth:`get_json` fails and isn't silenced.
+
+        If this method returns a value, it is used as the return value
+        for :meth:`get_json`. The default implementation raises
+        :exc:`~werkzeug.exceptions.BadRequest`.
+
+        :param e: If parsing failed, this is the exception. It will be
+            ``None`` if the content type wasn't ``application/json``.
+
+        .. versionchanged:: 2.3
+            Raise a 415 error instead of 400.
+        """
+        if e is not None:
+            raise BadRequest(f"Failed to decode JSON object: {e}")
+
+        raise UnsupportedMediaType(
+            "Did not attempt to load JSON data because the request"
+            " Content-Type was not 'application/json'."
+        )
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/response.py b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/response.py
new file mode 100644
index 00000000..7f01287c
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/wrappers/response.py
@@ -0,0 +1,831 @@
+from __future__ import annotations
+
+import json
+import typing as t
+from http import HTTPStatus
+from urllib.parse import urljoin
+
+from .._internal import _get_environ
+from ..datastructures import Headers
+from ..http import generate_etag
+from ..http import http_date
+from ..http import is_resource_modified
+from ..http import parse_etags
+from ..http import parse_range_header
+from ..http import remove_entity_headers
+from ..sansio.response import Response as _SansIOResponse
+from ..urls import iri_to_uri
+from ..utils import cached_property
+from ..wsgi import _RangeWrapper
+from ..wsgi import ClosingIterator
+from ..wsgi import get_current_url
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import StartResponse
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+    from .request import Request
+
+
+def _iter_encoded(iterable: t.Iterable[str | bytes]) -> t.Iterator[bytes]:
+    for item in iterable:
+        if isinstance(item, str):
+            yield item.encode()
+        else:
+            yield item
+
+
+class Response(_SansIOResponse):
+    """Represents an outgoing WSGI HTTP response with body, status, and
+    headers. Has properties and methods for using the functionality
+    defined by various HTTP specs.
+
+    The response body is flexible to support different use cases. The
+    simple form is passing bytes, or a string which will be encoded as
+    UTF-8. Passing an iterable of bytes or strings makes this a
+    streaming response. A generator is particularly useful for building
+    a CSV file in memory or using SSE (Server Sent Events). A file-like
+    object is also iterable, although the
+    :func:`~werkzeug.utils.send_file` helper should be used in that
+    case.
+
+    The response object is itself a WSGI application callable. When
+    called (:meth:`__call__`) with ``environ`` and ``start_response``,
+    it will pass its status and headers to ``start_response`` then
+    return its body as an iterable.
+
+    .. code-block:: python
+
+        from werkzeug.wrappers.response import Response
+
+        def index():
+            return Response("Hello, World!")
+
+        def application(environ, start_response):
+            path = environ.get("PATH_INFO") or "/"
+
+            if path == "/":
+                response = index()
+            else:
+                response = Response("Not Found", status=404)
+
+            return response(environ, start_response)
+
+    :param response: The data for the body of the response. A string or
+        bytes, or tuple or list of strings or bytes, for a fixed-length
+        response, or any other iterable of strings or bytes for a
+        streaming response. Defaults to an empty body.
+    :param status: The status code for the response. Either an int, in
+        which case the default status message is added, or a string in
+        the form ``{code} {message}``, like ``404 Not Found``. Defaults
+        to 200.
+    :param headers: A :class:`~werkzeug.datastructures.Headers` object,
+        or a list of ``(key, value)`` tuples that will be converted to a
+        ``Headers`` object.
+    :param mimetype: The mime type (content type without charset or
+        other parameters) of the response. If the value starts with
+        ``text/`` (or matches some other special cases), the charset
+        will be added to create the ``content_type``.
+    :param content_type: The full content type of the response.
+        Overrides building the value from ``mimetype``.
+    :param direct_passthrough: Pass the response body directly through
+        as the WSGI iterable. This can be used when the body is a binary
+        file or other iterator of bytes, to skip some unnecessary
+        checks. Use :func:`~werkzeug.utils.send_file` instead of setting
+        this manually.
+
+    .. versionchanged:: 2.1
+        Old ``BaseResponse`` and mixin classes were removed.
+
+    .. versionchanged:: 2.0
+        Combine ``BaseResponse`` and mixins into a single ``Response``
+        class.
+
+    .. versionchanged:: 0.5
+        The ``direct_passthrough`` parameter was added.
+    """
+
+    #: if set to `False` accessing properties on the response object will
+    #: not try to consume the response iterator and convert it into a list.
+    #:
+    #: .. versionadded:: 0.6.2
+    #:
+    #:    That attribute was previously called `implicit_seqence_conversion`.
+    #:    (Notice the typo).  If you did use this feature, you have to adapt
+    #:    your code to the name change.
+    implicit_sequence_conversion = True
+
+    #: If a redirect ``Location`` header is a relative URL, make it an
+    #: absolute URL, including scheme and domain.
+    #:
+    #: .. versionchanged:: 2.1
+    #:     This is disabled by default, so responses will send relative
+    #:     redirects.
+    #:
+    #: .. versionadded:: 0.8
+    autocorrect_location_header = False
+
+    #: Should this response object automatically set the content-length
+    #: header if possible?  This is true by default.
+    #:
+    #: .. versionadded:: 0.8
+    automatically_set_content_length = True
+
+    #: The response body to send as the WSGI iterable. A list of strings
+    #: or bytes represents a fixed-length response, any other iterable
+    #: is a streaming response. Strings are encoded to bytes as UTF-8.
+    #:
+    #: Do not set to a plain string or bytes, that will cause sending
+    #: the response to be very inefficient as it will iterate one byte
+    #: at a time.
+    response: t.Iterable[str] | t.Iterable[bytes]
+
+    def __init__(
+        self,
+        response: t.Iterable[bytes] | bytes | t.Iterable[str] | str | None = None,
+        status: int | str | HTTPStatus | None = None,
+        headers: t.Mapping[str, str | t.Iterable[str]]
+        | t.Iterable[tuple[str, str]]
+        | None = None,
+        mimetype: str | None = None,
+        content_type: str | None = None,
+        direct_passthrough: bool = False,
+    ) -> None:
+        super().__init__(
+            status=status,
+            headers=headers,
+            mimetype=mimetype,
+            content_type=content_type,
+        )
+
+        #: Pass the response body directly through as the WSGI iterable.
+        #: This can be used when the body is a binary file or other
+        #: iterator of bytes, to skip some unnecessary checks. Use
+        #: :func:`~werkzeug.utils.send_file` instead of setting this
+        #: manually.
+        self.direct_passthrough = direct_passthrough
+        self._on_close: list[t.Callable[[], t.Any]] = []
+
+        # we set the response after the headers so that if a class changes
+        # the charset attribute, the data is set in the correct charset.
+        if response is None:
+            self.response = []
+        elif isinstance(response, (str, bytes, bytearray)):
+            self.set_data(response)
+        else:
+            self.response = response
+
+    def call_on_close(self, func: t.Callable[[], t.Any]) -> t.Callable[[], t.Any]:
+        """Adds a function to the internal list of functions that should
+        be called as part of closing down the response.  Since 0.7 this
+        function also returns the function that was passed so that this
+        can be used as a decorator.
+
+        .. versionadded:: 0.6
+        """
+        self._on_close.append(func)
+        return func
+
+    def __repr__(self) -> str:
+        if self.is_sequence:
+            body_info = f"{sum(map(len, self.iter_encoded()))} bytes"
+        else:
+            body_info = "streamed" if self.is_streamed else "likely-streamed"
+        return f"<{type(self).__name__} {body_info} [{self.status}]>"
+
+    @classmethod
+    def force_type(
+        cls, response: Response, environ: WSGIEnvironment | None = None
+    ) -> Response:
+        """Enforce that the WSGI response is a response object of the current
+        type.  Werkzeug will use the :class:`Response` internally in many
+        situations like the exceptions.  If you call :meth:`get_response` on an
+        exception you will get back a regular :class:`Response` object, even
+        if you are using a custom subclass.
+
+        This method can enforce a given response type, and it will also
+        convert arbitrary WSGI callables into response objects if an environ
+        is provided::
+
+            # convert a Werkzeug response object into an instance of the
+            # MyResponseClass subclass.
+            response = MyResponseClass.force_type(response)
+
+            # convert any WSGI application into a response object
+            response = MyResponseClass.force_type(response, environ)
+
+        This is especially useful if you want to post-process responses in
+        the main dispatcher and use functionality provided by your subclass.
+
+        Keep in mind that this will modify response objects in place if
+        possible!
+
+        :param response: a response object or wsgi application.
+        :param environ: a WSGI environment object.
+        :return: a response object.
+        """
+        if not isinstance(response, Response):
+            if environ is None:
+                raise TypeError(
+                    "cannot convert WSGI application into response"
+                    " objects without an environ"
+                )
+
+            from ..test import run_wsgi_app
+
+            response = Response(*run_wsgi_app(response, environ))
+
+        response.__class__ = cls
+        return response
+
+    @classmethod
+    def from_app(
+        cls, app: WSGIApplication, environ: WSGIEnvironment, buffered: bool = False
+    ) -> Response:
+        """Create a new response object from an application output.  This
+        works best if you pass it an application that returns a generator all
+        the time.  Sometimes applications may use the `write()` callable
+        returned by the `start_response` function.  This tries to resolve such
+        edge cases automatically.  But if you don't get the expected output
+        you should set `buffered` to `True` which enforces buffering.
+
+        :param app: the WSGI application to execute.
+        :param environ: the WSGI environment to execute against.
+        :param buffered: set to `True` to enforce buffering.
+        :return: a response object.
+        """
+        from ..test import run_wsgi_app
+
+        return cls(*run_wsgi_app(app, environ, buffered))
+
+    @t.overload
+    def get_data(self, as_text: t.Literal[False] = False) -> bytes: ...
+
+    @t.overload
+    def get_data(self, as_text: t.Literal[True]) -> str: ...
+
+    def get_data(self, as_text: bool = False) -> bytes | str:
+        """The string representation of the response body.  Whenever you call
+        this property the response iterable is encoded and flattened.  This
+        can lead to unwanted behavior if you stream big data.
+
+        This behavior can be disabled by setting
+        :attr:`implicit_sequence_conversion` to `False`.
+
+        If `as_text` is set to `True` the return value will be a decoded
+        string.
+
+        .. versionadded:: 0.9
+        """
+        self._ensure_sequence()
+        rv = b"".join(self.iter_encoded())
+
+        if as_text:
+            return rv.decode()
+
+        return rv
+
+    def set_data(self, value: bytes | str) -> None:
+        """Sets a new string as response.  The value must be a string or
+        bytes. If a string is set it's encoded to the charset of the
+        response (utf-8 by default).
+
+        .. versionadded:: 0.9
+        """
+        if isinstance(value, str):
+            value = value.encode()
+        self.response = [value]
+        if self.automatically_set_content_length:
+            self.headers["Content-Length"] = str(len(value))
+
+    data = property(
+        get_data,
+        set_data,
+        doc="A descriptor that calls :meth:`get_data` and :meth:`set_data`.",
+    )
+
+    def calculate_content_length(self) -> int | None:
+        """Returns the content length if available or `None` otherwise."""
+        try:
+            self._ensure_sequence()
+        except RuntimeError:
+            return None
+        return sum(len(x) for x in self.iter_encoded())
+
+    def _ensure_sequence(self, mutable: bool = False) -> None:
+        """This method can be called by methods that need a sequence.  If
+        `mutable` is true, it will also ensure that the response sequence
+        is a standard Python list.
+
+        .. versionadded:: 0.6
+        """
+        if self.is_sequence:
+            # if we need a mutable object, we ensure it's a list.
+            if mutable and not isinstance(self.response, list):
+                self.response = list(self.response)  # type: ignore
+            return
+        if self.direct_passthrough:
+            raise RuntimeError(
+                "Attempted implicit sequence conversion but the"
+                " response object is in direct passthrough mode."
+            )
+        if not self.implicit_sequence_conversion:
+            raise RuntimeError(
+                "The response object required the iterable to be a"
+                " sequence, but the implicit conversion was disabled."
+                " Call make_sequence() yourself."
+            )
+        self.make_sequence()
+
+    def make_sequence(self) -> None:
+        """Converts the response iterator in a list.  By default this happens
+        automatically if required.  If `implicit_sequence_conversion` is
+        disabled, this method is not automatically called and some properties
+        might raise exceptions.  This also encodes all the items.
+
+        .. versionadded:: 0.6
+        """
+        if not self.is_sequence:
+            # if we consume an iterable we have to ensure that the close
+            # method of the iterable is called if available when we tear
+            # down the response
+            close = getattr(self.response, "close", None)
+            self.response = list(self.iter_encoded())
+            if close is not None:
+                self.call_on_close(close)
+
+    def iter_encoded(self) -> t.Iterator[bytes]:
+        """Iter the response encoded with the encoding of the response.
+        If the response object is invoked as WSGI application the return
+        value of this method is used as application iterator unless
+        :attr:`direct_passthrough` was activated.
+        """
+        # Encode in a separate function so that self.response is fetched
+        # early.  This allows us to wrap the response with the return
+        # value from get_app_iter or iter_encoded.
+        return _iter_encoded(self.response)
+
+    @property
+    def is_streamed(self) -> bool:
+        """If the response is streamed (the response is not an iterable with
+        a length information) this property is `True`.  In this case streamed
+        means that there is no information about the number of iterations.
+        This is usually `True` if a generator is passed to the response object.
+
+        This is useful for checking before applying some sort of post
+        filtering that should not take place for streamed responses.
+        """
+        try:
+            len(self.response)  # type: ignore
+        except (TypeError, AttributeError):
+            return True
+        return False
+
+    @property
+    def is_sequence(self) -> bool:
+        """If the iterator is buffered, this property will be `True`.  A
+        response object will consider an iterator to be buffered if the
+        response attribute is a list or tuple.
+
+        .. versionadded:: 0.6
+        """
+        return isinstance(self.response, (tuple, list))
+
+    def close(self) -> None:
+        """Close the wrapped response if possible.  You can also use the object
+        in a with statement which will automatically close it.
+
+        .. versionadded:: 0.9
+           Can now be used in a with statement.
+        """
+        if hasattr(self.response, "close"):
+            self.response.close()
+        for func in self._on_close:
+            func()
+
+    def __enter__(self) -> Response:
+        return self
+
+    def __exit__(self, exc_type, exc_value, tb):  # type: ignore
+        self.close()
+
+    def freeze(self) -> None:
+        """Make the response object ready to be pickled. Does the
+        following:
+
+        *   Buffer the response into a list, ignoring
+            :attr:`implicity_sequence_conversion` and
+            :attr:`direct_passthrough`.
+        *   Set the ``Content-Length`` header.
+        *   Generate an ``ETag`` header if one is not already set.
+
+        .. versionchanged:: 2.1
+            Removed the ``no_etag`` parameter.
+
+        .. versionchanged:: 2.0
+            An ``ETag`` header is always added.
+
+        .. versionchanged:: 0.6
+            The ``Content-Length`` header is set.
+        """
+        # Always freeze the encoded response body, ignore
+        # implicit_sequence_conversion and direct_passthrough.
+        self.response = list(self.iter_encoded())
+        self.headers["Content-Length"] = str(sum(map(len, self.response)))
+        self.add_etag()
+
+    def get_wsgi_headers(self, environ: WSGIEnvironment) -> Headers:
+        """This is automatically called right before the response is started
+        and returns headers modified for the given environment.  It returns a
+        copy of the headers from the response with some modifications applied
+        if necessary.
+
+        For example the location header (if present) is joined with the root
+        URL of the environment.  Also the content length is automatically set
+        to zero here for certain status codes.
+
+        .. versionchanged:: 0.6
+           Previously that function was called `fix_headers` and modified
+           the response object in place.  Also since 0.6, IRIs in location
+           and content-location headers are handled properly.
+
+           Also starting with 0.6, Werkzeug will attempt to set the content
+           length if it is able to figure it out on its own.  This is the
+           case if all the strings in the response iterable are already
+           encoded and the iterable is buffered.
+
+        :param environ: the WSGI environment of the request.
+        :return: returns a new :class:`~werkzeug.datastructures.Headers`
+                 object.
+        """
+        headers = Headers(self.headers)
+        location: str | None = None
+        content_location: str | None = None
+        content_length: str | int | None = None
+        status = self.status_code
+
+        # iterate over the headers to find all values in one go.  Because
+        # get_wsgi_headers is used each response that gives us a tiny
+        # speedup.
+        for key, value in headers:
+            ikey = key.lower()
+            if ikey == "location":
+                location = value
+            elif ikey == "content-location":
+                content_location = value
+            elif ikey == "content-length":
+                content_length = value
+
+        if location is not None:
+            location = iri_to_uri(location)
+
+            if self.autocorrect_location_header:
+                # Make the location header an absolute URL.
+                current_url = get_current_url(environ, strip_querystring=True)
+                current_url = iri_to_uri(current_url)
+                location = urljoin(current_url, location)
+
+            headers["Location"] = location
+
+        # make sure the content location is a URL
+        if content_location is not None:
+            headers["Content-Location"] = iri_to_uri(content_location)
+
+        if 100 <= status < 200 or status == 204:
+            # Per section 3.3.2 of RFC 7230, "a server MUST NOT send a
+            # Content-Length header field in any response with a status
+            # code of 1xx (Informational) or 204 (No Content)."
+            headers.remove("Content-Length")
+        elif status == 304:
+            remove_entity_headers(headers)
+
+        # if we can determine the content length automatically, we
+        # should try to do that.  But only if this does not involve
+        # flattening the iterator or encoding of strings in the
+        # response. We however should not do that if we have a 304
+        # response.
+        if (
+            self.automatically_set_content_length
+            and self.is_sequence
+            and content_length is None
+            and status not in (204, 304)
+            and not (100 <= status < 200)
+        ):
+            content_length = sum(len(x) for x in self.iter_encoded())
+            headers["Content-Length"] = str(content_length)
+
+        return headers
+
+    def get_app_iter(self, environ: WSGIEnvironment) -> t.Iterable[bytes]:
+        """Returns the application iterator for the given environ.  Depending
+        on the request method and the current status code the return value
+        might be an empty response rather than the one from the response.
+
+        If the request method is `HEAD` or the status code is in a range
+        where the HTTP specification requires an empty response, an empty
+        iterable is returned.
+
+        .. versionadded:: 0.6
+
+        :param environ: the WSGI environment of the request.
+        :return: a response iterable.
+        """
+        status = self.status_code
+        if (
+            environ["REQUEST_METHOD"] == "HEAD"
+            or 100 <= status < 200
+            or status in (204, 304)
+        ):
+            iterable: t.Iterable[bytes] = ()
+        elif self.direct_passthrough:
+            return self.response  # type: ignore
+        else:
+            iterable = self.iter_encoded()
+        return ClosingIterator(iterable, self.close)
+
+    def get_wsgi_response(
+        self, environ: WSGIEnvironment
+    ) -> tuple[t.Iterable[bytes], str, list[tuple[str, str]]]:
+        """Returns the final WSGI response as tuple.  The first item in
+        the tuple is the application iterator, the second the status and
+        the third the list of headers.  The response returned is created
+        specially for the given environment.  For example if the request
+        method in the WSGI environment is ``'HEAD'`` the response will
+        be empty and only the headers and status code will be present.
+
+        .. versionadded:: 0.6
+
+        :param environ: the WSGI environment of the request.
+        :return: an ``(app_iter, status, headers)`` tuple.
+        """
+        headers = self.get_wsgi_headers(environ)
+        app_iter = self.get_app_iter(environ)
+        return app_iter, self.status, headers.to_wsgi_list()
+
+    def __call__(
+        self, environ: WSGIEnvironment, start_response: StartResponse
+    ) -> t.Iterable[bytes]:
+        """Process this response as WSGI application.
+
+        :param environ: the WSGI environment.
+        :param start_response: the response callable provided by the WSGI
+                               server.
+        :return: an application iterator
+        """
+        app_iter, status, headers = self.get_wsgi_response(environ)
+        start_response(status, headers)
+        return app_iter
+
+    # JSON
+
+    #: A module or other object that has ``dumps`` and ``loads``
+    #: functions that match the API of the built-in :mod:`json` module.
+    json_module = json
+
+    @property
+    def json(self) -> t.Any | None:
+        """The parsed JSON data if :attr:`mimetype` indicates JSON
+        (:mimetype:`application/json`, see :attr:`is_json`).
+
+        Calls :meth:`get_json` with default arguments.
+        """
+        return self.get_json()
+
+    @t.overload
+    def get_json(self, force: bool = ..., silent: t.Literal[False] = ...) -> t.Any: ...
+
+    @t.overload
+    def get_json(self, force: bool = ..., silent: bool = ...) -> t.Any | None: ...
+
+    def get_json(self, force: bool = False, silent: bool = False) -> t.Any | None:
+        """Parse :attr:`data` as JSON. Useful during testing.
+
+        If the mimetype does not indicate JSON
+        (:mimetype:`application/json`, see :attr:`is_json`), this
+        returns ``None``.
+
+        Unlike :meth:`Request.get_json`, the result is not cached.
+
+        :param force: Ignore the mimetype and always try to parse JSON.
+        :param silent: Silence parsing errors and return ``None``
+            instead.
+        """
+        if not (force or self.is_json):
+            return None
+
+        data = self.get_data()
+
+        try:
+            return self.json_module.loads(data)
+        except ValueError:
+            if not silent:
+                raise
+
+            return None
+
+    # Stream
+
+    @cached_property
+    def stream(self) -> ResponseStream:
+        """The response iterable as write-only stream."""
+        return ResponseStream(self)
+
+    def _wrap_range_response(self, start: int, length: int) -> None:
+        """Wrap existing Response in case of Range Request context."""
+        if self.status_code == 206:
+            self.response = _RangeWrapper(self.response, start, length)  # type: ignore
+
+    def _is_range_request_processable(self, environ: WSGIEnvironment) -> bool:
+        """Return ``True`` if `Range` header is present and if underlying
+        resource is considered unchanged when compared with `If-Range` header.
+        """
+        return (
+            "HTTP_IF_RANGE" not in environ
+            or not is_resource_modified(
+                environ,
+                self.headers.get("etag"),
+                None,
+                self.headers.get("last-modified"),
+                ignore_if_range=False,
+            )
+        ) and "HTTP_RANGE" in environ
+
+    def _process_range_request(
+        self,
+        environ: WSGIEnvironment,
+        complete_length: int | None,
+        accept_ranges: bool | str,
+    ) -> bool:
+        """Handle Range Request related headers (RFC7233).  If `Accept-Ranges`
+        header is valid, and Range Request is processable, we set the headers
+        as described by the RFC, and wrap the underlying response in a
+        RangeWrapper.
+
+        Returns ``True`` if Range Request can be fulfilled, ``False`` otherwise.
+
+        :raises: :class:`~werkzeug.exceptions.RequestedRangeNotSatisfiable`
+                 if `Range` header could not be parsed or satisfied.
+
+        .. versionchanged:: 2.0
+            Returns ``False`` if the length is 0.
+        """
+        from ..exceptions import RequestedRangeNotSatisfiable
+
+        if (
+            not accept_ranges
+            or complete_length is None
+            or complete_length == 0
+            or not self._is_range_request_processable(environ)
+        ):
+            return False
+
+        if accept_ranges is True:
+            accept_ranges = "bytes"
+
+        parsed_range = parse_range_header(environ.get("HTTP_RANGE"))
+
+        if parsed_range is None:
+            raise RequestedRangeNotSatisfiable(complete_length)
+
+        range_tuple = parsed_range.range_for_length(complete_length)
+        content_range_header = parsed_range.to_content_range_header(complete_length)
+
+        if range_tuple is None or content_range_header is None:
+            raise RequestedRangeNotSatisfiable(complete_length)
+
+        content_length = range_tuple[1] - range_tuple[0]
+        self.headers["Content-Length"] = str(content_length)
+        self.headers["Accept-Ranges"] = accept_ranges
+        self.content_range = content_range_header  # type: ignore
+        self.status_code = 206
+        self._wrap_range_response(range_tuple[0], content_length)
+        return True
+
+    def make_conditional(
+        self,
+        request_or_environ: WSGIEnvironment | Request,
+        accept_ranges: bool | str = False,
+        complete_length: int | None = None,
+    ) -> Response:
+        """Make the response conditional to the request.  This method works
+        best if an etag was defined for the response already.  The `add_etag`
+        method can be used to do that.  If called without etag just the date
+        header is set.
+
+        This does nothing if the request method in the request or environ is
+        anything but GET or HEAD.
+
+        For optimal performance when handling range requests, it's recommended
+        that your response data object implements `seekable`, `seek` and `tell`
+        methods as described by :py:class:`io.IOBase`.  Objects returned by
+        :meth:`~werkzeug.wsgi.wrap_file` automatically implement those methods.
+
+        It does not remove the body of the response because that's something
+        the :meth:`__call__` function does for us automatically.
+
+        Returns self so that you can do ``return resp.make_conditional(req)``
+        but modifies the object in-place.
+
+        :param request_or_environ: a request object or WSGI environment to be
+                                   used to make the response conditional
+                                   against.
+        :param accept_ranges: This parameter dictates the value of
+                              `Accept-Ranges` header. If ``False`` (default),
+                              the header is not set. If ``True``, it will be set
+                              to ``"bytes"``. If it's a string, it will use this
+                              value.
+        :param complete_length: Will be used only in valid Range Requests.
+                                It will set `Content-Range` complete length
+                                value and compute `Content-Length` real value.
+                                This parameter is mandatory for successful
+                                Range Requests completion.
+        :raises: :class:`~werkzeug.exceptions.RequestedRangeNotSatisfiable`
+                 if `Range` header could not be parsed or satisfied.
+
+        .. versionchanged:: 2.0
+            Range processing is skipped if length is 0 instead of
+            raising a 416 Range Not Satisfiable error.
+        """
+        environ = _get_environ(request_or_environ)
+        if environ["REQUEST_METHOD"] in ("GET", "HEAD"):
+            # if the date is not in the headers, add it now.  We however
+            # will not override an already existing header.  Unfortunately
+            # this header will be overridden by many WSGI servers including
+            # wsgiref.
+            if "date" not in self.headers:
+                self.headers["Date"] = http_date()
+            is206 = self._process_range_request(environ, complete_length, accept_ranges)
+            if not is206 and not is_resource_modified(
+                environ,
+                self.headers.get("etag"),
+                None,
+                self.headers.get("last-modified"),
+            ):
+                if parse_etags(environ.get("HTTP_IF_MATCH")):
+                    self.status_code = 412
+                else:
+                    self.status_code = 304
+            if (
+                self.automatically_set_content_length
+                and "content-length" not in self.headers
+            ):
+                length = self.calculate_content_length()
+                if length is not None:
+                    self.headers["Content-Length"] = str(length)
+        return self
+
+    def add_etag(self, overwrite: bool = False, weak: bool = False) -> None:
+        """Add an etag for the current response if there is none yet.
+
+        .. versionchanged:: 2.0
+            SHA-1 is used to generate the value. MD5 may not be
+            available in some environments.
+        """
+        if overwrite or "etag" not in self.headers:
+            self.set_etag(generate_etag(self.get_data()), weak)
+
+
+class ResponseStream:
+    """A file descriptor like object used by :meth:`Response.stream` to
+    represent the body of the stream. It directly pushes into the
+    response iterable of the response object.
+    """
+
+    mode = "wb+"
+
+    def __init__(self, response: Response):
+        self.response = response
+        self.closed = False
+
+    def write(self, value: bytes) -> int:
+        if self.closed:
+            raise ValueError("I/O operation on closed file")
+        self.response._ensure_sequence(mutable=True)
+        self.response.response.append(value)  # type: ignore
+        self.response.headers.pop("Content-Length", None)
+        return len(value)
+
+    def writelines(self, seq: t.Iterable[bytes]) -> None:
+        for item in seq:
+            self.write(item)
+
+    def close(self) -> None:
+        self.closed = True
+
+    def flush(self) -> None:
+        if self.closed:
+            raise ValueError("I/O operation on closed file")
+
+    def isatty(self) -> bool:
+        if self.closed:
+            raise ValueError("I/O operation on closed file")
+        return False
+
+    def tell(self) -> int:
+        self.response._ensure_sequence()
+        return sum(map(len, self.response.response))
+
+    @property
+    def encoding(self) -> str:
+        return "utf-8"
diff --git a/backend/venv/lib/python3.12/site-packages/werkzeug/wsgi.py b/backend/venv/lib/python3.12/site-packages/werkzeug/wsgi.py
new file mode 100644
index 00000000..01d40af2
--- /dev/null
+++ b/backend/venv/lib/python3.12/site-packages/werkzeug/wsgi.py
@@ -0,0 +1,595 @@
+from __future__ import annotations
+
+import io
+import typing as t
+from functools import partial
+from functools import update_wrapper
+
+from .exceptions import ClientDisconnected
+from .exceptions import RequestEntityTooLarge
+from .sansio import utils as _sansio_utils
+from .sansio.utils import host_is_trusted  # noqa: F401 # Imported as part of API
+
+if t.TYPE_CHECKING:
+    from _typeshed.wsgi import WSGIApplication
+    from _typeshed.wsgi import WSGIEnvironment
+
+
+def responder(f: t.Callable[..., WSGIApplication]) -> WSGIApplication:
+    """Marks a function as responder.  Decorate a function with it and it
+    will automatically call the return value as WSGI application.
+
+    Example::
+
+        @responder
+        def application(environ, start_response):
+            return Response('Hello World!')
+    """
+    return update_wrapper(lambda *a: f(*a)(*a[-2:]), f)
+
+
+def get_current_url(
+    environ: WSGIEnvironment,
+    root_only: bool = False,
+    strip_querystring: bool = False,
+    host_only: bool = False,
+    trusted_hosts: t.Iterable[str] | None = None,
+) -> str:
+    """Recreate the URL for a request from the parts in a WSGI
+    environment.
+
+    The URL is an IRI, not a URI, so it may contain Unicode characters.
+    Use :func:`~werkzeug.urls.iri_to_uri` to convert it to ASCII.
+
+    :param environ: The WSGI environment to get the URL parts from.
+    :param root_only: Only build the root path, don't include the
+        remaining path or query string.
+    :param strip_querystring: Don't include the query string.
+    :param host_only: Only build the scheme and host.
+    :param trusted_hosts: A list of trusted host names to validate the
+        host against.
+    """
+    parts = {
+        "scheme": environ["wsgi.url_scheme"],
+        "host": get_host(environ, trusted_hosts),
+    }
+
+    if not host_only:
+        parts["root_path"] = environ.get("SCRIPT_NAME", "")
+
+        if not root_only:
+            parts["path"] = environ.get("PATH_INFO", "")
+
+            if not strip_querystring:
+                parts["query_string"] = environ.get("QUERY_STRING", "").encode("latin1")
+
+    return _sansio_utils.get_current_url(**parts)
+
+
+def _get_server(
+    environ: WSGIEnvironment,
+) -> tuple[str, int | None] | None:
+    name = environ.get("SERVER_NAME")
+
+    if name is None:
+        return None
+
+    try:
+        port: int | None = int(environ.get("SERVER_PORT", None))
+    except (TypeError, ValueError):
+        # unix socket
+        port = None
+
+    return name, port
+
+
+def get_host(
+    environ: WSGIEnvironment, trusted_hosts: t.Iterable[str] | None = None
+) -> str:
+    """Return the host for the given WSGI environment.
+
+    The ``Host`` header is preferred, then ``SERVER_NAME`` if it's not
+    set. The returned host will only contain the port if it is different
+    than the standard port for the protocol.
+
+    Optionally, verify that the host is trusted using
+    :func:`host_is_trusted` and raise a
+    :exc:`~werkzeug.exceptions.SecurityError` if it is not.
+
+    :param environ: A WSGI environment dict.
+    :param trusted_hosts: A list of trusted host names.
+
+    :return: Host, with port if necessary.
+    :raise ~werkzeug.exceptions.SecurityError: If the host is not
+        trusted.
+    """
+    return _sansio_utils.get_host(
+        environ["wsgi.url_scheme"],
+        environ.get("HTTP_HOST"),
+        _get_server(environ),
+        trusted_hosts,
+    )
+
+
+def get_content_length(environ: WSGIEnvironment) -> int | None:
+    """Return the ``Content-Length`` header value as an int. If the header is not given
+    or the ``Transfer-Encoding`` header is ``chunked``, ``None`` is returned to indicate
+    a streaming request. If the value is not an integer, or negative, 0 is returned.
+
+    :param environ: The WSGI environ to get the content length from.
+
+    .. versionadded:: 0.9
+    """
+    return _sansio_utils.get_content_length(
+        http_content_length=environ.get("CONTENT_LENGTH"),
+        http_transfer_encoding=environ.get("HTTP_TRANSFER_ENCODING"),
+    )
+
+
+def get_input_stream(
+    environ: WSGIEnvironment,
+    safe_fallback: bool = True,
+    max_content_length: int | None = None,
+) -> t.IO[bytes]:
+    """Return the WSGI input stream, wrapped so that it may be read safely without going
+    past the ``Content-Length`` header value or ``max_content_length``.
+
+    If ``Content-Length`` exceeds ``max_content_length``, a
+    :exc:`RequestEntityTooLarge`` ``413 Content Too Large`` error is raised.
+
+    If the WSGI server sets ``environ["wsgi.input_terminated"]``, it indicates that the
+    server handles terminating the stream, so it is safe to read directly. For example,
+    a server that knows how to handle chunked requests safely would set this.
+
+    If ``max_content_length`` is set, it can be enforced on streams if
+    ``wsgi.input_terminated`` is set. Otherwise, an empty stream is returned unless the
+    user explicitly disables this safe fallback.
+
+    If the limit is reached before the underlying stream is exhausted (such as a file
+    that is too large, or an infinite stream), the remaining contents of the stream
+    cannot be read safely. Depending on how the server handles this, clients may show a
+    "connection reset" failure instead of seeing the 413 response.
+
+    :param environ: The WSGI environ containing the stream.
+    :param safe_fallback: Return an empty stream when ``Content-Length`` is not set.
+        Disabling this allows infinite streams, which can be a denial-of-service risk.
+    :param max_content_length: The maximum length that content-length or streaming
+        requests may not exceed.
+
+    .. versionchanged:: 2.3.2
+        ``max_content_length`` is only applied to streaming requests if the server sets
+        ``wsgi.input_terminated``.
+
+    .. versionchanged:: 2.3
+        Check ``max_content_length`` and raise an error if it is exceeded.
+
+    .. versionadded:: 0.9
+    """
+    stream = t.cast(t.IO[bytes], environ["wsgi.input"])
+    content_length = get_content_length(environ)
+
+    if content_length is not None and max_content_length is not None:
+        if content_length > max_content_length:
+            raise RequestEntityTooLarge()
+
+    # A WSGI server can set this to indicate that it terminates the input stream. In
+    # that case the stream is safe without wrapping, or can enforce a max length.
+    if "wsgi.input_terminated" in environ:
+        if max_content_length is not None:
+            # If this is moved above, it can cause the stream to hang if a read attempt
+            # is made when the client sends no data. For example, the development server
+            # does not handle buffering except for chunked encoding.
+            return t.cast(
+                t.IO[bytes], LimitedStream(stream, max_content_length, is_max=True)
+            )
+
+        return stream
+
+    # No limit given, return an empty stream unless the user explicitly allows the
+    # potentially infinite stream. An infinite stream is dangerous if it's not expected,
+    # as it can tie up a worker indefinitely.
+    if content_length is None:
+        return io.BytesIO() if safe_fallback else stream
+
+    return t.cast(t.IO[bytes], LimitedStream(stream, content_length))
+
+
+def get_path_info(environ: WSGIEnvironment) -> str:
+    """Return ``PATH_INFO`` from  the WSGI environment.
+
+    :param environ: WSGI environment to get the path from.
+
+    .. versionchanged:: 3.0
+        The ``charset`` and ``errors`` parameters were removed.
+
+    .. versionadded:: 0.9
+    """
+    path: bytes = environ.get("PATH_INFO", "").encode("latin1")
+    return path.decode(errors="replace")
+
+
+class ClosingIterator:
+    """The WSGI specification requires that all middlewares and gateways
+    respect the `close` callback of the iterable returned by the application.
+    Because it is useful to add another close action to a returned iterable
+    and adding a custom iterable is a boring task this class can be used for
+    that::
+
+        return ClosingIterator(app(environ, start_response), [cleanup_session,
+                                                              cleanup_locals])
+
+    If there is just one close function it can be passed instead of the list.
+
+    A closing iterator is not needed if the application uses response objects
+    and finishes the processing if the response is started::
+
+        try:
+            return response(environ, start_response)
+        finally:
+            cleanup_session()
+            cleanup_locals()
+    """
+
+    def __init__(
+        self,
+        iterable: t.Iterable[bytes],
+        callbacks: None
+        | (t.Callable[[], None] | t.Iterable[t.Callable[[], None]]) = None,
+    ) -> None:
+        iterator = iter(iterable)
+        self._next = t.cast(t.Callable[[], bytes], partial(next, iterator))
+        if callbacks is None:
+            callbacks = []
+        elif callable(callbacks):
+            callbacks = [callbacks]
+        else:
+            callbacks = list(callbacks)
+        iterable_close = getattr(iterable, "close", None)
+        if iterable_close:
+            callbacks.insert(0, iterable_close)
+        self._callbacks = callbacks
+
+    def __iter__(self) -> ClosingIterator:
+        return self
+
+    def __next__(self) -> bytes:
+        return self._next()
+
+    def close(self) -> None:
+        for callback in self._callbacks:
+            callback()
+
+
+def wrap_file(
+    environ: WSGIEnvironment, file: t.IO[bytes], buffer_size: int = 8192
+) -> t.Iterable[bytes]:
+    """Wraps a file.  This uses the WSGI server's file wrapper if available
+    or otherwise the generic :class:`FileWrapper`.
+
+    .. versionadded:: 0.5
+
+    If the file wrapper from the WSGI server is used it's important to not
+    iterate over it from inside the application but to pass it through
+    unchanged.  If you want to pass out a file wrapper inside a response
+    object you have to set :attr:`Response.direct_passthrough` to `True`.
+
+    More information about file wrappers are available in :pep:`333`.
+
+    :param file: a :class:`file`-like object with a :meth:`~file.read` method.
+    :param buffer_size: number of bytes for one iteration.
+    """
+    return environ.get("wsgi.file_wrapper", FileWrapper)(  # type: ignore
+        file, buffer_size
+    )
+
+
+class FileWrapper:
+    """This class can be used to convert a :class:`file`-like object into
+    an iterable.  It yields `buffer_size` blocks until the file is fully
+    read.
+
+    You should not use this class directly but rather use the
+    :func:`wrap_file` function that uses the WSGI server's file wrapper
+    support if it's available.
+
+    .. versionadded:: 0.5
+
+    If you're using this object together with a :class:`Response` you have
+    to use the `direct_passthrough` mode.
+
+    :param file: a :class:`file`-like object with a :meth:`~file.read` method.
+    :param buffer_size: number of bytes for one iteration.
+    """
+
+    def __init__(self, file: t.IO[bytes], buffer_size: int = 8192) -> None:
+        self.file = file
+        self.buffer_size = buffer_size
+
+    def close(self) -> None:
+        if hasattr(self.file, "close"):
+            self.file.close()
+
+    def seekable(self) -> bool:
+        if hasattr(self.file, "seekable"):
+            return self.file.seekable()
+        if hasattr(self.file, "seek"):
+            return True
+        return False
+
+    def seek(self, *args: t.Any) -> None:
+        if hasattr(self.file, "seek"):
+            self.file.seek(*args)
+
+    def tell(self) -> int | None:
+        if hasattr(self.file, "tell"):
+            return self.file.tell()
+        return None
+
+    def __iter__(self) -> FileWrapper:
+        return self
+
+    def __next__(self) -> bytes:
+        data = self.file.read(self.buffer_size)
+        if data:
+            return data
+        raise StopIteration()
+
+
+class _RangeWrapper:
+    # private for now, but should we make it public in the future ?
+
+    """This class can be used to convert an iterable object into
+    an iterable that will only yield a piece of the underlying content.
+    It yields blocks until the underlying stream range is fully read.
+    The yielded blocks will have a size that can't exceed the original
+    iterator defined block size, but that can be smaller.
+
+    If you're using this object together with a :class:`Response` you have
+    to use the `direct_passthrough` mode.
+
+    :param iterable: an iterable object with a :meth:`__next__` method.
+    :param start_byte: byte from which read will start.
+    :param byte_range: how many bytes to read.
+    """
+
+    def __init__(
+        self,
+        iterable: t.Iterable[bytes] | t.IO[bytes],
+        start_byte: int = 0,
+        byte_range: int | None = None,
+    ):
+        self.iterable = iter(iterable)
+        self.byte_range = byte_range
+        self.start_byte = start_byte
+        self.end_byte = None
+
+        if byte_range is not None:
+            self.end_byte = start_byte + byte_range
+
+        self.read_length = 0
+        self.seekable = hasattr(iterable, "seekable") and iterable.seekable()
+        self.end_reached = False
+
+    def __iter__(self) -> _RangeWrapper:
+        return self
+
+    def _next_chunk(self) -> bytes:
+        try:
+            chunk = next(self.iterable)
+            self.read_length += len(chunk)
+            return chunk
+        except StopIteration:
+            self.end_reached = True
+            raise
+
+    def _first_iteration(self) -> tuple[bytes | None, int]:
+        chunk = None
+        if self.seekable:
+            self.iterable.seek(self.start_byte)  # type: ignore
+            self.read_length = self.iterable.tell()  # type: ignore
+            contextual_read_length = self.read_length
+        else:
+            while self.read_length <= self.start_byte:
+                chunk = self._next_chunk()
+            if chunk is not None:
+                chunk = chunk[self.start_byte - self.read_length :]
+            contextual_read_length = self.start_byte
+        return chunk, contextual_read_length
+
+    def _next(self) -> bytes:
+        if self.end_reached:
+            raise StopIteration()
+        chunk = None
+        contextual_read_length = self.read_length
+        if self.read_length == 0:
+            chunk, contextual_read_length = self._first_iteration()
+        if chunk is None:
+            chunk = self._next_chunk()
+        if self.end_byte is not None and self.read_length >= self.end_byte:
+            self.end_reached = True
+            return chunk[: self.end_byte - contextual_read_length]
+        return chunk
+
+    def __next__(self) -> bytes:
+        chunk = self._next()
+        if chunk:
+            return chunk
+        self.end_reached = True
+        raise StopIteration()
+
+    def close(self) -> None:
+        if hasattr(self.iterable, "close"):
+            self.iterable.close()
+
+
+class LimitedStream(io.RawIOBase):
+    """Wrap a stream so that it doesn't read more than a given limit. This is used to
+    limit ``wsgi.input`` to the ``Content-Length`` header value or
+    :attr:`.Request.max_content_length`.
+
+    When attempting to read after the limit has been reached, :meth:`on_exhausted` is
+    called. When the limit is a maximum, this raises :exc:`.RequestEntityTooLarge`.
+
+    If reading from the stream returns zero bytes or raises an error,
+    :meth:`on_disconnect` is called, which raises :exc:`.ClientDisconnected`. When the
+    limit is a maximum and zero bytes were read, no error is raised, since it may be the
+    end of the stream.
+
+    If the limit is reached before the underlying stream is exhausted (such as a file
+    that is too large, or an infinite stream), the remaining contents of the stream
+    cannot be read safely. Depending on how the server handles this, clients may show a
+    "connection reset" failure instead of seeing the 413 response.
+
+    :param stream: The stream to read from. Must be a readable binary IO object.
+    :param limit: The limit in bytes to not read past. Should be either the
+        ``Content-Length`` header value or ``request.max_content_length``.
+    :param is_max: Whether the given ``limit`` is ``request.max_content_length`` instead
+        of the ``Content-Length`` header value. This changes how exhausted and
+        disconnect events are handled.
+
+    .. versionchanged:: 2.3
+        Handle ``max_content_length`` differently than ``Content-Length``.
+
+    .. versionchanged:: 2.3
+        Implements ``io.RawIOBase`` rather than ``io.IOBase``.
+    """
+
+    def __init__(self, stream: t.IO[bytes], limit: int, is_max: bool = False) -> None:
+        self._stream = stream
+        self._pos = 0
+        self.limit = limit
+        self._limit_is_max = is_max
+
+    @property
+    def is_exhausted(self) -> bool:
+        """Whether the current stream position has reached the limit."""
+        return self._pos >= self.limit
+
+    def on_exhausted(self) -> None:
+        """Called when attempting to read after the limit has been reached.
+
+        The default behavior is to do nothing, unless the limit is a maximum, in which
+        case it raises :exc:`.RequestEntityTooLarge`.
+
+        .. versionchanged:: 2.3
+            Raises ``RequestEntityTooLarge`` if the limit is a maximum.
+
+        .. versionchanged:: 2.3
+            Any return value is ignored.
+        """
+        if self._limit_is_max:
+            raise RequestEntityTooLarge()
+
+    def on_disconnect(self, error: Exception | None = None) -> None:
+        """Called when an attempted read receives zero bytes before the limit was
+        reached. This indicates that the client disconnected before sending the full
+        request body.
+
+        The default behavior is to raise :exc:`.ClientDisconnected`, unless the limit is
+        a maximum and no error was raised.
+
+        .. versionchanged:: 2.3
+            Added the ``error`` parameter. Do nothing if the limit is a maximum and no
+            error was raised.
+
+        .. versionchanged:: 2.3
+            Any return value is ignored.
+        """
+        if not self._limit_is_max or error is not None:
+            raise ClientDisconnected()
+
+        # If the limit is a maximum, then we may have read zero bytes because the
+        # streaming body is complete. There's no way to distinguish that from the
+        # client disconnecting early.
+
+    def exhaust(self) -> bytes:
+        """Exhaust the stream by reading until the limit is reached or the client
+        disconnects, returning the remaining data.
+
+        .. versionchanged:: 2.3
+            Return the remaining data.
+
+        .. versionchanged:: 2.2.3
+            Handle case where wrapped stream returns fewer bytes than requested.
+        """
+        if not self.is_exhausted:
+            return self.readall()
+
+        return b""
+
+    def readinto(self, b: bytearray) -> int | None:  # type: ignore[override]
+        size = len(b)
+        remaining = self.limit - self._pos
+
+        if remaining <= 0:
+            self.on_exhausted()
+            return 0
+
+        if hasattr(self._stream, "readinto"):
+            # Use stream.readinto if it's available.
+            if size <= remaining:
+                # The size fits in the remaining limit, use the buffer directly.
+                try:
+                    out_size: int | None = self._stream.readinto(b)
+                except (OSError, ValueError) as e:
+                    self.on_disconnect(error=e)
+                    return 0
+            else:
+                # Use a temp buffer with the remaining limit as the size.
+                temp_b = bytearray(remaining)
+
+                try:
+                    out_size = self._stream.readinto(temp_b)
+                except (OSError, ValueError) as e:
+                    self.on_disconnect(error=e)
+                    return 0
+
+                if out_size:
+                    b[:out_size] = temp_b
+        else:
+            # WSGI requires that stream.read is available.
+            try:
+                data = self._stream.read(min(size, remaining))
+            except (OSError, ValueError) as e:
+                self.on_disconnect(error=e)
+                return 0
+
+            out_size = len(data)
+            b[:out_size] = data
+
+        if not out_size:
+            # Read zero bytes from the stream.
+            self.on_disconnect()
+            return 0
+
+        self._pos += out_size
+        return out_size
+
+    def readall(self) -> bytes:
+        if self.is_exhausted:
+            self.on_exhausted()
+            return b""
+
+        out = bytearray()
+
+        # The parent implementation uses "while True", which results in an extra read.
+        while not self.is_exhausted:
+            data = self.read(1024 * 64)
+
+            # Stream may return empty before a max limit is reached.
+            if not data:
+                break
+
+            out.extend(data)
+
+        return bytes(out)
+
+    def tell(self) -> int:
+        """Return the current stream position.
+
+        .. versionadded:: 0.9
+        """
+        return self._pos
+
+    def readable(self) -> bool:
+        return True
diff --git a/backend/venv/pyvenv.cfg b/backend/venv/pyvenv.cfg
new file mode 100644
index 00000000..96abf894
--- /dev/null
+++ b/backend/venv/pyvenv.cfg
@@ -0,0 +1,5 @@
+home = /Library/Frameworks/Python.framework/Versions/3.12/bin
+include-system-site-packages = false
+version = 3.12.6
+executable = /Library/Frameworks/Python.framework/Versions/3.12/bin/python3.12
+command = /Library/Frameworks/Python.framework/Versions/3.12/bin/python3 -m venv /Users/fernandapascoal/Desktop/2024.2-Equipe3/backend/venv
diff --git a/cookies.txt b/cookies.txt
new file mode 100644
index 00000000..40ce084c
--- /dev/null
+++ b/cookies.txt
@@ -0,0 +1,5 @@
+# Netscape HTTP Cookie File
+# https://curl.se/docs/http-cookies.html
+# This file was generated by libcurl! Edit at your own risk.
+
+#HttpOnly_127.0.0.1	FALSE	/	FALSE	0	access_token_cookie	eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJmcmVzaCI6ZmFsc2UsImlhdCI6MTczOTA2NDM2MSwianRpIjoiM2U1YWMzMGYtNTU1Mi00MzljLThjMjMtYzI5NjA5OTRjY2Q3IiwidHlwZSI6ImFjY2VzcyIsInN1YiI6IjEiLCJuYmYiOjE3MzkwNjQzNjEsImV4cCI6MTczOTA2NTI2MX0.1D4vgD0C04PyG-zbgw-iiKQzzkd21ZWShfyTTT2EC8w
diff --git a/cypress/e2e/features/tests/ReservaSalas.feature b/cypress/e2e/features/tests/ReservaSalas.feature
new file mode 100644
index 00000000..eab391ce
--- /dev/null
+++ b/cypress/e2e/features/tests/ReservaSalas.feature
@@ -0,0 +1,14 @@
+SCENARIO: Visualizar salas de reunião disponíveis em 15/01 das 14h às 15h com ar condicionado
+    GIVEN  eu estou logado como "professor" na página “Reservar"
+    AND no campo “tipo de sala” está selecionado “reunião"
+    AND no campo “data” está inserido “15/01/2025”
+    AND no campo “Hora Início” está selecionada a opção “14:00"
+    AND no campo “Hora Fim” está selecionada a opção “15:00"
+    AND no campo “Equipamentos” está selecionada a opção “Ar-condicionado"
+    WHEN eu seleciono a opção “Procurar"
+    THEN é exibida a sala "E001" como sala disponível e "bem" avaliada
+    THEN o user "aroldo@mail.com" volta á página de "login".
+
+
+SCENARIO: Sucesso ao reservar uma sala disponível
+    GIVEN
\ No newline at end of file
diff --git a/docs/cenarios.rb b/docs/cenarios.rb
new file mode 100644
index 00000000..1ff2741b
--- /dev/null
+++ b/docs/cenarios.rb
@@ -0,0 +1,86 @@
+FEATURE: Cadastrar e fazer a manutenção dos usuários
+    AS a usuário do SAGAA
+    I WANT TO cadastrar-me no novo sistema
+    SO THAT consiga acessar a plataforma
+
+
+SCENARIO: Sucesso no cadastro de usuário.
+    GIVEN eu estou na página de "Cadastro";
+    WHEN eu preencho o campo "Nome" com "Demosténes"
+    AND "CPF" com "123.456.789-00"
+    AND "Email" com "demostenes@example.com"
+    AND "Você é professor?" com "Não"
+    AND "Senha" com "SecurePassword123"
+    AND "Confirmar Senha" com "SecurePassword123";
+    AND clico em "Cadastrar";
+    THEN uma mensagem "Cadastro criado com sucesso!" deve ser exibida.
+
+
+SCENARIO: Sucesso no cadastro de professor.
+    GIVEN eu estou na página de "Cadastro";
+    WHEN eu preencho o campo "Nome" com "Paula"
+    AND "CPF" com "321.879.789-33"
+    AND "Email" com "vanessa@example.com"
+    AND "Você é professor?" com "Sim"
+    AND "SIAP" com "101010"
+    AND "Senha" com "12345678"
+    AND "Confirmar Senha" com "12345678";
+    AND clico em "Cadastrar";
+    THEN uma mensagem "Cadastro criado com sucesso!" deve ser exibida.
+
+
+
+
+SCENARIO: Fracasso no cadastro por campos obrigatórios não preenchidos
+    GIVEN eu estou na página de "Cadastro de usuário";
+    WHEN eu preencho o campo "Nome"
+    AND "CPF"
+    AND "Email"
+    AND "Você é professor?" com "Não"
+    AND "Senha"
+    AND  deixo o campo "Confirmar Senha" vazio;
+    AND clico em "Cadastrar"
+    THEN uma sinalização informando que o campo "Confirmar Senha" é      obrigatório deve ser exibida.
+
+
+SCENARIO: Fracasso no cadastro por duplicação de ID única.
+    GIVEN eu estou na página de "Cadastro de usuário";
+    WHEN eu preencho o campo "Nome" com "Carlos Mendes"
+    AND "CPF" com "456.123.789-10"
+    AND "Email" com "carlos.mendes@example.com"
+    AND "Você é professor?" com "Sim"
+    AND "SIAP" com "010101"
+    AND "Senha" com "Password456"
+    AND "Confirmar Senha" com "Password456";
+    AND clico em "Cadastrar";
+    THEN uma mensagem de erro "Erro: email/cpf/siap já está registrado." deve ser exibida.
+   
+   
+SCENARIO: Fracasso no cadastro por senhas que não coincidem
+    GIVEN eu estou na página de "Cadastro de usuário";
+    WHEN eu preencho o campo "Nome" com "Beatriz"
+    AND "CPF" com "789.456.123-00"
+    AND "Email" com "Beatriz.oliveira@example.com"
+    AND "Você é professor?" com "Não"
+    AND "Senha" com "MyPassword123"
+    AND "Confirmar Senha" com "DifferentPassword123";
+    AND clico em "Cadastrar";
+    THEN uma mensagem de erro "As senhas não coincidem." deve ser exibida.
+
+
+# Login
+SCENARIO: Sucesso no login
+GIVEN eu estou na página de "Login";
+WHEN eu preencho o campo "Email" com "demostenes@example.com"
+AND o campo "Senha" com "SecurePassword123";
+AND clico em "Entrar";
+THEN eu devo ser redirecionado para a página "reserva"
+
+
+SCENARIO: Fracasso no login
+GIVEN eu estou na página de "Login";
+WHEN eu preencho o campo "Email" com "demostenes@example.com"
+AND o campo "Senha" com "SenhaIncorreta123";
+AND clico em "Entrar";
+THEN uma sinalização de erro "Usuário ou senha incorretos." deve ser exibida.
+
diff --git a/instance/users.db b/instance/users.db
new file mode 100644
index 00000000..c2865699
Binary files /dev/null and b/instance/users.db differ
diff --git a/package-lock.json b/package-lock.json
index 5af242dc..02b50f05 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -18,7 +18,6 @@
       },
       "devDependencies": {
         "@badeball/cypress-cucumber-preprocessor": "^18.0.1",
-        "@bahmutov/cypress-code-coverage": "^2.6.1",
         "@bahmutov/cypress-esbuild-preprocessor": "^2.2.0",
         "@cucumber/cucumber": "^9.2.0",
         "@cypress/code-coverage": "^3.10.7",
@@ -48,34 +47,6 @@
         "vitest": "^0.31.1"
       }
     },
-    "node_modules/@actions/core": {
-      "version": "1.10.0",
-      "resolved": "https://registry.npmjs.org/@actions/core/-/core-1.10.0.tgz",
-      "integrity": "sha512-2aZDDa3zrrZbP5ZYg159sNoLRb61nQ7awl5pSvIq5Qpj81vwDzdMRKzkWJGJuwVvWpvZKx7vspJALyvaaIQyug==",
-      "dev": true,
-      "dependencies": {
-        "@actions/http-client": "^2.0.1",
-        "uuid": "^8.3.2"
-      }
-    },
-    "node_modules/@actions/core/node_modules/uuid": {
-      "version": "8.3.2",
-      "resolved": "https://registry.npmjs.org/uuid/-/uuid-8.3.2.tgz",
-      "integrity": "sha512-+NYs2QeMWy+GWFOEm9xnn6HCDp0l7QBD7ml8zLUmJ+93Q5NF0NocErnwkTkXVFNiX3/fpC6afS8Dhb/gz7R7eg==",
-      "dev": true,
-      "bin": {
-        "uuid": "dist/bin/uuid"
-      }
-    },
-    "node_modules/@actions/http-client": {
-      "version": "2.1.0",
-      "resolved": "https://registry.npmjs.org/@actions/http-client/-/http-client-2.1.0.tgz",
-      "integrity": "sha512-BonhODnXr3amchh4qkmjPMUO8mFi/zLaaCeCAJZqch8iQqyDnVIkySjB38VHAC8IJ+bnlgfOqlhpyCUZHlQsqw==",
-      "dev": true,
-      "dependencies": {
-        "tunnel": "^0.0.6"
-      }
-    },
     "node_modules/@adobe/css-tools": {
       "version": "4.2.0",
       "dev": true,
@@ -167,6 +138,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/types": "^7.22.5"
       },
@@ -178,6 +150,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/types": "^7.22.5"
       },
@@ -215,6 +188,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-environment-visitor": "^7.22.5",
@@ -237,6 +211,7 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -245,6 +220,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "regexpu-core": "^5.3.1",
@@ -261,6 +237,7 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -269,6 +246,7 @@
       "version": "0.4.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-compilation-targets": "^7.17.7",
         "@babel/helper-plugin-utils": "^7.16.7",
@@ -285,6 +263,7 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -324,6 +303,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/types": "^7.22.5"
       },
@@ -364,6 +344,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/types": "^7.22.5"
       },
@@ -383,6 +364,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-environment-visitor": "^7.22.5",
@@ -400,6 +382,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-environment-visitor": "^7.22.5",
         "@babel/helper-member-expression-to-functions": "^7.22.5",
@@ -427,6 +410,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/types": "^7.22.5"
       },
@@ -473,6 +457,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-function-name": "^7.22.5",
         "@babel/template": "^7.22.5",
@@ -580,6 +565,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -594,6 +580,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5",
@@ -610,6 +597,8 @@
       "version": "7.18.6",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-class-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -625,6 +614,8 @@
       "version": "7.20.7",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/compat-data": "^7.20.5",
         "@babel/helper-compilation-targets": "^7.20.7",
@@ -643,6 +634,7 @@
       "version": "7.21.0-placeholder-for-preset-env.2",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "engines": {
         "node": ">=6.9.0"
       },
@@ -654,6 +646,7 @@
       "version": "7.18.6",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -669,6 +662,7 @@
       "version": "7.8.4",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -680,6 +674,7 @@
       "version": "7.12.13",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.12.13"
       },
@@ -691,6 +686,7 @@
       "version": "7.14.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.14.5"
       },
@@ -705,6 +701,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -716,6 +713,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.3"
       },
@@ -727,6 +725,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -741,6 +740,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -755,6 +755,7 @@
       "version": "7.10.4",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.10.4"
       },
@@ -766,6 +767,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -777,6 +779,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -791,6 +795,7 @@
       "version": "7.10.4",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.10.4"
       },
@@ -802,6 +807,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -813,6 +819,7 @@
       "version": "7.10.4",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.10.4"
       },
@@ -824,6 +831,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -835,6 +843,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -846,6 +855,7 @@
       "version": "7.8.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.8.0"
       },
@@ -857,6 +867,7 @@
       "version": "7.14.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.14.5"
       },
@@ -871,6 +882,7 @@
       "version": "7.14.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.14.5"
       },
@@ -885,6 +897,7 @@
       "version": "7.18.6",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -900,6 +913,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -914,6 +928,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-environment-visitor": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -931,6 +946,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-module-imports": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -947,6 +963,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -961,6 +978,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -975,6 +993,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -990,6 +1009,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -1006,6 +1026,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -1028,6 +1049,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/template": "^7.22.5"
@@ -1043,6 +1065,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1057,6 +1080,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1072,6 +1096,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1086,6 +1111,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-dynamic-import": "^7.8.3"
@@ -1101,6 +1127,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-builder-binary-assignment-operator-visitor": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1116,6 +1143,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-export-namespace-from": "^7.8.3"
@@ -1131,6 +1159,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1145,6 +1174,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-compilation-targets": "^7.22.5",
         "@babel/helper-function-name": "^7.22.5",
@@ -1161,6 +1191,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-json-strings": "^7.8.3"
@@ -1176,6 +1207,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1190,6 +1222,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-logical-assignment-operators": "^7.10.4"
@@ -1205,6 +1238,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1219,6 +1253,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1234,6 +1269,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -1250,6 +1286,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-hoist-variables": "^7.22.5",
         "@babel/helper-module-transforms": "^7.22.5",
@@ -1267,6 +1304,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1282,6 +1320,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1297,6 +1336,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1311,6 +1351,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-nullish-coalescing-operator": "^7.8.3"
@@ -1326,6 +1367,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-numeric-separator": "^7.10.4"
@@ -1341,6 +1383,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/compat-data": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -1359,6 +1402,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-replace-supers": "^7.22.5"
@@ -1374,6 +1418,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-optional-catch-binding": "^7.8.3"
@@ -1389,6 +1434,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5",
@@ -1405,6 +1451,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1419,6 +1466,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1434,6 +1482,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-create-class-features-plugin": "^7.22.5",
@@ -1451,6 +1500,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1465,6 +1515,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1479,6 +1531,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-module-imports": "^7.22.5",
@@ -1497,6 +1551,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/plugin-transform-react-jsx": "^7.22.5"
       },
@@ -1539,6 +1595,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1554,6 +1612,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "regenerator-transform": "^0.15.1"
@@ -1569,6 +1628,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1583,6 +1643,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-module-imports": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -1602,6 +1664,8 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "optional": true,
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -1610,6 +1674,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1624,6 +1689,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5"
@@ -1639,6 +1705,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1653,6 +1720,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1667,6 +1735,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1681,6 +1750,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5"
       },
@@ -1695,6 +1765,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1710,6 +1781,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1725,6 +1797,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -1740,6 +1813,7 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/compat-data": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -1833,6 +1907,7 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -1841,6 +1916,7 @@
       "version": "0.1.5",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.0.0",
         "@babel/plugin-proposal-unicode-property-regex": "^7.4.4",
@@ -1856,6 +1932,8 @@
       "version": "7.22.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-validator-option": "^7.22.5",
@@ -1874,7 +1952,8 @@
     "node_modules/@babel/regjsgen": {
       "version": "0.8.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "peer": true
     },
     "node_modules/@babel/runtime": {
       "version": "7.22.5",
@@ -2202,161 +2281,6 @@
         "node": "*"
       }
     },
-    "node_modules/@bahmutov/cypress-code-coverage": {
-      "version": "2.6.1",
-      "resolved": "https://registry.npmjs.org/@bahmutov/cypress-code-coverage/-/cypress-code-coverage-2.6.1.tgz",
-      "integrity": "sha512-oK6RP+537hcIB90QdI4m/u95Pr7YYgbD4PwoDJr924t6viiN8hSc1f3BbD5D8Xuq9JagXLes68l7+lR0cBI9Hg==",
-      "dev": true,
-      "dependencies": {
-        "@actions/core": "^1.10.0",
-        "@cypress/browserify-preprocessor": "3.0.2",
-        "chalk": "4.1.2",
-        "console.table": "^0.10.0",
-        "dayjs": "1.10.7",
-        "debug": "4.3.3",
-        "execa": "4.1.0",
-        "globby": "11.1.0",
-        "istanbul-lib-coverage": "3.0.0",
-        "js-yaml": "3.14.1",
-        "nyc": "15.1.0",
-        "rimraf": "^4.4.1",
-        "sort-array": "^4.1.5"
-      },
-      "bin": {
-        "cc-merge": "bin/cc-merge.js"
-      },
-      "peerDependencies": {
-        "cypress": ">=10.0.0"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/argparse": {
-      "version": "1.0.10",
-      "resolved": "https://registry.npmjs.org/argparse/-/argparse-1.0.10.tgz",
-      "integrity": "sha512-o5Roy6tNG4SL/FOkCAN6RzjiakZS25RLYFrcMttJqbdd8BWrnA+fGz57iN5Pb06pvBGvl5gQ0B48dJlslXvoTg==",
-      "dev": true,
-      "dependencies": {
-        "sprintf-js": "~1.0.2"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/brace-expansion": {
-      "version": "2.0.1",
-      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz",
-      "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==",
-      "dev": true,
-      "dependencies": {
-        "balanced-match": "^1.0.0"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/debug": {
-      "version": "4.3.3",
-      "resolved": "https://registry.npmjs.org/debug/-/debug-4.3.3.tgz",
-      "integrity": "sha512-/zxw5+vh1Tfv+4Qn7a5nsbcJKPaSvCDhojn6FEl9vupwK2VCSDtEiEtqr8DFtzYFOdz63LBkxec7DYuc2jon6Q==",
-      "dev": true,
-      "dependencies": {
-        "ms": "2.1.2"
-      },
-      "engines": {
-        "node": ">=6.0"
-      },
-      "peerDependenciesMeta": {
-        "supports-color": {
-          "optional": true
-        }
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/glob": {
-      "version": "9.3.5",
-      "resolved": "https://registry.npmjs.org/glob/-/glob-9.3.5.tgz",
-      "integrity": "sha512-e1LleDykUz2Iu+MTYdkSsuWX8lvAjAcs0Xef0lNIu0S2wOAzuTxCJtcd9S3cijlwYF18EsU3rzb8jPVobxDh9Q==",
-      "dev": true,
-      "dependencies": {
-        "fs.realpath": "^1.0.0",
-        "minimatch": "^8.0.2",
-        "minipass": "^4.2.4",
-        "path-scurry": "^1.6.1"
-      },
-      "engines": {
-        "node": ">=16 || 14 >=14.17"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/globby": {
-      "version": "11.1.0",
-      "resolved": "https://registry.npmjs.org/globby/-/globby-11.1.0.tgz",
-      "integrity": "sha512-jhIXaOzy1sb8IyocaruWSn1TjmnBVs8Ayhcy83rmxNJ8q2uWKCAj3CnJY+KpGSXCueAPc0i05kVvVKtP1t9S3g==",
-      "dev": true,
-      "dependencies": {
-        "array-union": "^2.1.0",
-        "dir-glob": "^3.0.1",
-        "fast-glob": "^3.2.9",
-        "ignore": "^5.2.0",
-        "merge2": "^1.4.1",
-        "slash": "^3.0.0"
-      },
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/js-yaml": {
-      "version": "3.14.1",
-      "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.14.1.tgz",
-      "integrity": "sha512-okMH7OXXJ7YrN9Ok3/SXrnu4iX9yOk+25nqX4imS2npuvTYDmo/QEZoqwZkYaIDk3jVvBOTOIEgEhaLOynBS9g==",
-      "dev": true,
-      "dependencies": {
-        "argparse": "^1.0.7",
-        "esprima": "^4.0.0"
-      },
-      "bin": {
-        "js-yaml": "bin/js-yaml.js"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/minimatch": {
-      "version": "8.0.4",
-      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-8.0.4.tgz",
-      "integrity": "sha512-W0Wvr9HyFXZRGIDgCicunpQ299OKXs9RgZfaukz4qAW/pJhcpUfupc9c+OObPOFueNy8VSrZgEmDtk6Kh4WzDA==",
-      "dev": true,
-      "dependencies": {
-        "brace-expansion": "^2.0.1"
-      },
-      "engines": {
-        "node": ">=16 || 14 >=14.17"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/minipass": {
-      "version": "4.2.8",
-      "resolved": "https://registry.npmjs.org/minipass/-/minipass-4.2.8.tgz",
-      "integrity": "sha512-fNzuVyifolSLFL4NzpF+wEF4qrgqaaKX0haXPQEdQ7NKAN+WecoKMHV09YcuL/DHxrUsYQOK3MiuDf7Ip2OXfQ==",
-      "dev": true,
-      "engines": {
-        "node": ">=8"
-      }
-    },
-    "node_modules/@bahmutov/cypress-code-coverage/node_modules/rimraf": {
-      "version": "4.4.1",
-      "resolved": "https://registry.npmjs.org/rimraf/-/rimraf-4.4.1.tgz",
-      "integrity": "sha512-Gk8NlF062+T9CqNGn6h4tls3k6T1+/nXdOcSZVikNVtlRdYpA7wRJJMoXmuvOnLW844rPjdQ7JgXCYM6PPC/og==",
-      "dev": true,
-      "dependencies": {
-        "glob": "^9.2.0"
-      },
-      "bin": {
-        "rimraf": "dist/cjs/src/bin.js"
-      },
-      "engines": {
-        "node": ">=14"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
     "node_modules/@bahmutov/cypress-esbuild-preprocessor": {
       "version": "2.2.0",
       "dev": true,
@@ -2590,6 +2514,8 @@
       "version": "3.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "@babel/core": "^7.16.0",
         "@babel/plugin-proposal-class-properties": "^7.16.0",
@@ -4054,6 +3980,8 @@
       "version": "1.8.2",
       "dev": true,
       "license": "Apache-2.0",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "acorn": "^7.0.0",
         "acorn-walk": "^7.0.0",
@@ -4064,6 +3992,8 @@
       "version": "7.4.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "bin": {
         "acorn": "bin/acorn"
       },
@@ -4075,6 +4005,8 @@
       "version": "7.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.4.0"
       }
@@ -4279,15 +4211,6 @@
         "dequal": "^2.0.3"
       }
     },
-    "node_modules/array-back": {
-      "version": "5.0.0",
-      "resolved": "https://registry.npmjs.org/array-back/-/array-back-5.0.0.tgz",
-      "integrity": "sha512-kgVWwJReZWmVuWOQKEOohXKJX+nD02JAZ54D1RRWlv8L0NebauKAaFxACKzB74RTclt1+WNz5KHaLRDAPZbDEw==",
-      "dev": true,
-      "engines": {
-        "node": ">=10"
-      }
-    },
     "node_modules/array-buffer-byte-length": {
       "version": "1.0.0",
       "dev": true,
@@ -4320,6 +4243,8 @@
       "version": "5.4.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.0.0",
         "inherits": "^2.0.1",
@@ -4330,12 +4255,16 @@
     "node_modules/asn1.js/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/assert": {
       "version": "1.5.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "object-assign": "^4.1.1",
         "util": "0.10.3"
@@ -4352,12 +4281,16 @@
     "node_modules/assert/node_modules/inherits": {
       "version": "2.0.1",
       "dev": true,
-      "license": "ISC"
-    },
-    "node_modules/assert/node_modules/util": {
+      "license": "ISC",
+      "optional": true,
+      "peer": true
+    },
+    "node_modules/assert/node_modules/util": {
       "version": "0.10.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "2.0.1"
       }
@@ -4458,12 +4391,15 @@
     "node_modules/babel-plugin-add-module-exports": {
       "version": "1.0.4",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/babel-plugin-polyfill-corejs2": {
       "version": "0.4.3",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/compat-data": "^7.17.7",
         "@babel/helper-define-polyfill-provider": "^0.4.0",
@@ -4477,6 +4413,7 @@
       "version": "6.3.0",
       "dev": true,
       "license": "ISC",
+      "peer": true,
       "bin": {
         "semver": "bin/semver.js"
       }
@@ -4485,6 +4422,7 @@
       "version": "0.8.1",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-define-polyfill-provider": "^0.4.0",
         "core-js-compat": "^3.30.1"
@@ -4497,6 +4435,7 @@
       "version": "0.5.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/helper-define-polyfill-provider": "^0.4.0"
       },
@@ -4508,6 +4447,8 @@
       "version": "10.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=6.9.0"
       },
@@ -4573,7 +4514,9 @@
     "node_modules/bn.js": {
       "version": "5.2.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/brace-expansion": {
       "version": "1.1.11",
@@ -4598,12 +4541,16 @@
     "node_modules/brorand": {
       "version": "1.1.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/browser-pack": {
       "version": "6.1.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "combine-source-map": "~0.8.0",
         "defined": "^1.0.0",
@@ -4620,6 +4567,8 @@
       "version": "2.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "resolve": "^1.17.0"
       }
@@ -4633,6 +4582,8 @@
       "version": "16.5.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "assert": "^1.4.0",
         "browser-pack": "^6.0.1",
@@ -4694,6 +4645,8 @@
       "version": "1.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "buffer-xor": "^1.0.3",
         "cipher-base": "^1.0.0",
@@ -4707,6 +4660,8 @@
       "version": "1.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "browserify-aes": "^1.0.4",
         "browserify-des": "^1.0.0",
@@ -4717,6 +4672,8 @@
       "version": "1.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "cipher-base": "^1.0.1",
         "des.js": "^1.0.0",
@@ -4728,6 +4685,8 @@
       "version": "4.1.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^5.0.0",
         "randombytes": "^2.0.1"
@@ -4737,6 +4696,8 @@
       "version": "4.2.1",
       "dev": true,
       "license": "ISC",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^5.1.1",
         "browserify-rsa": "^4.0.1",
@@ -4753,6 +4714,8 @@
       "version": "3.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "string_decoder": "^1.1.1",
@@ -4766,6 +4729,8 @@
       "version": "0.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "pako": "~1.0.5"
       }
@@ -4805,6 +4770,8 @@
       "version": "5.2.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "base64-js": "^1.0.2",
         "ieee754": "^1.1.4"
@@ -4826,12 +4793,16 @@
     "node_modules/buffer-xor": {
       "version": "1.0.3",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/builtin-status-codes": {
       "version": "3.0.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/c8": {
       "version": "7.14.0",
@@ -4877,7 +4848,9 @@
     "node_modules/cached-path-relative": {
       "version": "1.1.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/cachedir": {
       "version": "2.3.0",
@@ -5026,6 +4999,8 @@
       "version": "3.5.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "anymatch": "~3.1.2",
         "braces": "~3.0.2",
@@ -5069,6 +5044,8 @@
       "version": "1.0.4",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.1",
         "safe-buffer": "^5.0.1"
@@ -5155,20 +5132,12 @@
         "wrap-ansi": "^7.0.0"
       }
     },
-    "node_modules/clone": {
-      "version": "1.0.4",
-      "resolved": "https://registry.npmjs.org/clone/-/clone-1.0.4.tgz",
-      "integrity": "sha512-JQHZ2QMW6l3aH/j6xCqQThY/9OH4D/9ls34cgkUBiEeocRTU04tHfKPBsUK1PqZCUQM7GiA0IIXJSuXHI64Kbg==",
-      "dev": true,
-      "optional": true,
-      "engines": {
-        "node": ">=0.8"
-      }
-    },
     "node_modules/coffeeify": {
       "version": "3.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "convert-source-map": "^1.3.0",
         "through2": "^2.0.0"
@@ -5181,6 +5150,8 @@
       "version": "1.12.7",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "bin": {
         "cake": "bin/cake",
         "coffee": "bin/coffee"
@@ -5214,6 +5185,8 @@
       "version": "0.8.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "convert-source-map": "~1.1.0",
         "inline-source-map": "~0.6.0",
@@ -5224,7 +5197,9 @@
     "node_modules/combine-source-map/node_modules/convert-source-map": {
       "version": "1.1.3",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/combined-stream": {
       "version": "1.0.8",
@@ -5274,6 +5249,8 @@
         "node >= 0.8"
       ],
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "buffer-from": "^1.0.0",
         "inherits": "^2.0.3",
@@ -5301,24 +5278,16 @@
     },
     "node_modules/console-browserify": {
       "version": "1.2.0",
-      "dev": true
-    },
-    "node_modules/console.table": {
-      "version": "0.10.0",
-      "resolved": "https://registry.npmjs.org/console.table/-/console.table-0.10.0.tgz",
-      "integrity": "sha512-dPyZofqggxuvSf7WXvNjuRfnsOk1YazkVP8FdxH4tcH2c37wc79/Yl6Bhr7Lsu00KMgy2ql/qCMuNu8xctZM8g==",
       "dev": true,
-      "dependencies": {
-        "easy-table": "1.1.0"
-      },
-      "engines": {
-        "node": "> 0.10"
-      }
+      "optional": true,
+      "peer": true
     },
     "node_modules/constants-browserify": {
       "version": "1.0.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/convert-source-map": {
       "version": "1.9.0",
@@ -5329,6 +5298,7 @@
       "version": "3.31.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "browserslist": "^4.21.5"
       },
@@ -5340,12 +5310,16 @@
     "node_modules/core-util-is": {
       "version": "1.0.3",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/create-ecdh": {
       "version": "4.0.4",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.1.0",
         "elliptic": "^6.5.3"
@@ -5354,12 +5328,16 @@
     "node_modules/create-ecdh/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/create-hash": {
       "version": "1.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "cipher-base": "^1.0.1",
         "inherits": "^2.0.1",
@@ -5372,6 +5350,8 @@
       "version": "1.1.7",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "cipher-base": "^1.0.3",
         "create-hash": "^1.1.0",
@@ -5398,6 +5378,8 @@
       "version": "3.12.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "browserify-cipher": "^1.0.0",
         "browserify-sign": "^4.0.0",
@@ -5536,7 +5518,9 @@
     "node_modules/dash-ast": {
       "version": "1.0.0",
       "dev": true,
-      "license": "Apache-2.0"
+      "license": "Apache-2.0",
+      "optional": true,
+      "peer": true
     },
     "node_modules/dashdash": {
       "version": "1.14.1",
@@ -5670,19 +5654,6 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
-    "node_modules/defaults": {
-      "version": "1.0.4",
-      "resolved": "https://registry.npmjs.org/defaults/-/defaults-1.0.4.tgz",
-      "integrity": "sha512-eFuaLoy/Rxalv2kr+lqMlUnrDWV+3j4pljOIJgLIhI058IQfWJ7vXhyEIHu+HtC738klGALYxOKDO0bQP3tg8A==",
-      "dev": true,
-      "optional": true,
-      "dependencies": {
-        "clone": "^1.0.2"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
     "node_modules/define-properties": {
       "version": "1.2.0",
       "dev": true,
@@ -5702,6 +5673,8 @@
       "version": "1.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -5717,6 +5690,8 @@
       "version": "2.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "JSONStream": "^1.0.3",
         "shasum-object": "^1.0.0",
@@ -5739,6 +5714,8 @@
       "version": "1.1.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.1",
         "minimalistic-assert": "^1.0.0"
@@ -5748,6 +5725,8 @@
       "version": "5.2.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "acorn-node": "^1.8.2",
         "defined": "^1.0.0",
@@ -5780,6 +5759,8 @@
       "version": "5.0.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.1.0",
         "miller-rabin": "^4.0.0",
@@ -5789,7 +5770,9 @@
     "node_modules/diffie-hellman/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/dir-glob": {
       "version": "3.0.1",
@@ -5822,6 +5805,8 @@
       "version": "1.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.4",
         "npm": ">=1.2"
@@ -5847,6 +5832,8 @@
       "version": "0.1.4",
       "dev": true,
       "license": "BSD-3-Clause",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "readable-stream": "^2.0.2"
       }
@@ -5856,15 +5843,6 @@
       "dev": true,
       "license": "MIT"
     },
-    "node_modules/easy-table": {
-      "version": "1.1.0",
-      "resolved": "https://registry.npmjs.org/easy-table/-/easy-table-1.1.0.tgz",
-      "integrity": "sha512-oq33hWOSSnl2Hoh00tZWaIPi1ievrD9aFG82/IgjlycAnW9hHx5PkJiXpxPsgEE+H7BsbVQXFVFST8TEXS6/pA==",
-      "dev": true,
-      "optionalDependencies": {
-        "wcwidth": ">=1.0.1"
-      }
-    },
     "node_modules/ecc-jsbn": {
       "version": "0.1.2",
       "dev": true,
@@ -5883,6 +5861,8 @@
       "version": "6.5.4",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.11.9",
         "brorand": "^1.1.0",
@@ -5896,7 +5876,9 @@
     "node_modules/elliptic/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/emoji-regex": {
       "version": "8.0.0",
@@ -6336,6 +6318,8 @@
       "version": "2.1.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.4.x"
       }
@@ -6344,6 +6328,8 @@
       "version": "1.0.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "md5.js": "^1.3.4",
         "safe-buffer": "^5.1.1"
@@ -6467,7 +6453,9 @@
     "node_modules/fast-safe-stringify": {
       "version": "2.1.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/fastq": {
       "version": "1.15.0",
@@ -6714,7 +6702,9 @@
     "node_modules/get-assigned-identifiers": {
       "version": "1.2.0",
       "dev": true,
-      "license": "Apache-2.0"
+      "license": "Apache-2.0",
+      "optional": true,
+      "peer": true
     },
     "node_modules/get-caller-file": {
       "version": "2.0.5",
@@ -6976,6 +6966,8 @@
       "version": "3.1.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.4",
         "readable-stream": "^3.6.0",
@@ -6989,6 +6981,8 @@
       "version": "3.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "string_decoder": "^1.1.1",
@@ -7002,6 +6996,8 @@
       "version": "1.1.7",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "minimalistic-assert": "^1.0.1"
@@ -7042,6 +7038,8 @@
       "version": "1.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "hash.js": "^1.0.3",
         "minimalistic-assert": "^1.0.0",
@@ -7068,6 +7066,8 @@
       "version": "1.1.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.10"
       }
@@ -7101,7 +7101,9 @@
     "node_modules/https-browserify": {
       "version": "1.0.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/https-proxy-agent": {
       "version": "5.0.1",
@@ -7232,6 +7234,8 @@
       "version": "0.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "source-map": "~0.5.3"
       }
@@ -7240,6 +7244,8 @@
       "version": "7.2.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "acorn-node": "^1.5.2",
         "combine-source-map": "^0.8.0",
@@ -7342,7 +7348,9 @@
     "node_modules/is-buffer": {
       "version": "1.1.6",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/is-callable": {
       "version": "1.2.7",
@@ -7370,6 +7378,7 @@
       "version": "2.12.1",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "has": "^1.0.3"
       },
@@ -7411,6 +7420,8 @@
       "version": "1.0.10",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "has-tostringtag": "^1.0.0"
       },
@@ -7636,7 +7647,9 @@
     "node_modules/isarray": {
       "version": "1.0.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/isexe": {
       "version": "2.0.0",
@@ -8119,6 +8132,8 @@
       "version": "0.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "jsonify": "~0.0.0"
       }
@@ -8164,6 +8179,8 @@
       "version": "0.0.1",
       "dev": true,
       "license": "Public Domain",
+      "optional": true,
+      "peer": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -8174,12 +8191,16 @@
       "engines": [
         "node >= 0.2.0"
       ],
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/JSONStream": {
       "version": "1.3.5",
       "dev": true,
       "license": "(MIT OR Apache-2.0)",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "jsonparse": "^1.2.0",
         "through": ">=2.2.7 <3"
@@ -8235,6 +8256,8 @@
       "version": "2.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.1",
         "stream-splicer": "^2.0.0"
@@ -8338,12 +8361,15 @@
     "node_modules/lodash.clonedeep": {
       "version": "4.5.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/lodash.debounce": {
       "version": "4.0.8",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "peer": true
     },
     "node_modules/lodash.flattendeep": {
       "version": "4.4.0",
@@ -8353,7 +8379,9 @@
     "node_modules/lodash.memoize": {
       "version": "3.0.4",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/lodash.merge": {
       "version": "4.6.2",
@@ -8533,6 +8561,8 @@
       "version": "1.3.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "hash-base": "^3.0.0",
         "inherits": "^2.0.1",
@@ -8568,6 +8598,8 @@
       "version": "4.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.0.0",
         "brorand": "^1.0.1"
@@ -8579,7 +8611,9 @@
     "node_modules/miller-rabin/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/mime-db": {
       "version": "1.52.0",
@@ -8617,12 +8651,16 @@
     "node_modules/minimalistic-assert": {
       "version": "1.0.1",
       "dev": true,
-      "license": "ISC"
+      "license": "ISC",
+      "optional": true,
+      "peer": true
     },
     "node_modules/minimalistic-crypto-utils": {
       "version": "1.0.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/minimatch": {
       "version": "3.1.2",
@@ -8668,7 +8706,9 @@
     "node_modules/mkdirp-classic": {
       "version": "0.5.3",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/mlly": {
       "version": "1.4.0",
@@ -8858,6 +8898,8 @@
       "version": "6.2.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "browser-resolve": "^2.0.0",
         "cached-path-relative": "^1.0.2",
@@ -9254,7 +9296,9 @@
     "node_modules/os-browserify": {
       "version": "0.3.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/ospath": {
       "version": "1.2.2",
@@ -9265,6 +9309,8 @@
       "version": "1.1.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "shell-quote": "^1.4.2"
       }
@@ -9347,7 +9393,9 @@
     "node_modules/pako": {
       "version": "1.0.11",
       "dev": true,
-      "license": "(MIT AND Zlib)"
+      "license": "(MIT AND Zlib)",
+      "optional": true,
+      "peer": true
     },
     "node_modules/parent-module": {
       "version": "1.0.1",
@@ -9364,6 +9412,8 @@
       "version": "1.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "path-platform": "~0.11.15"
       }
@@ -9372,6 +9422,8 @@
       "version": "5.1.6",
       "dev": true,
       "license": "ISC",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "asn1.js": "^5.2.0",
         "browserify-aes": "^1.0.0",
@@ -9394,7 +9446,9 @@
     "node_modules/path-browserify": {
       "version": "0.0.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/path-exists": {
       "version": "4.0.0",
@@ -9423,12 +9477,15 @@
     "node_modules/path-parse": {
       "version": "1.0.7",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "peer": true
     },
     "node_modules/path-platform": {
       "version": "0.11.15",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">= 0.8.0"
       }
@@ -9492,6 +9549,8 @@
       "version": "3.1.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "create-hash": "^1.1.2",
         "create-hmac": "^1.1.4",
@@ -9688,6 +9747,8 @@
       "version": "0.11.10",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">= 0.6.0"
       }
@@ -9695,7 +9756,9 @@
     "node_modules/process-nextick-args": {
       "version": "2.0.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/process-on-spawn": {
       "version": "1.0.0",
@@ -9748,6 +9811,8 @@
       "version": "4.0.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "bn.js": "^4.1.0",
         "browserify-rsa": "^4.0.0",
@@ -9760,7 +9825,9 @@
     "node_modules/public-encrypt/node_modules/bn.js": {
       "version": "4.12.0",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/pump": {
       "version": "3.0.0",
@@ -9774,7 +9841,9 @@
     "node_modules/punycode": {
       "version": "1.4.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/qs": {
       "version": "6.10.4",
@@ -9793,6 +9862,8 @@
     "node_modules/querystring-es3": {
       "version": "0.2.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.4.x"
       }
@@ -9833,6 +9904,8 @@
       "version": "1.0.4",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "randombytes": "^2.0.5",
         "safe-buffer": "^5.1.0"
@@ -9918,6 +9991,8 @@
       "version": "2.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "readable-stream": "^2.0.2"
       }
@@ -9926,6 +10001,8 @@
       "version": "2.3.8",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "core-util-is": "~1.0.0",
         "inherits": "~2.0.3",
@@ -9939,12 +10016,16 @@
     "node_modules/readable-stream/node_modules/safe-buffer": {
       "version": "5.1.2",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/readable-stream/node_modules/string_decoder": {
       "version": "1.1.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "safe-buffer": "~5.1.0"
       }
@@ -9980,12 +10061,14 @@
     "node_modules/regenerate": {
       "version": "1.4.2",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "peer": true
     },
     "node_modules/regenerate-unicode-properties": {
       "version": "10.1.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "regenerate": "^1.4.2"
       },
@@ -10002,6 +10085,7 @@
       "version": "0.15.1",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/runtime": "^7.8.4"
       }
@@ -10042,6 +10126,7 @@
       "version": "5.3.2",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "@babel/regjsgen": "^0.8.0",
         "regenerate": "^1.4.2",
@@ -10058,6 +10143,7 @@
       "version": "0.9.1",
       "dev": true,
       "license": "BSD-2-Clause",
+      "peer": true,
       "dependencies": {
         "jsesc": "~0.5.0"
       },
@@ -10068,6 +10154,7 @@
     "node_modules/regjsparser/node_modules/jsesc": {
       "version": "0.5.0",
       "dev": true,
+      "peer": true,
       "bin": {
         "jsesc": "bin/jsesc"
       }
@@ -10130,6 +10217,7 @@
       "version": "1.22.2",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "is-core-module": "^2.11.0",
         "path-parse": "^1.0.7",
@@ -10213,6 +10301,8 @@
       "version": "2.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "hash-base": "^3.0.0",
         "inherits": "^2.0.1"
@@ -10421,6 +10511,8 @@
       "version": "2.4.11",
       "dev": true,
       "license": "(MIT AND BSD-3-Clause)",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.1",
         "safe-buffer": "^5.0.1"
@@ -10433,6 +10525,8 @@
       "version": "1.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "json-stable-stringify": "~0.0.0",
         "sha.js": "~2.4.4"
@@ -10442,6 +10536,8 @@
       "version": "1.0.0",
       "dev": true,
       "license": "Apache-2.0",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "fast-safe-stringify": "^2.0.7"
       }
@@ -10469,6 +10565,8 @@
       "version": "1.8.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -10513,7 +10611,9 @@
           "url": "https://feross.org/support"
         }
       ],
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/slash": {
       "version": "3.0.0",
@@ -10536,23 +10636,12 @@
         "node": ">=8"
       }
     },
-    "node_modules/sort-array": {
-      "version": "4.1.5",
-      "resolved": "https://registry.npmjs.org/sort-array/-/sort-array-4.1.5.tgz",
-      "integrity": "sha512-Ya4peoS1fgFN42RN1REk2FgdNOeLIEMKFGJvs7VTP3OklF8+kl2SkpVliZ4tk/PurWsrWRsdNdU+tgyOBkB9sA==",
-      "dev": true,
-      "dependencies": {
-        "array-back": "^5.0.0",
-        "typical": "^6.0.1"
-      },
-      "engines": {
-        "node": ">=10"
-      }
-    },
     "node_modules/source-map": {
       "version": "0.5.7",
       "dev": true,
       "license": "BSD-3-Clause",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.10.0"
       }
@@ -10751,6 +10840,8 @@
       "version": "2.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "~2.0.1",
         "readable-stream": "^2.0.2"
@@ -10768,6 +10859,8 @@
       "version": "1.1.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "duplexer2": "~0.1.0",
         "readable-stream": "^2.0.2"
@@ -10777,6 +10870,8 @@
       "version": "3.2.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "builtin-status-codes": "^3.0.0",
         "inherits": "^2.0.4",
@@ -10788,6 +10883,8 @@
       "version": "3.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "string_decoder": "^1.1.1",
@@ -10801,6 +10898,8 @@
       "version": "2.0.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.1",
         "readable-stream": "^2.0.2"
@@ -10810,6 +10909,8 @@
       "version": "1.3.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "safe-buffer": "~5.2.0"
       }
@@ -10941,6 +11042,8 @@
       "version": "1.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "minimist": "^1.1.0"
       }
@@ -10963,6 +11066,7 @@
       "version": "1.0.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "engines": {
         "node": ">= 0.4"
       },
@@ -10979,6 +11083,8 @@
       "version": "1.4.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "acorn-node": "^1.2.0"
       }
@@ -11119,6 +11225,8 @@
       "version": "2.0.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "readable-stream": "~2.3.6",
         "xtend": "~4.0.1"
@@ -11135,6 +11243,8 @@
     "node_modules/timers-browserify": {
       "version": "1.4.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "process": "~0.11.0"
       },
@@ -11272,16 +11382,9 @@
     "node_modules/tty-browserify": {
       "version": "0.0.1",
       "dev": true,
-      "license": "MIT"
-    },
-    "node_modules/tunnel": {
-      "version": "0.0.6",
-      "resolved": "https://registry.npmjs.org/tunnel/-/tunnel-0.0.6.tgz",
-      "integrity": "sha512-1h/Lnq9yajKY2PEbBadPXj3VxsDDu844OnaAo52UVmIzIvwwtBPIuNvkjuzBlTWpfJyUbG3ez0KSBibQkj4ojg==",
-      "dev": true,
-      "engines": {
-        "node": ">=0.6.11 <=0.7.0 || >=0.7.3"
-      }
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/tunnel-agent": {
       "version": "0.6.0",
@@ -11332,7 +11435,9 @@
     "node_modules/typedarray": {
       "version": "0.0.6",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/typedarray-to-buffer": {
       "version": "3.1.5",
@@ -11354,15 +11459,6 @@
         "node": ">=14.17"
       }
     },
-    "node_modules/typical": {
-      "version": "6.0.1",
-      "resolved": "https://registry.npmjs.org/typical/-/typical-6.0.1.tgz",
-      "integrity": "sha512-+g3NEp7fJLe9DPa1TArHm9QAA7YciZmWnfAqEaFrBihQ7epOv9i99rjtgb6Iz0wh3WuQDjsCTDfgRoGnmHN81A==",
-      "dev": true,
-      "engines": {
-        "node": ">=10"
-      }
-    },
     "node_modules/ufo": {
       "version": "1.1.2",
       "dev": true,
@@ -11372,6 +11468,8 @@
       "version": "3.0.3",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "bin": {
         "umd": "bin/cli.js"
       }
@@ -11380,6 +11478,8 @@
       "version": "1.1.3",
       "dev": true,
       "license": "Apache-2.0",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "acorn-node": "^1.3.0",
         "dash-ast": "^1.0.0",
@@ -11395,6 +11495,7 @@
       "version": "2.0.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "engines": {
         "node": ">=4"
       }
@@ -11403,6 +11504,7 @@
       "version": "2.0.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "unicode-canonical-property-names-ecmascript": "^2.0.0",
         "unicode-property-aliases-ecmascript": "^2.0.0"
@@ -11415,6 +11517,7 @@
       "version": "2.1.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "engines": {
         "node": ">=4"
       }
@@ -11423,6 +11526,7 @@
       "version": "2.1.0",
       "dev": true,
       "license": "MIT",
+      "peer": true,
       "engines": {
         "node": ">=4"
       }
@@ -11500,6 +11604,8 @@
       "version": "0.11.1",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "punycode": "^1.4.1",
         "qs": "^6.11.0"
@@ -11518,6 +11624,8 @@
       "version": "6.11.2",
       "dev": true,
       "license": "BSD-3-Clause",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "side-channel": "^1.0.4"
       },
@@ -11532,6 +11640,8 @@
       "version": "0.10.4",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "2.0.3"
       }
@@ -11544,12 +11654,16 @@
     "node_modules/util-deprecate": {
       "version": "1.0.2",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/util/node_modules/inherits": {
       "version": "2.0.3",
       "dev": true,
-      "license": "ISC"
+      "license": "ISC",
+      "optional": true,
+      "peer": true
     },
     "node_modules/uuid": {
       "version": "9.0.0",
@@ -11843,7 +11957,9 @@
     "node_modules/vm-browserify": {
       "version": "1.1.2",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/w3c-xmlserializer": {
       "version": "4.0.0",
@@ -11887,6 +12003,8 @@
       "version": "4.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "anymatch": "^3.1.0",
         "browserify": "^17.0.0",
@@ -11907,6 +12025,8 @@
       "version": "17.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "assert": "^1.4.0",
         "browser-pack": "^6.0.1",
@@ -11968,6 +12088,8 @@
       "version": "2.0.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "readable-stream": "~2.3.6",
         "xtend": "~4.0.1"
@@ -11977,6 +12099,8 @@
       "version": "3.3.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.8.x"
       }
@@ -11984,12 +12108,16 @@
     "node_modules/watchify/node_modules/path-browserify": {
       "version": "1.0.1",
       "dev": true,
-      "license": "MIT"
+      "license": "MIT",
+      "optional": true,
+      "peer": true
     },
     "node_modules/watchify/node_modules/stream-browserify": {
       "version": "3.0.0",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "~2.0.4",
         "readable-stream": "^3.5.0"
@@ -11999,6 +12127,8 @@
       "version": "3.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "string_decoder": "^1.1.1",
@@ -12012,6 +12142,8 @@
       "version": "4.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "readable-stream": "3"
       }
@@ -12020,6 +12152,8 @@
       "version": "3.6.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "string_decoder": "^1.1.1",
@@ -12033,6 +12167,8 @@
       "version": "0.12.5",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "dependencies": {
         "inherits": "^2.0.3",
         "is-arguments": "^1.0.4",
@@ -12054,16 +12190,6 @@
         "node": ">=10.13.0"
       }
     },
-    "node_modules/wcwidth": {
-      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/wcwidth/-/wcwidth-1.0.1.tgz",
-      "integrity": "sha512-XHPEwS0q6TaxcvG85+8EYkbiCux2XtWG2mkc47Ng2A77BQu9+DqIOJldST4HgPkuea7dvKSj5VgX3P1d4rW8Tg==",
-      "dev": true,
-      "optional": true,
-      "dependencies": {
-        "defaults": "^1.0.3"
-      }
-    },
     "node_modules/webidl-conversions": {
       "version": "7.0.0",
       "dev": true,
@@ -12383,6 +12509,8 @@
       "version": "4.0.2",
       "dev": true,
       "license": "MIT",
+      "optional": true,
+      "peer": true,
       "engines": {
         "node": ">=0.4"
       }
@@ -12515,33 +12643,6 @@
     }
   },
   "dependencies": {
-    "@actions/core": {
-      "version": "1.10.0",
-      "resolved": "https://registry.npmjs.org/@actions/core/-/core-1.10.0.tgz",
-      "integrity": "sha512-2aZDDa3zrrZbP5ZYg159sNoLRb61nQ7awl5pSvIq5Qpj81vwDzdMRKzkWJGJuwVvWpvZKx7vspJALyvaaIQyug==",
-      "dev": true,
-      "requires": {
-        "@actions/http-client": "^2.0.1",
-        "uuid": "^8.3.2"
-      },
-      "dependencies": {
-        "uuid": {
-          "version": "8.3.2",
-          "resolved": "https://registry.npmjs.org/uuid/-/uuid-8.3.2.tgz",
-          "integrity": "sha512-+NYs2QeMWy+GWFOEm9xnn6HCDp0l7QBD7ml8zLUmJ+93Q5NF0NocErnwkTkXVFNiX3/fpC6afS8Dhb/gz7R7eg==",
-          "dev": true
-        }
-      }
-    },
-    "@actions/http-client": {
-      "version": "2.1.0",
-      "resolved": "https://registry.npmjs.org/@actions/http-client/-/http-client-2.1.0.tgz",
-      "integrity": "sha512-BonhODnXr3amchh4qkmjPMUO8mFi/zLaaCeCAJZqch8iQqyDnVIkySjB38VHAC8IJ+bnlgfOqlhpyCUZHlQsqw==",
-      "dev": true,
-      "requires": {
-        "tunnel": "^0.0.6"
-      }
-    },
     "@adobe/css-tools": {
       "version": "4.2.0",
       "dev": true
@@ -12605,6 +12706,7 @@
     "@babel/helper-annotate-as-pure": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/types": "^7.22.5"
       }
@@ -12612,6 +12714,7 @@
     "@babel/helper-builder-binary-assignment-operator-visitor": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/types": "^7.22.5"
       }
@@ -12636,6 +12739,7 @@
     "@babel/helper-create-class-features-plugin": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-environment-visitor": "^7.22.5",
@@ -12650,13 +12754,15 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
     "@babel/helper-create-regexp-features-plugin": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "regexpu-core": "^5.3.1",
@@ -12665,13 +12771,15 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
     "@babel/helper-define-polyfill-provider": {
       "version": "0.4.0",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-compilation-targets": "^7.17.7",
         "@babel/helper-plugin-utils": "^7.16.7",
@@ -12683,7 +12791,8 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
@@ -12709,6 +12818,7 @@
     "@babel/helper-member-expression-to-functions": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/types": "^7.22.5"
       }
@@ -12737,6 +12847,7 @@
     "@babel/helper-optimise-call-expression": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/types": "^7.22.5"
       }
@@ -12748,6 +12859,7 @@
     "@babel/helper-remap-async-to-generator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-environment-visitor": "^7.22.5",
@@ -12758,6 +12870,7 @@
     "@babel/helper-replace-supers": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-environment-visitor": "^7.22.5",
         "@babel/helper-member-expression-to-functions": "^7.22.5",
@@ -12777,6 +12890,7 @@
     "@babel/helper-skip-transparent-expression-wrappers": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/types": "^7.22.5"
       }
@@ -12803,6 +12917,7 @@
     "@babel/helper-wrap-function": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-function-name": "^7.22.5",
         "@babel/template": "^7.22.5",
@@ -12875,6 +12990,7 @@
     "@babel/plugin-bugfix-safari-id-destructuring-collision-in-function-expression": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -12882,6 +12998,7 @@
     "@babel/plugin-bugfix-v8-spread-parameters-in-optional-chaining": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5",
@@ -12891,6 +13008,8 @@
     "@babel/plugin-proposal-class-properties": {
       "version": "7.18.6",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-class-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -12899,6 +13018,8 @@
     "@babel/plugin-proposal-object-rest-spread": {
       "version": "7.20.7",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/compat-data": "^7.20.5",
         "@babel/helper-compilation-targets": "^7.20.7",
@@ -12910,11 +13031,13 @@
     "@babel/plugin-proposal-private-property-in-object": {
       "version": "7.21.0-placeholder-for-preset-env.2",
       "dev": true,
+      "peer": true,
       "requires": {}
     },
     "@babel/plugin-proposal-unicode-property-regex": {
       "version": "7.18.6",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -12923,6 +13046,7 @@
     "@babel/plugin-syntax-async-generators": {
       "version": "7.8.4",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -12930,6 +13054,7 @@
     "@babel/plugin-syntax-class-properties": {
       "version": "7.12.13",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.12.13"
       }
@@ -12937,6 +13062,7 @@
     "@babel/plugin-syntax-class-static-block": {
       "version": "7.14.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.14.5"
       }
@@ -12944,6 +13070,7 @@
     "@babel/plugin-syntax-dynamic-import": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -12951,6 +13078,7 @@
     "@babel/plugin-syntax-export-namespace-from": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.3"
       }
@@ -12958,6 +13086,7 @@
     "@babel/plugin-syntax-import-assertions": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -12965,6 +13094,7 @@
     "@babel/plugin-syntax-import-attributes": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -12972,6 +13102,7 @@
     "@babel/plugin-syntax-import-meta": {
       "version": "7.10.4",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.10.4"
       }
@@ -12979,6 +13110,7 @@
     "@babel/plugin-syntax-json-strings": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -12986,6 +13118,8 @@
     "@babel/plugin-syntax-jsx": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -12993,6 +13127,7 @@
     "@babel/plugin-syntax-logical-assignment-operators": {
       "version": "7.10.4",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.10.4"
       }
@@ -13000,6 +13135,7 @@
     "@babel/plugin-syntax-nullish-coalescing-operator": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -13007,6 +13143,7 @@
     "@babel/plugin-syntax-numeric-separator": {
       "version": "7.10.4",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.10.4"
       }
@@ -13014,6 +13151,7 @@
     "@babel/plugin-syntax-object-rest-spread": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -13021,6 +13159,7 @@
     "@babel/plugin-syntax-optional-catch-binding": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -13028,6 +13167,7 @@
     "@babel/plugin-syntax-optional-chaining": {
       "version": "7.8.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.8.0"
       }
@@ -13035,6 +13175,7 @@
     "@babel/plugin-syntax-private-property-in-object": {
       "version": "7.14.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.14.5"
       }
@@ -13042,6 +13183,7 @@
     "@babel/plugin-syntax-top-level-await": {
       "version": "7.14.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.14.5"
       }
@@ -13049,6 +13191,7 @@
     "@babel/plugin-syntax-unicode-sets-regex": {
       "version": "7.18.6",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.18.6",
         "@babel/helper-plugin-utils": "^7.18.6"
@@ -13057,6 +13200,7 @@
     "@babel/plugin-transform-arrow-functions": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13064,6 +13208,7 @@
     "@babel/plugin-transform-async-generator-functions": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-environment-visitor": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -13074,6 +13219,7 @@
     "@babel/plugin-transform-async-to-generator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-module-imports": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -13083,6 +13229,7 @@
     "@babel/plugin-transform-block-scoped-functions": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13090,6 +13237,7 @@
     "@babel/plugin-transform-block-scoping": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13097,6 +13245,7 @@
     "@babel/plugin-transform-class-properties": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13105,6 +13254,7 @@
     "@babel/plugin-transform-class-static-block": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -13114,6 +13264,7 @@
     "@babel/plugin-transform-classes": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -13129,6 +13280,7 @@
     "@babel/plugin-transform-computed-properties": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/template": "^7.22.5"
@@ -13137,6 +13289,7 @@
     "@babel/plugin-transform-destructuring": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13144,6 +13297,7 @@
     "@babel/plugin-transform-dotall-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13152,6 +13306,7 @@
     "@babel/plugin-transform-duplicate-keys": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13159,6 +13314,7 @@
     "@babel/plugin-transform-dynamic-import": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-dynamic-import": "^7.8.3"
@@ -13167,6 +13323,7 @@
     "@babel/plugin-transform-exponentiation-operator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-builder-binary-assignment-operator-visitor": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13175,6 +13332,7 @@
     "@babel/plugin-transform-export-namespace-from": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-export-namespace-from": "^7.8.3"
@@ -13183,6 +13341,7 @@
     "@babel/plugin-transform-for-of": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13190,6 +13349,7 @@
     "@babel/plugin-transform-function-name": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-compilation-targets": "^7.22.5",
         "@babel/helper-function-name": "^7.22.5",
@@ -13199,6 +13359,7 @@
     "@babel/plugin-transform-json-strings": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-json-strings": "^7.8.3"
@@ -13207,6 +13368,7 @@
     "@babel/plugin-transform-literals": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13214,6 +13376,7 @@
     "@babel/plugin-transform-logical-assignment-operators": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-logical-assignment-operators": "^7.10.4"
@@ -13222,6 +13385,7 @@
     "@babel/plugin-transform-member-expression-literals": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13229,6 +13393,7 @@
     "@babel/plugin-transform-modules-amd": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13237,6 +13402,7 @@
     "@babel/plugin-transform-modules-commonjs": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -13246,6 +13412,7 @@
     "@babel/plugin-transform-modules-systemjs": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-hoist-variables": "^7.22.5",
         "@babel/helper-module-transforms": "^7.22.5",
@@ -13256,6 +13423,7 @@
     "@babel/plugin-transform-modules-umd": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-module-transforms": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13264,6 +13432,7 @@
     "@babel/plugin-transform-named-capturing-groups-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13272,6 +13441,7 @@
     "@babel/plugin-transform-new-target": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13279,6 +13449,7 @@
     "@babel/plugin-transform-nullish-coalescing-operator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-nullish-coalescing-operator": "^7.8.3"
@@ -13287,6 +13458,7 @@
     "@babel/plugin-transform-numeric-separator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-numeric-separator": "^7.10.4"
@@ -13295,6 +13467,7 @@
     "@babel/plugin-transform-object-rest-spread": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/compat-data": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -13306,6 +13479,7 @@
     "@babel/plugin-transform-object-super": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-replace-supers": "^7.22.5"
@@ -13314,6 +13488,7 @@
     "@babel/plugin-transform-optional-catch-binding": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/plugin-syntax-optional-catch-binding": "^7.8.3"
@@ -13322,6 +13497,7 @@
     "@babel/plugin-transform-optional-chaining": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5",
@@ -13331,6 +13507,7 @@
     "@babel/plugin-transform-parameters": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13338,6 +13515,7 @@
     "@babel/plugin-transform-private-methods": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-class-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13346,6 +13524,7 @@
     "@babel/plugin-transform-private-property-in-object": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-create-class-features-plugin": "^7.22.5",
@@ -13356,6 +13535,7 @@
     "@babel/plugin-transform-property-literals": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13363,6 +13543,8 @@
     "@babel/plugin-transform-react-display-name": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13370,6 +13552,8 @@
     "@babel/plugin-transform-react-jsx": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-module-imports": "^7.22.5",
@@ -13381,6 +13565,8 @@
     "@babel/plugin-transform-react-jsx-development": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/plugin-transform-react-jsx": "^7.22.5"
       }
@@ -13402,6 +13588,8 @@
     "@babel/plugin-transform-react-pure-annotations": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-annotate-as-pure": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13410,6 +13598,7 @@
     "@babel/plugin-transform-regenerator": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "regenerator-transform": "^0.15.1"
@@ -13418,6 +13607,7 @@
     "@babel/plugin-transform-reserved-words": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13425,6 +13615,8 @@
     "@babel/plugin-transform-runtime": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-module-imports": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5",
@@ -13436,13 +13628,16 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
     "@babel/plugin-transform-shorthand-properties": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13450,6 +13645,7 @@
     "@babel/plugin-transform-spread": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-skip-transparent-expression-wrappers": "^7.22.5"
@@ -13458,6 +13654,7 @@
     "@babel/plugin-transform-sticky-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13465,6 +13662,7 @@
     "@babel/plugin-transform-template-literals": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13472,6 +13670,7 @@
     "@babel/plugin-transform-typeof-symbol": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13479,6 +13678,7 @@
     "@babel/plugin-transform-unicode-escapes": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5"
       }
@@ -13486,6 +13686,7 @@
     "@babel/plugin-transform-unicode-property-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13494,6 +13695,7 @@
     "@babel/plugin-transform-unicode-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13502,6 +13704,7 @@
     "@babel/plugin-transform-unicode-sets-regex": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-create-regexp-features-plugin": "^7.22.5",
         "@babel/helper-plugin-utils": "^7.22.5"
@@ -13510,6 +13713,7 @@
     "@babel/preset-env": {
       "version": "7.22.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/compat-data": "^7.22.5",
         "@babel/helper-compilation-targets": "^7.22.5",
@@ -13595,13 +13799,15 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
     "@babel/preset-modules": {
       "version": "0.1.5",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.0.0",
         "@babel/plugin-proposal-unicode-property-regex": "^7.4.4",
@@ -13613,6 +13819,8 @@
     "@babel/preset-react": {
       "version": "7.22.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/helper-plugin-utils": "^7.22.5",
         "@babel/helper-validator-option": "^7.22.5",
@@ -13624,7 +13832,8 @@
     },
     "@babel/regjsgen": {
       "version": "0.8.0",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "@babel/runtime": {
       "version": "7.22.5",
@@ -13841,116 +14050,6 @@
         }
       }
     },
-    "@bahmutov/cypress-code-coverage": {
-      "version": "2.6.1",
-      "resolved": "https://registry.npmjs.org/@bahmutov/cypress-code-coverage/-/cypress-code-coverage-2.6.1.tgz",
-      "integrity": "sha512-oK6RP+537hcIB90QdI4m/u95Pr7YYgbD4PwoDJr924t6viiN8hSc1f3BbD5D8Xuq9JagXLes68l7+lR0cBI9Hg==",
-      "dev": true,
-      "requires": {
-        "@actions/core": "^1.10.0",
-        "@cypress/browserify-preprocessor": "3.0.2",
-        "chalk": "4.1.2",
-        "console.table": "^0.10.0",
-        "dayjs": "1.10.7",
-        "debug": "4.3.3",
-        "execa": "4.1.0",
-        "globby": "11.1.0",
-        "istanbul-lib-coverage": "3.0.0",
-        "js-yaml": "3.14.1",
-        "nyc": "15.1.0",
-        "rimraf": "^4.4.1",
-        "sort-array": "^4.1.5"
-      },
-      "dependencies": {
-        "argparse": {
-          "version": "1.0.10",
-          "resolved": "https://registry.npmjs.org/argparse/-/argparse-1.0.10.tgz",
-          "integrity": "sha512-o5Roy6tNG4SL/FOkCAN6RzjiakZS25RLYFrcMttJqbdd8BWrnA+fGz57iN5Pb06pvBGvl5gQ0B48dJlslXvoTg==",
-          "dev": true,
-          "requires": {
-            "sprintf-js": "~1.0.2"
-          }
-        },
-        "brace-expansion": {
-          "version": "2.0.1",
-          "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz",
-          "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==",
-          "dev": true,
-          "requires": {
-            "balanced-match": "^1.0.0"
-          }
-        },
-        "debug": {
-          "version": "4.3.3",
-          "resolved": "https://registry.npmjs.org/debug/-/debug-4.3.3.tgz",
-          "integrity": "sha512-/zxw5+vh1Tfv+4Qn7a5nsbcJKPaSvCDhojn6FEl9vupwK2VCSDtEiEtqr8DFtzYFOdz63LBkxec7DYuc2jon6Q==",
-          "dev": true,
-          "requires": {
-            "ms": "2.1.2"
-          }
-        },
-        "glob": {
-          "version": "9.3.5",
-          "resolved": "https://registry.npmjs.org/glob/-/glob-9.3.5.tgz",
-          "integrity": "sha512-e1LleDykUz2Iu+MTYdkSsuWX8lvAjAcs0Xef0lNIu0S2wOAzuTxCJtcd9S3cijlwYF18EsU3rzb8jPVobxDh9Q==",
-          "dev": true,
-          "requires": {
-            "fs.realpath": "^1.0.0",
-            "minimatch": "^8.0.2",
-            "minipass": "^4.2.4",
-            "path-scurry": "^1.6.1"
-          }
-        },
-        "globby": {
-          "version": "11.1.0",
-          "resolved": "https://registry.npmjs.org/globby/-/globby-11.1.0.tgz",
-          "integrity": "sha512-jhIXaOzy1sb8IyocaruWSn1TjmnBVs8Ayhcy83rmxNJ8q2uWKCAj3CnJY+KpGSXCueAPc0i05kVvVKtP1t9S3g==",
-          "dev": true,
-          "requires": {
-            "array-union": "^2.1.0",
-            "dir-glob": "^3.0.1",
-            "fast-glob": "^3.2.9",
-            "ignore": "^5.2.0",
-            "merge2": "^1.4.1",
-            "slash": "^3.0.0"
-          }
-        },
-        "js-yaml": {
-          "version": "3.14.1",
-          "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.14.1.tgz",
-          "integrity": "sha512-okMH7OXXJ7YrN9Ok3/SXrnu4iX9yOk+25nqX4imS2npuvTYDmo/QEZoqwZkYaIDk3jVvBOTOIEgEhaLOynBS9g==",
-          "dev": true,
-          "requires": {
-            "argparse": "^1.0.7",
-            "esprima": "^4.0.0"
-          }
-        },
-        "minimatch": {
-          "version": "8.0.4",
-          "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-8.0.4.tgz",
-          "integrity": "sha512-W0Wvr9HyFXZRGIDgCicunpQ299OKXs9RgZfaukz4qAW/pJhcpUfupc9c+OObPOFueNy8VSrZgEmDtk6Kh4WzDA==",
-          "dev": true,
-          "requires": {
-            "brace-expansion": "^2.0.1"
-          }
-        },
-        "minipass": {
-          "version": "4.2.8",
-          "resolved": "https://registry.npmjs.org/minipass/-/minipass-4.2.8.tgz",
-          "integrity": "sha512-fNzuVyifolSLFL4NzpF+wEF4qrgqaaKX0haXPQEdQ7NKAN+WecoKMHV09YcuL/DHxrUsYQOK3MiuDf7Ip2OXfQ==",
-          "dev": true
-        },
-        "rimraf": {
-          "version": "4.4.1",
-          "resolved": "https://registry.npmjs.org/rimraf/-/rimraf-4.4.1.tgz",
-          "integrity": "sha512-Gk8NlF062+T9CqNGn6h4tls3k6T1+/nXdOcSZVikNVtlRdYpA7wRJJMoXmuvOnLW844rPjdQ7JgXCYM6PPC/og==",
-          "dev": true,
-          "requires": {
-            "glob": "^9.2.0"
-          }
-        }
-      }
-    },
     "@bahmutov/cypress-esbuild-preprocessor": {
       "version": "2.2.0",
       "dev": true,
@@ -14126,6 +14225,8 @@
     "@cypress/browserify-preprocessor": {
       "version": "3.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "@babel/core": "^7.16.0",
         "@babel/plugin-proposal-class-properties": "^7.16.0",
@@ -15130,6 +15231,8 @@
     "acorn-node": {
       "version": "1.8.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "acorn": "^7.0.0",
         "acorn-walk": "^7.0.0",
@@ -15138,13 +15241,17 @@
       "dependencies": {
         "acorn": {
           "version": "7.4.1",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
     "acorn-walk": {
       "version": "7.2.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "agent-base": {
       "version": "6.0.2",
@@ -15267,12 +15374,6 @@
         "dequal": "^2.0.3"
       }
     },
-    "array-back": {
-      "version": "5.0.0",
-      "resolved": "https://registry.npmjs.org/array-back/-/array-back-5.0.0.tgz",
-      "integrity": "sha512-kgVWwJReZWmVuWOQKEOohXKJX+nD02JAZ54D1RRWlv8L0NebauKAaFxACKzB74RTclt1+WNz5KHaLRDAPZbDEw==",
-      "dev": true
-    },
     "array-buffer-byte-length": {
       "version": "1.0.0",
       "dev": true,
@@ -15295,6 +15396,8 @@
     "asn1.js": {
       "version": "5.4.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.0.0",
         "inherits": "^2.0.1",
@@ -15304,13 +15407,17 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
     "assert": {
       "version": "1.5.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "object-assign": "^4.1.1",
         "util": "0.10.3"
@@ -15318,11 +15425,15 @@
       "dependencies": {
         "inherits": {
           "version": "2.0.1",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         },
         "util": {
           "version": "0.10.3",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "2.0.1"
           }
@@ -15392,11 +15503,14 @@
     },
     "babel-plugin-add-module-exports": {
       "version": "1.0.4",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "babel-plugin-polyfill-corejs2": {
       "version": "0.4.3",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/compat-data": "^7.17.7",
         "@babel/helper-define-polyfill-provider": "^0.4.0",
@@ -15405,13 +15519,15 @@
       "dependencies": {
         "semver": {
           "version": "6.3.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
     "babel-plugin-polyfill-corejs3": {
       "version": "0.8.1",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-define-polyfill-provider": "^0.4.0",
         "core-js-compat": "^3.30.1"
@@ -15420,6 +15536,7 @@
     "babel-plugin-polyfill-regenerator": {
       "version": "0.5.0",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/helper-define-polyfill-provider": "^0.4.0"
       }
@@ -15427,6 +15544,8 @@
     "babelify": {
       "version": "10.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {}
     },
     "balanced-match": {
@@ -15462,7 +15581,9 @@
     },
     "bn.js": {
       "version": "5.2.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "brace-expansion": {
       "version": "1.1.11",
@@ -15481,11 +15602,15 @@
     },
     "brorand": {
       "version": "1.1.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "browser-pack": {
       "version": "6.1.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "combine-source-map": "~0.8.0",
         "defined": "^1.0.0",
@@ -15498,6 +15623,8 @@
     "browser-resolve": {
       "version": "2.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "resolve": "^1.17.0"
       }
@@ -15509,6 +15636,8 @@
     "browserify": {
       "version": "16.5.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "assert": "^1.4.0",
         "browser-pack": "^6.0.1",
@@ -15563,6 +15692,8 @@
     "browserify-aes": {
       "version": "1.2.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "buffer-xor": "^1.0.3",
         "cipher-base": "^1.0.0",
@@ -15575,6 +15706,8 @@
     "browserify-cipher": {
       "version": "1.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "browserify-aes": "^1.0.4",
         "browserify-des": "^1.0.0",
@@ -15584,6 +15717,8 @@
     "browserify-des": {
       "version": "1.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "cipher-base": "^1.0.1",
         "des.js": "^1.0.0",
@@ -15594,6 +15729,8 @@
     "browserify-rsa": {
       "version": "4.1.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^5.0.0",
         "randombytes": "^2.0.1"
@@ -15602,6 +15739,8 @@
     "browserify-sign": {
       "version": "4.2.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^5.1.1",
         "browserify-rsa": "^4.0.1",
@@ -15617,6 +15756,8 @@
         "readable-stream": {
           "version": "3.6.2",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "^2.0.3",
             "string_decoder": "^1.1.1",
@@ -15628,6 +15769,8 @@
     "browserify-zlib": {
       "version": "0.2.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "pako": "~1.0.5"
       }
@@ -15645,6 +15788,8 @@
     "buffer": {
       "version": "5.2.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "base64-js": "^1.0.2",
         "ieee754": "^1.1.4"
@@ -15660,11 +15805,15 @@
     },
     "buffer-xor": {
       "version": "1.0.3",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "builtin-status-codes": {
       "version": "3.0.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "c8": {
       "version": "7.14.0",
@@ -15696,7 +15845,9 @@
     },
     "cached-path-relative": {
       "version": "1.1.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "cachedir": {
       "version": "2.3.0",
@@ -15786,6 +15937,8 @@
     "chokidar": {
       "version": "3.5.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "anymatch": "~3.1.2",
         "braces": "~3.0.2",
@@ -15809,6 +15962,8 @@
     "cipher-base": {
       "version": "1.0.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.1",
         "safe-buffer": "^5.0.1"
@@ -15867,16 +16022,11 @@
         "wrap-ansi": "^7.0.0"
       }
     },
-    "clone": {
-      "version": "1.0.4",
-      "resolved": "https://registry.npmjs.org/clone/-/clone-1.0.4.tgz",
-      "integrity": "sha512-JQHZ2QMW6l3aH/j6xCqQThY/9OH4D/9ls34cgkUBiEeocRTU04tHfKPBsUK1PqZCUQM7GiA0IIXJSuXHI64Kbg==",
-      "dev": true,
-      "optional": true
-    },
     "coffeeify": {
       "version": "3.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "convert-source-map": "^1.3.0",
         "through2": "^2.0.0"
@@ -15884,7 +16034,9 @@
     },
     "coffeescript": {
       "version": "1.12.7",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "color-convert": {
       "version": "2.0.1",
@@ -15904,6 +16056,8 @@
     "combine-source-map": {
       "version": "0.8.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "convert-source-map": "~1.1.0",
         "inline-source-map": "~0.6.0",
@@ -15913,7 +16067,9 @@
       "dependencies": {
         "convert-source-map": {
           "version": "1.1.3",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -15946,6 +16102,8 @@
     "concat-stream": {
       "version": "1.6.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "buffer-from": "^1.0.0",
         "inherits": "^2.0.3",
@@ -15969,20 +16127,15 @@
     },
     "console-browserify": {
       "version": "1.2.0",
-      "dev": true
-    },
-    "console.table": {
-      "version": "0.10.0",
-      "resolved": "https://registry.npmjs.org/console.table/-/console.table-0.10.0.tgz",
-      "integrity": "sha512-dPyZofqggxuvSf7WXvNjuRfnsOk1YazkVP8FdxH4tcH2c37wc79/Yl6Bhr7Lsu00KMgy2ql/qCMuNu8xctZM8g==",
       "dev": true,
-      "requires": {
-        "easy-table": "1.1.0"
-      }
+      "optional": true,
+      "peer": true
     },
     "constants-browserify": {
       "version": "1.0.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "convert-source-map": {
       "version": "1.9.0",
@@ -15991,17 +16144,22 @@
     "core-js-compat": {
       "version": "3.31.0",
       "dev": true,
+      "peer": true,
       "requires": {
         "browserslist": "^4.21.5"
       }
     },
     "core-util-is": {
       "version": "1.0.3",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "create-ecdh": {
       "version": "4.0.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.1.0",
         "elliptic": "^6.5.3"
@@ -16009,13 +16167,17 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
     "create-hash": {
       "version": "1.2.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "cipher-base": "^1.0.1",
         "inherits": "^2.0.1",
@@ -16027,6 +16189,8 @@
     "create-hmac": {
       "version": "1.1.7",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "cipher-base": "^1.0.3",
         "create-hash": "^1.1.0",
@@ -16048,6 +16212,8 @@
     "crypto-browserify": {
       "version": "3.12.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "browserify-cipher": "^1.0.0",
         "browserify-sign": "^4.0.0",
@@ -16149,7 +16315,9 @@
     },
     "dash-ast": {
       "version": "1.0.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "dashdash": {
       "version": "1.14.1",
@@ -16241,16 +16409,6 @@
         "strip-bom": "^4.0.0"
       }
     },
-    "defaults": {
-      "version": "1.0.4",
-      "resolved": "https://registry.npmjs.org/defaults/-/defaults-1.0.4.tgz",
-      "integrity": "sha512-eFuaLoy/Rxalv2kr+lqMlUnrDWV+3j4pljOIJgLIhI058IQfWJ7vXhyEIHu+HtC738klGALYxOKDO0bQP3tg8A==",
-      "dev": true,
-      "optional": true,
-      "requires": {
-        "clone": "^1.0.2"
-      }
-    },
     "define-properties": {
       "version": "1.2.0",
       "dev": true,
@@ -16261,7 +16419,9 @@
     },
     "defined": {
       "version": "1.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "delayed-stream": {
       "version": "1.0.0"
@@ -16269,6 +16429,8 @@
     "deps-sort": {
       "version": "2.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "JSONStream": "^1.0.3",
         "shasum-object": "^1.0.0",
@@ -16283,6 +16445,8 @@
     "des.js": {
       "version": "1.1.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.1",
         "minimalistic-assert": "^1.0.0"
@@ -16291,6 +16455,8 @@
     "detective": {
       "version": "5.2.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "acorn-node": "^1.8.2",
         "defined": "^1.0.0",
@@ -16308,6 +16474,8 @@
     "diffie-hellman": {
       "version": "5.0.3",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.1.0",
         "miller-rabin": "^4.0.0",
@@ -16316,7 +16484,9 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -16340,7 +16510,9 @@
     },
     "domain-browser": {
       "version": "1.2.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "domexception": {
       "version": "4.0.0",
@@ -16356,6 +16528,8 @@
     "duplexer2": {
       "version": "0.1.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "readable-stream": "^2.0.2"
       }
@@ -16364,15 +16538,6 @@
       "version": "0.2.0",
       "dev": true
     },
-    "easy-table": {
-      "version": "1.1.0",
-      "resolved": "https://registry.npmjs.org/easy-table/-/easy-table-1.1.0.tgz",
-      "integrity": "sha512-oq33hWOSSnl2Hoh00tZWaIPi1ievrD9aFG82/IgjlycAnW9hHx5PkJiXpxPsgEE+H7BsbVQXFVFST8TEXS6/pA==",
-      "dev": true,
-      "requires": {
-        "wcwidth": ">=1.0.1"
-      }
-    },
     "ecc-jsbn": {
       "version": "0.1.2",
       "dev": true,
@@ -16388,6 +16553,8 @@
     "elliptic": {
       "version": "6.5.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.11.9",
         "brorand": "^1.1.0",
@@ -16400,7 +16567,9 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -16689,11 +16858,15 @@
     },
     "events": {
       "version": "2.1.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "evp_bytestokey": {
       "version": "1.0.3",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "md5.js": "^1.3.4",
         "safe-buffer": "^5.1.1"
@@ -16779,7 +16952,9 @@
     },
     "fast-safe-stringify": {
       "version": "2.1.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "fastq": {
       "version": "1.15.0",
@@ -16920,7 +17095,9 @@
     },
     "get-assigned-identifiers": {
       "version": "1.2.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "get-caller-file": {
       "version": "2.0.5",
@@ -17078,6 +17255,8 @@
     "hash-base": {
       "version": "3.1.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.4",
         "readable-stream": "^3.6.0",
@@ -17087,6 +17266,8 @@
         "readable-stream": {
           "version": "3.6.2",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "^2.0.3",
             "string_decoder": "^1.1.1",
@@ -17098,6 +17279,8 @@
     "hash.js": {
       "version": "1.1.7",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.3",
         "minimalistic-assert": "^1.0.1"
@@ -17124,6 +17307,8 @@
     "hmac-drbg": {
       "version": "1.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "hash.js": "^1.0.3",
         "minimalistic-assert": "^1.0.0",
@@ -17143,7 +17328,9 @@
     },
     "htmlescape": {
       "version": "1.1.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "http-proxy-agent": {
       "version": "5.0.0",
@@ -17165,7 +17352,9 @@
     },
     "https-browserify": {
       "version": "1.0.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "https-proxy-agent": {
       "version": "5.0.1",
@@ -17233,6 +17422,8 @@
     "inline-source-map": {
       "version": "0.6.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "source-map": "~0.5.3"
       }
@@ -17240,6 +17431,8 @@
     "insert-module-globals": {
       "version": "7.2.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "acorn-node": "^1.5.2",
         "combine-source-map": "^0.8.0",
@@ -17307,7 +17500,9 @@
     },
     "is-buffer": {
       "version": "1.1.6",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "is-callable": {
       "version": "1.2.7",
@@ -17323,6 +17518,7 @@
     "is-core-module": {
       "version": "2.12.1",
       "dev": true,
+      "peer": true,
       "requires": {
         "has": "^1.0.3"
       }
@@ -17345,6 +17541,8 @@
     "is-generator-function": {
       "version": "1.0.10",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "has-tostringtag": "^1.0.0"
       }
@@ -17465,7 +17663,9 @@
     },
     "isarray": {
       "version": "1.0.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "isexe": {
       "version": "2.0.0",
@@ -17795,6 +17995,8 @@
     "json-stable-stringify": {
       "version": "0.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "jsonify": "~0.0.0"
       }
@@ -17825,15 +18027,21 @@
     },
     "jsonify": {
       "version": "0.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "jsonparse": {
       "version": "1.3.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "JSONStream": {
       "version": "1.3.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "jsonparse": "^1.2.0",
         "through": ">=2.2.7 <3"
@@ -17874,6 +18082,8 @@
     "labeled-stream-splicer": {
       "version": "2.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.1",
         "stream-splicer": "^2.0.0"
@@ -17935,11 +18145,14 @@
     },
     "lodash.clonedeep": {
       "version": "4.5.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "lodash.debounce": {
       "version": "4.0.8",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "lodash.flattendeep": {
       "version": "4.4.0",
@@ -17947,7 +18160,9 @@
     },
     "lodash.memoize": {
       "version": "3.0.4",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "lodash.merge": {
       "version": "4.6.2",
@@ -18068,6 +18283,8 @@
     "md5.js": {
       "version": "1.3.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "hash-base": "^3.0.0",
         "inherits": "^2.0.1",
@@ -18093,6 +18310,8 @@
     "miller-rabin": {
       "version": "4.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.0.0",
         "brorand": "^1.0.1"
@@ -18100,7 +18319,9 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -18123,11 +18344,15 @@
     },
     "minimalistic-assert": {
       "version": "1.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "minimalistic-crypto-utils": {
       "version": "1.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "minimatch": {
       "version": "3.1.2",
@@ -18150,7 +18375,9 @@
     },
     "mkdirp-classic": {
       "version": "0.5.3",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "mlly": {
       "version": "1.4.0",
@@ -18276,6 +18503,8 @@
     "module-deps": {
       "version": "6.2.3",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "browser-resolve": "^2.0.0",
         "cached-path-relative": "^1.0.2",
@@ -18546,7 +18775,9 @@
     },
     "os-browserify": {
       "version": "0.3.0",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "ospath": {
       "version": "1.2.2",
@@ -18555,6 +18786,8 @@
     "outpipe": {
       "version": "1.1.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "shell-quote": "^1.4.2"
       }
@@ -18603,7 +18836,9 @@
     },
     "pako": {
       "version": "1.0.11",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "parent-module": {
       "version": "1.0.1",
@@ -18615,6 +18850,8 @@
     "parents": {
       "version": "1.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "path-platform": "~0.11.15"
       }
@@ -18622,6 +18859,8 @@
     "parse-asn1": {
       "version": "5.1.6",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "asn1.js": "^5.2.0",
         "browserify-aes": "^1.0.0",
@@ -18639,7 +18878,9 @@
     },
     "path-browserify": {
       "version": "0.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "path-exists": {
       "version": "4.0.0",
@@ -18655,11 +18896,14 @@
     },
     "path-parse": {
       "version": "1.0.7",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "path-platform": {
       "version": "0.11.15",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "path-scurry": {
       "version": "1.9.2",
@@ -18697,6 +18941,8 @@
     "pbkdf2": {
       "version": "3.1.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "create-hash": "^1.1.2",
         "create-hmac": "^1.1.4",
@@ -18810,11 +19056,15 @@
     },
     "process": {
       "version": "0.11.10",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "process-nextick-args": {
       "version": "2.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "process-on-spawn": {
       "version": "1.0.0",
@@ -18848,6 +19098,8 @@
     "public-encrypt": {
       "version": "4.0.3",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "bn.js": "^4.1.0",
         "browserify-rsa": "^4.0.0",
@@ -18859,7 +19111,9 @@
       "dependencies": {
         "bn.js": {
           "version": "4.12.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -18873,7 +19127,9 @@
     },
     "punycode": {
       "version": "1.4.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "qs": {
       "version": "6.10.4",
@@ -18884,7 +19140,9 @@
     },
     "querystring-es3": {
       "version": "0.2.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "querystringify": {
       "version": "2.2.0",
@@ -18904,6 +19162,8 @@
     "randomfill": {
       "version": "1.0.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "randombytes": "^2.0.5",
         "safe-buffer": "^5.1.0"
@@ -18950,6 +19210,8 @@
     "read-only-stream": {
       "version": "2.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "readable-stream": "^2.0.2"
       }
@@ -18957,6 +19219,8 @@
     "readable-stream": {
       "version": "2.3.8",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "core-util-is": "~1.0.0",
         "inherits": "~2.0.3",
@@ -18969,11 +19233,15 @@
       "dependencies": {
         "safe-buffer": {
           "version": "5.1.2",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         },
         "string_decoder": {
           "version": "1.1.1",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "safe-buffer": "~5.1.0"
           }
@@ -19001,11 +19269,13 @@
     },
     "regenerate": {
       "version": "1.4.2",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "regenerate-unicode-properties": {
       "version": "10.1.0",
       "dev": true,
+      "peer": true,
       "requires": {
         "regenerate": "^1.4.2"
       }
@@ -19017,6 +19287,7 @@
     "regenerator-transform": {
       "version": "0.15.1",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/runtime": "^7.8.4"
       }
@@ -19044,6 +19315,7 @@
     "regexpu-core": {
       "version": "5.3.2",
       "dev": true,
+      "peer": true,
       "requires": {
         "@babel/regjsgen": "^0.8.0",
         "regenerate": "^1.4.2",
@@ -19056,13 +19328,15 @@
     "regjsparser": {
       "version": "0.9.1",
       "dev": true,
+      "peer": true,
       "requires": {
         "jsesc": "~0.5.0"
       },
       "dependencies": {
         "jsesc": {
           "version": "0.5.0",
-          "dev": true
+          "dev": true,
+          "peer": true
         }
       }
     },
@@ -19104,6 +19378,7 @@
     "resolve": {
       "version": "1.22.2",
       "dev": true,
+      "peer": true,
       "requires": {
         "is-core-module": "^2.11.0",
         "path-parse": "^1.0.7",
@@ -19153,6 +19428,8 @@
     "ripemd160": {
       "version": "2.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "hash-base": "^3.0.0",
         "inherits": "^2.0.1"
@@ -19284,6 +19561,8 @@
     "sha.js": {
       "version": "2.4.11",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.1",
         "safe-buffer": "^5.0.1"
@@ -19292,6 +19571,8 @@
     "shasum": {
       "version": "1.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "json-stable-stringify": "~0.0.0",
         "sha.js": "~2.4.4"
@@ -19300,6 +19581,8 @@
     "shasum-object": {
       "version": "1.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "fast-safe-stringify": "^2.0.7"
       }
@@ -19317,7 +19600,9 @@
     },
     "shell-quote": {
       "version": "1.8.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "side-channel": {
       "version": "1.0.4",
@@ -19338,7 +19623,9 @@
     },
     "simple-concat": {
       "version": "1.0.1",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "slash": {
       "version": "3.0.0",
@@ -19353,19 +19640,11 @@
         "is-fullwidth-code-point": "^3.0.0"
       }
     },
-    "sort-array": {
-      "version": "4.1.5",
-      "resolved": "https://registry.npmjs.org/sort-array/-/sort-array-4.1.5.tgz",
-      "integrity": "sha512-Ya4peoS1fgFN42RN1REk2FgdNOeLIEMKFGJvs7VTP3OklF8+kl2SkpVliZ4tk/PurWsrWRsdNdU+tgyOBkB9sA==",
-      "dev": true,
-      "requires": {
-        "array-back": "^5.0.0",
-        "typical": "^6.0.1"
-      }
-    },
     "source-map": {
       "version": "0.5.7",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "source-map-js": {
       "version": "1.0.2",
@@ -19497,6 +19776,8 @@
     "stream-browserify": {
       "version": "2.0.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "~2.0.1",
         "readable-stream": "^2.0.2"
@@ -19512,6 +19793,8 @@
     "stream-combiner2": {
       "version": "1.1.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "duplexer2": "~0.1.0",
         "readable-stream": "^2.0.2"
@@ -19520,6 +19803,8 @@
     "stream-http": {
       "version": "3.2.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "builtin-status-codes": "^3.0.0",
         "inherits": "^2.0.4",
@@ -19530,6 +19815,8 @@
         "readable-stream": {
           "version": "3.6.2",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "^2.0.3",
             "string_decoder": "^1.1.1",
@@ -19541,6 +19828,8 @@
     "stream-splicer": {
       "version": "2.0.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "^2.0.1",
         "readable-stream": "^2.0.2"
@@ -19549,6 +19838,8 @@
     "string_decoder": {
       "version": "1.3.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "safe-buffer": "~5.2.0"
       }
@@ -19630,6 +19921,8 @@
     "subarg": {
       "version": "1.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "minimist": "^1.1.0"
       }
@@ -19643,7 +19936,8 @@
     },
     "supports-preserve-symlinks-flag": {
       "version": "1.0.0",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "symbol-tree": {
       "version": "3.2.4",
@@ -19652,6 +19946,8 @@
     "syntax-error": {
       "version": "1.4.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "acorn-node": "^1.2.0"
       }
@@ -19741,6 +20037,8 @@
     "through2": {
       "version": "2.0.5",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "readable-stream": "~2.3.6",
         "xtend": "~4.0.1"
@@ -19753,6 +20051,8 @@
     "timers-browserify": {
       "version": "1.4.2",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "process": "~0.11.0"
       }
@@ -19841,13 +20141,9 @@
     },
     "tty-browserify": {
       "version": "0.0.1",
-      "dev": true
-    },
-    "tunnel": {
-      "version": "0.0.6",
-      "resolved": "https://registry.npmjs.org/tunnel/-/tunnel-0.0.6.tgz",
-      "integrity": "sha512-1h/Lnq9yajKY2PEbBadPXj3VxsDDu844OnaAo52UVmIzIvwwtBPIuNvkjuzBlTWpfJyUbG3ez0KSBibQkj4ojg==",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "tunnel-agent": {
       "version": "0.6.0",
@@ -19877,7 +20173,9 @@
     },
     "typedarray": {
       "version": "0.0.6",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "typedarray-to-buffer": {
       "version": "3.1.5",
@@ -19890,23 +20188,21 @@
       "version": "5.1.3",
       "dev": true
     },
-    "typical": {
-      "version": "6.0.1",
-      "resolved": "https://registry.npmjs.org/typical/-/typical-6.0.1.tgz",
-      "integrity": "sha512-+g3NEp7fJLe9DPa1TArHm9QAA7YciZmWnfAqEaFrBihQ7epOv9i99rjtgb6Iz0wh3WuQDjsCTDfgRoGnmHN81A==",
-      "dev": true
-    },
     "ufo": {
       "version": "1.1.2",
       "dev": true
     },
     "umd": {
       "version": "3.0.3",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "undeclared-identifiers": {
       "version": "1.1.3",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "acorn-node": "^1.3.0",
         "dash-ast": "^1.0.0",
@@ -19917,11 +20213,13 @@
     },
     "unicode-canonical-property-names-ecmascript": {
       "version": "2.0.0",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "unicode-match-property-ecmascript": {
       "version": "2.0.0",
       "dev": true,
+      "peer": true,
       "requires": {
         "unicode-canonical-property-names-ecmascript": "^2.0.0",
         "unicode-property-aliases-ecmascript": "^2.0.0"
@@ -19929,11 +20227,13 @@
     },
     "unicode-match-property-value-ecmascript": {
       "version": "2.1.0",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "unicode-property-aliases-ecmascript": {
       "version": "2.1.0",
-      "dev": true
+      "dev": true,
+      "peer": true
     },
     "universalify": {
       "version": "2.0.0",
@@ -19974,6 +20274,8 @@
     "url": {
       "version": "0.11.1",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "punycode": "^1.4.1",
         "qs": "^6.11.0"
@@ -19982,6 +20284,8 @@
         "qs": {
           "version": "6.11.2",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "side-channel": "^1.0.4"
           }
@@ -19999,13 +20303,17 @@
     "util": {
       "version": "0.10.4",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "inherits": "2.0.3"
       },
       "dependencies": {
         "inherits": {
           "version": "2.0.3",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         }
       }
     },
@@ -20015,7 +20323,9 @@
     },
     "util-deprecate": {
       "version": "1.0.2",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "uuid": {
       "version": "9.0.0",
@@ -20172,7 +20482,9 @@
     },
     "vm-browserify": {
       "version": "1.1.2",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "w3c-xmlserializer": {
       "version": "4.0.0",
@@ -20205,6 +20517,8 @@
     "watchify": {
       "version": "4.0.0",
       "dev": true,
+      "optional": true,
+      "peer": true,
       "requires": {
         "anymatch": "^3.1.0",
         "browserify": "^17.0.0",
@@ -20218,6 +20532,8 @@
         "browserify": {
           "version": "17.0.0",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "assert": "^1.4.0",
             "browser-pack": "^6.0.1",
@@ -20272,6 +20588,8 @@
             "through2": {
               "version": "2.0.5",
               "dev": true,
+              "optional": true,
+              "peer": true,
               "requires": {
                 "readable-stream": "~2.3.6",
                 "xtend": "~4.0.1"
@@ -20281,15 +20599,21 @@
         },
         "events": {
           "version": "3.3.0",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         },
         "path-browserify": {
           "version": "1.0.1",
-          "dev": true
+          "dev": true,
+          "optional": true,
+          "peer": true
         },
         "stream-browserify": {
           "version": "3.0.0",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "~2.0.4",
             "readable-stream": "^3.5.0"
@@ -20298,6 +20622,8 @@
             "readable-stream": {
               "version": "3.6.2",
               "dev": true,
+              "optional": true,
+              "peer": true,
               "requires": {
                 "inherits": "^2.0.3",
                 "string_decoder": "^1.1.1",
@@ -20309,6 +20635,8 @@
         "through2": {
           "version": "4.0.2",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "readable-stream": "3"
           },
@@ -20316,6 +20644,8 @@
             "readable-stream": {
               "version": "3.6.2",
               "dev": true,
+              "optional": true,
+              "peer": true,
               "requires": {
                 "inherits": "^2.0.3",
                 "string_decoder": "^1.1.1",
@@ -20327,6 +20657,8 @@
         "util": {
           "version": "0.12.5",
           "dev": true,
+          "optional": true,
+          "peer": true,
           "requires": {
             "inherits": "^2.0.3",
             "is-arguments": "^1.0.4",
@@ -20346,16 +20678,6 @@
         "graceful-fs": "^4.1.2"
       }
     },
-    "wcwidth": {
-      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/wcwidth/-/wcwidth-1.0.1.tgz",
-      "integrity": "sha512-XHPEwS0q6TaxcvG85+8EYkbiCux2XtWG2mkc47Ng2A77BQu9+DqIOJldST4HgPkuea7dvKSj5VgX3P1d4rW8Tg==",
-      "dev": true,
-      "optional": true,
-      "requires": {
-        "defaults": "^1.0.3"
-      }
-    },
     "webidl-conversions": {
       "version": "7.0.0",
       "dev": true
@@ -20547,7 +20869,9 @@
     },
     "xtend": {
       "version": "4.0.2",
-      "dev": true
+      "dev": true,
+      "optional": true,
+      "peer": true
     },
     "y18n": {
       "version": "5.0.8",
diff --git a/src/App.tsx b/src/App.tsx
index 0df674e2..f8af82b0 100644
--- a/src/App.tsx
+++ b/src/App.tsx
@@ -1,22 +1,27 @@
-import { createBrowserRouter, RouterProvider } from "react-router-dom";
-import CreateTest from "./app/home/pages/CreateTest";
-import ListTests from "./app/home/pages/ListTests";
+import { useEffect, useState } from "react";
 
-const router = createBrowserRouter([
-  {
-    path: "*",
-    Component: CreateTest,
-  },
-  {
-    path: "/create-test",
-    Component: CreateTest,
-  },
-  {
-    path: "/tests",
-    Component: ListTests,
-  },
-]);
+function App() {
+    const [message, setMessage] = useState("");
+    const [error, setError] = useState("");
 
-export default function App() {
-  return <RouterProvider router={router} fallbackElement={<p>Loading...</p>} />;
+    useEffect(() => {
+        fetch(import.meta.env.VITE_API_URL)
+            .then((response) => {
+                if (!response.ok) {
+                    throw new Error(`Erro na API: ${response.status}`);
+                }
+                return response.json();
+            })
+            .then((data) => setMessage(data.message))
+            .catch((error) => setError(error.message));
+    }, []);
+
+    return (
+        <div>
+            <h1>Testando API</h1>
+            {error ? <p style={{ color: "red" }}>{error}</p> : <p>{message}</p>}
+        </div>
+    );
 }
+
+export default App;
diff --git a/src/shared/services/ApiService.ts b/src/shared/services/ApiService.ts
index 3e94dded..b140f08e 100644
--- a/src/shared/services/ApiService.ts
+++ b/src/shared/services/ApiService.ts
@@ -116,3 +116,4 @@ export class ApiService {
     }
   }
 }
+